ALBERT

Description: Introduction: The optimal clinical setting and cell product characteristics for chimeric antigen receptor (CAR) T cell therapy in multiple myeloma (MM) are uncertain. In CLL patients treated with anti-CD19 CAR T cells (CART19), prevalence of an early memory (early-mem) T cell phenotype (CD27+ CD45RO- CD8+) at time of leukapheresis was predictive of clinical response independently of other patient- or disease-specific factors and was associated with enhanced capacity for in vitro T cell expansion and CD19-responsive activation (Fraietta et al. Nat Med 2018). T cell fitness is therefore a major determinant of response to CAR T cell therapy. In an accompanying abstract (Cohen et al.), we report that higher percentage of early-mem T cells and CD4/CD8 ratio within the leukapheresis product are associated with favorable clinical response to anti-BCMA CAR T cells (CART-BCMA) in relapsed/refractory MM. Here, we compare leukapheresis samples from MM patients obtained at completion of induction therapy (post-ind) with those obtained in relapsed/refractory (rel/ref) patients for frequency of early-mem T cells, CD4/CD8 ratio, and in vitro T cell expansion. Methods: Cryopreserved leukapheresis samples were analyzed for the percentage of early-mem T cells and CD4/CD8 ratio by flow cytometry and in vitro expansion kinetics during anti-CD3/anti-CD28 bead stimulation. Post-ind samples were obtained between 2007 and 2014 from previously reported MM trials in which ex-vivo-expanded autologous T cells were infused post-ASCT to facilitate immune reconstitution (NCT01245673, NCT01426828, NCT00046852); rel/ref samples were from MM patients treated in a phase-one study of CART-BCMA (NCT02546167). Results: The post-ind cohort includes 38 patients with median age 55y (range 41-68) and prior exposure to lenalidomide (22), bortezomib (21), dexamethasone (38), cyclophosphamide (8), vincristine (2), thalidomide (8), and doxorubicin (4); median time from first systemic therapy to leukapheresis was 152 days (range 53-1886) with a median of 1 prior line of therapy (range 1-4). The rel/ref cohort included 25 patients with median age 58y (range 44-75), median 7 prior lines of therapy (range 3-13), and previously exposed to lenalidomide (25), bortezomib (25), pomalidomide (23), carfilzomib/oprozomib (24), daratumumab (19), cyclophosphamide (25), autologous SCT (23), allogeneic SCT (1), and anti-PD1 (7). Median marrow plasma cell content at leukapheresis was lower in the post-ind cohort (12.5%, range 0-80, n=37) compared to the rel/ref cohort (65%, range 0-95%). Percentage of early-mem T cells was higher in the post-ind vs rel/ref cohort (median 43.9% vs 29.0%, p=0.001, left figure). Likewise, CD4/CD8 ratio was higher in the post-ind vs rel/ref cohort (median 2.6 vs 0.87, p2 lines of therapy prior to apheresis (n=3) compared to the rest of the cohort (n=35). Conclusion: In MM patients, frequency of the early-mem T cell phenotype, a functionally validated biomarker of fitness for CAR T cell manufacturing, was significantly higher in leukapheresis products obtained after induction therapy compared to the relapsed/refractory setting, as was CD4/CD8 ratio and magnitude of in vitro T cell expansion. This result suggests that CAR T cells for MM would yield better clinical responses at early points in the disease course, at periods of relatively low disease burden and before exposure to multiple lines of therapy. Figure. Figure. Disclosures Garfall: Novartis: Research Funding; Kite Pharma: Consultancy; Amgen: Research Funding; Bioinvent: Research Funding. Cohen:GlaxoSmithKline: Consultancy, Research Funding; Kite Pharma: Consultancy; Oncopeptides: Consultancy; Celgene: Consultancy; Novartis: Research Funding; Poseida Therapeutics, Inc.: Research Funding; Bristol Meyers Squibb: Consultancy, Research Funding; Janssen: Consultancy; Seattle Genetics: Consultancy. Fraietta:Novartis: Patents & Royalties: WO/2015/157252, WO/2016/164580, WO/2017/049166. Davis:Novartis Institutes for Biomedical Research, Inc.: Patents & Royalties. Levine:CRC Oncology: Consultancy; Brammer Bio: Consultancy; Cure Genetics: Consultancy; Incysus: Consultancy; Novartis: Consultancy, Patents & Royalties, Research Funding; Tmunity Therapeutics: Equity Ownership, Research Funding. Siegel:Novartis: Research Funding. Stadtmauer:Janssen: Consultancy; Amgen: Consultancy; Takeda: Consultancy; Celgene: Consultancy; AbbVie, Inc: Research Funding. Vogl:Karyopharm Therapeutics: Consultancy. Milone:Novartis: Patents & Royalties. June:Tmunity Therapeutics: Equity Ownership, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties, Research Funding; Tmunity Therapeutics: Equity Ownership, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties, Research Funding; Immune Design: Membership on an entity's Board of Directors or advisory committees; Novartis Pharmaceutical Corporation: Patents & Royalties, Research Funding; Celldex: Consultancy, Membership on an entity's Board of Directors or advisory committees; Immune Design: Membership on an entity's Board of Directors or advisory committees; Novartis Pharmaceutical Corporation: Patents & Royalties, Research Funding. Melenhorst:Novartis: Patents & Royalties, Research Funding; Incyte: Research Funding; Tmunity: Research Funding; Shanghai UNICAR Therapy, Inc: Consultancy; CASI Pharmaceuticals: Consultancy.

