Findings

Neurovirulent Venezuelan equine encephalitis virus (VEEV) is a member of the genus Alphavirus, in the family Togaviridae. VEEV causes lethal encephalitis in equines and occasionally infect humans [1, 2]. In our earlier studies, we have reported upregulation of several integrins (Itg) and integrin binding molecule genes such as nischarin (Nisch) and integrin alpha V (ItgaV) as well as genes that are implicated in the alteration of the blood brain barrier (BBB) such as MCP-1 [3] in the brains of VEEV infected mice [4, 5]. Other studies have also shown that adhesion molecules expressed on the surface of microvascular endothelial cells of the BBB play an important role in viral encephalitis [6, 7]. VEEV replicon particles have also been recently reported to disrupt BBB [8]. Though, VEEV has been known to enter the central nervous system (CNS) through the olfactory tract, these studies indicate that adhesion molecules expressed on the BBB may also play role in VEEV pathogenesis [2]. Therefore, in this study, we evaluated the expression of extracellular matrix (ECM) and adhesion molecules in VEEV infected CD-1 mice brain. Several ECM and adhesion molecules genes, such as integrins (ItgαX, Itg2, 3, and 7), cadherin (Cdh) 1 and 2, intracellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule (VCAM-1) were found to be upregulated in the brains of VEEV infected CD-1 mice (Table 1). Immunohistochemistry analysis showed ICAM-1 expression and its co-localization with inflammation, fibrinogen leakage and VEEV antigen in and around the brain microvessels (Figure 1). Pathway analysis of the modulated ECM and adhesion molecules genes using DAVID software [9, 10] indicated their involvement in leukocyte migration at BBB (Additional file 1 Figure S1). Of these several genes, ICAM-1 has been implicated in the pathogenesis of various other neurotropic viruses such as West Nile virus, Semliki Forest virus, Theiler's murine encephalomyelitis virus and lymphocyte choriomeningitis virus [11–14]. To test if ICAM-1 plays any role in VEEV pathogenesis, ICAM-1 knockout (IKO) mice (B6.129S4-Icam1tm1Jcgr/J, stock No. 002867, Jackson Laboratories, Bar Harbor, Maine, USA) were infected with 1000 pfu of VEEV in two separate studies. In both these studies, VEEV infected wild type (WT; C57BL/6J, stock No. 000664, Jackson Laboratories, Bar Harbor, Maine, USA) mice became sick earlier than the ICAM-1 knock-out (IKO) mice. IKO mice showed delayed appearance of the clinical symptoms such as shivering, excitability, and hind limb paralysis with a total loss of mobility. General health and appearance of the WT mice evaluated as ruffled fur, hunched back posture and lethargy in the early stages of the disease was severe than similarly infected IKO mice. IKO mice were more responsive to touch, less lethargic and were eating better than the WT mice. Though a 20% reduction in mortality was observed in both the studies (Figure 2), there was no difference in the mean survival time of the IKO mice that succumbed to VEEV infection as compared to the WT mice. Histopathological analysis of brain showed less severe single cell necrosis and perivascular cuffing in the IKO mice brain than the WT mice at 96 hr post infection (pi) (Figure 3). However, no significant difference in VEEV antigen was noticed (data not shown) in the brains of IKO and WT mice. The inflammatory cytokine specific focused microarray performed on the brain RNA samples of VEEV infected WT and IKO mice at 96 hr pi (Table 2) showed inflammatory gene expression differences in the brain of IKO mice. Basal level differences in the cytokine expression of uninfected IKO and WT brain are given in additional file 2 Table S1. Microarray analysis (Table 2) showed a complex inflammatory response to VEEV infection in the brain of IKO mice which can be classified based on pro- and anti-inflammatory response and may explain some of the differences seen early in the infection of IKO and WT mice. Several pro-inflammatory cytokines/ligands such as Ccl24, Cxcl11 and Cxcl13 were down regulated in IKO mice brain as compared to WT mice. Ccr2, a chemokine receptor that is involved in the migration of peripheral blood mononuclear cells into the brain [15] was also down regulated in the brains of VEEV infected IKO mice. Anti-inflammatory genes IL10, Ccl22 and IL22 were also down regulated in the brains of VEEV infected IKO mice. These cytokines and chemokines have been implicated in the regulation and suppression of inflammation in the tissues [16–19]. Other pro-inflammatory genes such as Ccr1 and Xcl1 that have been implicated in the recruitment of mononuclear phagocytes and T cells respectively into the brain [20–23] were either upregulated or induced in the VEEV infected IKO mice brain. Modulation of these anti-inflammatory and pro-inflammatory genes may contribute to the final outcome of inflammation observed in the brain of VEEV infected IKO mice. Gene expression evaluation at early and later time points than 96 hr pi may shed more light on the inflammatory gene expression kinetics in the brain of IKO mice. To further evaluate if inflammation is crucial, animals were treated with a known anti-inflammatory drug naproxen (40 mg/kg, once a day). Similar to the VEEV infected IKO mice, naproxen treated VEEV infected CD-1 mice (n = 10) early in the infection (0-5 days) showed lesser degree of ruffled fur, lethargy, hunched back, and delayed appearance of shivering and paralysis as compared to the untreated mice (n = 10), thereafter, these mice quickly became sick and there was no difference in the mean survival time of naproxen treated and untreated VEEV infected mice. A marked decrease but not total ablation in the vascular inflammation was observed in the naproxen treated VEEV-infected mice at 96 and 120 hr pi (n = 3 each group) (Figure 4). Viral load in the brain was evaluated by RT-PCR at 48, 72, 96 and 120 hr pi (n = 5 each group). There was no significant difference in the viral load in the brain of the naproxen treated and untreated VEEV infected mice groups. However, when taken as individual sets of mice, more viral load was observed in the brain of some of the naproxen treated mice as compared to the untreated ones (Figure 5). Treatment with ribavirin (80 mg/kg, once a day), a known anti-viral drug, alone or in combination with naproxen also did not affect the mortality or the viral load in VEEV infected mice. These results show a complex role of inflammation in VEEV pathology, initial better health of naproxen treated mice may be due to reduced inflammation in the brain. However, increase in brain viral load in naproxen treated animals may be due to unchecked viral replication due to reduced inflammation.

