Stress Triggers Expression of Bovine Herpesvirus 1 Infected Cell Protein 4 (bICP4) RNA during Early Stages of Reactivation from Latency in Pharyngeal Tonsil

ABSTRACT

Bovine herpesvirus 1 (BoHV-1), an important pathogen of cattle, establishes lifelong latency in sensory neurons within trigeminal ganglia (TG) after acute infection. The BoHV-1 latency-reactivation cycle, like other alphaherpesvirinae subfamily members, is essential for viral persistence and transmission. Notably, cells within pharyngeal tonsil (PT) also support a quiescent or latent BoHV-1 infection. The synthetic corticosteroid dexamethasone, which mimics the effects of stress, consistently induces BoHV-1 reactivation from latency allowing early stages of viral reactivation to be examined in the natural host. Based on previous studies, we hypothesized that stress-induced cellular factors trigger expression of key viral transcriptional regulatory genes. To explore this hypothesis, RNA-sequencing studies compared viral gene expression in PT during early stages of dexamethasone-induced reactivation from latency. Strikingly, RNA encoding infected cell protein 4 (bICP4), which is translated into an essential viral transcriptional regulatory protein, was detected 30 min after dexamethasone treatment. Ninety minutes after dexamethasone treatment bICP4 and, to a lesser extent, bICP0 RNA were detected in PT. All lytic cycle viral transcripts were detected within 3 h after dexamethasone treatment. Surprisingly, the latency related (LR) gene, the only viral gene abundantly expressed in latently infected TG neurons, was not detected in PT during latency. In TG neurons, bICP0 and the viral tegument protein VP16 are expressed before bICP4 during reactivation, suggesting distinct viral regulatory genes mediate reactivation from latency in PT versus TG neurons. Finally, these studies confirm PT is a biologically relevant site for BoHV-1 latency, reactivation from latency, and virus transmission.

IMPORTANCE BoHV-1, a neurotropic herpesvirus, establishes, maintains, and reactivates from latency in neurons. BoHV-1 DNA is also detected in pharyngeal tonsil (PT) from latently infected calves. RNA-sequencing studies revealed the viral infected cell protein 4 (bICP4) RNA was expressed in PT of latently infected calves within 30 min after dexamethasone was used to initiate reactivation. As expected, bICP4 RNA was not detected during latency. All lytic cycle viral genes were expressed within 3 h after dexamethasone treatment. Conversely, bICP0 and the viral tegument protein VP16 are expressed prior to bICP4 in trigeminal ganglionic neurons during reactivation. The viral latency related gene, which is abundantly expressed in latently infected neurons, was not abundantly expressed in PT during latency. These studies provide new evidence PT is a biologically relevant site for BoHV-1 latency and reactivation. Finally, we predict other alphaherpesvirinae subfamily members utilize PT as a site for latency and reactivation.

INTRODUCTION

Bovine herpesvirus 1 (BoHV-1), an alphaherpesvirinae subfamily member, causes respiratory disorders, conjunctivitis, genital infections, encephalitis, abortions, and multi-systemic fatal disease in neonates. Acute infection immunosuppresses cattle, which predisposes cattle to secondary bacterial colonization, severe pneumonia, and even death (1). Consequently, BoHV-1 is a cofactor of bovine respiratory disease complex (BRDC) (2, 3), which costs the US cattle industry approximately $540 million in direct costs and $5 billion in indirect costs each year (4). Furthermore, BoHV-1, including commercially modified live vaccines, can initiate reproductive complications in pregnant cows, including abortion (5).

During productive infection, herpesviruses exhibit a well-regulated temporal cascade of viral gene expression that is divided into 3 stages: immediate early (IE), early (E), and late (L), reviewed in (2). Three BoHV-1 proteins are encoded by immediate early (IE) mRNAs: infected cell protein 0 (bICP0), bICP4, and bICP22 (6–9). A BoHV-1 tegument protein, VP16, specifically transactivates IE promoters during productive infection (10, 11). BoHV-1 contains a promoter, immediate early transcription unit 1 (IEtu1), that controls IE expression of IE/2.9 and IE/4.2 mRNAs, which are generated by alternative splicing (6–8). IE/2.9 mRNA is translated into the bICP0 protein, and IE/4.2 mRNA is translated into the bICP4 protein (6–8). A separate E promoter drives subsequent expression of an early transcript that also encodes the bICP0 protein (6–8, 12). The bICP0 and bICP4 proteins are crucial for productive infection because they encode proteins that stimulate viral transcription and productive infection. Early, and late genes are generally organized as other alphaherpesvirinae subfamily members and have similar functions.

Following acute infection of oral, ocular, or nasal cavities, sensory neurons within trigeminal ganglia (TG) are an important site for BoHV-1 latency. Stressful stimuli, including environmental heat, transporting cattle long distances, and/or restriction of water and food, increases the incidence of BoHV-1 reactivation from latency in cattle, reviewed in (13–15). Notably, the synthetic corticosteroid dexamethasone (DEX), which mimics the effects of stress, consistently induces reactivation from latency (16–18). The only known BoHV-1 genes abundantly expressed in latently infected neurons are the latency related (LR) gene (15) and ORF-E (19), which is upstream of the LR gene. LR gene products establish and maintain neuronal latency in calves (20), in part because these products interact with cellular transcription factors (21–25), interfere with bICP0 protein synthesis (26), promote neuronal differentiation (27, 28), and inhibit apoptosis (24, 29). In TG neurons of latently infected calves, LR gene products promote latency by directly or indirectly enhancing expression of more than 100 cellular genes that encode proteins essential for the Wnt/β-catenin and Akt signaling pathways (22, 23). In contrast to dividing cells, these cellular signaling pathways promote neuronal survival, axonal growth and maintenance, and proper synaptic connections (30–36).

