MAX Mutations in Endometrial Cancer: Clinicopathologic Associations and Recurrent MAX p.His28Arg Functional Characterization

MAX Mutations in Endometrial Cancer: Clinicopathologic Associations and Recurrent MAX p.His28Arg... Abstract Background Genomic studies have revealed that multiple genes are mutated at varying frequency in endometrial cancer (EC); however, the relevance of many of these mutations is poorly understood. An EC-specific recurrent mutation in the MAX transcription factor p.His28Arg was recently discovered. We sought to assess the functional consequences of this hotspot mutation and determine its association with cancer-relevant phenotypes. Methods MAX was sequenced in 509 endometrioid ECs, and associations between mutation status and clinicopathologic features were assessed. EC cell lines stably expressing MAXH28R were established and used for functional experiments. DNA binding was examined using electrophoretic mobility shift assays and chromatin immunoprecipitation. Transcriptional profiling was performed with microarrays. Murine flank (six to 11 mice per group) and intraperitoneal tumor models were used for in vivo studies. Vascularity of xenografts was assessed by MECA-32 immunohistochemistry. The paracrine pro-angiogenic nature of MAXH28R-expressing EC cells was tested using microfluidic HUVEC sprouting assays and VEGFA enzyme-linked immunosorbent assays. All statistical tests were two-sided. Results Twenty-two of 509 tumors harbored mutations in MAX, including 12 tumors with the p.His28Arg mutation. Patients with a MAX mutation had statistically significantly reduced recurrence-free survival (hazard ratio = 4.00, 95% confidence interval = 1.15 to 13.91, P = .03). MAXH28R increased affinity for canonical E-box sequences, and MAXH28R-expressing EC cells dramatically altered transcriptional profiles. MAXH28R-derived xenografts statistically significantly increased vascular area compared with MAXWT and empty vector tumors (P = .003 and P = .008, respectively). MAXH28R-expressing EC cells secreted nearly double the levels of VEGFA compared with MAXWT cells (P = .03, .005, and .005 at 24, 48, and 72 hours, respectively), and conditioned media from MAXH28R cells increased sprouting when applied to HUVECs. Conclusion These data highlight the importance of MAX mutations in EC and point to increased vascularity as one mechanism contributing to clinical aggressiveness of EC. Endometrial cancer (EC) is the most common gynecologic cancer in the United States and one of a few cancer types for which both incidence and mortality are increasing (1,2). Endometrioid endometrial cancer (EEC) is the most common histologic subtype, accounting for approximately 85% of cases (3). Genetically, EEC is a highly mutated tumor type. Many driver mutations have been identified; however, our present understanding of gene defects that contribute to specific cancer processes such as invasion, metastasis, and angiogenesis is limited. Based on recent findings from The Cancer Genome Atlas (TCGA) and subsequent in silico analyses, the transcription factor MYC-associated factor X (MAX) emerged as a new genetic factor likely to influence EC tumor biology (4–6). MAX is the obligate binding partner of MYC, a master transcription factor with pro-proliferation, pro-growth, and oncogenic functions. MYC has low binding affinity for DNA, but MYC:MAX heterodimers bind E-box sequences to regulate gene expression (7,8). MAX plays an equally important role as a binding partner for members of the MAX dimerization proteins (MXDs) family including MNT, MXD1,3-4, and MGA. MAX:MXD family dimers oppose the pro-growth effects of MYC by promoting expression of differentiation and quiescence genes (9–11). Deregulation of MYC family members is seen in a variety of tumors (12). MAX abnormalities, however, are rare, with the notable exceptions of loss-of-function germline and somatic variants in pheochromocytoma and paraganglioma patients, and somatic loss-of-function mutations in small cell lung cancer (13–15). The missense MAX mutations reported by TCGA in ECs stand in sharp contrast to these loss-of-function MAX variants (4). One particular hotspot mutation discovered by TCGA, the MAX p.His28Arg mutation, has not been previously reported in primary specimens from other cancer types to date. We undertook studies to investigate MAX’s role in EC, particularly the functional consequences of the p.His28Arg mutation. Methods Patient Materials Uterine cancer samples were collected by the Division of Gynecologic Oncology, Washington University, St. Louis, from 1991 to 2010. Written informed consent for this study was obtained and approved by Washington University protocols HSC 91-0507 and HSC 93-0828 and Ohio State University protocol 2012C0116. Microsatellite instability (MSI) status and POLE mutation testing were performed previously (16–18). Nonendometrioid and POLE-mutated ECs were excluded. Targeted Sequencing All coding exons in MAX long and short isoforms (NM002382 and NM145112, respectively) were sequenced to an average of 170× in 509 EECs using the TruSeq Custom Amplicon Kit v1.5 and a MiSeq instrument with Reagent Kit v2 (Illumina, San Diego, CA). Variants were identified using Miseq Reporter software (v2.5.1) (19,20). Electrophoretic Mobility Shift Assays Oligonucleotide sequences used in electrophoretic mobility shift assays (EMSAs) are given in Supplementary Table 1 (available online). In vitro translated (IVT) protein or nuclear lysates were incubated at 42 °C for 10 minutes, followed by incubation with the indicated antibodies (Santa Cruz Biotechnology, Dallas, TX: c-MYC [N-262, sc-764] and MAX [C-17, sc-197]) at 25 °C for 15 minutes, then addition of oligonucleotides and incubation at 25 °C for 20 minutes. Gel shift assays were performed with the LightShift Chemiluminescent EMSA kit (Thermo Fisher Scientific, Waltham, MA). Results are representative of three experiments. Additional details are provided in the Supplementary Methods (available online). Cell Culture 293T and AN3CA (parental and MAX-expressing) cells were cultured in DMEM (Sigma Aldrich, St. Louis, MO) with 10% fetal bovine serum (FBS). RL95-2 and Ishikawa (parental and MAX-expressing) cells were cultured in F12:DMEM (Life Technologies, Carlsbad, CA) with 10% FBS. Human umbilical vein endothelial cells (HUVECs) were cultured in endothelial growth medium 2 (Lonza, Basel, Switzerland). Cell lines were confirmed mycoplasma negative. The estimated number of passages between authentication of cancer cell lines and completion of experiments is 25. Additional details regarding cell line origin, authentication, and forced expression of MAX are provided in the Supplementary Methods (available online). Conditioned Media Preparation and VEGFA Enzyme-Linked Immunosorbent Assay Cells were plated in duplicate in six-well plates at 1 × 106 cells/well (AN3CA-derived cell lines) or 3.5 × 106 cells/well (Ishikawa- and RL95-2-derived cell lines). After cells were attached, media was changed then collected either 24, 48, or 72 hours later. Media was centrifuged at 211 g (4 °C) for 10 minutes. Supernatants were aliquoted and snap frozen on dry ice. Cells were counted to normalize VEGFA protein quantification. VEGFA levels were measured using the Human VEGF Enzyme-Linked Immunosorbent Assay (ELISA) Kit (ThermoFisher Scientific). Absorbance at 450 nm was measured, and a four-parameter logistic standard curve was used to determine concentrations. Biologic duplicates were assayed in technical duplicate. Microfluidic Sprouting Assay Microfluidic HUVEC sprouting assays were performed as described previously (21). Conditioned media from MAXWT, MAXH28R, or empty vector (EV) AN3CA cells was introduced into HUVEC-lined microchannels, and HUVEC sprouting into a central collagen channel was observed. Images of HUVEC sprouting were acquired immediately before and 24 hours after treatment with conditioned medium. Images were processed and analyzed with ImageJ software to calculate the normalized sprouting ratio for each aperture. Additional details are provided in the Supplementary Methods (available online). In Vivo Tumor Models Six- to eight-week-old outbred, NCr-nu/nu females were utilized for flank (six to 11 mice per group) and intraperitoneal (two to three mice per group) xenograft studies. For flank xenografts, researchers blinded to tumor genotypes measured tumors twice weekly. Animal studies are covered by the Ohio State University IACUC Protocols No. A201300000141 and No. 2012A00000008-R1. Additional details are provided in the Supplementary Methods (available online). Immunohistochemistry and Xenograft Vessel Quantification Immunohistochemistry was performed using a Bond Rx autostainer (Leica, Wetzlar, Germany). Automated dewaxing, rehydration, antigen retrieval, blocking, primary antibody incubation, postprimary antibody incubation, detection, and counterstaining were performed using Bond reagents (Leica). Pan-endothelial cell antigen (MECA-32; 1:200, BD Pharmingen, 550563, San Diego, CA) with rabbit antirat IgG (1:200, Vector Laboratories, AI-4001, Burlingame, CA) and PECAM-1 (CD31; 1:1000, Santa Cruz Biotechnology, sc1506R) antibodies were used with Bond Polymer Refine Detection (Leica, S9800). Images were captured using the Vectra Intelligent Slide Analysis System (PerkinElmer, Waltham, MA). Quantification of vessel area was performed on MECA-32 stained slides for six random 20× images per xenograft using inform 2.1 (PerkinElmer) and imageJ. Additional information on sample preparation is provided in the Supplementary Methods (available online). Statistical Analysis Associations between MAX mutation status and MSI, stage, recurrence/progression, age, lymphovascular space invasion, race, and adjuvant therapy were calculated using two-sided Fisher's exact tests. P values for body mass index (BMI) and grade were calculated using two-sided chi-square tests. For univariate survival analysis, P value was determined by log-rank test, and hazard ratio (HR) and 95% confidence interval (CI) were obtained using the Mantel Haenszel approach. Cox proportional hazards multivariable analysis was performed using MAX mutation status and features known to be associated with outcome (18,22). The proportional hazards assumption was tested using the scaled Schoenfeld residuals and the Kaplan-Meier transformed survival times. None of the variables in the multivariable analysis was statistically significant (P> .05), indicating a lack of evidence for departure from the proportional hazards assumption. Ten patients with perioperative deaths (<30 days) and 24 with persistent (stage IVB) disease were excluded from analysis. T tests were used to determine the statistical significance of differences observed in chromatin immunoprecipitation (ChIP)-quantitative polymerase chain reaction (qPCR), quantitative reverse transcription polymerase chain reaction (qRT-PCR), and ELISA experiments. T tests with Welch’s correction when appropriate were used in the TCGA RNAseq data analysis. One-way analysis of variance (ANOVA) with Bonferroni’s multiple comparison test was used in the MECA-32+ area analysis. HUVEC sprouting assay was analyzed using one-way ANOVA with Tukey’s-Cramer post-test. P values of less than .05 were considered statistically significant. All statistical tests were two-sided. All calculations were performed using SPSS 22 (IBM, Armonk, NY), Prism 5 (GraphPad, La Jolla, CA), R Survival Package, and JMP (SAS Institute, Cary NC). Additional methods for mutation modeling, immunoprecipitation, microarray expression profiling, MAX and MYC genic occupancy, TCGA RNA-seq analysis, qRT-PCR, Sanger sequencing, TA cloning, and in vitro cell-based assays are provided in the Supplementary Methods (available online). Results MAX Mutations in EEC and Associations With Clinicopathologic Features Including Outcome MAX was sequenced in 509 primary EECs (18,22,23). The overall mutation rate was 4.3% (22 tumors with mutations). The p.His28Arg mutation was observed in 12 tumors, along with nine additional somatic variants in 11 tumors (Figure 1A). The hotspot p.His28Arg mutation was also observed by TCGA (Figure 1A) (4). Variant allele fractions (VAFs) were consistent with full clonality (Supplementary Table 2, available online). cDNA analysis for primary tumors showed expression of both wild-type (WT) and mutant alleles (corresponding to p.His28Arg), and immunoblot analysis showed comparable levels of MAX expression in wild-type and mutant tumors (data not shown). One tumor (specimen 2219) had two somatic mutations, p.His28Arg and the known cancer susceptibility allele p.Arg75*, which occurred in trans (Supplementary Figure 1, available online) (13). Two tumors (specimens 1122 and 1913) had apparent gain or loss of the MAX locus, as evidenced by single nucleotide polymorphism VAFs (Supplementary Table 2, available