Anti-CD30 CART cells as consolidation after autologous haematopoietic stem-cell transplantation in patients with high-risk CD30+ lymphoma: a phase 1 study

Abstract

Background:

Chimeric antigen receptor T cells targeting CD30 (CD30.CAR) are safe and with promising activity when preceded by lymphodepleting chemotherapy. We aimed to determine the safety of CD30.CAR-T as consolidation after autologous stem cell transplant (ASCT) in patients with CD30⁺ lymphoma at high risk of relapse.

Methods:

In a phase 1 dose-escalation study, patients with high risk classical Hodgkin lymphoma (HL) and CD30⁺ non-Hodgkin lymphoma (NHL) received CD30.CAR-T cell infusion as consolidation after carmustine, etoposide, cytarabine, and melphalan (BEAM) ASCT. Patients aged 3 and older with a Kanofsky performance score >60% planned for ASCT were eligible if they were considered high risk for relapse as defined by primary refractory disease or relapse within 12 months of initial therapy or extranodal involvement at start of pre-transplant salvage therapy. Patients received a single infusion of CAR-T cells (2×10⁷ CAR-T cells/m², 1×10⁸ CAR-T cells/m², or 2×10⁸ CAR-T cells/m²) after trilineage hematopoietic engraftment defined as absolute neutrophil count ≥500/mm³ for 3 days, platelets ≥25×10⁹/L and hemoglobin ≥ 8 g/dL without transfusion for 5 days. The primary endpoint was the determination of the maximum tolerated dose (MTD) which was based on the rate of dose limiting toxicity (DLT) in patients who received CAR-T cell infusion. This study is registered with ClinicalTrials.gov (#NCT02663297) and study enrollment is complete.

Findings:

Between June 7, 2016 and November 30, 2020 twenty-one patients were enrolled and 18 patients (11 HL, 6 T cell lymphoma, 1 grey zone lymphoma) were infused with CD30.CAR-T cells at a median of 21.5 days (range 17–44 days) after ASCT. There were no DLTs observed, so the highest dose tested, 2×10⁸ CAR-T cells/m², was determined to be the MTD. One patient had grade 1 cytokine release syndrome. The most common grade 3–4 adverse events were lymphopenia (2 (11%) of 18) and leukopenia (2 (11%) of 18). There were no treatment-related deaths. At a median follow up of 48 months (IQR: 27.5 to 61 months) post infusion, the median progression free survival (PFS) for all treated patients (n=18) was 32 months and the median PFS for treated HL patients (n=11) has not been reached. The 2-year PFS for HL patients was 73%. The median overall survival for all patients has not been reached.

Interpretation:

CD30.CAR-T cell infusion as consolidation after BEAM ASCT is safe, with low rates of toxicity and encouraging preliminary activity in HL patients at high risk of relapse, highlighting the need for larger studies to confirm these findings.

INTRODUCTION

Although the majority of patients with classical Hodgkin lymphoma (HL) are cured with frontline therapy, about 10–25% of patients have disease refractory to initial therapy or relapse after initial complete response (CR).¹ The standard of care for patients with HL after first relapse is high dose chemotherapy followed by autologous stem cell transplant (ASCT) based on randomized clinical trials which showed a significant improvement in progression free survival (PFS),² with about half of the patients treated cured with this approach.³ More recently, maintenance therapy with the anti-CD30 antibody drug conjugate brentuximab vedotin (BV) after ASCT improved PFS in patients with high-risk features, with a 2-year PFS of 63% compared to 51% with placebo.⁴ However, with the incorporation of BV in earlier lines of therapy,⁵ and toxicities frequently associated with maintenance BV after ASCT, including neuropathy and neutropenia,⁴ alternative approaches are needed as consolidation therapy after ASCT in patients at high risk of relapse.

We and others have shown high response rates with minimal toxicity in heavily pre-treated patients with relapsed/refractory HL who received chimeric antigen receptor T-cells targeting CD30 (CD30.CAR) after lymphodepleting chemotherapy.^6–8 In addition to HL, which universally expresses CD30, there are subsets of non-Hodgkin lymphomas (NHL) which express CD30, such as anaplastic large cell lymphoma (ALCL) and peripheral T cell lymphomas. These patients are also often treated with ASCT either in first remission or at relapse, although their prognosis is generally worse than for patients with HL.⁹ CD30 CAR-T cells have also demonstrated activity in patients with CD30⁺ T cell lymphomas.¹⁰

Given the excellent safety profile and promising activity of CD30.CAR-T cells in CD30⁺ lymphomas, we hypothesized that autologous CD30.CAR-T cells could be safely infused as consolidation therapy after carmustine, etoposide, cytarabine, and melphalan (BEAM) conditioning with ASCT and replace the use of BV in patients at high risk of relapse. We thus conducted a phase 1 clinical trial in which CD30.CAR-T cells were infused after stem cell engraftment in patients with HL and CD30⁺ NHL at high risk for disease recurrence after ASCT.⁴ Here, we demonstrate that this approach is safe and has comparable activity to BV maintenance therapy after ASCT. Furthermore, we observed that CD30.CAR-T cells infused after BEAM conditioning and stem cell engraftment expand in the peripheral blood as observed in HL patients in whom CD30.CAR-T cells are infused after lymphodepleting chemotherapy.

