Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 Molecular PharmacologyThis article hasFast not beenForward. copyedited Published and formatted. on The January final version 31, may 2013 differ as from doi:10.1124/mol.112.084111 this version.

MOL #84111

Title Page

Title: A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR,

ERBB2 and MET Drives Growth and Survival of Head and Neck

Squamous Carcinoma Cell Lines

Downloaded from

Authors: Katherine R. Singleton, Jihye Kim, Trista K. Hinz, Lindsay A.

Marek, Matias Casás-Selves, Clark Hatheway, Aik Choon Tan, molpharm.aspetjournals.org

James DeGregori, and Lynn E. Heasley

Affiliations: Departments of Craniofacial Biology (K.R.S., T.K.H., L.A.M., C.H.,

L.E.H.), Medicine (J.K., A.T.), and Biochemistry and Molecular at ASPET Journals on October 2, 2021

Genetics (M.C.S., J.D.), University of Colorado Anschutz Medical

Campus, Aurora, Colorado 80045

1

Copyright 2013 by the American Society for Pharmacology and Experimental Therapeutics. Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Running Title Page

Running Title: Synthetic lethal relations identify a RTK network in HNSCC

Corresponding Author: Katherine R. Singleton

Department of Craniofacial Biology Downloaded from

School of Dental Medicine, University of Colorado Anschutz

Medical Campus molpharm.aspetjournals.org

Mail Stop 8120, P.O. Box 6511, Aurora, CO 80045

Email: [email protected]

Number of Pages: Text – 21 pages (not including figure legends)

Tables – 0 + 2 Supplemental Tables at ASPET Journals on October 2, 2021

Figures – 7 + 6 Supplemental Figures

References – 49

Number of Words: Abstract – 248

Introduction – 747

Discussion – 923

List of Non-Standard Abbreviations:

AKT – protein kinase B

ALK – anaplastic lymphoma kinase

BiNGS! - Bioinformatics for Next Generation Sequencing

BSA – bovine serum albumin

2 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

CML – chronic myeloid leukemia

DMEM – Dulbecco’s modified Eagle medium

DMSO – dimethyl sulfoxide

EGFR - epidermal growth factor receptor

EPHB2 – ephrin receptor B2

ERBB2 - epidermal growth factor receptor 2

ERK – extracellular signal regulated kinase Downloaded from

ESCC – esophageal squamous cell carcinoma

FBS – fetal bovine serum molpharm.aspetjournals.org

FDR – false discovery rate

FGF2 – fibroblast growth factor 2

FGFR - fibroblast growth factor receptor

FZD – frizzled at ASPET Journals on October 2, 2021

HGF – hepatocyte growth factor

HNSCC - head and neck squamous cell carcinoma

INSRR – insulin receptor related receptor

JAK – Janus kinase

MAPK – mitogen-activated protein kinase

MEK2 – mitogen activated protein kinase kinase 2

MET- hepatocyte growth factor receptor

MOI – multiplicity of infection

NB – negative binomial

NSCLC – non-small cell lung cancer

3 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

p16INK4A – cyclin-dependent kinase inhibitor 2A

p53 – tumor protein 53

PARP – poly ADP ribose polymerase

PBS – phosphate buffered saline

PI3K – phosphoinositide 3-kinase

PTEN – phosphatase and tensin homolog

PTK7 – protein tyrosine kinase 7 Downloaded from

RAS – rat sarcoma

Rb – retinoblastoma protein molpharm.aspetjournals.org

RPMI-1640 – Roswell Park Memorial Institute-1640

RTK - receptor tyrosine kinase

SDS – sodium dodecyl sulfate

SHP2 – tyrosine-protein phosphatase non-receptor type 11 at ASPET Journals on October 2, 2021

SLAT - synthetic lethal with AZ8010 treatment

SOS2 – son of sevenless 2

STAT - signal transducer and activator of translation

TCF7 – T-cell factor

TKI - tyrosine kinase inhibitor

TNKS – tankyrase

TTBS – Tris-buffered saline with 0.1% Tween 20

4 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Abstract:

Our lab has previously shown that some gefitinib insensitive head and neck squamous cell carcinoma (HNSCC) cell lines exhibit dominant autocrine fibroblast growth factor receptor (FGFR) signaling. Herein, we deployed a whole genome loss-of-function screen to identify genes whose knockdown potentiated the inhibitory effect of the FGFR inhibitor, AZ8010, in HNSCC cell lines. Three HNSCC cell lines expressing a genome- wide shRNA library were treated with AZ8010 and the abundance of shRNA sequences Downloaded from was assessed by deep sequencing. Underrepresented shRNAs in treated cells are expected to target genes important for survival with AZ8010 treatment. Synthetic lethal molpharm.aspetjournals.org hits were validated with specific inhibitors and independent shRNAs. We found that multiple alternate receptors provided protection from FGFR inhibition, including receptor tyrosine kinases (RTKs), epidermal growth factor receptor 2 (ERBB2) and hepatocyte growth factor receptor (MET). We showed that specific knockdown of either ERBB2 or at ASPET Journals on October 2, 2021

MET in combination with FGFR inhibition led to increased inhibition of growth relative to

FGFR tyrosine kinase inhibitor (TKI) treatment alone. These results were confirmed using specific small molecule inhibitors of either ERBB family members or MET.

Moreover, the triple combination of FGFR, MET and ERBB family inhibitors showed the largest inhibition of growth and induction of apoptosis as compared to the double combinations. These results reveal a role for alternate RTKs in maintaining pro-growth and survival signaling in HNSCC cells in the setting of FGFR inhibition. Thus, improved therapies for HNSCC patients could involve rationally designed combinations of TKIs targeting FGFR, ERBB family members and MET.

5 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Introduction:

Over 500,000 patients worldwide are diagnosed with HNSCC yearly. With a 5- year survival of only 50% (Haddad et al., 2008), HNSCC exhibits one of the poorest survival rates among common cancer types. Therapy for HNSCC has seen little advancement in recent years and largely involves improved chemotherapy schedules and use of intensity-modulated radiation (Murdoch 2007). Unlike other cancers, the frequency of oncogenic driver mutations (Agrawal et al., 2011; Stransky et al., 2011) in Downloaded from

HNSCC is low. Rather, HNSCC tumors are characterized by mutations in tumor suppressors such as tumor protein 53 (p53), cyclin-dependent kinase inhibitor 2A molpharm.aspetjournals.org

(p16Ink4A), phosphatase and tensin homolog (PTEN) and retinoblastoma protein (Rb)

(reviewed in (Leemans et al., 2011)). Approximately 90% of HNSCC tumors overexpress epidermal growth factor receptor (EGFR) (Dassonville et al., 1993) and

EGFR ligands, correlating with poor prognosis (Dassonville et al., 1993; Grandis et al., at ASPET Journals on October 2, 2021

1996; Grandis et al., 1998; Ang et al., 2004). From this evidence, trials of the EGFR

TKIs, gefitinib and erlotinib, and the inhibitory antibody, cetuximab, were completed in

HNSCC patients. While no significant response to EGFR TKIs was observed, cetuximab yielded a modest increase in survival (Cohen et al., 2003; Soulieres et al., 2004; Cohen

2006; Specenier et al., 2011) and is approved for HNSCC treatment (Specenier et al.,

2011).

The hypothesis that tumor heterogeneity across a cancer type dictates therapeutic response forms the basis of personalized medicine and shows promise for improved treatment of non-small cell lung cancer (NSCLC). In trials of NSCLC patients treated with an EGFR TKI, responses are limited to tumors bearing EGFR mutations

6 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

(Shepherd et al., 2005; Gazdar 2009). However, EGFR mutations are rare in HNSCC

(Agrawal et al., 2011; Stransky et al., 2011) and attempts to match EGFR expression or gene copy number with cetuximab response have failed (Egloff et al., 2009; Licitra et al., 2011). Recently, we showed that a possible intrinsic resistance mechanism of

HNSCC cells to EGFR inhibition is mediated by an RTK autocrine loop comprised of

FGFRs and fibroblast growth factor 2 (FGF2) that could be inhibited with FGFR TKIs

(Marshall et al., 2011). This work highlighted an FGFR autocrine loop as a novel target Downloaded from for improved HNSCC therapy and illustrated how growth of tumors may be driven distinct from traditional oncogenic mutations. molpharm.aspetjournals.org

Evidence supports the activity of other alternate RTKs in growth of HNSCC.

ERBB2 is a dominant mediator of growth in a subset of breast cancers (Perou et al.,

2000). Also, ERBB2 is amplified in esophageal squamous cell carcinomas (ESCCs) and at ASPET Journals on October 2, 2021 overexpression is linked to lowered survival (Sato-Kuwabara et al., 2009). ERBB2 is overexpressed in 20-40% of HNSCC tumors with gene amplification in 5-10% of cases, correlating with decreased survival (reviewed in (Morgan et al., 2009)). A role for an additional RTK, MET, is emerging in HNSCC. MET is overexpressed in some HNSCC tumors and efficacy of MET TKIs to reduce survival and migration of HNSCC cells has been shown (Seiwert et al., 2009). However, the full role of this RTK in the survival of

HNSCC cells remains uncertain.

Following the success of the ABL inhibitor, imatinib mesylate, in the treatment of

BCR-ABL-positive chronic myeloid leukemia (CML) (Druker 2002), cancer research has increasingly focused on identifying and targeting similar driver events in other cancers

(Stambuk et al., 2010). The success of imatinib lies primarily in BCR-ABL being the sole

7 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 driving event for CML (Druker 2002), but other cancers have not proven as tractable.

NSCLC is not characterized by a single driver, but by many, resulting in the need for patient stratification before treatment (Ladanyi et al., 2008). Some drivers are difficult to identify, as illustrated by the response of HNSCC patients to EGFR inhibitors

(Kalyankrishna et al., 2006) while others, exemplified by RAS, remain difficult to target

(Hopkins et al., 2002). In addition, even with targeted therapy, a significant number of patients will develop resistance (O'Dwyer et al., 2004). As a result, attention has Downloaded from focused on discovering biomarkers for patient selection as well as exploring novel combination therapies (Morse et al., 2010). With this goal, we performed a genome- molpharm.aspetjournals.org wide shRNA loss-of-function screen to identify targets whose inhibition would improve the response of HNSCC cells to FGFR TKIs. We found that HNSCC cell lines rely on the activation of multiple alternate RTKs, including ERBB2 and MET, to fully establish growth and survival. Our findings are consistent with the hypothesis of RTK coactivation at ASPET Journals on October 2, 2021

(Xu et al., 2010) and responds to a call for development of therapies comprised of combinations of targeted inhibitors (Glickman et al., 2012).

8 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Materials and Methods:

Cell Culture: All HNSCC cell lines used in this study were submitted to fingerprint analysis by the University of Colorado Cancer Center DNA Sequencing and Analysis

Core confirming their authenticity. Cell lines were routinely cultured in Dulbecco’s modified Eagle medium (DMEM) or Roswell Park Memorial Institute-1640 (RPMI-1640) Downloaded from

(584-A2, MSK-921) growth medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) with 1% penicillin-streptomycin (Sigma-Aldrich, St. Louis, MO) molpharm.aspetjournals.org at 37°C in a humidified 5% CO2 incubator.

Lentivirus Preparation: The HIV based GeneNet Lentiviral Human 50K library (pSIH1-

H1-Puro, System Biosciences, SBI, Mountain View, CA) was packaged in 293T cells at ASPET Journals on October 2, 2021 with packaging component vectors coding for VSV-G, Gag, Pol, and Rev (generously provided by Dr. Douglas Graham, Dept. Pediatrics, UC Denver Anschutz Medical

Campus). Cells were incubated overnight with Turbofect transfection reagent

(Fermentas, Glen Burnie, MD), packaging vectors and library. The lentiviruses released into the medium were filtered using a 0.45 μm filter (Corning Inc., Corning, NY). 4x106 cells from each of the cells lines UMSCC25, 584-A2, and CCL30 were plated and transduced 24 hrs later with the SBI shRNA library containing viral media and 8 μg/mL polybrene (Sigma-Aldrich) at a multiplicity of infection (MOI) of ~0.2. Viral titer was determined by serial dilution. After 72 hrs, transfected cells were selected with puromycin (1 μg/mL) for 5 days. For validation of synthetic lethal hits, five independent

9 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 shRNAs (Supplemental Table 1) in the lentiviral vector pLKO.1 and the pCMV-VSV-G and pΔ8.9 packaging component vectors were obtained from the Functional Genomics

Facility at the University of Colorado at Boulder. All cultures were confirmed for knock down and submitted to growth assays within three passages of puromycin (1 μg/mL,

Sigma-Aldrich) selection.

Genome-Wide RNAi-Based Functional Screening: Library expressing HNSCC cells Downloaded from were divided into six groups of 5x106 cells per T-75 flask (BD Biosciences, Franklin

Lakes, NJ), where three were treated with vehicle (dimethyl sulfoxide, DMSO) and three molpharm.aspetjournals.org groups were treated with AZ8010 (1.0 μM for UMSCC25, 0.125 μM for 584-A2 and 0.5

μM for CCL30). After 72 hrs, the cells were placed in media without drug for an additional 72 hrs. Total RNA was purified from each replicate using an RNeasy Mini Kit

(Qiagen, Germantown, MD) and reverse transcribed using vector specific primers (SBI) at ASPET Journals on October 2, 2021 and MMLV reverse transcriptase (Invitrogen, Carlsbad, CA). The cDNA was amplified using nested PCR to isolate the shRNA sequences and to add Illumina adapter sequences. The samples were sequenced on an Illumina Genome Analyzer IIx

(Illumina, San Diego, CA).

Bioinformatics Analysis: To analyze and interpret the sequence data obtained from the Genome Analyzer, we used the bioinformatics pipeline, BiNGS! (Bioinformatics for

Next Generation Sequencing) (Kim et al., 2012), as described and validated in previous studies (Casas-Selves et al., 2012; Porter et al., 2012; Sullivan et al., 2012). Briefly, after filtering steps, Negative Binomial (NB) was then used to model the distribution of

10 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 read counts of the data using edgeR (Robinson et al., 2007). The q-value of the false discovery rate (FDR) was calculated for multiple comparisons. In order to combine multiple shRNAs into genes, we adapted weighted Z-transformation It is based on the product of q-value, which follows a chi-square distribution with 2k degrees of freedom

(where k is the number of p-values) (Whitlock 2005). Lists of genes with differentially represented shRNAs were ranked using the associated p-value [P(wZ)]. By multiplying the P(wZ) value by the rank of the gene, an E-value was obtained. E-value is an Downloaded from estimation of false positivity. The resulting data sets have been deposited in NCBI GEO

Database (Accession #GSE39305) molpharm.aspetjournals.org

Clonogenic Growth Assay: To measure the effect of inhibitors or shRNA-mediated knock down on cell growth, cells were seeded at 100 cells per well in six well tissue culture plates in full media. After 24 hrs, cells were treated as indicated and cultured for at ASPET Journals on October 2, 2021

14 days with the addition of fresh media containing inhibitors every 7 days. Plates were rinsed twice with phosphate buffered saline (PBS) and fixed and stained with 0.5%

(wt/vol) crystal violet in 6.0% (vol/vol) gluteraldehyde solution for 30 min at room temperature. Plates were rinsed in distilled H2O and photographed. The MetaMorph imaging software program (Molecular Devices, Downingtown, PA) was used to quantify total colony area.

