© INsTITuT PASTEURIELSEVIERParis 1997

Res. Immunol.1997, 148,247-256

In vivo study of interferon-alpha-secreting cellsin pig foetallymphohaematopoietic organs

following in utero TGEV coronavirus injection

I. Splfchal ( I ) (*), z. ReMkova (I), M . Sinkora (1), J. Sinkora (I), I. Trebichavsky (I),

H. Laude (2) and B. Charley (2)

(IJ Division of Immunology and Gnotobiology, Institute ofMicrobiology of the Academyof Sciences of the Czech Republic, 54922 Novy Hnidek (Czech Republic), and

(2) Laboratoire de Virologie et Immunologie moleculaires, INRA, 78350 Jouy-en-Josas (France)

SUMMARY

Non-infectious UV·inactivated transmissible gastroenteritis virus (TGEV) was previ­ously shown to induce interferon alpha (IFNa) secretion following in vitro incubationwith blood mononuclear cells. In this study, pig foetuses at different stages of gestationwere injected in utero with (a) partially UV·inactivated wild TGEV or (b) fUlly UV-inacti­vated wild or dm49-4 mutant TGEV coronavirus. Nucleated cells from foetal liver, bonemarrow, spleen and blood were isolated 10 or 20 h after injection and assayed ex vivofor IFNa secretion by EUSPOT and ELISA techniques. The administration of TGEVinduced IFNa-secreting cells in foetal Iymphohaematopoietic organs at mid-gestation. Incontrast, IFNa was not detected in control sham-operated foetuses. A specific pointmutation in the amino acid sequence of the viral membrane glycoprotein M of TGEVmutant dm49-4 was associated with lower or absent IFNa in utero inducibility by mutantvirus as compared with wild virus. Row cytometry analysis did not show differences inleukocyte surface marker expression between control and TGEV· or between dm49-4and wild virus-treated foetus cells, with the exception of a reduction in percentages ofpolymorphonuclear cells in TGEV-treated Iymphohaematopoietic tissues, which is prob­ably due to IFNa secretion. The present data provided in vivo evidence of IFNa secretionat the cell level in foetallymphohaematopoietic organs. Such IFNa-secreting cells in Iym­phohaematopoietic tissues may be the source of IFNa detected during foetal infections.

Key-words: Coronavirus, Transmissible gastroenteritis virus, IFNa; ELISA, ELISPOT,Foetus, Pig.

INTRODUCTION

Interferon-a (IFNa) is a critical component ofearly, host non-specific immune defence againstviral infections. It acts as both an antiviral agent

and an immunomodulator as well as a cell growthinhibitor. Several leukocyte populations are ableto secrete IFNa, in response to virus stimuli,depending on the virus used and on whether viralinfection of leukocytes is necessary or not.

Submitted January 9. 1997, accepted Apri124, 1997.

(*) Correspond ing author. present address : National Institute of Animal Health, Kannondai, Tsukuba, Ibaraki 305, Japan.

248 I. SPLicHAL ET AL.

Monocytes are most often associated with pro­duction of IFNa in response to infectious viruses(Roberts et al., 1979; Saksela et al., 1984),whereas a distinct population of leukocytesamong peripheral blood mononuclear cells(PBMC), referred to as "natural interferon-pro­ducing cells" (NIPCs), is able to secrete IFNafollowing exposure to non-infectious viral struc­tures (Lebon et al., 1982a; Capobianchi et al.,1985; Charley and Laude, 1988). Human andporcine NIPCs have been characterized as highlyinfrequent, non-phagocytic, non-adherent, non-B,non-T, MHC II+ and CD4+ cells (Cederblad andAIm, 1990; Charley and Lavenant, 1990; Sand­berg et al., 1990; Fitzgerald-Bocarsly, 1993;Nowacki and Charley, 1993).

Viral glycoproteins were suggested as beingresponsible for triggering synthesis of IFNa inNIPC. Monoclonal antibodies (mAbs) to viralglycoproteins of herpes simplex virus type I,transmissible gastroenteritis virus (TGEV) andAujeszky's disease virus were shown to blockvirus-induced IFNa. secretion (Lebon, 1985;Charley and Laude, 1988; Artursson, 1993).However, the precise nature of NIPCs as well asof the interactions between NIPCs and virus lead­ing to IFNa production remains to be elucidated.

TGEV is an enteric coronavirus which causesacute and fatal diarrhoea, as well as intense andearly IFNa production in newborn piglets (LaBonnardiere and Laude, 1981). In vitro studies onIFNa induction by TGEV have shown thatIFNa-secreting cells (IFNa-SCs) were detectedamong non-adherent porcine PBMCs after expo­sure to UV-inactivated TGEV or glutaraldehyde­fixed TGEV-infected cells (Charley and Laven­ant, 1990; Nowacki and Charley, 1993). Aspecific point mutation in the amino acidsequence of the viral membrane glycoprotein M

in dm49-4 mutant TGEV was shown to be asso­ciated with a defect in in vitro induction of IFNa(Laude et al., 1992).

