Peripheral blood mononuclear cells (PBMCs) were obtained from healthy donors after informed consent on protocols approved by the Baylor College of Medicine Institutional Review Board. The cell lines CAPAN1, K562, and 293T were obtained from the American Type Culture Collection (Rockville, MD) and were grown in Dulbecco’s modified eagle medium (DMEM, GE Healthcare Life Sciences, Pittsburgh, PA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Hyclone, Waltham, MA) and 2 mM L-GlutaMAX (Gibco BRL Life Technologies, Inc., Gaithersburg, MD). All cell lines were maintained in a humidified atmosphere containing 5% carbon dioxide (CO₂) at 37°C.

We synthesized a human, codon-optimized 2^nd generation CAR with specificity against PSCA using the published scFv sequences (18, 19), which was cloned in-frame with the IgG2-CH3 domain (spacer), CD28 co-stimulation domain, and the zeta (ζ) chain of the T cell receptor (TCR) CD3 complex in an SFG retroviral backbone to make the 2nd generation CAR. We generated a 2^nd generation CAR with specificity against MUC1 using the published scFv sequence (20–22), which was cloned in-frame with the IgG2-CH3 domain (spacer), 41BB co-stimulation domain, and the ζ chain of the TCR CD3 complex in an SFG retroviral backbone. To generate the IL-7 construct, we synthesized (DNA 2.0, Menlo Park, CA) a codon-optimized sequence encoding the signal peptide and extracellular domain of human IL-7, with restriction sites Xho1 and Sph1 incorporated up and downstream, respectively. To generate the IL-7Rα construct, we synthesized (DNA 2.0, Menlo Park, CA) a codon-optimized sequence encoding the signal peptide and extracellular domain of the human IL-7α, with restriction sites Xho1 and Sph1 incorporated up and downstream, respectively. The IL-7 and IL-7Rα DNA inserts were incorporated into SFG retroviral vectors that contained mOrange and GFP fluorescent markers, respectively. Retroviral supernatant for these vectors was generated by transfection of 293T cells. 3x10⁶ 293T cells were plated in 10 cm tissue culture treated dishes in 10 mL Iscove’s Modified Dulbecco’s Medium (IMDM, GE Healthcare Life Sciences, Pittsburgh, PA) supplemented with 10% heat-inactivated FBS and 2 mM L-GlutaMAX. A day later, cells were transfected with PegPam, RDF, and the DNA construct loaded into GeneJuice Transfection Reagent (Millipore, Burlington, MA) (23). Supernatant from the transfected cells were collected at 48 and 72 hrs, pooled together, and stored at -80°C until transduction.

To generate CAR T cells, 1x10⁶ PBMCs were plated in each well of a non-tissue culture-treated 24-well plate that had been pre-coated with OKT3 (1 mg/ml) (Ortho Biotech, Inc., Bridgewater, NJ) and CD28 (1 mg/ml) (Becton Dickinson & Co., Mountain View, CA). Cells were cultured in complete medium (RPMI-1640 containing 45% Clicks medium (Irvine Scientific, Inc., Santa Ana, CA), 10% FBS and 2 mM L-GlutaMAX), which was supplemented with recombinant human IL-2 (50 U/mL, NIH, Bethesda, VA) on day 1. On day 3 post OKT3/CD28 T blast generation, 1 mL of retroviral supernatant was added to a 24-well non-tissue culture-treated plate pre-coated with recombinant fibronectin fragment (FN CH-296; Retronectin; Takara Shuzo, Otsu, Japan) and centrifuged at 2000G for 90 minutes. OKT3/CD28 activated T cells (0.2 x 10⁶/mL) were resuspended in complete medium supplemented with IL-2 (100U/mL) and then added to the wells and centrifuged at 400G for 5 minutes. To co-express IL-7 and IL-7Rα with PSCA CAR and MUC1 CAR respectively, activated T cells were transduced sequentially on days 3 and 4 – first with the CAR to generate C.P and C.M cells, and then with either IL-7 or IL-7Rα respectively to generate C.P7 and C.M7R T cells. Transduction efficiency was measured 3 days after the final transduction by flow cytometry.

