Decreased Androgen-responsive Growth of Human Prostate Cancer Is Associated with Increased Genetic Alterations¹ Free

Androgens play an important role in the tumorigenesis and progression of prostate cancer, the most commonly diagnosed malignancy and the second leading cause of cancer death among the Western male population. During prostate cancer progression, there is a transition from the androgen-responsive to the androgen-unresponsive stage that determines the clinical course and outcome of the patients. Most of the patients with advanced prostate cancer are treated with androgen deprivation therapy that results in a dramatic regression of androgen-responsive cancer cells (1). About 70–80% of the patients initially respond to the androgen ablation therapy, but the tumor eventually progresses in to an androgen-unresponsive stage resulting in a poor prognosis (2, 3). There are no effective alternative therapies available currently for this lethal subset of cancer patients. Therefore, it is necessary to understand the genetic basis and molecular mechanisms involved in the transition from androgen-dependent to androgen-refractory stage to develop new treatment/therapy for prostate cancer.

Most of the human cancers are associated with genetic instability characterized as chromosomal instability (aneuploidy, deletions ,or amplification) or MSI⁴ (4, 5). The later is more pronounced in hereditary nonpolyposis colorectal cancer and is attributed to the loss of MMR genes (6). However, in sporadic cancers MSI is evidenced without the defect in MMR and, thus, some additional mechanisms might be involved (7). Like most cancers, prostate tumor progression and the acquisition of androgen-refractiveness may also be promoted by the accumulation of genetic changes. Recently, Chen et al. (8) observed loss of DNA MMR proteins among the prostatic cancer cell lines. Similarly, a down-regulation of MMR enzyme activity as well as a lower expression of MMR-related genes has been reported by Yeh et al. (9). Furthermore, Leach et al. (10) also reported the mutation in the hMSH2 MMR gene in the LNCaP cell line.

Cytogenetic studies of prostate cancer have revealed consistent chromosomal abnormalities (11, 12, 13). The altered chromosomal regions can be scrutinized additionally for MSI and AL or loss of heterozygosity with the help of the powerful microsatellite analysis assay. Frequently deleted regions from various chromosomes suggest the presence of TSGs involved in the carcinogenesis of prostate cancer (14). Therefore, identifying the genetic alterations in cancer-related genes such as oncogenes, TSGs, and DNA MMR genes is one main focus for the prostate cancer research. Thus far, no oncogenes or antioncogenes have been conclusively correlated with the initiation or progression of prostate cancer (15). This is largely attributable to two reasons: (a) the lack of suitable in vitro models; and (b) the astoundingly heterogeneous nature of the prostate tumor that makes it complicated in understanding the genetic alterations (16, 17).

Recently, Lin et al. (18) developed a unique prostate cancer cell model by using sublines of LNCaP cells originally established from a human prostate adenocarcinoma (19). During in vitro cell culturing, the LNCaP cells accumulate altered biological behaviors. They are characterized by expressing the same level of functional androgen receptors but exhibiting a different responsiveness to androgen stimulation at later stages (18). This tumor cell model may closely resemble the different stages of tumorigenesis with the acquisition of androgen-refractiveness as seen in prostate cancer patients. Hence, this model provides the opportunity to study the progressive changes in the primary and metastatic stages developed from the same LNCaP cells.

To investigate the genetic alterations involved in prostate tumor progression, we performed microsatellite and cytogenetic analyses in androgen-responsive and androgen-unresponsive stages of the LNCaP cell model. An increasing number of microsatellite DNA markers showed MSI and AL during this process, whereas the detected karyotypic changes remain minimal. Additional analyses of genetic alterations in clinical archival samples corresponded with the observations in the cell line model.

Female and male nude mice (nu/nu), 6–8 weeks of age, were obtained from Harlan Sprague Dawley Inc. (Wilmington, MA). The mice were housed in laminar flow cabinets under specific pathogen-free conditions. Animal protocols used in this study were approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee, which complied with the Public Health Service policy on Human Care and Use of Laboratory Animals.

The LNCaP cell model used in the present study was developed and characterized by Lin et al. (18). Briefly, three sublines of LNCaP were designated as C33 (LNCaP passage number <33), C51 (LNCaP passage number between 45 and 70), and C81 (LNCaP passage number between 81 and 120). Cell lines were maintained in RPMI 1640 supplemented with 5% fetal bovine serum and 1% penicillin-streptomycin (Life Technologies, Inc., Grand Island, NY). Cultured cells were fed twice a week and harvested with trypsin once a week; one passage is equal to one trypsinization.