Table 1 Mouse extracellular matrix and adhesion molecules gene(s) differentially expressed by two fold in VEEV infected CD 1 mouse brain.
Full size table
Figure 1
figure 1

ICAM-1 expression corresponds with perivascular cuffing, fibrinogen leakage, and the presence of VEEV in the brains of VEEV infected mice. (a) H&E staining of the brain of VEEV infected CD-1 mice exhibited prominent perivascular cuffs throughout the brain. (b) Endothelial cells are hypertrophied and exhibit positive immunoreactivity for ICAM-1. (c) Immunohistochemistry for fibrinogen demonstrated localized fibrinogen leakage around hypertrophied endothelial cells, indicating disruption of the BBB. (d) Extensive VEEV specific staining revealed numerous cells infected with VEEV around the affected vessels. (e) Perivascular cuffs were composed of moderate numbers of mononuclear cells, primarily lymphocytes and fewer monocytes which had transmigrated the endothelium and multifocally extended into the neuropil. (f) Neurons were primarily infected with VEEV; however, VEEV antigen was also localized in glia and fewer lymphocytes. (g) Vacuolation of the neuropil was observed around the affected vessels.

Full size image
Figure 2
figure 2

Survival study in IKO mice. Representative of two separate survival studies done in IKO mice upon VEEV infection is shown here. Mice were observed twice a day for fourteen days pi for signs of clinical disease. There was a 20% reduction in mortality in IKO mice over their WT controls in both the studies.

Full size image
Figure 3
figure 3

Inflammation in the brain of VEEV infected IKO and WT mice. (a) IKO 96 hr pi cerebrum, subventricular zone, H&E, 400X: There were low numbers of necrotic neurons, characterized by small amounts of karyorrhexis (arrows), when compared to the same region of the WT mice (b). (b) WT 96n hr pi, cerebrum, subventricular zone, H&E, 400X: There were high numbers of necrotic neurons, characterized by large amounts of karyorrhexis (arrows), when compared to the same region of IKO mice (a). (c) IKO 96 hr pi, brainstem, H&E, 400X: There was little to no endothelial swelling and no perivascular cuffing (arrows) compared to the WT mice (d). (d) WT 96 hr pi, brainstem, H&E, 400X: There was moderate endothelial swelling and mild perivascular cuffing (thick arrows) and rare necrotic cells (thin arrow) compared to the IKO mice (c). (e) IKO 96 hr pi, brain, thalamus, H&E, 400X: There was little to no endothelial swelling and no perivascular cuffing (arrows) compared to the WT mice (f). (f) WT 96 hr pi, brain, thalamus, H&E, 400X: There was moderate endothelial swelling and mild perivascular cuffing (thick arrow) and rare necrotic cells (thin arrows) compared to the IKO mice (e).

Full size image
Table 2 Differential gene expression in brains of VEEV infected ICAM-1 WT mice (n = 3) as compared to IKO mice (n = 3) at 96 hr pi.
Full size table
Figure 4
figure 4

Inflammation was reduced in VEEV infected mice treated with naproxen or naproxen plus ribavirin. Animals were treated with naproxen (40 mg/kg/once a day) only or naproxen plus ribavirin (80 mg/kg/once a day) at the time of infection (1000 pfu of VEEV inoculated in left rear foot pad). Animals were monitored twice a day and clinical symptoms of disease were monitored for two weeks.

Full size image
Figure 5
figure 5

Viral load in the brain of mice treated with naproxen and naproxen plus rivavirin. Animals were treated with naproxen and naproxen plus ribavirin as described in the text. Mice were sacrificed at each given time points (n = 5 each group) and brain tissues were evaluated for viral load by RT-PCR. Given is a representative picture of one of the set of the mice. (1) drug treated 120 hr pi; (2) untreated 120 hr pi; (3) drug treated 96 hr pi; (4) untreated 96 hr pi; (5) drug treated 72 hr pi; (6) untreated 72 hr pi; (7) drug treated 48 hr pi; and (8) untreated 48 hr pi.

Full size image

To our knowledge, this is the first study to evaluate the expression of ECM and adhesion molecules in brain during VEEV infection and to evaluate the inflammatory response in IKO mice during VEEV infection. The results show that adhesion molecules such as ICAM-1 may play an important role in VEEV disease pathology. These findings improve our present understanding of VEEV induced inflammation in the mouse brain and its implication in VEEV disease.