| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
Tumor Immunology Laboratory of the Urology Department, School of Medicine of the University of Washington, Seattle, WA 98195.
a Address correspondence to this author at: Mailstop 356510, Department of Urology, University of Washington, Seattle, WA 98195. Fax 206-543-1146; e-mail ecorey{at}u.washington.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin
(ACT). Linear epitope mapping with a polyclonal antibody indicated the
presence of six major antigenic regions in PSA. Examination of the
panel of mAbs established that three of them bind to linear epitopes.
Five of the mAbs inhibited >90% of PSA enzymatic activity. However,
inhibition of PSA enzymatic activity and hindrance of PSA-ACT
association by mAbs cannot be used to predict whether the mAbs bind to
free PSA, the PSA-ACT complex, or both. Some of the mAbs may block
PSA-ACT association through peripheral occlusion of the binding site,
or through induction of conformational changes in PSA.
Key Words: indexing terms: linear epitope • 1-antichymotrypsin • proteolytic activity
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin (ACT), and specific association
only with free (f)-PSA. Antigenic determinants as binding targets of antibodies can be divided into two categories: linear (sequential, continuous) and nonlinear (conformational, discontinuous). Linear epitopes consist of amino acid residues that are adjacent in the primary sequence; nonlinear epitopes consist of amino acid residues that are separated in the primary structure but are brought into proximity when the protein is in its native form. At present there is no simple way to identify conformational epitopes in the absence of three-dimensional structural information about the monoclonal antibody (mAb)-antigen complex. However, the identity of linear epitopes can be predicted by computer programs that calculate various parameters that have been found to be correlated with antigenic nature in previously studied antigens (e.g., hydrophilicity, flexibility, and surface probability (2)). The method postulates that (a) mAbs that bind to linear epitopes react with segments of 5–8 consecutive amino acid residues, and (b) these epitopes are on the surfaces of molecules that tend to be hydrophilic. However, computational techniques are not yet sophisticated enough to achieve the accuracy of experimental techniques. Other methods of identifying antibody binding sites involve (a) digestion of the antigen, (b) recombinant techniques, (c) phage display, (d) mass spectrometry, and (e) the use of synthetic peptides. Fragments from trypsin (3) or papain (4) digestion of antigens have been used to determine antibody binding targets. Numerous attempts with cyanogen bromide cleavage products have been published (4)(5)(6). Van Duijnhoven et al. reported the use of recombinant DNA techniques for epitope mapping (7). The phage display technique has found use in epitope mapping in determination of the antigenic regions of various proteins (8)(9)(10)(11). By using modern mass spectrometry techniques, epitopes can be determined as described by Zhao and Chait (12), or by Papac et al. (13). Geysen et al. in 1987 (14) published a further method of identifying linear epitopes as association targets of antibodies by using overlapping synthetic peptides of antigens of known sequence. Given the recent progress in methods for simultaneous synthesis of large numbers of peptides, creating the peptide arrays corresponding to all possible contiguous segments of the protein of interest has become practical. The peptides are designed with substantial overlapping regions. Linear epitopes are then defined by identifying the peptides that associate most strongly with antibodies developed against the full-sized antigen. This methodology has been used with success in many cases (15)(16)(17)(18)(19). We used the overlapping synthetic peptide method in combination with a commercially available rabbit polyclonal anti-PSA antibody and a panel of novel anti-PSA mAbs developed in our laboratory to characterize the linear epitopes of PSA.
