39554059

Single Nuclei Sequencing Reveals Intratumoral Cellular Heterogeneity and Replication Stress in Adrenocortical Carcinoma

Liudmila V. Popova, PhD1 ;* Elizabeth A. R. Garfinkle, PhD2 ;* Daniel M. Chopyk, MD, PHD1 ;* Jaye B. Navarro2, Adithe Rivaldi2, Yaoling Shu1, MD, MS; Elena Lomonosova, PhD4, John E. Phay, MD1; Barbra S. Miller, MD1; Swati Sattuwar, MD3; Mary Mullen, MD4; Elaine R. Mardis, PhD2, 5; Katherine E. Miller, PhD2, 5; Priya H. Dedhia, MD, PhD1, 6,7+

1Division of Surgical Oncology, The Ohio State University and Arthur G. James Comprehensive Cancer Center, Columbus, Ohio, USA.

2The Steve and Cindy Rasmussen Institute for Genomic Medicine at Nationwide Children’s Hospital, Columbus, Ohio, USA.

3Department of Pathology, The Ohio State University, Columbus, Ohio, USA

4Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, Alvin J. Siteman Cancer Center, Washington University School of Medicine, St Louis, MO, USA.

5Department of Pediatrics, The Ohio State University College of Medicine, Columbus, OH, USA

6Translational Therapeutics Program, The Ohio State University and Arthur G. James Comprehensive Cancer Center, Columbus, Ohio, USA.

7Center for Cancer Engineering, The Ohio State University, Columbus, Ohio, USA.

*Authors contributed equally to this manuscript.

¡Correspondence

Priya Dedhia, MD, PhD Division of Surgical Oncology The Ohio State University and Arthur G. James Comprehensive Cancer Center Department of Surgery Division of Surgical Oncology N924 Doan Hall, 410 W. 10th Avenue Columbus, OH 43210

priya.dedhia@osumc.edu

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 41

bioRxiv preprint doi: https://doi.org/10.1101/2024.09.30.615695; this version posted October 28, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

ABSTRACT

Adrenocortical carcinoma (ACC) is a rare endocrine malignancy with a poor prognosis and 45 limited treatment options. Bulk genomic characterization of ACC has not yielded obvious therapeutic or immunotherapeutic targets, yet novel therapies are needed. We hypothesized that elucidating the intratumoral cellular heterogeneity by single nuclei RNA sequencing analyses would yield insights into potential therapeutic vulnerabilities of this disease. In addition to characterizing the immune cell and fibroblast landscape, our analyses of single nuclei gene expression profiles identified an adrenal cortex cell cluster exhibiting a program of replication stress and DNA damage response in primary and metastatic ACC. In vitro assessment of replication stress and DNA damage response using an ACC cell line and a series of newly-derived hormonally active patient-derived tumor organoids revealed ATR sensitivity. These findings provide novel mechanistic insight into ACC biology and suggest that an underlying dependency on ATR may be leveraged therapeutically in advanced ACC.

149 / 150 words

42 43 44 46 47 48 49 50 51 52 53 54 55 56 57 58

INTRODUCTION

Adrenocortical carcinoma (ACC) is a rare, aggressive malignancy that originates in the adrenal cortex and has a 5-year survival of ~6% for patients with stage IV disease.1 While margin negative surgical resection is the most desirable frontline treatment, many patients present with metastatic disease necessitating treatment with systemic agents. In those able to undergo resection, most experience tumor recurrence.2-5 Mitotane, an adrenolytic agent that also blocks cortisol synthesis, remains the only FDA-approved therapeutic agent specifically for treatment of advanced ACC. However, attaining therapeutic levels of mitotane is challenging due to significant toxicity and harmful side effects.2,4 Administration of chemotherapeutic drugs - etoposide, doxorubicin, and cisplatin - in conjunction with mitotane improve outcomes, but combination therapy increases median progression-free survival by only 5 months and adds significant toxicity.2,6 Immunotherapy has shown promise in certain cancers, yet its efficacy in ACC remains limited.7 Thus, there is a critical need to understand molecular mechanisms driving ACC tumor progression to improve therapy.

Bulk sequencing strategies have allowed for comprehensive genomic characterization of ACC. Previous studies using whole exome sequencing, bulk mRNA sequencing, micro RNA sequencing, DNA copy number analysis, and DNA methylation analysis have enabled molecular classifications of ACC that correlate with distinct clinical outcomes.8,9 Despite significant advances in understanding the genomic landscape of ACC, effective treatment options are still lacking. Single-cell RNA sequencing offers a powerful, complementary technique to bulk sequencing strategies by profiling individual cells, and thus enabling analysis of intratumoral cellular heterogeneity, a key factor that contributes to poor outcomes in many other cancers.10-13 The rationale for this approach in ACC includes evidence of heterogeneity in nuclear localization of B-catenin in ACC14 and identification of ACC tumor cell subpopulations with unique metabolomic signatures (e.g. increased pentose phosphate pathway activity) using spatial metabolomics.15 In addition, recent work demonstrated intratumoral heterogeneity in ACC at single-nuclei resolution.16 These observations suggest important cellular differences in ACC and identify a prospective cell of origin; however, they have not identified new therapeutic targets. Furthermore, prior single nuclei RNA sequencing (snRNAseq) of ACC consisted primarily of early-stage ACC, which is often amenable to surgical resection. Thus, we hypothesized that elucidating the underlying cellular biology at the single cell level in advanced ACC has the potential to identify novel therapeutic

59 60 61 62 63 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 89 90 91 92

vulnerabilities in this disease. In addition, prior snRNAseq of ACC demonstrated depletion of fibroblasts and immune cells compared to normal adrenal tissue but did not fully characterize these populations to identify potential therapeutic targets.16 Our snRNAseq analyses of advanced ACC patient specimens further define the immune landscape and discover a population of fibroblasts associated with ACC. We also identify distinct population of adrenal cortex cells in ACC primary and metastatic samples, characterized by a unique transcriptomic program of proliferation, replication stress, and DNA damage response. Incorporating an ACC cell line and our novel ACC patient-derived tumor organoids (PTOs) in functional studies, we uncover ataxia telangiectasia and Rad3-related protein (ATR) dependency, a promising therapeutic strategy for this deadly disease.

RESULTS

snRNAseq reveals heterogeneous subpopulations in ACC

To profile cellular subpopulations in normal adrenal, adrenal adenoma, and ACC tissues, we performed snRNAseq on 11 specimens collected from 9 patients. The samples included normal adrenal tissue (n = 1), adrenal adenoma (AA, n = 2), primary ACC (PACC, n = 4), and metastatic ACC (MACC, n = 4, Figures 1A, 1B, and Table 1). After standard preprocessing steps, 4,735 nuclei from normal adrenal, 14,481 nuclei from AA, 23,837 nuclei from PACC samples, and 19,151 nuclei from MACC samples were used for analysis. Samples were integrated using the RPCA algorithm in Seurat and unbiased clustering resulted in 10 distinct cell populations (Figure 1C). To define adrenal cortex cells, we calculated average module scores (AC module scores) based on expression of adrenal cortex and ACC biomarkers, NR5A1 encoding the protein SF1, INHA, and MLANA.17-19 This analysis identified 6 adrenal cortex clusters with high AC module scores and 4 clusters with low AC module scores. The 4 clusters with low AC module scores were identified as lymphocytes (PTPRC)20, myeloid cells (ITGAM)21, fibroblasts (SERPINE1 and IGFBP7)22,23, and endothelial cells (VWF and FLT1).24,25 In addition to increased AC module scores, adrenal cortex clusters lacked expression of PTPRC, ITGAM, SERPINE1, IGFBP7, VWF, or FLT1 (Figure 1D and Supplementary Figure S1). We also calculated cell type proportions and determined that normal adrenal samples were characterized by the presence of 7 different cell clusters, AA samples were characterized by the presence of 9 different cell clusters, and both PACC and

93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126

MACC samples were defined by the presence of 10 different cell clusters. (Figure 1E). Overall, these data suggest that the cellular biology of ACC is defined by unique clusters when compared to normal adrenal or adrenal adenoma.

ACC exhibits a unique tumor immune microenvironment

To characterize the immune landscape in adrenal tissues, we subset the lymphocyte and myeloid cell clusters (Figure 1), reprocessed, integrated using Harmony26, and performed unbiased reclustering. Eight immune cell types were identified including macrophages, granulocytes, regulatory T cells, CD4+ and CD8+ T cells, NK cells, B cells, and plasma cells (Figure 2A). Across all specimens, macrophages were the most common immune cell (Figure 2B, Supplementary Tables S1A-2A). Key immunocyte-specific genes were plotted in a dot plot and used to guide immune cell typing (Figure 2C).

We first focused on the macrophage cluster, as these were the dominant immune population. 1,661 total macrophages were subset, reprocessed, and reclustered. From this subset, an additional 51 lymphocytes were identified and removed leaving 1,610 total macrophages for further analysis (Figure 2D). We assessed macrophage polarity by expression of proinflammatory or M1 markers, CD80, CD86, and FCGR1A (also known as CD64),27,28 and 146 immunosuppressive or M2 markers, MRC1, CD163, and MARCO29-31. Although we noted no significant differences in the proportion of M1 macrophages across all tissues, we found a significant decrease in M2 macrophages in PACC and MACC specimens compared to normal and AA specimens (false discovery rate (FDR) <0.05, Figure 2E). Prior work correlated increased M2 macrophage infiltration with worse prognosis in other solid tumors; however, more recent evidence indicates plasticity in M1 and M2 polarization.32 Furthermore, macrophage expression of SPP1 was strongly associated with clinical prognosis independent of classic M1 versus M2 macrophage markers.33,34 Additionally, SPP1+ macrophages have been associated with an immunosuppressive phenotype.32,35,36 Therefore, we assessed SPP1 expression and found a significant increase in SPP1+ macrophages (FDR = 0.03) in PACC and MACC specimens compared to normal and AA (Figure 2E, Supplementary Tables S1B- 2B). Our findings suggest that SPP1+ macrophages are associated with ACC tumor progression.

147 148 149 150 151 152 153 154 155 156 157 158 159

127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145

We next analyzed the expression of immune checkpoint molecules expressed by macrophages to identify potential strategies for immunotherapy. We plotted the average expression of pro-inflammatory (CD40 and CD80) and immunosuppressive (CD274 or PDL1, LGALS9, PVR, SIRPA, and CD80) markers (Figure 2G).37-44 CD80 has a dual role as either a costimulatory or coinhibitory ligand depending on whether it binds CD28 or CTLA-4, respectively, on neighboring T cells.44 We found that PACC-associated macrophages had higher average expression of CD274, and LGALS9, and MACC-associated macrophages had higher average expression of PVR, and SIRPA. PACC-associated macrophages also exhibited elevated expression of CD40 and CD80 relative to the benign tissues, suggesting these markers may serve as potential targets for immunotherapy. Taken together, our data suggest a changing macrophage landscape from normal adrenal to ACC that increasingly presents an immunosuppressive environment.

Analysis of lymphoid populations demonstrated that both T and NK cells were reduced or absent in most ACC specimens. Interestingly, PACC 4, a tumor from a patient without clinical or biochemical evidence of excess cortisol secretion, had the highest numbers of T and NK cells compared to other samples (Supplementary Table S2A). This finding is consistent with the association between hypercortisolism and reduced T cell infiltration in ACC (Table 1 and Figure 1B).8,45 Therefore, our analysis of CD8+, CD4+, and NK cells were primarily represented by PACC 4. First, we subset CD8+ T cells and plotted average expression of key immunoregulatory molecules (Supplementary Figure S2). Our analysis revealed that CD8+ T cells in the ACC samples compared to those in benign tissue had reduced expression of markers associated with immune activating functions (CD28 and CD40LG) and increased expression of markers of T cell exhaustion (CTLA-4, PDCD1, TIGIT, LAG3, and HAVCR2).46- 48 We next assessed expression of these markers within the CD4+ T cell subset and found that expression of CD28, CTLA-4, TIGIT, and LAG3 was increased whereas expression of CD40LG, PDCD1, and HAVCR2 was decreased in ACC relative to normal adrenal and AA (Supplementary Figure S2). Taken together, our analyses demonstrate signs of T cell exhaustion in ACC which was more pronounced in CD8+ compared to CD4+ T cells. Finally, analysis of the NK cells revealed increased expression of TIGIT49 and the NK inhibitory receptor KLRC150 in ACC relative to normal adrenal and AA specimens (Supplementary Figure S2). Collectively, these findings indicate that hypercortisolism in ACC reduces T cell infiltration, and thus ACC is characterized by an immunosuppressive microenvironment, which

160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192

may require precision therapeutic strategies that enhance immune infiltration in combination with immunotherapy to achieve efficacy.

ACC is associated with a unique fibroblast subpopulation

We next subset fibroblasts to determine if malignant specimens are associated with changes in fibroblast subpopulations. Initially we identified 10 fibroblast clusters. Four clusters including adipocytes were excluded from further analysis, and the remaining fibroblast data were reprocessed and reintegrated to identify 5 fibroblast clusters (Figure 3A-B). These clusters were annotated based on marker gene expression (Figure 3C) and fibroblast markers (Figure 3D).51,52 Fibroblast cluster 3.1 was present in all specimens and was the predominant cluster in PACC and MACC (Figure 3B). Characterized by high expression of ACTA2, MMP11, and HAS2, cluster 3.1 likely represents myofibroblastic cancer-associated fibroblasts (myCAFs).53,54 Further analysis of fibroblast cluster 3.1 in PACC specimens for known cancer- associated fibroblasts markers55 demonstrated increased levels of PDGFRB, FAP, VIM, POSTN, HAS2, CXCL8, CXCL2, LMNA, and HAS 1. MACC specimens expressed increased CAV1, CFD, MMP11, COL1A1, which are additional markers of cancer-associated fibroblasts (Figure 3E). These findings suggest that ACC is characterized by the expansion of a distinct fibroblast subpopulation at both primary and metastatic sites.