METHODS

Study design and participants

This was a phase 1, multi-centre, single arm study performed at 2 sites in the USA (see appendix, page 4, for participating sites). All subjects received their CAR-T cell infusion at the University of North Carolina. Patients (aged ≥ 3 years) with recurrent HL or CD30⁺ NHL who were planning to undergo high dose chemotherapy and ASCT were eligible for treatment. Patients were required to have CD30⁺ disease as documented by immunohistochemistry either at diagnosis or at relapse and at least one of the following high-risk features: failure to achieve CR post initial treatment, initial remission duration of <12 months, or extra-nodal involvement at start of pre-transplant salvage therapy, which were the risk factors included in the AETHERA trial⁴. Patients could not have received any investigational agents within 6 weeks prior to infusion or BV within 4 weeks prior to infusion and had to be on less 10 mg per day of prednisone or its equivalent. Patients were required to have adequate performance status (Karnofsky or Lansky score >60%) and organ function (cardiac and pulmonary function adequate for ASCT, bilirubin ≤ 1.5 times upper limit of normal (ULN), aspartate aminotransferase ≤ 3 times ULN, serum creatinine ≤ 1.5 times ULN, and pulse oximetry >90% on room air) at time of procurement and cell infusion to be eligible for ASCT. Patients could not be pregnant or lactating, due to concern for fetal toxicity of ASCT and unknown effects of CAR-T cells and had agreed to use effective contraception from the time of consent until 6 months post CAR T cell infusion. Patients were excluded if they had active infection with HIV, HTLV, Hepatitis B or Hepatitis C. Patients underwent cell collection for both transplant and CAR-T cell manufacturing. Collection for CAR-T cells was via whole blood for 18 patients and apheresis for 3 patients, and was performed prior to mobilization and collection for ASCT. CD30.CAR-T cells were manufactured in a Good Manufacturing Practice compliant facility at the University of North Carolina (UNC) as previously described using a gamma retroviral vector and including a CD28 costimulatory domain.⁶ Patients were eligible to receive CD30.CAR-T cell infusion post trilineage hematopoietic engraftment after BEAM ASCT, defined as ANC ≥ 500/mm³ x 3 days, platelets ≥ 25×10⁹/L and Hb ≥ 8 g/dL without transfusion for 5 days. This study was conducted in accordance with the Declaration of Helsinki and International Conference on Harmonization Guidelines for Good Clinical Practice and was approved by the institutional review board at UNC. All patients or parents/legal guardians provided written informed consent.

Procedures

After collection of blood for manufacturing, patients underwent BEAM ASCT and post hematopoietic engraftment, they received a single intravenous dose of CAR-T cells on one of 3 dose levels (DL1 = 2×10⁷ CAR-T cells/m², DL2 = 1×10⁸ CAR-T cells/m², DL3 = 2×10⁸ CAR-T cells/m²). Pediatric patients were not allowed to be enrolled on a dose level until 2 adult subjects completed follow up at that dose. A pre-specified maximum of 18 subjects were treated.

Patients were monitored clinically for adverse events and with a complete blood count and differential and comprehensive metabolic panel on week 1, week 2, week 3 (optional assessment), week 4, week 6, month 3, month 6, month 9, and month 12, followed by an annual physical exam up to year 5. CRS was graded according to the Lee criteria.¹² All other toxicities were graded according to the National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE, version 4.0). Imaging to assess disease status ¹³was performed at 6 weeks and then every 6 months until 2 years post CAR-T cell infusion. Disease status was assessed per the Lugano criteria¹³ by the investigator.

