at SciVerse ScienceDirect

Biologicals 40 (2012) 364e368

Contents lists available

Biologicals

journal homepage: www.elsevier .com/locate/biologicals

Meeting report

Determinants of immunogenic response to protein therapeutics

Satish K. Singh a, Leslie P. Cousens b, David Alvarez c, Pramod B. Mahajan d,*

aBioTherapeutics Pharmaceutical Sciences, Pharmaceutical Research & Development, Pfizer Inc., Chesterfield, MO 63017, USAbDivision of Protein Therapeutics, EpiVax, Inc., 146 Clifford Street, Providence, RI 02903, USAcDepartment of Microbiology and Immunobiology, Harvard Medical School, 77 Avenue Louis Pasteur, NRB Room 836, Boston, MA 02115, USAdDepartment of Pharmaceutical, Biomedical and Administrative Sciences, Drake University College of Pharmacy & Health Sciences, 2507 University Avenue,Des Moines, IA 50311, USA

a r t i c l e i n f o

Article history:Received 5 April 2012Received in revised form1 June 2012Accepted 2 June 2012

Keywords:BiotherapeuticsImmunogenicityProtein therapeuticsAntigenic determinantsTregitopeDendritic cells

* Corresponding author.E-mail address: Pramod.Mahajan@Drake.edu (P.B.

1045-1056/$36.00http://dx.doi.org/10.1016/j.biologicals.2012.06.001

a b s t r a c t

Protein therapeutics occupy a very significant position in the biopharmaceutical market. In addition tothe preclinical, clinical and post marketing challenges common to other drugs, unwanted immunoge-nicity is known to affect efficacy and/or safety of most biotherapeutics. A standard set of immunogenicityrisk factors are routinely used to inform monitoring strategies in clinical studies. A number of in-silico,in vivo and in vitro approaches have also been employed to predict immunogenicity of biotherapeutics,but with limited success. Emerging data also indicates the role of immune tolerance mechanisms andimpact of several product-related factors on modulating host immune responses. Thus, a comprehensivediscussion of the impact of innate and adaptive mechanisms and molecules involved in induction of hostimmune responses on immunogenicity of protein therapeutics is needed. A detailed understanding ofthese issues is essential in order to fully exploit the therapeutic potential of this class of drugs. ThisRoundtable Session was designed to provide a common platform for discussing basic immunobiologicaland pharmacological issues related to the role of biotherapeutic-associated risk factors, as well as hostimmune system in immunogenicity against protein therapeutics. The session included overviewpresentations from three speakers, followed by a panel discussion with audience participation.

1. Introduction

A basic feature of the immune system is tolerance to self-proteins acquired through mechanisms of central tolerance.However, a significant number of human or ‘self’-derived proteintherapeutics also exhibit immunogenicity, suggesting characteris-tics that are not recognized as self, and/or the existence of addi-tional pathways/mechanisms underlying these responses [1]. Inmost cases, immunogenicity manifests itself as the generation ofneutralizing and non-neutralizing polyclonal antibodies directedagainst the administered therapeutic rendering it less efficacious[2,3]. Similarly, responses can be generated in individuals who arenot tolerant because they do not produce a particular humanprotein or parts thereof. For example, patients with severe hemo-philia A involving large deletions or nonsense mutations of theFactor VIII gene are more likely to have an antibody response toexogenous Factor VIII than patients with less severe mutations,because their immune system views the therapeutic as a foreignprotein [4]. Thus, it is generally accepted that immunogenicity of

Mahajan).

biological therapeutics can potentially compromise their efficacy[5,6]. Moreover, immunogenicity to biological therapeutics alsoleads to a variety of adverse reactions such as hypersensitivity andallergic reactions [7,8], limiting their utility as therapeutic agents. Athorough understanding of the underlying cellular, biochemical, aswell as molecular mechanisms contributing to such immunogenicresponses is very valuable in devising strategies to overcome theselimitations. Therefore, a Roundtable Session was organized at theAnnual Meeting of the American Association of PharmaceuticalScientists held during October 22e27, 2011 in Washington D.C. Theobjectives of this session were to (i) review current understandingof various product-related factors leading to immunogenicity ofbiotherapeutics (ii) discuss utility of in-silicomethods for predictingand reducing immunogenicity of biotherapeutics and (iii)understand cellular events underlying the immunogenic responseto biotherapeutics.

