• Open Access

Markers of cellular senescence in zero hour biopsies predict outcome in renal transplantation

Authors

  • Christian Koppelstaetter,

    1. Division of Nephrology, Department of Internal Medicine, Medical University Innsbruck, Innsbruck, Austria
    Search for more papers by this author
  • Gabriele Schratzberger,

    1. Division of Nephrology, Department of Internal Medicine, Medical University Innsbruck, Innsbruck, Austria
    Search for more papers by this author
  • Paul Perco,

    1. Division of Nephrology and Dialysis, Department of Internal Medicine 3, Medical University Vienna, Vienna, Austria
    2. Department of Internal Medicine III, Krankenhaus der Elisabethinen, Linz, Austria
    Search for more papers by this author
  • Johannes Hofer,

    1. Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria
    Search for more papers by this author
  • Walter Mark,

    1. Division of General and Transplant Surgery, Department of Surgery, Medical University Innsbruck, Innsbruck, Austria
    Search for more papers by this author
  • Robert Öllinger,

    1. Division of General and Transplant Surgery, Department of Surgery, Medical University Innsbruck, Innsbruck, Austria
    Search for more papers by this author
  • Rainer Oberbauer,

    1. Division of Nephrology and Dialysis, Department of Internal Medicine 3, Medical University Vienna, Vienna, Austria
    2. Department of Internal Medicine III, Krankenhaus der Elisabethinen, Linz, Austria
    Search for more papers by this author
  • Christoph Schwarz,

    1. Division of Nephrology and Dialysis, Department of Internal Medicine 3, Medical University Vienna, Vienna, Austria
    2. Department of Internal Medicine III, Krankenhaus der Elisabethinen, Linz, Austria
    Search for more papers by this author
  • Christa Mitterbauer,

    1. Division of Nephrology and Dialysis, Department of Internal Medicine 3, Medical University Vienna, Vienna, Austria
    2. Department of Internal Medicine III, Krankenhaus der Elisabethinen, Linz, Austria
    Search for more papers by this author
  • Alexander Kainz,

    1. Division of Nephrology and Dialysis, Department of Internal Medicine 3, Medical University Vienna, Vienna, Austria
    2. Department of Internal Medicine III, Krankenhaus der Elisabethinen, Linz, Austria
    Search for more papers by this author
  • Henryk Karkoszka,

    1. Department of Nephrology, Endocrinology and Metabolic Diseases, Medical University of Silesia, Katowice, Poland
    Search for more papers by this author
  • Andrzej Wiecek,

    1. Department of Nephrology, Endocrinology and Metabolic Diseases, Medical University of Silesia, Katowice, Poland
    Search for more papers by this author
  • Bernd Mayer,

    1. Emergentec Biodevelopment, Vienna, Austria
    Search for more papers by this author
  • Gert Mayer

    1. Division of Nephrology, Department of Internal Medicine, Medical University Innsbruck, Innsbruck, Austria
    Search for more papers by this author

  • Christian Koppelstaetter and Gabriele Schratzberger contributed equally to this work.

  • We declare that we have no conflict of interest. The funding source of this study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. All authors had full access to all data of the study. This study was approved by the Ethical Committees of all contributing institutions.


Gert Mayer, MD, Division of Nephrology, Department of Internal Medicine, Medical University Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria. Tel.: +43 512 504 25855; fax: +43 512 504 25857; e-mail: gert.mayer@i-med.ac.at

Summary

Although chronological donor age is the most potent predictor of long-term outcome after renal transplantation, it does not incorporate individual differences of the aging-process itself. We therefore hypothesized that an estimate of biological organ age as derived from markers of cellular senescence in zero hour biopsies would be of higher predictive value. Telomere length and mRNA expression levels of the cell cycle inhibitors CDKN2A (p16INK4a) and CDKN1A (p21WAF1) were assessed in pre-implantation biopsies of 54 patients and the association of these and various other clinical parameters with serum creatinine after 1 year was determined. In a linear regression analysis, CDKN2A turned out to be the best single predictor followed by donor age and telomere length. A multiple linear regression analysis revealed that the combination of CDKN2A values and donor age yielded even higher predictive values for serum creatinine 1 year after transplantation. We conclude that the molecular aging marker CDKN2A in combination with chronological donor age predict renal allograft function after 1 year significantly better than chronological donor age alone.

