Molecular Markers in Upper Urothelial Carcinoma Associated to Balkan Endemic Nephropathy. Aristolochic Acid as the Major Risk Factor of the Worldwide Disease

The role of aristolochic acid in the etiology of Balkan endemic nephropathy (BEN) and associated upper urothelial carcinoma (UUC) was recently confirmed. The aim of this study was to determine the marker(s) specific for BEN-associated UUC. A total of 82 patients with UUC (38 from the BEN region and 44 control tumors) were included in the study. The Ki-67 index in BEN tumors correlated with the grade and multifocality (p < 0.05), but in regression analysis, only the grade of BEN tumor. The p53 index was significantly higher in BEN than in control tumors (p < 0.05), as well as the alteration of p53 (p < 0.05). BEN low-stage tumors, tumors without limphovascular invasion (LVI), and tumors of the renal pelvis had a higher p53 index than the control tumors (p < 0.05, 0.01, 0.05, respectively). The Ki-67 index was higher in control tumors with high-stage and solid growth than in BEN UUC (p < 0.050, 0.005). The Ki-67 correlated with the grade, growth, stage, LVI, and multifocality of UUC on the best way, but not with the group. In regression analysis, only multifocality of UUC had predictive influence on Ki-67 activity (p < 0.001). P53 correlated with the grade, growth, and group (p < 0.05). This investigation identifies the p53 pathway as the specific cell cycle marker involved in BEN-associated UUC.


INTRODUCTION
Balkan endemic nephropathy (BEN) is a chronic tubulointerstitial disease prevalent in Bosnia and Herzegovina, Bulgaria, Croatia, Romania, and Serbia. In addition to renal disease, an increased number of upper urothelial carcinomas (UUC) has been observed in the foci of BEN. Carcinoma may occur alone or in combination with BEN [1,2]. Urothelial malignancies of the renal pelvis and ureter are significantly more frequent, up to 100 times, in endemic than in nonendemic areas. They tend to cluster in families affected with BEN, indicating an association between these diseases and, probably, a common etiologic

Patient Population
We studied 82 consecutive patients with UUC who had undergone nephroureterectomy with removal of the bladder cuff. Extended lymphadenectomy was not routinely performed. All cases of UUC were diagnosed at the Institute of Pathology, Faculty of Medicine, University of Nis. The analysis was done on 65 pelvic and 17 ureteral tumors. Patients were divided into two groups: 38 patients were from endemic settlements, villages along the South Morava River basin (BEN tumors), and 44 control subjects included rural and city populations free of BEN. In seven tumors collected from the BEN region, tubulointerstitial lesions similar to BEN were observed.
Before quantifying the immunohistochemical results, the technique quality was assessed and those areas with greater positivity were selected, avoiding peripheral area measurement, necrosis, or artifact.
Slides were reviewed independently by three investigators. Interobserver discrepancies were resolved using a double-headed microscope. Only nuclear expression was recorded for p53, p16, cyclin D1, and Ki-67. The number of distinctly positive tumor cell nuclei was counted under high power (400) using a 1010 eyepiece grid. In total, 1000 tumor cells were assessed. The number of positive nuclei was expressed as a percentage of all tumor cell nuclei counted. In this way, we defined p53, cyclin D1, and Ki-67 indices. Regarding p53, tumors showing immunoreactivity of more than 10% of the tumor cell nuclei were considered altered [13].
The staining protocol for p16 nuclear protein includes heterogeneous, homogeneous , and negative findings. A tumor was considered to have a normal heterogeneous p16 pattern if it had relatively weak nuclear staining with considerable differences in nuclear intensity, including many negative cells. Strong p16 staining was considered if the majority of the malignant cells had intensive p16 nuclear expression and p16 negative tumor cells were rare. A tumor was termed p16 negative if no malignant cells had positive staining. A tumor with or without overexpression of p16 was categorized as altered [14].
For testing HER-2 (C-erbB2) status, we used the HercepTest scoring system devised by DAKO. HER-2 cell membrane-specific immunoreactivity was scored by estimating the percentage of positive tumor cells as follows: score 0, no immunoreactive cells; score +1, positivity in <5% cancer cells; score +2, positivity in 5-50% cancer cells; and score +3, positivity in >50% of cancer cells. The specimens were considered HER-2 positive when the score was ≥2+.