2. Summary of the session

Immunogenicity of a therapeutic protein depends largely on itsability to trigger either a cellular or humoral immune response. It iswell understood that an immune reaction to a foreign protein is

S.K. Singh et al. / Biologicals 40 (2012) 364e368 365

more likely because of the higher probability of a “foreign” epitopebeing recognized by T cells. The T cells recognize small linearpeptides derived from protein antigens upon their uptake,proteolytic processing, and presentation in the context of majorhistocompatibility complex (MHC) by an antigen-presenting cell.Thus, the affinity of epitopes for theMHC is a critical determinant ofimmunogenicity. A number of factors inherent in the bindingbetween a T-cell epitope and the MHC binding groove that makethis relationship amenable to computer-based prediction algo-rithms have been described [9]. Similarly, role of dendritic cells hasalso been investigated extensively to understand various factorsthat contribute to the breakdown of tolerance to a human proteintherapeutic. According to the general consensus in the field, factorsresponsible for this breakdown of tolerance to biotherapeutics maybe grouped under two broad categories: (a) Factors related to thechemistry, manufacture and control (CMC) of biotherapeutics, alsotermed ‘product-related factors’, and (b) clinic- or patient-relatedfactors [10]. The first talk by Dr. Satish Singh focused on theproduct-related factors.

Product-related factors are those that arise from the design ofthe molecule, and those related to quality (i.e. the manufacturingprocess and the final product). The design of the molecule involvesthe structure and the target. Presence or absence of human or non-human sequences or epitopes impacts the immunogenicityoutcome. The development of the therapeuticmonoclonal antibodyfield, moving from mouse to chimeric to humanized to human, hasbeen largely driven by the concept of self vs non-self immunity [11].The nature of the therapeutic (immunostimulatory vs immuno-suppressive, or agonist vs antagonist) proteins themselves canimpact the resulting immune response. For example, the adjuvant-like function of granulocyte-macrophage colony-stimulating factor(GM-CSF) is a likely explanation for its higher immunogenicity ascompared to granulocyte-colony-stimulating factor or G-CSF [1,12].Similarly the type 1 interferons are known potentiators of theimmune system and are also likely to enhance the immuneresponse to self [13]. Therapeutic antibodies that bind to cell surfacedeterminants may have greater potential to be immunogenic thanthose that interact with soluble targets [11,14]. For example, theprofound immunogenicity observed by the anti-CD28 humanizedmonoclonal antibody TGN1412, was determined to be its “super-agonist” activation of surface CD28 on memory T lymphocytesleading to a near fatal cytokine storm in humans [15]. Aside froma therapeutic drug’s direct mode of action (MOA) invoking immu-nogenicity, the potential also exists for a therapeutic protein drug topotentially act as a “superagonist” when cross-linked with anti-drug antibodies [16]. A strategy for tolerance induction has beenproposed based on this concept, whereby a cell-bindingmAb, whenminimally mutated to a monomeric non-cell binding version, losesimmunogenicity and can be used to induce tolerance against theoriginal cell-binding form [17]. Glycosylation patterns, determinedby the expression systems for recombinant proteins, can alsoimpact the risk of immunogenicity. The most likely effect of non-human glycosylation is an IgE reaction. However, lack of glycosyl-ation has been shown to lead to a higher immunogenic response.Thus, deglycosylated GM-CSF (1, 12) is more immunogenic than theglycosylated protein (1, 12). Similarly, unglycosylated recombinanthuman Interferon b (rhIFN-b) 1b gives rise to much greaterimmunogenicity than glycosylated rhIFN-b1a [10,18,19].

Product quality factors such as impurities, contaminants, frag-ments, aggregates, and other product-related degradants have allbeen implicated in the generation of immune responses. Of thesefactors, aggregation is considered to carry the greatest risk forpotentiating immunogenicity. A number of hypotheses have beenproposed to explain how aggregates could lead to a breakdown oftolerance, and thus leading to immunogenicity [20,21]. Studies in

animal models have shown that aggregation as a general risk factoris probably too simplistic, and all aggregates do not carry the samerisk. Aggregates resulting from different stresses have varyingcharacteristics in terms of size, morphology and protein structure.While all aggregatesmay potentially lead to an immune response, itis currently accepted that, aggregates with native-like structure arelikely to have the greatest potential for creating a neutralizingimmune response [20,22e24]. As mentioned above, aggregatescome in a wide range of sizes and morphologies. Concern seems tobe lower for aggregates in the oligomeric size ranges measurable byhigh-pressure size exclusion chromatography and greater for thelarger sizes that fall into the submicron to micron ranges [25].However, a couple of recent publications suggest that sucha generalization may also not always hold true [24,26].

