HER2-specific synthetic antigen receptor T cell therapy synergizes with radiotherapy to provide improved antitumor efficacy in non-small cell lung cancer.

1.
Background
While immune checkpoint inhibitors have improved the treatment of unresectable non-small cell lung cancer (NSCLC) when added to standard-of-care chemoradiotherapy, long-term survival remains poor,1 underscoring the need for developing novel and more effective therapeutic strategies.
NSCLC is a heterogeneous disease characterized by an array of oncogenic driver alterations.2 One such driver is the receptor tyrosine kinase, human epidermal growth factor receptor 2 (HER2), a member of the epidermal growth factor receptor (EGFR/HER1) family, which consists of 4 members (HER1–4).3 Although HER2 lacks a known ligand, it dimerizes with other ligand-bound EGFR family members or homodimerizes when overexpressed, activating downstream signaling cascades that drive tumor cell proliferation and survival such as phosphatidylinositol/protein kinase B (PI3 K/Akt) and rat sarcoma/mitogen-activated protein kinase (Ras/MAPK).4,5
In NSCLC, HER2 activation occurs through gene mutation (1% to 4% of cases), gene amplification (1% to 15% of cases), or protein overexpression (2% to 30% of cases).6,7 HER2 activation in NSCLC is associated with poor prognosis and resistance to cancer therapy.6,7 Radiotherapy has been shown to upregulate HER2 expression and signaling in breast cancer cells, promoting pro-survival pathways that contribute to radioresistance.8,9 Studies in lung cancer models also demonstrate radiation-induced activation of HER-family signaling, including increased HER2 phosphorylation and downstream Akt/MAPK activation, while genetic or pharmacological HER2 blockade augments radiosensitivity.10–12 These findings suggest that radiotherapy activates HER2-related adaptive survival pathways in lung tumors, creating a window for HER2-directed therapies.
From a therapeutic standpoint, HER2 has gained increasing attention as a targetable vulnerability in NSCLC. Early efforts focused on pan-HER family tyrosine kinase inhibitors (TKIs) and anti-HER2 monoclonal antibodies, with limited and variable success. More recently, the development of selective HER2 TKIs, such as pyrotinib and poziotinib, has renewed interest in targeting HER2-mutant NSCLC and has demonstrated clinical activity in this population.7 However, anti-HER2 antibody–drug conjugates (ADCs), which deliver a cytotoxic payload to HER2-expressing cells, have shown greater promise either as a monotherapy or in combination with immunotherapy. Notably, the trastuzumab–deruxtecan ADC was recently approved by the US Food and Drug Administration (FDA) for the treatment of HER2-expressing NSCLC, demonstrating some meaningful responses while also revealing inconsistent benefit across patients.6 The development of effective anti-HER2 therapies therefore represents an opportunity to overcome radiation-induced resistance and improve patient outcomes.
Adoptive cell transfer (ACT) using autologous T cells engineered to express synthetic antigen receptors (SARs) has achieved outstanding success in treating hematological malignancies.13,14 The most widely used SAR is the chimeric antigen receptor (CAR), with 7 CAR T cell therapies currently FDA-approved. Despite their clinical efficacy, CAR T therapies are often associated with severe toxicities,15 which must be managed to ensure patient safety. Mounting evidence suggests that these toxicities are linked to the synthetic receptor architecture.16 To address these limitations, we developed a DNAX-activating protein of 12 kDa (DAP12)-based multichain immunoreceptor complex as the foundation for a HER2-specific SAR, which has demonstrated an improved safety profile and potent antitumor activity in preclinical models.17,18
Applying engineered T cell therapies to solid tumors such as NSCLC remains challenging due to limited tumor-specific antigen availability, an immunosuppressive tumor microenvironment, rapid T cell exhaustion, and inadequate T cell trafficking and persistence.19 Radiotherapy has emerged as a promising combinatorial strategy to overcome these barriers by promoting the release of tumor-associated antigens and damage-associated molecular patterns, upregulating major histocompatibility complex class I expression, and enhancing cytokine and chemokine secretion.20 Combining radiotherapy with T cell therapy may therefore enhance antitumor immunity and improve treatment outcomes.
Indeed, Zhou and colleagues showed that EGFR-targeted CAR T cells combined with radiotherapy improved tumor control in murine models of triple-negative breast cancer (TNBC),21 and similar synergistic effects have been reported in prostate, glioblastoma, and pancreatic cancers.22
Based on this rationale, we hypothesized that the localized effects of radiotherapy can be leveraged to induce focal HER2 expression within tumors, thereby enhancing the safety and efficacy of DAP12-based HER2-SAR T cells against HER2-expressing NSCLC.

