Differential growth and transcriptomic profile of stem cell-derived midbrain astrocytes

Reviewer #1 (Public Review):

Summary:

In a previous study, the authors developed a human iPS cell line which expresses Cre under the control of the Lmx1a promoter in order to track, select for, and differentiate human dopamine neurons. In the manuscript under review, the authors are using methods which they have modified to generate astrocytes from the same cell line. The authors are interested in examining astrocytes which are derived from regionalized, floor plate progenitors.

The fundamental weakness of this paper is that the authors are making arguments about regional identity but their work is limited to experiments in vitro. Some of the claims that the authors make should be tested in vivo - ie, in sections, at least. Are floor plate markers or other ventral markers ever expressed in astrocytes or glial progenitors in the mammalian fetus? When do astrocytes emerge in the floor plate? All of the data here are based on an overly simplified in vitro platform.

Lmx1a expression is not limited to the ventral midbrain; it is also expressed in other parts of the developing, ventral CNS and in the roof plate and dorsal CNS (Millonig et al, Nature 2000). Indeed, many of the phenotypes of the Lmx1a mutant mouse (dreher) have little to do with the ventral midbrain. The authors are making an assumption that regional identity is fixed when they begin their astrocyte differentiation protocol - not necessarily true. After astrocytic differentiation is initiated, the authors have done little to demonstrate that floor plate identity is maintained even in selected cells; in fact, the transcriptomic data suggests that the cells are released from a floor plate fate. The authors seem to realize this but do not make any attempt to prove their thesis. If regional identity is not maintained, the authors need a better experiment.

If regional identity is not maintained, so what? Don't we already know that this can happen? The authors acknowledge that this is known in the discussion.

The authors have done transcriptomics studies to follow the changes in these cells but they have not told us very much that is meaningful. It would be useful to validate some of the new astrocytic markers that they have identified - Pax and Irx genes (Welle et al., Glia 2021) come quickly to mind. What about genes related to Shh and Wnt signaling that are prevalent in the floor plate? In particular, a lot of work has been done examining the role of Shh on the properties and lineage of astrocytes (Farmer et al., Science 2016; Hill et al., eLife 2019; Gingrich et al., Neural Dev 2022; Xie et al., Cell Rep 2022). There are a lot of stones which remain unturned, here, and the authors could actually tell us much more without doing an immense amount of work. These suggestions and criticisms are described in far greater detail in the confidential comments to the authors.

Work Cited:

Chizhikov et al., Mamm Genome 2006. https://pubmed.ncbi.nlm.nih.gov/17019651/

Chizhikov et al., Development 2004. https://journals.biologists.com/dev/article/131/11/2693/42269/Control-of-roof-plate-formation-by-Lmx1a-in-the

Chizhikov et al., PNAS 2010. https://pubmed.ncbi.nlm.nih.gov/20498066/

Emsley and Macklis. Neuron Glia Biol 2007. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1820889/

Farmer et al., Science 2016. https://pubmed.ncbi.nlm.nih.gov/26912893/

Gross et al., Development 2016. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4958331/

Hill et al., eLife 2019. https://pubmed.ncbi.nlm.nih.gov/31194676/

Gingrich et al., Neural Dev 2022. https://pubmed.ncbi.nlm.nih.gov/35027088/

Iskusnykh et al., eLife 2023. https://elifesciences.org/articles/84095

Millonig et al, Nature 2000. https://pubmed.ncbi.nlm.nih.gov/10693804/

Welle et al. Glia 2021. https://pubmed.ncbi.nlm.nih.gov/36342840/

Xie et al., Cell Rep 2022. https://pubmed.ncbi.nlm.nih.gov/35196485/

Reviewer #2 (Public Review):

