|
  
- 积分
- 0
- 威望
- 0
- 包包
- 483
|
作者：Andreas Httmanna,b, Ulrich Dhrsena, Katja Heydariana, Ludger Klein-Hitpassc, Tanja Boesd, Andrew W. Boydb, Chung-Leung Lib,e作者单位：aClinic of Hematology, University Hospital of Essen, Essen, Germany;bLeukaemia Foundation of Queensland Laboratory, Queensland Institute of Medical Research, Brisbane, Queensland, Australia;cInstitute of Cell Biology, University Hospital of Essen, Essen, Germany;dInstitute for Medical Informatics, B
5 o Z! O; ~1 V% o6 W) ^- h . l- n4 l+ }' Z5 ^1 y$ `0 t
5 ~ b3 t5 \! \2 q0 R 2 g5 A0 `) E/ Y9 g% T
" W; ]& a2 X* |" a. V
6 I) I2 V1 Y# @9 L7 d
2 _. b8 t. V, g" M# T" E 4 X1 f1 D7 `: _% d9 E, j8 a
4 ?; u* ]5 y W/ C$ h$ ~+ s
" N4 ]9 r/ _& a% c$ F$ o. c( i : G9 U3 e7 l0 g: N% `
b/ w& y3 m% n* f" Z+ L
5 i9 u+ f4 g' M; H- P$ p
【摘要】
1 E( D0 p2 k$ x0 ?0 K4 z Gene expression studies from hematopoietic stem cell (HSC) populations purified to variable degrees have defined a set of stemness genes. Unexpectedly, results also hinted toward a HSC chromatin poised in a wide-open state. With the aim of providing a robust tool for further studies into the molecular biology of HSCs, the studies herein describe the construction and comparative molecular analysis of -phage cDNA libraries from highly purified HSCs that retained their long-term repopulating activities (long-term HSCs ) and from short-term repopulating HSCs that were largely depleted of these activities. Microarray analysis of the libraries confirmed the previous results but also revealed an unforeseen preferential expression of translation- and metabolism-associated genes in the LT-HSCs. Therefore, these data indicate that HSCs are quiescent only in regard of proliferative activities but are in a state of readiness to provide the metabolic and translational activities required after induction of proliferation and exit from the HSC pool.
/ |/ S6 f' ^; H0 M# m0 ~ 【关键词】 Gene expression Quiescence Mouse Microarray Hematopoietic stem cells: n& ^' ]! E% A5 T
INTRODUCTION; l4 v6 C, L: d" R! X
7 v% `7 R! F! N$ J& Q$ E1 K, d6 J
Homeostasis of the hematopoietic system depends on the ordered replenishment of terminally differentiated cells. This process relies on rare hematopoietic stem cells (HSCs) and their progeny, which continuously enter tightly controlled pathways of lineage commitment, differentiation, and maturation. In steady-state hematopoiesis, HSCs are believed to be (a) relatively quiescent with respect to cell division frequency , evidence in support of the "traditional" model is provided by a large number of transplantation experiments using cell populations highly enriched for HSCs., p* I$ t. z* }0 P" M9 [
* P, O8 ~4 ?- G, g
Dissecting the hematopoietic hierarchy in functionally defined HSC and progenitor cell subpopulations allowed for analysis of the molecular phenotype underlying the HSC compartment. These studies combined elegant gene subtraction techniques, high throughput sequencing, and microarray and bioinformatics analyses to identify genes preferentially expressed in primitive murine HSCs of fetal and adult origin (Table 1). These large-scale surveys describe a global HSC gene expression profile in the purified cell populations investigated and identify numerous (novel) candidate HSC regulatory molecules. Confirmatory experiments pending, one of these studies led to a confined combination of receptor-type molecules allowing for isolation of a highly enriched HSC population by means of fluorescence-activated cell sorting (FACS) analysis or even in situ identification of HSCs by immunohistochemistry .# G, i5 i! S: D: z# w
0 p/ L: Q: ^: {7 \
Table 1. Gene expression studies of murine hematopoietic stem cells
4 u. {# b: t3 h" K) t3 D2 b
# l/ `- v8 E3 q; NIn the present investigation, we attempted to further these studies by constructing and genetically analyzing cDNA libraries from two functionally distinct adult murine HSC populations: . The cDNA libraries then served as templates for DNA-microarray analysis. In this study, we largely confirm the HSC gene expression profile observed in previous studies. Focusing on genes upregulated in the LT-HSC compartment, we herein report an unexpected preferential expression of ribosomal and oxidative phosphorylation-associated genes. These data indicate that LT-HSCs are metabolically highly active and appear to be poised in a state of readiness to rapidly boost the translational machinery when required to sustain rapid cell division and differentiation.9 K/ I; y: g3 Y2 J; J1 h5 }
' ^' h- S( @& l% Q" ^ y% KMATERIALS AND METHODS7 J3 {/ j" W+ o9 h1 h
2 M. y6 t4 c# a, M) s/ Y; m9 J
Animals
# Y( Y3 Z: W8 Y/ P5 [0 S1 h3 P4 _5 S" y
Mice of the C57B-L/6J(Ly5.1) and C57B-L/6J(Ly5.2) strain were used at 6¨C12 weeks of age. They were obtained from Animal Resources (Perth, WA, Australia, http://www.arc.wa.gov.au) and maintained in the Queensland Institute of Medical Research animal facility. All experiments were approved by the Institute¡¯s Animal Care Committee (docket no. A9605-99-029).3 A" F* Y# L, S4 v
9 ~- t1 G, }9 s6 m' }6 X+ e
HSC Isolation
* x- G7 j5 v* S) O% {8 m- D L% l5 k& t3 v/ r
