Targeting the Leukemia Antigen PR1 with Immunotherapy for the Treatment of Multiple Myeloma

Despite therapeutic advances, multiple myeloma remains an incurable disease. Patients with high-risk disease features have a median survival of approximately 3 years (1). While immunotherapy is currently not part of the standard regimens for the management of multiple myeloma, the role of immunotherapy and the immune system has been clearly demonstrated in multiple myeloma with allogeneic stem cell transplantation (allo-SCT), a prime example of immunotherapy and currently the only available curative option for multiple myeloma patients. However, the high treatment-related mortality associated with allo-SCT, in many cases, outweighs the graft-versus-myeloma (GvM) effect, and may limit the use of allo-SCT to selected patients who have failed standard-of-care treatments.

Instead of infusing a large number of cells that contain only a small fraction of cytotoxic T lymphocytes (CTL) with multiple myeloma specificity, thereby risking nonspecific reactivity of the infused cells with normal tissue as in the case with allo-SCT, a few studies have demonstrated promising results with immunotherapies that target distinct antigens that are expressed by multiple myeloma or with cellular immunotherapies using marrow-infiltrating lymphocytes (2–4). Because of the heterogeneity of antigen expression by multiple myeloma cells and immune evasion mechanisms (5), identifying and targeting novel antigens is critical to the success of curative immunotherapy for multiple myeloma.

PR1 (VLQELNVTV) is a human leukocyte antigen (HLA)-A*0201-specific nonameric peptide derived from the myeloid-restricted primary granule proteins (PGP) neutrophil elastase (NE) and proteinase 3 (P3; ref. 6). NE and P3 are normally expressed by myeloid progenitor cells, polymorphonuclear leukocytes (PMNs) and monocytes, and are overexpressed or aberrantly expressed in acute myeloid leukemia (AML) and chronic myelogenous leukemia (CML), making them ideal leukemia associated antigens (LAA; refs. 6–8). We discovered PR1 on the surface of AML and showed that CTL immunity to PR1 contributes to cytogenetic remission in patients with CML after allo-SCT (9, 10). We first developed a PR1-peptide vaccine and showed that it can induce immunity and objective clinical responses in patients with relapsed/refractory AML, CML, and myelodysplastic syndrome (MDS; ref. 11). As vaccines have shown the highest antitumor efficacy in the setting of minimal disease burden (12), we also developed a T-cell receptor (TCR)-like mAb, 8F4, which targets PR1/HLA-A2 for use in the high tumor burden setting. 8F4 has demonstrated anti-leukemia activity in vitro and in vivo against myeloid leukemia cell lines and primary samples from AML patients (13), and is currently being developed for clinical trials.

In addition to myeloid leukemia, published reports have shown that nonmyeloid tumors lacking endogenous P3 and NE expression (i.e., lack PR1) are able to take up NE and P3 from the extracellular environment (14–18). Furthermore, we showed that breast cancer, melanoma, lung cancer, and malignant lymphoid cells cross-present NE and P3 (i.e., present PR1 on surface HLA-A2), and are rendered susceptible to PR1-CTL and 8F4 (14, 15, 19). Cross-presentation is a mechanism predominantly found in antigen-presenting cells (APC), whereby exogenous antigens are taken up, processed, and presented on HLA class I to prime an immune response (20).

As a natural extension of our prior work and because (i) multiple myeloma originates from B cells, which are known APCs (15, 21, 22); and (ii) myeloid progenitor cells and PMNs (the sources for NE and P3) are abundantly found in the bone marrow and may play a role in the multiple myeloma microenvironment (6, 23), we hypothesized that multiple myeloma cross-presents NE and P3 and subsequently becomes susceptible to PR1-targeting immunotherapies.

