Increased Serum and Urinary Levels of a Parathyroid Hormone-related Protein COOH Terminus in Non-Small Cell Lung Cancer Patients

PTHrP² was originally isolated from cancer cell lines and a patient with the syndrome of HHM (1, 2). PTHrP is expressed in many human cancers. An antibody against the PTHrP NH₂-terminal portion immunohistochemically stained malignant cells in skin tumors (3); RIA detected PTHrP (1–34) in human prostate cancer cell lines (4); HPLC detected PTHrP in a human T-cell lymphotropic virus type 1-infected cell line (5); Northern and Southern blot analyses found PTHrP gene expression in a human renal carcinoma (6); immunohistochemical staining using an antibody against NH₂-terminal region localized PTHrP in a human breast cancer (7); and immunohistochemical staining using an anti-PTHrP monoclonal antibody found PTHrP in gastric cancers (8). Moreover, NH₂-terminal RIA detected PTHrP in major lung cancer cell types (9). The first 13 amino acids of PTHrP share 70% sequence homology with PTH (2, 10). PTHrP has a PTH-like bioactivity, which is mediated by binding PTH receptors (11), although there may be receptor subtypes that are specific for PTHrP (12). It has been suggested that PTHrP has its own functions that differ from those of PTH, and it exerts biological effects other than serum calcium level modulation; these are transforming growth factor-like properties (13), osteoclast inhibition by the COOH-terminal peptide (14), and placental calcium transport by the mid-region PTHrP (15).

The human PTHrP gene has seven exons. Three mature polypeptides with different COOH-terminal regions, PTHrP (1–139), PTHrP (1–173), and PTHrP (1–141) are translated from three different mRNAs through alternative splicing (16). PTHrP gene expression in vitro has been shown to be regulated by a variety of substances including phorbol esters, epidermal growth factor (17), 1,25-dihydroxyvitamin D₃ (18), glucocorticoids (19), 17β-estradiol (20), calcitonin, and chromogranin A (21).

A majority of squamous cell tumors synthesize PTHrP, even in the patients with normal serum calcium levels (22). PTHrP mRNA and peptide expression are analyzed by in situ hybridization and immunohistochemistry in squamous carcinomas and adenocarcinomas from patients who were either normocalcemic or hypercalcemic (22, 23). However, RIA hardly detected any serum PTHrP in these patients (24, 25). Antibodies that had been used in PTHrP RIA were raised against mature PTHrP; this may be one of the reasons for the low detectability of PTHrP in lung cancer patients. Posttranslational modifications of PTHrP probably changed the sensitivity of COOH-terminal RIA. The posttranslational processing steps might include O-glycosylation (26) and COOH-terminal amidation (27). Multiple COOH-terminal PTHrP secretory forms are generated by isoform-specific protein processing, and multiple NH₂-terminal PTHrP secretory forms are generated by cell-specific protein processing (28). PTHrP peptide is broken up by proteases to produce small fragments in the blood; of these, the COOH-terminal fragment (C-PTHrP) is stable and measurable in the urine (29, 30).

In the present study, we studied concentrations of circulating and urinary C-PTHrP in patients with primary lung cancer with normal serum calcium level to determine the significance of C-PTHrP measurement for lung cancer diagnosis. We used PCR to evaluate PTHrP mRNA expression and its alternative splicing types in 17 lung cancer tissues and 16 lung cancer cell lines. Four pairs of exon-specific oligonucleotide primers were used to evaluate the alternative RNA splicing. We measured C-PTHrP peptide levels in cell pellets and cell culture media from 16 human lung cancer cell lines. We compared the COOH-terminal RIA results with the PCR results in lung cancer cell lines as well as cancer tissues.

We studied 28 of 34 patients with primary lung cancer (16 adenocarcinoma, 6 squamous cell carcinoma, and 6 small cell carcinoma) with normal serum calcium levels who were treated in our department from June 1995 to May 1996 and 10 normal subjects. These subjects were well informed of the purpose of the present study and fully agreed to the study. No cancer patient had received any therapy before the measurements. Tables 1 and 2 show characteristics of the patients. Three cancer patients and one normal subject were women; the others were men and differences of age, serum calcium levels, and serum creatinine levels between lung cancer patients and normal subjects were not statistically significant. Two cancer patients had bone metastases, although their serum calcium levels were within normal limits.