An independent means of characterizing the binding targets of anti-PSA mAbs is by studying the effect of the mAbs on the enzymatic activity of PSA. This activity is held to be essential for formation of the PSA-ACT complex (20)(21), implying that antibodies that affect the activity may be associating with epitopes proximate to the ACT binding site, and may therefore have special properties, such as the ability to distinguish f-PSA from the complex. One certainly expects to find a correlation between inhibition of PSA enzymatic activity by mAbs and inhibition of the association of PSA with ACT. Our results show that although the two inhibitory functions are strongly correlated, the presence of these inhibitory properties did not necessarily imply that the mAb was associating directly with the PSA-ACT interface.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
0.1%=1.84).
anti-psa monoclonal antibodies
Ten anti-PSA mAbs were developed in our laboratory by immunization
of BALB mice with purified seminal plasma PSA (23).
synthetic peptides
Synthetic peptides were purchased from Zeneca (Cambridge
Biochemicals, Wilmington, DE). They were synthesized on 96-well format
plates by solid-phase methodology with Fmoc chemistry. Cysteine
residues in peptides were in the reduced form. The amide group was
chosen as the C terminus of the peptides to provide maximal resemblance
to internal sequences. Peptides were cleaved from the solid support and
dried. As a quality control, two peptides synthesized on the same plate
were evaluated, one of an arbitrary reference sequence and one from the
PSA sequence (peptide 155). These peptides were analyzed by HPLC to
determine purity, and amino acid analysis was used for estimation of
the average yield. Peptides used for screening were not purified; the
purity was ~60%, with an average yield of ~2 mg. The PSA sequence
was spanned by peptides 15 amino acid residues long with 10-residue
overlaps. Biotin was added to a four-residue extension ("handle")
of sequence SGSG at the N terminus to allow easy attachment to solid
supports through association with immobilized streptavidin. The
peptides were designated by the PSA sequence numbers of their first
residues. The first peptide (1) spanned the protein sequence from
position 0 (R) to position 14 (W), the second peptide (5) positions
5–19 (W-A), and so on.
linear epitope mapping
Enzyme immunoassay (EIA) plates (Costar 3590; Costar, Cambridge,
MA) were coated with 100 µL of 5 mg/L streptavidin (Sigma, St. Louis,
MO) in water. Plates were incubated at 37 °C until dry. After three
washes with 20 mmol/L PBS, pH 7.4, containing 1 mL/L Tween 20, the
plates were blocked for 1 h with 200 µL of 20 g/L bovine serum
albumin (BSA) in PBS (pH 7.4) at room temperature (RT). Approximately 2
mg of peptides (~60% pure) were dissolved in 200 µL of dimethyl
sulfoxide (~3 mmol/L peptide, stock solution); each peptide was
diluted 1:500 in PBS:1 g/L BSA:1 g/L NaN3 before use
(~6 µmol/L peptide). These dilutions (100 µL) were incubated with
streptavidin-coated wells for 1 h at RT with shaking. Each
antibody (100 µL; 10 µg/mL, ~62 nmol/L), either anti-PSA mAb or
rabbit anti-PSA polyclonal antibody (Dako Corp., Carpinteria, CA), were
incubated with the immobilized peptides overnight at 4 °C. After
three washes, the appropriate horse anti-mouse antibody (Zymed Labs.,
So. San Francisco, CA) or goat anti-rabbit antibody (Bio-Rad Labs.,
Hercules, CA), coupled with horseradish peroxidase (HRP), was added at
a dilution of 1:2000 in 20 g/L BSA:PBS and incubated for 1 h at
RT. Reactive peptides were detected through HRP with the
2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) substrate
(Zymed Labs.). As a negative control, plates with immobilized synthetic
peptides were treated in the same manner with primary anti-PSA
antibodies omitted.
inhibition of 60–1a2, 60–8a2, and 17–1a2 association with psa by
peptides derived from psa
PSA was immobilized on EIA plates (750 ng/mL, ~25 nmol/L, 100
µL/well, overnight at 4 °C). 60–1A2, 60–8A2, or 17–1A2 (90
µL; 10 µg/900 µL, ~69 nmol/L) were incubated with 10 µL of
peptide 155 or 160 (mAbs 60–1A2 and 60–8A2) or peptide 50 or 55 (mAb
17–1A2) (peptides at ~60 µmol/L in PBS:1 g/L BSA) for 30 min and
incubated with immobilized PSA for 20 h. The quantity of bound
antibody was determined as above. Control experiments with no peptides
present were run simultaneously.
