Characterization and immunological determination
of the complex between prostate-specific antigen
Wan-Ming Zhang,1*Patrik Finne,1Jari Leinonen,1Satu Vesalainen,2Stig Nordling,3
Sakari Rannikko,2and Ulf-H˚ akan Stenman1
Prostate-specific antigen (PSA) rapidly forms a complex
with ?2-macroglobulin (A2M) in vitro; however, PSA
complexed with A2M (PSA-A2M) is not detected by
conventional immunoassays for PSA because it is en-
capsulated by the A2M. In this study, we show that
denaturation of PSA-A2M at high pH renders PSA
immunoreactive. Part of the complexed PSA is released
in free form and part remains bound to denatured A2M.
These forms can be measured by a conventional immu-
noassay for PSA. This finding enabled us to design a
dissociation assay for the detection of PSA-A2M, which
was based on the removal of immunoreactive PSA in
serum by immunoadsorption, denaturation of PSA-A2M
at high pH, and measurement of the released PSA
immunoreactivity by a conventional PSA immunoassay.
This PSA-A2M assay was calibrated with PSA-A2M
formed in vitro. The detection limit of the assay was 0.14
?g/L. Inter- and intraassay coefficients variation were
4–9% and 8–14%, respectively. When purified PSA was
incubated with A2M, the loss of PSA immunoreactivity
was highly correlated with the PSA-A2M formed, as
measured by the dissociation assay for PSA-A2M (r ?
0.99; P <0.0001). The concentration of PSA-A2M in
serum correlated with that of total PSA both in prostate
cancer (PCa) and benign prostatic hyperplasia (BPH);
however, the ratio of PSA-A2M in relation to total PSA
was significantly higher in BPH than in PCa (P <0.0003).
ROC curve analysis suggested that measurement of the
ratio of PSA-A2M to total PSA in serum improves the
diagnostic accuracy for PCa compared with assays for
total PSA only.
Prostate-specific antigen (PSA)4is a 30-kDa single chain
glycoprotein produced mainly by the prostatic epithelium
(1–3). PSA is a serine protease with chymotrypsin-like
enzymatic activity and a member of the glandular kal-
likrein family (4–6). In vitro PSA forms complexes with
protease inhibitors such as ?2-macroglobulin (A2M), preg-
nancy zone protein, ?1-antichymotrypsin (ACT), ?1-pro-
tease inhibitor (API), and protein C inhibitor (6–9).
In serum, most of the immunoreactive PSA occurs in
complex with ACT (PSA-ACT); the rest is either free or in
complex with API [PSA-API; Zhang et al., manuscript
submitted, and Refs. (10, 11)]. Five major antigenic re-
gions have been identified on the PSA molecule, only one
of which is covered by ACT in PSA-ACT (12). The
PSA-ACT and PSA-API complexes are readily detected by
specific sandwich assays or by conventional PSA immu-
noassays [Zhang et al., manuscript submitted, and Refs.
(8–14)]. Specific measurement of complexed and free PSA
in serum improves the diagnostic accuracy for prostate
cancer (PCa) compared with assays of total PSA only
[Zhanget al., manuscript
(10, 13, 14)].
A2M is a tetramer assembled from pairwise disulfide-
bridged 180-kDa subunits, each subunit containing a bait
region, which is susceptible to cleavage by most pro-
teases, and a reactive internal ?-cysteinyl-?-glutamyl thiol
ester (15, 16). When A2M interacts with a protease, the
bait region of A2M is proteolytically cleaved, causing
activation of the thiol ester and covalent binding of the
protease to A2M, mainly through an ?-Lys-?-Glu bond
submitted, and Refs.
Departments of1Clinical Chemistry,2Urology, and3Pathology, Helsinki
University Central Hospital, FIN-00290, Helsinki, Finland.
*Author for correspondence. Fax 358-0-4714804; e-mail wmzhang@
Received April 15, 1998; revision accepted September 28, 1998.
4Nonstandard abbreviations: PSA, prostate-specific antigen; A2M, ?2-
macroglobulin; ACT, ?1-antichymotrypsin; API, ?1-protease inhibitor; PCa,
prostate cancer; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel
electrophoresis; BPH, benign prostatic hyperplasia; BSA, bovine serum albu-
min; MAb, monoclonal antibody; IFMA, immunofluorometric assay; and TBS,
Clinical Chemistry 44:12
Enzymes and Protein
(16–18). Simultaneously, the conformation of A2M is
changed from an electrophoretically “slow” S-form to a
“fast” F-form, and the protease is entrapped within the
A2M molecule (15, 16). One tetrameric A2M can bind two
protease molecules (17). The encapsulation of proteases
by A2M sterically hinders access of high-molecular weight
substance such as high-molecular weight inhibitors or
antibodies to the enzymes (19). Thus, the PSA-A2M com-
plex is not detected by conventional PSA immunoassays
(8, 11). When PSA-A2M is denatured with sodium dode-
cyl sulfate (SDS), PSA epitopes are exposed, rendering it
reactive with PSA antibodies (6, 20). PSA-A2M has been
detected qualitatively in male serum with high concentra-
tions of PSA by immunoblotting after SDS-polyacryl-
amide gel electrophoresis (SDS-PAGE) (20).