Blood was drawn into heparin-treated tubes containing protease inhibitors, EDTA 2Na, and aprotinin, at 8:00 a.m. Samples were immediately centrifuged at 3000 rpm for 5 min to separate out the plasma, which was then stored at −20°C until measurements were taken. Urine was collected between 6:00 and 8:00 a.m., and samples were immediately centrifuged at 3000 rpm for 15 min. The aqueous phase was stored at −20°C until measurements were taken. RIAs were done as follows. A COOH-terminal region-specific RIA for human PTHrP (109–141) was done using sheep antiserum immunized with synthetic human PTHrP (109–141) and ¹²⁵I-labeled synthetic PTHrP peptide (108–141) (C-PTHrP kit Daiichi; Daiichi Radioisotope Laboratories, Ltd., Tokyo, Japan). An NH₂-terminal region-specific RIA for human PTHrP (1–40) was done using goat antiserum immunized with a synthetic human PTHrP (1–40) and ¹²⁵I-labeled synthetic PTHrP peptide (1–34) (N-PTHrP kit; Incstar Corp., Stillwater, MN). Intact PTHrP was measured by the two-site immunoradiometric assay targeting PTHrP (1–87) using a solid-phase monoclonal antibody against PTHrP peptide (1–34) and liquid-phase polyclonal antibody against PTHrP peptide (50–83) (PTHrP IRMA kit; Mitsubishi Yuka Corp., Tokyo, Japan).

Urinary C-PTHrP immunoreactivity was corrected by urinary creatinine concentration and expressed as urinary C-PTHrP/urine creatinine ratio (10⁻¹ × pmol/mg creatinine).

Data were analyzed using Mann-Whitney’s U test and χ² test. Differences or correlations were considered significant at P < 0.05.

Tumor specimens had been obtained during surgery or autopsy from 17 primary lung cancer patients (median age, 64.7 years; range, 49–75 years; Table 1). Immediately after removal, samples were chilled on crushed ice, then frozen and stored in a freezer at −80°C until use. All patients had normal calcium levels. Serum and urine C-PTHrP levels were measured in 11 patients by COOH-terminal RIA.

We studied 16 human cancer cell lines, NCI-H69, NCI-H1963, NCI-H526, NCI-H433, NCI-H209, PC6, NCI-H146, HSRRB-PC3, PC14, HSRRB-A549, HSRRB-ABC1, HSRRB-EBC1, NCI-H838, NCI-H650, PC13, and NCI-H727 for PTHrP production (Table 3). PTHrP levels were measured in the cell extracts and culture media using COOH-terminal RIA. Cells were incubated in RPMI 1640 (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) supplemented with l-glutamine, 5000 units/ml penicillin G, 5000 μg/ml streptomycin, and 10% FBS (RPMI+). Cells were cultured for 5 days in RPMI+ in 5% CO₂, 37°C. Then cells were rinsed with serum-free RPMI medium (RPMI−) twice. Cells (1 × 10⁷) were incubated with RPMI− for 48 h in 5% CO₂ at 37°C. Culture media with cells were centrifuged at 200 × g for 5 min. Supernatant was collected, and 6 μl of acetic acid per 1 ml medium, 10% v/v protease inhibitors mixture (15 mg/ml EDTA, 0.69 mg/ml pepstatin A, and 0.35 mg/ml phenylmethylsulfonyl fluoride) were added. The culture media were stored at −20°C until analyses.

Log growth phase cells (1 ×10⁷) were washed twice in PBS. Cells were resuspended in RPMI−, then centrifuged at 200 × g for 5 min. Cells were resuspended in PBS with 0.1 m acetic acid and 10% v/v protease inhibitors mixture. Cells were homogenated by a tissue tearor (Polytron; Kinematica GmbH, Littau, Switzerland) on ice for 60 s. Cell suspensions were centrifuged for 20 min at 32,000 × g. The supernatant was collected and stored at −20°C until analyses. Cell extracts and condition media were passed through octadecylsilyl-silica cartridges (C18 Anprep; Amersham Life Science, Buckinghamshire, England), then washed with 0.1% trifluoroacetic acid and eluted with 80% acetonitrile.

Cell pellets and tissue were resuspended in guanidine-thiocyanate buffer and homogenized by the tissue tearor. Total RNA was extracted by the guanidine-thiocyanate and CsCl gradient ultracentrifugation method (31). RNA samples were stored at −80°C until analyses.

