Temporal transcriptome changes induced by MDV in marek's disease-resistant and -susceptible inbred chickens

MD is a serious lymphoproliferative disease in chickens caused by MDV and characterized by transformation of T cells that cause tumors in various organs including liver, spleen, gonads, heart, peripheral nerves, skin and muscle [1–3]. Chickens with MD exhibit over-expression of Hodgkin's disease antigen CD30 (CD30^hi) that makes it a natural model for studying the initiation and progression of CD30^hi lymphomas [4]. MDV is an alphaherpesvirus belonging to the Mardivirus genus which contains three members: MDV-1, MDV-2 and HVT (herpesvirus of turkeys) [5–7]. According to Calnek et al. [8, 9], MDV, like other herpesviruses, goes through a complex life cycle that includes cytolytic and latent phases in host cells. An early cytolytic infection is started at 2-7dpi characterized by the virus particles expressing large amounts of the early protein pp38. Subsequently, a latent phase is initiated at around 7-10dpi with the MDV genome persisting in the host cells. Following latency, a late cytolytic phase causes inflammation and transformation of latently infected lymphocytes into tumor cells and is triggered between 14-21dpi [8, 9]. During the first cytolytic phase, MDV first uses B cells as a target for its replication before targeting activated CD4⁺ T cells to enable a persistent latent infection [10–12].

Two highly inbred chicken lines 6₃ and 7₂, sub-lines of lines 6 and 7, have been bred since 1939 with line 6₃ chickens resistant to MD and line 7₂ chickens susceptible to MD [13]. To better understand the mechanisms underlying MD-resistance and -susceptibility, several studies have been made to ascertain the differences between these two chicken lines. A much higher virus copy number was observed in line 7 chickens indicating varying levels of virus replication [14]. Different proportions of CD4⁺ T cells and CD8⁺ T cells were found in MD-resistant and -susceptible chickens when infected by MDV. In MD-susceptible birds, as the CD4⁺ T cells increased in number, the number of CD8⁺ T cells decreased; the opposite occurred in MD-resistant chickens [15]. Lymphocyte surface markers such as Ly-4, Bu-1 and Th-1, were present in different levels in these two chicken lines [16, 17]. The expression of some cytokines, such as IL6 and IL18, was also found to differ between line 6 and line 7 chickens[18]. From an epigenetic perspective, differences in promoter DNA methylation levels between line 6₃ and line 7₂ chickens have been found in several candidate genes [19].

With the development of functional genomics technologies, some progress has been made towards investigating the mechanism of MD-resistance in a genome-wide manner. Quantitative trait loci associated with MD-resistance or -susceptibility have been mapped to chromosomes 1, 5, 7, 9, 15, 18, 26, Z, E21 and E16 [20–23]. Also, through the use of chicken immune-specific microarrays, immunoglobulin genes have been shown to have a higher expression in MD-resistant chicken lines as compared to MD-susceptible chicken lines [24]. However, the exact mechanisms behind resistance and susceptibility to MD are still unknown.

Researchers at the Avian Disease and Oncology Laboratory (ADOL, East Lansing, MI, USA) have developed nineteen recombinant congenic strains (RCS) with varying phenotypic traits from lines 6₃ and 7₂ to further investigate the mechanisms of MD-resistance and - susceptibility [13, 25]. One of these strains, RCS-M, which was developed from line 6₃ and line 7₂ and possesses around 87% genetic background of line 6₃, is genetically closer to the resistant line 6₃. Our previous study of tumor incidence rates induced by a partially attenuated very virulent plus (vv+) strain of MDV (648A, passage 40) in the different RCSs revealed that while only 0-3% of line 6₃ chickens and up to 100% of the line 7₂ chickens developed tumors after MDV infection, about 40% of RCS-M chickens developed tumors (Data not shown). Because of this intermediate response of RCS-M to MDV infection, it is suitable for us to use these three chicken lines to investigate the mechanism of MD-resistance and susceptibility. We performed a temporal trancriptome analysis with spleen samples from line 6₃, 7₂ and RCS-M chickens before and after MDV infection at 5dpi, 10dpi and 21dpi. Our main objective is to build on the current understanding of Marek's disease pathogenesis and immune response to MDV, and this genome-wide approach is used for this purpose. To our knowledge, this is the first comprehensive study combining a chicken line having intermediate resistance to MD together with highly-resistant and susceptible lines to more precisely identify the possible genetic factors behind MD resistance and susceptibility.

