| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
STEM CELLS IN THE NEWS |
1 Executive Director, Cell Therapy R & D Head, Hematopoiesis Department Holland Laboratory, American Red Cross hawleyr{at}usa.redcross.org
2 Research Communications Manager Holland Laboratory, American Red Cross sobieski{at}usa.redcross.org
"If it looks like a duck, and quacks like a duck, we have at least to consider the possibility that we have a small aquatic bird of the family Anatidae on our hands."
Douglas Adams, "Dirk Gently’s Holistic Detective Agency"
Multipotent hematopoietic stem cells (HSCs) have self-regenerative potential and the capacity to differentiate into all of the mature cells of the blood-forming system. It is vital to understand the underlying mechanisms governing these cell fate decisions in order to maximize the potential clinical use of HSCs in tissue regeneration and gene therapy applications. The ability to prospectively identify and isolate HSCs has been the subject of intense investigation since the first genetic evidence of their existence over 40 years ago. Despite significant advances in enrichment protocols, the continuous propagation (i.e., expansion) of human HSCs ex vivo has not yet been achieved. One reason for this failure is the difficulty in obtaining pure populations of HSCs. Because of this dilemma, the genetic programs involved in HSC self-renewal and lineage commitment have remained poorly understood [1]. Thus, determination of gene expression profiles for various murine and human populations enriched for HSCs and comparison with transcriptional profiles for other somatic stem cell (SC) populations—such as neural stem cells (NSCs)—and embryonic stem cells (ESCs) is expected to provide a better understanding of the SC phenotype. In the context of HSCs, it is hoped that this information will provide clues as to whether the apparently intrinsic asymmetric cell divisional behavior of these somatic SCs can be extrinsically manipulated to efficiently induce successive rounds of symmetric self-renewal divisions in vitro [2–5]. A number of reports over the past two and a half years have described Herculean efforts focused along these lines [6–15].
HSC GENETIC PROGRAM SNEAK PREVIEW
Phillips et al. [6] published an early analysis of the genetic program of HSCs in the June 20, 2000 issue of Science. The group purified mRNA from murine fetal liver HSCs (Sca+AA4.1+c-Kit+Lin-/lo) and from HSC-depleted AA4.1- fetal liver cells to develop cDNA- and RT-PCR-based subtracted fetal liver HSC libraries. The libraries were characterized by bioinformatics and microarray analyses, yielding 5,735 clones derived from the cDNA-based subtracted library, which represented 2,119 nonredundant gene products. The investigators used BLAST algorithm searches of 7 databases to analyze the 5' ends of the HSC-enriched clones. Conceptual translations of novel expressed sequence tags (ESTs) were analyzed for protein structural characteristics. The collected information has been deposited online in the Princeton Stem Cell Database (SCDb, http://stemcell.princeton.edu). Each entry includes an "executive summary" that encapsulates the predicted characteristics for the gene product. Although many of the proteins were previously undescribed, the researchers used their deduced properties to place them into known functional categories, for example, transcription factors, cell surface receptors, and signaling molecules.
Based on analyses of library complexity and functional tests for repopulating potential of the Sca+AA4.1+c-Kit+Lin-/lo and HSC-depleted AA4.1- fetal liver populations, the investigators suggested that SCDb represents the molecular phenotype of fetal HSCs. Among the gene products presented, at least 161 transcription factors, 174 cell-surface or membrane-associated proteins, and 147 signaling molecules were identified by FASTA comparisons, homology queries, or motif searches. Many of these were proposed to be novel HSC-specific genes. A preliminary comparison of sequences differentially expressed in fetal HSCs versus adult bone marrow HSCs revealed macroH2A1.2 and dnmt3b as being more highly expressed in HSCs from fetal liver. Both genes encode chromatin-modifying proteins likely to play roles in HSC biology. Among the gene products present in both SC types, the homeotic transcriptional factor TGIF (a known co-repressor of TGF-ß) was highly represented, underscoring the potential importance of TGF-ß in hematopoietic control mechanisms. SCDb is a rich source of information on HSC biology and provides the basis for continuing inquiry and analysis of HSC properties.
COMMONALITY OF HSC AND NSC GENETIC PROGRAMS
In the July 3, 2001 issue of the Proceedings of the National Academy of Sciences, Terskikh et al. [8] used a PCR procedure to create a subtracted cDNA library of genes expressed in adult murine HSCs as compared to total mouse bone marrow. The investigators used publicly available databases to conduct searches of the library’s sequences and then used SCDb to compare the adult HSC-enriched sequences with fetal liver HSC-enriched sequences. Transcripts coding for cell-surface proteins represented the largest group in the library, including the HSC marker, c-Kit. The investigators also identified several novel transmembrane proteins, including two members of the secretin subfamily of seven transmembrane-receptor proteins, mETL1 and Cyt28. Northern blot tissue analysis of mETL1 transcripts showed moderate expression in heart, and low levels in lung, kidney, brain, and liver. High levels of expression of Cyt28 were observed in brain, kidney, and heart and moderate expression in lung and testis.
