  
- 积分
- 0
- 威望
- 0
- 包包
- 483
|
作者:Mark A. Vickersa, Sarah J. Canninga, Wendy L. Craigb, Neil M. Massonb, Ian J. Wilsonc - v/ @( @- N* Q) n
7 w r7 L9 \7 w3 I2 K
* ^" Y% ~2 w; Y& W* F
4 x+ L7 ~/ C0 V* b8 B( m. ]4 Z6 B 8 @+ q2 q* r; W/ e: j: X
, a) S W' q6 e# U o6 t
8 t3 Y, J" x8 I- _' ]( n# H& K/ T
: ?' U8 j, e9 p5 F7 W
8 x* A4 Q7 W& `$ p) P# `( w0 K/ o
9 s1 N5 s, k9 k+ g j. H+ @( I, ~
/ ]6 \4 B$ o% j1 J, c
6 |7 D* b# N$ D* Y o- b. s8 C ( ~- s0 n7 S( p1 {
【摘要】 N. _% Z$ I/ o7 S/ ~
The early, random nature of X inactivation should cause related cells to have similar, but distinctive, active X chromosomes. We assessed the frequency of stem cell plasticity using X inactivation proportions (XIPs), determined at the human androgen receptor locus, in paired tissue samples from healthy individuals. Tissues sampled were stomach (n = 18 informative females), duodenum (n = 18), colon (n = 10) with corresponding peripheral blood samples (n = 33), and varicose veins (n = 28) with corresponding T cells (n = 26) and peripheral blood granulocytes (n = 25). XIPs from samples thought to have common stem cell origins were highly correlated: multiple samples from single vein, r = .80 (n = 24); T cells versus granulocytes, r = .67 (n = 23); duodenum versus stomach, r = .63 (n = 12). Blood cells and vessels are derived from a common hemangioblast, but XIP correlations were moderate or poor: vein versus T cells, r = .42 (n = 26); vein versus granulocytes, r = .11 (n = 25). X inactivation is believed to be a late process in gut, especially hind-gut, with corresponding independence from blood precursors. Correlations with blood cells were low: stomach, r = .23 (18); duodenum, r = .21 (18); colon, r = .034 (10). Any crossover of stem cells between different organs during adult life should increase correlations with age; no such increase was seen. This study confirms that XIPs can be used to track stem cell populations, provides a theoretical basis for the power of such studies, and indicates that hemopoietic stem cell plasticity is, at most, uncommon in normal humans. / V& P; v Y/ E& A8 Z6 t
【关键词】 Biomathematical modeling T cells Stem cell plasticity Somatic stem cells Pluripotent stem cells Multipotential differentiation Granulocytes' G2 p5 m. i f# `' n9 O: ~
INTRODUCTION
' Z4 R9 f. x% y: @4 s) O, Q9 s0 k, J, P: C i
It has been suggested that stem cells, previously believed to be committed to predetermined, organ-specific paths of differentiation, can differentiate into cells of other organs. This view has been supported predominantly by observations after transplantation. For instance, after bone marrow transplantation in experimental animals, donor-derived cells can be detected in vascular endothelium, liver, heart/cardiac muscle, skeletal muscle, glia, neurons, kidney, pancreas, lung, skin, and gastrointestinal (GI) tract .
