Thomas P. Monath Crozet BioPharma LLC, Devens, MA, United States
Chikungunya and Zika viruses invaded the New World in 2013–14 and are likely to persist as endemic and periodically epidemic diseases. Both are transmitted by viremic humans and blood-feeding Aedes aegypti vectors. The temporal and geographic recurrence of future outbreaks will be dependent principally on patterns of human immunity. There is a risk of escape into sylvatic cycles of transmission in the Americas, which could change the long-term epidemiology of both viruses. As RNA viruses are subject to evolutionary change, changes in virulence and transmission patterns are likely. Interventions (vaccines and antiviral drugs) are feasible technically but have significant commercial and regulatory challenges. Future areas for research are described.
Chikungunya; Zika; Epidemiology; Ecology; Interventions
Perhaps the most important questions to be addressed are whether chikungunya virus (CHIKV) and Zika virus (ZIKV) diseases will persist in the Western Hemisphere; what patterns of transmission will predominate; what the full scope of Zika disease expression is with neuropsychiatric follow-up of surviving babies infected in utero; and what kinds of public health and medical responses will be required to contain and control these diseases. Chikungunya and Zika can be considered together since they share similar transmission cycles, and they invaded (or in the case of CHIKV, reinvaded) the New World in the same year (2013) (Riou et al., 2017). Additional questions related to future trends include whether other syndromes and complications of infection will be elucidated; whether virus evolution or adaptation will result in new virus phenotypes; what other areas of the world are at risk, and is ZIKV a new global threat; what does the future hold with respect to our ability to control Aedes vectors; and what role vaccination (feasible technically) and antivirals will play in intervention strategies. Some of these questions are addressed in this chapter.
It is highly likely that both CHIKV and ZIKV will persist in the Western Hemisphere because they are transmitted in a cycle involving a ubiquitous vector (Aedes aegypti) and a vast susceptible population of human hosts that are highly mobile and capable of efficiently transporting these viruses to areas that are ecologically and climatically able to support the transmission cycles. The reproductive rate of CHIK and Zika virus infections will be primarily determined by the prevalence of immunity, a function of cumulative immunity with age and rapid acquisition during the sweep of epidemics. The latter can render isolated communities or island populations immune to future outbreaks for years, but many population groups inhabiting large land masses will remain susceptible to repeated outbreaks. This pattern is well established for CHIKV in Asia (see Chapter 5), for dengue viruses (DENV) worldwide, and for yellow fever virus (YFV) in Africa, all of which have similar transmission cycles. Zika and chikungunya outbreaks with high infection rates in island communities (e.g., Zika in the Caroline Islands, French Polynesia; and CHIK in the Comoros Is., La Réunion, and the West Indies) have left behind a temporary barrier to reintroduction.
Various models have been applied to estimate the impact of virgin soil outbreaks and the residual tail of immunity. In the case of Zika, using a compartmental model, it was estimated that without intervention a Zika epidemic affecting a population of 100,000 people would burn out after 42% of the population was infected (Lee et al., 2016). In concert with this model, attack rates during virgin soil chikungunya outbreaks in island populations were 63% in the Comoros (Sergon et al., 2007) and 38% in La Réunion, after which further transmission ceased (Gerardin et al., 2008). It is interesting and unexplained at present why CHIKV and ZIKV appear to have similar transmission potentials, (Riou et al., 2017) despite levels of virus in blood far higher for CHIKV. It may be that the longer duration of viremia in ZIKV infections is responsible for this evening out of transmission potential.
