Cancer Research Infection and Cancer: Biology, Therapeutics, and Prevention Tumor Immunology: New Perspectives
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Research 66, 10399-10407, November 1, 2006. doi: 10.1158/0008-5472.CAN-06-1564
© 2006 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tian, J.
Right arrow Articles by Locker, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tian, J.
Right arrow Articles by Locker, J.

Cell, Tumor, and Stem Cell Biology

Loss of Nkx2.8 Deregulates Progenitor Cells in the Large Airways and Leads to Dysplasia

Jianmin Tian, Radma Mahmood, Robert Hnasko and Joseph Locker

Department of Pathology, Albert Einstein College of Medicine, Bronx, New York

Requests for reprints: Joseph Locker, Department of Pathology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. Phone: 718-430-3422; Fax: 718-430-3483; E-mail: locker{at}aecom.yu.edu.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Nkx2.8, a homeodomain transcription factor, has been characterized in liver cancer and in the developing central nervous system. We now show that this factor is also expressed in the lung, where it localizes in adults to a discrete population of tracheobronchial basal cells. To target the mouse gene, the first exon was replaced by a LacZ marker gene joined to the intact 5'-untranslated region. Marker expression was observed throughout the lower respiratory tract, beginning on E11 in a few cells of the distal lung buds. The region of expression then spread upward. By neonatal day 1, expression was greatest in the large airways and the Nkx2.8–/– mice exhibited generalized tracheobronchial hyperplasia. Bromodeoxyuridine (BrdUrd) labeling studies showed that a higher rate of bronchial cell proliferation persisted at 6 to 8 months. In adults, Nkx2.8 marker expression decreased with progressive differentiation into ciliated and secretory cells. The cell localizations and patterns of coexpression with BrdUrd and differentiation markers suggest a progenitor relationship: the cells that most strongly express Nkx2.8 seem to function as tracheobronchial stem cells. Moreover, Nkx2.8 acts to limit the number of these progenitor cells because the marker-expressing population was greatly expanded in Nkx2.8–/– mice. Increased proliferation and an altered progenitor relationship caused progressive bronchial pathology, which manifested as widespread dysplasia in the large airways of 1-year-old Nkx2.8–/– mice. (Cancer Res 2006; 66(21): 10399-407)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Nkx2.8 is part of a family of vertebrate developmental regulators that are homologues of the Drosophila homeodomain transcription factor, NK2. Most of the Nkx2 factors were initially characterized from studies of development. In contrast, Nkx2.8 was found in a study of human neoplastic gene regulation (1). An Nkx2.8 cDNA was cloned from a hepatocellular carcinoma cell line via its binding to an {alpha}-fetoprotein (AFP) gene upstream regulatory element (2, 3). AFP is normally expressed in fetal liver and frequently reexpressed in hepatocellular carcinoma, and thus the new factor seems to be an appropriate developmental regulator to explain this pattern of oncofetal gene expression. Several lines of experimentation supported this relationship. Transcripts were readily shown in human fetal liver, and significant levels of Nkx2.8 protein were present in human, rat, and mouse hepatocellular carcinoma cell lines that express AFP. Moreover, mechanistic studies in human hepatocellular carcinoma line HuH7 established that Nkx2.8 functions as a significant transcriptional activator that regulates AFP in vivo (1, 4). All of these observations, along with the relationship of Nkx2.8 to other NK2 factors, led us to expect that Nkx2.8 would have an important role in liver development. We planned to show this role by establishing an Nkx2.8-null mutant mouse with an inserted marker gene to localize the cells that normally express Nkx2.8.

Concurrent with our own efforts, Pabst et al. (5) cloned the mouse homologue gene, which they independently named Nkx2.9, and showed strong ventral expression in the developing central nervous system. This group recently characterized an Nkx2.9-null mouse (6) and noted late gestational changes in the central nervous system (CNS), including attenuation of the accessory nerve (XI) and an increase in the number of ventromedial motor column cells. Analysis of our mutant mouse also showed significant expression in the CNS, which will be described elsewhere.

Surprisingly, the Nkx2.8–/– mice had no liver abnormalities and Nkx2.8 expression was not detected in mouse liver at any developmental stage. However, examination of marker gene expression indicated a novel relationship of Nkx2.8 to tracheobronchial epithelium, which led us to characterize Nkx2.8 during lung and pharyngeal development. Moreover, loss of Nkx2.8 caused extensive precancerous changes in the large bronchi of adults. The data suggest that Nkx2.8 is most strongly expressed in cells that seem to function as tracheobronchial stem cells, with Nkx2.8 acting to limit this population. Targeted inactivation of the gene causes expansion of this cell population and widespread dysplasia.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Construction of the targeting vector. To construct the targeting vector, two regions of Nkx2.8 were amplified from genomic DNA of mouse strain 129SvJ using the Expand Long Template PCR System (Roche, Indianapolis, IN), according to the protocols of the manufacturer, and buffer 3, at an annealing temperature of 55°C. A 2,670-bp segment from bp –1,875 to +795, including the 5' untranslated region (UTR), exon 1, and part of the intron, was amplified with primers TCAAGATTGTCCATTTTGCCAGAGG and AAGAGTAATCGGCGAGCCTGAGATGATGCCC and cloned into plasmid pCRscript SK(+) (Stratagene, La Jolla, CA). The region from –1,875 to the ATG start site (+203) was excised with SacII and NcoI and cloned into the same sites in plasmid pHM4 (7).1 A 2,576-bp downstream segment from +484 to +3,166 was amplified with primers GAAGTCTGTGGTCCTCAAACAGACAGGT and CACTCGGCCAGCTACAAATCCCAAC and cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, CA). The region containing the upstream Nkx2.8 region and the entire targeting cassette was excised from the pHM4 construct as a single NotI fragment, which was then cloned into a NotI site in the plasmid containing the downstream segment. The resultant targeting vector contained a 2,078-bp upstream genomic segment with a perfect fusion of the promoter to the ATG start codon of a nuclear-localizing LacZ. Exon 1 and part of the intron were replaced by the targeting cassette (LacZ and Neo driven by the mouse phosphoglycerate kinase promoter), which was followed by a 2,808-bp genomic segment containing exon 2, the 3'-UTR, and an additional 1,320 bp of downstream DNA.

