Roles of Tumor Necrosis Factor Alpha (TNF-) and th
http://www.100md.com
病菌学杂志 2005年第5期
Department of Pathology, University of Vermont, Burlington, Vermont
ABSTRACT
Giving C57BL/6 mice 104 PFU of coxsackievirus B3 (H3 variant) fails to induce myocarditis, but increasing the initial virus inoculum to 105 or 106 PFU causes significant cardiac disease. Virus titers in the heart were equivalent at days 3 and 7 in mice given all three virus doses, but day 3 titers in the pancreases of mice inoculated with 104 PFU were reduced. Tumor necrosis factor alpha (TNF-) concentrations in the heart were increased in all infected mice, but cytokine levels were highest in mice given the larger virus inocula. TNF-–/– and p55 TNF receptor-negative (TNFR–/–) mice developed minimal myocarditis compared to B6;129 or C57BL/6 control mice. p75 TNFR–/– mice were as disease susceptible as C57BL/6 animals. No significant differences in virus titers in heart or pancreas were observed between the groups, but C57BL/6 and p75 TNFR–/– animals showed 10-fold more inflammatory cells in the heart than p55 TNFR–/– mice, and the cell population was comprised of high concentrations of CD4+ gamma interferon-positive and V4+ cells. Cardiac endothelial cells isolated from C57BL/6 and p75 TNFR–/– mice upregulate CD1d, the molecule recognized by V4+ cells, but infection of TNF–/– or p55 TNFR–/– endothelial cells failed to upregulate CD1d. Infection of C57BL/6 endothelial cells with a nonmyocarditic coxsackievirus B3 variant, H310A1, which is a poor inducer of TNF-, failed to elicit CD1d expression, but TNF- treatment of H310A1-infected endothelial cells increased CD1d levels to those seen in H3-infected cells. TNF- treatment of uninfected endothelial cells had only a modest effect on CD1d expression, suggesting that optimal CD1d upregulation requires both infection and TNF- signaling.
INTRODUCTION
Cytokines clearly play an important role in myocarditis pathogenicity whether induced by immunization with cardiac myosin (experimental autoimmune myocarditis) or by virus infection (20-22, 24, 25, 29). Exogenous administration of tumor necrosis factor alpha (TNF-) and interleukin-1 (IL-1) promotes coxsackievirus B3 (CVB3)-induced myocarditis in otherwise disease-resistant mice (21, 22, 25), while neutralization of endogenous TNF- protects against myosin-induced myocarditis (27). The effect of TNF- in myocarditis is somewhat controversial, however, since TNF- protects against encephalomyocarditis virus (EMCV)-induced myocarditis (29) and EMCV-induced encephalitis (26). Differences in the role of TNF- between CVB3 and EMCV likely reflect the different roles of the immune system in diseases caused by the two picornaviruses. In the model of CVB3-induced myocarditis presented here, cardiac injury is dependent on V4+ cell responses which facilitate the activation of autoimmune CD8+ T cytolytic cells (18). These autoimmune CD8+ T cells are the dominant pathogenic effectors in myocarditis (13). Thus, any factors which promote CD1d expression in the heart should aggravate CVB3-induced heart damage. In contrast, CD1d-dependent mechanisms are protective in EMCV infection (5). In this communication, we demonstrate that TNF- is important in upregulating CD1d. This observation is consistent with TNF- promoting pathogenicity in CVB3 infections, where CD1d-dependent responses cause autoimmunity induction, and with TNF- protecting in EMCV infections, where CD1d-dependent responses are antiviral.
MATERIALS AND METHODS
Mice. C57BL/6 (B6), B6129SF2/J (B6;129), B6129-Tnfrsf1atm1Mak (p55 TNFR–/–), B6129S2-Tnfrsf1btm1Mwm (p75 TNFR–/–), and B6129S6-Tnftm1Gkl (TNF-–/–) mice were purchased from Jackson Laboratories, Bar Harbor, Maine. Male mice, 5 to 7 weeks of age, were used for infection with CVB3 and endothelial cell isolation.
Virus and virus titrations. The H3 and H310A1variants of CVB3 were used (19). Adult mice were infected by intraperitoneal injection with 104, 105, or 106 PFU of virus in phosphate-buffered saline (PBS) as indicated below. Animals were killed either 3 or 7 days after infection. Hearts and/or pancreases were removed. For some experiments, whole organs (heart and pancreas) from individual mice were homogenized in 0.9 ml of RPMI 1640 medium. Debris was removed by centrifugation at 300 x g, and the supernatant volume was measured. For virus titration, 0.1 ml of the supernatant was serially diluted (1:10), and titers were determined by using the plaque-forming assay (19). Virus titers per heart or pancreas were mathematically determined by multiplying the titer in 0.1 ml of supernatant by the total supernatant volume. The remaining supernatant was used for TNF- determination as described below. In other experiments, hearts were weighed and then divided in two. Half of the heart was weighed and then homogenized for virus titration, while the remaining half was formalin fixed for histology. Enhanced green fluorescent protein (eGFP)-H3 virus was kindly supplied by Lindsay Whitton, Scripps Research Institute, La Jolla, Calif. (8).
ELISA. Supernatant from homogenized heart was evaluated for TNF- by using a commercial enzyme-linked immunosorbent assay (ELISA) kit according to the directions of the manufacturer (Endogen; Pierce, Rockford, Ill.). Cytokine concentrations per heart were determined by multiplying the TNF- in the 50 μl of supernatant used for the ELISA by the total supernatant volume.
Histology. Hearts were fixed in 10% buffered formalin, sectioned, stained with hematoxylin and eosin, and evaluated for myocarditis by image analysis as described previously (19).
Intracellular cytokine staining. The details of intracellular cytokine staining have been published previously (11). Briefly, peripheral blood was taken from euthanatized mice by cardiac puncture into EDTA-PBS. For heart inflammatory cells, individual hearts were perfused with PBS and then minced finely and subjected to four sequential enzymatic digestions with 0.4% collagenase II (Sigma) and 0.25% pancreatin (Sigma) for 8 min at 37°C for each digestion. Lymphoid cells were isolated by centrifugation on Histopaque (Sigma Chemical Co., St. Louis, Mo.). Cells (105) were cultured for 4 h in RPMI 1640 medium containing 10% fetal bovine serum, 10 μg of brefeldin A (BFA) per ml, 50 ng of phorbol myristate acetate per ml, and 500 ng of ionomycin (Sigma) per ml. After culture, the cells were washed in PBS-1% bovine serum albumin (BSA) (Sigma) containing BFA and incubated on ice for 30 min in PBS-BSA-BFA containing a 1:100 dilution of Fc Block and Cy-chrome-conjugated anti-mouse CD4 (clone GK1.5 [PharMingen] or Cy-chrome-rat immunoglobulin G1 [IgG1] [isotype control; clone R3-34]). The cells were washed once with PBS-BSA-BFA; fixed in 2% paraformaldehyde for 10 min; resuspended in PBS-BSA containing 0.5% saponin, Fc Block, and 1:100 dilutions of fluorescein isothiocyanate (FITC)-conjugated anti-IFN- and phycoerythrin (PE)-conjugated anti-IL-4 (clones XMG 1.2 and BVD4-1D11) or FITC-conjugated and PE-conjugated rat IgG1 (clone R3-34); and incubated for 30 min on ice. The cells were washed once in PBS-BSA-saponin and once in PBS-BSA and then resuspended in 2% paraformaldehyde and analyzed with a Coulter Epics Elite flow cytometer with a single excitation wavelength (488 nm) and band filters for Cy-chrome (670 nm), FITC (525 nm), and PE (575 nm). The cell population was classified for cell size (forward scatter) and complexity (side scatter). At least 10,000 cells were evaluated. Positive staining was determined relative to isotype controls.
For analysis of cell surface molecule expression, inflammatory cells from the heart were labeled with the following antibodies from BD Biosciences/PharMingen: PE-Cy7-anti-NK1.1 (clone PK136), PE-anti-CD8a (clone OX-8), FITC-anti-V4 (clone UC3-10A6), and peridinin chlorophyll a protein-Cy5.5-anti-CD11b (Mac-1 chain; clone M1/70). Band filters were 780 nm for PE-Cy7 and 695 nm for peridinin chlorophyll a protein-Cy5.5 using long pass filters of 735 and 685 nm, respectively.