First-strand cDNA was synthesized from 5 μg of the total RNA using a Super Script Preamplification kit (Life Technologies, Inc., Gaithersburg, MD) and random hexamers. Five μg of total RNA, 250 ng of random hexamers, and 5 μl of diethyl pyrocarbonate-treated water were mixed in a plastic tube. Samples were incubated for 10 min at 70°C and chilled on ice for at least 1 min. Then, 2 μl of 10× PCR buffer solution, 2 μl of 25 mm MgCl₂ solution, 1 μl of 10 mm deoxynucleotide triphosphate mixture and 2 μl of 0.1 m DTT solution were added. After incubation at 25°C for 5 min, 200 units of SuperScript 2 reverse transcriptase were added and incubated at 25°C for 10 min, followed by incubation at 42°C for 50 min. The reaction was incubated at 70°C for 15 min and chilled on ice.

Two μl of cDNA solution were added to 48 μl of PCR buffer containing 100 pmol of upstream and downstream primers. These primers were designed to amplify cDNA from each alternative spliced mRNA. The upstream primer used for the amplification was specific for exons 3 and 4, and the downstream primers were specific for exons 4, 5, 6, and 7. Oligonucleotide sequences that were used in the present study are listed in Table 4. Oligonucleotide primers used for amplification of human β-actin were sense 5′-ATCATGTTTGAGACCTTCAA-3′, corresponded to β-actin cDNA 1854 to 1874, and antisense 5′-CATCTCTTGCTCGAAGTCCA-3′, which represented the antisense strand of the β-actin cDNA sequence 2152 to 2172 (32). Five units of Taq DNA polymerase, 5 μl of 10× PCR buffer solution, 3 μl of 25 mm MgCl₂, 1 μl of 10 mm deoxynucleotide triphosphate mixture (PCR Amplification kit; Takara, Tokyo, Japan), and 36.5 μl of distilled water were added. Amplification was performed in a thermal cycler (model PJ2000; Perkin-Elmer Corp., Branchburg, NJ) with an initial denaturation at 96°C for 3 min and an annealing temperature at 55°C for 2 min. This was followed by 25 cycles with a denaturing step at 95°C for 15 s, an annealing step at 55°C for 30 s, and a polymerization step at 72°C for 45 s.

The oligonucleotide sequence that was used in the detection of the PCR product was 5′-TCCAGAGTCTAACCAGGCAGAGCGAGTTCG-3′, representing the antisense strand of PTHrP cDNA sequence 424 to 453 (exon 4) (33). The oligonucleotide was labeled on the 5′ end using T4 polynucleotide kinase (Life Technologies, Inc.). Oligonucleotide was added to make 0.92 μg/μl in 2 μl of 10× kinase buffer solution, 5 μl of [γ-³²P]ATP, and 10.4 μl of distilled water. Eight units of T4 polynucleotide kinase were added, and samples were then incubated at 37°C for 45 min. The reaction was terminated by incubating at 68°C for 10 min. PCR products were capillary blotted onto Hybond N+ membrane (Amersham Life Science) and incubated at 80°C for 2 h. The membrane was prehybridized for 2 h at 65°C in 5× SSC containing 1% SDS and 1× Denhardt’s solution. The membrane was hybridized for 24 h at 65°C in the same solution containing radiolabeled oligonucleotide at 1.0 × 10⁷ cpm/filter and denatured salmon sperm. After washing at 65°C, the filters were exposed to X-OMAT film (Kodak, Rochester, NY) for 48 h at −80°C.

RIAs were done by the same method using ¹²⁵I-labeled peptides and an antibody targeting COOH-terminal PTHrP (109–141) as described in the previous section.

The average serum C-PTHrP level was 38.95 ± 19.41 pmol/l (mean ± SD) in 28 lung cancer patients, whereas that in 10 normal subjects was 26.53 ± 9.43; the difference in C-PTHrP levels was statistically significant (P = 0.0065). The average serum C-PTHrP level was 41.55 ± 21.08 in 22 non-small cell carcinoma patients and 29.45 ± 5.55 in 6 small cell carcinoma patients. The difference between non-small cell carcinoma patients and normal subjects was again statistically significant (P = 0.0041). However, the difference in serum levels of C-PTHrP between small cell carcinoma patients and normal subjects was not statistically significant (P = 0.2119; Fig. 1).

Average urinary C-PTHrP:urine creatinine ratio was 7.56 ± 5.17 × 10⁻¹ pmol/mg creatinine (mean ± SD) in 28 lung cancer patients, whereas it was 4.91 ± 1.77 in 10 normal subjects, a statistically significant difference in ratios (P = 0.0287). Average urinary C-PTHrP:urine creatinine ratio was 8.14 ± 5.69 in 22 non-small cell carcinoma patients and 5.41 ± 1.23 in 6 small cell carcinoma patients. The difference between non-small cell carcinoma patients and normal subjects was again statistically significant (P = 0.0253). However, the difference in ratios between small cell carcinoma patients and normal subjects was not statistically significant (P = 0.2328; Fig. 2).

