Monitoring melanoma patients on treatment reveals a distinct macrophage population driving targeted therapy resistance

Summary Resistance to targeted therapy remains a major clinical challenge in melanoma. To uncover resistance mechanisms, we perform single-cell RNA sequencing on fine-needle aspirates from resistant and responding tumors of patients undergoing BRAFi/MEKi treatment. Among the genes most prominently expressed in resistant tumors is POSTN, predicted to signal to a macrophage population associated with targeted therapy resistance (TTR). Accordingly, tumors from patients with fast disease progression after therapy exhibit high POSTN expression levels and high numbers of TTR macrophages. POSTN polarizes human macrophages toward a TTR phenotype and promotes resistance to targeted therapy in a melanoma mouse model, which is associated with a phenotype change in intratumoral macrophages. Finally, polarized TTR macrophages directly protect human melanoma cells from MEKi-induced killing via CD44 receptor expression on melanoma cells. Thus, interfering with the protective activity of TTR macrophages may offer a strategy to overcome resistance to targeted therapy in melanoma.

(C) Down-regulation of POSTN expression in melanoma cells after cell transfection with POSTN siRNA.The analysis was done with qRT-PCR.The data are normalized to expression of the housekeeping gene PPIA and presented as fold change to CTR siRNA (n=4-5 independent experiments, unpaired t-test, mean ± s.e.m).(D-E) Representative western blots of MAPK signaling pathway (pERK and total (tot)ERK antibody) after (D) POSTN expression was inhibited in cells via POSTN siRNA transfection for 24h and then treated with MEKi for 3 days and (E) cells were pretreated with recombinant (rec) POSTN for 24h and then treated with MEKi for 3 days.α-Tubulin antibody was used as loading control.Graphical analysis shows a quantification of pERK expression.Quantification was performed by ImageJ software.Results were normalized to totERK and quantified as fold change versus loading control α-Tubulin (n=3-4 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(F-G) Bar graphs showing percentage of alive cells analyzed by flow cytometry data (Annexin V-negative, 7AAD-negative) after (F) POSTN expression was inhibited in cells via POSTN siRNA transfection for 24h and then treated with MEKi for 3 days and (G) Cells were pretreated with recombinant (rec) POSTN for 24h and then treated with MEKi for 3 days (n=3 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(H) Bar graphs showing quantification of pERK expression and percentage of alive cells analyzed by flow cytometry data (Annexin V-negative, 7AAD-negative) after cells were pretreated with recombinant (rec) MDK for 24h and then treated with MEKi for 3 days.For the western blot, α-Tubulin antibody was used as loading control.Quantification was performed by ImageJ software.Results were normalized to total (tot)ERK and quantified as fold change versus loading control α-Tubulin.(n=3-4 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).

Figure S1 .
Figure S1.Cell populations detected by scRNAseq in Cohort 1 and its characterization.Related to Figure 1 and 2. (A) UMAP plot showing 31 clusters in the tumors of Cohort 1 after PCA and unsupervised Louvain-based clustering.(B) Heatmap of copy number analysis of tumors from Cohort 1 by InferCNV.CD4 + T-cells were used as reference.(C) UMAP plot represents clustering of malignant cells from tumors of Cohort 1. (D) Heatmap showing the enrichment scores of gene signatures of malignant cell populations from tumors of Cohort 1 with published resistance-related phenotypic markers.

Figure S2 .
Figure S2.POSTN, MDK and SPP1 expression.Related to Figure 2. (A) Heatmap indicating differential expressed genes in malignant cells of every tumor in Cohort 1 at every time point of biopsies.(B) Violin plots illustrating expression of MDK and SPP1 in the cell populations for each tumor in Cohort 1 identified by scRNAseq.(C) Violin plots showing expression of POSTN, MDK and SPP1 in malignant cells at every FNA time point for each tumor in Cohort 1. (D) Representative pictures of POSTN IHC staining in fast and slow progressing tumors.

Figure S3 .
Figure S3.POSTN and MDK could not regulate resistance formation in vitro.Related to Figure 2. (A) POSTN expression in melanoma cells (n=3-4 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(B) MDK expression in melanoma cells (n=4-6 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(C)Down-regulation of POSTN expression in melanoma cells after cell transfection with POSTN siRNA.The analysis was done with qRT-PCR.The data are normalized to expression of the housekeeping gene PPIA and presented as fold change to CTR siRNA (n=4-5 independent experiments, unpaired t-test, mean ± s.e.m).(D-E) Representative western blots of MAPK signaling pathway (pERK and total (tot)ERK antibody) after (D) POSTN expression was inhibited in cells via POSTN siRNA transfection for 24h and then treated with MEKi for 3 days and (E) cells were pretreated with recombinant (rec) POSTN for 24h and then treated with MEKi for 3 days.α-Tubulin antibody was used as loading control.Graphical analysis shows a quantification of pERK expression.Quantification was performed by ImageJ software.Results were normalized to totERK and quantified as fold change versus loading control α-Tubulin (n=3-4 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(F-G) Bar graphs showing percentage of alive cells analyzed by flow cytometry data (Annexin V-negative, 7AAD-negative) after (F) POSTN expression was inhibited in cells via POSTN siRNA transfection for 24h and then treated with MEKi for 3 days and (G) Cells were pretreated with recombinant (rec) POSTN for 24h and then treated with MEKi for 3 days (n=3 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(H) Bar graphs showing quantification of pERK expression and percentage of alive cells analyzed by flow cytometry data (Annexin V-negative, 7AAD-negative) after cells were pretreated with recombinant (rec) MDK for 24h and then treated with MEKi for 3 days.For the western blot, α-Tubulin antibody was used as loading control.Quantification was performed by ImageJ software.Results were normalized to total (tot)ERK and quantified as fold change versus loading control α-Tubulin.(n=3-4 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).

