Owing to the types of heterogeneity described earlier, it is currently difficult to obtain “homogeneous” populations of PSCs for proteomic analysis. Consequently, important nuances inherent to the control of self-renewal or commitment are possibly diluted in studies that average an entire culture. Thus, refined analyses on more homogeneous populations will be required to assess the potentially informative protein abundance changes and/or modifications that have been identified through the qualitative and quantitative proteomic analysis of heterogeneous populations so that they may be correlated with a specific functional relevance. The obvious question is how do we obtain more homogeneous populations of PSCs for future detailed analyses? The best example of how this can be achieved is from the HSC system, where cell surface accessible markers are used to identify and isolate functionally defined subsets of bone marrow-derived HSCs and their progeny [61
]. However, a similar “cell surface barcode” [63
] does not exist for the identification and purification of bona fide PSCs. Rather, the analysis of PSCs rely on surface markers like SSEA-1 (mouse), SSEA-3 and SSEA-4 (human), TRA-1-60, and TRA-1-81, none of which are suitable for the isolation of homogeneous PSC subpopulations or lineage-specific progeny [64
]. While it will likely not be possible to identify surface markers specific to every cell type present, the ability to identify and isolate functionally defined subpopulations of PSCs without the use of genetic manipulations would be beneficial to both therapeutic and research efforts.
As in business, a barcode is an optical and readable representation of data that is unique to each product. By analogy, a “cell surface barcode” should be uniquely informative of a biological state for specific cell types. Although the need for a unique classification of cells based on biological markers is well appreciated, the cell surface remains less well characterized than other subproteomes and there are currently only a limited number of candidate surface markers for PSCs. This can be attributed to the challenges faced when analyzing cell surface proteins, including the difficulty in solubilizing transmembrane proteins and enriching for a lesser abundant subproteome. Thus, the application of proteomic technologies that are especially well suited for the study of cell surface proteins is poised to make a significant impact on our understanding of how stem cells communicate with their microenvironment and on the development of new reagents for identifying and isolating PSCs and their progeny.
Creative approaches have been developed to circumvent the challenges in identifying cell surface proteins (selected reviews [66
]). In general, two categories of strategies have emerged, which have shown an ability to identify cell surface proteins more efficiently than the analysis of the whole cellular proteome and are expected to be useful in identifying surface markers on PSCs. These include “physical enrichment” and “affinity enrichment” of plasma membrane-(PM) associated proteins (). Briefly, in physical enrichment strategies, differential centrifugation is used to enrich for membrane proteins and several studies reviewed here have used variations of this approach on mouse and human PSCs [71
]. In affinity enrichment strategies, several types of chemical tags can be used to selectively label accessible regions of surface proteins and these methods are typically combined with physical enrichment strategies. The most common affinity enrichment strategies include protein biotinylation via cleavable or non-cleavable biotin onto accessible lysine residues (reviewed by [74
]), and oligosaccharide biotinylation via hydrazide chemistry (CSC-Technology), first described by Wollscheid et al. [63
]. One advantage of using cleavable biotin (for protein biotinylation) or the CSC-Technology (oligosaccharide biotinylation) is the data “tag” resulting from the cleaved biotin or oligosaccharide that remains on the peptide and can be identified by MS. This facilitates the data analysis by allowing for efficient removal of contaminating, non-surface proteins that may have been enriched during centrifugation, as it is not possible to purify the PM from intracellular membranes by centrifugation alone. Moreover, this data tag also provides experimental evidence regarding the extracellular domain for transmembrane proteins. Such information is critical for the future development of affinity reagents that recognize the extracellular domain, especially in instances where membrane topology predictions are ambiguous. There are technical challenges to consider for any cell surface protein enrichment strategy, including the specificity of the label for protein subsets, the need to optimize labeling conditions for each cell type examined, the potential for contamination from intracellular membranes, and the often large number of cells required. However, as the methods become more streamlined and the detection limit in modern mass spectrometers improves, this will enable the analysis of smaller, more homogeneous populations. Finally, each cell surface enrichment method targets a different characteristic of the cell surface and, consequently, they reveal complimentary views of the surface proteome [75
]. It is unlikely that a single strategy will be able to fully elucidate the cell surface proteome, though integrating several approaches is expected to greatly expand our view of this important subproteome.
Figure 4 Strategies and benefits of PM targeting. (A) General proteomic workflow for PM protein enrichment. (B) Chemical-based tagging methods for affinity enrichment. (C–E) Comparison of results from seven studies [63, 71–73, 76–78] that (more ...)
