Identification of DLK1, a Notch ligand, as an immunotherapeutic target and regulator of tumor cell plasticity and chemoresistance in adrenocortical carcinoma

Nai-Yun Sun1, Suresh Kumar1, Yoo Sun Kim1, Diana Varghese1, Arnulfo Mendoza2, Rosa Nguyen2, Reona Okada2, Karlyne Reilly2, Brigitte Widemann2, Yves Pommier1, Fathi Elloumi1, Anjali Dhall1, Mayank Patel3, Etan Aber2, Cristie Contreras-Burrola2, Rosie Kaplan2, Dan Martinez4, Jennifer Pogoriler4, Amber K. Hamilton5, Sharon J. Diskin5, John M. Maris5, Robert W. Robey6, Michael M. Gottesman6, Jaydira Del Rivero1 and Nitin Roper1#

1Developmental Therapeutics Branch, Center for Cancer Research, NCI, Bethesda, MD, USA 2Pediatric Oncology Branch, Center for Cancer Research, NCI, Bethesda, MD, USA

3Laboratory of Pathology, Center for Cancer Research, NCI, Bethesda, MD, USA

4Department of Pathology and Laboratory Medicine, The Children’s Hospital of Philadelphia and Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA.

5Department of Pediatrics, Perelman School of Medicine, University of Pennsylvania, and Children’s Hospital of Philadelphia, Philadelphia, PA, USA.

6Laboratory of Cell Biology, Center for Cancer Research, NCI, Bethesda, MD, USA

Corresponding author:

Nitin Roper M.D., M.Sc. Investigator

Lasker Clinical Research Scholar Developmental Therapeutics Branch

Center for Cancer Research

National Cancer Institute 37 Convent Drive

Building 37, Room 5056B Bethesda, MD 20892

Phone: 240-858-3571

Email: nitin.roper@nih.gov

Funding Sources: NIH Intramural Research Program, NCI, Center for Cancer Research; ADC Therapeutics (NR and JDR); Department of Defense Rare Cancers Research Program Concept Award (NR); the Cancer Moonshot (ZIA BC0118542) funding to My Pediatric and Adult Rare Tumor Network (MyPART)

Competing interests: Nitin Roper and Jaydira Del Rivero have received research funding from ADC Therapeutics for this study. The other authors have no competing interests to report.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

40 Abstract

Immunotherapeutic targeting of cell surface proteins is an increasingly effective cancer therapy. However, given the limited number of current targets, the identification of new surface proteins, particularly those with biological importance, is critical. Here, we uncover delta-like non- canonical Notch ligand 1 (DLK1) as a cell surface protein with limited normal tissue expression and high expression in multiple refractory adult metastatic cancers including small cell lung cancer (SCLC) and adrenocortical carcinoma (ACC), a rare cancer with few effective therapies. In ACC, ADCT-701, a DLK1 targeting antibody-drug conjugate (ADC), shows potent in vitro activity among established cell lines and a new cohort of patient-derived organoids as well as robust in vivo anti- tumor responses in cell line-derived and patient-derived xenografts. However, ADCT-701 efficacy is overall limited in ACC due to high expression and activity of the drug efflux protein ABCB1 (MDR1, P-glycoprotein). In contrast, ADCT-701 is extremely potent and induces complete responses in DLK1+ ACC and SCLC in vivo models with low or no ABCB1 expression. Genetic deletion of DLK1 in ACC dramatically downregulates ABCB1 and increases ADC payload and chemotherapy sensitivity through NOTCH1-mediated adrenocortical de-differentiation. Single cell RNA-seq of ACC metastatic tumors reveals significantly decreased adrenocortical differentiation in DLK low or negative cells compared to DLK1 positive cells. This works identifies DLK1 as a novel immunotherapeutic target that regulates tumor cell plasticity and chemoresistance in ACC. Our data support targeting DLK1 with an ADC in ACC and neuroendocrine neoplasms in an active first-in-human phase I clinical trial (NCT06041516).

41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63

Introduction

Targeting cell-surface antigens with antibody-drug conjugates (ADCs) is a promising immunotherapeutic approach in oncology with recent FDA approvals across a diverse set of malignancies. Nonetheless, identification of new tumor-specific targets is imperative, especially for refractory adult metastatic tumors with few treatment options, and for less common malignancies, such as neuroendocrine (NE) neoplasms, with unique biological features.

One defining feature of neuroendocrine neoplasms, as the categorization of these cancers implies, is NE differentiation, characterized by high expression of a coordinated set of genes, including synaptophysin and chromogranin, routinely used as clinical diagnostic markers for these tumors. The Notch pathway is a major negative regulator of neuroendocrine differentiation and suppression of this pathway is common across neuroendocrine tumors1. While mechanisms of Notch pathway suppression in neuroendocrine cancers are not entirely clear, it is known that the cell surface Notch ligands such as delta-like 3 (DLL3) inhibit Notch pathway activation in normal development2. Moreover, as DLL3 expression is restricted to the brain but aberrantly expressed in many neuroendocrine cancers, DLL3 was an early immunotherapeutic target in small cell lung cancer (SCLC)3 and neuroendocrine prostate cancer4. While initial efforts to target DLL3 with an ADC failed, more recent efforts targeting DLL3 via T-cell engager strategies in SCLC have demonstrated remarkable success5,6 with recent approval by the FDA for treatment of relapsed SCLC.

To our knowledge, there has been no systematic effort to assess whether Notch ligands beyond DLL3 may or may not be targetable cell surface proteins in cancer. Therefore, in this work, we screened normal tissue and metastatic cancer datasets for expression of Notch ligands (DLL1, DLL3, DLK1, JAG1, JAG2) and uncovered DLK1 (delta-like non-canonical Notch ligand 1) as a candidate cell surface immunotherapy target protein. Moreover, we show that DLK1 is targetable by an ADC, particularly in the rare cancer adrenocortical carcinoma (ACC) in which DLK1 is highly

64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88

expressed. Importantly, we find that DLK1 is a key driver of chemoresistance in ACC through maintenance of adrenocortical differentiation and expression of the drug efflux protein ABCB1 (MDR1, P-glycoprotein) thereby demonstrating an important biological function for this new immunotherapeutic target.

DLK1 has limited normal tissue expression and high expression in multiple metastatic cancers including adrenocortical carcinoma

To assess whether Notch ligands could be suitable cell surface immunotherapeutic targets, we compared normal tissue expression of DLL1, DLL4, DLK1, JAG1, and JAG2 with DLL3 using the adult Genotype-Tissue Expression (GTEx) Portal7. As expected, expression of DLL3 was restricted to the brain (Supplementary Fig. 1A). However, other Notch ligands (DLL1, DLL4, JAG1, and JAG2) were expressed across a wide span of normal tissues (Supplementary Fig. 1A) except DLK1, which had normal expression in the adrenal gland, pituitary, ovary, hypothalamus, and testis with low to no expression in other normal tissues (Supplementary Fig. 1B). We next assessed tumor expression of DLK1 using RNA-seq data from a cohort of ~1000 adult patients with treatment refractory metastatic cancers8. We observed high DLK1 expression in a subset of refractory cancers such as sarcomas, SCLC, germ cell tumors, and grade 2 neuroendocrine tumors (Fig. 1A). High DLK1 expression has also been recently observed in pediatric neuroblastoma9. Strikingly, almost all adrenal cancers, i.e. ACC and pheochromocytoma/paraganglioma (PCPG), expressed high levels of DLK1 (Fig. 1A), which we also observed in the TCGA PanCancer dataset10 (Fig. 1B). While ACC and PCPG are both rare tumors of the adrenal gland (ACC incidence of ~0.5-2 cases per million people per year and PCPG incidence of 2-8 cases per million people per year11), we focused further analysis on ACC as it is an aggressive, highly malignant cancer with an overall poor prognosis (5-year survival 20-

89 90 91 92 93 94 Results 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114

25%) with an urgent need for new treatment options12. DLK1 was the most highly expressed Notch ligand with little to no expression of DLL3 across multiple ACC cohorts13-15 including a new RNA-seq cohort generated from ACC metastases (n=50) at our institution (Fig. 1C, Supplementary Tables 1 and 2). To validate DLK1 expression, we performed DLK1 IHC across our cohort of ACC metastatic tumors and found 97% (n=28/29) of ACC patients were DLK1+ (mean H-score 147) with H-scores ranging from 10 to 300 (Fig. 1D). Thus, our data demonstrate DLK1 as a potential new surface immunotherapeutic target in multiple malignancies, particularly ACC.

ADCT-701, an antibody drug conjugate targeting DLK1, induces cytotoxicity in ACC through apoptosis and bystander killing

Given the high and near ubiquitous expression of DLK1 in ACC, we next sought to determine if DLK1 could be targeted in ACC using a DLK1-directed antibody-drug conjugate (ADCT-701). ADCT-701 consists of a humanized anti-DLK1 monoclonal IgG1 antibody coupled to SG3199, a pyrrolobenzodiazepine (PBD) dimer, which causes potent, cytotoxic DNA interstrand cross-linking of the minor groove of DNA16 (drug-to-antibody ratio~1.8) via a Val-Ala cleavable linker and HydraSpace™ utilizing the GlycoConnect™ technology (Fig. 2A). We determined the cytotoxicity of ADCT-701 in vitro using three established ACC cell lines with varying levels of DLK1 surface expression (Fig. 2B). Compared to the isotype-control ADC (B12- PL1601), ADCT-701 inhibited cell growth in DLK1+ CU-ACC1 and H295R cells, but not in DLK1- CU-ACC2 cells (Fig. 2C). However, DLK1 CU-ACC2 cells, similar to CU-ACC1 and H295R cells, were sensitive to the PBD payload of ADCT-701 (Supplementary Fig. 2A). We next used CRISPR-Cas9 gene editing in the CU-ACC1 cell line and established multiple single cell clones with complete loss of DLK1 (Supplementary Fig. 2B, C). In several CU-ACC1 DLK1 KO clones, ADCT-701 cytotoxicity was abrogated (Supplementary Fig. 2D) thereby validating the DLK1- specific cytotoxicity of ADCT-701. We observed similar findings in the DLK1- PCPG cell line

115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140

hPheo117 (Supplementary Fig. 2E, F). Taken together, these data demonstrate that ADCT-701 exhibits in vitro cytotoxic activity in ACC in a DLK1-dependent manner.

We next evaluated the mechanism by which ADCT-701 induces cell death. Cellular internalization of ADC after binding to the surface target is an essential step for ADC cytotoxicity18. To demonstrate that our anti-DLK1 mAb could be internalized efficiently, we quantified the cellular internalization rates of DLK1 antibody across ACC cell lines with varying DLK1 expression levels using imaging flow cytometry. DLK1 antibody was rapidly internalized in DLK1+ CU-ACC1 and H295R cells but not in DLK1- CU-ACC2 cells (Fig. 2D). Next, consistent with previous studies demonstrating PBD induced DNA interstrand crosslinks results in cell cycle arrest19, we found that CU-ACC1 and H295R cells treated with ADCT-701 were blocked in the G2/M phase (Supplementary Fig. 3A and Supplementary Fig. 4A). We then determined the presence of DNA double-strand breaks by yH2AX, as well as apoptosis by cleaved caspase-3 and cleaved poly (adenosine diphosphate-ribose) polymerase (PARP). YH2AX, cleaved caspase 3, and cleaved PARP were upregulated in CU-ACC1 and H295R cells after ADCT-701 but not after B12-PL1601 treatment (Supplementary Fig. 3B) and ADCT-701 significantly increased apoptosis (Annexin V+/PI+) compared to untreated and B12-PL1601 treated cells (Supplementary Fig. 3C and Supplementary Fig. 4B). Collectively, these results suggest that ADCT-701 treatment leads to DNA double-strand breaks, G2/M arrest, and ultimately apoptosis.

