"18 T cells in human decidua

Ruth Ditzian-Kadanoff, MD, PhD," Jack Garon, MD,b Marion S. Verp, MD,c and Moshe ZHberstein, MDC

Chicago, Illinois

OBJECTIVE: Our purpose was to assess the presence and frequency of 'VB T cells in the decidua of term and first-trimester pregnancies. STUDY DESIGN: Term and first-trimester placentas were obtained from normal subjects. Frozen sections and cell suspensions were prepared from decidual tissue and stained with monoclonal antibodies to T cell markers. Cell sorter analysis was performed. RESULTS: 'VB T cells in term decidual cell preparations were enriched 2.4-fold compared with peripheral blood. Immunohistochemical staining of term decidual tissue demonstrated many 'VB' and C03' T cells, fewer C08 + cells, and rare C04 + cells. In contrast, first-trimester decidua was found to be devoid of 'VB +

T cells, by both cell sorter analysis and immunohistochemical methods. CONCLUSIONS: Term, but not early, decidua harbors a resident T-cell population that is significantly enriched in 'VB T cells compared with blood. These lymphocytes may provide an added defense mechanism against infection during the peripartum. (AM J OBSTET GVNECOL 1993;168:831-6.)

Key words: Reproductive immunology, )'3 T cells, decidua

A newly identified subset of T cells, )'3 T cells, has been the recent subject of intensive study. 1, 2 These cells possess a distinct T-cell receptor. Like the previously characterized T-cell receptor of the afj T cells, the )'3 T cell receptor is closely associated with the CD3 surface T-cell marker. It is a heterodimer composed of two single chain subunits ()' and 3), each encoded by a separate gene. The gene loci always contain constant and variable regions of gene families. Diversity in cell receptor antigen specificity is obtained with gene rear­rangement. In )'3 and al3 T cells, a gene of each family is picked for the final rearranged gene that is ex­pressed. The mature )'3 T cells are mostly CD4 and CDS negative,3 but the presence of CDS + )'3 T cells has been described. 1. 4 In humans )'3 T cells constitute 6.7% ± 1.5% of the peripheral blood monocyte popu­lation. I

The specific functions of )'3 T cells have not yet been elucidated. They often exhibit cytotoxicity that is with­out major histocompatibility complex restriction but can recognize major histocompatibility complex class p. 6 Their presence in the epithelium lining hollow

From the Departments of Medicinea and Pathology, b Mount Sinai Hospital, and the Department of Obstetrics and Gynecology, Univer­sity of Chicago.' Supported by the Mount Sinai Research Foundation, Mount Sinai Hospital, Chicago, Illinois. Received for publication May 4, 1992; revised August 10, 1992; accepted September 22, 1992. Reprint requests: Ruth Ditzian-KadanofJ, MD, PhD, Department of Medicine, Mount Sinai Hospital, California and 15th St., Chicago, IL 60608. Copyright © 1993 by Mosby-Year Book, Inc. 0002-9378/93 $1.00 + .20 6/1/42800

organs such as uterus, vagina, intestine, and urinary bladder has been well documented. 7 The tissue distri­bution of )'3 T cells differs among species. The mouse has a large resident )'3 T population in the gut and dermis, 7 each with a dominant variable region type, 8

suggesting that their role is different from random antigen recognition. The human gut and dermis are normally devoid of)'3 T cells, but they appear in the gut in celiac disease. 9

, 10 Their numbers are increased in various infections.3

, 11 The )'3 T cells proliferate vigor­ously both in vitro and in vivo in response to either mycobacteria or to the heat shock protein HSP65, which is homologous to a component of mycobacterial cell walls. 12 Taken together, the current data suggest an immune surveillance function for )'3 T cells in hollow organs like uterus and intestine.7 These cells may func­tion by removing altered, specifically infected, or trans­formed cells. I2

