Upload
others
View
18
Download
0
Embed Size (px)
Citation preview
Research ArticleComparative Glycopeptide Analysis for ProteinGlycosylation by Liquid Chromatography and TandemMass Spectrometry Variation in Glycosylation Patterns ofSite-Directed Mutagenized Glycoprotein
Young Hye Hahm1 Sung Ho Hahm1 Hyoun Young Jo1 and Yeong Hee Ahn 2
1Rophibio Inc Cheongju 28160 Republic of Korea2Department of Biomedical Science Cheongju University Cheongju 28160 Republic of Korea
Correspondence should be addressed to Yeong Hee Ahn ahnyhcjuackr
Received 16 April 2018 Revised 29 June 2018 Accepted 31 July 2018 Published 2 September 2018
Academic Editor Gunther K Bonn
Copyright copy 2018 Young Hye Hahm et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited
Glycosylation is one of the most important posttranslational modifications for proteins including therapeutic antibodies andgreatly influences protein physiochemical properties In this study glycopeptidemapping of a reference and biosimilar recombinantantibodies (rAbs) was performed using liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI-MSMS) and an automated Glycoproteome Analyzer (GPA) algorithmThe tandemmass analyses for the reference and biosimilarsamples indicate that this approach proves to be highly efficient in reproducing consistent analytical results and discoveringthe implications of different rAb production methods on glycosylation patterns Furthermore the comparative analysis of amutagenized rAb glycoprotein proved that a single amino acid mutation in the Fc portion of the antibody molecule causedincreased variations in glycosylation patterns These variations were also detected by the mass spectrometry method efficientlyThis mapping method focusing on precise glycopeptide identification and comparison for the identified glycoforms can be usefulin differentiating aberrant glycosylation in biosimilar rAb products
1 Introduction
Glycosylation involves the covalent attachment of glycans tospecific amino acid residues on protein and is one of theimportant posttranslational protein modification processes[1] Glycosylation alters the properties of proteins includingpharmacokinetics effector functions solubility and stability[2] Aberrant glycosylation in glycoproteins has been relatedto the occurrence and progression of certain diseases [3]Careful observation of protein glycosylation is crucial for thedevelopment of stable and effective drugs Furthermore theyare crucial in comparing biosimilar products to referencedrugs as mandated by regulatory agencies [4] Changes inmanufacturing process conditions for biologics such as pro-cess optimization scale-up production and site changesmayimpact glycosylation patterns of the resulting recombinant
antibody (rAb) [5 6] As such glycosylation is considered acritical quality attribute (CQA) for rAb therapeutics as it hasthe potential to determinewhether the biosimilar candidate ishighly similar to the reference drugThe glycosylation patternof a rAb affects a wide spectrum of biological processesTherefore consistent glycosylation is necessary to prove thedrugrsquos ability to maintain safety and efficacy [7]
Mass spectrometry is a core technology in the field ofproteomics with high-throughput performance and accuratedigitalized informatics [8] One particular mass analysistechnique is matrix assisted laser desorptionionization massspectrometry (MALDI-MS) equipped with time-of-flight(TOF) analyzer [9ndash12] Another widely used method inmass spectrometry is liquid chromatography-electrosprayionization tandemmass spectrometry (LC-ESI-MSMS) [13ndash17] LC-ESI-MSMS is a robust high-throughput method
HindawiInternational Journal of Analytical ChemistryVolume 2018 Article ID 8605021 10 pageshttpsdoiorg10115520188605021
2 International Journal of Analytical Chemistry
in identifying and quantifying glycopeptides in enzymaticdigests from various proteomic samples High-throughputESI-MSMS for glycopeptides initially separated by LCallows for the identification and quantification of alldetectable features which in turn provides a more detailedaccount of protein glycosylation patterns Data analysis forlarge amounts of raw mass data resulting from LC-ESI-MSMS for the glycoproteome is a challenging task thereforevarious search engines have been developed such as Glyco-master DB [18] Byonic [19] MAGIC [20] and Glycopro-teome Analyzer (GPA) [21] GPA is capable of identifyingsite-specific N-glycopeptides efficiently and features the useof 3-top monoisotopic mass peak intensity of glycopeptides[21 22]The high-speedmapping of glycopeptides using GPAhas proven to display analytical efficiency with a false displayrate (FDR) le1
Here we have introduced a practicalmethod formappingand comparing the glycosylation patterns in rAb glycopro-teins in which glycopeptide samples prepared by in-solutionor in-gel protein digestion are analyzed by LC-ESI-MS-MSThe developed method is applied for comparative analysisof glycosylation patterns of biosimilar rAbs as well as amutagenized rAb glycoprotein
2 Experimental
21 Reagents and Chemicals The reference rAb glycopro-tein (Adalimumab commercially known as Humira) wasobtained fromGSamHospital (Gyeonggi-do Korea) HiTrapMabselect SuRe columns were purchased from GE Health-care and C
18trap column was purchased from Harvard
Apparatus (Holliston MA USA) Trypsin for protein diges-tion was obtained from Promega (Madison WI USA) 14-dithiothreitol (DTT) iodoacetamide (IAA) trifluoroaceticacid (TFA) and formic acid (FA) were purchased fromSigma Aldrich (St Louis MO USA) HPLC-grade waterand acetonitrile (ACN) were purchased from JT Baker(Phillipsburg NJ USA) CHO-k1 cells were purchased fromATCC (Manassas VA USA) and ExpiCHO-s cells werepurchased from Thermo Fisher Scientific (Waltham MAUSA)
22 Expression Vector for the Biosimilar The light chainand heavy chain genes for biosimilar rAb were synthesized(Genscript Piscataway NJ USA) and cloned into a customexpression vector pCPp2-CMV The expression vector con-tains both the heavy chain and the light chain genes forthe biosimilar rAb controlled by the human CMV promoterrespectively The vector also contains puromycin-resistancegene as a selectable marker under SV40 promoter
23 Transient Expression of Biosimilar rAbs For the tran-sient expression of biosimilar rAbs ExpiCHO-s cells weretransfected with the expression vector and maintained infed-batch culture following manufacturerrsquos protocol for maxtiter Briefly 50ml of Expi-CHO-s cells cultured in ExpiCHOculture medium in a 250 shaker flask was transfected with50 120583g of the expression vector DNA using Expifectamine
CHO reagent Transfected cells were maintained on shakerat 32∘C in a CO
2incubator and ExpiCHO enhancer and feed
were added as recommended Cells were monitored every 24hours for the cell growth and viability Cells were harvested7-10 days after transfection before cell viability drops below75 Cell supernatant was collected by centrifugation at3000 g for 30min and filtered through a vacuum filtrationunit (022 um) Filtered cell supernatant was subjected topurification for biosimilar rAbs by protein A affinity chro-matography using HiTrap Mabselect SuRe columns
24 Mutagenesis of Biosimilar rAb Generation A mutag-enized heavy chain gene containing tryptophan to valinesubstitution within the CH2 domain of the heavy chain (atposition C41 by IMGT codon numbering or 290 by Kabatnumbering) was synthesized and cloned into the expressionvector containing the light chain gene The newly clonedDNA was transfected into CHO-k1 cells and mutagenizedrAb was purified from cell supernatant using HiTrap Mab-select SuRe columns
25 Stable Expression of the Biosimilar rAbs after Cell LineGeneration Cell lines for the biosimilar rAbs were generatedusing suspension-adapted CHO-k1 cells CHO-k1 cells werepreadapted in serum-free suspension culture andmaintainedin CDM4CHO (Hyclone Little Chalfont UK) chemicallydefinedmedium 20 120583g of thewild-type expression vectorwasused to transfect 2 times 106 cells by electroporation followinga proprietary recombination-based transfection protocolTransfected cells were monitored daily and subject to mediachange and antibiotic selection with puromycin (8 ugmL)every 48 hours for 12-14 days Stably transfected cell pool wasthen used for the selection of high titer cell lines by FACScell sorting using MoFlo-XDP (Beckman Coulter Brea CAUSA) FACS cell sortingwas performed after staining the cellsusing R-PE conjugated F(abrsquo)
2fragment goat anti-human IgG
Fc120574 fragment specific antibody (Jackson ImmunoResearchLaboratories Inc West Grove PA USA) following theprotocol in Brezinsky et al (2003) [23] High titer cell lineswere sorted directly into 96-well plates containing 200120583l ofproprietary single cell growth medium per well based onthe R-PE fluorescence expression levels and after excludingdead cells and doublets Cell sorting was performed usingthe single cell sort mode to increase the efficiency sortingone cell per well Cells sorted into 96 well plates weremonitored for cell growth and clonality by scanning the plateusing Clone Select Imager (Molecular Devices San Jose CAUSA) Approximately 15 days after the sorting expressionlevels of the biosimilar rAb were determined with an ELISAassay and 20 clones with the highest expression levels wereexpanded for further analysis A final clone with the highestcell titer was selected from the 20 clones preserved inliquid nitrogen stocks and then used for the expression ofthe biosimilar in a fed-batch condition For the expressionof the biosimilar from the cell line cells were seeded in30ml of CDM4CHO with 8 120583gml puromycin and 800 120583l ofSheff Pulse II (Kerry Bioscience Ithaca NY USA) feedingmedium was added to the culture every 48 hours Cells were
International Journal of Analytical Chemistry 3
maintained at 32∘C for 10-12 days and harvested when cellviability dropped below 75 Biosimilar rAb was purifiedfrom the cell culture supernatant using HiTrap MabselectSuRe protein A columns
26 Protein Quantitation with Indirect ELISA Assay 96-well plates for ELISA assay were prepared by coating thewells with goat anti-human IgG whole molecule primaryantibody (5120583gmL Sigma Aldrich) and blocking with tris-buffered saline with BSA pH 8 Purified IgG from humanserum was used as a standard in serial dilutions in twofolds starting from 50 ng ml Biosimilar rAb samples werediluted so that the titer estimation can be made within thestandard curve range Standards and samples were added tothewells in duplicates After extensivewashing thewells wereprobed with goat anti-human IgG- (gamma-chain specific-)peroxidase antibody (120000 Sigma Aldrich) and colorizedwith 331015840551015840-Tetramethylbenzidine (TMB) liquid substrateColor change was induced with 1 HCl and the absorbancevalues were measured at 450 nm with VMax microplatereader (Molecular Devices)
27 Protein Characterization with SDS-PAGE and WesternBlotting Theprotein was characterized with sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) andWestern blotting Samples were reduced in a 1x samplereducing agent at 90-10∘C for 5minThe samples alongwith aprotein standard were loaded into a 4-20 Tris-PAG precastgel (Komabiotech Seoul Korea) with 1x Tris-Glycine SDSrunning buffer for 90min at a constant voltage of 120VThe resolved protein was either stained using Coomassieblue reagent overnight and destained with 50 ddH
2O 40
methanol and 10 acetic acid before imaging or transferredto a nitrocellulose membrane using the TransBlot-SemidryTransfer System (Bio-rad) for the Western blotting Themembranes were blocked in 5 NFDM and incubated ina primary antibody solution containing goat anti-humanIgG whole molecule antibody (1100000 Sigma Aldrich)The membrane was probed with rabbit anti-goat IgG HRP-conjugated antibody (120000 Sigma Aldrich) and detectedusing a LumiGLO chemiluminescent substrate kit (KPL)Therelative intensity of the protein bands was scanned usingFusionFx (Vilber Collegien France)
