Poly[ADP-Ribose] Polymerase-1 Expression Is Related To Cold Ischemia, Acute Tubular Necrosis, and Delayed Renal Function In Kidney Transplantation

Cold ischemia time especially impacts on outcomes of expanded-criteria donor (ECD) transplantation. Ischemia-reperfusion (IR) injury produces excessive poly[ADP-Ribose] Polymerase-1 (PARP-1) activation. The present study explored the hypothesis that increased tubular expression of PARP-1 contributes to delayed renal function in suboptimal ECD kidney allografts and in non-ECD allografts that develop posttransplant acute tubular necrosis (ATN). Materials and Methods Nuclear PARP-1 immunohistochemical expression was studied in 326 paraffin-embedded renal allograft biopsies (193 with different degrees of ATN and 133 controls) and in murine Parp-1 knockout model of IR injury. Results PARP-1 expression showed a significant relationship with cold ischemia time (r coefficient = 0.603), time to effective diuresis (r = 0.770), serum creatinine levels at biopsy (r = 0.649), and degree of ATN (r = 0.810) (p = 0.001, Pearson test). In the murine IR model, western blot showed an increase in PARP-1 that was blocked by Parp-1 inhibitor. Immunohistochemical study of PARP-1 in kidney allograft biopsies would allow early detection of possible delayed renal function, and the administration of PARP-1 inhibitors may offer a therapeutic option to reduce damage from IR in donor kidneys by preventing or minimizing ATN. In summary, these results suggest a pivotal role for PARP-1 in the ATN of renal transplantation. We propose the immunohistochemical assessment of PARP-1 in kidney allograft biopsies for early detection of a possible delayed renal function.


Introduction
Renal ischemia produced during transplantation or otherwise is a major cause of acute kidney injury, initiating a complex and interrelated sequence of events that result in the injury and eventual death of renal cells [1,2]. Prolonged cold ischemia contributes to organ damage and increases patient morbidity and mortality. Salahudeen et al. recently studied 6,465 kidney transplant patients using United Network for Organ Sharing (UNOS) data and concluded that prolonged cold ischemia is a strong risk factor for delayed graft function and a significant predictor of short-term [3] and long-term [4] graft loss, as reported by other authors [5,6]. Nevertheless, the pathogenic mechanism has yet to be fully elucidated.
Johnston et al. concluded that cold ischemia time has a major impact on the outcome of transplants from aged and expandedcriteria donors (ECDs). In non-ECD kidneys, very prolonged cold ischemia time is associated with an increase in primary non-function. ECD kidneys from older donors show a greater increase in delayed graft function with longer cold ischemia time. Thus, ECD grafts with cold ischemia time of .8 h have higher delayed graft function rates than do non-ECD grafts with cold ischemia time of .37 h [7].
Early renal transplant dysfunction is mainly caused by ischemic damage (acute tubular necrosis [ATN]), rejection, infection, or cyclosporin A toxicity [8]. The prognosis is complicated by the fact that reperfusion, although essential for the survival of ischemic renal tissue, causes additional damage (reperfusion injury) [9,10] that contributes to the renal dysfunction and injury associated with ischemia/reperfusion (IR) of the kidney [1,10].
Poly[ADP-Ribose] Polymerase-1 (PARP-1) (E.C. 2.4.2.30) is a nuclear zinc-finger DNA-binding protein with a molecular weight of 113 kDa that specifically detects DNA-strand breaks or nicks produced by different genotoxic agents in mammalian cells [11]. PARP-1 catalyzes the ADP ribosylation of proteins using NAD(+) as substrate [12]. PARP activation is a consequence of ischemic injury and results in a depletion of intracellular NAD(+) [13], which can only be replenished via a reaction that consumes ATP. DNA damage produced by IR injury requires cells to consume large amounts of ATP to support poly(ADP-ribosyl)ation. For this reason, whereas moderate PARP activity protects cellular genome integrity, excessive PARP activation can lead to cell death from ATP depletion [14][15][16]. Our group previously demonstrated that PARP1 expression in tubules of aged donors correlates with functional reserve parameters (serum creatinine and time required to achieve effective diuresis) [17].
The present study was designed to produce clinicopathological evidence to test the hypothesis that increased tubular expression of PARP-1 in human allograft kidneys that are suboptimal or develop ATN posttransplant might be one of the predictive factors for a subsequent delay in renal function.