Anchorage-Independent Growth Assay: In order to measure the effect of inhibitors on anchorage independent growth, 20,000 cells were suspended in 1.5 mL of growth media containing 0.35% Difco agar noble (Becton, Dickinson and Co., Franklin Lakes,

11 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

NJ) and overlaid on a base layer of growth media containing 0.5% agar noble. Drugs were added by overlay with 2 mL growth media. Media and drugs were replaced once a week. After 14 days, plates were stained for 24 hrs with 200 μL of 1 mg/mL nitroblue tetrazolium and photographed. MetaMorph was used to quantify total colony number.

Cell Proliferation Assay: Cells were plated at 50 cells per well in either 96 or 48 well tissue culture plates and treated with inhibitors at various doses. After the DMSO- Downloaded from treated control wells had become confluent (1-2 weeks), cell numbers were assessed using a CyQUANT Direct Cell Proliferation Assay (Invitrogen) according to the molpharm.aspetjournals.org manufacturer’s instructions.

Immunoblot Analysis: Phospho-extracellular signal regulated kinase (ERK), phospho-

MET, phospho-AKT, phospho-ERBB2, phospho-EGFR, phospho-STAT3, poly ADP at ASPET Journals on October 2, 2021 ribose polymerase (PARP) and caspase-3 levels in HNSCC cells were measured by immunoblotting. Cells in 10 cm tissue culture dishes were treated with different inhibitors in full growth media. After treatment, cells were rinsed once in PBS, collected in PBS and centrifuged at 2,000 x g for 3 min. The cell pellet was lysed in buffer containing 0.5% Triton X-100, 50 mM β-glycerophosphate (pH 7.2), 0.1 mM Na3VO4, 2 mM MgCl2, 1 mM EGTA, 1mM DTT, 0.3 M NaCl, 2 μg/mL leupeptin and 4 μg/mL aprotinin. Samples were centrifuged at 13,000 x g for 5 min and the particulate fraction was discarded. Aliquots of cell extracts were added to 25 μL of sodium dodecyl sulfate

(SDS) loading buffer and proteins were resolved by SDS-PAGE. After electrophoretic transfer to nitrocellulose, filters were blocked in 3% bovine serum albumin (BSA; Cohn

12 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Fraction V, ICN Biomedicals, Inc., Aurora, OH) in Tris-buffered saline with 0.1% Tween

20 (TTBS). The filters were then incubated overnight at 4°C with anti-phospho-Erk, phospho-MET (Tyr1234/1235), phospho-Akt (Thr308), phospho-ERBB2 (Tyr1248), phospho-STAT3 (Ser727) (Tyr705), PARP or caspase-3 (Cell Signaling Technology,

Inc., Danvers, MA, #9101, #3077, #4056, #2247, #3281, #9134, #9145, #9542 and

#9661) The filters were washed three times in TTBS and incubated for one hour at room temperature with alkaline phosphatase coupled goat anti-rabbit antibodies. The filters Downloaded from were developed using LumiPhos reagent (Pierce, Rockford, IL) according to the manufacturer’s instructions. The filters were stripped and reprobed for total ERK1 and 2 molpharm.aspetjournals.org or Na/K-ATPase using a mixture of ERK1 and ERK2 antibodies or Na/K-ATPase α- subunit antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, sc-93, sc-154 and sc-21712). The filters were also reprobed for total MET, total AKT, total STAT3, total

ERBB2 and total EGFR (Cell Signaling Technology, Inc., #3127, #9272, #9132, #2242 at ASPET Journals on October 2, 2021 and #2232).

Caspase 3 Activity Assay: Caspase 3 activity was assessed as previously described

(Matassa et al., 2001).

Chemicals: AZ8010 and AZD8931 were obtained by Material Transfer Agreements from AstraZeneca (Wilmington, DE). PF02341066 and lapatinib were obtained from

Selleck Chemicals (Houston, TX).

13 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Results:

A synthetic lethal screen identifies genes whose silencing enhances growth inhibition by FGFR-specific TKIs. In order to identify targets whose inhibition would sensitize HNSCC cells to FGFR TKIs, three cell lines expressing a genome-wide library of shRNAs were used in a loss-of-function screen (Fig. 1A). The HNSCC cell lines used to express the shRNA library were UMSCC25 and 584-A2, cell lines derived from tumors of the larynx, and CCL30, derived from a tumor of the nasal septum. Downloaded from

UMSCC25 cells are highly dependent on EGFR for growth and signaling while FGFRs function as an auxiliary growth pathway (Supplemental Figure 1A and (Marshall et al., molpharm.aspetjournals.org

2011)). By contrast, CCL30 and 584-A2 cells are dependent on FGFR alone for growth as evidenced by high sensitivity to treatment with the FGFR specific TKI, AZ8010

(Supplemental Figure 1A and B, (Marshall et al., 2011)). The levels of EGFR and

FGFRs in these cell lines were published previously (Marshall et al., 2011). at ASPET Journals on October 2, 2021

Cell lines stably expressing the lentiviral shRNA library with the majority of cells expressing only one shRNA were treated with DMSO for 72 hours followed by 72 hours of recovery in drug free media. Alternatively, the cell lines were treated with the FGFR inhibitor, AZ8010, for 72 hours at concentrations that reduced growth by approximately

30% as compared to DMSO treated controls, followed by a 72 hour recovery period without drug. Total RNA was then isolated from each treatment group, and the shRNAs were amplified, sequenced and analyzed as described in the Materials and Methods.

Nearly 28 million reads were generated for each replicate and the reads were found to cluster within treatment groups and cell lines via unsupervised hierarchical clustering

(Supplemental Figure 2). This indicates that the drug treatment is reproducibly affecting

14 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 the proportions of shRNAs present in the cell populations and that the knockdown of certain genes can modulate the response of cells to AZ8010 treatment.

In order to consider a gene a synthetic lethal with FGFR inhibition, we chose an

E value ≤2 as the cut-off (Supplemental Table 2) from the BiNGS! ranked genes. We termed the synthetic lethal hits Synthetic Lethal with AZ8010 Treatment (SLAT) and detected 504 SLATs in UMSCC25 cells, 292 SLATs in 584-A2 cells, and 104 SLATs in

CCL30 cells. While there was varying amounts of overlap between any two cell lines, Downloaded from with UMSCC25 and 584-A2 sharing the most SLATs (25), CCL30 sharing 9 with

UMSCC25 and 9 with 584-A2, there were no SLATs observed in common among all molpharm.aspetjournals.org three cell lines (Supplemental Figure 3). This likely reflects the distinct differences in growth drivers between 584-A2 and CCL30 cells (FGFR dominant) and UMSCC25 cells

(EGFR dominant).

For the remainder of the study, we have chosen to focus on the SLATs identified at ASPET Journals on October 2, 2021 in UMSCC25 cells that support a RTK network comprised of FGFRs, EGFR, ERBB2 and MET as the functional growth pathway. Where relevant, we use the functional genomics data from 584-A2 and CCL30 to highlight why ERBB2 and MET were not identified as SLATs in these cell lines. The ten highest ranking SLATs identified in

UMSCC25 cells are shown in Figure 1B with the raw sequencing counts in triplicate for each shRNA indicated. Notably, the RTK, MET, was identified as a SLAT based on three distinct shRNAs, while ERBB2 was identified based on a single shRNA. While no synthetic lethal genes were observed in common among the three cell lines, there were receptor tyrosine kinases, G-protein coupled receptors and specific downstream signaling components identified as SLATs in all the lines. For example, the insulin

15 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 receptor related receptor (INSRR) and the ephrin receptor B2 (EPHB2) were identified as highly ranked SLATs in 584-A2 and protein tyrosine kinase 7 (PTK7) and the

Wnt/Frizzled (FZD) pair, WNT16 and FZD7, were identified in CCL30 cells. Moreover, genes encoding signaling enzymes associated with WNT signaling (T-cell factor (TCF7) and tankyrase (TNKS)), phosphoinositide 3-kinase (PI3K) signaling (PIK3C3, PIK3R2,

PIK3CB and AKT2), Janus kinase-signal transducer and activator of translation (JAK-

STAT) signaling (STAT1 and STAT2), RAS/ERK signaling (son of sevenless 2 (SOS2), Downloaded from

KRAS, tyrosine-protein phosphatase non-receptor type 11 (SHP2) and mitogen activated protein kinase kinase 2 (MEK2)) were identified as SLATs with a less stringent molpharm.aspetjournals.org cut off of E value ≤20, which still corresponds to a p value ≤0.05 (Fig. 1C). Overall, the results from the synthetic lethal screen support the hypothesis of “RTK coactivation”, whereby cancer cells rely on activation of multiple RTKs to maintain flexible and vigorous signaling responses in the face of various insults such as drug treatment at ASPET Journals on October 2, 2021

(reviewed in (Xu et al., 2010)). To investigate whether HNSCC cell lines maintain growth in this manner, we have focused the remainder of the study on UMSCC25 cells and validation and further exploration of the SLATs, ERBB2 and MET.

Validation of ERBB2 and MET as synthetic lethal genes with FGFR-specific

TKIs in UMSCC25 cells. Both ERBB2 and MET are broadly expressed in a panel of

HNSCC cell lines including UMSCC25 cells (Figure 2). It is noteworthy that the FGFR dependent cell lines, 584-A2 and CCL30, express reduced levels of ERBB2 and MET, respectively, and neither of these RTKs were identified as SLATs in these cell lines in the initial screen. ERBB2 is a ligand independent member of the EGFR family that heterodimerizes with other members of the ERBB family to initiate downstream

16 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 signaling to mitogen-activated protein kinase (MAPK), protein kinase B (AKT) and JAK-

STAT pathways (reviewed in (Hynes et al., 2005)). MET is a RTK, where, upon binding of the receptor’s ligand, hepatocyte growth factor (HGF), MET is phosphorylated and participates in signaling for survival, motility and proliferation (reviewed in (Cecchi et al.,

2010)). To validate ERBB2 and MET, UMSCC25 cells were transduced with two independent shRNAs targeting ERBB2 and two independent shRNAs targeting MET.

The shRNAs were from the TRC library, distinct from those used in the original screen, Downloaded from and the level of target knockdown was determined by immunoblot (Fig. 3A). Consistent with the screen results, the combination of AZ8010 treatment and ERBB2 knockdown molpharm.aspetjournals.org resulted in greater growth inhibition as compared to either treatment alone (Fig. 3B).

Similarly, MET knockdown alone decreased clonogenic growth of UMSCC25 cells relative to the non-silencing control construct (NSC) expressing UMSCC25 cells.

Additionally, confirming the screen results, knockdown of MET in combination with at ASPET Journals on October 2, 2021

FGFR inhibition caused a greater reduction of clonogenic growth than AZ8010 treatment or MET knockdown alone (Fig. 3C). These data provide molecular validation of ERBB2 and MET as being synthetic lethal with an FGFR inhibitor in UMSCC25 cells.

A pan-ERBB inhibitor, AZD8931, yields synergistic inhibition of proliferation with FGFR-specific TKIs. Because genetic knockdown of ERBB2 sensitizes UMSCC25 cells to AZ8010 treatment, we investigated whether pharmacological inhibitors of the ERBB family have a similar effect. The small molecule

TKI, AZD8931, is a reversible, equipotent inhibitor of the ERBB family members EGFR,

ERBB2 and ERBB3 and has been shown to have significant anti-growth effects in

NSCLC and HNSCC xenografts in mice (Hickinson et al., 2010). AZD8931 potently

17 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 inhibited tyrosine phosphorylation of ERBB2 and EGFR in UMSCC25 cells, consistent with published results in other cell lines (Supplemental Figure 4A and B). UMSCC25 cells, in which ERBB2 was identified as a SLAT, showed a significant decrease in growth with ERBB inhibition alone and this inhibition was increased with the addition of

AZ8010 (Fig. 4A, Supplemental Figure 1A). When combinations of AZ8010 and

AZD8931 were examined over multiple concentrations, synergistic growth inhibition of

UMSCC25 cells (as evidenced by CI indices <0.5) was readily apparent (Fig. 4B). Downloaded from

Importantly, neither 584-A2 nor CCL30 cells responded to AZD8931 alone

(Supplemental Figure 1C) and the drug failed to increase growth inhibition combined molpharm.aspetjournals.org with AZ8010 (Fig. 4C). These findings demonstrate the specificity of AZD8931 and are consistent with the identification of ERBB2 as a SLAT only in UMSCC25 cells (Fig. 4B).

Similar results were observed with the pan-ERBB inhibitor, lapatinib (Supplemental

Figure 5). To investigate whether other HNSCC cell lines would similarly respond to the at ASPET Journals on October 2, 2021 combination of FGFR and ERBB inhibition, we tested additional HNSCC cell lines that we previously identified as exhibiting dual growth inputs through ERBB family members and FGFR (Marshall et al., 2011). As shown in Figure 4D, UMSCC1 and UMSCC8 cells showed a greater growth inhibition with combined AZ8010 and AZD8931 relative to either inhibitor alone. To confirm that the effect of AZD8931 was due to inhibition of

ERBB family members, ERBB2 was knocked down with two independent shRNAs in

UMSCC1 and UMSCC8 cells (Supplemental Figure 6A-C). A similar effect was seen with the combination of ERBB2 knockdown and AZ8010 treatment, suggesting that the targeting of EGFR in addition to other ERBB family members by AZD8931 is partially responsible for the effect. These results show that alternate RTKs such as ERBB2 can

18 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 provide protection from FGFR inhibition and that targeting these receptors with a combination of small molecule inhibitors can provide superior cell growth inhibition.