28 In-Solution Enzymatic Digestion 100 ug of protein sam-ple obtained from protein A affinity chromatography wasdenatured with 50mM ammonium bicarbonate (ABC) and2M urea The sample was reduced by 100mM DTT for 1 hat 35∘C After incubation the sample was alkylated with100mM IAA for 1 h in a darkroom and digested with trypsin(0125 uguL) overnight at 37∘C Digested sample was driedin a SpeedVac and stored in minus20∘C or reconstituted in 1FA for desalting SPE micro-spin column was equilibratedby centrifugation with wash buffer (99 H
2O 1 FA) and
elution buffer (80 ACN 1 FA) Aliquot of the trypticdigests was applied to the column washed twice and elutedtwicewith 30 uL elution bufferThe eluted peptide samplewasdried in a SpeedVac
29 In-Gel Enzymatic Digestion In-gel protein sampleexcised from SDS-PAGE was repeatedly dehydrated with 500uL ACN for 10min at room temperature The gel slices werereduced with 75 uL DTT for 30min at 56∘C alkylated withIAA for 20min in a darkroom and destained with 500 uL100mM ammonium bicarbonateacetonitrile (ABCACN)for 30min with periodic vortexing After dehydration thegel slices were rehydrated with 10mM ABC and 10 ACNbuffer containing trypsin and digested at 37∘C overnight 5FAACN (12 volvol) was added to the digested mixtureand incubated at 37∘C for 15min with slight vortex Thedigested sample was dried in a SpeedVac and either storedin minus20∘C or reconstituted in 01 FA for MSMS analysisAliquot of the tryptic digests was applied to SPE micro-spin column washed twice and eluted twice with 30 uLelution buffer The eluted peptide sample was dried in aSpeedVac
210 LC-ESI-MSMSAnalysis Tryptic peptide sample recon-stituted with water containing 01 formic acid was sepa-rated by a Nano Acquity UPLC system (Waters USA) 5uL aliquot of the peptide solution was applied to a C
18
trap column (id 180 um length 20mm and particle size5 um Waters USA) with an autosampler The peptides weredesalted in the trap column for 10min at a 5 uLminflow rate The trapped peptides were back-flushed into ahomemade C
18trap column (id 100 um length 200mm
and particle size 3 um) for separation Mobile phases Aand B were composed with 100 H
2O containing 01
FA and 100 ACN containing 01 FA respectively TheLC gradient began at 5 mobile phase B and was main-tained for 15min The gradient was ramped up to 15 Bfor 5min 50 B for 75min and 95 B for 1min 95B was maintained for 13min and decreased to 5 B for1min The column was finally re-equilibrated with 5 Bfor 10min LTQ Orbitrap Elite mass spectrometer (ThermoFisher Scientific) equipped with a nanoelectrospray sourcewas used to analyze the separated peptides The electrosprayvoltage was set to 22 kV and the LTQ Orbitrap Elite massspectrometer (Thermo Fisher Scientific) was operated indata-dependent mode during the chromatographic separa-tion The MS acquisition parameter for full-scan resolutionwas 60000 in the Orbitrap for each sample Data-dependentMSMS scans were acquired by collision-induced dissocia-tion (CID) and higher-energy collision dissociation (HCD)CID and HCD fragmentation scans were acquired in lineartrap quadrupole (LTQ) mode with a 30-ms activation timeand in Orbitrap at resolution 15000 with a 20-ms activationtime A 35 normalized collision energy (NCE) and 50Daisolation window were used for CID and HCD analysesPreviously fragmented ions were excluded for 300 s in allMSMS scans Raw mass data acquired were analyzed withthe automated Glycoproteome Analyzer (GPA) algorithmto identify glycopeptides and then the resulting identifiedglycopeptides were further confirmed manually with care-ful investigation in retention time chromatographic peakshape isotopicmass distribution pattern and fragmented ionmatching
4 International Journal of Analytical Chemistry
Table 1 List of tryptic glycopeptides from triplicate analysis of reference rAba
No Glycopeptides IonCharge (+)
DetectedMass (mz)
Run 1 Run 2 Run 3RT (min) MS Intensity RT (min) MS Intensity RT (min) MS Intensity
1 EEQYNSTYR 3 3 0 0 2 1142959 2770 92380184 2839 3270910 2840 10319768
2 EEQYNSTYR 3 3 1 0 2 1215989 2741 372193798 2831 72807858 2830 1887948683 810996
3 EEQYNSTYR 3 4 0 0 2 1244499 2770 91063963 2839 15496872 2842 154070823 830000
4 EEQYNSTYR 3 4 1 0 2 1318029 2740 1541540913 2828 398363214 2830 8154840543 878688
5 EEQYNSTYR 3 5 1 0 3 946380 2760 7447300 2832 5456578 2830 71994696 EEQYNSTYR 4 2 0 0 2 1122443 2760 126933115 2839 1113753 2835 67216327 EEQYNSTYR 4 3 0 1 3 913355 2828 188888 2888 2653427 2890 13007138 EEQYNSTYR 4 3 1 0 2 1297018 2740 50607743 2825 6312278 2821 20746171
9 EEQYNSTYR 4 3 1 1 2 1442557 2810 29760350 2881 55607576 2880 283562773 962043
10 EEQYNSTYR 4 4 1 0 2 1398554 2738 285767417 2825 58136525 2820 2056147173 932706
11 EEQYNSTYR 4 4 1 1 3 1029738 2810 5127014 2880 36633653 2885 1266473012 EEQYNSTYR 4 5 1 0 3 1000398 2754 2426554 2832 4640770 2830 6085787
13 EEQYNSTYR 5 2 0 0 2 1203473 2750 542316083 2828 16939021 2829 694627983 802652
14 EEQYNSTYR 5 4 1 0 3 986723 2736 3496656 2810 2860860 2810 1093770615 EEQYNSTYR 5 4 1 1 3 1083755 2810 8174096 2881 18861846 2880 2560383616 EEQYNSTYR 5 4 1 2 3 1180784 2852 1129042 2920 3397473 2929 118106017 EEQYNSTYR 6 2 0 0 2 1284497 2740 102132925 2818 733668 2821 1402451818 EEQYNSTYR 6 3 1 0 3 973049 2736 4256849 2818 638835 2820 472169919 EEQYNSTYR 6 3 1 1 3 1070081 2804 5703122 2874 18933930 2870 833695720 EEQYNSTYR 7 2 0 0 2 1365524 2740 55356342 2818 559700 2821 5262829aMS Intensity is based on the summation of 3-top monoisotopic mass peak intensity [21] If simultaneously detected MS Intensity values of 2+ and 3+ massions are summed
3 Results and Discussion
31 LC-ESI-MSMS Analysis of Reference rAb GlycoproteinThe reference rAb glycoprotein desalted with a HiTrapMabselect SuRe proteinA column (GEHealthcare) accordingto manufacturerrsquos protocols was quantified by UV spectrom-etry ELISA assay and BCA assay and characterized by SDS-PAGE and Western blotting The reference rAb glycoproteinsample desalted was reduced alkylated in a darkroom anddigested with trypsin overnight The digested sample wasdesalted with a SPE micro-spin column and analyzed byLC-ESI-MSMS coupled with CID and HCD fragmentationtechniques
Figure 1 shows a base peak chromatogram (BPC)obtained from the analysis of the tryptic digests of thereference rAb glycoprotein The obtained raw mass data wasanalyzed with GPA algorithm and the automatically assignedglycopeptides were further validated manually Amongthe identified glycopeptides EEQYNSTYR 3(hexose) 3(N-acetylhexosamine) 1(fucose) 0(sialic acid) was eluted atthe retention time (RT) 2825min The extracted ion
chromatogram of the monoisotopic peak pattern of theprecursor ion (2+) for glycopeptide EEQYNSTYR 3 3 1 0was illustrated in Figure 1 Tandem mass spectra of theglycopeptide EEQYNSTYR 3 3 1 0 obtained by CID andHCD fragmentation techniques were also shown in Figure 2We confirmed that fragment ions were well-matched withtheir structures
The analysis results of the reference rAb obtainedby triplicate sample preparation and mass analyses weresummarized in Table 1 LC-ESI-MSMSwith tryptic digestiondetected 20 consistent glycosylation patterns with similardistributionsThe glycosylation patterns normalized for eachrun were graphed together in Figure 3 We observed threemajor fucosylated glycopeptides EEQYNSTYR 3 3 1 0EEQYNSTYR 3 4 1 0 and EEQYNSTYR 4 4 1 0 which arereportedly predominant glycoforms found in recombinantantibodies produced in CHO cells [7 24 25] In additionhigh mannose forms with the relative ratios lt13 were alsoobservedmdashEEQYNSTYR 4 2 0 0 EEQYNSTYR 5 2 0 0EEQYNSTYR 6 2 0 0 and EEQYNSTYR 7 2 0 0 Theseglycosylation patterns have been frequently observed in rAb
International Journal of Analytical Chemistry 5
BPCRT 000 - 12000
36908129360
486486090813786
8129360
38018129359 4885
8609078
39195938265
41557140149 4940
860908045388384719 7289
92651156689
75807685170
9379667
54346036748
73319265115
74519270032
57096040029
282510541217 7671
9179984
26957800903 10429
812936281009480141
118258129346
84048725453
95047903803
21174351794
15984451174
4644451187
XIC
RT 2825AA 108570163BP 12159879
12164894
1215987912169907
12174919
1217993812184948
Mass peak(2+)
10 20 30 40 50 60 70 80 90 100 110 1200Time (min)
05
101520253035404550556065707580859095
100Re
lativ
e Abu
ndan
ce
Figure 1 Base peak chromatogram (BPC) by LC-tandemMS analysis of the tryptic digests of reference rAbThe extracted ion chromatogram(XIC) of the glycopeptide EEQYNSTYR 3 3 1 0 identified as one of major components at RT 2825min was also inserted together with themonoisotopic peak pattern of the glycopeptide
therapeutics due to the fact that recombinant rAbs are notfully modified and substantially affected by posttranslationalmodifications [26 27] After confirming the reliability andrepeatability in the analysis of the reference rAb glycoproteina comparative analysis was performed using biosimilar rAbsA and B which were prepared by two different methods ofproduction
32 Comparative Analysis of Glycosylation Patterns in Biosim-ilar rAbs Samples of biosimilar rAbs were obtained fromeither transiently transfected cells or a stably transfected cellline and referred to as biosimilar rAb A and B respectivelyfor clarification purposes The purified biosimilar rAbs Aand B were digested with trypsin and analyzed with LC-ESI-MSMS The obtained raw mass data was analyzed withGPA algorithm and the identified glycopeptides were furthervalidated manually The glycosylation patterns acquired forthe biosimilar rAb A and B were individually normalized andthemapping results were comparedwith those obtained fromthe analysis for the reference rAb sample (see Figure 4)
The major glycoforms for the biosimilar rAbA and rAb B as well as the reference rAb wereconsistentmdashEEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0 However we observed variationsin the minor glycosylation species not only between thereference rAb and biosimilar rAbs but also between thetwo biosimilar rAbs samples produced using differentmethods The biosimilar rAb A and B displayed a decreasedabundance for EEQYNSTYR 3 4 1 0 and an increased
abundance for EEQYNSTYR 4 4 1 0 with the addition ofgalactose in comparison to the reference rAbThe biosimilarrAb B displayed higher abundance for glycans with sialicacidsmdashEEQYNSTYR 4 4 1 1 EEQYNSTYR 5 4 1 1 andEEQYNSTYR 5 4 1 2mdashin comparison to the biosimilar rAbA These results suggest that LC-tandem mass spectrometrycan be a highly sensitive tool for distinguishing subtledifferences in minor glycosylation species
33 Glycosylation Patterns in Mutagenized rAb GlycoproteinWe tested the potential validity of using LC-tandem massspectrometry in the determination process for the similarityof biosimilar products to the referencematerials in biosimilarsamples with a single mutation in amino acid sequencesUnintentionalmutations that arise during protein productioncan cause changes in glycosylation patterns which in turncan impact the safety quality and potency of the recombinantrAbs [28] A tryptophan to valine amino acid mutation wascreated in the vicinity of the glycosylation site of Adalimumabto test its effect on the pattern of glycosylation to be detectedusing LC-tandemmass spectrometry Interestingly the singleamino acid mutation caused the inability for biosimilarrAb to bind to the protein A affinity column (data notshown) So instead of purifying the mutagenized rAb usingprotein A columns we performed the analysis using in-gelsamples Unpurified samples were analyzed by SDS-PAGEand the expected heavy chain and light chain bands wereobserved at 51 kDa and 26 kDa respectively (see Figure 5)The bands excised from the gel were subjected to in-gel
6 International Journal of Analytical Chemistry
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3240 RT 2830 AV 1 NL 279E5T ITMS + c NSI d Full ms2 121649cid3500 [32000-200000]
13926141
104155599606016
11431420
15386483
8795134
11144237
7985649
69015655281868
137569371596713013567330
192069853660828 150270741190462573132361758679212726949 19038585170177336743636
4393188 14856172
[1+]
[1+]
[1+]
[1+]
[1+]
[2+]
6968846
7690668
[2+]
