Josephin Domain Containing 2 (JOSD2) inhibition as Pan-KRAS-mutation-targeting strategy for colorectal cancer

Abstract

KRAS is the most common mutated oncogenes in colorectal cancer (CRC), yet effective therapeutic strategies for targeting multiple KRAS mutations remained challenging. The prolonged protein stability of KRAS mutants contribute to their robust tumor-promoting effects, but the underlying mechanism is elusive. Herein by screening deubiquitinases (DUBs) siRNA library, we identify Josephin domain containing 2 (JOSD2) functions as a potent DUB that regulates the protein stability of KRAS mutants. Mechanistically, JOSD2 directly interacts with and stabilizes KRAS variants across different mutants, by reverting their proteolytic ubiquitination; while KRAS mutants reciprocally inhibit the catalytic activity of CHIP, a bona fide E3 ubiquitin ligase for JOSD2, thus forming a JOSD2/KRAS positive feedback circuit that significantly accelerates KRAS-mutant CRC growth. Inhibition of JOSD2 by RNA interference or its pharmacological inhibitor promotes the polyubiquitination and proteasomal degradation of KRAS mutants, and preferentially impede the growth of KRAS-mutant CRC including patient-derived cells/xenografts/organoids (PDCs/PDXs/PDOs) over that harboring wild-type KRAS. Collectively, this study not only reveals the crucial roles of JOSD2/KRAS positive feedback circuit in KRAS-mutant CRC, but also provides a rationale to target JOSD2 as the promising pan-KRAS-mutation-targeting strategy for the treatment of a broad population of CRC patients with KRAS variant across different mutant types.

Fig. 1. DUBs siRNA library screening identifies JOSD2 as a crucial DUB regulating KRAS protein stability and activity.

a The schematic diagram of screening the candidate DUBs of regulating KRAS stability by using KRAS^G12V-luciferase reporter system and DUBs siRNA library (The schematic diagram was created in BioRender. https://BioRender.com/w35e177). The KRAS^G12V-luciferase reporter system was constructed by cloning the KRAS^G12V coding sequences in front of the firefly-luciferase sequence, thus forming a fusion gene of KRAS^G12V-Luciferase. CDS coding sequences, DUB deubiquitinases. b JOSD2 siRNA pools significantly minimized the luciferase activity of the KRAS^G12V-luciferase reporter system. c JOSD2 depletion inhibited the luciferase activity of the KRAS^G12V-luciferase reporter system in a dose-dependent manner. (means ± SD, n = 3; three independent experiments were performed). d, e JOSD2, but not JOSD1, positively regulated KRAS mutants’ protein levels. f JOSD2 depletion dramatically decreased KRAS mutants’ stability. g JOSD2 depletion downregulated KRAS levels through the ubiquitin-proteasome systems but not the autophagy-lysosome system. h, i JOSD2, but not JOSD1, positively regulated MAPK signaling pathways. WT, wild-type; Mut, mutation. The significance analysis of (c) was conducted by One-way ANOVA (Bonferroni method was utilized to correct for multiple comparisons). The samples derived from the same experiment and processed on different gels in parallel are shown below: d SW480(JOSD2): one gel for GAPDH/KRAS, another for JOSD2; e SW480(shJOSD2) & SW620(shJOSD2): one gel for GAPDH/JOSD2, another for KRAS. f Different gels for GAPDH and Myc(KRAS); g one gel for GAPDH/JOSD2, another for KRAS; h SW480(shJOSD1): one gel for p-MEK/GAPDH/MEK, another for p-ERK/JOSD1/ERK; SW620(shJOSD1): one gel for p-ERK/MEK, another for p-MEK/ERK, another for GAPDH/JOSD1; SW480(shJOSD2): one gel for p-MEK/JOSD2/GAPDH, another three for ERK, p-ERK and MEK, respectively; SW620(shJOSD2): one gel for p-MEK/GAPDH, another for p-ERK/JOSD2, another two for ERK and MEK, respectively; i SW480(JOSD1): one gel for p-MEK/KRAS/GAPDH/MEK, another for p-ERK/JOSD1, another for ERK; SW620(JOSD1): one gel for p-ERK/KRAS/p-MEK/GAPDH, another for MEK/JOSD1, another for ERK; SW480(JOSD2): one gel for ERK/KRAS, another for p-MEK/GAPDH, another for p-ERK/JOSD2, another for MEK; SW620(JOSD2): one for ERK/KRAS, another for p-MEK/GAPDH, another for p-ERK/JOSD2, another for MEK. Source data are provided as a Source Data file.

