Effect of Genetic Variability within 8q24 on Aggressiveness Patterns at Diagnosis and Familial Status of Prostate Cancer

Purpose: Recently, two independent loci located at 8q24 that contribute to prostate cancer risk in men of European origin were identified.

Experimental Design: Using Bayesian probability network and logistic regression model, we searched for associations between 34 single-nucleotide polymorphisms (SNP) located at 8q24 and the aggressiveness patterns of prostate adenocarcinoma or familial history of cancers in 823 White Caucasian French men.

Results: Probability network according to the Markov chain algorithm separated the SNPs into two main groups: The first one was linked to the locus marked by rs6983267 and the second one was linked to the locus marked by rs1447295. When the patients were stratified according to tumor stage and prostate-specific antigen value, the association between the variant genotypes from six SNPs located in the second network and prostate cancer risk was strongest or confined to the patients from the more aggressive classes. However, the association between prostate cancer risk and the CC genotype of rs7841264, which marked the recombination hotspot at 8q24, was confined to patients with the highest Gleason score (odds ratio, 2.15; 95% confidence interval, 1.27-3.64; P = 0.004). Interestingly, the G allele of rs6983267 was associated with familial prostate cancer risk.

Conclusions: Our data further support that variability at 8q24 is associated with a high risk of aggressive prostate cancer at diagnosis and is linked with familial history of prostate cancer. These results corroborate that 8q24 SNPs must be evaluated in terms of prostate cancer aggressiveness markers to optimize early diagnosis procedures and management of the disease.

Prostate cancer has killed 9,789 men in France in 2002.⁸

Its specific mortality is directly related to aggressive tumor patterns and age at diagnosis. A positive family history of cancer, especially prostate cancer (1), and ethnic background are the main established risk factors. The widespread use of prostate-specific antigen (PSA) testing to screen for prostate carcinoma has led to significant overdiagnosis due to the frequent detection of indolent malignancies, but on the other hand, ∼20% of aggressive prostate cancers (Gleason score >7) with a PSA <4 ng/mL are frequently missed at early diagnosis (2).

Independent studies have previously shown that a single-nucleotide polymorphism (SNP) located at 8q24, rs1447295, is associated with prostate cancer risk (3, 4). Two studies found that this association was stronger for aggressive prostate cancer (5, 6), whereas two other studies reported an increased risk for prostate cancer in Caucasian men, regardless of tumor stage or age at diagnosis (7, 8). Our goal was to determine whether 8q24 polymorphisms could be used to predict the aggressiveness of prostate cancer risk at diagnosis. Our cohort was nested in the Cancer Genetic Markers of Susceptibility project, which recently reported the presence of at least two independent loci within 8q24 that contributed to prostate cancer in men of European ancestry (9). Particularly, a new locus marked by rs6983267 was pointed out because it was associated with increased prostate cancer risk, with an attributable risk higher than the locus marked by rs1447295. In this study, we have analyzed the relationship between 8q24 variants and the clinical aggressiveness patterns of cancers and familial status.

Study population. Information on age; aggressiveness patterns of prostate adenocarcinoma at diagnosis (Gleason score, PSA, and stage); and familial history of prostate, breast, or other cancer were obtained from 343 individuals with confirmed prostate cancers (Table 1). All of the participants were Caucasian, were clinically detected, and have tumors with differentiation Gleason grade at 4 or 5. Their mean age at diagnosis was 66 y (range, 41-88 y). The median PSA at diagnosis was 18.3 ng/mL (range, 0.4-1,500 ng/mL). The distribution of clinical features of aggressiveness is reported in Table 1. Gleason scores were ranged into the following three classes: low 7 when the predominant grade was 3 (Gleason score 7 = 3 + 4); high 7 when the predominant grade was 4 (Gleason score 7 = 4 + 3); and 8-10, which included tumors with only grade 4 or 5. Clinical stage of the tumors was classified as confined disease (T₁ or T₂, N₀, M₀) or advanced (T₃ or T₄ or N₁ or M₁ according to the International Union against Cancer 2004 classification). Familial history of cancer was considered positive if at least one first-degree relative has a prostate, breast, or other cancer. The cases were compared with 480 control men (normal digital rectal examination and PSA <4 ng/μL) from the same geographic origin as previously described (9). Written informed consent for this study was obtained from all participants. The study was approved by the Paris-St Louis Internal Review Board.

Genotyping. Genotyping of the selected 34 SNPs located between rs10956365 and rs7824074 (between 128,473,068 and 128,658,004) was done as already described in ref. 9.

