Molecular Biology Reports

, Volume 35, Issue 3, pp 459–464

Association of HSP70-hom genetic variant with prostate cancer risk


    • Department of Molecular Immuno-OncologyFaculty of Medicine
  • Hamadi Saad
    • Department of UrologyEPS Fattouma Bourguiba
  • Faouzi Mosbah
    • Department of UrologyEPS Sahloul
  • Lotfi Chouchane
    • Department of Molecular Immuno-OncologyFaculty of Medicine
Original Paper

DOI: 10.1007/s11033-007-9107-1

Cite this article as:
Sfar, S., Saad, H., Mosbah, F. et al. Mol Biol Rep (2008) 35: 459. doi:10.1007/s11033-007-9107-1


Because of the importance of androgens to prostate cancer (PCa) development, several candidate genes along androgen pathway have been under intensive study. Given the role of the molecular chaperone HSP70 in the regulation of the androgen receptor (AR) transactivation function, we first chose to explore the association between the HSP70-hom functional genetic variant (+2437 T > C) and prostate cancer risk by genotyping DNA samples from 101 unselected PCa patients and 105 healthy men. There was a trend towards lower frequency of TC and CC genotypes among patients when compared with healthy controls, however the difference did not reach the statistical significance (TC genotype: OR = 0.53, = 0.05; CC genotype: OR = 0.42, = 0.16). Moreover, individuals carrying at least one C allele have a statistically significant lower susceptibility for PCa (OR = 0.51 (0.26–0.97); = 0.02). Since some factors may influence tumor progression rather than initiation, we also examined the relationship between the HSP70-hom polymorphism and the clinical characteristics of the malignancy at the time of diagnosis. The stratified analysis of the genotypes with the clinical stage and tumor grade showed that there was no significant difference in the risk estimates according to prognostic indicators of PCa disease in our population study. This is the first report on the studies of HSP70 SNPs in PCa and our data suggest that this genetic variant may be a genetic marker for PCa susceptibility in Tunisians.


Clinical outcomeGenetic polymorphismGenetic susceptibilityHSP70-homProstate cancer


Prostate cancer (PCa) is a common disease with a multifactorial and complex etiology. It is the most common male malignancy and the second leading cause of death in many western countries [1]. PCa worldwide incidence varies widely between different ethnic groups and geographical areas. It has been reported that African Americans have the highest incidence while the lowest is among Asian males. For the year 2005, its incidence was at ∼234/105 in the United States [2].

In 2003, PCa was ranked as the fourth most diagnosed cancer in Tunisia. The age-adjusted standardized incidence rate is 10/105 compared with 6,79/105 in 2000. Many factors contribute to PCa and genetic susceptibility is among the most important of these [3]. Approximately 5–10% of PCa cases and 45% of patients diagnosed under the age of 55 years may be influenced by a germ-line predisposition [4].

The growth of prostatic epithelial cells is regulated by several complexes including steroids, growth factors and cytokines [5, 6]. Androgens particularly regulate prostatic growth through an enhanced cellular proliferation rate [7, 8]. The central element of androgen signaling in prostatic cells is the androgen receptor (AR), a ligand-activated transcription factor that recognizes and binds specific response elements (ARE) of androgen responsive genes [9]. The standard therapy relies on removing the testicular androgens and blocking the androgen receptor (AR) actions resulting in slowing the disease progression. The refractory prostate cancer has been attributed to different mechanisms including mutations in the AR hormone-binding domain or amplification of the AR gene that permit transactivation of target genes with little or no steroid hormones [10, 11]. Since most cases of hormone insensitive PCa still retain a wild-type AR, alterations in the expression or function of the co-factors genes that control the AR signaling pathway could presumably play a major role in resistance to endocrine therapies [12]. Thus, the key genes implicated in the molecular mechanisms that govern AR regulatory pathways could be susceptibility candidate genes for this cancer.

In its inactive state, the AR is assembled into intracellular heteromeric complexes that contain a consortium of accessory proteins including several heat shock family molecular chaperones (HSP90, HSP70, and HSP56), co-chaperones and various types of transcription co-activators. Through their capacity to remodel the conformation of steroid hormone receptors, chaperones ensure a tight control, a dynamic range, and a rapid reversible transcriptional response. Some studies have reported that HSP70 chaperones function to coordinate BAG-1 protein, a DNA binding co-chaperone, into complexes with steroid nuclear receptors and possibly other components of the transcriptional machinery, resulting in transcriptional stimulation [1318].

