Papers in Press. Published December 14, 2015 as doi:10.1373/clinchem.2015.244012 The latest version is at http://hwmaint.clinchem.org/cgi/doi/10.1373/clinchem.2015.244012 Clinical Chemistry 62:1 218–226 (2016)
Endocrinology and Metabolism
LC-MS/MS Measurement of Parathyroid Hormone–Related Peptide Mark M. Kushnir,1,2* Alan L. Rockwood,1,2 Frederick G. Strathmann,1,2 Elizabeth L. Frank,1,2 Joely A. Straseski,1,2 and A. Wayne Meikle2,3
INTRODUCTION: Parathyroid hormone–related peptide (PTHrP) is involved in activating pathways, allowing tumor cells to form bone metastases. Measurement of PTHrP is used for the diagnosis and clinical management of patients suspected of hypercalcemia of malignancy. We developed an LC-MS/MS method for measuring PTHrP, established sex-specific reference intervals, and assessed the method’s performance. METHODS:
PTHrP was enriched from plasma samples with rabbit polyclonal anti-PTHrP antibody conjugated to magnetic beads. Enriched PTHrP was digested with trypsin, and PTHrP-specific tryptic peptide was analyzed with 2-dimensional LC-MS/MS in multiple reaction monitoring mode.
RESULTS:
The lower limit of quantification was 0.6 pmol/L, and the upper limit of linearity was 600 pmol/L. Total imprecision was ⬍10%. Very poor agreement was observed with the RIA (n ⫽ 207; Deming regression RIA ⫽ 0.059 ⫻ LC-MS/MS ⫺ 1.8, r ⫽ 0.483; Sy兩x ⫽ 3.9). Evaluation of the clinical performance of the assay using samples from patients with and without hypercalcemia (n ⫽ 199) resulted in an area under the ROC curve of 0.874. In sets of consecutively analyzed routine samples of patients assessed for hypercalcemia, the PTHrP positivity rate by RIA (n ⫽ 1376) was 1.9%, and 26.6% by LC-MS/MS (n ⫽ 1705). Concentrations were below the lower limit of quantification in 95.6% of the samples by RIA and 2.0% by LC-MS/MS. CONCLUSIONS: PTHrP is a normal constituent in circulating blood and its concentrations are substantially underestimated by commercial RIAs, causing falsenegative results in samples from patients suspected of hypercalcemia. Our observations suggest a link between increased concentrations of PTHrP in post-
1
ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT; 2 Department of Pathology and 3 Department of Medicine, University of Utah, Salt Lake City, UT. * Address correspondence to this author at: ARUP Institute for Clinical and Experimental Pathology, 500 Chipeta Way, Salt Lake City, UT 84108. Fax 801-584-5207; e-mail [email protected]. Received June 2, 2015; accepted October 19, 2015.
menopausal women with low body mass index and increased incidence of osteoporosis. © 2015 American Association for Clinical Chemistry
The N-terminal sequence of parathyroid hormone– related peptide (PTHrP)4 has close homology with that of parathyroid hormone (PTH). Because both hormones bind to the same receptors, some of the physiological functions of PTHrP are similar to those of PTH (1–5 ). PTHrP acts as an autocrine, paracrine, and endocrine hormone and is able to simulate most of the actions of PTH, including regulation of calcium ion homeostasis, bone resorption, distal tubular Ca reabsorption, and inhibition of proximal tubular phosphate transport. In health, PHTrP regulates bone development by maintaining the endochondral growth plate. It also plays a role during tooth eruption, development of mammary glands, pregnancy, fetal development, and regulation of Ca transfer to milk during lactation (5– 8 ). Circulating concentrations of PTHrP in health are very low but can be increased during pregnancy and lactation and in some nonmalignant diseases (7– 8 ). PTHrP may be present at increased concentrations in patients diagnosed with cancers of the breast, bladder, lung, uterus, and skin (3, 6, 9 ). PTHrP is known to activate pathways that allow tumor cells to form bone metastases, a condition known as hypercalcemia of malignancy (HCM) (6, 9 ). In patients with HCM, PTH is typically suppressed owing to increased blood Ca, whereas uncontrolled release of PTHrP by tumor cells is responsible for the hypercalcemia (8 ). RIA is commonly used for measurement of PTHrP. A number of RIA methods have been developed that use antibodies against 5 epitopes within the PTHrP sequence (10 –12 ). Pandian et al. (11 ) suggested that PTHrP is likely the causative agent of HCM and developed a diagnostic test for its measurement. Fraser et al. (13 ) evalu-
Previously published online at DOI: 10.1373/clinchem.2015.244012 © 2015 American Association for Clinical Chemistry 4 Nonstandard abbreviations: PTHrP, parathyroid hormone–related peptide; PTH, parathyroid hormone; HCM, hypercalcemia of malignancy; LOQ, limit of quantification; MRM, multiple reaction monitoring; PPACK, Phe-Pro-Arg-chloromethylketone.
218
Copyright (C) 2015 by The American Association for Clinical Chemistry
Measurement of PTHrP by LC-MS/MS
ated a laboratory-developed RIA and observed that some healthy individuals had measurable concentrations of PTHrP; the reference interval determined for healthy individuals was ⬍2.6 pmol/L. Concentrations of PTHrP were ⬎2.6 pmol/L in 46% of samples from patients with HCM (13 ). Ikeda et al. (10 ) measured PTHrP with a 2-site RIA in 110 healthy individuals; the observed concentrations were 0.8 (0.01) pmol/L [mean (SD)], with an established upper reference limit of 1.1 pmol/L. Considering the tight distribution of concentrations in healthy individuals (within ⫾0.01 pmol/L), performance of the method could be questioned. The reported limit of quantification (LOQ) of the RIA developed by Wu et al. was 0.1 pmol/L, and PTHrP was undetectable in 75% of individuals (14 ). In patients with HCM and solid tumors, PTHrP concentrations were ⬎2.0 pmol/L. De Miguel et al. (12 ) observed increased PTHrP concentrations in 75% of patients with bone metastases, whereas the method was insufficiently sensitive to measure PTHrP in healthy individuals. Two LC-MS/MS methods for PTHrP have been reported previously. Lu et al. (15 ) developed a method with a reported LOQ of 62 nmol/L, which is insufficient for diagnostic applications. Washam et al. (16 ) developed a qualitative SELDI-TOF MS method for detection of PTHrP peptide (amino acids 12– 48) in plasma of patients with breast cancer. The aim of this work was to develop an LCMS/MS method for measurement of PHTrP, validate the method, establish reference intervals of PTHrP in healthy adults, compare the LC-MS/MS assay with commercially available RIAs, and evaluate the method’s performance. Materials and Methods Detailed information is provided in Supplemental Materials, which accompanies the online version of this article at http://www.clinchem.org/content/vol62/issue1. REAGENTS AND STANDARDS
Calibration standards of PTHrP were prepared from recombinant PTHrP (amino acids 37–122; AVSEHQ LLHDKGKSIQDLRRRFFLHHLIAEIHTAEIRATSE VSPNSKPSPNTKNHPVRFGSDDEGRYLTQETNK VETYKEQPLKTP) purchased from PeproTech in 0.1% BSA at concentrations of 1, 2, 5, 10, and 30 pmol/L. The calibrators were standardized against calibration standards from the PTHrP RIA kit (Immunotech). Rabbit polyclonal anti-PTHrP antibody was purchased from PeproTech. The recombinant 15N-labeled internal standard was diluted in 0.1% BSA to a concentration of 1 nmol/L and frozen at ⫺20 °C until use.
SAMPLE PREPARATION AND ANALYSIS
To a 400-L aliquot of plasma sample, we added 400 L HEPES buffer (pH 7.4) and 20 L internal standard (recombinant [15N]PTHrP) and incubated the samples for 15 min. We added 5 L magnetic bead suspension and incubated the samples with agitation at 10 °C for 3 h. We washed the beads, added 200 L of 25 mmol/L bicarbonate buffer and 10 L of 4 g/L trypsin, and incubated the samples at 37 °C for 3 h. After digestion, we transferred the samples to a 96-well plate and injected 70-L aliquots into the LC-MS/MS. Two-dimensional HPLC separation was performed on an HPLC system consisting of series 1260 and 1290 pumps (Agilent Technologies). We used a Synergy Polar RP, 50 ⫻ 3-mm, 4-m HPLC column (Phenomenex) for first-dimension separation, with a gradient of 99%– 87% A in 2.7 min (A, 10 mmol/L formic acid in water; B, 10 mmol/L formic acid in methanol). For seconddimension separation, we used a Synergy Max-RP, 100 ⫻ 3-mm, 2.5-m column (Phenomenex), with a gradient of 97%– 85% A in 2 min (A, 5 mmol/L acetic acid in water; B, 5 mmol/L acetic acid in acetonitrile). Liquid chromatography separation was at 40 °C. We performed quantitative analysis on an AB5500 mass spectrometer (AB Sciex) with a V-spray ionization source in positive-ion, multiple reaction monitoring (MRM) mode. We monitored mass transitions m/z 498.753720.35 and 499.253721.35 for the YLTQETNK peptide and m/z 504.253729.35 and 504.753730.25 for the internal standard. Qualitative confirmation of PTHrP was assessed by the ratio of the concentrations determined from 2 mass transitions of the targeted peptide and the internal standard (17 ). The stated LOQ and the reference interval of the PTHrP RIA (Immunotech) were 2 pmol/L and ⬍4 pmol/L. Concentrations of intact PTH were measured with quantitative ELISA (Roche Diagnostics), and concentrations of total Ca with a Ca Gen.2 kit on a Cobas c502 (Roche Diagnostics). All tests were performed at ARUP Laboratories (Salt Lake City, UT). METHOD VALIDATION
Method validation consisted of evaluating imprecision; determining LOQ, limit of detection, linearity, carryover, recovery, ion suppression, reference intervals; and comparing the method with commercial RIAs. Detailed information on method validation is provided in online Supplemental Materials. The method was compared with a PTHrP RIA (Immunotech); samples used for evaluation were remaining aliquots of patient plasma samples submitted to ARUP Laboratories for routine testing. We used 3 sets of plasma samples for method evaluation: (a) samples previously analyzed for PTHrP by the Immunotech RIA (n ⫽ 207); (b) samples containing increased Ca (⬎10.5 mg/L) and Clinical Chemistry 62:1 (2016) 219
B
8000
Intensity (cps)
A
4000
Intensity (cps)
600
300
4.56
0 4.50
4.60
4.70
4.80
4.90
5.00
5.10
5.20
Time (min)
4.50
4.60
4.70
4.80 4.90 Time (min)
5.00
5.10
Fig. 1. MRM chromatograms of mass transitions of PTHrP (A) and internal standard (B) in a patient sample containing 1.8 pmol/L PTHrP. Solid lines correspond to the primary mass transitions, and dashed lines correspond to the secondary mass transitions.