Peripheral blood mononuclear cells (PBMC) were obtained at predefined time points to evaluate for the presence of CAR-T cells and monitor cytokines in the plasma. Expansion and persistence of CD30.CAR-T cells was determined by quantitative PCR (qPCR) of DNA extracted from PBMC, as previously described.⁶ Immune phenotype characterization of PBMCs was also performed using flow cytometry. Plasma cytokines were monitored using Luminex assay, while CCL17 (thymus and activation regulated chemokine [TARC]) levels were measured using a specific ELISA, as previously reported.⁶

Outcomes

The primary objective was to determine the safety and tolerability, and to estimate the maximum tolerated dose (MTD) of CD30.CAR-T cells in patients with CD30⁺ lymphomas at high risk for relapse. Dose limiting toxicities (DLTs) were evaluated from date of infusion to 6 weeks post infusion and defined as grade 3 or 4 cytokine release syndrome (CRS) lasting >72 hours, grade 3 or higher non-hematologic toxicity, or treatment-related death considered possibly, probably, or definitely related to study cellular products per the opinion of the study team. At initial study design, there was also a DLT definition for hematologic toxicity defined as failure to achieve stable engraftment of platelets >50K without transfusion and ANC > 1K by 6 weeks following CAR-T cell infusion, but this was removed in July 2nd, 2019, at the time that hematologic parameters were removed from DLT definitions for other protocols based on the experience of hematologic reconstitution obtained from other CAR-T cell clinical trials. However, none of the patients enrolled before and after this change would have had a DLT based on these criteria. MTD was defined as the dose that caused DLT in 10% of eligible cases.

Secondary objectives included PFS (defined from day of ASCT to relapse or progression, or death as a result of any cause), overall survival (OS) (defined from date of administration of CAR-T cells to date of death), and expansion and persistence of CD30.CAR-T cells in the peripheral blood. Persistence of CAR-T cells was determined by quantitative PCR . Exploratory objectives included measurement of patient reported outcomes in patients treated with CAR-T cells, which will be reported separately.

Post-hoc subgroup analyses included measurement of PFS and OS in patients with HL.

Statistical analyses

This was a phase 1 dose-escalation trial designed to evaluate the safety of ATLCAR.CD30 cells. Dose escalation was based on the continual reassessment method (CRM)¹¹, which was implemented with the power model as a working model with constants (0.0281, 0.0568, 0.1) and an exponential prior with mean 1 on the model parameter. The initial escalation was in cohorts of 2 until a DLT was observed. Skipping doses while escalating was not allowed. Based on the CRM framework designed for this study, it was determined that at least 12 patients would be treated, with an additional 6 patients possible, for a maximum of 18 treated patients. All analyses for this phase 1 study were per protocol and all patients who received CAR-T cell infusion were included in statistical analyses for safety. Patients’ baseline characteristics were summarized using descriptive statistics, and patient toxicities were reported by way of frequency tables. The reported toxicity represents the maximum grade per patient, per toxicity group, with no duplicates and that have been determined to have ‘definite’, ‘probable’, or ‘possible’ attribution to treatment. For secondary endpoints, Wilson’s method was used for calculating 95% confidence intervals for the response rates. The Kaplan-Meier method was used to estimate the time to event functions of OS and PFS. Median follow-up time for survivors was calculated and reported along with the 25^th and 75^th percentiles (IQR). . OS was calculated using the CAR-T infusion date to the date of death from any cause, or date of last contact (censored). Only patients who received their CAR-T cell infusion were included in OS calculations. PFS was calculated from the date of ASCT to either the date of progression, the date of death from any cause, or the date of last contact (censored). We performed PFS calculations for all patients who were collected with the intent to treat with CAR-T cells as well as just for patients who received the infusion. The ‘loglog’ method (based on the log of the hazard) was used for calculating 95% confidence intervals for median OS and PFS. Statistical analyses were performed using both SAS (version 9.4, Cary, NC) and R (2019): A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. This study is registered with ClinicalTrials.gov (#NCT02663297).

Role of the funding source

The funder of the study had no role in study design, data collection, analysis, or interpretation, or writing of the report.

RESULTS

Between June 7, 2016 and November 30, 2020, 21 patients were screened and enrolled, with 18 patients eventually receiving a single dose of CD30.CAR-T cells (Figure 1, Table 1). Three patients were collected for CAR-T cell manufacturing, but not treated (1 due to change in treatment plan, 1 due to product contamination at collection, and 1 due to failure of cells to expand). Only patients who received CAR-T cell infusion were included in the primary analysis. Four patients were treated on DL1 (including 2 pediatric patients), 5 patients were treated on DL2, and 9 patients were treated on DL3. Eleven patients had HL, 6 patients had CD30⁺ T cell lymphoma, and 1 patient had grey zone lymphoma. The median age for treated patients was 43 (range 16–76 years). Fifteen patients (83%) received one line of salvage therapy prior to ASCT, while 3 patients (17%) required a second line of therapy. Fifteen patients (83%) had received BV prior to ASCT. Nine patients (50%) were refractory to frontline therapy and 5 patients (28%) relapsed within 12 months of completing initial therapy. Ten patients (56%) had extra-nodal involvement at time of relapse. Sixteen patients (89%) were in CR prior to transplant with 2 patients (11%) in partial response (PR). Patients received CD30.CAR-T cells at a median of 21.5 days (range 16–44 days) following ASCT.