The epitheliochorial nature of placentation inpig prevents transfer of immunoglobulins or anti­gens from mother to foetus, which I?rec1udes anyimmune activation of foetuses (Sterzl et al.,1966). This type of placentation together withmultiparity, relatively long term of gestation andsize of foetuses makes this species suitable forstudies on the development of the immunesystem. In an earlier study on the prenatal ontog­eny of porcine IFNa-SCs, we detected in vitroinducible IFNa-SCs in pig foetal liver and other1ymphohaematopoietic organs, at very earlystages of gestation (Splichal et aI., 1994). Thepresent study was undertaken to evaluate in uteroviral induction of IFNa-SC in pig foetuses byTGEV, either partially or fully UV-inactivated, atdifferent stages of gestation. We found that 10 or20 h after in utero injection of TGEV in theumbilical cord of pig foetuses, IFNa-SCs weredetected ex vivo in Iymphohaematopoietic tissuesby the ELISPOT assay.

MATERIALS AND METHODS

Animals

Healthy pregnantgilts of miniature pig bred in theLaboratory of gnotobiology in Novy Hnidek wereused in our experiments. They had free access towater but were starved 12 hours before first surgery.The gilts were subcutaneously (s.c.) premedicatedwith 1 mg of atropin sulphate (Hoechst-Biotika, Slo­vakia) per 25 kg of body weight, and they wereanaesthetized with 1.5-2.5% of halothane (Leeiva,Czech Republic) mixed with 0 and Np. HCO0,500 U) (Leciva, Czech Republic) and acetate

FCMFCSFSCHCGIFNIFNa-SCmAbMOCNIPC

now cytometry (analysis).foetal calf serum.forward scatter.human chorionic gonadotropin.interferon.IFNa-secreting cell.monoclonal antibody.mononuclear cell.natural interferon-producing cell.

PBLPBMCPBSPFUPMNSSCSWC3TGEV

peripheral blood leukocyte.peripheral blood mononuclear cell.phosphate-buffered saline.plaque-fonning unit.polymorphonuclear (cell).side scatter.an antigen common for myelomonocyllc lineages.transmissible gastroenteritis vrrus,

IN UTERO INDUCTION OF IFNa-SECRETING CELLS 249

medroxyprogesterone (50 mg per 25 kg of bodyweight) (Upjohn, Netherlands) were intramuscularly(i .m.) injected. In the first series of experiments, four­teen foetuses of 52, 82 and 101 days of gestation wereinjected with partially UV-inactivated TGEV (500PRJ/ml in 50, 300 and 500 III of saline, respectively)via the umbilical vein when the umbilical cord wasexteriorized after laparotomy and uterotomy of thegilts. Control foetuses were subjected to the same sur­gery but treated with equivalent volumes of the salineonly. The uterine and abdominal walls were sutured,and gilts were placed in a postsurgical care unit Theyhad free access to water but a limited amount of food.Animals were treated with 1,500,000 U penicillin G(Spofa, Czech Republic) s.c. and 0.5 g streptomycin(Medexport, Russia) i.m. per 25 kg of body weight.

In the second series of experiments, twenty-onefoetuses at 75, 77, 91 and 105 days of gestation wereinjected via the umbilical vein (300-1,500 Ill) with afully UV-inactivated wild or dm49-4 mutant TGEV(initial titre of 6x 107 PFU per mI before inactiva­tion). The injected volume was proportionallyadjusted to expected body weight. The treatment ofanimals was the same as described above.

Cell suspensions

The second uterotomy was performed 20 or10 hours later, in the first or second series of experi­ments, respectively. The sows were anaesthetized,foetuses were bled via umbilical cord arteries, andblood samples containing 20 U of heparin/ml ofblood (Leeiva, Czech Republic) were collected. Cellsuspensions from liver, spleen, both femurs and ster­num were prepared by cutting these organs with scis­sors in cold PBS. Supernatants containing cells werecollected after 10 min sedimentation at I g to removedebris and clumps. PBMCs were prepared by centri­fugation of diluted blood on "Ficoll" density gradient(Pharmacia, Sweden), and red cells were depleted byhypotonic lysis with deionized water as described(Splfchal et al., 1994). In the case of a non-adherentcell fraction isolation, cells were incubated for 90min in RPMI-I640 with 20% foetal calf serum (FCS)(Seromed, Germany), L-glutamine (2 roM), penicillin(100 UlmI) and streptomycin (100 ug/ml) (Gibco,UK) in 5 % CO atmosphere at 37°C to remove plas­tic-adherent celts. The number of nucleated cells andcell viability were calculated before assays.

