Fax: (858) 822-5009
Parathyroid hormone-related protein varies with sex and androgen status in nonsmall cell lung cancer
Version of Record online: 3 AUG 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 6, pages 1313–1320, 15 September 2007
How to Cite
Montgrain, P. R., Quintana, R., Rascon, Y., Burton, D. W., Deftos, L. J., Casillas, A. and Hastings, R. H. (2007), Parathyroid hormone-related protein varies with sex and androgen status in nonsmall cell lung cancer. Cancer, 110: 1313–1320. doi: 10.1002/cncr.22922
- Issue online: 31 AUG 2007
- Version of Record online: 3 AUG 2007
- Manuscript Accepted: 17 MAY 2007
- Manuscript Revised: 7 MAY 2007
- Manuscript Received: 19 MAR 2007
- Flight Attendants Medical Research Institute (R.H.H. and L.J.D.), by a Department of Veterans Affairs Merit Review Award (R.H.H.). Grant Number: CA23100
- National Cancer Institute (L.J.D.)
- University of California Tobacco-Related Disease Research Program (R.H.H.)
- biologic tumor marker;
- growth substances;
- gonadal hormones;
- parathyroid hormone receptor type 1;
- steroid receptors
In nonsmall cell lung cancer, tumor parathyroid hormone-related protein (PTHrP) expression predicts longer survival in women but not in men. To explain the sex-dependent survival effect, the authors proposed that hormonal influences decrease PTHrP in men versus women, that PTHrP inhibits tumor growth, and that the effect is greater in women than in men. The objectives of this study were to compare lung carcinoma PTHrP expression and carcinoma growth in male and female mice and to determine whether gonadal steroids regulate PTHrP in lung cancer cells.
Tumor PTHrP content was measured by immunoassay, and tumor burden was assessed with multiple measures in BEN squamous cell orthotopic lung carcinomas in athymic mice. In addition, lung adenocarcinoma PTHrP messenger RNA (mRNA) values determined by microarray analyses were compared between men and women. Cultured lung cancer cells were assayed for PTHrP after treatment with estradiol or R1881, a synthetic androgen.
Lung carcinomas contained approximately 3 times more PTHrP in female mice than in male mice. Similarly, levels of PTHrP mRNA were significantly greater in adenocarcinomas from patients who were women than from patients who were men. Male mice had greater tumor burden than female mice. Androgen treatment reduced PTHrP in 3 lung cancer lines. Estradiol had no effect. Testosterone treatment also reduced lung carcinoma PTHrP in female mice.
Lung carcinomas in females expressed more PTHrP than in males possibly because of negative regulation by androgens in males. Female mice with higher tumor PTHrP content had significantly less tumor burden than male mice, supporting the hypothesis that PTHrP inhibits tumor growth. Cancer 2007. © 2007 American Cancer Society.
Lung cancer causes more deaths in men and women worldwide each year than any other type of cancer. In the United States, roughly 90,300 deaths in men and 72,300 deaths in women are expected from the disease in 2006.1 Although smoking is the most common cause in both sexes, clinicopathologic characteristics and patient survival vary between women and men in several categories. For example, adenocarcinomas are more common in women, and squamous carcinomas are more common in men.2 Some studies suggest that women who are smokers have an increased susceptibility to developing lung cancer compared with men who are smokers.3 Conversely, female sex is a favorable prognostic factor after diagnosis; women demonstrate significantly longer survival than men at all stages. Sex variation in pathogenesis may result from behavioral or environmental differences, genetic effects, or hormonal influences. Growth factors for lung cancer potentially may vary between men and women. For example, levels of gastrin-releasing peptide reportedly vary in lung cancer, depending on sex.4 Parathyroid hormone-related protein (PTHrP) also may fall into this category.
PTHrP is a paraneoplastic protein that takes its name from similarities between its amino-terminal portion, PTHrP 1–34, and the corresponding portion of parathyroid hormone. The 2 molecules bind the same receptor, PTH1R, with similar affinities.5 PTHrP was discovered because of its capacity to cause hypercalcemia of malignancy,6 but it also can alter carcinoma growth and have an impact on survival.7–10 For example, PTHrP 1–34 decreases the growth of prostate carcinoma xenografts, and high levels of tumor PTHrP immunoreactivity connote prolonged survival in some types of human cancer.9, 11, 12 Because roughly 66% of all nonsmall cell lung carcinomas express PTHrP,13–15 its effects on disease progression or survival may have a significant impact in lung cancer.
