Biology and Molecular Aspects of Herpes Simplex and Varicella- zoster Virus Infections Thomas]. Liesegang, MD
T
he herpes simplex and varicella-zoster viruses are members of the subfamily alpha herpesviruses with specific properties of the virion and with the capacity to establish latent infections in humans. The genome of each of these viruses has been determined with an estimate of the number of genes and proteins encoded. The biology and molecular events of the herpes simplex virus productive and latent infection have been detailed with the use of both in vitro and in vivo model systems. The neuron is the site of latency in the ganglia with a limited transcription of genes expressed during the latent period. The specific molecular regulation of latency and reactivation are not well established. There are co-cultivation, electron microscopy, and biochemical studies that support the concept of corneal latency, although this has not been proven conclusively. Details about the varicella-zoster virus biology and molecular events are not as well advanced since animal models have been lacking. The biology of the productive infection (varicella) is different from herpes simplex virus infection since the portal of entry is the respiratory system. Data support the concept of the maintenance of latency within satellite cells in the ganglia rather than within neurons. There are multiple genes expressed during this latency. These features may explain the different clinical presentations and course of reactivation (zoster) compared with herpes simplex virus reactivation. Ophthalmology 1992;99:781-799
The herpes simplex virus (HSV) had been the most in tensively studied of all viruses until the recent research on the human immunodeficiency virus. The HSV serves as a model to study synaptic connections in the nervous system, membrane structures, gene regulation, translo cation of protein, and many other biologic functions. Complex and sophisticated studies using human and an imal tissue in vitro and in vivo have begun to unravel the unique symbiosis that exists in nature between humans and the HSV. The biology and molecular aspects of the numerous stages of the host-viral interaction have been elucidated by these studies. Some of the stages explored have been the development of the primary infection at Originally received: October 14, 1991. Manuscript accepted: December 2, 1991 . Presented in part at the American Academy of Ophthalmology Annual Meeting, Anaheim, October 1991 . From the Mayo Clinic Jacksonville, 4500 San Pablo Rd, Jacksonville, FL 32224.
the epithelial site, the mode of spread to other sites with a preference for the sensory and autonomic ganglion, the host control of primary infection and the limitation of HSV from dissemination, the tissues that establish and then maintain latency, the nature of the viral genome during latency, the association of the viral genome with host cellular macromolecules, the mechanism of reacti vation of latent infection, the establishment of recurrent disease at the epithelial site, and the host control of re current epithelial disease. Animal models have served as a guide, 1 but, despite culture assays, biochemical analysis, immunofluorescent studies, and ultrastructural tech niques, we have not been able to define this host parasite relationship clearly. Because varicella-zoster virus (VZV) remains predom inantly cell-associated throughout the virus replication cycle in vitro, it has not been possible to obtain high titer infectious cell-free virus. There is also no easily susceptible animal, so model systems are not well developed. There fore, the biologic, molecular biologic, and immunologic
781
Ophthalmology
Volume 99, Number 5, May 1992
aspects of VZV pathogenesis have not advanced as rapidly as the knowledge of the HSV. Much of our knowledge of VZV has been gained from comparison and deduction from in vitro and in vivo HSV animal models. The immune mechanisms moderating HSV and VZV infection act at multiple stages and are even more complex than describing the biologic and molecular aspects of the infection. 2•3 The host response may influence the acqui sition ofdisease, the severity ofinfection, the development and maintenance oflatency, and the frequency of recur rences. The discussion here will be limited to a current working hypothesis of the biologic and molecular aspects ofHSV and VZV within the human host, albeit extending some mechanisms that have only been demonstrated in animal models or in in vitro studies. The acute ocular HSV disease appears to be comparable in the rabbit, mouse, and human. 4 Latency may not be identical since frequency of spontaneous HSV viral shedding differs among rabbits, mice, and humans. The trigger mecha nisms for reactivation also differ among species. 4 This report will focus on recent studies, but it is not meant to ignore the multiple investigators who, over the past few decades, have provided the foundation that current biol ogists are building on.
The Herpes Virus Viruses are acellular infectious particles that are incapable of metabolic activity outside living cells. Genetic infor mation is stored either in the form of RNA or DNA and can only be expressed within a living cell at the expense of that cell's own machinery and energy. The mature virus particle is known as a virion. Both HSV and VZV are DNA viruses approximately 150 to 200 nanometers in diameter and composed ofa nucleic acid and protein core, a protein capsid, a tegument layer, and a lipid bilayer envelope (Fig 1). The central nucleic acid is linear, double-
Herpesvirus
Nucleocapsid
Protein
Nucleocapsid
Figure 1. Schematic drawing of a herpesvirus. The protein of the her pesvirus is surrounded by the DNA, much like thread on a spool. A protein structure called the capsid is in the shape of an icosahedron and surrounds this DNA genome core. This combined structure is called a nucleocapsid. An additional phospholipoprotein envelope surrounds the nucleocapsid with glycoprotein spikes projecting from the surface. The tegument is an amorphous protein structure between the nucleocapsid and the envelope. The complete infectious particle is called a virion.
782
Symmetry and Rotational Axis of an Icosahedron
Figure 2. The icosahedron, showing the elements of icosahedral symmetry as viewed from the two-fold, three-fold, and five-fold axes of rotational symmetry. The icosahedron possesses 12 vertices, 20 triangular faces, and 30 edges. (Modified and published courtesy of ]oklik WK, Virology, 3d ed, chap. 2. Appleton & Lange, 1988 and Harrison SC. In: Fields BA, Knipe OM, eds. Virology, 2d ed, val. 1, chap. 3, Raven Press, 1990.)
stranded, and relatively guanine and cytosine rich, with a molecular weight of approximately 100 X 106 daltons with VZV being slightly smaller. Protein surrounds this DNA perhaps like thread (the DNA) on a spool (the pro tein). This nucleoprotein is surrounded by a protein coat, the capsid, which is made of repeated polypeptide chains. The capsid protects the nucleic acid, confers symmetry on the virus, and allows the introduction of viral genome into the host cell. The herpesvirus has a cubicle symmetry in which the surrounding capsomer building blocks form an icosahedron with a 5:3:2 symmetry, which is con structed from equilateral triangles (Fig 2). Although the virus appears spherical on electron microscopy, it is ac tually an icosahedron composed of symmetric subunits. Forming the icosahedron are 12 pentavalent capsomers present at all the vertices; 60 hexamers form the 20 faces, and 90 hexamers form the 30 edges-making a total of 162 capsomers. The herpesvirus also has an envelope, which is a complex phospholipoprotein structure derived from the cytoplasmic membrane of the host cell that has been modified by viral protein. Virus-encoded glycopro tein subunits, known as peptomers, are incorporated into the envelope and project from its outer surface. Both pep tamers and capsid proteins are antigenic and induce spe cific antibody response in the infected host. Only envel oped virions are capable of entry into cells and initiating replication. Between the capsid and the viral envelope is an amorphous protein structure called the tegument. This protein is extremely important in inducing active viral transcription in t!w nucleus and shutting off the host pro tein production. HSV Type I, HSV Type II, and VZV are members of the subfamily alpha herpesviruses, which are classified based on properties of the virion, host range, site of la tency, and the replication cycle (Table 1). Alpha viruses
Liesegang · Herpes Simplex and Varicella-zoster Virus Infections Table 1. Human Herpesvirus Virus
Subgroup
Designation
Disease
a a a
HHV1 HHV2 HHV3
'Y
HHV4 HHVS HHV6 HHV7
Oral ocular HSV Genital HSV Chickenpox Herpes zoster Infectious mononucleosis Mild infection Exanthem subitum Unknown
Herpes Simplex Virus Type I Herpes Simplex Virus Type II Varicella-zoster virus Epstein-Barr virus Human cytomegalovirus Human B-celllymphotropic virus Human herpes virus-7 HSV
=
{3 {3 {3
Guanine and Cytosine Content (mole%) 67% 69% 46% 60% 57% 40% Not known
herpes simplex virus.