2.
Methods
2.1
Cell lines and treatments
Human NSCLC cell lines A549, H1299, H1975, and SK-MES-1 (ATCC: Manassa, VA) were cultured in RPMI-1640 (A549, H1299), ATCC-modified RPMI-1640 (H1975), or DMEM (SK-MES-1) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic–antimycotic (Wisent Bioproducts: Saint-Jean Baptiste, Canada). Cell lines were authenticated and assessed for mycoplasma as described.23 A549 cells were transduced with a dual-promoter lentiviral vector encoding enhanced firefly luciferase and a puromycin resistance gene at a multiplicity of infection (MOI) of 5.24 HEK293T cells were cultured in DMEM (Gibco; Thermo Fisher Scientific: Waltham, MA) supplemented with 10% FBS, L-glutamine, HEPES, penicillin, streptomycin, or normocin at concentrations described in Moore et al.25 All cells were cultured at 37 °C in 5% CO2. NSCLC cells were irradiated with 6MV X-rays using a TrueBeam linear accelerator as previously described.26

2.2
Receptor generation, lentivirus production, and T cell engineering
The HER2 DAP12-SAR construct was generated as previously described.25 Briefly, the transgene was cloned into a third-generation pCCL vector downstream of the EF-1α promoter, with a bidirectional mCMV promoter driving truncated NGFR expression. The construct included full-length DAP12, a GSG-T2A-GTS linker, and the SAR (mlgκ leader, H10-2-G3 anti-HER2 DARPin,27,16 (G4S)2 linker, and the hinge/transmembrane/intracellular domains from KIR2DS2).
Lentivirus was produced as in Helsen et al.16 Briefly, HEK293T cells were transfected with pRSV-Rev, pMD2.G, pMDLg/pRRE, and the pCCL-HER2 DAP12-SAR vector using Lipofectamine 2,000 (Thermo Fisher Scientific) in Opti-MEM (Gibco). After 12 to 16 h, media were replaced with sodium butyrate (Sigma-Aldrich)-supplemented DMEM. Supernatants were harvested 24 to 36 h later, concentrated, and stored at −80 °C. Viral titer was determined by serial dilution, transduction of HEK203T cells, and determining percent NGFR+ with flow cytometry.25
Peripheral blood mononuclear cells were stimulated with ImmunoCult Human CD3/CD28/CD2 activators (25μL/mL; STEMCELL Technologies: Vancouver, Canada) in T cell media25 and transduced 16 to 24 h later with HER2 DAP12-SAR lentivirus (MOI = 2). For in vitro use, cells were expanded in T cell media for 14 d, and for in vivo use, cells were cultured for 8 d in ExCellerate Human T cell Expansion Media, Xeno-Free (Bio-Techne: Minneapolis, MN) supplemented with 100 IU/mL rhIL-2 and 10 ng/mL rhIL-7. Cells were cryopreserved in CryoStor CS10 (STEMCELL Technologies). HER2 DAP12-SAR surface expression was assessed by flow cytometry as previously described.25

2.3
T cell activation assay
Tumor cells were cocultured with HER2-SAR or NT T cells at a 1:1 effector-to-target (E:T) ratio for 6 h. Stimulation was stopped with 4 mM EDTA for 15 min at room temperature in the dark. Cells were stained with NGFR, CD69, and Live/Dead dye, fixed in 2% paraformaldehyde, and filtered prior to acquisition. Samples (1 × 105 events) were acquired on a CytoFLEX S flow Cytometer (Beckman Coulter) and analyzed using FCS Express (v7.18.0025; De Novo Software Inc, Pasadena, CA). Gating was performed as previously described.16 Antibodies are listed in Table S1.