To harvest bone marrow cells, six to nine mice were killed by cervical dislocation, and cells from both femora and tibiae were flushed out with balanced salt solution supplemented with 5% fetal calf serum (BSS-FCS). Bone marrow cells were pelleted, and red cells were lysed in 20 ml of hypotonic buffer solution (0.156 M NH4Cl, 0.1 M EDTA, and 0.01 M NaHCO3). For the rhodamine 123 and Hoechst 33342 staining, the marrow cells were resuspended in 10 ml of fresh BSS-FCS containing 6 µg/ml Hoechst 33342 (Hoe) and 0.1 µg/ml rhodamine 123 (Rho) (Molecular Probes, Eugene, OR, http://probes.invitrogen.com) and incubated at 37¡ãC for 30 minutes. The Hoe dose and duration of incubation have been optimized on murine thymocytes to give a tight G0/G1 peak. After pelleting, the cells were resuspended in 10 ml of fresh BSS-FCS containing 6 µg/ml Hoe and incubated for a further 30 minutes at 37¡ãC. For a more defined gate-setting on the Rho and Hoe parameters, 1 µg/ml Verapamil was added to a parallel control sample. Lineage depletion, Sca-1/c-Kit staining, sorting, and analyses using a FACS Vantage SE instrument (Becton, Dickinson and Company, San Jose, CA, http://www.bd.com) were performed as described . The gating strategy for the FACS is depicted in Figure 1./ Y' _! g9 H3 H- E1 f
7 o! e; E9 T. `2 W, l) {9 T
Figure 1. Gating strategy for isolation of HSCs. Viable cells were gated to identify the Hoechst side population (A). Cells positive for c-kit and Sca-1 expression (B) were then used to identify the Verapamil-sensitive gate (C). The Verapamil-sensitive gate was used to identify LT-HSC ST-HSC (D). Cells retaining low (, ST-HSC) levels of rhodamine 123 dye were sorted and used for subsequent manipulations. Abbreviations: FSC, forward scatter; HSC, hematopoietic stem cell; LT-HSC, long-term hematopoietic stem cell; ST-HSC, short-term hematopoietic stem cell.
`5 |/ n" | \( X
R/ ?3 K2 g5 }$ D# dLibrary Construction2 T: U; L% N1 p9 f2 y% C5 H0 J
9 L9 T) e2 ?& h9 l2 L
Forty-three thousand six hundred LT-HSCs and 151,600 ST-HSCs were isolated in 23 sequential FACS procedures (Fig. 1). Each of the 23 FACS procedures used 6¨C9 mice. After sorting, cells were lysed in RNAzol (Advanced Biotechnologies Ltd., Surrey, U.K., http://www.abgene.com) and kept at ¨C80¡ãC. For RNA extraction, the frozen samples were thawed on ice, pooled, and processed according to the manufacturer¡¯s instructions. The yield was less than 20 ng of total RNA from LT-HSCs and 79 ng from ST-HSCs. cDNA was generated by reverse transcription-polymerase chain reaction (PCR) with polyA-priming (Superscript II; Invitrogen Life Technologies, Carlsbad, CA, http://www.invitrogen.com), PCR-amplified (Advantage cDNA PCR Kit; Clontech, Palo Alto, CA, http://www.clontech.com), size-fractionated (Chroma Spin-400 Columns; Clontech), cloned directionally into the SfiIA and SfiIB sites of a -phage vector (TriplEx2; Clontech), and packed using a high-efficiency extract (Gigapack III Gold; Stratagene, La Jolla, CA, http://www.stratagene.com).9 j8 ]( Y% G, D0 Q
2 t1 \2 b3 h* K; uMicroarray Hybridization and Analysis# G# A0 W! g0 I V, A
6 t m0 ?- ~1 W
For gene expression profiling, library inserts were PCR-amplified (Platinum Pfx; Life Technologies, Karlsruhe, Germany, http://www.lifetech.com) using the T7 (primer: 5' ggc cag tga att gta ata cga ctc act ata ggg 3') and T3 (primer: 5'tcc gag atc tgg acg agc 3') promoter flanking regions as priming sites. DNA equivalent to 5 x 109 plaque forming units (pfu) from each library was used as a template. T7 RNA polymerase promoter tagged cDNA (0.5 µg) was used in an in vitro transcription reaction using the BioArray High Yield RNA Transcript Labeling Kit (Enzo Biochem, New York, http://www.enzo.com) to generate biotinylated cRNA. The cRNA was purified using RNeasy (Qiagen, Hilden, Germany, http://www1.qiagen.com). The quality of the cRNA was checked by agarose gel electrophoresis and found to have a comparable size distribution. cRNA was fragmented according to the Affymetrix manual, and 10 µg of fragmented biotinylated cRNA was hybridized to MOE430A arrays (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). Hybridization, washing and staining with streptavidin-phycoerythrin conjugate, and scanning (Agilent GeneArray Scanner 2500; Agilent Technologies, Palo Alto, CA, http://www.agilent.com) were done essentially as recommended by Affymetrix. After scanning, absolute analyses were done using the Affymetrix Microarray Suite 4.0 (MAS 4.0) to determine the signal intensities (average differences) and the absolute calls (present, P; absent, A; marginal, M). To facilitate comparison between the different analyses, the experiments were normalized to an arbitrarily chosen target intensity of 1,000 units, using the all-scaling option. Comparison analyses of the two experiments were done using MAS 4.0 to determine the fold change values as well as the difference call. Differential expression was assumed if a probe received a present call in one or both experiments and the log fold change call was 1 or ¨C1. In cases in which multiple probe sets covered a single Unigene cluster, the probe ID with the highest log fold change value was included in the analyses. Affymetrix probe set annotations were as of March 2005. Each library was analyzed by a single array experiment. The microarray data have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) and are accessible through the Gene Expression Omnibus series accession no. GSE4391.