. `8 w) W9 E6 ~6 T0 {9 c4 b8 Y9 o: z1 ]4 M! k4 t c+ t
However, stem cell plasticity remains controversial .5 r9 {. x7 k/ C; [: r) P; `
, ?: s. {5 I* b* ]: x T( w
Even if the data reflect genuine lineage plasticity, it is unclear whether normal human stem cells behave in this way or whether plasticity is an experimentally induced artifact. To address this question, a technique for marking populations of normal stem cells is required. Theoretically, the phenomenon of X inactivation provides such a natural marking technique. In females, only one of the two X chromosomes in each cell is transcriptionally active. The inactive chromosome is silenced randomly during the blastocyst stage of development , and so each organ system, has its own "signature" ratio, allowing discrimination between populations of cells. Sampling XIPs of cellular populations therefore provides a measure of similarity of origin. If there were free exchange of stem cells between organs, each sample would be derived from the same population, and the XIPs would be highly correlated. If there were no crossover, correlations would be lower.: q8 G" N1 G; u; d- Z. v0 }- G
$ |5 G* m" _" X8 ]* j, h; j l TSeveral factors important in determining ratios must be taken into account when interpreting values (Fig. 1). If X inactivation takes place before lineage commitment and the stem cells of two organs are derived from the same group of cells, the ratios should be similar but not identical. Even if allocation of precursor cells to organs is random, a hypergeometric distribution (sampling cells without replacement) should result. Further factors likely to cause XIPs to differ between tissues include measurement error, heterogeneity within individual tissues, and nonrandom allocation of primordial cells to certain organs. The mixture of all these distributions should cause the XIP of each organ to differ from one another at birth. If stem cells were shared between organs during adult life, whether by trans- or dedifferentiation and/or common precursor stem cells, then the ratios of relevant organs should become more similar with age. Other processes that affect the rate of change of correlation with time are selective pressure on certain X chromosomes, important in determining differences between granulocytes and T cells, and stochastic stem cell replication, which has been shown to approximate to a ß-distribution .! y4 G. A2 D' o, D* A% ^- k, j
2 ~3 A, Q) ~; y* K' j6 I" C8 Z
Figure 1. Illustration of determination of X inactivation patterns in different organs. Individual cells are shown as circles, with methylation of maternal or paternal alleles as light or dark shading. Unmethylated cells are shown as empty circles. The genome in the developing morula undergoes demethylation (clear cells on left). As differentiation proceeds, some organs may be derived from cells that are specified before X inactivation is initiated. In this case, the X inactivation ratios of organs will be drawn from two different binomial distributions and not be correlated (compare organ 1 with organs 2, 3, and 4). Alternatively, X inactivation may be initiated before lineage commitment. If sufficient cell amplification occurs before organ specification occurs and two organs sample from this population, then organs should share similar XIPs. On the other hand, the organs may differ, as the two samples may be drawn from spatially different parts or at random, in which case they will be equivalent to two samples drawn from a hypergeometric distribution (sampling with no replacement), which may differ considerably if the number of cells sampled to form each organ is low (compare organ 3 with organ 4). During life, stochastic asymmetric stem cell replacement will result in further randomization according to a ß-distribution, a minor effect in blood but one that may be larger for biopsies from solid organs, where fewer stem cells are sampled. Finally, there may be selective advantage for one X chromosome over another (not illustrated). b. E% B. C. [
6 H8 ]) a, }% E( Y+ h
Several groups have reported correlations between XIPs in different tissues. Gale et al. showed a correlation between buccal epithelia, urinary epithelia, and blood, but this correlation worsened with increasing age.
) C! z% M9 M; v. ^/ V$ r! {2 h: `" d2 k
In searching for stem cell sharing, blood cells are good candidates. Not only are hemopoietic stem cells (HSCs) known to circulate, but experimental evidence for stem cell sharing is probably better than for any other organ. Two solid organs were chosen to compare with blood, GI epithelium, and blood vessels. We posited that evidence of stem cell sharing might be most dramatic in organs with initially dissimilar ratios. A likely candidate for this situation is GI epithelium, where direct visualization has indicated that murine X chromosome lacZ transgene inactivation is late, with commitment preceding X inactivation , although the situation for other cells of vessel walls is less clear. To detect stem cell sharing in normal adult females, we therefore compared XIPs of GI epithelium and varicose veins with those of blood across a broad age range.9 A0 i8 U$ d7 M
# g% |) f; B- {0 e& ~MATERIALS AND METHODS. Y4 U: \1 C$ r3 r
@3 t: n+ i: M* _8 dSamples, @. s, i0 p( f) Y& B. o7 H2 x
9 h# a) w( o' w, ]4 O+ O* P2 |0 {All protocols were approved by Grampian Research Ethics Committee, and all patients donating tissues gave consent beforehand. Normal GI mucosa was sampled from females attending for diagnostic endoscopy. Prospective case note review excluded patients with known GI inflammatory diseases (inflammatory bowel disease, celiac disease, or gastritis), whereas patients demonstrating macroscopic pathology were excluded by experienced endoscopists. Patients were recruited to sample a wide range of ages (upper GI endoscopy: n = 29; range, 22¨C88 years; mean, 59; SD, 20; colonoscopy: n = 11; range, 32¨C73 years; mean, 53; SD, 16). At upper GI endoscopy, paired mucosal 3-mm biopsies of gastric antrum and duodenal cap were taken; at colonoscopy, sigmoid colonic mucosa was biopsied. Typical yields were approximately 800 ng of DNA.6 f7 o2 r9 r- V! l1 n. g
) _0 ^( B% V6 W- i/ v, m. [Varicose veins were obtained from patients undergoing removal for therapeutic or cosmetic reasons. Veins were washed in saline to remove contaminating blood, and then two or three full-thickness samples were cut from either end of sufficient size to yield approximately 30 µg of DNA. Thirty-two patients were sampled with a mean age of 55 years (range, 28¨C81; SD, 14).
