Transdifferentiation involves the conversion of one mature somatic cell type into another mature somatic cell type, without transition through a pluripotent stage. As such, transdifferentiation may involve driving gene expression programs by ectopic transcription factor expression and/or the use of growth factors and chemicals to promote differentiation into a specific cell type. Understanding the control of cell fate switches is becoming increasingly important in the field of disease modeling, particularly for diseases of aging and for cell types that are difficult to obtain via biopsy. Consequently, neurological research will benefit from developments in transdifferentiation approaches. However, caution should be applied in the interpretation of data using these cells for a variety of reasons. The following mini-review outlines transdifferentiation methods, as well as strengths and limitations pertinent to this rapidly advancing field.

Patient live cell-based assay platforms, combined with exogenous stressors/stimuli to emulate disease factors, offer an effective avenue to investigate disease or drug mechanisms, with human relevance and without animal experimentation. Understanding neurological disease factors across a myriad of genetic or environmental factors with larger scale human population studies not only assists distinguishing causative factors from untreatable hallmarks, but also accounts for inter-individual variability and pharmacogenetic factors. The advent of routine reprogramming of somatic cells into induced pluripotent stem cells (iPSCs), using lentiviral, transgene-free and cDNA-based protocols, has enabled the routine conversion of non-invasively acquired cells via pluripotent stem cells to cell types that are not otherwise readily accessible. Advances in efficient microplate, microfluidic (Beers et al., 2015; Gagliano et al., 2019) reprogramming and cost-effective maintenance methods (Kuo et al., 2020) can reduce the significant cost and time demand of iPSC generation and handling. However, variability can arise from iPSC lines as a result of selection and handling (Volpato and Webber, 2020), while growth factor differentiation methods can vary in outcomes with cell culture density (Tonge and Andrews, 2010; Julia et al., 2017). Experiments using panels of cell lines can thus cause variation in differentiation outcomes, which can lead to inconsistencies in downstream assays (Volpato and Webber, 2020).

Complementary to reprogramming of somatic cells, the transdifferentiation of human patient skin or blood biopsy samples offers a simple means to generate neurologically relevant cell types (Figure 1).

Transdifferentiation has previously been carried out using fibroblasts or peripheral blood mononuclear cells (PBMCs) to neural stem cells, neurons, astrocytes, oligodendrocytes, microglia, and skeletal muscle cells (Table 1). Direct differentiation is not limited to a terminally differentiated endpoint; for neurogenesis, reprogramming of fibroblasts to multipotent neural precursor cells can be performed with the overexpression of transcription factors SOX2, OCT3, KLF4, and MYC (Meyer et al., 2014), SOX2 and PAX6 (Maucksch et al., 2012), or even a single transcription factor, such as PTF1A (Xiao et al., 2018).

Along with many studies utilizing iPSCs in recent years as a patient derived disease modeling source, are studies indicating that the reprogramming process and subsequent clonal selection erases age-accumulated disease phenotypes, thus complicating their comparison with patient tissue studies (Mertens et al., 2018). An advantage of direct conversion methods is the retainment of age-accumulated factors, such as epigenetic DNA methylation, DNA damage lesions, heterochromatin organization, senescence associated β-galactosidase activity, and proteostasis, which are re-set when reprogramming cells to a pluripotent state (Mertens et al., 2015, 2018; Tang et al., 2017). Additionally, the reprogramming process results in heterogenous telomerase activation (Huang et al., 2011; Allsopp, 2012).

Tang et al. (2017) assessed age-related cellular hallmarks in motor neurons transdifferentiated from old and young human dermal fibroblasts, in comparison to motor neurons derived from iPSCs. The method of motor neuron differentiation from iPSCs was the same as used for transdifferentiating human fibroblasts, i.e., exogenous expression of the motor neuron-specifying factors, neurogenin-2 (NGN2), SOX11, islet-1 (ISL1), and LIM homeobox 3 (LHX3). Markers of heterochromatin and nuclear organization (LAP2α, H3K9m3 and HP1γ) were found to be reduced in induced motor neurons derived from the older fibroblasts, relative to motor neurons derived from younger fibroblasts or iPSCs. Another study using donors ranging from 3 days old to 96 years old, demonstrated that conversion of fibroblast samples to neurons using neuronal microRNAs (miRNAs) retained numerous age-associated features, including epigenetic CpG loci methylation signatures, elevated oxidative stress, DNA damage and telomere shortening (Huh et al., 2016). Similarly, age-associated factors observed in induced neurons from fibroblasts from aged donors but not iPSCs include increased DSB repair lesions (elevated γH2AX foci), and dysfunctional nuclear pore complex transport with poor nuclear-cytoplasmic localization and export correlated with loss of RanBP17 expression (Mertens et al., 2018). Striatal neurons generated by miRNA-based direct neuronal conversion of Huntington’s disease patient dermal fibroblasts retained age related phenotypes and spontaneously formed nuclear huntingtin protein aggregates, whereas neurons from iPSCs were aggregate-free (Victor et al., 2018). Given that many neurodegenerative disease phenotypes emerge with aging, the approach of transdifferentiation is complementary with reprogramming for the purposes of neurological disease modeling.

