Gasdermin B over-expression modulates HER2-targeted therapy resistance by inducing protective autophagy through Rab7 activation

GSDMB upregulation in response to anti-HER2 therapies associates with drug resistance

Our previous work proved that GSDMB over-expression is a maker of poor prognosis associated with trastuzumab resistance in HER2 breast carcinoma patients in both neoadjuvant and adjuvant treatment settings [12, 14]. Moreover, high levels of GSDMB decrease sensitivity to trastuzumab in vitro in HCC1954 and SKBR3 cells, and its expression increases during the acquisition of trastuzumab resistance in HER2 + breast cancer PDX models [12]. Here, to address whether GSDMB also plays a role in the clinical behavior of gastric tumors, we first observed that strong cytoplasmic and nuclear GSDMB staining (Fig. 1A and Supplementary Table 1) associated statistically with HER2 positive status (GSDMB high expression in 18/31 (58.1%) of HER2-positive and 8/28 (29%) of HER2-negative tumors; p = 0.023). In HER2 gastric carcinomas, alike HER2 breast tumors [12], GSDMB over-expression associates with relapse (p = 0.060, Supplementary Table 2), thus supporting the relationship between high levels of GSDMB and poor prognosis in gastric tumors. Next, to test if in HER2 breast and gastroesophageal cancers GSDMB is functionally involved in regulating drug response/resistance also to the HER2 tyrosine kinase inhibitor lapatinib, we used three HER2 + cancer cell lines that endogenously express GSDMB, HCC1954 (breast cancer), OE19 (esophageal) and NCI-N87 (gastric). First, we treated these cells for different time points (up to 72 h) with their corresponding IC50 of lapatinib (Fig. 1B-C and Supplementary Fig. 1A). Additionally, for comparison, trastuzumab treatment was carried out only in OE19 and NCI-N87 cells (Supplementary Fig. 1B-C) since HCC1954 cells are intrinsically highly resistant to this drug [14, 28]. Both lapatinib (Fig. 1B-C and Supplementary Fig. 1A) and trastuzumab (Supplementary Fig. 1B-C) provoke a sharp induction of GSDMB mRNA, which peaked at 24–48 h, in all tested models. HER2 upregulation was also detected as previously reported [29]. At the protein level, while GSDMB was strongly upregulated by lapatinib in HCC1954 cells, the total amount of GSDMB protein was not clearly increased in gastroesophageal cancer cells (OE19 and N87) neither after lapatinib nor trastuzumab treatment. This is due to the appearance of a processed form of GSDMB protein (p37) (Fig. 1B-C and Supplementary Fig. 1A-C) at the latest treatment time points, that corresponds to the previously identified C-terminal cleavage product generated by apoptotic caspases-3/6/7 [30] (Supplementary Fig. 1D). GSDMB processing by caspases-3/6/7 generates N- (p10) and C-terminal (p37) fragments [30] that do not have an effect on cell death induction [6].

