Establishment and Characterization of Novel Cell Lines from Sinonasal Undifferentiated Carcinoma

Sinonasal undifferentiated carcinoma (SNUC) is a rare, highly aggressive cancer that arises in the nasal cavity and paranasal sinuses. Initially described by Frierson and colleagues in 1986 (1), this tumor is on the spectrum of neuroendocrine sinonasal malignancies, which includes esthesioneuroblastoma, neuroendocrine carcinoma, and small cell carcinoma. There is a male predominance [2-3:1], and there are no known etiologic factors. SNUC is typically negative for Epstein-Barr virus (2, 3). In general, SNUC presents as large tumors involving multiple sinonasal structures, often extends into the orbital or cranial cavity, and can metastasize to the cervical lymph nodes, lungs, bones, brain, and liver (2, 4–6).

The treatment of SNUC includes aggressive multimodal therapy with radiotherapy and chemotherapy and, in some instances, surgery (4, 6–8). Despite aggressive management, the prognosis remains poor, and the median survival time from diagnosis is less than 18 months (2, 3).

To better understand the biologic features of SNUC and help develop new therapeutic strategies for this disease, establishing a reliable model for laboratory-based analysis is essential. While other sinonasal malignancy models have been described, these are not relevant to the biologic characteristics of SNUC (9, 10). In this report, we describe the establishment and characterization of the first human SNUC cell line and present the molecular and phenotypic behavior of this unique system.

A 74-year-old woman was diagnosed and treated at The University of Texas MD Anderson Cancer Center (MDACC; Houston, TX) for a T4N0M0 SNUC of the right maxillary sinus. She had received induction chemotherapy with 4 cycles of etoposide and carboplatin with a partial response before undergoing surgical resection of the residual tumor. Under an approved Institutional Review Board protocol, a portion of the resected specimen and matching normal tissue were collected.

Transmission electron microscopy was conducted by the High Resolution Electron Microscopy Facility at MDACC. A sample was taken from the central portion of the solid tumor from the sinonasal cavity specimen immediately after surgical extirpation and was preserved in a solution containing 3% glutaraldehyde, 2% formaldehyde, and 0.1 mol/L cacodylate (pH 7.3). Ultrathin sections were cut with an LKB Ultracut microtome (Leica), stained with uranyl acetate and lead citrate in an LKB Ultrostainer, and examined with a JEM 1010 transmission electron microscope (JEOL) at an accelerating voltage of 80 kV. Digital images were obtained using the AMT imaging system (Advanced Microscopy Techniques Corp.).

The resected specimen was rinsed in PBS containing 500 units of penicillin and 500 μg of streptomycin and was then rinsed in PBS 3 times. The specimen was minced into 1- to 2-mm³ pieces and cultured in Dulbecco's Modified Eagle's Medium (DMEM; Invitrogen Corporation) containing 10% FBS (Sigma-Aldrich), 25 units of penicillin, and 25 μg of streptomycin. After 1 month, about 40 tumor colonies were harvested using 3 MM Whatman paper soaked with 0.05% trypsin, 0.0076% EDTA 4Na solution in PBS. The colonies were then transferred to 24-well plates and designated as passage 0. Once the passage 0 cells became confluent, they were sequentially replated. Because they were sensitive to regular trypsin-EDTA solution, the cells were treated with TrypLE Express Stable (Invitrogen) when they were expanded. Cells were maintained in 10% FBS DMEM containing 0.1 mg/mL of Primocin (InvivoGen). Two populations originally from independent colonies were designated MDA8788-6 and MDA8788-7 on the basis of their initial colony numbers. Both the MDA8788-6 and MDA8788-7 lines were free from mycoplasma and murine pathogens by the Infectious Microbe PCR Amplification Test (University of Missouri, Research Animal Diagnostic Laboratory, Columbia, MO).

Three established human head and neck squamous carcinoma cells lines were used, UMSCC33, OSC19, MDA1386LN, and an established cancer-associated fibroblast cell line. All cell lines were grown on tissue culture dishes in DMEM supplemented with 10% heat-inactivated FBS.

The genomic DNA from the acquired cell lines and from the original fresh-frozen tissues were extracted using the Gentra Puregene Cell Kit (Qiagen, Inc.) following the manufacturer's instructions. Short tandem repeat (STR) genotyping was conducted by the Characterized Cell Line Core Facility at MD Anderson.

