Angiogenesis is believed to represent an early event in rheumatoid arthritis. However, the
contribution of different factors involved in the angiogenic process during various stages of
arthritis is still unclear. The aim of this study was to identify genes that are differentially
expressed in arthritic joint tissue, particularly genes with a role in angiogenesis.
Three different mouse models of collagen induced arthritis (CIA) were applied: namely,
CIA induced by heterologous collagen in DBA/1 and C57BL/6 mice, as well as the chronic CIA
model in DBA/1 mice induced by homologous collagen. Total RNA was extracted from mouse
paws and applied to perform gene expression analysis. Therefore two different real-time RT-PCR
approaches were employed: a pathway focused Angiogenesis Profiler Array analysing
simultaneously the expression of 84 genes, and gene specific RT-PCR. Immunohistochemistry for
CD31 (a well-known endothelial cell marker) and intravital fluorescence microscopy were
applied to analyse the microvessel density in arthritic paws and knees, respectively. To assess the
ability of arthritic joint homogenates to induce angiogenesis, different in vitro and in vivo
angiogenesis assays (endothelial chemotaxis assay, matrigel plug assay) were employed.
Immunohistochemical analysis revealed that highly inflamed synovial tissue of mouse
paw joints contains large numbers of CD31 positive vessels. Although arthritic synovium showed
regions of high vascular density, non-vascularised regions could be identified. Further, no
significant difference in the functional capillary density in arthritic and healthy knee synovium
was observed. Interestingly, there was a trend towards lower values in arthritic joints. This finding
suggests that neovascularisation occurs in CIA, but pannus growth outpaces new vessel
formation.
The different CIA models displayed distinct mRNA expression signatures. However,
arthritic tissue from DBA/1 and C57BL/6 mice did not exhibit a distinctive pro-angiogenic
expression profile. For example, increased mRNA expression could be observed for VEGF receptors (FLT-1, FLK-1, NRP-1, NRP-2), midkine, HGF, IGF-1, ANG-1, ANG-2, TIE-2,
whereas expression of VEGF/A and FGF-1 mRNA did not change during the course of acute
CIA. The mRNA levels of TSP-1 and endostatin, both involved in angiogenesis inhibition, were
also up-regulated during CIA, indicative of a balanced regulation of pro-angiogenic and antiangiogenic
factors. Overall, in the acute models, a simultaneous induction of different angiogenic
factors could be observed, which might indicate a concerted regulation of blood vessel formation
in the arthritic synovium. Gene expression pattern in chronic CIA (induced by homologous
collagen) varied between individual mice, reflecting the heterogeneous clinical features. But none
of the analysed genes was markedly altered. Indeed, only a few animals exhibited changes in
transcript levels of IL-1[beta], TNF[alpha], IL-6, HGF, IGF-1, midkine, leptin, and ANG-2. Although this requires further analyses, one might speculate that this is due to the fact that long-standing
(burned-out) lesions tone down their pro-inflammatory and pro-angiogenic environment.
A constant increase in HGF and NRP-1 mRNA levels suggested that both might be
important key players during the development of CIA. In a series of consecutive experiments, two
different therapeutic strategies were applied to block these molecules in acute CIA, namely the
administration of a HGF antagonist, NK4, and an anti-NRP-1 antibody, respectively. Both
inhibited angiogenesis induced by arthritic paw homogenates in vitro or in vivo. Treatment with
either NK4 or anti-NRP-1 significantly attenuated the severity and progression of CIA. However,
histological examination demonstrated that this clinical improvement was not necessarily
associated with a reduction in synovial vascularity. These observations suggest that inhibiting
NRP-1 or HGF signalling alone is not sufficient to treat arthritis.