Safety of Non-Replicative and Oncolytic Replication-Selective HSV Vectors

Abstract

Herpes simplex virus type 1 (HSV-1) is a DNA virus and human pathogen used to construct promising therapeutic vectors. HSV vectors fall into two classes: replication selective oncolytic vectors for cancer therapy and defective non-replicative vectors for gene therapy. Vectors from each class can accommodate ≥30 kb of inserts, have been approved clinically, and demonstrate a relatively benign safety profile. Despite oncolytic HSV replication in tumors and elicited immune responses, they are well-tolerated in cancer patients. Current non-replicative vectors elicit only limited immune responses. Seropositivity and immune responses against HSV don’t eliminate vectors or infected cells, so the vectors can be re-administered. In this review, we highlight vectors translated to the clinic, and hostvirus immune-interactions impacting safety and efficacy of HSVs.

HERPES SIMPLEX VIRUS (HSV) BIOLOGY

Herpes simplex virus type 1 (HSV-1) (Box1) is a prevalent neurotropic human pathogen with worldwide prevalence of about 67% [1]. HSV-1 normally induces immunity that prevents reinfections with the same serotype, but not with HSV-2 [2]. HSV-1 infection can be asymptomatic, mild, or life-threatening. In most immunocompetent individuals, HSV-1 causes only mild and self-resolving diseases, such as stomatitis, cold sores, or genital herpes [3]. In rare situations, HSV-1 infection is associated with diseases with high morbidity (herpes stromal keratitis, meningitis) and mortality (herpes simplex encephalitis) [3]. Why some individuals develop severe disease is not completely understood, but both innate and adaptive immune responses are fundamental to controlling HSV-1 and reducing pathogenesis [4]. Neonatal infection is more aggressive, leading to systemic dissemination with high morbidity if untreated [5].

Box 1. Molecular biology of wild-type HSV-1 infection.

The HSV-1 DNA genome is contained in an icosahedral capsid, coated with tegument, a layer of at least 20 proteins important for gene expression and virion assembly, and a lipid envelope studded with 11 different virus-encoded glycoproteins [2]. The envelope glycoproteins function in viral entry, spread, and immune evasion. Glycoproteins gB, gD, gH, and gL, are necessary for infection, interacting with Nectin-1, the main receptor, and inducing fusion of the virus and cellular membranes [91,92]. Following fusion, the tegument proteins and capsid dissociate, with the capsid transported to the nuclear membrane where the genome enters the nucleus [93]. Late tegument proteins in incoming virus particles exert immediate effects even before viral gene expression, thus counteracting cellular antiviral defenses [2]. The HSV-1 life cycle can then follow a lytic or latent pathway. During a lytic productive infection, a temporal cascade of gene expression occurs: synthesis of immediate-early (IE, α), early (E, β), and late (L, γ) proteins [2]. The IE infected cell proteins (ICP) ICP0, ICP4, ICP22, and ICP27 regulate expression of the E and L genes [2]. IE ICP47 is an inhibitor of human transporter associated with antigen processing (TAP), blocking MHC I antigen presentation [85]. Early proteins are required for replication of the virus genome, while L structural proteins are involved in virion morphogenesis and cell exit, as well as modulating immune responses.

Transmission of HSV occurs through close contact and typically follows primary infection of epithelial cells in the oral (HSV-1) or genital (HSV-2) mucosa, lifelong latency in sensory neurons, and reactivation, ending in reinfection of peripheral mucosa. During latent infection, ICP0, an E3 ubiquitin ligase, is not efficiently expressed and cannot counteract genome silencing so that the transcription cascade is not triggered, and the virus genome remains in an epigenetically repressed state, with the exception of untranslated latency-associated transcripts (LATs) [94]. Latency in sensory neurons is a compromise where the silent HSV-1 genome and infected cells are preserved from immune attack, while the host CNS is protected from virus spread. Danger signals such as stress, UV light, or fever, can stimulate reactivation from latency, with the lytic gene expression cascade resuming [2,94]. In regards to pathogenicity in humans, genetic polymorphisms/mutations in innate defenses are linked to a higher risk of encephalitis (e.g., TLR3, UNC93B1, TRIF, IRF3) [95]. Effective anti-viral drugs, such as acyclovir, are available, as opposed to most other viruses [3].

HSV-1 has evolved numerous functions to evade, moderate, and subvert immune surveillance [2]. Reactivations are rapidly controlled by the immune system without significant inflammation or strong immune reactions [4,6] (Box 1). Engineering the virus to exploit and manipulate these counteracting functions is key to designing optimized HSV-1 vectors, for example, allowing multiple vector administrations [7–10] in contrast to other vectors impeded by their immunogenicity.

Here, we review current knowledge from more than 30 years of preclinical and clinical studies. We briefly summarize the host responses induced and counteracted by wild-type HSV-1, and the immune responses prompted by oncolytic HSV (oHSV) and defective, non-replicative vectors. HSV vectors have attributes that make them very appealing for gene and cancer therapy (Table 1), as demonstrated with the approved HSV vectors. While difficulties working with or manufacturing HSV vectors are being overcome, the inaccurate view that they are inherently unsafe remains. Clinical studies have generally validated the safety profile observed in preclinical studies, with clinical adverse events being less frequent than originally postulated and also in comparison with other treatment modalities.

TABLE 1.

Advantages of HSV vectors.

