Autophagy in kidney transplants of sirolimus treated recipients

Background Mammalian target of rapamycin (mTOR) inhibitors are increasingly used as immunosuppressive agents in kidney transplantation. In the experimental setting it has been shown that mTOR inhibitors promote autophagy, but the concept that this might also occur in transplant patients has not been addressed. Objectives This study was designed to investigate the association between mTOR inhibition and autophagy in renal transplants under routine clinical conditions. Materials and Methods Protocol transplant biopsies of patients receiving sirolimus were compared to biopsies of patients treated without mTOR inhibitor. Electron microscopy was used for quantitative stereological analysis of autophagosomal volume fractions. Ultrastructural analysis was focused on podocytes to avoid cell type bias. Autophagy-related gene products were profiled by QPCR from laser assisted microdissected glomeruli and by immunohistochemistry for semiquantitative evaluation. Results By electron microscopy, we observed a significant > 50% increase in podocytic autophagosomal volume fractions in patients treated with sirolimus. Evaluation of biopsy material from the same patients using transcriptional profiling of laser capture microdissected glomeruli revealed no differences in autophagy-related gene expressions. Immunohistochemical evaluation of autophagic degradation product p62 was also unaltered whereas a significant increase was observed in podocytic LC3 positivity in biopsies of sirolimus treated patients. Conclusions These results indicate an association of sirolimus treatment and autophagosome formation in transplant patients. However, they might reflect autophagosomal buildup rather than increased autophagic flux. Further research is needed to investigate the potential functional consequences in short- and long-term outcome of patients treated with mTOR inhibitors.


Background
Autophagy is an evolutionary conserved mechanism that sequesters proteins and damaged organelles within double-membrane vesicles termed autophagosomes, which fuse with lysosomes to degrade their cargo. Thereby autophagy counteracts accumulation of damaged organelles and provides energy (1). Defects in autophagy have been linked to accelerated aging, reduced metabolic fitness, cancer, cardiovascular disease and several other human pathologies (2). Autophagy has been found to protect the kidney from acute tubular injury (3) and it was concluded that pharmaceutical autophagic activation may be beneficial for different diseases (4). One of the key regulators of autophagy is mammalian target of rapamycin (mTOR). mTOR is a serinethreonine kinase, which interacts with regulatory proteins to form the mTOR complex 1 (mTORC1). mTORC1 functions as a nutrient sensor, which when activated stimulates protein synthesis and cell growth and blocks autophagy. In the experimental setting, pharmaceutical inhibition of mTOR by sirolimus (rapamycin) or its derivative, everolimus, prevents the formation of mTORC1 and thus triggers the initiation of autophagic processes (1). mTOR inhibitors are used as immunosuppressive agents in transplantation medicine as they also block cytokine receptordependent signal transduction and thereby inhibit the activation of T lymphocytes (5).

Objectives
Despite the growing use of mTOR inhibitors there is no data about their potential impact on autophagy in the setting of transplantation. Based on experimental studies, it has been assumed that autophagy might be induced in sirolimus treated patients (6), but this has never been formally tested. It was the goal of this proof-of-principle study to fill this gap by examining renal transplant biopsies using different methodological approaches. All patients transplanted between 2002 and 2007 were included if they fulfilled the following criteria: transplant protocol biopsy taken at 3 months after transplantation; samples for transmission electron microscopy (TEM) available; routine pathological assessment of the biopsy showed no signs of rejection or acute tubular necrosis; immunosuppressive treatment included sirolimus at the time of biopsy. Protocol biopsies were performed after obtaining a written consent according to our center's policy, approved by the Hannover Medical School ethics committee. Urine samples were collected at the time of biopsies and patients were excluded if they showed signs of urinary tract infection. Fifteen patients who received sirolimus in combination with calcineurin inhibitors (CNIs) (cyclosporine n = 7; tacrolimus n = 5) and mycophenolate mofetil (MMF) or only with MMF (n = 3) fulfilled the inclusion criteria. As our analyses revealed no differences between sirolimus treatment alone or sirolimus treatment in combination with CNI, data from all 15 sirolimus treated patients were pooled (sirolimus group). Control patients with matching dates of transplantation were selected by the same inclusion criteria as above, when they were treated with cyclosporine and MMF alone (n = 20). Six patients who were treated with belatacept and MMF were included as an additional control group. All patients received anti-interleukin-2 receptor antibody induction, initial methylprednisolone, and prednisolone in the maintenance therapy. Demographical and clinical data were taken from patients charts and are shown in Table 1. Allograft function was assessed by the estimated glomerular filtration rate (eGFR) according to the Cockroft and Gault formula.