In this study, we developed a sensitive and quantita-
tive dissociation assay for PSA-A2M in serum, character-
ized PSA-A2M formed in vitro and endogenous PSA-A2M
in serum, and measured the concentrations of PSA-A2M
in serum from patients with PCa and benign prostatic
hyperplasia (BPH). The dissociation assay for PSA-A2M is
based on the removal of immunoreactive PSA, i.e., PSA-
ACT, PSA-API, and free PSA, by immunoadsorption,
denaturation of PSA-A2M at high pH, and measurement
of the PSA thus rendered immunoreactive.
Materials and Methods
Sera were obtained from 73 (median age, 67.6 years;
range, 50.3–96.1 years) patients referred with PCa, 58
(median age, 69.6 years; range, 51.9–81.9 years) referred
with BPH, and 46 healthy females. The diagnosis of PCa
was based on histological examination of tissues obtained
by biopsy or radical prostectomy; the diagnosis of BPH
was based on histological examination of tissues obtained
by transurethral resection of the prostate. The sera were
taken before initiation of therapy. All samples were kept
frozen at ?20 °C until analysis.
The Superdex-200 was from Pharmacia Biotech, the 3,3?-
diaminobenzidine tetrahydrochloride was from Sigma
Chemical Co., and the PVDF membrane (Immobilon P)
was from Millipore. The streptavidin-conjugated mag-
netic beads were from Promega. The NaOH was from
Reagena LTD, and the HCl was from Merck. The sulfo-
succinimidyl-6-(biotinamido)hexanoate was purchased
from Pierce. The trifluoroacetic acid was from Fluka. The
DELFIA assay buffer and enhancement solution used in
immunoassays were from Wallac.
The major isoenzyme of PSA was purified from human
seminal fluid (21); the A2M was purified from plasma as
described (22). Molecular mass protein markers were
from Amersham. The bovine serum albumin (BSA) was
A monoclonal antibody (MAb) to PSA (5E4) was pro-
duced by standard procedures (Leinonen et al., unpub-
lished results); H117 and H50 were kind gifts from B.
Dowell (Abbott Diagnostics, Abbott Park, IL). Polyclonal
antibodies to PSA and A2M and a peroxidase-conjugated
swine anti-rabbit IgG immunoglobulin were from Dako-
patts. MAb 5E4 and the polyclonal antibody to PSA were
biotinylated according to the manufacturer’s (Pierce) in-
structions. MAb H50 and the polyclonal IgG to A2M were
labeled with Eu3?as described (10).
Total PSA was determined by a time-resolved immun-
ofluorometric assay (IFMA) using MAb H117 as the
capture antibody and MAb H50 labeled with Eu3?as
tracer. The calibrators were prepared from pure PSA and
standardized as described (21). A sandwich assay utiliz-
ing the polyclonal antibody to PSA as the capture anti-
body and a polyclonal antibody to A2M labeled with Eu3?
as tracer was performed as described (10) and termed
“A2M antibody-based PSA-A2M assay” in this study. The
IFMA for A2M was performed as described (8).
Samples were loaded on a Superdex-200 column (60 ? 1.6
cm) and eluted with Tris-buffered saline (TBS; 50 mmol/L
Tris-HCl buffer, pH 7.4, containing 150 mmol/L NaCl and
8 mmol/L NaN3). The flow rate was 15 mL/h, and 2-mL
fractions were collected. The column was roughly cali-
brated by measuring the absorbance at 280 nm in the
fractions to identify the elution volumes of human IgG
(150 kDa) and human albumin (68 kDa).
electrophoresis and immunoblotting
SDS-PAGE was performed under reducing conditions
(23) in 10 ? 10 cm, 2 mm thick, 3–16% gradient polyacryl-
amide gels. After electrophoresis proteins were trans-
ferred eletrophoretically to immobilon P and incubated
with polyclonal antibodies to PSA or A2M (24). Bound
antibodies were detected with peroxidase-conjugated
swine anti-rabbit IgG immunoglobulin, using 3, 3?-diami-
nobenzidine tetrahydrochloride as the substrate.
complex formation between psa and a2m
Purified PSA (60 ?g) was incubated with 12 mg of A2M in
800 ?L of TBS containing 50 g/L BSA (1:10 molar ratio) at
37 °C. Aliquots of 100 ?L were taken at time intervals of 0,
1, 3, 6, 8, 24, 48, and 72 h. The aliquot taken at 72 h was
subjected to gel filtration. The fractions obtained and the
aliquots were further analyzed by immunoassays for PSA,
A2M, and PSA-A2M and by immunoblotting.
immunoadsorption of psa
Five micrograms of biotinylated MAb 5E4 (to PSA) was
incubated with 100 ?g of streptavidin-conjugated mag-
netic beads at 25 °C. After 30 min, unbound antibodies
Zhang et al.: Determination of PSA-A2M complex
were removed by washing the beads with assay buffer.
The beads (100 ?g) were incubated with 200 ?L of
PSA-A2M formed in vitro or male serum at 25 °C. After 30
min, the beads were collected with a magnet, and the
supernatants were analyzed for PSA and PSA-A2M.
denaturation of psa-a2m
PSA-A2M formed in vitro was purified by gel filtration
and then diluted in TBS containing 50 g/L BSA or pooled
female serum without PSA immunoreactivity. Diluted
PSA-A2M was incubated with NaOH at a final concentra-
tion range of 0–100 mmol/L at 25 °C. Aliquots withdrawn
at 0, 30, 60, and 180 min were measured for pH by an
electronic pH meter (PW4920; Phlips) and for PSA by the
PSA IFMA. As a control, purified PSA diluted in TBS
containing 50 g/L BSA was treated in the same way.
