B‐cell maturation antigen‐specific chimeric antigen receptor T cells for multiple myeloma: Clinical experience and future perspectives

Despite major advances in the treatment of multiple myeloma (MM), it remains a largely incurable disease with long‐term control often dependent on continuous therapy. More effective, better tolerated treatments are therefore required to achieve durable remissions and to improve the quality of life of MM patients. Adoptive immunotherapy employing T cells expressing chimeric antigen receptors (CAR) is currently among the most promising treatment approaches in cancer. Within the target portfolio for MM immunotherapy, B‐cell maturation antigen (BCMA) is among the most widely studied target antigens. BCMA is consistently expressed on MM cells and, importantly, is not expressed in critical healthy tissue. For this reason, it is an ideal target for MM immunotherapy. Several clinical trials evaluating different BCMA‐targeting CAR constructs have been initiated and early results are very promising. However, in this rapidly developing clinical landscape, the ultimate role of BCMA‐specific CAR‐T cell therapy remains unclear. In this review, we will summarize currently available clinical data on BCMA‐directed CAR‐T cells and discuss potential future perspective for this promising treatment approach in MM.


| INTRODUCTION
The outcomes of patients with multiple myeloma (MM) have significantly improved over recent decades following the widespread introduction of novel agents such as proteasome inhibitors, immunomodulatory drugs (IMiDs) and monoclonal antibodies into routine clinical care. Despite this progress, the duration of remissions achieved in certain patient populations, in particular, those classified as genetically "high-risk" (HR), can still be relatively short and almost all patients eventually experience relapse and finally succumb to the disease. Chimeric antigen receptor (CAR) T cells (CAR-Ts) are new tools in the growing armamentarium against MM. In other hematological malignancies, namely, acute lymphoblastic leukemia (ALL) and diffuse large B-cell lymphoma (DLBCL), CD19-specific CAR-Ts achieved promising response rates in heavily pretreated patients. [1][2][3] This led to FDA and EMA approval of Kymriah and Yescarta as the first two commercially available CAR-T products. CAR-T therapy target antigens that have been clinically evaluated in MM include CD19, 4,5 CD138, 6 the kappa light chain 7 and SLAMF7/CS1 (NCT03710421). Of the CAR-T targets assessed to date, B-cell maturation antigen (BCMA) is one of the most intensively studied and is probably the most promising with the potential to significantly influence the therapeutic landscape in MM.

| BCMA: EXPRESSION AND FUNCTION
Expression of the surface protein BCMA, also known as CD269 or tumor necrosis factor receptor superfamily member 17 (TNFRSF17), first occurs in late memory B cells and is ubiquitously expressed on plasma cells (PCs). 8,9 The B-cell activation and proliferation-inducing ligands, BAFF and APRIL, bind to BCMA and consequently promote NFκB and mitogen-activated protein kinases (MAPK) pathway activation. 10 BCMA seems to be important for long-lived PC maturation and survival but may be less critical to the overall humoral B-cell response. 11,12 BCMA is regularly expressed on MM cells 13 at a broad range of epitope densities and can promote MM growth and immunosuppression in the bone marrow (BM) microenvironment. 14 Membranebound BCMA can be cleaved by γ-secretase, a process which may lead to reduced overall BCMA cell surface expression as well as to the formation of a soluble form (sBCMA) that interferes with BCMAbinding therapeutic molecules. 15 Importantly, no significant BCMA expression has been observed on nonhematological tissues. 16

| BCMA-TARGETED CAR-T CONSTRUCTS IN CLINICAL EVALUATION
Several BCMA-specific CAR-T constructs are currently under clinical evaluation for the treatment of MM. The design of the most clinically advanced CAR constructs is summarized in Table 1. The CAR constructs developed at the National Cancer Institute (NCI) 17

