Commentary: Development of Therapeutics for Congenital Cytomegalovirus Infection

Abstract

In the mid 1980’s, I flew from Birmingham, Alabama to San Francisco, rented a car, and drove to Palo Alto so that I could meet with John Martin at Syntex. John, along with Julian Verheyden, synthesized ganciclovir, which had significant in vitro activity against cytomegalovirus (CMV) in vitro. This drug provided my colleagues and me an opportunity to evaluate it as a therapeutic agent for congenital CMV infection, knowing full well that it was mutagenic, teratogenic, and carcinogenic. John in his wisdom convinced the management of Syntex to provide ganciclovir for this disease, allowing me to study this drug in symptomatic congenitally infected children through the NIAID Collaborative Antiviral Study Group (CASG). Certainly, no other person or company would advocate for the use of such a medication in children, regardless of disease severity, because of its toxicity profile. Since these early days, ganciclovir, and subsequently its prodrug valganciclovir, have become the standard of care for the treatment of congenital cytomegalovirus infection. The following commentary defines the need and progress in the development of therapy

Introduction

In the mid 1980s, I flew from Birmingham, Alabama to San Francisco, rented a car, and drove to Palo Alto so that I could meet with John Martin at Syntex. John, along with Julian Verheyden, synthesized ganciclovir, which had significant in vitro activity against cytomegalovirus (CMV).^1–3 This drug provided my colleagues and me an opportunity to evaluate it as a therapeutic agent for congenital CMV infection, knowing full well that it was mutagenic, teratogenic, and carcinogenic. John in his wisdom convinced the management of Syntex to provide ganciclovir for this disease, allowing me to study this drug in symptomatic congenitally infected children through the NIAID Collaborative Antiviral Study Group (CASG). Certainly, no other person or company would advocate for the use of such a medication in children, regardless of disease severity, because of its toxicity profile. Since these early days, ganciclovir, and subsequently its prodrug valganciclovir, have become the standard of care for the treatment of congenital CMV infection. In the early days of ganciclovir discovery and development at Syntex, John Martin was the lone champion. He involved himself in all aspects from synthesis, biological screening, publications, and patent strategy. There is no doubt that without John’s DHPG program promotion the drug would have died at Syntex and perhaps elsewhere, too. The following commentary defines the need and progress in the development of therapy.

Background

Cytomegalovirs infections, ubiquitous in humans, are an important cause of congenital infection and a leading cause of sensorineural hearing loss (SNHL) worldwide.^4–6 The prevalence of maternal CMV infection is an important determinant of vertical CMV transmission. Congenital CMV infection rates are directly proportional to maternal seroprevalence in that highly CMV seropositive populations have higher rates of congenital infection.^7–9 Unlike rubella and toxoplasmosis, where intrauterine transmission occurs as a result of maternal infection acquired during pregnancy (primary infection), congenital CMV infection can occur in infants born to mothers who have had CMV infection prior to pregnancy (non-primary infection).^10–16 In fact, congenital CMV infection following a non-primary maternal infection accounts for two-thirds to three-quarters of all congenital CMV infections in highly seroimmune populations.^{8–12,17–19} This finding indicates the difficulty that will be encountered for vaccine development as congenital infection occurs in the presence of both humoral and cell-mediated immune responses. Young maternal age and non-Hispanic black race have been associated with an increased risk of congenital CMV infection.^6,20–25 The argument that maternal immunity to CMV limits intrauterine transmission is supported by the finding that the rate of intrauterine infection in women with pre-existing CMV immunity (non-primary maternal infection) is about 1–1.5%, which is about 20–30 fold less than in women with primary CMV infection during pregnancy.^8,4,10,26,27 However, the number of women experiencing primary CMV infections during pregnancy is considerably smaller than those with non-primary infections, especially in highly seropositive populations, as noted above. Thus, in populations with near-universal seroimmunity, infants with congenital CMV infection are born to women with non-primary infection.^8,28–30

Transmission to the developing fetus is believed to occur through hematogenous spread of infectious virus to fetal blood at the placental interface lying between maternal and fetal blood supplies. Virus presumably infects the fetal liver, and after amplification in the liver, or alternatively, during primary viremia, disseminates in the fetal circulation. However, the characteristics and mechanisms of virus infection within specific fetal organs, particularly the fetal brain, are unknown.