Introduction

The incidence of end-stage renal disease is rising dramatically and, as the prognosis of patients on dialysis is still poor, renal transplantation is the treatment of choice (Wolfe et al., 1999; Stengel et al., 2003; USRDS, 2004). Within the Eurotransplant organization, approximately 12 000 patients are currently registered on the renal waiting list. However, only about 3000 grafts can be allocated each year because of donor shortage, even though a significant expansion of the donor pool during the last decades has been achieved by extending the upper age limit (Mayer & Persijn, 2006). Although transplant outcome and even recipient mortality are negatively affected a significant number of organs from elderly donors maintain excellent function for a prolonged period of time (Lee et al., 1996; Morris et al., 1999; Prommool et al., 2000; Smits et al., 2002; Gjertson, 2004; Oppenheimer et al., 2004; Cohen et al., 2005; Frei et al., 2008). It is thus tempting to speculate that biological organ age rather than chronological donor age determines organ quality and long-term kidney function after transplantation. Over the last decades, research efforts resulted in the availability of several biological age markers. Telomeres, for example, the noncoding regions at the end of the chromosomes, are usually about 10–13 kilobase pairs long at birth and get shortened by each cell division as well as by environmental stress factors and reactive oxygen species (Okuda et al., 2002; Ben-Porath & Weinberg, 2004; Epel et al., 2004). Once a critical telomere length is reached, cells enter a state of stable replication arrest named cellular senescence. In addition, the cell cycle inhibitors CDKN2A (p16INK4a) and CDKN1A (p21WAF1) have also been associated with a senescent phenotype. While CDKN1A activity is related to telomere shortening via the p53 pathway, CDKN2A is linked to a telomere-shortening independent form of stress-induced senescence (Campisi, 1996; Shay & Wright, 2000).

In order to investigate the value of these markers to predict serum creatinine levels 1 year after transplantation, we determined telomere length and the gene expression levels of CDKN2A and CDKN1A in renal biopsies obtained immediately before engraftment and compared their power to other conventionally used parameters, including the ‘gold standard’ chronological donor age, by linear regression analysis.

Results

The characteristics of the study populations are given in Table 1.

Table 1.  Experimental and demographic data of the study population (n = 54)
ParameterMeanSD
Serum creatinine level at 1 year (mg%)1.51 0.52
Telomere length (T/S ratio)2.70 0.77
CDKN2A mRNA (fold expression compared to healthy 33-year-old control)9.30 7.10
CDKN1A mRNA (fold expression compared to healthy 33-year-old control)4.23 3.43
Donor age (years)46.3315.95
Recipient age (years)50.1115.31
Recipient gender (female/male)20/34 
Donor gender (female/male)23/29 
HLA mismatch (total count)2.32 1.62
Panel reactive antibodies (%)2 (0/4) 
Cold ischemia time (hours)13.18 5.44
Acute rejection (no/yes)38/10 
Acute allograft failure (no/yes)37/13 
Corticosteroids at hospital discharge (n)40 
Cyclosporine A at hospital discharge (n)27 
Tacrolimus at hospital discharge (n)25 
Azathioprine at hospital discharge (n) 3 
Mycophenolate mofetil at hospital discharge (n)29 
Rapamycin at hospital discharge (n) 6 
Interstitial inflammation at implantation (grade)1 × 2, 2 × 1, 41 × 0, 10 × NA
Interstitial fibrosis at implantation (grade)1 × 2, 1 × 1.5, 10 × 1, 33 × 0, 9 × NA
Arteriolar hyalinosis at implantation (grade)2 × 2, 17 × 1, 25 × 0, 10 × NA
Tubular atrophy at implantation (grade)1 × 3, 2 × 2.5, 5 × 2, 17 × 1, 19 × 0, 10 × NA
Sclerosed glomeruli (%)1.49 2.55

Associations between donor age, telomere length, CDKN2A and CDKN1A expression

A negative, linear and significant correlation was observed between telomere length and donor age and a weak positive correlation between donor age and CDKN2A expression (Fig. 1).