2.4
Xenograft model and tissue handling
Eight- to 10-week-old female nonobese diabetic rag gamma (NRG) mice (NOD.Cg-Rag1(tm1Mom)Il2rg(tm1Wjl)/SzJ; Jackson Laboratory: Bar Harbor, ME) were housed in a pathogen-free facility with ad libitum access to chow diet, water, and environmental enrichment. Anesthetized mice were injected subcutaneously with 1 × 106 luciferase-labeled A549 cells in 1:1 Matrigel (Corning Life Sciences: Corning, NY) to PBS (Wisent Bioproducts). Tumor volume was measured by caliper and calculated as V = 0.5 × (length × width2). Resected tumors were snap frozen in liquid nitrogen or formalin-fixed and paraffin-embedded (FFPE) by the McMaster Histology Core Facility. Frozen tumors were homogenized using the TissueRuptor handheld rotor-stator homogenizer (QIAGEN: Hilden, Germany) in 500μL/50 mg lysis buffer28 or TRIzol Reagent (Invitrogen: Waltham, MA). All procedures were approved by McMaster's Animal Research Ethics Board (AUP 20-12-47).

2.5
Combination index analysis
The highest single-agent model in the combPDX web application (https://licaih.shinyapps.io/CombPDX/) was used to predict synergistic, independent, or antagonistic effects of the combined treatments in vivo.29

2.6
Immunohistochemistry
For single CD3 staining, FFPE sections were deparaffinized, rehydrated, and blocked with hydrogen peroxide (Thermo Fisher Scientific) following standard protocols. Antigen retrieval was performed in citrate buffer (pH 6; Sigma-Aldrich), followed by blocking with 10% goat serum (Vector Laboratories Inc, Newark, CA). Sections were incubated with primary and biotinylated secondary antibodies (Table S1), developed with the NovaRed Substrate Kit (Vector Laboratories Inc), and counterstained with hematoxylin (Abcam: Cambridge, UK). CD3+ cells were quantified in 10 random fields per section using QuPath (v0.501).
For Ki-67/CD3 costaining and cleaved caspase-3 (CC3) staining, tissue sections were cut and processed at the SCSP Core Histology Facility, McMaster University, using a BOND RX automated stainer (Leica Biosystems: Nussloch, Germany). Sections were deparaffinized and stained with primary antibodies CC3 or Ki-67 followed by CD3 (Table S1). Antigen retrieval was performed with Leica ER2 for 20 min before staining Ki-67 and ER2 for 10 min before staining CD3. Detection was performed using Lecia BOND Polymer Refine Red and BOND Refine DAB detection kits according to the Core's standard rabbit antibody protocols. CC3+, Ki-67+, and CD3+ cells were quantified in 20 random fields per section using ImageJ (v1.53).

2.7
Immunoblotting
Cells were lysed, and proteins were quantified using the BCA Protein Assay Kit (Thermo Fisher Scientific).28 Lysates were resolved by 12% to 15% SDS-PAGE and transferred to polyvinylidene difluoride membranes (Bio-Rad). Membranes were blocked with 5% albumin (BioShop Canada Inc) and probed with primary and horseradish peroxidase (HRP)-linked secondary antibodies (see Table S1 for antibodies and dilutions used).28 Chemiluminescence was detected with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific) on the Fusion FX system (Vilber Lourmat: Collégien, France). Densitometry was performed using ImageJ.

2.8
Real-time quantitative polymerase chain reaction
Total RNA was purified using the RNease Mini Kit with on-column DNase digestion (QIAGEN). RNA quality and quantity were assessed with a NanoPhotometer (Implen Inc, Westlake Village, CA), and cDNA synthesis was performed using SuperScript IV (Invitrogen). qPCR was run on the Corbett Rotor Gene 6,000 (Montreal Biotech Inc, Dorval, Canada) using TaqMan Assay fluorogenic 5' nuclease chemistry (Invitrogen). Relative gene expression was calculated using the Livak method,30 normalizing to 18S. Probes are listed in Table S1.

2.9
Evaluation of cell surface HER2 expression
Half a million cells were harvested and stained with Live/Dead Fixable Near-IR Dead Cells Stain (1:500; Invitrogen) and BV421 mouse anti-human HER2 antibody (BD Biosciences) or isotype control (BD Biosciences). Cells were fixed in 2% paraformaldehyde and filtered prior to collecting 1 × 105 events/sample on the CytoFLEX LX Flow Cytometer (Beckman Coulter). FCS Express software was used for data analysis. See Table S1 for antibodies.

