The RAS-related GTPase RHOB confers resistance to EGFR-tyrosine kinase inhibitors in non-small-cell lung cancer via an AKT-dependent mechanism

Abstract

Although lung cancer patients harboring EGFR mutations benefit from treatment with EGFR-tyrosine kinase inhibitors (EGFR-TKI), most of them rapidly relapse. RHOB GTPase is a critical player in both lung carcinogenesis and the EGFR signaling pathway; therefore, we hypothesized that it could play a role in the response to EGFR-TKI In a series of samples from EGFR-mutated patients, we found that low RHOB expression correlated with a good response to EGFR-TKI treatment while a poor response correlated with high RHOB expression (15.3 versus 5.6 months of progression-free survival). Moreover, a better response to EGFR-TKI was associated with low RHOB levels in a panel of lung tumor cell lines and in a lung-specific tetracycline-inducible EGFR^L858R transgenic mouse model. High RHOB expression was also found to prevent erlotinib-induced AKT inhibition in vitro and in vivo Furthermore, a combination of the new-generation AKT inhibitor G594 with erlotinib induced tumor cell death in vitro and tumor regression in vivo in RHOB-positive cells. Our results support a role for RHOB/AKT signaling in the resistance to EGFR-TKI and propose RHOB as a potential predictor of patient response to EGFR-TKI treatment.

Keywords: AKT; EGFR; RhoB; TKI; resistance.

Figure 1. RHOB expression predicts the response to EGFR‐TKI in patients harboring EGFR‐activating mutations

1. Four representative images of RHOB immunostaining in human non‐small‐cell lung cancers. Figures correspond to the percentage of tumors analyzed.

2. Characteristics of 96 patients with an EGFR‐activating mutation who were treated with EGFR‐TKI. Each column corresponds to one patient. Data include general characteristics (stage at diagnosis, age, sex, tobacco usage, position of EGFR mutation, TKI type), progression‐free survival (PFS; lower graph), and intensity of RHOB staining as determined by immunohistochemistry (null: 0; weak: +; moderate: ++; and high: +++).

3. Representative scans of patients with low‐ or high‐RHOB‐expressing lung tumors, before and after erlotinib treatment. Red arrows indicate lung tumors.

4. Progression‐free survival of erlotinib‐treated patients with EGFR‐mutated lung tumors, according to RHOB expression, assessed by immunohistochemistry (low‐RHOB group = negative + weak staining; high‐RHOB group = moderate + high staining). P‐values were determined by the Kaplan–Meier method.

5. Progression‐free survival of patients who received EGFR‐TKI as first‐line therapy (n = 63). P‐values were determined by the Kaplan–Meier method.

6. Progression‐free survival of patients who received EGFR‐TKI as second‐line (n = 28), third‐line (n = 3), or fourth‐line (n = 2) therapy. P‐values were determined by the Kaplan–Meier method.

7. RHOB immunostaining score evolution in EGFR‐mutated lung tumors before treatment and after EGFR‐TKI relapse.

Figure EV1. RHOB expression predicts the response to EGFR‐TKI in patients with EGFR‐mutated lung cancer

1. A, B

   Progression‐free survival of EGFR‐mutated patients treated with EGFR‐TKI (A) for the whole population or (B) according to RHOB expression in treatment‐naive tumors (null: 0; weak: +; moderate: ++; and high: +++), as assessed by immunohistochemistry.

Figure 2. RHOB loss of expression increases sensitivity to erlotinib in mice with EGFR^L858R‐driven lung tumors

1. A

   Representative H&E staining of whole lungs from EGFR ^L858R/Rhob ^−/−, EGFR ^L858R/Rhob ^+/−, and EGFR ^L858R/Rhob ^+/+ mice treated or not with erlotinib (12.5 mg/kg/day) for 4 days. Scale bars: 5 mm.

2. B

   Quantification of the tumor/lung ratio. n = 7 for each group except for EGFR ^L858R/Rhob ^+/− placebo (n = 6).

3. C, D

   Representative Ki67 immunostaining of EGFR ^L858R/Rhob ^−/−, EGFR ^L858R/Rhob ^+/−, and EGFR ^L858R/Rhob ^+/+ mice treated or not with erlotinib (12.5 mg/kg/day) for 4 days (scale bars: 50 μm), and the corresponding quantification (D). Three independent zones per mouse lung were used for quantification. n = 21 (seven mice) for each group except for EGFR ^L858R/Rhob ^+/− placebo (n = 18; six mice).

4. E

   Immunostaining of cleaved caspase‐3 in lung tumors from EGFR ^L858R/Rhob ^+/+ or EGFR ^L858R/Rhob ^−/− mice treated for 24 h with erlotinib at 12.5 mg/kg. Black arrows point apoptotic cells. Scale bars: 50 μm.

