PHGDH Induction by MAPK Is Essential for Melanoma Formation and Creates an Actionable Metabolic Vulnerability

Abstract

Overexpression of phosphoglycerate dehydrogenase (PHGDH), the rate-limiting enzyme in the serine synthesis pathway, promotes melanomagenesis, melanoma cell proliferation, and survival of metastases in serine-low environments such as the brain. Here, we found that PHGDH is universally increased in melanoma cells and required for melanomagenesis. Although PHGDH amplification explained PHGDH overexpression in a subset of melanomas, oncogenic BRAFV600E also promoted PHGDH transcription through mTORC1-mediated translation of ATF4. Importantly, depletion of PHGDH in genetic mouse melanoma models blocked tumor formation. In addition to BRAFV600E-mediated upregulation, PHGDH was further induced by exogenous serine restriction. Surprisingly, BRAFV600E inhibition diminished serine restriction-mediated PHGDH expression by preventing ATF4 induction. Consequently, melanoma cells could be specifically starved of serine by combining BRAFV600E inhibition with exogenous serine restriction, which promoted cell death in vitro and attenuated melanoma growth in vivo. In summary, this study identified that PHGDH is essential for melanomagenesis and regulated by BRAFV600E, revealing a targetable vulnerability in BRAFV600E-mutant melanoma. Significance: BRAFV600E promotes the expression of the serine synthesis enzyme PHGDH, which is required for melanoma formation, and can be targeted to sensitize melanoma to dietary serine restriction, providing a melanoma cell-specific treatment strategy.

Fig. 1:. PHGDH is upregulated in melanoma irrespective of its copy number status.

(A) PHGDH mRNA expression and copy number status in human melanoma tumors from the Skin Cutaneous Melanoma TCGA Pan Cancer Atlas dataset. (B) PHGDH expression comparing Nevi (n=18) vs Melanoma (n=45) in the GSE3189 dataset. (C) Correlation analysis between PHGDH mRNA expression and copy number status in melanoma cell lines (n=69) from the DepMap dataset (22Q2). The cell lines marked in red are considered amplified. (D) PHGDH absolute copy number in melanocyte and melanoma cell lines was quantified by Droplet Digital PCR using RNaseP as endogenous control. Values shown are the Mean±SD of a representative experiment performed in triplicate. H1, Hermes1; H2, Hermes2; H3, Hermes3A; H4, Hermes4B. (E) PHGDH mRNA expression in melanocyte and melanoma cell lines determined using qRT-PCR. The values are relative to Hermes1 (H1) cells and normalized using 18S as internal control. (F) PHGDH protein expression quantified using ImageJ software using the image from Supplementary Fig. S1A. The values shown are relative to Hermes1 (H1) cells and normalized using β-Actin and represent the average of two independent experiments. (G-I) Simple linear correlation analysis between PHGDH copy number and mRNA levels (G), PHGDH copy number and PHGDH protein levels (H), and PHGDH mRNA and protein levels (I). R represents the correlation coefficient. N=2 biological replicates. ** p<0.01.

Fig. 2:. PHGDH but not extracellular S/G is critical for melanoma formation.

(A) Kaplan Meier curve showing the overall survival of Braf^V600E; Pten^FL/FL chimeras fed +S/G or −S/G diets. (B) Total number of tumors developed by mice shown in (A). (C) Phgdh expression in tumors of mice shown in (A). The quantification of Supplementary Fig. S1B is shown. (D) Kaplan Meier curve showing the overall survival of BPP^TRE-shPhgdh (shPhgdh) and BPP^TRE-shRen (shRen) chimeras. (E) Total number of tumors developed by mice shown in (D). (F) GFP and Phgdh expression in tumors of mice shown in (D). The quantification of Supplementary Fig. S2C is shown. (G) Kaplan Meier curve showing the overall survival of BP^{LSL-shPten-shPhgdh} (LSL-shPhgdh) and BP^{LSL-shPten-shRen} (LSL-shRen) chimeras. (H) Total number of tumors developed by mice shown in (G). (I) Western blot showing GFP and Phgdh expression in tumors of mice shown in (G). SE, short exposure; LE, long exposure. The survival curves were analyzed using Log-rank (Mantel-Cox) test. **** p<0.0001, *** p<0.001, * p<0.05; ns, not significant.

Fig. 3:. Mutant BRAF^V600E promotes PHGDH expression.

