Abdominal cystic lymphangioma

Lymphangioleiomyomatosis (LAM) is a disease of unknown etiology that is characterized by the proliferation of abnormal smooth muscle cells (LAM cells) in the lung, which leads to cystic parenchymal destruction and progressive respiratory failure. Recent evidence suggests that the proliferative and invasive nature of LAM cells may be due, in part, to somatic mutations in the TSC2 gene, which has been implicated in the pathogenesis of tuberous sclerosis complex. Here, we describe the clinical and molecular characteristics of LAM, as well as the efforts now under way to understand the genetic and biochemical factors that lead to progressive pulmonary destruction and, ultimately, to lung transplantation or death.

Key words: lymphangioleiomyomatosis; tuberous sclerosis complex

Abbreviations: DLCO = diffusing capacity of the lung for carbon monoxide; HRCT = high-resolution CT; HUMAR = human androgen receptor; LAM = lymphangioleiomyomatosis; LOH = loss of heterozygosity; MMPH = multifocal micronodular pneumocyte hyperplasia; PCR = polymerase chain reaction; TSC = tuberous sclerosis complex

**********

Lymphangioleiomyomatosis (LAM), a rare disease that is characterized by cystic destruction of the lung leading to chronic respiratory failure, is found primarily in women of childbearing age. It is a multisystem disorder and is also associated with abdominal tumors (eg, angiomyolipomas, lymphangioleiomyomas). The lung cysts and abdominal tumors are characterized by the presence of abnormal smooth muscle cells (ie, LAM cells). Epidemiologic, genetic, and molecular studies have demonstrated a link between sporadic LAM and tuberous sclerosis complex (TSC), an autosomal-dominant neurocutaneous disorder with variable penetrance caused by mutations in the TSC1 and TSC2 genes. Here, we discuss the clinical characteristics of LAM, the association of LAM and TSC, the morphologic characteristics of abnormal smooth muscle proliferative lesions, as well as gene and protein abnormalities that are characteristic of both sporadic LAM and LAM in patients with TSC.

CLINICAL CHARACTERISTICS

LAM is characterized by progressive respiratory failure and recurrent pneumothoraces. (1) The clinical characteristics of LAM were investigated in three detailed studies. (1-3) Patients presented with dyspnea due to small airway obstruction and/or chylous pleural effusion, chronic cough, or acute chest pain resulting from pneumothorax. Wheezing and hemoptysis occured less commonly, with 26% of patients having evidence of airway hyperreactivity. Asymptomatic lung disease may be discovered after the diagnosis of abdominal angiomyolipomas or axial lymphatic masses. (4) Symptoms arising from abdominal lesions include flank pain, hematuria, and abdominal distension. Retroperitoneal lymphatic involvement can give rise to significant lymphedema and neuropathies.

Physical examination of the lungs most commonly reveals crackles and decreased breath sounds, which are consistent with parenchymal destruction or chylous effusion. Wheezing is heard in 14% of patients, and clubbing is a rare finding. Cardiac examination may provide evidence of pulmonary hypertension secondary to chronic respiratory failure. Abdominal examination may reveal the presence of ascites or masses, indicating angiomyolipomas or axial lymphatic involvement.

On a chest radiograph, LAM is characterized by an increased number of interstitial markings in the presence of normal lung volumes or hyperinflation. (5) Although rarely seen on chest radiograph, thin-walled cysts surrounded by normal parenchyma, which are pathognomonic of LAM, are easily identified on high-resolution CT (HRCT) scanning of the chest (Fig 1). (6-8) Other diseases, such as eosinophilic granuloma, benign metastasizing leiomyoma, and Birt-Hogg-Dube syndrome, need to be included in the differential diagnosis. In addition, patients with TSC may have nodular lesions that are independent of LAM-related cysts, perhaps representing the multifocal micromodular pneumocyte hyperplasia (MMPH) seen on histopathologic examination. On abdominal imaging, angiomyolipomas and lymphatic masses can be detected in patients with sporadic LAM as well as in those with TSC. In patients with TSC, renal angiomyolipomas are more likely to be bilateral than in sporadic LAM, and involve a greater proportion of the kidney. Renal cell carcinomas are seen in patients with TSC, and renal masses should be investigated appropriately. (9)

[FIGURE 1 OMITTED]

