Prostate-Specific Antigen: A Prognostic
Indicator of Prostate Pathophysiology

Copyright © 1996, Richard J. Ablin, The Ablin Foundation
Abstract


Since the identification of prostate-specific antigen (PSA), continued
technological advances have provided highly sensitive assays for the
quantitation of PSA. Given its absence of disease-specificity, and recent
detection at low levels in an increasing number of non-prostatic tissues,
PSA is far from being the perfect "tumour" marker. However, the
positive predictive value of PSA for assessing cancer risk, nonetheless
presently makes PSA the most useful "tumour" marker for monitoring
progression and response to treatment among patients with prostate
cancer. Earlier detection through screening for elevated levels of PSA,
while controversial, has been proposed as a way to promote earlier
treatment with hopes to decrease prostate cancer mortality. Various
PSA-related concepts, including the ratio of "free" PSA and complexes
of PSA with the protease inhibitor, alpha1-antichymotrypsin, to total
PSA, have been proposed and placed within diagnostic and management
algorithms. Elevations of PSA in other irregularities of the prostate,
notably in benign prostatic hyperplasia, and the increasing frequency and
number of non-prostatic tissues, including those in the female, expressing
PSA, have implications on future immunoassays for PSA, as well as the
molecular basis for its anomalous expression and physiological
function(s).

Introduction

The identification and subsequent quantitation of tissue- or
cell-type-specific antigens may provide insight into the malignant potential
of tumours, i.e., their growth rate and metastatic capability. In endeavors
toward this approach, the identification of select antigens of the human
prostate have been applied to assess the usefulness of immunoassays for
these antigens in the blood, secretions and/or tissue fluids which may
augment or supersede conventional biochemical and cytochemical
methods, as markers for the diagnosis and management of prostatic
cancer (PCa).

Background

Of the literature known to this author, Flocks et al. (1) were the first to
demonstrate the antigenicity of human prostatic tissues. Subsequent
studies by several investigators working independently have led to the
identification and characterization of tissue-specific antigens and their
epitopes of the prostate, prostatic fluid and seminal plasma. (2) Of these,
two predominant antigens -prostate-specific antigen (PSA) and prostate
acid phosphatase (PAP) have played clinically significant roles as
markers of PCa (2). In keeping with the title of this overview, the focus
will be to highlight some of the salient aspects of the identification and
utilization of PSA relative to the diagnosis and management of PCa, with
brief commentary of its biological function(s).

Identification and Characterization

Concomitant with the immunochemical identification and diversity thereof
of one or more tissue-specific antigens of the normal, benign and
malignant human prostate as PAP by Ablin et al. (3, 4) an antigen distinct
from PAP, i.e., PSA, was also identified. (3, 5) In the absence however
at the time of these early observations by Ablin et al. in 1970 (3, 5) of the
yet to come zealous interest in PAP and PSA, they laid dormant until the
report, nine years later, of what the biomedical community has mistakenly
accepted as the de novo identification of PSA by Wang et al. (6)
Recognition of the inherent limitations in the application of immunoassays
of PAP for the diagnosis of PCa (7 and references therein) directed the
collective attention of several investigators, including Wang et al. (6), to
the earlier identified prostate tissue-specific antigen shown by Ablin et al.
(3, 5) to be distinct from PAP. An almost immediate result, was the
confirmation and purification of this antigen by Wang et al. (6), who
referred to it initially as prostate antigen, and subsequently as PSA. The
earlier demonstrated presence of PSA in normal, benign and malignant
prostatic tissue by Ablin et al. (3-5), was also confirmed by Wang et al.
(6)

Parenthetically, and for completeness, it is noted that subsequent to the
initial identification of PSA (3, 5),independent confirmation of PSA in
seminal plasma (3, 5), variously referred to as gamma-seminoprotein (8),
E1 (9), and p30 . (10), was made.

First identified by Ablin et al. (3, 5) and purified and characterized by
Wang et al. (6), PSA was demonstrated to be a single chain 240 amino
acid glycoprotein (11), with a molecular weight of 33,000-34,000 (6).
The primary gene structure and amino acid sequence of PSA exhibits a
high degree of similarity with other serine proteases of the kallikrein
family and participates in the liquefaction of semen following ejaculation
(11-13). PSA exists in multiple forms; as a protease, PSA in the serum
forms complexes with protease inhibitors, i.e., alpha1-antichymotrpysin
(ACT); alpha1-proteinase inhibitor, alpha2-macroglobulin, and
inter-alpha-trypsin inhibitor (14). The major proportion of PSA in serum
appears in complex with ACT (14, 15).

