Adding Help to an HLA-A*24:02 Tumor-Reactive γδTCR Increases Tumor Control

γδT cell receptors (γδTCRs) recognize a broad range of malignantly transformed cells in mainly a major histocompatibility complex (MHC)-independent manner, making them valuable additions to the engineered immune effector cell therapy that currently focuses primarily on αβTCRs and chimeric antigen receptors (CARs). As an exception to the rule, we have previously identified a γδTCR, which exerts antitumor reactivity against HLA-A*24:02-expressing malignant cells, however without the need for defined HLA-restricted peptides, and without exhibiting any sign of off-target toxicity in humanized HLA-A*24:02 transgenic NSG (NSG-A24:02) mouse models. This particular tumor-HLA-A*24:02-specific Vγ5Vδ1TCR required CD8αα co-receptor for its tumor reactive capacity when introduced into αβT cells engineered to express a defined γδTCR (TEG), referred to as TEG011; thus, it was only active in CD8+ TEG011. We subsequently explored the concept of additional redirection of CD4+ T cells through co-expression of the human CD8α gene into CD4+ and CD8+ TEG011 cells, later referred as TEG011_CD8α. Adoptive transfer of TEG011_CD8α cells in humanized HLA-A*24:02 transgenic NSG (NSG-A24:02) mice injected with tumor HLA-A*24:02+ cells showed superior tumor control in comparison to TEG011, and to mock control groups. The total percentage of mice with persisting TEG011_CD8α cells, as well as the total number of TEG011_CD8α cells per mice, was significantly improved over time, mainly due to a dominance of CD4+CD8+ double-positive TEG011_CD8α, which resulted in higher total counts of functional T cells in spleen and bone marrow. We observed that tumor clearance in the bone marrow of TEG011_CD8α-treated mice associated with better human T cell infiltration, which was not observed in the TEG011-treated group. Overall, introduction of transgenic human CD8α receptor on TEG011 improves antitumor reactivity against HLA-A*24:02+ tumor cells and further enhances in vivo tumor control.

gdT cell receptors (gdTCRs) recognize a broad range of malignantly transformed cells in mainly a major histocompatibility complex (MHC)-independent manner, making them valuable additions to the engineered immune effector cell therapy that currently focuses primarily on abTCRs and chimeric antigen receptors (CARs). As an exception to the rule, we have previously identified a gdTCR, which exerts antitumor reactivity against HLA-A*24:02-expressing malignant cells, however without the need for defined HLA-restricted peptides, and without exhibiting any sign of off-target toxicity in humanized HLA-A*24:02 transgenic NSG (NSG-A24:02) mouse models. This particular tumor-HLA-A*24:02specific Vg5Vd1TCR required CD8aa co-receptor for its tumor reactive capacity when introduced into abT cells engineered to express a defined gdTCR (TEG), referred to as TEG011; thus, it was only active in CD8 + TEG011. We subsequently explored the concept of additional redirection of CD4 + T cells through co-expression of the human CD8a gene into CD4 + and CD8 + TEG011 cells, later referred as TEG011_CD8a. Adoptive transfer of TEG011_CD8a cells in humanized HLA-A*24:02 transgenic NSG (NSG-A24:02) mice injected with tumor HLA-A*24:02 + cells showed superior tumor control in comparison to TEG011, and to mock control groups. The total percentage of mice with persisting TEG011_CD8a cells, as well as the total number of TEG011_CD8a cells per mice, was significantly improved over time, mainly due to a dominance of CD4 + CD8 + double-positive TEG011_CD8a, which resulted in higher total counts of functional T cells in spleen and bone marrow. We observed that tumor clearance in the bone marrow of TEG011_CD8atreated mice associated with better human T cell infiltration, which was not observed in the TEG011-treated group. Overall, introduction of transgenic human CD8a receptor on TEG011 improves antitumor reactivity against HLA-A*24:02 + tumor cells and further enhances in vivo tumor control.