Statistical analysis. For statistical analysis, the age at diagnosis, the aggressiveness patterns, and familial history of cancer were defined according to Table 1. Using the probabilistic learning in Bayesian networks according to the Markov chain model, we searched for a probabilistic profile and the relative importance of associations between genotypes and aggressiveness patterns (BayesiaLab 4.2, Bayesia SA). The SNPs harboring the best association according to Kullback-Leibler distance in the probabilistic model were selected for classic statistical analyses. Odds ratios (OR), as a measure of relative risk, and 95% confidence intervals (95% CI) were estimated using logistic regression models. All statistical analyses were done using Statview software version 5.0 (SPSS, Inc.).

When we analyzed the genotyping results using probability analysis according to the Markov chain model, we found two main networks of SNPs that correlated separately (Fig. 1). The first one included all markers located between rs10956365 and rs7841264, and the second one with all SNPs located between rs1447293 and rs6991990. Only the three most telomeric SNPs (rs4407842, rs6415495, and rs7824074) were not associated with those networks. Then, Bayesian probabilistic analysis was used to select the SNPs that had the best Kullback-Leibler distance with the more aggressive prostate cancer classes (PSA value >20, advanced tumor stage, or Gleason score >7). In this way, eight markers were selected for further analyses using the regression logistic model: one from the first network (rs7841264) and seven from the second one (rs1447295, rs4242382, rs4242384, rs7824776, rs9656967, rs7837688, and rs6999921). When the patients were stratified according to tumoral stage, the association between the variant genotypes from four SNPs located in the second network and prostate cancer risk was strongest among cases with advanced stage [AA + AC pooled genotypes from rs1447295: OR, 1.85 (P = 0.02); AA + AG from rs4242382: OR, 2.31 (P = 0.006); CC + AC from rs4242384: OR, 2.31 (P = 0.005); and TT + GT from rs7837688: OR, 1.94 (P = 0.01); Table 2]. Again, when we stratified the patients according to PSA classes, we observed that the association between prostate cancer risk and the variant genotypes from one of those polymorphisms (rs4242384) and from another polymorphism located in the second network (rs7824776) was strongest among cases with a PSA value >20 (Table 3). Moreover, for four other polymorphisms located in the same network (rs1447295, rs4242382, rs7837688, and rs9656967), the association was confined to patients with a PSA value >20 [AA + AC pooled genotypes from rs1447295: OR, 2.44 (P = 0.0003); AA + AG from rs4242382: OR, 2.71 (P = 0.0005); TT + GT from rs7837688: OR, 2.13 (P = 0.003); and TT + AT from rs9656967: OR, 1.79 (P = 0.02); Table 3]. Finally, stratification of the patients by Gleason score showed that the association between the CC genotype of rs7841264, which is located in the first network but marks the recombination hotspot at 8q24, and prostate cancer risk was restricted to patients with a Gleason score >7 (OR, 2.15; 95% CI, 1.27-3.64; P = 0.004; Table 4).

The same strategy using Bayesian probabilistic analysis was applied to identify SNPs associated with familial history of prostate, breast, or other cancers. Using the regression logistic model, we found that the G allele of rs6983267 was associated with familial prostate cancer risk (OR, 2.14; 95% CI, 1.25-3.68; P = 0.006; Table 5). No association between variants of the 8q24 markers and familial history of breast or other cancers was observed.

No significant association between any of those polymorphisms and the risk of prostate cancer was found when patients were stratified according to age at onset.

This study confirmed on a French Caucasian men population that the 8q24 region is a marker for the aggressiveness potential of a prostate carcinoma at diagnosis. The frequency of the A allele of rs1447295 in our population of prostate cancer (21%; 14% in control population) was similar to the one reported for a Caucasian population (7). Prevalence of the A allele at rs1447295 ranged from 8% to 17% in White controls across all studies (3, 6, 7) and reached 31% in African American controls in both the Flint Men's Health Study (3) and the Multiethnic Cohort Study (7). In our study, the ORs for prostate cancer risk did not differ significantly by age at onset. This result agrees with previous reports, which have shown that difference in risk by age at diagnosis was not present among Caucasians (7) and in the Australian population (8).

