Therapeutic targeting of syndecan-1 axis overcomes acquired resistance to KRAS-targeted therapy in gastrointestinal cancers

Abstract

The therapeutic benefit of recently developed mutant KRAS (KRAS∗) inhibitors remains limited by the rapid onset of resistance. Here, we aim to delineate mechanisms underlying acquired resistance and identify actionable targets for overcoming this clinical challenge. Previously, we identified syndecan-1 (SDC1) as a key effector for pancreatic cancer progression whose surface expression is driven by KRAS∗. By leveraging both pancreatic and colorectal cancer models, we show that surface SDC1 expression initially diminishes upon KRAS∗ inhibition but recovers in tumor cells that bypass KRAS∗ dependency. Mechanistically, we reveal that YAP1 activation drives the recovery of SDC1 surface localization to enhance macropinocytosis-mediated nutrient salvaging and activation of multiple receptor tyrosine kinases for tumor maintenance, promoting resistance to KRAS∗-targeted therapy. Overall, our study provides a strong rationale for targeting the YAP-SDC1 axis to overcome resistance to KRAS∗ inhibition, thereby revealing promising therapeutic opportunities for improving the clinical outcome of patients with KRAS∗-mutated cancers.

Keywords: KRAS inhibitor; colorectal cancer; macropinocytosis; pancreatic cancer; syndecan; therapy resistance.

Graphical abstract

Figure 1

SDC1 membrane expression recovers upon acquired resistance to KRAS∗-MAPK signaling blockade (A) Experimental design for the development of PDAC tumors resistant to KRAS∗ blockade from iKras p53^L/+ mouse models (top). H&E staining and immunohistochemistry for SDC1 and p-ERK of inducible Kras (iKras, left bottom) as well as iKras-inactivated relapsed tumors (Escaper, right bottom). Scale bar, 100 μm. (B) SDC1 expression levels in the iKras p53^L/+ tumor cells in the presence (iKras ON) or absence (iKras OFF) of doxycycline as well as in iKras-independent recurrent tumor cells (Escaper: E1–, E2–, E3–, E4–) and iKras-reactivated recurrent tumor cells (E5+, E6+). (C) Surface SDC1 expression levels of iKras p53^L/+ tumor cells (AK10965) upon doxycycline withdrawal for the indicated days. (D) Human PDAC cells, AsPC-1, HuP-T4, and MIA PaCa-2, were treated with trametinib (20 nM) for the indicated time periods, and surface protein levels of SDC1 were measured. (E) Human PDAC MIA PaCa-2 cells cultured were treated with AMG510 (20 nM) for the indicated time periods, and surface protein levels of SDC1 were measured. (F) Surface SDC1 expression levels of AMG510-R cells derived from MIA PaCa-2, SW837, and LIM2099 cells chronically treated with AMG510 or vehicle control. n = 3 biological replicates. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. Data are presented as mean ± SD.

Figure 2

SDC1 reconstitution enables tumor growth independent of the oncogenic KRAS-MAPK pathway (A) Orthotopic xenografts generated from iKras p53^L/+ PDAC cells. Post implantation, mice were kept on DOX for 7 days and then switched to off DOX (to induced Kras^G12D inactivation). Tumor growth was visualized by bioluminescent imaging at 4 weeks off DOX for GFP-, mouse wild-type SDC1-GFP, and mouse soluble (mSol) SDC1-GFP-expressing groups or at 2 weeks for the KRAS^G12V-expressing group. (B) Clonogenic assay of SDC1 knockdown with shSdc1 or scramble control (Scr) in SDC1-bypass cells derived from SDC1-bypass tumors shown in (A). n = 3 biological replicates. (C) Clonogenic assay of MIA PaCa-2 cells and AMG510-resistant derivative (AMG510-R), developed under continuous 500 nM AMG510 treatment, following shRNA-mediated knockdown of Sdc1 or scramble control (Scr) in the presence of 500 nM AMG510. n = 3 biological replicates. (D) Subcutaneous tumor growth of parental MIA PaCa-2 or AMG510-R cells transduced with shSDC1 or scramble control (Scr), treated with AMG510 (A, 30 mg/kg) or vehicle (V). Treatment started when tumors reached ∼150 mm³, and tumor sizes were measured every 2 days. n = 3 per group. (E) Representative photographs of dissected tumors from mice of Figure 2D. (F and G) Immunohistochemistry analysis for SDC1 (F) and Ki-67 (G) in tumors from Figure 2E. Scale bar, 50 μm. Quantification was from ten random fields. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n.s., non-significant. Data are presented as mean ± SD.

