| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
The Jackson Laboratory, Bar Harbor, Maine
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although the relative rarity of cases of juvenile-type GC tumors has precluded human genetic linkage studies to identify candidate susceptibility genes, mouse models may provide this missing information. To this end, we are pursuing tumor susceptibility genes in the SWR mouse model of spontaneous GC tumor development, which parallels the histological, endocrinological, and developmental timing characteristics of human juvenile-type GC tumors (4
, 5) . GC tumor initiation in SWR females is restricted to the time of ovarian maturation at puberty, within the window of 3–5 weeks of age. By 8 weeks of age, 
1% of SWR females will develop macroscopically visible unilateral or bilateral ovarian GC tumors. Beyond this time frame, females that have not developed GC tumors are no longer at risk, have normal reproductive cycles and normal fertility, and develop no other visible pathological changes in the ovary through 12 months of age.
The GC tumors found in SWR mice are endocrinologically active and secrete inhibin and estrogen in the presence of appropriate androgenic substrate (6 , 7) . This profile is identical to human juvenile-type GC tumors that also produce inhibin and estrogen, leading to symptoms of precocious puberty in young girls or menstrual irregularities in postpubertal women (1 , 2 , 8) . The mouse tumors are homogeneously comprised of GCs and appear as highly vascularized solid or cystic masses. In addition, these tumors are histologically similar to the juvenile-type human tumors by their lack of organized follicular structures (Call-Exner bodies) that exemplify adult-type GC tumors. Between 7 and 12 months of age, GC tumor-bearing mice may develop multiple metastases, and the most commonly seeded organs are kidney, renal lymph node, liver, and lung (9) . Human juvenile-type GC tumors also may become malignant, and the diagnosis of advanced-stage disease (stage III) with tumor spread outside the ovarian capsule confers a poor prognosis for the patient (1 , 10) . Overall, the common biological properties of the mouse and human juvenile-type GC tumors suggest that similar genetic pathways are involved; moreover, the identification of genetic determinants for GC tumor susceptibility in the mouse could provide valuable leads for finding genes involved in juvenile-type GC tumor etiology in humans.
Previous studies designed to identify genes associated with GC tumor development in SWR mice determined that tumor susceptibility is a polygenic trait involving several autosomal Gct susceptibility genes: Gct1 on Chr 4; Gct2 on Chr 12; Gct3 on Chr 15; and Gct5 on Chr 9 (11) . From both F2 intercross and recombinant inbred strain mapping data, it was evident that the SWR allele at the Gct1 locus is requisite for tumor development (11) . In addition to SWR-derived genes, an X-linked tumor incidence modifier gene (Gct4) was identified in the mapping partner strain SJL. The Gct4 gene also shows a strong parent-of-origin effect: if female mice of an otherwise SWR autosomal background inherit an SJL-derived Gct4 allele (Gct4J) from their father, there is a dramatically increased chance of developing an ovarian tumor, although the same is not true for maternal inheritance of the Gct4J allele (4 , 11) . To resolve the identity of Gct4, we constructed novel congenic mouse strains that have an autosomal SWR complement to render the mice susceptible to GC tumors, combined with Chr X segments derived from either the SJL or CAST strain. In this report, we present a high-resolution genetic map for Gct4 and evidence for a second, newly identified X-linked tumor incidence modifier gene Gct6.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
High-Resolution Genetic Map of Chr X.
As a framework for mapping X-linked GC tumor susceptibility genes, we produced 102 backcross mice by mating (CAST x SWXJ-9)F1 females to SWXJ-9 recombinant inbred males. These mice were genotyped for 38 loci within the region containing DXMit46 to DXMit35, a genetic distance spanning 24.5–62 cM (MGD).2
For routine genotyping assays, genomic DNA was isolated from a tail tip biopsy and used in a PCR assay with MIT primer pairs (Research Genetics, Huntsville AL; Integrated DNA Technologies, Coralville, IA). PCR products were separated by electrophoresis in 4% agarose gels and visualized by ethidium bromide staining. MapManager QT, version 3.0b28, was used to align haplotype information and to determine genetic distances between loci.
