|Home | About | Journals | Submit | Contact Us | Français|
The interferon-β (IFN-β) response is critical for protection against viral myocarditis in several mouse models, and IFN-α or -β treatment is beneficial against human viral myocarditis. The IFN-β response in cardiac myocytes and cardiac fibroblasts forms an integrated network for organ protection, however the different IFN-α subtypes have not been studied in cardiac cells. We developed a quantitative RT-PCR assay that distinguishes between thirteen highly conserved IFN-α subtypes, and found that reovirus T3D induces five IFN-α subtypes in primary cardiac myocyte and fibroblast cultures: IFN-α1, -α2, -α4, -α5, and -α8/6. Murine IFN-α1, -α2, -α4, or -α5 treatment induced IRF7 and ISG56 and inhibited reovirus T3D replication in both cell types. This first investigation of IFN-α subtypes in cardiac cells for any virus demonstrates that IFN-α is induced in cardiac cells, that it is both subtype- and cell type-specific, and that it is likely important in the antiviral cardiac response.
Many viruses infect the heart, and >5% of the human population has experienced some form of viral myocarditis (Esfandiarei and McManus, 2008; Woodruff, 1980). Moreover, >50% of sudden deaths in young adults are due to cardiac causes with >10% of those due to myocarditis (Doolan, Langlois, and Semsarian, 2004; Puranik et al., 2005). Unfortunately, cardiac myocytes are not readily replenished, with a recent report suggesting only 1% cell renewal in the average human lifetime (Bergmann et al., 2009). This cardiac vulnerability likely necessitates a uniquely effective cardiac response to limit virus spread through the heart until immune defenses can be deployed. The cytokine IFN-β could provide this critical first line of defense (Muller et al., 1994), and indeed, we have shown that IFN-β is a determinant of protection against reovirus-induced myocarditis in a mouse model (Sherry, Torres, and Blum, 1998). That is, differences in the capacity of reovirus strains to induce myocarditis correlate with differences in viral induction of IFN-β and viral sensitivity to the antiviral effects of IFN-α/β in primary cultures of cardiac myocytes. In addition, a non-myocarditic reovirus induces cardiac lesions in mice lacking IFN-α/β function (Sherry, Torres, and Blum, 1998). Reovirus-induced myocarditis reflects direct viral cytopathogenic effect in cardiac cells (Baty and Sherry, 1993), while coxsackievirus B (CVB)-induced myocarditis reflects both direct and immune-mediated pathology (Huber, 2008; Whitton, 2002), and murine cytomegalovirus-induced myocarditis (MCMV) likely reflects predominantly immune-mediated damage (Lenzo et al., 2003). IFN-α/β provides protection against myocarditis induced by CVB (Wang et al., 2007) and MCMV (Bartlett et al., 2004; Cull, Bartlett, and James, 2002) as well, suggesting that regardless of mechanism, the IFN response plays an important antiviral role. Importantly, IFN-β treatment clears virus from the heart and improves cardiac function in patients with persistent viral myocarditis (Kuhl et al., 2006; Kuhl et al., 2003).
IFN-β is expressed and secreted by most cell types in response to viral infection or other stimuli (Kawai and Akira, 2008). Binding to the IFN-α/β-receptor (IFNAR) stimulates a signal transduction cascade to induce hundreds of interferon-stimulated genes (ISGs) (Khabar et al., 2004), some of which have antiviral function (Sadler and Williams, 2008), and one of which is a transcription factor (IRF7) which can further induce IFN-β and IFN-β4 in a positive amplification loop (Honda et al., 2005; Marie, 1998; Sato et al., 1998). However this signaling cascade is clearly cell type-specific (Daffis et al., 2008; van Boxel-Dezaire, Rani, and Stark, 2006). Previously, we found that cardiac myocytes and cardiac fibroblasts differ in their basal expression of components in the IFN pathway, resulting in differences in their IFN-β response to viral infection (Stewart et al., 2005; Zurney, Howard, and Sherry, 2007). Our results suggested that high basal IFN-β expression in cardiac myocytes pre-arms this vulnerable, non-replenishable cell type, while high basal expression of latent components in the IFN response path in adjacent cardiac fibroblasts renders these cells more responsive to IFN and prevents them from inadvertently serving as a reservoir for viral replication and spread to cardiac myocytes. These studies provided the first indication of an integrated network of cell type-specific innate immune components for organ protection, and confirmed the importance of investigating the IFN response in cell types relevant to the disease of interest.
