Engineering T cells with a novel PD-L1-specific TCR targeting immune and cancer cells

Dear Editor

The immune system plays a central role in recognizing and eliminating cancer cells, yet many tumors escape immune surveillance by creating an immunosuppressive microenvironment driven by inhibitory cytkines, checkpoint molecules and tolerogenic immune cell subsets. Reversing this suppression has been pursued therapeutically, most notably through blockade of immune checkpoints such as the PD-1/PD-L1 axis. Complementing these approaches, our research group has identified naturally occurring proinflammatory T cells that recognize tumor microenvironment (TME) antigens expressed by key immunosuppressive cells in the TME.¹ Among these, PD-L1-specific T cells were detected in both healthy donors and cancer patients and were able to eliminate PD-L1-expressing tumor and immune cells, highlighting their immune modulatory potential.² Importantly, their therapeutic relevance was demonstrated in a clinical trial where vaccination with PD-L1- and indoleamine 2,3-dioxygenase (IDO)-derived peptides combined with PD-1 blockade achieved high response rates in patients with metastatic melanoma.³

While these vaccine-based strategies have shown clear therapeutic promise, their efficacy may be constrained by variable T cell activation and tumor immune escape. Adoptive transfer of engineered T cells provides a supplementary approach to amplify these immune modulatory T-cell responses. In this study, we aimed to translate the promising properties of PD-L1-specific T cells towards a defined cellular therapy. To do so, we reconstructed a PD-L1-specific TCR from a monoclonal T cell culture recognizing the PD-L1-derived epitope PDL101 and introduced it into the TRAC locus of primary CD8⁺ T cells from seven healthy donors using non-viral CRISPR-Cas9-based editing.⁴ Simultaneously, we knocked out the endogenous TRBC loci to minimize TCR mispairing. Both utilized guide RNAs mediated efficient knockout of the endogenous CD3-TCR complex (Fig. 1a left). Staining the engineered T cells with PDL101-loaded HLA-A2 tetramers confirmed the generation of PDL101-TCR-T cells and enabled the enrichment and expansion of cells carrying the transgenic TCR, yielding highly specific T cell cultures (Fig. 1a middle and on Figshare: 10.6084/m9.figshare.32270148). Upon stimulation with their cognate antigen, the PDL101-TCR-T cells mounted strong effector responses characterized by interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α) production as well as upregulation of the degranulation marker CD107a (Fig. 1a right and on Figshare). More importantly, the PDL101-TCR-T cells showed potent antigen-dependent cytotoxicity, efficiently lysing peptide-loaded, TAP-deficient T2 target cells in a concentration-dependent manner (Fig. 1b left). In addition, the T cells also eliminated targets pulsed with the longer PD-L1-derived peptide PDL1Long1, consistent with TAP-independent processing and cross-presentation of the embedded PDL101 epitope. Antigen-dependent cytotoxicity was dose-dependent, with robust target cell lysis sustained into the low picomolar range (available on Figshare).

Fig. 1. PD-L1-specific TCR-T cells show potent antigen-specific reactivity and respond to both PD-L1-expressing cancer- and immune cells.