inhibition of psa enzymatic activity by mabs
Inhibition of PSA enzymatic activity against the chymotrypsin
substrate S-2586 (MeO-Suc-Arg-Pro-Tyr-pNA; Kabi Pharmacia,
Franklin, OH) by the mAbs was tested. One microgram of PSA was
incubated with 1.5 or 4.5 equivalents of mAbs for 2 or 18 h in 80
µL of 100 mmol/L Tris-HCl, 100 mmol/L NaCl, pH 7.51 (PSA ~410
nmol/L, mAbs ~615 nmol/L, ~1.85 µmol/L respectively). After
addition of 20 µL of 5 mmol/L S-2586, absorbance at 410 nm was read
approximately every 5–10 min over 2 h. Experiments with PSA only,
MOPC-21 (unrelated antibody), and ACT were run simultaneously as
controls. Linear regression analysis was used to determine
A410/min, and the degree of inhibition was
calculated as the difference in these rates attributable to the
presence of the antibody.
antibody inhibition of act binding to psa
One microgram of PSA was incubated with 3 equivalents of each of
the mAbs for 3 h at 37 °C in 18 µL of 100 mmol/L Tris-HCl,
100 mmol/L NaCl, pH 7.51 (~1.85 µmol/L PSA, ~5.6 µmol/L mAb).
Three equivalents of ACT (Athens Scientific, Athens, GA) in 2 µL of
the same buffer were then added (final concentration of ACT ~5.6
µmol/L). From these reaction solutions, 1.5 µL was removed at times
0, 1, 3, 6, and 26 h and immediately diluted into 150 µL of
ice-cold PBS:10 g/L BSA. Additional dilutions of these solutions were
prepared for EIA determination of PSA-mAb, mAb-PSA-ACT, and PSA-ACT
complexes. The presence of these complexes was detected in three ways.
For detection of the PSA-ACT and mAb-PSA-ACT complexes, we used a PSA
capture process. Plates coated with the mouse anti-PSA mAb 22–8A2 (10
µg/mL, 62 nmol/L) were incubated with the diluted reaction mixtures
for 2 h at RT. The ACT complexes [PSA-ACT and (or) mAb-PSA-ACT]
were detected with a rabbit anti-ACT polyclonal antibody (Athens
Scientific, purified in-house; 3 µg/mL, ~18.8 nmol/L) in
combination with a mouse-adsorbed HRP-conjugated goat anti-rabbit
antibody (Bio-Rad Labs.) at a working dilution of 1:2000 in 1 g/L
BSA:PBS. MOPC-21 and PSA alone were used as controls. To confirm the
detection of mAb-PSA-ACT complexes, we captured them with a horse
anti-mouse antibody at 3 µg/mL (~18.8 nmol/L) (BA2080; Vector
Labs., Burlingame, CA), with detection by a rabbit anti-ACT polyclonal
antibody (3 µg/mL, ~18.8 nmol/L) in combination with a
mouse-adsorbed HRP-conjugated goat anti-rabbit antibody (Bio-Rad
Labs.). To check for anti-mouse antibody coating and capture of mAb-PSA
and (or) mAb-PSA-ACT complexes in cases of negative results for ACT, we
coated plates in the same way with anti-mouse antibody, but detected
the presence of PSA with a rabbit anti-PSA polyclonal antibody in
combination with a mouse-adsorbed HRP-conjugated goat anti-rabbit
antibody (Bio-Rad Labs.). A further technique for detection of
mAb-PSA-ACT was as follows. We captured the ACT complexes on EIA plates
coated with the rabbit anti-ACT polyclonal antibody at 10 µg/mL
(~62 nmol/L). After 2 h of incubation with the reaction mixture,
the presence of the mouse anti-PSA mAb was detected by reaction with an
HRP-conjugated rabbit anti-mouse antibody IgG+A+M (H+L) (Zymed Labs.)
at a 1:2000 working dilution in 1 g/L BSA:PBS.