Endothelial cell isolation and infection. Hearts and aortas were aseptically removed from eight euthanatized mice per group, pooled, perfused with 10 ml of warmed PBS, and then flushed slowly with 10 ml of 0.25% trypsin in PBS (37°C) over a 15-min period. The trypsinized cells were collected, washed once in RPMI 1640-10% fetal bovine serum (FBS), and then cultured on Matrigel (BD Biosciences)-coated 25-cm2 tissue culture flasks (Fischer) in medium with 10% FBS, 5% horse serum, antibiotic antimycotic solution (Sigma), and L-glutamine. When confluent, cells were freed by using Dispase (BD-Biosciences) and then reseeded onto Matrigel-coated flasks at a 1:4 dilution. Cultures were used only to the third passage. Cells (105)were plated in 96-well tissue culture plates coated with Matrigel overnight in a 37°C humidified CO2 incubator. Cells were infected with virus at a multiplicity of infection (MOI) of 100 for 30 min, washed once with medium, incubated with 1 μg of monoclonal anti-CVB3 (clone 8A6) (14) for 10 min, washed twice with medium, and then incubated for up to 24 h in medium with 10% FBS. Control cultures were not infected. Some cultures were also treated with 0 to 200 ng of TNF- (PharMingen) per ml. Cells were recovered by Dispase treatment; washed; incubated with a 1:100 dilution of PE-anti-CD1 (clone 1B1l PharMingen), biotinylated anti-CD31 (PE-CAM-1) (clone MEC 13.3; PharMingen), and Cy-chrome-streptavidin (PharMingen) for 30 min; washed once with PBS-BSA; and resuspended in 2% paraformaldehyde for flow analysis. CD31 is a marker used to identify vascular endothelial cells (2).
GFP-CVB3 can be used only when infecting cells with H3 virus, since the construct used the H3 plasmid (8). For experiments using the H310A1 virus, isolated endothelial cells were labeled with anti-CD31 and anti-CD1d and then fixed for 10 min with 2% paraformaldehyde. The cells were resuspended in PBS-BSA-saponin containing 1 μl of monoclonal anti-CVB3 (clone 8A6; IgM isotype (14) and Alexa 488 anti-mouse IgM.
Determination of cell death. The procedure for determination of cell death was adapted from that of Choi et al. (3). Endothelial cells were infected with H3 or H310A1 virus as described above. Uninfected and H310A1-infected cells were also treated with 200 ng of TNF- per ml for 24 h. Endothelial cells were isolated, washed with PBS, and resuspended in 200 μl of binding buffer (10 mM HEPES [pH 7.4], 140 mM NaCl, and 2.5 mM CaCl2). To each sample, 5 μl of FITC-annexin V (PharMingen) was added, and the samples were incubated at room temperature for 10 min. Propidium iodide (PI) was added to the cells at a final concentration of 1 μg/ml, and the cells were analyzed by flow cytometry.
Statistics. Data were analyzed for skewness and kurtosis by using the SPSS for Windows program (version 11.0, 2001; SPSS, Inc., Chicago, Ill.). Statistical analysis was done by Wilcoxon ranked score, since the variance was not normally distributed in many groups.
RESULTS
Increased virus inoculum in C57BL/6 mice enhances myocarditis susceptibility and correlates with TNF- induction but does not result in increased cardiac virus titers. Previously, we have reported that C57BL/6 mice are relatively resistant to H3 virus-induced myocarditis when infected with 104 PFU of H3 virus, a dose which causes severe myocarditis in BALB/c mice (9, 12). These studies investigated whether increasing the initial virus inoculum caused greater myocarditis susceptibility. Male C57BL/6 mice were injected with either 104, 105, or 106 PFU of H3 virus. Each group consisted of 8 mice per dose for uninfected animals and for animals 3 days after infection and of between 8 and 20 mice per group for mice 7 days after infection. Hearts were removed from four mice per group on days 0 and 3 for histology, and the remaining mice (four mice per group) provided hearts and pancreases, which were homogenized and used determination of virus titers and TNF- evaluation by ELISA. Half of the mice surviving to day 7 were evaluated for histology, and the remaining animals were used to provide hearts and pancreases for virus titer determination and TNF- analysis (Fig. 1). There was no mortality in any group by day 3 after infection. By day 7, 0 of 8 mice infected with 104 PFU, 9 of 20 mice infected with 105 PFU, and 11 of 20 mice infected with 106 PFU had died. Few inflammatory cells were observed in the myocardium before day 7. Mice injected with both 105 and 106 PFU of virus developed statistically equivalent amounts of myocarditis by day 7, while animals infected with the lowest dose (104 PFU) had significantly less myocarditis than animals given the higher inocula. TNF- concentrations were significantly increased in all virus-infected mice compared to uninfected animals, but the levels in hearts of mice given the dose of either 105 or 106 PFU were higher than those in mice injected with 104 PFU. TNF- levels were elevated by day 3 despite minimal cardiac inflammation and were decreasing by day 7. Despite the substantial differences in myocarditis, cardiac virus titers were statistically equivalent between the groups given 104, 105, and 106 PFU at both days 3 and 7 after infection. In contrast, there was significantly more virus in the pancreases of mice given 105 and 106 PFU than in animals receiving 104 PFU, suggesting that increasing the initial virus inoculum may have distinct effects in different organs.
Myocarditis in transgenic mice. The rapid induction of TNF- in the heart concurs with published reports that this cytokine is important in myocarditis susceptibility (21, 22). Male C57BL/6, B6;129, TNF-–/–, p55 TNFR–/–, and p75 TNFR–/– mice were infected with 105 PFU of H3 virus and killed either 3 (four mice per group) or 7 (10 or more mice per group) days later. B6;129F2 mice were used as the control for the TNF-–/– strain, while C57BL/6 mice were controls for the p55 TNFR–/– and p75 TNFR–/– strains. No mortality occurred in the day 3 groups. Mortality at day 7 was 9 of 18 for C57BL/6, 5 of 10 for B6;129F2, 2 of 10 for TNF-–/–, 1 of 10 for p55 TNFR–/–, and 3 of 10 for p75 TNFR–/– mice. Whole hearts and pancreases were weighed and used for determination of virus titers on day 3. Organs obtained on day 7 were weighed and divided for histology and determination of virus titers. For these studies, titers are reported as PFU per gram of tissue. Blood was obtained 7 days after infection. Lymphoid cells were isolated, stimulated with phorbol myristate acetate-ionomycin-brefeldin A, and then stained to determine the percentage that were positive for CD4, IFN-, and IL-4. TNF-–/– mice had significantly less myocarditis and fewer CD4+ IFN-+ cells than B6;129F2 controls (Fig. 2). p55 TNFR–/– mice also developed significantly less myocarditis than wild-type C57BL/6 mice, while there was no difference in myocarditis between C57BL/6 and p75 TNFR–/– animals, indicating that this receptor does not participate in the disease process. CD4+ IFN-+ cells were decreased in p55 TNFR–/– mice compared to either C57BL/6 (P < 0.05) or p75 TNFR–/– (P < 0.05) animals, which concurs with studies associating CD4+ IFN-+ responses with myocarditis susceptibility (16). Although TNF- has been reported to promote encephalomyocarditis virus clearance, no significant difference was observed between cardiac or pancreatic virus titers in the different animal groups.