Figure 1:

Trial Flow Chart

Table 1:

Characteristics of patients receiving CD30 CAR-T cell infusion following BEAM ASCT

Characteristic	All Collected (N= 21)	All Treated (N = 18)	HL only (N=11)
Age, years, Median (range)	45 (12–76)	43 (16 – 76)	24 (17–61)
Sex, male	16 (76%)	15 (83%)	9 (82%)
Race
Asian	2 (10%)	2 (11%)	2 (18%)
Black	1 (5%)	1 (6%)	0
White	18 (86%)	15 (83%)	9 (81%)
Disease
HL	13 (62%)	11 (61%)
ALK-positive ALCL	3 (14%)	3 (17%)
ALK-negative ALCL	1 (5%)	1 (6%)
AITL	1 (5%)	1 (6%)
PTCL, NOS	1 (5%)	1 (6%)
Grey zone lymphoma	1 (5%)	1 (6%)
CD30+ Large B-Cell Lymphoma	1 (5%)	0
ECOG performance status
0	15 (71%)	15 (83%)	10 (91%)
1	6 (29%)	3 (17%)	1 (9%)
Number of systemic salvage therapies prior to ASCT
1	17 (81%)	15 (83%)	9 (81%)
2	4 (19%	3 (17%)	2 (18%)
Frontline Therapy
ABVD*	10 (48%)	9 (50%)	8 (73%)
BV-AVD	1 (5%)	1 (6%)	1 (9%)
CHOEP	2 (10%)	2 (11%)	0
BV-CHP	2 (10%)	2 (11%)	0
Other**	6 (29%)	4 (22%)	2 (18%)
Salvage Therapy***
BV-benda	6 (24%)	5 (24%)	5 (38%)
ICE	6 (24%)	4 (19%)	3 (23%)
BV-nivo	2 (8%)	2 (10%)	2 (15%)
BV	3 (12%)	3 (14%)	0
BV-chemo	3 (12%)	2 (10%)	1 (7.7%)
Other chemo	4 (16%)	4 (19%)	2 (15%)
Crizotinib	1 (4%)	1 (5%)	0
Time from diagnosis to ASCT, median (range), months	11.8 (8.2–140)	12.7 (8.2–140)	13.2 (8.6–140)
Received prior BV	16 (76%)	15 (83%)	9 (81%)
Refractory to prior BV ****	2 (13%)	2 (13%)	1 (11%)
Received prior checkpoint inhibitor	2 (10%)	2 (11%)	2 (18%)
Disease Status after Frontline Therapy
Refractory	12 (57%)	9 (50%)	5 (45%)
Relapse <12 months	5 (24%)	5 (28%)	4 (36%)
Relapse ≥12 months	4 (19%)	4 (22%)	2 (18%)
Extranodal involvement at pre-ASCT relapse	10 (48%)	10 (56%)	5 (45%)
Disease Status at ASCT
CR	17 (81%)	16 (89%)	9 (81%)
PR	3 (14%)	2 (11%)	2 (18%)
PD	1 (5%)	0	0
Time to ANC ≥500/mm3 median (range), days	11 (9–12)	11 (9–12)	11 (10–11)
Time from ASCT to CAR-T, median (range), days		21.5 (16–44)	21 (16–27)
CAR-T Dose
2×107 cells/m2		4 (22%)	3 (27%)
1×108 cells/m2		5 (28%)	3 (27%)
2×108 cells/m2		9 (50%	5 (45%)

^*Patient with grey zone lymphoma also received ABVD in addition to HL patients

^**Children’s Oncology Group ANHL12P1, ABVE-PC, DA-EPOCH-R, CHOP +/− rituximab

^***This is out of 25 for all collected patients, 21 for all treated patients and 13 for HL patients since 3 treated patients (and one collected but not treated patient) received an additional line of therapy.

