SEARCH

SEARCH BY CITATION

Keywords:

  • Side population;
  • Hoechst 33342;
  • Stem cell;
  • Phenotype;
  • Dye efflux

Abstract

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

A defining property of murine hematopoietic stem cells (HSCs) is low fluorescence after staining with Hoechst 33342 and Rhodamine 123. These dyes have proven to be remarkably powerful tools in the purification and characterization of HSCs when used alone or in combination with antibodies directed against stem cell epitopes. Hoechst low cells are described as side population (SP) cells by virtue of their typical profiles in Hoechst red versus Hoechst blue bivariate fluorescent-activated cell sorting dot plots. Recently, excitement has been generated by the findings that putative stem cells from solid tissues may also possess this SP phenotype. SP cells have now been isolated from a wide variety of mammalian tissues based on this same dye efflux phenomenon, and in many cases this cell population has been shown to contain apparently multipotent stem cells. What is yet to be clearly addressed is whether cell fusion accounts for this perceived SP multipotency. Indeed, if low fluorescence after Hoechst staining is a phenotype shared by hematopoietic and organ-specific stem cells, do all resident tissue SP cells have bone marrow origins or might the SP phenotype be a property common to all stem cells? Subject to further analysis, the SP phenotype may prove invaluable for the initial isolation of resident tissue stem cells in the absence of definitive cell-surface markers and may have broad-ranging applications in stem cell biology, from the purification of novel stem cell populations to the development of autologous stem cell therapies.


Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

The ability to purify potential stem/progenitor cells relies largely on the expression of appropriate cell-surface antigens by the cells in question, such that they can be immunostained with fluorescently conjugated antibodies and isolated by fluorescent-activated cell sorting (FACS). In general, stem cell purification relies on judicious combinations of cell-surface markers, which can serve as either positive or negative markers for stem cell activity, none of which are by themselves specifically expressed on the stem cell. Although a major advance in the murine hematopoietic stem cell field came with development of phenotypic strategies to physically purify transplantable activity from bone marrow [1] and fetal liver [2], other hematopoietic stem cell (HSC) purification strategies have taken advantage of differential staining with vital dyes such as Rhodamine 123 or Hoechst 33342 [3]. In the hematopoietic field, the Hoechst low side population (SP) phenotype was originally described in murine bone marrow preparations, where this fraction was found to be greatly enriched for long-term repopulating hematopoietic stem cells [4]. Transplantation activity enrichment of mouse bone marrow–derived SP cells varies from 1,000- to 3,000-fold, which is similar to the enrichment achieved by purification of HSCs using combinations of cell-surface markers [5]. These SP cells are identified according to their ability to efflux the dye at a greater rate than other cells within the bone marrow. Moreover, the degree of efflux activity seems to correlate with the maturation state, such that cells exhibiting the highest efflux activity are the most primitive or least restricted in terms of differentiation potential [6].

Stem cell biology in general, but particularly in solid tissues, suffers from the lack of specific cell-surface markers that unambiguously label all stem cells and/or only stem cells. Although recent inroads have been made in this field, such as isolation of neural stem cells from mouse brains using a peanut agglutininlow/heat stable antigenlow profile [7], many organ systems (such as lung, pancreas, and kidney) still lack definitive markers for isolation of progenitor cells. The search for specific markers of resident tissue stem cells is a major theme in stem cell biology today. The Hoechst efflux phenomenon has proven to be a highly useful primary purification strategy for isolating potential stem/progenitor cells from various tissues in the absence of cell-surface markers. Cells with an SP phenotype have now been described in many solid tissues, including the skeletal muscle, lung, liver, heart, testis, kidney, skin, brain, and mammary gland.

Molecular Determinants of the Side Population Phenotype

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

How and why do these cells efflux dyes? The supravital stain Hoechst 33342 stoichometrically binds to AT-rich regions of the minor groove of DNA [8]. Hoechst fluorescence intensity is an index of DNA content, chromatin structure, and conformation and discriminates between cells in different stages of cell cycle [9]. The fidelity of Hoechst as a stem cell probe resides in the ability of the dye to be effluxed by membrane efflux pumps of the ATP-binding cassette (ABC) transporter superfamily, including multidrug resistance 1 (Mdr1a/1b, mouse; MDR1, human) [10] and breast cancer resistance protein 1 (Bcrp1)/ATP-binding cassette, subfamily G (WHITE), member 2 (ABCG2) [11]. The activity of these membrane pumps is blocked by verapamil, and the SP fraction is lost when this drug is included in the Hoechst incubation (Fig. 1). Enforced expression of these membrane transporters with retroviral vectors was shown to have direct functional effects on murine stem cells. Expression of ABCG2 blocked hematopoietic development [11], whereas overexpression of MDR1 resulted in HSC expansion and a myeloproliferative disease [12].

thumbnail image

Figure Figure 1.. Fluorescent-activated cell sorting profile of murine bone marrow after staining with Hoechst 33342. (A): The side population (SP) appears as the Hoechst low fraction capable of pumping out the dye and typically represents 0.05%–0.10% of viable cells from murine bone marrow. The non-SP cells that retain high levels of Hoechst staining are also referred to as main population cells. (B): The SP is ablated when verapamil is included in the Hoechst incubation. Verapamil blocks the activity of drug transport proteins, preventing them from effluxing the dye.