Previous studies have reported that various malignant tumors overexpress HSP70, which relates closely to tumorgenesis, malignant development, tumor immunity, resistance to apoptosis and a poor prognosis in the clinical course [1922]. Furthermore, it has previously been reported that human androgen independent PCa cell line PC3 has a high expression level of HSP70 which may be involved in the development of fatal androgen-independent tumors [23, 24]. Accordingly, some findings reported the induction of apoptosis and the inhibition of proliferation by abrogation of HSP70 expression in various tumor cells, such as PC3 and LNCaP [2022, 25].

A duplicated locus of three intronless genes encoding members of the 70 kDa heat shock protein family (HSP70) is located in class III region of the human major histocompatibility complex (6p21-3), between the complement C2 and TNFα genes [26, 27]. HSP70–1 and HSP70-2 genes encode the similar heat inducible proteins, whereas HSP70-hom encodes a non heat-inducible protein that shares high homology with the protein products of HSP70-1/2. The HSP70-hom gene is located 4 Kb telomeric to the HSP70-1 gene and has no heat shock consensus sequence in its 5′ flanking sequence. A common polymorphism in the coding region of the HSP70-hom gene (T > C transition at position 2437) was described by Caroline et al. [28]. This polymorphism results in a Met (non polar, hydrophobic) to Thr (polar, neutral) amino acid substitution at position 493 which lies within the HSP70 peptide-binding domain. This alteration at amino acid 493 could be associated with variation in the peptide binding specificity of the HSP70-hom protein that alters its biological function [29]. Interestingly, whereas HSP70 proteins were ubiquitously expressed throughout the organs of mice and humans, it has been reported that the HSP70-hom homologue is found to be predominantly expressed in mouse testis that are androgen dependent tissues.

Taken together, these findings promoted us to investigate whether the functional HSP70-hom polymorphism (+2437 T > C: rs2227956) might represent a risk marker for PCa carcinogenesis and aggressiveness.

Materials and methods

Subjects and controls

A total of 206 unrelated Tunisian men including 101 PCa patients and 105 age-matched male controls were selected from the same population living in the middle coast of Tunisia. PCa patients were recruited between September 2002 and August 2004 from the two departments of Urology of Monastir and Sousse Hospitals, Tunisia. The diagnosis has been histopathologically confirmed for 97 patients. The remaining four patients were included in the study according to their radiological evidence of metastasis and their high serum prostate specific antigen (PSA) levels (these criteria ensuring the presence of advanced disease). The serum PSA values were measured in all cases before treatment. Clinical characteristics including Gleason grade, TNM stage, age at diagnosis, and family history were obtained from medical records. Clinicopathological parameters were dichotomized as follows: The pathological stage at the time of diagnosis was classified according to TNM system into the localized group (T1−T2N0M0) and the advanced group (T3−T4N0M0 and T1−T4N0-1M1/T1−T4N1M0-1). The histopathological grade was recorded as the Gleason score and was classified into two groups: a Low grade group (Gleason score <7) and a high grade group (Gleason score ≥7).

The control group consisted of 105 healthy male subjects having no evidence of any personal or family history of cancer. Information was obtained from control participants through a standardized questionnaire including age, residence, dietary habits, and family history of cancer or other diseases. Their PSA levels were within the normal limit (<4 ng/ml) and showed no signs of prostate hyperplasia or prostate carcinoma by digital rectal examination. All subjects in both groups provided informed consent to participate in the study and to allow their biological samples to be genetically analyzed. Approval for the study was given by the National Ethical Committee.

Genotype analysis

Genomic DNA was isolated from whole blood using the salting out procedure [30]. Genotyping was performed by the PCR based restriction fragment length polymorphism method (RFLP-PCR), as described previously.

The primer sequences used to amplify the HSP70-hom fragment were: hom1 5′GTAACTTAGATTCAGGTCTGG3′ and hom2 5′CCGGATCCCATAGGCTCAGAGAG ACC3′. PCR was carried out in a final reaction volume of 20 μl containing, 50–100 ng of genomic DNA, 200 μmol/L dNTPs, 2 mM MgCl2, 1 × Taq DNA polymerase buffer, 1 μmol/L of each forward and reverse primers and 1 U of Taq DNA polymerase (Amersham, Paris, France). In the thermocycling procedure, initial denaturation at 94°C for 5 min was followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and extension at 72°C for 3 min before a final extension at 72°C for 10 min. PCR products were analyzed onto 1% agarose gel with ethidium bromide.

The PCR product was digested with 5 units of NcoI (Qbiogene, Europe) enzyme in a 20 μl reaction mixture, as suggested by the manufacture for overnight at 37 °C and separated on a 2% agarose gel. The T but not the C allele has an NcoI restriction site within the 2045 bp amplified PCR product. The three possible genotypes were defined by three distinct banding patterns: TT (1495 and 550 bp fragments), CT (2045, 1495 and 550 bp fragments), and CC (2045 bp fragment).