In addition to TG neurons, BoHV-1 DNA is consistently detected in pharyngeal tonsil (PT) of latently infected cattle (18, 37–39). In contrast to TG neurons, LR-RNA expression is only detected in PT of latently infected calves when nested PCR is performed, indicating this gene is not abundantly expressed (18). By 6 h after latently infected calves are treated with DEX to induce reactivation from latency, in situ hybridization revealed lytic cycle viral transcripts are detected (18), suggesting reactivation occurs in PT and not indirectly via TG neurons. The intensity of in situ hybridization and higher number of PT cells expressing viral transcripts increase as a function of time after DEX treatment, which correlates with detection of infectious virus from tonsil swabs (40). In addition to PT, BoHV-1 DNA is consistently detected in peripheral blood mononuclear cells (41), lymph nodes, and spleen where infectious virus is not readily detected (42). Viral DNA from other alphaherpesvirinae subfamily members, including Pseudorabies virus (43, 44), equine herpesvirus 4 (45), and canine herpesvirus 1 (46), is also detected in PT and other lymphoid tissue during latency. Furthermore, a report concluded HSV-1, but not HSV-2 or human cytomegalovirus DNA, is detected in a subset of people who underwent tonsillectomy or adenectomy because of chronic lymphoid hyperplasia without evidence of acute HSV-1 infections (47). As expected, this report also detected Epstein-Barr Virus in PT of these patients.

The objectives of this study were to identify viral and cellular genes expressed in PT during early stages of DEX-induced reactivation from latency. Using RNA-sequencing (RNA-seq), our studies revealed the bICP4 transcript was readily detected in PT within 30 min after DEX treatment but not prior to DEX treatment. These findings support the premise that PT is a biologically relevant site for BoHV-1 reactivation from latency and virus transmission. We predict PT is an important site for reactivation for other “neurotropic” herpesviruses.

RESULTS

Identification of viral transcripts expressed in PT during early stages of reactivation from latency.

RNA-seq studies were performed to identify viral genes expressed in PT during early stages of DEX-induced reactivation from latency. For these studies, male Holstein calves were infected with a wild type (wt) virulent BoHV-1 strain (Cooper) as described previously (20, 22); see Fig. 1A for schematic summarizing the calf studies. As expected, viral transcripts were not detected in uninfected calves: however, low levels of viral transcripts were detected during latency (calves not treated with DEX) (Fig. 1B). Notably, a region localized to the left terminus of the viral genome yielded low levels of transcripts during latency (Fig. 2A). All viral reads from this region during latency contained the repeat GCTCCTCCTCCCTC. Further inspection revealed these reads span the terminus but were not derived from a single transcript. The BoHV-1 HindIII K fragment spans the left hand terminus of the viral genome and exhibits size heterogeneity in viral strains because the GCTCCTCCTCCCTC repeat can be present at different copy numbers (48). Consistent with previous studies (18), the LR transcript was not readily detected in PT of latently infected calves. In stark contrast, the LR transcript is abundantly expressed during latency in TG neurons (16, 49).

FIG 1.

Schematic of infection and frequency of viral transcription identified in PT. (A) Summary of calf studies, including important milestones during these experiments. Abbreviations: minutes (‘), hours (H), and day (D). (B) Identification of viral reads and % of total viral reads during latency and the time points during DEX-induced reactivation from latency in PT of calves latently infected calves.

FIG 2.

Viral transcripts expressed in PT during latency and DEX-induced reactivation from latency. (A) Schematic of BoHV-1 genome and position of bICP0, bICP4, and LR gene. Frequency of viral genes during latency and reactivation from latency. During latency, a family of transcripts containing the denoted repeat were expressed. (B) Distribution of RNA reads within the bICP4 transcript across the bICP4 gene. (C) Distribution of RNA reads within the bICP0 transcript across the bICP0 gene.

Latently infected calves (60 days after infection but not treated with DEX) were given a single IV injection of water-soluble DEX to initiate reactivation from latency (Fig. 1A). Within 30 min after DEX treatment, the number of viral transcripts detected in PT were 300-fold higher relative to latency (Fig. 1B). Strikingly, nearly all induced viral transcripts mapped to the bICP4 gene at 30 min after DEX treatment (Fig. 2A; 30 m DEX) and 90 min after DEX treatment (90 m DEX). By 90 min after DEX treatment, bICP0 RNA was also detected at low levels. These proteins are predicted to be expressed because RNA reads span the bICP4 (Fig. 2B) and bICP0 (Fig. 2C) ORFs. At 30 and 90 min after DEX treatment, transcripts within the HindIII K fragment also increased. By 3 h after DEX treatment, there was more than a 1,000-fold increase in viral transcript abundance compared to latency (Fig. 1B), and these viral transcripts spanned the entire genome. Simultaneously bICP4 reads were reduced relative to the rest of the viral genome. The bICP4 protein stimulates early and late promoters but autoregulates its expression by impairing IEtu1 promoter activity (7, 50). These activities partially explain why bICP4 transcripts were reduced relative to other lytic cycle viral genes during early stages or DEX-induced reactivation. Collectively, these studies revealed DEX treatment of latently infected calves rapidly induced expression of bICP4 RNA in PT culminating in expression of all lytic cycle viral genes within 3 h after DEX treatment.

