Syndecan-1-targeted therapeutic antibody impairs macropinocytosis and elicits antitumor immunity in pancreatic cancer

Summary

Pancreatic ductal adenocarcinoma (PDAC) remains one of the deadliest malignancies, with a 5-year survival rate of just 13%. While the development and early clinical use of small molecules targeting oncogenic KRAS mutations, key drivers of PDAC, have shown promise, resistance to these targeted therapies remains a significant challenge. We recently identified Syndecan-1 (SDC1), a highly expressed heparan sulfate proteoglycan, as a critical KRAS effector protein that promotes nutrient salvage and tumor growth. Here, we report the development of a human-specific monoclonal antibody (anti-SDC1 mAb) that inhibits PDAC cell proliferation in vitro and suppresses PDAC tumor growth in vivo. Mechanistically, the anti-SDC1 mAb blocks macropinocytosis and induces antibody-dependent cellular cytotoxicity (ADCC). In vivo, anti-SDC1 mAb synergizes with standard chemotherapy, KRAS^∗ inhibitors, and immunotherapies, resulting in tumor regression and near-complete response. These findings highlight the anti-SDC1 mAb as a promising therapeutic strategy for PDAC and potentially other KRAS^∗ and SDC1-driven tumors.

Graphical abstract

Highlights

• •SDC1 is highly expressed and functionally essential across multiple cancer types
• •Anti-SDC1 mAb inhibits macropinocytosis and triggers antitumor immune response
• •Anti-SDC1 mAb produces strong tumor suppression across diverse preclinical models
• •Combining anti-SDC1 mAb with standard therapies further enhances tumor control

We identify SDC1 as a key driver of PDAC progression and resistance to KRAS inhibitors and develop a therapeutic-grade antibody that suppresses tumor growth. The anti-SDC1 mAb blocks macropinocytosis, induces ADCC, and synergizes with chemotherapy, KRAS inhibitors, and immunotherapies, establishing SDC1 targeting as a promising PDAC therapy.

Introduction

Mutations in KRAS that result in a constitutively activated oncoprotein (hereafter referred to as KRAS^∗) are present in over 90% of pancreatic ductal adenocarcinoma (PDAC) cases and play a pivotal role in cancer progression, making KRAS^∗ an ideal therapeutic target for treating PDAC.¹ Preclinical studies have demonstrated that genetic KRAS^∗ ablation and the administration of allele-specific pharmacological inhibitors can significantly suppress tumor malignancy in PDAC models.¹ While the recent approval of KRAS^G12C inhibitors sotorasib (Lumakras) and adagrasib (Krazati) for advanced non-small cell lung cancer has spurred the development of additional allele-specific KRAS^∗ inhibitors, such as KRAS^G12D inhibitors MRTX1133 and RMC-9805, none have yet been approved for the treatment of KRAS-mutant PDAC despite ongoing clinical trials.² Furthermore, evidence from preclinical studies assessing KRAS^∗ inhibitors in inducible Kras^∗ PDAC models, as well as clinical observations, reveals that pancreatic cancer cells can rapidly circumvent their dependency on KRAS^∗, thereby limiting the therapeutic benefit of KRAS^∗ inhibitors.^3,4 These findings highlight the urgent need to elucidate the mechanisms underlying acquired resistance to KRAS^∗ inhibition for developing combination strategies that effectively suppress KRAS^∗ signaling and its compensatory mechanisms to improve disease management and therapeutic outcomes.

Our previous studies have established the critical role of the heparan sulfate proteoglycan, Syndecan-1 (SDC1), in driving PDAC progression and maintenance.^5,6 Specifically, we demonstrated that SDC1, a downstream surrogate of KRAS^∗, drives macropinocytosis, enabling PDAC cells to scavenge extracellular proteins and nutrients that sustain tumor growth.⁵ Notably, we further reveal that SDC1 is highly expressed on the surface of PDAC cells with acquired resistance to either genetic or pharmacological inhibition of KRAS^∗.⁴ We found that SDC1 functions to sustain macropinocytic activity and tumor maintenance in these resistant cells, underscoring its critical role in overcoming therapeutic challenges. The universally high expression of SDC1 in PDAC, combined with its critical role in the progression of both KRAS^∗-dependent and -independent pancreatic cancer cells, highlights its potential as a therapeutic target for PDAC treatment.

Multiple therapeutic modalities, including antibody-drug conjugates (ADCs) and engineered immune cells, have been developed to target SDC1 or its related pathogenic signaling pathways. Most existing SDC1 antibodies trace back to the naked anti-SDC1 clone nBT062.^6,7 This clone binds SDC1 to deliver cytotoxic payloads and facilitate effector cell-mediated killing,^6,7 but it does not directly target SDC1 function. Critically, the impact of these agents on SDC1-mediated macropinocytosis, particularly in PDAC, remains largely undefined, and none has advanced to clinical use. Accordingly, there is a significant unmet need for therapies that directly target SDC1 and SDC1-mediated oncopathogenesis, especially within the context of pancreatic cancer treatment.

Therapeutic antibodies have emerged as a cornerstone in cancer therapeutics, offering highly specific and versatile approaches to target malignant cells.⁸ Moreover, their specificity is particularly advantageous for cell-surface antigens, such as SDC1, which represents a promising avenue for developing innovative cancer therapies. In this study, we developed and evaluated a monoclonal anti-SDC1 therapeutic antibody (anti-SDC1 mAb). The anti-SDC1 mAb demonstrated exceptional binding ability and specificity to human SDC1 protein and SDC1-expressing human PDAC cells. Notably, it substantially inhibited PDAC cell growth in vitro and exhibited desirable tumor-suppressive effects in vivo irrespective of KRAS mutation status or resistance to KRAS^∗ inhibitors. Moreover, its therapeutic effects could be further enhanced when combined with treatments such as KRAS^∗ inhibitors and immunotherapies. In summary, our anti-SDC1 mAb not only offers a promising alternative strategy for PDAC treatment but also illuminates the potential of targeting SDC1 in other malignancies characterized by mutant KRAS or aberrant SDC1 expression.

Results

SDC1 is highly expressed and critical for disease progression in human cancers

We previously demonstrated the essential role of SDC1 in driving tumor progression in both KRAS^∗-dependent and -independent PDAC mouse models.^4,5 Building on these findings, we sought to assess the broader clinical relevance of SDC1 by analyzing its expression patterns across various human cancer types. Immunohistochemistry (IHC) staining revealed high expression of SDC1 in approximately 90% of patient-derived pancreatic cancer xenograft (PDX) tissues (40 out of 45), while flow cytometry analysis demonstrated elevated SDC1 expression in approximately 95% of PDAC cell lines (17 out of 18) (Figures 1A–1C, S1A, and S1B). Functionally, SDC1 knockout or knockdown significantly reduced PDAC cell proliferation in vitro (Figure 1D) and suppressed tumor growth in vivo (Figure 1E). Beyond pancreatic cancer, IHC staining also detected high SDC1 expression in lung, colorectal cancers, and cholangiocarcinoma patient samples (Figures 1F–1H).⁶ Functional assays further confirmed that SDC1 knockdown markedly reduced colony formation in A549 (lung), MB231 (breast), and HT29 (colorectal) cancer cells, indicating a broader role for SDC1 in tumorigenesis (Figures S1C and S1D). Clinical data further supported these findings, as elevated SDC1 expression was associated with poorer overall survival in patients with pancreatic cancer, breast cancer, or cholangiocarcinoma from The Cancer Genome Atlas (TCGA) datasets (Figures 1I, 1J, and S1E). No significant correlation between elevated SDC1 expression and worse prognosis, however, was observed in colon cancer or lung adenocarcinoma TCGA cohorts (Figures S1F and S1G). Collectively, these findings highlight the prevalence of elevated SDC1 expression across multiple cancer types and demonstrate its functional role in promoting tumor progression, supporting its potential as a therapeutic target.

Figure 1.

SDC1 is highly expressed in pancreatic cancer and is critical for PDAC progression

(A) Quantification of scores of a PDAC PDX tissue microarray (TMA) stained with anti-human SDC1 antibody. SDC1 expression is classified as negative (score = 0), low (score = 1), intermediate (score = 2), and high (score = 3).

(B and C) (B) Quantification of the ratio of human SDC1⁺ cells and (C) median fluorescence intensity of human SDC1 expression of selected human PDAC cell lines.

(D) Representative images and histogram quantification of the colony formation assay of PATC53 and MiaPaca2 cells infected with scrambled short hairpin RNA (shRNA) or shRNA against human SDC1. Colonies were visualized with crystal violet.

(E) Growth curve of subcutaneous tumors in NCr nude mice inoculated with parental or SDC1-knockout AsPC-1 cells. n = 5 per group.

(F–H) Quantification of scores of (F) lung cancer patient TMA, (G) colorectal cancer patient TMA, and (H) cholangiocarcinoma patient TMA stained with anti-human SDC1.

(I) Overall survival of patients in the TCGA PDAC cohort.

(J) Overall survival of patients in the TCGA invasive carcinoma of the breast cohort. Survival data were stratified by human SDC1 mRNA level. Data are represented as mean ± SD. Statistical significance was analyzed by ANOVA or unpaired Student’s t test. p < 0.05 and p < 0.001 are noted with ∗ and ∗∗∗, respectively.

Characterization of the anti-SDC1 mAb with a unique epitope and high specificity for human SDC1

We previously showed that SDC1-mediated macropinocytosis in PDAC cells requires its extracellular domain and plasma membrane localization.⁵ To target the SDC1 extracellular domain and neutralize its function to elicit antitumor effects, we generated ectodomain-specific monoclonal antibodies by immunizing SDC1^−/− mice with recombinant SDC1 ectodomain and producing hybridoma. As a reference standard, we used clone nBT062, a clinically validated anti-SDC1 antibody developed for ADC applications, to benchmark and characterize our anti-SDC1 clones. We first examined the binding ability of these monoclonal antibody (mAb) clones to both mouse and human SDC1 by ELISA and flow cytometry analysis. Among all tested clones, the lead clone, clone 22B, exhibited the highest binding ability to recombinant human SDC1 protein and SDC1-expressing human PDAC cells, with affinity comparable to that of the commercial nBT062 mAb (Figures 2A, 2B, S2A, and S2B).⁷ The superior binding performance was further validated by bio-layer interferometry, in which clone 22B demonstrated nanomolar binding to human SDC1 (Figure S2C). Cross-reactivity testing revealed that, while clones 28C and 33A displayed modest cross-reactivity to both human and mouse SDC1, clone 22B showed superior specificity for only human SDC1 (Figure S2D). This specificity is likely attributable to the sequence divergence between human and mouse SDC1, which share only ∼76% similarity, with few contiguous sections of identical amino acids.⁶ Notably, clone 22B demonstrated superior binding to cynomolgus SDC1 (cynoSDC1) compared to that of nBT062 (Figure S2E), suggesting its suitability for preclinical evaluation of toxicity and efficacy in cynomolgus models prior to clinical development.

Figure 2.

Characterization of anti-SDC1 mAbs

(A and B) Binding of selected anti-SDC1 mAb clones to (A) recombinant human SDC1 protein analyzed by ELISA, or (B) PATC53 cells analyzed by flow cytometry.

(C) Comparison and linear regression of median fluorescence intensity of clone 22B to cells with differential expression of human SDC1 measured by MI15.

