The sVEGFR1-i13 splice variant regulates a β1 integrin/VEGFR autocrine loop involved in the progression and the response to anti-angiogenic therapies of squamous cell lung carcinoma

Abstract

Background: While lung adenocarcinoma patients can somewhat benefit from anti-angiogenic therapies, patients with squamous cell lung carcinoma (SQLC) cannot. The reasons for this discrepancy remain largely unknown. Soluble VEGF receptor-1, namely sVEGFR1-i13, is a truncated splice variant of the cell membrane-spanning VEGFR1 that has no transmembrane or tyrosine kinase domain. sVEGFR1-i13 is mainly viewed as an anti-angiogenic factor which counteracts VEGF-A/VEGFR signalling in endothelial cells. However, its role in tumour cells is poorly known.

Methods: mRNA and protein status were analysed by Real-Time qPCR, western blotting, ELISA assay, proximity ligation assay or immunohistochemistry in human tumour cell lines, murine tumourgrafts and non small cell lung carcinoma patients samples.

Results: We show that anti-angiogenic therapies specifically increase the levels of sVEGFR1-i13 in SQLC cell lines and chemically induced SQLC murine tumourgrafts. At the molecular level, we characterise a sVEGFR1-i13/β1 integrin/VEGFR autocrine loop which determines whether SQLC cells proliferate or go into apoptosis, in response to anti-angiogenic therapies. Furthermore, we show that high levels of both sVEGFR1-i13 and β1 integrin mRNAs and proteins are associated with advanced stages in SQLC patients and with a poor clinical outcome in patients with early stage SQLC.

Conclusions: Overall, these results reveal an unexpected pro-tumoural function of sVEGFR1-i13 in SQLC tumour cells, which contributes to their progression and escape from anti-angiogenic therapies. These data might help to understand why some SQLC patients do not respond to anti-angiogenic therapies.

Fig. 1

Anti-angiogenic therapies upregulate sVEGFR1-i13 expression in squamous cell lung carcinoma cell lines. The MGH7, H2170 or Calu-1 SQLC cell lines were treated with the indicated concentrations of bevacizumab (µg/ml) for 72 h (a, b, d, f, h) or with the indicated concentrations (µM) of KI8751 or SU5416 for 24 h (c, e, g, h). a–c ELISA assays were performed for quantification of sVEGFR1 protein level in the supernatants (a, c) or the cell pellets (b). d, e RT-qPCR analyses were performed to quantify sVEGFR1-i13 and VEGFR1 mRNA levels. GAPDH was used as an internal control. The value 1 was arbitrarily assigned to the specific signal/gapdh ratio obtained in non treated cells. f, g Western-blotting analyses of sVEGFR1-i13. Actin was used as a loading control. Histograms represent the quantification of the specific signal relative to actin. In all experiments, the means ± SD of three independent experiments are shown. (*p < 0.05; **p < 0.01; ***p < 0.001). (h) Western-blotting analyses of sVEGFR1-i13 in the supernatants of treated cells. Staining of the PVDF membrane using red ponceau was done to verify equal loading in all samples (not shown)

Fig. 2

Anti-angiogenic treatments upregulate sVEGFR1-i13 in murine SQLC tumourgrafts. a, b Left panels: representative immunostainings of paraffin-embedded sections of UN-SCC680 (a) or UN-ADC12 (b). The tumourgrafts had received PBS, control isotype, sunitinib or DC-101 as indicated. Negative immunostaining (irrelevant IgG, negative control) is also shown. Right panels: automatic quantification of tumour cell immunostaining (as described in the Materials and methods section) for each condition. The mean ± SD of 6 mice per condition is shown. Statistical analyses were performed using a student non paired t-test (*p < 0.05; **p < 0.01)

Fig. 3

a sVEGFR1-i13 regulates in an opposite way VEGFR1/VEGFR2 signalling pathways in response to anti-angiogenic therapies. (b) 96 h MTS proliferative assays were performed in MGH7 and H2170 cells treated with increasing concentrations of SU5416, KI8751 or bevacizumab as indicated. The mean ± SD of three independent experiments are shown. b–f MGH7 or H2170 cells were transfected for 72 h with two distinct sVEGFR1-i13 siRNA (or mismatch siRNA) and treated with 10 µg/ml bevacizumab (BVZ). Alternatively, the cells were transfected for 48 h with sVEGFR1-i13 siRNA and treated for 24 additional hours with 10 µM SU5416 or 10 µM KI8751. b Upper panel: RT-qPCR analyses were performed in MGH7 or H2170 cells to quantify sVEGFR1-i13 and VEGFR1 mRNA levels and assess the efficiency of the knock-down of sVEGFR1-i13. GAPDH was used as an internal control. The value 1 was arbitrarily assigned to the specific signal/gapdh ratio obtained in mismatch transfected cells. Lower panel: Western blot analyses in MGH7 cells of the indicated proteins are shown. Actin was used as a loading control. c Quantification of the signal for each indicated protein relative to actin in three different experiments. sVEGFR1-i13 (black histograms), P-VEGFR1 (white histograms), P-VEGFR2 (grey histograms) and active caspase-3 (hatched histograms). d Apoptosis was evaluated by counting apoptotic MGH7 cells after Hoechst staining in sVEGFR1-i13 (white bars) or mismatch (black bars) transfected cells. The mean ± SD of three independent experiments is shown. e, f The same experiments as in b, c were performed in H2170 cells. Statistical analyses (*p < 0.05, **p < 0.01, ***p < 0.001)