Vector Attributes	Oncolytic	Non-Replicative
Broad host range	✓	✓
-surface receptors widely expressed	✓	✓
Non-integrating	✓	✓
Redirect cellular tropism	✓	✓
-glycoprotein mutations and new receptor-ligand expression	✓	✓
Very Infectious	✓	✓
Cytotoxic at low MOI	✓	✕
Induces Inflammation	✓	✕
Deletion of essential genes	✕	✓
-for safety and to provide space for inserted sequences
Latency in neurons for life	?	✓
Genetic engineering enables rapid construction	✓	✓
Multiple administrations	✓	✓
Low immunogenicity	?	✓
Space for large sequence inserts	✓	✓
Anti-viral drugs available	✓	✕
-for endogenous HSV reactivation		✓

✓, positive attributes; ✕, negative attributes; ?, questionable.

IMMUNE RESPONSES AGAINST HSV-1

HSV-1 establishes multiple intricate interactions with host defense systems, intrinsic, innate, and adaptive immunity, with the goal of evading immune surveillance and maintaining life-long persistence [11,12]. Understanding these interactions is critical for vector design, ensuring safety and efficacy.

Intrinsic and innate immunity

Intrinsic and innate immunity, which are interconnected, are the first lines of defense that limit viral spread and regulate the ensuing adaptive response [13]. The intrinsic response is mediated by preexisting restriction proteins assembled in Promyelocytic leukemia (PML)-nuclear bodies (or ND10) (Glossary) that inhibit lytic infection without requiring interferon (IFN) expression [13]. This is counteracted by Infected cell polypeptide 0 (ICP0), allowing the infectious cycle to proceed [14]. Innate immunity is triggered by pathogen associated molecular patterns (PAMPs) inducing IFN-stimulated gene (ISG) expression with antiviral and inflammatory activities [15]. Several cell types contribute to immune surveillance: NK cells, important for cytokine production and recognition and killing of virally-infected cells [15], and plasmacytoid dendritic cells (pDCs), the major producers of IFNα in vivo [16].

Intrinsic and innate cellular defenses are effective against HSV-1 during primary infection and reactivation from latency. However, HSV-1 expresses a large number of factors that counteract/modulate this, including immediate early (IE) proteins ICP0 and ICP27, early (E) protein UL42, and late (L) proteins US3, US5, US11, UL21, UL31, UL36, UL41, UL46/VP11–12, UL48, UL49/VP22, and ICP34.5 [4,17]. The IE and E viral functions eventually overcome the cellular defenses, so the infection progresses, but with decreases and delays in virus production and spread, giving time to elicit adaptive immune responses that ultimately control the infection.

Adaptive immune responses to HSV-1

Antigen-specific adaptive immune responses take time to evolve and finally control virus spread [18]. Neutralizing antibodies against HSV-1 play a seemingly minor role compared to CD8⁺ T cell-mediated immunity [19]. Conflicting results are often due to differences in animal species, administration route, latency model, complex interactions between cell types, and HSV-1 versus HSV-2 [19,20].

Humoral responses.

Antibody levels do not determine the outcome of HSV-1 infection in humans, while in mice, transfer of seropositive serum had no appreciable effect on viral replication or establishment of latency [19]. However, B cell-deficient mice are more susceptible to herpetic encephalitis after ocular infection [21]. In human latency, B cells and HSV-specific antibodies are increased and respond to reactivation in the skin [22]. In acute infections, the combination of both antibodies and CD4⁺ T cells protects neuronal tissue after immunization [23].

Cellular responses.

CD8⁺ T cells are recruited to HSV lesions early in infection, contributing greatly to immune control and cytolysis [19]. In recurrent genital HSV-2 patients, T cells expanded after reactivation and then returned to a steady state, like an acute recall response [24]. IFNγ-mediated CD8⁺ T effector functions contribute to protection against secondary infections, maintenance of latency, and limiting viral spread [19]. IFNγ-producing CD8⁺ T cells persist during latency in areas of T cell infiltration, despite not clearing the infection [19]. IFNγ receptor depletion leads to impaired resolution of HSV infections [20] and significant mortality [25]. Depletion studies showed that both CD4⁺ and CD8⁺ T cells contribute to protection against HSV replication and shedding in the local mucosa [19]. Neuronal infection and reactivation are mainly controlled by noncytolytic CD8⁺ T cells, however, in CD8-depleted or deficient mice, CD4⁺ T cells are sufficient to clear virus from both the local mucosal and neural sites [26]. Responses are further complicated by the modulation of MHC-I and -II expression [26].

HSV modulates adaptive immunity. HSV infects human immature dendritic cells (DC) and blocks their maturation, impairs their activity, and triggers apoptosis [16]. HSV-1 IE ICP47 blocks the transporter associated with antigen presentation (TAP) [27], and L ICP34.5 inhibits autophagy, processes involved in antigen presentation [28]. Infection of activated T cells is not productive but represses T cell receptor signal transduction and cytolytic activity [29]. HSV glycoproteins gE and gC bind to the Fc region of antibodies and complement, respectively, inhibiting antibody-dependent cell-mediated cytotoxicity [30].

The interaction between the host immune system and HSV, involving about a quarter of the HSV genes, is a long-term battle that often ends in a stalemate [12,18]. Alterations in HSV innate and adaptive immune evasion gene products can influence vector safety and efficacy and are important considerations in vector design [11].

