T cell receptor gene-modified allogeneic T cells with siRNA for endogenous T cell receptor induce efficient tumor regression without graft-versus-host disease

2.1 Sample collection

PBMCs were isolated from healthy donors who provided informed consent. The study protocol was approved by the clinical research ethics committee of Nagasaki University Hospital. PBMCs were cultured in GT-T551 culture medium (Takara Bio) supplemented with 0.6% autologous plasma, 0.2% human serum albumin, and 600 IU/mL interleukin-2.

2.2 Retroviral and lentiviral transduction

We created a retroviral vector encoding NY-ESO-1-specific TCRα and TCRβ genes, including siRNA constructs that specifically downregulated endogenous TCR (siTCR retroviral vector; MS3-NY-ESO-1-siTCR). The vector includes DNA encoding TCR-α and TCR-β chains specific for NY-ESO-1_157-165 and HLA-A*02:01; the Gly50 and Ala51 residues in the CDR2 region of the TCR-β chain were replaced with Ala and Glu, respectively, resulting in a reduction in the K_d from 21.4 mM for the wild type to 1.9 mM.²¹ The coding sequences of the TCR-α and -β transgenes are codon-optimized and therefore resistant to siRNA against endogenous TCR.¹⁸ Human lymphocytes transduced with MS3-NY-ESO-1-siTCR showed NY-ESO-1-specific cytotoxicity in human tumor cells.²²

To target β-2 microglobulin, which is required for the expression of HLA class I, we created a guide RNA (GAGTAGCGCGAGCACAGCTAAGG) that was ligated into the pLVSIN-EF1α (Takara Bio) lentiviral vector.

PBMCs were stimulated with 30 ng/mL OKT-3(eBioscience) and 600 IU/mL interleukin-2 prior to transduction. Plates were coated with 20 μg/mL RetroNectin (Takara Bio) and incubated overnight at 4°C. Retroviral or lentiviral solutions were preloaded onto RetroNectin-coated plates, centrifuged at 2000 g for 2 h, and rinsed with PBS, according to the RetroNectin-bound virus infection method. The cells were then applied onto preloaded plates. We transduced viral vector twice on two consecutive days with MOI between three and five; PBMCs transduced with the retroviral or lentiviral vector were designated as gene-modified cells. Control PBMCs were stimulated with OKT3 in a similar manner and cultured in the same medium; these specimens were designated as unmodified cells.

2.3 Enrichment of Human leukocyte antigen-negative, CD8-positive, or CD4-positive T cells

Cells were washed with buffer (PBS containing 2% FBS) and incubated for 15 min with phycoerythrin-conjugated anti-HLA, -CD8, or -CD4 mAbs at 4°C. After two washes, the cells were incubated for 15 min with phycoerythrin microbeads at 4°C. After two more washes, the cells were passed through an LS Column (Miltenyi Biotec), and the flow-through fraction (HLA-negative cells) or non-flow through fractions (CD8-positive or CD4-positive cells) were collected for further use. All enrichment was over 98%.

2.4 Intracellular cytokine staining assay

The cells were plated at an effector-to-target ratio of 2:1 (2 × 10⁵ effectors: 1 × 10⁵ targets) in 200 μL of RPMI medium containing 10% FBS in a 96-well plate. The effector cells were unmodified PBMCs or gene-modified PBMCs, and the target cells were peptide-pulsed T2 cells. Next, allophycocyanin-labeled anti-CD107a antibody was added, and the plate was incubated at 37°C for 1 h before adding Golgi Stop, followed by incubation for another 4 h. Therefore, the effector cells and the target cells were co-cultured for a total of 5 h. The surface marker was stained using the following antibodies: anti-CD3 (Becton Dickinson Biosciences, Franklin Lakes, NJ, USA), anti-CD4 (BD Biosciences), and anti-CD8 (BD Biosciences). After washing with PBS, the cells were incubated in Cytofix/Cytoperm (BD Biosciences) for 20 min. After washing twice with Perm wash buffer (BD Biosciences), the cells were incubated with intracellular staining-antibodies (anti-IFNγ [BioLegend] and anti-TNFα [BioLegend]) for 30 min. After washing with fluorescence-activated cell sorting buffer, the samples were analyzed using a BD LSRFortessa flow cytometer (BD Bioscience). Data analysis was performed using FlowJo (Tree Star).

2.5 ELISA

Target cells were washed and suspended at a density of 1 × 10⁵ cells/mL in RPMI, and then 100 μL of each target cell type was added in duplicate to a 96-well round-bottom plate. Effector T cells were washed and resuspended at 1 × 10⁶ cells/mL in RPMI medium, and then 100 μL of T cells was combined with the target cells in the indicated wells. The plates were incubated at 37°C for 18–24 h and the supernatant was harvested. INF-γ ELISA was performed as previously described.²³

2.6 Mouse studies

The animal experiments were conducted using 8–9-week-old immunodeficient non-obese diabetic/SCID/γc^null (NOG) mice. The mice were irradiated (2.5 Gy). The next day, they were injected with 2.5 × 10⁶ NY-ESO-1-expressing human tumor cell line NW-MEL-38, followed by administration of 3.0 × 10⁶ TCR gene-modified or unmodified cells through the tail vain; 25%–35% tetramer-positive cells were detected in the non-selected gene-modified cells. We monitored the mice for tumor growth and the development of GVHD two to three times per week. The development of GVHD was evaluated as weight loss. The study protocol was approved by the ethics committee of animal experiments at Nagasaki University.