We generated CAPAN1 cell line that expressed transgenic GFP/FFLuc to facilitate bioluminescence imaging. To do this, GFP/FFLuc retroviral supernatant was plated in a non-tissue culture-treated 24-well plate (1 ml/well), which was pre-coated with a recombinant fibronectin fragment. CAPAN1 cells (0.2x10⁶/mL) were added to the plates (1 mL/well) and then transferred to a 37°C, 5% CO₂ incubator. Transgene expression was analyzed by flow 1-week post-transduction. Cells were subsequently sorted based on GFP using a MoFlo flow cytometer (Cytomation, Fort Collins, CO).

The following antibodies were used in this study for T cell phenotyping: CD3-APC, CD4-Krome Orange, CD8-Pacific Blue, CD127-AF00, CD3-AF700 (Beckman Coulter Inc. Brea, CA). PSCA and MUC1 antigen expression by tumor cells was measured using anti-PSCA and anti-MUC1 (Santa Cruz Biotechnology. Inc., Dallas, TX), respectively. CAR molecules were detected using Goat anti-human F(ab’)2 antibody conjugated with AlexaFluor647 (109-606-097) (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Cells were stained with saturating amounts of antibody (~5uL) for 20 min at 4°C, washed (PBS, Sigma-Alrich, St. Louis, MO), and then acquired on Gallios™ Flow Cytometer (Beckman Coulter Inc., Brea, CA). Analysis was performed using Kaluza^® Flow Analysis Software (Beckman Coulter Inc.).

The cytotoxicity and specificity of engineered T cells was evaluated in a standard 4–6 hr ⁵¹Cr-release assay, using E:T ratios of 40:1, 20:1, 10:1, and 5:1. Effector T cells were co-incubated in triplicate with target cells labeled with ⁵¹Cr in a V-bottomed 96-well plate. At the end of the incubation period at 37°C and 5% CO₂, supernatants were harvested, and radioactivity counted in a gamma counter. Percentage of specific lysis was calculated as follows: % specific cytotoxicity = [experimental release (cpm) – spontaneous release (cpm)]/[maximum release (cpm) – spontaneous release (cpm)] x 100.

To measure production of IL-7 by C.P7 T cells, irradiated K562 cells expressing PSCA antigen were used to stimulate C.P7 T cells at different E:T ratios. Supernatant samples were collected at different time points to measure IL-7 levels by ELISA. To evaluate growth of C.M and C.M7R T cells, 1x10⁶ irradiated CAPAN1 tumor cells were used to stimulate equal number of T cells in the absence or presence of IL-7 (10ng/mL). Cells were quantified manually and by flow cytometry. Mixed T cell stimulation assay was setup by mixing 1:1 (2.5x10⁵ cells each) of C.P T cells with C.M or C.M7R T cells, and C.P7 T cells with C.M or C.M7R T cells in presence of 0.5x10⁶ irradiated CAPAN1 tumor cells. Cells were quantified periodically using flow cytometer by using CountBright™ Absolute Counting Beads (Invitrogen, Eugene, OR) and 7-AAD was added to exclude dead cells. Acquisition was halted at 2,500 beads and manufacturer’s instructions were used to calculate total cell numbers.

Human Duoset ELISA kits from R&D Systems (Minneapolis, MN) were used to measure IL-7, IFN-γ, and TNF-α cytokines produced by T cells. Supernatant samples to measure IL-7 production by T cells were collected at indicated time points and stored at -20°C. IFN-γ and TNF-α were measured in supernatant obtained from T cells cultured with tumor cells overnight (~20 hours) and stored at -20°C. All samples were thawed and processed on ice for ELISA, which was performed according to the manufacturer’s instructions.

For co-culture experiments, GFP/FFLuc expressing CAPAN1 tumor cells (1x10⁶ cells) were inoculated into 3D algimatrix bioscaffold (Thermo Fisher Scientific, Inc., Waltham, MA) and cultured in 6-well G-Rex devices (Wilson Wolf Manufacturing, New Brighton, MN). Three days later, 5x10⁴ C.M, 5x10⁴ C.P, C.P+C.M (2.5x10⁴ each), or C.P7+C.M7R (2.5x10⁴ each) T cells were added to tumor cells. Anti-tumor activity was monitored using the IVIS Lumina In Vivo Imaging system (Caliper Life Sciences, Hopkinton, MA) 10 minutes after adding D-luciferin (PerkinElmer, Waltham, MA) (15 mg/mL) into the culture medium. Cells were then collected and recovered from the Algimatrix using Aligmatrix dissolving buffer (Thermo Fisher Scientific, Inc.). To quantify cells by flow cytometry we used CountBright™ Absolute Counting Beads (C36950; Invitrogen, Eugene, OR) and 7-AAD was added to exclude dead cells. Acquisition was halted at 2,500 beads.