Fig. 4

β1 integrin is a critical determinant of the contrasting effects of sVEGFR1-i13 in response to anti-angiogenic therapies in SQLC cells. a Western blot analysis for detection of β1 integrin was performed in MGH7 or H2170 cells as indicated. Actin was used as a loading control. b Flow cytometry analysis for the detection of β1 integrin was performed in non permeabilised MGH7 or H2170 cells. Irrelevant isotype IgG was used as a negative control. c Western blot experiments for the detection of the indicated proteins. MGH7 cells were transfected for 48 h with β1 integrin siRNA (or mismatch as indicated) and treated (or not - Co) for 24 additional hours with 10 µM KI8751 (KI) or 10 µM SU5416 (SU). For bevacizumab (BVZ), MGH7 cells were transfected during 72 h with β1 integrin (or mismatch) siRNA in the presence (or absence) of 10 µg/ml bevacizumab. Actin was used as a loading control. d Mean densitometric quantification of the indicated proteins relative to actin signal in 3 different experiments. Statistical analyses were performed using ANOVA test (*p < 0.05, **p < 0.01, ***p < 0.001). e Western blot analyses of the indicated proteins were performed in H2170 cells transfected for 48 h with a plasmid encoding β1 integrin (β1) and treated (+) or not (−) with 10 µM KI8751 for 24 additional hours (left panels) or 10 µg/ml bevacizumab for 72 h (right panels). Actin was used as a loading control. f Mean densitometric quantification of the indicated proteins relative to actin signal in 3 different experiments. Statistical analyses were performed using ANOVA test (*p < 0.05, **p < 0.01, ***p < 0.001). g VEGFR2 protein was immunoprecipitated using an anti-VEGFR2 antibody (clone 55Β11) from total protein extracts obtained from MGH7 cells that had been treated for 72 h with 10 µg/ml bevacizumab or for 24 h with 10 µM SU5416 as indicated. IgG was used as an irrelevant antibody. The presence of VEGFR2, β1 integrin or sVEGFR1-i13 protein in the immunoprecipitates was assessed by western blotting. Input represents 10% of the immunoprecipitates. h The presence of VEGFR2/sVEGFR1-i13 co-immunoprecipitates was assessed in H2170 cells that had been treated for 72 h with 10 µg/ml bevacizumab as mentioned in e. i VEGFR2 protein was immunoprecipitated in MGH7 cells transfected with mismatch or sVEGFR1-i13 siRNA during 48 h and treated (+) or not (−) with 10 µM SU5416 for 24 additional hours. The presence of VEGFR2, β1 integrin or sVEGFR1-i13 protein in the immunoprecipitates was assessed by western blotting. Input represents 10% of the immunoprecipitates. g–i Numbers represent the quantification of VEGFR2, β1 integrin and sVEGFR1-i13 signal intensities using Image J software. In immunoprecipitation experiments, β1 integrin and sVEGFR1-i13 signals were determined according to VEGFR2 signal in each condition. The value 1 was arbitrarily assigned to the untreated condition signal

