Ramalho-Santos et al. (1) and Ivanova et al. (2), comparing the same three “stem cells”— embryonic stem cells (ESCs); neural stem cells (NSCs), referred to as neural progenitor/stem cells (NPCs) in the present study; and hematopoietic stem cells (HSCs)—with their differentiated counterparts, each identified a list of commonly expressed “stemness” genes, proposed to be important for conferring the functional characteristics of stem cells. The ability to capture expression profiles of cells using microarrays offers the possibility of defining a stem cell by its constellation of active genes. An intriguing question, however, is whether the functional commonalities (self-renewal and pluripotency) (3) among stem cells can be defined at the genetic level. Do all stem cells express a similar set of “stemness” genes necessary for their unique properties, or do different stem cells express different sets of genes that confer stemness?
We have independently carried out gene expression profiling of three types of stem or immature progenitor cells: ESCs, NPCs, and retinal progenitor/stem cells, or RPCs (Fig. 1) (4, 5). The intersection of ESC-, NPC- and RPC-enriched genes defined a list of 385 genes that are collectively expressed by all three stem cells (6). It can be inferred that these genes may represent or include putative “stemness” genes. For the approach taken here to be able to define and support the notion of “stemness” genes, however, would also require that very similar sets of genes can be identified regardless of the type of stem cells used. To test the validity of this notion, we have collectively analyzed our results along with those from the studies of Ramalho-Santos et al. (1) and Ivanova et al. (2) (Figs. 2 and 3). To our surprise, a comparison of the three independently derived lists of “stemness” genes showed only one gene (integrin alpha-6) commonly identified in the three studies (Figs. 2A and 3A) (6). This finding raised serious concerns about the conclusions reported in (1) and (2), as was also critically highlighted by Burns and Zon (7).
We then examined whether the same extreme discrepancy was observed for the lists of genes expressed in one specific stem cell type. In marked contrast, there was a very significant overlap in the lists of stem cell–specific genes from the three studies. A total of 332 “ESC-enriched genes” (Fig. 2B) (6) and 236 “NPC-enriched genes” (Fig. 2C) (6) were identified by all three investigators. Statistical analysis (8) showed that these numbers are highly significant (P < 10–8). These results strongly go against major differences in cells and analytical methodologies between groups as an explanation for the lack of overlap in the three lists of “stemness” genes. However, when we computed for “stemness” genes, based on genes commonly expressed in two types of stem cells, we found that the three studies overlap by only 10 genes, with P = 1.4 × 10–4 (Fig. 2D). Therefore, as the number of stem cell types intersected to identify “stemness” genes is increased, the overlap between datasets from different investigators drops dramatically.
To examine why that may be the case, we plotted a comparison of significance scores (Fig. 4) for the different categories of genes derived by each group [fold change (FC) value or lower confidence bound (LCB) score] (9). It is quite apparent that as one increases the number of stem cell types for comparison, the genes that occur in the intersection are genes that show progressively lower differential expression between stem and differentiated cells (Fig. 4). For example, we found that the log of the significance scores for individual ESC or NPC-specific transcripts are typically in the range of 6 to 10 over differentiated progenies in all three groups. Transcripts of “stemness” genes from intersection of two or more stem cells, from all three groups, commonly showed log significance scores in the 2 to 4 level. In contrast to ESC and NPC, for which each study came up with overlapping sets of genes that are uniquely or highly expressed exclusively in each stem cell type, none of the putative “stemness” genes in the three studies were highly expressed genes compared with differentiated cells. These results indicate that the expression of “stemness” genes common to all stem cells, if such genes exist at all, is only relatively elevated compared with differentiated cells. This would create a significant variation in the genes identified by differential expression, amplified by subtle differences in experimental conditions between different studies. We propose this as the most important basis for the extreme discrepancies in the lists of putative “stemness” genes.
Our observation does not rule out the possibility that genes unique to all stem cells which are expressed at low levels may exist. However, it is clear that no one single study can confidently identify the bona fide genes that specify “stemness,” and cross-validation of lists generated independently by different investigators is crucial. For instance, the total of 24 genes commonly identified in two of the studies (Figs. 2A and 3A) is statistically significant and warrants further investigation (6). It is possible also that there are “stemness” genes that have not yet been identified and are not represented in the chips used. Genes that may be important for stem cell functions such as self-renewal but that are also expressed in non-stem cells (for example, Stat3, gp130) are unlikely to be identified by a comparative microarray approach. Another possibility is that different stem cell types may use different gene networks to achieve self-renewal or multipotency (10). Finally, “stemness” genes may only be transiently expressed, so that they are easily missed by comparing two homeostatic states. Further efforts to identify these genes will require different strategies.
In summary, speculations made from independent studies (1, 2) about identity of stemness genes do not hold up when the studies are compared. Our explanation for this also demonstrates the inherent problem of testing the stemness hypothesis using a profiling approach.