The human decidua has been shown to contain small cells that look like lymphocytes.'" 14 A transforming growth factor-13 (TGF-fj)-like molecule produced in the decidua of murine allopregnancy by small cells has been shown to be the main suppressing factor of cyto­toxic activity in vitro. 15, 16 When stimulated, )'3 T cells, like afj T cells, produce large quantities of TGF -fj in addition to other cytokines. 2 This TGF -13 is usually secreted in a latent inactive form that needs to be activated by cleavage from the carrier proteins. The activation can be achieved in many ways, including local lowering of pH. I7

We looked for )'3 T cells by immunohistochemical staining in situ both of decidua from normal first­trimester elective pregnancy terminations and decidua

831

832 Ditzian-Kadanoff et al.

adherent to placentas of normal term deliveries. We also measured the percentage of these cells by fluores­cence-activated cell sorter analysis of single-cell suspen­sions. Our data show an increase in the percentage of ",8 T cells in term decidual preparation. In contrast, in the first-trimester decidual preparation ",8 T cells were undetectable by our methods.

Material and methods Placentas. Placentas were collected after normal de­

livery (Mount Sinai Institution Review Board-approved as normally discarded tissue). A thin layer of tissue (l mm thick) from the maternal surface of the pla­centa, which is rich in decidual patches, was excised and processed. Paraffin-embedded, hematoxylin-eosin stained sections of decidua and contiguous trophoblast tissue were examined to assess for lymphocyte pres­ence. Several pieces, about 1 cm each in diameter, were coated in ornithine carbamyltransferase oil (Miles, Elkhart, Ind.) and flash frozen for 5 minutes in liquid nitrogen for frozen section immunohistochemical study. The remaining portion was reduced to single-cell suspension by compression through a thin wire mesh in Ca + + - and Mg+ + -free Hank's (Gibco, Grand Island, N.Y.) medium. First-trimester placentas (University of Chicago, in accordance with National Institutes of Health guidelines for human tissues) were examined under the dissecting microscope, and decidual-rich patches were excised and treated as described above.

Decidual lymphocytes. The large heavy trophoblast cells from the decidual preparation were removed by discarding the pellet after a lO-second slow centrifuga­tion. The cells in the supernatant were layered on lymphocyte separation medium and centrifuged at 1600 revolutions/min for 30 minutes at 20° C. The interface layer of mononuclear cells was aspirated and kept overnight in a 37° C, 5% carbon dioxide incubator in Rose Park Memorial Institute (Gibco) medium with added 5% fetal calf serum (Gibco), I % penicillin and streptomycin (Gibco), 1% L-glutamine (Gibco), 1 % 2-mercaptoethanol (Sigma, St. Louis), and interleu­kin-2 in the form of 5% supernatant from MLA 144 cells (ATCC, Rockville, Md.), thus allowing adherence and further elimination of macrophages, fibroblasts, and trophoblast cells.

Blood lymphocytes. Peripheral blood was collected in heparinized tubes from the same individuals as the placentas. The blood was diluted I: 3 in Hanks' me­dium devoid of Ca + + and Mg+ +. Mononuclear cells were obtained by layering over Ficoll-Hypaque (Phar­macia, Piscataway, N.J.) lymphocyte separation medium and centrifuging at 1600 revolutions/min for 30 min­utes at 20° C. The interface layer enriched in mononu­clear cells was aspirated and cultured overnight ill

similar conditions to the decidual-derived cells.

