| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Departments of Paediatric Gastroenterology and
Obstetrics and Gynaecology, St. Bartholomew’s and the Royal London School of Medicine and Dentistry, London, United Kingdom; Departments of
Histopathology, and
§
Rheumatology, Guy’s, King’s College and St. Thomas’ Hospital, Medical and Dental School, London, United Kingdom;
¶
Department of Histopathology, University College London, London, United Kingdom; and
||
Department of Clinical Immunology, Royal Free School of Medicine, London, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. There is also a substantial proportion of
CD7+ LPLs that do not express CD3
(CD3-7+, 14 ± 7% of all LPLs) in
the fetal gut that may be differentiating into CD3+ cells.
Rearranged TCR-ß transcripts of fetal LPLs, IELs, and paired blood
lymphocytes were cloned and sequenced, and virtually no overlap of
clonality was observed between blood and intestine, suggesting that gut
T cells may not be derived from the blood. In addition, 30 days after
engraftment of SCID mice with fetal intestine,
CD3-7+ cells, proliferating T cells, and
pre-TCR- transcripts were abundant, and there is a threefold
increase in CD3+ IELs. These data show that in the human
intestine before birth a population of precursor T cells exists that
may be differentiating into mature T cells in
situ |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
TCR+ IELs found
in normal euthymic mice but have few TCR-ß+ IELs, the
relative proportions of ß and IELs varying with the age of
the mice (3, 4, 5). Thymectomized and reconstituted mice (ATXBM, ATXFL),
which are reconstituted with either adult bone marrow or day 12 fetal
liver, can generate and TCR-ß+ IELs at about
40% of normal levels with a lag period of around 6 mo following bone
marrow or fetal liver transfer (6). These chimeric animals have no T
cells in their lymph nodes or Peyer’s patches. Neonatally
thymectomized mice contain CD8 TCRlow cells in their
intestinal epithelium, which are thought to represent intermediate
stages in T cell development (7). Thymectomized RAG-/-
mice reconstituted with bone marrow from nude mice generate
TCR-ß and IELs that exclusively express the CD8
homodimer (8). Evidence for a similar function for the human intestine is scarce. mRNA transcripts for RAG-1 and RAG-2 have been detected in human IEL but not lamina propria lymphocytes (LPLs) (9). Terminal deoxynucleotidyltransferase (TdT) is undetectable in adult human IEL (10), thus any role for RAG-1 and RAG-2 in the formation of the adult gut IEL repertoire must be quite small because it has been reported that for adult IEL, TCR ß-chain mRNAs contain numerous N additions (11, 12).
Human fetal intestinal mucosa contains T cells in the lamina propria
and epithelium from 12–14 wk gestation (13). Although both mice and
humans have intestinal IEL from early in ontogeny, there are
fundamental species differences between them. In mice,
+ IELs predominate in young animals (<8 wk) after
which both and ß IEL numbers increase; however, IELs
still predominate (40–80% of CD8+ IELs (14)). In humans,
+ IELs are a relatively small population accounting
for only 
10% of total IEL numbers throughout life (15, 16, 17); the
remaining IELs use the TCR-ß. Also in humans, there is no evidence
that the IEL population is thymus-independent.
+ IEL numbers only increase in humans during coeliac
disease (18).
It is difficult to reconcile the presence of numerous T cells in the human fetal intestine with the concept of the human gut mucosa being an immune effector organ if T cells are present when the exposure to gut Ags is minimal. An alternative theory could be that T cells in the human fetal gut develop locally from immature precursors.
A recent study has demonstrated the existence of TCR-ß rearrangements
in fetal IELs at 14 wk gestation in the absence of TCR-
rearrangement (19). This study also showed that "T early " was
abundantly expressed in early gestation fetuses along with a new
tentatively described "T early ß" mRNA with unknown function.
These data provide evidence that TCR rearrangement is ongoing in the
fetal intestine from an early stage, but do not address the issue of
whether the cells are differentiating into T cells locally or are
immature thymic emigrants that home to the gut.
