Estrogen and Cytochrome P450 1B1 Contribute to Both Early- and Late-Stage Head and Neck Carcinogenesis

Head and neck cancer, the sixth most common cancer in the United States, accounts for 650,000 new cancer cases each year worldwide (1). It is a heterogeneous group of malignancies that develop primarily in the squamous epithelium of the oral cavity, pharynx, larynx, nasal cavity, and paranasal sinuses (1, 2). A rise in the incidence of squamous cell carcinoma of the head and neck (HNSCC) in adults age 40 or less has been reported recently by several groups and attributed primarily to an increase in the prevalence of tongue cancers (3–5). These statistics dictate the need to identify early molecular events within the oral epithelium that may serve as targets for preventive intervention.

Exposure to tobacco smoke and use of alcohol are major risk factors for developing HNSCC (1). Infection with human papilloma virus (HPV) has been associated with a subset of HNSCCs, SCC of the oropharynx (6). However, the lack of an association of a substantial proportion of HNSCC cases with exposure to these established risk factors (7–9) suggests that additional genetic and/or environmental factors may contribute to disease susceptibility. A recent report indicates that 75% of young never-smoker/never-drinker HNSCC patients who develop primarily oral tongue SCC, (generally not associated with HPV) are females (9). These data suggest that, in addition to the major risk factors, female hormones may contribute to head and neck carcinogenesis.

Recent data from this group indicate that the estrogen metabolism pathway is altered in lung tissue following tobacco smoke exposure (10), suggesting that estrogen metabolism may play a role in the formation of other cancers of the aerodigestive tract. In this study, cytochrome P450 (CYP) 1B1 was identified as the gene whose expression was modulated to the greatest extent by smoke (10). CYP1B1 is the major enzyme that along with CYP1A1 and CYP3A4 catalyzes the formation of carcinogenic metabolites of both 17β-estradiol (E₂; ref. 11) and constituents of tobacco smoke (12) that are subsequently inactivated by one or more detoxification enzymes including catechol-o-methyltransferase (COMT), sulfotransferase (SULT)1A1, UDP-glucuronosyltransferase (UGT)1A1, and glutathione-S-transferase (GST)P1. Little attention has been given to the importance of the estrogen metabolism pathway in HNSCC.

The goals of the present study were to characterize the expression of estrogen synthesis and metabolism enzymes within the head and neck epithelium, assess the impact of E₂ exposure on the expression of these genes and evaluate the impact of both E₂ exposure and/or CYP1B1 depletion on the motility, apoptosis, and proliferation of cultured cells. To circumvent the inherent heterogeneity of HNSCCs, in vitro studies focused on oral tongue SCC, the most frequently diagnosed malignancy of the oral cavity (5)—one that is not associated with HPV infection (13). Using a cell line established from a human premalignant lesion (MSK-Leuk1; refs. 14, 15) and 5 human tongue SCC cell lines, we demonstrated that estrogen metabolism genes are expressed during both early and late stages of head and neck carcinogenesis. We report here that E₂ exposure induces CYP1B1 in MSK-Leuk1 but not in HNSCC cells. Knockdown of CYP1B1 by shRNA inhibits migration and proliferation of MSK-Leuk1 cells. Collectively, these data provide novel insight into the mechanisms underlying carcinogenesis of the head and neck and may improve our ability to identify individuals at increased risk for HNSCC, as well as facilitate the search for new targets for intervention.

MSK-Leuk1 cells were derived from a dysplastic leukoplakia lesion located adjacent to a SCC of the tongue (14, 15). MSK-Leuk1 cells were obtained from Dr. Peter Sacks (Memorial Sloan-Kettering Cancer Center) and cultured in KGM (keratinocyte growth medium; Lonza). MSK-Leuk1 cells (passage 33) were determined to be identical to the early passage MSK-Leuk1 cells (Identity Mapping Kit, Coriell Institute for Medical Research). All HNSCC cell lines were derived from patients with SCC of the tongue. SCC9 (male) and SCC15 (male) cells were provided by Dr. Andres Klein-Szanto and cultured in S-MEM medium, supplemented with 2 mmol/L of l-glutamine, 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 10% FBS (16). UPCI:SCC56 (male), UPCI:SCC103 (female), and UPCI:SCC122 (male) cells were obtained from Dr. Susanne Gollin (University of Pittsburgh Cancer Institute; ref. 17) and cultured in MEM medium, supplemented with 2 mmol/L of l-glutamine, 100 μmol/L of nonessential amino acids, 50 μg/mL of gentamycin (Gibco), and 10% FBS. All reagents for cell culture were provided by the Cell Culture Facility at Fox Chase Cancer Center, unless otherwise specified.

