Journal of General Virology (1992), 73, 375-381.

375

Printed in Great Britain

Protection of mice from lethal influenza by defective interfering virus: T cell responses L. McLain, D. J. Morgan and N. J. Dimmock* Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, U.K.

The immune-mediated lethal influenza in C3H/He-mg (H-2 k) mice infected with A/WSN influenza virus (H1N1) was investigated. A primary class I major histocompatibility complex-restricted, CD8 ÷ cytotoxic T lymphocyte (CTL) response was found in the lungs with a peak activity at 5 days post-infection. Monoclonal antibody depletion in vivo showed that a lethal CD8 + cell response as well as a lethal CD4 ÷ response was generated during infection. Mice survived infection only if both CD8 ÷ and CD4 ÷ cells were depleted. Mice infected with the same dose of virus, but treated with

defective interfering (DI) A / W S N virus develop only a transient sub-lethal respiratory disease even though multiplication of virus in the lungs is undiminished, and we have shown here that this correlates with a reduction in the local CTL response. The mechanism by which DI virus beneficially modulates the immune response is discussed; it is proposed that there is classical, but cell type-specific DI virus interference in lymphocytes but not in the cells of the lung in which virus multiplies productively.

Introduction

neuramidinase (NA) antibody in the lung is made in considerable amounts. Treatment of infected mice with DI virus allows partial expression of the anti-HA antibody repertoire, notably haemagglutination-inhibiting antibodies which are unusual in not being neutralizing, and also non-haemagglutination-inhibiting antibodies, which resemble the T cell receptor in recognizing a non-native form of the HA only in the context of major histocompatibility complex (MHC) class I (Dimmock et al., 1986; McLain & Dimmock, 1989, 1991). The latter, when adoptively transferred, are protective in very low (pg) amounts against a lethal infection (McLain & Dimmock, 1989, 1991). In this report we study the T lymphocyte responses in the lung.

Defective interfering (DI) viruses have genomes which are a deleted form of that of the infectious or standard virus; they appear to arise by an accident of replication and are dependent for their multiplication upon coinfection with standard virus. Under appropriate conditions they also interfere with the multiplication of standard virus in vitro (Perrault, 1981 ; Holland, 1990). In vivo, DI viruses can ameliorate the severity of infection and protect against lethal disease (Barrett & Dimmock, 1986; Roux et al., 1991). One such example is the protection conferred by DI virus against lethal type A influenza pneumonia in the mouse (Dimmock et al., 1986). However, the peculiarity of this system is that, unlike other virus infections in the mouse in which there is DI virus-mediated protection (Barrett & Dimmock, 1986), multiplication of virus in the lung, the main target site of virus infection, is not affected. Since influenza in the mouse is an immune T cell-mediated disease (Cate & Mold, 1975), it has been argued that DI influenza virus in some unknown way reduces the extent of the immune pathology (Dimmock et al., 1986). This paper continues the investigation into which aspects of the immune response are affected. Earlier studies on the influenza virus-infected mouse have shown that the antibody response to the viral haemagglutinin (HA) protein in the lung is completely suppressed by influenza virus, whereas the serum antibody response is unaffected. However, antiviral 0001-0590 © 1992 SGM

Methods Viruses. Influenza viruses A/WSN (H1NI), A/PR/8/34 (H1N1) and A/Japan/305/57 (H2N2) were grown by inoculating 103 p.f.u, into the allantoic cavity of a hen's egg and incubating for 48 h at 33 °C. Virus was quantified by plaque assay on chick embryo fibroblast monolayers and by HA assay. DI WSN virus was prepared by three sequential high multiplicity passages of standard virus in pro, starting with 109 p.f.u./egg. Before inoculation into mice, DI virus was pelleted from allantoic fluid and resuspended in PBS at a concentration of 104 HA units (HAU)/ml. Since DI and infectious influenza virus particles cannot be physically separated, infectivity was removed by u.v. irradiation (Dimmock & Kennedy, 1978). This is possible because infectivity has a target size about 20-fold greater than that of the interfering activity (data not shown). Irradiated DI virus was divided into two aliquots, one of which was incubated with 0.01% flpropiolactone (BPL; Sigma) for 10 h at 4 °C (Barrett et al., 1984) to