The apparent rate constants for the PSA-ACT association reaction in the
presence of mAbs 16–3A2 and 22–8A2 were calculated by nonlinear
regression of the data presented in Fig. 6A
to Eq. 1:
 | (1) |
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Reactivity of synthetic peptides with the anti-rabbit
antibody alone was used as a negative control; final data were
corrected for this nonspecific binding.
|
Further evaluation of linear epitopes was performed with a set of mAbs
of known specificity and sandwiching characteristics. Three of the 10
anti-PSA mAbs reacted with synthetic peptides derived from the known
primary sequence of PSA. 17–1A2 exhibited binding to
synthetic peptides 50 (sequence 50–64, LLGRHSLFHPEDTGQ) and 55
(sequence 55–69, SLFHPEDTGQVFQVS) (Fig. 2
A), whereas 60–1A2 and 60–8A2 bound to synthetic peptides 155
(sequence 155–169, ISNDVCAQVHPQKVT) and 160 (sequence 160–175,
CAQVHPQKVTKFML) (Figs. 3
A and 4A
). The binding signals of these antibodies under these
conditions vary greatly in intensity: The signal from 60–8A2 was
strong after 5–10 min, whereas those for 60–1A2 and 17–1A2 were
measured after 60 min. This characteristic is seen to be directly
related to the relative affinities of particular mAbs for native PSA
(23). To confirm the epitopic nature of these peptides we
performed competition experiments with PSA. Peptides 50 and 55 both
competed with PSA for association with 17–1A2. Peptide 50 exhibited
stronger inhibition of mAb-PSA binding (59%), whereas peptide 55
showed weaker inhibition (33%). Peptides 155 and 160 both competed
with PSA for association with 60–8A2; however, the inhibition of
mAb-PSA binding by peptide 155 was only ~10%. Peptide 160 inhibited
38% of mAb-PSA association. Synthetic peptides 155 and 160 did not
inhibit the association of 60–1A2 with PSA. The other seven antibodies
did not exhibit significant binding to PSA-derived synthetic peptides
(data not shown). Experiments were repeated at least twice; within each
experiment triplicate data were used for analysis. Negative controls
without primary antibody were performed simultaneously and used as
controls for reactivity of the peptides with the secondary antibody.
|
|
|
antibody inhibition of psa enzymatic activity
The influence of the mAbs on PSA enzymatic activity was determined
by measuring the ability of PSA to cleave the S-2586 chymotrypsin
substrate after interaction with the mAbs. Steady-state rates were
measured by monitoring A410 over ~2 h. These
data were reduced by linear regression analysis to yield
A410/min with correlation coefficients
>0.98, except in those cases when the mAbs reduced the rate by >80%.
The results for 2-h incubations with 1.5 and 4.5 molar equivalents of
mAbs (relative to PSA) are shown in Table 1
. The mAbs can be classified in this regard as follows:
(a) mAbs 15–19A2, 17–1A2, 60–1A2, 60–6A2, and 60–9A2
inhibited >86% of PSA enzymatic activity; (b) mAb 22–8A2
showed partial inhibition (~60%); (c) mAb 22–1A2 and
16–3A2 actually promoted PSA enzymatic activity (~120–140%), and
(d) mAbs 22–2A3 and 60–8A2 had minimal effects on PSA
enzymatic activity. Two further sets of conditions, 1.5 and 4.5 molar
equivalents of the mAbs with 18-h incubations, were tested to determine
the effect of incubation time on this inhibition. These experiments did
not give significantly different results, except that ACT (1.5
equivalents) inhibited PSA more strongly after 18 h (64.6%) than
after 2 h (33.5%).
