Back to Program


Genome-Wide Microarray Analysis of Semen RNAs from Men With Biopy-Proven Prostate Cancer or Benign Hyperplasia
Robert C. Eyre, MD1, Brian Desmarais, B.S.2, HZ Yin, M.D.2, Ann A. Kiessling, PhD1.
1Bedford Research Foundation, Beth Israel Deaconess Medical Center, Somerville and Boston, MA, USA, 2Bedford Research Foundation, Somerville, MA, USA.

Background
Prostatic secretions and cells in semen may reflect the health of the prostate gland, but semen screening tests for prostate disease are lacking. To begin to evaluate the utility of semen screening for prostate cancer, we performed whole genome microarray analysis of RNAs in semen specimens from vasectomized men to avoid potential confusion of the data by RNA contribution from cells of the germ cell compartment.
Methods
Two semen specimens were obtained over a 4-week interval from patient A two months following transrectal prostate biopsy that revealed Gleason 4+4 cancer (serum PSA 1.0 ng/dl). Two semen specimens were obtained one week apart from patient B two months after prostate biopsy that showed benign hyperplasia only (serum PSA 4.0)
Semen RNAs were isolated by Arcturus kit extraction, and RNA integrity determined by Agilent Bioanalyzer profile of ribosomal RNA 28S/18S ratio, which indicated RNA degradation. RNAs were shipped to MoGene, St. Louis, MO, for Agilent whole genome micro-array analysis following fluorescence labeling specific for degraded RNAs.
Each microarray contained 43,377 gene element probes representing the entire human genome. The fluorescence data for each microarray was assembled into a Filemaker Pro database, linked by Feature Number. To evaluate the reproducibility of the data, the mean and standard deviation of the ratios of each feature were calculated for the cancer semen specimens (ratio 1.00 ± 1.24) and the benign semen specimens (ratio 0.95 ± 0.3).
Results
The ratio of the average fluorescence values for the benign semen specimens to the cancer semen specimens was 1.0 ±2.1. These findings indicated most of the RNAs were present in both benign and cancer semen specimens, as expected.
The ratios of 122 gene elements in the cancer specimens relative to the benign specimens were more than two standard deviations from the mean, an indication they might be useful biomarkers for cancer. Thirty seven were underdetected 4-fold in the cancer specimens, and 85 were over-detected 4-fold. The variability between replicate samples was greater for this subset: average ratio of the two cancer specimens was 5.8±17 and for the two benign specimens was 1.0± 2.0, although the variability between cancer and benign specimens was greater (1.0 ± 36.4).
Of the 37 gene elements underdetected in the cancer specimens, 27 DAVID-mapped to known genes. Nine gene elements were the most consistently over-expressed in both benign specimens: VDR, vitamin D Receptor; DZIP1L, deleted in azospermia-like 1; OSTM1, osteopetrosis associated transmembrane protein 1; BAG5, BCL2-associated anti-apoptotic protein; GRWD1, ribosome biogenesis; PSMC2, 26Sproteosome subunit; and 3 unknown genes.
Of the 85 genes over-detected in the cancer semen, 75 DAVID-mapped to known genes, and 33 were over detected in both cancer specimens of which 19 (BAIAP2, BTN2A3, DEDD2, EIF3S8, GLIPR1L1, HOXB7, LOC497257, LOC644232, MRC2, NUDT22. POLR3GL, PPIL2, PTPRZ, RIPK2, SEC13, UHMK1, WDHD1, ZNF354A, ZNF618) DAVID-mapped to known genes, many of whose functions are unknown.
Conclusion
The panel of 9 genes underexpressed and the 19 genes overexpressed in the prostate cancer semen specimens provides the groundwork for developing potential prostate cancer biomarkers in semen specimens.


Back to Program