J. Comp. Path. 1996 Vol. 115, 373 383

Interferon-y Mediates Long-term Persistent Chlamydia psittaci Infection In Vitro

j . Brown and G. Entrican

Moredun Research Institute, 408 Gilmerton Road, Edinburgh, EH17 7JH, UK

S u m m a r y

The long-term anti-chlamydial effects of recombinant ovine interferon gamma (rOvlFN-y) were studied in ovine ST-6 fibroblasts infected with the $26/3 strain of Chlamydia psittaci. Chlamydial multiplication was assessed by enzyme- linked immunosorbent assay analysis ofsupernate lipopolysaccharide, titration of inclusion-forming units in culture supernates, and enumeration of inclusion bodies in cultured cells at 7-day intervals. Concentrations of 250 and 1000 U/ml of rOvIFN-y resulted in a microbistatic inhibition of C. psittaci growth, which appeared to become microbicidal when rOvIFN-y was maintained in the cultures for 14 days or more. There were no signs of C.psittacimultiplication when cultures were maintained in 25 or 100 U/ml of rOvlFN-y. However, subsequent removal of rOvIFN-y from these cultures resulted in a re- emergence of viable, infectious chlamydiae, which eventually killed all the fibroblasts. This re-emergence was more rapid in cultures initially treated with 25 U/ml of rOvIFN-y than in those treated with 100 U/ml.

�9 1996 W.B. Saunders Company Limited

Introduction

Chlamydia psittaci is an obligate intracellular pathogen which infects a wide variety of hosts, including man. The capacity of certain strains of C. psittaci to cause enzootic abortion of ewes (EAE) has a major economic and welfare impact on the agricultural industry. In addition, primary infection in pregnant women may be life-threatening (Aitken, 1991). In C. psittaci infection of ewes, the pathogen enters a persistent subclinical or latent state, which can be maintained for several months (McCafferty, 1990). It emerges from the latent phase during the first pregnancy after infection and colonizes the placenta, resulting in abortion (McEwen et al., 1951). Mild immunosuppression associated with pregnancy may be a factor in reactivation of the organism. There is no evidence in the literature to support the existence of a similar latency in human infections with EAE strains of C. psittaci. However, another species, Chlamydia trachomatis, is known to persist in human ocular and urogenital infections (Beatty et al., 1993, 1995).

Cytokines can influence the outcome of chlamydial infections. In particular, interest has focused on interferon-gamma (IFN-y), which has been shown to exert anti-chlamydial effects in cultures of a number of cell lines from a variety

Correspondence to: G. Entrican.

0021-9975/96/080373+ 11 $12.00/0 �9 1996 W.B. Saunders Company Limited

374 J. Brown and G. Entrican

of species (Byrne and Krueger, 1983; Rothermel et al., 1986; Byrne et al., 1989; Graham et al., 1995) and plays a role in the resolution of chlamydial infections in vivo (Williams et al., 1988; Zhong et al., 1988; McCafferty et al., 1994). However, IFN-y does not always eradicate Chlamydia spp. from culture and may even assist in maintaining persistent infection in vitro. The life cycle of chlamydiae has two distinct stages: an infectious extracellular elementary body (EB), and a non-infectious, metabolically active, intracellular reticulate body (RB) (Levitt and Barol, 1987). Treatment of HeLa 229 cells with low concentrations of IFN-y shortly after infection with C. trachomatis serovar A / H a r - 13 resulted in the formation of atypical RBs that failed to differentiate into EBs, and viable infectious chlamydiae were recovered after removal of IFN-y from the culture system (Beatty et al., 1993). More recent in-vitro studies indicate that this IFN-y-mediated persistence can be maintained for up to one month (Beatty et al., 1995).

Persistent, subclinical C. psittaci infection in non-pregnant sheep is a major feature of EAE, but nothing is known about the immune mechanisms associated with this phenomenon. Because chlamydiae cause different forms of disease in different hosts, the behaviour of a particular strain cannot always be predicted from the available literature. IFN-y can restrict C. psittaci mul- tiplication in ovine cells in the short term (Graham et al., 1995) and it is therefore important to define its role, if any, in maintaining the persistent state.