Immunoquantitation of Serum Levels

Initial evidence suggesting a possible clinical application of PSA was
presented by Papsidero et al. (16), who reported the identification of
PSA in the serum of PCa patients. Subsequent studies by Kuriyama et al.
(17), demonstrated the clinical usefulness of PSA in monitoring PCa,
wherein: i) pretreatment chemotherapy levels of PSA were of prognostic
value for survival in patients with advanced disease (Stage D2) and ii)
patients who underwent curative therapy for localized disease, who
subsequently developed metastases, showed increasing PSA levels
preceding to or corresponding with recurrence of disease. The further
prognostic value of PSA for tumour progression was demonstrated in
patients with Stages B2 to D1, by an elevation in PSA 12 months (mean
lead time) prior to tumour recurrence in 92% of the patients evaluated
(18).

Additionally, from a review of these early studies, comparative evaluation
of five serially measured suggested prognostic markers for progression of
PCa, i.e., PSA; PAP; acid phosphatase; total alkaline phosphatase and
bone alkaline phosphatase, showed PSA to be the most reliable (19).
Similarly, comparative evaluation of PSA, but at the tissue level, with
proposed "universal" tumour markers, i.e., carcinoembryonic antigen;
non-specific cross reacting antigen and beta-chorionic gonadotrophin,
showed PSA to be the most sensitive (20).

Immunohistochemical Localization

PSA initially, with rare instances of its detection at very low levels with
polyclonal, but not monoclonal PSA antibody in non-prostatic
malignancies (21), has been thought to be produced exclusively by the
epithelial cells lining the acini and ducts of the prostate. Based on this
specificity, the immunohistochemical localization of antibodies to PSA
(22), as well as PAP (23), were demonstrated to serve as cell-
type-specific markers for the identification of prostatic epithelial cells and
the prostatic histogenesis of poorly differentiated and/or unknown
metastases (22, 23). PSA has however, most recently been observed by
Diamandis et al. in 30% of female breast cancer tissue extracts in
association with steroid hormone-receptor positive tumours (24), a
primary ovarian carcinoma (25), and in the milk of lactating women (26).
The possible implications of these recent observations is addressed under
"Conclusion."

Perspective Considerations

Immunoquantitation of Serum Levels

Based on the demonstrated presumptive prognostic utilization of the
quantitation of serum levels of PSA in PCa, substantially large patient
populations have been evaluated. For this purpose, a monoclonal
solid-phase immunoradiometric assay (Tandem-R PSA [Hybritech, San
Diego, California]) or a polyclonal competitive radioimmunoassay
(Pros-Check PSA [Yang Laboratories, Bellevue, Washington]) have
been widely used. As the two assays rely on different calibration scales,
the upper limits of normal PSA are 4.0 ng/ml for the monoclonal
(m-PSA) assay and 2.5 ng/ml for the polyclonal (p-PSA) assay. In this
regard, it is important to note that while the two assays show close linear
correlation (27), the p-PSA assay yields values 1.5-1.85 times those of
the m-PSA assay (27, 28). It is therefore critical to making any
comparison between PSA values obtained with the two assays to note
the assay used and to make the necessary conversion. With rare
exception (29), quantitation of serum PSA by these assays has
demonstrated that PSA is the most sensitive and reliable marker
presently available for monitoring the progression of PCa and response
to therapy (30-33).

Given the developing significance of the detection of PSA-ACT
complexes as a means of differentiating PCa from benign prostatic
hyperplasia (BPH), it is noteworthy that comparison of PCa patient sera
with assays for PSA-ACT and the Tandem-R PSA assay have shown an
excellent correlation for the detection of PSA-ACT (14, 15). The ability
of the Tandem-R PSA assay to detect PSA-ACT is important when
making comparisons between assays for PSA.

A number of parameters have been evaluated in concert with PSA in an
endeavor to improve the ability to predict prognosis in PCa. Beyond the
scope of the present considerations, mention of the use of the relationship
between local nuclear deoxyribonucleic acid tumour ploidy and PSA,
although requisite of further investigation is noteworthy.

As evident from reports of the American Cancer Society National
Prostatic Cancer Detection Project (34, 35) and others (36-39), there is
tremendous interest and optimism in the use of PSA for screening and
staging of PCa. As stated by Catalona (37): "The result is objective,
quantitative, and obtainable independently of the examiner's skill, and the
procedure is more acceptable to patients than other screening
procedures".

In spite of the interest and optimism, the role of PSA as a solitary criteria
in screening and staging remains controversial. This controversy may be
viewed perhaps in two categories. In the first, the principal concerns are
inherently related to the fact that: i) PSA, although specific for prostatic
tissue (until the recent observations by Diamandis and co-workers
[24-26]) is not tumour-specific and ii) PCa is heterogeneous, with
subpopulations of cells that vary in their synthesis, as well as possibly
secretion of PSA. In the second, the concerns focus on: i) whether
screening will detect what are generally thought to be latent, or clinically
indolent cancers, present in approximately 30% of men >50 years of age
and ii) whether the morbidity and mortality associated with treatment of
these tumours will exceed those of the disease itself. A further concern,
mentioned for the purpose of completeness, but beyond the scope of the
present review, is the cost-related clinical utility of screening.