Due to the heterogeneous expression of DLK1 in ACC (Fig. 1E), we next assessed for potential bystander killing18 using a system in which DLK1 KO CU-ACC1 cells were cultured with DLK1-expressing parental CU-ACC1 cells at various ratios. We observed greater cytotoxicity of DLK1 KO CU-ACC1 cells than expected with no cytotoxicity observed with B12-PL1601 treatment (Supplementary Fig. 3D). We further investigated bystander killing by conditioned media transfer experiments in which DLK1+ CU-ACC1 cells were treated with ADCT-701 or B12-PL1601 before transferring the media to DLK1+ CU-ACC1 cells or DLK1 KO CU-ACC1 cells. ADCT-701 conditioned media induced cytotoxicity in DLK1+ CU-ACC1 cells similar to treatment with ADCT-

141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166

701 (Supplementary Fig. 3E). ADCT-701 conditioned media also elicited bystander killing, as demonstrated by greater cytotoxicity in DLK1 knockout CU-ACC1 cells compared with B12- PL1601 conditioned media or ADCT-701 treatment (Supplementary Fig. 3E). Overall, these results indicate that ADCT-701 not only can target DLK1+ cells but can also indirectly induce cytotoxicity in DLK1 cells.

ADCT-701 has potent in vitro activity in DLK1+ ACC patient-derived organoids and induces

robust anti-tumor responses in ACC cell line-derived and patient-derived xenografts

Since ACC is a rare cancer type with few available human cell lines20, we sought to validate the in vitro cytotoxicity of ADCT-701 in a newly developed cohort of ACC short-term patient-derived organoids (PDOs) (defined as less than 5 total passages) (Supplementary Table 3). Overall, we found 50% (n=6/12) of PDOs responded to ADCT-701 (Fig. 2E) and 50% (n=6/12) of PDOs had no response (Fig. 2F). As expected, all ADCT-701 responders were DLK1+ (Fig. 2G), However, among ADCT-701 non-responders, 50% (n=3/6) were still DLK1+ (Fig. 2H) suggesting that ADCT-701 sensitivity is influenced by factors other than DLK1 expression.

Next, to further explore the potential for targeting DLK1 in ACC, we evaluated responses to ADCT-701 among DLK1+ human ACC cell line-derived xenograft and ACC patient-derived (PDX) models (Fig. 2I, J). ADCT-701 treatment elicited durable anti-tumor responses and significantly prolonged the survival of both CU-ACC1 and H295R tumor-bearing mice compared with tumor-bearing mice treated with saline or B12-PL1601 (Fig. 2I and Supplementary Fig. 5A). However, H295R tumors eventually became resistant to ADCT-701, whereas CU-ACC1 tumors remained sensitive to ADCT-701 for up to 100 days. While the DLK1 antibody within ADCT-701 targets human not mouse DLK1, no body weight loss in mice was observed with ADCT-701 treatment (Supplementary Fig. 5B) suggesting minimal off-target payload activity.

We next investigated the anti-tumor activity of ADCT-701 among three DLK1+ ACC PDX models: 164165, 592788, and POBNCI_ACC004 (Fig. 2J and Supplementary Fig. 6A). All three

167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192

models were validated as ACC tumors based on IHC expression of the common neuroendocrine marker synaptophysin, the adrenal specific marker SF1, and the cell proliferation marker Ki67 (Supplementary Fig. 6B-D). ADCT-701 treatment led to tumor growth inhibition and significant lengthening of survival among 164165 PDX and 592788 PDX mice (Fig. 2J and Supplementary Fig. 5C). Strikingly, ADCT-701 induced complete responses in all treated POBNCI_ACC004 PDX tumors (5/5). However, 3/5 POBNCI_ACC004 PDX tumors quickly recurred and did not respond to ADCT-701 re-treatment (Fig. 2J). Similar to treated xenografts, ADCT-701 was well-tolerated in PDX models as determined by body weight measurements (Supplementary Fig. 5D).

ABCB1, a drug efflux protein, mediates intrinsic and acquired ADC and chemotherapy

resistance among DLK1+ ACC pre-clinical models

We next sought to decipher mechanisms of resistance to ADCT-701 in our DLK1+ ACC pre-clinical models. As payload insensitivity is known to mediate ADC resistance18, we first tested the cytotoxicity of PBD across DLK1+ ADCT-701 responder and non-responder PDOs. Strikingly, we observed extreme resistance to PBD among non-responder PDOs, with PBD average IC50s of 97 nM in NCI-ACC40, 31 nM in NCI-ACC48, and 54 nM in NCI-ACC54, which represents close to 1000x greater resistance than previously reported for PBD21 (Fig. 3A). We then explored the activity of common therapies used to treat ACC in the clinic12 (mitotane, etoposide, doxorubicin, and carboplatin) among two DLK1+ ACC PDOs with and without response to ADCT-701 and PBD (NCI-ACC51 and NCI-ACC48, respectively) that were able to grow in culture longer than other PDOs (greater than 5 passages). We observed resistance to chemotherapy in the NCI-ACC48 PDO compared to the NCI-ACC51 PDO but no difference in mitotane activity (Supplementary Fig. 7A). We next looked for potential mechanisms of resistance to PBD by analyzing our NCI-ACC RNA-seq patient dataset for expression of drug transporter genes previously implicated in resistance to PBD-ADCs22 as well as commonly upregulated drug efflux pumps of the ABC transporter family21. ABCB1 (MDR1 or p-glycoprotein) and ABCG2 had the greatest difference in

193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218

expression between NCI-ACC48 and NCI-ACC51 patient tumors (Fig. 3B) suggesting these drug efflux genes could explain the difference in PBD resistance in corresponding PDOs. Indeed, the NCI-ACC48 PDO had much higher surface expression of ABCB1 than the NCI-ACC51 PDO (Fig. 3C). Furthermore, co-treatment of 3 different ABCB1 inhibitors (valspodar, elacridar, and tariquidar) with PBD and ADCT-701 all showed dramatic reversal of resistance in the NCI-ACC48 PDO (Fig. 3D and Supplementary Fig. 7B). ABCB1 inhibitors more modestly increased sensitivity to ADCT-701 and PBD in the NCI-ACC51 PDO demonstrating a functionally lower level of ABCB1 activity in this model compared to the NCI-ACC48 PDO (Supplementary Fig. 7C). Thus, these results indicate that primary in vitro resistance to ADCT-701 can be mediated by high expression and activity of the drug efflux protein ABCB1.

We next sought to determine if ABCB1 expression and activity could also explain differences in ADCT-701 in vivo activity across our ACC cell line-derived xenograft and PDX models. Among our ACC xenografts, CU-ACC1 had lower surface ABCB1 expression than H295R (Supplementary Fig. 8A), which could at least partially explain the long-term tumor control with ADCT-701 treatment in CU-ACC1 but not H295R tumors (Fig. 2I). Among our ACC PDXs, there was considerably lower surface ABCB1 expression in POBNCI_ACC004 (Fig. 3E), which had initial complete responses to ADCT-701, compared to both 164165 and 592788 (Fig. 3E), which had partial but no complete anti-tumor responses to ADCT-701. To further test the role of ABCB1 in relation to ADCT-701 activity in these models, we developed PDX-derived organoids from untreated POBNCI_ACC004, 164165 and 592788 PDX tumors (i.e. PDXOs). Consistent with our in vivo data, POBNCI_ACC004 PDXO was much more sensitive to SG3199 than 164165 and 592788 PDXOs (Fig. 3F). ABCB1 inhibitors also increased sensitivity to SG3199 among 164165 and 592788 PDXOs (Fig. 3G) demonstrating that ABCB1 drug efflux activity is a mechanism of intrinsic resistance to ADCT-701 in these two models.

Although POBNCI_ACC004 PDX tumors had initial complete responses to ADCT-701 treatment, these tumors quickly relapsed and were unresponsive to additional ADCT-701 doses

219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244

245 (Fig. 2J). Therefore, to assess mechanisms of acquired resistance to ADCT-701, we performed RNA-seq on POBNCI_ACC004 PDX tumors resistant to ADCT-701 treatment (n=3) and saline treated control tumors (n=4). Using differential gene expression analysis, we observed upregulation of ABCB1 expression in POBNCI_ACC004 PDX tumors resistant to ADCT-701 compared to controls (Fig. 3H and Supplementary Table 4). Surface ABCB1 expression was also highly upregulated in an POBNCI_ACC004 resistant compared to an untreated POBNCI_ACC004 control tumor (Fig. 3I). We then developed a PDXO from a POBNCI_ACC004 resistant tumor and found that ABCB1 inhibitors re-sensitized the POBNCI_ACC004 resistant PDXO to PBD and ADCT-701 (Fig. 3J and Supplementary Fig. 8C) demonstrating the role of this drug efflux transporter in mediating acquired resistance to ADCT-701. Lastly, unlike in neuroblastoma23, we found no difference in DLK1 expression by IHC between pre- and post- ADCT-701 treated ACC PDX tumors (Supplementary Fig. 8D) suggesting that selection and outgrowth of DLK1 negative cells does not contribute to ADCT-701 acquired resistance in ACC.

ADCT-701 elicits complete, durable responses in DLK1+ small cell lung cancer tumors without ABCB1 expression

As we found DLK1 to be expressed in a subset of metastatic cancers apart from ACC (Fig. 1A), we hypothesized that ADCT-701 would be highly effective against DLK1+ tumors with low or no ABCB1 expression such as SCLC. We therefore screened SCLC cell lines for expression of DLK1 and found 22% (n=11/51) were DLK1+ (Supplementary Fig. 9A). We then selected three DLK1+ SCLC cell lines (H524, H146, and H1436) and confirmed that cell surface DLK1 expression was at a level equal to or higher than the known SCLC target DLL3 (Fig. 3K and Supplementary Fig. 9B). All three SCLC cell lines also lacked surface ABCB1 expression (Fig. 3L) and were highly sensitive to both SG3199 and ADCT-701 in vitro (Fig. 3M, N). In vivo, ADCT-701 treatment resulted in complete responses and long-term tumor-free survival compared to B12-PL1601 and saline in all three of the SCLC xenograft models (Fig. 3O and Supplementary Fig. 9C) without

246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270

appreciable body weight loss (Supplementary Fig. 9D). Notably, the SCLC H146 xenograft, which had very low DLK1 expression (H-score 30), also had long-term complete responses with ADCT- 701 treatment (Fig. 3O). Thus, our results indicate that ADCT-701 can effectively target DLK1+ tumors with low or no ABCB1 expression.