& F! y: W) u% z u6 D U, f4 c7 [; e8 {# o: t
For determining the statistical significance of an enrichment of overexpressed genes in one chromosome or one particular chromosomal region, the hypergeometric distribution was used. With this distribution, the probability of observing at least x probe sets, which represent genes lying on one chromosome is given by:7 A) x' E, q+ ^1 T
/ w% a% i# K7 @, w2 ^& A/ H/ a5 i# f
where N is the number of probe sets with known chromosomal location (here 14,428), n the number of overexpressed probe sets with known chromosomal location (here, 1,055 for LT-HSCs and 1,550 for ST-HSCs), and K the number of probe sets, which represent genes lying on the interesting chromosome. To calculate p values, the software SAS (version 8.02; SAS Institute, Inc., Cary, NC, http://www.sas.com) was used. At the time of analysis, the array manufacturer provided chromosomal location information only for a minority of probe sets. Thus, chromosomal location of probes was derived from the Ensembl Mouse Genome Server database (accessible at http://www.ensembl.org/Mus_musculus/).
) O# M5 x7 e/ Q6 ~
( p4 K9 A- j# T- t' Z! q$ EReal-Time PCR r7 a; }7 J7 r1 l( p8 g& }
6 F0 I) t. u( S3 Q) ]Real-time PCR was performed with the ABI Prism 7900HT Sequence Detector (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) according to the manufacturer¡¯s instructions. PCR was carried out in a 20-µl reaction volume using 2 µl of diluted library aliquots (dilution factor 1:1,000) with a target-specific assay (Assays-on-demand Gene expression system; Applied Biosystems). Five targets with strong signal intensity (Rps17, Rpl14, Rpl37a, Cox5a, Ndufa8) and six targets with weaker signal intensity (Rpl7a, Rpl41, Rpl38, Atp5l, Atp5o, Atp5e) were selected from the list of ribosomal and oxidative phosphorylation-related genes upregulated in the LT-HSCs (Mm01314921_g1, Mm00782569_s1, Mm01546394_s1, Mm02342562_gH, Mm02524711_g1, Mm01610969_m1, Mm00432638_m1, Mm00503351_m1, Mm01612097_g1, Mm01611861_g1, Mm00445969_m1; Applied Biosystems). For normalization, the expression level of the housekeeping gene actin-beta (ActB) was measured as an endogenous control (Mm00607939_s1; Applied Biosystems). For quantification of each PCR result, we calculated the Ct value between the target gene and its endogenous control (ActB). Ct values from analysis in the LT-HSCs were subtracted from the Ct values in the ST-HSCs (Ct). The 2¨CCt value (relative quantity) was expressed as log fold change.
% J& W0 m8 R3 J3 i8 y- ?
+ B8 E9 g" r( j) c% ~0 GRESULTS' P1 {4 j" J [; K s( `$ e' q
3 T; `' K- l O9 z* Q j: O* h
Characterization of cDNA Libraries
1 S/ ~, G8 X7 @; \9 @4 _
{2 t$ ] f: l! C3 tThe titer of the primary cDNA libraries was approximately 4 x 106 pfu per ml for each library. For further analyses, library aliquots after one amplification round were used. One hundred forty clones from each library were picked randomly, and after conversion into plasmid DNA, inserts were analyzed by gel electrophoresis and DNA sequencing. The analysis showed that both libraries contained high numbers of independent (>95%) and recombinant (>95%) clones with an average insert size of approximately 1 kb (0.4¨C5 kb) (data not shown).! s$ o: u- a% n$ T$ u/ B( V: g
) \, a7 H; t8 T3 h/ H3 W
To test whether the transcription and amplification procedures caused a bias against full-length transcripts, array probe sets targeted to both 5' and 3' regions of two abundantly expressed housekeeping genes, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ß-actin, were used as an additional quality control. The 3'/5' signal ratios of GAPDH and ß-actin were 0.71 and 2.07, and 0.70 and 1.61 for the amplified LT-HSC and ST-HSC cDNA libraries, respectively, indicating a good representation of 5' ends of the mRNA in the cDNA libraries prepared.) G5 U$ X+ A" {# u/ {
" U9 R1 @/ M! D( G6 [Additionally, we expected that the relative abundance of mRNA encoding the MDR1/Abcb1a was higher in the LT-HSCs, whereas relative signal intensities of the Sca-1/Ly6-A/E, c-kit-receptor, and GAPDH and ß-actin genes was preserved at even levels in both libraries throughout the library preparation and amplification procedures. As depicted in supplemental online Figure 1, the relative signal intensity of these low- and high-copy number gene products indicates that the relative abundance of tested genes has been conserved throughout the sample manipulation and array experiment.