Figure 1.

Correlation between donor age and (a) relative telomere length (T/S ratio), (b) CDKN2A and (c) CDKN1A. The correlation of donor age with telomere length and CDKN2A was significant. CDKN1A was not correlated with donor age. Telomere length showed a weak negative correlation with CDKN1A (e) but not with CDKN2A (d). The expression of CDKN1A was not correlated with the expression of CDKN2A (f). Gene expression of CDKN2A and CDKN1A is fold expression compared to 33-year-old control. T/S ratio refers to the DNA of a 74-year-old male.

Postoperative allograft function and donor/recipient parameters

Linear regression analysis was performed using serum creatinine at 12 months after transplantation as dependent variable. The expression of CDKN2A (p16) was the single most predictive parameter explaining 31.3% of the variability of 1 year allograft function (Table 2), followed by donor age (17.2%) and telomere length (6.6%). No other significant association was detected although a trend was noted for acute postoperative transplant failure, incidence of acute rejection and severity of arteriolar hyalinosis and tubular atrophy.

Table 2.  Linear regression analysis of the data set using serum creatinine at 12 months as the dependent variable (n = 54)
 Parameter estimate95% CIp-valueAdjusted R2
CDKN2A (p16) mRNA (fold expression compared to 33-year-old control) 0.0417 0.0250; 0.0584< 0.0010.313
Donor age (10 years) 0.141 0.0595; 0.2227  0.00110.172
Relative telomere length (T/S ratio)–0.195–0.3753; – 0.0137  0.03550.066
Acute rejection (no/yes)–0.3670–0.7380; 0.0040  0.05240.059
Arteriolar hyalinosis–0.2490–0.5170; 0.0190  0.06770.056
Tubular atrophy 0.1738–0.0118; 0.3593  0.06560.056
Recipient age (10 years) 0.068–0.0239; 0.1609  0.14260.022
Acute allograft failure (no/yes)–0.2341–0.5733; 0.1051  0.17170.018
Sclerosed glomeruli (%)–0.0422–0.6377; 0.0827  0.19890.017
CDKN1A mRNA (fold expression compared to 33-year-old control)–0.0286–0.0707; 0.0135  0.17870.016
Cold ischemia time (hours)–0.0158–0.0436; 0.0119  0.25860.006
Interstitial inflammation (grade)–0.1648–0.6127; 0.2830  0.4617< 0
HLA mismatch (total count) 0.0304–0.0637; 0.1245  0.5195< 0
Donor gender (female/male) 0.0896–0.2012; 0.3806  0.5386< 0
Panel reactive antibodies (%)–0.0025–0.0122; 0.0073  0.6126< 0
Recipient gender (female/male)–0.0717–0.3674; 0.2238  0.6281< 0
Interstitial fibrosis (grade)–0.0578–0.3648; 0.2493  0.7061< 0

In a multiple linear regression analysis, the combination of CDKN2A expression and donor age explained 37.2% of the variability of 1 year allograft function, which was significantly better than donor age alone (p < 0.001). In contrast, the combination of telomere length and donor age only reached a prognostic value of 15.3%. Combining the biological markers CDKN2A and telomere explained 32.9% of serum creatinine variability after 1 year.

