Skip To Content Skip To Left Navigation
NSF Logo Search GraphicGuide To Programs GraphicImage Library GraphicSite Map GraphicHelp GraphicPrivacy Policy Graphic
OLPA Header Graphic

Dr. Colwell's Remarks


"Cholera and the Environment: A Classic Model for Human Pathogens in the Environment"

Dr. Rita R. Colwell
National Science Foundation
Symposium: From Outside to Inside:
Environmental Microorganisms as Human Pathogens
American Association for the Advancement of Science Annual Meeting,
Seattle, Washington

February 14, 2004

See also slide presentation.

If you're interested in reproducing any of the slides, please contact
The Office of Legislative and Public Affairs: (703) 292-8070.

[title slide: Cholera and the Environment: A Classic Model for Human Pathogens in the Environment]
(Use "back" to return to the text.)

Good afternoon. I'm very pleased to help open this symposium exploring how pathogens interact with their environment. Jerry Cangelosi deserves special kudos for his organizing role—thank you, Jerry! Just the fact that this symposium is taking place is clear recognition that we are moving beyond the old reductionist model of viewing a pathogen in isolation, developing a much richer understanding of a microorganism in all its complexity. As John Donne might have said, "No pathogen is an island, entire of itself; every pathogen is a piece of the continent, a part of the main."

I've entitled my talk, "Cholera and the Environment: A Classic Model for Human Pathogens in the Environment." We have come a long way to the integrative point of view that Vibrio cholerae, the causative organism of the disease cholera, is a normal component of the ecosystem—an organism that can never be eradicated but only controlled. Although I can now use the word "classic" when describing the story of cholera, I can recall when the viable but non-culturable form of Vibrio cholerae—the dormant state that cholera can assume in the environment—was referred to by some as "Colwell's ghosts."

The connections between cholera--an ancient water-borne disease--and the environment provide a paradigm for looking at many organisms with multi-dimensional biographies, including others featured in today's symposium. A fully dimensional understanding of an infectious disease, whether cholera, Hantavirus, or malaria, now reaches from countries to continents and beyond, and enriches medicine with insights from across science and engineering.

[earth from space with cloud veil]
(Use "back" to return to the text.)

I would like to discuss the global context for a moment, and then move to some specific case examples of pathogens in the environment, concluding with some new observations on the case of cholera. A global context indisputably frames all human health issues in the 21st century. This context is comprised of several realities: the worldwide movement of people and goods, the new recognition that earth processes operate on a global scale, and a dynamic international scientific enterprise.

Health issues are no longer just a matter between patient and physician—if they ever were—but now encompass an individual's complex relationship with the global environment. As Gro Harlem Brundtland—former director of the World Health Organization—has said, "In the modern world, bacteria and viruses travel almost as fast as money. With globalization, a single microbial sea washes all of humankind. There are no health sanctuaries."

[Frequent flyers]
(Use "back" to return to the text.)

The graph toward the bottom shows how international travel has skyrocketed in the past half century, up to almost 500 million international arrivals per year, or more.

We can see, on the world map at the top, where all these travelers are arriving. The routes circumscribing the world map are the most popular air routes between continents. Note the large green arrows at the right of the map, which give the percentage of increased arrivals over the mid-1990s in different locations. International arrivals increased in every region, but in Africa and the Middle East, they jumped by almost half.

Our world—the world of infectious disease, the world of research, the social world—is now integrated and global. These connections make simplistic the notion that we can successfully eradicate a disease from the face of the planet. Infectious disease is a moving target—as climate shifts, any disease with an environmental stage or vector will be affected.

As we recognize signals from climate models and incorporate them into health measures, new opportunities arise for proactive—rather than reactive—approaches to public health. I'll turn now to a brief survey of a few cases of infectious diseases in their ecological contexts.

[photo: pitcher plant cluster]
(Use "back" to return to the text.)

Ecology has immediate lessons for epidemiology, and the National Science Foundation has supported studies that bring the two together. Take the mosquito that lays eggs in these beautiful North American carnivorous plants called pitcher plants. Although not a disease vector, this mosquito's evolution bears lessons on how vector-borne diseases can spread as climate shifts.

