We have read with great interest the recent paper published in Blood by Camargo et al.1 The authors used the Hoechst ABCG2-based efflux assay to isolate primitive stem cells from murine bone marrow (BM). They demonstrate that side population (SP) cells with a Hoechst 33342 (H33342) low-fluorescent profile (SP low fraction) have a higher clonogenic potential than the rest of the SP, showing that approximately 35% of single-transplanted SP low cells generate stable lymphohematopoietic grafts. In addition, they show that SP low cells are unable to reconstitute mice at absolute efficiencies. However, using similar strategies another group reported very high engraftment rates (> 90%).2

H33342 retention experiments use living cells, allowing the investigators to study an active functional process based on the differential efflux of H33342 by the ABCG2 multidrug transporter, which is responsible for the formation of the H33342 flow cytometric fluorescent profile.3 Within the SP, H33342 low retention has been associated with an increased ABCG2 expression and with a higher level of stemness and cell plasticity.4,5

Reduced accumulation of several fluorescent dyes (ie, rhodamine 123) has also been observed in stem cells, and increased dye efflux has been associated with long-term repopulating ability. Thus, the highest levels of multidrug transporters can be found in CD34+ pluripotent cells but also in CD34neg cells. When BM cells are stained with H33342, we are evaluating both the cell viability and the functionality of the cell systems. Moreover, the H33342 assay (as originally developed by Goodell et al6) would be better defined as a saturation rather than an efflux assay. Therefore, it is likely that we are selecting (under the appropriate efflux/saturation conditions) the most viable cells and then the most clonogenic, instead of a proposed series of different primitive stem cells within the SP able to actively exclude the H33342 dye at a different rate on the basis of the different ABCG2 expression.

For the SP assay, laser power and alignment, filter settings, sample collection, cell viability, H33342 concentration, and the whole staining procedure are crucial for the optimal resolution of the SP.7 Thus, different experimental conditions would contribute to explain controversial results. With little exceptions, the SP has been largely described as CD34neg, representing about 0.02% of BM nucleated cells. However, in our laboratory we have found an increased relative number of SP cells (median, 0.09%; range, 0.04%-0.96%, n = 100) both in human and murine normal BM samples, which are also CD34neg. These results could be explained by the fact that we have applied some modifications to the original H33342 staining protocol described by Goodell et al.6 Within those modifications we have lowered the power of the ultraviolet (UV) laser to 30 mW, giving a better resolution for the SP. It is likely that high laser power (50 mW to 100 mW) may result in DNA damage and thus influence SP clonogenicity. On the other hand, slight differences in the whole procedure could result in changes in the resolution of the SP, which might help explaining the differences observed between groups. Moreover, the H33342 dye has been reported to be toxic for the cells,8 so those cells presenting the low H33342 concentration such as the SP low fraction should be more viable and functional than the rest. We suggest that this constitutes an alternative explanation on why SP low cells display a higher repopulation capacity than the SP high subset.


  • Correspondence: Jordi Petriz, Laboratori 123, Institut de Recerca, Hospital Vall d'Hebron, P Vall d'Hebron 119-129, 08035 Barcelona, Spain; e-mail: jpetriz{at}


Hoechst-low side-population cells

We have described a method for isolating hematopoietic stem cells (HSCs) from murine bone marrow utilizing a Hoechst 33342 (H33342)-based stain.1,2 HSCs isolated by this method, named side population (SP) cells, can be further subfractionated on the basis of their differential Hoechst-dye efflux.2,3 Cells with the highest ability to efflux are enriched for long-term HSCs, whereas SP cells with less robust dye-extruding capacity are enriched for short-term HSCs. From these results we proposed that Hoechst 33342 dye exclusion is a functional property that directly correlates with the extent of HSC self-renewal.3 The letter by Sales-Pardo et al suggests that our ability to separate these 2 functionally distinct populations might not reveal inherent differences in self-renewal potential but rather represents increased viability and therefore functionality of SP cells that stain low for Hoechst. This is a valid concern to which we respond below.

Data on the toxicity of H33342 have been available for decades, as it is well known that H33342 is powerful inhibitor of DNA sythesis.4 From these early experiments, it was very evident that the effects on cell viability varied tremendously between different cell types and Hoechst dye concentration.4,5 Work by Van Zant and Fry5 found that most hematopoietic progenitors, with the exception of colony-forming unit-megakaryocyte (CFU-Meg) and CFU-erythroid (CFU-E), were functionally unaffected after H33342 staining. Interestingly, CFU-spleen (CFU-S), a heterogeneous population consisting of both short and long-term HSCs, was the least sensitive to Hoechst exposure under conditions similar to the ones used for our SP stain. In addition, experiments that we reported in 1997, in which we compared the long-term repopulation activity of whole SP cells with that of SP cells additionally exposed to Hoechst in the presence of the efflux inhibitor verapamil, demonstrated no differences in terms of function, as assayed by competitive transplantation experiments.2 We have also compared the clonogenicity of highly purified HSCs that were either stained or not with Hoechst 33342, and our results indicate that at least in vitro, there are no significant detrimental effects associated with Hoechst exposure.3

Even though Hoechst staining does involve a great deal of viability loss in nonhematopoietic tissues,6 as pointed out, it does not seem to cause any functional impairment within the primitive hematopoietic compartments. Therefore, we consider that our results do represent an inherent difference in the self-renewal capacity of cells within the SP subset that is directly correlated with their ability to efflux dye.


  • Correspondence: Fernando D. Camargo, Whitehead Institute for Biomedical Research, 9 Cambridge Center, Cambridge, MA 02142; e-mail: camargo{at}