March 1993 Am J Obstet Gynecol

Flow cytometer analysis. Peripheral blood mono­cytes and decidual cell preparations from early and term pregnancies were stained with monoclonal anti­bodies to various T cell markers, to assess the frequency of the different subsets, particularly the newly described ",8 T cells (fluorescein-conjugated Leu-2,3,4, Becton Dickinson, Bedford, Mass.) and T-cell receptor 8 from T Cell Science (Cambridge, Mass.). The stained cells were counted in a fluorescence-activated cell sorter. Gates were set to encompass the easily recognized peripheral blood lymphocytes and were kept constant for the decidual cells. Becton Dickinson G 1C 1 antibody was used for negative control. Percentages were calcu­lated by subtracting the nonspecific staining with the negative control from the raw percentage. The single­color stain had a nonspecific stain rate of < 1.5%. Two-color stain was performed by means of indirect staining with plain anti-CD3 followed by phycoerythrin antimouse second-step and fluorescein-co~ugated

anti-8 (T Cell Science). This method allows considerable nonspecific staining, and therefore the single staining was used for quantitative analysis. Direct double stain­ing was used with phycoerythrin-co~ugated anti-CD4 and anti-CDS (Becton Dickinson) and fluorescein iso­thiocyanate anti-T cell receptor 8 (T Cell Science).

Immunohistochemistry. Frozen sections 6 f.Lm thick were taken, air-dried for 2 minutes, then acetone fixed for 5 minutes and stored at 20° C overnight.

All immunochemical reagents were from Sigma, ex­cept when otherwise stated.

The sections were rehydrated in phosphate-buffered saline solution, blocked in phosphate-buffered saline solution with 0.2% bovine serum albumin, and stained with the designated antibodies for 30 minutes. The sections were next covered with biotinylated Fab2' goat antimouse antibody diluted in 20% human serum to optimal concentration for 30 minutes. Incubation with streptavidin-conjugated peroxidase for 30 minutes fol­lowed, and finally the peroxidase substrate, 3,3 diami­nobenzidine tetrachloride in 0.05 mol/L Tris and 0.01% hydrogen peroxide, was applied for 5 to 10 minutes until the section turned color. Counterstaining with water-soluble hematoxylin and fixation with water-sol­uble mounting medium (glycerol jelly) completed the process. T-cell receptor aJ3, T-cell receptor 81, Leu-2, Leu-3 (Becton Dickinson), and T-cell receptor 8 (T Cell Science) were used. Tonsils and thymus sections were used for positive controls. The G 1C 1 antibody was used for a negative control on an adjacent section as the first step in the sequence in every experiment.

Results

Fig. I depicts an example of a typical lymphocyte cluster in the decidua and shows in contrast that the contiguous trophoblast tissue is devoid of lymphocytes.

Volume 168, Number 3, Part I Am J Obstet Gynecol

Ditzian-Kadanoff et al. 833

Fig. 1. Hematoxylin-eosin-stained, paraffin-embedded decidual tissue from human term placenta. Lymphocyte cluster (scanty cytoplasm, small nuclei) is shown in center field, easily distinguishable from decidual cells (ample cytoplasm, large nuclei) mostly in upper right corner. (Original magnifi­cation x 40.)

Fig. 2. Immunohistochemical detection of "(8 cells in decidual tissue. Section stained with Identi-T T-cell receptor 81 (T Cell Science) as in Material and methods.

Fig. 2 shows positively staining cells for the 'YO marker at high magnification. The negative and positive con­trols are not shown. Fig. 3 shows sections stained for all the markers at low magnification to allow comparison of

their frequency. The data in Figs. 2 and 3 demonstrate that 'YO + and CD3 + T cells occur frequently in the decidua. CDS + T cells are less prevalent than CD3 +

and 'YO + cells, and CD4 + cells are rare.

834 Ditzian-Kadanoff et al. March 1993 Am J Obstet Gynecol

Fig. 3. Immunohistochemical staining of sections of decidual tissue. Monoclonal antibodies, Leu-4, Leu-2, Leu-3, and T-cell receptor al3 (Becton Dickinson), as detailed in Material and methods, were used to detect cells positive for markers: CD3 (A), CD4 (B), CD8 (C), T-cell receptor al3 (D), and "),1> (E).

In agreement with the work described by Dietl et aI., I" we did not detect CD3 + or 'YO + T cells in first­trimester decidua (data not shown).

