Background
Deer mice (
Peormyscus maniculatus) and other peromyscine species are the most abundant mammals in North America and found from the sub-arctic to Mexico[
1]. The deer mouse genome has recently been sequenced to 7x depth but it has not yet been annotated, which presents difficulties for leveraging it as a resource. The genus hosts a number of pathogens, including the agents that cause granulocytic ehrlichiae (
Ehrlichia spp.), Lyme disease (
Borrelia burgdorferi), babesiosis (
Babesia microti), cryptosporidia (
Cryptosporidium parvum), bartenellosis (
Bartonella vinsonii), cutaneous leishmaniasis in Mexico and Central America (
Leishmania mexicana), deer tick virus, Powassan virus, and Sin Nombre (SNV) and New York-1 (NYV) hantaviruses (etiologic agents of hantavirus cardiopulmonary syndrome [HCPS])[
2‐
17]. These diseases cause significant morbidity and mortality in North and Central America each year, and some are important agricultural pathogens. Despite the importance of this genus as reservoirs and vectors, little is known about the immunology of peromyscines.
Hantaviruses are tri-segmented negative-stranded viruses hosted by rodents or insectivores[
18,
19]. Several pathogenic hantaviruses, which are rodent-borne, cause HCPS in the Americas or hemorrhagic fever with renal syndrome (HFRS) in Eurasia, both of which have prominent immunopathologic components. The two most important HCPS-causing hantaviruses are SNV, hosted by the North American deer mouse, and Andes virus (ANDV), hosted by the South American long-tailed pygmy rice rat (
Oligoryzomys longicaudatus). Reservoirs remain persistently-infected, without pathology, despite mounting an immune response that generates high-titer neutralizing antibodies within one to two months after infection[
20‐
23].
The presence of virus-specific IgG antibodies in reservoir rodents infected with their hantaviruses suggests a prominent role for helper T cell (Th)-mediated class switching and affinity maturation. Several Th cell subsets have been identified that provide distinct effector functions that mediate host responses to pathogens[
24]. These cells are distinguished by epigenetic expression of specific cytokines, cell surface antigens and transcription factors that results in limited plasticity after differentiation[
25]. While these factors have been exhaustively examined in the commonly-used laboratory house mouse (
Mus musculus), virtually no work has been conducted on natural reservoirs of zoonotic viruses, thus, it is not clear how the coevolutionary adaptations between these viruses and their reservoir hosts have influenced the host responses mediated by helper T cells.
Spillover from natural reservoirs to other rodent species has occasionally been observed in nature. Pinion mice (
Peromyscus truei) are naturally susceptible to SNV[
26,
27] but it is unknown if they develop disease, or if they clear the virus or remain persistently infected. Deer mice are experimentally susceptible to ANDV infection without disease; however, unlike SNV infection they clear ANDV[
28]. This dichotomous outcome provides an opportunity to study how persistence or clearance occurs in a reservoir host species. The lack of peromyscine-specific reagents has limited the experimental value of deer mice; however, the availability of its genome presents a substantial opportunity for examination of the host response. We devised a computational method to parse the unannotated deer mouse genome for “virtual cDNAs” of interest, which were then used to generate primers for real-time PCR arrays to assess gene expression in lymph node cell cultures from deer mice experimentally infected with Andes virus.
Discussion
The deer mouse is among a growing list of non-model rodent species to have its genome deep sequenced (National Human Genome Research Institute, 2011). Unfortunately, its genome is not annotated and, thus, presents challenges for exploiting it as a research tool. Moreover, considering that more than 50 other species of Peromyscus are found in North America and Mexico, its genome is an important resource for other investigators that work with this genus. Thus, we developed a method for rapidly identifying real-time PCR primer candidates for cDNAs of interest for the deer mouse that can be used for quantitative expression analysis. We used this tool to produce a 94-gene immunoarray to examine expression of a variety of T cell differentiation and antiviral genes in lymph node cell recall assays from deer mice infected with ANDV. While transcription data have several limitations, such as an inability to assess translational or post-translational events (e.g., phosphorylation), they can provide important information about how cells respond to antigenic stimuli and what pathways may be candidates for further scrutiny with protein-specific assays. This array showed that several immunological pathways were activated during in vitro lymph node cell recall responses to ANDV antigen, with a preponderance of Th2-like gene expression.