low PTH (⬍15 pg/mL) (n ⫽ 88); and (c) samples with Ca within reference intervals and low PTH (⬍15 pg/mL) (n ⫽ 44). Differences in concentrations among the groups were evaluated with the Wilcoxon rank-sum test. We evaluated the suitability of 6 types of blood collection tubes with blood drawn from 5 volunteers. Blood was collected into potassium EDTA, sodium heparin, lithium heparin, serum, serum separation tubes, and Phe-Pro-Arg-chloromethylketone (PPACK) inhibitor tubes (Hematologic Technologies). PTHrP concentrations observed in the samples were compared within each individual. Samples were analyzed fresh and after refrigerated storage for 48 h. We evaluated ion suppression by postcolumn infusion (18 ). A reference interval study for PTHrP was performed with samples from self-reported healthy adult volunteers, 114 men (ages 20 –58 years, median 33) and 122 women (ages 20 – 67 years, median 32). Blood was collected in tubes with PPACK inhibitor, plasma was separated from blood cells within 1 h, and the samples were stored at ⫺70 °C until analysis. Results Product ion mass spectrum of the PTHrP-specific peptide (YLTQETNK) is shown in online Supplemental Fig. 1, and MRM chromatograms for a sample are shown in Fig. 1. Total assay imprecision at the evaluated concentrations was ⬍10% (see online Supplemental Table 1). The LOQ and limit of detection of the method were 0.6 and 0.3 pmol/L, respectively. The method was linear up to 600 pmol/L, with inaccuracy at the highest concentration of 6.0%. No carryover was detected immediately after a sample containing 5000 pmol/L PTHrP. Method recovery was 98%; recovery of the immunoaf220
Clinical Chemistry 62:1 (2016)
finity enrichment was 70%. Concentrations observed in the samples from the dilution integrity experiment agreed with each other within 7%. The quantitative performance and imprecision of the method were evaluated by use of 2 internal standards, recombinant [15N]PTHrP and the “winged” labeled peptide (DDEGRYL*TQETNKVETY), added to the samples before and after affinity enrichment. When the winged peptide was added before enrichment, it was binding to the antibody during the affinity enrichment. The mean imprecision of replicate measurements at 4 evaluated concentrations (9 –50 pmol/L) was 21%, with a mean bias of 61% (data not shown). The higher biases were observed in samples containing higher concentrations of PTHrP. PTHrP is known to be very unstable (11, 12 ). In samples of 4 of the 5 individuals, PTHrP concentrations in samples collected in PPACK tubes were higher than in the other 5 evaluated types of tubes (mean difference, 30%). After 48-h storage at 4 °C, PTHrP concentrations in PPACK plasma were 39% lower than in the fresh samples (P ⫽ 0.036), indicating poor stability of PTHrP in the plasma containing PPACK inhibitor when stored at 4 °C. Degradation was not observed in samples defrosted for aliquoting on ice and then refrozen immediately after aliquoting. The presence of hemolysis, icterus, and lipemia did not affect method performance; no ion suppression was observed at the retention time of the peptide. When we compared the LC-MS/MS method with a commercial PTHrP RIA kit (Immunotech), the Deming regression equation for the comparison was RIA ⫽ 0.059 ⫻ LC-MS/MS ⫺ 1.8; n ⫽ 207; r ⫽ 0.483; Sy兩x ⫽ 3.9 (Fig. 2A).
Measurement of PTHrP by LC-MS/MS
B
45
0.9
35
0.8 0.7
IA = 0.059 × LC-MS/MS − 1.8, r = 0.483, Sy|x = 3.9
30 RIA (pmol/L)
1.0
40
25
Sens i vity
A
20 15
Area Under Curve = 0.874
0.6 0.5 0.4 0.3
10
0.2
5
0.1 01 0.0
0 0
50
100
150
200
250
300
0.0 0.1 0.2
0.3 0.4
0.5 0.6
0.7 0.8
0.9 1.0
1 − Specificity
LC-MS/MS (pmol/L)
Fig. 2. LC-MS/MS method comparison with RIA (Immunotech) for analysis of PTHrP in plasma samples (n = 207) (A); ROC curve for the detection of increased PTHrP as a cause of hypercalcemia in a set of plasma samples from patients with hypercalcemia (n = 91) and samples from self-reported healthy adults with no hypercalcemia (n = 108) (B).
Nonparametric reference intervals for PTHrP established with this method were 0.6 –3.3 pmol/L and 0.6 – 2.2 pmol/L in adult women and men, respectively. Both PTH and PTHrP exert their action on Ca regulation in blood by binding to the same receptors. Considering this, we evaluated the distribution of PTHrP, PTH ⫹ PTHrP, and the ratio of PTHrP/PTH concentrations in plasma samples of healthy adults. Fig. 3 shows distributions of concentrations, and Fig. 4 shows the median concentrations of PTHrP, PTH, PTHrP ⫹ PTH, and the PTHrP/PTH ratio in healthy men and women. Reference intervals for PTH ⫹ PTHrP were 1.5– 8.3 pmol/L in women and 1.2– 8.2 pmol/L in men. Reference intervals for PTHrP/PTH were 0.2–1.4 in women and 0.2–1.0 in men. The data in Table 1 summarize median concentrations and distribution of the concentrations in men and women by age group (decade of life). No statistically significant difference in the distribution of the concentrations and ratios was observed among age groups in men; in women, however, PTHrP concentrations were significantly higher in the 20- to 30-year age group (P ⫽ 0.0025) and in women ⬎50 years old (P ⫽ 0.006) compared to the 30- to 50-year age group (Table 1). PTH ⫹ PTHrP was higher in women ⬎50 years old compared to younger women (P ⫽ 0.0033). In the 20- to 30-year age group, PTHrP/PTH was significantly higher than in the rest of the age groups (P ⫽ 0.0048). In postmenopausal women, PTHrP concentrations were higher than in premenopausal women, but the difference was not statistically significant (P ⫽ 0.14) (see online Supplemental Fig. 2); lower concentrations were observed in
women and men with higher body mass index (see online Supplemental Fig. 3). To assess the performance of the LC-MS/MS assay, we analyzed PTHrP in 2 sets of patient samples submitted for routine testing. In the set of samples with high Ca and low PTH, 42% of samples had concentrations of PTHrP above the established reference intervals, and PTHrP/PTH was above the reference interval in all samples. PTHrP ⫹ PTH was above the reference interval in 24% of samples. In the set of patient samples containing Ca within reference intervals and low PTH, 23% of samples had PTHrP concentrations above the reference interval. Clinical performance of the assay and its ability to detect increased PTHrP concentrations as a cause of hypercalcemia was evaluated by analysis of PTHrP in a set of plasma samples biochemically characterized as hypercalcemic (Ca ⬎10.5 mg/dL and PTH ⬍15 pg/mL; n ⫽ 91) and samples from self-reported healthy adults with concentrations of Ca and PTH within established reference intervals (n ⫽ 108). Logistic regression analysis resulted in an area under the ROC curve of 0.874 (2 test value ⫽ 113 for the model; P ⬍ 0.0001) (Fig. 2B). Data on the diagnostic utility of the assay are summarized in Table 2. Discussion In our method, we analyzed PTHrP by quantifying PTHrP-specific peptide released during tryptic digestion. The peptide 102YLTQETNK110 is in the midClinical Chemistry 62:1 (2016) 221
A
B
5
PTH P = 0.967
100
4
PTHrP P < 0.001
3
PTH H (pg/mL)
PTHrrP (pmol/L)
120
2
80 60 40 20 0
0
Women
D
10
PTHrP + PTH P = 0.040
8
PTHrP/PTH
PT THrP + PTH (pmol/L)
C
Women
Men
6
4
2
Men
2
PTHrP/PTH / P = 0.0016
1.5
1
0.5
0
0
Women
Women
Men
Men
Fig. 3. Distribution of concentrations of PTHrP (A), PTH (B), PTH + PTHrP (C), and PTHrP/PTH (D) in apparently healthy women (n = 120) and men (n = 109).
