Renal cancer accounts for 3% of adult malignancies. An estimated 35,710 new cases are expected to occur in the U.S. in the current year and 12,480 patients will die of this disease [1]. The majority of renal cancers arises from the proximal tubular epithelium, with a characteristic clear or granular cell appearance by light microscopy, and is referred to as renal cell carcinoma (RCC).

Recent evidence suggests that solid renal parenchymal tumors arising in the distal portion of the nephron, such as oncocytomas, chromophobe renal cell carcinomas and collecting duct carcinomas represent an heterogeneous group of neoplasm from both clinical and biological perspectives [2]. Collecting duct carcinoma (CDC), also known as Bellini duct carcinoma, is a rare but highly aggressive renal neoplasm arising from the distal portion of the nephron, and represents approximately 2% of all the RCCs [3, 4]. Clinically, CDC appears as a renal mass often accompanied by flank pain and hematuria and is frequently mistaken for RCC or transitional cell carcinoma of the renal pelvis [3, 5, 6]. CDC, however, can be identified based on gross, microscopic, histochemical, and immunohistochemical features. Macroscopically, CDCs are often located at the confluence of the medulla and renal pelvis, and show a characteristic gray-white-tan color, with absence of foci of necrosis and hemorrhage [7]. Histologically, CDC presents a tubulo-papillary morphology, often accompanied by desmoplasia, atypia in collecting ducts, and intratubular spread [7], features rarely seen in RCC.

Histochemically, CDC cells contain intracytoplasmic mucicarminophilic material, where RCCs do not [7]. CDCs are also positive by immunohistochemical staining with high-molecular-weight keratin and lectin, proteins typically expressed in the epithelium of the distal tubules [8, 9]. Conversely, almost all other renal carcinomas express antigens widely expressed in the cells of the proximal tubules, such as low molecular weight cytokeratins and vimentin [9]. Although little is known about the genetic profile of CDCs, a DNA flow cytometry study has demonstrated aneuploidy in 90% of these tumors [5], and cytogenetics has shown frequent monosomy of chromosomes 1, 6, 8, 14, 15, and 22 [10, 11].

Loss of heterozygosity (LOH) analysis in six CDCs revealed allelic loss at chromosomes 8p and 13q in 50% of the tumors, and rarely at the short arm of chromosome 3 [12, 13]. This data suggests that distinct genetic alterations from those observed in RCC, in which 3p LOH is common [14], occur in the development of this rare renal tumor.

To improve our understanding of the biology of CDC and to explore the possibility that different genes may be involved in the etiology and prognosis of this neoplasm, we analyzed by immunohistochemistry eleven cases of CDC for the expression of five genes (Fez1, Fhit, p53, p27, and bcl2) often involved in the development of many common cancers. These findings were correlated with conventional clinical-pathological features including clinical outcome.

The application of the most recent molecular cytogenetic techniques revealed that renal parenchymal tumors can be classified into distinct subtypes based on the combination of specific genetic alterations [18]. The pathological and immunohistochemical description as well as the cytogenetic abnormalities support the hypothesis that CDC is more similar to urothelial carcinoma than to clear cell carcinoma of the kidney. Indeed, whereas allelic deletion of the short arm of chromosome 3 is considered a genetic hallmark of clear cell carcinoma, LOH at chromosomes 8p, 9p, and 17p has been frequently described in both transitional cell carcinoma and CDC. Here, we have reported the results of our immunohistochemical analysis of the expression of five genes (FEZ1, FHIT, P53, P27 ^{kip 1}, and BCL2) in a relatively large series of CDCs (eleven cases), considering the rarity of this tumor.

FEZ1/LZTS1 (leucine zipper tumor suppressor 1) is a putative tumor suppressor gene located at 8p22 [19]. Studies have indicated this chromosomal region is the location of an important tumor suppressor gene (TSG) [20]. LOH at 8p has been described in 50% of the CDCs studied [12], suggesting that a TSG in this region may also play a role in the development of this rare tumor. In our study, we found loss of Fez1 expression in the majority of CDCs studied and a correlation with the presence of lymph node metastasis. Furthermore, lack of Fez1 protein correlated with a poorer prognosis in 90% of patients with median survival of 17 months.