Although peripheral blood CD4+ Th1 (IFN-+) responses correlate to myocarditis susceptibility, the numbers of these cells in peripheral blood are modest. Confirmation that p55 TNFR–/– show reduced CD4Th1 in the heart is shown in Fig. 3. Inflammatory cells were isolated from four individual mice per group at 7 days after infection. There were approximately 10-fold more inflammatory cells in infected C57BL/6 and p75 TNFR–/– mice (106 cells per heart) than in the p55 TNFR–/– mice (105 cells/heart). Approximately 25% of inflammatory cells in C57BL/6 and p75 TNFR–/– mice were CD4+ IFN-+, compared to fewer than 1% of p55 TNFR–/– inflammatory cells. In contrast, 61% of inflammatory cells in p55 TNFR–/– mice were CD11b+ (Mac-1), compared to approximately 25% of cells from C57BL/6 and p75 TNFR–/– animals. The other major difference is that high proportions of cells from C57BL/6 and p75 TNFR–/– animals were V4+ and CD8+, in contrast to the case for p55 TNFR–/– animals, which had none of these cells. All CD8+ cells were also V4+. Approximately half of the V4+ cells were negative for both CD8 (Fig. 3) and CD4 (data not shown).
Figure 4 gives representative histology of the cardiac inflammation in the control (C57BL/6 and B6;129F2) and transgenic (TNF-–/–, p55 TNFR–/–, and p75 TNFR–/–) mice. Both control strains and p75 TNFR–/– mice showed substantial interstitial inflammation, while few inflammatory cells were obvious in TNF-–/– or p55 TNFR–/– mice.
TNF- and the p55 TNFR are necessary for CD1d induction in infected endothelial cells. V4+ T cells selectively infiltrate the hearts of myocarditis-susceptible mice, and previous studies have shown that this population is CD1d restricted (15, 16). The lack of V4+ cells in p55 TNFR–/– mice provided circumstantial evidence that TNF- might regulate CD1d expression during CVB3 infection. The roles of TNF- and virus infection in CD1d induction were investigated. First, vascular endothelial cells were isolated from the hearts and aortas of C57BL/6 mice. The purity of the population was determined by staining with anti-CD31, an endothelial cell marker (2), and demonstrated that the cell population contained 89.5% CD31+ cells. Endothelial cells were either uninfected or infected with H3 or H310A1 virus at an MOI of 100 and then were cultured for 0, 12, and 24 h (Fig. 5). Supernatants were retrieved and assayed for TNF-. Uninfected endothelial cells released no cytokine during the 24 h. Both H3- and H310A1-infected cells produced TNF-, but approximately eightfold more cytokine was released by the H3- compared to the H310A1-infected cells. Next, endothelial cells were isolated from C57BL/6 and transgenic mice and then infected with eGFP-CVB3 at an MOI of 100. Figure 6 shows that virus infection does not upregulate CD1d expression in endothelial cell cultures from TNF-–/– or p55 TNFR–/– transgenic mice. Endothelial cells from C57BL/6 and p75 TNFR–/– mice were fully capable of upregulating CD1d when infected. Although this study shows that TNF- is necessary, it does not indicate whether this cytokine is sufficient for CD1d induction. To investigate this point, C57BL/6 endothelial cells were either uninfected or infected with either the H3 (myocarditic) or H310A1 (nonmyocarditic) variant at an MOI of 100. Half of endothelial cell cultures were treated with 200 ng of TNF- per ml. Figure 7 shows that either TNF- treatment or H310A1 virus infection of C57BL/6 endothelial cells alone only modestly upregulated expression of CD1d. However, TNF- treatment of H310A1-infected cells induced amounts of CD1 equivalent to those in H3 infection.
Figure 8 shows a dose response for TNF- induction of CD1d in H310A1-infected C57BL/6 endothelial cells. Cells were infected with H310A1 at an MOI of 100 and treated with between 0 and 200 ng of cytokine per ml for 24 h. Cells were retrieved and labeled with anti-CD31, anti-CD1d, and anti-CVB3. With gating on the CD31+ (endothelial cell) population, the percentage of endothelial cells which were infected and expressed CD1d was determined and is shown in the upper panel of Fig. 8; examples of the flow cytometry for infected cells treated with 0 and 200 ng of TNF- per ml are shown in the lower panel. Upregulation of CD1d was dependent on the dose of cytokine used, with the maximal effect at 200 ng/ml. The percentage of infected cells did not statistically differ irrespective of the amount of TNF- used.
Effect of infection and TNF- on cell death. Both infection and TNF- might induce cell death. If there were differences in the percentage of cells dying in each culture, this might give a false impression of increased CD1d expression when CD1d+ cells may survive preferentially to the CD1d– population. Uninfected, infected, uninfected and TNF--treated, and infected and TNF--treated cells were produced as described above. The cells were labeled with annexin V and PI to determine live and dead cells (Fig. 9). Only 3.7% of uninfected endothelial cells were dead (annexin V+ PI+), compared to 18.6 and 24.2% in H3- and H310A1-infected cultures, respectively. TNF- treatment, either alone or with H310A1, did cause greater cell death than infection alone, but this slight increase does not correlate with the pattern of CD1d expression.
DISCUSSION
TNF- and IL-1? are proinflammatory cytokines which strongly promote myocarditis susceptibility in mice (21, 22). Exogenous administration of these cytokines causes myocarditis-resistant mice to develop the disease. This communication shows that TNF- promotes susceptibility to coxsackievirus B3-induced myocarditis through the p55 TNFR. The effect on myocarditis susceptibility is independent of virus titers, either early (day 3) or late (day 7), in the pancreas or heart. One of the interesting observations is that while C57BL/6 mice infected with relatively small amounts of virus (104 PFU/mouse) are myocarditis resistant, as has been shown previously (9), increasing the initial virus inoculum to 105 or 106 PFU per mouse results in substantial myocarditis susceptibility. The increased disease susceptibility is not apparently due to increasing cardiac virus titers, since mice given all three virus doses showed equivalent virus titers in the heart at 3 and 7 days after infection. Rather, increasing susceptibility with inoculum size causes substantial augmentation of TNF- levels in the heart even 3 days after infection, when inflammatory cells are not extensive. Analysis of the infiltrating cell populations in C57BL/6, p55 TNFR–/–, and p75 TNFR–/– mice shows that the C57BL/6 and p75 TNFR–/– animals have significant infiltration with CD4+ IFN-+, Mac-1+, and V4+ cells. Only the Mac-1+ cells predominate in the myocarditis-resistant p55 TNFR–/– mice.
It is therefore intriguing that two closely related CVB3 variants differ markedly in their ability to induce myocarditis and also vary markedly in their ability to upregulate TNF- expression. The H3 and H310A1 variants differ by a single nonconserved mutation at amino acid 165 in the VP2 protein (19). This asparagine (H3)-to-aspartate (H310A1) mutation occurs in a VP2 puff region associated in CVB3 with virus receptor (decay-accelerating factor [DAF]) binding (1, 23). The H310A1 variant was derived from the H3 variant as an antibody escape mutant by using a monoclonal antibody (clone 10A1) to the virus receptor (28, 30). H310A1 has significantly reduced receptor binding avidity compared to the H3 virus. Since antibody cross-linking of DAF activates tyrosine kinase pathways (4), it is highly likely that the H3 and H310A1 variants differ in their ability to transduce signals through DAF. Both virus and TNF- signal pathways are necessary for CD1d induction, since treating endothelial cells with the cytokine alone poorly upregulated this molecule, while both TNF- treatment and H310A1 infection strongly induced CD1d expression. H3-infected cells already express high levels of CD1d compared to H310A1 virus-infected cells, and extra TNF- had minimal effect on expression.
T cells expressing the V4+ T-cell receptor have been previously shown to play an essential role in myocarditis susceptibility (15, 16, 18). These effectors selectively recognize CD1d, a major histocompatibility complex class 1-like molecule which is important in innate immunity (5-7). CD1d–/– mice infected with H3 virus develop minimal myocarditis compared to CD1d+/+ animals and fail to activate V4+ cells (15). The V4+ cells infiltrate the H3 virus-infected myocardium within 3 days of virus injection, making these effectors the earliest T cells entering the heart (17). They are highly effective in killing infected but not uninfected cardiac myocytes, and this cytotoxicity is blocked by antibody to CD1d but not by antibodies to the classical major histocompatibility complex class I (H2K/D) and class II (H2 IA/IE) antigens (10, 15). Thus, rapid infiltration of infected organs by these innate effectors could aid in rapid control of virus prior to development of the adaptive immune response. This may be a reason why increasing the virus inoculum fails to result in increased cardiac virus titers. If V4+ cells infiltrate the heart at the higher virus inocula but not at the lower dose, the V4+ cells may be effective in limiting virus expansion.