Twelve-well tissue-culture-treated plates (Corning, NY, USA, catalog no. 353043) were coated with 200 µL of 1% (W/V) agarose (Fisher Scientific, Pittsburgh, PA, catalog no. BP160-500). 1x10⁶ Capan-1 GFP/FFLuc-labeled tumor cells were resuspended in 1 mL of T-cell medium and seeded in a final volume of 2 mL. The plates were incubated at 37°C to allow spheroid formation for 48 hours. Subsequently, the spheroids were transferred to a 6-well G-Rex (Wilson Wolf Manufacturing, New Brighton, MN, catalog no. 80240M) preconditioned with 30 mL of T-cell medium per well. Spheroid growth was monitored using IVIS Lumina In Vivo Imaging system (Caliper Life Sciences, Hopkinton, MA) 10 minutes after adding D-luciferin (PerkinElmer, Waltham, MA) (15 mg/mL) into the culture medium at indicated time intervals.

To test the cytotoxicity of effector cells, co-culture experiments were performed to assess the in vitro anti-tumor activity of T cells against tumor target cells. Capan-1 GFP/FFLuc-labeled tumor spheroids were generated as previously described. Two days later, they were transferred to a 6-well G-Rex in a final volume of 29 mL of T-cell medium per well. Initial bioluminescence signal produced by spheroids was measured before T cell treatment using the IVIS Lumina In Vivo Imaging system 10 minutes after adding D-luciferin. GFP- or mOrange-labeled effector T cells were then added, resuspended in 1 mL of T-cell medium, resulting in a final volume of 30 mL per well. Subsequent luminescence signals were captured at the indicated intervals as described above. Cells from the wells were collected on day 14, labeled with a CD3-APC antibody (Beckman Coulter, Indianapolis, IN, catalog no. IM2467U) to distinguish GFP- or mOrange-labeled T cells, and mixed with CountBright™ Absolute Counting Beads (Invitrogen, Eugene, OR, catalog no. C36950) and 7-AAD (BD Biosciences, Franklin Lakes, NJ, catalog no. 51-68981E) to exclude dead cells. Absolute cell numbers were calculated according to CountBright™ beads manufacturer’s recommended protocol. Data analysis was performed using Kaluza^® Flow Analysis Software version 2.1.

The in vivo tumor study in mice was performed according to a protocol approved by the Baylor College of Medicine Institutional Animal Care and Use Committee (Protocol #AN-5639). To assess the in vivo activity of the binary approach against bulky pancreatic tumors, 6 to 8-week-old female and male NSG mice (NOD.Cg-Prkdcscid IL-2rgtm1Wjl/SzJ, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) were injected subcutaneously (s.c.) into the left flank with 5x10⁶ CAPAN-1 cells suspended in 1X DPBS. Once the tumor reached a size of approximately 300 mm3 (around 4–5 weeks), mice were divided into three groups. Animals were then intravenously (i.v.) injected with GFP/FFLuc⁺ non-functional (lacking intracellular signaling components) CAR-expressing control T cells (4x10⁶), C.P+C.M T cells (2x10⁶ each), or C.P7+C.M7R T cells (2x10⁶ each). Tumor volume was monitored by caliper measurements weekly and calculated using the formula: length x width x width/2. T cell expansion was monitored based on bioluminescence signal from the infused T cells using the IVIS Lumina In Vivo Imaging system, 10 minutes after intraperitoneal (i.p.) injection of a 100 µL working solution of 15 mg/ml D-Luciferin (Perkin Elmer, Boston, MA). Mice were euthanized once the tumor volume reached the protocol limit (1500 mm³) or in the event of tumor ulcerations that grew >2 mm in diameter despite treatment or recurred. Mice were monitored for manifestations of toxicities and infections by research personnel as well as BCM veterinary staff, including measuring body weight and observing visual symptoms of illness (e.g., ruffled coat, hair loss, and/or skin redness). T cell luminescence data was analyzed using Living Image software (Caliper Life Sciences, Hopkinton, MA).