The bICP4 protein is expressed in PT during early stages of reactivation.

Immunohistochemistry (IHC) studies were performed to confirm the bICP4 protein was expressed during early stages of reactivation from latency. The bICP4 antibody used for this study is directed against two bICP4 peptides, and bICP4 specific antibodies were affinity purified. This antibody specifically recognizes the bICP4 protein in a Western Blot and TG neurons by IHC (51). An infected calf treated with DEX for 90 min was examined for bICP4 antibody staining and to test whether bICP4 expression was localized to specific PT regions. PT contain numerous lymphoid follicles (Fig. 3A: denoted by F). The epithelium directly surrounding follicles is referred to as reticular epithelium (RE) (52), which contains epithelial cells, Micro-fold cells, lymphocytes, macrophages, dendritic cells, intraepithelial vasculature, and a discontinuous basement membrane. Non-reticular epithelium (N-RE) is present between follicles, and is comprised of pseudostratified columnar or squamous epithelium, occasional non-epithelial cells, and vascular structures. Studies focused on the boxed area of PT in Fig. 3A revealed weak bICP4 staining in follicles and RE 90 min after DEX treatment (Fig. 3B). Similarly, in situ hybridization study detected BoHV-1 transcripts in follicles and reticular epithelium (18). A couple of cell clusters exhibited higher levels of bICP4+ cells (denoted by yellow arrows). While weak background staining of PT cells from latently infected cows was present, there were no clusters of cells in latently infected calves that were bICP4+.

FIG 3.

Detection of bICP4 protein expression in PT. IHC was performed using the bICP4 peptide specific IgG as described in the materials and methods. (A) 90’ DEX is PT from a calf treated with DEX for 90 min. F denotes a follicle, N-RE is the non-reticular epithelium, and RE denotes reticulum epithelium. (B) is a higher magnification of the boxed area in (A). (C) PT from a calf latently infected with BoHV-1 (L) and incubated with the bICP4 antibody. Black arrows denote cells weakly stained by the bICP4 antibody. Yellow arrows denote a cluster of PT cells stained by the bICP4 antibody.

Additional studies were performed to confirm bICP4 was expressed during early stages of reactivation. Three hours after DEX treatment of latently infected calves, immunofluorescence (IF) was used to detect bICP4 expression. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and then IF was performed using the bICP4 antibody (Fig. 4). The rational for initially performing TUNEL is lymphoid cells are prone to apoptosis following corticosteroid treatment, reviewed in (53). Furthermore, published studies correlated induction of apoptosis and reactivation of certain herpesviruses, including HSV-1 latently infected TG (54) and cell lines latently infected with Kaposi’s sarcoma-associated herpesvirus (HHV-8), HHV-6A, HHV-6B, or Epstein-Barr virus (55, 56). TUNEL+ cells were detected in 2 different latently infected calves treated with DEX for 3 h (Fig. 4A and C; black arrows). Furthermore, bICP4+ cells were readily detected by 3 h after DEX treatment (Fig. 4B and D; white arrows). Notably, most of the bright staining bICP4+ cells appeared to be adjacent to TUNEL+ PT cells. PT sections from latently infected calves did not contain discernible TUNEL+ cells (Fig. 4E) or bICP4+ cells (Fig. 4F). In summary, these studies revealed the bICP4 protein was readily detected by 3 h after DEX treatment and there was a correlation between proximity of cells undergoing programmed cell death and cells expressing bICP4.

FIG 4.

Detection of bICP4 protein expression and TUNEL+ cells in PT 3 h after DEX treatment. TUNEL and IF using the bICP4 antibody were performed as described in Materials and Methods. TUNEL studies are shown in (A), (C), and (E). IF is shown (B), (D), and (F). PT from 2 different calves were treated with DEX for 3 h (A) and (B), and (C) and (D). PT sections from a latently infected calf are shown in panels E and F. For all panels, magnification was 40X.

Detection of bICP0 and VP16 protein 6 h after DEX treatment.

bICP0 and VP16 protein expression was weakly detected by IHC at 3 h after reactivation (data not shown). Additional studies examined bICP0 and VP16 expression in PT 6 h after calves were treated with DEX. PT from latently infected calves treated with DEX for 6 h were used for these studies because in situ hybridization studies detected BoHV-1 transcripts in follicles and reticular epithelium (18). bICP0 (Fig. 5A) and VP16 (Fig. 5B) positive PT cells were readily detected 6 h after DEX treatment (Fig. 5). As expected, staining of bICP0 (Fig. 5C) and VP16 (Fig. 5D) was not readily detected in PT sections from calves latently infected with BoHV-1. In summary, these studies revealed that numerous PT cells expressed bICP0 and VP16 by 6 h after DEX, which supports RNA-seq studies indicating the entire lytic cycle program of viral gene expression was activated by DEX treatment.

FIG 5.

Expression of bICP0 and VP16 protein expression in PT during reactivation from latency. IHC was performed using the bICP0 or VP16 peptide specific IgG as described in the materials and methods. IHC was performed using PT sections from latently infected calves 6 h (H) after DEX treatment using the bICP0 peptide specific antibody (A) or VP16 peptide specific antibody (B). IHC was also performed with calves latently infected with BoHV-1 (no DEX treatment) using the bICP0 antibody (C) or VP16 antibody (D).

