HLA-G transgenic mice

Journal of Reproductive Immunology43 (1999) 253–261

HLA-G transgenic mice�

Andrew Mellor *, Min Zhou, Simon J. Conway, David MunnProgram in Molecular Immunology, Institute of Molecular Medicine and Genetics,

Medical College of Georgia, 1120 15th Street, Augusta, GA 30912 2600, USA

Abstract

We have generated a number of transgenic mice using DNA segments derived from theHLA-G gene. Using these mice we have examined the pattern of expression dictated byHLA-G promoter elements in mice and shown that HLA-G functions both as a restrictionelement and a transplantation antigen recognized by murine T cells. In addition, we haveshown that trophoblast cells expressing H-2Kb under the control of HLA-G promoterelements affect maternal T cell phenotype and responsiveness during pregnancy. Using thesesame HLA-G/H-2Kb transgenic mice we have shown that trophoblast cells, expressing aninducible enzyme that degrades tryptophan, protects allogeneic conceptus expressing pater-nally-inherited transgenes from attack by maternal T cells that leads to fetal rejection.© 1999 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: HLA-G; Transgenic; Mice; Pregnancy

www.elsevier.com/locate/jreprimm

1. Introduction

Transgenes incorporating HLA-G promoter elements and/or HLA-Gstructural genes are useful in studying the pattern of HLA-G expression andthe immunology of HLA-G in mice. Thus, transgenes incorporating HLA-G promoter elements are expressed at a high level throughout mousetrophoblast beginning shortly after implantation (Horuzsko and Mellor,1994; Horuzsko et al., 1997; Zhou and Mellor, 1998). Furthermore, studies

� Presented at the first International Conference on HLA-G, Paris, July 1998.* Corresponding author. Tel.: +1-706-7218733; fax: +1-706-7218732.E-mail address: [email protected] (A. Mellor)

0165-0378/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.PII: S0165 -0378 (99 )00027 -3

A. Mellor et al. / Journal of Reproducti6e Immunology 43 (1999) 253–261254

on the immunology of transgenic mice in which a HLA-G structural genewas expressed under the control of the H-2Kb promoter reveal that HLA-Gfunctions as a transplantation antigen and as a restriction element recog-nized by murine T cells (Horuzsko et al., 1997). HLA-G-specific murine Tcells recognize HLA-G as a native MHC molecule (direct recognition) andas processed peptides associated with MHC molecules (indirect recogni-tion). The finding that murine T cells recognize HLA-G as a restrictionelement suggests that HLA-G mediates a positive selection of murinethymocytes and, hence, shapes the developing repertoire of thymocytes inmice. These studies show that HLA-G structural and functional characteris-tics are similar to those of classical HLA class I molecules (HLA-A,B,C)that have also been expressed in transgenic mice (for examples see Horuzskoet al., 1997). The lack of any clear differences between HLA-G and classicalHLA-class I molecules provide no hints that might support the concept thathigh level HLA-G expression at the maternal–fetal interface during humanpregnancies contributes to fetal survival by moderating maternal T cellresponses.

To examine other potential explanations for survival of fetal allografts,we studied the effect of trophoblast-specific H-2Kb expression driven byHLA-G promoter elements on (a) maternal H-2Kb-specific, CD8+ T cellsand (b) a novel mechanism by which the trophoblast protects allogeneicconceptus from attack by maternal T cells (Munn et al., 1999). We find thattrophoblast-specific expression of paternally-inherited HLA-G/H-2Kb (GK)transgenes induces expansion of H-2Kb-specific, CD8+ T cells in maternalspleen during pregnancy. Furthermore, expression of the same paternally-inherited GK transgene induces rapid and uniform rejection of fetal allo-grafts when mothers are treated with an inhibitor of thetryptophan-catabolizing enzyme indoleamine-2,3 dioxygenase.