From the point of view of staging, wide variations in PSA levels exist in
many patients with either localized or advanced metastatic disease (40,
41). As such, except within broad ranges, levels of serum PSA poorly
predict tumour stage on an individual basis.

In terms of the use of PSA for screening for early detection of PCa, a
major limitation has been false-positive elevations produced by BPH
(30). Additionally, other irregularities of the prostate, e.g., prostatitis (42)
and prostatic ischemia and/or infarction (43), can be associated with
elevated levels. Conversely, as pointed out by Oesterling et al. (44), up
to "... 40% of men with organ-confined cancer who undergo radical
prostatectomy ... have a normal serum PSA value."

Results of the quantitation of levels of PSA in BPH and PCa and their
relationship to gland weight may serve to illustrate the dilemma of
false-positive elevations produced by BPH. PSA levels are elevated
approximately 0.3 ng/ml/g of BPH tissue compared with 3.0 ng/ml/g of
cancer (45). As such, a patient with BPH, and a gland volume of 60 cc,
may thus have a PSA value indistinguishable from that produced by a
normal-sized, 30 cc gland in a patient with a 3 cc cancer (35).

Studies by Stamey (46) further demonstrated that gland weight is the
most important non-cancer variable in increases in PSA level. On this
basis it has been proposed (47-49) that measurement of gland volumes
by transrectal ultrasound (TRUS) may aid in discrimination of PSA level
increases due to cancer or BPH, since by TRUS the specific gravity of
the prostate approximates 1.0, and the volume estimation of TRUS may
be considered nearly equivalent to the weight of the gland (50).
Therefore it has been suggested that adjusting the range of normal PSA
levels to gland volumes, i.e., by estimating PSA density (PSAD) rather
then level, it may be possible to reduce the number of false-positive PSA
determinations (51, 52). PSAD has been particularly recommended for
ascertaining the clinical significance of modestly elevated levels of PSA,
i.e., in the range of 4.1-10.0 ng/ml (51).

Concomitant with introduction of the use of PSAD, several other
PSA-related concepts (indices) directed toward enhancement of the
early detection capability of PSA by improving its sensitivity and
specificity are worthy of consideration.

Observations of a change in PSA levels over a defined unit of time, i.e.,
PSA velocity (PSAV), has been suggested. The concept of PSAV
introduced by Carter et al. (53) is based on their demonstration that a
rate change in serum PSA of >0.75 ng/ml/year distinguished patients with
PCa from BPH with 90% specificity and 100% specificity from control
subjects, with a sensitivity of 60%. Using this concept, Brawer et al. (39)
noted that a 20% increase in serum PSA during one year indicated a
significant risk for PCa.

Investigating the distribution of PSA levels in a large population of healthy
men as a function of age and prostatic volume, Dalkin et al. (54) and
Oesterling et al. (44) independently found that levels of serum PSA
directly correlated with patient age in association with the increase in
prostatic volume with advancing age. On this basis, it has been
recommended (44) that rather than relying on a single-age related
reference range, it is more appropriate to have age-specific PSA
reference ranges. The use of age-specific PSA reference ranges is
suggested to increase the specificity for detecting more clinically
significant cancers in older men and increase the sensitivity for detecting
more potentially curable cancers in young men (44).

In investigating the observation that the major proportion of PSA in
serum appears in complex with ACT (14, 15), Stenman et al. (14)
observed that the ratio of PSA-ACT to total PSA is higher in PCa than
BPH. Following this observation, subsequent studies have confirmed and
extended the potential utility in applying the ratios of PSA-ACT:PSA and
"free" PSA:PSA to reduce the number of false-positives in differentiating
early stage PCa from BPH (55-57).

Therefore, with attention to the foregoing considerations, the use of PSA
as part of an initial evaluation of asymptomatic and symptomatic patients
becomes more realistic.

Immunohistochemical Quantitation

The capability to utilize immunohistochemical identification of PSA and
PAP as prostatic cell-type-specific markers have been employed and
extended to provide additional knowledge of the relationship of these
antigens and prostatic growth and function.

Staining for PSA (20, 58-60), and occasionally PAP (58, 60), decrease
in poorly differentiated primary tumours and in metastases, the latter in
which, they may even be absent. The progression of disease in a small
group of untreated PCa patients with Stage A2, whose biopsy specimens
exhibited decreased staining for PSA and PAP, has suggested (60) that
weak staining for these normally immunoreactive antigens might be
prognostic of potentially more aggressive neoplasms. Parenthetically, of
particular interest in this regard, are observations of decreasing PSA and
PAP in fine needle aspirate cytosols with increasing cytologic grade,
tumour stage and tumour ploidy from diploid to aneuploid (61).

