Journal Logo

EXPERIMENTAL TRANSPLANTATION

TISSUE TRANSGLUTAMINASE: A MEDIATOR AND PREDICTOR OF CHRONIC ALLOGRAFT NEPHROPATHY?

Johnson, Timothy S.1 4; Abo-Zenah, Hamdy1; Skill, James N.1 2; Bex, Samantha3; Wild, Graham3; Brown, Colin B.1; Griffin, Martin2; Nahas, A. Meguid El1

Author Information
doi: 10.1097/01.TP.0000131171.67671.3C
  • Free

Abstract

Despite the dramatic improvement in short-term renal-al-lograft survival (1), cadaveric kidney attrition rate exceeds 40% at 5 years posttransplantation (Tx) (2). Late graft loss is associated with a slow deterioration of the allograft function (3). The recognition of nonimmune factors in the pathogenesis of chronic allograft dysfunction and loss has led to the introduction of the term chronic allograft nephropathy (CAN) (4). CAN is detectable in up to 40% of grafts after the initial few months postTx and is the most common cause of late Tx failure (5).

CAN is characterized by an initial mesangial and interstitial expansion associated with arteriolar hyalinosis that progresses with time to glomerular and vascular sclerosis as well as interstitial fibrosis (6). The diagnosis of CAN is usually made late, when the histology of renal allografts reveals marked sclerotic and fibrotic changes (7). This hampers the understanding of the underlying pathogenic mechanisms. Early detection of allograft damage in the first few weeks after Tx would improve the understanding of its mediators and offer a possibility of treatment and prevention.

One theory to explain CAN suggests that changes to the allograft extracellular matrix (ECM) precede the subsequent development of tubulointerstitial fibrosis. Increased density of ECM and expansion has been attributed to both quantitative and qualitative changes (8).

Recently, tissue transglutaminase (tTg) has been proposed as a potential mediator of ECM accumulation in experimental diabetic (9) and nondiabetic (10) renal scarring. tTg is a calcium-dependent enzyme involved in crosslinking proteins through the formation of ε-(γ-glutamyl) lysine dipeptide bonds that increase the resistance of crosslinked proteins to proteolysis (11). At the cellular level, the insoluble protein polymers, resulting from ε-(γ-glutamyl) lysine crosslinking, have been reported to be involved in cells response to stress (12), cell adhesion (13), and death (14). In the extracellular compartment, evidence suggests that tTg has both the ability to increase ECM deposition by activating transforming growth factor (TGF)-β1 (15), crosslinking collagen fibrils (16), and by decreasing breakdown via conferring resistance to matrix metalloproteinase’s action (10). Changes in tTg levels have been seen in other types of fibrotic lesions including those affecting the lung, liver, and heart (17–19).

Because the characteristic interstitial fibrosis present in the chronically failing renal allograft may have its onset soon after Tx or even in the donor kidney (20), and because tTg was implicated in experimental renal scarring (10), the aim of this study is to investigate changes in tTg and its crosslink product (ε-(γ-glutamyl) lysine) in the early postTx period and determine whether these changes can be related to the development of CAN. This is done with a view both to using tTg as an early prognostic marker in the allograft and the development of tTg inhibitors to prevent CAN.

PATIENTS AND METHODS

Patients

This study involved a retrospective analysis of adult recipients having a first cadaver kidney Tx at the Sheffield Kidney Institute, Sheffield Teaching Hospital Trust, UK, between 1996 and 1998. Forty-three of the 104 recipients receiving Txs during this period had a graft biopsy during their early postTx period (27.5±18 days). Twenty-three patients had adequate biopsy material for analysis. Implantation biopsies were taken after release of the vascular clamps. The early follow-up biopsies (n=23) were obtained whenever clinically indicated or when there was an unexplained elevation of serum creatinine (sCr) by at least 30% above the baseline. Thirty-eight percent (8/23) of the recipients experienced clinical and histologic evidence of CAN over a period of 1.04±1.1 years (Table 1). These 8 patients had declining renal function (progressors) compared with the remaining 15, whose renal function remained stable (i.e., 1/sCr slope against time was not significantly negative) over a 3-year period (nonprogressors) (Table 1). The patients studied were immunosuppressed by a triple therapy protocol including cyclosporine A (CsA [Neoral, Sandoz, Basel, Switzerland]), azathioprine, and prednisolone.

T1-9
Table 1:
Donor and renal-allograft recipients’ demographic characteristics

Clinical Events

We evaluated changes in the interstitial fibrosis score in parallel with the postTx clinicobiochemical parameters. Variables included recipient’s age, sex, and body weight, mean arterial pressure (MAP), mean CsA level (Tx biopsy), acute rejection (AR), delayed graft function (DGF) (i.e., dialysis within 2 weeks postTx), sCr at 10 days postTx (sCr10d) and during CAN (sCrCAN), proteinuria, serum triglycerides, and total cholesterol. Of the donor’s characteristics, age, cause of death, and cold ischemia time were studied (Table 1).

