CLIN.CHEM.37/10, 1781-1787 (1991)
Immunoreactivity of Plasma Parathyrin-Related Peptide: Three Region-Specific
Radioimmunoassays and a Two-Site Immunoradiometric Assay Compared
Wendy A. Ratdliife,’ Sarah Norbury,’ Richard A. Stott,’ David A. Heath,2 and John G. Ratcliffe’
We measured parathyrin (parathyroid hormone)-related
peptide (PTHRP) in plasma by three region-specific RIAs
and compared them with an established two-site immunoradiometric assay (IRMA) of PTH APi -86 in samples
from control subjects and from patients with primary
hyperparathyroidism (PH) and humoral hypercalcemia of
malignancy (HHM). The two direct AlAs of PTHRP1-34
and PTHRP37-67 were specific for regions ,9-1 8 and
52-61, respectively. In the extraction AlA of PTHRP1 -34
we used an affinity gel containing a monoclonal antibody
specific for the 17-27 sequence; cross-reacting PTHRP
species eluted from the gel were assayed by the RIA of
PTHRP1-34. PTHRP1-86 plasma concentrations by IRMA
were <0.23 pmol/L in control subjects and patients with
PH, and were significantlyincreased in patients with HHM
(mean 6.1 pmol/L, P <0.001).
In contrast, plasma
PTH APi -34 concentrations were not significantly different in the three groups; in HHMpatients, the mean was
190 pmol/L. Plasma PTHRP37-67 concentrations were
similar in control and PH groups and, although significantly increased in HHM patients (mean 440 pmol/L, P
<0.002),
showed poor diagnostic discrimination.
PTHRP1-34 concentrations after affinity extraction of
plasma were also significantly higher in HHM patients
(mean 10.7 pmol/L, P <0.001), but showed incomplete
diagnostic discrimination.We conclude that the diagnostic
utility of the direct RlAs for quantifying PTHRP is markedly
inferior to the IRMA of PTH APi -86.
AddItIonal Keyphras.s: hyperparathyroidism
of malignancy
.
hypercalcemia
Parathyrin (PTH)-related peptide (PTHRP), a paraneoplastic factor first isolated in 1987 (1-3), shares
N-terminal homology with PTH (4), which allows it to
act via PTH receptors on classical target tissues such as
bone and kidney.3 P’FHRP can reproduce all of the major
biochemical features of the clinical syndrome of humoral hypercalcemia of malignancy (I{HM); e.g., hypercalcemia, increased nephrogenous AMP excretion, and
increased osteoclastic bone resorption (5-8). Since 1987,
overwhelming evidence has accumulated that PTHRP
produced by tumors is an important
mediator of the
Department
of Clinical Chemof Medicine, Queen Elizabeth Medical
‘Wolfson Research Laboratories,
istry,
and 2Department
Centre, Birmingham,
B15 2TH, U.K.
3Nonstandard
abbreviations: PTH, parathyrin
(parathyroid
hormone); IRMA, immunoradiometric
assay; PTHRP, parathyrinrelated protein; PBS, phosphate-buffered
saline; MAb, monoclonal
antibody; HHM, humoral hypercalcemia of malignancy; and PH,
primary hyperparathyroidism.
Received February 27, 1991; accepted August 16, 1991.
hypercalcemia associated with malignancy. Recent reviews have summarized information
on the tissuedistribution of PTHRP and its proposed pathophysiological
roles, gene structure and expression, and the relationship of structure to function (9, 10).
Little information
is available currently about the
major secreted and circulating forms of PTHRP (9).
Because the minimum sequence required for FFH bioactivity
is residues 1-30 (11), the bioactive species
circulating
in HHM should possessan intact N-terminus. Immunoassays developed to date have largely been
directed
at this region. Immunoassays of plasma
PTHRP include direct RIAs of PTHRP1-34 (12) and
PTHRP1O9-138 (13), RIAs of FFHRP1-34 after affinity
extraction of plasma (14) or after extraction of plasma
with C2 cartridges (15), and direct two-site immunoradiometric
assays (mMAs) of P’FHRP1-74 (13) and
P’FHRP1-86 (16). The concentrations
of PTHRP in
plasma measured in these different assays vary considerably, and increased concentrations of plasma FFHRP
have been reported in 40-90% of patients with HHM.
In this study, we measured concentrations of plasma
PTHRP with three region-specific RIAs and compared
our results with an established two-site IRMA of
PTHRP1-86 to assessthe diagnostic utility of the assays
in control subjects and patients with primary hyperparathyroidism (PH) and HHM. The BIAs, which involve the use of antisera and monoclonal antibodies
(MAbs) produced against the N-terminal 1-34 and midsequence 37-67 regions, were two direct BIAs for
PTHRP1-34 and 37-67, and one RIA for FFHRP1-34
after affinity extraction of plasma.
