Download Steel, O`Donoghue et al accepted - Spiral

1
MATERNAL ORIGIN OF INFLAMMATORY LEUKOCYTES IN PRETERM
FETAL MEMBRANES, SHOWN BY FLUORESCENCE IN SITU
HYBRIDISATION.
1J.H.
Steel, 1K. O’Donoghue, 2N.L. Kennea, 1M.H.F. Sullivan, 2,3A.D. Edwards.
1Department
of Obstetrics & Gynaecology, Institute of Reproductive and
2
3
Developmental Biology, Department of Paediatrics, and MRC Clinical Sciences
Centre, Imperial College London, Hammersmith Hospital, Du Cane Road, London W12
0NN, UK.
Corresponding author:
Dr J.H. Steel
Department of Molecular Endocrinology
Institute of Reproductive and Developmental Biology
Imperial College London
Hammersmith Hospital
Du Cane Road
London W12 0NN
Telephone: +44 0207 594 2180
Fax: +44 0207 594 2184
email: [email protected]
1
2
Abstract.
The aim of this study was to determine the maternal or fetal origin of inflammatory
leukocytes in fetal membranes from cases of chorioamnionitis. Fetal membranes were
collected from male preterm infants and chorioamnionitis was diagnosed histologically.
Fluorescence in situ hybridisation for X and Y chromosomes was used to determine the
gender of infiltrating leukocytes in the chorion and amnion. Leukocytes, trophoblast
and mesenchymal cells were identified using immunohistochemistry for CD45,
cytokeratin-7 and vimentin, respectively. Leukocytes present in the chorion and amnion
were labelled XX, indicating maternal origin, and these cells were immunoreactive for
the leukocyte marker CD45 but not for vimentin or cytokeratin-7. All other cells in the
chorion and amnion were labelled XY and of fetal origin. The results indicated that
maternal leukocytes invade the amnion and chorion in chorioamnionitis and we suggest
that this is part of the maternal inflammatory response to intrauterine infection.
Key words:
Fluorescence in situ hybridisation, chorioamnionitis, leukocytes, maternal, fetal.
3
Introduction.
Chorioamnionitis is a histological diagnosis based upon the infiltration of inflammatory
leukocytes into the fetal membranes. It is associated with a maternal inflammatory
reaction, often with fever and increased C-reactive protein (CRP) levels. There is
evidence that chorioamnionitis is also associated with activation of the fetal immune
system; in particular, elevated pro-inflammatory cytokines (such as IL-6) and increased
memory T cells in fetal blood, findings that have been linked with the incidence of brain
injury in preterm infants (1,2).
In chorioamnionitis, leukocyte infiltration is often considered to be maternal, although
no studies have demonstrated this definitively. However, a fetal origin for leukocytes in
amniotic fluid has been identified (3) and such cells could possibly access the amnion or
chorion of the fetal membranes. The aim of this study was to perform a preliminary
investigation using fluorescence in situ hybridisation (FISH) to detect X and Ychromosomes in leukocytes and surrounding cells, in paraffin sections of preterm fetal
membranes with chorioamnionitis from male infants. This was combined with
immunocytochemical analysis to identify cell types within the membranes.
Materials and Methods.
Patients and tissues.
This study was approved by the local ethics committee (Hammersmith, Queen
Charlotte’s & Chelsea and Acton Hospitals Research Ethics Committee), and informed
consent was obtained from parents of all the infants. Fetal membranes were collected
from consecutive preterm infants, born vaginally or by Caesarean section. The infants
were delivered following prolonged rupture of membranes, spontaneous preterm labour,
or by Caesarean section because of deteriorating fetal or maternal health linked to
preeclampsia (preterm not in labour).