The results from our studies have suggested that PTHrP slows lung carcinoma growth and may be a positive prognostic indicator. For example, neutralizing antibodies to PTHrP accelerate the growth of PTHrP-expressing orthotopic lung carcinomas in athymic mice, suggesting that the protein inhibits cancer cell proliferation.8 Recently, we showed that lung carcinoma PTHrP immunoreactivity was associated with longer survival in women with nonsmall cell lung cancer, but not in men.15 Further research is indicated to establish the etiology for the sex dependence and the mechanism for the survival effect.
Male and female sex hormones regulate PTHrP expression in a variety of tissues. Estrogen stimulates PTHrP production in cancers of the breast, kidney, pituitary, and uterus,16–19 whereas testosterone is inhibitory in cancer of the testis.20 Lung carcinoma PTHrP expression may vary between men and women if sex hormones exert differential effects. Differences in expression between tumors in men and women, in turn, may lead to a sex-dependent survival effect if PTHrP inhibits lung carcinoma growth, as our other work suggests. Therefore, the objectives of this research project were to evaluate lung carcinoma PTHrP expression as a function of sex, to compare lung carcinoma growth in male and female mice, and to determine whether estrogens and androgens regulate PTHrP in lung cancer cells.
MATERIALS AND METHODS
Human BEN lung squamous carcinoma cells were obtained from Dr. John Martin (Brisbane, Australia). The human A549 lung adenocarcinoma and H460 lung large cell carcinoma cell lines were purchased from the American Type Culture Collection (Manassas, Va). Cells were cultured in RPMI-1640 medium with 5% fetal bovine serum at 37°C in an atmosphere of 5% carbon dioxide and 95% air. Stable DsRed-expressing clones of BEN cells were generated by transfection with a retroviral expression plasmid. The pLNCX-DsRed2 plasmid (Clontech, Mountain View, Calif) was complexed with LipofectAMINE 2000 (Invitrogen, Carlsbad, Calif) in serum-free Dulbecco minimal essential medium. The DNA:liposome complexes were added to cells in a minimal volume and allowed to incubate from 4 to 6 hours. The DNA:liposome complexes were then removed and replaced with RPMI containing 10% fetal calf serum. Positive colonies were selected by fluorescence visualization 7 days later by pipette tip picking.
Orthotopic Lung Carcinoma Studies
For these studies, we used 5- to 6-week old male and female athymic nude mice (Harlan, Indianapolis, Ind) with the approval of the Veterans Affairs San Diego Healthcare System Institutional Animal Care and Use Committee. Standards in the National Institutes of Health Guide for Animal Use were followed.
DsRed-expressing BEN cells were harvested from subconfluent cultures by brief trypsinization, washed, and suspended at a density of 6 × 104 cells/μL in 250 μg/mL Matrigel in iced, serum-free media. Viability was >90% by Trypan blue staining before tumor implantation. Mice were anesthetized, the skin was incised, and the left rib cage was exposed with gentle blunt dissection. Then, 3 × 106 cells in 50 μL of media plus Matrigel were injected through the chest wall into the left lung.21 The incision was stapled closed. Body weights were obtained weekly for the first month and twice weekly thereafter. Blood was drawn from a saphenous vein every 2 weeks.
After 8 weeks, mice were anesthetized with an overdose of pentobarbital. Blood was drawn in heparinized tubes by a retro-orbital approach, placed on ice, and separated later into serum. The lungs were perfused with normal saline to clear residual blood then removed from the chest. The DsRed-expressing tumors were visualized in a fluorescent light box (540 nm/40 nm band-pass excitation filter, 590 nm broad-pass emission filter; Lightools Research, Encinitas, Calif). Fluorescent ventral and dorsal views of the lungs were photographed with a Spot RT 220-3 digital camera (Diagnostic Instruments, Sterling Heights, Mich) at constant magnification and exposure time.
