6.1 Introduction
The cluster of differentiation antigen 19 (CD19) is a 95 kD transmembrane glycoprotein ubiquitously expressed on B cells from pro-B to mature B-cell phenotypes, thus making it an optimal target for targeted cellular therapy against all B-cell non-Hodgkin lymphomas (B-NHL)/chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) and B-cell acute lymphoblastic leukemia (B-ALL). CD19 is not expressed on other hematopoietic, or organ, cell populations. Targeting CD19 can hypothetically result in prolonged B-cell aplasia. Given the clinical experience with the anti-CD20 monoclonal antibody rituximab with temporary B-cell aplasia, severe clinical consequence has not been observed. Intravenous gamma globulin has proven to effectively supplement humoral immunity in hypogammaglobulinemic patients. Thus, CD19 continues to serve as an acceptable tumor antigen to target with cellular therapy. Genetically engineered recombinant T-cell receptors directed against a specific tumor antigen (chimeric antigen receptors, CARs) can recognize and kill tumor cell targets. This review will focus on the clinical experience of targeting CD19 with CAR-modified T cells (19-CAR-T) for B-cell lymphomas, excluding CLL/SLL and multiple myeloma.
The initial CAR constructs consisted of a single-chain antigen recognizing variable fragment (scFv) extracellular domain from an antibody in conjunction with a transmembrane link to a functional CD3ζ intracellular signaling domain (Eshhar et al. 1993). While this initial design had demonstrable effector function, proliferation and expansion were greatly enhanced by the incorporation of signal transmembrane co-stimulatory domains into later-generation constructs (Hombach et al. 2001). This translated into marked improvement of antitumor efficacy in early animal models (Brentjens et al. 2003). The clinical experience of 19-CAR-T for relapsed and/or refractory (rel/ref) B-NHL and Hodgkin lymphoma (HL) reviewed in this chapter will largely focus on “second-generation” 19-CAR-T constructs with TCR/CD3 signal 1 coupled to signal 2 co-stimulation with either CD28 or 4-1BB. Additionally, there will be brief mention of data with other tumor targets including kappa light chain and CD30 as well as later-generation constructs.
6.1.1 Clinical Studies: 19-CAR-T for Patients with Rel/Ref B-NHL
The initial clinical experience in 19-CAR-T for patients with rel/ref follicular lymphoma (FL) n = 2 and diffuse large B-cell lymphoma (DLBCL) n = 2 was with a first-generation construct from the City of Hope (Jensen et al. 2010). The two DLBCL patients received 19-CAR-T 1 month following high-dose therapy and autologous stem cell transplantation (HDT-ASCT), and one of the two patients had progressive disease following CAR-T infusion. The two patients with FL experienced disease progression following therapy. Interleukin-2 (IL-2) was infused adjunctively in the two patients with FL with the intent of providing proliferative stimulus to the CAR-T cell. Significant toxicity was not observed, and 19-CAR-T failed to persist with only one of four patients demonstrating persistence of 19-CAR-T in the peripheral blood at 1 week post-infusion.