have a variable host range, a short growth cycle, spread rapidly with efficient destruction ofinfected cells, and have the capacity to establish latent infection primarily in gan glia and hence are neurotropic viruses. Beta viruses are typified by a narrow host range, a long reproductive cycle in culture with slow virus spread, the enlargement of in fected cells, and the capacity to establish latency in secre tory glands and lymphoreticular cells. Gamma viruses are associated with lymphoproliferative disease in their nat ural host and have a host range restricted to the natural host. To date, seven herpesviruses whose natural hosts are humans have been identified. The human herpesvirus Type 6 causes exanthem subitum, or roseola, a common childhood disease. HHV-7 has not yet been associated with a disease. The neurotropic herpesviruses cause similar infections although they are usually easily clinically distinguishable. Primary HSV infection is localized, albeit with some sys temic manifestations. Primary VZV infection (varicella) is a disseminated infection. Primary HSV infection is spread by contact with epithelial or mucous membranes and then it gains access to sensory nerve endings. The predominant mode of spreading VZV is by the respiratory route. After local replication in the respiratory tract, the virus is disseminated by the blood and lymphatics (pri mary viremia). The virus is then taken up by cells of the
reticuloendothelial system, where it undergoes multiple cycles of replication during the remainder of the incu bation (up to 17 days incubation). A more extensive sec ondary viremia develops followed by mucosal and skin lesions (varicella). The VZV then passes centripetally from the skin and mucosal lesions to the corresponding sensory ganglion by the contiguous sensory nerve endings and sensory nerve fibers. It is possible for virus to seed ganglia by hematogenous spread. Varicella is usually uneventful, but, in newborns and immunosuppressed children and adults, complications are common. 5 HSV Types I and II are distinct but antigenically re lated. HSV Type I generally involves infection above the waist (ocular and facial) and HSV Type II below the waist. There are unique biologic properties and distinctive dif ferences between these viruses in tropism and site-specific recurrence frequency 6 (Table 2). During the primary infection, both HSV and VZV es tablish latency in sensory ganglion. HSV produces fre quently recurring localized infections with mild discom fort, whereas VZV rarely recurs more than once but is characterized by more severe pain. The lesions of HSV are usually few and limited in distribution whereas zoster spreads to involve much of the cutaneous dermatome. These differential features suggest that the nature of the acute and latent infection in sensory ganglia differ and
Table 2. Herpes Simplex Virus Infection Ocular HSV Orofacial HSV Genital HSV Neonatal HSV Disseminated HSV Encephalitis HSV Meningoencephalitis HSV HSV
=
Predominant Virus Type 1 1> 2> 2> 1> 1 2
2 1 1 2
Outcome
Recurrence
Visual impairment Resolution Resolution Retardation Resolution or death Neurologic damage or death Resolution
Yes Yes Yes No No No No
herpes simplex virus.
783
Volume 99, Number 5, May 1992
Ophthalmology
the mechanisms of reactivation from latency may also differ.
The Viral Genomes The HSV genome is approximately 152,000 base pairs in size and codes for approximately 72 genes and 70 different proteins. 7 The VZV genome is approximately 125,000 base pairs in size and codes for approximately 68 genes and 80 proteins. 8 The HSV genome is composed of two portions: a unique long region and a unique short region with gene sequences that occur only once. Both these long and short unique sequences are flanked by repeat se quences that occur at the termini and internally in the genome (Fig 3). A hinge point exists between the internal repeat sequences such that each unique sequence can flip flop relative to the other, generating four possible isomeric forms of the complete DNA molecule (Fig 4). Thus, viral genomes extracted from infected cells consist of four equimolar isomers that differ solely in the orientation of the long and short components. 9 The biologic functions of these inverted repeats are unknown, although deletion of small portions does not effect the ability to replicate or establish latency. The characteristic genome anatomy with two longer unique components, both bordered by inverted repeat units, makes it possible for the genome to circu larize, which may be essential for viral DNA synthesis and for the establishment oflatency. 10 HSV I and II have approximately half the genome sequences homologous. There is one major exception with the insertion of some 1500 base pairs into the coding sequence in the HSV Type II unique short section. The genes are oriented in both directions. 7 The VZV genome also is composed of two portions: a unique long region of approximately 105,000 base pairs and a unique short region of5200 base pairs (Fig 3). Each of these pairs is flanked by inverted repeat elements of Isomeric forms, no.
Genomes
-
HSV 1, 2
vz:v 0
- 50
Molecular
4
2 100
wt, x 10 6
Figure 3. Schematic drawing of the herpes simplex virus and varicella zoster virus genomes. Each of these viral genomes consists of long and short unique sequences (Ul, Us) flanked by long stretches of bases that are duplicated as inverted repeat elements. These are designated as ter minal or internal repeats surrounding the long and short unique sequences (TRl, TRs, IRl, IRs, respectively). Portions of the genomes that are actively transcribed in latently infected ganglia are shown as heavy arrows below the genome maps. The relative molecular weights are shown as are the number of isomeric forms.
784
Figure 4. The isomers of HSV and VZV. The unique long and unique short regions (Ul, Us) of HSV can both invert at a hinge point in the internal repeat unit resulting in four isomers that occur with equal fre quency during an infection. The unique short region (Us) of VZV also can invert, giving two major isomers. Only about 5% of VZV virions have an inverted unique long section (Ul) resulting in two minor isomers since the Us region can also invert in either direction. The internal repeat regions are very important in allowing inversion, circularization of the genome in the nucleus, intragenomic recombination, and permitting the rolling circle mechanism of DNA production.
