Of Mice and not Humans: How Reliable are Animal Models for Evaluation of Herpes CD8⁺-T cell-Epitopes-Based Immunotherapeutic Vaccine Candidates?

Abstract

Herpes simplex virus type 1 and type 2 (HSV-1 and HSV-2) -specific CD8⁺ T cells that reside in sensory ganglia, appears to control recurrent herpetic disease by aborting or reducing spontaneous and sporadic reactivations of latent virus. A reliable animal model is the ultimate key factor to test the efficacy of therapeutic vaccines that boost the level and the quality of sensory ganglia-resident CD8⁺ T cells against spontaneous herpes reactivation from sensory neurons, yet its relevance has been often overlooked. Herpes vaccinologists are hesitant about using mouse as a model in pre-clinical development of therapeutic vaccines because they do not adequately mimic spontaneous viral shedding or recurrent symptomatic diseases, as occurs in human. Alternatives to mouse models are rabbits and guinea pigs in which reactivation arise spontaneously with clinical features relevant to human disease. However, while rabbits and guinea pigs develop spontaneous HSV reactivation and recurrent ocular and genital disease none of them can mount CD8⁺ T cell responses specific to Human Leukocyte Antigen- (HLA-) restricted epitopes. In this review, we discuss the advantages and limitations of these animal models and describe a novel “humanized” HLA transgenic rabbit, which shows spontaneous HSV-1 reactivation, recurrent ocular disease and mounts CD8⁺ T cell responses to HLA-restricted epitopes. Adequate investments are needed to develop reliable preclinical animal models, such as HLA class I and class II double transgenic rabbits and guinea pigs to balance the ethical and financial concerns associated with the rising number of unsuccessful clinical trials for therapeutic vaccine formulations tested in unreliable mouse models.

1. Introduction

Herpes simplex viruses (HSV-1 & HSV-2) are nearly ubiquitous pathogens that cause fatal disseminated disease in newborns to significant morbidity and mortality in adults ranging from skin lesion (cold sores), genital ulcerations, blinding eye keratitis, and encephalitis in immunocompromised individuals [1-4]. Most importantly, HSV-2 infected individuals are at higher risk (4 fold) of HIV-1 acquisition than HSV-2 non-infected individuals [5]. HSV-1 & HSV-2 establish latency in sensory ganglia followed by sporadic and spontaneous reactivations and recurrent shedding in mucocutaneous surfaces [3, 6-10]. The natural history of HSV-1 and HSV-2 infections is illustrated in Fig. 1. At present no FDA approved therapeutic vaccines are available for herpes. Although the standard antiviral drug regimens (e.g. acyclovir) reduce/suppress recurrent symptomatic disease, asymptomatic shedding, and lateral transmission to some extent; they do not clear the infection or stop recurrent disease [3, 4, 11-13].

Fig. 1. The natural history of genital, oro-facial and ocular herpes infections.

Left: (1) While most cases of genital herpes are caused by HSV-2, reports of HSV-1 genital infection are increasing. The virus is transmitted through the genital tract (GT) proliferate locally in this primary site of infection, enters the sensory nerves and travels along nerves to the lumbosacral dorsal root ganglia (sacral ganglia or SG) that innervate the GT, where it establishes a latent infection. (2). Sporadic spontaneous reactivation of HSV-2 from latently infected neurons leads to viral shedding in GT which can cause symptomatic recurrent genital herpes, one of the most prevalent sexually transmitted disease. Right: HSV-1 is transmitted through eyes, lips, mouth and cause orofacial and ocular herpes. (3) Ocular herpes is mainly caused by HSV-1, which infects the cornea and then establishes latency in sensory neurons of the trigeminal ganglia (TG). (4) Sporadic spontaneous reactivation of HSV-1 from latently infected neurons leads to viral shedding in saliva and tears which can cause symptomatic recurrent Herpes Stromal Keratitis (HSK), a blinding corneal disease.

Reliable animal models of herpes infection and disease are the key to the development of an effective therapeutic vaccine against HSV-1 and HSV-2. Because of the obvious ethical and practical considerations in assessing candidate herpes therapeutic vaccine directly in humans, a critical question is: which species would be the most appropriate animal model to investigate the protective efficacy of the growing numbers of therapeutic vaccine candidates? For most immunologists, mouse is the preferred model due to the availability of (i) unlimited inbred and transgenic strains, (ii) specific immune molecule knockout strains, and (iii) the well-characterized immunological probes to study the immune response to specific therapies. In that perspective, the majority of preclinical studies are evaluating HSV vaccine efficiencies on mouse models. Unfortunately, in contrast to humans, recurrent HSV shedding and recurrent herpetic disease does not occur in mice because HSV spontaneous reactivation is either extremely rare or does not occur in mice [14]. Therefore, although mouse studies have provided ample crucial information regarding immune response against primary HSV-1 and HSV-2 infections[15-19], the efficacy of human epitope-based therapeutic vaccines against recurrent shedding and disease cannot be assessed in mice. Several clinical trials have been carried out based on mice data, but almost none led to clinical success. HSV infection in the murine model preclude its use to study horizontal or vertical transmission or to evaluate strategies designed to prevent reactivation. The lack of an accurate animal model that translates human immune responses and diseases, certainly limits the proper preclinical assessment.