In the current manuscript Li et al., study the preservation of the regional identity during the process of astrocyte generation from pluripotent stem cells. More precisely, this work investigates if neural progenitor cells patterned for the ventral midbrain give rise to astrocytes with conserved regional specification, which could reflect the astrocytic heterogeneity in the brain. To this end, the authors utilized a previously generated reporter iPSC line in which the expression of introduced blue fluorescence protein (BFP) is subjacent to the activation of LMXA1, a ventral midbrain floor plate marker. The study reports that following a defined patterning protocol based on SHH and FGF8, over 90% of d19 cells, corresponding to a neural progenitor stage, acquired the midbrain floor plate identity. However, during the subsequent astrogenic induction and glial progenitor expansion, this identity is gradually lost, supposedly due to the growth advantage of cells deriving from the residual LMX1A- neural progenitors. Contrariwise, if the LMX1A+ progenitors were purified, regional identity would be maintained throughout the astrocytic generation and incur an early astrogenic switch and maturation of derived astrocytes. By using single-cell RNA sequencing, the authors further identified distinct transcriptomic signatures on the astrocytic progeny of LMX1A- and LMX1A- progenitors.

Strengths and weaknesses:

(1) The main model utilized was engineered from the KOLF2 human iPSC line into an elegant LMX1A-reporter line based on the expression of BFP. This results in an attractive model for studies tracing the fate of LMX1A cells. However, consideration should be given to the fact that the parental line, exhibits a splice disruption in the COL3A1 gene encoding type III collagen (Pantazis 2022, doi:10.1016/j.stem.2022.11.004 ), which has been identified as being enriched in certain ventral astrocytic populations (Bradley 2019, doi:10.1242/dev.170910).

(2) The authors argue that the depletion of BFP seen in the unsorted population immediately after the onset of astrogenic induction is due to the growth advantage of the derivatives of the residual LMX1A- population. However, no objective data supporting this idea is provided, and one could also hypothesize that the residual LMX1A- cells could affect the overall LMX1A expression in the culture through negative paracrine regulation. Therefore, cell cycle or proliferation studies of these cells are needed to prove the authors' assumption. Furthermore, on line 124 it is stated that: "Interestingly, the sorted BFP+ cells exhibited similar population growth rate to that of unsorted cultures...". In the face of the suggested growth disadvantage of those cells, this statement needs clarification.

(3) Regarding the fidelity of the model system, it is not clear to me how the TagBFP expression was detected in the BFP+ population supposedly in d87 and d136 pooled astrocytes (Fig S6C) while no LMX1A expression was observed in the same cells (Fig S6F).

(4) The generated single-cell RNASeq dataset is extremely valuable. However, given the number of conditions included in this study (i.e. early vs late astrocytes, BFP+ vs BFP-, sorted vs unsorted, plus non-patterned and neuronal samples) the resulting analysis lacks detail. For instance, from a developmental perspective and to better grasp the functional significance of astrocytic heterogeneity, it would be interesting to map the identified clusters to early vs late populations and to the BFP status. Moreover, although comprehensive, Figure S7 is complex to understand given that citations rather than the reference populations are depicted.

(5) Do the authors have any consideration regarding the morphology of the astrocytes obtained in this study? None of the late astrocyte images depict a prototypical stellate morphology, which is reported in many other studies involving the generation of iPSC-derived astrocytes and which is associated with the maturity status of the cell.

Author response:

(1) Rationale of the study and key finding

We respectively disagree with Reviewer #1’s comments on ‘the fundamental weakness of this paper … about regional identity ...’. We believe that they misunderstood the rationale and key message of the paper.

The rationale of the study stems from the increasing recognition of the importance of generating ‘regional-specific’ astrocytes from iPSCs for disease modelling, due to astrocyte heterogeneity and their region-specific involvement in disease pathology. Regional astrocytes are typically differentiated from neural progenitors (NPCs) that are ‘patterned’ to the desired fate during iPSC neural induction. While the efficiency is not 100%, it is nevertheless assumed that the initial lineage composition (%) of patterned NPCs is preserved during the course of astrocyte differentiation and hence that the derived astrocytes represent the intended regional fate.

We questioned this approach using genetic lineage tracing with ventral midbrain-patterned neural progenitors as an example. By monitoring astrocytic induction of purified BFP+ NPCs and unsorted ventral midbrain-patterned NPC (referred to as BFP- in the paper, line 154 submitted PDF), we demonstrate that despite BFP+ NPCs being the vast majority (>90% LMX1A+ and FOXA2+) at the onset of astrocytic induction, their derivatives were lost in the final astrocyte product unless BFP+ NPCs were purified prior to astrocytic induction and differentiation.

