A CD8αβ co-receptor modified to contain an intracellular CD28 signaling tail enhances TCR-engineered T cell function independent of solid-tumor-associated co-stimulatory ligands

Shihong Zhang; Tzu-Hao Tang; Sinéad Kinsella; Francesco Mazziotta; Michael T Schweizer; Megan S McAfee; Ariunaa Munkhbat; Yapeng Su; Valentin Voillet; Lauren E Martin; Colton W Smith; Yuta Asano; Menna Hailemariam; Jakob Bakhtiari; Bo Lee; Cecilia Yeung; Hung Chen; Alessandro M Rizzi; Daniel G Chen; Kelsey Furiya; Nick Horst; Tianzi Zhang; Phung Le; Kelly McKenna; Shannon K Oda; Anthony Rongvaux; Phillip D Greenberg; Thomas M Schmitt; Aude G Chapuis

doi:10.1038/s41467-025-67446-5

. 2026 Jan 3;17:753. doi: 10.1038/s41467-025-67446-5

A CD8αβ co-receptor modified to contain an intracellular CD28 signaling tail enhances TCR-engineered T cell function independent of solid-tumor-associated co-stimulatory ligands

Shihong Zhang ^1,^2,^#, Tzu-Hao Tang ^1,^#, Sinéad Kinsella ^1,^#, Francesco Mazziotta ¹, Michael T Schweizer ^3,⁴, Megan S McAfee ^1,⁵, Ariunaa Munkhbat ^1,⁶, Yapeng Su ^1,⁷, Valentin Voillet ⁸, Lauren E Martin ¹, Colton W Smith ¹, Yuta Asano ^1,⁹, Menna Hailemariam ¹, Jakob Bakhtiari ¹, Bo Lee ¹, Cecilia Yeung ¹, Hung Chen ^1,¹⁰, Alessandro M Rizzi ¹, Daniel G Chen ¹, Kelsey Furiya ¹, Nick Horst ^1,¹¹, Tianzi Zhang ¹, Phung Le ^1,¹², Kelly McKenna ^1,^13,¹⁴, Shannon K Oda ¹⁵, Anthony Rongvaux ^1,¹⁶, Phillip D Greenberg ^1,^3,¹⁶, Thomas M Schmitt ¹, Aude G Chapuis ^1,^3,^✉

PMCID: PMC12820041 PMID: 41484084

Abstract

Adoptive transfer of T cells engineered with tumor-specific T cell receptors (TCRs) has shown limited efficacy in solid tumors, hindered by insufficient persistence, tumor trafficking, and dependence on tumor-associated co-stimulatory ligands. In a phase I trial (NCT 04639245) for patients with metastatic MAGE-A1-expressing tumors and adequate organ function; one participant received treatment, which was well-tolerated. In this case and NSG murine models, infusion of CD4/CD8 T cells co-expressing a class-I MAGE-A1-specific TCR and CD8αβ, failed to control tumor progression. To enhance function downstream of TCR signaling, here we investigate the adaptability of TCR components to synthetic modification. Leveraging the obligate co-expression of CD8αβ required for class-I TCR function in CD4 T cells, we identify CD8β as a tractable site for engineering without loss of function. In vitro screening demonstrates incorporation of the CD28 intracellular tail, yielding a CD8/CD28 chimeric co-receptor, most effectively enhances cytokine production, T cell persistence, and tumor control in immunodeficient murine models while preserving stem-like transcriptional features compared to native CD8β. Further rational modification of the CD28 binding motifs improves tumor control in vivo with increased intratumoral accumulation and reduced exhaustion. This benefit also extends to PRAME and WT1-specific TCRs in vitro supporting generalizability.

Subject terms: Tumour immunology, Cancer immunotherapy, Cancer therapy, Immunotherapy

TCR-engineered T cells have shown limited efficacy in part due to the absence of co-stimulation leading to limited accumulation in solid tumors. The authors here show engineering the CD8β coreceptor with an intracellular CD28 domain enhances cytokine production, persistence, and tumor control in vivo independent of tumor-associated co-stimulatory ligand encounter.

Introduction

Utilizing T cell receptors (TCRs) transduced into patient T cells is an attractive therapeutic approach as it enables immunorecognition of tumors, with high sensitivity for low peptide/HLA densities of a diverse range of intracellularly expressed antigens¹. However, TCR-transduced T cells have shown modest efficacy against non-hematologic malignancies with sporadic anti-tumor effects^2,3. Larger clinical trials have yielded overall response rates of 44% in rare indications, such as synovial sarcoma, and as low as 9% in other solid tumors⁴. Multiple barriers have been shown to contribute to this lack of efficacy: in addition to tumor heterogeneity limiting antigenic encounter⁵ and the presence of immunosuppressive populations⁶ the solid tumor microenvironment (TME) is typically rich in inhibitory cytokines⁷, suppressive checkpoint ligands⁸ and the general lack of positive co-stimulatory signals required for robust T cell activation⁸. Together, these obstacles underscore the imminent need to improve T cell trafficking and function within the solid tumor TME.

TCR-based therapies have primarily exploited CD8 T cells with cytotoxic activity that recognize tumor antigens presented by HLA class I¹. However, the enhanced antitumor activity provided by concurrent CD4 T cell responses is well documented in murine models⁹ and upon co-transfer of tumor-reactive CD4 T cells with CD19-targeting chimeric antigen receptor-engineered (CAR) T cells¹⁰. Although most class I-restricted TCRs require concurrent CD8 co-receptor binding for optimal target avidity and downstream signaling¹¹, antigen-specific recognition can also be conveyed to CD4 T cells by co-expressing the class I-restricted TCRs with the CD8αβ co-receptor¹². This strategy is being explored in clinical trials to promote increased cytotoxic function and ultimately anti-tumor activity¹³.

Beyond functionally engaging CD4 T cells, effective T cell activation and associated proliferation typically also requires positive co-stimulation^8,14. As an immune evasion tactic, tumor cells commonly downregulate positive and upregulate inhibitory co-stimulatory ligands that compromise T cell activation¹⁵. As a result, antigen-specific T cells are deleted and/or become dysfunctional after encountering tumors, resulting in failure to sustain tumor control¹⁶. Unlike CAR constructs that integrate a co-stimulatory domain, TCRs require independent co-stimulatory binding for co-receptor triggering¹⁷. Strategies employing fusion proteins combining an inhibitory ectodomain with a costimulatory endodomain are being explored^18,19, however, these strategies depend on sufficient ligand density within the TME²⁰.

Here, we engineered a synthetic co-stimulatory signal directly into the class-I TCR complex by modifying the CD8β chain to include a CD28 intracellular domain, enabling ligand-independent co-stimulation with TCR engagement. This design enhances cytokine production, persistence, and intra-tumoral accumulation of both CD4 and CD8 T cells while preserving stem-like features and reducing exhaustion. Further modification of CD28 signaling motifs yields an optimized variant with increased in vivo anti-tumor activity. Together, these findings establish a generalizable framework to strengthen TCR-engineered T cell function, integrating synthetic co-stimulation into diverse TCR-based therapies.

Results

Functional CD4 T_TCR-MA1 cells enhance MAGE A1-specific CD8 T_TCR-MA1 cells

Five TCR clonotypes targeting the HLA-A2-restricted MAGE-A1_278-286 (KVLEYVIKV) epitope²¹, were identified (TCR_MA11-5) following in vitro stimulation and isolation using RACE-PCR²², and synthesized as codon-optimized P2A-linked constructs in a lentiviral backbone previously used for clinical transduction²³ (see Methods). CD8 T cells were transduced with each TCR and demonstrated peptide/HLA multimer (p/HLA) binding with similar mean fluorescence intensities (MFIs) (range 2672-3059) (Supplementary Fig. 1A) and produced TNFα and IFNγ when exposed to supraphysiologic levels of cognate peptide (Supplementary Fig. 1B). Concurrently transduced CD4 T cells had lower p/HLA binding (MFIs 960-2131) and limited cytokine production, suggesting reduced p/HLA binding without the CD8 co-receptor and functional TCR engagement²⁴. TCR_MA15 had a similar functional avidity towards the wild-type peptide compared to TCR_MA12, 3 and 4 (range 48.1-85.8 nM) (Supplementary Fig. 1C), and also recognized the previously described K278T mutated MAGE-A1_278-286 peptide (Supplementary Fig. 1D)²⁵. Based on these characteristics, TCR_MA15 was selected for future studies.

To assess the impact of the CD8αβ interaction with HLA Class I in TCR-transduced T cells, transduction with TCR_MA15 alone (T_TCR-MA1) or co-transduction with CD8αβ (T_{TCR-MA1-CD8αβ}) using a single multi-cistronic vector construct (Fig. 1A) produced CD4 T_TCR-MA1, CD4 T_{TCR-MA1-CD8αβ} and CD8 T_{TCR-MA1-CD8αβ} (CD8 T_TCR-MA1 were not generated as addition of CD8αβ to CD8 T cells does not impact function¹²). Compared with CD8 T cells, CD4 T_{TCR-MA1-CD8αβ} in part rescued the ability of CD4 T_TCR-MA1 to produce IFNγ, TNFα, and IL-2 upon cognate peptide exposure (mean EC50 of 1119.5, 807.25, and 1464.5, respectively, for CD4 T_TCR-MA1 compared with 20.65, 9.91, and 43.52 for CD4 T_{TCR-MA1-CD8αβ}) (Fig. 1B).

Fig. 1 — A Schematic of the multi-cistronic vector design used in pre-clinical studies. B Dose response curves (left panels) of IFNγ, TNFα, and IL-2 expression in response to decreasing peptide concentrations by CD8 T_{TCR-MA1-CD8αβ} (red), CD4 T_TCR-MA1 (blue, open circles) and CD4 T_{TCR-MA1-CD8αβ} (blue, full circles). Mean EC₅₀ (right panels) of IFNγ, TNFα, and IL-2 cytokine expression for the same previous 3 cell populations. C Growth kinetics (left panel) for the ME275 cell line in a live tumor visualization assay (Incucyte S3) in the absence (gray line) or presence of CD4 T_{TCR-MA1-CD8αβ} (blue line), CD8 T_{TCR-MA1-CD8αβ} (red line), or CD4 + CD8 T_{TCR-MA1-CD8αβ} (purple line). The E:T ratio was 2:1, and arrows indicate the addition of tumor cells to the culture. Final tumor integrated intensity (right panel) for the same experiment at 144 and 192 h. The mean ± standard deviation (SD) of triplicate wells is shown. p value determined by a one-way ANOVA with Tukey’s multiple comparison test. (n = 3 wells/group). D Counts of survived T cells were measured at 144 h post tumor simulation from Incucyte S3 live tumor assay **(C)**. Change in absolute numbers of CD4 T_{TCR-MA1-CD8αβ} (left panel) and CD8 T_{TCR-MA1-CD8αβ} (right panel) in either alone (CD4 - blue bars, CD8 - red bars) or combined (purple bars) conditions compared to seeded cells. The mean ± SD from triplicated wells is shown (n = 3 wells/group). p value determined by unpaired two-tailed t-test. E Schematic (left) and representative growth kinetics (right) of 1 × 10⁶ ME275 tumors engrafted subcutaneously into NSG mice with subsequent transfer of 1 × 10⁷ equal numbers of total irrelevant T cells (T_Irr.-CD8αβ), CD4 T_{TCR-MA1-CD8αβ} alone, CD8 T_{TCR-MA1-CD8αβ} alone, or a 1:1 ratio of combined CD4 and CD8 T_{TCR-MA1-CD8αβ}, 22 days after engraftment. The tumors were allowed to reach ~60mm³ in size before T cell infusion (n = 3 mice/group, 2 tumors/mouse). F Schematic (left) and representative growth kinetics (right) of 1 × 10⁵ A375F tumors engrafted into NSG mice with subsequent transfer of 1 × 10⁷ equal numbers of total T cells (T_Irr.-CD8αβ, CD4 T_{TCR-MA1-CD8αβ} alone, CD8 T_{TCR-MA1-CD8αβ} alone, or a 1:1 ratio of combined CD4 and CD8 T_{TCR-MA1-CD8αβ}) 12 days after engraftment. The tumors were allowed to reach ~60mm³ in size before T cell infusion (n = 3–4 mice/group, 2 tumors/mouse). G A375F tumor volume from (F) assessed at sacrifice 8 days after T cell infusions for each indicated condition. p value determined by one-way ANOVA with Tukey’s multiple comparison test. (n = 4 mice for T_Irr.-CD8αβ group and n = 3 mice for CD4 T_{TCR-MA1-CD8αβ} alone, CD8 T_{TCR-MA1-CD8αβ} alone, or a 1:1 ratio of combined CD4 and CD8 T_{TCR-MA1-CD8αβ} groups, 2 tumors/mouse averaged). All data are presented as mean ± SEM. Panels E, F created in BioRender. Tang, A. (2026) https://BioRender.com/6wg9k7f.

To determine the impact of CD4 and CD8 T_{TCR-MA1-CD8αβ} both alone and combined, in vitro survival of the MAGE-A1 and HLA-A*0201-expressing ME275 melanoma cell line was assessed in a real-time killing assay where T cells were repeatedly challenged with tumor (Fig. 1C left panel, Supplementary Table 1 and Supplementary Fig. 2). Although CD8 T_{TCR-MA1-CD8αβ} remained superior at repeated tumor killing compared to CD4⁺ T_{TCR-MA1-CD8αβ}, the 1:1 combination of CD4 and CD8 T_{TCR-MA1-CD8αβ} (with each subset contributing half of the total cell number) significantly outperformed an equal total number of CD8-only or CD4-only T_{TCR-MA1-CD8αβ} at 144 h and 192 h (Fig. 1C, right panel). At 144 h, CD4 and CD8 T_{TCR-MA1-CD8αβ} remaining in wells within the combination group were ~30x more abundant whereas CD4 and CD8 T_{TCR-MA1-CD8αβ} alone were ~3x more abundant (Fig. 1D). CD4 and CD8 T_{TCR-MA1-CD8αβ} combination group continued to control tumor for 13 days (Supplementary Fig. 3A).

To assess the proliferative capacity, CD4 and CD8 T_{TCR-MA1-CD8αβ} were exposed once to ME275 and their divisions assessed 7 days later (Supplementary Fig. 3B–G). CD8 T_{TCR-MA1-CD8αβ} divided more in the presence of CD4 T_{TCR-MA1-CD8αβ} (Supplementary Fig. 3B–D), while CD4 T_{TCR-MA1-CD8αβ} divided less when combined with CD8 T_{TCR-MA1-CD8αβ} (Supplementary Fig. 3E–G). Despite the increased ability of TCR-transgenic CD4 T cells co-transduced with CD8αβ to secrete helper cytokines when exposed to supraphysiological levels of peptide¹² and proliferate after tumor exposure (Supplementary Fig. 3E), they showed reduced tumor killing upon repeated challenge (Fig. 1C). In contrast, CD8 T cells divided less after single tumor exposure but were more effective at repeated killing (Supplementary Fig. 3B, 3E, Fig. 1C). Together, CD4 and CD8 T cells sustained tumor control (Supplementary Fig. 3A), likely through CD4-derived cytokine support that enhances CD8 T cells proliferation and cytotoxicity. The reduced proliferation of CD4 T cells in co-culture (Supplementary Fig. 3F), may reflect rapid tumor clearance by CD8 T cells and/or sequestration of CD4 T cell -derived cytokines.