PSA-A2M incubated with NaOH was further separated by
gel filtration, and fractions obtained were analyzed by the
PSA IFMA and the A2M antibody-based PSA-A2M assay.
In addition, 10 sera with high PSA (?20 ?g/L) from
patients with PCa were subjected to immunoadsorption,
treatment with NaOH, separation by gel filtration, and
analysis by the PSA IFMA and the A2M antibody-based
dissociation assay for psa-a2m in serum
Serum samples or calibrators dissolved in pooled female
serum (200 ?L) were subjected to immunoadsorption,
after which 180 ?L of the adsorbed calibrators or samples
were incubated with 20 ?L of NaOH (0.60 mol/L) at
25 °C. After 30 min, 340 ?L of assay buffer containing 20
?L of HCl (0.60 mol/L) was added to neutralize the
solution. Duplicates of 225 ?L, corresponding to the 75 ?L
of original serum or calibrators, were assayed for total
PSA by IFMA. The intra- or interassay coefficients of
variation (CVs) were determined by measuring five se-
rum samples with PSA concentrations from 2 to 10 ?g/L
10 times within the same or different analytical runs.
standardization of the dissociation assay
The loss of PSA immunoreactivity during the in vitro
complexation with A2M was assessed by gel filtration and
used to estimate the PSA content in the PSA-A2M calibra-
tor. The calibrators were prepared by dilution of purified
PSA-A2M formed in vitro in a pooled female serum
without PSA immunoreactivity to concentrations of 0, 0.1,
1, 4, 20, and 100 ?g/L.
stability of psa-a2m in serum
Purified PSA-A2M formed in vitro was added to pooled
female serum at five different concentrations from 1 to 100
?g/L and stored at 4 or 25 °C for 0, 24, 72, and 168 h. The
concentrations of PSA and PSA-A2M were measured by
the PSA IFMA and by the dissociation assay for PSA-
A2M, respectively. Alternatively, purified PSA-A2M was
added into 46 individual female sera at a final concentra-
tion of 25 ?g/L and stored for 2 h at 25 °C. The recovery
of PSA-A2M was analyzed by the dissociation assay for
PSA-A2M. We also measured the pH of the individual
sera after the addition of NaOH.
The detection limit of the dissociation assay for PSA-A2M
was defined as the concentration giving a fluorescence
signal equal to that of a female serum pool plus 2 SD
(calculated from 20 duplicates). The difference in the
concentrations and proportions of PSA-A2M between
serum from patients with PCa and BPH was tested by the
Wilcoxon rank-sum test. The diagnostic accuracy of the
various PSA assays were compared by ROC curves anal-
ysis as described (25).
formation and characterization of psa-a2m
Formation of PSA-A2M in vitro. Incubation of purified PSA
with A2M at 37 °C produced a gradual loss of immuno-
reactive PSA, as detected by the PSA IFMA (Fig. 1). Gel
filtration of the aliquot obtained at the 72-h time point
showed that very little PSA immunoreactivity (?1% of
added PSA) was detected in the fractions containing A2M,
although ?40% of the added PSA had complexed with
A2M (Fig. 2). In the A2M antibody-based PSA-A2M assay,
a low response was observed in the fractions containing
A2M (Fig. 2). However, a similar response was also
detected after gel filtration of pure A2M or female serum
without PSA immunoreactivity (not shown). The fractions
containing A2M obtained by gel filtration were further
subjected to immunoadsorption with biotinylated mono-
and polyclonal PSA antibodies and streptavidin-conju-
Fig. 1. Kinetics of the formation of PSA-A2M in vitro.
PSA (60 ?g) in 100 ?L of TBS was incubated with 12 mg of A2M (1:10 molar
ratio) in 700 ?L of TBS at 37 °C. Aliquots (100 ?L) were drawn at time intervals
of 0, 1, 3, 6, 8, 24, 48, and 72 h and analyzed by the PSA IFMA and by the
dissociation assay for PSA-A2M. A control containing pure PSA diluted in TBS
containing 50 g/L BSA was stable under the conditions used. E, PSA-A2M; F,
PSA; ‚, PSA control.
Clinical Chemistry 44, No. 12, 1998
gated magnetic beads. Only trace amounts of immunore-
active PSA (?1%) were adsorbed to the magnetic beads.
Elution of the adsorbed material with trifluoroacetic acid
and immunoblotting with polyclonal anti-PSA or anti-
A2M antibodies showed that the immunoreactive PSA
was in the free form and that no A2M immunoreactivity
was associated with this PSA (not shown).
Characterization of PSA-A2M by immunoblotting. When the
fractions containing A2M obtained by gel filtration were
analyzed by SDS-PAGE and immunoblotting, two bands
with molecular masses of 300–400 and 100–200 kDa and
which reacted with PSA and A2M antibodies were ob-
served (Fig. 3). The band at 300–400 kDa apparently was
dimeric A2M complexed with PSA, and the band of
100–200 was probably monomeric A2M complexed with
PSA. Several bands of molecular mass ?100 kDa reacted
only with antibodies to A2M, suggesting that they were
fragments of A2M (Fig. 3B). In addition, a weak 30-kDa
band, which reacted with PSA antibody, was also ob-
served (Fig. 3A), indicating that a minor portion of PSA
was released from PSA-A2M by treatment with SDS.
Endogenous PSA-A2M in serum displayed similar bands
that reacted with antibodies to PSA and A2M (not shown).