online). Figure 1. View largeDownload slide MAX mutations in endometrioid endometrial carcinoma (EEC). A) Schematic of MAX mutations identified in this study of 509 EEC samples (above) and by The Cancer Genome Atlas (below) (4), shown on the long isoform (160 amino acids). The hotspot p.His28Arg mutation seen in 12 tumors and the p.Arg60Gln mutation seen in three tumors both map to the helix-loop-helix domain (green). B) Kaplan-Meier plots show statistically significantly reduced recurrence-free survival (RFS) for women with MAX-mutant tumors. P value was determined by log-rank test. Hazard ratio and 95% confidence interval were obtained using the Mantel Haenszel approach. The survival curve was truncated at eight years. C) Multivariable analysis for RFS includes clinical variables commonly associated with outcome and MAX mutation status. P values and hazard ratios were calculated using Cox proportional hazards model. Variables included in model are those frequently shown to be prognostic in univariate analyses. Statistical significance was calculated using multivariable Cox proportional hazard tests. P values are two-sided. All statistical tests were two-sided. CI = confidence interval; HR = hazard ratio. Figure 1. View largeDownload slide MAX mutations in endometrioid endometrial carcinoma (EEC). A) Schematic of MAX mutations identified in this study of 509 EEC samples (above) and by The Cancer Genome Atlas (below) (4), shown on the long isoform (160 amino acids). The hotspot p.His28Arg mutation seen in 12 tumors and the p.Arg60Gln mutation seen in three tumors both map to the helix-loop-helix domain (green). B) Kaplan-Meier plots show statistically significantly reduced recurrence-free survival (RFS) for women with MAX-mutant tumors. P value was determined by log-rank test. Hazard ratio and 95% confidence interval were obtained using the Mantel Haenszel approach. The survival curve was truncated at eight years. C) Multivariable analysis for RFS includes clinical variables commonly associated with outcome and MAX mutation status. P values and hazard ratios were calculated using Cox proportional hazards model. Variables included in model are those frequently shown to be prognostic in univariate analyses. Statistical significance was calculated using multivariable Cox proportional hazard tests. P values are two-sided. All statistical tests were two-sided. CI = confidence interval; HR = hazard ratio. Patients with MAX-mutant tumors had reduced recurrence-free survival (RFS; HR = 4.00, 95% CI = 1.15 to 13.91, P = .03) (Figure 1B). Mutation was statistically significantly associated with microsatellite instability, tumor stage, and patient BMI (Supplementary Table 3, available online). MAX mutation status remained an independent predictor of reduced recurrence-free survival in a multivariable model that included age, grade, stage, and lymphovascular space invasion (HR = 2.95, 95% CI = 1.20 to 7.29, P = .02), suggesting that MAX defects contribute to clinical aggressiveness (Figure 1C). MAX p.His28Arg DNA Binding Histidine 28 makes a critical contact with the E-box sequence (5’-CACGTG-3’) by hydrogen bonding between NE2 and N7 and/or O6 of Gua(3’) (24). Based on in silico prediction, the His28Arg substitution likely increases affinity for Gua(3’) due to reduced bond length and/or bidentate interaction (Figure 2A). We therefore assessed DNA binding by EMSAs with IVT proteins. As MAX short isoform homodimers bind E-boxes poorly, we assayed long isoform MAXWT and MAXH28R using a canonical two-E-box sequence from the NPM1 promoter (25,26). MAXH28R had greater affinity compared with MAXWT and the supershifted complexes electrophoresed at different rates, consistent with altered protein:DNA interaction (Figure 2B;Supplementary Figure 2A, available online). This was confirmed using AN3CA EC cells overexpressing MAXWT or MAXH28R long isoform, and 293T cells co-expressing both isoforms of MAXWT or MAXH28R (Supplementary Figure 2, B and C, available online). The protein:DNA complex is likely a result of MAXH28R homodimers as co-incubation with an anti-cMYC antibody did not result in a supershift (Figure 2C;Supplementary Figure 2C, available online). Because protein levels of MAXH28R and MAXWT were similar (Figure 2, B and C), the observed results were not due to differences in MAX protein quantity. EMSA findings were similar using a previously studied four-E-box oligonucleotide from the CDK4 promoter (Supplementary Figure 2D) (27). Single E-box oligonucleotides, however, showed no differences between MAXH28R and MAXWT, suggesting that multimeric MAX complexes could account for the differences seen with four- and two-E-box oligonucleotides (data not shown). MAX short isoform homodimers bound E-boxes poorly (Supplementary Figure 2E, available online). When c-MYC was co-expressed with MAXH28R, the MAX shift disappeared, indicating that the p.His28Arg substitution does not hinder MAX:MYC interaction (the MAX:MYC:DNA interaction was not detectable under the conditions used) (Figure 2C). Immunoprecipitation and immunoblot demonstrated exogenous MAXWT, and MAXH28R interacted with endogenous c-MYC and endogenous MAX (Figure 2D). Figure 2. View largeDownload slide p.His28Arg alters MAX E-box binding and transcription. A) MAX/MYC heterodimer bound to a canonical E-box. MAXH28R is predicted to place η2 of the guanidine group in closer proximity to N7 of Gua (3’) (green dotted line) and could allow for bidentate interaction (purple dotted lines). B) Electrophoretic mobility shift assay (EMSA) shows differential binding of MAXH28R to NPM1 E-boxes. In vitro–translated MAXH28R protein created a shift (lane 5), whereas MAXWT did not (lane 3). Anti-MAX antibody supershifted the MAXH28R band (lane 6) and produced a different supershift with MAXWT (lane 4) (see also Supplementary Figure 2A, available online). C) EMSA with MAX-transfected AN3CA cell nuclear extracts points to MAX homodimer:DNA interaction. Anti-cMYC antibody did not supershift the MAXH28R band (lanes 1 and 3), and cotransfection of cMYC eliminated the shift band, consistent with reduced MAX homodimer levels due to increased MAX:cMYC heterodimerization. Immunoblot (right) shows the relative amounts of MAX and cMYC. D) Immunoprecipitation demonstrating myc-tagged-MAXH28R interaction with cMYC and endogenous MAX. E) Chromatin immunoprecipitation (ChIP)–quantitative polymerase chain reaction (qPCR) on AN3CA cells stably expressing MAXH28R and MAXWT (see Supplementary Figure 3, available online) for E-box regions in NPM1, CDK4, and CAD, and control regions in CTCF and KLHL13.F) Volcano plot (middle) of expression differences between patients with MAX-mutated (n = 8) and MAX-WT (n = 154) tumors from The Cancer Genome Atlas (3). Dot plots in side panels show biologic validation of GATM and SERPINB9 expression determined by qRT-PCR on MAXH28R-mutated tumors (n = 11) and matched MAXWT tumors (n = 24) from our series. Bar graphs in side panels show ChIP-qPCR performed as in (E). ChIP-qPCR and qRT-PCR experiments in (E and F) are shown as mean (SD). P values calculated by two-sided t tests; EMSAs and blots shown are representative of three experiments. ChIPs performed on two independent immunoprecipitations per cell line. EV = empty vector; WT = wild-type. Figure 2. View largeDownload slide p.His28Arg alters MAX E-box binding and transcription. A) MAX/MYC heterodimer bound to a canonical E-box. MAXH28R is predicted to place η2 of the guanidine group in closer proximity to N7 of Gua (3’) (green dotted line) and could allow for bidentate interaction (purple dotted lines). B) Electrophoretic mobility shift assay (EMSA) shows differential binding of MAXH28R to NPM1 E-boxes. In vitro–translated MAXH28R protein created a shift (lane 5), whereas MAXWT did not (lane 3). Anti-MAX antibody supershifted the MAXH28R band (lane 6) and produced a different supershift with MAXWT (lane 4) (see also Supplementary Figure 2A, available online). C) EMSA with MAX-transfected AN3CA cell nuclear extracts points to MAX homodimer:DNA interaction. Anti-cMYC antibody did not supershift the MAXH28R band (lanes 1 and 3), and cotransfection of cMYC eliminated the shift band, consistent with reduced MAX homodimer levels due to increased MAX:cMYC heterodimerization. Immunoblot (right) shows the relative amounts of MAX and cMYC. D) Immunoprecipitation demonstrating myc-tagged-MAXH28R interaction with cMYC and endogenous MAX. E) Chromatin immunoprecipitation (ChIP)–quantitative polymerase chain reaction (qPCR) on AN3CA cells stably expressing MAXH28R and MAXWT (see Supplementary Figure 3, available online) for E-box regions in NPM1, CDK4, and CAD, and control regions in CTCF and KLHL13.F) Volcano plot (middle) of expression differences between patients with MAX-mutated (n = 8) and MAX-WT (n = 154) tumors from The Cancer Genome Atlas (3). Dot plots in side panels show biologic validation of GATM and SERPINB9 expression determined by qRT-PCR on MAXH28R-mutated tumors (n = 11) and matched MAXWT tumors (n = 24) from our series. Bar graphs in side panels show ChIP-qPCR performed as in (E). ChIP-qPCR and qRT-PCR experiments in (E and F) are shown as mean (SD). P values calculated by two-sided t tests; EMSAs and blots shown are representative of three experiments. ChIPs performed on two independent immunoprecipitations per cell line. EV = empty vector; WT = wild-type. To further investigate MAXH28R:DNA interaction, we performed ChIP-qPCR using AN3CA cells stably expressing flag-tagged MAXH28R or MAXWT at levels comparable with endogenous MAX (Supplementary Figure 3, available online). We assessed binding to E-boxes in the MAX/MYC target genes NPM1, CDK4, and CAD, and control regions in CTCF and KLHL13 (25,27,28). MAXH28R showed statistically significantly enhanced affinity for the E-boxes in NPM1 and CDK4 compared with MAXWT (P < .001 and P = .04, respectively) (Figure 2E). Together, these findings demonstrate that the p.His28Arg mutation alters DNA binding and could, therefore, alter gene expression. MAX p.His28Arg Transcriptional Changes TCGA RNA-seq data for the eight MAX-mutant and 156 MAX-WT EECs were used to explore differentially expressed genes (DEGs) (Figure 2F) (4). We identified 64 candidate DEGs with P values of less than .01 and greater than twofold expression change (Supplementary Figure 4A, available online). The majority (71%) bound MAX and/or MYC based on ENCODE ChIP-seq data, implying that MAX mutation might directly alter transcription (Supplementary Figure 4B, available online) (29). Eight candidate DEGs were validated in our collection of EECs using qRT-PCR. Expression of GATM and SERPINB9 was statistically significantly different between mutant and WT tumors (P = .02 and P = .03, respectively) (Figure 2F), whereas expression of the other six candidates was not statistically significantly different (Supplementary Figure 4C, available online). ChIP-qPCR for cells stably expressing flag-tagged MAX showed statistically significantly increased MAXH28R binding to both GATM and SERPINB9 E-boxes compared with MAXWT (P = .002 and P = .009, respectively) (Figure 2F). The effect of MAXH28R on global gene expression was assessed by whole transcriptome profiling using AN3CA cell lines stably expressing exogenous short or long isoform MAXWT and MAXH28R. MAX protein levels were comparable between stable clones (Supplementary Figure 5A, available online). Principle component analysis proved that biologic duplicates clustered together and MAXH28R-long cell lines were the most different from the other cell types (Supplementary Figure 5B, available online). Both MAXH28R short and long isoform cell lines were clearly different from the cell lines as indicated by hierarchical clustering (Supplementary Figure 5C, available online), which could be a reflection of either direct or indirect effects on transcription. Vascular Phenotype of MAXH28R Xenograft Tumors In vitro assays showed no differences in proliferation, colony formation, or wound healing for the stable MAXH28R- and MAXWT-expressing AN3CA cells (Supplementary Figure 6, A–C, available online). Differences in proliferation rates were evident between MAXH28R- and MAXWT-expressing AN3CA clones but did not reach statistical significance. Similarly, no statistically significant differences in proliferation or colony formation were observed with MAXH28R- and MAXWT-expressing Ishikawa cells (Supplementary Figure 6, A and B, available online). Although the number of colonies did not differ between genotypes in either the AN3CA or Ishikawa cell lines, colony sizes varied. MAXWT-expressing AN3CA colonies were larger than the MAXH28R or empty vector (EV) colonies. On the other hand, MAXH28R-expressing Ishikawa colonies were larger than the MAXWT or EV colonies. Xenograft studies, however, showed striking differences in the flank tumors derived from MAXH28R and MAXWT EC