MHC-Associated Peptide Proteomics (MAPPs) as a Tool for Assessment of Immunogenicity Risk Potential of Therapeutic Monoclonal Antibodies

Biotherapeutic monoclonal antibodies (mAbs) can elicit unwanted immune responses in patients, potentially leading to anti-drug antibody (ADA) formation, reduced efficacy, or safety concerns. The major histocompatibility complex (MHC)-associated peptide proteomics (MAPPs) assay enables the identification of naturally presented human leucocyte antigen (HLA) class II peptides and provides a mechanistic readout of antigen processing and presentation. In this study, we applied a high-sensitivity MAPPs workflow using magnetic bead-based HLA-DR/DP/DQ immunoprecipitation and high-resolution mass spectrometry (HRMS) to assess the immunogenicity risk of six marketed mAbs across a panel of 10 HLA-typed donors. Peptide length distribution and MHC motif deconvolution were used as quality control measures. The number and clustering of drug-derived peptides were compared across antibodies and donors and correlated with reported clinical ADA incidence. The results highlight the value of MAPPs for early-stage candidate selection and HLA-aware risk stratification.

Biotherapeutics can provoke unwanted immunogenicity that may compromise the safety and efficacy of a drug. As such, immunogenicity risk potential assessments are increasingly integrated into early stages of drug development (1,2). The formation of anti-drug antibodies (ADAs) typically involves antigen processing and presentation by antigen-presenting cells (APCs), followed by activation of CD4+ T helper cells through recognition of biotherapeutic-derived peptides presented on HLA class II molecules. This interaction can contribute to the initiation of an adaptive immune response leading to the secretion of ADAs by B cells (Figure 1a) (3,4).

One in vitro assay that provides insight into this process is the MHC-associated peptide proteomics (MAPPs) assay. MAPPs identifies the naturally presented peptides that are generated after internalization and lysosomal processing of a protein by APCs and subsequently bound to MHC class II molecules (in humans HLA class II) (5,6). As peptide presentation is a prerequisite for eliciting a T-cell response, the MAPPs assay is a powerful tool to assess the immunogenicity potential of protein therapeutics. Unlike in silico prediction algorithms or HLA-binding assays, MAPPs reflects the full intracellular processing landscape and reveals which peptide fragments are actually presented on the cell surface, taking into account protein folding, post-translational modifications, and cellular proteolytic activity (4,7,8).

Several studies have shown that the number of drug-derived peptides presented, and the number of sequence regions (“clusters”) in which they occur, correlate approximately with clinical ADA incidence rates (4,6,9,10). MAPPs has also been used to rank similar biotherapeutic candidates, investigate mechanistic causes of immunogenicity, and identify or confirm immunodominant T cell epitopes (6,9,11,12). As reviewed by Karle and others, MAPPs has emerged as one of the best-characterized in vitro tools for preclinical immunogenicity risk potential assessment (4).

In the present study, we applied a high-sensitivity MAPPs workflow using magnetic beads coupled with a mixture of monoclonal antibodies targeting HLA-DR, HLA-DP, and HLA-DQ molecules, the three main human MHC class II molecules. Monocyte-derived dendritic cells (moDCs) from 10 HLA-typed healthy donors were loaded with up to six marketed monoclonal antibodies: trastuzumab, nivolumab, atezolizumab, infliximab, adalimumab, and ustekinumab. Peptides were eluted from affinity-purified HLA class II molecules and analyzed via liquid chromatography tandem mass spectrometry (LC–MS/MS)(Figure 1b). Identified peptides were annotated, compared across donors, and analyzed with respect to peptide length, clustering, and predicted HLA binding motifs. The results were interpreted in light of published clinical ADA incidence data to explore the correlation between naturally presented peptide profiles and known immunogenicity. This study further supports the utility of the MAPPs assay for the preclinical evaluation and ranking of biotherapeutic candidates based on their immunogenicity risk.

