Germline EGFR variants are over-represented in adolescents and young adults (AYA) with adrenocortical carcinoma

Authors/Affiliations

Sara Akhavanfard1,2, Lamis Yehia1, Roshan Padmanabhan1, Jordan P. Reynolds3, Ying Ni1,4, Charis Eng 1,2,5,6,7*

1 Genomic Medicine Institute, Lerner Research Institute, Cleveland Clinic, Cleveland, OH

2 Cleveland Clinic Lerner College of Medicine, Case Western Reserve University, Cleveland, OH

3 Department of Pathology, Cleveland Clinic, Cleveland, OH

4 Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland, OH

5 Center for Personalized Genetic Healthcare, Cleveland Clinic Community Care and Population Health, Cleveland, OH

@ The Author(s) 2020. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

6 Germline High-Risk Cancer Focus Group, Cancer Prevention, Control & Population Research

Program, Case Comprehensive Cancer Center, Case Western Reserve University School of Medicine, Cleveland, OH

Department of Genetics and Genome Sciences, Case Western Reserve University School of Medicine, Cleveland, OH

*Correspondence

Charis Eng, MD, PhD, FACP

Genomic Medicine Institute, NE5-314

Lerner Research Institute

Cleveland Clinic

9500 Euclid Avenue, Cleveland, Ohio 44195

Email: engc@ccf.org Phone: (216) 444-3440 Fax: (216) 636-0655

UNCORRECTED MANUSCRIT

Downloaded from https://academic.oup.com/hmg/advance-article/doi/10.1093/hmg/ddaa268/6035190 by guest on 21 December 2020

Abstract

Adrenocortical Carcinoma (ACC) is a rare endocrine tumor with poor overall prognosis and 1.5- fold overrepresentation in females. In children, ACC is associated with inherited cancer syndromes with 50-80% of childhood-ACC associated with TP53 germline variants. ACC in adolescents and young adults (AYA) is rarely due to germline TP53, IGF2, PRKARIA, and MEN1 variants. We analyzed exome sequencing data from 21 children (<15y), 32 AYA (15- 39y), and 60 adults (>39y) with ACC, and retained all pathogenic, likely pathogenic, and highly prioritized variants of uncertain significance. We engineered a stable lentiviral-mutant ACC cell line, harboring an EGFR variant (p.Asp1080Asn) from a 21-year-old female without germline- TP53-variant and with aggressive ACC. We found that 4.8% of the children (P = . 004) and 6.2% of AYA (P < . 0001), all female participants, harbored germline EGFR variants, compared to only 0.3% of the control group. Expanding our analysis to the RTK-RAS-MAPK pathway, we found that the RTK genes have the highest number of highly prioritized germline variants in these individuals amongst all three arms of this pathway. We showed EGFR mutant cells migrate faster and are characterized by a stem-like phenotype compared to wildtype cells. While EGFR inhibitors did not affect the stemness of mutant cells, Sunitinib, a multi-receptor tyrosine kinase inhibitor, significantly reduced their stem-like behavior. Our data suggest that EGFR could be a novel underlying germline predisposition factor for ACC, especially in the Childhood-AYA (C- AYA) population. Further clinical validation can improve precision oncology management of this disease, which is known to have limited therapeutic options.

Introduction

Adrenocortical carcinoma (ACC) is a rare endocrine tumor of the cortex of the adrenal gland, with a poor overall prognosis despite current treatment regimens including surgery (1,2). ACC has an annual incidence of 1-2 per million individuals with a slight overrepresentation (1.5 to 1) in females (1,3). ACCs can occur at all ages with a median age of 40 to 50 years (4). These tumors can either secrete hormones and cause symptoms (functional) as with Cushing syndrome, male feminization, and female virilization, or can be non-functional and typically discovered as an incidental finding (4). Complete resection can cure ACCs at the earlier stages (I and II), which are confined to the adrenal gland, but they are not as effective in more advanced stages (III and IV), with a 5-year survival of 6 to 13 percent for individuals with stage IV disease (5-7).

ACCs can be sporadic or heritable. Previous molecular studies connected somatic alterations of multiple driver genes, including CTNNB1 (beta-catenin; MIM: 116806), TP53 (MIM: 191170), ZNRF3 (MIM: 612062), ATRX (MIM: 300032), TERT (MIM: 187270), PRKARIA (MIM: 188830), and RPL22 (MIM: 180474), to sporadic ACC tumorigenesis (7-10). Additionally, somatic whole-genome doubling has been shown to be associated with aggressive clinical outcomes (7). In children, ACC is mostly associated with heritable cancer disorders, like Li- Fraumeni syndrome (MIM: 609265), with 50-80% of children with ACC having TP53 germline loss-of-function variants, Carney complex (MIM: 160980) with PRKARIA germline loss-of- function variants with the distinctive primary pigmented nodular adrenocortical disease (PPNAD) features, and Beckwith-Wiedemann syndrome (MIM: 130650) with imprinting defect of 11p15.5 locus (IGF2 [MIM:147470], CDKN1C [MIM:600856], H19 [MIM: 103280]) (11- 13). ACC in adolescents and young adults (AYA) is rarely due to germline TP53, IGF2,

PRKAR1A, or MEN1 (MIM: 613733) variants and their somatic genomic profile is not as well defined (14). Additionally, germline loss-of-function variants in DNA mismatch repair associated genes, like MSH2 (MIM: 609309), MSH6 (MIM:600678), MLH1 (MIM: 120436), and PMS2 (MIM: 600259), have been arguably associated with ACCs in Lynch Syndrome (MIM: 120435) (15,16). Despite all these advances, there is still a lack of understanding regarding the genetic predisposition, pathogenesis, and treatment of ACCs, particularly in AYA populations.

Results

A twenty-one-year-old female with metastatic bilateral adrenocortical carcinoma presented with an episode of cholecystitis and pulmonary embolism a few years ago. Clinical routine sent a portion of the resected ACC for Foundation One somatic genetic testing, which revealed CTNNB1 (MIM:116806) c.39_42delinsGTT (p.Ala13fs), CDKN2A/B (MIMs: 600160,600431) loss, and FLCN (MIM:607273) W306* driver somatic variants. Germline testing for TP53 came back wildtype, ruling out Li-Fraumeni syndrome given the clinical context of family history absent for Li-Fraumeni component malignancies. In order to identify other possible germline predisposing factors for her bilateral ACC tumors, research germline whole-exome sequencing was performed, and highly prioritized candidates selected using our variant reduction pipeline (Supplementary Material, Table S1). As part of our cancer-related gene variants, we found two potentially deleterious germline variants in one pathway: a heterozygous variant in the auto- phosphorylation domain of EGFR (c.3238 G>A, p.Asp1080Asn), and another heterozygous variant in SOSI (MIM: 182530) gene (c.1795C>A, p.Pro599Thr), both belonging to the EGFR pathway (Supplementary Material, Fig. S1). Considering the double-hit in this pathway and existence of multiple targeted treatments available for EGFR, in the setting of our 21-year-old

ACC case’s non-response to the current treatment regimen, we prioritized the EGFR variant for further investigation (Table 1).

Germline EGFR variants are overrepresented in C-AYA with Adrenocortical Carcinoma

We combined the gene list of our prioritized germline and somatic variants involved in our ACC case and performed an Ingenuity pathway analysis (IPA) to evaluate the interactions among our mutated genes. The connection between our target genes and molecules in the IPA knowledge database formed the basis of this network construction. Our IPA-predicted top network comprised 13 of our genes (10 germline and 3 somatic gene variants) centering around EGFR (right-tailed Fisher Exact Test P = 1x10-27; Figs. 1A and 1B), suggesting that the disruption of the EGFR pathway could play a role in ACC predisposition.