Materials and Methods
We studied 326 paraffin-embedded renal allograft biopsies distributed in four groups: two ECD groups, one with and one without the presence of ATN; and two non-ECD groups, one with and one without the presence of ATN. ECDs were selected as specified in UNOS policies and procedures as either $60 yrs of age or between 50 and 59 yrs of age with at least two of the following three risk factors: a history of hypertension, a history of cerebrovascular disease, and serum creatinine at any time $1.5 mg/dL) [7].
For the ECDs, 220 kidney wedge biopsies were fixed in Glyofix (Pacisa-Giralt, Barcelona, Spain) and embedded in paraffin by microwave-accelerated technique to determine renal lesions at 0-h pre-transplant. Out of these 220 ECD kidneys, 95 biopsies with some degree of ATN and 65 without ATN (serving as control group) were selected for the study. The remaining 60 kidneys were not transplanted due to severe vascular or glomerulo-interstitial renal lesions, and 20 of these were maintained as whole perfused kidneys at 4uC for 48 h in Wisconsin preservation solution. For the non-ECDs, 98 kidney cylinder biopsies were taken between days 5 and 11 post-transplant from oligoanuric recipients, fixed in 10% buffered formalin, and embedded in paraffin by standard procedure to identify ATN and classify its degree as: mild (1) [,10% of tubules with necrotic cells], moderate (2) [10 to 49%], or severe (3) [$50%]. A further 68 kidney sections (transplant protocol biopsies) from non-ECDs with stable renal function and without morphological evidence of ATN served as a control group. The study was conducted according to the Helsinki declaration and approved by the Ethics Committee of the hospital. All biopsies were taken after written informed consent was obtained and subjects were included in this investigation after agreeing to participate and signing the appropriate consent form.
Data were gathered on renal function parameters (serum creatinine [mg/dL]), creatinine clearance [mL/min./1.73 m 2 ]), donor and recipient age and sex, cold-ischemia time, reanastomosis time, time to effective diuresis (defining effective diuresis in terms of Cockcroft-Gault calculated creatinine clearance rather than need for posttransplantation dialysis), immunosuppression regimens, and number of hemodialysis sessions.

Ischemia-reperfusion (IR) mouse model
We used 20 male Parp1 +/+ wild-type and 20 male Parp1 2/2 knockout C57BL/6 mice (24 wks old and 20-30 g). Knockout mice were obtained according to a previously reported procedure [18]. The mice were kept under stable conditions at the Institute of Parasitology and Biomedicine in Granada with ad libitum access to food and water. All experiments were performed according to European Union and Spanish Government guidelines for the ethical care of animals (EU Directive 86/609, RD 223/1988).
Mice were anesthetized by intraperitoneal inoculation of equitensin (2 IU/20 g) and maintained at 37uC on a thermal plate. The left kidney was accessed by anterolateral abdominal horizontal incision of 1.5 cm, and the vascular pedicle was clamped with a model 2A S&T metallic clip (S&T Microlab AG, Rheinfall, Switzerland), maintaining the kidneys within the abdominal cavity under University of Wisconsin solution flow at 4uC. After 45 min of clamping, the clip was removed and the peritoneum and skin were sutured. After 6 and 48 h of reperfusion, the animals were killed with an overdose of sodium pentothal. There were no deaths during postoperative or reperfusion periods.

Administration of PARP inhibitor to mice
PARP-1 inhibitor 3-aminobenzamide (3-ABA) was purchased from Sigma Chemicals (St Louis, MO) and dissolved in saline at a concentration of 5 mg/mL 3-ABA (10 mg/kg) was administered intraperitoneally at 1 h before ischemic injury. Vehicle-treated mice received the saline injection without 3-ABA. In preliminary control experiments, it was determined that administration of 3-ABA to sham-operated mice had no morphological effect.