DEX treatment progressively induces cell death in PT during reactivation from latency.

To further explore programmed cell death during reactivation, TUNEL assays were performed using additional time points during DEX-induced reactivation and latency (Fig. 6). Low levels of TUNEL+ cells were detected in tonsil sections prepared from latently infected calves. Within 30 min after latently infected calves were treated with DEX (30’ DEX panel), less than a 2-fold increase in the number of TUNEL+ cells were detected. However, 3 h after DEX treatment (3 H DEX panel) an 11-fold increase in TUNEL+ cells was consistently observed in several sections from different calves. Staining of PT sections with a cleaved caspase 3 antibody detected weak staining at 3H after DEX treatment (data not shown) suggesting a subset of cells underwent apoptosis. This observation was supported by differential gene expression analyses, which demonstrated that apoptotic processes (P-value = 0.0010) and positive apoptotic processes (P-value 0.0037) in the Reactome pathways/processes (57, 58) (data not shown). In summary, these studies revealed DEX-induced higher levels of programmed cell death in PT within 3 h after, latently infected calves were treated with DEX, which correlated with expression of all viral genes in PT.

FIG 6.

Programmed cell death increases in PT following DEX treatment. TUNEL was performed on PT sections as described in the materials and methods. Samples were prepared from latently infected calves (latency), latently infected calves treated with DEX for 30 min (30’), or 3 h (3H). Black arrowheads denote TUNEL+ cells.

Identification of transcription factors stimulated in PT during early stages of reactivation from latency.

Since bICP4 was the first viral transcript detected during DEX-induced reactivation from latency, we hypothesized that cellular transcription factors stimulate bICP4 expression. Consequently, we identified differentially expressed genes (DEGs) at 30 min after DEX treatment compared to latency and specifically looked for DEGs that encode transcription factors. The rational for focusing on 30 min after DEX treatment is lytic cycle viral gene expression (bICP4) was readily detected suggesting differentially expressed cellular transcription factors at this time mediated bICP4 expression. The pituitary homeobox 1 (PITX1), the #1 most highly induced transcription factor 30 min after DEX treatment (Table 1), cooperates with GR to transactivate a simple promoter containing nuclear hormone receptor binding sites (59). Furthermore, this study demonstrated PITX1 did not cooperatively transactivate androgen, estrogen-α, estrogen-β, progesterone, or retinoic acid receptors suggesting PITX1 is specific for GR-mediated transactivation. The ovo-like zinc finger 1 (OVOL1) regulates epithelial differentiation during embryonic development in numerous tissues and also regulates stemness of cancer cells (60). Krüppel like factor 15 (KLF15) and Zinc finger and BTB domain containing 16 (ZBTB16), also known as the promyelocytic leukemia zinc finger protein (PLZF), are members of the KLF family of transcription factors and are DEGs in TG neurons of calves during stress-induced reactivation (61). Notably, GR and KLF15 form a feed-forward transcription loop (62, 63) and cooperatively transactivate the IEtu1 promoter (64) and increase bICP4 expression (7, 50). GR and KLF15 also cooperatively transactivate key HSV-1 promoters (ICP0, ICP4, and ICP27) and silencing KLF15 expression impairs productive infection (65–68). ZBTB16 stimulates BoHV-1 productive infection better than other stress-induced transcription factors identified in TG (69): however, the only viral promoter we have identified to be transactivated by ZBTB16 alone or with GR is the HSV-1 ICP4 promoter (68). The other induced transcription factors may mediate certain aspects of stress-induced reactivation from latency: however, their roles are not as clear as PITX1, ZBTB16, or KLF15. For example, Homeobox Protein Distal-Less 3 (DLX3), specifically binds DNA and is a member of the distal-less family of non-Antennapedia homeobox genes, which contributes to mammalian trophoblast functions (70, 71). Dual specificity phosphatase 26 (DUSP26) regulates the mitogen activated protein kinase (MAPK) signaling pathway, reviewed in (72), suggesting it mediates DEX-induced transcription. We predict a subset of cellular transcription factors upregulated by DEX directly or indirectly stimulate lytic cycle viral gene expression during early stages of reactivation in PT.

TABLE 1.

Top 20 list of transcription factors increased within 30 min after latently infected calves were treated with DEX. For details, see text

Gene	Fold-Change (log2)	p-value
PITX1	2.95522	5e-05
OVOL1	2.57038	5e-05
DUSP26	2.23067	5e-05
YBX2	1.80612	0.0001
ARSE	1.66924	5e-05
ZBTB16	1.62876	0.0007
TFCP2L1	1.48281	0.0003
NFE2	1.38612	0.00085
DLX3	1.36805	0.00455
KLF15	1.33602	5e-05
RARG	1.30147	5e-05
MYBPH	1.28599	0.0006
MYBPC1	1.22954	5e-05
PLEK2	1.2231	0.00105
ARHGEF19	1.21883	5e-05
HOPX	1.21118	5e-05
CARD14	1.17831	0.0005
IRX3	1.17204	0.0002
TUNAR	1.09796	0.00295
ELF3	1.05685	5e-05

Identification of transcription factors with reduced expression in PT during early stages of reactivation from latency.