Given that tissue PSA (and PAP) has been shown to be age-dependent
correlating with androgen levels (62) and that PCa is heterogeneous in its
responsiveness to antiandrogen therapy, one must interpret with caution
the relevancy of decreased staining for PSA and PAP, and in fact
perhaps question the value of staining for PSA in poorly differentiated
PCa. Of particular significance in this regard are recent studies by
Pretlow et al. (63) in which the heterogeneity in PSA expression
observed immunohistochemically in PCa, and BPH tissues, was
confirmed by quantitation of PSA in extracts prepared from the same
tissues examined microscopically. In accord with earlier comparative
studies of the diversity of PSA in normal, benign and malignant prostatic
tissues (4), PSA was found to be expressed at a significantly lower level
in malignant than the benign prostate (63). The possible significance of
the decrease of PSA is considered further under "Prospective
Considerations."

In further application of the principal of the cell-type-specificity of
antibodies to PSA, Hamdy et al. (64) detected circulating PSA-positive
cells by flow cytofluorography in the peripheral blood of patients with
metastatic PCa. The presence of PSA-positive cells showed a higher
degree of sensitivity and specificity in predicting a positive bone scan,
indicative of the presence of bony metastases, than serum levels of PSA
(64). The circulating PSA-positive cells may have represented a
subpopulation of tumour cells with distinct metastatic properties or, host
immunocytes which had taken-up PSA (64).

Directed toward a more sensitive means to identify circulating
haematogenous micrometastases of PCa, the initial observations by
Hamdy et al. (64) have been extended by the use of the reverse
transcriptase-polymerase chain reaction (RT-PCR) for PSA mRNA
expression (65-67) (65-67). RT-PCR assays for PSA mRNA
expression may prove particularly useful in providing: i) a more sensitive
means of staging compared to conventional modalities, e.g., computed
tomography and magnetic resonance imaging, to identify extraprostatic
disease prior to radical surgery for what is often mistaken as localized
PCa, ii) identification of patients requisite of other treatment modalities,
and iii) confirmation of the curative status of treatment and/or the possible
necessity for further treatment.

A possible caveat (68) in reference to the identification of circulating
PSA-positive cells and/or PSA mRNA are immunoelectron microscopic
observations by Sinha et al. (69) of the localization of PSA not only in
prostatic epithelial cells, but also in neutrophils and macrophages in the
prostate. The presence of PSA in neutrophils and macrophages suggests
that at least part of PSA escaping from the prostatic epithelium and
ductal system is phagocytosed and released in the serum at some other
site (69). As PSA thereby released may become re-phagocytosed in the
peripheral blood, differentiation of PSA-positive circulating prostatic
cancer cells from PSA-positive cells of the reticuloendothelial system,
and/or the source of PSA mRNA in the peripheral blood, is critical in
assessing their respective significance. This possible alternative means of
transport of extracellularly released PSA raises questions as to our
current understanding of the pathway of PSA transport, and perhaps
even the interpretation of serum PSA levels.

Other concerns impacting on the clinical utility of RT-PCR detection of
PSA mRNA and which await longer follow-up include: i) implications
from recent evidence of the expression of low levels of PSA mRNA in
non-prostatic tissues (70) and ii) the importance of the presence of the
identification of haematogenously shed PSA-positive cells relative to their
potential to develop distant metastases.

Prospective Considerations

PSA is an important marker, not only for PCa, but for other prostatic
irregularities (see comments below), and in most instances "... after
treatment is a harbinger of disease (PCa) recurrence" (71). However, the
role of PSA, "free" or in complex with protease inhibitors, notably ACT,
as the sole criteria in screening and staging remains controversial.

As emphasized earlier, the use of PSA as part of an in initial evaluation of
asymptomatic and symptomatic patients in concert with attention to
PSAD, PSAV, or age-specific PSA reference ranges to be followed by
the use of other criteria, e.g., digital rectal examination and TRUS, in
patients with PSA levels above a defined cutoff level, places the use of
PSA for PCa within a realistic framework for further evaluation. In this
regard, numerous presentations at the recent Annual Meeting of the
American Urological Association (72) focused on such further evaluation
of these PSA-related concepts. The relevant aspects of these
presentations are included below. However, their number in the majority
of cases, exceeded the practicality of individual referencing in this review.
The interested reader is therefore referred to the Proceedings issue of the
Annual Meeting in the Journal of Urology (72).