DLK1 is a major regulator of ABCB1, adrenocortical differentiation, and chemoresistance in ACC

We next sought to assess whether DLK1 has a functional role in ACC using our established CU-ACC1 DLK1 KO cells (Supplementary Fig. 2B, C). As DLK1 has been shown to activate or inhibit NOTCH1 signaling in several model systems24, we assessed expression of NOTCH1 in DLK1 KO compared to DLK1 WT parental cells and observed upregulation of total NOTCH1 and the active, intracellular domain (ICD) of NOTCH1 (N1ICD) (Fig. 4A). There was also a dramatic reduction in the NE protein synaptophysin (Fig. 4A) after DLK1 KO consistent with the known role of active Notch signaling in downregulating NE gene expression25. Correspondingly, we observed a significant negative correlation between NOTCH1 and DLK1 expression across TCGA ACC tumors (Fig. 4B). DLK1 and NOTCH1 expression were also significantly higher and lower, respectively, in ACC tumors compared to the normal adrenal gland (Fig. 4℃).

Given that prior work has shown Notch-active tumors with low NE gene expression (i.e. non-NE) to be chemoresistant26, we performed cytotoxicity assays with PBD in CU-ACC1 cells with and without DLK1 KO. Surprisingly, DLK1 KO cells were much more sensitive to PBD (Fig. 4D) and chemotherapeutics such as etoposide and doxorubicin (Supplementary Fig. 10A). However, DLK1 KO cells displayed no change in proliferation compared to parental cells. (Supplementary Fig. 10B). Strikingly, DLK1 KO had near complete loss of ABCB1 surface expression compared to parental cells, which showed a broad, bi-modal distribution of ABCB1 (Fig. 4E). In DLK1 KO cells, we also observed strong downregulation of steroidogenesis protein

271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296

CYP17A1 (Fig. 4A), which is highly expressed in the adrenal cortex and composes part of the Adrenocortical Differentiation Score (ADS)14. However, CYP17A1 expression was not decreased 299 with siRNA downregulation of DLK1 (Supplementary Fig. 10C) suggesting longer-term NOTCH1 300 signaling, known to be required to induce transdifferentiation26, is required to induce changes in 301 adrenocortical differentiation. CU-ACC1 DLK1 KO clones also had dramatically lower secretion of cortisol compared to parental cells (Fig. 4F). As we observed DLK1 KO cells to be completely adherent compared to a mixed phenotype (suspension and adherent) of CU-ACC1 parental cells (Supplementary Fig. 10D) we isolated suspension and adherent CU-ACC1 cells (Supplementary Fig. 10E). Suspension CU-ACC1 cells showed high expression of DLK1 and SYP and low expression of N1ICD (Supplementary Fig. 10F). In contrast, adherent CU-ACC1 cells showed low expression of DLK1 and SYP and high expression of N1ICD (Supplementary Fig. 10F). CU-ACC1 adherent cells also had lower expression of surface ABCB1 with modest differences in chemosensitivity compared to CU-ACC1 suspension cells (Supplementary Fig. 10G, H).

To validate our DLK1 KO results, we assessed expression of NOTCH1, NE and adrenocortical differentiation proteins in an ACC PDO with high DLK1 expression (NCI-ACC48) and an ACC PDO with no DLK1 expression (NCI-ACC49) (Figure 4G). We observed higher expression of N1ICD and much lower expression of SYP and CYP17A1 in the DLK1- NCI-ACC49 PDO compared to the DLK1+ NCI-ACC48 PDO (Figure 4G). The NCI-ACC49 PDO was also highly sensitive to SG3199, etoposide and doxorubicin (Supplementary Fig. 10I) and ABCB1 was not expressed (Fig. 4H). We next overexpressed N1ICD in CU-ACC1 cells, and similar to DLK1 KO cells, we observed decreased expression of synaptophysin and near complete downregulation of CYP17A1 and ABCB1 (Figure 4I, J). However, there were only minor differences in chemosensitivity between CU-ACC1 and N1ICD-overexpressing CU-ACC1 cells (Supplementary Fig. 10J) suggesting DLK1 KO may mediate chemosensitivity through additional mechanisms apart from ABCB1 expression and NOTCH1 signaling. Lastly, we analyzed single- cell RNA-seq data from 18 ACC metastatic tumors (Aber et al. manuscript in submission) and

297 298 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322

observed a significantly higher ADS among cells with high DLK1 expression compared to cells with low or no DLK1 expression (Fig. 4K) supporting our experimental results. Altogether, our data suggest a model by which DLK1, through inhibition of NOTCH1 signaling, maintains adrenocortical differentiation and high ABCB1 expression and imparts ADC and chemoresistance in ACC (Fig. 4L). Based on our data, DLK1-directed ADCs would also be expected to have greater activity in ACC tumors with positive but low DLK1 expression due to decreased adrenocortical differentiation and ABCB1 expression (Fig. 4L).

A first-in-human phase I clinical trial of ADCT-701 in patients with ACC and neuroendocrine neoplasms

Based on the pre-clinical efficacy of ADCT-701 in ACC and SCLC, as well as parallel data in neuroblastoma9, a first-in-human phase 1 clinical trial was developed to test the safety and preliminary efficacy of ADCT-701 in adult patients with ACC and neuroendocrine neoplasms. This trial (NCT06041516) is currently recruiting patients with the primary objective to determine the maximum tolerated dose (MTD) of ADCT-701.

Discussion

In this work, we have identified DLK1 as a cancer cell surface antigen that can be successfully targeted with an ADC in pre-clinical models of refractory metastatic cancers, namely ACC and SCLC. While ADCs are an effective and increasingly common cancer therapy, approval is currently limited to select malignancies (i.e., breast cancer, urothelial cancers, and ovarian cancers) with overall few antigen targets (i.e. TROP2, nectin-4, HER2, tissue factor, and folate receptor alpha18). Thus, identifying new and optimal cell surface targets, such as DLK1, is an important step towards broadening the therapeutic potential of ADCs. Indeed, based on our pre- clinical data, we have initiated a first-in-human phase 1 clinical trial with an ADC targeting DLK1 (NCT06041516). To our knowledge, this is the first ADC clinical trial for patients with ACC and

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348

neuroendocrine neoplasms including rare neuroendocrine malignancies such as PCPG and adult neuroblastoma.

In addition to identifying DLK1 as a new immunotherapeutic target, we demonstrate a novel role for DLK1 in conferring ADC and chemoresistance through high expression and activity of the multidrug efflux pump ABCB1. Intrinsic resistance to ADCs is common in the clinic with the ABCB1 drug transporter considered to be one potential mechanism27. Specifically in ACC, chemotherapy resistance has long been attributed to high expression of ABCB128,29, which is known to be one of several genes highly expressed in the adrenal cortex14. Our findings could thus have important implications for ACC treatment as inhibition of DLK1 could be a strategy to reduce resistance to chemotherapy in ACC, particularly the EDP (etoposide, doxorubicin, cisplatin) regimen which includes two chemotherapeutic drugs known to be ABCB1 transport substrates (etoposide and doxorubicin). EDP chemotherapy is widely used to treat advanced ACC but has not improved overall survival1,30.

A key new mechanistic finding from this work is the identification of DLK1 as a driver of adrenocortical. Previous work has identified several key regulators of adrenocortical differentiation such as the histone methyltransferase EZH231,32 9 and the deSUMOylating enzyme SENP233,34. Our data suggest that DLK1 maintains adrenal steroidogenic cell differentiation, which is concordant with recent data from a spatial transcriptomic analysis of DLK1+ and DLK1- ACC tumor regions35. Moreover, our data are consistent with the known role of DLK1 regulating cellular differentiation processes such as adipogenesis, hematopoiesis, stem cell homeostasis, neurogenesis, angiogenesis, and muscle regeneration36. In regard to ABCB1, we demonstrate that long-term NOTCH1 activation and de-differentiation of ACC are required to downregulate ABCB1. As prior work has demonstrated NOTCH1 as a positive regulator of ABCB1 expression in several other solid tumor models37, our data highlights the context dependent nature of Notch signaling30.

349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373

Our experimental data also uncovers a role for DLK1 in transdifferentiating cells from a NE to non-NE state, which to our knowledge, has not been previously known, although DLK1 has been observed to be upregulated in NE tumors31,32. Interestingly, adrenocortical carcinomas are not typically categorized as NE tumors as they are not of neuroepithelial origin and they generally lack expression of NE genes such as chromogranin A38. Indeed, neuroendocrine scoring systems have used gene expression data from the normal adrenal cortex to identify non-NE genes compared to NE genes in the normal adrenal medulla39. However, ACC tumors commonly express NE genes such as synaptophysin40 and our data demonstrate that synaptophysin is regulated by DLK1 through NOTCH1. Low or no expression of synaptophysin on routine ACC tumor specimens, albeit likely low in prevalence, could indicate a non-NE, less adrenocortical differentiated state (i.e. DLK1low/NOTCH1high/ABCB1low) with sensitivity to chemotherapy or an ADC. Counterintuitively, our data suggest that high DLK1 expression may not be an optimal biomarker for a DLK1-directed ADC as DLK1high tumors would be expected to have high ABCB1 expression and thereby demonstrate payload resistance. Rather, tumors with low DLK1 expression (which are able to bind and internalize a DLK1-directed ADC) may exhibit the most optimal ADC response due to low ABCB1 expression.

The direct link we propose between DLK1, NOTCH1 signaling, and ABCB1 expression also suggest that ABCB1 inhibition could improve anti-tumor responses to DLK1-directed ADCs. However, a clinical trial testing the addition of the ABCB1 antagonist, tariquidar, to chemotherapy among ACC patients was stopped prematurely due to toxicity41. Off-target toxicity of the combination of ABCB1 inhibitors and chemotherapy (such as to bone marrow whose stem cells are protected from chemotherapy by expression of ABC efflux transporters such as ABCG and ABCB1) should be much less of a concern when ABCB1 inhibitors are combined with targeted therapy such as ADCs, which have reduced off-target toxicity. One other strategy for future ADCs could be to use payloads that are not substrates for drug efflux transporters.

374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398

Beyond ADCs, degrader-antibody conjugates18 targeting DLK1 may be a strategy to downregulate DLK1, which could potentially sensitize ACC tumors to chemotherapy or other ADCs. Moreover, there are now multiple immunotherapeutic strategies to target cancer cell surface proteins such as CAR T cells and bi-specific T cell engagers (BiTEs). Indeed, DLK1- directed CAR T cells have been shown to have pre-clinical efficacy among DLK1-expressing hepatocellular carcinoma models42. CAR T cells may be a particularly attractive option in ACC given the high level of chemoresistance; however, CAR T cells generally have more toxicity than ADCs43 and thus it may advisable to accrue safety information on targeting DLK1 from our phase 1 study before pursuing clinical testing of CAR T cells. Although DLK1 has minimal expression across most normal tissues, there is high expression in several organs such as the adrenal gland, particularly the adrenal medulla compared to the adrenal cortex9,44. ADCT-701 treatment may thus lead to adrenal hormone deficiency requiring supplementation with mineralocorticoids and/or corticosteroids. Another active phase 1 trial targeting DLK1 with an ADC in advanced cancers (NCT06005740)45 using monomethyl auristatin E (MMAE) as the payload (also an ABCB1 substrate), may also provide additional safety information.