Combined treatment with FGFR and MET inhibitors synergistically inhibits

HNSCC cell line proliferation. Next, we tested if pharmacologic inhibition of both

FGFR and MET, another highly ranked SLAT (Fig. 1B), would yield additive or synergistic growth inhibition in UMSCC25, UMSCC1 and UMSCC8 cells. PF02341066

(PF1066; crizotinib) was originally developed as a MET inhibitor, but has other known Downloaded from targets including anaplastic lymphoma kinase (ALK) (Zou et al., 2007; Kwak et al.,

2010). PF1066 blocked phosphorylation of MET in UMSCC25 at concentrations molpharm.aspetjournals.org consistent with published data (Supplemental Figure 4B), indicating that MET is basally activated in untreated cells. HNSCC cell lines were treated with PF1066 and AZ8010 alone and in combination in anchorage independent or clonogenic growth assays. We found that PF1066 alone inhibited the growth of UMSCC25 cells, in which MET was at ASPET Journals on October 2, 2021 identified as a SLAT (Supplemental Figure 1A). Furthermore, AZ8010 alone modestly inhibited the growth of these cells, but enhanced the growth inhibition in combination with PF1066 (Fig. 5A). Over a range of doses, the combination of PF1066 and AZ8010 provoked a synergistic response in the cell line UMSCC25, with maximal synergy observed at higher doses of both drugs (Fig. 5B). To validate the screen results, we tested whether 584-A2 and CCL30 cells, in which MET was not identified as a SLAT, would respond to the combination treatment. 584-A2 cells but not CCL30 cells express

MET (Figure 2B). While PF1066 treatment moderately reduced growth of 584-A2 cells and CCL30 cells, the combination treatment of AZ8010 and PF1066 did not increase the effect of AZ8010 on these cells (Fig. 5C, Supplemental Figure 1D). UMSCC1 and

19 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

UMSCC8 cells, previously shown to respond to a combination of FGFR and ERBB family member inhibition (Fig. 4D), also showed an enhanced growth inhibition with the combination of FGFR and MET inhibition as compared to the single treatments (Fig.

5D). This effect of PF1066 was validated by knockdown of MET in UMSCC1 and

UMSCC8 cells (Supplemental Figure 6D-F). Thus, our results demonstrate that different

HNSCC cell lines rely on distinct repertoires of RTKs for growth and survival and their combined inhibition can achieve a superior growth inhibition relative to any one TKI Downloaded from alone.

We have shown that the combination of AZ8010 and AZD8931 or the molpharm.aspetjournals.org combination of AZ8010 and PF1066 can synergistically inhibit the clonogenic growth of a subset of HNSCC lines. To assess the mechanism of this growth inhibition, we used immunoblot analysis to measure pERK in extracts from UMSCC25 cells treated with different combinations of the TKIs. We have previously shown that decreased ERK at ASPET Journals on October 2, 2021 signaling occurs following inhibition of the dominant RTKs in HNSCC cell lines (Marshall et al., 2011). In Figure 6A, either AZ8010 or AZD8931 alone reduce the level of ERK phosphorylation, but the combination treatment provides a nearly complete ablation of

ERK phosphorylation (Fig. 6A, columns 1-7). In contrast, PF1066 treatment slightly increased pERK and the addition of AZ8010 lowers levels to control levels (Fig. 6A, columns 1, 8-11). Moreover, inhibition of MET phosphorylation with PF1066 slightly lowered pSTAT3 levels, suggesting MET may be one component of multiple pathways maintaining STAT3 signaling in these cells (Fig. 6A, columns 1, 8-11). We detected no effect of the TKIs on pAKT (Figure 6A).

20 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

We tested whether the combination therapies cause a decrease in clonogenic growth through an increase in apoptosis. Levels of cleaved PARP and caspase-3 increase with AZ8931 and PF1066 treatment while the combination of AZ8010 and

AZD8931 or the combination of AZ8010 and AZD8931 does not appreciably increase the levels (Fig. 6B). However, AZ8010 and PF1066 combination treatment did increase apoptosis as measured by a caspase-3 activity assay (Fig. 6C). Thus, we find that the two combination treatments function in slightly different manners to decrease Downloaded from clonogenic growth. AZ8010 and AZD8931 combination causes a decrease in proliferation through enhanced inhibition of ERK signaling while the combination of molpharm.aspetjournals.org

AZ8010 and PF1066 yields a decrease in STAT3 phosphorylation. Both treatments increase in apoptosis although it is unclear if the combination treatments cause significantly more apoptosis than the single treatments.

The triple combination of AZ8010, AZD8931 and PF1066 strongly inhibits at ASPET Journals on October 2, 2021

HNSCC cell growth accompanied by induction of apoptosis. The combined inhibition of FGFR and ERBB2 or MET provides a more complete decrease in growth than inhibition of one RTK. We postulated that UMSCC25, UMSCC1 and UMSCC8 depend on EGFR, ERBB2, MET and FGFR for maximal growth, and inhibition of all four

RTKs would be required to achieve maximal growth inhibition. UMSCC25, UMSCC1 and UMSCC8 cells were treated with AZ8010, AZD8931 and PF1066 alone, with the three distinct combinations of two inhibitors and with a combination of all three inhibitors, and the effects on clonogenic growth were measured. As shown in Figure 7A, the triple combination resulted in a significantly greater inhibition of growth relative to any of the double combinations. To determine whether the triple combination was

21 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 exerting a non-specific cytotoxic effect on the cells, we treated MSK-921 HNSCC cells in the same manner. MSK-921 cells are highly and exclusively dependent on EGFR

(data not shown). As expected, the cell line only exhibited a decrease in anchorage dependent growth when treated with the pan-ERBB inhibitor, AZD8931, with no evidence of additive effects with FGFR or MET inhibition (Fig. 7B). Likewise, growth of the highly FGFR dependent cell lines, 584-A2 and CCL30, was inhibited by AZ8010 with no evidence of additive effects by other TKIs (Fig. 7C). Importantly, the triple Downloaded from treatment lead to a significant induction in apoptosis as measured by PARP and caspase-3 cleavage (Fig. 6B) or caspase-3 activity (Fig. 6C). These results indicate a molpharm.aspetjournals.org subset of HNSCC cell lines engage multiple RTKs for growth and survival signaling and that simultaneous inhibition of multiple receptors is necessary to maximally inhibit growth.

at ASPET Journals on October 2, 2021

22 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Discussion:

Our results show that coactivation of RTK networks is critical for maintenance of growth and survival in a subset of HNSCC cell lines. Research has already identified a role for EGFR as a driver of growth in HNSCC tumors, leading to the approval of cetuximab for the treatment of the disease (Specenier et al., 2011). Also, ERBB2 and

MET have both been found to be overexpressed in HNSCC and in vitro experiments have indicated a role for these receptors as mediators of HNSCC growth (Morgan et al., Downloaded from

2009; Seiwert et al., 2009). This work shows that RTKs have the ability to function as a network such that effective growth inhibition requires their simultaneous blockade. molpharm.aspetjournals.org

Recent research is unveiling the general importance of RTK networks in cancer and the relevance to cancer cell response to targeted therapy. For example, glioblastoma cell lines and primary tumors show activation of multiple RTKs. While a dominant RTK preferentially drives downstream signaling, a secondary RTK can at ASPET Journals on October 2, 2021 maintain signaling through the pathway when the dominant RTK is inhibited (Huang et al., 2007; Stommel et al., 2007). Additionally, in breast cancer cells resistant to the

ERBB2 inhibiting antibody, trastuzumab, cells maintain growth and survival signaling through ERBB2, ERBB3 and insulin-like growth factor-I receptor (IGF-IR) heterotrimers that are formed exclusively in the resistant cells (Huang et al., 2010). Evidence from cell lines and primary tumor samples indicate that coactivation of multiple RTKs also functions in ovarian cancer (Jiao et al., 2011) and gastric cancer (Kataoka et al., 2011).

Proposed mechanisms to target these RTK networks include inhibitor combinations, multitargeted inhibitors or inhibition of common downstream signaling components (Xu et al., 2010).

23 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Our results raise significant questions about the strategies previously used to identify targets in HNSCC. Clearly, the search for effective targeted therapies for

HNSCC has been hampered by the general lack of oncogenic mutations that characterize the cancer type (Agrawal et al., 2011; Stransky et al., 2011). In fact, researchers have speculated whether the disease may, in fact, be untreatable using targeted therapies and that clinical progress can only occur through improved prevention and early detection (Agrawal et al., 2011). However, combined with Downloaded from observations that many cell lines respond with growth inhibition to EGFR inhibitors

(Marshall et al., 2011), but tumors do not respond well to the same therapy in the clinic, molpharm.aspetjournals.org these data may instead imply that non-mutational changes in the cell, such as coactivation of RTKs, are driving growth. A coupling of disruption of tumor suppressors such as p53 or p16 and the coactivation of multiple RTKs to maintain pro-growth and anti-apoptosis signaling robustness could be proposed to be driving tumorigenesis in, if at ASPET Journals on October 2, 2021 not all, then a large proportion of HNSCC tumors. Thus, we propose that developing approaches to identify and target non-mutated oncogene driver pathways and networks in HNSCC is critical for significant therapeutic advances in this cancer.

The question remains how to effectively and efficiently determine the subtype into which a particular HNSCC cell line will reside with the ultimate goal of identifying the most appropriate therapy approach for each individual patient. Our present study shows that a subset of HNSCC cell lines do respond to combined inhibition of multiple

RTKs while other cell lines seem to be highly dependent on a single RTK (Figure 7). It has been suggested that a method to overcome the robustness of a RTK network is to target a shared, downstream signaling protein (Xu et al., 2010). This strategy would

24 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 negate the need to individually tailor treatments to the specific network activated in a particular tumor. Unfortunately, we did not find a shared node downstream of the different RTKs in the three cell lines that were screened (Figure 1, Supplemental Table

2). To fully understand the heterogeneity of HNSCC, more cell lines would need to be screened, but these results suggest that downstream targets would also be subject to the drawbacks of targeting receptors. One of the strengths of an unbiased, functional screen is that it can identify pathways or genes crucial to the growth of the cancer cell Downloaded from that are not mutated or otherwise deregulated. Growing evidence indicates these sort of drivers are present in cancer and will be missed by more traditional screening methods molpharm.aspetjournals.org such as large scale sequencing (Luo et al., 2009). Unfortunately, while these pathways and genes may have vital roles and can be targeted clinically, by their very nature, they do not lend themselves to straightforward screening methods, such as SNaPshot

(Sequist et al., 2011). Therefore, work is needed to develop tractable biomarkers for at ASPET Journals on October 2, 2021 proper patient selection. It is conceivable that individual RNAi screens could be deployed in cell lines developed from each patient’s tumor in order to design a rational, personalized treatment regime.

The findings from our study provide in vitro, preclinical support for combinations of TKIs as a superior therapeutic approach to managing HNSCC. In this regard, the recent review by Glickman and Sawyers provides a compelling argument for therapy with rationally-defined combinations of targeted drugs that induce synergistic killing of cancer cells relative to the transient responses observed with monotherapies as both a practical and necessary step in the evolution of modern cancer therapeutics (Glickman et al., 2012). Their argument is based on the clear successes with combination

25 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 therapies in the management of HIV and tuberculosis infections in humans where monotherapies provide only transient responses. As many small molecule inhibitors of

RTKs are in various stages of clinical development, targeting multiple RTKs in the clinic with combinations of inhibitors is a feasible goal, albeit with technical and logistic hurdles as well.

Downloaded from

molpharm.aspetjournals.org

at ASPET Journals on October 2, 2021

26 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Authorship Contributions:

Participated in research design: Singleton, Casás-Selves, DeGregori, and Heasley.

Conducted experiments: Singleton, Hinz, Marek, and Hatheway.

Data analysis: Singleton, Kim, Tan, and Heasley.

Wrote and contributed to manuscript: Singleton, Kim, Tan, and Heasley. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021

27 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

References:

Agrawal N, Frederick MJ, Pickering CR, Bettegowda C, Chang K, Li RJ, Fakhry C, Xie TX, Zhang J, Wang J, Zhang N, El-Naggar AK, Jasser SA, Weinstein JN, Trevino L, Drummond JA, Muzny DM, Wu Y, Wood LD, Hruban RH, Westra WH, Koch WM, Califano JA, Gibbs RA, Sidransky D, Vogelstein B, Velculescu VE, Papadopoulos N, Wheeler DA, Kinzler KW and Myers JN (2011). Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1. Science 333: 1154-1157. Ang KK, Andratschke NH and Milas L (2004). Epidermal growth factor receptor and response of head-and-neck carcinoma to therapy. Int J Radiat Oncol Biol Phys 58: 959-965. Casas-Selves M, Kim J, Zhang Z, Helfrich BA, Gao D, Porter CC, Scarborough HA, Downloaded from Bunn PA, Jr., Chan DC, Tan AC and Degregori J (2012). Tankyrase and the canonical Wnt pathway protect lung cancer cells from EGFR inhibition. Cancer Res. Cecchi F, Rabe DC and Bottaro DP (2010). Targeting the HGF/Met signalling pathway in cancer. Eur J Cancer 46: 1260-1270. molpharm.aspetjournals.org Cohen EE (2006). Role of epidermal growth factor receptor pathway-targeted therapy in patients with recurrent and/or metastatic squamous cell carcinoma of the head and neck. J Clin Oncol 24: 2659-2665. Cohen EE, Rosen F, Stadler WM, Recant W, Stenson K, Huo D and Vokes EE (2003). Phase II trial of ZD1839 in recurrent or metastatic squamous cell carcinoma of the head and neck. J Clin Oncol 21: 1980-1987. Dassonville O, Formento JL, Francoual M, Ramaioli A, Santini J, Schneider M, Demard at ASPET Journals on October 2, 2021 F and Milano G (1993). Expression of epidermal growth factor receptor and survival in upper aerodigestive tract cancer. J Clin Oncol 11: 1873-1878. Druker BJ (2002). Perspectives on the development of a molecularly targeted agent. Cancer Cell 1: 31-36. Egloff AM and Grandis JR (2009). Improving Response Rates to EGFR-Targeted Therapies for Head and Neck Squamous Cell Carcinoma: Candidate Predictive Biomarkers and Combination Treatment with Src Inhibitors. J Oncol 2009: 896407. Gazdar AF (2009). Activating and resistance mutations of EGFR in non-small-cell lung cancer: role in clinical response to EGFR tyrosine kinase inhibitors. Oncogene 28 Suppl 1: S24-31. Glickman MS and Sawyers CL (2012). Converting cancer therapies into cures: lessons from infectious diseases. Cell 148: 1089-1098. Grandis JR, Melhem MF, Barnes EL and Tweardy DJ (1996). Quantitative immunohistochemical analysis of transforming growth factor-alpha and epidermal growth factor receptor in patients with squamous cell carcinoma of the head and neck. Cancer 78: 1284-1292. Grandis JR, Melhem MF, Gooding WE, Day R, Holst VA, Wagener MM, Drenning SD and Tweardy DJ (1998). Levels of TGF-alpha and EGFR protein in head and neck squamous cell carcinoma and patient survival. J Natl Cancer Inst 90: 824- 832.