5 t4 k7 X( p6 K! b1 }9 ^
' q; n- t3 Z7 YDNA was extracted from solid tissues by incubation in 1 ml of lysis buffer (10 mM Tris¡¤Cl, pH 8, 0.5% SDS, 10 mM EDTA, 10 mM NaCl). One-hundred micrograms of proteinase K (Sigma-Aldrich, Poole, U.K., http://www.sigmaaldrich.com) was added before incubation overnight at 37¡ãC and freezing at ¨C20¡ãC. On thawing, an additional 1 ml of lysis buffer was added, and then 200 µl of 5x ANE (50 mM sodium acetate, 500 mM NaCl, 5 mM EDTA, 2.5% SDS) was added. After two phenol extractions, a phenol:chloroform extraction, and ethanol precipitation, the pellet was resuspended in 50¨C100 µl of TE (10 mM Tris¡¤Cl, pH 8, 1 mM EDTA).
# P, z0 }" Y$ F4 U6 J! q+ B& o/ w3 u' k+ i
DNA was extracted using whole blood from patients in the GI survey, but cells were separated into granulocytes and T cells using density gradient centrifugation (Histopaque 1077 and 1119; Sigma-Aldrich) and either beads or an EasySep T cell enrichment kit (Stem Cell Technologies, U.K., http://www.stemcell.com) in the samples corresponding to varicose veins. In all cases, an extraction kit was used (Nucleon Biosciences, Manchester, U.K., http://www.tepnel.com).
; T" |; v5 ]2 j5 b2 F: e% _8 X6 \! l2 \
X inactivation ratios were measured at the human androgen receptor by PCR using labeled primers and digestion with HpaII as previously described , except that areas under the curve were analyzed rather than peak heights. Results are expressed as proportion longer allele.
, e0 z; X- q7 j
* q" @9 L. Q7 M6 K) A& AStatistical Analysis9 {. f8 A6 S* P5 ?8 h$ u
" w5 ]0 l: A2 |; O; t2 F: V
All correlation coefficients are Pearson product moment coefficients. Regression equations with 95% confidence intervals were calculated using the method of least squares.+ ?, \4 D8 N2 G: H+ n* @
}- ]6 e# x7 {) r; K; |/ C2 a2 _Power calculations to calculate the minimum degree of detectable crossover are not available. A simulation study was performed to approximate whether this crossover is detectable experimentally. For each individual in the simulation study, a pair of proportions from a bivariate normal distribution with an initial correlation coefficient of i, and mean and standard deviations of upper allele frequency similar to those for the data in Table 1 were generated. These represent the XIP values for two different organs. The simulations were repeated n times to get a sample of size n.; |' }/ ?. O: g
1 R _7 e; ~( G, ~% m
Table 1. Subject details and X inactivation proportion data expressed as percentage larger allele
/ X, N& ~0 ?; n1 a+ U% o/ j) [) m+ E- ~" K6 ?$ Q. b
A second set of n pairs of proportions was generated in the same way, generating initial XIP proportions with the same mean and i. We approximated stem cell sharing in this second set by modifying the XIP of organ 1, 1, using the XIP of the second organ, 2, and vice versa as a function of the proportion of shared stem cells, x. The use of proportions makes the analysis insensitive to different stem cell pool sizes in the two organs, as long as these do not change with age. Such sharing should be exponential, with an asymptotic limit of the average XIP of the two organs, and it was approximated by the following function.
