Cyclosporine A (CsA) is a calcineurin inhibitor (CNI) with powerful immunosuppressant properties and has been widely used in the last three decades to prevent allograft rejection (1).
Even though there is not one fully specific marker for the chronic nephrotoxicity of CsA, tubular atrophy and hypocellular interstitial fibrosis have been extensively reported in the native or transplanted kidneys from patients exposed to the drug (2–4). Whether it is a significant cause of graft loss is currently challenged, all the more because new biomarkers such as C4d and donor-specific antibodies have allowed for a better detection of humoral rejection, the burden of which is at present recognized as major (5–8).
What role remains for CNI chronic toxicity in the modern era, then? Although in retrospect we accept that this mechanism of fibrogenesis was too often diagnosed by default, it does not exclude that CsA will play a synergistic role and facilitate the progression of interstitial fibrosis in grafts with an (often immune-mediated) epithelial or endothelial injury. Paradoxically, now that we better discern other specific causes of graft loss, the role of CNI could become underestimated (9).
The aim of this study was to test whether CNI could induce phenotypic changes in tubular epithelial cells using biomarkers that have been validated in the context of fibrogenesis. Our group has demonstrated that the so-called epithelial to mesenchymal transition (EMT) markers have a good predictive value with respect to interstitial fibrosis and tubular atrophy in human kidney recipients (10–12). The existence of a fibrogenic EMT is itself a controversial issue in adult organs (13). By definition, EMT is a biological process that allows a polarized epithelial cell, which normally interacts with the basement membrane through its basal surface, to undergo changes that enable it to assume a mesenchymal cell phenotype and thus to produce extracellular matrix components. EMT is also useful to disperse cells in the context of embryogenesis and tumorigenesis (14). Whether it similarly provides tubular epithelial cells with the capacity to cross their basement membrane and swell the population of interstitial myofibroblasts is vividly debated, up to the point that the Banff working group has decided to coin a new term to describe the epithelial to mesenchymal switch that has been observed in injured kidneys, lungs, and livers.
Regardless, two biomarkers reminiscent of EMT have been found by us and others to reflect fibrogenesis in renal grafts (11, 15–18): the de novo expression of vimentin and the translocation of beta-catenin into the cytoplasm (10, 19). These studies were performed in transplanted patients who were all treated by CsA, and in vitro CsA has been reported to cause phenotypic and functional changes typical of EMT in renal epithelial cells (20). Full-blown EMT was never confirmed in vivo in rodents exposed to CsA, but the de novo expression of vimentin, a hallmark of EMT, has been observed (21).
Overall, it is reasonable to postulate that CsA could, either directly or indirectly (through ischemic signals induced by the vasoconstriction of the afferent arteriole or through activation of the renin-angiotensin-aldosterone and the nitric oxide-endothelin systems [22, 23]), induce an epithelial to mesenchymal phenotypic switch in tubular cells that will eventually contribute to fibrogenesis through tubular atrophy.
CsA Toxicity at 2 Weeks: Characterization of Renal Lesions
Rats were euthanized after 2 weeks of exposure to CsA. At that time, median concentration of CsA in the serum was 2307 ng/mL (1986–4800), and renal function was significantly impaired (median creatinine concentration was 49 μmol/L, as opposed to 24 μmol/L in the control group, P<0.05) (see SDC 1,http://links.lww.com/TP/A495). In CsA-exposed rats, histological evaluation of renal lesions by Masson's trichrome allowed for the detection of spots of proximal tubule vacuolization and occasionally tubules made of a thin epithelium with a seemingly enlarged lumen. Interstitial fibrosis and nodular hyalinization of the arterioles were not observed (of note, in another set of experiments, 4 weeks of exposure to CsA did not induce significant fibrosis either, data not shown).
Blood pressure and proteinuria were not significantly different although blood pressure tended to be lower in CsA-treated rats in line with previous studies using this model (24).