M a t e r i a l s a n d M e t h o d s

Cells and C. psittaci Stocks

The ovine ST-6 fibroblast cell line was derived from a naturally occurring adeno- carcinoma of the small intestine (Norval et al., 1981). The ceils were propagated in Iscove's modified Dulbecco's medium (IMDM; Gibco, Paisley, UK) supplemented with 5% fetal bovine serum (FBS; Boehringer Mannheim, GmbH, Germany) at 37~ in a humidified 5% CO2 environment, unless stated otherwise. ST-6 cultures were passaged by washing the confluent monolayers with warm phosphate-bufl~red saline, removing the cells with trypsin/EDTA (Gibco, Paisley, UK) and reseeding into fresh flasks. These cells were used to generate stocks of C psittaci EBs, titrate infectious EBs, and also measure the bioactivity of recombinant ovine IFN-y (rOvIFN-y). All cultures and experiments described in this paper were performed with antibiotic-free media.

Tissue culture-derived stocks of the $26/3 ovine abortion strain of C. psittaci (McClenagham et al., 1984) were obtained by inoculating sub-confluent monolayers of ST-6 fibroblasts with EBs, as previously described (Graham et al., 1995).

Infection of ST-6 Cells and Treatment with rOvIFN-7

ST-6 cells were initially seeded into vented tissue culture flasks (25 c m 2 surface area) (Costar, High Wycombe, UK) at a density of 3 • 10 ~ cells/flask in 6 ml of IMDM containing FBS 5%. When the monolayers became sub-confluent the medium was removed and the cells were infected with EBs at a multiplicity of infection of 1 EB/ 10 cells. Uninfected controls were incubated with medium alone. After 3 h, the liquid phase was removed and replaced by medium, either alone or containing rOvIFN-y at concentrations ranging from 5 to 1000U/ml. The rOvIFN-y was produced in Chinese hamster ovary cells (Graham et al., 1995) and its bioactivity on ST-6 cells was determined as previously described (Entrican et al., 1992). The cells were cultured

Persistent C. ps i t tac i Infection In Vitro 375

in the rOvIFN-y for 1 week and then passaged as described above. For each concentration of IFN-7 the cells were passaged into two flasks at a density of 3 x 105 cells/flask. The cells in one flask were maintained in the IFN-7 and those in the other flask were cultured without IFN-y. The following week, the cells from each flask that had been treated with IFN-7 were passaged into two flasks and treated exactly as described above. This pattern was continued on a weekly basis, resulting in the generation of cultures of infected cells which had been exposed to IFN-7 for various lengths of time. Those cells which were no longer being treated with IFN-y were passaged routinely under normal conditions. The cultures were maintained for the duration of the experiment or until lysed as a result of chlamydial growth. Samples of culture supernates were removed, clarified by centrifugation at 400g and then analysed for lipopolysaccharide (LPS) content and infectious EBs.

Measurement of C. psittaci Multiplication

The LPS concentrations in the culture supernates were determined by an enzyme- linked immunosorbent assay (ELISA), incorporating two monoclonal antibodies pro- duced against C. psittaci LPS. Supernates were tested in triplicate by the ELISA, as fully described elsewhere (Graham et al., 1995).

The infectivity of C. psittaci stocks and culture supernates was determined by titration on confluent monolayers of ST-6 fibroblasts in 96-well flat-bottomed microtitre plates (Costar, High Wycombe, UK). Briefly, 100-gl volumes of 10-fold dilutions in IMDM (supplemented with FBS 2%) of culture supernates were added to duplicate wells; the plates were then incubated for 3 days, after which the monolayers were fixed with methanol and stained for 15 min in Giemsa (BDH Chemicals, Poole, UK). The number of inclusion-forming units (IFU) in each sample was then determined by counting the inclusion bodies.

