| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Departments of Pathology, Tufts University–New England Medical Center (S.J.D.-C., R.T., H.J.W.), Boston, Massachusetts 02111; Barts and The London Queen Mary’s School of Medicine and Dentistry (S.J.D.-C.), E1 1BB London, United Kingdom; University Hospital of Seville (M.d.M.), 41009-Seville, Spain; and University Hospital of Malaga (A.B.), 29010-Malaga, Spain
Address all correspondence and requests for reprints to: Salvador J. Diaz-Cano, M.D., Ph.D., Department of Histopathology and Morbid Anatomy, The Royal London Hospital, Whitechapel, London E1 1BB, United Kingdom. E-mail: s.j.diaz-cano{at}mds.qmw.ac.uk
Abstract
C-cell hyperplasias are normally multifocal in multiple endocrine neoplasia type 2A. We compared clonality, microsatellite pattern of tumor suppressor genes, and cellular kinetics of C-cell hyperplasia foci in each thyroid lobe.
We selected 11 females from multiple endocrine neoplasia type 2A kindred treated with thyroidectomy due to hypercalcitoninemia. C-cell hyperplasia foci were microdissected for DNA extraction to analyze the methylation pattern of androgen receptor alleles and microsatellite regions (TP53, RB1, WT1, and NF1). Consecutive sections were selected for MIB-1, pRB1, p53, Mdm-2, and p21WAF1 immunostaining, DNA content analysis, and in situ end labeling. Appropriate tissue controls were run.
Only two patients had medullary thyroid carcinoma foci. Nine informative C-cell hyperplasia patients showed germline point mutation in RET, eight of them with the same androgen receptor allele preferentially methylated in both lobes. C-cell hyperplasia foci showed heterogeneous DNA deletions revealed by loss of heterozygosity of TP53 (12 of 20), RB1 (6 of 14), and WT1 (4 of 20) and hypodiploid G0/G1 cells (14 of 20), low cellular turnover (MIB-1 index 4.5%, in situ end labeling index 0.03%), and significantly high nuclear area to DNA index ratio.
MEN 2A (germline point mutation in RET codon 634) C-cell hyperplasias are monoclonal and genetically heterogeneous and show down-regulated apoptosis, findings consistent with an intraepithelial neoplasia. Concordant X-chromosome inactivation and interstitial gene deletions suggest clone expansions of precursors occurring at a point in embryonic development before divergence of each thyroid lobe and may represent a paradigm for other germline mutations.
THE IDENTIFICATION OF RET germline mutations in multiple endocrine neoplasia type 2 (MEN 2) has resulted in carrier detection and early diagnosis of C-cell hyperplasias (CCHs) and medullary thyroid carcinomas (MTCs) (1, 2, 3). RET mutation seems to be necessary, but other alterations should be required to explain the MEN 2 phenotypic heterogeneity and the differences between CCH and adrenal medullary hyperplasias.
Clonality is still the hallmark of neoplasia and strongly suggests acquired somatic mutations, providing proliferative advantage to a given cell population (4, 5). Clonality has been studied by X-chromosome inactivation (XCI) assays, which are based on the DNA methylation of many genes that render maternal and paternal chromosome functionally nonequivalent (6, 7, 8). The same point mutation or single nucleotide polymorphism (SNP) within all cells also implies a common progenitor, (4), and it has been found associated with loss of heterozygosity (LOH) of certain loci (9). LOH should be linked to inactivation of tumor suppressor genes (TSG) by DNA deletions, contributing to tumor cell selection (4). These non-X-linked markers test clonal expansions associated to selective growth advantages (high proliferation and/or abnormally low apoptosis) (4, 10, 11, 12, 13), but they fail to identify clones occurring prior to the presence of the genetic lesion (9, 10).
This study investigates the clonal patterns of CCH associated with MEN 2A (RET point mutation in codon 634) in both thyroid lobes, based on the analysis of both the methylation pattern of androgen receptor alleles and TSG microsatellite patterns using microdissected tissue samples. The kinetic features of such lesions are also analyzed.