~YNS~
[2+]
~YNS~~YNS~
]
YNS~YNNNNSS
~YNS~
[2+]
~YNS~
~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~
8709283
9526278
~YNS~[1+]
~YNS~
[1+]1+]
~YNS~
[1+]10336189
10620417
17419075
~YNS~ ~YNS~ ~YNS~YNS YNS YNS YNS YNS ~YNS~
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000400mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
-1(2O
-2(2O
[2+] [2+][2+] [2+] [2+] [2+] [2+]
[2+]
[1+] [1+][1+] [1+]
[1+] [1+] [1+]
(a) MSMS spectrum EEQYNSTYR 3 3 1 0 CID
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3241 RT 2830 AV 1 NL 518E5T FTMS + c NSI d Full ms2 121649hcd3500 [10000-200000]
1380549
2040865
1392670310537871
1272611115397529 19199423
131855215814420 1189562164186744677113 14736592 195405772872824 102980527817769
~YNS~ ~YNS~~YNS~
[1+] [1+]
[1+]
1260550
1680660
1860761
-HO
-2HO
-HO - Acetyl-HO-CHO
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000100mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
S9[1+]
(b) MSMS spectrum EEQYNSTYR 3 3 1 0 HCD
Figure 2 Tandem mass spectra of the glycopeptide EEQYNSTYR 3 3 1 0 obtained by (a) collision-induced dissociation and (b) higher-energy collision dissociation
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
2 International Journal of Analytical Chemistry
in identifying and quantifying glycopeptides in enzymaticdigests from various proteomic samples High-throughputESI-MSMS for glycopeptides initially separated by LCallows for the identification and quantification of alldetectable features which in turn provides a more detailedaccount of protein glycosylation patterns Data analysis forlarge amounts of raw mass data resulting from LC-ESI-MSMS for the glycoproteome is a challenging task thereforevarious search engines have been developed such as Glyco-master DB [18] Byonic [19] MAGIC [20] and Glycopro-teome Analyzer (GPA) [21] GPA is capable of identifyingsite-specific N-glycopeptides efficiently and features the useof 3-top monoisotopic mass peak intensity of glycopeptides[21 22]The high-speedmapping of glycopeptides using GPAhas proven to display analytical efficiency with a false displayrate (FDR) le1
Here we have introduced a practicalmethod formappingand comparing the glycosylation patterns in rAb glycopro-teins in which glycopeptide samples prepared by in-solutionor in-gel protein digestion are analyzed by LC-ESI-MS-MSThe developed method is applied for comparative analysisof glycosylation patterns of biosimilar rAbs as well as amutagenized rAb glycoprotein
2 Experimental
21 Reagents and Chemicals The reference rAb glycopro-tein (Adalimumab commercially known as Humira) wasobtained fromGSamHospital (Gyeonggi-do Korea) HiTrapMabselect SuRe columns were purchased from GE Health-care and C
18trap column was purchased from Harvard
Apparatus (Holliston MA USA) Trypsin for protein diges-tion was obtained from Promega (Madison WI USA) 14-dithiothreitol (DTT) iodoacetamide (IAA) trifluoroaceticacid (TFA) and formic acid (FA) were purchased fromSigma Aldrich (St Louis MO USA) HPLC-grade waterand acetonitrile (ACN) were purchased from JT Baker(Phillipsburg NJ USA) CHO-k1 cells were purchased fromATCC (Manassas VA USA) and ExpiCHO-s cells werepurchased from Thermo Fisher Scientific (Waltham MAUSA)
22 Expression Vector for the Biosimilar The light chainand heavy chain genes for biosimilar rAb were synthesized(Genscript Piscataway NJ USA) and cloned into a customexpression vector pCPp2-CMV The expression vector con-tains both the heavy chain and the light chain genes forthe biosimilar rAb controlled by the human CMV promoterrespectively The vector also contains puromycin-resistancegene as a selectable marker under SV40 promoter
23 Transient Expression of Biosimilar rAbs For the tran-sient expression of biosimilar rAbs ExpiCHO-s cells weretransfected with the expression vector and maintained infed-batch culture following manufacturerrsquos protocol for maxtiter Briefly 50ml of Expi-CHO-s cells cultured in ExpiCHOculture medium in a 250 shaker flask was transfected with50 120583g of the expression vector DNA using Expifectamine
CHO reagent Transfected cells were maintained on shakerat 32∘C in a CO
2incubator and ExpiCHO enhancer and feed
were added as recommended Cells were monitored every 24hours for the cell growth and viability Cells were harvested7-10 days after transfection before cell viability drops below75 Cell supernatant was collected by centrifugation at3000 g for 30min and filtered through a vacuum filtrationunit (022 um) Filtered cell supernatant was subjected topurification for biosimilar rAbs by protein A affinity chro-matography using HiTrap Mabselect SuRe columns
24 Mutagenesis of Biosimilar rAb Generation A mutag-enized heavy chain gene containing tryptophan to valinesubstitution within the CH2 domain of the heavy chain (atposition C41 by IMGT codon numbering or 290 by Kabatnumbering) was synthesized and cloned into the expressionvector containing the light chain gene The newly clonedDNA was transfected into CHO-k1 cells and mutagenizedrAb was purified from cell supernatant using HiTrap Mab-select SuRe columns
25 Stable Expression of the Biosimilar rAbs after Cell LineGeneration Cell lines for the biosimilar rAbs were generatedusing suspension-adapted CHO-k1 cells CHO-k1 cells werepreadapted in serum-free suspension culture andmaintainedin CDM4CHO (Hyclone Little Chalfont UK) chemicallydefinedmedium 20 120583g of thewild-type expression vectorwasused to transfect 2 times 106 cells by electroporation followinga proprietary recombination-based transfection protocolTransfected cells were monitored daily and subject to mediachange and antibiotic selection with puromycin (8 ugmL)every 48 hours for 12-14 days Stably transfected cell pool wasthen used for the selection of high titer cell lines by FACScell sorting using MoFlo-XDP (Beckman Coulter Brea CAUSA) FACS cell sortingwas performed after staining the cellsusing R-PE conjugated F(abrsquo)
2fragment goat anti-human IgG
Fc120574 fragment specific antibody (Jackson ImmunoResearchLaboratories Inc West Grove PA USA) following theprotocol in Brezinsky et al (2003) [23] High titer cell lineswere sorted directly into 96-well plates containing 200120583l ofproprietary single cell growth medium per well based onthe R-PE fluorescence expression levels and after excludingdead cells and doublets Cell sorting was performed usingthe single cell sort mode to increase the efficiency sortingone cell per well Cells sorted into 96 well plates weremonitored for cell growth and clonality by scanning the plateusing Clone Select Imager (Molecular Devices San Jose CAUSA) Approximately 15 days after the sorting expressionlevels of the biosimilar rAb were determined with an ELISAassay and 20 clones with the highest expression levels wereexpanded for further analysis A final clone with the highestcell titer was selected from the 20 clones preserved inliquid nitrogen stocks and then used for the expression ofthe biosimilar in a fed-batch condition For the expressionof the biosimilar from the cell line cells were seeded in30ml of CDM4CHO with 8 120583gml puromycin and 800 120583l ofSheff Pulse II (Kerry Bioscience Ithaca NY USA) feedingmedium was added to the culture every 48 hours Cells were
International Journal of Analytical Chemistry 3
maintained at 32∘C for 10-12 days and harvested when cellviability dropped below 75 Biosimilar rAb was purifiedfrom the cell culture supernatant using HiTrap MabselectSuRe protein A columns
26 Protein Quantitation with Indirect ELISA Assay 96-well plates for ELISA assay were prepared by coating thewells with goat anti-human IgG whole molecule primaryantibody (5120583gmL Sigma Aldrich) and blocking with tris-buffered saline with BSA pH 8 Purified IgG from humanserum was used as a standard in serial dilutions in twofolds starting from 50 ng ml Biosimilar rAb samples werediluted so that the titer estimation can be made within thestandard curve range Standards and samples were added tothewells in duplicates After extensivewashing thewells wereprobed with goat anti-human IgG- (gamma-chain specific-)peroxidase antibody (120000 Sigma Aldrich) and colorizedwith 331015840551015840-Tetramethylbenzidine (TMB) liquid substrateColor change was induced with 1 HCl and the absorbancevalues were measured at 450 nm with VMax microplatereader (Molecular Devices)
27 Protein Characterization with SDS-PAGE and WesternBlotting Theprotein was characterized with sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) andWestern blotting Samples were reduced in a 1x samplereducing agent at 90-10∘C for 5minThe samples alongwith aprotein standard were loaded into a 4-20 Tris-PAG precastgel (Komabiotech Seoul Korea) with 1x Tris-Glycine SDSrunning buffer for 90min at a constant voltage of 120VThe resolved protein was either stained using Coomassieblue reagent overnight and destained with 50 ddH
2O 40
methanol and 10 acetic acid before imaging or transferredto a nitrocellulose membrane using the TransBlot-SemidryTransfer System (Bio-rad) for the Western blotting Themembranes were blocked in 5 NFDM and incubated ina primary antibody solution containing goat anti-humanIgG whole molecule antibody (1100000 Sigma Aldrich)The membrane was probed with rabbit anti-goat IgG HRP-conjugated antibody (120000 Sigma Aldrich) and detectedusing a LumiGLO chemiluminescent substrate kit (KPL)Therelative intensity of the protein bands was scanned usingFusionFx (Vilber Collegien France)
28 In-Solution Enzymatic Digestion 100 ug of protein sam-ple obtained from protein A affinity chromatography wasdenatured with 50mM ammonium bicarbonate (ABC) and2M urea The sample was reduced by 100mM DTT for 1 hat 35∘C After incubation the sample was alkylated with100mM IAA for 1 h in a darkroom and digested with trypsin(0125 uguL) overnight at 37∘C Digested sample was driedin a SpeedVac and stored in minus20∘C or reconstituted in 1FA for desalting SPE micro-spin column was equilibratedby centrifugation with wash buffer (99 H
2O 1 FA) and
elution buffer (80 ACN 1 FA) Aliquot of the trypticdigests was applied to the column washed twice and elutedtwicewith 30 uL elution bufferThe eluted peptide samplewasdried in a SpeedVac
29 In-Gel Enzymatic Digestion In-gel protein sampleexcised from SDS-PAGE was repeatedly dehydrated with 500uL ACN for 10min at room temperature The gel slices werereduced with 75 uL DTT for 30min at 56∘C alkylated withIAA for 20min in a darkroom and destained with 500 uL100mM ammonium bicarbonateacetonitrile (ABCACN)for 30min with periodic vortexing After dehydration thegel slices were rehydrated with 10mM ABC and 10 ACNbuffer containing trypsin and digested at 37∘C overnight 5FAACN (12 volvol) was added to the digested mixtureand incubated at 37∘C for 15min with slight vortex Thedigested sample was dried in a SpeedVac and either storedin minus20∘C or reconstituted in 01 FA for MSMS analysisAliquot of the tryptic digests was applied to SPE micro-spin column washed twice and eluted twice with 30 uLelution buffer The eluted peptide sample was dried in aSpeedVac
210 LC-ESI-MSMSAnalysis Tryptic peptide sample recon-stituted with water containing 01 formic acid was sepa-rated by a Nano Acquity UPLC system (Waters USA) 5uL aliquot of the peptide solution was applied to a C
18
trap column (id 180 um length 20mm and particle size5 um Waters USA) with an autosampler The peptides weredesalted in the trap column for 10min at a 5 uLminflow rate The trapped peptides were back-flushed into ahomemade C
18trap column (id 100 um length 200mm
and particle size 3 um) for separation Mobile phases Aand B were composed with 100 H
2O containing 01