There are several limitations to our study. While we demonstrate that DLK1 is a regulator of ABCB1 expression and thereby sensitivity to a DLK1-directed ADC and chemotherapy, there are likely additional variables which affect ADC and chemotherapy sensitivity that we are unable to account for in this study. Moreover, while we focused on the role of DLK1 in mediating ADC resistance in our functional studies, DLK1 is known to regulate cancer stemness36,46,47 and tumor progression48. Thus, further investigation into the role of DLK1 in ACC tumorigenesis will be important. Also, although our study focused on ACC and SCLC, our screening data revealed high DLK1 expression across several additional metastatic tumor types such as germ cell tumors and sarcomas and recent parallel work has demonstrated DLK1 as an immunotherapeutic cell surface target in pediatric neuroblastoma9. Thus, a biomarker-based assessment of DLK1 across a

399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423

broader group of malignancies could be a future clinical approach for DLK1-directed immunotherapeutic clinical trials.

In summary, we have identified DLK1 as a new immunotherapeutic target in ACC and neuroendocrine neoplasms such as SCLC. We have also demonstrated DLK1 as an important driver of chemotherapy and ADC resistance through regulation of the drug efflux pump ABCB1. Our data support the clinical testing of targeting DLK1 with an ADC in ACC and other neuroendocrine neoplasms and identify DLK1 as an important cell surface target for future immunotherapeutic approaches.

424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445

446 447 448 449

Materials and Methods

ACC patient tumor specimens

ACC patient tumors were collected from surgical resection of metastatic sites at the NIH Clinical Center under NIH Institutional Review Board protocols (NCT05237934, NCT01109394, and NCT03739827). Tumor tissue from surgical resections was used for organoid experiments and DLK1 IHC. RNA-sequencing was performed from formalin-fixed tissue acquired either from surgical resected tissue or from archival tissue. A summary of ACC patient tumors with associated RNA-seq, DLK1 IHC and/or organoid assay data used in this study are shown in Supplementary Table 1.

Bulk and single-cell RNA sequencing

Formalin-fixed, paraffin-embedded (FFPE) tumor tissue samples were prepared for bulk RNA sequencing (RNA-Seq). RNA-seq libraries were prepared using Illumina TruSeq RNA Access Library Prep Kit or Total RNA Library Prep Kit according to the manufacturer’s protocol (Illumina). The NCI CCBR RNA-seq pipeline (https://github.com/skchronicles/RNA-seek.git) was used for further processing. In summary, STAR (2.7.6a) was run to map reads to hg38 (release 36) reference genome. Then, RSEM was used to generate gene expression values in log2(FPKM+1). We applied the “RemoveBatchEffect” function from the package Limma to remove the impact of the library preparation protocols (access or totalRNA). Single cell RNA-seq was performed as described in Aber et al. manuscript in submission. In brief, single cell suspensions from 18 ACC liver and/or lung metastases were sequenced on the 10x Genomics Chromium Platform targeting 6,000 cells per sample. Data was processed using the cellranger pipeline and downstream analysis performed in R using Seurat. Our analysis categorized DLK1 expression as high or low/negative in malignant cells only (identified by copy number variation using inferCNV).

450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474

An adrenocortical differentiation score was calculated using genes from a previously published adrenocortical differentiation gene set apart from DLK1, which was excluded14.

Tumor cells isolation and organoid culture

To generate organoid cultures, fresh ACC patient tumors were minced into tiny fragments in the sterile dish. Tumor fragments were performed to enzymatic digestion in advanced DMEM/F12 supplemented with 1x Glutamax and 10 mM HEPS buffer containing collagenase type IV (200 U/ml, Sigma-Aldrich) and DNase I (50 U/ml, Sigma-Aldrich) on an orbital shaker for 1 hr at 37 ℃ and filtered through 70 um strainers. The mixture was spun for 5 min at 1500 rpm. The cell pellet was treated with 1 x RBC lysis buffer (Sigma-Aldrich #) for 5 min at room temperature to remove the red blood cells and then spun for 5 min at 1500 rpm. Single cell suspensions were seeded on a Matrigel dome. Matrigel was mixed with tumor cells in minimum basal medium (MBM) consisting of 1x N2 supplement (Thermo Fisher Scientific #17502048), 1x B27 supplement (Thermo Fisher Scientific #17504044), 50 ng/ml EGF (Thermo Fisher Scientific PHG0311), 20 ng/ml bFGF (STEMCELL Technologies #78003), 100 ng/ml IGF-2 (STEMCELL Technologies #78023), and 10 µM Y-27632 (STEMCELL Technologies #72308) at a 1:1 ratio and added to a 6-well plate. Each well was overlayed with 2 ml MBM medium after Matrigel had solidified in a 37°C and 5% CO2 culture incubator for 20 min. ACC organoid culture MBM medium was refreshed once a week. Every 2-4 weeks organoids were passaged by mechanical pipetting Matrigel gently using Dispase in DMEM/F12 media (STEMCELL Technologies #) and several washes with PBS until Matrigel was cleared out. Organoid fragments were then re-suspended in Matrigel and seeded as described above.

In vitro short-term organoid culture cytotoxicity assays

Human ACC patient tumor single cells were embedded in 10 ul of MBM medium with 50% Growth Factor Reduced-Matrigel (Corning #354230) on 384-well white plate at a concentration of

475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500

2000 cells per well. After solidification of the Matrigel for 30 min at 37℃, 20 ul fresh MBM medium was added to each well, and the plates were further incubated for 2 days. After the 2 days of pre- culture, cells were treated with 30 ul ADCT-701 and B12-PL1601 for 7 days or with 30 ul SG3199 or other chemotherapeutic drugs for 3 days. For ADCT-701 with or without ABCB1 inhibitors cytotoxicity, NCI-ACC51 and NCI-ACC48 organoid single cells were embedded in Matrigel on 384-well white plates. After 2 days of incubation, cells were treated with different concentration of ADCT-701 or SG3199 combined with or without 1uM valspodar, 10 uM elacridar, and 1uM tariquidar for 7 days or 3 days respectively. For chemosensitivity of NCI-ACC51 and NCI-ACC48 organoids, cells were plated in 384-well white plates as in the previous seeding steps. After 2 days incubation, cells were treated with mitotane, etoposide, doxorubicin, or carboplatin for 3 days. 20 ul of CellTiter-Glo 2.0 reagent (Promega G9241) was added and the luminescence was quantified with a SpectraMax i3x reader (Molecular Devices).

IHC staining

4-5 um sections from formalin-fixed, paraffin-embedded blocks were stained using the Bond Refine polymer staining kit (Leica Biosystems DS9800) for DLK1 antibody (dilution 1:2000, Abcam, ab21682) antibody on the Bond Rx automated staining system (Leica Biosystems) following standard IHC protocols with some modifications. Briefly, the slides were deparaffinized and incubated with E1 (Leica Biosystems) retrieval solution for 20 minutes. Primary antibody was incubated for 1 hr at room temperature and no post-primary step was performed. Cover-slipped slides were scanned with a Aperio CS-O slide scanner (Leica Biosystems). DLK1 immunohistochemistry was scored by a pediatric pathologist. Each case was scored for the most prominent intensity (0-3 with 1 representing equivocal, 2 weak, and 3 strong positive staining) as well as for percentage of staining. A modified H-score was calculated as intensity multiplied by percentage of positively stained cells.

501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525

For CD56, Ki67, and synaptophysin immunohistochemistry, auto-stainers Ventana Benchmark Ultra (Ventana, Tucson, AZ), were used. Leica Bond Max (Leica Biosystems, Deerfield, IL) auto-stainer was used for SF1 immunohistochemistry. Validation of these stains was performed on daily clinical laboratory controls by the Anatomic Pathologist on clinical service at the Laboratory of Pathology, National Cancer Institute. A Roche Diagnostics anti-CD56 antibody (rabbit, monoclonal, 760-4596, clone MRQ-42, Indianapolis, IN) was used at a prediluted concentration. An Agilent Technologies anti-Ki67 antibody (mouse, monoclonal, # M7240, clone MIB-1, Santa Clara, CA) was used at a dilution of 1:200. A Perseus Proteomics anti-SF1 antibody (mouse, monoclonal, PP-N1665-00, clone N1665, Komaba, Japan) was used at a dilution of 1:200. A Roche Diagnostics anti-synaptophysin antibody (rabbit, monoclonal, # 790-4407, clone SP11, Indianapolis, IN) was used at a prediluted concentration.

Cell lines

Human ACC cell lines CU-ACC1 and CU-ACC2 were obtained from the University of Colorado. CU-ACC cells were cultured in 3:1 (v/v) Ham’s F-12 Nutrient Mixture (Gibco #)-DMEM (Gibco #) containing 10% heat-inactivated FBS (Gemini Bio 100-106), 0.4 µg/mL hydrocortisone (Sigma-Aldrich H6909), 5 µg/mL insulin (Sigma-Aldrich #), 8.4 ng/mL cholera toxin (Sigma- Aldrich #), 10 ng/ml epidermal growth factor (Invitrogen #), 24 ug/mL adenine (Sigma-Aldrich #) and 1% Penicillin-Streptomycin (Gibco #15140122)]. Human ACC cell line H295R (CRL-2128) was obtained from ATCC and cultured in 1:1 DMEM:F12 (Gibco #11320082) containing 2.5% Nu- Serum (Corning #355100), 1% ITS+ Premix Universal Culture Supplement (Corning #354352), and 1% Penicillin-Streptomycin. Human SCLC cell lines H146, H524, and H1618 were obtained from ATCC. Human SCLC cell line H1436 was obtained from Haobin Chen (Washington University). H146 and H524 cells were cultured in RPMI-1640 (Corning MT10040CM) supplemented with 10% FBS and 1% Penicillin-Streptomycin. H1618 and H1436 cells were cultured in HITES media DMEM/F12 (1:1) (Gibco #11320082) containing 5% FBS, 1x

526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551

GlutamaxTM (Gibco #35050061), 10 nM Hydrocortisone, 10 nM beta-estradiol (Sigma-Aldrich #E2758), Insulin-Transferrin-Selenium mix/solution (Invitrogen #41400045), and 1% Penicillin- Streptomycin. All cell lines were cultured at 37℃ in a humidified incubator with 5% CO2, regularly tested to be mycoplasma-negative (Lonza LT07-318) and authenticated by STR profiling (Laragen Inc.).