Statistical Analysis
For purposes of analysis, the pathological tumor stage (low vs. high), grade (low vs. high), growth pattern (papillary vs. solid), LVI (yes vs. no), and clinical parameters (sex [M vs. F], localization [pelvis vs. ureter], multifocality [yes vs. no]) were evaluated as dichotomized variables. The Fisher's exact test and the χ 2 test were used to evaluate the expression of p16 and HER-2 with pathological parameters (stage, grade, growth, lymphovascular invasion) and multifocality of tumors. Other molecular markers (p53, cyclin D1, Ki-67) are expressed as means ± standard deviation, and statistical significance between these groups was estimated according to the Student's t-test for unpaired samples. Correlation between molecular markers (p53, p16, cyclin D1, HER-2, Ki-67) and conventional pathological parameters was analyzed by Pearson correlation and regression analysis. The result was considered statistically significant if p < 0.05.
All analyses were performed with the SPSS statistical package (SPSS version 10.0 for Windows).
There was no difference in the expression of p16, HER-2, and cyclin D1 between BEN and control UUC, as well as in their relation to the conventional pathological parameters and localizationrenal pelvis or ureter (Tables 4 and 5). The Ki-67 labeling index was significantly higher in control tumors with the high-stage (p < 0.05) and solid growth (p < 0.005) than in BEN tumors with the same morphological characteristics; and localization did not have an influence on proliferative Ki-67 activity ( A possible correlation between p53, p16, cyclin D1, HER-2, Ki-67, and standard pathological features and group was investigated. Pearson's coefficient of correlation showed that tumor suppressor p53 correlated with the grade (p < 0.05), growth (p < 0.05), and with the group (p < 0.05). Tumor suppressor p53 was identified as a specific indicator of the UUC arising in patients with BEN. Proliferative marker Ki-67 correlated with the grade (p < 0.005), growth (p < 0.01), stage (p < 0.05), LVI (p < 0.05), and multifocality (p < 0.001) of UUC on the best way, but not with the group (Table 7). However, in the regression model, only multifocality of UUC had a predictive influence on Ki-67 expression (β = 0.338, t = 3.335, p < 0.001).