HSV-1 VECTORS

There are two main types of non-integrating therapeutic vectors derived from HSV-1: (a) replication-selective oncolytic viruses for the treatment of cancer, named virotherapy [31] (Box 2), and (b) defective, non-replicative vectors for therapeutic transgene delivery to treat genetic and non-genetic diseases, named gene therapy [32] (Box3). Oncolytic HSVs (oHSVs), genetically engineered to replicate only in cancer cells (Box 2), have been evaluated in numerous patients [10]. They are potentially more pathogenic because of replication competence, thus, providing insight into the most serious adverse events HSV vectors might elicit. Successful therapy with these two vector types and their interactions with host defense mechanisms are diametrically opposite. OHSVs should kill infected cancer cells in an immunogenic fashion and be highly inflammatory, while non-replicative HSV-1 vectors should not alter infected cells and be non-immunogenic (Figure 1). Both vector types must be safe and non-pathogenic in patients.

Box 2. Oncolytic HSV-1 (oHSV).

OHSVs are genetically engineered to be replication-selective such that virus infection and/or replication is limited to cancer cells. They are a distinct class of cancer therapeutics, virotherapy, with two unique mechanisms of action: (i) oncolytic / cytotoxic activity directly killing cancer cells while sparing normal cells and tissue and amplifying in situ, and (ii) inducing tumor inflammation and anti-tumor immune responses [96]. Because cancer is often lethal with a large commercial market, there have been many more oHSV clinical trials investigating safety (phase 1) and efficacy (phase 2 and 3) than with non-replicative HSV vectors. OHSV was the first genetically engineered oncolytic virus (OV) described and one of the most translated OVs to the clinic [97]. OHSVs can be engineered for oncolytic activity and safety by deleting/mutating viral genes necessary for pathogenicity (γ34.5, UL56, UL39/ICP6), replication in normal post-mitotic cells (UL39/ICP6, UL23/TK, UL2/UNG), inhibiting anti-viral immune responses (γ34.5, Us3, ICP27, ICP47) and/or apoptosis (Us3, gJ, gG, ICP6, ICP22, LAT) [10,31,96,98]. This targets dysregulated/defective cell physiology common to most cancer cells, such as cell cycle control, cell proliferation, apoptosis, and innate immunity [96]. OHSV can also be receptor-targeted to cancer-specific cell surface molecules by deleting/mutating viral glycoprotein genes (e.g., gC, gD, and gB) and exchanging their binding domains with single-chain antibodies or ligands to rewire tropism [96]. Finally, oHSV can be transcriptionally-targeted by driving expression of IE ICP4 or L γ34.5 genes with cancer cell-specific transcriptional regulatory sequences (see CAN-3110) or repressing transcription with normal cell miR target sequences [96].

HSV can be a lethal pathogen in humans, so it is essential to err on the side of safety in designing these vectors. With its large genome size, oHSV can be ‘armed’ with therapeutic transgenes or sequences for gene therapy. The only OV approved in the US is talimogene laherparepvec (T-Vec, Imlygic^®), a GM-CSF ‘armed’ oHSV, for the treatment of advanced melanoma [10]. G47Δ (Delytact^®) was recently approved in Japan for the treatment of GBM [9]. There have been an additional 22 different oHSVs in clinical trials, currently or previously, for a range of solid tumors: breast, colorectal, gallbladder, GBM, head and neck, liver, melanoma, neuroblastoma, prostate, soft tissue sarcoma, pancreatic, and metastatic (clinicaltrials.gov) [58,90]. OHSV treatment induces immune and inflammatory responses in the tumor microenvironment (TME), which can be both beneficial and detrimental to therapy [11,98]. Induction of an immunologically ‘hot’ TME, so-called in situ vaccination [98], is important for clinical success [9,59,99,100].

Box 3. Non-replicative gene therapy vectors.

There are two classes of non-replicative defective HSV1 vectors, (i) amplicon and (ii) recombinant [32]. Defective amplicon vectors are derived from plasmids containing no HSV genes and only cis-acting sequences for DNA replication and packaging (~2 kb), as well as therapeutic genes or sequences [101]. With helper HSV, the amplicon plasmid is amplified and packaged as an approximately 150-kb genome into infectious viral particles [78]. First-generation amplicon vectors contained contaminating helper virus, which could be immunogenic [75]. The development of systems that generate helper-free amplicons resulted in no or only faint inflammation [76]. Non-replicative recombinant HSV-1 vectors are deficient in at least one essential IE function and are produced in complementing cells expressing the missing function(s). They also lack some non-essential genes to create space for the insertion of exogenous sequences like transgenes. First-generation non-replicative HSV-1 vectors couldn’t multiply due to the absence of IE ICP4. However, they displayed cytotoxicity due to the expression of other IE functions, mainly ICP0 and ICP27 [102]. These early studies introduced the notion that non-replicative HSV-1 vectors were toxic and immunogenic, an outdated view that has persisted. Further engineering resulted in vectors deleted or functionally deficient in the expression of all the IE genes [79,80]. They are fully non-toxic in infected cells yet preserve robust and durable transgene expression when driven by adequate regulatory sequences and protected from epigenetic silencing [79,80].

Figure 1. HSV-1 vector types.

(Left). Replication-selective oncolytic HSV for cancer therapy. The virus can replicate in and kill target cancer cells, generating more oHSV that can spread in the tumor, repeating the cycle. (Right). Non-replicative HSV vectors for gene therapy. Vector transduction/infection of target normal cells is minimally intrusive except for the consequences of transgene expression. Transgene products, but not the virus, can spread to other cells, acting on receptors or taken up by cells, or function in the cells where they are expressed, depending on the product. Both vector types are structurally identical and packaged in the same virus particles unless genetically altered, and only differ in their genome.