The National Cancer Institute (NCI) was the first group to publish a case report of a second-generation 19-CAR-T incorporating a CD28 co-stimulatory domain (along with exogenous IL-2) in a patient with FL (Kochenderfer et al. 2010). The patient experienced a partial remission (PR) lasting approximately 10 months, and the 19-CAR-T product was noted to persist for >6 months. This group subsequently updated their prospective experience of 19-CAR-T incorporating CD28 co-stimulation for re/ref B-NHL preceded by lymphodepleting chemotherapy consisting of cyclophosphamide 60–120 mg/kg and fludarabine at a total dose of 125 mg/m2 (Kochenderfer et al. 2015). All six patients with indolent B-NHL (including splenic marginal zone lymphoma n = 1 and CLL/SLL n = 4) responded with either a partial (PR, n = 2) or complete remission (CR, n = 4). Six of the seven evaluable patients with DLBCL, including three patients with primary mediastinal B NHL, responded with either a PR (n = 2) or CR (n = 4). Two of the patients were not evaluable for a response and one patient experienced stable disease. The longest durations of responses were 12 and 23 months from treatment for DLBCL and indolent B-NHL, respectively. Peripheral expansion of the 19-CAR-T product peaked at 7–17 days. Because of toxicity, predominately in the form of cytokine-release syndrome (CRS), the dose of 19-CAR-T was subsequently lowered from 5 × 106/kg to 1 × 106/kg. Greater than or equal to grade 3 toxicity, again predominately CRS, was observed in 13 of 15 patients. A subsequent study by the same group of lower-dose chemotherapy (cyclophosphamide 900 mg/m2 and fludarabine 90 mg/m2) presented at the American Society of Hematology (ASH) meeting in 2014, noted less toxicity related to severe CRS with six of nine patients responding to 19-CAR-T therapy (Kochenderfer et al. 2014). They have subsequently updated their experience reporting increased IL-15 post-infusion correlated to CAR-T expansion and clinical response (Kochenderfer et al. 2017a) with remissions lasting greater than 4 years (Kochenderfer et al. 2017b). Following these studies, this CAR-T construct was licensed to Kite Pharma as KTE-C19 and subsequently axicabtagene ciloleucel (axi-cel), a 19-CAR-T with CD28 co-stimulation following fludarabine and cyclophosphamide conditioning, for rel/ref B-NHL with initial presentation by Locke et al. at ASH 2015 wherein six patients had been treated and three patients experienced >grade 3 toxicity including a grade 4 encephalopathy and grade 4 CRS. Three patients were evaluable for response, in short follow-up at 1 month, with ORR of 100% (CR, n = 2; PR, n = 1) (Locke et al. 2015). This leads to the multicenter phase II, Zuma-1 trial that led to FDA approval of the first CAR-T-cell therapy for the treatment of adult patients with rel/ref large B-cell lymphoma after at least two lines of standard therapy. In this study, axi-cel was administered to 101 patients with DLBCL (n = 77), primary mediastinal B-cell lymphoma, or transformed FL (n = 24) following low-dose fludarabine and cyclophosphamide conditioning. At 6-month follow-up, ORR was 82% (CR, n = 55 [54%]; PR, n = 28 [28%]. At median follow-up period (15.4 months), 42% of patients maintained a response (CR, 40%). The overall rate of survival at 18 months was 52%. Ninety-three percent of all patients experienced CRS, while 64% experienced neurotoxicity (NT), which were all resolved. Forty-three percent of patients were treated with tocilizumab, and 27% received glucocorticoids for the management of CRS and/or NT with no apparent effect on response rates (Neelapu et al. 2017).
The group at the University of Pennsylvania recently published their phase IIa study treating chemorefractory FL and DLBCL patients with 19-CAR-T (Schuster et al. 2017a). In contrast to a CD28 second-signal co-stimulatory domain, their construct incorporates a 4-1BB co-stimulatory transmembrane molecule. Patients were treated with variable lymphodepleting chemotherapy per treating physician prior to administration of 19-CAR-T. This report included 14 evaluable patients with FL and 14 evaluable patients with DLBCL. Of the 28 patients treated with 19-CAR-T, 18 (64%) had a response. CR was achieved with 6 of the 14 patients (43%) with DLBCL and 10 of the 14 patients (71%) with FL. Continuous response was observed at the median follow-up of 28.6 months in 86% of patients with DLBCL and in 89% of patients with FL. Following infusion, median peak expansion of 19-CAR-T cells was 8 days in patients with a response and 10 days in patients without a response. Fourteen of 16 patients who achieved CR had PCR-detectable levels of 19-CAR-T DNA between 6 and 24 months after infusion. Eight of 16 patients in CR proceeded to sustained B-cell recovery at median of 6.7 months. The investigators observed severe CRS in five patients (18%) and severe encephalopathy in three patients (11%) among which two cases were self-limiting and one case was fatal. This treatment was further explored in a multicenter, multinational phase II study (The JULIET Trial) presented at ASH 2017 (Schuster et al. 2017b). The investigators reported best ORR of 53% with 39.5% CR rate in 81 patients infused and evaluable. The CR rate at 6 months was 30% with CTL019 detectable in the peripheral blood up to 1 year post-infusion. CRS occurred in 58% of patient and 23% grade 3–4. They reported a 12% incidence of neurotoxicity. This study led to FDA approval of Novartis Pharmaceuticals licensed, tisagenlecleucel, on May 1, 2018 for rel/ref adult DLBCL following at least two lines of chemotherapy.