INTRODUCTION
gdT cells share the properties of both innate and adaptive immunity and play an essential role in cancer immunosurveillance (1,2). Unlike conventional abT cells, gdT cells recognize their cognate antigens in an MHC-unrestricted manner, targeting stress-induced and malignantly transformed self-antigens (3,4). As such, gdT cells represent an attractive cell subset to substantiate T cell-based immunotherapeutic strategies that still mainly focus on abT cells.
Based on their TCRd chain repertoire, two major subsets of gdT cells can be distinguished: Vd2 + and Vd2 − cells. Vd2 + cells mainly reside in the human peripheral blood, representing up to 5% of total circulating T cells, and sense metabolic changes in tumor cells with intracellular accumulation of phosphoantigens (pAgs) level. Vd2 + T cell recognition is facilitated by butyrophilin (BTN) family molecules, including BTN2A1 and BTN3A1 (5)(6)(7)(8)(9)(10). On the other hand, Vd2 − cells mainly localize in mucosal and epithelial tissues, but their antitumor properties are scarcely known (4). Vd2 − cells recognize a broad range of stress-induced ligands, such as the MHC-associated proteins MICA and MICB, foreign lipid antigens presented on CD1c/d molecules in classical HLA-like manner, and CMV-associated UL16-binding protein (ULBP) family members, that are upregulated in stressed or malignant cells (11)(12)(13)(14)(15).
Vd1 + T cells, one of the major Vd2 − subsets, have been shown to exert antitumor reactivity against leukemia and solid tumors (16)(17)(18)(19)(20)(21), indicating their potential in cancer immunotherapy. Adoptive transfer of in vitro expanded Vd2 + cells only showed marginal clinical responses to date (4,22), while adoptive transfer of Vd2 − cells is yet to be tested in the clinic (23). Translational efforts using gdT cells and their receptors outside the context of allogeneic stem cell transplantation (24,25) face substantial hurdles, due to their limited proliferative capacity, underestimated diversity in co-receptors expression and function, as well as scarce information on how gdTCRs interact with their targets.
To bypass these major drawbacks of translating gdT cellsbased immune therapies into clinical practice, we developed the concept of TEGs: abT cells engineered to express a defined gdTCR, allowing the introduction of highly tumor-reactive gdTCR, both Vd2 + (26,27) or Vd2 − (28, 29) subsets, into proliferatively-proficient abT cells (27,30,31). This concept did not only allow to select for highly tumor-reactive gdTCR, but also within the context of Vd2 + TCRs to reprogram both CD4 + and CD8 + abT cells (26,27). Professional help for TCRengineered CD8 + abT cells by also functionally engineering CD4 + abT cells has not only been shown to be important in vitro (32) but also to improve clinical responses (33). Within this context, we previously identified an allo-HLArestricted and tumor-specific Vg5Vd1TCR derived from clone FE11, introduced in the TEG concept as TEG011, which was, although not dependent on a defined peptide, selectively targeting HLA-A*24:02 + tumor cells without impairing the healthy tissues (34). Furthermore, we also highlighted that antitumor reactivity of Vg5Vd1TCR derived from clone FE11 requires CD8a as costimulatory receptor and showed that both CD8aa on the original clone FE11 and CD8ab on transduced abT cells are capable of providing costimulation to the Vg5Vd1TCR derived from clone FE11 (34). Thus, for this very particular Vg5Vd1TCR, the concept of TEGs would not benefit from reprogramming CD4 + abT cells when only a Vg5Vd1TCR is transferred as CD4-transduced TEG011 cells do not elicit antitumor reactivity.
Human CD8 is a membrane glycoprotein classified in an immunoglobulin-like superfamily consisting of hetero-or homodimer of a and b chains, making up for the CD8ab or CD8aa co-receptor on the cell surface. CD8ab predominantly expressed on abT cells, while CD8aa mainly expressed on the cell membrane of innate immune cells, including macrophages, dendritic cells, natural killer (NK) cells, and gdT cells (35). Transfer of CD8 receptor has been reported for abTCR engineered abT cells to functionally reprogram CD4 + abT cells, when low to intermediate affinity abTCRs are used for engineering (36). Within this context, we addressed the implication of CD8aa-dependency of FE11 gdTCR in relation to its tumor immunity. Based on this mechanistic basis of antitumor reactivity for TEG011 cells, we hypothesize that the transfer of CD8a receptor can functionally rescue Vg5Vd1TCR engineered CD4 + abT cells. Within this context, we explored now as additional approach to improve the efficacy of TEG011 therapy, the simultaneously co-expressing Vg5Vd1TCR derived from clone FE11 together with CD8a receptor in a TEG format, referred to as TEG011_CD8a. Importantly, we demonstrate that introduction of transgenic human CD8a co-receptor into CD4 + TEG011 cells successfully enhanced its antitumor efficacy in vitro and in vivo, and thus did not require CD8b. Furthermore, we show that the co-expression of CD8a in CD4 + TEG011 provides additional survival signal and facilitates better T-cell persistence and infiltration in vivo, both of which are essential to sustain long-term tumor control of adoptively transferred TCRbased immunotherapy.

Cell Lines
Daudi, SW480, and Phoenix-Ampho cell lines were obtained from ATCC. K562 with HLA-A*24:02-transduced cell line was kindly provided by Fred Falkenburg (Leiden University Medical Centre, Netherlands) and subsequently transduced with luciferase for in vivo imaging purposes. EBV-LCL was kindly provided by Phil Greenberg (Seattle, WA, USA). Phoenix-Ampho and SW480 cells were cultured in DMEM supplemented with 1% Pen/Strep (Invitrogen) and 10% FCS (Bodinco), whereas all other cell lines in RPMI with 1% Pen/ Strep and 10% FCS. All cell lines were authenticated by short tandem repeat profiling/karyotyping/isoenzyme analysis and were passaged for a maximum of 2 months, after which new cell line stocks were thawed for experimental use. Furthermore, all cell lines were routinely verified by growth rate, morphology, and/or flow cytometry and tested negative for mycoplasma using MycoAlert Mycoplasma Kit (Lonza, Breda, Netherlands). Peripheral blood mononuclear cells (PBMCs) from healthy donors were isolated by Ficoll-Paque (GE Healthcare, Eindhoven, Netherlands) from buffy coats supplied by Sanquin Blood Bank (Amsterdam, Netherlands).

