Introduction

Prostatic intraepithelial neoplasia (PIN) is characterized by intraluminar proliferation of epithelial cells in ducts and acini. PIN frequently coexists with prostatic carcinoma (PCA) (Bostwick, 1995; Bostwick and Brawer, 1987; Skjørten et al, 1997) and is commonly found in the nontransition zone, which is the predominant site for PCA (de la Torre et al, 1993; Qian et al, 1997). According to histologic and cytologic findings, PIN is divided into high-grade PIN (HGPIN) and low-grade PIN (LGPIN) (Montironi et al, 1996). Previous histopathologic study of whole-mount prostatectomy specimens showed the proximity of HGPIN lesions but not LGPIN lesions to PCA (Shin et al, 2000). Recent studies revealed that HGPIN and PCA share common cytogenetic features. Allelic loss of chromosome 8p is frequent in both HGPIN and invasive PCA (Emmert-Buck et al, 1995; Häggman et al, 1997). Mutations of the H-ras gene were closely associated with progression of HGPIN into invasive PCA in transgenic mice, although this was not confirmed in humans (Shibata et al, 1996). These findings indicated that HGPIN but not LGPIN is the most likely precursor lesion for PCA.

HGPIN lesions could be defined only at the microscopic level and therefore immunohistochemical procedures have been the main modalities used to analyze its biologic characteristics. Crowded cells in the HGPIN lesions were occasionally labeled with an antibody (MIB-1) against proliferating cells (Häussler et al, 1999). Meanwhile, foci of HGPIN were strongly immunoreactive with Bcl-2 (Häussler et al, 1999; Stattin et al, 1996), suggesting that programmed cell death (apoptosis) is prolonged in HGPIN.

Fas antigen is a 45-kDa transmembrane protein of the TNF receptor superfamily that can induce programmed cell death (apoptosis) through cross-linkage with Fas ligand (FasL) (Nagata, 1997; Suda et al, 1993). Fas is located on chromosome 10q24.1 and is comprised of 9 exons and 8 introns. The Fas extracellular domain contains three cysteine-rich motif repeats encoded by exons 2 to 5. Its transmembrane domain consists of the last 49 bp of exon 6 and the first 2 bp of exon 7. The remaining 81 bp in exon 7, the 25 bp in exon 8, and 329 bp in exon 9 comprise the cytoplasmic region (Cheng et al, 1995). Fas is expressed in a wide range of tissues including the thymus, liver, and activated T and B lymphocytes (Itoh et al, 1991; Watanabe-Fukunaga et al, 1992). A mutation in the death domain of the Fas gene leads to loss of its apoptosis function, a loss-of-function mutation, which might contribute to the pathogenesis of human malignancies. Indeed Fas gene mutations have been reported in both the lymphoid lineage and epithelial malignancies: approximately 10% of cases with multiple myeloma (Landowski et al, 1997) and sporadic non-Hodgkin’s lymphoma (Grønbæck et al, 1998), 7.7% in lung cancer (Lee et al, 1999a), and 28% in urinary bladder cancer (Lee et al, 1999b).

Laser capture microdissection under direct microscopic visualization enables rapid one-step procurement of selected human cell populations from histologic sections. This method facilitates microdissection of selected cells, and thus extensive studies of the objective lesions are now possible. Using this method on whole-mount samples, we selectively microdissected numerous PIN and PCA lesions from 27 patients with PCA. Genomic DNA extracted from each lesion was analyzed for Fas mutations.

Results

Fas Gene Mutations

As shown in Table 1, four mutations of the Fas gene were detected in four HGPIN lesions from four cases. There were no prominent differences in frequency of mutations or loss of heterozygosity (LOH) in HGPIN or PCA lesions between noncastrated and castrated cases. All mutations were point mutations: three missense and one nonsense mutation detected in exon 9, which encodes the death domain region of the Fas receptor (Nagata, 1997). Substitutions at codon 260 of the Fas cDNA sequence (Gene Bank accession No. M67454), Gln to Stop (Case 10) and Gln to Arg (Case 21), respectively, were found in the HGPIN lesions. All of the mutations were transitions: C to T and A to G. No G to C or C to A transversions were found. All of the HGPIN lesions with mutations were present in the nontransition zone. Neither PCA nor benign proliferative glands adjoining HGPIN and/or PCA showed Fas mutations.

Table 1 Mutations and LOH of the Fas Gene in High-Grade Prostatic Intraepithelial Neoplasia (HGPIN) and Prostatic Cancer (PCA)
Full size table

Allelic Status

Sixteen of 27 (59.6%) cases were heterozygous for one or more sites of the known biallelic polymorphisms, ie, at positions −1377, −670, 416, and 836. Of these 16 cases, 5 (31.3%) with PCA and 4 (25%) with HGPIN showed LOH in the promoter region (−670) or exon 7. The HGPIN lesion in Case 24 had a missense mutation at position 1099 and LOH at −670. In Case 21, the HGPIN lesion had a mutation at exon 9 but no LOH, and the PCA lesion had LOH but no mutation.