dle of the PTHrP sequence, and although its concentration does not represent the concentration of bioactive PTHrP, it serves as a marker of the rate of PTHrP biosynthesis. The method used immunocapture of PTHrP from plasma samples followed by tryptic digestion, while it was
B
7.0
7.0
6 6.0 0
6.0
5.0
5.0
pmol/L
pmol/L
A
4.0 3.0
bound to the antibody conjugated to magnetic beads, followed by direct analysis of the digests with triplequadrupole mass spectrometry. The on-bead digestion used in this method allowed for faster analysis without peptide elution, reduction, and alkylation before digestion. Considering that the enrichment takes place on the
4.0
PTHrP PTH
3.0
2.0
2.0
1.0
1.0
Total PTH + PTHrP
0.0
0.0 20–30
31–40
41–50
51–60
Years
61+
20–30
31–40
41–50
51–60
Years
Fig. 4. Median concentrations of PTHrP, PTH, and PTH + PTHrP in samples from apparently healthy women (n = 120) (A) and men (n = 109) (B) by age group (decade of life) (see Table 1 for the distribution of concentrations).
222
Clinical Chemistry 62:1 (2016)
Measurement of PTHrP by LC-MS/MS
Table 1. Concentrations of PTHrP, PTH plus PTHrP, and PTHrP/PTH in plasma samples of self-reported healthy adults by age group.a Age, years
n
Median age, years
PTHrP, pmol/L
PTH + PTHrP, pmol/L
PTHrP/PTH
20–30
52
25.5
1.6 (0.7–4.2)
4.3 (1.3–7.4)
0.60 (0.2–1.86)
31–40
37
34.0
1.3 (0.6–3.5)
4.3 (2.6–8.4)
0.47 (0.17–1.83)
41–50
16
45.0
1.4 (0.2–2.0)
4.5 (0.25–9)
0.49 (0.19–0.80)
51–60
14
56.5
1.9 (1–3.4)
5.3 (3.6–6.9)
0.53 (0.23–1.16)
4
63.0
2.0 (1–3.3)
6.5 (4.5–9.3)
0.59 (0.15–0.73)
6
34.5
2.4 (1.7–3.2)
5.3 (3.5–7.8)
0.79 (0.6–1.0)
20–30
35
27.0
1.2 (0.6–2.0)
3.8 (1.5–8.0)
0.49 (0.16–0.87)
31–40
42
33.0
1.2 (0.6–2.1)
4.3 (1.0–10.1)
0.35 (0.14–1.07)
41–50
24
43.5
1.2 (0.6–2.4)
4.1 (0.6–5.7)
0.45 (0.21–1.0)
51–60
14
53.0
1.3 (0.8–2.4)
4.3 (2.2–10.0)
0.39 (0.14–0.76)
Women
≥61 Breastfeeding women Men
a
Data are median (central 95% concentration) unless noted otherwise.
protein level, accurate quantification of PTHrP requires careful selection of the internal standard. Two stable isotope–labeled internal standards were evaluated, a winged labeled peptide (19 ) and recombinant [15N]PTHrP. The winged internal standard provided partial improvement in performance; quantitative performance substantially improved with the use of recombinant [15N]PTHrP as the internal standard. The YLTQETNK peptide used in this method is unique to PTHrP, has no homology with any other tryptic fragments of human proteins, and is a specific surrogate marker for quantification of PTHrP. Among ⬎8000 analyzed patient samples, we observed ⬍2% samples with a ratio of mass transitions outside 30% of the expected range among the samples containing ⬎2 pmol/L PTHrP. PTHrP in plasma samples is unstable and undergoes rapid degradation (11, 12 ). Of the evaluated types of blood collection tubes, PTHrP concentrations were highest in the tubes with PPACK inhibitor (analyzed immediately after defrosting). After the samples were stored refrigerated for 48 h, PTHrP degraded to concentrations similar to those observed in the other types of
collection tubes. The interindividual variation in the reduction of PTHrP concentrations in samples analyzed fresh and after storage could be related to the differences in concentrations of endogenous proteases causing PTHrP degradation. Similar observations have been reported in earlier publications (11, 12 ). In agreement with our data, the PTHrP degradation rate varied among the samples from different individuals. The data suggest that PTHrP degradation begins at the time of blood collection and plasma separation and imply the necessity to draw blood in tubes containing PPACK inhibitor (20, 21 ). In samples collected in tubes with PPACK inhibitor, PTHrP concentrations were above the LOQ of the method in 99.6% (Fig. 3), with higher concentrations observed in women than men. Concentrations were highest in breastfeeding women (Table 1), compared with the other groups of healthy adults, and in postmenopausal women. We did not observe an association in men between PTHrP concentrations and age. The gaussian distribution of the concentrations in samples of healthy individuals supports tight regulation of PTHrP. Considering that PTHrP plays a role
Table 2. Diagnostic utility of the LC-MS/MS test for PTHrP. Positive predictive value
Negative predictive value
Clinical sensitivity
Clinical specificity
Diagnostic efficiency
1.8
0.88
0.80
0.76
0.89
0.82
2.3 (upper normal for men)
0.94
0.73
0.59
0.96
0.79
3.3 (upper normal for women)
0.98
0.67
0.45
0.99
0.73
Cutoff concentration, pmol/L
Clinical Chemistry 62:1 (2016) 223
in development of mammary glands, lactation, pregnancy, and osteoporosis (5, 7 ), we expected that PTHrP concentrations would be higher in women than men, which is what our data showed. Higher concentrations of PTHrP were observed in individuals with lower body mass index (see online Supplemental Fig. 3) and postmenopausal women (see online Supplemental Fig. 2). On the basis of current knowledge, in health, Ca in blood is regulated by PTH, and PTHrP is often undetectable in blood of healthy individuals (8 ). Our data demonstrate that PTH and PTHrP are present in blood at comparable concentrations. Because both hormones bind to the same receptors, it is likely that they are both responsible for Ca regulation, and the total concentration of PTHrP ⫹ PTH (or possibly a weighted sum of PTHrP and PTH) could be more representative of the status of Ca regulation than the concentration of PTH alone. Our data (Table 1, Fig. 3, and online Supplemental Fig. 2) show that the concentrations of PTHrP and PTHrP ⫹ PTH are higher in postmenopausal women compared with men and premenopausal women. These observations could provide a partial explanation of the high prevalence of osteoporosis in postmenopausal women with low body mass index (22, 23 ). We evaluated the ability of the method to identify PTHrP as a cause of hypercalcemia by analyzing PTHrP in pathologic patient plasma samples. We hypothesized that in many of the patents with unexplained hypercalcemia that tested negative for PTHrP by RIA, PTHrP was not detected because of poor performance of the RIA. In a set of 88 patient samples of patients with hypercalcemia, 42% of samples had increased PTHrP, and the PTHrP/PTH ratio was above the reference interval in all samples of the set. In another set of plasma samples from patients with Ca within reference intervals and low PTH, PTHrP concentrations were above the established reference interval in 23% of patients. Considering that Ca is tightly regulated and its concentration in health is maintained within a narrow physiological range, the observed Ca concentrations in these 2 sets of samples could not be explained by PTH alone. One possibility for the hypercalcemia in some of the patients from the first cohort and normocalcemia in the second cohort could be related to an increase of PTHrP. The highest diagnostic efficiency was observed at a PTHrP cutoff of 1.8 pmol/L; the highest positive predictive value and clinical specificity were observed at the upper cutoff concentration for women (3.3 pmol/L) (Table 2), supporting the clinical value of the assay. In a review of causes of hypercalcemia, Jacobs and Bilezikian (8 ) stated that in many cases hypercalcemia was not associated with increased PTHrP. This observation could be related to the poor performance of available 224
Clinical Chemistry 62:1 (2016)
PTHrP assays, which underestimate PTHrP concentrations, resulting in a very low positivity rate in samples from patients clinically suspected of hypercalcemia. Because of the poor performance, previously available methods were not able to reliably measure PTHrP in healthy individuals (11–14 ), resulting in the assumption that PTHrP is not present in circulating blood in health. In this study, we identified a large number of patient samples with substantial disagreement in results between the commercial RIA and our LCMS/MS assay (Fig. 2). In routine use, among consecutive samples analyzed by RIA (n ⫽ 1376), the positivity rate was 1.9%; among consecutive samples analyzed by LC-MS/MS assay (n ⫽ 1705), the positivity rate was 26.6%. In the above sets of samples, when analyzed by RIA, concentrations were below the LOQ in 95.6% of the samples and in only 2.0% of samples by LC-MS/MS. These results suggest that in many patients with clinical symptoms of hypercalcemia, PTHrP was undetectable by RIA not because the concentrations were too low, but because of the poor performance of the RIA. Considering that samples submitted to reference laboratories for measurement of PTHrP are typically from patients in whom hypercalcemia is clinically suspected, the observed low positivity rate of the RIA (⬍2%) suggests that the RIA produces falsely low results in many samples. On the basis of our data, increased