The majority of expressed subset markers were suggestive of Th2 cells (9 of 15 genes, 60%). Th2 cells historically have been considered important for antibody responses to extracellular pathogens and toxins and less important for viral infections. However, most studies examining T cell responses have focused on pathogenic viruses and since ANDV is not pathogenic in deer mice it may be that such a response is protective from persistent infection. Th1 and Treg genes were also upregulated, but no indication of a Th17 response, suggesting a mixed Th1, Th2 and Treg cell response. We found similar results with primary T cell lines from deer mice at 10 days post infection with SNV, in which
Ifng,
Il4,
Il5,
Il10 and
Tgfb were elevated[
23]. However, with ANDV infection, neither
Il5 or
Il10 expression appear elevated. It is noteworthy that the T cell lines from SNV infected deer mice in this previous work were of spleen origin, collected at 10 dpi and propagated on bone-marrow derived APC that were likely dendritic cells. In the present work, lymphocytes from lymph nodes were used, thus it is possible that cell source, timing and culturing methods could alter the expression of genes. However, since the present work used freshly isolated lymph node cells it is likely to be more relevant to natural infection compared to T cell lines, which require months to generate, may select for rare clones and which may be influenced by developmental cues from the APCs. Examination of IL-5, which is closely linked to IL-4 in other mammalian species and likely so in deer mice, and IL-10 may be important in the different outcomes of deer mice infected with ANDV (clearance) or SNV (persistence).
Seven of the 10 genes of the IL-4 signaling pathway on the array were upregulated;
Il4,
Il4ra,
Jak1,
Stat6,
Nfatc2ip,
Il13 and
Irf4.
Socs1,
Gata3 and
Il4rc, also part of the pathway, were not modulated. Activation of Th2 cells results in the dimerization of Nfatc2ip with IRF4, which were in the same expression cluster group (Figure
4B), within those cells and which is a transactivator of the IL-4 promoter[
30]. IL-4 is then secreted from Th2 cells and binds to its receptors on target cells, including B cells, where it activates the Jak1 pathway and induces dimerization of IRF4/STAT6 (which also clustered with
Irf4 and
Nfatc2ip) that leads to B cell activation, germinal center formation, class switching, maturation to plasma cells, affinity maturation and antibody secretion[
31,
32]. IL-4-stimulated B cells that also have CD86 (B7-2) engaged (from T cell CD28) increase expression of
Pou2f2 which, in laboratory house mice, leads to increased synthesis and secretion of IgG[
33].
Pou2f2 is also a marker of Treg cells, thus we cannot discriminate its expression in lymph node cultures, which have T cells and B cells. Considering the number of genes in this pathway that are expressed during ANDV infection, additional genes of the pathway should be examined. The IL-4 intercellular communication pathway has more than 20 proteins involved in cognate interactions of T cells and B cells, thus expression analysis and examination of phosphorylation events should be a high priority for future work.
Interferon-γ (IFNγ) induces the expression of several transcription factors, including members of the interferon response factor (IRF) family[
34]. IFNγ ligation induces the activation of STAT1, expression of which was elevated in all of the deer mouse lymph node cell cultures (Figure
3A, Additional file
3: Table S2), that, in turn, activates a number of other factors, including IRF1. IRF1 directs the differentiation of antigen-stimulated primary T cells towards the Th1 phenotype. IRF4 controls a variety of maturation events in both T and B cells, including the maturation of T cells toward the Th2, Th9 or Th17 phenotypes[
35], and induction of class switching, which precedes somatic hypermutation leading to affinity maturation[
36], and plasma cell formation in B cells, which also requires IL-4 and IL-6[
37]. IL-21, which contributes to germinal center plasma cell formation in the laboratory house mouse, was not upregulated. It is also involved in expression of MHC class II molecules in dendritic cells that are critical for initiating helper T cell responses to antigens. IRF4 expression in follicular helper T (Tfh) cells is essential for germinal center formation[
35]. IRF8 is expressed by germinal center B cells, but not plasma cells, where it participates in the regulation of more than 50 genes involved in B cell activation and maturation, and interaction with Tfh cells and follicular dendritic cells that present antigen in its native confirmation to germinal center B cells[
38].
