Melanoma-Derived iPCCs Show Differential Tumorigenicity and Therapy Response

Abstract

A point mutation in the BRAF gene, leading to a constitutively active form of the protein, is present in 45%-60% of patients and acts as a key driver in melanoma. Shortly after therapy induction, resistance to MAPK pathway-specific inhibitors develops, indicating that pathway inhibition is circumvented by epigenetic mechanisms. Here, we mimicked epigenetic modifications in melanoma cells by reprogramming them into metastable induced pluripotent cancer cells (iPCCs) with the ability to terminally differentiate into non-tumorigenic lineages. iPCCs and their differentiated progeny were characterized by an increased resistance against targeted therapies, although the cells harbor the same oncogenic mutations and signaling activity as the parental melanoma cells. Furthermore, induction of a pluripotent state allowed the melanoma-derived cells to acquire a non-tumorigenic cell fate, further suggesting that tumorigenicity is influenced by the cell state.

Keywords: BRAF; cancer; iPCC; iPSC; induced pluripotency; inhibitor; melanoma; reprogramming; stem cells; therapy resistance.

Figure 1

Generation and Characterization of Metastable Reprogrammed Melanoma Cells (A) Scheme for tumor cell reprogramming. (B) Reprogrammed tumor cells form ESC-like alkaline phosphatase-positive colonies on feeder cells. (C) qPCR measurement shows reactivation of the endogenous loci of the pluripotency markers NANOG, SOX2, and SALL4 but only mild increase in endogenous OCT4 expression. GAPDH was used as endogenous control and hiPSCs as reference sample. Indicated is the mean ± SD. p Values were calculated by two-tailed, unpaired sample t test of technical triplicates in two clones of SKMEL147-and Mewo-iPCCs and in three independent experiments of HT-144-iPCCs. Asterisk indicates t test p value of ≤0.05 in comparison with the respective reference (^∗p ≤ 0.05, ^∗∗p ≤ 0.01, ^∗∗∗p ≤ 0.005). (D) Immunofluorescence staining of NANOG and TRA-1-60 in the parental melanoma cells (I) and melanoma iPCCs (II). DAPI was used for nuclear counterstaining. (E) qPCR analysis reveals loss of melanocytic markers in iPCCs compared with their parental melanoma cell lines. Gene expression levels were normalized to GAPDH. Error bars indicate 95% confidence intervals. p Values were calculated from three independent experiments by two-tailed, unpaired sample t test. Asterisk indicates t test p value of ≤0.05 in comparison with the respective reference (^∗p ≤ 0.05, ^∗∗p ≤ 0.01, ^∗∗∗p ≤ 0.005). (F) NANOG expression analyzed by qPCR in HT-144-iPCCs at indicated time points after doxycycline withdrawal compared with the parental HT-144. Nanog expression was normalized to internal GAPDH. Error bars indicate 95% confidence intervals. Dotted line, normalization to day 0. (G) Metastable melanoma iPCCs form teratomas in vivo showing tissue structures of mesodermal (I), ectodermal (II), and endodermal (III) origin. Paraffin-embedded tumor slices were stained with H&E. See also Figure S1.

Figure 2

Characterization of Melanoma iPCCs (A) Heatmap and dendrogram generated by unsupervised hierarchical clustering of differentially regulated genes (empirical Bayes moderated t test, p < 0.05) in HT-144, HT-144-iPCCs, and hiPSCs. Red indicates increased expression and green decreased expression relative to the control. (B–D) qPCR analysis of melanoma iPCCs compared with their reprogrammed progeny. iPSCs were used as reference sample. Data were obtained from three independent experiments. (B) Expression analysis of development-associated TGF-B superfamily members NODAL, LEFTY1, and LEFTY2. (C) Expression analysis of the epigenetic modifier DNMTL3. (D) Expression analysis of the MET marker E-cadherin. GAPDH was used as endogenous control and hiPSCs as reference sample. Error bars indicate 95% confidence intervals. Indicated is the mean + SD. p Values were calculated by two-tailed, unpaired sample t test. Asterisk indicates t test p value of ≤0.05 in comparison with the respective reference (^∗∗∗p ≤ 0.005).