ONCOLYTIC HSV-1

oHSV preclinical safety studies

For US Food and Drug Administration (FDA) investigational new drug (IND) approval to initiate a clinical trial, in-depth safety studies must be performed in appropriate preclinical animal models that are sensitive to HSV pathogenesis, such as BALB/c or A/J mice, immunodeficient mice, and sometimes Aotus nancymae new world primates [33,34]. However, pathogenicity can vary greatly in different inbred mouse strains and with different ‘wild-type’ HSV-1 strains [35]. Typically oHSV is administered via routes related to the therapeutic target, for example, (i) intracerebral (IC) for brain tumors or central nervous system (CNS) disorders; (ii) intravenous (IV) for systemic delivery to metastatic disease; (iii) intracerebroventricular (ICV) for leptomeningeal or disseminated CNS disease; (iv) intraperitoneal (IP) for ascites and abdominal cavity tumors; and (v) intrahepatic artery for metastatic tumors and liver disease [36]. HSV-sensitive models and oHSV dose escalation should reveal potential toxicity that can be avoided or managed. Toxicity is often measured by determining the number of animals that don’t survive oHSV injection, morbidity, and histopathology of tissue from challenged animals [37].

Intracerebral (IC) Injection.

The most common route of oHSV administration is intratumoral, so treatment of glioblastoma (GBM) and other brain tumors involves IC injection [38,39]. The brain is the most sensitive organ to HSV pathogenicity, often leading to death or severe sequelae [3], so a common target for identifying toxicity. All oHSVs in clinical trials for GBM have a deletion of the γ34.5 gene, the major contributor to viral neurovirulence [40]. G207 (Fig. 2A; γ34.5 deleted (Δ), ICP6-, LacZ+ [41]), was the first oHSV developed for GBM treatment and to go into clinical trial in the US [42]. IC injection of G207 (10⁷ plaque-forming units (pfu)) produced no apparent symptoms or evidence of disease, while 50% of mice receiving 10,000-fold less wild-type HSV-1 succumbed within 1 week [41,43]. G207 superinfection of mice surviving HSV-1 at the same location caused no disease or evidence of HSV-1 reactivation [43]. G47Δ (Fig. 2A; γ34.5Δ, ICP6-, ICP47Δ, LacZ+; Teserpaturev / Delytact^®;), a third-generation oHSV derived from G207 by the deletion of ICP47, which restores MHC I presentation and increases virus replication in tumor cells, was as safe as G207 in mice [44]. Chimeric oHSV C134 (Fig. 2A), expressing the HCMV IRS1 gene to overcome γ34.5Δ decreases in virus replication, had a lethal dose 50 (LD50) >10⁷ pfu after IC injection, showing that HCMV IRS1 did not increase toxicity [33].

Figure 2. Genetic structure of (A) oHSVs in clinical trials, completed or current, and (B) non-replicative HSV-1 vectors NP2 and KB103 (B-VEC).

A. The HSV genome consists of unique long (U_L) and unique short (Us) regions bracketed by terminal repeat long (TR_L) and internal repeat long (IR_L), and internal repeat short (IRs) and terminal repeat short (TRs). Transgene inserts are in blue font, and below the arrow, deletions are indicated with below the genome. OHSV names are indicated on right and genomic alterations are indicated: 1716 [60]; G207 [41]; G47Δ [44]; T-Vec [46]; M032 [57]; CAN-3110 [65]; C134 [103]; NV1020 [70]; and HF10 [72]. B. IE genes are indicated in red, deletion by Δ, and location indicated on the genome. COL7A1 and PENK driven by the HCMV IE promoter are inserted into the deleted ICP4 region. IE ICP22 is downregulated by expression from an E promoter.

Other routes of administration.

ICV injection of G207 (10⁷ pfu) resulted in no disease symptoms in any mice compared to 50% morbidity with 1000-fold less HSV-1 [43]. In a separate study of ICV G207 (10⁷ pfu), 3/10 mice had to be euthanized due to rapid weight loss, which could be blocked by low-dose ICV G207 (10⁴ pfu) or poly I:C [45]. This is a rationale for administering an initial low or priming dose of oHSV followed by high-dose injections, as with talimogene laherparepvec (T-Vec; Fig. 2A) in patients [46,47]. Concerns about ventricular toxicity in mice led to the exclusion of patients with GBM tumors <1 cm from the ventricles. Other routes of G207 delivery found to be non-toxic, include IV, intrahepatic artery [48], and intraprostatic [49]. IV delivery is the preference of oncologists and the pharmaceutical industry, but is not very efficient with oHSV due to neutralizing antibodies and innate factors [48]. Oral or intra-esophageal administration of G47Δ produced only a transient small decrease in toxicity scores [50], while intra-sciatic nerve injection produced no neuronal ultrastructural abnormalities [51]. In vitro, G207 infection of human hematopoietic stem cells had no effect on CD34⁺ stem cells [52]. Repeated doses of T-Vec were well tolerated in BALB/c mice, rats after intrahepatic artery injection, and dogs [53]. However, injection of subcutaneous tumors in nude (T cell-) and SCID (B and T cell-) mice resulted in lethal systemic infections in 20% and 100% of mice, respectively [53], indicating a role for T and B cells in limiting toxicity.

Nonhuman primates (NHPs).