FEZ1 encodes a 67-kDa leucine-zipper protein with a region of similarity to cAMP-dependent activated protein [19]. Mutations of FEZ1 gene have been reported in several solid tumors, including prostate, breast, esophageal, and gastric carcinomas [17, 19]. In addition, reduced Fez1 expression is associated with high-grade transitional cell carcinoma of the bladder [21]. Recent studies have shown that the introduction of FEZ1 into Fez1 null cancer cells reduced cell growth with the accumulation of cells at late S to G₂/M phase of the cell cycle. Conversely, inhibition of Fez1 expression stimulated cells growth [22]. Furthermore, LOH at the chromosomal region where the FEZ1 gene lies (8p21-22) has been also associated with the invasive behavior of breast cancer [23] and with prostate cancer progression [24]. These data are consistent with an important role of FEZ1 in several human cancers including CDC.

The tumor suppressor gene FHIT maps to the short arm of chromosome 3 (3p14.2), encompasses the common FRA3B fragile region, and encodes for a protein of 16.8 kDa, with diadenosine triphosphate hydrolase activity [25]. Reduction of Fhit protein expression as consequence of alteration of the FHIT gene has been observed by immunohistochemistry in many types of cancers [26, 27]. Although Shoemberg et al. did not detect LOH involving chromosome 3p [12], Hadaczek et al. reported LOH at 3p in two cases of CDC [13]. The same authors described a correlation between reduced Fhit expression and 3p allelic loss in renal carcinomas, particularly in CDCs [28]. While in our study Fhit inactivation does not seem to be a common event in CDC, an involvement of the FHIT gene in tumorigenesis of this rare tumor cannot be excluded.

The TP53 tumor suppressor gene maps to chromosome 17p13.1 and plays a major role in DNA transcription, cell growth, proliferation and apoptosis process [29]. In normal cells, expression of wild type p53 protein is generally below the detection level when studied by immunohistochemical method. However, p53 gene point mutations occur frequently (22–76%) in different solid neoplasms. Mutated p53 protein, being more resistant to degradation, accumulates in the cells and can be detected by immunohistochemistry. Although an association between p53 protein overexpression and tumor stage, grade and survival has been observed in RCC [30] our data suggest that involvement of p53 alterations does not occur with the same frequency in CDC.

P27 is a member of the universal cyclin-dependent kinase inhibitor (CDKI) family. The expression of this important protein is regulated by cell to cell contact inhibition as well as by specific growth factors, such as transforming growth factor (TGF-β). In addition to its role as a CDKI, p27 is considered a putative tumor suppressor gene, a major regulator of drug resistance in solid tumors, and a promoter of apoptosis. p27 acts also as a safeguard against inflammatory injury and has a role in cell differentiation [31]. It has been suggested that loss of the p27 negative cell cycle regulation may contribute to oncogenesis and tumor progression in several tumor types. In renal cell carcinoma, Kamai et al. reported that low level of p27 protein was associated with tumor invasion and unfavorable prognosis, suggesting p27 as a powerful prognostic marker for survival in urinary tract cancer [32]. Masuda et al. indicated that p27 has an independent predictive prognostic value for transitional cell carcinoma of the renal pelvis [33]. Our results show that p27 loss or subcellular compartmentalization represents a frequent feature in CDCs. Previous studies have noted that cytoplasmic localization of p27 lead to an inactivation of its normal function as negative cell cycle regulator [34].

Nevertheless, we did not find statistical correlation to assess its involvement in CDC biology, possibly due to the limited number of tumors studied.

The proto-oncogene BCL2, implicated in the regulation of cell death by inhibiting apoptosis, seems to be vital in normal kidney morphogenesis. In fact, Bcl2 deficient mice develop polycystic kidneys characterized by dilated proximal and distal tubules [35]. High levels of Bcl2 protein expression have been found in many different types of cancer, suggesting a possible role for Bcl2 deregulation of apoptosis and malignant tissue transformation. Expression of Bcl2 has also been associated with poor prognosis in patients with various cancers including prostate cancer [36]. In the present study, Bcl2 expression was not associated with any clinical-pathological variables.

Our results suggest a potential association between FEZ1 expression and CDC pathology and prognosis. No similar patterns were seen for any of the other markers studied. Even so, the statistical power of the study was limited and negative findings should not be construed as evidence that these markers are not important. Rather, a larger study would need to be carried out to further investigate their role in CDC.

Our results suggest that Fez1 expression may be associated to both clinical-pathological features and survival in patients with CDC. FEZ1 gene alterations may be linked to the high frequency of LOH found at 8p22, where FEZ1 resides. The lack of similar association for the other four genes studied may be due to the low statistical power of the study.