Introduction to the Zika virus:
The Zika virus (ZIKV) was discovered in 1947 in Uganda and named after the Zika forest. ZIKV is classified as a flavivirus, denoting a family of viruses that are arthropod-borne, enveloped, and contain RNA as their genomic material. The Zika virus is transmitted by the Aedes mosquito, commonly found along the equatorial belt stretching from Africa to Asia (Gorshkov et al, 2018). ZIKV replicates in the mosquitos’ gut epithelial cells and salivary glands and remains virulent within the salivary gland for over a week. When the mosquito bites, the virus is injected into human skin; infecting skin fibroblasts, keratinocytes, and Langerhans cells. The virus then travels to the lymph nodes and into the bloodstream to establish primary infection (Hamel et al., 2015, Olagnier et al., 2016, Kim et al., 2018). The first human cases were reported in 1952 and the disease has continued to spread to Southeast Asia, African tropics, and the Pacific islands. By 2015, the virus had reached Mexico, South America, Central America, and the Caribbean. Within the United States, the vast majority of cases are travel-related, and the remaining cases are either locally or sexually acquired (Gorshkov et al., 2018).
Once infected with ZIKV, most adults experience no symptoms. If symptoms do develop, patients suffer from sudden onset of fever, myalgia, joint pain, conjunctivitis (pink eye), headache, and skin (maculo-papular) rash. These symptoms only last for about a week, but the virus remains in the blood for much longer (Hamel et al., 2015, Peterson et al., 2016). Although the primary infection is not fatal, there are several serious neurologic complications that can occur because of a compromised immune system. A single mutation in the ZIKV genome has made the virus particularly dangerous for pregnant women during the first trimester. An infected mother can pass the viral RNA across the placental barrier and/or infect the baby at time of birth. The ZIKV structural protein, precursor membrane (prM), is important for maturation and assembly of the virion during replication. A single amino acid substitution from serine to asparagine in prM of ZIKV has been linked to fetal microcephaly in mouse and human cells (Yuan et al., 2017). Microcephaly is a neurodevelopmental disorder wherein the infant is born with a reduced head size, and therefore impaired intellectual development. Common effects of impaired brain maturation as caused by microcephaly include motor deficits and convulsions (CDC, 2019). Another neurologic consequence of ZIKV infection is Guillain-Barre’ syndrome. Guillain-Barre syndrome is an autoimmune disease affecting the peripheral nervous system by attacking peripheral nerves and their myelin insulation. What starts as weakness and tingling in the feet can spread to upper extremities, cause paralysis of eye muscles (ophthalmoplegia), ataxia, and in severe cases, paralysis of all extremities (Dimachkie and Barohn, 2014). All reported people with Guillain-Barre syndrome during a ZIKV outbreak were positive for ZIKV antibodies (Krauer et al., 2017).
Much of the current, serious pathophysiological and neurologic complications arising from ZIKV such as Guillain-Barre syndrome are relatively rare and were not associated with ZIKV until after 2013. It was not clear what may have changed in the virus to suddenly produce serious diseases in some patients, while presenting with self-resolving flu-like symptoms in most people. A recent study by Carbaugh et al., looked at the possibility that viral genetic changes could contribute to the new clinical pathogeneses. The study identified an N-linked glycosylation site on the envelope protein that was absent in the historical ZIKV isolates. The glycosylated virus was highly pathogenic in mice compared to non-glycosylated viruses. The non-glycosylated viruses presented lower viral serum loads in the blood and brain with subcutaneous inoculation but remained virulent when injected intracranially. This suggests that glycosylation of the envelope protein may be advantageous in the periphery but not in the central nervous system, suggesting the presence of differential genetic changes that distinguish virulence in the periphery versus the brain. This study suggests that envelope protein glycosylation may contribute to ZIKV pathogenesis by facilitating better attachment and infection of leukocytes (Carbaugh et al., 2019).
Human-to-human ZIKV transmission is also possible. There have been multiple reports of ZIKV cases in Brazil caused by contaminated blood transfusions, and sexual contact. Another prevalent method of human-to-human transmission could be through eye infection. ZIKV has been found to infect the eye, causing severe retinitis associated with blindness in newborns and adults. ZIKV RNA has been detected in the conjunctival fluid of the eye, raising the possibility that tears could be a possible method of transmission. ZIKV can cause a wide array of ocular abnormalities, including optic neuritis, chorioretinal atrophy, and blindness in newborns. Adults mostly suffer from conjunctivitis and uveitis (Ventura and Ventura, 2018). There have been several case studies presenting ocular abnormalities caused by ZIKV and research experiments in animals to better understand the molecular mechanisms of eye disease caused by ZIKV. This is an intriguing area of research because the primary target of ZIKV is not the eye, but the virus still reaches the eye somehow and virions persist there causing severe visual deficits even after clinical symptoms may have subsided (Ventura and Ventura, 2018).