Adsorption with PSA antibody did not remove the PSA or
A2M immunoreactivity from the PSA-A2M fractions (not
treatment of psa-a2m with NaOH
Denaturation of PSA-A2M in TBS buffer. When purified
PSA-A2M dissolved in TBS buffer containing 50 g/L BSA
was incubated with NaOH for 30 min at 25 °C, PSA
immunoreactivity was measured by the PSA IFMA after
neutralization with HCl. The recovery of PSA immunore-
activity increased with increasing pH up to 11.5 (Fig. 4),
corresponding to a NaOH concentration of 70 mmol/L in
a TBS buffer-based matrix. The recovery of PSA immuno-
reactivity was ?30% of the calculated PSA content in
PSA-A2M. Incubation of PSA-A2M in TBS at pH 11.5 for
up to 3 h at 25 °C marginally affected the recovery of PSA
immunoreactivity (not shown). Incubation of purified
Fig. 2. Purification of PSA-A2M formed in vitro by gel filtration.
PSA (60 ?g) and 12 mg of A2M (1:10 molar ratio) were incubated in 800 ?L of
TBS at 37 °C for 72 h and fractionated by gel filtration on a Superdex-200
column. The horizontal bar indicates fractions further characterized by immuno-
blotting and immunoadsorption. F, PSA measured by PSA IFMA after incubation
with A2M; E, PSA assayed with PSA IFMA without incubation with A2M; Œ, A2M
measured with the A2M IFMA after incubation with PSA; ‚, PSA-A2M measured
with the A2M antibody-based PSA-A2M assay after incubation of PSA with A2M.
Fig. 3. Identification of PSA-A2M formed in vitro by immunoblotting.
The fractions containing A2M obtained by gel filtration as indicated in Fig. 2 were
pooled and subjected to SDS-PAGE and immunoblotting with polyclonal anti-PSA
(A) and anti-A2M (B) antibodies. Lane M, molecular mass markers as follows:
myosin (220 kDa), phosphorylase (97 kDa), BSA (66 kDa), ovalbumin (46 kDa),
and carbonic anhydrase (30 kDa); lane 1, purified A2M; lane 2, PSA-A2M purified
by gel filtration; lane 3, purified PSA.
Fig. 4. Release of PSA immunoreactivity from PSA-A2M formed in vitro
at various pH values.
PSA-A2M purified by gel filtration was diluted in TBS containing 50 g/L BSA and
incubated with NaOH at a final concentration range of 0–100 mmol/L at 25 °C
for 30 min. The aliquots were analyzed with the PSA IFMA. The control consisted
of pure PSA diluted in the same buffer as above. E, PSA-A2M (%); F, total PSA;
Zhang et al.: Determination of PSA-A2M complex
PSA in TBS with increasing concentrations of NaOH
tended to reduce PSA immunoreactivity; however, the
loss was remarkable only above pH 12.5, which corre-
sponded to a NaOH concentration of 85 mmol/L (Fig. 4).
When PSA-A2M treated in TBS at pH 11.5 was subjected
to gel filtration, two components were detected by the
PSA IFMA (Fig. 5). A high-molecular weight peak (?400
kDa) comprised about 30% of the released PSA immuno-
reactivity, whereas the rest eluted as a 30-kDa component
(Fig. 5). The 400-kDa component was also detected by the
A2M antibody-based PSA-A2M assay (not shown), sug-
gesting that it represented PSA bound to denaturated
A2M. The 30-kDa peak apparently consisted of free PSA
released from PSA-A2M (Fig. 5). When added to fresh
female serum, the released PSA was able to form a
complex with A2M and ACT, indicating that it was
enzymatically active (not shown).
Denaturation of PSA-A2M in serum. When purified PSA-
A2M diluted in pooled female serum or when immu-
noadsorbed serum from PCa patients (n ? 10) was
incubated at high pH, the appearance of PSA immuno-
reactivity was detected by the PSA IFMA. Maximal
recovery was obtained at pH 11.4, corresponding to a
NaOH concentration of 60 mmol/L in this matrix. Gel
filtration of the NaOH-treated serum revealed a high-
molecular mass component (?400 kDa) and a 30-kDa
component, which contained ?30% and 70% of the
recovered PSA immunoreactivity, respectively (not
shown). The smaller component was detected only by
the PSA IFMA, whereas the 400-kDa component was
also detected by the A2M antibody-based PSA-A2M
assay (not shown). Thus, PSA-A2M formed in vitro and
endogenous PSA-A2M in serum showed similar pat-
terns after denaturation at high pH.
dissociation assay for psa-a2m
Immunoadsorption of PSA in serum. Immunoadsorption of
PCa serum with the PSA antibody removed ?99.9% of
PSA immunoreactivity if the concentration of PSA in the
serum was ?500 ?g/L. After the beads were washed with
assay buffer, the PSA immunoreactivity was eluted from
Fig. 5. Separation of PSA-A2M formed in vitro by gel filtration after
denaturation at pH 11.5.
Purified PSA-A2M formed in vitro was incubated at pH 11.5 for 30 min at 25 °C.
The samples before and after treatment with NaOH were fractionated by gel
filtration. Œ, PSA-A2M in fractions measured with the A2M antibody-based
PSA-A2M assay after denaturation; ‚, PSA-A2M in fractions measured with the
A2M antibody-based PSA-A2M assay before denaturation; F, PSA measured by
the PSA IFMA after denaturation; E, PSA measured by the PSA IFMA before
Fig. 6. Calibration curve for the dissociation assay for PSA-A2M.