Materials and Methods
Case selection
MEN 2A kindred with a Cys634 to Tyr substitution of RET proto-oncogene (177 members, five generations) were screened for CCH/MTC by basal and pentagastrin-induced calcitonin levels (14, 15, 16). All patients fulfilled the clinical and molecular criteria proposed by the International RET Mutation Consortium (17). Patients with pentagastrin-induced hypercalcitoninemia and/or RET point mutations underwent total thyroidectomy, the specimens being serially sectioned and completely embedded for histopathologic diagnosis.
Appropriate archival material from both lobes was available from 11 females. We initially selected cases showing 50 or more C cells in at least one x100 microscope field per lobe (18, 19). C cells were large, mildly to moderately atypical, and confined within the follicular basement membrane, but cytologically indistinguishable from invasive MTC (20, 21). One patient (case CCH-6) did not show RET mutation and was excluded from further analyses. The same areas in consecutive sections were used in each study, and their cellular composition was confirmed in adjacent hematoxylin-eosin-stained sections.
PCR analysis of clonality and TSG microsatellites
One CCH focus containing at least 100 C cells (0.5 mm; Ref. 2) was microdissected from each thyroid lobe (two 20-µm unstained sections/focus). Appropriate controls (follicular cells, lymph node, peripheral nerves, and thyroidal soft tissue) were carried out.
DNA was extracted using a modified phenol-chloroform protocol,
(22) and digested with HhaI (New England Biolabs, Inc., Beverly, MA). Half of each sample underwent
enzymatic digestion (0.8 U/µl), whereas the remaining half was kept
as undigested control. Both samples were equally processed, but
excluding HhaI in the undigested ones (10, 23, 24). A mimicker (0.3 µg of double-stranded and
XhoI-linearized 
X174-RII phage) (Life Technologies, Inc., Gaithersburg, MD) was included in each reaction for
digestion testing. Complete digestion was checked by gel
electrophoresis; incompletely digested samples were phenol chloroform
purified and redigested with higher HhaI concentration.
HhaI was then inactivated by phenol-chloroform extraction
(22). DNA was precipitated with ice-cold absolute ethanol
in the presence of 0.3 mol/liter sodium acetate (pH 5.2) and
resuspended in 10 µl of 10 mmol/liter Tris-HCl (pH 8.4), 50
mmol/liter KCl, 1.5 mmol/liter MgCl2, and 100 µg/ml BSA.
The CAG repeat in the first exon of the human androgen receptor (AR)
gene was amplified using both digested and undigested DNA templates
(13, 23, 25, 26). The undigested DNA was also used for
amplification of intron microsatellites at TSG (TP53,
RB1, WT1, and NF1) loci, according to
the optimized conditions shown in Table 1
(25, 27). The tests were run in a Perkin-Elmer thermal
cycler model 480 (Perkin-Elmer Corp., Norwalk, CT). The
whole PCR volume (10 µl) was electrophoresed into 0.75-mm thick 8%
polyacrylamide gels, nondenaturing for the AR product and 20–80%
denaturing gradient (from top to bottom) for TSG products (13, 25, 26). The gels were run at 5 V/cm until xylene cyanol band
was within the bottom inch of the gel. The gels were then fixed with
7% acetic acid (5 min), dried under vacuum (40 min, 80 C), and put
inside a developing cassette containing one intensifying screen and
preflashed films (Kodak XAR; Kodak,
Rochester, NY) facing the intensifying screen (16–48 h, -70 C). The
autoradiograms were developed using an automated processor
Kodak-Omat 100 (Kodak).
|
Nuclear DNA quantification by slide cytometry
Feulgen-stained sections were used for densitometric evaluation of DNA content (29). CAS model 200 and the Quantitative DNA Analysis software (Becton Dickinson and Co., Franklin Lakes, NJ) were used for that purpose, measuring at least 200 nuclei (or the whole lesion if smaller) in every CCH focus in the most cellular area until completion in consecutive microscope high power fields (HPFs; x400). Only complete, nonoverlapping and focused nuclei were quantified in each HPF.
Internal controls (both lymphocytes and histologically normal
follicular cells from the same section) were first normalized with
complete rat hepatocytes (external controls, one slide/staining holder)
(Becton Dickinson and Co.). They were then used for
setting the diploid controls and calculating the DNA index of
G0/G1 cells (
10% of
measured cells with evidence of G2+M cells)
(30). The proliferation rate was calculated from the DNA
histogram by subtracting the number of cells within
G0/G1 limits from the total
number of measured cells and expressed as percentage (29, 31). Both mean nuclear area and nuclear area to DNA index ratio
of G0/G1 cells were
recorded. The latter represents a morphometric parameter of apoptosis
when coupled with in situ end labeling (ISEL) (13, 30). Age- and sex-matched normal thyroids from 10 autopsies were
selected for nuclear area and DNA index analysis. At least 1000 C cells
were evaluated as controls for this purpose.