Biopsies

A wedge biopsy was taken during Tx (n=23). A Trucut 14 gauge needle was used for follow-up biopsies. Specimens were halved, with one piece being formalin-fixed, paraffin-embedded, and sectioned to 5 μm and the other placed in OCT, snap frozen, and sectioned at 10 μm. Paraffin sections were stained with Masson’s trichrome. A minimum of 10 glomeruli within the biopsy was considered adequate for analysis.

Morphometric Analysis

A standard point counting method was used to semiquantify Masson’s trichrome stain for the estimation of renal fibrosis using a 25-point grid on consecutive fields (magnification, ×400) across the entire biopsy.

Immunohistochemistry and Immunofluorescence

The distribution of immunoreactive tTg and its crosslink product, the ε-(γ-glutamyl) lysine dipeptide bond, was determined by immunolocalization on both paraffin-embedded and unfixed cryostat sections to provide a comprehensive evaluation of location. Paraffin sections provide optimal staining for soluble (not bound to substrate protein) tTg and intracellular ε-(γ-glutamyl) lysine, whereas the cryostat protocol is optimal for insoluble (protein bound) tTg and extracellular ε-(γ-glutamyl) lysine. This is because immunolocalization of tTg on paraffin-embedded fixed tissue is difficult because of epitope occlusion, such that the detection of insoluble tTg (bound to protein [including ECM]) is poor; moreover, probing unfixed cryostat tissue necessitates the removal of soluble tTg during the staining procedure (9, 10).

Soluble tTg and Intracellular ε-(γ-glutamyl) Lysine Immunohistochemistry

Neutral buffered formalin-fixed, paraffin-embedded sections (5 μm) were deparaffinized and the target antigen revealed using target unmasking fluid (Signet, Dedham, MA). Sections were probed with a 1:250 dilution of mouse anti-ε-(γ-glutamyl) lysine monoclonal antibody (Clone 81D4, Covalab, Oullins, France) or a 1:300 dilution of mouse anti tTg (Cub7042) monoclonal (Stratek Scientific, Luton, UK) antibody. Binding was revealed using an avidin-biotin-peroxidase immunohistochemical technique. Technique specificity was determined by the replacement of the primary antibody with an equal concentration of mouse nonimmune immunoglobulin (Ig)G.

Insoluble tTg and Extracellular ε-(γ-glutamyl) Lysine Immunofluorescence

Immunofluorescence for insoluble tTg and ε-(γ-glutamyl) lysine crosslinks in insoluble proteins was carried out as previously described (9). Ten micrometer cryostat sections were washed and blocked in 3% (w/v) bovine serum albumin, 0.05% (v/v) Triton X100, 5% goat serum in phosphate-buffered saline (antibody dilution buffer, ADB) and then immunoprobed with either a 1:300 dilution of a mouse monoclonal anti tTg antibody (Cub7042) (Stratek Scientific) or 1:150 dilution of a mouse anti-ε-(γ-glutamyl) lysine monoclonal antibody (Clone 81D4; Covalab) before fixation in methanol at −20°C. All the solutions before fixation were supplemented with protease inhibitors (9).

Primary antibody binding was revealed with 1:500 dilution of a goat anti-mouse Cy5 (indodicarbocyanine)-conjugated antibody (Stratek Scientific). Confocal microscopy was performed (Leica Wetzlar, Germany), using a Kr/AR laser for both the Cy5 (excitation 650 nm) and autofluorescence (467 nm) with computer imaging and analyses obtained at 665 nm and 530 nm for Cy5 and autofluorescence, respectively. Threshold visualization levels for the photomultiplier were set using implantation biopsies. Primary antibody replacement with mouse nonimmune serum was used for technique specificity.

In Situ Transglutaminase Activity

In situ activity (ISA) measurement of tTg was performed as previously described (10). Ten micrometer cryostat sections were washed with ADB containing a cocktail of protease inhibitors and then incubated with fluorescein isothiocyanate (FITC) cadaverine (Molecular probe, Leiden, The Netherlands) in 2.5 mM CaCl2 for 1 hour at room temperature (Cadaverine is a tTg substrate). For negative controls CaCl2 was replaced with 2 mM EDTA, (calcium chelator) or cystamine (tTg inhibitor). Sections were subsequently immunoprobed with a 1:250 dilution of mouse anti-FITC monoclonal antibody (Clone FITC4, Dako, UK) followed by a 1:500 dilution of a goat anti-mouse Cy5-conjugated antibody (Stratek Scientific). Visualization and quantification using confocal microscopy were performed as above.