Fig. 5

sVEGFR1-i13/β1 integrin autocrine cross-talk stimulates cell proliferation in MGH7 cells. a, b MGH7 cells were transfected and studied after 48 h with a plasmid encoding sVEGFR1-i13 (sVEGFR1-i13) or control, Co. a Left panels: quantification by ELISA of sVEGFR1-i13 in the supernatants. Right panels: cell number (x10⁶) was estimated following trypan blue staining in cells cultured for 48 h in the presence of the supernatants obtained from cells transfected with the plasmid encoding sVEGFR1-i13 (sVEGFR1-i13), or controls. The mean ± SD of three independent experiments is illustrated. b Western blot analyses of the indicated proteins in MGH7 cells either transfected (plasmid) with a plasmid encoding sVEGFR1-i13 (R1-i13) or control (Co), or cultured for 48 h in supernatants (supernatants) taken from cells transfected with a plasmid encoding sVEGFR1-i13 (R1-i13) or control (Co). Actin was used as a loading control. c MGH7 cells were transfected for the indicated times with either control (MoC) or sVEGFR1-i13 (MoFL2) morpholino. Cell number (x10⁶) was estimated following trypan blue staining. RT-qPCR in MGH7 cells demonstrated the efficiency of sVEGFR1-i13 morpholino as detected by the significant increase of sVEGFR1-i13 mRNA level. The mean ± SD of three independent experiments is illustrated. d MGH7 cells were cultured for 72 h in the presence of supernatants taken from cells transfected for 48 h with a plasmid encoding sVEGFR1-i13 (R1-i13) or control (Co) in the presence of 10 µg/ml of either a scramble peptide (sp12) or a peptide blocking the interaction between sVEGFR1-i13 and β1 integrin (p12) as indicated. Upper panel: cell number (x10⁶) was estimated following trypan blue staining. Lower panel: western blot analyses of the indicated proteins were performed. Actin was used as a loading control. e Western blot analyses of the indicated proteins were performed in MGH7 cells transfected for 72 h with either mismatch or β1 integrin siRNA as indicated and treated (+) or not (−) for 24 h with 1 ng/ml sVEGFR1 recombinant ligand. Actin was used as a loading control. f VEGFR2 protein was immunoprecipitated using an anti-VEGFR2 antibody (clone 55Β11) from total protein extracts obtained from MGH7 cells treated or not for 24 h with 1 ng/ml sVEGFR1 recombinant ligand. IgG was used as an irrelevant antibody. The presence of VEGFR2, β1 integrin or sVEGFR1-i13 protein in the immunoprecipitates was assessed by western blotting. The ‘Input’ represents 10% of the amount used for the immunoprecipitations. Numbers represent the quantification of VEGFR2 and β1 integrin signal intensities in immunoprecipitates using Image J software. β1 integrin signals were determined according to VEGFR2 signal in each condition. The value 1 was arbitrarily assigned to the untreated condition signal. In all experiments, statistical analyses were performed using a student non paired t-test (a) or ANOVA test (c, d) (*p < 0.05, **p < 0.01)

Fig. 6

sVEGFR1-i13/β1 integrin cross-talk is involved in SQLC tumour progression. a Representative sVEGFR1-i13 immunostainings from paraffin-embedded sections of normal lung tissues (upper panel: alveolus; lower panel: bronchus) as well as of two well-differentiated squamous cell lung carcinomas. Scores are indicated for each case. b Mean levels ± SD of sVEGFR1-i13 immunohistochemical scores according to the P-VEGFR1(Tyr1213) status in squamous cell lung carcinoma, where + and – represent tumours with high or low levels of P-VEGFR1(Tyr1213) compared to normal lung tissues respectively. Statistical analyses were performed using a non parametric Mann–Whitney test. c Mean levels ± SD of β1 integrin immunohistochemical scores according to the sVEGFR1-i13 status in ADC (white boxes) and SQLC (hatched boxes), where + and – represent tumours with high or low levels of sVEGFR1-i13 compared to normal lung tissues respectively. Statistical analyses were performed using a non parametric Mann–Whitney test. d Mean levels ± SD of β1 integrin+sVEGFR1-i13 immunohistochemical scores according to P-VEGFR1(Tyr1213) status in ADC (white boxes) and SQLC (hatched boxes), where + and – represent tumours with high or low levels of P-VEGFR1(Tyr1213) compared to normal lung tissues respectively. Statistical analyses were performed using a non parametric Mann–Whitney test. e Mean levels ± SD of MAS5-normalised sVEGFR1-i13 mRNA in SQLC patients taken from the GSE4573 database expressing either low (<25th percentile), medium (25–75th percentile) or high (>75th percentile) levels of β1 integrin (ITGB1) mRNA. Statistical analyses were performed using Kruskal–Wallis test. f Kaplan–Meier univariate survival analysis of SQLC patients with pTNM I/II stage according to high (>75th percentile; 4th quartile) or low/medium (<75th percentile; 1st–3rd quartiles) sVEGFR1-i13 and β1 integrin (ITGB1) mRNA levels. h Left panels: anti-angiogenic therapies increase the level of sVEGFR1-i13 in squamous lung tumour cells. However, whether cells respond or do not respond to anti-angiogenic therapies depends on the level of β1 integrin. In cells with high levels of β1 integrin (MGH7-like phenotype), the increase of sVEGFR1-i13 protein level upon anti-angiogenic treatment contributes to the activation of a β1 integrin-dependent VEGFR1/VEGFR2 autocrine loop. This loop promotes the tumour cell proliferation and survival, thereby escape from the treatment. In contrast, in cells with low β1 integrin expression (H2170-like phenotype), sVEGFR1-i13 heterodimerises with VEGFRs, prevents VEGFR signalling and induces apoptosis. We propose that this H2170-like phenotype might represent a subclass of SQLC patients who could be more responsive to ramucirumab treatment than the others. Right panel: In the absence of anti-angiogenic therapies, in SQLC tumours expressing high expression level of β1 integrin, the sVEGFR1-i13/β1 integrin/VEGFR cross-talk might contribute to the progression of squamous lung carcinoma patients