Single- and double-hit events in genes encoding immune targets before and after T cell–engaging antibody therapy in MM

Novel T cell–engaging therapies such as chimeric antigen receptor (CAR) T cells or bispecific antibodies (BsAb’s) have entered the treatment for multiple myeloma (MM) and show exceptionally high response rates in relapsed/refractory (RR) patients. Beyond BCMA, a plethora of other plasma cell targets is currently under investigation, including SLAMF7,¹  GPRC5D,²  CD44v,³  and FCRL5⁴  (also known as FcRH5), with promising preliminary results in clinical phase 1 trials. Yet patients continue to relapse, and in contrast to immunotherapy in aggressive lymphoma, there is no survival plateau after treatment with CAR T cells or BsAb’s in MM.

Recently, we described a first tumor-intrinsic resistance mechanism to TNFRSF17 (known as BCMA)-directed CAR T cell therapy. It helped explain the mystery of homozygous deletions on chromosome 16p encompassing the TNFRSF17 (BCMA) gene as the cause of antigen escape in a patient treated with the BCMA CAR T cell product ide-cel in the KarMMa trial.⁵  At the American Society of Hematology (ASH) 2020 meeting, other groups confirmed biallelic events (del/del and del/mut) as mechanisms of antigen escape after CAR T cell therapy.^6,7  Notably , biallelic inactivation usually requires 2 independent hits, and heterozygous TNFRSF17 deletions have been described in BCMA-naïve patients^8,9  at risk for a second hit. Thus, the pressure of an anti-BCMA–directed immunotherapy could lead to irreversible antigen loss and subsequent treatment failure.

Here, we report the first case of a biallelic BCMA gene deletion after BsAb therapy, highlighting that this genomic mechanism of resistance is relevant in both types of T cell treatment: CAR T cell therapy and BsAb therapy. Furthermore, by using whole-genome sequencing (WGS) data, we provide evidence that genomic alterations in genes encoding immunotherapy targets preexist in a relevant proportion of patients before immunotherapy, including heterozygous deletions, which may potentially serve as a predictive biomarker in future trials.

We report the case of a 56-year-old patient treated with a CD3xBCMA BsAb who was subject to WGS in compliance with the CARE guidelines. Furthermore, we report on WGS data derived from 100 patients with MM (50 newly diagnosed [NDMM]; 50 relapsed/refractory [RRMM]; supplemental Table 2). All patients gave their written informed consent for scientific evaluations. The study was approved by the Internal Review Board at the University of Würzburg (reference KFO216) and adhered to the tenets of the Declaration of Helsinki.

WGS and bulk RNA sequencing (RNA-seq) were performed as previously described.¹⁰  Briefly, WGS libraries were prepared from 1 µg DNA from CD138 purified cells with the TruSeq polymerase chain reaction free library prep kit, and 2 × 150-bp paired-end sequences were generated on a NovaSeq 6000 instrument with a median coverage of 106×. For bulk RNA-seq, 250 ng of total RNA per sample was used to produce stranded RNA libraries (TruSeq Total Stranded RNA, Illumina). Altogether, 2 × 100-bp paired-end reads were sequenced on the NovaSeq 6000 with a median of 64 million reads per sample (bioinformatic approach is provided in the supplemental Data).

BCMA protein expression was determined by immunohistochemistry using a polyclonal goat anti-BCMA antibody (dilution 1:10; target retrieval pH 6.1) (AF193, R&D Systems) on paraffin-embedded bone marrow sections according to standard procedures.

We detected a clonal biallelic deletion at 16p13.13 encompassing the BCMA gene in a 56-year-old patient with penta-refractory MM who was receiving daratumumab, pomalidomide, and dexamethasone in his eighth line of therapy. WGS showed a hypodiploid karyotype with monosomy 16 and a 586-kb homozygous deletion involving BCMA (Figure 1; supplemental Figure 1; supplemental Table 1). The patient was treated previously on the AMG420 trial, which evaluated a CD3xBCMA bispecific T cell–engager construct at the sixth line of therapy; he achieved a complete response but relapsed after 6 months of treatment. We retrospectively assessed the BCMA status using immunohistochemistry on bio-banked trephine biopsies that showed strong and consistent BCMA expression before the patient participated in the AMG420 trial but loss of BCMA expression at relapse. Notably, BCMA loss was persistent at a subsequent relapse from carfilzomib, lenalidomide, dexamethasone (KRD) salvage therapy 20 months later. Without knowing about BCMA loss at relapse at that time, the anti-BCMA antibody-drug conjugate belantamab mafodotin was administered at the ninth line of therapy, to which the patient was primary refractory (supplemental Figure 2). This is the first case study that shows homozygous BCMA gene deletion to be a mechanism of resistance to BCMA-directed BsAb’s. In this patient, antigen loss was irreversible, and subsequent BCMA therapy failed, suggesting that targeted immunotherapies, in theory, are an evolutionary bottleneck in the progression of MM. It is still unclear whether the BCMA deletion was a rare preexisting event or whether it developed during anti-BCMA therapy. Highly sensitive single-cell approaches and accounting for spatial heterogeneity will be required to address this question.