Materials and Methods
Materials
PTHRP1-34 and [Tyr0J PTHRP1 34were from Peninsula Labs., St. Helens, Merseyside, U.K. PTHRP1-86
was from
Bachem, Saffron Walden, Essex, U.K.
PTHRP37-67 and subfragments of PTHRP1-34 and
P’FHRP37-67 used in cross-reaction
studies were prepared by Alta Bioscience, University of Birmingham,
Birmingham, U.K. The preparation and characterization of MAbs to PTHRP1-34 (17) and rabbit antisera to
FFHRP1-34 (17) and PTHRP37-67 (16) have already
been described. Sac-Cel (anti-rabbit immunoglobulin)
was from IDS, Washington, Tyne and Wear, U.K.
PTH1-84 was assayed by “N-tact” IRMA (Incatar, Wokingham, Berkshire, U.K.). Disposable polystyrene minicolumns (7O-Mm pore-size filters) were from Pierce Warriner,
Chester, Cheshire, U.K. Conditioned medium
from keratinocytes
was provided by Ms. A. Blight,
CLINICALCHEMISTRY,Vol.37, No. 10, 1991 1781
Hospital, Birmingham, U.K. Human donor serum for recovery pools for the extraction
assay was obtained from the Blood Transfusion Service,
Birmingham,
U.K. Sep-Pak C18 cartridges were from
Millipore, Watford, Herts., U.K. Sources of other reagents have been described previously (16,17).
Birmingham
Accident
Patients
The following specimens were assayed in direct BIAs
of PTHRP1-34 and 37-67 and the IRMA of P’FHRP1-86.
The control subjects were 30 laboratory staff members.
In 19 patients with PH, mean serum calcium and
PTH1-84 concentrations were 2.90 mmol/L (range 2.7 13.63) and 29.4 pmol/L (range 4.1-179), respectively; the
FFH reference range is 0.9-4.0 pmolIL. The diagnosis of
PH was later confirmed in nine patients by the normalization of their serum calcium after parathyroidectomy.
In the HHM groUp, 31 patients with hypercalcemia had
solid tumors, suppressedconcentrations of plasma PTH,
and detectable PTHRP1-86 concentrations (16). At the
time of sampling, the mean corrected concentration of
serum calcium was 3.45 mmol/L (range 2.58-4.46).
P’FH1-84, measured in 29 of the patients, was <0.5
pmolJL in 28 patients and 2.7 pmol/L in one patient,
whose corrected calcium was 2.58 mmol/L after treatment to lower serum calcium. The majority of the HHM
patients had advanced malignancy: metastatic disease
was present in 17 (this may be an underestimate because a systematic search for metastases was not possible in many cases). Several patients had received treatments to lower the serum calcium concentration, including five who had received bisphosphonate therapy. The
primary
tumor sites included lung (eight; metastatic in
three, and by histology squamous carcinoma in four),
esophagus (one), renal cortex (three), ovary (two), bone
(one; synovial cell sarcoma), breast (two), malignant
pheochromocytoma (one), pancreas (three; islet cell tumor in two), abdominal mass (two), prostate (two), liver
(one), bladder (one), and an unknown primary (four).
PTHRP1-34 was measured by RIA after affinity extraction of plasma from control subjects (24 laboratory
staff members) and 15 patients with PH and 15 with
HHM. Because of the large volume of plasma required,
someof the specimens differed from those assayed by the
other methods. In the patients with PH, the mean serum
calcium and PTH1-84 concentrations were 2.89 mmol/L
(range 2.60-3.70) and 11.9 pmoIIL (range 4.1-42.3),
respectively. Seven of the 15 HHM patients were common to the group in which PTHRP was measured by
RIA and IRMA. The mean corrected concentration of
serum calcium was 3.51 mmol/L (range 2.81-4.42), and
the serum PTH1-84 concentrations measured in 13
patients were <0.5 pmolIL. The sites of the primary
tumor were liver (two), breast (three), lung (four; histology in two showed squamous carcinoma), pancreas
(one), bladder (one), myeloma (one), esophagus (one),
ovary (one), and unknown (one). Serum calcium concentrations were corrected for the effect of low concentrations of serum albumin: corrected calcium (minollL) =
total calcium (mmollL) + 0.02 [44) - albumin (g/L)]. The
1782
CLINICAL CHEMISTRY, Vol. 37,
No. 10, 1991
upper limit of the reference range for corrected calcium
is 2.60 mmol/L.