Up-Regulation of Mesenchymal Genes in the Renal Cortex
We studied the messenger RNA (mRNA) expression of key mesenchymal genes in the renal cortex from rats treated by CsA or its vehicle for 2 weeks (Fig. 1). Snail, vimentin, S100A4, and collagen III transcripts were found significantly increased under CsA.
Identification of EMT Markers in the Tubular Epithelium
We reasoned that the aforementioned mesenchymal genes could be up-regulated either by the activation of interstitial fibroblasts into myofibroblasts or by the reprogramming of tubular epithelial cells to a mesenchymal state. Comparative evaluation by Masson's trichrome and immunohistochemistry for α-smooth muscle actin, a marker for myofibroblasts, in the kidneys from rats treated with CsA or its vehicle showed no quantitative or qualitative change that could reflect the up-regulation of mesenchymal genes in particular. Supplemental Figure 2 (see SDC 2,http://links.lww.com/TP/A496) shows the absence of expression of α-smooth muscle actin, a marker of myofibroblasts, outside the vessels, both in control (CTL) and CsA groups. However, tubular epithelial cells from rats exposed to CsA displayed phenotypic changes compatible with an EMT-like repair response: vimentin, an intermediate filament, and HSP47, a marker for collagen secretion, were found to be abnormally expressed; likewise, the linear and peripheral expression of β-catenin that is characteristic of epithelial cells was often replaced by a diffuse expression in the cytoplasm (Fig. 2). This phenotypic switch affected all layers of the renal cortex but was predominant in the deeper half of the cortex (see SDC 3,http://links.lww.com/TP/A497). Although tubules containing intracytoplasmic vacuoles were constantly expressing these markers, most of the EMT+ tubules were morphologically normal and thus indistinguishable from the tubules of vehicle-treated rats when examined only by standard histological techniques. Similar findings were obtained in rats treated with a lower dose of CsA (7.5 mg/kg/day), although to a lesser extent (see SDC 4,http://links.lww.com/TP/A498). Such changes were not observed in epithelial cells from rats exposed to the vehicle of CsA or to sirolimus, used in this study as a second, non-CNI control group (see SDC 5,http://links.lww.com/TP/A499).
Neither Angiotensin II Nor Endothelin Receptor Blockade Prevents CsA-Induced EMT
Irbesartan (an angiotensin II receptor antagonist) and bosentan (an endothelin receptor antagonist), both of which are commercially available for humans and are potentially able to promote the regression of renal fibrosis, were given to rats to prevent or blunt the CsA-induced reprogramming of epithelial cells, but to no avail (see SDC 6,http://links.lww.com/TP/A500 and SDC 7,http://links.lww.com/TP/A501). Renal failure, overexpression of mesenchymal genes in the renal cortex, and expression of EMT markers in the renal tubules were all unaffected.
Regression of EMT Markers After CsA Withdrawal
We studied the kinetics of the clinical, histological, and molecular markers of CsA-induced injury every 2 weeks for 6 weeks after CsA wash-out (Figs. 3 and 4). Two weeks after CsA withdrawal, renal function had returned to normal, and both tubular vacuolization and tubular dilatation had markedly decreased in the CsA-treated animals. By contrast, EMT markers, both at the mRNA and protein levels, persisted in the CsA-treated rats. Snail mRNA even increased during the 2 weeks after the withdrawal of CsA. Four and 6 weeks after CsA withdrawal, the vacuolization of tubules had vanished, as had tubular dilatation, except for rare spots rounded with interstitial fibrosis (Fig. 4). Four weeks after withdrawal, mRNA expression of mesenchymal genes in the cortex was similar whether rats had been exposed to CsA or not, and the expression of EMT marker expression was no more seen in renal tubules. The reversibility of EMT was not seen in rats for which CsA was not withdrawn (see SDC 8,http://links.lww.com/TP/A502).