For all the experiments that involved E₂ exposure, MSK-Leuk1 cells were cultured in phenol red-free and serum-free DermaLife K Medium (Lifeline Cell Technology). SCC cells were cultured in their respective media with no phenol red, supplemented with charcoal-stripped serum (Gibco). Cells were incubated for 48 hours to remove endogenous estrogens and then plated at 70% confluence. After 24 hours, the medium was replaced with either control medium containing vehicle (0.01% ethanol) or medium supplemented with 1 nmol/L of E₂ (Sigma-Aldrich). Cells were harvested after the appropriate treatment period and analyzed.

A set of 5 lentivirus-encoded shRNA constructs specific for CYP1B1 (clone id TRCN0000062323-TRCN0000062327) and the empty pLKO.1 vector (control) were obtained from Open Biosystems. Each of the 5 constructs and the pLKO.1 vector were cotransfected along with the ViraPower Lentiviral Packaging Mix (Invitrogen) into 293FT producer cells (provided by Dr. R. Zhang, Fox Chase Cancer Center), using Lipofectamine 2000 (Invitrogen). The viral supernatants were harvested and viral titers [10⁵–10⁶ transduction units/mL] were determined using puromycin selection of normal human fibroblasts. MSK-Leuk1 cells were incubated with different dilutions of the viral supernatants and allowed to recover in complete medium. Transfection efficiency was estimated on the basis of transfecting cells with a construct carrying green fluorescent protein and approached 100%. Stable clones were selected using puromycin (10 μg/mL; Sigma-Aldrich) and analyzed for CYP1B1 levels by Western blot.

MSK-Leuk1 cells, expressing either vector or CYP1B1 shRNA, were cultured in phenol red-free and serum-free medium for 48 hours and then plated at 70% confluence. After 24 hours, the cells were treated with either vehicle or E₂ (1 nmol/L) in triplicate, as described above. When cells reached 100% confluence (48 hours later), the surface of the cell culture dish was carefully scratched using a micropipette tip, thus making an evenly distributed gap in the cell monolayer. The medium was replaced, and 5 representative images of each gap were acquired at 0 hours using a Nikon TE-2000U wide-field, inverted microscope (Optical Apparatus Co.) equipped with a Roper Scientific Cool Snap HQ camera. Another set of 5 to 10 representative images per gap was obtained following a 16-hour incubation. The area devoid of cells was measured on every image using MetaMorph 7.0 (Molecular Devices, Inc.). The gap closure percentages were calculated as (area at 0 hour − area at 16 hours)/(area at 0 hour).

In addition, a time-lapse movie capturing the process of gap closure in vector-expressing MSK-Leuk1 cells was obtained. The medium was replaced with fresh medium containing 25 mmol/L of HEPES buffer and cells were allowed to incubate for 1 hour at 37°C. A preset location was photographed every 10 minutes for a period of 16 hours using the same microscope and camera setup as above. The percentage of proliferating cells (those rounded up for cell division) was counted in this representative area.

Apoptosis was assessed using the Guava Nexin kit (Millipore). Fifty thousand cells were plated per well in 6-well plates. After the appropriate treatment, floating cells were collected, combined with attached cells following trypsinization, and resuspended in DermaLife K Medium (Lifeline Cell Technology) supplemented with 5% FBS. The cell suspension (100 μL) was incubated with 100 μL of Guava Nexin Reagent for 20 minutes, according to the manufacturer's instructions. Two thousand cells were analyzed from each sample using the Guava EasyCyte system and the resulting data were expressed as a percentage of apoptotic cells (annexin V-positive cells/total number of cells counted).