Here we report that multiple myeloma lacks endogenous expression of NE and P3. We show that multiple myeloma takes up NE and P3 and cross-presents PR1 in the context of HLA-A2 on the cell surface. We demonstrate that PR1 cross-presentation leads to specific killing of multiple myeloma by PR1-CTL and 8F4 in vitro and in vivo. Finally, we show that PR1 is expressed on the surface of multiple myeloma cells from primary HLA-A2⁺ patient samples. Our data point to a potential role for PR1-targeting immunotherapies for patients with multiple myeloma and highlight the importance of cross-presented antigens as potential clinical targets for immunotherapy.

The Arkansas (ARK), LP-1, ARP-1, IM-9, OPM-2, RPMI 8226, and U266 multiple myeloma cell lines, U937 histiocytic leukemia, T2 T/B cell hybridoma, H2023 lung cancer, and T-47D and MDA-MB-453 breast cancer cell lines were obtained from ATCC. Cell lines were grown in RPMI1640 media with 25 mmol/L HEPES + l-glutamine (Hyclone) supplemented with 10% FBS (Gemini Bio-Products), 100 U/mL penicillin, and 100 μg/mL streptomycin (Cellgro). All cell lines were cultured and maintained in 5% CO₂ at 37°C. Cell lines were validated using short tandem repeat DNA fingerprinting by our institutional sequencing facility.

mRNA was purified using RNA Stat 60 kit (TelTest). cDNA was synthesized using the Gene AMP RNA kit (Perkin Elmer). The following primers were used: NE forward primer 5′-CACGGAGGGGCAGAGACC-3′ and reverse primer 5′-TATTGTGCCAGATGCTGGAG-3′; P3 forward primer 5′-GACCCCACCATGGCTCAC-3′, and reverse primer 5′-ATGGGAAGGACAGACAGGAG-3′; actin forward primer 5′-CCAGAGCAAGAGAGCTATCC-3′ and reverse primer 5′-CTGTGGTGGTGAAGCTGTAG-3′ (14, 18). cDNA amplification was performed using an iCycler (Bio-Rad). Samples were separated on a 1.5% agarose gel. Bands were visualized using GelDoc2000 (Bio-Rad) and analyzed using Quantity One software (Bio-Rad).

Cell lysates were generated by resuspending cell pellets at 4°C for 30 minutes in lysis buffer [10 mmol/L HEPES (pH 7.9), 10 mmol/L KCl, 0.1 mmol/L EGTA 0.1 mmol/L EDTA, 1 mmol/L DTT] containing protease inhibitors (Thermo Fisher Scientific). Protein was loaded on 10% SDS gels (Bio-Rad), separated by electrophoresis under reducing conditions, transferred onto polyvinylidene fluoride membranes, and blocked with 5% milk. Membranes were probed using anti-NE (Santa Cruz Biotechnology), anti-P3 (NeoMarkers), anti-actin (Millipore) antibodies, and peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch). Chemiluminescence was captured on Kodak film (Kodak).

Cells were cultured with 10 μg NE, P3 (Athens Research & Technology), EndoGrade ovalbumin (Ova; Hyglos) or with irradiated (7500 cGy) PMNs or peripheral blood mononuclear cells (PBMCs; 1:1 ratio) at varying durations. Cells were permeabilized by washing in BD Biosciences Perm/Wash buffer and stained with fluorochrome-conjugated anti-P3 (Clone MCPR3-2; Thermo Scientific) or anti-NE (Santa Cruz Biotechnology) antibodies using Alexa Fluor-488 or -647 conjugation kits from Invitrogen. Cross-presentation of peptides was detected by staining cell surface with fluorescently conjugated, anti-PR1/HLA-A2 antibody (clone 8F4) as previously described (14, 15, 19). Costimulatory molecule surface expression was analyzed by staining myeloma patient bone marrow for CD19, CD33, CD34, CD38, CD138, CD40, CD80, CD86, and HLA-DR (all from Biolegend). Data were analyzed using FlowJo software (Tree Star). Inhibition of cross-presentation was accomplished by treating cell cultures with the endoplasmic reticulum (ER) to Golgi antegrade inhibitor brefeldin A (Sigma-Aldrich), or proteasome inhibitors lactacystin (Sigma-Aldrich) or bortezomib (Millennium Pharmaceuticals). To determine the effects of IMiDs on cross-presentation, cell cultures were treated with lenalidomide (Celgene).