(D) Human SDC1 level of parental or SDC1-knockout AsPC-1 cells stained with clone 22B and a PE-conjugated secondary antibody.

(E) Representative IHC images of human SDC1 expression in AsPC-1 and PATC53 PDAC tumors from NCr nude mice, or normal human pancreas, spleen, and thymus. Tissues were stained with nBT062 or clone 22B. Scale bars, 100 μm.

(F) Summary of epitope binning of nBT062, clone 22B, and MI15. Binding intensity was calculated by MFI (with blocking)/MFI (without blocking) × 100%.

(G) Binding of clone 22B to cyclic-restrained short peptides derived from human SDC1 by fluorescent ELISA.

(H) Binding of clone 22B to CHO-K1 cells expressing wild-type human SDC1 or mutant SDC1 with indicated truncations.

(I) Binding of clone 22B to PATC53 cells in the presence of various concentrations of soluble SDC1 analyzed by flow cytometry. Data are represented as mean ± SD.

To further evaluate the binding specificity of clone 22B to human SDC1, we manipulated SDC1 expression using CRISPR-Cas9-mediated SDC1 knockout to obtain different cell clones with various SDC1 expressions. We found that the binding intensity of clone 22B to PDAC cells showed a strong positive correlation with SDC1 expression levels, as measured using the validated commercial anti-SDC1 mAb MI15 (Figure 2C). As a control, SDC1 was not detected in human PDAC AsPC-1 cells with complete SDC1 knockout, confirming the specificity of clone 22B (Figures 2D and S2F). Further IHC staining revealed that clone 22B and clone nBT062 exhibited highly similar staining patterns in both human pancreatic cancer and normal tissues, indicating comparable SDC1 antigen recognition and specificity (Figures 2E and S2G).

To further characterize the binding profile of clone 22B, we conducted an epitope binning analysis to determine whether clone 22B binds a distinct epitope on SDC1 compared to nBT062 and MI15. Pre-staining SDC1-expressing PDAC cells with either nBT062 or MI15 did not diminish the binding intensity of clone 22B, indicating that it targets an epitope distinct from those recognized by nBT062 or MI15 (Figure 2F). To identify the specific epitope of clone 22B, we performed cyclic-restrained peptide-based conformational epitope mapping at the single-amino-acid resolution and identified two potential binding sites, DITLSQQ and DFTF (Figure 2G). Between these, the DITLSQQ site was associated with stronger binding by clone 22B. This was further validated using CHO-K1 cells expressing wild-type SDC1 or SDC1-truncated mutants lacking either the DITLSQQ or DFTF sequences. Flow cytometry analysis demonstrated that clone 22B binding was comparable between wild-type and DFTF-deleted SDC1-expressing cells but was strongly attenuated in cells lacking the DITLSQQ sequence (Figure 2H), confirming DITLSQQ as the primary epitope for clone 22B. Consistently, sequence alignment of human and mouse SDC1 demonstrated that the DITLSQQ epitope is present only in human SDC1 but absent in mouse SDC1,⁶ explaining the species specificity and lack of cross-reactivity of clone 22B (Figure S2D). Finally, as the binding site of clone 22B lies upstream of known SDC1 cleavage sites, we assessed whether soluble SDC1 can act as an antigen sink that interferes with the therapeutic efficacy of clone 22B. Using a flow cytometry-based binding assay, we found that a supraphysiological soluble SDC1 (500 ng/mL, 10-fold higher than reported patient levels⁹) produced only a modest reduction in clone 22B binding (Figure 2I), indicating limited interference by circulating soluble SDC1 under these conditions. In summary, anti-human SDC1 mAb clone 22B demonstrates exceptional binding ability and specificity to an epitope that is distinct from those recognized by commercial antibodies.

Clone 22B inhibits macropinocytosis-dependent cell proliferation of PDAC cells

Next, we investigated whether clone 22B could block SDC1-mediated macropinocytosis, a hallmark of PDAC cells driven by KRAS^∗ or with acquired resistance to KRAS inhibitors.^4,5 Treatment with clone 22B, but not isotype control or nBT062, significantly suppressed macropinocytosis, as evidenced by reduced uptake of tetramethylrhodamine-labeled 70 kDa dextran (TMR-dextran) in PDAC cells (Figures 3A and S3A). Given that PDAC cells exhibit a heightened dependency on glutamine for survival and proliferation,^5,10 glutamine deficiency is a known factor leading to the induction of macropinocytosis.¹¹ Accordingly, while PDAC cells exhibited decreased proliferation when cultured in medium with sub-optimal level of glutamine, the supplement of albumin, which can be uptaken by cells via macropinocytosis, rescued the cell proliferation defect (Figures 3B and S3B).^5,12 This thus serves as a functional assay for evaluating macropinocytic activity. Remarkably, treatment with clone 22B, but not nBT062 or the isotype control, completely abrogated this rescue effect (Figures 3B and S3B), indicating that clone 22B effectively disrupts macropinocytosis-dependent nutrient uptake. Notably, the disruption was specific to SDC1-mediated macropinocytosis, with no detectable effect on bulk fluid-phase endocytosis or on receptor-mediated endocytosis of other surface proteins, such as HER2 (Figures S3C and S3D). To further elucidate the mechanism underlying the growth defects induced by clone 22B, we assessed the cell cycle in antibody-treated PDAC cells. Clone 22B induced significant cell-cycle arrest under nutrient-deficient condition, characterized by a marked reduction in the proportion of cells in S-phase compared to the isotype group (Figure 3C). In line with this observation, mRNA sequencing (mRNA-seq) analysis revealed downregulation of genes associated with cell proliferation and cell cycle progression in cells treated with clone 22B (Figure 3D). Next, to determine whether the clone 22B binding motif DITLSQQ is required for the functional role of SDC1 in macropinocytosis, we performed a rescue experiment by re-expressing either wild-type SDC1 or a SDC1 mutant lacking the DITLSQQ motif in SDC1-knockdown PDAC cells. Wild-type SDC1 completely rescued the macropinocytosis defect induced by SDC1 knockdown, whereas the DITLSQQ-deleted mutant failed to restore macropinocytosis (Figure 3E). As a consequence, the mutant was also unable to rescue the impaired proliferation ability of SDC1-deficient cells, as demonstrated by both Incucyte-based proliferation assays and colony formation assays (Figure 3F). Collectively, these findings highlight the unique mechanism of action of clone 22B in inhibiting macropinocytosis and its downstream impact on cell cycle regulation and proliferation in PDAC cells.

Figure 3.

Anti-SDC1 mAb suppresses PDAC cell macropinocytosis and proliferation in a glutamine-dependent manner

(A) Representative images and quantification of macropinocytotic activity of PATC53 cells treated with clone 22B or isotype control. Macropinocytosis was visualized with 70 kDa TMR-dextran. Scale bars, 30 μm.

(B) Growth curve of MiaPaca2 cells cultured under nutrition-deficient conditions with 0.2 mM glutamine treated with indicated antibodies with or without supplementation of 2% BSA. EIPA (25 μM) treatment served as the positive control.

(C) BrdU cell cycle assay of AsPC-1 cells treated with clone 22B or isotype control. Cells were treated for 2 days before collecting for flow cytometry analysis.

(D) Gene set enrichment analysis of total mRNA extracted from AsPC-1 cells treated with clone 22B or isotype control. Cells were treated for 3 days before collecting for analysis.

(E) Representative images and quantification of macropinocytotic activity of SDC1-knockdown MiaPaca2 cells rescued by indicated SDC1 constructs. Macropinocytosis was visualized with TMR-dextran. Scale bars, 20 μm.

(F) Growth curve and colony formation of SDC1-knockdown MiaPaca2 cells rescued by indicated SDC1 constructs. Colonies were visualized by Incucyte.

(G–I) Growth curve of AsPC-1 cells cultured under (G) 0.2 mM, (H) 0.8 mM, or (I) 2.0 mM glutamine and treated with indicated antibodies, or with 25 μM EIPA. Data are represented as mean ± SD. Statistical significance was analyzed by ANOVA or unpaired Student’s t test. p < 0.01, p < 0.001, and p < 0.0001 are noted with ∗∗, ∗∗∗, and ∗∗∗∗, respectively. ns, non-significant.

We further assessed whether nutrient availability modulates the effect of clone 22B on cell proliferation. PDAC cells were cultured in medium containing 0.2 mM, 0.8 mM, and 2.0 mM glutamine, or in 0.8 mM glutamine with replenishment every 2 days. Under glutamine-restricted conditions (0.2 mM or 0.8 mM without replenishment), clone 22B significantly reduced cell proliferation compared to isotype or nBT062, with the effect pronounced at later time points under 0.8 mM glutamine, coincident with nutrient depletion (Figures 3G and 3H). No reduction was observed in 2.0 mM glutamine or in frequently replenished 0.8 mM glutamine (Figures 3I and S3E). These results indicate that nutrient depletion amplifies the anti-proliferative effects of clone 22B. Building on these findings, we speculated that concurrent treatment with clone 22B and a pharmacological inhibitor of glutamine metabolism under nutrient-replete conditions could phenocopy the effects observed in glutamine-limited media. Indeed, co-treatment with the glutamine antagonist DRP-104 and clone 22B in medium with 2.0 mM glutamine markedly suppressed cell proliferation relative to untreated controls (Figure S3F).^13,14 Taken together, our findings demonstrate that the anti-macropinocytic and anti-proliferative effects of clone 22B are potentiated by glutamine limitation, highlighting its potential therapeutic activity in the nutrient-deprived, glutamine-deficient microenvironment characteristic of PDAC.

Antitumor efficacy of anti-SDC1 mAb across diverse tumor models

The dependency of clone 22B’s anti-proliferative effects on glutamine concentration suggests that its in vivo therapeutic efficacy may be substantially influenced by the tumor microenvironment, prompting an evaluation of the antitumor efficacy of clone 22B in PDAC tumors in vivo. However, glutamine concentrations within PDAC mouse tumors derived from both genetically engineered mouse models and xenograft models were found to be comparable to physiological levels (0.6–0.8 mM).¹⁵ These levels in murine models align with the conditions under which our in vitro studies indicated that clone 22B has limited effects on cell proliferation (Figure S3E).¹⁵ Nevertheless, the inhibition of SDC1-mediated macropinocytosis resulted in significantly reduced internalization of clone 22B than nBT062 (Figures S4A and S4B). Reducing antibody internalization by inhibiting macropinocytosis was found to enhance its antibody-dependent cellular cytotoxicity (ADCC) effect (Figure S4C).^16,17,18 To leverage this unique characteristic and synergistically enhance the ADCC activity and thereby improve the antitumor efficacy of clone 22B, we engineered a defucosylated form of clone 22B (def-22B) by knocking out α1,6-fucosyltransferase (encoded by the FUT8 gene) in CHO-S cells (Figures S4D and S4E).¹⁹ The defucosylation of clone 22B by FUT8 deletion significantly improved its binding to FcγRs (Figure S4F), inducing stronger activation of ADCC reporter cells (Figure S4G) and boosting the cytotoxic killing of target cells compared to its fucosylated form (Figures 4A and S4H). Additionally, the supplement of soluble SDC1 at a pathophysiological concentration did not interfere with the ADCC activity of def-22B against target PDAC cells, indicating that circulating soluble SDC1 is insufficient to meaningfully impair 22B-mediated cytotoxicity under these conditions (Figure S4I). Importantly, def-22B retained its binding affinity and specificity to human SDC1 comparable to the parental clone (Figures S4J and S4K).