Tissues were fixed in 10% neutral buffered formalin for about 24 hours, rinsed and
stored in 70% ethanol. Membranes were then cut into small pieces (1 cm2) and
processed for paraffin embedding. Sections (5 µm) were mounted on poly-L-lysine
coated clean glass slides and dried at 37°C for a few hours. The presence of
chorioamnionitis was determined in a representative section stained with haematoxylin
4
and eosin and was confirmed in all cases on another sample of the same tissue sent to
the hospital Pathology department. Histological chorioamnionitis was defined as the
presence of infiltrating leukocytes in amnion, chorion or both, according to the
classification by Romero et al (4). Samples from male infants were selected for further
analysis if they contained more than five leukocytes per high power field and had good
tissue preservation, so that the position of the leukocytes could be determined
microscopically following FISH. The five suitable cases used for the study are listed in
Table 1, and were chosen to represent different gestational ages and clinical
characteristics. All fetal membranes used in the present study had been shown to
contain bacteria using a fluorescence in situ hybridisation procedure in a previous study
(5), therefore they were regarded as cases of intrauterine infection.
Fluorescence in situ hybridisation (FISH).
FISH was performed as previously described (6.7) and modifications for use on this
tissue were adapted from previous studies (8,9). Ten sections from each sample were
supplied to the person performing the FISH procedure, who was unaware of the clinical
history of the patients, and sections for FISH analysis were chosen at random.
Paraffin sections (5 m) were dewaxed by immersion in Histoclear™ and rehydrated in
an ethanol series. Tissue was heat-treated with 2x SSC (sodium saline citrate) for 15
minutes at 80°C, washed briefly in distilled water and twice in phosphate buffered saline
(PBS). Sections were then treated with 100l of 100g/ml pre-warmed Proteinase-K
(stock solution 2mg/ml, VWR Merck Eurolab Ltd., UK) for 15 minutes at 37°C, and
washed as before. Secondary fixation was with 2:1 v/v ice-cold methanol:acetone for 2
minutes. Chromosome-specific centromeric repeat probes DXZ1, labeled with
SpectrumOrange™, and DYZ1, labeled with SpectrumGreen™ (Vysis, Abbott
Laboratories Ltd., Maidenhead, U.K.) were used as standard XY probes to detect X and
Y chromosomes in the same section. The centromeric repeat probe DYZ1, labelled this
time with SpectrumOrange™, and a locus specific probe, (LSI SRY), also labelled with
SpectrumOrange™, were also applied separately to randomly-selected slides to confirm
Y chromosome localisation, as used previously (10). The hybridisation buffer contained
50% formamide and 10% dextran sulphate in 2x SSC at pH 7.0. Five microlitres of
5
probe, diluted 1:2 in hybridisation buffer, was added to each slide. Target DNA was
denatured directly on the in situ hybridisation block (Genetic Research Instrumentation
Ltd) at 71°C for 7 minutes followed by a 4-hour hybridisation at 37°C. Slides containing
male peripheral blood lymphocytes or sections of paraffin-embedded trophoblast were
used as hybridisation reaction controls. Following hybridisation, cells were washed
with 2x SSC for 2 minutes, incubated with 0.4x SSC in a waterbath at 72°C for 2
minutes and then incubated with 2x SSC/0.1% nonylphenoxy-polyethoxy-ethanol-40
(NP-40) at room temperature for 2 minutes. Slides were dehydrated through an
ascending ethanol series, air-dried and mounted in a fluorescence antifade medium
containing DAPI (diamidino-2-phenyl-indole; Vector Laboratories, Burlingame, CA).
Slides were analysed by epifluorescence microscopy using single band pass filters for
Aqua, Orange and DAPI and the triple band pass filter set (Zeiss Axioskope, Germany).
Images were captured using a cooled charge-coupled device camera, reviewed in Quipps
m-FISH software (Vysis). Nuclei with two red signals were categorised as XX (of
female origin) and those with one red signal and one green signal were categorised as
XY (of male origin). Each slide was checked for hybridisation efficiency and analysed
only if more than 75% of nuclei contained both signals. These were counted when the
intensity and size of the fluorescent signals was approximately equal and inside a
distinct nucleus with an intact border, as indicated by DAPI staining (7,9). The slides
were analysed by two separate observers (K.O’D, N.K), who were unaware of the
clinical history of the samples at the time of viewing.