To validate our findings with a larger dataset, we then analyzed germline exome data from 21 children (<15y), 32 AYA (15-39y), and 60 adult research participants (>39y) with ACC, originating from the Cleveland Clinic, TCGA and St. Jude’s datasets (Supplementary Material, Tables S2 and S3). We utilized the Integrative Variant Analysis platform to classify variants based on the American College of Medical Genetics and Genomics guidelines. We retained all pathogenic, likely pathogenic, and highly-prioritized variants of uncertain significance (VUS). We found that 4.8% of 21 children (P = . 004, OR=18.6, 95% CI= 2.5-139.9) and 6.2% of 32 AYA (P < . 0001, OR=24.7, 95% CI= 5.9-105.0), all-female participants, and 1.7% of 60 adult individuals with ACC (P = . 07, OR=6.2, 95% CI= 0.8-45.2), and overall, 3.5% of all 113 ACC cases (P < . 0001, OR=13.7, 95% CI= 5.0-37.6) harbored at least a highly prioritized VUS germline EGFR variant in the non-kinase-domain, compared to only 0.3% of non-TCGA ExAC control group (Table 2) (Supplementary Material, Table S4).

We performed the same analysis on an additional 1,386 individuals with non-ACC solid tumors from the St. Jude dataset, and showed that individuals with ACC tumors had the highest frequency of prioritized germline EGFR variants among other solid tumors (P = . 07, OR=6.9, 95% CI= 0.8-56.7), following by retinoblastoma, high-grade glioma, and Wilms tumor (Supplementary Material, Tables S5 and S6).

Biochemical Analysis of EGFR (c.3238 G>A, p.Asp1080Asn) Variant

We compared the amount of EGFR protein, in mutant and wildtype, in both stable (Fig. 2A) and transiently transfected (Fig. 2B) cell lines. We used immunofluorescence assay to visualize GFP- tagged EGFR (green fluorescent protein) in our cells. There was more prominent EGFR expression in our mutant HEK293T cells relative to the wildtype ones (Figs. 2A and 2B).

Protein extracted from whole-cell lysates of the mutagenized cell lines were used to confirm the activation of downstream signaling of the EGFR pathway. Total AKT and pAKT expression were higher in all the three compartments of the mutant cells: nucleus, cytoplasm, and membrane by Western blot (Figs. 2C and 2D) (Supplementary Material, Fig. S2A). The ratio of nuclear pAKT/AKT, without EGF stimulation, was higher (2.2 times) in EGFR mutant cells compared to the wildtype cells, but the difference was not statistically significant (P > .05) (Fig. 2D). Total STAT3 was higher in mutant cells, especially in the nucleus (Supplementary Material, Fig. S2B).

EGFR Mutant Cells Migrate Faster Compared to the Wildtype Cells

As our female CCF research participant’s tumor was very aggressive and metastasized to multiple sites in her body, we performed migration and invasion assays as well to evaluate the

effect of our EGFR variant. We performed these experiments using the transfected HEK293 and lentiviral stable cell lines of empty vector, wild-type, and p.Asp1080Asn EGFR plasmids using Boyden Chambers. Mutant cells showed increased migration even without EGF treatment (Fig. 3A). Invasion Assay showed no difference between cells with either genotype.

Evaluating Epithelial to Mesenchymal Transition (EMT) and Stem-like Nature of the

Mutant Cell Line

EMT and stem cell markers including FOXC2, CD133, PAX6, AFP, SOX17, GATA4, PDXI, MSX1, MAP2, OTX2, TP63, Brachyury, CDH1, and CDH2 were evaluated in P2, P4, and P7 wildtype and mutant cells following 4 hours serum starvation and no EGF treatment. Mutant cells characterized by having a significant stem-like phenotype compared to the wildtype cells (Fig. 3B).

Sunitinib Significantly Reduced the Stem-like Behavior of the Mutant Cells

To evaluate the effect of anti-EGFR treatment on the p.Asp1080Asn mutated cell line, we performed a cell survival assay using 10 uM Erlotinib, Lapatinib, Gefitinib, Pazopanib, Osimertinib, or Sunitinib, accompanying 100ng/ml/5min EGF treatment as described. Osimertinib, among all tested EGFR inhibitors, resulted in the highest growth inhibition of mutant cells (Fig. 3C). While EGFR inhibitors had no effect on the stemness of the mutant cells, Sunitinib, a multi-receptor tyrosine kinase inhibitor, and Rapamycin significantly reduced the stem-like behavior of those cells via inhibition of FOXC2 marker (two-sided Student’s t-test, Sunitinib P = . 00015, Rapamycin P = . 011) (Fig. 3D). The combination of Osimertinib and Sunitinib showed a promising effect on eradication of the mutant cells compared to the wildtype cells (luM concentration, two-sided Student’s t-test P = . 00027) (Fig. 3E).

Potential Involvement of the RTK-RAS Pathway in ACC Predisposition

Expanding our analysis to the RTK-RAS-MAPK pathway, we found that 55.7% (63 out of 113) of individuals with ACC showed to have highly prioritized germline variants in RAS-RTK (mainly RTK) genes compared to 24.4% (12979 out of 53,105) of our non-TCGA ExAC control group (children’s group, P < . 0001, OR=1193.5, 95% CI= 431.4-3302.0; AYA group, P < . 0001, OR=373.0, 95% CI= 183.0-760.4; Adult < 50 years-old, P < . 0001, OR=895 1, 95% CI= 311.3- 2574.0; Adult >= 50 years-old, P < . 0001, OR=248.6, 95% CI= 133.9-461.6), while there were no germline pathogenic TP53 variants detected in those ACC research participants. Interestingly, frequency of these RTK germline variants were highest in the AYA females with ACC (12 out of 26) compared to all the other age groups (P < . 0001, OR=319.7, 95% CI= 145.3-703.3). The presented data is at variant’s level and each individual may have more than one variant (Fig. 4) (Supplementary Material, Tables S7 and S8).

Discussion

Recent advancements in the understanding of the disrupted molecular pathways underlying ACC have guided the design of new clinical trials of targeted therapies; however, the results have not been promising, and the survival remains poor (11). While poorly studied, most genomic studies focus on somatic alterations. In 2016, eg, the TCGA ACC reported on the comprehensive somatic variant landscape of 91 adult-onset ACC, noting somatic variants in PRKARIA, RPL22, TERF2 (MIM: 602027), CCNE1 (MIM:123837), and NF1 (MIM: 613113) as drivers (7). As an

incidental fall-out of this somatic study, the authors looked at a targeted 177 cancer- predisposition genes, not systematically covering EGFR or other RTK genes, in their 91 ACCs.

Here, we agnostically evaluated 113 individuals with ACC, from all age groups, for all genes via exome, including the 114 RAS-RTK pathway genes and 89 additional genes compared to the previous study, and focusing on the germline.

Somatic dysregulation of the EGFR pathway through constitutive receptor activation or overexpression is a well-known mechanism in the development of many cancer types, including lung cancer, glioblastoma, colorectal and esophageal cancers, breast cancer, and head and neck cancers (17-19). EGFR mutant cancers have their own distinct epidemiological features compared to non-EGFR mutant cases, for example, EGFR somatic mutant non-small-cell lung cancer (NSCLC) cases are more likely to be seen in non-smoker, female, poor prognosis, and adenocarcinoma cases (20,21). Despite the well-established role of somatic EGFR variants in cancer, the impact of heritable EGFR variants in susceptibility, pathogenesis, and prognosis of human cancers, including ACCs, is not well known nor investigated. Most advancements in this regard belong to studies, mainly case reports, revealing an association between germline variants of EGFR and familial NSCLCs (21-25). Cheng et al. performed a metanalysis and showed that familial NSCLCs, like EGFR mutant sporadic NSCLCs, trended toward adenocarcinoma, females, non-smokers, occasionally harboring secondary somatic EGFR variants, and rarely multi-focal lesions (21), suggesting inherited susceptibility to a subset of lung cancers with germline alteration of EGFR (17,21-25). They proposed germline EGFR variants can be oncogenic by themselves, and the tumor progression can be considerably boosted with the coexistence of a second somatic EGFR activating variant (21,26,27).

In our study, following finding the germline EGFR pathway disruption in our 21-year-old female research participant with metastatic bilateral ACC, we evaluated the entire cohort for germline EGFR variants. The increased percentage of ACC individuals with germline EGFR variants in the AYA population suggested that, like children with ACC, AYA individuals with ACC seem to have a higher heritable fraction as well, but with an involvement of a distinct molecular pathway. Cancers, including ACC, in the AYA population are frequently overlooked, including off the main focus of pharmaceutical companies. Our study proposes EGFR, as a new predisposition gene, especially in AYA individuals harboring ACCs and may be a focused drug target.

