The ecology and evolution of the common coldThe failure to cure the common cold has become almost a clich´e about how science, and medical science in particular, can be defeated by the simplest of problems. Common colds are utterly familiar to all, and have predictable and annoying, but rarely dangerous, symptoms. Their apparent simplicity, however, masks a deeper complexity that takes us to the heart of fundamental issues in ecology, epidemiology and evolutionary biology. The diversity and ubiquity of the common cold, particularly when contrasted with its more lethal close relatives, provides a natural laboratory for deriving and testing mathematical models linking these three areas.
Common colds are caused by at least nine unrelated groups of viruses. Of these, the most common and among the least severe is the rhinovirus, cause of about half of all colds. Rhinovirus itself has a staggering level of diversity, boasting over 100 different serotypes (viral types that react to different antibodies). These serotypes are as genetically different from each other as are different species of animals. This diversity is partly a consequence of extremely high viral mutation rates. A rhinovirus makes on average one copying error each time it replicates its tiny genome consisting of only 8000 base pairs. Two observations, however, suggest that mutation is not the sole cause of diversity. First, the known serotypes have remained stable since first identified over 50 years ago, hardly what would be expected after decades of mutation. Second, many serotypes are present simultaneously even in fairly small towns, which may host 80 different serotypes during a single “cold season.”
These characteristics are particularly striking when contrasted with those of influenza. Many serotypes of influenza have infected humans during the past century, but these replace each other quickly over time. Only one or two circulate simultaneously, and physicians are kept guessing until the last minute about which new type will dominate. Influenza is a much more severe disease, sometimes endangering the lives of healthy young adults. Finally, influenza outbreaks peak in mid-winter, while rhinovirus prevalence peaks in fall and spring.
Explaining these characteristics requires looking at viruses at many scales simultane- ously: the biochemistry that determines how viruses infect host cells, the systems biology of interactions with the immune system, the evolutionary ecology of viral competition within a single host, and the dynamics of virus abundance in entire populations. This linking across scales of interaction characterizes the revolution of modern biology, and rhinoviruses provide an ideal but understudied system for tracing these links. Two tools have made this tracing possible, unifying biological sciences across scales: molecular genetics and mathematics. These are the tools we will bring to bear in this study.
The combination of genetics and mathematics has revolutionized our understanding of evolutionary processes, whose signature is writ large in the high diversity and avirulent infection characteristic of rhinoviruses. The selective forces promoting these traits come into clear focus in the context of their relatives. Rhinoviruses are picornaviruses, a large family that contains a remarkably diverse cast of characters, including the three serotypes of polio, the seven of foot-and-mouth disease, and the single serotype of hepatitis A. Rhinoviruses are most closely related to the 70 serotypes of human enterovirus (which infect the gut) which include the polioviruses and several dangerous Coxsackie viruses that damage the heart or cause diabetes. Rhinoviruses have arisen at least three times from the enterovirus group, with 76 classified as type A, 25 as type B, and one that remains genetically nearly identical to current enteroviruses.
The vast difference between the infections caused by the closely related rhinoviruses and polioviruses result from differences at the smallest scale, that of biochemistry. The proteins that coat rhinoviruses have lost the ability to persist in the acidic environment of the gut, or even tolerate the body’s core temperature, and are thus restricted to the nasal passages. The viral coat proteins also orchestrate the complicated process of attaching to particular receptors on host cells, and control the delicate shape changes required for the virus to inject its genetic material into the cell and begin an infection.
One would expect these delicate processes to be conserved by evolution. However, changes also create new opportunities for the virus. The virus coat proteins are precisely those most visible to the immune system, and changes can allow evasion of immune memory. In addition, these proteins determine which cells are attacked, and thus the severity and course of infection. The three polioviruses are the only picornaviruses that attach to one particular receptor on cells, and as a result share the ability to create neuroinvasive disease. All 25 type B rhinoviruses use another receptor, one that is common on cells in the upper respiratory tract. Strangely, most of the 76 type A rhinoviruses use the type B receptor, while a minority use an unrelated receptor, showing that receptor switches can occur relatively quickly.
The rhinoviruses thus combine complexity over a huge range of scales, and it is precisely the interlocking levels of complexity that may provide the key to understanding them. At the biochemical level, viruses interact in precise and subtle ways with host cells. At the tissue level, rhinoviruses interact with the immune system. At the evolutionary level, the human species is beset by a cloud of these tiny beasts and, thanks to their high mutation rates, each infected person has within them a unique cloud of viruses with different genetic sequences. At the ecological level, these viruses transmit themselves through populations structured by age, location, and immunological history, creating seasonal and spatial patterns of abundance. A change of one nucleotide (one letter in the 8000 letter genome of a rhinovirus) could alter how a virus binds to cells, which cells it can bind to, how it is recognized and attacked by the immune system, how successfully it transmits itself to new hosts, and how successfully it spreads through an entire population.
These properties make rhinovirus an ideal system for unraveling complexity. Thanks to their tiny genomes, viruses can be efficiently sequenced. Thanks to their abundance and high diversity, informative samples can be collected at a local health clinic or the neighborhood supermarket. These genetic data, supplemented by published genetic and epidemiological data, provide the basis for synthetic mathematical models. Although our preliminary models make fairly accurate predictions of the serotype diversity observed in the most extensive published study, they fail to predict the evolutionary stability of these serotypes over time. We will refine these models to predict serotype diversity based on the full complexity of these viruses: seasonality, high mutation rate, and interactions with individual cells and the immune system. The models can be tested on both type A and type B rhinoviruses, and promise insight into the closely related picornaviruses with radically lower serotype diversity (such as hepatitis A).
In addition to modeling biodiversity, we will mathematically model the evolutionary ecology of these viruses, focusing on the relationship between high mutation rates and mild symptoms. The genetically diverse viruses in a single host compete for cells, but may also cooperate in their evasion of the immune system. Because viruses share their environment with closely related but non-identical “conspecific” individuals, they have the potential for complex evolutionary interactions.
Finally, we will use mathematical models and modern phylogenetic methods to reconstruct the history of cell receptor changes in the picornavirus family. Evolution experiments in the laboratory have shown that viruses placed under strong selection can switch receptors in a matter of weeks, although almost certainly at a cost. Our models will estimate the frequency of switches, and thus the consequences for further evolutionary innovation and virulence. We can then return to consider current rhinovirus biodiversity as a snapshot of a long-term evolutionary process.
The common cold, often neglected in medical studies as little more than a nuisance, thus provides a unique window into viral evolution, and can be used to answer fundamental questions about ecology, epidemiology and evolutionary biology. Why are there so many recurrent types of rhinovirus? Among closely related viruses, why are some deadly and others almost harmless? Why are certain portions of evolutionary space (particular genetic sequences) off-limits? Rather than addressing these apparently very different questions in isolation, the rhinoviruses and their close relatives beckon us to address them simultaneously with the unifying methods of genetics and mathematics, linking these questions in a conversation that illuminates both micro and macro evolutionary processes.
