Attenuation and immunogenicity in primates of vaccinia virus recombinants expressing human interleukin-2 Charles Flexner *§, Bernard Moss*, William T. London ~;and Brian R. Murphy *~ Vector-directed lymphokine expression represents a novel approach to the attenuation of live recombinant viruses which might be used as vaccines. Expression of interleukin-2 (IL-2) by recombinant vaccinia virus has been shown to significantly attenuate virus virulence in rodent species without diminishing immunogenicity. Skin lesion formation and immunogenicity o f vaccinia/IL-2 recombinants in three species of primates was examined. IL-2 expression was associated with a 15-fold reduction in the area of induration after intradermal inoculation of recombinant viruses in patas monkeys. Wild type and a control vaccinia recombinant produced large ( > 5000 mm 2) skin ulcers in this species, but the 11[.-2 expressing recombinant produced no ulceration. Production of antibodies to vaccinia virus and to influenza A virus haemagglutinin expressed by recombinant vectors was examined in rhesus and squirrel monkeys. 11_.-2 expression accelerated the resolution of skin lesions in rhesus but not squirrel monkeys. Despite this, antibody production was equivalent in the presence or absence of IL-2. IL-2 expression can greatly reduce the skin lesions formed by live recombinant vaccinia vectors in primates, indicating significant attenuation, without reducing the immunogenicity of the vaccine. Keywords: Vaccinia virus; interleukin-2; virus attenuation

Introduction The expression of antigenic proteins by recombinant live virus vectors represent a promising new approach to immunization. Live viruses currently under investigation as vaccine vectors include adenovirus ~, herpesviruses z,3 and poxviruses 4. The advantage of this approach to vaccination includes authentic expression of complex eukaryotic proteins, ease and economy of vaccine preparation, and long-lived cellular and humoral immunity induced by the vaccine. Live recombinant vaccinia viruses in particular have been used to protect a number of animal species, including primates, from infection by pathogenic viruses *-6. The major drawback to such vaccines is the virulence of the vector itself, since all are live viruses related to or derived from human pathogens. Adenoviruses, herpesviruses, and poxviruses can all cause progressive or disseminated disease in immunocompromised hosts. Attenuation of virulence is therefore essential to the development of these recombinant vaccines. Virus attenuation can be achieved through spontaneous or purposeful inactivation of virus genes involved in virulence. Indeed, several vaccinia virus genes have been shown to be important for virulence in animals, and the inactivation of such genes has resulted in attenuation 7-9. *Laboratory of Viral Diseases and tLaboratory of Infectious Diseases, National Institute of Allergy and Infectious Diseases, and ~;Intramural Research Program, National Institute of Neurological and Communicative Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892, USA. §Present address: Division of Clinical Pharmacology, Department of Medicine, The Johns Hopkins Hospital, Baltimore, Maryland 21205, USA. =To whom correspondence should be addressed. (Received 10 January 1989; accepted 2 February 1989) 0264-410X/90/080017-05 $03.00 ((.~ 1990 Butterworth & Co. (Publishers) Ltd

An alternative approach to attenuation involves the incorporation of immunostimulatory substances, such as lymphokines, into live recombinant viruses. Thus, the expression of the lymphokine interleukin-2 (IL-2) by recombinant vaccinia leads to rapid clearance of virus and protection of athymic nude mice from death due to progressive vaccinia infection ~°'1 ~. Furthermore, viruses expressing IL-2 are attenuated in normal, immunocompetent rodents via intracranial or intradermal inoculation, but maintain their immunogenicity, stimulating strong humoral and cellular immunity to both vaccinia virus and foreign antigens expressed by recombinant virus x°. The relevance of such observations to human vaccination with live recombinant viruses remains speculative. We have therefore administered live recombinant vaccinia viruses expressing human IL-2 to several primate species and have addressed two questions: first, whether IL-2 expression attenuates vaccinia virus for primates; second, whether co-expression of IL-2 and a foreign viral protein affects the immunogenicity of the foreign protein.