Download figure to PowerPoint

It was hypothesized that Mdr1a/1b and Bcrp1 may have overlapping function and be redundant in determining the SP phenotype in mice. As such, the loss of either transporter alone would not disrupt the SP phenotype. Mdr1a/1b−/− mice contain normal numbers of SP cells in bone marrow, demonstrating that expression of Mdr1a/1b is not required for the SP phenotype [11]. In contrast, Bcrp1−/− mice have significant defects and reduced numbers of SP cells in the bone marrow and skeletal muscle [13]. There was an almost complete loss of SP cells with the LineageSca-1+c-kit+ phenotype in bone marrow of these mice, with the small number of residual SP cells representing more mature hematopoietic cells with a predominantly Lineage+c-kit phenotype [13]. This suggests that Mdr1a/1b may help confer the SP phenotype but is not absolutely required and its Hoechst efflux activity is not as potent as Bcrp1. Experimental evidence supports this notion; enforced expression of a MDR1 retroviral vector in murine bone marrow cells in vitro only increased the SP fraction to 3.6% of the total population [14] whereas an equivalent experiment with an ABCG2 vector gave 62.5% SP cells [11]. Molecular regulators of these membrane pumps are poorly characterized, although members of the serine/threonine kinase Akt family have recently been shown to regulate Bcrp1 translocation, and the SP from the bone marrow of Akt1−/− mice is greatly reduced [15].

Functional assays have shown Bcrp1−/− mice have normal numbers of HSCs in the total bone marrow, indicating that HSCs were maintained in these mice, but with the loss of this membrane transporter, they were located outside the SP region [13]. This again highlights a lack of direct correlation between SP and stem cell. Also, although greatly reduced in abundance, the bone marrow of these mice still contains residual SP cells, implying that more than one drug efflux pump is responsible for this phenotype. These transporters may not be essential in conferring stem cell activity but may act more in providing environmental protection. Mice lacking both Mdr1a/b and Bcrp1 expression (M−/− B−/−) have normal numbers of peripheral blood cells and bone marrow colony-forming cells and demonstrate normal hematopoietic development [16]. There was a near total elimination of SP cells in the bone marrow of M−/− B−/− mice, but hematopoietic progenitor cells from the bone marrow were more sensitive to mitoxantrone in vitro compared with either M−/− B+/+ or M+/+ B−/− mice [16]. This suggests functional redundancy between these transporters for HSC development and further clarifies their role in the SP phenotype in HSCs and to intrinsic drug resistance within hematopoietic progenitor cells. Furthermore, overexpression of Bcrp1 in MCF-7 breast cancer cells conferred resistance to mitoxantrone, doxorubicin, and daunorubicin and caused an ATP-dependent enhancement of the efflux of rhodamine 123 cells [17]. Mitoxantrone selectively kills Bcrp1−/− hematopoietic cells at a dose that does not affect wild-type hematopoietic cells [13]. Moreover, Mdr1a/1b−/− mice show altered pharmacokinetics for several anticancer agents and hypersensitivity to the pesticide ivermectin [10]. This would seem particularly relevant for stem cells regardless of their tissue of residence given their potential requirement for ongoing proliferation throughout the life of the organism.

Markers of Side Population Cells From Various Tissues

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

If the ABC transporters Bcrp1 and Mdr1 are responsible for the SP phenotype, then is it possible to use just these markers to sort prospective stem cells from your tissue of interest? It would appear not since different tissue SP cells may use different transporters and some level of Bcrp1 expression is often detected in various non-SP cells [11]. Thus, although these transporters may contribute to the SP phenotype, they are not likely to be exclusive markers. Functional data also do not support the sole use of these markers for stem cell purification; in human hematopoietic preparations, ABCG2 coexpression with the HSC markers CD34 and CD133 was observed at low or undetectable levels, and cells sorted solely for the marker ABCG2 had little colony-forming potential [18]. If this is the case, then what other markers are shared between SP cells from different tissues? Is an SP cell in the bone marrow the same cell in the muscle? Does the SP represent a homogenous population of stem/progenitor cells? These are all issues that need to be addressed when considering the SP in your tissue of interest.