Flow cytometric analysis

For surface DLK1 expression analysis, human ACC cells, ACC patient tumor cells, ACC PDX cells, and human SCLC cells were harvested and washed with FACS buffer (PBS containing 1% BSA and 0.1% sodium azide). Cells were incubated with the anti-human DLK1 primary antibody (AG-20A-0070-C100 AdipoGen; 1:100 per million cells) at 4℃ for 30 minutes in the dark. Cells were washed by FACS buffer. Cells were then incubated with secondary antibody (Invitrogen, P852; 1:500) at 4℃ for 30 minutes in the dark. For surface DLL3 expression analysis, human SCLC cells were collected and washed with FACS buffer. Cells were stained with human DLL3-PE (R&D Systems, FAB4315P; 10 ul per one million cells) or isotype control antibody (R&D Systems, IC108P; 10 ul per one million cells) at room temperature for 30 minutes in the dark. For surface MDR-1 expression analysis, human ACC cells, ACC patient tumor cells, ACC PDX cells, and human SCLC cells were collected and washed with FACS buffer. Cells were stained with human CD243-PE (Biolegend, #348606; 5 ul per one million cells in 100 ul wash buffer) or isotype control antibody (Biolegend, #400214; 5 ul per one million cells in 100 ul wash buffer) at 4℃ for 30 minutes in the dark. Cells were washed and then stained with PI (Biolegend, #421301;1:100) following above antibodies incubation. Living cells were separated as PI negative cells. To semi-quantitate DLK1 or DLL3 cell surface expression in ACC and SCLC cell lines, cell surface molecules of DLK1 or DLL3 per cell were calculated after subtracting background signal from DLK1 secondary antibody alone (Invitrogen, P852) or DLL3 isotype control antibody (R&D Systems, IC108P) respectively by BD Quantibrite Beads PE Fluorescence Quantitation Kit (BD

552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577

Bioscience, #340495) in accordance with the manufacturer’s protocol. Stained cells were acquired on LSR Fortessa (BD Biosciences), and data were analyzed using FlowJo software version 10.8.1. Flow cytometry gating strategies are shown in Supplementary Fig. 11A-C.

In vitro cell line cytotoxicity assays

ACC or SCLC cells were seeded into 384-well white plate at a concentration of 1500 cells per well in 30 ul medium and allowed to adhere overnight. The PCPG cell line, hPheo1, 300 cells per well in 30 ul medium were plated into 384-well white plate for overnight. 30 ul fresh medium containing different concentrations of antibody-drug conjugate (ADCT-701 and B12-PL-1601) or free payload (SG3199) was added to each well and the plates were further incubated for 7 days or 3 days respectively. After 7 days (ADCT-701 and B12-PL1601) or 3 days (SG3199) incubation, 20 ul of CellTiter-Glo 2.0 reagent (Promega G9241) was added and the luminescence was recorded using a SpectraMax i3x reader (Molecular Devices).

Imaging flow cytometry

A total of 1 x 106 CU-ACC1 or H295R cells were seeded in each well of six-well plate and allowed to attach overnight. Then, attached cells were incubated with APC-conjugated DLK1 antibody (R&D Systems, a bio-techne brand FAB1144A) or isotype control antibody (R&D Systems, a bio-techne brand IC0041A) for 1 hour in 2 ml of cell culture media at 37℃. Then, the cell monolayer was collected and rinsed with cold PBS twice and resuspended. The cellular internalization rate of DLK1 antibody in treated cells was evaluated using an Amnis ImageStreamX Mark II imaging flow cytometry (Luminex, Austin, TX, USA).

Cell cycle and apoptosis assays

For EdU incorporation studies, cells were processed as per the manufacturer’s instructions (invitrogen C10634). Briefly, 3 x 105 CU-ACC1 or H295R cells were plated in 6-well

578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603

plates and allowed to adhere overnight. CU-ACC1 or H295R cells were then treated with 0.02 ug/mL ADCT-701 or B12-PL1601 for 2 or 5 days, respectively. Cells were labeled with 10 uM Click-IT™ EdU in a 37℃ and 5% CO2 culture incubator for 1 hr. Cells were then fixed and permeabilized. Click-IT™ Plus reaction cocktail was added in cells. Cells were then stained with DAPI for DNA content and detected using a LSR Fortessa cytometer (BD Biosciences) and analyzed using FlowJo software version 10.8.1. For Annexin V staining, 2 x 106 CU-ACC1 or 1.5 x 106 H295R cells were seeded in 6 cm dish and treated with 20 µg/mL ADCT-701 or B12-PL1601 for 1 or 3 days respectively. Apoptosis was detected using an FITC Annexin V Apoptosis Detection Kit with PI (BioLegend #640914) following the manufacturer’s instruction. Briefly, cells were washed with cold PBS containing 1% BSA and 0.1% sodium azide and resuspended in Annexin V binding buffer and stained with Annexin V and PI at room temperature and then analyzed immediately. Annexin V-positive cells were detected using a LSR Fortessa cytometer (BD Biosciences) and analyzed using FlowJo software version 10.8.1.

Immunoblotting

Cells were lysed in RIPA buffer (Millipore 20-188) supplemented with protease inhibitor (Sigma-Aldrich #11836153001) and phosphatase inhibitors (Sigma-Aldrich #04906837001). Protein concentration was determined by the DC™ Protein Assay Reagents Package Kit (Bio- Rad #5000116). 20 µg of protein lysates were resolved on 4-15% Protein Gel (Bio-Rad #5671084) and transferred to nitrocellulose membrane. The membranes were blocked in 5% blotting grade blocker (Bio-Rad, 1706404XTU) in TBS with 0.1% Tween-20 and then incubated with the indicated primary antibodies. Primary antibodies (1:1000) included DLK1 (CST #2069), phospho- Histone H2A.X (Ser139) (Millipore 05-636), cleaved caspase-3 (Asp175) (CST #9661), cleaved PARP (Asp214) (CST #9541), total NOTCH1 (CST #3608), NOTCH1-ICD (CST #4147), SYP (CST #36406), and CYP17A1 (CST #17334). Primary antibody for detection of a-tubulin (Sigma-

604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628

Aldrich T9026) was used at a dilution of 1:1500. Secondary antibodies (1:5000) were from donkey anti-rabbit IgG-HRP (Cytiva NA934) and sheep anti-mouse IgG-HRP (Cytiva NA931).

In vitro bystander killing assays

Bystander activity was assessed by co-culturing WT CU-ACC1 and CU-ACC1 DLK1 KO (clone 10) cells at various ratios in white-walled 384-well tissue culture treated plates with complete media. The following day, cells were treated with 1 ug/mL ADCT-701 or B12-PL1601 and incubated in a humidified atmosphere with 5% CO2 at 37℃ for 4 days. Cell viability was measured by CellTiter-Glo Luminescent Cell Viability Assay kit (Promega). Bystander activity was also assessed using conditioned media assays in which CU-ACC1 cells were seeded at a density of 1500 cells/well. ADCT-701 was then added in triplicate the next day, in a dose titration ranging from 20 µg/mL to 20 pg/mL. Cells were subsequently incubated for 5 days. On day 5, 30 uL of conditioned media from these plates was removed from each well and transferred to a fresh plate containing CU-ACC1 DLK1 KO (clone 10) or WT CU-ACC1 cells, which were plated 24 hrs previously in 30 uL complete media (final volume 60 uL). These plates were incubated for 5 days before cell viability measurement. Cell viability was determined using the CellTiter-Glo Luminescent Cell Viability Assay kit and data were presented as percent cell viability relative to untreated controls.

Lentiviral constructs and lentivirus production

For the CRISPR-Cas9 system, a single target sequences for CRISPR interference were designed using the sgRNA designer (https://portals.broadinstitute.org/gppx/crispick/public) and subcloned into the lentiCas9-Blast (Addgene #52962). Viral transduction was performed in the presence of polybrene (5-10 µg/mL, Sigma-Aldrich TR-1003-G) and cells were centrifuged at 1200 x g for 4 hrs at 30 ℃ followed by removal of virus and polybrene. After 72 hrs, cells were selected with blasticidin (1-4 ug/mL) for 5 days.

629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654

siRNA-mediated knockdown of DLK1

DLK1-targeting siRNA (siDLK1-1: #4392420 ID s16740, siDLK1-2: #4392420 ID s16738, siDLK1-3: #4392420 ID s16739) and control siRNA (#4390843) were purchased from Invitrogen. CU-ACC1 (5 x 105 cells/well) cells were plated in 6-well plate overnight. Cells were then transfected with siRNA at a final concentration of 30 nM using Lipofectamine RNAiMAX (Invitrogen #13778150). Four days after transfection, whole cell lysates were collected and analyzed by Western blotting.

Enzyme-linked immunosorbent assay (ELISA)

Cortisol levels were quantified from conditioned media from CU-ACC1 parental and DLK1 KO cells using a cortisol ELISA kit (Enzo Life Science ADI-900-071). Conditioned media was collected from cells (1 x 106 cells/well) seeded in 12-well plate for 3 days.

Mice

All animal procedures reported in this study were approved by the NCI Animal Care and Use Committee (ACUC) and in accordance with federal regulatory requirements and standards. All components of the intramural NIH ACU program are accredited by AAALAC International. Seven-week-old female NSG mice were obtained from the CCR Animal Research Program. Seven-week-old female athymic nude mice were purchased from Jackson Labs. All mice were housed in accredited facilities on a 12 hrs light/dark cycle with free access to food and water under pathogen-free conditions.

In vivo efficacy studies

For ACC cell line-derived xenograft models, 2 x 106 CU-ACC1 or H295R cells were subcutaneously injected into the right flank of NSG mice. When the tumor volume reached approximately 100-150 mm3, mice were randomized to each group (3-4 mice per group). For ACC

655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680

PDX models, 164165 and 592788 were obtained from the NCI Patient-Derived Models Repository (PDMR) within the NCI Developmental Therapeutics Program. POBNCI_ACC004 PDX was developed by the NCI Pediatric Oncology Branch. 164165 and 592788 PDX tumor fragments were implanted subcutaneously into the right flank of NSG mice by using a trocar needle. 2 x 106 POBNCI_ACC004 PDX single cells with mouse cell depletion were injected subcutaneously into NSG mice. Recruitment of paired mice in equal numbers to treatment groups was staggered as necessary for any given study. Mice were randomized to each group (5-8 mice per group) once tumors reached 100-200 mm3. For SCLC cell line-derived xenograft models, 2 x 106 H524 or H1436 cells were implanted subcutaneously into the right flank of NSG mice. 2 x 106 H146 cells were injected subcutaneously into right flank of athymic nude mice. Tumor-bearing mice were randomized into three treatment groups (4-7 mice per group) once tumor volume reached 100- 150 mm3. All tumor cells used in vivo were suspended in 100 uL of PBS with 50% Matrigel (BD #356237). Normal saline, B12-PL-1601 (1 mg/kg, diluted in normal saline), and ADCT-701 (1 mg/kg, diluted in normal saline) were intravenously injected into the tail vein on day 0. PDX tumor- bearing mice that initially responded to ADCT-701 treatment were re-treated with ADCT-701 if re- growth tumor volume reached over 100 mm3 (otherwise no further doses were administered). Xenograft tumor-bearing mice that did not initially respond to ADCT-701 were re-treated at D7. Body weight and tumor size were measured once or twice weekly respectively, and tumor volume (mm3) were calculated with the formula as length x width2 x 0.5. Mice were euthanized when tumor volume reached 1500-2000 mm3 or 100 days after dosing. Tumors were collected for RNA- sequencing and IHC analysis.

Quantification and statistical analysis

All statistical tests between groups were unpaired two-tailed Student’s t-tests, unless otherwise stated, and p-values less than 0.05 were considered statistically significant. Survival analyses were conducted using Cox-proportional hazard models using the R survival package

681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706

(v3.1.7). Log-rank values were reported for survival analyses. For box plots, the horizontal line represents the median, the lower and upper boundaries correspond to the first and third quartiles and the lines extend up to 1.5 above or below the IQR (where IQR is the interquartile range, or distance between the first and third quartiles).

