Reader Comments

Post a new comment on this article

Referee comments: Referee 1 (David M. Sherry)

Posted by PLOS_ONE_Group on 15 Feb 2008 at 17:24 GMT

Referee 1's review (David M. Sherry):

Key questions in understanding the generation of complex tissues, and eventually for harnessing stem cells to repair damaged tissues, include how gene expression in progenitor cells is regulated and how gene expression changes in order to generate specific types of cells. The manuscript by Trimarchi, Stadler and Cepko ("Individual Retinal Progenitor cells Display Extensive Heterogeneity of Gene Expression") examines these questions in retinal progenitor cells (RPCs) using single cell PCR, gene array and analyses of gene expression clustering complemented by in situ hybridization methods to assess gene expression by mouse RPCs at several developmental stages. The results reveal an unexpected and surprisingly high level of heterogeneity in gene expression in RPCs. An interesting new technical contribution is the application of a novel method of gene expression profiling for single cells based on the Fisher's Exact Test. This study is the most comprehensive study of its kind to date and provides a large gene expression database will be an extremely useful tool for further study. The manuscript is well written. The findings are novel, timely, and important and will make a substantial contribution to the fields of retinal development, molecular regulation of cell differentiation, and stem cell biology.

The analyses and interpretation of the data in the manuscript are critically dependent on the appropriate identification of RPCs as RPCs. Considering the heterogeneity of gene expression present in the sample, there is potential for circular identification here. Appropriately, much of the first section of the Results is dedicated specifically to this key issue. The extensive control experiments, especially those combining 3H-thymidine labeling in conjunction with known RPC markers, together with the gene expression cluster analyses, provide confidence that the cells classified as RPCs are, in fact, RPCs. In the hierarchical clustering analysis shown in Supplemental Figure 3, it appears that the RPCs are divided into several subgroups, but how these subgroups may relate to subgroups identified by single cell profiling with Fisher's exact test is not discussed.

The authors do a commendable job of pointing out limitations in knowledge and technique. There are two additional points that deserve specific comment in an appropriate place in the text. The temporal resolution of the data is limited somewhat as the sample is based on three developmental timepoints spaced several days apart, and additional changes in gene expression patterns potentially could occur in the time between samples. A related point that deserves specific discussion is that the sample analyzed does not include RPCs from later in postnatal retinal development when many rods, bipolar cells and Müller cells are generated. It is likely that gene expression patterns in late RPCs will differ from the patterns observed in the earlier RPCs examined here and show additional heterogeneity in gene expression. This issue is particularly relevant given the likelihood that some Müller cells may de-differentiate and proliferate in the damaged retina.

It also would be worthwhile for the authors to comment briefly on some of the genes that do not appear within the co-expression clusters. For example, I was surprised that the glycine transporter 1 (GlyT1) gene did not appear in the amacrine cell gene cluster based on co-expression with TCFAP-2beta, and the brn3 transcription factors were not present in the retinal ganglion cell gene cluster based on co-expression with NF-68.

The absence of spatial resolution in the single cell sample is discussed and is addressed reasonably by the in situ hybridization studies. In several cases, it appears that genes were expressed in a central-to-peripheral gradient or were absent at the peripheral margin of the retina (for example, Figs 1E, 3E, 5B, 7A, 10E,12B and E). Considering the generation cells in a central-to-peripheral pattern over retinal development, these spatial patterns are worth noting in the text.

There is considerable speculation in data interpretation and the functional significance of some of the gene expression patterns observed, but it is appropriate and clearly identified. Some readers will certainly disagree with some of the authors' interpretation and speculation, but these differences in opinion will be healthy and should fuel other studies that in turn will lead to a better understanding of the functional roles of specific genes in retinal progenitor cells and specification of cell fate.

There is extensive referencing to supplemental data, but given the nature of those data this is probably the best way to present it. Critically, several of the Supplemental Tables (Tables 10-14) were missing.

For the purposes of the current report, the in situ hybridization micrographs of retinal sections illustrate the results adequately, although several show imperfections that are often associated with high throughput in situ hybridization methods. It would be extremely helpful to label the retinal layers in the top row of micrographs in each of the figures showing in situ hybridization of retina tissue (Figures 1, 3-5, 7, 9-12). This will be especially important to help readers from outside the retina development community interpret the data. The use of scale bars in the figures should be clarified in the accompanying legend. It generally appears that the micrographs in each column are at the same scale as the micrograph at the top of the column, but in some cases this doesn't appear to be true. For example, Fig. 1D and G do not appear to be at the same scale. The same is true for Fig 3D,G and J and several other panels.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.