Prostatic adenocarcinoma tissues were obtained from 15 patients. The tissue was fixed in 10% neutral formalin, embedded in paraffin, and 5-μm sections were stained with H&E for pathological evaluation. The combined Gleason grade and pathological stage was assessed by a pathologist (Dr. Sonny L. Johansson, Department of Pathology, University of Nebraska Medical Center, Omaha, NE) according to American Joint Committee on Cancer (20).

C33 and C81 stage cells were prepared for karyotype analysis by treating the subconfluent cells with Colcemid (final concentration, 0.01 mg/ml), washed with HBSS, and trypsinized. Single-cell suspensions were centrifuged, and the pellet was exposed to a hypotonic solution (0.06 m KCl). After centrifugation, the cells were fixed with ice-cold methanol and acetic acid (3:1) solution, washed twice with the fixative, and dropped on glass slides for air-drying. Slides were aged and G-banded with Wright’s stain procedure (21). Fifteen metaphase spreads from each sample were counted and analyzed to determine the numerical and structural chromosomal abnormalities.

Tissue sections of 5 μm, each containing cancerous and noncancerous cells on the same slide from formalin-fixed, paraffin-embedded tissue samples were deparaffinized and stained with H&E. Sections were treated with collagenase-trypsin buffer (0.6% collagenase-H; 0.5% trypsin type XI in PBS) for 2 h at 37°C, dehydrated in graded alcohols, kept in xylene, and dried before laser microdissection. Cells were microdissected by using a PixCell laser capture microscope with an infrared diode laser and were collected on CapSure (Arcturus Engineering Inc., Santa Clara, CA). The captured cells were incubated overnight at 37°C in 25 μl of Tris-EDTA buffer [10 mm Tris-HCl (pH 8.0), 1 mm EDTA, and 1% Tween 20] and 25 μl of proteinase K (0.1 mg/ml). After incubation, the samples were boiled for 10 min to inactivate the proteinase K. Because the amount of the extracted DNA from the microdissected samples was too small to be determined by spectrophotometry, the quantification of the normal and cancerous cell DNA before PCR amplification was not possible. The tubes were briefly spun, and 2 μl of the supernatant were used for PCR.

For the LNCaP sublines, the genomic DNA was purified using the guanidine isothiocyanate-cesium chloride cushion ultracentrifugation method as described previously (22). The genomic DNA was purified by SDS-proteinase K digestion and phenol:chloroform (1:1, v/v) extraction.

MSI and the AL were performed with simple sequence repeat microsatellite DNA markers using a PCR. One hundred sixty-five tandem repeat markers (Research Genetics, Huntsville, AL), equally distributed over all of the chromosomes with an average distance between two markers of 25 cM were chosen. Additional markers were applied in the regions, which showed MSI and AL. We additionally used mononucleotide markers BAT-25 and BAT-26 from a reference panel of 5 markers recommended by the National Cancer Institute (Bethesda, MD; Ref. 23) to have an idea about the MSI in the sublines of LNCaP cell model. For these mononucleotide markers, we used DU145, PC3, LNCaP-C33, -C51, and -C81 prostatic cancer cell lines. Genomic DNA from BPH tissues was used as a control for the size of the amplified band with BAT-25 and BAT-26 markers.

The PCR mixture contained 1 × PCR buffer [20 mm Tris-HCl (pH 8.4), 1.5 mm MgCl₂, and 50 mm KCl]; 2 mm of dGTP, dATP, and dTTP; 25 μm of dCTP; 0.4 μCi of [α-³²P]dCTP; 5 ng of each primer; 20 ng of genomic DNA from the cell lines; 0.25 units of Taq-Polymerase; and double distilled H₂O in a final volume of 10 μl. The PCR reaction involved an initial denaturation of 5 min at 95°C, 27 cycles of 30 s at 95°C, 1 min at 55°C, and 1 min at 72°C, and a final elongation of 5 min at 72°C. In case of microdissected samples, the PCR cycles were increased to 40, and the final extension was for 10 min at 72°C.

The amplified PCR products were denatured at 95°C for 10 min after the addition of 30 μl of formamide stop buffer (95% formamide and 0.25% bromphenol blue and xylene cyanol FF) and were fractionated onto a 6% polyacrylamide, 7.7 m urea gel (SequaGel; National Diagnostics). Gels were run at 2000 V for ∼2 h at 55°C, transferred to blotting paper, dried for 30–40 min, and exposed to Hyperfilm-MP (Kodak) for 18–48 h. Any change in length either because of insertion or deletion of repeating units in a microsatellite within a tumor as compared with the normal tissue was defined as the MSI. A change in the allele band intensity related to the control band was considered as AL or gain.

