key: cord-0008260-15zfvsur authors: Novella, Isabel S.; Domingo, Esteban; Holland, John J. title: Rapid viral quasispecies evolution: implications for vaccine and drug strategies date: 2004-11-02 journal: Mol Med Today DOI: 10.1016/s1357-4310(95)91551-6 sha: 630d1a0b0838908c8f5974b320847fdb9360489a doc_id: 8260 cord_uid: 15zfvsur High mutation rates occurring during replication allow RNA viruses to evolve rapidly and adapt continuously to new environments. This poses an enormous challenge to vaccine and drug development which, to be effective, must consid RNA virus variability and follow approaches that minimize the probability of escape or resistant mutants arising. RNA viruses are responsible for a large number of human dise~ such as polio, influenza, rabies and many others, including 'newly emerging' infections such as hantavirus hemorrhagic t and Ebola (see Table 1 ). The evolution of RNA viruses is strongly determined by remarkable variability. Extremely high mutation rates (10"3-10 "5 incorporations per base copied) 2 and both homologous and nonho ogous recombination allow extensive generation of new sequen Most mutations will be deleterious and eliminated by natural s tion, but many have selective value in the particular environme which the virus is replicating, and viruses harboring such muta may become dominant in the population. The continuous gener of variants during replication, independent of their selective v produces a complex distribution of nonidentical genomes te: quasispecies 4. Many factors can influence the long-term evolution of RNA vir one of which is population size. Thus, darwinian selection is us the driving force during transmission of large virus population~ during replication of viral quasispecies 5, whereas random dr Cryptic immunopathological responses -Hidden/subclinical tissue damage resulting from an inappropriate immune response to the pathogen. Defective interfering particle -Virus particles produced during an infectious cycle that are not able to replicate themselves. Escape mutants -Viruses that are resistant to the effect of drug or antibody as a result of sequence changes within the site of action of the antiviral agent. Homologous/non-homologous recombination -Transfer of genetic information from one genome or chromosome to another during DNA replication which is initiated by the pairing of complementary/non-complementary sequences within the genome. Live-attenuated virus vaccine-Vaccine based upon a replicationcompetent virus that has been crippled or weakened (usually by empirical methods such as repeated passage in tissue culture or the non-natural host) to reduce its pathological effects. Plantigens and plantibodies -Viral proteins or antiviral antibodies produced in genetically modified plants. Random drift -Gradual accumulation of mutations within the genome. Replicons -Unit of nucleic acid that contains all the sequence necessary for replication. Subacute sclerosing panencephalitis -A rare fatal disease of children or adolescents years after measles infection, resulting from slow replication of defective measles virus within the brain. Tautomers -A compound that differs from its isomer, with which it readily interconverts, only in the positions of a mobile atom, frequently hydrogen, and of double bonds. Virus reassortants -Novel progeny virus formed by random association and packaging of genome segments during mixed infection of a cell by different RNA virus species. prevalent during genetic bottlenecks 6'7. Natural selection is extremely effective. Extensive genetic heterogeneity has often been found not only among infected individuals during an outbreak, but also within single hosts or even particular tissues. Selection of drug-resistant mutants in vivo during treatment strongly suggests that specific adaptation does occur rapidly, as has been shown in cell culture 8. Obviously, evolution will depend on the environmental conditions under which viral populations replicate. These environmental conditions include, among others, the immune response of the host, antiviral intervention, fever and hormone status. Consequently, effective drug and vaccine development should focus on approaches that induce a broad variety of defenses, so that the probability of generating mutants that can escape all of them (i.e. multiple mutations in a single genome) is minimized. For vaccines, formulations based on single or a few viral peptide sequences (i.e. a limited number of epitopes) often induce good responses against the homologous sequence, and usually have the advantages of safety and stability. However, achievement of adequate levels of protection in a population will be problematic. Some vaccine trials of HIV and foot and mouth disease virus (FMDV) are showing this problem 9,t°. As a result of the changing and unpredictable dynamics of RNA virus quasispecies, these vaccines may not always prevent infection by related virus isolates, nor prevent selection of escape mutants during primary infection n. Peptide vaccines for RNA viruses have sometimes been successful when challenge or reactivity tests were performed with homologous strains, but results with heterologous strains have frequently given disappointing results. Examples are HIV (Ref. 12) and hepatitis B virus t3. No peptide vaccine is currently licensed to prevent diseases caused by RNA viruses. Subunit vaccines (i.e. vaccines consisting of viral proteins) have the same kinds of problem as peptide vaccines, and they are often poorly immunogenic. Efforts focusing on the use of a cloned HIV envelope glycoprotein in a vaccine for AIDS have faced this problem and other approaches must be examined t4. Additional concerns are stability and the possibility of denaturation during preparation, and the induction of an abnormal immune response leading to enhancement of disease after infection. This has been described for respiratory syncytial virus (RSV) (Ref. 15 ). One subunit vaccine that has been rather succesful is the vaccine for feline leukemia virus (FeLV), but escape mutants have already been reported '6. The problem of viable escape mutants is also influenced by factors other than mutation