Adrenal cortex clusters demonstrate unique transcriptional programs

To further characterize the 6 adrenal cortex clusters, we subset, reintegrated, and performed unbiased clustering of these data, which resulted in identification of 7 adrenal cortex clusters and one adrenal medulla cluster (defined by CHGA expression, Supplementary Figure S3).56 Adrenal medulla cells were predominantly found in normal adrenal tissue (150 cells). After excluding adrenal medulla cells, we reintegrated and reclustered adrenal cortex cells to reveal 7 final adrenal cortex clusters (AC 3.1-3.7) based on marker gene expression (Figures 4A and B). The majority of the cells in cluster 3.7 originated from PACC 3, the only specimen with excess DHEAS, androstenedione, and testosterone secretion. Consistent with these clinical findings, AC 3.7 expressed high levels of CYP11A1 and CYP17A1 which are involved in testosterone, DHEAS, and androstenedione production. (Figure 4C, Table 1, and Supplementary Figure S4).57,58 We then assessed subpopulation frequency across different sample types and found that clusters AC 3.5 and AC 3.6 were significantly enriched in PACC and MACC specimens (FDR < 0.0001, Figure 4D). In order to identify gene expression

193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226

changes associated with tumor progression, we calculated average module scores for each of the clusters for Hallmark pathways downloaded from MSigDB59,60 (Figure 4E and Supplementary Figure S5). In these clusters, in which cell cycle regression was performed, we found that cluster AC 3.6 was characterized by the highest module scores for the following hallmark pathways: E2F targets, G2M checkpoint, mitotic spindle, PI3K AKT mTOR signaling, and DNA repair. Hence, AC 3.6 bears a gene expression signature of dysfunctional DNA replication and repair, indicating potential mechanistic underpinnings of malignancy and tumor progression. Cluster AC 3.5 was defined by TGF-ß signaling, unfolded protein response, and glycolysis. The unfolded protein response and glycolysis signatures in cluster 3.5 were enriched only in MACC, whereas the DNA damage repair signature in cluster AC 3.6 was present in all PACC and MACC samples (Supplementary Figure S5). We also found that cluster AC 3.3 was defined by the Wnt/ß catenin signaling pathway, which is commonly dysregulated in ACC.2 Cluster AC 3.2 was defined by cholesterol homeostasis, indicating a potential role in steroid metabolism.

Replication stress and DNA damage repair signature is enriched in cluster AC 3.6

To better understand the role of replication stress and DNA damage repair pathways in ACC, we used REACTOME pathways downloaded from the MSigDB database to generate module scores for each of the identified clusters.59,61 We observed that cluster AC 3.6 in PACC and MACC was characterized by the highest module scores associated with activation of ATR in response to replication stress, homology-directed repair, non-homologous end joining, mismatch repair, and base excision repair (Figure 5A). Additionally, we observed that replication stress and DNA repair pathways were enriched in PACC and MACC specimens (Supplementary Figure S6). We next used QIAGEN Ingenuity Pathway Analysis (IPA)62 to compare cluster AC 3.6 to other clusters found in PACC and MACC specimens and again identified that cell cycle checkpoints and homology-directed repair through homologous recombination or single-strand annealing were among the top 10 upregulated pathways (Figure 5B). Replication stress and chromosome instability have been previously associated with tumorigenesis and tumor progression.63 To elucidate genomic instability in AC clusters, we used inferCNV64 to identify copy-number variation (CNV) in tumor specimens on a per- cluster basis. We found that cluster AC 3.6 had the highest total number of CNVs, including both copy number gains and losses compared to other clusters (Figure 5C-E). This observation suggests that cluster AC 3.6, which demonstrates dysregulated replication stress

227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260

and DNA damage repair response as well as genomic instability, may reflect the underlying biology driving ACC tumor development and progression.

ATR-dependent replication stress in ACC

To assess for the presence of replication stress and DNA damage in ACC, we performed immunofluorescence on patient specimens (MACC 2 and MACC 3) for the DNA damage marker, [H2AX and the replication stress marker phosphorylated RPA. Both proteins localize to form foci, or clusters, in regions of DNA damage or replication stress.65 Both ACC specimens exhibited [H2AX and RPA foci (Figure 5F, Supplementary Figure S7, Supplementary Table S3). We next sought to evaluate DNA damage in the human NCI- H295R ACC cell line to establish an in vitro model for further mechanistic studies and found [H2AX foci indicating a cellular response to DNA damage in the absence of cytotoxic treatments (Figure 6A, Supplementary Figure S8).

To assess for potential therapeutic vulnerabilities within the DNA damage and replication stress response in ACC, we assessed sensitivity of NCI-H295R to small molecule inhibitors of ATR kinase, elimusertib and VX-803, and of ATM kinase, AZ32 and KU60019. NCI-H295R cells were sensitive to both ATR inhibitors with IC50 values ranging from 23 to 42 nM (Figure 6B). In contrast, ATM inhibition with AZ32 and KU60019 had minimal cytotoxic effects, even at concentrations of 10,000 nM (Figure 6C). Prior studies have identified IC50 values ranging from 100 to 300 nM and 500 to 20,000 nM for ATR monotherapy and ATM monotherapy respectively in cell lines for other solid tumors.66-69 ATR is recruited to sites of replication stress and enacts downstream signaling via phosphorylation of specific targets to arrest the cell cycle and ensure repair of damaged replication forks.70 Therefore, we next assessed the response of NCI-H295R cells to inhibition of enzymes in the DNA repair response downstream of ATR, including checkpoint kinase 1 (CHK1, a direct target of ATR)70 and Rad51, a key regulator of homologous recombination that is not a direct ATR target.71 Inhibition of CHK1 with rabusertib or PF47736 resulted in a modest effect on viability with IC50 values between 1000 to 10,000 nM (Figure 6D), which is higher than previously reported IC50 values in cell lines considered sensitive to CHK1 monotherapy.69,72,73 Inhibitors of Rad51 including RI-1 (specifically inhibits Rad51) and SAHA (HDAC inhibitor which inhibits Rad51), had minimal effect on viability except at highest concentrations (50000 nM for RI-1 and 10000 nM for SAHA, Figure 6E). Finally, inhibition of poly-ADP-ribose polymerase (PARP), a DNA

261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294

repair protein critical for single strand break repair and base excision repair pathways, with olaparib had no effect on NCI-H295R viability (Figure 6F). Prior studies have identified IC50 values for olaparib ranging from 100 to 20000 nM in sensitive cell lines.74,75 Together, these data indicate the presence of replication stress and specific sensitivity to ATR inhibition in an ACC cell line.

PTOs are 3-D cultures comprised of multiple cell types that can recapitulate the heterogenetic nature of the intratumoral environment and can model many aspects of tumor complexity in vitro.76,77 To further assess the role of replication stress in ACC tumorigenesis, we generated hormonally active ACC PTOs from patient 5 (Table 1, OSU1) and a second ACC patient with biochemical evidence of hypercortisolism (OSU2). Neither patient received neoadjuvant cytotoxic therapies prior to surgery. Both ACC PTO lines proliferated long-term in culture, expressed high levels of the adrenal cortex marker SF1, and produced cortisol (Figures 7A- C). Similar to the NCI-H295R cell line, both untreated ACC PTO lines demonstrated H2AX foci (Figure 7B, Supplementary figure S9). To determine if ACC PTOs exhibit sensitivity to ATR inhibition, we treated the ACC PTOs with elimusertib and VX-803, which reduced viability at IC50 values of 139 to 326 nM and 2000 to 3000 nM, respectively (Figure 7D). We also treated the ACC PTOs with the ATM inhibitors, AZ32 and KU60019, neither of which caused any appreciable reduction in viability (Figure 7D). These findings reinforce a critical role of ATR in ACC and further indicate a potential role for ATR inhibition in treatment of ACC.

DISCUSSION

In this study, we use snRNA sequencing to demonstrate changes in immune, fibroblast and adrenal cortex landscape in PACC and MACC specimens relative to non-malignant adrenal cortex tissues. As such, we have identified several potential therapeutic approaches for further investigation in ACC clinical management, as outlined below. Immunotherapy options for ACC have been characterized by poor responses.7 Our data indicate that macrophages are the dominant immune cell and that SPP1+ macrophages are increased in ACC. Recent studies demonstrate that SPP1+ macrophages are associated with tumor progression and worse clinical outcomes.32,33,36,78,79 In addition, ACC macrophages expressed markers of immunosuppression. Collectively, these data suggest targeting macrophages with antibodies or small molecules, such as blockade of the CD47-SIRPa axis or anti-PDL1 antibodies38, to

295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328

induce activation and tumor cell phagocytosis may be a viable immunotherapy strategy in ACC. In contrast, we found lymphoid infiltration only in the absence of excess cortisol production, suggesting that immunotherapy alone may be more effective in non-functional ACCs and that concurrent cortisol blockade may improve effectiveness of immunotherapy in the presence of hypercortisolism.

Interestingly, our analysis of fibroblasts showed a loss of fibroblast subpopulations in ACC. In addition, the same fibroblast population was present in primary and metastatic ACC. These findings suggest that this fibroblast subpopulation may migrate from the primary tumor to the site of metastasis and serve as a premetastatic niche to support tumor cells as in other cancers.80 Fibroblasts in ACC expressed increased FAP and PDGFRB. Targeting FAP and PDGFRB resulted in tumor reduction in preclinical studies of other cancers.81-83 Further studies are necessary to elucidate the role of fibroblasts in ACC tumor progression and identify potential therapeutic targets.

We found that ACC was characterized by a unique adrenal cortex subpopulation defined by a signature for replication stress and DNA damage response pathways. This subpopulation may correspond to the ACC M subpopulation identified by Tourigny et al.16 Although Tourigny et al. identified a signature for mitosis in the ACC M subpopulation, they did not identify a signature for replication stress or DNA repair. Our extended findings may be attributed to differences in sequencing methodologies and patient population. While Tourigny et al.16 predominantly included early-stage tumors, which can provide insight into the developmental origin of ACC, our study included stage 4 tumors, which may provide insights for developing targeted therapies. Other studies have established that genes implicated in DNA damage and cell cycle regulation are overexpressed in ACC and are associated with worse overall survival in patients.8,84 In addition, Gara et al. have shown that APOBEC3B, which is known to induce DNA damage, is overexpressed in ACC tumors with TP53 mutations.85 Prior work has also shown that metastatic ACC exhibits a higher mutation burden compared to primary tumors,86 indicative of genomic instability. In addition to identifying replication stress signature using snRNAseq data, our work mechanistically evaluated these pathways in ACC.

Cancer cells experience chronic replication stress due to oncogene activation and sustained proliferative signaling. Furthermore, many chemotherapeutic agents induce replication stress by depleting nucleoside pools required for DNA replication or causing direct DNA damage in

329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362

tumor cells.88 ATR is a primary upstream orchestrator of the replications stress response pathway. By inducing G2-S phase cell arrest and DNA repair as well as replication fork stabilization, ATR allows cells to adapt to and prevent replication-stress induced cell death.70,89,90

Our findings suggest that ATR can be therapeutically targeted in ACC irrespective of cortisol secretion. ATR inhibitors drive cells with increased DNA damage to prematurely enter mitosis, resulting in cell death. In addition to preclinical studies, some ATR inhibitors have also been investigated in clinical trials as single agents or in conjunction with other therapeutic modalities, such as radiotherapy or chemotherapy. In advanced solid tumors, ATR inhibitors as monotherapy are well-tolerated and have resulted in partial or complete response. 93,9 93,94 Other preclinical studies and clinical trials show ATR inhibitors and chemotherapeutic agents can work synergistically to show efficacy in triple negative breast cancer, melanoma, platinum- resistant high-grade serous ovarian cancer, and other advanced solid tumors.91,94-99 Thus, by demonstrating ATR-dependent replication stress in ACC, we identify a promising novel treatment strategy for this deadly disease.

Identifying biomarkers to predict ATR sensitivity in ACC can improve potential efficacy. In non- small cell lung cancer (NSCLC), loss of BRG1, a chromatin remodeling enzyme, was identified as a biomarker for efficacy of ATR inhibition.100 Loss of ARID1A, genomic instability, ATM/G1 pathway abnormalities, and high tumor inflammation have also been identified as potential biomarkers of sensitivity to ATR inhibition.92 Although some preclinical models, including chronic lymphocytic leukemia, have demonstrated ATR sensitivity in conjunction with TP53 mutation, this relationship was not seen in the aforementioned clinical trial of ATR inhibition in solid tumors. 102-104 Importantly, TP53 inactivating mutations are present in 21-33% of ACC,8,105 so future work elucidating the role of p53 inactivation or other biomarkers in the context of ATR sensitivity are necessary.

CHK1 is a target of ATR that plays a prominent role in the ATR signaling pathway by activating checkpoints, delaying cell cycle progression, and enabling DNA repair to proceed.89,106 Our data show that the NCI-H295R cell line and ACC PTOs exhibit notable sensitivity to ATR inhibition and modest sensitivity to CHK1 inhibition. These data suggest that ATR-dependence in ACC may in part occur via alternative downstream targets, such as SMARCAL1, which functions to promote replication fork stability and fork restart.107 Additional

363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396

bioRxiv preprint doi: https://doi.org/10.1101/2024.09.30.615695; this version posted October 28, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

non-CHK1 targets of ATR, which are implicated in the replication stress response, include WRN and FANCI. 108,109 Further studies are needed to determine if these targets play a role in ACC tumor progression.

Taken together, our results demonstrate diversity of transcriptional signatures in ACC and implicate several possible strategies for targeted treatment of ACC. Moreover, our results exemplify how elucidation of cellular heterogeneity combined with mechanistic evaluation in PTOs might direct therapeutic decisions and improve patient outcomes in ACC and other rare cancers.

397 398 399 400 401 402 403 404 405 406

METHODS

Sample collection

Patients were consented for collection of specimens and clinical information using a study protocol approved by The Ohio State University Institutional Review Board. Normal adrenal and adrenal adenoma (AA) specimens were collected from patients diagnosed with adrenal hypercortisolism and included one normal adrenal-adrenal adenoma pair. Primary (PACC) and metastatic ACC (MACC) samples were collected from patients diagnosed with ACC and included one tumor-metastasis pair. After collection, samples were frozen for further processing. We performed snRNAseq as our samples included previously banked frozen samples.110 Somatic and germline mutations are reported for patients who underwent genetic testing using Tempus (Chicago, IL), Foundation Medicine (Boston, MA), or CustomNext- Cancer testing from Ambry Genetics (Aliso Viejo, CA). Sample details are summarized in Table 1.

Single nuclei RNA sequencing

Nuclei isolation and library preparation

Nuclei were isolated from frozen tissue as described in Lacar et al. and sorted using the Bigfoot Spectral Cell Sorter (Thermo Fisher Scientific, Waltham, MA). For each sample, 16,000 nuclei were sorted directly into the 10X Genomics master mix to load approximately 10,000 nuclei into a 10x Genomics Chromium device for microfluidic-based capture of single nuclei.111 Reverse transcription, cDNA amplification and library preparation were performed according to the manufacturer’s protocol for Chromium Next GEM Single Cell 3’ Gene Expression kit. Resulting libraries were sequenced on an Illumina NovaSeq 6000 instrument to generate paired-end sequencing data.