^****Only includes patients who received prior BV

ABVD, doxorubicin, bleomycin, vinblastine, dacarbazine; ABVE-PC, doxorubicin, bleomycin, vincristine, etoposide, prednisone, cyclophosphamide; AITL, angioimmunoblastic T cell lymphoma; ANHL12P1, ANC, absolute neutrophil count; ASCT, autologous stem cell transplant; benda, bendamustine; BV, brentuximab vedotin; BV-CHP, brentuximab vedotin, cyclophosphamide, doxorubicin, prednisone; CHOEP, cyclophosphadmie, doxorubicin, vincristine, etoposide, prednisone; CHOP, cyclophosphamide, doxorubicin, vincristine, prednisone; DA-EPOCH-R – dose adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, rituximab; HL, Hodgkin lymphoma; ICE, ifosfamide, carboplatin, etoposide; PTCL, NOS, peripheral T cell lymphoma not otherwise specified

There were no DLTs observed after infusion of CD30.CAR-T cells (Table 2, 3), thus DL3 (2×10⁸ CAR-T cells/m²), the highest dose level tested and validated in our prior CAR-T cell studies⁶, was established as the MTD for this trial. There were no treatment related deaths. One patient treated on DL3 had grade 1 CRS on day 8, consisting of fever, which lasted until day 10 and did not require treatment with tocilizumab or steroids. No patients had grade 2 or higher CRS. No immune effector cell associated neurotoxicity syndrome was observed. Importantly, no patients experienced graft failure (appendix p. 1).There were low rates of grade 3 or 4 cytopenias after CD30.CAR-T cell infusion (Table 2). One patient had grade 3 anemia, grade 4 neutropenia, and grade 3 thrombocytopenia. Although expected post ASCT, the pancytopenia was likely due to disease progression. One patient developed a grade 3 rash, which was determined possibly related to CAR-T cells, but also possibly related to trimethoprim/sulfamethoxazole taken for prophylaxis for pneumocystis jiroveci pneumonia post-transplant and resolved with drug discontinuation and no other interventions. There was one other grade 1 rash which resolved without treatment. There were no infections reported in the first 30 days post treatment. Two patients developed secondary malignancies post treatment. One patient developed stage IV non-small cell lung cancer, and 1 patient developed testicular cancer approximately 2 and 2.5 years post CD30.CAR-T cell infusion, respectively. Malignancies were determined unrelated to CAR-T cell treatment and replication-competent retrovirus testing was negative in both patients.

Table 2:

Adverse Events of CD30.CAR-T cell infusion post ASCT.

Toxicity	Number Grade 1–2	Grade 1–2 (%)	Number Grade 3	Grade 3 (%)	Number Grade 4	Grade 4 (%)	Total Patients	Total (%)
Anemia	1	6%	1	6%			2	11%
Aspartate aminotransferase increased	4	22%					4	22%
Cytokine Release Syndrome	1	6%					1	6%
Diarrhea	3	17%					3	17%
Dizziness	2	11%					2	11%
Fatigue	3	17%					3	17%
Headache	2	11%					2	11%
Hypocalcemia	2	11%					2	11%
Lymphocyte count decreased	6	33%	2	11%			8	44%
Nausea	6	33%					6	33%
Neutrophil count decreased	3	17%			1	6%	4	22%
Platelet count decreased	4	22%	1	6%			5	28%
Rash maculo-papular	1	6%	1	6%			2	11%
Vomiting	2	11%					2	11%
White blood cell decreased	4	22%	2	11%			6	33%

Table includes all grade 3 or higher adverse events and all grade 1–2 events occurring in ≥10% of patients as well as adverse events of special interest (cytokine release syndrome, rash)

Table 3:

Grade 3 or Higher Adverse Events and Adverse Events of Special Interest after CAR-T Infusion by Dose Level

Adverse Event	All (N=18)	DL1 (N=4)	DL2 (N=5)	DL3 (N=9)
Lymphopenia	2 (11%)	1 (25%)	0	1 (11%)
Leukopenia	2 (11%)	0	1 (20%)	1 (11%)
Neutropenia	1 (6%)	0	1 (20%)	0
Anemia	1 (6%)	0	1 (20%)	0
Thrombocytopenia	1 (6%)	0	1 (20%)	0
Rash (grade 3)	1 (6%)	0	0	1 (11%)
Rash (any grade)	2 (11%)	0	0	2 (22%)
Cytokine release syndrome (Grade 1)	1 (6%)	0	0	1 (11%)