PR1-specific CTL were generated from HLA-A2–positive healthy donor (HD) PBMC by stimulating with PR1 peptide (Bio-synthesis Inc.) in vitro, as previously described (14, 24). Briefly, PBMC from healthy donor leukapheresis were isolated using Histopaque 1077 gradient centrifugation (Sigma-Aldrich) and were cocultured with PR1 (20 μg/mL)-pulsed T2 cells at 1:1 ratio in RPMI1640 media supplemented with 10 % human AB serum (Gemini Bio-Products). Cell cultures were restimulated with PR1-pulsed T2 cells on days 7, 14, and 21, and 20 IU/mL of recombinant human IL2 (rhIL2; Biosource International).

Cytotoxicity assays were performed as previously described (14, 18). In brief, 1 × 10³ target cells/mL were fluorescently labeled with calcein-AM (Invitrogen) for 15 minutes at 37°C and thoroughly washed with RPMI1640 to remove free calcein-AM. Target cells were cocultured with peptide-specific CTL at the indicated effector-to-target (E:T) ratios for 4 hours at 37°C in 60-well Terasaki plates. Trypan blue was added to each well to stop the reaction and fluorescence was detected on a CytoFluor II plate reader (Applied Biosystems). The percent specific cytotoxicity was calculated as follows:

Complement-mediated cytotoxicity (CDC) assays were performed as previously described (13, 14). Briefly, U266 cells were cultured with NE or P3 for 24 hours and stained with calcein-AM. Labeled cells (1 × 10⁶) were resuspended in serum-free RPMI1640 media and treated with anti-PR1/HLA-A2 (clone 8F4) antibody or isotype control antibody for 10 minutes at 37°C. Standard rabbit complement (C') (Cedarlane Labs) was added and cells were incubated for 60 minutes at 37°C. Fluorescence was measured as described in the previous section.

Patient and HD peripheral blood (PB) samples and bone marrow aspirates were collected after informed consent to participate in an MD Anderson Cancer Center (MDACC) institutional review board-approved study. Peripheral blood mononuclear cells (PBMC) were isolated using Histopaque 1077 gradient centrifugation (Sigma-Aldrich). PBMC were stained using the following fluorescent antibodies: CD8 APC-H7 (BD Biosciences), CD3 FITC (BD Biosciences), PE-conjugated PR1/HLA-A2-dextramer (Immudex) or tetramer (Baylor College of Medicine MHC Tetramer Core, Houston, TX) and the following Pacific blue–conjugated lineage antibodies: CD4 (BD Biosciences), CD14 (BD Biosciences), CD16 (BD Biosciences), and CD19 (Biolegend). Samples were fixed with 4% paraformaldehyde. Data were acquired on a FACSCanto flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star). The frequency of PR1-CTL was determined as the percentage of live cells that were lineage⁻, CD3⁺, CD8⁺, and PR1-dextramer⁺ or PR1-tetramer⁺. Phenotype of PR1-CTL (PR1/HLA-A2-dextramer⁺) was investigated using CCR7 PE-Cy7 (Biolegend) and CD45RA PerCP-Cy5.5 (Biolegend) staining, and was classified as central memory (CCR7⁺/CD45RA⁻), effector memory (CCR7⁻/CD45RA⁻), naïve (CCR7⁺/CD45RA⁺), or terminally differentiated (CCR7⁻/CD45RA⁺).

Bone marrow smears and U266 cells were fixed with cold acetone and blocked with 5% normal mouse serum (Jackson ImmunoResearch). Fixed slides were washed with PBS and then double stained with Alexa-647–conjugated 8F4 antibody and Alexa-488–conjugated mouse anti-human HLA-A2 antibody (Serotec) or Alexa-488–conjugated rabbit anti-CD138 antibody (Bioss). Slides were stained with antibodies for 90 minutes at room temperature. After washing, ProLong Gold antifade reagent with DAPI (Invitrogen) was added. Confocal imaging was performed using Leica Microsystems SP2 SE confocal microscope with 10×/25 air, 63×/1.4 oil objectives. Leica LCS software (version 2.61) was used for image analysis.

NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) female mice were purchased from Jackson Laboratory and housed at MDACC following International Animal Care and Use Committee–approved protocol. Mice received sublethal irradiation 24 hours prior to intravenous injection with 2 × 10⁶ U266 cells (25). U266 engraftment and disease burden was measured by quantifying blood human IgE level by ELISA (Bethyl Laboratories). Mice were treated intravenously with 0.5 × 10⁶ PR1-CTL, irrelevant peptide (CG1, E75)-CTL or left untreated (PBS-treated) 14 days after U266 engraftment. Mice were treated intravenously with 8F4 antibody (10 mg/kg) or IgG2A isotype control (10 mg/kg; Jackson ImmunoResearch) three times per week beginning on day 28 for a total of 10 injections (26). Mice were sacrificed 35 days after CTL infusion or 3–4 days after the last antibody treatment, and bone marrow was harvested, stained with mouse CD45, human (h) CD45, hCD138, and HLA-A2 fluorescently conjugated antibodies, and then analyzed by flow cytometry.

Statistical analyses were performed using GraphPad Prism 6.0 software. P values less than 0.05 were considered significant.

To determine whether multiple myeloma cells express NE and P3, a panel of multiple myeloma cell lines was analyzed for endogenous expression of these PGP. Our data indicate that NE and P3 are absent in multiple myeloma cells at the protein and transcript levels (Fig. 1A and B). These findings are in agreement with reports from the Cancer Cell Line Encyclopedia (27), which confirm the lack of NE and P3 in multiple myeloma (Supplementary Fig. S1). The absence of NE and P3 in multiple myeloma is expected as multiple myeloma is of lymphoid origin, which is known to be deficient in myeloid PGP (6). The myelomonocytic U937 leukemia cell line that endogenously expresses NE and P3 was used as a positive control in these experiments.

Because we and others previously showed that solid tumors and B cells take up NE and P3 (14–19), we next tested whether multiple myeloma cells take up NE and P3, the first step in antigen cross-presentation. We cocultured a number of multiple myeloma cell lines with soluble NE or P3 over increasing durations. Flow cytometry analysis of intracellular NE and P3 staining demonstrates that NE and P3 are taken up by multiple myeloma (Fig. 1C and D). Time-dependent internalization of NE and P3 is observed, with a marked difference in the kinetics between NE and P3. The uptake of both proteins is sustained throughout the duration of the cultures. The kinetics and degree of NE and P3 uptake by the multiple myeloma cell lines closely resembles the pattern of NE and P3 uptake observed in breast cancer and melanoma (14).

As PR1 is an HLA-A2–restricted epitope, we focused our cross-presentation studies on the HLA-A2⁺ (i.e., HLA-A*0201) U266 multiple myeloma cell line. In agreement with the panel of multiple myeloma cell lines we examined (Fig. 1C and D), we confirmed the internalization of soluble and PMN-associated NE and P3 by U266 cells (Fig. 2A and B; Supplementary Fig. S2). Previously, we demonstrated that solid tumor cells can cross-present PR1 from both soluble and cell-associated sources of NE and P3 (14, 19). These findings were recapitulated in the U266 multiple myeloma cells that were cocultured with soluble NE, P3, or irradiated HLA-A2⁻ PMNs, the latter serving as the cellular source of NE and P3 that lack endogenous PR1 (Fig. 2C and D). As shown previously, cross-presentation was demonstrated by staining cells with anti-PR1/HLA-A2 Alexa-647–conjugated antibody (14). Because of the unique and inherent HLA-binding properties of TCR-like antibodies, we have observed low-affinity binding of 8F4 to HLA-A2 in other cell lines (13, 14), and also observed low background 8F4 staining in nonpulsed U266 cells. However, to highlight the specificity of 8F4 for PR1/HLA-A2, the HLA-A2⁻ multiple myeloma cell lines OPM-2 and RPMI-8226 were also cocultured with HLA-A2⁻ PMN and examined for cross-presentation (Supplementary Fig. S3).

