3.1. Anlotinib targets FGFR3 and inhibits FGFR3 phosphorylation in the treatment of OSCC
To explore the possible anlotinib targets, firstly, we detected the mRNA expression levels of VEGFR1-3, FGFR1-4, PDGFRβ and C-KIT genes in four OSCC cell lines by qRT-PCR. We found that HUVECs mainly expressed FGFR1 and VEGFR2 mRNA (Fig. 1A and Table 1). However, HOK and OSCC cells expressed low mRNA levels of VEGFR1-3, PDGFRβ and C-KIT genes, but with high mRNA levels of FGFR1-4 genes (Fig. 1A and Table 1). Furthermore, four OSCC cell lines mainly expressed FGFR3 mRNA among FGFR1-4 (Fig. 1A and Table 1). We further evaluated the protein levels of the above genes by western blotting and found that VEGFR1, c-Kit and PDGFRβ were barely detected, VEGFR2 and VEGFR3 protein basically not expressed in OSCC cells (Fig. 1B). Meanwhile, FGFR3 protein had the highest expression levels among FGFR1-4 protein in OSCC cells (Fig. 1B). Although the FGFR3 protein levels were not obviously changed after anlotinib treatment, we found the phosphorylation levels of FGFR3 were significantly reduced in the anlotinib group in a dose-dependent manner in OSCC cells compared with the non-anlotinib group (Fig. 1C and Supplementary Fig. 1). These above results implied that FGFR3 may act as the anlotinib target in OSCC cells.
Table 1
Relative mRNA expression of therapeutic target of Anlotinib in OSCC cell lines
Cell line | VEFGR1 | VEFGR2 | VEFGR3 | PDGFRβ | FGFR1 | FGFR2 | FGFR3 | FGFR4 | c-KIT |
HUVECs | 1.000 | 3.896 | 0.065 | 0.034 | 29.590 | 0.002 | 0.073 | 0.008 | 0.667 |
HOK | 0.335 | 0.255 | 0.226 | 0.401 | 106.328 | 2.551 | 1.861 | 0.628 | 0.709 |
SCC9 | 0.323 | 0.032 | 0.011 | 0.695 | 2.211 | 3.605 | 11.951 | 3.400 | 0.149 |
SCC15 | 0.095 | 0.426 | 0.074 | 0.740 | 1.648 | 7.160 | 12.289 | 3.144 | 0.005 |
SCC25 | 0.025 | 0.004 | 0.042 | 0.806 | 1.722 | 9.049 | 13.947 | 3.048 | 0.000 |
UM1 | 0.007 | 0.003 | 0.009 | 0.247 | 1.589 | 0.710 | 13.122 | 3.571 | 0.005 |
The VEGFR1 expression level of HUVEC was defined as 1. |
3.2. FGFR3 expression level affects antitumor activity of anlotinib in OSCC
To further examine the role of FGFR3 in anlotinib-induced antitumor effects in OSCC, three different sequences of siFGFR3 were used to transfer into the SCC9 and SCC15 cells, and the siRNA with the highest silencing efficacy (siFGFR3-3) was selected for subsequent experiments (Fig. 2A). As shown in Fig. 2B-D, anlotinib induced proliferation inhibition (Fig. 2B) and apoptosis (Fig. 2C) in SCC9 and SCC25 in a dose-dependent manner. The phosphorylation of FGFR3, AKT and mTOR was decreased after anlotinib treatment (Fig. 2D). The ability of proliferation inhibition and apoptosis stimulation effects of anlotinib was significantly decreased after knockdown FGFR3 in OSCC cells, moreover, the phosphorylation levels of FGFR3, AKT and mTOR were slightly increased after knockdown FGFR3, suggesting that lower FGFR3 expression levels led to less anlotinib sensitivity of OSCC. Recombinant human FGF-2 (rhFGF-2, 20 ng/mL), which can stimulate autophosphorylation of FGFR3 and the phosphorylation levels of AKT and mTOR, also significantly diminished the ability of proliferation inhibition and apoptosis stimulation effects of anlotinib in OSCC cells (Fig. 2B-C and Supplementary Fig. 2). After anlotinib treatment, the phosphorylation levels of FGFR3, AKT and mTOR were decreased in rhFGF-2-treated plus anlotinib group. These changes caused subsequent corresponding changes of downstream apoptosis proteins, including BAX and BCL-2 (Fig. 2D).