Pulmonary function testing demonstrates an obstructive ventilatory abnormality, with superimposed restriction in 17%o of patients. (2) The diffusing capacity of the lung for carbon monoxide (DLCO) often is decreased out of proportion to any abnormalities in spirometry or lung volumes, suggesting a primary gas-exchange defect. (3) In a recent study, (10) the rate of decline in DLCO, but not in FE[V.sub.1], correlated directly with the histologic severity of disease and inversely with the time to transplantation. Similarly, patients often demonstrate hypoxemia at rest or during exercise, which would not be expected to result from the limited decrease in ventilatory lung function. (10,11) Hypoxemia in the absence of ventilatory or cardiac limitation has been documented by cardiopulmonary exercise testing. Finally, 26% of patients with LAM responded significantly to therapy with bronchodilators, indicating a reversible component in the airways obstruction.

PATHOLOGY

Gross specimens from open lung biopsies or autopsies contain cysts throughout the lung parenchyma that are 0.5 to 9,.0 cm in size. (12) On microscopic examination, discrete foci of abnormal smooth muscle cells abut cystic structures lined with hyperplastic type II pneumocytes. The foci contain centrally located, small, spindle-shaped cells with larger epithelioid cells at the periphery, all of which are arranged in haphazard fashion. (13) In addition to LAM cell foci, patients with TSC may also exhibit MMPH, a feature characterized by ill-defined nodules consisting of alveoli with hyperplastic type II pneumocytes. (14) Despite clinical findings that are consistent with airway hyperreactivity, airway inflammation or invasion by LAM cells has not been demonstrated.

Abnormal smooth muscle cells in LAM cell foci, primarily of the epithelioid subtype, react with HMB45, a monoclonal antibody that reacts with gp100, a melanocyte antigen found in premelanosomes. (12) Although the epitope recognized in LAM cells has not been identified, HMB45 immunoreactivity has become the hallmark of diagnosis. Consistent with their classification as smooth muscle cells, LAM cells react with several smooth muscle-specific antibodies, including those against [alpha]-actin, vimentin, desmin, and smooth muscle myosin heavy chains I and II. (12) By immunohistochemistry, the intensity of staining for several proteins involved in cell proliferation (eg, Bcl-2, MCL-1, c-myc, and proliferating cell nuclear antigen) was greater in LAM cells than in adjacent lung cells. (15) Proliferating cell nuclear antigen immunoreactivity was highest in spindle-shaped cells, suggesting a higher rate of proliferation than that of peripheral epithelioid LAM cells. In tissue sections, apoptosis appeared to be rare among the abnormal smooth muscle cells in the LAM foci. (15) In support of a role for sex hormones in the pathogenesis of LAM, epithelioid LAM cells reacted strongly with antibodies to estrogen and progesterone receptors. In lung biopsies from five patients who had been treated with medroxyprogesterone, estrogen and progesterone receptor immunoreactivity was undetectable, which is perhaps a consequence of the down-regulation of receptors in end-stage disease and/or an effect of drug treatment. (16)

Given the absence of immune cells in LAM cell foci, extracellular matrix destruction and invasion would appear to be due to the abnormal smooth muscle cells themselves. LAM cells were reactive with antibodies directed against matrix metalloproteinases MMP-2 and MMP-9, which is consistent with the presence of proteins involved in extracellular matrix degradation. (17)

EPIDEMIOLOGY

The clinical characteristics of lung disease in patients with TSC and LAM are similar to those in patients with sporadic LAM. (18) As a consequence, an epidemiologic link between LAM and TSC was sought. There are currently approximately 450 women with LAM in North America, but the true incidence is unknown. In contrast, the prevalence of TSC is estimated to be approximately 1 in 6,000, with equal frequency in men and women. (19) In retrospective studies, the prevalence of LAM in patients with TSC was estimated to be < 4%. (20) However, recent studies (21-23) have demonstrated that one third of patients with TSC who had no pulmonary symptoms when screened by HRCT of the chest had pulmonary cysts that were consistent with LAM (Table 1, Group 1). In a prospective study (Table 1), newly diagnosed patients (group 1) had normal lung function and more mild disease on HRCT scans of the chest than did patients with TSC in whom LAM previously had been diagnosed (group 2). (21) In the latter group of patients, FE[V.sub.1] and DLCO were inversely proportional to disease severity, as assessed by HRCT of the chest.

