Introduction
Copy number increase involving chromosome 8q belong to the most frequent aberrations in human solid cancers, including for example tumors of the breast, ovary, endometrium, lung, colon, head and neck, urinary bladder, kidney and prostate (1,2). Whereas gains often affect large portions of 8q or even the entire q-arm, high-level amplifications are typically focused on the chromosomal bands 8q21 (3–5) and 8q24 (6,7), thus highlighting the loci of putative oncogenes, including MYC at 8q24. Several candidate oncogenes have been suggested to reside inside the 8q21 region, including tumor protein D52 (TPD52) (3,8–13). TPD52 has been suggested to play a role for vesicle trafficking and Ca2+ dependent exocytotic secretion, and has been shown to facilitate cytokinesis in rapidly proliferating cells (14,15).
In line with an oncogenic role, TPD52 overexpression has been described from a multitude of cancer types, including breast (9), lung prostate (2), ovarian (8), pancreatic cancer (16), multiple myeloma (17,18), Burkitt’s lymphoma (19), melanoma (20) and testicular germ cell tumors (21), and has been linked to poor prognosis in breast, medulloblastoma, lung and prostate cancer patients (22). Moreover, cell line experiments and in vivo analysis in mice support the implication of TPD52 in tumorigenesis and progression to metastasis (23,24). Accordingly, TPD52 has been suggested as a promising target for antitumor therapies in breast (25) and prostate cancer (24), and it seems likely that also other tumor types showing TPD52 overexpression could profit from a TPD52 specific therapy.
A systematic analysis of TPD52 expression in human cancers in order to identify tumor types that might benefit from potential anti-TPD52 therapies is lacking. In this study, we have employed quantitative real-time PCR in more than 900 tumor samples to compare the prevalence and expression levels of TPD52 across 29 important human cancer types and corresponding normal tissues.
Results
Technical issue
A total of 894 cancer samples from 29 different tumor types and 40 normal tissue samples from 20 different normal tissue types were included in the analysis (Table I). A total of 105 (10.5%) tumor samples and 3 (7.5%) normal tissue samples were excluded from analysis because either the reference gene TBP or the target gene TPD52 showed a Ct value exceeding 35, indicating that too little cDNA was generated for reliable TPD52 expression analysis.
TPD52 expression in normal tissues
TPD52 transcript was found in all analyzed normal tissues. Average ΔCt was 2.4±0.4. For a direct comparison of the TPD52 expression levels of the different normal tissues, the expression levels were normalized to the TPD52 expression level in testis, which had the lowest expression level of all analyzed normal tissues. Accordingly, highest TPD52 expression levels [≥32 (25) -fold overexpression as compared to normal testis] were found in normal oral cavity mucosa (104-fold overexpression), pancreas (91-fold), thyroid gland (79-fold), kidney (76-fold), prostate (69-fold), stomach (56-fold), vulva (45-fold), liver (39-fold), lymph node (38-fold), and breast (32-fold). All data are summarized in Fig. 1.
TPD52 expression in cancer
To compare TPD52 expression across the different tumor types, average expression levels in the cancer samples were normalized to the corresponding normal tissue. For cancer types without available corresponding normal tissues, the average expression level of all normal tissues (ΔCt=2.4±0.4) was used for normalization. These cancer types included tumors of the larynx, cervix, esophagus and colon. This analysis revealed ≥1.5-fold TPD52 overexpression in 18/29 (62%) analyzed tumor types, with highest levels in non-seminoma (56-fold overexpression compared to corresponding normal tissue), seminoma (42-fold), ductal breast cancer (28-fold) and lobular breast cancer (14-fold). Downregulation as compared to the corresponding normal tissues was found in 11 (38%) of the analyzed tumor types, including papillary renal cell cancer (-8-fold), leiomyosarcoma (-6-fold), clear cell renal cell cancer (-5-fold), liposarcoma (-5-fold) and lung cancer (-4-fold) as the tumor types with the strongest downregulation. All data are summarized in Fig. 2.
Prevalence of TPD52 expression in cancer
In order to estimate the variability of TPD52 expression in the analyzed tumor types, we determined the fraction of samples showing at least 2-fold up- or downregulation. This analysis revealed that TPD52 overexpression was particular frequent (80% of samples showing ≥2-fold overexpression) in germ cell tumors (non-seminoma and seminoma), bladder cancer, esophageal carcinoma and mamma carcinoma, whereas cancer types typically showing TPD52 downregulation included renal cell carcinoma (90% papillary and 68% clear cell), leiomyosarcoma (69%) and liposarcoma (52%). All data are shown in Fig. 3.