rates. Viruses that replicate very rapidly will be more likely to generate and select such mutants during an infection although the immune response clears many millions of infectious particles per day. This occurs with HIV ( Refs 17, 18) . In addition, viruses that can cause immunosupression or establish persistent infection (particularly when they can avoid displaying viral antigens at the cell surface), such as lymphocytic choriomeningitis virus, are difficult to control. Nevertheless, the most important factor that will affect the generation of viable escape mutants is the nature of the virus itself. Thus, different viruses show different antigenic epitopes, and each epitope may or may not tolerate extensive substitutions. Substitutions that confer antibody resistance to some conserved epitopes will be deleterious and escape mutants will be unable to survive selection. This explains the lack of escape mutants for successful RNA virus vaccines such as those for poliovirus, measles and mumps. Conversely, the more variable and adaptable antigenic epitopes in viruses such as FMDV and HIV-1 pose much greater challenges ( Fig. 1) . At the other end of the spectrum, live-attenuated virus vaccines inevitably include not only an ample repertoire of B-cell and T-cell epitopes, but also a broad spectrum of mutations, because they consist of diverse quasispecies populations. As attenuation may limit, but not prevent, viral replication, the number of sequences that the immune system encounters becomes even more extensive. Most of the successful vaccines against RNA (and DNA) viruses have been live-attenuated viruses, including those used to control poliomyelitis, smallpox, measles, rubella and mumps. Promising results have also been reported for RSV and parainfluenza virus 15. However, the use of these vaccines inevitably involves some risks. Use of live RNA virus always allows some possibility for virulent revertants to arise by mutation and/or by recombination of attenuated precursors, a phenomenon described for poliovirus ~. This problem might be circumvented by developing strains with a very high detrimental mutational load distributed along the genome and, of course, by in-depth knowledge of the factors that contribute to attenuation in vaccinees. The selection of viral strains to develop a vaccine is a critical, difficult and uncertain enterprise. Cloning of virus from natural isolates, or cell-culture replication of such isolates in the laboratory, is likely to select for sequences that are different from those in circulating strains, either because of random sampling or natural selection (see above). This seems to be happening with HIV vaccines 12. During the preparation of vaccine batches, the enormous ability of RNA viruses to recover fitness (and possibly virulence) during massive replication 5 must always be kept in mind. Studies of vaccine safety must in addition include studies of differences among age/sex/ethnic groups of vaccinees. Simian immunodeficiency virus deletion mutant vaccines developed as a model for HIV (Ref. 21) seem to be safe for adult monkeys, but Ruprecht and co-workers have found that they are not safe for newborn monkeys after mucosal infection n. Another potential problem is the rare mutational generation of genomes that can produce atypical infections or cryptic immunopathological responses, an effect observed for several viruses, including defective measles viruses, causing subacute sclerosing panencephalitis 23. Finally, the potential for inapparent (but damaging) long-term persistent infections and/or proviral DNA integration and host-cell chromosome alterations must be considered. Live vaccines for HIV or hepatitis C virus might have this potential. Some successful vaccination programs will lead to virus eradication but others may leave a fraction of the population susceptible or cryptically infected. For certain viruses, the existence of animal reservoirs can provide a continuous source of infection. Under these circumstances, RNA viruses replicating in susceptible hosts (or vectors) could occasionally generate new antibody-es~pe variants that can infect a previously immune population. A very well-known example of such behavior is the rare production of influenza A virus reassortants (Fig. 2) carrying segments from fowl and human influenza genomes. These include new epitopes that are not recognized by previously induced immune responses in the human population and they initiate human pandemics. This might also occur with a ntt0aber of other RNA viruses 24. Thus, during disease outbreaks, considerable effort must focus on the isolation and characterization of evolving strains involved in the disease process, and vaccine producers should be prepared to 1o replace formulations that may have become less effective . This is acknowledged by the annual review of prevalent viral subtypes included in vaccines against influenza. These need to take into account changes resulting from hypervariability in some epitopes (Fig. 3) . RNA virus variation is a constant hurdle for the development of effective antiviral drugs. As for vaccine design, drug therapies must deal with the rapid appearance of resistant virus mutants, but, with antiviral drugs, the problem is aggravated. Usually, adequate vaccination will induce an immune response against a variety of epitopes and different variants of each epitope, so that escape mutants must necessarily include multiple appropriate, non-debilitating mutations. Accordingly, their appearance is much less probable. However, most antiviral drugs exert their functions on protein domains that usually involve a very limited number of residues (for example, the catalytic site or contiguous sites of enzymes). The direct consequence is that resistant viruses are often single point mutants, and the probability for rapid takeover by such mutants is high. The use of monoclonal antibodies in the treatment of RNA virus infections is impaired for the same reason. There are several examples in the literature of resistant mutants being isolated after only weeks or even one or two days of antiviral treatment 24. Furthermore, such resistant variants