shorter base pairs (8000 base pairs). During DNA repli cation, the unique short region (and its repeats) inverts, producing two isomeric forms of VZV DNA. Approxi mately half of the viral particles contain the unique short region in one orientation and half in the other. Further studies have shown that a small proportion of the genome (perhaps 5%) also have the unique long region inverted (Fig 4 ). 11 Thus, VZV virus contains four genome isomers (two major and two minor). This contrasts with HSV in which both segments (unique short and unique long) in vert with equal frequency and result in four isomeric forms. All these VZV and HSV isomers are equally in fectious. The VZV and HSV genomes share several features: two unique regions (unique short and unique long) flanked by an inverted repeat and four genome isomers of virion DNA, albeit in different proportions. 8 The unique short and the repeat units (terminal repeat long, internal repeat long) are shorter in VZV, and the HSV genome also is terminally redundant with a sequence of 200 base pairs, which is repeated directly at the genome ends and is also present in the inverse orientation at the junction between the internal repeat long and the internal repeat short. The guanine and cytosine content in VZV is 46% compared with 67% for HSV I and 69% for HSV II. The complete genome sequence for HSV I and VZV have been determined using bacteriophage cloning and chain termination sequencing techniques as well as re striction endonuclease analysis. 8 • 12 The VZV genome is not a unique size but may vary between limits of ap proximately 124,000 to 126,000 base pairs. 8 The avail ability of DNA sequencing, the recognition of open read ing frames and coding proteins, and other transcriptional regulatory sequences, coupled with computer analysis, have permitted a much greater understanding of the re lationship between herpesviruses and allowed comparative study of individual genes. The genetic relationship and similar DNA arrangements between VZV and HSV Type
Liesegang · Herpes Simplex and Varicella-zoster Virus Infections I have allowed an assumption of the VZV gene functions based on the more extensive knowledge of the HSV gene function. 7 • 12 The localization of these coding sites for pro tein may be important because they are the targets for antiviral therapy. There are also unique reiterated ele ments of the genome, which have proven useful in in vestigating the epidemiology of VZV and HSV. Both VZV and HSV Type I have similar core genes. VZV has some genes that have no HSV counterpart and vice versa. Some regions of the genome exhibit relative rearrangement of genes. 7 A schema for the evolution of VZV and HSV Type I from a common ancestor by a series ofrecombinational events resulting in the expansion or contraction of terminal repeat short and internal repeat short segments has been suggested. 8 During the divergence of HSV-1 and VZV, very extensive processes of incor poration of mutations have evidently taken place to the degree that VZV has an overall genome base composition of 46% compared with 68.3% guanine plus cytosine of HSV 1. It is probably these genetic differences that play a major role in the disease pattern. Unfortunately, the lack of similar animal models of VZV have hindered proving several hypotheses. Both VZV and HSV Type I also have a close relationship to the Epstein-Barr virus. 13 The gene locations have been extensively studied in herpesviruses with special emphasis placed on defining the major glycoprotein families that are important in im mune recognition and perhaps future vaccines. The gly coproteins may be especially important in pathogenesis because they become incorporated into the envelope of the virus as well as brand the host cell and are involved in the first interaction between the virus and the cells that it can infect. The origin of replication of DNA has been mapped to a homologous sequence in both HSV and VZV. HSV Type I contains two origins of DNA repli cation, whereas VZV has one origin of replication. 11 There have been seven genes that appear essential for HSV DNA replication with each of these having a homologue in VZV. 7 Restriction-endonuclease analysis of viral DNA recovered from clinical isolates has proven that latent varicella virus reactivates to cause zoster. 14 Different viral isolates from varicella or zoster patients or patients with varicella with subsequent zoster in the same individual show slight variation in the mobility of certain DNA frag ments. Isolations from a single outbreak of varicella, however, have a stable genome structure. HSV demon strates the same basic genome stability. In laboratory animals, diverse clinical outcomes result from different HSV viral types and strains. These reactions are complex to assign but may depend on viral genetic constitution, 15 • 16 on the viral challenge dose, or on the host's genetically determined immune response. 17- 20 Viral strains differ in regard to neuroinvasiveness and neuro virulence, terms that are related but not identical. Neu rovirulence is the viral capacity to replicate and cause disease in the nervous system after direct inoculation into these tissues. Neuroinvasiveness relates to the capacity of the viruses to enter and move through the nervous system after inoculation at some peripheral site. There have been several investigations mapping HSV gene functions as
sociated with significant biologic effects and studies of the properties of the products encoded by these genes. 21 The HSV-I strain characterized by neuroinvasiveness has been linked to encephalitis. Gene mutations that occur in tem perature-sensitive mutants of HSV Type I can affect the penetration, the formation of functional capsids, and viral structural protein synthesis. 22 The specific nuclear base sequence of DNA polymerase genes can contribute to dif ferences in the capacity of HSV Type I and Type II to replicate in the trigeminal ganglion and spread from the eye to the central nervous system. 23 •24 Identifiable and reproducible ocular lesions of distinctly different size, lo cation, and shape can result with different strains of HSV in animals. 15 • 16•25 In animal models, some HSV strains produce consistently mild epithelial disease, some produce epithelial and stromal disease, and others produce uveitis. Strains of HSV that produce more severe stromal disease have greater amounts of envelope glycoprotein-D than other strains that produce less severe stromal disease. 26- 29 Steroid response to treatment of HSV also is genetically regulated. 30 The ability to reactivate the virus is strain specific in the rabbit and can be separated into sensitivity to endogenous (spontaneous) reactivation and exogenous (induced) reactivationY HSV thymidine kinase activity is important in determining the severity of keratitis, the extent of ganglionitis and encephalitis, and is important in promoting latency and reactivation from latency, al though it is not important in establishing latency. 10•32 Mutants with thymidine kinase deficiency have remark ably different patterns of disease in experimental mice. 33 The specific DNA map units responsible for many of these features have been determined. 27 •34 Caution is advised, however, in overinterpreting animal data because the nat ural infection in humans involves complex cellular inter actions that may have more relevance than some of these biologic features in animals.
Primary Herpes Simplex Virus Infection The natural history of HSV Type I infection in humans is generally characterized by a childhood infection that may present with a nonspecific upper respiratory infection or be asymptomatic. On rare occasions, primary HSV Type I infection may be associated with ocular tissues and present as a bilateral vesicular lid edema, ulcerative blepharitis, or epithelial keratitis. These cells are ultimately lysed. The primary infection defines the pathway of viral spread and the site of viral latency. 35 Primary infection in humans occurs most commonly within the body surface innervated by the trigeminal nerve (i.e., the mucocuta neous areas of the face, eye, and oral mucosa). HSV can be spread through the nasal passages to the olfactory sys tem of the central nervous system and also through the esophagogastric mucosa to the ganglion ofthe vagus nerve. It also has been implicated as a cause of recurrent peptic ulcer. 36 Autopsy studies reveal that nearly all patients in fected with HSV house DNA in selected nerve ganglia. The trigeminal ganglion are the most commonly infected site followed by the sacral root ganglia.
785
Ophthalmology
Volume 99, Number 5, May 1992
Although primary infection (when apparent) tends to cause more severe symptoms than recurrent infection, it is self-limited with complete recovery and disappearance ofHSV from the initial site of infection. In immunocom promised individuals and newborns, HSV can spread from the primary replicating site to other organs and may be fatal. 37 Primary HSV Type I infection involves primarily ectodermally derived tissue (i.e., skin, mucous mem branes, or corneal epithelium). Initial HSV Type II also effects ectodermally derived tissue, either from contact with the birth canal of an infected mother or later in life through contact with sexual partners. These mucocuta neous sites are unique in several aspects: they are con nected to the outside world, they are hormone sensitive, have a rich blood supply, and have a rich innervation by somatic, sensory, or autonomic nerves. The primary infection in laboratory animals involves an initial viral replication in mucocutaneous tissues. Vi ruses infect cells ifthey have an appropriate receptor mol ecule on their outer surface, and different receptors are probably recognized by different viruses. The wide range of hosts suggests that HSV receptors on cell surfaces are widely distributed. After a productive replication in the mucocutaneous site with spread of progeny virus to con tiguous cells, there is extension to the mucocutaneous projections of the sensory nerve probably within hours. 38 A receptor for HSV attachment to the·nerve termini may be structurally similar to heparan sulfate, a common con stituent of plasma membranes. 10 The herpes virion at taches to the receptor with its surface glycoproteins an choring the virus with a fusion reaction of the viral en velope with the plasma membrane of the neuritic extension of the sensory nerve. 39 ·40 The viral glycoprotein spikes on the herpes simplex virus envelope determine both the attachment and the fusion activity by which the virus is internalized. This fusion results in the release of naked viral nucleocapsids within the axoplasm 41 (Fig 5). After penetration by fusion, the naked viral nucleo capsids are transported by microtubular dependent ret rograde axonal flow or by membrane-bound organelles to the cell nucleus, at about the rate offast axonal transport of protein, which is 5 to 10 mm per hour. 40-42 As part of normal function, neurons make a variety of products that are packaged and transported along the axon. It appears that microtubule gliding and movement of organelles are mechanisms for HSV axonal transport toward the nu cleus.43 Because the virus has lost its envelope, it is non infectious and probably unable to leave the axon. There is probably loss of the tegument material that surrounds the nucleocapsid during transport, but otherwise it seems essential that the nucleocapsid be structurally maintained until it reaches the nucleus. Once inside the neuron, the noninfectious forms must undergo at least one cycle of replication to account for the infectious (enveloped) forms detected in neurons in the next phase of infection (i.e., when the infectious virion migrates centrifugally from the ganglion to the inoculation site). The virus replication at the inoculation site causes the development of the characteristic lesion. 44 Virus can be transported centrally to reach the ganglion nerve root and may leave the neuron through membranes
786
late
~
5) Assembly
Figure 5. Schematic replication ofHSV in a susceptible cell. 1) The virus fuses with the plasma membrane after attachment at specific receptor sites which recognize HSV Type I. The empty envelope is left at the plasma membrane. 2) The nucleocapsid is introduced into the cytoplasm and then transported to the nuclear pore by fast axoplasmic flow. Two proteins are released from the tegument of the virion. One shuts off host protein synthesis and the other is transported to the nucleus as a trans inducing factor (Vmw65). 3) The viral DNA is released into the nucleus and becomes circularized. The empty capsid coat is left at the nuclear pore. 4) In the presence of the tegument transinducing factor, the tran scription of the immediate early genes occurs with transport of mRNA to the cytoplasm and translation to protein products. These products induce the transcription of early genes with transport of mRNA to the cytoplasm and translation to beta proteins which are involved in DNA synthesis. DNA synthesis occurs by a rolling circle mechanism that yields viral DNA. Transcription of gamma genes results in gamma proteins which consist of structural proteins of the virus. 5) The capsid proteins are constructed into complex icosahedron structures, which are packaged with viral DNA cleaved from the rolling DNA concatamers. Viral gly coprotein and tegument protein accumulate and alter the internal nuclear membrane. The nucleocapsid is probably enveloped as it passes through the internal nuclear membrane, de-enveloped between the inner and outer nuclear membrane, and then leaves the nucleus. 6) The virus is transported to the endoplasmic reticulum and Golgi apparatus where it is enveloped and then transported via vesicles to the nerve periphery and released by exocytosis into the extracellular space or to contiguous cells. The glycoprotein spikes of the envelope are matured as the virus travels away from the nucleus (Modified from the concepts of Roizman B, Sears AE 9 ).