a CD8⁺ T cells isolated from peripheral blood mononuclear cells of four healthy donors were stimulated with anti-CD3/CD28 beads, IL-2 and IL-7, and transfected with TRAC- and TRBC-targeting ribonucleoprotein complexes to assess endogenous TCR knockout efficiency (left). Subsequently, the transfections were repeated in seven healthy donors with a double-stranded DNA homology-directed repair template, enabling insertion of the PDL101-TCR into the TRAC locus. Engineered T cells were purified as CD3⁺PDL101-tetramer⁺ cells using fluorescence-activated cell sorting and expanded with the rapid expansion protocol. Antigen specificity before and after enrichment was assessed by staining with control- (HLTV-1 Tax_11-19) or PDL101 peptide-loaded HLA-A2 tetramers (middle). Antigen-dependent upregulation of IFN-γ, TNF-α and CD107a was quantified using intracellular cytokine staining (ICS) following PDL101 peptide stimulation (5 µM, right). Each data point represents a separate donor and shows the fraction of PDL101-tetramer⁺ cells, IFN-γ⁺/TNF-α⁺ cells or CD107a⁺ cells within the CD8⁺ population. b Antigen-specific cytotoxicity was assessed by ⁵¹Cr-release assays using radioactively labeled T2 cells loaded with either DMSO, PDL101 or PDL1Long1 (4 h, 20 µM) as target cells for enriched PDL101-TCR-T cells from five donors (left). Thin lines represent mean target cell lysis of technical duplicate or triplicate measurements for individual donors, whereas the thick lines represent the mean of all donors. Statistical analysis at discrete E:T ratios was performed using paired, two-tailed t-tests comparing peptide-loaded target cells to DMSO controls. The cutaneous T cell lymphoma cell lines MAC-1 and MAC-2a were similarly used as targets for enriched PDL101-TCR-T cells (right). c Autologous CD14⁺ myeloid cells were purified from three healthy donors and differentiated towards either a TAM-like phenotype with M-CSF, IL-4, IL-10 and tumor-conditioned medium from MDA-MB-231, or an MDSC/M2-like phenotype using GM-CSF and IL-6. Surface expression of CD163, CD206, CD86, HLA-DR and PD-L1 on undifferentiated CD14-purified control cells (gray) and the differentiated TAM-like (light blue) and M2/MDSC-like (dark blue) myeloid subsets was quantified by flow cytometry (top). Bars show the fraction of CD33⁺CD11b⁺ cells expressing high levels of each marker, except for PD-L1 which is shown as the total fraction of PD-L1⁺ cells. Data points from individual donors are indicated with distinct shapes. Between 0.75 and 1.0 ×10⁵ enriched PDL101-TCR-T cells were either plated alone, with PDL101 peptide (5 µM), or co-cultured with either of the three myeloid subsets at an effector-to-target (E:T) ratio of 2:1 in IFN-γ ELISPOT (bottom). Bars represent mean IFN-γ spot count ±SD of technical triplicates. *p ≤ 0.05 according to the DFR rule. d Autologous monocyte-derived DCs were generated by differentiating CD14-purified cells with IL-4 and GM-CSF and subsequently maturing them with IL-6, IL-1β, TNF-α and prostaglandin E₂. Peptide-loaded (50 µg/mL, 4 h), irradiated DCs were used to expand CD8⁺ T cells that were transfected as in (a), over two rounds performed one week apart. The proportion of PDL101-TCR⁺ T cells was determined before DC stimulation and one week after each DC stimulation using tetramer staining (top). Then, autologous DCs were used as target cells for expanded PDL101-TCR-T cells from a single donor in IFNγ ELISPOT at an E:T ratio of 2:1 (bottom). DCs were either mock transfected or transfected with PD-L1-targeting siRNA (0.375 µM) prior to use as target cells. PD-L1 expression was assessed by flow cytometry (bottom). Bars represent mean IFN-γ spot count ±SD of technical triplicates, *p ≤ 0.05 according to the DFR rule

Having established the potency and specificity of the engineered T cells against peptide-loaded targets, we next investigated whether these cells could recognize and eliminate PD-L1-expressing cancer cells. To this end, we utilized two HLA-A2⁺ cutaneous T cell lymphoma cell lines, MAC-1 and MAC-2a, that both originate from the same patient. Interestingly, PDL101-TCR-T cells from five different donors preferentially lysed MAC-1 cells compared to MAC-2a cells, consistent with the observation that only MAC-1 cells displayed detectable surface levels of PD-L1 (Fig. 1b right and on Figshare).

Besides cancer cells, several immune subsets present in the TME also express PD-L1 and contribute to local immunosuppression. Accordingly, we investigated whether PDL101-TCR-T cells could recognize and respond to PD-L1-expressing myeloid cells. To this end, autologous CD14⁺ cells from three donors were purified and differentiated into either a TAM-like phenotype using macrophage colony-stimulating factor (M-CSF), interleukin-4 (IL-4), IL-10, and tumor-conditioned media from MDA-MB-231 cells, or an MDSC/M2-like phenotype using GM-CSF and IL-6. Importantly, both differentiation regimens increased PD-L1 expression compared to undifferentiated controls (Fig. 1c top). TAM-like cells displayed high CD163 and CD206 expression alongside low CD86 and HLA-DR levels, underlining their tolerogenic phenotype (Fig. 1c top). Meanwhile, the GM-CSF/IL-6 condition produced a mixed myeloid population consisting of both MDSC-like cells with lower HLA-DR, CD86 and CD163 expression as well as cells with activated M2-like macrophage characteristics including higher expression of HLA-DR, CD163, and CD206 (Fig. 1c top). Next, we assessed targeting of the differentiated myeloid subsets by autologous PDL101-TCR-T cells using the IFN-γ ELISPOT and intracellular cytokine staining (ICS) assay. Two of three donors showed statistically significant T-cell responses to the TAM-like myeloid cells compared to undifferentiated control cells with minimal PD-L1 expression, whereas all three donors exhibited robust responses towards the MDSC/M2-like myeloid cells (Fig. 1c bottom and on Figshare). These findings were recapitulated in the ICS assay, which showed upregulation of IFN-γ, TNF-α and CD107a under the same conditions (available on Figshare). Notably, while TAM-like cells showed higher PD-L1 expression, MDSC/M2-like cells induced stronger T-cell responses. Differences in antigen presentation or co-stimulatory capacity may explain these discrepancies, consistent with the observation of higher mean fluorescent intensity of CD86 in MDSC/M2-like cells (available on Figshare). Collectively, these findings demonstrated that PDL101-TCR-T cells can effectively target PD-L1-expressing cancer and immune cells.