Both IFN-α and IFN-β bind the IFN-α/β-receptor but with different affinities, resulting in induction of an overlapping but distinct set of ISGs (Jaks et al., 2007). While IFN-β is expressed by most cell types in response to viral infection or other stimuli, dendritic cells are likely the predominant cells expressing IFN-α as an early step in maturation of the adaptive immune response (Steinman and Hemmi, 2006). Nonetheless, IFN-α expression has been detected in many cell types, suggesting that direct protection against pathogens is also an important function for IFN-α. Indeed, IFN-α is a component of the most effective antiviral regimen currently available for treatment of Hepatitis C virus infection, and IFN-α has also been beneficial to patients with viral myocarditis (Daliento et al., 2003; Miric et al., 1996; Stille-Siegener, Heim, and Figulla, 1995). Given that IFN-α and IFN-β bind the same cell receptor and given the importance of both IFN-α and IFN-β in protection against viral myocarditis, we wished to investigate IFN-α expression and function in primary cultures of murine cardiac cells.
IFN-α is expressed in many if not all mammalian species (Woelk et al., 2007). While IFN-β is encoded by a single gene, mouse IFN-α is encoded by at least 13 functional genes which share >80% DNA and protein similarity (Hardy et al., 2004; van Pesch et al., 2004). This has limited the reagents and technologies available for discernment of individual IFN-α subtypes to determine their roles in protection against pathogens. To differentiate between thirteen highly conserved murine IFN-α subtypes, we developed a 3′-base mismatch approach to design primer pairs specific for each IFN-α subtype in quantitative RT-PCR. We used this novel assay and purified preparations of individual IFN-α subtypes to investigate IFN-α subtype expression and function in cardiac myocytes and fibroblasts. This first investigation of viral induction of IFN-α in cardiac cells for any virus demonstrates that IFN-α is induced in cardiac cells, that it is both subtype- and cell type-specific, and that it is likely important in the antiviral cardiac response.
To differentiate between thirteen highly conserved IFN-α subtypes, a 3′-base mismatch approach was adopted to design qRT-PCR primer pairs. This method is based on the extreme specificity of the 3′ terminal nucleotide for primer extension, i.e., cDNA synthesis can be extended only when its 3′ terminal nucleotide is complementary to the template nucleotide. Since IFN-α gene sequences for Cr:NIH(S) mice are unavailable, fifteen primer pairs were designed using alternative mouse sequences from GenBank, and then Cr:NIH(S) mouse genomic DNA was used in qPCR to confirm that the primer pairs could amplify products from this heterologous mouse strain (Table 1). Each primer pair generated a single peak in a standard melt-curve analysis (data not shown), indicating amplification of a single product. Next, total cell RNA was harvested from T3D-infected primary cardiac myocyte and primary cardiac fibroblast cultures, and was used as template for oligo-dT-primed cDNA synthesis and qRT-PCR. Finally, qRT-PCR products for IFN-α2, -α4 and α5 were sequenced using pyrosequencing technology, confirming the subtype-specificity for these primer pairs (examples in Fig. 1).
In order to determine whether T3D infection induces IFN-α and whether induction is subtype-specific in cardiac cells, primary cardiac myocyte cultures and primary cardiac fibroblast cultures were mock-infected or infected with T3D, and RNA was harvested at 8 hours post-infection for qRT-PCR. As reported previously (Sherry, Torres, and Blum, 1998; Stewart et al., 2005; Zurney, Howard, and Sherry, 2007), T3D induced IFN-β in both cell types. Results indicated that five of the thirteen IFN-α subtypes were induced in both cell types: IFN-α1, -β2, -β4, -β5, and -β8/6. Other IFN-α subtypes were either undetectable or the induction was less than 2-fold relative to mock-infected cultures. Therefore, additional experiments focused on these five IFN-β subtypes only.
IFN-β and IFN-α4 were both expressed at higher basal levels in cardiac myocytes than in cardiac fibroblasts (Fig. 2A), suggesting that the two genes may be regulated by a common cell type-specific mechanism. In contrast, IFN-α1, -α2, -α5, and -α8/6 were expressed at similar basal levels in the two cardiac cell types (Fig. 2A). Finally, IFN-α8/6 was expressed at significantly higher basal levels than IFN-β or any other IFN-α subtype (P values < 0.05 in all cases, except for IFN-α4 in myocytes), suggesting unique regulation of its basal expression by a mechanism common to both cardiac cell types.
At 8 hours post-infection with T3D, as previously seen for IFN-β (Zurney, Howard, and Sherry, 2007), all five IFN-α subtypes were induced to higher expression in cardiac myocytes than in cardiac fibroblasts (Fig. 2B). In both cell types, IFN-β expression was approximately 2- to 3-fold higher than IFN-α4 expression (fibroblasts and myocytes, respectively), and all other IFN-α subtypes were expressed at significantly lower levels. However, IFN-α2 was among the most highly expressed IFN-α subtypes in cardiac myocytes but among the least expressed in cardiac fibroblasts.