: ]% k0 g+ p( H: b# X6 N3 A6 [7 k; d) H6 C& u$ r* r$ r3 T- N! h) w
18 a9 Q; `; l' @3 N& ]( N
. G$ @4 U1 y% d* h I: C& VCorrelation coefficients were calculated for both sets of data. Let ri and rm be the correlation coefficients between XIPs in the two organs before and after stem cell mixing, respectively, and let n be the sample sizes for each set. If the proportion of upper alleles from each organ are normally distributed and the null hypothesis, that there is no stem cell mixing, is true, then2 |6 c1 D1 c/ e- P8 |
: h( \. P$ m! Y
28 Z, T) U3 i9 |' n7 M$ R. \
e7 S4 J5 A* x# S8 w- \0 x
is approximately normally distributed, with a mean of 0 and SD of 1. If the absolute value of z (the standardized normal deviate, measuring the distance from the mean in units of standard deviations) is greater than 1.96, then this test would be significant at a 5% level. We repeat this simulation a large number of times to get the proportion of tests that are significant at a 5% level, defined as the power.# @+ `, \# |0 t; O
- ?, U2 v+ @' [, @
RESULTS
1 b" o% D* N9 O3 H- {
3 `4 M, Q# a" ]5 ^Samples from 10 individuals (one colon/blood, five upper GI/blood, four veins/blood) were found to be noninformative at the human androgen receptor locus and were excluded from analysis.% M! k; T% h, ]9 A
" |. F) U' j: c8 V$ N9 _5 wData from informative individuals are shown in Table 1. As expected, XIPs from all organs were normally distributed. The mean, median, SD, and kurtosis were 47, 47, 12.5, and 0.81, respectively, for veins; 49, 45, 23, and ¨C0.55 for granulocytes; 48, 49, 18.5, and 0.21 for T cells; 54, 49, 15, and 0.8 for duodenum; 47, 44, 16, and ¨C0.2 for stomach; and 57, 54, 16, and ¨C1.3 for colon. Values from unrelated individuals were not correlated (analysis not shown). Measurement error has previously been shown to very low relative to interindividual variation .
D& B6 N& m& H: y+ o c
, U! r% R5 t9 K- n H1 P {8 N n1 [Correlations Between Samples from Organs Known to Have Similar Stem Cell Origins+ s5 y% Q- X( @' j4 C( q2 K
' D- s5 Z1 d' h* a( _! z- w
T Cells and Granulocytes. XIPs from T cells and granulocytes from the same subject were highly correlated (r = .61; n = 23) (Fig. 2A). This figure increases to r = .79 by excluding an outlier, which may represent either a premalignant clone or selection of an X chromosome .& h7 M7 a4 @ M, D1 P% K
' n7 q% C6 }/ ~! bFigure 2. Correlations of XIPs between samples from organs known to have similar stem cell origins. Each dot represents the XIPs from paired samples from a single individual. (A): T cells versus granulocytes on horizontal axis. (B): Vein sample 1 versus vein sample 2. (C): Stomach versus duodenum on horizontal axis.$ P+ z" d* ^0 H
9 H' r# X# Z5 e
Samples from Single Varicose Veins. To examine heterogeneity between samples from solid organs, we analyzed multiple biopsies from single varicose veins. The average distance between samples was 19 cm (SD 9.9). The correlations between samples from the veins showed no significant relationship with distance between the samples (r = .15; n = 39). Correlations between samples are shown in Figure 2B; the correlation coefficient between two samples from the same vein is .86 (n = 39)./ Z5 B6 @* a, s" y% U
5 [" X/ |, Q N* _( q9 M) T
Duodenal Mucosa and Gastric Mucosa. XIPs from paired duodenal and gastric biopsies were also correlated with one another (r = .63; n = 12) (Fig. 2C). Excluding one outlier increased the correlation coefficient to .89.
) S w% s3 S' ~% n5 g( n1 s4 U. t
* B% y; P* z8 Q. U0 r: T( eCorrelations Between Samples from Organs Known to Have Shared Only Distant Embryologic Stem Cell Precursors$ |0 ~( J3 J8 s8 i6 |; D) K1 x n
( K, I9 [1 `5 l5 d: ^1 N+ S i2 z( s
GI Tract and Blood. XIP values for gastric and duodenal samples were only moderately correlated with those from blood (gastric, r = .23 and n = 18; duodenal, r = .21 and n = 18) (Fig. 3A). Those from colon exhibited almost no correlation with blood (r = .04; n = 10) (Fig. 3B).4 K. H. n2 L! J+ ?) z# `( k* I
( l3 c; v6 w- `7 t+ [& H% g
Figure 3. Correlations of XIPs between samples from organs known to have shared only distant embryologic stem cell precursors. Each dot represents the XIPs from paired samples from a single individual. (A): Duodenum (open circles) and stomach (filled circles) versus blood on horizontal axis, with samples from single subjects linked by dotted lines. (B): Colon versus blood on horizontal axis. (C): Vein versus T cells (open circles) and granulocytes (filled circles) on horizontal axis.' w& j0 O5 S# f- }& [
& ]5 C' K `" `% i2 o$ S1 P
Vein and Blood. XIP values for vein samples were only moderately correlated with those from T cells (r = .42; n = 53 comparisons) and even less so with those from paired granulocytic samples (r = .18; n = 49 comparisons) (Fig. 3C).