In this study, we demonstrated that in rats, the administration of CsA for weeks leads to a marked up-regulation of mesenchymal genes in the renal cortex, which is explained by the reprogramming of epithelial cells, and not by the activation of interstitial fibroblasts. Incidentally, the biomarkers we have used in this study are also predictive of the progression of interstitial fibrosis in human kidney recipients, which is indirect evidence that CsA could (synergistically with other causes of injury) contribute to renal fibrogenesis. The scores for EMT marker expression were significantly correlated with tubular vacuolization and tubular atrophy, showing that epithelial changes reflect the intensity of the renal injury. However, EMT markers were expressed not only in tubules displaying tubular vacuolization and/or atrophy but also in tubules that were obviously exposed to CsA but morphologically normal. This suggests that EMT precedes (and thus potentially causes) tubular atrophy. We would like to stress out that these biomarkers probably reflect a generic “repair” response of the parenchyma to some injury and are in no way specific for CsA toxicity. Acute ischemia, rejection, and probably other factors will trigger them as well.
Another important finding of our study is the reversibility of the CsA-induced phenotypic changes, within a month after the withdrawal of the drug. Although we were unable to translate this into hard endpoints such as fibrosis, because rodents are resistant to fibrosis and 2 to 4 weeks of exposure to CsA is most probably insufficient to induce accumulation of collagens, this suggests that CNI-sparing or withdrawal strategies are relevant therapeutic options to protect the renal epithelium, at least in those patients who develop interstitial fibrosis without any apparent cause. In line with this, we recently reported that CsA withdrawal abolishes the negative predictive value of early EMT with respect to renal graft fibrogenesis in human recipients (25). At first sight, our results are contradictory with that reported in a signal study by Elzinga et al. (24), in which rats treated with CsA for 28 days presented with an advanced tubulointerstitial damage that even progressed after CsA withdrawal. In that study, however, the CsA plasmatic levels were still at 46 ng/mL at 28 days after CsA withdrawal. In contrast, we found undetectable levels of CsA 2 weeks after CsA withdrawal. Therefore, although small amounts of CsA will perpetuate tubular injury, strict cessation of CsA exposure does allow for the regression of tubular lesions and of EMT markers' expression. A prospective trial is currently ongoing that will address the usefulness at the bedside of measuring the EMT score in kidney recipients, to avoid CsA nephrotoxicity (clinical trial CERTITEM, #NCT01079143).
Of note, inhibitors of the renin angiotensin (admittedly nephroprotective in native kidneys) or of the NO-endothelin systems, both mechanistically involved in CsA renal effects, failed to prevent the CsA-induced phenotypic alterations in tubular cells. This finding is consistent with clinical studies performed in large cohorts of kidney transplanted patients treated either by angiotensin conversion enzyme inhibitors or by angiotensin II receptor antagonists, in which no specific benefit was observed with respect to the progression of graft failure (26, 27). With respect to endothelin, it is up-regulated by CsA in vascular and tubular cells exposed to CsA (28, 29) and it was shown to induce EMT in some tumors (30). In our hands, though, an endothelin receptor antagonist (bosentan) was ineffective as well.
Our study has limitations. First, rats were experimentally exposed to concentrations of CsA that are higher than the target range in humans. However, the doses we used are commonly used to induce immunosuppression in rats (31–33) and also in keeping with previously published model of CsA nephrotoxicity (24). Second, we studied the effect of CsA independently from any human leukocyte antigen mismatching. Better yet, our findings are not polluted by any superimposed effect the surgical procedure or some immunological conflict could have on the tubular epithelium.
In conclusion, we believe that in animals exposed in vivo to CsA, the expression of EMT markers by renal epithelial cells reflects a parenchymal injury that is reversible provided CsA is withdrawn on time.