ISEL of fragmented DNA
Extensive DNA fragmentation associated with apoptosis was detected by ISEL as previously reported (31, 32, 33). Sections were incubated in 2x SSC (20 min, 80 C) and digested with proteinase K [100 µg/ml in Tris-HCl (pH 7.6), 30 min] at room temperature in moist chamber. DNA fragments were digoxigenin-labeled on 5'-protuding termini using the Klenow fragment of Escherichia coli DNA polymerase I (1 h at 37 C), detected using antidigoxigenin Fab fragments labeled with alkaline phosphatase, and developed with nitroblue tetrazolium-X phosphate (31, 32, 33, 34). Appropriate controls were run, including positive (reactive lymph node), negative (omitting DNA polymerase I), and enzymatic (DNase I digestion before end labeling). The enzymatic controls allowed establishing a reliable positivity threshold in each sample. The ISEL index was expressed as percentage of positive nuclei compared with the total number of C cells in the same HPF (35, 36, 37). The whole lesion was screened.
Immunohistochemical expression of MIB-1 and cell cycle regulators
After quenching endogenous peroxidase activity with 0.5% H2O2-methanol (10 min.), the antigens were retrieved by microwaving sections in 10 mmol/liter citrate buffer (pH 6.0; 750 W, 2 cycles of 5 min) and the sections were transferred to a moist chamber for the remaining steps (34). The tissue sections were incubated overnight at 4 C with the corresponding primary monoclonal antibody (normal pRB1 at 5 µg/ml, abnormal p53 at 1 µg/ml, and cyclin D1, p21WAF1, MDM-2, and MIB-1 at 2.5 µg/ml) (Oncogene Science, Inc. and Calbiochem, Cambridge, MA). Nonspecific binding was blocked pretreating the sections with diluted horse serum (1:100) and the signal amplified with biotinylated antimouse secondary antibody (1:200) and peroxidase-labeled avidin-biotin complex (1:100) (Vector Laboratories, Inc., Burlingame, CA). All reactions were developed under microscopic control with 3,3'-diaminobenzidine as chromogen and counterstained with hematoxylin. Appropriate positive (reactive lymph node) and negative (omitting the primary antibody) controls were simultaneously run. The immunohistochemical expression was screened and scored as described for the ISEL analysis.
Statistical analysis of quantitative variables
ANOVA and Student’s t tests were applied to assess the differences of morphometric features in G0/G1 cells by DNA content and considered significant if P less than 0.05.
Results
At the time of thyroidectomy, the patients aged 4–8 yr and there
was no clinical or laboratory evidence of adrenal or parathyroid
pathology. All patients had bilateral CCH, associated with MTC in two
cases (CCH-10 and CCH-11). Two cases (four CCH foci) were excluded from
the clonality assay. Case CCH-6 showed no germline RET
mutation and a polyclonal secondary CCH due to chronic thyroiditis
(Fig. 1). Case CCH-2 was
noninformative due to AR locus LOH (Table 2).
|
|
). Another
patient (two CCH foci) displayed a polyclonal pattern. Assuming equal
probability of methylation of each AR allele at the cellular level, the
probability of concordant methylation patterns of AR alleles will be
0.5 (Fig. 2a). However, tissues from
informative females can be polyclonal, monoclonal with preferential
methylation of the larger allele, or monoclonal with the smaller allele
predominating (Fig. 2b) (38). Under those circumstances
and assuming independent methylation of each allele, equal a
priori probability should be expected for each chance and tissue
(P = 1/3). In tissue comparisons from informative
patients (two different alleles present), concordant allele patterns
would expected in 2(1/3)n, being 2
= number of alleles and n = number of lesions compared. Therefore,
identical patterns could be randomly found in
2(1/3)2 = 0.22 when two tissue samples are
compared (Fig. 2b). Therefore, the probability of randomly finding
identical allele patterns in tissues from eight females would be
0.228 = 5.49
10-6. Any tissue-specific
variability in DNA methylation was excluded using multiple matched
controls of the same sample size. The polyclonal case (CCH-4, Table 2)
could not be accurately assessed for this purpose, due to the absence
of gel pattern differences between true polyclonal proliferation and
proliferation of two clones revealing different X chromosome
inactivated (Fig. 2b).
