A Dual-Monoclonal Sandwich Assay forProstate-specific Membrane Antigen: Levels in

Tissues, Seminal Fluid and Urine

R.L. Sokoloff,* K.C. Norton, C.L. Gasior, K.M. Marker, and L.S. Grauer

Hybritech Incorporated, A Subsidiary of Beckman Coulter, Inc., San Diego, California

BACKGROUND. Prostate-specific membrane antigen (PSMA) is a 750-residue integral mem-brane glycoprotein and the target of an in-vivo imaging agent for metastatic prostate carci-noma (PCa). PSMA expression in normal and diseased prostatic tissues has previously beendemonstrated by immunohistochemical techniques. In order to quantify PSMA levels in tissuehomogenates and physiological fluids, we have developed a dual monoclonal antibody(mAb) sandwich assay which detects the antigen at a sensitivity <1 ng/mL and which is linearacross the working range 0–50 ng/mL.METHODS. The assay involves capture of the PSMA by a biotinylated mAb (7E11) immo-bilized onto a streptavidin-coated microtiter plate; this mAb binds to the N-terminus of theantigen. The captured PSMA is detected by an Eu-labelled mAb (PEQ226) which binds in theregion corresponding to Residues 134–437 of the antigen. PSMA was purified from LNCaPcells by immunoaffinity chromatography, and used as a calibrator, based on its concentrationby the bicinchoninic acid (BCA) protein assay.RESULTS. The assay was applied to a panel of normal and tumor tissues. Levels were highestin the prostate tissues (292–4254 ng/mg protein). Low levels (21–51 ng/mL) were observed inmembranes from ovary and breast, and neglible levels (1–10 ng/mg) in membranes from skin,liver, intestine, and kidney. Levels in the corresponding cytosol fractions were 20-to 50-foldlower. The average PSMA level in seminal fluid from 21 donors was 9,012 ng/mL. Onaverage, levels in normal-male urine (3.47 ng/mL) were ten-fold higher than in normal-female urine (0.3 ng/mL).CONCLUSIONS. This report is the first to describe absolute quantitation of PSMA in tissuesand fluids. Congruent with earlier tissue studies based on immunohistochemical staining andWestern-blot analysis, prostate tissue membranes expressed the highest levels of PSMA.Prostate 43:150–157, 2000. © 2000 Wiley-Liss, Inc.

KEY WORDS: prostate-specific membrane antigen (PSMA); immunoassay; 7E11; urine;seminal fluid; tissues; prostate cancer

INTRODUCTION

Prostate cancer (PCa) is the most common cancer inmales, and the second prevalent cause of male-deathsin the United States. Standard medical practice for theearly detection of PCa includes quantitation of pros-tate-specific antigen (PSA) serum levels in combina-tion with digital rectal exam (DRE). [1] The synergisticcombination of these two methods has resulted in in-creased detection rates of treatable PCa. However,three current issues in the screening detection of PCaare a) some prostate carcinomas still elude detectionby these methods until they become incurable, b) bi-

opsy findings support a cancer diagnosis in only 20%cases in which DRE is positive or PSA is between therange 4–10 ng/mL, and c) current detection methodspresent difficulties in discerning clinically significantfrom insignificant prostate tumors. New assays to aidin the early detection or characterization of PCa in-clude free-PSA and human kallikrein-2 (hK-2) [2,3].

*Correspondence to: Roger Sokoloff, Ph.D., Beckman-Coulter, POBox 269006, San Diego, CA 92196-9006.E-mail: RLSokoloff@Beckman.comReceived 29 September 1999; Accepted 13 December 2000

The Prostate 43:150–157 (2000)

© 2000 Wiley-Liss, Inc.

Although these assays are informative, a need still ex-ists for better detection methods.

One circulating protein that has potential utility forcancer detection is prostate-specific membrane anti-gen (PSMA), a 750-residue, 100-kDa Type 2 integralmembrane glycoprotein [4,5]. Immunohistochemical(IHC) staining studies with the anti-PSMA monoclo-nal antibody (mAb) 7E11-C5.3 on human tissue speci-mens from different organs demonstrated that PSMAexpression is highly restricted to the prostate, and thatstaining is more extensive in sections from high-grademalignant tumor or metastases compared to normal orbenign tissues [6–9]. High expression levels in prostatehave also been demonstrated by Western blot analysisand by ribonuclease protection assays [10,11].