Although aggregation garners the most attention, care must betaken to understand the impact of other degradation mechanismson the potential for immunogenicity. Common pathways such asoxidation and deamidation, while not often implicated directly, canresult in alteration of structure and exposure of novel epitopes orcan result in aggregation and thus indirectly impact the immuno-genicity. For example, oxidationwas shown to impact the structureof Fc mAb enhancing aggregation and deamidation [27], andmetal-catalyzed oxidation of IFN-b was shown to enhance immunoge-nicity in transgenic mice [28]. No published examples could befound for the clinical impact of deamidated drug product onimmunogenicity. However, the risk comes not only from chemicalchanges to a product after packaging and during storage, but alsoafter administration. In vivo post-translational modifications ofendogenous proteins (such as deamidation, oxidation, acetylation,deimination, isoaspartylation, dimerization, and phosphorylation)are implicated in several autoimmune diseases [29,30]. Deamida-tion is also a likely in vivo fate of biotherapeutics once injected[31,32]. The concern thus arises that a therapeutic which issusceptible to such post-translational modifications in vivo mayhave a greater ability to trigger an immunogenic response than onewhich does not carry such a liability in its structure.

Process-related impurities such as host cell proteins (HCPs) orDNA may have adjuvant-like effects. The first version of the bio-similar rhSomatotropin used in the development program led toalmost 60% of patients developing anti-growth hormone antibodiesand was traced to high levels of HCPs from Escherichia colienhancing antibody production against growth hormone [33].Wang et al. [34] showed significant induction of interleukin-6 ina whole-blood assay by preparations containing HCPs vs purepreparations. Bacterial DNA contains unmethylated CpGmotifs thatare known to activate toll-like receptors and thus may provideadjuvant activity [35].

Contaminants such as leachables also have the potential tofunction as adjuvants. This effect was shown for the plasticizer di-(2-ethylhexyl) phthalate (DEHP) in mice, when injected with theantigen ovalbumin where DEHP at 5 mg/kg enabled an IgG1response after a single subcutaneous injection [36]. Plasticizers aresmall-molecule compounds added to polymers to improve theirfunctional properties. DEHP or other plasticizers are commonlypresent in plastic components used for storage, dosing, andadministration of biological drugs, and may also be extracted byconstituents such as surfactants in the formulation of the product[37,38]. The prescribing information for Xyntha� (antihemophilicfactor, recombinant) (www.xyntha.com) refers to this effect.

These examples teach us that reduction of clinical immunoge-nicity requires a concerted effort from discovery and developmentscientists. Development scientists have to reduce risk by improvingthe quality aspects of the product. However, from the perspective ofthe discovery scientist, it is not sufficient that a molecule isdesigned only with the intended therapeutic effect and potency in

S.K. Singh et al. / Biologicals 40 (2012) 364e368366

mind. Considerations of factors important in CMC or ‘develop-ability’ must also play a role in the design and selection of thecandidate e a lesson learned by the small-molecule pharmaceuti-cals world more than a decade ago [39]. Potential aggregation-prone regions and hot spots for post-translational modificationsin vitro as well as in vivo must be examined as part of the design[40]. As discussed in the subsequent talk by Dr. Leslie Cousens fromEpiVax, computationally guided mitigation of T-cell epitopesshould be considered. Once a candidate is selected, there arelimited tools in the development toolbox with the ability to impactthe immunogenicity outcome.

While product-related factors play an important role in induc-tion of immune responses to a biotherapeutic, this phenomenonprovides only a partial explanation for the observed heterogeneityof immune responses to various protein therapeutics. Molecularanalysis of interactions between the antigen and the host immunesystem may lead to a better understanding of immune regulation,which in turn could prove valuable in developing practicalapproaches for clinical management of the immunogenic responseto biotherapeutics. One such approach is to develop methods foridentifying and modifying antigenic epitope[s] in biotherapeutics.Dr. Leslie Cousens, the next speaker, presented a unique approachfor epitope prediction and application of this approach to modulateantigenic responses.

There are a number of factors inherent in the binding betweena T-cell epitope and an MHC binding groove that make this rela-tionship amenable to computer-based prediction algorithms. Asseveral common HLA-DR types share largely overlapping peptidebinding repertoires, analysis focused on as few as seven MHCmolecules can “cover” the genetic background of most humansworldwide [9]. Amino acid preferences for interacting withdifferent parts of theMHC binding groove can be observed and thenextrapolated to predict interactions between unknown epitopesand certain MHC molecules. This method initially described bySturniolo et al. [41] has been adapted by Koita et al. to generateClass II prediction tools [42]. EpiVax’s EpiMatrix algorithm parsesprotein sequences into overlapping 9-mer peptide frames, each ofwhich is then evaluated for binding potential to each of the 8common MHC class II HLA alleles that generally represent thenatural breadth of genetic diversity in human HLA [9]. Thepredictive value of this algorithm has been extensively tested and issupported by published in vitro as well as in vivo studies [43e48].