Of 10 men with TSC, none had evidence of LAM, suggesting that, in addition to genetic factors, hormonal environment may play a role in the development of cystic lung disease. Unlike lung cysts, nodules were present equally in male and female patients with TSC, independent of the presence of LAM, suggesting that MMPH and lung cysts may have distinct etiologies. The high prevalence of LAM in patients with TSC, as well as the apparent dissociation of cysts and nodules, indicates that screening for cystic disease by HRCT of the chest may identify patients with TSC who are at risk for the development of LAM-related complications, and who might be candidates for therapy early in the course of their disease.

Although almost all cases have occurred in premenopausal women, LAM has been reported in men and postmenopausal women. (24,25) Because the vast majority of cases occur in women of childbearing age, estrogens are thought to play a role in the progression of the disease. This hypothesis has been supported by case reports (26) that described a worsening of the disease during pregnancy, menstruation, and oral contraceptive use. In a nationwide ease-control study, however, the risk of LAM did not increase with the use of oral contraceptive medications. (27)

MOLECULAR BIOLOGY OF LAM AND TSC

In addition to the epidemiologic association between TSC and LAM, the identification of gene and protein abnormalities in LAM cells and angiomyolipomas has strengthened the hypothesis that the two diseases have a common etiology. Genetic abnormalities were found in the TSC1 and TSC2 loci on chromosomes 9q34 and 16p13, respectively. (28) Mutations in TSC2 are common in patients with sporadic LAM as well as in those with TSC and LAM. In contrast, TSC1 mutations have been found in patients with TSC, but not in those with sporadic LAM. (29,30) The TSC1 gene encodes hamartin, a 1,164-amino acid protein that contains a myosin tail-like, intermediate filament, and adenosine triphosphate synthetase regions, as well as a putative transmembrane domain. These structural domains predict functions that are consistent with a role for hamartin in cytoskeletal rearrangement. (31) The TSC2 gene encodes tuberin, a 1,359-amino acid protein that functions as a guanosine triphosphatase-activating protein for Rapla and Rab5. (32,33) The macromolecular complex that includes hamartin and tuberin is primarily cytosolic, is regulated by phosphorylation, and appears to function in multiple tumor-suppressor roles in cell-cycle control. (34,35)

GLOBAL GENE EXPRESSION IN SPORADIC LAM

To identify and characterize the genes expressed in LAM proliferative lesions (genomics), we sought to develop cultures of cell lines in which to characterize the relevant gene products and potential therapeutic targets (proteomics). Global gene expression in the lungs of patients with LAM was examined using tissue collected at the time of lung transplantation. The foci of abnormal smooth muscle cells were identified in embedded frozen sections and were selected by laser capture microdissection. After the isolation of messenger RNA from LAM lesions, reverse transcriptase polymerase chain reaction (PCR) was performed with specific primers for the transcripts of gp100, [alpha]-smooth muscle actin, and SM22. After the reproducibility of the procedure was established, we assessed global gene expression with filters representing approximately 4,000 genes. The overall gene expression of primary pulmonary artery smooth muscle cells was compared with that of each microdissected LAM sample. The overall gene expression of LAM proliferative regions correlated best with cells from angiomyolipomas, and next best with that of smooth muscle cells, but not of fibroblasts, lung epithelial adenocarcinoma (A549) cells, or the melanoma cell line Malme 3M, which also reacts with the HMB45 antibody. The presence of similarities between LAM cells in lung nodules and angiomyolipomas is consistent with a common origin. These data support a model of metastatic spread of cells, as defined by the presence of identical TSC mutations in cells from the two sites. (30) A further evaluation of specific candidate etiologic genes is currently under way.

LONG-TERM CULTURE AND CHARACTERIZATION OF LAM CELLS

To establish a cell line in which to study the molecular and biochemical nature of LAM cells, samples of lung tissue from patients who have received a diagnosis of LAM were obtained after lung biopsy or transplantation. Small pieces of tissue were implanted on cell culture dishes and were grown in serum-supplemented culture media. To isolate homogeneous cell populations, cells were seeded at low density and were allowed to form colonies, which then were removed following capture in a cloning cylinder. The cell lines obtained appeared to express smooth muscle antigens and reacted with the HMB45 antibody. (36,37) Some isolated cell lines have retained these characteristics for 8 to 10 passages and are currently the subject of genetic, biochemical, and structural studies.

