Growth Hormone Upregulates Melanocyte-Inducing Transcription Factor Expression and Activity via JAK2-STAT5 and SRC Signaling in GH Receptor-Positive Human Melanoma

Abstract

Growth hormone (GH) facilitates therapy resistance in the cancers of breast, colon, endometrium, and melanoma. The GH-stimulated pathways responsible for this resistance were identified as suppression of apoptosis, induction of epithelial-to-mesenchymal transition (EMT), and upregulated drug efflux by increased expression of ATP-binding cassette containing multidrug efflux pumps (ABC-transporters). In extremely drug-resistant melanoma, ABC-transporters have also been reported to mediate drug sequestration in intracellular melanosomes, thereby reducing drug efficacy. Melanocyte-inducing transcription factor (MITF) is the master regulator of melanocyte and melanoma cell fate as well as the melanosomal machinery. MITF targets such as the oncogene MET, as well as MITF-mediated processes such as resistance to radiation therapy, are both known to be upregulated by GH. Therefore, we chose to query the direct effects of GH on MITF expression and activity towards conferring chemoresistance in melanoma. Here, we demonstrate that GH significantly upregulates MITF as well as the MITF target genes following treatment with multiple anticancer drug treatments such as chemotherapy, BRAF-inhibitors, as well as tyrosine-kinase inhibitors. GH action also upregulated MITF-regulated processes such as melanogenesis and tyrosinase activity. Significant elevation in MITF and MITF target gene expression was also observed in mouse B16F10 melanoma cells and xenografts in bovine GH transgenic (bGH) mice compared to wild-type littermates. Through pathway inhibitor analysis we identified that both the JAK2-STAT5 and SRC activities were critical for the observed effects. Additionally, a retrospective analysis of gene expression data from GTEx, NCI60, CCLE, and TCGA databases corroborated our observed correlation of MITF function and GH action. Therefore, we present in vitro, in vivo, and in silico evidence which strongly implicates the GH-GHR axis in inducing chemoresistance in human melanoma by driving MITF-regulated and ABC-transporter-mediated drug clearance pathways.

Keywords: JAK; MITF; SRC (sarcoma or c-Src); STAT; chemoresistance; growth hormone (GH); growth hormone receptor (GHR); melanoma.

Figure 1

Changes in gene expression following drug-treatment in human melanoma cells: (A–D) Human melanoma cells SK-MEL-28 were treated with anticancer drugs and RNA expression was compared with respective untreated controls, at 2, 6, 12, and 24 h timepoints by RT-qPCR with corresponding prevalidated primers (sequence in supplementary Table S1) for target genes. Changes in growth hormone (GH), growth hormone receptor (GHR), and melanocyte-inducing transcription factor (MITF) and ABC-transporters (supplementary Figure S1A–D) with time due to doxorubicin (A), vemurafenib (B), crizotinib (C), and cabozantinib (D) treatments reflected concomitant changes in MITF, GH–GHR axis (A–D) as well as in ABC-transporters (Supplementary Figure S1A–D). (E–H) Changes in RNA expressions of GH (E,F) and MITF (G,H) RNA expressions at 24 h treatment with doxorubicin (dox), or cisplatin (cis), or vemurafenib (vem) treatments on MDA-MB-435 (E,G) and MALME-3M (F,H) cells. RNA expressions were quantified by RT-qPCR and normalized against expression of TUBB5 and ACTB as reference genes (*, p < 0.05, Wilcoxon sign rank test, n = 3). (J–L) In SK-MEL-28, MDA-MB-435, and MALME-3M human melanoma cells, treatment with doxorubicin for 48 h increased the protein levels of GH and MITF as seen by western-blot (J) and subsequent quantification using ImageJ (NIH) (K,L) and expressions were normalized against expression of ACTB (β-actin) (*, p < 0.05, Students t test, n = 3).

Figure 2

GH treatment directly upregulates MITF and MITF target RNA expression in human melanoma cells: (A,B) Human melanotic melanoma cells SK-MEL-28 were treated with increasing doses (0, 50, 100, 200 ng/mL) of recombinant human growth hormone (GH) and heatmap showing changes in RNA expressions at 6, 24, and 48 h timepoints were analyzed for GH, GHR, MITF, and a number of MITF targets, as well as the ABC-transporters ABCB5 and ABCG2 (A); An identical experiment was performed in the presence of 200 nM doxorubicin (B). Numbers inside boxes indicate fold-change in gene expression compared to GH untreated control. Similar set of experiments for amelanotic melanoma cells SK-MEL-28 is shown in supplementary Figure S2. (C–E) SK-MEL-28 cells were further transfected either with scramble (scr) of GHR-targeted (GHR) siRNA for gene knockdown in presence or absence of 50 ng/mL GH treatment and MITF (C), TYRP1 (D), and MET (E) RNA expressions were analyzed. RNA expressions were quantified by RT-qPCR and normalized against expression of TUBB5 and ACTB as reference genes (*, p < 0.05, Wilcoxon sign rank test, n = 3).