|
antibody inhibition of act binding to psa
To look for a correlation with the results of the
enzyme-inhibition studies and learn more about the antigenic targets of
the mAbs, we examined the ability of the mAbs to block the association
of PSA with ACT. Five mAbs with various PSA-association characteristics
were tested in these experiments: 15–19A2 and 60–6A2, which inhibit
PSA enzymatic activity but bind to both f-PSA and PSA-ACT; 17–1A2,
which inhibits PSA enzymatic activity but exhibits a strong binding
preference for f-PSA; 22–8A2, which has a moderate effect on PSA
enzymatic activity and binds to both f-PSA and PSA-ACT; and 16–3A2,
which binds to both f-PSA and PSA-ACT and increases PSA enzymatic
activity. Aliquots of reaction mixtures were removed over 26-h
incubations to evaluate the extent of PSA-ACT association. Fig. 5
shows the quantities of mAb-PSA-ACT complex and (or) PSA-ACT
complex detected by a rabbit anti-ACT polyclonal antibody after being
captured by the mouse anti-PSA mAb 22–8A2; the other four mAbs were
evaluated as potential inhibitors of complex formation in these
experiments. The signal is normalized to the signal of PSA-ACT seen in
the reaction mixture in the presence of the unrelated antibody MOPC-21,
which does not associate with either PSA or ACT and is not expected to
influence PSA-ACT association. Two of the mAbs (15–19A2 and 60–6A2)
appeared to inhibit the association of ACT with PSA completely. The
other two mAbs behaved very differently. In the reaction with mAb
16–3A2, the concentrations of ACT-containing complexes increased
throughout the incubation period. The half-life of the increase
appeared to be identical to the half-life of the control MOPC-21
(uninhibited) reaction, but the asymptote was smaller. mAb 17–1A2 had
a dramatically different effect. This mAb increased the rate of
formation of ACT-containing complexes relative to the control reaction.
|
To distinguish between the formation of PSA-ACT and mAb-PSA-ACT in
these experiments, we captured mAb-PSA-ACT complexes by means of an
anti-mouse antibody and detected the presence of ACT. Results are shown
in Fig. 6
A. mAbs 15–19A2, 17–1A2, and 60–6A2 evidently did not form
mAb-PSA-ACT complexes. To ensure that the absence of the mAb-PSA-ACT
signal is not due to failure of the capture by the anti-mouse antibody,
we successfully used the same experimental setup for detection of
mAb-PSA complexes with an anti-PSA polyclonal antibody (Fig. 6B
). PSA
was detected in complexes with all of these antibodies, with the
exception of 17–1A2, which is known to be a weaker antibody. For this
antibody, we performed the same experiment with a fivefold higher
concentration of the mAb-PSA complex solution in the capture reaction.
Under these conditions we were able to detect the signal of PSA in the
mAb-PSA complex, but no mAb-PSA-ACT was detected (see Fig. 7
).
|
These results were confirmed by experiments in which an anti-ACT
polyclonal antibody was used for capture. mAb-PSA-ACT complexes were
detected by anti-mouse antibodies (Fig. 8
). mAb-PSA-ACT complexes were detected after PSA incubation with
mAbs 16–1A2 and 22–8A2, but not with 60–6A2, 15–19A2, or 17–1A2.
