IQGAP1 suppresses TβRII-mediated myofibroblastic activation and metastatic growth in liver

Abstract

In the tumor microenvironment, TGF-β induces transdifferentiation of quiescent pericytes and related stromal cells into myofibroblasts that promote tumor growth and metastasis. The mechanisms governing myofibroblastic activation remain poorly understood, and its role in the tumor microenvironment has not been explored. Here, we demonstrate that IQ motif containing GTPase activating protein 1 (IQGAP1) binds to TGF-β receptor II (TβRII) and suppresses TβRII-mediated signaling in pericytes to prevent myofibroblastic differentiation in the tumor microenvironment. We found that TGF-β1 recruited IQGAP1 to TβRII in hepatic stellate cells (HSCs), the resident liver pericytes. Iqgap1 knockdown inhibited the targeting of the E3 ubiquitin ligase SMAD ubiquitination regulatory factor 1 (SMURF1) to the plasma membrane and TβRII ubiquitination and degradation. Thus, Iqgap1 knockdown stabilized TβRII and potentiated TGF-β1 transdifferentiation of pericytes into myofibroblasts in vitro. Iqgap1 deficiency in HSCs promoted myofibroblast activation, tumor implantation, and metastatic growth in mice via upregulation of paracrine signaling molecules. Additionally, we found that IQGAP1 expression was downregulated in myofibroblasts associated with human colorectal liver metastases. Taken together, our studies demonstrate that IQGAP1 in the tumor microenvironment suppresses TβRII and TGF-β dependent myofibroblastic differentiation to constrain tumor growth.

Figure 1. IQGAP1 interacts with TβRII and regulates its stability.

(A) Left: HSCs that express TβRII-HA by retroviral transduction were transduced with lentiviruses encoding nontargeting shRNA (NT shRNA, control) or IQGAP1 shRNAs, and subjected to WB for TβRII. Knockdown of IQGAP1 by 3 different shRNAs consistently upregulated TβRII protein levels. Middle: cells were transduced with retroviruses encoding YFP (control) or IQGAP1-YFP. Overexpression of IQGAP in HSCs reduced TβRII protein. Right: endogenous TβRII protein levels increased in IQGAP1-knockdown cells. (B) HSCs transduced with lentiviruses encoding either NT shRNA or IQGAP1 shRNA were harvested for RNA extraction and SYBR green–based real-time RT-PCR. IQGAP1 knockdown did not change TβRII mRNA levels. n = 3 independent experiments. (C) IQGAP1 (red) and TβRII-HA (green) colocalized at the plasma membrane (arrowheads) and in intracellular vesicles (arrows) in HSCs by IF. Scale bars 20 μm. (D) Left: TβRII coprecipitated with IQGAP1 when IP was performed using anti-IQGAP1. Middle: IQGAP1 coprecipitated with TβRII-HA when IP was performed using anti-HA. Right: IQGAP1 coprecipitated with endogenous TβRII when IP was performed using anti-TβRII. Data are representative of multiple repeats with similar results.

Figure 2. IQGAP1 C terminus aa 1503–1657 is required for binding and suppressing TβRII.

(A) Top 4 rows: full-length (FL) IQGAP1 and GST-fused truncated IQGAP1 proteins are shown. Bottom, GST fused truncated IQGAP1 proteins extracted from bacteria were incubated with HSC lysates for GST pull-down assays. Both aa 746–1657 and aa 1503–1657 of IQGAP1 bound to TβRII. Ponceau S staining depicted the purity of the recombinant proteins. (B) Top: after the GST tag of GST-TRII was removed by thrombin treatment, detagged TβRII was incubated with GST-fused IQGAP1 proteins for in vitro binding assays. Both aa 746–1657 and aa 1503–1657 of IQGAP1 bound to TβRII directly in vitro. Ponceau S staining depicted the purity of GST and GST-fused IQGAP1 proteins. Bottom: detagged IQGAP1 aa 746–1657 was incubated with GST or GST-TβRII for in vitro binding assays. GST-TβRII bound to IQGAP1 aa 746-1657 directly in vitro. aa 746–1657 instead of aa 1503–1657 of IQGAP1 was used in this assay because IQGAP1 antibodies could not recognize aa 1503–1657 of IQGAP1. Ponceau S staining depicted the purity of GST and GST fusion proteins. (C) HSCs expressing TβRII-HA were transduced with lentiviruses encoding GFP, IQGAP1-FLAG, or IQGAP1 (1-1502)-FLAG, and subjected to WB. In contrast with IQGAP1, IQGAP1 (1-1502) mutant lacking the TβRII binding region failed to repress TβRII protein levels. Densitometric ratios are shown on the bottom. All data shown represent multiple repeats with similar results.