376

L. McLain, D. J. Morgan and N. J. Dimmock

inactivate interfering activity. BPL was removed by overnight dialysis against excess PBS. Mice under light ether anaesthesia were inoculated intranasally with 200 HAU DI virus or 200 HAU BPL-inactivated DI virus 2 h before infection, and then were co-inoculated with 10 LDs0 (4000p.f.u.) infectious WSN virus and 200 HAU of DI virus or 200 HAU BPL-inactivated DI virus. BPL-inactivated virus retains full HA and NA activity, and was used to control for any non-specific effects of the DI virus preparation. Vaccinia virus recombinants each expressing an influenza virus gene were obtained from Dr G. L. Smith (University of Oxford, U.K.) or Dr R. W. Compans (vacc-HA; Birmingham, Ala., U.S.A.) and were grown in HeLa cells (Table 1).

Mice. C3H/He-mg and BALB/c mice were originally obtained from B&K Universal. Both sexes proved equally susceptible to WSN virus and were therefore used interchangeably at 5 weeks of age. All experiments followed the guidelines laid down by the U.K. Coordinating Committee for Cancer Research. Isolation of lymphocytes from mouse lungs. Mice were killed by cervical dislocation. Lungs were removed aseptically and washed in RPMI 1640 medium. Lung tissue was then cut into small pieces, pressed through a stainless steel mesh into RPMI 1640 containing 10% (v/v) heat-inactivated foetal calf serum (FCS) and cell debris removed by brief low speed centrifugation. Lymphocytes were isolated by centrifugation on Ficoll-Paque (Pharmacia) for 45 min at 400g. Adherent cells were removed by passage through nylon wool. The majority of the cell preparation were T cells. Cell-mediated eytotoxicity assays. Suspensions of C3H1OT½ or BALB/c 3T3 cells were labelled by incubation for 1 h at 37 °C in medium containing 100p_Ci [51Cr]sodium chromate (Amersham)/ 2 x 105 cells. These were then seeded in the radioactive medium at 2 × 104 cells/well in a 96-well microtitre tray. After overnight incubation, confluent monolayers were washed and inoculated with A/WSN virus or vaccinia virus at 10 p.f.u./cell, and incubated for 2 h and 9 h respectively. Primary lung effector lymphocytes were added at various ratios and incubated at 37 °C for 6 h. Lysis of target cells was measured as specific 51Cr release. Depletion of CD8+and CD4÷cells in vivo. Rat hybridoma cells YTS 169.4 (anti-mouse CD8) and YTS 191.1 (anti-mouse CD4) (Cobbold et al., 1984) were obtained from the ECACC (Porton Down, Wiltshire, U.K.), and maintained in RPMI 1640 medium supplemented with 10% FCS. Antibody-rich ascitic fluid was produced from pristane-primed MFl-nu/nu mice (Harlan Olac). Naive C3H/He-mg mice were treated with two intravenous (i.v.) injections containing 100 ~tgpurified antibody 3 days and 1 day prior to infection. This resulted in the elimination of 99 % of CD8 ÷ and CD4 ÷ T cells from the mouse, as determined by immunofluorescence on peripheral blood smears.

Table 1. Recombinant vacc&ia viruses Influenza virus A/WSN A/CAM/46 A/PR8 A/PR8 A/PR8 A/PR8 A/PR8 A/PR8 A/Udorn/72 A/Udorn/72

H 1N 1 H1N1 H1N1 H 1N 1

H1N1 H 1N 1 H1N1 HIN1 H2N2 H2N2

Gene

Recombinant vaccinia virus

HA NA NP PA PB1 PB2 M1 NSI M2 NS2

vacc-HA vacc-NA vacc-NP vacc-PA vacc-PBl vacc-PB2 vacc-M1 vacc-NSl vacc-M2 vacc-NS2

Results T h e e x p e r i m e n t a l system is t h a t used by D i m m o c k et al. (1986). T h e r e are two e x p e r i m e n t a l groups o f mice, b o t h o f w h i c h are given a lethal dose o f influenza A / W S N virus intranasally. I n a d d i t i o n one g r o u p is g i v e n active D I A / W S N virus free o f i n f e c t i v i t y a n d the o t h e r the s a m e D I virus p r e p a r a t i o n t r e a t e d w i t h B P L , w h i c h i n a c t i v a t e s the i n t e r f e r i n g a c t i v i t y w i t h o u t affecting the H A or N A functions o f the e n v e l o p e glycoproteins. I n all respects, the course o f infection a n d i m m u n e responses in m i c e given the i n a c t i v a t e d D I virus a p p e a r s to be i d e n t i c a l to t h a t o f m i c e g i v e n infectious virus o n l y ; m i c e die 7 to 9 days after infection with t h e i r lungs c o m p l e t e l y c o n s o l i d a t e d . M i c e t r e a t e d w i t h active D I W S N virus show signs o f r e s p i r a t o r y d i s e a s e a n d d e v e l o p consolidation e x t e n d i n g to a b o u t 40~o o f the lung, but n e a r l y all ( > 8 0 % ) r e c o v e r ( D i m m o c k et al., 1986).