To examine this in vivo, NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice were engrafted with 1 × 10⁶ ME275 cells (Fig. 1E). CD4, CD8, or combined CD4/CD8 T_{TCR-MA1-CD8αβ} (1:1 ratio, 1 × 10⁷ total) mediated tumor control, whereas an irrelevant TCR (T_{TCR-Irr-CD8αβ}) did not, confirming specificity. We next engineered A375 melanoma cells to express both MAGE-A1 and HLA-A*0201 (A375F) (Supplementary Table 1 and Supplementary Fig. 2), which reached ~60 mm³ by 13 days post engraftment (Fig. 1F). In this model, CD4⁺ or CD8⁺ T_{TCR-MA1-CD8αβ} alone slowed but did not prevent tumor growth while the CD4/CD8 combination (1:1 ratio, 1 × 10⁷ total) had significantly reduced tumor volumes compared to controls (Fig. 1F and G), supporting advancement of the combined approach to clinical translation.

CD4 and CD8 T_{TCR-MA1-CD8αβ} are safely infused in one patient, require potency enhancement in stringent murine models

Thirty-two patients were screened (Supplementary Fig. 4 – consort diagram) and one adult with MAGE A1-expressing non-small cell lung cancer (NSCLC) (Supplementary Fig. 5A, Supplementary Note. 1) received two T_{TCR-MA1-CD8αβ} infusions (10⁹ cells, ratio CD4:CD8 1:1) 7 months apart. Infused T_{TCR-MA1-CD8αβ} expressed the CD28, CD62L and CD127 memory-associated markers (Fig. 2A) and produced IFNγ and TNFα upon cognate peptide exposure (Fig. 2B). Only low-grade adverse events (i.e. CTCAE grade ≤2) were documented with no post-infusion evidence of cytokine release syndrome (CRS) or immune effector cell-associated neurotoxicity (ICANS) (Supplementary Table 2). The patient progressed 60 days after infusions #1 and #2 (RECIST 1.1 criteria²⁶) and succumbed to his disease 4 months later. T_{TCR-MA1-CD8αβ} reached 1.2 and 2% of total CD3⁺ cells 1 day after each infusion but were detected at minimal levels 28 days later (Fig. 2C). Given the limitation of a single patient, this result cannot be conclusive but suggests additional measures may be needed to improve efficacy.

Indeed, mice infused with half doses of CD4 plus CD8 T_{TCR-MA1-CD8αβ} (5 × 10⁶ vs. 1 × 10⁷ total cells) (Fig. 2D) failed to control tumor growth (Fig. 2E). In the combined infusion group, only intratumoral CD4⁺ exhibited reduced expression of markers exhaustion-associated markers (Tim3, PD-1, CD39, and 2B4)^27–29, compared to either subset infused alone (Supplementary Fig. 5B, Fig. 2F), yet all T_{TCR-MA1-CD8αβ} subsets - either alone or combined - produced <10% median IFNγ and TNFα upon cognate peptide exposure (Fig. 2G). There were no Treg populations (CD4⁺FOXP3⁺) identified in either CD4-containing group (Supplementary Fig. 5C). These results show that in a more stringent murine model, the observed reduced expression of exhaustion-associated markers in CD4 occurred without gains in function or tumor control, highlighting the need for additional T cell optimization.

Patient tumors have limited and inconsistent expression of both positive and negative co-stimulatory ligands

T cells can receive co-stimulation either naturally through ligands in the TME or synthetically via immunomodulatory fusion proteins (IFPs) that couple inhibitory receptor ectodomains to costimulatory endodomains²⁰. The CD28 co-stimulatory molecule was naturally expressed on T_{TCR-MA1-CD8αβ} infused in a patient (Fig. 2A) and in our in vivo murine models³⁰ (Supplementary Fig. 6A), indicating their capacity to respond to positive co-stimulation. Unlike in hematologic malignancies, costimulatory ligands are less consistently present in solid tumors³¹. To examine the expression of both positive and negative co-stimulatory ligands across human TMEs, in silico analysis of previously published single-cell RNA sequencing (scRNA-seq) datasets from NSCLC (n = 42 patients) and melanoma patients (n = 8 patients)^32,33 was performed. Cell subsets were grouped and annotated according to prior reports^32,33 (Supplementary Fig. 6B and C). Differential gene expression of curated positive and inhibitory ligands (Supplementary Table 3) in the NSCLC and melanoma datasets (Fig. 2H and I, green and red boxes) revealed low/absent tumor-specific expression relative to myeloid and endothelial cells, with a positive co-stimulatory ligand gene score predominantly enriched in non-tumor cells (Fig. 2J and K). Surface expression of CD86, CD80 and 4-1BB-L was also low or absent in A375F and ME275 tumor lines (Supplementary Fig. 6D), confirming positive co-stimulatory ligands are rare in the TME and that a single IFP strategy is unlikely to be broadly applicable.

Intracellular CD28 domains integrated into the CD8β chain enhance CD4 T cell cytokine production, survival, and tumor killing in vitro

To achieve ligand-independent signaling, we tethered positive intracellular costimulatory domains to the requisite CD8αβ co-receptor chains, based on findings that TCR chain modifications impaired function (Supplementary Fig. 7A–C) and prior evidence that CD8α chain modifications can transmit activation signals^34–36. CD8α provides an LCK-binding domain, whereas CD8β provides a proximal palmitoylation site promoting membrane-lipid rafts association and downstream signaling^30,37. Although CD8α expression alone can drive signaling via dimerization³⁸, the CD8αβ heterodimer has ≥100 fold greater LCK affinity, favoring effective TCR activation and downstream signaling³⁹. First, we assessed CD8α alone and a chimeric CD8α-CD4 that retained the CD8α extracellular and transmembrane domains tethered to the intracellular CD4 tail, containing both LCK-binding and palmitoylation sites³⁷ (Supplementary Fig. 8A). TCR_MA1 CD4 T cells transduced with CD8α or CD8α-CD4 had similar surface expression as cells transduced with both CD8α and CD8β (Supplementary Fig. 8B). However, the functional avidity was lower for CD8α and CD8α-CD4 (EC₅₀ 169.9, 103.2, respectively, vs 7.58 for CD8αβ) (Supplementary Fig. 8C) and lytic capacity was reduced (Supplementary Fig. 8D), confirming that retaining the CD8β chain is necessary for T_TCR-MA1 function.

Next, to assess the impact of intracellular domain (ICD) modifications, we designed chimeric CD8αβ co-receptors (CCR) retaining the CD8α/β extracellular and transmembrane domains but replacing the entire CD8α ICD or CD8β ICD downstream of the palmitoylation site (Supplementary Table 4)⁴⁰. Substitution of the CD8α ICD reduced CD8 surface expression, whereas CD8β ICD modifications preserved it (Supplementary Fig. 8E). CCRs incorporating the ICDs of CD28, 4-1BB, ICOS, OX40 or GITR tethered to the CD8β chain downstream of the palmitoylation site were generated (Fig. 3A; Supplementary Tables 5 & 6). To standardize TCR expression, constructs were first titrated in CD4⁺ T cells, selected for their lack of endogenous CD8αβ expression, thus forcing the use of CCRs for class I-restricted responses, ensuring comparable transduction efficiency and transgene expression (Fig. 3B, first 3 panels to the left). Transduced T cells were then sorted on similar p/HLA+ binding and expanded (Fig. 3B, right panel).

Fig. 3 — A Schematic of the multi-cistronic vector designs used to integrate CCRs downstream of CD8β. Image created in BioRender. Tang, A. (2026) https://BioRender.com/6wg9k7f. B Histogram of CD8α (left), CD8β (middle), and MAGE-A1 p/HLA multimer (right) binding of CD4 transduced to express T_TCR-MA1 with the indicated CCRs after transduction for equal transduction copy numbers and post-REP (rapid expansion protocol) p/HLA multimer binding. C (left) Fit curves of IFNγ expression in response to decreasing peptide concentrations by CD4 T cells transduced with TTCR-MA1 and the indicated CCRs, as well as CD8αα (negative control), measured 18h post peptide stimulation. (right) Mean half maximal effective concentration (EC₅₀) of IFNγ expression for CD4 T cells transduced with the same constructs. D Fold change of IFNγ levels in T_TCR-MA1 cells transduced to express the different CCR constructs and incubated for 18 h with 1 µM cognate peptide alone (left panel) or incubated with 1 µM cognate peptide following 1 week of priming by co-culture with irradiated ME275 (right panel). E (left) Growth kinetics of the ME275 cell line in the absence (black lines) or presence of CD4 with the indicated CCR constructs from a live tumor visualization assay (Incucyte S3). An E:T ratio of 10:1 was used, and tumors were added to the culture every 72 h. (middle) Final tumor integrated intensity for the same experiment 16 days after start (384 h). (right) Final lymphocyte counts were obtained after 16 days. p-values determined by one-way ANOVA with Tukey’s multiple comparison test. (n = 3 wells/group). All data are presented as mean ± (SD).

CCRs incorporating CD28, 4-1BB, ICOS or OX40 showed functional avidities (EC_50s = 93.2, 128.6, 145.1 and 181.2, respectively) comparable to CD8αβ (125.2), whereas GITR and CD8α exhibited lower avidities (488.95 and 724.45) (Fig. 3C). All constructs produced similar levels of IFNγ after peptide exposure (1.02-1.19 fold vs T_{TCR-MA1-CD8αβ}) (Fig. 3D, left panel), but only CD28, 4-1BB, ICOS, and OX40 suggested sustained IFNγ secretion after sequential exposure to irradiated ME275 and peptide (1.53- 2.16 fold vs T_{TCR-MA1-CD8αβ}) (Fig. 3D, right panel). Among these, the CD8/CD28 co-receptor (CD8/28) most effectively restrained ME275 tumor growth in repeated killing assays (Fig. 3E, left and middle) and yielded greater live total remaining T cells in wells at 384 h (16 days) (Fig. 3E, right). No differences in tumor-killing were observed between T_{TCR-MA1-CD8/28} with or without CRISPR knockout of endogenous CD8β (Supplementary Fig. 8F) and the construct did not exhibit tonic signaling (Supplementary Fig. 9A). As such, subsequent adoptive transfer studies used CD4 and CD8 T cells transduced with TCR_MA15 and CD8/28 (T_{TCR-MA1-CD8/28}).

The chimeric CD8/CD28 co-receptor enhances TCR_MA1 cell function in vivo

To assess T cell-tumor interactions, we repeated the stringent xenograft model where 5 × 10⁶ T_{TCR-MA1-CD8αβ} (ratio CD4:CD8 1:1) failed to control tumor growth (Fig. 2D). Mice in the T_{TCR-MA1-CD8/28} group markedly reduced tumor size 9 days post-transfer (average tumor volume at euthanasia 37mm³) (Fig. 4A). Tumors from T_{TCR-MA1-CD8/28}-treated mice contained significantly more CD4 and CD8 T cells per mm² (Fig. 4B, quantified in Supplementary Fig. 9B), with higher T cell-to-tumor ratios overall (Fig. 4C), and within CD4 (Fig. 4D) and CD8 (Fig. 4E) T cell subsets. The CD4:CD8 ratio in this group approached 1 (mean 0.872) versus 0.381 in the T_{TCR-MA1-CD8αβ} group where CD8 T cells predominated (Fig. 4F). T_{TCR-MA1-CD8/28} cells showed reduced expression of surface markers associated with exhaustion (PD1⁺/Tim3⁺, CD39⁺/Tim3⁺, and PD1⁺/TIGIT⁺) (Fig. 4G, Supplementary Fig. 9C) and produced significantly more cytokines upon cognate peptide exposure ex vivo (Fig. 4H). Collectively, this data shows that T_{TCR-MA1-CD8/28} were detected at a higher frequency in tumors, potentially due to either preferential tumor accumulation or an earlier burst of intra-tumoral expansion. Notably, the remaining cells displayed fewer exhaustion-associated features, and exhibited enhanced function compared to T_{TCR-MA1-CD8αβ}.

Fig. 4 — A Representative of tumor burden measurements (left) for NSG mice engrafted with 1 × 10⁵ cells A375F and subsequently infused with a total of 5 × 10⁶ CD4 and CD8 T_{TCR-MA1-CD8αβ} CCR cells (T_{TCR-Irr.-CD8/CD28}, T_{TCR-MA1-CD8αβ}, T_{TCR-MA1-CD8/CD28)} at a ratio of 1:1 12 days post engraftment. Final tumor volume of tumor-bearing mice (right panel) between days 6–8 after infusion (n=9 mice/group, 2 tumors/mouse averaged, from 3 separate experiments). p value determined by Kruskal-Wallis with Dunn’s multiple comparison test. All data are presented as mean ± SEM. B Representative IHC images of A375F tumors resected from NSG mice (same experiment design as A) showing DAPI (blue), CD4 (red), and CD8 (green) T cells of 4 groups: No T cells, irrelevant TCR (T_{TCR-Irr.-CD8/CD28}), CD8αβ (T_{TCR-MA1-CD8αβ}), and CD8/CD28 CCR (T_{TCR-MA1-CD8/CD28}). 2X and 10X magnification. Scale bars represent 100 µm and 500 µm. The white box represents the area enlarged in the 500 µm images underneath the top images. **C–F** Tumor-infiltrating T cells to tumor cells ratio (C) from resected A375F tumors from NSG mice infused with T_{TCR-Irr.-CD8/CD28}, T_{TCR-MA1-CD8αβ}, or T_{TCR-MA1-CD8/CD28}. The indicated p-value was determined by a Mann–Whitney unpaired two-tailed t-test. CD4 TCR-T cells and tumor cells ratio (D), p value determined by Welch’s unpaired two-tailed t-test. CD8T cells and tumor cells ratio (E), p value determined by Mann-Whitney unpaired two-tailed t-test. CD4 and CD8 T cell ratio (F), p value determined by Welch’s unpaired two-tailed t-test. Analysis performed using HALO Link software on IHC slides. (n = 3 mice, 6 tumors/group, 2 tumors/mouse). All data are presented as mean ± SEM. G Bar graph showing PD-1⁺/Tim3⁺, CD39⁺/Tim3⁺, and PD-1⁺/TIGIT⁺ frequency of CD4 and CD8 TILs between T_{TCR-MA1-CD8αβ} and T_{TCR-MA1-CD8/CD28}. Single cell suspensions were obtained at sacrifice, 8 days after T cell infusion from (same experiment design as A). p value determined by unpaired two-tailed t-test. (n = 3 mice, 6 tumors/group, 2 tumors/mouse). The graph represents mean ± SD. H Quantification of IFNγ⁺/TNFα⁺ populations after 1 μM cognate peptide stimulation for 18 h ex vivo across the same experimental groups from (A) in CD4 (left panel) and CD8 (right panel) T_{TCR-MA1-CD8αβ} for the indicated conditions. Single cell suspensions were obtained at sacrifice, 6 days after T cell infusion. Graph represents mean ± SD. Unpaired two-tailed t-test. (n = 3 mice, 5-6 tumors/group, 2 tumors/mouse). I NSG mice were engrafted with 1 × 10⁵ cells A375F and subsequently infused with a total of 5 × 10⁶ CD4 and CD8 T_{TCR-MA1-CD8αβ} CCR cells (T_{TCR-Irr.-CD8/CD28}, T_{TCR-MA1-CD8αβ}, T_{TCR-MA1-CD8/CD28}) at a ratio of 1:1 6 days post engraftment. Survival analysis was carried out when tumor burden reached 500 mm³ and assessed tumor progression occurring 44 days vs. 23 days vs. 16 days post T cell infusion, respectively; the Kaplan-Meier test for analysis of survival was applied. p values determined by a log-rank test. (n = 6 mice/group). Panels A & I were created in BioRender. Tang, A. (2026) https://BioRender.com/6wg9k7f.

While the survival of all mice infused with T_{TCR-MA1-CD8/28} on day 13 after A375F engraftment was only modestly prolonged compared to controls (T_{TCR-Irr-CD8/28}) (Supplementary Fig. 9D), when administered on day 6 into mice with measurable tumors (Supplementary Fig. 9E), T_{TCR-MA1-CD8/28} significantly extended survival compared with T_{TCR-MA1-CD8αβ} (Fig. 4I).