To compare relative expression of the different IFN-α subtypes, expression for each IFN-α subtype was calculated as a percentage of IFN-β expression (Fig. 2C). Notably, while IFN-α1, IFN-α2, IFN-α5 and IFN-α8/6 were expressed at 6 – 26% the level of IFN-β in cardiac fibroblasts, these same IFN-α subtypes were expressed at only 0.5 – 2% the level of IFN-β in cardiac myocytes; a ten-fold difference between cell types. Together, the data suggest that the Type I IFNs induced by viral infection are predominantly IFN-β and IFN-α4 in cardiac myocytes, but IFN-β and multiple IFN-α subtypes in cardiac fibroblasts.
To identify possible differences in the cell response to viral infection, T3D induction of IFN was calculated as fold induction relative to basal expression (Fig. 2D). As previously seen for IFN-β (Zurney, Howard, and Sherry, 2007), T3D infection resulted in a significantly greater fold induction for each of the five IFN-α subtypes in cardiac myocytes than in cardiac fibroblasts. Moreover, while the range of induction in cardiac myocytes for IFN-β, -α2, -α4, and -α5 spanned a >16-fold difference (448-fold for IFN-α5 to 7467-fold for IFN-β), the range in cardiac fibroblasts spanned only a <3-fold difference (259-fold for IFN-α5 to 712-fold for IFN-β). Nonetheless, the relative order was very similar between the two cell types. Together, the data suggest that differences in cardiac cell responses to viral infection are likely more quantitative than qualitative.
To investigate IFN-α induction over time, RNA was harvested from primary cardiac cell cultures at 4 – 24 hours post-infection with T3D. The four most abundantly induced IFN-α subtypes were quantified by qRT-PCR (Fig. 3). The trends in relative expression levels for the different IFN-α subtypes were very similar to the single 8 hr time-points in Fig. 2. To investigate secreted IFN-α, supernatants were harvested between 4 and 24 hours post-infection of primary cardiac cell cultures, and total IFN-α was quantified by ELISA (Fig. 4). T3D induction of secreted IFN-α was significant by 8 hours post-infection for cardiac myocytes (P < 0.001) but only at 24 hours post-infection for cardiac fibroblasts (P = 0.002), and was significantly higher for cardiac myocytes than cardiac fibroblasts at 12 and 24 hours post-infection (P = 0.05 and 0.002, respectively). Therefore, reovirus T3D induction of secreted IFN-α, like IFN-α mRNA, is cell type-specific.
In some cell types, viral infection induces initial synthesis and secretion of IFN-β and IFN-α4, which then signal through the IFN-α/β receptor to induce IRF7 for subsequent further induction of IFN-β, IFN-α4 and other IFN-α subtypes (Honda et al., 2005; Marie, 1998; Sato et al., 1998). To determine whether reovirus T3D induction of IFN-α is direct or is mediated by IFN-α/β, primary cardiac cell cultures were generated from IFN-α/β-receptor-null mice, and reovirus T3D induction of IFN was quantified by qRT-PCR. Basal expression of IFN mRNA was either undetectable or very low in IFN-α/β-receptor-null cells (data not shown) relative to wild-type cells (Fig. 2A), and suggested a ≥5-fold decrease in IFN-β, IFN-α4, and IFN-α8/6 in cardiac myocytes and a ≥5-fold decrease in IFN-α8/6 in cardiac fibroblasts (data not shown). Basal expression of other IFN mRNA types was too low in wild-type cells (Fig. 2A) to extrapolate reduction in IFN-α/β-receptor-null cells. T3D induced IFN-β and IFN-α4 significantly in IFN-α/β-receptor null cultures, but expression was dramatically decreased relative to wild type cardiac cell cultures (Fig. 5). Interestingly, while the reduction in expression was greater in cardiac myocytes than cardiac fibroblasts for both IFN-β and IFN-α4, the reduction was most dramatic for T3D-induced IFN-α4 expression in cardiac myocytes (Fig. 5B), resulting in an inversion where cardiac fibroblasts expressed more IFN-α4 than cardiac myocytes. Reduction in expression for other IFN-α subtypes ranged from 10- to 100-fold (data not shown) or was not possible to extrapolate given the low expression in wild type cultures (Fig. 2B). Together the data suggest that IFN-mediated amplification is more critical for T3D-induced IFN-α expression, in particular for IFN-α4, than for IFN-β expression in cardiac myocytes. In contrast, T3D-induced type I IFN in cardiac fibroblasts is less dependent on IFN-mediated amplification.
To determine whether IFN-α or IFN-β induces Type I IFN in cardiac cells in the absence of viral infection, primary cardiac cell cultures were treated with murine IFN-β or individual IFN-α subtypes at 1000 U/ml. Total RNA was harvested at 2 and 5 hours post-treatment, and IFN expression was assessed by qRT-PCR (duplicate experiments, data not shown). Neither IFN-β nor any of the IFN-α subtypes tested induced Type I IFN in cardiac cells in the absence of viral infection.