SALE

Proliferative pathways can be initiated via EGFR signaling, and most of the known protooncogenes in ACC play downstream of EGFR. So, activation of EGFR is a crucial step in starting the signaling cascades in this cancer.(28) As proof-of-principle functional validation, we were able to confirm the damaging nature of the EGFR p.Asp1080Asn variant in our selected poor prognosis case with metastatic bilateral ACC. Our downstream analysis showed that our mutant EGFR selectively activates the STAT3 signaling pathway, perhaps in the absence of AKT upregulation, which enhances cell survival and activates anti-apoptotic pathways (29). Increased stemness of our mutant cells suggests these cancer cells have increased capability of self-renewal and dedifferentiation which can be responsible for the tumor migration, metastasis, yet relatively slow cell growth, and resistance to treatment (30). These findings together can explain the aggressive behavior and non-responsiveness to therapy of the EGFR mutant ACC in our research participant, especially towards the traditional cytotoxic therapies that depend on rapid cell doublings. Although our pre-clinical observations need to be further validated, our study suggests that the addition of Sunitinib to the EGFR inhibitor of choice can improve the

I

result of treatment by significantly reducing the stem-like behavior of tumor cells, and potentially decrease the risk of relapse and metastasis in those individuals. Osimertinib, among all four tested EGFR inhibitors showed the most promising inhibition. While this drug has been reported not to be effective in decreasing STAT3 upregulation due to mutant EGFR in lung cancer cells, this may not be true in ACC lines. Tissue-specificity, even for identical targeted therapies, must therefore be kept in mind.

Interestingly, not only EGFR but other RTK-pathway-related genes also showed the same pattern of germline involvement in our series (Fig. 4) (Supplementary Material, Table S7). Notably, the germline involvement of RAS-pathway-related genes was minimal compared to the RTK’s. These results are encouraging for a re-evaluation of the related targeted therapies in the AYA sub-group diagnosed with ACC, offering possibly improved treatment, replacing the extremely toxic current regimen of choice, specifically, mitotane as a single agent or combined with other cytotoxic drugs. PI3K/AKT/mTOR and RAS/RAF/MEK/ERK pathways have multiple promising targets, focusing on their anti-proliferative activity of PI3K/mTOR inhibitors in the treatment of ACCs (31-33).

Additionally, in our 21-year-old female research participant with metastatic bilateral ACC, we found a germline pathogenic variant in the phosphoglycerate dehydrogenase (PHGDH) gene that evolves in the production of a critical enzyme in the serine synthesis pathway. Somatic mutation of this gene has been previously shown to be associated with multiple cancers including melanoma, breast cancer, clear cell renal cell carcinoma, and bladder cancer (34-37). Since other

ACCs in our series did not have any prioritized variants in this gene, we did not peruse any further analysis on the detected PHGDH variant and its association with ACCs.

This study has some limitations. Because ACCs are very rare, the total number of patients evaluated in this study was limited and gathered from three different institutes. Therefore, the sequencing methods and quality of data were not exactly the same. Our functional analysis was performed using a lentivirus transduced stable HEK293T cell line, a human embryonic kidney with some component of neuronal and adrenal gland cells, and not a pure adrenal cell line.

In summary, our observations suggest that EGFR and possibly other RTK-pathway-related genes are novel underlying germline predisposition factors for ACC, especially in the C-AYA population. Identifying a targetable pathway for ACC has the potential to improve precision oncology management of this disease, which is known to have limited therapeutic options.

Materials and Methods

Research Participant Enrolment / Sample Selection

Genomic data for this study were obtained from three sources:

1. Cleveland Clinic Foundation (CCF): A 21-year-old female with metastatic adrenocortical carcinoma, presented to the Hematology-Oncology clinic, was enrolled in this study, under Cleveland Clinic-approved IRB protocol 8458. Electronic medical records (EMR) reviewed to confirm the final diagnosis and tumor type, using healthcare providers’ notes and reports.

2. St. Jude (StJ) Cloud: 21 of the participants’ germline/clinical data were attained from two datasets of the St. Jude Cloud, generated by St. Jude Children’s Research Hospital and McDonnell Genome Institute of Washington University School of Medicine, under legal agreement 4147653 (Supplementary Material, Table S2):

a. 19 participants from the Pediatric Cancer Genome Project (PCGP) (38)

b. Two participants from St. Jude Lifetime (SJLIFE) (39)

As controls, 1,386 non-ACC Tumors from both datasets of St. Jude were recruited and analyzed as well.

3. TCGA Dataset: Permission for access to the TCGA germline DNA-sequence and clinical data was approved by the database of Genotypes and Phenotypes (dbGaP Project 12067). A total of 91 individuals with ACCs were selected (31 AYA, 60 adult participants), and germline variants were extracted from the blood-derived WES (Supplementary Material, Table S3).

Sample Preparation and Sequencing

Genomic DNA was extracted from peripheral-blood leukocyte of CCF case (ReliaPrep Large Volume HT gDNA System / Promega, Madison, WI), by standard methods at the CCF Genomic Medicine Biorepository (Cleveland, OH). Nextera Rapid Capture Exome library prep kit (Illumina, San Diego, CA) was used to run Whole Exome Sequencing (WES) on the DNA sample. Qubit fluorometer was used for sample quantification and QC (Invitrogen, Carlsbad, CA) with the dsDNA broad-range assay kit and E-gel electrophoresis system (Invitrogen). A total of 50ng DNA input was used and sheared enzymatically. Libraries were validated using the Qubit dsDNA broad-range assay kit and evenly pooled using 500ng of each tagged library, using

ligating Illumina adapters and unique barcodes. Hybridization using Illumina capture probes was completed on the final pool and amplified by PCR. Validation of the final enriched library pool was completed using the Qubit fluorometer to derive concentration (ng/ul), Bioanalyzer for library quality and average bp size, and final quantification via qPCR (KAPA Biosystems, Illumina library quantification kit). Standard Illumina protocols for the HiSeq 2500 system were used for dilution of the final enriched pool, denaturation, and loading. Samples were run across two rapid run flowcells, 2x100bp (paired-end) run at an Illumina sequencing service center (Cleveland, OH).

Alignment and Variant Calling

The sequencing data of the CCF case was received in binary alignment map (BAM) format. Fast- Q file was re-generated, and raw reads were mapped to the human reference haploid genome sequence GRCH37/hg19 using Burrows-Wheeler Aligner (BWA v.0.6.1, RRID:SCR_010910, http://bio-bwa.sourceforge.net/) (40). Genome Analysis Toolkit (GATK 3.5, RRID:SCR_001876, https://software.broadinstitute.org/gatk/),(41, 42) Sequence Alignment/Map (SAMtools, RRID:SCR 002105). http://samtools.sourceforge.net/) (43), and Picard (RRID:SCR_006525, http:/broadinstitute.github.io/picard/) were used for indel-realignment, removal of PCR duplicates, and base- and quality-score recalibrations. GATK Haplotype Caller was used for variant calling of single-nucleotide variations (SNVs) and short (<50 bp) indels. In- house GATK pipeline produced a variant calling file.