MATERIALS AND METHODS
Four sets of rats were used for the study of cyclosporine renal toxicity. The first set included four groups of six rats treated by vehicle, irbesartan, CsA, or both. The second set included four groups of six rats treated by vehicle, bosentan, CsA, or both. The third set included a group of five rats treated for 2 weeks with vehicle, a group of eight rats treated for 2 weeks with CsA, and six groups of five rats treated either by vehicle or CsA for 2 weeks and killed 2, 4, or 6 weeks later. In a fourth set, rats were treated with CsA 7.5 mg/kg/day for 2 weeks (n=3) and with CsA 15 mg/kg/day for 30 days (n=6). In addition, six rats were given sirolimus 50 μg/day for 2 weeks as a supplementary control. Sprague-Dawley male rats with weights of 300 to 325 g were fed a low-salt diet (0.025% Na) for a week before and during the experiment. A low-salt diet is known to accelerate CsA-mediated tubulointerstitial injury in rats (24). For the wash-out studies, the rats were fed with standard laboratory chow from the time CsA was interrupted. CsA (Neoral; Novartis Pharma, Rueil-Malmaison, France) was administered subcutaneously every day for 2 weeks at a dose of 15 mg/kg, and ricin oil (Cooper, Melun, France), CsA's vehicle, was used as the control. The angiotensin II receptor (AT1) antagonist irbesartan (Aprovel; Bristol Myers Squibb, Rueil-Malmaison, France) and the endothelin receptor antagonist bosentan (a gift from Actelion, Allschwil, Switzerland) were administered daily by gavage as suspensions in 10% arabic gum at doses of 20 and 100 mg/kg, respectively. A suspension of 10% arabic gum was used as the control.
The rat was killed by exsanguination under general anesthesia with thiopental, blood was collected by aortic catheterization, and the kidneys were then perfused with saline and harvested for fixation in formalin-acetic acid-alcohol and congelation at −80°C.
All animal experimentations were performed in conformity with good practices for animal care and were authorized by the referee for ethics in animal experimentation at Inserm U702.
Renal function (assessed by creatininemia, uremia, and proteinuria) was assessed at the time of killing.
Renal mRNA expression was quantified by real-time polymerase chain reaction (PCR) from cortex fragments frozen in RNALater (Qiagen, Courtaboeuf, France). RNA extraction was performed with TRIzol solution (Life Technologies, BRL, Gaithersburg, MD). RNA quality was checked by measuring the ratio of optical densities at 260 and 280 nm. We used reverse transcription with Superscript II (Life Technologies BRL) to convert 1 μg of RNA into complementary DNA, which was then amplified by PCR using a LightCycler 480 (Roche Diagnostic, Meylan, France), Sybr Green (Fast Start DNA Master Sybr Green I; Roche Applied Science, Roche Diagnostic), and specific primers designed by Roche (Universal Probe Library) for snail, vimentin, S100A4, and collagen III under the following conditions: 95°C for 5 min, 45 cycles at 95°C for 15 sec, 60°C for 15 sec, and 72°C for 15 sec. PCR was also carried out for the hypoxanthine-guanine phosphoribosyltransferase (HPRT) housekeeping gene. We expressed the results as 2deltaCp, where Cp is the cycle threshold number and normalized these results with HPRT. We analyzed the dissociation curves after each run for each amplicon to assess the specificity of amplification when using Sybr Green.
All histological quantifications were performed blinded. Histological assessments of the kidneys were performed on 3-μm-thick paraffin slides with Masson's trichrome for morphology. Proteic expression of the EMT markers was studied by immunohistochemistry using antibodies directed against vimentin, β-catenin, and HSP 47 as previously described (11). Threshold of expression for each protein was established as to optimize the sensitivity and specificity of these markers in a preliminary experiment (data not shown). Vimentin is an intermediate filament normally absent in epithelial cells. A tubular section was counted as positive for vimentin if at least two cells within it expressed vimentin. β-catenin is a protein that can link membrane cadherins to the actin cytoskeleton under the cytoplasmic membrane in epithelial cells and act as a mesenchymal transcription factor when it is free to enter the nucleus. β-catenin change of expression is especially useful to detect the alteration of the epithelial phenotype, as other epithelial markers such as cadherins or cytokeratins are physiologically absent from a significant percentage of normal tubules, and thus their decrease is subject to doubt. β-catenin change was defined as an interruption in basal expression and intracytoplasmic redistribution of β-catenin in at least one cell in a tubular section. HSP47 is a chaperone protein for collagen and is normally absent in epithelial cells. The criterion for HSP47 positivity was the expression of HSP47 in at least three cells of a single tubular section.