28 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Haddad RI and Shin DM (2008). Recent advances in head and neck cancer. N Engl J Med 359: 1143-1154. Hickinson DM, Klinowska T, Speake G, Vincent J, Trigwell C, Anderton J, Beck S, Marshall G, Davenport S, Callis R, Mills E, Grosios K, Smith P, Barlaam B, Wilkinson RW and Ogilvie D (2010). AZD8931, an equipotent, reversible inhibitor of signaling by epidermal growth factor receptor, ERBB2 (HER2), and ERBB3: a unique agent for simultaneous ERBB receptor blockade in cancer. Clin Cancer Res 16: 1159-1169. Hopkins AL and Groom CR (2002). The druggable genome. Nat Rev Drug Discov 1: 727-730. Huang PH, Mukasa A, Bonavia R, Flynn RA, Brewer ZE, Cavenee WK, Furnari FB and White FM (2007). Quantitative analysis of EGFRvIII cellular signaling networks reveals a combinatorial therapeutic strategy for glioblastoma. Proc Natl Acad Sci Downloaded from U S A 104: 12867-12872. Huang X, Gao L, Wang S, McManaman JL, Thor AD, Yang X, Esteva FJ and Liu B (2010). Heterotrimerization of the growth factor receptors erbB2, erbB3, and insulin-like growth factor-i receptor in breast cancer cells resistant to herceptin. Cancer Res 70: 1204-1214. molpharm.aspetjournals.org Hynes NE and Lane HA (2005). ERBB receptors and cancer: the complexity of targeted inhibitors. Nat Rev Cancer 5: 341-354. Jiao Y, Ou W, Meng F, Zhou H and Wang A (2011). Targeting HSP90 in ovarian cancers with multiple receptor tyrosine kinase coactivation. Mol Cancer 10: 125. Kalyankrishna S and Grandis JR (2006). Epidermal growth factor receptor biology in head and neck cancer. J Clin Oncol 24: 2666-2672. Kataoka Y, Mukohara T, Tomioka H, Funakoshi Y, Kiyota N, Fujiwara Y, Yashiro M, at ASPET Journals on October 2, 2021 Hirakawa K, Hirai M and Minami H (2011). Foretinib (GSK1363089), a multi- kinase inhibitor of MET and VEGFRs, inhibits growth of gastric cancer cell lines by blocking inter-receptor tyrosine kinase networks. Invest New Drugs. Kim J and Tan AC (2012). BiNGS!SL-seq: a bioinformatics pipeline for the analysis and interpretation of deep sequencing genome-wide synthetic lethal screen. Methods Mol Biol 802: 389-398. Kwak EL, Bang YJ, Camidge DR, Shaw AT, Solomon B, Maki RG, Ou SH, Dezube BJ, Janne PA, Costa DB, Varella-Garcia M, Kim WH, Lynch TJ, Fidias P, Stubbs H, Engelman JA, Sequist LV, Tan W, Gandhi L, Mino-Kenudson M, Wei GC, Shreeve SM, Ratain MJ, Settleman J, Christensen JG, Haber DA, Wilner K, Salgia R, Shapiro GI, Clark JW and Iafrate AJ (2010). Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. N Engl J Med 363: 1693-1703. Ladanyi M and Pao W (2008). Lung adenocarcinoma: guiding EGFR-targeted therapy and beyond. Mod Pathol 21 Suppl 2: S16-22. Leemans CR, Braakhuis BJ and Brakenhoff RH (2011). The molecular biology of head and neck cancer. Nat Rev Cancer 11: 9-22. Licitra L, Mesia R, Rivera F, Remenar E, Hitt R, Erfan J, Rottey S, Kawecki A, Zabolotnyy D, Benasso M, Storkel S, Senger S, Stroh C and Vermorken JB (2011). Evaluation of EGFR gene copy number as a predictive biomarker for the efficacy of cetuximab in combination with chemotherapy in the first-line treatment

29 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

of recurrent and/or metastatic squamous cell carcinoma of the head and neck: EXTREME study. Ann Oncol 22: 1078-1087. Luo J, Solimini NL and Elledge SJ (2009). Principles of cancer therapy: oncogene and non-oncogene addiction. Cell 136: 823-837. Marshall ME, Hinz TK, Kono SA, Singleton KR, Bichon B, Ware KE, Marek L, Frederick BA, Raben D and Heasley LE (2011). Fibroblast growth factor receptors are components of autocrine signaling networks in head and neck squamous cell carcinoma cells. Clin Cancer Res 17: 5016-5025. Matassa AA, Carpenter L, Biden TJ, Humphries MJ and Reyland ME (2001). PKCdelta is required for mitochondrial-dependent apoptosis in salivary epithelial cells. J Biol Chem 276: 29719-29728. Morgan S and Grandis JR (2009). ErbB receptors in the biology and pathology of the aerodigestive tract. Exp Cell Res 315: 572-582. Downloaded from Morse DL and Gillies RJ (2010). Molecular imaging and targeted therapies. Biochem Pharmacol 80: 731-738. Murdoch D (2007). Standard, and novel cytotoxic and molecular-targeted, therapies for HNSCC: an evidence-based review. Curr Opin Oncol 19: 216-221. O'Dwyer ME, Mauro MJ, Blasdel C, Farnsworth M, Kurilik G, Hsieh YC, Mori M and molpharm.aspetjournals.org Druker BJ (2004). Clonal evolution and lack of cytogenetic response are adverse prognostic factors for hematologic relapse of chronic phase CML patients treated with imatinib mesylate. Blood 103: 451-455. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO and Botstein D (2000). Molecular portraits of human breast tumours. Nature 406: 747-752. at ASPET Journals on October 2, 2021 Porter CC, Kim J, Fosmire S, Gearheart CM, van Linden A, Baturin D, Zaberezhnyy V, Patel PR, Gao D, Tan AC and Degregori J (2012). Integrated genomic analyses identify WEE1 as a critical mediator of cell fate and a novel therapeutic target in acute myeloid leukemia. Leukemia 26: 1266-1276. Robinson MD and Smyth GK (2007). Moderated statistical tests for assessing differences in tag abundance. Bioinformatics 23: 2881-2887. Sato-Kuwabara Y, Neves JI, Fregnani JH, Sallum RA and Soares FA (2009). Evaluation of gene amplification and protein expression of HER-2/neu in esophageal squamous cell carcinoma using Fluorescence in situ Hybridization (FISH) and immunohistochemistry. BMC Cancer 9: 6. Seiwert TY, Jagadeeswaran R, Faoro L, Janamanchi V, Nallasura V, El Dinali M, Yala S, Kanteti R, Cohen EE, Lingen MW, Martin L, Krishnaswamy S, Klein-Szanto A, Christensen JG, Vokes EE and Salgia R (2009). The MET receptor tyrosine kinase is a potential novel therapeutic target for head and neck squamous cell carcinoma. Cancer Res 69: 3021-3031. Sequist LV, Heist RS, Shaw AT, Fidias P, Rosovsky R, Temel JS, Lennes IT, Digumarthy S, Waltman BA, Bast E, Tammireddy S, Morrissey L, Muzikansky A, Goldberg SB, Gainor J, Channick CL, Wain JC, Gaissert H, Donahue DM, Muniappan A, Wright C, Willers H, Mathisen DJ, Choi NC, Baselga J, Lynch TJ, Ellisen LW, Mino-Kenudson M, Lanuti M, Borger DR, Iafrate AJ, Engelman JA

30 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

and Dias-Santagata D (2011). Implementing multiplexed genotyping of non- small-cell lung cancers into routine clinical practice. Ann Oncol 22: 2616-2624. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, Tan EH, Hirsh V, Thongprasert S, Campos D, Maoleekoonpiroj S, Smylie M, Martins R, van Kooten M, Dediu M, Findlay B, Tu D, Johnston D, Bezjak A, Clark G, Santabarbara P and Seymour L (2005). Erlotinib in previously treated non-small-cell lung cancer. N Engl J Med 353: 123-132. Soulieres D, Senzer NN, Vokes EE, Hidalgo M, Agarwala SS and Siu LL (2004). Multicenter phase II study of erlotinib, an oral epidermal growth factor receptor tyrosine kinase inhibitor, in patients with recurrent or metastatic squamous cell cancer of the head and neck. J Clin Oncol 22: 77-85. Specenier P and Vermorken JB (2011). Cetuximab in the treatment of squamous cell carcinoma of the head and neck. Expert Rev Anticancer Ther 11: 511-524. Downloaded from Stambuk S, Sundov D, Kuret S, Beljan R and Andelinovic S (2010). Future perspectives of personalized oncology. Coll Antropol 34: 763-769. Stommel JM, Kimmelman AC, Ying H, Nabioullin R, Ponugoti AH, Wiedemeyer R, Stegh AH, Bradner JE, Ligon KL, Brennan C, Chin L and DePinho RA (2007). Coactivation of receptor tyrosine kinases affects the response of tumor cells to molpharm.aspetjournals.org targeted therapies. Science 318: 287-290. Stransky N, Egloff AM, Tward AD, Kostic AD, Cibulskis K, Sivachenko A, Kryukov GV, Lawrence MS, Sougnez C, McKenna A, Shefler E, Ramos AH, Stojanov P, Carter SL, Voet D, Cortes ML, Auclair D, Berger MF, Saksena G, Guiducci C, Onofrio RC, Parkin M, Romkes M, Weissfeld JL, Seethala RR, Wang L, Rangel- Escareno C, Fernandez-Lopez JC, Hidalgo-Miranda A, Melendez-Zajgla J, Winckler W, Ardlie K, Gabriel SB, Meyerson M, Lander ES, Getz G, Golub TR, at ASPET Journals on October 2, 2021 Garraway LA and Grandis JR (2011). The mutational landscape of head and neck squamous cell carcinoma. Science 333: 1157-1160. Sullivan KD, Padilla-Just N, Henry RE, Porter CC, Kim J, Tentler JJ, Eckhardt SG, Tan AC, Degregori J and Espinosa JM (2012). ATM and MET kinases are synthetic lethal with nongenotoxic activation of p53. Nat Chem Biol. Whitlock MC (2005). Combining probability from independent tests: the weighted Z- method is superior to Fisher's approach. J Evol Biol 18: 1368-1373. Xu AM and Huang PH (2010). Receptor tyrosine kinase coactivation networks in cancer. Cancer Res 70: 3857-3860. Zou HY, Li Q, Lee JH, Arango ME, McDonnell SR, Yamazaki S, Koudriakova TB, Alton G, Cui JJ, Kung PP, Nambu MD, Los G, Bender SL, Mroczkowski B and Christensen JG (2007). An orally available small-molecule inhibitor of c-Met, PF- 2341066, exhibits cytoreductive antitumor efficacy through antiproliferative and antiangiogenic mechanisms. Cancer Res 67: 4408-4417.

31 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Footnotes:

The studies were supported by a National Institutes of Health grant [R01 CA127105].

Downloaded from

molpharm.aspetjournals.org

at ASPET Journals on October 2, 2021

32 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

Figure Legends:

FIGURE 1: Synthetic lethal interactions with FGFR inhibition are identified using a genome-wide RNAi-based loss-of-function screen. A) As described in the Materials and Methods, the HNSCC cell lines UMSCC25, CCL30 and 584-A2 were transduced with a genome-wide shRNA library and either treated with the FGFR inhibitor, AZ8010 or DMSO control for 72 hours. After recovering for an additional 72 hours, RNA was Downloaded from isolated from the cells and nested PCR was performed to isolate the shRNA sequences and to add Illumina adaptor sequences. Samples were submitted to Illumina sequencing molpharm.aspetjournals.org and the reads were analyzed by BiNGS! B) Tabular representation of the shRNA counts obtained by deep sequencing for the ten most highly ranked SLATs in UMSCC25 cells as well as ERBB2. The intensity of red indicates the number of shRNA read counts in each replicate for both the control and AZ8010 treated groups. Each line corresponds to at ASPET Journals on October 2, 2021 an independent shRNA targeting the indicated gene. The top ten hits and the data for

ERBB2 for UMSCC25 cells are shown. C) Synthetic lethal hits (SLATs) (E value ≤ 20) among growth receptors and their common downstream signaling pathways. SLATs identified in the UMSCC25 cells are shaded red, SLATs in 584-A2 cells are shaded green, and SLATs in CCL30 cells are shaded blue. Pathway members that were not identified as SLATs in any cell line are unshaded. The number listed near each SLAT is the p-value (see Materials and Methods).

FIGURE 2: A panel of HNSCC cell lines shows variable expression of ERBB2 and

MET protein. Whole cell lysates of the indicated HNSCC cell lines were collected and

33 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 immunoblotted for A) ERBB2 or B) MET. Filters were stripped and reprobed for the α- subunit of Na/K ATPase as a loading control.

FIGURE3: Confirmation of ERBB2 and MET as synthetic lethal genes with FGFR inhibition by targeted knockdown. A) Pooled UMSCC25 cells expressing either a negative control shRNA targeting GFP or two independent shRNAs targeting either

ERBB2 or MET were generated and ERBB2 or MET protein expression was measured Downloaded from by immunoblotting. The protein level of the α-subunit of Na/K-ATPase was measured as a loading control. B) Pooled UMSCC25 cells expressing the control shRNA or one of molpharm.aspetjournals.org two shRNAs targeting ERBB2 were used in a clonogenic growth assay in the presence or absence of 1.0 μM AZ8010. After two weeks, colonies were stained and quantified as described in the Materials and Methods. C) Pooled UMSCC25 cells expressing the control shRNA or one of two shRNAs targeting MET were used in a clonogenic growth at ASPET Journals on October 2, 2021 assay in the presence or absence of 1.0 μM AZ8010. After two weeks, colonies were stained and quantified as described in the Materials and Methods. The data are the means and SEM of three replicates from three independent experiments; ns indicates not significant, * indicates p<0.05 and ** indicates p<0.005 using a two-tailed Student’s t-test.

FIGURE 4: Combined inhibition of ERBB family members and FGFR leads to synergistic impairment of the growth of HNSCC cell lines. A) UMSCC25 cells were treated in triplicate with DMSO, 3.0 nM AZD8931 or 10 nM AZD8931 alone or in combination with 0.5 μM AZ8010 or 1.0 μM AZ8010. The effect of the treatments on cell

34 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 growth was measured by clonogenic assay. B) UMSCC25 cells were plated in triplicate and treated with DMSO, AZ8010, AZD8931 and the combinations at the doses shown.