Methods
RIA of PTHRP1-34. This assay involves a specific,
high-avidity
rabbit
antiserum
to PTHRP1-34,
FFHRP1-34 as standard,
and I-1abeled
[Tyr#{176}]PTHRP1-34 as tracer. [Tyr#{176}]-P1’HRP1-34
was iodinated
with Chlorainine-T and purified by passage through a
C18Sep-Pak cartridge; the tracer obtained had a specific
activity
of -150 Ci/g and a shelflife of four weeks at 4#{176}C
(17). The antiserum cross-reacted with PThRP9-34, but
not with PTHRP18-34 or PTHRP12-21, which suggests
that it recognized the 9-18 region of PTHRP. Plasma
(100 L) or standard (0-1000 pg, 100 L) was incubated
for 16 h at 4#{176}C
with 1I-labeled [Tyr#{176}]Fl’HRP134
(15 000 counts/mm,
100 zL) and rabbit antiPTHRP1-34 diluted 100 000-fold (100 ML) in a total
incubation
volume of 500 zL. The diluent buffer was
phosphate-buffered
saline (PBS) consisting of sodium
chloride (120 mmoIJL), potassium chloride (2.7 mmol/L),
and phosphate buffer (10 mmol/L), pH 7.4, containing,
per liter, 2.5 g of Polypep, 1 mL of Triton X-100, and 0.1
g of sodium azide. We used Sac-Cel to separate antibodybound and free fractions and counted the radioactivity
of the bound fraction for 1 mm with a Model 1261
Multigamnia counter (LKB, Bromina, Sweden) with
data reduction by RIA Caic II Software (LKB).
RIA of PTHRP37-67. This assay involves a rabbit
antiserum to PTHRP37-67 (16), ‘I-labeled
PFHRP3767 as tracer, and PTHRP37-67 as standard. The antiserum cross-reacted with PTHRP52-61, but not with
PTHRP32-41, 37-46, 42-51, 45-54, 47-56, 50-59, or 5766, suggesting that it recognized the 52-61 region of
PTHRP. P’FHRP37-67 was iodinated in the presence of
Chlorainine-T
toa specific activity of 400-450 Ci/g and
was purified with a C18 Sep-Pak cartridge (17). Plasma
(100 zL) or standard (0-5000 pg, 100 iL) was incubated
for 16 h at 4#{176}C
with ‘25I-labeled PTHRP37-67 (15 000
counts/mm, 100 zL) and rabbit anti-PTHRP (diluted
4000-fold (100 ML) in a total incubation volume of 500
pL. The diluent buffer and separation method were the
same as in the RIA of PTHRP1-34.
Optimization
of immunoextraction
0fPTHRP1 -34. Because of the lack of a marked matrix effect in the
PTHRP1-34 RIA, we prepared recovery pools of human
serum obtained from the Blood Transfusion Service. We
added ‘25I-labeled [Tyr#{176}]-PTHRP1-34
or native PTHRP,
affinity-purified from conditioned medium from keratinocytes (17), to human donor serum (5 mL) and optimized
the efficiency of the extraction on the basis of the
recovery of radioactivity or PTHRP1-34 iminunoreactivity. Initially, we compared five affinity gels: MAbs 5A,
5C, 6C, 9D, or 1D5 coupled to Sepharose (17). MAb 9D
affinity gel [100 1zL of a 40% (by vol) gel suspension]
gave the highest recovery of iodinated PTHRP1-34 and
was selected for subsequent studies. Elution of
FFHRP1-34 iminunoreactivity
from the affinity gel was
greatest when we used 200 pL of HC1 (25 mmol/L, pH
2.5). We examined the effect of the volume extracted on
the recovery of PTHRP1-34. For 5, 7.5, or 10 mL of
serum containing affinity-purified PTHRP1-34 (100 ng/
L), recoveries were 65%, 63%, and 65%, respectively. In
a separate experiment, mean recoveries (n = 3) of
affinity-purified
PTHRP1-34 immunoreactivity
from 5
mL (500 pg of PTHRP1-34) and 10 mL of serum (1000 pg
of P’FHRP1-34) were 75% (71-79%) and 76% (74-78%),
respectively.
Affinity extraction RIA of PTHRP1-34. Plasma (5 niL)
was mixed for 1 h with 100 iL of a 400 mL’L suspension
of affinity gel consisting of a MAb 9D to PTHRP1-34
coupled to Sepharose. The affinity gel was recovered
the filter of a disposable polystyrene mini-column, and
washed twice with 1 mL of assay diluent. PTHRP was
eluted from the gel in 200 L of HC1 (25 mmol/L, pH 2.5),
and the gel was washed with an additional 200 tL of
assay diluent. PTHRP1-34 in the combined acid eluate
and wash (400 ML) was assayed as described above.