Figure S4 .
Figure S4.Expression of POSTN and MDK receptors and analysis of TTR and TTS macrophage cell types with classical M1 and M2 markers.Related to Figure 3. (A) CellChat analysis showing expression of POSTN receptors across tumors and cell populations of Cohort 1. (B) CellChat analysis showing expression of MDK receptors across tumors and cell populations of Cohort.(C) TTR and TTS macrophage module score for Martinez M1 and M2 macrophage gene signatures.(D) TTR and TTS macrophage module score for Newman M1 and M2 macrophage gene signatures.

Figure S5 .
Figure S5.POSTN and MDK receptor expression across TTR and TTS macrophages, their signaling pathway networks, and characterization of TTR macrophages.Related to Figure 3 and 4. (A) CellChat analysis showing expression of POSTN receptors between TTR (blue) and TTS macrophages (yellow) (Wilcoxon rank sum test) of Cohort 1. (B) CellChat analysis showing expression of MDK receptors between TTR (blue) and TTS macrophages (yellow) (Wilcoxon rank sum test) of Cohort 1. (C) POSTN and MDK signaling pathway network in Cohort 1. Malignant cells were used as senders and other cell populations used as receivers.(D) UMAP plot showing 6 clusters after multiplex immunohistochemistry staining of TTR and other macrophage markers in Cohort 3 generated by Leiden clustering.(E) Heatmap showing the average marker expression per Leiden cluster in Cohort 3. (F) Heatmap showing the average marker expression per subcluster of Leiden cluster 5 in Cohort 3. (G) Representative western blots of MAPK signaling pathway (pERK and total (tot)ERK antibody) in Yumm1.7 cell line after cells were treated with different concentrations of MEKi for 3 days.α-Tubulin antibody was used as loading control.(H)Expression of POSTN in Yumm1.7 cells infected with empty lentiviral vector (Yumm1.7 CTR) or POSTN overexpressing lentiviral vector (Yumm1.7 POSTN+) (flow cytometry data, n=3 independent experiments, unpaired t test, mean ± s.e.m).(I) Bar graphs show cell survival of Yumm1.7 CTR and Yumm1.7 POSTN+ cells after treatment with MEKi for 3 days (flow cytometry, Annexin V-negative, 7AAD-negative) (n=3 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).

Figure S6 .
Figure S6.Ligand-receptor-based CellChat communication of TTR and TTS macrophages with malignant cells.Related to Figure 7. (A) Expression of VEGF-A and TNFα in TTR and TTS macrophages.MФ = macrophages.(B)Bubble plot showing ligand-receptor-based CellChat analysis of the FNA scRNAseq dataset from Cohort 1, with TTR and TTS macrophages as senders and the malignant cells as receivers.MФ = macrophages.(C) Bubble plot showing ligand-receptor based CellChat analysis of Cohort 5 with the TTR, TTS and other macrophages as senders and the malignant cells as receivers.MФ = macrophages.

Figure S7 .
Figure S7.SPP1 and CD44 expression and their related pathways.Related to Figure 7. (A) Violin plots showing expression of SPP1 across the tumors and cell populations identified by scRNAseq of Cohort 1. (B) Violin plots showing expression of CD44 across the tumors and different cell populations of Cohort 1. (C) CD44 expression in melanoma cells (flow cytometry data, n=3 independent experiments, mean ± s.e.m).(D) Bar graphs show cell survival after treatment with MEKi or treatment with MEKi following a 2 hour pretreatment with CD44 blocking antibody (flow cytometry, Annexin V-negative, 7AAD-negative) (n=3 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(E) Bar graphs show cell survival after treatment with MEKi for 3 days following a 2-hour pretreatment with recombinant SPP1 or CD44 blocking antibody or combined recombinant SPP1+ CD44 blocking antibody (flow cytometry, Annexin V-negative, 7AAD-negative) (n=3 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).(F) Principal component analysis (PCA) plot for the transcriptomic data derived from the four experimental groups.M150543 cells were treated with MEKi for 20 hours following a 2 hour pretreatment with recombinant SPP1 or CD44 blocking antibody or combined recombinant SPP1+ CD44 blocking antibody.Points are color coded per treatment (MEKi = 3 N; CD44Ab + MEKi = 3 N; SPP1 + MEKi = 2 N; SPP1 + CD44Ab + MEKi = 2 N).(G) Quantification of canonical CD44 downstream signaling pathways.M150543 cells were treated with MEKi alone for 6 hours or following a 2-hour pretreatment with CD44 blocking antibody, recombinant SPP1 or combined recombinant SPP1+ CD44 blocking antibody (flow cytometry, n=4-5 independent experiments, ordinary one-way ANOVA test, mean ± s.e.m).