To produce OT1-iT cells converted from the OT1 pro-pre-B cells, we sorted OT1 pro-pre-B cells (CD3^-Mac1^-Ter119^-B220⁺CD19⁺CD93⁺IgM^-) from the bone marrow nucleated cells of OT1 C57BL/6 transgenic mice and transduced them with Hoxb5 retroviruses or green fluorescent protein (GFP) control following a previous protocol.16 Next, the transduced cells were retro-orbitally transplanted into sublethally irradiated Rag1^-/- mice (C57BL/6, 3.5 Gy, 5 million cells/mouse) to generate the OT1-iT cells (online supplementary figure S1a; figure 1A,B). Four to six weeks post-transplantation, the OT1-iT cells appeared in the PB, lymph node (LN), and spleen (SP) of the recipient OT1-iT-Rag1^-/- mice (figure 1C,D). Additionally, the OT1-TCR proteins were expressed on the surface of the stage 1 double-negative thymocytes (DN1 cells) in the thymus of the OT1-iT-Rag1^-/- mice (figure 1E). As expected, there were no iT generated in the PB of the Rag1^-/- recipients transplanted with GFP control transduced pro-pre-B cells (figure 1C). To validate that the OT1-iT cells were derived from the OT1 pro-pre-B cells rather than natural OT1 T-cell contaminants, we performed DNA sequencing of B cell receptor (BCR) heavy chain (IgH) rearrangements using the genome from the single OT1-iT cells which were sorted from the SP of the OT1-iT-Rag^-/- mouse using a previously reported protocol.15 As expected, the single OT1-iT cells contained B-cell antigen receptor immunoglobulin heavy-chain V(D)J rearrangements (online supplementary figure S1b), which signaled their B cell origin. Furthermore, donor-derived Lin^-Sca1⁺c-kit⁺ (LSK) and common lymphoid progenitor (CLP) cells were absent in the bone marrow of the recipients’ 6 weeks’ post-transplantation (online supplementary figure S1c), which further excludes the possibility of donor long-term HSC contamination. Collectively, these results indicate that OT1 pro-pre-B cells can be converted into OT1-iT cells in the presence of Hoxb5.

To establish the tumor targets of the OT1-iT cells, we constructed an ovalbumin (OVA)-expressing B16F10 melanoma cell line (B16F10-OVA), which presents the MHC-I restricted OVA antigen. Next, we cocultured SP-derived OT1-iT cells (effector (E) cell) with B16F10-OVA cells (target (T) cell) to examine their antitumor activity. We chose wild-type T cells (WT-T) from C57BL/6 mouse as the negative control for their natural TCR repertoire diversity. Particularly, primary splenic OT1-iT cells (1×10³, 1×10⁴, 5×10⁴, 1×10⁵, and 2×10⁵) or splenic WT-T cells (1×10³, 1×10⁴, 5×10⁴, 1×10⁵, and 2×10⁵) isolated from OT1-iT-Rag1^-/- or WT mice were cocultured with 1×10⁴ B16F10-OVA cells. The number of B16F10-OVA cells sharply decreased after 36 hours of coculture with the primary OT1-iT cells at various E:T ratios (figure 2A,C upper panel and figure 2D) compared with the WT-T control group (p<0.01, p<0.001). Moreover, CD8⁺ OT1-iT cells exhibited robust proliferation in the presence of the B16F10-OVA cells within 7 days, while control CD8⁺ WT-T cells proliferated much more slowly. This indicated direct tumor cell-stimulated activation of the OT1-iT cells (figure 2G, left panel and figure 2H). To test whether T-cell activation preceding the tumor cell coculture could enhance the tumor-killing ability of the OT1-iT cells, we stimulated the primary OT1-iT cells or the primary WT-T cells with CD3/CD28 antibodies for 4 days in vitro. Strikingly, a much lower number of the preactivated OT1-iT cells (1×10², 1×10³, 5×10³, 1×10⁴, and 2×10⁴) than the primary OT1-iT cells could significantly kill the B16F10-OVA cells (1×10⁴) in vitro, whereas the activated WT-T cells still exhibited no-killing behavior even at the highest E:T ratio (2:1) (p<0.01, p<0.001) (figure 2B,C lower panel and figure 2E). The proliferation experiments also showed that the activated OT1-iT cells also proliferated much faster than the activated WT-T cells after coculture with the B16F10-OVA cells (figure 2G, right panel and figure 2I). Moreover, the OT1-iT cells demonstrated significantly efficient cytotoxicity activity over WT-T cells as the E:T ratio increased (figure 2F), and the proliferation of the cocultured B16F10-OVA tumor cells was inhibited in the OT1-iT group (p<0.5) (online supplementary figure S2a and b). Taken together, the OT1-iT cells converted from B cells effectively killed tumor cells in vitro.