An expanding array of modeling exercises have shed light on the potential for global spread of ZIKV. These models considered vector distribution, human travel patterns, and multiple factors including climate, vector abundance, land cover, and socioeconomic variables (Bogoch et al., 2016; Samy et al., 2016; Carlson et al., 2016). The emergence of widespread ZIKV transmission in the Pacific and in the Americas demonstrates the global reach of the virus wherever Ae. aegypti is prevalent, a pattern previously seen with chikungunya and dengue. However, the appearance of limited outbreaks of Zika in Singapore and the identification of microcephaly in Guinée-Bissau may only represent an enhanced index of suspicion and surveillance. An important question is whether the future epidemiology of ZIKV in Africa (and possibly SE Asia), where the virus is endemic and has a zoonotic cycle involving nonhuman primates, will change in the future with the introduction of Asian lineage strains. In Africa, ZIKV spills over from the zoonotic cycle to involve humans and transmission by both sylvatic vectors and Ae. aegypti, but is not recognized because infection is clinically inapparent, undifferentiated, and mild. Zika virus infection in Ae. aegypti is more efficient than YFV, with a shorter extrinsic incubation period (Cornet et al., 1979), and high rates of cumulative immunity (neutralizing antibodies) with age have been noted in human populations (LeGonidec et al., 1973; Monath et al., 1974). Thus it is likely that in many parts of Africa, women of child-bearing age have specific or heterotypic immunity resulting from exposure to multiple Flaviviruses that protect against congenital Zika syndrome.
Neutralizing antibodies to CHIK are found in 65%–100% of adolescent and adult rural inhabitants in areas of Africa and SE Asia where the disease is not recognized, indicating the presence of a profound endemic immune barrier preventing the appearance of epidemics (Monath et al., 1974; Nitatpattana et al., 2014). In areas where the virus is introduced at longer intervals, such as the Philippines, more than 50% of the population remains susceptible at any point in time allowing the periodic appearance of outbreaks that recur at intervals of about 17 years (Salje et al., 2016). It is no surprise that immunity determines the epidemiological patterns of Zika and chikungunya infections, and thus the future prospects of these viruses in the Americas will be largely determined by this factor. Subclinical infections contribute significantly to this background of immunity (Yoon et al., 2015).
In light of the these findings, Ferguson et al. predicted that the current epidemic of ZIKV disease in the Americas would subside within 3 years, would be characterized by low-level endemic transmission thereafter, and would not reappear in large epidemic form for about 10 years (Ferguson et al., 2016). This pattern is typical of other Flavivirus diseases whether they involve a zoonotic cycle or depend on interhuman transmission. A similar cyclical pattern dependent on the prevalence of immunity will characterize chikungunya outbreaks, which will be followed by quiescent periods until susceptible populations are reconstituted. In addition to immunity, other factors may contribute to re-emergence of epidemics, including climate change (Carlson et al., 2016), the Southern Oscillation effects on rainfall patterns (Paz and Semenza, 2016), the solar cycle, and enigmatic ecological influences. An example of the latter occurred in 1974–75 when there was an upsurge of Flavivirus epidemic activity on multiple continents (St. Louis encephalitis in North America, Rocio encephalitis in Brazil, yellow fever in Brazil and West Africa, and Murray Valley encephalitis in Australia) that would be difficult to explain without some global ecological perturbance. It is likely that such influences on arbovirus emergence will be identified and better defined in the future.
Both CHIKV and ZIKV are RNA viruses lacking proof-reading enzymes and susceptible to mutation and adaptation. This feature was exemplified during the Indian Ocean CHIK epidemic in 2005. The outbreak was caused by an East-Central-South African (ECSA) lineage virus that underwent a mutational change from alanine to valine at position 226 of the E1 envelope glycoprotein that increased infection efficiency (reduced extrinsic incubation period) for Aedes albopictus, the most prevalent vector in the affected Indian Ocean region (Chapters 2 and 8) (Tsetsarkin et al., 2007). As the epidemic spread northeast to India in 2009, a second mutational event (leucine to glutamine at position 210 in the E2 glycoprotein) further increased transmissability by Ae. albopictus (Niyas et al., 2010). These mutations had no effect on infection or transmission by Ae. aegypti, however, and where this vector was abundant the old Asian lineage of CHIKV had not undergone evolutionary pressure to adapt to another vector. The CHIKV strains introduced into the Americas in 2013 were of the Asian lineage (lacked the albopictus mutations) and were efficiently transmitted by New World Ae. aegypti. (It should be noted that recent studies have called into question the effects of these E glycoprotein mutations and showed similar vector competence for both virus lineages in Ae. aegypti and Ae. albopictus (Christofferson et al., 2014). In the case of vector adaptations for ZIKV, less is known but it has been suggested that loss of the N154 glycosylation site in the E protein is a possible adaptive response to transmission by Aedes dalzieli, a sylvatic vector (Faye et al., 2014).