Runx3 is expressed by B cells and promotes proliferation, presumably during clonal expansion induced by Th1 cells[
39].
Both deer mice that had increased levels of
Ifng mRNA (DM3 and DM6) also had high levels of
Tnf mRNA; however, infection with ANDV does not cause conspicuous inflammation in deer mice[
28]. Considering that IFNγ induces activation of STAT1, which was significantly elevated in all four deer mice, it is unclear how its absence results in increased
Stat1 gene expression; however, by 3 days post culture it is possible that
Ifng expression had already subsided in the two cultures that were
Ifng-negative. In addition to its role in T cell activation, STAT1 is also the major signaler/transcription factor for type I and type III IFNs and is typically a marker for innate signaling. In the LN cultures there is probably little or no active virus replication (although viral RNA was detected), but the innate response might be triggered by addition of antigen. It is also noteworthy that the important regulators of Th1, Th2 and Treg cells (T-bet, GATA3, Fox-p3, respectively) for the laboratory house mouse are not upregulated in deer mice infected with ANDV, yet many genes associated with the expression of these regulators are expressed, nonetheless. It may be that expression of those genes is unnecessary (i.e., the proteins are already present), or that by 14 days post infection they are no longer needed and are subsequently repressed. Most transcription factors are tightly regulated because they have enzyme-like activity; small amounts can profoundly affect downstream signaling cascades and gene expression. Regardless, it is clear that genes associated with three helper T cell subsets are expressed during the immune response to ANDV. It will be necessary to examine the activities of these transcription factors to clarify their roles; fortunately, many of these proteins are highly conserved and antibodies to their orthologs in other species (e.g., laboratory house mouse) are likely cross-reactive with many deer mouse proteins. The small sample size presented here (4 deer mice) limits interpretation of these data; however, the number of genes examined should allow extensive examination of the host response to clarify the pathways involved in deer mouse responses to infectious agents.
The development of software that allows rapid deployment of real-time PCR expression arrays for the deer mouse should help clarify the immunological events that govern hantavirus persistence or clearance from rodent reservoirs. The data presented here demonstrates that such work can be done, and that it generates substantial amounts of data indicating which pathways may, or may not, be involved in the host response. This should permit more efficient examination of these pathways using commercially available, cross-reactive antibodies to specific signaling molecules that may play decisive roles in hantavirus clearance.
Methods
Ethics statement and infections. All procedures using deer mice were in compliance with the USA Animal Welfare Act and approved by the Rocky Mountain Laboratories institutional animal care and use committee and performed following the guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care, International (AAALAC) by certified staff in an AAALAC-approved facility. All work involving infectious materials was performed in the biosafety level-4 laboratory at RML. Six deer mice were intramuscularly-infected with 200 FFU ANDV-9717869 in the hind limbs. Fourteen days later, deer mice were anesthetized by respiratory delivery of isoflurane for cardiac blood collection, then euthanized by thoracotomy. Lymph nodes were not recovered from one deer mouse (DM1) and another deer mouse’s lymphocytes (DM5) failed to proliferate in vitro in response to antigen. These two animals were excluded from the analysis in this work. Thus, the data were from the four remaining deer mice (DM2, DM3, DM4, DM6).