[mosquito ovipositing in pitcher plant]
(Use "back" to return to the text.)

This mosquito uses day-length to regulate seasonal development, as explained by William Bradshaw and Christina Holzapfel of the University of Oregon, who study this phenomenon. Mosquito populations have now adapted to the climate of North America from Florida to Canada. Disease-carrying invaders, like the Asian tiger mosquito, must similarly adapt to cold and to different day-lengths. As spring comes earlier and growing season has lengthened over the latter half of the last century, the pitcher plant mosquito has adapted to shorter photoperiods, especially in the north.

(Use "back" to return to the text.)

In a rare example of a documented genetic shift due to warming, this graph shows, as latitude increases, the mosquito's genetic shift to shorter photoperiods (that is the left axis, in hours). Bradshaw has detected this genetic shift over as short as five years; as he says, "This is evolution at breakneck speed!"

Many of you will be familiar with another story that involves climate and a natural disease vector--the case of the Hantavirus, which was unknown in the New World until the 1993 outbreak in the Four Corners area of the United States. The mortality rate of those infected with the virus was 70% in the first few weeks. The carrier turned out to be a rodent.

Biologists working at an NSF-supported Long-Term Ecological Research site at Sevilleta, led by Terry Yates of the University of New Mexico and his team, were able to detect the deadly virus in mouse tissue that had been archived years before. Some of the study sites are shown here.

In addition, the investigators showed a link between climate and the outbreak of disease. Mild and wet winters associated with a periodic climate pattern, El Nino-Southern Oscillation, had provided more food for the rodents, whose populations had increased dramatically in 1993.

In this graph, the black line shows rodent density. The red line shows density of rodents infected with the hantavirus. The green bars plot human cases. A take-home message is that the key predictor of disease cases is not the increase in numbers of rodents, but the increase in infected rodents (in red).

The researchers are now pursuing a "trophic cascade hypothesis," developed to explain the changing levels of human risk for zoonotic diseases associated with climate variability. The cascade of disease through the trophic levels is apparently set off by El Nino.

I turn now to another infectious disease—campylobacter—which emerged as a leading cause of gastroenteritis in some countries about a quarter century ago. Public health interventions, focused on food-borne transmission, have not decreased disease incidence. In fact, we see--in the graphs at right—an annual rise over the ten-year period. Also, the disease has displayed a striking and consistent seasonal pattern. My student Valerie Louis observed a significant correlation between increased temperature and the seasonal peak of campylobacter infections in England and Wales.

These figures show campylobacter incidence by district during the 1990s. At left are annual cases per 100,000 people. At right are the annual rates in different districts.

The red line shows the weekly incidence rate for children aged 1-4; note the sharpness of the peak for that age class compared to all the other age classes. The seasonal effect, or sharpness of the spikes, becomes less pronounced in the older age groups. Such seasonality related to age has not been discerned for this disease before. Certainly, children under five are a prime target for public health measures.

The bottom graph shows that males—the blue line—have a higher rate of infections than females, no matter what the age group.

More broadly, campylobacter infects humans through a wide array of ecological pathways, forming a rich canvas for applying the biocomplexity approach to public health policy.

Tularemia is another disease with very complex environmental links. This graph shows an epidemic pattern of tularemia in Sweden, where the disease has been endemic for 70 years. Here, airborne tularemia was contracted from contaminated hay when rodent populations increased, then died off. (I am grateful to Andrew Pearson from Britain's Health Protection Agency for sharing his insights and graphics on tularemia.)

Tularemia actually occurs throughout the Northern Hemisphere and even beyond, having caused a million human cases over the past 70 years. Seven different forms—species, subspecies and biogroups—are recognized.

This plague-like disease occurs in more than 100 wild mammals, as well as in birds, insects, and humans, although human infections are really just a by-product of the organism's interactions with its environment. Like cholera, a couple of tularemia forms occur in water, and waterborne outbreaks have resulted from contaminated drinking water. A hypothesis is that tularemia is water-borne, with muskrats, hares and the like as the reservoirs in nature.

European forms of tularemia are spread through a greater variety of hosts and vectors than in North America. This table gives a snapshot of some animals from which the organism was isolated in Jamtland, Sweden. Tularemia in Scandinavia occurs across flood land, stream and tundra ecosystems.