Thus, these data confirm NE and P3 internalization by U266 multiple myeloma cells. Notably, there were kinetic differences in internalized NE or P3 based on the protein source: PMN-derived NE achieved higher intracellular levels than soluble NE and was comparable with the uptake of PMN-derived P3.

Cross-presentation involves distinct, well-defined pathways that utilize the proteasome and ER/Golgi (28–30). The proteasome plays an important role in antigen cross-presentation by cleaving intracellular proteins into small, 8–11 amino acid peptides, which are transported into the ER by the TAP1/2 complex. In the ER, peptides are loaded onto MHC-I molecules using components of the antigen processing machinery and are then exported to the cell surface via the Golgi. We hypothesized that NE and P3 cross-presentation involves proteasome and ER/Golgi shuttling, as previously shown for NE and P3 cross-presentation by solid tumors and APCs (14, 15). Our data confirm that multiple myeloma utilizes the ER/Golgi and proteasome for NE and P3 cross-presentation, as incubation of cells with lactacystin, a proteasome inhibitor, and brefeldin A, which inhibits ER to Golgi antegrade transport, both decreased PR1/HLA-A2 expression (Fig. 3A and B).

Next, we translated our in vitro studies into a clinically relevant system by examining PMN as the source for NE and P3 and using bortezomib, a proteasome inhibitor approved for treatment of multiple myeloma. In agreement with our findings with lactacystin, bortezomib reduced PR1/HLA-A2 on the U266 myeloma cell surface (Fig. 3C). In addition, the IMiD lenalidomide did not alter surface levels of PR1/HLA-A2 on U266 after coculture with PMNs (Fig. 3D). We also observed a consistent decrease in the surface expression of overall HLA-A2 due to inhibition of the proteasome (Supplementary Fig. S4). As conventional antigen cross-presentation mechanisms employ proteasome (20), it is not surprising that overall expression of HLA-A2 decreases with the use of lactacystin and bortezomib, and highlights the role of the proteasome in PR1 cross-presentation.

Immunotherapy targeting PR1 has shown promising efficacy in the treatment of myeloid malignancies (11, 13, 31, 32). Thus, we investigated whether PR1 cross-presentation by multiple myeloma cells could lead to their lysis by PR1-CTL and the complement-fixing, anti-PR1/HLA-A2 8F4 antibody. Calcein-AM cytotoxicity assays demonstrate that cross-presentation of NE and P3 by U266 cells renders them susceptible to killing by PR1-CTL and 8F4 antibody in a dose-dependent manner (Fig. 4). Specifically, NE enhanced the killing of U266 cells by PR1-CTL at 10:1 and 5:1 effector:target (E:T) ratios, when compared with untreated or ova-supplemented U266 cells (Fig. 4A). The efficacy of PR1-CTL in eliminating PR1-cross-presenting multiple myeloma cells was further validated by U266 cells that were supplemented with P3, where killing was demonstrated at the 10:1, 5:1, and 2.5:1 E:T ratios (Fig. 4A). As we have previously shown that the anti-PR1/HLA-A2 antibody (8F4) lyses malignant cells via CDC (13, 14), we tested whether multiple myeloma could be killed by 8F4. In a standard CDC assay, we demonstrate significantly higher 8F4-mediated killing of U266 target cells that were cultured with NE and P3 (Fig. 4B). Combined, these studies show that multiple myeloma is rendered susceptible to killing by PR1-targeting immunotherapies and further confirm PR1 cross-presentation by multiple myeloma.