Fig. 1figure 1

GSDMB is induced in response to anti-HER2 therapies, and its silencing increases the sensitive to lapatinib treatment. A Representative immunofluorescence images of GSDMB and HER2 expression in GSDMB/HER2-positive gastric carcinomas. Scale bar, 20 µm. Nuclei were counterstained with DAPI. Representative pie chart of GSDMB expression statistics in HER2 + gastric carcinomas (see Supplementary Table 1 for extra clinical features and statistics). B-C Relative mRNA (left) and protein levels (right) of GSDMB and HER2 in HCC1954 (B) and OE19 (C) cells treated with IC50 of lapatinib (2 µM and 0.7 µM, respectively) at indicated time points. D GSDMB and HER2 protein levels in lapatinib resistant HCC1954 and OE19 cells (LR) and their corresponding control cells (C) treated chronically with lapatinib and DMSO, respectively, and after ten days of drug removal. Study of the different cytotoxic effect of the chronic lapatinib treatment in these cells was analyzed by cell viability assays. E Relative mRNA levels of GSDMB and HER2 in trastuzumab (TRC1 and TRC2) and lapatinib (LRC1 and LRC1) resistant tumors, compared to the parental tumor (C) derived from a HER2 breast cancer PDXs. F GSDMB expression was reduced in HCC1954 LR (left) and OE19 LR cells (right) by two specific siRNAs (siGB1 and siGB2), in comparison with the control (siNTC). Cytotoxic effect of the presence of lapatinib (2 µM and 1.5 µM, respectively) in GSDMB-siRNAs-silenced cells HCC1954 LR (left) and OE19 LR (right) cells was assessed by cell viability assays. G Cytotoxic effect after 72 h treatment with IC50 of lapatinib (2 µM and 0.7 µM, respectively) in GSDMB-shRNAs-silenced cells HCC1954 and OE19 (right) cells was assessed by cell growth (viability assays, left panel) and death (Annexin V FITC and PI, right panel). Annexin V-FITC positive cells alone (A + /PI-) and Annexin V-FITC and PI doubled stained (A + /PI +) were defined as apoptotic cells. The number over the bars indicate the ratio of cell death relative to the shNTC condition. Statistical significance was determined by two-tailed unpaired t-test (*P < 0.05; **P < 0.01). Data are shown as the mean ± s.e.m. Three independent experiments with similar results were performed. In (B, C, E), gene expression was normalized to the mRNA levels of GAPDH. A, Annexin V; PI, propidium iodide. NTC, non-targeting control. DMSO, dimethyl sulfoxide. Lap, lapatinib

Next, to assess further if GSDMB induction correlates with acquired resistance to HER2-targeted therapies, we generated HCC1954 and OE19 cells long-term resistant (LR) to high doses (> IC50) of lapatinib. In both models, again, we observed an upregulation of GSDMB and HER2, and as expected, a strong decrease in HER2 phosphorylation [29], compared to their respective control cells (chronically grown with high doses of the vehicle DMSO) (Fig. 1D). Interestingly, GSDMB upregulation was dependent on lapatinib presence since drug removal for 10 days restored GSDMB protein levels (Fig. 1D). Moreover, as a complementary model, we used primary cultures from previously generated HER2 + breast cancer PDXs [31]. Thereby, compared to the parental cell line (C), which is sensitive to trastuzumab, the in vitro resistant clones to trastuzumab (TRC1 and TRC2) or lapatinib (LRC1 and LRC2) showed a significant GSDMB upregulation (Fig. 1E). These in vitro results in various cell models (breast and gastric cancer cell lines, breast PDX primary culture), together with our published observations in vivo with PDXs and HER2 breast carcinoma patients [12], suggest that the GSDMB over-expression would induce resistance to trastuzumab as well as lapatinib. Since TKIs are mainly utilized in trastuzumab pretreated patients [17] (indeed HCC1954 cells are intrinsically resistant to trastuzumab [14, 28]) and there is a need to understand the molecular mechanisms that underlies the TKI derived resistance, we selected lapatinib as the suitable (second line) anti-HER2 therapy to further decipher the connection between GSDMB and drug resistance.

To study if GSDMB is functionally involved in the early response and long-term resistance to lapatinib, we silenced GSDMB in OE19 and HCC1954 cells by shRNA (two different sequences shGB1 and shGB2), as previously reported [13, 14]. These shRNAs robustly decrease GSDMB expression at the protein (Supplementary Fig. 1E) and mRNA levels (all four GSDMB isoforms that translate to protein) in OE19 (Supplementary Fig. 1F) and HCC1954 cells [14], and do not target other GSDM genes (Supplementary Fig. 1G). Furthermore, in the resistant HCC1954 LR and OE19 LR cells, GSDMB expression was transiently reduced by two different GSDMB-specific siRNAs (siGB1/2, Fig. 1F). Stable GSDMB-silencing by lentiviral shRNA transduction could not be obtained in these models because they exhibited intrinsic resistance to the selection antibiotic puromycin.