In a xenograft animal model, C33 and C81 stage cells of the LNCaP cell model showed differential tumorigenicity when implanted s.c. in the nude mice (Fig. 1). To examine the androgen effect, female mice were used as a control. Both males and females (five animals per group) implanted with C81 passage cells showed 40% and 60% tumor formation on day 14 and day 21, respectively, and the tumor formation was 100% on day 21 or day 28. On the other hand, C33 passage cells showed much slower tumor development that was dependent on androgens both in male or female mice (Fig. 1). The data showed that the C81 stage cells are more tumorigenic, and androgens have no significant effect on their latent period of tumor formation.

Representative karyotype of the LNCaP cell model in C33 and C81 stage cells was shown in Fig. 2. Chromosomal studies from both stages exhibited an identical chromosomal pattern except a rearrangement on chromosome 3q, an interstitial deletion on 3p, and a gain of one copy of the Y chromosome in C81 stage cells (Fig. 2,B). In both cases, the numerical abnormalities involved one missing copy of chromosome 2, two missing copies of chromosome 6, and one missing copy of chromosome 13. Structural abnormalities involved unbalanced translocations of 1p and 15q, deletions of 2p and 10q, and the presence of an unidentifiable marker chromosome on 16q (Fig. 2). The karyotype nomenclature for C33 cells was: 85–89, XXYY, der (1)t(1;15; p22;q22)×2, del (2; p13p21),-6,-6, del (10; q22q24)×2,-13, der (15)t(1;15; p22;q22), add (16; q24)×2,+mar[cp15]. The karyotype nomenclature of C81 cells was: 85–90, XXYY,+Y, der (1)t(1;15; p22;q22)×2, del (2; p13p21), add (3)del (3; p13p21),-6,-6, del (10; q22q24)×2,-13, der (15)t(1;15; p22;q22), add (16; q24)×2,+mar[cp15].

Genomic DNA from prostatic cancer cell lines (DU145, PC3, LNCaP-C33, -C51, and -C81) was analyzed for MSI studies with mononucleotide repeats BAT-25 and BAT-26. Because of the lack of normal DNA from the same patient from which these cell lines were derived, we used two BPH samples as a control (Fig. 3). LNCaP cell lines as well as DU145 showed alteration in mobility with both BAT-25 and BAT-26 markers compared with normal controls. The PCR amplification products for PC3 cells did not show alteration in size (Fig. 3). This suggested that MSI was not detectable in the PC3 cell line, whereas other prostatic cell lines showed MSI.

Three sublines (C33, C51, and C81) were screened with di-, tri-, or tetranucleotide repeat markers for the associated genetic changes during the progression of this LNCaP cell model. One hundred forty-three of the tandem repeat primer sets were informative, whereas 22 primer sets were noninformative. Example of a noninformative and informative cases (without any change in genetic profile) is shown in Fig. 4,A. The early C33 (androgen-responsive) stage was considered as a control to compare its genetic profile with the mid C51 and the late C81 stages. Fig. 4,B represented the examples of MSI in C51 and C81 stages of the LNCaP cell model. Nine markers distributed on seven different chromosomes revealed MSI both in C51 and C81 stages (Table 1). Two markers (D1S1183 and D1S1612) showing MSI were located on chromosome 1 and two markers (D5S107 and APC II) on chromosome 5. One marker was located on each chromosome 4 (D4S2431), 11 (D11S1392), 17 (D17S1301), 18 (DCC), and 19 (D19S246). Additionally, 23 markers spread on 15 different chromosomes showed MSI only in the late C81 stage cells (Table 1). Chromosome 1 (D1S1597 and D1S534), 4 (D4S1625 and D4S2384), 5 (D5S816 and D5S1501), 8 (D8S1110 and D8S373), 10 (D10S667 and D10S1426), 12 (D12S1042 and D12S1064), and 17 (D17S1290, D17S1293, and TP53) revealed multiple markers for MSI in different regions of the same chromosome related to the late stage. Only one marker exhibited MSI each on chromosomes 2, 3, 6, 9, 11, 16, 18, and 19 in the later stage. Microsatellite DNA markers showing MSI in androgen-unresponsive C81 late-stage cells may be related to the acquisition of androgen-unresponsive growth as well as increased tumorigenicity.