To determine whether individual IFN-α subtypes can signal in cardiac cells, IFN-α subtype induction of two ISGs was assessed by qRT-PCR (Fig. 6). The ISG IRF-7 was chosen for its importance in the positive feedback loop for IFN induction (Honda et al., 2005; Marie, 1998; Sato et al., 1998), and the ISG 56 was chosen as one representative antiviral ISG. IFN-α induction of ISGs was greater in cardiac fibroblasts than in cardiac myocytes, consistent with previous results for IFN-β in these cell types (Stewart et al., 2005; Zurney, Howard, and Sherry, 2007). At 8 hours post-treatment, the trend for IRF7 expression was similar in the two cells types, with the greatest induction by IFN-β and IFN-α4, and the lowest induction by IFN-α2. A similar trend was seen for ISG 56 expression, except that expression decreased by 8 hours post-treatment in cardiac myocytes. Thus cardiac myocytes and cardiac fibroblasts are differentially sensitive to IFN-α treatment as previously seen for IFN-β (Zurney, Howard, and Sherry, 2007), and signaling in cardiac cells is IFN-α subtype-specific.
To determine whether individual IFN-α subtypes can provide antiviral protection to cardiac cells, primary cardiac cell cultures were pre-treated with IFN-α subtypes and then challenged with reovirus T3D (Fig. 7). In both cardiac myocytes and cardiac fibroblasts, each IFN-α subtype tested decreased viral replication at 1000 U/ml, with decreased effects at 100 U/ml. And in both cell types, IFN-β provided the greatest antiviral protection, while IFN-α1 and IFN-α2 provided the least protection.
IFN-α and IFN-β bind a shared receptor to induce expression of an overlapping set of antiviral genes (Jaks et al., 2007), and both cytokines have been used successfully in the treatment of human viral myocarditis (Daliento et al., 2003; Kuhl et al., 2006; Kuhl et al., 2003; Miric et al., 1996; Stille-Siegener, Heim, and Figulla, 1995). However, the only investigation measuring expression of different IFN-α subytpes in the heart during murine viral myocarditis used CVB3 (Baig and Fish, 2008), which induces both direct and immune-mediated damage (Huber, 2008; Whitton, 2002). Therefore, the IFN-α in that study likely reflected contributions from both cardiac and inflammatory cells. Similarly, there has been only one investigation comparing the efficacy of different IFN-α subtypes in protection against murine viral myocarditis (Bartlett et al., 2004; Cull, Bartlett, and James, 2002), and that study also used a virus that mediates cardiac damage primarily through immune-mediated rather than direct mechanisms (MCMV (Lenzo et al., 2003)). There have been no studies comparing cardiac expression of the different IFN-α subtypes using a virus that mediates direct damage to the heart, or using cardiac cells to assess expression of different IFN-α subtypes and their relative importance in the cardiac protective response. Results here demonstrate that IFN-α is induced in cardiac cells, that it is both subtype- and cell type-specific, and that it is likely important in the antiviral cardiac response.
Investigations comparing expression for the different IFN-α subtypes have been hindered by the >80% DNA homology between genes (Hardy et al., 2004; van Pesch et al., 2004). One recent study used fluorescein-labeled probes in a heteroduplex analysis of PCR products generated from consensus primers (Demoulins et al., 2009). Several other recent studies used IFN-α subtype-specific primers, one for Taqman-based qRT-PCR (Loseke et al., 2003) and two for Sybergreen-based qRT-PCR (Baig and Fish, 2008; Izaguirre et al., 2003). Here, we took advantage of the extreme specificity of the 3′ terminal nucleotide for primer extension to design primers that could be used for Sybergreen-based qRT-PCR, a much less expensive option than Taqman-based qRT-PCR. We also used pyrosequencing in a novel approach to validate the specificity of qRT-PCR products without the need for subcloning. Results provide not only insights into differential IFN-α subtype expression in cardiac cells, but also a method for quantitatively comparing and validating expression from closely related genes.
