Generation of a baculovirus recombinant prostate-specific membrane antigen and its use in the development of a novel protein biochip quantitative immunoassay.

Prostate-specific membrane antigen (PSMA) is a 100-kDa transmembrane glycoprotein identified by the monoclonal antibody 7E11-C5.3 from the human prostate tumor cell line LNCaP. Because of its significant upregulation in androgen refractory and metastatic prostate cancers, PSMA may be a useful prognostic biomarker and a target for developing novel therapeutic strategies. However, the lack of abundant pure PSMA protein and the low efficacy in immunoaffinity isolation from LNCaP cells have hampered the development of clinical assays. In order to obtain a renewable and reliable source of pure antigen, we used the baculovirus/insect cell system to express and purify a recombinant PSMA. A recombinant baculovirus containing a 6x histidine-tagged PSMA gene was generated, from which rPSMA was expressed and purified using cobalt-chelating affinity chromatography. The purity and correct molecular size of rPSMA were demonstrated by gel electrophoresis and mass spectrometry. Glycosidase digestions showed that the oligosaccharides on rPSMA are primarily N-linked high-mannose type. Although the glycosylation is different from the native PSMA, it did not affect the immunoreactivity of rPSMA to antibodies specific for either the intra- or the extracellular domains of PSMA. Finally, the purified rPSMA was successfully used to develop a quantitative PSMA immunoassay using the novel ProteinChip surface-enhanced laser desorption/ionization mass spectrometry technology.

Identification of a superimmunoglobulin gene family member overexpressed in benign prostatic hyperplasia .

BACKGROUND: Benign prostate hyperplasia (BPH), a nonmalignant disease with an increasing rate of occurrence associated with advancing age, requires auxiliary markers to help identify its presence and distinguish its progression from prostate cancer. METHODS: Hybridoma technology was used to generate an antibody against a BPH antigen, which was subsequently characterized by Western blot analysis, sequence homology, and RT-PCR. RESULTS: A BPH-associated protein, designated P25/26, was identified that showed a strong sequence similarity with superimmunoglobulin family members, overexpressed in BPH, with lower expression observed in both normal and prostate cancer tissues. CONCLUSIONS: Further studies appear warranted to assess the role that this and other superimmunoglobulin family members may have in the pathogenesis of BPH, and to determine if these glycoproteins have any clinical utility in the differential diagnosis or therapeutic monitoring of BPH.

Prostate-specific membrane antigen levels in sera from healthy men and patients with benign prostate hyperplasia or prostate cancer.

Prostate-specific membrane antigen (PSMA) serum levels have been proposed to be of prognostic significance in patients with advanced prostate disease. The objective of the present study was to confirm PSMA serum expression by Western blot techniques, to determine whether such data could assist in the differentiation of benign from malignant prostatic disease, and to determine the suitability of serum PSMA measurements in predicting recurrent or progressive prostate malignancies. We measured PSMA, a transmembrane glycoprotein identified in prostate epithelial cells, in the sera of 236 normal individuals and cancer patients by Western blot analysis. Within the normal male population, PSMA levels increase with age and were found to be significantly elevated in subjects more than 50 years of age when compared to those of younger men. We did not confirm previous reports that serum PSMA measurements could distinguish late-stage prostate carcinoma from early-stage prostate carcinoma, nor did we find PSMA to be more effective than prostate-specific antigen in monitoring prostate cancer patient prognosis. Furthermore, we found elevated serum PSMA in healthy females, and, similar to the healthy male population, the levels increased with age, with the highest levels found in the sera from breast cancer patients. These latter observations further support that PSMA is not a specific biomarker for prostate cancer and that a variety of normal and diseased tissue may contribute to the serum levels of PSMA.

Location of prostate-specific membrane antigen in the LNCaP prostate carcinoma cell line.