to infect astrocytes and oligodendroglia cells with subse quent damage and infection of other axons within the same nerve root giving a "backdoor" into other divisions of the ganglia. Myelinated nerves around Schwann cells may act as a physical barrier, but, in nonmyelinated fibers, there is no physical barrier preventing passage of virus from the axon to the Schwann cell cytoplasm. Adjacent myelinated and nonmyelinated Schwann cells can be in fected and enveloped virus may be transmitted by this mechanism the full length of the peripheral nerve within a few days. Other hypotheses for the mechanism of viral transport from the peripheral site include transport in the periaxonal space, by sequestered infection of Schwann cells, or the lymphatic channels of the epineurium, but all are unlikely. 45 When nucleocapsids have reached the membrane of the nucleus, the viral DNA is released into the nucleus, probably through the nuclear pores. The empty capsids are left outside the nuclear membrane.
Liesegang · Herpes Simplex and Varicella-zoster Virus Infections On entry into the nucleus, there is circularization of the viral DNA genome and initiation of viral replication. During the active infection, the viral nucleocapsids are assembled in the cell nucleus and then enveloped at the inner lamina of the nuclear membrane, de-enveloped on passage of the outer nuclear membrane, and leave the outer nuclear membrane as naked nucleocapsids. A sec ond and final envelopment of the virus occurs at the membranes ofthe endoplasmic reticulum, where the virus also receives a further enclosure in the transport vesicle. The vesicle is important for protection, transportation, fusion with the plasma membrane, and for the maturation of the glycoprotein spikes. The enveloped virus is trans ported along the nerve in transport vesicles probably de rived from both endoplasm reticulum and Golgi apparatus ultimately to arrive at the nerve endings. The glycoprotein spikes of the envelope that were acquired at the inner nuclear membrane become more mature when the virus passes through the endoplasmic reticulum and the Golgi apparatus. When it reaches the nerve endings, the plasma membrane of the nerve is changed morphologically and chemically during its fusion with the virus transport ves icle, and the virus is released probably by exocytosis after envelopment at the plasma membrane and by fusion of the vesicle carrying enveloped virus at the plasma mem brane.40 The host plasma membrane is branded with virus specific glycoproteins, which makes it a target for im munologic attack. The route of envelopment and release of herpesvirus is similar to the route of synthesis, pro cessing, and release of secretory glycoproteins. In the eye, the virus released from the corneal nerve termini may infect the corneal and conjunctival epithelium, and re establish active HSV infection. The duration of reactivated infection is usually shorter because of previous host cel lular and humoral anti-HSV response. In the eye, how-
The HSV "Round Trip"
Figure 6. The "round trip" hypothesis of HSV infection. The initial replication of HSV at the inoculation site may not be essential in causing clinical disease. If the virus has access to nerve endings, it can travel as a naked nucleocapsid to the nucleus where it can undergo a replicative cycle and then travel down the nerve as an enveloped mature form. Release can then either cause the clinical lesion or at least contribute to the progression of the clinical lesion. This may explain the incubation period for clinical disease and is compatible with the "backdoor" spread ofHSV.
"Back Door"
lnfection~wt
h HSV
Trigeminal ganglion Ophthalmic branch
f.---=-=
Maxillary branch
Mandibular branch
Figure 7. The "backdoor" approach to the ophthalmic nerve causing ocular herpes. The most common site of initial HSV infection is in the facial area served by the maxillary division of the trigeminal nerve. In· fection here can result in latency in the maxillary portion of the trigeminal ganglion. At the time of initial infection or with reactivation, virus can spread to the ophthalmic or mandibular portion of the trigeminal nerve and can perhaps cause HSV infection in the ophthalmic division of the trigeminal nerve without ever having had a prior skin or mucous mem· brane HSV infection in this distribution.
ever, repeated episodes may enhance the immune-in flammatory reaction causing structural alterations in the cornea. The extracellular release of virus from the peripheral sensory nerve now allows further infection of the der matome and also allows subsequent spread to satellite or support cells around the nerve. Indeed, the initial mul tiplication of virus at the inoculation site may not be es sential in causing clinical disease or even in establishing latency if the virus has access to the nerve endings. The virus travels up thy axon as an unenveloped particle and travels centrifugally as an enveloped mature form. The delay from time of inoculation to the development of clinical lesions (5 to 8 days) may be explained by this "round trip" theory with the clinical disease actually re sulting from viral release from nerve endings (Fig 6). 44·46 As a further extension of this hypothesis, HSV infection in the eye is probably more commonly the result of an initial HSV Type I infection at an orofacial site with spread through the maxillary or mandibular division to the tri geminal nerve with a "backdoor" spread to the ophthalmic nerve and subsequent infection or later reactivation with spread centrifugally down the ophthalmic nerve to the eye (Fig 7). 47 This clinical course has been demonstrated in mice after snout inoculation 48- 51 and is a better animal model for recurrent ocular disease than direct inoculation by scarification or partial thickness trephination. Contig uous spread of locally inoculated virus also may occur and allow further extension of the disease, but persistence ofthe virus at the inoculation site and thereby the severity of clinical disease may depend crucially on the degree of secondary virus input from the nerve itself. It is not clear ifany ocular infections in humans originate from the cor neal or conjunctival epithelium except in those rare sit uations of documented primary systemic HSV with as sociated eye involvement. It also is not clear whether the neuron is damaged (as are other cells) by the productive
787
Ophthalmology
Volume 99, Number 5, May 1992
infection, and it is not clear whether the same neuron can be a productive neuron and later a latent neuron. We also are unclear ofthe nature of virus interaction with Schwann cells (in peripheral nerves), with satellite cells (in the gan glion), or with support cells (in the central nervous system).