Our findings demonstrate that iPSC-derived astrocytes may not faithfully represent the antecedent neural progenitor pool in terms of lineage, and that the regionality of PSC-derived astrocytes should not be assumed based on the (dominant) NPC identity. We believe that this finding is important for iPSC disease modelling research, especially where disease pathophysiology concerns astrocytes of specific brain regions.

Reviewer #1 raised several interesting questions concerning floor plate marker expression during astrocytic induction and astrocyte differentiation in normal development. These are important outstanding questions in developmental neurobiology, but they are outside the scope of this in vitro study. Indeed, the approach taken by published PSC-astrocyte studies - such as assigning regional identity of PSC-derived astrocytes based on the starting NPC fate or validating PSC-astrocyte using regional markers defined in the developing embryo - is partly due to our limited knowledge about the developing and mature astrocytes in different brain regions. This knowledge gap consequently restricts a thorough characterisation of the regional identity of PSC-astrocytes in such cases.

(2) LMX1A expression in the brain and LMX1A-BFP lineage tracer line

We thank Reviewer #1 for highlighting the wider expression of LMX1A. We are fully aware of this consideration and hence the thorough examination of PSC-derived ventral midbrain-patterned NPCs by immunostaining and single cell RNA-sequencing in this and a previous study (PMID: 38132179). All LMX1A+ cells produced in our protocol exhibit ventral midbrain progenitor gene expression profiles when compared to dataset obtained from human fetal ventral midbrain.

Some of the comments give us the impression that there might be some confusion regarding the lineage tracing system used in this study. The LMX1A-Cre/AAVS1-BFP line is not a classic reporter line that mark LMX1A-expressing cells in real time. Instead, it was designed as a tracer line that expresses BFP in the derivatives of LMX1A+ cells as well as cells expressing LMX1A at the time of analysis.

(3) Is regional identity fixed?

We feel that Reviewer #1 misunderstood the paper in their comments ‘The authors are making an assumption that regional identity is fixed when they begin their astrocyte differentiation protocol - not necessarily true…’. We in fact pointed out in the paper that expression of LMX1A and FOXA2, a signature of midbrain floor plate progenitors, is lost in our BFP+ astrocytes. In this paper, ‘regional identity’ was loosely used to also refer to lineage identity and genetic traits, not just gene expression. We will consider alternative wording during revision to avoid potential confusion.

(4) Splice disruption in the COL3A1 gene and potential effect on astrocyte differentiation of Kolf2 iPSCs

We thank Reviewer #2 for highlighting the variations in KOLF2C1 hiPSCs and the study by Bradley et al. (2019) on differential COL3A1 expression in some ventral astrocytes. We noted that the progenitors produced by Bradley et al. were NKX2.1+ ventral forebrain cells, rather than the LMX1A+ ventral midbrain progenitors investigated in our study. Our scRNAseq data show that all three populations of astrocytes exhibit low levels of COL3A1 expression. While we will continue to examine astrocyte COL3A1 expression in publicly available gene expression datasets, we feel that a selective impairment in astrocyte differentiation of BFP+ cells is unlikely.

(5) Additional data analysis and validation of potential new markers

We will carefully consider the reviewers’ suggestions on further analysis of our single-cell RNA sequencing dada during revision. Regarding eLife’s assessment of validating differential gene expression in different brain regions, it is worth noting that both BFP+ and BFP- cells mapped to the published midbrain scRNAseq data set (La Manno et al, Cell 2016, PMID: 27716510), supporting their midbrain fate. We agree in principle that all single-cell RNA sequencing findings should be validated by immunostaining. It would be beneficial to experimentally verify that our candidate BFP+ differentially expressed genes indeed mark astrocytes derived from LMX1A+ NPCs in vivo, as opposed to other midbrain NPCs. However, this verification cannot be realistically performed in a human setting, but only in an analogous mouse tracer line.

The current eLife assessment nicely summarised part of our findings, in a sense secondary output of this work. We would appreciate a revised eLife assessment that include the message that iPSC-derived astrocytes, in terms of genetic lineage, can deviate greatly from the starting progenitor pool. We would be very happy to provide further information or clarification if it would be helpful. We are committed to doing our best as authors to enhance reader experience and support the continued success of eLife.