GENETIC CHARACTERISTICS OF LAM CELLS

To determine cell clonality, genomic DNA was prepared from cultured cells or lung tissue, and the methylation patterns of the human androgen receptor (HUMAR) gene were assessed. (38) HUMAR-specific PCR primers amplified HUMAR alleles both in lung cells or in tissue derived from patients with LAM. Restriction analysis was carried out using HpaII, an endonuclease that cleaves only nonmethylated segments. Amplified products were separated on a gel. When the restriction enzyme HpaII was employed to digest nonmethylated regions of the HUMAR, only the methylated allele was amplified, allowing for the assessment of clonality in LAM cells. Other studies including karyotyping and the identification of other markers should help to characterize better the cell lines derived from patients with LAM. (39)

Mutations in the TSC2 gene, as determined by a loss of heterozygosity (LOH) at the TSC2 locus, were previously identified in lung tissue from patients with sporadic LAM. (30) The TSC2 mutations were present in cells from LAM proliferative lesions but not from healthy lungs, indicating the contribution of localized mutagenesis in the pathogenesis of LAM. In addition, mutations were apparently identical in angiomyolipomas and lung lesions from the same patients, suggesting that LAM cells may migrate to lymphatic, abdominal, and/or thoracic organs after arising in a single location. We are currently screening for LOH at the TSC2 locus in LAM cell lines. To analyze for this genetic alteration, samples of DNA from LAM cells and the lungs of patients with LAM were amplified using PCR with specific oligonucleotide primers for microsatellite loci on chromosome 16p13.3. (40,41) An LOH at the TSC2 locus was present in a cell line derived from a patient with sporadic LAM, indicating that cultured cells retain genetic characteristics of the original founder cells. Consistent with a somatic mutagenesis step in the etiology of LAM, there are clonal cell lines from different tissues in the same patient that both have and do not have an LOH at the TSC2 locus.

CONCLUSION

Single base-pair mutations in the TSC2 gene, and a putative loss of tumor-suppressor activity, could, in large part, explain the uncontrolled proliferation of LAM cells. TSC genes play roles in transcription, signal transduction, cell cycle control, control of cell growth and proliferation, and cell adhesion. (31,42-45) The cellular components involved in metastasis, invasion, and angiogenesis also may be altered in the abnormal smooth muscle cells found in pulmonary LAM lesions and angiomyolipomas. In addition to the point mutations previously identified in the TSC2 gene, chromosomal and gene alterations leading to these phenotypes may include aneuploidy, translocations, deletions, and gene amplification. (46) The nature of the macromolecular complex containing hamartin and tuberin, and their interactions with other cellular proteins, is poorly understood. Future findings related to the genomics and proteomics of TSC and LAM using tissue and cell lines derived from patients may shed light on the cellular processes that lead to the LAM phenotype and will help to identify molecular targets for therapeutic intervention.

ACKNOWLEDGMENT: We thank Dr. Martha Vaughan for constructive comments and revision of the manuscript.

REFERENCES

(1) Kalassian KG, Doyle R, Kao P, et al. Lymphangioleiomyomatosis: new insights. Am J Respir Crit Care Med 1997; 155:1183-1186

(2) Chu SC, Horiba K, Usuki J, et al. Comprehensive evaluation of 35 patients with lymphangioleiomyomatosis. Chest 1999; 115:1041-1052

(3) Kitaichi M, Nishimura K, Itoh H, et al. Pulmonary lymphangioleiomyomatosis: a report of 46 patients including a clinicopathologic study of prognostic factors. Am J Respir Crit Care Med 1995; 151:527-533

(4) Merchant RN, Pearson MG, Rankin RN, et al. Computerized tomography in the diagnosis of lymphangioleiomyomatosis. Am Rev Respir Dis 1985; 131:295-297

(5) Carrington CB, Cugell DW, Gaensler EA, et al. Lymphangioleiomyomatosis: physiologic-pathologic-radiologic correlations. Am Rev Respir Dis 1977; 116:977-995

(6) Aberle DR, Hansell DM, Brown K, et al. Lymphangiomyomatosis: CT, chest radiographic, and functional correlations. Radiology 1990; 176:381-387

(7) Avila NA, Chen CC, Chu SC, et al. Pulmonary lymphangioleiomyomatosis: correlation of ventilation-perfusion scintigraphy, chest radiography, and CT with pulmonary function tests. Radiology 2000; 214:441-446