Renal samples and processing
Two groups of kidneys from C57BL/6 Parp1 +/+ mice and two groups from C57BL/6 Parp1 2/2 mice were formed, divided between kidneys with 6 h or 48 h of reperfusion (n = 10 for each time and mouse type) and kidneys with 6 h of reperfusion plus PARP-1 inhibitor 3-ABA (n = 10 for each group); two groups of control kidneys for each reperfusion time were also studied (n = 20 for each group). In all animals, the left kidney was subjected to ischemia-reperfusion by clamping the complete renal vascular pedicle, using the right kidney as control. After extraction, each kidney sample was divided longitudinally into two halves. One half, with separated cortex and medulla, was rapidly frozen in isopentane at 250uC and submerged in liquid nitrogen for 10 s to develop western-blotting. The other half was immediately fixed in 10% buffered formalin for 24 h and then paraffin-embedded for morphological study using hematoxylin-eosin and PAS staining.

Western blot analysis
Tissues extracted from the human and mouse kidney cortex samples were washed with PBS and resuspended in 100 ml lysis buffer (50 mM Tris-HCl pH 8, 0.1 mM EDTA, 0.5% Triton X-100, 12.5 mM b-mercaptoethanol) for 30 min on ice. Pellet was eliminated and sample buffer (50 mM Tris-HCl pH 6.8, 6 M urea, 6% b-mercaptoethanol, 3% SDS, 0.003% bromophenol blue) was added to the supernatant. Proteins were resolved on SDS-12% polyacrylamide gels and transferred onto Immun-Blot PVDF Membrane (Bio-Rad, Laboratories Irvine, CA, USA). The blot was blocked with 5% milk powder in PBS with 0.

Statistical analysis
The Kolmogorov-Smirnov test was used to assess the normality of the variables. A descriptive analysis was performed, and the Student's t-test, one way ANOVA with post hoc Bonferroni test, chi square test, and Pearson's correlation were applied to determine statistical significances. We constructed ROC curves for PARP-1 and calculated the area under the curve. The statistical analysis was performed using the SPSS-Windows 15.0 program (SPSS Inc, Chicago, IL, USA). The confidence interval was 95% (p,0.05).

Results
Overall PARP-1 expression in the 326 paraffin-embedded renal allograft biopsies studied was scored as follows: 0 (38.9%); 1 (19.2%); 2 (18.9%); and 3 (23.0%). A score of 0 was only detected in renal  PARP-1 appears to have played an important role in early kidney allogaft function (Table 2), with a statistically significant relationship between its expression and cold ischemia time (r coefficient = 0.603, p = 0.0001, Pearson test), time to effective diuresis (r = 0.770, p = 0.0001, Pearson test), and serum creatinine levels at time of biopsy (r = 0.649) and at three months (r = 0.403, p = 0.0001, Pearson test) but not at six months or one year. Likewise, the degree of ATN showed a significant correlation with the same parameters (cold ischemia time r = 0.456; time to effective diuresis r = 0.696; and creatinine levels at time of biopsy r = 0.520, one month r = 0.455, and six months r = 0.508, p = 0.0001, Pearson test). There was an even greater difference in PARP-1 expression intensity between kidneys from ATN patients who did not reduce serum creatinine levels to below 1.7 mg/dL after transplantation and those from ATN patients who did (creatinine ,1.7, PARP-1 1.7960.62 vs. creatinine .1.7, PARP-1 2.3361.04, p = 0.0001, Student's t-test). Figure 2 and Table 3 represent and summarize the statistical results of the ROC curve analyses, showing the high values of the area under curve for the variables ATN, cold ischemia time, time to effective diuresis, and serum creatinine level at biopsy.
In all 20 kidneys ruled out for transplantation but preserved as whole perfused kidneys, immunohistochemistry study revealed a marked increase in PARP-1 expression between the biopsy at 0 h and the renal cortex after 48 h of cold ischemia in Wisconsin preserved solution ( Figure 3A), and western blot study showed a mild activation of PARP-1 after the 48 h of cold ischemia ( Figure 3B).
In a Parp1 knockout mouse IR model, we used western-blot to demonstrate induction of PARP-1 expression in kidney cortex of  C57BL/6 Parp1 +/+ , which was more evident at 48 h ( Figure 4A). After 6 h of reperfusion, 3-ABA-treated Parp1 +/+ mice showed decreased PARP-1 expression and reduced polymer activity ( Figure 4B), and no PARP-1 protein was detected by westernblot in Parp1 2/2 knockout mice ( Figure 4A); renal injury appeared to be reduced in these two groups (data not shown).