Ar Gene Sequencing and RFLP.
From the original mapping data, the Chr X region containing Gct4 also contained the androgen receptor gene (Ar). Based on biological evidence that androgens support GC tumorigenesis, the Ar gene was considered a potential candidate for Gct4 (7
, 12) . The coding region of the Ar gene was sequenced in three strains of mice (SWR, SJL, and CAST) and compared with the published 2697 nucleotide transcripts of C57BL/6 and BALB/c strain mice (13)
. Polyadenylated mRNA was isolated from intact ovaries using a MicroFast Tract oligo(dT) kit (Invitrogen, Carlsbad, CA) and converted to cDNA as follows: 500 ng of mRNA, 2 µM gene-specific primer or 50 µg/ml oligo(dT), 1 mM deoxynucleotide triphosphates, 0.01 M DTT, 1x First Strand Buffer, and 200 units of Superscript II reverse transcriptase (Invitrogen) were incubated in a total volume of 10 µl. The Ar cDNA was sequenced in four overlapping segments with the adenosine residue of the ATG start site referenced as base position 1 and the following primer pairs: (a) segment 1 (nucleotide span -57 to +1142), 5'-CAAGAGACGAGGAGGCAGGATAAG (forward) and 5'-TTGATACGGGCGTGTGGATG (reverse); (b) segment 2 (nucleotide span +679 to +1614), 5'-GACAGTGCCAAGGAGTTGTGTAAAG (forward) and 5'-GGTCTTCTGGGGTGGAAAGTAATAG (reverse); (c) segment 3 (nucleotide span +702 to +2021), 5'- AGCAGTGTCTGTGTCCATGG (forward) and 5'- GTTGGTTGTTGTCATGTCGTCCG (reverse); and (d) segment 4 (nucleotide span +1555 to +2688), 5'-TTTGGACAGTACCAGGGACC (forward) and 5'-CCAAATCTTCACTGTGTGTGG (reverse).
Each PCR reaction contained 1 µl of cDNA, 0.2 mM deoxynucleotide triphosphates, 0.4 µM of each primer per pair, 1x PCR buffer, and 1x Advantage cDNA polymerase mix (Clontech, Palo Alto, CA) in a volume of 50 µl. Thermocycling conditions were as per manufacturer’s instructions for fragments < 5 kb for a total of 30 cycles. PCR fragments were cloned in a dual promoter TA cloning vector (K2030-01; Invitrogen) following the manufacturer’s instructions. Cloned DNA segments were sequenced in both directions and aligned with the published cDNA sequence using Sequencer software (Ver. 4.1; Gene Codes Corp., Ann Arbor, MI).
The DNA sequence of the CAST Ar gene revealed five differences compared with the Ar gene of other laboratory stains examined. The nucleotide substitution (C to A) at residue 1036 in exon 1 of CAST Ar resulted in the loss of a BfaI restriction enzyme site, and this distinction was used to develop a RFLP assay. Two primers (5'- GACAGTGCCAAGGAGTTGTGTAAAG and 5'-TTGATACGGGCGTGTGGATG) were used to PCR amplify a 464-bp region of exon 1 from genomic DNA. Resulting fragments were incubated at 37°C with 10 units of BfaI (New England Biolabs, Beverly MA) in appropriate buffer for 12 h. After enzymatic digestion, the 464-bp fragment of CAST remains uncut, whereas the SWR- or SJL-derived Ar is cut into two fragments (359 and 105 bp).
SWR.SJL-X Congenic Strains.