and by
Bluebird Bio (BB2121) 18 both utilize a murine single-chain variable fragment (scFv). For the earlier CD19-directed CAR-Ts, anti-murine CAR host immune responses were reported that may have limited efficacy by prevention of in vivo expansion. 19 As a result, there has been increasing use of human-derived scFvs in newer BCMA- Selection of BCMA-targeting CAR constructs in clinical development NCT03070327). 28 Treatment was conducted in four escalating dose cohorts (72 × 10 6 , 137 × 10 6 , 475 × 10 6 and 818 × 10 6 mean transfused CAR+ cells). The ORR was 64% with a median duration of response of 106 days. Importantly, patients treated at the higher dose levels (DL3 and DL4) all showed an objective response. Three of these five patients sustained response for more than 6 months.
The Fred Hutch presented clinical trial data on seven RRMM patients who had received BCMA-directed CAR-Ts (NCT03338972). 23 Patients had a median of eight treatment lines prior to CAR-T administration (range 6-11). They were divided into two cohorts based on the BM PC burden (A 10%-30%, B >30% BM PCs). Patients received between 5-15 × 10 7 CAR+ T cells in a fixed ratio of CD4+ and CD8+ cells. All evaluable patients responded to the treatment and all achieved MRD-negativity.
In parallel with these primarily academia-driven investigator-  An ORR of 100% with an improvement of remission after CAR-T administration from 40% CR to 70% CR/sCR was reported. In addition, MRD negativity increased from 44% to 60%. 35 Due to the limited patient numbers and short follow-up, formal evaluation of the efficacy of a CAR-T-based consolidation approach, especially when compared to standard maintenance strategies, is required.
In summary, these preliminary clinical results of BCMA-directed CAR-Ts are promising. However, cross-comparison of the studies to identify the most potent CAR-T product is difficult due to major differences in CAR design, cell production protocols and patient characteristics. In addition, despite the high rates of MRD-negativity, the duration of responses is relatively short. Therefore, treatment of larger patient populations and longer follow-up will be necessary to better understand the relative strengths and weaknesses of the different CAR-T constructs. Among all evaluated BCMA-CAR-Ts, bb2121 is the clinically most advanced cell product. It is expected that bb2121 will be the first CAR-T therapy to be approved for the treatment of MM and approval is expected in 2020. This first indication will probably be for RRMM patients with at least three previous treatment lines including IMiDs, PIs and CD38 antibodies. Furthermore, triplet as well as quadruplet regimens in earlier lines of treatment will challenged and even study concepts for first-line high-risk MM patients (eg, with R-ISS III) are initiated.

| Side effects of BCMA-targeting CAR-Ts
Typical side effects reported for CD19-targeting CAR-Ts, the most intensively evaluated CAR-T therapy, are cytokine release syndrome (CRS) and neurotoxicity ("CAR-T related encephalopathy syndrome"; CRES). 36 A recent consensus paper by the American Society for Transplantation and Cellular Therapy (ASTCT) provides recommendations for the grading of CRS and CRES. 37 The interleukin (IL)-6 inhibitor tocilizumab and corticosteroids are now standard of care for the treatment of higher grade CRS and CRES. 36 The frequency and severity of CRS and CRES observed in different BCMA-CAR-T trials are summarized in Table 4. CRS with BCMA-CAR-Ts is common but is mainly grade 1-2. CRS associated with BCMA-specific CAR-Ts so far appears to be manageable and the rates of CRES seem to be relatively modest compared to those seen with CD19-specific CAR-Ts.
In addition to these specific CAR-T related toxicities, prolonged cytopenia is frequently observed. This is thought to be principally due to the use of myelosuppressive conditioning therapy (Cy ± Flu) in this heavily pretreated patient population.
A major potential concern with gene therapy is insertional mutagenesis. In contrast to studies conducted with hematopoietic stem cells, T cells seem to be less susceptible to secondary malig-