Cell suspensions from cultures at the time of passage were centrifuged onto glass slides at 600 rpm for 5 rain with a cytocentrifuge (Cytospin 3; Shandon Scientific, Runcorn, UK), air dried, fixed with methanol and stained in Giemsa for 15 min. Slides were then dehydrated in a graded ethanol series, cleared in xylene and mounted with synthetic mountant (Shandon Scientific). The percentage of cells containing inclusion bodies was assessed by counting an average of 300 cells/sample.

R e s u l t s

Cul ture supernates of r O v I F N - y - t r e a t e d and un t rea ted cells were analysed for chlamydial LPS and infectious EBs 7 days after infection (Fig. 1). There was a reduction in chlamydial LPS in cultures which were incubated in rOvIFN-7 (all concentrat ions) , as c o m p a r e d with un t rea ted infected controls. T h e con- centra t ions of LPS in infected cultures incuba ted with r O v I F N - y were similar to those in uninfec ted controls. Infectious ch lamydiae were found only in the supernates f rom cells which had been cul tured ei ther in the absence of r O v I F N - y or in the presence o f no more than 5 U / m l . Inclusions were not found in any of the infected ST-6 cells themselves at the t ime of passage.

T h e growth of the chlamydiae after 21 days in cul ture is shown in Fig. 2. W h e n r O v I F N - 7 was r e m o v e d on day 7, there was a dose -dependen t increase in supernate LPS concent ra t ions in all cultures (Fig. 2a). W h e n r O v I F N - y was r e move d after 14 days, there was a small increase in LPS in cultures t rea ted with 25 U / m l , but not in those t rea ted with 1000 ,250 or 100 U / m l . Chlamydia l LPS was found at increased concent ra t ions in cultures incuba ted in 5 U / m l r O v I F N - y as c o m p a r e d with uninfec ted controls, but at r educed concent ra t ions

376

0.8

j . Brown and G. Entrican

0.4

�9

0 . 6 - -

0.2

0.0 1000 250 100 25 5 0

C o n c e n t r a t i o n of IFN-~/(U/ml)

Fig. 1. ST-6 cells infected with C. psittaci (V]) and uninfected controls (IN), incubated in the presence of various concentrations of rOvIFN-T. Culture supernates were analysed for the presence of chlamydial LPS by ELISA 7 days after infection. Values represent means of triplicate measurements, expressed as optical density (OD) at a wavelength of 492 nm. Bars representing standard error of the mean are present at all data points, but in most cases are too small to be visible. Supernates were also analysed for infectious elementary bodies, which were found in only two samples: *denotes 10 inclusion forming uni ts /ml ('+5) and **denotes 305 inclusion forming uni ts /ml (_+ 114.5).

as compared with infected untreated control cultures maintained for 21 days. LPS concentrations in the supernates of infected ST-6 cells cultured in the presence of 1000, 250, 100 or 25 U / m l rOvIFN-~{ for 21 days did not differ from those of uninfected controls. These patterns are reflected in the results of titration analysis (Fig. 2b) and enumeration of inclusion bodies within cultured cells (Fig. 2c). Viable EBs were recovered at this stage from the supernates of cells cultured in the presence of 100 U / m l of rOvIFN- 7 for the first 14 days, despite the absence of supernate LPS or inclusions in cells at the time of passage. The chlamydiae had lysed infected untreated control cultures by 21 days, precluding further analysis of these cells.

Incubation of cells with rOvIFN-y was terminated 63 days after infection. At 70 days all cells previously cultured in the presence of rOvIFN-7 at 1000,

P e r s i s t e n t C. ps i t tac i I n f e c t i o n In Vitro 3 7 7

~_ 10 9 2.5 ~ 10 s ~ (b)

lO~

1.o o

0.5 lo i - l o l -

0 ~ 1 ~ I 0 2 1 0 0 5 1000 250 0

Concentrat ion of IFN-y (U/ml)

Fig. 2.