Mapping putative epitopes within a candidate protein thera-peutic is a starting point for assessing and mitigating the potentialimmunogenicity of a whole protein. Proteins containing many, orconcentrated clusters of T-cell epitopes, are predicted to be highlyimmunogenic, while those containing few sparse epitopes aremorelikely to be less immunogenic. These characteristics can be factoredinto the estimate of a whole protein’s overall ‘immunogenicityscore’, a useful tool for consideration in the selection and bioengi-neering of protein therapeutics. The ability to assign this stan-dardized measure of immunogenicity to a protein facilitatesinformed decisions about the likelihood that a protein will provokean immune response [49]. This strategy can be applied for assessingimmunological risk of a potential protein therapeutic in the earlieststages of the drug development process. Prediction of immuno-genic epitopes has led to the strategy of deimmunization by epitopemodification to disrupt HLA binding, and thus T-cell stimulation.The ability to analyze, predict, and modify the immunogenicity ofa potential protein therapeutic has tremendous benefits at everystage of the drug development process.

The development of immune responses to recombinant humanor humanized proteins can be considered a form of ‘breaking’ ofimmunologic tolerance. The reduction and/or the overwhelming ofT regulatory cell (Treg) responses in the application of therapeutic

proteins are potential contributors to anti-therapy immuneresponses that may be considered. In 2007, EpiVax made thesurprising discovery that therewere highly conserved, promiscuousT-cell epitopes located in the Fc region and framework of the Fabregion of IgG that were hypothesized to be regulatory T-cellepitopes. Further examination indicated that they stimulated Tregup-regulation of FoxP3 in vitro, in vivo induced adaptive tolerance,and thus they were named Tregitopes [50]. Our recent data suggestthat 1] APC present Tregitopes to Treg, 2] Treg produce IL-10 andproliferate in response to Tregitope stimulation, and 3] the APCphenotype shifts towards tolerance. Thus, we speculate that thesesmall peptides provide a universal ‘off switch’ for immuneresponses, a finding thatmay have a range of applications, includinga new tool to mitigate therapeutic protein immunogenicity.

The ultimate solutions to regulating immunogenic responses tobiotherapeutics would necessarily have to come from the variouscellular components of the host immune system itself. Althoughthe mammalian cellular circuitry is quite complex, recent techno-logical advances have contributed greatly to our understanding ofthe role of various immune cell types in mounting an immuneresponse. Dr. David Alvarez summarized the role of dendritic cellsas major players in the initial decision-making process thatpromotes immunogenicity.

Most protein therapeutics can potentially generate immuneresponses that could impact drug safety and efficacy [2,3]. Theunderlying cause of this seems contrary to our understanding ofcentral tolerance [51,52]: that self-derived proteins re-introducedinto the host as an immunotherapeutic agent should be tolerizedand unable to evoke immune responses. Consequently, the fact thatantibodies are generated towards immunotherapeutics compels usto rethink how immunogenicity can arise at the cellular as well asmolecular level. Our current understanding of tolerance vs immu-nity stems from our view that the innate and adaptive immunesystem has evolved to discriminate between host and foreignentities [53e55]. Two prevailing views, the self/non-self-hypothesis [53,54] and danger model [55], provide compellingarguments that the immune system not only recognizes self fromnon-self antigens but also integrates additional cues at the time ofantigen encounter. It is in this context, that an in-depth look intothe initial cellular and molecular events engaged upon theadministration of an immunotherapeutic may yield some impor-tant insight into the onset of immunogenicity. Although at the siteof antigen delivery [or immunotherapeutic delivery] a number ofdifferent cell types may play a role in the ensuing response, a strongcase can be made for the professional antigen-presenting cell, thedendritic cell.

Dendritic cells are strategically positioned throughout the bodywhere they serve as immune sentinels poised to respond toincoming pathogens at the interface between host tissues and theexternal environment [56,57]. Dendritic cells are highly phagocyticcells that continuously sample their microenvironment for anti-genic material to be processed and packaged into MHC moleculesfor eventual presentation to cognate lymphocytes. A number ofdendritic cell subsets have been identified in mice and humans,with varying capacities in priming CD4 and CD8 T cells, antigencross-presentation, as well as direct anti-viral function [i.e. plas-macytoid dendritic cells] [58]. Their distribution in secondarylymphoid tissues and at peripheral mucosal and non-mucosal sitesalso varies, rendering them accessible to incoming antigens atseveral portals of entry [57,58]. Dendritic cells are important inbridging innate and adaptive immune responses, and act to primeCD4/CD8 T cells as well as B cells [56]. Since immunogenicity arisesbecause antibodies have been generated against the immunother-apeutic agent, most attention has focused on B cell responses andhow tolerance was subverted at this level. Indeed, direct cross-