Combined immunocytochemistry and FISH.
In order to identify the cell types within the tissue sections, immunocytochemistry for
markers of trophoblast and epithelial cells (Cytokeratin-7) (11), mesenchymal cells such
as fibroblasts (vimentin), and leukocytes (CD45) was performed on the same sections
following FISH. Tissue was heat-treated with 2x SSC for 5 minutes at 80°C, treated
with 100l of 20g/ml pre-warmed Proteinase-K for 5 minutes at 37°C, and washed
after each step. After fixation with ice-cold 2:1 v/v methanol: acetone for 2 minutes,
slides were hybridised with an XY probe cocktail diluted 1:1 in buffer for 4 hours.
Following post-hybridisation washes and an ethanol dehydration series, slides were air-
6
dried and incubated with 5% normal goat serum/5% bovine serum albumin in PBS for 1
hour. Slides were incubated with the appropriate anti-human murine IgG-class
monoclonal antibodies diluted 1:25 (Vimentin and CD45, DakoCytomation, Cambs.,
U.K.) or 1: 5 (Cytokeratin-7, DAKO labs.) for 60 minutes and washed twice.
Subsequent incubations were with biotinylated goat anti-mouse secondary antibody
(Vector Laboratories), detected with streptavidin conjugated with fluorescein (Vector
Laboratories), both diluted 1:50 and incubated for 30 minutes each. Slides were
dehydrated in 100% ethanol for 1 minute, air-dried, and mounted in DAPI, for analysis
by epifluorescence microscopy.
The quality of combined immunocytochemistry and XY FISH using the CD45 antibody
was not optimal because the pretreatment for the FISH procedure removed the antigen
for CD45. Adjacent sections were therefore used for each technique and matching areas
chosen for comparison
Results
FISH
In sections of fetal membranes from all the cases of chorioamnionitis, groups of cells
with the morphology of leukocytes were present in the chorion or amnion, or both. An
example is shown in Figure 1 (a,b). In the chorion and amnion from male deliveries,
the nucleated cells had both X and Y chromosomes (Figure 1; c,d). In the decidua, the
majority of cells were labelled XX indicating that these cells were maternal. In all
sections examined, the groups of leukocytes were invariably XX, with two red signals
per nucleus, indicating the presence of X chromosomes only (Figure 1; c,d,g,h), while
these cells were surrounded by cells labelled for X and Y. The distribution of XY and
XX cells was confirmed in all cases using both standard XY FISH and single FISH
using the alternative DNA probes for the Y chromosome (DYZ1 or SRY), labelled with
SpectrumOrangeTM (Vysis, UK). (Figure 1; e,f).
FISH results were obtained from a minimum of two sections from each case, in which
the results were concordant. In all the chorioamnionitis cases examined (n=5), XX cells
7
were visualised within the amnion and chorion, surrounded by XY labelled cells,
irrespective of delivery route or reason for preterm birth.
Combined immunocytochemistry and FISH.
Cells within the amnion and chorion with large nuclei labelled XY were
immunoreactive for cytokeratin 7 (Figure 2; a,b); indicating epithelial or trophoblast
cells, but the groups of cells with smaller nuclei identified as XX were negative for this
antibody (Figure 2; b). Fibroblasts within the amniotic connective tissue were
positively stained for vimentin (Figure2; c,e), but this antibody also did not stain the
groups of XX cells (Figure 2; d,f). In contrast, CD45 immunoreactivity, in adjacent
sections compared with the FISH results, was clearly seen in these groups of XX cells
(Figure 2; g,h), indicating that they were leukocytes.
Discussion.