Fifty thousand cells per well were plated in 6-well plates. After the appropriate treatment, the DNA content of the cells, an indirect measure of proliferation, was determined as described previously (18) using a Fluorescent DNA Quantitation kit (Bio-Rad Laboratories). In brief, cells were harvested, sonicated in 0.1× TEN assay buffer (Bio-Rad Laboratories) for 5 seconds and incubated with a Hoechst dye mixture (BioRad Laboratories) for 1 hour. Total DNA was measured using Fluoroscan Ascent FL (Thermo Fisher Scientific) at an excitation wavelength of 360 nm and an emission wavelength of 460 nm.

Tissue microarrays (TMA) were generously provided by Dr. Mark Lingen, University of Chicago. TMAs contained duplicate tissue cores of surgical head and neck specimens from 128 patients, including 116 samples of HNSCC, 20 samples of dysplasia, and 37 samples of normal epithelium from different sites within the head and neck. Mean age at diagnosis was 64 years (range: 30–90 years). Sixty-nine percent of the patients were males and 24% were females, with gender unknown for the remaining 7% of patients. The characteristics of this population are summarized in Table 1 of the Supplemental data.

Immunohistochemical analyses were performed on histologic sections of formalin-fixed, paraffin-embedded human head and neck TMAs and cytospins of cultured human head and neck cells. Sections were stained with antibodies against human CYP1B1 (raised in rabbit; Alpha Diagnostics International Inc.), estrogen receptor (ER) α (raised in mouse; Lab Vision Products), ERβ (raised in mouse, Serotec) and E₂ (raised in rabbit, BioGenex) using standard immunohistochemical procedures (QualTek Molecular Laboratories). Sodium citrate (pH 6) was used for antigen retrieval. Human breast carcinoma was used as a positive control for each antibody. TMA sections were scanned and images were captured using an Automated Cellular Imaging System (ACIS, ChromaVision). Pathologically confirmed regions of HNSCC, dysplasia, and normal epithelium were scanned using a 40× objective. Following normalization to a threshold (background staining), the staining intensity of each selected area was quantified and expressed in arbitrary units.

Cells were solubilized in RIPA buffer containing 150 mmol of NaCl, 1% Na Deoxycholate, 1% Triton X-100, 0.1% SDS, 10 mmol/L of Tris-Base, 50 mmol/L of NaF, 0.1 mmol/L of Na3VO4, supplemented with protease inhibitor cocktail (Roche). One hundred micrograms of total protein was separated on a 10% SDS PAGE (Bio-Rad) and electroblotted onto a polyvinylidene fluoride membrane. Membranes were blocked for 1 hour at room temperature in Tris-buffered saline with Tween-20 (TBST; 50 mmol/L of Tris-HCl, pH 7.5, 150 mmol/L of NaCl, 0.1% Tween-20) containing 5% nonfat milk and incubated overnight at 4°C with primary antibodies against either CYP1B1 (Imgenex Corp.; Catalog No: IMG-5988A), ERα (Santa-Cruz Biotechnology Inc.; H-184:sc-7207), ERβ (Millipore; Catalog No: 05-824) or HPRT (hypoxanthine phosphoribosyltransferase; Abcam, Inc.; Catalog No: ab10479). After washing 3 times with TBST, the membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG secondary antibody (Bio-Rad; 1 hour at room temperature), rinsed with TBST, and visualized using ECL Western Blotting Detection Reagents (GE Healthcare).

RNA was extracted from pelleted cultured cells using the RNAeasy Mini kit (Qiagen) and its quality (18S and 28S RNA) was evaluated by gel electrophoresis. cDNA was synthesized from 1 μg of total RNA, using the High Capacity cDNA Archive Kit (Applied Biosystems, Inc.). Three microliters of the resulting cDNA was mixed with 12.5 μL of 2× TaqMan Universal PCR master mix and 1.25 μL of 20× primer mix (Applied Biosystems) in a final reaction volume of 25 μL, according to the manufacturer's instructions. Reactions were performed in triplicate in an Applied Biosystems 7900HT Fast Real-Time PCR System using universal conditions. Transcript levels were quantified and expressed relative to those of the human transferrin receptor (TFRC) as 2^−ΔCt (19).

For the gene expression, cell proliferation, and apoptosis analyses, Student's t test (Excel) was used to analyze differences between the groups. For the cell migration assay, statistical analyses were performed using the 2-sided Mann–Whitney test (Instat Statistical Software, GraphPad Software).