Variant Classification

The variant annotation and interpretation analysis were generated through the use of Ingenuity® Variant Analysis™M (IVA) software (www.qiagenbioinformatics.com) from

Ingenuity Systems (version 5.4.20190121). We kept variants with call quality at least 20.0, read depth at least 20.0, genotype quality at least 30.0, and outside top 5.0% most exonically variable 100-base windows in healthy public genomes (1000 Genomes Project, RRID:SCR_006828, http://www.1000genomes.org/). Variants kept up to 20 bases to the intronic region if they were predicted to disrupt splicing by MaxEntScan (RRID:SCR_016707, http://genes.mit.edu/burgelab/maxent/Xmaxentscan_scoreseq.html) (44). Variants were excluded if the allele frequency was greater than or equal to 0.05% in any of the following population databases: 1000 Genomes Project (phase3v5b), NHLBI ESP exomes (ESP6500SI-V2, RRID:SCR_012761, http://evs.gs.washington.edu/EVS/), ExAC Frequency (0.3.1) ( RRID:SCR_004068, http://exac.broadinstitute.org/), and the gnomAD Maximum Frequency (2.0.1) (RRID:SCR_014964, http://gnomad.broadinstitute.org/). Variants with a Phred-scaled CADD (v1.3) (http://cadd.gs.washington.edu/info) (45) score <10, or tolerant SIFT prediction (2016-02-23) (RRID:SCR_012813, http://sift.bii.a-star.edu.sg/) were excluded as well unless there was an established pathogenic common variant. Subsequently, only variants that were classified by auto-classification of IVA, based on the American College of Medical Genetics and Genomics (ACMG, RRID:SCR_005769, http://www.acmg.net) guidelines, as pathogenic or likely pathogenic (P/LP), and retained variants of uncertain significance (VUS), which we call them highly prioritized VUS, were kept for further evaluation. The following databases were used in IVA classification: Allele Frequency Community (2018-09-06), RefSeq Gene Model (2018-07-10, RRID:SCR_003496, http://www.ncbi.nlm.nih.gov/RefSeq/), PolyPhen-2 (v2.2.2) (RRID:SCR_013189, http://genetics.bwh.harvard.edu/pph2/), PhyloP (2009-11), DbSNP (151)

(RRID:SCR_002338, http://www.ncbi.nlm.nih.gov/SNP/), TargetScan (6.2) (RRID:SCR_010845, http://targetscan.org/), GENCODE (Release 28) (RRID:SCR_014966,

Downloaded from https://academic.oup.com/hmg/advance-article/doi/10.1093/hmg/ddaa268/6035190 by guest on 21 December 2020

https://www.gencodegenes.org), CentoMD (5.0), Ingenuity Knowledge Base (Stepford 190106.000, http://www.ingenuity.com/), OMIM (2017-05-26) (RRID:SCR_006437,

http://omim.org), BSIFT (2016-02-23), TCGA (2013-09-05)

(RRID:SCR_003193,

http://cancergenome.nih.gov/),

Clin Var (2018-08-01) (RRID:SCR_006169,

http://www.ncbi.nlm.nih.gov/clinvar/),

DGV (2016-05-15)

(RRID:SCR_007000,

http://dgv.tcag.ca/),

COSMIC (v86)

(RRID:SCR_002260,

http://cancer.sanger.ac.uk/cancergenome/projects/cosmic/), HGMD (2018.3)

(RRID:SCR_001888, http://www.hgmd.cf.ac.uk/ac/index.php). P/LP variants were excluded if Clin Var called them benign or likely benign. All the prioritized variants were inspected through the Integrative Genomics Viewer (IGV, RRID:SCR_011793, http://www.broadinstitute.org/igv/) to rule out artifacts (46,47).

Variant analysis in the control population

As a control, we extracted germline exome data, for all the RTK-RAS-MAPK genes from 53,105 individuals from the exome aggregation consortium dataset (ExAC), excluding individuals belonging to The Cancer Genome Atlas (TCGA), known as non-TCGA ExAC dataset. We performed an independent IVA analysis, with the same parameters, on this data to compare the frequency of the prioritized germline EGFR and other RTK-RAS-MAPK variants between patients with ACC tumors and this non-cancer control population.

Pathway Analysis

Qiagen Ingenuity Pathway Analysis (IPA) (48) was used to generate networks for our prioritized genes. Networks were algorithmically generated based on the connectivity of our genes of

interest with other molecules existing in the Ingenuity’s knowledge base (version 5.4.20190121, RRID:SCR_008117, http://www.ingenuity.com/).

Sanger Sequencing Confirmation

We validated prioritized variants through Sanger sequencing. Gene-specific primers før each variant were designed. Targeted area sequenced using forward and reverse primers using a 3730xl DNA Analyzer. Mutation Surveyor DNA Variant Analysis Software (SoftGenetics, State College, PA, RRID:SCR_001247, http://www.softgenetics.com/mutationSurveyor.php) was used for the final analysis.

Cell Lines

HEK293T cells (Human embryonic kidney, neuronal and adrenal gland cells, ATCC Cat# CRL- 3216, RRID: CVCL_0063) were used for mutagenesis experiment, and were cultured in DMEM supplemented with 10% FBS and 1% penicillin and streptomycin. HEK293T cell line was purchased from ATCC after 2010, and has been molecularly authenticated by STRS analysis, and tested negative upon routine mycoplasma testing with the MycoAlert Mycoplasma Detection Kit (Lonza, Alpharetta, GA).

Plasmids, Mutagenesis, and Cell Line Transfection

EGFR (MIM: 131550) p.Asp1080Asn mutant plasmid was generated from GFP-tagged wild- type EGFR plasmid #32751 (Addgene, Watertown, MA, RRID:Addgene_32751) (49) via QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies). All the custom constructs tested with Sanger sequencing before transfection. Validated GFP-tagged EGFR wildtype, mutant, and empty vector plasmids used to transfect HEK293T cell lines using Lipofectamine

3000 reagent (Thermo Fisher Scientific, Grand Island, NY). Cells harvested after a minimum 24hrs of incubation. A lentivirus transduced stable cell line for the EGFR variant (empty vector, wildtype, mutant) generated under 5 mg/ml puromycin selection for 30 days based on approved CCF biosafety protocol.

RNA Extraction and qRT-PCR

RNA was extracted from the Lentiviral stable cell line (empty vector, wildtype, mutant EGFR) with the RNeasy Mini Kit (QIAGEN, Germantown, MD), purified with Turbo DNase treatment (Life Technologies, Grand Island, NY), and reverse-transcribed with Superscript III Reverse Transcriptase (Life Technologies). Primers corresponding to the target EMT and Stemness markers were used, and cDNA was quantified with SYBR Green (Life Technologies). The standard DDCT method was used to analyze the data.

Immunoblotting

Whole-cell lysates were used to extract protein using Mammalian Protein Extraction Reagent M- PER (Thermo Scientific Pierce, Grand Island, NY) complemented with a cocktail of protease and phosphatase inhibitors (Sigma-Aldrich, St. Louis, MO) and were quantified with the BCA Protein Assay Kit (Thermo Scientific Pierce). Nuclear, cytoplasmic, and membranous fractions were extracted using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific Pierce) and Mem-PERTM Plus Membrane Protein Extraction Kit (Thermo Scientific Pierce) according to the manufacturer’s instructions. Lysates were separated by SDS-PAGE and transferred onto nitrocellulose membranes. We probed for Antibodies against phospho-EGFR (Tyr1040), EGFR, phospho-ERK (Thr202/Tyr204), ERK, phospho-AKT (Ser473), AKT, phospho STAT3, and STAT3 (all from Cell Signaling Technology, Danvers, MA; 1:1000), and

GAPDH at 1:50,000 dilution. Blots were scanned digitally and quantified with the Odyssey Infrared Imaging System (Li-Cor Biosciences, Lincoln, NE).

Immunofluorescence

WT and mutant EGFR Cells (both transiently transfected and stable cell lines) were seeded on coverslips and were fixed with 4% paraformaldehyde for 10 min at room temperature. Coverslips were mounted with ProLong Gold Antifade mountant with DAPI (Invitrogen, Carlsbad, CA). Slides were visualized, and images were obtained with a TCS SP5 confocal microscope (Leica, Buffalo Grove, IL)

Growth Curve and Cell Viability

Countess Automated Cell Counter (Invitrogen) was used to count the trypsinized and homogenized cells. Trypan blue was used to assess cell viability. We also used the MTT [3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (Sigma-Aldrich) to assess cell viability according to the manufacturer’s instructions. Cell viability were accessed using following inhibitors: Sunitinib (Selleck USA, Houston, TX, Cat# S7781), Osimertinib (Selleck USA, Cat# S7781), Erlotinib (Cayman Chemical, Ann Arbor, MI, Cat# 10483), Lapatinib (Cayman Chemical, Cat# 11493), Gefitinib (Cayman Chemical, Cat# 13166), Pazopanib (Cayman Chemical, Cat# 12097), and Rapamycin (Sigma-Aldrich, Cat# R8781). All the inhibitors diluted based on manufacturer’s instructions.

