Methods section for supplemental data

A.) IC50 values were calculated using the PRISM software and non-variable slope analysis. Fold increase in sensitivity was calculated based on changes IC50 values in the DasRes + 2μM dasatinib in relation to the DasRes cell line IC50 values.

IC50 values were calculated using the PRISM software and non-variable slope. Fold change relative to control IC50 values are depicted as relative resistance.

Supplementary Fig. S4. Over expressed eIF4E induces luminal subtype tumors.

BCPAP, SW1736, C643, Cal62 control (C), dasatinib-resistant (DR), and dasatinib-resistant cells maintained in 2μM dasatinib in response to indicated concentrations of selumetinib (μM) for 72 hours. Growth was measured by Sulforhodamine B.

Supplementary Fig. S2. HA-eIF4E-induced pre-malignant lesions persist in post-weaning mammary glands.

Short Tandem Repeat profiling was performed on the Control and DasRes BCPAP, SW1736, C643, and Cal62 cell lines.

A.) Densitometry quantification of ERK 1/2 phosphorylation. Fold changes were calculated based on the control cell lines being normalized to 1. Signals were normalized to α-tubulin. Quantification represents the mean of at least 3 separate experiments for each cell line +/- SEM B.) Western blot analysis after c-Src knockdown using two independent shRNAs (1 & 2) in the BCPAP cell line. Analysis was performed in comparison to a non-transfected (WT) and scrambled control (-). Lysates were probed for c-Src, ppERK1/2, ERK1/2, and α-tubulin. C.) Densitometry quantification of BRAF co-immunoprecipitations after c-Raf pulldown in the control and dasatinib resistant cell lines. Fold Changes are relative to DMSO controls, and densitometry was normalized to CRAF pulldown. D.) Quantification of ERK 1/2 phosphorylation based on western blot analysis for ppERK 1/2 in whole cell lysates following treatment of cells with 100nM dasatinib for 0,2,4,8,24,48, or 46 hours plus a 2 hour challenge (48+2). Fold changes were calculated based on the DMSO (0hr) treatment being normalized to 1. Signals were normalized to α-tubulin or β-actin. Quantification represents the mean of at least 3 separate experiments for each cell line +/- SEM. E.) Cells transfected with the c-Fos-SRE-Luc and pRL-null vectors were assayed for luciferase activity using the Dual Luciferase Reporter Assay System (Promega) after being treated with 100nM dasatinib for 0,2,4,8,24,48, or 46 hours plus a 2 hour challenge (48+2). Signals were normalized to the DMSO (0hr) treatment. Quantification represents the mean of 3 separate experiments +/- SEM.

The Kegg pathway Melanoma was enriched (p-value: 0.1465), enriched in the DasRes cell lines. Portrayed are the genes in the melanoma pathway. Highlighted in dark grey are genes that are enriched in the dasatinib-resistant cell lines. Genes enriched include: PIK3R5, E2F3, IGF1R, PIK3CG, PDGFB, FGF7, FGF21, FGF22, BRAF, E2F1, FGF2, FGF14, FGF9, MAPK1, EGFR, PDGFC, FGF1, CCND1, MAP2K1.

PDF file - 66K, IHC control images.

Supplementary Fig. S3. Double K8/Ki-67staining of wild type (WT) and transgenic (WAP-4E) mammary glands (gestation 6).

PDF file - 116K, Supplemental Figure 4. Activation of FGFR (1) leads to increased gene expression and secretion of IL-6 family members (2). These bind to the gp130 receptor (3) which activates STAT3 (4). Phosphorylated STAT3 regulates HAS2 expression (5). HAS2 stimulates production of hyaluronan (6), which is then secreted and contributes to induced proliferation, migration and chemoresistance.

PDF file - 45K, Supplemental Figure 3. HC-11 or HC-11/R1 cells were injected into the mammary fat pads of Balb/c mice. Mice were palpated to determine the % tumor free.

Supplementary Fig. S1. Generation of WAP-4E transgenic mice.

Tumor weights were measured upon euthanasia for both the Cal62 DasRes and Parental cell lines treated for 29 days with 1mg/kg QD trametinib. Data as means +/- SEM (n=8; student t-test; ***, P < 0.0005)

A.) Cell lines C643, Cal62, BCPAP, and SW1736, were treated with increasing doses of dasatinib ranging from (0.019 μM to 1.25 μM) for 72 hours, in combination with increasing doses of selumetinib (0.025 μM to 1 μM). Cell growth was measured using the Sulforhodamine B assay. Synergy was measured by determining the combination index using the Calcusyn software. Combinations that elicited a synergistic response are depicted by their corresponding shade of grey. (0.3-0.7,Synergism; 0.1-0.3, Strong Synergism; <0.1, Very Strong Synergism). B.) Cell lines BCPAP and Cal62 were treated with increasing doses of dasatinib ranging from (0.019 μM to 1.25 μM) for 72 hours, in combination with increasing doses of the ERK1/2 inhibitor SCH772984 (0.005 μM to 0.5 μM). Cell growth was measured using the Sulforhodamine B assay. Synergy was measured by determining the combination index using the Calcusyn software. Combinations that elicited a synergistic response are depicted by their corresponding shade of grey. (0.3-0.7, Synergism; 0.1-0.3, Strong Synergism; <0.1, Very Strong Synergism).

Supplementary Fig. S5. HA-eIF4E rescues HMEC/hTERT cells from Ras-induced senescence and activates ATR signaling.