2.10
Luciferase-based cytotoxicity assay
T cell constructs and luciferase-expressing A549 cells were seeded into a 96-well solid white flat-bottom plate at different E:T ratios beginning at 8:1 and diluting to 0.25:1. After a 24 h incubation at 37 °C, XenoLight D-Luciferin Potassium Salt (PerkinElmer: Waltham, MA) was added to each well and incubated in the dark for 10 min at room temperature. Luminescence was measured with the BioTek Synergy H1 (Agilent Technologies Inc, Santa Clara, CA).

2.11
RNA sequencing and analysis
RNA sequencing (RNA-seq) was performed by the McMaster Genomics Facility. RNA quality and library preparation followed established protocols.23 Sequencing was conducted on the Illumina NextSeq 2,000 (1 × 50 bp, XLEAP-SBS chemistry), generating ∼31 million clusters per sample. Reads were assessed with FastQC (v0.12.1), trimmed using Cutadapt (v4.9),31 and aligned to human (hg38) and mouse (mm10) genomes using HISAT2 (v2.2.1).32 Disambiguate was employed to separate graft (human) and host (mouse) reads.33 Read counts were generated using featureCounts (v2.0.3)34 and differential expression analysis was done using the DESeq2 (v1.46.0), with significance defined as adjusted P < 0.05. Gene set enrichment analysis (GSEA) was conducted using Hallmark gene sets (MSigDB v2023.1) with a preranked approach (RANK=−log10(P-adj) × log2(Fold Change)) on the GSEA application (v4.2.3).35,36 Gene sets with a false discovery rate (FDR q-value) < 0.05 were considered significantly enriched.

2.12
Statistical analysis
Data are presented as mean ± standard error of the mean (SEM) of at least 3 independent experiments unless otherwise indicated. GraphPad Prism (v10.2.3; GraphPad: La Jolla, CA) was used for statistical analyses. Unpaired 2-tailed t tests were performed to determine the statistical significance of experimental data. For multiple group analysis, 1- or 2-way ANOVA followed by Tukey's post hoc test was used unless otherwise indicated. For survival data analysis, the Kaplan–Meier method and Log-rank (Mantel-Cox) test was used to detect differences between groups. Statistical significance was accepted at P-value < 0.05.

3.
Results
3.1
Anti-HER2-SAR T cells demonstrate potent antitumor activity against HER2-expressing xenografts and synergy with induction radiotherapy
We generated a DAP12-associated multichain immunoreceptor complex-targeting HER2, as described by Moore et al. (Fig. 1A).25 Compared to non-transduced (NT) primary human T cells, HER2-directed engineered SAR T cells exhibited robust activation when cocultured with HER2-expressing A549 NSCLC cells, marked by a 73% increase in CD69+ T cells (Fig. 1B).
Since in vitro assays cannot fully recapitulate the complexity of the tumor microenvironment, we employed an A549 xenograft model to evaluate the efficacy of HER2-SAR T cells alone or in combination with 5 Gy induction radiotherapy (Fig. 1C). A dose of 0.5 × 106 HER2-SAR T cells was selected for combination therapy, as a higher dose of 4.5 × 106 cells was poorly tolerated due to graft-vs-host disease (Fig. S1A, B) and 1.5 × 106 cells were too potent to detect additive effects with radiotherapy (Fig. S1C, D). Combination therapy significantly prolonged survival compared to the vehicle and monotherapy groups, with a median survival of 100 vs 41.5 d for vehicle (P < 0.0001 by log-rank test), 57 d for radiotherapy (P < 0.0001), and 56 d for HER2-SAR T cells alone (P = 0.005; Fig. 1D, E). Importantly, combination index (CI) analysis demonstrated that combined induction radiation and HER2-SAR T cell treatment enhanced tumor suppression in a synergistic fashion (global CI = 0.0459; Fig. 1F). Mice tolerated the 0.5 × 106 HER2-SAR T cell dose and combination therapy without graft-vs-host disease or significant weight loss (Fig. S1E).
Furthermore, 5 Gy induction radiotherapy significantly increased HER2-SAR T cell infiltration, identified as cells positive for CD3 (CD3+), in irradiated A549 tumors compared to nonirradiated tumors 8 d after HER2-SAR T cell infusion at the 0.5 × 106 dose level (P = 0.043; Fig. 2A, B). Dual IHC staining for CD3 and the proliferation marker Ki-67 showed that 57.7% of CD3+ cells also expressed Ki-67 (Ki-67+) in tumors treated with HER2-SAR T cells alone, and this proportion rose to 69.1% following combined irradiation and HER2-SAR T cell therapy (P = 0.038; Fig. 2C). In addition, irradiation reduced the number of Ki-67 + CD3− cells (Fig. 2D). While this decrease was not significant vs NT T cells (P = 0.095), it was significant in tumors treated with combined irradiation and HER2-SAR T cells (P = 0.025). Consistent with these findings, CC3 staining showed a statistically significant increase only in the 5 Gy plus HER2-SAR T cell combination therapy group compared to NT T cell-treated controls (P = 0.003; Fig. 2A, E).
Although no additional antitumor or survival benefit was observed with the higher 1.5 × 106 dose in combination with radiotherapy, irradiated tumors still exhibited a similar increase in HER2-SAR T cell infiltrate (Fig. S1F, G).