Herpes Simplex Virus Replication On entry into the nucleus, there is circularization of the viral DNA genome. In the presence of certain host cell factors and a component of the tegument (Vmw 65), which probably arrives at the nucleus independently, there is a coordinated and temporally related cascade of gene expression (Fig 5). Thus, not all genes are expressed at the same time, and the regulation of gene expression oc curs at the level of transcription, translation, or both. 9 This temporal control of virus gene replication is achieved by positive and negative regulation of gene expression by viral proteins acting on upstream promoter-regulatory se quences and adjacent noncoding sequences. In the absence of the inducer component of the tegument protein, the cascade is halted. During a productive HSV infection in the nerve nucleus, the cascade ofgene expression involves three classes: the immediate early (alpha) genes, the early (beta) genes, and the late (gamma) gt
T
he herpes simplex and varicella-zoster viruses are members of the subfamily alpha herpesviruses with specific properties of the virion and with the capacity to establish latent infections in humans. The genome of each of these viruses has been determined with an estimate of the number of genes and proteins encoded. The biology and molecular events of the herpes simplex virus productive and latent infection have been detailed with the use of both in vitro and in vivo model systems. The neuron is the site of latency in the ganglia with a limited transcription of genes expressed during the latent period. The specific molecular regulation of latency and reactivation are not well established. There are co-cultivation, electron microscopy, and biochemical studies that support the concept of corneal latency, although this has not been proven conclusively. Details about the varicella-zoster virus biology and molecular events are not as well advanced since animal models have been lacking. The biology of the productive infection (varicella) is different from herpes simplex virus infection since the portal of entry is the respiratory system. Data support the concept of the maintenance of latency within satellite cells in the ganglia rather than within neurons. There are multiple genes expressed during this latency. These features may explain the different clinical presentations and course of reactivation (zoster) compared with herpes simplex virus reactivation. Ophthalmology 1992;99:781-799
The herpes simplex virus (HSV) had been the most in tensively studied of all viruses until the recent research on the human immunodeficiency virus. The HSV serves as a model to study synaptic connections in the nervous system, membrane structures, gene regulation, translo cation of protein, and many other biologic functions. Complex and sophisticated studies using human and an imal tissue in vitro and in vivo have begun to unravel the unique symbiosis that exists in nature between humans and the HSV. The biology and molecular aspects of the numerous stages of the host-viral interaction have been elucidated by these studies. Some of the stages explored have been the development of the primary infection at Originally received: October 14, 1991. Manuscript accepted: December 2, 1991 . Presented in part at the American Academy of Ophthalmology Annual Meeting, Anaheim, October 1991 . From the Mayo Clinic Jacksonville, 4500 San Pablo Rd, Jacksonville, FL 32224.
the epithelial site, the mode of spread to other sites with a preference for the sensory and autonomic ganglion, the host control of primary infection and the limitation of HSV from dissemination, the tissues that establish and then maintain latency, the nature of the viral genome during latency, the association of the viral genome with host cellular macromolecules, the mechanism of reacti vation of latent infection, the establishment of recurrent disease at the epithelial site, and the host control of re current epithelial disease. Animal models have served as a guide, 1 but, despite culture assays, biochemical analysis, immunofluorescent studies, and ultrastructural tech niques, we have not been able to define this host parasite relationship clearly. Because varicella-zoster virus (VZV) remains predom inantly cell-associated throughout the virus replication cycle in vitro, it has not been possible to obtain high titer infectious cell-free virus. There is also no easily susceptible animal, so model systems are not well developed. There fore, the biologic, molecular biologic, and immunologic
781
Ophthalmology
Volume 99, Number 5, May 1992
aspects of VZV pathogenesis have not advanced as rapidly as the knowledge of the HSV. Much of our knowledge of VZV has been gained from comparison and deduction from in vitro and in vivo HSV animal models. The immune mechanisms moderating HSV and VZV infection act at multiple stages and are even more complex than describing the biologic and molecular aspects of the infection. 2•3 The host response may influence the acqui sition ofdisease, the severity ofinfection, the development and maintenance oflatency, and the frequency of recur rences. The discussion here will be limited to a current working hypothesis of the biologic and molecular aspects ofHSV and VZV within the human host, albeit extending some mechanisms that have only been demonstrated in animal models or in in vitro studies. The acute ocular HSV disease appears to be comparable in the rabbit, mouse, and human. 4 Latency may not be identical since frequency of spontaneous HSV viral shedding differs among rabbits, mice, and humans. The trigger mecha nisms for reactivation also differ among species. 4 This report will focus on recent studies, but it is not meant to ignore the multiple investigators who, over the past few decades, have provided the foundation that current biol ogists are building on.
The Herpes Virus Viruses are acellular infectious particles that are incapable of metabolic activity outside living cells. Genetic infor mation is stored either in the form of RNA or DNA and can only be expressed within a living cell at the expense of that cell's own machinery and energy. The mature virus particle is known as a virion. Both HSV and VZV are DNA viruses approximately 150 to 200 nanometers in diameter and composed ofa nucleic acid and protein core, a protein capsid, a tegument layer, and a lipid bilayer envelope (Fig 1). The central nucleic acid is linear, double-
Herpesvirus
Nucleocapsid
Protein
Nucleocapsid
Figure 1. Schematic drawing of a herpesvirus. The protein of the her pesvirus is surrounded by the DNA, much like thread on a spool. A protein structure called the capsid is in the shape of an icosahedron and surrounds this DNA genome core. This combined structure is called a nucleocapsid. An additional phospholipoprotein envelope surrounds the nucleocapsid with glycoprotein spikes projecting from the surface. The tegument is an amorphous protein structure between the nucleocapsid and the envelope. The complete infectious particle is called a virion.
782
Symmetry and Rotational Axis of an Icosahedron
Figure 2. The icosahedron, showing the elements of icosahedral symmetry as viewed from the two-fold, three-fold, and five-fold axes of rotational symmetry. The icosahedron possesses 12 vertices, 20 triangular faces, and 30 edges. (Modified and published courtesy of ]oklik WK, Virology, 3d ed, chap. 2. Appleton & Lange, 1988 and Harrison SC. In: Fields BA, Knipe OM, eds. Virology, 2d ed, val. 1, chap. 3, Raven Press, 1990.)
stranded, and relatively guanine and cytosine rich, with a molecular weight of approximately 100 X 106 daltons with VZV being slightly smaller. Protein surrounds this DNA perhaps like thread (the DNA) on a spool (the pro tein). This nucleoprotein is surrounded by a protein coat, the capsid, which is made of repeated polypeptide chains. The capsid protects the nucleic acid, confers symmetry on the virus, and allows the introduction of viral genome into the host cell. The herpesvirus has a cubicle symmetry in which the surrounding capsomer building blocks form an icosahedron with a 5:3:2 symmetry, which is con structed from equilateral triangles (Fig 2). Although the virus appears spherical on electron microscopy, it is ac tually an icosahedron composed of symmetric subunits. Forming the icosahedron are 12 pentavalent capsomers present at all the vertices; 60 hexamers form the 20 faces, and 90 hexamers form the 30 edges-making a total of 162 capsomers. The herpesvirus also has an envelope, which is a complex phospholipoprotein structure derived from the cytoplasmic membrane of the host cell that has been modified by viral protein. Virus-encoded glycopro tein subunits, known as peptomers, are incorporated into the envelope and project from its outer surface. Both pep tamers and capsid proteins are antigenic and induce spe cific antibody response in the infected host. Only envel oped virions are capable of entry into cells and initiating replication. Between the capsid and the viral envelope is an amorphous protein structure called the tegument. This protein is extremely important in inducing active viral transcription in t!w nucleus and shutting off the host pro tein production. HSV Type I, HSV Type II, and VZV are members of the subfamily alpha herpesviruses, which are classified based on properties of the virion, host range, site of la tency, and the replication cycle (Table 1). Alpha viruses
Liesegang · Herpes Simplex and Varicella-zoster Virus Infections Table 1. Human Herpesvirus Virus
Subgroup
Designation
Disease
a a a
HHV1 HHV2 HHV3
'Y
HHV4 HHVS HHV6 HHV7
Oral ocular HSV Genital HSV Chickenpox Herpes zoster Infectious mononucleosis Mild infection Exanthem subitum Unknown
Herpes Simplex Virus Type I Herpes Simplex Virus Type II Varicella-zoster virus Epstein-Barr virus Human cytomegalovirus Human B-celllymphotropic virus Human herpes virus-7 HSV
=
{3 {3 {3
Guanine and Cytosine Content (mole%) 67% 69% 46% 60% 57% 40% Not known
herpes simplex virus.