IFN-α is expressed at detectable levels in some tissues even without stimulation (Brandt, Linnane, and Devenish, 1994; Marie, 1998; Tovey et al., 1987; Zoumbos et al., 1985). High basal expression of IFN-α can directly protect cells from virus infection (Bautista et al., 2005) and can determine high basal expression of IRF7 and thereby increase viral induction of IFN through a positive amplification loop (Hata et al., 2001). High basal IFN-α can also stimulate maturation and activation of dendritic cells (Montoya et al., 2002) and can augment cell responses to IFN-γ (Takaoka et al., 2000) and IL-6 (Mitani et al., 2001). Here, we found that basal expression of IFN-α was both subtype- and cell type-specific in cardiac cells. We previously demonstrated that high basal Type I IFN expression in cardiac myocytes stimulates high basal ISG expression relative to cardiac fibroblasts, resulting in a pre-arming in cardiac myocytes from both protective ISGs and latent IRF7 expression (Zurney, Howard, and Sherry, 2007). Results here indicate that this Type I IFN could be comprised of both IFN-β and IFN-α4, but not IFN-α1, -α2, -α5 or -α8/6 since they were expressed at basal levels that were equivalent in the two cell types (Fig. 2). In addition, results here demonstrate that IRF7 is not active in cardiac myocytes in the absence of viral infection, consistent with results seen in other cell types. That is, high basal IRF7 expression in cardiac myocytes did not result in higher basal expression of downstream IFN-α subtypes in that cell type relative to cardiac fibroblasts (Fig. 2). The mechanism underlying high basal expression of IFN-β and IFN-α4 in cardiac myocytes remains unclear.
Type I IFN expression is enhanced by an IRF7-mediated positive amplification loop (Honda et al., 2005; Marie, 1998; Sato et al., 1998) and cell type-specific differences in IRF7 expression can therefore determine viral induction of IFN-α. Indeed, differences in virus-induced expression for two different human IFN-α subtypes reflects differences in organization of IRF elements in the promoters of their genes (Civas et al., 2006; Genin et al., 2009). Constitutive IRF7 levels correlate with levels of both IFN-α4 and IFN-non-α4 induced by influenza virus in several cell types, and high constitutive IRF7 overcomes the requirement for the positive amplification loop for expression of both IFN-α classes (Prakash et al., 2005). West Nile Virus induction of IFN-α is reduced in multiple primary cell types deficient in IRF7, but interestingly, IFN-β is not (Daffis et al., 2008). However, no previous studies have addressed the impact of IRF7 expression levels on viral induction of different IFN-α subtypes. Previously, we found that reovirus T3D induces greater IRF7 in cardiac fibroblasts than in cardiac myocytes (Stewart et al., 2005), consistent with the greater responsiveness of the former than the latter to IFN signaling and reflecting differences in basal expression of the JAK-Stat pathway (Zurney, Howard, and Sherry, 2007). Here, we found that while IFN-β and IFN-α4 comprised the majority of Type I IFN induced by reovirus in both cardiac cell types, the relative contribution from the other subtypes differed dramatically, providing only 4% of the total Type I IFN in cardiac myocytes but 27% of the total in cardiac fibroblasts (Fig. 2B). Moreover, reovirus induced IFN-non-α4 subtypes to a combined level of only 6% that of IFN-β in cardiac myocytes, but 56% the level of IFN-β in cardiac fibroblasts (Fig. 2C). These data would suggest that cell type-specific differences in virus-induced IRF7 levels (Stewart et al., 2005) correlate with differences in viral induction of IFN-α subtypes downstream of IRF7. However in contrast, viral induction of IFN-α4 and IFN-β in cells lacking IFN-α/β signaling was reduced much more in cardiac myocytes than in cardiac fibroblasts (Fig. 5). This may reflect the fact that basal IRF7 expression is much greater in cardiac myocytes than in cardiac fibroblasts, and that this basal expression is dependent on IFN-α/β signaling (Stewart et al., 2005). If basal IRF7 expression contributed to viral induction of IFN-β and IFN-α4, then viral induction of these two IFN types would be more dependent on IFN-α/β signaling in cardiac myocytes than in cardiac fibroblasts, as was seen here (Fig. 5). Finally, IFN-α/β signaling was more critical for viral induction of IFN-α4 than IFN-β in cardiac myocytes (Fig. 5). Induction of these two IFN types is generally co-incident, as they are both induced directly by virus and also through the amplification loop (Honda et al., 2005; Marie, 1998; Sato et al., 1998). Differential expression of IFN-β and IFN-α subtypes in human cells is determined, at least in part, by different promoter interactions with IRF3 and IRF7 (Genin et al., 2009). Indeed, IRF3 can repress IRF7 induction of some human IFN-α subtypes (Genin et al., 2009). If viral induction of mouse IFN-α4 was more dependent than IFN-β on IRF7, then, as was seen here (Fig. 5), IFN-α4 induction would be more dependent than IFN-β on IFN-α/β signaling (for both basal IRF7 expression (Stewart et al., 2005) and IFN-induced IRF7 expression). These data would suggest that, as for West Nile Virus comparisons between total IFN-α and IFN-β (Daffis et al., 2008), the role of IRF7 in reovirus induction of Type I IFN varies between IFN types and is cell type-specific.