Acknowledgments

We also would like to thank the technicians in the CCR Animal Research Program for their support of this study.

Author contributions

Study conception and design: NYS and NR; Data collection and experiments: NYS; Analysis and interpretation of experiments: NYS and NR; Manuscript writing: NYS and NR. All authors reviewed the results and approved the final version of the manuscript.

Data and materials availability

Previously published expression datasets re-analyzed in this study can be accessed at GSE10927 and https://portal.gdc.cancer.gov/. Data from Jain et al.13 was obtained directly from the authors. Normalized RNA-seq data from the newly generated NCI ACC cohort reported in this study can be found in Supplementary Table 2.

707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732

bioRxiv preprint doi: https://doi.org/10.1101/2024.10.09.617077; this version posted October 11, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

733	References
734	1	Kunnimalaiyaan, M. & Chen, H. Tumor suppressor role of Notch-1 signaling in
735		neuroendocrine tumors. The oncologist 12, 535-542 (2007).
736		https://doi.org/10.1634/theoncologist.12-5-535
737	2	Zhou, B. et al. Notch signaling pathway: architecture, disease, and therapeutics. Signal
738		Transduction and Targeted Therapy 7, 95 (2022). https://doi.org/10.1038/s41392-022-
739		00934-y
740	3	Saunders, L. R. et al. A DLL3-targeted antibody-drug conjugate eradicates high-grade
741		pulmonary neuroendocrine tumor-initiating cells in vivo. Science translational medicine
742		7, 302ra136 (2015). https://doi.org/10.1126/scitranslmed.aac9459
743	4	Puca, L. et al. Delta-like protein 3 expression and therapeutic targeting in neuroendocrine
744		prostate cancer. Science translational medicine 11 (2019).
745		https://doi.org/10.1126/scitranslmed.aav0891
746	5	Paz-Ares, L. et al. Tarlatamab, a First-in-Class DLL3-Targeted Bispecific T-Cell
747		Engager, in Recurrent Small-Cell Lung Cancer: An Open-Label, Phase I Study. Journal
748		of clinical oncology : official journal of the American Society of Clinical Oncology 41,
749		2893-2903 (2023). https://doi.org/10.1200/JCO.22.02823
750	6	Ahn, M. J. et al. Tarlatamab for Patients with Previously Treated Small-Cell Lung
751		Cancer. The New England journal of medicine 389, 2063-2075 (2023).
752		https://doi.org/10.1056/NEJMoa2307980
753	7	Consortium, G. T. The Genotype-Tissue Expression (GTEx) project. Nat Genet 45, 580-
754		585 (2013). https://doi.org/10.1038/ng.2653
755	8	Pradat, Y. et al. Integrative Pan-Cancer Genomic and Transcriptomic Analyses of
756		Refractory Metastatic Cancer. Cancer discovery 13, 1116-1143 (2023).
757		https://doi.org/10.1158/2159-8290.CD-22-0966
758	9	Weiner, A. K. et al. A proteogenomic surfaceome study identifies DLK1 as an
759		immunotherapeutic target in neuroblastoma. bioRxiv, 2023.2012.2006.570390 (2024).
760		https://doi.org/10.1101/2023.12.06.570390
761	10	Chang, K. et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nature Genetics
762		45, 1113-1120 (2013). https://doi.org/10.1038/ng.2764
763	11	Fassnacht, M. et al. Adrenocortical carcinomas and malignant phaeochromocytomas:
764		ESMO-EURACAN Clinical Practice Guidelines for diagnosis, treatment and follow-up.
765		Annals of oncology : official journal of the European Society for Medical Oncology 31,
766		1476-1490 (2020). https://doi.org/10.1016/j.annonc.2020.08.2099
767	12	Fassnacht, M. & Allolio, B. Clinical management of adrenocortical carcinoma. Best
768		practice & research. Clinical endocrinology & metabolism 23, 273-289 (2009).
769		https://doi.org/10.1016/j.beem.2008.10.008
770	13	Jain, M. et al. TOP2A is overexpressed and is a therapeutic target for adrenocortical
771		carcinoma. Endocr Relat Cancer 20, 361-370 (2013). https://doi.org/10.1530/ERC-12-
772		0403
773	14	Zheng, S. et al. Comprehensive Pan-Genomic Characterization of Adrenocortical
774		Carcinoma. Cancer cell 29, 723-736 (2016). https://doi.org/10.1016/j.ccell.2016.04.002
775	15	Giordano, T. J. et al. Molecular classification and prognostication of adrenocortical
776		tumors by transcriptome profiling. Clinical cancer research : an official journal of the
777		American Association for Cancer Research 15, 668-676 (2009).
778		https://doi.org/10.1158/1078-0432.Ccr-08-1067

779	16	Hartley, J. A. The development of pyrrolobenzodiazepines as antitumour agents. Expert
780		Opin Investig Drugs 20, 733-744 (2011). https://doi.org/10.1517/13543784.2011.573477
781	17	Ghayee, H. K. et al. Progenitor cell line (hPheo1) derived from a human
782		pheochromocytoma tumor. PloS one 8, e65624 (2013).
783		https://doi.org/10.1371/journal.pone.0065624
784	18	Dumontet, C., Reichert, J. M., Senter, P. D., Lambert, J. M. & Beck, A. Antibody-drug
785		conjugates come of age in oncology. Nature reviews. Drug discovery 22, 641-661 (2023).
786		https://doi.org/10.1038/s41573-023-00709-2
787	19	Flynn, M. J. et al. ADCT-301, a Pyrrolobenzodiazepine (PBD) Dimer-Containing
788		Antibody-Drug Conjugate (ADC) Targeting CD25-Expressing Hematological
789		Malignancies. Molecular cancer therapeutics 15, 2709-2721 (2016).
790		https://doi.org/10.1158/1535-7163.Mct-16-0233
791	20	Wang, T. & Rainey, W. E. Human adrenocortical carcinoma cell lines. Mol Cell
792		Endocrinol 351, 58-65 (2012). https://doi.org/10.1016/j.mce.2011.08.041
793	21	Corbett, S. et al. The Role of Specific ATP-Binding Cassette Transporters in the
794		Acquired Resistance to Pyrrolobenzodiazepine Dimer-Containing Antibody-Drug
795		Conjugates. Molecular cancer therapeutics 19, 1856-1865 (2020).
796		https://doi.org/10.1158/1535-7163.MCT-20-0222
797	22	Hartley, J. A. et al. Pre-clinical pharmacology and mechanism of action of SG3199, the
798		pyrrolobenzodiazepine (PBD) dimer warhead component of antibody-drug conjugate
799		(ADC) payload tesirine. Scientific reports 8, 10479 (2018).
800		https://doi.org/10.1038/s41598-018-28533-4
801	23	Weiner, A. K. et al. A proteogenomic surfaceome study identifies DLK1 as an
802		immunotherapeutic target in neuroblastoma. bioRxiv, 2023.2012.2006.570390 (2023).
803		https://doi.org/10.1101/2023.12.06.570390
804	24	Baladron, V. et al. dlk acts as a negative regulator of Notch1 activation through
805		interactions with specific EGF-like repeats. Exp Cell Res 303, 343-359 (2005).
806		https://doi.org/10.1016/j.yexcr.2004.10.001
807	25	Shan, L., Aster, J. C., Sklar, J. & Sunday, M. E. Notch-1 regulates pulmonary
808		neuroendocrine cell differentiation in cell lines and in transgenic mice. American journal
809		of physiology. Lung cellular and molecular physiology 292, L500-509 (2007).
810		https://doi.org/10.1152/ajplung.00052.2006
811	26	Lim, J. S. et al. Intratumoural heterogeneity generated by Notch signalling promotes
812		small-cell lung cancer. Nature 545, 360-364 (2017). https://doi.org/10.1038/nature22323
813	27	Loganzo, F., Sung, M. & Gerber, H. P. Mechanisms of Resistance to Antibody-Drug
814		Conjugates. Molecular cancer therapeutics 15, 2825-2834 (2016).
815		https://doi.org/10.1158/1535-7163.Mct-16-0408
816	28	Flynn, S. D. et al. P-glycoprotein expression and multidrug resistance in adrenocortical
817		carcinoma. Surgery 112, 981-986 (1992).
818	29	Robey, R. W. et al. Revisiting the role of ABC transporters in multidrug-resistant cancer.
819		Nature reviews. Cancer 18, 452-464 (2018). https://doi.org/10.1038/s41568-018-0005-8
820	30	Ranganathan, P., Weaver, K. L. & Capobianco, A. J. Notch signalling in solid tumours: a
821		little bit of everything but not all the time. Nature Reviews Cancer 11, 338-351 (2011).
822		https://doi.org/10.1038/nrc3035

31 Laborda, J., Sausville, E. A., Hoffman, T. & Notario, V. dlk, a putative mammalian homeotic gene differentially expressed in small cell lung carcinoma and neuroendocrine tumor cell line. The Journal of biological chemistry 268, 3817-3820 (1993).

32 George, J. et al. Comprehensive genomic profiles of small cell lung cancer. Nature 524, 47-53 (2015). https://doi.org/10.1038/nature14664

33 Fassnacht, M. et al. Combination chemotherapy in advanced adrenocortical carcinoma. The New England journal of medicine 366, 2189-2197 (2012). https://doi.org/10.1056/NEJMoa1200966

34 Tierney, J. F. et al. National Treatment Practice for Adrenocortical Carcinoma: Have They Changed and Have We Made Any Progress? The Journal of clinical endocrinology and metabolism 104, 5948-5956 (2019). https://doi.org/10.1210/jc.2019-00915

35 Benetatos, L. & Hatzimichael, E. Delta-like homologue 1 and its role in the bone marrow niche and hematologic malignancies. Clin Lymphoma Myeloma Leuk 14, 451-455 (2014). https://doi.org/10.1016/j.clml.2014.06.019

36 Grassi, E. S. & Pietras, A. Emerging Roles of DLK1 in the Stem Cell Niche and Cancer Stemness. J Histochem Cytochem 70, 17-28 (2022). https://doi.org/10.1369/00221554211048951

37 Lee, W. K. & Frank, T. Teaching an old dog new tricks: reactivated developmental signaling pathways regulate ABCB1 and chemoresistance in cancer. Cancer Drug Resist 4, 424-452 (2021). https://doi.org/10.20517/cdr.2020.114

38 Strosberg, J. R. Update on the management of unusual neuroendocrine tumors: pheochromocytoma and paraganglioma, medullary thyroid cancer and adrenocortical carcinoma. Semin Oncol 40, 120-133 (2013). https://doi.org/10.1053/j.seminoncol.2012.11.009

39 Zhang, W. et al. Small cell lung cancer tumors and preclinical models display heterogeneity of neuroendocrine phenotypes. Translational lung cancer research 7, 32- 49 (2018). https://doi.org/10.21037/tlcr.2018.02.02

40 Komminoth, P., Roth, J., Schröder, S., Saremaslani, P. & Heitz, P. U. Overlapping expression of immunohistochemical markers and synaptophysin mRNA in pheochromocytomas and adrenocortical carcinomas. Implications for the differential diagnosis of adrenal gland tumors. Laboratory investigation; a journal of technical methods and pathology 72, 424-431 (1995).