GSDMB-shRNA-silenced HCC1954 and OE19 cells were significantly more sensitive to lapatinib treatment, as they exhibited an important reduction in cell viability and an increase in cell death compared to shNTC control cells (Fig. 1G). Likewise, in the LR models with siGSDMB, we found a slight but statistically significant decrease (around 20%) in cell viability in the presence of lapatinib (Fig. 1F).

GSDMB increases pro-survival autophagy in response to lapatinib treatment

GSDMB silencing did not affect the levels of HER2 receptor in any of our cell models (Supplementary Fig. 2A), suggesting that GSDMB does not promote response/resistance to lapatinib through direct modulation of HER2 quantity. Therefore, to decipher the molecular mechanism by which GSDMB modulates lapatinib response, we focused on autophagy, as this process has been demonstrated to act as a resistance mechanism to anti-HER2 therapies both in vivo and in vitro [22]. Autophagy induction during tumor progression can lead to either survival (pro-tumor) or cell death (anti-tumor) depending on the stimulus and the cellular context [21]. Hence, we first tested in HCC1954 and OE19 parental cells if lapatinib treatment induced autophagy with survival or death consequences (Supplementary Fig. 2B-E). In both cell lines, lapatinib treatment induced autophagic response at 24–48 h, measured by the increase in the levels of LC3B-II (Supplementary Fig. 2B). Importantly, this autophagy is protective since blocking autophagic flux with chloroquine (CQ), which affects the completion of the latter stages of autophagy [32], significantly increased the cytotoxicity of lapatinib (Supplementary Fig. 2C). Similarly, blocking the formation of autophagosomes [33] by an ATG5-specific siRNA (Supplementary Fig. 2D) enhanced the effect on cell viability of lapatinib (Supplementary Fig. 2E).

Next, we assessed whether cells with high or low GSDMB expression could have different endogenous autophagic responses by measuring autophagic flux as the accumulation of the lipidated LC3B (LC3B-II) form in western blots [34, 35]. In this regard, we analyzed LC3B-II turnover in the presence and absence of lysosomal degradation using CQ. Therefore, higher LC3-II levels were observed in GSDMB-expressing (shNTC) HCC1954 and OE19 cells in comparison with GSDMB-silenced cells in basal autophagy, (Fig. 2A-B, see grey bars) and this effect was significantly exacerbated upon autophagy activation with lapatinib (Fig. 2A-B, see blue bars). It should be noted that due to the high efficacy of the combined treatment, the enhanced cell death resulted in an overall degradation of proteins, including GSDMB and GAPDH (Fig. 2A-B). Besides, a significant increase in the relative volume density of autophagic vacuoles was observed by transmission electron microscopy in GSDMB-expressing HCC1954 cells, compared to GSDMB-silenced cells, after the treatment with lapatinib and its combination with CQ (Fig. 2C). In the same way, HCC1954 LR and OE19 LR cells, which express high levels of GSDMB, exhibit increased LC3B-II accumulation by western blot and LC3B-II puncta by confocal imaging (Fig. 2D-E) compared to their respective control cells. Thus, these results suggest that high GSDMB expression somehow increases the intrinsic autophagic response to cell stress. Accordingly, we confirmed that the autophagic flux induced by lapatinib (Fig. 2A-C), or by serum starvation (Supplementary Fig. 3A-B) is reduced in GSDMB-silenced cells compared to control lines. Moreover, in HCC1954 LR cells grown in the presence of high doses of lapatinib, the reduction of GSDMB expression by siRNAs also diminishes the autophagic flux (Fig. 2F). Furthermore, we discarded that this observed autophagic response involved selective autophagy subtypes such as mitophagy (Supplementary Fig. 4A), aggrephagy (Supplementary Fig. 4B) or lipophagy (Supplementary Fig. 4C).