concentrations of PTHrP in samples of patients with suspected hypercalcemia are more common than was stated previously. Our observation on poor agreement between the methods for PTHrP is not unique; poor agreement between the commercial RIAs was described in other studies (11, 12 ). Because of the inadequate analytical sensitivity of the PTHrP RIA and poor stability of PTHrP, it has been assumed that PTHrP is not present in circulating blood in health (8 ) and that measurement of PTHrP is not diagnostically useful if HCM is not suspected. This LCMS/MS method allowed measurements of endogenous concentrations of PTHrP in healthy individuals, and our data showed that PTHrP is present in circulating blood in health. On the basis of our data, endogenous concentrations of PTHrP in blood are comparable to concentrations of PTH. Considering the difference in half-lives of PTH and PTHrP (approximately 4 min and a few hours, respectively) (24, 25 ), and that both PTH and PTHrP are involved in Ca regulation (6, 7 ), the total and relative concentrations of PTHrP and PTH could be important for the assessment of Ca regulation in patients with signs of osteoporosis, osteomalacia, and unexplained hypercalcemia. Data on the distribution of PTH ⫹ PTHrP and the PTHrP/PTH ratio in samples from healthy adults (Table 1 and Fig. 3) demonstrated relatively narrow distributions. Future studies will be required to assess associations
Measurement of PTHrP by LC-MS/MS
between PTHrP concentrations and various physiological and pathological conditions. Because increased secretion of PTHrP could be caused by tumors, this PTHrP method may potentially be used as a screening test for early detection of PTHrP-secreting malignancies and conditions associated with altered Ca regulation. In summary, we developed an LC-MS/MS method for the quantitative analysis of PTHrP in plasma samples. The immunoaffinity enrichment and on-bead digestion used in this method could be generally applicable to LC-MS methods for other protein biomarkers. Poor agreement was observed between this LC-MS/MS method and the Immunotech RIA, likely because of the poor analytical sensitivity of the RIA. We established reference intervals for PTHrP, the total concentration of PTHrP ⫹ PTH, and the ratio of the concentrations PTHrP/PTH in samples of self-reported healthy adults. Our data comparing the positivity rate between this LCMS/MS method and commercial RIAs indicate substantial underestimation of PTHrP concentrations by RIAs and underdiagnoses of PTHrP as a cause of hypercalcemia in many patients. Because both PTH and PTHrP have roles in Ca regulation, we believe that PTH ⫹ PTHrP and PTHrP/PTH could be useful markers for the assessment of Ca regulation. In women, concentrations were highest in breastfeeding women and in women ⬍30 and ⬎50 years old. We did not observe an association between PTHrP concentrations and age in men. Higher concentrations of PTHrP were observed in women and men with lower body mass index. These observations
may provide an explanation of the higher prevalence of osteoporosis in postmenopausal women with low body mass index. Finally, given the current state of PTHrP testing in clinical laboratories and lack of reference methodologies, there is a need for standardization among PTHrP methods.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest: Employment or Leadership: None declared. Consultant or Advisory Role: E.L. Frank, ARUP Laboratories, Inc. Stock Ownership: None declared. Honoraria: None declared. Research Funding: This work was supported by the ARUP® Institute for Clinical and Experimental Pathology. Expert Testimony: None declared. Patents: M.M. Kushnir, provisional patent application no. 62/179,488. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. Acknowledgments: We thank the ARUP Institute for Clinical and Experimental Pathology for supporting this project and Kimberly Kalp for technical assistance.
References 1. Moseley JM, Kubota M, Diefenbach-Jagger H, Wettenhall RE, Kemp BE, Suva LJ, et al. Parathyroid hormonerelated protein purified from a human lung cancer cell line. Proc Natl Acad Sci U S A 1987;84:5048 –52. 2. Stewart AF, Wu T, Goumas D, Burtis WJ, Broadus AE. N-terminal amino acid sequence of two novel tumorderived adenylate cyclase-stimulating proteins: identification of parathyroid hormone-like and parathyroid hormone-unlike domains. Biochem Biophys Res Commun 1987;146:672– 8. 3. Strewler GJ, Stern PH, Jacobs JW, Eveloff J, Klein RF, Leung SC, et al. Parathyroid hormonelike protein from human renal carcinoma cells: structural and functional homology with parathyroid hormone. J Clin Invest 1987;80:1803–7. 4. Broadus AE, Mangin M, Ikeda K, Insogna KL, Weir EC, Burtis WJ, Stewart AF. Humoral hypercalcemia of cancer: identification of a novel parathyroid hormonelike peptide. N Engl J Med 1988;319:556 – 63. 5. Wysolmerski JJ, Philbrick WM, Dunbar ME, Lanske B, Kronenberg H, Broadus AE. Rescue of the parathyroid hormone-related protein knockout mouse demonstrates that parathyroid hormone-related protein is essential for mammary gland development. Development 1998;125:1285–94. 6. Suva LJ, Winslow GA, Wettenhall RE, Hammonds RG, Moseley JM, Diefenbach-Jagger H, et al. A parathyroid hormone-related protein implicated in malignant
hypercalcemia: cloning and expression. Science 1987; 237:893– 6. 7. Mannstadt M, Ju¨ppner H, Gardella TJ. Receptors for PTH and PTHrP: their biological importance and functional properties. Am J Physiol 1999;277:665–75. 8. TP Jacobs, JP Bilezikian. Clinical review: Rare causes of hypercalcemia. J Clin Endocrinol Metab 2005;90: 6316 –22. 9. Mundy GR. Mechanisms of bone metastasis. Cancer 1997;80:1546 –56. 10. Ikeda K, Ohno H, Hane M, Yokoi H, Okada M, Honma T, et al. Development of a sensitive two-site immunoradiometric assay for parathyroid hormone-related peptide: evidence for elevated levels in plasma from patients with adult T-cell leukemia/lymphoma and B-cell lymphoma. J Clin Endocrinol Metab 1994;79: 1322–7. 11. Pandian MR, Morgan CH, Carlton E, Segre GV. Modified immunoradiometric assay of parathyroid hormonerelated protein: clinical application in the differential diagnosis of hypercalcemia. Clin Chem 1992;38: 282– 8. 12. de Miguel F, Motellue JL, Hurtado J, Jime´nez FJ, Esbrit P. Comparison of two immunoradiometric assays for parathyroid hormone-related protein in the evaluation of cancer patients with and without hypercalcemia. Clin Chim Acta 1998;277:171– 80. 13. Fraser WD, Robinson J, Lawton R, Durham B, Gal-
14.
15.
16.
17.
18.
19.
lacher SJ, Boyle IT, et al. Clinical and laboratory studies of a new immunoradiometric assay of parathyroid hormone-related protein. Clin Chem 1993;39: 414 –9. Wu TJ, Taylor RL, Kao PC. Parathyroid-hormone-related peptide immunochemiluminometric assay developed with polyclonal antisera produced from a single animal. Ann Clin Lab Sci 1997;27:384 –90. Lu CM, Burton WD, Fitzgerald RL, Deftos LJ, Buchholz BA, Vogel JS, Herold DA. Mass spectrometric immunoassay for parathyroid hormone-related protein. Anal Chem 2002;74:5507–12. Washam CL, Byrum SD, Leitzel K, Ali SM, Tackett AJ, Gaddy D, et al. Identification of PTHrP(12– 48) as a plasma biomarker associated with breast cancer bone metastasis. Cancer Epidemiol Biomarkers Prev 2013; 22:972– 83. Kushnir MM, Rockwood AL, Nelson GJ, Yue B, Urry FM. Assessing analytical specificity in quantitative analysis using tandem mass spectrometry. Clin Biochem 2005; 38:319 –27. Matuszewski BK, Constanzer ML, Chavez-Eng CM. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal Chem 2003;75:3019 –30. Kushnir MM, Rockwood AL, Roberts WL, Hoofnagle AN, Meikle AW. Accurate measurement of thyroglobulin in serum in presence of anti-thyroglobulin autoantibod-
Clinical Chemistry 62:1 (2016) 225
ies. Clin Chem 2013;59:982–90. 20. Hutchesson AC, Hughes SV, Bowden SJ, Ratcliffe WA. In vitro stability of endogenous parathyroid hormonerelated protein in blood and plasma. Ann Clin Biochem 1994;31:35–9. 21. Kushnir MM. Are samples in your freezer still good for biomarker discovery? Am J Clin Pathol 2013;140: 287– 8.
226
Clinical Chemistry 62:1 (2016)
22. Mpalaris V, Anagnostis P, Goulis DG, Iakovou I. Complex association between body weight and fracture risk in postmenopausal women. Obes Rev 2015;16:225– 33. 23. Unni S, Yao Y, Milne N, Gunning K, Curtis JR, LaFleur J. An evaluation of clinical risk factors for estimating fracture risk in postmenopausal osteoporosis using an electronic medical record database. Osteoporos Int
2015;26:581–7. 24. Sokoll LJ, Drew H, Udelsman R. Intraoperative parathyroid hormone analysis: a study of 200 consecutive cases. Clin Chem 2000;46:1662– 8. 25. Tov AB, Mandel D, Weissman Y, Dollberg S, Taxir T, Lubetzky R. Changes in serum parathyroid hormonerelated protein in breastfed preterm infants. Breastfeed Med 2012;7:50 –3.