S.K. Singh et al. / Biologicals 40 (2012) 364e368 367

linking of the B cell receptor by antigens can bypass T-cell help [i.e.the T-independent pathway] and lead to low affinity and low titerantibody production [59]. However, the presence of high-affinity,neutralizing antibodies in numerous patients [1,60] infers T-cellhelp occurred and thus presumably cognate interactions between Tcells and antigen-presenting dendritic cells. The requirement forantigen-specific CD4 T-cell help in the ensuing B cell response alsoexposes a breakdown in tolerance at the level of CD4 T cells.Conventional wisdom would argue that self-reactive T cells aredeleted from the T-cell pool through negative selection in thethymus, and therefore how self-reactive CD4 T cells can still persistin the host, become activated, and provide help to B cells leading toantibody production remains a mystery. One important andpotential underlying cause of immunogenicity deals with howsimilar or dissimilar the immunotherapeutic agent is, compared to‘self’. The introduction of protein-based biologics has been metwith considerable obstacles over the last several decades, in largepart because of use non-human monoclonal antibodies, contami-nants and impurities, sequence variation, formulation, and proteinmodifications [10]. Since an important function of dendritic cells isto capture antigens through a number of different classes ofreceptors [Toll-Like receptors, heat shock proteins, lectins, immu-noglobulin superfamily, Fc receptors], there is an incentive toexplore this pathway as a possible trigger to immunogenicity[60,61]. Though differences in amino acid sequence betweenimmunotherapeutic agent and ‘self’ are heavily scrutinized, subtledifferences in post-translation modifications like glycosylation orthe formation of aggregates are now receiving their due notice, asdiscussed in the earlier talk by Dr. Satish Singh. Aggregates andimmune complexes, in particular, can cross-link and activate B cellsas well as be readily taken up by dendritic cells [20]. In this regard,understanding or ‘visualizing’ the initial events of dendritic cell-antigen encounter may uncover important clues into how ‘seem-ingly self’-based immunotherapeutic proteins are interrogatedin vivo. Recently, advances in intravital multiphoton imaging at thesingle-cell level have afforded us the possibility to ‘eavesdrop’ onthe cellular communication between leukocytes and to visualizethe complex interplay of physical, cellular, biochemical, and otherfactors that influence immune cell behavior [62]. Application of thissensitive imaging technology has recently been utilized to examinehost-pathogen [63] or host-antigen [64,65] interactions, includingimmune complexes [66,67] with dendritic cells. In this RoundtableSession, preliminary imaging of dendritic cell interactions withprotein aggregates and particulates were presented, and unveiledhow uniquely sensitive dendritic cells are in “interrogating”different formulations of the same parent protein. Sensing ofaggregated self-proteins [compared to non-aggregated self-proteins] by dendritic cells highlights the subtle differences inproteins that can be distinguished by dendritic cells. The moleculardetails governing this recognition are not well understood, but itwill be important to determine what specific protein modificationsare being recognized and how this translates into generatingdownstream immune responses. Understanding how immunoge-nicity arises at the level of dendritic cell sensing of protein-basedbiologics may provide important insight and lead to the rationaldesign of the next generation of safer protein-based biologics.

A question and answer session followed the presentations, whichled to an open and lively discussion between the attendees and thethree speakers. There was general agreement about the need toimprove analyticalmethods for detection/removal of impurities frombiotherapeutics preparations and formulations. Importance of theneed to assess,maintain and enhance integrity of the biotherapeuticswas also emphasized, especially in light of the rapidly growingnumbers and significance of ‘biosimilars’ and ‘biobetters’. Continuedimprovements in the in vitro aswell as in-silico tools for assessing and

predicting immunogenicitywasalsostressed.One issue thatgarneredconsiderable attentionwas the question of how to promote toleranceto the biotherapeutics. For example, the common practice used bymany immunologists of employing an alternative mode of exposure(e.g. nasal spray) of mice to the biotherapeutic under study (e.g. anantibody formulation) has been shown to be quite effective inreducing or even eliminating the subsequent immunogenicity to thesame biotherapeutic formulation [68].While the clinical utility of thisapproach may have to await further research, there was generalagreement among the attendees about the value of this strategy forpre-clinical studies in developing and testing new formulations ofexisting biologicals as well as during discovery and development ofnew biotherapeutics. Finally, need for a continued dialog amongresearchers from academia, industry and regulatory agencies wereemphasized. Participants also called for expanding the scope of suchdiscussions by including experts fromvariousfields of study i.e. cell &computational biology, immunology and clinical pharmacology toname a few. In conclusion, all participants commended theProgramming Committee of AAPS for providing a forum for thisimportant topic by arranging a successful Roundtable Session.

