Loss of heterozygosity on chromosome 9q and p53 alterations in human bladder cancer†
All patients involved in this study provided written informed consent under the protocols approved by the appropriate institutional review boards.
Somatic loss of the 9q allele as well as alteration of the tumor suppressor p53 occurs commonly in bladder cancers. Although alteration of p53 has been strongly associated with invasive stage disease, the prognostic significance of 9q loss of heterozygosity (LOH) and the relations between these alterations are less well defined.
The 9q LOH was examined at five microsatellites and p53 alterations (mutation and persistent immunohistochemical staining) in a population-based case series of 271 newly diagnosed bladder cancer patients. Loss of heterozygosity was scored quantitatively and p53 mutation completed using single-strand conformation polymorphism screening followed by sequencing.
Overall, allelic loss at 9q was detected in 74.5% (202/271) of cases and allele loss was associated with invasive disease (P < 0.05). Although based on small numbers, all nine in situ lesions contained 9q LOH. Age, gender, and smoking were not significantly associated with chromosome 9q allele loss. Both intense persistent p53 staining and LOH at 9q were independently associated with invasive disease (P < 10−14 and P < 0.05, respectively).
These data, using a population-based sample, suggest a relation between 9q LOH and invasive stage bladder cancer, and thereby suggests that a tumor suppressor gene at this loci, in addition to p53, may be important in the development of this more aggressive form of the disease. Cancer 2005. © 2005 American Cancer Society.
In the U.S., approximately 60,000 new cases of bladder cancer occurred in 2004.1 Carcinoma of the urinary bladder tends to present as a superficial, lower-stage and lower-grade tumor at the time of diagnosis.2 These noninvasive tumors can be removed by transurethral resection and tend to have a significantly higher survival rate.3 Most early-stage bladder cancers remain superficial and noninvasive for extended periods. However, after surgical removal some patients develop one or more recurrences and there is a persistent risk that these tumors could invade the muscular walls of the bladder and become life threatening.4, 5 Developing molecular markers that define these at-risk lesions is clinically important for prognosis and improved treatment.
Alterations of the p53 gene are some of the most frequently documented somatic alterations in human bladder cancer, as well as other tumor types, and detection of this altered gene holds prognostic significance. Several studies, including our own, have clearly shown that inactivation of the p53 pathway is prevalent in this disease, and that its aberrant immunohistochemical staining correlates with the degree of invasiveness of the tumor.6, 7
Allelic loss on chromosome 9q is also a frequent event in bladder carcinogenesis, with multiple regions of deletion identified, including regions containing the candidate tumor suppressor genes, PTCH, FACC, and TSC1.8–13 The clinical correlates associated with 9q loss of heterozygosity (LOH), however, are less well defined, particularly in large, population-based studies, where the selection biases encountered in hospital-based studies or studies of particular tumor stages are reduced. Allelic loss at 9q has been reported as an early lesion in the development of bladder cancer,14 but has also been associated with disease recurrence in superficial bladder tumors.15 Other studies have not seen these associations and describe LOH at 9q as random and oligoclonal.16–18
It is clear that alterations of chromosome 9q are frequent in bladder cancer, but the relation of these alterations to clinical correlates of the disease, to inactivation of the p53 pathway, or to carcinogenic exposures, important to the risk of bladder cancer, have not been well established. Therefore, we identified the prevalence of LOH at several sites on chromosome 9q and examined their relation to alterations of p53, as well as clinical and demographic factors of the patients in a population-based case series study of bladder cancer to better clarify the role of 9q in disease occurrence and progression.
MATERIALS AND METHODS
Residents of New Hampshire, ages 25–74 years, diagnosed with incident bladder cancer from July 1, 1994 to June 30, 1998 were identified by a rapid reporting system of the New Hampshire State Cancer Registry.19 Only those subjects with incident disease were included in the study, and all participants provided written consent under the appropriate institutional review board protocols. Study participants completed an extensive interview to obtain information on demographic traits and carcinogen exposures. Pathology reports and paraffin-embedded tumor specimens were requested from the treating physician / pathology laboratories. Bladder tumors were reviewed by one pathologist and classified according to the 1973 and 2004 WHO guidelines for the classification of bladder tumors.
Both bladder tumor tissue and circulating peripheral blood samples were obtained for 271 cases. Tumors of mixed histology (2), tumors from nonwhite individuals (10), and samples (3) that were not pathologically confirmed as cancer were excluded from analysis.