As previously considered, the concept of PSAD may aid in distinguishing
early PCa from patients with BPH and other prostatic irregularities.
However, PSAD has been criticized (44) from the point of view that: i)
the epithelial:stromal ratio varies from one patient to another, and as it is
only the epithelium that produces PSA, two prostates of similar size can
produce markedly different amounts of PSA, ii) there is an error in
determining precise prostatic volume, and iii) other factors, e.g., age, may
influence serum PSA. In confirmation of this criticism, recent reports (72)
of the application of PSAD demonstrated that TRUS calculated volumes
(essential to PSAD), substantially underestimated, and to a lesser degree,
overestimated prostate tumour volume. Imprecise volume assessment,
and thereby inaccurate determination of PSAD, may have an effect on
patient management in cases where selection of treatment is PSAD
dependent. In addition, PSAD alone and in combination with
age-specific PSA reference ranges showed no advantage over the use of
the normal PSA level, defined as no greater than 4.0 ng/ml, in enhancing
the specificity of PSA, or in the staging of PCa.

In regard to PSAV, the difficulty with rate of change in serum PSA,
according to Oesterling et al. (73), is that on an individual basis, there can
be substantial assay and biological variation in the level of PSA from one
determination to the next. For example, Riehmann et al. (74, as cited by
Oesterling et al. [73]) found the biological variation of PSA was as much
as 55% over a one-year period in patients without PCa. Recently
reported studies of PSAV (72) have upheld this initially expressed
concern. This variability may necessitate that the periodic evaluation of
changes in PSA be extended to >2 years to permit distinction of changes
in PSA associated with PCa from those due to BPH, before
recommending a biopsy based on PSA. Given this period of observation,
a 20% increase/year in PSA, independent of age, would indicate the
necessity for a biopsy.

When PSAV was evaluated in combination with age-specific PSA
references ranges, "age-specific PSAV" (a new PSA-related concept)
yielded a 91% sensitivity and 83% specificity among men <74 years of
age for PCa (75).

Placed within a diagnostic algorithm, the use of age-specific PSA
reference ranges and digital rectal examination have been suggested (73)
to increase the sensitivity for detecting PCa at an earlier, potentially
curable stage in young men and increase the specificity for detecting more
clinically significant PCa in older men.

As with PSAD and PSAV, several investigators have independently
evaluated and reported (72) the utility of age-specific PSA reference
ranges alone, vs. PSA, and in combination with PSAD and/or PSAV in
substantially large patient and control populations. The general consensus
from these studies was that, with the possible exception of "age-specific
PSAV," age-specific PSA ranges alone or in combination with PSAD,
had no advantage over PSA alone in detecting PCa and that 4.0 ng/ml
should be the cutoff used irrespective of age. In patients >70 years of
age, a decrease in false-positives was achieved by increasing the upper
limit of normal to 7.5 ng/ml vs. the suggested 6.5 ng/ml (44), but only
with a decrease in sensitivity.

Utilization of the ratio of PSA-ACT:PSA reduces the number of
false-positives, and may facilitate improved discrimination between PCa
and BPH. However, the further evaluation of larger patient cohorts,
including, as pointed out from a recent study showing a decrease of
immunoreactivity of PSA with stored sera (57), the use of fresh serum
samples, and an understanding of the pathophysiology of the observed
differences in the degree of PSA-ACT complex formation, are needed.

In reference to differences in PSA-ACT complex formation, comparative
immunohistochemical and in situ hybridization studies of normal, benign
and malignant human prostatic tissues have been made (76, (77).
Co-localization of ACT in PSA-containing epithelium in normal and
malignant prostate, with the latter characterized by greater variation and
decrease in staining in high vs. low grade tumours have been observed
(76, 77). No staining, or only occasional focal staining for ACT were
seen, nor were ACT transcripts, in PSA-producing epithelium of benign
tissues (77). The absence of ACT in PSA-containing benign tissue may
provide some explanation as to the lower proportion of PSA-ACT
complexes noted in the serum of some patients with BPH (14, 55, 56).

In the long run, only further prospective studies and evaluation therein of
the foregoing PSA-related concepts will provide definitive answers to
their utility. In this regard, it is important to view current and subsequent
PSA-related concepts as dynamic and evolving phenomena, as well as,
occasionally of semantics. Exemplifying this, is the recent suggestion,
based on the observation of lower levels of "free" PSA in patients with
PCa than with BPH, of the measurement (or expression) of "free" PSA in
serum as a means to increase the specificity of PSA screening for PCa
(78). As it has been noted (14, 15, 55) that the greater proportion of
PSA in patients with PCa compared to those with BPH appears in
complex with ACT, it follows that patients with PCa will have lower
levels of "free" PSA.

