Mice carrying the hypomorphic mutation which reduces the transcription factor GATA1 in megakaryocytes (Gata1low mice), develop by 8-months myelofibrosis a phenotype resembling primary myelofibrosis, the most severe of myeloproliferative neoplasms (Vannucchi et al Blood 2002;100:1123). The high levels of TGF-β expressed by the abnormal Gata1low megakaryocytes that drive myelofibrosis in these mice alter the bone marrow (BM) expression profiling up-regulating expression of the transcription factor c-Jun (Ciaffoni et al BCD 2015; 54:234). Recently, Dr. Weissman reported that c-Jun over-expression in response to TGF-β activation induces in mice fibrosis in multiple organs (Werning et al PNAS 2017, 114, 4757), suggesting that also Gata1low mice may develop multi-organ fibrosis. To test this hypothesis, we compared morphology (by haematoxylin-eosin staining) and fibrosis (reticulin fibres by trichrome Mallory and collagen fibres by Gomory or Sirius Red Picrate staining) of organs from Gata 1low and wild-type (WT) mice at 1-, 8- and 15-months of age (3 mice/group).

With age, the organs from WT mice presented histological abnormalities consistent with the mild one expected for being associated with aging and were seldom positive for fibrosis. By contrast, all the organs from Gata1low mice analysed had profound abnormal morphologies with reticulin or collagen fibers detectable, in addition to BM and spleen, in skin, lung, and kidney in an age-specific fashion (Fig 1). Gata1lowskin was thicker than normal and the connective layer presented numerous reticulin strikes already by 1-month and heavily dense connective regions strongly positive for collagen fibres by 8-15 months, resembling the skin from scleroderma patients. In lung, the alveoli had thickened walls with reticulin fibers detectable by 8-months and collagen bundles by 15-months near the bronchus walls. The abnormal morphology of kidney included reduced numbers of glomeruli and poorly organized cortical parenchyma. Reticulin and collagen fibres were observed in the medullary and nephron region by 8- and 15-months, respectively. In heart, cardiomyocytes presented a strong reduction of intercalary disks by 8-months and reticulin fibers were detectable at 15-months, suggesting that this mild fibrosis is driven by lung insufficiency. The liver presented abnormal localization and morphology of hepatocytes and presence of extramedullary hematopoiesis in perisinusoid areas. Sirius Red Picrate staining revealed few reticulin fibres mostly within erythroid islands.

To clarify the mechanisms leading to multi-organ fibrosis in Gata1low mice, the transcription signature of 8-months Gata1low BM was compared with that published for murine liver, lung and kidney fibrosis (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? token = ijgpwgsyjhwpdoj&acc = GSE89630). Gata1low BM presented abnormal expression of 1724 genes (821 up- and 903 down). Pathway analyses indicated that most of the down genes were in the Gata1 pathway while the most prominent up genes were in the c-Jun, EZH2, SCL and p53 pathways. Some of the gene abnormal in Gata1low BM were also abnormal in liver (64), lung (29) and kidney (432) fibrosis. Only 28 genes were abnormal in more than one organ (25 liver and kidney, 7 lung). These common genes were obvious markers for fibrosis (5 collagen genes, fibronectin, TGF-β) and did not include c-Jun. Among genes up in Gata1low BM, kidney and liver there was lipocalin 2 (LCN2), a growth factor overexpressed in primary myelofibrosis patients (Lu et al Blood 2015;126:972) that exerts positive and negative effects, respectively, on fibrosis in BM and liver. By comparing the plasma levels of LCN2 in Gata1low and WT mice at 8-9 -and 15-17-months (8-29 mice/group), we determined that levels of LNC2 do not change with age in WT mice but increase by 3-fold in old Gata1low mice (p=0.028), explaining why liver was not one of the organs in which the mutants develop fibrosis.

In conclusion, in addition to myelofibrosis, Gata1low mice develop fibrosis in skin, lung and kidney but not in heart and liver and represent genetic models for studies on the pathogenesis of fibrosis in multiple organs. Moreover, these results suggest that although the initiation factor(s) for fibrosis in the various organs are likely different, most of them are expressed by Gata1low megakaryocytes highlighting the importance for studies on the secretome profile of these cells.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.