The ligand dependent nuclear receptor corepressor like (LCoRL) gene was the #1 downregulated cellular transcription factor at 30 min after DEX treatment (Table 2). LCoRL is a nuclear protein reported to impair agonist-activated GR (69). LCoRL interacts with certain histone deacetylases and C-terminal binding protein (CtBP) corepressors and recruits this complex to GREs or GR bound DNA. LCoRL expresses 5 to 7 isoforms, varying from 319 to 1904 amino acids and is strongly associated with body weight in livestock via interactions with NCAPG (non-SMC condensin I complex subunit G), reviewed in (73). NCAPG is a subunit of chromatin condensin1 and critical for mitotic chromatin condensation. Interestingly, 6 zinc finger proteins (ZC2HCA1, ZNF713, ZNF711, ZNF367, ZNF148, ZNF148) were downregulated in PT at 30 min after DEX treatment. Zinc finger proteins interact with RNA, DNA, poly-ADP-ribose, and/or proteins. Certain zinc finger transcription factors are linked with DEX-induced cell differentiation, reviewed in (74), which correlates with DEX-induced activation of keratinocyte differentiation pathways in PT (data not shown). We suggest that developmental changes in PT induced by DEX treatment (directly or indirectly) promotes viral reactivation from latency.

TABLE 2.

Top 20 list of transcription factors decreased within 30 min after latently infected calves were treated with DEX. For details, see text

Gene	Fold-Change (log2)	p-value
LCORL	−1.96098	5e-05
SPESP1	−1.77104	0.00335
LARP4	−1.65137	5e-05
ZC2HC1A	−1.59249	5e-05
ZNF713	−1.58335	5e-05
MARCH7	−1.45943	5e-05
MIER1	−1.44925	5e-05
ZNF711	−1.43753	5e-05
AP1AR	−1.40514	5e-05
PLEKHF2	−1.35821	5e-05
ZNF367	−1.35634	5e-05
ZNF148	−1.32711	5e-05
ARL6	−1.31382	5e-05
RERGL	−1.31376	0.00495
HLTF	−1.30188	5e-05
ZNF644	−1.27672	5e-05
KLF10	−1.26841	5e-05
MIS18BP1	−1.26382	5e-05
AEBP2	−1.24726	5e-05
IKZF5	−1.19533	5e-05

La-related protein-4 (LARP4) interacts with poly(A)-binding protein (PABP) and mediates T cell activation by stabilizing NFκB1 mRNA levels, which culminates in stimulating expression of interleukin-2 and interferon gamma (75). Interestingly, BoHV-1 was previously shown to infect activated CD4⁺ T cells in culture (76) and in acutely infected calves (38). The membrane-associated RING-CH-type finger (MARCHF7) is an E3 ubiquitin ligase, reviewed in (77). Up-regulated ubiquitination genes are enriched in samples from 30 and 90 min after DEX treatment supporting MARCHF7 mediated ubiquitination of NLRP3 (NOD LRR- and pyrin domain containing protein 3), a critical mediator of an inflammasome activated by numerous pathogens, reviewed in (78). Mesoderm induction early response 1 (MIERI) expresses 2 major isoforms that recruit histone deacetylase 1 (HDAC1) to repress transcription (79). Furthermore, MIERI interacts with Creb-binding protein (CBP) and inhibits the histone acetyltransferase activity of CREB. KLF10 overexpression is sufficient to induce apoptosis in epithelial cells and lymphoid cells and can impair growth of tumors, reviewed in (80). In general, transcription factors downregulated by DEX during early stages of reactivation in PT repress transcription of genes that stimulate innate immune responses, inflammation, and cell death.

Summary of differentially expressed cellular signaling pathways during early stages of reactivation from latency in PT.

To better understand changes that occur in PT within 30 min after DEX treatment, we identified differentially expressed genes that regulate key cellular signaling pathways. At 30 min after DEX treatment, neutrophil degranulation was strongly induced in PT (Fig. 7). Neutrophils are present on the surface of PT (81), and are present in mucosal secretions of PT (82). Interestingly, corticosteroids prevent apoptosis in neutrophils and exhibit pro-inflammatory and anti-inflammatory effects, reviewed in (83, 84). While neutrophil degranulation may not have a direct effect on viral gene expression during early stages of reactivation from latency, it is predicted to impact the PT environment after DEX treatment.

FIG 7.

Reactome of immune system and neutrophil degranulation pathway. Overrepresentation chart of differentially expressed genes in PT at 30 min after latently infected calves were treated with DEX. Neutrophil degranulation exhibited a highly signficant increase compared to PT in latently infected calves. Activation of RAS in B-cells was also overrepresented. The color legend in the top right of the figure reflects P-values.

RhoV, a member of the Rho family of GTPases, is a family of small G proteins that belongs to the Ras superfamily of GTPases, reviewed in (85). Although it is not clear whether RhoV exchanges GDP or GTP, it is primarily bound to GTP. Rho signaling pathways are significantly overrepresented in PT 30 min after DEX treatment, which supports its role in stimulating cell death (Fig. 8). The exact role that Rho GTPases is currently being examined in the context of BoHV-1 viral replication and gene expression.

FIG 8.

Reactome of Rho GTPase signaling pathways. Overrepresentation chart of differentially expressed genes in PT at 30 min after latently infected calves were treated with DEX. Color legend reflecting P-values is shown in top right.