The group from the Fred Hutchinson Cancer Research Center (FHCRC) published their experience of treating rel/ref B-NHL with lentivirus transduced 19-CAR-T with a secondary 4-1BB co-stimulatory molecule in a fixed 1:1 ratio of CD4:CD8, based largely upon preclinical modeling of improved persistence and efficacy of the 19-CAR-T product (Riddell et al. 2014). These investigators tested the effects of adding fludarabine to conditioning regimen preceding variable doses of 19-CAR-T infusion. In the cyclophosphamide and fludarabine conditioning arm, 18 B-NHL patients achieved an ORR of 72% (CR, n = 9; PR, n = 4). In the cyclophosphamide-based conditioning without fludarabine arm, 12 B-NHL patients achieved an ORR of 50% (CR, n = 1; PR, n = 5) (Turtle et al. 2016). Of the total 32 evaluable patients, 20 developed any grade CRS, and 9 developed severe NT associated with treatment. Severe CRS was observed in four patients, all of whom had received cyclophosphamide and fludarabine conditioning. There was also a correlation between 19-CAR-T dose and severe CRS and NT. Thus, adding fludarabine to the conditioning regimen and treatment with higher doses of 19-CAR-T correlated with increasing toxicities. Three of six patients (50%) treated at the highest dose (2 × 107/kg) after cyclophosphamide and fludarabine conditioning developed severe CRS, and four of six patients (67%) developed severe NT. Of note, peak serum concentrations of IL-6, interferon gamma (IFN-γ), ferritin, and C-reactive protein correlated with development and severity of CRS. The highest IL-6 and IFN-γ levels were seen in patients who received cyclophosphamide and fludarabine conditioning followed by infusion of the highest 19-CAR-T dose. This technology was subsequently licensed to Juno Therapeutics, and a multi-institutional trial was initiated. In a preliminary update at the 2017 ASH meeting, investigators reported all grade CRS rate at 30% (12/69), severe CRS rate at 1% (1/69), and NT rate at 20% (14/69) (Abramson et al. 2017). At the time of the report, best ORR was 75% (51/68), with a CR at 56% (38/68).
Clinical studies of 19-CAR-T for relapsed or refractory B-NHL
Institution | Viral vector | Co-stimulatory molecule | Conditioning | Patients | Clinical responses | Notes |
|---|---|---|---|---|---|---|
NCI (Kochenderfer et al. 2015) | Gamma retrovirus | CD28 | Cyc 60–120 mg/kg Flu 125 mg/m2 | iB-NHL n = 6, aB-NHL n = 9 | iB-NHL 6/6 (CR = 4, PR = 2) aB-NHL 6/7 (CR = 4, PR = 2) inevaluable n = 3 | 13/15 >grade 3 toxicity |
NCI (Kochenderfer et al. 2014) | Gamma retrovirus | CD28 | Cyc 90 mg/m2 flu 90 mg/m2 | FL n = 1 DLBCL n = 8 | FL PR DLBCL 6/8 (CR = 1, PR = 5) | No patients required vasopressor or MV |
Gamma retrovirus | CD28 | Cyc 900-1500 mg/m2 flu 90 mg/m2 | DLBCL n = 19 FL n = 2 MCL n = 1 | 73% ORR (55% CR; 18% PR) | IL-15 levels correlated with response | |
UPenn (Schuster et al. 2017a) | Lentivirus | 4-1BB | Variable | FL n = 14 DLBCL n = 14 | FL and DLBCL 64% ORR (FL 71% CR DLBCL 43% CR) | 8/28 >grade 3 toxicity |