The transdifferentiation of neural cells from patient blood cells affords similar investigative opportunities as iPSC-derived neural cells, including the study of neurological mechanisms in health and disease, drug screening and neurodevelopment (Li L. et al., 2018). Multiple laboratories have demonstrated the potential to transdifferentiate human blood cells into various neural cell types, including neurons, microglia, astrocytes and oligodendrocytes, which may also facilitate the development and study of more complex 3D culture models, in addition to the study of individual neural cell types or co-cultures of different cell types. One seminal study reported the generation of CD34⁺ neural progenitor cells from adult peripheral blood and cord blood, which could be subsequently differentiated into dopaminergic or nociceptive neurons, astrocytes or, in very limited cases, oligodendrocytes (Lee et al., 2015). The functional characterization of either glial subtype was not explored in this study. However, the nociceptive neurons were used to develop a model of chemotherapy-induced neuropathy suitable for drug screening. Moreover, human adult peripheral blood T cells, following IL-2 and CD3/CD28 receptor stimulation, can be transdifferentiated directly into neurons, which bypassed the need to generate neural progenitor cells and which were capable of forming action potentials and immature but functional synapses (Tanabe et al., 2018). IL-2 and CD3 receptor stimulation has also been used to expand human adult PBMCs that can be transdifferentiated into photoreceptor-like cells (Komuta et al., 2016), allowing for future study of the retina and visual neurosensory system. In relation to glia, human adult PBMCs can be transdifferentiated into microglia with an immune phenotype characteristic of resident microglia, including phagocytic activity and tumor necrosis factor-α release (Ohgidani et al., 2014). Notably, this study revealed that microglia transdifferentiated from monocytes from a patient with Nasu-Hakola disease had an altered inflammatory cytokine profile, compared to microglia from a healthy subject, highlighting the potential of this approach to study microglial phenotypes in neurological disorders. The possibility remains that astrocytes can also be transdifferentiated from human blood cells. A novel population of CD45⁺ peripheral blood cells, termed insulin-producing cells, can give rise to astrocyte-like cells (Li et al., 2015) but the functional activity of these cells is yet to be reported. Nevertheless, this study is reminiscent of early studies with human blood monocyte-derived microglia (Etemad et al., 2012; Noto et al., 2014) that contributed to the aforementioned transdifferentiation of microglia from monocytes (Ohgidani et al., 2014). Finally, human CD34⁺ cells, obtained from granulocyte colony stimulating factor (GM-CSF)-mobilized peripheral blood, can be transdifferentiated to oligodendrocytes (Venkatesh et al., 2015). Although functional characterization of these cells was lacking, the future development of such cells could be considered as a transplant option to remyelinate neurons in neurological disorders, such as multiple sclerosis, as well as in spinal cord injuries. These studies demonstrate the utility of blood cells for neural cell generation for neurological disease modeling.

Chemical transdifferentiation methods offer a simple approach for modulation of relevant pathways for direct conversion. Commonly antagonized pathways for enhancement of neuronal differentiation include the bone morphogenetic protein (BMP) pathway, (e.g., commonly used inhibitors dorsomorphin, LDN193189), pro-proliferative Notch pathway (e.g., DAPT), TGF-β signaling (e.g., RepSox), and inhibition of GSK-3 kinase activity leading to WNT activation (e.g., CHIR99021), often combined with elevation of cAMP levels (e.g., dibutyryl-cAMP, forskolin) (Crawford and Roelink, 2007; Hu et al., 2015). Such cocktails have been demonstrated to convert human fibroblasts to functional forebrain glutamatergic neurons within 2 weeks (Yang et al., 2019). PBMCs have been demonstrated to transdifferentiate to functional microglia within 2 weeks in the presence of the cytokines, GM-CSF and interleukin (IL)-34 (Ohgidani et al., 2014).