References

[1] Baker MP, Reynolds HM, Lumicisi B, Bryson CJ. Immunogenicity of proteintherapeutics: the key causes, consequences and challenges. Self Nonself 2010;1:314e22.

[2] Schellekens H. Bioequivalence and the immunogenicity of bio-pharmaceuticals. Nat Rev Drug Discov 2002;1:457e62.

[3] De Groot AS, Scott DW. Immunogenicity of protein therapeutics. TrendsImmunol 2007;28:482e90.

[4] Fakharzadeh SS, Kazazian HH. Correlation between factor VIII genotype andinhibitor development in hemophilia A. Semin Thromb Hemost 2000;26:167e71.

[5] Chamuleau ME, van de Loosdrecht AA, Huijgens PC. Monoclonal antibodytherapy in haematological malignancies. Curr Clin Pharmacol 2010;5:148e59.

[6] Prajapati DS, Shah JS, Dave JB, Patel CN. Therapeutic applications of mono-clonal antibodies. Syst Rev Pharm 2011;2:37e42.

[7] Niebecker R, Kloft C. Safety of therapeutic monoclonal antibodies. Curr DrugSaf 2010;5:275e86.

[8] Hansel TT, Kropshofer H, Singer T, Mitchell JA, George AJ. The safety and sideeffects of monoclonal antibodies. Nat Rev Drug Discov 2010;9:326e38.

[9] Southwood S, Sidney J, Kondo A, del Guercio MF, Appella E, Hoffman S, et al.Several common HLA-DR types share largely overlapping peptide bindingrepertoires. J Immunol 1998;160:3363e73.

[10] Singh SK. Impact of product-related factors on immunogenicity of bio-therapeutics. J Pharm Sci 2011;100:354e87.

[11] Clark M. Antibody humanization: a case of the emperor’s new clothes?Immunol Today 2000;21:397e402.

[12] Wadhwa M, Skog AH, Bird C, Ragnhammar P, Lilljefors M, Gaines-Das R, et al.Immunogenicity of granulocyte-macrophage colony-stimulating factor [GM-CSF] products in patients undergoing combination therapy with GM-CSF.Clin Cancer Res 1999;5:1353e61.

[13] Gribble EJ, Sivakumar PV, Ponce RA, Hughes SD. Toxicity as a result ofimmunostimulation by biologics. Expert Opin Drug Metab Toxicol 2007;3:209e34.

[14] Schellekens H. How to predict and prevent the immunogenicity of therapeuticproteins. Biotechnol Annu Rev 2008;14:191e202.

[15] Suntharalingam G, Perry MR, Ward S, Brett SJ, Castello-Cortes A, Brunner MD,et al. Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibodyTGN1412. N Engl J Med 2006;355:1018e28.

[16] Shankar G, Pendley C, Stein KE. A risk-based bioanalytical strategy for theassessment of antibody immune responses against biological drugs. Nat Bio-technol 2007;25:555e61.

[17] Gilliland LK, Walsh LA, Frewin MR, Wise MP, Tone M, Hale G, et al. Eliminationof the immunogenicity of therapeutic antibodies. J Immunol 1999;162:3663e71.

[18] Gribben JG, Devereux S, Thomas NSB, Keim M, Jones HM, Goldstone AH, et al.Development of antibodies to unprotected glycosylation sites on recombinanthuman GM-CSF. Lancet 1990;335:434e7.

[19] van Beers MMC, Jiskoot W, Schellekens H. On the role of aggregates in theimmunogenicity of recombinant human interferon beta in patients withmultiple sclerosis. J Interferon Cytokine Res 2010;30:767e75.

[20] Rosenberg AS. Effect of protein aggregates: an immunologic perspective. TheAPS J 2006;8:E501e7.

[21] Sauerborn M, Brinks V, Jiskoot W, Schellekens H. Immunological mechanismunderlying the immune response to recombinant human protein therapeu-tics. Trends Pharmacol Sci 2010;31:53e9.

S.K. Singh et al. / Biologicals 40 (2012) 364e368368

[22] Hermeling S, Crommelin DJA, Schellekens H, Jiskoot W. Structure-immuno-genicity relationships of therapeutic proteins. Pharm Res 2006;21:897e903.

[23] Fradkin AH, Carpenter JF, Randolph TW. Immunogenicity of aggregates ofrecombinant human growth hormone in mouse models. J Pharm Sci 2009;98:3247e64.

[24] Pisal DS, Kosloski MP, Middaugh CR, Bankert RB, Balu-Iyer SV. Native-likeaggregates of factor VIII [FVIII] are immunogenic in von Willebrand factordeficient and hemophilia A mice. J Pharm Sci 2012;101:2055e65.