As PR1 cross-presentation by U266 increases susceptibility to PR1-targeting immunotherapy in vitro, we next investigated whether PR1 can be targeted in vivo using multiple myeloma U266 xenograft mouse models (25). To test this hypothesis, we utilized NSG mice, which contain bone marrow–resident PMNs, providing an available source of NE and P3 for cross-presentation. In addition, the murine PR1 sequence is homologous to the human sequence (VLQELNVTV), and murine CTL can recognize PR1/HLA-A2 and are able to be expanded after vaccination (33).

After confirming U266 engraftment in the bone marrow and following treatment with 8F4, we demonstrate a significantly decreased U266 multiple myeloma burden, as shown by a decrease in the concentration of human IgE in mouse serum in comparison with isotype and untreated groups (Fig. 5A). Furthermore, 8F4 treatment also noticeably reduced the percent of multiple myeloma cells in mouse bone marrow in comparison with mice treated with isotype and untreated mice (Fig. 5B). Similar results were also seen using PR1-CTL (Fig. 5C). U266 multiple myeloma cells were identified as human CD45⁺ and mouse CD45⁻ cells (Fig. 5D and E). These data suggest that 8F4 and PR1-CTL are a feasible and effective therapy for HLA-A2⁺ multiple myeloma.

We next investigated whether PR1 could be detected in the bone marrow from patients with multiple myeloma, and if immunity to PR1 (i.e., PR1-CTL) could be detected in peripheral blood (PB) from patients with multiple myeloma following allo-SCT. We were able to detect PR1/HLA-A2 on the surface of 4 of 8 HLA-A2⁺ patients with multiple myeloma (Supplementary Table S1; Fig. 6; control staining is shown in Supplementary Fig. S5).

To determine whether PR1-CTL could be detected in PB from HLA-A2⁺ patients with multiple myeloma, we used PR1/HLA-A2 dextramer staining and showed PR1-CTL in the PB from 10 of 14 multiple myeloma patients who received allo-SCT (Supplementary Fig. S6A and S6B; Supplementary Table S2). The median frequency of PR1-CTL in these multiple myeloma patients was 0.053% of CD8⁺ T cells (range, 0%–1.67%), which is higher than the frequency of PR1-CTL reported in HLA-A2⁺ healthy donors (24). Furthermore, using CD45RA and CCR7 staining, PR1-CTL phenotype was analyzed in 4 of the patients and demonstrated primarily an effector memory phenotype of the PR1-CTL (Supplementary Fig. S7). In addition, we investigated whether PR1-CTL are present in the PB of HLA-A2⁺ multiple myeloma patients after receiving autologous stem cell transplant (auto-SCT; Supplementary Fig. S8; Supplementary Table S3). PR1-CTL frequencies were detected by PR1/HLA-A2 tetramer staining in all 18 multiple myeloma patients following auto-SCT. The median frequency of PR1-CTL in these patients was 0.205% (range, 0.076%–1.31%). These patient data suggest that bone marrow–derived NE and P3 is taken up and cross-presented by multiple myeloma, and that immunity to PR1 is elicited in multiple myeloma patients following SCT.

Finally, we used flow cytometry to study the expression of PR1 on CD138⁺ multiple myeloma cells from patient bone marrow [to corroborate confocal microscopy data (Supplementary Table S1; Fig. 6)] and to also investigate the expression of costimulatory molecules on PR1/CD138⁺ multiple myeloma cells. Our data show surface expression of PR1 and costimulatory molecules HLA-DR, CD40, CD80, and CD86 on CD138⁺ multiple myeloma cells in 12 (Supplementary Table S4; Supplementary S9) patients, further supporting the observation that patient multiple myeloma cells are capable of PR1 cross-presentation and possibly cross-priming.

This work demonstrates that multiple myeloma cells lacking endogenous NE and P3 have the ability to take up NE and P3 from the extracellular microenvironment and cross-present PR1 in the context of HLA-A2. Furthermore, PR1 cross-presentation renders multiple myeloma susceptible to killing by PR1-CTL and anti-PR1/HLA-A2 antibody. Our data illustrate a novel mechanism by which multiple myeloma is able to present antigens to the immune system. These findings suggest that cross-presentation may broaden the multiple myeloma tumor antigen repertoire, thus expanding the immunotherapeutic targets that could be exploited to treat multiple myeloma. Also, this study highlights the therapeutic potential of PR1-based therapies in multiple myeloma.