Figure 3. IQGAP1 suppresses TGF-β–mediated activation of HSCs into myofibroblasts.

(A) HSCs that were transfected with control or IQGAP1 siRNA were serum-starved and stimulated with TGF-β1 (5 ng/ml) or PDGF-BB (20 ng/ml). HSC activation was assessed by WB for activation markers such as α-SMA, fibronectin, and p-SMAD2. TGF-β1 but not PDGF-BB induced upregulation of α-SMA and fibronectin. Two different IQGAP1 siRNAs consistently potentiated the effect of TGF-β1. (B) HSCs that were transfected with IQGAP1 siRNA were mixed with control cells for double IF for IQGAP1 (green) and α-SMA (red). As compared with nontransfected HSCs (asterisks), IQGAP1 knockdown induced the formation of α-SMA–positive stress fibers (arrows), indicative of myofibroblastic transdifferentiation of HSCs. Scale bar: 50 μm. (C) Left: HSCs treated as described in A were subjected to IF for α-SMA. TGF-β1 promoted the formation of α-SMA–positive stress fibers, and this effect was potentiated by IQGAP1 siRNA. Scale bar: 50 μm. Right: quantitative data of α-SMA IF showing that IQGAP1 knockdown significantly increased TGF-β1 activation of HSCs into myofibroblasts. *P < 0.05 by ANOVA. n = 7 randomly picked microscopic fields, each containing 100–200 cells. (D) HSCs that express GFP, IQGAP1-FLAG, or IQGAP1 (1-1502)-FLAG were stimulated with TGF-β1 and harvested for WB. IQGAP1 downregulated TβRII and inhibited HSC activation. In contrast, IQGAP1 (1-1502) mutant failed to suppress TβRII and TGF-β signaling. Data represent 3 independent experiments with similar results.

Figure 4. TGF-β1 increases IQGAP1/TβRII binding, and IQGAP1 knockdown inhibits lysosomal targeting of TβRII.

(A) HSCs that express TβRII-HA were serum starved and stimulated with TGF-β1 for indicated times. Cell lysates were subjected to IP using anti-IQGAP1, and coprecipitated TβRII was detected by WB. Densitometric ratios are shown on the bottom. TGF-β1 increased IQGAP1/TβRII binding. Blots represent 3 independent experiments. (B and C) HSCs that were serum starved and pretreated with cycloheximide (40 μg/ml) for 1 hour were incubated with TGF-β1 at 4°C for ligand/receptor binding. After cells were incubated at 37°C for indicated times, cells were fixed for double IF for HA (green) and LAMP1 (red). IQGAP1 knockdown significantly reduced TβRII in late endosome/lysosomes at both 30 and 60 minutes after TGF-β1 stimulation (arrowheads). Scale bar: 20 μm. **P < 0.01 by ANOVA. n = 6 cells each group. Data represent 3 independent experiments with identical results. (D and E) Cells treated as described in B and C were stained for HA and EEA-1. IQGAP1 knockdown induced the accumulation of TβRII in EEA-1–positive endosomes at both 30 and 60 minutes after TGF-β1 stimulation (arrows). Scale bar: 20 μm. **P < 0.01 by ANOVA. n = 6 cells each group. Data are representative of 3 independent repeats with identical results.

Figure 5. IQGAP1 knockdown inhibits TGF-β1 downregulation of TβRII, TβRII ubiquitination, and the plasma membrane targeting of SMURF1.