A.) Clonogenic growth was detected by crystal violet staining in the BCPAP and Cal62 cell lines following 6 days of treatment with indicated inhibitors and 6 days of treatment release. B.) Colony area signal intensity was measured using Odyssey CLx imager (Li-Cor), and presented as percent fold change relative to the DMSO treated wells. Data as means +/- SEM (n=3; student t-test; *, P < 0.05, **, P < 0.005). C.) C. Comparison of the Cal62 Parental tumor volumes after 42 days of treatment with vehicle, 12.5mg/kg dasatinib BD, 0.5mg/kg trametinib QD, or the combination. Data as means +/- SEM (n=8; student t-test; **, P < 0.005)

PDF file - 38K, Proportional odds logistic regression results evaluating the association between pSTAT3 and pFRS2.

S1: Neither mitochondrial load nor membrane potential differ between MNX and wild-type strains. Mouse embryonic fibroblasts were isolated from each MNX and wild-type strain and stained with MitoTracker Green FM to identify overall mitochondrial load and with MitoTracker Red CMXRos to identify membrane potential. Probe fluorescence in each cell was analyzed by flow cytometry. A: Relative fluorescence for each probe showing mitochondrial load and raw membrane potential. B: MitoTracker Red is normalized to MitoTracker Green to show membrane potential per mitochondrion. Error bars represent standard error of the mean. S2: OCR and ECAR profiles vary between C57BL/6J and C3H/HeN MNX and wild-type strains. Mouse embryonic fibroblast from CC, CH, HH, and HC mouse strains were passaged once then analyzed for ECAR (Extracellular acidification rate) under serum starved conditions with no glucose to determine non-glycolytic acidification, after addition of glucose (glycolysis), after addition of oligomycin to determine glycolytic capacity (flux), and after addition of 2-DG (glycolytic reserve). * denotes p<0.05 and error bars represent standard error of the mean. CH and HH strains differ significantly in non-glycolytic acidification while CC and HC differ significantly in glycolytic reserve. S3: Mitochondrial flux of MNX and wild-type mice is variable. MNX and wild-type mouse embryonic fibroblasts from each strain were plated and the Seahorse analyzer was used to measure oxygen consumption and extracellular acidification prior to compound addition (Basal respiration), after oligomycin addition (ATP-Linked respiration), and after addition of the uncoupler FCCP (Maximal respiration). * denotes p<0.05 and error bars represent standard error of the mean. A: Basal respiration differs significantly between * FC and BB, and ** FF and BB. Maximal respiration differs significantly between * FC and BB, ** FC and FB, as well as ***FF and BB. B: Basal, ATP-linked, and Maximal respiration does not differ significantly between CC, CH, HH, and HC strains S4: Mitochondrial DNA content does not differ between C57BL/6J and C3H/HeN MNX and wild-type strains. Lungs were isolated from four-week old CC (n=5), CH (n=7), HH (n=5), and HC (n=8) mice. Total DNA was isolated from each lung taking care to preserve mtDNA. qPCR utilizing Taqman primers for mitochondrial encoded mtCO2 and ND1 as well as nuclear encoded 18s was performed and fold change for mitochondrial genes was calculated relative to nuclear 18s. Error bars represent standard error of the mean. S5. Analysis of Metastatic size and area. Images (10x) from Figure 5B were analyzed using Photoshop and Image J for relative size and number of metastases. Analysis of formalin-fixed paraffin-embedded experimental lung metastases derived from K1735-M2 Melanoma cell intravenous injections. DMSO vehicle (-) or MitoTEMPO (+) was IP injected into 4-week-old wild-type or MNX mice 24 hours and again 1 hour prior to intravenous injection of cancer cells. Mice were euthanized two weeks post cell injection and lungs were harvested. Data as medians (n = 9-16; One-way ANOVA and multiple comparisons tests). S6: MtDNA affects expression of select nuclear genes. RNA was isolated from lungs of 4-week old CC, CH, HH, and HC mice. qPCR was performed with Taqman primers to determine expression of nuclear genes normalized to the nuclear 18s housekeeping gene. Error bars represent standard error of the mean. Blue bars represent gene expression levels of HH and CH mice relative to CC while green bars represent gene expression levels of CC and HC mice relative to HH mice. S7: Mitochondrial superoxide scavenging selectively alters nuclear gene expression. RNA was isolated from lungs of 4-week old CC, CH, HH and HC mice treated with DMSO vehicle (-) or MitoTEMPO (+) 24 hours and 1 hour prior to euthanasia. qPCR was performed with Taqman primers to determine gene expression of A: miR199a B: miR125b C: dnmt1 D: ago2 E: sod2 F: fam120a G: rab6b H: sgtb and I: scai. All expression was normalized to 18s. S8: Mitochondria stromal haplotype does not alter cancer cell extravasation: Male and female wild-type (HH) and MNX (HC) mice were injected into the lateral tail vein with CFSE labeled K1735-M2 melanoma cells. Lungs were harvested 24 hours post-injection. A. The right lung lobes were analyzed by flow cytometry for CFSE positive cancer cells. Seven hundred thousand events were collected using the Acurri C6 cytometer and dead cells were identified using propidium iodide and excluded from further analysis. B. The left lung lobe was fixed in 10% formalin. Sections (3 x 20 µm) approximately 100 µm apart were analyzed from each mouse for detection of CFSE+ cancer cells. The number obtained from all three sections were then averaged and graphed in comparison to the uninjected controls