/ D& j2 j! b6 u& V5 ^! N
, Q/ z+ j. d/ ]Age Dependence of XIP Correlations
2 H1 A) v0 Q! }3 }+ L0 L' ]7 V5 f- d+ d1 j* c
Initially, the age dependencies of XIPs were investigated within organs. The correlation of XIPs between T cells and granulocytes decreased with age, as has been previously reported with age, r = ¨C.38; p = .017; n = 39) (Fig. 4A).& L: h. \6 Y k2 U4 U
6 m- N/ r8 H Q( w) O8 lFigure 4. Dependence of differences between XIPs from paired samples on age. The modulus of one XIP subtracted from the other (|XIP1 ¨C XIP2|) is plotted against age of subject on the horizontal axis. (A): Comparisons between different samples from single vein. (B): Stomach (filled circles), duodenum (open circles), and colon (plus signs) minus the corresponding XIPs from blood samples. (C): Vein samples minus the corresponding XIPs from granulocytes (filled circles) and T cells (open circles). Abbreviation: GI, gastrointestinal tract.; p: E# V4 R! [1 O( c5 w/ L
3 n" D+ E9 D; X1 f# R% c. ?We next investigated the relationships of the correlations between solid organs and blood cells with age, again analyzed as the modulus of the differences in XIPs. It can be seen in Figure 4B that for most blood-tissue pairs the correlations showed no significant changes with age. Pooling the comparisons for GI tissue shows the differences between blood and GI tissue tend to increase slightly with age, but the effect is small and statistically insignificant (p = .69). The comparisons of granulocytes or T cells with venous tissue also yielded no significant differences (Fig. 4C) (p = .56 and 0.2, respectively; p = .66 pooled).
1 A$ y9 {9 z8 V" J% M- H$ z! U2 S1 n. p8 s5 ]/ f8 a) M( U
Analysis of Exclusion of Stem Cell Sharing6 T4 D3 u2 d' ?6 x
) |9 |- D/ m' B8 [1 aWe modeled the expected effects of stem cell sharing on the XIPs of paired organs as functions of two parameters: initial correlation coefficients and percentage of stem cell sharing. In Table 2, we give the power of tests to exclude stem cell mixing for n = 20 and n = 100, for 50,000 replicate simulations. The power of tests depends very little on the initial value of the correlation coefficient, at least for the range of values here. Although any given degree of stem cell sharing causes a larger change in correlation coefficient when the initial values are low, it can be seen that the smaller changes seen at higher initial correlations are equivalent from the point of view of statistical power. It is also apparent that this technique is insensitive to low degrees of stem cell sharing, for example, 1% cumulative. However, the numbers presented in this study are sensitive to large degrees (>30%) of stem cell sharing, which is equivalent to 0.6% sharing per annum over 50 years.
7 m# f q6 u% ?% v [' }% }- `6 q( `2 j: h1 O
Table 2. Analysis of power to detect stem cell sharing by analyzing XIPs from two cell populations within individuals: W4 J7 C$ {+ n6 x
. l) M- y$ b1 x$ rDISCUSSION
/ k; M- c; ^: D: N* a3 U) z' t; \0 j2 \) G# t
This study was prompted by the possibility of using XIPs to detect stem cell sharing between organs in normal individuals. Our main findings are, first, that samples from organs known to have similar stem cell origins exhibited similar XIPs. Second, organs known to have shared only distant, embryological stem cell precursors exhibited lesser or no significant similarities. Third, correlations between tissues did not become more similar with age.: V1 B) Z1 B' L2 P
# F9 I, a, I" q( z! n; e
XIPs from T cells and granulocytes are highly correlated, as previously reported. XIPs in different samples from the same tubular organ, both vein and gut, are also highly correlated and so closely related to one another. This may result from either derivation of cells from a common precursor population in utero and/or stem cell sharing between the biopsy sites later in life. The observation that XIPs from vein samples became significantly more similar with increasing age supports the view that there is some sort of mobility of venous stem cells in adult life.* C" f, F, W7 J, c5 w
9 |$ E2 C* x z' V5 x0 T, a
The standard deviation of values from solid organs is less than that observed from peripheral blood samples, which probably indicates derivation from a larger number of precursor cells at the time of X inactivation. It is also possible that solid organs are not subject to processes that increase the heterogeneity of blood samples as discussed in the introduction. In any event, it is clear that a sufficiently large number of stem cells is being sampled in our biopsies to give values that are representative of that organ. It should be stressed that the DNA was derived from full-thickness samples and therefore comprises mainly connective tissue and muscle cells; endothelial cells are a small minority of the total. Venous wall would also include some lymphocytes and myeloid cells, which might explain at least part of the moderate correlations that we observed. If so, there would be less, and perhaps no, correlation between XIPs of the connective tissue of the vessel wall and peripheral blood.