Data preprocessing

Fastq files were generated using the mkfastq command from 10x Genomic CellRanger. Alignment to the GRCh38 human reference, filtering, barcode counting, and UMI counting steps were performed using the 10x Genomics CellRanger software suite v7.0 following the default parameters.

407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440

Quality control, filtering, and doublet removal

Downstream analyses were performed using Seurat v4 for R 4,1,1 and 4.2.2.19 For quality control purposes, individual objects were filtered to remove nuclei that contained more than 5% of mitochondrial transcripts. In addition, data were filtered to only include nuclei that contained a minimum of 200 detected genes for all the samples and a maximum of 8000 genes (for Normal and AA 1 samples), a maximum of 7000 genes (for AA 2 sample), a maximum of 13000 genes (for PACC 1 sample), a maximum of 12000 genes (for PACC 2 sample), a maximum of 9000 genes (for PACC 3, PACC 4, and MACC 4 samples), a maximum of 15000 genes (for MACC 1 sample), a maximum of 11500 genes (for MACC 2 sample), and a maximum of 14000 (for MACC 3 sample). Upper filtering cutoffs were selected based on differences in data distribution between samples. As a result of filtering, the following numbers of nuclei were excluded from further analysis: 17 nuclei (Normal sample), 49 nuclei (AA 1 and MACC 1 samples), 29 nuclei (AA 2 sample), 70 nuclei (PACC 1 sample), 381 nuclei (PACC 2 sample), 328 nuclei (PACC 3) sample, 171 nuclei (PACC 4 sample), 31 nuclei (MACC 2 sample), 553 nuclei (MACC 3 sample), and 5 nuclei (MACC 4 sample). Doublets were detected and removed from individual objects using the DoubletFinder package v2.0 for R.112

Normalization and cell cycle regression

To control for cell-cycle effects on transcriptional programs, we performed cell cycle regression. After DoubletFinder, cell cycle scores were calculated for the individual objects using the CellCycleScoring() function based on the list of cell cycle marker genes published in Tirosh et al.113 Normalization and variance stabilization were performed on the individual objects using the SCTransform() function from the sctransform package v0.3 for R with “vst.flavor” argument set to v2 and “vars.to.regress” argument set to regress the percentage of mitochondrial content, G2/M and S cell cycle scores. 114

Data integration

To minimize patient-to-patient variability, samples were integrated using the RPCA algorithm in Seurat. 115

Dimensionality reduction and clustering

Principal component analysis (PCA) was performed on the integrated Seurat object. The first 30 principal components were used for dimensionality reduction via the Uniform Manifold

441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474

Approximation and Projection (UMAP) method. Clusters were defined using the FindClusters() function for a series of resolutions between 0.1 and 1. Optimal clustering resolution of 0.1 was selected based on cluster stability, as in Long et al., using the clustree package v0.5 in R. 116,117

Cell type annotation

FindAllMarkers() function in Seurat was used to define marker genes for further cluster annotation. Dot plots were generated using the DotPlot() function in Seurat with the argument “scale” set to the default value. To complement the FindAllMarkers() approach and annotate adrenal cortex clusters more precisely, we used the AddModuleScore() function in Seurat to calculate module scores based on the expression of NR5A1 (which encodes the protein steroidogenic factor 1, SF1), INHA, and MLANA, which are all adrenal cortex biomarkers. 17,18 Adrenal cortex clusters were characterized by higher adrenal cortex biomarker module scores when compared to lymphocyte, myeloid cell, endothelial, and fibroblast clusters.

Subclustering of immune cells

Myeloid- and lymphocyte-typed clusters were subset from the initial integrated Seurat object and re-processed using SCTransform as described above. Samples were then merged and re-integrated using Harmony26 at a resolution of 0.2. Dimensionality reduction and clustering steps were followed as described above. This process was repeated after non-immune cells were subset out. ggplot2118 was used to generate a histogram of immunocyte proportions and the FindAllMarkers() function in Seurat was used to define marker genes for cluster annotations displayed using the DotPlot() function. The macrophage-typed clusters were then further subset, re-processed, and re-integrated using Harmony at a resolution of 0.2. This process was repeated after excluding a small number of CD3-expressing lymphocytes. Dimensionality reduction, clustering steps, and cell type annotation were performed as described above. ggplot2118 was used to generate a histogram of macrophage subtypes for each sample type (normal adrenal, AA, PACC, MACC). The function propeller() from the speckle119 package for R was used to compare subtype proportions across sample types. False discovery rate (FDR) is reported. Average gene expression was calculated across each sample type and scaled average expression was displayed in a heatmap using the DoHeatmap() function. There were not enough cells for the CD8+ T cell, CD4+ T cell, and natural killer (NK) cell clusters to subset and reprocess, re-integrate, and recluster. Therefore, these lymphocytes were subset from the macrophage and leukocyte re-integrated object and

475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508

average gene expression was calculated across each sample type and displayed in a heatmap as described above.

Subclustering of fibroblasts

The fibroblast cluster was subset from the initial integrated Seurat object, reprocessed, and re-integrated using Harmony to increase clustering resolution. After dimensionality reduction and clustering, optimal resolution of 0.2 was selected as described above. As a result, 10 clusters were identified. Four clusters including adipocytes were excluded, after which the fibroblasts were reprocessed and re-integrating using Harmony. PACC 2 was excluded from analysis due to having only 5 fibroblasts. After dimensionality reduction and clustering, optimal resolution of 0.1 was selected as described above. As a result, 5 clusters of fibroblasts were identified. UMAP projection was displayed using the DimPlot() function in Seurat. Marker genes were identified using the FindAllMarkers() function and visualized using the DotPlot() function in Seurat. Average gene expression was calculated across sample types and displayed in a heatmap as described above.

Subclustering of adrenal cortex cells

To increase clustering resolution, adrenal cortex cells were subset from the integrated Seurat object and reprocessed using normalization, cell-cycle regression, integration using RPCA, dimensionality reduction, and clustering steps as in processing the initial Seurat object. Optimal clustering resolution of 0.1 was selected as described above. FindAllMarkers() function was used to identify marker genes for each of the clusters. Marker genes were visualized using the DotPlot() function as described above. In addition, adrenal cortex module scores were calculated for each of the clusters as described above. As a result, 6 adrenal cortex clusters and 1 cluster of adrenal medulla were identified. To further increase clustering resolution, the adrenal medulla cluster was excluded, and adrenal cortex cells were again subset from the integrated Seurat object and reprocessed using normalization, cell-cycle regression, integration, dimensionality reduction, and clustering steps as described above. Optimal resolution of 0.1 was selected as described above. As a result, 7 clusters of adrenal cortex cells were identified; these cluster assignments were used for all downstream analyses.

Cell type proportions

To compare cell type proportions of adrenal cortex cells across the samples, we used the propeller method from the speckle package v0.0.3 in R.119

509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542

Pathway analysis

To characterize adrenal cortex clusters based on gene expression patterns, we used the AddModuleScore() function from the Seurat package to calculate module scores using a selection of Hallmark pathways downloaded from MSigDB using the msigdbr package v7.5 in R.59,60 We further calculated average module scores for each of the clusters to generate the heatmap for Figure 4C. We also calculated average module scores for each of the clusters on a per-tumor type basis to generate the heatmap for Supplementary Figure S4. We performed a similar analysis using a selection of replication stress and DNA damage response REACTOME pathways downloaded from MsigDB59-61 We calculated module scores as in Wu et al.35 for each of the selected REACTOME pathways, and then calculated average module scores for each cluster on a per-tumor type basis (for Figure 5A) and on a per-sample basis (for supplementary Figure S5), which were displayed on a heatmap. To generate heatmaps, we used the pheatmap() function from the pheatmap package v.1.0 in R with the argument “scale” set to “row”.120 For all heatmaps, clusters with fewer than 50 cells were excluded from analysis. Additionally, we used QIAGEN Ingenuity Pathways Analysis (IPA) (QIAGEN Inc., https://digitalinsights.qiagen.com/IPD) to carry out pathway enrichment analysis for Figure 5B.62

InferCNV

We used inferCNV64 v1.8 to infer copy number alterations in the tumor samples. We used the HMM_CNV_predictions.HMMi6.hmm_mode-samples.Pnorm_0.5.pred_cnv_regions.datfile to quantify the number of chromosomal aberrations per sample. To compare the number of chromosomal aberrations between clusters, we used the chisq.test() function in R.

Data and code availability

snRNAseq data generated for this study has been deposited to dbGAP, accession phs003764. Code that was used to analyze the snRNAseq data is available upon request.

Immunofluorescence of ACC

Hematoxylin and eosin-stained slides of formalin fixed paraffin embedded (FFPE) samples were screened for diagnosis, cellularity, and necrosis by a board-certified pathologist. FFPE blocks were cut into 4-um sections and deparaffinized in organic solvents. Slides were

543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576

dehydrated, submerged in DAKO Antigen Retrieval Buffer (pH 9.0) (Agilent Technologies, Santa Clara, CA), and incubated at 110℃ in a steam rice cooker for 30 minutes. The slides were then cooled on ice for 30 minutes, followed by 5 minutes in distilled water. As previously described121, samples were permeabilized with DAKO Wash Buffer (Agilent Technologies), then blocked with blocking buffer (DAKO Wash buffer; 1% BSA). Sections were stained with mouse anti-0H2AX (Millipore-Sigma, St.Louis, MO, 05-636) at 1:500 and rabbit anti-RPA (Bethyl Laboratories, Montgomery, TX, A300-245A-M) at 1:200. Secondary incubation was performed with goat anti-mouse Alexa fluor 488 and goat anti-rabbit Alexa fluor 647. DAPI (Millipore-Sigma) was added, and the slides were dehydrated with increasing concentrations of ethanol. Samples were then mounted with ProLong Gold anti-fade reagent (Thermo Fisher Scientific) and stored at -20℃. Stained samples were imaged on a Leica Thunder Imager Microscope.

Cell culture

NCI-H295R

The NCI-H295R ACC cell line was originally purchased from American Type Culture Collection (ATCC, Manassas, VA). NCI-H295R cells were grown in DMEM/F12 with L- glutamine and 15 mM HEPES without phenol red (Thermo Fisher Scientific) supplemented with 2.5% Nu-Serum™ (Corning, Corning, NY), 1% penicillin-streptomycin (Sigma-Millipore), and 1x ITS Premix Universal Cell Culture Supplement (Corning). Cells were split approximately once per week and all experiments utilized cells below 20 passages.

ACC patient-derived tumor organoid (PTO) generation

PTOs were generated from a tumor harvested from a patient diagnosed with ACC (patient 5 in Table 1). Briefly, tumor tissue was minced then digested with 0.5 mg/mL collagenase II (Worthington Biochemical, Freehold, NJ) in DMEM/F12 (Cytiva, Marlborough, MA). Tumor cells were then seeded at 5-10x104 cells per well in 24-well plates in 80% Matrigel® (Corning), with DMEM/F12 supplemented with 10 nM HEPES (Cytiva), 1x Glutamax, 1x B27 Supplement (both Thermo Fisher Scientific), 1.25 mM n-acetylcysteine, 10 mM nicotinamide, 10 nM [Leu15]-Gastrin, 0.5 µM A-83-01, 1 µM PGE2 (all Sigma-Millipore), 0.5 µg/mL R-Spondin-1, 50 ng/ml EGF (both R&D Systems, Minneapolis, MN), 0.1 ug/mL FGF-10, 30 ng/ml recombinant murine Wnt3a, and 0.1 ug/mL recombinant human Noggin (all Thermo Fisher

577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609

Scientific). PTOs were routinely passaged upon confluency. Brightfield images of organoid colonies were captured using an EVOS™ M5000 microscope (Thermo Fisher Scientific).

Immunofluorescence of cultured cells and ACC PTOs

NCI-H295R cells were seeded at 1.5 x 105 cells per cm2 on top of 12 mm diameter poly-L- lysine coated glass coverslips (Electron Microscopy Sciences, Hatfield, PA) placed in wells of a 24-well cell culture plate. After 24 hours, cells were washed with PBS, fixed with 4% paraformaldehyde (Thermo Fisher Scientific), and permeabilized with 0.5% Triton X-100. After blocking with 10% normal goat serum (Abcam, Cambridge, UK) in 0.1% Triton X-100, cells were stained with mouse anti-yH2AX (Cell Signaling Technologies, Danvers, MA, CST#80312) at 1:250 followed by secondary incubation with goat anti-mouse Alexa Fluor® 488 antibody at 1:500 (Abcam ab150117) and phalloidin-Alexa Fluor® 594 at 1:1000 (Abcam #ab176757). After mounting with Fluoroshield mounting medium with DAPI (Abcam), images were obtained on a Nikon AXR confocal microscope (Nikon, Tokyo, Japan). Cells incubated without primary antibodies were used as negative controls for NCI-H295R immunofluorescence experiments.

ACC PTOs were removed from Matrigel domes and stained in a similar manner, and then mounted on the 0.2-0.4 um double cavity well microscope slides (Electron Microscopy ciences #71883-79) and imaged similarly to NCI-H295R cells. Primary antibodies were incubated at the following concentrations overnight at 4C: rabbit anti-SF-1 1:250 (Abcam ab51074), mouse anti-yH2AX 1:250 (Cell Signaling Technologies CST#80312), or equivalent isotype controls (Abcam ab172730, Cell Signaling Technologies CST#5415). The following secondary antibodies were used: goat anti-rabbit Alexa Fluor® 488 1:1000 (Abcam #ab150081), goat anti-mouse Alexa Fluor® 546 1:500 (Thermo Fisher Scientific A-11003) phalloidin-Alexa Fluor® 647 1:400 (Thermo Fisher Scientific A2287). Comparable isotype antibodies (Cell Signaling Technologies #5145 and Abcam ab172730) were used as negative controls for ACC PTO immunofluorescence experiments.