The finding that many transcripts, in addition to Cyt28 and mETL1, were also expressed in central and peripheral nervous tissues prompted the researchers to investigate whether there were expression products that were shared between bone marrow HSCs and neurospheres derived from the subventricular zone of newborn mice, a population enriched for NSCs. cDNAs from these two cell types were cohybridized with an array of NSC-specific transcripts. Transcripts identified as common to both HSCs and NSCs were then used for in situ hybridization on tissue sections from developing mice ranging from embryonic day 13 to adult. The researchers found that the HSC/NSC-common probes hybridized in the germinal zones at all stages, coincident with the presence of neural progenitors. They found that Cyt28 was prominently expressed in the proliferative layers of cells immediately adjacent to the ventricles, which have been shown to contain at least some types of NSCs. The authors suggested that Cyt28 is a potential marker for neural stem/progenitor cells and that genes like Cyt28, which are preferentially expressed in two or more types of SCs, may form a nucleus of SC-specific gene products. In this regard, it is noteworthy that Cyt28, which is the murine ortholog of human GPR56/TM7XN1, has subsequently been found to be expressed in the HSC-supportive AFT024 stromal line [16], which raises the possibility of its involvement in cell-cell signaling between SCs and their niches.
A UNIFIED THEORY OF "STEMNESS"
In a paper published in the October 18, 2002 issue of Science, Ramalho-Santos et al. [9] prepared amplified RNA probes from murine HSCs (c-Kit+Lin-Sca-1+ "side population" cells from bone marrow), neurospheres (
3% NSCs) and a C57BL/6 ESC line, and hybridized the amplification products to oligonucleotide microarrays (Affymetrix MG-U74Av2 GeneChip® arrays). To provide a baseline comparison to differentiated cell types, they similarly prepared RNA amplification products from cells from adult mouse bone marrow and from lateral ventricles of the adult mouse brain. The arrays were analyzed with Affymetrix MAS 4.0 software (http://www.affymetrix.com/index.affx) in combination with dChip software developed by the Wong laboratory at Harvard (http://www.dchip.org). The resulting sequences were then compared among the various cell types examined and to sequences in publicly available databases to determine coincidence of genes in the newly derived and existing data sets (http://mcb.harvard.edu/melton/index.html). Previously known markers for each of the SC types were confirmed in the group’s results and each SC type showed a population of highly enriched gene products distinct from the other types. They also noted that the overlap of genes enriched in all three SCs contains more than 50% ESTs.
Significantly, the researchers observed a group of 216 transcription products enriched among all the SC types compared with the differentiated tissues. Four of these gene products did not appear at all in either differentiated cell sample: uridine phosphorylase, suppressor of Lec15, and two ESTs. Although the majority of the 216 SC-enriched genes were also expressed, but at lower levels, in differentiated tissues, the researchers suggested that their shared preponderance in a variety of SC types may indicate that they constitute a central group of gene products that are essential in imparting SC properties. The authors offered a list of key attributes of SCs: A) specific signal transduction pathways (e.g., JAK/STAT, TGF-ß, Yes kinase, and Notch); B) capacity to sense growth hormone and thrombin; C) interaction with extracellular matrix via integrin 
6/ß1, Adam9, and bystin; D) engagement in the cell cycle; E) high resistance to stress; F) a remodeled chromatin, and G) regulation of translation by RNA helicases of the Vasa type. Of the 216 SC-enriched genes, 12 of the 60 that mapped to a chromosomal location in the mouse genome via LocusLink (National Center for Biotechnology Information, http://www.ncbi.nlm.nih.gov/LocusLink) are found on chromosome 17, nearly four times the number of SC-enriched genes than expected by chance.
In a companion article in the same issue of Science, Ivanova et al. [10] compared expression profiles for human and murine HSCs and progenitors to elucidate evolutionarily conserved HSC-enriched gene products, and then extended the comparison to murine NSCs and ESCs to establish SC transcripts that extend across mouse tissue types. The adult murine HSCs, obtained from bone marrow, were long-term HSCs (LT-HSCs, Lin-c-Kit+Sca-1+Rholo, capable of life-long repopulation of the hematopoietic system), short-term HSCs (ST-HSCs, Lin-c-Kit+Sca-1+Rhohi, providing transient hematopoietic reconstitution), lineage-committed progenitors (LCPs, Lin-c-Kit+Sca-1-), and mature blood cells (MBCs, Lin+), and the murine fetal HSCs included Lin-AA4.1+c-Kit+Sca-1+ cells, LCPs with a Lin-AA4.1+c-Kit+Sca-1- phenotype, and Lin+ MBCs. The murine nonhematopoietic SCs examined were neurosphere side population cells (highly enriched for NSCs) and the ESC line CCE. The human cells profiled were Lin-CD34+CD38- HSCs, Lin-CD34+CD38+ LCPs, and Lin+ MBCs from fetal liver.