& f0 Q: {- r- o2 z6 @
! h) D- W9 C, |& P3 ?: e1 BIf there is large-scale, continuous sharing of stem cells between organs, their XIPs should be similar at all ages. If there were a lesser degree of ongoing sharing, initial differences would be expected because of unequal allocations of embryonic stem cells during embryogenesis but gradual, subsequent equalization of ratios with time. We found that correlations between samples from different organs are low and do not become closer with increasing age, in accordance with the findings of Sharp et al. . The simplest explanation of these data is that there is little or no significant sharing of stem cells between gut or veins and blood. It should be stressed that our data do not address the question whether stem cells have the ability to be shared between organs; they merely support the view that if any such phenomenon occurs, it cannot be widespread. We studied multiple samples from 60 individuals and so estimate that our study can exclude a cumulative 10%¨C20% (Table 2) sharing of stem cells, equivalent to about 1 in 400 per year. Furthermore, we analyzed whole tissue biopsies, not individual cell types, and we confined our study to tissues that had not undergone any obvious repair to injury, a process that has been associated with stem cell plasticity.
" b x, F% z& b0 r" V4 `
( i* ^) r4 n) q# s- U% V+ \( i& zAn alternative explanation for our findings is that stem cell sharing may be an ongoing process, but other processes that cause XIPs to become more different with time mask its effects. For instance, there may be genetic selection for one X chromosome over the other, with the direction of selection being different in the two organs. Another alternative explanation is that stem cells are mobile between organs but different sets of stem cells contribute to different organs at different times. Although the stability of XIPs has not been studied exhaustively, observations over a few years give stable ratios, indicating that this explanation is unlikely .' Z6 d0 ]9 C8 \* U$ h
# w# d& v2 A7 C! E, i9 nCONCLUSION
3 N' Y K( i2 Q3 t$ Z
( `! H; L, ?0 r) D7 zX inactivation proportions can be exploited to track cell lineages. Using this approach, we find tissues with close, but not distant, stem cell origins have closely related XIPs. However, the degree of similarity does not increase with age, implying that stem cell sharing between tissues is uncommon in healthy human volunteers.
( L- }1 n* g4 j# Y; }* y' ~5 B5 H1 F1 k' w4 M0 z8 k& {3 r% X4 l
DISCLOSURES* G( l; I) E, P; V8 v8 \
) Y; \! Q3 s) f9 u4 g4 p
The authors indicate no potential conflicts of interest.4 ?* U& K8 b1 Z! P, }: ?! e1 T4 i
% p" a0 \5 N! c* ~/ M& }% R* \: ]4 Q
ACKNOWLEDGMENTS, m: r# B& i& R8 X6 C0 ?
; j" F- B; Z9 _: U! O! `
This work was funded by Aberdeen Royal Infirmary Leukemia Research Fund.
, V- o! k5 e% U/ Q' |7 e 【参考文献】 R9 l2 a! _" n- q6 Q9 d( w" W; c
5 m; |2 ~4 s9 R v% S2 A1 Q& b' O# _9 {( k# y- g! d5 ]
Herzog EL, Chai L, Krause DS. Plasticity of marrow-derived stem cells. Blood 2003;102:3483¨C3493.- J7 }4 a0 t0 x( z
, B8 \9 d- Y8 \
Horwitz EM, Prockop DJ, Fitzpatrick LA et al. Transplantability and therapeutic effects of bone marrow¨Cderived mesenchymal cells in children with osteogenesis imperfecta. Nat Med 1999;5:309¨C313.8 \4 E* Y l; @9 B3 e4 |/ g
8 [, ]9 Q. q3 I! W& _Theise ND, Badve S, Saxena R et al. Derivation of hepatocytes from bone marrow cells in mice after radiation-induced myeloablation. Hepatology 2000;231:235¨C240.