cells. AN3CA MAXH28R tumors had markedly increased vascularity and were much more hemorrhagic than MAXWT tumors (Figure 3A;Supplementary Figure 7, available online). Pan-endothelial cell antigen PVLAP (MECA-32) staining proved that the AN3CA MAXH28R tumors had approximately twice the vascular area compared with MAXWT and EV (P = .003 and P = .008, respectively) (Figure 3, B and C). Similarly, tumors derived from MAXH28R-expressing Ishikawa cells also had statistically significantly increased vascular area compared with MAXWT and EV (P = .03 and P < .001, respectively) (Supplementary Figure 8, available online). Figure 3. View largeDownload slide Assessment of vascularity in AN3CA xenografts. A) Representative flank xenografts derived from MAXWT (n = 22 mice), MAXH28R (n = 22), and empty vector (EV; n = 11) cells. Additional images for xenografts are provided in Supplementary Figure 7 (available online). Scale bars = 5 mm. B) Immunohistochemical staining for endothelial cells in xenografts. Representative images for MECA-32 pan-endothelial marker stained tissues. Scale bars = 100 μm. C) Quantification of MECA-32+ area (percent of total area of image). Three random tumors assessed per condition. Six 20× images quantified per tumor. Data shown as mean ± SD. P values calculated using one-way analysis of variance (P < .01) with Bonferroni multiple comparison test are shown. D) Representative images (left) of intraperitoneal AN3CA xenografts derived from MAXWT clone 5 (n = 3) and MAXH28R clone 2 (n = 2) (see Supplementary Figure 3, available online). Scale bars = 5 mm. Representative images of immunohistochemical staining for endothelial cells by MECA-32 (middle) and CD31 (right) in adjacent sections. Scale bars = 100 μm. EV = empty vector; WT = wild-type. Figure 3. View largeDownload slide Assessment of vascularity in AN3CA xenografts. A) Representative flank xenografts derived from MAXWT (n = 22 mice), MAXH28R (n = 22), and empty vector (EV; n = 11) cells. Additional images for xenografts are provided in Supplementary Figure 7 (available online). Scale bars = 5 mm. B) Immunohistochemical staining for endothelial cells in xenografts. Representative images for MECA-32 pan-endothelial marker stained tissues. Scale bars = 100 μm. C) Quantification of MECA-32+ area (percent of total area of image). Three random tumors assessed per condition. Six 20× images quantified per tumor. Data shown as mean ± SD. P values calculated using one-way analysis of variance (P < .01) with Bonferroni multiple comparison test are shown. D) Representative images (left) of intraperitoneal AN3CA xenografts derived from MAXWT clone 5 (n = 3) and MAXH28R clone 2 (n = 2) (see Supplementary Figure 3, available online). Scale bars = 5 mm. Representative images of immunohistochemical staining for endothelial cells by MECA-32 (middle) and CD31 (right) in adjacent sections. Scale bars = 100 μm. EV = empty vector; WT = wild-type. In a separate intraperitoneal (IP) AN3CA xenograft model, MAXH28R-expressing tumors were markedly more hemorrhagic than MAXWT tumors and had increased vascularization (Figure 3D;Supplementary Figure 9, available online). Immunohistochemistry for PECAM-1 (CD31), another endothelial cell marker, displayed similar staining compared with MECA-32, confirming our earlier results (Figure 3D;Supplementary Figure 10, available online). The tumor volumes for the MAXH28R and MAXWT flank xenografts were similar (Supplementary Figure 11, available online), making it unlikely that tumor size accounted for the vascularity differences. MAXH28R Pro-angiogenic Paracrine Signaling To test the hypothesis that paracrine factors elaborated by MAXH28R-expressing cells contribute to the vascular phenotype in xenografts, we assessed HUVEC sprouting in vitro using a previously described microfluidic model (21). We observed statistically significantly increased HUVEC sprouting with MAXH28R conditioned medium compared with MAXWT and EV (P = .02 and P = .005, respectively) (Figure 4A). As the expression array data (Supplementary Figure 5D, available online) identified VEGFA as a candidate pro-angiogenic factor increased by MAXH28R, VEGFA levels in conditioned media from the EV, MAXWT, and MAXH28R EC cells were determined by ELISA. AN3CA MAXH28R conditioned media had nearly double the concentration of VEGFA as MAXWT at all assessed time points (P = .03, .005, and .005 at 24, 48, and 72 hours, respectively) (Figure 4B). Secreted VEGFA was also nearly doubled in Ishikawa and RL95-2 EC cell lines stably expressing MAXH28R compared with MAXWT (Supplementary Figure 12, available online). Taken together, these data demonstrate that MAXH28R in cancer cell lines promotes increased vascularity and that increased VEGFA secretion could contribute to the pro-angiogenic phenotype. Figure 4. View largeDownload slide Differences in HUVEC sprouting and VEGFA secretion in vitro. A) Representative images (top) of HUVEC sprouting through apertures in the polydimethylsiloxane (PDMS) microdevice into 3D collagen gel with conditioned media from AN3CA empty vector (EV) cells and cells stably expressing MAXWT (clone 7) or MAXH28R (clone 2). Scale bar = 50 μm. Quantification (bottom) of HUVEC sprouting. Bar graphs represent mean (SD) for three independent microdevice experiments per condition. One-way analysis of variance for statistical significance was performed (P < .01), and Tukey’s post hoc pair-wise test for statistical significance is shown. B) Quantification of VEGFA by enzyme-linked immunosorbent assay in media conditioned by AN3CA EV cells and cells stably expressing MAXWT or MAXH28R (see Supplementary Figure 3, available online). Bar graphs represent duplicate experiments, shown as mean ± SD. Statistical significance from the two-sided t test. EV = empty vector; WT = wild-type. Figure 4. View largeDownload slide Differences in HUVEC sprouting and VEGFA secretion in vitro. A) Representative images (top) of HUVEC sprouting through apertures in the polydimethylsiloxane (PDMS) microdevice into 3D collagen gel with conditioned media from AN3CA empty vector (EV) cells and cells stably expressing MAXWT (clone 7) or MAXH28R (clone 2). Scale bar = 50 μm. Quantification (bottom) of HUVEC sprouting. Bar graphs represent mean (SD) for three independent microdevice experiments per condition. One-way analysis of variance for statistical significance was performed (P < .01), and Tukey’s post hoc pair-wise test for statistical significance is shown. B) Quantification of VEGFA by enzyme-linked immunosorbent assay in media conditioned by AN3CA EV cells and cells stably expressing MAXWT or MAXH28R (see Supplementary Figure 3, available online). Bar graphs represent duplicate experiments, shown as mean ± SD. Statistical significance from the two-sided t test. EV = empty vector; WT = wild-type. Discussion Determining how specific and presumably biologically relevant gene defects contribute to cancer phenotypes remains a challenge. Our functional characterization of the recurrent and EC-specific MAX p.His28Arg mutation provides important evidence that the mutation alters MAX function and contributes to a cancer-related phenotype, angiogenesis. The hotspot EC MAX mutation p.His28Arg has altered DNA binding and is associated with marked changes in transcription. EMSAs and ChIP assays showed that MAXH28R had increased E-box binding compared with MAXWT, which is consistent with structural modeling that predicted that the p.His28Arg substitution increases MAX’s affinity for DNA. Our xenograft studies implicated MAXH28R as playing a role in angiogenesis, and the in vitro angiogenesis assays validated the pro-angiogenic effect seen for MAXH28R and pointed to a role for secreted factors. Our demonstration that conditioned media from three different EC cell lines expressing MAXH28R had elevated VEGFA levels further supports a model for secreted factors mediating MAXH28R-associated pro-angiogenesis. MAX-mediated increases in vascularity and VEGFA levels, both of which are associated with poor clinical outcome in EC patients, could in part explain selection for MAX mutation in EC (30–32). The mutation pattern of MAX that we observed strongly suggests a dominant or dominant-negative activity in EC (Figure 1A), consistent with our functional data. The paucity of stop and frameshift mutations and absence of second-hit mutations in EEC is in stark contrast to the loss-of-function abnormalities seen in neuroendocrine tumors, suggesting context-dependent roles for MAX (13–15). Two amino acids in MAX appear to display substitutions that are specific to cancer type and that may have different functional consequences. Val9Met and Arg60Gln are observed in EECs, while Val9Leu and Arg60Trp are observed in pheochromocytomas (4,14). The two pheochromocytoma missense variants have been shown to be loss-of-function in terms of ability to repress MYC-driven expression of a reporter assay in rat PC12 cells null for MAX (33). Although Val9Met and Arg60Gln are yet to be functionally tested in EC, the possibility exists that they are loss-of-function alleles, similar to the corresponding pheochromocytoma variants. However, the effect of these variants must be experimentally determined in additional systems given the complexity of MAX’s transcription regulatory function. There are several examples of context-dependent roles for cancer genes. FGFR2 is an oncogene in EC but acts as a tumor suppressor in melanoma (34,35). Loss- and gain-of-function defects occur in TP53; tumors that have an inherited loss-of-function mutation and a second somatic mutation point to p53’s role as a tumor suppressor, but oncogenic missense changes that alter signaling cascades and chromatin modification also drive cancer in some instances (36–39). Most TP53 mutations abrogate p53’s tumor suppressive function, acting either as cellular recessives or dominant-negatives (38). Gain-of-function TP53 missense mutations in the DNA binding domain cause loss of wild-type p53 tumor suppressive function, alter genomic binding, and confer novel oncogenic activity (36–39). MAX defects may parallel the complex nature of TP53 mutations. Clear loss-of-function deletion and nonsense alleles exist for MAX, in addition to potentially more biologically complex missense mutations involving the DNA binding domain. Some mutations could be tissue-dependent loss-of-function alleles for certain activities, while at the same time conferring novel oncogenic properties. Our finding that a recurrent cancer-specific mutation in MAX acts in a dominant manner is not unexpected. An endogenous truncated MAX protein called dMAX is a naturally occurring example of a dominant-negative MAX. dMAX can dimerize with MYC and MXD members but cannot associate with DNA, and it has been associated with increased aerobic metabolism (40–42). MAX and MAX-like-protein X (MLX) are known to regulate transcriptional networks essential to tumor cell metabolism/nutrient availability, and it is possible that the increased vascularity seen in xenografts expressing MAXH28R is part of a central theme of MAX’s important role in coordination of nutrient availability (43). Further, the effect of the p.His28Arg mutation in DNA binding is likely to influence activities of many MAX-interacting proteins, not just its canonical partner MYC. This study is limited by characterization of a single hotspot mutation in the MAX gene. Further efforts will be required to determine whether the other observed MAX mutations alter DNA binding and/or increase angiogenesis. Although our study found that MAX mutation in EC patients was associated with poor outcome, analysis of additional cohorts will be required to validate the prognostic significance. Our results link the MAX p.His28Arg mutation and tumor angiogenesis. The global effect on transcriptional profiles observed in cells expressing MAXH28R also implies there are additional gene expression changes that could contribute to biologic aggressiveness in EEC. Taken together, the findings from our mutation and functionalization studies implicate mutant MAX as a driver of aggressive EEC. Funding This study was funded by the National Institutes of Health (R21 CA155674 to P. J. Goodfellow), the National Cancer Institute (P30 CA016058 supporting the Genomics and Biostatistics shared resources at the Ohio State University Comprehensive Cancer Center), the National Institute of General Medical Sciences (T32 GM068412 to C. M. Rush), the Pelotonia Fellowship Program (C. J. Walker), The American Heart Association (15SDG25480000 to J. W. Song), and The American Cancer Society (IRG-67-003-50 to J. W. Song). Notes The study funders had no role in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication. The authors have declared that no conflict of interest exists. PJG and PD devised the initial concept of the study. DGM provided primary tumor specimens. CJW, MJO, and