TMAs were analyzed by comparing the staining intensity of each antibody in HNSCCs, dysplasias, and normal head and neck epithelium using a pairwise approach. When appropriate, the staining intensities of each antibody were compared within each tissue type by gender. All TMA statistical analyses were done using the 2-sided Mann–Whitney test. The Benjamini–Hochberg false discovery rate approach was used to account for multiple testing (20). All P values listed for TMA analyses were corrected P values on the basis of this approach. The R statistical language and environment was used for these analyses (21).

Immunohistochemical staining of sections from formalin-fixed, paraffin-embedded pellets of MSK-Leuk1 cells and 5 HNSCC cell lines was performed using antibodies specific for ERα, ERβ, and CYP1B1. ERβ and CYP1B1 were detected in MSK-Leuk1 cells and all HNSCC cell lines at comparable levels, with staining for both proteins localized to the nucleus. ERα was not detected in any of the cell lines evaluated (Fig. 1A). Consistent with immunohistochemical staining data, ERβ and CYP1B1 were detected by Western blot in all head and neck lines. Whereas ERα was detected in MCF-7 cells (positive control), it was not detected in any of the head and neck cell lines (Fig. 1B).

The finding that ERβ and CYP1B1 are present in cultured head and neck cells was extended by examining the expression profile of CYP19 (aromatase), which encodes the rate-limiting enzyme in estrogen synthesis, and several estrogen metabolism genes in both MSK-Leuk1 cells and 5 HNSCC cell lines. Transcripts for CYP1B1, CYP1A1, COMT, UGT1A, and GSTP1 were detected in all cell lines (Fig. 1B), whereas transcripts for CYP3A4 and SULT1A1 were near the limits of detection (not shown). The most abundant transcripts in all cell lines were those encoding the conjugation genes COMT and GSTP1. The level of CYP19 transcripts was below the limit of detection in all cell lines evaluated (data not shown).

The effect of E₂ exposure on transcript levels of ERβ, CYP1B1, and COMT, which encodes the major estrogen and xenobiotic conjugation enzyme, was examined in cultured MSK-Leuk1 and HNSCC cells. In MSK-Leuk1 cells, E₂ treatment (24 hours) increased the levels of CYP1B1 transcripts 2.3-fold as compared with those of vehicle-treated controls (P = 0.0004; Fig. 2A). The induction of CYP1B1 by E₂ appeared to be time dependent in MSK-Leuk1 cells and peaked after 6 hours of exposure (Fig. 2B). In contrast, the levels of COMT, ERβ, as well as the reference gene TFRC, remained unaltered in MSK-Leuk1 cells posttreatment. Treatment of HNSCC cells with E₂ did not alter the level of any of the transcripts of interest (ERβ, CYP1B1, and COMT).

To investigate the contribution of CYP1B1 to cancer progression, MSK-Leuk1 cells deficient in CYP1B1 were constructed using a lentivirus system to express shRNA specific to CYP1B1 mRNA. Western blot analyses indicated that CYP1B1 levels were decreased in cells expressing CYP1B1 shRNA, relative to control cells that expressed the vector (Fig. 3A).

The motility of CYP1B1-deficient MSK-Leuk1 cells was compared with that of cells expressing control vector (treated with either vehicle or E₂). The rate of motility of CYP1B1-deficient cells measured as the ability of the cells to repopulate a scratched area of a previously confluent monolayer was 54% to 57% lower than that of control cells expressing the basic vector (P < 0.0001; Fig. 3B and C). Motility was not affected by E₂ treatment. Rates of proliferation and apoptosis were comparable in CYP1B1 shRNA- and vector-expressing cells during the time period when cell migration was analyzed (16 hours; Fig. 3D and E).

To further confirm that the observed gap closure was due to the migration and not proliferation of the cells, the motility of vector-expressing MSK-Leuk1 cells was observed in real time over a 16-hour period. The cells were motile, with approximately 20% dividing during the observation period. No difference was observed in the rate of proliferation of cells infiltrating the gap, as compared with those in the cell monolayer outside of the gap (Supplemental data, Movie 1).