The results were obtained using the JMP 8.0.1 software, and medians were given with their interquartile range. Comparisons between two groups were performed using the Wilcoxon rank-sum test. Results were considered significant with P less than 0.05.
See Supplemental Methods 1 and 2 for PCR primers and immunohistochemistry procedures (SDC 9,http://links.lww.com/TP/A503).
The authors thank Edith Baugey for her excellent technical assistance in immunohistochemistry.
1. Calne RY, White DJ, Thiru S, et al. Cyclosporin A in patients receiving renal allografts from cadaver donors. Lancet
1978; 2: 1323.
2. Myers BD, Ross J, Newton L, et al. Cyclosporine-associated chronic nephropathy. N Engl J Med
1984; 311: 699.
3. Naesens M, Kuypers DR, Sarwal M. Calcineurin inhibitor nephrotoxicity. Clin J Am Soc Nephrol
2009; 4: 481.
4. Nankivell BJ, Borrows RJ, Fung CL, et al. The natural history of chronic allograft nephropathy. N Engl J Med
2003; 349: 2326.
5. El-Zoghby ZM, Stegall MD, Lager DJ, et al. Identifying specific causes of kidney allograft loss. Am J Transplant
2009; 9: 527.
6. Gaston RS, Cecka JM, Kasiske BL, et al. Evidence for antibody-mediated injury as a major determinant of late kidney allograft failure. Transplantation
2010; 90: 68.
7. Gourishankar S, Leduc R, Connett J, et al. Pathological and clinical characterization of the ‘troubled transplant': Data from the DeKAF study. Am J Transplant
2010; 10: 324.
8. Matas AJ, Leduc R, Rush D, et al. Histopathologic clusters differentiate subgroups within the nonspecific diagnoses of CAN or CR: Preliminary data from the DeKAF study. Am J Transplant
2010; 10: 315.
9. Solez K, Colvin RB, Racusen LC, et al. Banff '05 Meeting Report: Differential diagnosis of chronic allograft injury and elimination of chronic allograft nephropathy (‘CAN'). Am J Transplant
2007; 7: 518.
10. Hertig A, Anglicheau D, Verine J, et al. Early epithelial phenotypic changes predict graft fibrosis. J Am Soc Nephrol
2008; 19: 1584.
11. Hertig A, Verine J, Mougenot B, et al. Risk factors for early epithelial to mesenchymal transition in renal grafts. Am J Transplant
2006; 6: 2937.
12. Galichon P, Hertig A. Epithelial to mesenchymal transition as a biomarker
in renal fibrosis: Are we ready for the bedside? Fibrogenesis Tissue Repair
2011; 4: 11.
13. Zeisberg M, Duffield JS. Resolved: EMT produces fibroblasts in the kidney. J Am Soc Nephrol
2010; 21: 1247.
14. Kalluri R, Weinberg RA. The basics of epithelial-mesenchymal transition. J Clin Invest
2009; 119: 1420.
15. Djamali A, Reese S, Yracheta J, et al. Epithelial-to-mesenchymal transition and oxidative stress in chronic allograft nephropathy. Am J Transplant
2005; 5: 500.
16. Djamali A, Vidyasagar A, Adulla M, et al. Nox-2 is a modulator of fibrogenesis in kidney allografts. Am J Transplant
2009; 9: 74.