An important question that must be addressed is, “Are there differences between CHIKV genotypes in virulence for humans (or susceptible animal models)?” A comparison of a Caribbean CHIKV with an Indian Ocean strain showed that the former had reduced in vitro replication fitness and reduced joint pathology and induction of proinflammatory cytokines in a mouse model (Teo et al., 2015). This seemed to fit with clinical observations that there was a higher proportion of asymptomatic human infections in the Caribbean than in the La Réunion outbreak; however, further observations to confirm and extend viral factors in virulence are needed. In future studies, it is likely that these can be best addressed by studying gene activation profiles and correlations with CHIK clinical manifestations in patients infected with different virus genotypes.
As described in Chapter 8, there are two genotypes or lineages of ZIKV—African and Asian—the latter evolved from the former following introduction from Uganda to Malaysia in ca. 1945 (Teo et al., 2015). As more isolates are studied from the Americas, minor changes in sequence have been observed (Wang et al., 2016), but it is unknown whether these play a role in pathogenesis, tropism, or transmission.
The vectors associated with transmission of ZIKV in Africa include multiple tree-hole breeding Aedes spp. that are also involved in YFV transmission between nonhuman primates and humans (Diagne et al., 2015). These species are not present in the New World, where only Ae. aegypti and Ae. albopictus are available for transmission. Neither ZIKV nor YFV efficiently infect Ae. albopictus, and in the Americas there is no selective pressure on ZIKV to adapt to transmission by this species. An unsettled question is whether the Asian ZIKV lineage underwent any genetic adaptation for efficient interhuman transmission by Ae. aegypti as a factor in the emergence of the pandemic. It is unlikely that any adaptation to Ae. aegypti was required, since even the original African virus is well suited to this mosquito (Cornet et al., 1979). Did the virus evolve a phenotype adapted to transmission by human hosts, perhaps altering viremia, neurotropism, the potential for congenital infection, shedding in urine and saliva, direct sexual transmission, and frequency of Guillain-Barre syndrome? Some comparisons between strains have been done in knockout mice (Stat2−/−) with compromised innate immunity showing strain differences in neurovirulence and spread to gonads (Tripathi et al., 2017), but these studies were not well controlled for mouse adaptation on laboratory passage. Future studies in animal models, as described in Chapter 10, and vectors may evaluate the effects on pathogenesis of individual sequence differences using infectious clone technology and site-directed mutagenesis. Could the virus adapt to more efficient transmission by mucosal routes, including aerosol transmission?
The pathogenesis of CHIKV (chronic effects on joint tissues), and of ZIKV [infection of placenta and neural progenitor cells, and persistent infection of (and damage to) testis] is a subject of intense study and details are beyond the scope of this chapter. Features of Zika—the high frequency of viral shedding, sexual transmission and persistence of virus in immunologically privileged sites, congenital infection and microcephaly, and parainfectious Guillain-Barré syndrome—were surprising in light of prior experience. However, all these features occur in natural and experimental infections in animals and humans with one or more of the following Flaviviruses: Japanese encephalitis virus—JEV (e.g., congenital infection, abortion, and reduced spermatogenesis in swine), St. Louis encephalitis virus—SLEV (congenital infection in mice and mental retardation in mice), West Nile virus—WNV (shedding in urine, Guillain-Barré, congenital infection, microcephaly in humans), Wesselsbron virus (congenital infection and arthrogryposis in sheep), tick-borne encephalitis virus (persistent infections in humans and monkeys), and YFV (persistent infection, mucosal shedding) (Monath, 1986). Exploiting some of these virus-host relationships in animal models may shed light on the pathogenesis of ZIKV and should be the subject of future research.