3.2
Radiotherapy transiently induces the expression of HER2 in NSCLC
Studies have shown that HER2 expression is upregulated in breast cancer 24 h after irradiation.8 Additionally, increased expression of the target antigen has been associated with enhanced CAR T cell efficacy.37 Based on these findings, we investigated whether radiotherapy similarly upregulates HER2 in NSCLC, which may present an opportunity for temporal optimization of the combined therapy. We found that human NSCLC cells express varying levels of HER2 with H1975 cells expressing the highest levels, followed by A549, H1299, and SK-MES-1 cells, respectively (Fig. S2A, B). A single 5 Gy fraction significantly increased total HER2 protein 24 h postirradiation across all lines (Fig. 3A, B). A trypan blue exclusion assay confirmed that >95% of cells remained viable 24 h after 5 Gy across all 4 cell lines, indicating that the observed increases in HER2 expression were not due to radiation-induced cell death (Fig. S2C). To determine the temporal dynamics of HER2 following irradiation, A549 NSCLC cells were exposed to a single dose of 5 Gy, and HER2 mRNA and protein were assessed at various time points. HER2 mRNA rose rapidly postirradiation (0.5 h) and decreased by 2 h, falling below basal levels at 4 h and reverting to the baseline by 24 h (Fig. 3C). Total HER2 protein rose as early as 1 h postirradiation but did not reach statistical significance until 4 h (Fig. 3D, E). HER2 protein remained elevated at the 48 h timepoint and returned to basal levels by 7 d (Fig. 3D, E). Surface HER2 also peaked at 24 h and subsided by 48 h postirradiation (Fig. 3F, G).
Induction of HER2 expression by radiation was also examined in A549 xenografts 4, 24, 48 h, and 7 d after treatment with a single fraction of 5 Gy (Fig. 4A). The tumoricidal effect of this treatment was detected within 4 h, as indicated by increased levels of CC3 in tumor lysates (Fig. 4B, C). Compared with time-matched nonirradiated tumors, HER2 protein was significantly elevated at 4 and 24 h but not at 48 h or 7 d postirradiation (Fig. 4D, E).

3.3
Reducing the interval between radiotherapy and T cell infusion does not enhance the therapeutic efficacy of the combined therapy
Based on the findings above, we hypothesized that infusing HER2-SAR T cells during peak HER2 expression postirradiation would enhance therapeutic outcomes. To test this, HER2-SAR T cells were administered to mice 4, 24, 48 h, and 7 d following a single fraction of 5 Gy (Fig. 5A). All treatment schedules yielded comparable tumor control (Fig. 5B, C). Although tumoral HER2 levels remained unchanged at 48 h and 7 d postirradiation (Fig. 4D, E), therapeutic enhancement was still observed when T cell infusion was delivered at these time points after radiation. This suggests that the improved efficacy of HER2-SAR T cells in conjunction with radiotherapy is likely driven by factors other than HER2 density.
To eliminate potential confounding effects of the tumor microenvironment on the therapeutic efficacy of HER2-SAR T cells, we conducted in vitro functional assays to assess T cell cytotoxicity against nonirradiated and irradiated A549 cells. We observed that HER2-SAR T cells triggered robust cytotoxic activity against A549 cells (Fig. 5D). While the cytotoxic capacity of HER2-SAR T cells showed trends for enhancement when cocultured with irradiated A549 cells (24-h postirradiation), this did not reach statistical significance (Fig. 5D). These data indicate that the enhanced in vivo efficacy of HER2-SAR T cells in combination with radiotherapy may be driven by a mechanism external to the tumor cells.