T_{TCR-MA1-CD8/28-}mediated tumor reduction is associated with cytotoxicity and preservation of self-renewal features

To profile transcriptional programs, we performed scRNAseq on tumor-infiltrating T cells 9 days after infusion of CD4 or CD8 T_{TCR-MA1-CD8αβ} alone or combined (combo) CD4 and CD8 T_{TCR-MA1-CD8αβ} or T_{TCR-MA1-CD8/28} (Figs. 2D, 4A). A total of 3,266 T cells were identified (2,445 CD8 and 821 CD4). Because CD4 T cells were relatively scarce – particularly in the T_{TCR-MA1-CD8αβ} group (Supplementary Fig. 9B and Fig. 4F), we aggregated them across groups for pseudo-bulk analysis.

CD4 T cells from the CD8αβ-only and combo CD8αβ groups clustered together, marked by expression of activation (IRF7, ISG20, IFITM1, ISG15, CD38, IFI6, CCL3) and exhaustion (LAG3, TIGIT) gene signatures. By contrast, CD4 cells from the combo CD8/28 group formed a distinct cluster, enriched for replication-associated genes (VIM, TUBB, TUBB4B) and IL7R (self-renewal) without expression of activation- and exhaustion-associated signatures (Supplementary Fig. 10A).

Unsupervised clustering of CD8 T cells across groups revealed six main clusters (Fig. 5A): proliferative clusters (2, 5) defined by proliferation-related (MKI67, TOP2A), tubulin- and histone-related (TUBA1B, HIST1H2AC, HIST1H1B) genes; cytotoxic clusters (0, 1) defined by cytotoxicity-related (CTSA, CTSW, FCER1G, GNLY), interferon-associated (IFNG, ISG15) as well as exhaustion-associated (TIGIT, DUSP2)^41,42genes; early Effector-Memory/Early Tem clusters (3, 4) defined by central-memory/stem-like (LEF1, IL7R, SELL, NELL2) and effector (GZMK, CCL5) genes, indicating cells with self-renewing capacity transitioning to a more effector-like state (Fig. 5A–C, Supplementary Fig. 10B). A distinct “exhausted” cluster was not identified, consistent with prior studies showing that exhaustion markers can also be expressed in activated T cells, complicating their distinction in scRNA-seq data-sets^43–45.

Relative to CD8-only cells, CD8 T cells in the combo CD8αβ group showed enriched cytotoxic and proliferative but reduced stem-like signatures (Fig. 5D), aligning with their failure to achieve tumor volume reduction (Fig. 5E). In contrast, CD8 T cells in the combo CD8/28 group showed enriched cytotoxic and stem-like but reduced proliferative signatures (Fig. 5F), correlating with tumor volume reduction (Fig. 5G). Although speculative, the lower proliferative signature may reflect diminished antigen availability after tumor shrinkage, favoring the acquisition of self-renewing programs. Comparisons across groups suggest that CD8-only and CD8 combo CD8/28 proliferate and differentiate less than CD8 combo CD8αβ cells, thereby maintaining stem-like features. For CD8-only, this may reflect absence of CD4⁺ help (less proliferative stimulation), whereas for CD8 combo CD8/28 this may be linked to tumor burden reduction (Supplementary Fig. 10C). Cell proportions (Supplementary Fig. 10D) and absolute numbers (Supplementary Fig. 10E) further support this model: CD8 in the combo CD8/28 group showed a predominance of cytotoxic and stem-like rather than proliferative signatures (tumor burden reduction), whereas the CD8 in the combo CD8αβ group displayed cytotoxicity and proliferative rather than stem-like signatures (no tumor burden reduction compared to control).

Taken together, these observations suggest that combining CD8 and CD4 T cells enhances proliferation of both subsets, while CD28 co-stimulation promotes cytotoxic signatures, consistent with their increased ability to produce cytokines (Fig. 4H) and preserves early memory programs, consistent with their reduced expression of exhaustion-associated surface markers (Fig. 4G).

CD8-CD28 CCR tail mutations further enhance anti-tumor activity in MAGE tumor models

Based on studies demonstrating that modifications CARs co-stimulatory motifs can enhance function^46,47, we generated T_{TCR-MA1-CD8/28} constructs with distinct CD28 intracellular domain (ICD) motif mutation combinations. These included LL > GG shown to increase CD28 CAR surface expression⁴⁸, and YMNM > AMNM, PYAP > AYAA and PRRP > ARRA shown to abrogate PI3K³⁰, LCK⁴⁹ and ITK⁴⁹ binding sites, respectively (Supplementary Fig. 11A). CD4 T cells were co-transduced with one of 16 constructs (Supplementary Table 7). TNFα production after cognate peptide exposure was similar across constructs (EC₅₀ 12.93–56.98 nM) (Supplementary Fig. 11B), suggesting preserved functional avidity. However, differences emerged after repeated A375F tumor challenge (Supplementary Fig. 11C). CD4 T_{TCR-MA1-CD8/28} with all four domains mutated (E7, hereafter T_{TCR-MA1-CD8/28-mut}), demonstrated superior tumor control relative to other variants (Supplementary Fig. 11C). Grouped analysis showed LL > GG and PYAP > AYAA enhanced tumor control and T cell proliferation, YMNM > AMNM and PRRP > ARRA showed mixed effects (Supplementary Fig. 11D, E), and concurrent YMNM + PRRP mutations significantly improved tumor control compared to other combinations (Supplementary Fig. 11E). Overall, T_{TCR-MA1-CD8/28-mut} showed the strongest activity and was selected for further study (Fig. 6A).

Fig. 6 — A Schematic of the four motif mutations on the CD28 intracellular domain. B Dose response curve showing IFNγ and TNFα levels in CD4 T_TCR-MA1 CCR cells incubated with decreasing concentrations of cognate peptide overnight. C NSG mice were engrafted with 1 × 10⁵ A375F and subsequently infused with total 5 × 10⁶ CD4 and CD8 T cells (T_{TCR-MA1-CD8αβ}, T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut}) at a ratio of 1:1 on day 12 post engraftment. Graph represents mean ± SEM. (n = 3 mice/group, 2 tumors/mouse, data shown are representative of two independent experiments). D NSG mice were engrafted with 1 × 10⁵ A375F and subsequently infused with total 2.5 × 10⁶ CD4 and CD8 T cells (T_{TCR-Irr.-CD8/28}, T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut}) at a ratio of 1:1 on day 12 post engraftment. (n = 3mice/group, 2 tumors/mouse, data shown are representative of two independent experiments.) The graph represents mean ± SEM. E A375F tumor volume assessed at sacrifice 9 days after T cell infusions for each indicated condition same as (D). p value determined by one-way ANOVA with Tukey’s multiple comparison test. The graph represents mean ± SEM. (n = 4 mice for T_{TCR-Irr.-CD8/28} group and n = 5 mice for T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut} groups, data shown are representative of two independent experiments)**. F** Growth kinetics of the SK-MEL-37 cell line (left) in the absence (black lines) or presence of CD4 T cells with the indicated CCR constructs (T_{TCR-Irr.-CD8/CD28}, T_{TCR-MA1-CD8αβ}, T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut}) at a ratio of E:T 5:1 in a live tumor visualization assay (Incucyte S3). Final lymphocyte counts (right) were obtained after 84 h. The graph represents mean ± SD. p-values determined by one-way ANOVA with Tukey’s multiple comparison test. (n = 3 wells/group). G Non-humanized MISTRG mice were engrafted with 1 × 10⁶ SK-MEL-37 (left) and subsequently infused with a total of 2 × 10⁷ CD4 and CD8 T cells (T_{TCR-Irr.-CD8/28}, T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut}) at a ratio of 1:1 on day 21 post engraftment (n = 4 mice, 2 tumors per mouse). SK-MEL-37 tumor volume (right) assessed at 62 days after T cell infusions for each indicated condition. The graph represents mean ± SEM. p-value determined by a two-tailed unpaired t-test. (n = 4 mice/group, 2 tumors/mouse averaged). Panels C,D, and G were created in BioRender. Tang, A. (2026) https://BioRender.com/6wg9k7f.

T_{TCR-MA1-CD8/28mut} and T_{TCR-MA1-CD8/28} had similar functional avidities measured by production of TNFα and IFNγ after exposure to decreasing concentrations of cognate peptide (Fig. 6B, Supplementary Fig. 12A) and mediated similar A375F tumor control in the stringent in vivo model (Fig. 6C). However, when the total infused dose was halved (2.5 × 10⁶ T cells), T_{TCR-MA1-CD8/28mut} achieved more robust tumor burden reduction relative to controls (Fig. 6D, E). Cells preferentially localized to tumors while remaining detectable in concurrent spleens (Supplementary Fig. 12B). CD4 and CD8 T_{TCR-MA1-CD8/28mut} had higher expression of the Ki-67 proliferation marker than T_{TCR-MA1-CD8/28} by IHC (Supplementary Fig. 12C). T_{TCR-MA1-CD8/28mut} also suppressed growth of the cell death-resistant HLA-A2⁺ MAGE-A1⁺ SK-MEL-37 melanoma cell line in vitro (Supplementary Fig. 2, Fig. 6F)⁵⁰, and mediated superior tumor reduction in vivo in a MISTRG mouse model (see Methods) at 83 days compared to T_{TCR-MA1-CD8/28} (Fig. 6G). Thus, T_{TCR-MA1-CD8/28mut} conferred modest but consistent improvements in both short- (10 days) and long-term (83 days) tumor control, potentially due to an increased proliferative capacity.

To test generalizability, we applied the CD28-mut strategy to PRAME- and WT1-specific TCRs (Supplementary Fig. 13). T_{TCR-PRAME-CD8/28mut} displayed slightly improved functional avidity (EC₅₀ IFNγ 86.78 vs 91.79; TNFα 24.2 vs 38.37 nM) compared with T_{TCR-PRAME-CD8/28} (Supplementary Fig. 13A, B), while T_{TCR-WT1-CD8/28mut} showed similar functional avidity to T_{TCR-WT1-CD8/28} (EC₅₀ IFNγ 2.25 vs 0.57; TNFα 1.92 vs 0.26 nM) (Supplementary Fig. 13C, D). However, both T_{TCR-PRAME-CD8/28mut} and T_{TCR-WT1-CD8/28mut} achieved superior tumor control after repeated tumor challenge in vitro against HLA-A2⁺ PRAME⁺ PANC1 (pancreatic ductal adenocarcinoma), HLA-A2⁺ PRAME⁺ A375F, and HLA-A2⁺ WT1⁺ SW480 (colon adenocarcinoma) lines compared with their non-mutated counterparts (Supplementary Fig. 13E–G). These findings indicate that CD28 tail mutations may enhance efficacy beyond TCR_MA1, across to other TCR therapies.

CD28mut with less evidence of unrelenting activation maintains responsiveness following repeated antigenic stimulation

To assess short-term effects of the mutant CD28 tail, we exposed T_{TCR-MA1-CD8αβ}, T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut} CD4⁺ T cells to cognate peptide and performed bulk RNA sequencing (RNAseq) with Gene Set Enrichment Analysis (GSEA). Compared to CD8αβ, both CD28 and CD28mut conditions showed similar enrichment in cytokine-related pathways including Th1/Th2 signaling (Fig. 7A), indicating comparable rapid activation of cytokine-driven programs. Exposure to cognate peptide-loaded TAP-deficient T2 B lymphoblastoid cell lines (T2-BLCL) revealed higher phosphorylated downstream TCR and CD28 signaling molecules (p)-ERK and (p)-AKT in T_{TCR-MA1-CD8/CD28mut} than T_{TCR-MA1-CD8/CD28} at 5 minutes, with both exceeding T_{TCR-MA1-CD8αβ} (Supplementary Fig. 14A), consistent with more pronounced early CD28-mediated costimulatory signaling and TCR-driven signaling⁵¹.

Fig. 7 — A GSEA pathway analysis of cytokine pathways (Biocarta), cytokine activity (GOMF), cytokine and inflammatory response (WP), and the Th1/Th2 pathway (Biocarta) of CD4 T_{TCR-MA1-CD8/28}, T_{TCR-MA1-CD8/28mut} normalized by T_{TCR-MA1-CD8αβ}. T_{TCR-MA1-CD8αβ}, T_{TCR-MA1-CD8/CD28}, and T_{TCR-MA1-CD8/CD28mut} were stimulated with 1 µM cognate peptide for 4 h and further analyzed by bulk RNA sequencing. p values were determined by two-sided Mann–Whitney U-test. B Schematic of experimental layout showing CD4 T_{TCR-MA1-CD8/28} or T_{TCR-MA1-CD8/28mut} incubated with irradiated ME275 for 21days, with tumor rechallenge every 7 days. Cells were stimulated with 1 µM cognate peptide for 18 h or without stimulation and further analyzed by flow cytometry and bulk RNA-seq downstream analysis. Image created in BioRender. Tang, A. (2026) https://BioRender.com/f64l203. C PCA plot of bulk RNA-seq data, showing the first two principal components (PC1 and PC2). Data points represent individual samples, color-coded by condition to visualize any separation based on gene expression profiles. Ellipses, drawn at a 90% confidence level, indicate the range of variation within each condition, highlighting how closely related or distinct the three groups are from each other. D Heatmap showing the expression levels of differentially expressed genes across constructs (T_{TCR-MA1-CD8/CD28}, T_{TCR-MA1-CD8/CD28mut}) and stimulations (unstimulated vs. stimulated) in the rlog-transformed data. Rows represent genes, and columns represent individual samples (3 biological replicates per condition). Expression values are color-coded from low (blue) to high (red) based on z-scores. The dendrograms on the axes indicate hierarchical clustering of genes and samples. E Boxplots showing the z-scores for the expression of manually curated gene signatures of proliferation in unstimulated and stimulated conditions. The z-scores were calculated from the averaged normalized counts of significant genes (p_adj <0.05, log₂ fold-change > 0.5). Each box represents the distribution of z-scores for a specific gene signature in the unstimulated and stimulated groups for T_{TCR-MA1-CD8/CD28} and T_{TCR-MA1-CD8/CD28mut}. Boxplots show the IQR; lines denote medians. The bounds of the box represent the 25th and 75th percentiles; whiskers span 1.5x IQR. F Boxplots showing the z-scores for the expression of manually curated gene signatures of exhaustion across unstimulated and stimulated conditions. The z-scores were calculated from the averaged normalized counts of significant genes (padj <0.05, log2 fold-change > 0.5). Each box represents the distribution of z-scores for a specific gene signature in the unstimulated and stimulated groups for T_{TCR-MA1-CD8/CD28} and T_{TCR-MA1-CD8/CD28mut}. Boxplots show the IQR; lines denote medians. The bounds of the box represent the 25th and 75th percentiles; whiskers span 1.5x IQR. G Dot plot summarizing GSEA results for the ‘Apoptosis’ pathway across multiple comparisons. The x-axis represents the pathway of interest, while the y-axis denotes the pairwise comparisons between experimental conditions. Dot size corresponds to the -log10-transformed p-value, indicating the significance of enrichment, and dot color represents the normalized enrichment score (NES), with red indicating positive enrichment and blue indicating negative enrichment. Enrichment p-values were calculated by two-sided permutation testing. H Flow cytometry quantification of Tim3⁺ PD-1⁺ double positive populations from T_{TCR-MA1-CD8/28} and T_{TCR-MA1-CD8/28mut} groups was measured at the end of the 21 days stimulation with irradiated ME275 incubation. The graph represents mean ± SD. p-value determined by a two-tailed unpaired t-test. (n = 3 wells/group). I Flow cytometry quantification of IFNγ⁺ and TNFα⁺ expression in T_{TCR-MA1-CD8/28} and T_{TCR-MA1-CD8/28mut} group was measured at the end of the 7 days stimulation with irradiated ME275 incubation. The graph represents mean ± SD. p value determined by one-way ANOVA with Tukey’s multiple comparison test. (n = 3 wells/group).