MS performed mutation screening and clinical correlation studies. CMR, CJW, and PJG performed predictive mutation modeling. CJW and PD performed electrophoretic mobility shift assays. MJO and CJW performed co-immunoprecipitations. CMR performed chromatin immunoprecipitation. CJW and CMR were responsible for TCGA data analysis and other bioinformatics analyses. CJW, PD, and CMR generated stable cell lines, RS, KS, PD, CMR, BS, CJW, RAZ, CMC, JLG, and MJO performed in vivo studies, including cell line implantation and animal husbandry (RS), harvesting (CMR, RS, CMC, RAZ, JLG, and BS), and vessel quantification (CJW, CMR, and MJO). Cell phenotype assessment was performed by PD, CMR, and CJW, including cell growth (PD and CMR), colony forming ability (CMR), and migratory potential (CJW). Angiogenic sprouting assays were performed by CWC and JWS. Enzyme-linked immunosorbent assays were performed by CMR. DGM and DEC provided indispensable guidance in study design and implementation. CJW, CMR, PD, and PJG wrote the manuscript with input from other authors. The final manuscript was approved by all authors. We would like to thank Mark Foster for assistance with mutation modeling and interpretation of chemical interactions, and we thank Qianben Wang and Benjamin Sunkel for their assistance with chromatin immunoprecipitation–quantitative polymerase chain reaction experiments. We also thank Alexis Chassen for manuscript editing and data curation. We thank Joseph McElroy for bioinformatics/statistical counsel. We acknowledge Alex Seibel for assisting with the sprouting measurements. We would like to acknowledge Raleigh Kladney, Cynthia Timmers, and the Solid Tumor Translational Science Core Facility. We acknowledge the Target Validation Core Facility. We would like to acknowledge Pearlly Yan, David Symer, and Sarah Warner with the Ohio State University Genomics Core Facility, and Tea Meulia with the Ohio State University Molecular and Cellular Imagine Center, a CFAES/OARDX core facility in Wooster, Ohio. We are very grateful to all of the patients who contributed specimens to this study and all of the attending physicians and staff at the Washington University School of Medicine Division of Gynecologic Oncology, as well as The Ohio State University College of Medicine Division of Gynecologic Oncology. References 1 Siegel RL , Miller KD , Jemal A. Cancer statistics , 2017 . CA Cancer J Clin . 2017; 67 ( 1 ): 7 – 30 . http://dx.doi.org/10.3322/caac.21387 Google Scholar CrossRef Search ADS PubMed 2 Morice P , Leary A , Creutzberg C , Abu-Rustum N , Darai E. Endometrial cancer . Lancet. 2016 ; 387 ( 10023 ): 1094 – 1108 . Google Scholar CrossRef Search ADS PubMed 3 Creasman WT , Odicino F , Maisonneuve P et al. , Carcinoma of the corpus uteri. FIGO 26th annual report on the results of treatment in gynecological cancer . Int J Gynaecol Obstet . 2006 ; 95 (suppl 1): S105 – S143 . Google Scholar CrossRef Search ADS 4 Cancer Genome Atlas Research Network , Kandoth C , Schultz N et al. , Integrated genomic characterization of endometrial carcinoma . Nature . 2013 ; 497 ( 7447 ): 67 – 73 . http://dx.doi.org/10.1038/nature12113 Google Scholar CrossRef Search ADS PubMed 5 Chang MT , Asthana S , Gao SP et al. , Identifying recurrent mutations in cancer reveals widespread lineage diversity and mutational specificity . Nat Biotechnol . 2016 ; 34 ( 2 ): 155 – 163 . Google Scholar CrossRef Search ADS PubMed 6 Kamburov A , Lawrence MS , Polak P et al. , Comprehensive assessment of cancer missense mutation clustering in protein structures . Proc Natl Acad Sci U S A . 2015 ; 112 ( 40 ): E5486 – E5495 . Google Scholar CrossRef Search ADS PubMed 7 Atchley WR , Fitch WM. Myc and Max: Molecular evolution of a family of proto-oncogene products and their dimerization partner . Proc Natl Acad Sci U S A. 1995 ; 92 ( 22 ): 10217 – 10221 . http://dx.doi.org/10.1073/pnas.92.22.10217 Google Scholar CrossRef Search ADS PubMed 8 Meyer N , Penn LZ. Reflecting on 25 years with MYC . Nat Rev Cancer . 2008 ; 8 ( 12 ): 976 – 990 . http://dx.doi.org/10.1038/nrc2231 Google Scholar CrossRef Search ADS PubMed 9 Grinberg AV , Hu CD , Kerppola TK. Visualization of Myc/Max/Mad family dimers and the competition for dimerization in living cells . Mol Cell Biol. 2004 ; 24 ( 10 ): 4294 – 4308 . http://dx.doi.org/10.1128/MCB.24.10.4294-4308.2004 Google Scholar CrossRef Search ADS PubMed 10 Hurlin PJ , Queva C , Eisenman RN. Mnt, a novel Max-interacting protein is coexpressed with Myc in proliferating cells and mediates repression at Myc binding sites . Genes Dev. 1997 ; 11 ( 1 ): 44 – 458 . http://dx.doi.org/10.1101/gad.11.1.44 Google Scholar CrossRef Search ADS PubMed 11 Walker W , Zhou ZQ , Ota S , Wynshaw-Boris A , Hurlin PJ. Mnt-Max to Myc-Max complex switching regulates cell cycle entry . J Cell Biol. 2005 ; 169 ( 3 ): 405 – 413 . Google Scholar CrossRef Search ADS PubMed 12 Tansey WP. Mammalian MYC proteins and cancer . N J Sci. 2014 ; 2014 : 1 – 27 . http://dx.doi.org/10.1155/2014/757534 Google Scholar CrossRef Search ADS 13 Comino-Mendez I , Gracia-Aznarez FJ , Schiavi F et al. , Exome sequencing identifies MAX mutations as a cause of hereditary pheochromocytoma . Nat Genet . 2011 ; 43 ( 7 ): 663 – 667 . Google Scholar CrossRef Search ADS PubMed 14 Burnichon N , Cascon A , Schiavi F et al. , MAX mutations cause hereditary and sporadic pheochromocytoma and paraganglioma . Clin Cancer Res. 2012 ; 18 ( 10 ): 2828 – 2837 . http://dx.doi.org/10.1158/1078-0432.CCR-12-0160 Google Scholar CrossRef Search ADS PubMed 15 Romero OA , Torres-Diz M , Pros E et al. , MAX inactivation in small cell lung cancer disrupts MYC-SWI/SNF programs and is synthetic lethal with BRG1 . Cancer Discov . 2014 ; 4 ( 3 ): 292 – 303 . http://dx.doi.org/10.1158/2159-8290.CD-13-0799 Google Scholar CrossRef Search ADS PubMed 16 Zighelboim I , Goodfellow PJ , Gao F et al. , Microsatellite instability and epigenetic inactivation of MLH1 and outcome of patients with endometrial carcinomas of the endometrioid type . J Clin Oncol. 2007 ; 25 ( 15 ): 2042 – 2048 . http://dx.doi.org/10.1200/JCO.2006.08.2107 Google Scholar CrossRef Search ADS PubMed 17 Zighelboim I , Schmidt AP , Gao F et al. , ATR mutation in endometrioid endometrial cancer is associated with poor clinical outcomes . J Clin Oncol. 2009 ; 27 ( 19 ): 3091 – 3096 . http://dx.doi.org/10.1200/JCO.2008.19.9802 Google Scholar CrossRef Search ADS PubMed 18 Billingsley CC , Cohn DE , Mutch DG , Stephens JA , Suarez AA , Goodfellow PJ. Polymerase varepsilon (POLE) mutations in endometrial cancer: Clinical outcomes and implications for Lynch syndrome testing . Cancer. 2015 ; 121 ( 3 ): 386 – 394 . Google Scholar CrossRef Search ADS PubMed 19 DePristo MA , Banks E , Poplin R et al. , A framework for variation discovery and genotyping using next-generation DNA sequencing data . Nat Genet. 2011 ; 43 ( 5 ): 491 – 498 . http://dx.doi.org/10.1038/ng.806 Google Scholar CrossRef Search ADS PubMed 20 McKenna A , Hanna M , Banks E et al. , The Genome Analysis Toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010 ; 20 ( 9 ): 1297 – 1303 . http://dx.doi.org/10.1101/gr.107524.110 21 Song JW , Munn LL. Fluid forces control endothelial sprouting . Proc Natl Acad Sci U S A. 2011 ; 108 ( 37 ): 15342 – 15347 . http://dx.doi.org/10.1073/pnas.1105316108 Google Scholar CrossRef Search ADS PubMed 22 Walker CJ , Miranda MA , O'Hern MJ et al. , Patterns of CTCF and ZFHX3 mutation and associated outcomes in endometrial cancer . J Natl Cancer Inst . 2015 ; 107 ( 11 ):djv249. 23 Zighelboim I , Mutch DG , Knapp A et al. , High frequency strand slippage mutations in CTCF in MSI-positive endometrial cancers . Hum Mutat. 2014 ; 35 ( 1 ): 63 – 65 . http://dx.doi.org/10.1002/humu.22463 Google Scholar CrossRef Search ADS PubMed 24 Nair SK , Burley SK. X-ray structures of Myc-Max and Mad-Max recognizing DNA. Molecular bases of regulation by proto-oncogenic transcription factors . Cell. 2003 ; 112 ( 2 ): 193 – 205 . Google Scholar CrossRef Search ADS PubMed 25 Zeller KI , Haggerty TJ , Barrett JF , Guo Q , Wonsey DR , Dang CV. Characterization of nucleophosmin (B23) as a Myc target by scanning chromatin immunoprecipitation . J Biol Chem. 2001 ; 276 ( 51 ): 48285 – 48291 . Google Scholar CrossRef Search ADS PubMed 26 Prochownik EV , VanAntwerp ME. Differential patterns of DNA binding by myc and max proteins . Proc Natl Acad Sci U S A. 1993 ; 90 ( 3 ): 960 – 964 . http://dx.doi.org/10.1073/pnas.90.3.960 Google Scholar CrossRef Search ADS PubMed 27 Hermeking H , Rago C , Schuhmacher M et al. , Identification of CDK4 as a target of c-MYC . Proc Natl Acad Sci U S A. 2000 ; 97 ( 5 ): 2229 – 2234 . http://dx.doi.org/10.1073/pnas.050586197 Google Scholar CrossRef Search ADS PubMed 28 Miltenberger RJ , Sukow KA , Farnham PJ. An E-box-mediated increase in cad transcription at the G1/S-phase boundary is suppressed by inhibitory c-Myc mutants . Mol Cell Biol. 1995 ; 15 ( 5 ): 2527 – 2535 . http://dx.doi.org/10.1128/MCB.15.5.2527 Google Scholar CrossRef Search ADS PubMed 29 Rosenbloom KR , Sloan CA , Malladi VS et al. , ENCODE data in the UCSC Genome Browser: Year 5 update . Nucleic Acids Res. 2013 ; 41 (database issue): D56 – D63 . Google Scholar CrossRef Search ADS PubMed 30 Chen CA , Cheng WF , Lee CN et al. , Cytosol vascular endothelial growth factor in endometrial carcinoma: Correlation with disease-free survival . Gynecol Oncol. 2001 ; 80 ( 2 ): 207 – 212 . http://dx.doi.org/10.1006/gyno.2000.6048 Google Scholar CrossRef Search ADS PubMed 31 Dobrzycka B , Terlikowski SJ , Kowalczuk O , Kulikowski M , Niklinski J. Serum levels of VEGF and VEGF-C in patients with endometrial cancer . Eur Cytokine Netw . 2011 ; 22 ( 1 ): 45 – 51 . Google Scholar PubMed 32 Kamat AA , Merritt WM , Coffey D et al. , Clinical and biological significance of vascular endothelial growth factor in endometrial cancer . Clin Cancer Res . 2007 ; 13 ( 24 ): 7487 – 7495 . http://dx.doi.org/10.1158/1078-0432.CCR-07-1017 Google Scholar CrossRef Search ADS PubMed 33 Comino-Mendez I , Leandro-Garcia LJ , Montoya G et al. , Functional and in silico assessment of MAX variants of unknown significance . J Mol Med (Berl) . 2015 ; 93 ( 11 ): 1247 – 1255 . Google Scholar CrossRef Search ADS PubMed 34 Pollock PM , Gartside MG , Dejeza LC et al. , Frequent activating FGFR2 mutations in endometrial carcinomas parallel germline mutations associated with craniosynostosis and skeletal dysplasia syndromes . Oncogene. 2007 ; 26 ( 50 ): 7158 – 7162 . http://dx.doi.org/10.1038/sj.onc.1210529 Google Scholar CrossRef Search ADS PubMed 35 Gartside MG , Chen H , Ibrahimi OA et al. , Loss-of-function fibroblast growth factor receptor-2 mutations in melanoma . Mol Cancer Res. 2009 ; 7 ( 1 ): 41 – 54 . http://dx.doi.org/10.1158/1541-7786.MCR-08-0021 Google Scholar CrossRef Search ADS PubMed 36 Lang GA , Iwakuma T , Suh YA et al. , Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome . Cell. 2004 ; 119 ( 6 ): 861 – 872 . http://dx.doi.org/10.1016/j.cell.2004.11.006 Google Scholar CrossRef Search ADS PubMed 37 Olive KP , Tuveson DA , Ruhe ZC et al. , Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome . Cell. 2004 ; 119 ( 6 ): 847 – 860 . http://dx.doi.org/10.1016/j.cell.2004.11.004 Google Scholar CrossRef Search ADS PubMed 38 Oren M , Rotter V. Mutant p53 gain-of-function in cancer . Cold Spring Harb Perspect Biol. 2010 ; 2 ( 2 ): a001107 . Google Scholar CrossRef Search ADS PubMed 39 Zhu J , Sammons MA , Donahue G et al. , Gain-of-function p53 mutants co-opt chromatin pathways to drive cancer growth . Nature. 2015 ; 525 ( 7568 ): 206 – 211 . http://dx.doi.org/10.1038/nature15251 Google Scholar CrossRef Search ADS PubMed 40 Makela TP , Koskinen PJ , Vastrik I , Alitalo K. Alternative forms of Max as enhancers or suppressors of Myc-ras cotransformation . Science. 1992 ; 256 ( 5055 ): 373 – 377 . http://dx.doi.org/10.1126/science.256.5055.373 Google Scholar CrossRef Search ADS PubMed 41 FitzGerald MJ , Arsura M , Bellas RE et al. , Differential effects of the widely expressed dMax splice variant of Max on E-box vs initiator element-mediated regulation by c-Myc . Oncogene. 1999 ; 18 ( 15 ): 2489 – 2498 . http://dx.doi.org/10.1038/sj.onc.1202611 Google Scholar CrossRef Search ADS PubMed 42 Babic I , Anderson ES , Tanaka K et al. , EGFR mutation-induced alternative splicing of Max contributes to growth of glycolytic tumors in brain cancer . Cell Metab. 2013 ; 17 ( 6 ): 1000 – 1008 . http://dx.doi.org/10.1016/j.cmet.2013.04.013 Google Scholar CrossRef Search ADS PubMed 43 O'Shea JM , Ayer DE. Coordination of nutrient availability and utilization by MAX- and MLX-centered transcription networks . Cold Spring Harbor Perspect Med . 2013 ; 3 ( 9 ): a014258 . Google Scholar CrossRef Search ADS © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JNCI: Journal of the National Cancer Institute Oxford University Press

Loading next page...