Generation and characterization of heterozygous-null embryonic stem cell lines. Embryonic stem cell line WW6 was grown as described (8). After electroporation with targeting vector linearized with NsiI, G-418-resistant clones were selected for analysis. Homologous recombination was detected by PCR and Southern blot (Supplementary Fig. S1). Insertion of the targeting cassette was detected by PCR with primer 1 from outside of the targeting vector (GGAGGTAGATCGGAGAGAGAGAATCTC) and primer 4 (GCCATGCGGGGGTCTTCTACCTTTCTC) from the LacZ gene. For Southern blot, 3 µg of genomic DNA were digested with HindIII, BamHI, or ApaI and resolved on an 0.6% agarose gel. The DNA was transferred to nitrocellulose and probed with a 1.5-kb cloned gene segment located 3' of the knockout construct using standard methods.

Generation of Nkx2.8-null mice and genotyping. We injected positive embryonic stem cell clones into blastocysts from C57BL/6 mice, which were implanted into 129SvJ females. Chimeric agouti mice were then bred with C57BL/6 mice. The offspring were genotyped and heterozygous-null mice were continuously bred into the C57BL/6 strain. Genotyping was carried out on DNA extracted from tail segments using a mixture of Nkx2.8 primer 2 (GGCTTCGCTCCTGCACTCATT), Nkx2.8 primer 3 (TGAAGCCGAGGCGTCCAGAGGTGGCCAT), and LacZ primer 4.

Transcript analysis. Nkx2.8 mRNA was detected using the SuperScript III One-Step RT-PCR kit (Invitrogen), 5 µg of total RNA per reaction, and primers GTGAGGCGCGAGCAACAGACG and CTTCAGCTTGTAGCGGTGGTTCTGG. cDNA synthesis for 30 minutes at 65°C was followed by touchdown PCR to a final annealing temperature of 60°C for a total of 40 cycles. LacZ and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA were detected by conventional oligo-dT-primed reverse transcription-PCR (RT-PCR) using CGAGCGATACACCGCATCCG and CCAGGAGTCGTCGCCACCAAT for LacZ and GGTCATCATCTCCGCCCCTTCTGC and GAGTGGGAGTTGCTGTTGAAGTCG for GAPDH.

Morphologic studies. For most studies, embryos and neonates were fixed whole for 16 hours in 10% buffered neutral formalin. Adult lungs were dissected out, gently inflated with formalin via the trachea, and then further fixed for 16 hours. For long-term cell labeling studies, bromodeoxyuridine (BrdUrd) was dissolved in drinking water at 0.5 mg/mL.

Tissues were embedded in paraffin and studied by routine histology and immunohistochemistry. For immunohistochemistry, slides were deparaffinized, hydrated, and incubated in a steamer for 20 minutes in 10 mmol/L sodium citrate (pH 6.0) for antigen retrieval. The following antibodies were diluted in blocking solution and usually incubated for 1 hour at room temperature: rabbit anti-ß-galactosidase (LacZ; Rockland, Gilbertsville, PA), 1:10,000; mouse monoclonal anti-TTF1 (Lab Vision, Fremont, CA), 1:200; mouse monoclonal anti-Ki67 (Novocastra, Newcastle upon Tyne, United Kingdom), 1:2,000; mouse monoclonal anti-BrdUrd (Roche), 1:100; goat polyclonal anti-CC10 (Santa Cruz Biotechnology, Santa Cruz, CA), 1:1,000; mouse monoclonal anti-ß-tubulin IV (BioGenex, San Ramon, CA), 1:500; and a monoclonal anti-pancytokeratin mixture (Sigma, St. Louis, MO), 1:300. For more sensitive but less quantitative detection of LacZ, the primary antibody was incubated at 4°C for 16 hours. Secondary peroxidase-3,3'-diaminobenzidine (sometimes with nickel intensification) or alkaline phosphatase-new fuchsin detections used mouse or rabbit Vectastain ABC or M.O.M. kits (Vector Laboratories, Burlingame, CA). Double-label immunofluorescence studies of paraffin sections used the same antibody concentrations, detection with Cy2-antirabbit and goat TRITC-antimouse secondary antibodies, and 4',6-diamidino-2-phenylindole counterstain. Sudan Black B was used to block autofluorescence (9).

For detection of ß-galactosidase activity, embryos were dissected free of extraembryonic membranes and fixed in 1% formaldehyde, 0.2% glutaraldehyde, 0.02% Triton X-100 in PBS, for 30 minutes at 4°C. Embryos were washed and stained for 40 hours in 1 mg/mL 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside (X-Gal), 5 mmol/L K3Fe(CN)6, 5 mmol/L K4Fe(CN)6, and 2 mmol/L MgCl2, all dissolved in PBS. Fixed embryos were photographed and submitted for routine paraffin embedding and histologic sectioning.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Characterization of gene targeting. The Nkx2.8 gene was targeted in embryonic stem cells by homologous recombination with a plasmid construct. In this targeting vector, the first exon was replaced by a LacZ-NLS gene, which was joined at its ATG start codon to the intact 5-UTR (Supplementary Fig. S1). Following blastocyst injection, chimeras were mated with C57BL/6 mice and heterozygous offspring were identified by genotyping. These were progressively bred into both C57BL/6 and 129SvJ backgrounds. Genotype and correct gene targeting were verified by genomic Southern blot hybridization and PCR analysis. Heterozygous matings subsequently produced offspring with a normal Mendelian ratio of genotypes, which were morphologically indistinguishable. Representative genotypes compiled from seven litters were 10 (23%) +/+, 23 (53%) +/–, and 10 (23%) –/–. The –/– mice were fertile and generated litters of five to seven pups.