To evaluate the long-term effect(s), CCR-expressing cells were challenged with irradiated tumor cells every 7 days for 21 days and analyzed by bulk RNAseq with or without peptide stimulation (Fig. 7B). Principal component analysis (PCA) showed transcriptional separation along PC1, distinguishing T_{TCR-MA1-CD8αβ} (Fig. 7C, brown ellipse) from the CD28 and CD28mut groups (Fig. 7C, green and purple ellipses) irrespective of peptide stimulation (Fig. 7C, Supplementary Fig. 14B). Differential gene expression and unsupervised clustering separated CD28mut (Fig. 7D, left side of the heatmap) from CD28 (right side of the heatmap), with each grouping together independent of stimulation, suggesting greater similarities within each construct, regardless of stimulation. Unstimulated T_{TCR-MA1-CD8/CD28mut} showed no/low activation/proliferation gene expression, which increased after peptide stimulation, while T_{TCR-MA1-CD8/CD28} displayed higher basal activation further enhanced by peptide stimulation (Fig. 7D, insert). Z-score analysis of selected genes (Supplementary Table 8) revealed upregulation of proliferation and exhaustion programs in both CD28 and CD28mut, but at comparatively lower levels in CD28mut pre- and post-stimulation (Fig. 7E, F), along with reduced apoptosis gene expression (Fig. 7G). T_{TCR-MA1-CD8/CD28mut} also showed lower PD1/Tim3 co-expression (Fig. 7H), indicating reduced cell exhaustion, while IFNγ and TNFα production remained comparable to CD28 (Fig. 7I).

Together these results suggest that T_{TCR-MA1-CD8/CD28mut} retains responsiveness to stimulation while limiting basal activation, reduces exhaustion under repeated stimulation, and maintains self-renewal features, potentially supporting more durable anti-tumor activity than T_{TCR-MA1-CD8/CD28wt}.

Discussion

Clinical trials of adoptive T cell therapies in solid tumors have shown limited efficacy^52–54. Barriers include limited persistence of transferred T cells, reduced function, and restricted accumulation at the tumor site, as well as tumor-associated immunosuppressive microenvironments, HLA downregulation, and heterogeneous expression of costimulatory ligands. Overcoming these challenges is critical for effective and persistent anti-tumor responses in patients. Our study focused on synthetic strategies to enhance the potency and durability of TCR-engineered T cells in the solid tumor setting.

CD4 T cells can be transduced with a Class I TCR by co-expressing the CD8αβ co-receptor, thus enhancing the persistence and function of engineered CD8^12,13 and CD4 T cells that recognize the same tumor target when class II expression is often limited⁵⁵. However, both our experience in one patient-case - although hampered by limited interpretability - and stringent murine models, point to the insufficiency of this intervention alone in consistently controlling solid tumors.

The dependency of memory CD4 T cells on co-stimulation, particularly CD28, for robust TCR activation⁵⁶ likely contributes to the lack of observed efficacy, particularly in immunosuppressive microenvironments where expression of co-stimulatory ligands are restricted⁸, and inhibitory ligands enriched. IFPs that convert inhibitory to positive costimulatory signals can enhance immune function¹⁹, but their variable expression across tumors limits individual applicability.

Tethering the CD28 costimulatory intracellular tail to CD8β may bypass the absence of positive or suppressive co-stimulatory ligands in our studies yielding greater benefit than the 4-1BB costimulatory domain. In the context of sparsely presented peptide/HLA antigen complexes compared with CAR T cell targets, CD28 signaling which elicits faster, larger magnitude changes in downstream protein phosphorylation compared to 41BB’s slower, memory-type activities⁵⁷ could represent a major determinant for eliciting TCR-driven solid tumor regression.

When tethered to CD8β, modifications in the dileucine motif (LL- > GG) are consistent with its previously reported benefit in CD19-CD28ζ CARs⁴⁸, alongside modifications of the CD28 cytoplasmic tail signaling motifs YMNM (mediates PI3K binding and Grb2 activation⁵⁸), PRRP (interacts with ITK to drive ERK and PLCγ signaling⁴⁶ and PYAP (engages LCK to initiate downstream signaling⁵⁹), which moderately but consistently enhanced anti-tumor killing and downstream TCR and CD28 signaling compared to unmodified CD28. Although these modifications would intuitively have been predicted to dampen function, prior studies in endogenous CD28-mutant mouse models showed that YMNM and PYAP domain mutations retain GRB2 binding and CD28-dependent anti-tumor activity in murine models⁶⁰, consistent with our findings and suggestive of preferential signaling through GRB2-dependent pathways. For CD28-based CARs, the PRRP → ARRA mutation was shown to abrogate ITK binding and reduce GADS association, with no impairment to CD28-dependent proliferation⁶¹, and downstream NF-κB activation can proceed independently of ITK binding. These findings suggest pathway redundancy, consistent with the minimal impact of this modification on tumor-killing. As the CD4⁺ CD8β/CD28 CCR incorporates TCR–MHC recognition to trigger CD28 co-stimulation in contrast to an independent CAR construct, this could potentially result in distinct affinity and signaling thresholds. Clarifying these differences may further elucidate how individual mutation motifs shape the balance of CD28 signaling.

Exhaustion remains major challenge for effective T cell therapy⁶². Our results show both T_{TCR-MA1-CD8/CD28} and T_{TCR-MA1-CD8/CD28mut} retain function, but T_{TCR-MA1-CD8/CD28mut} exhibit more resistance to exhaustion following serial antigen-specific activation. These findings suggest T_{TCR-MA1-CD8/CD28mut} may achieve a more optimal balance between T cell activation and preserved stemness in this setting. Next steps will include determining whether analogous motif modifications in the CD28 tail, when positioned downstream of IFPs or CARs can mediate similar benefits⁴⁷.

This study relies on immunodeficient NSG models, which lack critical components of the human immune system and do not fully recapitulate the complexity of the TME. Furthermore, an insufficient number of patients received T_{TCR-MA1-CD8αβ} which impedes drawing meaningful conclusions of the effect of co-engaging CD4 T cells towards the same Class I target. While human studies further assessing whether the T_{TCR-MA1-CD8/CD28} and T_{TCR-MA1-CD8/CD28mut} mediate similar effects in patients are being implemented, humanized mouse models may provide a reflective platform, as they replicate key constraints of the solid tumor microenvironment such as stromal barriers, immunosuppressive cells, and cytokines that are likely to prevent infiltration and negatively impact the efficacy of transferred T cells^63,64.

Overall, our findings lay the groundwork for enhancing and refining synthetic co-stimulatory strategies and support the continued development of next-generation TCR-engineered T cell therapies.

Method

Isolation of TCR_MA1 and lentiviral vector construction

T cell lines recognizing MAGE-A1_278-86 peptide (KVLEYVIKV) were generated from 12 donors, cloned by limiting dilution, and screened for high functional avidity for an HLA-A*0201⁺ TAP-deficient B-lymphoblastoid cell line [T2 B-LCL] pulsed with the MAGE-A1_278-86 peptide^65,66. The 5 clones selected as TCR donors, TCR_MA1-TCR_MA5, demonstrated the lower K_D values for tetramer binding. TCR_MA11-TCR_MA15 Vα1.3 and Vβ17 TCR chains were isolated using Rapid Amplification of cDNA Ends-Polymerase Chain Reaction (RACE-PCR)²² and expanded and cloned by limiting dilution. Based on the wild-type sequences, the final construct was synthesized by GeneArt (Regensburg, Germany) in a codon-optimized format to promote high-level protein expression in human cells as previously described²³.

Functional avidity screening of 5 different clones of TCR_MA1

Donor PBMC were thawed and rested in RPMI-1640 supplemented with 10% immune cell serum (A2596101, Gibco) overnight before transducing with the five clones. T cells were identified by flow cytometry using HLA-A2:MAGEA1 peptide/HLA-multimer-APC (p/HLA-multimer) (1:200, KVLEYVIKV, FHCRC in-house production), anti-CD8-FITC (1:200, RPA-T8, BioLegend, 301050), and anti-CD4-BV421 (1:200, RPA-T4, BioLegend, 300532).

Intracellular cytokine staining (ICS) was performed by surface staining with anti-CD4-APC-Cy7 (1:200, OKT4, BioLegend, 300518), anti-CD8-FITC (1:200 RPA-T8, BioLegend, 301050), and Live/Dead-BV510 (1:1000, Invitrogen, L34957). Cells were fixed and permeabilized with Cytofix buffer (BD Biosciences, 554722) and Perm/Wash (BD Biosciences, 554723) per manufacturer’s protocol and stained with anti-TNFα-PE-Cy7 (1:200, MAb11, BioLegend, 502930) and anti-IFNγ-APC (1:200, B27, BioLegend, 506510), then analyzed by flow cytometry.

Production of TCR-deficient Jurkat cells and Nur77-mNeonGreen reporter cells

0.625 μL of 160 μM CRISPR RNAs targeting TRAC (sequence: AGAGTCTCTCAGCTGGTACA, IDT) and TRBC1/2 (sequence: GGAGAATGACGAGTGGACCC, IDT) were each duplexed with 1.25 µL of tracrRNAs (IDT, 1072533) in duplex buffer (IDT, 11-01-03-01) for 30 minutes at 37 °C. Resulting duplexes were then incubated with Cas9 protein (61 μM, 1 μL, IDT) and polyglutamic acid (100 mg/mL, 1 μL) for 15 minutes at 37 °C. Following this, 4 μL of the mixture was added to 17 μL of P3 buffer (Lonza) and used to resuspend 1×10⁶ Jurkat cells (ATCC, TIB-152). The 20 μL cell suspension was then electroporated using the 4D-Nucleofector X Unit (Lonza, V4SP-3096) with the EH-115 program. Subsequent CD3-negative cells were sorted by FACS. To generate Nur77 reporter cells, the CD3-KO Jurkat cells were further modified by knocking out class-I HLA genes using CRISPR with a pan-class-I-HLA gRNA (sequence: AAAAGGAGGGAGCTACTCTC), overexpressing codon-optimized CD8β-P2A-CD8α (pRRLSIN backbone) via lentiviral transduction, and knocking in mNeonGreen into the NR4A1 (Nur77) locus using CRISPR/homology-directed repair (guide: ATGAAGATCTTGTCAATGAT).

Recognition of MAGE-A1-K278T variant

Nur77 reporter Jurkat cells were transduced with the 5 TCR clones (TCR_MA1-5) using protamine sulfate (1:1000, 10 μg/mL stock, MP Biomedicals, S4480). TCR transduced Jurkats were cocultured with peptide-loaded T2 B-LCLs (1 µM MAGE-A1₂₇₈ or MAGE-A1_K278T) overnight and analyzed by flow cytometry for Nur77-mNeonGreen reporter activation.

Selection of the MA1-specific TCR for clinical translation

While other somatic mutations exist within the MAGE-A1 gene, only the K278T missense mutation falls within the targeted epitope (MAGE-A1₍₂₇₈₋₂₈₆₎). The K278T mutation has been in publicly available datasets and is included in the breast cancer cell line HCC1008²⁵. Of the 5 isolated TCRs and to choose one for clinical translation, we proceeded by first excluding the least avid TCR_MA11 (mean EC50 of 134 nM). The remaining 4 TCRs had very similar functional avidity (ranging from 48.1-85.8 nM). TCR_MA15 was chosen based on its ability to recognize the wild-type and mutated peptide, thus broadening the targeted epitope to include the K278T mutation. All 5 TCRs (TCR_MA11-5) were isolated from healthy HLA-A2+ donors. To avoid safety concerns, none were modified or affinity enhanced.

Generation of TCR_MA1 lentivirus used in pre-clinical studies

Sequenced TCR_MA15 was synthesized (Genscript) and cloned into 3^rd generation lentivirus vector pRRLSIN.cPPT.MSCV/TCRMA1β-P2A-CD8α-P2A-CD8β-P2A-TCRMA1α/wPRRE(in-house generation, FHCRC). Lentivirus was generated by transfecting HEK293T cells (ATCC, CRL-3216) with the TCR_MA15 lentiviral construct using Effectene Transfection Reagent (Qiagen, 301425). The virus was concentrated using the Lenti-X Concentrator (Takara Bio, 631231).

Generation of T_TCR-MA1: T cell isolation, activation and lentiviral transduction

PBMCs were thawed and rested for 1 h in X-VIVO15 media (Lonza, 04-418Q) supplemented with 5% fetal bovine serum (FBS)(Cytiva, SH30071.02), 5% CTS (Gibco, A2596101), and 1% antibiotic-antimycotic (Gibco, 15240062). CD4 or CD8 T cells were isolated using the CD4+ (Miltenyi, 130045101) or CD8+ T cell (Miltenyi, 130045201) positive selection kit per the manufacturer's protocol. The cells were activated for 48 h using TransAct (1:50, Miltenyi, 130111160) at 2×10⁶ cells/mL with 50 IU/mL recombinant interleukin-2 (IL-2) (Peprotech, 200-02) in X-VIVO15 media supplemented with 10% FBS and 1% antibiotic-antimycotic solution. Activated CD4 and CD8 T cells were lentivirally transduced using protamine sulfate (1:1000, 10 μg/mL stock) and supplemented with IL-2 (50 IU/mL). CD4 and CD8 T cells were expanded for one week, then stained with p/HLA-multimer-APC, anti-CD8-FITC and anti-CD4-PE (1:200, SK3, BioLegend, 980804), then FACS sorted by BD Symphony S6 sorter and cryopreserved with Cryostor CS10 (Stem cell, 07930). Chimeric Co-receptor (CCR) construct generation: chimeric co-receptor sequences replaced the CD8β chain but retained the first 6 amino acids (Supplementary Table 6) and were transduced into CD4 or CD8 T cells. CCR T cells were validated with anti-CD8α (1:200, BV421, RPA-T8, BioLegend, 301035), CD8β (1:200, PE, H35-17.2, Invitrogen, 12-0083-82) and p/HLA-multimer-APC. T cell identification with surface antibodies follows the surface gating strategy in the supplementary information (Supplementary Fig. 15A).

T cell rapid expansion protocol (REP)

CD4 and CD8 T cells were FACS sorted using p/HLA-multimer, anti-CD8-FITC, and anti-CD4-PE⁶⁵(Supplementary Fig. 15C). For each condition to expand, 5 × 10⁴ cells were combined with 10 × 10⁶ mixed PBMCs (3 healthy donors, irradiated with 40 Gy) and 2 × 10⁶ TM-LCL cells⁶⁵ (Irradiated with 80 Gy) into a 6-well GREX plate (Wilson Wolf). Each well was supplemented with anti-OKT3 (30 ng/ml, Miltenyi, 130-093-387), IL-2 (50 IU/ml, PeproTech, 200-02), IL-7 (10 ng/ml, PeproTech, 200-07), and IL-15 (10 ng/ml, PeproTech, 200-15). Cells were cultured for 10–12 days and then checked for purity with the p/HLA-multimer-APC, anti-CD4-PE, and anti-CD8-FITC as described above. T_TCR-MA1 CD28 expression assessment: REP T cells were stained with anti-CD28 (1:200, PE-Cy7, 28.2, Invitrogen, 25-0289-42), anti-CD4-APC-Cy7, and anti-CD8α-BV421.