Figure 3

SDC1 mediates the activation of multiple receptor tyrosine kinases in KRAS-bypass tumors (A) RPPA heatmap of genes enriched in SDC1-bypass (SB) cells derived from AK192 or AK10965 tumors compared to parental cells GN (GFP-no DOX for 24 h), GD (GFP-DOX), SN (SDC1-no DOX for 24 h), and SDC1-DOX. Expression levels shown are representative of log2 values of each replicate. (B) Phospho-RTK array analysis of parental and SB cells as described in (A). (C) Phospho-RTK array analysis of MIA PaCa-2 and AMG510-R cells infected with scrambled shRNA or shRNA against Sdc1. (D) Western blot analysis of MIA PaCa-2 and AMG510-R cells infected with scramble shRNA or shRNA against Sdc1. (E) Western blot analysis of SB cells derived from individual AK192 SB-3 and AK192 SB-4 tumors and infected with scrambled shRNA or shRNA against Sdc1.

Figure 4

YAP1 drives the recovery of SDC1 expression through repressing a guanine-nucleotide-activating protein to reactivate ARF6 activity (A and B) ARF6 activity was measured in iKras p53^L/+ tumor cells (AK192) cultured with (ON) or without (OFF) doxycycline for the indicated days (A), as well as in their derived SB tumor cells and Escaper tumor cells (E1– and E2– are KRAS∗ independent, and E5+ is KRAS∗ reactivated). Input lysates were immunoblotted to validate expression of ARF6 and β-actin. (C) Western blot analysis of iKras p53^L/+ tumor cells (AK192) grown in the absence of DOX for indicated time periods. (D) mRNA expression of ARF6 GAPs and GEFs in the E+ and E− cell microarray dataset during KRAS-reactivated and -independent relapse. n = 4 biological replicates. (E) Western blot analysis of MIA PaCa-2 cells treated with AMG510 (20 nM) for indicated time periods. (F) Cell lysates of MIA PaCa-2 AMG510-R cells expressing ectopic Asap2 (Asap-OE) or empty vector (Vec) were immunoblotted for ASAP2 and β-actin. (G) Surface SDC1 expression levels of MIA PaCa-2 AMG510-R cells expressing ectopic Asap2 (blue peak) or empty vector (gray peak) were measured. n = 3 biological replicates. (H) Western blot analysis of iKras p53^L/+ tumor cells (AK192, DOX ON and OFF for ≥2 weeks) infected with scrambled shRNA or shRNA against Yap1. (I) Surface SDC1 expression levels of iKras p53^L/+ tumor cells (AK192, DOX ON and OFF for ≥2 weeks) infected with scrambled shRNA (gray peak) or shRNAs against Yap1 (blue and red peaks). n = 3 biological replicates. (J) Clonogenic growth assay of iKras p53^L/+ tumor cells (AK192) infected with scrambled shRNA or shRNA against Yap1 upon KRAS∗ activation (DOX ON) or KRAS∗ inactivation (DOX OFF for ≥2 weeks). n = 3 biological replicates. (K) Western blot analysis of iKras p53^L/+ tumor cells (AK14838) with CRISPR-Cas9-mediated Yap1 knockout (sgYap1) or with non-targeting sgRNA (sgRNA-Ctrl) grown in absence of DOX for indicated time periods. (L) Surface SDC1 expression levels of iKras p53^L/+ tumor cells (AK14838) or their derived cells with CRISPR-Cas9-mediated Yap1 knockout (sgYap1) or control (sgRNA-Ctrl) grown in the absence of DOX for indicated days. n = 3 biological replicates. (M) iKras p53^L/+ tumor cells (AK14838) with Cas9-mediated Yap1 knockout (sgYap1) or sgRNA control (sgRNA-Ctrl) were grown in the presence (ON) or absence (OFF) of DOX. Cell proliferation was measured using the live-cell time-lapse imaging module of Incucyte. Black line: sgRNA-Ctrl ON; red line: sgRNA-Ctrl OFF; gray line: sgYap1 ON; blue line: sgYap1 OFF. n = 6 biological replicates. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n.s., non-significant. Data are presented as mean ± SD.