2.7 Mixed lymphocyte reaction

We labeled allogeneic effector lymphocytes with 5 μM carboxyfluorescein succinimidyl ester (CFSE) and cultured with mitomycin C-treated target cells at an effector-to-target ratio of 1:1 (2 × 10⁵ cells) in 200 μL of RPMI medium containing 10% FBS in a 96-well plate. After 5 days, cells were harvested and stained using the following antibodies: anti-CD3 (Becton Dickinson Biosciences), anti-CD4 (BD Biosciences), and anti-CD8 (BD Biosciences). Proliferation of effector lymphocytes were evaluated by CFSE intensity using a BD LSRFortessa flow cytometer (BD Bioscience).

2.8 Immunohistochemistry

Bone marrow tissues were harvested from NOG mice and were processed into paraffin-embedded and frozen tissue sections for hematoxylin and eosin staining. Frozen tissues were stained by anti-human CD4 (BioLegend) and anti-human CD8 (BioLegend) mAbs.

2.9 Statistical analysis

Categorical values were compared using the χ² test. Nonparametric continuous values were compared using the Mann–Whitney U-test. Statistical significance was defined as a p-value of <0.05.

3.1 Gene-modified cells with siTCR vector showed an anti-tumor effect and reduced reactivity to allogeneic lymphocytes

We previously reported that human lymphocytes transduced with a retroviral vector encoding siRNA constructs specific to endogenous TCR reduced the expression of endogenous TCR and exhibited high surface expression of the introduced tumor-specific TCR.^18-20 Human lymphocytes transduced with a high-affinity TCR specific to a cancer/testis antigen, NY-ESO-1, by a retrovirus vector with siRNA specific to endogenous TCR (siTCR vector) were reported to induce significant tumor response in patients with NY-ESO-1 expressing synovial sarcoma.²²

Figure 1A shows a representative flow cytometry analysis result of NY-ESO-1-specific TCR expression in the human lymphocytes transduced with siTCR retroviral vector: MS3-NY-ESO-1-siTCR. By staining with an NY-ESO-1_157-165/HLA-A*02:01 tetramer, the expression of NY-ESO-1-specific TCR was found to be 29.2% in CD4⁺ cells and 33.8% in CD8⁺ lymphocytes (Figure 1A). As this TCR showed high affinity for and bound to the tetramer in a CD8-independent manner, we concluded that the tetramer could detect TCR on CD4⁺ T cells.

To test the tumor antigen-specific activity of these gene-modified cells, we evaluated the release of the degranulation marker CD107a and cytokine secretion in an intracellular staining assay and ELISAs. Release of CD107a and secretion of IFNγ and TNFα were observed only when siTCR gene-engineered T cells were stimulated with the NY-ESO-1 peptide, indicating specific anti-tumor activity (Figure 1B,C).

Next, we evaluated the expression of endogenous TCR in gene-modified cells. Lymphocytes were stained with an A2 NY-ESO-1-specific tetramer and endogenous TCR Vβ mixture (Vβ1, 2, 3, 17) antibody. The percentage of endogenous TCR-positive cells was approximately 30% in unmodified lymphocytes, whereas this value in gene-modified lymphocytes was reduced to approximately 20%. Furthermore, in NY-ESO-1 tetramer-positive gene-modified lymphocytes, the percentage significantly decreased to 6% (Figure 1D). This result was similar to that observed for CD8⁺ lymphocytes. These results indicated that the expression of endogenous TCR was downregulated in TCR gene-engineered T cells with siTCR vector.

The reactivity of siTCR gene-modified cells to allogeneic lymphocytes was examined in a thymidine uptake proliferation assay. Along with decreased expression of endogenous TCR, NY-ESO-1 tetramer-positive gene-modified cells exhibited dramatically reduced reactivity to allogeneic lymphocytes (Figure 1E).

3.2 Adoptive transfer of T cell receptor gene-modified T cells into immunodeficient non-obese diabetic/SCID/γcnull mice induced complete tumor regression without graft-versus-host disease development

To investigate whether human lymphocytes genetically engineered to express NY-ESO-1-specific TCR could inhibit the growth of NY-ESO-1-expressing human tumors, we adoptively transferred TCR gene-modified T cells into immunodeficient NOG mice. Eight- to nine-week-old irradiated NOG mice were inoculated with the NY-ESO-1-expressing HLA-A*02:01-positive human tumor cell line NW-MEL-38 and then administered TCR gene-transduced T cells. We monitored the mice for tumor growth and the development of GVHD two to three times per week. The development of GVHD was evaluated by weight loss (Figure 2A).