Generation o f an influenza virus-specific cytotoxic T lymphocyte (CTL) response in the lungs o f C3H/He-mg mice L y m p h o c y t e s purified from the lungs o f b o t h groups o f infected m i c e lysed 51Cr-labelled W S N virus-infected syngeneic cells in a d o s e - d e p e n d e n t (Fig. 1 a) a n d M H C class I - d e p e n d e n t (Fig. 1 b) m a n n e r . T h i s C T L response was d e t e c t a b l e 3 d a y s p o s t - i n f e c t i o n (p.i.), p e a k e d at 5 d a y s p.i. a n d was still very m u c h p r e s e n t at t h e t i m e o f d e a t h (Fig. 1 c). A s i m i l a r p r i m a r y C T L response was seen w h e n only infectious virus was a d m i n i s t e r e d to mice. B P L - i n a c t i v a t e d D I virus a l o n e s t i m u l a t e d no d e t e c t a b l e response ( d a t a n o t shown). T h e response in infected m i c e t r e a t e d w i t h active D I virus was s i m i l a r e x c e p t t h a t it was a p p r o x i m a t e l y twofold lower at e a c h t i m e p o i n t (Fig. 1 c). Since b y all c r i t e r i a e x a m i n e d there was no difference in virus m u l t i p l i c a t i o n in the lungs o f i n f e c t e d m i c e ( D i m m o c k et al., 1986), we c o n c l u d e d t h a t the cytotoxic T cell response to the influenza virus infection is i n h i b i t e d in some w a y b y the a d m i n i s t r a t i o n o f D I virus. A d m i n i s t r a t i o n o f D I virus alone s t i m u l a t e d no d e t e c t a b l e C T L r e s p o n s e ( d a t a n o t shown).

Antigen specificity o f the primary C T L response in lungs during influenza virus infection P r i m a r y C T L s isolated f r o m the lungs o f i n f e c t e d m i c e p r o t e c t e d b y D I virus, lethally i n f e c t e d m i c e g i v e n i n a c t i v a t e d D I virus a n d n o n - i n f e c t e d m i c e were tested a g a i n s t i n d i v i d u a l influenza virus a n t i g e n s e x p r e s s e d by r e c o m b i n a n t v a c c i n i a virus vectors in syngeneic L929 cells. O p t i m a l e x p r e s s i o n o f b o t h H A a n d N A b y t a r g e t

T cell responses in D I virus-treated mice

I(a )

I

I

(b) C,H

BALB/c 4

60

40

--

3

_

20 b

3 .~ 40

10 E :T ratio I

I

lll

i4

30 I

I

I

30 20

___.

(

c

)

1-

I _-__-

10 0 1

3

5

7

9

1l

Time p.i. (days) Fig. 1. Generation of M H C class I primary C T L activity in the lungs of C3H/He-mg mice after intranasal inoculation with 10 LDso A / W S N virus, and BPL-inactivated DI W S N virus ( A ) or active DI W S N virus (O). (a) Lysis at different effector :target (E :T) ratios 5 days p.i. (b) M H C class I-restricted killing of syngeneic C3H (H-2 k) W S N virusinfected target cells but not of infected BALB/c (H-2 d) target cells. I to 3 represent C3H mice; effector cells from (1) non-infected mice, (2) from mice infected with W S N virus and active DI W S N virus, (3) from mice infected with W S N virus and BPL-inactivated DI W S N virus and (4) from BALB/c mice infected with W S N virus. (c) Time course of the production of lung CTLs. All mice inoculated with W S N virus and BPL-inactivated DI W S N virus were dead (t) by 9 days p.i. Unless otherwise stated, target cells were 5~Cr-labelled C3H10T ~ cell monolayers infected with A / W S N virus at a multiplicity of 10. Effector cells were isolated 5 days p.i. and used at an E :T ratio of 30 : 1. The bars represent 1 S.D.; no error bar m e a n s that the S.D. falls within the dimensions of the d a t u m point.