A study of the then-Soviet Union showed that a great diversity of animal vectors are associated with different forms of tularemia, in different geographic environments and in different seasons. Russian epidemics have mainly been water-, vector-, and airborne. The region saw 100,000 cases a year in the 1940s.

This map indicates where tularemia has been found in North America (the line marks the lowest latitude at which epidemics or epizootics have been found). The North American form, Francisella tularensis subspecies tularensis, is more virulent to human beings than its European and Asian counterparts.

Tularemia was actually first recognized in squirrels around San Francisco in 1911. Later, water transmission was established. Tularemia cases in the US have decreased, numbering just 200 per year since 1960. Recent US outbreaks occurred from captive prairie dogs, as well as two outbreaks on Martha's Vineyard and pulmonary cases in South Dakota.

[the case of cholera]
(Use "back" to return to the text.)

From campylobacter to hantavirus to tularemia, we are beginning to trace the connections between a pathogen and its environment. I will turn now to some new results from my own research on the case of cholera.

[Large chromosome]
(Use "back" to return to the text.)

The 50 toxin genes of Vibrio cholerae reside on its large chromosome, seen here on the right. Sequencing data confirm that V. cholerae is a versatile organism, able to live in several habitat types, as well as to infect the human gastrointestinal tract.

Lateral transfer of genetic material is clearly occurring in this organism. Virulence genes are distributed in environmental strains of V. cholerae from various serogroups—apparently an environmental reservoir of such genes. We know that 0139, the new serogroup, acquired novel DNA from other cholera bacteria in its environment. This underscores the versatility of cholera, and gives even greater credence to the significance of the environment in understanding cholera's complexity. We also know, however, that it is difficult to isolate V. cholerae 01 from the environment, where it is competing with some 250 other serotypes.

In fact, bacterial viruses have been found to play an important role in the microbial ecology of aquatic ecosystems. Lysogeny, integration of the phage into the host bacterium, occurs more frequently in 01 El Tor and 0139 strains than in 01 classical strains, and is believed to impart antibiotic resistance to the host bacterium. Phage infection may give rise to new toxigenic variants, and polylysogeny can occur, as Erin Lipp, a postdoctoral fellow in my lab has shown.

In that recent study, Erin, Eric Espeland, and I observed classical and El Tor strains of cholera infected with a temperate phage. At least three prophages (inserted DNA) were found to exist in one strain of El Tor. Should such a mechanism operate in the wild, genetic material such as toxin genes could be transmitted in the environment via multiple temperate phages.

[slide not available]
(Use "back" to return to the text.)

Cholera's interactions with its environment are another major line of inquiry. The ecological relationship of cholera and planktonic copepods was first established in 1983 when Anwar Huq and I showed that V. cholerae attach to live copepods in the Chesapeake Bay and around Bangladesh. This strongly suggests that cholera is probably not an eradicable disease, because its causative organism lives naturally in riverine, brackish and estuarine ecosystems.

Currently, my student, Tonya Rawlings at the University of Maryland, is investigating whether the seasonality of cholera in endemic areas is associated purely with temperature and salinity, or whether V. cholerae's interactions with given species of hosts—namely these copepods—also affect its seasonal abundance.

At present, she is investigating whether two epidemic variants of cholera, 01 and 0139, may have preference for attaching to specific genera of copepods. Here we see three copepod genera: Oithona, Eurytemora, and Acartia. Not all bacteria compete well for space on surfaces. The pathogenic species of V. cholerae attach best. Here is another example of cholera's environmental capabilities, and further evidence of its autochthonous aquatic nature.

Eurytemora is more prevalent in the upper Chesapeake Bay, with more Acartia in the lower Bay. We have also noted seasonal fluctuations in the prevalence of these species. We need to do intensive work in various environments to see if copepod species selectivity—for example, the ability of Vibrio cholerae 01 to attach better to a particular copepod species—might be linked to a more severe cholera epidemic when that particular copepod is more abundant.