ACKNOWLEDGMENTS
We thank Lindsay Whitton of Scripps Research Institute, La Jolla, Calif., for kindly supplying the eGFP-CVB3. We also thank Colette Charland for help with the flow cytometry and Kevin Kolinich for help in preparing the manuscript.
This work was supported by grants HL58583 and P01 AI45666 from the National Institutes of Health.
REFERENCES
Bergelson, J. M., J. G. Mohanty, R. L. Crowell, N. F. St. John, D. M. Lublin, and R. W. Finberg. 1995. Coxsackievirus B3 adapted to growth in RD cells binds to decay-accelerating factor (CD55). J. Virol. 69:1903-1906.
Chen, S., M. Sega, and A. Agarwal. 2004. "Lumen digestion" technique for isolation of aortic endothelial cells from heme oxygenase-1 knockout mice. BioTechniques 37:84-86, 88-89.
Choi, J. J., C. F. Reich III, and D. S. Pisetsky. 2004. Release of DNA from dead and dying lymphocyte and monocyte cell lines in vitro. Scand J Immunol 60:159-166.
Davis, L. S., S. S. Patel, J. P. Atkinson, and P. E. Lipsky. 1988. Decay-accelerating factor functions as a signal transducing molecule for human T cells. J. Immunol. 141:2246-2252.
Exley, M., N. Bigley, O. Cheng, S. Tahir, S. Smiley, Q. Carter, H. Stills, M. Grusby, Y. Koezuka, M. Taniguchi, and S. Balk. 2001. CD1d-reactive T-cell activation leads to amelioration of disease caused by diabetogenic encephalomyocarditis virus. J. Leukoc. Biol. 69:713-718.
Exley, M., S. Porcelli, M. Furman, J. Garcia, and S. Balk. 1998. CD161 (NKR-P1A) costimulation of CD1d-dependent activation of human T cells expressing the invariant Valpha 24Jalpha Q T cell receptor alpha chains. J. Exp. Med. 188:867-876.
Exley, M. A., Q. He, O. Cheng, R. J. Wang, C. P. Cheney, S. P. Balk, and M. J. Koziel. 2002. Cutting edge: compartmentalization of Th1-like noninvariant CD1d-reactive T cells in hepatitis C virus-infected liver. J. Immunol. 168:1519-1523.
Feuer, R., I. Mena, R. Pagarigan, M. K. Slifka, and J. L. Whitton. 2002. Cell cycle status affects coxsackievirus replication, persistence, and reactivation in vitro. J. Virol. 76:4430-4440.
Henke, A., S. Huber, A. Stelzner, and J. Whitton. 1995. The role of CD8+ T lymphocytes in coxsackievirus B3-induced myocarditis. J. Virol. 69:6720-6728.
Huber, S. 2000. T cells expressing the gamma-delta T cell receptor induce apoptosis in cardiac myocytes. Cardiovasc. Res. 45:579-587.
Huber, S., D. Graveline, W. Born, and R. O'Brien. 2001. Cytokine production by V+ T-cell subsets is an important factor determining CD4+ Th cell phenotype and susceptibility of BALB/c mice to coxsackievirus B3-induced myocarditis. J. Virol. 75:5860-5868.
Huber, S., D. Graveline, M. Newell, W. Born, and R. O'Brien. 2000. Vgamma1+ T cells suppress and Vgamma4+ T cells promote susceptibility to coxsackievirus B3-induced myocarditis in mice. J. Immunol. 165:4174-4181.
Huber, S., and P. Lodge. 1984. Coxsackievirus B3 myocarditis in Balb/c mice: evidence for autoimmunity to myocyte antigens. Am. J. Pathol. 116:21.
Huber, S., A. Moraska, and M. Cunningham. 1994. Alterations in major histocompatibility complex association of myocarditis induced by coxsackievirus B3 (CVB3) mutants selected with monoclonal antibodies to group A streptococci. Proc. Natl. Acad. Sci. USA 91:5543-5547.
Huber, S., D. Sartini, and M. Exley. 2003. Role of CD1d in coxsackievirus B3-induced myocarditis. J. Immunol. 170:3147-3153.
Huber, S., C. Shi, and R. C. Budd. 2002. Gammadelta T cells promote a Th1 response during coxsackievirus B3 infection in vivo: role of Fas and Fas ligand. J. Virol. 76:6487-6494.
Huber, S. A. 2004. Increased susceptibility of male BALB/c mice to coxsackievirus B3-induced myocarditis: role for CD1d. Med Microbiol Immunol (Berlin), in press.
Huber, S. A., D. Sartini, and M. Exley. 2002. V4+ T cells promote autoimmune CD8+ cytolytic T-lymphocyte activation in coxsackievirus B3-induced myocarditis in mice: role for CD4+ Th1 cells. J. Virol. 76:10785-10790.
Knowlton, K. U., E. S. Jeon, N. Berkley, R. Wessely, and S. Huber. 1996. A mutation in the puff region of VP2 attenuates the myocarditic phenotype of an infectious cDNA of the Woodruff variant of coxsackievirus B3. J. Virol. 70:7811-7818.
Kumar, A., V. Thota, L. Dee, J. Olson, E. Uretz, and J. E. Parrillo. 1996. Tumor necrosis factor alpha and interleukin 1beta are responsible for in vitro myocardial cell depression induced by human septic shock serum. J. Exp. Med. 183:949-958.
Lane, J., D. Neumann, A. Lafond-Walker, A. Herkowitz, and N. Rose. 1992. Interleukin 1 or tumor necrosis factor can promote coxsackie B3-induced myocarditis in resistant B10.A mice. J. Exp. Med. 175:1123-1129.
Lane, J., D. Neumann, A. Lanfond-Walker, A. Herskowitz, and N. Rose. 1993. Role of IL-1 and tumor necrosis factor in coxsackievirus-induced autoimmune myocarditis. J. Immunol. 151:1682-1690.
Lindberg, A., R. Crowell, R. Zell, R. Kandolf, and U. Pettersson. 1992. Mapping the RD phenotype of the Nancy strain of coxsackievirus B3. Virus Res. 24:187-196.
Meldrum, D. R. 1998. Tumor necrosis factor in the heart. Am. J. Physiol. 274:R577-R595.
Neumann, D., G. Allen, C. Narins, N. Rose, and A. Herskowitz. 1993. Heart autoantibodies in human myocarditis and cardiomyopathy: virus and skeletal muscle cross-reactivity, p. 325-334. In H. Figulla, R. Kandolf, and B. McManus (ed.), Idiopathic dilated cardiomyopathy. Springer-Verlag, Berlin, Germany.
Su, X.-M., and S. Sriram. 1991. Treatment of chronic relapsing experimental allergic encephalomyelitis with the intravenous administration of splenocytes coupled to encephalitogenic peptide 91-103 of myelin basic protein. J. Neuroimmunol. 34:181.
Takehana, H., T. Inomata, H. Niwano, M. Nishii, C. Matsuda, K. Kohno, Y. Machida, and T. Izumi. 2002. Immunomodulatory effect of pentoxifylline in suppressing experimental autoimmune myocarditis. Circ. J. 66:499-504.
Van Houten, N., P. Bouchard, A. Moraska, and S. Huber. 1991. Selection of an attenuated coxsackievirus B3 variant using a monoclonal antibody reactive to myocyte antigen. J. Virol. 65:1286-1290.
Wada, H., K. Saito, T. Kanda, I. Kobayashi, H. Fujii, S. Fujigaki, N. Maekawa, H. Takatsu, H. Fujiwara, K. Sekikawa, and M. Seishima. 2001. Tumor necrosis factor-alpha (TNF-alpha) plays a protective role in acute viral myocarditis in mice: a study using mice lacking TNF-alpha. Circulation 103:743-749.
