Fig 2: Effects of inhibition of c-Jun, galectin-3, and HDAC3 on ARSB-mediated changes in PD-L1 expression in A375 melanoma cells and normal melanocytes. (a) In the A375 cells, PD-L1 protein is markedly increased by ARSB siRNA, and this increase is inhibited by cJP and galectin-3 siRNA. PHPS1 does not affect the PD-L1 expression. The HDAC3 inhibitor, RGFP966, further increases the ARSB siRNA-induced increase in PD-L1. (b) In normal melanocytes, similar effects of ARSB siRNA, cJP, galectin-3 siRNA, and RGFP are observed. The total production of PD-L1 is less than in the A375 cells (642 ± 34 pg/mg protein vs. 796 ± 50 pg/mg protein; p < 0.0001, n = 9). (c,d) Similar effects are observed in PD-L1 mRNA expression in the A375 malignant cells and the normal melanocytes (p < 0.0001). Galectin-3 siRNA inhibits the effect of ARSB silencing, consistent with dependence on galectin-3 for the observed effects. (e) JQ1, the inhibitor of the BET family of bromodomain proteins, blocks the effects of ARSB siRNA and RGFP on PD-L1 expression (p < 0.0001). p-values are determined by an unpaired t-test, two-tailed, with unequal variance. [ARSB = arylsulfatase B = N-acetylgalactosamine-4-sulfatase; consi = control siRNA; gal-3 = galectin-3; cJP = c-Jun mimetic peptide; JQ1 = inhibitor of the BET family of bromodomain proteins; PD-L1 = programmed death ligand-1; PHPS1 = SHP2 inhibitor; RGFP = RGFP966 = HDAC3 inhibitor; rh = recombinant human; si = small interfering siRNA; 0.2 = 0.2 mg/kg treatment dose]. **** represents p ≤ 0.0001.

Fig 3: Schematic of overall pathway. Following exposure to bioactive, recombinant human ARSB, the sulfate group of chondroitin 4-sulfate is removed and galectin-3 binding with C4S is enhanced. This leads to a decline in the free galectin-3 and reduced impact on c-Jun DNA binding. Reciprocal effects lead to an increase in HDAC3, reduced acetyl-H3, increased availability of histone lysines to bind with PD-L1 promoter DNA, leading to closed chromatin, and reduced expression of PD-L1. When ARSB is silenced, opposite effects occur due to the increased availability of galectin-3.

Fig 4: Impact of methyl-β-d-xylopyranoside (BDX) on mediators of PD-L1 expression. (a) Methyl-β-d-xylopyranoside (BDX), an inhibitor of chondroitin sulfate proteoglycan biosynthesis, increases free galectin-3 in the A375 cells (p ≤ 0.001, n = 3). The combination of ARSB siRNA and BDX has a slightly greater effect than ARSB siRNA alone (p < 0.05, n = 3). (b) Nuclear c-Jun is increased by BDX (p ≤ 0.001, n = 3) and ARSB siRNA (p ≤ 0.001, n = 3). cJP inhibits the effects of BDX and ARSB siRNA. (c) HDAC3 activity is reduced by both ARSB siRNA (p = 0.002, n = 3) and BDX (p = 0.0002, n = 3) and further reduced by their combination (p = 0.002, p = 0.003). These effects are reversed by cJP. (d) PD-L1 expression is increased by both ARSB siRNA and BDX, and these increases are inhibited by cJP. p-values are determined by unpaired t-test, two-tailed, with unequal variance. [ARSB = arylsulfatase B = N-acetylgalactosamine-4-sulfatase; BDX = methyl-β-d-xylopyranoside; consi = control siRNA; cJP = c-Jun mimetic peptide; HDAC = histone deacetylase; NS = not significant; PD-L1 = programmed death ligand-1; si = small interfering siRNA; 0.2 = 0.2 mg/kg treatment dose]. **** represents p ≤ 0.0001, *** represents p ≤ 0.001, ** represents p ≤ 0.01, and * is for p ≤ 0.05.

Fig 5: PD-L1 expression and protein in B16F10 subcutaneous melanomas, normal melanocytes, and A375 melanoma cells following ARSB silencing and exposure to exogenous, bioactive ARSB. (a,b) In the mouse melanomas, PD-L1 protein (n = 9) and mRNA expression (n = 6) are reduced following treatment with exogenous ARSB (p < 0.0001). (c,d) In normal human melanocytes, ARSB silencing increases the protein and mRNA expression of PD-L1 (p < 0.0001, n = 9). (e,f) Similarly, in the malignant A375 cells, PD-L1 protein and mRNA are increased by ARSB silencing and reduced by rhARSB (p < 0.0001, n = 9). The baseline PD-L1 value is greater in the malignant than in the normal cells. (g–j) Immunohistochemistry of cultured A375 cells stained for PD-L1 demonstrates marked increase following ARSB siRNA (g) and decrease following exogenous ARSB (i), compared to the control siRNA (h) and control (j). (k,l) In the normal melanocytes and the A375 cells, silencing of the mannose-6-phosphate receptor (M6PR) by siRNA did not block the effect of exogenous ARSB on PD-L1 expression. In contrast, silencing of the calcium-independent insulin-like growth factor 2 receptor (IGF2R) by siRNA almost completely blocked the rhARSB-induced decline in PD-L1. p-values are <0.0001, determined by an unpaired t-test, two-tailed, with unequal variance and n ≥ 6. [ARSB = arylsulfatase B = N-acetylgalactosamine-4-sulfatase; consi = control siRNA; IGF2R = insulin-like growth factor 2 receptor; M6PR = mannose-6-phosphate receptor; PD-L1 = programmed death ligand-1; rh = recombinant human; si = small interfering siRNA; 0.2 = 0.2 mg/kg treatment dose]. **** represents p ≤ 0.0001, *** represents p ≤ 0.001, and * is for p ≤ 0.05. Scale bar is 150 um.