More than two third of the world’s population is infected with HSV-1 & HSV-2, and many people suffer from psychological problems related to herpes [20] even though multiple therapeutic approaches including topical and/or oral administration of anti viral drugs have been available for more than 75 years. While, a majority of human population is infected with HSV-1 and/or HSV-2, more than half of them do not show any clinical symptoms (asymptomatic individuals). Thus, the immune systems of those individuals are likely capable of controlling the clinical symptoms, but cannot prevent viral shedding. Therefore, a therapeutic HSV vaccine, which can boost the immune system of both asymptomatic and symptomatic individuals to decrease the viral shedding, transmission and disease, is highly desirable. In an ideal herpes animal model: (i) the infection would be initiated via a mucocutaneous route similar to that by which humans are commonly infected, i.e., inoculation of the ocular, oro-facial or genital mucosal epithelium; (ii) a small proportion of the animals would develop spontaneous disease symptoms similar in both pathology and severity to those seen in the minority of symptomatic humans, (iii) a large proportion of the animals, would develop immune responses after infection that would protect the animal from the disease, similar to those seen in the majority of asymptomatic humans. In these “asymptomatic” animals the immune response can be scrutinized to determine which aspects are important for protection, including innate and adaptive immune responses. The “symptomatic” animals, which cannot spontaneously control the pathogen, provide an opportunity for the investigation of potential protective therapeutic immunization. In this review, we will talk about the advantages and limitations of different animal models for herpes infection and disease that has been used in preclinical development of therapeutic vaccines. We will also provide a summary with key issues involved in the pre-clinical selection of new vaccines for clinical trials.

2. CD8⁺ T cells decrease HSV reactivation

Increasing neutralizing antibody titers in herpes clinical vaccine trials did not reduce viral reactivation; shedding and recurrent disease suggesting the necessity of cellular immunity [21-23]. (1) In mice, CD8⁺ T cells accumulate in sensory ganglia from 7-10 days following herpes infection and become the predominant T-cell type during latency [24]. HSV-specific CD8⁺ T-cells producing IFN-γ and GrB appear to suppress (or abort) induced viral reactivation in explanted mouse sensory ganglia [24, 25]. It has been proposed that CD8⁺ T-cells may similarly reduce detectable HSV reactivation in vivo [26-29]. If correct, a human therapeutic vaccine that increases HSV-specific IFN-γ(+) and GrB+ CD8⁺ T cells in latently infected sensory ganglia should significantly decrease HSV spontaneous reactivation (as measured by shedding in tears) and reduce recurrent eye disease, which is dependent on virus reactivation. (2) Significant numbers of activated CD8⁺ T cells producing IFN-γ and TNF-α, were found in latently infected TG of human cadavers indicating the existence of an antigen-driven immune response [30-32]. While HSV-specific CD8⁺ T-cells appear to be the key factor to protect spontaneous reactivation and recurrent herpes, a challenge for herpes immunologists is: “To find an appropriate preclinical animal model to study the efficacy of human HLA-restricted T cell epitopes based vaccine to decrease recurrent herpes disease and spontaneous viral reactivation?” The ideal animal model should able to produce an immune response specific to human HSV epitopes (such as HLA-A*0201-restricted epitopes), while mimicking all aspects of viral pathogenesis, neuropathology, neuro-invasiveness, and latency that occur in human.

3. The mouse model lacks spontaneous HSV-1 and HSV-2 reactivation

The mouse is the premier mammalian system used for modeling human disease and understanding the immune mechanism of specific genes in normal and diseased states. However, in the case of HSV-1 and HSV-2, sporadic spontaneous reactivation and recurrent disease do not occur in mice. During primary herpes infection, the virus: (i) infects mucocutaneous surfaces of skin, mouth, genital tract, and eyes, (ii) invades the local sensory nerves by propagating via neurons, and (iii) establishes life-long latency in the neuron bodies of sensory ganglia (sacral ganglia for genital herpes and trigeminal ganglia for ocular and oral herpes). Factors such as hormonal changes, stress, and ultraviolet (UV) exposure trigger virus reactivation. The reactivated viruses travel back to re-infect the primary site of infection (i.e. the skin, the mouth, the genital tract or the eyes). A representative diagram of HSV infection, development of latency and the subsequent viral reactivation are shown in Fig. 2. A good example of recurrent herpes disease is Herpes Stromal Keratitis (HSK), an inflammatory disease of the cornea, which may result in corneal opacity and blindness. HSK in human occurs long after the primary infection, when reactivated virus returns to the eye from latency. The destructive consequences of ocular HSV-1 infection appear to be immune related [33-42]. The immune system of the infected individual but not the virus causes the majority of corneal damages [43].

Fig. 2.

A simple step-by-step representation showing the route of HSV attack (A), establishment of latency (A), spontaneous reactivation (B) and hormone induced reactivation (C). Panel B shows the “spontaneous reactivation” of HSV from latency and the path of anterograde transport from nerve ganglia to primary site of infection. Panel C shows the hormone induced “artificial reactivation” of HSV from latency and the path of anterograde transport from nerve ganglia to primary site of infection. Spontaneous HSV reactivation in mice is a rare event. Statistically, about 1/100 eyes or less shows “spontaneous reactivation” upon a period of extended sampling. Also <1% of TG have infectious virus but too low for useful analysis. Using RT-PCR 1 neuron/10 TG has lytic transcripts detectable.

HSV-1 infection in mice eye can lead to a localized infection followed by infection of sensory neurons of the trigeminal ganglia (TG) and virus spread in the central nervous system (CNS) and ultimately to the brain. Mice survived from lethal HSV-1 infection, maintain latency mainly in the sensory nerve of TG. During this latent infection period, no infectious virus can be recovered from the nerve tissue but if the nerve tissues are explanted and maintained on “feeder layer” of cultured cells, virus will eventually appear to replicate. This observation indicates that the viral genome remains intact during latency. However, evidence of spontaneous reactivation of HSV-1 does not occur in mice [14] (Fig. 2c). In 2002 Feldman et al and in 2006 Margolis et al showed the episode of HSV-1 spontaneous “molecular reactivation” from latency in mice TG [44, 45]. However, the amount of infectious virus they detected in latently infected TG was at the lower limit of the sensitivity. Statistically, about 1/100 TG or less shows “spontaneous reactivation” upon a period of extended sampling and RT-PCR able to detect only 1 neuron/10 TG with lytic transcripts [45-48]. It is possible that the mouse model is deficient in effective transport of the virus to the periphery. Besides that, the systemic immune responses (passively transferred neutralizing antibody, adoptively transferred CD8⁺ T cells) do protect mouse eye, but not human eye, against ocular HSV-1 disease. Therefore, while the mice model is quite useful to study the genetic parameters during latent infection, the physiological process of reactivation cannot be effectively studied in mice. Hence, mice cannot be used to assess the efficacy of candidate therapeutic vaccines against spontaneous virus reactivation. The “adaptation” of the “defense mechanism” and “pathogenic mechanism” of herpes virus in mouse models are more effective than human. For instance, it has been proposed that immune responses to ocular herpes infection in susceptible 129 strain mice lead to central nervous system immunopathology and fatal encephalitis [49-51]. Whether, the mechanism reported in these models applies to humans is yet to be proven.