Weller, A., K. Simpson, M. Herzum, N. Van Houten, and S. Huber. 1989. Coxsackievirus B3 induced myocarditis: virus receptor antibodies modulate myocarditis. J. Immunol. 143:1843-1850.(S. A. Huber and D. Sartin)
ABSTRACT
Giving C57BL/6 mice 104 PFU of coxsackievirus B3 (H3 variant) fails to induce myocarditis, but increasing the initial virus inoculum to 105 or 106 PFU causes significant cardiac disease. Virus titers in the heart were equivalent at days 3 and 7 in mice given all three virus doses, but day 3 titers in the pancreases of mice inoculated with 104 PFU were reduced. Tumor necrosis factor alpha (TNF-) concentrations in the heart were increased in all infected mice, but cytokine levels were highest in mice given the larger virus inocula. TNF-–/– and p55 TNF receptor-negative (TNFR–/–) mice developed minimal myocarditis compared to B6;129 or C57BL/6 control mice. p75 TNFR–/– mice were as disease susceptible as C57BL/6 animals. No significant differences in virus titers in heart or pancreas were observed between the groups, but C57BL/6 and p75 TNFR–/– animals showed 10-fold more inflammatory cells in the heart than p55 TNFR–/– mice, and the cell population was comprised of high concentrations of CD4+ gamma interferon-positive and V4+ cells. Cardiac endothelial cells isolated from C57BL/6 and p75 TNFR–/– mice upregulate CD1d, the molecule recognized by V4+ cells, but infection of TNF–/– or p55 TNFR–/– endothelial cells failed to upregulate CD1d. Infection of C57BL/6 endothelial cells with a nonmyocarditic coxsackievirus B3 variant, H310A1, which is a poor inducer of TNF-, failed to elicit CD1d expression, but TNF- treatment of H310A1-infected endothelial cells increased CD1d levels to those seen in H3-infected cells. TNF- treatment of uninfected endothelial cells had only a modest effect on CD1d expression, suggesting that optimal CD1d upregulation requires both infection and TNF- signaling.
INTRODUCTION
Cytokines clearly play an important role in myocarditis pathogenicity whether induced by immunization with cardiac myosin (experimental autoimmune myocarditis) or by virus infection (20-22, 24, 25, 29). Exogenous administration of tumor necrosis factor alpha (TNF-) and interleukin-1 (IL-1) promotes coxsackievirus B3 (CVB3)-induced myocarditis in otherwise disease-resistant mice (21, 22, 25), while neutralization of endogenous TNF- protects against myosin-induced myocarditis (27). The effect of TNF- in myocarditis is somewhat controversial, however, since TNF- protects against encephalomyocarditis virus (EMCV)-induced myocarditis (29) and EMCV-induced encephalitis (26). Differences in the role of TNF- between CVB3 and EMCV likely reflect the different roles of the immune system in diseases caused by the two picornaviruses. In the model of CVB3-induced myocarditis presented here, cardiac injury is dependent on V4+ cell responses which facilitate the activation of autoimmune CD8+ T cytolytic cells (18). These autoimmune CD8+ T cells are the dominant pathogenic effectors in myocarditis (13). Thus, any factors which promote CD1d expression in the heart should aggravate CVB3-induced heart damage. In contrast, CD1d-dependent mechanisms are protective in EMCV infection (5). In this communication, we demonstrate that TNF- is important in upregulating CD1d. This observation is consistent with TNF- promoting pathogenicity in CVB3 infections, where CD1d-dependent responses cause autoimmunity induction, and with TNF- protecting in EMCV infections, where CD1d-dependent responses are antiviral.
MATERIALS AND METHODS
Mice. C57BL/6 (B6), B6129SF2/J (B6;129), B6129-Tnfrsf1atm1Mak (p55 TNFR–/–), B6129S2-Tnfrsf1btm1Mwm (p75 TNFR–/–), and B6129S6-Tnftm1Gkl (TNF-–/–) mice were purchased from Jackson Laboratories, Bar Harbor, Maine. Male mice, 5 to 7 weeks of age, were used for infection with CVB3 and endothelial cell isolation.
Virus and virus titrations. The H3 and H310A1variants of CVB3 were used (19). Adult mice were infected by intraperitoneal injection with 104, 105, or 106 PFU of virus in phosphate-buffered saline (PBS) as indicated below. Animals were killed either 3 or 7 days after infection. Hearts and/or pancreases were removed. For some experiments, whole organs (heart and pancreas) from individual mice were homogenized in 0.9 ml of RPMI 1640 medium. Debris was removed by centrifugation at 300 x g, and the supernatant volume was measured. For virus titration, 0.1 ml of the supernatant was serially diluted (1:10), and titers were determined by using the plaque-forming assay (19). Virus titers per heart or pancreas were mathematically determined by multiplying the titer in 0.1 ml of supernatant by the total supernatant volume. The remaining supernatant was used for TNF- determination as described below. In other experiments, hearts were weighed and then divided in two. Half of the heart was weighed and then homogenized for virus titration, while the remaining half was formalin fixed for histology. Enhanced green fluorescent protein (eGFP)-H3 virus was kindly supplied by Lindsay Whitton, Scripps Research Institute, La Jolla, Calif. (8).
ELISA. Supernatant from homogenized heart was evaluated for TNF- by using a commercial enzyme-linked immunosorbent assay (ELISA) kit according to the directions of the manufacturer (Endogen; Pierce, Rockford, Ill.). Cytokine concentrations per heart were determined by multiplying the TNF- in the 50 μl of supernatant used for the ELISA by the total supernatant volume.
Histology. Hearts were fixed in 10% buffered formalin, sectioned, stained with hematoxylin and eosin, and evaluated for myocarditis by image analysis as described previously (19).
Intracellular cytokine staining. The details of intracellular cytokine staining have been published previously (11). Briefly, peripheral blood was taken from euthanatized mice by cardiac puncture into EDTA-PBS. For heart inflammatory cells, individual hearts were perfused with PBS and then minced finely and subjected to four sequential enzymatic digestions with 0.4% collagenase II (Sigma) and 0.25% pancreatin (Sigma) for 8 min at 37°C for each digestion. Lymphoid cells were isolated by centrifugation on Histopaque (Sigma Chemical Co., St. Louis, Mo.). Cells (105) were cultured for 4 h in RPMI 1640 medium containing 10% fetal bovine serum, 10 μg of brefeldin A (BFA) per ml, 50 ng of phorbol myristate acetate per ml, and 500 ng of ionomycin (Sigma) per ml. After culture, the cells were washed in PBS-1% bovine serum albumin (BSA) (Sigma) containing BFA and incubated on ice for 30 min in PBS-BSA-BFA containing a 1:100 dilution of Fc Block and Cy-chrome-conjugated anti-mouse CD4 (clone GK1.5 [PharMingen] or Cy-chrome-rat immunoglobulin G1 [IgG1] [isotype control; clone R3-34]). The cells were washed once with PBS-BSA-BFA; fixed in 2% paraformaldehyde for 10 min; resuspended in PBS-BSA containing 0.5% saponin, Fc Block, and 1:100 dilutions of fluorescein isothiocyanate (FITC)-conjugated anti-IFN- and phycoerythrin (PE)-conjugated anti-IL-4 (clones XMG 1.2 and BVD4-1D11) or FITC-conjugated and PE-conjugated rat IgG1 (clone R3-34); and incubated for 30 min on ice. The cells were washed once in PBS-BSA-saponin and once in PBS-BSA and then resuspended in 2% paraformaldehyde and analyzed with a Coulter Epics Elite flow cytometer with a single excitation wavelength (488 nm) and band filters for Cy-chrome (670 nm), FITC (525 nm), and PE (575 nm). The cell population was classified for cell size (forward scatter) and complexity (side scatter). At least 10,000 cells were evaluated. Positive staining was determined relative to isotype controls.
For analysis of cell surface molecule expression, inflammatory cells from the heart were labeled with the following antibodies from BD Biosciences/PharMingen: PE-Cy7-anti-NK1.1 (clone PK136), PE-anti-CD8a (clone OX-8), FITC-anti-V4 (clone UC3-10A6), and peridinin chlorophyll a protein-Cy5.5-anti-CD11b (Mac-1 chain; clone M1/70). Band filters were 780 nm for PE-Cy7 and 695 nm for peridinin chlorophyll a protein-Cy5.5 using long pass filters of 735 and 685 nm, respectively.