Generation of research tumor cell lines

A375 (ATCC, CRL-1619), ME275 (Research Resource Identifier [RRID]: CVCL_S597), and SK-MEL-37 (RRID: CVCL_3878) melanoma cell lines were obtained from ATCC and Ludwig Institute for Cancer Research and used as target cell lines for T_TCR-MA1 cells. The A375 cell line was modified to increase expression of HLA-A2 and MAGE-A1 (Supplementary Fig. 2A). Briefly, A375 cells were transduced with the pRRLSIN.cPPT.MSCV/HLAA2-MAGEA1/wPRE (in-house generation, FHCRC) lentiviral vector and later stained with anti-HLA-A2-FITC (1:200, BB7.2, BioLegend, 343303) and FACs sorted on HLA-A2^hi and subsequently referred to as A375F. A375F and ME275 were stained with anti-41BBL-PE (1:200, 5F4, BioLegend, 311503), anti-CD80-BV650 (1:200, 2D10, BioLegend, 305227), and anti-CD86-BV421 (1:200, GL-1, BioLegend, 105031) to confirm costimulatory ligand expression. PANC-1 and SW480 were obtained from ATCC (CRL-1469 and CCL-228) and used as a target of T_TCR-PRAME or T_TCR-WT1, respectively, in Incucyte killing assays. Tumor cell lines were transduced with luciferase-GFP and mCherry in the 3^rd generation lentivirus vector (see above) for Incucyte killing assays and in vivo experiments.

T_TCR-MA1 cell cytotoxicity assays: Intracellular Cytokine Staining

T cells were seeded at 1 × 10⁵ cells/well in 100 µl X-VIVO-15 medium supplemented with 5% CTS and 5% FBS with 6 different peptide concentrations for 18 h (10 µM, 1 µM, 100 µM, 10 µM, 1 µM, 0 µM). Golgi stop (BD Biosciences, 554724) and Golgi plug (BD Biosciences, 555092) were added 18 h before surface staining with anti-CD4 (1:200, PE, SK3, BioLegend, 980804), anti-CD8 (1:200, FITC, RPA-T8, BioLegend, 301050), and Live/Dead-BV510 (1:1000, Invitrogen, L34957) (Supplementary Fig. 15B). Cells were analyzed with BD Fortessa-X50.

Real-Time Killing Assay

CD4 and/or CD8 T cells purity was confirmed following the REP protocol using anti-CD4-PE, anti-CD8-FITC, and p/HLA-multimer-APC, and were seeded at 1 × 104T cells/well in a clear, 96-well tissue-treated flat-bottom plate with 5 × 103 ME275 cells expressing mCherry. Kinetic image analysis was performed using the Incucyte S3 (Sartorious, 51540). Images were recorded every 2 h, and 5 × 103 ME275 cells were added every 48 h, depending on tumor killing efficiency. Serial rounds of tumor challenges were repeated until loss of tumor control was identified. Other tumor cell lines used in this study, A375F, SK-MEL-37, PANC-1, and SW480, were seeded as before with respective CCR TCR-T cells (TCR_MA1, TCR_PRAME, TCR_WT1). E:T ratio varied from 2:1 to 10:1 depending on the effector and target cell types. 5 × 103 tumor cells were added between 48 and 72 h, depending on the lytic efficiency of the T cells. Analysis of T cell counts in Real-time killing assays: At indicated timepoints remaining T cells in the wells were counted via flow cytometry with count bright counting beads (Invitrogen, C36950).

CFSE proliferation assay

CD4⁺ or CD8⁺ T cells were labeled with 5 µM CFSE (carboxyfluorescein diacetate succinimidyl ester; Invitrogen, C34554) in PBS for 20 min at 37 °C, followed by quenching with complete culture medium and washing twice. Labeled cells were then stimulated with ME275 cells at an effector-to-target ratio of 5:1 in complete X-VIVO 15 media. Half of the media was changed 3 days after, and cells were analyzed by flow cytometry 8 days after. Proliferation was assessed by sequential dilution of CFSE intensity in viable CD4⁺ or CD8⁺ T cells.

In vivo murine models

NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ mice (NSG) were bred in-house at the Fred Hutchinson Cancer Center from mice obtained from The Jackson Laboratory (Strain#:00555, RRID:IMSR_JAX:005557) and M-CSF^h/h IL-3/GM-CSF^h/h hSIRPA^tg TPO^h/h Rag2^-/- Il2rg^-/- (MISTRG) (VG #5097 / 5090 / 5155/ 5089 / 5079 / 5078, Regeneron)⁶⁴ were bred in Rongvaux lab (Fred Hutch Cancer Center) and provided by Rongvaux lab. All mice were maintained in pathogen-free conditions at the Fred Hutchinson Cancer Center (Seattle, WA), and acclimatized for at least 2 days before experimentation, which was performed per Institutional Animal Care and Use Committee guidelines. All experimental mice were female and ranged in age from 8 to 12 weeks. All mouse experiments were approved by the Institute for Animal Care and Use Committee (IACUC Protocol #50898 and #2400037) at the Fred Hutchinson Cancer Center. All mice were euthanized using CO₂ according to American Veterinary Medical Association (AVMA) guidelines.

1 × 10⁶ ME275 or 1 × 10⁵ A375F modified tumor cells were injected subcutaneously (SC) in 100 µl of solution containing 75% PBS, and 25% Matrigel (Corning, 356237). Tumors were monitored for 13 days or 22 days for A375F or ME275 respectively, at which point tumor volumes were recorded and mice were randomized accordingly. 2.5 × 10⁶, 5 × 10⁶, and 1 × 10⁷T cells with different TCR_MA1 constructs were intravenously injected into mice between days 9–12. Groups included an irrelevant TCR (TCR_Irr.-CD8αβ, TCR_Irr.-CD8/28) as negative controls. After infusion, tumors were measured 3 times a week and mice were euthanized 10–14 days later. Tumors, spleens, and blood were harvested at approximately day 21 for A375F and day 33 for ME275. Infiltrated T Cells were isolated from tissues by the Human Tumor Dissociation Kit (Miltenyi, 130095929). ACK Lysis Buffer (Gibco, A1049201) was used to remove erythrocytes. TILs were counted with CountBright absolute counting beads (Invitrogen) from the tumor single cell suspension. The single cell suspensions were analyzed with flow cytometry (see In vivo experiment flow cytometry panel) or cryopreserved for 10X sc-RNA seq preparation. Undigested tumors were fixed in 10% neutral buffered formalin (Azer Scientific, PFNBF-20) and paraffin-embedded for immune-histochemistry staining and analysis. Survival experiments: tumors were monitored for 6 days until all palpable. A total of 5 × 10⁶T cells were injected on day 6, and tumor growth rates were recorded until reaching 500 mm³ in volume. Mouse tumor measurement guidelines: All mice engrafted with tumors were monitored 3 times per week and the every day after tumors became palpable (5-7 mm in diameter). In compliance with institutional (Fred Hutchinson Ethics Committee) regulations, tumors were measured using digital calipers and not allowed to exceed the maximum size of 20 mm in largest diameter (2000 mm³) for 1 implanted tumor, or 15 mm in diameter for 2 implanted tumors (1500 mm³). Tumor volume was determined by digital caliper measurement with tumor length x width x depth (ME275 and A375F models) and tumor length x width x ½ width (SKMEL-37 model).

Clinical protocol

The clinical trial (NCT04639245) was approved by the Fred Hutchinson Cancer Center (FHCC) Institutional Review Board (FWA00001920 / IORG0000057) under protocol FHCC #10420/RG1007463 (Supplementary Note 1), the U.S. Food and Drug Administration, and the Recombinant DNA Advisory Committee. The study design and conduct complied with all relevant regulations regarding the use of human study participants and was conducted in accordance with the criteria set by the Declaration of Helsinki. HLA A*0201 (HLA-A2)-expressing patients with triple negative breast cancer (TNBC), non-small cell lung cancer (NSCLC) or urothelial carcinoma were eligible for treatment. All participants provided informed consent.

Patient selection

HLA-A2 genotype was confirmed by high-resolution typing and MAGE-A1 expression was confirmed by immunohistochemistry defined as any tumor with >10%, 1+ or higher staining intensity, membranous, nuclear or cytoplasmic expression of MAGE-A1, prior to enrollment. Inclusion criteria further included measurable disease by RECISIT 1.1 criteria, ≥ 18 years old, ECOG performance status ≤ 2. Exclusion criteria included autoimmune diseases requiring immunosuppressive therapy and other tumor-specific concurrent treatments.

Treatment plan

The study was originally designed first as a Phase I 3 + 3 dose escalation (Supplementary Note 1). The first participant enrolled in each cohort was planned to receive 1 × 10⁹ T_TCR-MA1 alone, standard lymphodepleting chemotherapy of cyclophosphamide and fludarabine could be administered to subsequent patient within the cohort. Participants were eligible to receive a first infusion of 1 × 10⁹ T_TCR-MA1 with acceptable organ function (eGFR > 30 mL/min, tBili <3.0 mg/dL, AST, and ALT <5x upper limit of normal, SaO2 ≥ 92% on ambient air, left ventricular ejection fraction (LVEF) ≥ 35%, ANC > 500 cells/mm³). Second infusions at the same dose level could be administered as soon as 12 weeks after the first and if no dose limiting toxicities had occurred. If indicated, the protocol allowed for interval bridging treatments between infusions. Patients were monitored for toxicities based on Common Toxicity Criteria v5.0⁶⁷. Once the Phase I was completed, the protocol planned that subsequent participants could be enrolled on the Phase I/II in which the maximum tolerated dose was going to be combined with atezolizumab or another PD1-axis inhibitor. However, the protocol was closed before the Phase I/II could enroll. Thirty-two patients were screened for this protocol. 8, 10, and 14 had TNBC, NSCLC, and urothelial cancer, respectively (Supplementary Fig. 5). Fifteen expressed HLA A*0201, and of those, 2 had tumors that expressed MAGE-A1. Two participants were enrolled in the study; one did not receive T_TCR-MA1 due to persistently undetectable disease on prior bridging therapy and one received 2 infusions of 1 × 10⁹ T_TCR-MA1 (dose level 1). Due to external factors (COVID pandemic, loss of corporate funding), the study was closed to accrual and treatment on 10/01/2022.

Patient 1 case report

This lifetime never-smoker was diagnosed with metastatic non-small cell lung cancer. A CT-guided transthoracic biopsy of a lung nodule showed a poorly differentiated squamous cell carcinoma that was TTF1 negative. Molecular testing was negative for actionable mutations, including EGFR, ALK, ROS1, MET amplification, RET rearrangement, PTEN depletion, HER2, and PD–L1. The patient received prior treatments of radiation (30Gy to a vertebral lesion), node-removal surgery, and expanded tumor-infiltrating lymphocytes (TILs) in combination with the PD-1 inhibitor nivolumab and a STAT3 inhibitor, all of which resulted in disease progression. After leukapheresis to generate the T_{TCR-MA1-CD8αβ} product, additional palliative radiation to a rib lesion was given as bridging therapy. Immediately before infusion, computed tomography evidenced at least 5 lung and pleural masses measuring 2.4 × 2.9 cm (smallest) to 3.2 × 6.1 cm (largest). The infusion was well tolerated without a cytokine release syndrome (CRS) or immune effector cell-associated neurotoxicity syndrome (ICANS) (Table 2). Disease progression occurred 60 days later. Following clinical delays, including a DVT, COVID infection, and after additional bridging chemotherapy of carboplatin and Taxol, a second infusion of 1 × 109 TTCR-MA1 was infused seven months later. Progressive disease again developed sixty days later, and the patient succumbed to NSCLC four months later.

Generation of T_TCR-MA1 for clinical use

All ex vivo procedures involving the processing of products intended for infusion were performed in the cGMP Cell Processing Facility (CPF) at FHCC. CD4 and CD8 T cells, used to generate the TCR MAGE-A1_278-86 transduced T cells (T_TCR-MA1), were sourced from an autologous PBMC donation through leukapheresis of patients enrolled in Protocol FHCC #10420. The cells were stimulated using polymer-bound CD3 and CD28 (TransAct, Miltenyi). On days +1 and +2 of stimulation, the cells were transduced with the GMP-grade pRRLSIN.cPPT.MSCV/TCRMA1β-P2A-TCRMA1α/wPRE lentiviral vector. Protamine sulfate (10 μg/mL), IL-2 (50 IU/mL), IL-21 (30 ng/mL), IL-7 (5 ng/mL), and IL-15 (1 ng/mL) were added, and the cells were centrifuged at 800 g for 90 minutes at 30 °C before being incubated overnight at 37 °C. The cells were harvested on day +12 ( + /−2) for infusion. Quality control of the infused products included assessing the lentiviral vector copy number per cell (targeting 5 copies/cell), envelope VSV-G DNA as a surrogate marker for replication-competent lentivirus (RCL) ( < 10 copies/50 ng DNA), and binding to the MAGE-A1₂₇₈ HLA-A2-restricted p/HLA multimer (targeting ≥ 30% of live cells).

Patient infusion product phenotype staining and intracellular cytokine staining

The patient T_{TCR-MA1-CD8αβ} were analyzed by flow cytometry using p/HLA-multimer-APC, anti-CD8-APC/Fire 750 (1:200, SK1, BioLegend, 344746), anti-CD4-PerCP-E710 (1:200, SK3, Invitrogen, 46004742), anti-CD28-BV605 (1:100, CD28.2, BioLegend, 302967), anti-CD127(IL7RA)-BV711 (1:50, A019D5, BioLegend, 351327), and anti-CD62L-BV785 (1:200, DREG-56, BioLegend, 304829) with Brilliant stain buffer (BD Bioscience, 563794). Followed by the ICS protocol as previously described.

Transgenic T cell tracking by p/HLA multimer

Patient PBMCs collected throughout the indicated timepoints after the first and second infusions were thawed and allowed to rest in RPMI-1640 supplemented with 10% CTS for 2 h. Cells were analyzed by flow cytometry by staining with anti-CD3-APC-Cy7 (1:200, SK7, BD, 561800), anti-CD8-FITC, and Live/Dead-BV510 stain.

MAGE A1 detection by Immunohistochemistry

Anti-MAGE A1 immunohistochemistry staining was performed on 4 µm sections of formalin fixed paraffin embedded tumor biopsies. Sections were baked at 60-65 °C for a minimum of 30 minutes followed by routine deparaffinization using xylenes and graded ethanol. Antigen retrieval was performed using a Tris-EDTA-based buffer at 95 °C for 20 min. After blocking for endogenous peroxidase and nonspecific proteins, antibody incubation was performed at room temperature using anti-MAGE A1 mouse monoclonal antibody MA454 (Invitrogen, MA454) at 8 µg/ml for 30 min. Visualization was performed using horseradish peroxidase followed by DAB substrate and a hematoxylin counterstain.

Flow cytometry for in vivo murine models

All in vivo processing of the infusion product T-TCR_MA1 were stained with anti-CD19-BV510 (1:200, HIB19, BioLegend, 302242), Live/Dead-BV510, anti-hCD45-PE (1:200, 2D1, BioLegend, 368510), anti-mCD45-PE-Cy5 (1:200, 30-F11, BioLegend, 103110), anti-CD3-PerCP-Cy5.5 (1:200, SK7, BioLegend, 344808), anti-CD4-APC-Cy7 (1:200, RPA-T4, BioLegend, 300518), anti-CD8-AF700 (1:200, RPA-T8, BioLegend, 301027), anti-Tim3-BV650 (1:200, F38-2E2, BioLegend, 345028), anti-PD-1-BV785 (1:200, EH12.2H7, BioLegend, 329930), anti-CD39-BV605 (1:200 dilution, A1, BioLegend, 328214), anti-TIGIT-BV421 (1:200 dilution, A15153G, BioLegend, 372712),or replaced anti-CD8-AF700 with anti-2B4-AF700 (1:200 dilution, C1.7, Biolegend, 329525) and anti-CD8-BUV737(1:200, RPA-T8, Invitrogen, 367008842) to identify exhaustion phenotype.

In vivo tumor immunohistochemistry

Formalin-fixed paraffin-embedded tissues, cut at 5 µm, were baked for 1 h at 65 °C. Slides were dewaxed and stained on a Leica BOND RX stainer (Leica, Buffalo Grove, IL) using Leica Bond reagents for dewaxing (Dewax Solution), antigen retrieval/antibody stripping (Epitope Retrieval Solution 2), and rinsing after each step (Bond Wash Solution). Antigen retrieval and antibody stripping steps were performed at 100 °C, with all other steps at ambient temperature.