Some of the microsatellite DNA markers showed AL on chromosomes 1, 3, 5, 8, 9, and 13 in C51 and C81 stage cells (Table 2). Representative autoradiographs for the AL have been shown in Fig. 4 C. AL on 3pter-qter (D3S2460) correlated well with the findings of the cytogenetic analysis showing rearrangement on chromosome 3q and interstitial deletion on 3p. Two markers (D1S1660 and D1S2757) on 1q showed AL in C51 and the late C81 stage cells that was not observed under cytogenetic analyses. Possibly this chromosomal region is too small to be detected by conventional cytogenetic studies. One marker showed a loss of allele each on 5p (D5S1457), 8q (D8S1132), 9q (D9S938), and 2 markers on 13q (D13S317 and D13S796) in C51 and C81 stage cells. Other tested markers within the same chromosomal regions (D8S383, D8S1047, D9S272, D9S173, D9S271, D13S1255, and D13S264) did not reveal AL, leading to the assumption that these regions with AL may be very small and covered only by one marker. Such loss of chromosomal regions during the progression of the LNCaP cell model suggests that these regions may be critical to understand the genetic basis of the human prostate cancer.

To additionally extend our observations from this LNCaP cell model, we analyzed the genetic changes in 15 clinical archival specimens that had a Gleason’s score between 5 and 7. Table 3 showed the clinical features and the genetic instability of the prostate cancer cases. We observed MSI in 11 of 15 cases (73%) by using six polymorphic DNA markers. At chromosomes 1, 5, and 19, 20%, 23%, and 26.6% cases, respectively, showed MSI using at least one DNA marker. At chromosome 17, one marker (D17S1290) revealed MSI in 40% cases, whereas another DNA marker at the TP53 locus showed MSI in only 6% cases. No patient showed MSI on chromosome 18 at the DCC locus. The statistical analysis for MSI and clinicopathological parameters (i.e., tumor stage and Gleason’s grade) did not reveal any significant correlation between MSI and grade or stage of prostate cancer.

We observed a total of 24 AL in the same group of 15 clinical samples (Table 4). A 40% AL was found on 1q, 20% on 3pter-3qter, 26% on 5p and 8q, and 33% on 13q. Of the samples, 26% (4/15) showed AL for 1 locus, 40% (6/15) for 2 loci, and 13% (2/15) for 3 chromosomal loci. Statistical analysis showed a positive correlation (r = 0.50) for AL with a Gleason’s score suggesting that these chromosomal regions might be of special importance, harboring some TSG(s).

The genetic mechanisms associated with prostate tumor progression to its androgen-unresponsive stage are currently an issue of major investigation because of an increased number of patients and deaths every year. The main objective of the present study was to investigate the genetic alterations associated with the increased tumorigenicity as well as with the acquisition of androgen-unresponsiveness by using an in vitro LNCaP cell model. Early stage (C33) cells showed androgen-responsiveness and slow tumor growth compared with late stage (C81) cells when implanted s.c. into the nude mice. Gleave et al. (24) also observed that the early stage of LNCaP cells did not make tumor without Matrigel when implanted s.c. in the nude mice. We observed a higher tumorigenicity of the late C81 stage cells in both sexes of nude mice. These results suggest that the cells at the later stage become more aggressive and are unresponsive to androgen. Therefore, we used the early C33 stage cells as a control and defined the genetic alterations related to the tumor cell progression and to the acquisition of androgen-unresponsive phenotype.

In the majority of the investigated prostate tumors, especially during the early stages of carcinogenesis, cytogenetic analyses were characterized by a low frequency of aberrant karyotypes (25, 26, 27, 28). The originally established LNCaP cell line showed an aneuploidy (88 chromosomes) with multiple copies of several marker chromosomes (19). In the present study, a comparison of the C33 and C81 stage cells of the LNCaP cell model with the original report in LNCaP cell line did not show major differences in their chromosomal patterns. The observed differences between C33 and C81 stage cells (such as structural rearrangements on chromosome 3 and a gain of one copy of Y chromosome in C81 stage cells) may be attributable to increased population and cell aging in the late stage cells or may be because of the preexisting subpopulations of LNCaP cells containing genetic alterations. The gain of the Y chromosome is a less frequent event than the loss (13, 25, 27, 28, 29). However, the gain of the Y chromosome together with aneusomy has been associated with an increased mortality rate (30). Therefore, gain of one copy of the Y chromosome and structural abnormality on chromosome 3 in the late-stage cells (C81) seems to be remarkably associated with the advanced stage of prostate cancer.