For each cell type, mRNA was prepared, amplified, hybridized to Affymetrix chips (MG-U74v2 and HG-U95 murine and human GeneChip® arrays, respectively), and analyzed using Affymetrix MAS 4.0 software. SC-enriched genes were subjected to bioinformatics analysis (http://www.sciencemag.org/cgi/content/full/1073823/DC1). The investigators assigned the HSC-enriched gene products to clusters corresponding to stages of hematopoiesis from LT-HSCs through terminally differentiated cells. When the investigators grouped the HSC-enriched genes by function, they found that 45% of the HSC-enriched genes fell into the category of regulatory molecules, including transcription factors, signaling molecules, cell-surface receptors, and ligands. They suggested that the preponderance of ligands, receptors, and ligand-receptor pairs implies possible importance of interaction of the cells with extracellular matrix as well as possible autocrine pathways. They also observed that ST-HSCs and early progenitors showed increased expression of molecules associated with entry into the cell cycle, including replication and proliferation proteins and DNA repair molecules. The researchers noted that, overall, fetal and adult HSCs shared 70% of HSC-enriched genes, a finding consistent with similarity of HSC function in fetal liver and adult bone marrow.
On comparing murine and human HSC genetic programs, the investigators found 822 human orthologs for murine HSC-enriched genes that are expressed in fetal liver, 322 of which were enriched in human fetal HSCs. These genes likely represent conserved functions in HSCs; the remaining orthologs did not show similar enrichment in both human and murine fetal HSCs. Finally, the researchers investigated the appearance of common enriched gene products across murine hematopoietic, neural, and embryonic SCs. The group found that HSCs and NSCs shared 644 enriched gene products; NSCs and ESCs shared 783; and ESCs and HSCs shared 538. A group of 283 sequences were found to be enriched in all three SC types. Some of these SC-enriched genes have previously been associated with SCs, including the transcription factors Edr1, Tcf3, EfnB2, and Hes1, but many are also expressed in non-SC cell types. Interestingly, one transcript of note, Abcg2 (also known as Bcrp1), which encodes an ATP-binding cassette transporter family member that was previously shown to be expressed in HSCs, NSCs, and ESCs [8, 17, 18] (see, e.g., Supplemental Table 2 of reference 8), was not identified as a common transcript in either study. This is especially curious considering that the side population assay based on Hoechst 33342 dye efflux activity mediated by Abcg2 was used to enrich for HSCs [9] and NSCs [10] in these studies. The trivial explanation that the Abcg2 probe set is not present on the oligonucleotide microarrays used appears to be excluded (93626_at MG-U74Av2 GeneChip® array).
FUTURE PERSPECTIVES
Clearly, therefore, not all expected SC-related genes are identified by current global gene expression profiling technology. Likewise, Ivanova et al. [10] indicated that only about half of the HSC-associated genes they identified had been previously detected by their group using subtractive hybridization methods [6]. As discussed in a recent review article [1], a major outstanding issue is the necessity for a centralized public repository to store and compare all SC-related gene expression data. The National Cancer Institute’s Cancer Gene Anatomy Project (http://www.ncbi.nlm.nih.gov/ncicgap) is a possibility; bioinformatics tools like the Gene Library Summarizer could then be utilized to find common genes expressed in multiple SC data sets that meet the chosen criteria.
Despite the limitations, collectively, genome-wide screening approaches represent important first steps toward defining an SC "molecular blueprint." In the future, it should be possible to conduct large-scale perturbation of SC-specific genes in the mouse by mutagenesis screens to determine the function of the unknown ESTs uncovered [19, 20]. Information gained from this type of system-level regulatory network analysis is anticipated to provide unprecedented insights into the genomic biology of SCs.
DISCLAIMER
Any views and opinions expressed herein are those of the authors. They do not necessarily reflect the policies or position of the American Red Cross.
|
ACKNOWLEDGMENT
"Life, the Universe, and Everything" by Douglas Adams. New York, NY: Pocket Books, Simon & Schuster, Inc., 1983;1-227.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| STEM CELLS | THE ONCOLOGIST | CME | ALPHAMED PRESS JOURNALS |