6 q8 ^* u3 t' B
. \+ X, R1 S; a0 i6 SKörbling M, Katz RL, Khanna A et al. Hepatocytes and epithelial cells of donor origin in recipients of peripheral-blood stem cells. N Engl J Med 2002;36:738¨C746.+ t! r& e3 ?7 N+ w3 `
% [$ M9 A2 Z# g/ O+ a; z+ D& l% gOkamoto R, Yajima T, Yamazaki M et al. Damaged epithelia regenerated by bone marrow-derived cells in the human gastrointestinal tract. Nat Med 2002;8:1011¨C1017.! G0 x- |' l- `% t- y" W5 C% u% P
" |' z( n y# @0 t- ]/ c' v1 i2 U
Shih C-C, Weng Y, Mamelak A et al. Identification of a candidate human neurohematopoietic stem cell population. Blood 2001;98:2412¨C2422.
2 c8 [9 y" j, D/ a- J9 u. o' X7 U$ Y& T$ R
Weissman IL. Normal and neoplastic stem cells. Novartis Found Symp 2005;265:35¨C50.) C9 }+ L& g9 J8 M5 f9 F1 M" D
: z3 l! v* }+ q. _/ i' n% g
Kucia M, Ratajczak J, Ratajczak MZ. Are bone marrow stem cells plastic or heterogenous¡ªThat is the question. Exp. Hematol 2005;33:613¨C623.
4 c+ e3 t/ q; C) \$ g0 J4 M. E
& E7 l) t+ Y9 C- Z6 HKrause DS, Theise ND, Collector MI et al. Multi-organ, multi-lineage engraftment by a single bone marrow¨Cderived stem cell. Cell 2001;105:369¨C377.5 l5 a* G& Z7 U2 N; r( X# q
7 K) D; [) T* x% O2 b- U
Wagers AJ, Sherwood RI, Christensen JL et al. Little evidence for developmental plasticity of adult hematopoietic stem cells. Science 2002;297:2256¨C2259.
- Z" `" A+ ~, `" X' W6 \. M9 a9 f6 Q
9 c1 R3 Q: i3 x* C0 y5 \6 }9 mBjornson CR, Rietze RL, Reynolds BA et al. Turning brain into blood: A hematopoietic fate adopted by adult neural stem cells in vivo. Science 1999;283:534¨C537.0 L' Z- n# @6 w |) x O3 e9 ~# e) |
; B. F9 X/ @! L/ s7 @Morshead CM, Benveniste P, Iscove NN et al. Hematopoietic competence is a rare property of neural stem cells that may depend on genetic and epigenetic alterations. Nat Med 2002;8:268¨C273.
) Z' f3 V( _. I' T. o4 a( ?
: A- P! V, _9 s7 [4 ITerada N, Hamazaki T, Oka M et al. Bone marrow cells adopt the phenotype of other cells by spontaneous cell fusion. Nature 2002;416:542¨C545.
% @& I1 w* F+ V; D! n8 L" o |/ y( \" I' y
Ying QL, Nichols J, Evans EP et al. Changing potency by spontaneous fusion. Nature 2002;416:545¨C548.6 T' l" U: f: ~8 ^5 I
8 k8 y) j. p' K3 v3 k2 ~5 ~Monteiro J, Derom C, Vlietinck R et al. Commitment to X inactivation precedes the twinning event in monochorionic MZ twins. Am J Hum Genet 1998;63:339¨C346.% C. @/ C. K& R3 K; e2 w
, e' p" u! q* r/ c4 L7 y; ^7 h
Boumil RM, Lee JT. Forty years of decoding the silence in X chromosome inactivation. Hum Mol Genet 2001;20:2225¨C2232.7 b+ V4 _- z1 b9 B! M# D0 o
) n( X8 v( \, o$ I& [$ @
Sado T, Ferguson-Smith AC. Imprinted X inactivation and reprogramming in the preimplantation mouse embryo. Hum Mol Genet 2005;14:R59¨CR64.% r9 u+ A; o; r3 k* l. o2 h0 u
e) X* R6 w; E; X+ ?) a' T! {& m3 bTan SS, Williams EA, Tarm PPL. X chromosome inactivation occurs at different times in different tissues of the post implantation mouse embryo. Nat Genet 1993;3:170¨C174.