FA and 100 ACN containing 01 FA respectively TheLC gradient began at 5 mobile phase B and was main-tained for 15min The gradient was ramped up to 15 Bfor 5min 50 B for 75min and 95 B for 1min 95B was maintained for 13min and decreased to 5 B for1min The column was finally re-equilibrated with 5 Bfor 10min LTQ Orbitrap Elite mass spectrometer (ThermoFisher Scientific) equipped with a nanoelectrospray sourcewas used to analyze the separated peptides The electrosprayvoltage was set to 22 kV and the LTQ Orbitrap Elite massspectrometer (Thermo Fisher Scientific) was operated indata-dependent mode during the chromatographic separa-tion The MS acquisition parameter for full-scan resolutionwas 60000 in the Orbitrap for each sample Data-dependentMSMS scans were acquired by collision-induced dissocia-tion (CID) and higher-energy collision dissociation (HCD)CID and HCD fragmentation scans were acquired in lineartrap quadrupole (LTQ) mode with a 30-ms activation timeand in Orbitrap at resolution 15000 with a 20-ms activationtime A 35 normalized collision energy (NCE) and 50Daisolation window were used for CID and HCD analysesPreviously fragmented ions were excluded for 300 s in allMSMS scans Raw mass data acquired were analyzed withthe automated Glycoproteome Analyzer (GPA) algorithmto identify glycopeptides and then the resulting identifiedglycopeptides were further confirmed manually with care-ful investigation in retention time chromatographic peakshape isotopicmass distribution pattern and fragmented ionmatching
4 International Journal of Analytical Chemistry
Table 1 List of tryptic glycopeptides from triplicate analysis of reference rAba
No Glycopeptides IonCharge (+)
DetectedMass (mz)
Run 1 Run 2 Run 3RT (min) MS Intensity RT (min) MS Intensity RT (min) MS Intensity
1 EEQYNSTYR 3 3 0 0 2 1142959 2770 92380184 2839 3270910 2840 10319768
2 EEQYNSTYR 3 3 1 0 2 1215989 2741 372193798 2831 72807858 2830 1887948683 810996
3 EEQYNSTYR 3 4 0 0 2 1244499 2770 91063963 2839 15496872 2842 154070823 830000
4 EEQYNSTYR 3 4 1 0 2 1318029 2740 1541540913 2828 398363214 2830 8154840543 878688
5 EEQYNSTYR 3 5 1 0 3 946380 2760 7447300 2832 5456578 2830 71994696 EEQYNSTYR 4 2 0 0 2 1122443 2760 126933115 2839 1113753 2835 67216327 EEQYNSTYR 4 3 0 1 3 913355 2828 188888 2888 2653427 2890 13007138 EEQYNSTYR 4 3 1 0 2 1297018 2740 50607743 2825 6312278 2821 20746171
9 EEQYNSTYR 4 3 1 1 2 1442557 2810 29760350 2881 55607576 2880 283562773 962043
10 EEQYNSTYR 4 4 1 0 2 1398554 2738 285767417 2825 58136525 2820 2056147173 932706
11 EEQYNSTYR 4 4 1 1 3 1029738 2810 5127014 2880 36633653 2885 1266473012 EEQYNSTYR 4 5 1 0 3 1000398 2754 2426554 2832 4640770 2830 6085787
13 EEQYNSTYR 5 2 0 0 2 1203473 2750 542316083 2828 16939021 2829 694627983 802652
14 EEQYNSTYR 5 4 1 0 3 986723 2736 3496656 2810 2860860 2810 1093770615 EEQYNSTYR 5 4 1 1 3 1083755 2810 8174096 2881 18861846 2880 2560383616 EEQYNSTYR 5 4 1 2 3 1180784 2852 1129042 2920 3397473 2929 118106017 EEQYNSTYR 6 2 0 0 2 1284497 2740 102132925 2818 733668 2821 1402451818 EEQYNSTYR 6 3 1 0 3 973049 2736 4256849 2818 638835 2820 472169919 EEQYNSTYR 6 3 1 1 3 1070081 2804 5703122 2874 18933930 2870 833695720 EEQYNSTYR 7 2 0 0 2 1365524 2740 55356342 2818 559700 2821 5262829aMS Intensity is based on the summation of 3-top monoisotopic mass peak intensity [21] If simultaneously detected MS Intensity values of 2+ and 3+ massions are summed
3 Results and Discussion
31 LC-ESI-MSMS Analysis of Reference rAb GlycoproteinThe reference rAb glycoprotein desalted with a HiTrapMabselect SuRe proteinA column (GEHealthcare) accordingto manufacturerrsquos protocols was quantified by UV spectrom-etry ELISA assay and BCA assay and characterized by SDS-PAGE and Western blotting The reference rAb glycoproteinsample desalted was reduced alkylated in a darkroom anddigested with trypsin overnight The digested sample wasdesalted with a SPE micro-spin column and analyzed byLC-ESI-MSMS coupled with CID and HCD fragmentationtechniques
Figure 1 shows a base peak chromatogram (BPC)obtained from the analysis of the tryptic digests of thereference rAb glycoprotein The obtained raw mass data wasanalyzed with GPA algorithm and the automatically assignedglycopeptides were further validated manually Amongthe identified glycopeptides EEQYNSTYR 3(hexose) 3(N-acetylhexosamine) 1(fucose) 0(sialic acid) was eluted atthe retention time (RT) 2825min The extracted ion
chromatogram of the monoisotopic peak pattern of theprecursor ion (2+) for glycopeptide EEQYNSTYR 3 3 1 0was illustrated in Figure 1 Tandem mass spectra of theglycopeptide EEQYNSTYR 3 3 1 0 obtained by CID andHCD fragmentation techniques were also shown in Figure 2We confirmed that fragment ions were well-matched withtheir structures
The analysis results of the reference rAb obtainedby triplicate sample preparation and mass analyses weresummarized in Table 1 LC-ESI-MSMSwith tryptic digestiondetected 20 consistent glycosylation patterns with similardistributionsThe glycosylation patterns normalized for eachrun were graphed together in Figure 3 We observed threemajor fucosylated glycopeptides EEQYNSTYR 3 3 1 0EEQYNSTYR 3 4 1 0 and EEQYNSTYR 4 4 1 0 which arereportedly predominant glycoforms found in recombinantantibodies produced in CHO cells [7 24 25] In additionhigh mannose forms with the relative ratios lt13 were alsoobservedmdashEEQYNSTYR 4 2 0 0 EEQYNSTYR 5 2 0 0EEQYNSTYR 6 2 0 0 and EEQYNSTYR 7 2 0 0 Theseglycosylation patterns have been frequently observed in rAb
International Journal of Analytical Chemistry 5
BPCRT 000 - 12000
36908129360
486486090813786
8129360
38018129359 4885
8609078
39195938265
41557140149 4940
860908045388384719 7289
92651156689
75807685170
9379667
54346036748
73319265115
74519270032
57096040029
282510541217 7671
9179984
26957800903 10429
812936281009480141
118258129346
84048725453
95047903803
21174351794
15984451174
4644451187
XIC
RT 2825AA 108570163BP 12159879
12164894
1215987912169907
12174919
1217993812184948
Mass peak(2+)
10 20 30 40 50 60 70 80 90 100 110 1200Time (min)
05
101520253035404550556065707580859095
100Re
lativ
e Abu
ndan
ce
Figure 1 Base peak chromatogram (BPC) by LC-tandemMS analysis of the tryptic digests of reference rAbThe extracted ion chromatogram(XIC) of the glycopeptide EEQYNSTYR 3 3 1 0 identified as one of major components at RT 2825min was also inserted together with themonoisotopic peak pattern of the glycopeptide
therapeutics due to the fact that recombinant rAbs are notfully modified and substantially affected by posttranslationalmodifications [26 27] After confirming the reliability andrepeatability in the analysis of the reference rAb glycoproteina comparative analysis was performed using biosimilar rAbsA and B which were prepared by two different methods ofproduction
32 Comparative Analysis of Glycosylation Patterns in Biosim-ilar rAbs Samples of biosimilar rAbs were obtained fromeither transiently transfected cells or a stably transfected cellline and referred to as biosimilar rAb A and B respectivelyfor clarification purposes The purified biosimilar rAbs Aand B were digested with trypsin and analyzed with LC-ESI-MSMS The obtained raw mass data was analyzed withGPA algorithm and the identified glycopeptides were furthervalidated manually The glycosylation patterns acquired forthe biosimilar rAb A and B were individually normalized andthemapping results were comparedwith those obtained fromthe analysis for the reference rAb sample (see Figure 4)
The major glycoforms for the biosimilar rAbA and rAb B as well as the reference rAb wereconsistentmdashEEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0 However we observed variationsin the minor glycosylation species not only between thereference rAb and biosimilar rAbs but also between thetwo biosimilar rAbs samples produced using differentmethods The biosimilar rAb A and B displayed a decreasedabundance for EEQYNSTYR 3 4 1 0 and an increased
abundance for EEQYNSTYR 4 4 1 0 with the addition ofgalactose in comparison to the reference rAbThe biosimilarrAb B displayed higher abundance for glycans with sialicacidsmdashEEQYNSTYR 4 4 1 1 EEQYNSTYR 5 4 1 1 andEEQYNSTYR 5 4 1 2mdashin comparison to the biosimilar rAbA These results suggest that LC-tandem mass spectrometrycan be a highly sensitive tool for distinguishing subtledifferences in minor glycosylation species
33 Glycosylation Patterns in Mutagenized rAb GlycoproteinWe tested the potential validity of using LC-tandem massspectrometry in the determination process for the similarityof biosimilar products to the referencematerials in biosimilarsamples with a single mutation in amino acid sequencesUnintentionalmutations that arise during protein productioncan cause changes in glycosylation patterns which in turncan impact the safety quality and potency of the recombinantrAbs [28] A tryptophan to valine amino acid mutation wascreated in the vicinity of the glycosylation site of Adalimumabto test its effect on the pattern of glycosylation to be detectedusing LC-tandemmass spectrometry Interestingly the singleamino acid mutation caused the inability for biosimilarrAb to bind to the protein A affinity column (data notshown) So instead of purifying the mutagenized rAb usingprotein A columns we performed the analysis using in-gelsamples Unpurified samples were analyzed by SDS-PAGEand the expected heavy chain and light chain bands wereobserved at 51 kDa and 26 kDa respectively (see Figure 5)The bands excised from the gel were subjected to in-gel
6 International Journal of Analytical Chemistry
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3240 RT 2830 AV 1 NL 279E5T ITMS + c NSI d Full ms2 121649cid3500 [32000-200000]
13926141
104155599606016
11431420
15386483
8795134
11144237
7985649
69015655281868
137569371596713013567330
192069853660828 150270741190462573132361758679212726949 19038585170177336743636
4393188 14856172
[1+]
[1+]
[1+]
[1+]
[1+]
[2+]
6968846
7690668
[2+]
~YNS~
[2+]
~YNS~~YNS~
]
YNS~YNNNNSS
~YNS~
[2+]
~YNS~
~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~
8709283
9526278
~YNS~[1+]
~YNS~
[1+]1+]
~YNS~
[1+]10336189
10620417
17419075
~YNS~ ~YNS~ ~YNS~YNS YNS YNS YNS YNS ~YNS~
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000400mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
-1(2O
-2(2O
[2+] [2+][2+] [2+] [2+] [2+] [2+]
[2+]
[1+] [1+][1+] [1+]
[1+] [1+] [1+]
(a) MSMS spectrum EEQYNSTYR 3 3 1 0 CID
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3241 RT 2830 AV 1 NL 518E5T FTMS + c NSI d Full ms2 121649hcd3500 [10000-200000]
1380549
2040865
1392670310537871
1272611115397529 19199423
131855215814420 1189562164186744677113 14736592 195405772872824 102980527817769
~YNS~ ~YNS~~YNS~
[1+] [1+]
[1+]
1260550
1680660
1860761
-HO
-2HO
-HO - Acetyl-HO-CHO
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000100mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
S9[1+]
(b) MSMS spectrum EEQYNSTYR 3 3 1 0 HCD
Figure 2 Tandem mass spectra of the glycopeptide EEQYNSTYR 3 3 1 0 obtained by (a) collision-induced dissociation and (b) higher-energy collision dissociation
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
International Journal of Analytical Chemistry 3
maintained at 32∘C for 10-12 days and harvested when cellviability dropped below 75 Biosimilar rAb was purifiedfrom the cell culture supernatant using HiTrap MabselectSuRe protein A columns
26 Protein Quantitation with Indirect ELISA Assay 96-well plates for ELISA assay were prepared by coating thewells with goat anti-human IgG whole molecule primaryantibody (5120583gmL Sigma Aldrich) and blocking with tris-buffered saline with BSA pH 8 Purified IgG from humanserum was used as a standard in serial dilutions in twofolds starting from 50 ng ml Biosimilar rAb samples werediluted so that the titer estimation can be made within thestandard curve range Standards and samples were added tothewells in duplicates After extensivewashing thewells wereprobed with goat anti-human IgG- (gamma-chain specific-)peroxidase antibody (120000 Sigma Aldrich) and colorizedwith 331015840551015840-Tetramethylbenzidine (TMB) liquid substrateColor change was induced with 1 HCl and the absorbancevalues were measured at 450 nm with VMax microplatereader (Molecular Devices)