Description: Neurologic toxicity has been observed with anti-CD19 chimeric antigen receptor (CAR) T cells and the anti-CD19 BiTE blinatumomab. Both focal (e.g., cranial nerve palsy) and global (e.g., generalized seizures) abnormalities have been reported, often associated with systemic cytokine release syndrome (CRS) but also observed after recovery from or in absence of CRS. CART-BCMA consists of expanded autologous T cells transduced with a 4-1BB:CD3-zeta-based CAR specific for B Cell Maturation Antigen. Here, we report clinical features and management of a severe neurotoxicity observed on a phase 1 trial of CART-BCMA for multiple myeloma (MM) (NCT02546167). The subject is a 55-year-old female with high-risk IgA lambda MM. At time of CART-BCMA infusion, her MM manifestations included cytopenias and plasmacytomas of the pleura and paravertebral muscles. Bone marrow (BM) was 〉95% BCMA+ plasma cells. Pre-treatment brain MRI showed pachymeningeal thickening and enhancement over the left cerebral convexity, possibly due to extension of calvarial MM lesions. There was no evidence of CNS MM on a neurologist's exam or by CSF cytology. The subject received 2x108 CART-BCMA cells, 40% of the planned dose, over two consecutive days, without lymphodepleting chemotherapy; a third planned infusion was held due to fevers. Over the next 4 days, fevers persisted, hypoxia and delirium developed, and cytopenias worsened. Brain MRI and lumbar puncture on day 4 showed no new abnormalities. Evaluation for infection was negative. These symptoms coincided with rise in serum IL-6 (nl range

Description: Select tumor tag sequences, informed by the probability of generating the same sequence in an independent recombination event, are likely to be useful in tracking mature B cell malignancies where the IGH locus has been deleted. We considered the use of IGK and IGL for this purpose, but found that 2-4% of these rearrangements were shared between different individuals. The relatively high frequencies of such public rearrangements at IGK and IGL prompted us to search for other forms of sequence diversity that could be used as private clonotypic tags. Somatic hypermutations (SHM) may serve such a function. Our analysis of IGK and IGL suggests that tumor tracking sequences for detecting minimal residual disease should be selected with care, and these loci may be best suited for lymphoid malignancies that are characterized by high levels of SHM. Tracking minimal residual disease for B cell malignancies is an established technology, traditionally using either flow cytometry or a custom quantitative PCR assay for each patient. Recent technical developments in the massively parallel sequencing of somatically rearranged IG loci allow for a standard assay to be applied to screen for residual tumor burden in all patients, by first identifying the clonal IG rearrangements tagging the tumor in an index sample taken from the patient during active disease, and then screening for these tumor tagging sequences in follow up samples. A crucial assumption in these tagging strategies is that the tumor tagging sequences are idiosyncratic to the tumor, and unlikely to be generated independently in a recurrent rearrangement. In order to screen for recurrent sequences between two healthy individuals, we generated IG heavy and light chain libraries from 100,000 antigen experienced B cells (CD19+CD27+) isolated from whole blood by FACS. 130 bp reads were collected, starting within the J segment and extending across the CDR3 into the V segment. Unique sequences were compared between individuals to assess the frequency of nucleotide identical, “public” rearrangements shared between individuals. Less than 0.01% of unique IGH sequences overlapped between individuals, so the risk of a false positive MRD result from recurrent recombination at IGH is minimal. However, 4.3% and 1.9% of unique sequences at IGK and IGL, respectively, were shared between individuals. The shared sequences had significantly higher average copy numbers than unshared sequences, accounting for 20% of total sequences at IGK and 12% of total sequences at IGL. These data suggest that B cells carrying public sequences undergo higher levels of clonal expansion, and/or they are recurrently produced. Public sequences carried by B cell malignancies are likely to be of limited utility as tumor-tagging sequences, as it may be impossible to distinguish