Cytotoxicity assays

NCI-H295R Cytotoxicity Assays

610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642

NCI-H295R cells were seeded at 1x104 cells per well in a 96-well plate and allowed to adhere for 24 hours. Media was then replaced with fresh media containing specified drugs or equivalent DMSO vehicle control (Corning). Cell viability was assessed by water-soluble tertozolium-8 (WST-8) based Cell Counting Kit-8 assay (Dojindo, Mashiki, Kumamoto, Japan) per the manufacturer’s instructions at 4 days. Absorbance readings were conducted using a SpectraMax® iD5 plate reader and SoftMax® Pro Software (Molecular Devices, San Jose, CA). All assays were performed in biologic triplicates with 3-5 replicates per condition. The following drugs were used at concentrations ranging from 1nM to 50uM: KU60019 (Selleck Chemicals, Houston, TX), RI-1 (Thermo Fisher Scientific), suberoylanilide hydroxamic acid (SAHA, aka vorinostat, Thermo Fisher Scientific), rabusertib (Selleck Chemicals), PF47736 (Selleck Chemicals), olaparib (Cell Signaling Technologies). Elimusertib (Selleck Chemicals), VX-803 (Selleck Chemicals), and AZ32 (Thermo Fisher Scientific). Mean viability of each drug treatment group was measured as a percentage of baseline DMSO vehicle treated cell absorbance. The half-maximal inhibitory concentration (IC50) for each drug was calculated by performing a four-parameter nonlinear best fit analysis of log concentration by percent viability on GraphPad Prism® 7.0a software (GraphPad Software, La Jolla, CA). Differences in viability across each concentration of a particular drug were compared using one-factor ANOVA with post-hoc multiple comparison analysis by Dunnett’s test in GraphPad Prism® 7.0a software (GraphPad Software). All data are reported as means ± SEM with statistical significance set as p <0.05.

ACC PTO cytotoxicity assays

Single-cell suspensions were generated by incubating well-established ACC organoid domes with 1 mg/mL dispase II (Thermo Fisher Scientific). 1.5x104 cells per well were seeded in 96- well plates in 80% Matrigel® as described above. Three days after seeding, media was replaced with media containing specified drugs or DMSO vehicle control. Organoid viability was assessed at 4 days by WST-8 assay as described above using the cell free Matrigel® domes for background absorbance readings. Differences in viability across each concentration of a drug were compared using one-factor ANOVA with post-hoc multiple comparison analysis by Dunnett’s test in GraphPad Prism® 7.0a software (GraphPad Software, La Jolla, CA). Each drug was tested in at least 3 independent experiments with 5 replicates per experiment.

643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676

Single cell suspensions of NCI-H295R cells and ACC organoid-derived cells were generated as described above. 3x105 cells were plated and medium was collected at 72-hours. Cortisol was measured by ELISA per manufacturer’s instructions (Abcam ab108665). Each experiment was conducted in triplicate and read on BioTek Synergy H1 microplate reader using Gen5 v3.10 software (Agilent Technologies). Cortisol concentrations were compared between groups using two-tailed independent student’s t-test. All data are reported as means ± SEM with statistical significance set was p <0.05.

AUTHOR CONTRIBUTION STATEMENT

Conceptualization: P.H.D.

Methodology: P.H.D. and K.E.M.

Investigation: L.V.P., E.A.R.G., D.M.C., J.B.N., A.R., Y.S., and E.L.

Specimen Acquisition: P.H.D., J.E.P., and B.S.M.

Specimen Pathology Review: S.S.

Formal analysis: L.V.P., E.A.R.G., D.M.C., Y.S., E.L.

Data curation: L.V.P., E.A.R.G., and K.E.M.

Funding acquisition: P.H.D.

Project administration: P.H.D.

Supervision: E.R.M., K.E.M. and P.H.D.

Validation: K.E.M. and P.H.D.

Visualization: L.V.P., P.H.D.

Writing-original draft: L.V.P., E.A.R.G., and D.M.C.

Writing-review and editing: L.V.P., E.A.R.G., D.M.C., B.S.M., M.M., K.E.M., and P.H.D.

ACKNOWLEDGMENTS

This project was supported by Ohio Cancer Research, The Victory Over Cancer Foundation, DOD W81XWH-22-PRCRP-CDA-SO, NIH R01CA279997, and NIH R21CA277083 (PHD). LVP is supported by the Pelotonia Scholars Program. DMC is supported by training grant PF- 23-1036284-01-CDP via the American Cancer Society. Any opinions, findings, and conclusions expressed in this material are those of the author(s) and do not necessarily reflect those of the NIH, Pelotonia Scholars Program, OSU, or the American Cancer Society. We

677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710

bioRxiv preprint doi: https://doi.org/10.1101/2024.09.30.615695; this version posted October 28, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

acknowledge resources from the Campus Microscopy and Imaging Facility (CMIF), Biospecimen Services Shared Resource and the Microscopy Shared Resource. These shared resources are supported in part by the Cancer Center Support Grant P30 CA016058, National Cancer Institute, Bethesda, MD. We also acknowledge Matthew Ringel, MD, Gary Hammer, MD, PhD, Timothy Frankel, MD, PHD, Wayne Miles, PhD, Abby Green, MD, PHD, and Vineeth Sukrithan, MD for insightful discussions.

711 712 713 714 715 716 717

bioRxiv preprint doi: https://doi.org/10.1101/2024.09.30.615695; this version posted October 28, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

REFERENCES

1. Else T, Williams AR, Sabolch A, Jolly S, Miller BS, Hammer GD. Adjuvant therapies and patient and tumor characteristics associated with survival of adult patients with adrenocortical carcinoma. J Clin Endocrinol Metab. 2014;99(2):455-461. doi:10.1210/jc.2013-2856

2. Else T, Kim AC, Sabolch A, et al. Adrenocortical carcinoma. Endocr Rev. 2014;35(2):282- 326. doi:10.1210/er.2013-1029

3. Bilimoria KY, Shen WT, Elaraj D, et al. Adrenocortical carcinoma in the United States: treatment utilization and prognostic factors. Cancer. 2008;113(11):3130-3136. doi:10.1002/cncr.23886

4. Vaidya A, Nehs M, Kilbridge K. Treatment of Adrenocortical Carcinoma. Surg Pathol Clin. 2019;12(4):997-1006. doi:10.1016/j.path.2019.08.010

5. Kenney L, Hughes M. Adrenocortical Carcinoma: Role of Adjuvant and Neoadjuvant Therapy. Surg Oncol Clin N Am. 2023;32(2):279-287. doi:10.1016/j.soc.2022.10.005

6. Fassnacht M, Terzolo M, Allolio B, et al. Combination chemotherapy in advanced adrenocortical carcinoma. N Engl J Med. 2012;366(23):2189-2197. doi:10.1056/NEJMoa1200966

7. Araujo-Castro M, Pascual-Corrales E, Molina-Cerrillo J, Alonso-Gordoa T. Immunotherapy in Adrenocortical Carcinoma: Predictors of Response, Efficacy, Safety, and Mechanisms of Resistance. Biomedicines. 2021;9(3):304. doi:10.3390/biomedicines9030304

8. Zheng S, Cherniack AD, Dewal N, et al. Comprehensive Pan-Genomic Characterization of Adrenocortical Carcinoma. Cancer Cell. 2016;29(5):723-736. doi:10.1016/j.ccell.2016.04.002

9. Assié G, Letouzé E, Fassnacht M, et al. Integrated genomic characterization of adrenocortical carcinoma. Nat Genet. 2014;46(6):607-612. doi:10.1038/ng.2953

10. McGranahan N, Swanton C. Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future. Cell. 2017;168(4):613-628. doi:10.1016/j.cell.2017.01.018

11. Lei Y, Tang R, Xu J, et al. Applications of single-cell sequencing in cancer research: progress and perspectives. J Hematol OncolJ Hematol Oncol. 2021;14(1):91. doi:10.1186/s13045-021-01105-2

12. Bischoff P, Trinks A, Wiederspahn J, et al. The single-cell transcriptional landscape of lung carcinoid tumors. Int J Cancer. 2022;150(12):2058-2071. doi:10.1002/ijc.33995

718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751

13. Pu W, Shi X, Yu P, et al. Single-cell transcriptomic analysis of the tumor ecosystems underlying initiation and progression of papillary thyroid carcinoma. Nat Commun. 2021;12(1):6058. doi:10.1038/s41467-021-26343-3

14. Penny MK, Lerario AM, Basham KJ, et al. Targeting Oncogenic Wnt/ß-Catenin Signaling in Adrenocortical Carcinoma Disrupts ECM Expression and Impairs Tumor Growth. Cancers. 2023;15(14):3559. doi:10.3390/cancers15143559

15. Wang Q, Sun N, Meixner R, et al. Metabolic heterogeneity in adrenocortical carcinoma impacts patient outcomes. JCI Insight. 2023;8(16):e167007. doi:10.1172/jci.insight.167007

16. Tourigny DS, Altieri B, Secener KA, et al. Cellular landscape of adrenocortical carcinoma at single-nuclei resolution. Mol Cell Endocrinol. 2024;590:112272. doi:10.1016/j.mce.2024.112272

17. Sbiera S, Schmull S, Assie G, et al. High diagnostic and prognostic value of steroidogenic factor-1 expression in adrenal tumors. J Clin Endocrinol Metab. 2010;95(10):E161-171. doi:10.1210/jc.2010-0653

18. Mizdrak M, Tičinović Kurir T, Božić J. The Role of Biomarkers in Adrenocortical Carcinoma: A Review of Current Evidence and Future Perspectives. Biomedicines. 2021;9(2):174. doi:10.3390/biomedicines9020174

19. Hao Y, Hao S, Andersen-Nissen E, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573-3587.e29. doi:10.1016/j.cell.2021.04.048

20. Al Barashdi MA, Ali A, McMullin MF, Mills K. Protein tyrosine phosphatase receptor type C (PTPRC or CD45). J Clin Pathol. 2021;74(9):548-552. doi:10.1136/jclinpath-2020- 206927

21. Broz ML, Krummel MF. The emerging understanding of myeloid cells as partners and targets in tumor rejection. Cancer Immunol Res. 2015;3(4):313-319. doi:10.1158/2326- 6066.CIR-15-0041

22. Li D, Xia L, Huang P, et al. Cancer-associated fibroblast-secreted IGFBP7 promotes gastric cancer by enhancing tumor associated macrophage infiltration via FGF2/FGFR1/PI3K/AKT axis. Cell Death Discov. 2023;9(1):17. doi:10.1038/s41420-023- 01336-x

23. Sakamoto H, Koma YI, Higashino N, et al. PAI-1 derived from cancer-associated fibroblasts in esophageal squamous cell carcinoma promotes the invasion of cancer cells and the migration of macrophages. Lab Investig J Tech Methods Pathol. 2021;101(3):353- 368. doi:10.1038/s41374-020-00512-2

24. Sawano A, Iwai S, Sakurai Y, et al. Flt-1, vascular endothelial growth factor receptor 1, is a novel cell surface marker for the lineage of monocyte-macrophages in humans. Blood. 2001;97(3):785-791. doi:10.1182/blood.v97.3.785

752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787

25. Lopes da Silva M, Cutler DF. von Willebrand factor multimerization and the polarity of secretory pathways in endothelial cells. Blood. 2016;128(2):277-285. doi:10.1182/blood- 2015-10-677054

26. Korsunsky I, Millard N, Fan J, et al. Fast, sensitive and accurate integration of single-cell data with Harmony. Nat Methods. 2019;16(12):1289-1296. doi:10.1038/s41592-019-0619- 0

27. Prenen H, Mazzone M. Tumor-associated macrophages: a short compendium. Cell Mol Life Sci CMLS. 2019;76(8):1447-1458. doi:10.1007/s00018-018-2997-3

28. Akinrinmade OA, Chetty S, Daramola AK, Islam MU, Thepen T, Barth S. CD64: An Attractive Immunotherapeutic Target for M1-type Macrophage Mediated Chronic Inflammatory Diseases. Biomedicines. 2017;5(3):56. doi:10.3390/biomedicines5030056

29. Martinez FO, Helming L, Milde R, et al. Genetic programs expressed in resting and IL-4 alternatively activated mouse and human macrophages: similarities and differences. Blood. 2013;121(9):e57-69. doi:10.1182/blood-2012-06-436212

30. Lu J, Ma L. The role of tumor-associated macrophages in the development, metastasis and treatment of breast cancer. Pathol Res Pract. 2020;216(9):153085. doi:10.1016/j.prp.2020.153085

31. Georgoudaki AM, Prokopec KE, Boura VF, et al. Reprogramming Tumor-Associated Macrophages by Antibody Targeting Inhibits Cancer Progression and Metastasis. Cell Rep. 2016;15(9):2000-2011. doi:10.1016/j.celrep.2016.04.084

32. Nasir I, McGuinness C, Poh AR, Ernst M, Darcy PK, Britt KL. Tumor macrophage functional heterogeneity can inform the development of novel cancer therapies. Trends Immunol. 2023;44(12):971-985. doi:10.1016/j.it.2023.10.007

33. Bill R, Wirapati P, Messemaker M, et al. CXCL9:SPP1 macrophage polarity identifies a network of cellular programs that control human cancers. Science. 2023;381(6657):515- 524. doi:10.1126/science.ade2292

34. Cheng S, Li Z, Gao R, et al. A pan-cancer single-cell transcriptional atlas of tumor infiltrating myeloid cells. Cell. 2021;184(3):792-809.e23. doi:10.1016/j.cell.2021.01.010

35. Wu SZ, Al-Eryani G, Roden DL, et al. A single-cell and spatially resolved atlas of human breast cancers. Nat Genet. 2021;53(9):1334-1347. doi:10.1038/s41588-021-00911-1

36. Ozato Y, Kojima Y, Kobayashi Y, et al. Spatial and single-cell transcriptomics decipher the cellular environment containing HLA-G+ cancer cells and SPP1+ macrophages in colorectal cancer. Cell Rep. 2023;42(1):111929. doi:10.1016/j.celrep.2022.111929

37. Shinchi Y, Ishizuka S, Komohara Y, et al. The expression of PD-1 ligand 1 on macrophages and its clinical impacts and mechanisms in lung adenocarcinoma. Cancer Immunol Immunother CII. 2022;71(11):2645-2661. doi:10.1007/s00262-022-03187-4

788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823

38. Mantovani A, Allavena P, Marchesi F, Garlanda C. Macrophages as tools and targets in cancer therapy. Nat Rev Drug Discov. 2022;21(11):799-820. doi:10.1038/s41573-022- 00520-5

39. Steele NG, Carpenter ES, Kemp SB, et al. Multimodal Mapping of the Tumor and Peripheral Blood Immune Landscape in Human Pancreatic Cancer. Nat Cancer. 2020;1(11):1097-1112. doi:10.1038/s43018-020-00121-4

40. Zhang H, Liu L, Liu J, et al. Roles of tumor-associated macrophages in anti-PD-1/PD-L1 immunotherapy for solid cancers. Mol Cancer. 2023;22(1):58. doi:10.1186/s12943-023- 01725-x

41. Ji K, Zhang Y, Jiang S, et al. SIRPa blockade improves the antitumor immunity of radiotherapy in colorectal cancer. Cell Death Discov. 2023;9(1):180. doi:10.1038/s41420- 023-01472-4