Expression of intact and targeted Nkx2.8 alleles. Embryonic gene expression was characterized by X-Gal staining. At 11.5 days of gestation (not illustrated), whole embryos showed staining of ventral brain and spinal cord in a pattern previously described in both transient transgenic mice (10) and an Nkx2.9-null mouse (6). To study day 12, we extended staining times and again showed staining of ventral CNS structures. In addition, the lungs also stained with X-Gal and were visible as triangular regions above the red liver (Fig. 1A ). In histologic sections (Fig. 1B), X-Gal staining was apparent in the most distal bronchial tubules, but only in the inferior lung segments. Staining was stronger in –/– than in +/– embryos, but the expression patterns were similar. At this age, strong X-Gal staining was also apparent in the lateral pharynx and thyroid, and the thymus was weakly positive (not illustrated).


Figure 1
View larger version (59K):
[in this window]
[in a new window]
  		Figure 1. Expression in the null mouse. A, expression of the inserted LacZ gene is shown by X-Gal staining in E 12.5 –/– and +/– embryo littermates. In both genotypes, the lungs (arrows) appear as triangular regions above the red liver. There is also staining throughout the ventral brain and spinal cord. B, frontal sections of X-Gal-stained –/– and +/– embryos. Note the staining in inferior distal bronchiolar tubules and the absence of staining in the liver. T, trachea; Br, main stem bronchi; LL, left lung; RL1, RL2, and RL3, three lobes of the right lung; Liv, liver. Whole embryo was stained with X-Gal, paraffin embedded and sectioned, stained with nuclear fast red and orange G, and photographed at x4 magnification. C, RT-PCR characterizations of Nkx2.8 and LacZ expression. RNA was extracted from total lung and total brain from day l neonates and from livers of day 16 embryos. RNA from Hepa1-6 cells was used as a positive control. A 303-bp product of Nkx2.8 mRNA is apparent in lung and brain from wild-type and heterozygous-null mice, and a 380-bp LacZ product is apparent only in mice with a targeted allele. Note that two of the brain preparations also show an artifactual band at a higher molecular weight. All RNA preparations showed positive detection of a 498-bp band amplified from GAPDH mRNA. Marker lanes show a {Phi}X174-HaeIII digest.

 
RT-PCR showed appropriate expression of both the intact and targeted alleles at several gestational ages. A representative study of lung and brain RNA from 1-day neonates is presented in Fig. 1C. Neither positive staining for LacZ nor Nkx2.8 mRNA was detected in liver at any developmental stage, although Nkx2.8 mRNA was readily detected in Hepa1-6 cells, which have a phenotype like fetal liver and were used as a positive control. The illustrated liver study used RNA from 16 days of gestation, the time of peak AFP expression. Additional RT-PCR studies showed Nkx2.8 mRNA in adult lung and brain, but not in liver (not illustrated).

Lung development. The expression of Nkx2.8 was characterized throughout development and compared with TTF1/Nkx2.1, a closely related factor and known regulator of lung development (Fig. 2 ; refs. 11, 12). For these detailed studies, we used conventional fixation, paraffin sections, and immunoperoxidase detection of the nuclear-localized LacZ protein. In 11-day embryos, TTF1 was expressed uniformly throughout the trachea and lung buds. Lung staining was present in all nuclei of the multilayered epithelium. LacZ-Nkx2.8 staining was detectable in a few scattered cells in the lung bud, mostly on the basement membrane. By day 12, TTF1 was generally reduced, with the weakest staining in the upper trachea. The number of Lac-Nkx2.8-positive cells was increased and, although still peripheral, had spread upward into larger bronchial branches. Staining of heterozygotes was weaker and we did not show LacZ staining in 11-day +/– embryos. At day 12, however, except for more intense staining in –/– mice, there were no clear differences between +/– and –/– embryos.


Figure 2
View larger version (41K):
[in this window]
[in a new window]
  		Figure 2. Relationship of Nkx2.8 and TTF1 to embryonic lung development. A, consecutive parasagittal sections from an E11.5 –/– embryo. B, consecutive parasagittal sections from an E12.5 –/– embryo. C, parasagittal sections from an E12.5 +/– littermate of the embryo in (B). Ph, pharynx; Thy, thyroid; A, aorta; Tr, trachea; LB, lung bud; Liv, liver. Top, x4 magnification; bottom, x20 magnification. All photographs are oriented with anterior to the left and superior to the top.

 
In 1-day neonates (Fig. 3 ), TTF1 staining was greatly reduced but could still be shown in small bronchi and bronchioles; epithelial cells were uniformly stained. LacZ-Nkx2.8-positive cells, never a uniform contiguous cell population, were most prominent in the large bronchi and trachea. Thus, appearance and expansion of Nkx2.8-expressing cells occurred after formation of the lungs and had opposite polarity from the expression of TTF1. The latter protein appears proximally in the pharynx at the very beginning of trachea and lung formation and the zone of expression moves distally during development. Nkx2.8 appears distally at a later developmental stage and the main zone of expression moves proximally into the large airways (Fig. 4C ). Moreover, TTF1 was expressed in continuous cell fields whereas Nkx2.8-positive cells were scattered.


Figure 3
View larger version (73K):
[in this window]
[in a new window]
  		Figure 3. Nkx2.8 in the 1-day neonatal respiratory tract. A, neonatal lung from an Nkx2.8+/– mouse. Top, three right lung lobes and main stem bronchus photographed at x4 magnification; bottom, pulmonary zones stained for LacZ-Nkx2.8 or TTF1, photographed at x40 magnification. Alv, alveoli; P bronch, peripheral brionchioles; M bronch, medium bronchi; C bronchi, central bronchi. Immunoperoxidase localization of LacZ with hematoxylin counterstain or photographically amplified immunoalkaline phosphatase localization of TTF1. B, matched tracheal sections in +/– and –/– neonates (x40). C, model for the relationship between Nkx2.8 and TTF1 in lung development.

 

Figure 4
View larger version (103K):
[in this window]
[in a new window]
  		Figure 4. Altered cell populations in the bronchi of 12-month-old adults. A, matched sections of extrapulmonary bronchi in 12-month-old +/– and –/– littermates. B, cellular detail of the bronchi in (A). Arrows in the –/– figure show densely stained LacZ-positive cells that are comparable to those observed in the +/– figure. C, smaller bronchi in the same animals. Arrows show two Nkx2.8-LacZ-positive basal cells in the +/– figure. The example from the –/– figure shows a dysplastic lesion that extends through the entire photograph. Immunoperoxidase detections of LacZ with hematoxylin counterstain, photographed at x20 (A) or x40 (B and C) magnification.

