Perhaps the most controversial area of debate concerns the use of Ddx4/DDX4 to select putative OSCs. DDX4 is considered to be a specific marker of PGCs and developing oocytes within the ovary and is generally believed to be intracytoplasmic in location (Castrillon et al., 2000). As sorting methods for intact cells require the use of cell surface markers to which antibodies can bind, it does not therefore seem possible that DDX4 could be utilised in this manner. However, it has been postulated that the protein has an external epitope in OSCs, which is internalised in more mature germ cells (White et al., 2012) (Fig. 1.7), as bioinformatics has suggested that the protein does in fact have transmembrane-spanning domains, thus justifying its use (Abban and Johnson, 2009, Zou et al., 2009). A study has demonstrated that some cells do express DDX4 on the cell surface: a porcine cell line transfected with a DDX4-containing plasmid expressed DDX4 on its cell surface as confirmed by flow cytometry, showing that it is possible for DDX4 to be expressed externally (Kakiuchi et al., 2014). Furthermore, flow cytometry of disaggregated pre-pubertal porcine testes detected a population of small cells with the DDX4 antibody used by White et al. (White et al., 2012) which expressed PRDM1 and IFITM3, although they did not appear to form germ cells and the authors concluded that they were not GSCs (Kakiuchi et al., 2014). Further information on the nature of the external epitope has come from White et al.’s study which demonstrated that only an antibody against the C-terminus of the protein is capable of selecting a Ddx4-positive cell population, whereas Ddx4-positive cells can only be detected by a N-terminus antibody if the cells are permeabilised (White et al., 2012).

Figure 4

Yet, conflicting data has been reported demonstrating that Ddx4/DDX4 is not expressed by putative mouse and human OSCs (Zhang et al., 2012, Hernandez et al., 2015). Zhang et al. used a fluorescent germline reporter mouse (Rosa26rbw/+;Ddx4-Cre) in which Ddx4-expressing cells change from fluorescing green to fluorescing either red (RFP), orange (OFP) or cyan (CFP) (Zhang et al., 2012). This allows the Ddx4-positive cells to be traced. No proliferation of isolated, in vitro cultured female RFP-expressing cells was found over 72 hours, in contrast to male cells which mitosed 1-3 times; indeed, these female cells did not establish in culture at all. However, as the cells were selected only on their expression of RFP and no further characterisation of these cells was apparently performed, it is possible the cells the authors were examining were oocytes, which clearly would not be expected to proliferate (Woods et al., 2013). In reply, Tilly’s group subsequently used a similar transgenic mouse strain (Rosa26tdTm/tdTm;Ddx4-Cre) whereby Ddx4 promoted the expression of tomato red (tdTM) to repeat the proliferation analysis experiment and demonstrated that oocytes were indeed included in the likely cell population that Zhang et al. examined (Zhang et al., 2012, Park and Tilly, 2015). They also demonstrated that a subpopulation of tdTM-positive cells could be isolated during FACS using a DDX4-antibody and this population behaved like OSCs; e.g. they established in culture and expressed germline markers (Prdm1, Dppa3, Ifitm3, Ddx4) (Park and Tilly, 2015).

Tilly’s group defended the use of DDX4, suggesting that technical constraints may be the reason for the evident lack of DDX4 expression (Woods and Tilly, 2015). Furthermore, a monoclonal DDX4 antibody has also been used to isolate OSCs (Fakih et al., 2015) and the sequence targeted by the C-terminus antibody is only homologous with a sequence in ATP-binding cassette subfamily C member 12 (ABCC12) that is located intracellularly (Woods and Tilly, 2013).

However, although the use of DDX4 is contentious, OSCs have been isolated using other putative markers (e.g. Ifitm3 (Zou et al., 2011, Xiong et al., 2015, Lu et al., 2016)); therefore the DDX4 argument alone is not sufficient to prove that OSCs do not exist.