DISCUSSION
The p53 gene product (located at 17p12), called "the guardian of the genome", plays a role in cell cycle regulation at the G1/S-phase checkpoint, DNA repair, and apoptosis [15]. Normally, p53 levels in a cell are low because the protein is rapidly degraded by its inhibitor, MDM 2 protein. p53 may be knocked out by deletion, by mutation, or by the action of inhibitor MDM 2 (which binds p53 and targets it for degradation).
On the other hand, cyclin D1 is an important positive regulator of the G1/S transition and is expressed in the early G1 in response to mitogenic signals [16]. Forced overexpression of G1 phase cyclins can overcome control of the G1/S checkpoint, leading to uncontrolled proliferation.
The amplification of HER-2 or overexpression of its product in urothelial carcinomas has been associated with tumor grade, tumor stage, and the patient's outcome [17]. The level of expression and the prognostic significance of HER-2 protein in urothelial cancer vary among different studies, with some revealing no prognostic relevance and other suggesting a better or worse prognosis [8,9]. This study demonstrated that analysis of p53 provides additional information about UUC from the BEN region. BEN tumors have a significantly higher p53 index than control tumors. Correlation analysis of expression-investigated molecular markers p53, p16, cyclin D1, HER-2, Ki-67, and group showed that only tumor suppressor p53 was a specific indicator of group. This investigation detected the p53 pathway as the specific cell cycle marker involved in BEN-related UUC.  Some authors suggested that accumulation of p53 protein in 10% or more of the tumor cell nuclei strongly correlates with mutations in the p53 gene [13]. Our investigation showed that 42% the BEN tumors and 20% of the control tumors had alteration of nuclear p53 immunoreactivity, with statistically significant difference. Krasteva et al. [18] screened 90 Bulgarian BEN patients for p53 gene mutations and found that tumor suppressor gene mutations were present in BEN patients. The results obtained are in support of their hypothesis that p53 gene alterations are possibly involved in BEN genetic pathways. Also, Grollman et al. demonstrated AA adducts associated with p53 mutation in UUC from a BEN region in Croatia [6]. In AAN, there is an increased incidence of UUC. Rodent models of AAN may be used to study biotransformation of AA and repair of AL-DNA adducts and, importantly, to validate the use of AL-DNA adducts as biomarkers of disease. The murine model will be particularly useful for determining genetic susceptibility to AAN and, by implication, to BEN [19].
As in AAN, patients with BEN are at increased risk for the development of UUC after kidney transplantation. Thus, 33.3% of patients with BEN developed UUC, compared with a 0.67% prevalence of urinary tract tumors among transplanted patients with other causes of end-stage renal disease [20].
Our investigation showed a significant difference in p53 expression between BEN and control tumors that was connected with the stage of UUC and renal pelvis localization. BEN low-stage UUC have a higher p53 index than control superficial tumors. We did not find any difference in p53 activity between high-stage BEN and control tumors. Renal pelvis UUC from endemic settlements had a significantly higher p53 index than control; the opposite was found in tumors of the ureter, however, without statistical significance.
Mutations in the p53 gene were frequently found in invasive UUC as well as in high-grade superficial UUC (including carcinoma in situ, the putative precursor of invasive UUC), whereas these mutations were rare in well-differentiated superficial UUC [21,22].
Our comparative study of E-cadherin expression in UUC between patients from BEN and nonendemic areas showed that BEN low-grade and low-stage tumors have important reduction of Ecadherin expression. The growth pattern had a predominant influence on E-cadherin expression in BEN tumors, as previously described [23]. Decreased E-cadherin expression is indicative for invasion and loss of local control. It is of particular interest to predict recurrence in patients with low-grade noninvasive tumors [24]. Deregulation of p53 and E-cadherin in low-stage UUC from BEN areas may be indicative for a potentially more aggressive tumor.
Low incidence of p53 mutations in Ta/T1 bladder tumors may suggest that p53 alterations in superficial cancer are related to a more aggressive phenotype and a higher risk of recurrence. Loss of p53 is a late event in tumor development. The earlier stages of tumor evolution are not prevented by p53 [25,26,27].
Different tissues have different susceptibility to p53 deficiency. The urothelium is less susceptible to tumorigenesis during p53 deficiency than fast-renewing cells, such as lymphocytes and fibroblasts [28]. Cycling cells in any given environment are subjected to a different kind of stress, and that could compromise DNA synthesis. Normal human cells elicit the p53-dependent and p53-independent responses, resulting in suppression of cyclin-dependent kinase 2 (CDK2) activities and inhibiting centrosome duplication. Centrosome amplification is one of the causes of carcinoma in situ, and is associated with a loss or mutational inactivation of p53. In human cells lacking functional p53, when DNA synthesis is blocked, p53-independent pathways suppress CDK2 kinase activity and do not allow centrosome reduplication [29,30,31].
On the other hand, investigation of proliferate Ki-67 activity showed that control high stage, as well as control tumors with solid growth, had a higher Ki-67 index than BEN tumors with the same morphological characteristics. There was no difference in Ki-67 activity between BEN low-stage and control low-stage tumors, as well as depending on localization (renal pelvis or ureter).
In BEN tumors with extreme cell atypia, a higher degree of apoptosis was shown, but proliferation does not dramatically increase with UUC progression to high grade. Chronic ischemia in surrounding tissue may facilitate apoptosis in BEN-related tumors [32]. Investigation of the pathogenesis of BEN has suggested that apoptosis in tubular epithelia has a part in the silent development of the disease [33,34].
Our earlier analysis of the morphological characteristics of UUC in BEN and nonendemic regions showed that the best characteristic that differentiated them was growth pattern; i.e., solid growth for BEN tumors and papillary for control tumors [35]. In this study, control UUC with solid growth had a higher Ki-67 index than BEN solid tumors. In UUC with a papillary configuration, there was no difference in proliferate Ki-67 activity.
In conclusion, investigation of multiple molecular markers indentifies tumor suppressor p53 as an indicator of the group, where BEN UUC had a higher p53 index than control tumors. Differentiation of BEN tumors had predictive influence on Ki-67 activity, and their proliferation was lower with stage progression, as well as with solid growth in comparison with control tumors.