Proliferation was measured after one week by the CyQuant assay as described in the

Materials and Methods and plotted on the top chart. Data are the means of two independent experiments. Combination Indices (CI) were calculated using the Calcusyn program and shown on the bottom table where the strength of the synergism is indicated by the darkness of shading. C) As in A, except 584-A2 cells were treated with Downloaded from

DMSO, 30 nM AZD8931, 30 nM AZ8010 and the combination. CCL30 cells were treated with DMSO, 100 nM AZD8931, 100 nM AZ8010 and the combination. The molpharm.aspetjournals.org growth of 584-A2 cells was measured by clonogenic assay and the growth of CCL30 cells was measured using an anchorage independent growth assay as described in the

Materials and Methods. D) As in A, except UMSCC1 and UMSCC8 cells were treated with DMSO, 30 nM AZD8931, 30 nM AZ8010 and the combination. Data are the means at ASPET Journals on October 2, 2021 and SEM of three independent experiments. Statistical analysis using a two-tailed

Student’s t-test showed significant differences in clonogenic growth where ns indicates not significant, * indicates p<0.05, ** indicates p<0.005 and **** indicates p<0.0001.

FIGURE 5: Combined inhibition of MET and FGFR leads to synergistic impairment of the growth of HNSCC cell lines. A) UMSCC25 cells were treated in triplicate with

DMSO, 0.1 μM PF1066 or 0.5 μM PF1066 alone or in combination with 0.5 μM AZ8010 or 1.0 μM AZ8010. The effect of the treatments on cell growth was measured by clonogenic assay. B) UMSCC25 cells were plated in triplicate and treated with DMSO,

AZ8010, PF1066 and the combinations at the doses shown. Proliferation was measured

35 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111 by CyQuant assay as described in the Materials and Methods and plotted on the top chart. Data are the means of two independent experiments. Combination Indices (CI) were calculated and shown on the bottom table. C) As in A, except 584-A2 cells were treated with DMSO, 0.5 μM PF1066, 30 nM AZ8010 and the combination and growth was measured by clonogenic assay. CCL30 cells were treated with DMSO, 0.3 μM

PF1066, 0.1 μM AZ8010 and the combination and growth was measured by anchorage independent growth assay. D) As in A, except UMSCC1 and UMSCC8 cells were Downloaded from treated with DMSO, 0.3 μM PF1066, 0.3 μM AZ8010 and the combination and growth was measured by clonogenic assay. Data are the means and SEM of three independent molpharm.aspetjournals.org experiments. Statistical analysis using a two-tailed Student’s t-test was used to show significant differences in growth where ns indicates not significant, * indicates p<0.05, ** indicates p<0.005, *** indicates p<0.0005 and **** indicates p<0.0001.

at ASPET Journals on October 2, 2021

FIGURE 6: Effects of combined FGFR, ERBB family and MET inhibition on ERK signaling and PARP cleavage. A) UMSCC25 cells were treated with the indicated doses (μM) of DMSO, AZ8010, AZD8931 or PF1066 alone and in combination for 6 hours in full media. Extracts were prepared and proteins resolved by SDS-PAGE. The filters were immunoblotted for pERK, pSTAT3 and pAKT. The filters were then stripped and reprobed for total ERK1/2, total STAT3, total AKT and the α-subunit of NA/K

ATPase as a loading control. B) UMSCC25 cells were treated as with the drugs indicated in A) with the doses (μM) shown for 72 hours. Extracts were prepared and proteins resolved by SDS-PAGE. The filters were immunoblotted for PARP. The filters were then stripped and reprobed for the α-subunit of NA/K ATPase as a loading control.

36 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version.

MOL #84111

C) UMSCC25 cells were treated as described in B) and caspase activity was measured as described in the Materials and Methods.

FIGURE 7: Simultaneous inhibition of FGFR, ERBB family receptors and MET leads to greater growth inhibition compared to the double combinations in a subset of HNSCC cell lines. A) UMSCC25, UMSCC1 and UMSCC25 were treated in triplicate with 0.5 μM AZ8010, 10 nM AZD8931, 0.1 μM PF1066, the respective double Downloaded from combinations and the triple combination. Growth was measured by clonogenic assay and quantified after 2 weeks. B) MSK-921 cells were treated in triplicate as in A. Growth molpharm.aspetjournals.org was measured via anchorage independent growth assay. C) 584-A2 cells were treated in triplicate with 0.1 μM AZ8010, 10 nM AZD8931, 0.1 μM PF1066, the respective double combinations and the triple combination. Growth was measured after 2 weeks by clonogenic assay. CCL30 cells were treated in triplicate with 0.1 μM AZ8010, 10 nM at ASPET Journals on October 2, 2021

AZD8931, 0.1 μM PF1066, the respective double combinations and the triple combination. Growth was measured after 2 weeks using an anchorage independent growth assay. Data are the means and SEM of three independent experiments.

Statistical analysis using a two-tailed Student’s t-test revealed significant differences in growth where ns indicates not significant, * indicates p<0.05, ** indicates p<0.005, *** indicates p<0.0005 and **** indicates p<0.0001.

37 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Molecular Pharmacology Fast Forward. Published on January 31, 2013 as DOI: 10.1124/mol.112.084111 This article has not been copyedited and formatted. The final version may differ from this version. Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 2, 2021 Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL TABLE 1: List of shRNAs used for validation.

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL TABLE 2: SLATs (E-value cut off of E value ≤2) in UMSCC25, 584-A2, and CCL30. UMSCC25 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 1 CLUAP1 5.15E-07 5.15E-07 48 FHL1 4.84E-04 2.32E-02 2 ZFR 8.78E-07 1.76E-06 49 RABGGTB 5.03E-04 2.46E-02 3 MET 1.94E-06 5.83E-06 50 THPO 5.28E-04 2.64E-02 4 MEMO1 1.95E-06 7.79E-06 50 BCAP29 5.28E-04 2.64E-02 5 NPAT 3.00E-06 1.50E-05 52 KBTBD4 5.28E-04 2.75E-02 6 EPB41L4B 3.18E-06 1.91E-05 53 MEX3B 5.28E-04 2.80E-02 7 G2E3 3.43E-06 2.40E-05 54 EML4 5.29E-04 2.85E-02 8 PSMD13 4.10E-06 3.28E-05 55 C20orf7 5.29E-04 2.91E-02 9 MUC7 5.60E-06 5.04E-05 56 GPR126 5.29E-04 2.96E-02 10 ARSD 5.67E-06 5.67E-05 56 PNPO 5.29E-04 2.96E-02 11 TMEM163 1.13E-05 1.24E-04 58 TJAP1 5.29E-04 3.07E-02 12 PCTK2 1.38E-05 1.65E-04 58 SLC26A2 5.29E-04 3.07E-02 13 ANP32E 1.38E-05 1.80E-04 58 NETO2 5.29E-04 3.07E-02 14 ITGA6 2.62E-05 3.66E-04 61 MAML3 5.29E-04 3.23E-02 15 C14orf147 2.74E-05 4.11E-04 62 HLA-DRB3 5.30E-04 3.29E-02 16 MDH1B 3.18E-05 5.08E-04 62 HLA-DRB1 5.30E-04 3.29E-02 17 SP1 3.68E-05 6.25E-04 62 HLA-DRB5 5.30E-04 3.29E-02 18 ALDOB 4.96E-05 8.92E-04 62 SPRYD5 5.30E-04 3.29E-02 19 ANAPC5 5.12E-05 9.72E-04 66 SYT14L 5.30E-04 3.50E-02 20 FCGR3B 5.17E-05 1.03E-03 67 DGAT2 5.31E-04 3.56E-02 21 TBXA2R 5.39E-05 1.13E-03 68 ZNF10 5.32E-04 3.62E-02 22 MT1X 5.80E-05 1.28E-03 69 ZNF764 5.32E-04 3.67E-02 23 SCRN3 6.54E-05 1.51E-03 70 E4F1 5.32E-04 3.73E-02 24 PAPOLA 7.53E-05 1.81E-03 70 NOV 5.32E-04 3.73E-02 25 FGFR1OP2 7.54E-05 1.88E-03 70 SSBP3 5.32E-04 3.73E-02 26 KCTD14 1.06E-04 2.75E-03 73 COL6A1 5.33E-04 3.89E-02 27 LLGL2 1.39E-04 3.76E-03 74 PLEKHA6 5.33E-04 3.94E-02 28 PYGM 1.45E-04 4.06E-03 75 TMEM196 5.33E-04 4.00E-02 29 MFI2 1.52E-04 4.40E-03 76 DCN 5.33E-04 4.05E-02 30 FAM136A 1.67E-04 5.02E-03 77 ZNF280B 5.34E-04 4.11E-02 31 ZNF787 1.68E-04 5.22E-03 78 PDCD11 5.34E-04 4.17E-02 32 FPR2 1.89E-04 6.04E-03 79 CXADR 5.35E-04 4.22E-02 33 TMED5 2.07E-04 6.83E-03 79 HLA-G 5.35E-04 4.22E-02 34 PTP4A1 2.10E-04 7.16E-03 81 RUNX2 5.35E-04 4.33E-02 35 LSM4 2.16E-04 7.57E-03 82 ERBB2 5.35E-04 4.39E-02 36 FARSA 2.26E-04 8.13E-03 82 TRIB2 5.35E-04 4.39E-02 37 AGRN 2.37E-04 8.76E-03 84 C12orf48 5.36E-04 4.50E-02 38 FAM172A 2.52E-04 9.58E-03 85 MAST1 5.37E-04 4.56E-02 39 PTPN11 2.58E-04 1.01E-02 86 THAP10 5.37E-04 4.62E-02 40 TBX5 2.74E-04 1.10E-02 87 BCL11A 5.38E-04 4.68E-02 41 MBD6 3.55E-04 1.46E-02 88 CCDC88A 5.38E-04 4.74E-02 42 TRAM1 3.99E-04 1.68E-02 89 MARS2 5.39E-04 4.80E-02 43 VEGFA 4.12E-04 1.77E-02 90 C3AR1 5.40E-04 4.86E-02 44 ZNF79 4.13E-04 1.82E-02 91 KIAA0892 5.41E-04 4.92E-02 45 SCOC 4.14E-04 1.86E-02 92 POLD3 5.41E-04 4.98E-02 46 SEMA4D 4.42E-04 2.03E-02 93 C2orf52 5.44E-04 5.06E-02 47 CALM1 4.44E-04 2.09E-02 94 FSTL5 5.44E-04 5.12E-02

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

UMSCC25 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 94 ZNF511 5.44E-04 5.12E-02 142 PLEKHO1 5.79E-04 8.22E-02 94 CD55 5.44E-04 5.12E-02 143 ZFP64 5.83E-04 8.33E-02 94 ST5 5.44E-04 5.12E-02 144 RPL28 5.83E-04 8.39E-02 94 NYX 5.44E-04 5.12E-02 145 C5orf45 5.86E-04 8.50E-02 99 FLJ22662 5.45E-04 5.39E-02 145 SQSTM1 5.86E-04 8.50E-02 100 BCAS2 5.46E-04 5.46E-02 147 COL5A3 5.88E-04 8.65E-02 100 FRAT2 5.46E-04 5.46E-02 148 CDV3 5.92E-04 8.76E-02 102 MAP7 5.46E-04 5.57E-02 148 SF3B14 5.92E-04 8.76E-02 103 KRT222P 5.46E-04 5.63E-02 150 HGD 5.95E-04 8.92E-02 104 AGPAT9 5.46E-04 5.68E-02 150 NIACR2 5.95E-04 8.92E-02 105 FAT2 5.46E-04 5.74E-02 150 RNF20 5.95E-04 8.92E-02 106 ITIH5 5.46E-04 5.79E-02 153 VWA5A 5.97E-04 9.13E-02 106 RAB3IP 5.46E-04 5.79E-02 153 PNPLA8 5.97E-04 9.13E-02 108 CYP2A7 5.48E-04 5.91E-02 155 FOXN3 6.00E-04 9.30E-02 109 FBXO31 5.48E-04 5.98E-02 156 LFNG 6.00E-04 9.37E-02 109 JCLN 5.48E-04 5.98E-02 157 PRKAB2 6.01E-04 9.44E-02 109 RPL24 5.48E-04 5.98E-02 158 CISH 6.07E-04 9.59E-02 112 MLH3 5.52E-04 6.18E-02 159 ACOX3 6.08E-04 9.66E-02 113 TGOLN2 5.53E-04 6.25E-02 159 FSCN3 6.08E-04 9.66E-02 114 C21orf62 5.54E-04 6.31E-02 161 PIP4K2C 6.09E-04 9.80E-02 115 ZNF407 5.55E-04 6.38E-02 162 SORBS2 6.11E-04 9.89E-02 116 KIR2DL1 5.56E-04 6.45E-02 162 SKP2 6.11E-04 9.89E-02 116 GTF2I 5.56E-04 6.45E-02 164 MGLL 6.11E-04 0.1003 116 KIR2DL2 5.56E-04 6.45E-02 165 CEACAM7 6.12E-04 0.1010 119 ARL6IP5 5.57E-04 6.63E-02 166 GNB4 6.12E-04 0.1016 120 YAF2 5.57E-04 6.69E-02 167 COL5A2 6.12E-04 0.1023 121 RSF1 5.59E-04 6.76E-02 168 PHB2 6.15E-04 0.1033 122 MMP14 5.63E-04 6.87E-02 169 PTBP2 6.16E-04 0.1041 123 PPAP2B 5.64E-04 6.94E-02 170 DCUN1D1 6.17E-04 0.1048 123 BCOR 5.64E-04 6.94E-02 170 OSTM1 6.17E-04 0.1048 125 ABTB2 5.65E-04 7.06E-02 172 CLCN7 6.17E-04 0.1061 125 OGFR 5.65E-04 7.06E-02 173 PTP4A2 6.21E-04 0.1075 127 TMPO 5.67E-04 7.20E-02 173 TACR2 6.21E-04 0.1075 128 FASTKD1 5.70E-04 7.30E-02 173 WWC2 6.21E-04 0.1075 129 DOK7 5.71E-04 7.36E-02 176 GNRHR 6.26E-04 0.1101 130 TRIOBP 5.71E-04 7.42E-02 176 PCDHB5 6.26E-04 0.1101 130 PDE1A 5.71E-04 7.42E-02 178 CBX3 6.26E-04 0.1114 132 TRAM2 5.71E-04 7.54E-02 179 PDS5A 6.30E-04 0.1127 133 THAP11 5.72E-04 7.61E-02 180 FAM62B 6.34E-04 0.1141 134 CNOT6L 5.74E-04 7.69E-02 180 CRYAA 6.34E-04 0.1141 134 TEX2 5.74E-04 7.69E-02 182 TRIM27 6.36E-04 0.1157 136 SEPT6 5.75E-04 7.82E-02 183 MLXIP 6.36E-04 0.1165 137 HSP90B1 5.76E-04 7.90E-02 184 HRASLS 6.40E-04 0.1178 138 HNRNPR 5.76E-04 7.96E-02 185 AVL9 6.40E-04 0.1185 138 PITPNM2 5.76E-04 7.96E-02 185 IGF2BP3 6.40E-04 0.1185 140 NFIB 5.77E-04 8.08E-02 187 CSNK1G2 6.42E-04 0.1200 141 SYT14 5.78E-04 8.16E-02 188 EPO 6.47E-04 0.1216

