Embed Size (px)
Citation preview
Quality assurance of RNA derived fromlaser microdissected tissue samplesobtained by the PALM® MicroBeam Systemusing the RNA 6000 Pico LabChip® kit.
Application
AbstractGene expression profiling of material isolated by microdissection has
become a very popular method for analyzing cellular behavior in a micro
scale and is used in a wide range of research and clinical applications.
Laser-assisted microdissection allows isolation of RNA from specific cell
populations out of their surrounding tissue environment. Usually, the
sample yield that can be obtained by this method is too low for quality
control of RNA using a standard technique. On the other hand, extraction
of high quality RNA is crucial for all subsequent steps and the success of
the overall experiment. Since the introduction of the Agilent RNA 6000
Pico LabChip® kit, it is possible for the first time to analyze total RNA
samples in concentrations down to 200 pg/µl, which is in the range of the
sample concentration one can expect from microdissected tissue or cells.
In this study we analyzed RNA samples that were obtained by laser
microdissection and pressure catapulting (LMPC; PALM® MicroBeam
System) from tissue sections. The influence of different staining proce-
dures and RNA extraction kits on RNA quality was investigated. RNA was
tested for quality and approximate yield, as well as the reproducibility
within homologous sample replicates. The data revealed clear differences
between the different isolation kits and also between the staining
procedures that were used.
Marcus Gassmann
Introduction
The analysis of gene expressionhas become one of the major scientific tools for understandingthe behavior of cells in vivo and in vitro and is used in many bio-logical and medical applications.Using microarray hybridization,the abundance of a large numberof transcripts can be analyzed in a single experiment. Comparisonof different expression patternsenables scientists to correlate thechanges in the transcriptome tovarious external factors influencingthe cellular behavior. Especially inclinical research there is the needfor methods that allow expressionanalysis from very specific cells. Prerequisite for meaningful resultsare samples of high purity withoutcontamination from unwantedcells that could potentially inter-fere with detection of importanttranscripts. Laser-assistedmicrodissection with thePALM®MicroBeam system allowsisolation of individual cells in asimple and fast way. Using thistechnology, researchers are ableto procure pure cell populationsfrom specific areas of tissue sections for isolation of DNA,RNA, and proteins with visual control using a microscope. AnUV-A-laser-mediated process dissects selected specimen fromvarious sources and transfersthem contact free directly into collection vessels for subsequentextraction (see figure 1). Togetherwith novel methods for linearRNA amplification it is possible toperform gene expression profilinganalyses even from very limitedamount of cell material. The most
2
Figure 1LMPC procedure (example)A Laser microdissectionB Laser pressure catapulting: Section after catapulting of selected areaC Catapulted area in the collection device
A
B
C
critical parameter for the successof such an experiment is theintegrity and purity of the RNA.
Experimental case study
Tissue preparation and microdissectionSnap-frozen mouse liver tissuestored at –80 °C was cut in 7 µmserial sections on a cryotome at–25 °C. The sections were trans-ferred to PALM® MembraneSlides(1mm glass slides covered with a1.35 µm thin Polyethylene-naph-thalate-membrane to facilitate thelaser pressure catapulting proce-dure) and air dried for 10 seconds.After a fixation step of 5 minutesin 70% ethanol at -20 °C and ashort wash in RNase free water(10 seconds) the sections werefurther processed according tostandard histochemical proce-dures. Depending on the subsequenttests either Hematoxylin/Eosin(HE), Nuclear Fast Red (NFR),Methylgreen (MG) or MethyleneBlue (MB) staining was applied.After staining and a short increas-ing ethanol series the sectionswere air-dried and either usedimmediately or refrozen at –20 °Cfor up to 2 days. The PALM® Micro-Beam System was used to preciselyexcise the selected tissue areaswith the UV-A-laser and then tocatapult the areas of interest contact free against gravity intocollecting caps filled with 12 µl of RNase-free water. For lysis the isolated material was thenimmediately mixed by inversionwith the respective lysis buffersfor RNA extraction (see below).