Statistical Analysis

Results are given as mean and standard deviation (mean±SD). Student’s t and the Mann-Whitney U tests were used for comparison between two variables. Analysis of variance was used to test differences between more than two quantitative variables with different variance. Linear regression analysis was used to detect the significance of relationships between two quantitative variables (Pearson correlation) or qualitative ones (nonparametric, Spearman’s correlation). Multivariate analysis was used to determine the predictive value of different, possibly interdependent, quantitative parameters. Cox regression analysis was used to assess the effect of variables on time to graft failure. Only graft loss through CAN was considered as Tx failure and patient death was censored. Correlation between graft survival and variables were calculated by logistic regression fitted to the data using a forward stepwise selection procedure. P <0.05 was statistically significant. Analysis was performed using the Statistical Package for Social Science (SPSS, Chicago, IL) version 9.

RESULTS

General Observations

The demographic and clinical characteristics of the study group are shown in Table 1.

Conventional Histology

Implantation histology.

At implantation, mild glomerular histologic changes were noted in five patients. These included mild glomerulopathy reflecting agonal changes, which took the form of intraglomerular capillary polymorphonu-clear leukocytes in one patient, increased mesangial matrix and cellularity in two patients, and IgA nephropathy (mesangial hypercellularity with IgA mesangial deposition) in the remaining two patients.

Tubular changes described at implantation suggested some degree of preservation injury and were characterized by mild tubular atrophy in three patients. In addition, some biopsies showed variation in tubular epithelial cell height with mild to moderate vacuolation.

Histologic changes suggestive of systemic hypertension (vasculopathy and hyalinosis) were noted in 14 patients. Three of these showed intimal thickening and elastic replication.

Early follow-up biopsy histology.

There was histologically proven AR in 14 of 23 biopsies. This was cellular in nature in five patients, predominantly vascular in another four, and combined in the remaining. Regenerative phase of acute tubular necrosis was noted in 8 of 23 patients, CsA nephrotoxicity was confirmed in 1 (marked isometric proximal tubules vacuolations) and suspected in another patient, and background IgA nephropathy in the remaining patients. Significant glomerulosclerosis was not detectable at implantation or at follow-up in any of the patients.

Late CAN biopsies.

The histology of the eight patients with CAN revealed the characteristic features of glomerulosclerosis, tubulointerstitial fibrosis, and severe arteriolopathy. In addition, significant tubular atrophy was noted with dilated tubular profiles, reflecting the loss of epithelial cell height.

Soluble Tissue Transglutaminase Immunolocalization

At implantation (Fig. 1, A to D), soluble tTg was hardly detectable by light microscopy. Vascular staining was evident, as expected, in the intimal endothelium and in smooth muscle cells of the media (Fig. 1D). Some proximal tubular epithelial cells demonstrated some weak and scattered staining (Fig. 1B), whereas minimal glomerular staining for tTg could be located in the capillary endothelium and in few mesangial cells (Fig. 1A). No staining was evident in the medulla (Fig. 1C).

F1-9
Figure 1:
Soluble tissue transglutaminase (A to L) and ε (γ-glutamyl) lysine crosslink (M to T) immunohistochemistry in paraffin-embedded sections from renal allograft recipients that subsequently developed chronic allograft nephropathy (CAN). Representative sections stained (brown) for tTg and ε (γ-glutamyl) lysine crosslink, respectively, at the time of implantation (A to D and M to P), within 3 months posttransplantation (E to H and Q to T) during CAN (I to L).

Follow-up biopsies showed increased soluble tTg in all renal compartments (Fig. 1, E to H). In proximal tubules, this was more widespread and intense. Staining was particularly evident in the loops of Henle (Fig. 1F) and the collecting ducts (Fig. 1G). At higher magnification, tTg staining was seen in the interstitial compartment of the kidney, predominantly within infiltrating cells (Fig. 1H). Glomerular staining was markedly increased in the mesangium (Fig. 1E). However, in heavily sclerosed glomeruli, there was no intraglomerular staining (Fig. 1E), although periglomerular staining occurred.

In CAN biopsies, we noted tTg glomerular staining similar to that seen in early biopsies. The intensity of glomerular expression of tTg decreased in parallel with the loss of glomerular structure (Fig. 1I). The most marked changes in soluble tTg in CAN were detectable within the tubulointerstitial compartment (Fig. 1, J to L) where increased immunostaining occurred in abnormal tubules, whether collapsing (Fig. 1J) or dilated because of flattened epithelium. Increases in interstitial tTg were observed in areas that were heavily infiltrated by inflammatory cells throughout both the cortex and the medulla (Fig. 1, K and L). This was typically seen around the vessels displaying obliterative arteriopathy.