3.4
Transcriptional landscape of irradiated tumors
To explore mechanisms underlying synergy, we performed RNA-seq on A549 tumors 7 d after a single 5 Gy dose. Reads were aligned to human and mouse genomes to resolve tumor cell responses from those of the murine stroma. In irradiated tumors, 134 human genes were upregulated and 143 downregulated, while 128 murine genes were upregulated and 144 downregulated (Fig. 6A; see Figs S3 and S4 for gene lists).
The volcano plots in Figs 6B and 6C highlight the most significantly differentially expressed genes (DEGs) in human tumor cells and murine stroma, respectively, following radiotherapy. In tumor cells, radiation suppressed genes involved in protein degradation (NEURL1B), extracellular matrix (ECM) remodeling and cell adhesion (SPOCK1, COL5A1), cell migration (SLIT3), and signaling regulation (SHISAL1) (Fig. 6B). Notably, FAS, a gene encoding for Fas death receptor, was the topmost upregulated gene and showed concordant upregulation at the protein level (Fig. 6D, E). Other upregulated genes included those associated with cellular stress and repair (SESN1, TFF2), tumor necrosis factor (TNF) receptor signaling (EDA2R), and immune modulation (ISG15) (Fig. 6B). In the stroma, radiation downregulated ECM-associated genes (Tnn, Col6a3, Col6a2, Eln, and Reln) and upregulated genes related to oxygen transport (Hbb-b1), cell cycle regulation (Ccng1), lipid metabolism (Apoc1), mitochondrial protein synthesis (Mtln), and glycogen synthesis (Gyg) (Fig. 6C). These data uncover compartment-specific transcriptional programs activated by radiation, reflecting processes such as apoptosis and stress adaptation in tumor cells, and ECM remodeling and metabolic reprogramming in the stromal microenvironment.
GSEA revealed several enriched pathways 7 d following radiotherapy (Fig. 6F, G). In tumor cells, radiation enhanced gene sets related to oxidative phosphorylation, fatty acid metabolism, Myc targets, p53 signaling, DNA repair, reactive oxygen species (ROS) pathway, and interferon (IFN) alpha response (Fig. 6F). Pathways associated with epithelial-mesenchymal transition, cell cycle progression (G2/M checkpoint and mitotic spindle), hypoxia, TNFα signaling via nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), and key oncogenic signaling pathways including transforming growth factor (TGF)-β, Wnt/β-catenin, and hedgehog were downregulated, suggesting suppression of proliferative and invasive transcriptional programs in response to radiation.
In the tumor stroma, radiation similarly enriched gene sets associated with oxidative phosphorylation, DNA repair, Myc targets, adipogenesis, fatty acid metabolism, ROS pathway, and interferon responses (Fig. 6G). Downregulated pathways included epithelial–mesenchymal transition, hypoxia, angiogenesis, glycolysis, mitotic spindle, TNFα signaling via NF-κB, Hedgehog, and Wnt/β-catenin signaling. Together, these findings reveal both distinct and shared transcriptional responses to radiation across tumor and stromal compartments, characterized by immune modulation and metabolic reprogramming that may influence therapeutic responses to HER2-SAR T cells.
While the 7-d time point is the focus of our study, we also examined earlier radiation-induced changes. RNA-seq analysis of A549 tumors at 24 h post-5 Gy revealed modulation of similar pathways in both tumor and stromal compartments (Fig. S5), including a consistent interferon and stress response signature, along with suppression of proliferative and adaptive signaling. Notably, FAS was again among the top upregulated genes at 24 h postirradiation, and protein immunoblotting confirmed increased Fas expression at 4, 24, 48 h, and 7 d, indicating early and sustained upregulation of Fas postirradiation (Figs 6D, E and Fig. S6) that could contribute to the increased cytotoxicity of HER2-SAR T cells.