IFN-α subtypes differ in their antiviral activity both in vitro and in vivo. When COS cells were transfected with constructs expressing murine IFN-α1, -α2, -α4 and -α6, and then supernatants were tested for protection of L929 and CHO cells against challenge with vesicular stomatitis virus, IFN-α4 was most effective (Van Heuvel et al., 1986). But when purified recombinant IFN-α1 and -α4 were tested for protection of L929 and J2E cells against encephalomyocarditis virus challenge, the two subtypes were equivalent (Swaminathan et al., 1992). IFN-α1, -α4 or -α9 transgenes introduced into regenerating muscles of mice each reduced MCMV replication, with IFN-α1 providing the most protection (Yeow, Lawson, and Beilharz, 1998). When IFN transgenes were applied vaginally to mice before challenge with HSV-2, IFN-α1, -α5 and -β were protective while -α4, -α6 and -α9 were not (Austin et al., 2006). Finally when mice were injected with constructs expressing IFN-α subtypes and challenged with influenza virus, -α5 and -α6 were most and -α1 least protective (James et al., 2007). Together, the data suggest that the antiviral activity of murine IFN-α subtypes may be both virus- and tissue-specific. Similar differences in antiviral activity have been seen for human IFN-α subtypes (Foster et al., 1996; Koyama et al., 2006; Larrea et al., 2004; Schanen et al., 2006). Interestingly, IFN-α subtypes may differ in antiviral activity but not anti-proliferative function, or vice versa (Foster et al., 2004; Swaminathan et al., 1992). Here, we found that the relative antiviral activity for Type I IFNs was the same for cardiac myocytes and cardiac fibroblasts: IFN-β provided the greatest protection, -α4 and -α5 were intermediate, and -α1 and -α2 provided the least protection (Fig. 7). These relative antiviral effects parallel relative IFN subtype-specific induction of ISGs (Fig.6; 8 hours post-treatment), perhaps reflecting differences in affinity for the IFN-α/β-receptor (Jaks et al., 2007). Finally, IFN-β, -α4 and -α5 were expressed at higher levels than IFN-α1 and -α2 in cardiac fibroblasts after reovirus infection (Fig. 2B), potentially contributing even further to their protective functions in the heart.
Viral induction of IFN (Fig. 2), IFN induction of ISGs (Fig. 6), and IFN antiviral activity (Fig. 7) was always greater for IFN-β than for any IFN-α subtype. This is consistent with previous evidence for greater activation of the JAK-Stat pathway (Grumbach et al., 1999) and greater antiviral activity of IFN-β than IFN-α (pooled preparations) against CVB3 (Heim and Weiss, 2004) in human cardiac fibroblasts. Indeed, IFN-β provides protection against influenza virus in mice that cannot be provided by IFN-α alone (Koerner et al., 2007). Nonetheless, while the combined expression of IFN-α subtypes was only 26% of the total IFN in infected cardiac myocytes, it was 52% of the total in infected cardiac fibroblasts (Fig. 2B). Given that cardiac fibroblasts are more responsive than cardiac myocytes to both IFN-α (Fig. 6) and IFN-β (Stewart et al., 2005; Zurney, Howard, and Sherry, 2007), this contribution of IFN-α could be particularly important in cardiac fibroblasts. Finally, although combined antiviral effects cannot be extrapolated from the data, IFN-α4 and -α5 each expressed 15% of the antiviral activity of IFN-β in cardiac myocytes, and 16 to 23% the activity in cardiac fibroblasts (Fig. 7B). Together, the data indicate that IFN-α subtypes are likely to play an important role in protecting the heart from viral infection.
Timed-pregnant Cr:NIH(S) mice were purchased from the National Cancer Institute. IFN-α/β-receptor-null mice (Muller et al., 1994) were maintained as breeding colonies to generate neonates and fetuses for generation of primary cell cultures. Mouse facilities were accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, and all procedures were approved by the North Carolina State University institutional animal care and use committee.
Primary cardiac myocyte cultures and primary cardiac fibroblast cultures were generated from 1- or 2-day-old neonatal or term fetal mice according to the method described previously (Baty and Sherry, 1993). Briefly, the apical two-thirds of the hearts from euthanized neonates or fetuses were removed and trypsinized. Cells were suspended in Dulbecco's modified Eagle's medium (DMEM; Gibco BRL, Gaithersburg, MD) supplemented with 7% fetal calf serum and 10 μg of gentamicin (Sigma Co.) per ml (cDMEM). Cells were plated in 6-well clusters, and incubated for 2 hours for separation by differential adhesion. Cardiac myocytes were harvested from the supernatant and the adherent cardiac fibroblasts were harvested by trypsinization. Following centrifugation, cardiac myocytes were suspended in cDMEM plus 0.06% thymidine (Sigma Co., St. Louis, MO), cardiac fibroblasts were suspended in cDMEM, and both cultures were incubated in a 37°C, 5% CO2 incubator. Cells were never passaged before use. By immunofluorescent staining, myocyte cultures contained <5% fibroblasts while fibroblast cultures contained <1% myocytes (Zurney, Howard, and Sherry, 2007) .
