Evidence of Hepatitis E Virus Breaking Through the Blood–brain Barrier and Replicating in the Central Nervous System

R. Shi; M. H. Soomro; R. She; Y. Yang; T. Wang; Q. Wu; H. Li; W. Hao


J Viral Hepat. 2016;23(11):930-939. 

In This Article


Neurological manifestations involved with HEV infection were reported by Sood[25] firstly, regarding Guillain–Barre' syndrome associated with acute hepatitis E. It has drawn more attentions towards cases of HEV infection complicated with neurological disorders, and the number of similar reports has been increasing afterwards. Patients presenting neurological signs are suspected with hepatotropic virus infection due to the elevated liver function index levels,[26] and HEV infection was diagnosed via serum anti-HEV IgM and HEV RNA detection, although the pathogenesis of HEV neuroinvasion remains elusive. It is speculated that direct affection of HEV on the nervous tissues played a significant role given the following reasons: similar cases occurred with geographical isolation; CSF samples collected from some patients were detected with HEV RNA positive; neurological signs terminated after administration of ribavirin, interferon-α and specific immunoglobulin to the patients.[12,14,27]

Peripheral nervous system diseases such as GBS, neuralgic amyotrophy and Bell's palsy are mostly involved with acute HEV infection, while central nervous system diseases such as meningitis, encephalitis, ataxia and acute transverse myelitis can occur in both acute and chronic hepatitis E patients.[12] Serum and CSF samples collected from the patients developing central nervous system disorders were all detected as HEV RNA positive, which suggested that HEV replication may occur in this compartment.[12,14] In our research, HEV RNA was detected in brain and spinal cord from 7 dpi to 28 dpi, with concurrence of liver function indexes elevation. The viral load in spinal cord was higher than that in the brain, and it is thought provoking that for some gerbils detected with brain HEV negative, HEV RNA was detectable in the spinal cord samples, but not vice versa. Moreover, for gerbils at 7 dpi to 21 dpi, serum HEV antigen positiveness was detected if the brain and spinal cord were HEV RNA positive, while this was not supported by the results at 28 dpi. The present study indicated that HEV is neuroinvasive and spinal cord tissue may be better reservoir than the brain. HEV invading into nervous tissues may originate from viremia, and the inconsistency of serum and brain, spinal cord HEV antigen test result verified that HEV replication took place in the nervous sections.

HEV ORF2 encodes the capsid protein which is believed to involve in the assembly of the HEV particle and its interaction with host cells. Immunohistochemical staining targeting HEV ORF2 protein showed the virions distribution in the central nervous tissues with positive staining interspersed in the perivascular area involving endothelium and astrocyte endfeet, both being comprised of BBB.[28] Choroid plexus epithelium and ependymal epithelium in brain and spinal central canal associated with CSF produce were detected with intensive HEV ORF2-positive signals. The HE staining and TEM observation revealed massive lymphocytes infiltrated, especially in the choroid plexus and perivascular area, and formed 'perivascular cuff', which is one of the typical symptoms of viral encephalitis.[29] Moreover, the cytoplasm of some neurons in brain and spinal and Purkinje cells in cerebellum also presented to be ORF2 antigen positive; hence, degeneration, necrosis and neuronophagia of these cells could be attributed to the HEV attack. These findings suggest that haematogenous HEV traversed the BBB and invaded in the brain and spinal cord, especially CSF producing structures.

It's well known that brain endothelial cells and astrocytes are the two main components of the BBB, of which the tight junctions play a critical role in preserving the integrity of the BBB. There are three approaches suggested for the BBB traverse by viruses: transcytosis, paracellular entry and 'Trojan horse' mechanism, of which the first and the last pathways may not incur blood–brain barrier lesions.[30–33] According to our TEM observations, compromised BBB structure was evidently determined by the tight junction defects, degenerative endothelial cells and disorganized basal membrane layers, in the HEV-positive brain and spinal cord tissues. These findings implied that HEV broke through the BBB by sabotaging the relative junctional complex and the endothelial cell structure; then, the increasing permeability of the BBB facilitated the passive entry of the virions paracellularly. This is in accord with Deroux et al.'s report,[23] in which moderate BBB damage was determined with increased CSF protein and glucose levels in a patient with concurrent of HEV infection and neurological manifestation. In addition, ZO-1 is described as one of the key tight junction proteins presented between the cerebral endothelium, and the astrocyte endfeet, which is specifically marked by GFAP, also take an important part in the barrier mechanism.[34,35] In our study, it is found that the integrity of BBB is damaged in view of the reduced ZO-1 expression, as a result of HEV infection. It then caused the upregulation of GFAP expression, which might function as the compensatory adjustment. The IHC analysis of ZO-1 and GFAP in our study has provided evidence in support of the passive entry of HEV into the CNS.

Besides the paracellular pathway, for some neurotropic viruses like Japanese encephalitis virus (JEV) and West Nile virus (WNV), transcytosis across both cerebral endothelial cells and pericytes via endocytic vesicles was achievable.[31,32] To determine the transcytosis of HEV invading CNS parenchyma, further study regarding the precise localization of HEV around perivascular region should be conducted by means of immune-electron microscope and fluorescent co-location techniques. Moreover, some immune cells are likely to assist some virus like hepatitis C virus into the CNS via 'Trojan horse' mechanism.[36] In the HEV-infected rabbit experiment we conducted, HEV suspension was injected in the ligated section of rabbit sacculus rotundus and appendix. HEV ORF2 antigen deposition was found in some of the mononuclear cells disseminated in the granular layer of cerebellum which was excised merely 3 hours later (unpublished). This coincides with the IHC analysis of HEV-positive gerbils in this study. We speculate that the HEV ORF2-positive mononuclear cells might be a differentiation of immunocytes originated from blood, which is similar with the microglia, and it serves as a 'Trojan horse' for the HEV invading the CNS. Recent discovery of HEV replicating in the immunologic tissues like tonsil, spleen and lymph nodes also supports this hypothesis.[7,8]

All of the HEV strains detected from cases exhibiting neurologic signs in developed countries were HEV genotype 3, which together with HEV genotype 4 is believed to be locally acquired porcine zoonotic.[17,37] Swine HEV strains are antigenically and genetically related to human HEV strains, and transmission between human and pigs of swine HEV isolates is widely reported.[38] Most of the autochthonous hepatitis E virus infection in developed countries like UK, France and USA was caused by HEV genotype 3, which has been broadly detected in domestic pigs.[11,12,14] However, swine genotype 4 is found mainly in Asian countries like Japan and China.[39,40] In our research, potential invasiveness of genotype 4 swine HEV to the neurological system was also identified.