Certain nonprostatic tissues, i.e. kidney and intes-tine, and bladder carcinoma, have demonstrated weak7E11-IHC staining [8]. Very low–yet significant–expression levels have also been detected in brain,salivary gland, kidney and small intestine by Westernblot [10]. In serum, correlation of PSMA serum levelswith aggressive, androgen-dependent PCa has beendemonstrated by semiquantitative immunoassay andWestern blot analysis [4,12]. However, one group wasunable to replicate these results [10].

The development, characteristics and potentialclinical utility of PSMA and mAb7E11 have recentlybeen reviewed [13–15]. As described in this report, wehave incorporated mAb-7E11 along with a newPSMA-specific antibody into a dual-mAb sandwichassay for quantitating PSMA in human fluids and tis-sue extracts.

MATERIALS AND METHODS

Monoclonal Antibodies

The preparation and properties of mAbs 7E11 andPEQ226 have previously been described. MAb 7E11,was developed from mice immunized with mem-branes from LNCaP cells–a prostate cancer (PCa) cellline established from cells obtained from a lymphnode of a patient with metastatic PCa-and binds to theintracellular, N-terminus of PSMA [4,16]. We biotinyl-ated 7E11 (purified product from Cytogen Corpora-tion, Princeton, NJ) using ImmunoPuret NHS-LC-Biotin II reagent (Pierce Chemical Company,Rockford, IL) at a 20:1 molar presentation ratio (bioti-nylation reagent:mAb) in 100 mM NaHCO3 based onpreviously published procedures [17].

MAb PEQ226, was developed from immunizationswith plasma membranes isolated from human PCa tu-mors , and subsequently shown to react with the ex-tracellular region of PSMA [18] Studies have demon-strated that PEQ226 binds the polypeptide backbone

between residues 134–437, and that indium-labelledPEQ226 can also image PCa metastases (unpublishedin-house studies). PEQ226 used in this assay was la-belled using Europium Labelling Reagent (Wallac OY,Turku, Finland) [19]. The resulting europium (Eu) la-belling density was 7.1 Eu atoms per mAb molecule.

PSMA Calibrator

PSMA was isolated from LNCap cells that weregrown and harvested as previously described [18].Harvested cells were sedimented, and the 5-mL pelletwas washed and homogenized in Extraction Solution(25 mM Trizma/150 mM NaCl/1% reduced TritonX-100, pH 7.4) with the aid of a Dounce homogenizer.The homogenate was stirred at 4–8°C for 12–18 h andcentrifuged, and the supernatant collected after pas-sage through a 0.2-mm pore size diameter membrane.PSMA from the filtrate was then immunoadsorbed toa matrix previously prepared by conjugating mAb-7E11 to Amino-Link affinity beads (Pierce ChemicalCompany) according to manufacturer’s instructions[20]. This adsorption was performed by incubating thefiltrate with 0.5 mL settled beads 12–18 h at 4°C withrotation. The beads were exhaustively washed andequilibrated in Extraction Solution, and then trans-ferred to an open column. PSMA was eluted with 100mM glycine/150 mM NaCl/1% reduced Triton X-100,pH 2.5. One- milliliter fractions were collected intotubes containing 100 mL 1.0 M Trizma, pH 8.0. Proteinconcentrations were determined by the method basedon bicinchoninic acid (BCA assay) [21]. Based on silverstained SDS-PAGE gels (Pharmacia PHAST-system),the first collected fraction contained the bulk of elutedPSMA at >90% estimated purity and was used to pre-pare the calibrators.

Specimens and Control

The assay was used to quantitate PSMA in humantissue extracts, seminal fluid and urine specimens.Prostate tissues were obtained either at autopsy orafter tissue microdissection following radical prosta-tectomy. Of the ten prostate tissues assayed five werepathologically confirmed cancers, one was benign dis-ease, and four were normal. Nonprostatic tissues wereobtained at autopsy within ten hours post-mortem,immediately flash-frozen in liquid nitrogen, andstored at −80°C until subsequent processing. Crudemembrane and cytosol fractions from these tissueswere prepared according to a previously publishedprocedure [22]. Specimens were homogenized in 5volumes 30 mM NaCl/1 mM EDTA/1 mM PMSF/10mM Tris-HCl, pH 7.2 at 4°C in a polytron homog-enizer. Crude fractions were then prepared by differ-

Sandwich Assay for PSMA 151

ential centrifugation, aliquoted and stored at −70°C.Protein concentrations in these fractions were subse-quently assessed by the BCA assay.

Urine and seminal fluid specimens from normal in-dividuals were collected, aliquoted and frozen. In ad-dition, a seminal fluid pool was prepared and alsoaliquoted and stored at −70°C for use as an assay con-trol.