We have previously reported immunohistochemical staining in these tumors.6 Immunohistochemical staining of paraffin-embedded slides was performed using the avidin-biotin complex technique. For each case, a single representative slide was selected for staining and histologic evaluation. Briefly, slides were deparaffinized and hydrated into water. Slides underwent antigen retrieval in Citra solution using the Biocare Decloaking Chamber (Biocare Medical, Walnut Creek, CA). Staining of p53 was performed using a monoclonal antibody, clone DO7, against wildtype p53 (BioGenex, San Ramon, CA) at a 1:100 dilution on the Optimax I-6000 Immunostainer (BioGenex). An appropriate positive control was used in each staining run and each slide was stained with a negative control. The percent of positively staining tumor cells was scored (negative, 1–9%, 10–49%, and =50%), and the intensity of nuclear staining was graded on a semiquantitative scale (0–3), rating intensity in the dominant pattern within the tumor.
DNA Extraction and p53 Mutation Analysis
DNA was extracted from paraffin-embedded tumor samples as previously described.6 Single-stranded conformational polymorphism (SSCP) analysis of TP53 exons 5–9 was performed on all bladder tumor DNA samples. Each sample was individually rated for the percentage of the sample on the slide/block that was tumor; on average, 70% of each of the specimens was tumorous. Exons were amplified by polymerase chain reaction (PCR) containing fluorescently labeled primers. Previously reported primer sequences for each exon were used.8 One microliter of PCR product and 1 μL of size standard TAMRA-350 (Applied Biosystems, Foster City, CA) were denatured in 4 μL of formamide/blue-dextran denaturing buffer at 95 °C for 5 minutes and then loaded onto MDE gels. Gel electrophoresis was carried out on a DNA sequencer (ABI Prism 377, Applied Biosystems) with an external cooling system (Thermo NESLAB Portsmouth, NH) attached for gels run at 25 °C. Genescan 3.1 software (Applied Biosystems) was used for fragment analysis. Samples with variant SSCP bands were reamplified and purified using Centri-Sep columns (Princeton Separations, Adelphia, NJ). These were then directly sequenced in both directions by a DNA autosequencer (ABI Prism 377) using the Big Dye Terminator v3.0 sequencing kit (Applied Biosystems) according to the manufacturer's instructions. The data were analyzed with Sequencing Analysis 3.3 software (Applied Biosystems) and Sequencher 4.1 software (Gene Codes, Ann Arbor, MI).
Microsatellite analysis was performed on 299 tumors using five microsatellites markers (D9S12, D9S149, D9S176, D9S302, and GSN). Primer pairs for each locus, with the forward primer labeled with the fluorescent dyes HEX or FAM, were obtained from Invitrogen (Carlsbad, CA). Each PCR mixture contained 0.5 μL of template DNA, 0.33 μM of each primer, 0.2 mM dNTPs, 1.5 μL of 10× reaction buffer, and 0.12 units of AmpliTaq Gold DNA polymerase (Applied Biosystems) in a final solution of 15 μL. PCR amplification was performed as follows: Initial denaturation step at 95 °C for 10 minutes and then 40 cycles of 96 °C for 10 seconds, annealing at 55–64 °C (depending on primer set) for 30 seconds and extension at 70 °C for 3 minutes, with a final extension at 70 °C for 30 minutes. The loading mix was prepared with 2.5 μL of deionized formamide, 0.5 μL internal size standard (Genescan-350 TAMRA) and 0.25 μL loading buffer (blue dextran, 50 mg/mL and EDTA, 25 mM). Amplified PCR products (1.0 μL) and 3.0 μL of the loading mix were denatured at 95 °C for 5 minutes and then immediately chilled on ice. One microliter of mixture was loaded on 5% Long Ranger / 6M urea gel and electrophoresis was carried out for 2 hours in the ABI PRISM 377 (Applied Biosystems). The collected data were then analyzed with the Genescan 3.1 software (Applied Biosystems).
For each locus, LOH was scored using the Genescan 3.1 software (Applied Biosystems). The allelic ratio (AR) was derived as the ratio of the peak heights in the blood-derived samples divided by the peak heights of the alleles from the tumor. Ratio values of <0.5 or >1.5 were taken to be indicative of LOH.
The association of LOH with patient demographics, carcinogen exposures, tumor traits, and p53 status was first assessed using individual unconditional logistic regression analyses controlling for age and sex. Loss of heterozygosity at individual loci and its association with tumor stage was examined with chi-square analysis. To examine the effect of multiple predictors on tumor stage (invasive vs. noninvasive), multivariate unconditional logistic regression was again used. All P values represent two-sided statistical tests with statistical significance at P < 0.05.