DISCUSSION

Stress, as mimicked by DEX treatment, rapidly induced bICP4 RNA expression in PT of latently infected calves. The HSV-1 ICP4 protein is essential for productive infection (86) because it activates early (E) and late (L) gene expression (87). ICP4 specifically binds many sites on the viral genome (88) and interacts with the TATA-binding protein and RNA polymerase II coactivators (89, 90) to stimulate E and L viral gene expression (87, 90, 91). There is also a correlation between the ability of ICP4 to stimulate viral transcription and increase histone dynamics (92). Since bICP4 is the functional analogue of HSV-1 ICP4, we predict bICP4 has similar functions and efficiently triggers BoHV-1 lytic cycle gene expression resulting in virus production and successful reactivation from latency. This prediction is supported by the finding that all viral genes were detected in PT at 3 h after DEX treatment and infectious virus is readily detected in PT swabs during DEX-induced reactivation (40).

The initial events that promote DEX-induced reactivation from latency (host and virus transcripts) in PT appear to be unique relative to TG (Fig. 9A). For example, bICP4 was readily detected in PT at 30 min after DEX treatment, which was prior to bICP0 and VP16 detection. Conversely, bICP0 and VP16 expression are readily detected in TG neurons within 1 h after latently infected calves are treated with DEX (93, 94). Furthermore, bICP4 and the other IE protein (bICP22) are detected after bICP0 and VP16 in TG neurons (51). Late viral gene expression was readily detected in PT by 3 h after DEX treatment whereas in TG 2 late viral proteins, glycoprotein C (gC) and gD, are not readily detected until 6 h after DEX treatment (93). Studies focused on identifying viral genes expressed in TG relied on IHC studies and viral specific antibodies because RNA-seq studies of whole TG did not readily detect lytic cycle viral gene expression. When compared to PT, we suggest only a subset of latently infected neurons express low levels of viral genes following DEX treatment.

FIG 9.

Schematic of reactivation from latency in PT versus TG and schematic of BoHV-1 IEtu1. (A) Summarizes the differences observed during reactivation from latency in TG versus PT. (B) BoHV-1 genome and location of unique long (UL) region, direct repeats (open rectangles), and unique short region (U_S). IE/4.2 mRNA encodes the bICP4 protein and IE/2.9 mRNA encodes the bICP0 protein. An IE promoter activates expression of IE/4.2 and IE/2.9 and is designated IEtu1 (black rectangle) (7, 8). E/2.6 is the early bICP0 mRNA and its expression is driven by the bICP0 early promoter (E pro; gray rectangle). bICP0 protein coding sequences are in Exon 2 (e2). An origin of replication (ORI) separates IEtu1 from IEtu2. IEtu2 promoter (IEtu2 pro) drives IE1.7 mRNA expression, which is translated into the bICP22 protein. Solid lines in IE/2.9, IE/4.2, and IE/1.7 are exons (e1, e2, or e3) and dashed lines introns. (C) Full length IEtu1 promoter showing direction of transcription (arrow) start site of IEtu1 transcription (black triangle), TATA box, binding site for VP16/Oct1 complex (TAATGARAT), and location of GRE#1 plus GRE#2. Additional transcription factor binding sites in the IEtu1 promoter are denoted. ETs-1 belongs to the ETS (erythroblast transformation-specific) transcription factor family. The alternative bICP4 (blue triangle) and bICP0 (green triangle) mRNA start sites are upstream of the initiating methionine (96, 97).

Based on studies in this report and previously published studies (51, 61, 93, 94), we predict different anatomic specific cellular factors quickly activate bICP4 expression in PT relative to TG. For example, PITX1 and LCoR were differentially expressed in PT but were not identified in TG during early stages of reactivation from latency (61) (Fig. 9A). Downregulation of LCoR, a GR transcriptional repressor, during early stages of reactivation in PT implies this is an important step during reactivation from latency in PT but not TG. LCoR is also predicted to directly participate in maintaining latency by repressing IEtu1 promoter activity. Two stress-induced transcription factors, KLF15 and ZBTB16, were differentially expressed in TG and PT suggesting these common factors mediate reactivation in TG and PT. However, KLF4 and Slug expression were induced in TG (61), but not PT, during DEX-induced reactivation from latency. Unlike reactivation from latency in PT, VP16 is readily detected in TG neurons during early stage of reactivation from latency (93, 94), suggesting VP16 plays a prominent role with respect to driving IE gene expression in TG, but not PT.

IEtu1 is a crucial BoHV-1 regulatory locus because it contains coding sequences for bICP0, bICP4 (6–8), and the LR gene (15). The finding that bICP4 was the most abundant viral gene expressed at 30 and 90 min after DEX treatment was unexpected because the IEtu1 promoter was initially reported to drive bICP4 and bICP0 expression from a common transcriptional start site under immediate early conditions during productive infection (6–8) (Fig. 9B; denoted by arrow). This initial study concluded the start site for the IETu1 transcript is 24–31 nucleotides downstream of the TATA box (black diamond). In contrast, a recent study concluded bICP0 and bICP4 contain distinct transcriptional start sites (Fig. 9C; bICP4 and bICP0 start sites are denoted by blue and green diamonds) (95). With respect to this recent study, the bICP0 RNA start site is 726 bases downstream from the TATA box in IETu1 and the bICP4 RNA start site is 826 bp downstream of the TATA box. Both RNA start sites are upstream of the initiating methionine residue reported for these proteins (denoted by ATG in Fig. 9C). Although there is not a consensus TATA box in close proximity to the bICP0 or bICP4 start sites, promoters lacking a TATA box have been described and the cellular transcription factor specificity protein 1 (Sp1) is linked to regulating transcription (96, 97). Three consensus Sp1 binding sites (CCCGCC) are downstream of the TATA box, and prior to the initiating methionine for bICP0 and bICP4. The only TAATGARAT motif (Oct-1 binding site) in the IEtu1 promoter region is upstream of the TATA box. Regardless of how bICP4 is selectively expressed during early stages of reactivation in PT, 2 functional glucocorticoid response elements (GREs) in the IEtu1 promoter/regulatory region (65, 98) are predicted to mediate stress-induced promoter activation of bICP4 RNA expression during early stages of reactivation from latency in PT. Although GRE#1 is 1,006 bases upstream of the TATA box (98), certain GREs can stimulate transcription when 5–19 kb from the start site of transcription (99).

