Development of Engineered T Cells Expressing a Chimeric CD16-CD3ζ Receptor to Improve the Clinical Efficacy of Mogamulizumab Therapy Against Adult T-Cell Leukemia

To date, except for allogeneic hematopoietic stem cell transplantation (allo-HSCT) for some patients with aggressive adult T-cell leukemia (ATL), an intractable peripheral T-cell malignancy caused by human T-lymphotropic virus type-1 (HTLV-1) infection, no reliable treatment has yet been established (1, 2). Against this background, based on promising results of phase I and II clinical trials in the setting of refractory ATL (3, 4), mogamulizmab (Mog), a humanized mAb against CC chemokine receptor 4 (CCR4) has recently been approved for the treatment of aggressive ATL in Japan with high expectation, as a large population of ATL tumor cells substantially express CCR4 (5). However, in daily practice, failure of Mog treatment has sometimes been observed in patients who have concomitant chemotherapy-associated lymphocytopenia. As is the case for other antitumor mAbs whose central tumoricidal activity is based on antibody-dependent cellular cytotoxicity (ADCC) against opsonized tumor cells without HLA restriction (6), such as anti-CD20 mAb (rituximab) for CD20⁺ B-cell malignancy (7) and anti-human epidermal growth factor receptor-2 (Her2/Neu) mAb (herceptin) for Her2/Neu⁺ solid tumors (8), Mog exerts potentiated ADCC activity mediated by FcγR IIIa(CD16)-bearing endogenous immune cells, such as natural killer (NK) cells, through increased binding affinity to FcγR IIIa achieved by defucosylation of the Fc portion (9). Recently, it has been reported that both circulating CD16⁺ α/β- T cells (10) and CD16⁺ γ/δ- T cells (11) are also involved in ADCC-mediated antiviral activity in vivo. In terms of antitumor mAb therapy with Mog, we have often observed in daily practice that clinical efficacy declines as treatment proceeds. This loss of efficacy is considered attributable to both chemotherapy-induced leukopenia and exhaustion of NK cells evoked by the ADCC process, that is, a substantial loss of active effector cells (12). Accordingly, therapeutic replenishment of CD16-expressing ADCC effector cells together with Mog would seem a logical approach for patients with aggressive ATL who have received intensive chemotherapy. By analogy, adoptive transfer of ex vivo–expanded NK cells (13, 14) and γδ T cells (15) in combination with antitumor mAbs is another possibility, but as yet, no reliable procedure has been established for ex vivo expansion of those autologous effector cells to therapeutically sufficient numbers (16). Alternatively, clinical trials for the treatment of refractory CD20⁺ lymphoma using combined administration of rituximab and culture-induced killer (CIK) cells (17) or lymphokine-activated killer (LAK) cells (NCT 01329354, ClinicalTrials.gov), both of which comprise substantial populations of NK cells, are underway, but no conclusive results have yet been reported. On the other hand, it has been revealed that FcγRIIIa-158 amino acid valine/phenylalanine (V/F) biallelic polymorphism influences the efficacy of NK cell activation; the binding affinity of the Fcγ RIIIa-158V/V homodimer (158V/V) to IgG1 and IgG3 is much higher than that of the FcγRIIIa-158F/F homodimer (158F/F; refs. 18–20). In addition, accumulated evidence has confirmed the long-term safety of retro- or lentivirally gene-modified T cells for adoptive cell therapy in terms of the risk of leukemogenesis (21, 22). In this context, we have previously examined whether lentivirally gene-modified CD3⁺ T cells can express a newly designed FcγRIIIa-158(V/V)-CD3ζ chimeric receptor (cCD16ζ-T cell) as an alternative ADCC effector (23).

In the current study, to expand the clinical versatility of this concept, we preclinically examined both in vitro and in vivo the feasibility of a novel form of adoptive immunotherapy for aggressive ATL using Mog in combination with cCD16ζ-T cells. Our experimental observations suggest that this approach could be a potentially realistic option for not only improving the efficacy of Mog, but also for providing a novel treatment for patients with aggressive ATL, particularly those who are ineligible for allo-HSCT.