|
Apparent rate constants were calculated for the PSA-ACT association
reaction in the presence of mAbs 16–3A2 and 22–8A2, with nonlinear
regression of the data presented in Fig. 6A
to Eq. 1
(Materials and Methods). The calculated rate constants of
the reaction in the presence of 16–3A2 and 22–8A2 were 0.92 ±
0.15 h-1 and 0.56 ± 0.13 h-1,
respectively.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There is conflict in the literature about the utility and validity of various methods of epitope characterization (26)(27). There is general agreement that the most precise method is x-ray crystallographic analysis of the three-dimensional structure of antibody–antigen complexes. However, only a few of these structures have been solved to date (28)(29)(30)(31)(32). Data from these studies indicate that a large surface of the antigen is usually involved in antibody association. On average 16–22 contact residues contribute to the energy of binding, but at widely varying levels. Laver et al. (26) in their review assert that the widely held view that epitopes in native proteins consist of segments of about six amino acid residues, and that they can therefore be mapped by utilizing synthetic peptides of similar length, is a misconception, since epitopes of this size often result from immune induction by denatured protein molecules. However, the work of Geysen et al. (14), with synthetic peptides, involved the study of various lengths of peptides, and their results indicated that octapeptides are the longest needed for detection of linear epitopes. Moreover, in their amino acid residue replacement study they observed no epitopes in which more than five residues were in contact with the antibody. Continuing reports of the successful use of the short peptides produced by phage display or by chemical synthesis in determination of binding sequences of various antibodies are evidence for the utility of small peptides in linear epitope mapping (6)(7)(8)(9)(10)(11)(14)(15)(16)(17)(18). Data from x-ray studies also indicate that 3–5 residues generally contribute most of the free energy of association (26)(27). The residues are not necessarily adjacent to each other in the primary structure; however, the antibodies in these studies were associating with nonlinear epitopes, whereas the most important contact residues of linear epitopes are likely to be proximate in the primary structure. A compromise hypothesis subsuming all of these studies might hold that in linear epitopes there are 3–6 adjacent residues that contribute the greatest part of the energy of interaction, but that there exist further contacts that stabilize the interaction, involving antigen residues that cannot be identified by short-peptide mapping.
After evaluation of the various available methods for determining the locations of the epitopes, we decided to use linear epitope mapping with synthetic overlapping peptides because of the volume of useful information provided and the high spatial resolution of the technique. Since only 5–10% of mAbs raised against native antigens bind to linear epitopes, we decided to maximize the probability of incorporating both such linear epitopes and any epitopes that might depend on local secondary structure of the protein by using peptides 15 amino acid residues long, which may be partially folded in solution. Linear epitope mapping with a rabbit polyclonal antibody indicated the presence of six major antigenic regions in PSA, in agreement with the report by Pettersson et al. (24). The apparent concurrence between the six antigenic regions seen in Pettersson's sandwich assays and the number determined by our linear epitope mapping can be either real or coincidental. However, we chose a polyclonal antibody for these experiments specifically because the multiplicity of clones present greatly increases the likelihood of detecting any given antigenic region. Signals seen in our procedure probably do not arise only from clones that associate with linear epitopes. Studies (33) indicate that part of the signal is due in some instances to the presence of clones that bind strongly to nonlinear epitopes in the native protein, and also associate less strongly with the portions of these nonlinear epitopes that are represented in the synthetic peptide library we studied.
On the basis only of linear epitope mapping with polyclonal sera, we cannot determine which if any of the highly antigenic regions is specific for f-PSA. Additional experiments, including immunization with peptides from these parts of the protein, could help to identify such regions. Using a set of anti-PSA mAbs that we previously characterized (23), we looked for a correlation between their sandwiching capabilities and the physical locations of their binding targets. The mAbs were divided into five groups on the basis of their sandwiching capabilities: (a) 15–19A2, 60–1A2, 60–6A2, and 60–9A2; (b) 16–3A2; (c) 17–1A2; (d) 22–1A1, 22–2A3, and 22–8A2; and (e) 60–8A2. All of the mAbs recognize both f-PSA and PSA-ACT, although 17–1A2 has a marked preference for f-PSA. 60–8A2 was most probably raised against denatured PSA; it does not bind to PSA in solution. Of this set, three mAbs bound to synthetic peptides. mAbs 60–1A2 and 60–8A2 both bind to sequence 155–174 of PSA. However, judging by the inhibition and competition data, these two mAbs appear to bind to different epitopes within this sequence. Synthetic peptides 155 and 160 competed with immobilized PSA for association with 60–8A2, but not with 60–1A2. Frequently antibodies that bind to linear epitopes still exhibit binding preference for the native antigen, either because of the presence of additional interactions with the protein or because the secondary structure of the epitope is imperfectly modeled by the isolated peptide. Evidently in the case of mAb 60–1A2 the preference for PSA is strong enough to prevent the peptide from competing effectively with the native epitope. Peptide 160 competes more strongly with PSA for association with 60–8A2 than does peptide 155, indicating that this epitope is located in the C-terminal portion of sequence 155–174. mAb 17–1A2 associates with PSA sequence 50–69, represented by peptides 50 and 55. Peptide 50 competes more strongly than peptide 55, suggesting that the binding target is in the N-terminal portion of sequence 50–69. The other seven antibodies did not exhibit significant binding to the synthetic peptides, implying that their targets are nonlinear epitopes.