(A) Top: HSCs with their cell-surface proteins prelabeled with biotin were incubated with TGF-1 for indicated times and cells were harvested for streptavidin pull-down and TβRII WB to determine internalized TβRII. Bottom: TβRII degradation curves generated by densitometric analysis are shown. IQGAP1 knockdown inhibited TGF-β1 downregulation of cell surface TβRII. Chlo, chloroquine; Tν, half-life of TβRII. Data are representative of multiple independent experiments. Asterisks designate a point where TβRII was down to 50%. (B) HSCs expressing TβRI-FLAG were transduced with lentiviruses encoding either NT shRNA or IQGAP1 shRNA, and TβRI protein levels were detected by Flag WB. IQGAP1 knockdown increased TβRI-Flag in HSCs. n = 3 experiments with similar results. (C) TβRII-HA was precipitated from HSCs by IP using anti-HA; TβRII ubiquitination was detected by WB. IQGAP1 knockdown markedly inhibited TβRII ubiquitination. (D) Double IF for IQGAP1 (red) and SMURF1 (green) revealed that IQGAP1 and SMURF1 colocalized at the periphery plasma membrane in control cells (arrows, upper panels), and that IQGAP1 knockdown reduced the localization of SMURF1 at the plasma membrane (lower panels). Scale bar: 20 μm. (E) TβRII and SMURF1 colocalized at the peripheral plasma membrane (arrowheads, upper panels), and IQGAP1 knockdown reduced TβRII/SMURF1 colocalization at the plasma membrane (lower panels). Scale bar: 20 μm. (F) IQGAP1 knockdown reduced SMURF1 protein levels in HSCs by WB. β-actin WB was used as a loading control. n = 3 independent experiments with identical results.

Figure 6. Basal activation phenotype of HSCs of Iqgap1^–/– mice.

(A) Left: livers of 1-year-old Iqgap1^–/– and matched Iqgap1^+/+ mice were subjected to H&E staining, and double IF for desmin (red, HSC marker) and α-SMA (green, marker of activated HSCs). Cell nuclei were counterstained by TOTO-3 (blue). Arrows indicate colocalization of these 2 proteins. Scale bar: 50 μm. Right: quantitative data analyzed by ImageJ software revealed that α-SMA–positive HSCs were significantly increased in Iqgap1^–/– livers compared with Iqgap1^+/+ livers. **P < 0.01 by t test. (B) Left: liver samples as described in A were analyzed by WB for α-SMA and collagen I. Middle: densitometric analysis revealed that the average level of α-SMA or collagen I of Iqgap1^–/– livers was significantly higher than that of Iqgap1^+/+ livers. *P < 0.05; **P < 0.01 by ANOVA. Right: representative images of Sirius red staining are shown. Scale bar: 50 μm. (C) HSCs of mice were treated with TGF-β1 at 72 hours after isolation and harvested for WB. Iqgap1^–/– HSCs exhibited an enhanced activation phenotype as compared with Iqgap1^+/+ HSCs in vitro. n = 2 independent cell preparations using 4 mouse livers for each prep with similar results from both cell preparations.

Figure 7. IQGAP1 deficiency in the liver promotes myofibroblastic activation and lung liver metastases in mice.

(A) Depiction of portal vein implantation of LLCs into the livers of mice. (B) Left: average tumor weight of Iqgap1^–/– livers was significantly higher than that of Iqgap1^+/+ livers at 10 days after tumor implantation. *P < 0.05 by t test. Right: representative photographs of liver and liver metastases (mets) of mice are shown. (C) WB on isolated liver metastases revealed that the average level of α-SMA or TβRII of the liver metastases of Iqgap1^–/– mice was significantly higher than that of Iqgap1^+/+ mice. GAPDH WB was used as a protein loading control. *P < 0.05; **P < 0.01 by ANOVA. (D) Representative images of α-SMA IF (green) and H&E staining revealing more tumor-associated myofibroblasts in the liver metastases of Iqgap1^–/– mice as compared with Iqgap1^+/+ mice. Cell nuclei were counterstained by TOTO-3 (blue). MFs, tumor-associated myofibroblasts. Scale bar: 50 μM.

Figure 8. IQGAP1 deficiency in the liver promotes colorectal liver metastases in mice.

(A) 2 × 10⁶ MC38 mouse colorectal cancer cells were implanted into the livers of Iqgap1^–/– and matched Iqgap1^+/+ mice by portal vein injection, and mouse survival was analyzed by GraphPad Prism 5 software. **P < 0.01 by ANOVA. (B and C) MC38 cells tagged by firefly luciferase were injected into the livers of mice and bioluminescence of MC38 cells was quantitated by in vivo xenogen imaging at different days after tumor implantation. Data were analyzed by the GraphPad Prism 5 software, and representative images are shown. *P < 0.05 by ANOVA.