BACKGROUND: Prostate-specific membrane antigen (PSMA) is a novel prostate biomarker overexpressed in poorly differentiated and metastatic prostate carcinomas and apparently upregulated following hormone-ablation therapy. PSMA appears to be a satisfactory target for antibody-directed imaging of prostate carcinomas despite the recent finding that the antigenic epitope recognized by monoclonal antibody (MAb) 7E11-C5 is found in the cytoplasmic domain of this transmembrane glycoprotein [Troyer et al.: Urol Oncol 1:29-37, 1995]. This finding prompted the present investigation to precisely define the cellular location of PSMA in the LNCaP prostate carcinoma cell line, the line used to generate MAb 7E11-C5. METHODS: Subcellular fractionation, immunofluorescence and immunoperoxidase staining of live and fixed cells, and immunoelectron microscopy were used to determine the localization of PSMA in LNCaP cells. RESULTS: PSMA was found to be localized at the inner face of the plasma membrane as well as being associated with mitochondria. Staining of LNCaP cells, treated by serum starvation followed by serum stimulation, showed no changes in the typical cytoplasmic staining pattern. CONCLUSIONS: The data suggest that the PSMA target epitope for antibody-directed imaging with MAb 7E11-C5 only becomes accessible upon apoptosis or necrosis. This further suggests that antibodies directed at the extracellular domain may enhance the sensitivity of antibody-directed imaging and therapy of prostate carcinomas by recognizing surface epitopes of PSMA on living cancer cells.

Detection and characterization of the prostate-specific membrane antigen (PSMA) in tissue extracts and body fluids.

The prostate-specific membrane antigen (PSMA) glycoprotein is recognized by the monoclonal antibody (MAb) 7E11-C5.3 as a predominant 100 kDa and minor 180 kDa component in LNCaP cell line extracts and its expression has been shown by immunohistochemistry to be highly restricted to prostate epithelium. The aim of the present study was to utilize Western blot analysis to determine if PSMA could be detected in human tissue extracts and body fluids and if so, which molecular forms were present. PSMA was detected as 120 and 200 kDa bands in normal, benign and malignant prostate tissues and seminal plasma. Further analysis demonstrated that the larger molecular form of PSMA may be a dimer of the lower m.w. species. The PSMA glycoprotein was not detected in the majority of non-prostate tissue extracts examined except for a low yet significant amount in normal salivary gland, brain and small intestine, suggesting that PSMA may not be as prostate-specific as originally thought. Since the prostate-specific antigen (PSA) has been shown to be maximally shed into the serum in high-grade and metastatic prostate carcinomas, it was surprising that PSMA could not be detected in serum by Western blot analysis even in patients with actively progressive metastatic disease. Second generation antibodies generated against different epitopes may be required to determine if PSMA is shed into serum. Our results support the hypothesis that PSMA is a novel prostate biomarker.

Characterization of a prostate carcinoma mucin-like antigen (PMA).

A recently identified, high m.w. human tumor antigen, reactive with monoclonal antibody (MAb) PD41 and designated prostate mucin antigen (PMA), was found to have expression highly restricted to prostate carcinomas. Both biochemical and immunological characteristics of this antigen and its relationship to other tumor-associated mucins and various species of submaxillary mucins were evaluated. Immunohistochemical examination of submaxillary tissues revealed differential expression of this antigen and other human tumor-associated mucins, with MAb PD41 expression found only in bovine submaxillary gland serous cells. Neuraminidase treatment enhanced antibody binding by 50%, suggesting partial masking of the PD41 epitope by sialic acid. Antigenicity was reduced by treatment with alkaline-borohydride, sodium metaperiodate, beta-galactosidase, O-glycanase, and various proteolytic enzymes. MAb PD41 antibody binding also could be significantly reduced by selected lectins and sugars suggesting that the immunodominant carbohydrate involved in the epitope was an O-linked oligosaccharide containing N-acetyl galactosamine as a major component. This antigen was further distinguished from T, Tn, or Sialyl-Tn antigens and blood group carbohydrate antigens by radioimmunometric analyses. Cross-blocking and double determinant RIA experiments also showed a distinction between the PD41 epitope and several well-characterized tumor-associated mucin antigens such as MUCI, CEA, M344, OCI25, and AR3 as well as bovine submaxillary core protein. Our results indicate that the PD41-reactive epitope is a non-acidic O-linked carbohydrate or glycopeptide epitope with restricted expression in prostate carcinomas and bovine submaxillary glands. This expression is distinct from other mucin-like tumor-associated antigens identified in human prostate carcinomas.

Expression of prostate-specific membrane antigen in normal, benign, and malignant prostate tissues.