have a variable host range, a short growth cycle, spread rapidly with efficient destruction ofinfected cells, and have the capacity to establish latent infection primarily in gan glia and hence are neurotropic viruses. Beta viruses are typified by a narrow host range, a long reproductive cycle in culture with slow virus spread, the enlargement of in fected cells, and the capacity to establish latency in secre tory glands and lymphoreticular cells. Gamma viruses are associated with lymphoproliferative disease in their nat ural host and have a host range restricted to the natural host. To date, seven herpesviruses whose natural hosts are humans have been identified. The human herpesvirus Type 6 causes exanthem subitum, or roseola, a common childhood disease. HHV-7 has not yet been associated with a disease. The neurotropic herpesviruses cause similar infections although they are usually easily clinically distinguishable. Primary HSV infection is localized, albeit with some sys temic manifestations. Primary VZV infection (varicella) is a disseminated infection. Primary HSV infection is spread by contact with epithelial or mucous membranes and then it gains access to sensory nerve endings. The predominant mode of spreading VZV is by the respiratory route. After local replication in the respiratory tract, the virus is disseminated by the blood and lymphatics (pri mary viremia). The virus is then taken up by cells of the
reticuloendothelial system, where it undergoes multiple cycles of replication during the remainder of the incu bation (up to 17 days incubation). A more extensive sec ondary viremia develops followed by mucosal and skin lesions (varicella). The VZV then passes centripetally from the skin and mucosal lesions to the corresponding sensory ganglion by the contiguous sensory nerve endings and sensory nerve fibers. It is possible for virus to seed ganglia by hematogenous spread. Varicella is usually uneventful, but, in newborns and immunosuppressed children and adults, complications are common. 5 HSV Types I and II are distinct but antigenically re lated. HSV Type I generally involves infection above the waist (ocular and facial) and HSV Type II below the waist. There are unique biologic properties and distinctive dif ferences between these viruses in tropism and site-specific recurrence frequency 6 (Table 2). During the primary infection, both HSV and VZV es tablish latency in sensory ganglion. HSV produces fre quently recurring localized infections with mild discom fort, whereas VZV rarely recurs more than once but is characterized by more severe pain. The lesions of HSV are usually few and limited in distribution whereas zoster spreads to involve much of the cutaneous dermatome. These differential features suggest that the nature of the acute and latent infection in sensory ganglia differ and
Table 2. Herpes Simplex Virus Infection Ocular HSV Orofacial HSV Genital HSV Neonatal HSV Disseminated HSV Encephalitis HSV Meningoencephalitis HSV HSV
=
Predominant Virus Type 1 1> 2> 2> 1> 1 2
2 1 1 2
Outcome
Recurrence
Visual impairment Resolution Resolution Retardation Resolution or death Neurologic damage or death Resolution
Yes Yes Yes No No No No
herpes simplex virus.
783
Volume 99, Number 5, May 1992
Ophthalmology
the mechanisms of reactivation from latency may also differ.
The Viral Genomes The HSV genome is approximately 152,000 base pairs in size and codes for approximately 72 genes and 70 different proteins. 7 The VZV genome is approximately 125,000 base pairs in size and codes for approximately 68 genes and 80 proteins. 8 The HSV genome is composed of two portions: a unique long region and a unique short region with gene sequences that occur only once. Both these long and short unique sequences are flanked by repeat se quences that occur at the termini and internally in the genome (Fig 3). A hinge point exists between the internal repeat sequences such that each unique sequence can flip flop relative to the other, generating four possible isomeric forms of the complete DNA molecule (Fig 4). Thus, viral genomes extracted from infected cells consist of four equimolar isomers that differ solely in the orientation of the long and short components. 9 The biologic functions of these inverted repeats are unknown, although deletion of small portions does not effect the ability to replicate or establish latency. The characteristic genome anatomy with two longer unique components, both bordered by inverted repeat units, makes it possible for the genome to circu larize, which may be essential for viral DNA synthesis and for the establishment oflatency. 10 HSV I and II have approximately half the genome sequences homologous. There is one major exception with the insertion of some 1500 base pairs into the coding sequence in the HSV Type II unique short section. The genes are oriented in both directions. 7 The VZV genome also is composed of two portions: a unique long region of approximately 105,000 base pairs and a unique short region of5200 base pairs (Fig 3). Each of these pairs is flanked by inverted repeat elements of Isomeric forms, no.
Genomes
-
HSV 1, 2
vz:v 0
- 50
Molecular
4
2 100
wt, x 10 6
Figure 3. Schematic drawing of the herpes simplex virus and varicella zoster virus genomes. Each of these viral genomes consists of long and short unique sequences (Ul, Us) flanked by long stretches of bases that are duplicated as inverted repeat elements. These are designated as ter minal or internal repeats surrounding the long and short unique sequences (TRl, TRs, IRl, IRs, respectively). Portions of the genomes that are actively transcribed in latently infected ganglia are shown as heavy arrows below the genome maps. The relative molecular weights are shown as are the number of isomeric forms.
784
Figure 4. The isomers of HSV and VZV. The unique long and unique short regions (Ul, Us) of HSV can both invert at a hinge point in the internal repeat unit resulting in four isomers that occur with equal fre quency during an infection. The unique short region (Us) of VZV also can invert, giving two major isomers. Only about 5% of VZV virions have an inverted unique long section (Ul) resulting in two minor isomers since the Us region can also invert in either direction. The internal repeat regions are very important in allowing inversion, circularization of the genome in the nucleus, intragenomic recombination, and permitting the rolling circle mechanism of DNA production.