(8) Muller NL, Chiles C, Kullnig P. Pulmonary lymphangiomyomatosis: correlation of CT with radiographic and functional findings. Radiology 1990; 175:335-339

(9) Neumann HP, Schwarzkopf G, Henske EP. Renal angiomyolipomas, cysts, and cancer in tuberous sclerosis complex. Semin Pediatr Neurol 1998; 5:269-275

(10) Taveira-DaSilva AM, Hedin CM, Stylianou MP, et al. Reversible airflow obstruction, proliferation of abnormal smooth muscle cells, and impairment of gas exchange as predictors of outcome in lymphangioleiomyomatosis. Am J Respir Crit Care Med 2001; 164:1072-1076

(11) Crausman RS, Jennings CA, Mortenson RL, et al. Lymphangioleiomyomatosis: the pathophysiology of diminished exercise capacity. Am J Respir Crit Care Med 1996; 153: 1368-1376

(12) Ferrans VJ, Yu ZX, Nelson WK, et al. Lymphangioleiomyomatosis (LAM): a review of clinical and morphological features. J Nippon Med Sch 2000; 67:311-329

(13) Bonetti F, Chiodera PL, Pea M, et al. Transbronchial biopsy in lymphangiomyomatosis of the lung: HMB45 for diagnosis. Am J Surg Pathol 1993; 17:1092-1102

(14) Muir TE, Leslie KO, Popper H, et al. Micronodular pneumocyte hyperplasia. Am J Surg Pathol 1998; 22:465-472

(15) Usuki J, Horiba K, Chu SC, et al. Immunohistochemical analysis of proteins of the Bcl-2 family in pulmonary lymphangioleiomyomatosis: association of Bcl-2 expression with hormone receptor status. Arch Pathol Lab Med 1998; 122: 895-902

(16) Matsui K, Takeda K, Yu ZX, et al. Downregulation of estrogen and progesterone receptors in the abnormal smooth muscle cells in pulmonary lymphangioleiomyomatosis following therapy: an immunohistochemical study. Am J Respir Crit Care Med 2000; 161:1002-1009

(17) Hayashi T, Fleming MV, Stetler-Stevenson WG, et al. Immunohistochemical study of matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs) in pulmonary lymphangioleiomyomatosis (LAM). Hum Pathol 1997; 28: 1071-1078

(18) Castro M, Shepherd CW, Gomez MR, et al. Pulmonary tuberous sclerosis. Chest 1995; 107:189-195

(19) Webb DW, Osborne JP. Tuberous sclerosis. Arch Dis Child 1995; 72:471-474

(20) Ryu J. Tuberous sclerosis complex and LAM. In: Moss J, ed. LAM and other diseases characterized by smooth muscle proliferation. New York, NY: Marcel Dekker, 1999; 407-418

(21) Moss J, Avila NA, Barnes PM, et al. Prevalence and clinical characteristics of lymphangioleiomyomatosis (LAM) in patients with tuberous sclerosis complex. Am J Respir Crit Care Med 2001; 164:669-671

(22) Costello LC, Hartman TE, Ryu JH. High frequency of pulmonary lymphangioleiomyomatosis in women with tuberous sclerosis complex. Mayo Clin Proc 2000; 75:591-594

(23) Franz DN, Brody A, Meyer C, et al. Pulmonary cystic and nodular changes consistent with LAM and MMPH are common in women with tuberous sclerosis [abstract]. Am J Respir Crit Care Med 2001; 163:A559

(24) Aubry MC, Myers JL, Ryu JH, et al. Pulmonary lymphangioleiomyomatosis in a man. Am J Respir Crit Care Med 2000; 162:749-752

(25) Johnson SR, Tattersfield AE. Clinical experience of lymphangioleiomyomatosis in the UK. Thorax 2000; 55:1052-1057

(26) Kelly J, Moss J. Lymphangioleiomyomatosis [review]. Am J Med Sci 2001; 321:17-25

(27) Wahedna I, Cooper S, Williams J, et al. Relation of pulmonary lymphangioleiomyomatosis to use of the oral contraceptive pill and fertility in the UK: a national case control study. Thorax 1994; 49:910-914

(28) Sampson JR, Harris PC. The molecular genetics of tuberous sclerosis. Hum Mol Genet 1994; 3:1477-1480