G-banding and spectral karyotyping were conducted by MD Anderson's Molecular Cytogenetics Facility by using the following procedures.

Exponentially growing cells were exposed to colcemid (0.04 μg/mL) for 25 minutes at 37°C and hypotonic treatment (0.075 mol/L KCl) for 20 minutes at room temperature. Cells were fixed in a methanol and acetic acid (3:1 by volume) mixture for 15 minutes and washed 3 times in the fixative. The slides were air dried and processed for G-banding and spectral karyotyping.

G-banding was conducted by treating the slides in trypsin and staining with Giemsa stain using a routine laboratory procedure (11) to characterize the cytogenetic alterations. Slides were analyzed using an Eclipse E400 microscope (Nikon Inc.), and images of metaphases were captured using an Applied Spectral Imaging system equipped with camera and karyotyping software. A minimum of 15 to 20 metaphases were karyotyped from each sample.

Spectral karyotyping was conducted according to the manufacturer's protocol using human SKY Paint probes (Applied Spectral Imaging). Images were captured using a Nikon 80i microscope equipped with SKY software (Applied Spectral Imaging). At least 15 to 20 metaphases were analyzed in detail.

MDA8788 cells were plated on a 96-well plate at a density of 1 × 10⁴ to 2 × 10⁴ per well. After 3, 24, 48, 72, 96, and 120 hours, the culture media were removed; 100 μL of double distilled water was added; and the plates were frozen and thawed 3 times. A total of 100 μL of Hoechst 33342 solution (20 μmol/L in 10 mmol/L Tris-HCl, 2 mol/L NaCl, 1 mmol/L EDTA, pH 7.4; Invitrogen) was added and measured at 360 nm excitation and 460 nm emission. The proliferation data were analyzed using GraphPad Prism 5 (GraphPad Software, Inc.).

Ten thousand MDA8788 cells were suspended in 0.3% agarose (Lonza Walkersville, Inc.) in DMEM containing 20% FBS. This suspension was overlaid onto a solid layer of 0.6% agarose in a 6-well plate. The cells were treated with fresh DMEM containing 20% FBS every other day. Three weeks later, the cells were fixed with methanol, stained with 0.02% crystal violet, and photographed at ×10 magnification.

For determination of anoikis resistance, 2 × 10⁶ viable MDA8788 cells were cultured as previously described (12).

Migration and invasion assays were conducted as previously described (13) using OSC19 cells as a positive control. Briefly, 4 × 10⁴ cells were plated on cell culture insert wells or Matrigel-coated wells (BD Biosciences) in DMEM containing 10% FBS, and the degree of migration was determined after 48 hours.

The cells from all lines (MDA8788-6, MDA8788-7, MDA1368LN, OSC19, UMSCC33, and cancer-associated fibroblasts) were lysed in lysis buffer (50 mmol/L HEPES, 150 mmol/L NaCl, 10 mmol/L EDTA, 10% glycerol, 1% Triton X-100, pH 7.5) with proteinase inhibitors (1 mmol/L sodium orthovanadate, 1 mmol/L phenylmethylsulfonyl fluoride, ×1 Complete protease inhibitor cocktail; Roche Holding, Inc.). Organ homogenates in TNN buffer [50 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 5 mmol/L EDTA, 0.5% Nonidet P-40] were clarified by centrifugation (total lysate). Antibodies were obtained as follows: pan-keratin (clone 80; Thermo Fisher Scientific), cytokeratin 8 (N1N3; GeneTex, Inc.), cytokeratin 19 (Ab-1, clone A53-B/A2.26; Thermo Fisher Scientific), vimentin (clone V9; DAKO), E-cadherin (clone 36/E-cadherin; BD Biosciences), N-cadherin (clone 32; BD Biosciences), synaptophysin (clone SVP-38; Millipore), and actin and α-smooth muscle (SMA; clone 1A4; Sigma-Aldrich).

Immunohistochemistry was conducted as previously described (14), using Cytokeratin Cocktail (mixture of clone AE1/AE3, DAKO; clone CAM 5.2, BD Biosciences; MNF116, DAKO; and clone Zym 5.2, Invitrogen), anti-E-cadherin (clone HECD-1; Invitrogen), anti-α-SMA (clone 1A4; Sigma-Aldrich), and anti-vimentin (clone V9; DAKO).