2. Materials and methods

Methods, reagents and transgenic mice described here have been pub-lished previously (Horuzsko and Mellor, 1994; Horuzsko et al., 1997; Zhouand Mellor, 1998). Analyses of pregnant mice treated with the indoleamine2,3 dioxygenase (IDO) inhibitor, 1-methyl-tryptophan, are described inMunn et al. (1999). In brief, slow release polymer pellets impregnated with1-methyl-tryptophan (0.9 mg/h) or placebo pellets were inserted surgicallyunder dorsal skin at 4.5 dpc. Pregnant mice were sacrificed at gestationtimes indicated and their uterus dissected to ascertain the number, develop-mental status and macroscopic and histological appearance of eachconceptus.

A. Mellor et al. / Journal of Reproducti6e Immunology 43 (1999) 253–261 255

3. Results

3.1. Targeting expression of H-2Kb to trophoblast using the HLA-Gpromoter does not induce fetal loss

In mice, polymorphic classical H-2K and H-2D molecules are expressedat high level in trophoblast from mid-gestation (Philpott, et al., 1988). Toinvestigate whether the HLA-G promoter targeted expression of H-2Kb todistinct trophoblast cell types and whether this interfered with fetal develop-ment, we prepared a recombinant GK transgene containing the HLA-Gpromoter linked to the H-2Kb structural gene (Zhou and Mellor, 1998). GKtransgenic mice were generated on the inbred CBA strain background.Matings were set up so that the GK transgene was inherited paternally([CBA×GK] matings). Fecundity rates were identical for [CBA×GK] andsyngeneic [CBA×CBA] matings and were slightly more than allogeneic[CBA×B6] matings. Of pups of [CBA×GK] matings, 57% inherited theGK transgene when males were heterozygous for the GK transgene (Table1). Immunohistological staining of trophoblast tissue sections revealed thatH-2Kb expression (a) was much higher than in control trophoblast fromCBA×B6 matings and (b) occurred throughout the trophoblast tissue andwas not localized to the maternal–fetal interface (Zhou and Mellor, 1998).These data show that high level expression of H-2Kb in trophoblast drivenby HLA-G promoter elements has no detrimental effect on development orsurvival of allogeneic fetus. Furthermore, HLA-G promoter activity is notconfined to cells lining the maternal–fetal interface of murine trophoblast.

3.2. Targeting expression of H-2Kb to trophoblast using the HLA-Gpromoter affects H-2Kb-specific T cells in maternal spleen

To assess the effect on maternal CD8+ T cells of targeting H-2Kb

expression to trophoblast using the HLA-G promoter, female BM3 T cell

Table 1Paternally-inherited expression of the GK transgene does not compromise fetal develop-ment

FecundityaMating genotype % GK transgenic

CBA×CBA 6.4 –CBA×GKb 576.4CBA×B6 –5.4

a Expressed as mean number of pups/litter at parturition.b GK males were heterozygous (�50% transgenic pups expected).

A.

Mellor

etal./

Journalof

Reproducti6e

Imm

unology43

(1999)253

–261

256

Fig. 1.


Table 2H-2Kb expression in trophoblast induces expansion of maternal H-2Kb-specific, CD8+ Tcells in maternal spleen

Splenocyte numbers (×106)aMice

Femaleb (×male) CD4+Total CD8+

15 (100)43 (100)BM3 (virgin) 6 (100)12 (80)BM3 (×CBA) 50 (116) 6.7 (112)17 (113)21 (350)BM3 (×GK) 78 (181)

15 (250)70 (163) 13 (87)BM3 (×B6)

a S.D.B10%; figures in parentheses are normalized to BM3 (virgin) females (100%).b Pregnant females were analyzed at 10.5 dpc.

receptor (TCR) transgenic mice were mated with male GK (homozygous)transgenic, B6 or CBA mice. All T cells in BM3 mice express TCRmolecules that confer recognition of H-2Kb (Sponaas et al., 1994) so thatphenotypic or functional effects on T cells caused by interactions betweencells expressing fetal H-2Kb molecules and maternal T cells can be readilyassessed. The number and phenotype of maternal splenic T cells wereassessed by flow cytometry at 10.5dpc (Zhou and Mellor, 1998). T cellswere stained with monoclonal antibody conjugates specific for murineCD8, CD4 and the clonotypic TCR expressed in BM3 TCR-transgenicmice.