shorter base pairs (8000 base pairs). During DNA repli cation, the unique short region (and its repeats) inverts, producing two isomeric forms of VZV DNA. Approxi mately half of the viral particles contain the unique short region in one orientation and half in the other. Further studies have shown that a small proportion of the genome (perhaps 5%) also have the unique long region inverted (Fig 4 ). 11 Thus, VZV virus contains four genome isomers (two major and two minor). This contrasts with HSV in which both segments (unique short and unique long) in vert with equal frequency and result in four isomeric forms. All these VZV and HSV isomers are equally in fectious. The VZV and HSV genomes share several features: two unique regions (unique short and unique long) flanked by an inverted repeat and four genome isomers of virion DNA, albeit in different proportions. 8 The unique short and the repeat units (terminal repeat long, internal repeat long) are shorter in VZV, and the HSV genome also is terminally redundant with a sequence of 200 base pairs, which is repeated directly at the genome ends and is also present in the inverse orientation at the junction between the internal repeat long and the internal repeat short. The guanine and cytosine content in VZV is 46% compared with 67% for HSV I and 69% for HSV II. The complete genome sequence for HSV I and VZV have been determined using bacteriophage cloning and chain termination sequencing techniques as well as re striction endonuclease analysis. 8 • 12 The VZV genome is not a unique size but may vary between limits of ap proximately 124,000 to 126,000 base pairs. 8 The avail ability of DNA sequencing, the recognition of open read ing frames and coding proteins, and other transcriptional regulatory sequences, coupled with computer analysis, have permitted a much greater understanding of the re lationship between herpesviruses and allowed comparative study of individual genes. The genetic relationship and similar DNA arrangements between VZV and HSV Type
Liesegang · Herpes Simplex and Varicella-zoster Virus Infections I have allowed an assumption of the VZV gene functions based on the more extensive knowledge of the HSV gene function. 7 • 12 The localization of these coding sites for pro tein may be important because they are the targets for antiviral therapy. There are also unique reiterated ele ments of the genome, which have proven useful in in vestigating the epidemiology of VZV and HSV. Both VZV and HSV Type I have similar core genes. VZV has some genes that have no HSV counterpart and vice versa. Some regions of the genome exhibit relative rearrangement of genes. 7 A schema for the evolution of VZV and HSV Type I from a common ancestor by a series ofrecombinational events resulting in the expansion or contraction of terminal repeat short and internal repeat short segments has been suggested. 8 During the divergence of HSV-1 and VZV, very extensive processes of incor poration of mutations have evidently taken place to the degree that VZV has an overall genome base composition of 46% compared with 68.3% guanine plus cytosine of HSV 1. It is probably these genetic differences that play a major role in the disease pattern. Unfortunately, the lack of similar animal models of VZV have hindered proving several hypotheses. Both VZV and HSV Type I also have a close relationship to the Epstein-Barr virus. 13 The gene locations have been extensively studied in herpesviruses with special emphasis placed on defining the major glycoprotein families that are important in im mune recognition and perhaps future vaccines. The gly coproteins may be especially important in pathogenesis because they become incorporated into the envelope of the virus as well as brand the host cell and are involved in the first interaction between the virus and the cells that it can infect. The origin of replication of DNA has been mapped to a homologous sequence in both HSV and VZV. HSV Type I contains two origins of DNA repli cation, whereas VZV has one origin of replication. 11 There have been seven genes that appear essential for HSV DNA replication with each of these having a homologue in VZV. 7 Restriction-endonuclease analysis of viral DNA recovered from clinical isolates has proven that latent varicella virus reactivates to cause zoster. 14 Different viral isolates from varicella or zoster patients or patients with varicella with subsequent zoster in the same individual show slight variation in the mobility of certain DNA frag ments. Isolations from a single outbreak of varicella, however, have a stable genome structure. HSV demon strates the same basic genome stability. In laboratory animals, diverse clinical outcomes result from different HSV viral types and strains. These reactions are complex to assign but may depend on viral genetic constitution, 15 • 16 on the viral challenge dose, or on the host's genetically determined immune response. 17- 20 Viral strains differ in regard to neuroinvasiveness and neuro virulence, terms that are related but not identical. Neu rovirulence is the viral capacity to replicate and cause disease in the nervous system after direct inoculation into these tissues. Neuroinvasiveness relates to the capacity of the viruses to enter and move through the nervous system after inoculation at some peripheral site. There have been several investigations mapping HSV gene functions as
sociated with significant biologic effects and studies of the properties of the products encoded by these genes. 21 The HSV-I strain characterized by neuroinvasiveness has been linked to encephalitis. Gene mutations that occur in tem perature-sensitive mutants of HSV Type I can affect the penetration, the formation of functional capsids, and viral structural protein synthesis. 22 The specific nuclear base sequence of DNA polymerase genes can contribute to dif ferences in the capacity of HSV Type I and Type II to replicate in the trigeminal ganglion and spread from the eye to the central nervous system. 23 •24 Identifiable and reproducible ocular lesions of distinctly different size, lo cation, and shape can result with different strains of HSV in animals. 15 • 16•25 In animal models, some HSV strains produce consistently mild epithelial disease, some produce epithelial and stromal disease, and others produce uveitis. Strains of HSV that produce more severe stromal disease have greater amounts of envelope glycoprotein-D than other strains that produce less severe stromal disease. 26- 29 Steroid response to treatment of HSV also is genetically regulated. 30 The ability to reactivate the virus is strain specific in the rabbit and can be separated into sensitivity to endogenous (spontaneous) reactivation and exogenous (induced) reactivationY HSV thymidine kinase activity is important in determining the severity of keratitis, the extent of ganglionitis and encephalitis, and is important in promoting latency and reactivation from latency, al though it is not important in establishing latency. 10•32 Mutants with thymidine kinase deficiency have remark ably different patterns of disease in experimental mice. 33 The specific DNA map units responsible for many of these features have been determined. 27 •34 Caution is advised, however, in overinterpreting animal data because the nat ural infection in humans involves complex cellular inter actions that may have more relevance than some of these biologic features in animals.
Primary Herpes Simplex Virus Infection The natural history of HSV Type I infection in humans is generally characterized by a childhood infection that may present with a nonspecific upper respiratory infection or be asymptomatic. On rare occasions, primary HSV Type I infection may be associated with ocular tissues and present as a bilateral vesicular lid edema, ulcerative blepharitis, or epithelial keratitis. These cells are ultimately lysed. The primary infection defines the pathway of viral spread and the site of viral latency. 35 Primary infection in humans occurs most commonly within the body surface innervated by the trigeminal nerve (i.e., the mucocuta neous areas of the face, eye, and oral mucosa). HSV can be spread through the nasal passages to the olfactory sys tem of the central nervous system and also through the esophagogastric mucosa to the ganglion ofthe vagus nerve. It also has been implicated as a cause of recurrent peptic ulcer. 36 Autopsy studies reveal that nearly all patients in fected with HSV house DNA in selected nerve ganglia. The trigeminal ganglion are the most commonly infected site followed by the sacral root ganglia.
785
Ophthalmology
Volume 99, Number 5, May 1992
Although primary infection (when apparent) tends to cause more severe symptoms than recurrent infection, it is self-limited with complete recovery and disappearance ofHSV from the initial site of infection. In immunocom promised individuals and newborns, HSV can spread from the primary replicating site to other organs and may be fatal. 37 Primary HSV Type I infection involves primarily ectodermally derived tissue (i.e., skin, mucous mem branes, or corneal epithelium). Initial HSV Type II also effects ectodermally derived tissue, either from contact with the birth canal of an infected mother or later in life through contact with sexual partners. These mucocuta neous sites are unique in several aspects: they are con nected to the outside world, they are hormone sensitive, have a rich blood supply, and have a rich innervation by somatic, sensory, or autonomic nerves. The primary infection in laboratory animals involves an initial viral replication in mucocutaneous tissues. Vi ruses infect cells ifthey have an appropriate receptor mol ecule on their outer surface, and different receptors are probably recognized by different viruses. The wide range of hosts suggests that HSV receptors on cell surfaces are widely distributed. After a productive replication in the mucocutaneous site with spread of progeny virus to con tiguous cells, there is extension to the mucocutaneous projections of the sensory nerve probably within hours. 38 A receptor for HSV attachment to the·nerve termini may be structurally similar to heparan sulfate, a common con stituent of plasma membranes. 10 The herpes virion at taches to the receptor with its surface glycoproteins an choring the virus with a fusion reaction of the viral en velope with the plasma membrane of the neuritic extension of the sensory nerve. 39 ·40 The viral glycoprotein spikes on the herpes simplex virus envelope determine both the attachment and the fusion activity by which the virus is internalized. This fusion results in the release of naked viral nucleocapsids within the axoplasm 41 (Fig 5). After penetration by fusion, the naked viral nucleo capsids are transported by microtubular dependent ret rograde axonal flow or by membrane-bound organelles to the cell nucleus, at about the rate offast axonal transport of protein, which is 5 to 10 mm per hour. 40-42 As part of normal function, neurons make a variety of products that are packaged and transported along the axon. It appears that microtubule gliding and movement of organelles are mechanisms for HSV axonal transport toward the nu cleus.43 Because the virus has lost its envelope, it is non infectious and probably unable to leave the axon. There is probably loss of the tegument material that surrounds the nucleocapsid during transport, but otherwise it seems essential that the nucleocapsid be structurally maintained until it reaches the nucleus. Once inside the neuron, the noninfectious forms must undergo at least one cycle of replication to account for the infectious (enveloped) forms detected in neurons in the next phase of infection (i.e., when the infectious virion migrates centrifugally from the ganglion to the inoculation site). The virus replication at the inoculation site causes the development of the characteristic lesion. 44 Virus can be transported centrally to reach the ganglion nerve root and may leave the neuron through membranes
786
late
~
5) Assembly
Figure 5. Schematic replication ofHSV in a susceptible cell. 1) The virus fuses with the plasma membrane after attachment at specific receptor sites which recognize HSV Type I. The empty envelope is left at the plasma membrane. 2) The nucleocapsid is introduced into the cytoplasm and then transported to the nuclear pore by fast axoplasmic flow. Two proteins are released from the tegument of the virion. One shuts off host protein synthesis and the other is transported to the nucleus as a trans inducing factor (Vmw65). 3) The viral DNA is released into the nucleus and becomes circularized. The empty capsid coat is left at the nuclear pore. 4) In the presence of the tegument transinducing factor, the tran scription of the immediate early genes occurs with transport of mRNA to the cytoplasm and translation to protein products. These products induce the transcription of early genes with transport of mRNA to the cytoplasm and translation to beta proteins which are involved in DNA synthesis. DNA synthesis occurs by a rolling circle mechanism that yields viral DNA. Transcription of gamma genes results in gamma proteins which consist of structural proteins of the virus. 5) The capsid proteins are constructed into complex icosahedron structures, which are packaged with viral DNA cleaved from the rolling DNA concatamers. Viral gly coprotein and tegument protein accumulate and alter the internal nuclear membrane. The nucleocapsid is probably enveloped as it passes through the internal nuclear membrane, de-enveloped between the inner and outer nuclear membrane, and then leaves the nucleus. 6) The virus is transported to the endoplasmic reticulum and Golgi apparatus where it is enveloped and then transported via vesicles to the nerve periphery and released by exocytosis into the extracellular space or to contiguous cells. The glycoprotein spikes of the envelope are matured as the virus travels away from the nucleus (Modified from the concepts of Roizman B, Sears AE 9 ).