cells was demonstrated by haemadsorption of chick red blood cells and reaction with fetuin respectively (data not shown). The CTLs lysed target cells expressing PA, PB 1, PB2, HA, NA, NP, NS1 and NS2, but not M1 and M2, with greatest activity against PB1, HA and NP (selected data shown in Fig. 2). However, it is not possible to compare CTL lysis of different targets quantitatively because the relative efficiency of expression of different influenza virus antigens is not known. Failure to detect CTLs to M l and M2 has to be viewed with some caution because although they are evolutionarily conserved, they originated from other influenza viruses [A/PR8 (H 1N 1) and A/Udorn (H2N2) respectively]. However, others have found that M1 is not recognized by either H-2k or H-2d mice, although it is recognized by human CTLs (Reay et al., 1989; Yewdell & Hackett, 1989). In every case in which there was a positive CTL response, lethally infected mice gave greater lysis than infected mice protected with DI virus, despite the fact that the number of effector cells was normalized and fluorescenceactivated cell sorter analysis showed that the ratio of CD4 +:CD8 + cells was approximately equal in each group (D. J. Morgan & N. J. Dimmock, unpublished data). C T L activity is largely virus strain-specific

Comparison of CTL killing of syngeneic target cells infected with either the homologous strain of virus [A/WSN (H1N1)] or a heterologous virus [A/ Japan/305/57 (H2N2)] showed that the CTL activity in both experimental groups was predominantly strainspecific (Fig. 3). This is unlike the situation reported by others in which the majority of CTLs are cross-reactive (Wraith, 1987; Yewdell & Hackett, 1989).

Table 2. Summary of the effects of D I W S N virus on immunologically mutant infected mice or mice depleted o f CD4 + and~or CD8 + cells using MAbs* C D 4 ÷, CD8 +

Survival 7 to 9 days Clearance of lung virus Consolidation 7 to 9 days Consolidation 12 days

377

C D 4 - , CD8 ÷

C D 4 ÷, C D 8 -

CD4-, CD8-

-DIt

+DI~:

-DI

+DI

-DI

+DI

-DI

+DI

NA§ + + + + NA

+ + + + +

NA + + + + NA

+ + + + + +

NA + + + + NA

+ + + + +

+ ---

+ + --

* Data from this report and D i m m o c k et al. (1986). t - D I , Mice inoculated with 10 LDso W S N virus and BPL-inactivated DI W S N virus. :~ + D I , Mice inoculated with 10 LDso W S N virus and active DI W S N virus. § NA, Not applicable.

L. McLain, D. J. Morgan and N. J. Dimmock

378 50

100

40

m A

I

8060--

3O O

Y. 4 0 -

,-~ 20

20-

10

Time p.i. (days) E:T30

3

Non-infected

30 3

30 3

wt

30 3

30 3

30 3

vacc-NP vacc-HA vacc-PB 1vacc-M 1

Fig. 2. Activity of primary lung CTLs against syngeneic target cells expressing individual influenza virus antigens. CTLs were obtained 5 days p.i. from mock-infected C3H/He-mg mice (l'q), mice infected with I0 LDso A/WSN virus and active DI WSN virus (1~) and mice infected with 10 LDs0 A/WSN virus and BPL-inactivated DI WSN virus (m), and used at the E :T ratio shown. L929 target cells were infected with vaccinia virus recombinants, each expressing a different influenza virus antigen, at a multiplicity of 10 or with vaccinia virus (wt) itself.

(a)

(b) 3 T

60--

4

~

-L

40--

2

20--

2

m

3

4 I 1

r--rFig. 3. Lysis of syngeneic C3H 10T½ target cells infected with A/WSN (H1N1) virus (a) or with a different subtype of influenza A virus, A/Japan/305/57 (H2N2) (b), by lung cytotoxic cells. CTLs were obtained from (1) mock-infected mice, (2) mice infected with I0 LDso A/WSN virus together with DI WSN virus and (3) mice infected with 10 LDs0 A/WSN virus together with BPL-inactivated DI WSN virus. Lysis by cytotoxic cells obtained from mice infected with A/Japan virus is shown in (4). All CTLs were prepared 5 days p.i. and used in an E :T ratio of 30 : 1.