Fig. 2. JOSD2 deubiquitinates and stabilizes KRAS in an enzyme activity-dependent manner.

a The ability of KRAS^{G12C/G12D/G12V} to bind to JOSD2 was stronger than that of KRAS^WT. WT, wild-type; IP, immunoprecipitation. b The interaction between JOSD2 and KRAS was independent of the enzyme activity of JOSD2. CA&HY, JOSD2-C24A&H125Y. c–e JOSD2, not JOSD1, interacted with KRAS at semi-endogenous levels and endogenous levels. f In vitro GST pull-down assays to examine the interaction between JOSD2 and KRAS^{WT/G12C/G12D/G12V}. 293T cells were transfected with the indicated expression plasmids for 24 h. Cell lysates were incubated with the Streptavidin magnetic beads conjugated with bacterial-expressed recombinant GST or GST-JOSD2 protein. Proteins retained on Streptavidin magnetic beads were subjected to WB analysis. g, h JOSD2, not JOSD1, removed the polyubiquitination chains of KRAS^{WT/G12C/G12D/G12V} in an enzyme activity-dependent manner. i JOSD2 depletion promoted the polyubiquitination of KRAS^{WT/G12C/G12D/G12V}. j JOSD2 removed the polyubiquitination chains of KRAS^{WT/G12C/G12D/G12V} in vitro (The schematic diagram was created in BioRender. https://BioRender.com/w35e177). 293T cells were co-transfected with indicated expression plasmids, and cell lysates were incubated with anti-Flag affinity gel overnight. Then, proteins retained on anti-Flag affinity gel were incubated with bacterial-expressed recombinant human JOSD2 (rhJOSD2) protein at 37 °C for 3 h, followed by WB analysis. k JOSD2, not JOSD1, removed the polyubiquitination chains of KRAS^G12V promoted by LZTR1. The samples derived from the same experiment and processed on different gels in parallel were shown below: d Input: one gel for JOSD1/JOSD2/GAPDH, another for Flag(G12V), IP: one gel for JOSD1/Flag(G12V), another for JOSD2; e SW620: Input: one gel for JOSD1/KRAS/GAPDH, another for JOSD2, IP: one gel for JOSD1/KRAS, another for JOSD2; HCT-116: Input: one gel for JOSD2/GAPDH, another for JOSD1/KRAS, IP: different gels for JOSD1, KRAS and JOSD2; f one gel for KRAS/GST-JOSD2, another for GST; g Input of KRAS-WT/G12C: one gel for HA(ub)/GAPDH/ Flag(KRAS), another for JOSD1&2, Input of KRAS-G12D/G12V: one gel for HA(ub)/GAPDH/JOSD1&2, another for Flag(KRAS); i Input of KRAS-WT/G12C/G12D/G12V: one gel for HA(ub)/GAPDH/JOSD2, another for Flag(KRAS); k Input: one gel for HA(ub)/GAPDH, another for Flag(KRAS), another for Myc(LZTR1)/JOSD1&JOSD2. Source data are provided as a Source Data file.

Fig. 3. JOSD2 inhibition significantly impedes KRAS-mutant patient-derived CRC growth in vitro/in vivo.

a The schematic diagram of PDCs and PDXs models (The schematic diagram was created in BioRender. https://BioRender.com/w35e177). CRC, colorectal cancer; PDC, patient-derived cells; PDX, patient-derived xenografts. b–e Knocking down JOSD2 significantly inhibited the tumor growth of PDXs in vivo. The tumors were intratumorally injected with lentivirus-encoded scramble shRNA (Ctrl) or JOSD2 shRNA (shJOSD2) every two days (means ± S.E.M., n = 6 mice/group for PDX-1 xenografts; n = 7 mice/group for PDX-2 xenografts). b The bearing-tumor mouse and tumor images. c The relative tumor volume of indicated groups. Tumors were measured every two days. d The tumor weight of the indicated groups. e Knocking down JOSD2 downregulated intratumor KRAS levels and inhibited MAPK signaling pathways of PDXs. The protein isolated from the tumor was subjected to WB analysis (means ± S.E.M., n = 6 mice/group). The significance analysis of c–e was conducted by Two-tailed unpaired Student’s t-tests. The samples derived from the same experiment and processed on different gels in parallel are shown below: e one gel for p-MEK/GAPDH/JOSD2/MEK, another for p-ERK/ERK/KRAS. Source data are provided as a Source Data file.