Materials and Methods

Donor Selection and moDC Generation

Whole blood and leukapheresis samples were collected from healthy donors as described in the ethical protocol/amendment IXP-009_V1 (Belgium; Reg. Nr. B0792020000009) or license number DE_TH_01H_MIA_2022_0012 (Germany; Ref. Nr. 24-2528.03-003). Peripheral blood mononuclear cells (PBMCs) were obtained from 10 healthy donors covering the most common HLA-DRB1 alleles. Monocytes were isolated using CD14 magnetic selection (Miltenyi Biotec) and differentiated into immature monocyte-derived dendritic cells (moDCs) over five days in the presence of IL-4 and GM-CSF. On day 5, moDCs were pulsed with 0.33 µM of the different test monoclonal antibodies (trastuzumab, nivolumab, atezolizumab, infliximab, adalimumab, or ustekinumab) and matured by addition of lipopolysaccharide (10 µg/mL) overnight.

MHC Class II Peptide Isolation

On day 6, cells were harvested, washed twice with PBS, and stored at –80 °C. For each sample, 1 million mature moDCs were lysed in a non-denaturing lysis buffer supplemented with protease inhibitors (Roche). Lysates were cleared by centrifugation and pre-cleared using control beads. HLA class II:peptide complexes were immunoprecipitated using magnetic beads conjugated with a mixture of monoclonal antibodies against HLA-DR (clone L243), HLA-DP (clone B7.21), and HLA-DQ (clone 1a3) (pan-HLA class II capture) in a KingFisher Apex magnetic robot (Thermo Fisher). Following extensive washing, bound peptides were eluted with mild acid and purified by C18-based solid-phase extraction (SPE).

LC–MS/MS Analysis

Peptides were analyzed on a TimsTOF SCP mass spectrometer (Bruker) coupled to an Evosep One LC system (Evosep Biosystems). The DDA PASEF method was used with one TIMS-MS survey scan and 10 PASEF MS/MS scans per cycle. The ion mobility scan range was 0.7–1.4 Vs/cm², and precursor mass range was 100–1700 m/z.

MS Data Analysis

Peptides were identified using Peaks Online 11 (Bioinformatics Solutions Inc.) against a concatenated FASTA file containing the UniProt human proteome and sequences of the six test monoclonal antibodies. A contaminant database including reagent-derived proteins was also included. Searches allowed for unspecific cleavage and included methionine oxidation and deamidation as variable modifications. A 1% false discovery rate (FDR) threshold was applied. Peptides ranging from 9 to 30 amino acids were considered for downstream analysis.

Quality Control and Motif Analysis

Peptide length distribution histograms were generated for each donor to assess enrichment of HLA class II-presented peptides. Motif deconvolution was performed using the MHC Motif Decon tool (13), applying NetMHCIIpan-4.3 (14) to assign likely HLA class II restrictions based on each donor’s genotype. Only high-confidence peptides mapping to the test mAb sequences were annotated as drug-derived.

Results

Peptide Yield and Quality Control

All samples yielded high numbers of HLA class II-bound peptides, ranging from 7437 to 13,595 unique peptides per donor-mAb combination (see Table I). Total peptide spectrum matches (PSMs) ranged from 15,910 to 28,501 per sample. Peptide length distributions displayed the expected unimodal peak centered at 15–16 amino acids across all donors and mAbs (Figure 2a), consistent with canonical HLA class II presentation. Motif deconvolution confirmed strong alignment with donor-specific HLA-DR, -DP, and -DQ allele motifs (Figure 2b), supporting the specificity and quality of peptide enrichment.

Identification of Drug-Derived Peptides

Drug-derived peptides were identified for all six tested mAbs with notable variability in abundance, sequence coverage, and clustering across donors. Infliximab and adalimumab, both known to be highly immunogenic clinically, yielded the most extensive peptide presentation (Figure 3). For infliximab (analyzed in six donors, Figure 3a), dominant peptide clusters were observed in the heavy chain framework region between CDR2 and CDR3 and in the CDR3 region, with consistent clustering across donors D, E, H, and I. Adalimumab (analyzed in four donors, Figure 3b) also demonstrated strong clustering in the heavy chain framework region between CDR2 and CDR3 (in donors D and E), and to a lesser extent in the CDR2 region and light chain CDR2.

Ustekinumab showed high presentation in donors I and J (the only two donors in which this monoclonal antibody was loaded, Figure 3c), with overlapping clusters in the heavy chain CDR3.

Trastuzumab (loaded in nine donors, Figure 3d) presented a more limited set of peptides. Most drug-derived peptides were identified in the heavy chain framework region between CDR2 and CDR3 and to a lesser extent in light chain CDR2. Presentation was consistent in donors B, C, and F, while absent in donor G.