Because each extract was assayed in duplicate at two
dilutions representing 25% and 8.3% of the extract, the
acidity of the extract was neutralized
by dilution and by
addition of assay reagents in diluent buffer. We determined the efficiency of extraction for each batch by using
two recovery pools, each consisting of 5 niL of donor
serum (Blood Transfusion
Service) supplemented with
1000 or 500 pg of immunoreactive PTHRP1-34 isolated
by affinity purification
of conditioned medium from keratinocytes. Plasma PTHRP1-34 results were corrected
for the efficiency of the extraction.
Two-site IRMA of plasma PTHRP1-86.
This IRMA has
been described in detail previously (16). Plasma (200
ML) or standard PTHRP1-86 (0-505 pmolJL, 200 zL)
was incubated for 1 h with 100 1zL of a solid-phase
suspension consisting of cellulose coupled to MAb 1D5
specific for FflIRP1-34. The solid phase was washed
and incubated with excess rabbit antiserum (code no.
3592) to PTHRP37-67, diluted 200-fold (50 zL), for 16 h
at 4#{176}C.
After washing the solid phase, we quantified the
bound immunoreactivity
by incubating the solid phase
with 1I-labeled PTHRP37-67 (120 000 counts/mm, 50
1L). The assay diluent consisted of PBS, pH 7.4, contaming, per liter, 20 mL of heat-inactivated horse serum, 10 niL of mouse serum, 1 mL of Triton X-100, and
0.1 g of sodium azide. In one RIA, MAb 1D5 crossreacted with P’FHRP9-34, 18-34, and 17-27, but not with
PFHRP 22-34, suggesting that it recognized the 17-27
agulant.
was patients
separated
immediately
and
stored
region
Patients
ofPlasma
PTHRP.
and plasma
specimens.
We
collected
blood
from
volunteers
and
either
by
syringe
or
into
evacuated
tubes
with
use
of lithium
heparin
as anticoin aliquots at -20 #{176}C.
Gel filtration chromatography.
We used a 100 x 1 cm
Bio-gel P-100 column for gel chromatography of plasma,
eluting with 1 mol/L acetic acid at a flow rate of 2-3
mJ.ih (18). Fractions of 0.55 niL were evaporated to
dryness and reconstituted in 0.3 mL of the appropriate
diluent for the RIA used. The fractionation
range of the
column was calibrated with the following molecular
mass (Da) markers: ovalbumin (45 000), carbonic anhydrase (29 000), cytochrome c (12 400), aprotinin (6500),
and 1I-labeled
PTHRP1-34 (4180).
Statistics.
Because of the skewed distribution
of
PTHRP concentrations,we used nonparametric
methods
for statistical comparison of results between
groups
(Mann-Whitney
U-test). For associations between
groups, we used the Spearman rank correlation.
Resu
RIA of PTHRP1-34. The detection limit of the RIA,
defined as the mass of PTHRP1-34 corresponding to 2
SD from the zero standard, was 5 pg/tube (12pmolfL).
Mean recoveries of 31.3 and 62.6 pg of FFHRP1-34
added to plasma at the time of assay were 93 (SEM
4.8)% and 105 (SEM 4.0)%, respectively. Within-batch
precision (CV), determined by assaying plasma from
normal subjects, was 12.3%, 2.9%, and 2.1% (n = 5) at
P’I’HRP1-34 contents of 22, 99, and 450 pmol/L, respectively. The standard curve for PTHRP1-34 is shown in
Figure 1. Dilution curves for plasma from control subjects or from patients with PH or HHM were approximately parallel to the standard curve (Figure 1).
No significant differences were found in the plasma
PTHRP1-34 concentrations in control subjects (mean
140 pmolJL, range 13-760), patients with PH (mean 360
pmolfL, range 32-1530), and patients with HHM (mean
190 pmol/L, range 21-1110) (Figure 2).
RIA of PTHRP37-67. This RIA has a detection limit
of 20 pg/tube (57 pmol/L), also corresponding
to 2 SD
from the zero standard. When we added 156 and 312 pg
of PTHRP37-67 to two plasma samples at the time of
assay and assayed each twice in duplicate, analytical
recovery was 113% and 117%, and 118% and 119%,
respectively. The within-batch precision (CV), determined with three plasma samples from normal subjects,
was 2.6%, 2.3%, and 3.1% (n = 6) at 260, 370, and 380
C
PH
.
5000
X4
4000
X2
HU
.
NX4X2
N
X4
X2
N
#{149}
COUNTS
BOUND
cpu
300
iooo
2000
.