Adhesive ability is an important attribute of V. cholerae, whether in the environment or in the human gut. The "environmental" capabilities of V. cholerae include its ability to secrete a powerful chitinase, which assists its growth on chitin surfaces. Besides colonizing copepods, V. cholerae is also present in shellfish.

Our hypothesis is that cholera originally evolved commensally with marine animals such as copepods, which provided them a surface to grow, nutrition and perhaps other mutual benefits. In the ocean ecosystem generally, the breakdown of chitin is an important ecological function. Another interesting capability is V. cholerae's production of a mucinase, which enables it to penetrate the mucus barrier that covers the gastrointestinal epithelium.

Carla Pruzzo of the Universita Politecnica delle Marche has reported preliminary data that serum from Mytilus hemolymph, a mussel, increases attachment in the intestinal epithelial cells. The upshot is that when Vibrio is ingested with seafood, cholera acquires "bridging molecules" which make cholera very adhesive in the intestine. As Pruzzo explains, "Both virulence and infectivity depend on both bacterial properties and environmental factors."

[3 graphs from Estelle]
(Use "back" to return to the text.)

One more point about copepods: Here we see the results of a three-year study of ponds in Matlab to see if plankton numbers or other measures of water quality could predict the presence of Vibrio cholerae as detected by fluorescent antibody. The top graph shows copepod nauplii (a young lifestage); the center shows copepod adults; and the bottom is V. cholerae. Copepod nauplii and adults, whether used singly or with other variables, were the best predictors of V. cholerae's presence.

[biofilm "cartoon"]
(Use "back" to return to the text.)

This artist's rendition shows another environmental function: the formation of a bacterial biofilm. John Mekalanos of Harvard University has studied how biofilms enhance colonization in V. cholerae. This protective clustering behavior is yet another mechanism for cholera's life cycle in the aquatic environment, which also serves to protect it from stomach acid when ingested by humans.

Biofilm formation is an intriguing part of the complexity of cholera's lifestyle. It is one of the many bacterial processes—along with production of virulence factors and bioluminescence—that are regulated by the special bacterial communication called quorum sensing.

[V. Louis table 8: conditions that favor occurrence of]
(Use "back" to return to the text.)

Other environmental attributes of V. cholerae are its salinity and temperature tolerances. This table, from Valerie Louis' work, gives a broad overview of temperature and salinity ranges in various experimental settings from the laboratory to the Chesapeake Bay, and from California to Louisiana, Florida, England and Japan. The salinity most favorable for V. cholerae was between 2 and 14 ppt.

[C. Bay sampling sites]
(Use "back" to return to the text.)

Here we see sampling sites in the Chesapeake Bay, which were used to investigate how salinity affects cholera concentrations. V. cholerae is more common in the northern part of the bay where salinity is low and when the weather is warmer. In fact, temperature and salinity—both of which display seasonal patterns—combined to predict the presence of V. cholerae with an accuracy between 75.5 and 88.5%.

[Fig 3, V. Louis salinity paper]
(Use "back" to return to the text.)

Here we see results from sampling sites along the north-south transect in the bay just shown. The top graph shows the percentage of samples testing positive for V. cholerae. The bottom graph plots temperature and salinity from north to south (left to right).

Changes in salinity from year to year, due to the influx of freshwater from the Susquehanna River at the head of the bay, may cause V. cholerae populations to fluctuate greatly. We find that salinity variation is a useful indicator of cholera variability in the Chesapeake Bay.

With climate change, shifting rainfall patterns in the mid-Atlantic region could modulate V. cholerae populations in the Bay. With carbon dioxide increasing in the atmosphere, global warming and more rainfall, river flow could increase, lowering salinity and driving up V. cholerae populations. This is a parallel here with the Bay of Bengal, monsoons and cholera.

[Peruvian coast]
(Use "back" to return to the text.)

In the Southern Hemisphere, the coast of Peru, shown on this map, has added new insights to the cholera story. Here, cholera surfaced in 1991 after a century of absence in Latin America. Cholera has recurred in Peru since then, following a seasonal pattern, with the greatest number of cases in summer (June-March) in Lima and other major cities along the coast.

[Table 4-Lipp et al]
(Use "back" to return to the text.)