between low-level residual disease and benign, recurrent rearrangements in the patient. Therefore we assessed if we could predict whether a given sequence in the memory repertoire would be public using solely information derived from that sequence. We used logistic regression to screen for variables to predict the likelihood of a given sequence to be public, and identified a number of expected variables as significant predictors, including the identity of the V and J segments, the length of the non-templated insertion at the junction, and the number of somatic hypermutations within the V or J segment. By far the most important of these factors was the number of SHM events in the clone; consequently, the most useful light chains for tumor tracking will be those with significant SHM. We continue to explore factors contributing to the public IG repertoire. Particularly at IGK, there is an unexpectedly narrow range of CDR3 lengths, and we are determining if this might be attributable to low diversity in the primary repertoire, or due to positive selection in favor of this length in the mature naïve or mature repertoires. In conclusion, a high frequency of public IG light chain sequences in the antigen- experienced peripheral B cell repertoire suggests that naïve application of light chain clones for tracking MRD can generate false positive results, but that careful selection of tumor tracking sequences with SHMs can minimize this risk. Disclosures: Carlson: Adaptive Biotechnologies: Consultancy, Equity Ownership, Patents & Royalties. Howie:Adaptive Biotechnologies: Employment, Equity Ownership.

Description: Background: Though CD19 is expressed only rarely on multiple myeloma (MM) plasma cells (PC), rare CD19+ B cells can be identified in MM patients that are clonally related to the MM PC. These clonotypic B cells may exhibit properties of cancer stem cells (enhanced MM-propagating properties and drug resistance compared to MM PC) and thus be a potential therapeutic target in conjunction with therapies that target MM PC. CTL019 consists of autologous T cells transduced via lentiviral vector with an anti-CD19 scFv coupled to CD3-zeta and 4-1BB signaling domains and expanded ex vivo with anti-CD3/CD28-conjugated beads. To target both clonotypic B cells and MM PC, we conducted a pilot clinical trial of CTL019 administered after high-dose melphalan and autologous stem cell transplantation (ASCT) in relapsed/refractory MM patients who had previously undergone first-line ASCT with short progression-free survival (PFS). Methods: Subjects were required to be medically fit for ASCT and have progressed within 1 year of a prior ASCT performed as part of first-line therapy. Study therapy consisted of ASCT with melphalan 140-200 mg/m2 followed by 1-5x107 CTL019 cells 12-14 days later. The primary endpoint was safety and feasibility of CTL019 manufacturing and administration in this clinical setting. Secondary endpoints included assessments of CTL019 in vivo persistence and activity against normal B cells, plasma cell immunophenotype as a response biomarker, and PFS after ASCT + CTL019 in comparison to PFS after initial ASCT. Results: Twelve subjects enrolled, and 10 received study therapy; autologous T cells failed to expand ex vivo in one enrolled subject, and one enrolled subject elected to pursue off-study therapy. Median age was 61 (range 48-68). Median prior lines of therapy was 6 (range 2-10). Poor-prognosis features were present in 8/10 subjects (6/10 with poor-prognosis cytogenetics, 2/10 with BRAF V600E mutations, 1/10 with secondary plasma cell leukemia). Median PFS after first-line ASCT was 258 days (range 100-342). In pre-ASCT bone marrow (BM), the dominant MM PC population was CD19-negative by flow cytometry in 9/9 evaluable subjects, though 7/9 exhibited rare CD19+ subsets comprising 0.05-1.5% of MM PC. Melphalan dose was 140 (N=7) or 200 (N=3) mg/m2. All subjects infused received the maximum target dose of 5x107 CTL019 cells. Adverse events (AE) consisted mostly of expected ASCT toxicities. Grade ³3 AE that were at least possibly related to CTL019 included grade 3 autologous GVHD (N=1, resolved with corticosteroids) and oral mucositis (N=1). Grade 1 cytokine release syndrome occurred in 1 subject. There was no ASCT-related mortality. After infusion, CTL019 cells were detectable in peripheral blood (PB) of all subjects and persisted for median of 44 days (range 