Establishment of an enzootic cycle of transmission of CHIKV or ZIKV in the Americas has obvious implications for the long-term control of these diseases. The primary transmission cycle of CHIKV and ZIKV in Africa is similar to that of YFV, involving sylvatic Aedes spp. and nonhuman primates, which have asymptomatic infections (Hanley et al., 2013). In tropical America, YFV is transmitted between neotropical monkeys (some of which develop fatal infections) and forest mosquitoes, of which Haemagogus spp. are predominant (with some Aedes spp. and Sabethes chloropterus as secondary vectors). This sylvatic cycle began relatively close in time to the introduction of YFV into the New World by humans during the slave trade 400 years ago, demonstrating conclusively that an arbovirus can escape from a human to a sylvatic cycle. To date, no studies have been reported on the susceptibility of Haemagogus mosquitoes (or even of Haemagogus cells in vitro) to ZIKV infection.
Althouse and colleagues modeled the numbers of monkeys and mosquitoes required to maintain a sylvatic cycle of ZIKV and estimated that 6000 primates and 10,000 mosquitoes would be sufficient (Althouse et al., 2016). Thus, spillover of ZIKV or CHIKV, for that matter, to wild neotropical primates is realistic possibility. There is an unconfirmed report of finding ZIKV RNA in a capuchin monkey in the state of Ceará, northeastern Brazil (Favoretto et al., 2016). Experimental infection of neotropical squirrel monkeys (Saimiri spp.) demonstrated viremia (Vanchiere et al., 2016). No other neotropical monkeys have been evaluated to date, and a high priority should be placed on this avenue of research. Ideally, YFV would be used as a positive control benchmark in such studies.
Aside from nonhuman primates, could other animals be involved in transmission cycles and amplification of ZIKV or CHIKV? This is unlikely in the case of Zika based on the ecology of closely related Flaviviruses like YFV and DENV that appear to be restricted in their host range to primate species. Experimental studies of ZIKV in multiple avian, mammalian, and some lower vertebrate species revealed only low levels of viremia in unlikely hosts, frogs and armadillos (Ragan et al., 2017). However, the host range of alphaviruses related to CHIKV, such as Ross River virus (RRV), Mayaro virus (MAYV), and Semliki Forest virus (SLEV), is quite broad involving marsupials, birds, and potentially bats. Experimental infection studies of CHIKV in North American mammals and birds revealed that big brown bats (Eptesicus fuscus) developed viremia (levels near the threshold of infection of mosquitoes) (Bosco-Lauth et al., 2016). Further investigation, particularly of neotropical vertebrates, is warranted in the future.
Althouse et al. laid out a series of priorities for research on the potential for sylvatic ZIKV cycles in the Americas (Althouse et al., 2016). The reader is encouraged to refer to this roadmap for future studies.
This is a subject for future research relevant to ZIKV pathogenesis and vaccine development. It is well accepted that enhancement of infection of Fc-receptor-bearing monocytes plays a role in the pathogenesis of dengue hemorrhagic fever. Enhancement of virus replication in vitro in the presence of subneutralizing levels of heterotypic antibody is easily demonstrated with all flaviviruses including ZIKV, but has been implicated only in the pathogenesis of dengue in vivo. The possibility that ZIKV antibody enhances DENV infection or that DENV immunity enhances ZIKV viremia (and thus disease and congenital infection) is strictly hypothetical. Crossreactive antibodies between ZIKV and DENV occur frequently (Keasey et al., 2017) and DENV-specific antibodies are known to enhance ZIKV replication in vitro (Dejnirattisai et al., 2016). However, in a study of passive transfer of ZIKV antibody from monkeys given experimental ZIKV vaccine, no enhancement of ZIKV replication was observed at levels of antibody that was subprotective (Abbink et al., 2016). In other words, ZIKV did not enhance itself. Moreover, there are many examples of Flaviviruses that crossprotect rather than enhance in vivo (including ZIKV and DENV prior infection protecting against challenge with YFV). Indeed, the incidence of ZIKV-associated congenital infection in Brazil appears to have been lower in yellow fever enzootic interior states, where YFV vaccination is practiced, than along the coast where vaccination is not practiced.