Figure 4.

Anti-SDC1 mAb demonstrates significant antitumor effects in various tumor models

(A) NK92-CD16-mediated killing assay to Panc02-hSDC1 cells. Target cells were treated with parental clone 22B, def-22B, isotype, or vehicle control.

(B) Human SDC1 expression of parental and human SDC1-overexpressed Panc02 cells stained with clone 22B or MI15.

(C and D) Growth curve of (C) Panc02-hSDC1 and (D) HY55582-hSDC1 tumors subcutaneously inoculated in C57BL/6NJ mice, treated with def-22B or vehicle control. n = 5 for each group.

(E) Quantification of the weights of HY55582-hSDC1 subcutaneous tumors treated with def-22B or vehicle. Tumors were collected 21 days post-inoculation. n = 5 per group.

(F) Representative images of subcutaneously inoculated Panc02-hSDC1 tumors, treated with def-22B or vehicle. Tumors were collected 30 days post-inoculation.

(G and H) Growth curve of (G) PATC53 and (H) AsPC-1 tumors subcutaneously inoculated in NCr nude mice, treated with def-22B or vehicle. n = 5 per group.

(I) Representative MRI images of orthotopic AsPC-1 tumors treated with def-22B or vehicle. Tumors were inoculated into the pancreata of NCr nude mice. MRI was performed on day 37 post-tumor inoculation. Data are represented as mean ± SD. Statistical significance was analyzed by ANOVA or unpaired Student’s t test. p < 0.05, p < 0.01, and p < 0.001 are noted with ∗, ∗∗, and ∗∗∗, respectively.

To assess the in vivo therapeutic potential of def-22B, we first engineered murine PDAC cell lines, Panc02 (Kras^WT), HY55582 (Kras^G12D), and HY50760 (Kras^G12C),²⁰ to express human SDC1, thereby permitting evaluation in syngeneic allograft models with intact immune compartments. Flow cytometry analysis confirmed strong binding of clone 22B to human SDC1-overexpressing Panc02 cells (Panc02-hSDC1) but no binding to parental cells lacking human SDC1 expression (Figure 4B). C57BL/6NJ mice inoculated subcutaneously with human SDC1-expressing PDAC cells were subsequently randomized into treatment groups after tumor establishment. Treatment with 100 μg def-22B twice weekly significantly suppressed tumor growth across all three murine PDAC models, as evidenced by substantially reduced tumor weights at the study endpoint compared to those in the isotype control group (Figures 4C–4F and 7I). Further in vivo studies were conducted in NCr nude mice inoculated with AsPC-1 or patient-derived tumor cells (PATC53). In NCr nude mice, def-22B demonstrated robust antitumor efficacy, with some mice achieving complete tumor regression by the study endpoint (Figures 4G and 4H). MRI analysis of AsPC-1-derived orthotopic tumors in NCr nude mice further confirmed significantly reduced tumor volume in the def-22B-treated mice compared to the isotype-treated mice by day 37 (Figure 4I). Conversely, clone 22B showed no therapeutic effect in NSG mice (Figure 5A), indicating the potential contribution of specific immune cell populations to its antitumor activity. On-target specificity and dependency on human SDC1 expression were further confirmed in vivo showing that def-22B had no effect on tumor growth in Panc02 allografts in C57BL/6NJ mice (Figure 5B) or in AsPC-1 xenografts with complete human SDC1 knockout in NCr nude mice (Figure S6A). Additionally, given the aberrant overexpression of SDC1 in other human cancers, such as triple-negative breast cancer (TNBC), we tested def-22B in MDA-MB231 TNBC cell-derived tumor models. Consistent with findings in PDAC models, def-22B significantly suppressed MDA-MB231 tumor growth in NCr nude mice compared to isotype control-treated animals (Figure S4L).

Figure 7.

Combinatorial drug regimens enhance the therapeutic effect of anti-SDC1 mAb

(A–D) Quantification of the ratio of (A) PD-1⁺ tumor-infiltrating NK cells, (B) PD-1⁺ tumor-infiltrating CD3⁺ T cells, (C) PD-L1⁺ CD45⁻ non-immune cells, and (D) PD-L1⁺ Panc02-hSDC1 tumor cells from subcutaneous Panc02-hSDC1 tumors inoculated in C57BL/6NJ mice treated with def-22B or vehicle. Tumors were collected 1 day after the 5^th treatment.

(E) Growth curve of subcutaneous Panc02-hSDC1 tumors in C57BL/6NJ mice, treated with def-22B, anti-PD-1, def-22B + anti-PD-1, or vehicle. n = 5 per group.

(F) Quantification of 4-1BB⁺ tumor-infiltrating NK cells in Panc02-hSDC1-bearing mice treated with def-22B or vehicle. Tumors were collected 1 day after the 5^th treatment.

(G) Growth curve of subcutaneous Panc02-hSDC1 tumors in C57BL/6NJ mice treated with def-22B, anti-4-1BB, def-22B + anti-4-1BB, or vehicle. n = 5 per group.

(H) Growth curve of subcutaneous PATC153 tumors in NCr nude mice treated with def-22B, gemcitabine, def-22B + gemcitabine, or vehicle. n = 5 per group.

(I) Growth curve of subcutaneous HY50760-hSDC1 tumors in C57BL/6NJ mice treated with def-22B, AMG510, def-22B + AMG510, or vehicle. n = 5 per group.

(J) Growth curve of subcutaneous AsPC-1 tumors in NCr nude mice treated with def-22B, MRTX1133, def-22B + MRTX1133, or vehicle. n = 5 per group.

(K) Schematic of the structure of chimeric def-22B-hIgG1 antibody.

(L) Growth curve of subcutaneous AsPC-1 tumors in NSG mice adoptively transferred with NK92-CD16-IL15 and treated with either hIgG1 isotype or def-22B-hIgG1. n = 5 per group.

(M) Quantification of tumor-infiltrating NK92-CD16-IL15 cells in the outer region of AsPC-1 tumors. NK92 cells were immunostained with anti-human CD45 antibody. Tumors were collected 1 day after the last antibody treatment. Data are represented as mean ± SD. Statistical significance was analyzed by ANOVA or unpaired Student’s t test. p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 are noted with ∗, ∗∗, ∗∗∗, and ∗∗∗∗, respectively. ns, non-significant.

Figure 5.

The therapeutic effect of anti-SDC1 mAb in vivo requires NK cells

(A) Growth curve of subcutaneous AsPC-1 tumors inoculated in NSG mice, treated with def-22B or vehicle. n = 5 per group.

(B) Growth curve of subcutaneous Panc02-GFP tumors in C57BL/6NJ mice, treated with def-22B or vehicle. n = 5 per group.

(C) Growth curve of Panc02-hSDC1 tumors in C57BL/6NJ mice, treated with def-22B, Fc-silenced 22B-LALAPG, or vehicle. n = 5 per group.

(D) Growth curve of subcutaneous Panc02-hSDC1 tumors in C57BL/6NJ mice, treated with def-22B or vehicle, with or without the depletion of macrophages by anti-CSF1R antibody. n = 5 per group.

(E) Luminescence-based ADCP reporter assay of U266 cells treated with def-22B, parental fucosylated clone 22B, or isotype control.

(F) Growth curve of subcutaneous Panc02-hSDC1 tumors in C57BL/6NJ mice, treated with def-22B or vehicle, with or without the depletion of NK cells by anti-NK1.1 antibody. n = 5 per group. Data are represented as mean ± SD. Statistical significance was analyzed by ANOVA or unpaired Student’s t test. p < 0.05 and p < 0.001 are noted with ∗ and ∗∗∗, respectively. ns, non-significant.

To further assess the translational potential of def-22B, we examined the toxicologic profile of def-22B in murine models and across normal human tissues. Compared with vehicle, def-22B did not affect body weight in either AsPC-1 (NCr nude) or Panc02-hSDC1 (C57BL/6NJ) models (Figures S5A and S5B). Histopathological analysis revealed no abnormalities in major organs, with the exception of enlarged, darkened splenic white pulp consistent with immune activation following def-22B treatment (Figures S5C and S5D). In vitro, def-22B induced minimal ADCC effects against human SDC1-negative normal human cells (HPNE and HUVECs; Figures S5E–S5H), suggesting limited risk to human SDC1-low or negative tissues. By contrast, def-22B elicited ADCC against human SDC1-positive human mammary epithelial cell MCF10A, indicating potential on-target, off-tumor (OTOT) activity (Figures S5F and S5I). Notably, clone 22B exhibited IHC staining patterns in normal human tissues comparable to nBT062 (Figures 2E and S2G). Clinical experience with nBT062-based therapies, for example ADCs, indicates an acceptable safety profile, characterized primarily by anemia and mucosal inflammation at the highest doses, and no severe adverse events.^21,22 Taken together, while OTOT effects remain possible, the aggregate nonclinical and reference clinical data from nBT062 support a manageable safety risk for def-22B, warranting continued translational evaluation with appropriate monitoring.

NK cells mediate the therapeutic efficacy of anti-SDC1 mAb

Given that def-22B induced robust antitumor efficacy in C57BL/6NJ and NCr nude mice (Figures 4C–4H) but not in NSG mice (Figure 5A), we hypothesized that antitumor immunity may play a role in the therapeutic effect of anti-SDC1 targeting. Importantly, we found that clone 22B harboring an LALAPG mutation,^23,24 which silences the function of the immunoglobulin Fc-domain (Figures S6B and S6C), completely abolished the in vivo therapeutic efficacy of clone 22B within an immunocompetent background (Figure 5C). This underscores the importance of immune system involvement, specifically macrophages or natural killer (NK) cells, in the antitumor effects of def-22B. To assess the involvement of macrophages in the response to def-22B in vivo, we treated Panc02-hSDC1-bearing C57BL/6NJ mice with an anti-CSF1R antibody 1 day before each def-22B injection to deplete macrophages (Figure S6D). We found that macrophage depletion did not diminish the antitumor efficacy of def-22B compared to def-22B treatment alone (Figure 5D), suggesting that the therapeutic effect of def-22B is macrophage independent. Interestingly, co-administration of def-22B and anti-CSF1R achieved superior disease control compared to def-22B treatment alone (Figure 5D), likely due to the alleviation of the local immunosuppressive microenvironment through the depletion of tumor-associated macrophages.²⁵ Moreover, antibody-dependent cellular phagocytosis (ADCP) reporter assay showed no significant enhancement of phagocytosis with def-22B and parental clone 22B compared with isotype treatment (Figure 5E). The role of NK cells in mediating the antitumor effect of def-22B was assessed by treating C57BL/6NJ mice with anti-NK1.1 antibody, PK136, prior to each def-22B injection (Figure S6E). Strikingly, NK cell depletion abolished the antitumor effect of def-22B (Figure 5F). This aligns with our findings that def-22B demonstrated dramatic ADCC effects in both luciferase-based ADCC reporter assay and NK92-CD16-mediated cell-killing assay (Figures 4A, S4G, and S4H). Together, our data strongly support that the antitumor effect of anti-SDC1 mAb is immune cell dependent, with NK cells playing a central role in its mechanism of action.