 
In the large airways, there were clear morphologic differences between the +/– and –/– genotypes (Fig. 4B). The tracheobronchial epithelium in the null mouse was taller and more crowded, with greater numbers of LacZ-positive and proliferating Ki67-positive cells. Mouse large-airway epithelium has a very low proliferation fraction from late gestation onward (13), and thus the degree of proliferation in the Nkx2.8-null neonate indicates significant hyperplasia.

Examination of adult lungs. Twelve-month-old adults (generation F2) showed clear differences between lungs of homozygous-null and heterozygous littermates (Fig. 4). The heterozygote lungs were indistinguishable from +/+ lungs (not illustrated) and LacZ expression was limited to a small number of epithelial cells in the large airways. Marker expression was most prominent in the extrapulmonary bronchi as single small basal cells with densely stained nuclei beneath the columnar bronchial epithelium. Very few positive cells were observed in more distal bronchi.

The homozygous-null mice had a substantial increase in the number of positive cells and significant pathologic changes compared with the heterozygotes. In the extrapulmonary segments of the main bronchi, groups of positive cells were observed in different positions within the epithelium. A few basal cells had densely staining nuclei but most positive cells had more open nuclei. The increase in LacZ-positive cells was associated with disruption of the orderly pattern of columnar epithelium observed in the heterozygote and wild type. More proximal bronchi showed irregular thickening and many more LacZ-positive cells throughout (not illustrated). Moreover, there were discrete dysplastic lesions in which these changes were more pronounced (Fig. 4C). In these lesions, the epithelium was thickened to five or more layers, the cells had lost their normal basement membrane-luminal polarity, and many were LacZ positive with variable staining intensity.

In addition to the changes in central airways, LacZ immunostaining was identified in peripheral lung as scattered positive cells in terminal bronchioles and in some type II pneumocytes (see below). This staining was weaker than in peripheral lung and there were no clear differences between homozygous-null and heterozygous-null mice.

LacZ immunostaining also showed discrete minor cell populations in several other parapharyngeal tissues: thyroid, parathyroid, thymus, tracheal epithelium, and tracheobronchial submucosal glands. The null mouse had increased numbers of positive cells in all of these tissues (not illustrated).

Altered bronchial cell proliferation in adult null mice. Because proliferation was increased in neonates, further studies were carried out to see if these differences persisted in mature adults of a homogeneous genetic background (i.e., generation N7 in C57BL/6; Fig. 5 ). In adults, however, bronchial epithelium has a low proliferation rate, and staining with Ki67 or proliferating cell nuclear antigen markers shows too few positive cells to allow meaningful analysis. Mice were therefore subjected to 1 week of continuous DNA labeling with BrdUrd in their drinking water to cumulate labeled nuclei (14). Three sets (+/+, +/–, and –/– genotypes) of littermates, 6 to 8 months old, were analyzed. The morphologic differences between the +/– and –/– mice were much more subtle than in 1-year-old lungs (Fig. 4) but became more extreme with aging. In heterozygotes, the number of LacZ-positive cells decreased with aging and the bronchial epithelium retained its orderly cell architecture, whereas in homozygotes, the number of LacZ positive cells increased, the epithelium became more irregular throughout, and there were frequent dysplastic lesions. These progressive differences were observed in both hybrid C57BL/6-129SvJ (F2) and inbred C57BL/6 (N7) backgrounds.


Figure 5
View larger version (71K):
[in this window]
[in a new window]
  		Figure 5. Relationship between bronchial cell proliferation and Nkx2.8 genotype in adult mice. A, matched sections of large bronchi of Nkx2.8+/– and Nkx2.8–/– 6-month-old littermates. B, terminal bronchioles in the same sections. Immunoperoxidase staining of LacZ and BrdUrd with hematoxylin counterstain, photographed at x40 magnification. C and D, counts of LacZ- and BrdUrd-positive cells from main bronchi proximal to the first bifurcation (Central) and from terminal bronchioles (Peripheral). Counts were made from three sets of male littermates ages 6 to 8 months. More than 20 high-power fields were counted from central and peripheral zones of each animal. Columns, mean of three mice; bars, SD. P values were of Student's t test comparisons.

 
Examination of standard LacZ staining in the heterozygotes showed a progressive decrease in the number of positive cells from central to peripheral bronchi. However, with more sensitive immunoperoxidase detection, separate central and peripheral zones of positive staining were apparent, with fewer positive cells in the intermediate bronchi between the two zones (not illustrated). The positive cells in the terminal bronchioles were morphologically different from those in central bronchi, and occasional type 2–like cells were also positive in the alveoli (Fig. 5B). With more sensitive staining, there was less distinction between cells that expressed different levels of Nkx2.8 but clearer discrimination of positive from negative cells. The two zones of positive cells overlapped more in the homozygous-null mouse but sequential counting showed the same general distribution: higher counts in the main bronchi and terminal bronchioles and lower counts in intermediate bronchi (not illustrated).

In the 6-month-old mice, the cell patterns in large airways suggested stem/progenitor cell relationships. In heterozygotes, the range of staining intensities suggested a gradual transition of staining intensity from densely stained basal cells to unstained columnar cells, and columnar cells had a clear polar orientation from basement membrane to lumen (Fig. 5A). The null mouse showed a greater number of positive cells with less clear polarity.

Proliferating and LacZ-positive cells showed a similar distribution in the bronchial tree, although the terminal bronchi had the highest proportion of proliferating cells. For uniform comparison, counts were compared in two areas: central bronchi, defined as the main bronchi to the first intrapulmonary bifurcations, and terminal bronchioles (Fig. 5C and D). The central bronchial cells were significantly altered in the null mouse but the peripheral bronchiolar cells were comparable to those in the heterozygotes. Thus, Nkx2.8 is expressed in two anatomically distinct cell compartments in large bronchi and terminal bronchioles, but only the former is deregulated in the null mouse.