Prostate-specific membrane antigen (PSMA) is a trans-membrane glycoprotein recognized by the murine monoclonal antibody (MAb) 7EII-C5.3 both in its native (CYT-351) and immunoconjugate form (CYT-356). Previous studies have shown that tissue expression of PSMA is highly restricted to prostate tissues. In this study, a definitive immunohistochemistry evaluation was performed to assess PSMA expression in prostate tissues. A stain index was established by multiplying the percentage of stained cells by the intensity of the stained cells to provide a quantitative measurement of PSMA expression in the various tissue types. The cellular location of PSMA, its correlation with clinical status, and its comparison with the expression of prostate-specific antigen (PSA) were evaluated. Prostate-specific membrane antigen was found to be highly expressed in most of the normal intraepithelial neoplasia, and the primary and metastatic prostate tumor specimens evaluated. In contrast to PSA, PSMA expression was often heterogeneous with variable staining patterns, ranging from a low-level diffuse cytoplasmic staining in normal prostate epithelium to very intense cytoplasmic and focal membrane staining in high-grade primary carcinomas and metastatic tissues. The predominant cytoplasmic staining was expected because the antigenic epitope of the PSMA transmembrane glycoprotein recognized by MAb 7EII-C5.3 is located in the cytoplasmic domain. Benign prostate tumors, ie, hypertrophy, showed the lowest expression of PSMA with a stain index of 52, compared with stain indexes of 146 and 258 for normal prostate and bone metastatic tissues, respectively. The reason for the apparent down-regulation of PSMA in benign prostate tissue is unknown but may be related to a splicing variant or post-translational modification of PSMA. Expression of PSMA was observed to increase with increasing pathologic grade, but not with clinical stage. Although PSMA was overexpressed in poorly differentiated and metastatic prostate tumors, expression in the primary tumor did not correlate with nodal status, extracapsular penetration, or seminal vesicle invasion. These results suggest that PSMA is not a useful biomarker of disease progression; however, high expression does appear to be associated with the more aggressive prostate carcinoma phenotype. The restricted specificity, differential prostate tissue expression, and overexpression of PSMA in metastatic tissues support the continued study of this unique prostate tumor-associated biomarker for developing new strategies for diagnostic and therapy of prostate cancer.

Biochemical characterization and mapping of the 7E11-C5.3 epitope of the prostate-specific membrane antigen.

The expression of the prostate-specific membrane antigen (PSMA) glycoprotein recognized by the murine monoclonal antibody (MAb) 7E11-C5.3 has been shown to be highly restricted to prostate epithelium. Although the conjugated form of this MAb (CYT-356) may soon be used clinically for in vivo imaging of extraprostatic disease, few details regarding the nature of the antigenic epitope of PSMA have been reported. This study was carried out to analyze the MAb 7E11-C5.3 epitope on PSMA using standard biochemical techniques, and the antigenic epitope was mapped with synthetic peptides. The MAb 7E11-C5.3 epitope was susceptible to both periodic acid oxidation and proteolytic digestion, which indicated that the antigen consisted of a glycoprotein. However, additional biochemical assays such as sodium borohydride, tunicamycin treatment, and digestion with glycosidases failed to abrogate MAb 7E11-C5.3 binding. Epitope mapping with synthetic peptides demonstrated the epitope to be localized to the intracellular domain at the N-terminus of the PSMA molecule with a minimal reactive peptide consisting of six amino acids (MWNLLH). The synthetic peptides were treated with periodic acid, which resulted in inhibition of antibody binding, suggesting that treatment of the PSMA antigen resulted in damage to the peptide chain. These data suggest that the MAb 7E11-C5.3 does not recognize a glycopeptide as was initially thought, but recognizes an intracellular epitope consisting of only the primary polypeptide chain. Further studies are needed to determine how CYT-356 is able to image tumors in vivo when the antigenic epitope is intracellular.

Two-site monoclonal antibody-based immunoradiometric assay for measuring prostate secretory protein in serum.