Approval for this study was obtained from the Institutional Review Board of Ehime University Hospital. Written informed consent was provided by all healthy volunteers and patients with ATL in accordance with the Declaration of Helsinki. The HEK 293T cell line (RIKEN BioResource Center, Tsukuba, Japan) was maintained in DMEM (Sigma Aldrich) supplemented with 10% FBS (Biowest) and 2 mmol/L l-glutamine (Sigma Aldrich), and employed for production of infectious lentivirus particles. K562 (ATCC), Jurkat (ATCC), the human T-lymphotropic virus type-I (HTLV-1) immortalized and IL-2-independent T-cell lines MT-1, MT-2, MT-4, TL-MAT, TL-Su, and Hut 102, (25) and the IL-2-independent ATL cell line ATN-1 (25) were cultured in RPMI1640 (Sigma Aldrich) supplemented with 10% FBS, 2 mmol/L l-glutamine, and penicillin/streptomycin (Biowest). B-lymphoblastoid cell lines (LCL) were established by transformation of peripheral blood B lymphocytes with Epstein Barr Virus (EBV). The HLA-A*24:02 gene-transduced K562 (K562-A24) was maintained in culture medium supplemented with 1.0 μg/mL puromycin (Sigma-Aldrich). Peripheral blood mononuclear cells (PBMC) from healthy volunteers and patients with ATL in remission after chemotherapy, and CD4⁺ autologous ATL tumor cells from samples of peripheral blood taken on admission were stored in liquid nitrogen until use.

A lentiviral vector expressing a chimeric CD16 with a 158V/V-CD3ζ(cCD16ζ) construct (ref. 23; Supplementary Fig. S1A) was synthesized and inserted into the lentiviral vector, pRRLSIN.cPPT.MSCV/GFP.WPRE (26), as described elsewhere (23).

Mogamulizumab (Poteligeo) was purchased from Kyowa-Hakko-Kirin, Ltd.

All in vivo mouse experiments were approved by the Ehime University animal care committee. Six-week-old NOD/scid/γc^null (NOG) female mice were purchased from the Central Institute for Experimental Animals (Kawasaki, Japan; ref. 27) and maintained in the institutional animal facility at Ehime University (Ehime, Japan).

T cells that had been gene-modified using cCD16ζ gene transfer, non-gene–modified T cells, natural killer (NK) cells isolated from PBMCs using MACS beads (Miltenyi Biotec), and cell lines were labeled with anti-CD3, anti-CD4, anti-CD8, anti-CD20, anti-CD45, anti-CD56, anti-CD62L, anti-CD194 (also known as CCR4; BD Biosciences), anti-CD45RA (eBioscience), anti-CD16, and anti-CD69 (BioLegend) mAbs. Flow cytometry was conducted using a Gallios flow cytometer (Coulter), and data analysis was performed using Flow Jo Version 7.2.2 software (TreeStar).

cCD16ζ gene-transduced T cells (cCD16ζ-T cells) were established as described elsewhere (23). Briefly, CD3⁺ T cells isolated from PBMCs of healthy volunteers and CD8⁺ T cells from patients with ATL using MACS beads (Miltenyi Biotec) were cultured with 50 U/mL rhIL2 (R&D Systems), and stimulated with anti-CD2/CD3/CD28 MACSiBeads (Miltenyi Biotec). Because ATL tumor cells reside in patients' CD4⁺ T cells, we employed nontumor CD8⁺ T cells to generate these ADCC effector T cells regarding ATL patients. Four days later, the cCD16ζ gene was lentivirally transduced into the cultured T cells using 10 μg/mL protamine sulfate (Mochida Pharmaceutical Co. Ltd.). One week later, CD16⁺ transfectants were further isolated using MACS beads (Miltenyi Biotec). After examination of the phenotypes of the CD16⁺ transfectants, these cells were subjected to further expansion and functional analyses. In some experiments, cCD16ζ-CD8⁺ T cells were additionally gene-modified using the HLA-A*24:02–restricted and human telomerase reverse transcriptase (hTERT) peptide 461-469;VYGFVRACL–specific T-cell receptor (TCR) gene (25) and subsequently subjected to functional assays.