In Vitro Migration and Invasion Assays

EGFR WT and mutant lentivirus transduced stable cell lines were grown for 24hr, serum-starved for 4 hours, and migration assay was performed using CytoSelectTM 24-Well Cell Migration

Assay (8um, Colorimetric) per manufacturer’s instructions (Cell Biolab Inc., San Diego, CA). CytoSelectTM 24-Well Cell Invasion Assay (Basement Membrane, Colorimetric) was used to perform the invasion assay based on the manufacturer’s recommendation. Cells that migrated or invaded were fixed with 4% paraformaldehyde for 5 min, permeabilized with 100% methanol for 20 min, and stained with Giemsa for 15 min all at room temperature. Cells were photographed under 10x magnification using Leica MZ16FA stereomicroscope.

Statistical Analysis

All data collected from the in vitro functional studies were analyzed by two-tailed Student’s t- test, and a P value <. 05 was considered statistically significant. All the error bars are demonstrating standard deviation for sample mean. All the p-values, odds ratios (ORs), and 95%-confidence intervals (CIs) were calculated with a two-sided Fisher’s Exact test, comparing the frequency of prioritized variants in our ACC groups versus non-TCGA ExAC control population.

Data Availability

Whole Exome data for C-AYA case with ACC from Cleveland Clinic deposited into the NCBI Sequence Read Archive (SRA) database under accession number PRJNA559601. Whole Exome data for O-AYA cases with ACC from St. Jude research hospital is accessible at

https://www.stjude.cloud/. Whole Exome data for C-AYA cases with ACC from_TCGA is

accessible at https://portal.gdc.cancer.gov/.

Genome Analysis Toolkit (GATK 3.5) (40), https://software.broadinstitute.org/gatk/

gnomAD Maximum Frequency (2.0.1) McArthur Lab, (https://www.biorxiv.org/content/10.1101/531210v2)

https://macarthurlab.org/2018/10/17/gnomad-v2-1/ CADD (v1.3) (45), http://cadd.gs.washington.edu/info SIFT prediction (2016-02-23) (50), https://sift.bii.a-star.edu.sg/

ClinVar (2018-08-01) (51), https://preview.ncbi.nlm.nih.gov/clinvar/ Integrative Genomics Viewer (IGV) (46,47), https://software.broadinstitute.org/software/igv/ Maftools (52),

https://bioconductor.org/packages/release/bioc/vignettes/maftools/inst/doc/maftools.html

Qiagen Ingenuity Pathway Analysis (IPA) (48), https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis/

OMIM, http://www.omim.org/

Author Contributions

Conceptualization: S.A., L.Y., and C.E .; Methodology: S.A., L.Y., R.P., T.R., Y.N., and C.E .; Variant Analysis & Investigation: S.A .; Data Curation: S.A., R.P .; Bioinformatic Analysis & Visualization: R.P., S.A .; Functional Analysis: S.A .; Data interpretation: S.A. and C.E .; Pathology slide review and interpretation: J.P.R .; Validation, Supervision & Funding Acquisition: C.E .; Writing - Original Draft, Review & Editing: S.A., L.Y., and C.E. All authors critically reviewed and approved the final manuscript.

Acknowledgments

We thank St. Jude Children’s Research Hospital for their generosity in making their genomic data publicly available, Peter Anderson MD, Kaitlin Sesock MS, LGC and Ryan Noss MS, LGC for their help in recruiting the CCF research participants, Phyllis Harbor (Genomic Medicine Biorepository) for her help with biorepositing and processing the samples, Madhav Sankunny PhD, and Ritika Jaini PhD, for helpful discussions, and Todd Romigh MS for technical advice in the early stages of this study (MS, RJ and TR: all Eng lab). This work was supported, in part, by the VeloSano Pilot Program (Cleveland Clinic Taussig Cancer Institute). L. Yehia is an Ambrose Monell Foundation Cancer Genomic Medicine Fellow at the Cleveland Clinic Genomic Medicine Institute. C. Eng is the Sondra J. and Stephen R. Hardis Chair of Cancer Genomic Medicine at the Cleveland Clinic and an American Cancer Society Clinical Research Professor.

Conflict of Interests Disclosures

The authors declare no competing interests.

UNCORRECTE

References

1 Bilimoria, K.Y., Shen, W.T., Elaraj, D., Bentrem, D.J., Winchester, D.J., Kebebew, E. and Sturgeon, C. (2008) Adrenocortical carcinoma in the United States: treatment utilization and prognostic factors. Cancer, 113, 3130-3136.

2 Else, T., Kim, A.C., Sabolch, A., Raymond, V.M., Kandathil, A., Caoili, E.M., Jolly, S., Miller, B.S., Giordano, T.J. and Hammer, G.D. (2014) Adrenocortical carcinoma. Endocr. Rev., 35, 282-326.

3 Ayala-Ramirez, M., Jasim, S., Feng, L., Ejaz, S., Deniz, F., Busaidy, N., Waguespack, S.G., Naing, A., Sircar, K., Wood, C.G. et al. (2013) Adrenocortical carcinoma: clinical outcomes and prognosis of 330 patients at a tertiary care center. Eur. J. Endocrinol., 169, 891-899.

4 Elfiky, A.A. and Krishnan Nair, H.K. (2016) Assessment and management of advanced adrenocortical carcinoma using a precision oncology care model. Discov. Med., 21, 49- 56.

5 Else, T., Williams, A.R., Sabolch, A., Jolly, S., Miller, B.S. and Hammer, G.D. (2014) Adjuvant therapies and patient and tumor characteristics associated with survival of adult patients with adrenocortical carcinoma. J. Clin. Endocrinol. Metab., 99, 455-461.

6 Fassnacht, M., Johanssen, S., Quinkler, M., Bucsky, P., Willenberg, H.S., Beuschlein, F., Terzolo, M., Mueller, H.H., Hahner, S., Allolio, B. et al. (2009) Limited prognostic value of the 2004 International Union Against Cancer staging classification for adrenocortical carcinoma: proposal for a Revised TNM Classification. Cancer, 115, 243-250.

7 Zheng, S., Cherniack, A.D., Dewal, N., Moffitt, R.A., Danilova, L., Murray, B.A., Lerario, A.M., Else, T., Knijnenburg, T.A., Ciriello, G. et al. (2016) Comprehensive Pan- Genomic Characterization of Adrenocortical Carcinoma. Cancer Cell, 29, 723-736.

8 Giordano, T.J., Thomas, D.G., Kuick, R., Lizyness, M., Misek, D.E., Smith, A.L., Sanders, D., Aljundi, R.T., Gauger, P.G., Thompson, N.W. et al. (2003) Distinct transcriptional profiles of adrenocortical tumors uncovered by DNA microarray analysis. Am. J. Pathol., 162, 521-531.

9 Tissier, F., Cavard, C., Groussin, L., Perlemoine, K., Fumey, G., Hagnere, A.M., Rene- Corail, F., Jullian, E., Gicquel, C., Bertagna, X. et al. (2005) Mutations of beta-catenin in adrenocortical tumors: activation of the Wnt signaling pathway is a frequent event in both benign and malignant adrenocortical tumors. Cancer Res., 65, 7622-7627.

10 Pinto, E.M., Chen, X., Easton, J., Finkelstein, D., Liu, Z., Pounds, S., Rodriguez-Galindo, C., Lund, T.C., Mardis, E.R., Wilson, R.K. et al. (2015) Genomic landscape of paediatric adrenocortical tumours. Nat. Commun., 6, 6302. ☒

11 Szyszka, P., Grossman, A.B., Diaz-Cano, S., Sworczak, K. and Dworakowska, D. (2016) Molecular pathways of human adrenocortical carcinoma - translating cell signalling knowledge into diagnostic and treatment options. Endokrynol Pol, 67, 427-450.

