Lineage tracing: The gold standard to claim direct reprogramming in vivo

Abstract

The discovery that many types of fully differentiated somatic cells can be “reprogrammed” back to a pluripotent state through epigenetic modifications driven by specific transcription factors opened the door for cell alchemy. In theory, any cell could be switched from one fate to another, and it seems logical that a closer relationship between the cells would confer easier reprogramming.1Sareen D. Svendsen C.N. Stem cell biologists sure play a mean pinball.Nat. Biotechnol. 2010; 28: 333-335Google Scholar Astrocytes and neurons derive from a common progenitor, and so it is perhaps not surprising that astrocytes could be turned into neurons in vitro using similar transcription factors, as shown over a decade ago.2Berninger B. Costa M.R. Koch U. Schroeder T. Sutor B. Grothe B. Götz M. Functional properties of neurons derived from in vitro reprogrammed postnatal astroglia.J. Neurosci. 2007; 27: 8654-8664Google Scholar But providing a compelling demonstration of mature adult astrocyte conversion into neurons in the living brain is very different from the tissue culture environment where the astrocyte is often in an immature or even artificial state. There have now been several papers that used retrovirus to deliver single specific neurogenic transcription factors to the brain and claim that endogenous astrocytes can be driven directly to a neuronal fate, albeit at relatively inefficient levels. A recent paper by Chen et al.3Chen Y.-C. Ma N.-X. Pei Z.-F. Wu Z. Do-Monte F.H. Keefe S. Yellin E. Chen M.S. Yin J.C. Lee G. et al.A NeuroD1 AAV-based gene Therapy for functional brain repair after ischemic injury through in vivo astrocyte-to-neuron conversion.Mol. Ther. 2020; 28: 217-234Google Scholar used a more efficient adeno-associated virus (AAV)-based method of delivering a single transcription factor, NeuroD1, and reported much larger-scale conversion of astrocytes to neurons. Remarkably, these newly reprogrammed neurons were also reported to repopulate a cerebral cortex lesion that had been massively depleted of neurons in a rodent stroke model and to grow axonal projections through the adult central nervous system (CNS) to appropriate targets, thereby restoring performance in functional tests. Similarly, a recent paper by Qian et al.4Qian H. Kang X. Hu J. Zhang D. Liang Z. Meng F. Zhang X. Xue Y. Maimon R. Dowdy S.F. et al.Reversing a model of Parkinson’s disease with in situ converted nigral neurons.Nature. 2020; 582: 550-556Google Scholar reported use of AAVs to deplete RNA-binding protein PTB (PTBP1) and directly convert midbrain astrocytes into dopamine neurons, which reconstructed the nigrostriatal circuit and provided recovery in a rodent model of Parkinson's disease. Notably, neither study reported lineage tracing or fate mapping to support their claims, and both studies are generating considerable controversy centered around the question of what is required to provide compelling and rigorous experimental evidence of direct conversion of one cell type to another in the in vivo CNS. Three separate studies are now published or are online challenging the claims by Chen et al.3Chen Y.-C. Ma N.-X. Pei Z.-F. Wu Z. Do-Monte F.H. Keefe S. Yellin E. Chen M.S. Yin J.C. Lee G. et al.A NeuroD1 AAV-based gene Therapy for functional brain repair after ischemic injury through in vivo astrocyte-to-neuron conversion.Mol. Ther. 2020; 28: 217-234Google Scholar and Qian et al.4Qian H. Kang X. Hu J. Zhang D. Liang Z. Meng F. Zhang X. Xue Y. Maimon R. Dowdy S.F. et al.Reversing a model of Parkinson’s disease with in situ converted nigral neurons.Nature. 2020; 582: 550-556Google Scholar Using rigorous lineage tracing strategies, these subsequent studies find no evidence for direct conversion of astrocytes into neurons in vivo, either by delivering NeurD1 or by depleting PTBP1. The first of these studies, by Wang et al.5Wang L.-L. Serrano C. Zhong X. Ma S. Zou Y. Zhang C.-L. Revisiting astrocyte to neuron conversion with lineage tracing in vivo.Cell. 2021; 184: 5465-5481.e16Google Scholar and published in Cell, challenges the assumption of AAV-mediated direct conversion by using lineage tracing experiments where endogenous astrocytes are tagged with a constitutive genetic reporter using Cre-loxP technology prior to attempted conversion. Their data show essentially no direct conversion of astrocytes to neurons expressing the lineage tracing reporter, but instead show aberrant expression of the NeuroD1 transgene used to label “converted” cells in endogenous neurons. Other studies available in bioRxiv used similar lineage tracing strategies and found that depleting PTBP1 was not able to convert astrocytes to neurons in the midbrain or elsewhere6Hoang T. Kim D.W. Appel H. Pannullo N.A. Leavey P. Ozawa M. Zheng S. Yu M. Peachey N.S. Kim J. Blackshaw S. Ptbp1 deletion does not induce glia-to-neuron conversion in adult mouse retina and brain.Preprint at bioRxiv. 2021; (10.04.462784)Google Scholar,7Chen W. Zheng Q. Huang Q. Ma S. Li M. Repressing PTBP1 is incapable to convert reactive astrocytes to dopaminergic neurons in a mouse model of Parkinson’s disease.Preprint at bioRxiv. 2021; (11.12.468309)Google Scholar. These observations raise serious questions about what type of evidence is needed to rigorously demonstrate direct conversion or cell reprogramming. Opposing arguments have been presented in a somewhat contentious exchange of correspondence between the Chen lab8Chen G. In vivo confusion over in vivo conversion.Mol. Ther. 2021; 29: 3097-3098Google Scholar published as a letter in Molecular Therapy and the Wang lab reply published in this issue of Molecular Therapy. A commentary by Calzolari and Berninger9Calzolari F. Berninger B. cAAVe phaenomena: beware of appearances!.Cell. 2021; 184: 5303-5305Google Scholar in Cell accompanying the study by Wang et al.5Wang L.