There have been several studies looking at gene expression changes related to disease in general [28–30] and MD in particular [24, 27, 31–34], although results tend to vary a lot. MD is a complex disease with the disease phenotype in susceptible individuals depending on the location and frequency of tumors. Any single gene with differential expression cannot fully explain the phenomenon of host resistance or susceptibility. Therefore, we tried to use a genome-wide approach to build on the current understanding of Marek's disease pathogenesis and immune response to MDV. Upon close examination of the transcriptional responses, dramatically increased numbers of significant genes were observed at 10dpi in RCS-M and at 21dpi in line 7₂ at lower FDR levels (FDR < 0.2) which indicated a strongly enhanced transcriptional response. At a more relaxed FDR level (FDR < 0.5), we find comparable numbers of differentially expressed genes at 5, 10 and 21dpi in line 6₃ and RCS-M. Line 7₂ has similar numbers of significant genes at 10dpi but there is a definite increase in the transcriptional response at 21dpi with close to twice as many differentially expressed genes. However, even at this level, we do not find any significantly expressed genes at 5dpi in line 7₂ (FDR<0.5), indicating a much muted transcriptional response (Table 1).

Over the years, several attempts have been made to identify the gene profiles that change as a result of MDV infection. For example, using microarray analysis, studies have identified some genes related to MDV infection by using different chicken lines and MDV strains [24, 31–34]. When chicken embryo fibroblasts were infected with MDV, genes related to inflammation, cell-growth and differentiation and antigen presentation, such as MIP, IL-13R, MHC I and MHC II were induced both at 2dpi and 4dpi [33]. In contrast, in spleen tissue, several other genes were found to be affected at an early stage, including TLR-15, IL-6 and Mx1[31]. In chickens with major histocompatibility complex (MHC)-associated MD resistance, the immunoglobulin genes IgG and IgM were differentially expressed after MDV infection at 7dpi and 14dpi[24], whereas in lines 6 and 7 from ADOL, that carry the same MHC haplotype (B²) but differ in their response to MDV infection[35], various alloantigens like Ly-4 [16] and Bu-1 [17] were differentially expressed. Linkage and association studies as well as integrated analyses using genetic mapping and microarrays have revealed some genes that may be responsible for MD progression or resistance, such as GH, IFNγ and SULT[27]. However, it is difficult to find a consensus amongst these studies due to variation in experimental parameters such as, virus strain or in vitro derived samples. By using a genome-wide approach and three chicken lines with varying resistance to MD, we were able to generate a comprehensive list of candidate genes that can be used for studying MD-resistance and susceptibility. Besides finding some genes that were reported in previous studies, such as Mx1, we also found several genes that have not been reported before in this context, such as CD8α, IL8, USP18, and CTLA4. CD8α, present in the putative resistant gene list at 10dpi, codes a surface glycoprotein expressed on a subpopulation of cytotoxic T lymphocytes (CTLs) [36], which binds to the α3 domain or membrane-proximal domain of most of the known HLA class I molecules to enhance CTL activation [37–41]. It has been shown that CD8α was up-regulated by MDV infection at the early cytolytic stage (4dpi and 7dpi), whereas IgM and CD3 were down-regulated [34]. These are similar to our microarray results, the slight difference being possibly due to the differences in virus strains and genetic background of the chickens. The CD8α gene was significantly up-regulated at 10dpi in the MD-resistant chicken line (line 6₃) and RCS-M, but down-regulated in the MD-susceptible chicken line (line 7₂). In chickens vaccinated against MDV an increase of CD8α cells was found after MDV infection compared to unvaccinated chickens [32]. The vaccinated birds were phenotypically similar to line 6₃ and hence, this result is consistent with our finding. The above evidence, taken all together, leads us to speculate that CD8α plays an important role in MD resistance. The induction of CD8α gene may result from an increase of the CD8⁺ T cells that eliminate MDV infected cells in the resistant chickens. However, this scenario needs further validation.