The studies in this report and a previous study (18), revealed cells in PT follicles and reticular epithelium contain viral genomes and support viral gene expression during early stages of reactivation from latency. Several distinct cell types are present in PT follicles and reticular epithelium. For example, the follicle is generally comprised of lymphoid cells whereas reticular epithelium contains epithelial cells, Micro-fold cells, lymphocytes, macrophages, and dendritic cells. Micro-fold cells transport antigens to follicles; thus, initiating immune responses, reviewed in (100). Previous studies indicated BoHV-1 infects circulating CD4⁺ T cells during acute infection of calves suggesting CD4⁺ T cells in PT may be a site for BoHV-1 latency (38). However, a preliminary study suggests CD4⁺ and CD8⁺ T cells in PT of latently infected calves contained no viral DNA when compared to total PT suggesting infected T cells do not survive and are not the reservoir for latency (Harrison and Jones, unpublished data). Since acutely infected CD4⁺ T cells frequently undergo apoptosis (38), this finding was somewhat expected. Antibodies that specifically identify single populations of other bovine lymphocytes, epithelial cells, and Micro-fold cells will be used to identify specific cell population(s) that harbor viral DNA.

The LR gene is abundantly expressed in TG neurons during latency (49, 101, 102): however, this was not the case in PT as judged by RNA-seq and RT-PCR studies (18). This was not surprising because the LR promoter is regulated by neuronal specific factors (103–105) that are apparently absent or not abundantly expressed in PT. LR gene products have multiple functions designed to maintain latency: for example, impairing apoptosis, productive infection and bICP0 expression, Notch-mediated transcription, and stress-induced transactivation of IEtu1 promoter activity (24, 26, 29, 106–110). Hence, the absence of LR gene products in PT will not influence viral gene expression and the number of cells that successfully reactivate from latency following DEX treatment. The family of transcripts with the repetitive element expressed in PT of latently infected calves are not likely to express a protein and their putative role in the latency-reactivation cycle is currently unknown.

In conclusion, these studies provide compelling evidence that reactivation from latency occurs rapidly in PT following DEX treatment. Furthermore, reactivation from latency in PT may be more efficient than TG with respect to virus transmission following stressful stimuli. This statement is supported by the knowledge that reactivation from latency via TG neurons or other neurons requires viral particles to travel anterogradely from neurons toward nerve terminals and then reinfect dermis or other innervated cell types prior to virus transmission to another host. However, virus transmission during reactivation from latency in PT can occur directly because PT is at the base of the oral cavity.

MATERIALS AND METHODS

Infection of calves and DEX treatment to induce reactivation from latency.

For these studies, weaned male Holstein calves (~440 lbs.) were acclimated for 7 days prior to infection with BoHV-1 (Cooper strain, GenBank Z48053.1) as previously described (18, 38, 40, 111, 112). In brief, BoHV-1-free calves were randomly assigned and housed in isolation rooms (2-3 calves in each room) in a BSL2 facility for large animals. During the acclimation period, calves were treated with the antibiotic Liquamycin (oxytetracycline; Zoetis) using 4.5 mL for each 100 lbs. of the calf every 4 days. Calves were anesthetized with Xylazine (approximately 50 mg/50 kg body weight; Bayer Corp.). Calves were inoculated with 1 mL of a solution containing 1x10⁷ PFU/mL of BoHV-1 in each nostril and eye, without scarification, for a total of 4x10⁷ PFU per animal. Instillation of the virus in the ocular cavity was performed by adding 0.1 to 0.2 mL of virus and then closing the eyelid and rubbing gently to facilitate infection. Infectious virus in ocular and nasal swabs were monitored after infection to ensure acute infection occurred as previously observed (40, 111). Calves were treated with Liquamycin until 8 days after infection to prevent bacterial pneumonia. TG and PT were collected during latency (≥ 60 days after infection). Latently infected calves were injected intravenously (jugular vein) with 100 mg water-soluble DEX (Sigma) suspended in Minimal Essential Media (MEM) that did not contain antibiotics or fetal bovine serum to initiate reactivation from latency (18). Calves were then loaded onto a cattle trailer and transported to the University of Nebraska-Lincoln Veterinary Diagnostic Laboratory (approximately ¼ of a mile from the BSL-2 large animal facility to the Diagnostic Laboratory). Loading calves onto the trailer, transporting the calves, and unloading them took approximately 10 min. For the 30 min treatment with DEX calves, they were transported to the Diagnostic Lab, placed in a pen, and then injected with DEX. Prior to euthanasia by electrocution, calves were heavily sedated with Xylazine. It is unlikely that the stress of transporting calves ¼ of a mile to the diagnostic center initiated lytic cycle viral gene expression. This observation is supported by the finding that latently infected calves are not expressing detectable levels of lytic cycle viral genes in PT (Fig. 1, 4, and 5) using the same methods to transport these latently infected calves (no DEX treatment) from the large animal facility to the Diagnostic Lab. Other studies also revealed that lytic cycle viral genes are not detected in TG of latently infected calves unless DEX is given to these calves and transported as described above (22, 51, 61, 93, 94, 113). After decapitation, TG and PT were collected, and samples from each were formalin fixed and then paraffin embedded. The remainder of TG or pharyngeal tonsil were minced into small pieces, placed into 50 mL conical tubes, tubes placed in a dry ice ethanol bath, and stored at −80°C. Following decapitation, it took approximately 10 min to process tissues and submerge tubes containing TG or pharyngeal tonsil in a dry ice ethanol bath. All protocols were approved by the Institutional Animal Care and Use Committee of the University of Nebraska.