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

UMSCC25 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 188 PDZRN3 6.47E-04 0.1216 234 SLC38A3 7.58E-04 0.1774 188 CBX7 6.47E-04 0.1216 234 CXorf48 7.58E-04 0.1774 188 PHF8 6.47E-04 0.1216 234 ATF5 7.58E-04 0.1774 188 SLC1A4 6.47E-04 0.1216 239 SNCA 7.59E-04 0.1813 193 TMEM49 6.53E-04 0.1261 240 SRR 7.63E-04 0.1832 193 CC2D1A 6.53E-04 0.1261 240 TSR1 7.63E-04 0.1832 195 PPIG 6.53E-04 0.1274 242 SHB 7.65E-04 0.1851 196 SNX25 6.59E-04 0.1291 243 PCDHGA1 7.71E-04 0.1873 197 RAB9A 6.59E-04 0.1299 244 GNL3 7.73E-04 0.1886 198 PFTK1 6.60E-04 0.1307 244 KCNQ5 7.73E-04 0.1886 198 NCLN 6.60E-04 0.1307 244 ST3GAL3 7.73E-04 0.1886 198 TMEM53 6.60E-04 0.1307 244 CACYBP 7.73E-04 0.1886 201 RNF114 6.66E-04 0.1339 244 VGLL4 7.73E-04 0.1886 202 DHX8 6.67E-04 0.1348 249 PSME3 7.79E-04 0.1939 203 USP38 6.68E-04 0.1355 250 ISY1 7.79E-04 0.1948 204 TINAGL1 6.75E-04 0.1377 250 MED17 7.79E-04 0.1948 205 PIP5K1A 6.77E-04 0.1387 252 TTC22 7.82E-04 0.1970 206 PCM1 6.81E-04 0.1403 253 ICA1L 7.82E-04 0.1978 207 SPIN2B 6.82E-04 0.1411 254 ATF6B 7.90E-04 0.2006 207 SPIN2A 6.82E-04 0.1411 255 UTRN 7.90E-04 0.2014 209 OR5L1 6.86E-04 0.1435 256 NET1 7.92E-04 0.2028 209 OR5L2 6.86E-04 0.1435 257 EHD1 7.94E-04 0.2041 209 REV1 6.86E-04 0.1435 258 SPAM1 7.95E-04 0.2051 212 TAF6 6.87E-04 0.1457 259 MLLT10 8.03E-04 0.2080 213 DPY19L4 6.90E-04 0.1470 260 STRADB 8.08E-04 0.2102 214 PDLIM7 6.91E-04 0.1478 261 ARL15 8.10E-04 0.2115 215 KCNQ1 6.98E-04 0.1502 262 DECR2 8.14E-04 0.2132 216 LATS2 7.09E-04 0.1532 262 ALOX15B 8.14E-04 0.2132 216 PLA2G2E 7.09E-04 0.1532 264 AP1S2 8.30E-04 0.2191 218 PPAT 7.17E-04 0.1562 264 ZBTB41 8.30E-04 0.2191 219 NFX1 7.18E-04 0.1573 264 ZNF19 8.30E-04 0.2191 220 KREMEN2 7.19E-04 0.1582 264 KAZALD1 8.30E-04 0.2191 220 RAB1A 7.19E-04 0.1582 268 TM9SF4 8.33E-04 0.2232 220 LRP4 7.19E-04 0.1582 269 IL15RA 8.33E-04 0.2242 223 C1orf116 7.23E-04 0.1613 270 ZNF160 8.41E-04 0.2271 224 HEXIM1 7.31E-04 0.1638 271 ZNF804A 8.43E-04 0.2285 225 FNDC3B 7.32E-04 0.1647 272 SLC11A2 8.51E-04 0.2314 225 AQP1 7.32E-04 0.1647 273 NSUN7 8.53E-04 0.2327 227 RHOQ 7.32E-04 0.1663 274 HISPPD2A 8.57E-04 0.2349 228 P2RY14 7.40E-04 0.1688 274 MYOZ3 8.57E-04 0.2349 229 SERP1 7.42E-04 0.1700 276 LUZP2 8.58E-04 0.2369 230 P4HTM 7.45E-04 0.1714 277 FHOD3 8.61E-04 0.2385 230 PDX1 7.45E-04 0.1714 278 SLC2A12 8.73E-04 0.2428 230 HIF3A 7.45E-04 0.1714 279 MC3R 8.77E-04 0.2448 233 KCNJ1 7.54E-04 0.1758 280 ATM 8.84E-04 0.2475 234 RPL3 7.58E-04 0.1774 280 NBS 8.84E-04 0.2475 234 CSNK2A1 7.58E-04 0.1774 280 LRRC61 8.84E-04 0.2475

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

UMSCC25 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 280 NBN 8.84E-04 0.2475 330 C12orf41 1.14E-03 0.3766 284 PGRMC1 8.91E-04 0.2530 330 P76 1.14E-03 0.3766 285 CHRNB3 9.10E-04 0.2594 332 ZNF184 1.14E-03 0.3792 286 C22orf29 9.12E-04 0.2609 333 ADRBK1 1.15E-03 0.3817 287 LCE1E 9.12E-04 0.2619 334 C20orf20 1.15E-03 0.3836 287 OGT 9.12E-04 0.2619 335 DUSP6 1.20E-03 0.4016 287 STAT2 9.12E-04 0.2619 336 PSME4 1.20E-03 0.4033 290 LRRC66 9.25E-04 0.2683 337 SERPINB3 1.20E-03 0.4047 290 RUSC2 9.25E-04 0.2683 338 PIM2 1.21E-03 0.4073 292 PIAS2 9.40E-04 0.2745 339 NUBPL 1.21E-03 0.4114 293 ACTR8 9.45E-04 0.2770 339 TRPS1 1.21E-03 0.4114 293 SMURF2 9.45E-04 0.2770 339 CYP2C19 1.21E-03 0.4114 293 COX5A 9.45E-04 0.2770 339 ARFIP1 1.21E-03 0.4114 293 ARMC10 9.45E-04 0.2770 339 RASGRP2 1.21E-03 0.4114 293 PDE4C 9.45E-04 0.2770 339 FAM35A 1.21E-03 0.4114 298 CNGA3 9.46E-04 0.2819 345 TBX2 1.22E-03 0.4199 298 HTR5A 9.46E-04 0.2819 346 MFF 1.23E-03 0.4257 300 ADCY7 9.66E-04 0.2899 347 ENOX2 1.25E-03 0.4331 301 ATPAF2 9.68E-04 0.2914 348 CDK10 1.26E-03 0.4382 302 SLC2A4RG 9.69E-04 0.2925 349 HIF1AN 1.27E-03 0.4448 303 AIFM1 9.83E-04 0.2979 350 SECISBP2 1.28E-03 0.4464 303 KRT19 9.83E-04 0.2979 351 DNAH7 1.30E-03 0.4565 305 NR5A2 9.88E-04 0.3013 352 PRKG2 1.30E-03 0.4579 305 LEPROT 9.88E-04 0.3013 353 SLC1A1 1.30E-03 0.4603 307 CPEB3 9.94E-04 0.3050 354 TOX4 1.31E-03 0.4635 308 LOC57228 9.94E-04 0.3063 354 CD300A 1.31E-03 0.4635 309 GJD3 1.00E-03 0.3092 354 BCAR3 1.31E-03 0.4635 310 STAB1 1.01E-03 0.3136 354 TAF6L 1.31E-03 0.4635 311 FANCC 1.02E-03 0.3161 354 MAG 1.31E-03 0.4635 311 BST1 1.02E-03 0.3161 359 SNX26 1.33E-03 0.4786 311 ARHGDIB 1.02E-03 0.3161 360 PEBP1 1.33E-03 0.4806 311 ADA 1.02E-03 0.3161 361 HTR3A 1.34E-03 0.4852 315 CENPB 1.02E-03 0.3228 362 FBXO24 1.38E-03 0.4989 315 UGT2B15 1.02E-03 0.3228 363 PIGF 1.38E-03 0.5026 315 PJA1 1.02E-03 0.3228 364 CYP2C9 1.39E-03 0.5046 318 CUL3 1.05E-03 0.3330 365 RBM39 1.39E-03 0.5065 319 CALCB 1.07E-03 0.3410 366 CDC2L1 1.42E-03 0.5203 320 CCDC117 1.08E-03 0.3455 366 CDC2L2 1.42E-03 0.5203 320 BIVM 1.08E-03 0.3455 368 MIPOL1 1.43E-03 0.5245 320 TUBGCP5 1.08E-03 0.3455 368 ZC3HAV1L 1.43E-03 0.5245 320 ETS2 1.08E-03 0.3455 370 GPR18 1.45E-03 0.5374 320 AQP4 1.08E-03 0.3455 371 MSRB2 1.50E-03 0.5581 320 GABRG1 1.08E-03 0.3455 372 HMGB1 1.53E-03 0.5678 320 TRPC4AP 1.08E-03 0.3455 372 HMGB1L1 1.53E-03 0.5678 327 STAT1 1.09E-03 0.3569 372 SP100 1.53E-03 0.5678 328 GFOD1 1.10E-03 0.3622 372 BAT4 1.53E-03 0.5678 329 PHF6 1.11E-03 0.3647 372 LSM1 1.53E-03 0.5678

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

UMSCC25 p-value p-value Rank Gene Symbol [p(wZ)] E value Rank Gene Symbol [p(wZ)] E value 377 LRRC50 1.53E-03 0.5768 424 CHKA 1.92E-03 0.8123 378 WDR17 1.53E-03 0.5798 425 SUPT6H 1.93E-03 0.8195 379 TMEM14C 1.54E-03 0.5824 426 CBLC 1.95E-03 0.8326 380 ECHDC1 1.55E-03 0.5890 427 NOP56 1.96E-03 0.8364 381 UBOX5 1.56E-03 0.5928 427 ATP5SL 1.96E-03 0.8364 381 C16orf72 1.56E-03 0.5928 429 FUSIP1 1.99E-03 0.8519 381 SHMT2 1.56E-03 0.5928 430 MGAT1 2.04E-03 0.8757 381 MED8 1.56E-03 0.5928 431 KRCC1 2.04E-03 0.8798 381 CACNA1H 1.56E-03 0.5928 432 NEDD4 2.11E-03 0.9122 381 GALNT2 1.56E-03 0.5928 432 CST1 2.11E-03 0.9122 381 SOX2 1.56E-03 0.5928 432 RUNDC3B 2.11E-03 0.9122 381 CFH 1.56E-03 0.5928 432 FBXL8 2.11E-03 0.9122 381 RPSA 1.56E-03 0.5928 432 TIMM44 2.11E-03 0.9122 390 EYA3 1.57E-03 0.6119 432 HSD17B7 2.11E-03 0.9122 391 FTHL17 1.61E-03 0.6276 432 OR51E2 2.11E-03 0.9122 392 FAM134B 1.62E-03 0.6350 432 GATAD1 2.11E-03 0.9122 393 GRLF1 1.68E-03 0.6587 432 FGFRL1 2.11E-03 0.9122 394 TRPV2 1.68E-03 0.6614 432 DENR 2.11E-03 0.9122 394 PDIA2 1.68E-03 0.6614 432 MS4A7 2.11E-03 0.9122 396 NME1-NME2 1.68E-03 0.6654 443 OASL 2.12E-03 0.9386 396 NME1 1.68E-03 0.6654 444 PRRT3 2.18E-03 0.9665 398 SETDB2 1.70E-03 0.6748 444 C12orf54 2.18E-03 0.9665 399 RTN4IP1 1.70E-03 0.6784 446 UFM1 2.21E-03 0.9852 399 FBLIM1 1.70E-03 0.6784 447 RACGAP1 2.22E-03 0.9914 399 ZSCAN2 1.70E-03 0.6784 448 CABC1 2.22E-03 0.9951 399 ROD1 1.70E-03 0.6784 449 HFE 2.25E-03 1.0115 399 GABRA1 1.70E-03 0.6784 450 PRUNE 2.27E-03 1.0193 399 MUTED 1.70E-03 0.6784 451 MCAM 2.29E-03 1.0316 405 CCDC90A 1.72E-03 0.6972 452 APLP2 2.32E-03 1.0498 406 OFD1 1.73E-03 0.7034 453 H2AFB2 2.32E-03 1.0522 406 EDEM3 1.73E-03 0.7034 453 H2AFB3 2.32E-03 1.0522 408 ATP6V0E1 1.74E-03 0.7080 455 ADAMTS2 2.34E-03 1.0646 409 NIN 1.77E-03 0.7229 456 RYBP 2.34E-03 1.0686 410 NR2F2 1.78E-03 0.7312 457 CACNB3 2.35E-03 1.0717 411 SMEK1 1.80E-03 0.7382 458 MBD4 2.36E-03 1.0810 412 NT5C1B 1.82E-03 0.7482 459 H2AFX 2.39E-03 1.0981 413 HTRA1 1.82E-03 0.7504 460 LMAN1 2.39E-03 1.1006 414 RECQL 1.82E-03 0.7539 460 ZNF446 2.39E-03 1.1006 415 ZNF286A 1.83E-03 0.7582 460 AK7 2.39E-03 1.1006 416 DKFZP564O0823 1.83E-03 0.7623 460 RTP4 2.39E-03 1.1006 417 SPG7 1.84E-03 0.7677 460 SP140L 2.39E-03 1.1006 417 IFI44 1.84E-03 0.7677 460 ACTR10 2.39E-03 1.1006 419 SLAMF7 1.85E-03 0.7731 460 RWDD2B 2.39E-03 1.1006 420 ANXA10 1.87E-03 0.7851 460 GRM7 2.39E-03 1.1006 421 ANK1 1.87E-03 0.7892 460 C7orf58 2.39E-03 1.1006 422 PSAT1 1.88E-03 0.7926 460 ZNF644 2.39E-03 1.1006 422 IRAK4 1.88E-03 0.7926 470 IL17B 2.42E-03 1.1359