The New World owl monkey (Aotus nancymae) is exquisitely sensitive to HSV infection, with clinical symptoms and histopathology similar to human neonates [54]. After IC HSV-1 (10³ pfu), encephalitis, with classic histopathologic features, rapidly develops [55]. In contrast, IC injection of G207 (10⁹ pfu) resulted in no adverse effects except a seizure that resolved in a few days, a million-fold difference in neuropathogenicity [55]. All monkeys receiving IC G207 (3 × 10⁷ pfu) developed anti-HSV serum antibodies, and no virus was shed [56]. This is similar to what has been seen in patients [39,46]. After intraprostatic injection of G207 (10⁷ pfu), none of the monkeys displayed any disease symptoms, virus shedding, virus spread to other organs, or significant histological changes [49]. M032 (Fig. 2A) was evaluated in Aotus because human IL-12 is not active in mice and was found to be safe except for 1 high-dose animal [57].

OHSV clinical safety studies

Twenty-five different oHSVs have progressed to clinical trials for a range of malignancies, 16 expressing therapeutic transgenes, providing important safety data from patients (clinicaltrials.gov). Many of the clinical trials involve combinations with other therapies, which can contribute to toxicity and/or mask oHSV effects [58]. Here, we focus on a representative oHSV cohort with different genetic alterations and published clinical results. While safety is a key consideration in early phase 1 trials, the overall goal is to demonstrate antitumor efficacy at a safe dose. Importantly, this has been accomplished with two oHSVs that have been approved for clinical use (T-Vec and G47Δ) [9,47].

G207 (1×10⁶-3×10⁹ pfu) was stereotactically injected into GBM tumors, and found to be safe and potentially efficacious [42] (NCT00028158ⁱ). Two of 4 resections were positive for G207 DNA, while 5 autopsies had no viral histopathology or HSV detected [42]. G207 shifted tumor-associated immune cell subtypes, with post-treatment expression of CXCL10 and IDO1 correlating with survival [59]. A recent clinical trial combining 5 Gy radiation in pediatric high-grade gliomas resulted in a median overall survival (mOS) of 12.2 versus 5.6 months for historical controls [38]. There were no serious adverse events (AEs) attributable to G207 or virus shedding, and seroconversion only occurred in 3/5 patients receiving a high dose, indicating dose dependency [38].

HSV1716 (Fig. 2A; Sephrevir^®) was the first oHSV in clinical trial in Europe [60]. Because of animal testing regulations in the UK regarding non-human primates, much lower doses (≤10⁵ pfu) were inoculated. In 3 GBM trials, no clinical toxicity was attributable to the virus [61]. HSV1716 has been evaluated in several extracranial tumors, including in pediatric patients [62]. No dose-limiting toxicities were seen after IV administration in 9 young seronegative patients, while all patients with data seroconverted, and 4 had HSV-1 DNA in their blood on day 4, possibly representing virus replication [63].

M032 (Fig. 2A) was evaluated in a canine clinical trial in 21 pet dogs with glioma, as human IL-12 is active in dogs. There were no AEs attributable to M032, while modulation of the tumor microenvironment (TME) and some clinical benefits were observed [64].

In a pivotal G47Δ phase 2 clinical trial in residual or recurrent GBM, 19 patients were treated with up to 6 intratumoral injections (10⁹ pfu each) over about 5 monthsⁱⁱ [9]. All patients experienced virus-related symptoms, mostly low-grade AEs (fever, vomiting, nausea, and lymphocyte/WBC/neutrophil count decrease), with transient grade 3 or lymphocyte decreases occurring in 5 patients [9]. There was a significant increase in the 1-year survival primary endpoint, 84.2% in treated patients versus 15% in historical controls [9], leading to G47Δ approval for recurrent GBM in Japan [9]. There was a large increase in tumor-infiltrating CD8⁺ and CD4⁺ T cells, but not T regulatory cells, that persisted for months after treatment [9]. In the prior phase 1 (dose escalation)/2 clinical trial with 2 intratumoral injections of G47Δ, one year survival was 38.5%, possibly due to fewer injections. All seronegative patients seroconverted by 1 week after treatment [39]. A similar AE profile was seen, except for a coincidental increased frequency of seizures in 3/3 patients in the phase 1 high-dose cohort and none in 7 same-dose patients in phase 2, and small single intratumoral hemorrhages in 3/13 patients, likely due to biopsy [39]. It is difficult to know whether seizures were due to the virus or physical trauma from multiple injections, or tumor progression.

CAN-3110 (Fig. 2A; rQNestin34.5v2 [65]). In a dose-escalating (1×10⁶-1×10¹⁰ pfu) phase 1 trial of CAN-3110 in recurrent GBM/high-grade glioma (HGG) patients (NCT03152318ⁱⁱⁱ), HSV-1 but not HSV-2 seropositivity significantly correlated with better survival (mOS=14.2 versus 7.8 months), which may be due to enhanced anti-tumor and - virus immunity, while virus persistence correlated with seronegativity [66]. There were no obvious dose effects on survival and no dose-limiting toxicities, although 5 serious AEs, including 2 prolonged seizures, were possibly related to the virus [66].