(A, B) PHGDH mRNA (A) and protein (B) expression upon acute overexpression of mutant BRAF^V600E in Hermes1 melanocytes. BRAF^V600E expression was induced by 0.5 μg/ml Doxycycline for 72 hours. (C, D) PHGDH mRNA (C) and protein (D) expression in response to MAPK pathway inhibition using BRAF^V600E inhibitor (PLX4032), MEK inhibitor (AZD6244), or ERK inhibitor (SCH772984) at 0.25 μM in human 1205Lu cells after 72 hours. (E, F) Phgdh mRNA (E) and protein (F) expression in response to MAPK pathway inhibition using BRAFV600E inhibitor (PLX4032), MEK inhibitor (AZD6244), or ERK inhibitor (SCH772984) at 0.25 μM in murine M10M3 cells after 72 hours. (G) Quantification of PHGDH expression in M10M3 allograft tumors after acute treatment with PLX4032 orally once daily for 3 days. Tumors were collected 6 hours after last dose. The quantification of Supplementary Fig. S3H is shown. (H) Correlation between PHGDH expression and MAPK activation (MAPK signature score) in melanoma cell lines from DepMap. Cell lines were categorized into top and bottom quartiles based on PHGDH expression levels. **** p<0.0001, * p<0.05; ns, not significant.

Fig. 4:. BRAF^V600E promotes PHGDH expression via the mTORC1-ATF4 axis.

(A, B) Western blot showing the effect of treatment with 0.25 μM PLX4032, AZD6244, or SCH772984 for 72 hours on PHGDH, ATF4, pTSC2, pS6, and p4EBP1 expression in 1205Lu cells (A) and M10M3 cells (B). (C) Western blot showing the effect of acute overexpression of BRAF^V600E on PHGDH, ATF4, pTSC2, pS6, and p4EBP1 expression in Hermes1 cells after 72 hours of 0.5 μg/ml Doxycycline. (D, E) PHGDH mRNA (D) and PHGDH and ATF4 protein expression (E) in 1205Lu cells in response to 0.1 μM Temsirolimus treatment for 72 hours. (F, G) PHGDH and ATF4 mRNA (F) and protein (G) expression in 1205Lu cells upon ATF4 silencing using siRNA. Tem, Temsirolimus. ** p<0.005, * p<0.05.

Fig. 5:. The MAPK pathway regulates the translation of ATF4.

(A) Western blot showing the effect of 48 hours of treatment with PLX4032 (0.25 μM), Temsirolimus (Tem, 0.1 μM), or the GNC2 inhibitor GCN2iB (20 μM) on the expression of the uORF-ATF4-GFP reporter, endogenous ATF4, and pEIF2α in 1205Lu cells grown in +S/G or −S/G medium. (B) Western blot and quantification showing the effect of 48 hours of PLX4032 treatment (0.25 μM) on ATF4 accumulation upon proteasome inhibition with MG132 (10 μg/ml) in 1205Lu cells. The quantification graph shows the average of two independent experiments. (C) Western blot showing the effect of 72 hours of PLX4032 (0.25 μM) or Temsirolimus (0.1 μM) treatment on the doxycycline inducible expression of endogenous ATF4 and mTORC1-independent ectopic HA-ATF4 in 1205Lu cells. (D) Western blot showing the effect of combined PLX4032 treatment (0.25 μM) for 72 hours and S/G depletion on PHGDH and ATF4 protein levels in 1205Lu and M10M3 cells.

Fig. 6:. BRAF inhibition sensitizes melanoma to S/G starvation.

(A) Quantification of dead 1205Lu and A375 cells cultured in +S/G or −S/G medium and treated with PLX4032 (1205Lu, 0.25 μM; A375, 0.5 μM) ± ZVAD-FMK (100 μM) for 72hrs. SyGr-pos., SYTOX Green-positive. (B) Quantification of colony growth assays of cells treated with PLX4032 (1205Lu, 0.025 μM; A375, 0.5 μM; SKMel28, 0.05 μM) and grown in +S/G or −S/G medium for 10-15 days. (C) Graph showing the effect of different diets on melanoma growth in Braf^V600E, Pten^FL/FL mice. Tumor growth was normalized to the tumor size at the start of treatment. (D) Kaplan Meier curve showing the overall survival of Braf^V600E, Pten^FL/FL mice treated on diets with or without S/G and with or without PLX4720. (E) Quantification of serine and glycine in melanomas treated with ± PLX4720 ± S/G for 7 days. * p<0.05, ** p<0.005, *** p<0.001, **** p<0.0001