4.
Discussion
The goal of our study was to develop a T cell therapy guided by conformal radiotherapy to direct cytotoxic activity within the tumor microenvironment. We demonstrate that HER2-targeting engineered SAR T cells synergize with induction radiotherapy to inhibit the growth of HER2-expressing NSCLC. By titrating down the HER2-SAR T cell dose, we achieved safe and effective antitumor responses when combined with induction radiotherapy.
Radiotherapy mediated a rapid but transient increase of HER2 expression at both the transcript and protein levels in A549 cells and tumors. However, this upregulation did not enhance the antitumor action of HER2-SAR T cells, suggesting that baseline HER2 expression was adequate for T cell engagement in this model. These findings are consistent with prior data showing that the relationship between antigen density and CAR T cell efficacy is not always linear. Some studies report improved CAR T cell function with higher antigen density,37 while others report a peak in CAR T cell activity within a specific antigen range.38 These differences are likely attributed to the selected target antigen, CAR design, T cell fitness, and the experimental model used.
Importantly, the therapeutic efficacy of the combination therapy was not dependent on a specific infusion interval in our preclinical model. Despite measurable fluctuations in HER2 expression after irradiation, HER2-SAR T cells produced comparable tumor control when administered 4, 24, 48 h, or 7 d postirradiation (Fig. 5B). These results indicate that synergy between the 2 therapies does not rely on a narrow window of transient HER2 upregulation but instead reflects radiation-induced remodeling of the tumor cells and/or the tumor microenvironment that enhances T cell infiltration and function. Within the range tested in our model, the timing of T cell infusion relative to radiotherapy therefore appears flexible and does not require precise coordination to achieve therapeutic benefit.
Our in vitro and in vivo data confirm that the antitumor effects of HER2-SAR T cells were specifically mediated through HER2, as NT T cells showed no activity (Figs 1 and 5, Fig. S1). Consistent with this, our previously published work demonstrated that DAP12-based HER2-SAR T cells exhibited robust cytotoxicity against HER2-positive U251 glioblastoma cells but no activity against HER2-negative U251 cells, supporting the antigen specificity of this SAR platform.25 However, because HER2 is also expressed in normal tissues of the lungs and other vital organs, anti-HER2 T cells carry a significant risk of on-target, off-tumor toxicity. Preclinical studies in mice harboring syngeneic HER2-positive tumors demonstrated that high-affinity CAR T cells targeting murine HER2 elicited moderate-to-severe toxicity, T cell exhaustion, and poor tumor control.39 In contrast, low-affinity HER2-CAR T cells were effective without causing toxicity. To mitigate the risk for normal organ toxicity, we focused our work on low HER2-SAR T cell doses with limited cytotoxic capacity when administered alone. Indeed, our data support the notion that titrated HER2-SAR T cells, when combined with conformal induction radiotherapy, can achieve tumor-specific synergy against HER2-expressing lung cancer. Additional preclinical studies using HER2-SAR constructs that also recognize murine HER2 or employ humanized immune system models could further validate this concept.
Allowing sufficient time between radiotherapy and HER2-SAR T cell infusion may help reduce potential side effects of combination therapy. This also aligns with clinical scenarios, where radiotherapy can serve as a bridging therapy during T-cell manufacturing. In our study, the combination therapy was well-tolerated and showed comparable antitumor effects regardless of whether HER2-SAR T cells were infused early (4 h) or later (7 d) after irradiation. However, standard radiotherapy for localized NSCLC typically involves fractionated dosing over several weeks. Future studies are needed to evaluate how this clinically relevant schedule affects tumoral HER2 expression and the efficacy of HER2-SAR T cells.