Nivolumab (loaded in all 10 donors, Figure 3e) yielded the lowest number of peptides with no heavy chain CDR-derived peptides, while only donors D and E showed dense overlapping peptide clusters in the light chain CDR2. Donors F and G showed no drug-derived peptides. This pattern with limited CDR-derived peptides is consistent with the antibody’s low clinical ADA incidence.

Atezolizumab-derived peptides were identified in all seven tested donors (A, C, D, E, H, I, J; Figure 3f), with limited peptides identified in the CDR3 and clustering mainly localized in the framework region between CDR2 and CDR3 and the C-terminal region. Light chain peptides were observed in CDR2 in donor D. These patterns suggest moderate presentability with donor-specific variation.

Overall, the distribution of presented peptides mapped predominantly to both germline (framework regions) and non-germline regions, including CDR2 and CDR3 of the heavy chain and CDR2 of the light chain. These sequence regions are more likely to escape central tolerance and have been repeatedly implicated in ADA responses in both in vitro and clinical data sets.

Correlation with Clinical Immunogenicity

A qualitative ranking of drug-derived peptide loads correlated well with reported clinical ADA incidence. Infliximab showed a wide range of reported ADA incidence rates depending on indication, patient population, and treatment regimen (typically 25–60% ADA without co-administration of immunomodulators) (15), just like adalimumab, for which ADA incidence was generally reported in the 10–40% range depending on indication and treatment regimen (16). Both these therapeutic monoclonal antibodies, associated with high clinical ADA rates, had the greatest number of peptides and cluster regions identified by MAPPs. Ustekinumab and atezolizumab, with moderate ADA incidence (~10–20%) (17,18), showed intermediate presentation. Trastuzumab and nivolumab, with low ADA rates (<5%) (19,20), consistently presented the lowest number of drug-derived peptides.

These results align with recent findings from ADA surveillance studies, confirming the immunodominance of CDR-derived peptides and their relevance for preclinical evaluation. For example, infliximab has previously been shown to elicit strong CD4+ T cell responses corresponding to CDR2 and framework regions (10, 21), and peptides derived from these regions were mapped in our data set. Conversely, trastuzumab and nivolumab, despite being widely administered, showed limited T cell activation potential, which was consistent with the sparse peptide presentation observed.

These findings support the utility of MAPPs in stratifying immunogenicity risk potential based on naturally processed peptide visibility and highlight its complementarity with clinical ADA data and T cell assays. The data suggest that MAPPs-derived presentation profiles, particularly of non-germline CDR regions, can be used as an early indicator of ADA liability and inform rational de-immunization strategies or candidate selection.

Discussion

This study demonstrates the utility of the MHC-associated peptide proteomics (MAPPs) assay, which models key immunological processes including dendritic cell-mediated uptake of biotherapeutics, intracellular antigen processing, and MHC class II peptide presentation. In this study, we specifically focused on MHC class II MAPPs, as we investigated the immunogenicity risk potential of therapeutic proteins taken up and processed by professional antigen-presenting cells. In contrast, MHC class I MAPPs were more applicable to assessing the immunogenicity risk of novel biotherapeutic modalities that result in intracellular protein expression and in situ MHC class I presentation following in vitro translation, such as mRNA-based therapeutics, viral vectors, or DNA constructs (22).

By directly identifying naturally presented HLA class II peptides from moDCs loaded with six marketed mAbs, we observed distinct differences in peptide yield, clustering, and inter-donor variability that aligned closely with known clinical ADA incidence rates. The correlation between the number and diversity of drug-derived peptides and clinical immunogenicity data underscores the biological relevance of the MAPPs assay. Infliximab and adalimumab, which are known to be highly immunogenic in patients, yielded the highest number of unique peptides across donors and showed consistent clustering in antigenically relevant regions. Conversely, trastuzumab and nivolumab exhibited low and inconsistent peptide presentation, matching their low clinical ADA incidence. This alignment supports the notion that natural antigen processing and HLA-mediated presentation captured by the MAPPs assay reflect key early steps in the anti-drug immune response.