+ae~

~ D + a

1000 250 100 25 5

Concentration of IFN- 7 (U/ml)

Chlamydial growth, assessed at 21 days after infection by: (a) LPS-ELISA analysis of supernates; values represent means of triplicate measurements, expressed as optical density (OD) at a wavelength of 492 nm; (b) titration of inclusion-forming units in supernates; (c) percentage of cells containing inclusion bodies in cytocentrifuge preparations of cell suspensions. IFN-y was maintained in cultures for 21 days ([]), for the first 14 days (NI) or for the first 7 days (mR). Infected control cultures received only IMDM with FBS 5% ([-1). Uninfected control cultures were not inoculated with C. psittaci elementary bodies (~11). Bars representing standard error of the mean are present at all data points, but in most cases are too small to be visible.

250 or 100U /ml exhibited concentrations of supernate LPS equivalent to those of the uninfected controls (Fig. 3). Cultures treated with 25 U / m l showed a small increase in supernate LPS when the rOvIFN-y was removed for the final 7 days of culture and a much larger increase when it was absent for 14 days. Cell monolayers were completely lysed when rOvIFN-y was absent for more than 14 days from cultures treated with 25 U/ml . Supernate LPS values showed a correlation with the number of cells containing inclusion bodies, and also with the number of IFUs in the supernates of cultures (Table 1). Infectious EBs and low numbers of inclusions were found in cultures treated with 100 U / m l of rOvIFN-7 for 35 days of a 70-day period (Table 1), but chlamydial LPS was absent.

By 91 days after infection, cultures treated with 1000 or 2 5 0 U / m l of rOvIFN-y for 14 days or more at the start of the experiment had only baseline concentrations of LPS (Fig. 4) and no viable EBs in the supernate (Table 2). Inclusions were not visible in cytocentrifuge preparations. All supernates from cells exposed to 100 U / m l of rOvIFN-7 had increased LPS concentrations (Fig. 4) and viable EBs (Table 2) after rOvIFN-y was removed from cultures,

378

2.5

J. Brown and G. Entrican

2.0

1.5

r

�9 1.0

0.5

Fig. 3.

0 ~ 1000 250 100 25 250 0

C o n c e n t r a t i o n o f IFN-T (U/ml)

Chlamydial growth, in ST-6 cells assessed by LPS-ELISA analysis of culture supernates 70 days after infection. Values represent means of triplicate measurements, expressed as optical density (OD) at a wavelength of 492 nm. Cultures were maintained in IFN- T for the first 63 days ( . ) , 49 days ([-]), 35 days ([]), 28 days ( l ) , 21 days (m) or 14 days ([~). There are no data for cultures treated with 250 U / m l for 21 days or 100 U / m l for 49 days. In addition, cultures treated with 1000 U / m l for 49 days or more were not continued. LPS concentrations in uninfected cultures are also shown (Ira). Bars representing standard error of the mean are present at all data points, but in most cases are too small to he visible.

which by that time had become lysed. Cultures originally incubated with rOvIFN-T at concentrations of 1000 or 250 U / m l for 14 days or more were maintained for up to 112 days after infection without any further evidence of recrudescence of the organism (data not shown).

D i s c u s s i o n

The phenomenon of C. psittaci persistence in relation to EAE has been recognized for half a century (Stamp et al., 1950), but the factors that control it remain largely unresolved. Persistence is most commonly established in ewes infected for the first time when non-pregnant (McEwen et al., 1951) and may even be a result of primary infection during the perinatal period (Wilsmore et al., 1984).

P e r s i s t e n t C. ps i t t ac i I n f e c t i o n In Vitro

Table 1 A s s e s s m e n t o f c h l a m y d i a l g r o w t h in 70-day cu l tures

379

Concentration Duration of interferon Percentage of Mean inclusion Standard of interferon gamma treatment cells containing forming error gamma (U/ml) (days) inclusions units/tH

100 63 0 0 0 100 35 0.2 6.3 8 100 28 0 0 0 100 21 0 0 0 25 63 6.41 0.2 0.005 25 49 71.97 136 7.5

*250 63 0 0 0 *0 - - 0 0 0

Chlamydial growth assessed by titration of inclusion-forming units in culture supernates and the percentage of cells containing inclusions in cytocentrifuge preparations of cell suspensions 70 days after infection. No inclusion-forming units or inclusion bodies were found in any of the cultures incubated with rOvIFN.y at concentrations of 250 or 1000 U/ml . * Uninfected controls.