3
RNA PurificationFor this study RNA isolation kitsfrom 3 different manufacturerswere used. Two are based on acolumn-purification, while thethird one uses an extraction-/pre-cipitation-based approach. RNAcleanup was carried out accordingto the manufacturers’ protocolswith the only variation that uni-formly a volume of 30 µl of elutionbuffer or water was used in thefinal step to collect the purifiedRNA. For the column-proceduresthe elution process was repeatedonce with the primary eluate of 30 µl to improve the yield. The scope of this work was thedemonstration of feasibility to per-form RNA sample quality controlfrom microdissected material andnot the evaluation of commerciallyavailable RNA extraction kits.Therefore, the manufacturers arenot explicitly named here, butinstead referred to as manufacturerA, B, and C.
RT-PCRSelected RNA samples werereversely transcribed with theFirst Strand cDNA Synthesis Kitfor RT PCR (AMV) cDNA Synthe-sis Kit for RT-PCR (Roche)according to the manufacturers’protocol. Briefly, 5-8 µl of eachRNA solution were transcribed byAMV-RTase with random primersin a total volume of 20 µl for onehour at 42 °C. For the subsequentquantitative RT-PCR analyses 1-2 µl of each cDNA-solution were used as templates. The PCRamplification of the cDNA wasperformed in a LightCycler®
instrument (Roche) in 20 µl reactionsusing protocols and components ofthe QuantiTect SYBR Green PCR-Kit(Qiagen). cDNA-specific primers for murine PBGD (porphobilinogendeaminase) were used as model system producing a PCR-fragment of 154 bp. Resulting crossing-pointvalues were used to compare extraction efficiencies of differentkits or influences of different staining procedures.
RNA 6000 Pico Assay ProtocolAll chips were prepared according to the instructions provided with the RNA 6000 Pico LabChip Kit.In brief, 550 µl of the RNA 6000 Picogel matrix were placed on a spin filter, centrifuged at 1500 g and divided in 65 µl aliquots. After addition of 1µl RNA 6000 Pico dyeconcentrate to one gel aliquot, the mixture was vortexed and centrifuged at 13000 g for 10 min. A RNA 6000 Pico chip was filled with gel-dye mix using the chip priming station, followed by addition of Conditioning Solutionand Marker. 1 µl of RNA 6000 ladder (Ambion) and RNA sampleswere added in the designated wells,the chip was vortexed for 1 minuteand run on an Agilent 2100 bio-analyzer.
Results and Discussion
RNA quality and yield obtained by dif-ferent isolation kitsAll 3 kits tested in this studyappeared to work well togetherwith the RNA 6000 Pico assay (figure 2). For the comparisonof the different kits we used astandard HE tissue staining anddissected areas corresponding to1000, 2000 and 3000 cells. Qualityand yield of the isolated RNAshowed a high variability withrespect to the different kits andalso within the replicates of onespecific method. This may becaused either by differences in thetissue samples (different areaswere isolated from the same section) or by some variances insample handling. On average theextraction-/precipitation-basedmethod (manufacturer C) showeda higher RNA recovery than thetwo column-based methods,which is most probably caused byirreversible RNA-binding at thecolumn material. The averageyield (manufacturer C) obtainedfrom a 1000 cell area was approxi-mately 700 pg/µl when resolvingthe pellet in 30 µl of water. Thiswould result roughly in a theoreti-cal average RNA recovery rate ofapprox 21 pg per cell. In contrastto this the yield obtained by thetwo column-based kits variedbetween approx. 130 and 380 pg/µl(1000 cells; 30 µl elution volume).Note that the quantitative resultsfrom an analysis with the RNA6000 Pico kit are somewhatdependent on the salt contentof asample. Only a rough estimationof the RNA concentration of asample can be obtained.