We observed that the association between variants from four SNPs linked to the locus marked by rs1447295 and prostate cancer risk was confined to patients with a PSA value >20 ng/mL. For two other SNPs (rs4242384 and rs7824776), the variant genotypes showed the strongest association for risk of prostate cancer in the same group of patients. Similarly, variants from four of those SNPs (rs1447295, rs4242382, rs4242384, and rs7837688) showed the strongest association for prostate cancer risk in patients possessing an advanced tumor stage. At this moment, no other study has analyzed the association between those polymorphisms and the risk of aggressive prostate cancer stratified by PSA value, but our results, showing the strongest association for patients with advanced tumor stage, are consistent with the findings obtained on a Caucasian population by Schumacher et al. (7). In this report, a greater association between rs1447295 and prostate cancer risk was observed in more advanced cases. However, they differed from the report of Severi et al. (8) conducted on the Australian population, which found that the risk did not differ by tumor stage. When we compared patients with moderate and aggressive prostate cancer, no statistically significant difference was observed among cases with advanced or localized tumor stage, but the variant genotype from rs1447295 was associated with a higher risk of PSA >20 at diagnosis (OR, 1.96; 95% CI, 1.14-3.37; P = 0.016). This might suggest that part of the insignificant comparisons simply resemble the association between variants and prostate cancer in general.

If results obtained by stratifying the patients by PSA value and tumor stage are quite similar, interestingly, stratification by Gleason score showed that the association between the CC genotype from the rs7841264 marker, which is located in the first network but points out the recombination hotspot at 8q24, and prostate cancer risk was confined to patients with a higher Gleason score (>7). When we directly compared patients with moderate versus aggressive disease, this genotype was associated with a higher risk of Gleason score >7 at diagnosis (OR, 2.24; 95% CI, 1.26-3.98; P = 0.006). Until now, no other study analyzed the association between this SNP and prostate cancer aggressiveness. On the contrary, an association between prostate cancer risk and the A allele from rs1447295 in men with a higher Gleason score was reported by several authors (3, 5–7), but others did not find any difference (4, 8, 10). These contradictory results might result from differences in the studied population or from differences in Gleason score cutoff. Further larger studies will be needed to confirm these associations.

Molecular somatic alterations in prostate cancer showed that the 8q24 chromosomal band is commonly gained in prostate carcinoma (11). The c-MYC gene, a well-known regulator of cell proliferation and programmed cell death, maps to this region and is overrepresented in prostate carcinoma with metastases (12). Moreover, the frequency of overrepresentation of the 8q24 locus in fluorescence in situ hybridization studies increases from prostate intraepithelial neoplasia to invasive primary carcinoma (13) and was associated with poor survival (14). However, no association between SNPs in the MYC gene (263 kb from rs1447295 in the telomeric direction) and prostate cancer risk was observed in our population (9). These results confirm additional work done by deCODE genetics on variation in c-MYC (3, 15).

Our results also show that rs6983267 was associated with familial history of prostate cancer but not of breast or other cancers. This result can be brought closer to the initial report of Amundadottir et al. (3) who first identified a prostate cancer susceptibility locus at 8q24 (marker DG8S737) in a genome-wide linkage scan conducted in a large Icelandic family. Wang et al. investigated the DG8S737 microsatellite repeat and the rs1447295 SNP in a large population of Caucasian men with prostate cancer enriched for hereditary disease according to Carter's criteria and in population-based controls. They found that the A allele of rs1447295 and the −8 allele from DG8S737 were significantly more common in familial prostate cancer than in controls, whereas those associations were not observed in sporadic cases (6). Recently, Christensen et al. (16) showed that a subset of Utah high-risk pedigrees with a familial history of prostate cancer showed nominal linkage evidence on chromosome 8q (heterogeneity LOD score, 1.67). Our results, which included a familial prostate cancer history but was not restricted to a high-penetrance model to a major gene (Carter's criteria), could be compatible with our previous observation of a brother-brother dependence inheritance (17) and the transmission of a susceptibility locus with Mendelian recessive inheritance as recently suggested by Pakkanen et al. (18).

In summary, this report confirms evidence of an association between 8q24 common variants and aggressive patterns at diagnosis of prostate cancer. The knowledge of 8q24 variant genotypes may provide important information for early diagnosis procedure and screening strategies for prostate cancer. With other low-penetrance susceptibility genes, 8q24 SNPs could provide adjunctive tests based on selected polymorphisms to increase the positive predictive value of a usual test, such as PSA screening, for early diagnosis of aggressive forms of prostate cancers.

No potential conflicts of interest were disclosed.

We thank the men who volunteered to participate in the PROGENE study; the researchers and core genotyping facilities from the prostate cancer program of the National Cancer Institute Cancer Genetic Markers of Susceptibility Study Project (http://cgems.cancer.gov/data) for providing the genotyping data; and in particular, Gilles Thomas and Stephen J Chanock from the Division of Cancer Epidemiology and Genetics (National Cancer Institute, NIH, Department of Health and Human Services, Bethesda, MD) for their help.