Reduced Pathogen‐Specific T Cell Response With BCMA‐Targeted T Cell‐Engaging Therapy to Multiple Myeloma Ida Steiro, Synne Stokke Tryggestad, Eli Svorkdal Hess, Hanne Hella, Pegah Abdollahi, Kari Lenita Falck Moore, Margrete Friestad, Magne Børset, Tobias S Slørdahl, Anne Marit Sponaas Abstract ABSTRACTIntroductionBispecific antibody (BsAb) treatment can lead to increased infections in myeloma patients.MethodsWe used ELISpot and flow cytometry to test the T cell response to BCMA‐CD3‐bispecific T cell engager (BiTE) pavurutamab (PAV) and BCMA‐expressing myeloma cells after in vitro culture.ResultsReduced pathogen‐specific responses were found in T cells cultured with PAV yet they killed myeloma cells. In an exploratory study, reduced pathogen‐specific memory T cell responses were detected in PBMCs from four myeloma patients responding to BCMA‐targeted BsAb therapy.ConclusionOur preliminary observation and its link to infection should be followed up in a larger study.Trial RegistrationThe authors have confirmed clinical trial registration is not needed for this submission. Introduction Bispecific antibodies (BsAb) are effective in treatment of relapsed and refractory multiple myeloma, but can lead to adverse effects including increased rates and severity of infections, particularly where BCMA is targeted []. This has been attributed to BCMA expression on B cells, but could also be due to loss of T cell fitness []. Little is known about how BsAb treatment affects circulating memory responses. We tested antigen‐specific responses to common pathogens in T cells cocultured with BCMA‐expressing human myeloma cells lines (HMCLs) and pavurutamab (PAV), an anti‐CD3‐BCMA bispecific T cell engager (BiTE) []. To avoid confounding effects of differential T cell fitness in patient samples, we used T cells from healthy donors. We investigated how well the T cells were able to respond to common pathogen antigens in an ELISpot assay and their ability to kill myeloma cells in vitro. Materials and Methods IFN‐γ T cell ELISpot assay: T cells isolated from the PBMC‐HMCL cocultures were seeded in R4 medium with or without APCs (antigen presenting cells) in a T cell:APC ratio of 1:0.5 with 1 μg/mL CEFX peptide pool and 20 U/mL IL‐2. After 2 days IFN‐γ‐producing T cells were quantified by ELISpot assay, using manufacturer's protocol; 5 × 104 cells/well (T cells and APCs or T cells only) were seeded in 96‐well ELISpot plates, precoated with anti‐IFN‐γ antibody in R2 medium. Cells were stimulated with CEFX peptide pool 0.75 μg/mL, 0.3 μg/mL anti‐CD3 mAb or media. Plates were incubated at 37°C with 5% CO2 for 21–24 h followed by ALP‐linked immunosorbent detection. The AID vSpot Spectrum was used for counting spots. Results are expressed as number of spot‐forming colonies (SFCs)/5 × 104 seeded cells. PBMCs from a healthy donor were added with each experiment as positive assay control (Figure ). Details of methods are provided in . Results PAV Stimulation Reduced Pathogen‐Specific T Cell Responses in Vitro But the T Cells Killed Myeloma Cells Healthy individuals have circulating memory T cells to pathogens they have been exposed to or vaccinated against []. Pathogen‐specific, IFN‐γ‐producing CD8+ T cells can be detected in PBMCs with ELISpot assays after stimulation with nonameric peptides from many pathogens (Figure ) []. To determine whether BiTE stimulation influences antigen‐specific T cell responses, PBMCs were incubated with PAV or BiTE isotype control antibody (control) and HMCLs as described in Methods. Three BCMA‐expressing HMCLs were used; JJN‐3, RPMI‐8226, and U226 (Figure ). BsAb is given with weekly and forthnightly intervals. To model the response after a week, T cells were isolated from cultures after 7 days, seeded with syngeneic APCs and CEFX peptides before they were tested in an ELISpot assay (Figure ). Although more T cells were harvested from cultures stimulated with PAV than control (Figure ), PAV significantly reduced the number of pathogen‐specific T cells (Figure ). The PAV‐stimulated T cells responded to anti‐CD3, albeit with fewer SFCs than those cultured with control (Figure ). A much lower level of response to both CEFX and anti‐CD3 was observed on Day 10 in the absence of APCs at Day 7 (Figure ), suggesting that restimulation with peptide on APCs was necessary to obtain adequate numbers of IFN‐γ‐producing CD8+ T cells. It was difficult to model the clinical response to repeated injections as in vitro stimulations with PAV beyond Day 7 also led to T cell death. On Day 7 after a second restimulation, there were only 7%–8% viable T cells (Figure ). This made it very difficult to recover enough T cells to test functional responses. Exhausted or activation‐induced dysfunctional T cells have attenuated effector functions, including impaired killing []. We investigated the ability to kill myeloma cells before and after culture and restimulation with PAV (Figure , Figure ). All cell lines were lysed after 24 h incubation with PAV and T cells (Figure ), and cells isolated from 7‐day cocultures with PAV killed HMCLs more efficiently than control when JJN‐3 and U266 cells were targeted (Figure , Table ). RPMI‐8226 were lysed by restimulated T cells regardless of exposure to PAV or control. PAV Upregulated