Endocrinology and Metabolism
LC-MS/MS Measurement of Parathyroid Hormone–Related Peptide Mark M. Kushnir,1,2* Alan L. Rockwood,1,2 Frederick G. Strathmann,1,2 Elizabeth L. Frank,1,2 Joely A. Straseski,1,2 and A. Wayne Meikle2,3
INTRODUCTION: Parathyroid hormone–related peptide (PTHrP) is involved in activating pathways, allowing tumor cells to form bone metastases. Measurement of PTHrP is used for the diagnosis and clinical management of patients suspected of hypercalcemia of malignancy. We developed an LC-MS/MS method for measuring PTHrP, established sex-specific reference intervals, and assessed the method’s performance. METHODS:
PTHrP was enriched from plasma samples with rabbit polyclonal anti-PTHrP antibody conjugated to magnetic beads. Enriched PTHrP was digested with trypsin, and PTHrP-specific tryptic peptide was analyzed with 2-dimensional LC-MS/MS in multiple reaction monitoring mode.
RESULTS:
The lower limit of quantification was 0.6 pmol/L, and the upper limit of linearity was 600 pmol/L. Total imprecision was ⬍10%. Very poor agreement was observed with the RIA (n ⫽ 207; Deming regression RIA ⫽ 0.059 ⫻ LC-MS/MS ⫺ 1.8, r ⫽ 0.483; Sy兩x ⫽ 3.9). Evaluation of the clinical performance of the assay using samples from patients with and without hypercalcemia (n ⫽ 199) resulted in an area under the ROC curve of 0.874. In sets of consecutively analyzed routine samples of patients assessed for hypercalcemia, the PTHrP positivity rate by RIA (n ⫽ 1376) was 1.9%, and 26.6% by LC-MS/MS (n ⫽ 1705). Concentrations were below the lower limit of quantification in 95.6% of the samples by RIA and 2.0% by LC-MS/MS. CONCLUSIONS: PTHrP is a normal constituent in circulating blood and its concentrations are substantially underestimated by commercial RIAs, causing falsenegative results in samples from patients suspected of hypercalcemia. Our observations suggest a link between increased concentrations of PTHrP in post-
1
ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT; 2 Department of Pathology and 3 Department of Medicine, University of Utah, Salt Lake City, UT. * Address correspondence to this author at: ARUP Institute for Clinical and Experimental Pathology, 500 Chipeta Way, Salt Lake City, UT 84108. Fax 801-584-5207; e-mail [email protected]. Received June 2, 2015; accepted October 19, 2015.
menopausal women with low body mass index and increased incidence of osteoporosis. © 2015 American Association for Clinical Chemistry
The N-terminal sequence of parathyroid hormone– related peptide (PTHrP)4 has close homology with that of parathyroid hormone (PTH). Because both hormones bind to the same receptors, some of the physiological functions of PTHrP are similar to those of PTH (1–5 ). PTHrP acts as an autocrine, paracrine, and endocrine hormone and is able to simulate most of the actions of PTH, including regulation of calcium ion homeostasis, bone resorption, distal tubular Ca reabsorption, and inhibition of proximal tubular phosphate transport. In health, PHTrP regulates bone development by maintaining the endochondral growth plate. It also plays a role during tooth eruption, development of mammary glands, pregnancy, fetal development, and regulation of Ca transfer to milk during lactation (5– 8 ). Circulating concentrations of PTHrP in health are very low but can be increased during pregnancy and lactation and in some nonmalignant diseases (7– 8 ). PTHrP may be present at increased concentrations in patients diagnosed with cancers of the breast, bladder, lung, uterus, and skin (3, 6, 9 ). PTHrP is known to activate pathways that allow tumor cells to form bone metastases, a condition known as hypercalcemia of malignancy (HCM) (6, 9 ). In patients with HCM, PTH is typically suppressed owing to increased blood Ca, whereas uncontrolled release of PTHrP by tumor cells is responsible for the hypercalcemia (8 ). RIA is commonly used for measurement of PTHrP. A number of RIA methods have been developed that use antibodies against 5 epitopes within the PTHrP sequence (10 –12 ). Pandian et al. (11 ) suggested that PTHrP is likely the causative agent of HCM and developed a diagnostic test for its measurement. Fraser et al. (13 ) evalu-
Previously published online at DOI: 10.1373/clinchem.2015.244012 © 2015 American Association for Clinical Chemistry 4 Nonstandard abbreviations: PTHrP, parathyroid hormone–related peptide; PTH, parathyroid hormone; HCM, hypercalcemia of malignancy; LOQ, limit of quantification; MRM, multiple reaction monitoring; PPACK, Phe-Pro-Arg-chloromethylketone.
218
Copyright (C) 2015 by The American Association for Clinical Chemistry
Measurement of PTHrP by LC-MS/MS
ated a laboratory-developed RIA and observed that some healthy individuals had measurable concentrations of PTHrP; the reference interval determined for healthy individuals was ⬍2.6 pmol/L. Concentrations of PTHrP were ⬎2.6 pmol/L in 46% of samples from patients with HCM (13 ). Ikeda et al. (10 ) measured PTHrP with a 2-site RIA in 110 healthy individuals; the observed concentrations were 0.8 (0.01) pmol/L [mean (SD)], with an established upper reference limit of 1.1 pmol/L. Considering the tight distribution of concentrations in healthy individuals (within ⫾0.01 pmol/L), performance of the method could be questioned. The reported limit of quantification (LOQ) of the RIA developed by Wu et al. was 0.1 pmol/L, and PTHrP was undetectable in 75% of individuals (14 ). In patients with HCM and solid tumors, PTHrP concentrations were ⬎2.0 pmol/L. De Miguel et al. (12 ) observed increased PTHrP concentrations in 75% of patients with bone metastases, whereas the method was insufficiently sensitive to measure PTHrP in healthy individuals. Two LC-MS/MS methods for PTHrP have been reported previously. Lu et al. (15 ) developed a method with a reported LOQ of 62 nmol/L, which is insufficient for diagnostic applications. Washam et al. (16 ) developed a qualitative SELDI-TOF MS method for detection of PTHrP peptide (amino acids 12– 48) in plasma of patients with breast cancer. The aim of this work was to develop an LCMS/MS method for measurement of PHTrP, validate the method, establish reference intervals of PTHrP in healthy adults, compare the LC-MS/MS assay with commercially available RIAs, and evaluate the method’s performance. Materials and Methods Detailed information is provided in Supplemental Materials, which accompanies the online version of this article at http://www.clinchem.org/content/vol62/issue1. REAGENTS AND STANDARDS
Calibration standards of PTHrP were prepared from recombinant PTHrP (amino acids 37–122; AVSEHQ LLHDKGKSIQDLRRRFFLHHLIAEIHTAEIRATSE VSPNSKPSPNTKNHPVRFGSDDEGRYLTQETNK VETYKEQPLKTP) purchased from PeproTech in 0.1% BSA at concentrations of 1, 2, 5, 10, and 30 pmol/L. The calibrators were standardized against calibration standards from the PTHrP RIA kit (Immunotech). Rabbit polyclonal anti-PTHrP antibody was purchased from PeproTech. The recombinant 15N-labeled internal standard was diluted in 0.1% BSA to a concentration of 1 nmol/L and frozen at ⫺20 °C until use.
SAMPLE PREPARATION AND ANALYSIS
To a 400-L aliquot of plasma sample, we added 400 L HEPES buffer (pH 7.4) and 20 L internal standard (recombinant [15N]PTHrP) and incubated the samples for 15 min. We added 5 L magnetic bead suspension and incubated the samples with agitation at 10 °C for 3 h. We washed the beads, added 200 L of 25 mmol/L bicarbonate buffer and 10 L of 4 g/L trypsin, and incubated the samples at 37 °C for 3 h. After digestion, we transferred the samples to a 96-well plate and injected 70-L aliquots into the LC-MS/MS. Two-dimensional HPLC separation was performed on an HPLC system consisting of series 1260 and 1290 pumps (Agilent Technologies). We used a Synergy Polar RP, 50 ⫻ 3-mm, 4-m HPLC column (Phenomenex) for first-dimension separation, with a gradient of 99%– 87% A in 2.7 min (A, 10 mmol/L formic acid in water; B, 10 mmol/L formic acid in methanol). For seconddimension separation, we used a Synergy Max-RP, 100 ⫻ 3-mm, 2.5-m column (Phenomenex), with a gradient of 97%– 85% A in 2 min (A, 5 mmol/L acetic acid in water; B, 5 mmol/L acetic acid in acetonitrile). Liquid chromatography separation was at 40 °C. We performed quantitative analysis on an AB5500 mass spectrometer (AB Sciex) with a V-spray ionization source in positive-ion, multiple reaction monitoring (MRM) mode. We monitored mass transitions m/z 498.753720.35 and 499.253721.35 for the YLTQETNK peptide and m/z 504.253729.35 and 504.753730.25 for the internal standard. Qualitative confirmation of PTHrP was assessed by the ratio of the concentrations determined from 2 mass transitions of the targeted peptide and the internal standard (17 ). The stated LOQ and the reference interval of the PTHrP RIA (Immunotech) were 2 pmol/L and ⬍4 pmol/L. Concentrations of intact PTH were measured with quantitative ELISA (Roche Diagnostics), and concentrations of total Ca with a Ca Gen.2 kit on a Cobas c502 (Roche Diagnostics). All tests were performed at ARUP Laboratories (Salt Lake City, UT). METHOD VALIDATION
Method validation consisted of evaluating imprecision; determining LOQ, limit of detection, linearity, carryover, recovery, ion suppression, reference intervals; and comparing the method with commercial RIAs. Detailed information on method validation is provided in online Supplemental Materials. The method was compared with a PTHrP RIA (Immunotech); samples used for evaluation were remaining aliquots of patient plasma samples submitted to ARUP Laboratories for routine testing. We used 3 sets of plasma samples for method evaluation: (a) samples previously analyzed for PTHrP by the Immunotech RIA (n ⫽ 207); (b) samples containing increased Ca (⬎10.5 mg/L) and Clinical Chemistry 62:1 (2016) 219
B
8000
Intensity (cps)
A
4000
Intensity (cps)
600
300
4.56
0 4.50
4.60
4.70
4.80
4.90
5.00
5.10
5.20
Time (min)
4.50
4.60
4.70
4.80 4.90 Time (min)
5.00
5.10
Fig. 1. MRM chromatograms of mass transitions of PTHrP (A) and internal standard (B) in a patient sample containing 1.8 pmol/L PTHrP. Solid lines correspond to the primary mass transitions, and dashed lines correspond to the secondary mass transitions.