Figure 9. IQGAP1-knockdown HSCs promote colorectal tumor implantation and growth in HSC/tumor coimplantation model.

(A) 0.5 × 10⁶ HT-29 human colorectal tumor cells were mixed with 0.5 × 10⁶ control HSCs (HSC-NTshRNA) or 0.5 × 10⁶ IQGAP1-knockdown HSCs (HSC-IQGAP1shRNA), respectively, and coimplanted into nude mice via subcutaneous injection. Tumor nodules were measured by a caliper at different days after implantation, and data were analyzed by the GraphPad Prism 5 software. IQGAP1-knockdown HSCs exhibited a greater tumor-promoting effect as compared with control HSCs. *P < 0.05 by ANOVA. (B) 0.5 × 10⁶ HT-29 cells tagged by firefly luciferase were mixed with 0.5 × 10⁶ control HSCs or 0.5 × 10⁶ IQGAP1-knockdown HSCs, respectively, and coimplanted into nude mice via subcutaneous injection. Bioluminescence of HT-29 cells was quantitated by in vivo xenogen imaging at indicated days after tumor implantation, and data were analyzed by GraphPad Prism 5 software. Imaging of representative mice and quantitative data are shown. IQGAP1-knockdown HSCs promoted the implantation of HT-29 cells in mice as compared with control HSCs. *P < 0.05 by ANOVA. (C) HSCs tagged by firefly luciferase were implanted into nude mice alone or with HT-29 tumor cells via subcutaneous injection. Bioluminescence of HSCs was quantitated by in vivo xenogen imaging at different days after implantation. Data are representative of 6 mice with consistent results. HSCs were able to survive up to 23 days in mice after HSC/tumor coimplantation.

Figure 10. IQGAP1 knockdown in HSCs promotes TβRII levels, myofibroblastic activation of HSCs in a HSC/tumor coimplantation model.

(A) HSCs expressing TβRII-HA were mixed with HT-29 tumor cells and coimplanted into nude mice via subcutaneous injection. Tumor nodules were subjected to IF for HA (green) and α-SMA (red) and H&E staining. Cells positive for both HA and α-SMA were detected in the tumor nodules (arrows). Cell nuclei were counterstained by TOTO-3 (blue). S, stroma; T, tumor cells. Scale bar: 50 μm. (B) Tumor nodules were subjected to WB and densitometric analysis. The average level of TβRII-HA or α-SMA in tumors arising from HT-29/HSC-IQGAP1shRNA coimplantation was significantly higher than that in tumors arising from HT-29/HSC-NTshRNA coimplantation. **P < 0.01 by ANOVA. (C) Representative HA IF (green) and H&E staining revealing more TβRII-HA–positive tumor-associated myofibroblasts in tumor nodules arising from HT-29/HSC-IQGAP1shRNA coimplantation as compared with tumors arising from HT-29/HSC-NTshRNA coimplantation. Cell nuclei were counterstained by TOTO-3 (blue). Scale bar: 50 μm. (D) Tumor nodules were subjected to α-SMA IF (green), H&E staining, and quantitative analysis. The average α-SMA IF density in tumors arising from HT-29/HSC-IQGAP1shRNA coimplantation was significantly higher than that in tumors from HT-29/HSC-NTshRNA coimplantation. *P < 0.05 by t test. Scale bar: 100 μm.

Figure 11. HSCs are activated into tumor-associated myofibroblasts of liver metastases.

(A) L3.6 human gastrointestinal cancer cells were implanted into the livers of SCID mice by portal vein injection. Established liver metastases were isolated for H&E staining and double IF for Stem121 (green), a specific marker of human engraftments, and α-SMA (red). Tumor-associated myofibroblasts of liver metastases were negative for Stem121 (arrows). Cell nuclei were counterstained by TOTO-3 (blue). Scale bar: 100 μm. (B) Adjacent sections of L3.6 liver metastases were subjected to H&E staining, immunostaining for Stem121, α-SMA, or desmin. Some tumor-associated myofibroblasts were positive for desmin (arrows). Scale bar: 50 μm. (C) Liver sections containing L3.6 micrometastases were subjected to H&E staining and double IF for α-SMA (green) and desmin (red). Some HSCs at the sinusoids adjacent to L3.6 tumor cells were activated to express α-SMA (arrowheads). Cell nuclei were counterstained by TOTO-3 (blue). Scale bars: 50 μm. (D) Adjacent sections of micrometastases shown in C were subjected to double IF for Stem 121 and α-SMA or desmin. The activated HSCs adjacent to the tumor cells were negative for Stem 121 (arrows). Cell nuclei were counterstained by TOTO-3 (blue). Scale bar: 50 μm.