[25] Carpenter JF, Randolph TW, Jiskoot W, Crommelin DJ, Middaugh CR, Winter G,et al. Overlooking subvisible particles in therapeutic protein products: gapsthat may compromise product quality. J Pharm Sci 2009;98:1202e5.

[26] Seidl A, Hainzl O, Richter M, Fischer R, Bohm S, Deutel B, et al. Tungsten-induced denaturation and aggregation of epoetin alfa during primary pack-aging as a cause of immunogenicity. Pharm Res 2012;29:1454e67.

[27] Liu DJ, Ren D, Huang H, Dankberg J, Rosenfield R, Cocco MJ, et al. Structure andstability changes of human IgG1 Fc as a consequence of methionine oxidation.Biochemistry 2008;47:5088e100.

[28] van Beers MM, Sauerborn M, Gilli F, Brinks V, Schellekens H, Jiskoot W.Oxidized and aggregated recombinant human interferon beta is immunogenicin human interferon beta transgenic mice. Pharm Res 2011;98:2393e402.

[29] Anderton SM. Post-translational modifications of self-antigens: implicationsfor autoimmunity. Curr Opin Immunol 2004;16:753e8.

[30] Doyle H, Mamula MJ. Post-translational protein modifications in antigenrecognition and autoimmunity. Trends Immunol 2001;22:443e9.

[31] Huang L, Lu J, Wroblewski VJ, Beals JM, Riggin RM. In vivo deamidation charac-terization of monoclonal antibody by LC/MS/MS. Anal Chem 2005;77:1432e9.

[32] Liu YD, van Enk JZ, Glynn GC. Human antibody Fc deamidation in vivo. Bio-logicals 2009;37:313e22.

[33] Thakrar K, Bodalia P, Grosso A. Assessing safety and efficacy of Omnitrope�. BrJ Clin Pharmacol 2010;2:298e301.

[34] Wang X, Hunter AK, Mozier ND. Host cell proteins in biologics development:identification, quantitation and risk assessment. Biotechnol Bioeng 2009;103:446e58.

[35] Krieg AM. CpG motifs in bacterial DNA and their immune effects. Annu RevImmunol 2002;20:709e60.

[36] Larsen ST, Lund RM, Nielsen GD, Thygesen P, Poulson OM. Di-[2-ethylhexyl]phthalate possess an adjuvant effect in a subcutaneous injection model withBALB/c mice. Toxicol Lett 2001;125:11e8.

[37] Hanawa T, Muramatsu E, Asakawa K, Suzuki M, Tanaka M, Kawano K, et al.Investigation of the release behavior of diethylhexylphthalate from thepolyvinyl-chloride tubing for intravenous administration. Int J Pharm 2000;210:109e15.

[38] Chiellini F, Ferri M, Latini G. Physical-chemical assessment of di-(2-ethylhexyl)-phthalate leakage from poly(vinyl chloride) endotracheal tubesafter application in high risk newborns. Int J Pharm 2011;409:57e61.

[39] Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Experimental and compu-tational approaches to estimate solubility and permeability in drug discoveryand development settings. Adv Drug Deliv Rev 1997;23:3e25.

[40] Kumar S, Singh SK, Wang X, Rup B, Gill D. Coupling of aggregation andimmunogenicity in biotherapeutics: T- and B-cell immune epitopes maycontain aggregation prone regions. Pharm Res 2011;28:949e61.

[41] Sturniolo T, Bono E, Ding J, Raddrizzani L, Tuereci O, Sahin U, et al. Generationof tissue-specific and promiscuous HLA ligand databases using DNA micro-arrays and virtual HLA class II matrices. Nat Biotechnol 1999;17:555e61.

[42] Koita OA, Dabitao D, Mahamadou I, Tall M, Dao S, Tounkara A, et al. Confir-mation of immunogenic consensus sequence HIV-1 T-cell epitopes in Bamako,Mali and Providence. Rhode Island Hum Vaccin 2006;2:119e28.

[43] De Groot AS, Bosma A, Chinai N, Frost J, Jesdale BM, Gonzalez MA, et al. Fromgenome to vaccine: in silico predictions, ex vivo verification. Vaccine 2001;19:4385e95.

[44] Bond KB, Sriwanthana B, Hodge TW, De Groot AS, Mastro TD, Young NL, et al.An HLA-directed molecular and bioinformatics approach identifies new HLA-

A11 HIV-1 subtype E cytotoxic T lymphocyte epitopes in HIV-1-infected Thais.AIDS Res Hum Retroviruses 2001;17:703e17.