4
18S
28S
Fluo
resc
ence
Time (seconds)
0.00.51.01.52.02.53.03.54.0
Fluo
resc
ence
0. 0
0.5
1.01.5
2.02.5
3.0
3.54.0
Fluo
resc
ence
Fluo
resc
ence
0. 00.51.01.52.02.53.03.54.0
19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
18S
28S
0
1
2
3
4
5
6
7
B
C
D
18S
28S
A
Figure 2RNA was isolated from mouse liver tissue sections using different RNA isolation kits. Histologicalstaining was carried out using a standard HE procedure. Microdissected areas represent approx.1000 cells each. A: Manufacturer A (column-based; no DNAse digest ) B: Manufacturer A (column-based; including DNAse digest on the column during cleanup as recommended by the supplier).C: Manufacturer B (column-based; no DNAse digest) D: Manufacturer C (extraction-/precipitation-based, no DNAse digest)
5
RNA quality and yield obtained fromtissue stained with different methodsIn this test, we compared RNAquality and yield when using different standard histochemicalstaining methods. Therefore, wedissected areas corresponding to2000 cells and purified the RNAusing the kit of manufacturer C(3 replicates for each sample).Serial tissue sections were stainedusing either Hematoxylin/Eosin(HE), Nuclear Fast Red (NFR),Methylgreen (MG) or MethyleneBlue (MB). In general, all fourstaining procedures were compati-ble with the RNA Pico Assay, butsignificant differences in RNAyield and quality were observed. Best results were obtained usingthe MG stain (average yield:approx. 3100 pg; highly intactRNA; see fig. 3A). HE (figure 3B)and NFR staining (figure 3C) gavevery similar results in terms ofRNA yield and quality (averageyield: approx. 1600 pg; slightlydegraded RNA). In the case of MBstaining, ambiguous results wereobserved (average yield: approx.730 pg; partially degraded RNA;figure 3D). The 28 S band isalmost entirely missing, whichcould be due to RNA degradationor due to a salt content thatexceeds the salt specifications ofthe Pico kit. RNA-extraction withthe kit of manufacturer B from the same sections with all fourstaining methods yielded slightlydifferent relations (see in RT-PCRresults below), and overall lowerconcentration levels as expectedfrom the comparison of the kitsfrom above. It should be notedthat this short investigation doesnot put a value on the use of certain cell stains for work withmicrodissected material but ratherpoints out the value of the RNA6000 Pico kit for optimization of experimental protocols.
18S
28S
Fluo
resc
ence
Fluo
resc
ence
Fluo
resc
ence
Fluo
resc
ence
Time (seconds)
0
5
10
15
20
25
19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
0
5
10
15
20
25
30
0.02.55.07.5
10.012.515.017.520.0
18S
28S0
123456789
B
C
D
18S
28S
18S
28S
A
Figure 3RNA was isolated from mouse liver tissue sections using Manufacturer C (extraction-/precipitation-based). Each microdissected area was chosen to represent approx. 2000 cells. A: Methylgreen stained tissueB: Hematoxylin/Eosin stained tissueC: Nuclear Fast Red stained tissueD: Methylene Blue stained tissue
6
1 2
Fluo
resc
ence
Time (seconds)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1 9 24 2 9 34 3 9 44 4 9 54 5 9 64 6 9
18S
28S
Fluo
resc
ence
Time (seconds)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
1 9 24 2 9 34 3 9 44 4 9 54 5 9 64 6 9
18S
28S
Fluo
resc
ence
Time (seconds)
0
1
2
3
4
5
6
7
8
9
19 24 29 34 39 44 49 54 59 64 69
A B C
A
Figure 4RNA was isolated from microdissected mouse liver tissue sections using different procedures(Manufacturer A, B, C) and used for real-time RT-PCR with the LightCycler®. Histological stainingwas carried out using a standard HE procedure. Microdissected areas represented approx. 1000cells each.A: Bioanalyzer results (RNA from A, B, C)B: specific melting curves (PCR products from A, B, C)C: LightCycler growth curves (PCR products from A, B, C)
Correlation of the results obtained by the RNA Pico Assay to real-time RT-PCRComparison of RNA extraction kits:All 3 sample extraction kits thatwere tested yielded good qualityRNA that could be used for RT-PCR (figure 4). For the com-parison we used HE-stained sec-tions to microdissect and catapultareas of 1000 cells each (3 repli-cates per sample). After RNA isolation with the three differentkits real-time RT-PCR with theLightCycler‚ was performed. Theamount of cDNA used as templatewas half of the RNA used for the2100 bioanalyzer analysis (fig. 4A).The specificity of the PCR-frag-ments was proven by the meltingcurve (figure 4B). The crossingpoints (Cp)* of the growth curves (figure 4C) being a scale for the initial specific RNA-amount wereclose together (about one cycledistance each) correlating to theBioanalyzer values although Manufacturer A was overestimat-ed there probably due to the DNA-contamination (figure.1).