Previously developed SWR.SJL-X congenic mice that carry large segments of SJL Chr X were used as starting material for the derivation of new congenic mouse strains carrying smaller, overlapping segments of SJL Chr X (11)
. Males from the congenic strains SWR.SJL-X2 or -X3, characterized by the phenotype of high GC tumor incidence in female siblings (20%), were backcrossed to SWR progenitors to produce females heterozygous for SWR and SJL alleles in the Gct4 candidate region. F1 females, heterozygous for the central region of the Chr X, were backcrossed to SWR males, and the progeny were genotyped to uncover additional informative recombinations. Females and males that carried smaller segments of SJL-X within the region bounded by DXMit96 and DXMit158 were subsequently intercrossed to generate three new SWR.SJL-X congenic strains, denoted as SWR.SJL-X5, -X6, and -X7.
SWR.CAST-X Congenic Strains.
A large segment of Chr X from the CAST strain was transferred onto the SWR.SJL-X3 strain background through 10 consecutive backcross generations to generate the unique strain SWR.CAST-X (Fig. 1)
. The CAST segment spans the region DXMit109 to DXMit35, a genetic distance of 18 cM that includes Gct4. The N10 backcross scheme also replenished the SWR genomic background (estimated return of 99% SWR genomic material). Established mating pairs of SWR.CAST-X were confirmed to possess SWR allelic background at critical regions for Gct susceptibility loci on Chrs 4, 9, 12, and 15. This colony was established to (a) examine the GC tumor frequency in female progeny homozygous for this segment and (b) use the polymorphism potential of the CAST strain to facilitate high-resolution genetic mapping of Gct4. Although the genetic background of this congenic strain is effectively SWR.CAST-X, we used SWR.SJL-X3 as the recipient strain because it would be more informative than the SWR progenitor strain. Because SWR.SJL-X3 females have a higher GC tumor frequency phenotype (20% of the population affected) than SWR females (1%), a reduction in GC tumor frequency in females of the SWR.CAST-X strain to 1% or less would (a) indicate successful displacement of the modifier allele Gct4J with the CAST allele and (b) determine that the CAST allele of Gct4 did not support high GC tumor frequency.
|
. These females were backcrossed to SWR.CAST-X males, and their F1 sons were genotyped for maternal inheritance of a recombinant Chr X. Recombinant F1 male offspring were mated to SWR females, and their daughters were examined for GC tumors. This test was established using only male Chr X recombinants because all evidence indicated that the parent-of-origin effect for Gct4 to support high frequency tumor development was restricted to paternal inheritance (11)
.
|
2 analysis for proportions, with a chosen significance level of P < 0.05 (Statview, version 4.5; SAS Institute, Cary, NC). |
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The marker order and recombination distances are in good agreement with genetic mapping data listed on the MGD, Mouse Genome Informatics Web Site,2
The Jackson Laboratory (Bar Harbor, ME) and the physical map available at Ensembl database.3
|
. GC tumor frequencies also are shown for congenic females that carry homozygous Chr X segments of SJL-X. The SWR.SJL-X5 and -X6 females have a significantly higher GC tumor incidence (13.1% and 20.5%, respectively) than SWR females (1%; P < 0.05). This indicates that both SWR.SJL-X5 and -X6 females carry the Gct4J modifier allele, whereas the SWR.SJL-X7 congenic does not, and that Gct4 resides in the region between DXMit45 and DXMit170, a 3 cM interval that includes the Ar gene (Fig. 3)
.
|
. The missense mutation at nucleotide 1036 in exon 1 of CAST results in the loss of a BfaI restriction enzyme site, which allowed us to develop a RFLP assay for the identification of CAST versus SWR or SJL Ar alleles (see "Materials and Methods"). This assay was used to map the Ar gene on the SWXJ-9.CAST N2F1 backcross panel (Fig. 3)
.
|
, GC tumor incidence in F1 females revealed that the modifier action and the paternal parent-of-origin effect were still associated, despite the narrowed genetic interval for Gct4 in the SWR.SJL-X5 strain. In each cross, 20% of F1 females developed a GC tumor if the Gct4J allele was paternally inherited, but only 2% of females developed a GC tumor if the Gct4J allele was maternally inherited.
|
Chr X Recombinant Males.