2.5

2 . 0 - -

1 . 5 -

r

�9 1 . 0 -

0 . 5 -

Fig. 4.

1000 250 100 250 0

C o n c e n t r a t i o n of IFN-~ (U/ml)

Chlamydial growth in ST-6 cells assessed by LPS-ELISA analysis of culture supernates 91 days after infection. Values represent means of triplicate measurements, expressed as optical density (OD) at a wavelength of 492 nm. Cultures were maintained in IFN-y for the first 63 days (m), 49 days ([]), 35 days (IN), 28 days ([]), 21 days (kN) and 14 days (tN). There are no data for cultures treated with 250 U / m l for 21 days or 100 U / m l for 49 days. In addition, cultures treated with 1000 U / m l for 49 days or more were not continued. LPS concentrations in uninfected cultures are also shown ([[ml). Bars representing standard error of the mean are present at all data points, but in most cases are too small to be visible.

3 8 0 J . B r o w n a n d G. E n t r i c a n

Tab le 2 A s s e s s m e n t o f c h l a m y d i a l g r o w t h in 91-day c u l t u r e s

Concentration Duration of inteoreron Mean inclusion Standard of interferon gamma treatment forming error gamma (U/ml) (days) units/ #l

100 63 3480 895 100 35 1160 118 100 28 467 58

*250 63 0 0 *0 - - 0 0

Chlamydial growth assessed by titration of inclusion-forming units in culture supernates 91 days after infection. No inclusion-forming units or inclusion bodies were found in any of the cultures incubated with IFN- 7 at concentrations of 250 or 1000 U/ml . * Uninfected controls.

Experimental infection of pregnant ewes has highlighted the importance of delayed type hypersensitivity (DTH) on resistance to EAE (Dawson et al., 1986). IFN-7, a cytokine with an important role in DTH responses, has been shown to assist in controlling chlamydial infections in vivo. For example, mice experimentally infected with C. psittaci or C. trachomatis have exacerbated lesions and harbour increased numbers of chlamydiae when treated with monoclonal antibodies that neutralize IFN-y (Zhong et al., 1989; McCafferty et al., 1994). For IFN-y to play a direct role in maintaining persistence of C. psittaci in relation to EAE, its effects must be non-lethal to the pathogen and must also be reversible. The present study showed that concentrations of rOvIFN-y in the range of 25 to 100 U / m l had a microbistatic effect on C. psittaci growth in ST-6 cells, regardless of the duration of exposure. However, at 250 or 1000 U/ml , the effect, although initially microbistatic, appeared to become microbicidal when the cultures were maintained in rOvIFN-y for 14 days or more. IFN-y might, therefore, play a role as a direct effector mechanism in the generation of solid immunity, as well as maintaining persistence in sheep. The amount of IFN-7 produced is directly related to antigenic stimulation, and this may be a factor in the immunopathology of EAE. Low-level cytokine production as a result of antigen pulsing might maintain the persistent state, whereas the solid immunity that follows abortion might be a result of a massive antigenic stimulus associated with multiplication of C. psittaci in the placenta. The concentration of IFN-7 that appeared to eradicate the organism from cultures (250 U/ml) has been found in vivo in post-abortion sheep after challenge with live C. psittaci (Graham et al., 1995).

These experiments provide information on the role that IFN-y can play in maintaining persistence, but do not explain how C. psittaci can become reactivated during pregnancy. The concept that pregnancy is associated with generalized immunosuppression is not new (Talwar, 1980; Lloyd, 1983). However, recent findings indicate that the situation may be more complex than originally anticipated. Cytokines appear to fall into two groups with respect to their effects on the outcome of pregnancy: beneficial cytokines, whose effects are characteristic of the TH2 subset; and detrimental cytokines, whose effects fit the profile of the TH1 subset (Chaouat et al., 1990; Wegmann

Persistent C. ps i t tac i Infection In Vitro 381

et al., 1993). It therefore seems likely that successful pregnancy is dependent not on generalized immunosuppression, but on biasing the maternal immune system towards a TH2 type of response. TH2 cytokines, particularly IL-4 and IL-10, antagonize and "downregulate" TH 1 cytokines. It is therefore theoretically possible that the downregulation of cell-mediated immunity would allow chlamydiae that had previously been subject to IFN-y microbistasis to multiply and colonize the placenta, resulting in fetal damage and subsequent abortion.