Next, tumors were dissected from surrounding lung, guided by their fluorescence, and weighed. Multiple tumors were combined. The tumor collection and contralateral tumor-free lung were homogenized in tissue lysis buffer with protease inhibitors (50 mM Tris-HCl, 0.25 M NaCl, 0.1% IGEPAL CA-630, 1% Triton X-100, 5 mM ethylenediamine tetracetic acid, and 5 mM NaF with 1 mM phenyl methyl sulfonyl fluoride, 0.2 U/mL aprotinin, 1 μM leupeptin, and 1 μM pepstatin) in a 4:1 ratio of buffer volume:tissue mass. Clear supernatants were obtained from the homogenates by centrifugation at × 16,000 g for 1 hour. Serum and supernatants were frozen at −70°C for subsequent assay of PTHrP, as described previously,22, 23 and for tumor burden assessment, as documented below. In addition, serum calcium was measured using a kit (Sigma Diagnostics, St. Louis, Mo).
An additional experiment was carried out as described above using only female mice. At the time of tumor cell injection, a pellet designed to release 0.2 mg of testosterone daily (Innovative Research of America, Sarasota, Fla) was inserted subcutaneously. Control females received a placebo pellet. Serum testosterone was measured using an enzyme immunoassay kit (BioCheck, Inc., Foster City, Calif) at baseline and at the conclusion of the experiment 2 weeks later. PTHrP was measured in tumor.
Tumor Burden Assessment
Tumor homogenate DsRed fluorescence was measured in 0.05-mL samples in 96-well plates with a SpectraMax 190 plate reader (Molecular Devices, Sunnyvale, Calif) set to 540 nm for excitation wavelength and 590 nm for emission wavelength. Tumor homogenate total protein was measured using the BCA Protein Assay (Pierce Chemical Company, Rockford, Ill). Serum was assayed for interleukin 8 (IL-8) by using a sandwich enzyme-linked immunosorbent assay (BioSource International, Camarillo, Calif) in microtiter plates coated with primary antibody, as described by Gujral and colleagues.24 Biotinylated detection antibody was added for 18 hours at 4°C, followed by streptavidin-conjugated to β-D-galactosidase. Color was developed with 4-methyl umbelliferyl-β-D-galactopyranoside (Calbiochem, San Diego, Calif) as the substrate. The lower limit of detection was 3 pg/mL IL-8.
Fluorescent images were analyzed on a Dell computer equipped with a Pentium 4 processor, 512 MB RAM, and Windows XP Pro operating system (Microsoft Corporation, Redmond, Wash). Tumor image areas and tumor image integrated pixel intensities were quantitated from the digitized images using Image-Pro Plus 4.5.1 software (Media Cybernetics, Silver Spring, Md).
Human Microarray Data
Human lung carcinoma PTHrP expression data were obtained from a published and publicly available data set. The set includes messenger RNA (mRNA) measurements from 51 men and 74 women with lung adenocarcinoma obtained with a U95Av2 microarray (Affymetrix, Inc.). This chip contains 2 oligonucleotide probes specific for the common coding region of PTHrP. The data set is available at www.pnsas.org/cont/vol0/issue2001/images/data/191502998/DC1/DatasetA_12600gene_floor_fig1order.xls.25 Expression values were scaled, and negative values were set to 0. For each patient, results from the 2 PTHrP probes were averaged. Finally, lung carcinoma PTHrP expression was compared between the men and the women.