Cross-presentation is an important mechanism in the initiation of the CD8⁺ T-cell immune response. A number of studies have shown that B cells can cross-present antigens and prime an immune response, while other reports have suggested that B cells lack the ability to cross-present (14, 34–36). This study is the first to report antigen cross-presentation by multiple myeloma cells, a finding supported by a previous study from outside our laboratory, demonstrating that multiple myeloma cells can act as APCs (37). In that study, CD38⁺ plasma cells from the bone marrow of patients with multiple myeloma were shown to stimulate a T-cell response to tetanus toxoid and purified protein derivative. Furthermore, that study demonstrated that multiple myeloma plasma cells had low expression of the costimulatory molecules CD40, CD80, CD86, and HLA-DR, which are known to be critical for priming the immune system. Our data show that a small subpopulation of multiple myeloma cells can express these costimulatory molecules (Supplementary Table S4). Although in our report we focused on the capacity of multiple myeloma cells to cross-present antigens in the context of becoming a target for a CD8⁺ T-cell immune response, together with published data showing that multiple myeloma cells express costimulatory molecules under specific conditions, it is possible that multiple myeloma could function as APCs and prime a CD8⁺ T-cell immune response.

The immune system has clear implications in multiple myeloma. For example, multiple myeloma cases have been reported in the setting of immunodeficiency states (38, 39). Furthermore, deficiencies in the immune system have been reported in multiple myeloma patients, while higher frequencies of distinct populations of immune cells have been correlated with favorable outcomes in multiple myeloma patients (40, 41). Recent studies have begun to characterize the surface antigen repertoire of multiple myeloma to discover effective antigen targets for immunotherapeutic development (42). Although some tumors downregulate surface HLA class I molecules to evade the immune system (43), this does not appear to be the case in multiple myeloma, where there is an abundance of surface peptide/HLA class I molecules. A number of peptides have been shown to elicit effective CD8⁺ T-cell immune responses in multiple myeloma and are promising immunotherapeutic targets (42, 44, 45). However, in many peptide discovery approaches, antigens that are known to be absent from the tissues that are being analyzed are oftentimes excluded from further development as immunotherapeutic targets. Our studies highlight the potential importance of cross-presented antigens and critically evaluate the antigen repertoire from a clinically relevant perspective, as they could provide an entire class of antigens that could be targeted with immunotherapeutic approaches.

The proteasome plays an integral role in the process of cross-presentation (46). The involvement of the proteasome in NE and P3 cross-presentation by multiple myeloma is especially relevant in the context of bortezomib, which is part of the first-line therapy for multiple myeloma patients. However, the effects of bortezomib on the peptide repertoire and antigen presentation has yet to be fully determined. There is a preponderance of data that demonstrate negative effects of proteasomal inhibition on HLA-class I antigen presentation (46). In the setting of multiple myeloma, one study reported a decrease in the presentation of endogenous antigens by multiple myeloma after cells were exposed to bortezomib (47). Other studies, however, have demonstrated an increase in the presentation of some antigens following treatment with bortezomib (48–50). The variable effects that are seen following cell treatment with bortezomib may be specific for bortezomib and not generalizable to all proteasome inhibitors. Although our data presented here and our previously published data support an inhibitory effect on cross-presentation following proteasomal inhibition (14, 15), further studies using antigens other than NE and P3 need to be conducted to more conclusively determine the effects of proteasome inhibition on antigen cross-presentation.