Having established the functional capacity of the PDL101-TCR-T cells, we next examined their exhaustion- and activation state following antigen engagement. After 48 h of co-culture with peptide-loaded T2 cells, we observed antigen-driven upregulation of PD-1, LAG-3 and CD25 on the PDL101-TCR-T cells (available on Figshare). Notably, the addition of anti-PD-1 treatment during co-culture showed variable effects on IFN-γ secretion by the T cells, with modest increases in donors with higher PD-1 expression (available on Figshare). These findings point to potential synergy between immune checkpoint blockade and PDL101-TCR-T cells, with possible therapeutic benefit.

Beyond checkpoint synergy, a key translational consideration is the lower integration and expression rate of the PDL101-TCR with CRISPR–Cas9 editing compared to viral transduction (Fig. 1a left). To address this challenge, we explored whether PDL101-TCR-T cells could be selectively expanded through stimulation with peptide-loaded dendritic cells (DCs) in vitro, mirroring a potential in vivo vaccination strategy. To this end, CRISPR-Cas9-based gene editing of T cells from three healthy donors initially yielded 9–12% PDL101-TCR⁺ T cells, which increased 5–6-fold to 50–65% following two rounds of stimulation with autologous peptide-loaded DCs (Fig. 1d top). These results demonstrate the feasibility of DC-mediated expansion and support the potential for in vivo boosting of this population through peptide-based vaccination.³ Interestingly, enriched PDL101-TCR-T cells were also able to recognize autologous, non-peptide-loaded DCs in a PD-L1-dependent manner (Fig. 1d bottom). This may enable preferential targeting of PD-L1⁺ tolerogenic DCs and thereby limit immunosuppression, but could also risk depletion of antigen-presenting cells required for effective anti-tumor immunity.

From a translational perspective, safety remains an important factor when targeting PD-L1-expressing cell types. Cellular therapies directed against PD-L1 are currently being explored, particularly in the form of CAR-T cells, but concerns regarding on-target, off-tumor toxicity have been raised.⁵ Specifically, Bajor et al. reported cytotoxic effects against several PD-L1⁺ non-malignant cell lines.⁵ In contrast to artificially engineered PD-L1–specific CARs, our approach utilizes a physiological, MHC-restricted TCR specificity that likely operates with lower functional avidity and a higher activation threshold, potentially providing an intrinsic safety mechanism.

In summary, our study provides a proof-of-concept demonstration of PD-L1-specific TCR-T cells and highlights their capacity to target PD-L1-expressing cancer and immune cells in vitro. By combining the physiological specificity of naturally occurring T cells with the precision of CRISPR-Cas9 genome engineering, this work represents initial steps toward a preclinical framework, warranting further investigation of PD-L1-targeted TCR-based strategies.

Author contributions

T.L.L., E.B.J. and M.H.A. designed the project. T.L.L., E.B.J. and S.M.A. performed the experiments. T.L.L., E.B.J., S.M.A. and Ö.M. performed the data analysis. All authors contributed to the interpretation of acquired data. M.H.A., Ö.M. and R.O.B. provided reagents. T.L.L. and E.B.J. wrote the first manuscript draft, while all authors contributed, read, and approved the final manuscript.

Data availability

All data supporting the findings of this study are included within the research article and the supplementary figures, which are available on the Figshare repository (10.6084/m9.figshare.32270148). The utilized primers, single-guide RNAs, antibodies and reagents as well as the PDL101-TCR HDR template structure are available in the Supplementary Information.

Competing interests

The authors declare no competing interests.

Ethics approval

Biological material from healthy anonymized blood donors were acquired from Rigshospitalet, Copenhagen, Denmark. According to the Danish Law on Research Ethics § 14, section 3, the use of anonymized biological material does not require approval from an ethics committee. The PDL101-specific clone giving rise to the PDL101-TCR originated from a breast cancer patient enrolled in a clinical study at our center. Written informed consent was obtained before study enrollment, and the protocol was approved by the Scientific Ethics Committee for the Capital Region of Denmark (approval number: H-A-2009-013) in accordance with the Declaration of Helsinki.

Supplementary information

The online version contains supplementary material available at 10.1038/s41392-026-02798-y.