As determined by immunophenotyping, SP cells from non-hematopoietic tissues such as skeletal muscle, mammary gland, and testis share some phenotypic features with bone marrow, such as expression of Sca-1 and absence of mature hematopoietic lineage markers (Table 1). However, the expression of common stem cell epitopes is very heterogeneous throughout SP cells from various tissues, and each organ seems to contain a phenotypically distinct SP in terms of cell surface profile. Moreover, in many cases there is a large degree of immunophenotypic variation between SP cells from the same organ. For example, approximately 75% of murine liver SP cells express CD45, although the cells with the highest efflux capacity found at the lower tip of the SP tail were enriched for CD45 cells [19]. Similarly, distinct subpopulations of hepatic SP cells express the stem cell markers CD34, c-kit, Sca-1, and Thy-1 on both CD45+ and CD45 cells [19]. In any case, all hepatic SP cells were found to be negative for the mature hepatocyte marker fumaryl acetoacetate and the biliary epithelium markers cytokeratin 19 and A6, indicating that they are not mature liver cells or committed to a particular hepatic lineage. Two distinct subpopulations of SP cells, CD45+ and CD45, have been identified in lung. The overall SP fraction represents 0.05%–0.07% of the total viable cell population, with 60%–70% of these cells expressing CD45 and the remainder being CD45 [20]. Although both lung SP fractions did not express lineage markers and strongly expressed Sca-1, the CD45+ fraction was CD34+CD31+ whereas the CD45 fraction was CD34CD31+/− [20]. Phenotyping data from the two aforementioned tissues show the SP from these organs is a heterogeneous population that may encompass contributions from both hematopoietic and resident tissue cells to the total SP pool.

A general issue with all stem cell purification strategies is that whereas assayable activity can be greatly enriched, the purified cell populations are still heterogeneous. In most cases, it is difficult to separate the most primitive cell compartment from closely related progenitors. The same is true of SP cells, and the need to rely on complex in vivo assays suggests that it may not be possible to rigorously address the exact phenotype of the cells responsible for stem cell activity in any SP fraction. It should also be noted that in the literature there are often large discrepancies between SP abundances from the same tissue. This relates to the many variables involved in the preparation and staining for isolation of SP cells by FACS. Because efflux of a dye is a dynamic process, slight variations in tissue dissociation, cell counting, Hoechst concentration, staining time and temperature, and stringency in selection of SP cells by FACS gating can dramatically affect the viability, homogeneity, and apparent yield of SP cells [21]. Variations in preparation and staining protocols also can affect the observed phenotype of SP cells. For example, there are conflicting reports about the Sca-1 phenotype of murine cardiac SP cells (Table 1). In muscle, the SP cells at the lower tip of the SP tail are almost exclusively Sca-1+CD45, but as the gate size is increased to include SP cells closer to the main population (MP), the proportion of Sca-1CD45 and Sca-1CD45+ cells increases noticeably [21]. In the same study, a similar trend was witnessed in bone marrow, where the SP cells shifted from a Sca-1+CD45+ phenotype to a Sca-1CD45+ as the size of the SP gate increased and the position moved closer to the MP [21]. In contrast to this, the original more rigid protocols for isolation of SP cells from bone marrow show they are phenotypically (LineageCD45+Sca-1+) and functionally very homogenous in terms of long-term reconstituting HSCs [4, 6]. Hence, although dye efflux is effective in obtaining a population of cells containing progenitor potential, in most cases the SP is still quite heterogeneous and shows tremendous variation caused largely by different preparation and purification protocols. Thus, the terms side population cell and stem cell should not be used interchangeably. Assays to determine the relative enrichment of in vitro and in vivo stem cell capacity of sorted SP cells must be used to determine optimal conditions for isolation of SP cells for individual tissue. Each organ possesses a unique set of conditions, and SP protocols must be optimized accordingly. Ultimately, the highest enrichment of stem cell activity from the SP fraction of any tissue will be achieved by purification using various cell-surface markers in combination with dye efflux.

Origins of Side Population Cells Throughout Development

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

There is accumulating evidence to suggest hematopoietic cells are recruited to tissues such as muscle [22], heart [23], liver [24, 25], and kidney [26, 27] during regeneration from certain types of damage. If tissue SP cells represent resident stem cell populations, do the SP cells in particular organs originate from a common pool of cells in the bone marrow and adopt tissue-specific characteristics upon seeding within a specific local environment? There is some experimental evidence that supports this notion. In lethally irradiated recipients transplanted with bone marrow SP cells isolated from Rosa26 transgenic mice, 5 months after transplant, approximately 40% of host muscle SP cells were LacZ+, thus being derived from donor cells [28]. In Bcrp1−/− mice transplanted with wild-type bone marrow, a SP of normal abundance and phenotype was detected in recipient lung but none was detected from lungs of Bcrp1−/− mice transplanted with bone marrow from Bcrp1−/− mice [29]. Also, total bone marrow and bone marrow SP cells transplanted into lethally irradiated mice produced lung SP cells that were donor derived and contained both CD45+ and CD45 fractions (although further analysis showed that bone marrow SP, CD45+ lung SP, and CD45 lung SP cells represent three phenotypically distinct cell populations despite having a common origin) [29]. In lethally irradiated mice transplanted with bone marrow SP cells, donor cells contributed to formation of both CD45+ and CD45 hepatic SP populations after 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) treatment [19].

