Tumor-associated genetic amplifications impact extracellular vesicle miRNA cargo and their recruitment of nerves in head and neck cancer

Abstract

Cancer neuroscience is an emerging field of cancer biology focused on defining the interactions and relationships between the nervous system, developing malignancies, and their environments. Our previous work demonstrates that small extracellular vesicles (sEVs) released by head and neck squamous cell carcinomas (HNSCCs) recruit loco-regional nerves to the tumor. sEVs contain a diverse collection of biological cargo, including microRNAs (miRNAs). Here, we asked whether two genes commonly amplified in HNSCC, CCND1, and PIK3CA, impact the sEV miRNA cargo and, subsequently, sEV-mediated tumor innervation. To test this, we individually overexpressed these genes in a syngeneic murine HNSCC cell line, purified their sEVs, and tested their neurite outgrowth activity on dorsal root ganglia (DRG) neurons in vitro. sEVs purified from Ccnd1-overexpressing cells significantly increased neurite outgrowth of DRG compared to sEVs from parental or Pik3ca over-expressing cells. When implanted into C57BL/6 mice, Ccnd1 over-expressing tumor cells promoted significantly more tumor innervation in vivo. qPCR analysis of sEVs shows that increased expression of Ccnd1 altered the packaging of miRNAs (miR-15-5p, miR-17-5p, and miR-21-5p), many of which target transcripts important in regulating axonogenesis. These data indicate that genetic amplifications harbored by malignancies impose changes in sEV miRNA cargo, which can influence tumorc innervation.

Figure 1.. Patient survival graphs.

Overall survival data from The Cancer Genome Atlas of head and neck cancer patients with amplification of CCND1 (a) or PIK3CA (b). Statistical analysis by logrank test with p values shown, analyzed via GEPIA2 (http://gepia2.cancer-pku.cn). These results are based upon data generated by the TCGA Research Network (https://www.cancer.gov/tcga ).

Figure 2.. Generation and analysis of mEER-P and mEER-C transgenic cells.

Lentiviral transduction of mEER cells with murine Ccnd1 or PIK3ca was assessed by western blot (a, b). Several positive clones were identified. Studies proceeded with CCND1 clone 7–6 and PIK3CA clone 2–1-3 (asterisks). Densitometric quantification of CCND1 (c) and PIK3CA (d) western blots; n=3 biological replicates. (e) MTS assays were conducted to assess proliferation differences between positively identified clones and parental cells. Statistical analysis by one-way ANOVA. *, p<0.05; ns, not significant. In vivo studies were conducted by injecting 100,000 (f) mEER-C or (g) mEER-P cells orthotopically into male C57Bl/6 mice to assess growth differences. Control mice were implanted with the same number of mEER (parental) cells. N= 10 mice/group. Statistical analysis by simple nonlinear/logarithmic fit test. *, p<0.05; ns, not significant.

Figure 3.. Assessment of intra-tumoral nerve density in transduced cell lines.

Orthotopic (a) mEER, (b) mEER-P, and (c) mEER-C tumors were collected at endpoint, fixed and stained for β-III tubulin. Stained sections were (d) scored based on staining intensity and distribution of intra-tumoral nerves and statistically analyzed for significance. N=10 tumors per groups were analyzed. Statistical analysis by Mann-Whitney test. **, p<0.01; ***, p<0.005; ns, not significant. Scale bar 100um.

Figure 4.. Increased gene expression impacts sEV-mediated neurite outgrowth.

Sholl analysis of DRG (control or treated with the indicated sEVs) was conducted by NeuriteJ, an ImageJ plug-in for neurite growth analysis). DRG were evaluated for maximum intersections of neurites (a) and maximum neurite distance (b). N = 4 DRG/group were analyzed. NGF, positive control; PBS, negative control; purified sEVs as listed from each cell line. Statistical analysis by one-way ANOVA. *, p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001; ns, not significant.

Figure 5.. Single genetic changes alter packaged miRNAs.