Future research will explore in vivo interactions of ZIKV, DENV, and YFV immunity. Ideally, this should include (i) determination of immune responses and clinical signs/symptoms in dengue immune versus naïve individuals who experience primary ZIKV infections; (ii) immune responses and clinical signs/symptoms in ZIKV-immune subjects who experience primary versus secondary dengue infections; (iii) immune responses, viremia, and adverse events in ZIKV immune versus naïve subjects who are vaccinated against DENV (i.e., with Dengvaxia) or with YFV 17D vaccine; and immune responses, viremia, and clinical signs and symptoms in yellow fever immune subjects who experience ZIKV infections.
There are multiple investigational vaccines against CHIKV and ZIKV in various stages of clinical development and there are also several antivirals in preclinical development. The detailed technologies being investigated are beyond the scope of this chapter. However, what can be considered is the future prospects for commercial vaccines and antiviral drugs.
The next few years will provide significant insights into the technical feasibility and commercial viability of general-use prophylaxis (GUP) vaccines and into the feasibility of antiviral interventions. In the case of vaccines, technical feasibility has been established for both CHIKV and ZIKV in multiple nonclinical studies and by analogy to commercially or clinically proven human and/or veterinary vaccines against closely related Flavi- and Alphaviruses. Less clear for CHIKV and ZIKV vaccines are (i) market size and profitability; (ii) the regulatory pathway to approval, which translates into the feasibility of demonstrating efficacy using either the traditional (demonstrating protection against disease) or accelerated approval pathway (defining an immune correlate and demonstrating it is reached). Market size (profitability) is probably more questionable for CHIKV than for ZIKV, because the consequences of CHIKV infection are generally self-limited or temporarily debilitating, whereas for ZIKV the risk of acquiring a mosquito-borne teratogen and the cost to individuals and society of such infections are much higher than for CHIKV. CHIKV may occur in outbreaks that are difficult to predict temporally or geographically; intervention with a fire-fighting (reactive-use campaign) vaccine may be difficult and delayed; and the risk: benefit ratio and pharmacoeconomics are not especially compelling. In contrast, prevention of ZIKV congenital infection is not dissimilar to cytomegalovirus (long a high priority for new vaccines) or rubella. Additional possible indications for a Zika vaccine are prevention of male infection, sexual transmission, testicular damage, and Guillain-Barré syndrome. These indications apply when ZIKV is transmitted at a low (endemic) level as well as during epidemics. For these reasons, it is highly likely that the next 10 years will lead to an active body of research on and clinical development of ZIKV vaccines. A significant issue to be resolved, however, is the regulatory endpoint for pivotal trials (prevention of Zika fever or of Zika congenital infection), the latter being much more difficult and costly.
Antiviral drugs will also be an area of active future work. There are a number of promising candidates for ZIKV antivirals. Intervention during the early stage of infection for women who are pregnant or possibly pregnant is an indication of considerable interest; nevertheless, treatment options are limited for compounds that would be safe for use in pregnancy. In the case of CHIKV, the evidence suggests that the chronic synovitis associated with CHIKV infections is due to an active infection and associated inflammatory responses in affected joints. Quality of life is reduced for months to years after the acute phase of chikungunya due to these complications. Thus, one can envision antiviral treatment during the acute disease to prevent the complication of chronic synovitis and arthritis. Indeed, antiviral treatment combined with anti-T cell therapy (e.g., CTLA4-Ig) provided benefit in a mouse model of chronic chikungunya arthritis (Miner et al., 2017). The pathogenesis of this disease is still unclear, however, and documentation of acute or persistent viral infection of joints and an immunopathological response are reasonable hypotheses to be tested by future research (Goupil and Mores, 2016). As mentioned earlier, studies on gene activation to clarify biomarkers of immunopathology of arthritis will be instructive.