We developed a double-determinant immunoradiometric assay for measuring serum prostate secretory protein (PSP), using monoclonal antibodies (MAb) against two different epitopes: MAb PSP-19 was the capture antibody and MAb PSP-6 was the tracer antibody. Assay sensitivity was 0.1 microgram/L. Analytical recovery of PSP was 93.5-104.6%, whereas the intra- and interassay mean CVs were 4.2% and 6.9%, respectively. In 92 normal men, ages greater than 50 years, the mean PSP concentration was 5.7 micrograms/L, with 10 (10.9%) men having concentrations greater than 10 micrograms/L. In contrast, 20 of 49 (40.8%) patients with benign prostate hyperplasia (BPH; mean PSP concentration 9.4 micrograms/L) and 46 of 100 (46%) patients with prostate cancer (mean PSP concentration 22.2 micrograms/L) had PSP concentrations greater than 10 micrograms/L. Mean serum PSP concentrations of the BPH (P less than 0.05) and prostate cancer (P less than 0.01) groups were significantly different from those of age-matched normal men. In a small group of patients, serial PSP concentrations correlated with the clinical course during therapy. Thus, PSP may be a useful marker for evaluating patients with prostate cancer.

A novel prostate carcinoma-associated glycoprotein complex (PAC) recognized by monoclonal antibody TURP-27.

A prostate carcinoma-associated antigen recognized by MAb TURP-27 was characterized immunohistochemically and biochemically. TURP-27 antigen was found localized in the cell membrane and cytoplasm of the ductal epithelial cells of normal (10%), benign (75-100%) and malignant (20-100%) prostate cells. Fetal prostate tissues were also found to express the TURP-27 antigen, suggesting expression early in development. This antigen was not expressed by non-prostate tumors examined, but significant cross-reactivity was observed in myelinated nerves while minor cross-reactivity was seen in certain lymphocyte subsets, cells in the adrenal medulla and chief cells of stomach. Immunoblotting and biochemical data demonstrated that the TURP-27 antigen is a sialic-acid-containing glycoprotein complex with major molecular species in prostate tissues of 310-250, 180, 140, 115, 95-90, 69, and 40- to 35-kDa. Immunoblotting patterns similar to those observed for prostate tissues were also seen in CNS extracts with the exception of the 69 and 40- to 35-kDa proteins. This prostate carcinoma-associated sialoglycoprotein complex (PAC) recognized by MAb TURP-27 is likely to represent a novel tumor antigen expressed by prostate tumors.

Monoclonal antibody PD41 recognizes an antigen restricted to prostate adenocarcinomas.

A monoclonal antibody (MAb) designated PD41 (IgG1k) was generated by hyperimmunizing BALB/c mice with a membrane preparation prepared from a moderately to poorly differentiated prostate carcinoma surgical specimen. The immunohistochemical reactivity of MAb PD41 was shown to be highly restricted to the ductal epithelia and secretions of prostate adenocarcinoma tissues. Sixty-five % of the prostate tumor specimens were stained with MAb PD41, whereas no staining of the fetal or benign prostate specimens was observed. PD41 reacted minimally with normal prostate tissues, with less than 1% of the epithelial cells staining. This MAb did not react with nonprostate carcinomas or to a variety of normal human tissues. Using both radioimmunoassay and immunofluorescent procedures, several cultured human tumor cell lines, human blood cells, and purified antigens to prostate-specific antigen and prostatic acid phosphatase also were found not to express the PD41 antigen. MAb PD41 also was shown to bind to the target antigen present in seminal plasma obtained from prostate carcinoma patients but not to seminal plasma from normal donors. Immunoblots of gel-separated components of prostate carcinoma tissue extracts indicate that the molecular weight of the proteins carrying the PD41 antigenic determinant can differ among individual tumors, ranging from Mr 90,000 to greater than 400,000. However, in seminal plasma from prostate cancer patients, the predominant component recognized by PD41 is the diffuse Mr greater than 400,000 band. It appears that this monoclonal antibody may recognize a prostate carcinoma-associated mucin-like antigen, which is preferentially expressed on prostate carcinomas, and therefore, may be a useful marker to distinguish benign prostate hyperplasia from prostate carcinoma.

Generation and characterization of monoclonal antibodies to prostate secretory protein.