Cloning of TEG011_CD8a and TEGLM1_CD8a
Clone FE11 was generated as previously described (28). FE11 and LM1 [non-functional g9d2TCR with length mutation on the complementary determining region 3 (CDR3) of the d2-chain (31)] gdTCRs were subcloned to pMP71 retroviral vectors containing both gTCR and dTCR chains, separated by a ribosomal skipping T2A sequence. pU57 constructs containing a ribosomal skipping P2A sequence, followed by full-length human CD8a, were purchased from Baseclear (Leiden, Netherlands). Thereafter, CD8a was subcloned into pMP71 vector using XhoI and HindIII restriction sites downstream of g115TCR-T2A-d115_LM1 sequence to generate a TEGLM1_CD8a (Supplementary Table 2) construct that contained NcoI and XhoI restriction sites up-and downstream of LM1 gdTCR chains. NcoI and XhoI restriction sites were then inserted up-and downstream of FE11 gdTCR sequences by sitedirected mutagenesis PCR, after which this sequence was ligated to P2A-CD8a sequence in pMP71 vector using the introduced NcoI and XhoI sites, generating a TEG011_CD8a construct (Supplementary Table 1). Where indicated, CD4 + , CD8 + , CD4 + CD8aa + , and CD4 + CD8ab + TCR-transduced T cells were sorted using a FACSAria II (BD) flow cytometry to >99% purity. Expression levels of CD8a mutants were measured by flow cytometry using anti-CD8a antibody (clones RPA-T8).

Depletion of Non-Engineered T Cells
Non-engineered T cells were depleted as previously described (27). In brief, transduced T cells were incubated with a biotinlabeled anti-abTCR antibody (clone BW242/412; Miltenyi Biotec, Leiden, Netherlands) and then incubated with an antibiotin antibody coupled to magnetic beads (anti-biotin MicroBeads; Miltenyi Biotec), most recently reported to preferentially bind to the bTCR chain (37). Thereafter, the cell suspension was loaded onto an LD column, and abTCR + T cells were depleted by MACS cell separation per the manufacturer's protocol (Miltenyi Biotec). After depletion, TEGs were expanded using a T-cell rapid expansion protocol (REP) (30).

Animal Model
The NOD.Cg-Prkdc scid Il2rg tm1Wjl Tg(HLA-A24)3Dvs/Sz (NSG-A24:02) mice (38) were bred and housed in the breeding unit of the Central Animal Facility of Utrecht University. Experiments were conducted per institutional guidelines after obtaining permission from the local ethical committee, and performed in accordance with the current Dutch laws on animal experimentation. Mice were housed in individually ventilated cage (IVC) system to maintain sterile conditions and fed with sterile food and water. After irradiation, mice were given the antibiotic ciproxin in the sterile water throughout the duration of the experiment. Both male and female mice were randomized with equal distribution among the different groups, based on age and initial weight (measure on Day −1) into 10 mice/group. Adult NSG-A24:02 mice (11-20 weeks old) received sublethal total body irradiation (1,75 Gy) on day −1 followed by intravenous injection of 1×10 5 K562-HLA-A*24:02 luciferase tumor cells on day 0, and received 2 intravenous injections of TEG011, TEG011_CD8a, or TEGLM1_CD8a cells on days 1 and 6 as previously reported (34). Together with the first TEGs injection, all mice received 0,6 × 10 6 IU of IL-2 (Proleukin; Novartis) in 100 μl incomplete Freund's adjuvant (IFA) subcutaneously and subsequently administered every 3 weeks until the end of the experiment. Mice were monitored at least twice a week for any symptoms of disease (sign of paralysis, weakness, and reduced motility), weight loss, and clinical appearance scoring (scoring parameter included hunched appearance, activity, fur texture, and piloerection). The humane endpoint was reached when mice showed the aforementioned symptoms of disease, experienced a 20% weight loss from the initial weight (measured on day −1), developed extramedullary solid tumor masses (if any) reached 2 cm³ in volume, and when clinical appearance score 2 was reached for an individual parameter or a total score of 4.

Assessment for TEGs Persistence
Mouse peripheral blood samples were obtained via cheek vein (max. 50-70 μl/mouse) once a week. Red blood cells were lysed using 1× RBC lysis buffer (Biolegend) and were then stained with a mixture of antibody panels as listed above. The persistence of TEG cells was counted as absolute cell number tumor-reactive TEG cells expressing following cell surface markers huCD45 + gdTCR + CD8 + and huCD45 + gdTCR + CD4 + CD8 + populations or non-reactive TEG cells expressing huCD45 + gdTCR + CD4 + marker observed in mouse peripheral blood using Flow-count Fluorospheres (Beckman Coulter) and measured by flow cytometry.