Alternatively, organ- and tissue-specific SP cells may arise solely as a consequence of the normal development of that specific organ but share the SP phenotype as a function of their inherent biological characteristics. A report where bone marrow from transgenic GFP+ donors was transplanted into the jugular vein of newborn W54/WV mice (genetic defect of c-kit receptor allows donor HSCs to colonize host bone marrow without irradiation) showed 8 months after transplant that the recipient testis SP was similar in size to previous studies (1.2%) but GFP+ cells were negligible (< 5%) [30]. In contrast, analysis of recipients showed 50%–60% of CD45+ cells in spleen and bone marrow were GFP+, indicating successful engraftment, and >70% of the bone marrow SP cells were GFP+ [30]. These results indicate that, at least in the case of the testis, the SP is not of bone marrow origin. In the kidneys of rats transplanted with bone marrow from transgenic GFP+ donors, approximately 10% of the kidney SP cells were GFP+, indicating that there is some bone marrow contribution to the renal SP [31]. However, previous reports have demonstrated bone marrow–derived cells can give rise to mesangial cells in vivo in response to mesangiolysis [32] in addition to playing a role in normal renal cell turnover [33]. In the aforementioned study, GFP+ cells were very rarely observed in glomeruli regardless of the induction of Thy1 nephritis or in tubular segments after induction of gentamycin nephrotoxicity [31]. The GFP+ cells seen in recipient kidney SP could have been circulating hematopoietic cells trapped in the renal vasculature, although the kidneys were perfused before staining.

Cellular Potential of Side Population Cells

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

With respect to whether an SP cell in one tissue is the same as an SP cell in another, this can also be investigated in terms of their cellular potential. SP cells from a range of nonhematopoietic tissues have been shown to have hematopoietic potential in vitro (with varying degrees of potency) when cultured under hematopoietic colony-forming conditions [34]; however, expression of the panhematopoietic marker CD45 is rarely observed in the SP cells of any tissue except bone marrow. The observation that tissue SP cells lack CD45 expression but have HSC activity might suggest that these cells have hematopoietic potential but are not committed to the hematopoietic lineage. They may lie above HSCs in terms of differentiation potential and in fact form either hematopoietic or mesenchymal lineages given appropriate stimuli. Despite the fact that many tissue SP cells have hematopoietic potential, they also generally lack expression of CD34, the marker classically associated with HSCs. However, the use of CD34 as a marker of murine long-term repopulating stem cells is a subject of considerable conjecture [35, 36]. The in vivo models harnessed to investigate stem cell potential typically involve some type of tissue injury as a priming event for proliferative expansion of transplanted cells to regenerate the damaged organ or expression of tissue-specific genes by donor cells in mutant mice. The methods used to induce and assess engraftment of donor cells will be stated in each of the following examples.

SP cells initially drew great interest in the stem cell community due to the potency of long-term repopulating HSCs contained in the murine bone marrow SP fraction. In addition to the ability to reconstitute the hematopoietic system, murine bone marrow SP cells have now been shown to have nonhematopoietic lineage potential in vivo. In lethally irradiated female mdx (dystrophin gene mutation) recipient mice injected with bone marrow SP cells of male origin, 4% of myofibers were dystrophin+, with 10%–30% of these fibers containing Y-chromosome+ donor nuclei, showing that bone marrow SP cells have myogenic potential in vivo [37]. Analysis of the bone marrow also confirmed that the female host was completely reconstituted with male donor cells [37]. Stable hematopoietic chimeras transplanted with bone marrow SP cells treated with DDC for 10 days showed donor cells in the hepatocytes of the parenchyma and periportal fields of the liver [19]. DDC is a drug that severely impedes liver parenchymal cell–driven regeneration, allowing more primitive cells to contribute to tissue repair. Thus, after stable engraftment, the progeny of bone marrow SP cells partially replenished the liver stem cell pool, from where they were recruited for regeneration of the damaged liver. In stable chimeras transplanted with bone marrow SP cells from Rosa26 transgenic mice and subsequently subject to cardiac ischemia and reperfusion, lacZ+ cells were observed in vessel structures of various types (mostly capillaries) and cardiac muscle [38]. In vessels, the LacZ+ donor cells colocalized with the endothelial cell markers Flt-1 and ICAM-1, demonstrating that the SP cells or their progeny had migrated to the injured heart via the circulation, localized to newly forming vessels, and seemed to have integrated into the surface lining as differentiated endothelial cells [38]. Lethally irradiated mice transplanted with bone marrow SP cells demonstrated an abundance of donor-derived osteoblastic cells lining trabecular bone near the growth plate, suggesting osteoprogenitors required for bone remodeling are recruited from a pool of SP cells residing in the bone marrow [39].