|
). LOH/SNP analyses showed
TP53 abnormalities in 6 of 10 informative cases (12 of 20
CCH foci, 60%), RB1 alterations in 4 of 7 informative
patients (6 of 14 CCH foci, 43%), and WT1 deletions in 2 of
10 informative patients (2 of 20 CCH foci, 10%) (Table 2, Fig. 3). No alterations were detected in
NF1 locus. TP53 abnormalities were concordantly
detected in both lobes in all patients in at least one polymorphic
region. Discordant genetic results were observed only in two patients
for RB1 (SNP) and in two patients for WT1 (LOH),
and coexistent TP53 and RB1 genetic alterations
in two patients (Table 2).
|
). Hypodiploid
G0/G1 cells revealed a
relatively decreased nuclear area (35.11 ± 4.20 µm
(2) and a significantly higher nuclear area to DNA index
ratio [43.12 ± 3.59 µm (2), P <
0.003) than control C cells [36.48 ± 3.98 µm (2),
37.22 ± 4.06 µm2, respectively]. Low
proliferation indices were obtained from the DNA histogram analysis
(<5% for monoclonal foci and 7.6–8.3% for polyclonal foci).
MIB-1 and cell cycle regulator immunoexpressions and ISEL revealed a
heterogeneous marker distribution (Fig. 4). Both proliferation and apoptosis
indices were low, especially the ISEL index (Table 3).
Remarkably, no p53 expression was detected in CCH foci (Fig. 4).
|
|
CCHs in MEN 2A patients were found monoclonal and showed the same inactivated X-chromosome in both lobes and multiple TSG abnormalities. These findings and the down-regulation of apoptosis are consistent with an intraepithelial neoplasia and suggest that progressional genetic events occur early in thyroid development.
The same AR allele was preferentially methylated in both lobes in
monoclonal CCHs (eight of nine informative patients, 89%; Table 2),
suggesting that a common progenitor contributed to those lesions
(25, 39, 40), and support a neoplastic nature. These
concordant patterns also support a clonal expansion of C-cell
precursors between the random XCI and the precursor migration into the
thyroid anlage (
15 wk, Fig. 5)
(11, 41). Supporting this point, patients of these kindred
also revealed concordant monoclonal patterns in 90% nodules from
adrenal medullary hyperplasias and microsatellite abnormalities in at
least one TSG in 75% pheochromocytomas (26, 42). Early
clonal expansions in neural crest derivatives could explain those
findings, but other alternative explanations for the monoclonal pattern
must be excluded.
|
1 kb (data not shown), excluding
degraded DNA as cause (13, 22, 25, 28, 44). Our PCR design
also included long denaturation and extension in the first three cycles
and 7-deaza-dGTP in the amplification mixture to avoid defective
amplification of CG-rich sequences (Table 1) (10, 13, 24, 25, 40). These technical considerations could reasonably exclude the
PCR bias as cause of concordant monoclonal patterns, as revealed by the
two cases showing the larger allele methylated (Fig. 1).
The polyclonal pattern in patients with germline RET
mutation (CCH-4) needs some considerations. XCI assays can result in
polyclonal patterns if the restriction endonuclease digestion is
incomplete or the target DNA is hypermethylated (10, 11, 13, 25, 28). Suboptimal enzymatic digestion may change the clonality
pattern of monoclonal tissues (10, 28), but that
possibility was excluded keeping internal controls of endonuclease
digestion. Although some DNA denaturation must be expected during
embedding and extraction, both the long digestion (16 h) and the
HhaI activity on single-stranded DNA assure complete
digestion. Repeated polyclonal patterns were obtained using nonboiled
templates. We are currently testing methylation in these lesions.