Anti-SDC1 mAb enhances intratumoral immune cell activation and modulates the tumor microenvironment

To further define how def-22B engages immune cells in vivo, we profiled intratumoral immune cell populations following def-22B treatment. Mice bearing subcutaneous Panc02-hSDC1 tumors were administered five doses of def-22B or vehicle from days 7 to 20, and tumors were analyzed 24 h after the final dose (Figure 6A). Although total immune infiltration was comparable between groups (Figures S7A and S7B), def-22B significantly increased both the frequency and absolute number of tumor-infiltrating NK cells (Figures 6B and 6C). In addition, def-22B treatment activated intratumoral NK cells, as evidenced by a higher proportion and expression of NK cell activation markers CD69 and NKp46 compared to vehicle-treated controls (Figures 6D–6F, S7C, and S7D). Concordantly, intratumoral cytokines were examined by a cytokine bead array, revealing a significantly elevated interferon-γ level in def-22B-treated tumors (Figure 6G). In vitro co-cultures of NK cells and Panc02-hSDC1 cells demonstrated that def-22B significantly stimulated interferon-γ secretion relative to vehicle, isotype, and parental clone 22B (Figure S7E). Interestingly, def-22B treatment was also associated with higher intratumoral CD3⁺ T cells and CD8⁺ T cell abundance (Figures 6H and 6I). While T cells lack either SDC1 or the Fc receptor required for direct interaction with the anti-SDC1 mAb and are not primary effectors of def-22B, this observation likely reflects a secondary response mediated by NK cell activation and associated pro-inflammatory cytokines such as interferon-γ. These findings were further corroborated in tumors derived from mouse PDAC HY55582 cells with human SDC1 expression (HY55582-hSDC1): six doses of def-22B from days 3 to 20 yielded significantly increased intratumoral NK cell frequency, absolute number, and activation (Figures 6J–6L, S7F, S7G, and S7H), along with expansion of intratumoral CD3⁺, CD4⁺, and CD8⁺ T cells and reduced Tregs (Figures 6K, 6M–6P, S7I, and S7J). Additionally, def-22B also led to an increase in intratumoral myeloid-derived suppressor cells (MDSCs) and a decrease in CD11c⁺ MHCII⁺ dendritic cells (Figure 6K), further suggesting a broader immune remodeling elicited by def-22B. Collectively, def-22B suppresses PDAC tumors by promoting NK cell activation and expansion and downstream T cell responses, accompanied by reshaping of the tumor microenvironment.

Figure 6.

Anti-SDC1 mAb activates intratumoral immune cell populations

(A) Schematic of the treatment schedule of C57BL/6NJ mice with subcutaneous Panc02-hSDC1 tumors for (B)–(I). Tumors were collected for flow cytometry 1 day after the 5^th treatment.

(B and C) Quantification of (B) the ratio and (C) absolute numbers of tumor-infiltrating NK cells in Panc02-hSDC1 tumors, treated with def-22B or vehicle.

(D) Quantification of the ratio of CD69⁺ tumor-infiltrating NK cells.

(E) Median fluorescence intensity of CD69 expression on tumor-infiltrating NK cells.

(F) Quantification of the ratio of NKp46⁺ tumor-infiltrating NK cells.

(G) Quantification of intratumoral interferon-γ levels in Panc02-hSDC1 tumors.

(H) Quantification of the ratio of tumor-infiltrating CD3⁺ T cells to CD45⁺ tumor-infiltrating immune cells.

(I) Quantification of the ratio of tumor-infiltrating CD8⁺ T cells to tumor-infiltrating CD3⁺ T cells.

(J) Schematic of the treatment schedule of subcutaneous HY55582-hSDC1 tumors inoculated in C57BL/6NJ mice for (K)–(P). Tumors were collected for flow cytometry or CyTOF analysis 1 day after the 6^th treatment.

(K) t-SNE plot of HY55582-hSDC1 tumor-infiltrating living CD45⁺ immune cells. Overlaid color-coded immune cell subsets were identified by FlowSOM.

(L–P) Quantification of the ratio of (L) tumor-infiltrating NK cells to CD45⁺ tumor-infiltrating immune cells, (M) tumor-infiltrating CD3⁺ T cells to CD45⁺ tumor-infiltrating immune cells, (N) tumor-infiltrating CD8⁺ T cells to tumor-infiltrating CD3⁺ T cells, (O) tumor-infiltrating non-Treg CD4⁺ T cells to tumor-infiltrating CD3⁺ T cells, and (P) tumor-infiltrating Treg cells to tumor-infiltrating CD4⁺ T cells in HY55582-hSDC1 tumors inoculated in C57BL/6NJ mice treated with def-22B or vehicle. Data are represented as mean ± SD. Statistical significance was analyzed by ANOVA or unpaired Student’s t test. p < 0.05, p < 0.01, and p < 0.001 are noted with ∗, ∗∗, and ∗∗∗, respectively.

Combination strategies enhance the therapeutic efficacy of anti-SDC1 mAb

Rational combination approaches have been demonstrated to induce more potent and durable responses than monotherapy approaches in several cancers.²⁶ Accordingly, we sought to identify rational combinations to enhance the antitumor efficacy of def-22B. Repeated activation of intratumoral immune cells often leads to immune exhaustion.²⁷ Indeed, flow cytometry and CyTOF analysis revealed increased expression of exhaustion markers PD-1 and PD-L1 on both immune and tumor cells following def-22B treatment in vivo and in NK92-tumor cell co-cultures (Figures 7A–7D, S8A, and S8B). Given that elevated PD-1/PD-L1 expression is associated with improved responses to immune checkpoint blockade (ICB),²⁸ we tested the combinatory effects of ICB with def-22B. In C57BL/6NJ mice bearing Panc02-hSDC1 tumors, def-22B plus anti-PD-1 significantly inhibited tumor growth versus either agent alone or vehicle, with minimal tumor progression observed at the study endpoint (Figure 7E). Furthermore, given the high 4-1BB expression on NK cells observed in vivo and in vitro (Figures 7F, S8C, and S8D) and the ability of anti-4-1BB to potentiate NK/T cell cytotoxicity,²⁹ combining def-22B with anti-4-1BB antibody produced greater tumor inhibition than monotherapy (Figure 7G). These data support def-22B as a backbone for ICB and co-stimulatory combinations.

Gemcitabine, a first-line PDAC chemotherapy agent, is known for exerting tumor-suppressive effects and can induce immunogenic cell death, which can enhance the efficacy of immune-based therapies.^30,31 Furthermore, previous studies have demonstrated that gemcitabine treatment sensitizes PDAC cells to NK cell-mediated cytotoxicity.³⁰ Based on these findings, we investigated the therapeutic potential of combining def-22B with gemcitabine in PATC153 patient-derived tumors harboring KRAS^WT, a model unresponsive to KRAS^∗ inhibitors. In this setting, while gemcitabine monotherapy unexpectedly accelerated progression versus vehicle, def-22B combined with gemcitabine produced a significantly greater antitumor activity, including marked regressions approaching complete response in several cases (Figure 7H), supporting a strong combination benefit in PDAC. Our previous work also showed that SDC1 upregulation is an adaptive mechanism of resistance to KRAS^∗ inhibition. On this basis, we tested the synergistic effects of def-22B with the KRAS^G12D inhibitor MRTX1133 and the KRAS^G12C inhibitor AMG510, respectively. In vitro, our anti-SDC1 mAb effectively suppressed cell proliferation and induced robust ADCC activity in PDAC cells resistant to KRAS^∗ inhibitor (Figures S8E–S8H). In vivo, def-22B and AMG510 displayed comparable antitumor effects in murine PDAC cells HY50760-hSDC1 (Kras^G12C), and the combination further enhanced tumor control (Figure 7I). Similarly, def-22B plus MRTX1133 yielded superior inhibition in the AsPC-1 model (KRAS^G12D) relative to either agent alone (Figure 7J). Collectively, these findings position def-22B as a versatile therapeutic that, when combined with chemotherapy, KRAS-targeted agents, or immune-based therapies, can achieve superior tumor control across diverse PDAC contexts.

Chimeric humanized anti-SDC1 mAb suppresses tumor growth in a humanized tumor model

To evaluate anti-SDC1 efficacy in a clinically relevant format, we engineered the clone 22B variable domains into a human IgG1 backbone, generating the chimeric 22B-hIgG1 antibody (def-22B-hIgG1) (Figure 7K). The chimeric antibody preserved high affinity and specificity for human SDC1, elicited robust ADCC effect, and exhibited a satisfactory in vivo half-life (Figures S8I–S8L). For in vivo studies, we employed NK92-CD16 effector cells due to their known cytotoxicity against human PDAC cells. To improve the survival and cytotoxicity of NK92-CD16 cells in NSG mice, the cells were retrovirally transfected to express human IL-15,³² enabling IL-2-independent growth, which was confirmed by MTT assay (Figure S8M). NSG mice bearing AsPC-1 tumors received adoptive transfer of IL-15-expressing NK92-CD16 cells together with def-22B-hIgG1, hIgG1 isotype, or vehicle (Figure S8N). Def-22B-hIgG1 significantly inhibited tumor growth in this humanized model compared to vehicle and isotype controls (Figure 7L). The restoration of efficacy of def-22B-hIgG1 in the NK-repopulated NSG mice, contrasting with minimal activity in naive NSG models (Figures 5A and S8O), supports a critical role of NK cells in mediating def-22B activity in vivo; concordantly, IHC analysis revealed increased NK cell infiltration in def-22B-hIgG1-treated tumors (Figures 7M and S8P). In summary, these data demonstrate potent antitumor efficacy of anti-SDC1 therapeutic targeting in PDAC, augmented by recruitment and activation of intratumoral NK cells.

Discussion

In this study, we reported a mAb targeting human SDC1 and demonstrated its therapeutic potential in various preclinical models, with a primary focus on PDAC. Our findings reveal that clone 22B exhibits exceptional binding and specificity to human SDC1, akin to the clinically evaluated antibody nBT062, but uniquely inhibits SDC1 function, resulting in robust suppression of macropinocytosis and significant reduction in cell proliferation under nutrient-deficient conditions. Furthermore, the optimized def-22B elicited potent antitumor immune responses, particularly through NK cell-mediated ADCC, leading to effective tumor growth inhibition in vivo. Importantly, def-22B demonstrated therapeutic efficacy in SDC1-positive PDAC tumors, irrespective of KRAS mutation status, and synergized with first-line chemotherapy, KRAS^∗-targeted agents, and immune-based therapies to achieve profound tumor regression. These findings establish the translational potential of def-22B, demonstrating a preclinical strategy in which an anti-SDC1 mAb directly targets SDC1 function while eliciting a robust antitumor immune response.

SDC1 is highly expressed across various human cancer types, including pancreatic cancer, lung cancer, breast cancer, colorectal cancer, cholangiocarcinoma, ovarian cancer, and multiple myeloma.^6,33 This widespread dysregulation highlights the potential for SDC1-targeted therapies to have broad applicability beyond pancreatic cancer. As a proof of concept, we demonstrated the efficacy of our SDC1-targeting antibody in TNBC, a malignancy characterized by high SDC1 expression and limited treatment options. Beyond its expression on cancer cells, SDC1 is also highly expressed in tumor-associated stromal fibroblasts, where it facilitates cancer cell proliferation and metastasis.^34,35,36 Given the critical role of fibroblasts in driving cancer progression, further investigation is warranted to determine whether targeting SDC1 in this stromal compartment contributes to the therapeutic efficacy of SDC1-targeted therapies. The role of SDC1 in cancer biology also extends to mediating drug resistance across various cancer types. Our recent study identified the YAP1-SDC1 axis as a key mechanism that sustains nutrient salvage and drives bypass of KRAS^∗ dependency in pancreatic and colorectal cancers,⁴ further supporting SDC1 as a promising therapeutic target across oncologic contexts.