PSMA Assay

All steps of the assay were performed at ambienttemperature. Mixtures were covered and agitated on amicroplate shaker during all incubation steps. Washsteps were performed 4 to 8 times with a solution of0.1% Tween-20 in PBS. All samples were assayed intriplicate on 96-well streptavidin-coated microplates(Wallac OY), and each microplate included a set ofcalibrators and controls in triplicate against which thecorresponding samples on the plate were quantitated.

Sequential steps were as follows: A 100-mL aliquotof the biotinylated 7E11 solution at 5 mg/mL in AssayDiluent (10% bovine serum albumin/0.02% Poly-MAK (Boehringer-Mannheim, Indianapolis IN)/0.1%mannitol/0/1% sodium azide/100 mM sodium ci-trate-phosphate, pH 7.0) was added to each well of thestreptavidin-coated plates. Following 1-h incubation,mixtures were decanted and discarded, and plateswashed and blotted to remove excess fluid. Freshlythawed aliquots of specimens, calibrators and seminalfluid control were diluted with 1% NP40 detergent inAssay Diluent, and a 100-mL aliquot of the appropriatemixture was added to each well in triplicate. After a3-h incubation, mixtures were decanted and dis-carded, and plates washed and blotted. A solution ofeuropim-labelled PEQ226 (2 mg/mL) in DELFIAt As-say Buffer (Wallac OY) was prepared and a 100-mLaliquot added to each well. Mixtures were incubated1-h, then mixtures decanted and discarded, and plateswashed and blotted. A 125-mL aliquot of DELFIAtEnhancement Buffer (Wallac) was added to each well.Following a minimum 5-minute incubation, fluores-cence was assessed in a Model 1234 DELFIAt Re-search Fluorometer.

Linearity over the working range, longitudinal re-covery of the diluted seminal fluid control, and mini-mal detectable concentration (MDC) were measuresused to assess assay performance. To determine MDC,84 replicate aliquots of calibrator diluent–i.e. 1% NP40in BSA/PBS containing no PSMA–were assayed on asingle plate. MDC was calculated as the concentrationcorresponding to two standard deviations of the 84replicates.

RESULTS

Assay Performance

The assay was consistently linear within its work-ing range of 0 to 50 ng PSMA/mL (Fig. 1). Based ondeterminations on replicate aliquots of the calibratordiluent, MDC was estimated to be 80 pg/mL. Re-peated quantitation on the pooled seminal fluid ali-quots demonstrated the longitudinal reproducibilityof the assay (Fig. 2). The average intraassay variabilityfor triplicate mixtures was generally less than 5%.

PSMA in Tissues and Fluids

PSMA levels relative to total protein were higher inprostate membranes (292–4,254 ng/mg protein) thanin membranes from nonprostatic tissues (1.8 to 51 ng/mL). (See Table I.) There was no statistically signifi-cant difference between normal, cancerous and be-nign-diseased prostates. Of the nonprostaticmembranes, ovary displayed the highest levels (51ng/mg), followed by cancerous breast and normalbreast (43 and 21 ng/mg respectively). No significantPSMA levels were detected in the other normal tissuesexamined by the assay.

Seminal fluid PSMA concentrations from 21 normaldonors ranged from 3,523 to 18,780 ng/mL, with a

Fig. 1. Calibration curves from five representative assays. Ali-quots of stock calibrator stock (LNCaP PSMA) were prepared andstored at −70°C. Prior to each assay, an aliquot was thawed,diluted to five working levels in triplicate, and assayed on each offive separate days. The calibration curve was consistently linearover the working range (0–50 ng PSMA/mL) of the assay; in allcases, the correlation coefficient (r) of PSMA (x) compared tofluorescent counts (y) was $ 0.993.

152 Sokoloff et al.

mean of 9,012 ng/mL and standard deviation of 4,190ng/mL (Fig. 3). Prostate-specific antigen (PSA) con-centrations ranged from 0.19 to 3.02 mg/mL with amean of 1.29 mg/mL and standard deviation of 0.69mg/mL (Fig. 3). PSMA levels loosely correlated withthose of PSA (r = 0.64), although in general PSA levelswere 100- to 200-fold higher than correspondingPSMA levels (Fig. 3).

On average, urine PSMA levels were about 10-foldhigher for normal males (3.47 ng/mL) compared tofemales (0.3 ng/mL, Fig. 4). The difference was statis-tically significant (P < 0.0001).