Male nude mice (aged 6–8 weeks) were purchased from the Animal Production Area of the National Cancer Institute-Frederick Cancer Research and Development Center (Frederick, MD). The mice were maintained in a pathogen-free environment and fed irradiated mouse chow and autoclaved reverse osmosis–treated water at facilities in accordance with current regulations and standards of the U.S. Department of Agriculture, U.S. Department of Health and Human Services, and the NIH. The mice were used in accordance with MD Anderson's Animal Care and Use Guidelines, and all animal procedures were done in accordance with a protocol approved by the institution's Institutional Animal Care and Use Committee.

A total of 7 or 8 mice per cell line underwent subcutaneous injection with 1 × 10⁶ cells suspended in a volume of 200 μL of PBS directly into the right flank with a 1-mL tuberculin syringe (Hamilton Co.) and a 30-gauge hypodermic needle. The same number of the cells suspended in a volume of 100 μL of PBS with 100 μL of Matrigel Matrix Growth Factor Reduced (BD Bioscience) were also injected into the left flank of the same mice because Matrigel is known to enhance the tumorigenicity of a wide range of cancer cell lines in vivo (15). Injected mice were then examined twice weekly for tumor development. When present, tumors were measured using calipers in cephalad-to-caudad and left-to-right dimensions. Tumor volume was calculated as V = AB² (π/6), where A is the longest dimension of the tumor and B is the dimension of the tumor perpendicular to A. When the longest dimension reached 15 mm, the animals were euthanized with CO₂. At the time of death, the tumors and lungs were harvested and placed in 10% buffered formalin solution overnight for fixation. Hematoxylin and eosin (H&E) staining was used on histologic sections.

We used the soft palate model that was established in our group (16). Briefly, 1 × 10⁶ MDA8788-6 cells were suspended in a volume of 15 μL of PBS with 15 μL Matrigel and injected into the muscle of the soft palate. To monitor orthotopic tumor growth in vivo, mice were examined by MRI after 12 days of the injection in the Small Animal Imaging Facility at MDACC. Anesthesia was induced using 5% isoflurane in oxygen, maintained using 1% to 3% in oxygen, and monitored according to respiratory rate using respiratory bellows and a small animal physiological monitoring system (Small Animal Instruments Inc.). All MRIs were acquired using a 7-T Biospec small animal imaging system (Bruker BioSpin) and imaging gradients with an inside diameter (ID) of 116 mm. A linear volume resonator (72-mm ID) was used for excitation, and a custom electronically tuned and actively decoupled RF surface coil (13 mm ID) was used for signal detection. Scout images to verify animal positioning were followed by high-resolution sagittal T₂-weighted RARE scans [TE/TR 60/3,000 ms, 20 × 15 mm² field of view (FOV), 256 × 128 image matrix, 1 mm slices, RARE factor 8, 5 averages], geometrically matched axial T₁-weighted spin echo (TE/TR 11.5/1,000 ms, 20 × 15 mm² FOV, 256 × 128 image matrix, 1 mm slices) and T₂-weighted RARE images (TE/TR 60/3,000 ms, RARE factor 8, 5 averages) of the paramedian dorsal skin.

Images were viewed using ParaVision software (Bruker BioSpin) and ImageJ analysis software (rsb.info.nih.gov/ij/). Measurements were taken after comparison of T₁ and T₂ images to confirm consistency of cutaneous microanatomic strata. All measurements were taken using ParaVision software on representative T₂-weighted axial slices.

We euthanized mice with CO₂ when they became moribund. At the time of death, the full heads of mice were obtained, fixed in formalin solution for 48 hours, decalcified in 5% formic acid for 4 days, and paraffin-embedded. H&E staining was done on histologic sections of the head to determine the extent of tumor growth and the degree of invasion into surrounding structures.

The unpaired 2-tailed t test was used to compare the differences in mean tumor volume between 2 groups (with or without Matrigel for the tumor injection) with GraphPad Prism 5. P < 0.05 was considered statistically significant.