Transmission electron microscopy and stereological analysis of autophagosomes
Small samples of the renal cortex from the biopsies were fixed in glutaraldehyde and embedded in Araldit CY212 (PLANO, Wetzlar, Germany) and analyzed with a TEM Morgagni 268 (FEI Company, Eindhoven, NL). Glomeruli from each biopsy were scanned by a blinded investigator who selected podocytes by ultrastructural criteria. Sixty to 80 digital images (1024 × 1024 pixels) per sample were captured in a zig-zag fashion at 22 000X magnification throughout glomeruli with a constant distance between every two images by adjusting the co-ordinates according to principles of systematic uniform random sampling. Autophagosomes were identified using previously established criteria (7). Briefly, vacuoles were classified as autophagosomes when they met with two or more

Immunohistochemistry
Four micrometer sections of paraffin-embedded biopsies were de-paraffinized and boiled in a citrate buffer for p62 staining and Tris-EDTA buffer for LC3 staining for 20 minutes at 100°C for antigen-retrieval,

Statistical analysis
Results are expressed as means ± SD. Statistical significance was determined by Mann-Whitney U test (GraphPad Prism software, San Diego, CA, USA). P < 0.05 was considered to be statistically significant.

Quantification of autophagosomes by transmission electron microscopy
By virtue of its high resolution, TEM can be used to visualize autophagosomes directly in tissue samples (7). Transplant biopsies were compared using TEM between sirolimus and cyclosporine treated patients. Clinical characteristics are summarized in Table 1. In order to reach optimal intersample comparability we specifically restricted our analysis on podocytes; a homogenous cell type that can be unambiguously identified by TEM and is known for its strong autophagic involvement (10)(11)(12). Pilot studies revealed that the analysis of other renal cell types, e.g. tubular cells, introduced a strong selection bias as they comprise different subtypes which possess individual rates of baseline autophagy (12). Autophagosomes were found in podocytes of all patients ( Figure 1B-D) but morphometric stereological quantification indicated significantly higher autophagosomal volume fraction in the sirolimus treated group ( Figure 1E). Since cyclosporine has also been linked to autophagy induction in some experimental and clinical studies (13), podocytic autophagosomes were also quantified in transplant biopsies from a small cohort of belatacept treated recipients. The mean podocytic autophagosomal volume fraction in biopsies of these patients, who were CNI free was similar to the cyclosporine treatment group and significantly lower than in sirolimus treated patients (Figure 1 E). These findings indicate that sirolimus treatment increases the load of autophagosomes in podocytes.

Assessment of autophagy by expression analysis and immunohistochemistry
While the rapid activation of autophagy is thought to occur mainly via changes in kinase activity, the chronic administration of mTOR inhibitors can induce transcriptional changes of autophagy related genes (14,15). In order to investigate whether podocytic changes seen with TEM were paralleled by changes in gene expression we performed laser capture microdissection of glomeruli (Figure 2A,B), thereby enriching podocytic transcripts for expression analysis by quantitative RT-PCR (16). Expression levels of all tested genes (Atg5, Beclin1 and Gabarapl1) showed no www.nephropathol.com 94 discernible differences between the groups ( Figure   2C) indicating that changes seen in TEM were not paralleled by increased transcription of autophagy pathway genes. Biopsy paraffin sections from the same patients were used for immunohistological analysis of p62. p62 is a typical autophagy substrate which accumulates when autophagic activity is low (1). p62 immunoreactivity was evaluated using a semiquantitative scoring system for podocyte staining intensity. There was a trend for less p62 in sirolimus treated patients but no statistically significant differences were detected ( Figure 2D-F). Another important hallmark of autophagy is the formation of cellular autophagosomal punctae containing light chain 3 (LC3). Assessment of immunhistochemical LC3 staining revealed a significantly increased area fraction of granular LC3 positivity in glomeruli of sirolimus treated patients as compared to cyclosporine controls ( Figure 2G-I).