Talimogene laherparepvec (T-Vec) (Fig. 2A; Imlygic^®) is the only OV approved in the US. In the pivotal phase 3 clinical trial (OPTiM; NCT00769704^iv), advanced melanoma patients were randomized to intralesional T-Vec or subcutaneous GM-CSF [67]. Due to low-grade ‘flu-like’ symptoms in seronegative patients in the phase 1 trial [46], the first dose was low (10⁶ pfu/ml) to seroconvert patients and was followed with 10⁸ pfu/ml once every 2 weeks [67]. The durable response rate (DRR) was 19.0% versus 1.4% for GM-CSF, and mOS was 23.3 versus 18.9 months [67], leading to FDA approval in 2015 for advanced melanoma and subsequently by the European Medicines Agency [47]. The main AEs, highest in the first cycle, were fatigue (50%), chills (49%), pyrexia (43%), nausea (36%), ‘flu-like’ illness (31%), and injection site pain (28%), of which <2% were grade 3–4 [67]. Biodistribution and shedding following treatment were minimal, with T-Vec DNA only detected on 11% of swabs of herpetic lesions [68]. T-Vec is currently in 16 active clinical trials for a variety of cancers, many with drug combinations (clinicaltrials.gov). Overall, no dose-limiting toxicities have been identified so far [47]. The safety profile in ‘real-world’ use is comparable to the OPTiM trial [69].

NV1020 (Fig. 2A) is derived from R7020, an HSV-1/2 intertypic recombinant developed as an HSV-2 vaccine [70]. NV1020 was administered via hepatic artery followed by conventional chemotherapy in a phase 1/2 dose-escalating clinical trial for metastatic colorectal carcinoma to the liver [71] (NCT00149396^v). A transient febrile reaction was observed after each virus infusion, but no virus-related grade 3/4 toxicities, and no shedding [71].

HF10 (Fig. 2A; Canerpaturev/C-REV/TBI-1401) is a spontaneously arising mutant oHSV that has been used in clinical trials for refractory solid tumors [72]. In a phase 1/2 trial in patients with superficial tumors or melanoma (NCT02428036^vi), it caused low levels of HF10-related AEs, mainly chills, fatigue, and pyrexia.

In summary, the clinical trial results have elucidated important safety features of oHSV immunovirotherapy: (i) a maximum tolerated dose (MTD) has not been reached in any of the trials; (ii) adverse events, mostly not serious, could be ascribed to oHSV, with the majority being expected “flu-like” symptoms [46,67,71], but seizures were observed in some GBM patients that could have been due to oHSV or tumor progression; (iii) there is limited if any shedding of oHSV; (iv) even in the brain, seroconversion was common among seronegative patients, indicative of an immune response to oHSV; (v) oHSV was likely replicating in the tumor but not in normal tissue; and (vi) the immune responses arising from oHSV infection don’t seem to be associated with toxicity. However, the inflammation and immunity induced by oHSV play a large role in the efficacy seen in patients, and in mouse models [9,66,73].

NON-REPLICATIVE GENE THERAPY VECTORS.

There are 2 general classes of non-replicative defective HSV: recombinant (virus genome lacking essential genes) and amplicon (helper virus dependent plasmid-based) [32] (Box 3). We will describe the immune responses elicited by these vectors and then the main features of their clinical trials. The immune responses elicited by non-replicative vectors, unable to spread and induce disease, and expressing no or only a few viral proteins, are expected to be mild and mostly limited to non-specific intrinsic and innate cellular responses. However, this is understudied and mostly in rodent brains [74]. A first-generation HSV-1 amplicon vector expressing LacZ and produced with an IE ICP4-deleted (Δ) mutant helper HSV was very inflammatory after IC injection, with activated microglial infiltration followed by activated lymphocytes and macrophages, which persisted [75]. A delayed inflammatory response was also seen at secondary sites projecting to the injection site [75]. This was probably due to helper virus contamination of amplicon stocks [76], with toxicity arising from IE proteins other than ICP4. IC injection of recombinant ICP4Δ HSV-1 vector (similar to helper virus above) activated immune responses that were dramatically lower than with a replicating vector [77]. Modifications to the amplicon plasmid (bacterial DNA deletion and insulator-like insertions) and diminished helper virus, greatly improved safety, and transgene expression [78]. Inactivation of all IE genes fully eliminated recombinant HSV vector cytotoxicity in cultured cells [79] and neurotoxicity and lymphocyte infiltration in the brain [80].

HSV-1 seropositivity did not eliminate inoculated virus or infected cells, induce significant immune state modifications, nor significantly affect transgene expression [81–83], likely due to disappearance of incoming viral structural proteins and lack of viral gene expression. The real concerns are immune reactions against the therapeutic transgene, an issue common to all gene therapy vectors, which depends on therapeutic protein immunogenicity, particularly for proteins not normally expressed by the host.

Gene therapy for cancer-related pain.

A first clinical trial exploited the ability of HSV-1 vectors to express transgenes after establishing a latent infection in the dorsal root ganglia (DRG). A dose-escalating, phase 1 clinical trial of NP2 (NCT00804076^vii), a non-replicative HSV-1 recombinant vector expressing human pre-proenkephalin ((PENK) Fig. 2B), was conducted in terminal cancer patients with intractable focal pain [81]. NP2 was injected into the dermatomes, with virus taken up by nerve terminals and transported to the DRG, where the vector institutes a persistent, quasi-latent state with transgene expression. No treatment-related serious AEs were reported and no subject seroconverted [81]. Pain relief was reported for the middle and high doses [81]. Since NP2 establishes a silent latent infection in the DRG, no expression of toxic functions is expected to take place. This clinical trial showed that intradermal delivery of NP2 was safe and potentially efficacious, the primary endpoints. A phase 2 clinical trial (NCT01291901^viii) was unfortunately discontinued for financial reasons.

Gene therapy for severe skin diseases.