Endothelial cell isolation and infection. Hearts and aortas were aseptically removed from eight euthanatized mice per group, pooled, perfused with 10 ml of warmed PBS, and then flushed slowly with 10 ml of 0.25% trypsin in PBS (37°C) over a 15-min period. The trypsinized cells were collected, washed once in RPMI 1640-10% fetal bovine serum (FBS), and then cultured on Matrigel (BD Biosciences)-coated 25-cm2 tissue culture flasks (Fischer) in medium with 10% FBS, 5% horse serum, antibiotic antimycotic solution (Sigma), and L-glutamine. When confluent, cells were freed by using Dispase (BD-Biosciences) and then reseeded onto Matrigel-coated flasks at a 1:4 dilution. Cultures were used only to the third passage. Cells (105)were plated in 96-well tissue culture plates coated with Matrigel overnight in a 37°C humidified CO2 incubator. Cells were infected with virus at a multiplicity of infection (MOI) of 100 for 30 min, washed once with medium, incubated with 1 μg of monoclonal anti-CVB3 (clone 8A6) (14) for 10 min, washed twice with medium, and then incubated for up to 24 h in medium with 10% FBS. Control cultures were not infected. Some cultures were also treated with 0 to 200 ng of TNF- (PharMingen) per ml. Cells were recovered by Dispase treatment; washed; incubated with a 1:100 dilution of PE-anti-CD1 (clone 1B1l PharMingen), biotinylated anti-CD31 (PE-CAM-1) (clone MEC 13.3; PharMingen), and Cy-chrome-streptavidin (PharMingen) for 30 min; washed once with PBS-BSA; and resuspended in 2% paraformaldehyde for flow analysis. CD31 is a marker used to identify vascular endothelial cells (2).
GFP-CVB3 can be used only when infecting cells with H3 virus, since the construct used the H3 plasmid (8). For experiments using the H310A1 virus, isolated endothelial cells were labeled with anti-CD31 and anti-CD1d and then fixed for 10 min with 2% paraformaldehyde. The cells were resuspended in PBS-BSA-saponin containing 1 μl of monoclonal anti-CVB3 (clone 8A6; IgM isotype (14) and Alexa 488 anti-mouse IgM.
Determination of cell death. The procedure for determination of cell death was adapted from that of Choi et al. (3). Endothelial cells were infected with H3 or H310A1 virus as described above. Uninfected and H310A1-infected cells were also treated with 200 ng of TNF- per ml for 24 h. Endothelial cells were isolated, washed with PBS, and resuspended in 200 μl of binding buffer (10 mM HEPES [pH 7.4], 140 mM NaCl, and 2.5 mM CaCl2). To each sample, 5 μl of FITC-annexin V (PharMingen) was added, and the samples were incubated at room temperature for 10 min. Propidium iodide (PI) was added to the cells at a final concentration of 1 μg/ml, and the cells were analyzed by flow cytometry.
Statistics. Data were analyzed for skewness and kurtosis by using the SPSS for Windows program (version 11.0, 2001; SPSS, Inc., Chicago, Ill.). Statistical analysis was done by Wilcoxon ranked score, since the variance was not normally distributed in many groups.
RESULTS
Increased virus inoculum in C57BL/6 mice enhances myocarditis susceptibility and correlates with TNF- induction but does not result in increased cardiac virus titers. Previously, we have reported that C57BL/6 mice are relatively resistant to H3 virus-induced myocarditis when infected with 104 PFU of H3 virus, a dose which causes severe myocarditis in BALB/c mice (9, 12). These studies investigated whether increasing the initial virus inoculum caused greater myocarditis susceptibility. Male C57BL/6 mice were injected with either 104, 105, or 106 PFU of H3 virus. Each group consisted of 8 mice per dose for uninfected animals and for animals 3 days after infection and of between 8 and 20 mice per group for mice 7 days after infection. Hearts were removed from four mice per group on days 0 and 3 for histology, and the remaining mice (four mice per group) provided hearts and pancreases, which were homogenized and used determination of virus titers and TNF- evaluation by ELISA. Half of the mice surviving to day 7 were evaluated for histology, and the remaining animals were used to provide hearts and pancreases for virus titer determination and TNF- analysis (Fig. 1). There was no mortality in any group by day 3 after infection. By day 7, 0 of 8 mice infected with 104 PFU, 9 of 20 mice infected with 105 PFU, and 11 of 20 mice infected with 106 PFU had died. Few inflammatory cells were observed in the myocardium before day 7. Mice injected with both 105 and 106 PFU of virus developed statistically equivalent amounts of myocarditis by day 7, while animals infected with the lowest dose (104 PFU) had significantly less myocarditis than animals given the higher inocula. TNF- concentrations were significantly increased in all virus-infected mice compared to uninfected animals, but the levels in hearts of mice given the dose of either 105 or 106 PFU were higher than those in mice injected with 104 PFU. TNF- levels were elevated by day 3 despite minimal cardiac inflammation and were decreasing by day 7. Despite the substantial differences in myocarditis, cardiac virus titers were statistically equivalent between the groups given 104, 105, and 106 PFU at both days 3 and 7 after infection. In contrast, there was significantly more virus in the pancreases of mice given 105 and 106 PFU than in animals receiving 104 PFU, suggesting that increasing the initial virus inoculum may have distinct effects in different organs.
Myocarditis in transgenic mice. The rapid induction of TNF- in the heart concurs with published reports that this cytokine is important in myocarditis susceptibility (21, 22). Male C57BL/6, B6;129, TNF-–/–, p55 TNFR–/–, and p75 TNFR–/– mice were infected with 105 PFU of H3 virus and killed either 3 (four mice per group) or 7 (10 or more mice per group) days later. B6;129F2 mice were used as the control for the TNF-–/– strain, while C57BL/6 mice were controls for the p55 TNFR–/– and p75 TNFR–/– strains. No mortality occurred in the day 3 groups. Mortality at day 7 was 9 of 18 for C57BL/6, 5 of 10 for B6;129F2, 2 of 10 for TNF-–/–, 1 of 10 for p55 TNFR–/–, and 3 of 10 for p75 TNFR–/– mice. Whole hearts and pancreases were weighed and used for determination of virus titers on day 3. Organs obtained on day 7 were weighed and divided for histology and determination of virus titers. For these studies, titers are reported as PFU per gram of tissue. Blood was obtained 7 days after infection. Lymphoid cells were isolated, stimulated with phorbol myristate acetate-ionomycin-brefeldin A, and then stained to determine the percentage that were positive for CD4, IFN-, and IL-4. TNF-–/– mice had significantly less myocarditis and fewer CD4+ IFN-+ cells than B6;129F2 controls (Fig. 2). p55 TNFR–/– mice also developed significantly less myocarditis than wild-type C57BL/6 mice, while there was no difference in myocarditis between C57BL/6 and p75 TNFR–/– animals, indicating that this receptor does not participate in the disease process. CD4+ IFN-+ cells were decreased in p55 TNFR–/– mice compared to either C57BL/6 (P < 0.05) or p75 TNFR–/– (P < 0.05) animals, which concurs with studies associating CD4+ IFN-+ responses with myocarditis susceptibility (16). Although TNF- has been reported to promote encephalomyocarditis virus clearance, no significant difference was observed between cardiac or pancreatic virus titers in the different animal groups.
Although peripheral blood CD4+ Th1 (IFN-+) responses correlate to myocarditis susceptibility, the numbers of these cells in peripheral blood are modest. Confirmation that p55 TNFR–/– show reduced CD4Th1 in the heart is shown in Fig. 3. Inflammatory cells were isolated from four individual mice per group at 7 days after infection. There were approximately 10-fold more inflammatory cells in infected C57BL/6 and p75 TNFR–/– mice (106 cells per heart) than in the p55 TNFR–/– mice (105 cells/heart). Approximately 25% of inflammatory cells in C57BL/6 and p75 TNFR–/– mice were CD4+ IFN-+, compared to fewer than 1% of p55 TNFR–/– inflammatory cells. In contrast, 61% of inflammatory cells in p55 TNFR–/– mice were CD11b+ (Mac-1), compared to approximately 25% of cells from C57BL/6 and p75 TNFR–/– animals. The other major difference is that high proportions of cells from C57BL/6 and p75 TNFR–/– animals were V4+ and CD8+, in contrast to the case for p55 TNFR–/– animals, which had none of these cells. All CD8+ cells were also V4+. Approximately half of the V4+ cells were negative for both CD8 (Fig. 3) and CD4 (data not shown).