In this study, we have used immunohistochemistry, analysis of
pre-TCR- (pT) and TCR Vß gene expression, and transfer
of fetal human intestine into SCID mice to provide compelling evidence
that T cells that are found in human fetal gut are markedly different
from those seen in the postnatal intestine and most probably arise from
the CD3-7+ cells abundant in the fetal gut.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study received ethical committee approval from the City and East London Health Authority. Human fetal intestine was obtained from the Medical Research Council Tissue Bank, Hammersmith Hospital London, or from the Homerton Hospital, London. Fetal blood was obtained by cordocentesis at the Homerton Hospital, London. All specimens were between 12 and 27 wk gestation as assessed by crown/rump measurements. Tissue was snap frozen in liquid nitrogen for immunohistochemistry and dissected into 1- to 2-mm2 fragments for organ culture or 1-cm segments for transplantation into SCID mice. Jejunal biopsies were obtained from nine infants aged 3 mo to 6 yr attending Queen Elizabeth Hospital Hackney. These children had chronic diarrhea of 2-wk duration, but the biopsies were histologically normal.
Mice
SCID (BALB/c) mice (20) were kept in specific pathogen-free conditions at St. Thomas’ Hospital, London. Mice were transplanted with 1-cm segments of fetal intestine at 6 wk of age as described elsewhere for rheumatoid synovium (21). Animals were killed 4 wk later and the transplanted fetal intestine was then snap-frozen for histology and RT-PCR analysis.
Immunohistochemistry and immunofluorescence
CD3-7+,
CD3+4-8- and CD3+,
Ki67+ cells were detected by sequential double
immunocytochemistry (22) using a peroxidase-conjugated secondary and
diaminobenzidine with the first Ab to give a brown reaction
product and an alkaline phosphatase-conjugated Ab and fast blue with
the second mAb to give a blue reaction product. Dual immunofluorescence
staining was used to detect CD3 cells expressing CD4, CD45RO, or HLA-DR
using a rhodamine-conjugated secondary Ab to detect CD4, CD45RO, or
HLA-DR immunoreactivity followed by a biotinylated anti-CD3 Ab and
streptavidin-conjugated FITC to detect CD3+ T cells. T
cells staining green and cells expressing CD4, CD45RO, and HLA-DR
staining red were detected using the green and red fluorescence filters
of a Leitz Diaplan microscope with a fluorescent light source
(Leitz, Wetzlar, Germany). CD103+ and CD62L+ T
cells were enumerated using single immunoperoxidase staining. Abs were
UCHT-1 (CD3; Dako, High Wycombe, U.K.), UCHL-1 (CD45RO; Dako), CR3–43
(HLA-DR; Dako), Leu-3a (CD4; Becton Dickinson, Oxford, U.K.), Leu-8
(CD62-L; Becton Dickinson), RFT-8 (CD8; Royal Free Hospital, London,
U.K.), SN-130 (CD45-RA; Royal Free Hospital), HML-1
(Eß7; Dako), biotinylated anti-CD3
(Serotec; Oxford, U.K.), streptavidin-FITC (Serotec), horseradish
peroxidase rabbit anti-mouse IgG (Dako), alkaline phosphatase
rabbit anti-mouse IgG (Dako), and rhodamine conjugated rabbit
anti-mouse IgG (Dako). For all immunocytochemistry experiments,
negative controls were included using secondary Abs alone or irrelevant
isotype controls. Apoptotic T cells were identified by staining
sections with anti-CD3 followed by FITC-conjugated rabbit anti-
mouse IgG. Sections were then fixed in 10% formal saline for 10 min
and washed in TBS for 10 min. They were then incubated with 0.01%
propidium iodide (Sigma, Poole, U.K.) in Tris-buffered saline (TBS) for
3 min, washed in TBS, and mounted.
Image analysis
For quantification of T cell subset density by immunocytochemistry, areas of lamina propria were mapped and measured using a computer-controlled cursor with a Seescan image analysis package (Seescan, Cambridge, U.K.). In this manner, T cell numbers in the lamina propria were calculated per square millimeter. Similarly, the length of epithelium was measured using a calibrated mouse-controlled cursor, and IEL numbers calculated per millimeter of epithelium.
RT-PCR and Southern hybridization
Total RNA was extracted from tissues using TRIzol (Life
Technologies, Paisley, U.K.) according to the manufacturers
instructions. In some cases, before reverse transcription the RNA was
treated with 3 U DNase-I (Promega, London, U.K.) for 30 min at 37°C
(pT, Cß, and C) and PCR reactions performed with RNA alone to
control for the possibility of contaminating DNA in the RNA
preparations. RNA was reverse transcribed using 100 U of Moloney murine
leukemia virus reverse transcriptase (M-MLV RT, Life
Technologies) in a 20-µL reaction containing 50 mM Tris-HCl, 75 mM
KCl, 3 mM MgCl2, 3 mM DTT, 500 µM each 2'-deoxynucleoside
5'-triphosphate (dNTP), and 100 ng oligo thymidine
5'-triphosphate (dTTP).