The absolute number of splenic CD8+ T cells increased significantly inpregnant BM3 females mated with GK or B6 males (Table 2 and Fig. 1).A marginal increase in the absolute number of maternal CD8+ T cellswas observed during syngeneic pregnancies. Increased numbers of CD8+

T cells accounted for some of the increased spleen cellularity in pregnantfemales mated with GK or B10 (43, 33% of increase in spleen cellularity,respectively) males. No significant increase in absolute numbers of splenicCD4+ T cells were detected (Table 2) even though they also expressTCR transgenes conferring recognition of H-2Kb (Zhou and Mellor,1998).

Fig. 1. Expanded cohorts of maternal T cells specific for fetal H-2Kb appear in spleen duringpregnancy. Pregnant BM3 TCR transgenic mice mated to CBA (syngeneic), GK transgenicor B6 male mice were sacrificed at 10.5 dpc as described (Zhou and Mellor, 1998). (2) Colorhistograms show populations of CD8+(lower right quadrant) in maternal spleen. (1) Colorhistograms show level of CD8 (middle row) or TCR (lower row) on gated CD8+ T cells inspleen. A marker (M1) is included to permit comparisons of CD* and TCR levels.


Table 3Expression of paternally-inherited GK transgenes induces T cell mediated fetal rejectionwhen mice are treated with IDO inhibitor

Gestation Mean number of conceptus/female treatedMatingwith

IDO inhibitorFemale (×male) Stage Placeboa

7.36.5–8.5 7.5CBA (×CBA)6.5 6.79.5–15.5nt 5.8Term

7.38.5CBA (×B6) 6.55.5*7.5–8.5 8.30.2**5.49.5–15.57.8CBA-RAG-1−/− (×B6) 11.5 nt0ntCBA (×GK) 11.5

8.5 nt 0BM3 (×GK)

a nt, not tested.* PB0.01 or ** PB0.001 by ANOVA relative to placebo controls at each time point.

The amount of CD8 and TCR expressed by maternal CD8+ T cells wasreduced when H-2Kb was expressed on trophoblast cells during pregnancy(Fig. 1). The extent of this effect correlated with the amount of H-2Kb

expressed on trophoblast, since splenic T cells expressed less surface CD8and TCR in pregnant BM3 females mated with GK males than theircounterparts mated with B6 males. No decrease in the amount of surfaceCD4 or TCR expressed by CD4+ T cells from spleens of pregnant BM3female mice was observed (data not shown).

3.3. Targeting expression of H-2Kb to trophoblast using the HLA-Gpromoter induces rapid and specific loss of allogeneic conceptus when IDOacti6ity is blocked

Recently, we have shown that treating pregnant CBA mice with aninhibitor (1-methyl-tryptophan) of the enzyme IDO induces specific, uni-form loss of allogeneic conceptus in the period 7.5–8.5 dpc (Table 3). IDOdegrades the essential amino-acid tryptophan and the IDO gene is expressedat high level in mouse (Munn et al., 1999) and human trophoblast(Kamimura et al., 1991). We have shown that allogeneic fetal loss ismediated by maternal lymphocytes since pregnant females with an inducedmutation in the RAG-1 gene (on the CBA background) carry allogeneicconceptus to term when treated with IDO inhibitor (Table 3). However,trophoblast-specific expression H-2Kb from a paternally inherited GK


transgene is sufficient to provoke T cell mediated fetal rejection since CBAfemale mice mated with GK (homozygous) transgenic males lost all theirfetus when treated with IDO inhibitor (Table 3). The same outcome wasobserved when female BM3 were mated to GK males and exposed to IDOinhibitor.