Sequence 155–174 involved in 60–1A2 and 60–8A2 binding did not exhibit strong antigenicity in experiments with the polyclonal antibody. We hypothesize that this is due to differences in procedures used for antibody production and the quality and purity of the PSA used for immunization. The sequence 50–70, the binding target of 17–1A2, had strong antigenicity in experiments with the polyclonal antibody. Since this mAb exhibits binding preference for f-PSA, and its target region appears to have substantial antigenicity, this region may correspond to the antigenic region observed by Petterson et al., which was preferential or specific for f-PSA (24). In that report, determination of f-PSA specificity was based on the sandwiching capabilities of the antibodies; according to the Western-blot data, all antibodies specific for f-PSA bound to nonlinear epitopes. However, 17–1A2 is also specific for f-PSA in sandwiching experiments with 16–3A2, 22–8A2, and 60–6A2, but this antibody reacts with PSA-ACT on Western blots and in EIA experiments in which PSA-ACT is captured by anti-PSA and anti-ACT polyclonal antibodies, while retaining a strong preference for f-PSA.
Analysis of the effects of mAb association on PSA enzymatic activity is an independent way of inferring additional information about the binding targets. Our data show that mAbs may differ in their specific binding targets, even when sandwich assays have shown that they bind to the same antigenic region. In a particularly interesting example, mAb 22–1A2 actually increases PSA enzymatic activity to >120% of that of the uncomplexed enzyme. This mAb does not sandwich with 22–2A3 or 22–8A2, of which 22–2A3 inhibits only ~20% of the PSA enzymatic activity and 22–8A2 inhibits some 60% of the activity. These results were replicated with various concentrations of mAb and incubation times, indicating that they were not due to artifacts related to incorrect concentrations of mAbs or to the presence of inactive mAb molecules in the preparation. All of these mAbs likely bind in proximity to each other near the active site, but the effects of their binding on protein function are dramatically different. mAb 22–8A2 could be sterically hindering the ACT binding site or occluding the binding pocket, or association of the mAb could be inducing slight changes in PSA tertiary structure that displace the groups necessary for its enzymatic activity.
mAbs that bind only to f-PSA might be expected to inhibit the association of ACT with PSA. However, our data show that inhibition of this association does not imply that the mAb is specific for f-PSA. Five mAbs from the panel inhibit PSA enzymatic activity; of these, four (15–19A2, 60–1A2, 60–6A2, 60–9A2) bind to f-PSA as well as to PSA-ACT. Two of these four (15–19A2, 60–6A2) were tested for the ability to interfere with PSA-ACT association, and both of them did so. Because PSA enzymatic activity is necessary for ACT binding, this phenomenon could be due either to steric hindrance of the ACT binding pocket or to conformational changes of PSA brought about by the interaction that interfere with the catalytic machinery. Thus association of certain mAbs with PSA abolishes ACT binding, yet these mAbs are capable of binding to the PSA-ACT complex. This confirms that the requirements for ACT association, which involves proper functioning of the PSA active site, are much more strict than for binding of mAbs.
The other two mAbs, 16–3A2 and 22–8A2, evaluated for inhibition of
PSA-ACT association had different effects on PSA enzymatic activity.