Statistical analysis

χ2 tests for hardy Weinberg equilibrium were carried out in each group (patients and controls). The HSP70-hom genotype frequencies between the study and control group were compared using 2 × 2 tables by means of χ2 analyses or Fisher’s exact test if one or more variables in 2 × 2 tables were <5. The odds ratio (OR) and 95% confidence interval (CI) were calculated to assess the relationship between the polymorphism and PCa. < 0.05 was required for statistical significance.

SEM-STATISTQUES software (centre Jean Perrin, Clermont-Ferrand, France) was used for statistical analysis.


The present study included 101 prostate cancer patients and 105 male controls. The mean age for patients and controls was 70.0 (range: 49–91), and 65.0 (range: 42–80) years, respectively. No significant differences in the mean age were found between the PCa patients and controls (P = 0.3). No deviation from Hardy-Weinberg’s equilibrium was observed in the patient and the control groups (P > 0.05).

Genotype distribution and risk estimates are summarized in Table 1. Among healthy controls, the prevalence of the TT, TC, and CC genotypes were 61.9, 30.47 and 7.63%, respectively, whereas the TT, TC, and CC were observed in 76.2, 19.8 and 4% of patients respectively. There was a trend towards lower frequency of TC and CC genotypes among patients when compared with healthy controls, however the difference did not reach the statistical significance (TC genotype: OR = 0.53, = 0.05; CC genotype: OR = 0.42, = 0.16). Nevertheless, subjects carrying at least one C allele (TC + CC) exhibited twofold reduction in the risk of developing PCa compared to those with TT genotype (OR = 0.51; = 0.02). The C allele frequency in the control group (23%) was greater than in the patient group (14%), however no significant effects were observed (OR = 0.54; = 0.18).
Table 1

Association of HSP70-hom genotype frequencies with PCa risk


Cases (n = 101)

Controls (n = 105)

OR (95%CI)

P value



77 (76.23%)

65 (61.9%)




20 (19.8%)

32 (30.47%)

0.53 (0.26–1.06)



4 (3.97%)

8 (7.63%)

0.42 (0.1–0.64)



24 (23.77%)

40 (38.1%)

0.51 (0.26–0.97)




174 (86%)

162 (77%)




28 (14%)

48 (23%)

0.54 (0.25–1.2)


OR: Odds Ratio, CI: Confidence Interval

Regarding prognostic values, an additional subgroup analysis was performed according to the clinical parameters of the malignancy (histopathological grade and tumor stage) at the time of diagnosis. Table 2 shows the principal patient characteristics according to HSP70-hom genotypes. When low and high grade groups were compared, the Fisher’s exact test revealed that the HSP70-hom polymorphism had no significant relationship with the pathological grade of the disease (OR = 0.7, = 0.55). However, there was a remarkable decreased risk of developing advanced disease among patients carrying the CC genotype compared with those having the TT genotype, but this was not statistically significant (OR = 0.38, = 0.3). Similarly, considering the TC and CC genotypes together, no significant prognostic impact of the combined genotype (TC + CC) was found when the patients were analyzed separately with regards to tumor grade (OR = 0.7, = 0.4) and clinical stage (OR = 0.63, = 0.33). With regard to allelic frequencies, similar results were found among different pathological subgroups of patients. As shown in Table 2, we failed to find any significant association between the C allele frequency and the pathological features of PCa such as Gleason score (OR = 0.73, = 0.45) and tumor stage (OR = 0.61, = 0.23).
Table 2

Distribution of the HSP70-hom polymorphism according to clinicopathologic parameters


Histopathological grade

Clinical stage

Low grade (n = 42)

High grade (n = 55)

OR (95% CI)


Localized (n = 30)

Advanced (n = 71)

OR (95% CI)




31 (73.8%)

44 (80%)



21 (70%)

56 (78.9%)




9 (21.4%)

9 (16.4%)

0.7 (0.22–2.22)


7 (23.3%)

13 (18.3%)

0.7 (0.22–2.25)



2 (4.8%)

2 (3.6%)

0.7 (0.07–7.51)


2 (6.7%)

2 (2.8%)

0.38 (0.03–4.05)



11 (26.2%)

11 (20%)

0.7 (0.24–2.02)


9 (30%)

15 (21.1%)

0.63 (0.21–1.83)




71 (84%)

97 (88%)



49 (81%)

125 (88%)




13 (16%)

13 (12%)

0.73 (0.3–1.8)


11 (19%)

17 (12%)

0.61 (0.25–1.5)


OR: Odds Ratio, CI: Confidence Interval

P value was obtained by Fisher’s exact test


High penetrance and a combination of low penetrance susceptibility genes are likely to be involved in the PCa pathogenesis [31]. Minor susceptibility genes may play a larger role in PCa risk. These genes do not cause carcinoma, but in certain environments or in interaction with other genetic alterations, they influence disease development.