Reovirus type 3 Dearing (T3D (Schiff, 2007)) was plaque purified and amplified in mouse L929 cells, which were maintained in a suspension system in sMEM (SAFC Biosciences, Denver, PA) supplemented with 5% fetal calf serum (Atlanta Biologicals, Atalanta, GA) and 2mM Lglutamine (Mediatech Inc., Herndon, VA). Virus was purified by CsCl gradient centrifugation (Smith, Zweerink, and Joklik, 1969) for use in all experiments.
Cardiac myocytes from Cr:NIH(S) mice were plated at 1 × 106 cells per well and cardiac fibroblasts were plated at 5 × 105 cells per well in 2 ml media in 24-well clusters. Cardiac fibroblasts were assumed to double after two day's incubation. All cultures were washed twice, and duplicate wells were infected at an moi of 10 pfu per cell two days post-plating. Virus suspended in 700 μl of cDMEM with or without thymidine was added to each well, and cells were incubated in 37°C, 5% CO2 incubator for 1 hour. Then an additional 1000 μl cDMEM with or without thymidine was added to each well. Mock-infected cells were treated identically except the inocula lacked virus.
Cardiac myocytes from IFN-α/β-receptor-null mice were plated at 5 × 105 cells per well and cardiac fibroblasts were plated at 2.5 × 105 cells per well in 1 ml media in 48-well clusters. Infections were as above, except that virus inoculum was 350 μl per well, and an additional 500 ml media was added after the 1 hour incubation.
Cardiac myocytes from Cr:NIH(S) mice were plated at 1.5 × 105 cells per well and cardiac fibroblasts were plated at 7.5 × 104 cells per well in 300μl media in 96-well clusters. Cardiac fibroblasts were assumed to double after two day's incubation. Twenty-four hours post-plating, cells were washed once, and overlying medium was replaced with 300 μl of IFN-β, -α1, -α2, -α4 or -α5 at 100 or 1000U per ml in cDMEM with or without thymidine in triplicate wells (IFN-β: cat # 12400-1; IFN-α subtypes: a gift of purified recombinant preparations quantified in a standard IFN assay measuring protection of L929 cells against encephalomyocarditis virus challenge; both from PBL Biomedical Laboratories, Piscataway, NJ). . After an additional overnight incubation, cells were washed twice, infected with T3D at an moi of 10 pfu per cell in 100 μl inoculum, and incubated for 1 hour in a 37°C, 5% CO2 incubator. Cells were washed once, and the appropriate IFN at 100 or 1000U per ml in 250μl cDMEM with or without thymidine was added to each well. Control cells lacked IFN treatment, but were infected with T3D as above. After 20 hours incubation, cultures were stored at −80°C, frozen and thawed an additional two times, and supplemented to a final 0.5% Nonidet P-40 (Sigma). Serial dilutions were used to infect L929 cells in a standard plaque assay (Sherry, Baty, and Blum, 1996).
Supernatants and RNA were harvested at indicated times post-infection or post-treatment with IFN (from PBL as above). Supernatants were stored at −35°C, and remaining cells were lysed directly on the plate using cell lysis buffer from an RNeasy Kit (Qiagen, Inc., Valencia, CA) supplemented with 1% β-mercaptoethanol. Cell lysates were homogenized using Qiashredders (Qiagen, Inc.), and total RNA was isolated using the RNeasy Kit according to the manufacturer's instructions. Genomic DNA was removed using RNase-free DNase I (Qiagen, Inc.). The isolated RNA was stored at −80°C.
To generate cDNA, one third of the RNA harvested from each well in a 24-well cluster or two thirds from a 48-well cluster was used as template in a 100 μl reaction containing 5μM oligo(dT) (Invitrogen Corp., Carlsbad, CA), 1× Taq buffer (Promega corp., Madison, WI), 7.5 mM MgCl2 (Promega corp.), 1 mM dithiothreitol (Promega corp.), 1 mM each dNTP (Roche, Indianapolis, IN), 0.67 U/μl RNasin (Promega Corp., Madison, WI), and 0.20 U/μl of AMV reverse transcriptase (Promega Corp.).