To determine the ADCC activity mediated by cCD16ζ-T cells or NK cells, a standard ⁵¹Cr-release assay was performed, as described elsewhere (28). Briefly, 10⁴ target cells were labeled with ⁵¹Cr (Na₂CrO₄; MP Bio Japan) and incubated with various concentrations of Mog or at different effector to target (E:T) ratios with effector cells in 200 μL of culture medium in 96-well round-bottom plates. After 4 hours of incubation with the effector cells at 37°C, 100 μL of the supernatant was collected from each well. The percentage of specific lysis was calculated as: (experimental release cpm − spontaneous release cpm)/(maximal release cpm − spontaneous release cpm) × 100 (%). In some experiments, we conducted ADCC assays with a fixed E:T ratio of 5:1 or 2:1 or a 0.1 μg/mL concentration of Mog. For blocking experiments, 10 μg/mL the F(ab′)2 fragment of the anti-human CD16-specific mAb 3G8 (Ancell) was employed.

Fifty thousand cCD16ζ-T cells were coincubated with ATN-1, MT-4, or K562 cells for 48 hours with or without 1 μg/mL Mog. cCD16ζ-T cells treated with 1 μg/mL OKT-3 (BioLegend) and 1 μmol/L/L ionomycin (Cell Signaling Technology) were employed as a positive control. IFNγ, TNFα, IL2, IL10, and IL4 in each culture supernatant were measured using an ELISA Kit (Thermo Scientific).

CD107a expression mediated by cCD16ζ-T cells in response to Mog-opsonized tumor cells was examined as described previously (29). Briefly, 1 × 10⁵ target cells preincubated with or without 1 μg/mL Mog for 20 minutes were incubated with 2 × 10⁵ cCD16ζ-T cells for 3 hours in a 96-well round-bottom plate. After being labeled with FITC-conjugated CD107a mAb (BioLegend), anti-CD3, anti-CD8 (BD Biosciences), and anti-CD16 (Biolegend) mAbs, the cells were analyzed by flow cytometry. For blocking experiments, 3G8 mAb was added at several concentrations (10, 25, 50 μg/mL).

To examine the in vivo tumor-suppressive activity mediated by the combined adoptive therapy using cCD16ζ-T cells and Mog, we employed luciferase gene–transduced ATN-1 cells (ATN-1/luc; ref. 25) whose CCR4 expression was equal to that of the parental ATN-1 cells. All NOG mice aged 6 weeks (n = 16) were exposed to 1 Gy of irradiation, and then intravenously inoculated with 5 × 10⁵ ATN-1/luc cells via the tail vein on day 0 (leukemia model; n = 8), or subcutaneously inoculated at the frontal abdominal wall (subcutaneous tumor model; n = 8). On day 4 in leukemia model and on day 7 when tumor became larger than 3 mm × 3 mm in subcutaneous tumor model, 5 × 10⁶ cCD16ζ-T cells combined with (n = 5) or without (n = 3 as a control) 40 μg Mog/mouse were intravenously administered once. In this series of experiments, no exogenous rhIL2 was administered. Serial acquisition of the luciferase photon counts of inoculated ATN-1/luc cells was carried out as described previously (30). Photon counts relative to that on day 4 for leukemia model and on day 7 for subcutaneous tumor model before each cell therapy, which indicated the residual tumor mass burden, was calculated for each mouse. Two series of similar experiments were conducted independently.

Data were analyzed using the statistical computing package SPSS 15.0J for Windows. Data were expressed as mean ± SD. Differences were assessed using ANOVA analysis and the Mann–Whitney U test. Overall survival rate was estimated by the Kaplan–Meier method and analyzed using the log-rank test. Differences at P < 0.05 were considered significant.