 
/lp/ou_press/max-mutations-in-endometrial-cancer-clinicopathologic-associations-and-prqeAnFTED
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.
ISSN
0027-8874
eISSN
1460-2105
D.O.I.
10.1093/jnci/djx238
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Genomic studies have revealed that multiple genes are mutated at varying frequency in endometrial cancer (EC); however, the relevance of many of these mutations is poorly understood. An EC-specific recurrent mutation in the MAX transcription factor p.His28Arg was recently discovered. We sought to assess the functional consequences of this hotspot mutation and determine its association with cancer-relevant phenotypes. Methods MAX was sequenced in 509 endometrioid ECs, and associations between mutation status and clinicopathologic features were assessed. EC cell lines stably expressing MAXH28R were established and used for functional experiments. DNA binding was examined using electrophoretic mobility shift assays and chromatin immunoprecipitation. Transcriptional profiling was performed with microarrays. Murine flank (six to 11 mice per group) and intraperitoneal tumor models were used for in vivo studies. Vascularity of xenografts was assessed by MECA-32 immunohistochemistry. The paracrine pro-angiogenic nature of MAXH28R-expressing EC cells was tested using microfluidic HUVEC sprouting assays and VEGFA enzyme-linked immunosorbent assays. All statistical tests were two-sided. Results Twenty-two of 509 tumors harbored mutations in MAX, including 12 tumors with the p.His28Arg mutation. Patients with a MAX mutation had statistically significantly reduced recurrence-free survival (hazard ratio = 4.00, 95% confidence interval = 1.15 to 13.91, P = .03). MAXH28R increased affinity for canonical E-box sequences, and MAXH28R-expressing EC cells dramatically altered transcriptional profiles. MAXH28R-derived xenografts statistically significantly increased vascular area compared with MAXWT and empty vector tumors (P = .003 and P = .008, respectively). MAXH28R-expressing EC cells secreted nearly double the levels of VEGFA compared with MAXWT cells (P = .03, .005, and .005 at 24, 48, and 72 hours, respectively), and conditioned media from MAXH28R cells increased sprouting when applied to HUVECs. Conclusion These data highlight the importance of MAX mutations in EC and point to increased vascularity as one mechanism contributing to clinical aggressiveness of EC. Endometrial cancer (EC) is the most common gynecologic cancer in the United States and one of a few cancer types for which both incidence and mortality are increasing (1,2). Endometrioid endometrial cancer (EEC) is the most common histologic subtype, accounting for approximately 85% of cases (3). Genetically, EEC is a highly mutated tumor type. Many driver mutations have been identified; however, our present understanding of gene defects that contribute to specific cancer processes such as invasion, metastasis, and angiogenesis is limited. Based on recent findings from The Cancer Genome Atlas (TCGA) and subsequent in silico analyses, the transcription factor MYC-associated factor X (MAX) emerged as a new genetic factor likely to influence EC tumor biology (4–6). MAX is the obligate binding partner of MYC, a master transcription factor with pro-proliferation, pro-growth, and oncogenic functions. MYC has low binding affinity for DNA, but MYC:MAX heterodimers bind E-box sequences to regulate gene expression (7,8). MAX plays an equally important role as a binding partner for members of the MAX dimerization proteins (MXDs) family including MNT, MXD1,3-4, and MGA. MAX:MXD family dimers oppose the pro-growth effects of MYC by promoting expression of differentiation and quiescence genes (9–11). Deregulation of MYC family members is seen in a variety of tumors (12). MAX abnormalities, however, are rare, with the notable exceptions of loss-of-function germline and somatic variants in pheochromocytoma and paraganglioma patients, and somatic loss-of-function mutations in small cell lung cancer (13–15). The missense MAX mutations reported by TCGA in ECs stand in sharp contrast to these loss-of-function MAX variants (4). One particular hotspot mutation discovered by TCGA, the MAX p.His28Arg mutation, has not been previously reported in primary specimens from other cancer types to date. We undertook studies to investigate MAX’s role in EC, particularly the functional consequences of the p.His28Arg mutation. Methods Patient Materials Uterine cancer samples were collected by the Division of Gynecologic Oncology, Washington University, St. Louis, from 1991 to 2010. Written informed consent for this study was obtained and approved by Washington University protocols HSC 91-0507 and HSC 93-0828 and Ohio State University protocol 2012C0116. Microsatellite instability (MSI) status and POLE mutation testing were performed previously (16–18). Nonendometrioid and POLE-mutated ECs were excluded. Targeted Sequencing All coding exons in MAX long and short isoforms (NM002382 and NM145112, respectively) were sequenced to an average of 170× in 509 EECs using the TruSeq Custom Amplicon Kit v1.5 and a MiSeq instrument with Reagent Kit v2 (Illumina, San Diego, CA). Variants were identified using Miseq Reporter software (v2.5.1) (19,20). Electrophoretic Mobility Shift Assays Oligonucleotide sequences used in electrophoretic mobility shift assays (EMSAs) are given in Supplementary Table 1 (available online). In vitro translated (IVT) protein or nuclear lysates were incubated at 42 °C for 10 minutes, followed by incubation with the indicated antibodies (Santa Cruz Biotechnology, Dallas, TX: c-MYC [N-262, sc-764] and MAX [C-17, sc-197]) at 25 °C for 15 minutes, then addition of oligonucleotides and incubation at 25 °C for 20 minutes. Gel shift assays were performed with the LightShift Chemiluminescent EMSA kit (Thermo Fisher Scientific, Waltham, MA). Results are representative of three experiments. Additional details are provided in the Supplementary Methods (available online). Cell Culture 293T and AN3CA (parental and MAX-expressing) cells were cultured in DMEM (Sigma Aldrich, St. Louis, MO) with 10% fetal bovine serum (FBS). RL95-2 and Ishikawa (parental and MAX-expressing) cells were cultured in F12:DMEM (Life Technologies, Carlsbad, CA) with 10% FBS. Human umbilical vein endothelial cells (HUVECs) were cultured in endothelial growth medium 2 (Lonza, Basel, Switzerland). Cell lines were confirmed mycoplasma negative. The estimated number of passages between authentication of cancer cell lines and completion of experiments is 25. Additional details regarding cell line origin, authentication, and forced expression of MAX are provided in the Supplementary Methods (available online). Conditioned Media Preparation and VEGFA Enzyme-Linked Immunosorbent Assay Cells were plated in duplicate in six-well plates at 1 × 106 cells/well (AN3CA-derived cell lines) or 3.5 × 106 cells/well (Ishikawa- and RL95-2-derived cell lines). After cells were attached, media was changed then collected either 24, 48, or 72 hours later. Media was centrifuged at 211 g (4 °C) for 10 minutes. Supernatants were aliquoted and snap frozen on dry ice. Cells were counted to normalize VEGFA protein quantification. VEGFA levels were measured using the Human VEGF Enzyme-Linked Immunosorbent Assay (ELISA) Kit (ThermoFisher Scientific). Absorbance at 450 nm was measured, and a four-parameter logistic standard curve was used to determine concentrations. Biologic duplicates were assayed in technical duplicate. Microfluidic Sprouting Assay Microfluidic HUVEC sprouting assays were performed as described previously (21). Conditioned media from MAXWT, MAXH28R, or empty vector (EV) AN3CA cells was introduced into HUVEC-lined microchannels, and HUVEC sprouting into a central collagen channel was observed. Images of HUVEC sprouting were acquired immediately before and 24 hours after treatment with conditioned medium. Images were processed and analyzed with ImageJ software to calculate the normalized sprouting ratio for each aperture. Additional details are provided in the Supplementary Methods (available online). In Vivo Tumor Models Six- to eight-week-old outbred, NCr-nu/nu females were utilized for flank (six to 11 mice per group) and intraperitoneal (two to three mice per group) xenograft studies. For flank xenografts, researchers blinded to tumor genotypes measured tumors twice weekly. Animal studies are covered by the Ohio State University IACUC Protocols No. A201300000141 and No. 2012A00000008-R1. Additional details are provided in the Supplementary Methods (available online). Immunohistochemistry and Xenograft Vessel Quantification Immunohistochemistry was performed using a Bond Rx autostainer (Leica, Wetzlar, Germany). Automated dewaxing, rehydration, antigen retrieval, blocking, primary antibody incubation, postprimary antibody incubation, detection, and counterstaining were performed using Bond reagents (Leica). Pan-endothelial cell antigen (MECA-32; 1:200, BD Pharmingen, 550563, San Diego, CA) with rabbit antirat IgG (1:200, Vector Laboratories, AI-4001, Burlingame, CA) and PECAM-1 (CD31; 1:1000, Santa Cruz Biotechnology, sc1506R) antibodies were used with Bond Polymer Refine Detection (Leica, S9800). Images were captured using the Vectra Intelligent Slide Analysis System (PerkinElmer, Waltham, MA). Quantification of vessel area was performed on MECA-32 stained slides for six random 20× images per xenograft using inform 2.1 (PerkinElmer) and imageJ. Additional information on sample preparation is provided in the Supplementary Methods (available online). Statistical Analysis Associations between MAX mutation status and MSI, stage, recurrence/progression, age, lymphovascular space invasion, race, and adjuvant therapy were calculated using two-sided Fisher's exact tests. P values for body mass index (BMI) and grade were calculated using two-sided chi-square tests. For univariate survival analysis, P value was determined by log-rank test, and hazard ratio (HR) and 95% confidence interval (CI) were obtained using the Mantel Haenszel approach. Cox proportional hazards multivariable analysis was performed using MAX mutation status and features known to be associated with outcome (18,22). The proportional hazards assumption was tested using the scaled Schoenfeld residuals and the Kaplan-Meier transformed survival times. None of the variables in the multivariable analysis was statistically significant (P> .05), indicating a lack of evidence for departure from the proportional hazards assumption. Ten patients with perioperative deaths (<30 days) and 24 with persistent (stage IVB) disease were excluded from analysis. T tests were used to determine the statistical significance of differences observed in chromatin immunoprecipitation (ChIP)-quantitative polymerase chain reaction (qPCR), quantitative reverse transcription polymerase chain reaction (qRT-PCR), and ELISA experiments. T tests with Welch’s correction when appropriate were used in the TCGA RNAseq data analysis. One-way analysis of variance (ANOVA) with Bonferroni’s multiple comparison test was used in the MECA-32+ area analysis. HUVEC sprouting assay was analyzed using one-way ANOVA with Tukey’s-Cramer post-test. P values of less than .05 were considered statistically significant. All statistical tests were two-sided. All calculations were performed using SPSS 22 (IBM, Armonk, NY), Prism 5 (GraphPad, La Jolla, CA), R Survival Package, and JMP (SAS Institute, Cary NC). Additional methods for mutation modeling, immunoprecipitation, microarray expression profiling, MAX and MYC genic occupancy, TCGA RNA-seq analysis, qRT-PCR, Sanger sequencing, TA cloning, and in vitro cell-based assays are provided in the Supplementary Methods (available online). Results MAX Mutations in EEC and Associations With Clinicopathologic Features Including Outcome MAX was sequenced in 509 primary EECs (18,22,23). The overall mutation rate was 4.3% (22 tumors with mutations). The p.His28Arg mutation was observed in 12 tumors, along with nine additional somatic variants in 11 tumors (Figure 1A). The hotspot p.His28Arg mutation was also observed by TCGA (Figure 1A) (4). Variant allele fractions (VAFs) were consistent with full clonality (Supplementary Table 2, available online). cDNA analysis for primary tumors showed expression of both wild-type (WT) and mutant alleles (corresponding to p.His28Arg), and immunoblot analysis showed comparable levels of MAX expression in wild-type and mutant tumors (data not shown). One tumor (specimen 2219) had two somatic mutations, p.His28Arg and the known cancer susceptibility allele p.Arg75*, which occurred in trans (Supplementary Figure 1, available online) (13). Two tumors (specimens 1122 and 1913) had apparent gain or loss of the MAX locus, as evidenced by single nucleotide polymorphism VAFs (Supplementary Table 2, available