To investigate the MSI in early stage cells of the LNCaP cell model, we used BAT-25 and BAT-26 markers. A National Cancer Institute Workshop organized on Microsatellite Instability recommended using these markers for investigating MSI (23). Furthermore, recent studies have demonstrated that deletions within the mononucleotide markers are sufficient to give an idea about the MSI in most of the tumors (31, 32, 33). The presence or absence of MSI was detected based on altered mobility of PCR amplified bands by both BAT-25 and BAT-26 markers. On the other hand, a study by Pyatt et al. (34) demonstrated the allelic size variations at both the quasimonomorphic loci BAT-25 (18.4%) and BAT-26 (12.6%). Overall, 2.9% variability was revealed for the amplified PCR products at both the loci. This may be the reason that we could detect polymorphism in DU145 for BAT-25 and BAT-26 markers. However, the sublines from the LNCaP cell model as well as PC3 and BPH samples did not show any polymorphism among the BAT-25 and BAT-26 markers. Therefore, we can conclude, based on the observed altered mobility of bands in the sublines of the LNCaP cell model compared with BPH samples, that the parental LNCaP-C33 stage cells may contain MSI before going to the later stages. These results on MSI in LNCaP cells are supported by a known mutation in the hMSH2 gene in LNCaP cells (10). Additionally, a down-regulation of MMR enzyme activity as well as a lower expression of MMR-related genes has also been reported in LNCaP cells (9).

We observed that a higher percentage (22.4%) of microsatellite DNA markers exhibited MSI in C81 late stage than in C51 mid stage (6.3%) of the LNCaP cell model. This showed that C81 androgen-unresponsive cells are clonally selected from C33 androgen-responsive cells containing the preexisting cells of a mutator phenotype (Fig. 3), i.e., mutations in the MMR genes causing the higher MSI in C81 stage cells. Such a progressive accumulation of genetic changes might be associated with the transition from early androgen-responsive to the late androgen-unresponsive phenotype. However, the studies related to MSI and androgen-unresponsiveness with tumor progression have been reported independently, but for the first time this paper contributes to correlate the association of increased genetic alterations with decreased androgen-responsive phenotype in the LNCaP cell model. It has also been known that progression of human cancer is associated with a higher number of genetic alterations detected in tumor cells, i.e., most tumors are clonal expansion of a single cell (35).

Analyses of genetic alterations in prostate cancer have documented frequent ALs on different chromosomes (36, 37). For example, 45% AL on chromosome 3 (38), 32% on chromosome 5 (39), 1.2–74% on the short arm of chromosome 8 (40), and 50–55% on chromosome 13 (39, 41) have been reported. In the present study, the chromosomal regions 1q, 3pter-qter, 5p, 8q, and 13q showed AL of 40, 20, 20, 26, and 33%, respectively, in the clinical archival samples. Chromosome 13 is one of the most frequently altered chromosomes in cancer and contains 3 putative TSGs: RB1, BRCA2, and DBM (42, 43). Therefore, the observed AL on 13q in LNCaP-C81 cells seems to be associated with tumor progression.

Loss on chromosome 1 has been observed frequently in breast cancer (44, 45), pancreatic cancer (46), and gastric carcinoma (47). The studies of AL on chromosome 1 have not been implicated previously in prostate carcinogenesis. However, Cunningham et al. (36) and Latil et al. (48) reported 23% and 16.3% of AL on chromosome 1 in prostate cancer patients, respectively. In the present work, we observed 40% AL on the chromosome 1q32.1 region in clinical samples. In the last 5 years, three regions, i.e., 1q24–25 (HPC1); 1q42.2–43 (PCaP); and 1p36 on chromosome 1, were identified as harboring a prostate cancer susceptibility locus (49, 50, 51).

Conclusively, our LNCaP cell model with its stepwise progression toward its malignancy and the acquisition of androgen-unresponsiveness provides an opportunity to investigate genetic alterations involved in this process. Molecular genetic studies from the LNCaP cell model during the tumor clonality and evolution suggests that increased genetic alterations may be associated with the decreased androgen-responsive growth of the LNCaP cells and may harbor some unknown gene(s) likely to be involved in progression of prostate cancer. We hypothesize that these findings have implications for understanding the molecular basis of prostate cancer progression as this in vitro LNCaP cell model showed the unique characteristics (e.g., functional androgen receptors, elevated PSA level, and increased androgen-unresponsiveness) related to the in vivo conditions of the human prostate cancer.

We thank Jin King and Fen-Fen Lin for their excellent technical assistance, Dr. Prathibha Rao for the animal work, and Kristi, L. W. Berger for editorial assistance. We also acknowledge the support of NCI Cancer Center support grant (P30 CA36727) to UNMC for tumor tissues and CHTN Western Division, Case Western Reserve University, Cleveland, Ohio, for providing BPH tissues.