12 Memon, S.S., Thakkar, K., Patil, V., Jadhav, S., Lila, A.R., Fernandes, G., Bandgar, T.R. and Shah, N.S. (2019) Primary pigmented nodular adrenocortical disease (PPNAD): single centre experience. J. Pediat.r Endocrinol. Metab., 32, 391-397.

13 Zhang, C.D., Pichurin, P.N., Bobr, A., Lyden, M.L., Young, W.F. and Bancos, I. (2019) Cushing syndrome: uncovering Carney complex due to novel PRKAR1A mutation. Endocrinol. Diabetes Metab. Case Rep., 18, 0150.

14 Lerario, A.M., Moraitis, A. and Hammer, G.D. (2014) Genetics and epigenetics of adrenocortical tumors. Mol. Cel.l Endocrinol., 386, 67-84.

15 Karamurzin, Y., Zeng, Z., Stadler, Z.K., Zhang, L., Ouansafi, I., Al-Ahmadie, H.A., Sempoux, C., Saltz, L.B., Soslow, R.A., O’Reilly, E.M. et al. (2012) Unusual DNA mismatch repair-deficient tumors in Lynch syndrome: a report of new cases and review of the literature. Hum. Pathol., 43, 1677-1687.

16 Walczak, E.M. and Hammer, G.D. (2015) Regulation of the adrenocortical stem cell niche: implications for disease. Nat. Rev. Endocrinol., 11, 14-28.

17 Kim, J.S., Cho, M.S., Nam, J.H., Kim, H.J., Choi, K.W. and Ryu, J.S. (2017) Prognostic impact of EGFR mutation in non-small-cell lung cancer patients with family history of lung cancer. PLoS One, 12, e0177015.

18 Roskoski, R., Jr. (2014) The ErbB/HER family of protein-tyrosine kinases and cancer. Pharmacol. Res., 79, 34-74.

19 Wang, Z. (2017) ErbB Receptors and Cancer. Methods Mol. Biol., 1652, 3-35.

20 Shigematsu, H., Lin, L., Takahashi, T., Nomura, M., Suzuki, M., Wistuba, II, Fong, K.M., Lee, H., Toyooka, S., Shimizu, N. et al. (2005) Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J. Natl. Cancer Inst., 97, 339-346.

21 Cheng, Y.I., Gan, Y.C., Liu, D., Davies, M.P.A., Li, W.M. and Field, J.K. (2019) Potential genetic modifiers for somatic EGFR mutation in lung cancer: a meta-analysis and literature review. BMC. Cancer, 19, 1068.

22 Bell, D.W., Gore, I., Okimoto, R.A., Godin-Heymann, N., Sordella, R., Mulloy, R., Sharma, S.V., Brannigan, B.W., Mohapatra, G., Settleman, J. et al. (2005) Inherited susceptibility to lung cancer may be associated with the T790M drug resistance mutation in EGFR. Nat. Genet., 37, 1315-1316.

23 Centeno, I., Blay, P., Santamaria, I., Astudillo, A., Pitiot, A.S., Osorio, F.G., Gonzalez- Arriaga, P., Iglesias, F., Menendez, P,, Tardon, A. et al. (2011) Germ-line mutations in epidermal growth factor receptor (EGFR) are rare but may contribute to oncogenesis: a novel germ-line mutation in EGFR detected in a patient with lung adenocarcinoma. BMC. Cancer, 11, 172.

24 Gazdar, A., Robinson, L., Oliver, D., Xing, C., Travis, W.D., Soh, J., Toyooka, S., Watumull, L., Xie, Y., Kernstine, K. et al. (2014) Hereditary lung cancer syndrome targets never smokers with germline EGFR gene T790M mutations. J. Thorac. Oncol., 9, 456-463.

25 Lu, S., Yu, Y., Li, Z., Yu, R., Wu, X., Bao, H., Ding, Y., Shao, Y.W. and Jian, H. (2019) EGFR and ERBB2 Germline Mutations in Chinese Lung Cancer Patients and Their Roles in Genetic Susceptibility to Cancer. J. Thorac. Oncol., 14, 732-736.

26 van Noesel, J., van der Ven, W.H., van Os, T.A., Kunst, P.W., Weegenaar, J., Reinten, R.J., Kancha, R.K., Duyster, J. and van Noesel, C.J. (2013) Activating germline R776H mutation in the epidermal growth factor receptor associated with lung cancer with squamous differentiation. J. Clin. Oncol., 31, e161-164.

27 Lou, Y., Pecot, C.V., Tran, H.T., DeVito, V.J., Tang, X.M., Heymach, J.V., Luthra, R., Wistuba, II, Zuo, Z. and Tsao, A.S. (2016) Germline Mutation of T790M and Dual/Multiple EGFR Mutations in Patients With Lung Adenocarcinoma. Clin. Lung Cancer, 17, e5-11.

28 Rahane, C.S., Kutzner, A. and Heese, K. (2019) Establishing a human adrenocortical carcinoma (ACC)-specific gene mutation signature. Cancer Genet., 230, 1-12.

29 Sordella, R., Bell, D.W., Haber, D.A. and Settleman, J. (2004) Gefitinib-sensitizing EGFR mutations in lung cancer activate anti-apoptotic pathways. Science, 305, 1163- 1167.

30 Prasad, S., Ramachandran, S., Gupta, N., Kaushik, I. and Srivastava, S.K. (2020) Cancer cells stemness: A doorstep to targeted therapy. Biochim. Biophys. Acta. Mol. Basis Dis, 1866, 165424.

31 Maira, S.M., Stauffer, F., Brueggen, J., Furet, P., Schnell, C., Fritsch, C., Brachmann, S., Chene, P., De Pover, A., Schoemaker, K. et al. (2008) Identification and characterization of NVP-BEZ235, a new orally available dual phosphatidylinositol 3-kinase/mammalian target of rapamycin inhibitor with potent in vivo antitumor activity. Mol. Cancer Ther., 7, 1851-1863.

32 Mariniello, B., Rosato, A., Zuccolotto, G., Rubin, B., Cicala, M.V., Finco, I., Iacobone, M., Frigo, A.C., Fassina, A., Pezzani, R. et al. (2012) Combination of sorafenib and everolimus impacts therapeutically on adrenocortical tumor models. Endocr. Relat. Cancer, 19, 527-539.

33 Pezzani, R., Rubin, B., Redaelli, M., Radu, C., Barollo, S., Cicala, M.V., Salva, M., Mian, C., Mucignat-Caretta, C., Simioni, P. et al. (2014) The antiproliferative effects of ouabain and everolimus on adrenocortical tumor cells. Endocr. J., 61, 41-53.

34 Locasale, J.W., Grassian, A.R., Melman, T., Lyssiotis, C.A., Mattaini, K.R., Bass, A.J., Heffron, G., Metallo, C.M., Muranen, T., Sharfi, H. et al. (2011) Phosphoglycerate dehydrogenase diverts glycolytic flux and contributes to oncogenesis. Nat. Genet., 43, 869-874.

35 Nguyen, M.Q., Teh, J.L.F., Purwin, T.J., Chervoneva, I., Davies, M.A., Nathanson, K.L., Cheng, P.F., Levesque, M.P., Dummer, R. and Aplin, A.E. (2020) Targeting PHGDH Upregulation Reduces Glutathione Levels and Resensitizes Resistant NRAS-Mutant Melanoma to MAPK Kinase Inhibition. J. Invest. Dermatol., 140, 2242-2252.

36 Possemato, R., Marks, K.M., Shaul, Y.D., Pacold, M.E., Kim, D., Birsoy, K., Sethumadhavan, S., Woo, H.K., Jang, H.G., Jha, A.K. et al. (2011) Functional genomics reveal that the serine synthesis pathway is essential in breast cancer. Nature, 476, 346- 350.

37 Yoshino, H., Enokida, H., Osako, Y., Nohata, N., Yonemori, M., Sugita, S., Kuroshima, K., Tsuruda, M., Tatarano, S. and Nakagawa, M. (2020) Characterization of PHGDH expression in bladder cancer: potential targeting therapy with gemcitabine/cisplatin and the contribution of promoter DNA hypomethylation. Mol. Oncol., 14, 2190-2202.