Stem cell relationships. Double-marker immunohistochemical studies were carried out to substantiate that the LacZ-expressing cells were indeed stem cells by showing two essential properties: self-renewal and multipotency (15). We therefore investigated the proliferation of specific types of LacZ-positive cells by colocalization of LacZ and BrdUrd. A second group of analysis characterized the relationship of LacZ to specific markers of bronchial differentiation.

Because the overall rate of cell proliferation was low, the long-term labeling with BrdUrd showed mostly daughter-cell pairs. Colocalization studies of 6-month-old animals showed that most bronchial cell proliferation was in LacZ-negative, differentiated epithelial cells. However, some LacZ-positive cell pairs were also BrdUrd positive. Two representative pairs suggest different aspects of stem cell regulation (Fig. 6 ). An asymmetrical daughter-cell pair is consistent with downstream maturation (Fig. 6A). In this example, the paired nuclei are arranged on an axis that is perpendicular to the basement membrane; the more densely LacZ-positive nucleus is adjacent to the basement membrane and the nucleus with weaker staining is suprabasal. In contrast, a symmetrical pair of densely-staining basal cells lies on an axis that is parallel to the basement membrane (Fig. 6B). The increased proliferation and abundance of LacZ-positive cells in the Nkx2.8–/– provided more examples of these cell relationships, but similar pairs were also observed in Nkx2.8+/– lungs (not illustrated).


Figure 6
View larger version (68K):
[in this window]
[in a new window]
  		Figure 6. Cell proliferation and differentiation markers. Studies are on consecutive sections from the 6-month-old Nkx2.8–/– and Nkx2.8+/– lungs illustrated in Fig. 5. A and B, 1-week feeding of BrdUrd predominantly labeled daughter-cell pairs. Note that, by immunofluorescence, the LacZ protein was mostly restricted to the central region of the nucleus whereas BrdUrd localized to a larger region that included the peripheral chromatin. A, a LacZ-positive asymmetrical daughter-cell pair. The more densely LacZ-positive nucleus is near the basement membrane and the more weakly staining nucleus is parabasal. B, a LacZ-positive pair in which both daughter cells have identical morphology and are aligned on the basement membrane. Magnification, x40. C, double immunostaining with markers of bronchial differentiation. LacZ and pancytokeratin: Tangential section of a large bronchus shows that cells on the basement membrane with intensely LacZ-staining of nuclei (black arrows) lack keratinization whereas some keratinized cells show weaker LacZ staining (white arrows). LacZ and ß-tubulin IV: Staining shows weak LacZ-positive nuclei in tubulin-positive ciliated cells (arrows). LacZ and CC10: Staining show positive nuclei in mature Clara cells. LacZ and p63 in a +/– heterozygote: The markers colocalize in cells near the basement membrane where the majority of LacZ-positive cells are also p63 positive. LacZ and p63 in a –/– homozygote: the most basilar region is LacZ positive and p63 negative, and these singly positive cells have displaced the double-labeled nuclei to a parabasal location. In all micrographs, LacZ was detected by immunoalkaline phosphatase using a new fuchsin substrate (red), and the second antibody was detected with immunoperoxidase using a nickel-intensified 3,3'-diaminobenzidine reaction (black). Except for p63, patterns of coexpression were similar in Nkx2.8+/– and Nkx2.8–/– lungs. Magnification, x40.

 
The mature bronchial epithelium of the mouse consists predominantly of ciliated cells, Clara cells, and basal cells. The differentiation of LacZ-positive cells was evaluated by localization with cytokeratin, ciliated cell marker ß-tubulin IV, Clara cell marker CC10, and basal cell marker p63.

A pancytokeratin antibody mixture gave uniform peripheral staining of Clara and ciliated cells. Strongly LacZ-positive cells on the basement membrane were negative for keratin, but some keratinized cells were weakly LacZ positive (Fig. 6C). ß-Tubulin IV was localized to the cilia and the upper cytoplasmic regions of mature ciliated cells. Some of these cells also showed weak LacZ-positive staining. CC10 was localized to the apical zone of Clara cells, some of which were also positive for LacZ. ß-tubulin IV/LacZ and CC10/LacZ double labeling was also observed in some immature cells that did not project to the bronchial lumen (not illustrated). Although single representative examples are illustrated, the relationships of keratin, ß-tubulin IV, and CC10 were qualitatively similar in Nkx2.8+/– and Nkx2.8–/– bronchi. The patterns of colabeling were all consistent with a stem cell relationship in which LacZ-Nkx2.8 expression was inverse to the degree of maturation. Strong Nkx2.8 expression was observed in a specialized population of small cells on the basement membrane whereas weaker expression was apparent in both ciliated and Clara cell lineages.

The relationship between p63 and LacZ-Nkx2.8 was more complex (Fig. 6C, bottom). p63 is characteristically expressed in tracheobronchial basal cells. However, in adult mice, basal cells show the strongest staining but there is weaker staining of most other epithelial nuclei (16), similar to the pattern that we observed. In Nkx2.8+/–, strong staining of LacZ and p63 colocalized in most basal cells near the lung hilum. Other basal cells, however, were positive only for p63, particularly in more peripheral bronchi. The single- and double-positive basal cells were otherwise indistinguishable as small cells with scant cytoplasm and a dense nucleus, in contact with the basement membrane but not the bronchial lumen. LacZ+p63– basal cells were very infrequent. In contrast to the heterozygotes, Nkx2.8–/– mice had numerous LacZ+p63– basal cells, and in many areas these single positive cells had displaced the LacZ+p63+ upward into a parabasal location. This abnormal pattern apparently represents an early stage of dysplasia.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Bronchial progenitor cells. The tracheobronchial epithelium of the mouse is complex. It consists of ciliated, Clara (secretory), neuroendocrine, and basal cells. The latter are short cells that contact the basement membrane but do not extend to the lumen, and comprise ~4% of the cells at the mouse tracheal bifurcation (17). In old heterozygous mice, we observed selective LacZ staining of basal cells in the trachea and main bronchi, predominantly outside of the lung (Fig. 3). In younger heterozygotes, similar cells were stained, along with additional cells that showed a transition to columnar epithelium (Figs. 5 and 6). These cells have previously been reported to function as tracheobronchial stem cells (18, 19), although "stem cell" best describes a range of cell types. In the small intestine, for example, Marshman et al. (15) have proposed a three-level hierarchy with gradual transitions from actual steady-state stem cells to potential clonogenic stem cells to dividing transit cells over a sequence of five to six cell divisions. The variety of LacZ-positive cells in the bronchus suggests a similar hierarchy although bronchial cells have an exceptionally low rate of proliferation (13, 20).