" Y m' q6 F4 ?- j$ O* u% z4 G+ n9 ^9 {1 g8 I! K Z9 Y( x
Comparative Analysis of Different Array Generations and Hybridization Events
5 c8 S4 t# v9 c$ L3 B+ K! Z! [; {( P; e& T6 e' b$ J! F, t# h
Because the present study used the MOE430A array and the majority of HSC gene expression studies listed in Table 1 used the Affymetrix MGU74 array, we compared the gene coverage of these sets with respect to absolute numbers of genes present and overlap between these two array generations. According to the Ensembl Mouse Genome Server, the current release of 31.33 g (as of June 2005) predicts 28,069 genes in the mouse genome (including 3,608 pseudogenes). As judged by nonredundant Unigene identifiers (Id), none of the commercially available microarrays covered the predicted number of genes (supplemental online Table 1). Combining the MGU74A, B, and C microarrays allowed for analysis of 38,028 probe sets, 15,384 (41%) of which had assigned a nonredundant Unigene Id. The MOE430A microarray used in the present study contained 22,690 probe sets, 12,867 (57%) of which were assigned to a nonredundant Unigene cluster. Furthermore, employing the Affymetrix annotation list as of March 2005, a substantial proportion of Unigene clusters were uniquely present on either the MOE430 or the MGU74 arrays. Five thousand two hundred eighty-three Unigene Ids represented by probe sets on the MOE430A array were not found on the MGU74A array (2,439 Unigene Ids when compared with the whole MGU74 A, B, and C set). Likewise, 3,618 Unigene Ids present on the MGU74A array were not included in the MOE430A array (5,283 Unigene Ids when compared with the whole MGU74 A, B, and C set). When a list of corresponding probe sets from different array generations was used (Affymetrix comparison spreadsheets, accessible at http://www.affymetrix.com/support/technical/comparison_spreadsheets.affx), similar discrepancies in gene coverage were observed. Two thousand one hundred seventy-one nonredundant Unigene clusters present on the MOE430A array were not represented on the MGU74 array set. Likewise, 9,767 probe sets from the MGU74 array set were not present on the MOE430A array. Taken together, the MOE430 array used in the present study, although still falling short of a comprehensive survey of all genes, allowed for analysis of genes not included in previous studies of HSC gene expression.
- R. F7 \+ A/ P8 K- A2 ^( a: }9 ^6 m
Out of 22,690 MOE430A probe sets, 4,676 (21%) received a present call in the LT-HSC and 7,384 (32%) in the ST-HSC experiment. The hybridization data and the raw comparison analyses are provided in supplemental online Excel Table 1. Excluding redundancy by means of filtering out Unigene Ids present more than once left 3,122 probes in the LT-HSC and 3,617 probes in the ST-HSC analysis. Differential expression (i.e., upregulation) was noted for 1,055 unique probes in the LT-HSCs and 1,550 unique probes in the ST-HSCs. Two thousand sixty-seven unique probes received a present call in the analysis of both LT-HSC and ST-HSC cell populations but failed the criteria for differential expression (supplemental online Excel Table 2).! M9 Z# r3 O* p) e: Y
1 T, \2 b! L" C/ W$ Q: w: GAccordance of Array Studies' H# G& f/ ?& ~* H5 q- [* k( y
7 k: N0 M2 W4 R: {- h5 q3 \6 ^To test whether the present results are in accordance with previous studies of gene expression in HSCs, we compared results from the landmark studies by Ivanova et al. . Of these 69 genes, 50 were found to be upregulated in the LT-HSCs in our study. Taken together, these data are in good agreement with the "common" HSC gene expression profile.
( G0 g8 M0 k; U
3 J3 U' ^% d* H1 d5 D$ jFigure 2. HSC gene expression overlap in studies from different investigators. Three thousand one hundred and twenty-two unique Unigene Ids receiving a present call (P) in the LT-HSC library were compared with 1,441 unique Unigene Ids present in the Rho_lo analysis by Ivanova et al. . Abbreviations: HSC, hematopoietic stem cell; lcb, lower confidence bound; LT-HSC, long-term hematopoietic stem cell.
# j% Z7 [/ R3 S+ j; T; q5 o; R" P& M/ E F% L
Chromosomal Clustering; }5 A& c+ }1 Z0 x. T
! p! d0 H7 ?/ F5 [" X% @4 TGroups of genes representing a functional entity (e.g., the major histocompatibility complex) may be spatially connected. Likewise, hematopoietic cell turnover is associated with a quantitative trait locus on chromosome 11 . Statistical analysis of our data led to different results (supplemental online Excel Table 4). A common hematopoietic cluster was identified on chromosome 3, in particular at 56,210K¨C60,570K bp (p = .007 for LT-HSCs, p = .02 for ST-HSCs). Clustering of LT-HSC-associated genes was observed on chromosome 4 (p = .04), and ST-HSC-associated genes were over-represented on chromosomes 12 (p = .04), 16 (p = .03), and 18 (p = .02). Whereas the hematopoietic cell turnover locus on chromosome 11 is underpinned by in vivo data, confirmatory experiments for the HSC loci identified in silico are still pending. As outlined above, coverage of the entire genome by available array sets is incomplete and a more comprehensive analysis may shift relative and absolute hybridization frequencies, if at all, to other loci.