Lymphocyte recall assay. Cervical lymph nodes were collected in PBS and made into single cell suspensions in serum-free Hank’s balanced salt solution (HBSS). Cells were washed twice in HBSS, then once in complete medium (CM, 5% FBS RPMI-1640) and adjusted to 107/ml in CM. For each deer mouse, 500 μl of cells (5x106) were pipetted into wells of a 24-well plate in duplicate; one well received 500 μl of CM (basal gene expression levels) and the other received 500 μl of recombinant ANDV nucleocapsid (N) antigen in CM for a final concentration of 10 μg/ml. Cells were incubated for 72 hours at 37°C and 5% CO2 then collected and treated with RLT buffer for RNA extraction (RNA Easy kit, Qiagen). Following inactivation (according to approved standard operating procedures) the material was transferred from the BSL-4 to BSL-2 and frozen at −70°C.
Genome mining and primer design. Eighty-four genes of immunologic and antiviral relevance were selected to expand the deer mouse array previously describe by us[
22]. Laboratory house house mouse (
M. musculus) cDNA sequences were downloaded from Jackson Laboratories Informatics (
http://www.informatics.jax.org/) and BLASTed against the unannotated deer mouse genome (
ftp://ftp.ncbi.nih.gov/pub/TraceDB/peromyscus_maniculatus/) to generate cDNAs using software developed by us (
http://dna.publichealth.uga.edu/BlastPrimer/BlastPrimer.php). A stringent threshold cutoff of 1e
-10 was used in this software to minimize erroneous hits in the database. The BLAST results were then aligned and merged into “virtual cDNAs”, then compared to the house mouse sequence. Gaps in the peromyscus sequences were filled with ‘N’ to maintain approximate transcript size based upon the laboratory house mouse transcript length. Where nucleotide differences occurred, which could have been polymorphisms or sequencing errors, the most common nucleotide was used (“majority rule”). The virtual cDNAs were piped to Primer3 software (
http://frodo.wi.mit.edu/) to generate candidate primer lists and product size estimates (product size 75–120 bp, percent G +C 47–53%) as previously described[
29,
40]. Candidate primers were selected and validated for SYBR green PCR using deer mouse splenocytes stimulated with 2 μg/ml concanavalin A (C-0412, Sigma, St. Louis, MO) in CM overnight prior to RNA extraction and real-time PCR of single products as described in the next section.
PCR array. Primers (Additional file
1: Table S1) were purchased in two 96-well plates, with forward primers in one and reverse primers in the other (Operon, Huntsville, AL). Primers were resuspended to 100 μM in water and combined in a third 96-deep well plate (Axygen P-1ML-SQ-C) to 10 μM (500 μl per well final volume) that was sealed with a silicone mat (Axygen CM-96-RD). Two μg of total RNA were reverse transcribed into cDNA using the RT2 cDNA Synthesis Kit (SABiosciences), which contains a genomic DNA elimination step. The cDNA was added to RT2 SYBR green I master mix (SABiosciences) and mixed thoroughly. Twenty microliters were dispensed into a 96-well real-time PCR plate using an 8 channel pipettor, then 5 μl of primers (2 μM final concentration) of primers were added. Cycling was 95°C for 30 sec and 60°C for 30 sec for 40 cycles, followed by an 80-step melt curve analysis (Bio-Rad iQ5 thermal cycler). The ΔΔCt method[
41] was employed using the mean of
Gapdh as the reference within samples for normalization, and comparison of normalized samples between antigen-stimulated and CM only to calculate fold change. Heat maps and hierarchical clusters were generated using the heatmap.2 (“gplots” package) from R statistical software (
http://www.r-project.org/). Distance matrix was calculated using the Euclidean method, and a complete linkage clustering method was used for generating the hierarchical clusters. To allow identical color codes across cell gene sets, maximum and minimum logfold changes across all gene expression were included as references indicated by Ref(Max) and Ref(Min).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
TS, HF and JP designed experimental protocols and executed laboratory work. TIS and TG developed the software for mining the deer mouse genome, parsing Primer3 and statistical analysis of gene expression levels. TS wrote the manuscript. All authors read and approved the final manuscript.