Expression of immunoregulatory cytokines by recombinant coxsackievirus B3 variants confers protection against virus-caused myocarditis

Abstract

Clinical and laboratory investigations have demonstrated the involvement of viruses and bacteria as potential causative agents in cardiovascular disease and have specifically found coxsackievirus B3 (CVB3) to be a leading cause. Experimental data indicate that cytokines are involved in controlling CVB3 replication. Therefore, recombinant CVB3 (CVB3rec) variants expressing the T-helper-1 (T(H)1)-specific gamma interferon (IFN-gamma) or the T(H)2-specific interleukin-10 (IL-10) as well as the control virus CVB3(muIL-10), which produce only biologically inactive IL-10, were established. Coding regions of murine cytokines were cloned into the 5' end of the CVB3 wild type (CVB3wt) open reading frame and were supplied with an artificial viral 3Cpro-specific Q-G cleavage site. Correct processing releases active cytokines, and the concentration of IFN-gamma and IL-10 was analyzed by enzyme-linked immunosorbent assay and bioassays. In mice, CVB3wt was detectable in pancreas and heart tissue, causing massive destruction of the exocrine pancreas as well as myocardial inflammation and heart cell lysis. Most of the CVB3wt-infected mice revealed virus-associated symptoms, and some died within 28 days postinfection. In contrast, CVB3rec variants were present only in the pancreas of infected mice, causing local inflammation with subsequent healing. Four weeks after the first infection, surviving mice were challenged with the lethal CVB3H3 variant, causing casualties in the CVB3wt- and CVB3(muIL-10)-infected groups, whereas almost none of the CVB3(IFN-gamma)- and CVB3(IL-10)-infected mice died and no pathological disorders were detectable. This study demonstrates that expression of immunoregulatory cytokines during CVB3 replication simultaneously protects mice against a lethal disease and prevents virus-caused tissue destruction.

FIG. 1

Sequences encoding murine IFN-γ or IL-10 were cloned at the start of the translation together with an artificial cleavage site for 3Cpro (A). The insertion of additional nucleotide sequences caused reduced growth capacities of the CVB3rec variants, which were analyzed by standard plaque formation assays on GMK cell monolayers (B) and by one-step growth curve experiments (C).

FIG. 2

One day p.i., the expression of virus-produced cytokines was determined by immunohistochemistry. By using pancreas paraffin sections of BALB/c mice, the presence of viral protein (A), viral genomes (B), virus-produced IFN-γ (C), and virus-produced IL-10 (D) was analyzed and compared to that of noninfected controls (magnification, ×440). (E) Three days p.i., the CVB3rec variants remained genetically stable under in vivo conditions as shown in pancreas tissue samples by use of RT-PCR. The detection of viral genomes was performed using a primer pair which corresponds to positions 471 to 494 and 848 to 872 of the original cDNA, resulting in the following PCR products: CVB3wt, 340 bp; CVB3(IFN-γ), 740 bp; CVB3(IL-10), 820 bp; and CVB3(muIL-10), 838 bp.

FIG. 3

(A) Paraffin sections of pancreas tissue obtained from BALB/c mice were stained with hematoxylin and eosin 2 and 7 days following infection with CVB3wt or CVB3rec variants (magnification, ×200). (B) RT-PCR was performed. Replication of CVB3wt in heart tissue (lanes 1 to 3, virus) was accompanied by myocardial inflammation (magnification, ×200), whereas no viral genome was detectable in the heart tissue of CVB3rec-infected mice. Additional lanes: M, DNA molecular weight markers; 4 to 6 CVB3(IFN-γ); 7 to 9, CVB3(IL-10); 10 to 12, CVB3(muIL-10).

FIG. 4

Paraffin sections of pancreas tissue (stained with hematoxylin and eosin) as well as of heart tissue (stained with Sirius red; connective tissue is stained red) of BALB/c mice that were originally infected with CVB3wt (A) or CVB3(muIL-10) (D) but survived the lethal CVB3H3 challenge revealed massive destruction of the exocrine pancreas, as well as fibrosis in heart tissue 28 days postchallenge, whereas no pathological disorders were detectable in mice originally infected with CVB3(IFN-γ) (B) or CVB3(IL-10) (C). Magnification, ×220.

FIG. 5

Amount of infectious virus particles in pancreas (A) and heart tissue (B) of BALB/c mice 3 days after a lethal challenge with five LD₅₀s of CVB3H3 characterized by TCID₅₀ assays. Four weeks prior to challenge, mice were inoculated either with CVB3wt (column 2), CVB3(IFN-γ) (column 3), CVB3(IL-10) (column 4), or CVB3(muIL-10) (column 5) or remained noninfected (1). Experimental groups consisted of five mice, and experiments were repeated three times. The mean ± standard deviation is shown.