DISCUSSION

We report here a dual-monoclonal sandwich immu-noassay for the absolute quantitation of PSMA. Theassay employs two mAbs with previously demon-strated high specificity for PSMA. One of these, mAb-7E11, is a component of a commercially available im-aging reagent for locating PCa metastases. This mAbwas developed against membranes isolated fromLNCaP cells, and subsequently shown on Western

Fig. 2. Recovery of the seminal fluidcontrol over a 12-month period. Aliquotsof pooled seminal fluid for use as a controlwere prepared and stored at −70°C. Priorto each assay, an aliquot of pooled seminalfluid was removed from storage, thawed,diluted to within the working range of theassay (1/500) and assayed against theworking calibrators. The assay reportedconsistent PSMA concentrations in theseminal fluid control throughout the year.

TABLE I. PSMA Levels in Tissue Membranes

Organ N[PSMA]/[protein]

(ng/mg)

Prostate, normal 4 519–4,254Prostate, cancer 5 292–4,196Prostate, benign disease 1 500Ovary 1 51Breast, cancer 1 43Breast, normal 1 21Small Intestine 1 16Large intestine 1 8.3Skin 1 8.1Liver 1 2.7Bone 1 2.6Kidney 1 1.8

Fig. 3. PSA versus PSMA in the seminal fluid specimens from 21patients. Seminal fluid specimens were collected from 21 normal-male donors and placed in −70°C storage within 3 h of collection.Samples were later thawed and centrifuged in a microfuge (Rainin,10,000 rpm, 5 min). The supernatants were recovered, and ali-quots were removed, diluted into the working ranges of the theassays for PSA and PSMA. Each point on the graph represents thePSA concentration and corresponding PSMA concentration for aparticular seminal fluid specimen. PSMA concentrations rangedfrom 3,712 to 18,780 ng/mL, with a mean of 9,012 ng/mL, medianof 8,326 ng/mL, standard deviation of 4,190 ng/mL, and 95%-confidence interval of 7,261 to 10,762 ng/mL. PSA concentrationswere determined by the Tandem PSA-MP assay (Hybritech, Incor-porated). PSA concentrations ranged from 0.19 to 3.02 mg/mLwith a mean of 1.29 mg/mL, median of 1.28 mg/mL, standarddeviation of 0.69 mg/mL, and 95%-confidence interval of 1.00 to1.58 mg/mL. Although a line of best fit was determined by linearregression ([PSMA] = 0.00304 × [PSA] + 0.051), PSMA correlatedonly loosely with PSA (r = 0.504).

Sandwich Assay for PSMA 153

blots to react against a 100-kDa protein subsequentlyidentified as PSMA [4,12]. IHC staining studies dem-onstrated strong reactivity of 7E11 with prostate epi-thelium, and negligible reactivity with other organsand tissue; these findings suggested that PSMA ishighly restricted to the prostate. Many studies havereported the use of 7E11 to explore the expressionparameters of PSMA [4,6–12].

The other mAb component, PEQ226, was devel-oped in-house and also reacts with PSMA and hasbeen applied to imaging PCa metastases in researchstudies. Yet, in contrast to 7E11, PEQ226 was devel-oped in mice immunized with PCa cells and reactswith the extracellular domain of PSMA. The highlyrestricted activity of these mAbs, when partnered to-gether, imparts high specificity to the sandwich assay.

Previous qualitative IHC and Western-blot studieswith mAb-7E11 on tissue sections from various organsdemonstrated that PSMA protein expression is highly