-L. Serrano C. Zhong X. Ma S. Zou Y. Zhang C.-L. Revisiting astrocyte to neuron conversion with lineage tracing in vivo.Cell. 2021; 184: 5465-5481.e16Google Scholar elegantly highlights the key issues and points of controversy. In their letter, Chen et al.8Chen G. In vivo confusion over in vivo conversion.Mol. Ther. 2021; 29: 3097-3098Google Scholar challenge the failure of Wang et al.5Wang L.-L. Serrano C. Zhong X. Ma S. Zou Y. Zhang C.-L. Revisiting astrocyte to neuron conversion with lineage tracing in vivo.Cell. 2021; 184: 5465-5481.e16Google Scholar to replicate experiments with two counterarguments based on (1) AAV dose or (2) non-specificity of Cre-loxP technology. However, these counterarguments do not seem entirely convincing. First, a high dose of virus could theoretically drive aberrant gene expression. But, even if this was the case, there would be a penumbral gradient of gradually declining virus concentration away from the infusion site and, along this gradient, optimal concentrations should exist to induce lineage-traced astrocyte conversion even in the study by Wang et al.5Wang L.-L. Serrano C. Zhong X. Ma S. Zou Y. Zhang C.-L. Revisiting astrocyte to neuron conversion with lineage tracing in vivo.Cell. 2021; 184: 5465-5481.e16Google Scholar There was none. Second, Cre-loxP technology has been used in thousands of lineage tracing studies, many combined with AAV administration with no aberrant expression. Indeed, the arguments against Cre would seem to apply equally, if not more, to the Chen lab studies that used AAV to deliver Cre, which would result in much higher cellular levels compared with the transgenic delivery used by Wang et al.5Wang L.-L. Serrano C. Zhong X. Ma S. Zou Y. Zhang C.-L. Revisiting astrocyte to neuron conversion with lineage tracing in vivo.Cell. 2021; 184: 5465-5481.e16Google Scholar As noted by Calzolari and Berninger,9Calzolari F. Berninger B. cAAVe phaenomena: beware of appearances!.Cell. 2021; 184: 5303-5305Google Scholar the ultimate experiment was the pre-labeling of cortical neurons by retrograde transport and the demonstration that NeuroD AAV was in neurons already present prior to the AAV delivery, ruling out that these were neurons converted from astrocytes. Another point to consider is solely using loss of immunohistochemical NeuN labeling in the Chen lab studies as a measure of presumed neuronal death. This is not a definitive analysis or categorical conclusion. NeuN expression can be down-regulated after cerebral ischemia without death of the neurons, and so manipulations that restore NeuN staining do not necessarily indicate generation of new neurons.10Unal-Cevik I. Kilinç M. Gürsoy-Ozdemir Y. Gurer G. Dalkara T. Loss of NeuN immunoreactivity after cerebral ischemia does not indicate neuronal cell loss: a cautionary note.Brain Res. 2004; 1015: 169-174Google Scholar More definitive markers of neuronal death are available and should be used when claiming to replace experimentally ablated neurons.11Fricker M. Tolkovsky A.M. Borutaite V. Coleman M. Brown G.C. Neuronal cell death.Physiol. Rev. 2018; 98: 813-880Google Scholar Finally, the Chen lab recently published another paper that used lineage tracing and reported conversion of labeled astrocytes to neurons.12Xiang Z. Xu L. Liu M. Wang Q. Li W. Lei W. Chen G. Lineage tracing of direct astrocyte-to-neuron conversion in the mouse cortex.Neural Regen. Res. 2021; 16: 750-756Google Scholar In our view, the supporting images are not compelling. Moreover, a critical experiment using retroviral constructs to label dividing cells (thus avoiding post-mitotic neurons) was not performed with Cre animals to rigorously test whether the new neurons came from mature astrocytes that had divided. Beyond these thoughts we have another discussion to put forth. Regardless of who is right or wrong, is converting astrocytes to neurons really a sensible thing to do in the CNS? Astrocytes are not merely passive bystanders in the nervous system, but rather fully integrated into the circuitry of the brain with exquisite anatomical and physiological functions critical for normal brain function. Once reactive, they are essential for brain repair by monitoring and maintaining homeostasis around neurons. All studies to date that have experimentally depleted mature or proliferating astrocytes from the injured CNS have shown detrimental effects. To directly convert astrocytes into neurons without replacing them seems counterintuitive and is likely to disrupt efficient brain function. Also, it seems unlikely that a single transcription factor is able to reprogram an astrocyte into a regionally specific neuron and fully restore that neuron's complex and often distant connections, which occurs during brain development and requires an amazingly orchestrated series of sequential multifactorial events. When the implications for a new discovery are potentially transformative, as is the case here, the data supporting the argument must be very strong. The most parsimonious explanation for the current argument is that a few reactive and dividing astrocytes may convert to cells with neuronal features with very low efficiency based on retroviral labeling studies (proof of concept that it can occur and resembling in vitro data); however, AAV-based viral delivery has known issues with leaky promoters and specificity and cannot be used to prove lineage. There should be consensus about a rigorous means to demonstrate conversion of astrocytes to neurons in vivo. Given the 20-year history of using linage tracing and fate mapping to establish cell conversion, we contend these methods should be the gold standard to claim direct reprogramming in vivo.

Topics & Concepts