(29) Jones AC, Shyamsundar MM, Thomas MW, et al. Comprehensive mutation analysis of TSC1 and TSC2- and phenotypic correlations in 150 families with tuberous sclerosis. Am J Hum Genet 1999; 64:1305-1315

(30) Carsillo T, Astrinidis A, Henske EP. Mutations in the tuberous sclerosis complex gene TSC2 are a cause of sporadic pulmonary lymphangioleiomyomatosis. Proc Natl Acad Sci U S A 2000; 97:6085-6090

(31) Lamb RF, Roy C, Diefenbach TJ, et al. The TSC1 tumour suppressor hamartin regulates cell adhesion through ERM proteins and the GTPase Rho. Nat Cell Biol 2000; 2:281-287

(32) Xiao GH, Shoarinejad F, Jin F, et al. The tuberous sclerosis 2 gene product, tuberin, functions as a Rab5 GTPase activating protein (GAP) in modulating endocytosis. J Biol Chem 1997; 272:6097-6100

(33) Wienecke R, Konig A, DeClue JE. Identification of tuberin, the tuberous sclerosis-2 product: tuberin possesses specific Rap1GAP activity. J Biol Chem 1995; 270:16409-16414

(34) Aicher LD, Campbell JS, Yeung RS. Tuberin phosphorylation regulates its interaction with hamartin: two proteins involved in tuberous sclerosis. J Biol Chem 2001; 276: 21017-21021

(35) Nellist M, van Slegtenhorst MA, Goedbloed M, et al. Characterization of the cytosolic tuberin-hamartin complex: tuberin is a cytosolic chaperone for hamartin. J Biol Chem 1999; 274:35647-35652

(36) Yu Z-X, Pacheco-Rodriguez G, Takeda K, et al. Isolation of HMB45 positive smooth muscle cells (LAM cells) from lung tissue from patients of lymphangioleimyomatosis [abstract]. Am J Respir Crit Care Med 1999; 161:A376

(37) Pacheco-Rodriguez G, Stevens LA, Valencia J, et al. Molecular Characterization of Lymphangioleiomyomatosis (LAM): histological profile of the abnormal smooth muscle cells ("LAM cells") derived from explants in long-term cultures [abstract]. Am J Respir Crit Care Med 161:A16

(38) Allen RC, Zoghbi HY, Moseley AB, et al. Methylation of HpaII and HhaI sites near the polymorphic CAG repeat in the human androgen-receptor gene correlates with X chromosome inactivation. Am J Hum Genet 1992; 51: 1229-1239

(39) Kaplan J, Hukku B. Cell line characterization and authentication. Methods Cell Biol 1998; 57:203-216

(40) Shen Y, Kozman HM, Thompson A, et al. A PCR-based genetic linkage map of human chromosome 16. Genomics 1994; 22:68-76

(41) Snarey A, Thomas S, Schneider MC, et al. Linkage disequilibrium in the region of the autosomal dominant polycystic kidney disease gene (PKD1). Am J Hum Genet 1994; 55:365-371

(42) Soucek T, Pusch O, Wienecke R, et al. Role of the tuberous sclerosis gene-2 product in cell cycle control: loss of the tuberous sclerosis gene-2 induces quiescent cells to enter S phase. J Biol Chem 1997; 272:29301-29308

(43) Tapon N, Ito N, Dickson BJ, et al. The drosophila tuberous sclerosis complex gene homologs restrict cell growth and cell proliferation. Cell 2001; 105:345-355

(44) Potter CJ, Huang H, Xu T. Drosophila tsc1 functions with tsc2 to antagonize insulin signaling in regulating cell growth, cell proliferation, and organ size. Cell 2001; 105:357-368

(45) Henry KW, Yuan X, Koszewski NJ, et al. Tuberous sclerosis gene 2 product modulates transcription mediated by steroid hormone receptor family members. J Biol Chem 1998; 273:20535-20539

(46) Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature 1998; 396:643-649

* From the Pulmonary-Critical Care Medicine Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD.

Correspondence to: Joel Moss, MD, PhD, Chief Pulmonary-Critical Care Medicine Branch, National Heart, Lung, and Blood Institute, National Institutes of Health, 10 Center Dr, Bldg 10, Room 6 D05, MSC 1590, Bethesda, MD 20892-1590; e-mail: mossj@nhlbi.nih.gov

COPYRIGHT 2002 American College of Chest Physicians
COPYRIGHT 2002 Gale Group