online). Figure 1. View largeDownload slide MAX mutations in endometrioid endometrial carcinoma (EEC). A) Schematic of MAX mutations identified in this study of 509 EEC samples (above) and by The Cancer Genome Atlas (below) (4), shown on the long isoform (160 amino acids). The hotspot p.His28Arg mutation seen in 12 tumors and the p.Arg60Gln mutation seen in three tumors both map to the helix-loop-helix domain (green). B) Kaplan-Meier plots show statistically significantly reduced recurrence-free survival (RFS) for women with MAX-mutant tumors. P value was determined by log-rank test. Hazard ratio and 95% confidence interval were obtained using the Mantel Haenszel approach. The survival curve was truncated at eight years. C) Multivariable analysis for RFS includes clinical variables commonly associated with outcome and MAX mutation status. P values and hazard ratios were calculated using Cox proportional hazards model. Variables included in model are those frequently shown to be prognostic in univariate analyses. Statistical significance was calculated using multivariable Cox proportional hazard tests. P values are two-sided. All statistical tests were two-sided. CI = confidence interval; HR = hazard ratio. Figure 1. View largeDownload slide MAX mutations in endometrioid endometrial carcinoma (EEC). A) Schematic of MAX mutations identified in this study of 509 EEC samples (above) and by The Cancer Genome Atlas (below) (4), shown on the long isoform (160 amino acids). The hotspot p.His28Arg mutation seen in 12 tumors and the p.Arg60Gln mutation seen in three tumors both map to the helix-loop-helix domain (green). B) Kaplan-Meier plots show statistically significantly reduced recurrence-free survival (RFS) for women with MAX-mutant tumors. P value was determined by log-rank test. Hazard ratio and 95% confidence interval were obtained using the Mantel Haenszel approach. The survival curve was truncated at eight years. C) Multivariable analysis for RFS includes clinical variables commonly associated with outcome and MAX mutation status. P values and hazard ratios were calculated using Cox proportional hazards model. Variables included in model are those frequently shown to be prognostic in univariate analyses. Statistical significance was calculated using multivariable Cox proportional hazard tests. P values are two-sided. All statistical tests were two-sided. CI = confidence interval; HR = hazard ratio. Patients with MAX-mutant tumors had reduced recurrence-free survival (RFS; HR = 4.00, 95% CI = 1.15 to 13.91, P = .03) (Figure 1B). Mutation was statistically significantly associated with microsatellite instability, tumor stage, and patient BMI (Supplementary Table 3, available online). MAX mutation status remained an independent predictor of reduced recurrence-free survival in a multivariable model that included age, grade, stage, and lymphovascular space invasion (HR = 2.95, 95% CI = 1.20 to 7.29, P = .02), suggesting that MAX defects contribute to clinical aggressiveness (Figure 1C). MAX p.His28Arg DNA Binding Histidine 28 makes a critical contact with the E-box sequence (5’-CACGTG-3’) by hydrogen bonding between NE2 and N7 and/or O6 of Gua(3’) (24). Based on in silico prediction, the His28Arg substitution likely increases affinity for Gua(3’) due to reduced bond length and/or bidentate interaction (Figure 2A). We therefore assessed DNA binding by EMSAs with IVT proteins. As MAX short isoform homodimers bind E-boxes poorly, we assayed long isoform MAXWT and MAXH28R using a canonical two-E-box sequence from the NPM1 promoter (25,26). MAXH28R had greater affinity compared with MAXWT and the supershifted complexes electrophoresed at different rates, consistent with altered protein:DNA interaction (Figure 2B;Supplementary Figure 2A, available online). This was confirmed using AN3CA EC cells overexpressing MAXWT or MAXH28R long isoform, and 293T cells co-expressing both isoforms of MAXWT or MAXH28R (Supplementary Figure 2, B and C, available online). The protein:DNA complex is likely a result of MAXH28R homodimers as co-incubation with an anti-cMYC antibody did not result in a supershift (Figure 2C;Supplementary Figure 2C, available online). Because protein levels of MAXH28R and MAXWT were similar (Figure 2, B and C), the observed results were not due to differences in MAX protein quantity. EMSA findings were similar using a previously studied four-E-box oligonucleotide from the CDK4 promoter (Supplementary Figure 2D) (27). Single E-box oligonucleotides, however, showed no differences between MAXH28R and MAXWT, suggesting that multimeric MAX complexes could account for the differences seen with four- and two-E-box oligonucleotides (data not shown). MAX short isoform homodimers bound E-boxes poorly (Supplementary Figure 2E, available online). When c-MYC was co-expressed with MAXH28R, the MAX shift disappeared, indicating that the p.His28Arg substitution does not hinder MAX:MYC interaction (the MAX:MYC:DNA interaction was not detectable under the conditions used) (Figure 2C). Immunoprecipitation and immunoblot demonstrated exogenous MAXWT, and MAXH28R interacted with endogenous c-MYC and endogenous MAX (Figure 2D). Figure 2. View largeDownload slide p.His28Arg alters MAX E-box binding and transcription. A) MAX/MYC heterodimer bound to a canonical E-box. MAXH28R is predicted to place η2 of the guanidine group in closer proximity to N7 of Gua (3’) (green dotted line) and could allow for bidentate interaction (purple dotted lines). B) Electrophoretic mobility shift assay (EMSA) shows differential binding of MAXH28R to NPM1 E-boxes. In vitro–translated MAXH28R protein created a shift (lane 5), whereas MAXWT did not (lane 3). Anti-MAX antibody supershifted the MAXH28R band (lane 6) and produced a different supershift with MAXWT (lane 4) (see also Supplementary Figure 2A, available online). C) EMSA with MAX-transfected AN3CA cell nuclear extracts points to MAX homodimer:DNA interaction. Anti-cMYC antibody did not supershift the MAXH28R band (lanes 1 and 3), and cotransfection of cMYC eliminated the shift band, consistent with reduced MAX homodimer levels due to increased MAX:cMYC heterodimerization. Immunoblot (right) shows the relative amounts of MAX and cMYC. D) Immunoprecipitation demonstrating myc-tagged-MAXH28R interaction with cMYC and endogenous MAX. E) Chromatin immunoprecipitation (ChIP)–quantitative polymerase chain reaction (qPCR) on AN3CA cells stably expressing MAXH28R and MAXWT (see Supplementary Figure 3, available online) for E-box regions in NPM1, CDK4, and CAD, and control regions in CTCF and KLHL13.F) Volcano plot (middle) of expression differences between patients with MAX-mutated (n = 8) and MAX-WT (n = 154) tumors from The Cancer Genome Atlas (3). Dot plots in side panels show biologic validation of GATM and SERPINB9 expression determined by qRT-PCR on MAXH28R-mutated tumors (n = 11) and matched MAXWT tumors (n = 24) from our series. Bar graphs in side panels show ChIP-qPCR performed as in (E). ChIP-qPCR and qRT-PCR experiments in (E and F) are shown as mean (SD). P values calculated by two-sided t tests; EMSAs and blots shown are representative of three experiments. ChIPs performed on two independent immunoprecipitations per cell line. EV = empty vector; WT = wild-type. Figure 2. View largeDownload slide p.His28Arg alters MAX E-box binding and transcription. A) MAX/MYC heterodimer bound to a canonical E-box. MAXH28R is predicted to place η2 of the guanidine group in closer proximity to N7 of Gua (3’) (green dotted line) and could allow for bidentate interaction (purple dotted lines). B) Electrophoretic mobility shift assay (EMSA) shows differential binding of MAXH28R to NPM1 E-boxes. In vitro–translated MAXH28R protein created a shift (lane 5), whereas MAXWT did not (lane 3). Anti-MAX antibody supershifted the MAXH28R band (lane 6) and produced a different supershift with MAXWT (lane 4) (see also Supplementary Figure 2A, available online). C) EMSA with MAX-transfected AN3CA cell nuclear extracts points to MAX homodimer:DNA interaction. Anti-cMYC antibody did not supershift the MAXH28R band (lanes 1 and 3), and cotransfection of cMYC eliminated the shift band, consistent with reduced MAX homodimer levels due to increased MAX:cMYC heterodimerization. Immunoblot (right) shows the relative amounts of MAX and cMYC. D) Immunoprecipitation demonstrating myc-tagged-MAXH28R interaction with cMYC and endogenous MAX. E) Chromatin immunoprecipitation (ChIP)–quantitative polymerase chain reaction (qPCR) on AN3CA cells stably expressing MAXH28R and MAXWT (see Supplementary Figure 3, available online) for E-box regions in NPM1, CDK4, and CAD, and control regions in CTCF and KLHL13.F) Volcano plot (middle) of expression differences between patients with MAX-mutated (n = 8) and MAX-WT (n = 154) tumors from The Cancer Genome Atlas (3). Dot plots in side panels show biologic validation of GATM and SERPINB9 expression determined by qRT-PCR on MAXH28R-mutated tumors (n = 11) and matched MAXWT tumors (n = 24) from our series. Bar graphs in side panels show ChIP-qPCR performed as in (E). ChIP-qPCR and qRT-PCR experiments in (E and F) are shown as mean (SD). P values calculated by two-sided t tests; EMSAs and blots shown are representative of three experiments. ChIPs performed on two independent immunoprecipitations per cell line. EV = empty vector; WT = wild-type. To further investigate MAXH28R:DNA interaction, we performed ChIP-qPCR using AN3CA cells stably expressing flag-tagged MAXH28R or MAXWT at levels comparable with endogenous MAX (Supplementary Figure 3, available online). We assessed binding to E-boxes in the MAX/MYC target genes NPM1, CDK4, and CAD, and control regions in CTCF and KLHL13 (25,27,28). MAXH28R showed statistically significantly enhanced affinity for the E-boxes in NPM1 and CDK4 compared with MAXWT (P < .001 and P = .04, respectively) (Figure 2E). Together, these findings demonstrate that the p.His28Arg mutation alters DNA binding and could, therefore, alter gene expression. MAX p.His28Arg Transcriptional Changes TCGA RNA-seq data for the eight MAX-mutant and 156 MAX-WT EECs were used to explore differentially expressed genes (DEGs) (Figure 2F) (4). We identified 64 candidate DEGs with P values of less than .01 and greater than twofold expression change (Supplementary Figure 4A, available online). The majority (71%) bound MAX and/or MYC based on ENCODE ChIP-seq data, implying that MAX mutation might directly alter transcription (Supplementary Figure 4B, available online) (29). Eight candidate DEGs were validated in our collection of EECs using qRT-PCR. Expression of GATM and SERPINB9 was statistically significantly different between mutant and WT tumors (P = .02 and P = .03, respectively) (Figure 2F), whereas expression of the other six candidates was not statistically significantly different (Supplementary Figure 4C, available online). ChIP-qPCR for cells stably expressing flag-tagged MAX showed statistically significantly increased MAXH28R binding to both GATM and SERPINB9 E-boxes compared with MAXWT (P = .002 and P = .009, respectively) (Figure 2F). The effect of MAXH28R on global gene expression was assessed by whole transcriptome profiling using AN3CA cell lines stably expressing exogenous short or long isoform MAXWT and MAXH28R. MAX protein levels were comparable between stable clones (Supplementary Figure 5A, available online). Principle component analysis proved that biologic duplicates clustered together and MAXH28R-long cell lines were the most different from the other cell types (Supplementary Figure 5B, available online). Both MAXH28R short and long isoform cell lines were clearly different from the cell lines as indicated by hierarchical clustering (Supplementary Figure 5C, available online), which could be a reflection of either direct or indirect effects on transcription. Vascular Phenotype of MAXH28R Xenograft Tumors In vitro assays showed no differences in proliferation, colony formation, or wound healing for the stable MAXH28R- and MAXWT-expressing AN3CA cells (Supplementary Figure 6, A–C, available online). Differences in proliferation rates were evident between MAXH28R- and MAXWT-expressing AN3CA clones but did not reach statistical significance. Similarly, no statistically significant differences in proliferation or colony formation were observed with MAXH28R- and MAXWT-expressing Ishikawa cells (Supplementary Figure 6, A and B, available online). Although the number of colonies did not differ between genotypes in either the AN3CA or Ishikawa cell lines, colony sizes varied. MAXWT-expressing AN3CA colonies were larger than the MAXH28R or empty vector (EV) colonies. On the other hand, MAXH28R-expressing Ishikawa colonies were larger than the MAXWT or EV colonies. Xenograft studies, however, showed striking differences in the flank tumors derived from MAXH28R and MAXWT EC cells. AN3CA MAXH28R