Figure 3

GH treatment directly upregulates MITF and MITF target proteins’ expression in human melanoma cells: (A,B) Human melanotic melanoma cells SK-MEL-30 (A) and SK-MEL-28 (B) were treated with/without 50 ng/mL recombinant human growth hormone in presence/absence of doxorubicin and siRNA-mediated GHR knockdown (GHRKD). After 48 h, protein was extracted and western-blot was performed for MITF, PMEL (PMEL17/gp100), MET, and GAPDH followed by densitometry analysis performed using ImageJ (NIH) and expression were normalized against expression of GAPDH as loading control (#,*, p < 0.05, Students t test, n = 3; # indicates comparison against untreated (-GH, -dox, -GHRKD) control).

Figure 4

GH enhances MITF regulated melanogenesis activity in human melanoma cells. Changes in amounts of melanin formed inside SK-MEL-30 cells following treatment with different treatments for 72 h in presence of 0, 50, 100 ng/mL recombinant human growth hormone as well as siRNA-mediated GHR knockdown (GHRKD) in presence of 50 ng/mL GH was analyzed. (A) alpha-MSH analog melanotan (100 nM), doxorubicin (1 μM); (B) mutant BRAF targeted therapeutics—vemurafenib (2 μM), and dabrafenib (3 μM); or (C) tyrosine kinase inhibitors—linsitinib (5uM), crizotinib (10 μM), sorafenib (5 μM) were used. Similar treatments with doxorubicin, cisplatin, and vemurafenib was done with BRAF-mutant cell-lines MALME-3M (D) and MDA-MB-435 (E) as well as the mouse melanoma cell line B16F10 (F). Cell pellets were dissolved by heating in NaOH and absorbance was measured at 450 nm. Reads were normalized by cell counts at the end of 72 h treatments. Inset shows cell pellets from doxorubicin treatment. Data represents mean +/− SD. Student’s t-test was performed using MS-EXCEL to identify significant differences at p < 0.05. (#,*, p < 0.05, Students t test, n = 3; * indicates comparison against untreated (-GH, -dox, -GHRKD) control).

Figure 5

GH enhances MITF-regulated tyrosinase activity in human melanoma cells. Changes in tyrosinase activity in 25 μg cell lysates of (A) MDA-MB-435, (B) MALME-3m, (C) SK-MEL-30, and (D) SK-MEL-28 cells treated with/without 50 ng/mL recombinant human growth hormone in presence/absence of doxorubicin and siRNA-mediated GHRKD. After 48 h treatment, total protein was extracted and quantified by Bradford assay. Tyrosinase assay was performed using a commercial kit (BioVision) following manufacturer’s protocol (#,*, p < 0.05, Students t test, n = 3; # indicates comparison against untreated (-GH, -dox, -GHRKD) control).

Figure 6

B16F10 mouse melanoma xenografts in bovine growth hormone transgenic (bGH) mice had elevated expression of MITF and MITF targets compared to the same in WT mice. Mouse melanoma B16F10 cells were xenografted in immunocompetent bovine growth hormone transgenic mice (bGH) as well as its wild-type littermates (WT) of both sexes (n = 6). Tumors grown for three weeks were extracted and levels of RNA and protein were analyzed by RT-qPCR and western-blot. Excess GH in bGH mice showed upregulation of multiple target RNA and protein levels compared to WT mice. (A,B) Changes in RNA levels of MITF and MITF-target genes in (A) male and (B) female bGH and WT mice. (C–E) Changes in protein levels of MITF and MITF-targets in male (D) and female (E) bGH and WT mice as seen by western-blot (D) and quantified by densitometry analysis performed using ImageJ (NIH) and expression were normalized against expression of TUBB5 as loading control (*, p < 0.05, Students t test, n = 3).

Figure 7

Bioinformatic analysis: GHR, MITF, and MITF target genes coexpress and strongly cluster in the human sample datasets. (A) Heatmap shows normalized RNA expression (FRKM) values of GHR, MITF, and 48 MITF target gene expression in all the samples of National Cancer Institute’s NCI-60 cancer cell panel. Green box indicates clustering of GHR, MITF, and MITF targets. Supplementary Figure S7 shows similar analyses for the CCLE (Cancer Cell Line Encyclopedia) dataset. (B) Correlation (Pearson’s) analysis of GHR and MITF RNA expression in the GTEx (Genotype Tissue Expression) human post-mortem dataset containing samples from 53 non-diseased tissue sites for nearly 1000 individuals using the GEPIA platform. A positive and highly significant correlation between GHR and MITF expression in sun-exposed (lower-leg; Pearson coefficient = 0.19, p = 0.00071) and not sun-exposed (suprapubic; Pearson coefficient = 0.25, p = 0.00011) normal human skin was observed.

Figure 8

GH-regulated MITF and MITF target gene regulation proceeds via JAK2–STAT5- and SRC-regulated pathways: Human melanoma cell MDA-MB-435 (here), SK-MEL-28 (supplementary Figure S9) and SK-MEL-30 (supplementary Figure S10) were treated with/without doxorubicin in the presence of GH as well as different intracellular signaling pathway inhibitors. After 24 h treatment, RNA expression for target genes (GH, MITF, and MITF targets) was quantified by RT-qPCR and normalized against expression of TUBB5 and ACTB as reference genes (#,*, p < 0.05, Wilcoxon sign rank test, n = 3; * indicates comparison against corresponding -GH controls while # indicates comparison against corresponding +GH controls in DMSO and doxorubicin treated groups).