Current anti-SDC1 agents, primarily derived from the nBT062 (B-B4) mAb, bind SDC1 to deliver cytotoxic agents or facilitate effector cell-mediated killing⁶ but do not directly inhibit SDC1 function. While the anti-SDC1 therapies have primarily been explored in the context of multiple myeloma, their applications in solid tumors have been largely overlooked. Our therapeutic antibody demonstrates dual functionality by exerting immune-dependent cytotoxicity via triggering NK-mediated killing and directly inhibiting SDC1-mediated macropinocytosis that is crucial for KRAS^∗-driven PDAC progression and maintenance. As a major nutrition salvage mechanism, macropinocytosis represents metabolic vulnerabilities that can be exploited to develop novel cancer therapies and prevention strategies.⁵ Existing pharmacological inhibitors, including the widely recognized EIPA, exhibited significant toxicity and poor specificity,³⁷ which markedly restricts their translational potential in oncology. In contrast, our anti-SDC1 mAb exhibits both remarkable specificity and potent inhibition of SDC1-mediated macropinocytosis, presenting a compelling alternative for targeting this critical metabolic pathway. In this study, we revealed that the clone 22B binding motif, DITLSQQ, is necessary for SDC1-mediated macropinocytosis and cell proliferation. Structural rigidity prediction indicates that this epitope is relatively rigid compared with other regions of the ectodomain (Figure S9). We postulate that clone 22B binding may disrupt protein-protein interactions required for macropinocytic signaling, thereby inhibiting the process. Definitive elucidation of the sequence- and structural-level determinants will require high-resolution structural and systematic mutational studies, which could inform next-generation strategies to therapeutically target SDC1-driven macropinocytosis.

Nonetheless, pancreatic cancer is notoriously considered an “immunologically cold” tumor due to its profoundly immunosuppressive microenvironment and poor infiltration of immune effector cells,³⁸ which severely impairs the efficacy of immunotherapies. Despite this challenging landscape, administration of def-22B demonstrated a capacity to modulate the tumor immune microenvironment, as evidenced by increased infiltration and activation of effector cells, including NK and T cells, accompanied by upregulated activation markers such as CD69, NKp46, and interferon signaling pathways. These findings suggest that def-22B has the potential to reshape the immune landscape and activate antitumor immune responses. However, the immunosuppressive tumor microenvironment of pancreatic cancer remains a significant barrier, characterized by abundant regulatory T cells, tumor-associated macrophages, MDSCs, as well as upregulated PD-1/PD-L1 expression. To address this, we conducted proof-of-concept studies combining def-22B with anti-PD-1 checkpoint blockade and a costimulatory agonist, anti-4-1BB, to further potentiate the antitumor immune response. These combinatorial approaches significantly enhanced tumor inhibition efficacies compared to def-22B monotherapy, albeit without achieving complete tumor regression. Thus, pairing def-22B with complementary immunotherapeutic agents is likely required to maximize antitumor activity and more fully realize the therapeutic potential of antibody-based strategies. Despite the development of targeted therapies and immunotherapies, chemotherapy remains as the mainstay for the treatment of most cancer types. Here, we demonstrated the synergy between def-22B and gemcitabine in suppressing tumor growth in PDAC preclinical models. SDC1 has been implicated in chemotherapy resistance in liver,³⁹ colorectal,⁴⁰ and prostate cancers,⁴¹ suggesting the potential of def-22B to overcome chemoresistance in multiple cancer types. Moreover, gemcitabine can alleviate the immunosuppressive tumor microenvironment by reducing Tregs and MDSCs while promoting NK and T cell infiltration and activation,^30,31 which may contribute to its combinatorial benefit with def-22B in vivo.

KRAS^∗ inhibitors are emerging as promising targeted therapies for PDAC. Our recent study revealed a pivotal role of SDC1 in mediating acquired resistance of KRAS^∗ inhibitors,⁴ highlighting the unique therapeutic value of combining SDC1-targeted therapies with KRAS^∗ inhibitors. It has been well established that the KRAS∗ oncogene plays an essential role in dictating the suppressive immune microenvironment of PDAC.⁴² Accordingly, KRAS^∗ inhibitor treatment elicits a pro-inflammatory microenvironment characterized by elevated interferon signaling, increased infiltration of effector cells and antigen-presenting cells, as well as reduction in M2-like macrophages and MDSCs.^43,44 This immunologically favorable landscape aligns with the mechanism of action of def-22B, further supporting the synergistic potential of def-22B in combination with KRAS^∗-targeted therapies.

While our anti-SDC1 mAb demonstrated compelling antitumor efficacy, alternative modalities such as ADCs that deliver cytotoxic payloads, pathway inhibitors, or radioisotopes via SDC1-targeted antibodies,^6,7 may also warrant evaluation. Moreover, SDC1-targeted immunotherapies, including bispecific antibodies, immunocytokines, or CAR-T or CAR-NK cell therapies, hold promise for improving antitumor immune responses by enhancing effector cell infiltration and activation within the tumor microenvironment.^38,45 Beyond oncology, SDC1 has been implicated in various inflammatory diseases and infectious conditions: it facilitates leukocyte recruitment and plays a significant role in resolving inflammation,⁴⁶ and it can act as a receptor or co-receptor for microbial pathogens during early infection.⁴⁷ Despite these insights, the therapeutic potential of SDC1-targeted agents in non-cancer contexts remains underexplored. The ability of SDC1-targeted therapies, such as def-22B, to modulate immune responses or inhibit pathogen interactions raises intriguing possibilities for their application in inflammatory diseases or infections where SDC1 is critically involved. Future studies could expand the therapeutic efficacy and safety of SDC1-targeted agents in these pathological settings.

Limitations of the study

A primary limitation is the absence of a comprehensive in vivo toxicity assessment for def-22B. The lack of cross-reactivity of clone 22B to murine SDC1 limits the evaluation of target-dependent biodistribution and toxicity in murine models, complicating preclinical risk evaluation. However, this species specificity constraint is partly mitigated by clone 22B’s cross-reactivity with cynomolgus SDC1, which enables toxicity and biodistribution studies in non-human primates. Additionally, although IHC shows staining patterns for clone 22B in human tissues that are comparable to those of nBT062, and prior clinical experience with nBT062/B-B4-based ADCs suggests an acceptable safety profile,^6,21,22 these data are contextual rather than determinative. Formal good laboratory practice (GLP) toxicology, safety pharmacology, and biodistribution studies in an appropriate species may be required to define the safety margins of def-22B and support clinical development.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Wantong Yao (wyao2@mdanderson.org).

Materials availability

All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •All data reported in this paper are included in the article and supplemental information. RNA-seq data are publicly available at the SRA portal of NCBI under the accession number GEO: GSE311267.
• •This paper does not report any original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We gratefully appreciate Dr. Anirban Maitra of the MD Anderson Cancer Center for his invaluable support in facilitating this research. We also thank Dr. Weiyi Peng of the University of Houston and Dr. Dihua Yu of the MD Anderson Cancer Center for their constructive suggestions. The MDA-MB-231 cell line was kindly provided by Dr. Dihua Yu, and the NK92-CD16-noGFP cell line was generously shared by Dr. Larry Anderson of Emory University. The phCMV-BaEV-TR plasmid was kindly provided by Dr. Richard Eric Davis of the MD Anderson Cancer Center. We further appreciate Dr. Yimo Sun and Dr. Xinru Jiang for their valuable guidance on NK92 cell transduction. The research was supported by CPRIT RP230050 (Cancer Prevention and Research Institute of Texas, CPRIT); R37CA272744 (NIH/NCI); RSG-22-017-01-CCB (American Cancer Society, ACS); R21CA2888804 (NIH/NCI); AACR-Pancreatic Cancer Action Network Pathway to Leadership Grant (16-70-25-YAO); Pancreatic Cancer Action Network-Translational Research Grant (19-65-YAO); W81XWH-20-1-0598 (Department of Defense, DOD); generous philanthropic donations to the MD Anderson Cancer Center Dallas Living Legend Fund and Moon Shots Program; Shelby-Lavine Pancreatic Scholar and University Cancer Foundation via the Institutional Research Grant (IRG) program at The University of Texas MD Anderson Cancer Center to W.Y.; the Uehara Memorial Foundation postdoctoral fellowship, the Shinya Foundation for International Exchange of Osaka University Graduate School of Medicine Grant, and the Cell Science Research Foundation Grant to M.T.; 1R01CA214793 (NCI) and HT94252311082 (DOD) to H.Y.; and 5P01CA11796 (NCI) to H.W. and H.Y. We also acknowledge the support of the Flow Cytometry and Cellular Imaging Core Facility, the Research Histology Core Laboratory, the Advanced Technology Genomics Core, and the Department of Veterinary Medicine at the MD Anderson Cancer Center (supported by CCSG NIH NCI grant P30CA016672 and CPRIT RP121010), as well as the Interdisciplinary Translational Education and Research Training (ITERT) program.

Author contributions

Z.Y. designed the studies, conducted the experiments, and analyzed the data; M.S.T. and S.C. performed portions of the flow cytometry and colony formation experiments; L.T.V. and L.B. generated antibody hybridomas; J.L. constructed the LALAPG-mutated antibody; J.Y. analyzed mRNA-seq data; P.T. analyzed immunofluorescence data; Y.W., M.T., Y. Zeng, X.W., J.P., and Y. Zheng assisted with experiments; K.M.W., H.W., L.K., T.H., and I.I.W. provided TMA slides; G.F.D. provided valuable insights and intellectual input; Z.Y., S.G., H.Y., and W.Y. wrote the manuscript; and H.Y. and W.Y. supervised the studies.