Insoluble Tissue Transglutaminase Immunolocalization

Immunofluorescence on cryostat sections revealed minimal, but detectable, levels of substrate-bound (insoluble) tTg at implantation (Fig. 2, A and B). The staining was evident in the interstitium, particularly in areas where cellular infiltration occurred.

F2-9
Figure 2:
Insoluble tissue transglutaminase (tTg) (A to F) and ε(γ-glutamyl) lysine crosslink (G to L) immunofluorescence (Cy5 label, red) in unfixed cryostat renal allograft biopsy specimens taken at implantation (A, B, G, H) within 3 months posttransplant (C, D, I, J) or during CAN (E, F, K, L). All biopsies shown were from recipients that subsequently developed CAN. Sections were visualized and quantified by confocal microscopy using Cy5 emissions at 665 nm (red) and autofluorescence emission at 530 nm (green). Graphs show semiquantification of staining for insoluble tTg (1), ε (γ-glutamyl) lysine crosslink (2), and the correlation of ε (γ-glutamyl) lysine crosslink with Masson’s trichrome staining (3). Data represent mean emission intensity at 665 nm (mV/μm2)±SD from at least 10 fields.

Early postTx, tTg staining increased in intensity and distribution in 12 of the 23 patients including all that later developed CAN (Table 2a). This was predominantly extracellular and detected in glomerular and tubulointerstitial areas (Fig. 2, C and D). Glomerular insoluble tTg (Fig. 2C) was again in the mesangium of glomeruli, especially those showing significant damage. With more severe injury, the expression of tTg was also noticeable in the periglomerular areas along with periglomerular fibrosis. Tubulointerstitial staining of insoluble tTg was largely peritubular, with expanded interstitium displaying more intense staining (Fig. 2D).

T2-9
Table 2:
Patient numbers and proportional risk

In CAN kidneys (Fig. 2, E and F), extensive and intense insoluble tTg staining occurred. Glomerular capillary staining was of such intensity that it masked the glomerular structure (Fig. 2E). In the tubulointerstitium, insoluble tTg was seen in the expanded interstitial areas surrounding damaged tubules (Fig. 2F) and in infiltrating inflammatory cells.

Semiquantification of Insoluble Tissue Transglutaminase

Immunostainable tTg (Cy5 emission intensity by confocal microscopy) increased significantly from implantation to follow-up (Fig. 2), with a correlation between the two values (r=0.472, P =0.031). This increase was detected not only in the early biopsies compared with implantation but also in CAN compared with the early changes (Fig. 2). Multiple linear regression analysis highlighted potential predictors for the increased tTg at follow-up, including tTg at implantation (R2=0.426, P =0.025), serum creatinine 10 days postTx (sCr10d) (R2 = 0.45, P =0.05), and proteinuria of more than 1 g/day during the follow-up period (R2=0.451, P =0.05).

Tissue Transglutaminase In Situ Activity (tTg-ISA)

Distribution of tTg ISA.

At implantation, the tTg-ISA was minimal in the glomeruli (Fig. 3A) and tubulointerstitial space (Fig. 3B). In early biopsies, glomeruli displayed increased staining for tTg-activity, mainly in what appeared to be the mesangium, basement membrane, and Bowman’s capsule (Fig. 3C). In the tubulointerstitial compartment, similar increases in Tg activity were noted, particularly in areas of expanded interstitium (Fig. 3D). In CAN biopsies, after the tTg-ISA assay, the morphology was particularly poor. However, it was still possible to demonstrate higher tTg-ISA in both glomerular (Fig. 3E) and tubulointerstitial (Fig. 3F) areas. These changes in both early and CAN biopsies were consistent with those seen for insoluble tTg. Inhibition of Tg activity using either EDTA or cystamine prevented the incorporation of fluorescein cadaverine (Fig. 3, G and H)

F3-9
Figure 3:
Tissue transglutaminase in situ activity (tTg-ISA) imaging at implantation (A and B), within 3 months posttransplantation (C and D), and during CAN (E and F). (G and H) Sections preincubated with EDTA to inhibit Tg activity. Sections were visualized and quantified by confocal microscopy using Cy5 emissions at 665 nm (red) and autofluorescence emission at 530 nm (green).

Semiquantification of Tg-ISA.

In the early biopsies, there was a marked increase in the level of tTg-ISA from implantation (Fig. 3). In the eight allografts that progressed to CAN, the level of tTg-ISA increased significantly compared with the earlier biopsies. Preincubation of the sections with EDTA reduced the emission intensity to a level that was significantly lower than the level at implantation. There was a strong correlation between insoluble tTg immunofluorescence and Tg-ISA (r=0.83, P =0.005).