Menefee, M. E. et al. Effects of the P-glycoprotein (Pgp) antagonist tariquidar (XR-9576; TQD) on Pgp function as well as the toxicity and efficacy of combined chemotherapy in patients with metastatic adrenocortical cancer (mACC). Journal of Clinical Oncology 26, 2543-2543 (2008). https://doi.org/10.1200/jco.2008.26.15_suppl.2543

42 Zhai, Y. et al. DLK1-directed chimeric antigen receptor T-cell therapy for hepatocellular carcinoma. Liver Int 42, 2524-2537 (2022). https://doi.org/10.1111/liv.15411

43 Brudno, J. N. & Kochenderfer, J. N. Current understanding and management of CAR T cell-associated toxicities. Nature Reviews Clinical Oncology 21, 501-521 (2024). https://doi.org/10.1038/s41571-024-00903-0

44 Hadjidemetriou, I. et al. DLK1/PREF1 marks a novel cell population in the human adrenal cortex. J Steroid Biochem Mol Biol 193, 105422 (2019). https://doi.org/10.1016/j.jsbmb.2019.105422

45 McDermott, M. S. et al. Abstract 1896: Therapeutic potential of TORL-4-500, an antibody-drug conjugate directed against delta like non-canonical notch ligand 1 (DLK1).

823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868

Cancer research 84, 1896-1896 (2024). https://doi.org/10.1158/1538-7445.Am2024- 1896

Begum, A., Kim, Y., Lin, Q. & Yun, Z. DLK1, delta-like 1 homolog (Drosophila), regulates tumor cell differentiation in vivo. Cancer Lett 318, 26-33 (2012). https://doi.org/10.1016/j.canlet.2011.11.032

Kim, Y., Lin, Q., Zelterman, D. & Yun, Z. Hypoxia-regulated delta-like 1 homologue enhances cancer cell stemness and tumorigenicity. Cancer research 69, 9271-9280 (2009). https://doi.org/10.1158/0008-5472.CAN-09-1605

Cai, C. M., Xiao, X., Wu, B. H., Wei, B. F. & Han, Z. G. Targeting endogenous DLK1 exerts antitumor effect on hepatocellular carcinoma through initiating cell differentiation. Oncotarget 7, 71466-71476 (2016). https://doi.org/10.18632/oncotarget.12214

869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917

Main Figure Legends

Figure 1. Identification of DLK1 as the most highly expressed Notch ligand in adrenocortical carcinoma. (A) DLK1 mRNA expression across adult refractory metastatic cancers (n=948). Tumor types with high DLK1 expression are highlighted in the colors shown. The percentage of each tumor type with high DLK1 expression is shown on the right. (B) DLK1 mRNA expression in the TCGA PanCancer dataset. (C) Expression of Notch ligands from four independent bulk ACC RNA-seq datasets. (D) Quantification of DLK1 IHC staining in ACC tumors with IHC images of four representation tumors with varying levels of DLK1 expression. Scale bars represent 200 µM. IHC: immunohistochemistry.

Figure 2. ADCT-701, a DLK1 targeting antibody-drug conjugate, has potent in vitro activity and induces robust in vivo anti-tumor responses in adrenocortical carcinoma. (A) Schematic structure of ADCT-701, a DLK1 targeting antibody drug conjugate. (B) Representative surface expression of DLK1 among ACC cell lines: CU-ACC1, CU-ACC2, and H295R. (C) ADCT- 701 cytotoxicity among CU-ACC1, CU-ACC2, and H295R cells. Cells were treated with ADCT- 701 and B12-PL1601 (non-targeted control ADC) for 7 days. Each point represents the meantSEM. (D) Representative imaging flow cytometry images and signal intensity analysis (n=3 biological replicates) showing cellular internalization of DLK1 antibodies in CU-ACC1, CU-ACC2, and H295R. (E) Cytotoxic activity of ADCT-701 responsive (n=6) and (F) non-responsive (n=6) ACC patient-derived organoids (PDOs). Flow cytometry histograms assessing DLK1 among ADCT-701 (G) responsive and (H) non-responsive ACC PDOs. Shaded gray histograms represent unstained controls for each condition. (I) CU-ACC1 and H295R xenograft tumor growth curves after treatment with saline, B12-PL1601, or ADCT-701 (1 mg/kg). Additional doses of ADCT-701 indicated by arrows. (J) ACC PDXs 164165, 592788, and POBNCI_ACC004 tumor growth curves after treatment with saline, B12-PL1601 or ADCT-701 (1 mg/kg). X symbols indicate the administration of ADCT-701 re-dosing. Arrow indicates unexpected death of 1 POBNCI_ACC004 tumor-bearing mouse prior to endpoint. DLK1 immunohistochemistry with H- scores shown above each individual xenograft or PDX tumor. Scale bars represent 200 uM.

947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968

Figure 3. ABCB1, a drug efflux protein, mediates intrinsic and acquired resistance to ADCT-701. (A) Cytotoxicity of SG3199 among ADCT-701 responsive and non-responsive DLK1+ ACC patient-derived organoids (PDOs). (B) Drug transporter mRNA expression in NCI-ACC patients as measured by RNA-seq. (C) Flow cytometry histograms assessing ABCB1 of DLK1+ ADCT-701 responsive (NCI-ACC51) and non-responsive (NCI-ACC48) ACC PDOs. (D) SG3199 cytotoxicity in the NCI-ACC48 PDO with and without treatment with ABCB1 inhibitors (1 uM valspodar, 10 uM elacridar and 1 uM tariquidar). (E) Flow cytometry histograms assessing ABCB1 among 164165, 592788 and POBNCI_ACC004 PDXs. (F) SG3199 cytotoxicity in 164165, 592788 and POBNCI_ACC004 PDX-derived organoids. (G) SG3199 cytotoxicity in the 164165 and 592788 PDX-derived organoids treated with or without ABCB1 inhibitors (1 uM valspodar, 10 uM elacridar and 1 uM tariquidar). (H) Volcano plot of differentially expressed genes in control tumors versus post-ADCT-701 acquired resistant tumors in POBNCI_ACC004 PDX. (I) Flow cytometry histograms assessing ABCB1 among ADCT-701 resistant and control POBNCI_ACC004 PDX tumors. (J) SG3199 cytotoxicity in the ADCT-701 resistant POBNCI_ACC004 PDX-derived organoid treated with or without ABCB1 inhibitors (1 uM valspodar, 10 uM elacridar and 1 uM tariquidar). (K) DLK1 molecules/cell relative to DLL3 among small cell lung cancer (SCLC) cell lines (H524, H146, and H1436). (L) Flow cytometry histograms assessing ABCB1 in SCLC cell lines. (M) SG3199 cytotoxicity in SCLC cell lines. Cells were treated with SG3199 for 3 days. (N) ADCT-701 cytotoxicity among SCLC cell lines. Cells were treated with ADCT-701 or B12-PL1601

918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946

for 7 days. Each point represents the meantSEM. (O) Tumor growth curves of SCLC xenograft models after treatment with saline, B12-PL1601 or ADCT-701 (1 mg/kg) treatment. Arrows indicate re-treatment with ADCT-701. Scale bars represent 200 uM. For flow cytometry histograms, shaded gray histograms represent isotype controls for each condition.

Figure 4. DLK1 is a major regulator of ABCB1, adrenocortical differentiation, and chemoresistance in adrenocortical carcinoma. (A) Immunoblot analysis of NOTCH1 signaling, total NOTCH1 and NOTCH1 intracellular domain (ICD), NE marker synaptophysin (SYP), and loading control (a-tubulin) proteins with and without DLK1 KO in CU-ACC1 cells. Two single-cell KO clones are shown. (B) Correlation between NOTCH1 and DLK1 expression among TCGA ACC tumors. (C) DLK1 and NOTCH1 expression in TCGA ACC tumors and normal adrenals. (D) SG3199 cytotoxicity in CU-ACC1 parental and DLK1 KO clones. (E) Flow cytometry histograms assessing ABCB1 in CU-ACC1 cells with and without DLK1 KO. (F) Concentration of cortisol in conditioned media from CU-ACC1 parental and DLK1 KO clones. (G) Immunoblot analysis of DLK1, total NOTCH1 and NOTCH1-ICD, SYP, the steroidogenic enzyme CYP17A1, and a-tubulin proteins in DLK1+ NCI-ACC48 and DLK1 ACC49 patient-derived organoids. (H) Flow cytometry histograms assessing ABCB1 in DLK1 negative NCI-ACC49 patient-derived organoids. (I) Immunoblot analysis of DLK1, total NOTCH1 (to detect the NOTCH1-ICD plasmid expression), SYP, CYP17A1, and a-tubulin proteins in CU-ACC1 cells with and without NOTCH1-ICD overexpression. (J) Flow cytometry histograms assessing ABCB1 in CU-ACC1 cells with and without N1ICD overexpression. (K) Single cell RNA-seq adrenocortical differentiation score comparing DLK1 high cells to DLK1 low or negative cells from 18 ACC metastatic tumors. (L) Model summarizing the findings of the current study. For flow cytometry histograms, shaded gray histograms represent isotype controls for each condition.

969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992

Relative mRNA expression

DLK1

DLL1

DLL3

DLL4.

(n=33)

(n=79)

ACC TCGA

Relative mRNA expression

log2(FPKM)

ம்

15.

10-

15.

DLK1

ACC Jain et al.

DLK1

DLL1

(n=16)

DLL1

DLL3.

DLL3

(n=50)

ACC NCI

DLL4

JAG1.

JAG1-

JAG2

JAG2-

(n=908)

expression

DLK1

Low

(n=40)

expression

DLK1

High

Adenocarcinoma, NOS

Hepatocellular carcinoma

Endometrioid carcinoma

Signet ring cell carcinoma

Sarcomas

Small cell lung cancer

Neuroendocrine tumor, grade II

Breast carcinoma

Germ cell tumor

Lung adenocarcinoma

Adenoid cystic carcinoma

Pheochromocytoma/Paraganglioma

Adrenocortical carcinoma

Tumor type

Other

<1%, n=1/477

8%, n=1/12

100%, n=1/1

9%, n=3/33

36%, n=4/11

20%, n=3/15

1%, n=1/99

50%, n=4/8

1%, n=4/384

4.5%, n=1/22

100%, n=3/3

92%, n=11/12

expression High DLK1

DLK1 H Score

100

200

300

ACC35

ACC12

ACC1

ACC22

ACC30

ACC33

ACC25

ACC9

ACC6

ACC20

ACC49

ACC2

ACC11

ACC NCI

ACC8

ACC42

ACC32

ACC37

ACC31

ACC45

ACC54

ACC34

ACC44

ACC40

ACC21

ACC46

ACC17

ACC41

ACC38

ACC26

H-score: 270

ACC21

H-score: 10

ACC22

H-score: 300

ACC26

H-score: 100

ACC42

DLK1 mRNA

Expression (RSEM)

Figure

200k

400k

600k

800k

ESCA

STAD

BLCA

CESC

COAD

DLBC

HNSC

KICH

KIRC KIRP-

LUAD

MESO-

SKCM-

PanCancer

THCA-

UCEC

UVM -

LUSC

PRAD

BRCA-

LGG

LAML

LIHC

CHOL

SARC

GBM-

THYM-

PAAD

TGCT

UCS

PCPG

ACC

DLK1 expression log2(TPM+1)

2.5

5.0

7.5

10.0

12.5

DLK1

(n=948)

metastatic tumors

Refractory

JAG2-

ACC Giordano et al.