To narrow the location of Gct4, 15 males carrying unique Chr X recombinations between the SJL and CAST congenic segments were mated to SWR females, and their F1 daughters were examined for GC tumor development. Fig. 5
shows the Chr X haplotypes of the unique males tested (3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
and the GC tumor incidence in their F1 female progeny that were heterozygous for Chr X. Males 1 and 2 represent negative controls taken directly from the SWR.CAST-X strain; as expected, these males did not support high-frequency GC tumor development in F1 daughters.
|
20%. This was surprising because the only difference between SWR.SJL-X3 congenic males that support 20% GC tumor frequency and males 4, 5, and 6 was the presence of a CAST-derived segment distal to Gct4. Second, the presence of another SJL-derived allele in the region containing DXMit38 supported high-frequency GC tumor development. F1 daughters of males 7–14 had a mean tumor incidence of 35% (29–47%), and the haplotype analysis defined the region containing the second modifier gene to a 1 cM interval between DXMit79 and DXMit4. This second gene is designated Gct6 and lies 6.5 cM distal to Gct4 on Chr X. As indicated by our progeny test cross panel, the presence of CAST regions lying proximally or distally to the Gct6 gene revealed this second GC tumor modifier locus; however, the introduction of CAST segments also subdued the modifier activity of Gct4. |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In juvenile-type GC tumors not associated with bone syndromes, there are multiple cases reported with trisomy for Chr 12 (17, 18, 19, 20, 21, 22) . Generally, the accumulated evidence does not suggest that this is an early or necessary step in tumorigenesis because not every tumor presents with this trisomic condition, and the tumor cell populations that do are heterogeneous. Full cytogenetic analysis of juvenile GC tumors and ploidy assessments by flow cytometric methods have indicated that these tumors can harbor multiple chromosomal abnormalities and are generally aneuploid, so the specific role of Chr 12 trisomy in tumor development remains to be established (23 , 24) . In the SWR mouse model for GC tumor development, cytogenetic abnormalities at the level of gross chromosomal duplication or loss are not observed (4) . Thus, the identification of individual GC tumor susceptibility genes represents a promising approach to test for further common genetic etiologies between the mouse and human GC tumors.
All evidence gathered from the SWR mouse model indicates that the site of action for Gct susceptibility genes lies within the GC cells. Alterations of the hormonal signaling profiles within the hypothalamic-pituitary-ovarian axis were ruled out by the fact that grafted, genetically susceptible ovaries can form tumors in fertile female recipients that are not genetically susceptible (12) . GC tumors do not form, however, if transplant recipients are gonadotropin-deficient (hypogonadal, hpg/hpg) mutants that do not undergo sexual maturation (12) . The ovarian transfer experiments thus indicate that the normal endocrine stimulation occurring during ovarian maturation is necessary and sufficient for GC tumor initiation and that the maturing SWR hypothalamic-pituitary-ovarian axis is not defective in this regard. GC tumor initiation is also restricted to the time of ovarian maturation at puberty, suggesting that the Gct susceptibility genes have a short window of influence over GC fate (12) . During puberty in mice, the first wave of ovarian follicles mature, but these growing follicles subsequently undergo atresia rather than maturation and ovulation (25) . Follicular atresia also occurs in subsets of follicles in any subsequent reproductive cycle, but this does not correlate with GC tumor initiation potential, suggesting that other GC cell attributes are unique during puberty.