Declaration of interests

Z.Y., L.B., H.Y., and W.Y. are co-inventors of a patent (United States Provisional Patent Application serial no. 63/503,785). T.H. receives advisory fees from Cullgen Inc., Psivant Therapeutics, Isomorphic Labs, Quotient Therapeutics, and Proxygen.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-SDC1 monoclonal antibody (Clone 22B)	This paper	N/A
Anti-SDC1 monoclonal antibody (Clone nBT062)	This paper	N/A
Anti-mouse NK1.1 antibody (Clone PK136)	This paper	N/A
Anti-mouse CSF1R antibody (Clone AFS98)	This paper	N/A
Anti-mouse PD-1 antibody (Clone RMP1-14)	This paper	N/A
Anti-mouse 4-1BB antibody (Clone 3H3)	This paper	N/A
Anti-human HER2 antibody (Trastuzumab, Clone 4D5-8)	Leinco Technologies	Cat# LT1500 RRID: AB_2893910
Mouse monoclonal anti-human CD45 (Clone HI30)	BioLegend	Cat# 304002; RRID:AB_314390
PE/Dazzle594 anti-mouse CD69 (Clone H1.2F3)	BioLegend	Cat# 104536; RRID: AB_2565583
BV650 anti-mouse NKp46 (Clone 29A1.4)	BioLegend	Cat# 137635; RRID: AB_2734200
PerCP-eFluor710 anti-mouse CD137 (Clone 17B5)	Invitrogen	Cat# 46-1371-82 AB_2531219
APC anti-mouse CD45 (Clone 30-F11)	BioLegend	Cat# 103112; RRID: AB_312977
PE/Cyanine5 anti-mouse CD3e (Clone 145-2C11)	BioLegend	Cat# 100310; RRID: AB_312675
Alexa Fluor 700 anti-mouse CD8a (Clone 53-6.7)	BioLegend	Cat# 100730; RRID:AB_493703
PE anti-mouse NK1.1 (Clone S17016D)	BioLegend	Cat# 156504; RRID: AB_2783136
PE/Cyanine7 anti-mouse PD-1 (Clone 29F.1A12)	BioLegend	Cat# 135215; RRID: AB_10696422
BV785 anti-mouse PD-L1 (Clone 10F.9G2)	BioLegend	Cat# 124331; AB_2629659
APC anti-human PD-L1 (Clone 29F.2A3)	BioLegend	Cat# 329708; RRID: AB_940360
APC anti-human PD-1 (Clone EH12.2H7)	BioLegend	Cat# 329908; RRID: AB_94047
APC anti-human CD137 (Clone 4B4-1)	BioLegend	Cat# 309810; RRID: AB_830672
APC anti-human CD138 (Clone MI15)	BioLegend	Cat# 356506; RRID: AB_2561880
Purified anti-human CD138 (Clone MI15)	BioLegend	Cat# 356502; RRID: AB_2561790
APC mouse IgG1, κ isotype control (Clone MOPC-21)	BioLegend	Cat# 400120 RRID: AB_2034027
APC rat IgG2a, κ isotype control (Clone RTK2758)	BioLegend	Cat# 400512; RRID: AB_2814702
PE goat anti-rat IgG (Clone Poly4054)	BioLegend	Cat# 405406; RRID: AB_315017
PE mouse IgG2a, κ isotype control (Clone MOPC-173)	BioLegend	Cat# 400212; RRID: AB_326460
PE goat anti-mouse IgG (Clone Poly4053)	BioLegend	Cat# 405307; RRID: AB_315010
PE anti-mouse IgG2a (Clone RMG2a-62)	BioLegend	Cat# 407108; RRID: AB_10549456
TruStain FcX (anti-mouse CD16/32) (Clone 93)	BioLegend	Cat# 156604; RRID:AB_2783138
HRP-conjugated donkey anti-mouse IgG (H + L) secondary antibody	Jackson ImmunoResearch Labs	Cat #715-035-150; RRID: AB_2340770
Mouse monoclonal APC mouse IgG1 isotype (Clone MOPC-21)	BioLegend	Cat# 400119, RRID: AB_2888687
89Y anti-mouse CD45 (Clone 30-F11)	DVS-Fluidigm	Cat# 3089005B; RRID: AB_2651152
115In anti-mouse CD4 (Clone RM4-5)	BioLegend	Cat# 100506; RRID: AB_312709
139La anti-mouse CD11b (Clone M1/70)	BioLegend	Cat# 101249; RRID: AB_2562797
141Pr anti-mouse Ly-6G/Ly-6C (Clone RB6-8C5)	DVS-Fluidigm	Cat# 3141005B RRID:AB_3677773
142ND anti-mouse CD11c (Clone N418)	BioLegend	Cat# 117302; RRID: AB_313771
145ND anti-human CD138 (Clone DL101)	DVS-Fluidigm	Cat# 3145003B RRID:AB_3677805
146ND anti-mouse CD8a (Clone 53-6.7)	BioLegend	Cat# 100702; RRID: AB_312741
149Sm anti-mouse CD19 (Clone 6D5)	BioLegend	Cat# 115502; RRID: AB_313637
150ND anti-mouse CD279 (Clone RMP1-30)	Tonbo Biosciences	Cat# 70-9981-U500; RRID:AB_3714652
152Sm anti-mouse CD3e (Clone 145-2C11)	BioLegend	Cat# 100302; RRID: AB_312667
153 Eu anti-Arginase-1 (Polyclonal)	R&D Systems	Cat# AF5868; RRID: AB_1964500
154Sm anti-mouse CD274 (Clone 10F.9G2)	BioLegend	Cat# 124303; RRID: AB_961230
158Gd anti-mouse Foxp3 (Clone FJK-16s)	DVS-Fluidigm	Cat# 3158003A; RRID: AB_2814740
161Dy anti-mouse iNOS (Clone CXNFT)	DVS-Fluidigm	Cat# 3161011B; RRID: AB_2922920
167Er anti-mouse CD25 (Clone PC61.5)	Tonbo Biosciences	Cat# 70-0251-U100 RRID:AB_2621489
169Tm anti-GFP (Clone 5F12.4)	DVS-Fluidigm	Cat# 3169009B; RRID: AB_2814899
170Er anti-mouse NK1.1 (Clone PK136)	DVS-Fluidigm	Cat# 3170002B; RRID: AB_2885023
171Yb anti-mouse CD80 (Clone 16-10A10)	DVS-Fluidigm	Cat# 3171008B; RRID: AB_2885024
172Yb anti-mouse CD86 (Clone GL1)	DVS-Fluidigm	Cat# 3172016B; RRID: AB_2922923
174Yb anti-mouse F4/80 (Clone BM8.1)	Tonbo Biosciences	Cat# 70-4801-U100 RRID: AB_2621519
Chemicals, peptides, and recombinant proteins
Recombinant human SDC1	R&D Systems	Cat# 2780-SD-CF
Recombinant mouse SDC1	R&D Systems	Cat# 3190-SD
Recombinant human SDC1, biotinylated	Sino Biological	Cat# 11429-H49H-B
Recombinant cynomolgus SDC1	Sino Biological	Cat# 90938-C02H
Recombinant mouse CD16.2	R&D Systems	Cat# 1974-CD
Recombinant human CD16a	R&D Systems	Cat# 4325-FC
Cell-ID™ Cisplatin-195Pt	Fluidigm	Cat# 201195
Cell-ID™ Intercalator	Fluidigm	Cat# 201192A
Recombinant human IL-2	GenScript	Cat# Z00368
Recombinant human IL-15	GenScript	Cat# Z03308
Endothelial Cell Growth Supplement	R&D Systems	Cat# 390599
Recombinant human EGF	GenScript	Cat# Z00333
Hydrocortisone	Sigma Aldrich	Cat# H0888
Cholera Toxin	Sigma Aldrich	Cat# C8052
Insulin, human recombinant, zinc solution	Gibco	Cat# 12585014
TMR-Dextran (70 kDa)	Fina Biosolutions	N/A
pHrodo-Green Dextran (10 kDa)	Thermo Fisher	Cat# P35368
EIPA	Selleckchem	Cat# S9849
DRP-104	MedChemExpress	Cat# HY-132832
AMG510	MedChemExpress	Cat# HY-114277
MRTX1133	MedChemExpress	Cat# HY-134813
Gemcitabine hydrochloride	MedChemExpress	Cat# HY-B0003
Critical commercial assays
IFNγ matched ELISA antibody pair set	Sino Biological	Cat# SEKA11725
Mouse Tumor Dissociation Kit	Miltenyi Biotec	Cat# 130-096-730
ADCC Reporter Bioassay, F Variant	Promega	Cat# G9790
FcγRIIa-H ADCP Bioassay	Promega	Cat# G9991
Incucyte Mouse IgG2a Fabfluor-pH Antibody Labeling Dye	Sartorius	Cat# 4750
Incucyte Human Fabfluor-pH Antibody Labeling Dye	Sartorius	Cat# 4722
BD™ Cytometric Bead Array Mouse Inflammation Kit	BD Biosciences	Cat# AB_2868960
BD Pharmingen™ FITC BrdU Flow Kit	BD Biosciences	Cat# 559619
Experimental models: Cell lines
HY55582	This paper	N/A
HY50760	This paper	N/A
PATC53	MDACC	RRID: CVCL_VR69
PATC153	MDACC	N/A
MDA-MB231	MDACC	RRID: CVCL_0062
Panc02	ATCC	RRID: CVCL_D627
hTERT-HPNE	ATCC	RRID: CVCL_C466
HUVEC	BCM	RRID: CVCL_2959
MCF10A	ATCC	RRID: CVCL_0598
U266	ATCC	RRID: CVCL_0566
AsPC-1	ATCC	RRID: CVCL_0152
MiaPaca2	ATCC	RRID: CVCL_0428
A549	ATCC	RRID: CVCL_0023
HT29	ATCC	RRID: CVCL_A8EZ
NK92-CD16-noGFP	ATCC	RRID: CVCL_V429
293FT	ATCC	RRID: CVCL_6911
CHO-S	ATCC	RRID: CVCL_7183
Experimental models: Organisms/strains
Mouse: C57BL/6NJ	Jackson Laboratory	Strain# 005304
Mouse: NCr nude	Taconic Biosciences	Model# NCRNU-F
Mouse: Nod Scid Gamma (NSG)	Jackson Laboratory	Strain# 005557
Recombinant DNA
pXC17.4	Genprice	Cat# PVT13313
pXC18.4	Genprice	Cat# PVT13184
pcDNA3.4	Genscript	N/A
psPAX2	Addgene	Cat# 12260; RRID: Addgene_12260
pMD2.G	Addgene	Cat# 12259; RRID: Addgene_12259
pHIT60	MDACC	N/A
phCMV-BaEV-TR	Dr. Richard Eric Davis	N/A
pLentiCRISPRv2-puro_sgRNA control (sgCTR) (Target sequence: GTACAGCTAAGTTAAACTCG)	This paper	N/A
pLentiCRISPRv2-puro_sg human SDC1-1 (Target sequence: GTCATTGCCGGAGGCCTCGT)	This paper	N/A
pLentiCRISPRv2-puro_sg human SDC1-2 (Target sequence: TCATTGCCGGAGGCCTCGTG)	This paper	N/A
pLentiCRISPRv2-puro_sg hamster FUT8 (Target sequence: AGCAGCCTTCCATCCCATTG)	This paper	N/A
pLKO.1-puro_sh-scramble (Scr)	Cellecta	N/A
pLKO.1-puro_sh human SDC1-#46	Sigma Aldrich	TRCN0000414046
pLKO.1-puro_sh human SDC1-#78	Sigma Aldrich	TRCN0000072578
pHAGE-GFP_Human SDC1 ORF (NM_002997.4)	GenScript	N/A
pGenLenti-Bla_Human SDC1 ORF (NM_002997.4)	GenScript	N/A
pGenLenti-Bla_Human SDC1 ORFΔDITLSQQ (NM_002997.4)	GenScript	N/A
pBABE-GFP_Human IL15 ORF (NM_000585.5)	GenScript	N/A
Software and algorithms
BD FACSDiva software	BD Biosciences	RRID:SCR_001456
Helios system	Fluidigm	RRID:SCR_019916
FlowJo v.10	BD Biosciences	RRID:SCR_008520
Cytobank	Cytobank	RRID:SCR_014043
IncuCyte S3 Live Cell Analysis System	Sartorius	RRID:SCR_023147
GraphPad Prism 10	GraphPad	RRID:SCR_002798
DESeq2	DESeq2	RRID:SCR_015687
clusterProfiler package	clusterProfiler	RRID:SCR_016884
Deposited data
Raw and analyzed mRNA-seq data	This paper	GEO: GSE311267

Experimental models and study participant details

Mice

C57BL/6NJ and NSG mice (female, 8-week-old) were obtained from the Jackson Laboratory. NCr nude mice were obtained from Taconic Biosciences. All mice were maintained in pathogen-free condition at MD Anderson Cancer Center. All manipulations were approved under Institutional Animal Care and Use Committee (IACUC) protocol 00001549.