27 Protein Characterization with SDS-PAGE and WesternBlotting Theprotein was characterized with sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) andWestern blotting Samples were reduced in a 1x samplereducing agent at 90-10∘C for 5minThe samples alongwith aprotein standard were loaded into a 4-20 Tris-PAG precastgel (Komabiotech Seoul Korea) with 1x Tris-Glycine SDSrunning buffer for 90min at a constant voltage of 120VThe resolved protein was either stained using Coomassieblue reagent overnight and destained with 50 ddH
2O 40
methanol and 10 acetic acid before imaging or transferredto a nitrocellulose membrane using the TransBlot-SemidryTransfer System (Bio-rad) for the Western blotting Themembranes were blocked in 5 NFDM and incubated ina primary antibody solution containing goat anti-humanIgG whole molecule antibody (1100000 Sigma Aldrich)The membrane was probed with rabbit anti-goat IgG HRP-conjugated antibody (120000 Sigma Aldrich) and detectedusing a LumiGLO chemiluminescent substrate kit (KPL)Therelative intensity of the protein bands was scanned usingFusionFx (Vilber Collegien France)
28 In-Solution Enzymatic Digestion 100 ug of protein sam-ple obtained from protein A affinity chromatography wasdenatured with 50mM ammonium bicarbonate (ABC) and2M urea The sample was reduced by 100mM DTT for 1 hat 35∘C After incubation the sample was alkylated with100mM IAA for 1 h in a darkroom and digested with trypsin(0125 uguL) overnight at 37∘C Digested sample was driedin a SpeedVac and stored in minus20∘C or reconstituted in 1FA for desalting SPE micro-spin column was equilibratedby centrifugation with wash buffer (99 H
2O 1 FA) and
elution buffer (80 ACN 1 FA) Aliquot of the trypticdigests was applied to the column washed twice and elutedtwicewith 30 uL elution bufferThe eluted peptide samplewasdried in a SpeedVac
29 In-Gel Enzymatic Digestion In-gel protein sampleexcised from SDS-PAGE was repeatedly dehydrated with 500uL ACN for 10min at room temperature The gel slices werereduced with 75 uL DTT for 30min at 56∘C alkylated withIAA for 20min in a darkroom and destained with 500 uL100mM ammonium bicarbonateacetonitrile (ABCACN)for 30min with periodic vortexing After dehydration thegel slices were rehydrated with 10mM ABC and 10 ACNbuffer containing trypsin and digested at 37∘C overnight 5FAACN (12 volvol) was added to the digested mixtureand incubated at 37∘C for 15min with slight vortex Thedigested sample was dried in a SpeedVac and either storedin minus20∘C or reconstituted in 01 FA for MSMS analysisAliquot of the tryptic digests was applied to SPE micro-spin column washed twice and eluted twice with 30 uLelution buffer The eluted peptide sample was dried in aSpeedVac
210 LC-ESI-MSMSAnalysis Tryptic peptide sample recon-stituted with water containing 01 formic acid was sepa-rated by a Nano Acquity UPLC system (Waters USA) 5uL aliquot of the peptide solution was applied to a C
18
trap column (id 180 um length 20mm and particle size5 um Waters USA) with an autosampler The peptides weredesalted in the trap column for 10min at a 5 uLminflow rate The trapped peptides were back-flushed into ahomemade C
18trap column (id 100 um length 200mm
and particle size 3 um) for separation Mobile phases Aand B were composed with 100 H
2O containing 01
FA and 100 ACN containing 01 FA respectively TheLC gradient began at 5 mobile phase B and was main-tained for 15min The gradient was ramped up to 15 Bfor 5min 50 B for 75min and 95 B for 1min 95B was maintained for 13min and decreased to 5 B for1min The column was finally re-equilibrated with 5 Bfor 10min LTQ Orbitrap Elite mass spectrometer (ThermoFisher Scientific) equipped with a nanoelectrospray sourcewas used to analyze the separated peptides The electrosprayvoltage was set to 22 kV and the LTQ Orbitrap Elite massspectrometer (Thermo Fisher Scientific) was operated indata-dependent mode during the chromatographic separa-tion The MS acquisition parameter for full-scan resolutionwas 60000 in the Orbitrap for each sample Data-dependentMSMS scans were acquired by collision-induced dissocia-tion (CID) and higher-energy collision dissociation (HCD)CID and HCD fragmentation scans were acquired in lineartrap quadrupole (LTQ) mode with a 30-ms activation timeand in Orbitrap at resolution 15000 with a 20-ms activationtime A 35 normalized collision energy (NCE) and 50Daisolation window were used for CID and HCD analysesPreviously fragmented ions were excluded for 300 s in allMSMS scans Raw mass data acquired were analyzed withthe automated Glycoproteome Analyzer (GPA) algorithmto identify glycopeptides and then the resulting identifiedglycopeptides were further confirmed manually with care-ful investigation in retention time chromatographic peakshape isotopicmass distribution pattern and fragmented ionmatching
4 International Journal of Analytical Chemistry
Table 1 List of tryptic glycopeptides from triplicate analysis of reference rAba
No Glycopeptides IonCharge (+)
DetectedMass (mz)
Run 1 Run 2 Run 3RT (min) MS Intensity RT (min) MS Intensity RT (min) MS Intensity
1 EEQYNSTYR 3 3 0 0 2 1142959 2770 92380184 2839 3270910 2840 10319768
2 EEQYNSTYR 3 3 1 0 2 1215989 2741 372193798 2831 72807858 2830 1887948683 810996
3 EEQYNSTYR 3 4 0 0 2 1244499 2770 91063963 2839 15496872 2842 154070823 830000
4 EEQYNSTYR 3 4 1 0 2 1318029 2740 1541540913 2828 398363214 2830 8154840543 878688
5 EEQYNSTYR 3 5 1 0 3 946380 2760 7447300 2832 5456578 2830 71994696 EEQYNSTYR 4 2 0 0 2 1122443 2760 126933115 2839 1113753 2835 67216327 EEQYNSTYR 4 3 0 1 3 913355 2828 188888 2888 2653427 2890 13007138 EEQYNSTYR 4 3 1 0 2 1297018 2740 50607743 2825 6312278 2821 20746171
9 EEQYNSTYR 4 3 1 1 2 1442557 2810 29760350 2881 55607576 2880 283562773 962043
10 EEQYNSTYR 4 4 1 0 2 1398554 2738 285767417 2825 58136525 2820 2056147173 932706
11 EEQYNSTYR 4 4 1 1 3 1029738 2810 5127014 2880 36633653 2885 1266473012 EEQYNSTYR 4 5 1 0 3 1000398 2754 2426554 2832 4640770 2830 6085787
13 EEQYNSTYR 5 2 0 0 2 1203473 2750 542316083 2828 16939021 2829 694627983 802652
14 EEQYNSTYR 5 4 1 0 3 986723 2736 3496656 2810 2860860 2810 1093770615 EEQYNSTYR 5 4 1 1 3 1083755 2810 8174096 2881 18861846 2880 2560383616 EEQYNSTYR 5 4 1 2 3 1180784 2852 1129042 2920 3397473 2929 118106017 EEQYNSTYR 6 2 0 0 2 1284497 2740 102132925 2818 733668 2821 1402451818 EEQYNSTYR 6 3 1 0 3 973049 2736 4256849 2818 638835 2820 472169919 EEQYNSTYR 6 3 1 1 3 1070081 2804 5703122 2874 18933930 2870 833695720 EEQYNSTYR 7 2 0 0 2 1365524 2740 55356342 2818 559700 2821 5262829aMS Intensity is based on the summation of 3-top monoisotopic mass peak intensity [21] If simultaneously detected MS Intensity values of 2+ and 3+ massions are summed
3 Results and Discussion
31 LC-ESI-MSMS Analysis of Reference rAb GlycoproteinThe reference rAb glycoprotein desalted with a HiTrapMabselect SuRe proteinA column (GEHealthcare) accordingto manufacturerrsquos protocols was quantified by UV spectrom-etry ELISA assay and BCA assay and characterized by SDS-PAGE and Western blotting The reference rAb glycoproteinsample desalted was reduced alkylated in a darkroom anddigested with trypsin overnight The digested sample wasdesalted with a SPE micro-spin column and analyzed byLC-ESI-MSMS coupled with CID and HCD fragmentationtechniques
Figure 1 shows a base peak chromatogram (BPC)obtained from the analysis of the tryptic digests of thereference rAb glycoprotein The obtained raw mass data wasanalyzed with GPA algorithm and the automatically assignedglycopeptides were further validated manually Amongthe identified glycopeptides EEQYNSTYR 3(hexose) 3(N-acetylhexosamine) 1(fucose) 0(sialic acid) was eluted atthe retention time (RT) 2825min The extracted ion
chromatogram of the monoisotopic peak pattern of theprecursor ion (2+) for glycopeptide EEQYNSTYR 3 3 1 0was illustrated in Figure 1 Tandem mass spectra of theglycopeptide EEQYNSTYR 3 3 1 0 obtained by CID andHCD fragmentation techniques were also shown in Figure 2We confirmed that fragment ions were well-matched withtheir structures
The analysis results of the reference rAb obtainedby triplicate sample preparation and mass analyses weresummarized in Table 1 LC-ESI-MSMSwith tryptic digestiondetected 20 consistent glycosylation patterns with similardistributionsThe glycosylation patterns normalized for eachrun were graphed together in Figure 3 We observed threemajor fucosylated glycopeptides EEQYNSTYR 3 3 1 0EEQYNSTYR 3 4 1 0 and EEQYNSTYR 4 4 1 0 which arereportedly predominant glycoforms found in recombinantantibodies produced in CHO cells [7 24 25] In additionhigh mannose forms with the relative ratios lt13 were alsoobservedmdashEEQYNSTYR 4 2 0 0 EEQYNSTYR 5 2 0 0EEQYNSTYR 6 2 0 0 and EEQYNSTYR 7 2 0 0 Theseglycosylation patterns have been frequently observed in rAb
International Journal of Analytical Chemistry 5
BPCRT 000 - 12000
36908129360
486486090813786
8129360
38018129359 4885
8609078
39195938265
41557140149 4940
860908045388384719 7289
92651156689
75807685170
9379667
54346036748
73319265115
74519270032
57096040029
282510541217 7671
9179984
26957800903 10429
812936281009480141
118258129346
84048725453
95047903803
21174351794
15984451174
4644451187
XIC
RT 2825AA 108570163BP 12159879
12164894
1215987912169907
12174919
1217993812184948
Mass peak(2+)
10 20 30 40 50 60 70 80 90 100 110 1200Time (min)
05
101520253035404550556065707580859095
100Re
lativ
e Abu
ndan
ce
Figure 1 Base peak chromatogram (BPC) by LC-tandemMS analysis of the tryptic digests of reference rAbThe extracted ion chromatogram(XIC) of the glycopeptide EEQYNSTYR 3 3 1 0 identified as one of major components at RT 2825min was also inserted together with themonoisotopic peak pattern of the glycopeptide
therapeutics due to the fact that recombinant rAbs are notfully modified and substantially affected by posttranslationalmodifications [26 27] After confirming the reliability andrepeatability in the analysis of the reference rAb glycoproteina comparative analysis was performed using biosimilar rAbsA and B which were prepared by two different methods ofproduction
32 Comparative Analysis of Glycosylation Patterns in Biosim-ilar rAbs Samples of biosimilar rAbs were obtained fromeither transiently transfected cells or a stably transfected cellline and referred to as biosimilar rAb A and B respectivelyfor clarification purposes The purified biosimilar rAbs Aand B were digested with trypsin and analyzed with LC-ESI-MSMS The obtained raw mass data was analyzed withGPA algorithm and the identified glycopeptides were furthervalidated manually The glycosylation patterns acquired forthe biosimilar rAb A and B were individually normalized andthemapping results were comparedwith those obtained fromthe analysis for the reference rAb sample (see Figure 4)
The major glycoforms for the biosimilar rAbA and rAb B as well as the reference rAb wereconsistentmdashEEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0 However we observed variationsin the minor glycosylation species not only between thereference rAb and biosimilar rAbs but also between thetwo biosimilar rAbs samples produced using differentmethods The biosimilar rAb A and B displayed a decreasedabundance for EEQYNSTYR 3 4 1 0 and an increased
abundance for EEQYNSTYR 4 4 1 0 with the addition ofgalactose in comparison to the reference rAbThe biosimilarrAb B displayed higher abundance for glycans with sialicacidsmdashEEQYNSTYR 4 4 1 1 EEQYNSTYR 5 4 1 1 andEEQYNSTYR 5 4 1 2mdashin comparison to the biosimilar rAbA These results suggest that LC-tandem mass spectrometrycan be a highly sensitive tool for distinguishing subtledifferences in minor glycosylation species
33 Glycosylation Patterns in Mutagenized rAb GlycoproteinWe tested the potential validity of using LC-tandem massspectrometry in the determination process for the similarityof biosimilar products to the referencematerials in biosimilarsamples with a single mutation in amino acid sequencesUnintentionalmutations that arise during protein productioncan cause changes in glycosylation patterns which in turncan impact the safety quality and potency of the recombinantrAbs [28] A tryptophan to valine amino acid mutation wascreated in the vicinity of the glycosylation site of Adalimumabto test its effect on the pattern of glycosylation to be detectedusing LC-tandemmass spectrometry Interestingly the singleamino acid mutation caused the inability for biosimilarrAb to bind to the protein A affinity column (data notshown) So instead of purifying the mutagenized rAb usingprotein A columns we performed the analysis using in-gelsamples Unpurified samples were analyzed by SDS-PAGEand the expected heavy chain and light chain bands wereobserved at 51 kDa and 26 kDa respectively (see Figure 5)The bands excised from the gel were subjected to in-gel
