Superior Therapeutic Index in Lymphoma Therapy: CD30(+) CD34(+) Hematopoietic Stem Cells Resist a Chimeric Antigen Receptor T-cell Attack

Abstract

Recent clinical trials with chimeric antigen receptor (CAR) redirected T cells targeting CD19 revealed particular efficacy in the treatment of leukemia/lymphoma, however, were accompanied by a lasting depletion of healthy B cells. We here explored CD30 as an alternative target, which is validated in lymphoma therapy and expressed by a broad variety of Hodgkin's and non-Hodgkin's lymphomas. As a safty concern, however, CD30 is also expressed by lymphocytes and hematopoietic stem and progenitor cells (HSPCs) during activation. We revealed that HRS3scFv-derived CAR T cells are superior since they were not blocked by soluble CD30 and did not attack CD30(+) HSPCs while eliminating CD30(+) lymphoma cells. Consequently, normal hemato- and lymphopoiesis was not affected in the long-term in the humanized mouse; the number of blood B and T cells remained unchanged. We provide evidence that the CD30(+) HSPCs are protected against a CAR T-cell attack by substantially lower CD30 levels than lymphoma cells and higher levels of the granzyme B inactivating SP6/PI9 serine protease, which furthermore increased upon activation. Taken together, adoptive cell therapy with anti-CD30 CAR T cells displays a superior therapeutic index in the treatment of CD30(+) malignancies leaving healthy activated lymphocytes and HSPCs unaffected.

Figure 1

Anti-CD30 chimeric antigen receptor (CAR) T cells mediate a specific response against CD30⁺ lymphoma cells in vitro and in vivo. Anti-CD30 CAR T cells lyse CD30⁺ lymphoma cells of different origin in a dose-dependent manner. (a) T cells from the peripheral blood were engineered with the anti-CD30 CAR and cocultivated in serial dilutions (0.5–5 × 10⁴/well) with CD30⁺ MyLa cutaneous T lymphoma cells or CD30^– Colo320 tumor cells (5 × 10⁴/well) for 48 hours in 96-well round bottom plates. Coincubation of lymphoma cells with T cells without CAR (w/o CAR) served as control. The assay was done in triplicates and mean values ± standard deviation (SD) were determined. Data represent mean values of one representative out of three experiments. (b) CAR T cells (2.5 × 10⁴/well) were cocultivated with CD30⁺ MyLa, L1236 Hodgkin's lymphoma cells, and CD30^– Colo320 tumor cells (each 5 × 10⁴/well), respectively. T cells without CAR (w/o) served as control. Specific lysis of tumor cells was determined by an XTT-based assay. The assay was done in triplicates and mean values (±SD) were determined. Data represent mean values of one representative out of five experiments. P values were calculated by Student's T-test. Asterisks indicate significant differences between groups (P < 0.05). (c) Anti-CD30 CAR T cells suppress tumor growth in vivo in a dose-dependent fashion. CAR T cells (2.5 × 10⁵–1 × 10⁷ cells/mouse) were coinjected with MyLa cutaneous T lymphoma cells (10⁷/mouse, groups of five or six animals) and tumor growth was monitored every 2–3 days. Mean values of the tumor volume (left) and the area under curve values (right) (±SD) were determined. P values were calculated by Student's T-test. Asterisks indicate significant differences (P < 0.05) compared with the control group. (d) Blocking of CAR T cells by soluble CD30-Fc. MD45 T cells with anti-CD30 CAR were cocultivated with CD30⁺ L540cy cells (each 5 × 10⁴/well) in presence of increasing amounts of CD30-Fc, the HRS3-specific anti-idiotypic mAb 9G10 and an isotype control mAb, respectively. After 48 hours, culture supernatnants were removed, analyzed for IL-2 secretion and blocking of CAR T-cell inhibition was determined as described in Materials and Methods. The assay was done in triplicate samples and the mean values ± SD were determined.

Figure 2

CD34⁺ hematopoietic stem and progenitor cells (HPSCs) express CD30 upon activation and differentiation. (a) CD30 surface expression on freshly isolated (day 0) and cultured CD34⁺ cells. (b) Quantification of CD30 on the cell surface over time based on flow-cytometric data (as shown in a) of umbilical cord blood derived CD34⁺ cells from three donors, presented as relative values of all cells or of CD34⁺ cells (± SD). (c) Human hematopoietic tree and definition of HSPC subsets. (d) CD30 expression on HSPC subsets. We used the following marker combinations to identify fractions enriched for HSCs/MPPs (CD34⁺CD133⁺CD45RA^–CD38^lowCD10^–), lymphoid-primed multipotents (LMPPs) (CD34⁺CD133⁺CD45RA⁺ CD38^lowCD10^–), GMPs (CD34⁺CD133⁺ CD45RA⁺CD38⁺CD10^–); multi-lymphoid primed progenitors (CD34⁺CD133⁺CD45RA⁺CD38^lowCD10⁺) and erythro-myeloid primed progenitors (CD34⁺CD133^lowCD45RA^–). Since the CD38 expression after initiation of culture is not indicative to discriminate LMPP and GMP-enriched fractions, we neglected its expression for HSPC subset gating on cultured samples and pooled LMPP and GMP fractions to an LMPP fraction.^, A dump channel excluding 7-AAD-positive dead cells and lineage (CD3, CD19, CD56, CD66b) positive cells and a CD45^dim gate were used for all experiments (see Supplementary Figure S1 for details). The flow-cytometric data show one representative out of three independent analyses; cells from three donors were analyzed.