IFNa assays

Cell suspensions (200 Ill) in RPMI with 10 %FCS and antibiotics were log , diluted in 3-6 wells of96-well cell culture microplate (Costar, UK). ELlS­POT was performed as described (Nowacki andCharley, 1993). Cells in ELlSPOT plates or in tissue

culture microplates were incubated in 5 % COatmosphere for 16 and 18 h, respectively. IFNa-SCfrequency was estimated by an ELISPOT assayusing peroxidase-labelled anti-pig IFNa mAb F17.Spots were counted under binocular microscopy,and the frequency was calculated from the totalnumber of living nucleated cells and number ofspots . IFNa titres in supernatants from tissue culturemicroplates incubated for 18 h, and in plasma, wereestimated by pig IFNa-specific ELISA using thesame antibodies as in the case of ELISPOT (DeArce et al., 1992; Nowacki and Charley, 1993).Results arc expressed as IFNa unit/m!. Productionof IFNaJlFNa-SC (yield) was estimated from theIFNa titre and the number of IFNa-specific spots.

Flow cytometry analysis

The following mouse mAbs directed against por­cine leukocyte surface markers were used: K252.1E4(anti-CD45), 74-22-15 (anti-SWC3), an antigen com­mon for myelomonocytic lineages), 1O-2H2 (anti­CD4) and MSA3 (anti-SLA-DR). Erythrocytes wereremoved from cell suspensions by hypotonic lysis ofthe cell pellet with water. Leukocytes were washedand stained as described recently (Cukrowska et aI.,1996): briefly, cells were treated with a primarymouse mAb and then with fluorescein-conjugatedswine anti-mouse Ig polyclonal antibodies (Sevac,Czech Republic); for double staining, biotinylatedmAbs were revealed by streptavidin-phycoerythrinconjugate (Immunotech, France). Cells were dividedinto two major populations on the basis of their sizeand internal complexity: polymorphonuclear cells(PMN) with higher SSC parameter gated separatelyfrom mononuclear leukocytes (MOC) with lowerinternal complexity. Flow cytometry data wereobtained using a "FACSort" flow cytometer (Becton­Dickinson, CA). Propidium iodide was added to cellsjust before cytometry to prevent counting of dead anddamaged cells, and at least 10,000 events were col­lected. Data were analysed using PC-LYSYS 1.0software (Becton-Dickinson, CAl.

Virus

High-passage Purdue 115 "wild strain" of TGEVand dm49-4 mutant TGEV (Laude et al., 1992)were used as virus sources. Procedures for viruspreparation have been described previously (Char­ley and Laude. 1988). Viruses were inactivated byUV irradiation. In the first series of experiments,TGEV was partially UV -irradiated to obtain a resid­ual infectivity titre of 5x 104 PFUlml. In the secondseries of experiments, both wild and dm 49-4mutant virus, at initial titres of 6x 107 PFU/ml, werefully inactivated before injection.

250 I. SPLiCHAL ET AL

RESULTS

Induction of IFNa-SC following in uteroinjection of partially inactivated wild TGEV

The optimal amount of partially inactivatedwild TGEV allowing foetus survival after intrave­nous injection at 54 days of gestation was deter­mined in preliminary experiments. Foetuses werethereafter infected with this amount of partiallyinactivated TGEV, and other foetuses were sham­operated as controls. At 52 days of gestation onlythree foetuses (two infected and one control) sur­vived till the next day; dead foetuses infected byvirus did not exhibit any macroscopic pathologi­cal lesions when compared with the dead controlfoetus. Only cell suspensions and plasma of sur­viving foetuses were used for IFNa determina-

tions.IFNa-SCs or IFNa were detected in liverand bone marrow cells or plasma in only one oftwo TGEV-treated foetuses already at mid-gesta­tion (day 52) (table I). In the spleen they werefound at later stages of gestation. The highestIFNa-SC frequency was observed in foetal liver.A slightly higher IFNa-SC frequency wasdetected in non-adherent cells compared withthat in total cells (table I, 101 days). The absenceof lFNa-SCs in spleen cells at 52 days may bedue to the limited number of cells. The highestIFNa-SC frequency was observed by 101 days ofgestation. whilst IFNa yield (production of IFNaper cell) was roughly constant (table II). No IFNasecretion in bone marrow cell culture supernatantscould be detected at 52 days of gestation (table II),although low numbers of IFNa-specific spotswere detected in one stimulated foetus (table I).

Table I. IFNa-secreting cells in pig foetal Iymphohaematopoietic organs, and plasma IFNa levels 20 h afterexperimental in utero injection of partially UV-inactivated wild TGEV coronavirus.