Overall, these results suggested that anlotinib targeted and inhibited FGFR3 phosphorylation levels and then affected the subsequent AKT/mTOR phosphorylation to cause apoptosis of OSCC cells.
METTL3 regulates FGFR3 mRNA m 6 A modification and decreases FGFR3 mRNA stability
To investigate how METTL3 regulate FGFR3 in OSCC cells, firstly, we knocked down METTL3 in SCC9 and SCC25 cells and found that the m6A levels were significantly decreased in METTL3-knockdown group by dot-blot analysis (Fig. 3A). We then analyzed the data of m6A MeRIP-seq from our previous study29 and found that the m6A peaks of FGFR3 was enriched near the 3’UTR regions (Fig. 3B). MeRIP-qPCR assay also showed that FGFR3 m6A modification was significantly decreased after METTL3 knockdown in SCC9 and SCC25 cells (Fig. 3C). Western blotting showed that METTL3 knockdown increased FGFR3 protein and mRNA expression levels (Fig. 3D). These data suggested that METTL3 might regulate FGFR3 expression at the post-transcriptional level.
Moreover, we compared FGFR3 mRNA and protein stability in METTL3-knockdown group with control cells by actinomycin D and cycloheximide assay. We found that the stability of FGFR3 mRNA level was remarkably increased in METTL3-knockdown cells compared with control cells, but without changing the degradation rate of FGFR3 protein level (Fig. 4E and F). The results manifested that FGFR3 mRNA m6A modification by METTL3 could accelerate its degradation.
METTL3 are inversely associated with anlotinib sensitivity in OSCC cells
To investigate the role of METTL3 in anlotinib treated OSCC cells, we explored the correlation between anlotinib sensitivity (IC50 values) and METTL3 expression level in OSCC cell lines. Four OSCC cell lines (SCC9, SCC15, SCC25 and UM1) were treated with anlotinib (Fig. 4A), and the expression of endogenous METTL3 was assessed by qRT-PCR and western blotting in OSCC cell lines (Fig. 4B-C). Strong correlation between IC50 and METTL3 expression level was identified in OSCC cells. The OSCC cells with higher METTL3 expression were less sensitive to anlotinib (higher IC50 values) (Table 2). METTL3 knockdown in OSCC cells sensitized the inhibitory effects of anlotinib in OSCC cells. The IC50 value was decreased and the apoptosis was increased after anlotinib treatment in METTL3 knockdown OSCC cells (Fig. 4D-E and Supplementary Fig. 3).
Table 2
Correlation analysis of IC50 and METTL3 expression of different OSCC cells
Cell line | SCC9 | SCC15 | SCC25 | UM1 | Pearson |
IC50 | 5.157 | 3.315 | 3.524 | 6.420 | |
METTL3 mRNA level | 1.778 | 1.528 | 1.715 | 1.959 | 0.967 |
METTL3 protein level* | 0.780 | 0.838 | 0.722 | 0.868 | 0.926 |
*The radio between the gray value of METTL3 protein and GAPDH protein. Pearson correlation coefficient was METTL3 expression level relative to IC50. |
METTL3 affects the FGFR3 expression and antitumor efficacy of anlotinib in PDX models
To verify the relationship of METTL3, FGFR3 and antitumor sensitivity of anlotinib in vivo, we explored the expression level of METTL3 and FGFR3 of tumor tissues from eight previously established OSCC PDX models16. After 30 days of anlotinib treatment, the TGI rate of anlotinib was evaluated to measure the antitumor sensitivity of anlotinib of OSCC PDX. TGI values in eight PDX models were 95.90 (#005), 92.28 (#010), 78.07 (#022), 92.85 (#024), 92.89 (#030), 92.89 (#032), 88.25 (#034) and 89.18% (#040). Histopathological examination (H&E) and IHC assay were used to detect the expression level of METTL3/FGFR3 in the control group of each PDX models (Fig. 5A-C). As shown in Fig. 5D, the expression level of METTL3/FGFR3 had significant correlationship with TGI rate in PDX models. PDX models with the lower expression level of METTL3 or the higher expression level of FGFR3 displayed more sensitive to anlotinib treatment with higher TGI rate (Fig. 5D). Meanwhile, the expression levels of METTL3 and FGFR were significantly negatively correlated in each PDX sample (Fig. 5D).