PCR for TCR Vß family members was performed as described elsewhere
(11). Southern hybridization for TCR Vß families was performed after
35 PCR cycles, which was previously assessed to be in the linear phase
of amplification. Primers for TCR Cß were sense
5'-GTCCACTCGTCATTCTCCG-3' and anti-sense
5'-GGCTCAAACACAGCGACCT-3', C sense 5'-CCAGAACCCTGACCCTGCCGTG-3'
and anti-sense 5'TATGGATCCCAGGGAGCACAGGCTGTCTT-3', and pT sense
5'-GGCACACCCTTTCCTTCTCTG-3' and anti-sense
5'-GCAGGTCCTGGCTGTAGAAGC-3'. For Cß, C, and pT, PCR was
performed with a hot start of 94°C for 2 min followed by annealing at
56°C for 10 min, followed by 35 cycles of 94°C for 30 s,
annealing at 56°C for 30 s, and extension at 72°C for 1 min.
PCR products were run on 1% agarose gels, transferred to nylon
membrane (Magna, Sartorius, Epsom, U.K.) using standard techniques and
hybridized to 32P end-labeled (>109 cpm/µg)
internal oligonucleotide probes. Internal probes used were TCR Cß
5'-GCCTTTTCCCTGTGGGAGAT-3', C 3'-ACTGTGCTAGACATGAGGTCTA-3' and
pT 5'-CAATGGCAGTGCACTGGATGCC-3'. Hybridization was conducted in
6x SSC with 0.5% SDS at 66°C overnight followed by a 5-min wash at
58°C in 6x SSC and a 5-min wash in 2x SSC 0.5% SDS at 58°C.
Filters were then exposed to x-ray film overnight at -70°C. Negative
control PCR products were included in all hybridizations to control for
probe specificity. In addition, to confirm the identity of the
amplified pT RT-PCR products these were purified over Wizard
PCR-Prep columns (Promega) and digested with Alu-I (Promega), which
gave characteristic 110-bp and 226-bp fragments on agarose gel
electrophoresis. Purified pT PCR products were also cloned and
sequenced to confirm their identity.
Cloning and sequencing of TCR Vß PCR products
TCR RT-PCR products were ligated into the pMOSBlue T vector using the pMOSblue T vector kit (Amersham, Buckingham, U.K.) according to the manufacturers instructions and cloned into MOSblue competent Escherichia coli. Individual recombinant colonies were PCR screened using T7 and universal primers flanking the multiple cloning site. Clones were chosen for sequencing based on PCR product size, the TCR PCR product with short 5' and 3' plasmid flanking sequences being around 370 bp.
PCR products of the recombinant colonies were directly sequenced using the dideoxy chain termination method using T7 Sequenase (Amersham) in the forward and reverse direction. Sequencing reactions were run on a Sequigen II manual sequencing apparatus (Bio-Rad, Hertfordshire, U.K.).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
IEL
Immunohistochemical staining of cryostat sections from human fetal
intestine aged 12–27 wk gestation reveals the abundance of
CD3+ cells in fetal gut, both in the lamina propria and
epithelium. This is despite the lack of lumenal stimulation from
enteric Ags (Figs. 1
and
2a). IEL density increases
steadily with age up to the oldest specimen examined, where there was
around two-thirds of the number seen in postnatal specimens (27 wk
gestation; Fig. 1). The Ab TCR-1 stains around 30% of
epithelial T cells, showing the majority to be TCR-ß+
(data not shown and 15 . CD3-7+ cells,
the phenotype of marrow-derived pre-T cells, are rare in the fetal
epithelium (0.6 ± 0.3% of total CD7+ cells) (23).