Figure 12. IQGAP1-knockdown HSCs promote the proliferation, migration, and survival of tumor cells.

(A) Conditioned medium collected from control and IQGAP1-knockdown HSCs were used as a growth stimulant for HT-29 in nonradioactive cell proliferation assays. Conditioned medium of IQGAP1-knockdown HSCs promoted the proliferation of HT-29 cells as compared with that of control shRNA–transduced HSCs. CM, conditioned medium. *P < 0.05 by ANOVA; n = 3 repeats with similar results. (B) Conditioned medium collected as described in A were used as a chemoattractant for HT-29 in Boyden chamber assays. Conditioned medium of IQGAP1-knockdown HSCs promoted HT-29 migration as compared with that of control shRNA–transduced HSCs. *P < 0.05 by ANOVA; n = 3 repeats with similar results. Scale bar: 100 μm. (C) MC38 cells were suspended in basal medium or conditioned medium as described in A and seeded onto polyhydroxyethylmethacrylate (poly-HEMA) precoated culture dishes. After cells were incubated for 24 hours with gentle shaking, anoikis was assessed by DAPI staining of unfixed cells (top) and WB for PARP cleavage (lower right). Conditioned medium of IQGAP1-knockdown HSCs protected MC38 cells from anoikis as compared with that of control HSCs. *P < 0.05 by ANOVA. n = 3 repeats. Scale bar: 50 μm. (D) Control and IQGAP1-knockdown HSCs were harvested for RNA extraction and SYBR green–based real-time RT-PCR for SDF-1/CXCL12 and HGF. The mRNA level of SDF-1/CXCL12 or HGF was significantly increased by IQGAP1 knockdown in HSCs. *P < 0.05 by t test; n = 3 independent experiments.

Figure 13. IQGAP1 in the myofibroblasts of human colorectal liver metastases is downregulated.

(A) Double IF for IQGAP1 and α-SMA was performed on liver biopsies of 29 patients with metastatic colorectal cancer. IQGAP1 IF intensities in the myofibroblasts of liver metastases and matched control liver were quantitated by ImageJ software. MFs, myofibroblasts. (B) Box and whisker plots revealing that IQGAP1 in the myofibroblasts of patient colorectal liver metastases was significantly lower than that of the matched livers. **P < 0.01 by t test. (C) IF and H&E staining of a representative patient are shown. Cell nuclei were counterstained by TOTO-3 (blue). Scale bar: 100 μm. (D) Conditioned media collected from HT-29, CT26, and MC38 colorectal cancer cells were incubated with HSCs for 24 hours. IQGAP1 protein levels of HSCs were determined by WB and densitometric analyses. Conditioned medium of colorectal tumor cells downregulated IQGAP1 of HSCs. TGF-β1 (5 ng/ml) recapitulated the effect of the conditioned medium. Data represent multiple experiments with similar results.

Figure 14. IQGAP1 of HSCs suppresses TGF-β activation of HSCs into myofibroblasts, and this effect is counterbalanced by tumor-derived factors.

The C-terminal aa 1503–1657 of IQGAP1 binds to TβRII in a manner that is enhanced by TGF-β stimulation. IQGAP1 recruited to the TGF-β receptor complexes promotes SMURF1/TβRII colocalization at the plasma membrane, TβRII ubiquitination, and lysosomal and proteasomal degradation of TβRII. Thus IQGAP1 regulates TβRII degradation and cellular protein abundance. IQGAP1 binding and repression of TβRII suppresses activation of HSCs into myofibroblasts, thus limiting liver metastatic growth. Tumor-derived factors including TGF-β1, however, are able to downregulate IQGAP1 of HSCs, thereby amplifying the TGF-β1 activation of HSCs into tumor-associated myofibroblasts, which in turn, further promote liver metastatic growth by upregulating paracrine factors such as SDF-1/CXCL12 and HGF.