Preparation of Single-Cell Suspensions
At the end of the study period, bone marrow (mixed from tibia and femur) and spleen sections were isolated and processed into single-cell suspension. Femur and tibia from the hind legs were collected; bone marrow cells were collected by centrifugation of the bones at 10,000 rpm for 15 s and resuspension of the cells in phosphate buffer saline (PBS).
A small section of the spleen was minced and filtered through a 70 μm cell strainer (BD); incubated with 1× RBC lysis buffer cells for maximum 4 min, and subsequently cells were washed and resuspended in PBS.
Absolute cell number of TEG cells were quantified using Flow-count Fluorospheres and measured from a total of 10 6 cells stained for the presence of TEG cells in spleen and bone marrow by flow cytometry analysis (BD LSRFortessa).

Histology Staining and Analysis
Formalin-fixed femur for bone marrow sections were embedded in paraffin and cut into 4 mm sections. Hematoxylin and eosin (H&E) staining was performed for the femur, for bone marrow section. Tissue sections were evaluated to assess for any differences in the presence, distribution, and extension of neoplastic foci indicating tumor tissue. Tissue sections of the femur were evaluated for quantification of tumor tissue by dividing the area covered by the tumor cells by the total area of bone marrow tissue visible in the section using the ImageJ analysis system software (NHI, Bethesda, Maryland, USA) and expressed as a percentage. Images were taken using an Olympus BX45 microscope with the Olympus DP25 camera and analyzed using DP2-BSW (version 2.2) or ImageJ software.

Statistical Analyses
Experimental data were analyzed using GraphPad Prism (GraphPad Software Inc., La Jolla, CA, USA) and shown as mean ± standard deviation (SD) or standard error of mean (SEM) with *P < 0.05; **P < 0.01; ***P < 0.001; and ****P < 0.0001. Statistical significances between groups were assessed using a non-parametric Kruskal-Wallis test, a two-way ANOVA, and a mixed-effects model with repeated measures where indicated.