Materials and methods In order to determine if IL-2 expression by a vaccinia virus can attenuate the recombinant viruses for primates, wild type and recombinant viruses were evaluated in a primate species (patus monkey, Eyphrocebus paras) known to develop large skin lesions with ulceration after intradermal (i.d.) inoculation with vaccinia 12. The virus recombinants chosen were made by insertions into a gene encoding a non-essential 38kDa protein and located within the HindlII C region of the genome 1°. Such recombinants have been shown previously to be nearly as virulent as the wild type strain of vaccinia in laboratory animals ~°. Construction and characterization

Vaccine, Vol. 8, February 1990

17

Vaccinia recombinant expressing interleukin-2: C. Flexner et al.

of recombinant viruses vTFCLZ-I (vacc//~-gal) and vCF13 (vacc/IL-2) which contain the E. coil fl-galactosidase (/%gal) gene with or without the human IL-2 gene inserted into the HindIII C region have been reported previously ~°. In both cases,/~-gal expression is regulated by the vaccinia late Pl I promoter, and IL-2 expression is controlled by the vaccinia early-late P7.5 promoter. All viruses were grown, titred, and purified as previously described ~3

Results Threc patas monkeys were anaesthetized, shaved on the back, and inoculated i.d. with 10 6 plaque-formingunits (p.f.u.) of wild type virus, vacc//3-gal, and vacc/IL-2 at widely separated sites on the back. Lesion size was measured on horizontal and vertical axes at three or four day intervals, and the surface area of induration and ulceration for each lesion was estimated by using the average lesion radius. All animals were bled prior to study and determined to be seronegative for vaccinia by ELISA, performed as described ~°. Vaccinia stocks were titred, aliquoted, and frozen at - 7 0 ° C ; all titres were confirmed on the day of inoculation. Lesion formation over time in patas monkeys is shown in Figure 1. Control recombinant vacc//%gal and wild type virus caused large areas of skin induration and ulceration, peaking at day 10 to 14, with gradual resolution thereafter; ulcers were 6000 -

a

,EE q000

g "~

2000

I

20

I0

6000

40

30

50

b

¢'4

q000 '~

. I

J

2000

0

~'~'~ 0

"r" I0

-I

~

[~

I

20

"~'I 30

I

I ~ 40

I 50

Time ( d a y s ) Figure I Mean area of skin induration (a) or ulceration (b) after inoculation of patas monkeys with vaccinia virus wild type (wt) or reeombinants. Surface area of lesions was measured as described in the text; n : 3 animals for each point. Standard errors are indicated by bars. Statistically significant differences between values for vacclHAlflgal and vacclHAllL-2 are indicated (*p
18