restricted to the prostate gland, although low-level ex-traprostatic expression has also been reported. For ex-ample, Horoszewicz et al., in the earliest report onPSMA studied 122 individual specimen sections from28 different, normal human organs in IHC studieswith 7E11 [4]. Epithelial staining occurred in all PCafoci in prostate (9/9), PCa metastases from lymphnode (2/2), and normal prostate (9/9 ); only a portionbenign prostate hyperplasia specimens stained (5/7).†Of the remaining 95 nonprostatic specimens, nonestained. Examining 33 normal and tumor cell lines, thesame authors reported that only the LNCaP cell linewas reactive [4]. Silver et al, also observed 7E11-IHCepithelial staining in prostate tissues [8]. Light, diffuse7E11-staining of normal non-prostatic tissues were ob-served only in duodenal mucosa and in portions ofproximal renal tubules and neuroendocrine cells incolonic crypts. However, endothelial staining was ob-served in a fraction of malignant specimens from renalcell carcinoma (8/17), bladder transitional cell carci-noma (7/13) and colon adenocarcinoma (3/19) [8].Congruent with Silver’s observations in nonprostatictumors, Liu, et al., applied a panel of mAbs developedagainst the extracellular domain of PSMA, and ob-served higher IHC staining in vascular endothelium incarcinomas from lung, colon, breast and others; nor-mal tissues did not stain [23]. Thus, it appears thatincreased expression of a PSMA ubiquitous in vascu-lar endothelium from a wide variety of cancers.Troyer, et al. applied Western blot analysis, probingwith 7E11, to extracts of various organs and body flu-ids, and observed a high molecular staining (>120kDa) in normal, benign and malignant prostate tis-sues, and in seminal fluid. (10) Less intense, yet sig-nificant bands were observed in sections from brain,small intestine, and salivary gland. Bands were ob-served on Western blots of malignant tissues from co-lon, lung, bladder, liver and breast, but were absent incorresponding normal tissues. Using a ribonucleaseprotection assay, Israeli, et al., demonstrated that ex-pression of PSMA mRNA was highly restricted toprostate, although faint bands were seen on the North-ern blots in extracts from brain and salivary gland [11].No evidence for PSMA mRNA expression was ob-served in extracts from kidney, liver, lung, mammarygland, pancreas, placenta, salivary gland, skeletalmuscle, spleen or testis.

Whereas both normal, benign and early-stage can-cerous prostate tissues have displayed similar 7E11-IHC staining characteristics, more intense or extensivestaining has been demonstrated in advanced PCaspecimens. Wright, et al., devised a stain index param-eter, based on the number of stained cells and stainingintensity to compare IHC results. For prostate speci-mens, the average index was lowest for BPH (Index =

Fig. 4. PSMA in male-urine specimens compared to female-urinespecimens. Morning first-void urine specimens were collectedfrom 58 normal-female and 62 normal-male donors, and an aliquotwas removed and stored at −70°C. On the day of the assay,aliquots were removed, thawed, diluted 1/2 and assayed for PSMA.PSMA levels were about 10-fold higher in the male group com-pared to the female group. Summary statistics for the male-urinegroup are as follows: mean, 3.47 ng/mL; median, 2.28 ng/mL; range,0.23 to 23.15 ng/mL. PSMA levels for the female-urine group aresummarized as follows: mean, 0.21 ng/mL; median, 0.14 ng/mL;range, 0.00 to 1.02 ng/mL. The difference between the two groupswas statistically significant (P < 0.005).

154 Sokoloff et al.

52); comparable for PIN (130), primary PCa (133), andnormal prostate (146); and highest for metastases fromlymph node (194) and bone (258) [6]. Silver, et al.,noted intense 7E11-IHC staining in most primary PCaspecimens (33/35) and lymph node metastases (7/8),but in only 44% of bone metastases (8/18) [8]. Basedon sections from 184 radical prostatectomies, Bostwicket al., estimated the proportion of 7E11-reactive cells tobe lowest in benign conditions (69.5% cells stained),compared to high grade PIN (77.9%) and PCa (80.2%)[7]. However, virtually all epithelial cells stained inhigh grade PCa (i.e., Gleason primary of 4 and 5).Sweat et al., in specimens from 232 patients with node-positive adenocarcinoma, tabulated separately thenumber of 7E11-staining cells and staining intensityfrom both prostate and lymph node metastases [9].Observed order of intensity (lowest to highest) waslymph node metastases, to benign epithelium, to pri-mary cancer. Benign epithelium displayed the lowestnumber of 7E11-stained cells; primary cancer andlymph node metastases, the same number.

Congruent with these earlier studies, the PSMA as-say described here demonstrated highest PSMA levelsin membranes from prostate vis-a-vis other organ tis-sues. Moreover, we were able to assess PSMA levels inabsolute amounts (i.e., ng/mg protein) against a cali-bration standard derived from PSMA purified fromLNCaP cells. Our results demonstrated that PSMAlevels in normal and disease prostate (relative to totalprotein) are at least 50-fold higher than in nonprostatictissues we studied.

PSMA levels in seminal fluid and urine are alsoreported here for the first time. PSMA in male-urinespecimens were measurable and significantly higherthan in female-urine specimens. The likely source ofurinary PSMA in males is the prostate, although onecannot rule out the possibility that other tissues rep-resenting low-density, high tissue volume expressionof PSMA may contribute to urinary total PSMA levels.Extraprostatic production of the analyte may explainour finding of PSMA–albeit at low levels–in normal-female urine. One possible extraprostatic source maybe bladder. Indeed, Gala et al., has demonstrated byNorthern blot hybridization studies the presence ofPSMA mRNA in extracts from normal bladder, blad-der carcinoma, and bladder tumor cell lines [24]. How-ever, we have examined bladder cancer cell lines byWestern-blots (probed with mAb-7E11 or PEQ226),and were unable to detect PSMA (unpublished data).It is also unclear whether prostate PSMA is introduceddirectly into the urine or is cleared from the blood viathe kidneys.

