Tissue samples were obtained at the time of the surgery from the City hospital, Nottingham and Queen’s Medical Centre, Nottingham. All tumour tissue samples were verified histologically by observation of tissue sections by a clinical pathologist. Blood samples were obtained from patients and healthy male volunteers with no known evidence of benign or malignant prostatic disease and the serum collected was stored at -80°C. All patients gave informed consent and this study received local ethical approval. PC3 and DU-145 human prostate carcinoma cell lines were maintainedin Dulbecco's modified Eagle's medium supplemented with 10% heatinactivated foetal calf serum (FCS); the LNCaP human prostate carcinoma cell line was cultured in RPMI 1640 medium with 10% FCS, 2mM L-glutamine, testosterone (5ng/ml) and hydrocortisone (5ng/ml).

CT26 (murine colon carcinoma) cells were maintained in DMEM media supplemented with 2mM L-glutamine and 10% FCS; B16-F1 (murine melanoma) and CMT 93 (murine rectal carcinoma) cells, A20 (B cell lymphoma) cells and RENCA, a BALB/c renal carcinoma cell line were maintained by serial in vitro passage in RPMI 1640 tissue culturemedium supplemented with 2 mM L-glutamine, 10% FCS, and 0.05mM 2-ME.Normal mouse tissues were harvested from naïve Balb/c mice and were immediately snap frozen in liquid nitrogen and stored at -80˚ C.

Two cDNA expression libraries were constructed from two tumour specimens of a well/moderately differentiated prostatic adenocarcinoma and an undifferentiated prostatic adenocarcinoma, respectively. These two specimens were obtained from Professor F Hamdy from patients attending Newcastle University under local ethical approval. Total RNA was isolated using RNA STAT 60 (ams Biotechnology, UK) according to the manufacturer’s protocol. Poly(A)⁺ RNA was purified from total RNA using a mRNA isolation kit (Stratagene UK) and cDNA was ligated into the ZAP Expression vector using Gigapack III Gold cloning kit (Stratagene UK). After in vitro packaging, two libraries containing 1.2(10⁶ and 2.0(10⁶ primary cDNA clones respectively were obtained.

Immunoscreening of the cDNA library was performed as described by Sahin et al. [3]. Briefly, 1:10 diluted autologous patient’s serum and allogeneic sera, were preabsorbed with E.coli-phage lysate to remove antibodies reactive with antigens related to the vector system. In order to eliminate cDNA clones encoding human immunoglobulins, membranes were pre-screened with AP-conjugated rabbit anti-human IgG, Fcγ fragment specific secondary antibody (Pierce, USA) prior to incubation with sera. Reactive plaques were detected with 5-bromo-4-chloro-3-indolyl-phosphate/ nitroblue tetrazolium (BCIP/NBT), marked and excluded from further study. The membranes were then incubated with pre-absorbed 1:200 diluted patient’s serum and serum-reactive clones were detected with AP-conjugated secondary antibody and visualized by incubating with BCIP/NBT. The reactive phage clones were subcloned to monoclonality and converted to pBK-CMV phagemids. To assess frequencies of antibody responses to the SEREX defined antigens in individual allogeneic sera from prostate cancer patients and normal individuals, E. coli were transfected with approximately equal numbers of positive phage and non-recombinant phage. These were then screened with 1:200 diluted individual allogeneic serum samples as described above, excluding the IgG pre screening step.

Phagemid DNA was purified using QIAprep Spin Miniprep kit (QIAGEN, West Sussex, UK), analysed by EcoRI/XhoI restriction enzyme digestion and clones representing different cDNA inserts were sequenced using an ABI PRISM 310 automatic sequencer (Applied Biosystems UK). Sequence alignments were performed using the GenBank database (www.ncbi.nlm.nih.gov). Genes identical with entries in the GenBank were classified as known genes, whereas those that shared sequence similarity only with ESTs and those that had no similarity to entries in either the GenBank or EST databases were classified as unknown genes. Multiple sequence alignments were performed with DNASIS (Hitachi Software Engineering Co Ltd) and MACAW (NCBI) software. Chromosomal localisation and exon-intron organisation for uncharacterised cDNAs was determined by comparison to the working draft of the human genome.