Figure 3

Reduced Tumor Marker Expression in Reprogrammed Melanoma Cells (A) Histological staining for the melanoma marker S100B in melanomas derived from HT-144, Mewo, and SKMEL147 cells and their respective reprogrammed progeny. (B) In contrast to melanomas, S100B expression is restricted to neuronal-like cells in HT-144-iPCC-derived teratomas (indicated by arrows). (C) Histological staining for epithelial cytokeratins (PanCK) in HT-144-iPCC-derived tumors and melanomas derived from the parental HT-144 cells. (D) Histological analysis of the expression of the proliferation marker KI67 in melanomas derived from HT-144, Mewo, and SKMEL147 cells and their respective reprogrammed progeny.

Figure 4

HT-144-iPCCs Efficiently Differentiate into Neuronal Cells (A) Schematic illustration of the neuronal differentiation protocol. (B) Morphology of HT-144-iPCCs subjected to neuronal induction at indicated time points. (C and D) qPCR analysis of pluripotency marker (C) as well as early and late neuronal marker expression (D). Data were pooled from three independent experiments. NANOG and OCT4 were normalized to hiPSCs. The neuronal markers MAP2, PAX6, and RBFOX3 were normalized to hiPSC-derived neuronal cells. GAPDH served as internal control. Error bars indicate 95% confidence intervals. p Values were calculated by two-tailed, unpaired sample t test. Asterisk indicates t test p value of ≤0.05 in comparison with the respective reference (^∗∗∗p ≤ 0.005). (E) Immunofluorescence staining of B3-tubulin in HT-144-iPCCs (I) and iPCC-derived neuronal cells (II). iPCCs used for the staining constitutively express GFP.

Figure 5

Melanoma iPCCs Gain the Potential to Terminally Differentiate into Non-tumorigenic Cells (A) Schematic protocol for the differentiation of iPCC-derived fibroblast-like cells. (B) qPCR analysis of NANOG and MITF expression in HT-144, reprogrammed HT-144-iPCCs, and iPCC-derived fibroblast-like cells (HT-144-dFLC-I). GAPDH served as endogenous control. Error bars indicate 95% confidence intervals. p Values were calculated from three independent experiments by two-tailed, unpaired sample t test. Asterisk indicates t test p value of ≤0.05 in comparison with the respective reference (^∗p ≤ 0.05, ^∗∗∗p ≤ 0.005). (C) Morphological comparison of the parental cell line HT-144 and fibroblast-like cells from two differentiations (HT-144-dFLC) with NHFs. (D) Global methylation profiles of A375 and HT-144 compared with the profiles of their respective iPCCs and HT-144-iPCC-derived fibroblast-like cells from three independent differentiations were used for a multidimensional scaling. Human fibroblasts (pink squares represent preparations from two different individuals), hiPSCs, and a publicly available dataset of hESCs were added as controls. (E) Survival of mice after subcutaneous injection of 1 × 10⁶ parental HT-144 melanoma cells or HT-144-iPCC-derived fibroblasts, respectively. Parental HT-144 melanoma cells gave rise to tumors in all five cases while no tumor growth was observed in five mice injected with HT-144-dFLC-II even after 20 weeks (one mouse was euthanized for other reasons).

Figure 6

Epigenetic Modifications Induced by Nuclear Reprogramming Lead to Therapy Resistance against MAPK Inhibitors in Melanoma Cells (A) Oncogene-specific PCR analysis of BRAF^V600E in the parental HT-144 cells, HT-144-iPCCs, and iPCC-derived fibroblast-like cells. Human fibroblasts were used as a negative control. (B) Methylation analysis of the BRAF locus in HT-144, HT-144-iPCCs, and HT-144-dFLCs. (C) ERK activity in reprogrammed HT-144-iPCCs and fibroblast-like cells compared with the parental cell line determined by western blot analysis of phosphorylated ERK. GAPDH was used as an internal reference control for semi-quantitative protein analysis. (D) FISH of HT144 and HT144-iPCCs shows no gain or loss of gene copy numbers related to genes of the BRAF-MEK-ERK signaling pathway such as BRAF (shown are nuclei with signals for the BRAF gene locus [red] and a centromeric reference probe [green]). (E) HT-144-iPCCs give rise to tumors with reduced ERK activity in distinct differentiated structures. Histological staining of total ERK (tERK) and phosphorylated ERK (pERK) in HT-144-iPCC-derived tumors. (F) Treatment of parental melanoma cells and their reprogrammed counterparts with the MEK inhibitor trametinib and the BRAF^V600E-specific inhibitor vemurafenib. (G) Therapy response of MAPK inhibitor-treated HT-144 and two fibroblast-like in vitro differentiations. Cells were incubated with 1,000 nM trametinib and 100 nM vemurafenib and analyzed for their metabolic activity at indicated time points. See also Figure S5.