low PTH (⬍15 pg/mL) (n ⫽ 88); and (c) samples with Ca within reference intervals and low PTH (⬍15 pg/mL) (n ⫽ 44). Differences in concentrations among the groups were evaluated with the Wilcoxon rank-sum test. We evaluated the suitability of 6 types of blood collection tubes with blood drawn from 5 volunteers. Blood was collected into potassium EDTA, sodium heparin, lithium heparin, serum, serum separation tubes, and Phe-Pro-Arg-chloromethylketone (PPACK) inhibitor tubes (Hematologic Technologies). PTHrP concentrations observed in the samples were compared within each individual. Samples were analyzed fresh and after refrigerated storage for 48 h. We evaluated ion suppression by postcolumn infusion (18 ). A reference interval study for PTHrP was performed with samples from self-reported healthy adult volunteers, 114 men (ages 20 –58 years, median 33) and 122 women (ages 20 – 67 years, median 32). Blood was collected in tubes with PPACK inhibitor, plasma was separated from blood cells within 1 h, and the samples were stored at ⫺70 °C until analysis. Results Product ion mass spectrum of the PTHrP-specific peptide (YLTQETNK) is shown in online Supplemental Fig. 1, and MRM chromatograms for a sample are shown in Fig. 1. Total assay imprecision at the evaluated concentrations was ⬍10% (see online Supplemental Table 1). The LOQ and limit of detection of the method were 0.6 and 0.3 pmol/L, respectively. The method was linear up to 600 pmol/L, with inaccuracy at the highest concentration of 6.0%. No carryover was detected immediately after a sample containing 5000 pmol/L PTHrP. Method recovery was 98%; recovery of the immunoaf220
Clinical Chemistry 62:1 (2016)
finity enrichment was 70%. Concentrations observed in the samples from the dilution integrity experiment agreed with each other within 7%. The quantitative performance and imprecision of the method were evaluated by use of 2 internal standards, recombinant [15N]PTHrP and the “winged” labeled peptide (DDEGRYL*TQETNKVETY), added to the samples before and after affinity enrichment. When the winged peptide was added before enrichment, it was binding to the antibody during the affinity enrichment. The mean imprecision of replicate measurements at 4 evaluated concentrations (9 –50 pmol/L) was 21%, with a mean bias of 61% (data not shown). The higher biases were observed in samples containing higher concentrations of PTHrP. PTHrP is known to be very unstable (11, 12 ). In samples of 4 of the 5 individuals, PTHrP concentrations in samples collected in PPACK tubes were higher than in the other 5 evaluated types of tubes (mean difference, 30%). After 48-h storage at 4 °C, PTHrP concentrations in PPACK plasma were 39% lower than in the fresh samples (P ⫽ 0.036), indicating poor stability of PTHrP in the plasma containing PPACK inhibitor when stored at 4 °C. Degradation was not observed in samples defrosted for aliquoting on ice and then refrozen immediately after aliquoting. The presence of hemolysis, icterus, and lipemia did not affect method performance; no ion suppression was observed at the retention time of the peptide. When we compared the LC-MS/MS method with a commercial PTHrP RIA kit (Immunotech), the Deming regression equation for the comparison was RIA ⫽ 0.059 ⫻ LC-MS/MS ⫺ 1.8; n ⫽ 207; r ⫽ 0.483; Sy兩x ⫽ 3.9 (Fig. 2A).
Measurement of PTHrP by LC-MS/MS
B
45
0.9
35
0.8 0.7
IA = 0.059 × LC-MS/MS − 1.8, r = 0.483, Sy|x = 3.9
30 RIA (pmol/L)
1.0
40
25
Sens i vity
A
20 15
Area Under Curve = 0.874
0.6 0.5 0.4 0.3
10
0.2
5
0.1 01 0.0
0 0
50
100
150
200
250
300
0.0 0.1 0.2
0.3 0.4
0.5 0.6
0.7 0.8
0.9 1.0
1 − Specificity
LC-MS/MS (pmol/L)
Fig. 2. LC-MS/MS method comparison with RIA (Immunotech) for analysis of PTHrP in plasma samples (n = 207) (A); ROC curve for the detection of increased PTHrP as a cause of hypercalcemia in a set of plasma samples from patients with hypercalcemia (n = 91) and samples from self-reported healthy adults with no hypercalcemia (n = 108) (B).
Nonparametric reference intervals for PTHrP established with this method were 0.6 –3.3 pmol/L and 0.6 – 2.2 pmol/L in adult women and men, respectively. Both PTH and PTHrP exert their action on Ca regulation in blood by binding to the same receptors. Considering this, we evaluated the distribution of PTHrP, PTH ⫹ PTHrP, and the ratio of PTHrP/PTH concentrations in plasma samples of healthy adults. Fig. 3 shows distributions of concentrations, and Fig. 4 shows the median concentrations of PTHrP, PTH, PTHrP ⫹ PTH, and the PTHrP/PTH ratio in healthy men and women. Reference intervals for PTH ⫹ PTHrP were 1.5– 8.3 pmol/L in women and 1.2– 8.2 pmol/L in men. Reference intervals for PTHrP/PTH were 0.2–1.4 in women and 0.2–1.0 in men. The data in Table 1 summarize median concentrations and distribution of the concentrations in men and women by age group (decade of life). No statistically significant difference in the distribution of the concentrations and ratios was observed among age groups in men; in women, however, PTHrP concentrations were significantly higher in the 20- to 30-year age group (P ⫽ 0.0025) and in women ⬎50 years old (P ⫽ 0.006) compared to the 30- to 50-year age group (Table 1). PTH ⫹ PTHrP was higher in women ⬎50 years old compared to younger women (P ⫽ 0.0033). In the 20- to 30-year age group, PTHrP/PTH was significantly higher than in the rest of the age groups (P ⫽ 0.0048). In postmenopausal women, PTHrP concentrations were higher than in premenopausal women, but the difference was not statistically significant (P ⫽ 0.14) (see online Supplemental Fig. 2); lower concentrations were observed in
women and men with higher body mass index (see online Supplemental Fig. 3). To assess the performance of the LC-MS/MS assay, we analyzed PTHrP in 2 sets of patient samples submitted for routine testing. In the set of samples with high Ca and low PTH, 42% of samples had concentrations of PTHrP above the established reference intervals, and PTHrP/PTH was above the reference interval in all samples. PTHrP ⫹ PTH was above the reference interval in 24% of samples. In the set of patient samples containing Ca within reference intervals and low PTH, 23% of samples had PTHrP concentrations above the reference interval. Clinical performance of the assay and its ability to detect increased PTHrP concentrations as a cause of hypercalcemia was evaluated by analysis of PTHrP in a set of plasma samples biochemically characterized as hypercalcemic (Ca ⬎10.5 mg/dL and PTH ⬍15 pg/mL; n ⫽ 91) and samples from self-reported healthy adults with concentrations of Ca and PTH within established reference intervals (n ⫽ 108). Logistic regression analysis resulted in an area under the ROC curve of 0.874 (2 test value ⫽ 113 for the model; P ⬍ 0.0001) (Fig. 2B). Data on the diagnostic utility of the assay are summarized in Table 2. Discussion In our method, we analyzed PTHrP by quantifying PTHrP-specific peptide released during tryptic digestion. The peptide 102YLTQETNK110 is in the midClinical Chemistry 62:1 (2016) 221
A
B
5
PTH P = 0.967
100
4
PTHrP P < 0.001
3
PTH H (pg/mL)
PTHrrP (pmol/L)
120
2
80 60 40 20 0
0
Women
D
10
PTHrP + PTH P = 0.040
8
PTHrP/PTH
PT THrP + PTH (pmol/L)
C
Women
Men
6
4
2
Men
2
PTHrP/PTH / P = 0.0016
1.5
1
0.5
0
0
Women
Women
Men
Men
Fig. 3. Distribution of concentrations of PTHrP (A), PTH (B), PTH + PTHrP (C), and PTHrP/PTH (D) in apparently healthy women (n = 120) and men (n = 109).
dle of the PTHrP sequence, and although its concentration does not represent the concentration of bioactive PTHrP, it serves as a marker of the rate of PTHrP biosynthesis. The method used immunocapture of PTHrP from plasma samples followed by tryptic digestion, while it was
B
7.0
7.0
6 6.0 0
6.0
5.0
5.0
pmol/L
pmol/L
A
4.0 3.0
bound to the antibody conjugated to magnetic beads, followed by direct analysis of the digests with triplequadrupole mass spectrometry. The on-bead digestion used in this method allowed for faster analysis without peptide elution, reduction, and alkylation before digestion. Considering that the enrichment takes place on the
4.0
PTHrP PTH
3.0
2.0
2.0
1.0
1.0
Total PTH + PTHrP
0.0
0.0 20–30
31–40
41–50
51–60
Years
61+
20–30
31–40
41–50
51–60
Years
Fig. 4. Median concentrations of PTHrP, PTH, and PTH + PTHrP in samples from apparently healthy women (n = 120) (A) and men (n = 109) (B) by age group (decade of life) (see Table 1 for the distribution of concentrations).