Similar cell relationships have been shown in terminal bronchioles (21–23), and our studies suggest that the LacZ-positive cells in this region are part of a morphologically distinct and anatomically separate compartment of bronchioloalveolar stem cells. Moreover, the Nkx2.8-null mice show that regulation of cell proliferation differs between the tracheobronchial stem cell and bronchioloalveolar stem cell compartments. Although Nkx2.8 is expressed in both, the loss of this protein changes the dynamics of proliferation and maturation in the large proximal airways, but not the terminal bronchioles. Thus, Nkx2.8 seems to have a unique role in large airways as a negative regulator that limits the number of tracheobronchial stem cells.

Establishment of these stem cell populations occurs late in development, after the lungs have formed, and reflects a reciprocal relationship between Nkx2.8 and TTF1/Nkx2.1 (Fig. 4D). Both factors have identical DNA binding (1, 4) and we have found that human Nkx2.8 has a selective repressing function in some cells where TTF1 is a transcriptional activator. The repressing function is mediated by a distinctive COOH-terminal domain that is 96% conserved between human and mouse Nkx2.8.2 Because TTF1 can autoregulate its own transcription (24), Nkx2.8 might down-regulate expression of TTF1 as well as antagonize its transcriptional activation of target genes. In the embryo, TTF1 expression precedes the formation of the lungs, and the zone of expression gradually moves downward from the pharynx into the most peripheral lungs. Expression of TTF1 is uniform in continuous epithelial layers. In contrast, Nkx2.8 appears later in the most distal lung buds, always as a limited noncontiguous cell population. These cells gradually shift upward and eventually populate the large airways, although a residual population of positive cells remains in the terminal bronchioles.

Nkx2.8 expression. The Nkx2.8-LacZ gene fusion has provided a very sensitive marker for minor cell populations in the lung, and we confirmed expression of Nkx2.8 with RT-PCR. We also found that expression in other tissues, including thyroid, thymus, lung, and parathyroid, always associated with minor cell populations. Previously, Nkx2.8 has been infrequently detected for two distinct reasons. First, the expression occurs in a small fraction of cells in tissue samples. Second, the gene has an exceptionally high G + C content and is a very difficult template for standard RT-PCR conditions, causing us to adapt experimental conditions to facilitate detection (see Materials and Methods).

As illustrated, the central and peripheral LacZ-positive cells seem to have similar levels of Nkx2.8 expression. With less sensitive detection, however, the cells in the large airways were still obvious but the peripheral cells were no longer visible, suggesting a lower level of Nkx2.8 expression in the latter. The LacZ-expressing Nkx2.8-null mice thus have significant utility. The Nkx2.8+/– heterozygotes mark specific progenitor cell populations and can be used to study the normal proliferation and injury responses in both large and small airways. The Nkx2.8–/– homozygotes have distinct pathologic changes apparently due to deregulation of tracheobronchial stem cells.

Cancer relationships. The studies began as an investigation of Nkx2.8 in hepatocellular carcinoma, where the transcription factor is a significant positive regulator. However, the absence of Nkx2.8 from mouse fetal liver indicates that its expression in hepatocellular carcinoma does not directly recapitulate liver development but is instead a characteristic abnormality of liver oncogenesis. Presumably, an unidentified but related transcriptional activator has an equivalent role in fetal liver.

In contrast to an oncogenic relationship in hepatocellular carcinoma, the null mouse model relates Nkx2.8 to suppression of lung cancer. The null mice had bronchial epithelial hyperplasia from birth, which eventually led to dysplasia in large airways. The numerous lesions showed increased numbers of epithelial cells, loss of polarity, and accumulation of differentiated cells that lacked an association with the basement membrane. The progressive changes caused by loss of Nkx2.8 are analogous to changes in the prostate gland caused by loss of Nkx3.1 (reviewed in ref. 25). Nkx3.1-null mice develop early prostate hyperplasia that gradually progresses to intraepithelial neoplasia at 1 to 2 years of age. The null mice do not develop invasive cancer but Nkx3.1 is considered a tumor suppressor because its absence potentiates spontaneous carcinogenesis in Pten+/– mouse. Surprisingly, Nkx3.1 is reexpressed in prostate cancer cells where it no longer has an inhibitory function (26–28). Thus, Nkx3.1 suppresses precancerous changes, not cancer cells, and this suppression has been successfully modeled only in intact mice. Such experimental limitations may also apply to modeling of the suppressive function of Nkx2.8.

In humans, lung cancer and bronchogenic carcinoma are synonymous because most cancers occur in the large bronchi (29, 30). These tumors may be adenocarcinoma, squamous carcinoma, large-cell anaplastic carcinoma, or small-cell carcinoma, and are preceded by dysplasias that have the same bronchial distribution (31, 32). Tumors that arise from terminal bronchioles or alveoli (i.e., bronchioloaveolar carcinomas) are relatively uncommon and are a histologically and biologically distinct form of adenocarcinoma (33).

In mouse lung, carcinogenesis is almost entirely restricted to the peripheral region. The vast majority of spontaneous and carcinogen-induced cancers arise in terminal bronchioles or alveoli and correspond to human bronchioloalveolar carcinoma (34–39). Indeed, bronchial cancer is virtually nonexistent in the mouse (39). Recent studies have shown that these peripheral mouse lung tumors result from transformation of bronchioloalveolar stem cells (22, 40). Although a similar link between tracheobronchial stem cells and bronchial lung cancer might be expected, our studies of Nkx2.8 seem to be the first substantial evidence of such a relationship.

