In the 2011 report ‘Drosophila male germline stem cells do not asymmetrically segregate chromosome strands’ by Yadlapalli, Cheng, and Yamashita (J. Cell Science124, 933-939) the authors report results from their investigation of DNA-strand inheritance in the well-described germinal stem cells (GSCs) that renew sperm in newly hatched Drosophila melanogaster males. They conclude that their findings do not support the hypothesis of non-random segregation of sister chromatids in these cells, unlike reports of several examples in mammalian cells (Lark, 1969; Merok et al., 2002; Rambhatla et al., 2005; Pine et al., 2010; Hari et al., 2011), and including adult tissue stem cells (Potten et al., 2002; Karpowicz et al., 2005; Smith, 2005; Capuco, 2007; Conboy et al., 2007).
There are two aspects of the report by Yadlapalli and colleagues that warrant attention, which was not given by the authors. The first concerns the integrity of the applied experimental system and the nature of the results obtained when using it. Whereas the highlighted experimental advantage of studies that use transgenetically modified Drosophila spermatogonial GSCs is the ability to delineate the cell products of asymmetric self-renewal divisions of GSCs, the authors did not address that the transgenic GFP-fusions with centrosome-associated and microtubule-associated proteins they used might be able to disrupt non-random segregation if it occurred.
A ‘label-release’ approach was used to investigate the inheritance pattern of BrdU-labeled DNA in those GSCs that undergo asymmetric self-renewal divisions defined by the orientation of their division plane with respect to spermatogonial hub cells. This method has two signature findings that are indicative of non-random sister chromatid segregation. However, only one was considered in the report. It was the emergence of unlabeled–labeled, respectively, GSC–gonialblast (GB) cell pairs during the third cell cycle after BrdU-labeling had been stopped. A crucial technical requirement in order to detect these signature unequal GSC division pairs is an effective chase to prevent continued BrdU incorporation. The effectiveness of the chase period was evaluated for 4 hours after a 12-hour labeling period that achieved 50% of GSC being labeled. However, GSC–GB labeling patterns were evaluated after 36 hours and 48 hours (i.e. at approximately 2.6 and 3.4 cell cycles, respectively). Therefore, an appropriate test of chase effectiveness should have been evaluated at the same times, especially because BrdU-positive cells are scored qualitatively, not quantitatively.
The authors appear to have also overlooked the second signature finding of the label-release approach, despite its prominence in their data. At the third and subsequent post-labeling GSC divisions and – from then on – thereafter, non-random sister chromatid segregation produces unlabeled GSC–GB pairs from initially labeled GSCs. As indicated in their stochastic simulation (see Fig. 4A in Yadlapalli et al. 2011), the probability of such unlabeled pairs to occur by chance after three cell cycles is ~3%. However, their reported frequency was nearly tenfold greater after 36 hours of chase (approximately 2.6 cell cycles). The authors failed to discuss this obvious discrepancy between their observed data and their simulation for random sister chromatid segregation.
The second aspect that warrants attention is the authors' representation of their simulation for the expected frequency of observations of unequal inheritance of BrdU-labeled DNA strands due to chance as a function of chase time. By omitting two important qualifications, they misrepresent the significance of the predicted 50% frequency for mouse and human cells after about seven generations of chase. This low level of BrdU would be undetectable when using the methods to detect non-random chromosome segregation because it corresponds to less than one labeled chromosome per cell. Moreover, when asymmetrically inherited BrdU-labeled DNA was quantified, levels were quantitatively indistinguishable from those detected in cells that had initially been labeled (Merok et al., 2002).
On the basis of their reported findings, Yadlapalli and colleagues conclude that sister chromatid segregation in spermatogonial GSCs is, essentially, random. However, the results are not without important technical caveats that lessen the strength of this conclusion. Moreover, the authors' implicit representation that the detection of randomized unequal inheritance of BrdU-labeled chromosomes in an organism such as Drosophila (which has few chromosomes) – may have also occurred in previous studies that used mouse and human cells (which have a higher number of chromosomes) is ill-founded. The potential errors of interpretation suggested by the authors are readily avoided when the frequency of non-random sister chromatid segregation is significant and the non-randomly inherited BrdU content after four or more successive cell divisions is determined to be quantitatively similar to the pre-chase content.