Spontaneous GC tumor development in SWR mice is a polygenic trait associated with multiple SWR-derived genes on Chrs 4, 9, 12, and 15 (11) . The gene with the strongest genetic influence on tumor development is Gct1 on distal Chr 4 near D4Mit232. This region of mouse Chr 4 is orthologous to human Chr 1p36, which is intriguing because 1p36 harbors multiple genes associated with various cancers (26, 27, 28) . Only 1% of SWR females develop GC tumors, but if their otherwise SWR genetic background is combined with the SJL allele of the X-linked modifier gene Gct4, these females have a dramatically increased risk of tumor development (11) . In this study, we refined the location of Gct4 to a 3 cM interval between DXMit45 and DXMit170 that includes the Ar gene. The biological studies with GC tumor-susceptible ovaries had already implicated androgenic stimulation as an important signal for both tumor initiation and promotion, making the Ar gene an excellent candidate for Gct4 (7 , 29) . To further explore this possibility, we sequenced the eight exons of the Ar gene in strains SWR, SJL, and CAST but did not find differences among SWR, SJL and the published consensus sequences. We conclude that mutation in the Ar coding regions did not explain the different biological activities of Gct4SW and Gct4J alleles, but our results do not rule out sequence differences in the regulatory regions that may affect Ar gene expression patterns in the ovary. We did identify three missense mutations and two silent mutations in the CAST Ar gene. It is unclear whether these nucleotide changes alter the function of Ar protein in CAST mice, but given that CAST males are fertile, and females appear normal, it suggests no overt loss or gain of function is present in receptor activity in these mice. The missense mutations in the CAST Ar gene are clustered in the NH2-terminal transactivation domain of the protein that mediates interactions with the basal transcription machinery and coactivating factors, suggesting that amino acid substitutions are tolerated and do not interfere with normal Ar signaling pathways. The unique missense mutation in exon 1 of the CAST Ar gene was used to map the Ar gene to the region of DXMit41, 96, 169, a position in good agreement with that reported on public databases (MGD2 and Ensembl3 ).
SWR.SJL-X5 and -X6 congenic mice were confirmed carriers of the Gct4J allele based on a high GC tumor incidence phenotype (20%), and they were further investigated for differences between paternal and maternal inheritance of the Gct4J allele. From reciprocal crosses carried out with the SWR progenitor strain, it was clear that the modifier gene’s activity to support increased GC tumor incidence was maintained by paternal but not maternal inheritance. It was interesting that these two distinct qualities, high tumor incidence and the parent-of-origin effect, were still associated despite the minimized segment of SJL chromosomal material carried in SWR.SJL-X5 congenic strain, suggesting that the parental imprint is linked to Gct4 activity.
Our mapping strategy with SWR.SJL-Xn strains successfully narrowed the region containing the SJL-derived modifier allele Gct4J between DXMit45 and DXMit170, which represents a 3 cM genetic distance, or approximately 26 Mb of DNA. This region of the Chr X contains 75 known genes and >100 predicted novel genes.3 By generating the congenic strain SWR.CAST-X, we successfully introduced a higher degree of genetic variance for mapping purposes. The SWR.CAST-X strain also provided new information about Chr X and GC tumor susceptibility. First, we uncovered another X-linked modifier locus (Gct6) within a 1-cM interval that lies 6.5 cM distal to Gct4 on Chr X and strongly supports GC tumorigenesis. It has not yet been determined whether the newly discovered Gct6 gene has a similar parent-of-origin effect as that observed for Gct4, but our data show that paternal inheritance is sufficient for high tumor frequency. Second, we found evidence that the tumor incidence-modifying activity of Gct4 is dependent on the allelic background of the distal Chr X region that includes Gct6. To explain this complexity, we hypothesize that the genes Gct4 and Gct6 interact in a common biological pathway. In this scenario, SWR and SJL alleles of Gct6 are supportive for the modifier activity of Gct4J, but the CAST allele of Gct6 is suppressive. In the case of the Gct6J allele, the SJL and CAST alleles of Gct4 are supportive of the modifier activity, but the SWR allele is suppressive. Evidence for this hypothesis awaits the generation of additional males carrying recombinant Chr X to test each possible combination of Gct4 and Gct6 alleles.