To explore the relevance of this resistance mechanism in MM, we screened for genomic alterations in BCMA and 20 alternative immunotherapy targets (supplemental Table 2) by using WGS in a group of 100 T cell immunotherapy-naïve patients (supplemental Tables 3 and ⁴ ). Indeed, we found heterozygous deletions in GPRC5D (15%), CD38 (10%), SDC1 (5%), TNFRSF17 (4%), and NCAM1 (3%). Overall, heterozygous deletions occurred in 30% of patients, and some patients showed deletions in up to 3 different immunotarget genes. Considering the emerging dual-antigen targeting immunotherapies,^11,12  it is worth mentioning that 1 patient presented with both heterozygous deletion of BCMA and heterozygous deletion of GPRC5D. Of note, 15 of 21 targets were encoded on chromosomes involved in hyperdiploid karyotypes or on 1q and were consequently amplified in about half the patients. Single-nucleotide variants (SNVs) in genes encoding immune targets were found at low frequency (Figure 2). With the exception of 3 frameshift mutations in GPRC5D, no mutation was nonsense or frameshift, and the functional consequences of gene mutations remain elusive. Nevertheless, we observed 1 biallelic event (del/mut) for GPRC5D.

The genomic makeup of MM changes from baseline to relapse,⁹  and 1q gain or amplification is more common at relapse (supplemental Figure 3). Consequently, in our cohort, gain of FCRL5 (19 vs 30 patients; P = .045) and SLAMF7 (19 vs 31 patients; P = .027, Fisher’s exact test) were significantly enriched in RRMM. Furthermore, as expected, we observed a clear trend for more deletions in pretreated patients (Figure 2) because karyotypes become more complex at later stages of the disease.¹³  This trend was confirmed when comparing our data to the data set from the CoMMpass trial, which included NDMM patients (supplemental Table 5).

In addition to using WGS, we performed transcriptome analysis for 72 of the 100 patients for whom genetic material was available. Strikingly, gains, heterozygous deletions, and SNVs did not result in significant changes in the gene expression of the immunotherapy targets (Figure 2), suggesting that expression of these genes is not solely regulated by gene copy number; with respect to loss of expression, both alleles have to be dysfunctional. In addition, we found tumor suppressor genes with implications in hematologic malignancies within common deleted regions such as CDKN1B¹⁴  and RERG¹⁵  on chromosome 12 deleted in 100% and 93% of patients with del(12p), respectively, or LITAF¹⁶  deleted in all patients with del(16p); thus, one might speculate whether loss of them may provide a growth advantage to the affected clones.

The current treatment landscape of MM is rapidly changing, and we expect therapy to move toward the use of T cell–based immunotherapies upfront. Our study supports the use of immunotherapies earlier in the disease process, because the frequency of deletions and mutations in genes encoding immunotherapy targets was lower in patients with NDMM vs those with RRMM. Our early data suggest that biallelic BCMA gene deletion is a key immune evasion mechanism for BCMA-targeting T cell therapies. It also argues for BCMA expression testing before BCMA therapy, especially in patients pretreated with anti-BCMA. Similarly, this mechanism of resistance to T cell–based therapies will affect alternative antigens, and we detected genomic vulnerability in their encoding genes in a significant number of treatment-naïve patients. Combination of different immunotargets (eg, multispecific CAR T cells) might be a promising approach for overcoming drug resistance caused by loss of a single antigen.^11,17 

The authors thank Sabine Roth and Rabea Konietschke for their excellent technical support.

This work was supported by German Cancer Aid via the Mildred Scheel Early Career Center (MSNZ Würzburg) Programme (L.R. and K.M.K.), by the patient advocacy group ‘Forschung hilft’-Stiftung zur Förderung der Krebsforschung an der Universität Würzburg and the Nationales Centrum für Tumorerkrankungen (H.E., K.M.K., and L.R.), and by Sonderforschungsbereich (SFB)/Transregio TR221 (projects A03 and A04) (H.E.).

Contribution: M.S. Truger, L.R., and K.M.K. conceived of and designed the study; J.D., X.Z., L.H., A. Ruckdeschel, S.H., L.H., J.P., M.S. Topp, A. Riedel, L.R., K.M.K., and H.E. provided study material or patients; W.W., M.M., and M.S. Truger analyzed bioinformatic and statistical data; L.R., M.S. Truger, K.M.K., J.D., X.Z., A. Ruckdeschel, M.J., A. Riedel, H.E., C.H., and N.W. interpreted data; M.C.D.V. and N.B. analyzed additional data sets; L.R., M.S. Truger, and K.M.K. wrote the paper; and all authors reviewed and approved the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Leo Rasche, Department of Internal Medicine 2, University Hospital Würzburg, Oberdürrbacherstr. 6, 97080 Würzburg, Germany; e-mail: rasche_l@ukw.de.