Recent findings have demonstrated that multiple viral genotypes are likely present within a single infection, suggesting that it is unlikely that specific viral genotypes of CMV account for phenotypic variations that follow fetal infection.^31,32 Developing an understanding of the parameters for dissemination of virus to organs in the fetus, particularly the CNS, is critical for both understanding the pathogenesis of the fetal infection, and, perhaps more importantly, for the rational design of prophylactic vaccines and potentially other interventions such as targeted biologics/therapeutics.

Congenital CMV infection is a leading cause of childhood permanent hearing loss and neurodevelopmental disabilities.^4,5,33 Congenital CMV-associated SNHL accounts for about 25% of all cases of SNHL in children . ³⁴ The number of children with congenital CMV-related disabilities is similar to or exceeds the number of children with better-known conditions such as Down’s syndrome or spina bifida. ³⁵ Approximately 85–90% of the 20,000 to 30,000 children born with congenital CMV infection each year in the United States do not exhibit any clinical abnormalities at birth (asymptomatic congenital CMV infection).^11,13,36 The remaining 10%–15% born with clinical abnormalities are categorized as having clinically apparent or symptomatic congenital infection. The infection may involve multiple organ systems with particular predilection for the reticuloendothelial and central nervous system. ³⁷ The most commonly observed clinical findings are petechial rash, jaundice, hepatomegaly, splenomegaly, and microcephaly. Ophthalmologic examination is abnormal in approximately 10% of infants with symptomatic congenital CMV infection, with chorioretinitis and/or optic atrophy most commonly observed.^38–40 Laboratory abnormalities in children with symptomatic infection reflect the involvement of the hepatobiliary and reticuloendothelial systems and include conjugated hyperbilirubinemia, thrombocytopenia and elevations of hepatic transaminases in over half of symptomatic newborns.^38–40 Neuroimaging is abnormal in approximately 50%–70% of children with symptomatic infection at birth, and intracerebral calcifications are the most common abnormality.^41,42 Although a number of non-specific neuroimaging findings have been reported including ventricular dilatation, cysts, and lenticulostriate vasculopathy in infants with congenital CMV, the significance of these findings is not clear.

Approximately half of the infants with symptomatic infection will develop sequelae including SNHL and cognitive and motor deficits.^{13,38–40,43–48} Predictors of adverse neurological outcome in children with symptomatic congenital CMV infection include microcephaly, the presence of other neurologic abnormalities at birth or in early infancy, neuroimaging abnormalities detected within the first month of life, and the presence of multiple clinical findings.^{39,41,45,48–53} Approximately 7%–15% of asymptomatically infected children will develop SNHL. Among those with hearing loss, one-half of children with asymptomatic infection will have bilateral deficits which can vary from mild high frequency loss to profound impairment.^{13,47,54–57} In addition, hearing loss in these children is often progressive and/or late in onset, requiring ongoing clinical monitoring.^47,54,56,57 Other neurological complications may also occur with asymptomatic congenital CMV infection, and while the data are sparse, occur at a much lower frequency than in symptomatic infection. ⁵⁷ All of these findings indicate the chronicity of congenital CMV infection.

Because of the significant morbidity associated with congenital CMV infection, it has become an important target for antiviral therapy. Against this backdrop, first ganciclovir, and thenvalganciclovir have been evaluated for the treatment of babies with symptomatic disease.

Ganciclovir

The development of therapeutics has been reviewed. (16) The initial studies of ganciclovir were performed in babies who had symptomatic congenital CMV infection. In a Phase IB study, intravenous (IV) ganciclovir was administered either at a dosage of 8 mg/kg or 12 mg/kg once per day for 6 weeks. The purpose of this study was to determine whether there was any impact whatsoever on clearance of virus from the urine and any preliminary evidence of improved hearing. By 6 weeks after the onset of therapy, the majority of babies who received 12 mg/kg/day had a significant decrease of virus load in urine and to a greater extent, statistically significantly so, than those who received 8/mg/kg/day. ⁵⁸ At the initiation of therapy, virus load was approximately five logs. At the high dose, most babies had a reduction to less than one log of virus in the urine, but still detectable after 6 weeks. Within a month of completion of therapy, excretion virus returned in the urine to approximately three logs. Thus, total elimination was not achieved. However, these findings were not without evidence of toxicity. Many either required adjustment in their treatment dosage because of hematopoietic toxicity or had had medication stopped and re-started, as detailed in the Phase III study below. Further, the necessity of maintaining a peripheral intravenous catheter for 6 weeks was problematic in the treatment group because of the development of secondary line infections.