Measurement of the relative number of 'YO T cells by fluorescence-activated cell sorter analysis of enriched decidual cell preparations from early and term preg­nancies were performed. Results are shown in Table I. Values are the percentage of cells within the range of size and fluorescence of the peripheral blood lympho­cytes that are positively stained. This gated population

contains also the smaller placenta cells, monocytes, B cells, and natural killer cells. Therefore the meaningful number is the percentage of T cells (by definition CD3 positive cells) that are 'YO positive: (percentage of 'YO +

staining cells)/(percentage of CD3 + staining cells). The percentage of peripheral blood monocyte cells that are 'Yo T cells is known to vary widely among individuals,2 in our laboratory between 0.2% and 9%. We therefore compared the percentage of 'YO T cells in each decidual preparation to the percentage of 'YO T cells in periph-

Volume 168. Number 3. Part I Am J Obstet Gynecol

Ditzian-Kadanoff et al. 835

Table I. Enrichment of resident -y& T cells in human term decidua

Case No.

1 2 3 4

Column 1* (CDJ) (%)

58 43 25.1 30.3

I

Decidua

Column 1I* (8) (%)

6 2.7 2.3 1.8

I Column lIlt (8ICD3) (%)

10.3 6.3 9.2 5.9

Column IV* (CD3) (%)

78 74 52.9 65.5

I

Peripheral blood

Column V* (8) (%)

2.5 2 2.5 1.7

I Column Vlt (8ICD3) (%)

3.0 2.7 4.7 2.6

Ratio

Column VII::

3.4 2.3 2.0 2.2

*Fluorescence-activated cell sorter analysis results for CD3 and /) percentages in the decidual and blood preparation, as in Material and Methods.

tCalculated percentage of CD3' cells that are /)' in decidua and blood, respectively. tTo compare decidua and blood we used (column VII) the ratio of column III/column VI. The statistical analysis of (/)/CD3

decidua)/(/)/CD3 blood) was performed by calculating for average, deviation, and one-sided tail Student t test on the logarithm of the ratios. The result is 2.42 (+ 0.64, -0.50), which differs from 1 with p < 0.03.

eral blood monocytes of the same individual. In Table I we demonstrate a 2.4-fold enrichment of -y& + T cells in the term decidua compared with the peripheral blood (jJ < 0.03).

As expected, the combined percentages of aj3 + and -y& + cells equal the CD3 + percentage, in both the decidua and the blood (data not shown). The CD4 C and CD8 + cells in the decidual cell preparation were mostly aj3 +, because -y& + cells were a small number. The percentages of CD4 + and CD8 + cells were similar to those in the peripheral blood (data not shown). This suggests either that most of these lymphocytes were blood-derived rather than tissue-resident or that there was no preferential enrichment of these subsets of the aj3 + T cells in the term decidua.

Two-color flow cytometer analysis for &/CD4 and &/CD8 showed that the decidual -y& T cells were all CD4 negative; 50% were CD8 + and 50% CDS - (results not shown).

An interesting exception was a mother who had a very high percentage of -y& T cells in the decidua (&/CD3 18.8%) and an equally high one in the circula­tion (&/CD3 19.8%). This individual developed leukocy­tosis 1 day post partum (white blood cell count doubled to 14), a presumptive indication of infection, and there­fore had to be excluded from this study of normal, healthy deliveries. The enrichment remains statistically significant even when this subject is included. However, this individual's data suggest a possible role for these cells in infection control (see Comment).

Fluorescence-activated cell sorter analysis of first­trimester decidual cell preparations revealed a very small total number of CD3 + cells and no significant -y& population (data not shown).