A hallmark statement made by a herpes immunologist, that the validity of HSK mouse models “relies on the assumption (however great) that similar events occur in the human condition” [52]. In mice, HSK is a one-time event, which initiates within ~14-21 days of the primary ocular infection. Initial ocular infection is followed by leukocyte infiltration, virus clearance, and a second phase of leukocyte infiltration and chronic inflammation in the absence of live virus [19, 53]. Thus mouse HSK is a (delayed) response to the acute (primary) ocular infection rather than a recurrent disease due to return of reactivated virus from the TG. In contrast, human HSK can recur numerous times over decades, associated with reactivation of HSV-1 from latency, and thus appears to be a truly recurrent disease. In addition, mouse T cells do not recognize human leukocyte antigen- (HLA-) restricted epitopes [27] and therefore the immunogenicity and protective efficacy of human T cell epitope-based vaccine and immunotherapy cannot be assessed in normal strains of mice. This problem can be solved by using Human leukocyte antigen (HLA) transgenic mice that develop T-cell responses against HLA-restricted T cell epitopes [15-19]. In fact, we have recently used HLA transgenic mice to study the immunogenicity and protective efficacy of several HLA-A*020-restricted epitopes against primary HSV-1 infection and disease [17, 54]. Nonetheless, the HLA transgenic mice model still do not resemble the human scenario and cannot be used to assess therapeutic efficacy of HLA-restricted human T cell epitopes.

Additional disadvantages of mouse model are: (i) Disparity in appearance of mouse vs. human disease by clinical exam; (ii) HSV-1 Ag is not detected in mouse cornea during HSK whereas viral Ag has been detected in 74% of human HSK cornea although the specific Ag(s) remain unknown. The detection of viral Ag is higher in those corneas that have current active episodes of HSK [55]; (iii) In mice, HSK can only be induced by molecular mimicry [56-58] but in human it is a natural event [19]. These differences make mouse a non-reliable experimental model to evaluate and relate to human HSK. The severity of HSV-1 corneal disease in mouse model is affected by (i) host strain, (ii) virus strain and (iii) host immune system. It is well documented that strains of inbred mice differ in their susceptibility to corneal HSV1 infection [59, 60]. For example, while C57BL6 mice being most resistant, DBA/2 mice being most susceptible, and BALB/C mice being intermediate. This order of resistance matches the resistance to latency and susceptibility to encephalitis [61, 62]. Thus, although mouse studies have provided plenty of useful and important information regarding ocular HSV-1 infection and despite the tremendous amount of knowledge gained about mouse immunology in general, the mouse is not an ideal model to extrapolate the mechanisms of recurrent-HSK in humans.

The mouse genital infection model is derived from Parr et al. [63-67] where female mice are treated with medroxyprogesterone to thin the genital epithelium lining and make the mice uniformly susceptible to HSV-2 vaginal infection. However, spontaneous reactivation of the virus from sensory ganglia and shedding of the virus in genital tract do not seem to occur in mice (as opposed to guinea pigs and humans) [68]. Therefore, the potential effect of vaccination against HSV recurrences in mice can only be inferred from the levels of latent viral DNA within neurons [69], since these levels have been shown to correlate with the number of ex vivo recurrence rates in mice [70-74].

4. Rabbit model closely mimics human ocular HSV-1 infection and disease

In contrast to mice, rabbits are relatively better choice for studying ocular HSV-1 infection, disease and immunity. Similar to humans, rabbit can develop primary ocular disease, undergoes latent infection and establishes life-long latency in the neuron of TG (Fig. 2A & B). Rabbits, HSV-1 intermittently reactivates from latently infected primary sensory neurons, leads to peripheral shedding of infectious virus and recurrent corneal disease similar to humans [75]. Increasing evidence suggests that in most humans, HSV sheds spontaneously (~35% of the time for HSV-1 and 15% of the time for HSV-2) [76-80]. In rabbit ocular model, HSV-1 sheds through tears both spontaneously and in response to local or systemic stimuli [81-83]. Certainly, this pattern of viral reactivation in rabbit closely resembles that in humans. In addition, most HSV-1 strains cause acute ocular infection, latency in TG and spontaneous shedding in rabbit. Many researchers use rabbit as a recurrent model of HSV1 infection and eye disease bearing in mind that: (i) numerous similarities exist between rabbit and human ocular mucosal immune system [84-86]; (ii) several T-cell-mediated ocular diseases, including herpetic conjunctivitis and recurrent corneal herpetic stromal keratitis (HSK), have been reported to be similar in rabbits and humans [87-89]; (iii) compared to mice, rabbit conjunctiva associate lymphoid tissue (CALT) resembles human CALT [90-93]. Microanatomy and immuno-histological studies indicate that rabbit conjunctival mucosa is comparable to that of humans and has a typical follicular ultra-structure with an abundance of “conjunctival lymphoid follicles” (CLF), whereas no lymphoid tissue was identified in mice [93-97]; (iv) from a practical standpoint, rabbits possess a relatively large cornea and conjunctival surface offering abundant mucosal associated lymphoid tissue for in vitro studies; (v) finally, the recent availability of many monoclonal and polyclonal antibodies specific to rabbit immune cell CD markers, cytokines and growth factors provide useful immunological tools for an unprecedented phenotypic and functional analysis of rabbit T cell repertoire and function.