# \3 a7 v$ {% i+ |! R$ [7 q
3 x+ f! p s5 S; s, U. x9 ?" uHoffman LM, Hall L, Batten JL et al. X Inactivation status varies in human embryonic stem cell lines. STEM CELLS 2005;23:1468¨C1478.
8 _) K' z0 z' v& y; b+ v! p8 {9 \" J! v3 t5 \
Vickers MA, McLeod E, Spector P et al. Assessment of mechanism of acquired skewed X inactivation by analysis of twins. Blood 2001;97:1274¨C1281.3 Q* o$ V5 a( c3 |' x! w8 R* d
, t, j1 ^. z' G" X* S4 qGale RE, Wheadon H, Boulos P et al. Tissue specificity of X chromosome inactivation patterns. Blood 1994;83:2899¨C2905.
! ]" `' g+ q8 d1 E
# Q2 Q: s4 h$ N8 ~9 j7 uAzofeifa J, Waldherr R, Cremer M. X chromosome methylation ratios as indicators of chromosomal activity: Evidence of intraindividual divergencies among tissues of different embryonal origin. Hum Genet 1996;97:330¨C333.
4 A" G7 Y2 }- J: k2 L9 ?* O
0 q9 q* b1 n# t1 A& P8 B% w. YTonon L, Bergamaschi G, Dellavecchia C et al. Unbalanced X chromosome inactivation in haemopoietic cells from normal women. Br J Haematol 1998;102:996¨C1003.
) Q' u c1 ~' w E, s L" x4 H$ L% G. y: H0 D/ T
Sharp A, Robinson D, Jacobs P. Age- and tissue-specific variation of X chromosome inactivation ratios in normal women. Hum Genet 2000;107:343¨C349.7 r' Z- X# h/ L- l0 O( O6 m3 c: n' B
7 Y( C3 d2 o3 F+ t1 X$ V8 n
Forrai A, Robb L. The hemangioblast-between blood and vessels. Cell Cycle 2003;2:86¨C90.
7 R. e3 F! M9 u; x C
z8 r# ^$ o* a/ t( w% R9 ?0 KShi Q, Rafii S, Wu MH et al. Evidence for circulating bone marrow¨Cderived endothelial cells. Blood 1998;92:362¨C367.
- P/ u$ r9 k% H) ?+ d/ [0 X6 S1 Q9 k8 S8 O3 K0 t3 }8 b
Gale RE, Fielding AK, Harrison CN et al. Acquired skewing of X chromosome inactivation patterns in myeloid cells of the elderly suggests stochastic clonal loss with age. Br J Haematol 1997;98:512¨C519. B% q9 Y. j: z$ U
( l3 J: b9 x; J- e7 s( _Champion KM, Gilbert JG, Asimakopoulos EA et al. Clonal haemopoiesis in normal elderly women: Implications for the myeloproliferative disorders and myelodysplastic syndromes. Br J Haematol 1997;97:920¨C926./ g# e, C4 {% Q- x* q
/ M0 z L: r4 q
Busque L, Mio R, Mattioli J et al. Nonrandom X inactivation patterns in normal females: Lyonization ratios vary with age. Blood 1996;88:59¨C65. k/ ]; U4 E* I/ i
B; i0 q! \% C# cSmith DM, Tabin CJ. Clonally related cells are restricted to organ boundaries early in the development of the chicken gut to form compartment boundaries. Dev Biol 2000;227:422¨C431.
9 L& `9 p, h! q$ ~1 ^% i. C9 B, [1 N0 h& L( n. x N' k& D
van Dijk JP, Heuver L, Stevens-Linders E et al. Acquired skewing of Lyonization remains stable for a prolonged period in healthy blood donors. Leukemia 2002;16:362¨C367.
# I7 e) ]1 |8 y# I- i+ b' C0 A, m6 |" ?
Thornley I, Sutherland R, Wynn R et al. Early hematopoietic reconstitution after clinical stem cell transplantation: Evidence for stochastic stem cell behavior and limited acceleration in telomere loss. Blood 2002;99:2387¨C2396. |
|