Figure 3

Anti-CD30 chimeric antigen receptor (CAR) T cells do not lyse activated hematopoietic stem and progenitor cells (HSPCs). (a) Flow cytometric cell sorting of CAR⁺ and CAR^– T cells. T cells were engineered with the anti-CD30 CAR, stained with anti-hIgG1 and anti-CD3 antibodies to detect CAR⁺ T cells and flow sorted to obtain the CD3⁺ CAR^– and CD3⁺ CAR⁺ T-cell subsets. (b,c) Cocultivation of CAR T cells with HPSCs and CD30⁺ lymphoma cells. HSPCs from two healthy donors were isolated by magnetic cell separation procedures from umbilical cord blood and activated in the presence of FLT3L, SCF, and TPO for 72 hours. Carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled CD30⁺ MyLa lymphoma cells (2 × 10⁴ /well) and activated HSPCs 1 × 10⁴ (donor UCB612) and 2 × 10⁴ cells/well (donor UCB615), were cocultivated for 24 hours each with sorted CAR^– and CAR⁺ T cells (2 × 10⁴/well), respectively, in 96-well round bottom plates. Cells were recovered and HSPCs labeled with an anti-CD34-PE antibody. Cells were analyzed by flow cytometry and dead cells were identified by 7-AAD staining (0.5 µg/ml). Representative dot blots demonstrating the gating strategy of CD34⁺ HSPCs and CFSE-labeled MyLa lymphoma cells and histograms of 7-AAD staining indicating target-cell lysis upon cocultivation with CAR T cells are shown (b). 7-AAD staining data were used to calculate specific lysis as described in Materials and Methods (c). The assay was done in triplicates and the mean values ± SD were determined. Data show representative results of three analyses with different donors. P values were calculated by Student's T-test and significant differences between relevant groups are indicated.

Figure 4

Anti-CD30 chimeric antigen receptor (CAR) T cells preferentially lyse CD30⁺ lymphoma cells but not purified CD30⁺ hematopoietic stem and progenitor cell (HSPC). (a) CD34⁺ HSPCs were isolated by magnetic cell separation from umbilical cord blood of healthy donors, activated for 3 days, stained with anti-CD34-PE and anti-CD30-FITC antibodies and sorted by flow-cytometry into CD30⁺ and CD30^– subsets. (b) Anti-CD30 CAR T cells were flow sorted into CAR⁺ and CAR^– subsets and coincubated (2 × 10⁴ cells/well) with carboxyfluorescein diacetate succinimidyl ester-labeled MyLa cells (2 × 10⁴/well) or with flow-sorted CD30⁺ and CD30^– HSPCs (each 1 × 10⁴ cells/well) for 24 hours. Cells were recovered and stained with anti-CD34 antibody and 7-AAD. The numbers of viable and dead cells were determined and specific lysis was calculated as described in Materials and Methods. The assay was done in triplicate samples and mean values ± SD of a representative analysis out of three are demonstrated. P values were calculated by Student's T-test and significant differences between relevant groups are indicated.

Figure 5

Anti-CD30 chimeric antigen receptor (CAR) T cells do not affect myeloid and lymphocyte cell differentiation. (a) To assay colony-forming capacity, 200 CD34⁺ hematopoietic stem and progenitor cells (HSPCs), which had been stimulated for 2 days with FLT3L, SCF, and TPO to induce CD30, were cultured in the presence or absence of CAR⁺ or CAR^– T cells (200/well) in 96-well round-bottom plates. After 24 hours, cells were harvested and seeded in MethoCult H4434 medium to induce erythro-myeloid cell differentiation. Hematopoietic colonies were scored after 12–14 days as described in Materials and Methods. Data of three HSPC donors are shown. (b) Newborn Rag2^–/– cγ^–/– mice were grafted with human CD34⁺ HSPCs (1–3 × 10⁵/mouse). After successful engraftment of human hematopoiesis autologous T cells with or without anti-CD30 or anti-carcinoembryonic antigen (CEA) CAR were transferred into mice at day 0 (2.5 × 10⁶ cells/mouse; six mice with anti-CD30 CAR T cells; three mice with anti-CEA CAR T cells). Blood was recovered once per week and B and T cells were identified by flow cytometry upon staining with anti-CD3 and anti-CD19 antibodies, respectively. Data represent mean values ± SD. P values were calculated by the Student's T-test. Asterisks indicate significant differences between groups (P < 0.05).