The calibrators were prepared by dilution of purified PSA-A2M formed in vitro in
female serum. The background of ?500 cps has been subtracted.
Fig. 7. Correlation between the loss of PSA immunoreactivity and
PSA-A2M formation when PSA was incubated with A2M in vitro.
Purified PSA was incubated with A2M at 37 °C as indicated in Fig. 1. Aliquots
drawn at 0, 1, 3, 6, 8, 24, 48, and 72 h were analyzed by the PSA IFMA and the
dissociation assay for PSA-A2M. The correlation between the loss of immunore-
active PSA (x) and PSA-A2M formed (y) is: y ? ?7.81 ? 1.07x; r ? 0.99.
Clinical Chemistry 44, No. 12, 1998
the beads with 1 mL/L trifluoroacetic acid and subjected
to SDS-PAGE and immunoblotting. The extracted PSA
consisted of complexed PSA (90 kDa) and free PSA. No
A2M immunoreactivity was detected (not shown).
Characteristics of the assay procedure. The dissociation assay
for PSA-A2M included immunoadsorption, treatment at
pH 11.4, neutralization, and determination of the released
PSA immunoreactivity. The quantification range of PSA-
A2M was 0–100 ?g/L (Fig. 6), and the detection limit was
0.14 ?g/L. The values in Fig. 6 represent PSA and
disregard the content of A2M. The intraassay CV was
4–9%, and the interassay CV was 8–14% in samples with
PSA concentrations in the range of 2–10 ?g/L.
Validation of the dissociation assay for PSA-A2M. When
purified PSA was incubated with A2M at 37 °C, the
concentrations of PSA-A2M measured by the assay for
PSA-A2M increased with time, whereas the concentration
of free PSA decreased (Fig. 1). The loss of free PSA (x)
correlated with the concentration of PSA-A2M formed (y;
Fig. 7). The equation for the line in Fig. 7 is: y ? ?7.81 ?
1.07x; r ? 0.99).
Recovery of PSA-A2M added to female serum. Purified PSA-
A2M formed in vitro was added to pooled female serum
to give concentrations of 1–100 ?g/L. After storage for 7
days at 4 °C, the mean recovery was 97% (range, 94–
101%); after storage at 25 °C, it was 95% (range, 90–98%).
When purified PSA-A2M was added to 46 individual
female sera and incubated for 2 h at 25 °C, the median
recovery of PSA-A2M was 96% (95% confidence interval,
92–97%). No PSA immunoreactivity was detected in se-
rum before denaturation. Addition of NaOH to 46 female
sera to a final concentration of 60 mmol/L increased the
pH to 11.40 ? 0.13 (mean ? SD).
determination of psa-a2m and total psa in serum
from healthy females and pca and bph patients
The concentration of PSA-A2M was below the detection
limit of the assay in 44 (96%) of the 46 female sera. Two
samples had apparent PSA-A2M concentrations of 0.18
and 0.43 ?g/L. The female serum used as a matrix for the
calibration had no detectable PSA immunoreactivity, as
determined by the IFMA for PSA. The median concentra-
tion of total PSA was 13.5 ?g/L (range, 0.4–432 ?g/L) in
sera from 73 patients with PCa and 4.9 ?g/L (range, 1–73
Fig. 8. The concentration of PSA-A2M in serum from patients with BPH
(E; n ? 58) and PCa (F; n ? 73) as a function of the concentration of
The concentration of total PSA was determined by the PSA IFMA, and the
concentration of PSA-A2M was analyzed by the dissociation assay for PSA-A2M.
Fig. 9. The ratio of PSA-A2M in relation to total PSA in serum from
patients with BPH (E; n ? 58) and PCa (F; n ? 73) as a function of the
concentration of total PSA.
The concentrations of total PSA and PSA-A2M were determined by the PSA IFMA
and the dissociation assay for PSA-A2M, respectively.
Fig. 10. ROC curves of total PSA and the ratio of PSA-A2M in serum with
the concentration of total PSA between 4 and 10 ?g/L.
The area under the curve for the ratio of PSA-A2M to total PSA (E) was 0.78; the
area under the curve for total PSA (F) was 0.66.
Zhang et al.: Determination of PSA-A2M complex
?g/L) in sera from 58 BPH patients (P ? 0.001). In
patients with PCa, the concentrations of PSA-A2M ranged
from 0 to 49 ?g/L (median, 1.2 ?g/L); in BPH patients, it
ranged from 0 to 14 ?g/L (median, 0.7 ?g/L; P ?0.001).
The PSA-A2M concentrations correlated with those of
total PSA both in PCa and BPH (Fig. 8). The ratio of
PSA-A2M in relation to total PSA was higher in BPH
(median, 17%; range, 0–60%) than in PCa (median, 12%;
range, 0–63%; P ?0.001; Fig. 9). In samples with PSA
concentrations of 4–10 ?g/L, the median ratio of PSA-
A2M to total PSA in serum was also significantly higher in
BPH (19.5%; n ? 28) than in PCa (14.5%; n ? 23; P ?
0.002). ROC curve analysis of samples with PSA concen-
trations in the range 4–10 ?g/L showed that the area
under the curve was 0.78 for the ratio of PSA-A2M to total
PSA, whereas for total PSA, the area under the curve was
0.66 (Fig. 10).