IFNa-SC fr~uency IFNa level(spots per 1 cells) (units per ml)

Day of gestation Liver Spleen Bone marrow Plasma

52nd (n = 2) 26; 0 o» 2.1; 0 4800; 082nd (n = 4) 84.5 ± 57.7 2.2 ± 1.1 14.8 ± 13.5 9200 ± 2870IOlst (n = 3) 47.3 ±23.4 6.3 ± 0.2 7.7 ± 4.7 10 190 ± 2 920IOlst<a) (n = 3) 66.3 ± 23.7 ND ND

Results are expressed as individual data or as means ± standard deviation. Day of gestation =day of stimulation by TGEV; ( a) non­adherent fraction; (b) low number of cells. ND=non-detected.

Numbers of sham-operated controls on S2nd, 82nd and JOist day of gestation were 1,2 and 2, respectively. No IFNa-SC or IFNatitres were detected in Iymphohaematopoietic organs or plasma of controls.

Table II. IFNa yield per cell in pig foetallymphohaematopoietic cell cultures 20 h after experimental in uteroinjection of partially UV-inactivated wild TGEV coronavirus.

Day of gestation

52 nd (n = 2)82nd (n =4)l Olst (n=3)1OIst(a) (n =3)

Liver

0.7 ; 01.9 ± 0.80.3 ± 0.20.2 ± 0.]

IFNa secretion(units per IFNa-SC)

Spleen

O(b)

0.3 ± 0.10.2 ± 0.1

ND

Bone marrow

o0.9 ± 0.80.2 ± 0.1

ND

(aJ Non -adherent fraction ; (b) low number of cells . ND=non-detected.No IFNa titres were detected in cell culture supernatants of Iymphohaematopoietic organs of controls.

IN UTERO INDUCTION OF IFNa-SECRETING CELLS 251

As controls, IFNa titres were determined inthe plasma of TGEV-injected foetuses. HighIFNa titres were found in foetal plasmas at 82and 101 days of gestation (table I). Only oneTGEV-injected foetus had plasma IFNa at 52days of gestation.

No lFNa-specific spots or IFNa secretion wasfound in control sham-operated foetuses, whichwere subjected to in utero injection of saline only(data not shown).

Flow cytometry (FCM) analysis was per­formed in 52- and IOI-day-old foetuses.SSCIFSC dot plot analysis showed a lower per­centage of PMN cells in virus-injected foetal pigorgans (fig. 1). FCM of leukocyte cell markersdid not show significant differences between con­trol and TGEV-injected foetuses (data notshown).

After evaluation of the results from this firstseries of experiments, we reduced the length of inutero stimulation from 20 to 10 hours in order toanalyse IFNa secretion at its expected time ofmaximal production.

Induction of IFNa-SC following in uteroinjection of fully inactivated wild and dm49-4mutant TGEV

In a second series of experiments using fullyinactivated virus, differences in IFNa inductionin pig foetuses stimulated by wild or dm49-4mutant TGEV were analysed. In order toexclude any possible IFNa secretion by cells ofmonocytic/macrophage lineage, only non­adherent cells were used (as performed withliver cells at 101 days: table I). The main dif­ferences observed between the two series ofexperiments were a much lower plasma IFNalevel, a reduced IFNa-SC frequency in liverand spleen and the absence of IFNa-SCs in foe­tal bone marrow (table III compared withtable I). IFNa yields per cell were similar(tables II and IV). Neither plasma IFNa norIFNa-SCs in organs were found in dm49-4­injected foetuses, with the exception of verylow IFNa-SC numbers in liver and spleen of75-day-old foetuses (tables III and IV).

Flow cytometry showed lower percentages ofPMN cells in liver and bone marrow of wildTGEV-infected foetuses than in dm49-4-treatedfoetuses, with one exception - the liver of 105­day-old foetuses (fig. 2). No significant changesin leukocyte marker expression were observed(data not shown).

DISCUSSION

Intraamniotic infections provoke abortion,infertility, foetal death and abnormal foetal devel­opment and are associated with increased levelsof amniotic inflammatory cytokines (Gravett etal., 1994; Romero et al., 1994; Dudley et aI.,1996). J.FN was found not only in human foetusesduring intraamniotic rubella infection (Lebon etal., 1985) but also in the amniotic fluid of preg­nant women without clinical signs of congenitalvirus infection (Lebon et al., 1982b). Followingour previous demonstration that in vitro inducibleIFNa-SCs were present at early stages of gesta­tion in porcine foetal lymphohaematopoieticorgans (SpHchal et al., 1994), the aim of thepresent study was therefore to evaluate, at the celllevel, the secretion of IFNa following in uterointravenous injection of TGEV in pig foetuses atdifferent stages of gestation. The study wasdivided into two parts: (a) induction of IFNasecretion by partially inactivated wild TGEV, forwhich IFNa secretion was analysed in total cellpopulations (first series of experiments) and (b)induction of IFNa secretion in the non-adherentIFNa-SC fraction (presumably NIPC-like cells)after exposure to non-infectious TGEV. Wild anddm49-4 mutant TGEVs were compared in orderto evaluate the influence of an amino acid pointmutation in protein M on the IFNa-inducingproperties in vivo (second series of experiments).