Discussion
The results of our study show that TPD52 is overexpressed in a multitude of human solid cancer types. Lower TPD52 levels were also found in the corresponding normal tissues, which is in concordance with previous studies reporting TPD52 expression in normal tissues derived from breast (25–28), prostate (29–31), ovarian (8), lung (32–34), bladder (35), brain (36), thyroid (37), endometrium (38,39), adrenal gland (40) and liver (41). Such low-level expression was expected given the essential role of TPD52 for vesicle trafficking and exocytotic secretion (14).
Many cancer types analyzed in our study were characterized by massive TPD52 overexpression. Overexpression was strongest and also most prevalent in breast cancers, urinary bladder cancers and in testicular cancers. Our data are in agreement with previous studies reporting TPD52 overexpression in breast cancer [7–47%, (3,42–50)]; bladder cancer [21%, (4)], prostate cancer [44–68%, (2,13,29)] and ovarian cancer [61%, (8)] (Table II). Of note, these tumor types are frequently affected by gains of 8q or even high-level amplification of the TPD52 locus at chromosome 8q21 (3,29), suggesting that at least part of the observed overexpression in these cancers may be driven by genomic copy number gains of TPD52. In line with this assumption, numerous previous studies reported associations between TPD52 overexpression and gene copy number increase in breast (9,12,42,43,45), prostate (2) and ovarian cancers (8).
We found substantial TPD52 downregulation in papillary and clear cell kidney cancers, leio- and liposarcomas. Since these cancers do not typically show genomic loss of 8q, it seems obvious that downregulation is not primarily due to TPD52 deletions but may be a result of modified transcription due to genetic or epigenetic regulation. TPD52 downregulation was also reported from leukemia (51–53), lymphoma (54,55), brain cancer (36,56–58) and sarcoma (59).
The good concordance of our results as compared to published studies underlines the validity of our analysis. Only few tumor types analyzed in our study showed discrepant results as compared to the literature, including pancreatic cancer, melanoma and colon cancer. It is possible, that these discrepancies are due to technical differences in the first place. For example, Loukopoulos et al (16) demonstrated upregulation of TPD52 in all 42 analyzed tumors, but analyzed xenografts instead of primary tumors in order to artificially maximize the fraction of tumor cells in the sample. Riker et al (60), Skrzypczak et al (61) and Hong et al (62) reported downregulation of TPD52 expression in colon cancer, whereas upregulation was found in our analysis. We did not include normal colon as a reference in our study but used an average expression value across all analyzed normal tissues as a surrogate. It is possible, that this strategy resulted in bias. The same may also apply for our results obtained from larynx, cervix and esophageal carcinomas, where we used the same ‘average’ reference.
The usage of matched normal tissues for reference in the vast majority of our samples enabled us to estimate the relative impact of TPD52 in the individual tumor types. Testicular germ cell cancers showed the highest levels and frequency of TPD52 expression, suggesting that TPD52 upregulation might play a particularly important role in this tumor type. Comparative genomic hybridization data suggest that large fractions of chromosome 8q, including the TPD52 locus, can be frequently [42–70%, (21,63,64)] gained or amplified (21) in testicular germ cell cancers, supporting a role of genomic gains for the overexpression also in this cancer type. This is consistent with overexpression of TPD52 in testicular germ cell tumors with CGH-confirmed 8q copy number gains reported by Scotheim et al (21).
We identified several additional tumor types with frequent (≥50% of samples) TPD52 overexpression that had not been reported before, including endometrial carcinomas, cervix cancer, and stomach cancers. Also these cancers are frequently affected by 8q gains or even amplifications (5), further supporting the concept of TPD52 copy number alterations represent an important mechanism of TPD52 overexpression (Fig. 3).
The large number of tumor types showing TPD52 overexpression in a significant fraction of samples encourages clinical testing of anti-TPD52 therapies. Payton et al have shown that TPD52 protein-based vaccination in mice induced an adaptive immune response capable of rejecting TPD52-overexpression induced tumorigenesis (65). Lewis et al have shown that TRAMP (transgenic adenocarcinoma of the mouse prostate) mice immunized with cDNA for mD52 as a DNA-based vaccine survived tumor cell challenge through a specific TH1-type T cell response (66). Our data suggest that such a therapy, if effective, could be applied also to a broad range of other tumor types. This is particularly true for cancers with strong overexpression in the tumor cells as compared to the corresponding normal tissue, including for example testicular germ cell cancers. Our finding, that breast cancer ranks second as a tumor type with strong cancer-related TPD52 expression level changes emphasized the potential of anti-TPD52 therapy in this tumor type.
In summary, our data demonstrate that TPD52 overexpression is common in many tumor types. Particularly strong TPD52 upregulation was found in cancers of the breast, urinary bladder cancer and testicular germ cell cancers, which frequently harbor 8q gains. These tumor types may be particularly promising candidates for potential anti-TPD52 therapies.