B-VEC (beremagene geperpavec, Vyjuvek^™; KB103) (Fig. 2B) is a non-replicative HSV-1 vector expressing the COL7A1 (~9 Kb) gene, used to treat recessive dystrophic epidermolysis bullosa (RDEB), a shattering skin disease, resulting from collagen VII (C7) mutations impairing anchoring fibrils [82,84]. Repeated doses of B-VEC were topically administered to freshly renewed skin to treat the symptoms but not cure the disease. Unlike the NP2 study, this trial exploits the ability of B-VEC to strongly express transgenes in superficial layers of the skin and be readministered as many times as required. Latency was not tested and is irrelevant, as is the fact that B-VEC can express some toxic functions, such as ICP0 and ICP27, since skin cells are constantly being renewed. B-VEC expresses IE ICP47, which inhibits antigen presentation to reduce immune recognition [85], facilitating multiple administrations [82].

A randomized, placebo-controlled, phase 1/2 clinical trial (NCT03536143^ix), matching wounds from RDEB patients receiving topical B-VEC and placebo repeatedly over 12 weeks, met primary and secondary objectives [82]. No grade 2 or above B-VEC-related AEs, vector shedding, or tissue-bound skin immune reactions were noted [82]. This was the first non-replicative HSV clinical trial targeting non-neuronal cells. A phase 3, double-blind, intra-patient randomized, placebo-controlled clinical trial of weekly applications of B-VEC gel in RDEB patients (NCT04491604^x) was completed with similar results for safety and lack of significant immune responses [84]. The primary endpoint was complete wound healing at 6 months, reached in 67% of wounds and significantly better than the placebo 22%. Six of eight seronegative patients seroconverted, and 13 of 18 developed antibodies to C7 without significant immunologic reaction, and no association between HSV-1 serostatus or C7 seroconversion [84]. Based on this, B-VEC was the first topical gene therapy approved in the US, in 2023 [86]. The ease of topical application has advantages over invasive procedures, but more data is necessary to confirm long-term efficacy [87].

Following a similar approach, KB105, a non-replicative HSV-1 vector encoding human transglutaminase I (TGM1), was developed to treat autosomal recessive congenital ichthyosis (ARCI) [83]. Preclinical studies demonstrated that repeated topical administration induced TGM1 protein in the target epidermal layer only at the dose site, without fibrosis, necrosis, or acute inflammation [83].

To summarize, these clinical trials and preclinical studies clearly indicate the absence of severe AEs and confirm the safety of the vectors, while displaying varying levels of efficacy. Importantly, in no case did vector administration or transgene expression result in immune reactions that might prevent treatment efficacy.

CONCLUDING REMARKS

The major conclusion stemming from these studies is that both non-replicative and oncolytic HSV-1 vectors are safe and potentially efficacious. HSV-1 vectors possess numerous salient features that make them very appealing for gene therapy (Table 1). The evidence from preclinical and clinical studies indicates that immune responses against HSV-1 vectors, either oncolytic or non-replicative, do not constitute a major impediment to efficacy or safety, and multiple dosing is likely better than a higher single dose [8,9,39], even with HSV seropositivity (Clinician’s Corner). In the case of non-replicative HSV vectors, they elicit no or very low levels of immunity, too weak to impact therapeutic outcomes.

What is next? Studies with non-replicative HSV-1 vectors demonstrated that they can be used for stable transgene expression from latently infected DRG neurons or strong but transient expression in non-neuronal cells in the skin. These vectors have not yet been used in more mainstream gene therapy applications in non-neuronal cells, such as in diabetes or muscular dystrophies. The next challenge is to demonstrate that these vectors can produce safe and stable transgene expression in non-neuronal cells, by establishing a sort of latent-like infection. This will require the absence of toxic functions but also preventing long-term epigenetic silencing of the vector genome (Outstanding Questions). While this goal seems ambitious, results suggest that we are not that far from reaching it [79,80,88]. Obtaining such results in preclinical models of non-neurologic or CNS disease would represent important progress toward uncovering the therapeutic potential of non-replicative HSV-1 vectors [89].

For oHSV, where immune responses are larger and a key component of efficacy, immunity does not seem to have a serious negative impact on safety and possibly has a positive impact on efficacy. In contrast to early safety concerns about using HSV as a vector, overall safety and tolerability in patients is much better than expected. The spectrum of AEs attributable to the virus has been relatively modest, predictable for a viral infection, and less than those seen with most cancer therapies [90]. A maximally tolerated dose (MTD) for oHSV has not been reported in any of the clinical trials so far. Whether HSV vectors exhibit a typical dose-response or if an MTD is more efficacious than a lower dose is unclear. For HSV-1, a human pathogen, the ability to genetically eliminate pathogenicity, due to our understanding of virus-host interactions, is an important advance that enabled HSV’s use as a successful and safe viral vector for gene and cancer therapy [58].

Clinician’s Corner.

• HSV-1, a neurotropic human pathogen, is a large, enveloped virus with a 152 kb double-stranded DNA genome, containing about 82 genes. Anti-viral drugs are available, but a vaccine is not.

• HSV-1 has evolved a complex and balanced interaction with humans, developing numerous strategies to evade host defenses. After lytic infection of epithelial cells, the virus typically enters a latent state in sensory neurons from which it can reactivate, even in the face of adaptive immune responses, which limit virus pathology but don’t clear the infection. However, the virus has the capacity to cause significant and lethal disease, especially in the central nervous system.

• HSV-1 vectors, both attenuated oncolytic and non-replicative gene therapy vectors, have been found to be safe so far and effective in clinical trials.

• Both oncolytic and non-replicative HSV vectors have been approved in the US: T-Vec for the treatment of advanced melanoma and B-VEC for dystrophic epidermolysis bullosa, respectively.