Discussion

Even though advanced chronological donor age is on average the most important single negative predictor of long-term outcome after renal transplantation, a respectable number of organs from elderly donors maintain excellent function for many years (Lee et al., 1996; Morris et al., 1999; Prommool et al., 2000; Smits et al., 2002; Oppenheimer et al., 2004; Cohen et al., 2005; Frei et al., 2008). As the rising demand for donor kidneys can only be met by accepting organs from elderly subjects more frequently, transplant physicians and their patients are searching for superior diagnostics to assess organ quality prior to engraftment. Here we examined whether markers of cellular senescence obtained in zero hour biopsies would provide additional information for predicting the allograft function 1 year after transplantation when compared to other, conventionally used parameters. Serum creatinine values after 12 months were chosen as the dependent variable of interest as they are affected by the same risk factors which have been associated with the incidence of chronic allograft failure (Siddiqi et al., 2004; Dominguez et al., 2005).

Historically, telomere shortening has been considered a hallmark of the classical form of replicative senescence, which was first described by Hayflick & Moorhead (1961). Once an individual chromosome is no longer ‘capped’ by a telomere exceeding a critical length, or the associated multicomponent protein assembly is disturbed, the situation is sensed as a DNA strand break and p53 levels increase, resulting in the transcriptional activation of CDKN1A, a cyclin kinase inhibitor, which mediates a cell cycle arrest in the G1 phase (Ben-Porath & Weinberg, 2004; Herbig et al., 2004). However, it should be mentioned that telomere length is difficult to determine accurately independent of the method used (Southern blot, real-time PCR) for which reason telomere results always have to be assessed somehow critically.

The concept of cellular senescence has been expanded by including cell cycle inhibitors, which are induced by DNA damaging agents, oxidative or ‘oncogenic’ stress, and other metabolic perturbations (Lloyd, 2002). Although telomere shortening has been observed under these circumstances as well, the activation of CDKN2A seems to act as the dominant factor to limit growth (Beausejour et al., 2003).

In our study, CDKN2A expression and, to a lesser extent, telomere length correlated with chronological donor age, which is in accordance with data published by Melk et al. (2000, 2004). Other researchers have described a strong correlation of p16INK4a and ARF with acceleration of aging by exogenic stress in several different tissues, including the kidney, in rodents (Zindy et al., 1997; Krishnamurthy et al., 2004; Edwards et al., 2007).

Melk et al. (2004) suggested that only p16INK4a but not the alternative isoform ARF can be set into relation to accelerated aging in the human kidney. Even though we could prove that also the indifferentiated p16INK4a is of high predictive power for biological age, we state that as a minor limitation of our study we did not determine the expression of p16INK4a and ARF separately.

After testing all parameters, only CDKN2A expression, chronological donor age and telomere length were significantly associated with serum creatinine values 12 months after transplantation. However, the incidence of acute rejection and acute postoperative transplant failure as well as several other parameters determined by light microscopy showed a high tendency to predict allograft function, matching well previous findings from other groups (Tanaka et al., 2003; Arias et al., 2007). CDKN2A has also been described by others to be associated with the progression of glomerular diseases as well as with histological features of chronic kidney injury (Melk et al., 2005; Sis et al., 2007). Also in the neuronal and hematopoetic systems, similar findings suggest a major role of p16INK4a in the lowering of renewal capacity and, furthermore, a decrease in function (Janzen et al., 2006; Molofsky et al., 2006).

The predictive power of the combination of chronological donor age and CDKN2A expression values was significantly better than the one achieved by chronological donor age alone, making this approach an attractive procedure for future prospective studies.

Since cellular senescence has been associated pathophysiologically with chronic allograft failure per se as well as with calcineurin inhibitor therapy, biological age of the organ at the time of transplantation might play a crucial role (Ferlicot et al., 2003; Joosten et al., 2003; Jennings et al., 2007).