The prevalence of LOH at any loci on chromosome 9q was 74.5% (202/271). For each of the five individual markers, LOH ranged from 31.5–50.0% of the patients (D9S12: 31.5%, D9S149: 49.1%, D9S176: 37.9%, D9S302: 50.0%, and GSN: 47.8%). There was no statistically significant difference in LOH prevalence by gender or age, although more LOH was observed among those diagnosed at an age older than 55 years old (Table 1). The prevalence of LOH did not differ markedly by any measure of tobacco use. Other important exposures associated with bladder cancer, including hazardous occupational exposures, inorganic arsenic, or use of hair dye, also showed no significant relation to LOH (data not shown). Loss of heterozygosity on 9q was significantly associated with tumor invasiveness; 83% (63/76) of the invasive tumors had LOH at any marker in 9q, compared with 70% (130/186) of the noninvasive tumors (P < 0.03, Table 1). We sought to more precisely define specific loci associated with the invasive phenotype; thus, we examined the association of LOH at each individual with tumor invasiveness, excluding carcinoma-in-situ, which is thought to have a biologically distinct phenotype. Loss of heterozygosity at the D9S149 marker was the only site showing significantly greater prevalence in invasive tumors (62%, 30/49 in invasive tumors compared with 42%, 49/113 in noninvasive tumors, P = 0.02; Table 2).
Table 1. Selected Patient and Tumor Characteristics with Any LOH in 5 Markers at Chromosome 9q
|Sex|| || || || || || |
| Female||45||76||14||24||1.0 (ref)|| |
| Male||157||74||55||26||1.22 (0.61–2.42)||0.6|
|Age, yrs|| || || || || || |
| ≤ 55||34||65||18||35||1.0 (ref)|| |
| > 55||168||77||51||23||1.79 (0.93–3.46)||0.08|
|Smoking status at entry|| || || || || || |
| Never smoker||32||70||14||30||1.0 (ref)|| |
| Former smoker||116||79||31||21||1.59 (0.73–3.43)||0.2|
| Current smoker|| || || || || || |
| 1–20 cigarettes/d||31||67||15||33||0.85 (0.35–2.08)||0.7|
| > 20 cigarettes/d||23||72||9||28||1.12 (0.41–3.04)||0.8|
|Duration of smoking, yrs|| || || || || || |
| Never smoked||32||70||14||30||1.0 (ref)|| |
| ≤ 20||35||74||12||26||1.29 (0.51–3.24)||0.6|
| 21–30||33||79||9||21||1.56 (0.59–4.16)||0.4|
| > 30||101||76||32||24||1.29 (0.59–2.78)||0.5|
|Age first smoked|| || || || || || |
|Never smoked||32||70||14||30||1.0 (ref)|| |
|Teenage (< 18)||120||77||35||23||1.46 (0.69–3.10)||0.3|
|Adult (≥ 18)||49||71||20||29||0.93 (0.40–2.18)||0.9|
|Stage|| || || || || || |
|Carcinoma in situ||9||100||0||0||not estimable|| |
|Noninvasive (low and high grade)||130||70||56||30||1.0 (ref)|| |
| Invasive||63||83||13||17||2.08 (1.05–4.11)||0.03|
|P53 mutation|| || || || || || |
|No||171||73||64||27||1.0 (ref)|| |
| P53 staining percent|| || || || || || |
| < 10%||43||90||5||10||1.0 (ref)|| |
| = 10%||158||71||64||29||0.29 (0.11–0.76)||0.01|
|P53 intensity 3+|| || || || || || |
|No||144||72||55||28||1.0 (ref)|| |
Table 2. Associations between LOH at Individual Loci on Chromosome 9q and Tumor Invasiveness
|D9S149|| || || |
| Noninvasive||45 (42)||62 (58)|| |
| Invasive||30 (62)||18 (38)||0.02|
|D9S12|| || || |
| Noninvasive||40 (31)||87 (69)|| |
| Invasive||15 (33)||30 (67)||0.8|
|D9S176|| || || |
| Noninvasive||49 (36)||87 (64)|| |
| Invasive||23 (41)||33 (59)||0.5|
|D9S302|| || || |
| Noninvasive||80 (47)||89 (53)|| |
| Invasive||41 (56)||32 (44)||0.2|
|GSN|| || || |
| Noninvasive||57 (50)||57 (50)|| |
| Invasive||15 (39)||23 (61)||0.3|
Both p53 mutation and p53 immunohistochemistry (IHC) staining intensity showed a positive, although not statistically significant, association with any LOH at 9q. Conversely, LOH was inversely associated with the percentage of tumor cells staining for p53 (>10% stained vs. ≤10%; odds ratio [OR], 0.29, 95% confidence interval [CI], 0.11–0.76).
In univariate analysis, both LOH at 9q and IHC staining intensity of p53 were associated with invasive staged disease. Therefore, we examined the simultaneous effect of these alterations on tumor invasiveness, using tumor invasiveness as the dependent variable in the model. As shown in Table 3, any LOH at 9q and p53 staining intensity were independently associated with invasive disease. As suggested by univariate analysis and our previous report on measures of p53 alterations in these tumors,6 the percentage of cells staining for p53 was not significantly associated with invasive stage disease in a model controlling for age, sex, and 9q LOH status (data not shown). Performing the same modeling with each individual loci revealed that this association, as expected, was driven by LOH at D9S149, which had an OR for invasive disease (controlled for p53 staining intensity, age, and sex) of 2.11 (95% CI, 0.93–4.78, Table 3).