Ocular immune privilege:
How exactly ZIKV enters the eye and causes infection is still incompletely understood. One place to start would be to understand the mechanisms of ocular immune privilege, so research can be focused on how viruses such as ZIKV can infiltrate to cause disease. Immune privilege is mediated by systemic and local functions, protecting the eye from daily inflammatory assaults. However, the specific mechanisms of how the eye accomplishes this is still being discovered. Immune privilege in the eye was first noticed by Sir Peter Edwards in 1941 when he placed a foreign tissue graft in the anterior chamber of the eye and did not observe any signs of rejection. There are several mechanisms that work together to mediate ocular immune privilege. The first line of defense consists of physical barriers, namely the blood-retinal-barrier (BRB). The BRB, much like the blood-brain-barrier (BBB) prevents free entry and exit of large, potentially pathogenic molecules. However, follow-up studies question the concept of protection of the eye from immune cells on the basis of anterior chamber-associated immune deviation (ACAID). ACAID describes an immune response to antigens injected in the anterior chamber of the eye. Antigen presenting cells from the eye migrate to the spleen and interact with killer T-cells and B-cells. This overall elicits a systemic immune response by inducing activity of CD4+ and CD8+ regulatory T-cells. Cellular fragments and proteins were shown to pass from the anterior chamber into the blood through a porous membrane known as the trabecular meshwork. Some studies present this phenomenon as negative evidence of antigen sequestration in the eye, but the ACAID response results from puncturing the eye with a needle or causing some disruption of ocular integrity. This suggests that ACAID may be more representative of a trauma-induced response rather than a mechanism of tolerance (Zhou and Caspi, 2010).
Another mechanism by which the eye maintains immune privilege is through a chemical microenvironment. The ocular microenvironment consists of soluble and cell-bound immunosuppressive factors. Some of these factors include, transforming growth factor-beta (TGF-β), neuropeptides such as alpha-melanocyte-stimulating hormone (α-MSH), among others. Ocular cells release these factors to suppress immune cell activity via soluble secretions and/or via contact-dependent mechanisms. The pigmented epithelia of the retina have an interesting feature—they not only inhibit T-cells but can convert them into regulatory T-cells (Zhou and Caspi, 2010).
Among the methods described of ocular immune privilege, the most convincing argument is the corneal transplantation. Corneal grafts are up to 90% successful without tissue matching or immunosuppressive drugs. Given all the evidence declaring the eye as an immune privileged site, how exactly does a virus get into it and cause severe abnormalities? The very “privilege” that may protect the eye in cases of daily assault or transplantation, leaves the eye vulnerable to autoimmune attacks. Ocular immune privilege may impede peripheral tolerance to antigens barricaded behind the blood-retinal barrier (BRB) (Zhou and Caspi, 2010).
The BRB is an extension of the blood-brain-barrier (BBB). The BRB consists of two segments: an inner barrier formed by retinal endothelial cells, and an outer layer formed by retinal pigmental epithelium, choriocapillaris, and Bruch’s membrane, both consisting of tight junctions. The outer BRB shields the neuroretina from pathogens including viruses. Disruption of the retinal pigmental epithelium could pave way for virus entry from choroidal capillaries into the tissues of the eye. ZIKV is permissive to all the cell types of the different layers of the eye except for photoreceptors (Singh and Kumar, 2018).