6 International Journal of Analytical Chemistry
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3240 RT 2830 AV 1 NL 279E5T ITMS + c NSI d Full ms2 121649cid3500 [32000-200000]
13926141
104155599606016
11431420
15386483
8795134
11144237
7985649
69015655281868
137569371596713013567330
192069853660828 150270741190462573132361758679212726949 19038585170177336743636
4393188 14856172
[1+]
[1+]
[1+]
[1+]
[1+]
[2+]
6968846
7690668
[2+]
~YNS~
[2+]
~YNS~~YNS~
]
YNS~YNNNNSS
~YNS~
[2+]
~YNS~
~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~
8709283
9526278
~YNS~[1+]
~YNS~
[1+]1+]
~YNS~
[1+]10336189
10620417
17419075
~YNS~ ~YNS~ ~YNS~YNS YNS YNS YNS YNS ~YNS~
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000400mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
-1(2O
-2(2O
[2+] [2+][2+] [2+] [2+] [2+] [2+]
[2+]
[1+] [1+][1+] [1+]
[1+] [1+] [1+]
(a) MSMS spectrum EEQYNSTYR 3 3 1 0 CID
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3241 RT 2830 AV 1 NL 518E5T FTMS + c NSI d Full ms2 121649hcd3500 [10000-200000]
1380549
2040865
1392670310537871
1272611115397529 19199423
131855215814420 1189562164186744677113 14736592 195405772872824 102980527817769
~YNS~ ~YNS~~YNS~
[1+] [1+]
[1+]
1260550
1680660
1860761
-HO
-2HO
-HO - Acetyl-HO-CHO
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000100mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
S9[1+]
(b) MSMS spectrum EEQYNSTYR 3 3 1 0 HCD
Figure 2 Tandem mass spectra of the glycopeptide EEQYNSTYR 3 3 1 0 obtained by (a) collision-induced dissociation and (b) higher-energy collision dissociation
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
4 International Journal of Analytical Chemistry
Table 1 List of tryptic glycopeptides from triplicate analysis of reference rAba
No Glycopeptides IonCharge (+)
DetectedMass (mz)
Run 1 Run 2 Run 3RT (min) MS Intensity RT (min) MS Intensity RT (min) MS Intensity
1 EEQYNSTYR 3 3 0 0 2 1142959 2770 92380184 2839 3270910 2840 10319768
2 EEQYNSTYR 3 3 1 0 2 1215989 2741 372193798 2831 72807858 2830 1887948683 810996
3 EEQYNSTYR 3 4 0 0 2 1244499 2770 91063963 2839 15496872 2842 154070823 830000
4 EEQYNSTYR 3 4 1 0 2 1318029 2740 1541540913 2828 398363214 2830 8154840543 878688
5 EEQYNSTYR 3 5 1 0 3 946380 2760 7447300 2832 5456578 2830 71994696 EEQYNSTYR 4 2 0 0 2 1122443 2760 126933115 2839 1113753 2835 67216327 EEQYNSTYR 4 3 0 1 3 913355 2828 188888 2888 2653427 2890 13007138 EEQYNSTYR 4 3 1 0 2 1297018 2740 50607743 2825 6312278 2821 20746171
9 EEQYNSTYR 4 3 1 1 2 1442557 2810 29760350 2881 55607576 2880 283562773 962043
10 EEQYNSTYR 4 4 1 0 2 1398554 2738 285767417 2825 58136525 2820 2056147173 932706
11 EEQYNSTYR 4 4 1 1 3 1029738 2810 5127014 2880 36633653 2885 1266473012 EEQYNSTYR 4 5 1 0 3 1000398 2754 2426554 2832 4640770 2830 6085787
13 EEQYNSTYR 5 2 0 0 2 1203473 2750 542316083 2828 16939021 2829 694627983 802652
14 EEQYNSTYR 5 4 1 0 3 986723 2736 3496656 2810 2860860 2810 1093770615 EEQYNSTYR 5 4 1 1 3 1083755 2810 8174096 2881 18861846 2880 2560383616 EEQYNSTYR 5 4 1 2 3 1180784 2852 1129042 2920 3397473 2929 118106017 EEQYNSTYR 6 2 0 0 2 1284497 2740 102132925 2818 733668 2821 1402451818 EEQYNSTYR 6 3 1 0 3 973049 2736 4256849 2818 638835 2820 472169919 EEQYNSTYR 6 3 1 1 3 1070081 2804 5703122 2874 18933930 2870 833695720 EEQYNSTYR 7 2 0 0 2 1365524 2740 55356342 2818 559700 2821 5262829aMS Intensity is based on the summation of 3-top monoisotopic mass peak intensity [21] If simultaneously detected MS Intensity values of 2+ and 3+ massions are summed
3 Results and Discussion
31 LC-ESI-MSMS Analysis of Reference rAb GlycoproteinThe reference rAb glycoprotein desalted with a HiTrapMabselect SuRe proteinA column (GEHealthcare) accordingto manufacturerrsquos protocols was quantified by UV spectrom-etry ELISA assay and BCA assay and characterized by SDS-PAGE and Western blotting The reference rAb glycoproteinsample desalted was reduced alkylated in a darkroom anddigested with trypsin overnight The digested sample wasdesalted with a SPE micro-spin column and analyzed byLC-ESI-MSMS coupled with CID and HCD fragmentationtechniques
Figure 1 shows a base peak chromatogram (BPC)obtained from the analysis of the tryptic digests of thereference rAb glycoprotein The obtained raw mass data wasanalyzed with GPA algorithm and the automatically assignedglycopeptides were further validated manually Amongthe identified glycopeptides EEQYNSTYR 3(hexose) 3(N-acetylhexosamine) 1(fucose) 0(sialic acid) was eluted atthe retention time (RT) 2825min The extracted ion
chromatogram of the monoisotopic peak pattern of theprecursor ion (2+) for glycopeptide EEQYNSTYR 3 3 1 0was illustrated in Figure 1 Tandem mass spectra of theglycopeptide EEQYNSTYR 3 3 1 0 obtained by CID andHCD fragmentation techniques were also shown in Figure 2We confirmed that fragment ions were well-matched withtheir structures
The analysis results of the reference rAb obtainedby triplicate sample preparation and mass analyses weresummarized in Table 1 LC-ESI-MSMSwith tryptic digestiondetected 20 consistent glycosylation patterns with similardistributionsThe glycosylation patterns normalized for eachrun were graphed together in Figure 3 We observed threemajor fucosylated glycopeptides EEQYNSTYR 3 3 1 0EEQYNSTYR 3 4 1 0 and EEQYNSTYR 4 4 1 0 which arereportedly predominant glycoforms found in recombinantantibodies produced in CHO cells [7 24 25] In additionhigh mannose forms with the relative ratios lt13 were alsoobservedmdashEEQYNSTYR 4 2 0 0 EEQYNSTYR 5 2 0 0EEQYNSTYR 6 2 0 0 and EEQYNSTYR 7 2 0 0 Theseglycosylation patterns have been frequently observed in rAb
International Journal of Analytical Chemistry 5
BPCRT 000 - 12000
36908129360
486486090813786
8129360
38018129359 4885
8609078
39195938265
41557140149 4940
860908045388384719 7289
92651156689
75807685170
9379667
54346036748
73319265115
74519270032
57096040029
282510541217 7671
9179984
26957800903 10429
812936281009480141
118258129346
84048725453
95047903803
21174351794
15984451174
4644451187
XIC
RT 2825AA 108570163BP 12159879
12164894
1215987912169907
12174919
1217993812184948
Mass peak(2+)
10 20 30 40 50 60 70 80 90 100 110 1200Time (min)
05
101520253035404550556065707580859095
100Re
lativ
e Abu
ndan
ce
Figure 1 Base peak chromatogram (BPC) by LC-tandemMS analysis of the tryptic digests of reference rAbThe extracted ion chromatogram(XIC) of the glycopeptide EEQYNSTYR 3 3 1 0 identified as one of major components at RT 2825min was also inserted together with themonoisotopic peak pattern of the glycopeptide
therapeutics due to the fact that recombinant rAbs are notfully modified and substantially affected by posttranslationalmodifications [26 27] After confirming the reliability andrepeatability in the analysis of the reference rAb glycoproteina comparative analysis was performed using biosimilar rAbsA and B which were prepared by two different methods ofproduction
32 Comparative Analysis of Glycosylation Patterns in Biosim-ilar rAbs Samples of biosimilar rAbs were obtained fromeither transiently transfected cells or a stably transfected cellline and referred to as biosimilar rAb A and B respectivelyfor clarification purposes The purified biosimilar rAbs Aand B were digested with trypsin and analyzed with LC-ESI-MSMS The obtained raw mass data was analyzed withGPA algorithm and the identified glycopeptides were furthervalidated manually The glycosylation patterns acquired forthe biosimilar rAb A and B were individually normalized andthemapping results were comparedwith those obtained fromthe analysis for the reference rAb sample (see Figure 4)
The major glycoforms for the biosimilar rAbA and rAb B as well as the reference rAb wereconsistentmdashEEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0 However we observed variationsin the minor glycosylation species not only between thereference rAb and biosimilar rAbs but also between thetwo biosimilar rAbs samples produced using differentmethods The biosimilar rAb A and B displayed a decreasedabundance for EEQYNSTYR 3 4 1 0 and an increased
abundance for EEQYNSTYR 4 4 1 0 with the addition ofgalactose in comparison to the reference rAbThe biosimilarrAb B displayed higher abundance for glycans with sialicacidsmdashEEQYNSTYR 4 4 1 1 EEQYNSTYR 5 4 1 1 andEEQYNSTYR 5 4 1 2mdashin comparison to the biosimilar rAbA These results suggest that LC-tandem mass spectrometrycan be a highly sensitive tool for distinguishing subtledifferences in minor glycosylation species
33 Glycosylation Patterns in Mutagenized rAb GlycoproteinWe tested the potential validity of using LC-tandem massspectrometry in the determination process for the similarityof biosimilar products to the referencematerials in biosimilarsamples with a single mutation in amino acid sequencesUnintentionalmutations that arise during protein productioncan cause changes in glycosylation patterns which in turncan impact the safety quality and potency of the recombinantrAbs [28] A tryptophan to valine amino acid mutation wascreated in the vicinity of the glycosylation site of Adalimumabto test its effect on the pattern of glycosylation to be detectedusing LC-tandemmass spectrometry Interestingly the singleamino acid mutation caused the inability for biosimilarrAb to bind to the protein A affinity column (data notshown) So instead of purifying the mutagenized rAb usingprotein A columns we performed the analysis using in-gelsamples Unpurified samples were analyzed by SDS-PAGEand the expected heavy chain and light chain bands wereobserved at 51 kDa and 26 kDa respectively (see Figure 5)The bands excised from the gel were subjected to in-gel
6 International Journal of Analytical Chemistry
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3240 RT 2830 AV 1 NL 279E5T ITMS + c NSI d Full ms2 121649cid3500 [32000-200000]
13926141
104155599606016
11431420
15386483
8795134
11144237
7985649
69015655281868
137569371596713013567330
192069853660828 150270741190462573132361758679212726949 19038585170177336743636
4393188 14856172
[1+]
[1+]
[1+]
[1+]
[1+]
[2+]
6968846
7690668
[2+]
~YNS~
[2+]
~YNS~~YNS~
]
YNS~YNNNNSS
~YNS~
[2+]
~YNS~
~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~
8709283
9526278
~YNS~[1+]
~YNS~
[1+]1+]
~YNS~
[1+]10336189
10620417
17419075
~YNS~ ~YNS~ ~YNS~YNS YNS YNS YNS YNS ~YNS~
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000400mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
-1(2O
-2(2O
[2+] [2+][2+] [2+] [2+] [2+] [2+]
[2+]
[1+] [1+][1+] [1+]
[1+] [1+] [1+]
(a) MSMS spectrum EEQYNSTYR 3 3 1 0 CID
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3241 RT 2830 AV 1 NL 518E5T FTMS + c NSI d Full ms2 121649hcd3500 [10000-200000]
1380549
2040865
1392670310537871
1272611115397529 19199423
131855215814420 1189562164186744677113 14736592 195405772872824 102980527817769
~YNS~ ~YNS~~YNS~
[1+] [1+]
[1+]
1260550
1680660
1860761
-HO
-2HO
-HO - Acetyl-HO-CHO
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000100mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
S9[1+]
(b) MSMS spectrum EEQYNSTYR 3 3 1 0 HCD
Figure 2 Tandem mass spectra of the glycopeptide EEQYNSTYR 3 3 1 0 obtained by (a) collision-induced dissociation and (b) higher-energy collision dissociation
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
International Journal of Analytical Chemistry 5
BPCRT 000 - 12000
36908129360
486486090813786
8129360
38018129359 4885
8609078
39195938265
41557140149 4940
860908045388384719 7289
92651156689
75807685170
9379667
54346036748
73319265115
74519270032
57096040029
282510541217 7671
9179984
26957800903 10429
812936281009480141
118258129346