Transcription factor over-expression methods typically include transgene delivery by plasmid cDNA, membrane permeable factor conjugated protein (e.g., human immunodeficiency virus-1 trans-activator of transcription, TAT) (Pan et al., 2010) or viral transduction. Transgenic over-expression enables directed subtype specification, as opposed to retrospective identification of a differentiated subtype, such as the specification of cholinergic motor neurons with the transcription factor NGN2, in combination with motor neuron transcription factors, ISL1 and LHX3 (Fernandopulle et al., 2018). Over-expression of transcription factors, SRY-box transcription factor 9 (SOX9), nuclear factor IA (NFIA) and NFIB have been investigated for their capacity to differentiate stem cells to astrocytes (Canals et al., 2018; Li X. et al., 2018), as well as direct transdifferentiation from murine fibroblasts (Caiazzo et al., 2015) within 2–4 weeks. To facilitate transgene over-expression differentiation protocols, lentiviral particle delivery systems can be cost effective and simple to set up and perform routinely in basic cell culture laboratories (Newbery et al., 2021). An added advantage of viral transduction-based methods is that poorly transduced high density cell clusters that may provide a source of heterogeneity in a non-selective differentiation protocol are removed following antibiotic selection. However, in application ineffective antibiotic selection can result in proliferative non-transduced cell populations. These issues may be overcome by increasing antibiotic titer or utilizing an alternative antibiotic. In combination with transcription factor over-expression, differentiation methods can also be modulated with small molecules to improve maturity or alter non-homogenous culture outcomes. For example dibutyryl-cAMP (Taylor and Reh, 1990; Vidaltamayo et al., 1996) and DAPT (Crawford and Roelink, 2007) are often included in NGN2-based neuronal differentiation protocols (Zhang et al., 2013; De Santis et al., 2018) given their effect on yield and maturity. Cytotoxic additives, such as cytosine arabinoside or CultureOne can also be included in neuronal differentiations to remove proliferating neural/glial progenitors from a post-mitotic neuronal population (Wallace and Johnson, 1989; Fernandopulle et al., 2018). The addition of BMP pathway inhibition (e.g., dorsomorphin or LDN193189) to neuronal differentiation medium could also potentially influence the fate of “bystander” glial cells from an astrocytic lineage to oligodendrocyte lineage for co-culture studies (Hu et al., 2010).

Although lentiviral vectors can be simple and cost effective to produce, the possibility of an integration-induced phenotype limits their clinical application. Furthermore forced exogenous transcription factor overexpression could also induce artifactual epigenetic signatures (Mollinari and Merlo, 2021). Use of small molecules alone may circumvent these limitations, while the cost of small molecules can become a prohibitive factor to efficient scalability. Furthermore, identification of novel small molecules to replace exogenous factor expression is challenging given their indirect control of transcription factors that control lineage (Mertens et al., 2018). The application of non-viral vectors for direct conversion of fibroblasts has been demonstrated in alternative methods involving polysaccharide cationic polymers, electroporation, and combinations of lipid, polycation and magnetofection, however, these non-viral methods remain challenged with low efficiencies (Gong et al., 2018).

Neurological disease relevant transdifferentiation protocols over the past decade include conversion of fibroblasts or PBMCs to multipotent neural stem cells or functional neurons, astrocytes, microglia, oligodendrocytes and skeletal muscle. Retaining potential age-associated neurological disease factors, these models can be implemented to complement reprogrammed somatic cell studies that explore disease modeling from nascent pluripotent stages. However, transdifferentiation approaches are not without practical setbacks, largely owing to the limited protocols available for specific differentiation outcomes and practical yield.

Similar to the considerations pertinent to the reprogramming of somatic cells to iPSCs, PBMCs are readily accessible from blood, however, have much lower proliferative potential than the more invasively acquired dermal fibroblasts (Panda and Ravindran, 2013). Access to a proliferative dermal cell population may be feasible via hair follicle multipotent neural crest stem cells (Gho et al., 2016; Molina and Finol, 2020). The direct conversion of PBMCs is challenged by yield and low proliferative capacity, for example the transdifferentiation of blood to functional neurons produced a modest yield of 50,000 cells per mL (∼3% conversion rate) (Tanabe et al., 2018). In either case, for studies requiring larger cell populations, it may be more efficient to generate progenitor populations from PBMCs or dermal fibroblasts (Meyer et al., 2014; Lee et al., 2015). The advantage of generating progenitors is that they can be expanded and frozen, which may provide a more efficient and long-term resource for generating differentiated cell types, rather than relying on limited stocks of primary cells.

Challenges in neurological cell-based models (not limited to transdifferentiation protocols), include homogeneity, maturity and subtype specification. Neuronal differentiation by NGN2 over-expression has been observed to generate homogenous MAP2⁺ neuronal populations, but generates neurons with mixed gene expression signatures of central and peripheral nervous system markers that differ depending on duration of over-expression (Lin et al., 2021), while synaptic formation is known to be limited when transdifferentiating from fibroblasts, as opposed to iPSCs or neural precursor cells (Zhang et al., 2013). Microarray analysis of SOX9, NFIA and NFIB astrocytic transdifferentiated fibroblasts show similarities to primary cortical astrocytes, relative to fibroblasts, but with many distinct heterogeneous populations. Additionally, as with many transdifferentiation protocols, further investigation is required to identify the precise anatomical region these cells represent and their maturity, compared to astrocytes in human tissue (Caiazzo et al., 2015).

All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.