To evaluate the mRNA distribution of SEREX-defined antigens, total RNA from a panel of normal human tissues (lung, liver, brain, trachea, heart, and kidney) was purchased from Clontech, (UK). Total RNA from normal testes, oesophageal cancer and adjacent non-tumour tissues, head and neck tumour and adjacent non-tumour tissues, prostate cancer tissues, stomach cancer and kidney cancer tissues and cell lines (DU-145, LNCaP and PC3) was isolated using RNA STAT 60 reagent (ams Biotechnology, UK) according to the manufacturer’s protocol. For mouse tissue screening, several tissues were dissected from mice and immediately snap frozen. Tissues were then ground in liquid nitrogen before proceeding to total RNA isolation using RNA STAT-60. The first-strand cDNA was synthesised from 2μg of total RNA primed with oligo-dT(15) (Promega UK) as described in the manufacturer’s protocol. Human MTA1 (5’-CGCTCAAGTCCTACCTGGAG and 3’-TGGTACCGGT TTCCTACTCG) and GAPDH (5’-ACCACCAACTGCTTA GCACC and 3’-CCATCCACAGTCTTCTGGGT) specific primers were used to verify expression of hMTA1. The mouse MTA1 gene specific primers (5’-GCGAGAGCTGT TACACCACA and 3’-ACTGCTGAGCACACTGGATG) and murine GAPDH primers (5’-GGTGAAGGTCGGAGT CAACGGA and 3’-GAGGGATCTCGCTCCTGGAAGA) were designed with the assistance of the Primer3 website and obtained lyophilized from Sigma Genosys (UK). The gene specific PCR primers were designed to ensure that the primer sequences were located within different exons and that the PCR products obtained were 300-500bp in length. GAPDH was used as an internal control to observe the relative amounts of cDNA from each sample that was used in the RT-PCR reaction. The cDNA templates were used in the PCR in a final volume of 50 µl containing 0.4 μM of each primer, 1.5 mM MgCl₂, 0.2 mM deoxynucleotide triphosphates (dNTPs) and 2.5 U Taq polymerase (Promega, UK).

PCR conditions were as follows for MTA1: Initial denaturation at 95ºC for 5 mins, followed by 34 cycles of denaturation at 95ºC for 1 min, annealing at 58ºC for 1 min and extension at 72°C for 1 min. This was followed by a final extension step at 72°C for 5 mins. GAPDH PCR conditions consisted of 25 cycles of denaturation at 94ºC for 45 secs, annealing at 58ºC for 45 secs and extension at 72°C for 45 secs. Semi-quantitative analysis of RT PCR products was performed by 1.5 % (w/v) agarose gel electrophoresis and ethidium bromide visualization. After normalization to GAPDH mRNA as a control for each cDNA template, the expression of gene-specific mRNA in the cDNA samples tested was compared.

After immunoscreening of 8 x 10⁵ pfu from each of two prostate carcinoma cDNA expression libraries with the autologous serum, a total of 26 and 73 immunoreactive clones were identified from first and second cDNA expression libraries, respectively. These clones were purified, in vivo excised, and converted to pBK-CMV plasmid forms. The cDNA inserts were analysed using restriction mapping and sequencing. Sequence comparison against GenBank and EST databases showed that these clones were derived from 54 distinct genes, which were designated as Pr1-Pr54. Pr1-Pr11 were derived from the first cDNA library and the remaining clones from the second library. Eight genes showed no sequence similarity or only partial sequence similarity with genes in the GenBank database but shared sequence similarity with ESTs derived from various tissues (referred to as “unknown genes”) (Table 1).

We therefore investigated these genes further and EST analysis of Pr3 (BC017874), Pr9 (BC062419), Pr10 (XM_496394), Pr11 (AC011375), Pr43 (NM_014929) and Pr50 (BC007379), showed that they shared sequence identity with reported ESTs derived from various normal tissues, e.g. lung, liver, kidney, brain and B cells, suggesting that these genes are widely expressed in human tissues (Table 1). Pr6 (AC104841) transcripts on the other hand looked more promising as no similarities with normal tissue were found which was confirmed in a panel of RNAs isolated from normal human tissues as determined by RT-PCR after 40 cycles of amplification. Unfortunately we were also unable to detect it in a panel of tumour tissues or tumour cell lines. The last one, Pr52, is located on chromosome 7p21 and is similar only to an EST derived from human placenta (DA832074). The tissue expression pattern of this gene was therefore evaluated by RT-PCR and results showed that Pr52 was expressed in normal brain, kidney, thyroid, testis and placenta (data not shown). It was however not found at the time of this investigation whether the expression of these genes was higher in cancer tissues compared to normal tissue. Further work is now being carried out to evaluate this using real-time RT-PCR.