Vaccine, Vol. 8, February 1990

still not completely healed by day 42. Vacc//3-gal lesions were slightly but not significantly smaller than wild-type virus in areas of induration and ulceration (p>0.05, Student's paired t-test). In contrast, vacc/IL-2 inoculation resulted in a small area of induration, about 15-fold less than vacc//~-gal at the peak of lesion size (day 10, Figure la); this difference was statistically significant (p <0.01). Vacc/IL-2 produced no skin ulceration in any of the patas monkeys (Figure Ib). Since each patas monkey received three different vaccinia viruses, including wild type virus, the relative immunogenicities of the three viruses could not be assessed in this experiment. These findings clearly demonstrated that expression of the IL-2 gene by a vaccinia recombinant could greatly attenuate the virus for primates. Since live recombinant vaccinia is under consideration as a vaccine vector, it was felt important to examine the impact of IL-2 expression on the immune response in primates to a foreign viral antigen expressed by a recombinant virus. The immune response to respiratory syncytial virus (RSV) glycoproteins expressed by recombinant vaccinia viruses was proportional to the size of skin lesions formed in primate species 12. The authors were therefore concerned that the reduction in skin lesion size associated with lymphokine expression might be accompanied by a corresponding decrease in the immune response to the vector. Because vaccinia virus induces a large lesion in patas monkeys as revealed in a previous study ~z and confirmed in the present study, we chose not to perform the immunogenicity study in this species. Therefore the immunogenicity of IL-2 expressing recombinant virus was evaluated in two primate species already known to form intermediate or small skin lesions after i.d. inoculation of vaccinia recombinants: rhesus (Macaca mullata) and squirrel (Saimiri sciureus) monkeys, respectively. These monkeys also demonstrated a correspondingly intermediate or low antibody response to RSV glycoproteins expressed by recombinant vaccinia viruses ~2. Four animals of each species were anaesthetized, shaved on the back, and inoculated i.d. with 108 p.f.u, of vacc/HA/fl-gal, a control recombinant containing the influenza A/PR/8/34 haemagglutinin (HA) inserted into the vaccinia thymidine kinase (TK) gene, and fl-gal inserted into the 38 kDa gene in HindIII C (vCF20, described previously~°), or vacc/HA/IL-2, which contains HA inserted into the T K gene and fl-gal plus human IL-2 into the 38kDa gene (vCF19, described previously~°). Skin lesions were measured every 3-4 days and serum samples were collected from animals on days 0, 28, and 42. Each animal was seronegative for vaccinia by ELISA prior to the study. Each recombinant vaccinia virus had the T K - phenotype, a phenotype that is associated with decreased lesion size in primates ~2. Thus, the effect of IL-2 expression on lesion size in the vacc/HA/IL-2 and vacc/HA/fl-gal infected rhesus and squirrel monkeys occurs on a background of reduced lesion size imposed by the T K - phenotype. Average skin lesion formation over time in rhesus and squirrel monkeys is shown in Figure 2. As expected, both species developed much smaller skin lesions than patas monkeys, and skin lesions were cleared more rapidly. This occurred despite the 100-fold greater amount of virus administered to the rhesus and squirrel monkeys. This small lesion size reflects the relative resistance of the rhesus and squirrel monkeys to vaccinia skin lesion formation ~2, as well as

Vaccinia recombinant expressing interleukin-2: C. Flexner et al.

200

C

200

C'4 E

E E

I00

V,

100

o~ _J

_1

I

1 l0

20o

I

I0 0

b

10

20

10

20

I

[ 30

d

200

E E ¢) N

o

30

2O

E E I00

.~_ --I

I

0 0

I0

20

I

o

30

Time

I0 0

1

[ 30

(days)

Figure 2

Mean area of skin induration (a,c) or ulceration (b,d) after inoculation of rhesus (a,b) or squirrel (c,d) monkeys with vaccinia virus recombinants. Surface area of lesions was measured as described in the text; n = 4 animals for each point. Standard errors are indicated by bars. Statistically significant differences between values for vacc/HA/~-gal and vacc/HA/IL-2 are indicated (*p<0.05, Student's two-tailed t test). O, vacc/HA/,8-gal; O, vacc/HA/IL-2

the T K - phenotype of the double recombinant vectors, conferring on them increased attenuation 7'12. Rhesus monkeys developed skin lesions which were about twice as large as those seen in squirrel monkeys (Figure 2), and their lesions persisted ,~7 days longer. This was consistent with dermal lesions previously seen in these species xz. Vace/HA/IL-2 and vacc/HA/fl-gal produced nearly identical lesion sizes, peaking at day 10. However, resolution of lesions produced by the IL-2 expressing vector was significantly accelerated in rhesus monkeys (p<0.05, Figure 2a and b), with vacc/HA/IL-2 lesions cleared 4-7 days earlier than those in vacc/HA//?-gal infected monkeys. Lesion size and duration in squirrel monkeys were similar with or without IL-2 (Figure 2c and d). Serum antibodies to influenza A virus were analysed by a haemagglutination-inhibition (HI) assay employing chicken erythrocytes, and antibodies to vaccinia virus were analysed by ELISA. Summary data are presented in Table 1. In general, the antibody responses to influenza and vaccinia were equivalent or slightly less in monkeys inoculated with vacc/HA/IL-2 compared to monkeys inoculated with vacc/HA//%gal. These differences were not statistically significant (p>0.05). Rhesus monkeys had approximately 10-fold higher antibody titres to vaccinia virus than did squirrel monkeys with either vector; these differences showed a tendency to but did not achieve statistical significance (0.10 > p > 0.05). This finding is consistent with data obtained, previously in

comparing serum antibody responses to vaccinia vectors expressing RSV glycoproteins in rhesus versus squirrel monkeys lz, in which the larger skin lesion size was associated with an enhanced antibody response to RS glycoproteins. Antibody titres to influenza HA were not significantly different in the two species in the present study, however (Table 1). Discussion