Not to delude oneself in failing to recognize the necessity, if PSA is to be
used for screening for distinguishing early PCa from elevated levels of
serum PSA in BPH and other prostatic irregularities, we should also
perhaps give consideration to the use of PSA as a marker of prostate
pathophysiology in general. Most certainly, prostatitis and BPH are
significantly sufficient medical problems to warrant such consideration.
With continued endeavors toward achieving appropriate stratification of
reference ranges for serum PSA for nonmalignant and malignant diseases
of the prostate this may become a reality. Particularly in BPH, where one
finds two types, i.e., the fibromuscular stromal form vs. the glandular
(epithelial) form, the ability to determine abnormal epithelial growth, for
which PSA is a unique marker, may facilitate the selection of more
appropriate therapy. Of perhaps equal, or even more concern to the
question of screening with PSA, is the PCa patient who has a normal
serum PSA level.

Towards a possible solution to the foregoing, further utilization of PSA as
a marker may not only be derived from its continuing evaluation in
various clinical settings alone, and in concert with other parameters, but
from an understanding of its biological (physiological) function(s). In this
regard, brief comment of the function(s) of PSA, beyond that of the
liquefaction of semen, although not definitive, may be somewhat
revealing.

The proteolytic activity of PSA has been demonstrated by its ability to
hydrolyze, e.g., insulin A and B chains; recombinant interleukin-2;
ovalbumin and fibrinogen, and to be inhibited by protease inhibitors such
as phenylmethylsulfonyl fluoride, leupeptin and zinc (11, 79).

Inhibition by zinc is perhaps particularly interesting in view of the inverse
relationship between zinc concentration and prostate pathology, e.g., the
level of zinc is reduced to about one-third of normal in the malignant
prostate (80, 81). In view of the association of proteases with tumour
progression and metastases, reduced concentrations of zinc, contributing
to increased proteolysis by PSA, may be functionally most significant
(79, 82), e.g., in metastatic PCa there is increased fibrinolysis and
inhibition of fibrin formation (83).

Observations of the progressive decrease in tissue levels of PSA in the
benign vs. malignant prostate, and in association with a progression from
well to poorly differentiated tumours, when compared to levels in the
normal prostate may be reflective of a normal growth controlling function
of PSA. Also noted in other tumours, the loss of normal antigens with
dedifferentiation from the normal to the malignant state has been
suggested as possibly indicating the absence of a particular growth
controlling (homeostatic) factor from neoplastic cells (84). If maintenance
of normal growth requires such a "self-marker," the absence of this
marker may indicate loss of the property of "contact inhibition" of the cell,
resulting in uncontrolled proliferation (85).

Conclusion

PSA is not a tumour-specific marker, yet as reviewed herein a great deal,
perhaps too much has been expected from PSA. Even when PSA is
placed in various diagnostic and treatment algorithms, it suffers problems.
Specifically, and as stated by the late Willet F. Whitmore, Jr. (86) "...they
(the algorithms) cannot define metastatic potential, the principal concern
and dominant cause of prostatic cancer death." Nonetheless, as
exemplified from the continued investigations of its clinical applicability,
design of PSA-related concepts and never ceasing statistical
manipulations (often beyond one's imagination) found in the current
literature, even more is yet anticipated.

Perhaps, what is needed for PCa more than the design of additional
PSA-related concepts and statistical manipulations thereon, is the
application of improved methods of detection, e.g., the RT-PCR assays
for PSA mRNA recently reported, directed toward providing
stage-related prognostic information and most of all, delineation of
prognostic factors that define metastatic potential.

Some of the zealousness for PSA may be tempered by the recent
demonstration of PSA in female breast cancer and milk of lactating
women. In particular, this apparently more than coincidental
non-prostatic expression of PSA has implications on future
immunoassays for PSA, notably its heretofore thought of cell- type
specificity, as well as its physiological function(s).

In consideration of the molecular basis of the apparent anomalous
expression of PSA, a possible caveat is the existence in women of the
male counterpart of the prostate (also known as the paraurethral or
Skene's glands) shown to have PSA (87). Given observations of the
association of PSA in breast cancer with steroid hormone-receptor
positive tumours, one may envision the existence of a complex regulatory
gene network controlling the expression of PSA. As such, a given tissue
may, depending on the state of cellular differentiation, express previously
repressed genes after neoplastic transformation. Also, and not mutually
exclusive, somatic mutations may lead to specific changes in PSA genes
in cancer cell clones.

With possibly a needed refocus on PSA, attention should be directed
toward further studies defining its physiological function(s) and therein its
role in the pathophysiology of the prostate, for which PSA may serve as
a marker of prostatic disease in general.