Perhaps the most striking difference between the persistent C. psittaci infection of ovine cells described here and persistent C. trachomatis infection of human cells is the nature of the latent organisms and the pathology of the diseases they cause. Persistent C. trachomatis infection is associated with chronic im- munopathological effects, thought to arise as a result of changes in the relative expression of potentially protective and immunopathogenetic chlamydial an- tigens (Beatty et al., 1993). In contrast, persistent C. psittaci infection in the non-pregnant ewe appears to be without pathological consequence (McCafferty, 1990). Persistent C. trachomatis (serovar A) infection has been established in human HeLa 229 cells by treating cultures with low con- centrations ofIFN-y (Beatty et al., t993). The inclusion bodies that subsequently developed contained large aberrant RBs, which failed to differentiate into infectious EBs. When IFN-7 was removed from the culture system for more than 48 h, viable EBs were recovered. More recent work by the same group demonstrated that C. trachomatis could be maintained in this persistent state for up to 30 days of culture (Beatty et al., 1995). In contrast, we were unable to demonstrate inclusion bodies by Giemsa staining in ST-6 cells infected with the $26/3 strain of C. psittaci when it was treated with rOvIFN-7 at concentrations of 1000, 250, 100 or 25U/ml. However, inclusion bodies developed shortly after rOvIFN- 7 was removed from cultures receiving 100 or 25 U/ml. These data may help to explain why C. psittaci is difficult to detect in vivo in the persistent state in sheep (Huang et al., 1990).

The study indicated that C. psittaci could persist in ovine cells in vitro for up to 63 days in the presence of moderately low concentrations of rOvIFN-y. This has important implications for the persistence of C. psittaci in ewes infected when non-pregnant. In addition, this work has provided an in-vitro model of long-term persistent infection which can now be used to examine other features of the disease, including the action of pregnancy hormones on immune function.

Acknowledgments

The authors thank Gareth Jones and Morag Livingstone for the LPS-ELISA reagents. This work was funded by the Scottish Office Agriculture, Environment and Fisheries Department.

References

Aitken, I. D. (1991). Human infection with Chlarnydia psittaci. In: Diseases of Sheep, 2nd Edit., W. B. Martin and I. D. Aitken, Eds, Blackwell Scientific Publications, Oxford, pp. 369 370.

382 J. Brown and G. Entrican

Beatty, W. L., Byrne, G. I. and Morrison, R. P. (1993). Morphologic and antigenic characterisation of interferon 3,-mediated persistent Chlamydia trachomatis infection in vitro. Proceedings of the National Academy of Science USA, 90, 3998-4002.

Beatty, W. L., Morrison, R. P. and Byrne, G. I. (1995). Reactivation of persistent Chlamydia trachomatis infection in cell culture. Infection and Immunity, 63, 199-205.

Byrne, G: I., Carlin, J. M., Merkert, T. P. and Arter, D. L. (1989). Long-term effects of gamma interferon on chlamydia-infected host cells: microbicidal activity follows microbistasis. Infection and Immunity, 57, 1318-1320.

Byrne, G. I. and Krueger, D. A. (1983). Lymphokine-mediated inhibition of Chlamydia replication in mouse fibroblasts is neutralised by anti-gamma interferon im- munoglobulin. Infection and Immunity, 42, 1152-1158.

Chaouat, G., Menu, E., Clarke, D. A., Dy, M., Minkowski, M. and Wegmann, T. G. (1990). Control of fetal survival in CBA x DBA/2 mice by lymphokine therapy. Journal of Reproduction and Fertility, 89, 447-458.