The lymphokine IL-2 is secreted by C D 4 + lymphocytes in response to antigenic stimulation, and plays a central role in the activation of both humoral and cellular immune responses 14. The administration of IL-2 has been shown to protect animals against otherwise lethal viral infections 15"16 and can enhance the immune response to vaccination 17'1s. The expression of IL-2 by live recombinant vaccinia virus protects immunocompromised athymic nude mice from lethal progressive vaccinia infection, apparently as a result of accelerated virus clearance 1°'~. IL-2 expression also attenuates vaccinia in immunocompetent rodents, as measured by intracranial lethality and skin lesion formation a°. A reduction in the size of vaccinia skin lesions in rodents 19 and primates 12 as a result of inactivation of the virus T K gene has been previously reported. A reduction in vaccinia skin lesion size in primate species as a result of vectordirected IL-2 expression has now been demonstrated. IL-2 expression eliminated skin ulceration and reduced

V a c c i n e , Vol. 8, F e b r u a r y 1990

19

Vaccinia r e c o m b i n a n t e x p r e s s i n g interleukin-2: C. F l e x n e r et al.

Table 1

Serum antibody response to influenza and vaccinia viruses in primates immunized with recombinant vaccinia viruses"

Influenza HI titresb (reciprocal mean Iog,o) Rhesus monkeys

Day 28 Day 42

Squirrel monkeys

vacc/HA/,6-gal

vacc/HA/IL-2

vacc/HA/,8-gal

vacc/HA/IL-2

2.334-0.14 2.334-0.14

2.11 +0.12 1.96__.0.19

2.33±0.19 2.03±0.19

1.96±0.26 1.81 ±0.27

vacc/HA/#-gal

vacc/HA/IL-2

vacc/HA//Y-gal

vacc/HA/IL-2

3.91 +0.29 3.76___0.25

3.91 +0.42 3.76±0.29

3.16+0.15 3.01 +0.25

2.71 ±0.52 2.86_+0.62

Vaccinia ELISA titres ' (reciprocal mean Iog,o)

Day 28 Day 42

• Reciprocal mean Iog,o titres -+ standard error; n = 4 for each group ~Haemagglutination-inhibiting (HI) titres in preqmmunization sera were ~<0.3 (reciprocal Iog,o) in each animal CELISA antibody titres in pre-immunization sera were ~<2.0 (reciprocal Ioglo) in each animal

peak induration 15-fold in patas monkeys, and significantly enhanced the healing of virus lesions in rhesus monkeys. Since the size of skin lesions formed by vaccinia reflects both virulence and extent of virus replication I o. ~9, the authors believe that the diminished skin lesions seen in primates reflects both diminished virus replication and enhanced virus clearance, although direct evidence for this would be difficult to obtain in primates. We chose the patas monkey for this study because of its susceptibility to vaccinia infection, as indicated by the large skin lesions formed by wild type vaccinia virus in this species I 2. The reduction in skin induration associated with IL-2 expression equals or surpasses the degree of attenuation of vaccinia associated with a T K - phenotype in patas monkeys I 2. However, the fact that no ulceration occurred at the site of inoculation with vacc/IL-2 in any of these animals suggests that IL-2 expression may attenuate vaccinia even more than T K inactivation in this primate species. The impact of IL-2 on antibody response was examined in rhesus and squirrel monkeys. Serum antibody production to both vaccinia and H A was equivalent in the presence or absence of IL-2. In the case of rhesus monkeys, this was achieved despite significant acceleration in the clearance of skin lesions. Interestingly, vacc/HA/IL-2 caused skin ulceration in rhesus and squirrel monkeys, even though ulcers were not observed with vacc/IL-2 in patas monkeys. This difference in ulceration between patas and rhesus or squirrel monkeys may reflect the larger virus inoculum given rhesus and squirrel monkeys, or may represent a differential effect of human IL-2 in different primate species. Rhesus and squirrel monkeys received double recombinant vectors which were T K - , and hence already somewhat attenuated. Expression of IL-2 had only a mild additional attenuating effect on the virus in rhesus monkeys, as indicated by accelerated clearance of virus lesions. This is consistent with published data on the virulence of T K - IL-2 recombinants in mice and rabbits x°, and suggests that in immunocompetent animals, IL-2 expression may not further attenuate an already attenuated virus. Inactivation of the viral T K gene did not adversely affect replication or immunogenicity of an already attenuated vaccine strain of vaccinia 2°. These data suggest that combining measures to attenuate a virus (e.g., gene inactivation and lymphokine expression) may not have an additive effect on attenuation.

