Embryonic/fetal-type globin genes are known to be up-regulated in models of adult erythropoiesis following exposure to short-chain fatty acids (SCFAs, butyrate or propionate), in a subset of beta-globin gene disorder patients treated with SCFAs, and in children with metabolic disorders that result in high levels of endogenous SCFAs. Here, we investigate a role for SCFAs during development, using a murine embryonic stem (ES) cell model.

Globin gene expression and cell-surface marker presentation were examined in murine embryoid bodies (EBs), which differentiate from ES cells and which recapitulate many features of early murine development. ES cells were allowed to differentiate (through the removal of both LIF and contact inhibition, without cytokines), in the absence or presence of SCFAs. Expression of alpha- and beta-type globin gene RNA and transcription factor RNA in the developing EBs was analyzed daily or every-other-day for 10–14 days by real-time PCR with gene-specific primers and probes. At day +2 and +6 /7 EBs were disaggregated, labeled, and analyzed by fluorescent-activated cell sorting (FACS) with labeled antibodies that had been raised against cell-surface markers of hematopoietic and endothelial differentiation.

Globin gene expression was augmented by SCFAs in all experiments, with maximal induction, when compared with untreated cells, at approximately day + 7 of differentiation; Comparative globin gene expression, between propionate-exposed and untreated, in 3 separate EB experiments and normalized to 18S expression, was variable, but was elevated for all globin genes examined, as shown in Table I. Absolute RNA expression, normalized to 18S in a representative experiment testing butyrate or propionate as inducers, indicated that alpha, zeta, and beta H1/0 were effected most by SCFAs, (in which expression ranged between 30- and 40-fold relative to 18S) while Beta major and Epsilon-y (ranging between 1- and 7-fold relative to 18S) were effected to a much lesser extent; This trend, in which alpha, zeta, and Beta H1/0 induction by SCFAs was more dramatic than that seen for Epsilon-y and beta major induction, was noted in all experiments.

In a single analysis, RNA for BMP-4 and GATA-2 was up-regulated on D + 7 by SCFAs at, respectively, 7- and 18- fold, which was not true for other hematopoietic transcription factors examined, such as LMO-2, GATA-1, EKLF-1, or AML-1, FACS analyses of untreated vs. SCFA-treated EBs in three separate experiments were analyzed with a screening gate for hematopoietic cells (CKit). These experiments showed a modest increase in the hematopoietic marker CD34 in SCFA-treated vs. un-treated cells; a doubling in CKit-positive calls that expressed late erythroid markers on Day + 6, from 5.6 +/− 2% to 12.1 +/− 6.0 % TER-119 positive (P<.05), and a 3-fold, but variable increase, from 1.4 +/− 1.5 % to 4.2 +/− 4.1 % EPO-Receptor positive (n.s.). Expression of CD 45, CD 41, PECAM, and FLK-1 was not consistently different between EBs differentiated in the absence or presence of SCFAs.

These data suggest that SCFAs minimally enhance hematopoiesis, but markedly enhance erythropoiesis, during differentiation in a murine developmental model; we speculate that SCFAs and related compounds could play a similar role in vivo during development.

Range of globin gene expression, SCFA- vs. un-treated EBs

Globin GenesFold-expression
Alpha 5- to 22-fold 
Zeta 4.2- to 35-fold 
Beta H1/0 5- to 22.5-fold 
Epsilon-y 9-to 96-fold 
Beta Major 2.5- to 22-fold 
Globin GenesFold-expression
Alpha 5- to 22-fold 
Zeta 4.2- to 35-fold 
Beta H1/0 5- to 22.5-fold 
Epsilon-y 9-to 96-fold 
Beta Major 2.5- to 22-fold 

Author notes

Corresponding author

Sign in via your Institution