Human nephrosclerosis triggers a hypoxia-related glomerulopathy

Abstract

In the kidney, hypoxia contributes to tubulointerstitial fibrosis, but little is known about its implications for glomerular damage and glomerulosclerosis. Chronic hypoxia was hypothesized to be involved in nephrosclerosis (NSC) or "hypertensive nephropathy." In the present study genome-wide expression data from microdissected glomeruli were studied to examine the role of hypoxia in glomerulosclerosis of human NSC. Functional annotation analysis revealed prominent regulation of hypoxia-associated biological processes in NSC, including angiogenesis, fibrosis, and inflammation. Glomerular expression levels of a majority of genes regulated by the hypoxia-inducible factors (HIFs) were significantly altered in NSC. Among these HIF targets, chemokine C-X-C motif receptor 4 (CXCR4) was prominently induced. Glomerular CXCR4 mRNA induction was confirmed by quantitative RT-PCR in an independent cohort with NSC but not in those with other glomerulopathies. By immunohistological analysis, CXCR4 showed enhanced positivity in podocytes in NSC biopsy specimens. This CXCR4 positivity was associated with nuclear localization of HIF1alpha only in podocytes of NSC, indicating transcriptional activity of HIF. As the CXCR4 ligand CXCL12/SDF-1 is constitutively expressed in podocytes, autocrine signaling may contribute to NSC. In addition, a blocking CXCR4 antibody caused significant inhibition of wound closure by podocytes in an in vitro scratch assay. These data support a role for CXCR4/CXCL12 in human NSC and indicate that hypoxia not only is involved in tubulointerstitial fibrosis but also contributes to glomerular damage in NSC.

Figure 1

Cluster of HIF target genes in NSC. Unbiased and unsupervised clustering was used to visualize the expressed HIF target genes in a dendrogram. Cluster dendrograms group the patients with the most similar gene expression profiles. The length of the branches represents the dissimilarity. Probe set abundance is displayed on a red-green color scale, with red indicating expression above and green below the mean. This cluster dendrogram uses the HIF target genes to segregate NSC from the healthy controls. Two distinct branches could be identified: one corresponding to the NSC group and the other one to the TN group.

Figure 2

Confirmation of LOXL2, FN1, and CXCR4 mRNA expression in glomeruli with NSC and controls. A: LOXL2 mRNA measured by RT-PCR is induced in glomeruli from NSC biopsy specimens. LOXL2 mRNA showed increased expression by quantitative RT-PCR in glomeruli from an independent group of NSC biopsy specimens (n = 13). Values were normalized to pretransplant LD control biopsy specimens (n = 6) and the housekeeper gene 18S rRNA. In FSGS, IgAN, and MCD, LOXL2 mRNA was not significantly different compared with that in living donor (LD) controls. **P < 0.01 compared with controls. B: FN1 mRNA is overexpressed in glomeruli from NSC biopsy specimens. Same biopsy cohorts as mentioned in A. **P < 0.01 compared with controls (LD). C: CXCR4 mRNA expression levels are higher in glomeruli from NSC biopsy specimens. Same biopsy cohorts as mentioned in A. *P < 0.05 compared with controls (LD). FSGS, focal segmental glomerusclerosis; IgAN, IgA nephropathy; MCD, minimal change disease.

Figure 3

CXCR4 and HIF1α staining in NSC and controls. A: CXCR4 and HIF1α staining in glomeruli with NSC. CXCR4 (green staining) and HIF1α (red staining) showed robust expression in podocytes of NSC (short arrows). As expected some tubular epithelial cells stained positive for HIF1α (long arrow). Interstitial mononuclear cells also stained positive for CXCR4 (not shown). B: High magnification of CXCR4 and HIF1α staining in NSC. Higher magnification of glomerular epithelial cells with surface staining for CXCR4 (short arrows) and nuclear staining for HIF1α (long arrow). C: CXCR4 and HIF1α staining in glomeruli from controls. Healthy tissue from TN served as a control, with podocytes showing a faint green CXCR4 background staining (arrows) without any HIF1α positivity.

Figure 4

CXCL12 staining in glomeruli with NSC. CXCL12 was localized by immunohistochemical analysis in biopsy specimens from patients with NSC and controls. A prominent CXCL12 expression was found, in a pattern consistent with podocytes. Note the staining of cells on the outer surface of the glomerular basement membrane (arrows). Original magnification, ×400. Inset: Higher magnification of CXCL12 staining of glomeruli with NSC confirmed expression by podocytes (arrows). Original magnification, ×1000.

Figure 5

Hypoxia induces HIF1α and HIF2α protein in cultured podocytes. A: By Western blot a dose-dependent induction of HIF1α under graded hypoxic conditions (0.2 and 2% O₂) was demonstrated in human podocytes compared with normoxic time controls (20% O₂, at 4 hours). B: In addition, HIF2α was found to be stabilized in human podocytes at 0.2 and 2% O₂ compared with 20% O₂ (at 24 hours). β-Actin was used as loading control.

Figure 6

Hypoxia induces Cxcr4 mRNA in podocytes in vitro. In cultured murine podocytes, expression of Cxcr4 mRNA was increased after 24 hours at 0.2% O₂ compared with podocytes incubated at 2% O₂ after 24 hours and normoxic (20% O₂) controls (n = 4 each). *P < 0.05, **P < 0.01, significant differences compared with 20% O₂. **P < 0.01, significant difference comparing 0.2 with 2% O₂. For expression data of CXCR4 and VEGFA in human podocyte cell lines exposed to hypoxia, see Table 3.

Figure 7

CXCR4 blockade reduces podocyte migration in vitro in a wound-healing assay. To assess potential functions of CXCR4 in podocyte migration, a wound-healing assay was performed. Confluent murine podocytes (K5P5) were treated with either blocking anti-CXCR4 antibody or immunoglobulin IgG as control. At 0 hours and after 22 hours, the width of each scratch inflicted at 0 hours was measured in a blinded fashion. The CXCR4 antibody significantly impaired the extent of wound closure compared with controls (23 ± 6 versus 41 ± 13%, **P < 0.01, n = 12 each). mAb, monoclonal antibody.