The timing of immunotherapies has been shown to be critical to clinical outcomes. PR1-targeting immunotherapy may be best integrated into standard-of-care myeloma regimens. Our in vitro data suggest that lenalidomide does not interfere with the ability of myeloma to cross-present PR1, implicating a possible benefit using these two therapies in combination. However, our data with bortezomib suggest that the activity of PR1-targeting immunotherapy may be attenuated in the setting of proteasome inhibition. Furthermore, 8F4 may be useful as a purging strategy in the setting of stem cell collection in preparation for auto-SCT. As patients with multiple myeloma often receive granulocyte colony stimulating factor (G-CSF) as part of their mobilization, which increases the expression of NE/P3 in the bone marrow (7, 51) and thereby increases the source for PR1 cross-presentation, 8F4 may be applicable in this setting to reduce the multiple myeloma burden in multiple myeloma cells that cross-present PR1. Our data using preclinical xenograft animal models, where 8F4 treatment alone markedly reduced multiple myeloma in the bone marrow, supports this hypothesis.

We recognize that normal hematopoietic cells express the source proteins from which PR1 is derived, raising a concern regarding the potential toxicity of PR1-targeting immunotherapies. Furthermore, we note that fine epitope mapping demonstrated that 8F4 has contact residues with the HLA-A2 molecule (13). As HLA-A2 is part of the conformational epitope of PR1/HLA-A2, naturally we expect some binding of 8F4 to HLA-A2, as shown previously (14, 15, 19). However, to date, preclinical studies (13, 26) and clinical trials (11, 32) have demonstrated PR1-targeting immunotherapies to be safe and efficacious despite the shared expression of NE and P3 between malignant cells and normal counterpart, and the binding of 8F4 to HLA-A2.

Although we have shown PR1 expression by primary multiple myeloma, we recognize that we have not directly shown killing of primary multiple myeloma cells by PR1-CTL or 8F4. However, taken together, our data support the continued investigation of PR1-targeting immunotherapies, including 8F4 antibody, PR1-peptide vaccine, and PR1-CTL in the treatment of multiple myeloma patients. Our results also emphasize the need to evaluate the role of cross-presentation as a mechanism for the generation of novel tumor antigens in multiple myeloma.

E.A. Mittendorf is a consultant/advisory board member for Merck and AstraZeneca. J.J. Molldrem holds ownership interest (including patents) in Astellas Pharma, for patent royalty income related to the 8F4 monoclonal antibody. No potential conflicts of interest were disclosed by the other authors.

Conception and design: G. Alatrash, A.A. Perakis, A.M. Cernosek, J.J. Molldrem

Development of methodology: G. Alatrash, A.A. Perakis, H.L. Peters, R. Patenia, A.V. Philips, A.M. Cernosek, N. Qiao, J. Weng, S. Lu, K. Clise-Dwyer, J.J. Molldrem

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): G. Alatrash, A.A. Perakis, C. Kerros, P. Sukhumalchandra, H. Jakher, M. Zope, A. Sergeeva, S. Yi, K.H. Young, Q. Ma, J.J. Molldrem

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): G. Alatrash, A.A. Perakis, C. Kerros, H.L. Peters, P. Sukhumalchandra, H. Jakher, M. Zope, K.H. Young, A.V. Philips, H.R. Garber, E.A. Mittendorf, J.J. Molldrem

Writing, review, and/or revision of the manuscript: G. Alatrash, A.A. Perakis, C. Kerros, H.L. Peters, M. Zope, K.H. Young, H.R. Garber, S. Lu, E.A. Mittendorf, J.J. Molldrem

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G. Alatrash, A.A. Perakis, M. Zhang, A.M. Cernosek, L.S. St John, S. Lu, Q. Ma

Study supervision: G. Alatrash, J.J. Molldrem

This research was funded by NIH, National Cancer Institute (NCI) grant P50CA142509 (to G. Alatrash), NIH T32 Immunobiology Training grant 5T32CA009598-24 (to H.L. Peters and C. Kerros), and NIH/NCI grant P30CA16672 (Flow Cytometry, Cell Sorting and Cell Imaging Core Facilities), and was supported by the generous philanthropic contributions to The University of Texas MD Anderson Moon Shots Program.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.