Fig. 4. Treatment of mice in vivo with anti-CD4 and anti-CD8 MAbs. C3H/He-mg mice were injected i.v. with PBS (O), or I00 Ixg anti-CD4 (V), anti-CD8 (A), or anti-CD4 and anti-CD8 (11) MAbs 3 days and 1 day prior to infection with 10 LDso WSN virus and BPL-inactivated DI WSN virus (closed symbols) or WSN virus and active DI WSN virus (open symbols).

Role of CD8 ÷ and CD4 ÷ T cells in the pathogenesis of influenza pneumonia

The role of CD4 ÷ and CD8 ÷ T cells in the pathogenesis of lethal influenza was determined by depletion in vivo prior to infection with the appropriate monoclonal antibodies (MAbs) either alone or together (Table 2). Fig. 4 shows that the course of disease in mice depleted of CD8 ÷ cells was unchanged relative to that of controls. Depletion of CD4 ÷ cells resulted in the same overall mortality, although signs of infection appeared 1 to 2 days earlier and deaths commenced earlier. However, with simultaneous depletion of CD8 + and CD4 ÷ cells, the majority (60 %) of mice survived. This group had very high levels of lung infectivity at 12 days p.i., a failure in clearance which parallels the situation in congenitally athymic mice (Wells et al., 1981, 1983). We conclude that in this system the CD8 ÷ cell response alone is lethal, as is the CD4 ÷ cell response (Table 2). This contrasts with adoptive transfer experiments in which it has been shown that CD8 ÷ cells are beneficial and CD4 ÷ cells are deterimental to the survival of mice sublethally infected with influenza virus (see Discussion). Role o f CD8 + and CD4 ÷ T cells in D I virus-mediated protection

Although the data in Fig. 5 indicate that recovery was not T cell-mediated, there was a striking increase in residual pathology in mice depleted of CD8 + or CD4 + T cells. At 12 days p.i., when there were no longer any clinical signs of respiratory disease, mice which had been injected with PBS showed about 7 % consolidation of the lungs, those depleted of CD8 + cells had 20% consolida-

T cell responses in D I virus-treated mice 50

40O

30cO

20o

10-

PBS

AntiCD4

AntiCD8

AntiCD4 +

AntiCD8 Fig. 5. Lung consolidation at 12 days p.i. in mice treated with 10 LDs0 A/WSN and active DI WSN virus. Mice were injected i.v. with 100111 PBS, I00 p.g anti-CD4 or 100 ixg anti-CD8 MAb or anti-CD4 and antiCD8 MAbs 3 days and 1 day prior to infection.

100

!

F

!

A

' I

!

I

I

I

regulate the others' activities reciprocally. However, we have not yet determined which cell(s) are responsible for consolidation under these conditions. These data also show that CD4 ÷ and CD8 ÷ cell populations are each able to clear infectivity. This agrees with the data of Allan et al. (1990), who depleted CD4 ÷ cells in vivo, and others who reconstituted immunodeficient mice with already specifically activated T cell populations (reviewed by Yewdell & Hackett, 1989). Somewhat surprising was the clearance of infectivity in doubly depleted CD4-/CD8animals infected with WSN virus and treated with active DI virus (< 10 p.f.u./ml in each of four mice 12 days p.i.), because in the parallel group without active DI virus virus was not cleared at all (2 x 106, 1"5 × 107 and 3-5 × 107 p.f.u./ml in three mice 12 days p.i.). It is possible that DI virus acts late in infection to inhibit multiplication, although there is no sign of this up to day 7 (Dimmock et al., 1986), or that there is a compensating host response. Failure of D I A / W S N (H1N1) virus to protect mice from &thai infection with A/PR/8/34 (HIN1) virus

Fig. 6 shows that DI WSN virus exerted no protection against lethal A/PR/8/34 virus infection in mice.