FHCRC (Turtle et al. 2016) | Gamma retrovirus | 4-1BB | cyc 2–4 g/m2 cyc 2–4 g/m2 + VP-16 300–600 mg/m2 cyc 60 mg/kg + flu 75 mg/m2 | B-NHL flu(+) n = 18 flu(−) n = 12 | flu(+) 72% ORR flu(−) 50% ORR | Improved efficacy and increased toxicity with fludarabine |
Moffitt (Locke et al. 2015) | Gamma retrovirus | CD28 | cyc + flu | n = 6 | All 3 evaluable at one month (CR = 2, PR = 1) | n = 3 >grade 3 toxicity attributable to CAR-T |
Multicenter (Novartis) (Neelapu et al., 2017) | Gamma retrovirus | CD28 | cyc + flu | n = 101 | 82% ORR (54% CR; 28% PR) | 93% CRS; 64% neurotoxicity, all of which resolved |
Multicenter (Kite) (Schuster et al. 2017b) | Lentivirus | 4-1BB | Variable | n = 81 | 53.1 % ORR (39.5% CR, 13.6% PR) | 58% CRS; 12% neurotoxicity |
Multicenter (Juno) (Abramson et al. 2017) | Gamma retrovirus | 4-1BB | cyc + flu | n = 68 | 75% ORR (56% CR) | 30% CRS; 20% neurotoxicity |
6.1.2 Clinical Studies: 19-CAR-T Following Hematopoietic Cell Transplantation (HCT)
6.1.2.1 19-CAR-T in Consolidation Following High-Dose Therapy and Autologous HCT (HDT-AHCT)
Clinical studies of 19-CAR-T for relapsed or refractory B-NHL in consolidation following HDT-ASCT
Institution | Viral vector | Co-stimulatory molecule | Conditioning | Patients | PFS | Notes |
|---|---|---|---|---|---|---|
City of Hope (Wang et al. 2016) | Lentivirus | “NHL1,” none “NHL2,” CD28 | BEAM | “NHL1” DBLCL n = 7, MCL n = 1 “NHL2” DLBCL n = 4, MCL n = 4 | “NHL1” 50% at 1-year “NHL2” 75% at 2 years | Improved expansion with second-generation “NHL2” 19-CAR-T |
MDACC (Kebriaei et al. 2016) | Transposon system | CD28 | Auto: BEAM Allo: variable | FL n = 3, DLBCL n = 4, HL n = 1, MCL n = 1, B-ALL n = 17 (auto, n = 7; allo, n = 19) | Auto-HSCT, 83% at 30 mos Allo-HSCT, 53% at 12 mos | 19-CAR-T persisted in blood on average 201 d (auto), 51 d (allo) |
MSKCC (Sauter et al. 2015) | Gamma retrovirus | CD28 | BEAM | n = 11 rel/ref DLBCL/transformed B-NHL in PR or marrow involvement | 40% at 12 months | 7/11 patients with >grade 3 neurotoxicity |
Two additional centers have presented their prospective experience, of 19-CAR-T following HDT-AHCT for B-NHL, MD Anderson Cancer Center (MDACC) and Memorial Sloan Kettering Cancer Center (MSKCC). The group from MDACC published their prospective experience with 19-CAR-T constructed by the Sleeping Beauty transposon with two phase 1 studies that enrolled 26 patients with advanced NHL or ALL (Kebriaei et al. 2016). Patients underwent HCT in autologous (n = 7) or allogenic (n = 19) settings followed by 19-CAR-T infusion as an adjuvant therapy. Patients who underwent auto-AHCT experienced progression-free and overall survival rates of 83% and 100% at 30 months, while the same end points for patients who received allo-HSCT were 53% and 63%, respectively, at 12 months. Three patients that received allo-HCT developed GvHD. One patient that developed GvHD died of preexisting conditions, while GvHD of the other two patients was resolved. Genomic analysis of 19-CAR-T cells generated by transposition showed uniform and stable insertion events through the genome with low rates of aberrant recombination. Of note, investigators could detect 19-CAR-T cells in blood for an average of 201 days in auto-HCT and 51 days in allo-HCT recipients.