0
100
10
PTHRP
1-34
Fig. 1. Standard curveforthe direct
AlA
1000
pg/tubs
of PTHRP1-34
Plasma from controlsubjects(C), patients with piimary hyperparathyroldism
(PH), and hypercalcemia associated with malignancy (HM) was assayed neat
(N) and diluted two- and fourfold In assay diluent
CLINICALCHEMISTRY,Vol.37, No. 10, 1991 1783
respectively.
The standard curve is shown in
Figure 3. Plasma from the three groups studied gave
dilution curves approximately parallel to the standard.
Plasma PTHRP37-67 concentrations measured in patients with HHM were significantly higher (mean 440
pmoIIL, range 150-1570; P = 0.002) than in control
subjects (mean 320 pmollL, range 270-680) or in patients with PH (mean 270 pmolIL, range 180-380)
(Figure 4). There was a significant correlation between
plasma PTHRP37-67 and 1-86 concentrations in patients with HHM (r = 0.53; P <0.001).
Affinity extraction and RIA of PTHRP1 -34. The within-batch precision (CV) of affinity-purified FFHRP1-34
(1000 and 500 pg) extracted from 5 mL of serum was
7.5% and 6.5%, respectively (n = 5). In 14 consecutive
assays, the extraction efficiency, based on the mean
recovery of 500 and 1000 pg of affinity-purified
PTHRP1-34 from serum, was 66 (SEM 2.6)% and 68
pmol/L,
(SEM 2.0)%, respectively.
All extracts when diluted
yielded curves approximately parallel to the standard
curve. The detection limit of the assay depended on the
efficiency of the extraction and was in the range ol
0.5-1.5 pmol/L. PTHRP purified from plasma by an
affinity gel directed toward the 17-27 region was assayed by the RIA of PTHRP1-34, which is specific for the
9-18 region of PTHRP. Mean concentrations of plasma
PTHRP1-34 in patients with HHM were significantly
higher (10.7 pmol/L, range 2.7-41.3; P = 0.001) than in
control subjects (2.9 pmol/L, range 0.35-5.7) or in patients with PH (1.7 pmoIIL, range 0.65-4.0) (Figure 5).
RIA of affinity-extracted plasma with either rabbit
anti-PTHRP1-34 (specific for residues 9-18) or MAlI
1D5 (17) (specificfor residues 17-27) showed that >95%
of the initial 1-34 immunoreactivity remained in extracted plasma, suggesting that the extraction step was
selective for a minor speciesof PTHRP. This contrasts
1600
1800
1400
1400
1200
#{149}
1000
1200
PTHRP 37-67
pMOL/L
1000
.
PTHRP 1-34
pMOL/L
800
.
S
600
.
400
S
.
.
S
S
:
400
200
.L 11i1’.
200
C
I.
600,
#{149}
S
PH
0
C
HHM
Fig.2. Plasma PTHRP1-34 concentrations
by directAlA in control
subjects(C)and Inpatients
withPH or HHM
Fig. 4. Plasma PTHRP37-67
Abbreviations as in Fig. 2
PH
by direct
concentrations
HHM
AlA
50
5000
X4
X2
HM
PH
C
NX4
X’2
?i
X4
)2
40
4000
PTHRP 1-34
pMOt./L
COUNTS
BOUND
30
CPM
3000
20
2000
1000
0
100
PTHRP
Fig.
1000
37-57
pg/tubs
3. Standardcurveforthe direct AlAof PTHRP37-67
Abbreviations as In FIg. 1
1784
CLINICALCHEMISTRY,Vol.37, No. 10, 1991
10000
C
PH
HHM
Fig.5. PlasmaPTHRP1-34 concentrations
measuredby AlAafter
affinity extractionof plasma
Abbreviations as in FIg 2; 0, undetectable results
with the high efficiency (>90%) of extraction of
PTHRP1-34 immunoreactivity
from culture fluid from
cells secreting PTHRP (17). Furthermore,
97% of
PTHRP1-34 and 1-86 added to culture fluid was extracted under identical conditions as used for extraction
of plasma.
Two-site IRMA of PTHRP1 -86. The analytical performance of this IRMA has already been described (16).
Concentrations
of plasma PTHRP1-86 were undetectable (<0.23 pmo]JL) in control subjects and in patients
with PH, and were significantly increased in patients
with HHM (mean 6.1 pmolfL, range 0.46-26.5) (Figure
6).
Gel filtration
of plasma.