Data are reported as average ± standard error of mean (mean± SEM) unless stated otherwise in the figure legends. All statistical analyses were performed using GraphPad PRISM software version 10.3.1 (GraphPad Software, Inc.). Statistical differences between groups were analyzed by t-tests, one-way ANOVA, or two-way ANOVA as indicated in the figure legends. P value less than 0.05 was regarded as statistically significant.

To target PSCA and MUC1, both TAAs known to be frequently expressed in pancreatic tumors (24, 25), we generated two human codon-optimized second generation CARs with specificity for PSCA (C.P) and MUC1 (C.M), containing CD28 and 41BB co-stimulatory endodomains, respectively, which were fused to the CD3ζ activation motif ( Figure 1A ) (26, 27). Retroviral transduction led to efficient expression of both CARs in healthy donor-derived primary human T cells, as illustrated in Figure 1B . Next, using a standard 5-hr ⁵¹Cr-release assay, we evaluated the cytolytic activity of these CAR-modified cells against CAPAN1 pancreatic cancer cells, which express both target TAAs, albeit at different levels (PSCA – 95.1% and MUC1 – 68.5%; Figure 1C ). CAPAN1 cells were killed by both the C.P ( Figure 1D ) and C.M ( Figure 1E ) while demonstrating minimal reactivity against antigen negative 293T cells. As anticipated, superior killing was associated with higher target antigen expression levels. To next assess the long-term anti-tumor activity of these T cells, either alone or in combination, we utilized a previously established extended in vitro co-culture system (27). GFP/FFLuc-modified CAPAN1 tumor cells (CAPAN1.GFP/FFLuc) embedded in Algimatrix were cultured with C.P, C.M, or both (C.P:C.M; 1:1) in 6-well G-Rex devices, and monitored by bioluminescence imaging. As shown in Figure 1F , untreated tumor cells grew unimpeded, while treatment with individual CAR products each demonstrated some level of tumor control. However, co-administration of both C.P and C.M T cells at 1:1 ratio with total (C. P+C.M) T cell numbers equal to single-treated conditions, demonstrated superior anti-tumor effects. Nevertheless, tumor regrowth was eventually detected in all conditions. We confirmed that this was not due to antigen loss by evaluating PSCA and MUC1 surface levels on outgrowing tumor cells (on day 20), which demonstrated expression of both antigens comparable to the untreated condition ( Figure 1G ). Interestingly, a second administration of T cells (on day 27) resulted in a decline in tumor bioluminescence ( Figure 1H ). This outcome illustrates the targetability of the residual tumor cells and implicates limited T cell persistence as the likely cause of tumor regrowth. Taken together, these results demonstrate that targeting more than one TAA in heterogeneous tumors improves overall tumor killing. However, suboptimal T cell persistence limits this benefit, resulting in a lack of durable anti-tumor responses.

Potent and sustained T cell activation, persistence and anti-tumor activity requires three stimulatory signals (signal 1 – antigen recognition, signal 2 - co-stimulation, and signal 3 - cytokine) (28). While our C.P and C.M cells receive signals 1 (CD3ζ) and 2 (CD28 or 41BB co-stimulatory domains, respectively) through the CAR, they lack signal 3. Thus, we explored whether the incorporation of transgenic cytokine support would promote superior anti-tumor effects. We evaluated the potential of various cytokines including IL-2 and IL-15, both of which have been shown to effectively promote T cell expansion (29, 30). However, IL-2 and IL-15 can be utilized by inhibitory cells including tumor-infiltrating regulatory T cells, which are known to suppress the anti-tumor function of infused CAR T cells (31–34). Furthermore, systemic administration of these cytokines can cause abnormalities such as capillary leak syndrome, hypotension, fever, and vomiting due to systemic toxicity affecting multiple organ systems (35, 36). To balance the need to support T cell expansion without promoting toxic side effects, we therefore designed a binary T cell strategy in which C.P cells were modified to secrete IL-7 (C.P7) (a homeostatic cytokine that augments the proliferation and survival of T cells, particularly naïve and memory subsets), and C.M cells to overexpress the IL-7Rα (C.M7R).