DLL4.

JAG1-

JAG1

JAG2

log2(FPKM+1)

105 and is also made available for use under a CC0 license.

which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC bRxiv preprint doi: https://doi.org/10.1101/2024.10.09.617077; this version posted October 11, 2024. The copyright holder for this preprint

Figure

DRxiv preprint doi: https://doi.org/10.1101/2024.10.09.617077; this version posted October 11, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

ADCT-701

BCN

Val-Ala

PABC

= N-acetyl-D-glucosamine (GlcNAc)

= N-acetyl-D-galactosamine (GaINAc)

= L-fucose

SG3199

= PL1601

CU-ACC1

Cell Viability (%)

150-

100-

50.

Nº

-1

CU-ACC2

150

100

ADCT-701

B12-PL1601

-2

-1

Concentration (log nM)

CU-ACC1

CU-ACC2

NCI-H295R

Normized to mode

100

Unstained

Secondary only

40-

anti-DLK1

103

104

105

103 104 105

103

104

105

CU-ACC1

CU-ACC2

NCI-H295R

APC

Merged

APC Merged

APC

Merged

Cell internalized intensity (FIU)

2×104

APC-DLK1 antibody

APC-IgG

1.5×104

1×104

5×103.

CU-ACC1

CU-ACC2

NCI-H295R

PDOs with response to ADCT-701

200-

NCI-ACC-17

200-

NCI-ACC-44

200

NCI-ACC-47

150

100

100-

₮

100

Cell Viability (%)

-1

-3

-2

-1

-2

-1

NCI-ACC-51

200

200-

NCI-ACC-52

200

NCI-ACC-56

150

100

ADCT-701

B12-PL1601

-1

-4

-3

-2

-1

Concentration (log nM)

PDOs without response to ADCT-701

NCI-ACC-40

NCI-ACC-42

200-

200

150-

150

150-

100-

100

-1

-2

-1

-4

-1

NCI-ACC-49

NCI-ACC-54

200

150

100

ADCT-701

B12-PL1601

-3

-2

-1

-4

-1

Concentration (log nM)

PDOs with response to ADCT-701

NCI-ACC17

NCI-ACC44

NCI-ACC47

NCI-ACC51

NCI-ACC52

NCI-ACC56

103

104

105

anti-DLK1 PE

PDOs without response to ADCT-701

NCI-ACC40

NCI-ACC42

(recurrent)

NCI-ACC48

NCI-ACC49

NCI-ACC54

103

104

105

anti-DLK1 PE

592788 PDX

H-score: 300

2000

n=7

1500

n=7

n=6

1000

3/6 PR

500

100

POBNCI_ACC004 PDX

H-score: 140

2000

n=5

1500

n=5

Saline

1000

B12-PL1601

ADCT-701

500

1/5 CR

100

Days after treatment

CU-ACC1

H-score: 180

NCI-H295R H-score: 100

Tumor volume (mm3)

2000

Saline, n=3

Saline, n=4

1500

B12-PL1601, n=3

ADCT-701, n=3

1500

B12-PL1601, n=4

ADCT-701, n=3

1000

500

3/3 PR

2/3 PR

100

Days after treatment

164165 PDX H-score: 100

Tumor volume (mm3)

2000

1500

n=8

n=7

n=8

1000

500

0/8 PR or CR

100

Hydraspace™

Linker

OMe

MeQ

NCI-H295R

150

100

-4

-1

200-

NCI-ACC-42 (recurrent)

100-

50-

Cell Viability (%)

200-

NCI-ACC-48

150

100

-4

-3

-2

-1

Figure 3i BioRxiv preprint doi: https://doi.org/10.1101/2024.10.09.617077; this version posted October 11, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

DLK1+ ACC PDOS

NCI-ACC48 PDO

IC50 (nM)

No response to ADCT-701

· NCI-ACC48

NCI-ACC51

anti-ABCB1

SG3199

32.05

150

NCI-ACC51

Cell Viability (%)

200

SG3199+valspodar

0.81

PDO

150

SG3199+elacridar

0.76

SG3199 IC50 (nM)

log2(FPKM)

100-

NCI-ACC48

PDO

100

SG3199+tariquidar-

0.77

103

104

105

Response

to ADCT-701

-3

-2

-1

W-

Concentration (log nM)

NCI-ACC51

NCI-ACC56

NCI-ACC40

NCI-ACC48

NCI-ACC54

ABCB1

ABCC1

ABCG2

ABCC2

SLCO2B1

SLC22A3

SLC7A7

164165 PDXO

IC50 (nM)

592788 PDXO

anti-ABCB1

SG3199

IC50 (nM)

SG3199

2.71

164165

Cell Viability (%)

150

164165 PDXO

7.33

Cell Viability (%)

150

SG3199+valspodar

0.49

Cell Viability (%)

150-

PDX

592788 PDXO

7.19

SG3199+elacridar

POBNCI_ACC004

100-

0.17

100-

0.10

SG3199+tariquidar

0.82

100-

592788

PDXO

50.

PDX

POBNCI

ACC004

-3

-2

-1

PDX

Concentration (log nM)

102

103

104

POBNCI_ACC004

POBNCI_ACC004 resistant PDXO

Small cell lung cancer

Enriched in control tumors

Enriched in resistant tumors

POBNCI_ACC004

anti-ABCB1

IC50 (nM)

2000

DLK1

DLL3

SG3199

3.57

Control tumor

SG3199+valspodar

150-

0.45

SG3199+elacridar

0.58

1500

17.5

Resistant tumor

Cell Viability (%)

molecules/cell

100-

SG3199+tariquidar

0.55

1000

15.0

102

103

104

50.

500

12.5

-Log, P

10.0

-3

-1

Concentration (log nM)

H524

H146

H1436

H524

H146

H1436

7.5

ABCB1

5.0

SG3199

2.5

150-

IC50 (nM)

H524

H146

H1436

Cell Viability (%)

H524

0.19

Cell Viability (%)

150-

150

0.0

100-

H146

0.25

100-

100.

-7.5

5.0

-2.5

2.5

5.0

7.5

H1436 0.28

Log2 fold change

50.

ADCT-701

B12-PL1601

-3

-2

-1

anti-ABCB1

Concentration (log nM)

H524

H524 H-score: 210

H146

H1436 H-score: 120

H-score: 30

H146

H1436

103

104

105

Tumor volume (mm3)

2000

Saline, n=5

2000

Saline, n=5

2000

B12-PL1601, n=4

Saline, n=6

1500

B12-PL1601, n=5

+ ADCT-701, n=4

1500

ADCT-701, n=6

1500

B12-PL1601, n=7

+ ADCT-701, n=7

1000

500

3/6 CR

500

4/4 CR

3/6 PR

7/7 CR

100

IC50 (nM)

7.7

1.57

1.93

1.73

Days after treatment

Figuren4

bioRxiv preprint doi: https://doi.org/10.1101/2024.10.09.617077; this version posted October 11, 2024. The copyright holder for this preprint was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC 105 and is also made available for use under a CC0 license.

DLK1 KO clones

DLK1

NOTCH1

CU-ACC1

kDa

ACC TCGA

15-

p<0.0001

10.

p<0.0001

SG3199

150

CU-ACC1

IC50 (nM)

-150

Pearson r = - 0.40

NOTCH1

Parental

1.46

p=0.0003

10-

-100

Clone 9

DLK1 0.21

NOTCH1

log2(RPKM+1)

Cell Viability (%)

100-

Clone 10

0.18

NOTCH1-ICD

-100

SYP

TCGA GTEx ACC Normal (n=78) adrenal (n=145)

TCGA

GTEx

-2

-1

ACC

Normal

Concentration (log nM)

2.5

7.5

CYP17A1

DLK1 log2(RPKM+1)

(n=78)

adrenal (n=145)

a-tubulin

-50

NCI-ACC48

NCI-ACC49

Cortisol

kDa

anti-ABCB1

1000-

CU-ACC1 parental

DLK1

NCI-ACC49

100

Concentration (ng/ml)

-37

Normalized to mode

500-

NOTCH1

Isotype

DLK1 KO

CU-ACC1 clone 9

-100

anti-ABCB1

CU-ACC1 clone 10

NOTCH1-ICD

-100

CU-ACC1

CU-ACC1 clone 9 CU-ACC1 clone 10

SYP

102

103

104

102

103

104

CYP17A1

-50

a-tubulin

DLK1 KO

CU-ACC1

Parental

+N1ICD

ACC scRNA-seq (n=18 metastases)

Adrenocortical carcinoma

anti-ABCB1

kDa

CU-ACC1 parental

1.5

p<0.0001

ADSHi

DLK1

ADSLow

NOTCH1

100

CU-ACC1 +N1ICD

Adrenocortical Differentiation Score

1.0

NOTCH1

SYP

0.5

DLK1Hi

102

103

104

0.0

DLK1Low

CYP17A1

Partial Response

Complete Response

¥ ADC Payload

-0.5

a-tubulin

-1.0

Chemosensitivity

DLK1 high

DLK1 low

or negative

ABCB1

Acquired resistance

Quartz 4

Explorer

39416174

Identification of DLK1, a Notch ligand, as an immunotherapeutic target and regulator of tumor cell plasticity and chemoresistance in adrenocortical carcinoma

40 Abstract

Introduction

DLK1 has limited normal tissue expression and high expression in multiple metastatic cancers including adrenocortical carcinoma

ADCT-701, an antibody drug conjugate targeting DLK1, induces cytotoxicity in ACC through apoptosis and bystander killing

ADCT-701 has potent in vitro activity in DLK1+ ACC patient-derived organoids and induces

robust anti-tumor responses in ACC cell line-derived and patient-derived xenografts

ABCB1, a drug efflux protein, mediates intrinsic and acquired ADC and chemotherapy

resistance among DLK1+ ACC pre-clinical models

ADCT-701 elicits complete, durable responses in DLK1+ small cell lung cancer tumors without ABCB1 expression

DLK1 is a major regulator of ABCB1, adrenocortical differentiation, and chemoresistance in ACC

A first-in-human phase I clinical trial of ADCT-701 in patients with ACC and neuroendocrine neoplasms

Discussion

Materials and Methods

ACC patient tumor specimens

Bulk and single-cell RNA sequencing

Tumor cells isolation and organoid culture

In vitro short-term organoid culture cytotoxicity assays

IHC staining

Cell lines

Flow cytometric analysis

In vitro cell line cytotoxicity assays

Imaging flow cytometry

Cell cycle and apoptosis assays

Immunoblotting

In vitro bystander killing assays

Lentiviral constructs and lentivirus production

siRNA-mediated knockdown of DLK1

Enzyme-linked immunosorbent assay (ELISA)

Mice

In vivo efficacy studies

Quantification and statistical analysis

Acknowledgments

Author contributions

Data and materials availability

Main Figure Legends

Figuren4

Graph View

Table of Contents