5. Circulating HSV-specific CD8⁺ T cells do access HSV-1 infected trigeminal ganglia in mice and rabbits

We have reproducibly demonstrated that circulating HSV-1 CD8⁺ T cells induced following parenteral and mucosal immunizations can efficiently migrate from the periphery to sensory ganglia and protect from acute infection following an ocular challenge in mice [13, 98-101] and rabbits [102]. An important goal of a therapeutic HSV-1 vaccine would be to enhance HSV-specific CD8⁺ T cells population that would migrate to latently infected TG and stop or reduce virus reactivation. Whether the size and the function of CD8⁺ memory T cell pool can be increased in relevant animal models of spontaneous latency/reactivation and humans following therapeutic vaccines remains to be determined. Mice can be infected with HSV via several routes including footpad, flank, ocular, intravaginal, and each route has its own unique properties. HSV infection at any of these sites initiates with local viral replication, which leads to the development of lesions, establishment of latent infection within sensory neurons and persistence of latent viral DNA within local sensory ganglia and in some cases, central nervous disease. Mice can certainly mount CD8⁺ T-cell responses in TG following acute infection, but spontaneous reactivation of HSV-1 and recurrent corneal disease are either rare or absent making the mouse an unreliable model to assess therapeutic vaccines [14]. While spontaneous reactivation of the virus from sensory ganglia and shedding of the virus in tears or in genital tract do not seem to occur in mice (as opposed to rabbits, guinea pigs and humans), reactivation of HSV from latent infection is readily observed in vitro when ganglia are explanted in culture [68]. HSV reactivation in mice can be induced to a limited extent by ultraviolet irradiation [103] or elevated body temperature or hormone [104, 105]. A recent study used an artificial restraint “stress” mouse model of virus reactivation [106]. In this model, mice are subjected to stress by restraining them in aerated plastic tubes for 12 hrs and depriving them of food and water for the same period of time. If this condition is translated to humans, it would be equivalent to a 12 hours stress caused by a “9-scale earthquake” similar to what recently happened in Japan. Obviously, this protocol of extreme and acute stress does not represent the daily “physiological” stress situations in humans (be that a physical or chemical stress). Interestingly, the authors recognized that “Exposure of HSV-1 latently infected mice to restraint stress diminishes the TG-resident CD8⁺ T cell population, compromises the function of the remaining CD8⁺ T cells [107-110], and induces HSV-1 reactivation from latency [111].” The study also claimed that exposing latently infected mice to restraint stress and treating them with TAK-779 blocked CD8⁺ T cell migration from blood to TG and conclude “Our findings suggest that augmenting the number of circulating HSV-specific CD8⁺ T cells is not sufficient to bolster the HSV-specific memory T cell population in latently infected sensory ganglia”. One cannot expect CD8⁺ T cells to migrate from the blood to latently infected TG if the number of these same CD8⁺ T cells is reduced in the first place, in both the blood and TG following the restraint stress itself [107-110]. Therefore, the mouse may not represent an appropriate animal model to investigate whether therapeutic vaccines will/or will nor boost the number and function of CD8⁺ T cells in the TG that would reduce spontaneous virus reactivation.

6. Human Leukocyte Antigen (HLA) transgenic rabbits: a “humanized” animal model that develops spontaneous recurrent ocular herpes and responds to human T cell epitopes

HLA transgenic rabbit model of ocular herpes has been proposed recently in our laboratory [112, 113]. Like their wild type counterparts, “humanized” HLA-A*0201 transgenic rabbits: (i) develop acute corneal disease; (ii) develop spontaneous reactivation; (iii) produces recurrent HSK, which is clinically similar to human HSK. In addition, “humanized” HLA-A*0201 transgenic rabbits (i) expresses human HLA class I molecules. One major component of the rabbit immune system is replaced by the identical component taken from human counterpart (i.e. HLA-A*0201 class I molecules) [112]; (ii) recognize and mount immune responses to HSV-1 human CD8+ T cell epitopes [114]. (iii) Although the state of the art in rabbit immunology still lags behind that of the mouse and human, several monoclonal and polyclonal antibodies specific to rabbit CD markers, cytokines, chemokines, and growth factors are now commercially available. Anti-rabbit CD8⁺ T cell mAbs and human tetramers have proven success to analyze HSV-specific CD8⁺ T cell infiltrates in TG and conjunctiva of acutely infected HLA Tg rabbits [12].

Taken together, the striking similarities between HLA Tg rabbit and human in terms of HSV-1 infection and immunity suggests that the HLA Tg rabbit is a preferred model to study the role of “symptomatic” vs. “asymptomatic” CD8⁺ T cells in exacerbating and controlling spontaneous HSV-1 reactivation and recurrent HSK. To our knowledge, this HLA Tg rabbit model is the only animal model with spontaneous HSV-1 reactivation that can develop “humanized” CD8⁺ T cell responses to human HSV epitopes [114]. This HLA Tg rabbit model will allow us for the first time to investigate the role of human specific CD8+ T-cells in reducing HSV-1 spontaneous reactivation. It will now allow for the first time to test the hypothesis that a therapeutic vaccine encompassing asymptomatic human CD8+ T-cell epitopes from frequently recognized HSV-1 antigens (such as ICP-0, ICP-4, gB, gD, VP11/12, and VP13/14) can decrease the effects of spontaneous reactivation (virus shedding in eyes and recurrent herpetic eye disease). A schematic diagram of HSV-1 spontaneous reactivation from the TGs and the possible role of “asymptomatic” CD8⁺ T cells (specific to human “asymptomatic” HSV-1 CD8⁺ T cell epitopes) in decreasing ocular herpes infection and disease in HLA-Tg rabbit are shown in Fig. 4. In addition, the availability of an HSV-1 mutant virus (CJLAT) that produces high levels of recurrent HSK in rabbits (up to 70% of eyes, compared to ~1-2% of wild type HSV-1 infected eyes) [115, 116] will allow for the first time to assess the therapeutic efficacy of human CD8⁺ T-cell epitope-based vaccines against recurrent HSK.