Dual staining with CD2 and CD7 on cryostat sections also reveals many
CD2-7+ IELs (15.5 ± 2.6% of
CD7+ IELs are CD2-; Fig. 3A). Significantly, 50% of
fetal IELs are CD8+, the remainder being subset negative
(CD3+4-8-). The majority of fetal
IELs are CD45RO+ (63 ± 36%), the phenotype of memory
cells. Virtually no fetal IELs are CD62L+ and 100% express
HML-1 (Eß7 integrin). Interestingly,
10% of fetal IELs are in cell cycle (9.1 ± 1.4%) as revealed
by double staining with Ki67. IELs of all subsets analyzed are dividing
(Fig. 3B). This is in marked contrast to healthy adult gut,
which contains virtually no dividing T cells in the epithelium or
lamina propria in vivo (24). Because the rate of proliferation seen for
IEL in normal fetal intestine is exceptionally high compared with the
relatively slow increase in IEL number seen during gestation, the
cells must either die or leave the intestine. The former is
likely to be the case because CD3+ IELs with fragmented
apoptotic nuclei can be seen and these account for up to 5% of all IEL
(4.5 ± 2.1%).
|
|
). The density of LPLs increases steadily
during gestation from 12 wk up to around 18 wk where their number is
approximately one third of that seen in postnatal specimens (Fig. 1).
This is in contrast to IEL, which increase in number steadily up to 27
wk of gestation. T cell density in fetal gut lamina propria is variable
within a single specimen, and there are many accumulations of large
blast-like CD3+ cells in the villi (Fig. 2c),
quite different from the organized T cell zones of the Peyer’s
patches. The Ab TCR-1 only stains 5% of lamina propria
T cells, indicating that the large majority of fetal LPLs are
TCR-ß+ (data not shown and 15 .
|
a, insert) in specimens of all ages
in contrast to the epithelium. These cells are very rare in Peyer’s
patches (<0.01%, data not shown), and in postnatal gut are only
occasionally seen in the epithelium at the tips of the villi (23). Dual
staining with CD2 and CD7 also reveals many LPLs to be
CD2-7+ (25 ± 3.4% of CD7+
LPLs are CD2-; Fig. 3A).
There are also many cells with weak cytoplasmic CD3 staining, quite
unlike the dense membrane staining seen in mature T cells (Fig. 2b), and, more rarely, it is possible to see LPLs with
membrane CD7 and cytoplasmic CD3 (Fig. 2b). CD4+
T cells outnumber CD8+ cells in the lamina propria of all
specimens, regardless of gestational age (Fig. 3A). However,
there is a large population of
CD3+4-8- LPLs in all specimens
(30 ± 7.3% of CD3+ LPLs are
CD4-8-). The majority of fetal LPLs bear
activation markers, a few are CD25+ (Fig. 2d),
and the majority are CD45RO+ (65 ± 2.8%).
There is also expression of CD103 (HML-1,
Eß7 integrin, 36 ± 2.8% of
CD3+ LPLs), HLA-DR (26 ± 4.1%), and virtually no LPL
are L-selectin positive.
In addition to IEL, LPL are also actively dividing as revealed by Ki67
staining (Fig. 2e). The number of dividing LPL is greater
than IEL (22 ± 1.8% LPL, cf 9.1 ± 1.4% IEL). LPLs of all
subsets are dividing (Fig. 3B). Fetal LPL with apoptotic
nuclei are also visible in specimens of all gestational ages examined
with 8 ± 1.8% of CD3+ LPL having apoptotic nuclei
(Fig. 2f).
Fetal LPL, IEL, and blood lymphocyte TCR-ß repertoires are diverse and separate
Analysis of the expressed TCR Vß usage by fetal LPL, IEL, and
matched blood populations by semiquantitative RT-PCR and Southern
hybridization reveals a diverse repertoire of Vß expression with most
of the Vß families analyzed being represented (Figs. 4 and 5).
Two representative 15 wk gestation specimens of four analyzed are
shown. The relative contribution to the TCR repertoire by individual
Vß family members varies between the three sites in individual
specimens. In addition, RT-PCR and Southern hybridization analysis was
performed on IEL and LPL derived from distal segments of the same gut
samples as shown in Figs. 4 and 5, and very similar patterns of TCR
Vß expression were observed (data not shown). The polyclonal nature
of the IEL population shown here by RT-PCR confirms our earlier studies
based on immunohistochemical analysis of fetal intestine (25). These
data contrast with analyses of postnatal IEL that are markedly
oligoclonal (11, 12). If CD3+ T cells in the fetal
intestine are derived from blood, it should be possible to observe
overlap in the CDR3 usage between blood and intestinal T cells;
however, we have found that this is not the case. We performed random
cloning and sequencing of TCR CDR3 regions from the PCR products of the
Southern analysis (samples 1 and 2). Vß families were chosen for
sequencing based upon their relative expression in the Southern
analyses. Families that had either relatively high expression in blood
compared with IEL/LPL (BV5, sample 1) or high expression in blood and
IEL compared with LPL (BV18, sample 1) or equal expression between the
three sites (BV12 and BV17, sample 2, data not shown) were chosen for
sequencing. This was done to control for the varying amount of TCR mRNA
between the three sites, which could have an effect on the number of
identical clones isolated. The results of this analysis are shown in
Tables I and
II. Sequences isolated
from blood, LPL, and IEL were unique, indicating that the IEL and LPL
populations had been isolated successfully. Table I shows the results
of CDR3 sequencing of BV18 from sample 1. Of 63 recombined BV18
sequences, only one sequence was isolated more than once, two from PBLs
and two from IELs; the remaining 61 sequences were unique. The majority
of isolated sequences were productively rearranged and in frame as
shown in Table I and II. Similarly, TCR BV5 transcripts isolated from
sample 1 were diverse, with one clone isolated once from LPL and IEL
but not blood. N-region diversity was extensive in blood, IEL, and LPL,
indicating that TdT activity is present at 15 wk gestation in the human
fetus. It is also noteworthy that there was a biased usage of Dß1.