The experimental approach using "open" sur­gery is an efficient procedure for in vivo infectionof foetuses via the umbilical cord vein. It has,however, several disadvantages such as loss ofamniotic fluid, damage of foetal membranes andpossible damage of the umbilical cord (Kovaru etal., 1971). In the present study we observed highmortality of pig foetuses at 52 days of gestation

252 I. SPLiCHAL ET AL

o 1000

Stimu lated fetusoO .r-- - ---,- - - - - - - - -.o

oollCl

A,: !§~ .0;:' c:a= 0, ....C-l

'"..... 0

~~I

C-l

'"lI'I o+--~--r~~....._-_r--r__.....,J

: ' ", -- -r

.' ,- :.':

Non-stimuiated fetus

aog ..-- - - - - - "7""'"- -:-- -,...,-,

P orC lnt of PMN CIUS (%17'0 ,--- - - - - - - - - - - - - - - - - - - ..,c

50

n

10

umbl~c.1

bloodbone

m ilrTC'",'-liver umbll,cal

bloodbone

marro-...: ~r

5:lnd d3y of g_st. toon

o NO ,STlMU TEO 0 STIMUl ATEO BY LNE TGEV

01 SI day ~f ;o st al'on

Fig. 1. FCM of umbilical cord blood. bone marrow and liver cells of 52- and IOl-day-old foetusesafter in utero injection of control medium or TGEV .

Dot plot analysis of foetal bone marrow cells (101 days of gestation) from control (a) or TGEV­injected (b) foetuses. Percentages of PMN cells in the different cell suspensions at 52 and 101 daysof gestation (c). Number of foetuses as indicated in table 1.

even in the case of one control foetus treated bysaline only, which could be due to easier damageof the umbilical cord at this period of gestationthan with older foetuses.

The main finding of the present study was thatIFNa-SCs were found as early as 52 days of ges­tation in foetal liver and bone marrow followingin utero injection of partially UV-inactivated

IN UTERO INDUCTION OF IFNa-SECRETING CELLS 253

Table Ill. IFNa-secreting cells in pig foetallymphohaematopoietic organs and plasma IFNa levels to h afterexperimental in utero injection of UV-inactivated wild and dm49-4 mutant TGEV coronavirus.

LiverWild virus Mutant virus

Day of gestation(wild/mutant)

IFNa-SC fre~uency

(spots per 10 cells)Spleen

Wild virus Mutant virus

IFNalevels(units per ml)

PlasmaWild virus Mutant virus

75th (n =3/3)77th (n = 211)91st (n =3/3)105th (n =3/3)

3.4 ± 2.90.7; 1.73.7 ± 1.40.1 ± 0.1

0.1 ± 0.1ooo

1.9 ± 1.40; 2.4

28.8 ± ll.518.1 ± 1.7

0.1 ± 0.1ooo

990 ± 120ND

1480 ± 11801400 ± 780

oooo

IFNa ELISPOT assay was performed with non-adherent cells. ND = non-detected .

Table IV. IFNa yield per cell in pig foetal lymphohaematopoietic organs 10 h after experimental in uteroinjection of U'V-inactivated wild and dm49-4 mutant TGEV coronavirus.

IFNa secretion(units per IFNa-SC)

LiverDay of gestation (wild/mutant) Wild virus Mutant virus

SpleenWild virus Mutant virus

75th (n =3/3)77th (n =2/J)91rd (n = 3/3)105th (n =3/3)

0.8 ± 0.40.2 ; 1.20.8 ± 0.40.1 ± 0.1

0.1 ± 0.1ooo

0.1 ± 0.10; 0.3

0.4 ± 0.10.3±O.1

oooo

IFNa ELISPOT assay was performed with non-adherent cells .

virus . This is, to our knowledge, the first demon­stration of the presence of IFNa-SCs in foetallymphohaematopoietic tissues after in vivo viralinduction. The fact that IFNa-SCs were found inonly one virus-injected foetus may be due to anunsuccessful virus inoculation. At 52 days of ges­tation, a low frequency of spots but no IFNa insupernatants were observed in bone marrow cellsof one foetus. This discrepancy may be due to alower sensitivity of ELISA in comparison withthat of EllSPOT.