Discussion
A link between mTOR inhibition and autophagy has so far only been shown in experimental studies. One important reason for the lack of clinical studies might be the limited applicability of autophagy detection methods for human tissues (17,18). In our study we analyzed the impact of mTOR inhibition on autophagy in the clinical setting of renal transplantation. By stereological TEM analysis we found that sirolimus treatment was associated with a significantly elevated volume fraction of podocytic autophagosomes in routinely performed kidney transplant protocol biopsies. TEM is the gold standard to detect autophagosomes and it is a powerful tool for studying autophagy in mammalian tissues (1). However, TEM is not always available and TEM-based stereological analysis is very labor-intensive and requires welltrained personnel. Thus the use of TEM is limited for larger cohort studies and for clinical routine application. Elaboration of alternative methods for monitoring autophagy in routine paraffin-embedded tissues is still evolving (17). We used laser-dissected glomeruli to analyze transcription of autophagy proteins Atg5, Beclin1 and Gabarapl1 but found no differences between the different treatment groups. This lack in transcriptional changes is consistent with previous biopsy microarray studies in mTOR inhibitor treated patients (19,20). It is important to note, however, that unaltered expression levels can not exclude sirolimus-associated autophagic activation since autophagic activity can be regulated without transcriptional changes. Another important method to monitor autophagy is the conversion of LC3-I to LC3-II which is a crucial part of the initiation process of autophagy and leads to the association of LC3 with the autophagosomal membrane (21). Immunoblot is the gold standard to quantify the conversion of LC3 under experimental conditions and fluorescent microscopy can be used to detect autophagosomal LC3 in vivo if LC3 has been genetically tagged with markers like green fluorescent protein (10). As these methods can not be used for patient biopsy material it has been tried to evaluate LC3 indirectly by immunohistochemistry which yielded heterogeneous results (18,22). Using sensitive immunhistochemical staining we observed an increased area fraction of granular LC3 immunopositivity in glomeruli of sirolimus treated patients. This is in agreement with a recent report in which a novel LC3 reporter mouse revealed that systemic administration of sirolimus leads to significantly increased LC3 activation (12). Several limitations of our study should be acknowledged. First, as an observational study, only associations, but no cause-and-effect relationships, can be established. This problem might have been aggravated by the small cohort size of our study. As kidney transplant recipients represent a very heterogeneous group with a high load of patientindividual covariates renal transplant studies usually require larger patient numbers (23). However, labour intensity of stereometric TEM analysis prevented a better powered cohort size and the option to perform functional correlation analysis with clinical data. Despite this shortcoming, our EM results unequivocally show that sirolimus treatment is significantly associated with increased podocyte autophagosomal volume fractions. It is important to note that our data can not provide an unequivocal answer whether this increase represents enhanced autophagic flux. It could also result from blockade of downstream steps in the autophagic pathway. Normally, lysosomes are restored from autophagosomes in a process that has been described as autophagic lysosomal reformation (24). Failure of this process can disrupt the autophagic equilibrium leading to a cellular build-up of autophagolysosomes which cannot be distinguished from earlier autophagosomes by TEM (25). Therefore, our results might reflect altered autophagosomal cycling rather than enhanced autophagic flux which could explain unaltered gene expression of Atg5, Beclin1 and Gabarapl1 and unchanged immunostaining for p62 which is an indicator of autophagic degradation. In that case, the intensified LC3 immunostaining might correspond to an autophagosomal buildup without reflecting changes in the net autophagic process of substrate degradation.

Conclusions
In summary, our TEM data focusing on podocytes in kidney transplant biopsies of patients who received sirolimus as part of their routine care show an increased relative volume density of autophagosomal vesicles. This indicates for the first time an association of sirolimus treatment and autophagosomal processing in the clinical setting of transplantation. Parallel biochemical findings suggest that mTOR inhibition might induce autophagosomal buildup rather than increased autophagic flux. Further research is needed to investigate potential functional consequences on short-and long-term outcome in patients treated with mTOR inhibitors.

Limitations of the study
This is a retrospective single-center study with a limited number of patients. Biochemical monitoring of autophagic flux was not possible since archived biopsy samples were used.