A major advancement highlighted in this study is the increased sensitivity and practicality of the MAPPs assay, driven by recent improvements in both mass spectrometry hardware and immunopeptidomics data analysis. Unlike classical proteomics workflows, which rely on well-defined cleavage by proteases such as trypsin, immunopeptidomics deals with naturally processed peptides of variable lengths and cleavage sites. This inherent complexity historically limited sensitivity, requiring large cell numbers and long acquisition times. However, the adoption of high-resolution instruments, combined with improved bioinformatics tools tailored to non-tryptic peptide identification, now enables confident detection and quantification of naturally presented peptides from as few as one million moDCs per condition.

This enhanced sensitivity is crucial for making MAPPs a feasible tool in early drug development workflows, where material availability is often limited. In this study, hundreds to thousands of unique HLA class II-bound peptides per sample were detected, including relevant drug-derived peptides, even with low input material. These advances reduce technical barriers and establish MAPPs as a practical and scalable immunogenicity assessment tool.

In parallel with increased analytical sensitivity, the robustness of the sample preparation workflow remains a critical determinant of data quality in MAPPs assays. Because the goal is to comprehensively profile all peptides bound within the MHC class II cleft at the time of sampling, any losses during cell lysis, solubilization, or immunoprecipitation can introduce significant bias. In contrast to proteomic workflows, where missed cleavages can be algorithmically tolerated, the omission of certain naturally processed peptides in immunopeptidomics can lead to an incomplete or distorted picture of T cell epitope presentation. Therefore, meticulous attention to solubilization efficiency, detergent removal, and peptide recovery was essential.

Loss of MHC-peptide complexes or displacement of peptides from the binding groove can occur if conditions are too harsh or inconsistent, potentially obscuring immunodominant epitopes or underestimating peptide cluster breadth. In our workflow, careful optimization of buffer systems, detergent stringency, and magnetic bead-based affinity capture helped ensure that peptide recovery was both efficient and reproducible. These refinements are especially important when working with small sample volumes and further support the feasibility of applying MAPPs as a screening tool in early-phase biotherapeutic development.

Equally critical is the quality of the dendritic cells used for antigen processing and presentation. The differentiation state, viability, and activation profile of moDCs directly influence their capacity to internalize, process, and present antigens via HLA class II molecules. Variability in dendritic cell quality can impact peptide yield and the diversity of presented epitopes, potentially masking true immunogenic hotspots or introducing donor-specific bias. In our experience, rigorous control of moDC culture conditions and standardized maturation protocols are essential to ensure consistent, high-quality APC function and reproducible MAPPs results.

Taken together, our results demonstrate that MAPPs can support data-driven ranking of candidate biotherapeutics based on their HLA class II presentation profiles. The ability to compare different molecules across a diverse donor cohort enables early identification of sequence regions prone to broad HLA-mediated presentation, and thereby potential T cell recognition. This information can guide de-immunization strategies, inform lead candidate selection, and help flag HLA-linked immunogenicity risks before entering clinical trials. Furthermore, MAPPs can be integrated with complementary in vitro T cell assays, where the naturally presented peptides identified here can be tested for functional relevance using proliferation or cytokine readouts.

One of the most effective strategies for immunogenicity risk assessment may be to begin with a MAPPs analysis to identify naturally presented peptides, particularly those derived from non-germline regions. These candidate epitopes can then be subjected to follow-up T cell assays to confirm their ability to activate CD4+ T cells (23). This stepwise approach combines the mechanistic strength of natural processing and HLA presentation with functional readouts of immunogenic potential, providing a robust and biologically grounded method for preclinical risk evaluation.

The combination of natural presentation profiling and functional T cell confirmation provides a powerful and mechanistically grounded immunogenicity risk potential assessment platform. In addition to identifying presented regions, the MAPPs assay enables classification of peptides as either germline-derived or non-germline (that is, product-specific sequence variants such as CDRs). This distinction is critical as non-germline regions are more likely to escape central tolerance and thus elicit T cell responses. While some germline-derived peptides may still be recognized, particularly in allotypic mismatch scenarios or under inflammatory conditions, non-germline sequences typically present a higher immunogenic risk. In this study, most dominant peptide clusters mapped to non-germline regions, particularly within the CDR2 and CDR3 domains, reinforcing their relevance as candidate T cell epitopes. For non-germline regions identified as naturally presented by HLA class II molecules, it is essential to investigate whether they can elicit CD4+ T cell responses in vitro (10,4), which would further support their classification as immunodominant epitopes and inform risk mitigation strategies. Such findings may also guide targeted de-immunization approaches, such as sequence modification in CDRs to reduce HLA binding potential while preserving antigen specificity (24, 25).