To evaluate the impact of OT1-iT cells on B16F10-OVA tumor proliferation in OT1-iT-Rag1^-/- mice, B16F10-OVA tumor cells were subcutaneously injected into OT1-iT-Rag1^-/- mice 6 weeks’ post-transplantation (figure 3A). Consequently, the tumor sizes were less than 160 mm² on day 28 post-B16F10-OVA injection, while those in the control (Rag1^-/- transplanted with B16F10-OVA) group were around 400 mm² on day 22 postinjection, leading to euthanasia to comply with experimental animal ethics procedures (figure 3B). Comparatively, the B16F10-OVA tumor-bearing OT1-iT-Rag1 ^-/- mice survived up to 40 days (figure 3C), demonstrating a prolonged survival than the untreated control. In addition, we observed that the tumor-infiltrated OT1-iT cells (CD45.2⁺CD8⁺) secreted interferon gamma (IFNγ) and granzyme B (GzmB) (figure 3D,E), which indicated their tumor cell killing behavior. Thus, these results demonstrated that the OT1-iT cells can reduce tumor development in OT1-iT-Rag1^-/- mouse.

To mimic the treatment scenario of patients with tumor, we performed an adoptive transfer assay of the SP-derived OT1-T cells (OT1 transgenic mouse, positive control) and OT1-iT cells to allogenic mice bearing tumors. We isolated splenic OT1-T cells from OT1 mice, OT1-iT cells from primary OT1-iT-Rag1^-/- mice, and WT-T cells from the WT mice followed by expansion and activation, and then adoptively transferred them into tumor-bearing mice (figure 4A). As expected, the tumor burden in the OT1-iT cell-treated B16F10-OVA tumor-bearing mice was significantly reduced, and these mice survived up to 43 days post-B16F10-OVA tumor cell injection. This achieved comparable therapeutic effects as the OT1-T cell treatment (up to 42 days). In contrast, tumor sizes of WT-T cell-treated mice reached the limit of ethic allowance within 28 days’ postinjection, resulting in their sacrifice (figure 4B,C). We further analyzed the activation status of the OT1-iT and OT1-T cells infiltrated in the tumors and observed that both the OT1-iT and OT1-T cells (CD45.2⁺CD8⁺) were completely activated (CD44^hiCD69⁺CD62L^-), comparing with primary T cells. Notably, OT1-iT cells had more activated CD44^hiCD69⁺CD62L^- cells than OT1-T cells (figure 4D,E). In addition, the activated OT1-iT cells secreted comparable levels of IFNγ and GzmB to OT1-T cells (figure 4F,G), indexing their tumor-eradicating behavior. Furthermore, programmed cell death protein 1 (PD-1) expression was upregulated in the infiltrated OT1-T and OT1-iT cells in the mice with tumor recurrence (figure 4H,I). These results illustrate the antitumor capacity of the reprogrammed OT1-iT cells in vivo.

Pro-pre-B cells’ (C57BL/6 mouse) or OT1 pro-pre-B cells’ (OT1 transgenic mouse) isolation, infection, and transfer were performed as previously described.16 Briefly, pro-pre-B cells from WT mice or OT1 transgenic mice were first enriched via positive magnetic affinity cell sorter selection using B220-biotin and anti-biotin MicroBeads (Miltenyi Biotec), and then sorted from the enriched B220⁺ cells by Aria III (BD). The sorted cells were subsequently stimulated with the pro-pre-B cell medium for 12–16 hours prior to retroviral transduction. Pro-pre-B cells (Hoxb5 retrovirus or OT1-Hoxb5 retrovirus) and OT1 pro-pre-B cells (GFP retrovirus or Hoxb5 retrovirus) were transduced with retrovirus by two rounds of spin transfection (800 g, 90 min, 35℃) at a density of 1 million/mL. For transplantation, 5 million GFP⁺ pro-pre-B cells or OT1 pro-pre-B cells were injected into the retro-orbital veins of the irradiated Rag1^-/- recipients (3.5 Gy, RS2000; Rad Source). All recipients were given water supplemented with trimethoprim–sulfamethoxazole for 2 weeks to prevent infection. OT1-iT lymphocytes were analyzed 4–6 weeks’ post-transplantation.