Cancer growth depends on the ratio of cells proliferating to those dying. In the prostate gland, androgens are the main regulator of this ratio by both stimulating proliferation and inhibiting apoptosis. Therefore, the genes implicated in androgen pathway could be attractive susceptibility candidate genes for this malignancy. Given the role of the molecular chaperone HSP70 in the regulation of the AR transactivation function, HSP70 polymorphisms are thought to play an important role in PCa risk and progression. Moreover, a relatively low degree of polymorphisms is reported in the two intronless genes HSP70-1 and HSP70-2, some of which correspond to silent mutations. Nevertheless, to our knowledge no studies have focused on correlating PCa risk with genetically determined high levels or affected functionality of HSP70 proteins.

To cope with this lack of knowledge, we examined the impact of the functional HSP70-hom polymorphism (T > C transition at position 2437) in the process of prostate carcinogenesis and aggressiveness. In this study, the patient-control results revealed that TC and CC genotypes were slightly but not significantly more frequent in the control group (30.47 and 7.63%, respectively) than in the at-risk study population (19.8 and 3.97%, respectively). However, no significant effect was observed, presumably because of insufficient statistical power. Alternatively, a significant association was found between the occurrence of the C allele and the protection against developing PCa (OR = 0.42; P = 0.02) in our population. On the other hand, we found that the allelic frequencies differed between patients (14%) and controls (23%) (OR = 0.54; = 0.1), but this difference did not reach any statistical significance possibly because of the small sample size or the low C allele prevalence. These findings suggest that the HSP70-hom polymorphism may be associated with a protection against PCa development.

Since some factors may influence tumor progression rather than initiation, we also examined this polymorphism in relation to PCa prognostic indicators at the time of diagnostic. While no significant impact on the pathological or clinical outcome of the malignancy was observed, there was a trend towards lower frequency of CC genotype among the advanced PCa patients compared to localized group (2.8 vs. 6.7%), but this result was not statistically significant (OR = 0.38, = 0.3), which suggests that a larger study sample may be required to observe a statistically significant effect.

As suggested by earlier reports, the neutrality of the Thr residue (C allele) may affect the efficiency of the HSP70-hom protein in acting as a molecular chaperone by lowering the strength of hydrophobic interaction between the chaperone and the target protein. In fact, in the hypothetical model of the peptide based on that of the HLA class I molecule, the 493 residue is located in the β sheets which form the floor of the HSP70-hom peptide binding site of the protein [29, 32]. In line with these findings, we suggest that this alteration may be associated with some variation in the peptide binding specificity and may affect interaction of the HSP70 molecule chaperone with hydrophobic proteins. On the basis of these data, we can hypothesize that this altered amino acid may decrease its ability to assemble the AR-chaperones complexes and to provide the correct conformation for hormone binding by the AR. Therefore, it is biologically plausible that the decreased functionality of this chaperone protein may provide better protection against developing PCa, presumably by modulating the DNA binding and the transactivation function of the AR.

To our knowledge, several studies about contribution of this HSP70-hom SNP in susceptibility of common diseases have been reported [3240]. To date, there is only one published genetic cancer study reporting a significant association of the HSP70-hom polymorphism with the increased risk of breast carcinoma in a Tunisian population [41]. The discrepancy between these studies reflects that the implications of this HSP70-hom polymorphism in carcinogenesis seem to vary greatly in different types of human malignancies. On the other hand, in view of the cytogenetic location of the HSP70-hom gene within the class III region of the HLA complex, there remains the possibility that the observed association may be due to the linkage disequilibrium between this HSP70-hom polymorphism and other polymorphisms elsewhere within the HLA region.

In summary, the present population-based study suggests that the functional polymorphism in the HSP70-hom gene may be a genetic biomarker for PCa onset in our Tunisian population. Because the incidence of PCa in Tunisia is low, the major analytical limitation of the study remains the small sample size. The study should therefore be viewed as hypothesis generating and should be followed by larger prospective and multiethnic studies to confirm the significance of our findings. Further studies of other genetic variations in the HSP70 genes are also required to understand their role in determining PCa risk and progression.


We would like to thank all Tunisian subjects for their participation in this study. This work was supported by the ministry of Higher Education and Scientific and Technological Research and the ministry of Public Health of Tunisia. We would like to thank Mr. Adel Rdissi for English revision.

Copyright information

© Springer Science+Business Media B.V. 2007