At a median follow up of 48 months (IQR: 27.5 to 61 months) post CD30.CAR-T infusion, 8 out of 18 patients (44%, 95% CI: 25% to 66%) have relapsed (5 with T-cell lymphoma, 3 with HL). Three out of 4 patients treated on DL1, 1 out of 5 treated on DL2, and 4 out of 9 treated on DL3 relapsed. One patient was in PR at the time of transplant, while the other 7 were in CR. As of the last assessment, the remaining 10 patients continue to be in CR. For the total number of patients (n=21) which included 3 patients that were not treated, the median PFS was not reached (Figure 2A). PFS at 2 years was 57% (95%CI: 40% to 83%). For all the treated patients (n=18) the median PFS was 32 months (95%CI: 4.6 months to not estimable) and the 2 year PFS was 56% (95%CI: 37% to 84%) (Figure 2B). Among the 13 patients with HL, which included 2 patients that were not treated, median PFS was not reached. The 2 year PFS for these patients was 69% (95%CI: 48% to 99%). For the 11 treated HL patients, the median PFS was also not reached (Figure 2C). The 2 year PFS for these patients was 73% (95%CI: 51% to 99%). The median OS for all treated patients has not been reached (Figure 2D). The 2-year OS for all patients is 78% (95% CI: 61% to 99%). All patients in the HL cohort are still alive, while 5 patients with T cell lymphoma have died (4 due to relapse and 1 due to an unrelated lung cancer). Out of the 3 patients with HL who relapsed, one patient went on to receive nivolumab and is now off therapy in CR, one patient was salvaged with BV and bendamustine and is now in CR post allogeneic stem cell transplant, and one patient received multiple therapies including nivolumab and BV and subsequently received bendamustine prior to allogeneic stem cell transplant.

Figure 2.

Activity of CD30.CAR-T cells post ASCT. (A) Kaplan Meier Progression free survival plot for all 21 patients collected. (B) Kaplan Meier Progression free survival plot for the 18 treated patients. (C) Kaplan Meier Progression free survival plot for the 11 treated HL patients. (D) Overall survival plot for 18 treated patients.

Twenty-one individuals were collected, and products were successfully manufactured for 19 patients. The characteristics of the 18 CD30.CAR-T cell products infused are summarized in appendix, page 2. Production of CD30.CAR-T cells required a median of 16 days (range 9 to 24 days). CD30.CAR expression ranged from 86.5% to 99.8%, with no differences between products generated from patients with HL or T cell lymphomas (appendix p. 2). CD30.CAR-T cell products contained a variable percent of CD4⁺ and CD8⁺ T cells, with 5 products composed of >70% of CD4⁺ T cells. The majority of CD30.CAR-T cells showed an effector phenotype¹⁴ (CCR7^–CD45RA^–, 71.5% ± 14.6%), with lower fractions of cells expressing CCR7 (CD45RA⁺CCR7⁺ = 2.1% ± 1.9% and CD45RA^–CCR7⁺ = 19.2% ± 12.24%) (appendix p. 2). No correlation was observed between the time required for the generation of CD30.CAR-T cells and phenotypic characteristics. All CD30.CAR-T cell products demonstrated excellent potency against CD30⁺ targets in vitro, as assessed by standard ⁵¹Cr release (62.5% ± 5% killing at 20:1 E:T ratio), with negligible cytotoxic activity against CD30^– targets (Supplementary Figure 2C).

CD30.CAR-T cells expanded in the peripheral blood peaking between week 1 and 2 post infusion (appendix p. 3) and remained detectable for 6 weeks with >15 copies/μg of DNA in 11 patients (61%). Only one of these 11 patients relapsed within 6 months of infusion. CD30.CAR-T cells were still detectable in 7 of the 12 subjects still in remission 6 months post infusion (58%), and in 3 of the 10 patients (with available “material”) still in remission 1 year post infusion (33%). No patients had detectable molecular signal 2 years post infusion, even if they remained in remission. CD30.CAR-T cell expansion and persistence, measured as area under the curve, did not differ by dose level and was overall comparable to that seen in our previous trial in patients treated with CD30.CAR-T cells post lymphodepletion with fludarabine/bendamustine^6,10. IL-15 and IL-7 cytokines measured in the plasma at the time of CD30.CAR-T cell infusion were 10.3 ± 2.5 pg/ml and 9.5 ± 1.2 pg/ml, respectively. These plasma levels of IL-15 and IL-7 were lower compared to those measured in patients infused with CD30.CAR-T cells after lymphodepletion with fludarabine/bendamustine⁶ (appendix p. 3B). Although patients had low absolute lymphocytes counts (ALC, 0.87 ± 0.5 ×10³/uL) at the time of infusion of the CD30.CAR-T cells (appendix p. 3), ALC were higher than those observed in patients infused with CD30.CAR-T cells immediately after fludarabine/bendamustine lymphodepletion.⁶ We also measured the plasma levels of CCL17, as surrogate markers of tumor response. When compared to patients with active disease,⁶ CCL17 plasma levels were low post ASCT, at the time of the CD30.CAR-T cell infusion, and remained low 6 weeks after therapy (appendix p. 3). Six patients were biopsied at time of relapse (1 patient with HL and 5 patients with T cell lymphoma) with 5 patients still having CD30 positive disease at relapse, while 1 patient with T cell lymphoma relapsed with CD30 negative disease.