Fig. 2figure 2

GSDMB-high cells show an increased autophagic flux in response to lapatinib. A-B GSDMB and LC3B protein levels in shNTC, shGB1 and shGB2 HCC1954 (A) and OE19 (B) cells treated with lapatinib (Lap, 2 µM and 0.7 µM, respectively) and/or CQ (10 µM and 50 µM, respectively) for 72 h. Quantification of the relative LC3B-II expression was conducted as described before [34, 35]. C Representative transmission electron microscopy images of shNTC and shGB2 HCC1954 cells treated with the treatment regimens indicated in (A). Quantification of the relative volume density of autophagic vacuoles is shown on the right. At least 25 cells were analyzed per experimental condition. D-E Western blot analysis of GSDMB and LC3B (left panels) in HCC1954 LR (D) and OE19 LR (E) cells and their respective controls (C) treated with or without CQ (10 µM and 50 µM, respectively) for 72 h. LC3B expression (green) analysis by confocal microscopy (right panels) in HCC1954 LR (D) and OE19 LR (E) cells and their controls (C) treated with or without CQ at the concentrations indicated in (A-B). Representative confocal microscopy images were shown, scale bar, 10 µm. Nuclei were counterstained with DAPI. F GSDMB and LC3B protein levels in GSDMB-siRNA-silenced HCC1954 LR cells treated with or without 10 µM CQ for 72 h. Quantification of LC3B-II expression (showed on the right of panels, A-B, D-F) was carried out by densitometric scanning and normalized to GAPDH expression following previous methods [34, 35]. Statistical significance was determined by two-tailed unpaired t-test (*P < 0.05; **P < 0.01). Data are shown as the mean ± s.e.m. Three independent experiments with similar results were performed. NTC, non-targeting control. LR, Lapatinib resistant cells. CQ, chloroquine. Lap, lapatinib

Then, given that GSDMB-high cells exhibit increased autophagic lapatinib response and that this process promotes survival in our cell models (Supplementary Fig. 2), we postulated that GSDMB-high cells could be particularly sensitive to the combination of lapatinib with autophagy inhibitors. Indeed, the autophagy blockage with CQ in GSDMB-high (shNTC) HCC1954 and OE19 cells treated with lapatinib produces a significant reduction in the cell viability (Fig. 3A-C) as well as an increase in cell death rates (Fig. 3B-D) compared to lapatinib alone. In contrast, no such dramatic effect was observed in the GSDMB-silenced cells. Importantly, in HCC1954 LR and OE19 LR cells, which over-express GSDMB, the addition of CQ strongly increases the cytotoxic effect of lapatinib, and thus revert their resistance to this anti-HER2 drug (Fig. 3E-H). Consistent with these results, the autophagy blockage by siATG5 confirmed the increased sensitivity to lapatinib in GSDMB-high cells in all cell models (Supplementary Fig. S5a-d). It should be noted that particularly in OE19 LR cells, which show the highest GSDMB levels, the autophagy inhibition alone (either by CQ or siATG5) has a significant effect on cell viability (Fig. 3G and Supplementary Fig. 5D), supporting that after long-term challenge with this anti-HER2 therapy the subsistence of these cells mostly relies on the endogenous pro-survival autophagic process.

Fig. 3figure 3

GSDMB-high cells are significantly more sensitive to the combination of lapatinib plus chloroquine. A-D The cytotoxic effect of the treatment with lapatinib and/or chloroquine in shNTC, shGB1 and shGB2 HCC1954 (A, B) and OE19 (C, D) cells was evaluated by cell viability assays and Annexin V-FITC plus PI. E–H The outcome in terms of cell viability and apoptosis of the different treatment regimens was analyzed in HCC1954 LR (E–F) and OE19 LR (G-H) cells compared to their corresponding control cells C. Statistical significance was determined by two-tailed unpaired t-test (*P < 0.05; **P < 0.01; ***P < 0.001). Data are shown as the mean ± s.e.m. Annexin V-FITC positive cells alone (A + /PI-) and Annexin V-FITC and PI doubled stained (A + /PI +) were defined as apoptotic cells. The number over the bars indicate the fold increase in cell death between the indicated conditions (B, D, F, H). Three independent experiments with similar results were performed. A, Annexin V; PI, propidium iodide; CQ, chloroquine. NTC, non-targeting control. LR, Lapatinib resistant cells. Lap, lapatinib, CQ, chloroquine