Modulation of retinal adaptive and cellular immunity play a critical role in ZIKV infection and progression. Retinal epithelial cells are the first line of defense against pathogens and is involved in both innate and adaptive immunity. These retinal cells are the primary targets for viruses such as cytomegalovirus and ZIKV, and produce chemokines, growth factors, and cytokines. Once infected with a virus, peripheral immune cells enter the eye to cause inflammation. Exacerbated ocular pathology due to inflammation is caused by increased expression of inflammatory genes such as TNFα and IL1B, as well as elevations in perforin, granzyme B, and interferon genes. ZIKV infection of retinal cells or Muller cells causes an increase in innate immune response and increased gene expression of interferons, causing chorioretinal atrophy. In addition, a growing body of research has implicated glial cells in evoking innate immune responses in addition to structural support of neurons (Singh and Kumar, 2018). A study by Manageeswaram et al., 2018 presented that subcutaneous infection can present with symptomatic posterior uveitis in mice. There is preferential infection of the cornea and retina with increased levels of chemokine expression, and infiltration of natural killer cells and neutrophils. The study showed that cytotoxic T-cells remained in the eyes of mice 30 days later even after infection subsides. Over all, this study provided a useful model for teasing apart the mechanism behind BRB breach during ZIKV infection and how various ocular tissues change to cause lesions and atrophy. Another strength of this study is that it provides a physiologically relevant mode of infection by breaking of skin (bite) and subsequent eye inflammation as seen in humans (Manangeeswaran et al., 2018). The specific mechanisms of how retinal pigment cells participate in regulating the blood-retinal barrier (BRB) and how they regulate innate immunity continues to be explored.
As of 2018, there are two proposed models for ZIKV entry and infection of the eye. The first is that ZIKV causes cell death in the outer layer of the BRB forming a point of entry into the eye to infect the inner barrier, resulting in inflammation and vision loss. ZIKV could also enter through the retinal arteries of the inner BRB leading to infection of retinal endothelial cells before entering the outer BRB through the choroid capillaries (Singh and Kumar, 2018). In order to address and alleviate symptoms associated with ocular virus infections, animal models have proved immensely valuable for research on therapeutic strategies and disease management.
Animal research describing ZIKV mechanisms of action underlying ocular abnormalities:
Early attempts to use animal models for mimicking ZIKV pathogenesis discovered that some strains of wild-type mice are resistant to flaviviruses. Therefore, most of the studies used immunocompromised, interferon knock-out mice (Ifnar-/-) for efficient and reliable viral infection and proliferation to simulate human pathogenesis/symptoms (Lazaer et al., 2016). Miner et al., reported that subcutaneous injection of ZIKV in IFNAR-/- caused panuveitis and viral RNA shedding in tears, but no detection of active virus in ocular tissues as reported in humans. The researchers did not find any abnormalities in the eyes of congenitally infected fetuses either, in contrast to several reports of the presence of histological abnormalities in human babies (Miner et al., 2016). Another method of inoculation involves direct injection into the tissue of interest. Singh et al., say that they have developed a method of injecting ZIKV intravitreally into the eyes to mimic human disease conditions. Although this may result in robust histological abnormalities like that seen in humans (retinal pigment epithelium atrophy, pigment mottling) the method of entry does not mimic human conditions (Singh et al., 2017). A major part of understanding ZIKV induced eye disease has to do with primary method of entry and infection so proper precautions are taken and effective therapies discovered. A successful mouse model that is more in line with a plausible mode of infection showed viral persistence in ocular tissues when injected subcutaneously with ZIKV (Manangeeswaran et al., 2018). Yet another mouse model has shown ZIKV to infect several, previously unexplored cell types of the eye such as the cornea, optic nerve, iris, retina, and viral RNA detection in tears. Specifically, these studies propose two modes of viral spread: axonal or hematogenic. The virus could travel retrograde via the optic nerve to infect the eye, or through fetal circulation from the placenta to infect the retina and its endothelial and pigment cells. The hematogenous route could lead to a host of abnormalities including loss of the ganglion cell layer, chorioretinal abnormalities and foveal maldevelopment (Miner et al., 2016, Singh et al., 2017). A combination of these models, taken for their strengths and weaknesses are useful for better understanding ZIKV pathogenesis in ocular tissue.
Because wild-type mice are not naturally susceptible to ZIKV, potentially more relevant research methods include in-vitro models using human cells or established cell lines. Singh et al., reported that retinal cell types of the BRB become susceptible to ZIKV infection by cell death caused by activation of Caspase 3. Interestingly, Aleman et al., and Zhao et al., have proposed that Muller cells may be the primary target of ZIKV. Muller cells are retinal glial cells, and infection of these cells cause increased release of pro-inflammatory cytokines and decreased neurotropic functions. Infection of retinal pigment epithelium has also shown disruption of cell-cell adhesion and barrier properties. Overall, the mechanisms behind how the BRB is breached is still under development (Singh et al., 2017). The family of TAM receptors (TYRO3, AXL, and MER) have been shown to be upregulated in endothelial and epithelial retinal cells during ZIKV infection, but there is continued debate on the specific role of these tyrosine kinase receptors (Singh and Kumar, 2018).