" E3 h6 o9 B7 V( W8 T& k% g* a- X( T
1 n$ \* E. u8 _) DGene Expression Analysis by Functional Groups
; [$ Q/ C/ l. Z: y2 W0 L' `/ ~3 O7 s7 ], C
We next analyzed the expression profiles of cell cycle regulatory molecules, apoptosis-related molecules, hematopoietic transcription factors, hematopoietic cytokine receptors, and molecules involved in Wnt signaling. This approach aimed primarily at confirming whether this study was in line with the current notion of gene expression patterns of HSCs. Given that most lines of evidence support the view that the hierarchy of hematopoiesis is ordered on the basis of quiescence, the majority of LT-HSCs are in the G0 phase of the cell cycle .) f) i) u6 v& L1 |' @# y
, ~1 o; J y' s+ F$ x5 O5 }+ l
Figure 3. Relative signal intensities of cell cycle-related genes. A selection of genes active in certain phases of the cell cycle (box insert) as well as genes related to cell cycling in general are displayed. Bar length is indicative of relative signal intensity. Gene families are displayed as groups: the cyclin family (top), the CDKs (middle), and the CKIs (bottom). The emerging pattern suggests that, at the transcriptional level, more members of cell cycle-related gene families were active in the ST-HSC compartment as compared with the LT-HSC compartment. The inhibitory p21cip1/waf and p27kip1 molecules were overexpressed in the LT-HSCs. Abbreviations: CDK, cyclin-dependent kinase; CKI, cyclin-dependent kinase inhibitor; LT-HSC, long-term hematopoietic stem cell; ST-HSC, short-term hematopoietic stem cell.5 F7 J/ h8 W6 Q, \: u, k' t
' h" Z, V0 w, y7 G$ M- C3 YFor obvious reasons, maintenance of genomic integrity is of particular importance in any type of stem cell and DNA damage may lead to programmed cell death. Consequently, we examined the expression pattern of key players in apoptosis (Fig. 4). Of the Bcl-2 homology region-containing proteins, anti-apoptotic Bfl-1 was preferentially expressed in LT-HSCs. Of the pro-apoptotic family members, Bax and Bad were found upregulated in LT-HSCs. Mcl-1, whose indispensable role in HSC homeostasis has recently been demonstrated , was expressed at high levels in both libraries. Numerous caspase and IAP (inhibitor of apoptosis proteins) family members gave strong signals in the analysis of the ST-HSC library. Apaf-1 was highly expressed in LT-HSCs only.
" V6 }6 y* D& O; S$ |. T
' W; k$ l% f/ \( F1 m4 A5 X; JFigure 4. Relative signal intensities of apoptosis-related genes. Genes involved in apoptosis were grouped by gene family affiliation. The Bcl-2 family genes are depicted by anti- and pro-apoptotic function, respectively. Caspases with initiator and effector function, the IAP family, and members of the death pathway are likewise presented. Strong overexpression of the Bcl-2 family members Bfl1, Bax, and Bad was observed in the LT-HSCs, whereas IAP family genes were predominantly detected in the ST-HSCs. Abbreviations: IAP, inhibitor of apoptosis proteins; LT-HSC, long-term hematopoietic stem cell; ST-HSC, short-term hematopoietic stem cell.
! s$ z, f" G1 {; C' I$ L& V# B5 A: F% X. ^2 t) n# C
Two transcription factors indispensable for the onset of definitive hematopoiesis have been identified: the homeo domain-containing Lmo-2 gene and the basic helix-loop-helix (bHLH) SCL/Tal-1 gene. Both were present in the two libraries (Fig. 5). A spatial expression pattern was found for the transcriptional repressors of the inhibitor of DNA-binding family. Id-1, -2, and -3 were expressed in the LT-HSCs only. Because double and triple Id knockout mice die at midgestation due to cardiac defects, the hematopoietic phenotype of targeted Id-protein disruption has not been investigated .( w$ U6 f' j e5 D; z. j
( w" U! W1 h' I* O
Figure 5. Relative signal intensities of hematopoietic transcription factors. Hematopoietic transcription factors were analyzed and sorted by either functional group (lymphoid and myeloid) or by gene family/pathway (Idb, Gata, Notch). Transcription factors not categorized in one of these groups are displayed separately (top). The expression pattern of Idb family genes points toward a prominent role in LT-HSCs. Abbreviations: Idb, inhibitor of DNA binding; LT-HSC, long-term hematopoietic stem cell; ST-HSC, short-term hematopoietic stem cell.
, ?* }& |3 j8 F: g2 m( E7 T# r9 W& r% Y+ U& F: b
Of the hematopoietic cytokine receptors (Fig. 6), the c-Kit receptor, which was used as a selection marker for both cell populations investigated, was equally expressed at high levels. The gp 130 subunit, part of various heteromeric receptor complexes, was expressed in the LT-HSCs only. Likewise, the GM-CSF-receptor -chain, interleukin (IL)-4, and IL-7 receptors were exclusively found in the LT-HSCs. The thrombopoietin receptor MPL and the IL-6 receptor were found in both libraries, but expression was notably stronger in the LT-HSCs. A signal from either of the IL-3 receptor subunits was not detected in the libraries." h% S9 t. ?/ }, u; F) _' E9 R
7 k5 q/ E' D( G, d i" u+ T
Figure 6. Relative signal intensities of hematopoietic receptor-type molecules. Analyzing the expression pattern of hematopoietic receptor-type molecules using gp130, ß-chain, or -chain signaling did not reveal a common pattern in either or both of the libraries. In regard to the other receptor-type molecules depicted, it is of note that a Flt3 signal was restricted to the ST-HSCs. Abbreviations: LT-HSC, long-term hematopoietic stem cell; ST-HSC, short-term hematopoietic stem cell.
E$ N8 j4 V f- x% j [- |2 J) C) n3 M2 A4 P
As for the Wnt signaling components, studies investigating their role in maintaining HSC homeostasis have led to conflicting results , a demand for stem cell pool expansion may rely on enforced ß-catenin expression, whereas steady-state hematopoiesis may require axin-mediated ß-catenin degradation. Furthermore, numerous other elements of the Wnt cascade were predominantly found in the ST-HSCs.8 f3 g1 E7 b" u& y3 q
[/ R) ` F. @6 h" R+ _, SFigure 7. Relative signal intensities of molecules involved in Wnt-signaling. Separating the Wnt pathways into extra- and intracellular components revealed that, except for Fzd7 and Dkk3, all extracellular Wnts, Wnt-receptors, and Wnt-activity-modifying molecules were absent in both libraries. Of the intracellular molecules, Axin1 was notably more strongly expressed in the LT-HSCs. Abbreviations: LT-HSC, long-term hematopoietic stem cell; ST-HSC, short-term hematopoietic stem cell.