42. Zhang CX, Huang DJ, Baloche V, et al. Galectin-9 promotes a suppressive microenvironment in human cancer by enhancing STING degradation. Oncogenesis. 2020;9(7):65. doi:10.1038/s41389-020-00248-0

43. Dong Y, Sun Q, Zhang X. PD-1 and its ligands are important immune checkpoints in cancer. Oncotarget. 2017;8(2):2171-2186. doi:10.18632/oncotarget.13895

44. Thiel M, Wolfs MJ, Bauer S, et al. Efficiency of T-cell costimulation by CD80 and CD86 cross-linking correlates with calcium entry. Immunology. 2010;129(1):28-40. doi:10.1111/j.1365-2567.2009.03155.x

45. Landwehr LS, Altieri B, Schreiner J, et al. Interplay between glucocorticoids and tumor- infiltrating lymphocytes on the prognosis of adrenocortical carcinoma. J Immunother Cancer. 2020;8(1):e000469. doi:10.1136/jitc-2019-000469

46. Friedlaender A, Addeo A, Banna G. New emerging targets in cancer immunotherapy: the role of TIM3. ESMO Open. 2019;4(Suppl 3):e000497. doi:10.1136/esmoopen-2019- 000497

47. Thommen DS, Schumacher TN. T Cell Dysfunction in Cancer. Cancer Cell. 2018;33(4):547-562. doi:10.1016/j.ccell.2018.03.012

48. Liu PS, Chen YT, Li X, et al. CD40 signal rewires fatty acid and glutamine metabolism for stimulating macrophage anti-tumorigenic functions. Nat Immunol. 2023;24(3):452-462. doi:10.1038/s41590-023-01430-3

49. Zhang Q, Bi J, Zheng X, et al. Blockade of the checkpoint receptor TIGIT prevents NK cell exhaustion and elicits potent anti-tumor immunity. Nat Immunol. 2018;19(7):723-732. doi:10.1038/s41590-018-0132-0

50. Crinier A, Dumas PY, Escalière B, et al. Single-cell profiling reveals the trajectories of natural killer cell differentiation in bone marrow and a stress signature induced by acute

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

myeloid leukemia. Cell Mol Immunol. 2021;18(5):1290-1304. doi:10.1038/s41423-020- 00574-8

51. Lendahl U, Muhl L, Betsholtz C. Identification, discrimination and heterogeneity of fibroblasts. Nat Commun. 2022;13(1):3409. doi:10.1038/s41467-022-30633-9

52. Li S, Lu R, Shu L, et al. An integrated map of fibroblastic populations in human colon mucosa and cancer tissues. Commun Biol. 2022;5(1):1326. doi:10.1038/s42003-022- 04298-5

53. Yang D, Liu J, Qian H, Zhuang Q. Cancer-associated fibroblasts: from basic science to anticancer therapy. Exp Mol Med. 2023;55(7):1322-1332. doi:10.1038/s12276-023-01013- 0

54. Menezes S, Okail MH, Jalil SMA, Kocher HM, Cameron AJM. Cancer-associated fibroblasts in pancreatic cancer: new subtypes, new markers, new targets. J Pathol. 2022;257(4):526-544. doi:10.1002/path.5926

55. Han C, Liu T, Yin R. Biomarkers for cancer-associated fibroblasts. Biomark Res. 2020;8(1):64. doi:10.1186/s40364-020-00245-w

56. Bedoya-Reina OC, Li W, Arceo M, et al. Single-nuclei transcriptomes from human adrenal gland reveal distinct cellular identities of low and high-risk neuroblastoma tumors. Nat Commun. 2021;12(1):5309. doi:10.1038/s41467-021-24870-7

57. Ye L, Su ZJ, Ge RS. Inhibitors of testosterone biosynthetic and metabolic activation enzymes. Mol Basel Switz. 2011;16(12):9983-10001. doi:10.3390/molecules16129983

58. Neunzig J, Bernhardt R. Dehydroepiandrosterone sulfate (DHEAS) stimulates the first step in the biosynthesis of steroid hormones. PloS One. 2014;9(2):e89727. doi:10.1371/journal.pone.0089727

59. Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinforma Oxf Engl. 2011;27(12):1739- 1740. doi:10.1093/bioinformatics/btr260

60. Liberzon A, Birger C, Thorvaldsdóttir H, Ghandi M, Mesirov JP, Tamayo P. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst. 2015;1(6):417-425. doi:10.1016/j.cels.2015.12.004

61. Gillespie M, Jassal B, Stephan R, et al. The reactome pathway knowledgebase 2022. Nucleic Acids Res. 2022;50(D1):D687-D692. doi:10.1093/nar/gkab1028

62. Krämer A, Green J, Pollard J, Tugendreich S. Causal analysis approaches in Ingenuity Pathway Analysis. Bioinforma Oxf Engl. 2014;30(4):523-530. doi:10.1093/bioinformatics/btt703

860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893

63. Gaillard H, García-Muse T, Aguilera A. Replication stress and cancer. Nat Rev Cancer. 2015;15(5):276-289. doi:10.1038/nrc3916

64. Tickle T, Tirosh I, Georgescu C, Brown M, Haas B. inferCNV of the Trinity CTAT Project. Published online 2019. https://github.com/broadinstitute/inferCNV

65. Yang SF, Nelson CB, Wells JK, et al. ZNF827 is a single-stranded DNA binding protein that regulates the ATR-CHK1 DNA damage response pathway. Nat Commun. 2024;15(1):2210. doi:10.1038/s41467-024-46578-0

66. Wengner AM, Siemeister G, Lücking U, et al. The Novel ATR Inhibitor BAY 1895344 Is Efficacious as Monotherapy and Combined with DNA Damage-Inducing or Repair- Compromising Therapies in Preclinical Cancer Models. Mol Cancer Ther. 2020;19(1):26- 38. doi:10.1158/1535-7163.MCT-19-0019

67. Shu J, Wang X, Yang X, Zhao G. ATM inhibitor KU60019 synergistically sensitizes lung cancer cells to topoisomerase II poisons by multiple mechanisms. Sci Rep. 2023;13(1):882. doi:10.1038/s41598-023-28185-z

68. Scotto L, Serrano XJ, Zullo K, et al. ATM inhibition overcomes resistance to histone deacetylase inhibitor due to p21 induction and cell cycle arrest. Oncotarget. 2020;11(37):3432-3442. doi:10.18632/oncotarget.27723

69. Wang C, Tang M, Chen Z, et al. Genetic vulnerabilities upon inhibition of DNA damage response. Nucleic Acids Res. 2021;49(14):8214-8231. doi:10.1093/nar/gkab643

70. da Costa AABA, Chowdhury D, Shapiro GI, D’Andrea AD, Konstantinopoulos PA. Targeting replication stress in cancer therapy. Nat Rev Drug Discov. 2023;22(1):38-58. doi:10.1038/s41573-022-00558-5

71. Ito M, Fujita Y, Shinohara A. Positive and negative regulators of RAD51/DMC1 in homologous recombination and DNA replication. DNA Repair. 2024;134:103613. doi:10.1016/j.dnarep.2023.103613

72. Yang X, Pan Y, Qiu Z, et al. RNF126 as a Biomarker of a Poor Prognosis in Invasive Breast Cancer and CHEK1 Inhibitor Efficacy in Breast Cancer Cells. Clin Cancer Res Off J Am Assoc Cancer Res. 2018;24(7):1629-1643. doi:10.1158/1078-0432.CCR-17-2242

73. Corrales-Guerrero S, Cui T, Castro-Aceituno V, et al. Inhibition of RRM2 radiosensitizes glioblastoma and uncovers synthetic lethality in combination with targeting CHK1. Cancer Lett. 2023;570:216308. doi:10.1016/j.canlet.2023.216308

74. Fleury H, Malaquin N, Tu V, et al. Exploiting interconnected synthetic lethal interactions between PARP inhibition and cancer cell reversible senescence. Nat Commun. 2019;10(1):2556. doi:10.1038/s41467-019-10460-1

894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927

75. Cahuzac M, Langlois P, Péant B, Fleury H, Mes-Masson AM, Saad F. Pre-activation of autophagy impacts response to olaparib in prostate cancer cells. Commun Biol. 2022;5(1):251. doi:10.1038/s42003-022-03210-5

76. Tuveson D, Clevers H. Cancer modeling meets human organoid technology. Science. 2019;364(6444):952-955. doi:10.1126/science.aaw6985

77. Dedhia PH, Bertaux-Skeirik N, Zavros Y, Spence JR. Organoid Models of Human Gastrointestinal Development and Disease. Gastroenterology. 2016;150(5):1098-1112. doi:10.1053/j.gastro.2015.12.042

78. Pittet MJ, Michielin O, Migliorini D. Clinical relevance of tumour-associated macrophages. Nat Rev Clin Oncol. 2022;19(6):402-421. doi:10.1038/s41571-022-00620-6

79. Wu J, Shen Y, Zeng G, Liang Y, Liao G. SPP1+ TAM subpopulations in tumor microenvironment promote intravasation and metastasis of head and neck squamous cell carcinoma. Cancer Gene Ther. 2024;31(2):311-321. doi:10.1038/s41417-023-00704-0

80. Wang Z, Liu J, Huang H, et al. Metastasis-associated fibroblasts: an emerging target for metastatic cancer. Biomark Res. 2021;9(1):47. doi:10.1186/s40364-021-00305-9

81. Lee HO, Mullins SR, Franco-Barraza J, Valianou M, Cukierman E, Cheng JD. FAP- overexpressing fibroblasts produce an extracellular matrix that enhances invasive velocity and directionality of pancreatic cancer cells. BMC Cancer. 2011;11:245. doi:10.1186/1471-2407-11-245

82. Lo A, Wang LCS, Scholler J, et al. Tumor-Promoting Desmoplasia Is Disrupted by Depleting FAP-Expressing Stromal Cells. Cancer Res. 2015;75(14):2800-2810. doi:10.1158/0008-5472.CAN-14-3041

83. Voltà-Durán E, Alba-Castellón L, Serna N, et al. High-precision targeting and destruction of cancer-associated PDGFR-ß+ stromal fibroblasts through self-assembling, protein-only nanoparticles. Acta Biomater. 2023;170:543-555. doi:10.1016/j.actbio.2023.09.001

84. Subramanian C, Cohen MS. Over expression of DNA damage and cell cycle dependent proteins are associated with poor survival in patients with adrenocortical carcinoma. Surgery. 2019;165(1):202-210. doi:10.1016/j.surg.2018.04.080

85. Gara SK, Tyagi MV, Patel DT, et al. GATA3 and APOBEC3B are prognostic markers in adrenocortical carcinoma and APOBEC3B is directly transcriptionally regulated by GATA3. Oncotarget. 2020;11(36):3354-3370. doi:10.18632/oncotarget.27703

86. Gara SK, Lack J, Zhang L, Harris E, Cam M, Kebebew E. Metastatic adrenocortical carcinoma displays higher mutation rate and tumor heterogeneity than primary tumors. Nat Commun. 2018;9(1):4172. doi:10.1038/s41467-018-06366-z

87. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-674. doi:10.1016/j.cell.2011.02.013

928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963

88. Ubhi T, Brown GW. Exploiting DNA Replication Stress for Cancer Treatment. Cancer Res. 2019;79(8):1730-1739. doi:10.1158/0008-5472.CAN-18-3631

89. Saldivar JC, Cortez D, Cimprich KA. The essential kinase ATR: ensuring faithful duplication of a challenging genome. Nat Rev Mol Cell Biol. 2017;18(10):622-636. doi:10.1038/nrm.2017.67

90. Lecona E, Fernandez-Capetillo O. Targeting ATR in cancer. Nat Rev Cancer. 2018;18(9):586-595. doi:10.1038/s41568-018-0034-3

91. Yano K, Shiotani B. Emerging strategies for cancer therapy by ATR inhibitors. Cancer Sci. 2023;114(7):2709-2721. doi:10.1111/cas.15845

92. Dillon MT, Guevara J, Mohammed K, et al. Durable responses to ATR inhibition with ceralasertib in tumors with genomic defects and high inflammation. J Clin Invest. 2024;134(2):e175369. doi:10.1172/JCI175369

93. Yap TA, Tan DSP, Terbuch A, et al. First-in-Human Trial of the Oral Ataxia Telangiectasia and RAD3-Related (ATR) Inhibitor BAY 1895344 in Patients with Advanced Solid Tumors. Cancer Discov. 2021;11(1):80-91. doi:10.1158/2159-8290.CD- 20-0868

94. Yap TA, O’Carrigan B, Penney MS, et al. Phase I Trial of First-in-Class ATR Inhibitor M6620 (VX-970) as Monotherapy or in Combination With Carboplatin in Patients With Advanced Solid Tumors. J Clin Oncol Off J Am Soc Clin Oncol. 2020;38(27):3195-3204. doi:10.1200/JCO.19.02404

95. Middleton MR, Dean E, Evans TRJ, et al. Phase 1 study of the ATR inhibitor berzosertib (formerly M6620, VX-970) combined with gemcitabine ± cisplatin in patients with advanced solid tumours. Br J Cancer. 2021;125(4):510-519. doi:10.1038/s41416-021- 01405-x

96. Shapiro GI, Wesolowski R, Devoe C, et al. Phase 1 study of the ATR inhibitor berzosertib in combination with cisplatin in patients with advanced solid tumours. Br J Cancer. 2021;125(4):520-527. doi:10.1038/s41416-021-01406-w

97. Telli ML, Tolaney SM, Shapiro GI, et al. Phase 1b study of berzosertib and cisplatin in patients with advanced triple-negative breast cancer. NPJ Breast Cancer. 2022;8(1):45. doi:10.1038/s41523-022-00406-0

98. Kim ST, Smith SA, Mortimer P, et al. Phase I Study of Ceralasertib (AZD6738), a Novel DNA Damage Repair Agent, in Combination with Weekly Paclitaxel in Refractory Cancer. Clin Cancer Res Off J Am Assoc Cancer Res. 2021;27(17):4700-4709. doi:10.1158/1078- 0432.CCR-21-0251

99. Konstantinopoulos PA, Cheng SC, Wahner Hendrickson AE, et al. Berzosertib plus gemcitabine versus gemcitabine alone in platinum-resistant high-grade serous ovarian

964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999

cancer: a multicentre, open-label, randomised, phase 2 trial. Lancet Oncol. 2020;21(7):957- 968. doi:10.1016/S1470-2045(20)30180-7