14-156). Presence of PB CTL019 cells was associated with absence of PB B cells. Notably, CTL019 cells were detected in BM in 9/10 subjects at day 42 and/or 100 post-ASCT. Median PFS after ASCT + CTL019 was 185 days (range 42-479); all subjects have progressed. The peak BM CTL019 frequency correlated significantly with favorable PFS (SpearmanÕs rho=0.77, P=0.009). There was no association between PFS and peak frequency of CTL019 or duration of CTL019 persistence in PB. In 3/10 subjects, PFS after ASCT + CTL019 met or exceeded PFS after first-line ASCT (Figure). For comparison, in a historical cohort of 18 patients who received first-line and salvage ASCT at our institution since 2008, no patients exhibited longer PFS after salvage ASCT. Conclusion: CTL019 manufacturing and administration post-ASCT is safe and feasible in patients with advanced MM. Correlation of PFS with CTL019 frequency in BM and prolonged PFS in 3 subjects is suggestive of clinical efficacy. A phase-two study of CTL019 using a 10-fold higher dose after first-line ASCT in high-risk MM patients is ongoing. Figure. Figure. Disclosures Garfall: Medimmune: Consultancy; Bioinvent: Research Funding; Novartis: Consultancy, Research Funding. Stadtmauer:Novartis: Consultancy; Takada: Consultancy; Janssen: Consultancy; Amgen: Consultancy; Teva: Consultancy; Celgene: Consultancy. Maus:Novartis: Patents & Royalties: related to CTL019, Research Funding. Hwang:Novartis: Research Funding. Vogl:Takeda: Consultancy, Research Funding; GSK: Research Funding; Calithera: Research Funding; Teva: Consultancy; Constellation: Research Funding; Celgene: Consultancy; Acetylon: Research Funding; Karyopharm: Consultancy. Cohen:Bristol-Meyers Squibb: Consultancy, Research Funding; Janssen: Consultancy. Weiss:Novartis: Consultancy. Porter:Genentech: Employment; Novartis: Patents & Royalties, Research Funding. Frey:Novartis: Research Funding; Amgen: Consultancy. Milone:Novartis: Patents & Royalties, Research Funding. Mangan:Novartis: Speakers Bureau. Lacey:Novartis: Research Funding. Melenhorst:Novartis: Patents & Royalties: Novartis, Research Funding. Ambrose:Novartis: Research Funding. Chen:Novartis: Research Funding. Kulikovskaya:Novartis: Research Funding. Levine:Novartis: Patents & Royalties, Research Funding; GE Healthcare Bio-Sciences: Consultancy. June:Johnson & Johnson: Honoraria; Tmunity Therapeutics: Equity Ownership; Novartis: Honoraria, Patents & Royalties, Research Funding; Immune Design: Consultancy, Equity Ownership; Celldex: Consultancy, Equity Ownership; Novartis: Honoraria, Patents & Royalties, Research Funding; Pfizer: Honoraria.

Description: Matching at HLA-C has been shown to influence outcomes and has been incorporated in selection of unrelated adult donors. In contrast, selection of UCB grafts has historically considered HLA-A and B at antigen and DRB1 at allele level resolution. Recent data in single, myeloablative pediatric UCB transplantation demonstrated that antigen level matching at HLA-C was associated with lower NRM and improved survival in patients receiving better HLA-matched units. Whether or not these same principles apply in the setting of double UCB (dUCB) transplantation has not been addressed. Thus, we retrospectively studied whether HLA-C matching would influence outcomes in 490 patients with hematological malignancies at two centers undergoing myeloablative and reduced intensity dUCB transplantation. We considered the worst HLA-matching of the 2 donor units with 316 (64%) patients receiving at least one 4/6 matched unit and 144 (29%) one or two 5/6 units, and 30 (6%) receiving two 6/6 HLA-matched units. Patients were scored considering the number of HLA-C antigen matches as 0-1 (23%), 2 (40%), 3 (18%), and 4 (19%), of 4 possible matches. The median age was 47 yrs. (range, 2-72), 59% were male, 285 (58%) had acute leukemia, 291 (59%) were CMV seropositive, 319 (66%) received RIC regimen, and 400 (82%) had CsA/MMF immunosuppression. In the overall study population, we observed no significant influence of HLA-C matching on the risk of death, treatment failure, non-relapse death, relapse, GVHD and hematopoietic recovery. However, we recognized an interaction between conventional HLA-matching at A, B, and DRB1 and number of matches at HLA-C in the survival endpoints. Thus, we analyzed two groups based on conventional HLA-matching at A, B and DRB1: those receiving at least one 4/6 HLA-matched