] and affinity [63
] cell surface enrichment strategies have been applied to mouse and human PSCs, and altogether revealed 4415 proteins. Approaches that enrich for the cell surface report 1651 unique proteins and identified proteins with multiple transmembrane domains at a higher efficiency () when compared to non-PM targeting strategies. As shown in , it is clear that the enrichment strategies are more efficient for selectively identifying cell surface proteins over intracellular proteins than non-targeted strategies; although several studies in the “non-PM targeted” group were designed to enrich for other cellular subproteomes and thus not expected to reveal cell surface proteins.
With the implementation of these targeted strategies, the view of the cell surface proteome of PSCs is beginning to emerge, but it is not clear that the cell surface has been completely mapped and functional studies using any putatively informative markers are pending. However, optimism for the future significant impact of proteomic analyses of PSCs is well-founded. For example, the “cell surface barcode” useful for identifying cells in the HSC system currently employs a number of cluster of differentiation (CD) molecules. illustrates a qualitative CD molecule barcode, including 146 CD molecules, identified thus far in human and mouse PSCs based upon the studies reviewed here. Additionally, using the CSC-Technology on mouse PSCs (preliminary data), our labs have identified additional CD molecules and non-CD molecules not described in the data reviewed here. We expect that this CD molecule barcode will be a useful starting point for immunophenotyping PSCs using commercially available antibodies, and integrating quantitative measures and specific protein modification information in the future is likely to provide additional specificity to this evolving “cell surface barcode”.
Figure 5 Qualitative proteomic CD molecule barcode for human and mouse PSCs and an example of how a CD molecule can be used for live cell FACS of subpopulations of PSCs. (A) All protein CD molecules are listed. Solid circle indicates that the CD molecule has been (more ...)
We envision that an iterative strategy combining targeted surface accessible protein discovery and quantification, affinity reagent development, live cell sorting, and downstream functional characterizations of the sorted subpopulations will contribute to a better understanding of the extent of heterogeneity present in PSC cultures. Moreover and analogous to HSCs, antibodies against accessible cell surface proteins should assist in the immunophenotyping of these cells as well as the isolation of specific populations suitable for developmental studies or regenerative medicine. To illustrate a practical example of how surface markers can be used to sort subpopulations of PSCs, we sorted live SSEA-1pos
(CD15) and SSEA-1neg
cells from ESCs with suboptimal colony morphology (). To achieve this, live ESCs were obtained using a non-enzymatic cell dissociation solution (Millipore) to dissociate the colonies and were subsequently stained with SSEA-1 (eBioscience, E025199) prior to sorting. Sorted cells were plated and expanded and the SSEA-1pos
colonies showed better colony morphology and higher OCT4 abundance than the SSEA-1neg
population (). While SSEA-1 is neither a protein nor is it specific to the pluripotent state, this example highlights the ability of sorting live PSCs by surface accessible markers for subsequent expansion and downstream functional characterization. In another example, CD326 was recently shown to be useful for the isolation and subsequent expansion of live iPSCs [79
]. Thus, it is reasonable to predict that the application of appropriate proteomic technologies to expand the repertoire of cell surface markers for PSCs will lead to the development of affinity-based strategies useful for sorting defined subpopulations without relying on genetic modifications.
Finally and based upon these meta-analyses, a better view of the cell surface landscape is beginning to emerge for PSCs. However, several key challenges and questions still remain. First, to implement the newly identified surface markers into a cell sorting strategy, it is likely that new affinity reagents suitable for live cell sorting will need to be developed, which is especially difficult and time-intensive for integral membrane proteins. Second, the true value of any stem cell marker (natural or transgenic) will ultimately be determined by its ability to isolate defined stem cell populations with known differentiation potentials. Only through this process will it be possible to isolate therapeutically relevant stem cell populations useful in regenerative medicine. Thus, new and improved methods and reagents must be employed to facilitate the visualization and functional characterization of normal single stem cells within a population, and this is one area where proteomic technologies are expected to make a significant impact. Especially, proteomic strategies that provide insight into surface accessible protein domains are expected to facilitate these efforts by accelerating the epitope design and affinity reagent development for live cell immunophenotyping. Third, from a biological perspective, it may not be possible to develop a panel of surface markers that can isolate a truly homogeneous population of PSCs due to the extensive heterogeneity that is continually being discovered, but the development of antibody panels for live cell sorting is likely to be informative in evaluating heterogeneity among lines and isolating more therapeutically viable iPSCs. Finally, as cell surface proteins play critical roles in cell signaling, proliferation, and adhesion (reviewed in [80
]), and represent a large percentage of current drug targets [83
], the analysis of cell surface proteins is expected to greatly facilitate our understanding of how stem cells interact with their microenvironment as well as facilitate the study of stem cell-derived disease models, apart from their role in developing surface marker panels for sorting and identifying PSC populations.