Reconstitution of the lymphohematopoietic system of lethally irradiated mice by limiting numbers of HSCs is the gold standard for in vivo stem cell assays, and at this point in time, direct stem cell activity from highly purified SP cells has only been demonstrated for those derived from bone marrow. Although the SP fraction from numerous tissues has now been associated with stem cell activity, this generally has not been demonstrated to the same degree of stringency as HSC activity from bone marrow SP cells. This again reinforces the point that SP cells cannot be directly equated with stem cells until proven so with limiting numbers in a functional assay. Nevertheless, isolation of SP cells from various tissues does seem to enrich for resident tissue stem/progenitor cells in many cases (Table 2). Muscle SP cells clearly have hematopoietic potential in vivo, although they seem to be less potent than bone marrow SP cells. This was shown via the analysis of the bone marrow of irradiated mdx females into which muscle SP cells from male donors were injected [37]. Variable engraftment of donor cells (30%–91%) was observed 1 month after transplant [37]. Y-chromosomes were detected by fluorescent in situ hybridization in metaphase spreads of recipient bone marrow, indicating that the introduced muscle cells could divide in vivo, and Giemsa staining revealed diverse types of hematopoietic cells. In addition, more than 90% of spleen cells in recipient animals were Y-chromosome+, and many expressed the hematopoietic markers CD43 and CD45, markers not expressed by muscle SP cells. Y-chromosome+ nuclei were also found in dystrophin+ muscle fibers of recipient mice, showing the myogenic potential of donor muscle SP cells. Recipients also showed Y-chromosome+ nuclei juxtaposed to myofibers, a position consistent with that of satellite cells, the myogenic precursors [40, 41].

Kidney SP cells also seem to have hematopoietic potential, as well as a capacity to become other mature cell types. Ten weeks after transplant of SP cells from GFP transgenic rat kidneys injected into irradiated recipients, approximately 0.03% of bone marrow cells of recipients expressed GFP, indicating donor cells had engrafted into host bone marrow and produced progeny (although their developmental stages and lineages were not analyzed). Donor cells were also localized in laminin+ skeletal muscle fibers and albumin+ hepatocytes. In the kidneys of transplanted rats, proximal tubules showed green fluorescence that was brighter than autofluorescence observed in wild-type rats and than fluorescence observed in distal tubules or collecting ducts of SP transplanted rats [31].

Liver SP cells certainly have the ability to give rise to a variety of cell types within the liver. Liver SP cells from Rosa26 transgenic mice transplanted into livers of recipient mice treated with DDC for 10 days before and after transplant engrafted as mature hepatocytes and bile duct epithelium in host livers, showing the hepatic SP cells had local regenerative capacity [19]. In addition to hepatic potential, liver SP cells also showed limited hematopoietic differentiation potential in vitro [19].

Other solid-tissue SP cells do not show apparent hematopoietic potential but do show multipotency. Skin SP cells from male donor mice injected into female mdx mice generated dystrophin-expressing Y-chromosome+ muscle fibers 3 months after transplant [42]. No donor cells were found in the spleen, and donor nuclei were not associated with regions of mononuclear cell infiltration, suggesting that these were not circulating cells. Thus, in nonirradiated hosts, skin SP cells did not differentiate into hematopoietic cells and likely did not remain in circulation, although they infiltrated damaged muscles and took residence there.

To test the potential of the mammary SP, SP cells were isolated from the mammary glands of Rosa26 transgenic mice and transplanted into cleared fat pads of immunocompromised Rag-1−/− recipient mice in limiting numbers [43]. Recipient mice were bred after 6 weeks to induce lobuloaveolar development in the mammary gland. Analysis of the mammary outgrowths showed donor cells could be detected in outgrowths from as few as 2.5 × 104 cells and robust staining was detected in outgrowths transplanted with 7.5 × 104 cells. Donor cells were present in both ductal and alveolar epithelium [43].

In recipient mice treated with busulfan to destroy endogenous spermatogenesis, donor testis SP cells transplanted via the efferent ducts showed a 13-fold increase in colonization efficiency compared with transplanted main population cells [44]. Moreover, testis SP-derived colonies were organized in cell associations qualitatively similar to those of normal adult mouse, although with a lower number of haploid germ cells [44].