Figure 4 gives representative histology of the cardiac inflammation in the control (C57BL/6 and B6;129F2) and transgenic (TNF-–/–, p55 TNFR–/–, and p75 TNFR–/–) mice. Both control strains and p75 TNFR–/– mice showed substantial interstitial inflammation, while few inflammatory cells were obvious in TNF-–/– or p55 TNFR–/– mice.
TNF- and the p55 TNFR are necessary for CD1d induction in infected endothelial cells. V4+ T cells selectively infiltrate the hearts of myocarditis-susceptible mice, and previous studies have shown that this population is CD1d restricted (15, 16). The lack of V4+ cells in p55 TNFR–/– mice provided circumstantial evidence that TNF- might regulate CD1d expression during CVB3 infection. The roles of TNF- and virus infection in CD1d induction were investigated. First, vascular endothelial cells were isolated from the hearts and aortas of C57BL/6 mice. The purity of the population was determined by staining with anti-CD31, an endothelial cell marker (2), and demonstrated that the cell population contained 89.5% CD31+ cells. Endothelial cells were either uninfected or infected with H3 or H310A1 virus at an MOI of 100 and then were cultured for 0, 12, and 24 h (Fig. 5). Supernatants were retrieved and assayed for TNF-. Uninfected endothelial cells released no cytokine during the 24 h. Both H3- and H310A1-infected cells produced TNF-, but approximately eightfold more cytokine was released by the H3- compared to the H310A1-infected cells. Next, endothelial cells were isolated from C57BL/6 and transgenic mice and then infected with eGFP-CVB3 at an MOI of 100. Figure 6 shows that virus infection does not upregulate CD1d expression in endothelial cell cultures from TNF-–/– or p55 TNFR–/– transgenic mice. Endothelial cells from C57BL/6 and p75 TNFR–/– mice were fully capable of upregulating CD1d when infected. Although this study shows that TNF- is necessary, it does not indicate whether this cytokine is sufficient for CD1d induction. To investigate this point, C57BL/6 endothelial cells were either uninfected or infected with either the H3 (myocarditic) or H310A1 (nonmyocarditic) variant at an MOI of 100. Half of endothelial cell cultures were treated with 200 ng of TNF- per ml. Figure 7 shows that either TNF- treatment or H310A1 virus infection of C57BL/6 endothelial cells alone only modestly upregulated expression of CD1d. However, TNF- treatment of H310A1-infected cells induced amounts of CD1 equivalent to those in H3 infection.
Figure 8 shows a dose response for TNF- induction of CD1d in H310A1-infected C57BL/6 endothelial cells. Cells were infected with H310A1 at an MOI of 100 and treated with between 0 and 200 ng of cytokine per ml for 24 h. Cells were retrieved and labeled with anti-CD31, anti-CD1d, and anti-CVB3. With gating on the CD31+ (endothelial cell) population, the percentage of endothelial cells which were infected and expressed CD1d was determined and is shown in the upper panel of Fig. 8; examples of the flow cytometry for infected cells treated with 0 and 200 ng of TNF- per ml are shown in the lower panel. Upregulation of CD1d was dependent on the dose of cytokine used, with the maximal effect at 200 ng/ml. The percentage of infected cells did not statistically differ irrespective of the amount of TNF- used.
Effect of infection and TNF- on cell death. Both infection and TNF- might induce cell death. If there were differences in the percentage of cells dying in each culture, this might give a false impression of increased CD1d expression when CD1d+ cells may survive preferentially to the CD1d– population. Uninfected, infected, uninfected and TNF--treated, and infected and TNF--treated cells were produced as described above. The cells were labeled with annexin V and PI to determine live and dead cells (Fig. 9). Only 3.7% of uninfected endothelial cells were dead (annexin V+ PI+), compared to 18.6 and 24.2% in H3- and H310A1-infected cultures, respectively. TNF- treatment, either alone or with H310A1, did cause greater cell death than infection alone, but this slight increase does not correlate with the pattern of CD1d expression.
DISCUSSION
TNF- and IL-1? are proinflammatory cytokines which strongly promote myocarditis susceptibility in mice (21, 22). Exogenous administration of these cytokines causes myocarditis-resistant mice to develop the disease. This communication shows that TNF- promotes susceptibility to coxsackievirus B3-induced myocarditis through the p55 TNFR. The effect on myocarditis susceptibility is independent of virus titers, either early (day 3) or late (day 7), in the pancreas or heart. One of the interesting observations is that while C57BL/6 mice infected with relatively small amounts of virus (104 PFU/mouse) are myocarditis resistant, as has been shown previously (9), increasing the initial virus inoculum to 105 or 106 PFU per mouse results in substantial myocarditis susceptibility. The increased disease susceptibility is not apparently due to increasing cardiac virus titers, since mice given all three virus doses showed equivalent virus titers in the heart at 3 and 7 days after infection. Rather, increasing susceptibility with inoculum size causes substantial augmentation of TNF- levels in the heart even 3 days after infection, when inflammatory cells are not extensive. Analysis of the infiltrating cell populations in C57BL/6, p55 TNFR–/–, and p75 TNFR–/– mice shows that the C57BL/6 and p75 TNFR–/– animals have significant infiltration with CD4+ IFN-+, Mac-1+, and V4+ cells. Only the Mac-1+ cells predominate in the myocarditis-resistant p55 TNFR–/– mice.
It is therefore intriguing that two closely related CVB3 variants differ markedly in their ability to induce myocarditis and also vary markedly in their ability to upregulate TNF- expression. The H3 and H310A1 variants differ by a single nonconserved mutation at amino acid 165 in the VP2 protein (19). This asparagine (H3)-to-aspartate (H310A1) mutation occurs in a VP2 puff region associated in CVB3 with virus receptor (decay-accelerating factor [DAF]) binding (1, 23). The H310A1 variant was derived from the H3 variant as an antibody escape mutant by using a monoclonal antibody (clone 10A1) to the virus receptor (28, 30). H310A1 has significantly reduced receptor binding avidity compared to the H3 virus. Since antibody cross-linking of DAF activates tyrosine kinase pathways (4), it is highly likely that the H3 and H310A1 variants differ in their ability to transduce signals through DAF. Both virus and TNF- signal pathways are necessary for CD1d induction, since treating endothelial cells with the cytokine alone poorly upregulated this molecule, while both TNF- treatment and H310A1 infection strongly induced CD1d expression. H3-infected cells already express high levels of CD1d compared to H310A1 virus-infected cells, and extra TNF- had minimal effect on expression.
T cells expressing the V4+ T-cell receptor have been previously shown to play an essential role in myocarditis susceptibility (15, 16, 18). These effectors selectively recognize CD1d, a major histocompatibility complex class 1-like molecule which is important in innate immunity (5-7). CD1d–/– mice infected with H3 virus develop minimal myocarditis compared to CD1d+/+ animals and fail to activate V4+ cells (15). The V4+ cells infiltrate the H3 virus-infected myocardium within 3 days of virus injection, making these effectors the earliest T cells entering the heart (17). They are highly effective in killing infected but not uninfected cardiac myocytes, and this cytotoxicity is blocked by antibody to CD1d but not by antibodies to the classical major histocompatibility complex class I (H2K/D) and class II (H2 IA/IE) antigens (10, 15). Thus, rapid infiltration of infected organs by these innate effectors could aid in rapid control of virus prior to development of the adaptive immune response. This may be a reason why increasing the virus inoculum fails to result in increased cardiac virus titers. If V4+ cells infiltrate the heart at the higher virus inocula but not at the lower dose, the V4+ cells may be effective in limiting virus expansion.