17. Robertson H, Ali S, McDonnell BJ, et al. Chronic renal allograft dysfunction: The role of T cell-mediated tubular epithelial to mesenchymal cell transition. J Am Soc Nephrol
2004; 15: 390.
18. Vongwiwatana A, Tasanarong A, Rayner DC, et al. Epithelial to mesenchymal transition during late deterioration of human kidney transplants: The role of tubular cells in fibrogenesis. Am J Transplant
2005; 5: 1367.
19. de Matos AC, Câmara NO, Tonato EJ, et al. Vimentin expression and myofibroblast infiltration are early markers of renal dysfunction in kidney transplantation: An early stage of chronic allograft dysfunction
? Transplant Proc
2010; 42: 3482.
20. Slattery C, Campbell E, McMorrow T, et al. Cyclosporine A-induced renal fibrosis: A role for epithelial-mesenchymal transition. Am J Pathol
2005; 167: 395.
21. Pallet N, Bouvier N, Bendjallabah A, et al. Cyclosporine-induced endoplasmic reticulum stress triggers tubular phenotypic changes and death. Am J Transplant
2008; 8: 2283.
22. Conger JD, Kim GE, Robinette JB. Effects of ANG II, ETA, and TxA2 receptor antagonists on cyclosporin A renal vasoconstriction. Am J Physiol
1994; 267(3 pt 2): F443.
23. Kopp JB, Klotman PE. Cellular and molecular mechanisms of cyclosporin nephrotoxicity. J Am Soc Nephrol
1990; 1: 162.
24. Elzinga LW, Rosen S, Bennett WM. Dissociation of glomerular filtration rate from tubulointerstitial fibrosis in experimental chronic cyclosporine nephropathy: Role of sodium intake. J Am Soc Nephrol
1993; 4: 214.
25. Hazzan M, Hertig A, Buob D, et al. Epithelial-to-mesenchymal transition predicts cyclosporine nephrotoxicity
in renal transplant recipients. J Am Soc Nephrol
2011; 22: 1375.
26. Jevnikar AM, Mannon RB. Late kidney allograft loss: What we know about it, and what we can do about it. Clin J Am Soc Nephrol
2008; 3(suppl 2): S56.
27. Opelz G, Zeier M, Laux G, et al. No improvement of patient or graft survival in transplant recipients treated with angiotensin-converting enzyme inhibitors or angiotensin II type 1 receptor blockers: A collaborative transplant study report. J Am Soc Nephrol
2006; 17: 3257.
28. Lanese DM, Conger JD. Effects of endothelin receptor antagonist on cyclosporine-induced vasoconstriction in isolated rat renal arterioles. J Clin Invest
1993; 91: 2144.
29. Nakahama H. Stimulatory effect of cyclosporine A on endothelin secretion by a cultured renal epithelial cell line, LLC-PK1 cells. Eur J Pharmacol
1990; 180: 191.
30. Bagnato A, Rosanò L. Epithelial-mesenchymal transition in ovarian cancer progression: A crucial role for the endothelin axis. Cells Tissues Organs
2007; 185: 85.
31. Shi B, Cai M, Qian Y, et al. Additive immunosuppressive effect of hydroxycamptothecin and cyclosporine on rejection of heart transplantation in rats. Transplant Proc
2007; 39: 3429.
32. Lijkwan MA, Cooke DT, Martens JM, et al. Cyclosporine treatment of high dose and long duration reduces the severity of graft coronary artery disease in rodent cardiac allografts. J Heart Lung Transplant
2005; 24: 439.
33. Geerling RA, de Bruin RW, Scheringa M, et al. Suppression of acute rejection prevents graft arteriosclerosis after allogeneic aorta transplantation in the rat. Transplantation
1994; 58: 1258.
Cyclosporine nephrotoxicity; Biomarker; Chronic allograft dysfunction
Supplemental Digital Content
© 2011 Lippincott Williams & Wilkins, Inc.