Haematopoiesis in the bone marrow of pigf~etuses starts at around 50 days of gestation(Sterzl and Kovai'u, 1977), and the finding ofIFNa.-SCs at that time indicates the early appear­ance of these cells in lymphohaematopoieticorgans. This finding is in accordance with ourprevious data on in vitro inducible foetal lFNa-

SCs (SplIchal et al., 1994) and supports the likelyhaematopoietic origin of IFNa-SCs (Charley etal., 1995).

At later stages of gestation, in utero TGEVinjection induced IFNa-SCs in foetal liver, spleenand bone marrow. When fully UV-inactivatedTGEV was used, IFNa-SCs were not found inbone marrow, which may reflect the inability ofnon-replicating virus to reach that organ. A muchlower IFNa production was observed after injec­tion of fully inactivated virus compared with thatafter injection of infectious TGEV, as evidencedby both lower plasma IFNa titres and lFNa-SCfrequency. This lower IFNa induction very prob­ably reflects a lower amount of virus availablewhen non-infectious TGEV is injected.

lFNa-SCs or IFNa was never found in bloodcell suspensions (negative data not shown), but

254 1. SPLiCHAL ET AL.

Percent of PM cells I")

BONEMARROW LIVER

DAY OF GESTATIO

o LD sus fEl 'wl lJT,lNT VIRUS

Fig. Z. Percentages of PMN cells in bone marrow and liver cell suspensions of wild or dm49-4mutant TGEV-injected foetuses of different ages. Numbers of foetuses as indicated in table III.

due to the low number of PBMC obtained, wecannot conclude whether these negative resultscould only be due to an insufficient number ofcells tested or to a real absence of IFNa-SCs.

The lower IFNa inducibility by the UV-inac­tivated dm49-4 TGEV mutant characterized bya mutation in glycoprotein M, in comparisonwith inducibility by wild virus, has alreadybeen shown in in vitro experiments (Laude etal., 1992). In the present work. IFNa wasinduced by the dm49-4 mutant to very low lev­els at only one period of gestation (75 days),while IFNa was observed in liver and spleen offoetuses at all ages tested, after induction bywild virus . This result implies that in uteroIFNa induction by inactivated TGEV largelydepends upon the presence of a native envelopeprotein M. as expected from in vitro studies(Laude et al., 1992).

The presence of IFNa-SCs in pig lyrnphohae­matopoietic tissues, together with plasma IFNa,

confirms recent data obtained in mice within abroader framework : briefly, early and transientserum IFNa production was detectable as well asthe presence of IFNa-producing cells localized inspecific areas of spleen and lymph nodes, but notin bone marrow, after in vivo inactivated virusinjection (Eloranta et al., 1996; Riffault et al.,1996).

The aim of FCM was to study changes incell populations during induction of IFNa pro­duction. NIPC size and granularity as well assurface markers have been described (Sand­berg et al., 1990; Charley and Lavenant,1990; Cederblad and Aim, 1990; Fitzgerald­Bocarsly, 1993 ; Nowacki and Charley, 1993).A more comprehensive characterizat ion ofhuman NIPC surface markers has recentlybeen published (Svensson et al., 1996) . Thelatter study co ncluded that NIPC could beimmature dendritic cells. In the present study,cell s isolated from fo etal Iymphohaemato­poietic org ans were analy sed on the basis of

IN UTERO INDUCTION OF IFNa-SECRETING CELLS 255

their size, internal complexity (SSC/FSC) andsurface markers . PMN with a higher sseparameter could be distinguished from mono­nuclear leukocytes (MaC) with lower internalcomplexity. The PMN/MOC ratio was foundto be reduced in virus-treated foetal organs(figs. l a, b and c). In most cases studied, PMNpercentages were lower in wild-virus-treatedthan in dm49-4-treated foetuses (fig. 2).Because dm49-4 did not induce IFNa in vivo,it is likely that TGEV-induced reduction ofPMN percentages is IFNa-mediated. Previousstudies also showed that immunization of pigfoetuses with corpuscular antigen resulted indepletion of myeloid cells from foetallympho­haematopoietic organs (Sterzl and Kovafu,1977).

No significant differences in leukocyte sur­face marker expression between stimulatedand non -stimulated animals, including SLA­DR and CD4 previously described on adult pigblood IFNa-SCs (Nowacki and Charley,1993), were observed in any tested cell sus­pensions.

In conclusion, our present results provide invivo evidence of IFNa secretion, at the celllevel, in pig foetal lymphohaematopoieticorgans, at early stages of lymphohaernato­poietic development. IFNa detected in foetalinfected tissues (Lebon et al., 1985) may indeedoriginate from NIPCs localized in foetal lym­phohaematopoietic tissues. In addition, ourexperimental model of in utero induction ofIFNa in pig foetuses will prove to be useful forfuture studies on prenatal cytokine networkdevelopment.