To test the feasibility of this binary system, we first transduced C.P cells with a retroviral vector encoding the human IL-7 gene linked to mOrange fluorescent protein by an IRES sequence ( Figure 2A , top). T cells were efficiently transduced to express both CAR and IL-7 transgenes ( Figure 2A , bottom, representative donor), with an average of >85% double-transduced cells ( Figure 2B ). As measured by ELISA, antigen (K562 cells engineered to express PSCA) stimulated C.P7 T cells secreted IL-7 ( Figure 2C ), which progressively increased with increasing antigen exposure ( Supplementary Figure S1A ) and duration of culture ( Supplementary Figure S1B ). We confirmed that C.P7 T cells retained their target antigen specificity and ability to produce effector cytokines. In a standard chromium release assay, C.P7 T cells were able to selectively kill PSCA-expressing targets ( Supplementary Figure S1C ) and exhibited a cytokine profile that was comparable to C.P cells ( Supplementary Figure S1D ). Taken together, our results demonstrate that C.P cells can be engineered to efficiently secrete IL-7, which is produced in an antigen stimulation-dependent manner without compromising the antigen specificity or anti-tumor function of the CAR.

While both chains of the IL-7R complex – IL-7Rα (CD127) and common γ chain (CD132) are endogenously expressed by lymphocytes, including T cells, native IL-7Rα expression is known to be downregulated upon exposure to IL-7 (37, 38). Additionally, antigenic stimulation has also been reported to reduce IL-7Rα expression by T cells in a non-IL-7 dependent manner (39). We explored the regulation of endogenous IL-7Rα in our C.M T cells and, consistent with previous findings, confirmed that upon IL-7 cytokine or antigen stimulation, the receptor was downregulated ( Supplementary Figure S2A ). To enhance the capacity of C.M cells to utilize C.P7-produced IL-7 ( Figures 2A–C , S1 ), we transduced C.M cells with a retroviral vector encoding human IL-7Rα linked to GFP by an IRES sequence ( Figure 2D , top). This resulted in the production of dual-transgenic T cells ( Figure 2D , bottom) with superior IL-7Rα expression compared to C.M cells ( Figure 2E ). Notably, this receptor overexpression resulted in greater expansion of C.M7R T cells compared to C.M cells when stimulated with irradiated CAPAN1 tumor cells in the presence of recombinant IL-7 ( Figure 2F ), consistent with enhanced IL-7 utilization.

Finally, to ensure that overexpression of IL-7Rα did not negatively impact the functional capacity of our CAR T cells, we assessed antigen-specificity, effector cytokine production, and responsiveness to other common γ chain family cytokines. As illustrated in Supplementary Figures S2B , S2C , the antigen specificity and cytolytic function of C.M7R T cells was similar to that of C.M T cells, as was the secretion of effector cytokines IFN-γ and TNF-α upon exposure to the tumor cells. Furthermore, C.M7R cells retained their ability to respond to other common γ chain family cytokines such as IL-2 and IL-15, illustrated by the similar growth pattern of C.M and C.M7R T cells ( Supplementary Figures S2D, E ). In summary, these results demonstrate that engineering C.M cells to overexpress IL-7Rα improves their ability to utilize IL-7 without adversely impacting their antigen specificity, effector cytokine production capacity, and ability to utilize other common γ chain cytokines such as IL-2 and IL-15.

We next evaluated whether the combination of C.P7 and C.M7R cells supported superior T cell expansion. First, we explored whether the quantity of IL-7 produced by C.P7 cells following antigen exposure was sufficient to drive T cell expansion. To assess this, we cultured C.M7R cells in conditioned medium harvested 48 hours after stimulating C.P7 cells or C.P (control lacking IL-7) with irradiated CAPAN1 tumor cells at E:T of 1:1. As shown in Figure 3A , expansion of C.M7R cells in C.P7 conditioned medium was significantly greater than that seen in cytokine-free medium or C.P conditioned medium. Next, to assess if antigen-activated C.P7 cells were able to support C.M7R cell expansion, we co-cultured both T cell populations with irradiated CAPAN1 tumor cells (E:E:T ratio of 1:1:2) and evaluated cell outgrowth by flow cytometry. As illustrated in Figure 3B , co-stimulating C.M or C.M7R cells with non-cytokine-producing C.P T cells (green lines) resulted in only short-term expansion followed by rapid T cell contraction during the second week of culture. In contrast, co-stimulation of C.M or C.M7R cells with C.P7 T cells produced a marked increase in total T cell numbers (red lines), with superior output seen in the binary condition (solid red line). Further in-depth analysis of the binary condition and separation of C.P7 from C.M7R cells based on GFP (surrogate for transgenic IL-7Rα) expression revealed the presence of both cell types, though the C.M7R population dominated, consistent with their ability to utilize C.P7 produced cytokine ( Figure 3C ).