Lastly, investigators from MSKCC have tested CAR-T in consolidation for high-risk rel/ref DLBCL/aggressive histology B-NHL in partial chemosensitive remission following a HDT-AHCT (Sauter et al. 2015). All patients on this study had either functional imaging-positive disease and/or bone marrow involvement characterizing them as high-risk per phase I study eligibility criteria. The 19-CAR-T utilized by this group includes a CD28 co-stimulatory molecule. Interim data presented at the 2015 ASCO meeting revealed 4 of 10 evaluable patients in continuous complete remission at a median of 14 months post-HDT-ASCT and 19-CAR-T and up to nearly 2 years in two patients, following study treatment (Sauter et al. 2015). The most common grade >3 toxicity attributable to 19-CAR-T was NT in 7/11 patients that was fully reversible.
6.1.2.2 19-CAR-T Following Allogeneic (Allo-HCT)
To test safety of 19-CAR-T post-allo-HCT, the group at the NCI reported a phase I dose escalation trial in B-NHL patients that progressed following allo-HCT (Brudno et al. 2016). The protocol did not include a lymphodepleting conditioning regimen before infusion. Of 20 patients, none developed GvHD after 19-CAR-T infusion, and an ORR of 40% (CR, n = 6; PR, n = 2) was achieved. The response rate was highest for ALL, and the longest ongoing CR was greater than 30 months in a patient with CLL. Anti-malignancy response of 19-CAR-T infusion was rapid; blood B-lymphocytes decreased from 3372 to 0/μl over 11 days in a case of CLL, compared to several weeks observed in standard donor lymphocyte infusion (DLI) treatment. Of note, after infusion 19-CAR-T cells had a significant elevation in programmed cell death protein 1 (PD-1) expression before reaching peak blood levels.
6.1.2.3 Later-Generation CAR Products
Two abstracts were presented at ASH 2015 with third- and fourth-generation constructs, respectively. Investigators from Sweden reported short interim follow-up on a phase I/II study testing a third-generation 19-CAR-T construct incorporating CD28 and IL-2 for rel/ref B-cell malignancies (including CLL and ALL). They reported 6/14 initial complete responses in lymphoma (n = 11) and ALL (n = 3) with investigation ongoing (Enblad et al. 2015). Lastly, investigators from China reported on a phase I/II clinical trial of a fourth-generation 19-CAR-T construct consisting of CD28/CD137/CD27 and iCasp9 apoptosis-inducible safety switch (4SCAR19) (Chang et al. 2015). Thirteen patients with rel/ref B-NHL (including 12 with either DLBCL or Burkitt’s lymphoma) were treated with 4SCAR19 preceded by fludarabine and cyclophosphamide conditioning, and eight experienced a CR at 3–10 months post-4SCAR19 treatment. Three patients died of non-disease-related causes with or without severe CRS, and two patients died of progressive lymphoma with a 120-day disease-free survival of 53% (95% CI: 36–69%) (Chang et al. 2015).
6.1.3 Clinical Studies: Alternate Tumor Antigen Targets for CAR Therapy (CD20, Kappa Light Chain, CD30, and CD22)
The earliest clinical experience with CAR-T therapy for B-NHL was with a scFv targeting CD20 in a first-generation construct (Till et al. 2008). In conjunction with adjunctive IL-2, the 20-CAR-T persisted in vivo up to 9 weeks post-infusion. Of the seven patients treated with MCL or FL, two patients achieved a CR, one patient a PR, and four had experienced stable disease. A second study utilized a first-generation neomycin-resistance selected 20-CAR-T following HDT-ASCT detected 20-CAR-T up to only 1 week post-infusion by quantitative polymerase chain reaction (qPCR), and no clinical responses were detected (Jensen et al. 2010). Given that B cell NHL is clonally restricted to either kappa (κ) or lambda (λ) immunoglobulin light chain, the group from Baylor College of Medicine investigated targeting κ-light chain with CAR-T and presented results on seven patients with rel/ref B-NHL at the 2013 ASH meeting (Ramos et al. 2013). Per PCR, the κ-CAR-T peaked in the periphery at 1–2 weeks post-infusion and persisted for up to 6 months. Three of the seven patients responded to κ-CAR-T (CR, n = 2, PR, n = 1).