Likewise, any significant contamination can explain polyclonal results
as those reported in MEN 2 MTC (45), opposed to the clonal
origin previously reported in inherited MTC and pheochromocytomas
(46, 47). Contamination was excluded by careful
microdissection under microscopic control and multiple sampling
(10, 11). Finally, the coexistence of TP53 and
RB1 abnormalities (Table 2) support a neoplastic nature and
agrees with the high MTC incidence (60%) in double heterozygous
animals for TP53 and RB1 (48). In
this scenario, the polyclonal pattern could be the result of either
C-cell precursor expansions before the random XCI or clonal
proliferation of cells showing different inactivated X-chromosome (Fig. 5) (11, 45).
The inherited genetic defect can determine an accelerated process of
somatic alterations and DNA deletions (hypodiploid
G0/G1 cells and TSG-LOH)
(11, 13), favoring cellular heterogeneity and molecular
progression (11, 25). Our results suggest that progression
in CCHs involves multiple genes, especially TP53 and
RB1 as reported in MTC (48, 49). Although the
number of cases is not sufficient to exclude chance, the concordant
TP53 deletion found in three patients (Table 2) suggests
that TP53 deletions have occurred at a point in embryonic
development before divergence of the C-cell precursors of each thyroid
lobe in this patient subset. Therefore, these mutations should occur
earlier than previously thought and correlate with the clinical
detection of MTC in patients younger than 5 yr. Conversely, the
heterogeneity of RB1 and WT1 alterations and
intragenetic polymorphism of TP53 mutations support their
somatic nature and a sort of "hot spot" for such mutations.
WT1 and RET have already shown collaborative
effects, as revealed by the presence of genitourinary abnormalities in
RET or GDNF knockout mice (50, 51).
In contrast, no patient in this series showed NF1
alterations, although this locus has revealed alterations in neural
crest tumors (especially pheochromocytomas) (52, 53).
Early clonal expansions in neural crest derivatives accumulating
genetic alterations would be the expression of multistep tumorigenesis
and is also supported by the reported abnormalities of loci on
chromosomes 1 and 3 (54, 55). The presence of additional
somatic RET mutations has been reported in MTC and should be
considered expression of genetic progression and tumor cell selection
rather than evidence against the clonality of those lesions
(54). Another aspect to consider is the origin of all
patients from a single family and the influence of familial effects on
the results. However, the familial genetic background cannot explain
the association of unrelated aspects such as concordant patterns of
clonality and the presence of multiple TSG abnormalities.
A TP53- and RB1-knockout model revealed MTC within the first 9–15 months of mouse life (60–83%) with the highest incidence in double heterozygous animals (mutated RB1 and TP53) (48). The absence of normal TP53 function would protect tumor cells from apoptosis (56) and contributes to an increased RB1 mutation rate in advanced tumors (48). Remarkably low ISEL indices and low-to-normal proliferation indices were observed in this CCH series. The reduced cellular loss associated with longer survival of initiated/mutated cells would promote progression from preinvasive stages over an extended period of time (11, 13, 57), and genetic instability (49). Transforming mutations that escape cell repair systems will accumulate and promote early preneoplastic foci only after inhibiting apoptosis (11, 13, 58). The postmitotic repairing systems are less effective in low-proliferation lesions (59), explaining the presence of heterogeneous and nonrandomly distributed mutations (60). In contrast, high proliferation and mutation rates determine homogeneous DNA damage and activation of apoptosis (57, 60, 61). Normal pRB1 and low p21WAF1 expressions found in this series are consistent with low proliferation and apoptotic indices. Only higher proliferation rates and increased p21WAF1 would inactivate pRB1 resulting in higher apoptosis (59). The combined delivery of two cooperative genes, like p21WAF1 and TP53, would lead the cyclin D/CDK-pRB1 pathway toward growth arrest and apoptosis (62, 63).
Some clinically relevant topic still remains unknown. RET encodes a receptor tyrosine kinase that is expressed in neural crest derivatives and their corresponding tumors, MTC, and pheochromocytoma (64). RET mutations of codon 634 cause receptor dimerization and constitutively activate the tyrosine kinase (65), thereby mimicking the effect caused by the binding of a ligand to the receptor. In general, a single mutation causing a neoplasm is unlikely, and other genetic alterations (TP53 and RB1 in this series) would be required (11). Although normal tissues are heterogeneous and show random LOH and DNA deletions in 4–20% (66, 67, 68), that level of somatic genetic alterations cannot explain this series incidence of TSG abnormalities, especially for CCH foci with alterations in two loci (CCH-4 and CCH-8). Similarly, somatic missense mutations at codon 918 of RET have been reported in 20% MTC and in a sample of CCH from MEN 2 patients (54). These factors and the preexisting germline mutation of RET may contribute to C-cell tumorigenesis. Those findings suggest that CCHs in MEN 2A are intraepithelial neoplasias.