20

Vaccine, Vol. 8, F e b r u a r y 1990

Conclusions Ongoing human trials with vaccinia recombinants expressing the human immunodeficiency virus envelope glycoprotein z1'22, coupled with a recent report of progressive disease in an HIV-seropositive army recruit vaccinated with vaccinia 23 accentuate the need to develop further means of enhancing the safety of live virus vaccine vectors. It is unknown if the inactivation of genes such as the T K will make live vaccinia safer for humans. Vaccinia vectors lacking a functional T K were still capable of causing progressive disease and death in athymic nude mice 1°. The ability of IL-2 to attenuate vaccinia in both rodent and primate species suggests that lymphokine expression is an alternative to gene inactivation as a means of enhancing the safety of live recombinant viruses, and may hold some advantages, particularly for immunocompromised individuals. The expression of IL-2 by live recombinant vaccinia virus vectors represents a novel means of attenuating the virus in primates without adversely affecting immunogenicity. These data may have important implications for the further development of recombinant vaccinia as a vaccine vector for man.

Acknowledgements The authors are grateful to Clarence Banks and Joe Jackson for assistance with ELISA and influenza HI assays, and Norman Cooper for growth and purification of viruses. References 1

2

3

4 5

Davis, A.R., Kostek, B., Mason, B.B., Hsiao, C.L., Morin, J., Dheer, S.K. and Hung, P.P. Expression of hepatitis B surface antigen with a recombinant adenovirus. Proc. Natl. Acad. Sci. USA 1985, 82, 7560 Shih, M-F., Arsenakis, M., Tiollais, P. and Roizman, B. Expression of hepatitis B virus S genes by herpes simplex virus type 1 vectors carrying c~- and ,8-regulated gene chimeras. Proc. Natl. Acad. Sci. USA 1984, 81, 5867 Lowe, R.S., Keller, P.M., Keech, B.J., Davison, A.J., Whang, Y., Morgan, A.J. et al. Varicella-zoster virus as a live vector for the expression of foreign genes. Proc. Natl. Acad. Sci. USA 1987, 84, 3896 Moss, B. and Flexner, C. Vaccinia virus expression vectors. Ann. Rev. Immunol. 1987, 5, 305 Moss, B., Smith, G.L., Gerin, J.L and Purcell, R.H. Live recombinant vaccinia virus protects chimpanzees against hepatitis B. Nature 1984, 311, 67

Vacci ni a r e c o m b i n a n t e x p r e s s i n g i n t e r l e u k i n - 2 : C. F l e x n e r et al.