Endogenous peroxidase was blocked with 3% H₂O₂ for 5 min followed by protein blocking with TCT buffer (0.05 M Tris, 0.15 M NaCl, 0.25% Casein, 0.1% Tween-20, 0.05% ProClin300 pH 7.6) for 10 minutes. Anti-CD8(Agilent #M710301-2, 1:800 dilution) was applied for 60 minutes, followed by the secondary antibody (1X Opal Anti-Ms + Rb HRP Polymer, Akoya Biosciences) for 20 minutes, then the tertiary TSA-amplification reagent (Akoya Biosciences OPAL fluor) for 20 minutes. A high-stringency wash was performed after the secondary and tertiary applications using a high-salt TBS-T solution (0.05 M Tris, 0.3 M NaCl, and 0.1% Tween-20, pH 7.2–7.6).

The primary and secondary antibodies were stripped with retrieval solution for 20 minutes before repeating the process with anti-CD4 (Sigma Aldrich #AC173, 1:100 dilution), starting with a new application of 3% H₂O₂. The stripping step was not performed after the final position.

Slides were removed from the stainer and stained with a 5µg/mL concentration of DAPI (Sigma, D8417) for 5 min, rinsed, and a cover slip was placed over it along with Prolong Gold Antifade reagent (Invitrogen/Life Technologies). After curing at room temperature, slide images were acquired on the Aperio Scanscope FL. Images were analyzed using HALO software (Indica Labs).

TCR tonic signaling

Jurkat triple reporter (TPR) as described by Rosskopft et al.⁶⁸ were used to identify tonic signaling on day 5 post-transduction with lentiviral construct, including T_{TCR-MA1-CD8ab}, T_{TCR-MA1-CD8/28}, and CD19-CD28z CAR constructs. On day 5 post-transduction, Jurkat TPR were stained with p/HLA-multimer-APC for MAGE constructs Jurkat TPR and EGFR-APC (1:200, AY13, BioLegend, 352905) for the CD19-CD28 CAR-T Jurkat TPR. Reporter fluorescence was analyzed by flow cytometry.

Single cell RNA sequencing

Cryopreserved tumor single-cell suspension from in vivo murine experiment were thawed, washed and processed using a 10x Chromium Controller following the 5’ Chromium Single Cell V(D)J Reagent Kit manual (10x Genomics). Library preparation was carried out according to the manufacturer’s protocol without modifications. Library quality was assessed using TapeStation High Sensitivity (Agilent) to evaluate library size, Qubit (Thermo Fisher) to measure dsDNA quantity, and KAPA qPCR analysis (KAPA Biosystems) to determine the quantity of amplifiable transcript.

Samples were pooled in equimolar proportions and sequenced on an Illumina HiSeq 2500 in rapid run mode, following the standard 10x Genomics protocol. TCR target enrichment, 5’ gene expression library preparation, and TCR library construction were performed according to the 5’ Chromium Single Cell V(D)J Reagent Kit manual (10x Genomics).

Raw FASTAQ files were processed using the 10x Genomics Cell Ranger software (v7.1.0) with default parameters. The EmptyDrops⁶⁹ method was applied to filter out cells with low RNA content. The “count” function was used for alignment, filtering, barcode counting, and UMI counting. Reads were aligned to the human reference genome (hg38, Ensembl) and the codon-optimized transgene sequence using Spliced Transcripts Alignment to a Reference (STAR)⁷⁰.

The filtered feature matrices generated by the CellRanger pipeline were used for downstream quality control (QC) and analyses. We used the function read10xCounts from the R package DropletUtils (version 1.14.2) to load the CellRanger output in R as a SingleCellExperiment⁴⁶ object. Doublet cells filtering was performed on each sample using the scds package (v1.10.0)⁴⁷. QC and filtering were conducted using the scater R package (v1.22.0)⁴⁸. Genes not detected across all the cells were removed and cells were filtered based on feature counts, the percentage of mitochondrial and ribosomal genes, and the number of expressed features. Cells with values beyond the median absolute deviations (MAD) specific threshold of 2.5 from the median were excluded. Features with a count greater than 1 in at least 3 cells were retained for downstream analysis. We then normalized, found the 2000 most variable genes and scaled for each dataset using Seurat (v4.3.0.9001) NormalizeData, FindVariableFeatures and ScaleData respectively^49,50. Dimensionality reduction (PCA and UMAP) was computed using RunPCA and RunUMAP, respectively. UMAP dimension reduction and clustering were computed using the first 20 principal components (PCs). The number of PCs capturing most of the variation in our data was selected using Seurat function ElbowPlot which visualizes the standard deviation of each PC. Clusters were identified via shared-nearest-neighbor-based (SNN) clustering and further analyzed at a resolution of 0.4. CD8 and CD4 cell subsets were identified using scGate (v1.0.1)⁵¹ and subsequently extracted with Seurat subset function. Differential gene expression was performed using FindAllMarkers Seurat function. ComplexHeatmap (https://github.com/jokergoo/ComplexHeatmap) and SCP were used for visualization purposes.

T_TCR-WT1 and T_TCR-PRAME cells generation

Activated CD4 and CD8 T cells were transduced with lentivirus (see above methods) containing T-TCR_WT1 (recognize WT1_37-45) and T-TCR_PRAME (recognize PRAME_425-433) constructs, both with CD8αβ. T_{TCR-WT1-CD8/CD28}, T_{TCR-WT1-CD8/CD28mu}, T_{TCR-PRAME-CD8/CD28}, T_{TCR-PRAME-CD8/CD28mut} were generated with the protocol described above.

Phospho-flow cytometry analysis

Expanded TTCR-MA1-CD8/28mut, TTCR-MA1-CD8/28, and TTCR-MA1-CD8αβ CD4 T cells were seeded at 1 × 106 in 50 µl in a 96-well plate and rested for one hour in a 37 °C incubator with complete RPMI-1640 media. 1 µM of cognate peptide[loaded on TAP-deficient T2 B lymphoblastoid cell lines (T2-BLCL)] were generated in small volume for stimulation. Cells were stimulated by adding 2 × 105 of the peptide-loaded T2-BLCL in 50µl at three different time points (0, 5, 10 minutes for p-AKT and 0, 10, 20 minutes for p-ERK). At the end of the stimulation, 100 µl of 4% of warmed PFA were added directly into the wells and incubated in 37 °C for 20 mins. Cells were then washed with PBS twice and permeabilized with BD Phospho-flow PERM buffer (BD, 558050) followed by the manufacturer’s protocol for overnight permeabilization. Fix and permeabilized cells were washed with FACS buffer and stained with anti-(p)-ERK-AF700 (1:50, Monoclonal Rabbit IgG Clone #269434, R and D System, IC1018N-100UG) and anti-(p)-AKT-Alexa Flour 647 (1:50, A21001C, BioLegend, 606556) for 30 mins under room temperature followed by two washes with FACS buffer and analyzed on flow cytometer.

Bulk RNA sequencing

Peptide stimulation assay

CD4 and CD8 T_TCR-MA1 were seeded at 1 × 10⁶ cells per well in 24-well flat-bottom plates and stimulated with cognate peptide for 4 h prior to preparation for bulk RNA sequencing (bulk RNAseq). Raw reads (FASTQs) were trimmed of their adapters using cutadapt (v2.9) and subsequently quality-checked before and after adapter trimming with FastQC (v0.11.9). Trimmed reads were aligned and quantified via STAR (v2.7.7a). Differential gene expression was computed by the Mann-Whitney U-test. Gene set enrichment analysis (GSEA) was run based on log fold change of RPKM values with the following analyses: cytokine pathways (Biocarta), cytokine activity (GOMF), cytokine and inflammatory response (WP), and the Th1/Th2 pathway (Biocarta).

Irradiated tumor rechallenge assay

CD4 and CD8 T_TCR-MA1 were seeded and cocultured at 1 × 10⁶ cells per well in 24-well flat-bottom plates with irradiated A375F (60 Gy) or ME275 (80 Gy) at an E:T ratio of 1:1 or 2:1, respectively, for 21 days. T cells were analyzed by flow cytometry with anti-IFNγ-APC, anti-TNFα-PE-Cy7, anti-PD-1-BV421(1:200, EH12.2H7, BioLegend, 329919), and anti-Tim-3-FITC (1:200, F38-2E2, BioLegend, 345021). The cells were counted and 1 × 10⁶ cells were transferred to a new well with the same E:T ratio of tumor cells to rechallenge T cells. Endpoint was defined as less than 1 × 10⁶ total cells remaining and were cryostored for subsequent bulk RNAseq and analysis. Bulk RNA sequencing was performed using BGISEQ-500 platform at BGI Genomics. Briefly, total RNA was extracted using the Qiagen RNeasy Micro Kit according to the manufacturer’s instructions. For the construction of low input polyA mRNA-seq libraries, the SMARTseq (v4) Package was used. Sequencing was performed on a DNBseq T7 machine (MGI) with paired-end 150 bp reads, generating 30 M raw reads per sample. Raw sequencing data were filtered and trimmed using the software Soapnuke developed by BGI Genomics. The filtered reads were then aligned to the reference transcriptome using Bowtie2 (v2.2.5). Gene read counts were subsequently generated from the alignment results using RSEM (v1.2.8). To identify differentially expressed genes across conditions while accounting for potential batch affect, we used the likelihood ratio test (LRT) with a full model of ~Batch+Condition + Stim and a reduced model of ~Batch. Raw count data were filtered to include genes with a minimum count of five across samples using the filter_counts function from the hciR package (v1.7). Normalized counts were obtained using the regularized log (rlog) transformation, and differential expression analysis was performed using the DESeq2 package (v1.42.0). Significant genes were identified by filtering results to retain only those with an adjusted p-value (padj) <0.05 and an absolute log2 fold-change > 0.5. The top 200 differentially expressed genes were visualized using plot_genes from hciR. All analyses and visualizations were conducted in R.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(53.5MB, pdf)}

Reporting Summary^{(125.6KB, pdf)}

Transparent Peer Review file^{(4MB, pdf)}

Source data

Source Data^{(263.8KB, xlsx)}

Acknowledgements

We received funding from the Emerson Collective (2021-2023), the National Institute of Health (P01CA225517 and P01CA18029-41), the Immunotherapy Integrated Research Center at the Fred Hutchinson Cancer Research Center, and Elevate Bio. The clinical study (NCT04639245) was investigator-initiated and conducted under an IND held by the FHCC. We thank Regeneron for the use of MISTRG mouse model for our in vivo experiments (Regeneron Contract No.: 2025_089439 Fred Hutch Ref: MTA250605). We thank FHCC investigators and the Cellular Immunotherapy Clinical Operations unit for participating in study design, patient recruitment, clinical care and data collection. We thank ElevateBio for providing financial support for the trial. Collection of human and murine data was supported by the Comparative Medicine Shared Resource (RRID: SCR_022610), Experimental Histopathology Shared Resource (RRID: SCR_022612), Flow Cytometry Shared Resource (RRID: SCR_022613), Genomics & Bioinformatics Shared Resource (RRID: SCR_022606), Immune Monitoring Shared Resource (RRID: SCR_022615) of the Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium (P30 CA015704).

Author contributions

Conceptualization: S.Z., T.M.S., M.M, A.G.C.; Methodology: S.Z, T.T., A.M., M.M.; Investigation: S.Z., T.T., S.K., F.M., M.T.S., A.M., Y.S., V.V., L.M., C.W.S., Y.A., M.H., J.B., B.L., C.Y., H.C., A.M.R., D.G.C., K.F., N.H., T.Z., P.L., K.M., S.K.O.; Formal analysis: S.Z, T.T., S.K., F.M., Y.S., V.V., D.G.C., S.K.O.; Resources: A.R., P.D.G., T.M.S, M.M, A.G.C.; Writing- original draft: T.T., S.K., A.G.C.; Writing -review and editing: T.T., S.K., F.M., Y.S., C.W.S., B.L., A.R., P.D.G., T.M.S., A.G.C.; Supervision: P.D.G., T.M.S., A.G.C. Funding acquisition: A.R., P.D.G., T.M.S., M.M, A.G.C. First author contribution: S.Z., T.T., and S.K. contributed equally, order reflects chronological involvement, with T.T. carrying out the majority of the revision experiments and manuscript completion.

Peer review

Peer review information

Nature Communications thanks Giulia Casorati and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All data associated with this paper are in the paper, or supplementary materials. The clinical protocol, including the statistical plan and consent, is listed at the end of the supplementary materials. All requests for raw data, code, and materials are promptly reviewed by the Fred Hutchinson Cancer Research Center to verify if the request is subject to any intellectual property or confidential obligations. Constructs are available from Fred Hutchinson Cancer Research Center under a material transfer agreement for preclinical research purposes only. Patient-related data not included in the paper were generated as part of a clinical trial and may be subject to patient confidentiality. Any data and materials that can be shared will be released via a Material Transfer Agreement. The sc-RNA gene expression data, including count matrices and raw sequencing data, generated in this study have been deposited into the NCBI Gene Expression Omnibus (GEO) under accession number GSE288522.(https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE288522). Previously published solid tumor scRNAseq data are available in the GEO repository under accession number GSE215121 (melanoma), and GSE148071 (lung). All data are included in the Supplementary Information or available from the authors, as are unique reagents used in this Article. The raw numbers for charts and graphs are available in the Source Data file whenever possible. Source data are provided with this paper.

Code availability

Data was generated using R. Data now available on GEO with accession number GSE288522.

Competing interests

A.G.C. has received support from Elevate Bio, Juno Therapeutics, Lonza, and Affini-T. P.D.G. is on the Scientific Advisory Board of Celsius, Earli, Elpiscience, Immunoscape, Rapt, and Nextech, was a scientific founder of Juno Therapeutics, and receives research support from Lonza. A.G.C., T.M.S., and P.D.G. are co-founders of, have equity in, and receive research support from Affini-T. A.G.C., T.M.S., and M.M. are listed as inventors on patents related to MA1 TCR and CD8/28 constructs. The remaining authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Shihong Zhang, Tzu-Hao Tang, Sinéad Kinsella.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-67446-5.