Reverse Transcriptase-Polymerase Chain Reaction Analysis
RNA from cultured lung cancer cells was isolated with the RNeasy mini kit (Qiagen, Valencia, Calif). RNA was eluted with 40 μL of RNase-free water and quantified based on optical density readings; 1 μg of total RNA was reverse transcribed in a final volume of 25 μL using Superscript II reverse transcriptase (Invitrogen) for 1 hour at 42°C. The primer sequences were as follows: forward 5′-TGGACACGACAACAAC CAGCC-3′, reverse 5′-CTGGTAGAAGCGTCTTGAGC-3′ for the androgen receptor (AR); forward 5′-CTG TTACTGGTCCAGGTTCA-3′, reverse 5′-CCAGCTGAT CATGTGTACCA-3′ for estrogen receptor β (ER-β); and forward 5′-GGAGACATGAGAGCTGGCCAA-3′, reverse 5′-CACCACGTTCTTGCACTTCA-3′ for ER-α. The complementary DNAs were amplified using 0.1 μL of Hot Star Taq DNA Polymerase (Qiagen) for 40 polymerase chain reaction (PCR) cycles (initial Taq activation at 95°C for 15 minutes, denaturing at 94°C for 30 seconds, annealing at 55°C for 30 seconds, extension at 72°C for 1 minute, followed by an additional extension cycle at 72°C for 10 minutes). Settings were the same for ER-β, except that the annealing phase took place at 56°C. The PCR products were separated by electrophoresis in a 2.5% agarose gel containing 0.5 μg/mL ethidium bromide for 1 hour and 20 minutes at 85 V and were identified using the Flurochem 8900 ChemiImager (Alpha Innotech, San Leandro, Calif). Products matched the expected sizes of 514 nucleotides for AR, 529 nucleotides for ER-β, 752 nucleotides for ER-α, and 176 nucleotides for the glyceraldehyde 3-phosphate dehydrogenase loading control. LNCaP prostate carcinoma cells, MCF7 breast carcinoma cells, and T47 breast carcinoma cells provided positive controls.
Western Blot Analysis
BEN, A549, and H460 cell lysate proteins were separated by electrophoresis on Criterion 4–12% Bis-Tris gels (Bio-Rad, Hercules, Calif) and transferred to polyvinyl difluoride membranes. Blocking was done with 10% nonfat dry milk for 1 hour at room temperature. Antibodies came from Santa Cruz Biotechnology (Santa Cruz, Calif). Primary antibodies were mouse monoclonal antibody AR (441) against the AR, rabbit polyclonal antibody ER-β (H-150) against ER-β, and rabbit polyclonal antibody ER-α (HC-20) against ER-α. Blots were incubated with primary antibodies overnight at 4°C, then exposed to goat antimouse or goat antirabbit immunoglobulin G-horseradish peroxidase secondary antibodies for 1 hour at room temperature. Chemiluminescence was generated with SuperSignal West Pico Chemiluminescent Substrate (Pierce, Rockford, Ill), and the light was recorded on film.
Treatment of Lung Cancer Cells with Sex Steroids
BEN, A549, and H460 cells were grown in 6-well plates. Cells were cultured for 48 hours prior to steroid treatment in phenol red-free RPMI 1640 with 10% charcoal-stripped serum to avoid the estrogenic effects of phenol red and to eliminate steroids contained in serum. Cells were treated for 24 hours with from 0.1 to 10 nM 17β-estradiol, testosterone, 5-dihydrotestosterone (5-DHT), or R1881, a synthetic AR agonist that does not undergo 5α reduction.26 PTHrP was measured by radioimmunoassay in conditioned media and cell lysates.
Data were compared among groups by 2-tailed unpaired t tests or analyses of variance with StatView software (version 4.5; Abacus Concepts Inc., Berkeley, Calif). Statistical significance was accepted if the probability of a type II error was <.05.
Orthotopic Lung Carcinomas
Of the 39 mice that were injected with BEN cells (Fig. 1A), 38 mice (97%) developed parenchymal lung carcinomas. The tumors were composed of sheets of pleomorphic cells resembling cultured BEN cells and had a squamous morphology (Fig. 1E,F). The tumors were visible through the chest wall because of the DsRed (Fig. 1B). Normal lung, heart, and chest wall had minimal autofluorescence, allowing the tumors to be recognized easily in the opened chest cavity (Fig. 1C,D). On fluorescent photographs (Fig. 1C), tumor boundaries were demarcated readily for quantitation of tumor area and integrated pixel intensity as estimates of tumor burden.