*(CpA: 31.9; CpB: 33.1; CpC: 30.7)
7
Comparison of staining proceduresIn this test serial tissue sections were stained usingeither Hematoxylin/Eosin(HE), Nuclear Fast Red (NFR),Methylgreen (MG) or MethyleneBlue (MB). Areas correspondingto 2000 cells were dissected,the RNA purified using Manu-facturer B (3 replicates each)and analyzed with the 2100 bioanalyzer (figure 5A) and the Light-Cycler® as above. The cDNA input for the PCRcorresponded to a fourth of the RNA amount used for the2100 bioanalyzer run. The melt-ing curve proves the specificityof the products (figure 5B),while the growth curve reflectsthe starting amount of RNA(figure 5C). In contrast to thetest with Manufacturer C (seeabove) NFR was the optimalstain, though the Cp-values layvery close together (only half a cycle difference) except for MB which showed worseresults. (CpNFR: 34,7; CpHE:35,2; CpMG: 35,8; CpMB: 37,6).This real-time RT-PCR resultsagain correlated well with the2100 bioanalyzer values.
Fluo
resc
ence
Time (seconds)
0.00.51.01.52.02.53.03.54.04.55.05.56.06.5
19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
Time (seconds)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
19 24 29 34 39 44 49 54 59 64 69
Time (seconds) 19 24 29 34 39 44 49 54 59 64 69
Fluo
resc
ence
Fluo
resc
ence
Fluo
resc
ence
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
NFR MG
MB HE
A
Figure 5RNA was isolated from mouse liver tissue sections using Manufacturer B and used for real-timePCR with the LightCycler®. Four different stainings (HE, NFR, MG, MB) were compared. Eachmicrodissected area was chosen to represent approx. 2000 cells.A: 2100 bioanalzyer resultsB: specific melting curvesC: LightCycler growth curves
Conclusion
RNA derived from laser-microdis-sected tissue isolated by thePALM®MicroBeam system is ofhigh quality and can be analyzedefficiently using the RNA 6000Pico assay with the Agilent 2100bioanalyzer. RNA-purification kitsof different manufacturers andvarious common staining proce-dures have been tested and werecompatible with the Pico Assay.Nevertheless, it seems to be ad-visable that any extraction protocoland staining method should betested in combination to find the optimal procedure for tissuemicrodissection experiments. The RNA 6000 Pico kit is well suited to show differences in RNAquality and, therefore, is an idealtool to optimize experimental conditions. Its unprecedented sensitivity allows for the first time quality control in the contextof microdissection experimentsensuring successful gene ex-pression profiling experiments.
Marcus Gassmann is Research
Chemist at Agilent Technologies,
Waldbronn, Germany.
We would like to acknowledge
the work of Dr. Ulrich Sauer of
P.A.L.M. Microlaser Technologies
AG for his support of this appli-
cation note. More information on
P.A.L.M. Microlaser Technologies
AG can be obtained at
www.agilent.com/chem/labonachip
The information in this publication is subject to change without notice.
Copyright © 2003 Agilent TechnologiesAll Rights Reserved. Reproduction, adaptationor translation without prior written permissionis prohibited, except as allowed under thecopyright laws.
Published March 1, 2003Publication Number 5988-9128EN
LabChip® and the LabChip Logo are registered trademarks of Caliper Technologies Corp. in the U.S. and other countries.