Data information: **P < 0.001 versus placebo; ***P < 0.0001 versus placebo. Data are expressed as mean ± SEM, P‐values were determined by Mann–Whitney two‐tailed t‐test.

Figure 3. The modulation of RHOB expression determines the level of resistance to erlotinib in EGFR‐mutated lung cancer cell lines

1. A–C

   HCC4006 cells were either (A) transfected with two siRNA against RHOB (siB1, siB2), (B) transduced with control (AdCont) or RHOB‐overexpressing adenoviruses (AdRHOB), or (C) both transfected with siRNA and transduced by adenoviruses, and then treated with increasing doses of erlotinib. The surviving cell fraction was determined by an MTS assay after 72 h and compared to untreated cells.

2. D–F

   Erlotinib IC₅₀ values were quantified in RHOB‐overexpressing or RHOB‐depleted (D) HCC827 cells, (E) HCC2935 cells, and (F) H3255 cells, as determined by an MTS assay after 72‐h treatment. RHOB overexpression or inhibition was monitored by Western blotting for each condition.

3. G

   HCC4006 cells were transduced with control (AdCont) or RHOB‐overexpressing adenoviruses (AdRHOB) at an increasing multiplicity of infection (MOI); then, erlotinib IC₅₀ values were determined after 72 h by an MTS assay, and a correlation analysis was performed. RHOB overexpression was monitored by Western blotting.

Data information: ***P < 0.0001 versus control cells. Data are representative of at least three independent experiments. Data are expressed as mean ± SEM, P‐values were determined by unpaired two‐tailed Student's t‐test.Source data are available online for this figure.

Figure EV2. Adenoviral‐mediated RHOB overexpression reverses erlotinib sensitivity induced by RHOB depletion

1. A–D

   HCC827 (A), HCC4006 (B), H3255 (C), or HCC2935 (D) cells were transfected with control (siNeg) or RHOB‐targeting siRNA (siB1) and then transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated with increasing doses of erlotinib. The surviving cell fraction was determined by an MTS assay after 72 h and compared to untreated cells. RHOB overexpression or inhibition was monitored by Western blotting. Data are expressed as mean ± SEM from three independent experiments.

Source data are available online for this figure.

Figure 4. RHOB induces resistance to erlotinib through the AKT pathway

1. HCC4006, HCC827, HCC2935, and H3255 cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 4 h with erlotinib at concentrations corresponding to the respective IC₅₀ values determined for each control cell line. The phosphorylation status of AKT, ERK1/2, and EGFR was assessed by Western blotting and normalized according to total protein levels. RHOB overexpression was also monitored by Western blotting.

2. Representative immunostaining of phospho‐AKT (Ser473) and phospho‐ERK1/2 and their total protein amounts in lung tumors from EGFR ^L858R/Rhob ^−/− or EGFR ^L858R/Rhob ^+/+ mice treated or not with erlotinib (12.5 mg/kg/day) for 4 days. The remaining hyperplastic areas were selected in erlotinib‐treated mice to efficiently characterize the effect of erlotinib on ERK and AKT pathways in both Rhob genotypes. Scale bars: 100 μm.

3. HCC4006 cells were transfected with a plasmid coding for a constitutively active AKT mutant (AKTmyr, myristoylated) or an empty vector (ø) and treated for 72 h with increasing concentrations of erlotinib. The surviving cell fraction was determined by an MTS assay, and AKT overexpression and phosphorylation at Ser473 were assessed by Western blotting. Data are representative of at least three independent experiments. Data are expressed as mean ± SEM from three independent experiments.

Source data are available online for this figure.

Figure EV3. RHOB overexpression does not affect response to erlotinib in EGFR WT cell lines

1. A549 or H1299 cells were transduced with control (AdCont) or RHOB‐overexpressing adenoviruses (AdRHOB) and treated with increasing doses of erlotinib. The surviving cell fraction was determined by an MTS assay after 72 h and compared to untreated cells. Data are expressed as mean ± SEM from three independent experiments.

2. A549 and H1299 cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 4 h with erlotinib at 1 μM. The phosphorylation status of AKT, ERK1/2, and EGFR was assessed by Western blotting and normalized according to total protein levels. RHOB overexpression was also monitored by Western blotting. EGFR‐mutated HCC4006 cells were used to monitor erlotinib efficiency (right panel).

Source data are available online for this figure.

Figure 5. AKT inhibition sensitizes RHOB‐expressing cells to erlotinib

1. A

   HCC4006 cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 4 h with erlotinib (100 nM), G594 (100 nM), or a combination of both drugs. The phosphorylation status of GSK3β (Ser9), ERK1/2, and EGFR (Tyr1173) was assessed by Western blotting and normalized according to the total protein levels. RHOB overexpression was also monitored by Western blotting.