Using a lentiviral vector carrying the cCD16ζ gene construct (Supplementary Fig. S1A), HEK 293 T cells were successfully gene-modified to express high amounts of CD16 on the cell surface (Supplementary Fig. S1B). By day 21 of the expansion culture, most of these effector gene-modified T cells had become CD3⁺CD8⁺CD56⁻CD16⁺ T cells (75.2 ± 3.1%, n = 10; Fig. 1A). The various levels of CCR4 expression displayed by the ATL cell line cells employed in this study are listed in Fig. 1B. K562 and K562-A24 (myeloid leukemia cell lines), Jurkat (a T-cell leukemia cell line), and LCL (an EBV-immortalized B-cell line), all negative for CCR4, were assessed as negative controls (Supplementary Fig. S1C). Using CCR4-high MT-4 (MFI = 8.92) and CCR4-intermediate ATN-1 (MFI = 2.77) as targets, cCD16ζ-T cells exerted similar tumoricidal activities against the target cells opsonized with Mog in an antibody–dose-dependent manner. This tumoricidal activity was maximal at 1.0 μg/mL Mog (Fig. 1C). Again, using a pharmacologic dose of 0.1 μg/mL Mog observed in clinical trials (3, 4), despite various levels of CCR4 expression, Mog-opsonized ATL cell line cells were similarly and sufficiently killed by cCD16ζ-T cells (Fig. 1D). Because, in practice, ATL tumor cells also variably express CCR4 on the cell surface (5), this observation appears to be of considerable interest in the context of treatment for ATL.

During this ADCC process mediated by cCD16ζ-T cells, cCD16ζ-T cells produced IFNγ (Fig. 2A) and TNFα (Fig. 2B), and released cytotoxic granules (Fig. 2C) in response to Mog-opsonized ATN-1 and MT-4 cells, but not K562 cells in the presence of Mog. Purified cCD16ζ-CD4⁺T cells similarly produced IL2 in response to Mog-opsonized ATN-1 cells (Fig. 2D), which was clearly inhibited by 3G8, the F(ab′)2 fragment of the anti-human CD16-specific mAb in the dose-dependent manner (Fig. 2E). Similarly IFNγ production mediated by cCD16ζ-T cells in response to Mog-opsonized ATN-1 and MT-4 were inhibited by 3G8 (data not shown). In addition, cCD16ζ-CD4⁺T cells never produced either IL10 or IL4 (data not shown). Collectively, among cCD16ζ-T cells, predominant cCD16ζ-CD8⁺ T cells executed cytocidal activity and concomitant cCD16ζ-CD4⁺ T cells displayed the type I helper T-cell function during this ADCC process in response to Mog-opsonized target cells. In addition, blockade with 3G8 mAb confirmed this ADCC activity executed by cCD16ζ-T cells was evoked by the introduced chimeric CD16-CD3ζ receptor-mediated T-cell activation signal following the recognition of Mog-opsonized target cells.

Furthermore, such tumoricidal activity against ATN-1 mediated by cCD16ζ-T cells (Fig. 3C) was comparable with that mediated by CD69^high activated NK cells using rhIL2 (Fig. 3B), but not freshly isolated resting CD69^dimNK cells (Fig. 3A). Thus, collectively, cCD16ζ-T cells might potentially be able to substitute for activated NK cells as ADCC effector cells in vivo.

Because of the limited number of ATL tumor cells stored in liquid nitrogen after recovery from patients, we employed CD107a assay instead of the standard ⁵¹Cr-release assay for measurement of ADCC activity mediated by cCD16ζ-T cells against patients' ATL tumor cells opsonized with Mog. On the premise of clinical application, as ATL tumor cells displayed a CD4⁺ T-cell phenotype, normal CD8⁺ T cells from PBMC sampled from ATL patients and healthy donors were isolated for cCD16ζ gene modification. As shown in Fig. 4A, cCD16ζ-CD8⁺ T cells generated from three allogeneic healthy donors released significant amounts of cytotoxic granules in response to Mog-opsonized ATL tumor cells from 4 patients: the CCR4 expression (MFI) of the ATL tumor cells was 13.7 for patient #1, 10.6 for patient #2, 7.37 for patient #3, and 3.35 for patient #4. The ATL tumor cell line ATN-1 for which the MFI for CCR4 was 2.77 (Fig. 1B) was employed as a positive control. As shown in Fig. 4B, except for patient #4, cCD16ζ-CD8⁺ T cells successfully generated from 3 of the patients tended to display comparable cytocidal degranulation in response to Mog-opsonized autologous ATL tumor cells to that mediated by allogeneic cCD16ζ-CD8⁺ T cells (Fig. 4A). These findings strongly suggested that even heavily treated ATL patients might be eligible for this approach.