Fig. 4. KRAS mutants significantly upregulate JOSD2 by inhibiting the E3 ligase activity of CHIP.

a The intratumor JOSD2 levels of KRAS-mutant CRC patients were significantly higher than those of KRAS wild-type CRC patients. IHC, immunochemistry; WT, wild-type; Mut, mutation. (means ± SD, n = 6 patients/group; Scale bar, 100 μm). b, c KRAS mutants, not KRAS wild-type, positively regulated JOSD2 levels. d KRAS mutants posed little effect on JOSD2 mRNA levels (means ± SD, n = 3; three independent experiments were performed). e Overexpression of KRAS mutants, not KRAS wild-type, increased JOSD2 stability. CHX, cycloheximide. f KRAS mutants decreased the polyubiquitination levels of JOSD2. g There was no significant difference in the ubiquitination levels of JOSD2-WT and C24A when immunoprecipitated with KRAS-G12V. WB, western blot. h K48, K142, and K181 are three main ubiquitination sites on JOSD2 (The schematic diagram was created in BioRender. https://BioRender.com/w35e177). i The stability of JOSD2-3KR was significantly higher than JOSD2-WT. j KRAS mutants significantly increased the protein levels and stability of JOSD2-WT, but not that of JOSD2-3KR. DOX, doxycycline. k, l JOSD2 interacted with CHIP in vivo/in vitro. m, n CHIP positively regulated the polyubiquitination levels of JOSD2. o CHIP negatively regulated JOSD2 levels. p Overexpression of CHIP significantly decreased the protein levels of JOSD2-WT, but not of JOSD2-3KR. q Depletion of CHIP prevented the further upregulation of JOSD2 levels induced by KRAS^G12V. (means ± SD, n = 3; three independent experiments were performed). The significance analysis of (a) was conducted by Two-tailed unpaired Student’s t-tests, and (d, q) were conducted by One-way ANOVA (Bonferroni method was utilized to correct for multiple comparisons). The samples derived from the same experiment and processed on different gels in parallel were shown below: b one gel for Flag(KRAS)/GAPDH, another for JOSD2; f Input: one gel for HA(ub)/GAPDH/Flag(JOSD2), another for Myc(KRAS); k Input: different gels for GAPDH and Flag(CHIP); IP: different gels for JOSD2 and Flag(CHIP); l different gels for GST and Myc(CHIP); q one gel for CHIP/Flag(G12V), another for GAPDH/JOSD2. Source data are provided as a Source Data file.

Fig. 5. Targeting JOSD2 poses a more potent anti-tumor effect on KRAS-mutant CRC over that of KRAS wild-type in vitro/in vivo.

a, b JOSD2 depletion inhibits the cell proliferation of KRAS-mutant cells strongly than KRAS wild-type cells in vitro. HT-29^KRAS-WT/Mut cells were infected with indicated lentiviruses for 96 h, then these cells were seeded with 2000 cells per well in 96-well/6-well plates, followed by SRB staining. Ctrl, control. a The representative cells proliferation images of the indicated group; Scale bar, 6 mm. b The quantitative analyses of cells proliferation, P = 0.0965, WT-Ctrl vs. WT-shJOSD2; P = 4E-8, G12C-Ctrl vs. G12C-shJOSD2; P = 6E-8, G12D-Ctrl vs. G12D-shJOSD2; P = 1E-6, G12V-Ctrl vs. G12D-shJOSD2. (means ± SD, n = 3; three independent experiments were performed). c, d JOSD2 depletion inhibits the colony formation of KRAS-mutant cells strongly than KRAS wild-type cells in vitro (means ± SD, n = 3; three independent experiments were performed). c The representative colony formation images of the indicated group; Scale bar, 5 mm. d The quantitative analyses of colony formation. e–h JOSD2 depletion posed a more potent anti-tumor effect on KRAS-mutant xenografts compared to KRAS wild-type in vivo (means ± S.E.M., n = 7 mice/group). e The tumor images of the indicated groups. f The tumor volume of the indicated groups. Tumors were measured every two days. g The tumor weight of the indicated groups. h Knocking down JOSD2 downregulated intratumor KRAS levels of KRAS-mutant xenografts are stronger than those of KRAS wild-type xenografts. The proteins isolated from the tumor were subjected to WB analysis. The significance analysis of (b, d, f, g) was conducted by One-way ANOVA (Bonferroni method was utilized to correct for multiple comparisons). The samples derived from the same experiment and processed on different gels in parallel are shown below: h one gel for JOSD2/GAPDH and another for KRAS. Source data are provided as a Source Data file.