• Non-replicative recombinant HSV-1 vectors can deliver long and multiple transgenes and/or sequences (~30 kb), including genes that cannot be delivered by most other vectors, such as the cDNA of dystrophin, von Willebrand factor, type 1 neurofibromatosis (NF1), as well as genes involved in ophthalmic diseases such as Stargardt disease and Usher syndrome. Defective amplicon vectors can carry up to 150 kb of sequence or multiples of shorter sequences. As such, non-replicative HSV-1 vectors are emerging as appealing and powerful vector systems.

• Non-replicative HSV vectors used in gene therapy elicit no or very low levels of virus immunity, too weak to impact therapeutic outcomes. They can be readministered several times, without inducing rejection of the virus or of the infected cells.

• For oHSV, where immune responses are larger and a key component of efficacy, both anti-viral and anti-tumor immunity don’t seem to have serious adverse safety effects, while HSV seropositivity may have a positive impact on efficacy. The adverse events profile in patients is typical of natural virus infections.

Outstanding Questions.

• How do non-replicative HSV-1 vectors perform in tissues other than neurons or renewable skin cells, which are naturally infected by HSV-1 and for which there is data?

  • -what happens in other tissues (liver, pancreas, muscle, kidneys, etc.) with epigenetic silencing, transgene expression, and immune responses, for example?
• How can we exploit the large size of the HSV-1 genome for delivering very long transgenes and/or regulatory sequences providing physiological control?

  • -how can we regulate the level and timing of transgene expression from HSV vectors?
  • -how do we overcome amplicon vector manufacturing difficulties to enable delivering sequences > 100 kb?

• How can we improve the delivery / targeting of HSV vectors, especially systemically, whether they are oncolytic for tumors or non-replicative for specific cell types or organs?

• Why has oHSV been so ‘safe’ in patients with cancer?

  • -is it related to the limited and localized infections typically seen with HSV-1?
  • -are we making oHSVs too attenuated for optimal efficacy and safety?
• What aspects of oHSV biology have the greatest impact on anti-tumor efficacy?

  • -what is the balance between oncolytic / cytotoxic activity, inflammation / innate and adaptive immune responses?

• How representative and/or prognostic of the targeted human disease are the preclinical models used to evaluate vector activity and safety?

Highlights.

• Preclinical and clinical studies indicate that immune responses against HSV-1 vectors, oncolytic or non-replicative, do not constitute a major impediment to efficacy or safety.

• Despite oHSV replication in tumors and their elicited immune responses, they have proven very safe and well-tolerated in cancer patients.

• Current non-replicative gene therapy vectors, which often express only the therapeutic transgenes(s), are safe and elicit only limited immune responses, in contrast to early safety concerns.

• Seropositivity and immune responses against HSV do not lead to elimination of HSV or infected cells, and the vectors can be readministered several times.

• Three HSV-1 vectors have already been approved, two of them for cancer therapy and one for gene therapy.

• Other HSV-1 vectors are currently being clinically tested, and serious adverse events linked to the vector have not been observed.

Glossary

Human cytomegalovirus IRS1 (HCMV IRS1)

Human cytomegalovirus IRS1 gene encodes a protein kinase R (PKR) inhibitor that antagonizes host protein synthesis shutoff and can complement this activity lacking in γ34.5Δ oHSVs

Infected cell polypeptide 0 (ICP0)

an E3-ubiquitin ligase is an IE protein that plays a critical role in counteracting many of the cellular innate antiviral defenses

Infected cell polypeptide 4 and 27 (ICP4 and ICP27)

two immediate-early (IE) HSV-1 proteins that play critical roles in regulating the expression of viral and cellular genes. Both are essential proteins, which means that lack of expression or mutation of either of them results in non-replicative viruses or vectors that can only be produced in cells expressing complementing amounts of these proteins

Infected cell polypeptide 34.5 (ICP34.5) gene or γ34.5

the major neurovirulence factor, is a late (L) viral protein that blocks host protein shutoff and autophagy, and is required for virus multiplication in neurons and other non-proliferative cells. Virus strains not expressing this protein can replicate in cancer cells but minimally in normal cells. For this reason, the gene encoding this protein has been deleted in most oHSVs

Lethal dose 50, LD50

the dose of virus or other agent at which only 50% of treated animals survive

Non-replicative HSV-1 vectors

: fall into two classes; (i) defective amplicon, derived from plasmids containing no HSV genes, only a viral replication origin and packaging sequences, and exogenous sequences that are packaged as an ~150kb genome in the presence of a helper HSV, and (ii) recombinant, lacking at least one essential IE gene or it’s expression

Plaque-forming units (pfu)

a measure of infectious virus based on the ability of a single virion to replicate and kill susceptible cells in a monolayer forming a plaque that can be counted

Promyelocytic leukemia (PML)

a gene involved in an oncogenic chromosomal translocation, is a key organizer of the so-called PML nuclear bodies

PML nuclear bodies or ND10 (nuclear domains 10)

matrix-associated domains that recruit an astonishing variety of seemingly unrelated proteins, including several that contribute to the innate antiviral cellular defenses and are targeted for proteolysis by ICP0

Unique long and short (UL and US)

HSV-1 genome contains two unique regions, long (UL) and short (US), bordered by inverted repeat regions (IR and TR). Genes located in UL are numbered as UL1, UL2 - UL56, while genes located in US are numbered as US1, US2 - US12

UL39 or ICP6

The large subunit of ribonucleotide reductase is essential for virus growth in post-mitotic non-dividing cells and contributes to neuropathogenicity