Gene expression was quantified using a Sybergreen system on an iCycler iQ fluorescence thermocycler (Bio-Rad Laboratories, Hercules, CA). Each 25 μl reaction contained 5% of the RT product, 1× Quantitech master mix (Qiagen, Inc.), 10 nM fluorescein (Invitrogen Corp., Carlsbad, CA), and 0.3μM each of the forward and reverse primers (IFN-β forward: 5′- GGAGATGACGGAGAAGATGC-3′ and reverse: 5′- CCCAGTGCTGGAGAAATTGT; GAPDH forward: 5′- GGGTGTGAACCACGAGAAAT-3′ and reverse: 5′-CCTTCCACAATGCCAAAGTT-3′; IFN-α as indicated in Table 1). iCyclerTM iQ Optional System Software, version 3.0 (Bio-Rad Laboratories) was used to analyze the data. Standard curves (generated from serial dilutions of known concentrations of purified PCR products) were used to quantify copy number. Individual cDNA standards for each IFN-α subtype were purified PCR products, using primers to generate products that contain the primers indicated in Table 1, as follows: IFN-α1 (334bp) Forward: 5′-GGAACAAGAGAGCCTTGACA-3′ and reverse: 5′-CTCACAGCCAGCAGGGCAT-3′; IFN-α2 (626bp) forward: 5′-TCTGTGCTTTCCTCGTGATG-3′ and reverse: 5′-GATGCAGTTTCTAGTCCAGG -3′; IFN-α4 (295bp) forward: 5′-CTGCTGGCTGTGAGGACATA-3′ and reverse: 5′-TTGCTCAAGATTGCTGAAACA-3′; IFN-α5 (603bp) forward: 5′-AGAGCCTTAACCCTCCTGGT-3′ and reverse: 5′- CGCTCAAGATTGCTGAAACA - 3′; IFN-α8/6 (240bp) forward: 5′-CCTGATGGTTTTGGTGGTGT-3′ and reverse: 5′-ATCTGCTGGGTCAGCTCAG-3′. Gene-of-interest expression was normalized to GAPDH gene expression, and then further normalized to account for differences in GAPDH expression in myocytes and fibroblasts. For example, normalized copies IFN-β in myocyte sample = [(copies IFN-β in myocyte sample) / (copies GAPDH in myocyte sample)] × (average copies GAPDH in all myocyte samples) / (average copies GAPDH in all fibroblast samples). Normalized copies IFN-β in fibroblasts = (copies IFN-β in fibroblast sample) / (copies GAPDH in fibroblast sample) with no further normalization for differences between cell types since that was already accounted for in myocytes.
qRT-PCR products were generated as above, except that one primer of the primer pair was biotinylated at its 5′ terminus (Invitrogen). A 50% slurry of Streptavidin Sepharose™ High Performance beads (GE Heathcare, Piscataway, NJ) was suspended in Binding buffer (10 mM Tris-HCl pH 7.6, 2 M NaCl, 1 mM EDTA, 0.1% Tween 20; Biotage, Charlottesville, VA) for a final 5% slurry. Primers for pyrosequencing were suspended in annealing buffer (20 mM Tris-acetate pH 7.6, 2 mM Mg-acetate; Biotage) at a final concentration of 0.4μM. Ten μl (40%) of qRT-PCR products were mixed with the streptavidin beads in 96-well clusters (Corning, NY), and then transferred to PSQ 96 Sample Prep Thermoplates (Low; Biotage) for Pyrosequencing, using Pyro Gold Reagents (Biotage) according to the manufacturer's instructions. Primers for pyrosequencing were: GAPDH (5′-GGTCTACATGTTCCAGTATG-3′); IFN-β (5′-CGGAGAAGATGCAGAAGAGT-3′); IFN-α2 (5′- ATTCCCCCTGGAGAAGGTG -3′); IFN-α4 (5′-CTGGAGAGCCCTCTCTTCCT-3′), and IFN-α5 (5′-AGCCTGTGTGATGCAACAGG-3′).
Expression of IFN-α protein was determined using a mouse IFN-α ELISA kit (PBL Biomedical Laboratories) according to the manufacturer's instructions. Briefly, 100 μl supernatant from each sample (24-well clusters) was measured. A standard curve was constructed ranging from 0-500 pg/ml, using serial dilutions of a mouse IFN-α standard provided with the kit. Absorbance was detected at 450nm on a Tecan Sunrise Microplate Reader (Tecan Systems Inc., San Jose, CA).
A Student's two-sample t-test (pooled variance) was applied for all cases except for Fig. 7, where ANOVA (Tukey's test) was applied (Systat 9.0). Results were considered significant at P < 0.05.
We thank Lindsey Jones, Susan Irvin and Jennifer Zurney for insightful discussions, Wrennie Edwards for technical assistance, and Ronald G. Jubin and colleagues at PBL Biomedical Laboratories for their gift of IFN-α preparations. This research was supported by NIH award R01 AI062657 and graduate student support from the North Carolina State University Genomics Graduate Program (L.L.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.