ACKNOWLEDGMENTS
We thank Lindsay Whitton of Scripps Research Institute, La Jolla, Calif., for kindly supplying the eGFP-CVB3. We also thank Colette Charland for help with the flow cytometry and Kevin Kolinich for help in preparing the manuscript.
This work was supported by grants HL58583 and P01 AI45666 from the National Institutes of Health.
REFERENCES
Bergelson, J. M., J. G. Mohanty, R. L. Crowell, N. F. St. John, D. M. Lublin, and R. W. Finberg. 1995. Coxsackievirus B3 adapted to growth in RD cells binds to decay-accelerating factor (CD55). J. Virol. 69:1903-1906.
Chen, S., M. Sega, and A. Agarwal. 2004. "Lumen digestion" technique for isolation of aortic endothelial cells from heme oxygenase-1 knockout mice. BioTechniques 37:84-86, 88-89.
Choi, J. J., C. F. Reich III, and D. S. Pisetsky. 2004. Release of DNA from dead and dying lymphocyte and monocyte cell lines in vitro. Scand J Immunol 60:159-166.
Davis, L. S., S. S. Patel, J. P. Atkinson, and P. E. Lipsky. 1988. Decay-accelerating factor functions as a signal transducing molecule for human T cells. J. Immunol. 141:2246-2252.
Exley, M., N. Bigley, O. Cheng, S. Tahir, S. Smiley, Q. Carter, H. Stills, M. Grusby, Y. Koezuka, M. Taniguchi, and S. Balk. 2001. CD1d-reactive T-cell activation leads to amelioration of disease caused by diabetogenic encephalomyocarditis virus. J. Leukoc. Biol. 69:713-718.
Exley, M., S. Porcelli, M. Furman, J. Garcia, and S. Balk. 1998. CD161 (NKR-P1A) costimulation of CD1d-dependent activation of human T cells expressing the invariant Valpha 24Jalpha Q T cell receptor alpha chains. J. Exp. Med. 188:867-876.
Exley, M. A., Q. He, O. Cheng, R. J. Wang, C. P. Cheney, S. P. Balk, and M. J. Koziel. 2002. Cutting edge: compartmentalization of Th1-like noninvariant CD1d-reactive T cells in hepatitis C virus-infected liver. J. Immunol. 168:1519-1523.
Feuer, R., I. Mena, R. Pagarigan, M. K. Slifka, and J. L. Whitton. 2002. Cell cycle status affects coxsackievirus replication, persistence, and reactivation in vitro. J. Virol. 76:4430-4440.
Henke, A., S. Huber, A. Stelzner, and J. Whitton. 1995. The role of CD8+ T lymphocytes in coxsackievirus B3-induced myocarditis. J. Virol. 69:6720-6728.
Huber, S. 2000. T cells expressing the gamma-delta T cell receptor induce apoptosis in cardiac myocytes. Cardiovasc. Res. 45:579-587.
Huber, S., D. Graveline, W. Born, and R. O'Brien. 2001. Cytokine production by V+ T-cell subsets is an important factor determining CD4+ Th cell phenotype and susceptibility of BALB/c mice to coxsackievirus B3-induced myocarditis. J. Virol. 75:5860-5868.
Huber, S., D. Graveline, M. Newell, W. Born, and R. O'Brien. 2000. Vgamma1+ T cells suppress and Vgamma4+ T cells promote susceptibility to coxsackievirus B3-induced myocarditis in mice. J. Immunol. 165:4174-4181.
Huber, S., and P. Lodge. 1984. Coxsackievirus B3 myocarditis in Balb/c mice: evidence for autoimmunity to myocyte antigens. Am. J. Pathol. 116:21.
Huber, S., A. Moraska, and M. Cunningham. 1994. Alterations in major histocompatibility complex association of myocarditis induced by coxsackievirus B3 (CVB3) mutants selected with monoclonal antibodies to group A streptococci. Proc. Natl. Acad. Sci. USA 91:5543-5547.
Huber, S., D. Sartini, and M. Exley. 2003. Role of CD1d in coxsackievirus B3-induced myocarditis. J. Immunol. 170:3147-3153.
Huber, S., C. Shi, and R. C. Budd. 2002. Gammadelta T cells promote a Th1 response during coxsackievirus B3 infection in vivo: role of Fas and Fas ligand. J. Virol. 76:6487-6494.
Huber, S. A. 2004. Increased susceptibility of male BALB/c mice to coxsackievirus B3-induced myocarditis: role for CD1d. Med Microbiol Immunol (Berlin), in press.
Huber, S. A., D. Sartini, and M. Exley. 2002. V4+ T cells promote autoimmune CD8+ cytolytic T-lymphocyte activation in coxsackievirus B3-induced myocarditis in mice: role for CD4+ Th1 cells. J. Virol. 76:10785-10790.
Knowlton, K. U., E. S. Jeon, N. Berkley, R. Wessely, and S. Huber. 1996. A mutation in the puff region of VP2 attenuates the myocarditic phenotype of an infectious cDNA of the Woodruff variant of coxsackievirus B3. J. Virol. 70:7811-7818.
Kumar, A., V. Thota, L. Dee, J. Olson, E. Uretz, and J. E. Parrillo. 1996. Tumor necrosis factor alpha and interleukin 1beta are responsible for in vitro myocardial cell depression induced by human septic shock serum. J. Exp. Med. 183:949-958.
Lane, J., D. Neumann, A. Lafond-Walker, A. Herkowitz, and N. Rose. 1992. Interleukin 1 or tumor necrosis factor can promote coxsackie B3-induced myocarditis in resistant B10.A mice. J. Exp. Med. 175:1123-1129.
Lane, J., D. Neumann, A. Lanfond-Walker, A. Herskowitz, and N. Rose. 1993. Role of IL-1 and tumor necrosis factor in coxsackievirus-induced autoimmune myocarditis. J. Immunol. 151:1682-1690.
Lindberg, A., R. Crowell, R. Zell, R. Kandolf, and U. Pettersson. 1992. Mapping the RD phenotype of the Nancy strain of coxsackievirus B3. Virus Res. 24:187-196.
Meldrum, D. R. 1998. Tumor necrosis factor in the heart. Am. J. Physiol. 274:R577-R595.
Neumann, D., G. Allen, C. Narins, N. Rose, and A. Herskowitz. 1993. Heart autoantibodies in human myocarditis and cardiomyopathy: virus and skeletal muscle cross-reactivity, p. 325-334. In H. Figulla, R. Kandolf, and B. McManus (ed.), Idiopathic dilated cardiomyopathy. Springer-Verlag, Berlin, Germany.
Su, X.-M., and S. Sriram. 1991. Treatment of chronic relapsing experimental allergic encephalomyelitis with the intravenous administration of splenocytes coupled to encephalitogenic peptide 91-103 of myelin basic protein. J. Neuroimmunol. 34:181.
Takehana, H., T. Inomata, H. Niwano, M. Nishii, C. Matsuda, K. Kohno, Y. Machida, and T. Izumi. 2002. Immunomodulatory effect of pentoxifylline in suppressing experimental autoimmune myocarditis. Circ. J. 66:499-504.
Van Houten, N., P. Bouchard, A. Moraska, and S. Huber. 1991. Selection of an attenuated coxsackievirus B3 variant using a monoclonal antibody reactive to myocyte antigen. J. Virol. 65:1286-1290.
Wada, H., K. Saito, T. Kanda, I. Kobayashi, H. Fujii, S. Fujigaki, N. Maekawa, H. Takatsu, H. Fujiwara, K. Sekikawa, and M. Seishima. 2001. Tumor necrosis factor-alpha (TNF-alpha) plays a protective role in acute viral myocarditis in mice: a study using mice lacking TNF-alpha. Circulation 103:743-749.
Weller, A., K. Simpson, M. Herzum, N. Van Houten, and S. Huber. 1989. Coxsackievirus B3 induced myocarditis: virus receptor antibodies modulate myocarditis. J. Immunol. 143:1843-1850.(S. A. Huber and D. Sartin)