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

UMSCC25 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 471 GPRC5B 2.47E-03 1.1637 518 DDX19A 3.27E-03 1.6952 472 NCKAP1L 2.51E-03 1.1824 519 FGR 3.34E-03 1.7315 473 MT1F 2.51E-03 1.1886 520 CD81 3.34E-03 1.7383 474 IGF1 2.54E-03 1.2026 521 NR1H2 3.40E-03 1.7738 475 PDCD1LG2 2.56E-03 1.2156 522 GBP3 3.46E-03 1.8061 475 RAC3 2.56E-03 1.2156 523 PTK2 3.49E-03 1.8231 477 DBC1 2.57E-03 1.2260 523 YWHAZ 3.49E-03 1.8231 478 FBXO9 2.62E-03 1.2546 523 GFRA1 3.49E-03 1.8231 479 PEX3 2.63E-03 1.2615 523 ATP2A1 3.49E-03 1.8231 480 C18orf54 2.67E-03 1.2814 527 MFN2 3.53E-03 1.8589 481 SELENBP1 2.72E-03 1.3086 527 C9orf5 3.53E-03 1.8589 481 RALA 2.72E-03 1.3086 527 CASC1 3.53E-03 1.8589 483 BAZ2A 2.73E-03 1.3187 527 JMJD2A 3.53E-03 1.8589 484 TIMM8B 2.74E-03 1.3260 531 SLC35E3 3.55E-03 1.8855 485 PPFIBP1 2.74E-03 1.3293 532 ESR1 3.55E-03 1.8908 486 ARHGEF9 2.75E-03 1.3378 533 PROK1 3.55E-03 1.8945 486 ACLY 2.75E-03 1.3378 534 ENTPD6 3.57E-03 1.9054 486 ZMYND10 2.75E-03 1.3378 535 REST 3.59E-03 1.9207 486 SIRT4 2.75E-03 1.3378 536 TNS1 3.67E-03 1.9681 486 EDF1 2.75E-03 1.3378 537 GMPR 3.70E-03 1.9887 486 UBQLN1 2.75E-03 1.3378 537 RARG 3.70E-03 1.9887 486 POLR2C 2.75E-03 1.3378 486 NDUFS8 2.75E-03 1.3378 486 ZNF771 2.75E-03 1.3378 495 CENPT 2.76E-03 1.3649 496 ATP2B3 2.77E-03 1.3742 496 SMN2 2.77E-03 1.3742 496 TBX1 2.77E-03 1.3742 496 SMN1 2.77E-03 1.3742 500 RGS12 2.79E-03 1.3946 501 ZNF268 2.90E-03 1.4522 502 CUEDC1 2.92E-03 1.4665 503 CWC22 3.00E-03 1.5098 504 KCNJ13 3.10E-03 1.5640 505 PAPD4 3.12E-03 1.5742 506 CLIP2 3.14E-03 1.5887 507 SLC5A4 3.15E-03 1.5960 508 TCIRG1 3.22E-03 1.6368 509 ZNF623 3.25E-03 1.6545 509 DNHD1 3.25E-03 1.6545 509 BPI 3.25E-03 1.6545 509 CATSPER2 3.25E-03 1.6545 509 CARKD 3.25E-03 1.6545 509 SYNCRIP 3.25E-03 1.6545 509 SERPINH1 3.25E-03 1.6545 516 SPCS1 3.26E-03 1.6825 517 TMEM40 3.27E-03 1.6900

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

584-A2 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 1 NAP1L1 8.86E-06 8.86E-06 48 HNRPDL 8.14E-04 3.91E-02 2 ARHGDIB 1.14E-05 2.28E-05 49 WAC 8.14E-04 3.99E-02 3 TOX4 1.39E-05 4.18E-05 50 SBNO1 8.16E-04 4.08E-02 4 PPIL1 3.84E-05 1.54E-04 51 PPP5C 8.17E-04 4.17E-02 5 MPPED2 3.89E-05 1.95E-04 51 MCM6 8.17E-04 4.17E-02 6 TTC6 7.91E-05 4.75E-04 53 IKZF1 8.18E-04 4.33E-02 7 LAMB4 8.05E-05 5.63E-04 54 LSR 8.20E-04 4.43E-02 8 NRP2 8.12E-05 6.49E-04 55 MAG 8.21E-04 4.51E-02 9 HFE 1.15E-04 1.03E-03 56 KLF3 8.21E-04 4.60E-02 10 HOXA3 1.17E-04 1.17E-03 57 TIMP1 8.22E-04 4.69E-02 11 WDR1 1.17E-04 1.29E-03 58 CHD1L 8.23E-04 4.78E-02 12 ADAMTS8 1.30E-04 1.56E-03 59 NKIRAS2 8.36E-04 4.93E-02 13 SEMA6D 2.01E-04 2.61E-03 60 C14orf119 8.36E-04 5.01E-02 14 SLCO4C1 2.23E-04 3.12E-03 61 SRP72 8.39E-04 5.12E-02 15 SCOC 2.41E-04 3.62E-03 62 ACO1 8.43E-04 5.23E-02 16 ZNF460 2.61E-04 4.17E-03 63 TP63 8.47E-04 5.33E-02 17 EHF 2.65E-04 4.50E-03 64 C21orf66 8.49E-04 5.43E-02 18 CADM3 2.90E-04 5.22E-03 65 FARP2 8.49E-04 5.52E-02 19 DNAJC10 4.34E-04 8.25E-03 66 APOBEC3F 8.57E-04 5.65E-02 20 TRIM5 4.51E-04 9.02E-03 66 ZNF763 8.57E-04 5.65E-02 21 HLF 5.02E-04 1.05E-02 66 APOBEC3G 8.57E-04 5.65E-02 22 VPS13A 5.22E-04 1.15E-02 69 HNRNPU 8.77E-04 6.05E-02 23 KLK13 6.70E-04 1.54E-02 69 SOS2 8.77E-04 6.05E-02 24 TAF9B 7.88E-04 1.89E-02 71 NUDT21 8.80E-04 6.25E-02 25 ACVR1C 8.02E-04 2.00E-02 72 GNRH1 8.81E-04 6.34E-02 25 IGSF1 8.02E-04 2.00E-02 73 FER 9.00E-04 6.57E-02 27 POLK 8.02E-04 2.17E-02 74 SCO1 9.00E-04 6.66E-02 27 CREB5 8.02E-04 2.17E-02 75 ATP11B 9.01E-04 6.76E-02 29 NADK 8.02E-04 2.33E-02 76 ATRNL1 9.06E-04 6.88E-02 30 MAGT1 8.02E-04 2.41E-02 76 EID1 9.06E-04 6.88E-02 30 ABHD6 8.02E-04 2.41E-02 78 SLC2A12 9.10E-04 7.10E-02 30 DNAJC13 8.02E-04 2.41E-02 79 ACTR10 9.15E-04 7.23E-02 33 SOST 8.02E-04 2.65E-02 79 DNAL1 9.15E-04 7.23E-02 34 SENP2 8.02E-04 2.73E-02 81 NT5C1B 9.20E-04 7.45E-02 35 VGLL4 8.02E-04 2.81E-02 82 P2RY12 9.27E-04 7.60E-02 36 LRP11 8.02E-04 2.89E-02 83 SSSCA1 9.28E-04 7.70E-02 37 KCNIP4 8.03E-04 2.97E-02 84 YBX1 9.34E-04 7.85E-02 38 FAM172A 8.03E-04 3.05E-02 85 FAM122B 9.37E-04 7.96E-02 39 NTN4 8.03E-04 3.13E-02 86 DPH5 9.49E-04 8.16E-02 40 SMG7 8.03E-04 3.21E-02 87 RTP4 9.50E-04 8.26E-02 41 KIAA1530 8.04E-04 3.30E-02 88 INSRR 9.69E-04 8.53E-02 42 IGSF2 8.04E-04 3.38E-02 89 THUMPD1 9.73E-04 8.66E-02 43 FILIP1L 8.05E-04 3.46E-02 90 ZSCAN5A 9.85E-04 8.86E-02 44 UIMC1 8.07E-04 3.55E-02 91 CACNG6 9.91E-04 9.02E-02 45 TBC1D9B 8.07E-04 3.63E-02 92 HCFC2 9.93E-04 9.13E-02 45 CTSS 8.07E-04 3.63E-02 93 TNS1 1.01E-03 9.37E-02 47 OR10R2 8.11E-04 3.81E-02 94 ST8SIA2 1.02E-03 9.60E-02

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

584-A2 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 95 PPPDE1 1.02E-03 9.72E-02 142 DST 1.40E-03 0.1993 96 LDLRAD2 1.03E-03 9.85E-02 143 MOBKL1B 1.41E-03 0.2011 96 HSPG2 1.03E-03 9.85E-02 144 MAP3K6 1.41E-03 0.2026 98 IL1RAPL1 1.03E-03 0.1012 144 KCTD14 1.41E-03 0.2026 99 SRPK2 1.05E-03 0.1039 144 NCR1 1.41E-03 0.2026 100 ZMYM4 1.07E-03 0.1072 147 ABCB1 1.41E-03 0.2075 101 RNF122 1.08E-03 0.1091 147 ABCB4 1.41E-03 0.2075 101 ROBO3 1.08E-03 0.1091 149 TMCC1 1.44E-03 0.2142 101 MPP7 1.08E-03 0.1091 150 MAPK1IP1L 1.44E-03 0.2161 104 NF2 1.08E-03 0.1126 151 MYO1B 1.47E-03 0.2213 104 GNA12 1.08E-03 0.1126 152 FLRT1 1.47E-03 0.2240 106 EIF5B 1.08E-03 0.1150 153 HBG2 1.49E-03 0.2284 107 FBXO28 1.09E-03 0.1166 153 HBG1 1.49E-03 0.2284 108 SDCCAG10 1.09E-03 0.1179 155 TKTL2 1.51E-03 0.2343 109 NRXN3 1.09E-03 0.1192 156 ILF3 1.51E-03 0.2360 110 USP21 1.10E-03 0.1206 157 CAPRIN2 1.52E-03 0.2388 110 ZNF781 1.10E-03 0.1206 158 ASMTL 1.56E-03 0.2458 110 PKD2L1 1.10E-03 0.1206 159 TBX10 1.57E-03 0.2498 113 ETV1 1.10E-03 0.1242 159 PPBP 1.57E-03 0.2498 114 EXOSC2 1.11E-03 0.1263 159 PMP2 1.57E-03 0.2498 115 ISCA1 1.11E-03 0.1278 159 ARF3 1.57E-03 0.2498 116 DAB2 1.12E-03 0.1296 163 FAM149A 1.59E-03 0.2589 117 TMEM155 1.13E-03 0.1327 164 CALCA 1.61E-03 0.2637 118 DYRK4 1.14E-03 0.1339 165 PARK7 1.62E-03 0.2680 119 GRK4 1.15E-03 0.1365 166 KIAA0284 1.66E-03 0.2750 120 TRIB2 1.16E-03 0.1390 167 DPPA4 1.68E-03 0.2798 121 PANX2 1.18E-03 0.1425 167 PFTK2 1.68E-03 0.2798 122 NUCKS1 1.18E-03 0.1439 169 MED24 1.74E-03 0.2948 123 KDELC1 1.18E-03 0.1451 170 FLT3LG 1.76E-03 0.2989 124 PFKFB2 1.19E-03 0.1475 171 PAX8 1.77E-03 0.3020 125 SURF1 1.21E-03 0.1513 172 PTGDR 1.80E-03 0.3087 126 AQP1 1.21E-03 0.1530 172 CRMP1 1.80E-03 0.3087 127 PALM2-AKAP2 1.25E-03 0.1582 174 TMPRSS2 1.85E-03 0.3226 127 AKAP2 1.25E-03 0.1582 175 HS6ST1 1.93E-03 0.3371 129 RAD1 1.25E-03 0.1609 175 PRB4 1.93E-03 0.3371 130 C15orf54 1.26E-03 0.1643 175 NCOA7 1.93E-03 0.3371 130 RPL14 1.26E-03 0.1643 175 FYCO1 1.93E-03 0.3371 132 C10orf92 1.30E-03 0.1713 175 PRB3 1.93E-03 0.3371 133 PLEKHG6 1.32E-03 0.1753 180 CD37 1.93E-03 0.3469 134 BMPR2 1.35E-03 0.1803 181 METTL11A 1.95E-03 0.3522 135 LECT2 1.35E-03 0.1821 182 COQ3 1.98E-03 0.3596 136 LIMA1 1.37E-03 0.1859 183 DNAJC4 1.99E-03 0.3635 137 RBBP6 1.38E-03 0.1884 184 ZXDC 2.08E-03 0.3828 137 IQGAP3 1.38E-03 0.1884 185 C6orf106 2.12E-03 0.3916 137 PPP2R2D 1.38E-03 0.1884 186 UQCC 2.12E-03 0.3942 137 CHAC2 1.38E-03 0.1884 187 ATG12 2.14E-03 0.3998 141 VCAN 1.40E-03 0.1973 188 C14orf126 2.14E-03 0.4025