Urinary C-PTHrP:urine creatinine ratio positively correlated with serum C-PTHrP level (P = 0.0002; Fig. 3). Urinary C-PTHrP:urine creatinine ratio was higher than average level + 2SD of normal subjects, which was 8.45 × 10⁻¹ pmol/mg creatinine in 7 cases of 28 lung cancer patients (25%) and in 7 cases of 22 non-small cell lung cancer patients (32%). Urinary C-PTHrP:urine creatinine ratio did not correlate with Tumor-Node-Metastasis factors. However, more cases with higher urinary C-PTHrP:urine creatinine ratio were seen in N₃ cases than N₀, N₁, and N₂ cases (χ² test; P = 0.0255).

Serum N-PTHrP were detected in 4 cases of 20 lung cancer patients who were measured, 1 of which was small cell lung cancer patient and 3 were non-small cell lung cancer patients. Serum N-PTHrP level in case 17 was 10.94 pmol/l, which was the highest concentration in the lung cancer patients. Serum intact-PTHrP level was 2.3 pmol/l in case 17, although intact-PTHrP was not detectable in other patients.

PTHrP cDNA from nucleotide number 10 to 525, which included exons 3 and 4 of the PTHrP gene, was amplified in all non-small cell lung cancer tissues and non-small cell cancer cell lines, and in two small lung cancer tissues, cases 2 and 4, and one small cell cancer cell line, HCI-H146 (Fig. 4). PTHrP m-RNA was not amplified in five normal lung tissues from non-small cell cancer patients (data not shown). Three different primer sets were then used which selectively amplified PTHrP cDNA from PTHrP (1–139), PTHrP (1–173), and PTHrP (1–141).

When a primer set, which was selected to amplify PTHrP from nucleotide number 10 to 884 for PTHrP (1–139), was used, 875-bp signals were seen in 11 cancer tissues and 10 cancer cell lines that had detectable PTHrP mRNA. A primer set for nucleotide number from 343 to 738 for PTHrP (1–173) detected a 396-bp signal in 10 cancer tissues and 9 cancer cell lines, and a primer set for nucleotide number from 343 to 636 for PTHrP (1–141) detected a 294-bp signal in 13 cancer tissues and 10 cancer cell lines (Fig. 5). A primer set for β-actin detected a 319-bp signal in all samples.

Tables 1 and 3 show PTHrP concentrations in the serum and urine from lung cancer patients and the cell extracts, and culture media of the lung cancer cell lines by COOH-terminal PTHrP RIA. Serum C-PTHrP was 112.0 pmol/l, and urinary C-PTHrP:urine creatinine ratio was 12.21 × 10⁻¹ pmol/mg creatinine in case 13, which was the highest concentration in lung cancer patients. C-PTHrP was detected in the media in NCI-H146, HSRRB-PC3, HSRRB-EBC1, NCI-H650, and NCI-H727. The highest concentration was 19.90 pmol/l in NCI-H727. C-PTHrP was detected in the cell extract from NCI-H727.

In the present study, we showed that lung cancer patients had higher serum and urinary C-PTHrP levels than normal subjects, and this difference was more marked when only non-small cell lung cancer patients were considered. Also, urinary C-PTHrP levels correlated with blood C-PTHrP levels. However, the difference in serum and urinary C-PTHrP levels between small cell lung cancer patients and normal subjects was not significant. These results suggested that non-small cell lung cancers produced PTHrP more frequently than small cell lung cancer.

Differences among levels of serum N-PTHrP, serum C-PTHrP, and urinary C-PTHrP were seen in seven lung cancer cases. The probable reasons for this are difference of sensitivity between N-PTHrP assay and C-PTHrP assay and difference of protein processing between N-PTHrP and C-PTHrP. The urinary C-PTHrP levels might be influenced by small changes in renal function, even though we had selected the patients who had normal renal function.