Acknowledgements

We gratefully thank Lydia Besnardeau for preparingviruses, Christiane de Vaureix for anti-IFNa mAb, Kve­toslava Rybinova and Marie Zahradnfckova for theirexcellent technical assistance during surgery, and MartaSrnolova and Eva Cervinkova for preparing cell suspen­sions. We thank Christine Young of the Translation Ser­vice. INRA, for revising the English.

This work was partly supported by INRA (DRI) andgrants no. 508/94/0153 and 524/97/0533 of the GrantAgency of the Czech Republic.

L'injection in utero de coronavirus VGETinduit la secretion d'lnterferon alpha

dans les organes lymphohematopoletiquesde foetus de pore

Nous avons montre precedemment que I'incubationin vitro de cellules rnononucleees sanguines avec ducoronavirus de la gastroenterite transmissible (VGET)inactive par irradiation UV, induisait la secretiond'interferon alpha (IFNex). Le travail presente ici aconsiste a realiser sur des foetus de pores a differentsstades de gestation une injection in utero (a) de VGETsauvage partiellement inactive par les UV, ou (b) devirus sauvage ou de mutant dm49-4 totalementinactives. La secretion d'IFNa par les cellulesnucleees isolees, 10 ou 20 h apres injection, de la rate,du foie, de la moelle osseuse et du sang fcetaux a eteanalysee ex vivo par technique ELISPOT et ELISA.L'injection de VGET a induit la presence de cellulessecretrices d'IFNa dans les organes lyrnpho­hernatopoletiquesfoetaux ami-gestation. Une secretiond'IFNa plus faible, voice absente, a ete observee a lasuite de l'injection du mutant viral dm49-4 caracterisepar une mutation ponctuelle dans la sequence de laglycoproteine M. L'etude en cytometrie de flux n'apas pennis de montrer de differences d'expression desmarqueurs de surface leucocytaires entre les cellulesde fcetus traites par Ie VGET ou par du milieu, outraites par Ie virus sauvage ou Ie mutant dm49-4, al'exception d'une reduction du pourcentage depolynucleaires, apres injection par Ie VGET, probable­ment due aI'INFa secrete.

Ces resultats montrent, in vivo, la secretion auniveau unicellulaire d'IFNa dans les tissuslymphohematopo'ietiques foetaux. De telles cellulessecretrices d'IFNa localisees dans les tissus pourraientconstituer la source de l'IFNa qui est detecte au coursd'infections fcetales,

Mots-des: Coronavirus, Virus de la gastro­enterite transmissible, IFNa; ELISA, ELISPOT,Foetus, Pore.

References

Artursson, K. (1993). Studies on the interferon-a/B systemof pigs. Ph.D. thesis. Uppsala.

Capobianchi, M.R.. Facchini. J., Di Marco. P.. Antonelli.G. & Dianzani. F. (1985 ). Induction of alpha inter­feron by membrane interaction between viral surfaceand peripheral blood mononuclear cells. Proc. Soc.Exp. Bioi. Med., 178. 551-556.

Cederblad, B. & Aim, G. (1990). Infrequent hut efficientinterferon-rr-producing human mononuclear leuko-

256 I. SPLiCHAL ET AL

cytes induced by herpes simplex virus in vitro studiedby immunoplaque and limiting dilution assays. J.Interferon Res., 10, 65-73.

Charley, B. & Laude, H. (1988), Induction of alpha inter­feron by transmissible gastroenteritis coronavirus: roleof transmembrane glycoprotein E1. J. Yirol., 62, 8-11.

Charley, B. & Lavenant, L. (1990), Characterization ofblood mononuclear cells producing IFN a followinginduction by coronavirus-infected cells (porcinetransmissible gastroenteris virus). Res. Immunol.,141, 141-151.

Charley, B., Nowacki, W. & Vaiman, M. (1995), Fre­quency of interferon-ex-secreting leukocytes in irradi­ated and bone-marrow-grafted pigs. Vet. Res., 26,292-299.

Cukrowska, B., Sinkora, J., Rehakova, Z., Sinkora, M.,Splichal, I., Tuckova, L., Avrameas, S., Saalmuller,A., Barot-Ciorbaru, R. & Tlaskalova-Hogenova, H.(1996), Isotype and antibody specificity of spontane­ously formed immunoglobulins in pig fetuses andgerm-free piglets: production by CD5- B cells.Immunology, 88, 611-617.

De Arce, H.D., Artursson, K., L'Haridon, R., Perers, A.,La Bonnardiere, C. & Aim, G.V. (1992), A sensitiveimmunoassay for porcine interferon-a. Vet. Immunol.Immunopathol., 30, 319-327.

Dudley, D.J., Hunter, c., Mitchell, M.D. & Varner, M.W.(1996), Elevations of amniotic fluid macrophageinflammatory protein-I alpha concentrations inwomen during term and preterm labor. Obstet. Gyne­col., 87, 94-98.