Acknowledgement

The interest and unsolicited initiative of Mark Haythorn, an Information
Scientist, in communicating with OncoLink on my behalf and informing
me of their possible interest in this review and support in part, from the
Robert Benjamin Ablin Foundation for Cancer Research for the
acquisition of literature references used in this review, are most gratefully
acknowledged.


References

1. Flocks RH, Urich VC, Patel CA et al. J. Urol., 84:134, 1960.
2. Ablin RJ, Whyard TC. In Farnsworth WE, Ablin RJ (Eds.): The
Prostate as an Endocrine Gland. CRC Press, Inc., Boca Raton, 149,
1990.
3. Ablin RJ, Bronson P, Soanes WA et al. J. Immunol., 104:1329, 1970.

4. Ablin RJ. Cancer, 29:1570, 1972.
5. Ablin RJ, Soanes WA, Bronson P et al. J. Reprod. Fertil., 22:573,
1970.
6. Wang MC, Valenzuela LA, Murphy GP et al. Invest. Urol., 17:159,
1979.
7. Ablin RJ. IRCS Med. Sci., 13:1042, 1985.
8. Hara M, Koyanagi Y, Inoue T et al. Jpn. J. Legal Med., 25:322,
1971.
9. Li TS, Beling CG. Fertil. Steril., 24:134, 1973.
10. Sensabaugh GF, Crim D. J. Forensic Sci., 23:106, 1978.
11. Watt KWK, Lee PL, M'Timkulu T et al. Proc. Nat'l. Acad. Sci.
(USA), 83:3166, 1986.
12. Lilja H. J. Clin. Invest., 76:1899, 1985.
13. Lilja H, Oldbring J, Rannevik G et al. J. Clin. Invest., 80:281, 1987.
14. Stenman U-H, Leinonen J, Alfthan H et al. Cancer Res., 51:222,
1991.
15. Lilja H, Christensson A, Dahl(n U et al. Clin. Chem. 37:1618, 1991.
16. Papsidero LD, Wang MC, Valenzuela LA et al. Cancer Res.,
40:2428, 1980.
17. Kuriyama M, Wang MC, Lee C-L et al. Cancer Res., 41:3874,
1981.
18. Killian CS, Yang N, Emrich LJ et al. Cancer Res., 45:886, 1985.
19. Killian CS, Emrich LJ, Vargas FP et al. J. Nat'l. Cancer Inst.,
76:179, 1986.
20. Purnell DM, Heatfield BM, Trump BF. Cancer Res., 44:285, 1984.
21. Bilgrami S, Singh NT, Shafi N et al. Lancet, 344:1371, 1994.
22. Nadji M, Tabei SZ, Castro A et al. Cancer, 48:1229, 1981.
23. Ablin RJ. In Peeters H (Ed.): Protides of Biological Fluids. Pergamon
Press, Oxford, 225, 1979.
24. Diamandis EP, Yu H, Sutherland JA. Breast Cancer Res. Treat.,
32:301, 1994.
25. Yu H, Diamandis P, Levesque M et al. Cancer Res., 55:1603, 1995.

26. Yu H, Diamandis EP. Clin. Chem., 41:54, 1995.
27. Hortin GL, Bahnson RR, Draft M et al. J. Urol., 139:762, 1988.
28. Chan DW, Bruzek DJ, Oesterling JE et al. Clin. Chem., 33:1916,
1987.
29. Goldrath DE, Messing EM. J. Urol., 142:1082, 1989.
30. Stamey TA, Yang N, Hay AR et al. N. Eng. J. Med., 317:909,
1987.
31. Oesterling JE, Chan DW, Epstein JI et al. J. Urol., 139:766, 1988.
32. Hudson MA, Bahnson RR, Catalona WJ. J. Urol., 142:1011, 1989.
33. Partin AW, Carter HB, Chan DW et al. J. Urol., 143:747, 1990.
34. Mettlin CJ, Lee F, Drago J et al. The American Cancer Society
National Prostatic Detection Project. Cancer, 67:2949, 1991.
35. Littrup PJ, Lee F, Mettlin C. Ca- A Cancer Journal for Clinicians,
42:198, 1992.
36. Cooner WH, Mosley BR, Rutherford CL, Jr. et al. J. Urol.,
143:1146, 1990.
37. Catalona WJ, Smith DS, Ratliff TL et al. N. Eng. J. Med., 324:1156,
1991.
38. Labrie F, Dupont A, Suburu R et al. J. Urol., 147:846, 1992.
39. Brawer MK, Beatie J, Wener MH et al. J. Urol., 150:106, 1993.
40. Brawer MK, Lange PH. J. Endourol., 3:227,1989.
41. Scardino PT. Urol. Clin. North Amer., 16:635, 1989.
42. Robles JM, Morell AR, Redorta JP et al. Eur. Urol., 14:360, 1988.
43. Glenski WJ, Malek RS, Myrtle JF et al. Mayo Clin. Proc., 67:249,
1992.
44. Oesterling JE, Jacobsen SJ, Chute CG et al. J. Amer. Med. Assoc.,
270:860, 1993.
45. Stamey TA, Kabalin JN. J. Urol., 141:1070, 1989.
46. Stamey TA, Kabalin JN, McNeal JE et al. J. Urol., 141:1076, 1989.