The variability in potential of SP from different tissues can be interpreted as not supporting a common bone marrow origin for these cells. It can also be interpreted as suggesting a wide degree of differentiative potential for such cells upon exposure to various inductive environments. It is also highly likely that the apparent variability from one source of SP cells to another again reflects the heterogeneity of SP isolation from organ to organ. Tissue-specific SP cells can only be regarded as stem cells by clearly demonstrating that a single cell can give rise to multiple differentiated cell types of their tissue of origin.

What has not been examined in close scrutiny in many examples of SP transplantation is the concept that spontaneous cell fusion rather than transdifferentiation might be the mechanism by which the donor cells acquire unexpected phenotypes and supposedly cross lineage boundaries. In a recent study where whole bone marrow from genetically marked reporter mice was transplanted into lethally irradiated recipients, engraftment of donor cells into recipient organs occurred not through transdifferentiation of the transplanted cell but by fusion of host and donor cells [45]. Transplanted bone marrow cells were shown to readily engraft in recipient organs by fusion events, and multinucleated cells containing donor nuclei were detected in Purkinje neurons in the brain, hepatocytes in the liver, and cardiac muscle in the heart [45]. An elegant study used lineage tracing to show that the myelomonocytic progeny of a single transplanted bone marrow SP cell were the fusion partners of recipient cells and that engraftment did not occur through direct differentiation of the transplanted cell or generation of tissue-specific progenitors by the transplanted cell [46]. These findings are particularly relevant to studies reporting functional transdifferentiation of donor cells to myofibers because fusion is a natural step in regeneration of skeletal muscle (myogenic stem cells fuse with damaged myofibers). One or many host nuclei present in the same cytoplasm as donor nuclei may have been the source of tissue-specific gene products, and multinucleated myofibers make fusion events difficult to track in this organ. Dystrophin expression in mdx mice by myofibers containing donor nuclei has been regularly cited as evidence of donor cell muscle differentiation; however, it is known that in untransplanted mdx mice, there is a population of revertant myofibers that is capable of expressing dystrophin due to exon skipping [47]. In the case of transplanted SP cells, fusion has not been eliminated as a possible mechanism of apparent transdifferentiation. In a more recent study where irradiated Scgd−/− mice (mutation that causes loss of δ-sarcoglycan expression, producing progressive cardiomyopathy and muscular dystrophy) were transplanted with bone marrow SP cells, engraftment of donor cells in host myofibers was reported at a similar incidence as previous studies but expression of δ-sarcoglycan was almost never seen [48]. Not only has this result and the fusion phenomenon cast suspicion about the validity of previous transplantation studies claiming expression of muscle gene products lacking in recipient cells after stem cell transplantation, but it now casts doubt over any transplantation study where fusion and/or reprogramming of the recipient genome was not examined.

Therapeutic Options Using the Side Population Phenotype

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

The great hope for stem cell research is the potential development of cellular therapies for the treatment of various human diseases. One of the most widely used animal models in stem cell research is the mdx mouse, a model of Duchenne's muscular dystrophy that has a mutation in the dystrophin gene. Previous therapeutic strategies aimed at restoring dystrophin expression by cell transplantation [49] or in vivo gene transfer with viral vectors [50] have been moderately successful but have resulted in only local restoration of dystrophin expression. SP cells seem to provide a means for systemic repair of muscle as a consequence of delivery of these cells through the vascular system. One study attempted autologous transplantation of ex vivo genetically modified muscle SP cells in nonirradiated mdx mice [51]. Fifteen thousand SP cells from muscle of mdx5cv mice (the 5cv allele of mdx has a lower spontaneous revertant phenotype and slightly more severe disease) were isolated and lentivirally transduced with a construct containing the human dystrophin gene. After reintroduction of modified SP cells into hosts, human dystrophin expression was detected in the muscles only of mdx5cv mice injected with mdx5cv SP cells and not in mdx5cv mice injected with transduced main population cells or wild-type mice injected with transduced mdx5cv SP cells [51]. This supports the hypothesis that disease-damaged muscle could attract muscle SP cells from the circulation. The fact that lentiviral vectors can efficiently transduce a putative muscle stem cell and those cells can then contribute to the repair of a damaged muscle has substantial implications for possible autologous cell therapy of muscular dystrophy. Similar strategies could possibly see analogous therapies developed for a wide range of human single gene disorders, such as cystic fibrosis, Huntington's disease, and sickle-cell anemia.