Pathologic analysis of the original tumor was conducted and compared with the established cell lines. H&E staining of the primary tumor from the patient's biopsy showed small- to medium-sized polygonal cells that formed nests, sheets, ribbons, and trabeculae (Fig. 1A), characteristic of the classical SNUC morphology. (1) After the induction chemotherapy, the tumor had developed a slight squamoid appearance (Fig. 1B). Electron microscopy indicated undifferentiated polygonal cells with sparse intracellular membrane structures, as well as ribosomes, neurosecretory granules, and lipid-filled vacuoles (Fig. 1C, left); higher magnification showed that microtubules were present along with rough endoplasmic reticulum, polyribosomes, and membrane-bound, dense-core neurosecretory granules (Fig. 1C, right). The established cell lines grew as adherent and tightly packed monolayers with a polygonal shape and large nuclei (Fig. 1D). The morphology was maintained across all cell passages. STR fingerprinting was conducted to characterize the isogenic nature of the MDA8788 cell lines to the parent tumor. All DNA extracted from the original specimen and the 2 new lines had identical STRs (Supplementary Table). Comparison to the American Type Culture Collection STR database and STR data from 49 cell lines in our laboratory revealed that MDA8788 is a unique cell line, distinct from all others in our collection.

Cytogenetic analysis of the MDA8788 cells was conducted by using G-banding and spectral karyotyping analysis. G-banding analysis revealed that 12 clonal markers (m1–m12; Fig. 2A and Supplementary Fig. S1) were shared by both cell lines. By combining G-banding and spectral karyotyping analysis (Fig. 2B), the following translocations were identified: m1 t (1:17), m2 t (1:22), m3 t (1:15), m4 t (2:12), m5 t (4:15), m6 t (3:6), m7 t (8:19), m8 t (9:10), m9 t (9:16), m10 t (13:13), m11 t (14, 14), and m12 t (18:19).

We next assessed the replicative behavior of the established cell lines using Hoechst 33342 fluorescence dye. The doubling time of both MDA8788 cell lines was found to be approximately 40 hours, which was slower than a median doubling time of 26.5 hours in head and neck squamous cancer cell lines (10). To examine the ability of the tumors to grow under anchorage-independent conditions, we conducted soft agar assays on both cell lines. Neither cell line formed colonies after 3 weeks, whereas MDA1386LN cells formed large colonies (Fig. 3A). To test whether the MDA8788 cells are resistant to anoikis (anchorage-dependent programmed cell death), the cells were grown in suspension culture at various time points. Within 72 hours, most of the MDA8788-6 cells died, whereas the MDA8788-7 cells displayed a 30% viability (Fig. 3B). These results suggested that while the SNUC cell lines had behavior characteristics that were reminiscent of squamous cell carcinomas, the lack of growth in soft agar and the relative anoikis sensitivity suggested a more epithelial-like phenotype. To examine the migratory and invasive potential of the SNUC cells, the MDA8788-6 cells were evaluated by in vitro Transwell migration and invasion assays. MDA8788-6 did not migrate or invade, even after 48 hours, whereas OSC19 showed both invasion and migration (Supplementary Fig. S2).

As SNUC tumors display both epithelial and mesenchymal biologic features, we next sought to examine whether the established SNUC cells express epithelial and mesenchymal markers by Western blotting (Fig. 4A). Both MDA8788 cells displayed high expression of pan-keratin, whereas the original tumor specimen had minimum expression compared with the SNUC cell lines, OSC19 (used as a positive control) and UMSCC33 (used as a representative of squamous cell carcinoma from the maxillary sinus). To specify the types of cytokeratin the SNUC cells were expressing, 2 antibodies recognizing cytokeratin 8 and 19 were used for Western blotting. Our SNUC tumor specimen and cell lines were positive for both cytokeratins.

The MDA8778 cell lines also expressed the epithelial markers E-cadherin and β-catenin but not the mesenchymal markers N-cadherin, vimentin, or α-SMA (Fig. 4A). The expression pattern of the EMT markers did not change during the course of cell culture (Supplementary Fig. S3). On the other hand, the SNUC tumor was positive for all the markers including the mesenchymal markers (Fig. 4A). To determine the origin of these mesenchymal markers seen in Western blotting, we conducted an immunohistochemical analysis of surgical specimen. Examination of H&E stained SNUC specimen revealed areas of epithelial tumor as well as nontumor components including stroma (Supplementary Fig. S4A). The areas of tumor expressed epithelial markers (keratins and E-cadherin; Supplementary Fig. S4B and S4C, respectively) but did not express mesenchymal markers. On the other hand, the nontumor area expressed mesenchymal markers (α-SMA and vimentin, Supplementary Fig. S4D and S4E, respectively) but did not express epithelial markers. Because SNUC is a neuroendocrine tumor, we also studied the expression level of synaptophysin. Neither the original tumor nor the MDA8788 cell lines showed any synaptophysin staining (Fig. 4B), which is consistent with the immunohistochemical profile conducted on the original tumor (data not shown).