Co-Transfer of Transgenic CD8a Receptor Is Sufficient to Re-Establish Tumor Reactivity of CD4 + TEG011 Cells
We previously identified an allo-restricted CD8a-dependent Vg5Vd1TCR clone FE11 (28), which showed in vitro antitumor reactivity against HLA-A*24:02-expressing tumor cells (34). We therefore investigated whether introduction of CD8aa or CD8ab along with Vg5Vd1TCR derived from clone FE11 could enhance antitumor reactivity of CD8 + , and also functionally reprogram CD4 + TEG011 cells. Hence, we cotransduced T cells with the FE11 gdTCR, and with either CD8a alone or CD8a together with CD8b ( Figure S1). Subsequently, we sorted separate sets of CD4 + TEG011 cells that co-expressed either exogenous CD8aa (CD4 + CD8a + ) or CD8ab (CD4 + CD8ab + ) as well as TEG011 cells expressing only endogenous CD4 and CD8 as negative and positive controls for tumor recognition, respectively ( Figure 1A). Thereafter, TEG cells were co-cultured with SW480 and EBV-LCL target cells or healthy PBMCs as mock control. Both CD4 + CD8a + and CD4 + CD8ab + TEG011 cells secreted significantly higher levels of IFNg upon exposure to tumor targets than CD4 + TEG011 cells. The acquired antitumor reactivity of CD4 + CD8a + and CD4 + CD8ab + TEG011 cells could be blocked by CD8a and CD8b blocking antibodies ( Figure 1B), confirming the strict dependence of FE11 gdTCR on introduced CD8 molecules. Taken together, we showed that introduction of CD8a alone is sufficient to re-establish antitumor reactivity of CD4 + T cells expressing FE11 gdTCR. Introduction of CD8b did not further enhance tumor recognition but was functionally involved in the molecular interaction with its target when present.
For clinical administration, co-expression of both CD8a and the gdTCR in one vector is preferred to allow reproducible and cost-effective production processes (26,27,39). Moreover, coexpressing both CD8a and the gdTCR in one vector can also overcome the difference in transduction efficiency when they were transduced separately. Therefore, we generated new retroviral constructs carrying either FE11 gdTCR or a nonfunctional length mutant clone LM1 gdTCR [ (31); served as mock control] followed by full-length human CD8a receptor sequences (TEG011_CD8a and TEGLM1_CD8a, Figure 1C). The complete sequence of transgenes for these retroviral constructs is listed in Supplementary Tables 1, 2, respectively. Subsequently, abT cells were transduced with either FE11 gdTCR without human CD8a receptor (TEG011), FE11 gdTCR with human CD8a receptor (TEG011_CD8a), or LM1 gdTCR with human CD8a receptor (TEGLM1_CD8a).
After TEG expansion, we performed magnetic selection of CD4 + T cells for each TEG constructs. To elucidate whether introduction of transgenic CD8a receptor adequately rescues TEG011 reactivity of non-tumor reactive CD4-transduced cells once delivered by the very same vector, we co-cultured tumor target HLA-A*24:02-transduced CML tumor cells (K562), SW480, and EBV-LVL cells with either CD4 + TEG011_CD8a, CD4 + TEGLM1_CD8a, or CD4 + TEG011 (without introduction of the CD8a receptor). Healthy T cells and TEG011 bulk cells (with CD4:CD8 1:1 ratio) were used as the untransformed mock target and positive effector control, respectively ( Figure 1D). CD4 + TEG011_CD8a cells produced a significantly higher IFNg level compared to CD4 + TEG011, which was equivalent to those of TEG011 bulk cells against all tumor targets, without affecting C D B A FIGURE 1 | Introduction of transgenic CD8a receptor on TEG011 improves T cell activation. (A) TEG011 were retrovirally transduced with either CD8a alone or CD8a in combination with CD8b. CD4 + , CD8 + , CD4 + CD8a + , and CD4 + CD8ab + subsets of T cells were subsequently sorted (left panel is a representative sorting plot for CD4 + , CD8 + , and CD4 + CD8a + cells; CD4 + CD8ab + cells were sorted in a similar manner) and tested for recognition of SW480 and EBV-LCL target cells by IFNg ELISPOT (right panel). Healthy PBMCs were included as untransformed mock control target cells. Data are of representative of four independent experiments, and error bars represent mean ± SEM (**P < 0.01; ***P < 0.001) calculated by two-way ANOVA. (B) CD8a and CD8b blocking on CD4 + T cells were transduced with the FE11 gdTCR and CD8a alone, or CD8a with CD8b. TEG011 was co-incubated with SW480 target cells in the presence of a control antibody, or CD8a or CD8b blocking antibodies. IFNg production was measured by ELISPOT. Data represent mean ± SD of replicates for each effector (**P < 0.01; ***P < 0.001; ****P < 0.0001) calculated by two-way ANOVA. (C) Schematic diagram of pMP71 retroviral vector constructs containing codon-optimized human gdTCR sequences from either clone FE11 (referred as TEG011_CD8a) or non-functional LM1 chains (referred as TEGLM1_CD8a) in combination with full length of human CD8a receptor (top panel). Within the transgene cassettes, individual gTCR and dTCR chains have been linked with a self-cleaving thosea asigna virus 2A (T2A; black box) ribosomal skipping sequence, while the CD8a sequence was connected with a porcine teschovirus-1-derived 2A (P2A; gray box) ribosomal skipping sequence. (D) CD4 + abT cells were transduced with either TEGLM1_CD8a, TEG011, or TEG011_CD8a gdTCR (as effector cells) and subsequently co-cultured with HLA-A*24:02-expressing target cell lines or healthy T cells (E:T ratio is 1:3) for 18-24 h. TEG011 bulk population with 50:50 ratio of both CD4 + and CD8 + TEGs and T cells from healthy donor were used as positive and untransformed mock controls, respectively. Antitumor reactivity was measured by IFNg ELISPOT, where 50 spots/15,000 cells were considered as a positive antitumor response and indicated by the dashed horizontal line. Data are representative of three independent experiments with replicates for each target, and error bars represent mean ± SD (*P < 0.05; **P < 0.01; ****P < 0.0001) calculated by two-way ANOVA. healthy cells. The equivalent IFNg level between CD4 + TEG011_CD8a and TEG011 bulk cells comprised of only 50% CD8 + TEG011 implied that reprogrammed CD4 + TEG011_CD8a are surprisingly poorer cytokine secretors. Importantly, enhanced tumor recognition was restricted to CD4 + TEG011_CD8a cells and not CD4 + TEGLM1_CD8a mock cells, highlighting the specific role of CD8a as costimulation for the introduced FE11 gdTCR. We concluded that introduction of transgenic CD8a receptor in combination with Vg5Vd1TCR derived from clone FE11 allowed reprogramming of CD4 + T cells towards HLA-A*24:02expressing tumor cells in vitro, though activity was lower when compared to CD8 + TEG011.