84048725453
95047903803
21174351794
15984451174
4644451187
XIC
RT 2825AA 108570163BP 12159879
12164894
1215987912169907
12174919
1217993812184948
Mass peak(2+)
10 20 30 40 50 60 70 80 90 100 110 1200Time (min)
05
101520253035404550556065707580859095
100Re
lativ
e Abu
ndan
ce
Figure 1 Base peak chromatogram (BPC) by LC-tandemMS analysis of the tryptic digests of reference rAbThe extracted ion chromatogram(XIC) of the glycopeptide EEQYNSTYR 3 3 1 0 identified as one of major components at RT 2825min was also inserted together with themonoisotopic peak pattern of the glycopeptide
therapeutics due to the fact that recombinant rAbs are notfully modified and substantially affected by posttranslationalmodifications [26 27] After confirming the reliability andrepeatability in the analysis of the reference rAb glycoproteina comparative analysis was performed using biosimilar rAbsA and B which were prepared by two different methods ofproduction
32 Comparative Analysis of Glycosylation Patterns in Biosim-ilar rAbs Samples of biosimilar rAbs were obtained fromeither transiently transfected cells or a stably transfected cellline and referred to as biosimilar rAb A and B respectivelyfor clarification purposes The purified biosimilar rAbs Aand B were digested with trypsin and analyzed with LC-ESI-MSMS The obtained raw mass data was analyzed withGPA algorithm and the identified glycopeptides were furthervalidated manually The glycosylation patterns acquired forthe biosimilar rAb A and B were individually normalized andthemapping results were comparedwith those obtained fromthe analysis for the reference rAb sample (see Figure 4)
The major glycoforms for the biosimilar rAbA and rAb B as well as the reference rAb wereconsistentmdashEEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0 However we observed variationsin the minor glycosylation species not only between thereference rAb and biosimilar rAbs but also between thetwo biosimilar rAbs samples produced using differentmethods The biosimilar rAb A and B displayed a decreasedabundance for EEQYNSTYR 3 4 1 0 and an increased
abundance for EEQYNSTYR 4 4 1 0 with the addition ofgalactose in comparison to the reference rAbThe biosimilarrAb B displayed higher abundance for glycans with sialicacidsmdashEEQYNSTYR 4 4 1 1 EEQYNSTYR 5 4 1 1 andEEQYNSTYR 5 4 1 2mdashin comparison to the biosimilar rAbA These results suggest that LC-tandem mass spectrometrycan be a highly sensitive tool for distinguishing subtledifferences in minor glycosylation species
33 Glycosylation Patterns in Mutagenized rAb GlycoproteinWe tested the potential validity of using LC-tandem massspectrometry in the determination process for the similarityof biosimilar products to the referencematerials in biosimilarsamples with a single mutation in amino acid sequencesUnintentionalmutations that arise during protein productioncan cause changes in glycosylation patterns which in turncan impact the safety quality and potency of the recombinantrAbs [28] A tryptophan to valine amino acid mutation wascreated in the vicinity of the glycosylation site of Adalimumabto test its effect on the pattern of glycosylation to be detectedusing LC-tandemmass spectrometry Interestingly the singleamino acid mutation caused the inability for biosimilarrAb to bind to the protein A affinity column (data notshown) So instead of purifying the mutagenized rAb usingprotein A columns we performed the analysis using in-gelsamples Unpurified samples were analyzed by SDS-PAGEand the expected heavy chain and light chain bands wereobserved at 51 kDa and 26 kDa respectively (see Figure 5)The bands excised from the gel were subjected to in-gel
6 International Journal of Analytical Chemistry
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3240 RT 2830 AV 1 NL 279E5T ITMS + c NSI d Full ms2 121649cid3500 [32000-200000]
13926141
104155599606016
11431420
15386483
8795134
11144237
7985649
69015655281868
137569371596713013567330
192069853660828 150270741190462573132361758679212726949 19038585170177336743636
4393188 14856172
[1+]
[1+]
[1+]
[1+]
[1+]
[2+]
6968846
7690668
[2+]
~YNS~
[2+]
~YNS~~YNS~
]
YNS~YNNNNSS
~YNS~
[2+]
~YNS~
~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~
8709283
9526278
~YNS~[1+]
~YNS~
[1+]1+]
~YNS~
[1+]10336189
10620417
17419075
~YNS~ ~YNS~ ~YNS~YNS YNS YNS YNS YNS ~YNS~
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000400mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
-1(2O
-2(2O
[2+] [2+][2+] [2+] [2+] [2+] [2+]
[2+]
[1+] [1+][1+] [1+]
[1+] [1+] [1+]
(a) MSMS spectrum EEQYNSTYR 3 3 1 0 CID
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3241 RT 2830 AV 1 NL 518E5T FTMS + c NSI d Full ms2 121649hcd3500 [10000-200000]
1380549
2040865
1392670310537871
1272611115397529 19199423
131855215814420 1189562164186744677113 14736592 195405772872824 102980527817769
~YNS~ ~YNS~~YNS~
[1+] [1+]
[1+]
1260550
1680660
1860761
-HO
-2HO
-HO - Acetyl-HO-CHO
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000100mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
S9[1+]
(b) MSMS spectrum EEQYNSTYR 3 3 1 0 HCD
Figure 2 Tandem mass spectra of the glycopeptide EEQYNSTYR 3 3 1 0 obtained by (a) collision-induced dissociation and (b) higher-energy collision dissociation
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
6 International Journal of Analytical Chemistry
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3240 RT 2830 AV 1 NL 279E5T ITMS + c NSI d Full ms2 121649cid3500 [32000-200000]
13926141
104155599606016
11431420
15386483
8795134
11144237
7985649
69015655281868
137569371596713013567330
192069853660828 150270741190462573132361758679212726949 19038585170177336743636
4393188 14856172
[1+]
[1+]
[1+]
[1+]
[1+]
[2+]
6968846
7690668
[2+]
~YNS~
[2+]
~YNS~~YNS~
]
YNS~YNNNNSS
~YNS~
[2+]
~YNS~
~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~~YNS~ ~YNS~ ~YNS~ ~YNS~ ~YNS~
8709283
9526278
~YNS~[1+]
~YNS~
[1+]1+]
~YNS~
[1+]10336189
10620417
17419075
~YNS~ ~YNS~ ~YNS~YNS YNS YNS YNS YNS ~YNS~
500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000400mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
-1(2O
-2(2O
[2+] [2+][2+] [2+] [2+] [2+] [2+]
[2+]
[1+] [1+][1+] [1+]
[1+] [1+] [1+]
(a) MSMS spectrum EEQYNSTYR 3 3 1 0 CID
20170120_CJU_AYH_ABH1202_20ug_Rs120000_CIDHCD 3241 RT 2830 AV 1 NL 518E5T FTMS + c NSI d Full ms2 121649hcd3500 [10000-200000]
1380549
2040865
1392670310537871
1272611115397529 19199423
131855215814420 1189562164186744677113 14736592 195405772872824 102980527817769
~YNS~ ~YNS~~YNS~
[1+] [1+]
[1+]
1260550
1680660
1860761
-HO
-2HO
-HO - Acetyl-HO-CHO
200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000100mz
05
101520253035404550556065707580859095
100
Rela
tive A
bund
ance
S9[1+]
(b) MSMS spectrum EEQYNSTYR 3 3 1 0 HCD
Figure 2 Tandem mass spectra of the glycopeptide EEQYNSTYR 3 3 1 0 obtained by (a) collision-induced dissociation and (b) higher-energy collision dissociation
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
International Journal of Analytical Chemistry 7
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_3
_5_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_0
_1
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_3_1
_1
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_4
_4_1
_1
EEQ
YNST
YR_4
_5_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_3_1
_0
EEQ
YNST
YR_6
_3_1
_1
EEQ
YNST
YR_7
_2_0
_0
REF rAb run1 REF rAb run2 REF rAb run3
0
10
20
30
40
50
60
Reference rAb sample
Figure 3 Triplicate analysis for glycosylation patterns of the reference rAb sample using LC-ESI-MSMS Each acquired mass dataset wasnormalized before comparative analysis The sugar symbols diamond (magenta) = sialic acid circle (yellow) = galactose square (blue) =GlcNAc circle (green) = mannose triangle (magenta) = fucose
0
10
20
30
40
50
60
EEQ
YNST
YR_3
_2_0
_0EE
QYN
STYR
_3_2
_1_0
EEQ
YNST
YR_3
_3_0
_0EE
QYN
STYR
_3_3
_1_0
EEQ
YNST
YR_3
_4_0
_0EE
QYN
STYR
_3_4
_1_0
EEQ
YNST
YR_3
_5_1
_0EE
QYN
STYR
_4_2
_0_0
EEQ
YNST
YR_4
_3_0
_0EE
QYN
STYR
_4_3
_0_1
EEQ
YNST
YR_4
_3_1
_0EE
QYN
STYR
_4_3
_1_1
EEQ
YNST
YR_4
_4_1
_0EE
QYN
STYR
_4_4
_1_1
EEQ
YNST
YR_4
_5_1
_0EE
QYN
STYR
_5_2
_0_0
EEQ
YNST
YR_5
_3_0
_1EE
QYN
STYR
_5_3
_1_1
EEQ
YNST
YR_5
_4_1
_0EE
QYN
STYR
_5_4
_1_1
EEQ
YNST
YR_5
_4_1
_2EE
QYN
STYR
_5_5
_1_1
EEQ
YNST
YR_6
_2_0
_0EE
QYN
STYR
_6_2
_1_0
EEQ
YNST
YR_6
_3_0
_1EE
QYN
STYR
_6_3
_1_0
EEQ
YNST
YR_6
_3_1
_1EE
QYN
STYR
_6_4
_0_0
EEQ
YNST
YR_7
_2_0
_0EE
QYN
STYR
_7_2
_1_0
EEQ
YNST
YR_8
_2_0
_0EE
QYN
STYR
_9_2
_0_0
Reference rAb vs Biosimilar rAbs
Ref rAb (averaged) Biosimilar rAb A Biosimilar rAb B
Figure 4 Comparison of glycosylation patterns of the reference rAb and the biosimilar rAb samples obtained by LC-ESI-MSMS Theglycosylation patterns for the reference rAb sample were obtained by triplicate analysis and averaged All datasets acquired from eachglycoprotein sample were normalized before comparative analysis
tryptic digestionThe digested sample was desalted by ZipTipfor mass spectrometry
Prior to analysis of the tryptic digests of the mutagenizedrAb glycoprotein a comparative analysis was performed
between the in-solution reference rAb and the in-gel refer-ence rAb glycoprotein to ensure the reliability and repeata-bility of the developed glycopeptide mapping methods (seeFigure 6) The results in Figure 7 show that the mass
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
8 International Journal of Analytical Chemistry
250
150
100
75
50
25
37
20
15
10
1
Heavy chain (HC)
2 3 4
Light chain (LC)
Figure 5 SDS-PAGE result for mutagenized rAb samples (lanes 1-3)and reference rAb glycoprotein (lane 4) Heavy chain and light chainbands were observed at 51 kDa and 26 kDa respectively
Recombinant Ab Sample
Peptides amp Glycopeptides
Comparative glycopeptide mapping
SDS gel separation
Coomassie staining
Tryptic digestion
In-Gel
Protein Column(desalting or buffer exchange)
Quantification
Tryptic digestion
In-Solution
Peptides amp Glycopeptides
LC-ESI-MSMS LC-ESI-MSMS
Figure 6 Comparative analysis for in-solution and in-gel rAbsamples In-solution sample is applied to a protein column andsubject to tryptic digestion In-gel sample is separated from SDS-PAGE gel and destained before tryptic digestion
spectrometry analysis of both the in-solution and the in-gelreference rAb samples yielded similar results with consistentglycosylation patterns proving the reliability of the method
We then conducted the analysis of the in-gel trypticdigests of the mutagenized rAb glycoprotein and
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_2_1
_0
EEQ
YNST
YR_3
_3_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_0
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_2_0
_0
EEQ
YNST
YR_4
_3_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_7
_2_0
_0
REF rAb (in-sol) REF rAb (in-gel)
REF rAb (in-solution) vs REF rAb (in-gel)
0
10
20
30
40
50
60
70
80
90
100
Figure 7 Comparison of the glycosylation patterns of the referencerAb obtained respectively from the in-solution and the in-gelmethodThe two methods for sample preparation provided compa-rable results implying that the in-gel method can be complimentaryto the in-solutionmethod for mapping protein glycosylation for ourpurposes
compared its pattern against the in-gel tryptic digestsof the reference rAb glycoprotein As shown in Figure 8we observed more substantiated variations in overallglycosylation patterns Again the major glycoforms(including EEQYNSTYR 3 3 1 0 EEQYNSTYR 3 4 1 0and EEQYNSTYR 4 4 1 0) maintain similarity betweensamples to a certain extent consistent with the previousanalysis results of the reference rAb and the wild-typebiosimilar samples prepared by the in-solution method Onthe other hand the differences in glycosylation patternsin minor species of the mutagenized rAb were moresubstantiated in comparison to those of the referencerAb The mapping of the mutagenized rAb glycoproteindisplayed highly sialylated and galactosylated glycoformssuch as EEQYNSTYR 5 4 1 1 EEQYNSTYR 5 4 1 2EEQYNSTYR 6 5 1 1 and EEQYNSTYR 6 5 1 2 Wealso observed the increase in high-branched glycoformssuch as EEQYNSTYR 6 5 1 0 EEQYNSTYR 6 5 1 1EEQYNSTYR 6 5 1 2 and EEQYNSTYR 7 6 1 2Overall we observed a significant increase of sialylatedgalactosylated and branched glycoforms in the mutagenizedrAb glycoprotein These substantial differences inglycosylation patterns can be attributable to the singleamino acid mutagenesis and are as expected fromthe observation that the single amino acid changealso caused the rAbrsquos inability to bind to protein Acolumn