RT-PCR detected PTHrP mRNA expression in 21 of 21 non-small cell lung cancer samples, 3 of 12 small cell lung cancer samples, and 0 of 5 normal lung tissues. These patients did not have elevated serum calcium levels. These data possibly explain why serum and urinary C-PTHrP levels were high in non-small cell cancer patients but not in small cell lung cancer patients. C-PTHrP RIAs detected more serum C-PTHrP and urinary C-PTHrP:urine creatinine ratio than average + 2SD of normal subjects in 23 and 32%, respectively, in non-small cell lung cancer patients, although all samples from non-small cell lung cancer patients had detectable PTHrP mRNA. These results suggested that the present assay was not sufficiently sensitive for use in cancer detection.

Four of nine non-small cell lung cancer cell lines and one of seven small cell lung cancer cell lines had detectable C-PTHrP in the cell extracts and culture medium, although all non-small cell lung cancer cell lines had detectable PTHrP mRNA. Possible reasons for these results are that C-PTHrP was underestimated because sensitivity of the RIA was not good enough, or the translation rate of PTHrP was different from cancer to cancer.

The human PTHrP gene encodes three different COOH-terminal PTHrPs, PTHrP (1–139), PTHrP (1–173), and PTHrP (1–141). Each PTHrP has different protease cleavage sites. The monobasic arginine at position 37 is a posttranslational processing site (34), and the 88–106 region is rich in multibasic endoproteolytic sites (35). Three different C-PTHrPs are made by RNA alternative splicing. Our study attempted to clarify the dominant C-PTHrP type in lung cancer, because such information may help to increase the sensitivity of detection of C-PTHrP in the serum and urine of lung cancer patients. In the present study, all three types of mRNA by the alternative splicing were found in 18 of 24 (75%) of these samples. Interestingly, PTHrP (1–139) mRNA was found in 21 of 24 cancer tissues and cancer cell lines that had detectable PTHrP mRNA, 19 of 24 lung cancer samples that had detectable PTHrP (1–173) mRNA, and 23 of 24 lung cancer samples with detectable PTHrP (1–141) mRNA. These results suggest that expression of all three PTHrP mRNAs is common in lung cancer tissues, whereas, PTHrP (1–139), PTHrP (1–173), and PTHrP (1–141) expression was different in each cancer tissue.

RT-PCR analyses of PTHrP mRNA in small cell lung cancer revealed that in case 1 and case 4, H1963 and H433, respectively, had one to three types of indistinct PTHrP mRNA expression by alternative splicing. These results probably indicate that many pathological types of lung cancer including small cell lung cancer had PTHrP mRNA expression.

Lung cancer patients with mediastinal lymph node metastases had higher C-PTHrP levels in their urine in the present study. PTHrP has been reported as forming an autocrine growth stimulation loop like TGF-α (36). PTHrP (1–34) has been reported to be an autocrine growth factor in a variety of tumors; the addition of exogenous PTHrP causes the proliferation of prostate cancer cells (4) and lung alveolar carcinoma cells (37). In these cell lines, antibodies against PTHrP (1–34) inhibit growth stimulation by PTHrP (1–34). Strong PTHrP gene expression is associated with bone metastases in breast cancer (38). Coamplification of the PTHrP gene and K-ras-2in BEN cells has been reported (39). PTHrP gene transcription is accelerated in Ha-ras/v-src-transformed cells (40). Therefore, PTHrP production in lung cancer cells has a potential role in cancer cell growth and metastases. PTHrP mRNA was detected in cancer tissues, and even from lung cancer patients with normal serum calcium levels. Yano et al. and Bandou et al. (published in Japanese) reported that more than 230.3 ± 314.6 pmol/l in the serum and 60.0 ± 17.9 × 10⁻¹ pmol/mg creatinine in the urine was necessary to cause HHM using C-PTHrP RIA; in our patients, its production might be insufficient to cause HHM.

We found that non-small cell lung cancers had detectable PTHrP mRNA, and serum and urinary C-PTHrP levels in non-small cell lung cancer patients were significantly higher than those in normal subjects. From the present study, we concluded that non-small cell lung cancer frequently produces PTHrP. The significance of C-PTHrP measurement in lung cancer patients for cancer detection is not clear, partly due to insufficient sensitivity of the present RIA. PTHrP probably plays a role similar to a cell proliferation factor in lung cancer, especially in non-small cell lung cancer, at levels that do not cause HHM.

We thank Herbert Oie and Bruce E. Johnson of Navy-Medical Oncology Branch, National Cancer Institute, NIH, and Japan Health Sciences Foundation for kind distribution of the cancer cell lines; Drs. Sunao Yachiku and Mutsuo Ishikawa of Asahikawa Medical College for helpful advice; and Simon N. Bayley of Asahikawa Medical College for the English revision.