TEG011_CD8a Improves In Vivo Tumor Control and Associates With Higher Persistence of Functional T Cells
In previous studies, we have shown TEG011 efficacy against HLA-A*24:02-expressing tumor cells in vitro and an extended in vivo safety profile, as well as peripheral persistence of TEG011, where long-term persistence of TEG associated with reduced probability for developing extramedullary solid tumor masses in vivo (34,40). To assess the consequence of the additional expression of TEG011_CD8a, NSG transgenic mice expressing human HLA-A*24:02 (NSG-A24:02) were irradiated, received luciferase-labeled K562 HLA-A*24:02 + cells, and subsequently received two intravenous injections of either mock control TEGLM1_CD8a, TEG011_CD8a, or TEG011 cells. All infused TEG variants showed comparable gdTCR expression, where the transduced abT cells expressed Vd1 + TCR for TEG011 and TEG011_CD8a ( Figure S2). Mice were monitored for tumor burden assessed by bioluminescent imaging, T cell persistence and infiltration, as well as any other signs of discomfort. Mice were sacrificed when the humane endpoints were reached (experimental outline Figure 2A). TEG011_CD8a-treated mice had a significantly lower tumor burden over time compared to the mock control TEGLM1_CD8a and TEG011-treated groups ( Figure 2B), indicating superior tumor control in vivo by TEG011_CD8a. All tumor-bearing mice eventually developed tumor, and measurement of individual mouse indicating tumor growth over time for each treatment group is shown in Figures 2C, D. Despite the significant in vivo tumor control, we observed only a trend towards an improved overall survival for TEG011_CD8a-treated mice ( Figure S3). This could be due to limited treatment window of this mouse model contributed by aggressive tumor growth of K562 HLA-A*24:02-transduced cells.
As TEG011 cells carry CD8a-dependent Vg5Vd1TCR, we focused our in vivo analysis to tumor-reactive CD8-expressing TEG cells (as validated by in vitro functional T cell assay in Figure 1D) while taking into account the non-tumor reactive CD4 + TEG cells. Therefore, we assessed CD8-expressing TEG cell product properties and persistence by measuring viable huCD45 + gdTCR + CD8 + singlepositive and huCD45 + gdTCR + CD4 + CD8 + double-positive cells (present in mock control TEGLM1_CD8a and TEG011_CD8a only) in mouse peripheral blood using flow cytometry (gating strategy depicted in Figure S4). TEG cells persisted up to 4 weeks after infusion in the mouse peripheral blood with biological variations between mice ( Figure 3A). To address this interindividual variation in T-cell persistence, we analyzed separately the percentage of mice where CD4 + and CD8 + T cells reached at least 500 cells/ml in the peripheral blood over time, a threshold described previously (41) ( Figure S5A). We observed a higher percentage of mice with persisting CD4 + and CD8 + T cells in TEG011_CD8a group when compared to mock TEGLM1_CD8a and TEG011 group. Despite some imbalance in the CD4:CD8 ratio with lower numbers for CD8 + TEG011 infused ( Figure S2), more CD8 + TEG011 persisted over time when compared to CD8 + single-positive TEG011_CD8a. Vice versa, endogenous CD4 T cells for TEG011_CD8a were lower before infusion when compared to TEG011 prior to infusion, while CD4 + CD8 + double-positive TEG011_CD8a were higher in numbers over time when compared to both CD4 + CD8 + doublepositive TEGLM1_CD8a and CD4 + TEG011 cells ( Figure S5B). As a net effect, we observed more CD8-expressing T cells for TEG011_CD8a cells when compared to TEG011 ( Figure 3B). Next, we investigated the expression of PD1 and TIM3 on CD8 + single-positive cells and CD4 + single-positive or CD4 + CD8 + doublepositive cells. Higher numbers of T cells expressing PD1 or TIM3 were observed on TEG011_CD8a cells, as compared to mock TEGLM1_CD8a and TEG011 cells (Figures S6A, B). CD8 + single-positive TEG011 and TEG011_CD8a showed an increased PD1 expression when compared to CD8 + single-positive TEG_LM1 ( Figure S6A). A partial decline of TIM3 expression was most pronounced over time in CD8 + single-positive TEG011_CD8a ( Figure S6B).
Next, we investigated infiltration of TEG cells into spleen and bone marrow on weeks 1 and 2 after infusion. Specifically, we compared the TEG011 and TEG011_CD8a groups to elucidate the contribution of transgenic CD8a co-expression in TEG011 infiltration in vivo, and focused on the total sum of CD8expressing TEG011 cells. We detected a significantly higher number of CD8-expressing TEG cells infiltrating in the spleen and bone marrow of TEG011_CD8a-treated mice at both time points ( Figure 3B). Importantly, we did not observe rapid clearance of CD4 + CD8 + double-positive TEG011_CD8a cells in these tissues within these time points, whereas CD8 + single-positive TEG011 cells were barely detected. Thus, we conclude that CD8a costimulation with TEG011 improves overall in vivo tumor control, T cell persistence, and infiltration of CD8-expressing TEG011 cells.

TEG011_CD8a Enhanced T Cell Infiltration and Effectively Cleared Tumor Cells in Bone Marrow
We previously reported an extensive in vivo safety profile of TEG011 against healthy tissues that express HLA-A*24:02 molecules, in which no significant histological lesions were observed in major organs, including liver, spleen, and intestine (40). For histopathology analysis, we collected a femur bone marrow section from each treatment group at the end of the study period to further evaluate antitumor efficacy of the new TEG011_CD8a cells ( Figure 4A). Tissue sections were assessed for the presence and extension of the neoplastic foci composed by round, large, undifferentiated tumor cells. The mock control TEGLM1_CD8a-treated group showed evident 19,2% neoplastic infiltration, whereas the TEG011-treated group showed up to 3,4% neoplastic infiltration of a homogeneous population of neoplastic cells in the bone marrow. Interestingly, we did not observe any neoplastic infiltration in the bone marrow of mice in the TEG011_CD8a group, and the appearance of bone marrow cell composition and cellularity was normal ( Figure 4B). In conclusion, although the number of analyzed bone marrows was limited, our data imply that TEG011_CD8a effectively cleared tumor cells in bone marrow, emphasizing the role of CD8a costimulation for better in vivo tumor control of TEG011 cells. Overall, our data indicate that introduction of transgenic CD8a on TEG011 cells effectively improves in vivo tumor control and better T cell infiltration into bone marrow. . Statistical significances were calculated by a mixed-effects model with repeated measure up to week 3 as comparison all treatment group (indicated next to legends) and only between TEG011 and TEG011_CD8a group for week 4 (indicated on the graph); (*P < 0.05; **P < 0.01). (C) Tumor burden for individual mouse for each treatment group measured by integrated signal density per total surface area (count/mm 2 ) using BLI. (D) Tumor load for individual mouse was evaluated by bioluminescence imaging on week 1 to week 4 using Milabs Optical Imaging (OI) Acquisition and OI-Post processing software (version 2.0). Anesthetized mice were injected intraperitoneally with 25 mg/ml Beetle-luciferin (Promega). Calibrated units were calculated from integrated density of bioluminescence signal (electron/s) as shown by the right bar. The animals were imaged 10 min after luciferin injection. Black areas indicate loss of mice.