Immunohistochemistry

Results of immunohistochemical study in lesions with Fas gene mutations and LOH are summarized in Table 1. Fas protein was expressed in 11 of 15 (73.3%) HGPIN lesions and 5 of 6 PCA (83.3%) lesions. No relationship was found between immunoreactivity for anti-Fas antibody and mutation or LOH in HGPIN and PCA lesions.

Discussion

Information regarding the molecular genetic characteristics of PIN was quite limited until development of the microdissection technique. Previous studies showed that allelic loss of chromosomes 8p, 10q, and 16q was frequent in both HGPIN and invasive PCA, suggesting the involvement of tumor suppressor genes or oncogenes at these loci (Emmert-Buck et al, 1995; Gray et al, 1995; Häggman et al, 1997; Strup et al, 1999). Through construction of a detailed deletion map spanning 10q23–25, Gray et al (1995) suggested the presence of prostate tumor suppressor genes near the 10q23–24 boundary, which is close to the location of the Fas gene, 10q24.1. Using the laser capture microdissection method, we analyzed the Fas gene mutations in numerous HGPIN lesions; Fas gene mutations were detected in 4 of 27 (14.8%) cases or 4 of 111 (3.6%) lesions with HGPIN. In contrast, none of 55 lesions with PCA had Fas gene mutations, indicating that PCA develops in HGPIN without Fas gene mutations. Fas gene mutations were never detected in benign proliferative glands adjoining HGPIN and/or PCA. Taken together, these observations suggest that Fas gene mutation might not contribute to the pathogenesis of PCA, in contrast to lung cancer, hematologic malignancies, and bladder cancer (Lee et al, 1999a, 1999b).

With regard to the sites and patterns of mutations in the Fas gene, a point mutation at codon 253 was reported in two patients with multiple myeloma (Landowski et al, 1997). In non-Hodgkin’s lymphomas, mutations at codons 248 and 251 were identified (Grønbæck et al, 1998). Lee et al (1999b) reported that 8 of 12 mutations found in bladder cancer showed a G to A transition at codon 251, suggesting that this might be a mutational hotspot. Two of the current HGPIN lesions showed mutations at codon 260. All of the mutations in the current series were transitions, suggesting that some “endogenous” mutagens act in the pathogenesis of HGPIN.

Missense mutations in the death domain have been suggested to affect receptor function in a dominant-negative fashion (Fisher et al, 1995), ie, mutant Fas protein derived from one mutated Fas allele might bind with normal Fas protein derived from the other normal allele to construct a structurally abnormal Fas trimer, which might have a defect in binding to adapter proteins. Among four HGPIN lesions with Fas gene mutations, one informative case (Case 24) had LOH at position −670. As the distance between the site of LOH and the mutation was approximately 10 kb, LOH might involve the mutation site in the same allele. Therefore, it is reasonable to consider that the LOH and mutation found in the HGPIN lesion in Case 24 occurred in different alleles, thus resulting predominantly in production of mutant Fas protein. The occurrence of LOH was unknown in the remaining three cases (Case 3, 10, 21) with Fas mutations. Fas function might have been lost or reduced due to a dominant-negative effect of mutant Fas protein in cases without LOH or because of predominant production of mutant Fas protein in cases with LOH. Fas-mediated apoptosis was suggested to be disrupted in these four HGPIN lesions.

The normal DNA repair mechanism is important for maintaining the integrity of the genome. Humans are frequently exposed to naturally occurring DNA-damaging agents, and the combined occurrence of DNA damage and impaired DNA repair function results in development of neoplasia. Indeed, replication errors, as revealed by microsatellite instability (MSI), were reported in cases with PCA (Dahiya et al, 1997; Rohrbach et al, 1999). The occurrence of LOH also indicates the underlying genetic instability in lesional proliferating cells. Rohrbach et al (1999) reported that MSI and LOH were found in 35% and 16%, respectively, of their PCA cases. In the present study, LOH at four sites in the Fas gene was found in 31.6% of PCA and 25% of HGPIN lesions. LOH of the Fas gene was never found in the benign lesions. Previous studies suggested that HGPIN is the precursor lesion of PCA (Bostwick, 1995; Bostwick and Brawer, 1987; Skjørten et al, 1997), and the present results showed that genetic instability occurred during the early phase of prostatic carcinogenesis.

In conclusion, Fas gene mutations were exclusively found in the HGPIN lesions, which might underlie the development of these lesions.

Materials and Methods

Patients

Twenty-seven patients, who had received total prostatectomy under the diagnosis of PCA, were selected for the present study. All of the patients were admitted to our hospitals between 1996 and 1998. Fourteen patients received androgen deprivation therapy (castration): LH-RH agonist (leuprolide) for 2 to 6 months (mean 3.4 months) together with the antiandrogen agents (flutamide or chlormadinone). The remaining 13 patients did not receive this treatment (noncastrated). The present cases were conventionally subdivided into the preoperatively noncastrated group (13 cases) and medically castrated group (14 cases). None of the 27 patients received preoperative chemotherapy or radiation therapy.