Taken together, these results confirm that GSDMB enhances the pro-survival autophagy in response to lapatinib in HER2 + breast and gastric cancer cells, and thus GSDMB-over-expressing cells are more sensitive to the addition of autophagy inhibitors both in the early response to HER2-targeted treatment and in drug resistant cells.

Autophagy inhibition enhances lapatinib efficacy in vivo specifically in GSDMB-expressing breast cancer cells

Next, to validate if the combination of lapatinib and CQ on HER2/GSDMB positive tumors would be an effective therapeutic approach, we assayed its functional effect on tumor growth in vivo using two different preclinical models, zebrafish, and mice (Fig. 4). In zebrafish, which allows testing the drug response in a large number of biological replicates [36], we first calculated the acute dose-dependent toxicity of both lapatinib and CQ by analyzing the embryo mortality (Supplementary Table 3). Next, the different HCC1954 models mentioned above (shNTC, shGB1/2 as well as LR and the corresponding control cells) that stably express GFP were inoculated into the yolk sac of zebrafish embryos and treated with either lapatinib, CQ or the combination of both (Fig. 4A-B and Supplementary Fig. 6A-B). As expected, lapatinib treatment produced a significant reduction in cancer cell growth of GSDMB-silenced tumors (shGB1 and shGB2), but not in HCC1954 shNTC (Fig. 4A). Remarkably, autophagy blockage by CQ provoked an increase in the effect of lapatinib on tumor growth reduction (p < 0.001) only in GSDMB-expressing (shNTC) tumors, compared to lapatinib alone, while no such effect was observed in GSDMB-silenced (shGB1, shGB2) tumors. Furthermore, similar results were found on lapatinib resistant cells (LR), where the combined treatment (CQ plus lapatinib) practically abolished the resistant phenotype of HCC1954 LR cells (Fig. 4B and Supplementary Fig. 6B). These findings support that the combination therapy was effective specifically in high GSDMB-expressing tumors.

Fig. 4figure 4

Autophagy blockade with chloroquine improves lapatinib efficacy in vivo in zebrafish and mouse xenografts of GSDMB-expressing tumors. A-B Representative fluorescence stereomicroscope images (left panel) of GFP expressing control (shNTC), and GSDMB-silenced (shGB1) HCC1954 xenografts (A) or HCC1954 LR and control (C) tumors (B) treated with the LC50 of lapatinib and chloroquine (35,1 mM and 116,4 mM, respectively). Insets represent an augmented image of GFP-positive tumor. Tumor proliferation rate (right panel) of the different HCC1954 xenografts was analyzed by measuring the fluorescence intensity ratio (48 hpt/0 hpt). Statistical significance was determined by two-tailed unpaired t-test. At least, n = 20 per each indicated condition. C Experimental design of the mouse xenograft model (n = 5 per condition). Mice were inoculated with either HCC1954-mCherry-luc control (shNTC), or GSDMB-silenced cells (shGB1 and shGB2) and treated with lapatinib (100 mg/kg, orally, once daily), CQ (50 mg/kg, intraperitoneally, once daily), or a combination of both (lapatinib + CQ). An aqueous solution containing 0.1% Tween 80 and 0.5% Hypromellose was used as vehicle. The experiment was performed for 30 days, according to the approved protocol and conditions of animal research (detailed in Supplementary Methods). D-G Quantification of the tumor weight (D) tumor volume evolution of shNTC (E), shGB1 (F) and shGB2 (G) HCC1954 xenografts, treated with the indicated regimens. Statistical significance was determined by multiple unpaired t-test – comparing vehicle with each of the other conditions at every time point. H Representative images of GSDMB immunohistochemical analysis and hematoxylin and eosin staining in shNTC tumors, treated with the different therapeutic strategies indicated in (C). Immunohistochemical images were taken on 10X and 40X (insets) magnification. (*P < 0.05; **P < 0.01; ***P < 0.001; ns, nonsignificant). Data are shown as the mean ± s.e.m. LC50, 50% lethal concentration; hpt, hours post-treatment. NTC, non-targeting control. LR, Lapatinib resistant cells, CQ, chloroquine. Lap, lapatinib