unit (4/6 & 4-6/6; n=316) or those receiving ≥ 5-6/6 matched unit (5/6 & 5-6/6; n=174). In the ≥5/6 UCB transplants, there was no significant influence of HLA-C matching on the risk of death, treatment failure, non-relapse death, and relapse. However, in 4/6 & 4-6/6 transplants, better matching at HLA-C was associated with lower risk of death, treatment failure, and non-relapse death (Table), but there remained no association with risk of relapse. These data contrast with those reported with single UCB grafts and suggest that with 4/6 HLA-matched UCB units, additional matching at HLA-C reduces treatment failure and improves survival, and should be included in the match strategy. In better matched (≥5/6) dUCB grafts further matching at HLA-C offers no additional benefit. Table 1.Table shows multivariate analysis results in 316 patient who received 4/6 & 4-6/6 dUCB graftsMatching at HLA CNOverall mortalityTreatment failureNon-relapse deathRelapseGrade II-IV acute GVHDRelative Risk (95% CI)PRelative Risk (95% CI)PRelative Risk (95% CI)PRelative Risk (95% CI)PRelative Risk (95% CI)P4/4331.01.01.01.01.03/4521.70.121.50.163.00.100.90.781.10.83(0.8-3.5)(0.8-2.8)(0.8-10.9)(0.4-1.9)(0.4-3.2)2/41362.30.011.70.065.4

Description: Background: Despite major advances in the treatment of multiple myeloma (MM) only a minority of patients achieve long-term disease control. Immunotherapy combined with autologous hematopoietic stem cell transplantation (auto-HCT) may reduce relapse rates. Immunoglobulin idiotype (Id) conjugated with a carrier protein, Keyhole limpet hemocyanin (KLH), is a tumor-specific antigen in MM. Vaccine-primed, anti-CD3/anti-CD28 costimulated adoptive T-cell transfer can augment humoral and cellular immune responses to vaccination despite cytotoxic therapy. We hypothesized that Id-KLH vaccine + the vaccine-primed costimulated T cells will result in a robust Id-specific humoral and cellular response, compared to a control vaccine (KLH only). Methods: In this randomized, phase II trial, the primary objective was to determine if Id-KLH primed, costimulated T cells will induce a more robust Id-specific immunity than KLH-primed T cells. Eligible patients had IgG monoclonal protein. Patients were randomized 1:1 to receive either Id-KLH vaccine or KLH-only vaccine, followed by auto-HCT, and then vaccine- primed costimulated T cells followed by two booster doses of the vaccine to which they were randomized. This study was supported by The MD Anderson Cancer Center SPORE in Multiple Myeloma. Results: A total of 36 patients were enrolled between 1/2013 and 5/2015. Sixteen (44%) were randomized to Id-KLH and 20 (55%) to KLH-only. The Table below summarizes the patient characteristics. There was no significant difference between the two groups in terms of age, risk status or induction therapy. No treatment-related mortality, infusion reactions or dose-limiting toxicity was seen in either arm. Five (31%) and 3 (15%) patients achieved complete remission (CR) by day+180 in the Id-KLH and KLH arms, respectively (p=0.42). Initial analysis of a subgroup of patients revealed a significantly higher mRNA expression of immune activation genes IL-2, CCR6 and CD40 by NanoString nCounter in the Id-KLH group compared with the KLH only group (Figure). Eleven (68%) and 17 (85%) went on to receive maintenance therapy with lenalidomide or lenalidomide + ixazomib in the Id-KLH and KLH arms, respectively (p=0.42). After a median follow up of 26 months, 2-year PFS was 81% and 83% in Id-KLH and KLH arms, respectively (p=0.35). Conclusion: Id-KLH vaccine and vaccine-primed costimulated T cells can be safely administered in the setting of auto-HCT.There was a more robust immune response and a trend towards higher CR rate in the Id-KLH group. Disclosures Stadtmauer: Teva: Consultancy; Amgen: Consultancy; Novartis: Consultancy; Takada: Consultancy; Janssen: Consultancy; Celgene: Consultancy. Garfall:Medimmune: Consultancy; Bioinvent: Research Funding; Novartis: Consultancy, Research Funding. Cohen:Janssen: Consultancy; Bristol-Meyers Squibb: Consultancy, Research Funding. June:Johnson & Johnson: Research Funding; Immune Design: Consultancy, Equity Ownership; Novartis: Honoraria, Patents & Royalties: Immunology, Research Funding; University of Pennsylvania: Patents & Royalties; Celldex: Consultancy, Equity Ownership; Tmunity: Equity Ownership, Other: Founder, stockholder ; Pfizer: Honoraria.