tumors had markedly increased vascularity and were much more hemorrhagic than MAXWT tumors (Figure 3A;Supplementary Figure 7, available online). Pan-endothelial cell antigen PVLAP (MECA-32) staining proved that the AN3CA MAXH28R tumors had approximately twice the vascular area compared with MAXWT and EV (P = .003 and P = .008, respectively) (Figure 3, B and C). Similarly, tumors derived from MAXH28R-expressing Ishikawa cells also had statistically significantly increased vascular area compared with MAXWT and EV (P = .03 and P < .001, respectively) (Supplementary Figure 8, available online). Figure 3. View largeDownload slide Assessment of vascularity in AN3CA xenografts. A) Representative flank xenografts derived from MAXWT (n = 22 mice), MAXH28R (n = 22), and empty vector (EV; n = 11) cells. Additional images for xenografts are provided in Supplementary Figure 7 (available online). Scale bars = 5 mm. B) Immunohistochemical staining for endothelial cells in xenografts. Representative images for MECA-32 pan-endothelial marker stained tissues. Scale bars = 100 μm. C) Quantification of MECA-32+ area (percent of total area of image). Three random tumors assessed per condition. Six 20× images quantified per tumor. Data shown as mean ± SD. P values calculated using one-way analysis of variance (P < .01) with Bonferroni multiple comparison test are shown. D) Representative images (left) of intraperitoneal AN3CA xenografts derived from MAXWT clone 5 (n = 3) and MAXH28R clone 2 (n = 2) (see Supplementary Figure 3, available online). Scale bars = 5 mm. Representative images of immunohistochemical staining for endothelial cells by MECA-32 (middle) and CD31 (right) in adjacent sections. Scale bars = 100 μm. EV = empty vector; WT = wild-type. Figure 3. View largeDownload slide Assessment of vascularity in AN3CA xenografts. A) Representative flank xenografts derived from MAXWT (n = 22 mice), MAXH28R (n = 22), and empty vector (EV; n = 11) cells. Additional images for xenografts are provided in Supplementary Figure 7 (available online). Scale bars = 5 mm. B) Immunohistochemical staining for endothelial cells in xenografts. Representative images for MECA-32 pan-endothelial marker stained tissues. Scale bars = 100 μm. C) Quantification of MECA-32+ area (percent of total area of image). Three random tumors assessed per condition. Six 20× images quantified per tumor. Data shown as mean ± SD. P values calculated using one-way analysis of variance (P < .01) with Bonferroni multiple comparison test are shown. D) Representative images (left) of intraperitoneal AN3CA xenografts derived from MAXWT clone 5 (n = 3) and MAXH28R clone 2 (n = 2) (see Supplementary Figure 3, available online). Scale bars = 5 mm. Representative images of immunohistochemical staining for endothelial cells by MECA-32 (middle) and CD31 (right) in adjacent sections. Scale bars = 100 μm. EV = empty vector; WT = wild-type. In a separate intraperitoneal (IP) AN3CA xenograft model, MAXH28R-expressing tumors were markedly more hemorrhagic than MAXWT tumors and had increased vascularization (Figure 3D;Supplementary Figure 9, available online). Immunohistochemistry for PECAM-1 (CD31), another endothelial cell marker, displayed similar staining compared with MECA-32, confirming our earlier results (Figure 3D;Supplementary Figure 10, available online). The tumor volumes for the MAXH28R and MAXWT flank xenografts were similar (Supplementary Figure 11, available online), making it unlikely that tumor size accounted for the vascularity differences. MAXH28R Pro-angiogenic Paracrine Signaling To test the hypothesis that paracrine factors elaborated by MAXH28R-expressing cells contribute to the vascular phenotype in xenografts, we assessed HUVEC sprouting in vitro using a previously described microfluidic model (21). We observed statistically significantly increased HUVEC sprouting with MAXH28R conditioned medium compared with MAXWT and EV (P = .02 and P = .005, respectively) (Figure 4A). As the expression array data (Supplementary Figure 5D, available online) identified VEGFA as a candidate pro-angiogenic factor increased by MAXH28R, VEGFA levels in conditioned media from the EV, MAXWT, and MAXH28R EC cells were determined by ELISA. AN3CA MAXH28R conditioned media had nearly double the concentration of VEGFA as MAXWT at all assessed time points (P = .03, .005, and .005 at 24, 48, and 72 hours, respectively) (Figure 4B). Secreted VEGFA was also nearly doubled in Ishikawa and RL95-2 EC cell lines stably expressing MAXH28R compared with MAXWT (Supplementary Figure 12, available online). Taken together, these data demonstrate that MAXH28R in cancer cell lines promotes increased vascularity and that increased VEGFA secretion could contribute to the pro-angiogenic phenotype. Figure 4. View largeDownload slide Differences in HUVEC sprouting and VEGFA secretion in vitro. A) Representative images (top) of HUVEC sprouting through apertures in the polydimethylsiloxane (PDMS) microdevice into 3D collagen gel with conditioned media from AN3CA empty vector (EV) cells and cells stably expressing MAXWT (clone 7) or MAXH28R (clone 2). Scale bar = 50 μm. Quantification (bottom) of HUVEC sprouting. Bar graphs represent mean (SD) for three independent microdevice experiments per condition. One-way analysis of variance for statistical significance was performed (P < .01), and Tukey’s post hoc pair-wise test for statistical significance is shown. B) Quantification of VEGFA by enzyme-linked immunosorbent assay in media conditioned by AN3CA EV cells and cells stably expressing MAXWT or MAXH28R (see Supplementary Figure 3, available online). Bar graphs represent duplicate experiments, shown as mean ± SD. Statistical significance from the two-sided t test. EV = empty vector; WT = wild-type. Figure 4. View largeDownload slide Differences in HUVEC sprouting and VEGFA secretion in vitro. A) Representative images (top) of HUVEC sprouting through apertures in the polydimethylsiloxane (PDMS) microdevice into 3D collagen gel with conditioned media from AN3CA empty vector (EV) cells and cells stably expressing MAXWT (clone 7) or MAXH28R (clone 2). Scale bar = 50 μm. Quantification (bottom) of HUVEC sprouting. Bar graphs represent mean (SD) for three independent microdevice experiments per condition. One-way analysis of variance for statistical significance was performed (P < .01), and Tukey’s post hoc pair-wise test for statistical significance is shown. B) Quantification of VEGFA by enzyme-linked immunosorbent assay in media conditioned by AN3CA EV cells and cells stably expressing MAXWT or MAXH28R (see Supplementary Figure 3, available online). Bar graphs represent duplicate experiments, shown as mean ± SD. Statistical significance from the two-sided t test. EV = empty vector; WT = wild-type. Discussion Determining how specific and presumably biologically relevant gene defects contribute to cancer phenotypes remains a challenge. Our functional characterization of the recurrent and EC-specific MAX p.His28Arg mutation provides important evidence that the mutation alters MAX function and contributes to a cancer-related phenotype, angiogenesis. The hotspot EC MAX mutation p.His28Arg has altered DNA binding and is associated with marked changes in transcription. EMSAs and ChIP assays showed that MAXH28R had increased E-box binding compared with MAXWT, which is consistent with structural modeling that predicted that the p.His28Arg substitution increases MAX’s affinity for DNA. Our xenograft studies implicated MAXH28R as playing a role in angiogenesis, and the in vitro angiogenesis assays validated the pro-angiogenic effect seen for MAXH28R and pointed to a role for secreted factors. Our demonstration that conditioned media from three different EC cell lines expressing MAXH28R had elevated VEGFA levels further supports a model for secreted factors mediating MAXH28R-associated pro-angiogenesis. MAX-mediated increases in vascularity and VEGFA levels, both of which are associated with poor clinical outcome in EC patients, could in part explain selection for MAX mutation in EC (30–32). The mutation pattern of MAX that we observed strongly suggests a dominant or dominant-negative activity in EC (Figure 1A), consistent with our functional data. The paucity of stop and frameshift mutations and absence of second-hit mutations in EEC is in stark contrast to the loss-of-function abnormalities seen in neuroendocrine tumors, suggesting context-dependent roles for MAX (13–15). Two amino acids in MAX appear to display substitutions that are specific to cancer type and that may have different functional consequences. Val9Met and Arg60Gln are observed in EECs, while Val9Leu and Arg60Trp are observed in pheochromocytomas (4,14). The two pheochromocytoma missense variants have been shown to be loss-of-function in terms of ability to repress MYC-driven expression of a reporter assay in rat PC12 cells null for MAX (33). Although Val9Met and Arg60Gln are yet to be functionally tested in EC, the possibility exists that they are loss-of-function alleles, similar to the corresponding pheochromocytoma variants. However, the effect of these variants must be experimentally determined in additional systems given the complexity of MAX’s transcription regulatory function. There are several examples of context-dependent roles for cancer genes. FGFR2 is an oncogene in EC but acts as a tumor suppressor in melanoma (34,35). Loss- and gain-of-function defects occur in TP53; tumors that have an inherited loss-of-function mutation and a second somatic mutation point to p53’s role as a tumor suppressor, but oncogenic missense changes that alter signaling cascades and chromatin modification also drive cancer in some instances (36–39). Most TP53 mutations abrogate p53’s tumor suppressive function, acting either as cellular recessives or dominant-negatives (38). Gain-of-function TP53 missense mutations in the DNA binding domain cause loss of wild-type p53 tumor suppressive function, alter genomic binding, and confer novel oncogenic activity (36–39). MAX defects may parallel the complex nature of TP53 mutations. Clear loss-of-function deletion and nonsense alleles exist for MAX, in addition to potentially more biologically complex missense mutations involving the DNA binding domain. Some mutations could be tissue-dependent loss-of-function alleles for certain activities, while at the same time conferring novel oncogenic properties. Our finding that a recurrent cancer-specific mutation in MAX acts in a dominant manner is not unexpected. An endogenous truncated MAX protein called dMAX is a naturally occurring example of a dominant-negative MAX. dMAX can dimerize with MYC and MXD members but cannot associate with DNA, and it has been associated with increased aerobic metabolism (40–42). MAX and MAX-like-protein X (MLX) are known to regulate transcriptional networks essential to tumor cell metabolism/nutrient availability, and it is possible that the increased vascularity seen in xenografts expressing MAXH28R is part of a central theme of MAX’s important role in coordination of nutrient availability (43). Further, the effect of the p.His28Arg mutation in DNA binding is likely to influence activities of many MAX-interacting proteins, not just its canonical partner MYC. This study is limited by characterization of a single hotspot mutation in the MAX gene. Further efforts will be required to determine whether the other observed MAX mutations alter DNA binding and/or increase angiogenesis. Although our study found that MAX mutation in EC patients was associated with poor outcome, analysis of additional cohorts will be required to validate the prognostic significance. Our results link the MAX p.His28Arg mutation and tumor angiogenesis. The global effect on transcriptional profiles observed in cells expressing MAXH28R also implies there are additional gene expression changes that could contribute to biologic aggressiveness in EEC. Taken together, the findings from our mutation and functionalization studies implicate mutant MAX as a driver of aggressive EEC. Funding This study was funded by the National Institutes of Health (R21 CA155674 to P. J. Goodfellow), the National Cancer Institute (P30 CA016058 supporting the Genomics and Biostatistics shared resources at the Ohio State University Comprehensive Cancer Center), the National Institute of General Medical Sciences (T32 GM068412 to C. M. Rush), the Pelotonia Fellowship Program (C. J. Walker), The American Heart Association (15SDG25480000 to J. W. Song), and The American Cancer Society (IRG-67-003-50 to J. W. Song). Notes The study funders had no role in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication. The authors have declared that no conflict of interest exists. PJG and PD devised the initial concept of the study. DGM provided primary tumor specimens. CJW, MJO, and MS performed mutation screening and