4. Discussion

Transgenic mice carrying DNA segments derived from the human HLA-G gene have proved to be useful for studying the HLA-G promoter andimmunological characteristics of HLA-G molecules in mice. The HLA-Gpromoter is transcriptionally active in mouse trophoblast when inheritedpaternally (Horuzsko and Mellor, 1994; Zhou and Mellor, 1998). Wedetected very low levels of H-2Kb expression driven by the recombinant GKtransgene in adult tissues using a sensitive T cell bioassay to detect H-2Kb

expression (Zhou and Mellor, 1998). This may be of biological significancebut may arise from ectopic expression due to position effects at thetransgene integration site.

Using mice carrying a recombinant transgene in which an H-2Kb pro-moter was linked to a HLA-G structural gene, we have shown that HLA-Ginteracts with murine T cells and functions as both a transplantationantigen and as a restriction element (Horuzsko et al., 1997). However,surface expression of HLA-G was achieved in transgenic mice only whenhuman 2-microglobulin was co-microinjected with the recombinant H-2Kb/HLA-G transgene (unpublished results).

We have used transgenic mice carrying a recombinant HLA-G promoter/H-2Kb structural gene (the GK transgene) to examine the pattern of H-2Kb

expression dictated by the HLA-G promoter in mice (Zhou and Mellor,1998). We have shown that the GK transgene drives expression of H-2Kb

on the entire trophoblast and expression is not restricted to the maternal-fe-tal interface. This result shows that HLA-G promoter activity is not limitedto a single layer of cells at the maternal–fetal interface in mice. Thiscontrasts with the observation that HLA-G is expressed only in spongiotro-phoblast cells at the maternal-fetal interface during human pregnancy.These broad differences in the pattern of HLA-G promoter activity inmouse and human trophoblast are likely to arise from differential availabil-ity of trans-acting transcriptional factors that act on the HLA-G promoter.

We have used GK transgenic mice to investigate the effect of high leveltrophoblast-specific expression of the GK transgene on maternal T cells(Zhou and Mellor, 1998). In a related study (Tafuri et al., 1995), otherinvestigators found that maternal T cells were transiently tolerised to


paternally-inherited H-2Kb during pregnancy and this correlated withdown-regulation of CD8 and TCR on the surface of maternal T cellsspecific for H-2Kb. We also detected down-regulation of CD8 and TCR onH-2Kb-specific maternal T cells. However, we also found expanded cohortsof H-2Kb-specific T cells in maternal spleen during pregnancy. This expan-sion was driven by paternally-inherited H-2Kb alloantigen expression ontrophoblast since expanded cohorts of H-2Kb-specific T cells were notpresent in spleen when mothers carried syngeneic conceptus. We concludethat the maternal peripheral T cell repertoire is exposed to paternally-inher-ited MHC class I alloantigens during pregnancy, that this leads to antigen-specific expansion of maternal T cells, but this does not compromise fetaldevelopment (Table 1). This suggests that the trophoblast does not preventinteraction between maternal T cells and trophoblast cells expressing fetalalloantigens. However, it is not known whether these interactions occur soas to induce T cell unresponsiveness rather than activation that leads to thedevelopment of effector T cells capable of attacking trophoblast tissue. Theresults of Tafuri et al. (1995) provide indirect support for the view thatmaternal T cell unresponsiveness is the outcome of these interactions butthis is difficult to demonstrate unequivocally.

