Induction of bone marrow stromal cells to neurons: Differentiation, transdifferentiation, or artifact?

Authors


Abstract

Differentiation of stem cells toward a neuronal lineage normally involves a gradually progressive restriction in developmental potential and is regulated by a diverse set of specific and temporally precise genetic events. However, recent studies have indicated that both rodent and human bone marrow stromal cells (MSCs) can be rapidly (within minutes to hours) induced to differentiate into neurons in vitro by relatively simple chemical means (using β-mercaptoethanol [BME] or dimethylsulfoxide [DMSO] and butylated hydroxyanisol [BHA]; Woodbury et al. [ 2000] J. Neurosci. Res. 61:364–370). The ability to transdifferentiate an easily accessible cell source into neurons could have substantial potential for promoting neural repair. We therefore explored the potential of simple chemical methods to transdifferentiate other cell types, including primary rat fibroblasts, primary human keratinocytes, HEK293 cells, rat PC-12 cells, and as positive control rat bone marrow stromal (BMS) cells. Surprisingly, all cells except for keratinocytes adopted at least partial “neuron-like” pyramidal cell morphology with fine-cellular extensions resembling neurites upon stimulation with BME or DMSO/BHA. However, time-lapse microscopy indicated that the chemical exposure of MSCs did not result in new neurite growth but rather cellular shrinkage, with retraction of the majority of existing cell extensions, leaving only few, fine neurite-like processes. To determine whether the chemically induced transdifferentiation resulted from simple cellular toxicity, MSCs were exposed to various stressors, including detergents, high-molarity sodium chloride, and extremes of pH. In all cases, cellular shrinkage and adoption of pseudoneuronal morphology were observed. Concomitantly with cellular shrinkage, apparent increases in immunolabeling for the neuronal markers NSE and NeuN were detected in the cell soma that could not be confirmed by RT-PCR. Furthermore, blockade of protein synthesis with cycloheximide did not prevent cells from adopting “neuron-like” morphology after chemical induction. Thus, morphological changes and increases in immunolabeling for certain cellular markers upon “chemical induction” of MSCs are likely the result of cellular toxicity, cell shrinkage, and changes in the cytoskeleton and do not represent regulated steps in a complicated cellular differentiation process. © 2004 Wiley-Liss, Inc.

Ancillary