To explore the role of E₂ in head and neck carcinogenesis, MSK-Leuk1 cells expressing either vector or CYP1B1 shRNA were incubated in the presence or absence of E₂ for 72 hours. The proliferation of cells expressing CYP1B1 shRNA was decreased as compared with that of vector-expressing cells irrespective of E₂ exposure (44.6% for vehicle-treated cells; P = 0.025) and 47.6% for E₂-treated cells (P = 0.006; Fig. 4A). E₂ exposure induced cell proliferation in vector-expressing cells by 10%; however, this increase was not statistically significant (Fig. 4A). CYP1B1 depletion did not affect apoptosis (Fig. 4B). However, exposure to E₂ decreased apoptosis in both vector-expressing (25.5%; P = 0.030) and CYP1B1 shRNA-expressing (30.1%; P = 0.015) cells (Fig. 4B). This E₂-mediated decrease in apoptosis was restored by the addition of the pure antiestrogen fulvestrant (Fig. 4C).

Representation of the estrogen pathway in human head and neck tissue was examined using TMAs of head and neck surgical specimens. TMAs were stained with antibodies against ERα, ERβ, CYP1B1, and E₂. The majority of samples stained positive for ERβ (91.9%), CYP1B1 (99.4%), and E₂ (88.4%), irrespective of gender. Staining of ERβ and CYP1B1 was localized to the nucleus, whereas staining of E₂ was observed in both the nucleus and cytoplasm (Fig. 5). Staining of ERα was detected in only a few cases (1.7%).

The staining intensity of ERβ, CYP1B1, and E₂ was quantified in pathologically confirmed areas of cancer, dysplasia, and normal epithelium using the ACIS (ChromaVision). Staining intensity of CYP1B1 and ERβ were both higher in HNSCCs as compared with normal epithelium (P = 0.024 and 0.008, respectively; Fig. 5; Table 1). No difference in the intensity of E₂ staining was observed between HNSCCs, dysplasias, or normal epithelium (Fig. 5; Table 1). In addition, no difference between males and females was observed in the staining intensity of any of the antibodies when either normal epithelium or HNSCCs were analyzed. Because HPV infection is associated with a better prognosis in HNSCC patients (22), the analyses were next restricted to sites of the head and neck not routinely associated with HPV infection. Similar to the results obtained when evaluating all specimens, the intensity of ERβ staining in the potentially non–HPV-associated tumors was elevated as compared with normal epithelium (P = 0.007). CYP1B1 staining intensity was also elevated in HNSCCs relative to normal epithelium and approached statistical significance (P = 0.07). No difference in the intensity of E₂ staining was observed between HNSCC, dysplasia, or normal epithelium for the potentially non–HPV-associated cases.

Results from the present study demonstrate for the first time that a panel of estrogen metabolism genes is expressed in cultured human head and neck cells. Detection of transcripts for these genes in both premalignant lesions and HNSCCs suggests that these enzymes contribute to cellular metabolism throughout tumorigenesis. Although the importance of E₂ signaling and ERβ in HNSCC has been suggested previously (23–25), the contribution of the estrogen pathway to the premalignant stage of head and neck tumorigenesis has not been evaluated. Furthermore, even though the expression of individual xenobiotic/hormone metabolism genes has been evaluated in head and neck cells (26–28), the potential role of estrogen in head and neck carcinogenesis has not been investigated in a comprehensive manner.

We demonstrate here for the first time that CYP1B1 is upregulated in MSK-Leuk1 but not in HNSCC cells following E₂ exposure. The mechanistic basis for this differential upregulation of CYP1B1 remains unclear. However, it has been shown for lung cancer that the timing of hormone exposure relative to a diagnosis of lung cancer may make a difference with respect to whether the hormonal effect is protective or adverse (29). CYP1B1 metabolizes hormones, including E₂, and xenobiotics, including tobacco-associated carcinogens, to species that can cause DNA damage (11, 30). Enhanced expression of CYP1B1, in the absence of an elevation in the expression of the conjugation gene COMT, could potentially promote the accumulation of mutagenic DNA damage and contribute to the formation of HNSCCs. Indeed, the CYP1B1*3 allele, which confers increased catalytic activity, has been associated with increased susceptibility for HNSCC (31, 32).