(Dß2.1 was only isolated from 32/240 clones). Thus, there is very
little overlap in clonality between blood and intestinal T cells in the
fetus, suggesting that these populations may arise separately.
|
|
|
transcripts are abundant in both fetal lamina propria and
epithelium, and fetal intestine transplanted into SCID mice contains
proliferating T cells and immature T cell subsets
RT-PCR analysis reveals that mRNA transcripts for pT are
detectable in fetal lamina propria and epithelium (Fig. 6A). The PCR products for
pT were cloned and sequenced to confirm their identity in addition
to Southern hybridization. To answer the question of whether the pT
mRNA is from cells that are blood-derived thymic emigrants or
represents a T cell progenitor population maturing in situ, we
transplanted segments of fetal intestine into SCID mice and analyzed
the T cell subpopulations before and after transplantation. It would be
expected that if the cells containing the pT message were recent
thymic emigrants, then after 30 days in a SCID mouse any
pT-expressing cells would have arisen locally because, in the thymus
at least, these cells are short-lived (26). Fig. 6B shows
that after 30 days engraftment into SCID mice pT transcripts were
still detectable, suggesting the presence of a pool of T cell
precursors in the fetal gut. Phenotypic analysis of lymphocytes in the
transplants before and after transplantation shows that after
engraftment there are still abundant
CD3+4-8- cells in the epithelium
and lamina propria (Fig. 7). In addition,
the CD3+ T cells present after engraftment continue to
divide (Fig. 7), and cells with a T cell precursor phenotype,
CD3-7+, are still present in the lamina
propria (Table III). The most
startling observation in these experiments is that following
transplantation the density of IELs increased approximately threefold
(Fig. 8 and Table IV); these IELs are mostly
TCR-ß+ (Fig. 7) because the proportion of
TCR-+ IELs decreases during the 30 day engraftment
period.
|
|
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
transcripts can be detected before
and after 30 days of engraftment. T cells continue to divide in the
grafts, and following engraftment there is a threefold increase in IEL
numbers over day 0 levels. In recent years, there has been intense interest in the phenomenon of extrathymic maturation of T cells, and many experimental mouse models have shown that gut T cell populations, especially in the epithelium, can mature in the absence of thymic influences. Embryologically and phylogenically, it would make sense for the intestine to be able to support T cell maturation because the intestine, like the thymus is endodermally derived, arising from the third pharyngeal pouch during embryologic development (27). It has been suggested that T cell generation in the gut associated lymphoid tissue (GALT) predates the emergence of the thymus in evolution with the gut, playing a major hematopoietic role before the development of the thymus (2). The evolutionary importance of the gut in T cell development is well-illustrated by the hagfish, which has no thymus, but T cell maturation and education takes place in the intestinal lamina propria (28).
Two recent studies have reported that production of T and B cells
may take place in novel crypt lamina propria lymphoid aggregates named
cryptopatches in mice (29, 30). Notably these structures are present in
nude, SCID, and RAG-2-/- mice but are absent in
IL-7-/- mice, suggesting their independence from the
thymus and hematopoietic characteristics, respectively.
c-kit+ cells isolated from cryptopatches are
capable of reconstituting the ß and IEL populations of SCID
mice upon adoptive transfer (30). However, cryptopatches cannot be
identified in human fetal intestine.