DISCUSSION
TEG011 has been reported to specifically recognize HLA-A*24:02 + malignant cells while sparing the HLA-A*24:02expressing healthy tissues with the requirement of CD8a costimulation (34,40). While TEG011 has shown a favorable efficacy profile in vivo, we only observed in approximately 50% of the mice long-term persistence of CD8 + TEG011 cells, which could be due to the lack of support by antigen-specific CD4 + T cells (29,40). The presence of both tumor-specific CD4 + and CD8 + abT cells has been reported to significantly improve clinical responses compared to tumor-specific CD8 + abT cells alone (33). To further improve the antitumor efficacy of TEG011, we co-expressed a CD8a co-receptor together with the Vg5Vd1TCR derived from clone FE11 in TEG format, referred to as TEG011_CD8a cells. Introduction of CD8a receptor is particularly beneficial for TEG011 as this particular gdTCR requires the presence of CD8a as co-receptor for their antitumor reactivity, as we published previously (34,40). CD8a expression has been reported as common feature of gdTCR after CMV infection (28). These insights imply that also other Vd1TCR might functionally depend on CD8a, which we could, however, not investigate in a broader context. Thus, when exploring tumor reactivity with selected Vd1TCR for the development of gdT cell-based immunotherapies (20), the absence of functional reactivity by an introduced Vd1TCR might not necessarily reflect the absence of binding of the Vd1TCR to its target but rather the lack of a co-stimulation through, e.g., CD8a or other co-stimulatory molecules. In this study, we reported on the capacity of the introduced CD8a co-receptor to successfully redirect non-tumor reactive CD4 + TEG011 cells in vivo and in vitro against tumor targets that express HLA-A*24:02 molecules. We now report on more than 80% of mice showing persistence of CD8-expressing T cells after 4 weeks. TEG011_CD8a cells showed also in absolute numbers higher T cell counts and stable peripheral persistence in vivo, which was, however, mainly a consequence of the persistence of CD4 + CD8 + double-positive TEG011_CD8a and not an improved persistence of CD8 + single-positive TEG011_CD8a. This finding supports the notions that co-expression of CD4 + and CD8 + T cells provides an additional survival signal for TEG011 cells. This observation is in line with clinical studies for CD19 CAR T cells that reported that a mixture of both CD4 + and CD8 + T cells with 1:1 ratio improved tumor remission in B-ALL patients (42,43). Regardless of the precise underlying molecular mechanism, for the first time we observed tumor clearance in the bone marrow by TEG011_CD8a, but not by TEG011 alone.
Using humanized transgenic mice expressing human HLA-A*24:02, we could study the implication of CD8a introduction to TEG011, referred to as TEG011_CD8a, elucidating their improved efficacy in vivo. We provide evidence that TEG011_CD8a effectively cleared tumor cells in bone marrow and elicited better tumor control against human HLA-A*24:02expressing tumor cells. We cannot entirely exclude that superior tumor control in TEG011_CD8a may have been caused initially by more CD8 single-positive cells in the TEG011_CD8a product compared to TEG011 product, as CD4 + /CD8 + ratios could not be entirely controlled in the experimental setup prior to infusion. However, our mouse model also allowed us to investigate TEG011_CD8a kinetics in the presence of tumor cells; and we observed sustained long-term TEG persistence mainly for gdTCR + CD4 + CD8 + double-positive and a decline in gdTCR + CD8 + single-positive TEG011_CD8a cells. Importantly, the sustained peripheral TEG persistence was only observed for TEG011_CD8a but not TEGLM1_CD8a, highlighting the key role of a functional tumor-reactive gdTCR. This observation rather argues against the classical helper function of gdTCR + CD4 + CD8 + double-positive TEG011_CD8a cells within the context of TEG011_CD8a. Hence, the concurrent expression of CD4 + and CD8 + co-receptor most likely provided additional survival signal for tumor-specific CD4 + T cells, which did not, however, translate into classical helper functions towards CD8 + T cells (40,44,45). CD4 + T cells have been reported to avoid expression of inhibitory receptors on CD8 + T cells (46) and as an important cell subset to induce memory T cell formation (47). Along this line we observed over time reduced expression of TIM3 in CD8 + single-positive TEG011_CD8a cells compared to mock and TEG011 group. CD4 + CD8 + double-positive TEG011_CD8a cells had lower levels of TIM3 when compared to CD8 + single-positive TEG011_CD8a cells. These data remain difficult to interpret, and most likely simply reflect different regulation and activation of non-tumor reactive CD4 + and tumor-reactive CD8 + TEG011 cells, respectively. We also acknowledge that xenograft mouse models do not allow to completely mimic all potential helper roles of human CD4 + T cells, due to the lack of human professional antigen-presenting cells.