Conclusion

The results of this study have demonstrated that MAPPs is a practical, sensitive, and informative method to evaluate the immunogenicity potential of therapeutic monoclonal antibodies. Using a high-resolution workflow, naturally presented HLA class II peptides from six marketed mAbs were detected across a diverse panel of 10 donors. The data revealed distinct, reproducible peptide clusters—predominantly in non-germline regions such as CDR2 and CDR3—that aligned with known clinical immunogenicity profiles. These insights suggest that MAPPs-derived peptide visibility, particularly in the context of HLA diversity, can inform candidate selection, sequence optimization, and risk mitigation in early-stage biologic development.

The ability to perform these analyses with limited cell input broadens the assay’s utility in standard preclinical workflows. We propose a tiered approach in which MAPPs is used to identify presented epitopes, followed by T cell assays to confirm their immunogenic potential. This strategy provides a robust, mechanism-based platform for ADA risk assessment and supports the rational design of safer and more effective biologics.

References

1. Karle, A. C. Applying MAPPs Assays to Assess Drug Immunogenicity. Front. Immunol. 2020,11, 698. DOI: 10.3389/fimmu.2020.00698
2. De Groot, A. S.; Moise, L. Prediction of Immunogenicity for Therapeutic Proteins: State of the Art. Curr. Opin. Drug Discov. Devel.2007, 10 (3), 332–340.
3. Rosenberg, A. S. Effects of Protein Aggregates: An Immunologic Perspective. AAPS J. 2006,8 (3), E501–E507. DOI: 10.1208/aapsj080359
4. Karle, A. C. MAPPs Assay: Applications, Mechanistic Insights and Impact on Understanding Biotherapeutic Immunogenicity. Front. Immunol. 2020, 11, 698. DOI: 10.3389/fimmu.2020.00698
5. Quarmby, V.; Phung, Q. T.; Lill, J. R. MAPPs for the Identification of Immunogenic Hotspots of Biotherapeutics. Expert Rev. Proteomics2018,15 (9), 733–748. DOI: 10.1080/14789450.2018.1521279
6. Karle, A.; Spindeldreher, S.; Kolbinger, F. Secukinumab, a Novel Anti–IL-17A Antibody, Shows Low Immunogenicity Potential in Human In Vitro Assays. mAbs2016, 8 (3), 536–550. DOI: 10.1080/19420862.2015.1136761
7. Faridi, P.; Purcell, A. W.; Croft, N. P. In Immunopeptidomics We Need a Sniper Instead of a Shotgun. Proteomics2018, 18 (5–6), e1700464. DOI: 10.1002/pmic.201700464
8. Bassani-Sternberg, M.; Bräunlein, E.; Klar, R.; et al. Direct Identification of Clinically Relevant Neoepitopes Presented on Native Human Melanoma Tissue by Mass Spectrometry. Nat. Commun. 2016, 7, 13404. DOI: 10.1038/ncomms13404
9. Meunier, S.; Hamze, M.; Karle, A.; et al. Impact of Human Sequences in Variable Domains of Therapeutic Antibodies on CD4 T-Cell Epitopes. Cell. Mol. Immunol.2020, 17 (4), 656–658. DOI: 10.1038/s41423-019-0304-3
10. Hamze, M.; Meunier, S.; Karle, A.; et al. Characterization of CD4 T Cell Epitopes of Infliximab and Rituximab. Front. Immunol.2017, 8, 500. DOI: 10.3389/fimmu.2017.00500
11. Cassotta, A.; Mikol, V.; Bertrand, T.; et al. A Single T Cell Epitope Drives the Neutralizing Anti-Drug Antibody Response to Natalizumab in Multiple Sclerosis Patients. Nat. Med. 2019, 25 (9), 1402–1407. DOI: 10.1038/s41591-019-0568-2