This preliminary study has shown a maternal origin for infiltrating cells with the
morphology of leukocytes within the chorion and amnion of fetal membranes from male
preterm infants with chorioamnionitis, with a variety of clinical characteristics and
reasons for preterm delivery. In addition, the XX-labelled cells were confirmed as
leukocytes, since they were not immunoreactive for cytokeratin 7 or vimentin and were
CD45 immunoreactive. In the past, the maternal origin of the inflammatory cells which
invade the chorion and amnion has been inferred but has not been shown definitively the origin of the cells suggests that chorioamnionitis involves a maternal inflammatory
response. Such inflammatory reactions are regarded as an important cause of preterm
labour, so investigating the contribution of maternal and fetal immune responses to this
process is essential to improve strategies for prevention as well as therapeutic
approaches.
8
Both materno-fetal and feto-maternal transfer of cells occurs during pregnancy. In
villitis of unknown etiology, maternal cells have been shown to cross into the placenta
using XY FISH (12); and maternal cells have been identified in chorionic villi using
HLA markers specific to mother or fetus (13). It is well established that fetal leukocytes
can be retrieved from the maternal circulation many years following a pregnancy (14).
In chorioamnionitis, maternal leukocytes probably enter the chorion and amnion from
maternal blood vessels in the decidua, since these are the closest vessels to the maternofetal interface in the fetal membranes. Chorioamnionitis is often used as a surrogate
marker for intrauterine infection (15), although the source and mechanism of infection
that provokes the maternal inflammatory response is complex and warrants further
investigation.
Microbiological analysis sometimes fails to demonstrate live bacteria in the tissues
even when inflammation is present, leading to confusion over causes of
chorioamnionitis. We have shown previously, using fluorescence in situ hybridisation
(5), that bacteria can be demonstrated in tissue sections from fetal membranes in cases
of preterm and term delivery, in most cases of chorioamnionitis and even in
histologically normal tissues collected at term from Caesarean section deliveries. In
most preterm deliveries there are bacteria present, whether or not there is any
histological chorioamnionitis (5). This suggests that the inflammatory response to
infection, rather than the actual presence of bacteria in the fetal membranes, may be the
crucial factor determining whether a pregnancy is vulnerable to preterm labour. Even
so, identification of specific organisms present in utero, possibly by means of sensitive
methods such as PCR, will allow therapy to be targeted more precisely.
FISH for identifying chromosomes has been used successfully upon cells smears and
frozen tissue sections, but paraffin sections have proved to be demanding (16), giving
results of variable quality. We have optimised the FISH method for use in paraffin
sections, thus taking advantage of the good morphological preservation necessary to
identify and localise particular cells in the tissue. Providing that the X and Y
chromosomes are both visible in a cell, the gender of that cell can be identified, and
sectioning of cell nuclei does not appear to prevent such identification.
9
Leukocytes in fetal membranes in patients with preterm delivery have been shown to be
activated, as shown by morphological and enzyme histochemical markers (17), and
leukocyte activation may have a key role in the pathogenesis of chorioamnionitis. In
preterm labour, it is thought that cytokines produced by inflammatory cells may
precipitate the cascade of events leading to labour itself.
Although we have shown that the leukocytes are of maternal origin, there is
considerable evidence that the fetal immune system is also activated in cases of preterm
labour (18) and the fetus is also likely to be involved in the inflammatory process of
chorioamnionitis (2). Fetal leukocytes are functionally active and are capable of
responding to infection, producing cytokines such as IL-8 (19), and neonatal leukocytes
produce significantly more interleukin-1 than do adult leukocytes (20). Activation of
the fetal inflammatory response is associated with other consequences harmful to the
fetus - fetal T-cell activation and cytokine levels are linked with evidence of brain injury
(1,21) and so the treatment and prevention of this condition will require further
investigation of maternal-fetal interactions. Analysis of the status of maternal immune
activation could aid prediction of patients at risk from preterm labour.
Acknowledgements
The authors would like to thank Dr. P. Duggan, Dr. P. Reynolds and Dr. E. Adams for
their help with collecting tissues for this study.
This work was supported by Action Research and Wellbeing.