Gel ifitration
of 1 mL of
plasma from a patient with PH with a PTHRP1-34
concentration (by direct RIA) of 360 pmol/L gave multiple peaks of PTHRP1-34 immunoreactivity,
with approximate molecular masses ranging from 5 to 16 kDa
(Figure 7). The recovery of PTHRP 1-34 immunoreactivity was 78% of that applied to the column.
Discussion
Because of the lack of information
on the molecular
forms of PTHRP in plasma, synthetic fragments of
PTHRP have been used to produce region-specific antibodies for the immunoassays reported to date. Our study
emphasizes the wide variation in absolute concentrations of plasma PTHRP determined by immunoassays of
different design and specificity, and highlights differences in their potential to discriminate clinically defined groups of patients.
Despite the low detection limit of our direct RIA of
PTHRP1-34, plasma concentrations in normal subjects
(mean = 140 pmoIJL) were higher than those of a
previous study in which plasma PTHRP1-34 was widetectable (<25 pmolJL) in the majority of patients (12).
Furthermore, Henderson et a!. (12) found increased
plasma PTHRP1-34 concentrations in patients with
hypercalcemia associated with malignancy and also in a
proportion of patients with primary
and secondary
30
S
PTHRP 1-86
pMOL./L
.
in the present study, no marked
differences in plasma PTHRP1-34 concentrations were
found in the clinical groups studied. These discrepancies
and the limited diagnostic value of these direct RIAs
may be explained in part by the specificities of the
antisera, and may also reflect the presenceof circulating
N-terminal fragments of F1’HRP. The PTHRP1-34 antiseruin used in the present study recognized residues in
the 9-18 region; in the earlier study, evidence was
presented that the antiserum recognized residues at the
extreme C-terminus of the 1-34 peptide (12).
The direct RIA of PTHRP37-67 is the first RIA to be
described with specificity directed at a midregion sequence. The same antiserum
was also used in the
two-site IRMA of PTHRP1-86. The analytical validation
of the RIA was satisfactory, apart from slight overrecovery of PTHRP37-67 added to plasma. The high concentrations of PTHRP37-67 measured in plasma suggest
that peptides containing this midregion sequence(e.g.,
residues 50-60) may be less susceptible to proteolytic
cleavage or metabolism, or may have a longer half-life
in the circulation than do peptides from the N- and
C-termii
(13). Although mean concentrations of
plasma PTHRP37-67 were greater in HHM patients
than in control subjects and patients with PH, the poor
discrimination obtained in individual patients suggests
that this RIA is unsuitable for diagnostic use.
In the affinity-extraction RIA, PTHRP was selectively
extracted
from 5 mL of plasma by an MAb directed
against the 17-27 region of PTHRP, and was subsequently quantified by an RIA that recognizes the 9-18
amino acid sequence. The efficiency of extraction, monitored by using recovery pools containing affinity-punfled PTHRP, was both reproducible and precise. Plasma
PTHRP1-34 concentrations found in normal subjects
were -50-fold lower than those measured by direct RIA,
but at least 10-fold higher than those measured by the
IRMA. By assaying plasma directly by RIA before and
after immunoextraction, we confirmed that the affinity
gel selectivelyextracted a minor proportion of the total
immunoreactivity present. This affinity gel has previously been shown to extract bioactive and immunoreacfive PTHRP efficiently from human and bovine milk
hyperparathyroidism;
PTHRP t-34
20
1
10
2
I I
pglfr.ctlan
3
4
1
II
5
$
.
0.
tO
20
30
60
40
10
So
PflACTION
C
PH
HHM
Fig. 6. PlasmaPTHRP1-86 concentrations
measuredby two-site
IRMA
Abbreviations
as
In FIg. 2; 0, undetectable results (<0.23 pmol/L)
Fig.7. Gelfiltration
chromatography
ofplasmafroma patientwithPH
The positions of elution of ovalbumin (1), carbonic anhydrase (2), cytochrome
c(3), aprodnin (4), and 1251-labeledPTHRP1-34 (5) are shown.The detection
limit
of the assay corresponds
to 15 pg
of PTHRP1.34per fraction
CLINICALCHEMISTRY,Vol.37, No.10, 1991 1785
(18), culture fluid from keratinocytes, and the BEN lung
cancer cell line (17); our results therefore suggest that
only a small proportion of the total N-terminal PTHRP
iminunoreactivity in plasma is structurally similar to
PTHRP from these other sources. Western blotting of
affinity-extracted P’FHRP from culture fluid from keratinocytes and the BEN cell line identified a major
species with a mobility similar to that of recombinant
PTHRP1-141 (J. Emly and W. A. Ratcliffe, unpublished
observations).
Using a similar
RIA, Budayr et a!.