* R7 A' v) {6 u3 L$ B1 U" r9 q$ x' Y+ l( w, i1 w* L" s
Translation- and Metabolism-Associated Genes* e# u0 i5 e4 i2 `9 A& I' ?
4 M# p( p8 J* S6 Z! pWe next implemented pathway analyses in our study. A search of the Kyoto Encyclopedia of Genes and Genomes (KEGG) through the Database For Annotation, Visualization and Integrated Discovery (DAVID, accessible at http://david.niaid.nih.gov/david/version2/index.htm) revealed a preferential expression of ribosome genes in the LT-HSC library (supplemental online Excel Table 5). Of 76 genes listed as part of the murine ribosome complex, 59 were induced in the LT-HSC library (p = 4.58¨C41, as determined by KEGG). Along with the presence of genes coding for the ribosomal complex, several eukaryotic translation initiation factors were upregulated in the LT-HSC compartment (e.g., Eif3s1, Eif3s2, Eif2c2, and Eif5b). Furthermore, a high level of metabolic activity in the LT-HSCs was indicated by coordinate expression of genes involved in the oxidative phosphorylation pathway. Of 106 oxidative phosphorylation-related genes in the KEGG database, 26 were induced in the LT-HSCs (p = 5.74¨C7, as determined by KEGG) (supplemental online Excel Table 5). The vast majority of translation- and metabolism-associated genes also received a present call in the microarray analysis of the ST-HSCs, but the observed differences in signal strength allowed the conclusion that high transcription levels of these genes are an LT-HSC-specific characteristic. Confirmatory real-time PCR experiments of translation- and metabolism-associated genes further supported this notion (supplemental online Table 2).9 ], A- g( b# D( e1 [
: s: X2 v" j/ S. }5 G' j3 b
DISCUSSION; ^& a6 q ~1 v2 G
) D* O F" R" a3 X7 G
We constructed cDNA libraries from minute amounts of RNA isolated from highly enriched and functionally defined HSC populations. The rationale for this approach was to make available a robust cDNA resource from HSCs which could be repeatedly used in downstream applications. Analysis of the primary libraries revealed a high degree of recombinant and independent clones. Amplification-related severe bias against representation of 5'-ends was not observed. Collectively, the characteristics of the libraries allowed for further analysis in microarray experiments.0 ]* j" r" I, T- z# y7 q3 _2 I
4 K5 q% I0 _$ D T8 x M" O' D
Adding to our understanding of the mechanisms related to stem cell-specific properties, the molecular phenotype of HSCs purified to various degrees has been described (Table 1). The majority of these studies employed microarrays, and the results hinted toward a wide-open chromatin (reviewed in . Previous reports of HSC chromosomal clusters on chromosomes 11 and 17 were not reproduced in our study. As for chromosome 11, the cluster was identified by analyzing mouse strain-specific differences in hematopoietic cell turnover. Hence, the mouse strain used in our study may control hematopoietic cell turnover through different loci. In contrast, a putative cluster of HSC-related genes was found on chromosome 3, in particular at the chromosomal region from 56,210K¨C60,570K bp.( ^0 d) y. H( o7 D
2 B% X% Y) ^7 w8 o8 x7 n
As for the analysis of gene expression by functional groups, a partial expression pattern of genes involved in cell cycle regulation, apoptosis, transcriptional activation, and receptor-mediated signaling is provided (Figs. 3¨C7). Particular emphasis was put on the Wnt signaling pathway(s), given that a role in HSC self-renewal is increasingly recognized (reviewed in ). Dividing the Wnt-signaling pathway into extracellular and intracellular components, the extracellular components were almost completely absent in both libraries (Fig. 7). This finding suggests that paracrine signaling from the microenvironment is quantitatively more important than autocrine loops." ?# ~. O" |6 l+ h7 L) S% {
1 c" L; g! q: T4 xWe next employed gene ontology (GO) analysis. Despite the principal problem that the available software is prone to miss preponderance of a GO group or pathway , a positive result for differential expression of a whole GO group or pathway would argue for biological relevance. Herein, we report that pathway analysis revealed high coordinate expression of translation- and energy metabolism-associated genes in the LT-HSC compartment. In particular, we observed that almost the complete set of ribosomal genes, along with RNA encoding translation initiation factors, was upregulated in the LT-HSCs, whereas expression levels for these genes in the proliferatively more active ST-HSCs were at least twofold lower. Extending the previous notion that the wide-open chromatin of HSCs reflects a state of readiness, we conclude that the tools of the translational machinery are likewise present at high levels to allow rapid response to differentiation induction by signaling cues or stress events. Furthermore, the upregulation of oxidative phosphorylation pathway genes in the LT-HSCs argues for stem cell quiescence only in regard to proliferation, but not with respect to metabolic activities.
0 ~3 x2 k1 }( }9 I; y) M2 s, N2 u" w! Z: x' [5 U+ r
In summary, we herein characterize cDNA libraries from LT-HSCs and ST-HSCs, providing tools for further studies into HSC self-renewal and differentiation. Gene expression analysis showed good agreement with the described molecular phenotype of HSCs. We also show that genes coding for translational processes and energy metabolism pathways are enriched in the LT-HSCs despite their quiescent cell cycle status.9 c3 n" ?6 i" C- Y9 l2 h9 b3 g2 o
& y4 {" _6 O- ?
