Background

Chimeric antigen receptor (CAR) T cells showed remarkable efficacy for the treatment of B-cell malignancies and have been approved by the US Food and Drug Administration for the treatment of relapsed/refractory B-cell acute lymphoblastic leukemia (B-ALL) and diffuse large B-cell lymphoma (DLBCL) [1, 2]. However, the development of CAR-T cells against T-cell malignancies seems more challenging due to the similarities between the normal, malignant and therapeutic T cells, which could result into CAR-T cell fratricide, T cell aplasia, and contamination of CAR-T cell products with malignant T cells [3, 4].

An innovative treatment option for T-cell malignancy was proposed that targeting T cell receptor β-chain constant region 1 (TRBC1) CAR-T could specifically identify and kill TRBC1+ T-cell malignancies, since either TRBC1 or TRBC2 is mutually exclusively expressed in T cells and moreover proportion of TRBC1+ T cells varies between 25 and 47% in healthy individuals, but malignant T cells are clonally TRBC1 positive or negative [5, 6]. Thus, anti-TRBC1 CAR-T cells could specifically kill TRBC1+ malignant T cells while sparing TRBC2+ normal T cells. However, anti-TRBC1 CAR gene could probably be inadvertently transferred into TRBC1+ malignant T cells during CAR-T cell manufacturing, and its product could in cis bind to autologous TRBC1 on the surface of malignant T cells, which could result into masking TRBC1 from identification by and mediating resistance to anti-TRBC1 CAR-T and meanwhile weaken effector function of anti-TRBC1 CAR transduced TRBC1+ cells. Following transduction of T cells with lentivirus encoding anti-TRBC1 CAR, all T cells could be categorized into TRBC1+ cells (C1), TRBC2+ cells (C2), anti-TRBC1 CAR transduced C1 cells (CAR-C1) and anti-TRBC1 CAR transduced C2 cells (CAR-C2) (Fig. 1a). Thus, it is interesting to evaluate whether both C1 and CAR-C1 could be identified and killed by CAR-C1 and CAR-C2 (Fig. 1a).

Fig. 1
figure 1

Effector functions of TRBC1+ and TRBC2+ cells genetically engineered with anti-TRBC1 CAR. a The categories and relationship of T cells following transduction with anti-TRBC1 CAR. TRBC1+ cells, C1; TRBC2+ cells, C2; anti-TRBC1 CAR transduced TRBC1+ cells, CAR-C1; anti-TRBC1 CAR transduced TRBC2+ cells, CAR-C2. b TRBC1 expression and CAR transduction efficacy of TRBC1-sorted and TRBC1-depleted T cells as well as CAR and TRBC1 expression of CAR-C1 and CAR-C2 analyzed by flow cytometry. c IFN-γ secretion by CAR-C1 and CAR-C2 against C1, CAR-C1 or C2 after 24-h co-culture. d-e Left, representive FACS profile of CD137 and C107a expression on CAR-C1 and CAR-C2 co-cultured with C1, CAR-C1 or C2. Right, percentages of CD137- and C107a-positive CAR-C1 and CAR-C2 following co-culture with C1, CAR-C1 or C2. f Cytotoxic activities of CAR-C1 and CAR-C2 against C1, CAR-C1 or C2 were examined by standard CFSE-based cytotoxity assays at several effector/target (E/T) ratios. g Scheme of the xenograft model. NOG mice (n = 5/group) were IV injected with 3 × 106 Luc/GFP–expressing Jurkat cells followed 3 days after by a single IV injection of 5 × 105 MOCK, CAR-C1 or CAR-C2. h IVIS imaging of tumor burden monitored by BLI at the indicated time points following MOCK, CAR-C1 or CAR-C2 T cell injection (day 0). i Radiance of individual mice at day 20 following MOCK, CAR-C1 or CAR-C2 T cell injection. n = 5 mice per group. j Kaplan-Meier survival curve of mice injected with mock, CAR-C1 or CAR-C2 T cells. ***P < 0.001 and n.s., not significant

Full size image

Results and discussions

To evaluate whether C1 and CAR-C1 could be identified and killed by CAR-C1 and CAR-C2, we first sorted donor T cells into TRBC1+ and TRBC1− (designated as C2) fractions using magnetic beads. A portion of C1 or C2 were used as target cells and other C1 and C2 from the same donor were genetically engineered with anti-TRBC1 CAR to obtain CAR-C1 and CAR-C2 as effect cells. We confirmed that transduction efficacy of anti-TRBC1 CAR was similar on C1 and C2, and moreover TRBC1 was not detected on CAR-C1 through flow cytometry (Fig. 1b). Since primed T cells could increase CD137 expression and IFN-γ secretion, and moreover cytotoxic T cells could express CD107 and mediated killing of target cells, these markers could be used to detect activation and cytolytic activity of T cells. We found that CAR-C2 than CAR-C1 showed higher level of IFN-γ production and CD137 expression when co-cultured with C1 but not CAR-C1 or C2 (Fig. 1c and d). In flow cytometry–based cytotoxicity assays, CAR-C2 and CAR-C1 both specifically killed C1 but not CAR-C1 or C2, more so in CAR-C2 than CAR-C1 (Fig. 1e and f).

We next evaluated the anti-tumour activity of CAR-C1 and CAR-C2 in vivo using Luc-expressing Jurkat T-ALL cells. NOG mice were transplanted with 3 × 106 Luc-expressing Jurkat cells 3 days before IV infusion of 5 × 105 CAR-C1, CAR-C2 or MOCK T cells (Fig. 1g). Consistent with the in vitro observation, CAR-C1 induced transient tumour regression, but tumours re-progressed rapidly. In contrast, mice treated with an equal number of CAR-C2 exhibited significantly higher ani-tumour ability with significantly prolonged survival (P < 0.001) (Fig. 1h-j).