I

I

80-

A

7" r'

379

/

! I 60I I I ! I I I ! I I

O

40-

Discussion

E 20 --

, ..{J?, w

2

4

m

w

i

6 8 Time p.i. (days)

10

12

Fig. 6. Lack of heterologous protection of mice by DI WSN virus. Mice were inoculated with active DI WSN virus (11) or BPL-inactivated DI WSN virus (O) together with 10 LDso A/WSN (H1N1) virus (solid lines) or A/PR/8/34 (H1Nt) (dashed lines),

tion and those depleted of CD4 ÷ cells 40 ~ consolidation, whereas mice depleted of both CD8 ÷ and CD4 + cells showed no consolidation (Fig. 5). This was indistinguishable from consolidation observed during acute infection in non-depleted mice. The increase in lung pathology was not reflected in any change in the rate of decline in signs of malaise and respiratory distress, or any other demonstrable aspect of the recovery process. Failure to resolve consolidation in mice depleted with anti-CD4 or anti-CD8 MAbs suggests that CD4 ÷ and CD8 ÷ cells

We have found a significant CTL response in the lungs of lethally infected mice, the cells responsible being MHC class I-restricted, CD8 + (McLain & Dimmock, 1991) and largely virus-strain specific. The CD8 + response was shown to be lethal after MAb depletion in vivo. These data differ from the conclusions of others in three ways. (i) Although Allan et al. (1990) have reported a high level of primary CTL activity in the lung exudates of mice given a sublethal influenza virus infection, most workers have used CTLs expanded by secondary in vitro stimulation because this gives a more consistent response (Yap & Ada, 1977), and most use spleen rather than lung as the source of lymphocytes (see reviews by Ada & Jones, 1986; Yewdell & Hackett, 1989). We also found a low-level primary CTL response in the spleen (data not shown). (ii) Most splenic CTLs are cross-reactive with different influenza A virus subtypes (Wraith, 1987; Wysocka & Bennink, 1988; Yewdell & Hackett, 1989; Reay et al., 1989). Although our overall lung CTL population was predominantly virus strain-specific and therefore presumably directed to epitopes of HA or NA, we also found CTL activity against most of the conserved antigens (PB1, PB2, PA, NP, NS1 and NS2).

380

L. M c L a i n , D. J. M o r g a n and N. J. D i m m o c k

(iii) It has been found consistently that the adoptive transfer of MHC class I-restricted and/or CD8 ÷ CTLs is beneficial to the influenza virus-infected mouse. This reduced lung virus titres, decreased mortality and promoted tissue repair (Yap et al., 1978; Webster & Askonas, 1980; Lin & Askonas, 1981 ; Wells et al., 1983; Lukacher et al., 1984; Askonas et al., 1988; MacKenzie et al., 1989). Therefore it would appear that preexisting CTLs protect against infection, whereas those generated during infection are pathogenic. Finally our data and those of others agree that delayed type hypersensitivitymediating and/or CD4 ÷ T cells, whether administered before infection or generated during the course of it, can be pathogenic in the influenza virus-infected mouse (Leung & Ada, 1980, 1982; Liew & Russell, 1980, 1983), although other data show them to be marginally beneficial (Lukacher et al., 1986; Askonas et al., 1988). In conclusion, our data clearly show the immunopathogenic contribution of T cells during fatal influenza pneumonia in the mouse. Although few systems have been studied, treatment of mice lethally infected with vesicular stomatitis virus or Semliki Forest virus with their respective DI viruses greatly reduces virus multiplication in target tissue, as it does m vitro (Barrett & Dimmock, 1986). In the influenza system there is no detectable reduction in infectivity (Dimmock et al., 1986) and, as the pathology is immunemediated, we turned to the immune system for an explanation. Earlier we showed that infected mice treated with DI virus develop a delayed and less severe lung consolidation from which they recover (Dimmock et al., 1986) and now we find that this correlates with the reduction in extent of the primary CD8 ÷ CTL response, which we also show is lethal in its effect. The CD4 ÷ cell response generated during the infection is also lethal, so this too must be reduced by treatment with DI virus. Currently the mechanism by which this comes about is a matter of speculation, but further data (D. J. Morgan & N. Dimmock, unpublished results) suggest that the lethal immune response results from virus entering T cells and modifying their activity along the lines suggested by Casali et al. (1984). Influenza virus multiplication in lymphocytes is abortive and non-cytolytic (Brownson et al., 1979; Casali el al., 1984). We suggest that DI virus also enters lymphocytes and prevents manifestation of the lethal response by conventional DI virus-mediated interference. Therefore what we are seeing is cell typespecific interference, a phenomenon well documented from the in vitro study of cell lines, and known to operate at both quantitative and qualitative levels (Holland, 1990; Dimmock, 1991). One aspect of this system as yet unexplored is the effect of administration of DI virus on immunoregulatory lymphokines, such as tumour necrosis factor ~ and ~-interferon, which can have a profound influence on immunopathology.