Table 3. LOH at 9q and p53 1HC Staining Intensity are Independently Associated with Invasive Stage Bladder
|Model 1 (n = 261)||Any LOH at 9q|| || || |
| || No||69||1.0 (ref)|| |
| || Yes||192||2.27 (1.01–5.10)||0.05|
| ||P53 staining intensity|| || || |
| || < 3||198||1.0 (ref)|| |
| || 3 +||63||16.37 (8.08–33.18)||<10−14|
|Model 2 (n = 154)||LOH at D9S149|| || || |
| || No||80||1.0 (ref)|| |
| || Yes||74||2.11 (0.93–4.78)||0.08|
| ||P53 staining intensity|| || || |
| || < 3||110||1.0 (ref)|| |
| || 3 +||44||10.65 (4.62–24.52)||<10−16|
Using a population-based series of incident tumors, thus reducing the bias that would be observed in hospital-based designs, we observed the prevalence of LOH at any loci on chromosome 9q to be approximately 75%. This is consistent with a previous hospital-based study of bladder cancer in the United Kingdom that observed a similar prevalence.20
It has been suggested that 9q loss is an early event in bladder cancer development.14, 21 Our data are consistent with this observation in several ways. First, all in situ carcinomas contained 9q LOH, albeit with small numbers. Second, in our multivariate model, 9q LOH related to a 2-fold relative risk of invasive disease; however, not as strongly as p53 alteration (i.e., p53 IHC staining intensity), which was associated with a greater than 16-fold risk. Furthermore, we observed an inverse association between the percentage of cells staining for persistent p53 and LOH at 9q; this measure of p53 inactivation has been shown to occur later in disease progression.7, 22, 23 However, neither mutation of TP53 nor intensity of p53 staining was associated with LOH at 9q. These observations are consistent with our previous observation that these different measures of p53 inactivation, i.e., IHC staining intensity, percent of stained cells, and genetic mutation, may represent distinct phenotypes of the disease, as there is not complete concordance between these alterations.6 In the absence of prospective monitoring, it is difficult to determine the precise sequence of the molecular-genetic event. Thus, loss at the 9q site may reflect the genetic instability that is often associated with p53 inactivation, and that may contribute to the pathway to invasive disease of bladder cancer development.24 In addition, careful study of these alterations in premalignant lesions of the bladder as well as in adjacent nontumorous bladder tissue may help to clarify if loss at this region is clonal and selected for in development of invasive disease, or if it is merely a consequence of the genetic instability associated with this aggressive phenotype. This would have obvious clinical as well as therapeutic implications.
Contrary to some previous findings, our results suggest that specific areas, particularly the area near microsatellite marker D9S149, may be targeted for loss in clones that eventually present as invasive disease. Further characterization of this region may be necessary to understand how loss of genetic material around D9S149 may be involved in the presentation of invasive disease. A candidate tumor suppressor gene located within several hundred kilobases of this marker is the tuberous sclerosis complex 1 (TSC1) gene, a putative tumor suppressor with observed mutation in bladder cancer.25–27 It is clearly possible that bladder tumors with LOH in 9q34 might result in loss of the TSC1 gene, which may play a mechanistic role in determining the invasive phenotype of these tumors. Additional investigation into the inactivation of this gene is warranted for bladder cancer.
There are few studies examining the relation between etiologic factors of bladder cancer and 9q LOH. Unlike the findings of Simoneau et al.,15 derived from a series of low-grade tumors, we did not observe any gender difference in the occurrence of LOH at 9q, even in analyses stratified by tumor stage. This disparity may result from uncontrolled confounding in the Simoneau et al. study, which did not use a multivariate model in describing predictors of LOH at 9q. Furthermore, unlike the case for lung cancer, where LOH at numerous loci have been associated with smoking,28, 29 we did not find any measure of smoking to be associated with LOH at 9q. This may reflect a biological difference in the type of or in the way that tobacco carcinogens interact with the lung compared with the bladder.
Our investigation confirms the prevalent role of 9q LOH using a population-based study of bladder cancer. Using this approach, we can present a more general picture of the prevalence of this alteration in this disease and a better understanding of the relation of this alteration to the etiology and characteristics of bladder cancer. The association of LOH at 9q and particularly at D9S149 with invasive stage disease, independent of p53 status, enhances the hypothesis that the TSC1 gene may represent a bladder tumor suppressor gene important in predisposing cells to the invasive phenotype. Further study is warranted to examine if this gene, or others in this region, may be targets for therapeutic intervention in this aggressive form of the disease.