Comment

We have shown a 2.4-fold increase in -y& T cells in human term decidua compared with peripheral blood. We therefore suggest that the resident -y& T cells in term decidua may play a role in decidual function at

that time. This 2.4-fold enrichment is probably only a lower limit estimate, with the true increase likely to be even higher. Because the placenta is engorged with blood, the decidual cell suspension also contains blood­derived mononuclear cells, which dilute the resident cells and bring any ratio of decidua to blood closer to 1. We also found 50% of these -y& T cells to be positive for the CD8 marker, which is usually associated with cyto­toxicity.

A possible role for -y& T cells could be the protection of the maternal uterus from aggressive trophoblast invasion. This could be achieved by their cytotoxic capacity and by killing invading trophoblastic cells. Another way of limiting the aggressiveness of the tro­phoblast could be achieved through the well known differentiation-promoting capacity of TGF-j3.19 It has been reported that the -y& T cells, when stimulated, may produce TGF-j3.2 Exogenous TGF-j3 causes induction of its own expression at both transcriptional and posttran­scriptional levels. 17. 20 Therefore a small amount of TGF-j3 secreted by -y& T cells could initiate and amplify its production locally by other cells in a paracrine fashion. Although the secreted TGF-j3 is usually in the latent, inactive form, it can be activated by many mech­anisms, including acidification. Aggressive cells that exhibit increased metabolic activity are likely to provide an acid local environment.

Another possible function of the decidua is protec­tion of the fetal allograft from maternal rejection by cytotoxic cell attack. However, this role is unlikely to be undertaken by resident -y& T cells in view of their absence from first-trimester decidua. This role is more likely to be undertaken by other cells early in preg­nancy. The presence of many lymphocyte-like, CD3-cells with no rearrangement of -y and & genes has been documented in early pregnancy.11 Later in pregnancy -y& T cells, by means of their cytotoxicity, may be instrumental in effecting a reduction in the number of fetal white blood cells passing through to the maternal circulation. Fetal white blood cells display major histo-

836 Ditzian-Kadanoff et al.

compatibility complex class I and II of paternal origin, which could stimulate maternal immune response of immunoglobulin G antibodies that can cross the pla­centa. Mter repeated exposure of multiple pregnancies, the anamnestic response is much stronger and could harm the fetal tissues.

The documented response of 'Y8 T cells to heat shock protein provides the most plausible mechanism to ex­plain their presence in the decidua specifically at term. Heat shock proteins are very likely to be produced in the decidua during the cellular trauma caused by labor. Peripheral blood 'YB T cells may respond to heat shock proteins, migrate into the decidual tissue to become resident, and proliferate. We found that about 50% 'YB T cells in the term decidua are CD8 +. Previous reports indicated that in peripheral blood only a small percent­age of 'YB T cells are CD8 + ,3 but that human spleen has 50% CD8 + 'YB T cells. I This may be significant, because CD8 + 'YB T cells may have a more specific cytotoxic function than do CD8 - cells. One patient in our study had a very high percentage of 'YB T cells, almost 20% of all T cells, both in decidua and blood. Her white blood cell count doubled the day after delivery, a clue to possible infection. Most likely, 'YB T cells contribute to fetal and maternal protection from infection during labor, a time when risk of infection is high.

We thank Drs. J. Quintans and J. Bluestone of the University of Chicago for helpful advice and for pro­viding access to a flow cytometer; Drs. N. Gleicher and E. Confino and the Mount Sinai obstetrics-gynecology residents for providing the placentas; and Dr. Gilman­Sachs of Chicago Medical School for kindly providing the phycoerythrin-conjugated antibodies.

REFERENCES 1. Bucy RP, Chen CL, Coope MD. Tissue localization and

CD8 accessory molecule expression ofT gamma delta cells in humans. J Immunol 1989;142:3045-9.

2. Bluestone JA, Cron RQ, Barrett TA, et al. Repertoire development and ligand specificity of murine TCR gamma delta cells. Immunol Rev 1991;120:5-33.

3. Modlin RL, Pirmez C, Hofman, et al. Lymphocytes bear­ing antigen-specific gamma delta T-cell receptors accumu­late in human infectious disease lesions. Nature 1989;339: 544-8.