In this article, the Results section focused on changes that lead to dysplasia in 1-year-old mice because our detailed analysis in early (F2) generations was confirmed up to this age in fully inbred mice. However, we also studied a limited number of older mice from the early F2 breeding. These findings must be considered preliminary because they were observed on a variable genetic background in mice that were only two generations removed from the genetic manipulations of gene targeting. Nevertheless, these preliminary findings are appended because they suggest that the dysplasia in the Nkx2.8–/– genotype progresses to invasive cancer. Mice of ages 18 to 22 months developed spontaneous lung cancer, with an incidence of 5/7 in –/–, 2/4 in +/–, and 1/4 in +/+ littermates (see Supplementary Table S1 and Supplementary Fig. S2). In addition, older null mice generally showed more advanced bronchial dysplasia than was apparent at 1 year. This high cancer incidence is not characteristic of either parental mouse strain. Spontaneous lung cancer is uncommon in aged C57BL/6 mice (41) and 129SvJ is also reported to have a low susceptibility to lung cancer (42).

The full relationship of Nkx2.8 to cancer will require more definitive studies of spontaneous and induced carcinogenesis in fully inbred mice, and the process may be accelerated within a more cancer-sensitive genetic background. At this time, however, the studies establish several intriguing relationships. Nkx2.8-expressing cells appear late in development, after the lungs have formed, and show a retrograde expansion from distal lung tubules to the large airways. In the mature lung, Nkx2.8 is a negative regulator of epithelial proliferation in the large airways and may act by limiting the expansion of tracheobronchial stem cells. A separate Nkx2.8-positive cell population is present in terminal bronchioles. Finally, the absence of Nkx2.8 leads to early bronchial hyperplasia, which gradually progresses to dysplasia.


   Acknowledgments
 
Grant support: NIH grants CA68440, CA76354, CA104292.

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.

We thank Bernice Morrow, Winfried Edelman, and Len Augenlicht for discussions and comments on the manuscript.