An alternate but plausible hypothesis to explain the background-dependent activity of Gct4 and Gct6 stems from consideration of the unique properties of Chr X of the wild-derived CAST strain that is evolutionarily and genetically divergent from the SWR and SJL laboratory strains. Our results show that introduction of a CAST Chr X segment into the SWR.SJL-X genome either proximal or distal to the DXMi233-DXMit38 region revealed Gct6, a gene that strongly supports GC tumorigenesis. However, in the absence of this CAST segment, there was no evidence for a Gct6 gene, even in SWR.SJL-X congenics (SWR.SJL-X3 and -X4) suspected of carrying the appropriate SJL allele. Similarly, the introduction of a CAST segment "masked" the activity of Gct4 as a modifier of GC tumor frequency, even when CAST segments were 6 cM distal to Gct4. This raises the question of whether Chr X regions of CAST are acting in an unconventional way in the ovary to affect Chr X gene expression that is ultimately reflected in the threshold for GC tumor initiation. For instance, it is recognized that a unique CAST allele of the X-controlling element (XceC), which is part of the X-inactivation center (Xic) locus, has a very strong impact on the choice of the active Chr X during the process of X inactivation (30
, 31)
. Generally, X inactivation is random in female somatic cells, such that the probability is equal as to whether the maternally derived or the paternally derived X Chr is transcriptionally inactive (Xi) or active (Xa). Inheritance of the XceC allele from the CAST strain, however, can dramatically skew the choice in favor of the XceC-carrying Chr, irrespective of paternal or maternal route of inheritance. In our progeny test panel, inheritance of the CAST XceC allele per se is not a requirement for the modifier activity of Gct6 because the effective CAST segments can lie outside the Xic locus located between DXMit170 and DXMit171. This suggests that this exact mechanism of skewed X-inactivation is not a factor in GC tumor initiation but does raise the possibility that novel genetic features of the CAST Chr X were introduced coincidentally with the advantageous increase in polymorphic markers.
Given the essentially normal reproductive function of SWR, SWR.SJL-Xn, or SWR.CAST-X females (excepting those females that develop GC tumors), there is no obvious ovarian phenotype associated with the Gct4 and/or Gct6 genes. We speculate both Gct4 and Gct6 genes regulate normal ovarian activity that interacts synergistically with SWR-derived autosomal mutations to support high-frequency GC tumorigenesis. In sheep, a Chr X-linked locus named FecX2Wshows a strong paternal, parent-of-origin effect and supports high ovulation rates (32) . Given the high degree of evolutionary conservation for Chr X genes among mammals and the common parent-of-origin effect, it is possible that Gct4 and FecX2W are the same gene, although their reproductive functions may differ between species. This is a realistic scenario because another X-linked gene, bone morphogenetic protein 15 (Bmp15), has been shown to have distinct effects on the reproductive function of the ovaries of mouse and sheep (33) . Conversely, Gct4 and Gct6 may be novel genes, adding to a growing list of X-linked genes that affect ovarian function.
In summary, we have mapped two modifier genes on the mouse Chr X that support high-frequency, juvenile-type GC tumor development in female mice. Understanding the function of these genes will provide novel directions for the pursuit of tumor susceptibility genes in human cases of juvenile-type GC tumors.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
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.
Requests for reprints: Ann M. Dorward, The Jackson Laboratory, 600 Main Street, Bar Harbor, Maine 04609.
1 The abbreviations used are: GC, granulosa cell; SWR, SWR/Bm; Chr, chromosome; SJL, SJL/Bm; CAST, CAST/Ei; MGD, Mouse Genome Database. 
Received 6/ 9/03. Revised 8/14/03. Accepted 9/17/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. A. Edson, A. K. Nagaraja, and M. M. Matzuk The Mammalian Ovary from Genesis to Revelation Endocr. Rev., October 1, 2009; 30(6): 624 - 712. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M Dorward, K. L Shultz, and W. G Beamer LH analog and dietary isoflavones support ovarian granulosa cell tumor development in a spontaneous mouse model Endocr. Relat. Cancer, June 1, 2007; 14(2): 369 - 379. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Dorward, K. L. Shultz, L. G. Horton, R. Li, G. A. Churchill, and W. G. Beamer Distal Chr 4 Harbors a Genetic Locus (Gct1) Fundamental for Spontaneous Ovarian Granulosa Cell Tumorigenesis in a Mouse Model Cancer Res., February 15, 2005; 65(4): 1259 - 1264. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