Multi-Drug Resistance, SP Cells, and Stem Cell Properties

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

SP cells efflux dyes, and this property seems to be useful in enriching for stem cells. Why would a stem cell efflux vital dyes at a greater rate than other cells? One hypothesis to explain this phenomenon is that the overexpression of the membrane transporters responsible for the dye efflux mechanism in SP cells provides a mechanism for long-term survival of progenitor populations via an enhanced ability to pump cytotoxic compounds out of the cell [52]. A recent study showed that heme molecules (porphyrins) are detrimental to bone marrow progenitor cells from Bcrp1−/− mice under hypoxia [53]. Bcrp1/ABCG2 specifically binds heme, and cells lacking this molecule accumulate porphyrins [53]. Therefore, it seems that progenitor cells can use Bcrp1/ABCG2 to reduce intracellular heme/porphyrin accumulation and overexpression of this membrane transporter confers a strong survival advantage under hypoxic conditions.

What should also be considered is the known function of proteins such as MDR1. Some tumors are inherently resistant to most chemotherapeutic drugs (intrinsic resistance), and many others exhibit broad-spectrum or multidrug resistance after several rounds of chemotherapy (acquired resistance). A major cause of intrinsic or acquired multidrug resistance in humans is increased expression of the large (170-kD) cell-surface P-glycoprotein encoded by the MDR1 gene [54, 55]. For many years, multidrug resistance was thought to be caused simply by up-regulation of MDR1 in certain tumor cells upon exposure to drugs. As for dye efflux in HSCs, cytotoxic agents delivered as chemotherapy agents are actively transported out of cells that express MDR1 against a concentration gradient, thereby reducing intracellular drug concentration and inhibiting drug-mediated cell death. However, the mechanism by which MDR1 confers this protection may be more complex than simply an increased ability to efflux cytotoxic substrates. Interestingly, many solid tumors represent a hypoxic environment and, as noted previously, Bcrp1 is also upregulated in response to hypoxia. In addition, recent work has shown that MDR1 might play a fundamental role in regulating cell death. The functional MDR1 P-glycoprotein confers resistance to the apoptosis induced by chemotherapeutic drugs, UV irradiation, Fas crosslinking, or the binding of tumor necrosis factor-α to its cell-surface receptor [56]. These stimuli induce cell death by activating the common cell death cascade mediated by a family of cysteine proteases known as caspases. Biochemical analyses have shown that upon Fas ligation, functional MDR1 inhibits the activation of downstream caspases 3 and 8, an effect that could be completely reversed by the addition of verapamil or anti-MDR1 monoclonal antibodies [57]. MDR1 has also been shown to confer resistance to cell lysis induced by activated complement [58]. This growing body of evidence implicating MDR1 in the protection of cells against a diverse range of cell death stimuli makes a lot of sense in terms of a stem cell population. Increased MDR1 expression in SP cells may confer a strong resistance to apoptosis as well as enhanced protection from xenobiotic compounds and be a contributing factor in long-term survival of these cells.

If MDR1 is at least partly responsible for the SP phenotype, does that suggest a strong relationship between cancer cells and stem cells? Solid tumors contain cells that are heterogeneous in phenotype and proliferative potential. Because these malignancies are considered clonal in origin, it has been suggested that cancer cells in general may undergo processes that are analogous to the self-renewal and multilineage differentiation of normal stem cells [59]. One study hypothesized that malignancies may also contain a SP whose intrinsic dye efflux capacity might be anticipated to confer the ability to export many cytotoxic drugs and hence to increase the chance of early relapse [56]. By definition, a tumor showing multidrug resistance should have an SP. Indeed, Hoechst staining can be used to identify significant numbers of SP cells in both primary neuroblastomas and neuroblastoma cell lines, a tumor type with considerable intrinsic resistance. In both cases, increased efflux of mitoxantrone with the Hoechst dye could be demonstrated in the SP cell fraction compared with the non-SP cells. SPs have also been identified in Ewing sarcoma, teratocarcinoma, small-cell lung cancer, breast adenocarcinoma, and glioblastoma cell lines. The high drug efflux capacity of these cancer SP cells correlated with strong expression of the drug-transporter proteins ABCG2 and ABCA3 [60], as would be expected given that the ability to efflux vital dyes is the basis upon which the cells were sorted from these tumors. The argument that this proves that the SP is a component of tumors and hence tumors are related to stem cells is a circular one. The SP phenotype should not be considered as more than one marker that can be used for the enrichment of putative stem cell populations. However, the assessment of relative SP numbers in tumor types may give us more information regarding the etiology of that tumor type.