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
International Journal of Analytical Chemistry 9
EEQ
YNST
YR_3
_2_0
_0
EEQ
YNST
YR_3
_3_1
_0
EEQ
YNST
YR_3
_4_1
_0
EEQ
YNST
YR_4
_4_1
_0
EEQ
YNST
YR_5
_2_0
_0
EEQ
YNST
YR_5
_4_1
_0
EEQ
YNST
YR_5
_4_1
_1
EEQ
YNST
YR_5
_4_1
_2
EEQ
YNST
YR_6
_2_0
_0
EEQ
YNST
YR_6
_5_1
_0
EEQ
YNST
YR_6
_5_1
_1
EEQ
YNST
YR_6
_5_1
_2
EEQ
YNST
YR_7
_6_1
_2
REF rAb (in-gel) Mutagenized rAb (in-gel)
REF rAB vs Mutagenized rAb
0
10
20
30
40
50
60
Figure 8 Comparative glycopeptide mapping of the reference rAband the mutagenized rAb glycoprotein by in-gel digestion and LC-ESI-MSMS Each dataset was normalized before comparative anal-ysis Glycosylation patterns were notably different between the ref-erence rAb and the mutagenized rAb samples Increased abundanceof sialylation galactosylation and branching was observed in themutagenized rAb in comparison to the reference rAb glycoprotein
4 Conclusions
The replicate analyses of the reference rAb glycoproteinallowed us to conclude that the comparative glycopeptidemapping method using LC-tandem mass spectrometry andautomated Glycoproteome Analyzer (GPA) software is ahighly robust method for monitoring glycosylation patternsWhen applying this comparative mapping method to theanalyses of the reference and biosimilar rAbs variationsin the distribution of galactosylated and sialylated glyco-forms in minor glycoform species were detected withoutany significant changes in the major glycoform speciesThese variations on the minor glycoform species of thebiosimilar rAb compared to the reference rAb likely reflectthe inherent nature of variations in biosimilar developmentsuch as variations in cell lines used and various conditionsfor culturing the cells in bioreactors for the production ofbiosimilars The ability of LC-tandem mass spectrometry-based system to readily detect such delicate variations inminor glycoform species suggests that the method of analysiscan be an effective tool to be used in determining similar-ities of biosimilars to the reference materials in terms ofglycosylation patterns with increased accuracy and possiblysupplementing results obtained from the more widely used2AB-based methods It is not clear at this point howeverif these subtle variations in the minor glycoform specieswithout any significant variations on the major glycoform
species can become the basis for the failure of the biosimilarrAb in proving similarity with the reference rAb withinthe scope of the FDA guideline The possible variationsin functionality andor immunogenicity of the biosimilarrAb over the reference rAb can only be determined withadditional tests following this report
We also tested the effect of a single amino acid mutationon the overall pattern of glycosylation of the biosimilar rAbCoincidently this single amino acid substitution introducedwithin the CH2 domain of the heavy chain (at position C41by IMGT codon numbering or 290 by Kabat numbering)of IgG molecule caused the inability of the mutagenizedbiosimilar rAb to bind to proteinA columnWithout a readilyavailable method to purify the mutagenized biosimilar rAbwe used in-gel based analysis of the glycoproteins The in-gel based method was first proven to be effective whencompared with in-solution based method in the analysesperformed using the reference rAb sample The in-gel basedanalysis of the mutagenized rAb glycoprotein showed thata single amino acid mutation in the vicinity of the IgG Fcglycosylation site can cause an increased level of variationson the overall pattern of glycoforms and with more pro-nounced variations on the minor species The fact that asingle mutation in the amino acid sequence of biosimilarcan cause a significant change in the resulting glycosylationpatterns andor physiochemical properties emphasizes theimportance of detailed glycopeptide mapping methods suchas LC-tandemmass spectrometry-basedmethod as describedour experiment which can distinguish even slight variationson the minor glycoform species as an added accuracy inbiosimilar development
Data Availability
The data used to support the findings of this study areavailable from the corresponding author upon request
Conflicts of Interest
The authors declare that they have no conflicts of interest
Acknowledgments
This research was funded by the National Research Founda-tion [Grant noNRF-2017R1D1A3B05029756] and theKoreanInstitute for Regional Program Evaluation of the Ministry ofTrade Industry and Energy (Grant no R0003808) throughthe Regional Promotion of RampD Technology DevelopmentProject
References
[1] PM Rudd T Elliott P Cresswell I AWilson and R A DwekldquoGlycosylation and the immune systemrdquo Science vol 291 no5512 pp 2370ndash2376 2001
[2] J Batra and A S Rathore ldquoGlycosylation of monoclonal anti-body products Current status and future prospectsrdquoBiotechnol-ogy Progress vol 32 no 5 pp 1091ndash1102 2016
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
10 International Journal of Analytical Chemistry
[3] M M Fuster and J D Esko ldquoThe sweet and sour of cancerglycans as novel therapeutic targetsrdquo Nature Reviews Cancervol 5 no 7 pp 526ndash542 2005
[4] A Beck and JM Reichert ldquoApproval of the first biosimilar anti-bodies in Europe a major landmark for the biopharmaceuticalindustryrdquomAbs vol 5 no 5 pp 621ndash623 2013
[5] A Beck S Sanglier-Cianferani and A Van DorsselaerldquoBiosimilar biobetter and next generation antibody characteri-zation by mass spectrometryrdquo Analytical Chemistry vol 84 no11 pp 4637ndash4646 2012
[6] R Jefferis ldquoRecombinant antibody therapeutics the impact ofglycosylation on mechanisms of actionrdquo Trends in Pharmaco-logical Sciences vol 30 no 7 pp 356ndash362 2009
[7] L Liu ldquoAntibody glycosylation and its impact on the pharma-cokinetics and pharmacodynamics of monoclonal antibodiesand Fc-fusion proteinsrdquo Journal of Pharmaceutical Sciences vol104 no 6 pp 1866ndash1884 2015
[8] J Pan S Zhang and C H Borchers ldquoComparative higher-order structure analysis of antibody biosimilars using combinedbottom-up and top-down hydrogen-deuterium exchange massspectrometryrdquo Biochimica et Biophysica Acta (BBA) - Proteinsand Proteomics vol 1864 no 12 pp 1801ndash1808 2016
[9] P Juhasz C E Costello and K Biemann ldquoMatrix-assistedlaser desorption ionization mass spectrometry with 2-(4-hydroxyphenylazo)benzoic acid matrixrdquo Journal of The Amer-ican Society for Mass Spectrometry vol 4 no 5 pp 399ndash4091993
[10] B Domon and R Aebersold ldquoMass spectrometry and proteinanalysisrdquo Science vol 312 no 5771 pp 212ndash217 2006
[11] Y Wada M Tajiri and S Yoshida ldquoHydrophilic affinity isola-tion and MALDI multiple-stage tandem mass spectrometry ofglycopeptides for glycoproteomicsrdquo Analytical Chemistry vol76 no 22 pp 6560ndash6565 2004
[12] X Yang S M Kim R Ruzanski et al ldquoUltrafast and high-throughput N-glycan analysis for monoclonal antibodiesrdquomAbs vol 8 no 4 pp 706ndash717 2016
[13] J B Fenn M Mann C K Meng S F Wong and C MWhitehouse ldquoElectrospray ionization for mass spectrometry oflarge biomoleculesrdquo Science vol 246 no 4926 pp 64ndash71 1989
[14] Y H Ahn J Y Kim and J S Yoo ldquoQuantitative mass spec-trometric analysis of glycoproteins combined with enrichmentmethodsrdquoMass Spectrometry Reviews vol 34 no 2 pp 148ndash1652015
[15] J Stadlmann M Pabst D Kolarich R Kunert and F AltmannldquoAnalysis of immunoglobulin glycosylation by LC-ESI-MS ofglycopeptides and oligosaccharidesrdquo Proteomics vol 8 no 14pp 2858ndash2871 2008
[16] M Wuhrer J C Stam F E Van De Geijn et al ldquoGlycosylationprofiling of immunoglobulin G (IgG) subclasses from humanserumrdquo Proteomics vol 7 no 22 pp 4070ndash4081 2007
[17] K Hirayama R Yuji N Yamada K Kato Y Arata andI Shimada ldquoComplete and Rapid Peptide and GlycopeptideMapping of Mouse Monoclonal Antibody by LCMSMS UsingIon Trap Mass Spectrometryrdquo Analytical Chemistry vol 70 no13 pp 2718ndash2725 1998
[18] L He L Xin B Shan G A Lajoie and B Ma ldquoGlycoMasterDB Software to assist the automated identification of N-linked glycopeptides by tandem mass spectrometryrdquo Journal ofProteome Research vol 13 no 9 pp 3881ndash3895 2014
[19] M Bern Y J Kil and C Becker Current Protocols in Bioinfor-matics A D Baxevanis G A Petsko L D Stein and G DStormo Eds JohnWiley amp Sons Inc Hoboken NJ USA 2012
[20] K Lynn C Chen T M Lih et al ldquoMAGIC An Automated N-Linked Glycoprotein Identification Tool Using a Y1-Ion PatternMatching Algorithm andrdquo Analytical Chemistry vol 87 no 4pp 2466ndash2473 2015
[21] G W Park J Y Kim H Hwang et al ldquoIntegrated Gly-coProteome Analyzer (I-GPA) for Automated Identificationand Quantitation of Site-Specific N-Glycosylationrdquo ScientificReports vol 6 no 1 2016
[22] N Y Choi H Hwang E S Ji et al ldquoDirect analysis of site-specific N-glycopeptides of serological proteins in dried bloodspot samplesrdquo Analytical and Bioanalytical Chemistry vol 409no 21 pp 4971ndash4981 2017
[23] S Brezinsky G Chiang A Szilvasi et al ldquoA simple methodfor enriching populations of transfected CHO cells for cells ofhigher specific productivityrdquo Journal of ImmunologicalMethodsvol 277 no 1-2 pp 141ndash155 2003
[24] M Nakano D Higo E Arai et al ldquoCapillary electrophoresis-electrospray ionization mass spectrometry for rapid and sen-sitive N-glycan analysis of glycoproteins as 9-fluorenylmethylderivativesrdquo Glycobiology vol 19 no 2 pp 135ndash143 2009
[25] S Sha C Agarabi K Brorson D-Y Lee and S Yoon ldquoN-Glycosylation Design and Control of Therapeutic MonoclonalAntibodiesrdquo Trends in Biotechnology vol 34 no 10 pp 835ndash846 2016
[26] G C Flynn X Chen Y D Liu B Shah and Z ZhangldquoNaturally occurring glycan forms of human immunoglobulinsG1 and G2rdquoMolecular Immunology vol 47 no 11-12 pp 2074ndash2082 2010
[27] X Chen and G C Flynn ldquoAnalysis of N-glycans from recom-binant immunoglobulin G by on-line reversed-phase high-performance liquid chromatographymass spectrometryrdquo Ana-lytical Biochemistry vol 370 no 2 pp 147ndash161 2007
[28] Y Li T Fu T Liu et al ldquoCharacterization of alanine to valinesequence variants in the Fc region of nivolumab biosimilarproduced in Chinese hamster ovary cellsrdquo mAbs vol 8 no 5pp 951ndash960 2016
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom
TribologyAdvances in
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
International Journal ofInternational Journal ofPhotoenergy
Hindawiwwwhindawicom Volume 2018
Journal of
Chemistry
Hindawiwwwhindawicom Volume 2018
Advances inPhysical Chemistry
Hindawiwwwhindawicom
Analytical Methods in Chemistry
Journal of
Volume 2018
Bioinorganic Chemistry and ApplicationsHindawiwwwhindawicom Volume 2018
SpectroscopyInternational Journal of
Hindawiwwwhindawicom Volume 2018
Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom
The Scientific World Journal
Volume 2018
Medicinal ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
NanotechnologyHindawiwwwhindawicom Volume 2018
Journal of
Applied ChemistryJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
Biochemistry Research International
Hindawiwwwhindawicom Volume 2018
Enzyme Research
Hindawiwwwhindawicom Volume 2018
Journal of
SpectroscopyAnalytical ChemistryInternational Journal of
Hindawiwwwhindawicom Volume 2018
MaterialsJournal of
Hindawiwwwhindawicom Volume 2018
Hindawiwwwhindawicom Volume 2018
BioMed Research International Electrochemistry
International Journal of
Hindawiwwwhindawicom Volume 2018
Na
nom
ate
ria
ls
Hindawiwwwhindawicom Volume 2018
Journal ofNanomaterials
Submit your manuscripts atwwwhindawicom