Herpes Simplex Keratitis

Latency and Reactivation

The spectrum of ocular disease caused by HSV is broad and is underscored by the ability of HSV to infect a host and establish an indefinite and latent presence in ganglionic neurons. Latency allows for spontaneous and recurrent reactivation of the disease and provides a viral reservoir within the population. The clinical sequelae of HSV infection are largely a result of recurrent disease and the immunologic response associated with each episode. This section describes the mechanisms of latency and reactivation and reviews the available information in this developing area.

After peripheral entry into the host and primary infection with viral replication within an end organ, HSV travels in a retrograde fashion to various ganglia, including the trigeminal, cervical, and sympathetic ganglia, and possibly the brain stem. Here it resides during the life span of the host. This process usually begins within 1–2 days of the primary infection and may take several weeks to complete. Despite an early immune response by the host, ganglionic infection occurs rapidly and does not require virus replication. Once ganglionic presence has been established, active replication in neurons and surrounding cells leads to cell death. It has been suggested that virus replication within the trigeminal ganglion is of paramount importance for viral spread to sites other than the inoculation site; no periocular virus activity was seen in mice whose corneas were inoculated with a virus mutant defective in its replication in nerve tissue. Regardless, latency can be established and is thought to represent the presence of the viral genome within the neuronal cells. ^{, ,} Using a variety of molecular biologic techniques, HSV DNA has been detected in the culture-negative ganglia of latently infected hosts. ^, The cornea itself has also been found to host latent HSV virus, with potential for reactivation.

Latently infected neurons have not been found to produce infectious virus. However, a region of the viral genome that is retained within the host cell nucleus during latency is responsible for RNA transcripts termed latency-associated transcripts (LATs). This region of the viral genome also contains part of the gene encoding an important early transcriptional protein of HSV termed IE0. These LATs are made from the opposite (complementary) DNA strand that is responsible for the normal IE0 message. The role of LATs is unknown. They are thought to be produced in large amounts during reactivation from latency and ultimately act as a type of regulator in the production of viral proteins and infectious virions. Reactivation of HSV results in virus shedding with or without clinical signs.

Although the cornea itself can be a source of virus in recurrent disease, ^, the trigeminal ganglion is the most common source of recurrent HSV infection. Recurrent infection can occur in a different end organ than the primary site. A primary oral or facial infection may subsequently reactivate by traveling via the ophthalmic division of the fifth cranial nerve to the eye instead of recurring at the site of primary infection. The presence of preexisting antigen-specific T-cells and humoral antibodies from primary infection can result in a less severe infection during recurrence; moreover, first-time ocular involvement after infection in another end organ resembles recurrent ocular disease.

Systemic antibodies have no known role in the development of recurrent disease despite their role in the host response to active primary and recurrent infection. It has been demonstrated that antibody titers remain unchanged during and between episodes of recurrent HSV. These titers may even increase without evidence of active infection. ^,

The study of HSV, particularly the differences in strains, has been a frequently updated area of investigation. Some strains have been reported to cause predominantly epithelial disease and even specific patterns of dendritic ulcers. Other strains have a tendency to produce more severe stromal disease, and some are more likely to cause recurrences. ^, In one study by Centifanto et al. DNA analysis revealed that differences in disease patterns may be determined genetically by a specific site on the genome of a virus substrain. A clinical isolate may be composed of several substrains and therefore be competent to produce a spectrum of disease. It also has been demonstrated that some strains cause an epithelial keratitis that is made worse by corticosteroids. There is no clear explanation of this effect, which is further confounded by the observation that the same effect can be seen in vitro where no host-mediated immune system is present.

It has been proposed that the first viral strain to infect the host and establish latency can prevent subsequent infections by a different strain of HSV. This effect was thought to be due to a blocking mechanism whereby other strains could infect the end organ but could not superinfect the associated ganglion. Evidence to the contrary developed when more than one strain of HSV was found in the same trigeminal ganglion of several animals. ^, Previous inoculation with a less virulent strain of virus did provide protection from a more virulent strain, including a decrease in the severity of keratitis and a decrease in the frequency of recovery of latent virions. However, this protection is incomplete and is demonstrated by the recovery of multiple strains of HSV from the same host. In another study, superinfection with a different viral strain, as opposed to recurrence of the original strain, was seen in one-third of cases of recrudescent herpetic keratitis. The implications of such findings would make the development of vaccinations for HSV difficult to perfect because of both the large number of strains and the possibility of several substrains of HSV existing in a single human host.