One potential advantage of this binary T cell approach is the ability to control and support proliferation and persistence of the receptor-expressing population, with the strategic delivery of the cytokine-producing cell component. To test whether repeated administration of C.P7 would indeed support this outcome, we cultured C.M7R cells with either non-transduced (NT) control T cells, C.P, or C.P7 T cells stimulated with irradiated CAPAN1 tumor cells for 48 hours at the E:T of 1:1 and monitored the expansion profile of C.M7R cells. As shown in Figure 3D , we observed robust C.M7R T cell expansion on day 5 post co-culture with activated C.P7 cells, while limited or no expansion was detected when co-cultured with NT or C.P cells. Furthermore, periodic addition of activated C.P7 T cells (every 5–6 days) resulted in a boosting effect, producing an increase in C.M7R cell numbers, whereas a single administration of stimulated C.P7 cells resulted in an eventual decrease in the number of C.M7R T cells. In summary, these results demonstrate that transgenic IL-7 produced by C.P7 cells supports T cell expansion, particularly of cells engineered to overexpress IL-7Rα. Furthermore, strategic administration of C.P7 cells can be used to prolong the expansion and persistence of C.M7R cells.

To evaluate if the binary strategy results in enhanced and durable anti-tumor effects, we compared the tumor killing ability of single (C.P or C.M, individually), dual (C.P+C.M), and the binary T cells (C.P7+C.M7R) in the long-term co-culture assay with CAPAN1 tumor cells described in Figure 1 . An initial decrease in tumor luminescence was observed in all treatment conditions, indicating tumor-killing by CAR T cells ( Figure 4A , representative tumor wells and Figure 4B , summarized results). However, the tumors eventually outgrew by day 20 ( Figure 4B ) in the C.P, C.M, and C.P+C.M treatment conditions as seen previously ( Figure 1F ). In contrast, the C.P7+C.M7R T cell combination produced superior anti-tumor effects with no detectable tumor outgrowth ( Figures 4A, B ). This enhanced anti-tumor activity, as anticipated, coincided with increased total T cell expansion/persistence in the binary treatment group ( Figure 4C ).

3D culture systems more closely replicate in vivo characteristics with tumor spheroids providing a model that better mimics the solid tumor architecture and physiology (40). Hence, we next explored the benefits of our binary strategy with pancreatic cancer cell-derived 3D tumor spheroids. We first established spheroids from GFP/FFLuc-labeled CAPAN1 tumor cells and optimized the growth conditions in 6-well G-Rex devices to ensure spheroid viability for an extended period of time ( Supplementary Figures S3A, B ). To confirm that CAPAN1 TAA expression was maintained in 3D cultures, we isolated tumor cells at two different time points (day 4 and 25) after spheroid formation and measured the expression of PSCA and MUC1, both of which were maintained throughout the culture period of 25 days ( Supplementary Figure S3C ). To assess anti-tumor activity we next co-cultured tumor spheroids alone (Tumor only), and with C.P+C.M, or C.P7+C.M7R T cells at the calculated E:T (based on initial number of tumor cells used to generate spheroids) of 1:40. Bioluminescence imaging of the spheroids demonstrated a decline in tumor burden in both T cell treated conditions ( Figure 5A , representative wells and Figure 5B , summarized results). However, the anti-tumor response in the C.P+C.M treated conditions, as observed in the previous co-cultures ( Figure 4 ), was less robust compared to the C.P7+C.M7R treatment conditions ( Figures 5A, B ). To confirm if this enhanced tumor control in the binary condition was due to improved T cell expansion/persistence, we isolated and quantified T cells from the co-cultures on day 14. Similar to our observations in the Algimatrix co-culture system ( Figure 4C ), higher numbers of total T cells were quantified in the C.P7+C.M7R combination compared to C.P+C.M treatment ( Figure 5C ). These results were validated in a pancreatic cancer xenograft model ( Supplementary Figure S4 ) in which the binary strategy resulted in improved anti-tumor activity against bulky subcutaneous CAPAN-1 tumors ( Supplementary Figure S4A ), coincident with enhanced total T cell expansion ( Supplementary Figure S4B ). In summary, these data demonstrate the binary system we developed to augment CAR T cell activity against heterogeneous tumors is effective against the physiologically more relevant 3D spheroids and confirm the results of our previous co-culture assessments.