Activation Markers on T Cells Reduction in the number of CEFX specific IFNγ spots (Figure ) could be caused by inadequate activation or activation‐induced dysfunction of T cells. Reduced T cell function with altered expression of surface markers were demonstrated for BsAbs []. We determined activation and exhaustion markers at baseline and after 3‐ and 7‐day culture with PAV or control and HMCLs. PBMCs cultured without HMCLs did not upregulate any of the markers (Figure ). Compared with control, PAV increased expression (% positive cells and MFI) of ICOS, CD38, TIGIT, and PD‐1 (Figure , Figure ). Higher proportions of CD8 T cells expressed these markers than CD4+ cells, possibly reflecting a dominant role of CD8+ cells upon BiTE stimulation []. When presenting the data as fold change MFI, for some of the cell lines and markers, there was a trend in which expression peaked at Day 3 of stimulation, followed by a decline at Day 7 (Figure , Figure ). TIGIT and PD‐1 peaked at Day 3 and either declined or was unchanged after 1‐week. Therefore, PAV activated the T cells and the suppression of the IFN‐γ response on Day 7 was not due to lack of stimulation. Soluble BCMA (sBCMA) reduces the response to BCMA‐targeted BsAb‐therapy []. sBCMA was found in culture supernatants from all cell lines (Figure ) and in the cocultures (Figure ). However, very little sBCMA was detected in the cultures with PAV. Importantly, PAV expanded T cell numbers (Figure ), and upregulated activation molecules on T cells, which was not seen in cultures with control. Therefore, sBCMA was unlikely to inhibit PAV and generation of CD8+ IFN‐γ producing T cells in these in vitro studies. Myeloma Patients Receiving BCMA‐BsAb Have Reduced Numbers of Antigen‐Specific T Cells in Circulation Antigen‐specific memory T cell responses can be detected in human PBMCs with an ELIspot assay after in vitro culture with IL‐2, IL‐7, and antigen stimulation (Figure ) [ ]. CEFX‐specific memory T cell responses were measured in PBMCs collected from four MM patients treated with BCMA‐directed BsAbs at the start, after 30 and 90 days (± 7 days). The number of pathogen‐specific IFN‐γ‐secreting CD8+ T cells dropped during treatment (Figure ). A drop at Day 30 (Figure ) could be explained by T cell margination during the first cycle [] as suggested from the decrease in total lymphocyte count (Figure ), but the T cells still responded to anti‐CD3. Nevertheless, at Day 90 the lymphocyte number increased and the pathogen‐specific T cells remained low. Incidentally, all patients suffered from infections (Table ). Although only exploratory, these data may suggest that pathogen‐specific memory T cell responses decline in patients treated with at least two cycles of BCMA‐directed therapy. Discussion T cell exhaustion and dysfunction after repetitive BsAb treatment is known, but linkage to infectious complications is not clear []. Although 1‐week PAV stimulation reduced CEFX‐specific IFN‐γ CD8+ T cell responses in vitro, the T cells were still able to kill myeloma cell lines. One reason for this could be that whilst IFN‐γ production requires recognition of peptide on MHC on the APCs together with coreceptors such as CD86 and ICAM, BsAbs activate T cells and force T and myeloma cells together leading to transfer of perforin and granzyme B at the synapse. Thus, we can envisage that partially suppressed T cells would be able to kill myeloma cells, and conversely, T cells killing tumor may not be able to remove pathogens. Other explanations such as apoptosis of pathogen‐specific T cells and expansion of tumor‐specific T cells are possible, but unlikely as the frequency of tumor‐specific T cells would be very low in T cells from healthy individuals []. The T cells expanded, and upregulated activation and inhibitory molecules in vitro, suggesting that PAV was active and not inhibited by sBCMA. In a small cohort of patients, all responding to BCMA‐targeted BsAbs, the number of pathogen‐specific T cells dropped. T cells were still in blood as they responded to anti‐CD3 stimulation. At Day 90, however, both pathogen‐specific and polyclonal T cell activation were significantly reduced compared to baseline. Susceptibility to infection in patients treated with BCMA‐targeting BsAbs has been attributed to deletion of B cells [], but our exploratory observations, although limited, could imply that T cell immunity to common pathogens is also affected. Whether this contributes to the increased frequency and severity of infections requires further investigations of possible mechanisms and more patients. Conclusion T cells cultured with a BCMA‐BiTEs and HMCLs killed myeloma cells, but showed reduced numbers of pathogen‐specific IFNγ+ CD8+ T cells. Pathogen‐specific memory T cells declined in the blood of four patients responding to BCMA‐targeted BsAb therapy. This initial observation should be followed up in a larger cohort of patients to investigate if and how reduction of T cell memory could be related to increased incidence of infections. Ethics Statement The use of buffy coat from blood donors was approved by the Regional Ethics Committee (REK 2009/2245). Collection of patient samples was approved by the Regional Ethics Committee (REK 2023/562858). All donors gave informed consent.