100. Gupta M, Concepcion CP, Fahey CG, et al. BRG1 Loss Predisposes Lung Cancers to Replicative Stress and ATR Dependency. Cancer Res. 2020;80(18):3841-3854. doi:10.1158/0008-5472.CAN-20-1744

101. Borenäs M, Umapathy G, Lind DE, et al. ALK signaling primes the DNA damage response sensitizing ALK-driven neuroblastoma to therapeutic ATR inhibition. Proc Natl Acad Sci U S A. 2024;121(1):e2315242121. doi:10.1073/pnas.2315242121

102. Kwok M, Davies N, Agathanggelou A, et al. ATR inhibition induces synthetic lethality and overcomes chemoresistance in TP53- or ATM-defective chronic lymphocytic leukemia cells. Blood. 2016;127(5):582-595. doi:10.1182/blood-2015-05-644872

103. Toledo LI, Murga M, Zur R, et al. A cell-based screen identifies ATR inhibitors with synthetic lethal properties for cancer-associated mutations. Nat Struct Mol Biol. 2011;18(6):721-727. doi:10.1038/nsmb.2076

104. Reaper PM, Griffiths MR, Long JM, et al. Selective killing of ATM- or p53-deficient cancer cells through inhibition of ATR. Nat Chem Biol. 2011;7(7):428-430. doi:10.1038/nchembio.573

105. Libè R, Groussin L, Tissier F, et al. Somatic TP53 mutations are relatively rare among adrenocortical cancers with the frequent 17p13 loss of heterozygosity. Clin Cancer Res Off J Am Assoc Cancer Res. 2007;13(3):844-850. doi:10.1158/1078-0432.CCR-06-2085

106. Neizer-Ashun F, Bhattacharya R. Reality CHEK: Understanding the biology and clinical potential of CHK1. Cancer Lett. 2021;497:202-211. doi:10.1016/j.canlet.2020.09.016

107. Couch FB, Bansbach CE, Driscoll R, et al. ATR phosphorylates SMARCAL1 to prevent replication fork collapse. Genes Dev. 2013;27(14):1610-1623. doi:10.1101/gad.214080.113

108. Chen YH, Jones MJK, Yin Y, et al. ATR-mediated phosphorylation of FANCI regulates dormant origin firing in response to replication stress. Mol Cell. 2015;58(2):323- 338. doi:10.1016/j.molcel.2015.02.031

109. . Ammazzalorso F, Pirzio LM, Bignami M, Franchitto A, Pichierri P. ATR and ATM differently regulate WRN to prevent DSBs at stalled replication forks and promote replication fork recovery. EMBO J. 2010;29(18):3156-3169. doi:10.1038/emboj.2010.205

110. Slyper M, Porter CBM, Ashenberg O, et al. A single-cell and single-nucleus RNA-Seq toolbox for fresh and frozen human tumors. Nat Med. 2020;26(5):792-802. doi:10.1038/s41591-020-0844-1

1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034

111. Lacar B, Linker SB, Jaeger BN, et al. Nuclear RNA-seq of single neurons reveals molecular signatures of activation. Nat Commun. 2016;7:11022. doi:10.1038/ncomms11022

112. McGinnis CS, Murrow LM, Gartner ZJ. DoubletFinder: Doublet Detection in Single- Cell RNA Sequencing Data Using Artificial Nearest Neighbors. Cell Syst. 2019;8(4):329- 337.e4. doi:10.1016/j.cels.2019.03.003

113. Tirosh I, Izar B, Prakadan SM, et al. Dissecting the multicellular ecosystem of metastatic melanoma by single-cell RNA-seq. Science. 2016;352(6282):189-196. doi:10.1126/science.aad0501

114. Hafemeister C, Satija R. Normalization and variance stabilization of single-cell RNA- seq data using regularized negative binomial regression. Genome Biol. 2019;20(1):296. doi:10.1186/s13059-019-1874-1

115. Stuart T, Butler A, Hoffman P, et al. Comprehensive Integration of Single-Cell Data. Cell. 2019;177(7):1888-1902.e21. doi:10.1016/j.cell.2019.05.031

116. Zappia L, Oshlack A. Clustering trees: a visualization for evaluating clusterings at multiple resolutions. GigaScience. 2018;7(7):giy083. doi:10.1093/gigascience/giy083

117. Long Z, Sun C, Tang M, et al. Single-cell multiomics analysis reveals regulatory programs in clear cell renal cell carcinoma. Cell Discov. 2022;8(1):68. doi:10.1038/s41421-022-00415-0

118. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Published online 2016. https://ggplot2.tidyverse.org.

119. Phipson B, Sim CB, Porrello ER, Hewitt AW, Powell J, Oshlack A. propeller: testing for differences in cell type proportions in single cell data. Bioinforma Oxf Engl. 2022;38(20):4720-4726. doi:10.1093/bioinformatics/btac582

120. Kolde R. pheatmap: Pretty Heatmaps. Published online 2019. https://CRAN.R- project.org/package=pheatmap

121. van Biljon L, Fashemi B, Rodriguez J, et al. Visualizing DNA Damage Repair Proteins in Patient-Derived Ovarian Cancer Organoids via Immunofluorescence Assays. J Vis Exp JoVE. 2023;(192). doi:10.3791/64881

1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064

1065 1066

FIGURE LEGENDS

Sample ID	Sample Name	Sample Description	Paired	Organ Site	Age at Surgery	Sex	Excess hormone secretion	Neoadjuvant Treatment	ENSAT Stage	KI67 Index	Somatic Mutations	Germline Mutations
Patient 1	NORMAL	Normal Adrenal	Yes	Adrenal Gland	35	F	None	N/A	N/A	N/A	N/A	N/A
	AA 1	Adrenal Adenoma	Yes	Adrenal Gland			Cortisol	N/A	N/A	N/A	N/A	N/A
Patient 2	AA 2	Adrenal Adenoma	No	Adrenal Gland	34	M	Cortisol	N/A	N/A	N/A	N/A	N/A
Patient 3	PACC 1	Primary ACC	Yes	Adrenal Gland	61	M	Cortisol			No data available	No data available	No data available
	MACC 1	Metastatic ACC	Yes	Liver				Etoposide, cisplatin	4	50%	1 CHEK2 p.Ile157Thr CTNNB1 p.Ser33Cys TP53 p.Arg267Trp DUSP16-ASXL1 fusion (D3;A11*)	No data available
Patient 4	PACC 2	Primary ACC	No	Adrenal Gland	54	F	Cortisol, DHEAS	Mitotane started 12 days prior to surgery	4	15-20%	DAXX p.Glu583fs*19 1 TP53 p.Tyr220fs*27 BCORL1 p.Trp14*	No data available
Patient 5	PACC 3	Primary ACC	No	Adrenal Gland	25	F	Cortisol, androstenedione, testosterone, DHEAS	None	4	20%	TERT promoter c.124C>T 2 CDKN2A Copy Number Loss CDKN2B Copy Number Loss	No data available
Patient 6	PACC 4	Primary ACC	No	Adrenal Gland	54	F	None	None	4	No data available	CTNNB1 p.Ser545Pro 2 TP53 p.Asn200fs LOF	No data available
Patient 7	MACC 2	Metastatic ACC	No	Liver	57	M	Cortisol	Mitotane	4	20% (for primary tumor)	No data available	No clinically significant variants detected
Patient 8	MACC 3	Metastatic ACC	No	Lung	45	F	Cortisol	Mitotane	4	35% (for primary tumor)	No data available	No clinically significant variants detected
Patient 9	MACC 4	Metastatic ACC	No	Cervical Lymph Node	38	M	Cortisol, Androstenedione	2 months of mitotane treatment 2 years prior to surgery	4	1-10% (for primary tumor)	No data available	Variants of unknown significance detected DICER p.Trp32Cys EGLN1 p.Pro378Ser

Table 1. Summary of Patient Specimens. AA - adrenal adenoma, PACC - primary ACC, MACC - metastatic ACC, DHEAS - dehydroepiandrosterone sulfate. 1 - Foundation Medicine Testing, 2 - Tempus Testing.

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085

Figure 1. Single-Nuclei RNA Sequencing Reveals Intratumoral Heterogeneity in Normal Adrenal, Adrenal Adenoma, Primary ACC, and Metastatic ACC. A. Overview of the experimental workflow. B. Summary of the samples and their hormone secretion profiles. C. UMAP projection of the integrated Seurat object showing that 10 clusters of 5 distinct cell types were identified. D. Dot plot demonstrating marker genes that were used to annotate the identified clusters. For the dot plot, scaled average expression is displayed. E. Bar plot showing cell type proportions of the identified cell populations across different sample types. DHEAS - dehydroepiandrosterone sulfate, AC - adrenal cortex.

A

30

20

-

Normal Adrenal n = 1

Primary ACC n = 4

UMAP_2

30

A

10

20

Tissue Dissociation and Nuclei Isolation

Single-Nuclei RNA Sequencing

30

30

-20

-10

0

10

20

30

Adrenal Adenoma n = 2

Metastatic ACC n = 4

UMAP_1

Specimen Collection

Clustering and Gene Expression Analysis

B

C

Type

Normal Adrenal

Adrenal Adenoma

Primary ACC Endothelial Cells

Metastatic ACC Endothelial Cells

15-

Androstenedione

Cortisol DHEAS

Endotheli

Cells

Endothel

Cells Fibroblasts

Fibroblasts

Fibroblasts

10

Fibroblasts

NORMAL

AA 1

AA 2

PACC 1

PACC 2

PACC 3

PACC 4

Testosterone

Myeloid Col

Myekild Cells

Myeloid Cells

Myoloid Cels

MACC 1

MACC 2

MACC 3

MACC 4

AC 1.5

1.5

UMAP_2

5

AC 1,5

AC 1.6

AC 1.6

AC 1.2

AC 1.2

AC 1.2

1.4

Type

Androstenedione

DHEAS

AC 1.4

AC 1.2.

0

AC 1.4

Normal Adrenal

☐ Yes

Yes

AC 1.1

AC

1.

AC 1.1

AC 1.1

Adrenal Adenoma

☐ No

☐ No

-5

AC 1.3

AC 1.3

AC 1.3

Primary ACC Metastatic ACC

Cortisol

Testosterone

Lymphocytes

Lymphocytes

Lymphocytes

☐ Yes

☐ Yes

Lymphocytes

-10

☐ No

☐ No

-5

0

5

10

-5

0

5

10

-5

0

5

10

-5

0

5

10

UMAP_

D

E

100

IGFBP7

SERPINE1

Average Expression

Clusters

VWF

2

1

75

AC 1.1

AC 1.2

FLT1

0

Genes

Percentage of Cells

AC 1.3

PTPRC

-1

AC 1.4

ITGAM

Percent Expressed

50

AC 1.5

NR5A1

0

AC 1.6

25

Myeloid Cells

MLANA

50

Lymphocytes

75

25

Endothelial Cells

INHA

Fibroblasts

AC 1.1

AC 1.2

AC 1.3

AC 1.4

AC 1.5

AC 1.6

Myeloid Cells

Lymphocytes Endothelial Cells

Fibroblasts

0

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

Clusters

1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101

Figure 2. Single-Nuclei RNA-Sequencing Uncovers Infiltrating Immunocyte Populations. A. UMAP projection of subset lymphocyte and macrophage clusters from Figure 1C. B. Bar plot display of percent immunocyte composition per sample type. C. Dot plot displaying key differentially expressed cell typing markers. Scaled average expression is displayed. D. UMAP projection of subset macrophage clusters from Figure 2A. E. Bar plot display of percent macrophage subtypes per sample type. F. Dot plot displaying key differentially expressed cell typing markers. Scaled average expression is displayed. G. Heatmap displaying average expression of key macrophage immune checkpoint ligands for each sample type. Scaled average expression is displayed.

A

B

100%

5

Macrophages

75%

Macrophages

0

CD8 T cells

UMAP_2

Regulatory, T cells

COB Ticets

Regulatory T cells

A

Granulocytes

50%

CD4 T cells

-5

Natural Kiter cells

Granulocytes

CO47

cells

Natural Killer cells

Plasma cells

-10

25%

B cells

B cells

Plasma cells

-15

0%

-10

-5

0

5

10

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

UMAP_1

C

Macrophages

D

Percent Expressed

5

Granulocytes

0

Regulatory T cells

25

50

CD4 T cells

75

100

UMAP_2

0

CD8 T cells

Average Expression

Natural Killer cells

2

1

-5

B cells

0

Plasma cells

-1

SPP1+ Macrophages

FPP=

MS4A6

SIGPE SIGLES

VSIG.

V

HLA DE

APO

SP

IGA

ANDE

FORD

RCA

CAN

ANKF

BAN

CRI

STAP

JCHAIN

TNFRSE

SDC1

-5

0

5

UMAP_1

10

15

E

100%

F

G

CD274

M1 Macrophages

Average Expression

75%

1.0

CD80

Expression

0.5

1.0

CD40

0.5

M1 Macrophages

0.0

-0.5

0.0

50%

M2 Macrophages

M2 Macrophages

-0.5

-1.0

1.0

SPP1+ Macrophages

LGALS9

Percent Expressed

25%

25

PVR

50

SPP1+ Macrophages

75

SIRPA

0%

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

CD80

CD86

FCGR1A

MRC1

CD163

MARCO

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

SPP1

1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116

Figure 3. Single-Nuclei Sequencing Identifies Cancer-Associated Fibroblasts in ACC. A. UMAP projection demonstrating that further subclustering of fibroblasts identified in Figure 1C results in identification of 5 fibroblast clusters, denoted 3.1 through 3.5. B. Bar plot demonstrating cell type proportions of each of the identified fibroblast clusters ( * - FDR <= 0.05, ***- FDR <= 0.001), C. Dot plot demonstrating top marker genes which were used to annotate the identified clusters. Top 5 markers per cluster are displayed. D. Dot plot demonstrating expression of known of fibroblast marker genes. For both dot plots, scaled average expression is displayed. E. Heatmap showing expression of selected fibroblast markers in fibroblast cluster 3.1. For the heatmap, scaled average expression is displayed. FDR - false discovery rate.