Eloranta, M.L., Sandberg, K. & Aim, G.V. (1996), Theinterferon-aJl3 responses of mice to herpes simplexvirus studied at the blood and tissue level in vitro andin vivo. Scand. J. lmmunol., 43, 355-360.

Fitzgerald-Bocarsly, P. (1993), Human natural interferon­a-producing cells. Pharmacol. Ther., 60, 39-62.

Gravett, M.G., Witkin, S.S., Haluska, G.J., Edwards, J.L.,Cook. MJ. & Novy, M.MJ. (1994), An experimentalmodel for intraamniotic infection and preterm laborin rhesus monkeys. Am. J. Obstet. Gynecol., 171,1660-1667.

KovlitiJ, F., Stozicky, V., Krurnl, J., Dlabac, V., Donat, J.& Novotna, J. (1971), Experimental surgery in thefoetal period of mammals. Acta. Vet. Bmo, S3, 1-68.

La Bonnardiere. C. & Laude, H. (1981), High interferontiter in newborn pig intestine during experimentallyinduced viral enteritis. Infect.Tmmun., 32, 28-31.

Laude, H., Gelfi, J., Lavenant, L. & Charley, B. (1992),Single amino acid changes in the viral glycopro­tein M affect induction of alpha interferon by thecoronavirus transmissible gastroenteritis virus. J.Virol., 66, 743-749.

Lebon, P. (1985), Inhibition of herpes simplex virustype t-induced interferon synthesis by monoclonal

antibodies against viral glycoprotein 0 and by lyso­somotropic drugs. J. Gen. Yirol., 6, 2781-2785.

Lebon, P., Commoy-Chevalier, M.J., Robert-Galliot, B. &Chany, C. (1982a), Different mechanisms for exand Pinterferon induction. Virology, 119,504-507.

Lebon, P., Daffos, F., Checoury, A., Grangeot-Keros, L.,Forestier, F. & Toublanc, J.E. (1985), Presence of anacid-labile alpha-interferon in sera from fetuses andchildren with congenital rubella. J. Clin. Microbiol.,21, 775-778.

Lebon, P., Girard, S., Thepot, F. & Chany, Ch. (1982b),The presence of a-interferon in human amnioticfluid. J. Gen. Virol., 59, 393-396.

Nowacki, W. & Charley, B. (1993), Enrichment of coron­avirus-induced interferon-producing blood leuko­cytes increases the interferon yield per cell: a studywith pig leukocytes. Res. Immunol., 144, 111-120.

Riffault, S., Eloranta, M.-L., Carrat, Ch., Sandberg, K.,Charley, B. & AIm, G. (1996), Herpes simplex virusinduces appearance of interferon-o/S producing cellsand partially interferon-c/p dependent accumulationof leukocytes in murine regional lymph nodes. J.Interferon Cytok: Res., 16, 1007-1014.

Roberts, N.J., Douglas, RG., Simons, R.M. & Diamond,M.E. (1979), Virus induced interferon production byhuman macrophages. J.lmmunol., 123,365-369.

Romero, R., Gomez, R., Galasso, M., Munoz, H., Acosta,L., Yoon, B.H., Svinarich, D. & Cotton, D.B. (1994),Macrophage inflammatory protein-I alpha in termand preterm parturition: effect of microbial invasionof the amniotic cavity. Am. J. Reprod. lmmunol., 32,108-113.

Saksela, E., Virtanen, I., Hovi, T., Secher, D.S. & Cantell,K. (1984), Monocyte is the main producer of humanalpha interferons following Sendai virus induction.Prog. Med. Yirol., 30, 78-86.

Sandberg, K., Matsson, P. & Aim, G.V. (1990), A distinctpopulation of nonphagocytic and CD4 + null lympho­cytes produce interferon-a after stimulation byherpes simplex virus-infected cells. J. Immunol., 145,1015-1020.

Splichal, I., Bonneau, M. & Charley, B. (1994), Ontogenyof interferon alpha secreting cells in the porcine fetalhematopoietic organs. lmmunol, Lett., 43, 203-208.

Sterzl, J. & Kovaru, F. (1977), Development of lymphatictissue and immunocompetency in pig foetus andgerm-free piglets. Acta. Vet. Bmo, 46, suppl. 3, 13-53.

Sterzl, J., Rejnek, J. & Tclvnirek, J. (1966), Impermeabil­ity of pig placenta for antibodies. Folia Microbiol.,ll,7-1O.

Svensson, H., Johannisson, A., Nikkila, T., Aim, G.V. &Cederblad, B. (1996), The cell surface phenotype ofhuman natural interferon-a producing cells as deter­mined by flow cytometry. Scand. J. lmmunol., 44,164-172.