Histologic findings in a subpopulation of these cases were reported previously (Shin et al, 2000). Based on the American staging system (modified by Whitmore-Jewett) (Bostwick et al, 1994), pathologic stage was determined as follows: 11 (41%) cases were in stage T2 and 16 (59%) were in stage T3. Histologic specimens were fixed in 10% neutral buffered formalin and routinely processed for paraffin embedding. Serial sections 5-μm thick were cut and stained with hematoxylin and eosin, and reviewed independently by two pathologists (HT and KA). The mean number of sections examined was 9.3 per case. In problematic cases, preservation of the basal cell layer, a reliable criterion for benign prostatic lesions, was immunohistochemically examined using basal cell–specific monoclonal anti-high molecular weight cytokeratin antibody (34βE12; Enzo Diagnostics, New York). Diagnosis of HGPIN was made according to histologic and cytologic features; ie, intraluminal proliferation of glandular epithelial cells with large and prominent nuclei. The basal layer was partially disrupted in HGPIN.

Laser Capture Microdissection and DNA Extraction

Microdissection of each lesion was performed using a PixCell laser capture microscope (Arcturus Engineering, Santa Clara, California) according to previously described methods with some modifications (Bonner et al, 1997; Emmert-Buck et al, 1996). Briefly, histologic sections were dehydrated, and then the histologic fields of interest were selected, overlaid with thermoplastic film mounted on transparent cap, and captured on the film through laser energy (Fig. 1). The dissected pieces were allowed to adhere to the transparent cap and were collected in 0.5 ml Eppendorf tubes. The cells were subsequently resuspended in 20 to 50 μl of extraction buffer containing 10 mm Tris (pH 8.0), 2 mm EDTA, 0.2% Tween 20, and 200 μg/ml proteinase K, and incubated overnight at 37° C. The mixture was heated at 100° C for 10 minutes to inactivate proteinase K, and 3% to 5% of the solution was used as a template for each PCR. The total number of microdissected lesions from 27 cases was 193: 111 lesions from HGPIN, 55 from PCA, and 27 from benign glands.

Figure 1
figure 1

High-grade prostatic intraepithelial neoplasia (HGPIN) lesion (A) in a representative case was successfully microdissected (B). Note successful resection of intraductal epithelial cells. Hematoxylin and eosin staining; original magnification, ×200.

Full size image

Detection of Mutations and LOH

The death domain is necessary for transduction of the apoptotic signal (Cheng et al, 1995; Itoh and Nagata, 1993; Itoh et al, 1991). Therefore, we examined mutations in exons 7, 8, and 332 bp of exon 9. The primers used for PCR analysis are listed in Table 2. DNA was subjected to first-round PCR consisting of 10 cycles with oligonucleotide primers, followed by a second PCR consisting of 35 cycles using 0.1% of the first-round PCR products as the template, then denaturation for 30 seconds at 95° C, annealing for 30 seconds at variable temperatures, and extension for 30 seconds at 72° C in a 9700 Applied Biosystems Thermocycler (Foster City, California). PCR products were purified using a QIAquick PCR Purification Kit (Qiagen, Valencia, California), and were sequenced by the dideoxy chain termination method using a DNA sequencing kit (Applied Biosystems). The samples were analyzed using a Genetic Analyzer (ABI PRISM 310′; Applied Biosystems). PCR products with suspected mutations were cloned in pCR 2.1-TOPO (Invitrogen, Carlsbad, California), and sequenced to confirm the mutation.

Table 2 Amplification Primers for Mutation Analysis of the Fas Gene
Full size table

LOH was examined at four sites of known polymorphisms, ie, at position −1377, −670 (promoter region), 416 (exon 3), and 836 (exon 7). DNA was amplified using primers flanking the four polymorphic sites (Table 1). Polymorphisms at −1377, 416, and 836 were examined by direct sequencing, and that at position −670 by restriction fragment length polymorphism by digestion with Mva I (Fermentas, Vilnius, Lithuania) (Huang et al, 1997).

Immunohistochemistry

Immunohistochemical study of paraffin sections was carried out using the avidin-biotin peroxidase complex method. For detection of Fas protein, mouse antihuman Fas antibody (4B4-B3) that recognizes extracellular domain of Fas was prepared by Dr. S. Nagata (unpublished data). The histologic sections were deparaffinized and rinsed in three changes of PBS. Then the antibody, diluted 1:200, was applied at 37° C for 45 minutes. Positive cells showed an intracytoplasmic dot-like staining. Staining manner in each case was distinct, ie, almost all cells were positive in positive cases and negative in negative cases.