How does the DI virus-infected mouse recover from infection? Clearly neither CD8 ÷ nor CD4 ÷ cells are responsible since their deletion did not increase lethality or alter the clinical signs of disease, although they evidently contributed to the resolution of pathology in the lungs; failure to clear consolidation was particularly marked in the absence of CD4 ÷ cells. DI virus may act to inhibit virus multiplication late in infection or ther e may be compensating host responses (see Results). However, in considering the effects of T cell depletion it should be remembered that this is a systemic phenomenon, that influenza virus is found systemically and thus survival may not be exclusively dependent on events taking place in the lung. Another possibility is that locally produced HA-specific IgG, which is produced only in the lungs of DI virus-treated infected mice (Dimmock et al., 1986; McLain & Dimmock, 1989, 1991), assists in the recovery process. (We have reported that the production of HAspecific IgG in the lung is suppressed in lethally infected mice, although there is the same amount of anti-NA antibody produced as in DI virus-treated infected mice.) The HA-specific IgG is unusual as none of the three specificities identified is neutralizing, but one was shown to be protective when passively transferred prior to lethal challenge. This antibody has the unusual property of having the specificity of a T cell receptor, reacting only with non-native HA in the context of its autologous MHC class I antigen. How it acts in vivo is not known. We thank Dr R. W. Compans and Dr G. L. Smith for generously providing viruses. Financial assistance was provided by the Cancer Research Campaign and the SERC.

References ADA, G. L. & JONES, P. D. (1986). The immune response to influenza infection. Current Topics in Microbiology and Immunology 128, 1 54. ALLAN, W., TABI, Z., CLEARY, A. & DOHERTY, P. C. (1990). Cellular events in the lymph node and lung of mice with influenza: consequences of depleting CD + T cells. Journal of Immunology 144, 3980 3986. ASKONAS, B. A., TAYLOR, P. M. & ESQUIVEL, F. (1988). Cytoxic T cells in influenza infection. Annals of the New York Academy of Sciences 532, 230-237. BARRETT, A D. T. & D~MMOCK, N. J. (1986). Defective interfering viruses and infections of animals. Current Topicsin Microbiologyand Immunology 128, 55-84. BARRETT, A. D. T., HUNT, N. & DIMMOCK, N. J. (1984). A rapid method for the inactivation of virus infectivity prior to assay for interferons. Journal of Virological Methods 8, 349-351. BROWNSON, J. M., MAHY, B. W. J. & HAZLEMAN, B. L. (1979). Interaction of influenza A virus with human peripheral blood lymphocytes. Infection and Immunity 25, 749-756. CASALI, P., RICE, G. P. A. & OLDSTONE, M. B. A. (1984). Viruses disrupt functions of human lymphocytes. Effects of measles virus and influenza virus on lymphocyte-mediated killing and antibody production. Journal of Experimental Medicine 159, 1322-1337. CATE, T. R. & MOLD, N. G. (1975). Increased influenza mortality of mice adoptively immunized with node and spleen cells sensitized by inactivated but not live virus. Infection and Immunity 11, 908914.