2. B

   HCC4006 cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 72 h with erlotinib alone (black and red curves) or in combination with the AKT inhibitor G594 at 100 nM (green and blue curves). The surviving cell fraction was determined by an MTS assay. Data are expressed as mean ± SEM from three independent experiments.

3. C–E

   HCC4006 cells were transduced with control (AdCont) or RHOB‐overexpressing adenoviruses (AdRHOB) and treated for 72 h with increasing concentrations of erlotinib in the absence or presence of increasing doses of G594. The surviving cell fraction was determined by an MTS assay, and erlotinib IC₅₀ values were determined for each condition (**P < 0.001 versus AdCont cells; ***P < 0.0001 versus AdCont cells). HCC4006 cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 48 h with erlotinib (100 nM), G594 (100 nM), or a combination of both drugs. Apoptosis was then determined by either quantification of the subG1 cell population (D) or detection of cleaved PARP and caspase‐3 (E) (***P < 0.0001 versus untreated cells; ^### P < 0.0001 versus G594 treated cells; ^††† P < 0.0001 versus erlotinib‐treated AdCont cells). In vitro data are representative of at least three independent experiments. Data are expressed as mean ± SEM from three independent experiments, P‐values were determined by unpaired two‐tailed Student's t‐test.

Source data are available online for this figure.

Figure EV4. G594 prevents GSK3β phosphorylation in RHOB‐overexpressing cells treated with erlotinib and reverses RHOB‐induced resistance

1. A–C

   H3255 (A), HCC2935 (B), and HCC827 (C) cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 4 h with erlotinib (100 nM), G594 (100 nM), or a combination of both drugs. The phosphorylation status of GSK3β (Ser9), ERK1/2, and EGFR (Tyr1173) was assessed by Western blotting and normalized according to the total protein levels. RHOB overexpression was also monitored by Western blotting.

2. D–F

   H3255 (D), HCC2935 (E), or HCC827 (F) cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 72 h with erlotinib alone (black and red curves) or in combination with the AKT inhibitor G594 at 100 nM (green and blue curves). The surviving cell fraction was determined by an MTS assay. Data are expressed as mean ± SEM from three independent experiments.

Source data are available online for this figure.

Figure EV5. AKT inhibition reverses RHOB‐induced resistance to erlotinib‐mediated apoptosis

1. A–F

   H3255 (A and B), HCC2935 (C and D), or HCC827 (E and F) cells were transduced with control (AdCont) or RHOB‐overexpressing (AdRHOB) adenoviruses and treated for 48 h with erlotinib (100 nM), G594 (100 nM), or a combination of both drugs. Apoptosis was then determined by either detection of cleaved PARP and caspase‐3 (A, C, and E) or quantification of the subG1 cell population (B, D, and F). ***P < 0.0001 versus untreated cells; ^### P < 0.0001 versus G594 treated cells; ^††† P < 0.0001 versus erlotinib‐treated AdCont cells. Data are expressed as mean ± SEM from three independent experiments, P‐values were determined by unpaired two‐tailed Student's t‐test.

Source data are available online for this figure.

Figure 6. The AKT inhibitor G594 resensitizes EGFR^L858R mouse tumors to erlotinib

1. A

   Representative H&E staining of lung tumors from EGFR ^L858R/Rhob ^−/−, EGFR ^L858R/Rhob ^+/−, and EGFR ^L858R/Rhob ^+/+ mice treated or not during 4 days with erlotinib (12.5 mg/kg/day), the AKT inhibitor G594 (25 mg/kg/day), or a combination of both drugs. Scale bars: 500 μm.

2. B

   Quantification of the tumor/lung ratio of mice treated or not with the individual drugs or with a combination of both.

3. C, D

   Representative Ki67 immunostaining of lung tumors from EGFR ^L858R/Rhob ^−/−, EGFR ^L858R/Rhob ^+/−, and EGFR ^L858R/Rhob ^+/+ mice treated or not with erlotinib alone (12.5 mg/kg/day) or in combination with the AKT inhibitor G594 (25 mg/kg/day) for 4 days (scale bars: 50 μm), and the corresponding quantification (D).

Data information: **P < 0.001; ***P < 0.0001. EGFR ^L858R/Rhob ^+/+ (placebo: n = 6; G594: n = 5; erlotinib: n = 7; G594 + erlotinib: n = 8); EGFR ^L858R/Rhob ^+/− (placebo: n = 7; G594: n = 6; erlotinib: n = 8; G594 + erlotinib: n = 7); EGFR ^L858R/Rhob ^−/− (placebo: n = 6; G594: n = 6; erlotinib: n = 6; G594 + erlotinib: n = 6). Three independent zones per mouse lung were used for Ki67 quantification. Data are expressed as mean ± SEM, P‐values were determined by Mann–Whitney two‐tailed t‐test.