A comprehensive research study conducted by Miner at al., 2016 looked at ZIKV-induced pan-uveitis and virus shedding in tears of mice. It evaluated ocular abnormalities caused by ZIKV and characterization of entry receptors. It is known that ZIKV causes conjunctivitis in about 15% of patients and uveitis occurs several weeks after primary infection, suggesting the persistence and replication of virus within the eyes even after initial symptoms subside (Furtado et al., 2016). For research in animal models however, ZIKV does not replicate in wild-type mice probably because ZIKV can antagonize human but not mouse STAT2—a molecule activated by type I and III interferon receptors (Grant et al., 2016). Knockout mice (Ifnar-/-) were developed so that these animals can be used to study successful ZIKV infection and pathogenesis—mice lacking the Ifnar gene developed neuroinvasive disease after ZIKV infection and died much sooner than animals with the Ifnar gene. Several studies have deemed TAM receptors involved in flaviviral attachment and entry either by promoting binding or activating TAM receptors (Hamel et al., 2015, Meertens et al., 2012, Battacharya et al., 2013). Recently, focus has shifted to be on particularly the Axl component of TAM receptors because it is highly expressed on trophoblasts and neurprogenitor cells—some of the key targets of ZIKV infection (Tabata et al., 2016). The study found that high levels of ZIKV RNA was found in eyes with acute uveitis, and infectious virus particles were detected in the lacrimal glands and tears. This study used two different low-passage strains of ZIKV (French-Polynesia [H/PF/2013] or Brazil [Paraiba 2015]). In Figure 1, panel A, it shows detection of RNA of the French Polynesia and Brazil strain at day 2 that dramatically increased by day 6 in anti-IFnar1 mAb- treated animals. A similar increase in ZIKV RNA levels were also detected with Paraiba 2015 by day 7 in Ifnar-/- mice (Figure 1, panel B). Figure 1, panel C presents the viral occupancy in the brain and eyes 28 days post infection with Paraiba 2015 in Ifnar-/- animals. Figure 1, panel D and E show levels of ZIKV RNA in tears and lacrimal glands of Ifnar-/- mice that received the Paraiba 2015 strain. To determine whether to RNA detected in the eyes and tears at day 7 and 28 were infectious, the researchers inoculated intraperitoneally AG29 mice with ocular homogenates or tears. AG29 mice lack receptors for interferon I and II and are highly vulnerable to infection. Inoculation with eye homogenates from infected mice at day 7 killed AG29 mice by day 10 post-inoculation (Figure 2, panel A). However, AG29 mice inoculated with eye homogenates (day 28) or tears (day 7) did not present with ZIKV-like symptoms (Figure 2, panel A), suggesting that infectious virus may not be produced in the eye after acute phase of infection. Figure 2, panel B shows a representative image of gross ocular pathology in mice inoculated with parental Paraiba 2015 or eye homogenates from Ifnar-/- mice.
Next, the study looked at the necessity of Axl and Mertk components of the TAM receptor and found that may not be necessary for ZIKV infection of the brain and eyes. These components are in the tyrosine kinase family whose ligands are Gas6 and Protein S and bind to the surface of enveloped viruses and apoptotic cells. DKO mice treated with anti-Ifnar1 antibodies had similar levels of ZIKV RNA in the spleen, brain, serum, and eyes on day 6 after infection compared to wild-type controls (Figure 3, panel A-D). Together the data suggests that in interferon-deficient mice, Axl and Mertk may not be required for ZIKV infection in different tissues. Finally, Figure 4 presents an informative and summative flowchart of the proposed ZIKV infection of the eye. Footpad inoculation of ZIKV in Ifnar-/- or wild-type neonatal mice, causes systemic spread (primary viremia), and the infection spreads to the eye during initial stages of infection, causing release of active particles in tears. Once the eye is infected, several areas are inflamed (conjunctivitis, uveitis, neuroretinitis) (Miner et al., 2016). Being aware of this general sequence of events is critical for understanding the molecular mechanisms behind Zika infection of the eye and for evaluating antiviral therapies for ocular abnormalities. In summary, this study has described tropism of ZIKV to infect certain regions of the eye and shedding and persistence of RNA in tears. Given this data, future studies could look at specific viral and host factors that facilitate ocular infection and how clearance by the immune system may pave way for the development of interventions to better eliminate viruses from immune-privileged sites such as the eye.