With the success of the antibody-drug conjugate brentuximab vedotin (BV) targeting CD30 in patients with CD30+ hematologic malignancies including Hodgkin lymphoma (HL) and anaplastic large cell lymphoma (ALCL) (Younes et al. 2010), the group from Baylor has developed a CAR construct targeting CD30 (30-CAR-T) in conjunction with a functional CD3ζ and CD28 co-stimulatory domain transduced via a retrovirus. They reported interim results of a prospective phase I study of 30-CAR-T wherein 18 products were transduced and 9 patients were treated (HL, n = 7, ALCL, n = 2). Eight of the patients had previously failed BV. Importantly, no patients received lymphodepleting chemotherapy prior to infusion of 30-CAR-T on study. Following safe administration to dose level #3 on the phase I study, at 6-week evaluation post-infusion, n = 1 CR and n = 1 PR, four patients with stable disease and three patients with progression of disease (Ramos et al. 2015). Given the safe administration, the Baylor investigators plan to incorporate 30-CAR-T following HDT-ASCT in the subsequent study.
CD22 emerged as a target tumor antigen for CAR-T therapy in B-ALL, as its expression is largely restricted to B-cell lineage. Recently, the first clinical experience of 22-CAR-T therapy in 21 patients (age range 7–30 years) with B-ALL was reported by the investigators at the NIH (Fry et al. 2018). The 22-CAR-T construct included a 4-1BB co-stimulatory domain, and patients were administered doses ≥1 × 106 per kg. All patients had at least one bone marrow transplantation, and notably 17 patients had received CD19-targeted immunotherapy wherein 15 had received 19-CAR-T therapy prior to this trial. A dose-dependent anti-malignant response was observed. Of 21 patients, 12 (57%) achieved CR, 9 of whom had received prior CD19-directed immunotherapy and had CD19 diminished or CD19-negative B-cell populations. Eight patients in CR experienced relapse at a median of 6 months following 22-CAR-T infusion. Given their experience with CD22 as a valid CAR-T therapy antigen, the investigators plan to extend their studies to CD19-CD22 multispecific CAR therapy to decrease the possibility of relapses associated with antigen escape.
6.2 Expert Point of View
Despite the above encouraging data, 19-CAR-T therapy for B-NHL appears less active than in B-ALL wherein the vast majority of patients achieve CR (Davila et al. 2014; Maude et al. 2014). Whether this is due to differences in the microenvironment (marrow- versus nodal-based disease) or other biologic features between B-ALL and B-NHL histologies remains unknown. Most recently, to analyze the genomic, phenotypic, and functional mechanisms of success or failure of CAR-T-cell therapy, the group at the University of Pennsylvania presented a trial of 41 patients with advanced and high-risk CLL, who received at least one dose of 19-CAR-T cells (Fraietta et al. 2018). They reported that intrinsic properties of 19-CAR-T isolated from patients who responded to CAR-T-cell therapy were markedly different than 19-CAR-T cells isolated from patients who were unresponsive. 19-CAR-T cells of the responders had elevated expression profiles of early memory differentiation, as well as comparatively enriched IL-6 signatures, and these cells had superior expansion during clinical manufacturing, while the 19-CAR-T cells of the unresponsive patients had elevated expression of late memory, apoptosis, and aerobic glycosylation that are associated with T-cell exhaustion, as well as poor expansion profiles. Of note, CD27+PD1−CD8+ 19-CAR-T-cell population expressing high levels of the IL-6 receptor correlated with a therapeutic response.
Currently, the two major toxicities of 19-CAR-T therapy include CRS and NT manifestations including, but not limited to, seizures, seizure-like activity, focal motor deficits, aphasia, and global encephalopathy (Lee et al. 2014). These toxicities temper the encouraging activity of this treatment modality, and strategies to abrogate the associated morbidity (and potential mortality) are mentioned in the section below. Lastly, it is important to note that many of the previously reviewed studies are in short follow-up. To this end, it is important to await longer follow-up from phase II studies. Given the time and resource necessary for autologous leukapheresis and CAR-T production, it will be imperative to analyze forthcoming efficacy data in later phase studies by intention to treat. This is particularly relevant in the rel/ref setting of aggressive histology disease, i.e., DLBCL, wherein patients’ disease phenotype and natural history may preclude proceeding to CAR-T treatment.