In contrast, CTL-associated antigen-4 (CTLA-4), present in our putative susceptible gene list at 10dpi, is a member of the immunoglobulin superfamily expressed on the surface of an activated T cell [42]. It has been reported that the knockout of CTLA-4 resulted in a lymphoproliferative disorder and death in mice, which indicated a very important role of CTLA-4 in negative regulation of T cell activation [43]. The blockade of the CTLA-4 pathway results in a rejection of tumor [44, 45], indicating that a lower CTLA-4 expression may be important for antitumor response. Therefore, in humans, a current strategy of immunotherapy focuses on the blockade of the CTLA-4 pathway [46, 47]. A higher expression of CTLA-4 was detected at lymphoma lesions in MD-susceptible chickens at 21dpi, although no significant difference was found in the whole tissue [48]. Importantly, a similar result existed in our data: the fold change of CTLA-4 at 10dpi after MDV infection is much less in line 6₃ than in line 7₂ and RCS-M (Table 4), indicating a lower level of CTLA-4 involved in antitumor immune response.

In addition to the above genes, some networks and biofunctions were also observed to be different between MD-resistant and susceptible chickens. It is interesting to note that most of the differentially expressed genes were not enriched in biofunctions of immune related diseases, but with other diseases or metabolism. This is consistent with the fact that apart from the generation of tumors, MD-susceptible chickens also exhibit weight loss, paralysis and other symptoms. However, some immune response-related biofunctions were enriched only in line 6₃ and RCS-M chickens. It was thought for a long time that tumor cells have no antigen and this enables them to escape the host immune system. While the finding of the melanoma antigen in the late 1980's shed light on the role of immune system to fight against tumors [49], the tumor cells are known to also have immunosuppressive agents that help them evade detection and killing by the immune system [50, 51]. The networks related to immune response found in line 6₃ chickens, suggests that in these chickens the immune system is activated to counteract the development of tumor. In contrast, the transformed cells in susceptible chickens are able to escape the natural resistance of the immune system to generate tumors, although at present it is still unclear if this is due to a larger initial damage to the immune system [52] or the immunosuppression induced by MDV in line 7₂ chickens. NOS2 is an enzyme that catalyzes the generation of NO [53] which in turn increases the virulence of MDV by immunosuppression [54]. However, it has also been shown that NO has inhibitory effects on MDV replication [55] and NO production in MDV-infected susceptible chickens (MHC, B¹³B¹³) is the lowest in comparison to MD-infected resistant birds (MHC, B¹⁹B¹⁹) [55, 56]. Interestingly, we found that the NOS2 gene was up-regulated in susceptible chicken lines. Therefore, it remains to be seen whether the up-regulation of NOS2 in line 7₂ could induce immunosuppression and increase the risk of tumor generation in MD-susceptible chickens. The above results indicate that different immune response in resistant and susceptible chickens lead to the vastly different responses to MDV infection.

To minimize transcriptional variations and take full advantage of a similar genetic background in the inbred lines, we paired birds by line and dpi, respectively, and tested the difference between infected and non-infected individuals. This procedure not only led us to identify the genes most likely related to MD resistance and susceptibility, but also revealed a common broad influence of MDV infection on the hosts. By using IPA to analyze differentially expressed gene sets, we found focused pathways enriched in metabolism, tissue development, gene expression and cell cycle along with other immune-related pathways preferentially enriched in resistant chickens. These results suggested possible mechanisms and specific genes related to MD-resistance or -susceptibility. We hypothesized that there are four possible causes behind MD-resistance: some genes activated in resistant chickens can (i) cause loss of the MDV receptor, (ii) help to clear infected cells, (iii) affect the viral life cycle or (iv) prevent transformation of infected cells. Our observations and previous research have showed that the virus load in both resistant and susceptible chickens was similar at early stages of infection [18, 57], suggesting the presence of receptors for MDV in both resistant and susceptible chickens. Thus, the latter three of the aforementioned possibilities are more likely to be the main reasons for MD-resistance although it is not easy to say which of these play a big role in non-MHC associated resistance.

Using a comprehensive genome-wide study of gene expression in chicken lines with varying resistance to MD, we were able to identify pathways and genes that may be involved in MD-resistance and susceptibility. Phenotypic similarities between the chicken lines enabled us to narrow the list of putative genes to 108 genes associated with MD-resistance and 121 genes associated with susceptibility. Combining network analysis with differential gene expression analysis helped uncover high-confidence genes such as CD8α, IL8, USP18, and CTLA-4 and several immune-related biofunctions with potentially important consequences to MDV infection. Our findings add to the current understanding of the mechanism behind resistance and susceptibility to MD while expanding the scope of future studies with a comprehensive list of putative genes. Our approach also underlines the importance of comprehensive functional studies to gain valuable biological insight into the genetic factors behind complex disease.