In summary, our study suggests that combining the expression of the cellular marker CDKN2A in zero hour biopsies with chronological donor age provides significantly more information about organ quality and long-term renal allograft outcome than donor age alone. Senescence-associated markers in the near future will be of enormous value in organ quality assessment in transplantation and clinical medicine in general.

Experimental procedures

All patients included received a kidney transplant from deceased donor. From 54 allografts transplanted at the Medical University Hospitals of Innsbruck, Vienna and Katowice, a pre-implantation biopsy core, which was obtained as part of the routine renal transplant procedure, was bisected and either stored in 4% formaldehyde solution for subsequent histological evaluation or in RNAlater® (Ambion, Austin, TX, USA). The formaldehyde fixed, paraffin embedded biopsies were processed according to standard procedures and 4 serial 2-µm sections were stained with hematoxylin and eosin, periodic acid–Schiff, Jones's methenamin silver and a trichrome stain (acidic fuchsin orange-G). The percentage of sclerotic glomeruli was calculated and the extent of arteriolar hyalinosis, interstitial inflammation, interstitial fibrosis and tubular atrophy was determined using a semiquantitative 4 grade scoring system (0 = absent, 1 = mild, 2 = moderate, 3 = severe) (Racusen et al., 1999). The RNAlater-stored sample was used for determining telomere length and gene expression.

In addition to the molecular and histological markers, the following data were obtained for each patient (see Table 1): serum creatinine 1 year after transplantation, donor/recipient age and gender, percentage of preformed panel reactive antibodies at transplantation, cold ischemia time and number of HLA A–, B– and Dr– mismatches. Postoperative acute allograft failure was defined as the need for dialysis within the first week after transplantation. The incidence of a biopsy proven acute rejection as well as the type of immunosuppressive medication during the follow-up time was also recorded.

Isolation and quantification of DNA and RNA

Genomic DNA and total RNA were extracted using commercially available kits (DNeasy®Tissue Kit, Qiagen, Hilden, Germany; TRI Reagent®, Molecular Research Center, Cincinnati, OH, USA). Prior to quantitative PCR experiments, the concentration of genomic DNA obtained from each sample was adjusted to 2.5 ng µL−1 by using a fluorescence-based kit (PicoGreen® dsDNAQuantitation Kit, Molecular Probes, Eugene, OR, USA). Total RNA was determined by absorbance at 260 nm and equated to a concentration of 0.1 µg µL−1. Total RNA extracted from the healthy part of a kidney, which had to be removed from a 33-year-old male patient because of renal cell cancer, was used as a standard for gene expression experiments. For telomere experiments, renal cortical tissue from 74-year-old male was applied.

Determination of telomere length by real-time PCR

Telomere length values were measured by quantitative PCR determining the relative ratio of telomere repeat copy number to single-copy gene copy number (T/S ratio) in experimental samples as compared with a reference DNA sample as described in detail by Cawthon (2002). In brief, separate duplicate PCR experiments were performed for telomere (T) and 36B4 (a single copy gene, S). The primer pair sequences for telomere PCR were (5’ to 3’): tel1b CGGTTTGTTTGGGTTTGGGTTTGGGTTTGGGTTTGGGTT and tel2b GGCTTGCCTTACCCTTACCCTTACCCTTACCCTTACCCT. Sequences for 36B4: 36B4 forward CAGCAAGTGGGAAGGTGTAATCC; 36B4 reverse CCCATTCTATCATCAACGGGTACAA. Telomere length was also measured by Southern blot in seven samples (TeloTAGGG Telomere Length Assay, Roche Diagnostics, Mannheim, Germany). The results correlated highly with the ones obtained by real-time PCR (R2= 0.89).