Subsequently, to validate these results, we performed similar treatment experiments in mice bearing orthotopically injected HCC1954 (shNTC, shGB1 and shGB2) breast cancer xenografts (Fig. 4C-H and Supplementary Fig. 7). Once again, while lapatinib alone mainly decreased tumor weight and volume in GSDMB-silenced tumors (Fig. 4D,F,G and Supplementary Fig. 7A), the addition of CQ enhanced the effect of lapatinib on the reduction of tumor growth exclusively in GSDMB-expressing (shNTC) tumors (Fig. 4D-E). The therapeutic response was independent of cell proliferation since no differences in PCNA immunohistochemical expression were detected in any of the experimental conditions (Supplementary Fig. 7B). As observed in vitro, GSDMB was also up-regulated in vivo in GSDMB-expressing tumors (shNTC) by lapatinib, noting an increase in the necrotic areas in the case of the combined treatment (Fig. 4H and Supplementary Fig. 7B), and again it was associated with a reduced response to the treatment compared to GSDMB-silenced tumors (Fig. 4D,F,G and Supplementary Fig. 7C). These in vivo data prove that the autophagy blockage is essential to improve the anti-HER2 therapy response in HER2/GSDMB tumors and reinforce the role of GSDMB in the promotion of pro-survival autophagy as a resistance mechanism to these therapies.

GSDMB acts as an autophagy adaptor inducing Rab7 activation with a predictive value to anti-HER2 therapies

To unravel further the role of GSDMB in autophagy, HCC1954 cells exogenously over-expressing myc-tagged full length GSDMB (GB) after treatment with lapatinib, CQ, or both drugs were immunoprecipitated with a Myc-tag antibody followed by mass-spectrometry studies. Among potential GSDMB interacting cancer and autophagy proteins, some Rab GTPases were found (Rab5C, Rab7A, Rab9A and Rab15, Fig. 5A and Supplementary Table 4). These have been implicated in different steps of the autophagy process [37]; especially, Rab7 which participates in the autophagosome-lysosome fusion, the autolysosome maturation and transport [25, 38] and has shown a pivotal role in resistance to chemotherapeutic agents [39]. Rab7a was found with a higher score both in CQ and combined treatment in the mass-spectrometry assay compared to untreated cells (Fig. 5A). To identify the interaction region between GSDMB and Rab7a, an in silico prediction assay was carried out, revealing an energetically feasible GSDMB/Rab7a complex (ΔG = -2.73 kJ/mol), in which the most probable GSDMB residues involved in this interaction are distributed through its N-terminal region (amino acids V7, D30, F46, Q60 and E185 from GSDMB, Fig. 5B). Interestingly, we also found in the GSDMB N-terminal domain two putative LC3B-interacting region (LIR) motifs (“LIR1” corresponding to 3SVFEEI8 sequence and “LIR2” covering 82AEFQIL87 amino acids, Supplementary Fig. 8A). Despite LC3B was not detected in the proteomic analysis, we hypothesized that a GSDMB-Rab7a-LC3B multiprotein complex could occur. In fact, the HawkDock web server [40], predicted a highly possible GSDMB/Rab7a/LC3B complex (ΔG = -2.17 kJ/mol, Supplementary Fig. 8B). By contrast, a similar complex was not feasible when the rest of human GSDMs were analyzed (data not shown), supporting the idea that although other GSDMs have been previously implicated in autophagic processes [4, 41, 42], only GSDMB might regulate autophagy through Rab7a/LC3B interaction. Thus, to validate these results experimentally, we utilized HCC1954 parental cells with endogenous GSDMB levels (Control, C), cells with exogenous GSDMB overexpression (GB) and a model exogenously over-expressing the C-terminal (GB92−416) fragment produced by caspase 3/6/7 processing [30], which lacks both LIR motifs.