10
References:
[1] Duggan PJ, Maalouf EF, Watts TL, Sullivan MH, Counsell SJ, Allsop J, AlNakib L, Rutherford MA, Battin M, Roberts I & Edwards AD Intrauterine T-cell
activation and increased proinflammatory cytokine concentrations in preterm infants
with cerebral lesions. Lancet 2001; 358:1699-1700.
[2] Chaiworapongsa T, Romero R, Kim JC, Kim YM, Blackwell SC & Yoon BH,
Gomez R. Evidence for fetal involvement in the pathologic process of clinical
chorioamnionitis. Am J Obstet Gynecol 2002; 186:1178-1182.
[3] Sampson JE, Theve RP, Blatman RN, Shipp TD, Bianchi DW, Ward BE &
Jack RM Fetal origin of amniotic fluid polymorphonuclear leukocytes. Am J Obstet
Gynecol 1997; 176:77-81.
[4] Romero R, Salafia CM, Athanassiadis AP, Hanaoka S, Mazor M, Sepulveda W
& Bracken MB The relationship between acute inflammatory lesions of the preterm
placenta and amniotic fluid microbiology. Am J Obstet Gynaecol 166:1392-1388, 1992.
[5] Steel JH, Malatos S, Edwards AD, Miles L, Duggan P, Kennea N, Reynolds PR,
Feldman RG & Sullivan MHF Bacteria and inflammatory cells in the fetal membranes
do not always cause preterm labour. Pediatric Res 2004; (in press)
[6] Choolani M, O'Donnell H, Campagnoli C, Kumar S, Roberts I, Bennett PR &
Fisk NM Simultaneous fetal cell identification and diagnosis by epsilon-globin chain
immunophenotyping and chromosomal fluorescence in situ hybridisation. Blood 2001;
98:554-557.
[7] O'Donoghue K, Choolani M, Chan J, de la Fuente J, Kumar S, Campagnoli C,
Bennett PR, Roberts IA & Fisk NM Identification of fetal mesenchymal stem cells in
maternal blood: implications for non-invasive prenatal diagnosis. Mol Hum Reprod
2003; 9:497-502.
11
[8] Dundas SR, Boyle S, Bellamy CO, Hawkins W, Garden OJ, Ross JA &
Bickmore W Dual Y-chromosome painting and immunofluorescence staining of
archival human liver transplant biopsies. J Histochem Cytochem 2001; 49:1321-1322.
[9] Khosrotehrani K, Stroh H, Bianchi DW & Johnson K Combined FISH and
immunolabeling on paraffin-embedded tissue sections for the study of microchimerism.
Biotechniques 2003; 34:242-244.
[10] O'Donoghue K, Chan J, de la Fuente J, Kennea N, Sandison A, Anderson JR,
Roberts IA & Fisk NM Microchimerism in female bone marrow and bone decades
after fetal mesenchymal stem-cell trafficking in pregnancy. Lancet 2004; 364:179-182.
[11] Blaschitz A, Weiss U, Dohr G & Desoye G Antibody reaction patterns in first
trimester placenta: implications for trophoblast and purity screening. Placenta 2000;
21:733-741.
[12] Redline RW & Patterson P Villitis of unknown etiology is associated with major
infiltration of fetal tissue by maternal inflammatory cells. Am J Pathol 1993; 143: 473479.
[13] Labarrere CA & Faulk WP Maternal cells in chorionic villi from placentae of
normal and abnormal human pregnancies. Am J Reprod Immunol 1995; 33:54-59.
[14] Bianchi DW, Zickwolf GK, Weil GJ, Sylvester S & DeMaria MA Male fetal
progenitor cells persist in maternal blood for as long as 27 years postpartum. Proc Natl
Acad Sci USA, 1996; 93:705-708.
[15] Guzick DS & Winn K The association of chorioamnionitis with preterm delivery.
Obstet Gynecol, 1985; 65:11-15.
[16] Haralambieva E, Banham AH, Bastard C, Delsol G, Gaulard P, Orr G, Pileri
S, Fletcher JA & Mason DY Detection by the fluorescence in situ hybridization
technique of MYC translocations in paraffin-embedded lymphoma biopsy samples. Br J
Haematol 2003; 121:49-56.