(a) Diagram of pathways influenced by miRNAs. (b) cDNA was generated from RNA isolated from parental and transduced cell line sEVs. qPCR analysis of mEE, mEER, mEER-P, and mEER-C cell lines shows the fold change in expression relative to mEER cells and reveals alterations in specific miRNAs. (N = 4 per group). Statistical analysis by two-way ANOVA. *, p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001; ns, not significant.

Figure 6.. miRNA transfection alters expression of miRNAs at the whole cell and sEV level.

mEER cells were transfected with miRNA mimic or scrambled vectors and incubated for 48 hours before undergoing GFP sorting. Cells were incubated for 72 hours and then sEVs were isolated from condition media. miRNA was measured by qPCR from cDNA generated from (a) whole cells or (b) sEVs (N = 4 per group). Fold change in expression is relative to untransfected mEER cells and sEVs. Statistical analysis by two-way ANOVA. *, p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001; ns, not significant.

Figure 7.. Overexpression of miR-15b-5p promotes neurite outgrowth.

Quantification of DRG neurites following the indicated treatments. DRG from C57BL/6 mice were isolated and placed in Matrigel and then treated with 20ul of 100ng/mL NGF (NGF), 20ul PBS (PBS), or 3ug of sEVs from parental mEER cells (parental sEVs), parental mEER cells transfected with a scrambled vector (scrambled), or sEVs from parental mEER cells transfected with a miR-15b-5p expressing vector (miR-15b-5p), a miR-17–5p expressing vector (miR-17–5p), or a miR-21–5p expressing vector (miR-21–5p) or their combinations. 48 hours after treatment, DRG were fixed, stained, and neurite outgrowth was measured by Sholl analysis. Maximum distance measurements are graphed, (n = 3 DRG analyzed per condition; each experiment was repeated two times). Statistical analysis by one-way ANOVA. *, p<0.05; ** p<0.01; *** p<0.001; **** p<0.0001; ns, not significant.

Figure 8.. sEV miRNAs impact PTEN-AKT pathway.

DRG were treated with 3μg of sEV from mEER cells over-expressing each indicated miRNA for 12 hours and then neurons were lysed and analyzed by western blot for PTEN (a), AKT (b) and ROCK1 (c). Densitometric quantification of the bands of interest was used to quantify changes in protein expression relative to the β-actin loading control. N= 4 biological replicates/group. Statistical analysis by one-way ANOVA with comparisons to the NGF control group. *, p<0.05; **, p<0.01; ***, p<0.005; ***, p<0.0001.

Figure 9.. mEER-CCND1 tumors potentiate lymph node metastasis.

The draining lymph nodes from mEER, mEER-C and mEER-P tumor-bearing animals were harvested (n= 5 lymph nodes/group), fixed, and IHC stained for cytokeratin (brown), an epithelial cell marker that stains tumor cells. a) Representative IHC-stained sections of lymph nodes from each group are shown. Scale bar, 500 μm. B) Quantification of metastases in lymph nodes in each group. Statistical analysis by Mann-Whitney nonparametric t-test. *, p< 0.05.

Figure 10.. sEV-packaged miR-15b-5p localizes to neurites in DRG cultures.

DRG from C57BL/6 mice were isolated, dissociated, and plated on PDL/laminin coated 8-well chamber slides, and then treated with 25ug of sEVs from miR-15b-5p transfected mEER cells for 24 hours. DRG were then fixed and probed for miR-15b-5p. (a) Photomicrograph of DRG treated with miR-15b-5p sEVs and probed for miR-15b-5p. Purple staining (arrows) is positive for miR-15b-5p. A region of interest is shown in higher magnification (inset). Semi-quantitative scoring of miR-15b-5p probe, separated by positive puncta located in either nerve soma, or within neurites. Quantification of the negative control is also included (b). Scale bars, 10 μm. N= 10 DRG neurons were analyzed with n=3 technical replicates for each. Statistical analysis by Welch’s t-test. *, p<0.05; **, p< 0.01; ***, p< 0.005.