Asahina and colleagues 1 have to be congratulated for the very interesting data and for the enormous amount of work they spent to strengthen their conclusions. The authors demonstrate that fetal mouse liver contains at least three populations of mesenchymal cells.
(A) The mesothelial cells which are desmin-positive but negative for smooth muscle actin-alpha (SMA␣).
(B) The perivascular mesenchymal cells which are desmin and SMA-positive with single desmin-positive SMA-negative cells.
(C) The parenchymal mesenchymal cells (hepatic stellate cells [HSCs]) which are desmin-positive and SMA␣-negative with single exceptions (SMA␣-positive and desmin-positive).
The latter cells were enriched by fluorescence-activated cell sorting for HSCs. Real-time polymerase chain reaction analysis and complementary DNA microarray analysis using total RNA from enriched cells confirmed the expression of the expected genes. Authors selected activated leukocyte cell adhesion molecule (ALCAM)-positive mesothelial cells and cultured them under different conditions. All cells became SMA␣-positive (myofibroblastic phenotype). ALCAM-expressing cells were cultured on collagen in the presence of retinol and palmitic acid. These cells showed the capacity to develop fat droplets, suggesting that they could acquire the HSC phenotype. Because we could make an observation similar to that mentioned under cell population (B) using rat fetal livers, 2 we would like to ask following questions:
(1) Could the single desmin-positive and SMA␣-positive cells of the parenchyma (in Fig. 2E,F of Asahina et al.) represent the cells of the arterial capillary of the sinusoids?
(2) Did the enriched HSC population contain SMA␣ RNA?
(3) Is it possible that the desmin-positive, SMA␣-negative cells (HSCs) do not survive in culture?
(4) If they survive (ALCAM-high) acquiring the myofibroblastic phenotype, does the SMA␣ gene expression decrease when they are cultured on collagen in the presence of retinol and palmitic acid, and does ALCAM remain positive (immunocytology)?
(5) Is it possible to make similar experiments using enriched HSCs?