Transfer of unmodified cells failed to control the growth of human tumor cells, and lethal GVHD developed. Tumor growth was slightly suppressed in mice receiving unmodified cells compared to the untreated mice because of the unhealthy states caused by lethal GVHD. In the mouse model into which gene-modified whole cells were transferred, tumor growth and weight loss were slower than those in the untreated model. Furthermore, transfer of TCR-positive enriched cells induced complete tumor regression without GVHD development (Figure 2B). We evaluated peripheral bloods harvested from the mice after administration of human T cells. Human CD3⁺ lymphocytes were detectable in both the groups receiving unmodified cells and gene-modified cells, indicating that the transferred T cells resided in mice during the monitor period (Figure 2C).

We transferred purified CD8⁺ (Figure 2D) or CD4⁺ (Figure 2E) lymphocytes into a similar mouse model and evaluated the tumor size and body weight. Transfer of CD8⁺ TCR gene-modified cells suppressed tumor growth, whereas CD4⁺ TCR gene-modified cells did not. In addition, both CD8⁺ and CD4⁺ TCR gene-modified cells suppressed GVHD development. Namely, CD8⁺ T cells contributed to tumor rejection, whereas both CD8⁺ and CD4⁺ T cells contributed to GVHD development. The NOG mice transduced with unmodified human PBMCs developed severe anemia and thrombocytopenia on day 47. In contrast, the mice receiving TCR-positive enriched cells maintained normal red blood cell and platelets counts (Figure 2F). The bone marrow showed severe fibrosis and low cellularity in mice transferred with unmodified human PBMCs, whereas mice transferred with TCR-positive enriched cells maintained a normal bone marrow tissue structure (Figure 2G), suggesting that destruction of the bone marrow caused the anemia and thrombocytopenia. Increased infiltration of human CD4⁺ and CD8⁺ T cells were found in mice transferred with unmodified human PBMCs compared with that of mice that received TCR-positive enriched cells (Figure 2H).

3.3 β-2 microglobulin gene-edited T cells lost the capacity to stimulate allogeneic lymphocytes with retained anti-tumor activity

To suppress the expression of HLA class I molecules, we designed guide RNA constructs targeting β-2 microglobulin, which is required for the expression of HLA class I. Using a lentiviral vector containing these guide RNA constructs, we generated HLA class I-deficient T cells. We analyzed the expression of HLA in gene-edited T cells using flow cytometry. HLA-deficient cells were detected in approximately 15% of lymphocytes by staining with HLA-A, B, and C antibodies (Figure 3A). To enrich HLA class I-deficient cells, we performed bead selection and sorted HLA-deficient cells with high purity (99%; Figure 3B).

To evaluate whether HLA-deficient T cells were recognized by allogeneic lymphocytes, we labeled allogeneic effector lymphocytes with carboxyfluorescein succinimidyl ester (CFSE) and investigated the proliferation of effector lymphocytes in a mixed lymphocyte reaction (MLR). CD8⁺ cells were sorted as effector cells and labeled with CFSE. Autologous lymphocytes, allogeneic lymphocytes, and HLA class I-deficient allogeneic lymphocytes were evaluated. Autologous lymphocytes did not stimulate effector cells, whereas allogeneic lymphocytes stimulated 50% of the effector cell fraction. However, when stimulated with HLA-negative allogeneic lymphocytes, this reactivity decreased to the same level as that observed for autologous cells (Figure 3C). Therefore, HLA class I-deficient T cells were considered to have lost the capacity to stimulate allogeneic lymphocytes. The efficiency of β-2 microglobulin knockout was not high, which was expected due to the low titer of the lentivirus. In the MLR analysis, HLA class I-deficient cells showed reduced antigenicity, indicating that the design of the gRNA sequence was appropriate and that effective HLA class I knockout was achieved.

Next, we generated T cells deficient in endogenous TCR and HLA class I molecules by serial transduction of both viral vectors: the siTCR retrovirus vector and the lentivirus vector targeting β-2 microglobulin. We conducted flow cytometry analysis of NY-ESO-1-specific TCR expression and HLA expression. After staining with a specific tetramer, the expression of NY-ESO-1-specific TCR was approximately 40%, and the percentage of HLA-A-, B-, and C-negative cells was approximately 10% in lymphocytes (Figure 4A). NY-ESO-1-specific TCR-positive T cells deficient in endogenous TCR and HLA class I accounted for approximately 3% of cells. To enrich these T cells, we performed bead selection and sorted HLA-deficient cells.

To test the anti-tumor activity of these T cells, we evaluated the release of the degranulation marker CD107a and the secretion of cytokines using an intracellular staining assay. CD107a release and IFNγ and TNFα secretion were confirmed following stimulation with NY-ESO-1 peptide, indicating that the specific anti-tumor activity had been retained (Figure 4B).

By using T cells deficient in endogenous TCR and HLA class I molecules, we evaluated the in vivo anti-tumor effect in NOG mice. T cells deficient in endogenous TCR and HLA class I showed similar in vivo anti-tumor effects as HLA class I unmodified cells in mice based on tumor growth (Figure 4C).