In seminal fluid, PSMA-although much lower thancorresponding PSA levels-loosely correlated withPSA. Reasons for this relationship are unknown; ex-

pression of both PSA and PSMA is highly restricted toprostate, and some of the co-variation in these ana-lytes across specimens may represent differences inthe proportion of prostatic fluid contributing to wholesemen. Despite similarities in name and specificity oforgan expression, PSA and PSMA are structurally andfunctionally different and display no sequence homol-ogy. PSA is a 30-kDa glycoprotein whose main func-tion appears to be cleavage of semenogelin in theejaculate; PSMA, which is three-fold larger, may func-tion is a carboxypeptidase [26–28]. In addition, theconcentration of PSA in seminal fluid is approxi-mately ten thousand-fold higher than correspondingPSMA levels

Interestingly, the single breast-cancer specimen inour study displayed a higher PSMA level compared tothe single normal specimen from breast. Although notstatistically significant, this observation is particularlyinteresting in view of earlier observations that PSA, isalso expressed in male and female breast tissues, andthat PSA-postive tumors correlated with progesteronereceptors [29]. Another reputed prostate-specific pro-tein, human glandular kallikrein protein (hK2), is alsoexpressed in the breast cancer cell line T47-D [30].

The physiological role of PSMA in prostate is anopen question, and some possibilities are suggested byreported roles of PSMA-like proteins studied in othertissues. Recent evidence presented by Luthi-Carter etal. has indicated that PSMA is highly homologous toNAALADase–an enzyme in brain that appears tomodulate glutamate-based neurotransmissionthrough its peptidase activity on N–acetylaspartylglu-tamic acid (NAAG) with release of glutamic acid [28].Using RTPCR and Northern blot techinques, these au-thors characterized a mRNA from brain that was in-distinguishable from the mRNA of LNCaP PSMA.Moreover, detergent extracts of brain membranes andcloned LNCaP PSMA displayed similar enzyme kinet-ics on NAAG, and were similarly inhibited by a panelof inhibitors. In addition, NAALADase activity wasimmunoprecipitated from brain-membrane extractsby mAb-7E11 that had been immobilized onto mi-crobead particles. The authors proposed that NAALA-Dase/PSMA may have potential value as a therapeu-tic target in various neurological diseases, e.g.amyolateralsclerosis (ALS).

The intestinal, membrane-bound enzyme foly-(poly-g-glutamate) carboxypeptidase (FpGC) appearsis similar to PSMA in its immunoreactivity, structural,and enzyme characteristics [26,27]. This enzyme cata-lyzes the hydolyzes of FpGC resulting–as the casewith NAALADase–in the release of free glutamate.However, intestinal FGC’s importance appears to tobe its involvement in the adsorption, intracellular re-tention and metabolism of FpG. Although the role of

Sandwich Assay for PSMA 155

PSMA in normal prostate health and in the etiology ofprostate cancer is still unresolved, Heston suggeststhat the FpGC-like activity of prostate PSMA may rep-resent its primary role in prostate [27]. Folate-deficiency has been implicated in the etiology of somecancers; some prostate cancers may result from imbal-ances in concentrations of the mono/poly-g-glutamylfolate forms, as mediated by PSMA [27]. Con-versely, Pinto, et al., have reported that PSMA inprostate tumor cell lines also displays folate hydrolaseactivity [31].

These reports suggest that prostatic PSMA, brainNAALADase, intestinal FpGC and the vascular pro-tein directly observed in cancerous tissue, may repre-sent the same gene product–i.e., single protein with adifferent role in each of these organs. The neglibleIHC-staining of various nonprostatic tissue speci-mens, and the evidence indicating identity ofNAALADase and FpGC to PSMA suggests that extra-prostatic sources may represent a low-density, large-mass contributor to the total level of circulatingPSMA. Therefore, methods to quantitate serum PSMAas an adjuvant to assess PCa must include the possi-bility that non-prostatic sources contribute to serumlevels. Serum measurements based on competitive im-munoassay and Western-blot analysis that have dem-onstrated a correlation between PSMA levels and PCaaggression, have also contended with high baselinelevels, as exemplified by measurable amounts in nor-mal patients [10]. Using the sandwich assay describedhere, we have observed comparable analyte concen-trations in serum specimen from normal males andfemales (unpublished). The potential utility of serumPSMA measurements may subsequently be maxi-mized by an assay which specificly recognizes theprostate-form of PSMA, and is unreactive with ex-tracts from other organs.