6

Spriggs, M.K., Collins, P.L., Tierney, E., London, W.T. and Murphy, B.R. Immunization with vaccinia virus recombinants that express the surface glycoproteins of human parainfluenza virus type 3 (PIV3) protects Patas monkeys against PIV3 infection. J. Virol. 1988, 82, 1293 7 Buller, R.ML., Smith, G.L., Cremer, K., Notkins, A.L. and Moss, B. Decreased virulence of recombinant vaccinia virus vectors is associated with a thymidine kinase-negative phenotype. Nature 1985, 317, 813 8 Buller, R.ML., Chakrabarti, S., Cooper, J., Twardzik, D.R. and Moss, B. Deletion of the vaccinia virus growth factor gene reduces virus virulence. J. Virol. 1988, 62, 866 9 Dallo, S. and Esteban, M. Isolation and characterization of attenuated mutants of vaccinia virus. Virology 1987, 159, 408 10 Flexner, C., Hugin, A. and Moss, B. Prevention of vaccinia virus infection in immunodeficient mice by vector-directed IL-2 expression. Nature 1987, 330, 259 11 Ramshaw, I.A., Andrew, M.E., Phillips, S.M., Boyle, D.B. and Coupar, B.E. Recovery of immunodeficient mice from a vaccinia virus/IL-2 recombinant infection. Nature 1987, 329, 545 12 Olmsted, R.A., Buller, R.M.L., Murphy, B.R., Beeler, J.A., Collins, P.L. and London, W.T. Evaluation in nonhuman primates of the safety, immunogenicity, and efficacy of recombinant vaccinia viruses expressing the F and G glycoproteins of respiratory syncytial virus. In: Vaccines 88, New Chemical and Genetic Approaches to Vaccination (Eds Ginsberg, H., Brown, F., Lerner, R.A. and Chanock, R.M.) Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, 1988, pp. 205-210 13 Mackett, M., Smith, G.L. and Moss, B. The construction and characterization of vaccinia virus recombinants expressing foreign genes. In: DNA Cloning, Vol. II (Ed. Glover, D.M.) IRL Press, Oxford, 1985, pp. 191-212 14 Rosenberg, S.A. and Lotze, M.T. Cancer immunotherapy using

15

16

17

18 19

20

21

22

23

interleukin-2 and interleukin-2-activated lymphocytes. Ann. Rev. Immunol. 1986, 4, 681 Rouse, B.T., Miller, L.S., Turtinen, T. and Moore, R.N. Augmentation of immunity to herpes simplex virus by in vivo administration of interleukin 2. J. Immunol. 1985, 134, 926 Weinberg, A., Konrad, M. and Merigan, T.C. Regulation by recombinant interleukin-2 of protective immunity against recurrent herpes simplex virus type 2 genital infection in guinea pigs. J. Virol. 1987, 81, 2120 Kawamura, H., Rosenberg, S.A. and Berzofsky, J.A. Immunization with antigen and interleukin 2 in vivo overcomes Ir gene low responsiveness. J. Exp. Med. 1985, 162, 381 Weinberg, A. and Merigan, T.C. Recombinant interleukin 2 as an adjuvant for vaccine-induced protection. J. Immunol. 1988,140, 294 Buller, R.M.L. and Moss, B. Genetic basis for vaccinia virus virulence. In: Vaccinia Viruses as Vectors for Vaccine Antigens (Ed. Quinnan, G.V.) Elsevier, New York, 1985, pp.37~47 Morita, M., Suzuki, K., Yasuda, A., Kojima, A., Sugimoto, M., Watanabe, K. et al. Recombinant vaccinia virus LC16m0 or LC16m8 that expresses hepatitis B surface antigen while preserving the attenuation of the parental virus strain. Vaccine 1987, 5, 65 Zagury, D., Leonard, R., Fouchard, M., Reveil, B., Bernard, J., Ittele, D. et al. Immunization against AIDS in humans. Nature 1987, 328, 249 Zagury, D., Bernard, J., Cheynier, R., Desportes, I., Leonard, R., Fouchard, M. et al. A group specific anamnestic immune reaction against HIV-1 induced by a candidate vaccine against AIDS. Nature 1988, 322, 728 Redfield, R.R., Wright, D.C., James, W.D., Jones, T.S., Brown, C. and Burke, D.S. Disseminated vaccinia in a military recruit with human immunodeficiency virus (HIV) disease. N. Engl. J. Med. 1987, 318, 673

Vaccine, Vol. 8, F e b r u a r y 1990

21