Efforts to find similar populations of extrathymically derived T cell
populations in humans have been less successful. We have previously
described a population of CD8ß- (CD8) T cells in
human fetal gut epithelium and lamina propria that is scarce in adult
gut (31). This phenotype has been shown to be thymus-independent in
mice (5). RAG-1 and RAG-2 transcripts have been detected in adult human
gut epithelium, but no TdT or pT has been reported at this site (9, 10). However, the study reporting RAG in the adult gut was solely based
on RT-PCR and consequently gives no indication of the number of cells
expressing RAG. In addition, TdT is not expressed in adult intestine,
but the T cells in the adult gut have extensive TCR junctional
diversity. These observations suggest that any TCR recombination that
may be occurring in the adult gut does not contribute a great deal to
the expressed TCR repertoire, and the T cells in the adult intestinal
epithelium are most likely derived from the thymus. RAG-1 and -2 along
with pT have been detected in adult liver; however, it is not
understood to what extent T cell generation at this site contributes to
the adult T cell repertoire or what function T cell generation in the
liver serves (32). Evidence for the extrathymic generation of a
population of T cells has been reported in a patient with thymic
aplasia who had circulating T cells expressing the V2 chain of the
TCR (33). Because T cells expressing V2 are not present in
the thymus but V2 is expressed in a minority of adult
TCR+ IEL, and is the sole TCR type in fetal IEL, it has
been suggested that this population in the epithelium represents
bone marrow emigrants educated in the intestinal epithelium (15).
The human thymus is populated by fetal liver and fetal bone marrow
CD7+ emigrants between 7–8 wk of gestation. The
CD3-7+ T lymphocytes identified in abundance
in fetal gut in this study may represent a progenitor T cell population
yet to undergo TCR rearrangement. CD7 is an early T lineage marker not
found on myeloid or erythroid lineages and as such is a good marker for
T cells that have not yet expressed markers of later T cell subsets
such as CD3, CD4, or CD8 (34). Interestingly, in the fetal gut
epithelium there are more CD2-7+ IELs than
CD3-7+. As CD2 is expressed on developing T
cells before CD3 (35), we expected to see less
CD2-7+ IELs and more
CD3-7+. However, it has been reported in mice
that IEL are bimodal with respect to CD2 expression (36), with 50% of
TCR ß+ IEL being CD2- and TCR
+ IEL being predominantly CD2-.
Therefore, it is likely that subsets of human fetal IELs also do not
express CD2. Developing thymocytes at various stages of their
maturation display activation markers (37) and CD45RO (38). Expression
of the CD45RO isoform is associated with progression into programmed
cell death during negative selection (38). T cells in the fetal
intestine express markers of activation such as HLA-DR, CD25,
Eß7, and CD62L-, and the
majority also express CD45RO. Expression of a large array of activation
markers and CD45RO in the absence of antigenic stimulation from the gut
lumen makes it likely that these cells are acquiring these markers via
a pathway other than classical Ag stimulation and may be acquiring them
in a manner similar to developing thymocytes. Further evidence that the
T cells of the fetal intestine are not thymic emigrants comes from the
observation that they do not express 4ß7
integrin (Act-1), which is normally required for entry into the gut via
mucosal address in cell adhesion molecule (MAdCAM) expressed on
gut endothelial cells (A. Dogan, unpublished observation).
Evidence that maturation and TCR recombination are occurring in situ in fetal intestine comes from our analysis of TCR Vß CDR3 regions from multiple randomly cloned Vß PCR products from LPL, IEL, and blood. Because the fetal intestinal lumen is essentially sterile until birth and organized Peyer’s patches do not form until around 19 wk gestation, thus excluding the possibility of recirculation of T cells from Peyer’s patch efferent lymphatics, we hypothesized that if TCR Vß recombination occurs in situ in the fetal gut, clones of TCR should be detectable in gut that are not found in the blood. Extensive sequence analysis shows this to be the case. Out of over 200 clones analyzed, all but three in this study were restricted to one site alone (LPL, IEL, and blood). It is also notable that of the four Vß families that we analyzed in two specimens of fetal intestine, IEL populations were polyclonal. However, this is negative data with no clones being identified between the three sites analyzed; previous studies that showed obvious clonality of IELs in adult intestine found dominant clonality of certain CDR3 regions after analysis of under 30 random clones (11, 12). Because IELs express a restricted TCR repertoire in adult intestinal mucosa, quite different to the situation described here, it is likely that the driving force behind this restriction of the TCR repertoire in adult gut is Ag driven. The polyclonality of TCR usage in fetal intestinal IEL may be due to selection before birth on a wide array of selecting Ags either in the thymus or the intestinal microenvironment. The gut is exposed to dietary and microbial Ags at birth, and it is likely that this causes expansion of T cells bearing TCRs capable of responding to the predominant Ags in the gut lumen in the neonatal period.