DISCUSSION

Here we present the long-term outcome of patients with CD30⁺ lymphomas at high risk of relapse after ASCT who received CD30.CAR-T cell infusion as consolidation therapy in a phase 1 dose escalation study. Our findings indicate that the infusion of CD30.CAR-T cells following high dose chemotherapy with BEAM and ASCT is safe since CD30.CAR-T cells did not compromise long-term stem cell engraftment or lead to prolonged cytopenias, as seen with other CAR-T cell targets¹⁵, and did not cause CRS or neurological toxicities, in line with previously reported data with CD30.CAR-T cells where only low grade CRS and no neurotoxicity was reported^6,7. Furthermore, promising activity was observed, particularly in patients with HL, who had a 2-year PFS of 73% and 2 year OS of 100%. Two patients treated on this clinical trial developed secondary malignancies, both of solid tumor origin. This patient population is at baseline increased risk of secondary malignancy^16,17. Replication-competent retrovirus testing was negative in these patients, supporting that the secondary malignancies were not caused by the CAR-T cells; however, we cannot completely rule out CAR-T cells as a potential contributor of risk.

Prior to our initiating this protocol, Moskowitz et al. in the AETHERA study evaluated the use of BV as maintenance therapy after ASCT in patients at high risk for relapse⁴. For the current trial, we utilized the same definition of high-risk patients as the inclusion criteria for that study. While it is difficult to compare treatments between different clinical trials, outcomes in the current study appear comparable to those of the AETHERA trial, where the patients receiving maintenance BV post ASCT had a 2-year PFS of 63%⁴ However, in the current clinical scenario many patients receive BV either as frontline⁵ or as part of second line salvage therapy prior to ASCT¹⁸; thus in our study, 83% of patients had received prior BV, including 81% of patients with HL. As observed in our previous clinical study,⁶ prior treatment with BV did not seem to affect the therapeutic activity of CD30.CAR-T cells infused after ASCT, although some patients’ biopsies used to determine CD30 expression were obtained prior to last dose of BV. This may be because CD30 expression is generally preserved in HL and loss of CD30 antigen does not appear to be a major mechanism of resistance to BV¹⁹, although there have been some recent studies suggesting downregulation of CD30 in patients with T-cell lymphoma after treatment with BV²⁰. Further study of the impact of prior CD30 targeted therapies and level of CD30 expression on response to CD30.CAR-T cells is needed.

Maintenance BV after ASCT in HL patients can be challenging to complete for many patients. Approximately one third of patients in the AETHERA trial discontinued BV prior to completion of therapy due to cumulative toxicity, particularly peripheral neuropathy, and discontinuation rates are even higher in the real world setting.²¹ A single infusion of CD30.CAR-T cells after ASCT appears to compare favorably, with respect to toxicity, to multiple cycles of BV and may greatly enhance patient compliance, which is of particular importance in the adolescent and young adult population. The results of our study also compare well with the results of a clinical trial investigating maintenance therapy with pembrolizumab after ASCT in patients with HL, which found a 19 month PFS of 81%.²² However, pembrolizumab administration was associated with high rates of toxicity, with 40% of patients having at least one grade 2 or higher immune related adverse event. How CD30.CAR-T cell maintenance compares to BV or checkpoint inhibitors would require a randomized clinical trial comparing these treatments after ASCT in patients with HL.

The treatment landscape for HL is rapidly changing with both BV and checkpoint inhibitors now being moved to earlier lines of therapy with promising results^5,23. Outcomes in cHL patients post ASCT in patients previously exposed to PD-1 inhibitors were improved compared to PD-1 naïve patients in a retrospective analysis with a post ASCT 2 year PFS of 93%²⁴. In another multi-institutional retrospective study, there was no PFS benefit seen with BV maintenance in patients previously exposed to BV or PD-1 inhibitors²⁵. Due to the time period of patient enrollment, no treated patients received PD-1 inhibitors prior to ASCT and CAR-T cell infusion so it is not possible to make conclusions on the role of CD30.CAR-T cell maintenance in this population. However, in this setting, CAR-T cells may provide patients a therapy they have not previously been exposed to.