Tumor PTHrP Expression Varies with Sex
Orthotopic BEN lung tumors that were formed contained approximately 3 times more PTHrP per mass of tumor in female mice than in male mice (Fig. 2A). Serum PTHrP concentrations averaged 1296 ± 148 pg/mL, and levels did not differ significantly between female and male mice. The mean ± standard deviation serum PTHrP concentrations in normal athymic mice without lung cancer were 130 ± 50 pg/mL (n = 6), roughly 10% of the levels in animals bearing orthotopic lung tumors. Like in the mice, PTHrP was related to sex in lung carcinomas in humans. Levels of PTHrP mRNA were significantly higher, over 2-fold greater, in adenocarcinomas from women than from men (Fig. 2B).
Effect of Sex on Lung Carcinoma Growth
Table 1 compares several measures of orthotopic lung carcinoma burden in male and female mice. Tumor mass, tumor total protein, and tumor DsRed fluorescence all were significantly greater in males than in females (P < .05). Tumor area and tumor sum intensity on digital fluorescent images also were greater in males than in females, approaching but not reaching the cut-off established prospectively for significance (P = .07 and P = .06, respectively). In addition, serum IL-8 levels were significantly greater in males than in females. Our IL-8 assay measures a specific human protein that is not present in mice and could be derived only from human tumor cells.
|Variable||Mean ± SE||P|
|Male mice (n = 6)||Female mice (n = 6)|
|Tumor mass, g||0.25 ± 0.04||0.10 ± 0.02||.01|
|Tumor total protein, mg/mL||49 ± 5||31 ± 4||.02|
|Tumor fluorescence, RFU||938 ± 133||298 ± 88||.003|
|Tumor image area, pixels||27 ± 5||14 ± 3||.07|
|Tumor image sum intensity, kilovoxels||604 ± 138||267 ± 72||.06|
|Serum IL-8, pg/mL||240 ± 65||90 ± 21||.01|
We evaluated whether the markers of tumor burden followed similar trends across mice with differing quantities of tumor. The measures tracked each other well. For example, the relations between tumor mass and tumor total protein, tumor DsRed fluorescence, tumor image area, and tumor image sum intensity were linear with Pearson correlation coefficients (R) of 0.89, 0.90, 0.86, and 0.88, respectively (P < .0001). The relation were similar for male and female mice.
Serum Calcium Levels
Terminal serum calcium levels in animals that carried orthotopic lung tumors were 11.5 ± 0.2 mg/dL in female mice and 11.4 ± 0.2 mg/dL in male mice, which were not significantly different from the values measured in mice without tumors (11.2 ± 0.2 mg/dL; n = 15 mice).
AR and ER Status
BEN, A549, and H460 lung carcinoma cell lines all expressed the AR and ER-β based on PCR and Western blot analyses (Fig. 3). Neither cell line expressed ER-α (data not shown).
Effects of Sex Hormones on PTHrP Expression
Exposure of A549 cells, BEN cells, and H460 cells to 17β-estradiol had no effect on PTHrP secretion, cell PTHrP content, or total PTHrP (Fig. 4). In contrast, R1881, a synthetic AR agonist, significantly reduced total PTHrP levels in all 3 types of human lung cancer cells. In separate experiments (Table 2), treatment with testosterone mimicked the effects of R1881, with 5-DHT demonstrating similar, lesser effects. The magnitude of the inhibitory effects varied with cell line and androgen. For example, H460 cells responded to R1881 less than the other cell lines but demonstrated greater effects in response to 5-DHT and testosterone.
|BEN||1.5 ± 0.1||1.3 ± 0.1†||1.2 ± 0.1†|
|A549||2.9 ± 0.3||2.8 ± 0.1||1.8 ± 0.1†|
|H460||2.3 ± 0.1||1.9 ± 0.1†||1.8 ± 0.1†|
Returning to our orthotopic carcinoma model, female mice were treated with exogenous testosterone. Serum testosterone levels at baseline averaged 0.07 ± 0.01 ng/mL with no difference between control animals and experimental animals. After 2 weeks, serum testosterone levels in treated females were almost 50 times greater than in controls (Fig. 5A). Lung carcinoma PTHrP concentrations were significantly lower in the testosterone-treated female mice compared with the control female mice (Fig. 5B).