We have used pT as a tool to analyze whether TCR recombination
may be on-going in fetal intestinal T cells. pT is a 33-kDa
glycoprotein that associates with the TCR ß-chain at the cell surface
before germline recombination of the TCR -chain (39). It is
exquisitely T-lineage-specific and is only expressed on prothymocytes
and at sites that support extrathymic development (40, 41) and not on
myeloid, NK, or B cells. Our finding that pT is expressed in fetal
intestine suggests that TCR recombination is occurring in situ. This
conclusion would agree with Koningsberger and colleagues who identified
in very early gestation fetuses TCR-ß transcripts in the absence of
rearranged TCR -chains (19). Our observation that following
engraftment of human fetal gut into SCID mice pT mRNA is still
detectable may be explained if the human fetal gut contains an as yet
unidentified T precursor population that is self-renewing in vivo.
pT message is expressed in pro-T cells, which in the thymus have a
lifespan of 3 days (26); thus, the pT we detect here in the
transplants after 30 days is representative of on-going TCR
rearrangement during this period. One of the most remarkable
observations in these experiments was the marked increase in IEL
numbers after transplantation of fetal gut into SCID mice. We have not
excluded the possibility that the increase in IEL numbers in engrafted
fetal intestine is due to an anti-mouse response; however, this is
not a major problem in the human PBL-SCID model, which is
well-established, and problems of graft-vs-host disease in this
system are only seen when large numbers of PBLs in association with NK
cells are transferred (42). We cannot discount the possibility that
some growth factors in the SCID mice recipients may induce this
expansion of human IEL. However, over a 30-day period during fetal
development IEL numbers also increase by two- to threefold during
normal ontogeny (Fig. 1).
It has been shown that the earliest thymic T lymphocyte precursors
(thymic lymphoid progenitors) are dependent on IL-7 and stem cell
factor for survival and maturation (43). The necessity for IL-7 for
thymic lymphoid progenitor survival is well-demonstrated by IL-7 and
IL-7R knockout mice, which have a greatly reduced thymic mass and
virtually no peripheral T cells (44, 45). We have performed
immunocytochemistry for both IL-7 and CD117 (stem cell factor receptor)
in fetal intestine and can detect IL-7 abundantly in the intestinal
epithelial cells and CD117 on occasional cells in the lamina propria
(T.T.M. and D.H., unpublished observations). If TCR
recombination is indeed occurring in the
CD7+3- T cell population in the gut, it should
be possible to detect partially rearranged or incomplete rearrangements
of the or TCR-ß genes in this population as the germline
TCR gene segments stochastically rearrange. Also, in situ hybridization
experiments may be necessary to confirm the position and cell types
expressing pT in the fetal gut. These studies are currently underway
in our laboratory.
Finally, because our analysis of intestinal T cell subsets is
restricted to specimens in which thymus development is complete, we
cannot rule out the possibility that the
CD3-7+ intestinal T cells reported here
represent early thymus emigrants. It is conceivable that early
"leakage" occurs from the thymus of
CD3-7+/- pT- T cell
progenitors that seed the fetal intestine to complete their
developmental program in the gut.
In conclusion, the human intestine shares many features with primary lymphoid organs before birth; therefore, it will be of considerable interest to examine whether the education of progenitor T lymphocytes on gut-specific ligands occurs at this time and what, if any, role development of T lymphocytes in the intestine before birth plays in the repertoire of Ag reactivity at birth.
|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Thomas T. MacDonald, Paediatric Gastroenterology, Suite 3.1, 59 Bartholomew Close, St. Bartholomew’s Hospital, London, U.K. EC1A 7BE; E-mail address:
3 Abbreviations used in this paper: LPL, lamina propria lymphocyte; IEL, intraepithelial lymphocyte; pT, pre-TCR-; TdT, terminal deoxynucleotidyltransferase.
Received for publication April 8, 1998. Accepted for publication July 22, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Proc. Natl. Acad. Sci. USA 88:43. T cell receptor. Eur. J. Immunol. 20:291.[Medline]
/ß+ lymphocytes reveals a major extrathymic pathway of T cell differentiation. J. Exp. Med. 173:483./ß in the intestinal epithelium: evidence for distinct circulation patterns of gut and thymus derived T lymphocytes. J. Exp. Med. 176:187./ heterodimer in human intestinal intraepithelial lymphocytes. Eur. J. Immunol. 19:1335.[Medline]
/ complex with monoclonal antibodies. J. Immunol. 141:335.