More recently, we have focused on a novel hypothesis to explain theparadox of fetal allograft survival in mammals (Medawar, 1953). We havedeveloped the hypothesis that cells expressing the tryptophan degradingenzyme IDO are involved in suppressing maternal T cell responses to fetalcells expressing paternally-inherited alloantigens. This hypothesis is basedon the unexpected discovery that cultured macrophages that acquire theability to suppress T cell proliferation in vitro do so by expressing IDO inresponse to signals from activated T cells which results in rapid consump-tion of tryptophan (Munn et al., 1999). Currently, we are investigating whyT cells do not proliferate when tryptophan levels are reduced. We wereintrigued to learn that the IDO gene is expressed in human placenta duringpregnancy (Kamimura et al., 1991) and that tryptophan levels are reducedin serum of pregnant women (Schrocksnadel et al., 1996). To test thehypothesis that cells expressing IDO protect the developing conceptus weblocked IDO activity in vivo using an IDO inhibitor and found that thisinduced specific loss of allogeneic conceptus while syngeneic conceptus alldeveloped to term. We used GK transgenic mice to demonstrate that fetalloss is T cell mediated since the only paternally-inherited alloantigenicdifference between mother and fetus in the CBA×GK mating is H-2Kb

(GK transgenic mice were made directly on the CBA background). Weconclude that cells expressing IDO in the trophoblast actively protect theconceptus from lethal attack by maternal T cells that recognize fetal MHCalloantigens. Maternal T cells may mediate fetal rejection directly or


indirectly; studies are underway to resolve this issue. We propose that theparadox of fetal allograft survival posed by Medawar almost 50 years ago(Medawar, 1953) is explained by trophoblast specific expression of IDO. Weare currently conducting experiments to examine how tryptophan consump-tion at the maternal–fetal interface prevents T cell mediated rejection ofmammalian fetus.

Acknowledgements

We thank Angie Compton for preparing the manuscript. Support forthese studies was provided by the MCG Department of Medicine and theCarlos and Marguerite Mason Trust.

References

Horuzsko, A., Mellor, A.L., 1994. Transcription of HLA-G transgenes commences shortlyafter implantation during embryonic development in mice. Immunology 83, 324–328.

Horuzsko, A., Antoniou, J., Tomlinson, P., Portik-Dobos, V., Mellor, A.L., 1997. HLA-Gfunctions as a restriction element and a transplantation antigen in mice. Int. Immunol. 9,645–653.

Kamimura, S., Eguchi, K., Yonezawa, M., Sekiba, K., 1991. Localization and developmentalchange of indoleamine 2,3-dioxygenase activity in the human placenta. Acta Med.Okayama 45, 135–139.

Medawar, P.B., 1953. Some immunological and endocrinological problems raised by evolu-tion of viviparity in vertebrates. Symp. Soc. Exp. Biol. 7, 320–328.

Munn, D.H., Shafizadeh, E., Attwood, J.T., Bondarev, I., Pashine, A., Mellor, A.L., 1999.Inhibition of T cell proliferation by macrophage tryptophan catabolism. J. Exp. Med.189, 1363–1372.

Philpott, K.L., Rastan, S., Brown, S., Mellor, A.L., 1988. Expression of H-2 class I genes inmurine extra-embryonic tissues. Immunology 64, 479–485.

Schrocksnadel, H., Baier-Bitterlich, G., Dapunt, O., Wachter, H., Fuchs, D., 1996. De-creased plasma tryptophan in pregnancy. Obstet. Gynecol. 88, 47–50.

Sponaas, A.-M., Tomlinson, P.D., Schulz, R., Antoniou, J., Ize-Iyamu, P., Schmitt-Verhulst,A.-M., Mellor, A.L., 1994. Tolerance induction by elimination of subsets of self-reactivethymocytes. Int. Immunol. 6, 1593–1604.

Tafuri, A., Alferink, J., Moller, P., Hammerling, G.J., Arnold, B., 1995. T cell awareness ofpaternal alloantigens during pregnancy. Science 270, 630–633.

Zhou, M., Mellor, A.L., 1998. Expanded cohorts of maternal CD8+ T cells specific forpaternal MHC class I accumulate during pregnancy. J. Reprod. Immunol. 40, 47–62.

.

HLA-G transgenic mice

Documents

Transcript of HLA-G transgenic mice