PSA rapidly forms a complex with A2M in vitro (6, 8, 11),
and PSA-A2M can be immunochemically detected in male
serum by immunoblotting but not by conventional immu-
noassays for PSA (6, 20). In this study, we developed a
quantitative dissociation assay for PSA-A2M in serum and
characterized PSA-A2M formed in vitro and endogenous
PSA-A2M in serum. The assay is based on the removal of
free PSA, PSA-ACT, and PSA-API; denaturation of PSA-
A2M at high pH; and measurement of released PSA
immunoreactivity by a conventional PSA immunoassay.
Using the new dissociation assay for PSA-A2M, we were
able to quantify PSA-A2M in serum containing PSA at
clinically relevant concentrations and to show that the
concentration of PSA-A2M in serum is correlated to total
PSA and that the ratio of PSA-A2M to total PSA in serum
is higher in BPH than in PCa.
Treatment at high pH causes partial dissociation of
PSA-A2M. Both the free PSA released and that remaining
covalently bound to denatured A2M can be detected by a
conventional PSA immunoassay. The recovery of PSA
immunoreactivity after denaturation is ?30% of the cal-
culated PSA content in PSA-A2M. Because the denatur-
ation at pH 11.5 does not cause any considerable loss of
the immunoreactivity of free PSA, the low recovery is
probably attributable to the reduced immunoreactivity of
the PSA bound to denatured A2M. The recovery of PSA
immunoreactivity is to some extent dependent on the
final pH during the treatment of PSA-A2M with NaOH. In
different sera, the final pH may vary because of differ-
ences in buffering capacity. However, treatment of 46
female sera with NaOH produced very little variation in
the final pH, and there was little variation in the recovery
of PSA immunoreactivity of PSA-A2M added to these
sera. Thus, differences in the buffering capacity of serum
had a negligible effect on the recovery of PSA from
The PSA IFMA used in the new dissociation assay for
PSA-A2M recognizes PSA-ACT and free PSA in an
equimolar fashion (8); however, it underestimates PSA
remaining bound to denatured A2M. This is to be ex-
pected because proteases captured by A2M are covalently
bound in a random fashion by a reactive thiol ester in
A2M, which binds mainly to lysine residues in the pro-
teases to form stable amide cross-links (16). A2M also
forms base-labile ester cross-links to serine, threonine,
tyrosine, or carbohydrate groups in captured proteases.
The free PSA released by NaOH treatment has probably
been bound by such bonds (16). The recognition of PSA in
complex with A2M after denaturation was not assay-
specific. Various assays using other PSA antibodies, in-
cluding those specific for free PSA, measured PSA bound
to denatured A2M in a similar way (not shown).
It is possible that the proportion of PSA released from
and bound to denaturated A2M varies from one serum to
another. This would affect the recovery of PSA immuno-
reactivity because the two forms of PSA measurable after
the denaturation of PSA-A2M have different immunore-
activities. We observed a ratio of ?70% to 30% between
free PSA released and that bound to A2M in 10 sera with
high PSA concentrations that had been separated by gel
filtration after treatment with NaOH. A similar ratio was
also observed when PSA-A2M formed in vitro was sub-
jected to denaturation and separation by gel filtration.
NaOH can be neutralized with HCl with little sample
dilution. After treatment of PSA-A2M with NaOH and
whereas PSA retained its immunoreactivity. With a sam-
ple volume of 225 ?L, corresponding to 75 ?L of serum or
calibrators, the detection limit was 0.14 ?g/L. This en-
abled us to reliably analyze PSA-A2M in sera with total
PSA concentrations ?4 ?g/L.
Treatment of PSA-A2M with SDS has been used to
expose PSA encapsulated by A2M (6, 20). However, SDS
interferes with antibody binding. This effect can be re-
duced by dilution; dilution however, causes considerable
reduction in assay sensitivity (26). We found that a
portion of the encapsulated PSA was released by SDS.
This is compatible with the observation that portions of
the proteases captured by A2M are not covalently bound
(16). The proportion of PSA released by SDS was much
lower than that released by NaOH (not shown).
Denaturation of PSA-A2M at high pH rendered it
immunoreactive in the A2M antibody-based PSA-A2M
assay as well as in the conventional PSA assay. Denatur-
ation of pure A2M at high pH also induced a small
response in the A2M-antibody-based PSA-A2M assay,
which probably was a result of increased nonspecific
background. This effect was not present in the new
immunoassay for PSA-A2M, which measured PSA re-
leased from the complex and bound to denaturated A2M.
Immunoblotting of PSA-A2M formed in vitro showed
that two bands with molecular masses of 300–400 and
100–200 kDa, respectively, reacted with PSA antibody.
The 300- to 400-kDa band apparently represents PSA
complexed with a dimer of A2M subunits, as also ob-
Clinical Chemistry 44, No. 12, 1998
served in a recent study (27). The low-molecular mass
band probably represents PSA bound to a monomeric
A2M subunit (27).
Because PSA-A2M formed in vitro and endogenous
PSA-A2M in serum showed similar patterns after dena-
turation at high pH, we used PSA-A2M formed in vitro to
calibrate the PSA-A2M assay. To obtain the same buffer-
ing capacity as in serum, the calibrators were prepared by
dilution of purified PSA-A2M in female serum. PSA-A2M
was stable in female serum; thus it was a suitable matrix
for the PSA-A2M calibrators.
Recently Espana et al. (28) measured PSA-A2M in
serum, using an A2M antibody-based PSA-A2M assay
consisting of a capture antibody to PSA and a tracer
antibody to A2M; they found that the serum concentra-
tions were not related to the concentrations of total PSA.