ε-(γ-glutamyl) Lysine Immunolocalization

Localization of ε-(γ-glutamyl) lysine on fixed paraffin sections is effective at visualizing intracellular ε-(γ-glutamyl) lysine crosslinking (9). At implantation, little staining was evident (Fig. 1, M to P), with limited to weak staining in a few dispersed proximal tubules (Fig. 1O). After implantation, intracellular ε-(γ-glutamyl) lysine was increased (Fig. 1, Q to T) in 11 of the 23 biopsies, including all that went on to develop CAN (Table 2a). This was mainly tubular, occurring typically in areas of damage rich in infiltrating cells. Loop of Henle and collecting duct epithelia demonstrated particularly intense staining (Fig. 1, S and T). There was little difference in intracellular ε-(γ-glutamyl) lysine staining between early biopsy and CAN.

Immunofluorescence on cryostat sections for extracellular ε-(γ-glutamyl) lysine crosslinking was minimal at implantation (Fig. 2, G and H). Early postTx, staining became stronger and extensive. In damaged glomeruli, ε-(γ-glutamyl) lysine staining was seen in the mesangium and periglomerular (Fig. 2I). Tubulointerstitial staining was peritubular (Fig. 2J) and clearly extracellular most likely within the basement membrane. Biopsies with expanded interstitium had more intense staining that was predominantly surrounding infiltrating cells in a similar manner to that of tTg. In CAN, both glomerular and tubulointerstitial staining was further amplified (Fig. 2, K and L), with extensive staining throughout both compartments.

Semiquantification of Extracellular ε-(γ-glutamyl) Lysine Crosslink Staining

With the use of Cy5 emission intensity, increases in ε-(γ-glutamyl) lysine occurred from implantation to early biopsies (Fig. 2). There was also a correlation between changes of both insoluble tTg and its crosslink product (r=0.578, P <0.01) as well as between Masson’s trichrome staining and ε-(γ-glutamyl) lysine (r=0.781, P <0.001) (Fig. 2). Quantification of ε-(γ-glutamyl) lysine in CAN biopsies showed an increase in ε-(γ-glutamyl) lysine between early and CAN biopsies (P <0.0001) (Fig. 2). Both insoluble tTg (r=0.79, P <0.01) and tTg-ISA (r=0.808, P <0.01) correlated with ε-(γ-glutamyl) lysine crosslinks.

Predictors of ε-(γ-glutamyl) Lysine Crosslink Accumulation

sCr 10 days postTx (R2=0.9, P =0.046) and the serum trig-lycerides (R2=0.9, P =0.033) were predictors of the elevated ε-(γ-glutamyl) lysine crosslink. Insoluble tTg and ε-(γ-glutamyl) lysine in the donor’s kidney predicted the increased ε-(γ-glutamyl) lysine at early follow-up (R2=0.462, coefficient±SE=0.826±0.259, P =0.006, and=1.041±0.367, P =0.013, respectively). Increased levels of insoluble tTg and ε-(γ-glutamyl) lysine were associated with a significant reduction in the renal function (R2=0.484, coefficient±SE=0.255±0.110, P =0.034 and=1.437±0.385 and P =0.002, respectively).

Of the clinical postTx factors, three were risk factors for increased ε-(γ-glutamyl) lysine at follow-up. These were elevated MAP (Tx to early postTx period, R2=0.724, coefficient±SE=0.170±0.043, P =0.003), DGF (R2=0.518, P =0.05), and AR (R2=0.735, P =0.007). Donors age was not associated with either ε-(γ-glutamyl) lysine (R2=0.018, NS) or tTg (R2=0.008, NS) levels in the implant or subsequent biopsies.

Tissue Transglutaminase, ε-(γ-glutamyl) Lysine, and Graft Outcome

Computer imaging and quantification indicated the eight recipients who developed CAN had higher levels of both insoluble tTg and ε-(γ-glutamyl) lysine bonds in the implantation biopsy compared with those with stable renal function (Fig. 4). Furthermore, higher levels of tTg and ε-(γ-glutamyl) lysine occurred in biopsies taken within the first few weeks of Tx from recipients who progressed to CAN compared with the nonprogressors (Fig. 4). Computer analysis was confirmed by manual examination of the sections by two of the authors (HAZ and JS) blinded to the specimen code using a 5-point scoring scheme on a standard fluorescent microscope (data not shown).