can be present and represent a significant fraction of the quasispecies population even before treatment "~6. One way to decelerate rapid generation of escape mutants might be to use combination therapies directed toward different steps of the virus life cycle. This might be achieved by the use of antiviral drug cocktails or by the use of drugs with multiple targets. The anti-HIV ribozyme approach developed by Wong-Staal and colleagues 27 might provide one such antiviral therapy if a number of target sequences can be included. Another novel approach might involve attempts toward specific manipulation of RNA virus mutation rates. These are normally within the narrow range that provides optimal biological adaptability (near to the error threshold). Much lower mutation rates could minimize the appearance of mutants resistant to other antivirals, while higher mutation rates would cause the quasispecies to exceed the error threshold and to lose meaningful genetic information (mutational meltdown) 4. Eventually, rational drug design approaches based on knowledge of the structure and function of viral polymerase domains might allow the development of virus-specific mutagenic drugs. Viral kinases are another potential target for design of specific ribonucleotide analogs that can be efficiently phosphorylated only in virus-infected cells. The first widely employed antiviral compound, iododeoxyuridine, was developed in the late 1950s. It is mutagenic due to increased formation of rare tautomers, and abundant viral kinases promote its preferential phosphorylation within herpesvirus-infected cells ~. As we enter the next millenium, specific drugdesign approaches will surely be capable of producing much more specific antivral mutagens. Vaccination campaigns and antiviral drug therapy are new selective forces shaping viral evolution. Although relative success has been achieved, RNA viruses have shown an inmense ability to respond to such changes in selective pressures. Many other relevant selective changes are rapidly occurring in our biosphere, largely due to human activities and human population growth. This not only affects the evolution of viral quasispecies in natural hosts, but also provides new opportunities for RNA viruses to expand their diverse quasispecies populations, or even to jump to new niches and new hosts. Novel virus diseases are emerging and more will emerge in the future, as some old ones disappear (and sometimes reappear). Any attempt to control and eradicate RNA viruses must deal with the fact of extreme variation and rapid evolution. The better we understand RNA viruses as complex quasispecies populations, the closer we will come to effective measures for their control. • Can vaccine protection against RNA viruses be provided by direct intramuscular injection of DNA preparations engineered to produce highly mutagenized mRNA (and diverse proteins)? • Can intestinal colonization with engineered avirulent bacterial vectors expressing multiple viral antigens provide effective vaccination? • Can multiple diverse viral proteins, expressed in dietary vegetables (plantigens) provide protection against intestinal (and other) RNA viruses? Can plantibodies? • Can improved technology with human hybridomas eventually produce cost-effective, concentrated, highly diverse polyclonal antiviral antibody preparations for use in Gammaglobulin prophylaxis and/or treatment? • Can we use the principles of interference by defective interfering particles for the design of antiviral replicons which replicate and interfere only during virus infections? • How will RNA virus evolution be affected by widespread employment of many new vaccines, antiviral drugs and other antiviral therapies? • How can we employ diversity to combat the problems of RNA virus diversity? Rates of spontuneous mutations among RNA viruses Mutation rates and rapid evolution of RNA viruses The Hypercycle. A Principle of Natural Self-Organization Genetic bottleneck and population passages cause profound fitness differences in RNA viruses Fitness of RNA virus decreased by Muller's ratchet Rapid fitness losses in mammalian RNA virus clones due to Muller's ratchet Viral drug resistance The immune response to HIV: implications for vaccine development Complications of RNA heterogeneity for the engineering of virus vaccines and antivirai agents Dilemma of neutralization resistance of HIV-1 field isolates and vaccine development In vitro T cell immune response to the PreS2 antigen of the hepatitis B virus envelope protein in preS2+S vaccine recipients. Absence of cross-reactivity of subtypes at a major T ceil recognition site The next step toward a global AIDS vaccine An update on approaches to the development of respiratory syncitial virus fRSV) and parainfiuenza virus type 3 (PIV3) vaccines The HIV vaccine dilemma: lessons from the cat Rapid turnover of plasma virians and CD4 lymphocytes in HIV infection Viral dynamics in human immunodeficiency virus type 1 infection Antigenic oscillations and shifts in immunodominance in HIV-I infections Genetics of poliovirus Protective effects of a live-attenuated SIV vaccine with a deletion in the nefgene Pathogenicity of live attenuated SIV after mucosal infection of neonatal macaques Mutations and A/I hypermutations in .measles virus persistence RNA virus quasispecies: significance for viral disease and epidemiology Structural basis of immune recognition of influenza virus hemagglutinin pol gene quasispecies of human immunodeficiency virus: mutations associated with drug resistance in virus from patients undergoing no drug therapy A hairpin ribo~me inhibits the" expression of diverse strains of human immunodeficiency virus type 1 Antiviral agents, in Hrology Acknowledgements. We thank Estelle Bussey for excellent technical assistance. Work in Madrid was supported by Fondo de Investigacioaes Sanitarias de la Seguridad Social grant n 93/0003-01, Direccion General de Investigacion Cientifica Y Tecnologica grant PB91-0051 and Fundacion Ramon Areces. Work in La Jolla was supported by National Intitutes of Health grant AI14627. I.S. Novella is a posdoctoral fellow of CIRIT (Generalitat de Catalunya).