A

B

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

Fibroblasts 3.4

5

Fibroblasts

Fibroblasts 3.4

*

Sample Type

Fibroblasts

100

90

Normal Adrenal

Fibroblasts 3.3

80

Adrenal Adenoma

UMAP_2

0

Percent

70

Primary ACC

---

Metastatic ACC

Fibroblasts 3.3

:

Fibroblasts 3.3

60

Fibroblasts 3.5

50

40

-5

Fibroblasts 3.1

Fibroblasts 3.1

Fibroblasts 3.1

30

---

Fibroblasts 3.1

20

*

*

10

0

5

10-15

-10

0

5

10-15

-5

0

5

10-15

0

5

10

0

-15

-10

-5

-5

-10

-10

-5

Fibroblasts 3.2

Fibroblasts 3.4

UMAP_1

Fibroblasts 3.1

Fibroblasts 3.3

Fibroblasts 3.5

Clusters

C

D

CAV1

SLPI

E

CD74

AFF3 -

S100A4

Fibroblasts 3.1

SHISA6

HAS1

PTGDS

AGTR1

LRP1B TENM2

Percent Expressed

LMNA

PDGFRA

0

DORO COLIAS

CADM2

DPT

ZNF385D

FAP

MĘCOM

25

CFD

CCL2

POPN

ERBB4

50

CXCL12

Percent Expressed

Identity

DEN

NRXN1

75

CXCL2

0

ACTS

ACTA2

Genes

NEGR1

100

CXCL1

MUDS

Average

CPM-

AC020637.1

CXCL8

25

50

HOPX

NRK

LINCO2388 PDE4B

Average Expression

Genes

IL6

POSIN IPMI

HAS2

75

TPM2

Expression

LRRC15

Average Expression

CON2

LARG15

1.5

PLXNA4

CCN2

1

HAS2 IL6

1.0

STEAP4

IL1RL1

TPM2

CXCL8

TPM1

1

GXCL1

0.5

FGF14

POSTN HOPX

CXCL2

PDE3A

0

0

CXCL12 CCL2

0.0

SGIP1

-0.5

MMP11

-1

CFD

EPHA3

DPT

-1.0

PLA2G5

TAGLN

LUNA

FN1

ACTA2

AGTRI HAS!

Fibroblasts 3.1

Fibroblasts 3.2

Fibroblasts 3.3

Fibroblasts 3.4

Fibroblasts 3.5

VIM

S10044

CO74

DCN

SIPI

PDPN

CAVI

FAP

COL1A1

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

Clusters

DDR2

PDGFRB

PDGFRA

Fibroblasts 3.1

Fibroblasts 3.2

Fibroblasts 3.3

Fibroblasts 3.4

Fibroblasts 3.5

Clusters

1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132

Figure 4. Single-Nuclei RNA Sequencing Demonstrates Intratumoral Heterogeneity of Adrenal Cortex Cells. A. UMAP projection showing that further subclustering of adrenal cortex cells results in identification of 7 adrenal cortex clusters, denoted AC 3.1 through AC 3.7. B. Dot plot demonstrating top 5 marker genes which were used to annotate each of the identified clusters. For the dot plot, scaled average expression is displayed. C. Bar plot showing cell type proportions of each of the adrenal cortex clusters ( *** - FDR <= 0.001). D. Bar plot demonstrating cell type proportions of each of the identified clusters on a per-sample basis. E. Heatmap demonstrating average module scores for a selection of Hallmark pathways downloaded from MSigDB for each of the identified clusters. In the heatmap, average module scores are scaled on a per-row basis.

A

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

6

AC 3.59

AC 3.1

AC 3.5

3

AC 3.4

AC 3.1

AC 3.1

AC 3.4

AC 3.4

AC 3.1

UMAP_2

AC 3 4

AC 3 5

0

AO 32

AC 3.2 AC 3.6

AC 3.7

AC

AC 36

AC 3.3

AC 3.6

-3

AC 3.2

AC 3.2

AC 3.3

AC 33

-4

0

4

8

-4

0

4

8

-4

0

4

8

-4

0

4

8

UMAP_1

B

AL136962.

C

NRG3

100

TSPAN8

NRXN3

TPX2

MIR924HG

ASPM

75

DIAPH3

APOLD1

Percent Expressed

Clusters

Percentage of Cells

AC 3.1

PURPL

CENPF

*

0

AC 3.2

50

AC 3.3

C9

25

AC 3.4

PHEX

AC 3.5

FYB1

50

AC 3.6

MTRNR2L12

75

AC 3.7

3

RPL41

25

RPL28

100

0

RPS27

RPL36

Average Expression

AFF3

PLPPR1

2

0

AC104078.2

NORMAL

AA 1

AA 2

PACC 2

PACC 3

PACC 4

MACC 2

MACC 3

MACC 4

PACO 1

MACC 1

NKD1

1

CADPS

Sample

FILIP1L

0

PDE10A

ATP1B3

E

NOTCH SIGNALING

Scaled Average Module Score Per Cluster

AC073593.2

IL6 JAK STAT3 SIGNALING

PGAP1

ΗΥΡΟΧΙΑ

ZBTB16

UV RESPONSE DOWN

PRKCA

IL2 STAT5 SIGNALING

CCBE1

ESTROGEN RESPONSE EARLY

ESTROGEN RESPONSE LATE

2

HS6ST3

ACSM3

CHOLESTEROL HOMEOSTASIS

AC 3.3

AC 3.4

1

AC 3.1

AC 3.2

AC 3.5

AC 3.6

AC 3.7

XENOBIOTIC METABOLISM

MYC TARGETS V2

MTORC1 SIGNALING

PEROXISOME

0

Clusters

TNFA SIGNALING VIA NFKB

INFLAMMATORY RESPONSE

-1

WNT BETA CATENIN SIGNALING

D

40

HEDGEHOG SIGNALING

KRAS SIGNALING UP

-2

OXIDATIVE PHOSPHORYLATION

Sample Type

MYC TARGETS V1

P53 PATHWAY

30

Normal Adrenal

UV RESPONSE UP FATTY ACID METABOLISM

Adrenal Adenoma

APOPTOSIS

Percent

Primary ACC

REACTIVE OXYGEN SPECIES PATHWAY

ANGIOGENESIS

20

Metastatic ACC

EPITHELIAL MESENCHYMAL TRANSITION

TGF BETA SIGNALING

UNFOLDED PROTEIN RESPONSE

GLYCOLYSIS

E2F TARGETS

10

G2M CHECKPOINT

MITOTIC SPINDLE

---

PI3K AKT MTOR SIGNALING

---

DNA REPAIR

KRAS SIGNALING DOWN

PROTEIN SECRETION

0

AC 3.1

AC 3.2

AC 3.3

AC 3.4

AC 3.5

AC 3.6

AC 3.7

AC 3.1

AC 3.2

AC 3.3

AC 3.4

AC 3.5

AC 3.6

AC 3.7

Clusters

Clusters

1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145

Figure 5. DNA Repair and Replication Pathways Are Enriched in Cluster AC 3.6. A. Heatmap demonstrating average module scores for REACTOME replication stress and DNA repair pathways downloaded from MSigDB. Clusters with fewer than 50 cells were excluded from the analysis. In the heatmap, average module scores are scaled per row. B. Bar plot demonstrating Ingenuity Pathway Analysis pathways upregulated and downregulated in cluster AC 3.6. Pathways are ordered by the z-score, and top 10 upregulated and top 3 downregulated pathways are displayed. Only 3 pathways were found to be downregulated in cluster AC 3.6. C. Bar plot showing total numbers of copy number variations (CNVs) inferred by InferCNV on a per-cluster basis. Chi-square test was used to calculate statistical significance ( **** - FDR <= 0.0001). D and E. Bar plots showing the total number of gains (D) and losses (E) inferred by inferCNV on a per-cluster basis. For both plots, chi-square test was used to calculate statistical significance ( **** - FDR <= 0.0001). F. Representative immunofluorescence images of formalin fixed paraffin embedded (FFPE) tissues from two patients stained for yH2AX (green), pRPA (red), and DAPI nuclear stain (blue). Scale bars, 10 um. Zoom-in panels show details of nuclei, gH2AX foci and pRPA foci.

A

Normal

Benign

Primary ACC

Metastatic ACC

Cluster

3

Cluster

ACTIVATION OF ATR IN RESPONSE TO REPLICATION STRESS

AC 3.1

2

AC 3.2

HOMOLOGY DIRECTED REPAIR

1

AC 3.3

AC 3.4

HDR THROUGH HOMOLOGOUS RECOMBINATION HRR

0

AC 3.5

HDR THROUGH MMEJ ALT NHEJ

AC 3.6

-1

AC 3.7

HDR THROUGH SINGLE STRAND ANNEALING SSA

-2

NONHOMOLOGOUS END JOINING NHEJ

-3

MISMATCH REPAIR

NUCLEOTIDE EXCISION REPAIR

BASE EXCISION REPAIR

B

C

D

E

All CNVs

300

Gains

Losses

Mitotic Prometaphase

200

100

Cell Cycle Checkpoints

250

---

---

---

Ingenuity Canonical Pathways

Total Number of CNVs

Mitotic Metaphase and Anaphase

200

CNV Type

150

Gain of > 2 oopies Gain of 2 copies

Total Number of CNVs

Total Number of CNVs

Mitotic G2-G2/M phases

150

Gain of 1 copy

100

50

RHO GTPases Activate Formins

Loss of 1 copy

100

Complete loss

RHO GTPase cycle

50

50

Cilium Assembly

HDR through HRR or SSA

0

0

0

AC 3.

AC 3.2

AC 3.3

AC 3.4

AC 3.5

AC 3.6

AC 3.7

AC 3.1

AC 3.2

AC 3.3

AC 3,4

AC 3.5

AC 3.6

AC 3.7

AC 3.

AC 3.2

AC 3.3

AC 3.4

AC 3.5

AC 3.6

AC 3.7

Kinesins

Clusters

Clusters

Clusters

Synthesis of DNA

MHC class II antigen presentation

F

DAPI

YH2AX

pRPA

Merged

Metabolism of steroid hormones

PPARalpha/RXRalpha Activation

PXR/RXR Activation

MACC 3

-2.5

0.0

2.5

5.0

7.5

z-score

MACC 2

1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162

Figure 6. NCI-H295R Adrenocortical Cell Line Exhibits DNA Replication Stress Response. A. Immunofluorescence demonstrating that DNA damage or replication stress is present in NCI-H295R cells in the absence of cytotoxic agents as determined by the presence of intranuclear H2AX foci. In addition to the anti- H2AX antibody, the cells were stained with DAPI as nuclear stain and phalloidin. Images were obtained using a 20X objective on a Nikon® AXR confocal microscope. Insets represent a cropped portion of the same fields at a 5X higher magnification. B through F. Water-soluble tetrazolium dye cell viability assays demonstrating sensitivity of the NCI-H29R5 cell line to various drugs targeting replication stress and DNA damage repair pathways: elimusertib and VX-803 (ATR inhibitors, 6B), AZ32 and KU60019 (ATM inhibitors, 6C), rabusertib and PF47736 (CHK1 inhibitors, 6D), RI-1 and SAHA (RAD51 inhibitors, 6E) and olaparib (PARP inhibitor, 6F). For each of the drug treatments, top panels display raw data and bottom panels display 4-parameter non-linear lines of best fit. G. Schematic summarizing replication stress and DNA damage repair pathways and drug treatments targeting those pathways. nd - non-determined. SAHA -

A

DAPI

YH2AX

Phalloidin

Merged

G

RPA-coated ssDNA (ex: replication arrest/replication stress)

DSBs

SSBs

YH2AX

SSB

YH2AX

repair

PARP

ATM

Olaparib -

ATR

AZ32, KU60019 -

Elimusertb, VX-803 +++

CHK2

CHK1

B

ATR inhibition

C

ATM inhibition

Rabusertib

PF47736

Elim usertib

VX-803

AZ 32

KU60019

175

175

175

175

ns

Cell cycle arrest and pathway decision point

% viability

150

125

% viability

150

150

150

125

% viability

125

% viability

125

·

100

100

100

100

-

I

S

7 5

75

5

+

s .

NHEJ

HR

Apoptosis

0

25

RI-1, SAHA + +/-

·

.

0

0

DMSO

… . M

DMSO

1 .M

10 .M

100 MM

DMSO

1 .M

10 .M

100 nM

… . M

DMSO

1 .M

… . M

Elim usertib

VX803

AZ32

KU60019

150

0 10 = 23.17 nM

150

10 4 = 41.97 MM

150

150

% viability

% viability

% viability

% viability

100

1.@

100

100

5 0

…

…

-1

.

1

2

,

4

%

+1

Q

1

2

,

4

5

.1

0

¥

2

*

4

S

·1

·

1

2

J

4

$

Concentration (log mM )

Concentration (log nM )

Concentration (log nM )

Concentration (log nM )

D

CHK1 inhibition

E

Rad51 inhibition

F

PARP inhibition

Rabusertib

PF 47736

RI-1

SAHA

O lap arib

175

175

175

175

175

150

150

150

ns

% viability

150

150

125

viability

125

% viability

125

% viability

% viability

125

100

100

100

100

100

75

75

7 5

A

s

50

*

25

25

25

25

0

.

0

0

0

DMSO

198 nM

DMSO

1 RM

18 .M

100 .M

1000 nM

DMSO

198 nM

1008 mM

DMSO

1 .M

10 AM

100 mM

DMSO

10 AM

100 mM

1

Rabusertib

PF 47736

R I-1

SAHA

O laparib

150

Css * - 1390 &M

150

1C .. . - 7842 .M

150

150

C.s . - 1318 &M

150

% viability

% viability

% viability

% viability

% viability

100

F

100

100

100

100

5 0

-1

·

1

2

3

4

%

-1

·

1

2

,

4

3

-1

Đ

”

2

J

4

5

-1

@

1

2

3

4

$

-1

Đ

1

2

1

4

5

Concentration (log mM )

Concentration (log nM )

Concentration (log nM )

Concentration (log nM )

Concentration (log nM )

1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178

1179 1180

suberoylanilide hydroxamic acid, ns - not significant, * - p<0.05, ** - p < 0.01, *** - p<0.001, **** - p<0.0001.

Figure 7. ACC PTOs exhibit DNA Replication Stress Response. A. ACC PTOs proliferate in culture long-term. Brightfield images of two ACC PTO lines (OSU1 and OSU2) were taken using a 20X objective >6 months after original tumor cell isolation. B. Immunofluorescence demonstrating that ACC PTOs express steroidogenic factor 1 (SF-1) and also showing that DNA damage or replication stress is present in ACC PTOs in the absence of cytotoxic agents as determined by the presence of intranuclear [H2AX foci. In addition to the anti- H2AX ☐ antibody and anti-SF-1 antibody, the cells were stained with DAPI as nuclear stain and phalloidin. Images were taken using a 20X objective on a Nikon® AXR confocal microscope. Insets represent a cropped portion of the same fields at a 5X higher magnification. C. ELISA demonstrates that ACC PTOs secrete cortisol. D. Water-soluble tetrazolium dye cell viability assays demonstrate that ACC PTOs exhibit sensitivity to ATR but not ATM inhibition. For each of the drug treatments, left panel displays raw data and right panel displays 4-parameter non-linear lines of best fit. nd - not determined, ns - not significant, * - p<0.05, ** - p < 0.01, *** - p<0.001, **** - p<0.0001.