Description: Background Circulating plasma cells (PCs) have been identified as a prognostic factor in patients with myeloma precursor states (MGUS and SMM) and active multiple myeloma (MM). Enumeration of circulating PCs by available methods is not suitable for widespread use and does not provide molecular characterization. We developed and evaluated a novel method for enumeration and molecular characterization of circulating PCs (circulating multiple myeloma cells, “CMMC”), based on the CELLSEARCH® System (Janssen Diagnostics LLC, Raritan, NJ), an automated technology for the capture, enumeration and characterization of rare cells in the peripheral blood. Methods We are performing a prospective study of patients with MGUS and SMM to evaluate CMMCs as biomarker for progression to active MM. Utilizing the CELLSEARCH® System CMMCs were captured by CD138 ferrofluid magnetic particles and identification was defined as CD38+ and CD19-, CD45-. Nonviable cells were excluded by DAPI. Isolated CMMCs were stored and FISH for t(4:14), t(14;16) and del17 was performed. Results We have enrolled 16 patients, MGUS = 3, SMM = 11, and newly diagnosed MM = 2. The Mayo Risk stratification for MGUS patients was: low risk = 2, low-intermediate = 1. All SMM patients were low risk by Mayo Model incorporating serum free light chains. The median number of bone marrow plasma cells for MGUS patients was 7 (range 7-9) and for SMM patients was 15 (range 10-40). The median CMMCs for MGUS = 6 (range 2-55), median CMMCs for SMM = 31 (5-1918). The two patients with NDMM had 5870 and 5 CMMCs, respectively. A single patient with SMM progressed with a symptomatic solitary lumbar plasmacytoma and had CMMCs of 5 and 3 at baseline and progression, respectively. Abnormalities by FISH were detected in both bone marrow and CMMCs. Accrual is ongoing and additional data will be presented at the meeting. Conclusions The CELLSEARCH® CMMC assay can detect, quantify and provide molecular characterization of circulating PCs in MGUS/SMM/MM; longer prospective follow-up is needed to test the prognostic value of CMMCs. Disclosures Weiss: Janssen: Consultancy, Research Funding. Sasser:Janssen: Employment. Rao:Janssen: Employment, Equity Ownership. Foulk:Janssen: Employment. Gross:Johnson & Johnson: Employment, Equity Ownership. Cohen:Janssen: Membership on an entity's Board of Directors or advisory committees. Vogl:Celgene Corporation: Consultancy; Amgen: Consultancy; Millennium/Takeda: Research Funding; GSK: Research Funding; Acetylon: Research Funding. Stadtmauer:Janssen: Consultancy.

Description: Despite intense efforts, multiple myeloma remains incurable in most patients with the standard of care therapies. The plasma cell surface receptor B cell maturation antigen (BMCA) is highly expressed by myeloma cells and we recently demonstrated that 12 out of 25 heavily pretreated myeloma patients achieved a partial response or better after anti-BCMA CAR T cell treatment (VGPR, n=5; CR, n=1; sCR, n=1; Cohen et al., 2019, JCI 129(6):2210). To better understand the biological basis of this therapy, we identified key correlates of response using the pre-manufacturing apheresed T cells, the infusion product, and post-infusion T cells from the 25 patients in this cohort. As reported before, the disease characteristics, tumor burden, and CAR transduction efficiency did not correlate with therapy response. CAR T cell expansion, measured by the area under the curve of CAR qPCR in the first 21 days (AUC[0-21]), was highest in responding, lowest in non-responding patients (Jonckheere-Terpstra test, JT = 38, p=1.8x10^-6)(Fig.1A,B). Soluble BCMA, a biomarker of disease burden, shows a similar trend with response (Jonckheere-Terpstra test, JT = 54, p=1.2x10^-4). Furthermore, AUC[0-21] for CAR T cell expansion and soluble BCMA decline also strongly correlated (Spearman's rank correlation test, rho=0.82; p=2.41x10^-6), underscoring the quantitative relationship between CAR T cell expansion and tumor reduction. We have previously shown that response to CAR T cell therapy in CLL is largely determined by T cell memory function. To find if this extends to myeloma, we immunophenotyped apheresed T cells (or CAR-T precursor cells) and infusion product from the 25 patients. Phenotypically distinct T cell subpopulations were identified using shared-nearest-neighbor clustering method (PMID: 31178118) and their correlation with response to CAR T cell treatment was evaluated. This analysis revealed that among CD4+ and CD8+ CAR-T precursor cells, subpopulations