References

1.Klebanoff, C. A., Chandran, S. S., Baker, B. M., Quezada, S. A. & Ribas, A. T cell receptor therapeutics: immunological targeting of the intracellular cancer proteome. Nat. Rev. Drug Discov.22, 996–1017 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Dougé, A., El Ghazzi, N., Lemal, R. & Rouzaire, P. Adoptive T cell therapy in solid tumors: state-of-the art, current challenges, and upcoming improvements. Mol. Cancer Ther.23, 272–284 (2024). [DOI] [PubMed] [Google Scholar]
3.Leidner, R. et al. Neoantigen T-Cell Receptor Gene Therapy in Pancreatic Cancer. N. Engl. J. Med386, 2112–2119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Hong, D. S. et al. Autologous T cell therapy for MAGE-A4(+) solid cancers in HLA-A*02(+) patients: a phase 1 trial. Nat. Med29, 104–114 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Yan, T., Zhu, L. & Chen, J. Current advances and challenges in CAR T-Cell therapy for solid tumors: tumor-associated antigens and the tumor microenvironment. Exp. Hematol. Oncol.12, 14 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Liu, Z. et al. Immunosuppression in tumor immune microenvironment and its optimization from CAR-T cell therapy. Theranostics12, 6273–6290 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Lopresti, L., Tatangelo, V., Baldari, C. T. & Patrussi, L. Rewiring the T cell-suppressive cytokine landscape of the tumor microenvironment: a new frontier for precision anti-cancer therapy. Front Immunol.15, 1418527 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Tufail, M., Jiang, C. H. & Li, N. Immune evasion in cancer: mechanisms and cutting-edge therapeutic approaches. Signal Transduct. Target Ther.10, 227 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Alspach, E. et al. MHC-II neoantigens shape tumour immunity and response to immunotherapy. Nature574, 696–701 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Turtle, C. J. et al. CD19 CAR-T cells of defined CD4+:CD8+ composition in adult B cell ALL patients. J. Clin. Invest126, 2123–2138 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Wooldridge, L. et al. Interaction between the CD8 coreceptor and major histocompatibility complex class I stabilizes T cell receptor-antigen complexes at the cell surface. J. Biol. Chem.280, 27491–27501 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Lahman, M. C. et al. Targeting an alternate Wilms’ tumor antigen 1 peptide bypasses immunoproteasome dependency. Sci. Transl. Med14, eabg8070 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Anderson, V. E. et al. Enhancing efficacy of TCR-engineered CD4 + T cells via coexpression of CD8alpha. J. Immunother.46, 132–144 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Laletin, V., Bernard, P. L., Costa da Silva, C., Guittard, G. & Nunes, J. A. Negative intracellular regulators of T-cell receptor (TCR) signaling as potential antitumor immunotherapy targets. J. Immunother. Cancer1110.1136/jitc-2022-005845 (2023). [DOI] [PMC free article] [PubMed]
15.Driessens, G., Kline, J. & Gajewski, T. F. Costimulatory and coinhibitory receptors in anti-tumor immunity. Immunol. Rev.229, 126–144 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.van der Leun, A. M., Thommen, D. S. & Schumacher, T. N. CD8(+) T cell states in human cancer: insights from single-cell analysis. Nat. Rev. Cancer20, 218–232 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Sharpe, A. H. Mechanisms of costimulation. Immunol. Rev.229, 5–11 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Oda, S. K. et al. A Fas-4-1BB fusion protein converts a death to a pro-survival signal and enhances T cell therapy. J. Exp. Med.21710.1084/jem.20191166 (2020). [DOI] [PMC free article] [PubMed]
19.Oda, S. K. et al. A CD200R-CD28 fusion protein appropriates an inhibitory signal to enhance T-cell function and therapy of murine leukemia. Blood130, 2410–2419 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Guo, J., Kent, A. & Davila, E. Chimeric non-antigen receptors in T cell-based cancer therapy. J. Immunother. Cancer910.1136/jitc-2021-002628 (2021). [DOI] [PMC free article] [PubMed]
21.Pascolo, S. et al. A MAGE-A1 HLA-A A*0201 epitope identified by mass spectrometry. Cancer Res61, 4072–4077 (2001). [PubMed] [Google Scholar]
22.Frohman, M. A. Rapid amplification of complementary DNA ends for generation of full-length complementary DNAs: thermal RACE. Methods Enzymol.218, 340–356 (1993). [DOI] [PubMed] [Google Scholar]
23.Chapuis, A. G. et al. T cell receptor gene therapy targeting WT1 prevents acute myeloid leukemia relapse post-transplant. Nat. Med25, 1064–1072 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Davari, K. et al. Development of a CD8 co-receptor independent T-cell receptor specific for tumor-associated antigen MAGE-A4 for next generation T-cell-based immunotherapy. J. Immunother. Cancer910.1136/jitc-2020-002035 (2021). [DOI] [PMC free article] [PubMed]
25.Caballero, O. L. et al. Frequent MAGE mutations in human melanoma. PLoS One510.1371/journal.pone.0012773 (2010). [DOI] [PMC free article] [PubMed]
26.Eisenhauer, E. A. et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur. J. Cancer45, 228–247 (2009). [DOI] [PubMed] [Google Scholar]
27.Zhou, Q. et al. Coexpression of Tim-3 and PD-1 identifies a CD8+ T-cell exhaustion phenotype in mice with disseminated acute myelogenous leukemia. Blood117, 4501–4510 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Canale, F. P. et al. CD39 expression defines cell exhaustion in tumor-infiltrating CD8(+) T Cells. Cancer Res78, 115–128 (2018). [DOI] [PubMed] [Google Scholar]
29.Liu, D. et al. 2B4 (CD244) induced by selective CD28 blockade functionally regulates allograft-specific CD8+ T cell responses. J. Exp. Med211, 297–311 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Esensten, J. H., Helou, Y. A., Chopra, G., Weiss, A. & Bluestone, J. A. CD28 Costimulation: From Mechanism to Therapy. Immunity44, 973–988 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Kakarla, S. & Gottschalk, S. CAR T cells for solid tumors: armed and ready to go?. Cancer J.20, 151–155 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Zhang, C. et al. A single-cell analysis reveals tumor heterogeneity and immune environment of acral melanoma. Nat. Commun.13, 7250 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Wu, F. et al. Single-cell profiling of tumor heterogeneity and the microenvironment in advanced non-small cell lung cancer. Nat. Commun.12, 2540 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Irving, B. A. & Weiss, A. The cytoplasmic domain of the T cell receptor zeta chain is sufficient to couple to receptor-associated signal transduction pathways. Cell64, 891–901 (1991). [DOI] [PubMed] [Google Scholar]
35.Kaczanowska, S. et al. A synthetic CD8 alpha:MyD88 coreceptor enhances CD8(+) T-cell responses to weakly immunogenic and lowly expressed tumor antigens. Cancer Res.77, 7049–7058 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Stein, P. H., Fraser, J. D. & Weiss, A. The cytoplasmic domain of CD28 is both necessary and sufficient for costimulation of interleukin-2 secretion and association with phosphatidylinositol 3’-kinase. Mol. Cell Biol.14, 3392–3402 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Fragoso, R. et al. Lipid raft distribution of CD4 depends on its palmitoylation and association with Lck, and evidence for CD4-induced lipid raft aggregation as an additional mechanism to enhance CD3 signaling. J. Immunol.170, 913–921 (2003). [DOI] [PubMed] [Google Scholar]
38.Hwang, J. R., Byeon, Y., Kim, D. & Park, S. G. Recent insights of T cell receptor-mediated signaling pathways for T cell activation and development. Exp. Mol. Med52, 750–761 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Rybakin, V., Clamme, J. P., Ampudia, J., Yachi, P. P. & Gascoigne, N. R. CD8αα and -αβ isotypes are equally recruited to the immunological synapse through their ability to bind to MHC class I. EMBO Rep.12, 1251–1256 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Cole, D. K. et al. The molecular determinants of CD8 co-receptor function. Immunology137, 139–148 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Sun, F. et al. The MAPK dual specific phosphatase (DUSP) proteins: A versatile wrestler in T cell functionality. Int Immunopharmacol.98, 107906 (2021). [DOI] [PubMed] [Google Scholar]
42.Joller, N., Anderson, A. C. & Kuchroo, V. K. LAG-3, TIM-3, and TIGIT: Distinct functions in immune regulation. Immunity57, 206–222 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Gros, A. et al. PD-1 identifies the patient-specific CD8+ tumor-reactive repertoire infiltrating human tumors. J. Clin. Investig.124, 2246–2259 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Inozume, T. et al. Selection of CD8+PD-1+ lymphocytes in fresh human melanomas enriches for tumor-reactive T cells. J. Immunother.33, 956–964 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Mietz, J. et al. Human effector CD8(+) T cells with an activated and exhausted-like phenotype control tumour growth in vivo in a humanized tumour model. EBioMedicine106, 105240 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Boucher, J. C. et al. CD28 costimulatory domain-targeted mutations enhance chimeric antigen receptor t-cell function. Cancer Immunol. Res9, 62–74 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Guedan, S. et al. Single residue in CD28-costimulated CAR-T cells limits long-term persistence and antitumor durability. J. Clin. Invest130, 3087–3097 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Nguyen, P., Moisini, I. & Geiger, T. L. Identification of a murine CD28 dileucine motif that suppresses single-chain chimeric T-cell receptor expression and function. Blood102, 4320–4325 (2003). [DOI] [PubMed] [Google Scholar]
49.Kofler, D. M. et al. CD28 costimulation Impairs the efficacy of a redirected t-cell antitumor attack in the presence of regulatory t cells which can be overcome by preventing Lck activation. Mol. Ther.19, 760–767 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Tortelli, T. C. J. et al. Accumulation of prohibitin is a common cellular response to different stressing stimuli and protects melanoma cells from ER stress and chemotherapy-induced cell death. Oncotarget8, 43114–43129 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Adachi, K. & Davis, M. M. T-cell receptor ligation induces distinct signaling pathways in naive vs. antigen-experienced T cells. Proc. Natl. Acad. Sci. USA108, 1549–1554 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Drougkas, K. et al. Comprehensive clinical evaluation of CAR-T cell immunotherapy for solid tumors: a path moving forward or a dead end?. J. Cancer Res Clin. Oncol.149, 2709–2734 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Yarza, R. et al. Efficacy of T-cell receptor-based adoptive cell therapy in cutaneous melanoma: a meta-analysis. Oncologist28, e406–e415 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Tsimberidou, A. M. et al. Feasibility and safety of personalized, multi-target, adoptive cell therapy (ima101): first-in-human clinical trial in patients with advanced metastatic cancer. Cancer Immunol. Res.11, 925–945 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Thibodeau, J., Bourgeois-Daigneault, M. C. & Lapointe, R. Targeting the MHC Class II antigen presentation pathway in cancer immunotherapy. Oncoimmunology1, 908–916 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Ndejembi, M. P. et al. Control of memory CD4 T cell recall by the CD28/B7 costimulatory pathway. J. Immunol.177, 7698–7706 (2006). [DOI] [PubMed] [Google Scholar]
57.Salter, A. I. et al. Phosphoproteomic analysis of chimeric antigen receptor signaling reveals kinetic and quantitative differences that affect cell function. Sci. Signal1110.1126/scisignal.aat6753 (2018). [DOI] [PMC free article] [PubMed]
58.Parry, R. V. et al. Ligation of the T cell co-stimulatory receptor CD28 activates the serine-threonine protein kinase protein kinase B. Eur. J. Immunol.27, 2495–2501 (1997). [DOI] [PubMed] [Google Scholar]
59.Holdorf, A. D. et al. Proline residues in CD28 and the Src homology (SH)3 domain of Lck are required for T cell costimulation. J. Exp. Med190, 375–384 (1999). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Boomer, J. S., Deppong, C. M., Shah, D. D., Bricker, T. L. & Green, J. M. Cutting edge: A double-mutant knockin of the CD28 YMNM and PYAP motifs reveals a critical role for the YMNM motif in regulation of T cell proliferation and Bcl-xL expression. J. Immunol.192, 3465–3469 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Gogishvili, T. et al. Proliferative signals mediated by CD28 superagonists require the exchange factor Vav1 but not phosphoinositide 3-kinase in primary peripheral T cells. Eur. J. Immunol.38, 2528–2533 (2008). [DOI] [PubMed] [Google Scholar]
62.Jenkins, E., Whitehead, T., Fellermeyer, M., Davis, S. J. & Sharma, S. The current state and future of T-cell exhaustion research. Oxf. Open Immunol.4, iqad006 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Chiorazzi, M. et al. Autologous humanized PDX modeling for immuno-oncology recapitulates features of the human tumor microenvironment. J. Immunother. Cancer1110.1136/jitc-2023-006921 (2023). [DOI] [PMC free article] [PubMed]
64.Rongvaux, A. et al. Development and function of human innate immune cells in a humanized mouse model. Nat. Biotechnol.32, 364–372 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Chapuis, A. G. et al. Transferred WT1-reactive CD8+ T cells can mediate antileukemic activity and persist in post-transplant patients. Sci. Transl. Med.5, 174ra127 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Schmitt, T. M. et al. Abstract A51: High-throughput method identifies rare, high-affinity, thymus-vetted T cell receptors (TCRs) for clinical translation. Cancer Immunol. Res.6, A51–A51 (2018). [Google Scholar]
67.National Cancer Institute, N. I. o. H. (Cancer Therapy Evaluation Program, 2009).
68.Rosskopf, S. et al. A Jurkat 76 based triple parameter reporter system to evaluate TCR functions and adoptive T cell strategies. Oncotarget9, 17608–17619 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
69.Lun, A. T. L. et al. EmptyDrops: distinguishing cells from empty droplets in droplet-based single-cell RNA sequencing data. Genome Biol.20, 63–63 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
70.Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics29, 15–21 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(53.5MB, pdf)}

Reporting Summary^{(125.6KB, pdf)}

Transparent Peer Review file^{(4MB, pdf)}

Source Data^{(263.8KB, xlsx)}

Data Availability Statement

Data was generated using R. Data now available on GEO with accession number GSE288522.

[CR1] 1.Klebanoff, C. A., Chandran, S. S., Baker, B. M., Quezada, S. A. & Ribas, A. T cell receptor therapeutics: immunological targeting of the intracellular cancer proteome. Nat. Rev. Drug Discov.22, 996–1017 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Dougé, A., El Ghazzi, N., Lemal, R. & Rouzaire, P. Adoptive T cell therapy in solid tumors: state-of-the art, current challenges, and upcoming improvements. Mol. Cancer Ther.23, 272–284 (2024). [DOI] [PubMed] [Google Scholar]

[CR3] 3.Leidner, R. et al. Neoantigen T-Cell Receptor Gene Therapy in Pancreatic Cancer. N. Engl. J. Med386, 2112–2119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Hong, D. S. et al. Autologous T cell therapy for MAGE-A4(+) solid cancers in HLA-A*02(+) patients: a phase 1 trial. Nat. Med29, 104–114 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Yan, T., Zhu, L. & Chen, J. Current advances and challenges in CAR T-Cell therapy for solid tumors: tumor-associated antigens and the tumor microenvironment. Exp. Hematol. Oncol.12, 14 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Liu, Z. et al. Immunosuppression in tumor immune microenvironment and its optimization from CAR-T cell therapy. Theranostics12, 6273–6290 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Lopresti, L., Tatangelo, V., Baldari, C. T. & Patrussi, L. Rewiring the T cell-suppressive cytokine landscape of the tumor microenvironment: a new frontier for precision anti-cancer therapy. Front Immunol.15, 1418527 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Tufail, M., Jiang, C. H. & Li, N. Immune evasion in cancer: mechanisms and cutting-edge therapeutic approaches. Signal Transduct. Target Ther.10, 227 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Alspach, E. et al. MHC-II neoantigens shape tumour immunity and response to immunotherapy. Nature574, 696–701 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Turtle, C. J. et al. CD19 CAR-T cells of defined CD4+:CD8+ composition in adult B cell ALL patients. J. Clin. Invest126, 2123–2138 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Wooldridge, L. et al. Interaction between the CD8 coreceptor and major histocompatibility complex class I stabilizes T cell receptor-antigen complexes at the cell surface. J. Biol. Chem.280, 27491–27501 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Lahman, M. C. et al. Targeting an alternate Wilms’ tumor antigen 1 peptide bypasses immunoproteasome dependency. Sci. Transl. Med14, eabg8070 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Anderson, V. E. et al. Enhancing efficacy of TCR-engineered CD4 + T cells via coexpression of CD8alpha. J. Immunother.46, 132–144 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Laletin, V., Bernard, P. L., Costa da Silva, C., Guittard, G. & Nunes, J. A. Negative intracellular regulators of T-cell receptor (TCR) signaling as potential antitumor immunotherapy targets. J. Immunother. Cancer1110.1136/jitc-2022-005845 (2023). [DOI] [PMC free article] [PubMed]

[CR15] 15.Driessens, G., Kline, J. & Gajewski, T. F. Costimulatory and coinhibitory receptors in anti-tumor immunity. Immunol. Rev.229, 126–144 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.van der Leun, A. M., Thommen, D. S. & Schumacher, T. N. CD8(+) T cell states in human cancer: insights from single-cell analysis. Nat. Rev. Cancer20, 218–232 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Sharpe, A. H. Mechanisms of costimulation. Immunol. Rev.229, 5–11 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Oda, S. K. et al. A Fas-4-1BB fusion protein converts a death to a pro-survival signal and enhances T cell therapy. J. Exp. Med.21710.1084/jem.20191166 (2020). [DOI] [PMC free article] [PubMed]