(14) reported PTHRP1-34 concentrationsof <2.5 pmol/L
in normal subjects, and increased concentrations in 55%
of patients with HHM. An extraction RIA in which
plasma was extracted by C2 cartridge measured mean
PTHRP1-34 concentrations of 3.1 pmolfL in normal
subjectsand increased concentrations in 40% of patients
with malignancy and hypercalcemia
(15). These relatively consistent findings obtained by RIA of plasma
extracts suggest that the extracted form of PTHRP may
be similar to bioactive PTHRP in terms of activity in
patients with HHM.
Plasma PTHRP concentrations in patients with PH
were similar to those in normal subjects by all methods
studied, thus confirming earlier reports (13-15). Although mRNA for PTHRP has been shown to be overexpressed in adenomatous and hyperplastic parathyroid
glands (19), and immunoreactive PTHRP has been identified in extracts of abnormal parathyroid tissue (20),
there is no convincing evidence that plasma P’FHRP
contributes to the increased PTH-like activity in patients with PH.
To minimize potential in vitro degradation of PTHRP
before assay (16), we separated the blood rapidly, froze
the plasma at -20 #{176}C,
and performed gel chromatography of plasma in the presence of acetic acid, 1 molIL.
Although we cannot exclude the possibility of a minor
degree of degradation in vitro, the excellent diagnostic
discrimination of the IRMA suggests that sample handling was satisfactory and preserved speciesof biological importance.
The precise molecular basis of the large differences in
plasma PTHRP immunoreactivities
measured by the
RIA compared with those by the IRMA remains to be
established. The high values obtained by direct BIAs,
intermediate values with the extracted RIA, and relatively low values with the IRMA suggest that the BIAs
cross-react with many circulating immunoactive species, whereas the IRMA has highly restricted crossreactivity. Nonspecific interference is unlikely to be the
sole explanation of these variations, given the detailed
analytical validation of the BIAs with different reagents. Furthermore, there is increasing evidence that
the conformation and metabolism of PTHRP is complex.
Nuclear magnetic resonance studies have shown interactions between N- and C-terminal
residues of
PTHRP1-34 that result in a highly compact structure
with differentsurfaces of high positive charge and
hydrophobic character (21). Many potential proteolytic
affinity-extraction
1786 CLINICALCHEMISTRY,Vol.37, No. 10, 1991
sites exist throughout
the peptide. Preliminary chromatographic studies on plasma (Figure 7) revealed
great heterogeneity of peptides of relatively low molecular mass (e.g., 5-16 kDa) and with N-terminal
immunoreactivity,
analogous to the extensive processing we demonstrated in milk (18). In addition, the
extraction of an insignificant proportion of the total
immunoreactivity of plasma PTHRP1-34 by a range of
MAbs coupled to different solid-phase matrices(systems
extensively validated for synthetic fragments of PFHRP
and native forms of PTHRP present in culture media)
suggests that intact or bioactive PTH.RP is a minor
species in plasma in the clinical conditions studied.
There may be important differences in the conformation
of naturally occurring peptide fragments,
in comparison
with synthetic peptides, which restrict epitope recognition by the affinity gels.
Our data from the PTHRP1-86 IRMA suggest that the
true circulating concentrations of intact PTHRP in
normal subjects are extremely low (picomolar or less).
This concurs with studies of an independent two-site
IRMA (13) and cytochemical bioassays, in which we
reported concentrations of PTH-like activity <0.001
pmol/L in maternal plasma (22) and in the low picomolan range in patients with HHM (23). Such values are
compatible with the known bioactivity of PTHRP.
In summary, we have shown that the apparent
iinmunoreactivity of PTHRP in plasma depends critically on assay specificity, thus highlighting
the complexity of circulating molecular species. The diagnostic
utility of the two-site IRMA was markedly superior to the
direct BIAs and, to a lesser extent, to the affinityextraction RIA. Only the IRMA and the affinity-extraction RIA appear to measure molecular forms of the
protein appropriate to its postulated action as an important mediator of HHM.
We thank the Department
of Health, London, and the Trustees
of the Former United Birmingham Hospitals for financial support
and Mr. R. Holderfor statistical advice.
References
1. Moseley JM, Kubota M, Diefenbach-Jagger HD, et al. Parathyroid hormone-related protein purified from a human lung cancer
cell line. Proc Natl Aced Sd USA 1987;84:5048-52.
2. Strewler
like protein
functional
GJ, Stern PH, Jacobs JW, et al. Parathyroid hormonefrom human renal carcinoma cells: structural
and
homology with parathyroid
hormone.J Clin Invest
1987;80:1803-7.