In vivo anti-ATL activity mediated by combined administration of cCD16ζ-T cells and Mog was measured using a xenografted mouse model and an in vivo bioluminescence assay. Because ATL clinically display both lymphoma phenotype and leukemia phenotype, we employed subcutaneous tumor model (n = 8; Fig. 5A–C) and leukemia model (n = 8; Fig. 5D–F). As shown in Fig. 5A, coadministered cCD16ζ-T cells and Mog (n = 4, excluded #5 that accidentally died soon after inoculation of tumor cells, shown by the red line) obviously suppressed the tumor growth in comparison with cCD16ζ-T cells alone as a control (n = 3, blue line). This suppression of tumor-growth led to significantly longer survival of the treated mice (median 54 days, range 49–85 days, vs. median 42 days, range 41–45 days; P = 0.02 by log-rank analysis; Fig. 5B). Serial images of the bioluminescence assay until day 32 demonstrated the suppression of subcutaneous tumor growth in treated cohort (Fig. 5C). In a leukemia model, combined administration of Mog and cCD16ζ-T cells also suppressed the leukemia progression (shown by red line; Fig. 5D), resulting in a significantly longer median survival time (median 45 days, range 38–61 days, vs. median 35 days, range 35–40 days; P < 0.01 by log-rank analysis; Fig. 5E). Serial images of the bioluminescence assay also demonstrated the suppression of leukemia progression in the therapeutic cohort (Fig. 5F). Together, these data suggested that coadministration of cCD16ζ-T cells with Mog exerted tumor-suppressive activity against both the lymphoma and leukemia phenotypes of ATL tumor cells in vivo.

To derive effector cells with stronger antitumor activity, we finally examined whether double gene-modified CD8⁺ T cells using the cCD16ζ gene and hTERT-specific TCR gene transfer would be able to exert both ADCC activity in combination with Mog and CTL activity simultaneously against ATL tumor cells. Representative data from three independent experiments are shown in Fig. 6, double gene-modified CD8⁺ T cells positive for both cell surface CD16 and HLA-A24/hTERT tetramer (Fig. 6A) were able to release cytotoxic granules upon recognition of ATN-1 via hTERT-specific TCR in the absence of Mog (28.7% being positive for CD107a vs. 8.4% for K562 as a HLA-A*24:02-negative control), which was obviously impeded in the presence of anti-HLA class I mAb (w6/32; Fig. 6B). At the same time, as shown in Fig. 6C, in the presence of Mog, these double gene-modified CD8⁺ T cells more actively released cytotoxic granules against ATN-1 (91% for CD107a vs. 4.7% for K562 as a CCR4-negative control), and in the presence of both Mog and w6/32 mAb, this release was reduced to 63.9%, the reduction of 27.1% being almost equal to that mediated by hTERT-specific TCR (28.7%: see Fig. 6B). Collectively, these double gene-modified CD8⁺ T cells additively exerted ADCC-mediated tumoricidal activity in the presence of Mog, and CTL-mediated tumoricidal activity via the introduced hTERT-specific TCR against ATN-1 at the same time. Therefore, use of these gene-modified CD8⁺ T cells, which are simultaneously capable of executing both ADCC- and CTL activity-mediated cytocidal activity against the same target tumor cells, seems potentially feasible.