To investigate why CAR-C1 than CAR-C2 demonstrated lesser killing ability against C1 and moreover neither of them could identify and kill CAR-C1, we hypothesize that since expression abundance of anti-TRBC1 CAR is significantly higher than TRBC1 on CAR-C1, a proportion of CARs in cis bind to autologous TRBC1 on CAR-C1, masking TRBC1 from identification by other anti-TRBC1 CAR-T, and meanwhile only the remaining unoccupied CARs identify C1, weakening effector function of CAR-C1 (Fig. 2a).

Fig. 2
figure 2

The cause for undetected TRBC1 and lesser effector function of CAR-C1. a Due to higher expression level of CAR than TRBC1 on CAR-C1, some CARs in cis bind to autologous TRBC1 on CAR-C1, resulting into masking TRBC1 from identification by other anti-TRBC1 CAR-T and meanwhile occupying these CARs, and thus only the remaining unoccupied CARs target TRBC1. b TRBC1 mRNA expression is maintained in CAR-C1 as compared to C1, as determined by qRT-PCR (ΔΔ Ct normalized to C1). c TRBC1 on C1 is detectable using both mAb 8A3 targeting TCRβ-chain constant region and mAb JOVI-1 from which the anti-TRBC1 CAR was derived, but TRBC1 on CAR-C1 cells is only recognized by mAb 8A3. d Left, representive FACS profile of TRBC1 expression on CAR-C1 and C1. Right, MFI of TRBC1 on CAR-C1 and C1. e Expression level of CAR was significantly higher than TRBC1 on CAR-C1 determined by qRT-PCR analysis (ΔΔ Ct normalized to TRBC1). f Confocal imaging of CAR-C1 using FITC-conjugated anti-TRBC1 antibody (green), TRITIK-conjugated anti-FLAG antibody (red), and DAPI (blue). Scale bars, 5 μm. g Left, representive FACS profile of CD45RA and CCR7 expression on CAR-C2 after 6-day culture alone or co-culture with C1. Right, percentages of naïve (CD45RA+ CCR7+), effector (CD45RA+ CCR7−), effector memory (CD45RA− CCR7−) and central memory (CD45RA− CCR7+) CAR-C2 cells. h-j Left, representive FACS profile of PD-1 (h), TIM-3 (i) and LAG-3 (j) expression on CAR-C2 after 6-day culture alone or co-culture with C1. Right, percentage of PD-1 (h), TIM-3 (i) and LAG-3 (j) positive CAR-C2. *P < 0.05, ***P < 0.001. Data are representative of three independent experiments

Full size image

We first found that TRBC1 mRNA expression was preserved in CAR-C1 as compared to C1 determined by qRT-PCR analysis (Fig. 2b). We further confirmed via flow cytometry that TRBC1 on CAR-C1 was detectable by anti-TRBC monoclonal antibody (mAb) 8A3 targeting not the same epitope recognized by mAb JOVI-1 from which the anti-TRBC1 CAR was derived (Fig. 2c), and moreover expression level of TRBC1 protein was similar on CAR-C1 and C1 (Fig. 2d). Meanwhile, qRT-PCR analysis demonstrated that expression level of CAR was significantly higher than TRBC1 in CAR-C1 and moreover confocal microscopy further confirmed that colocalization of anti-TRBC1 CAR and TRBC1 on the cell surface of CAR-C1 (Fig. 2e and f). These findings supported that TRBC1 molecules were still expressed on the surface of CAR-C1 but in cis bound by a proportion of anti-TRBC1 CARs, masking TRBC1 from identification by other anti-TRBC1 CAR-T, and meanwhile only the remaining unoccupied CARs identified C1, weakening effector function of CAR-C1.

In addition, contaminating TRBC1+ malignant cells during anti-TRBC1 CAR-T manufacturing not only produced CAR-C1 which was resistant to anti-TRBC1 CAR-T and had lesser killing ability, but were expected to accelerate exhaustion and terminal differentiation of anti-TRBC1 CAR-T with limited in vivo persistence due to continuous (tonic) ligand-driven CAR stimulation [7, 8]. Co-culture of CAR-C2 with C1 in a 2:1 ratio (physiological condition) for 6 days revealed lower and higher percent of naïve and effect CAR-C2 cells, respectively, compared to solo culture of CAR-C2 (Fig. 2g). In addition, the co-culture of CAR-C2 and C1 exhibited increasing expression of PD-1, TIM-3 and LAG-3 in CAR-C2 (Fig. 2h-j). These findings suggested that compared with unfractionated T cells, TRBC1-depleted T cells genetically engineered with anti-TRBC1 CAR not only avoided resistance to anti-TRBC1 CAR-T, but reduced exhaustion and terminal differentiation.

Conclusions

Although anti-TRBC1 CAR-T appeared a promising approach for T-cell malignancy, unfractioned T cells transduced to express anti-TRBC1 CAR could not only produce CAR-C1 cells which had lesser killing ability against TRBC1+ malignant T cells and moreover were resistant to anti-TRBC1 CAR-T, but contaminate TRBC1+ cells which promoted exhaustion and terminal differentiation of anti-TRBC1 CAR-T. Therefore, it was necessary to pre-deplete TRBC1+ T cells, even if allogeneic T cells were used for anti-TRBC1 CAR-T manufacturing for patients without sufficient autologous T cells.