[45] Dong Y, Demaria S, Sun X, Santori FR, Jesdale BM, De Groot AS, et al. HLA-A2-restricted CD8þ-cytotoxic-T-cell responses to novel epitopes in Mycobacte-rium tuberculosis superoxide dismutase, alanine dehydrogenase, and gluta-mine synthetase. Infect Immun 2004;72:2412e5.

[46] McMurry J, Sbai H, Gennaro ML, Carter EJ, Martin W, De Groot AS. AnalyzingMycobacterium tuberculosis proteomes for candidate vaccine epitopes.Tuberculosis 2005;85:95e105.

[47] CohenT,Moise L, ArditoM,MartinW,DeGroot AS. Amethod for individualizingtheprediction of immunogenicity of protein vaccines andbiologic therapeutics:individualized T cell epitope measure [iTEM]. J Biomed Biotechnol 2010.

[48] Moise L, Buller RM, Schriewer J, Lee J, Frey SE, Weiner DB, et al. VennVax,a DNA-prime, peptide-boost multi-T-cell epitope poxvirus vaccine, inducesprotective immunity against vaccinia infection by T cell response alone.Vaccine 2011;29:501e11.

[49] De Groot AS, Moise L. Prediction of immunogenicity for therapeutic proteins:state of the art. Curr Opin Drug Discov Devel 2007;10:332e40.

[50] De Groot AS, Moise L, McMurry JA, Wambre E, Van Overtvelt L, Moingeon P,et al. Activation of natural regulatory T cells by IgG Fc-derived peptide “tre-gitopes”. Blood 2008;112:3303e11.

[51] Nurieva RI, Liu X, Dong C. Molecular mechanisms of T-cell tolerance. ImmunolRev 2011;241:133e44.

[52] Mathis D, Benoist C. Back to central tolerance. Immunity 2004;20:509e16.[53] Janeway Jr CA. Approaching the asymptote? Evolution and revolution in

immunology cold spring Harb. Symp Quant Biol 1989;54:1e13.[54] Janeway Jr CA. The immune system evolved to discriminate infectious nonself

from noninfectious self. Immunol Today 1992;13:11e6.[55] Matzinger P. Essay 1: the danger model in its historical context. Scand J

Immunol 2001;54:4e9.[56] Banchereau J, Steinman RM. Dendritic cells and the control of immunity.

Nature 1998;392:245e52.[57] Alvarez D, Vollmann EH, von Andrian UH. Mechanisms and consequences of

dendritic cell migration. Immunity 2008;29:325e42.[58] Heath WR, Carbone FR. Dendritic cell subsets in primary and secondary T cell

responses at body surfaces. Nat Immunol 2009;10:1237e44.[59] Zubler RH. Naive and memory B cells in T-cell-dependent and T-independent

responses. Springer Semin Immunopathol 2001;23:405e19.[60] van Vliet SJ, García-Vallejo JJ, van Kooyk Y. Dendritic cells and C-type lectin

receptors: coupling innate to adaptive immune responses. Immunol Cell Biol2008;86:580e7.

[61] Savina A, Amigorena S. Phagocytosis and antigen presentation in dendriticcells. Immunol Rev 2007;219:143e56.

[62] Sumen C, Mempel TR, Mazo IB, von Andrian UH. Intravital microscopy:visualizing immunity in context. Immunity 2004;21:315e29.

[63] Junt T, Moseman EA, Iannacone M, Massberg S, Lang PA, Boes M, et al.Subcapsular sinus macrophages in lymph nodes clear lymph-borne virusesand present them to antiviral B cells. Nature 2007;450:110e4.

[64] Carrasco YR, Batista FD. B cells acquire particulate antigen in a macrophage-rich area at the boundary between the follicle and the subcapsular sinus ofthe lymph node. Immunity 2007;27:160e71.

[65] Roozendaal R, Mempel TR, Pitcher LA, Gonzalez SF, Verschoor A, Mebius RE,et al. Conduits mediate transport of low-molecular-weight antigen to lymphnode follicles. Immunity 2009;30:264e76.

[66] Phan TG, Grigorova I, Okada T, Cyster JG. Subcapsular encounter andcomplement-dependent transport of immune complexes by lymph node Bcells. Nat Immunol 2007;8:992e1000.

[67] Phan TG, Green JA, Gray EE, Xu Y, Cyster JG. Immune complex relay bysubcapsular sinus macrophages and noncognate B cells drives antibodyaffinity maturation. Nat Immunol 2009;10:786e93.

[68] Meritet JF, Maury C, Tovey MG. Induction of tolerance to recombinant ther-apeutic proteins. J Interferon Cytokine Res 2001;21:1031e8.