222
Clinical Chemistry 62:1 (2016)
Measurement of PTHrP by LC-MS/MS
Table 1. Concentrations of PTHrP, PTH plus PTHrP, and PTHrP/PTH in plasma samples of self-reported healthy adults by age group.a Age, years
n
Median age, years
PTHrP, pmol/L
PTH + PTHrP, pmol/L
PTHrP/PTH
20–30
52
25.5
1.6 (0.7–4.2)
4.3 (1.3–7.4)
0.60 (0.2–1.86)
31–40
37
34.0
1.3 (0.6–3.5)
4.3 (2.6–8.4)
0.47 (0.17–1.83)
41–50
16
45.0
1.4 (0.2–2.0)
4.5 (0.25–9)
0.49 (0.19–0.80)
51–60
14
56.5
1.9 (1–3.4)
5.3 (3.6–6.9)
0.53 (0.23–1.16)
4
63.0
2.0 (1–3.3)
6.5 (4.5–9.3)
0.59 (0.15–0.73)
6
34.5
2.4 (1.7–3.2)
5.3 (3.5–7.8)
0.79 (0.6–1.0)
20–30
35
27.0
1.2 (0.6–2.0)
3.8 (1.5–8.0)
0.49 (0.16–0.87)
31–40
42
33.0
1.2 (0.6–2.1)
4.3 (1.0–10.1)
0.35 (0.14–1.07)
41–50
24
43.5
1.2 (0.6–2.4)
4.1 (0.6–5.7)
0.45 (0.21–1.0)
51–60
14
53.0
1.3 (0.8–2.4)
4.3 (2.2–10.0)
0.39 (0.14–0.76)
Women
≥61 Breastfeeding women Men
a
Data are median (central 95% concentration) unless noted otherwise.
protein level, accurate quantification of PTHrP requires careful selection of the internal standard. Two stable isotope–labeled internal standards were evaluated, a winged labeled peptide (19 ) and recombinant [15N]PTHrP. The winged internal standard provided partial improvement in performance; quantitative performance substantially improved with the use of recombinant [15N]PTHrP as the internal standard. The YLTQETNK peptide used in this method is unique to PTHrP, has no homology with any other tryptic fragments of human proteins, and is a specific surrogate marker for quantification of PTHrP. Among ⬎8000 analyzed patient samples, we observed ⬍2% samples with a ratio of mass transitions outside 30% of the expected range among the samples containing ⬎2 pmol/L PTHrP. PTHrP in plasma samples is unstable and undergoes rapid degradation (11, 12 ). Of the evaluated types of blood collection tubes, PTHrP concentrations were highest in the tubes with PPACK inhibitor (analyzed immediately after defrosting). After the samples were stored refrigerated for 48 h, PTHrP degraded to concentrations similar to those observed in the other types of
collection tubes. The interindividual variation in the reduction of PTHrP concentrations in samples analyzed fresh and after storage could be related to the differences in concentrations of endogenous proteases causing PTHrP degradation. Similar observations have been reported in earlier publications (11, 12 ). In agreement with our data, the PTHrP degradation rate varied among the samples from different individuals. The data suggest that PTHrP degradation begins at the time of blood collection and plasma separation and imply the necessity to draw blood in tubes containing PPACK inhibitor (20, 21 ). In samples collected in tubes with PPACK inhibitor, PTHrP concentrations were above the LOQ of the method in 99.6% (Fig. 3), with higher concentrations observed in women than men. Concentrations were highest in breastfeeding women (Table 1), compared with the other groups of healthy adults, and in postmenopausal women. We did not observe an association in men between PTHrP concentrations and age. The gaussian distribution of the concentrations in samples of healthy individuals supports tight regulation of PTHrP. Considering that PTHrP plays a role
Table 2. Diagnostic utility of the LC-MS/MS test for PTHrP. Positive predictive value
Negative predictive value
Clinical sensitivity
Clinical specificity
Diagnostic efficiency
1.8
0.88
0.80
0.76
0.89
0.82
2.3 (upper normal for men)
0.94
0.73
0.59
0.96
0.79
3.3 (upper normal for women)
0.98
0.67
0.45
0.99
0.73
Cutoff concentration, pmol/L
Clinical Chemistry 62:1 (2016) 223
in development of mammary glands, lactation, pregnancy, and osteoporosis (5, 7 ), we expected that PTHrP concentrations would be higher in women than men, which is what our data showed. Higher concentrations of PTHrP were observed in individuals with lower body mass index (see online Supplemental Fig. 3) and postmenopausal women (see online Supplemental Fig. 2). On the basis of current knowledge, in health, Ca in blood is regulated by PTH, and PTHrP is often undetectable in blood of healthy individuals (8 ). Our data demonstrate that PTH and PTHrP are present in blood at comparable concentrations. Because both hormones bind to the same receptors, it is likely that they are both responsible for Ca regulation, and the total concentration of PTHrP ⫹ PTH (or possibly a weighted sum of PTHrP and PTH) could be more representative of the status of Ca regulation than the concentration of PTH alone. Our data (Table 1, Fig. 3, and online Supplemental Fig. 2) show that the concentrations of PTHrP and PTHrP ⫹ PTH are higher in postmenopausal women compared with men and premenopausal women. These observations could provide a partial explanation of the high prevalence of osteoporosis in postmenopausal women with low body mass index (22, 23 ). We evaluated the ability of the method to identify PTHrP as a cause of hypercalcemia by analyzing PTHrP in pathologic patient plasma samples. We hypothesized that in many of the patents with unexplained hypercalcemia that tested negative for PTHrP by RIA, PTHrP was not detected because of poor performance of the RIA. In a set of 88 patient samples of patients with hypercalcemia, 42% of samples had increased PTHrP, and the PTHrP/PTH ratio was above the reference interval in all samples of the set. In another set of plasma samples from patients with Ca within reference intervals and low PTH, PTHrP concentrations were above the established reference interval in 23% of patients. Considering that Ca is tightly regulated and its concentration in health is maintained within a narrow physiological range, the observed Ca concentrations in these 2 sets of samples could not be explained by PTH alone. One possibility for the hypercalcemia in some of the patients from the first cohort and normocalcemia in the second cohort could be related to an increase of PTHrP. The highest diagnostic efficiency was observed at a PTHrP cutoff of 1.8 pmol/L; the highest positive predictive value and clinical specificity were observed at the upper cutoff concentration for women (3.3 pmol/L) (Table 2), supporting the clinical value of the assay. In a review of causes of hypercalcemia, Jacobs and Bilezikian (8 ) stated that in many cases hypercalcemia was not associated with increased PTHrP. This observation could be related to the poor performance of available 224
Clinical Chemistry 62:1 (2016)
PTHrP assays, which underestimate PTHrP concentrations, resulting in a very low positivity rate in samples from patients clinically suspected of hypercalcemia. Because of the poor performance, previously available methods were not able to reliably measure PTHrP in healthy individuals (11–14 ), resulting in the assumption that PTHrP is not present in circulating blood in health. In this study, we identified a large number of patient samples with substantial disagreement in results between the commercial RIA and our LCMS/MS assay (Fig. 2). In routine use, among consecutive samples analyzed by RIA (n ⫽ 1376), the positivity rate was 1.9%; among consecutive samples analyzed by LC-MS/MS assay (n ⫽ 1705), the positivity rate was 26.6%. In the above sets of samples, when analyzed by RIA, concentrations were below the LOQ in 95.6% of the samples and in only 2.0% of samples by LC-MS/MS. These results suggest that in many patients with clinical symptoms of hypercalcemia, PTHrP was undetectable by RIA not because the concentrations were too low, but because of the poor performance of the RIA. Considering that samples submitted to reference laboratories for measurement of PTHrP are typically from patients in whom hypercalcemia is clinically suspected, the observed low positivity rate of the RIA (⬍2%) suggests that the RIA produces falsely low results in many samples. On the basis of our data, increased concentrations of PTHrP in samples of patients with suspected hypercalcemia are more common than was stated previously. Our observation on poor agreement between the methods for PTHrP is not unique; poor agreement between the commercial RIAs was described in other studies (11, 12 ). Because of the inadequate analytical sensitivity of the PTHrP RIA and poor stability of PTHrP, it has been assumed that PTHrP is not present in circulating blood in health (8 ) and that measurement of PTHrP is not diagnostically useful if HCM is not suspected. This LCMS/MS method allowed measurements of endogenous concentrations of PTHrP in healthy individuals, and our data showed that PTHrP is present in circulating blood in health. On the basis of our data, endogenous concentrations of PTHrP in blood are comparable to concentrations of PTH. Considering the difference in half-lives of PTH and PTHrP (approximately 4 min and a few hours, respectively) (24, 25 ), and that both PTH and PTHrP are involved in Ca regulation (6, 7 ), the total and relative concentrations of PTHrP and PTH could be important for the assessment of Ca regulation in patients with signs of osteoporosis, osteomalacia, and unexplained hypercalcemia. Data on the distribution of PTH ⫹ PTHrP and the PTHrP/PTH ratio in samples from healthy adults (Table 1 and Fig. 3) demonstrated relatively narrow distributions. Future studies will be required to assess associations