representing naive and central memory T cells were enriched in T cells from responding patients, while non-responders displayed a distinctly activated effector phenotype at baseline. Additional analyses showed that apheresed CD8+ and CD4+ T cells from responder patients were non-cycling, granzyme B-negative, CTLA4[low] but otherwise largely immune checkpoint inhibitor-negative. CD8+ CAR-T precursor cells isolated from non-responders exhibited high expression levels of TIM3 or LAG3, and/or granzyme B, but not PD1, CTLA4, CD45RO or CD27. These data confirm the high activation, potential exhaustion and end-stage differentiation state of CAR-T precursor cells in this group. Similar analyses of infusion product CAR T cells did not reveal subpopulations associated with response. Clustering analysis of CD8+ CAR T cells within 20 days after infusion revealed a BCMA CAR-expressing cluster enriched in responding patients: a non-cycling, negatively regulated, Eomes-expressing central memory subset (cluster 0; Fig. 1E). Non-responding patients CAR-T cells displayed high levels of granzyme B and PD1 expression but were otherwise devoid of signs of activation (cluster 8; Fig. 1F). Furthermore, the abundance of CD8+ CAR-T cells with cluster 0 and 8 phenotype correlated significantly with in vivo expansion (AUC[0-21]; Fig. 1C). Four patients with a sufficiently high proportion of CAR expressing cells were phenotyped up to 125 days post-infusion. This analysis showed that the highly activated CAR T cell clusters 2 and 5 dominated at early phases post infusion but was rapidly replaced by non-cycling CAR T cells with downregulated CTLA4 and LAG3 but maintained expression of PD1 and TIM3 (cluster 0; Fig. 1D). Patient 27 with VGPR had a prominent effector population four months after infusion. BCMA-redirected CD4+ CAR T cells showed an enrichment of central memory phenotype CAR T cells in responding patients early after infusion, with high expression of Eomes, TIM3, and other immune checkpoint inhibitor molecules. This cluster also dominated the CD4 T cell repertoire in the first four months after infusion in the four responding patients. In conclusion, our data suggest that strategies to promote expression of Eomes and central memory function and reduce exhaustion in BCMA CAR T cells will enhance clinical activity. Further, these results underscore the "self-sustaining" feature of successful CAR T cell therapies in myeloma. Disclosures Pruteanu: Novartis: Employment. Cohen:Poseida Therapeutics, Inc.: Research Funding. Garfall:Tmunity: Honoraria, Research Funding; Amgen: Research Funding; Novartis: Patents & Royalties: inventor on patents related to tisagenlecleucel (CTL019) and CART-BCMA, Research Funding; Janssen: Research Funding; Surface Oncology: Consultancy. Lacey:Novartis: Patents & Royalties: Patents related to CAR T cell biomarkers; Tmunity: Research Funding; Novartis: Research Funding. Fraietta:Tmunity: Research Funding; Cabaletta: Research Funding; LEK Consulting: Consultancy. Brogdon:Novartis: Employment. Davis:Tmunity: Research Funding; Cabaletta: Research Funding. Levine:Tmunity Therapeutics: Equity Ownership; Avectas: Membership on an entity's Board of Directors or advisory committees; Vycellix: Membership on an entity's Board of Directors or advisory committees; Novartis: Consultancy; Novartis: Consultancy, Patents & Royalties, Research Funding; Cure Genetics: Consultancy; Incysus: Membership on an entity's Board of Directors or advisory committees; Brammer Bio: Membership on an entity's Board of Directors or advisory committees; CRC Oncology: Consultancy. Milone:Novartis: Research Funding; Novartis: Patents & Royalties: patents related to tisagenlecleucel (CTL019) and CART-BCMA. Stadtmauer:Janssen: Consultancy; Tmunity: Research Funding; Amgen: Consultancy; Abbvie: Research Funding; Novartis: Consultancy, Research Funding; Takeda: Consultancy; Celgene: Consultancy. June:Novartis: Research Funding; Tmunity: Other: scientific founder, for which he has founders stock but no income, Patents & Royalties. Melenhorst:National Institutes of Health: Research Funding; Parker Institute for Cancer Immunotherapy: Research Funding; Novartis: Research Funding, Speakers Bureau; Colorado Clinical and Translational Sciences Institute: Membership on an entity's Board of Directors or advisory committees; Stand Up to Cancer: Research Funding; Incyte: Research Funding; IASO Biotherapeutics, Co: Consultancy; Simcere of America, Inc: Consultancy; Shanghai Unicar Therapy, Co: Consultancy; Genentech: Speakers Bureau.