3. Burtis WJ, Wu T, Bunch CH, et al. Identification of a novel
17,000-dalton
parathyroid hormone-like adenylate cyclase-stimulating protein from tumour associated with humoral hypercalcemia of malignancy. J Biol Chem 1987;262:7151-6.
4. Suva IA, Winslow GA, WettenhallREH, et al. A parathyroid
hormone-related protein implicated in malignant hypercalcemia
cloning and expression. Science 1987;237:893-6.
5. Rabbani SA, Mitchell J, Roy DR, Hendy GN, Goltzman D.
Influence of the amino-terminus
on in vitro and in vivo biological
activity of synthetic parathyroid hormone-like peptides of malignancy. Endocrinology 1988;123:2709-16.
6. Thompson DD, Seedor JG, Fisher JE, Rosenblatt M, Rodan GA.
Direct action of the parathyroid hormone-like human hypercalcemic factor on bone. Proc Natl Acad Sci USA 1988;85:5673-7.
7. Pizurki L, Rizzoli R, Moseley J, Martin TJ, Caverzasio J,
Bonjour JP. Effect of synthetic tumoral PTH-related peptide on
iAMP production and Na-dependent Pi transport. Am J Physiol
1988;225:P957-61.
PR, Adam WE, Moseley JM, Martin TJ. Actions of
Bynthetic
parathyroid hormone-related protein (1-34) onthe isolatedrat kidney.J Endocrunol1989;120:45-50.
B. Stewart AF, Broadus AE. Clinical review 16: parathyroid
B. Ebeling
hormone-relatedproteins: coming of age in the 1990s. J Cliii
Endocrinol Metab 1990;71:1410-4.
10. Martin TJ, Suva LI. Parathyroidhormone-related
protein in
hypercalcaemia
of malignancy.
Cliii Endocrinol1989;31:631-47.
11. KempBE, Moseley JM, Ebeling PR, et al. Parathyroidhorrnone-related protein of malignancy, active synthetic fragments.
Science1987;238:1568-70.
12. HendersonJE, Shustik C, Kremer
R, et al. Circulating conientrations of parathyroidhormone-likepeptide in malignancy
snd in hyperparathyroidism.
J Bone Miner Res 1990;5:105-13.
13. Burtis WJ, Brady TG, Orloff JJ, et al. Immunochemical
iharacterization of circulatingparathyroidhormone-related
proLainin patients with humoral hypercalcemia of cancer. N Engi J
Med 1990;322:1106-12.
14. Budayr AA, Nis8enson RA, Klein RF, et al. Increased serum
Levelsof a parathyroid hormone-like protein in malignancy-assonated hypercalcemia. Ann Intern Med 1989;111:807-12.
15. Kao PC, Klee GO,TaylorRL,HeathH. Parathyroidhormonerelated peptide in plasmaof patientswith hypercalcemia
tnalignant lesions. Mayo Clin Proc 1990;65:1399-.407.
and
16. Ratcliffe WA. Norbury 5, HeathDA, Ratcliffe JG. Development
and validation of an immunoradiometric
assay of parathyrin-related protein in unextracted plasma. Clin Chem 1991;37:678-85.
17. Ratdliffe WA. Hughes S, Gilligan MG, Heath DA, RatcliffeJG.
Productionandcharacterizationofmonoclonal antibodies to parathyroid hormone-related protein. J Immunol Methods 1990;127:109-16.
18. Ratcliffe WA, Green E, Emly J, et al. Identification and partial
characterisation
ofparathyroid hormone-related protein in human
and bovinemilk. J Endocrinol 1990;127:167-76.
19. Ikeda K, Arnold A, ManginM, et a]. Expression of transcripts
encoding a parathyroid hormone-related
peptide in abnormalhuman parathyroid tissues. J Clin Endocrinol
Metab1989;69:1240-8.
20. Docherty HM, Ratdliffe WA, Heath DA, Docherty K. Expressionof parathyroidhormone-related
proteinin abnormalhuman
parathyroids. J Endocrinol 1991;129:431-8.
21. Barden JA, KempBE. NMR studiesof a 34-residueN-terminal fragmentofthe parathyroidhormone-related
protein secreted
during humoral hypercalcemiaof malignancy.Eur J Biochem
1989;184:379-94.
22. MonizC, BurtonPBJ,Malik AN, et al. Parathyroidhormone
related peptidein normal human fetal development.
J Mol Endocrinol 1990;5:259-66.
23. GoltzmanD, Stewart AF, BroadusAE. Malignancy-associated
hypercalcemia:
evaluation with a cytochemicalbioassayforpara-
thyroid hormone. J Cliii EndocrinolMetab 1981;53:899-905.
CLINICALCHEMISTRY,Vol.37, No. 10, 1991 1787