The remaining 46 genes showed sequences identical with or highly similar to known GenBank entries (“known genes”) (Table 2). Three of these had been previously isolated from other tumour tissues by SEREX (Pr 17, 48 and 54). Clone Pr54 is identical to human transcriptional coactivator tubedown-100 (TBDN100), which is involved in regulation of angiogenesis and was isolated by us from gastric cancer and is overexpressed in gastric cancer tissues compared with adjacent non-cancer tissues (Ga19, GenBank accession number AY039242) [4]. Clone Pr48 is similar to the gene encoding human ankyrin repeat domain 17, which has also been isolated from breast cancer (NY-BR-16, Genbank accession number AF308285) [8]. Clone Pr17 is identical to human NEDD8 ultimate buster-1 (NUB1) mRNA, which was detected in renal cell carcinoma (NY-REN-18, Genbank accession number AF155099) [9].

To determine whether immune recognition of antigens encoded by SEREX genes was tumour-related, allogeneic sera from normal individuals and patients with prostate cancer were tested for antibody reactivity against selected antigens using a protein array system. Of the 35 antigens screened, 17 antigens reacted with a subset of sera from both normal individuals and cancer patients; thus, approximately half (17/35) of the antigens identified in this study were non-cancer related and were classified as naturally occurring autoantigens. Nine proteins had a cancer-related serological profile, reacting only or mainly with sera from prostate cancer patients. The remaining antigens only reacted with autologous serum and not with any allogeneic sera from normal individuals or patients with prostate cancer (Table 3).

From the panel of genes identified, we decided to focus on genes which did not generate any reactivity to serum obtained from normal patients, since these are more likely to demonstrate cancer-specific immunological responses. These genes included the aarF domain containing kinase 4 (ADCK4), for which very little information could be found; the ribosomal protein S10, which has previously been shown to be deregulated in colorectal cancer samples, along with larger transcripts being produced in addition [10]; the cell death regulatory protein GRIM19, which appears to be an important cell death regulatory protein and is involved in mitochondrial metabolism. Loss of GRIM-19 expression has been shown to be associated with renal cell carcinomas. Interestingly, GRIM-19 mutations have been observed in mitochondrial rich tumours of thyroid [11]. Considering that it is one of the critical regulatory proteins and its expression is responsible for tumour cell death, it seems an unlikely antigen for immunotherapy. And finally, MTA1 which was found to represent the most promising gene based on previously published literature reporting on its overexpression in various cancers. Moreover, being highly expressed in metastatic cells suggests that it would be a highly specific therapeutic target which could be potentially used to prevent disease relapse in advanced disease states. Also, several studies have clearly established that MTA1 overexpression leads to cell transformation and by suppressing MTA1 expression you get an inhibition of cell growth [12,13].

The expression pattern of MTA1 transcripts in a panel of normal human tissues, matched human non-tumour and tumour tissues, human prostate cancer tissues and cell lines was analysed by RT-PCR. Weak expression of MTA1 transcript was detected in normal human tissues except testis, where MTA1 was highly expressed (Fig. 1A). Human head and neck tumours and oesophageal cancer tissues expressed higher levels of MTA1 transcripts than matched non-tumour tissues (data not shown). MTA1 mRNA is highly expressed in human prostate cancer cell lines and a significant number of prostate, renal and stomach cancer tissues (Fig. 1B). We further analysed the expression of MTA1 in murine tissues and tumour cell lines since MTA1 is highly conserved in humans and mouse demonstrating 96% similarity at the protein level. MTA1 was found to be over-expressed in all murine tumour cell lines as compared to normal tissues, except the testis (Fig. 1C), which showed a similar level of expression to tumour tissues (data not shown). In comparison, all the normal tissues examined expressed MTA1 at very low levels (Fig. 1C). Interestingly, we found that MTA1 was expressed at low levels in muscle tissue compared to other tissues, which contradict previously published work where no MTA1 expression was found [14].

Fig. (1). (A) Expression of MTA1 in normal human tissues analysed by RT-PCR. Cycling conditions were optimized so that the RT-PCR products were analyzed when amplification is within the linear phase. Lanes 1-9 from left to right, normal brain, kidney, heart, liver, lung, PBMC, testis, negative control and DNA ladder. (B) Tissue distribution of MTA1 transcript in malignant tissues and cell lines was analysed by RT-PCR. Lanes 1-4: prostate cancer tissues; lanes 5-7: prostate cancer cell lines (lane 5: DU-145; lane 6: LNCaP; lane 7: PC3); lanes 8-11: renal cancer; lanes 12-16: gastric cancer. Amplification of GAPDH was also performed in parallel as a control in all samples. (C) Agarose gel electrophoresis of mGAPDH and mMTA1 following RT- PCR from murine tumour cell lines and tissues. Lanes 1-10 from left to right, CT26, A20, RENCA, CMT, B16, liver, lung, muscle, spleen, kidney.