To determine whether the SNUC cell lines were tumorigenic in vivo, viable cells were injected into the flanks of nude mice with or without Matrigel. Both MDA8788-6 and MDA8688-7 cells formed tumors (Fig. 5A and Supplementary Fig. S5A, respectively), whereas Matrigel substantially enhanced the tumor growth statistically significantly by day 39 in MDA8788-6 and day 28 in MDA8788-7 (P < 0.005; Fig. 5B–D and Supplementary Fig. S5B and S5C, respectively). Histopathologic examination of the tumors revealed minimal differences in morphologic characteristics between tumors of mice injected with or without Matrigel (data not shown). As expected, the tumors grown from both MDA8788-6 and MDA8788-7 cells (Fig. 5E and Supplementary Fig. S5D, respectively) had a histopathologic appearance similar to that of the original surgical specimen (Fig. 1B), although the tumors from the MDA8788-7 line had a slightly more squamoid appearance than that of the MDA8788-6 line. Metastases were not identified in any of the mice at necropsy (data not shown).

To examine whether the SNUC cells represent clinical features, we injected MDA8788-6 cells into the soft palate of mice. The implanted cells developed visible tumors in nude mice and tumor invasion into the brain was observed by a T₁-weghted image (Fig. 6A and B). Staining with H&E showed local invasion into muscle (Fig. 6C), bone (Fig. 6D), nerve (Fig. 6E), blood vessels (Fig. 6F), and lymphatic vessels (Fig. 6G). The in vivo behavior accurately reflected the invasive nature of SNUC tumors in patients.

The aim of this study was to establish a reliable and phenotypically accurate tumor model system for SNUC, and we were successful in establishing and characterizing 2 novel isogenic SNUC cell lines. Malignant sinonasal tumors are clinically challenging because of their rarity, their proximity to vital structures, and their histologic variety within a complex anatomic region. These tumors comprise fewer than 3.6% of all malignancies seen in the head and neck (17). Among them, SNUC is particularly rare and aggressive. It was first described by Frierson and colleagues in 1986 (1), and fewer than 100 cases had been reported by 2005 (18). Despite multimodality therapy, including surgical resection, chemotherapy, and radiotherapy, the prognosis for patients with SNUC is poor; the median survival is less than 18 months (2, 3) and the 5-year survival rate is less than 20% (19).

The development and evaluation of new therapeutic approaches for SNUC has been limited by a lack of available models. To better understand and overcome the poor outcomes of patients diagnosed with this disease, establishment of a tumor model of SNUC is essential. Currently, there are few cell lines derived from sinonasal tumors; most of them were isolated from squamous cell carcinomas (9, 10) and one was established from an intestinal-type sinonasal adenocarcinoma (20). No stable SNUC cell lines had been reported before the development and characterization of the cell lines reported here. These new cell lines are designated MDA8788-6 and MDA8788-7, and they were derived from a T4N0M0 SNUC of the right maxillary sinus of a 74-year-old female who was diagnosed and treated at our center. STR genotyping analysis revealed the isogenic nature of the 2 MDA8788 cell lines and the parent tumor. The morphologic features of the cells were stable throughout cell culture. These results indicate that our newly established cell lines are representative of the disease, making them useful for preclinical studies of the biology and treatment of SNUC.

An analysis by Gil and colleagues showed that 30% of the SNUCs have an abnormal karyotype. In our study, 12 chromosomal translocations were identified in both MDA8788 cell lines. None of these translocations matched with previous cytogenetic studies of SNUCs (21, 22). It will be interesting to investigate the functional impact of the observed chromosomal translocations as studies in other tumor types have identified oncogenic activation of proto-oncogenes at chromosomal breakpoints. Further analysis of the observed cytogenetic abnormalities with mutational studies may yield potential molecular targets in this disease.