PSM’ is a truncated splice variant of PSMA that hasbeen reported in LNCaP cells and prostate tissue[18,32]. RT-PCR studies indicate that PSM’ m-RNAwould translate into a protein corresponding to PSMAresidues 59–750 [32]. The expression ratio of PSM’ tofull-length may be a useful prognostic marker to dis-cern malignant from non-malignant disease [32]. TheN-terminus region of full-length PSMA, which is lack-ing in PSM’, includes both the membrane-anchoringdomain and the intracellular binding region for 7E11.PSM’ is not detectable by the assay described in thisreport. Recently, Grauer et al. have isolated a trun-cated protein from LNCaP cells that lacks the aminoterminus of the full-length PSMA [18].

In conclusion, an immunoassay has been developedthat is capable of detecting PSMA in biological fluidsand extracts at subnanogram levels. This assay wassuccessfully applied to quantitate PSMA levels in

crude membrane and cytosol fractions from varioustissues, in highly diluted seminal fluid specimensfrom normal patients, and in urine specimens. Theassay has potential utility in exploring PSMA expres-sion in diseases of the prostate.

REFERENCES

1. Catalona WJ, Richie JP, Ahmann FR, Hudson MA, Scardino PT,Flanigan RC, deKernion JB, Ratliff TL, Kavoussi LR, Dalkin BL,Waters WB, MacFarlane MT, Southwick PC. Comparison ofdigital rectal examination and serum prostate-specific antigen inthe early detection or prostate cancer: results of a multicenterclinical trial of 6,630 men. J Urol 1994;151:1283–1290.

2. Catalona WJ, Partin AW, Slawin KM, Brawer MK, Flanigan RC,Patel A, Richie JP, deKernion JB, Walsh PC, Scardino PT, LangePH, Subong ENP, Parson RE, Gasior GH, Loveland KG, South-wick PC. Use of the Percentage of free prostate-specific antigento enhance differentiation of prostate cancer from benign pros-tatic disease. JAMA 1998;279:1542–1547.

3. Catalona WJ, Partin AW, Chan DW, Tindall DJ, Young CY-F,Klee GG, Darte C, Finlay JA, Rittenhouse HG, Wolfert RL, Woo-drum DL. Detection of prostate cancer with %FPSA and hK2when PSA is 2–4 ng/mL. J Urol 1999;161 Suppl 4:207.

4. Horoszewicz JS, Kawinski E, Murphy GP. Monoclonal antibod-ies to a new antigenic marker in epithelial prostatic cells andserum of prostatic cancer patients. Anticancer Res 1993;7:927–936.

5. Israeli RS, Powell CT, Fair WR, Heston WDW. Molecular clon-ing of a complementary DNA encoding a prostate-specificmembrane antigen. Cancer Res 1993;53:227–230.

6. Wright G, Haley C, Beckett ML, Schellhammer, PF. Expressionof prostate-specific membrane antigen in normal, benign, andmalignant prostate tissues. Urol Oncol 1995;1:18–28.

7. Bostwick DG, Pacelli A, Blute M, Roche P, Murphy GP. Prostatespecific membrane antigen expression in prostatic intraepithe-lial neoplasia and adenocarcinoma. Cancer 1997;82:2256–2261.

8. Silver DA, Pellicer I, Fair WR, Heston WDW, Cordon-Cardo C.Prostate-specific membrane antigen expression in normal andmalignant human tissues. Clin Cancer Res 1997;3:81–85.

9. Sweat SD, Pacelli A, Murphy GP, Bostwick DG. Prostate-specificmembrane antigen expression is greatest in prostate adeoncar-cinoma and lymph node metastases. Adult Urol 1998;52:1–4.

10. Troyer JK, Beckett ML, Wright GL. Detection and characteriza-tion of the prostate-specific membrane antigen (PSMA) in tissueextracts and body fluids. Int J Cancer 1995;62:552–558.

11. Israeli RS, Powell CT, Corr JG, Fair WR, Heston WDW. Expres-sion of the prostate-specific membrane antigen. Cancer Res1994;54:1807–1811.