Reprogramming CD4 + T cells by genetic engineering has been reported to clinically impact efficacy and toxicity by high affinity receptors, like CARs (48). Vg9Vd2TCR (30) and CD8abindependent abTCRs (32) have been also reported to reprogram CD4 + T cells, which not only have the ability to exert tumor cell killing but also induce maturation of professional antigenpresenting cells. Transfer of CD8ab in combination with intermediate affinity tumor reactive abTCR has been reported to support tumor control in vitro and in vivo (49,50), and for high affinity abTCR with artificial signaling domains adding CD8a alone has been shown to reprogram CD4 + T cells (36). Within this context, our data show that CD8aa in combination with a natural gdTCR serves as costimulatory receptor, as opposed to the well-described inhibitory function of CD8aa on abT cells within the context of a natural abTCR. Expression of that CD8aa on activated CD4 + and CD8ab + abT cells has been reported to act as corepressor by competing with CD8ab + cells for p56 lck signaling molecule (51). Though we investigated the role of CD8aa in the TEG concept, our data support the notion that CD8aa in combination with a gdTCR is synergistic on natural gdT cells, as activated CD8aa + gdT cells were reported in supporting control of HIV infection (52). We have also previously reported significant increases in circulating CD8aa + gdT cells in CMV-positive population (28). Thus, CD8aa appears to have opposing functions on innate and adaptive immune cells, where it acts as costimulatory receptor in the context of a gdTCR.
The precise molecular interaction between CD8aa and its specific ligand in our context remains yet to be unraveled. The CD8aa receptor has been shown to bind to MHC Class I molecules, including HLA-A*02:01, HLA-A*11:01, HLA-B*35:01, HLA-C*07:02, via protruding a3 domain loop of MHC molecules with lower affinity than the binding of a TCR-pMHC complex (53)(54)(55)(56). Polymorphisms in the MHC a3 domain contributes to a binding variation of CD8aa to different HLA molecules, such as HLA-A*24:02. In this context, HLA-A*24:02 is one of the possible ligands for CD8aa on TEG011, in line with an earlier study that reported CD8aa interaction with HLA-A*24:02 in a similar way with HLA-A*02:01, involving binding to the a2 and a3 domains, as well as to the b2m domain of pMHC complex, but with different conformation that suggests CD8aa plasticity (57). The nonclassical MHC molecules are also reported to interact with CD8a, such as HLA-G and HLA-E (58). HLA-G is a known ligand for CD8aa, which is expressed on some colorectal cancer (59)(60)(61), while HLA-E is mainly expressed in human endothelial cells and is highly expressed in tumor cells (58). Other studies also demonstrated the interaction between CD8 and CEACAM5, which support the possibility of CEACAM5 as CD8a ligands (62).
Overall, we demonstrate that TEG011 equipped with human CD8a coreceptor elicits superior tumor control and long-term persistence, which mainly impacted numbers of gdTCR + CD4 + CD8 + double-positive TEG011_CD8a cells, and associated with better T-cell infiltration. In addition, TEG011_CD8a cells successfully cleared tumor cells in the bone marrow. In contrast to currently emerging immunotherapy approach using CAR T cells, our strategy allows tumor-specific targeting of HLA-A*24:02positive cancer patients, irrespective of antigen-specific expression on cell surface and the type of cancer, and thus TEG011_CD8a therapy has broader applicability towards a substantial amount of cancer patients with HLA-A*24:02positive haplotype highlighting its therapeutic potential for further clinical application.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The animal study was reviewed and approved by Utrecht Animal Welfare Body (IvD) and Central Authority for Scientific Procedures on Animals (CCD). Written informed consent was obtained from the owners for the participation of their animals in this study.

AUTHOR CONTRIBUTIONS
IJ, TS, ZS, and JK conceptualized, designed, and developed the in vivo models. IJ, PH, WS, and SH performed the in vitro and in vivo experiments. LB and AB performed the histopathology examination of the mouse tissues. DB and RO contributed vital components. IJ analyzed all in vitro and in vivo data and was a major contributor in writing the manuscript. IJ, ZS, and JK interpreted all in vitro and in vivo data. IJ and JK wrote the manuscript. All authors read, reviewed, and approved the final manuscript.

FUNDING
Funding for this study was provided by ZonMW 43400003 and VIDI-ZonMW 917.11.337, KWF 6426, 6790 and 7601 to JK; 12586 to TS and JK; 11393 and 13043 to ZS and JK; 11979 to JK and DB.