F4-9
Figure 4:
Representative images of insoluble tTg staining in four recipients whose graft survived for 3 years and four that failed at both implantation (1) and 3 month posttransplantation (2). Graphs show semiquantification by image analysis of all studied allografts at the same time points showing the differences in tTg (A and B) and ε (γ-glutamyl) lysine crosslink (C and D) in surviving and failed kidneys. Data represents the mean Cy5 emission intensity (mV/μ (2))±SD at 665 nm from at least 10 fields.

In the CAN-recipients, tTg was correlated with the Masson’s trichrome staining (r=0.843, P =0.004) as was ε-(γ-glutamyl) lysine bonds (r=0.622, P =0.05). The percent change of tTg and that of ε-(γ-glutamyl) lysine were significantly correlated with poorer graft outcome (logistic P =0.0297 and 0.0013, respectively). To determine whether tTg and ε-(γ-glutamyl) lysine were superior to other recognised histologic predictors of CAN measured previously in this patient group (21), a multivariate Cox regression analysis was performed and a relative hazard ratio of graft failure calculated (Table 2b). This demonstrated that, on both implantation and early biopsies, ε-(γ-glutamyl) lysine and particularly tTg levels are superior to other predictors, including both Masson’s trichrome and collagen III staining. However, the highest risk of graft failure appears to be associated with the change in tTg levels between implantation and first biopsy.

DISCUSSION

Changes in tTg in experimental models of fibrosis (10,17–19) have highlighted the need to evaluate tTg’s involvement in human disease. Sequential biopsies taken postrenal Tx offer a unique opportunity to study changes in tTg in the same kidney, while also offering a new insight into noninflammatory pathways underlying CAN. Here, we have evaluated both tTg and ε (γ-glutamyl) lysine crosslink in biopsies from renal-allograft recipients taken at implantation, within 3 months postTx, and years after Tx to determine the potential role of this enzyme and its product in allograft scarring. Although the number of the patients studied is relatively small, the large number of sequential biopsies allowed us to explore changes in the expression of tTg and its product with time and correlate it with histologic changes.

Increased levels of tTg and its ε-(γ-glutamyl) lysine product occur in renal allografts within a few weeks from Tx and before established fibrosis is detectable by conventional methods. Extracellular renal tTg and ε (γ-glutamyl) lysine increase as CAN develops, with changes correlating with pathology and function both in early and late biopsies. Becaused we did not undertake routine protocol biopsies and sample the allografts of patients whose renal function were stable, we cannot speculate as to the nature of tTg changes in these allografts. However, the CAN specimens biopsied had a range of severity of renal pathology allowing correlation analyses. Our previous work on native human kidney biopsies support a close association of renal fibrosis and tTg expression and crosslinked ECM deposition (22).

The intracellular accumulation of tTg in endothelial, mesangial, tubular, and infiltrating cells highlighted these as major sources of the enzyme. The predominant extracellular location of the ε (γ-glutamyl) lysine crosslink indicated that the externalized tTg was active; however, a number of tubular epithelial cells had intense intracellular ε (γ-glutamyl) lysine staining, suggesting loss of Ca2+ homeostasis leading to activation of tTg. These results are consistent with our previous findings in experimental renal scarring models (10); however, the infiltrating cell staining here is considerably greater.

The infiltrating cell population in allografts represents a mix of inflammatory and fibroblast cell types. We had previously noted the presence of tTg in infiltrating cells in experimental renal scarring, although they were few and given little importance (10). Studies in a range of human nephropathies have suggested greater interstitial tTg expression in native human kidney disease that was confirmed by in situ hybridisation (22). However, interstitial levels were considerably less than that described here in allografts and may represent a specific allograft biology. The distribution of the interstitial tTg-positive cells is reminiscent of that of myofibroblasts associated with progressive interstitial fibrosis. In addition, a proportion tTg-positive cells may be of monocytic origin. Macrophages are known to up-regulate tTg once stimulated and have a clearly defined role for tTg in cell adhesion (23).

The glomerular distribution of soluble tTg varied with glomerular damage. There was essentially no staining in globally sclerosed glomeruli or apparently normal glomeruli. However, strong expression occurred in glomeruli with mild to moderate glomerulosclerosis, suggesting that tTg is up-regulated in response to damage but shut down when the remodeling process is complete. In contrast, insoluble (extra-cellular) tTg and ε(γ-glutamyl) lysine increased with the level of glomerular damage. Unlike soluble Tg, which has a half-life of approximately 12 hours (24), insoluble tTg is not cleared but incorporated into the insoluble protein matrix that encapsulates the glomerulus and appears to provide a good indicator of glomerular damage. The translocation of insoluble tTg from the mesangial and endothelial cells is likely to cause this accumulation of extracellular tTg. Subsequently, this would lead to ε (γ-glutamyl) lysine dipeptide bonds within the glomerular basement membrane and within the mesangium. The irreversible crosslinking of the glomerular ECM may lead to stabilization of the matrix and ultimately progressive glomerulosclerosis. Glomerular over-expression of tTg in the mesangium may be the forerunner of allograft glomerulopathy reported in CAN, where glomerular hypertrophy and mesangial matrix expansion are well documented (25). The over-expression of tTg within mesangial cells during the course of CAN is reminiscent of changes observed during the course of atherosclerosis in vascular smooth muscle cells (26).