38 Downing, J.R., Wilson, R.K., Zhang, J., Mardis, E.R., Pui, C.H., Ding, L., Ley, T.J. and Evans, W.E. (2012) The Pediatric Cancer Genome Project. Nat. Genet., 44, 619-622.

39 Wang, Z., Wilson, C.L., Easton, J., Thrasher, A., Mulder, H., Liu, Q., Hedges, D.J., Wang, S., Rusch, M.C., Edmonson, M.N. et al. (2018) Genetic Risk for Subsequent Neoplasms Among Long-Term Survivors of Childhood Cancer. J. Clin. Oncol., 36, 2078-2087.

40 Li, H. and Durbin, R. (2009) Fast and accurate short read alignment with Burrows- Wheeler transform. Bioinformatics, 25, 1754-1760.

41 McKenna, A., Hanna, M., Banks, E., Sivachenko, A., Cibulskis, K., Kernytsky, A., Garimella, K., Altshuler, D., Gabriel, S., Daly, M. et al. (2010) The Genome Analysis

Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res., 20, 1297-1303.

42 DePristo, M.A., Banks, E., Poplin, R., Garimella, K.V., Maguire, J.R., Hartl, C., Philippakis, A.A., del Angel, G., Rivas, M.A., Hanna, M. et al. (2011) A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat. Genet., 43, 491-498.

43 Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R. and Genome Project Data Processing, S. (2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics, 25, 2078-2079.

44 Yeo, G. and Burge, C.B. (2004) Maximum entropy modeling of short sequence motifs with applications to RNA splicing signals. J. Comput. Biol., 11, 377-394.

45 Rentzsch, P., Witten, D., Cooper, G.M., Shendure, J. and Kircher, M. (2019) CADD: predicting the deleteriousness of variants throughout the human genome. Nucleic Acids Res., 47, D886-D894.

46 Thorvaldsdottir, H., Robinson, J.T. and Mesirov, J.P. (2013) Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. Brief Bioinform., 14, 178-192.

47 Robinson, J.T., Thorvaldsdottir, H., Winckler, W., Guttman, M., Lander, E.S., Getz, G. and Mesirov, J.P. (2011) Integrative genomics viewer. Nat. Biotechnol., 29, 24-26.

48 Kramer, A., Green, J., Pollard, J., Jr. and Tugendreich, S. (2014) Causal analysis approaches in Ingenuity Pathway Analysis. Bioinformatics, 30, 523-530.

49 Carter, R.E. and Sorkin, A. (1998) Endocytosis of functional epidermal growth factor receptor-green fluorescent protein chimera. J. Biol. Chem., 273, 35000-35007.

50 Vaser, R., Adusumalli, S., Leng, S.N., Sikic, M. and Ng, P.C. (2016) SIFT missense predictions for genomes. Nat. Protoc., 11, 1-9.

51 Landrum, M.J., Lee, J.M., Benson, M., Brown, G., Chao, C., Chitipiralla, S., Gu, B., Hart, J., Hoffman, D., Hoover, J. et al. (2016) ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res., 44, D862-868.

52 Mayakonda, A., Lin, D.C., Assenov, Y., Plass, C. and Koeffler, H.P. (2018) Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res., 28, 1747-1756.

UNCOK Ni … . .

Figure Legends

UNCORRECTED MANUSCRIPT

Downloaded from https://academic.oup.com/hmg/advance-article/doi/10.1093/hmg/ddaa268/6035190 by guest on 21 December 2020

Figure 1

A
Germline VariantsSomatic Variants
PHGDHPathogenicCTNNB1Pathogenic
EGFRVUSCDKN2A/BPathogenic
SOS1VUSFLCNPathogenic
LMNAVUSCCNE1VUS
RGS2VUSCDH1VUS
TTC21BVUSEPHA7VUS
ECT2LVUSFLT4VUS
MPHOSPH8VUSGATA3VUS
ANO2VUSGPR124VUS
NRXN3VUSMLL3VUS
TPCN1VUSPRKDCVUS
ARHGEF38VUS
PCDH17VUS
CYSTM1VUS
ARL5BVUS
CCDC80VUS
HECTD1VUS
SHANK2VUS
ADCY4VUS
MEAF6VUS
METTL5VUS
NRCAMVUS
RPRMLVUS

B

Akt

CAPG

CCNE1

AGK

CDH1

ADGRA2

CDKN2A

estrogen receptor

CMTM8

Vegf

CTNNB1

STAT53/b

ERK

SOS1

ERK1/2

RGS2

GATA3

RAS

EGFR

Histone h3

PTPRS

Histone h4

PRKDC

lgG

PI3K p85

KMT2C

PI3K (complex)

LINC00963-209

PACSIN2

P38 MAPK

LMNA

NSD1

Mek

NRCAM

NFKB (complex)

Figure 1. Pathway Analysis of the Representative ACC Female Participant Based on her Prioritized Germline and Somatic Variants.

(A) Prioritized germline and somatic variants of the CCF participant with ACC.

(B) Pathway analysis resulting from the combination of the genes harboring germline and somatic variants in the ACC participant.

UNCORRECTED MANUSCRIPT

Figure 2

A

DAPI

EGFR

pEGFR

Merge

WT EGFR

Mutant EGFR

B ☒

B

WT EGFR

Mutant EGFR

C

QV

WT EGFR

Mut EGFR

Min of 100ng/ml EGF

0

0

5

45

0

5

45

Nuclear pAKT (S473)

Nuclear AKT

Nuclear a-Actinin

Total GAPDH

D

Nuclear AKT & pAKT

1.60

1.49

1.40

1.34

1.20

1.14

1.13

1.00

1.00

0.80

0.60

0.44

0.49

0.49

0.40

0.36

0.26

0.27

0.20

0.17

0.13

0.18

0.00

QV no EGF

WT EG FR no EGF

WT EGFR 5 min EGF

WT EGFR 45 min EGF

Mut EGFR no EGF

Mut EGFR 5 min EGF

Mut EGFR 45 min EGF

Total AKT PAKT

Figure 2. EGFR Downstream Signaling Pathway Activation in Mutant EGFR Cells.

(A) Immunofluorescence assay of EGFR (green) and pEGFR (red) in Stable mutant and wildtype cells. Magnification is 63x.

(B) Immunofluorescence assay of EGFR (green) and pEGFR (red) in transiently transfected mutant and wildtype cells. Magnification is 252x.

(C) Immunoblotting of AKT and pAKT in empty vector, wildtype, or mutant cells, at 3 different time points after 100ng/ml epidermal growth factor (EGF) treatment.

(D) Quantification related to C. Nuclear AKT and pAKT were higher in the mutant cells compared to wildtype cells, while the pAKT/AKT ratio was not significantly different between the two groups. Two-sided Student’s t-test P > .05.

UNCORRECTED M.

Figure 3

Erlotinib 10uM
Lapatinib 10uM Gefitinib 10uM
pazapanib10uM
Osimertinib 10uM
Sunitinib10uM
EGF 100ng/ml/5min

A

Replica 1

Replica 2

Replica 3

C

Mutant EGFR

Wildtype EGFR

No EGF

Mutant EGFR

pEGFR(Y1045)

Total EGFR

PERK (Thr202,Tyr204)

ERK

B

Ectoderm markers

Mesoderm markers

Endoderm markers

pAKT (S473)

8.0

AKT

Relative expression

7.0

6.7

6.0

pSTAT3 (Y705)

5.2

5,2

5.4

5.0

4.0

3.8

STAT3

3.0

2.2

2,5

2,5

2.0

1.4

1,0

GAPDH

1.0

D.O

FoxG2

CD133

CCHI

Map2

Otx2

TPS3

Brachyury

Msx1

Pdet

AFP

P7 WT EGFR

P7 Mut EGFR

D

E

Relative expression

Sunitinib10UM

Rapamycin10uM

.

% Cell Viability after 72hrs of Treatment

1

6

120

$

100

2

İ

1

.

