• Potential conflict of interest: Nothing to report.

We thank Chen et al. for their interest in our study1 identifying microRNA (miR) 19b as an inhibitor of hepatic stellate cell (HSC)-mediated fibrogenesis. We also welcome the opportunity to address reported discrepancies in miR profiles between quiescent and activated HSCs. Although Chen et al. state the same approaches were used in several studies,1-5 resulting in different miR profiles, several noticeable differences in experimental designs exist. Each of the above-mentioned studies used different platforms for miR profiling: 1) Exiqon miRCURY LNA (spotted) array v. 8.03; 2) Ncode Multi-Species miR (spotted) microarray V25; 3) Agilent Rat miR (spotted) microarray (version not specified)1; 4) Affymetrix GeneChip miR 1.0 (oligo) array2; 5) Quantitative reverse-transcription polymerase chain reaction (qRT-PCR).4 Moreover, not all miR assays interrogate the same set of miRs. Additionally, expression miR profiles should be verified by qRT-PCR. For further verification, an in vivo assessment by in situ hybridization would strengthen the findings demonstrated in our study.1 Another technical difference likely to introduce variability is methods and criteria for sample size and data and statistical analyses, which are not always reported.

In addition to methodology/technical differences, the day of HSC harvest is important, and it is clear there are variations in what is considered quiescent and activated cells. Although Chen et al. state that all studies used the same quiescent and activated HSCs, in actuality this was not the case. Among the studies described in this Letter to the Editor, variations in the state of quiescence and activation were reported. Some studies reported freshly isolated, day 2 or day 3 (culture-activated) as a quiescent cell population. Similarly, the activated state of HSCs varied as well; day 7, 10, and 14 (culture-activated) were used. We have shown previously that daily variations in gene expression exist throughout the in vitro HSC transdifferentiation/activation process.6 Our data indicated the specific day of culture activation used can significantly impact studies examining differences between selected populations of quiescent and activated cells. Thus, it is not surprising that different miR array profiles were generated among studies considering the variation in what researchers deemed quiescent and activated. Furthermore, several profiling experiments using miR arrays have been conducted comparing quiescent and in vivo activated HSCs (i.e., carbon tetrachloride [CCl4]7, 8 or bile duct ligation [BDL]9 fibrotic models). Therefore, taking into account differences observed in gene expression profiles between in vitro and in vivo activated HSCs,10 miR array analyses may also produce disparate profiles depending on the activation process selected (i.e., in vitro versus in vivo activation).

Overall, the authors should also note that while some variation does exist in the number/identity of miRs that are differently expressed in quiescent versus activated HSCs, by and large each of the referenced array studies as well as other recent publications do seem to support each other's findings (supported by your table). We hope our response has offered some clarity and insight on miR expression profile discrepancies in HSCs that exist among different research groups/studies.

Ashley M. Lakner Ph.D.*, Nury M. Steuerwald Ph.D.*, Laura W. Schrum Ph.D.*, * Carolinas Medical Center, Charlotte, NC.