Conclusions

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

There is still a great amount of uncertainty about the origin, phenotype, and potential of SP cells from various tissues. Clearly the Hoechst efflux phenomenon has proven to be an extremely powerful tool to obtain enriched HSC populations from murine bone marrow, and this seems to correlate with the identification of putative stem cells in several solid organs. However, the dye efflux phenomenon is an active biological process, and because these prospective stem cells are not being isolated by a definitive cell surface profile, in many cases the purity of the population being isolated is very heterogeneous. Indeed, there are likely to be greater technical variations in the accurate isolation and quantitation of SP cells than cells isolated by direct immunopurification due to the dynamic nature of the parameter being measured (dye efflux). The fact that the proteins involved in the process of vital dye efflux itself may be used in other circumstances should always be taken into consideration before concluding that a cell with this phenotype is a stem cell. Despite this, there is evidence indicating that the SP is a good place to start in the search for resident stem cells in a particular organ when the phenotype of the cells in question is not known.

Table Table 1.. Phenotypes of murine side population cells from various tissues
  1. a

    Abbreviations: Lin, Mature hematopoietic lineage marker; ND, Not determined; ?, Conflicting data.

TissueLinCD45CD34Sca-1c-kitThy-1CD43Reference
Bone marrow++/−++/−+[6]
Skeletal muscle+/−+ND[37]
Mammary glandND+NDND[43]
TestisND++/−[30, 44]
RetinaNDNDND+/−ND[61]
SkinND++ND[42]
HeartND+/− (?)ND[62, 63]
BrainNDND+NDND[64]
LiverVery heterogeneous[19]
LungContains two phenotypically distinct side population subpopulations[20]
Table Table 2.. Stem cell activity of side population cells isolated from various mouse tissues
  1. a

    Abbreviations: DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; SP, side population.

TissueSP abundanceStem cell activityEnrichment for stem cell activityReference
Bone marrow0.05%–1%Reconstitution of hematopoietic system of lethally irradiated mice1,000- to 3,000-fold[4]
  Generation of dytrophin+ muscle fibers in mdx miceSimilar engraftment levels of 2,000 to 5,000 SP cells compared with 5 × 107 whole bone marrow cells[37]
  Form mature hepatocytes and bile duct epithelium in livers of DDC-treated miceNot determined[19]
  Generation of functional osteoclasts13% of recipient osteoblasts donor-derived in SP transplants (3,000 cells), 8% of osteoblasts donor-derived in whole bone marrow transplants (1 × 106 cells)[39]
  Engraftment and formation of functional tissue in lethally irradiated/coronary ischemic miceNot determined[38]
Skeletal muscle2%–3%Reconstitution of hematopoietic system of lethally irradiated miceNot determined (although much less efficient than bone marrow SP cells)[37]
  Generation of dytrophin+ muscle fibers in mdx miceVariable engraftment of 15,000 SP cells, no engraftment from injection of 15,000 MP cells[51]
Liver1%Form mature hepatocytes and bile duct epithelium in livers of DDC-treated miceVarying levels of engraftment from SP cells, no or negligible engraftment from non-SP cells[19]
Brain3.6%SP fraction contained >98% of all neurosphere activity in vitroFrequency of neural stem cell isolation 1 in 12 SP cells versus 1 in 3,333 non-SP cells (278-fold enrichment)[65]
Mammary gland2%–3%Generation of mammary outgrowths when transplanted into cleared fat padsSimilar outgrowth from <4,000 mammary gland SP cells compared with 100,000 total mammary epithelial cells (□ 25-fold enrichment)[43]
Kidney (rat)0.03–0.1%Formation of bone marrow, skeletal muscle, and liver and kidney cells in irradiated recipientsNot determined[31]
Skin0.5%–2.5%Generation of dytrophin+ muscle fibers in mdx mice9% of dystrophin+ fibers donor-derived in SP transplants (12,000 cells), <0.5% of dystrophin+ fibers donor-derived in MP transplants (800,000 cells)[42]
Testis0.1%–1.3%Generate full range of spermatogenic stages when in testes of busulfan-treated mice13-fold[44]
Retina0.1%Differentiated into neurons, glia, and late-born retinal cell types in vitroNot determined[61]

Acknowledgements

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References

The authors would like to thank Professor Margaret A. Goodell (Centre for Cell and Gene Therapy, Baylor College of Medicine, Houston) for critical reading of the manuscript. Grant Challen holds an Australian Postgraduate Award. Melissa Little is an NHMRC Principal Research Fellow.

Disclosures

M.L. founded and acts as a Director on the Board of Nephrogenix Pty Ltd.

References

  1. Top of page
  2. Abstract
  3. Purification of Stem Cells via Hoechst Efflux: The Side Population Phenotype
  4. Molecular Determinants of the Side Population Phenotype
  5. Markers of Side Population Cells From Various Tissues
  6. Origins of Side Population Cells Throughout Development
  7. Cellular Potential of Side Population Cells
  8. Therapeutic Options Using the Side Population Phenotype
  9. Multi-Drug Resistance, SP Cells, and Stem Cell Properties
  10. Conclusions
  11. Acknowledgements
  12. References