Many factors have been implicated in the activation of recurrent HSV ocular disease. The mechanisms that control HSV latency are key to viral shedding and reactivation of disease. Intuitively, the immune system should be the core of regulation in infections, and this has been demonstrated to be a factor in reactivation of latent infection. Recent studies of CD8+ T-cell inhibition of HSV-1 reactivation show viral inactivation via the use of lytic granules degrading precursors to viral gene expression. These CD8+ T-cells maintain latency without causing neuronal apoptosis. Sunlight, trauma (including surgery), heat, abnormal body temperature, menstruation, other infectious diseases, and emotional stress have all been implicated in the activation of herpetic disease in human beings. ^, Prostaglandin F2 α analog and prostamide glaucoma medications latanoprost and bimatoprost have also been implicated in ocular or even periocular HSV reactivation. Although some type of immunoregulation may exist in all of these circumstances, it has not been clearly demonstrated. Reactivation of disease has been achieved in experimental animals with immunosuppression. However, this is generally not well demonstrated in human beings. Nonetheless, the incidence of bilateral disease is higher in immunocompromised patients. ^, Psychologic stress, systemic infection, sunlight exposure, menstrual period, contact lens wear, and eye injury were studied as potential triggers of ocular HSV recurrences; 308 immunocompetent adults were observed for 18 months, and weekly logs were kept on those five factors and were considered only if completed before the onset of symptoms. With 33 valid recurrences, none of these factors was associated with a recurrence of ocular HSV. When the 26 recurrences excluded for being reported late were examined, high stress and systemic infection were found to have been reported significantly more frequently than in the 33 valid responses. This was a very characteristic example of recall bias in a prospective cohort study.

Overall, the severity and frequency of disease are multifactorial and may depend not only on emotional, physical, and immunologic stress but also on the viral genome and its virulence.

Immune Defense Mechanisms

Both humoral and cellular responses are part of the host response to HSV infection and may indeed be responsible for limiting spread of disease and for the pathologic sequelae in the local tissue. In ocular infection, viral replication and immunologic response from infection are responsible for destructive changes associated with necrotizing stromal keratitis, immune stromal keratitis, and endotheliitis. ^{, ,} The pathogenesis associated with these manifestations of HSV infection is not clearly understood and has been the focus of many investigations.

Stromal inflammation from HSV keratitis is thought to be due to either replicating virus or the alteration of antigenicity of the stromal cells, allowing for recurrent immune-mediated cycles of inflammation. Details of this process were made more clear by Easty, who summarized his studies of HSV stromal keratitis. He showed that HSV may enter stromal keratocytes, where it can be eliminated or escape the immune response and persist within this layer. The virus also can exist within the stroma as a whole virion shed into the stroma from neurons to incite an inflammatory response, or as a slowly proliferating intracellular virus while altering the antigenicity of the cell wall and eliciting an immune response. Latent existence, during which the cell remains antigenically unchanged and there is no active clinical disease, also may occur. In this case a reactivation would result in disease. In some patients the disease process is caused by an alteration in cell antigenicity that can lead to an autoimmune-mediated response. Because intact HSV is rarely found in human corneal transplant specimens of patients with a history of HSV keratitis, it has been proposed that residual viral antigens, along with the viral immune alteration of cell membranes in the corneal stroma, can produce an inflammatory response. This response involves lymphocytes, bystander activation of CD4+ T-cells, antiviral antibodies, serum complement, polymorphonuclear leukocytes (PMNs), and macrophages. ^{, , ,}

Molecular mimicry mechanisms have also been proposed for autoimmune disease after HSV infection; viral determinants, such as coat proteins, mimic host antigens, to trigger self-reactive T-cells and destroy host tissue.

The endotheliitis seen with HSV and described in a later section also has been studied. Sundmacher et al. proposed that endothelial cells become infected with HSV and therefore elicit a cellular and humoral response. In this theory, cell lysis is thought to be a product of the immune response, with release of infectious virus into the aqueous. However, microscopic examination of endothelium in such cases has failed to show viral particles. A delayed-type hypersensitivity response to herpes antigens within the stroma or endothelium also has been proposed as a mechanism responsible for endotheliitis and overlying stromal edema. Furthermore, the delayed-type hypersensitivity reaction not only plays a protective role but also has been implicated in the cause of corneal opacification following HSV infection.

Secretion of glycoproteins by HSV also has been implicated in the immune response. Strains of HSV capable of causing stromal disease release greater amounts of glycoprotein than those that cause primarily epithelial disease. ^, The exact types of glycoprotein and their roles may be unclear; however, the antigenicity of this protein is thought to be one factor in stimulating the host immune response.

One other puzzling and clinically pivotal aspect of stromal inflammation is neovascularization. In contrast with human herpes virus 8, responsible for Kaposi sarcoma, HSV-1 and HSV-2 are not thought to produce proteins that are directly angiogenic. Vascular endothelial growth factor (VEGF), a group of proteins that are the most important thus far studied specific mediators of angiogenesis, have been detected in various inflammatory corneal conditions. VEGF is expressed rapidly, within 24 hours of HSV-1 virus inoculation in mice, by cells other than those directly infected ; severity of stromal inflammation was reduced with a VEGF inhibitor in that same experiment and with an endothelial cell inhibitor in another study by the same investigators. Similarly, matrix metalloproteinase-9, produced by neutrophils and upregulated by HSV infection, induces angiogenesis in mice and its inhibition appeared to diminish the rate of neovascularization. The role of and the potential for inhibition of angiogenesis in the pathogenesis of stromal keratitis unravels a fascinating area of research, as previously done with retinal neovascularization.