Monoclonal antibodies (MAbs) were produced against a highly purified preparation of prostate secretory protein (PSP) isolated from normal seminal plasma. Fifteen antibodies were selected for further evaluation based on their strong reactivity and specificity for PSP. All the MAbs had a specificity for prostate epithelial cells and none reacted to any of a variety of normal tissues as determined by immunoperoxidase staining. Six of the MAbs were selected for further immunohistochemical evaluation based on their ability to recognize different antigenic determinants. Using competitive binding immunoassays, a variety of overlapping specificities were observed with at least 2 distinct epitopes identified. Although some staining variability was noted, the 6 antibodies, in general, gave the same pattern of tissue reactivity. Both the normal prostate and the benign prostate hyperplastic ductal epithelial cells stained intensely, with 78 to 100% and 50-100% of the cells staining, respectively. The number and often the staining intensity of the tumor cells decreased as the tumor became more undifferentiated. Approximately 40 to 100% and 15 to 70% of the tumor cells stained in the moderately-differentiated and well-differentiated carcinoma tissues, respectively, whereas either no staining was observed or less than 20% of the tumor cells stained in the poorly-differentiated and undifferentiated tumors. Most of the metastatic prostate tumors showed either no staining or scattered staining in a few cells (i.e., less than 20%).

Radiolocalization of human prostate tumor in a mouse subrenal capsule model by monoclonal antibody TURP-27.

The subrenal capsule assay was used to determine if 125I-labeled anti-prostate monoclonal antibody TURP-27 could target human prostate tumor fragments implanted under the renal capsule of normal immunocompetent C57 BL/6 mice. Maximal binding and optimal tumor to non-tumor tissue ratios occurred within 24-48 hours postadministration of 125I-TURP-27. No significant localization was observed in mice bearing TURP-27 antigen-negative human colon tumor tissue implants or with an isotype-matched control monoclonal antibody. These preclinical data suggest that TURP-27 may have clinical application for imaging metastatic prostate tumors and further application in immunoconjugate and/or radiotherapy of prostate cancer.

Binding of human IgG to cultured urogenital tumor cells.

Binding of human IgG by the Fc portion of the immunoglobulin molecule was detected on established human tumor cell lines by indirect immunofluorescence microscopy, cytofluorography, quantitative absorption, and rosette formation with the use of antibody-coated erythrocytes. Of the nonlymphoid tumors tested, IgG binding was restricted to the cell membranes of certain prostate and urinary bladder tumor cell lines. Although most cell lines tested shared a common antigenic determinant with monocytes and granulocytes, these cells did not express T- and B-cell antigens, the complement 3b receptor, or bind a monoclonal antibody specific for the Fc receptor expressed on human neutrophils. The facts that IgG binding was present on long-term established tumor lines and was not influenced by in vitro passage provide evidence that these properties are intrinsic to the tumor cells and may play some role in the pathophysiology of these tumors.

Human prostate tissue antigens defined by murine monoclonal antibodies.

BALB/c mice were hyperimmunized against a membrane preparation derived from a pool of transurethral resection specimens which included three benign prostatic hyperplasia and one prostate adenocarcinoma tissue samples. The activated lymphocytes were fused with the NS-1 mouse myeloma cell line, and supernatants from immunogen-reactive hybridomas were screened for antibody binding activity using a solid-phase radioimmunoassay against the Calu-1 human lung adenocarcinoma cell line and several membrane preparations derived from various normal human tissues. Hybridoma cultures secreting antibodies which did not appear cross-reactive were doubly cloned by limiting dilution and screened against a large panel of membrane preparations derived from normal prostate, benign prostatic hyperplasia, and prostate adenocarcinoma tissues as well as samples obtained from a variety of normal human tissues. The monoclonal antibodies were also evaluated against 24 normal, virally transformed, and malignant human cell lines. Two monoclonal antibodies were isolated which demonstrated a restricted binding activity to prostate antigens and were not widely cross-reactive with nonprostate normal tissues or cell lines. These antibodies were designated TURP-27 (IgG3, k) and TURP-73 (IgG2a, k). Both of these monoclonal antibodies were reactive against formalin-fixed, paraffin-embedded tissues in the immunoperoxidase assay and were subsequently tested against a variety of normal, hyperplastic, and malignant human tissues. These studies indicated that TURP-27 may be directed against a new prostate organ-associated marker and that TURP-73 is directed against an antigen expressed on prostate and a limited number of other tissues.

Antiprostate carcinoma monoclonal antibody (D83.21) cross reacts with a membrane antigen expressed on cytomegalovirus-transformed human fibroblasts.