Erin Lipp and others studied the seasonal distribution in coastal Peru of total V. cholerae, V. cholerae 01, and ctxA. The percentages for each are shown in this table. V. cholerae detection followed ambient temperature increases and coincided with or preceded annual outbreaks of cholera in summer.

[El Nino-SST slide]
(Use "back" to return to the text.)

The climate-cholera link seen in the 1997-98 El Nino year suggests that an early warning system for cholera risk can be established for Peru and neighboring countries.

Interestingly, though not related directly, both El Nino events and cholera outbreaks have increased since the 1970s. This pattern surfaces in both Peruvian waters and the Bay of Bengal. We suggest that sea surface temperature and height, as well as plankton blooms, can be remotely sensed and used to forecast outbreaks.

The actual mechanism that sets off sudden, explosive outbreaks of cholera remains a puzzle. Cholera became a re-emerging disease in Africa in 1970, and in Latin America (as I've already mentioned), in 1991, following absence of the epidemic form for a century. Added to this is the appearance of the new serogroup, 0139, in India, Bangladesh and beyond.

As this model suggests, we believe environmental factors could trigger an epidemic, play a role in seroconversion, and help determine the severity of epidemics. Our recent, four-year study covered four geographically separate locations in Bangladesh, and further defined the organism's ecology there.

The model here summarizes environmental factors directly associated with the presence and multiplication of V. cholerae in the environment. Water temperature, rainfall, water depth and zooplankton all have a direct impact on cholera in the environment. Water temperature also affects the concentration of organic and inorganic chemicals. All of these also affect zooplankton and V. cholerae.

This graph shows the association between the number of cholera cases and water temperature. In fact, water temperature appears to have the most distinct relationship with cases of cholera in Matlab, Bangladesh. Just as an example, at one site, the risk of cholera is between 2.38-4.59 times higher six weeks after a 5 degrees C increase in water temperature.

Another environmental variable, water depth, showed a 2-4 week lag in an inverse relationship with cholera cases.

Here is a closely related variable, rainfall, plotted against the number of cholera cases. As rainfall increased, the number of cases declined, probably due to a decrease in salinity.

This graph shows the relationship of conductivity and cholera cases. Conductivity is a measurement of the concentration of ionic components in solution, providing a measure of potential contaminants in the system. Unlike the lag-time associated with other environmental factors, increased conductivity of the lake-water is associated with a simultaneous increase in cholera cases.

This graph of environmental variables versus cholera cases in a pond in Bakerganj pulls the various results together. The bars show the observed number of cholera cases. The solid line traces the number of cases predicted by the Poisson regression model two weeks into the future. The dashed line is the 95% upper prediction limit. The model employs water depth, copepod number, conductivity and rainfall, all with appropriate lag times—and there is a close match.

More broadly, at each of the four surveillance sites we studied, there was a significant association between physical and biological properties of the water samples and the occurrence of cholera. Incidentally, the association of cholera and water temperature in the ponds corroborates a previous strong correlation with sea surface temperature in the Bay of Bengal. At the same time, culturable V. cholera did not help to predict cholera cases.

This environmental study was done in association with a clinical study, published elsewhere. Environmental isolates of cholera proved more variable than the clinical isolates, suggesting that the clinical strains were most likely a subset of the totality of strains found.

[summary slide]
(Use "back" to return to the text.)

Today I have discussed some infectious diseases with very distinctive biographies. Each of them has a complex and still somewhat mysterious relationship with the environment. I believe that many human pathogens will reveal similar complexities, given new paradigms and new tools to probe infectious disease. In an era where people, pathogens and invading pests travel around the world through both natural and man-made means, both purposefully and in unintended ways, we can no longer circumscribe the dynamics of an infectious disease with a neat and orderly framework and expect to contain and understand its complexity. In a world of ever-more-rapid change, the patterns of disease expand across scales, and explanations must draw upon biological, physical and social science. For the first time we can begin to integrate the complexity of these patterns, if we step out beyond old paradigms to explore these many new dimensions.



National Science Foundation
Office of Legislative and Public Affairs
4201 Wilson Boulevard
Arlington, Virginia 22230, USA
Tel: 703-292-8070
FIRS: 800-877-8339 | TDD: 703-292-5090

NSF Logo Graphic