Figure 6

CD30⁺ hematopoietic stem and progenitor cells (HSPCs) and B cells express CD30 below the activation threshold of anti-CD30 chimeric antigen receptor (CAR) T cells. (a–c) Comparison of CD30 levels on target cells. (a) MyLa lymphoma cells were stained with the anti-CD30 antibody or an isotype control antibody and recorded by flow cytometry. (b) HSPCs from four donors were activated and flow sorted into the CD30^– and CD30⁺ subpopulations utilizing the same anti-CD30 antibody; the mean fluorescence intensity (mfi) was recorded. (c) Mouse MC38 tumor cells which lack CD30 were engineered to express human CD30. Cells were stained with the anti-CD30 antibody and flow sorted into subsets with different CD30 levels. Numbers in histograms and above columns represent mfi of CD30 staining in arbitrary units. (d,e) Sorted MC38 tumor cells (2.5 × 10⁴/well) were cocultivated for 24 hours with T cells with or without anti-CD30 CAR (2.5 × 10⁴/well). Target-cell lysis was determined by the XTT-assay (d) and IFN-γ concentration in the supernatants by ELISA (e). The assay was done in triplicate, mean values ± SD determined and plotted against the mfi of CD30 of the respective target-cell subset. P values were determined by the Student's T-test and significant differences (P < 0.05) indicated by asterisks. The assay was performed twice with similar results. (f) B cells were activated and flow sorted into CD30^– and CD30⁺ subpopulations utilizing the anti-CD30 mAb as described in Materials and Methods. (g) Unseparated and CD30⁺ or CD30^– B cells (each 1 × 10⁴/well) were coincubated with CAR⁺ or CAR^– T cells (2.5 × 10⁴/well) for 48 hours. B cells were identified by staining with the anti-CD19 mAb, dead cells by staining with 7-AAD (0.5 µg/ml) and the number of dead and live B cells was recorded by flow cytometry. Specific B-cell lysis was calculated as described. For comparison, cytolysis of MyLa cells by CAR T cells was tested in parallel. The assay was done in triplicates and mean values ± SD were determined. P values of relevant groups were determined by the Student's T-test. Data of one representative analysis out of two are shown.

Figure 7

Hematopoietic stem and progenitor cells (HSPCs) did not induce CD30-specific degranulation of chimeric antigen receptor (CAR) T cells, are protected against bystander lysis and express high levels of the granzyme B inhibitor PI-9. (a) CAR T cells specifically degranulate in the presence of CD30⁺ lymphoma cells but not CD30⁺ HSPC cells. Anti-CD30 CAR T cells (4 × 10⁴ cells/well) were coincubated with CD30⁺ MyLa lymphoma cells and flow-sorted CD30⁺ or CD30^– HSPCs (each 2 × 10⁴ cells/well), respectively. Degranulating T cells were identified by staining for CD107a. After 12 hours, T cells were recovered, stained additionally with anti-CD3 and anti-human IgG1 antibodies to detect CD3⁺ CAR⁺ and CD3⁺ CAR^– T cells, respectively, and analyzed by flow cytometry. Mean values ± SD of triplicates were determined and p values calculated by the Student's T-test. (b) HSPC bystander lysis by CAR T cells in the presence of CD30⁺ MyLa lymphoma cells. Carboxyfluorescein diacetate succinimidyl ester-(CFSE) labeled MyLa cells (2 × 10⁴ cells/well) and CD30⁺ HSPCs (1 × 10⁴ cells/well) were together cocultivated with T cells with or without CAR. Cells were harvested, labeled with an anti-CD34 antibody and 7-AAD and the number of viable and dead CD34⁺ HSPCs was determined; specific target-cell lysis was calculated as described in Materials and Methods. Data represent mean ± SD of triplicate samples. Statistic analysis was performed by using the Student's T-test. Data of a representative analysis out of two are shown. (c,d) HSPCs express high levels of the granzyme B inactivating PI-9 protease. HSPCs from umbilical cord blood of healthy donors (n = 6) were activated for 3 days, stained for PI-9 or with an isotype control antibody (iso) and CD30, respectively. For comparison, peripheral blood lymphocytes of healthy donors (n = 4) and cells from lymphoma cell lines (n = 3) were stained and analyzed by flow cytometry. (c) Comparison of PI-9 expression between HSPC, peripheral blood lymphocytes and lymphoma cell lines, respectively. (d) Comparison of PI-9 expression between CD30⁺ and CD30^– HSPC. Data represent mean values ± SD where applicable. (e) HSPCs upregulate PI-9 upon cytokine-mediated activation. Isolated HSPCs (n = 3) were cultivated either without activation or activated for 3 days as described above. Cells were recovered and stained for PI9 and analyzed by flow cytometry. Data represent mean values ± SD. Statistical analysis was performed by using the Student's T-test and significant differences were indicated.