Primary OT1-iT cells or WT-T cells derived from the SP of the OT1-iT-Rag1^-/- mice or C57BL/6 mice were enriched by depletion of Ter119⁺CD11b⁺Gr1⁺B220⁺NK1.1⁺CD11c⁺ cells and cultured in the T-cell medium without interleukin 2 at a density of 1×10⁶/mL. Activated OT-iT cells were obtained by coculturing primary OT1-iT cells with Dynabeads Mouse T-Activator CD3/CD28 (CD3/CD28 Gibco) for T-cell expansion and activation for 4 days. Primary or activated OT1-iT (E) cells were incubated with 1×10⁴ B16F10-OVA (T) cells in 96-well plates for 36 hours at respective E:T ratios (primary T cells, E:T=0.1:1, 1:1, 5:1, 10:1, 20:1; activated T cells, E:T=0.01:1, 0.1:1, 0.5:1, 1:1, 2:1). The number of B16F10-OVA tumor cells (DsRed⁺) were enumerated using the CountBright Absolute Counting Beads (Thermo Fisher) by LSRFortessa-X20 (BD). Microphotographs were taken to assess the antitumor effect of primary or activated OT1-iT using the ImageXpress Micro Confocal (Molecular Devices). The Non-Radioactive Cytotoxicity Assay kit (Promega) was used to analyze the lactate dehydrogenase released from the coculture cells, and the cytotoxicity of the OT1-iT cells were analyzed following the instructions of the assay kit.

The proliferation analysis of the OT1-iT cells was performed as described.22 Prior to coculture, OT1-iT cells or WT-T cells were stained with the Cell Proliferation Dye eFluor 670 according to the manufacturer’s directions. The stained T cells and B16F10-OVA were cocultured with T-cell medium. The proliferation status of the OT1-iT cells (CD45.2⁺CD8⁺) were analyzed at day 0, day 3 and day 7 after coculture by the LSRFortessa-X20 (BD).

OT1-iT-Rag1^-/- mice or Rag1^-/- mice, which were used as the tumor model, were transplanted with B16F10-OVA cells (0.15 million/mouse) in the groin by subcutaneous injection. For the OT1-iT-Rag1^-/- tumor model, Rag1^-/- mice were transplanted with the OT1 pro-pre-B cells transduced with Hoxb5 retroviruses (5 million/mouse) or pro-pre-B cells transduced with OT1-Hoxb5 retroviruses 6 weeks prior to the B16F10-OVA cell injection. For adoptive transfer, the splenic T cells (WT-T, OT1-T, OT1-iT) were first expanded and activated for 7 days in vitro using the Dynabeads Mouse T-Activator CD3/CD28 for T-cell expansion and activation. The expanded and activated T cells (10 million/mouse) were transplanted into the tumor-bearing Rag1^-/- mice 10 days after B16F10-OVA cell injection. The tumor size was measured every other day using calipers and calculated as length ×width (mm²). Mice with tumor sizes larger than 400 mm² were euthanized for ethical consideration.

OT1-iT cells were isolated from the melanoma tumors from the tumor-bearing mice as previously described.23 The isolated cells were stained with antibodies against CD45.2, CD8a, TCRVα2, and TCRVβ5. For the intracellular staining, cells isolated from the tumors were first stained with the surface antibodies (CD45.2, CD11b, Gr1 and CD8), fixed, and then stained with Allophycocyanin (APC)-conjugated IFNγ and PE-conjugated GzmB.

Flow cytometry data were analyzed by the FlowJo software. Prism7 (GraphPad) and SPSS (V.22.0) were used for the statistical analysis.

Extended experimental procedures including regents, cell culture, retrovirus preparation, and BCR analysis are described in online supplementary file.