to infect astrocytes and oligodendroglia cells with subse quent damage and infection of other axons within the same nerve root giving a "backdoor" into other divisions of the ganglia. Myelinated nerves around Schwann cells may act as a physical barrier, but, in nonmyelinated fibers, there is no physical barrier preventing passage of virus from the axon to the Schwann cell cytoplasm. Adjacent myelinated and nonmyelinated Schwann cells can be in fected and enveloped virus may be transmitted by this mechanism the full length of the peripheral nerve within a few days. Other hypotheses for the mechanism of viral transport from the peripheral site include transport in the periaxonal space, by sequestered infection of Schwann cells, or the lymphatic channels of the epineurium, but all are unlikely. 45 When nucleocapsids have reached the membrane of the nucleus, the viral DNA is released into the nucleus, probably through the nuclear pores. The empty capsids are left outside the nuclear membrane.
Liesegang · Herpes Simplex and Varicella-zoster Virus Infections On entry into the nucleus, there is circularization of the viral DNA genome and initiation of viral replication. During the active infection, the viral nucleocapsids are assembled in the cell nucleus and then enveloped at the inner lamina of the nuclear membrane, de-enveloped on passage of the outer nuclear membrane, and leave the outer nuclear membrane as naked nucleocapsids. A sec ond and final envelopment of the virus occurs at the membranes ofthe endoplasmic reticulum, where the virus also receives a further enclosure in the transport vesicle. The vesicle is important for protection, transportation, fusion with the plasma membrane, and for the maturation of the glycoprotein spikes. The enveloped virus is trans ported along the nerve in transport vesicles probably de rived from both endoplasm reticulum and Golgi apparatus ultimately to arrive at the nerve endings. The glycoprotein spikes of the envelope that were acquired at the inner nuclear membrane become more mature when the virus passes through the endoplasmic reticulum and the Golgi apparatus. When it reaches the nerve endings, the plasma membrane of the nerve is changed morphologically and chemically during its fusion with the virus transport ves icle, and the virus is released probably by exocytosis after envelopment at the plasma membrane and by fusion of the vesicle carrying enveloped virus at the plasma mem brane.40 The host plasma membrane is branded with virus specific glycoproteins, which makes it a target for im munologic attack. The route of envelopment and release of herpesvirus is similar to the route of synthesis, pro cessing, and release of secretory glycoproteins. In the eye, the virus released from the corneal nerve termini may infect the corneal and conjunctival epithelium, and re establish active HSV infection. The duration of reactivated infection is usually shorter because of previous host cel lular and humoral anti-HSV response. In the eye, how-
The HSV "Round Trip"
Figure 6. The "round trip" hypothesis of HSV infection. The initial replication of HSV at the inoculation site may not be essential in causing clinical disease. If the virus has access to nerve endings, it can travel as a naked nucleocapsid to the nucleus where it can undergo a replicative cycle and then travel down the nerve as an enveloped mature form. Release can then either cause the clinical lesion or at least contribute to the progression of the clinical lesion. This may explain the incubation period for clinical disease and is compatible with the "backdoor" spread ofHSV.
"Back Door"
lnfection~wt
h HSV
Trigeminal ganglion Ophthalmic branch
f.---=-=
Maxillary branch
Mandibular branch
Figure 7. The "backdoor" approach to the ophthalmic nerve causing ocular herpes. The most common site of initial HSV infection is in the facial area served by the maxillary division of the trigeminal nerve. In· fection here can result in latency in the maxillary portion of the trigeminal ganglion. At the time of initial infection or with reactivation, virus can spread to the ophthalmic or mandibular portion of the trigeminal nerve and can perhaps cause HSV infection in the ophthalmic division of the trigeminal nerve without ever having had a prior skin or mucous mem· brane HSV infection in this distribution.
ever, repeated episodes may enhance the immune-in flammatory reaction causing structural alterations in the cornea. The extracellular release of virus from the peripheral sensory nerve now allows further infection of the der matome and also allows subsequent spread to satellite or support cells around the nerve. Indeed, the initial mul tiplication of virus at the inoculation site may not be es sential in causing clinical disease or even in establishing latency if the virus has access to the nerve endings. The virus travels up thy axon as an unenveloped particle and travels centrifugally as an enveloped mature form. The delay from time of inoculation to the development of clinical lesions (5 to 8 days) may be explained by this "round trip" theory with the clinical disease actually re sulting from viral release from nerve endings (Fig 6). 44·46 As a further extension of this hypothesis, HSV infection in the eye is probably more commonly the result of an initial HSV Type I infection at an orofacial site with spread through the maxillary or mandibular division to the tri geminal nerve with a "backdoor" spread to the ophthalmic nerve and subsequent infection or later reactivation with spread centrifugally down the ophthalmic nerve to the eye (Fig 7). 47 This clinical course has been demonstrated in mice after snout inoculation 48- 51 and is a better animal model for recurrent ocular disease than direct inoculation by scarification or partial thickness trephination. Contig uous spread of locally inoculated virus also may occur and allow further extension of the disease, but persistence ofthe virus at the inoculation site and thereby the severity of clinical disease may depend crucially on the degree of secondary virus input from the nerve itself. It is not clear ifany ocular infections in humans originate from the cor neal or conjunctival epithelium except in those rare sit uations of documented primary systemic HSV with as sociated eye involvement. It also is not clear whether the neuron is damaged (as are other cells) by the productive
787
Ophthalmology
Volume 99, Number 5, May 1992
infection, and it is not clear whether the same neuron can be a productive neuron and later a latent neuron. We also are unclear ofthe nature of virus interaction with Schwann cells (in peripheral nerves), with satellite cells (in the gan glion), or with support cells (in the central nervous system).
Herpes Simplex Virus Replication On entry into the nucleus, there is circularization of the viral DNA genome. In the presence of certain host cell factors and a component of the tegument (Vmw 65), which probably arrives at the nucleus independently, there is a coordinated and temporally related cascade of gene expression (Fig 5). Thus, not all genes are expressed at the same time, and the regulation of gene expression oc curs at the level of transcription, translation, or both. 9 This temporal control of virus gene replication is achieved by positive and negative regulation of gene expression by viral proteins acting on upstream promoter-regulatory se quences and adjacent noncoding sequences. In the absence of the inducer component of the tegument protein, the cascade is halted. During a productive HSV infection in the nerve nucleus, the cascade ofgene expression involves three classes: the immediate early (alpha) genes, the early (beta) genes, and the late (gamma) gt