[CR19] 19.Oda, S. K. et al. A CD200R-CD28 fusion protein appropriates an inhibitory signal to enhance T-cell function and therapy of murine leukemia. Blood130, 2410–2419 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Guo, J., Kent, A. & Davila, E. Chimeric non-antigen receptors in T cell-based cancer therapy. J. Immunother. Cancer910.1136/jitc-2021-002628 (2021). [DOI] [PMC free article] [PubMed]

[CR21] 21.Pascolo, S. et al. A MAGE-A1 HLA-A A*0201 epitope identified by mass spectrometry. Cancer Res61, 4072–4077 (2001). [PubMed] [Google Scholar]

[CR22] 22.Frohman, M. A. Rapid amplification of complementary DNA ends for generation of full-length complementary DNAs: thermal RACE. Methods Enzymol.218, 340–356 (1993). [DOI] [PubMed] [Google Scholar]

[CR23] 23.Chapuis, A. G. et al. T cell receptor gene therapy targeting WT1 prevents acute myeloid leukemia relapse post-transplant. Nat. Med25, 1064–1072 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Davari, K. et al. Development of a CD8 co-receptor independent T-cell receptor specific for tumor-associated antigen MAGE-A4 for next generation T-cell-based immunotherapy. J. Immunother. Cancer910.1136/jitc-2020-002035 (2021). [DOI] [PMC free article] [PubMed]

[CR25] 25.Caballero, O. L. et al. Frequent MAGE mutations in human melanoma. PLoS One510.1371/journal.pone.0012773 (2010). [DOI] [PMC free article] [PubMed]

[CR26] 26.Eisenhauer, E. A. et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur. J. Cancer45, 228–247 (2009). [DOI] [PubMed] [Google Scholar]

[CR27] 27.Zhou, Q. et al. Coexpression of Tim-3 and PD-1 identifies a CD8+ T-cell exhaustion phenotype in mice with disseminated acute myelogenous leukemia. Blood117, 4501–4510 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Canale, F. P. et al. CD39 expression defines cell exhaustion in tumor-infiltrating CD8(+) T Cells. Cancer Res78, 115–128 (2018). [DOI] [PubMed] [Google Scholar]

[CR29] 29.Liu, D. et al. 2B4 (CD244) induced by selective CD28 blockade functionally regulates allograft-specific CD8+ T cell responses. J. Exp. Med211, 297–311 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Esensten, J. H., Helou, Y. A., Chopra, G., Weiss, A. & Bluestone, J. A. CD28 Costimulation: From Mechanism to Therapy. Immunity44, 973–988 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Kakarla, S. & Gottschalk, S. CAR T cells for solid tumors: armed and ready to go?. Cancer J.20, 151–155 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Zhang, C. et al. A single-cell analysis reveals tumor heterogeneity and immune environment of acral melanoma. Nat. Commun.13, 7250 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Wu, F. et al. Single-cell profiling of tumor heterogeneity and the microenvironment in advanced non-small cell lung cancer. Nat. Commun.12, 2540 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Irving, B. A. & Weiss, A. The cytoplasmic domain of the T cell receptor zeta chain is sufficient to couple to receptor-associated signal transduction pathways. Cell64, 891–901 (1991). [DOI] [PubMed] [Google Scholar]

[CR35] 35.Kaczanowska, S. et al. A synthetic CD8 alpha:MyD88 coreceptor enhances CD8(+) T-cell responses to weakly immunogenic and lowly expressed tumor antigens. Cancer Res.77, 7049–7058 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Stein, P. H., Fraser, J. D. & Weiss, A. The cytoplasmic domain of CD28 is both necessary and sufficient for costimulation of interleukin-2 secretion and association with phosphatidylinositol 3’-kinase. Mol. Cell Biol.14, 3392–3402 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Fragoso, R. et al. Lipid raft distribution of CD4 depends on its palmitoylation and association with Lck, and evidence for CD4-induced lipid raft aggregation as an additional mechanism to enhance CD3 signaling. J. Immunol.170, 913–921 (2003). [DOI] [PubMed] [Google Scholar]

[CR38] 38.Hwang, J. R., Byeon, Y., Kim, D. & Park, S. G. Recent insights of T cell receptor-mediated signaling pathways for T cell activation and development. Exp. Mol. Med52, 750–761 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Rybakin, V., Clamme, J. P., Ampudia, J., Yachi, P. P. & Gascoigne, N. R. CD8αα and -αβ isotypes are equally recruited to the immunological synapse through their ability to bind to MHC class I. EMBO Rep.12, 1251–1256 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Cole, D. K. et al. The molecular determinants of CD8 co-receptor function. Immunology137, 139–148 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Sun, F. et al. The MAPK dual specific phosphatase (DUSP) proteins: A versatile wrestler in T cell functionality. Int Immunopharmacol.98, 107906 (2021). [DOI] [PubMed] [Google Scholar]

[CR42] 42.Joller, N., Anderson, A. C. & Kuchroo, V. K. LAG-3, TIM-3, and TIGIT: Distinct functions in immune regulation. Immunity57, 206–222 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Gros, A. et al. PD-1 identifies the patient-specific CD8+ tumor-reactive repertoire infiltrating human tumors. J. Clin. Investig.124, 2246–2259 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Inozume, T. et al. Selection of CD8+PD-1+ lymphocytes in fresh human melanomas enriches for tumor-reactive T cells. J. Immunother.33, 956–964 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Mietz, J. et al. Human effector CD8(+) T cells with an activated and exhausted-like phenotype control tumour growth in vivo in a humanized tumour model. EBioMedicine106, 105240 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Boucher, J. C. et al. CD28 costimulatory domain-targeted mutations enhance chimeric antigen receptor t-cell function. Cancer Immunol. Res9, 62–74 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Guedan, S. et al. Single residue in CD28-costimulated CAR-T cells limits long-term persistence and antitumor durability. J. Clin. Invest130, 3087–3097 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Nguyen, P., Moisini, I. & Geiger, T. L. Identification of a murine CD28 dileucine motif that suppresses single-chain chimeric T-cell receptor expression and function. Blood102, 4320–4325 (2003). [DOI] [PubMed] [Google Scholar]

[CR49] 49.Kofler, D. M. et al. CD28 costimulation Impairs the efficacy of a redirected t-cell antitumor attack in the presence of regulatory t cells which can be overcome by preventing Lck activation. Mol. Ther.19, 760–767 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Tortelli, T. C. J. et al. Accumulation of prohibitin is a common cellular response to different stressing stimuli and protects melanoma cells from ER stress and chemotherapy-induced cell death. Oncotarget8, 43114–43129 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Adachi, K. & Davis, M. M. T-cell receptor ligation induces distinct signaling pathways in naive vs. antigen-experienced T cells. Proc. Natl. Acad. Sci. USA108, 1549–1554 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Drougkas, K. et al. Comprehensive clinical evaluation of CAR-T cell immunotherapy for solid tumors: a path moving forward or a dead end?. J. Cancer Res Clin. Oncol.149, 2709–2734 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Yarza, R. et al. Efficacy of T-cell receptor-based adoptive cell therapy in cutaneous melanoma: a meta-analysis. Oncologist28, e406–e415 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Tsimberidou, A. M. et al. Feasibility and safety of personalized, multi-target, adoptive cell therapy (ima101): first-in-human clinical trial in patients with advanced metastatic cancer. Cancer Immunol. Res.11, 925–945 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Thibodeau, J., Bourgeois-Daigneault, M. C. & Lapointe, R. Targeting the MHC Class II antigen presentation pathway in cancer immunotherapy. Oncoimmunology1, 908–916 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Ndejembi, M. P. et al. Control of memory CD4 T cell recall by the CD28/B7 costimulatory pathway. J. Immunol.177, 7698–7706 (2006). [DOI] [PubMed] [Google Scholar]

[CR57] 57.Salter, A. I. et al. Phosphoproteomic analysis of chimeric antigen receptor signaling reveals kinetic and quantitative differences that affect cell function. Sci. Signal1110.1126/scisignal.aat6753 (2018). [DOI] [PMC free article] [PubMed]

[CR58] 58.Parry, R. V. et al. Ligation of the T cell co-stimulatory receptor CD28 activates the serine-threonine protein kinase protein kinase B. Eur. J. Immunol.27, 2495–2501 (1997). [DOI] [PubMed] [Google Scholar]

[CR59] 59.Holdorf, A. D. et al. Proline residues in CD28 and the Src homology (SH)3 domain of Lck are required for T cell costimulation. J. Exp. Med190, 375–384 (1999). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Boomer, J. S., Deppong, C. M., Shah, D. D., Bricker, T. L. & Green, J. M. Cutting edge: A double-mutant knockin of the CD28 YMNM and PYAP motifs reveals a critical role for the YMNM motif in regulation of T cell proliferation and Bcl-xL expression. J. Immunol.192, 3465–3469 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Gogishvili, T. et al. Proliferative signals mediated by CD28 superagonists require the exchange factor Vav1 but not phosphoinositide 3-kinase in primary peripheral T cells. Eur. J. Immunol.38, 2528–2533 (2008). [DOI] [PubMed] [Google Scholar]

[CR62] 62.Jenkins, E., Whitehead, T., Fellermeyer, M., Davis, S. J. & Sharma, S. The current state and future of T-cell exhaustion research. Oxf. Open Immunol.4, iqad006 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR63] 63.Chiorazzi, M. et al. Autologous humanized PDX modeling for immuno-oncology recapitulates features of the human tumor microenvironment. J. Immunother. Cancer1110.1136/jitc-2023-006921 (2023). [DOI] [PMC free article] [PubMed]

[CR64] 64.Rongvaux, A. et al. Development and function of human innate immune cells in a humanized mouse model. Nat. Biotechnol.32, 364–372 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR65] 65.Chapuis, A. G. et al. Transferred WT1-reactive CD8+ T cells can mediate antileukemic activity and persist in post-transplant patients. Sci. Transl. Med.5, 174ra127 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR66] 66.Schmitt, T. M. et al. Abstract A51: High-throughput method identifies rare, high-affinity, thymus-vetted T cell receptors (TCRs) for clinical translation. Cancer Immunol. Res.6, A51–A51 (2018). [Google Scholar]

[CR67] 67.National Cancer Institute, N. I. o. H. (Cancer Therapy Evaluation Program, 2009).

[CR68] 68.Rosskopf, S. et al. A Jurkat 76 based triple parameter reporter system to evaluate TCR functions and adoptive T cell strategies. Oncotarget9, 17608–17619 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR69] 69.Lun, A. T. L. et al. EmptyDrops: distinguishing cells from empty droplets in droplet-based single-cell RNA sequencing data. Genome Biol.20, 63–63 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR70] 70.Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics29, 15–21 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

A CD8αβ co-receptor modified to contain an intracellular CD28 signaling tail enhances TCR-engineered T cell function independent of solid-tumor-associated co-stimulatory ligands

Shihong Zhang

Tzu-Hao Tang

Sinéad Kinsella

Francesco Mazziotta

Michael T Schweizer

Megan S McAfee

Ariunaa Munkhbat

Yapeng Su

Valentin Voillet

Lauren E Martin

Colton W Smith

Yuta Asano

Menna Hailemariam

Jakob Bakhtiari

Bo Lee

Cecilia Yeung

Hung Chen

Alessandro M Rizzi

Daniel G Chen

Kelsey Furiya

Nick Horst

Tianzi Zhang

Phung Le

Kelly McKenna

Shannon K Oda

Anthony Rongvaux

Phillip D Greenberg

Thomas M Schmitt

Aude G Chapuis

Abstract

Introduction

Results

Functional CD4 TTCR-MA1 cells enhance MAGE A1-specific CD8 TTCR-MA1 cells

Fig. 1. Functional CD4 TTCR-MA1 cells enhance MAGE A1-specific CD8+ TTCR-MA1 cells in vitro and in vivo.

CD4 and CD8 TTCR-MA1-CD8αβ are safely infused in one patient, require potency enhancement in stringent murine models

Fig. 2. Infusion of functional CD4 and CD8 TTCR-MA1-CD8αβ are not associated with excessive toxicity in one patient and require enhancement to achieve tumor burden reduction.

Patient tumors have limited and inconsistent expression of both positive and negative co-stimulatory ligands

Intracellular CD28 domains integrated into the CD8β chain enhance CD4 T cell cytokine production, survival, and tumor killing in vitro

Fig. 3. Generation of chimeric CD8β chains functionally integrated into the immune synapse.

The chimeric CD8/CD28 co-receptor enhances TCRMA1 cell function in vivo

Fig. 4. CCR enhances tumor control and T cell survival in vivo.

TTCR-MA1-CD8/28-mediated tumor reduction is associated with cytotoxicity and preservation of self-renewal features

Fig. 5. The chimeric CD8/CD28 co-receptor TCR-T cells have a reduced exhaustion(-like) phenotype in vivo.

CD8-CD28 CCR tail mutations further enhance anti-tumor activity in MAGE tumor models

Fig. 6. CD8-CD28 CCR tail mutations enhance anti-tumor activity in MAGE tumor models.

CD28mut with less evidence of unrelenting activation maintains responsiveness following repeated antigenic stimulation

Fig. 7. The gene expression profile of TTCR-MA1CD8/CD28mut identified a balance between functional, exhaustion and persistence.

Discussion

Method

Isolation of TCRMA1 and lentiviral vector construction

Functional avidity screening of 5 different clones of TCRMA1

Production of TCR-deficient Jurkat cells and Nur77-mNeonGreen reporter cells

Recognition of MAGE-A1-K278T variant

Selection of the MA1-specific TCR for clinical translation

Generation of TCRMA1 lentivirus used in pre-clinical studies

Generation of TTCR-MA1: T cell isolation, activation and lentiviral transduction

T cell rapid expansion protocol (REP)

Generation of research tumor cell lines

TTCR-MA1 cell cytotoxicity assays: Intracellular Cytokine Staining

Real-Time Killing Assay

CFSE proliferation assay

In vivo murine models

Clinical protocol

Patient selection

Treatment plan

Patient 1 case report

Generation of TTCR-MA1 for clinical use

Patient infusion product phenotype staining and intracellular cytokine staining

Transgenic T cell tracking by p/HLA multimer

MAGE A1 detection by Immunohistochemistry

Flow cytometry for in vivo murine models

In vivo tumor immunohistochemistry

TCR tonic signaling

Single cell RNA sequencing

TTCR-WT1 and TTCR-PRAME cells generation

Phospho-flow cytometry analysis

Bulk RNA sequencing

Peptide stimulation assay

Functional CD4 T_TCR-MA1 cells enhance MAGE A1-specific CD8 T_TCR-MA1 cells

Fig. 1. Functional CD4 T_TCR-MA1 cells enhance MAGE A1-specific CD8⁺ T_TCR-MA1 cells in vitro and in vivo.

CD4 and CD8 T_{TCR-MA1-CD8αβ} are safely infused in one patient, require potency enhancement in stringent murine models

Fig. 2. Infusion of functional CD4 and CD8 T_{TCR-MA1-CD8αβ} are not associated with excessive toxicity in one patient and require enhancement to achieve tumor burden reduction.

The chimeric CD8/CD28 co-receptor enhances TCR_MA1 cell function in vivo

T_{TCR-MA1-CD8/28-}mediated tumor reduction is associated with cytotoxicity and preservation of self-renewal features

Fig. 7. The gene expression profile of T_{TCR-MA1CD8/CD28mut} identified a balance between functional, exhaustion and persistence.

Isolation of TCR_MA1 and lentiviral vector construction

Functional avidity screening of 5 different clones of TCR_MA1

Generation of TCR_MA1 lentivirus used in pre-clinical studies

Generation of T_TCR-MA1: T cell isolation, activation and lentiviral transduction

T_TCR-MA1 cell cytotoxicity assays: Intracellular Cytokine Staining

Generation of T_TCR-MA1 for clinical use

T_TCR-WT1 and T_TCR-PRAME cells generation