Regardless of the renal compartment where changes are observed, changes in insoluble tTg and ε (γ-glutamyl) lysine occur extracellularly and likely within the ECM. ECM crosslinking may cause ECM accumulation by increasing the rate of ECM deposition by way of nonconventional deposition pathways (16, 27) and by stabilizing the ECM to the action of proteinases (10). This extracellular action of tTg also plays an important role in the activation of the large latent TGFβ1 complex (15), resulting in local increases in active TGFβ1 with its well-characterized fibrotic potential (28).

The tubular intracellular localization of tTg may suggest a role in cell death and tubular atrophy. tTg has been associated with the formation of the apoptotic envelope (14). However, few of the ε (γ-glutamyl) lysine staining cells showed morphologic characteristics of a apoptosis. Previous studies have shown this crosslinking can occur independently of apoptosis (29) and may contribute to cell death through an alternative mechanism termed transglutaminase-mediated cell death (30). The high intracellular ε (γ-glutamyl) lysine crosslink observed in this study suggests that this is a common occurrence in CAN, contributing significantly to tubular atrophy associated with scarring and loss of renal function.

An unexpected finding of this study is the potential use of tTg and ε (γ-glutamyl) lysine as early and statistically better histologic markers of graft outcome than previously reported in this patient group (21). If the sample population is divided into kidneys that develop CAN and those that continue to function, we noted that at both implantation and first biopsy, higher levels of insoluble tTg and ε (γ-glutamyl) lysine were detectable in allografts that subsequently developed CAN.

The early Tx biopsy values may be influenced by AR and DGF, with both associated with increased levels of ε-(γ- glutamyl) lysine crosslinking in this analysis. Interestingly, levels of tTg and ε (γ-glutamyl) lysine in the allograft at the time of Tx are almost as good predictors of graft outcome as in the early postTx biopsy and indicates that kidneys with higher tTg and ε (γ-glutamyl) lysine may be less suitable for Tx. This may simply reflect early fibrosis or damage of the kidney before Tx, although there was no obvious histologic evidence. Alternatively, the role of tTg as an acute-phase wound-response enzyme raises the possibility that some kidneys release more tTg in response to the stress of ischemia-reperfusion injury and Tx, causing the initiation of the scarring process at Tx. Increases in tTg and crosslink between implantation and first biopsy are a particularly useful indicator of graft failure.

The data presented in this study clearly demonstrates that there is an increase of both tTg and ε-(γ-glutamyl) lysine dipeptide crosslinks in renal-allograft biopsies that are associated with the development of renal fibrosis. tTg and ε-(γ-glutamyl) lysine dipeptide are good predictors of graft outcome and may serve as valuable prognostic markers and therapeutic targets in the future.