Preparation of total RNA from PT and RNA-seq.

PT were collected from 3 calves per time point in the study: latency, 30 min, 90 min, or 3 h after DEX treatment. Total RNA was prepared from approximately 2 g of tissue using TRIzol (Life Technologies) according to the manufacturer's instructions and as previously described (22). RNA integrity and concentration were quantified with a 5200 Fragment Analyzer System (Agilent Technologies). An Illumina TruSeq stranded mRNA library preparation kit was used to construct cDNA libraries, which were sequenced as 75-bp paired-end reads on an Illumina NextSeq 500 (Illumina) to the depth of approximately 100 million reads per library.

Read alignment of the RNA-seq data was carried out as previously described (22). In brief, raw paired-end sequence reads in individual fastq files were trimmed to remove adapter sequences and low-quality bases using Trimmomatic software (version 0.35) (114). The remaining reads were mapped to the UMD 3.1 genome assembly with Tophat2 (version 2.1.1) (115) using the National Center for Biotechnology Information (NCBI) Bos taurus and BoHV-1 reference annotation to guide the alignment. The default parameters for Tophat2 were used.

Cuffdiff software (version 2.2.1) (116) was employed to process the aligned sequence read files and test for differential gene expression. Transcript abundance was estimated for each transcript in the NCBI Bos taurus reference annotation (UMD3.1) or BoHV-1 reference genome for each sample. A gene was defined as being expressed provided it had an average fragment per kilobase of transcript per million mapped reads (FPKM) of 1.0 in at least 1 of the 2 groups in the comparison. Unexpressed genes were removed from further downstream analysis. Differential gene expression was detected by comparing the log (base 2) ratios of the FPKM values for every gene and transcript. The resulting P-values were corrected for multiple testing using Benjamini-Hochberg correction (117). P values were considered statistically significant at a Benjamini-Hochberg adjusted P value of ≤ 0.05.

Immunohistochemistry and immunofluorescence.

PT from latently infected or DEX treated calves were harvested and directly placed in formalin. Slide preparation and IHC was performed as previously described (22, 23, 118). Briefly, 4-5μm sections of paraffin embedded PT were cut and mounted onto glass slides. Slides were incubated at 65°C for ~20 min before deparaffinization in Xylene and serial rehydration using decreasing concentrations of Ethanol. Endogenous peroxidases were blocked by incubating slides in 0.045% H₂O₂ for 20 min. Antigen retrieval was performed using Proteinase K (Agilent Dako) and slides were blocked using Animal Free Blocking Solution (AFBS; Cell Signaling Technologies) for 45 min at room temperature (RT). Avidin and Biotin were blocked using Vector Labs Avidin/Biotin blocking kit as per manufacturer’s instructions (Vector Labs). Slides were incubated overnight at 4°C with the bICP0, VP16, or bICP4 peptide specific antibody in AFBS. These are peptide specific antibodies that specifically recognize these proteins and they were previously described (51). The following day, Vectastain ABC HRP Kit was used according to manufacturer’s instructions (Vector Labs). Biotinylated secondary antibody was prepared in AFBS, and slides were incubated for 30 min at RT. Slides were then developed with NovaRed substrate (Vector Labs) and lightly counterstained with Mayer’s hematoxylin (Sigma-Aldrich). Slides were imaged using an Olympus BX43 microscope and captured using CellSense software. The bICP0 or bICP4+ PT cells were examined in a blinded fashion.

For IF studies, slides were dual stained with bICP4 overnight at 4°C using a 1:200 dilution after completion of TUNEL assay as described below. The following day slides were washed and stained with a secondary Alexa 488 (Invitrogen, Catalog # A32731) diluted 5 μg/mL for 1 h at room temperature. All images were captured as described above using the 40 X objective. Brightfield images were obtained with autoexposure times ~0.01 s; fluorescent images were obtained with equal manual exposure of 0.1 s.

Detection of apoptosis in fixed tissue.

The TUNEL assay was performed to detect cells undergoing programmed cell death in PT tissue sections as described previously (17, 38). For these studies, the DeadEnd Colorimetric TUNEL System was used (Promega, catalogue # G7360). Positive control slides were included by treating with OPTIZYME DNase I (Fisher) at 5U/mL for 10 min at RT prior to TdT staining (data not shown).

Data availability.

The RNA-seq data was deposited into the NCBI GEO repository as series record GSE216685.