16–3A2 actually enhanced the PSA hydrolytic rate by ~30–40%,
whereas 22–8A2 decreased the activity by ~60%. A similar difference
was seen in the results of the PSA-ACT formation experiments (see Fig. 6A
), where the observed rate constant for the appearance of the PSA-ACT
complex was ~60% higher in the presence of mAb 16–3A2 than in the
reaction containing 22–8A2. This enhancement of PSA enzymatic activity
is therefore seen in two independent types of reactions. The effect is
probably not exerted on the product-off step, since what is being
measured in the case of PSA-ACT complex formation is the appearance of
the enzyme–ACT complex, a process that could not be hastened by
increasing the dissociation rate. Our data are insufficient to
determine whether 16–3A2 exerts its enhancement effect on the
encounter step or on the chemical processes of the enzyme.
mAb 17–1A2 binds preferentially to f-PSA, and it inhibits >86% of PSA enzymatic activity. This mAb probably binds to a region of PSA that is near the site of ACT association. In studying the effect of 17–1A2 on ACT binding, we found no (17–1A2)-PSA-ACT complex, but the presence of the mAb was seen to retard the course of formation of PSA-ACT. The most likely explanation is that 17–1A2 inhibits the association of PSA with ACT, but because of the relatively weak interaction between 17–1A2 and PSA, ACT gradually replaces the mAb in the complex with PSA.
This is the first description of the specific regions of PSA involved in antibody association. Epitope mapping experiments with an anti-PSA polyclonal antibody have identified six strongly antigenic regions of PSA; these results are in agreement with published data. Inhibition of PSA enzymatic activity and (or) ACT association by anti-PSA mAbs cannot be used to determine whether the mAbs are specific for f-PSA. However, these data could serve as a starting point for a high-resolution study of the epitopes.
We hope that the localization and characterization of the association targets of anti-PSA antibodies will aid in classification and evaluation of new mAbs, and will lead to further structural insights into PSA-mAb complexes, as well as eventual development of more sensitive and specific assays for prostate cancer.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1-antichymotrypsin; f-PSA, free PSA; mAb,
monoclonal antibody; EIA, enzyme immunoassay; BSA, bovine serum
albumin; RT, room temperature; and HRP, horseradish peroxidase. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1-antichymotrypsin. Clin Chem 1991;37:1618-1625.
1-antichymotrypsin as an indicator of prostate cancer. J Urol 1993;150:100-105.
[Web of Science][Medline]
[Order article via Infotrieve]
1-antichymotrypsin complex. Clin Chem 1995;41:1480-1488.
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  Y. Baumgart, A. Otto, A. Schafer, E. Usbeck, C. Cott, A. Schott, M. Tornack, A. Wenzel, A. Mossie, and G. Birkenmeier Characterization of Novel Monoclonal Antibodies for Prostate-Specific Antigen (PSA) with Potency to Recognize PSA Bound to {alpha}2-Macroglobulin Clin. Chem., January 1, 2005; 51(1): 84 - 92. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Leinonen, P. Wu, and U.-H. Stenman Epitope Mapping of Antibodies against Prostate-specific Antigen with Use of Peptide Libraries Clin. Chem., December 1, 2002; 48(12): 2208 - 2216. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Michel, G. Deleage, J.-P. Charrier, J. Passagot, N. Battail-Poirot, G. Sibai, M. Jolivet, and C. Jolivet-Reynaud Anti-Free Prostate-specific Antigen Monoclonal Antibody Epitopes Defined by Mimotopes and Molecular Modeling Clin. Chem., May 1, 1999; 45(5): 638 - 650. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. H. Black, C. L. Grass, J. Leinonen, U.-H. Stenman, and E. P. Diamandis Characterization of Monoclonal Antibodies for Prostate-specific Antigen and Development of Highly Sensitive Free Prostate-specific Antigen Assays Clin. Chem., March 1, 1999; 45(3): 347 - 354. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