4. Groh V, Porcelli S, Fabbi SM, et al. Human lymphocytes bearing T cell receptor '(/8 are phenotypically diverse and evenly distributed throughout the lymphoid system. J Exp Med 1989;169:1277-94.

March 1993 Am J Obstet Gynecol

5. Matis LA, Fry AM, Cron RQ, Cotterman MM, Dick RF, Bluestone JA. Structure and specificity of a class II MHC alloreactive '(8 T cell receptor heterodimer. Science 1989; 245:746-9.

6. Kabelitz D, Bender A, Schondelmaier S, daSilva-Lobo ML, Janssen O. Human cytotoxic lymphocytes. V. Frequency and specificity of gamma delta + cytotoxic lymphocyte precursors activated by allogeneic or autologous stimula­tor cells. J Immunol 1990;145:2827-32.

7. Itohara S, Farr A, Lafaille JJ, et al. Homing of a '(8 thymocyte subset with homogeneous T cell receptors to mucosal epithelia. Nature 1990;343:754-7.

8. Asarnow DM, Kuziel WA, Bonyhadi M, Tigelaar RE, Tucker PW, Allison JP. Limited diversity of gamma delta antigen receptor genes of Thy-l + dendritic epidermal cells. Cell 1988;55:837-47.

9. Savilahti E, Aralo A, Verkasalo M. Intestinal gamma/delta receptor-bearing T lymphocytes in celiac disease and in­flammatory bowel disease in children. Constant increase in celiac disease. Pediatr Res 1990;28:579-81.

10. Trejdosiewicz LK, Calabrese A, Smart CJ, et al. Gamma delta T cell receptor positive cells of the human gas­trointestinal mucosa: occurrence and V region gene ex­pression in Heliobacter pylori-associated gastritis, coeliac disease and inflammatory bowel disease. Clin Exp Immu­nol 1991;84:440-4.

11. Carding SR. A role for gamma/delta T cells in the primary response to influenza virus. Res ImmunoI1990;141:603-6.

12. Kaufman STE. Heat shock proteins and the immune response. Immunol Today 1990;11:129-36.

13. Daya S, Clark DA, Chaput A, Devlin C, JarrellJ. Suppres­sor cells in human decidua. AM J OBSTET GVNECOL 1985; 151:267-70.

14. Christmas SE, Bulmer IN, Meager A, Johnson PM. Phe­notypic and functional analysis of human CD3 - decidual leukocyte clones. Immunol 1990;71:182-9.

15. Clark DA, Falbo M, Rowley RB, Banwatt D, Stedronska­Clark J. Active suppression of host-vs-graft reaction in pregnant mice. J Immunol 1988;141:3833-40.

16. Clark DA, Flanders KC, Banwat D, et al. Murine preg­nancy decidua produces a unique immunosuppressive molecule related to transforming growth factor 132. J Im­munol 1990;144:3008-14.

17. Van Obberghen-Schilling E, Roche NS, Flanders KC, Sport MB, Roberts AB. Transforming growth factor beta 1 positively regulates its own expression in normal and transformed cells. J Bioi Chern 1988;263:7741-6.

18. DietlJ, Horny HP, Ruck P, et al. Intradecidual T lympho­cytes lack immunohistochemically detectable T cell recep­tors. Am J Reprod Immunol 1990;24:33-6.

19. Matsumoto K, Hashimoto K, Hashiro M, Yoshimasha H, Yoshikawa K. Modulation of growth and differentiation in normal human keratinocytes by transforming growth fac­tor-beta. J Cell Physiol 1990;145:95-101.

20. McCartney-Francis N, Mizel D, Wong H, Wahl L, Wahl S. TGF-beta regulates production of growth factors and TGF-beta by human peripheral blood monocytes. Growth Factors 1990;4:27-35.