| Predictors of response
When making the case for cost and labor-intensive novel treatment approaches such as CAR-Ts, reliable predictive markers would be very useful for identifying the patients who are likely to benefit the most.
However, due to the small numbers of patients treated with BCMAspecific CAR-Ts and the limited follow-up so far, it is difficult to define reliable predictors of response.
In several phase I studies, a dose-dependent increase in response rates was seen at higher DLs (Table 2). 26,27,29 In addition, it appears that peak peripheral blood CAR-T expansion levels may correlate with response. [26][27][28]32 However, the underlying mechanisms promoting these peak expansion levels is not yet clearly understood. Conditioning may not be mandatory for response to BCMA-CAR-Ts in MM patients but may be associated with more durable expansion and is therefore generally recommended. 27 24,29 In addition, although a high tumor burden in the BM can be associated with an increased risk of severe CRS, this does not seem to influence response. 26,27 Reports on more patients and likely meta-analyses may be helpful to identify and validate predictive biomarkers of CAR-T treatment outcome. In a meta-analysis, clinical results from 285 patients treated in 15 clinical trials were pooled: higher dose levels and absence of high-risk cytogenetics were associated with higher response rates. 46 Further analysis is necessary to proof these hypotheses.
Albeit prior treatment and BCMA expression may not be direct biomarkers of response, refractoriness and very late CAR-T therapy application rather than earlier has been discussed as less advantageous for prolonged efficacy. In our experience with 40 CD19+ lymphoma patients, multiple treatment with cytostatic drugs, particularly bendamustine, exerted a negative effect on quantity and quality of CD3+ T cells for further production into CARTs. Currently, the BELINDA as well as the ZUMA-7 trials are ongoing to clarify for CD19+ lymphoma patients whether the use of CAR-T cell therapy in second line is better than in third line. The same applies to myeloma patients and makes early use of this treatment modality to achieve a very deep remission upfront very temping.
To achieve optimal safety for CAR T cell treatment of myeloma patients, the expertise of both, auto-and allo-teams, should be combined. 47

| Mechanisms of resistance to BCMA-targeting CAR-Ts
The limited number of patients treated so far precludes meaningful analysis of the mechanisms underlying primary and secondary resistance to BCMA-specific CAR-Ts. One mechanism could be the down- Strategies to modulate the immunosuppressive nature of the MM microenvironment may therefore be warranted. Furthermore, the quality and fitness of T cells used for CAR-T production seem to be an important factor in adoptive immunotherapy. Moreover, a higher proportion of less differentiated T cells in the leukapheresis product appears to correlate with the clinical response. 27 However, it is currently unclear if an optimization of production protocols to enrich for less-differentiated CAR-Ts affect product efficacy.
In summary, further investigation of resistance mechanisms is required to develop strategies to overcome treatment failure.