A

B

OSU1

OSU2

DAPI

SF-1

YH2AX

Phalloidin

Merged

OSU1

C

Cortisol (ng/m L)

2000

1500

OSU2

1000

300

200

1

1P

100

.

NCI-H 295R

OSU1 PTOS

OSU2 PTOS

D

ATR inhibition

ATM inhibition

OSU1

OSU2

OSU1

OSU2

VX-403

Elin usertib

VX-403

AZ32

KU …

AZ32

…

…

…

…

…

…

… ..

…

…

…

…

…

… ..

…

… … .

… ..

E lim u sertib

VX-803

1 … . …

150

VX-803

AZ32

KU60019

AZ32

1507

% viability

” viability

% viability

s visbility

% wish Hity

sa

$-1

Concentration (log nM )

Concentration (log nM )

Concentration (leg nte )

Concentration (log at )

Concentration (log aM )

Concentration (log aM )

1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196

bioRxiv preprint doi: https://doi.org/10.1101/2024.09.30.615695; this version posted October 28, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1198 1199

SUPPLEMENTARY FIGURES

Adrenal Cortex Module Score

Supplementary Figure S1. Identifying Adrenal Cortex Cells. To identify adrenal cortex clusters after initial integration, dimensionality reduction, and clustering, average module scores were calculated based on the expression of ACC markers NR5A1, MLANA, and INHA for each of the identified clusters. Clusters classified as adrenal cortex had higher average module scores compared to other cell types. Clusters are ordered left to right based on decreasing average module scores.

2.0

1.5

AC 1.6

1.0

AC 1.2

AC 1.1

AC 1.4

AC 1.5

AC 1.3

0.5

Lymphocytes

Endothelial Cells

Myeloid Cells

Fibroblasts

0.0

-0.5

AC 1.6

AC 1.2

AC 1.1

AC 1.4

AC 1.5

AC 1.3

Lymphocytes

Endothelial Cells

Myeloid Cells

Fibroblasts

Clusters

1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215

Supplementary Figure S2. Immune Ligands Expressed by CD8+ T cells, CD4+ T cells, and Natural Killer Cells. A. Heatmaps demonstrating expression of ligands by CD8+ T cells (A), CD4+ T cells (B), and natural killer cells (C).

A

B

CD8+ T cells

CD4+ T cells

CD47

CTLA4

CD40LG

CSF1

CD28

TNFRSF18

CSF1

Expression

HAVCR2

Expression

LAG3

1.0

ICOS

TIGIT

0.5

1.0

0.0

PDCD1

0.5

CTLA4

-0.5

CD40LG

0.0

-0.5

TNFRSF18

-1.0

LAG3

-1.0

HAVCR2

TNFRSF4

PDCD1

TIGIT

GZMB

CD27

ICOS

GZMB

CD27

CD28

CD47

Normal Adrenal

Adrenal Adenoma

Malignant ACC

Normal Adrenal

Adrenal Adenoma

Malignant ACC

C

Natural killer cells

KLRC2

KLRC1

GZMA

HLA-DRB1

LAIR2

STAT1

Expression

GZMK

1.0

FLNA

0.5

EMP3

0.0

TIGIT

-0.5

GNLY

-1.0

CCL3

ACTB

SELL

SLFN5

CCL4

IGFBP7

TGFBR3

Normal Adrenal

Adrenal Adenoma

Malignant ACC

1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227

Supplementary Figure S3. snRNAseq Demonstrates Intratumoral Heterogeneity of Adrenal Cortex Cells. A. UMAP projection showing that subclustering of adrenal cortex cells results in identification of 7 adrenal cortex clusters and 1 adrenal medulla cluster in normal adrenal, adrenal adenoma, primary ACC, and metastatic ACC specimens. B. Dotplot demonstrating marker genes that were used to annotate each of the identified clusters. C. Violin plot demonstrating average module scores calculated based on ACC marker expression as in Supplementary Figure S1. Clusters are ordered left to right based on decreasing average module scores.

A

B

Normal Adrenal

Adrenal Adenoma

Primary ACC

Metastatic ACC

AC 2.2

CHGA

4

AC 2.2

Average Expression

AC 2.2

AC 22

2

AC 2.4

AC 2

AC 2.5

CAC 2.7

AC 24

AC 24

AC 2.5

1

UMAP_2

NR5A1

0

0

AC

Genes

-1

AC 2.1

AC 2.6

AC 26

AC 2.6

AC 2.5

AM/

AC 2.1

AM

-2

AC 2.1

AC 2.1

MLANA

Percent Expressed

0

-4

AC 2.3

AC 2,3

AC 2.3

AC 2.3

25

50

INHA

75

-8

-5

0

5

10

-5

0

5

10

-5

0

5

10

-5

0

5

10

UMAP_1

AC 2.1

AC 2.2

AC 2.3

AC 2.4

AC 2.5

AC 2.6

AC 2.7

AM

C

Clusters

Adrenal Cortex Module Score

2.0

1.5

1.0

AC 2.7

AC 2.2

AC 2.4

AC 2.1

AC 2.3

0.5

AC 2.5

AC 2.6

AM

0.0

-0.5

AC 2.7

AC 2.2

AC 2.4

AC 2.1

AC 2.3

AC 2.5

AC 2.6

AM

Clusters

1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245

Percent Expressed

HSD17B3

0

☒ 25

☒ 50

☒

Genes

HSD3B2

☒

☒

☒

☒

☒

☒

☒ 75

☒

100

CYP17A1

☒

☒

Average Expression

☒

☒

☒

☒

☒

1.5

1.0

0.5

CYP11A1

☒

☒

☒

☒

☒

☒

☒

0.0

-0.5

-1.0

AC 3.1

AC 3.2

AC 3.3

AC 3.4

AC 3.5

AC 3.6

AC 3.7

Clusters

1246 1247 1248

Supplementary Figure S4. Dotplot demonstrating expression of enzymes involved in 1249 testosterone, DHEAS, and androstenedione synthesis. Cluster AC 3.7 demonstrates higher expression of CYP11A1 and CYP17A1, which are involved in testosterone, DHEAS, and androstenedione production. For the dot plot, scaled average expression is displayed.

1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268

Normal

Benign

Primary ACC

Metastatic ACC

																					Cluster	3	Cluster
																					NOTCH SIGNALING		AC 3.1
																					IL6 JAK STAT3 SIGNALING	2	AC 3.2
																					HYPOXIA
																					UV RESPONSE DOWN	1	AC 3.3
																					IL2 STAT5 SIGNALING		AC 3.4
																					ESTROGEN RESPONSE EARLY
																					ESTROGEN RESPONSE LATE	0	AC 3.5
																					CHOLESTEROL HOMEOSTASIS		AC 3.6
																					XENOBIOTIC METABOLISM	-1
																					MYC TARGETS V2		AC 3.7
																					MTORC1 SIGNALING
																					PEROXISOME	-2
																					TNFA SIGNALING VIA NFKB
																					INFLAMMATORY RESPONSE	-3
																					WNT BETA CATENIN SIGNALING
																					HEDGEHOG SIGNALING
																					KRAS SIGNALING UP
																					OXIDATIVE PHOSPHORYLATION
																					MYC TARGETS V1
																					P53 PATHWAY
																					UV RESPONSE UP
																					FATTY ACID METABOLISM
																					APOPTOSIS
																					REACTIVE OXYGEN SPECIES PATHWAY
																					ANGIOGENESIS
																					EPITHELIAL MESENCHYMAL TRANSITION
																					TGF BETA SIGNALING
																					UNFOLDED PROTEIN RESPONSE
																					GLYCOLYSIS
																					E2F TARGETS
																					G2M CHECKPOINT
																					MITOTIC SPINDLE
																					PI3K AKT MTOR SIGNALING
																					DNA REPAIR
																					KRAS SIGNALING DOWN
																					PROTEIN SECRETION

Supplementary Figure S5. Hallmark Pathway Analysis Identifies Gene Expression Signatures on a Per-Cluster Level. Heatmap demonstrating average module scores for a selection of Hallmark pathways downloaded from MSigDB for each of the identified clusters split by sample type, expanded from Figure 2. Clusters with fewer than 50 cells were excluded from analysis. In the heatmap, average module scores are scaled on a per-row basis.

1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289

Supplementary Figure S6. DNA Repair Pathways are Enriched in Cluster AC 3.6. A. Heatmap demonstrating average module scores for a selection of REACTOME pathways downloaded from MSigDB for all adrenal cortex clusters split by individual sample, extended from Figure 5. Clusters with fewer than 50 cells were removed from analysis. In the heatmap, average module scores are scaled on a per-row basis.

Normal

AA 1

AA 2

PACC 1

PACC 2

PACC 3

PACC 4

MACC 1

MACC 2

MACC 3

MACC 4

Cluster

4 Cluster AC 3.1 AC 3.2 AC 3.3 AC 3.4 AC 3.5 AC 3.6 AC 3.7

ACTIVATION OF ATR IN RESPONSE TO REPLICATION STRESS

3

HOMOLOGY DIRECTED REPAIR

2

HDR THROUGH HOMOLOGOUS RECOMBINATION HAR

0

HDR THROUGH MMEJ ALT NHEJ

HDR THROUGH SINGLE STRAND ANNEALING SSA

-2

NONHOMOLOGOUS END JOINING NHEJ

-4

MISMATCH REPAIR

NUCLEOTIDE EXCISION REPAIR

BASE EXCISION REPAIR

1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318

Supplementary Figure S7. H&E Staining of FFPE Patient Tissues. A. H&E staining of ACC with necrosis (sample MACC 2) under 10X magnification. B. H&E staining of ACC with necrosis (sample MACC 2) under 20X magnification. C. H&E staining of ACC (sample PACC 2) under 40X magnification. Mitoses are highlighted by green circles. D. H&E staining demonstrating ACC with lymphovascular invasion (sample MACC 3) under 10X magnification. E. H&E staining of ACC lung metastasis (sample MACC 3) under 1X magnification.

A

MACC 2

B

MACC 2

C

PACC 2

D

MACC 3

E

MACC 3

1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332

Supplementary Figure S8. Negative Controls for NCI-H295R Cell Line Immunofluorescence. NCI-H295R cells were processed alongside those shown in Figure 6 but without primary mouse anti-y-H2AX antibody. The green channel represents background produced by Alexa Fluor® 488-conjugated secondary antibody.

DAPI

AF488

Phalloidin

Merged

H295R cells

100 um

100 pm

100 pm

100 yım

1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360

Supplementary Figure S9. Isotype Controls ACC patient-derived tumor organoid (PTO) Immunofluorescence. ACC PTOs were processed alongside those shown in Figure 7, but instead with anti-rabbit (corresponding to SF-1) and anti-mouse (corresponding to y-H2AX) primary antibody isotype controls. The red and green channels background staining produced by anti-rabbit Alexa Fluor® 488-and anti-mouse- Alexa Fluor® 546 conjugated secondary antibodies.

Mouse isotype / AF488

Rabbit isotype / AF546

DAPI

Phalloidin

Merged

OSU1

100 g

100 pm

100 pm

100 um

100 gh

OSU2

100 g

100 pm

100 pm

100 ph

100 1

1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383

Supplementary Table S1A

Cell Type	Normal Adrenal	Adrenal Adenoma	Primary ACC	Metastatic ACC	Total per cell type
Macrophages	134	784	570	122	1610
Granulocytes	15	29	78	19	141
Regulatory T cells	14	12	186	0	212
CD4 T cells	31	51	88	3	173
CD8 T cells	44	79	656	10	789
Natural Killer cells	8	6	61	20	95
B cells	1	3	18	1	23
Plasma cells	0	0	72	0	72
Total per sample type	247	964	1,729	175

Supplementary Table S1B

Macrophage Subtype	Normal Adrenal	Adrenal Adenoma	Primary ACC	Metastatic ACC
M1 Macrophages	80	447	424	110
M2 Macrophages	54	325	118	6
SPP+ Macrophages	0	12	28	6

Supplementary Table S1. Immunocyte Counts per Sample Type. A. Immunocyte counts per sample type. B. Macrophage subtype counts per sample type.

1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398

Supplementary Table S2A

Cell Type	Normal	AA1	AA 2	PACC 1	PACC 2	PACC 3	PACC 4	MACC 1	MACC 2	MACC 3	MACC4	Total per cell type
Total Macrophages	134	339	445	13	20	306	231	7	9	28	78	1610
Granulocytes	15	6	23	6	8	37	27	1	1	11	6	141
Regulatory T cells	14	1	11	1	1	3	181	0	0	0	0	212
CD4 T cells	31	5	46	1	3	7	77	0	0	3	0	173
CD8 T cells	44	25	54	9	2	15	630	2	1	6	1	789
Natural Killer cells	8	2	4	2	6	5	48	1	1	14	4	95
B cells	1	1	2	0	0	1	17	0	0	1	0	23
Plasma cells	0	0	0	0	0	1	71	0	0	0	0	72
Total per sample	247	379	585	32	40	375	1282	11	12	63	89

Supplementary Table S2B

Macrophage Subtype	Normal	AA1	AA 2	PACC 1	PACC 2	PACC 3	PACC 4	MACC 1	MACC 2	MACC 3	MACC4
M1 Macrophages	80	199	248	10	13	213	188	7	6	21	76
M2 Macrophages	54	135	190	2	5	84	27	0	3	2	1
SPP+ Macrophages	0	5	7	1	2	9	16	0	0	5	1

Supplementary Table S2. Immunocyte Counts per Sample. A. Immunocyte counts per individual sample. B. Macrophage subtype counts per individual sample.

1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416

bioRxiv preprint doi: https://doi.org/10.1101/2024.09.30.615695; this version posted October 28, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Samples	Pathology Findings
PACC 2	Mitotic count 25/50HPF, Conventional ACC, Capsular, extragInadular and LVI, necrosis 10- 20%
MACC 2	Mitotic count 74/ 50 hpf; Oncocytic and myxoid morphology, necrosis (5-10%); capsular and extra-glandular invasion
MACC 3	Mitotic count 25/50HPF, Conventional ACC, Capsula invasion, necrosis 5-10%

1417

1418

1419

1420

Supplementary Table S3. Summary of Pathology Findings in ACC Specimens. Pathology findings are summarized for samples PACC 2, MACC 2, and MACC 3.