REFERENCES

1. Gjerston D. Multifactorial analysis of renal transplants reported to the United Network for Organ Sharing Registry. In: Terasaki PI, Cacka A, eds. Clinical transplantation. Los Angeles, UCLA Tissue Typing Laboratory 1992, p. 299.
2. Tullis S, Tilney N. Both alloantigen-dependent and independent factors influence chronic allograft rejection. Transplantation 1995; 59: 313.
3. Hayry P, Isoneime H, Yilmaz S, et al. Chronic allograft rejection. Immunology 1993; 134: 33.
4. Tilney N, Paul L. Alloantigen-dependent events leading to chronic graft dysfunction. In: Tilney NL, Strom TB, Paul LC, eds. Transplantation biology: Cellular and molecular aspects. Philadelphia. New York, Lippincott-Raven 1996, p. 567.
5. Paul LC. Chronic allograft nephropathy: an update. Kidney Int 1999; 56: 783.
6. Racusen LC, Solez K, Colvin RB, et al. The Banff 97 working classification of renal allograft pathology. Kidney Int 1999; 55: 713.
7. Solez K, Axelsen RA, Benediktsson H, et al. International standardization of criteria for the histologic diagnosis of renal allograft rejection: the Banff working classification of kidney transplant pathology. Kidney Int 1993; 44: 411.
8. Eddy AA. Molecular basis of renal fibrosis. Pediatr Nephrol 2000; 15: 290.
9. Skill NJ, Griffin M, El Nahas AM, et al. Increases in renal epsilon-(gamma-glutamyl)-lysine crosslinks result from compartment-specific changes in tissue transglutaminase in early experimental diabetic nephropathy: pathologic implications. Lab Invest 2001; 81: 705.
10. Johnson TS, Skill NJ, El Nahas AM, et al. Transglutaminase transcription and antigen translocation in experimental renal scarring. J Am Soc Nephrol 1999; 10: 2146.
11. Folk JE, Finlayson JS. The epsilon-(gamma-glutamyl)lysine crosslink and the catalytic role of transglutaminases. Adv Protein Chem 1977; 31: 1.
12. Aeschlimann D, Thomazy V. Protein crosslinking in assembly and remodelling of extracellular matrices: the role of transglutaminases. Connect Tissue Res 2000; 41: 1.
13. Johnson TS, Knight CR, el-Alaoui S, et al. Transfection of tissue transglu-taminase into a highly malignant hamster fibrosarcoma leads to a reduced incidence of primary tumour growth. Oncogene 1994; 9: 2935.
14. Fesus L, Madi A, Balajthy Z, et al. Transglutaminase induction by various cell death and apoptosis pathways. Experientia 1996; 52: 942.
15. Nunes I, Gleizes PE, Metz CN, et al. Latent transforming growth factor-beta binding protein domains involved in activation and transglutaminase-dependent cross-linking of latent transforming growth factor-beta. J Cell Biol 1997; 136: 1151.
16. Kleman JP, Aeschlimann D, Paulsson M, et al. Transglutaminase-catalyzed cross-linking of fibrils of collagen V/XI in A204 rhabdomyosar-coma cells. Biochemistry 1995; 34: 13768.
17. Small K, Feng JF, Lorenz J, et al. Cardiac specific overexpression of trans-glutaminase II (G(h)) results in a unique hypertrophy phenotype independent of phospholipase C activation. J Biol Chem 1999; 274: 21291.
18. Mirza A, Liu SL, Frizell E, et al. A role for tissue transglutaminase in hepatic injury and fibrogenesis, and its regulation by NF-kappaB. Am J Physiol 1997; 272: G281.
19. Griffin M, Smith LL, Wynne J. Changes in transglutaminase activity in an experimental model of pulmonary fibrosis induced by paraquat. Br J Exp Pathol 1979; 60: 653.
20. Nicholson ML, Harper SJ, Wheatley TJ, et al. Renal transplant fibrosis: histomorphometric assessment of early renal transplant biopsies for markers of chronic rejection. Transplant Proc 1997; 29: 2793.
21. Abo-Zenah H, Katsoudas S, de Takats D, et al. Early progressive interstitial fibrosis in human renal allografts. Clin Nephrol 2002; 57: 9.
22. Johnson TS, El-Koraie AF, Skill NJ, et al. Tissue transglutaminase and the progression of human renal scarring. J Am Soc Nephrol 2003; 14: 2052.
23. Akimov SS, Belkin AM. Cell surface tissue transglutaminase is involved in adhesion and migration of monocytic cells on fibronectin. Blood 2001; 98: 1567.
24. Verderio E, Nicholas B, Gross S, et al. Regulated expression of tissue trans-glutaminase in Swiss 3T3 fibroblasts: effects on the processing of fibronectin, cell attachment, and cell death. Exp Cell Res 1998; 239: 119.
25. Cheigh JS, Mouradian J, Soliman M, et al. Focal segmental glomerulosclerosis in renal transplants. Am J Kidney Dis 1983; 2: 449.
26. Goumenos DS, Brown CB, Shortland J, et al. Myofibroblasts, predictors of progression of mesangial IgA nephropathy? Nephrol Dial Transplant 1994; 9: 1418.
27. Akimov S, Belkin AM. Cell-surface transglutaminase promotes fibronectin assembly via interaction with the gelatin-binding domain of fibronectin: a role in TGF beta-dependent matrix deposition. J Cell Sci 2001; 114: 2989.
28. Border WA, Noble NA. TGF-beta in kidney fibrosis: a target for gene therapy. Kidney Int 1997; 51: 1388.
29. Johnson TS, Griffin M, Thomas GL, et al. The role of transglutaminase in the rat subtotal nephrectomy model of renal fibrosis. J Clin Invest 1997; 99: 2950.
30. Johnson TS, Scholfield CI, Parry J, et al. Induction of tissue transglutaminase by dexamethasone: its correlation to receptor number and transglutaminase-mediated cell death in a series of malignant hamster fibrosarcomas. Biochem J 1998; 331: 105.
© 2004 Lippincott Williams & Wilkins, Inc.