Fig. 4.

A schematic representation showing the possible role of asymptomatic CD8⁺ T cells (specific to human HSV-1 CD8⁺ T cell epitope) in decreasing ocular herpes infection and disease in the newly identified HLA-A*0201 transgenic (HLA Tg) rabbit model of spontaneous reactivation, recurrent corneal disease. Unlike what happens in the mouse model, in the rabbit model, similar to humans, circulating HSV-1-specific CD8⁺ T cells do access HSV-1 latently infected trigeminal ganglia where they encounter newly made antigens made following spontaneous reactivation of the virus from latency.

A recent publication by Chentoufi et al. from our laboratory [112] showed that HSV-1 (McKrae) infected HLA Tg rabbits recognize human CD8+ T cell epitopes from HSV-1 gD. The frequencies of CD8+ T-cells specific to a panel of 10 potential HLA-A*0201-restricted human gD epitopes in the draining lymph nodes (DLN) of HLA Tg rabbits have been measured by standard HLA-A*0201 tetramer-staining assays. In parallel, the frequencies of CD8+ T-cells specific to the same 10 peptide panel in the PBMC of HSV-1 asymptomatic HLA-A*0201(+) human individuals are also compared. Three gD epitopes (gD_53-61, gD_70-78, and gD_278-286) are highly recognized by both humans and HLA Tg rabbits but not by non-transgenic (wild type New Zealand rabbits). These results conveyed the message that HLA-A*0201 Tg rabbits and HLA-A*0201(+) human individuals share similar CD8⁺ T-cell responses when the epitopes were presented by the same HLA-A*0201 molecule. This highlights the potential of HLA-A*0201 Tg rabbits as a valuable animal model for the pre-clinical assessment of human HLA-A*0201-restricted epitope-based therapeutic vaccines.

When HLA Tg rabbits latently infected with wild type HSV-1, only 2% of rabbit eyes develop HSK which is virtually identical to human HSK by clinical exam. Interestingly, HSV-1 mutant (CJLAT), which was constructed from HSV-1 strain McKrae, showed significantly higher spontaneous reactivation rate than other HSV-1 viruses [117]. Over 70% of CJLAT-infected rabbit eyes develop severe HSK [115] which are indistinguishable from human HSK. HSK in CJLAT infected rabbits are evidenced by: (i) Clinical exam [115] (ii) histopathology [115]; (iii) Confocal microscopy. Therefore, the CJLAT-infected HLA Tg rabbit model would represent a realistic animal model for systematic study of recurrent-HSK.

7. Guinea pig model closely mimics human genital HSV-2 infection and disease

The guinea pig represents the gold standard and the most clinically relevant animal model for evaluation of therapeutic vaccine candidates against genital herpes caused by HSV-2 infection [118-120]. Many clinical and pathologic features of acute and recurrent genital disease of guinea pigs inoculated with low doses of HSV-2 are similar to those seen in human infection. Intravaginal inoculation of HSV-2 into guinea pigs leads to acute replication and disease at the site of infection, establishment of a latent viral reservoir within the innervating sensory neurons, and periodic reactivation leading to viral shedding and even recurrent clinical lesions [73, 74, 121, 122]. The development of both acute and recurrent disease provides the opportunity to test both the prophylactic and the therapeutic effects of potential vaccines in guinea pig model. Since late-1980s, the guinea pig model of genital herpes has allowed investigators to evaluate several vaccine and immunotherapeutic strategies for the control of recurrent herpes virus infections [123]. The guinea pig model is useful in testing therapeutic vaccine candidates against recurrent HSV infection for several reasons [124]. (i) This animal model mimics human infection, with self-limited primary vulvovaginitis developing after inoculation with HSV. (ii) Despite a full range of host immune responses, guinea pigs infected intravaginally with HSV-2 develop primary genital disease and establish latency in sacral ganglia (SG) (Fig. 2A & B). After recovery from primary illness, animals exhibit both spontaneous and UV induced recurrent disease, and establish latent infection in the lumbosacral dorsal root ganglia [125]. (iii) Periodic reactivation from latency produces recurrent genital infections, similar to human genital herpes. The infectious virus is usually detected in the genital tract along with high levels of recurrent genital disease [126, 127]. However, the disadvantages are the lack of inbred strains and limited availability of immunologic reagents to follow the immune responses. Due to the current paucity of immunological reagents for the guinea pig, the full range of immune effector mechanisms cannot be determined [73, 74, 128-130]. In addition, guinea pig T cells do not recognize human leukocyte antigen- (HLA-) restricted T cell epitopes and therefore the immunogenicity and protective efficacy of human T cell epitope-based vaccine and immunotherapy cannot be assessed in guinea pigs. The development of HLA transgenic guinea pig would be able to generate T-cell responses specific to HLA-restricted T cell epitopes. Moreover, with the development of defined immunologic reagents, this model should prove useful for exploring the immune responses that are important in the control of primary, latent, and recurrent HSV-2 infection and disease in humans.

