Neuronal system-dependent facilitation of tumor angiogenesis and tumor growth by calcitonin gene-related peptide

Abstract

A neuropeptide, calcitonin gene-related peptide (CGRP), is widely distributed in neuronal systems and exhibits numerous biological activities. Using CGRP-knockout mice (CGRP(-/-)), we examined whether or not endogenous CGRP facilitates angiogenesis indispensable to tumor growth. CGRP increased tube formation by endothelial cells in vitro and enhanced sponge-induced angiogenesis in vivo. Tumor growth and tumor-associated angiogenesis in CGRP(-/-) implanted with Lewis lung carcinoma (LLC) cells were significantly reduced compared with those in wild-type (WT) mice. A CGRP antagonist, CGRP8-37 or denervation of sciatic nerves (L(1-5)) suppressed LLC growth in the sites of denervation compared with vehicle infusion or sham operation. CGRP precursor mRNA levels in the dorsal root ganglion in LLC-bearing WT were increased compared with those in non-LLC-bearing mice. This increase was abolished by denervation. The expression of VEGF in tumor stroma was down-regulated in CGRP(-/-). These results indicate that endogenous CGRP facilitates tumor-associated angiogenesis and tumor growth and suggest that relevant CGRP may be derived from neuronal systems including primary sensory neurons and may become a therapeutic target for cancers.

Fig. 1.

Reduced tumor-associated angiogenesis and tumor growth in CGRP-knockout (KO) mice. (A) Tumor growth after implantation in CGRP-knockout mice and their WT counterparts. LLC cells were implanted s.c. Results were compared with growth in WT counterparts on the same day and are mean ± SEM from 14 animals. ANOVA was used. *, P < 0.05. NS, not significant. (B) Typical H&E staining of tumors (day 14) in CGRP-KO mice (Left) and their WT counterparts (Right). Arrowheads indicate newly formed blood vessels. (C) Tumor growth in CGRP^−/− infused with CGRP continuously (0.1 nmol/h) by using miniosmotic pumps. LLC cells were implanted s.c. to the site of CGRP infusion. Results were compared with growth in vehicles-infused CGRP^−/− on day 7 and are mean ± SEM from 5∼7 animals. Student's t test was used. *, P < 0.05. (D) In the histologic tumor samples isolated from the mice 14 days after implantation, microvessel density was determined. Results from the sham operation group (blue column) were compared with those of denervation (red column) and are mean ± SEM from 10∼13 animals. Student's t test was used. *, P < 0.05. (E) Microvessel area was determined in the same specimen as in C. Results from the CGRP-KO mice (red column) were compared with those of WT (blue column) and are mean ± SEM from 10∼13 animals. Student's t test was used. *, P < 0.05.

Fig. 2.

Effects of continuous s.c. infusion of CGRP antagonist on tumor growth and tumor-associated angiogenesis in an LLC implantation model. (A) Tumor growth after implantation. LLC cells were implanted s.c. into WT. Subcutaneous infusion of a CGRP antagonist, CGRP8-37, with miniosmotic pumps (2 nmol/h) suppressed tumor growth 12 and 14 days after implantation (red column). Results were compared with those from vehicle-infused mice and are mean ± SEM from 10∼11 animals. ANOVA was used. *, P < 0.05. NS, not significant. (B) Typical H&E staining of tumor receiving vehicle (Left) or CGRP antagonist (Right). Arrowheads indicate newly formed blood vessels. (C) In the histologic tumor samples isolated from the mice 14 days after implantation, microvessel density was determined. Results obtained with the CGRP antagonist (red column) were compared with those obtained with vehicle and are mean ± SEM from 5 animals. Student's t test was used. *, P < 0.05. (D) Microvessel area was determined in the same specimen as in C. Results from CGRP antagonist (red column) were compared with those from vehicle and are mean ± SEM from 5 animals. Student's t test was used. *, P < 0.05.

Fig. 3.

Effect of sciatic nerve denervation on tumor growth and tumor-associated angiogenesis in an LLC implantation model. (A) Tumor growth after implantation. LLC cells were implanted s.c. into WT. Sciatic nerve denervation was performed as described in Materials and Methods. Results were compared with those from sham-operated mice and are mean ± SEM from 9∼11 animals. ANOVA was used. *, P < 0.05. NS, not significant. (B) Typical H&E staining of a tumor (day 14) after sham operation (Left) or denervation (Right). Arrowheads indicate newly formed blood vessels. (C) In the histologic tumor samples isolated from the mice 14 days after implantation, microvessel density was determined. Results from the sham-operation group (blue column) were compared with those after denervation (red column) and are mean ± SEM from 4∼6 animals. Student's t test was used. *, P < 0.05. (D) Microvessel area was determined in the same specimen as in C. Results from the sham operation group (blue column) were compared with those after denervation (red column) and are mean ± SEM from 4∼6 animals. Student's t test was used. *, P < 0.05. (E) Effects of sciatic nerve denervation on pro-CGRP mRNA levels in DRGs in tumor-bearing mice. Tumor implantation to the area innervated by L_1–5 resulted in increased expression of pro-CGRP in sham-operated WT (blue column). Sciatic nerve denervation reduced expression of pro-CGRP (red column). Mean ± SEM from 4∼11 animals is shown. ANOVA was used. *, P < 0.05.

Fig. 4.

Expressions of growth factors in tumor stromal tissues in CGRP-knockout mice. (A) VEGF dependence in LLC tumor growth used in the present work. Daily administration of the VEGF receptor tyrosine kinase inhibitor ZD6474 for 2 weeks suppressed the growth of LLC tumors. Results were compared with vehicle-treated mice and are mean ± SEM from 6∼10 animals. Student's t test was used. **, P < 0.05. (B) VEGF dependence in angiogenesis in LLC tumors. Daily administration of ZD6474 for 2 weeks suppressed angiogenesis. Results were compared with vehicle-treated mice and are mean ± SEM from 6∼10 animals. Student's t test was used. **, P < 0.05. (C) Reduced expressions of VEGF in tumor stromal tissues in CGRP-knockout mice 2 weeks after LLC implantation. Real-time PCR was performed as described in Materials and Methods. Results were compared with WT mice and are mean ± SEM from 7∼9 animals. Student's t test was used. **, P < 0.05. (D) Expression of bFGF in tumor stromal tissues. Experimental conditions were the same as those in C. NS, not significant. (E) Expression of CTGF in tumor stromal tissues. Experimental conditions were the same as those in C. (F) Expression of TGF-β in tumor stromal tissues. Experimental conditions were the same as those in C. (G) Immunohistochemical localization of VEGF in tumor tissues including stroma. The samples were isolated from CGRP^−/− and their WT counterparts.