Fig. 5figure 5

GSDMB modulates autophagic response through Rab7 and LC3B interaction. A Heatmap representation of the mass spectrometry analysis performed on HCC1954 GB cells treated with the different treatment regimens, after co-immunoprecipitation assay using a Myc-tag antibody. Normalization to untreated cells was performed. Color key indicates protein expression score: dark blue: highest; dark orange: lowest. B In silico protein interaction prediction using the InterEvDock2 and PPCheck web servers identifies a potential interaction between Rab7a (yellow) and the N-terminal domain of GSDMB (green). Inset magnifies the predicting interacting region (pink). 3D structures were obtained from Uniprot data bases (Q8TAX9: GSDMB and P51149: Rab7a). C-D Co-immunoprecipitation assay of Rab7-GSDMB (C) and LC3B-GSDMB (D) interaction after CQ and lapatinib plus CQ treatment in HCC1954 cells exogenously expressing full length GSDMB (GB), a construct lacking the N-terminal domain (GB(92–416)) or control (C; empty vector) cells. E Representative images of the colocalization between Rab7 (red), LC3B (cyan) and GSDMB-Myc-tag (green) by confocal microscopy in HCC1954 GB and GB(92–416) cells after lapatinib (2 µM) plus CQ (10 µM) treatment. Nuclei were counterstained with DAPI. Quantification of the Manders’ Overlap Coefficient (Rab7 overlapping Myc-tag and LC3B overlapping Myc-tag) is shown on the right. Five independent experiments were performed obtaining at least 60 cells, per experimental condition. Scale bar, 10 µm. F Representative images of the colocalization between Rab7 (red) and LC3B (green) by confocal microscopy in shNTC, shGB1 and shGB2 HCC1954 cells after lapatinib (2 µM) plus CQ (10 µM) treatment. Nuclei were counterstained with DAPI. Quantification of the Manders’ Overlap Coefficient (LC3B overlapping Rab7) is shown on the right. Five independent experiments were performed obtaining at least 60 cells, per experimental condition. Scale bar, 10 µm. G-H The cytotoxic effect of the treatment with lapatinib plus/or siRab7a-silencing was evaluated by cell viability assays in shNTC, shGB1 and shGB2 HCC1954 (G) and HCC1954 LR (H) cells. I-J Comparative cytotoxic effect by cell viability assays (I) and Annexin V-FITC plus PI (J) in HCC1954 C, GB and GB(92–416) cells after 72 h treatment with lapatinib (2 µM) and/or CQ (10 µM). Annexin V-FITC positive cells alone (A + /PI-) and Annexin V-FITC and PI doubled stained (A + /PI +) were defined as apoptotic cells. Statistical significance was determined by two-tailed unpaired t-test (*P < 0.05; **P < 0.01; ***P < 0.001; ns, nonsignificant). Data are shown as the mean ± s.e.m. In (E-J), three independent experiments with similar results were performed. Lap, lapatinib, CQ, chloroquine

First, we noticed that neither GB nor GB92−416 over-expression significantly modifies Rab7 and LC3B endogenous levels (Supplementary Fig. 8C). In the co-immunoprecipitation assays, the interaction between GSDMB and Rab7 was detected only in HCC1954 GB cells but not in GB92−416 or control cells (Fig. 5C and Supplementary Fig. S8D). As expected, LC3B and Rab7 co-immunoprecipitation was mainly found after autophagy inhibition by CQ treatment in HCC1954 GB cells (Fig. 5C). Similar results were observed in the binding assays between GSDMB and LC3B (Fig. 5D and Supplementary Fig. 8

留言 (0)

沒有登入
gif