47. Littrup PJ, Kane RA, Williams CR et al. Radiology, 178:537, 1991.
48. Littrup, PJ, Williams CR, Egglin TK et al. Radiology, 179:49, 1991.
49. Lee F, Littrup PJ, Loft-Christensen L et al. Cancer, 70(Suppl.):211,
1992.
50. Miyashita H, Watanabe H, Ohe H et al. Prostate, 5:453, 1984.
51. Benson MC, Whang IS, Olsson CA et al. J. Urol., 147:817, 1992.
52. Kane RA, Littrup PJ, Babaian R et al. Cancer, 69:1201, 1992.
53. Carter HB, Pearson JD, Metter J et al. J. Amer. Med. Assoc.,
267:2215, 1992.
54. Dalkin BL, Ahmann F, Southwick P et al. J. Urol.,
149(Suppl.):413A, 1993.
55. Christensson A, Bj(rk T, Nilsson O et al. J. Urol., 150:100, 1993.
56. Leinonen J, L(vgren T, Vornanen T et al. Clin. Chem., 39:2098,
1993.
57. Stenman U-H, Hakama M, Knekt P et al. Lancet, 344:1594, 1994.
58. Stein BS, Vangore S, Peterson RO. Urology, 24:146, 1984.
59. Keillor JS, Aterman KJ. Urol., 137:894, 1993.
60. Epstein JI, Eggleston JC. Hum. Pathol., 15:853, 1984.
61. Stege R, Lundh B, Tribukait B et al. J. Urol., 144:299, 1990.
62. Goldfarb DA, Stein BS, Shamszadeh M et al. J. Urol., 136:1266,
1986.
63. Pretlow TG, Pretlow TP, Yang B et al. Int. J. Cancer, 49:645, 1991.

64. Hamdy FC, Lawry J, Anderson JB et al. Brit. J. Urol., 69:392,
1992.
65. Moreno JG, Croce CM, Fischer R et al. Cancer Res., 52:6110,
1992.
66. Deguchi T, Doi T, Ehara H et al. Cancer Res., 53:5350, 1993.
67. Katz AE, Olsson CA, Raffo AJ et al. Urology, 43:765, 1994.
68. Ablin RJ. Brit. J. Urol., 71:761, 1993.
69. Sinha AA, Wilson MJ, Gleason DF. Cancer, 60:1288, 1987.
70. Smith MR, Biggar S, Hussain M. Cancer Res., 55:2640, 1995.
71. Zagars GK, Pollack A. Cancer, 72:832, 1993.
72 Proceedings American Urological Association. J. Urol.,
153(Suppl.4):1A, 1995.
73. Oesterling JE, Cooner WH, Jacobsen SJ et al. Urol. Clin. North
Amer., 20:671, 1993.
74. Riehmann M, Rhodes PR, Cook TD et al. Urology, 42:390, 1993.
75. Pearson JD, Carter HB, Metter EJ et al. J. Urol.,
153(Suppl.4):465A, 1995.
76. Bjartell A, Bj(rk T, Matikainen M-T et al. Urology, 42:502, 1993.
77. Björk T, Bjartell A, Abrahamsson PA et al. Urology, 43:427, 1994.
78. Catalona WJ, Smith DS, Wolfert RL et al. J. Urol.,
153(Suppl.4):312A, 1995.
79. Ban Y, Wang MC, Watt KWK et al. Biochem. Biophy. Res.
Comm., 123:482, 1984.
80. Bush IM, Berman E, Nourkayhan S et al. American Urological
Association, 69th Annual Meeting, St. Louis, Missouri, 110, 1974.
81. Muntzing J, Nilsson T, Polacek J. Scand. J. Urol. Nephrol., 8:87,
1974.
82. Stein RS, Ablin RJ. Unpublished observations.
83. Kreisle W, Ahmann R, Feinberg W et al. Proc. Amer. Assoc.
Cancer Res., 31:168,1990.
84. Nairn RC, Richmond HG, McEntegart MG et al. Brit. Med. J.,
2:1335, 1960.
85. Ablin RJ. In Tannenbaum M (Ed.): Urologic Pathology: The
Prostate. Lea & Febiger, Philadelphia, 33, 1977.
86. Whitmore WF, Jr. Lancet, 343:1263, 1994.
87. Ablin RJ. Clin. Chem., 35:507, 1989.