12. Rochon YP, Horoszewicz JS, Boynton AL, Holmes EH, BarrenRJ, Erickson SJ, Kenny GM, Murphy GP. Western blot assay forprostate-specific membrane antigen in serum of prostate cancerpatients. Prostate 1994;25:219–223.

13. Fair WR, Israeli RS, Heston WDW. Prostate-specific membraneantigen. Prostate 1997;32:140–148.

14. Gregorakis AK, Holmes EH, Murphy GP. Prostate-specificmembrane antigen: current and future utility. Sem Urol Oncol1998;16:2–12.

15. Murphy GP, Elgamal A-Z A, Su S, Bostwick DG, Holmes EH.Current evaluation of the tissue localization and diagnostic util-ity of prostate specific membrane antigen. Cancer 1998;8:2259–2269.

16. Troyer JK, Qi F, Beckett ML, Wright GL. Biochemical character-

156 Sokoloff et al.

ization and mapping of the 7E11-C5.3 epitope of the prostate-specific membrane antigen. Urol Oncol 1995;1:29–37.

17. Gretch DR, Suter M, Stinski MF. The use of biotinylated mono-clonal antibodies and streptavidin-affinity chromatography toisolate herpes virus hydrophobic proteins or glycoproteins.Anal Biochem 1987;162:270–277.

18. Grauer LS, Lawler KD, Marignac JL, Kumar A, Goel AS, WolfertRL. Identification, purification and subcellular localization ofPSM’ protein in the LNCaP prostatic carcinoma cell line. CancerRes 1998;58:4787–4789.

19. Hemmila I, Dakubu S, Mukkala V-M, Siitari H, Lovgren T. Eu-ropium as a label in time-resolved immunofluorometric assays.Anal Biochem 1984;137:335–343.

20. Instructions, AminoLinkt Coupling Gel: Pierce Chemical Com-pany.

21. Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH,Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk DC.Measurement of protein using bicinchoninic acid. Anal Biochem1985;150:76–85.

22. Howard FD, Ledbetter JA, Mehdi SQ, Herzenberge LA. A rapidmethod for the detection of antibodies to cell surface antigens: Asolid phase radioimmunoassay using cell membranes. J Immu-nol Methods 1980;38:75–84.

23. Liu H, Rajasekaran AK, Moy P, Xia Y, Kim S, Navarro V, Rah-mati R, Bander NH. Constitutive and antibody induced inter-nalization of prostate-specific membrane antigen. Cancer Res1998;58:4055–4060.

24. Gala JL, Loric S, Brasseur, Delhez H, Heusterpreute M, DumasF, Berteau P, Tombal B, Van Cangh P, Eschwege P, de Nayer P,

Philippe M. PSMA expression in bladder cell carcinoma. J Urol1998;159 Suppl 5:287.

25. Lilja H. A kallikrein-like serine protease in prostatic fluidcleaves the predominant seminal vesicle protein. J Clin Invest1985;76:1899–1903.

26. Heston WDW. Characterization and glutamyl preferring carbo-cypeptidase function of prostate specific membrane antigen: Anovel folate hydrolase. Urology 1997; 49 Suppl 3A:104–112.

27. Halsted CH, Ling E-h, Luthi-Carter R, Villanueva JA, GardnerJM, Coyle JT. Folylpoly-g-glutamate carboxypeptidase from pigjejunum. J Biol Chem 1998;273:20417–20424.

28. Luthi-Carter R, Barczak AK, Speno H, Coyle JT. Molecular char-acterization of human brain N-acetylated a-linked acidic dipep-tidase (NAALADase). J Pharmacology Exp Thera 1998;286:1020–1025.

29. Diamandis EP, Yu H, Sutherland DJA. Detection of prostatespecific antigen immunoreactivity in breast tumors. Breast Can-cer Res Treat 1994;32:301–310.

30. Hsieh M-L, Charlesworth MC, Goodmanson M, Zhang S, SeayT, Klee GG, Tindall DJ, Young CYF. Expression of human pros-tate-specific glandular kallikrein protein (hK2) in the breast can-cer cell line T47-D. Cancer Res 1997;57:2651–2656.

31. Pinto JT, Suffoletto BP, Berzin TM. Prostate-specific membraneantigen: A novel folate hydrolase in human prostatic carcinomacells. Clin Cancer Res 1996; 2:1445–1451.

32. Su SL, Huang I-P, Fair WR, Powell CT, Heston WDW. Alterna-tively spliced variants of prostate-specific membrane antigenRNA: ratio of expression as a potential measurement of pro-gression. Cancer Res 1995;55:1441–1443.

Sandwich Assay for PSMA 157