At present, allo-HSCT can only prolong the survival of specific patients with aggressive ATL, as a large proportion of ATL patients are ineligible for allo-HSCT because of advanced age-related comorbidities which would evoke life-threatening treatment-related adverse events and severe GVHD due to allo-immunity (2, 31). On the basis of the promising results of phase I and II clinical trials (3, 4), Mog has been applied clinically with high expectation, expanding the opportunity for successful treatment of many patients with aggressive ATL who would otherwise be ineligible for allo-HSCT. In this context, the decline in the clinical efficacy of Mog is an important issue that needs to be urgently addressed. Like other anticancer mAbs, the tumoricidal activity mediated by Mog against ATL tumor cells depends on ADCC activity of endogenous immune cells expressing CD16 (3, 4, 9). Then, as treatment proceeds, repeated cycles of intensive chemotherapy to suppress the disease activity of aggressive ATL inevitably decrease the number of lymphocytes, which are the effector cells for Mog (Supplementary Fig. S3). In addition to the case shown in Fig. S3, our limited experience with Mog monotherapy for relapsed/refractory ATL, excluding patients who received allo-HSCT and combined therapy with Mog and intensified chemotherapy mSLG15 (24, 32), had shown that the lymphocyte count during Mog monotherapy in patients for whom the treatment failed to suppress disease progression tended to be lower than that when Mog monotherapy successfully controlled disease (Supplementary Table S1). Furthermore, in vitro experiments showed that at a pharmacologic dose of Mog (0.1 μg/mL), the decline in ADCC activity against Mog-opsonized MT-4 and ATN-1 mediated by cCD16ζ-T cells was more obviously dependent on the number of effector cells (E:T ratio ≤ 2.5:1; Supplementary Fig. S2A and S2B). Accordingly, therapeutic replenishment of ADCC effector cells would seem reasonable when Mog is used for treatment of aggressive ATL. In our previous study, as adoptively transferable ADCC effector cells, NK cells seemed less advantageous, mainly because they did not proliferate after execution of ADCC activity, leading to their shorter survival after infusion and limited efficacy both in vitro and in vivo (23). On the other hand, cCD16ζ-T cells proliferated transiently for about a week following execution of ADCC, resulting in longer survival in vivo and greater tumor suppression (23).

Against the above background, we examined the feasibility of cCD16ζ-T cells in the context of Mog treatment for patients with aggressive ATL, and found that the cells successfully exerted ADCC activity against ATL cell line cells expressing a wide range of CCR4 on their surface (Fig. 1B and D). This may have been at least partly attributable to enhanced binding of the Fc portion of Mog to FcγR IIIa (CD16) as a result of the unique defucosylation treatment (9) as well as augmented binding of the introduced Fc fragment of mutated CD16 with 158V/V on cCD16ζ-T cells (18–20). cCD16ζ-T cells exerted ADCC activity against Mog-opsonized ATN-1 and MT-4 cells, but not K562 cells (Fig. 1C), through production of IFNγ and TNFα, and cytotoxic degranulation (Fig. 2A–C). In addition, cCD16ζ-CD4⁺T cells were potentially capable of exerting Th1 helper function (Fig. 2D and E). We then examined whether autologous cCD16ζ-T cells generated from heavily treated patients with aggressive ATL would be able to kill autologous ATL tumor cells in the presence of Mog. Because ATL is a CD4⁺ T cell malignancy, in this setting, we employed normal autologous CD8⁺ T cells from patients to exclude tumor cells. These cCD16ζ-CD8⁺ T cells were able to successfully recognize and release cytotoxic granules against Mog-opsonized autologous ATL tumor cells, to a degree comparable with that mediated by cCD16ζ-CD8⁺ T cells using allogeneic CD8⁺ T cells from healthy volunteers (Fig. 4A and B). Utilization of chemo-naïve normal T cells from healthy volunteers seems reasonably advantageous to acquire a sufficient number of cCD16ζ-T cells for clinical use. However, in the setting of allo-HSCT for patients with aggressive ATL, a recent single-institution retrospective study from Japan has raised concern as to whether pretransplant Mog treatment might increase the risk of acute GVHD and nonrelapse mortality (33). Therefore, it will be necessary to carefully monitor any GVHD mediated by these cCD16ζ-T cells in well-designed clinical trials. Finally, we conducted therapy-oriented in vivo experiments using NOG mice inoculated with ATN-1/luc cells subcutaneously in the abdominal wall (as a lymphoma model; Fig. 5A–C) or intravenously (as a leukemia model; Fig. 5D–F). Although tumor cells were not eradicated in model, as expected, concomitantly infused cCD16ζ-T cells in combination with Mog successfully suppressed both types of disease and significantly prolonged the survival of these mice. In this setting, regrowth of tumor cells in treated mice largely seemed attributable to limited persistence of cCD16ζ-T cells in vivo, but not to the loss of CCR4 expression (ref. 34; data not shown).