Cell lines

Established cell lines were obtained from the American Type Culture Collection (ATCC), Leibniz Institute DSMZ, or Expasy and were cultured according to the manufacturer’s protocols. HY50760 and HY55582 were derived from KPC mice with a C57BL/6 background. Panc02, HY50760, HY55582, and U266 were maintained in RPMI-1640 medium supplemented with 10% heat-inactivated (HI) fetal bovine serum (FBS) and 100U/mL penicillin-streptomycin. Human cell lines, including AsPC-1, PATC53, PATC102, PATC153, MiaPaca2, MDA-MB231, A549, and HT29, were cultured in DMEM (high glucose) medium supplemented with 10% HI FBS and 100U/mL penicillin–streptomycin. AMG510-resistant MiaPaca2 cells were maintained with 500 nM AMG510. MRTX1133-resistant PATC53 cells were maintained with 500 nM MRTX1133. NK92 cells were cultured in RPMI-1640 medium supplemented with 25% HI FBS, 100U/mL penicillin–streptomycin, 100 μM 2-mercaptoethanol, and 200 U/mL recombinant human IL-2 (GenScript). HUVEC cells were cultured in F12K medium supplemented with endothelial cell growth supplement (R&D, 390599). hTERT-HPNE cells were cultured with 75% DMEM (high glucose), 25% Medium M3 Base (Incell, M300F) supplemented with 5% HI FBS, 10 ng/mL human EGF (GenScript, Z00333). MCF10A cells were cultured with DMEM/F12 medium supplemented with 5% horse serum, 20 ng/mL human EGF, 0.5 μg/mL Hydrocortisone (Sigma, H0888), 100ng/mL Cholera Toxin (Sigma, C0852), and 10 μg/mL Insulin (Gibco, 12585014). 293FT cells were maintained in DMEM (high glucose) medium supplemented with 10% HI FBS. CHO-S cells were cultured in FreeStyle CHO medium (Gibco). Hybridoma cells were cultured in RPMI-1640 medium supplemented with 10% HI FBS and 100U/mL penicillin-streptomycin, or CD Hybridoma Medium (Gibco) with 100U/mL penicillin-streptomycin. All human cell lines were fingerprinted and verified at the Characterized Cell Line Core Facility at MD Anderson Cancer Center. All cell lines were routinely tested for mycoplasma contamination.

Method details

Antibody production and purification

Anti-SDC1 monoclonal antibodies were generated by immunizing mouse SDC1-knockout mice with human SDC1 recombinant protein (R&D Systems, 2780-SD-CF). B cell-derived hybridoma cells were subcloned and screened based on the reactivity to human SDC1 protein by ELISA. Leading candidates were selected for the subsequent evaluation. Antibodies produced by hybridoma were purified by a protein A agarose column (GenScript) and stored in pH 7.4 DPBS. To produce recombinant anti-SDC1 mAbs, pXC17.4/18.4 or pcDNA3.4 plasmids encoding antibody light chain and heavy chain were transiently transfected in CHO-S cells or FUT8-KO CHO-S cells. Culture medium containing antibodies was collected 10–12 days after transfection. The recombinant antibody was purified by a protein A agarose column (GenScript) and stored in pH 7.4 DPBS. To produce PK136 and AFS98 depletion antibodies, PK136 or AFS98 hybridoma cells were cultured in CD Hybridoma Medium (Gibco). Depletion antibodies were purified by protein G agarose column (GenScript) from the culture supernatant and stored in pH 7.4 DPBS. The production of RMP1-14 anti-mouse PD-1 antibody and 3H3 anti-mouse 4-1BB antibody followed the same protocol as the production of recombinant anti-SDC1 mAbs using pcDNA3.4 plasmids encoding antibody light chain and heavy chain and purified by protein G agarose column. For quality control, the binding of purified antibodies to respective antigens was measured by ELISA or flow cytometry. The purity of produced antibody was examined by NanoDrop and PAGE electrophoresis. All batches of antibodies used in this research passed quality control.

Construction of chimeric 22B-hIgG1

The VH and VL sequences of anti-SDC1 mAb clone 22B were subcloned into the human IgG1 (kappa) backbone. The chimeric 22B-hIgG1 was produced by CHO-S or FUT8-KO CHO-S cells following the same protocol as described above.

Detection of fucosylation by FITC-LCA

The fucosylation of CHO-S cells was examined by staining parental or FUT8-KO CHO-S cells with 20 μg/mL FITC-LCA (Vector Laboratories, FL-1041-5) in DPBS and visualized by fluorescence microscope. Fucosylation of purified antibodies was detected by western blot. Specifically, 50 ng of each antibody was loaded on 4–12% Bis-Tris protein gels (Invitrogen) for electrophoresis under reducing conditions and transferred to nitrocellulose membrane (Bio-Rad, 1704158). The membrane was blocked with a carbo-free blocking solution (Vector Laboratories, SP-5040) and stained with 10 μg/mL FITC-LCA. Images were captured using a Bio-Rad ChemiDoc imaging system.

In vitro binding assays of anti-SDC1 mAbs by ELISA

The binding of anti-SDC1 mAbs to recombinant SDC1 and Fc receptors was evaluated by ELISA. Specifically, 0.2 μg/mL recombinant human SDC1 (R&D Systems, 2780-SD), 0.2 μg/mL mouse SDC1 (R&D Systems, 3190-SD), 0.2 μg/mL cynomolgus SDC1 (Sino Biological, 90938-C02H), 2 μg/mL mouse CD16.2 (R&D Systems, 1974-CD), or 2 μg/mL human CD16a (R&D Systems, 4325-FC) was coated on high binding 96-well plates (Corning, 9018) at 4°C for overnight. Plates were subsequently blocked by 5% non-fat milk in PBS. Serially diluted antibodies were added into the plate after blocking and incubated at 37°C for 2 h, then incubated with 1:5,000 diluted HRP-conjugated donkey anti-mouse IgG (H + L) secondary antibody (Jackson ImmunoResearch, 715-035-150). The reactions were developed with TMB and stopped by 1 N H₂SO₄. Optical density at 450 nm (OD₄₅₀) and 630 nm (OD₆₃₀) were read by a FLUOstar Omega microplate reader (BMG LABTECH).

Flow cytometry analysis

The binding of anti-SDC1 mAbs to cells was evaluated by flow cytometry. Cells were dissociated by non-enzymatic cell dissociation solution (Thermo Fisher Scientific) and resuspended in flow staining buffer (DPBS supplemented with 2% FBS, 5 mM EDTA, and 0.01% NaN₃). Resuspended cells were then incubated with serially diluted antibodies for 30 min on ice, washed with cold flow staining buffer twice, and stained with PE-conjugated anti-mouse IgG antibody (BioLegend). To evaluate the binding with the presence of soluble SDC1, 5, 50, or 500 ng/mL recombinant human SDC1 (R&D, 2780-SD-CF) was added together with serial diluted anti-SDC1 antibodies during staining. Fluorescence-based binding intensities were examined with a BD FACSCanto II flow cytometer.

For epitope binning of anti-SDC1 mAbs by flow cytometry, PATC53 cells were pre-stained by 100 μg/mL “blocking” antibody (hIgG1 isotype) for 30 min 1 μg/mL staining antibody (mIgG isotype) was then added into each sample without washing. Cells were subsequently stained with PE-conjugated anti-mouse IgG antibody (BioLegend). Samples were analyzed with the BD FACSCanto II flow cytometer.

To quantify the expression levels of surface proteins, cells were dissociated by non-enzymatic cell dissociation solution (Thermo Fisher Scientific) and resuspended in flow staining buffer. Cells were then stained with fluorescence-labeled antibodies for 30 min on ice, followed by DAPI staining, or fixed by eBioscience IC Fixation Buffer (Thermo Fisher Scientific). Protein expressions were detected with the BD FACSCanto II flow cytometer.

For intratumoral immune cell profiling by flow cytometry, tumors were collected 24 h after the final treatment and dissociated using mouse Tumor Dissociation Kit (Miltenyi Biotec, 130-096-730) and gentleMACS Octo dissociator (Miltenyi Biotec, 130-096-427). Dissociated cells were then resuspended in flow staining buffer, stained with Zombie Violet (BioLegend), blocked with anti-CD16/CD32 (BioLegend, 156604) antibody, followed by incubation with mixed fluorescence-labeled antibodies for 30 min on ice, and fixed by eBioscience IC Fixation Buffer (Thermo Fisher Scientific). Data were acquired with BD LSRFortessa and analyzed by FlowJo software.

CyTOF

The protocol of CyTOF staining was described previously.²⁰ In brief, 24 h after the final treatment, tumors were harvested and dissociated using the mouse Tumor Dissociation Kit (Miltenyi Biotec, 130-096-730) and the gentleMACS Octo dissociator (Miltenyi Biotec, 130-096-427). Dissociated cells were then resuspended in flow staining buffer and blocked with anti-CD16/CD32 (BioLegend, 156604) antibody followed by incubation with mixed metal-labeled antibodies for 30 min at room temperature. Cisplatin (195Pt, Fluidigm) and Cell-ID Intercalator (Fluidigm) were used to identify live cells. The samples were analyzed using a Helios system (Fluidigm) from the MD Anderson Flow Cytometry and Cellular Imaging Core Facility. Data were processed by FlowJo and Cytobank.

Quantification of IFNγ

IFNγ level in the in vitro NK92 co-cultured system was quantified by IFNγ matched ELISA antibody pair set (Sino Biological, SEKA11725) following the manual’s instructions. For the detection of intratumoral IFNγ after antibody treatment, tumors were harvested 24 h after the final injection and immediately frozen by liquid nitrogen. Tumors were then homogenized in the presence of protease inhibitor and centrifuged to collect tissue supernatant. The concentration of IFNγ was quantified with a BD Cytometric Bead Array Mouse Inflammation Kit (AB_2868960) following the manual’s instructions.

Immunohistochemistry (IHC) staining

The IHC staining protocol was described previously.^4,20 In brief, tumors were fixed in 10% formalin and transferred to 70% ethanol on the next day before paraffin embedding. Tissues were then sectioned at 5 μm. Slides were deparaffinized and rehydrated through a series of xylene and ethanol washes. Antigen retrieval was carried out by heating the slides for 12 min in DAKO Target Retrieval Solution using a microwave. Tissue sections were incubated with 3% H₂O₂ for 30 min at room temperature to block endogenous peroxidase activity. After incubating in 10% normal goat or horse serum for 1 h, sections were incubated with primary antibody at 4°C overnight. For the detection of human SDC1 expression, the primary antibody was B-A38 (Abcam, ab34164). For the comparison of antibody specificity, the primary antibodies were 22B-mIgG2a and nBT062-mIgG2a. For the staining of intratumoral NK92-IL15 in AsPC-1-NSG model, the primary antibody was HI30 (BioLegend, 304002). After washing with PBS, sections were then incubated for 1 h with ImmPRESS-HRP polymer reagent (Vector Laboratories). Samples were developed with DAB Quanto chromogen substrate (Epredia) and counterstained with hematoxylin.

Epitope mapping

Conformational epitope mapping was performed by PEPperPRINT GmbH. In brief, 7, 10, and 13 amino acid cyclic-constrained peptides designed from human SDC1 protein sequence with overlaps of 6, 9, and 12 amino acids were synthesized. The binding of clone 22B to the cyclic-constrained peptides was measured by fluorescence ELISA.