With an identical assay, we previously detected very low
PSA-A2M immunoreactivity in male serum with high
PSA concentrations (10). In the present study, we found
that the apparent immunoreactivity detected by the A2M
antibody-based PSA-A2M assay is caused by A2M be-
cause it was also observed with purified A2M and female
serum. A similar background problem hampers assays for
PSA-ACT (10). This suggests that the apparent immuno-
reactivity represents a nonspecific background caused by
adsorption of A2M to the solid phase. This explanation is
supported by the finding that ?1% of the PSA-A2M in
serum and that formed in vitro could be recovered by
immunoadsorption when either mono- or polyclonal anti-
PSA antibodies were used.
Otto et al. (27) demonstrated recently that PSA can
bind to methylamine-transformed A2M without bait re-
gion cleavage, and the PSA-methylamine-transformed-
A2M complex is detectable in a dual PSA antibody-based
immunoassay. However, proteases are thought to com-
plex with A2M in vivo mainly via bait region cleavage
caused by the high concentration of native A2M in blood
(29). This is supported by the observation that PSA-A2M
occurring in vivo is not measurable by conventional PSA
immunoassays without denaturation.
The ratio of PSA-A2M to total PSA in sera from BPH
patients was higher in BPH than in PCa sera, which is
contrary to the behavior of PSA-ACT (10, 13). This is
probably explained by differences in the mechanism of
complex formation of PSA with A2M and ACT, respec-
tively, and in their clearance from circulation. PSA forms
complexes more rapidly with A2M than with ACT, and
even the proteolytically cleaved or “nicked” PSA isoen-
zymes can bind to A2M (8). A2M-protease complexes have
half-lives of only 2–5 min (16), which are much shorter
than that of total PSA, i.e., 2–3 days (30, 31). Although
most of the enzymatically active PSA released into circu-
lation may be expected to form complexes with A2M,
PSA-ACT predominates in plasma because of its slow
clearance (32). Because enzymatically active PSA is rap-
idly complexed with A2M (8), most of the free PSA
present in blood at the time of sampling may be assumed
to have low enzyme activity. It could consist of enzymat-
ically inactive proenzyme and nicked isoenzymes that
bind only with A2M (8). Nicked PSA could gradually
form complexes with A2M after sampling. The high
proportion of free PSA in serum from BPH patients could
therefore lead to preferential formation of PSA-A2M after
sampling. Thus, the concentration of PSA-A2M may re-
flect the concentration of free PSA at the time of sampling.
This notion is actually supported by our preliminary
results, which suggest that the sum of PSA-A2M and free
PSA in serum may improve the diagnostic accuracy for
PCa when compared with free PSA alone. Formation of
PSA-A2M in vitro is also thought to cause loss of PSA
immunoreactivity during long-term storage of serum
Measurement of the ratio of PSA-A2M to total PSA in
serum improved the diagnostic accuracy compared with
total PSA alone, as evidenced by ROC curve analysis.
Thus, the immunoassay for PSA-A2M has the potential to
improve the clinical usefulness of the PSA determination
for detection of PCa. We are presently evaluating the
clinical utility of the combination of PSA-A2M and free
and total PSA by analyzing samples from a screening
In conclusion, we have developed a quantitative assay for
PSA-A2M in serum that is based on immunoadsorption of
PSA and PSA-ACT, denaturation of PSA-A2M at high pH,
and determination of the PSA released by a conventional
PSA immunoassay. Our first results with samples from
referred patients indicate that the ratio of PSA-A2M in
relation to total PSA in serum is higher in BPH than in
PCa and that this can be used to improve the validity of
the PSA assay for detection of PCa.
This work was supported by grants from the Academy of
Finland, the Finnish Cancer Society, Sigird Juse ´lius Foun-
dation, Helsinki University Central Hospital, and The
Centre for International Mobility in Finland (CIMO).
MAbs to PSA, H117, and H50 were kind gifts from Abbott
Diagnostics (Abbott Park, IL).
1. Wang MC, Valenzuela LA, Murphy GP, Chu TM. Purification of a
human prostate-specific antigen. Investig Urol 1979;17:159–63.
2. Watt KW, Lee PJ, M’Timkulu T, Chan WP, Loor R. Human prostate-
specific antigen: structural and functional similarity with serine
proteases. Proc Natl Acad Sci U S A 1986;83:3166–70.
3. Schaller J, Akiyama K, Tsuda R, Marti T, Rickli EE. Isolation,
characterization and amino-acid sequence of ?-seminaprotein, a
glycoprotein from human seminal plasma. Eur J Biochem 1987;
4. Lundwall A, Lilja H. Molecular cloning of human prostate-specific
antigen cDNA. FEBS Lett 1987;214:317–22.
5. Henttu P, Vihko P. cDNA coding for the entire human prostate-
specific antigen shows high homologies to the human tissue
Zhang et al.: Determination of PSA-A2M complex
kallikrein genes. Biochem Biophys Res Commun 1989;160:903– Download full-text
6. Christensson A, Laurell CB, Lilja H. Enzymatic activity of prostate-
specific antigen and its reactions with extracellular serine protein-
ase inhibitors. Eur J Biochem 1990;194:755–63.
7. Christensson A, Lilja H. Complex formation between protein C
inhibitor and prostate-specific antigen in vitro and in human
semen. Eur J Biochem 1994;220:45–53.