Description: Introduction: Venetoclax is an oral BCL2 inhibitor, effective in CLL and under investigation in multiple clinical trials to treat relapsed or refractory multiple myeloma (RRMM) in combination with a proteasome inhibitor (VPI) - either carfilzomib or bortezomib (Moreau P, 2017; Kumar S, 2017). VPI has shown promising initial results in a phase 2 trial with an overall response rate of 78% and a very good partial response or better (≥VGPR) rate of 56% (Costa L J, 2018; Terpos E, 2019). However, a separate phase 3 trial evaluating bortezomib given with venetoclax or placebo found improved response rates and progression free survival but a decrease in overall survival rate in the venetoclax arm of the study (data not published). The primary association with increased mortality was infection. Better characterizing infections in patients treated with VPI may give insight into the pathophysiology and suggest strategies for safe and effective use of this therapy in RRMM. Methods: We retrospectively reviewed charts of patients treated with VPI for infectious complications from initiation of treatment until one month after progression treated at the Hospital of the University of Pennsylvania. Infections were classified by site, pathogen and severity of infection. Additional data collected included demographics, blood cell counts, quantitative immunoglobulins, SPEP M-spike, light chain analysis, cytogenetics, prior lines of therapy, prophylactic antibiotics and IVIG use. We determined non-monoclonal IgG (NM-IgG) via the difference between serum M-spike and IgG for IgG patients. The two-sample Wilcoxon rank-sum was used to compare different subgroups within our study population. Results: We identified 18 patients treated with a VPI combination regimen of which 78% were male. The median age was 64.5 years (range 47-76) and a median of 3 (1-10) prior regimens. 14 were treated with carfilzomib and 4 with bortezomib combinations. 4 patients progressed by the time of data collection. 8 patients were positive for t(11;14). 11 patients experienced 35 discrete infectious episodes resulting in 4 hospitalizations. Respiratory infections predominated (29/35) with 24 upper respiratory infections or sinusitis. 4 were lower respiratory infections and comprised all the hospitalizations. Among respiratory infections, viruses were the only pathogens identified, including influenza, rhinovirus, coronavirus and RSV. Other sites of infection were gastrointestinal (including recurrent C. difficile) and urinary tract. No CNS, blood, or intra-abdominal infections were identified. Reductions in lymphocyte counts and non-monoclonal IgG were near universal but significant reductions in neutrophil counts were not observed. A greater proportional reduction of NM-IgG was noted in good-responders, defined as CR and VGPR (≥VGPR, n=9), with a mean of 78% decrease compared to 47% in poor-responders (PR/MR/SD/PD, n=4, p=0.045). Importantly, treatment of those with t(11;14) positive status demonstrated no significant difference in NM-IgG. Additionally, there was a relative risk ratio favoring a reduction in infections among those with t(11;14). Conclusion: Patients with RRMM treated with VPI experience frequent infectious complications, some severe, with sustained reductions in serum IgG and lymphocyte counts, but without neutropenia. Reduction in NM-IgG is associated with improved response depth. This phenomenon could explain the paradoxical improvement of response rate and progression free survival with concomitant reduction in overall survival due to infections noted by other reports. Interestingly, the presence of t(11;14) in this analysis was not associated with as great a drop in NM-IgG, as was seen in responders as whole, indicating a more differential effect. The absence of significant reduction in NM-IgG is a potential mechanism for relative reduction of infections in this subgroup. Suggestions of survival benefit among t(11;14) positive subgroups in other reports may be related to this preservation of NM-IgG. Replenishing NM-IgG with IVIG is a hypothetical mitigating strategy that could be of benefit in other subgroups treated with this combination of therapy. Disclosures Cohen: Poseida Therapeutics, Inc.: Research Funding. Garfall:Surface Oncology: Consultancy; Novartis: Patents & Royalties: inventor on patents related to tisagenlecleucel (CTL019) and CART-BCMA, Research Funding; Janssen: Research Funding; Amgen: Research Funding; Tmunity: Honoraria, Research Funding. Vogl:Karyopharm Therapeutics: Consultancy; Takeda: Consultancy; Celgene: Consultancy; Amgen: Consultancy; Janssen: Consultancy; Active Biotech: Consultancy. Mangan:janssen: Speakers Bureau; celgene: Speakers Bureau; takeda: Speakers Bureau; amgen: Speakers Bureau. Stadtmauer:Celgene: Consultancy; Takeda: Consultancy; Janssen: Consultancy; Amgen: Consultancy; Novartis: Consultancy, Research Funding; Tmunity: Research Funding; Abbvie: Research Funding. OffLabel Disclosure: venetoclax for myeloma