Binding kinetics measurement by OCTET

Binding kinetics of anti-SDC1 mAbs was measured using the OCTET RED96 platform. In brief, biotinylated recombinant human SDC1 (Sino Biological 11429-H49H-B) was immobilized on OCTET SA biosensors (Sartorius) at a fixed concentration of 10 μg/mL and then interacted with serial dilutions of the anti-SDC1 mAbs ranging from 500 nM to 100 nM. The binding was calculated based on the kinetic constants measured by OCTET.

Detection of macropinocytotic and endocytic activity

To measure macropinocytotic activity after antibody treatment, PDAC cells were treated with 10 μg/mL antibody under normal culture medium for 6 h and followed by 18 h of treatment under serum-free conditions. After treatment, cells were incubated with 1 mg/mL 70 kDa TMR-Dextran (Fina Biosolutions) for 35–60 min to allow dextran uptake. Cells were washed three times with DPBS, then fixed by polyformaldehyde and stained with DAPI. TMR-Dextran uptake was measured using a confocal fluorescence microscope.

To measure endocytic activity after antibody treatment, PDAC cells were treated with 10 μg/mL antibody under normal culture medium for 24 h. After treatment, cells were incubated with 25 μg/mL 10kDa pHrodo-Green-Dextran (Thermo Fisher, P35368) for 45 min or 24 h to allow dextran uptake. Cells were washed three times with DPBS, then proceed for flow cytometry or confocal fluorescence microscope analysis.

Antibody internalization assay

Antibody internalization was monitored by Incucyte. In brief, anti-SDC1 mAbs or mIgG2a isotype were labeled with Incucyte Mouse IgG2a Fabfluor-pH Antibody Labeling Dye (Sartorius, 4750) following manufacturer’s instruction. PDAC cells were then incubated with 4 μg/mL labeled antibodies under normal culture medium. For the internalization of trastuzumab with the presence of anti-SDC1 antibodies, trastuzumab (Leinco Technologies, LT1500) was labeled with Incucyte Human Fabfluor-pH Antibody Labeling Dye (Sartorius, 4722) following manufacturer’s instruction. PDAC cells were then incubated with 4 μg/mL labeled trastuzumab with the treatment of 10 μg/mL unlabeled anti-SDC1 mAbs or mIgG2a isotype. The internalization of Fabfluor-pH labeled antibodies was monitored by Incucyte.

Cell proliferation assay

To monitor cell proliferation after antibody treatment, PDAC cells were resuspended in the assay medium (DMEM supplemented with 10% dialyzed FBS, 0.2mM glutamine, and 25mM HEPES) and seeded in a 48-well plate with a density of 6,000 cells per 300 μL per well. 24 h later, culture medium was replenished with fresh assay medium supplemented with 15 μg/mL antibodies or 25 μM EIPA (Selleckchem, S9849). For cell proliferation assay with the presence of DRP-104, MiaPaca2 cells were resuspended in the assay medium supplemented with 2.0 mM glutamine and seeded in a 48-well plate with a density of 6,000 cells per 300 μL per well and treated with 0.5 μM DRP-104 (Medchemexpress, HY-132832). 24 h later, the culture medium was replenished with fresh assay medium containing 2.0 mM glutamine, 0.5 μM DRP-104, with or without 15 μg/mL clone 22B. For cell proliferation assay of SDC1 knockout or knockdown cells, PDAC cells were resuspended in regular culture medium and seeded in a 96-well plate with a density of 2,000 cells per 100 μL per well. Cells were then cultured for 7 days without changing the medium unless otherwise indicated. The growth of treated cells was monitored by Incucyte.

Colony formation assay

500–1,000 cells were re-suspended in regular culture medium and seeded into 6-well plates or 12-well plates in replicates and cultured for 10–14 days to allow colony formation. The colonies were stained with 0.2% crystal violet. Images were captured with an Epson Perfection V39 II scanner. Crystal violet-stained colonies were eluted with 33% acetic acid for quantification. The absorbance at 595 nm was read by the FLUOstar Omega microplate reader (BMG LABTECH).

MTT cell proliferation assay

IL-2 independent growth of IL-15 transduced NK92-CD16 cells was evaluated by MTT assay. In brief, parental or IL-15 transduced NK92-CD16 cells were seeded in a 96-well plate at a density of 2,000 cells per well with or without 200 U/mL human IL-2 (GenScript) and cultured for 96 h. MTT reagent (Abcam) was added into each well at the endpoint and incubated for 2 h. Culture medium was removed by centrifugation. Precipitated crystal was dissolved by DMSO. The absorbance of OD₅₉₂ was read by the FLUOstar Omega microplate reader (BMG LABTECH).

Cell cycle analysis by BrdU incorporation

PDAC cells were cultured in DMEM supplemented with 10% dialyzed FBS, 0.2 mM glutamine, and 25 mM HEPES, and treated with clone 22B or isotype antibody for 2 days. Cells then were treated with BrdU for 90 min and stained with a FITC BrdU Flow Kit (BD Biosciences, 559619) following the manufacturer’s protocol.

mRNA sequencing

Total RNA was isolated using a RNeasy Mini Kit (Qiagen). Library preparation, sample clustering, and sequencing were performed by Healgen Scientific. Next-generation sequencing was carried out on an Illumina NovaSeq platform. Differential expression analysis of two groups in RNA sequencing was performed using the DESeq2 R package. The resulting p value was adjusted using the Benjamini–Hochberg approach for controlling the false discovery rate. Genes with an adjusted p < 0.05 found by DESeq2 were assigned as differentially expressed. Gene function enrichment analysis was performed with the clusterProfiler package in R.

Virus production and transduction

psPAX2 (Addgene plasmid 12260) and pMD2.G (Addgene plasmid 12259) were used as packaging plasmids for lentivirus production. For retrovirus production, pHIT60 and phCMV-BaEV-TR were used instead. The expression plasmids and the packaging plasmids were mixed with polyethyleneimine (PEI) and transfected in 293FT cells. Culture medium containing viral particles was harvested 48–72 h after transfection and filtered with a 0.45-μm filter. Virus was used fresh or stored at −80°C. To transduce cells by lentivirus, cells were incubated with lentivirus solution and 10 μg/mL polybrene (Millipore). Transduced cells were selected by antibiotics or sorted by flow cytometry 48–72 h after infection. For the transduction of NK92 cells, retrovirus was concentrated by PEG8000-based virus concentrator. Concentrated retrovirus was pre-coated on retronectin-coated plate (Takara Bio). NK92 cells were then added into the plate and centrifuged at 1,000g for 90 min at 32°C. 2 μM BX795 (InvivoGen) was then added to improve the transduction efficiency.

ADCC and ADCP assay

For the NK92-CD16-mediated killing assay, NK92-CD16 cells were pre-activated with 10 ng/mL human IL-15 (GenScript, Z03308) overnight. CFSE (Millipore)-labeled target cells were coated with 10 μg/mL antibodies for 30 min. Then, unbound antibodies were removed by centrifugation. Target cells were then mixed with NK92-CD16 cells at different effector: target ratios as indicated and cultured in U-bottom 96-well plates for 6 h. Cells were dissociated by trypsin and stained with DAPI. Target cell death was examined by flow cytometry by quantifying the ratio of DAPI⁺ CFSE⁺ cells. For luminescence-based ADCC reporter assay, target cells were 1:10 mixed with Promega FcγRIIIa (V158) ADCC reporter cells (Promega, G9790) with the presence of serially diluted antibodies in a V-bottom 96-well plate for 20 h. For the ADCC reporter assay with the presence of soluble SDC1, cells were co-treated with 50 ng/mL recombinant human SDC1 (R&D, 2780-SD-CF). For the ADCC reporter assay with the presence of EIPA, cells were co-treated with 10 μM EIPA (Selleckchem, S9849). Bio-Glo reagent was added into each well and developed for 5 min 100 μL supernatant was then transferred into a flat-bottom white plate to read luminescence by a FLUOstar Omega microplate reader (BMG LABTECH). The ADCP reporter assay follows a similar protocol to the ADCC reporter assay using Promega FcγRIIa-H ADCP reporter cells (Promega, G9991).

Animal study

For subcutaneous mouse xenograft studies, 2×10⁵ HY55582-hSDC1, 4×10⁵ HY50760-hSDC1, or 5×10⁵ Panc02-hSDC1 cells were inoculated in the right flank of the C57BL/6NJ mice (female, 8-week-old). For subcutaneous human xenograft studies in NCr nude mice, 3×10⁶ AsPC-1, 5×10⁶ PATC53, 5×10⁶ PATC153, or 5×10⁶ MDA-MB231 cells were inoculated in the right flank of the NCr nude mice (female, 8-week-old). Specifically, MDA-MB231 cells were inoculated with 50% Matrigel (Corning). For orthotopic xenograft, 5×10⁶ AsPC-1 cells (in 50% Matrigel) were inoculated into the pancreas of NCr nude mice (female, 8-week-old). Tumor-bearing mice were randomly assigned to each group and received treatment as indicated: For HY55582-hSDC1 and HY50760-hSDC1 subcutaneous tumor and AsPC-1 orthotopic tumor, treatment began 3 days after inoculation. Def-22B (5 mg/kg for mIgG2a isotype, or 10 mg/kg for hIgG1 isotype in NSG mice with NK92 cell transfer) was adminstrated intraperitoneally twice per week until the end of the experiments. Depletion antibodies PK136 (10 mg/kg) and AFS98 (20 mg/kg) were administrated 24 h before every def-22B injection. RMP1-14 or 3H3 (5 mg/kg) was administrated in combination with def-22B. AMG510 (MedChemExpress, HY-114277) was administered via oral gavage at 20 mg/kg daily until the end of the experiment. MRTX1133 (MedChemExpress, HY-134813) was injected intraperitoneally at 30 mg/kg BID 2×weekly until the end of the experiment. Gemcitabine (MedChemExpress, HY-B0003) was injected intraperitoneally at 50 mg/kg twice per week until the end of the experiment. For the adoptive transfer of NK92-CD16 cells in NSG mice, 1×10⁷ NK92-CD16-IL15 cells were intraperitoneally injected weekly. Control mice received vehicle or isotype treatment. Mice were euthanized when tumors were ulcerating, according to approved guidelines of the institution’s Animal Ethics Committee. Tumor growth of subcutaneous tumors was measured by caliper. The growth of orthotopic tumors was monitored by magnetic resonance imaging. Body weight was measured by digital balance. The volume of the subcutaneous tumor was calculated by Volume = Length × Width ×Width/2.

For the determination of circulation half-life of anti-SDC1 antibody, 10 mg/kg def-22B-hIgG1 was injected intravenously into naive C57BL/6NJ mice. Serum was collected at different timepoint as indicated. The serum concentration of def-22B-hIgG1 was quantified by ELISA. Half-life was calculated by non-compartmental analysis.

Quantification and statistical analysis

Data are shown as the mean ± SD. Analyses were performed using GraphPad Prism software. Statistical analyses for comparison of grouped samples were performed using one way or two-way analysis of variance (ANOVA). Other comparisons were performed using unpaired Student’s t tests. A p-value less than 0.05 was considered statistically significant. Statistically significant differences of p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 are noted with ∗, ∗∗, ∗∗∗ and ∗∗∗∗, respectively.

Published: February 17, 2026