In this study, we employed the cCD16–CD3ζ construct. The fact that first-generation CAR constructs containing only CD3-ζ had previously failed to fully exert the anticipated antitumor effect (35, 36) prompted us to incorporate a costimulatory molecule, such as 4.1BB (37), into the cytoplasmic portion of our chimeric CD16-CD3ζ construct, in expectation of better persistence and tumor trafficking in vivo, and possibly providing a better clinical outcome (38). In fact, we are now also devising a “second and third generation” chimeric CD16–CD3ζ construct encompassing 4.1-BB with or without CD28. However, because CCR4 is also naturally expressed by normal regulatory CD4⁺ T cells (Tregs; refs. 39, 40), this raises concern as to whether profound and prolonged depletion of Tregs induced by more powerful cCD16ζ-T cells in combination with Mog might cause severe autoimmune disease. Furthermore, no proven procedure for expanding autologous Tregs from patients with T-cell malignancy who have received intensive chemotherapy has yet been established. In clinical practice, several cases with severe skin damage probably caused by autoimmunity have been observed in ATL patients who have received Mog (3, 4, 41). Accordingly, we preformed this preclinical study to acquire proof-of-concept, beginning with a careful test of this cCD16–CD3ζ construct. On the other hand, a recent study has proposed a novel concept that CCR4 expression in normal human Tregs is limited to the “effector Treg” subset, which produces high amounts of Foxp3 and displays exclusive immunosuppressive activity (39). At the same time, the authors claimed that the “precursor Treg” subset, being negative for CCR4, would be preserved despite Mog treatment, yielding a limited potential risk of life-threatening autoimmune disease evoked by Mog treatment (39) Therefore, this issue will need to be investigated further in well-designed clinical trials.

Finally, we examined the possibility of applying double gene-modified effector T cells displaying tumor antigen (hTERT)-specific TCR and the cCD16ζ receptor, being capable of simultaneously exerting CTL activity and ADCC activity in response to the same target tumor cells, making them powerful effector cells for treatment of ATL (Fig. 6A–C). We found that these double gene-modified effector cells seemed to increase cytocidal activity in the form of ADCC activity targeting Mog-opsonized ATN-1 via the cCD16ζ receptor and CTL activity via the introduced HLA-A*24:02–restricted hTERT-specific TCR against HLA-A*24:02⁺ and hTERT overexpressing ATN-1 (25). Similar results were obtained using another set of double gene-modified T cells expressing the cCD16ζ receptor and HLA-A*24:02–restricted WT1-specific TCR (ref. 29; data not shown). To our knowledge, feasibility of such double gene-modified effector T cells using TCR and CAR gene transfer has not been much discussed yet (42). Thus, we are now planning further studies to expand the possibilities of this approach.

Although further studies are warranted, our experimental observations strongly suggest that cCD16ζ-T cells might provide another realistic option for improving the clinical efficacy of Mog, thus offering a novel approach for adoptive therapy in patients with aggressive ATL, especially those who are ineligible for allo-HSCT or for whom suitable donors are lacking.

No potential conflicts of interest were disclosed.

Conception and design: H. Tanaka, H. Fujiwara, H. Shiku, M. Yasukawa

Development of methodology: H. Tanaka, H. Fujiwara, F. Ochi, K. Tanimoto, N. Casey, J. Barrett

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Tanaka, H. Fujiwara, F. Ochi, N. Casey

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): H. Tanaka, H. Fujiwara, F. Ochi

Writing, review, and/or revision of the manuscript: H. Tanaka, H. Fujiwara, M. Yasukawa

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Tanimoto, S. Okamoto, J. Mineno, K. Kuzushima

Study supervision: H. Fujiwara, T. Sugiyama, M. Yasukawa

The authors are grateful for the technical assistance of Dr. Kenji Kameda, Ehime University and Dr. Hiroo Saji, HLA Laboratory, Japan, for HLA typing.

This work was supported in part by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (KAKEN 24591426 and 15K09506; to H. Fujiwara; and KAKEN 15H04858, to M. Yasukawa), a grant from Project for Development of Innovative Research on Cancer Therapeutics (to M. Yasukawa), a Grant-in-Aid for Cancer Research from the Ministry of Health, Labor and Welfare (to M. Yasukawa), a grant from Takeda science Foundation (to M. Yasukawa), a grant from the Uehara Memorial Foundation (to M. Yasukawa), and a grant from Princess Takamatsu Cancer Research Fund (to M. Yasukawa).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.