Lack of a HLA transgenic guinea pig model has been a significant impediment in the development therapeutic CD8⁺ T-cell based vaccine against genital herpes. The development of a “humanized” guinea pig model using a transgenic approach has long been awaited. The technical challenges including: (i) a longer gestation time (60–75 days) relative to that of mice (20–30 days); (ii) a smaller average litter size (4 vs. ≥7); (iii) a limited number of litters/ year (5 vs. 10); (iv) a difficulty in obtaining fertilizable eggs; and (v) the unavailability of embryonic stem cells, certainly obstruct the progress in making transgenic guinea pigs. In efforts to address these issues, our lab is applying novel transgenic technologies, such as physical and chemical delivery methods of naked plasmid DNA directly to spermatogonium, to develop a guinea pig model carrying HLA class I and class II (HLA-A*0201 and HLA-DR) genes. A transgenic guinea pig over-expressing the human HLA-A*0201 and HLA-DR would serve as a valuable research model to test the efficacy of T-cell based therapeutic human vaccines against genital herpes. Cotton rat as an alternative model for HSV-2 infection and disease has been proposed by others to test the efficacy and safety of candidate therapeutic vaccine [131]. The salient feature of this model is that, unlike mice, vaginal infection does not require hormonal manipulation. Primary infection of cotton rat is associated with spread of virus to other organs such as the liver. Rats recover from primary infection are prone to spontaneous clinical recurrences that mimic human disease and infection. Additional advantages include the availability of immunological reagents to study the immune response to HSV in rats.

8. Protective immune mechanisms against herpes in various animal models versus human

The herpes studies on mouse model have absorbed decades of research and considerable amounts of resources, yet cannot guarantee the efficacy of vaccines in humans. Some advantages and disadvantages of preclinical animal models used for the evaluation of herpes therapeutic vaccine study are listed in Table I. Fig. 3 illustrates the HSV infection-reactivation cycle and the steps where a therapeutic vaccine can be targeted to interrupt virus cycle. Immunization with gD induces full protection against primary herpes infection and disease in various mouse models [132-134]. However, therapeutic immunization with gD induce a 74% protection against genital herpes in HSV-1 seronegative women, while failing to induce significant protection in men [135]. The reason for this partial protection and the immune mechanism behind it remains to be elicited. Usually, human individuals need higher number of doses to elicit protection. For example, while transient moderate protection was inconsistently achieved after 3 doses of vaccination lasting over 6 months in humans, often a single dose of vaccination is sufficient to induce strong protection in mice [133, 136]. Immunization schedules that are effective in mice (either BALB/c or C57BL/6) failed to protect rabbits and guinea pigs [137]. A series of 2 immunizations with gB or gD provided protection in only a fraction of immunized rabbits and guinea pigs and did not prevent death [138, 139]. Unlike mice, both rabbits and guinea pigs are extremely difficult to protect [138, 140, 141]. Rabbits and guinea pigs are exquisitely susceptible to infection since < 10² pfu could induce ocular and genital infection respectively [138, 139]. Rabbits and guinea pigs develop latent infection and shedding of virus at low density [138]. The same is observed in humans. In contrast, in mice, high virus loads and death are the most frequent outcome following herpes infection [133, 136]. The similarity of herpes infection between rabbits, guinea pigs and humans appears as a valuable advantage, as compared to mouse models. Therefore, the mouse model of herpes infection is quite different from human/ rabbit and guinea pig in terms of the spontaneous herpes reactivation issue (Fig. 2A, 2B & 2C). This suggests that a distinct pattern of protective immune mechanisms might play in mouse. This is well illustrated in the recent history of success of pre-clinical herpes vaccine using mouse models that were failed in clinical trials [4]. Initially it was thought that the protection mechanism was operated by the induction of virus neutralizing antibodies (based on mAb passive transfer experiments in mice) rather than by the induction of HSV-specific CD8 T cells. Later reports however have claimed IFN-gamma-producing and granzyme B (GrB) positive CD8⁺ T cells as a critical component of defense [24, 142]. Recent pre-clinical and clinical vaccine studies clearly suggest that the immune mechanisms of protection in murine models differ from those in humans.

Table I.

A list of advantages and disadvantages for preclinical animal models generally used for herpes vaccine research.

Animal	Advantages	Disadvantages
Mouse	Cheaper Easy to handle due to small size Well characterized immune system Commercially available immunological reagents Commercially available inbred and transgenic variety	Rare spontaneous HSV reactivation A few report of only molecular reactivation of HSV1 in the TG No recurrent viral shedding or recurrent HSK Not suitable for HSV1preclinical therapeutic vaccination study
Guinea Pig	Cheaper Easy to handle Do show spontaneous HSV reactivation More closely resemble human genital disease Good preclinical model for HSV2 therapeutic vaccination study	Not well characterized immune system HLA transgenic species not available yet Poor availability of immunological reagents are poorer than mice
Rabbit	Do show spontaneous HSV1 reactivation HLA transgenic species are established Ocular infection by all HSV1 strains Good preclinical model to study HSV1 therapeutic vaccine efficacy More tissues are available for in vitro studies Anatomically similar eye morphology with human eye	Not well characterized immune system Immunological reagents are now available but less than those for mice
Cotton Rat	A biologically superior model of human genital disease compare to mouse and guinea pig Both spontaneous and induced genital HSV2 reactivation is established Excellent model to study oral (lip) HSV1 infection (herpes labialis) Commercial availability of immunological reagents to study mucosal response against HSV	HLA transgenic species are not available yet Immune system is not well characterized

Fig. 3.

Schematic flow chart representing the HSV infection-reactivation cycle and the potential targeted step where therapeutic vaccine should intervene this cycle. The blue arrow indicates infections, latency, spontaneous reactivation, viral shedding, recurrent infection and eye disease. The brown arrow and the brown cross -indicated inhibition of spontaneous reactivation, viral shedding, recurrent infection and eye disease.