CONCLUSION
; }5 u$ R. Q9 H" a! l* Q) k. }- w( J% U M1 S1 l
Gene expression profiling studies from murine hematopoietic, otherwise somatic, and embryonic stem cells have defined a putative set of stemness genes. Validation of these results by means of hematopoiesis-specific gain- and loss-of-function analyses is presumably being pursued by many groups. In this regard, the availability of a resource like the cDNA libraries described herein may further HSC research in many ways. The value of this resource is underscored by its accordance with previous HSC gene expression studies. Furthermore, careful analysis of the cDNA libraries contents suggested that steady-state HSCs are far more "active" than anticipated.! l- t) z8 p# ^ M- ^6 {+ Q
4 t1 G! N% s3 Y$ E, i, l; EDISCLOSURES
4 ~9 g `6 ]' s+ L+ c- \+ ?) D, `( P# }5 J1 v2 a
The authors indicate no potential conflicts of interest.
# I) D! ]3 i( d/ R7 \
7 ?( l, {- T2 V) L$ zACKNOWLEDGMENTS! a7 v L3 u+ [; A" a! I# Y
8 v# Q* E! h5 R) N p9 I
We thank Paula Hall and Grace Chojnowski for assistance with FACS. This work was supported by the Leukemia Foundation of Queensland and the National Health and Medical Research Council of Australia. A.H. was supported by the Dr. Mildred Scheel Stiftung f¨¹r Krebsforschung, Germany. This paper is dedicated to Prof. G¨¹nter Brittinger on the occasion of his 75th birthday.8 }7 w4 y: u R) F: E* P
【参考文献】
2 C5 }3 h6 b( C; O5 o
* H- g3 Z, {$ f; p7 |/ Q# Z6 p3 b$ p
3 P4 Z& ?5 @. g# bHodgson GS, Bradley TR, Radley JM. The organization of hemopoietic tissue as inferred from the effects of 5-fluorouracil. Exp Hematol 1982; 10: 26¨C35.
& ]+ k7 k( p( W! G2 j. Z; Y) j/ F
H¨¹ttmann A, Liu SL, Boyd AWet al. Functional heterogeneity within rhodamine123(lo) Hoechst33342(lo/sp) primitive hemopoietic stem cells revealed by pyronin Y. Exp Hematol 2001; 29: 1109¨C1116.
' e1 ]$ t, g+ b% `+ w' w/ ^) t- z/ v" ^9 ?8 P# R' `
Siminovitch L, McCulloch EA, Till JE. The distribution of colony-forming cells among spleen colonies. J Cell Physiol 1963; 62: 327¨C336.+ K. V" J- @2 E8 J& E, D6 ?2 p
0 c1 K5 y8 X$ ~' W" K
Till JE, McCulloch EA. A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat Res 1961; 14: 213¨C222.
6 _/ k, U2 Q! I+ ?8 ^. u1 k
2 Q$ R: E: V# h2 }/ ]Adolfsson J, Mansson R, Buza-Vidas Net al. Identification of Flt3 lympho-myeloid stem cells lacking erythro-megakaryocytic potential a revised road map for adult blood lineage commitment. Cell 2005; 121: 295¨C306.
6 ?! R4 A, U1 [# G _6 U# T0 ?2 i3 F9 H p0 a6 X O- b
Kiel MJ, Yilmaz OH, Iwashita Tet al. SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell 2005; 121: 1109¨C1121.3 M& Y# ~; N o! y- Z8 D, z
7 {! D7 ^0 M; gLi CL, Johnson GR. Murine hematopoietic stem and progenitor cells: I. Enrichment and biologic characterization. Blood 1995; 85: 1472¨C1479.2 I) K+ H$ A; B. l
4 W' F1 B- m/ p* ]3 f1 GIvanova NB, Dimos JT, Schaniel Cet al. A stem cell molecular signature. Science 2002; 298: 601¨C604.4 M9 j b" F3 V& o$ J& t1 J
' d' n! }: y8 z1 i0 j5 b/ n- Y: @
Ramalho-Santos M, Yoon S, Matsuzaki Yet al. "Stemness": Transcriptional profiling of embryonic and adult stem cells. Science 2002; 298: 597¨C600.* |- x1 ^5 Q; s+ _8 B' R
, q, }: E* {" r9 u3 ^Zhong JF, Zhao Y, Sutton Set al. Gene expression profile of murine long-term reconstituting vs. short-term reconstituting hematopoietic stem cells. Proc Natl Acad Sci U S A 2005; 102: 2448¨C2453.
# t0 Z- W, |- K7 _- M( |) C8 ?+ q4 ]% V# _1 E
de Haan G, Bystrykh LV, Weersing Eet al. A genetic and genomic analysis identifies a cluster of genes associated with hematopoietic cell turnover. Blood 2002; 100: 2056¨C2062.
! H6 q) b3 n) N9 t) C7 c, A2 F2 i0 O3 g
+ v+ e4 \* O* a3 j) q; \Cheshier SH, Morrison SJ, Liao Xet al. In vivo proliferation and cell cycle kinetics of long-term self-renewing hematopoietic stem cells. Proc Natl Acad Sci U S A 1999; 96: 3120¨C3125.& _; k7 Y1 V( a" [
+ X, c% J s" D8 i$ a$ t
Cheng T, Rodrigues N, Shen Het al. Hematopoietic stem cell quiescence maintained by p21cip1/waf1. Science 2000; 287: 1804¨C1808." b J/ k4 V' Y) o
0 j1 U; ^* f+ v0 d; E' L# {, Z, a' D
Cheng T, Rodrigues N, Dombkowski Det al. Stem cell repopulation efficiency but not pool size is governed by p27(kip1). Nat Med 2000; 6: 1235¨C1240.