12. Spindeldreher, S.; Karle, A.; Correia, E.; et al. T Cell Epitope Mapping of Secukinumab and Ixekizumab in Healthy Donors. mAbs 2020, 12 (1), 1707418. DOI: 10.1080/19420862.2019.1707418
13. Kaabinejadian, S.; Barra, C.; Alvarez, B.; Nielsen, M. Accurate MHC Motif Deconvolution of Immunopeptidomics Data Reveals a Significant Contribution of DRB3, 4 and 5 to the Total DR Immunopeptidome. Front. Immunol. 2022, 13, 835454. DOI: 10.3389/fimmu.2022.835454
14. Reynisson, B.; Alvarez, B.; Paul, S.; Peters, B.; Nielsen, M. NetMHCpan-4.1 and NetMHCIIpan-4.0: Improved Predictions of MHC Antigen Presentation by Concurrent Motif Deconvolution and Integration of MS MHC Eluted Ligand Data. Nucleic Acids Res. 2020, 48 (W1), W449–W454. DOI: 10.1093/nar/gkaa379
15. Ben-Horin, S.; Chowers, Y. Review Article: Loss of Response to Anti-TNF Treatments in Crohn’s Disease. Aliment. Pharmacol. Ther.2011,33 (9), 987–995. DOI: 10.1111/j.1365-2036.2011.04612.x
16. Strand, V.; Balsa, A.; Al-Saleh, J.; et al. Immunogenicity of Biologics in Chronic Inflammatory Diseases: A Systematic Review. BioDrugs2017, 31 (4), 299–316. DOI: 10.1007/s40259-017-0233-4
17. Blauvelt, A.; Reich, K.; Tsai, T. F.; et al. Secukinumab Is Superior to Ustekinumab in Clearing Skin in Psoriasis. N. Engl. J. Med. 2017, 376 (14), 1311–1320. DOI: 10.1056/NEJMoa1615972
18. Herbst, R. S.; Soria, J. C.; Kowanetz, M.; et al. Predictive Correlates of Response to the Anti-PD-L1 Antibody Atezolizumab in NSCLC. Nature2014, 515 (7528), 563–567. DOI: 10.1038/nature14011
19. Jackisch, C.; Kim, S.-B.; Semiglazov, V.; Melichar, B.; Pivot, X.; Hillenbach, C.; Stroyakovskiy, D.; Lum, B. L.; Elliott, R.; Weber, H. A.; Ismael, G. Subcutaneous versus Intravenous Formulation of Trastuzumab for HER2-Positive Early Breast Cancer: Updated Results from the Phase III HannaH Study. Ann. Oncol. 2015,26 (2), 320–325. DOI: 10.1093/annonc/mdu524
20. Weber, J. S.; D’Angelo, S. P.; Minor, D.; et al. Nivolumab versus Chemotherapy in Patients with Advanced Melanoma Who Progressed after Anti–CTLA-4 Treatment (CheckMate 037). Lancet Oncol. 2015, 16 (4), 375–384. DOI: 10.1016/S1470-2045(15)70076-8
21. Barra, C.; Ackaert, C.; Reynisson, B.; et al. Immunopeptidomic Data Integration to Artificial Neural Networks Enhances Protein-Drug Immunogenicity Prediction. Front. Immunol. 2020, 11, 1304. DOI: 10.3389/fimmu.2020.01304
22. Bing, S. J.; Justesen, S.; Wu, W. W.; et al. Differential T Cell Immune Responses to Deamidated Adeno-Associated Virus Vector. Mol. Ther.2022,24, 255–267. DOI: 10.1016/j.omtm.2022.01.005
23. Harris, C. T.; Cohen, S. Reducing Immunogenicity by Design: Approaches to Minimize Immunogenicity of Monoclonal Antibodies. BioDrugs 2024, 38 (2), 205–226. DOI: 10.1007/s40259-023-00641-2
24. Baker, M. P.; Jones, T. D. Identification and Removal of Immunogenicity in Therapeutic Proteins. Curr. Opin. Drug Discov. Devel.2007, 10 (2), 219–227.
25. Harding, F. A.; Stickler, M. M.; Razo, J.; DuBridge, R. B. The Immunogenicity of Humanized and Fully Human Antibodies: Residual Immunogenicity Resides in the CDR Regions. mAbs2010, 2 (3), 256–265. DOI: 10.4161/mabs.2.3.11641

About the Column Editor

About the Authors