Since hypoxia may lead to changes in red cell indices, we asked whether there was a relationship between these indices and severity of lung disease. Pulmonary function, cardiopulmonary exercise data, and red blood cell indices from 277 LAM patients, grouped according to use of oxygen, were analyzed. Patients who used supplemental oxygen intermittently or continuously had higher hematocrit and hemoglobin levels than those who did not. Those using supplemental oxygen continuously also had higher red blood cell counts than patients who did not use oxygen. Red blood cell count was significantly correlated with DLCO for patients not using supplemental oxygen, those on supplemental oxygen, and for both groups combined. Lower resting PaO2 while breathing room air was also associated with higher red blood cell count, hematocrit, and hemoglobin. Resting PaO2 was a marker of lung disease severity, and was significantly correlated negatively with the LAM histology scores, a measure of severity of lung disease that is a predictor of survival or time to transplantation. Based on a multivariate analysis, DLCO was a significant predictor of the hematocrit, hemoglobin, and red blood cell count. SaO2 at peak exercise was also significantly correlated with red blood cell count. To assess the relationship between red blood cell count and rate of progression of lung disease, we correlated red blood cell count with the yearly rate of DLCO decline. Although this relationship did not reach statistical significance for patients not receiving supplemental oxygen, the correlation between red blood cell count and the yearly rate of decline in DLCO was significant for patients who used supplemental oxygen and for all patients combined. Thus, higher red blood cell indices were associated with greater severity of lung disease and greater rate of decline in lung function. Since hypoxia increases red blood cell indices, through increased synthesis of EPO, the accelerated loss of lung function in those with erythrocytosis could be due to a growth-enhancing effect of EPO upon LAM cells. The next question was, therefore, whether EPO effects could be found on LAM lung nodules or on LAM cells in culture. Both epithelioid and spindle-shaped LAM cells in LAM lung nodules exhibited immunoreactivity for the EPO receptor (EPOR) as, to a lesser extent, did normal vascular smooth muscle and endothelial cells. Ca. 90% of cultured cells grown from explanted LAM lungs reacted with the anti-EPOR antibody as did pulmonary artery smooth muscle cells (PASM). Microscopically sections of explanted LAM lungs showed reactivity with antibodies against EPO in LAM lung nodules. Anti-EPO antibodies did not react with proliferating fibroblasts sections of lung tissue from patients with idiopathic pulmonary fibrosis, which is characterized by fibroblast proliferation and large depositions of collagen. However, we observed immunoreactivity of anti-EPO anitbodies with alveolar macrophages and type II pneumocytes in these sections. We then investigated the activation state of EPOR in LAM cells. EPOR is phosphorylated at multiple sites after binding of EPO. Phosphorylation of Tyr479 is responsible for EPOR activation and plays a role in cell proliferation. Although smooth muscle cells reacted weakly with anti-phospho-EPOR(Tyr479) antibodies, staining of alveolar cells was stronger. Greatest immunoreactivity with antibodies against activated EPOR(Tyr479) was observed, however, in cells within the LAM lung nodules and in adjacent type II pneumocytes. These data suggest that the activated EPOR and pathway necessary for EPO-induced cell proliferation are activated in LAM cells. EPOR mRNA was detected in total RNA from LAM cells collected by laser-capture microdissection from LAM nodules, and in total RNA extracted from lung, kidney, PASM, and A549 cells (human lung epithelial adenocarcinoma cells), but EPO was not produced by microdiseccted LAM cells. Because EPO might be associated with the collagen that is abundant in LAM lesions, we next tested the ability of recombinant human EPO to bind collagen or Matrigel. EPO binding to collagen was higher than the binding to Matrigel, suggesting that EPO was associated with the extracellular matrix. We had shown that cells from LAM lung explants that react with antibodies against the membrane-associated CD44v6 molecule have dysfunctional TSC2. Therefore, we assessed the presence of EPO and EPOR in cells separated by fluorescence-activated cell sorting using CD44 and CD44v6 antibodies to collect four cell populations (CD44+/CD44v6+, CD44-/CD44v6-, CD44-/CD44v6+, and CD44+/CD44v6-) from the heterogeneous cultured LAM cells. RNA from each population contained EPOR mRNA, but cultured LAM cells (CD44+/CD44v6+, and CD44-/CD44v6+), like the laser-captured microdissected LAM cells, did not produce EPO, which was produced by CD44+/CD44v6- cells. Thus, although LAM cells contained EPOR and could be sensitive to EPO from non-LAM cells in the lung, they did not produce EPO. Because EPO appears to increase the proliferation of cells lacking the TSC2 gene, its effects on cells with loss of heterozygosity for TSC2 grown from LAM lung explants, were tested. Increases of 50-100% in rates of proliferation were seen with the addition of EPO. Because these cultures were not homogeneous and had non-LAM cells as well a substantial loss of TSC2 heterozygosity, EPO may have affected the proliferation of non-LAM as well as LAM cells. To verify an effect of EPO on rate of proliferation of TSC2-/- cells, we used homogeneous TSC2-/- cells grown from a TSC skin tumor (periungual fibroma). These cells lack tuberin due to inactivating mutations in both TSC2 alleles, as has been reported in lung LAM cells. We found that TSC2+/- and TSC2-/- cells produced EPOR but not EPO. Comparison of gene expression by TSC2+/- and TSC2-/- cells showed an amount of EPOR in TSC2-/- cells three-fold that in TSC2+/-. In the presence of EPO, the TSC2-/- skin tumor cells proliferated at two- to three-times the rate of fibroblasts grown from normal-appearing skin of the same patient, upon which EPO had little effect.This study demonstrated the presence of activated EPOR in LAM lung lesions and in LAM cells in culture, along with the existence, by immunohistochemistry, of an activated EPO signaling pathway, which could lead to LAM cell proliferation and the continued growth of LAM lesions. Further, we showed that EPO promoted the growth of lung LAM cells and TSC2 -/- skin tumor cells. Although we found no evidence for the production of EPO by LAM cells, our data suggest that EPO, produced by either the renal cortex or non-LAM cells in the lung, could enhance LAM cell proliferation and thereby accelerate disease progression. Elevated red cell count, probably in response to hypoxia, was associated with more rapid decline in lung function, establishing a link between the pathophysiology of LAM and activation of signaling pathways that promote LAM cell growth. This finding is unique because it demonstrates a correlation between hypoxia-enhanced erythrocytosis and accelerated loss of lung function.