We found it interesting that despite the highly aggressive nature of SNUC in patients, the SNUC cell lines showed a less malignant phenotype in tumor models. For example, the doubling time of the cells was relatively slow, they did not form colonies in soft agar, and they were relatively sensitive to detachment. They did not migrate or invade in in vitro assays. On the other hand, the SNUC cells were highly tumorigenic in the mice. To determine the tumorigenic potential of the MDA8788 cell lines, the cells were heterotransplanted into nude mice. The cells formed solid tumors histologically identical to the original surgical specimen, indicating that our SNUC cell lines can be used as reliable reagents for in vivo preclinical studies. These findings indicate that the microenvironment of the tumor may play an important role for the development of SNUCs. Moreover, in vitro experiments may be less revealing, vis-à-vis potential therapeutic targets and, thus, well-designed animal studies may be required in this unique disease. It is known that SNUC presents as large tumors involving multiple sinonasal structures, often extends into the orbital or cranial cavity, and can metastasize to the cervical lymph nodes, lungs, bones, brain, and liver (2, 4–6). To determine the reliability of our SNUC cell lines in replicating phenotype of SNUC in patients, the cells were orthotopically implanted to mice. Histologic examination showed that developed tumors could invade into the nasal cavity and cranial cavity, indicating that these SNUC cell lines and the orthotopic model may work as powerful preclinical tools.

Immunoblotting assay revealed that MDA8788 cells were positive for pan-keratin, cytokeratin 8, and cytokeratin 19, which are positive in SNUC specimens at high rates (4, 8, 23–26). Other epithelial markers and mesenchymal markers were also tested. The SNUC cell lines were also found to express the epithelial markers E-cadherin and β-catenin, but they do not express the mesenchymal markers N-cadherin, vimentin, or α-SMA, indicating that the SNUC cell lines have epithelial characteristics and not mesenchymal features. Interestingly, our SNUC specimen expressed the epithelial markers at lower levels than the SNUC cell lines and was positive for the mesenchymal markers. This may be the result of contamination of non tumor cells, such as cancer-associated fibroblasts. Neither the surgical specimen nor the SNUC cells expressed synaptophysin, despite this tumor's neuroendocrine classification. Some researchers have already reported that SNUC can often be synaptophysin-negative consistent with our finding (4, 8, 23, 27).

In summary, this is the first report of establishing stable human-derived SNUC cell lines. The cell lines are highly tumorigenic and maintain the histologic and molecular features of the original tumor when grown in murine models. Therefore, these lines may serve as useful tools for future studies of SNUC tumor biology, as well as in the development and testing of novel therapies for this deadly disease.

No potential conflicts of interest were disclosed.

Conception and design: Y. Takahashi, M.E. Kupferman, M.J. Frederick, J.N. Myers, E.Y. Hanna

Development of methodology: Y. Takahashi, M.E. Kupferman, M. Zhao, J.N. Myers, E.Y. Hanna

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Takahashi, M.E. Kupferman, D. Bell, T. Jiffar, J.G. Lee, T.-X. Xie, N.-W. Li, A. Gelbard, J.N. Myers, E.Y. Hanna

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Takahashi, M.E. Kupferman, D. Bell, J.G. Lee, T.-X. Xie, N.-W. Li, A. Gelbard, J.N. Myers

Writing, review, and/or revision of the manuscript: Y. Takahashi, M.E. Kupferman, T. Jiffar, N.-W. Li, J.N. Myers, E.Y. Hanna

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.E. Kupferman, T. Jiffar, M. Zhao, J.N. Myers

Study supervision: M.E. Kupferman, J.N. Myers, E.Y. Hanna

The authors thank Drs. T.J. Liu and Yoshitsugu Mitani for providing technical support, Samar Jasser and Terri Astin for administrative assistance, and Lizzie Hess for her important editing of the manuscript.

This work was supported by the Pittsburgh Foundation Study of Sinonasal Malignancies, Various Donors Sinus Cancer Research Fund, The University of Texas MD Anderson Cancer Center institutional start-up funds (179475 and 179495), a National Cancer Institute-Cancer Center Core Grant (NCI CA16672), NIH Grant K08-DE019185 (M.E. Kupferman), Joint American-College of Surgeons – Triological Society Clinical Scientist Development Award (M.E. Kupferman), and Sheila Newar Cancer Research Fund (M.E. Kupferman).

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.