Since we have limitations to study the human immunity against herpes, we have to rely on the best fitting animal models that reflect the desired pattern of immune responses. Table II provides the list of all experimental animal models used so far for HSV infection. Models are considered valuable until the test results have been ascertained using the same virus in its natural host. The first and fore most criteria of any successful vaccination study at the pre clinical level is to choose the appropriate animal model to demonstrate the safety, immunogenicity, and protective efficacy of the candidate therapeutic vaccine. Table III offers a recent update of preclinical animal models used for herpes vaccine research. We recently uncover that the HLA-Tg rabbits are the candid animal model for studying therapeutic vaccines against HSV-1 mediated ocular infection and disease. In the same note, we anticipate that HLA transgenic guinea pig would be the other perfect model to study the recurrent genital herpes shedding and disease. However, HLA transgenic guinea pig model is not currently available but the concept certainly could be a future objective to study recurrent genital herpes.

Table II.

Current list of experimental laboratory animals and non-human primates used to study HSV-1 and HSV-2 infection and disease.

Experimental animals used for HSV infection
Animals	HSV-1	HSV-2	Route of infection	Disease
Mice	HSV-1 (all strains)	HSV-2 (all strains)	All possible routes
Cotton Rat	N/A	HSV-2 (strain G)	Topical through vagina	Genital herpes
Wister Rat	HSV-1(strain F & 16)	N/A	Topical through eyes
Wister Rat	HSV-1 (strain F)		Intra nasal
Lewis Rat	HSV-1 strain McKrae)		Topical through eyes
Rabbit	HSV-1 (strain McKrae)	N/A	Topical through eyes	Herpetic keratitis
Guinea Pig	N/A	HSV-2(strain 17)	Topical through vagina	Genital herpes
Zebra fish*	HSV-1 (strain KOS)	N/A	Intra Peritonial	Central nervous system
Non-human Primates
Rhesus Macaques	N/A	HSV-2 (strain G)
Owl Monkey**	HSV-1 (strain F)		Intracerebral injections

^*Zebrafish. 2008, Burgos JS;

^**Gene Ther. 2003, Deisboeck TS

Table III.

A recent update summarizing the list of preclinical animal models used to study the efficacy of herpes vaccine in treating primary ocular, primary genital and recurrent genital infections.

Recent Preclinical animal models used for herpes vaccine study
Animal	Transgenic	Virus	Ocular/genital	Infection	Outcome
Mice1	RAG2(-/-) gamma (-/-)	HSV-2	Genital	Primary	1. Protective innate immune response
					2. Reduced viral replication in genital tract
Rabbit2	HLA-A2	HSV-1	Ocular	Primary	1. Reduced HSV-1 replication in tears
					2. Reduced corneal eye disease
Guinea Pig 3	Wild type	HSV-2	Genital	Primary	1. Less genital skin disease
					2. Reduced viral replication in genital tract
Guinea Pig4	Wild Type	HSV-2	Genital	Recurrent	1. Reduced recurrent genital lesions

¹Kwant-Mitchell, A. 2009, J. Virol;

²Chentoufi, AA. 2010, J. Immunol.;

³Fotouhi, F. 2008, FEMS Immunol Med. Microbiol.

⁴Rose, WA 2^nd, 2008, Int. J. Antimicrob Agents

9. Pre-clinical investigation of viral vectors for prime-boost based therapeutic ocular herpes vaccines in rabbits

The concept of using an unrelated recombinant virus to deliver viral antigens for the purpose of therapeutic vaccination was developed more than 25 years ago [143]. In 1983, a vaccinia virus (VACV) recombinant for the hepatitis B surface antigen (HBsAg) was the first viral vector shown to successfully induce protective immunity against HBV [143]. Since then, a number of unrelated viral-vector vaccines using adenovirus, HSV, vaccinia virus, retrovirus, New-castle disease virus, vesicular stomatitis virus, measles virus, poliovirus and West Nile virus, have been designed and tested in both preclinical and clinical trials (reviewed in [144]). Among the advantages of adenoviral vectors are: (i) their natural tropism for mucosal surfaces, which makes them ideal for the purpose of mucosal vaccination against HSV-1 ocular infection for which parenteral immunization strategies have failed to confer protection [145-147]; and (ii) their remarkable ability to significantly boost T cell responses that were primed by a sub-unit vaccine (e.g. DNA and peptides) [148]. Recombinant rAd5 vector-based vaccines have been used successfully in rabbit eyes [149], and thus are suitable for the prime-boost pre-clinical studies against ocular (in rabbits) and genital (in guinea pigs) herpes. Currently there are six active clinical trial studies in the US that use rAd5 vector-based vaccines for non-herpes conditions (ClinicalTrial.gov. To minimize the potential rAd5-induced keratoconjunctivitis in rabbits, we have recently constructed replication incompetent rAd5 vectors expressing human “asymptomatic” CD4⁺ and CD8⁺ T cell epitopes from gB and gD (BenMohamed. in preparation). While the replication incompetent rAd5 is a promising clinical vaccine vector (reviewed in [148, 150, 151]), safety and pre-existing immunity are still concerns, and may require adoption of additional measures: (1) Of the 58 different human adenovirus (Adv) types belonging to species A-G, Adv8, -19, -37, -53 -54 and -56 (all within species D) are associated with acute keratoconjunctivitis in humans [152-159]. Some adenovirus strains present specific tropism to the eyes others are restricted [160, 161]. While the majority of serotypes utilize coxsackievirus-adenovirus receptor (CAR) as their primary attachment receptor, the subgroup D Adv19 and -37 associated with epidemic keratoconjunctivitis in humans use CD46 receptor (known as membrane cofactor protein) [162-164]. Adv37 also uses sialic acid linked to galactose via alpha2,3 glycosidic bonds as a cellular receptor [160]. In rabbits eyes, subgroup C (Adv1, -2, -5, and -6) appeared to be the most successful as vectors [165]. Topical ocular delivery of live Ad5, uses CD46 receptor that is ubiquitously expressed by a variety of cells from rabbit ocular surface [162-164], and may induce keratoconjunctivitis [161, 166]. Therefore, caution may be necessary when applying rAdv5 to rabbit eyes. (2) In addition, the majority of adults in developed world have inherent systemic Ad5 nAbs that may dampen its high transduction efficiency and immunogenicity when delivered parenterally [167, 168]. Replication-incompetent rAd5 vectors elicit potent T-cell responses in humans, but they have been shown to be substantially less immunogenic when administered intramuscularly to seropositive individuals who have preexisting anti Ad5 nAbs [168]. However, such limitation may not apply when the topical ocular route is used to deliver the rAd5 vaccine [169]. Indeed, pre-existing systemic anti-Ad5 nAbs did not seem to block transgene expression in the eye [149, 170-173]. Moreover, repeated ocular or mucosal administrations of rAd5 vectors result in efficient transgene delivery, expression and immunogenicity, even in the presence of preexisting systemic anti-Ad5 nAbs [169, 174-178]. Together, the above reports suggest that even with preexisting systemic anti-Ad5 nAbs, topical ocular mucosal route in HLA transgenic rabbits may still induce strong local HSV-specific CD8⁺ T cell responses. Once a replication incompetent rAd5 vector-based vaccine induced immunity in test-of-concept pre-clinical studies in humanized HLA Tg rabbits, future rAd vector-based clinical vaccines will be constructed using a vector derived from an Ad serotype that is rare in the developing world and distinct from Ad5, such as Ad26 and Ad35. This may effectively circumvent the problem linked to pre-existing anti-Ad5 nAbs. Those clinical studies will also very likely consider heterologous rAd35/rAd26 prime-boost regimens since this combination elicited remarkably potent T cell responses in non-human primates [150, 179-182]. Ad26 and Ad35 vectors infect a variety of human cells that ubiquitously express CD46, their primary receptor [167]. Furthermore, rAd26 and rAd35 efficiently transduce in the presence of anti-Ad5 antibodies, and seroprevalence of Ad35 in adults is much lower than that of Ad5 [148, 150, 167, 183].