We recently observed that tumor PTHrP was associated with increased survival in patients with nonsmall cell lung cancer in a sex-dependent manner.15 We observed that women with carcinomas that displayed PTHrP immunoreactivity had a median survival almost 3 years longer than patients with PTHrP-immunonegative carcinomas. In contrast, PTHrP expression was unrelated to survival in men with lung cancer. Because PTHrP slows the growth of PTHrP-expressing orthotopic lung carcinomas in nude mice,8 the protein may have a causal relation with prolonged patient survival. The difference between the sexes suggested that a sex-related mechanism permitted the prosurvival effect in women or prevented it in men. In the current study, we explored 1 possible cause for the sex-dependence. We proposed that tumors in women express greater PTHrP than tumors in men. The survival effect would then occur preferentially in women because of higher levels of PTHrP to inhibit tumor growth.
Data from 2 separate models, a mouse orthotopic lung carcinoma model and a database of gene expression measurements from human lung adenocarcinomas, supported the hypothesis that tumor PTHrP varies with sex. Female mice demonstrated excess tumor PTHrP content over males, even though the carcinomas were formed from the same number of BEN lung cancer cells in each mouse. Thus, tumor PTHrP levels were modified by something related to the sex of the animal carrying the carcinoma. Similarly, PTHrP mRNA levels were increased in lung adenocarcinomas from women compared with men. These results are consistent with sex effects that have been observed for PTHrP expression in other systems. For example, PTHrP levels are greater in the ventricular myocardium of women compared with men.27
We also investigated whether lung carcinomas in female mice, with their greater level of PTHrP, grew at a slower rate than tumors in male mice. Several measures of tumor burden suggested that carcinomas were smaller in females than males after the same period of growth. Additional work will be necessary to determine whether the difference in PTHrP levels contributed to the difference in tumor growth and to verify the effects with multiple carcinoma cell lines. Variation in lung carcinoma PTHrP expression with sex may contribute to differential effects of the protein on patient survival. Other sex-related factors also may play a role and should be investigated.
Sex hormones regulate PTHrP in other tissues,28, 29 and human lung cancers commonly express ARs and ERs.30 Hence, circulating gonadal hormones may cause lung carcinoma PTHrP expression to vary with sex. We studied 3 AR-positive and ER-β-positive human lung cancer cell lines that represented squamous, adenocarcinoma, and large cell histologies. In all 3 cell lines, R1881, a potent synthetic AR ligand, inhibited PTHrP expression, but estradiol had no effect. The lesser effects in H460 cells suggest that individual lung cancer cell lines may vary in their androgen responsiveness and reflect the finding that multiple factors regulate PTHrP. Testosterone had similar effects to R1881, although they were smaller in magnitude. The minimal effects of 5-DHT in A549 cells may be caused by the known, rapid inactivation of 5-DHT by those cells.31 Male athymic mice have serum testosterone levels some 10-fold higher than the levels in female mice.32 Thus, the lower levels of PTHrP in tumors in the male mice may be caused by the effects of circulating testosterone. Our last experiment tested this hypothesis by administering exogenous testosterone to female mice to achieve serum values similar to those in males. The treatment decreased lung carcinoma PTHrP levels, as predicted. Similarly, lung carcinoma PTHrP mRNA levels may be lower in men because of the effects of testosterone.
In summary, these studies provide evidence that lung cancer PTHrP expression varies with sex. Lung tumors in females express more PTHrP than in males. Androgens may be responsible for the differential expression by sex, because the androgen receptor agonist R1881 inhibits PTHrP expression in cultured human lung cancer cells, and female mice have lower tumor PTHrP levels when treated with testosterone. We also present data supporting the hypothesis that PTHrP inhibits lung carcinoma growth. Female mice, which have greater tumor PTHrP content than males, also have significantly less tumor burden than males, as evidenced by multiple independent markers. Additional work will be necessary to determine the etiology for the actions of sex steroids on PTHrP expression, to learn the conditions under which PTHrP inhibits lung carcinoma growth, and to discern the mechanisms for the effect.
We thank Kathy Smith and Cheryl Chalberg for assistance with radioimmunoassays and Su Tu for cell culture work.
- 23Immunoassays for PTH and PTHrP. In: BilezikianJP, MarcusR, LevineA, eds. The Parathyroids. San Diego, Calif: Academic Press; 2001: 143–165..