Discussion
Increased PARP-1 expression was observed in nuclei of human renal tubular cells after a variable period of cold ischemia and in the same nuclei of patients who developed ATN. PARP-1 expression was inversely correlated with the time to recovery of renal function. PARP-1 induction in tubular cells was previously found in different experimental models of ischemic renal injury [19][20][21][22], and our group recently demonstrated PARP-1 expression in tubules of aged donors [17]. However, the present study of a large series of kidney transplant patients supports aclose relationship between delayed renal function and tubular nuclear PARP-1 expression. According to our results, PARP-1 protein upregulation may be a possible and previously unconsidered pathway for delayed graft function (DGF) [23,24].
A prolonged cold ischemia time is a strong risk factor for DGF, graft loss [4,6], and long-term changes after kidney transplantation [25]. Donor kidneys inevitably undergo a period of cold ischemia. In our series, the periods of cold ischemia ranged from one to thirty-six hours, with significantly longer times for non-ECDs biopsied for suspicion of ATN (see Table 1). A gradual increase in DGF was significantly correlated with longer cold ischemia time (r coefficient: 0.666, p = 0.01, Pearson test). DGF varied according to cold ischemia time and donor age. It has been reported that ECD transplant performed with a short cold ischemia time (0-8 h) has a DGF rate equivalent to a non-ECD transplant performed with a 20-h cold ischemia time [7].
The present study revealed a significant difference in the intensity of PARP-1 immunohistochemical expression in renal tubular cells between kidneys from transplant protocol biopsies with stable renal function or ECDs without ATN and those from non-ECD transplant patients with ATN or from ECDs with ATN (see Table 1). Results also suggested that a decreased tubular expression of PARP-1 was related to an earlier recovery of renal function. In addition, transplanted kidneys in patients with serum creatinine levels that did not fall below 1.7 mg/dL showed double the intensity of PARP-1 expression. DGF after kidney transplantation may be due to various factors [25], such as the condition of the transplanted kidney and the compliance of the vascular system in the renal graft or recipient. Nevertheless, these findings indicate that the degree of PARP-1 activation may be related to the extent of human renal tubular injury and to renal function, suggesting a role for this enzyme in the pathogenic mechanism of ATN due to IR. In a previous study of patients with ATN, our group found that the kidneys tolerating a long period of cold ischemia had the highest levels of PARP-1 (cold ischemia ,24 h, PARP-1 = 1.7160.62 vs. cold ischemia $24 h, PARP-1 = 2.8660.35). In fact, the lowest PARP-1 expression level (1-9% of tubular nuclei positive) was only observed in kidneys with less than 20 h of cold ischemia (mean, 16.36 h; range, 12-20 h) [26].
The functional capacity of renal tubular cells significantly contributes to an adequate renal function. Hence, measures taken to ameliorate the condition of these cells may also improve the outcome of kidney transplantation [27]. Although the chronic inhibition of PARP activity is likely to be harmful to the cell, it has been proposed that its transient inhibition after IR injury may prevent cell death [15].
Identification of a specific histological biomarker for the early diagnosis of tubular injury in renal biopsies is a current research challenge. Zhang et al. [28] recently used immunohistochemistry to characterize the expression of kidney injury molecule-1 (KIM-1) in renal transplant biopsies, finding a significant correlation between renal functional indices and KIM-1 staining intensity. They suggested that evaluation of KIM-1 staining may serve to optimize the diagnosis of tubular injury in allograft biopsies, similar to our proposal for PARP-1.
In summary, these results suggest a pivotal role for PARP-1 in the ATN of renal transplantation. We propose the immunohistochemical assessment of PARP-1 in kidney allograft biopsies as a risk marker for early detection of a possible delayed renal function.