   Footnotes
 
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

1 K. Kaestner, University of Pennsylvania, personal communication. Back

2 Y. Kajiyama and J. Locker, unpublished results. Back

Received 5/ 1/06. Revised 8/24/06. Accepted 9/ 1/06.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Apergis GA, Crawford N, Ghosh D, et al. A novel nk-2-related transcription factor associated with human fetal liver and hepatocellular carcinoma. J Biol Chem 1998;273:2917–25.[Abstract/Free Full Text]
  2. Wen P, Locker J. A novel hepatocytic transcription factor that binds the {alpha}-fetoprotein promoter-linked coupling element. Mol Cell Biol 1994;14:6616–26.[Abstract/Free Full Text]
  3. Wen P, Crawford N, Locker J. A promoter-linked coupling region required for stimulation of {alpha}-fetoprotein transcription by distant enhancers. Nucleic Acids Res 1993;21:1911–8.[Abstract/Free Full Text]
  4. Kajiyama Y, Tian J, Locker J. Regulation of {alpha}-fetoprotein gene expression by Nkx2.8. Mol Cell Biol 2002;22:6122–30.[Abstract/Free Full Text]
  5. Pabst O, Herbrand H, Arnold HH. Nkx2-9 is a novel homeobox transcription factor which demarcates ventral domains in the developing mouse CNS. Mech Dev 1998;73:85–93.[CrossRef][Medline]
  6. Pabst O, Rummelies J, Winter B, Arnold HH. Targeted disruption of the homeobox gene Nkx2.9 reveals a role in development of the spinal accessory nerve. Development 2003;130:1193–202.[Abstract/Free Full Text]
  7. Kaestner KH, Montoliu L, Kern H, Thulke M, Schutz G. Universal ß-galactosidase cloning vectors for promoter analysis and gene targeting. Gene 1994;148:67–70.[CrossRef][Medline]
  8. Ioffe E, Liu Y, Bhaumik M, Poirier F, Factor SM, Stanley P. WW6: an embryonic stem cell line with an inert genetic marker that can be traced in chimeras. Proc Natl Acad Sci U S A 1995;92:7357–61.[Abstract/Free Full Text]
  9. Baschong W, Suetterlin R, Laeng RH. Control of autofluorescence of archival formaldehyde-fixed, paraffin-embedded tissue in confocal laser scanning microscopy (CLSM). J Histochem Cytochem 2001;49:1565–72.[Abstract/Free Full Text]
  10. Santagati F, Abe K, Schmidt V, et al. Identification of Cis-regulatory elements in the mouse Pax9/Nkx2-9 genomic region: implication for evolutionary conserved synteny. Genetics 2003;165:235–42.[Abstract/Free Full Text]
  11. Yuan B, Li C, Kimura S, Engelhardt RT, Smith BR, Minoo P. Inhibition of distal lung morphogenesis in Nkx2.1(–/–) embryos. Dev Dyn 2000;217:180–90.[CrossRef][Medline]
  12. Wert SE, Dey CR, Blair PA, Kimura S, Whitsett JA. Increased expression of thyroid transcription factor-1 (TTF-1) in respiratory epithelial cells inhibits alveolarization and causes pulmonary inflammation. Dev Biol 2002;242:75–87.[CrossRef][Medline]
  13. Kauffman SL. Kinetics of pulmonary epithelial proliferation during prenatal growth of the mouse lung. Anat Rec 1975;183:393–403.[CrossRef][Medline]
  14. Ledda-Columbano GM, Pibiri M, Loi R, Perra A, Shinozuka H, Columbano A. Early increase in cyclin-D1 expression and accelerated entry of mouse hepatocytes into S phase after administration of the mitogen 1, 4-bis[2-(3,5-dichloropyridyloxy)] benzene. Am J Pathol 2000;156:91–7.[Abstract/Free Full Text]
  15. Marshman E, Booth C, Potten CS. The intestinal epithelial stem cell. Bioessays 2002;24:91–8.[CrossRef][Medline]
  16. Daniely Y, Liao G, Dixon D, et al. Critical role of p63 in the development of a normal esophageal and tracheobronchial epithelium. Am J Physiol Cell Physiol 2004;287:C171–81.[Abstract/Free Full Text]
  17. Pack RJ, Al-Ugaily LH, Morris G, Widdicombe JG. The distribution and structure of cells in the tracheal epithelium of the mouse. Cell Tissue Res 1980;208:65–84.[Medline]
  18. Boers JE, Ambergen AW, Thunnissen FB. Number and proliferation of basal and parabasal cells in normal human airway epithelium. Am J Respir Crit Care Med 1998;157:2000–6.[Medline]
  19. Otto WR. Lung epithelial stem cells. J Pathol 2002;197:527–35.[CrossRef][Medline]
  20. Kauffman SL. Cell proliferation in the mammalian lung. Int Rev Exp Pathol 1980;22:131–91.[Medline]
  21. Perl AK, Wert SE, Nagy A, Lobe CG, Whitsett JA. Early restriction of peripheral and proximal cell lineages during formation of the lung. Proc Natl Acad Sci U S A 2002;99:10482–7.[Abstract/Free Full Text]
  22. Kim CF, Jackson EL, Woolfenden AE, et al. Identification of bronchioalveolar stem cells in normal lung and lung cancer. Cell 2005;121:823–35.[CrossRef][Medline]
  23. Hong KU, Reynolds SD, Watkins S, Fuchs E, Stripp BR. Basal cells are a multipotent progenitor capable of renewing the bronchial epithelium. Am J Pathol 2004;164:577–88.[Abstract/Free Full Text]
  24. Nakazato M, Endo T, Saito T, Harii N, Onaya T. Transcription of the thyroid transcription factor-1 (TTF-1) gene from a newly defined start site: positive regulation by TTF-1 in the thyroid. Biochem Biophys Res Commun 1997;238:748–52.[CrossRef][Medline]
  25. Shen MM, Abate-Shen C. Roles of the Nkx3.1 homeobox gene in prostate organogenesis and carcinogenesis. Dev Dyn 2003;228:767–78.[CrossRef][Medline]
  26. Li X, Guan B, Maghami S, Bieberich CJ. NKX3.1 is regulated by protein kinase CK2 in prostate tumor cells. Mol Cell Biol 2006;26:3008–17.[Abstract/Free Full Text]
  27. Ornstein DK, Cinquanta M, Weiler S, et al. Expression studies and mutational analysis of the androgen regulated homeobox gene NKX3.1 in benign and malignant prostate epithelium. J Urol 2001;165:1329–34.[CrossRef][Medline]
  28. Xu LL, Srikantan V, Sesterhenn IA, et al. Expression profile of an androgen regulated prostate specific homeobox gene NKX3.1 in primary prostate cancer. J Urol 2000;163:972–9.[CrossRef][Medline]
  29. Buccheri G, Barberis P, Delfino MS. Diagnostic, morphologic, and histopathologic correlates in bronchogenic carcinoma. A review of 1,045 bronchoscopic examinations. Chest 1991;99:809–14.[Medline]
  30. Paris C, Benichou J, Saunier F, et al. Smoking status, occupational asbestos exposure and bronchial location of lung cancer. Lung Cancer 2003;40:17–24.[CrossRef][Medline]
  31. Bota S, Auliac JB, Paris C, et al. Follow-up of bronchial precancerous lesions and carcinoma in situ using fluorescence endoscopy. Am J Respir Crit Care Med 2001;164:1688–93.[Abstract/Free Full Text]
  32. Wang GF, Lai MD, Yang RR, et al. Histological types and significance of bronchial epithelial dysplasia. Mod Pathol 2006;19:429–37.[CrossRef]
  33. Laskin JJ, Sandler AB, Johnson DH. Redefining bronchioloalveolar carcinoma. Semin Oncol 2005;32:329–35.[CrossRef][Medline]
  34. Malkinson AM. Genetic studies on lung tumor susceptibility and histogenesis in mice. Environ Health Perspect 1991;93:149–59.[Medline]
  35. Rehm S, Devor DE, Henneman JR, Ward JM. Origin of spontaneous and transplacentally induced mouse lung tumors from alveolar type II cells. Exp Lung Res 1991;17:181–95.[Medline]
  36. Belinsky SA, Stefanski SA, Anderson MW. The A/J mouse lung as a model for developing new chemointervention strategies. Cancer Res 1993;53:410–6.[Abstract/Free Full Text]
  37. Tripodis N, Hart AA, Fijneman RJ, Demant P. Complexity of lung cancer modifiers: mapping of thirty genes and twenty-five interactions in half of the mouse genome. J Natl Cancer Inst 2001;93:1484–91.[Abstract/Free Full Text]
  38. Manenti G, Acevedo A, Galbiati F, et al. Cancer modifier alleles inhibiting lung tumorigenesis are common in inbred mouse strains. Int J Cancer 2002;99:555–9.[CrossRef][Medline]
  39. Nikitin AY, Alcaraz A, Anver MR, et al. Classification of proliferative pulmonary lesions of the mouse: recommendations of the mouse models of human cancers consortium. Cancer Res 2004;64:2307–16.[Abstract/Free Full Text]
  40. Jackson EL, Willis N, Mercer K, et al. Analysis of lung tumor initiation and progression using conditional expression of oncogenic K-ras. Genes Dev 2001;15:3243–8.[Abstract/Free Full Text]
  41. Frith CH, Highman B, Burger G, Sheldon WD. Spontaneous lesions in virgin and retired breeder BALB/c and C57BL/6 mice. Lab Anim Sci 1983;33:273–86.[Medline]
  42. Wang Y, Zhang Z, Yan Y, et al. A chemically induced model for squamous cell carcinoma of the lung in mice: histopathology and strain susceptibility. Cancer Res 2004;64:1647–54.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Proc Am Thorac SocHome page
D. J. Weiss, J. K. Kolls, L. A. Ortiz, A. Panoskaltsis-Mortari, and D. J. Prockop
Stem Cells and Cell Therapies in Lung Biology and Lung Diseases
Proceedings of the ATS, July 15, 2008; 5(5): 637 - 667.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tian, J.
Right arrow Articles by Locker, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tian, J.
Right arrow Articles by Locker, J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online