T cell responses in D I virus-treated mice

COBBOLD, S. P., JAYASURIYA, A., NASH, A., PROSPERO, T. D. & WALDMANN, H. (1984). Therapy with monoclonal antibodies by elimination of T cell subsets in vitro. Nature, London 312, 548-551. DIMMOCK, N. J. (1991). The biological significance of defective interfering viruses. Reviews in Medical Virology (in press). DIMMOCK, N. J. & KENNEDY, S. I. T. (1978). Prevention of death in Semliki Forest virus-infected mice by administration of defectiveinterfering Semliki Forest virus. Journal of General Virology 39, 231 242. DIMMOCK, N. J., BECK, S. & MCLA~, L. (1986). Protection of mice from lethal influenza: evidence that defective interfering virus modulates the immune response and not virus multiplication. Journal of General Virology 67, 839-850. HOLLAND, J. J. (1990). Defective viral genomes. In Virology, 2nd edn., vol. 1, pp. 151-165. Edited by B. N. Fields & D. M. Knipe. New York: Raven Press. LEUNG, K. N. & ADA, G. L. (1980). Cells mediating delayed-type hypersensitivity in the lungs of mice infected with an influenza A virus. Scandinavian Journal of Immunology 12, 393-400. LEUNG, K. N. & ADA, G. L. (1982). Different functions of helper T cells on the primary in vitro production of delayed hypersensitivity to influenza virus. Cellular Immunology 67, 312 324. LIEW, F. Y. & RUSSELL, S. M. (1980). Delayed-type hypersensitivity to influenza virus. Journal of Experimental Medicine 151, 799~14. LIEW, F. Y. & RUSSELL, S. M. (1983). Inhibition of pathogenic effect of effector T cells by specific suppressor T cells during influenza infection of mice. Nature, London 304, 541 543. LIN, Y.-L. & ASKONAS, B. A. (1981). Biological properties of an influenza virus-specific killer T cell clone. Journal of Experimental Medicine 154, 225 234. LUKACHER, A. E., BRACIALE,V. L. & BRACIALE, T. J. (1984). In vivo effector function of influenza virus-specific cytotoxic T lymphocyte clones is highly specific. Journal of Experimental Medicine 160, 814 826. LUKACHER, A. E., MORRISON, L. A., BRACIALE,V. L. & BRACIALE,T. J. (1986). T lymphocyte function in recovery from experimental viral infection: the influenza model. In Mechanisms of Host Resistance to lnJbctious Agents, Tumors, and Allografts, pp. 233-254. Edited by R. M. Steinman & R. J. North. New York: Rockefeller University Press. MACKENZIE, C. D., TAYLOR, P. M. & ASKONAS, B. A. (1989). Rapid recovery of lung histology correlates with clearance of influenza virus by specific CD8 + cytotoxic cells. Immunology 67, 376-381. McLAIN, L. & DIMMOCK, N. J. (1989). Protection of mice from lethal influenza by adoptive transfer of non-neutralizing haemagglutination-inhibiting lgG obtained from the lungs of infected animals

381

treated with defective interfering virus. Journal of General Virology 70, 2615-2624. MCLAIN, L. & DIMMOCK, N. J. (1991). An influenza hemagglutininspecific IgG enhances class I MHC-restricted CTL killing in vitro. Immunology 73, 12-18. PERRAULX, J. (1981). Origin and replication of defective interfering particles. Current Topics in Microbiology and Immunology 93, 151207. REAY, P. A., JONES, I. M., Goxcn, F. M., McMICnAEL, A. J. & BROWNLEE, G. G. (1989). Recognition of the PAl, neuraminidase and matrix proteins of influenza virus A/NT/60/68 by cytotoxic T lymphocytes. Virology 170, 477-485. Roux, L., SIMON, A. E. & HOLLAND, J. J. (1991). Effects of defective interfering viruses on viral replication and pathogenesis in vitro and in vivo. Advances in Virus Research (in press). WE~TER, R. G. & ASKONAS,B. A. (1980). Cross-protection and crossreactive cytotoxic T cells induced by influenza virus vaccines in mice. European Journal of lmmunology 10, 396-401. WELLS, M. A., DANIEL, S., DJEU, J. Y., KILEY, S. C. & ENNIS, F. A. (1983). Recovery from a viral respiratory infection. IV. Specificity of protection by cytotoxic T lymphocytes. Journal oflmmunology 130, 2908-2914. WELLS, M. A., ENNIS, F. A. & ALBRECHT, P. (1981). Recovery from a viral respiratory tract infection. 2. Passive transfer of immune cells to mice with influenza pneumonia. Journal of Immunology 126, 10421046. WRAITH,D. C. (1987). The recognition of influenza virus-infected cells by cytotoxic T lymphocytes. Immunology" Today 8, 239-245. WYSOCKA, M. & BENNINCK, J. R. (1988). Limiting dilution analysis of memory cytotoxic T lymphocytes specific for individual influenza virus gene products. Cellular Immunology 112, 425-429. YAP, K. L. & ADA, G. L. (1977). Cytotoxic T cells in the lungs of mice infected with influenza virus. Scandinavian Journal of lmmunology 7, 73-80. YAP, K. L., ADA, G. L. & MCKENZIE, I. F. C. (1978). Transfer of specific cytotoxic T lymphocytes protect mice inoculated with influenza virus. Nature, London 273, 238-239. YEWDELL, J. W. & HACKETT, C. J. (1989). The specificity and function of T lymphocytes induced by influenza A viruses. In The Influenza Viruses, pp. 361-429. Edited by R. M. Krug. New York: Plenum Press.

(Receh,ed 3 September 1991 : Accepted 1 November 1991)