In conclusion, genetic (monoclonal proliferation with multiple TSG abnormalities) and kinetic (down-regulated apoptosis) evidences suggest that MEN 2A CCHs are intraepithelial neoplasias. Concordant XCI pattens in both lobes support early clone expansions preceding divergence of C-cell precursors of each thyroid lobe; this may represent a paradigm for other germline mutations during embryogenesis.
Footnotes
This work was presented in part as abstract forms at the United States and Canadian Academy of Pathology (USCAP) Meetings in Orlando, Florida (March 1997), and Boston, Massachusetts (February 1998). This work was performed at the Departments of Pathology, Tufts University–New England Medical Center (Boston, MA) and St. Bartholomew’s and the London NHS Trust (London, UK).
Abbreviations: AR, Androgen receptor; CCH, C cell hyperplasia; HPF, high-power field; ISEL, in situ end labeling; LOH, loss of heterozygosity; MEN 2, multiple endocrine neoplasia type 2; MTC, medullary thyroid carcinoma; NF1, neurofibromatosis 1; TP53, tumor protein p53; RB1, retinoblastoma; SNP, single nucleotide polymorphism; TSG, tumor suppressor genes; WT1, Wilms’ tumor 1; XCI, X-chromosome inactivation.
Received April 12, 2000.
Accepted April 2, 2001.
References
signaling complex in a
kindred with vesicoureteral reflux. Hum Genet 102:474–478[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  L Foppiani, F Forzano, I Ceccherini, W Bruno, P Ghiorzo, F Caroli, P Quilici, R Bandelloni, A Arlandini, G Sartini, et al. Uncommon association of germline mutations of RET proto-oncogene and CDKN2A gene Eur. J. Endocrinol., March 1, 2008; 158(3): 417 - 422. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Saggiorato, I. Rapa, F. Garino, G. Bussolati, F. Orlandi, M. Papotti, and M. Volante Absence of RET Gene Point Mutations in Sporadic Thyroid C-Cell Hyperplasia J. Mol. Diagn., April 1, 2007; 9(2): 214 - 219. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. B. de Groot, T. P. Links, J. T. M. Plukker, C. J. M. Lips, and R. M. W. Hofstra RET as a Diagnostic and Therapeutic Target in Sporadic and Hereditary Endocrine Tumors Endocr. Rev., August 1, 2006; 27(5): 535 - 560. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Blanes, J. J. Sanchez-Carrillo, and S. J. Diaz-Cano Topographic Molecular Profile of Pheochromocytomas: Role of Somatic Down-Regulation of Mismatch Repair J. Clin. Endocrinol. Metab., March 1, 2006; 91(3): 1150 - 1158. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Marx and W. F. Simonds Hereditary Hormone Excess: Genes, Molecular Pathways, and Syndromes Endocr. Rev., August 1, 2005; 26(5): 615 - 661. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Machens, P. Niccoli-Sire, J. Hoegel, K. Frank-Raue, T. J. van Vroonhoven, H.-D. Roeher, R. A. Wahl, P. Lamesch, F. Raue, B. Conte-Devolx, et al. Early Malignant Progression of Hereditary Medullary Thyroid Cancer N. Engl. J. Med., October 16, 2003; 349(16): 1517 - 1525. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Pozo-Garcia, S. J. Diaz-Cano, B. Taback, and D. S.B. Hoon Clonal Origin and Expansions in Neoplasms: Biologic and Technical Aspects Must Be Considered Together Am. J. Pathol., January 1, 2003; 162(1): 353 - 355. [Full Text] [PDF]
|
|
|
|
 |
|
 H Chen, S Hirota, K Isozaki, H Sun, A Ohashi, K Kinoshita, P O'Brien, L Kapusta, I Dardick, T Obayashi, et al. Polyclonal nature of diffuse proliferation of interstitial cells of Cajal in patients with familial and multiple gastrointestinal stromal tumours Gut, December 1, 2002; 51(6): 793 - 796. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