Fig. 6. JOSD2 catalytic inhibitor HY041004 preferentially impedes KRAS-mutant CRC growth in vitro/in vivo.

a HY041004 reversed the effects of JOSD2 deubiquitinating KRAS mutants. b HY041004 downregulated KRAS mutants’ protein levels in a concentration-dependent manner. c HY041004 inhibited the cell proliferation of KRAS-mutant CRC cells in vitro. WT, wild-type; Mut, mutation. (means ± SD, n = 3; three independent experiments were performed). d HY041004 inhibited the colony formation of KRAS-mutant CRC cells in a concentration-dependent manner in vitro. SW480/SW620 cells were treated with the indicated concentration of HY041004 for 72 h (SW480: 0, 0.03, 0.06, 0.12 μM; SW620: 0, 0.06, 0.09, 0.12 μM), then replaced with fresh medium and cultured for another 14 days followed by colony formation assays; Scale bar, 5 mm. e The inhibitory effects of HY041004 on the cells proliferation of KRAS-mutant cells are stronger than KRAS wild-type cells in vitro. P, vs. KRAS-WT. (means ± SD, n = 3; three independent experiments were performed). f The inhibitory effects of HY041004 on the colony formation of KRAS-mutant cells are stronger than KRAS wild-type cells in vitro; Scale bar, 5 mm. g–i HY041004 posed more potent anti-tumor effects on KRAS-mutant xenografts compared to KRAS wild-type in vivo (means ± S.E.M., n = 7 mice/group). g The tumor volume of the indicated groups. h The tumor weight of the indicated groups. i HY041004 downregulated intratumor KRAS levels of KRAS-mutant xenografts stronger than those of KRAS wild-type xenografts. j The represent images of CRC PDO. Scale bar, 200 μm. k HY041004 significantly downregulated KRAS levels in PDOs (patient-derived organoids). CRC, colorectal cancer; Scale bar, 100 μm. l HY041004 posed more potent anti-tumor activity on KRAS-mutant CRC PDO in vitro. (means ± SD, n = 3; three independent experiments were performed). The significance analysis of (e, g, h) was conducted by One-way ANOVA (Bonferroni method was utilized to correct for multiple comparisons), and (i, l) was conducted by Two-tailed unpaired Student’s t-tests. Source data are provided as a Source Data file.

Fig. 7. JOSD2 is elevated in KRAS-mutant CRC tissues and positively correlated with KRAS expression.

a, b KRAS-mutant (Mut) and wild-type (WT) colorectal cancer (CRC) tissues were performed IHC (immunohistochemistry) assays and stained with KRAS and JOSD2. Representative images of IHC staining of KRAS and JOSD2 in KRAS-Mut and WT CRC tissues (n = 36 patients/group) were shown; Scale bar, 1.5 mm. c The correlation analysis between JOSD2 and KRAS was performed, and the r value was calculated by the two-tailed Spearman correlation analysis via GraphPad Prism (P < 0.001, r = 0.6147, 95% confidence interval 0.4410~0.7440). IHC score was evaluated by multiplication of positive staining proportions (1 score, <25%; 2 score, 25%–50%; 3 score, 50%-75%; 4 score, 75%–100%) and intensity of protein expressions (1 score, weak staining; 2 score, moderate staining; 3 score, high staining). –, negative expression (1–3 score); +, low expression (4–6 score); ++, medium expression (7–9 score); +++, high positive expression (10–12 score). d JOSD2-high expression CRC displayed higher KRAS levels compared to JOSD2-low expression tissues. e KRAS-mutant CRC tissues displayed higher JOSD2 levels compared to KRAS wild-type CRC tissues. f The schematic diagram of JOSD2/KRAS positive feedback loop promoting the proliferation of KRAS-mutant CRC (The schematic diagram was created in BioRender. https://BioRender.com/w35e177). Source data are provided as a Source Data file.