clinical correlation studies. CMR, CJW, and PJG performed predictive mutation modeling. CJW and PD performed electrophoretic mobility shift assays. MJO and CJW performed co-immunoprecipitations. CMR performed chromatin immunoprecipitation. CJW and CMR were responsible for TCGA data analysis and other bioinformatics analyses. CJW, PD, and CMR generated stable cell lines, RS, KS, PD, CMR, BS, CJW, RAZ, CMC, JLG, and MJO performed in vivo studies, including cell line implantation and animal husbandry (RS), harvesting (CMR, RS, CMC, RAZ, JLG, and BS), and vessel quantification (CJW, CMR, and MJO). Cell phenotype assessment was performed by PD, CMR, and CJW, including cell growth (PD and CMR), colony forming ability (CMR), and migratory potential (CJW). Angiogenic sprouting assays were performed by CWC and JWS. Enzyme-linked immunosorbent assays were performed by CMR. DGM and DEC provided indispensable guidance in study design and implementation. CJW, CMR, PD, and PJG wrote the manuscript with input from other authors. The final manuscript was approved by all authors. We would like to thank Mark Foster for assistance with mutation modeling and interpretation of chemical interactions, and we thank Qianben Wang and Benjamin Sunkel for their assistance with chromatin immunoprecipitation–quantitative polymerase chain reaction experiments. We also thank Alexis Chassen for manuscript editing and data curation. We thank Joseph McElroy for bioinformatics/statistical counsel. We acknowledge Alex Seibel for assisting with the sprouting measurements. We would like to acknowledge Raleigh Kladney, Cynthia Timmers, and the Solid Tumor Translational Science Core Facility. We acknowledge the Target Validation Core Facility. We would like to acknowledge Pearlly Yan, David Symer, and Sarah Warner with the Ohio State University Genomics Core Facility, and Tea Meulia with the Ohio State University Molecular and Cellular Imagine Center, a CFAES/OARDX core facility in Wooster, Ohio. We are very grateful to all of the patients who contributed specimens to this study and all of the attending physicians and staff at the Washington University School of Medicine Division of Gynecologic Oncology, as well as The Ohio State University College of Medicine Division of Gynecologic Oncology. References 1 Siegel RL , Miller KD , Jemal A. Cancer statistics , 2017 . CA Cancer J Clin . 2017; 67 ( 1 ): 7 – 30 . http://dx.doi.org/10.3322/caac.21387 Google Scholar CrossRef Search ADS PubMed 2 Morice P , Leary A , Creutzberg C , Abu-Rustum N , Darai E. Endometrial cancer . Lancet. 2016 ; 387 ( 10023 ): 1094 – 1108 . Google Scholar CrossRef Search ADS PubMed 3 Creasman WT , Odicino F , Maisonneuve P et al. , Carcinoma of the corpus uteri. FIGO 26th annual report on the results of treatment in gynecological cancer . Int J Gynaecol Obstet . 2006 ; 95 (suppl 1): S105 – S143 . Google Scholar CrossRef Search ADS 4 Cancer Genome Atlas Research Network , Kandoth C , Schultz N et al. , Integrated genomic characterization of endometrial carcinoma . Nature . 2013 ; 497 ( 7447 ): 67 – 73 . http://dx.doi.org/10.1038/nature12113 Google Scholar CrossRef Search ADS PubMed 5 Chang MT , Asthana S , Gao SP et al. , Identifying recurrent mutations in cancer reveals widespread lineage diversity and mutational specificity . Nat Biotechnol . 2016 ; 34 ( 2 ): 155 – 163 . Google Scholar CrossRef Search ADS PubMed 6 Kamburov A , Lawrence MS , Polak P et al. , Comprehensive assessment of cancer missense mutation clustering in protein structures . Proc Natl Acad Sci U S A . 2015 ; 112 ( 40 ): E5486 – E5495 . Google Scholar CrossRef Search ADS PubMed 7 Atchley WR , Fitch WM. Myc and Max: Molecular evolution of a family of proto-oncogene products and their dimerization partner . Proc Natl Acad Sci U S A. 1995 ; 92 ( 22 ): 10217 – 10221 . http://dx.doi.org/10.1073/pnas.92.22.10217 Google Scholar CrossRef Search ADS PubMed 8 Meyer N , Penn LZ. Reflecting on 25 years with MYC . Nat Rev Cancer . 2008 ; 8 ( 12 ): 976 – 990 . http://dx.doi.org/10.1038/nrc2231 Google Scholar CrossRef Search ADS PubMed 9 Grinberg AV , Hu CD , Kerppola TK. Visualization of Myc/Max/Mad family dimers and the competition for dimerization in living cells . Mol Cell Biol. 2004 ; 24 ( 10 ): 4294 – 4308 . http://dx.doi.org/10.1128/MCB.24.10.4294-4308.2004 Google Scholar CrossRef Search ADS PubMed 10 Hurlin PJ , Queva C , Eisenman RN. Mnt, a novel Max-interacting protein is coexpressed with Myc in proliferating cells and mediates repression at Myc binding sites . Genes Dev. 1997 ; 11 ( 1 ): 44 – 458 . http://dx.doi.org/10.1101/gad.11.1.44 Google Scholar CrossRef Search ADS PubMed 11 Walker W , Zhou ZQ , Ota S , Wynshaw-Boris A , Hurlin PJ. Mnt-Max to Myc-Max complex switching regulates cell cycle entry . J Cell Biol. 2005 ; 169 ( 3 ): 405 – 413 . Google Scholar CrossRef Search ADS PubMed 12 Tansey WP. Mammalian MYC proteins and cancer . N J Sci. 2014 ; 2014 : 1 – 27 . http://dx.doi.org/10.1155/2014/757534 Google Scholar CrossRef Search ADS 13 Comino-Mendez I , Gracia-Aznarez FJ , Schiavi F et al. , Exome sequencing identifies MAX mutations as a cause of hereditary pheochromocytoma . Nat Genet . 2011 ; 43 ( 7 ): 663 – 667 . Google Scholar CrossRef Search ADS PubMed 14 Burnichon N , Cascon A , Schiavi F et al. , MAX mutations cause hereditary and sporadic pheochromocytoma and paraganglioma . Clin Cancer Res. 2012 ; 18 ( 10 ): 2828 – 2837 . http://dx.doi.org/10.1158/1078-0432.CCR-12-0160 Google Scholar CrossRef Search ADS PubMed 15 Romero OA , Torres-Diz M , Pros E et al. , MAX inactivation in small cell lung cancer disrupts MYC-SWI/SNF programs and is synthetic lethal with BRG1 . Cancer Discov . 2014 ; 4 ( 3 ): 292 – 303 . http://dx.doi.org/10.1158/2159-8290.CD-13-0799 Google Scholar CrossRef Search ADS PubMed 16 Zighelboim I , Goodfellow PJ , Gao F et al. , Microsatellite instability and epigenetic inactivation of MLH1 and outcome of patients with endometrial carcinomas of the endometrioid type . J Clin Oncol. 2007 ; 25 ( 15 ): 2042 – 2048 . http://dx.doi.org/10.1200/JCO.2006.08.2107 Google Scholar CrossRef Search ADS PubMed 17 Zighelboim I , Schmidt AP , Gao F et al. , ATR mutation in endometrioid endometrial cancer is associated with poor clinical outcomes . J Clin Oncol. 2009 ; 27 ( 19 ): 3091 – 3096 . http://dx.doi.org/10.1200/JCO.2008.19.9802 Google Scholar CrossRef Search ADS PubMed 18 Billingsley CC , Cohn DE , Mutch DG , Stephens JA , Suarez AA , Goodfellow PJ. Polymerase varepsilon (POLE) mutations in endometrial cancer: Clinical outcomes and implications for Lynch syndrome testing . Cancer. 2015 ; 121 ( 3 ): 386 – 394 . Google Scholar CrossRef Search ADS PubMed 19 DePristo MA , Banks E , Poplin R et al. , A framework for variation discovery and genotyping using next-generation DNA sequencing data . Nat Genet. 2011 ; 43 ( 5 ): 491 – 498 . http://dx.doi.org/10.1038/ng.806 Google Scholar CrossRef Search ADS PubMed 20 McKenna A , Hanna M , Banks E et al. , The Genome Analysis Toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010 ; 20 ( 9 ): 1297 – 1303 . http://dx.doi.org/10.1101/gr.107524.110 21 Song JW , Munn LL. Fluid forces control endothelial sprouting . Proc Natl Acad Sci U S A. 2011 ; 108 ( 37 ): 15342 – 15347 . http://dx.doi.org/10.1073/pnas.1105316108 Google Scholar CrossRef Search ADS PubMed 22 Walker CJ , Miranda MA , O'Hern MJ et al. , Patterns of CTCF and ZFHX3 mutation and associated outcomes in endometrial cancer . J Natl Cancer Inst . 2015 ; 107 ( 11 ):djv249. 23 Zighelboim I , Mutch DG , Knapp A et al. , High frequency strand slippage mutations in CTCF in MSI-positive endometrial cancers . Hum Mutat. 2014 ; 35 ( 1 ): 63 – 65 . http://dx.doi.org/10.1002/humu.22463 Google Scholar CrossRef Search ADS PubMed 24 Nair SK , Burley SK. X-ray structures of Myc-Max and Mad-Max recognizing DNA. Molecular bases of regulation by proto-oncogenic transcription factors . Cell. 2003 ; 112 ( 2 ): 193 – 205 . Google Scholar CrossRef Search ADS PubMed 25 Zeller KI , Haggerty TJ , Barrett JF , Guo Q , Wonsey DR , Dang CV. Characterization of nucleophosmin (B23) as a Myc target by scanning chromatin immunoprecipitation . J Biol Chem. 2001 ; 276 ( 51 ): 48285 – 48291 . Google Scholar CrossRef Search ADS PubMed 26 Prochownik EV , VanAntwerp ME. Differential patterns of DNA binding by myc and max proteins . Proc Natl Acad Sci U S A. 1993 ; 90 ( 3 ): 960 – 964 . http://dx.doi.org/10.1073/pnas.90.3.960 Google Scholar CrossRef Search ADS PubMed 27 Hermeking H , Rago C , Schuhmacher M et al. , Identification of CDK4 as a target of c-MYC . Proc Natl Acad Sci U S A. 2000 ; 97 ( 5 ): 2229 – 2234 . http://dx.doi.org/10.1073/pnas.050586197 Google Scholar CrossRef Search ADS PubMed 28 Miltenberger RJ , Sukow KA , Farnham PJ. An E-box-mediated increase in cad transcription at the G1/S-phase boundary is suppressed by inhibitory c-Myc mutants . Mol Cell Biol. 1995 ; 15 ( 5 ): 2527 – 2535 . http://dx.doi.org/10.1128/MCB.15.5.2527 Google Scholar CrossRef Search ADS PubMed 29 Rosenbloom KR , Sloan CA , Malladi VS et al. , ENCODE data in the UCSC Genome Browser: Year 5 update . Nucleic Acids Res. 2013 ; 41 (database issue): D56 – D63 . Google Scholar CrossRef Search ADS PubMed 30 Chen CA , Cheng WF , Lee CN et al. , Cytosol vascular endothelial growth factor in endometrial carcinoma: Correlation with disease-free survival . Gynecol Oncol. 2001 ; 80 ( 2 ): 207 – 212 . http://dx.doi.org/10.1006/gyno.2000.6048 Google Scholar CrossRef Search ADS PubMed 31 Dobrzycka B , Terlikowski SJ , Kowalczuk O , Kulikowski M , Niklinski J. Serum levels of VEGF and VEGF-C in patients with endometrial cancer . Eur Cytokine Netw . 2011 ; 22 ( 1 ): 45 – 51 . Google Scholar PubMed 32 Kamat AA , Merritt WM , Coffey D et al. , Clinical and biological significance of vascular endothelial growth factor in endometrial cancer . Clin Cancer Res . 2007 ; 13 ( 24 ): 7487 – 7495 . http://dx.doi.org/10.1158/1078-0432.CCR-07-1017 Google Scholar CrossRef Search ADS PubMed 33 Comino-Mendez I , Leandro-Garcia LJ , Montoya G et al. , Functional and in silico assessment of MAX variants of unknown significance . J Mol Med (Berl) . 2015 ; 93 ( 11 ): 1247 – 1255 . Google Scholar CrossRef Search ADS PubMed 34 Pollock PM , Gartside MG , Dejeza LC et al. , Frequent activating FGFR2 mutations in endometrial carcinomas parallel germline mutations associated with craniosynostosis and skeletal dysplasia syndromes . Oncogene. 2007 ; 26 ( 50 ): 7158 – 7162 . http://dx.doi.org/10.1038/sj.onc.1210529 Google Scholar CrossRef Search ADS PubMed 35 Gartside MG , Chen H , Ibrahimi OA et al. , Loss-of-function fibroblast growth factor receptor-2 mutations in melanoma . Mol Cancer Res. 2009 ; 7 ( 1 ): 41 – 54 . http://dx.doi.org/10.1158/1541-7786.MCR-08-0021 Google Scholar CrossRef Search ADS PubMed 36 Lang GA , Iwakuma T , Suh YA et al. , Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome . Cell. 2004 ; 119 ( 6 ): 861 – 872 . http://dx.doi.org/10.1016/j.cell.2004.11.006 Google Scholar CrossRef Search ADS PubMed 37 Olive KP , Tuveson DA , Ruhe ZC et al. , Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome . Cell. 2004 ; 119 ( 6 ): 847 – 860 . http://dx.doi.org/10.1016/j.cell.2004.11.004 Google Scholar CrossRef Search ADS PubMed 38 Oren M , Rotter V. Mutant p53 gain-of-function in cancer . Cold Spring Harb Perspect Biol. 2010 ; 2 ( 2 ): a001107 . Google Scholar CrossRef Search ADS PubMed 39 Zhu J , Sammons MA , Donahue G et al. , Gain-of-function p53 mutants co-opt chromatin pathways to drive cancer growth . Nature. 2015 ; 525 ( 7568 ): 206 – 211 . http://dx.doi.org/10.1038/nature15251 Google Scholar CrossRef Search ADS PubMed 40 Makela TP , Koskinen PJ , Vastrik I , Alitalo K. Alternative forms of Max as enhancers or suppressors of Myc-ras cotransformation . Science. 1992 ; 256 ( 5055 ): 373 – 377 . http://dx.doi.org/10.1126/science.256.5055.373 Google Scholar CrossRef Search ADS PubMed 41 FitzGerald MJ , Arsura M , Bellas RE et al. , Differential effects of the widely expressed dMax splice variant of Max on E-box vs initiator element-mediated regulation by c-Myc . Oncogene. 1999 ; 18 ( 15 ): 2489 – 2498 . http://dx.doi.org/10.1038/sj.onc.1202611 Google Scholar CrossRef Search ADS PubMed 42 Babic I , Anderson ES , Tanaka K et al. , EGFR mutation-induced alternative splicing of Max contributes to growth of glycolytic tumors in brain cancer . Cell Metab. 2013 ; 17 ( 6 ): 1000 – 1008 . http://dx.doi.org/10.1016/j.cmet.2013.04.013 Google Scholar CrossRef Search ADS PubMed 43 O'Shea JM , Ayer DE. Coordination of nutrient availability and utilization by MAX- and MLX-centered transcription networks . Cold Spring Harbor Perspect Med . 2013 ; 3 ( 9 ): a014258 . Google Scholar CrossRef Search ADS © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

JNCI: Journal of the National Cancer InstituteOxford University Press

Published: Nov 15, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off