Two Distinct Isoforms of Matrix Metalloproteinase-2 Are Associated with Human Delayed Kidney Graft Function

Abstract

Delayed graft function (DGF) is a frequent complication of renal transplantation, particularly in the setting of transplantation of kidneys derived from deceased donors and expanded-criteria donors. DGF results from tubular epithelial cell injury and has immediate and long term consequences. These include requirement for post-transplantation dialysis, increased incidence of acute rejection, and poorer long-term outcomes. DGF represents one of the clearest clinical examples of renal acute ischemia/reperfusion injury. Experimental studies have demonstrated that ischemia/reperfusion injury induces the synthesis of the full length secreted isoform of matrix metalloproteinase-2 (FL-MMP-2), as well as an intracellular N-terminal truncated MMP-2 isoform (NTT-MMP-2) that initiates an innate immune response. We hypothesized that the two MMP-2 isoforms mediate tubular epithelial cell injury in DGF. Archival renal biopsy sections from 10 protocol biopsy controls and 41 cases with a clinical diagnosis of DGF were analyzed for the extent of tubular injury, expression of the FL-MMP-2 and NTT-MMP-2 isoforms by immunohistochemistry (IHC), in situ hybridization, and qPCR to determine isoform abundance. Differences in transcript abundance were related to tubular injury score. Markers of MMP-2-mediated injury included TUNEL staining and assessment of peritubular capillary density. There was a clear relationship between tubular epithelial cell expression of both FL-MMP-2 and NTT-MMP-2 IHC with the extent of tubular injury. The MMP-2 isoforms were detected in the same tubular segments and were present at sites of tubular injury. qPCR demonstrated highly significant increases in both the FL-MMP-2 and NTT-MMP-2 transcripts. Statistical analysis revealed highly significant associations between FL-MMP-2 and NTT-MMP-2 transcript abundance and the extent of tubular injury, with NTT-MMP-2 having the strongest association. We conclude that two distinct MMP-2 isoforms are associated with tubular injury in DGF and offer novel therapeutic targets for the prevention of this disorder.

Fig 1. Scoring of injury in H/E stained delayed graft function kidney biopsies.

Panel A: Injury score of 1-mild focal tubular dilatation; Panel B: Injury score of 2-biopsy shows more extensive tubular dilatation, mild interstitial edema, and occasional necrotic tubular epithelial cell in lumen; Panel C: Injury score of 3-moderate to severe tubular dilatation, extensive interstitial edema, and frequent necrotic tubular epithelial cells in lumen. (Final magnification X 400).

Fig 2. FL-MMP-2 and NTT-MMP-2 expression in control protocol renal biopsies.

Panel A: Control protocol biopsy stained for FL-MMP-2. There is trace, focal immunohistochemical staining for FL-MMP-2 in proximal tubules (arrows) Panel B: Control protocol biopsy stained for NTT-MMP-2. There is no detectable immunohistochemical staining. (Final magnification X 200).

Fig 3. Ranking of serial biopsy sections for degree of tubular injury and immunohistochemistry staining.

Serial sections of protocol and DGF biopsies were stained for NTT-MMP-2 and FL-MMP-2 with hematoxylin/eosin and ranked according to injury score. Representative results for injury score 0. (A, B, C X 300; D, E, F X 600).

Fig 4. Ranking of serial biopsy sections for degree of tubular injury and immunohistochemistry staining.

Serial sections of protocol and DGF biopsies were stained for NTT-MMP-2 and FL-MMP-2 with hematoxylin/eosin and ranked according to injury score. Representative results for injury score 1. (A, B, C X 300; D, E, F X 600).

Fig 5. Ranking of serial biopsy sections for degree of tubular injury and immunohistochemistry staining.

Serial sections of protocol and DGF biopsies were stained for NTT-MMP-2 and FL-MMP-2 with hematoxylin/eosin and ranked according to injury score. Representative results for injury score 2. (A, B, C X 300; D, E, F X 600).

Fig 6. Ranking of serial biopsy sections for degree of tubular injury and immunohistochemistry staining.

Serial sections of protocol and DGF biopsies were stained for NTT-MMP-2 and FL-MMP-2 with hematoxylin/eosin and ranked according to injury score. Representative results for injury score 3. (A, B, C X 300; D, E, F X 600). There is a correlation between the overall tubular injury scale and the immunohistochemical staining scales for both NTT-MMP-2 and FL-MMP-2. Significantly, immunohistochemical staining for both NTT-MMP-2 and FL-MMP-2 occurs within the same tubular segments and there is a direct correlation between NTT-MMP-2 and FL-MMP-2 staining tubular injury.

Fig 7. FL-MMP-2 and NTT-MMP-2 are found within discrete tubular epithelial cellular compartments.

Images of DGF biopsies were acquired using Nomarksi optics and processed with pseudocolor to enhance contrast as detailed in Materials and Methods. Panel A: Immunohistochemical staining for FL-MMP-2 is confined to the cytoplasmic compartment (N, nucleus). Panel B: Immunohistochemical staining for NTT-MMP-2 is concentrated at the basolateral surfaces of the proximal tubular epithelial cells adjacent to the tubular basement membrane. The insert demonstrates NTT-MMP-2 staining within extended filamentous structures contained within basolateral infoldings characteristic of mitochondria. (Final magnification Panels A, B: X600, insert X1200).

Fig 8. In situ hybridization confirms NTT-MMP-2 expression in DGF tubular epithelial cells.

Control protocol biopsies and DGF biopsies were processed for in situ hybridization to detect the NTT-MMP-2 mRNA transcript as detailed in Materials and Methods. Panel A: Control renal biopsy without the targeting NTT-MMP-2 oligonucleotides. Panel B: Control renal biopsy including the targeting NTT-MMP-2 oligonucleotides. There is no detectable NTT-MMP-2 transcript. Panel C: DGF biopsy without the targeting NTT-MMP-2 oligonucleotides. Panel D: DGF biopsy with the targeting NTT-MMP-2 oligonucleotides. There is readily detectable NTT-MMP-2 mRNA transcript signal within tubular epithelial cells (arrows). (Final magnification X 300).

Fig 9. qPCR measurement of FL-MMP-2 and NTT-MMP-2 transcript abundance in control protocol biopsies and DGF biopsies.

mRNA was extracted from paraffin-embedded formalin-fixed renal biopsy specimens and FL-MMP-2 and NTT-MMP-2 transcript abundance was determined by qPCR as detailed in Materials and Methods. Both isoforms were normalized to a ribosomal protein, 36B4. There was a low, but detectable level of FL-MMP-2 transcript in the control protocol biopsies, consistent with the level of FL-MMP-2 immunohistochemical staining shown in Fig 2. FL-MMP-2 transcript abundance was increased approximately twelve-fold in the DGF samples. NTT-MMP-2 transcript abundance was nearly undetectable in the control protocol biopsies, consistent with the absence of NTT-MMP-2 immunohistochemical staining in controls (Fig 2). In contrast, NTT-MMP-2 transcripts were readily detectable in the DGF biopsies (* p<0.01).

Fig 10. Delayed graft function is associated with decreased peritubular capillary density.

Capillary endothelial cells were detected by CD-31 staining as detailed in Materials and Methods. Panel A, B: CD-31 staining of control protocol biopsy. Tubular structures are surrounded by multiple peritubular capillaries (arrows). Panel C, D: CD-31 staining of DGF biopsy. There is an evident decrease in the abundance and density of the peritubular capillaries, particularly in a patchy, non-uniform distribution (arrows). (Final magnifications: Panels A, C: X250; Panels B, D: X 300).