Measurement of PTHrP by LC-MS/MS
between PTHrP concentrations and various physiological and pathological conditions. Because increased secretion of PTHrP could be caused by tumors, this PTHrP method may potentially be used as a screening test for early detection of PTHrP-secreting malignancies and conditions associated with altered Ca regulation. In summary, we developed an LC-MS/MS method for the quantitative analysis of PTHrP in plasma samples. The immunoaffinity enrichment and on-bead digestion used in this method could be generally applicable to LC-MS methods for other protein biomarkers. Poor agreement was observed between this LC-MS/MS method and the Immunotech RIA, likely because of the poor analytical sensitivity of the RIA. We established reference intervals for PTHrP, the total concentration of PTHrP ⫹ PTH, and the ratio of the concentrations PTHrP/PTH in samples of self-reported healthy adults. Our data comparing the positivity rate between this LCMS/MS method and commercial RIAs indicate substantial underestimation of PTHrP concentrations by RIAs and underdiagnoses of PTHrP as a cause of hypercalcemia in many patients. Because both PTH and PTHrP have roles in Ca regulation, we believe that PTH ⫹ PTHrP and PTHrP/PTH could be useful markers for the assessment of Ca regulation. In women, concentrations were highest in breastfeeding women and in women ⬍30 and ⬎50 years old. We did not observe an association between PTHrP concentrations and age in men. Higher concentrations of PTHrP were observed in women and men with lower body mass index. These observations
may provide an explanation of the higher prevalence of osteoporosis in postmenopausal women with low body mass index. Finally, given the current state of PTHrP testing in clinical laboratories and lack of reference methodologies, there is a need for standardization among PTHrP methods.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest: Employment or Leadership: None declared. Consultant or Advisory Role: E.L. Frank, ARUP Laboratories, Inc. Stock Ownership: None declared. Honoraria: None declared. Research Funding: This work was supported by the ARUP® Institute for Clinical and Experimental Pathology. Expert Testimony: None declared. Patents: M.M. Kushnir, provisional patent application no. 62/179,488. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript. Acknowledgments: We thank the ARUP Institute for Clinical and Experimental Pathology for supporting this project and Kimberly Kalp for technical assistance.
References 1. Moseley JM, Kubota M, Diefenbach-Jagger H, Wettenhall RE, Kemp BE, Suva LJ, et al. Parathyroid hormonerelated protein purified from a human lung cancer cell line. Proc Natl Acad Sci U S A 1987;84:5048 –52. 2. Stewart AF, Wu T, Goumas D, Burtis WJ, Broadus AE. N-terminal amino acid sequence of two novel tumorderived adenylate cyclase-stimulating proteins: identification of parathyroid hormone-like and parathyroid hormone-unlike domains. Biochem Biophys Res Commun 1987;146:672– 8. 3. Strewler GJ, Stern PH, Jacobs JW, Eveloff J, Klein RF, Leung SC, et al. Parathyroid hormonelike protein from human renal carcinoma cells: structural and functional homology with parathyroid hormone. J Clin Invest 1987;80:1803–7. 4. Broadus AE, Mangin M, Ikeda K, Insogna KL, Weir EC, Burtis WJ, Stewart AF. Humoral hypercalcemia of cancer: identification of a novel parathyroid hormonelike peptide. N Engl J Med 1988;319:556 – 63. 5. Wysolmerski JJ, Philbrick WM, Dunbar ME, Lanske B, Kronenberg H, Broadus AE. Rescue of the parathyroid hormone-related protein knockout mouse demonstrates that parathyroid hormone-related protein is essential for mammary gland development. Development 1998;125:1285–94. 6. Suva LJ, Winslow GA, Wettenhall RE, Hammonds RG, Moseley JM, Diefenbach-Jagger H, et al. A parathyroid hormone-related protein implicated in malignant
hypercalcemia: cloning and expression. Science 1987; 237:893– 6. 7. Mannstadt M, Ju¨ppner H, Gardella TJ. Receptors for PTH and PTHrP: their biological importance and functional properties. Am J Physiol 1999;277:665–75. 8. TP Jacobs, JP Bilezikian. Clinical review: Rare causes of hypercalcemia. J Clin Endocrinol Metab 2005;90: 6316 –22. 9. Mundy GR. Mechanisms of bone metastasis. Cancer 1997;80:1546 –56. 10. Ikeda K, Ohno H, Hane M, Yokoi H, Okada M, Honma T, et al. Development of a sensitive two-site immunoradiometric assay for parathyroid hormone-related peptide: evidence for elevated levels in plasma from patients with adult T-cell leukemia/lymphoma and B-cell lymphoma. J Clin Endocrinol Metab 1994;79: 1322–7. 11. Pandian MR, Morgan CH, Carlton E, Segre GV. Modified immunoradiometric assay of parathyroid hormonerelated protein: clinical application in the differential diagnosis of hypercalcemia. Clin Chem 1992;38: 282– 8. 12. de Miguel F, Motellue JL, Hurtado J, Jime´nez FJ, Esbrit P. Comparison of two immunoradiometric assays for parathyroid hormone-related protein in the evaluation of cancer patients with and without hypercalcemia. Clin Chim Acta 1998;277:171– 80. 13. Fraser WD, Robinson J, Lawton R, Durham B, Gal-
14.
15.
16.
17.
18.
19.
lacher SJ, Boyle IT, et al. Clinical and laboratory studies of a new immunoradiometric assay of parathyroid hormone-related protein. Clin Chem 1993;39: 414 –9. Wu TJ, Taylor RL, Kao PC. Parathyroid-hormone-related peptide immunochemiluminometric assay developed with polyclonal antisera produced from a single animal. Ann Clin Lab Sci 1997;27:384 –90. Lu CM, Burton WD, Fitzgerald RL, Deftos LJ, Buchholz BA, Vogel JS, Herold DA. Mass spectrometric immunoassay for parathyroid hormone-related protein. Anal Chem 2002;74:5507–12. Washam CL, Byrum SD, Leitzel K, Ali SM, Tackett AJ, Gaddy D, et al. Identification of PTHrP(12– 48) as a plasma biomarker associated with breast cancer bone metastasis. Cancer Epidemiol Biomarkers Prev 2013; 22:972– 83. Kushnir MM, Rockwood AL, Nelson GJ, Yue B, Urry FM. Assessing analytical specificity in quantitative analysis using tandem mass spectrometry. Clin Biochem 2005; 38:319 –27. Matuszewski BK, Constanzer ML, Chavez-Eng CM. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal Chem 2003;75:3019 –30. Kushnir MM, Rockwood AL, Roberts WL, Hoofnagle AN, Meikle AW. Accurate measurement of thyroglobulin in serum in presence of anti-thyroglobulin autoantibod-
Clinical Chemistry 62:1 (2016) 225
ies. Clin Chem 2013;59:982–90. 20. Hutchesson AC, Hughes SV, Bowden SJ, Ratcliffe WA. In vitro stability of endogenous parathyroid hormonerelated protein in blood and plasma. Ann Clin Biochem 1994;31:35–9. 21. Kushnir MM. Are samples in your freezer still good for biomarker discovery? Am J Clin Pathol 2013;140: 287– 8.
226
Clinical Chemistry 62:1 (2016)
22. Mpalaris V, Anagnostis P, Goulis DG, Iakovou I. Complex association between body weight and fracture risk in postmenopausal women. Obes Rev 2015;16:225– 33. 23. Unni S, Yao Y, Milne N, Gunning K, Curtis JR, LaFleur J. An evaluation of clinical risk factors for estimating fracture risk in postmenopausal osteoporosis using an electronic medical record database. Osteoporos Int
2015;26:581–7. 24. Sokoll LJ, Drew H, Udelsman R. Intraoperative parathyroid hormone analysis: a study of 200 consecutive cases. Clin Chem 2000;46:1662– 8. 25. Tov AB, Mandel D, Weissman Y, Dollberg S, Taxir T, Lubetzky R. Changes in serum parathyroid hormonerelated protein in breastfed preterm infants. Breastfeed Med 2012;7:50 –3.