/ cells and the diagnosis of coeliac disease. Clin. Exp. Immunol. 85:109.[Medline]
homodimer-bearing T cells in human fetal intestine. Eur. J. Immunol. 24:1703.[Medline]
chain expression by adult human hepatic cells: evidence for extrathymic T cell maturation. Eur. J. Immunol. 26:3114.[Medline]
/ T cell receptor. J. Clin. Pathol. 42:705.-chain gene. Proc. Natl. Acad. Sci. USA 92:12105. gene. Eur. J. Immunol. 25:1877.[Medline]
This article has been cited by other articles:
|
|
 |
|
  X. Wang, A. Das, A. A. Lackner, R. S. Veazey, and B. Pahar Intestinal double-positive CD4+CD8+ T cells of neonatal rhesus macaques are proliferating, activated memory cells and primary targets for SIVMAC251 infection Blood, December 15, 2008; 112(13): 4981 - 4990. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Yang, B. Madison, D. L. Gumucio, and D. H. Teitelbaum Specific overexpression of IL-7 in the intestinal mucosa: the role in intestinal intraepithelial lymphocyte development Am J Physiol Gastrointest Liver Physiol, June 1, 2008; 294(6): G1421 - G1430. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Gunther, J. A. Holloway, J. G. Gordon, A. Knight, V. Chance, N. A. Hanley, D. I. Wilson, R. French, J. Spencer, H. Steer, et al. Phenotypic Characterization of CD3-7+ Cells in Developing Human Intestine and an Analysis of Their Ability to Differentiate into T Cells J. Immunol., May 1, 2005; 174(9): 5414 - 5422. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Nonaka, T. Naito, H. Chen, M. Yamamoto, K. Moro, H. Kiyono, H. Hamada, and H. Ishikawa Intestinal {gamma}{delta} T Cells Develop in Mice Lacking Thymus, All Lymph Nodes, Peyer's Patches, and Isolated Lymphoid Follicles J. Immunol., February 15, 2005; 174(4): 1906 - 1912. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Williams, P. W. Bland, A. C. Phillips, S. Turner, T. Brooklyn, G. Shaya, R. D. Spicer, and C. S. J. Probert Intestinal {alpha}{beta} T Cells Differentiate and Rearrange Antigen Receptor Genes In Situ in the Human Infant J. Immunol., December 15, 2004; 173(12): 7190 - 7199. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. Wheeler, K. L. Black, G. Liu, H. Ying, J. S. Yu, W. Zhang, and P. K. Lee Thymic CD8+ T Cell Production Strongly Influences Tumor Antigen Recognition and Age-Dependent Glioma Mortality J. Immunol., November 1, 2003; 171(9): 4927 - 4933. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Bas, S. G. Hammarstrom, and M.-L. K. C. Hammarstrom Extrathymic TCR Gene Rearrangement in Human Small Intestine: Identification of New Splice Forms of Recombination Activating Gene-1 mRNA with Selective Tissue Expression J. Immunol., October 1, 2003; 171(7): 3359 - 3371. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. da Silva Menezes, D. de Sousa Mucida, D. C. Cara, J. I. Alvarez-Leite, M. Russo, N. M. Vaz, and A. M. C. de Faria Stimulation by food proteins plays a critical role in the maturation of the immune system Int. Immunol., March 1, 2003; 15(3): 447 - 455. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Golby, M. Hackett, L. Boursier, D. Dunn-Walters, S. Thiagamoorthy, and J. Spencer B cell development and proliferation of mature B cells in human fetal intestine J. Leukoc. Biol., August 1, 2002; 72(2): 279 - 284. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R C Fitzgerald, B A Onwuegbusi, M Bajaj-Elliott, I T Saeed, W R Burnham, and M J G Farthing Diversity in the oesophageal phenotypic response to gastro-oesophageal reflux: immunological determinants Gut, April 1, 2002; 50(4): 451 - 459. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F.-k. Kong, C.-l. H. Chen, A. Six, R. D. Hockett, and M. D. Cooper T cell receptor gene deletion circles identify recent thymic emigrants in the peripheral T cell pool PNAS, February 16, 1999; 96(4): 1536 - 1540. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