' k# Y0 s! S/ w2 B1 S' \
+ f2 V" P# f% I' ?# x# p0 s5 ^Opferman JT, Iwasaki H, Ong CCet al. Obligate role of anti-apoptotic MCL-1 in the survival of hematopoietic stem cells. Science 2005; 307: 1101¨C1104.2 z2 u/ Y8 G+ S0 ~7 N$ e) S; Y
# I0 _$ ^3 P2 T
Fraidenraich D, Stillwell E, Romero Eet al. Rescue of cardiac defects in id knockout embryos by injection of embryonic stem cells. Science 2004; 306: 247¨C252.
3 f4 G$ k+ | K+ e* |# ?
/ }$ j: O8 }. u) zAkashi K, He X, Chen Jet al. Transcriptional accessibility for genes of multiple tissues and hematopoietic lineages is hierarchically controlled during early hematopoiesis. Blood 2003; 101: 383¨C389.! L, u4 ~7 ?1 M2 X$ ]% a
. b# g, l* D3 T- d5 R
Reya T, Duncan AW, Ailles Let al. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 2003; 423: 409¨C414./ w5 z7 |+ r% t; O% N
- c, I6 H2 k3 B6 Q% x
Willert K, Brown JD, Danenberg Eet al. Wnt proteins are lipid-modified and can act as stem cell growth factors. Nature 2003; 423: 448¨C452.; s; x, {2 f3 n" [* p+ D' t
: w! e) B- S3 oCobas M, Wilson A, Ernst Bet al. Beta-catenin is dispensable for hematopoiesis and lymphopoiesis. J Exp Med 2004; 199: 221¨C229.
% K0 ]8 o" X& U$ O& [, ^' m2 w, h/ O. z
Dennis G Jr., Sherman BT, Hosack DAet al. DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol 2003; 4: 3." T$ D2 M, T n: h
1 F) {+ o4 |* C7 z$ }6 KLi L, Akashi K. Unraveling the molecular components and genetic blueprints of stem cells. Biotechniques 2003; 35: 1233¨C1239.
$ c) h6 J. m1 v% L. o9 \4 @% W5 {0 {$ G
Vogel G. Stem cells. ¡®Stemness¡¯ genes still elusive. Science 2003; 302: 371.- \; p: @6 P) c% _7 C" B- W
1 F6 D: @) p5 QEvsikov AV, Solter D. Comment on "¡®Stemness¡¯: Transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 2003; 302: 393.( Z) J9 i$ M- y! t& B+ X- d
0 F: a" c& P0 t3 U% `5 O
Fortunel NO, Otu HH, Ng HHet al. Comment on "¡®Stemness¡¯: Transcriptional profiling of embryonic and adult stem cells" and "a stem cell molecular signature". Science 2003; 302: 393.- P6 n7 l1 g6 n2 I9 l$ E2 _" P
8 a& Q& }* Y+ P7 W
H¨¹ttmann A, Li CL, D¨¹hrsen U. Bone marrow-derived stem cells and "plasticity". Ann Hematol 2003; 82: 599¨C604.
s4 H: W; B/ y0 \0 E( T& u8 u u! c! |7 Q& V; q
Rattis FM, Voermans C, Reya T. Wnt signaling in the stem cell niche. Curr Opin Hematol 2004; 11: 88¨C94.0 F: ^. ^ m E. a8 o" c
/ c& X! @" E" ^4 k
Reya T, Clevers H. Wnt signalling in stem cells and cancer. Nature 2005; 434: 843¨C850.
& y4 d/ p5 L8 X5 d' v' c# N( m+ d ?; ]! D6 {' r+ ]- z0 W
Khatri P, Draghici S. Ontological analysis of gene expression data: Current tools, limitations, and open problems. Bioinformatics 2005; 21: 3587¨C3595.
0 z: S- B+ M- F) {+ f
- o: S( m( U1 i1 l ?/ f2 h: kPhillips RL, Ernst RE, Brunk Bet al. The genetic program of hematopoietic stem cells. Science 2000; 288: 1635¨C1640.
4 Q* S `' C% E6 M6 n7 q* h! \+ K$ c9 ]' n
Park IK, He Y, Lin Fet al. Differential gene expression profiling of adult murine hematopoietic stem cells. Blood 2002; 99: 488¨C498.- O) `" m) w3 [$ v4 D' X, X
- b# n) ~7 A9 n# o8 u" N5 x
Terskikh AV, Miyamoto T, Chang Cet al. Gene expression analysis of purified hematopoietic stem cells and committed progenitors. Blood 2003; 102: 94¨C101.
" J) v/ C. v! ?& v
9 l! V) f# m% n# B* ?& H6 @Venezia TA, Merchant AA, Ramos CAet al. Molecular signatures of proliferation and quiescence in hematopoietic stem cells. PLoS Biol 2004; 2: e301.
9 _1 j# T' V! G8 f7 {* M a1 `6 G( H
Rossi DJ, Bryder D, Zahn JMet al. Cell intrinsic alterations underlie hematopoietic stem cell aging. Proc Natl Acad Sci U S A 2005; 102: 9194¨C9199. |
|
发表于 2009-3-5 00:04
发表于 2015-6-5 11:01