10. Summary

1. Current concepts of sub-unit based prophylactic vaccines for ocular and genital herpes are promising but not successful in therapeutic settings [6, 7]. To achieve a successful therapeutic vaccine against recurrent virus shedding and recurrent disease, we are currently working on a novel prime-boost homologous and heterologous vaccine strategy that uses “asymptomatic” lipopeptide epitopes priming followed by replication incompetent recombinant adenovirus vector 5, 26 or 35 (rAd5, rAd26 and/or rAd35) boosting, with the lipopeptides and the rAds delivering the same “asymptomatic” CD8⁺ T cells epitopes.

2. Development of new and improved therapeutic vaccines and their delivery systems is constantly being pursued for the treatment of human herpes diseases. However, prior to their use in humans, all new vaccine candidates must undergo pre-clinical evaluation in a reliable animal model. The scientific community must overcome the challenge of finding such reliable animal model for pre-clinical studies that are important not only to establish the safety, immunogenicity and protective efficacy of the vaccine compound, but also to plan protocols for subsequent clinical trials from which safety and efficacy can be evaluated.

3. The mouse is the premier mammalian system for modeling human diseases and understanding the immune mechanism of specific genes in normal and diseased states. Mouse studies have provided plenty of useful and important information regarding ocular HSV-1 infection and despite the tremendous amount of knowledge gained about mouse immunology in general; the mouse is not an ideal model to extrapolate the mechanisms of recurrent HSV-1 and HSV-2 in humans. Spontaneous reactivations, sporadic shedding of the virus in the eyes and GT and recurrent disease do not occur in mice. Therefore, circulating HSV-specific CD8⁺ T cells may not access HSV latently infected sensory ganglia simply because there is no spontaneous antigens expressed as occurred in humans, rabbits and guinea pigs. Thus mouse in not a reliable model for assessing the efficacy of therapeutic vaccine against spontaneous reactivation or recurrent herpetic disease.

4. Besides the availability of numerous mouse specific immunological tools and mouse strains such as knock-outs, knock-ins and transgenic mice, the other advantages of mice are their faster breeding process and cheaper housing that made them praised as prime experimental animal model. But an increasing number of studies have revealed substantial differences between the immune responses induced by a vaccine candidate of mice and humans. Therapeutic vaccine candidates that show promise in a given mouse strain might fail in outbreed human population.

5. There has been a tendency to ignore differences in herpes infection and immunity between mice and humans, and in many cases, perhaps, make the assumption that what is true in mice is necessarily true in humans. By making such assumptions we run the risk of overlooking aspects of human herpes infection and immunity that do not occur, or cannot be modeled, in mice.

6. Persistent effort on developing herpes therapeutic vaccine using mouse as a preclinical model only increase the failure rate pointlessly. It is worth considering the possibility that any given immune response in a mouse may not occur in precisely the same way in humans.

7. So the current gamble is in between rabbit and guinea pig versus mouse model for the continuation of herpes therapeutic vaccine study at the preclinical level. While the HLA-transgenic rabbit and HLA-transgenic guinea pig models are difficult to generate, the easily available mouse model generate plentiful of data but may not be extrapolated to human.

8. We hope that the newly introduced HLA transgenic rabbit model (and maybe one day HLA transgenic guinea pigs one day) with spontaneous reactivation and recurrent herpes disease that mount a T-cell response to HLA-restricted human CD8+ T-cell epitopes will solve many hurdles associated with the preclinical phase of herpes vaccine development. These two animal models will help to address two questions (i) Can spontaneous herpes shedding and recurrent ocular or genital disease be reduced by induction of a vigorous “protective” HSV-specific CD8⁺ T cells in the sensory ganglia induced by “asymptomatic” epitopes? (ii) Conversely, can recurrent herpetic disease and spontaneous shedding be exacerbated by “pathogenic” CD8⁺ T-cells induced by “symptomatic” epitopes?

Abbreviations

HLA

Human Leukocyte Antigen

HSV

Herpes Simplex Virus

IFN-γ

Interferon gamma

Tg

Transgenic

TG

Trigeminal ganglia