With the establishment of the 12/mg/kg dose of intravenous ganciclovir,^59,60,63 the CASG conducted a Phase III randomized controlled trial study of 6 weeks of intravenous ganciclovir versus no therapy for neonates with symptomatic congenital CMV disease. ⁶¹ From 1991 through 1999, 100 patients with symptomatic congenital CMV disease involving the central nervous system (CNS) were enrolled. Patients were randomized to ganciclovir treatment at 6 mg/kg/dose IV every 12 h (12 mg/kg/day) for 6 weeks or to no treatment. Infants randomized to the no treatment arm were managed in an identical fashion to those receiving active drug: however, a placebo was not employed, as an indwelling catheter would be required for 6 weeks, making it unethical. In addition to viral clearance, other endpoints included an improved brain stem evoked response (BSER) by one gradation, or maintenance of normal hearing between baseline and 6-month follow-up. This endpoint included: (1) a biologic assessment that evaluated all the ears, and (2) a functional assessment of only the best ear. Brain stem evoked response audiometry was determined by one gradation between baseline and the 6-month follow-up (or, for those patients with normal hearing at baseline, normal BSER at both time points). Secondary endpoints included improvements of laboratory abnormalities 2 weeks after the onset of therapy, as well as clinical improvement, particularly weight gain and improvement in head growth. Follow-up audiologic analyses were performed on best evaluable ear (“functional” assessment) and on total evaluable ears (“biologic” assessment).

Of these 100 subjects, 42 met all study entry criteria for evaluation, had both a baseline and 6-month follow-up BSER audiometric exam, and thus were evaluable for the primary endpoint. Two additional subjects who had both baseline and 6-month follow-up BSERs did not meet all inclusion criteria (1 enrolled at 29 weeks gestational age and randomized to no treatment, one enrolled at 33 days of life and randomized to ganciclovir). Comparisons of the baseline demographic characteristics by treatment category of evaluable and non-evaluable patients for the primary endpoint revealed no significant differences with respect to indicators of CMV disease severity, including baseline BSER assessments.

Twenty-one (84%) of 25 ganciclovir recipients either had improvement in hearing in their best ear between baseline and 6 months or had normal hearing at both time points, compared with 10 (59%) of 17 subjects in the no treatment group [adjusted p-value = 0.06; OR 5.03 (95% CI: 0.84,45.94)]. Inclusion in the best-ear analysis of the two additional patients who did not meet all entry criteria yielded an adjusted p-value of 0.03. None of the 25 ganciclovir recipients had worsening in hearing in their best ear between baseline and 6 months, compared with 7 (41%) of 17 subjects in the no treatment group [adjusted p-value < 0.001; OR 21.11 (95% CI: 2.84,∞)]. Five (21%) of 24 ganciclovir recipients had worsening in hearing in their best ear between baseline and ≥1 year, compared with 13 (68%) of 19 subjects in the no treatment group [adjusted p-value = 0.002; OR 10.26 (95% CI: 1.79, 81.92)]. Ganciclovir: -treated subjects also had a more rapid median time to normalization of ALT (19 days) compared with subjects in the no treatment group (66 days) (p = 0.03).

As was noted in the Phase 1B study toxicity was significant as 29 (63%) of 46 ganciclovir-treated subjects developed Grade 3 or 4 neutropenia [utilizing the Division of AIDS (DAIDS) Toxicity Tables] during the 6 weeks of study drug administration, compared with 9 (21%) of 43 subjects in the no treatment group over the same period of time (p < 0.01). Fourteen (48%) of the 29 required dosage adjustments, although only four patients had the drug permanently discontinued. The mean time (±SD) of onset of grade 3 or 4 neutropenia for subjects receiving ganciclovir was 14.2 (±12.3) days, and for the no treatment group was 14.3 (±13.1) days. Neutropenia in ganciclovir-treated subjects resolved in 12.8 (±13.6) days, and in the no treatment group in 14.2 (±13.5) days. All affected subjects resolved their neutropenia. The incidence of Grade 3–4 thrombocytopenia was comparable across both study arms, as was the incidence of Grade 3–4 increases in serum creatinine, alanine amino transferase (ALT), and total bilirubin levels.