8. Leinonen J, Zhang W-M, Stenman U-H. Complex formation be-
tween PSA isoenzymes and protease inhibitors. J Urol 1996;155:
9. Zhang W-M, Leinonen J, Kalkkinen N, Stenman U-H. Prostate-
specific antigen forms a complex with and cleaves alpha-1-
protease inhibitor in vitro. Prostate 1997;33:87–96.
10. Stenman U-H, Leinonen J, Alfthan H, Rannikko S, Tuhkanen K,
Alfthan O. A complex between prostate-specific antigen and
alpha-1-antichymotrypsin is the major form of prostate-specific
antigen in serum of patients with prostatic cancer: assay of the
complex improves clinical sensitivity for cancer. Cancer Res
11. Lilja H, Christensson A, Dahlen U, Matikainen MT, Nilsson O,
Pettersson K, Lo ¨vgren T. Prostate-specific antigen in serum
occurs predominantly in complex with alpha-1-antichymotrypsin.
Clin Chem 1991;37:1618–25.
12. Pettersson K, Piironen T, Seppa ¨la ¨ M, Liukkonen L, Christensson
A, Matikainen MT, et al. Free and complexed prostate-specific
antigen (PSA): in vitro stability, epitope map, and development of
immunofluorometric assays for specific and sensitive detection of
free PSA and PSA-alpha-1-antichymotrypsin complex. Clin Chem
13. Leinonen J, Lo ¨vgren T, Vornanen T, Stenman U-H. Double-label
time-resolved immunofluorometric assay of prostate-specific anti-
gen and of its complex with alpha-1-antichymotrypsin. Clin Chem
14. Christensson A, Bjo ¨rk T, Nilsson O, Dahlen U, Matikainen MT,
Cockett AT, et al. Serum prostate-specific antigen complexed to
alpha-1-antichymotrypsin as an indicator of prostate cancer. J Urol
15. Barrett AJ, Brown MA, Sayers CA. The eletrophoretically ’slow’ and
’fast’ forms of the alpha-2-macroglobulin molecule. Biochem J
16. Sottrup-Jensen L. Alpha-macroglobulins: structure, shape, and
mechanism of proteinase complex formation. J Biol Chem 1989;
17. Sottrup-Jensen L, Petersen TE, Magnusson S. Mechanism of
proteinase complex formation with ?2-macroglobulins. Three
modes of trypsin binding. FEBS Lett 1981;128:127–32.
18. Barrett AJ, Starkey PM. The interaction of alpha-2-macroglobulin
with proteinases: characteristics and specificity of the reaction
and a hypothesis concerning its molecular mechanism. Biochem
19. Miyata K, Nakamura M, Tomoda K. Interaction between serratia
protease and human plasma alpha-2-macroglobulin. J Biochem
20. Zhou AM, Tewari PC, Bluestein BI, Caldwell GW, Larsen FL.
Multiple forms of prostate-specific antigen in serum: differences
in immunorecognition by monoclonal and polyclonal assays. Clin
21. Zhang W-M, Leinonen J, Kalkkinen N, Dowell B, Stenman U-H.
Purification and characterization of different molecular forms of
prostate-specific antigen in human seminal fluid. Clin Chem
22. Kurecki T, Kress LF, Laskowski M. Purification of human alpha-2-
macroglobulin and alpha-1-proteinase inhibitor using zinc chelate
chromatography. Anal Biochem 1979;99:415–20.
23. Laemmli UK. Cleavage of structure proteins during the assemble
of the head bacteriophage T4. Nature 1970;227:680–5.
24. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of
proteins from polyacrylamide gels to nitrocellulose sheets: proce-
dure and some applications. Proc Natl Acad Sci U S A 1979;76:
25. Stenman U-H, Hakama M, Knekt P, Aromaa A, Teppo L, Leinonen
J. Serum concentrations of prostate-specific antigen and its
complex with alpha-1-antichymotrypsin before diagnosis of pros-
tate cancer. Lancet 1994;344:1594–8.
26. Platt EJ, Karlsen K, Lopez-Valdivieso A, Cook PW, Firestone GL.
Highly sensitive immunoadsorption procedure for detection of
low-abundance proteins. Anal Biochem 1986;156:126–35.
27. Otto A, Ba ¨r J, Birkenmeier G. Prostate-specific antigen forms
complexes with human ?2-macroglobulin and binds to the ?2-
macroglobulin receptor/LDL receptor-related protein. J Urol 1998;
28. Espana F, Sanchez CJ, Estelles A, Gilabert J, Griffin JH, Heeb MJ.
Quantitative immunoassay for complexes of prostate-specific
antigen with alpha-2-macroglobulin. Clin Chem 1996;42:545–50.
29. Birkenmeier G, Stigbrand T. Production of conformation-specific
monoclonal antibodies against ?2-macroglobulin and their use for
quantitation of total and transformed ?2-macroglobulin in human
blood. J Immunol Methods 1993;162:59–64.
30. Stamey TA, Yang N, Hay AR, McNeal JE, Freiha FS, Redwine E.
Prostate-specific antigen as a serum marker for adenocarcinoma
of the prostate. N Engl J Med 1987;317:909–16.
31. Oesterling JE, Chan DW, Epstein JI, Kimball AW, Bruzek DJ, Rock
RC, et al. Prostate specific antigen in the preoperative and
postoperative evaluation of localized prostatic cancer treated with
radical prostatectomy. J Urol 1988;139:766–72.
32. Stenman UH. Prostate-specific antigen, clinical use and staging:
an overview. Br J Urol 1997;79(Suppl 1):53–60.
Clinical Chemistry 44, No. 12, 1998