Although the immune defense mechanisms of HSV infection require further investigation, it is clear that there are both beneficial and harmful effects. We know that host defense is mandatory in eliminating virions and thus in preventing the direct toxicity associated with infection. The detrimental effects of this response include local tissue destruction, scarring, and recurrent inflammation. Although this phenomenon may occur in many organs and with many infections, ophthalmologists have the capability of monitoring these secondary effects under microscopic view. Unfortunately, the detrimental effects of the immune response associated with HSV keratitis can result in devastating visual results. Therefore we are challenged to distinguish the dual role of immunity; to do so allows for the most timely and appropriate medical management.

Congenital and Neonatal Ocular Herpes

Fortunately, congenital ocular herpetic disease is rare. When all cases of HSV ocular disease are considered, HSV-1 accounts for most cases. Because most congenital cases are acquired in conjunction with genital herpes in the mother and during parturition, HSV-2 accounts for 80% of cases. In general, HSV infects between 1500 and 2000 neonates each year in the United States : 4% are acquired congenitally, 86% at birth, and 10% postnatally. In most cases, infection of the mother is a primary genital infection occurring during early to midgestation, with 60%–80% being asymptomatic or unrecognized at the time of delivery. It is important to recognize that other routes of acquisition are also possible and include oral lesions, maternal breast lesions, and nosocomial transmission. In addition, infection localized to the skin, eye, or mouth will progress to disseminated disease and/or central nervous system involvement in 75% of cases if not treated.

Ocular manifestations include periocular skin lesions, conjunctivitis, epithelial keratitis, stromal keratitis, and cataracts. ^, Posterior ocular findings are discussed elsewhere. Although maternal immunoglobulin G (IgG) to HSV may cross the placenta, it does not appear to be sufficient to prevent ocular disease completely. However, maternal IgG may quell the clinical course in both systemic and ocular infection. The use of antibody titers for diagnosis is not useful because of preexisting maternal antibody and the delayed production of IgM. The clinical course of congenital and neonatal ocular herpes is similar to primary disease and is discussed later.

Because of the potentially serious complications of congenital and neonatal ocular HSV disease, prompt medical treatment is mandated in conjunction with amblyopia therapy. Moreover, skin vesicles, mouth ulcers, conjunctivitis, or other ocular manifestations suggestive of HSV are all indications for empiric treatment with acyclovir, which should be started in conjunction with a pediatric infectious disease consultation.

Primary Ocular Herpes

The passage of maternal anti-HSV IgG allows for partial immunity for 6 months. At this point, the passive immunity has diminished to levels that allow for a primary infection of HSV. By the age of 5 years, nearly 60% of the population has been infected with HSV. Latent infection is the usual course and a viral carrier state is established. Only 6% of those infected actually develop clinical manifestations, which typically affect the perioral region rather than the eye.

The presentation of primary ocular HSV is varied and may include acute follicular conjunctivitis, keratoconjunctivitis, preauricular adenopathy, and periocular and eyelid skin vesicles ( Fig. 78.1 ). Adenoviral infections can mimic these findings except where typical vesicular skin lesions are present. Pseudomembranes also may be present and can further confuse the diagnosis in some cases.

Although rarely a cause, primary HSV should be considered in all cases of keratitis. Early involvement of the cornea may present as a diffuse punctate keratopathy as well as corneal vesicles. During the early stages of infection, epithelial cysts may form and stain negatively with fluorescein. These findings are best appreciated with broad slit lamp illumination. It is thought that this stage of HSV keratitis is similar to the early vesicular lesions of the skin. Eventually these microcysts evolve to erode the overlying epithelium and form microdendritic lesions. Although primary disease is thought to be confined to the epithelium because of a lack of previous immunologic stimulus, it also may be the reason why a large and diffuse involvement of this layer is common.

Recurrent Ocular Herpes

In Liesegang’s review of the epidemiology of the clinical manifestations of ocular HSV, recurrence rates were 36% at 5 years and 63% at 20 years after a primary episode. After a second episode, 70%–80% of patients had another recurrence within 10 years. Schuster et al. reported a 33% recurrence within 2 years in patients with two prior infectious epithelial keratitis episodes. The HEDS reported an 18% recurrence rate for epithelial keratitis and for stromal keratitis within 18 months in 346 patients with history of ocular HSV within the previous year.

Blepharitis

Herpetic blepharitis can result from a primary infection or recrudescence. The clinical appearance is a vesicular lesion involving a focal area of the eyelid with surrounding erythema similar to that seen at the mucocutaneous junction of the mouth or nose ( Fig. 78.2A ). The typical lesion progresses to ulceration and crusting and heals without a scar unless secondarily infected. Many patients with herpes blepharitis experience recurrences of their lid disease only. However, some patients may develop subsequent keratitis after any episode of HSV blepharitis; therefore this lesion should be recognized and properly treated. Because this lesion is relatively uncommon, it is often misdiagnosed. However, it can be separated from more common causes of blepharitis, such as staphylococci, seborrhea, or meibomian gland disease, which tend to involve the entire lid and do not cause vesicles. Diagnosis of HSV blepharitis can be made on the typical clinical appearance and course of a recurrent, focal, vesicular lesion and through viral culture or PCR testing.