Dawson, M., Zaghloul, A. and Wilsmore, A. J. (1986). Ovine enzootic abortion: experimental studies of immune responses. Research in Veterinary Science, 40, 59-64.

Entrican, G., McInnes, C. J., Rothel, J. S. and Haig, D. M. (1992). Kinetics of ovine interferon-gamma production: detection of mRNA and characterisation of biological activity. Veterinary Immunology and Immunopathology, 33, 171-178.

Graham, S. P.,Jones, G. E., Maclean, M., Livingstone, M. and Entrican, G. (1995). Recombinant ovine interferon gamma inhibits the multiplication of Chlamydia psittaci in ovine cells. Journal of Comparative Pathology, 119, 185-195.

Huang, H. S., Buxton, D. and Anderson, I. E. (1990). The ovine immune response to Chlamydia psittaci; histopathology of the lymph node. Journal of Comparative Pathology, 102, 89-97.

Levitt, D. and Barol, J. (1987). The immunobiology of Chlamydia. Immunology Today, 8, 246-251.

Lloyd, S. (1983). Effect of pregnancy and lactation upon infection. Veterinary Immunology and Immunopathology, 4, 153-176.

McCafferty, M. C. (1990). Immunity to Chlamydia psittaci with particular reference to sheep. Veterinary Microbiology, 25, 87 99.

McCafferty, M. C., Maley, S. W., Entrican, G. and Buxton, D. (1994). The importance of interferon-y in an early infection of Chlamydia psittaci in mice. Immunology, 81, 631-636.

McClenaghan, M., Herring, A.J. and Aitken, I. D. (1984). Comparison of Chlamydia psittaci isolates by DNA restriction endonuclease analysis. Infectio n and Immunity, 45, 384-389.

McEwen, A. D., Littlejohn, A. I. and Foggie, A. (1951). Enzootic abortion of ewes: some aspects of infection and resistance. Veterinary Record, 63, 489-492.

Norval, M., Head, K. W., Else, R. W., Hart, H. and Neill, W. A. (1981). Growth in culture of adenocarcinoma cells from the small intestine of sheep. British Journal of Experimental Pathology, 62, 270-282.

Rothermel, C. D., Rubin, B. Y., Jaffe, E. A. and Murray, H. W. (1986). Oxygen- independent inhibition of intracellular Chlamydia psittaci growth by human mono- cytes and interferon-y-activated macrophages. Journal of Immunology, 137, 689-692.

Stamp, J. T., McEwen, A. D., Watt, J. A. A. and Nisbet, D. I. (1950). Enzootic abortion of ewes: transmission of the disease. Veterinary Record, 62, 251-254.

Talwar, G. P. (1980). Immunobiology of pregnancy. Current Topics in Immunology, 13, 23-39.

Wegmann, T. G., Lin, H., Guilbert, L. and Mosmann, T. R. (1993). Bidirectional cytokine interactions in the maternal-fetal relationship: is successful pregnancy a TH2 phenomenon? Immunology Today, 14, 353-356.

Williams, D. M., Byrne, G. I., Grubbs, B., Marshal, T.J . and Schachter, J. (1988). Role in vivo for gamma interferon in control of pneumonia caused by Chlamydia trachomatis in mice. Infection and Immunity, 56, 3004-3006.

Wilsmore, A. J., Parsons, V. and Dawson, M. (1984). Experiments to demonstrate

Pers i s tent C. ps i t tae i Infect ion In Vitro 383

routes of transmission of ovine enzootic abortion. British Veterinary Journal, 140, 380-391.

Zhong, G., Peterson, E. M., Czarniecki, C. W. and De La Maza, L. M. (1988). Recombinant murine gamma interferon inhibits Chlarnydia trachomatis serovar L 1 in vivo. Infection and Immunity, 56, 283-286.

Zhong, G., Peterson, E. M., Czarniecki, C. W., Schreiber, R. D. and De La Maza, L. M. (1989). Role of endogenous gamma interferon in host defence against Chlamydia trachomatis infections. Infection and Immunity, 57, 152-157.

Received, April 25th, 1996] Accepted, July 5th, 1996 J