Virological and epidemiological studies have implicated human cytomegalovirus (HCMV) as a possible etiological agent of prostate cancer. Because of the suspected associations, this laboratory tested the reactivity of a prostate-associated monoclonal antibody with HCMV-transformed cells. This mouse monoclonal antibody, D83.21, reacts with a membrane antigen on prostate and bladder tumor cells and does not bind to a variety of other malignant or normal cells. The results of this study indicated that the prostate-associated antibody bound to a membrane antigen on HCMV-transformed cells as detected by radioimmunoassay, immunofluorescence, and complemented-dependent cytotoxicity. This cross-reactivity appeared to be specific for HCMV-transformed cells and did not react with HCMV-infected cells or those transformed by other viruses. Antibody affinity chromatography, used to isolate the D83.21-reactive protein, revealed two peptides of 60 and 28 kd on both prostate tumor and HCMV-transformed cells. The results suggest that D83.21 reacts with a common cell surface protein expressed on HCMV-transformed cells and urogenital tumors.

Immunohistochemical localization of prostate carcinoma-associated antigens.

The immunoperoxidase technique was used to study the localization and distribution of antigens reactive with two monoclonal antibodies, D83.21 and P6.2, produced against cultured prostate tumor cells, in formalin-fixed, paraffin-embedded histological sections of human tissues. Monoclonal D83.21 reacted with 11 of 19 (58%) primary prostate carcinomas and 1 of 6 (17%) metastatic tumors, whereas monoclonal P6.2 reacted with 14 of 19 (68%) primary and 4 of 6 (67%) metastatic prostate tumors. Neither antibody reacted with five primary prostate tumors and one metastatic prostate tumor. In some tumor cells, the antigens recognized by these monoclonals were localized in either the cytoplasm or cell membrane, while in other tumor cells, both diffuse cytoplasmic and membrane or focal staining patterns were observed. In addition to the variable staining patterns, antigenic heterogeneity was also noted within most prostate tumors examined. Two types of staining variability were observed: (a) tumor cells in one area of the tissue section stained positive, but in another area they did not react with the antibody; and (b) both stained and unstained tumor cells were adjacent to each other. These results would suggest that a panel of monoclonals will be required to detect the different subpopulations of prostate tumor cells. Neither antibody reacted with 6 normal or 12 benign prostate tissues, nor any of a variety of other normal human tissues except for staining of the proximal tubules of normal kidneys. The antigen detected by P6.2 demonstrated a wider tissue distribution being found on bladder, breast, lung, and pancreatic tumors, whereas the antigen recognized by D83.21 was restricted to prostate and bladder carcinomas. These antibodies may have clinical applicability for the identification of prostate tumor cells in biopsy specimens and for immunohistopathological classification of prostate carcinomas.

Monoclonal antibodies to human prostate and bladder tumor-associated antigens.

Monoclonal antibodies to human prostate adenocarcinoma membrane antigens were produced by fusion of P3X63/Ag8 mouse myeloma cells with spleen cells from BALB/c mice immunized against the prostate cancer cell line DU145. The hybrids were screened for antibody production using glutaraldehyde-fixed cells in a solid-phase radioimmunoassay. Antibody-binding specificity was also checked by quantitative adsorption, membrane immunofluorescence, and complement-dependent cytotoxicity assays. A hybridoma clone (83.21) was isolated that secreted antibodies which preferentially bound to several prostate and bladder cancer cell lines but did not bind to a variety of other normal and malignant human cell lines. This antibody also reacted with a cytomegalovirus-transformed human embryonic lung cell line but not to normal human embryonic lung cells. Quantitative adsorption studies demonstrated that the 83.21 monoclonal antibody was strongly reactive to membrane preparations from human prostate adenocarcinoma tissue and a liver metastasis of prostate carcinoma. Little or no binding activity was observed against two other prostate carcinomas, bening prostatic hyperplasia, normal prostate, or normal liver. Binding studies indicate that the 83.21 monoclonal antibody does not bind to alpha-fetoprotein, carcinoembryonic antigen, prostatic acid phosphatase, human leukocyte antigen, beta 2-microglobulin, HLA-Dr antigens, fibronectin, or prostate antigen. The data indicate that we have isolated a monoclonal antibody that binds to an antigen(s) expressed by several urogenital carcinoma cell lines as well as human prostate tumor tissue and that the antibody is not directed against well-known human tumor cell markers.