80

0

FORL2

CO1 33

CDH1

Map2

0tx2

TP63

Brachyury

Fax:2

CD6 33

COHI

Map2

Obx2

TP63

Beachyury

WENO TX WT EGFR W no TYMut EGFR: 1 Sontini BOUM WT EGFR . Senitrib 104M Mor EGFR

Wino TX WT EGFR W no TX Mac EGFR & Rapamycin I DUM WT EGFR & Rapanych IDUM MUT EGFR

60

Relative expression

8

Erlotinib 10uM

Osimertinib 10uM

40

,

6

20

$

0

100 uM

2

0 uM

0.01 uM

0.1 uM

1 uM

10 uM

1

0

CD1 33

Brachyury

8

WT-Osimertinib+Sunitinib

Mut-Osimertinib+Sunitinib

CDH1

Mag?

TP63

Farc2

CD1.33

COHI

Map2

Ots2

TP63

Bradıyury

Figure 3. EGFR Mutant Cells Migrate Faster and Have More Stem-like Behavior Compared to the Wildtype Cells.

(A) EGFR mutant cells migrate faster compared to wildtype cells. This experiment was performed in duplicate.

(B) EGFR mutant cells have more stem-like behavior compared to the wildtype cells. Data represent mean values + SD. This experiment was performed in triplicate.

(C) Comparing the effect of 10 uM Erlotinib, Lapatinib, Gefitinib, Pazopanib, Osimertinib, and Sunitinib, accompanying 100ng/ml/5min EGF treatment on mutant vs. wildtype cells.

(D) Comparing the effect of 10 uM Erlotinib, Osimertinib, Rapamycin, and Sunitinib on the stem-like behavior of the EGFR mutant cells. Data represent mean values ± SD. ** P= . 00015,

*P= . 011, calculated by two-sided Student’s t-test.

(E) Effect of Sunitinib and Osimertinib combination therapy on the mutant EGFR cells. P = .00027, calculated by two-sided Student’s t-test.

UNCORRECT

Figure 4
ChildrenAYAAdult <50Adult >=50
RTK-RAS GenesNTRK118%12%4%
FGFR4 KIT
FGFR1
EGFR25%
FGFR315%6%6%12%
PDGFRB
JAK2
INSR
ROS1
FGFR2
INSRR
RET
RASGRP3 SHC3 RAPGEF1 RAPGEF2 SOS1 SHC1 RASGRP2 ABL1 IRS110%9% 3%6% 18%6% 2%
PLXNB118% 3%24% 6%6%
DAB2IP5% 10%
RASA1
CBL
ERRFI1 RASA2 FNTB RASAL3 RCE1 RASAL2
HRAS3%
SCRIB2% 2%
100 Patients100 Patients100 Patients100 Patients
30%45%36%12%
25%12%24%18%

Downloaded from https://academic.oup.com/hmg/advance-article/doi/10.1093/hmg/ddaa268/6035190 by guest on 21 December 2020

Figure 4. RTK-RAS Pathway Variant Distribution in Patients with ACC. Frequency of prioritized germline RTK-RAS pathway variants in individuals with ACC. Percentage of variants in each age group calculated based on potential 100 patients. Children’s group, P <. 0001, OR=1193.5, 95% CI=431.4-3302.0; AYA group, P <. 0001, OR=373.0, 95% CI= 183.0-760.4; Adult < 50 years-old, P <. 0001, OR=895.1, 95% CI=311.3-2574.0; Adult >= 50 years-old, P < . 0001, OR=248.6, 95% CI= 133.9-461.6.

Table 1. Details of EGFR Variant in the Female Case with ACC Tumor
ChromosomePositionREF SequenceALT SequenceVariant typeCytobandHGVSGene SymbolRS ID
Chr755270285GASNV7p11.2NM_005228.5:c.3238G>A (p.Asp1080Asn)EGFRrs1023185448
PM2Pathogenic ModerateGnomAD exomes allele frequency = 0.000004 (Threshold for recessive gene EGFR=0.0001)
PP2Pathogenic Supporting64 out of 82 non-VUS missense variants in gene EGFR are pathogenic = 78.0% (threshold=51.0%) 94 out of 343 clinically reported variants in gene EGFR are pathogenic = 27.4% (threshold=12.0%)
PP3Pathogenic Supporting6 pathogenic predictions from DANN, FATHMM-MKL, M-CAP, MutationAssessor, Mutation Taster and SIFT

Abbreviations: REF, Reference; ALT, Altered; HGVS, Human Genome variation Society; RS ID, Reference SNP Cluster Identification; Chr, Chromosome; SNV, Single Nucleotide Variant; PM2, ACMG criteria meaning Variant Absent from Population Databases; PP2, ACMG Criteria Meaning Missense Variant in a Gene with a Low Rate of Benign Missense Variation and where Missense Variants are a Common Mechanism of Disease; PP3, ACMG criteria meaning Variant with Multiple Lines of Computational Evidence; ACMG, American College of Medical Genetics and Genomics; VUS, Variant of Uncertain Significance; DANN, Deleterious Annotation of genetic variants using Neural Networks; FATHMM-MKL, Functional Analysis through Hidden Markov Models- Multiple Kernel Learning Algorithm; M-CAP, Mendelian Clinically Applicable Pathogenicity Score; SIFT, Sorts Intolerant From Tolerant Amino Acid Substitutions Tool.

UNCORRECTED MANUSCRIPT

Table 2. Prevalence of prioritized germline EGFR variants in 113 individuals with ACC
Total Number of PatientsData Source% of EGFR VariantsP value
Children21StJ4.8%.004
AYA32CCF + TCGA6.2%<. 0001
Adults60TCGA1.7%.07
TotalUNCORRECTED 113CCF+ StJ+ TCGA3.5%MANUSCRIPT <. 0001

Abbreviations: AYA, Adolescent and Young Adult; StJ, St. Jude Children’s Research Hospital; CCF, Cleveland Clinic Foundation; TCGA, The Cancer Genome Atlas.

Downloaded from https://academic.oup.com/hmg/advance-article/doi/10.1093/hmg/ddaa268/6035190 by guest on 21 December 2020

Abbreviations

AYA ACC

Adolescents and Young Adults

Adrenocortical Carcinoma

ALT Altered

ACMG American College of Medical Genetics and Genomics

BAM

Binary Alignment Map

CADD

Combined Annotation Dependent Depletion

C-AYA

Childhood-AYA

Chr Chromosome

CCF

Cleveland Clinic Foundation

EGF

Epidermal Growth Factor

EGFR

Epidermal Growth Factor Receptor

EMT

Epithelial to Mesenchymal Transition

ExAC

Exome Aggregation Consortium

FATHMM-MK

Functional Analysis through Hidden Markov Models- Multiple Kernel

Learning Algorithm

GATK

Genome Analysis Toolkit

GFP

Green Fluorescent Protein

HGVS

Human Genome Variation Society

IGV

Integrative Genomics Viewer

IPA

Ingenuity Pathway Analysis

IVA

Ingenuity® Variant Analysis™M

ENCORE NEM ROD MANUSCRIPT

M-CAP

Mendelian Clinically Applicable Pathogenicity Score

NSCLS

Non-Small-Cell Lung Cancer

P/LP

Pathogenic or Likely Pathogenic

PCGP PHGDH

Pediatric Cancer Genome Project

Phosphoglycerate Dehydrogenase

PM2

ACMG Criteria Meaning Variant Absent from Population Databases

PP2

ACMG Criteria Meaning Missense Variant in a Gene with a Low Rate of Benign Missense Variation and where Missense Variants are a Common Mechanism of Disease

PP3

ACMG Criteria Meaning Variant with Multiple Lines of Computational Evidence

DANN

Deleterious Annotation of Genetic Variants Using Neural Network

REF

Reference

RS ID

Reference SNP Cluster Identification

RTK

Receptor Tyrosine Kinase

SIFT

Sorts Intolerant from Tolerant Amino Acid Substitutions Tool

SNV

Single Nucleotide Variation

STJ

St. Jude

SJLIFE

St. Jude Lifetime

TCGA

The Cancer Genome Atlas

VUS

Variants of Uncertain Significance

WES

Whole Exome Sequencing