Determination of CDKN2A and CDKN1A mRNA expression by real-time PCR

Samples were transcribed into cDNA and determination of gene expression levels of CDKN2A (Assay ID: Hs00233365_m1) and CDKN1A (Assay ID: Hs99999142_m1) were performed utilizing a commercially available, predesigned primer set and TaqMan Mix (Applied Biosystems, Foster City, CA, USA). Thermal cycling profiles started with 2 min at 50 °C (RNAse inhibitors activation) and 10 min at 95 °C to activate the polymerase. Repeated cycles were performed 40 times at 95 °C for 15 s, followed by 60 °C for 1 min. Samples were run in duplicates for CDKN2A and CDKN1A and relative ratios to 18 s RNA were determined following the ddCt method. Except for 18 s RNA (Applied Biosystems) all other tested housekeeping genes (GAPDH, β-Actin and β-2-microglobulin) showed signs of coregulation and were therefore not considered as appropriate for conclusive experiments.

All experiments were carried out using the ABI PRISM 7500 and evaluated by the SDS 1.9.1 software package and GraphPad Prism®.

Statistical analysis

The demographic data given in Table 1 are presented as mean ± SD or median and interquartile range (IQR). One telomere value exceeding a T/S ratio of 6 was excluded from the data set in the multivariate analysis and the univariate telomere analysis. Pairwise scatterplots were constructed for donor age, telomere length, and expression values of CDKN2A and CDKN1A (Fig. 1). The Spearman correlation coefficient was used as quantitative measure to describe correlations between investigated variables.

Linear regression analysis was performed to determine the associations of the independent variables in the study population with serum creatinine values 12 months post-transplant (Table 2).

As we were interested in markers that provide additional information to donor age for the prediction of creatinine values 1 year post-transplant we performed a sample size calculation. For a multiple linear regression model that already includes one covariate (donor age) with an adjusted R2 of 0.25, a sample size of 51 will provide a power > 80% to detect an increase in R2 of 0.1 (α= 0.05).

The combination of donor age and either telomere length or the expression of CDKN2A was tested in multiple linear regression models. In addition, the predictive power of the model holding the two biological markers – namely, the telomere length and CDKN2A values – was calculated (Table 3).

Table 3.  Multiple linear regression analyses using donor age, telomere length, and the expression of CDKN2A
 Creatinine at 12 months (n = 53)
Parameter estimate95% CIp-valueAdjusted R2
Donor age (10 years) 0.095 0.021, 0.169  0.0134 
CDKN2A (p16) mRNA (fold expression compared to 33-year-old control) 0.036 0.019, 0.052< 0.001 
Total  < 0.0010.373
Donor age (10 years) 0.123 0.024, 0.2122  0.0155 
Relative telomere length (T/S ratio)–0.059–0.263, 0.144  0.5627 
Total    0.00580.153
Relative telomere length (T/S ratio)–0.131–0.287, 0.024  0.0967 
CDKN2A (p16) mRNA (fold expression compared to 33-year-old control) 0.039 0.022, 0.056< 0.001 
Total  < 0.0010.329

Contributors

G. Mayer, C. Koppelstaetter and P. Perco were responsible for the conception and design. G. Schratzberger and B. Mayer extended the study design. C. Koppelstaetter did the experiments. P. Perco, R. Oberbauer and B. Mayer did the statistical analysis and interpretation of data. W. Mark and R. Öllinger conducted the biopsy collection in Innsbruck, and Andrzej Wiecek and Henryk Karkoszka in Katowice. R. Oberbauer, C. Schwarz and C. Mitterbauer were responsible for data and biopsy collection in Vienna and primary analysis. J. Hofer did the histological evaluation. G. Mayer, C. Koppelstaetter, P. Perco and B. Mayer contributed to data interpretation and wrote the report. All authors gave their final approval of the version to be published.

Acknowledgments

We would like to thank Prim. Dr Reinhard Kramar from the Austrian Dialysis and Transplantation Registry and all colleagues contributing to the registry for their help in providing data on the clinical follow-up. We also thank Dr Cawthon from the University of Utah for sending us updates to his protocol for quantitative PCR-based determination of telomere length.

This study was in part supported by a research grant from the Genzyme Renal Innovations Programme.

Ancillary