| Optimization of production protocols
For CAR-T therapy, both the CAR constructs and the CAR-T production process affect treatment efficacy. However, cell production protocols are often treated as commercially sensitive information and are not made publically available. This limits our collective ability to meaningfully compare the different CAR-T studies.
Less-differentiated T cells in the final cell product, especially those with a naïve-like (T N ) or stem cell memory-like (T SCM ) phenotype, are thought to affect long-term engraftment and sustained antitumor activity. 49 Cytokines such as IL-2, IL-7 and IL-15 can influence the composition of different T-cell subsets during CAR-T generation. [50][51][52] In addition, pharmacological pathway inhibition during ex vivo T-cell expansion 53  The application of CAR-Ts in a defined CD4:CD8 ratio has been proposed to be superior to the application of bulk generated CAR-Ts. 19,63 The first results with BCMA-directed CAR-Ts in a defined CD4:CD8 ratio were reported by the FHCC and are encouraging with six of six patients achieving MRD-negative responses. 23 However, this information is too preliminary to allow for any more general recommendations regarding the value of defined CD4:CD8 ratios.
It is highly desirable to launch more CAR-T cell studies combined with in-house production of CAR-Ts developed by academic institutions. This will make all data available to the scientific community straight from the leukapheresis and the CAR T-cell product till the clinical response data, the frequency and immunophenotype of the circulating CAR-Ts. We think that only an algorithm comprising all these data layer will allow us to define the type of patients who will eventually profit from CAR-T cell therapy. Moreover, early phase I investigator-initiated trials allow to test innovative targets and CAR constructs which will be overtaken by the pharmaceutical industry for further evaluation in phase II or III trials. National and international third-party funding for early phase trials is highly desirable, particularly in Europe where gene immuno-cell therapy is so much behind the US and China. Besides the addition of drugs stimulating the immune system, targeting more than one antigen may reduce the risk of target antigen escape. These additional CARs can either be included in the same construct as BCMA to obtain bi-specific CAR-Ts or can be on two differ- In summary, combination therapy is probably an important strategy for successful BCMA-directed CAR-T therapy and several new combination strategies are or soon will be clinically evaluated. The advantage of adding small molecules (eg, IMiDs and proteasome inhibitors) or monoclonal antibodies (eg, against CD38 and SLAMF7) to CAR-Ts is that translation of this approach to the clinic will be relatively simple. On the other hand, antigen escape can be prevented by dual targeting. The combination of several approaches including targeting of the tumor microenvironment, inhibition of immune evasion and achievement of sustained antitumor activity by CAR-Ts may therefore be necessary for long-term remissions.

| Perspective
BCMA-directed CAR-Ts have so far mainly been evaluated as salvage treatment in the highly refractory "last line" setting ( Figure 2A). However, impaired T-cell fitness in this heavily pretreated patient population may be in part responsible for limited long-term disease control.
Therefore, BCMA-targeted CAR-Ts may be more effective in earlier lines of treatment, for example, in first relapse after high-dose chemotherapy ( Figure 2B). Furthermore, their administration as consolidation treatment after high-dose chemotherapy, even in the first-line setting, especially for patients with insufficient response, such as persistent minimal residual disease, or the presence of HR cytogenetics, may be a promising strategy to achieve deep and durable responses ( Figure 2C). MRD testing could be used as a trigger for CAR-T-based consolidation. Long-term remissions of more than 24 months and even cure of MM by MRD eradication should be the goal of this complex, labor-intensive and expensive technology.
In DLBCL, salvage autologous transplantation is already being directly compared to CD19-CAR-Ts in the ZUMA-7 clinical trial (NCT03391466). In MM, BCMA-directed CAR-Ts will definitely be evaluated in the first-line setting and may even challenge front-line high-dose chemotherapy as a possibly less toxic treatment approach that can better preserve the quality of life of MM patients ( Figure 2D).
CAR-Ts may even be an effective consolidation treatment for elderly and frail patients who would not be eligible for autologous transplantation in the first-line setting.
Maintenance therapy after high-dose chemotherapy is currently accepted as the standard-of-care in MM. However, a "oneshot" treatment approach with CAR-T cells may achieve sufficiently deep remissions that maintenance therapy could be omitted ( Figure 2E). This would remove the requirement of continuous treatment and thereby improve the quality of life as well as treatment costs.
However, whether CAR-T cell treatment will be able to overcome the adverse prognosis of biologically high-risk MM, whether epitope exposure in an MRD-positive disease status will be sufficient for T-cell expansion, and whether only costly, relative toxic combination regimens can deliver meaningful long-term remissions in a complex disease like MM are challenging questions that need to be assessed and evaluated in many future clinical trials.

| CONCLUSIONS
In conclusion, the field of adoptive immunotherapy is developing rapidly. For the treatment of MM, there are currently several targets under investigation. BCMA is among the most studied as well as being one of the most promising targets for CAR-T therapy. It is expected that one of the BCMA-directed CAR-T constructs will be the first genetically modified cellular product approved for the treatment of MM. This may significantly change the treatment landscape of MM, leading to more durable disease control and represents an important addition to the growing armamentarium of therapies in our quest to ultimately cure patients with MM.