6.3 Future Directions
Much of the clinical development around CAR-T therapy is strategies to prevent or treat toxicity related to treatment, most notably the use of anti-IL-6 receptor blockade to abrogate CRS (Davila et al. 2014). Additionally, engineering suicide genetic elements to “turn off” the activated cellular product when toxicity is observed are being developed (Di Stasi et al. 2011). Safety and management cohort of the ZUMA-1 trial reported results of an IL-6 receptor blocker, tocilizumab, used as CRS prophylaxis at day 2 of infusion. Rates of patients with grade ≥3 CRS were lower in the prophylaxis cohort, 1 of 34 (3%), compared to 13 of 101 (13%) in the main cohort (Locke et al. 2017). However, a concern of potentially greater severe NT in the experimental group may provide insight into the pathophysiology of NT in the setting of IL-6 receptor blockade. A recent report from MSKCC described factors associated with NT in ALL patients receiving 19-28z CAR-T including cytokines such as IL-6, as well as others that may be produced by other cellular lineages (Santomasso et al. 2018). Another complication of CAR-T therapy are infections that are potentially related to prior cytotoxic treatments and/or lymphodepleting conditioning regimens. A recent phase I/II study that enrolled 133 patients reported that incidence of infections after 19-CAR-T-cell therapy was comparable to other salvage chemo-immunotherapies. Prior cytotoxic treatments, 19-CAR-T dose, CRS severity, and ALL malignancy were associated with more frequent infections (Hill et al. 2018).
Future investigation toward improvement in 19-CAR-T efficacy for B-NHL may involve additional pharmacologic adjuncts to catalyze the therapeutic potential of this adoptive cellular therapy. The Bruton’s tyrosine kinase inhibitor ibrutinib, which has demonstrated impressive single-agent activity in many histologies of B-NHL (Smith 2015), has been found to inhibit Th2 responses while enhancing Th1-based immunity via inhibition of the interleukin-2-inducible kinase (ITK) in preclinical models. Subsequent to this discovery, CLL patients previously exposed to ibrutinib demonstrated enhanced ex vivo and in vivo expansion of 19-CAR-T in addition to decreased expression of programmed cell death 1 (PD-1) on the product (Fraietta et al. 2016; Long et al. 2017). PD-1 is a T-cell exhaustion receptor serving as a downregulator of T cells upon engagement of PD-1 ligand-1 (PD-L1) or PD-L2 and is upregulated on adoptive transfer of CAR-T (Abate-Daga et al. 2013). This receptor is pharmacologically targetable by checkpoint inhibitors in active clinical investigation for various lymphoma histologies (Matsuki and Younes 2016). Clinical trials combining checkpoint inhibitors and 19-CAR-T are being designed. A clinical experience of 19-CAR-T in CLL patients that had received ibrutinib was reported by the FHCRC investigators wherein 24 patients achieved an ORR of 71% (Turtle et al. 2017). Furthermore, in a case report, ongoing treatment with a PD-1 inhibitor antibody after 19-CAR-T therapy resulted in a durable CR response in a DLBCL patient that was otherwise unresponsive to 19-CAR-T (Chong et al. 2017). Ibrutinib has also been shown to enhance 19-CAR-T cytotoxic killing of MCL in cell lines, in vivo, as well as in xenograft mouse models (Ruella et al. 2016). Additional potential combinatorial strategies could include immune modulatory agents, such as lenalidomide, which has previously been shown to enhance T-cell synapse formation and downregulation of tumor cell inhibitory molecules (Ramsay et al. 2012) and has demonstrated enhanced antitumor efficacy of 19-CAR-T and 20-CAR-T in animal models (Otahal et al. 2016).
Lastly, active development of third- and later-generation CAR-T is ongoing. Included in these investigations is development of constructs with co-stimulatory elements and/or lymphoproliferative cytokine genes engineered into the 19-CAR product (Pegram et al. 2012). Additionally, combinatorial antigen specificity is under active investigation (Kloss et al. 2013; Zah et al. 2016) toward the goal of circumnavigating antigen escape (Jackson and Brentjens 2015; Gardner et al. 2016).