12
[17] Matsubara S, Yamada T, Minakami H, Watanabe T, Takizawa T & Sato I
Polymorphonuclear leukocytes in the fetal membranes are activated in patients
with preterm delivery: ultrastructural and enzyme-histochemical evidence. Placenta
1999; 20:185-188.
[18] Berry SM, Romero R, Gomez R, Puder KS, Ghezzi F, Cotton DB & Bianchi
DW Premature parturition is characterized by in utero activation of the fetal immune
system. Am J Obstet Gynecol 1995;173:1315-20.
[19] Taniguchi T, Matsuzaki N, Shimoya K, Neki R, Okada T, Kitajima H, Saji F
& Tanizawa O Fetal mononuclear cells show a comparable capacity with maternal
mononuclear cells to produce IL-8 in response to lipopolysaccharide in
chorioamnionitis. J Reprod Immunol1993; 23:1-12.
[20] Contrino J, Krause PJ, Slover N & Kreutzer D Elevated interleukin-1
expression in human neonatal neutrophils. Pediatr Res 1993; 34:249-252.
[21] Shalak LF, Laptook AR, Jafri HS, Ramilo O & Perlman JM Clinical
chorioamnionitis, elevated cytokines, and brain injury in term infants.
Pediatrics 2002;110:673-680.
13
Figure legends
Figure 1:
A representative example of inflamed fetal membranes stained with haematoxylin and
eosin, showing the amnion (A), chorion (C) and decidua (D), with infiltration of
inflammatory cells at the margin of the chorion and amnion (arrows) (a & b).
In inflamed fetal membranes from patient 3, the majority of amniotic and chorionic cells
have X and Y chromosomes shown by FISH (c); higher magnification shows groups of
inflammatory cells, bearing two X chromosomes labelled with SpectrumOrange™,
within the chorion (arrow) (d). The distribution of Y chromosome signals is confirmed
using a different probe for the Y chromosome labelled with SpectrumOrange™ (arrows)
(e); higher magnification shows the absence of male signals in the inflammatory cell
layer within the chorion (small arrows), while male cells with larger nuclei are still
labelled (large arrowheads) (f). XX cells can be seen in areas of inflammatory
infiltration in fetal membranes from patient 5 (g), and patient 1(h); cells bearing a Y
chromosome are labelled with SpectrumGreen (large arrowheads) and in XX cells
both X chromosomes are labelled with SpectrumOrange (small arrows).
Scale bar: (a) 100 µm, (b,c,d,e,h) 50 µm, (f,g) 25 µm.
Figure 2: (a) Epithelial cells (inset) within the amnion and trophoblast cells within the
chorion, all labelled XY, were immunoreactive for cytokeratin-7 (FITC), but the clusters
of XX inflammatory cells were negative for this antibody (arrows)(b). Fibroblasts
within the amniotic connective tissue were vimentin-FITC positive (c) but the
inflammatory cells labelled XX on FISH (both chromosomes labelled with
SpectrumOrange™) were negative for this antibody (d). Vimentin-FITC negative
inflammatory cells (arrows) (e) are shown to be labelled XX on FISH in adjacent
sections (arrows) (f) and similarly CD45-FITC immunoreactivity is shown in
inflammatory cell clusters (arrows) (g) which are shown to be labelled XX in adjacent
sections (arrows) (h).
Scale bar: (a) 100 µm, (b-h) 25 µm.
14
Table 1: Patients.
Patient number
Gestational age
Delivery route
(weeks)
Reason for
Gender
delivery
1
23.5
V
PROM
Male
2
26
CS
PROM
Male
3
29.4
CS
PROM
Male
4
31
CS
PNIL
Male
5
30.3
V
PIL
Male
Key:
CS = Caesarean section, PIL = preterm in labour, PNIL = preterm not in labour (preeclampsia), PROM = preterm labour with prolonged rupture of membranes, V =
vaginal delivery,