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

584-A2 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 189 TMED2 2.17E-03 0.4097 236 STARD10 3.29E-03 0.7767 190 KIAA0467 2.19E-03 0.4163 237 NEIL3 3.32E-03 0.7863 191 TMC5 2.23E-03 0.4254 238 TMLHE 3.33E-03 0.7914 192 RPA1 2.24E-03 0.4307 239 AGFG1 3.46E-03 0.8263 193 IKZF2 2.27E-03 0.4372 240 FMR1 3.48E-03 0.8340 194 MUTYH 2.29E-03 0.4438 241 TMEM49 3.53E-03 0.8495 194 ACSL6 2.29E-03 0.4438 241 USP38 3.53E-03 0.8495 196 LUZP2 2.33E-03 0.4566 241 KCTD20 3.53E-03 0.8495 196 HPS6 2.33E-03 0.4566 241 PLAC8 3.53E-03 0.8495 198 PSRC1 2.34E-03 0.4638 245 TAF11 3.57E-03 0.8744 199 ZNF562 2.38E-03 0.4745 245 UHRF1BP1 3.57E-03 0.8744 200 TMEM9 2.40E-03 0.4797 247 TGFB3 3.59E-03 0.8875 200 LIFR 2.40E-03 0.4797 248 ALB 3.61E-03 0.8962 202 ACE2 2.40E-03 0.4845 249 PSMD6 3.76E-03 0.9364 203 YPEL1 2.50E-03 0.5069 249 RPRD1A 3.76E-03 0.9364 203 TLR10 2.50E-03 0.5069 249 MED16 3.76E-03 0.9364 203 CDC37 2.50E-03 0.5069 252 TOMM70A 3.84E-03 0.9666 203 ARPC4 2.50E-03 0.5069 253 DSN1 3.85E-03 0.9737 203 EIF4G1 2.50E-03 0.5069 253 SP100 3.85E-03 0.9737 203 ADM2 2.50E-03 0.5069 253 EFCAB2 3.85E-03 0.9737 203 APOBEC3C 2.50E-03 0.5069 253 RAD21 3.85E-03 0.9737 210 C12orf65 2.52E-03 0.5297 257 RNGTT 3.88E-03 0.9969 210 EPB41L2 2.52E-03 0.5297 258 SPTBN1 4.03E-03 1.0408 212 C19orf6 2.53E-03 0.5373 259 PI4K2A 4.04E-03 1.0459 212 DEFA5 2.53E-03 0.5373 260 KLRG1 4.08E-03 1.0598 214 NDRG3 2.55E-03 0.5465 261 KIAA1217 4.13E-03 1.0789 215 METTL2A 2.59E-03 0.5575 261 SH3BP4 4.13E-03 1.0789 216 SOX2 2.60E-03 0.5624 263 VANGL1 4.21E-03 1.1078 217 RGS3 2.69E-03 0.5839 264 FGF1 4.22E-03 1.1153 218 MED28 2.72E-03 0.5933 265 ANKRD20A3 4.23E-03 1.1221 219 SLC26A1 2.77E-03 0.6077 265 RHEBL1 4.23E-03 1.1221 220 AFTPH 2.79E-03 0.6142 265 ANKRD20A1 4.23E-03 1.1221 221 PPIL2 2.81E-03 0.6213 265 ANKRD20A2 4.23E-03 1.1221 222 RBM6 2.83E-03 0.6273 269 EDA 4.26E-03 1.1448 222 C3orf37 2.83E-03 0.6273 270 TCEAL2 4.31E-03 1.1642 224 TFAM 2.83E-03 0.6335 270 C16orf58 4.31E-03 1.1642 225 CNOT2 2.83E-03 0.6373 270 ETNK1 4.31E-03 1.1642 226 SDCCAG1 2.86E-03 0.6462 273 PRKCE 4.36E-03 1.1893 226 ANKRD12 2.86E-03 0.6462 274 TMEM27 4.36E-03 1.1960 226 TNFSF13B 2.86E-03 0.6462 275 RYR1 4.60E-03 1.2643 229 SLC16A3 2.95E-03 0.6763 276 PIK3C3 4.64E-03 1.2807 230 NRXN1 3.10E-03 0.7127 276 SAC3D1 4.64E-03 1.2807 231 FSCN3 3.16E-03 0.7302 276 RRP15 4.64E-03 1.2807 232 SFRP1 3.17E-03 0.7346 279 ADH4 4.72E-03 1.3161 233 SPDEF 3.21E-03 0.7486 280 OLFML2A 4.72E-03 1.3215 234 ETV4 3.22E-03 0.7529 281 NFIB 4.76E-03 1.3367 235 ABI3 3.28E-03 0.7710 282 RPS11 4.76E-03 1.3415

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

584-A2 Rank Gene Symbol p-value [p(wZ)] E value 282 FUT4 4.76E-03 1.3415 282 POP4 4.76E-03 1.3415 282 RPL9 4.76E-03 1.3415 282 KIF18A 4.76E-03 1.3415 287 KBTBD4 4.86E-03 1.3952 288 SEMA4G 4.91E-03 1.4149 289 7-Mar 4.92E-03 1.4214 289 SF3A2 4.92E-03 1.4214 291 UBE2J1 4.97E-03 1.4469 292 PRICKLE4 5.03E-03 1.4681 293 CHD7 5.09E-03 1.4922 294 MCM10 5.10E-03 1.4985 295 TRAM1 5.17E-03 1.5254 296 CA11 5.26E-03 1.5571 296 MFF 5.26E-03 1.5571 296 ALDH4A1 5.26E-03 1.5571 296 TNKS 5.26E-03 1.5571 296 PIGO 5.26E-03 1.5571 296 TFPI2 5.26E-03 1.5571 296 RAB7A 5.26E-03 1.5571 296 ADAM12 5.26E-03 1.5571 296 MAPK8IP3 5.26E-03 1.5571 296 ZFP82 5.26E-03 1.5571 296 PLEK2 5.26E-03 1.5571 307 CTSZ 5.27E-03 1.6174 308 MTMR1 5.28E-03 1.6257 309 PER2 5.31E-03 1.6408 310 NF1 5.32E-03 1.6490 311 PDZD11 5.52E-03 1.7174 312 TCEA1 5.53E-03 1.7264 313 C16orf35 5.55E-03 1.7370 314 SLC6A12 5.61E-03 1.7613 315 RAP2B 5.75E-03 1.8107 316 PSMF1 5.76E-03 1.8197 317 PROL1 5.79E-03 1.8340 318 KBTBD10 5.84E-03 1.8579 319 STRN 5.90E-03 1.8814 320 TRIM38 5.93E-03 1.8974 321 KIAA0020 5.94E-03 1.9080 322 NPAL2 6.01E-03 1.9348 323 UBAP2L 6.08E-03 1.9650 324 OAS2 6.14E-03 1.9878 324 EPPK1 6.14E-03 1.9878 324 COL14A1 6.14E-03 1.9878

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

CCL30 Rank Gene Symbol p-value [p(wZ)] E value Rank Gene Symbol p-value [p(wZ)] E value 1 BCLAF1 9.25E-04 9.25E-04 48 VNN2 6.65E-03 0.3190 2 FN1 1.14E-03 2.27E-03 49 MPPED2 6.75E-03 0.3310 3 FGF2 1.23E-03 3.70E-03 50 YWHAZ 6.87E-03 0.3437 4 LRPPRC 2.05E-03 8.20E-03 51 CBX4 7.01E-03 0.3575 5 GAPDH 2.09E-03 1.05E-02 52 ESR2 7.08E-03 0.3681 6 ATP6AP2 2.18E-03 1.31E-02 53 PTP4A1 7.21E-03 0.3822 7 DBF4 2.22E-03 1.55E-02 54 TMED2 7.29E-03 0.3934 8 GYG1 2.29E-03 1.83E-02 55 RPL22 7.30E-03 0.4013 9 AK2 2.42E-03 2.18E-02 56 CLUL1 7.40E-03 0.4143 10 IL5RA 2.44E-03 2.44E-02 57 CSNK2A1 7.41E-03 0.4226 11 OTUB2 2.57E-03 2.82E-02 58 MAX 7.72E-03 0.4478 12 CSF2 2.57E-03 3.09E-02 59 SGK1 7.81E-03 0.4608 13 CHD4 2.59E-03 3.37E-02 60 RAP2B 8.03E-03 0.4821 14 PAX8 2.61E-03 3.66E-02 61 HMG20B 8.19E-03 0.4998 15 LPIN1 3.01E-03 4.51E-02 62 PPP1CB 8.23E-03 0.5100 16 DKK3 3.04E-03 4.86E-02 63 PDE8A 8.26E-03 0.5207 17 FABP4 3.06E-03 5.21E-02 64 RUNX1 8.35E-03 0.5343 18 PEX5 3.23E-03 5.81E-02 65 KCNK7 8.38E-03 0.5445 19 EEF1E1 3.30E-03 6.27E-02 66 RRH 8.47E-03 0.5587 20 TRIM44 3.60E-03 7.21E-02 67 FXR1 8.62E-03 0.5773 21 CXCR6 3.73E-03 7.84E-02 68 KLHL20 8.85E-03 0.6019 22 LSM14A 3.74E-03 8.24E-02 69 CASP1 8.85E-03 0.6110 23 DNAJC6 3.86E-03 8.89E-02 70 SDF4 9.23E-03 0.6463 24 RPS7 4.11E-03 9.85E-02 71 IPO8 9.25E-03 0.6564 25 MAPK9 4.16E-03 0.1039 72 EPB41L3 9.76E-03 0.7024 26 DYNC1LI2 4.17E-03 0.1084 73 CCDC6 9.90E-03 0.7226 27 SERPINB3 4.32E-03 0.1165 74 TUBA1A 9.94E-03 0.7358 28 C1QBP 4.38E-03 0.1227 75 DNAJC24 9.96E-03 0.7471 29 PDCD10 4.50E-03 0.1306 76 ZMYND11 1.01E-02 0.7654 30 VIPR2 4.73E-03 0.1418 77 HYAL1 1.02E-02 0.7861 31 CSRP2 5.06E-03 0.1568 78 MALT1 1.05E-02 0.8172 32 EZR 5.06E-03 0.1620 79 SIKE 1.06E-02 0.8369 33 RBM15B 5.10E-03 0.1682 80 TRIM8 1.10E-02 0.8769 34 KCNA3 5.16E-03 0.1754 81 ZNF208 1.10E-02 0.8895 35 THAP4 5.16E-03 0.1807 82 FAM83E 1.12E-02 0.9157 36 SLC35B1 5.21E-03 0.1875 83 CCR3 1.13E-02 0.9343 37 TPT1 5.27E-03 0.1949 84 CSTF2T 1.14E-02 0.9572 38 PGM1 5.37E-03 0.2039 85 PSG9 1.15E-02 0.9792 39 SFRS10 5.56E-03 0.2168 86 NUP50 1.16E-02 0.9994 40 TLE4 5.76E-03 0.2305 87 CYP2C18 1.17E-02 1.0149 41 CLIC2 5.89E-03 0.2414 88 CACYBP 1.19E-02 1.0451 42 NRXN1 5.94E-03 0.2494 89 C1orf116 1.19E-02 1.0597 43 KPNA4 6.01E-03 0.2584 90 SH2B3 1.19E-02 1.0726 44 ATP6V0E1 6.24E-03 0.2746 91 DBT 1.21E-02 1.0994 45 HIST1H2AE 6.26E-03 0.2817 92 UBE2L3 1.23E-02 1.1344 46 PCDHGA1 6.44E-03 0.2963 93 ZFP106 1.24E-02 1.1577 47 TMPRSS5 6.44E-03 0.3028 94 GBF1 1.25E-02 1.1738

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

CCL30 Rank Gene Symbol p-value [p(wZ)] E value 95 TRIM2 1.27E-02 1.2025 96 FGF3 1.29E-02 1.2339 97 FGFR3 1.29E-02 1.2541 98 GTF3C2 1.30E-02 1.2729 99 GPBP1L1 1.31E-02 1.3003 100 PCDH12 1.34E-02 1.3366 101 MOBKL1B 1.34E-02 1.3530 102 PTEN 1.37E-02 1.4023 103 GCN1L1 1.38E-02 1.4205 104 EIF5B 1.40E-02 1.4509 105 TMSL3 1.40E-02 1.4693 105 TMSB4X 1.40E-02 1.4693 107 GLUL 1.41E-02 1.5096 108 ZFX 1.42E-02 1.5361 109 DNAJC13 1.42E-02 1.5506 110 ZNF217 1.43E-02 1.5778 111 PAFAH2 1.45E-02 1.6143 112 KIF5B 1.46E-02 1.6339 113 XPNPEP1 1.46E-02 1.6513 114 TMEM30B 1.47E-02 1.6760 115 IRF9 1.48E-02 1.6977 116 KCNJ1 1.50E-02 1.7385 117 GCDH 1.50E-02 1.7539 118 PGM3 1.55E-02 1.8277 119 USP9X 1.56E-02 1.8522 120 PI4K2A 1.60E-02 1.9189 121 AKAP4 1.63E-02 1.9689 122 PHTF1 1.63E-02 1.9873

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL FIGURE 1: Dose response of cell lines to RTK inhibition. The indicated cell lines were treated with AZ8010 (A, B), AZ8931 (A, C) or PF1066 (A, D) at a range of doses and proliferation was measured by CyQuant assay. The IC50 was calculated for each inhibitor and cell line.

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL FIGURE 2: Unsupervised hierarchical clustering of replicates and treatment groups within cell lines treated with AZ8010.

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL FIGURE 3: Overlap of SLATs between cell lines. The figure displays the number of SLATs identified in common between UMSCC25, 584-A2 and

CCL30 cells and their gene names. Note that no SLAT was identified in all three cell lines.

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL FIGURE 4: Inhibition of RTK phosphorylation by TKIs.

UMSCC25 cells were treated with AZ8010, AZ8931, PF1066 and the combinations at the doses indicated (µM) for 6 hrs. Whole cell lysates were collected and immunoblotted for A) phospho-ERBB2, B) phospho-EGFR or C) phospho-MET. Filters were stripped and reprobed for levels of total protein and the α-subunit of Na/K-ATPase as a loading control.

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL FIGURE 5: Combined inhibition of FGFR and ERBB family members by the TKI, lapatinib, leads to impairment of cell line growth. The cell lines indicated were treated in triplicate with DMSO, 0.1 µM lapatinib, 0.5 µM labatinib alone or in combination with the indicated doses of AZ8010. The effects of the treatment on cell growth were measured by clonogenic assay (UMSCC25, 584-A2) or anchorage independent growth assay (CCL30). Statistical analysis using a two-tailed Student’s t- test was used to show significant differences in growth where ns indicates not significant, * indicates p<0.05, ** indicates p<0.005.

Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111

SUPPLEMENTAL FIGURE 6: Knockdown of ERBB2 or MET inhibits growth of two

HNSCC cell lines and improves response to AZ8010 treatment. A) Pooled UMSCC1 or UMSCC8 cells expressing either a negative control shRNA targeting GFP or two independent shRNAs targeting ERBB2 were measured for ERBB2 protein expression by immunoblot. The protein level of the α-subunit of Na/K-ATPase was measured by immunoblot as a loading control. B) Pooled UMSCC1 cells expressing the control shRNA or one of two shRNAs targeting ERBB2 were used in triplicate in a clonogenic growth assay in the presence or absence of 1.0 µM AZ8010. C) Pooled UMSCC8 cells expressing the control shRNA or one of two shRNAs targeting ERBB2 were used in triplicate in a clonogenic growth assay in the presence or absence of 1.0 µM AZ8010.

D) Pooled UMSCC1 or UMSCC8 cells expressing either a negative control shRNA targeting GFP or two independent shRNAs targeting MET were measured for MET protein expression by immunoblot. The protein level of the α-subunit of Na/K-ATPase was measured by immunoblot as a loading control. E) Pooled UMSCC1 cells expressing the control shRNA or one of two shRNAs targeting MET were used in triplicate in a clonogenic growth assay in the presence or absence of 1.0 µM AZ8010.

After two weeks, colonies were stained and quantified as described in the Materials and

Methods. F) Pooled UMSCC8 cells expressing the control shRNA or one of two shRNAs targeting MET were used in triplicate in a clonogenic growth assay in the presence or absence of 1.0 µM AZ8010. After two weeks, colonies were stained and quantified as described in the Materials and Methods. Data are the means and SEM of three independent experiments. Statistical analysis using a two-tailed Student’s t-test was used to show significant differences in growth where ns indicates not significant, * indicates p<0.05, ** indicates p<0.005, *** indicates p<0.0005 and **** indicates p<0.0001. Singleton, Kim, Hinz, Marek, Casas-Selves, Hatheway, Tan, DeGregori and Heasley. A Receptor Tyrosine Kinase Network Comprised of FGFRs, EGFR, ERBB2 and MET Drives Growth and Survival of Head and Neck Squamous Carcinoma Cell Lines. MOLPHARM #084111