This study provided many important lessons. First, regarding the primary endpoint, ganciclovir recipients had a 79% stabilization in hearing, or ultimately, improvement. In the counterpart no treatment group, only 32% achieved this same endpoint. The decibel difference, respectively, between the two groups was 25, indicating a significant change. These changes in hearing were maintained for greater than 1 year. ⁶² Importantly, the data indicate that a chronic infection was amenable to therapy.

One of the main unanswered questions from the results of this study was whether ganciclovir treatment for 6 weeks improved neurodevelopmental outcomes of babies with symptomatic congenital CMV disease involving the CNS. Denver Developmental Evaluations had been performed at 6 weeks, 6 months, and 12 months in this trial, and were analyzed posthoc. ⁶² Of the 100 subjects enrolled, Denver assessments were available for 74, 74, and 72 subjects at 6 weeks, 6 months, and 12 months, respectively. At the 6-week assessment, the average number of developmental delays per subject were 1.5 for ganciclovir recipients and 2.05 for no treatment subjects (p = 0.13). At 6 months, the average delays were 4.46 and 7.51, respectively (p = 0.06). At 12 months, the average delays were 9.78 delays versus 17.14 delays, respectively (p = 0.007). Multivariate ANOVAs were utilized to test independent factors known to be related to poor developmental outcomes in congenital CMV infection. After adjustment for these factors, the effect of ganciclovir therapy remained statistically significant at the 12-month assessment (p = 0.007), and approached statistical significance at 6 weeks (p = 0.08) and 6 months (p = 0.08).

Valganciclovir

These results of the IV ganciclovir study suggested that the benefit of antiviral therapy could wane over the first 2 years of life,^61,63 implying that a new strategy of extending the duration of therapy might be beneficial. The prodrug of ganciclovir, valganciclovir, had recently been developed as a research-grade liquid formulation. The CASG therefore developed a pharmacokinetic (PK) and pharmacodynamic (PD) study to determine the dose of oral valganciclovir that achieved systemic ganciclovir exposure equivalent to IV ganciclovir. ⁶⁴ Previous studies had indicated that the area-under-the-curve (AUC) of ganciclovir is most closely related to virologic treatment success, and data from the CASG’s prior Phase II study in the 1980s of IV ganciclovir (6 mg/kg/dose q12h) in neonates with symptomatic congenital CMV disease showed a median AUC₁₂ of 27 µgxh/mL (mean AUC₁₂ = 32.3 + 13.7 µgxh/mL, range 17.2–55.9 µgxh/mL, n = 13). ⁵⁹ The mean ganciclovir AUC₁₂ following IV administration of 5 mg/kg in adults was 28.6 µgxh/mL at 1 week of therapy. By comparison, adult pharmacokinetic data of ganciclovir following 1 week of oral valganciclovir therapy have documented an AUC₁₂ of 32.8 µgxh/mL. ⁶⁵ Taken together, pharmacokinetic data from past adult, pediatric, and neonatal studies showed that subjects receiving ganciclovir, whether as oral valganciclovir or IV ganciclovir, generally achieve an AUC₁₂ in the range of 27–32 µgxh/mL, with a coefficient of variation in the range of 30–40%.

Twenty-four subjects under 1 month of age with symptomatic congenital CMV disease were evaluated and were with or without CNS involvement. This was based upon data from a long-term follow-up study of children with symptomatic congenital CMV infection over a 30 year period of time which suggested that disseminated CMV disease at birth (as manifest by petechiae, hepatosplenomegaly, intrauterine growth restriction, thrombocytopenia, or hepatitis) with or without the presence of neurologic involvement at birth, is predictive of hearing loss. ⁵⁴ Pharmacokinetic and pharmacodynamic data (above) indicated that a dose of 15.62 mg/kg provides an AUC₁₂ of 27.4 µgxh/mL, and the oral bioavailability of valganciclovir oral solution increased in early infancy from 48% at approximately 4 weeks of life to 64% at approximately 7 weeks of life; this increase is proportionate to the increase in ganciclovir clearance during the same period.

