Screening for cancer: are biomarkers of value?

– February/March 2011 14 Tumour markers
Molecular forms of prostate specific
antigen (PSA) in serum: clinical and
analytical implications
Prostate specific Antigen (PSA) is widely used as a disease biomarker
for diagnosis and monitoring of prostate cancer (PCa). Numerous
different immunoassays are available for the measurement of
PSA and its subforms in serum. The assays can be referenced to
different laboratory standards and are not interchangeable. Patients
and physicians should be aware of which assay was used, and
longitudinal monitoring should be performed with the same test.
by Dr Katharina Braun, Dr David Ulmert and Dr Hans Lilja
Prostate-specific antigen (PSA) is a kallikrein-related
peptidase encoded by a five
exon gene 7.1 kb (KLK3), one of fifteen
genes clustered in a 280 kb locus on the long
arm of chromosome 19 in the cytogenic
region q13.3-4 [1]. KLK3 (encoding PSA)
and KLK2 (encoding kallikrein-related
peptidase 2 or hK2) share approximately
80% amino acid sequence identity and the
two proteins are produced and secreted at
highly abundant levels by prostate epithelium
although some expression can also
be detected in certain other extra-prostatic
tissues [2].
PSA is synthesised as a 261-amino-acid
(aa) pre-pro precursor that is processed to a
non-catalytic zymogen through removal of
a ≈17-aa signal peptide upon transfer to the
endoplasmic reticulum, whereas the short
activation peptide must be released, e.g. by
hK2, to convert the non-catalytic ≈244-aa
zymogen to the mature 237-aa catalytic
single-chain PSA [2].
Originally called gamma-seminoprotein,
a seminal fluid protein was identified in
1966 and characterised in 1971 by Hara et
al [3]. The authors anticipated that the protein
would be a potential marker for seminal
fluid applicable in the field of forensic
medicine. In 1979, PSA was purified from
prostatic tissue, and was later found to be
identical to gamma-seminoprotein [4]. Subsequently,
several studies recognised PSA
as a potential marker for PCa [5]. The first
assay for PSA in serum was developed by
Kuriyama et al [6] shortly after Papsidero
and coworkers [5] identified PSA in blood.
PSA is synthesised in normal prostate epithelium,
benign prostate hyperplasia (BPH)
and all stages of prostate adenocarcinoma.
The concentration of PSA in seminal fluid
is up to 10⁶ fold higher than in blood [7].
The median concentration of tPSA in blood
is ≈0.7 ng/mL in healthy men at early middle
age [8], whereas in advanced cancer the
amount of PSA in the blood can increase up
to 10,000 fold [7].
Although recent data from the large population-based
randomised trials in Europe
and the US have demonstrated that PSAbased
prostate cancer screening can reduce
mortality from prostate cancer by about half
after fourteen years, these important benefits
are tempered by considerable overdetection
and consequential risks for overtreatment
associated with current screening modalities
[9]. Risk of prostate cancer diagnosis,
metastasis and death from prostate cancer
are very strongly associated with concentration
of PSA in blood [8]. This strong rationale
explains the widespread use of PSA as a
key biomarker to assess disease risk, monitor
therapeutic intervention and disease
recurrence and as a key component in
various prognostic models.
Molecular forms of PSA in serum
PSA added to blood in vitro exists in three
forms: one fraction will occur complexed with
inactivating protease inhibitors, one portion
as non-complexed non-catalytic PSA, and a
third as active PSA entrapped by macroglobulins
[10]. However, the “total PSA” (tPSA)
detected in clinical samples comprises the
sum of the concentration of both free PSA
and PSA complexed to protease inhibitor
ACT [11]. Data from the original discovery
and characterisation of the proportion of free
PSA versus PSA-ACT complexes suggested a
mean free-to-total PSA ratio of 22% (range
7-50%) in patient’s serum samples [11]. Based
on PSA-measurements at early middle age
in a large, highly representative populationbased
cohort of men, the median proportion
of free-to-total PSA in blood has later been
shown to be ≈33% (IQR 28%; 38%) [12].
Complexed PSA
In the blood circulation, the majority of noncatalytic
PSA is covalently complexed with
the protease inihibitor α1-antichymotrypsin
(ACT or SERPINA5). Active PSA can also
be enveloped by α-macroglobulins such as
α2-macroglobulin (A2M) and pregnancy
zone protein (PZP) [10]. Unlike the interactions
with ACT, the complex-formation
with A2M or PZP does not inactivate PSA
although it blocks catalytic PSA from access
to protein substrates [10]. It is noteworthy
that such macromolecules mask epitopes recognised
by commercially available assays and
thus stay undetected by these methods [1,11].
Since the original discovery in the early 1990s
it has been carefully documented that the proportion
of PSA-ACT is higher in men with
PCa compared to men with BPH [6,10], that
the free-to-total PSA ratio is an independent
predictor of prostate cancer risk [9], and that
the free-to-total PSA ratio enhances discrimination
of men with BPH from those with evidence
of PCa beyond that of total PSA alone
[13]. A systematic review and meta-analysis of
66 subsequent studies found that the free-tototal
PSA ratio (“%fPSA”) enhanced the accuracy
in predicting the diagnostic outcome of
a prostate biopsy compared to that based on
tPSA alone [14].
Free PSA and subforms
The non-complexed, free PSA in blood
is a mixture of different inactive forms
circulating unattached to any plasma
proteins. These inactive forms can be
separated into two main fractions: single
chain “intact” forms with or without
truncated remainders of the short activation
peptide, and forms that are inactive
due to internal cleavages. The most