Our clinical trial also enrolled patients with other CD30⁺ lymphomas. Although numbers are small, the results for patients with T-cell lymphoma in our trial were modest as 5 of 6 patients experienced disease relapse. This may reflect the fact that T-cell lymphomas are less responsive to high dose chemotherapy with historically worse outcomes after ASCT compared to HL.²⁶ In addition, this study included a higher risk population of T-cell lymphomas who were transplanted in second or later remission. These patients generally experience worse outcomes, with a retrospective study finding a 5 year PFS of 12% for patients transplanted in second remission.²⁷ All of the patients with T-cell lymphoma enrolled on this trial were considered more suitable candidates for ASCT over allogeneic stem cell transplant according to their treating physician. In high risk T-cell lymphoma patients treated with ASCT, CD30.CAR-T cell maintenance did not appear to overcome the poor prognosis. It is possible that there could be greater clinical benefit in patients with T-cell lymphoma transplanted in first remission, who were excluded from this trial.

In the current study, we monitored the expansion and persistence of CD30.CAR-T cells in vivo. Chemotherapy-based lymphodepletion and, in particular, fludarabine plays a critical role in promoting CAR-T cell expansion after infusion.⁶ CAR-T cells are generally infused within a few days after lymphodepletion at the peak levels of the homeostatic cytokines IL-7 and IL-15 that contribute to the rapid expansion of CAR-T cells.⁶ The BEAM regimen does not include fludarabine, and CD30.CAR-T cells were infused more than 3 weeks after chemotherapy to allow for hematopoietic reconstitution. Both these factors likely contribute to the lower levels of IL-7 and IL-15 detectable in the plasma at the time of CD30.CAR-T cell infusion. However, since patients remained relatively lymphopenic after BEAM, we still observed significant expansion of CD30.CAR-T cells within the first 3 weeks after infusion, similar to those observed in patients receiving CD30.CAR-T cells after bendamustine and fludarabine.⁶ In a recent study in which CD30.CAR-T cells were administered in tandem with ASCT in 5 patients with HL and 1 patient with ALCL a similar expansion to CD30.CAR-T cells in vivo was observed.²⁸ However, studies of CD19 CAR-T cells given two days after ASCT in patients with B-cell NHL showed lower persistence compared to CAR-T cells after lymphodepletion^29,30. It is unclear whether this is related to the timing of treatment after ASCT or CAR-T or disease intrinsic factors.

Study limitations include small sample size and heterogeneous patient population. In addition, there were three patients not included in the analysis because they never received treatment either due to inability to manufacture CAR-T cells or change in treatment plan. However, the majority of patients enrolled were able to receive CAR-T cells, highlighting the feasibility of this approach. In addition, there were unfortunately limited biopsy samples pre and post treatment to assess CD30 expression and other features.

In summary, we demonstrated the feasibility and safety of CD30.CAR-T cells as consolidation after BEAM ASCT in CD30⁺ lymphoma, with a preliminary indication of activity in HL patients at high risk of relapse. Future studies should evaluate the efficacy of this approach compared to maintenance therapy with BV or checkpoint inhibitors.

Research in context.

Evidence before this study

We searched PubMed for clinical trials published from database inception to March 21, 2016, using the terms (“chimeric antigen receptor” or “CAR-T”) and “CD30” and did not find any published clinical trials of CD30 CAR-T cells at that time. We also searched PubMed for clinical trials published from database inception to March 21, 2016, investigating consolidation post autologous stem cell transplant for Hodgkin lymphoma and T-cell lymphoma, using the terms (“Hodgkin” or “Hodgkin’s” or “T-cell lymphoma”) and “consolidation” and “autologous stem cell transplantation.” Most of the 191 studies found on the search were related to transplant as consolidation or review papers. There was a small single arm study investigating post transplant radiation and chemotherapy in Hodgkin lymphoma but no further follow up studies of this approach. There was also a randomized study across lymphoma subtypes which found that interferon α−2b post autologous stem cell transplant did not improve outcomes. More recently, a randomized control study found a progression free survival benefit with the addition of brentuximab vedotin consolidation post autologous stem cell transplant in patients with Hodgkin lymphoma.

Added value of this study

We found that autologous CD30 CAR-T cells can be safely infused as consolidation therapy after autologous stem cell transplantation with promising activity in patients with Hodgkin lymphoma at high risk of relapse. This study introduces another potential indication for CD30 CAR-T cells for further investigation.

Implications of all the available evidence

Since this study was written and completed, there have also been some clinical trials demonstrating promising activity of checkpoint inhibitors as consolidation post autologous stem cell transplant in Hodgkin lymphoma. In addition, the treatment paradigm for Hodgkin lymphoma is rapidly changing with brentuximab vedotin and checkpoint inhibitors being moved to earlier lines of therapy, with a majority of patients being exposed to both treatments prior to autologous stem cell transplantation. The promising results of this study raise interest for an alternative consolidation approach that does not include agents that patients may have previously been exposed to.

Data Sharing Statement

Requests for data sharing may be submitted to Barbara Savoldo (bsavoldo@med.unc.edu).