To assess the functional role of CYP1B1 in premalignant head and neck cells, the migratory potential of MSK-Leuk1 cells deficient in CYP1B1 was compared with that of the same cell line carrying control vector. The novel finding that CYP1B1 deficiency leads to the decreased migration of oral leukoplakia cells is consistent with a prior report of an association between CYP1B1 depletion and the decreased invasiveness of cultured endometrial cancer cells (33). The invasive properties of MSK-Leuk1 cells were previously shown to be enhanced by tobacco smoke via induced expression of the urokinase-type plasminogen activator (34). Tobacco smoke also induces CYP1B1 (14, 26); however, the role of this pathway in the invasiveness of MSK-Leuk1 cells was not investigated. In the present study, the observed decrease in cell migration was not attributed to decreased proliferation (Fig. 3). However, when a longer exposure time was investigated (72 hours; Fig. 4), cell proliferation was decreased in CYP1B1-deficient cells, as compared with vector-expressing cells. The ability of CYP1B1, independent of E₂, to promote the migration and proliferation of oral premalignant cells suggests that it may play a role in the clonal spread of leukoplakic lesions within the oral mucosa and facilitate cancer progression within the head and neck.

Exposure to E₂ failed to alter the rate of cell proliferation in MSK-Leuk1 cells, irrespective of CYP1B1 levels. In contrast, E₂ inhibited apoptosis in both control and CYP1B1-deficient MSK-Leuk1 cells. A previous report indicates that E₂ exposure increased the proliferation of cultured HNSCC cells derived from locally advanced tumors marginally; however, the effect of E₂ on apoptosis was not analyzed (23). The observed ability of E₂ to decrease apoptosis in premalignant cultured cells suggests that estrogens may be involved in the progression of premalignant lesions to HNSCCs. The ability of the pure antiestrogen fulvestrant to antagonize E₂-mediated inhibition of apoptosis, suggests that this effect is ER mediated and that antiestrogens may be beneficial as chemopreventive agents for HNSCC.

Using TMAs of surgical specimens from 128 patients, we have demonstrated that CYP1B1 protein is present at detectable levels in normal, dysplastic, and tumor tissues of the head and neck. Previously, CYP1B1 mRNA and/or protein have been detected in HNSCC cell lines (35, 36) and MSK-Leuk1 cells (37, 38); however, human head and neck tissues have not been analyzed for CYP1B1 protein previously. The present results further demonstrate that CYP1B1 is overexpressed in HNSCCs, as compared with the normal epithelium of the head and neck. CYP1B1 overexpression in tumors relative to normal tissue has been demonstrated for a number of organs, including breast, uterus, skin, lung, and esophagus, but not for head and neck (39, 40). Our observation of enhanced expression of CYP1B1 in HNSCCs suggests that CYP1B1 may be a marker of head and neck tumorigenesis.

MSK-Leuk1 cells, as well as cultured HNSCC cells, stained positive for ERβ expression. Consistent with these data, ERβ was detected in human HNSCCs and dysplastic tissues as well as in the normal epithelium. ER isoforms have been detected previously in human HNSCC tissues and cell lines (23–25), but not in premalignant cells, with ERβ being the predominant transcriptionally active form (23). The detection of ERβ in both dysplastic and HNSCC cells suggests the potential contribution of estrogen signaling to the development of HNSCCs at both the premalignant and malignant stages. The absence of a gender difference in the intensity of immunohistochemical staining for CYP1B1, ERβ, or E₂ indicates that the estrogen pathway may contribute to head and neck carcinogenesis in both males and females.

In summary, we have shown that a number of estrogen metabolism genes are expressed in cells derived from both early- and late-stage lesions of the head and neck. This study is the first to report the detection of estrogen within human head and neck tissue and demonstrate that both estrogen and CYP1B1 may contribute to the progression of HNSCCs. These data suggest that CYP1B1 may be an important biomarker of tumorigenesis in the head and neck and may represent a novel target for chemopreventive intervention in patients with premalignant lesions.

No potential conflicts of interest were disclosed.

We thank Drs. Rugang Zhang and Barbara Burtness for their careful review of this manuscript, Dr. John A. Ridge for helpful discussions, and Maureen Climaldi for her expert assistance in preparing this manuscript for publication. Excellent technical support was provided by the Genomics Facility, the Cell Culture Facility, the Experimental Histopathology Facility, and the Biostatistics and Bioinformatics Facility at the Fox Chase Cancer Center.

Supported by grants CA-006927 and CA113451 (EC) from the National Cancer Institute, an appropriation from the Commonwealth of Pennsylvania, including a grant from the PA Department of Health.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.