Enhanced T cell infiltration and local expansion, as evidenced by increased Ki-67 + CD3+ cells, likely underpin the observed synergy between HER2-SAR T cells and radiotherapy (Fig. 2A–C). In tissue cultures, irradiation of A549 cells only minimally increased HER2-SAR T cell cytotoxicity (Fig. 5D). Since these assays measure overall target cell lysis over a 24-h period, they were not designed to assess relative acute killing and do not distinguish the contributions of specific T cell-mediated killing mechanisms such as degranulation or death receptor-mediated apoptosis, nor delayed apoptotic responses that may emerge in vivo. The lack of a measurable difference in overall cytotoxicity in vitro suggests that the observed synergy is unlikely due to substantial enhancement of HER2-SAR T cell killing. Instead, we hypothesize that synergy is driven by radiation-induced changes in tumor cell susceptibility and the tumor microenvironment, potentially via gene expression reprogramming that facilitates T cell recruitment, proliferation, and/or retention.
Our RNA-seq analysis of irradiated A549 tumors revealed that radiotherapy profoundly reshaped the transcriptional landscape, producing both immediate and sustained effects on gene expression that favor cellular stress responses and immune modulation. Radiation-induced regulation of genes associated with DNA repair, metabolic stress, cell cycle arrest, apoptosis, and immune activation, which may increase tumor cell sensitivity to HER2-SAR T cell attack. Notably, the modulation of key immunomodulatory pathways including IFN-α, IFN-γ, TGF-β, TNF-α, and IL-2/STAT5 signaling 7-d postirradiation reflects a delayed and complex remodeling of the tumor microenvironment. These observations emphasize the importance of timing in combination therapies and suggest that the optimal window for HER2-SAR T cell intervention may be several days after radiotherapy, when the tumor microenvironment is potentially more receptive to immune attack.
Studies have shown that radiation induces proapoptotic transcriptional programs in tumor cells, enhancing death receptor activity and sensitizing tumors to T cell-mediated killing.40–43 For instance, low-dose total tumor irradiation has been shown to induce a “death receptor score” and prime leukemia cells for CAR T cell cytotoxicity, with upregulation of FAS, CASP8, and other death pathway components.41 In line with these findings, our RNA-seq dataset revealed FAS upregulation at 24 h and 7 d postirradiation, and immunoblotting confirmed increased Fas protein at 4, 24, 48 h, and 7 d. These data indicate sustained sensitization of tumor cells to HER2-SAR T cell-mediated apoptosis via Fas/FasL engagement. This mechanism is further supported by enhanced CC3 expression in tumors treated with the combination of 5 Gy and HER2-SAR T cells. While prior reports have documented transient increases in inflammatory cytokines and chemokines following irradiation,44,42,45 we observed limited modulation of cytokine transcripts such as CX3CL1 and GDF15, possibly due to sampling timing, tumor type, or treatment context. Notably, at 7-d postirradiation, we observed significant modulation of ECM-related genes (eg elastin, collagens, and adhesion molecules; Fig. S4B) in host stromal cells, indicating robust tumor microenvironment remodeling that may also facilitate HER2-SAR T cell infiltration or retention. While findings support that radiation-induced apoptotic priming and structural reprogramming are involved in augmenting the efficacy of HER2-SAR T cells, further studies are needed to fully elucidate these mechanisms.
HER2-directed engineered T cells are generally well-tolerated in humans.46–49 In a phase I clinical trial, researchers found that HER2-CAR T cells directed against HER2-positive sarcomas exhibited a favorable safety profile with improved HER2-CAR T cell expansion and persistence with lymphodepletion and repeat cycles of treatment.48 Efforts to evaluate the safety of this therapy in combination with anti-PD-1 checkpoint blockade in sarcomas are underway (NCT04995003). In HER2-positive lung cancer, phase I trials examined the safety and efficacy of HER2-CAR T cells (NCT03198052 and NCT03740256), though one trial was withdrawn to change the CAR structure of an unspecified CAR design due to safety concerns (NCT02713984). To date, there are no reported clinical outcomes for the use of anti-HER2 CAR T cells in treating HER2-positive lung cancer alone or in combination with radiotherapy.
In conclusion, induction radiotherapy prior to HER2-SAR T cell infusion generates synergistic antitumor efficacy beyond either treatment alone. This likely results from enhanced expression of the immune-mediated apoptotic pathway and remodeling of the tumor microenvironment induced by radiation. Combining radiotherapy with HER2-SAR T cells is a promising therapeutic strategy for HER2-positive NSCLC and merits further investigation.

Supplementary Material
qiag030_Supplementary_Data