Figure 5

The YAP1-SDC1 axis is required for macropinocytic activity in tumor cells resistant to KRAS∗ inhibition (A and B) Macropinocytosis was visualized with TMR-dextran (red fluorescence) (A; scale bar, 20 μm) and quantified (B; n = 10 areas) in iKras p53^L/+ tumor cells as well as Escaper cells (E1– is KRAS∗ independent and E+ is KRAS∗ reactivated) grown in the presence (ON) or absence (OFF) of doxycycline for 3 days. (C) Glutamine deprivation assay of MIA PaCa-2 parental and AMG510-R cells. Cells are grown in either one of the following conditions: 0.2Q (0.2 mM glutamine), 0.2 mM glutamine with 2% albumin (Alb), or 0.2 mM glutamine and 2% Alb with 25 μM EIPA treatment. Values are presented as relative number of viable cells at the time endpoint of the assay. (D) Clonogenic assay of parental and AMG510-R MIA PaCa-2 cells treated with either DMSO (Ctrl) or 50 μM EIPA. (E and F) Macropinocytosis was visualized with TMR-dextran (red fluorescence) (E; scale bar, 20 μm) and quantified (F; n = 10 areas) in E1– cells infected with two different Sdc1 and Yap1-targeting shRNA or scrambled shRNA control. (G and H) MIA PaCa-2 and AMG510-R cells stably expressing ectopic SDC1 or empty vector were infected with scrambled shRNA or shRNA against Sdc1 or Yap1. Macropinocytosis was visualized with TMR-dextran (red fluorescence) (G; scale bar, 20 μm) and quantified (H; n = 10 areas). (I) RAC1 activity in MIA PaCa-2 and AMG510-R cells stably expressing Rac1-dominant negative (RAC1-DN) or empty vector (Vec). (J and K) Macropinocytosis in MIA PaCa-2 and AMG510-R cells stably expressing RAC1-DN or empty vector (Vec) was visualized with TMR-dextran (J; red fluorescence, scale bar, 20 μm) and quantified (K; n = 10 areas). n = 3 biological replicates. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; n.s., non-significant. Data are presented as mean ± SD.

Figure 6

Targeting the YAP1-SDC1 axis abolishes acquired resistance to mutant KRAS-targeted therapy (A) Western blot analysis of MIA PaCa-2 or AMG510-R cells treated with 1 or 5 μM of YAP1 ASOs or control ASO for 72 h. (B) Surface SDC1 expression levels of MIA PaCa-2 and AMG510-R cells treated with 5 μM of YAP1-ASOs or control ASO. (C and D) Clonogenic assay of MIA PaCa-2 (C) and AMG510-R (D) cells treated with control ASO (Ctrl) or YAP1-ASOs. n = 3 biological replicates. (E) MIA PaCa-2 and AMG510-R cells were subcutaneously injected into 5-week-old female NGS mice. Mice were treated with control ASOs (Ctrl, 50 mg/kg, SQ, QD [5 days ON, 2 days OFF]) or YAP1-ASOs (ASO, 50 mg/kg, SQ, QD [5 days ON, 2 days OFF]) when tumor size reached 100 mm³. Tumor volumes were measured every 2 days (left), and representative photographs of dissected tumors are shown (right). Results are presented as the mean ± SD (n = 5). ∗∗p < 0.01, ∗∗∗p < 0.001; n.s., non-significant. Data are presented as mean ± SD.

Figure 7

Targeting YAP1 reverses acquired resistance to mutant KRAS-targeted therapy in CRC PDX models (A) Immunohistochemistry analysis of SDC1 and YAP1 expression in CRC PDX tumor at 3 days after initiation of AMG510 treatment or upon tumor progression (tumor volume change >100% from baseline) after 150 days of AMG510 treatment. Scale bar, 50 μm. (B) Indicated PDX models with acquired resistance to AMG510 were treated with vehicle (Veh), YAP1-ASO (ION537; 50 mg/kg, SQ, QD [5 days ON, 2 days OFF]), AMG510 (30 mg/kg, PO, QD), or ION537 and AMG510 combination treatment. n = 10 per group. (C) Immunohistochemistry for SDC1, YAP1, and Ki-67 expression in tumor tissue collected from models described in Figure 7B at 21 days after initiation of treatment. Scale bar, 100 μm. (D) Quantification of Ki-67 staining from Figure 7C. n = 3 per group. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. Data are presented as mean ± SD. (E and F) RPPA heatmap of B8026R tumors at 2 days (E) for 21 days (F) of treatment. Protein levels shown are log2 values.