Using the oral dose of valganciclovir of 16 mg/kg/dose BID, the next CASG study assessed whether longer-term therapy provided superior durability of beneficial effects. Subjects were randomized to receive either 6 weeks or 6 months of oral valganciclovir therapy. A total of 109 subjects from 31 study sites were enrolled, and 96 subjects were randomized to blinded study medication (47 active drug, 49 placebo) after receiving 6 weeks of valganciclovir. Total ears from subjects receiving 6 months of valganciclovir were more likely to have improved hearing or to maintain normal hearing between baseline and 12 months after a priori adjustment for CNS involvement at baseline [aOR (95% CI): 3.04 (1.26, 7.35); p = 0.01]. Similar results were evident when prematurity and age at treatment initiation were added to the model (p = 0.01). The relative risk for improved or protected total ear hearing between baseline and 12 months for the 53 subjects with baseline CNS involvement receiving 6 months of treatment and 6 weeks of treatment was 1.66 (95% CI: 0.92, 2.4), and the risk difference was 0.27 (95% CI: 0.09, 0.45). The benefit of longer-term therapy in the total ears analysis was maintained at 24 months, with improved outcomes after adjusting for CNS involvement at baseline [aOR (95% CI): 2.61 (1.05, 6.43); p = 0.04]. Similar results were evident when prematurity and age at treatment initiation were added to the model (p = 0.04). The relative risk for improved or protected total ear hearing between baseline and 24 months for the 42 subjects with baseline CNS involvement receiving 6 months of treatment and 6 weeks of treatment was 1.46 (95% CI: 0.87, 2.05), and the risk difference was 0.23 (95% CI: 0.05, 0.41).

Adjusting a priori for CNS involvement at baseline, subjects randomized to 6 months of valganciclovir also had higher Bayley III Language Composite (p = 0.0046) and Receptive Communication Scale (p = 0.0031) developmental scores at 24 months compared with subjects randomized to 6 weeks of treatment. No significant interaction effects were found when outcome and baseline CNS involvement were incorporated in a single model, indicating similar treatment benefits for both groups (with and without CNS involvement). These differences were maintained when age at treatment initiation and prematurity were added to the model (p-values of 0.0037 and 0.0027, respectively). All other components of the Bayley assessments trended toward improved outcomes for subjects randomized to 6 months of therapy.

These benefits of treatment were noted whether therapy was initiated earlier or later during the first month of life. ⁶⁶ As a result of these CASG trials, the U.S. Food and Drug Administration (FDA) modified the valganciclovir package insert to cite the dose used in CASG studies, ⁶⁷ and the American Academy of Pediatrics now recommends 6 months of oral valganciclovir as therapy for infants born with symptomatic congenital CMV disease. ⁶³ Importantly, significant hematologic toxicity was not encountered with valganciclovir treatment, likely attributable to the differences in peak plasma concentrations when compared to intravenous therapy.

While the advances detailed in the studies above are important, the benefit of valganciclovir therapy overall is modest. This raises the possibility that more effective drugs or treatment regimens could provide additional benefit in audiological and developmental outcomes of treated babies (72). Subjects with symptomatic congenital CMV disease who achieve complete viral suppression (defined as ≤ 2.5 log) by day 14 of valganciclovir therapy and then maintain it over the next 4 months are statistically more likely to have improved hearing across the first 2 years of life. ⁶⁸ However, the ability of ganciclovir/valganciclovir to cause such a rapid decline is limited, with typical viral load decreases of only one log over the first week and then <0.5 log over subsequent weeks. ⁶⁶ Combination therapy with the addition of another antiviral drug with CMV activity, a different mechanism of action, and an acceptable safety profile has the potential to significantly advance our management options in the treatment of symptomatic congenital CMV disease, as has been the case with combination therapy in the treatment of human immunodeficiency virus and hepatitis C virus infections. These data foreshadow the potential for linking CMV biomarker(s) to drug exposure. These relationships would serve to isolate the most important pharmacokinetic parameters linked to treatment response, and more precisely identify the optimum dosing strategy. Importantly, for potential future combination therapy, establishing these relationships will ascertain whether synergism or additivity exists, necessitating lower doses/exposures and the potential for reduced toxicities.