FIGURE 2-1 Bangladesh border, barrier islands, and location of Dacca, Matlab, Math-baria, and Bakerganj.

SOURCE: Printed with permission from Google.

FIGURE 2-1 Bangladesh border, barrier islands, and location of Dacca, Matlab, Math-baria, and Bakerganj.

SOURCE: Printed with permission from Google.

Research in Dacca admitted about a thousand new cases per day for almost 30 days and had to use temporary space to house cholera victims. We are working to create predictive models to provide advance warning of conditions that produce severe epidemics in this region of the world.

However, V. cholerae, the bacterium, is a natural inhabitant of rivers, estuaries, and coastal waters throughout the world. Currently, we are sequencing approximately 50 different strains of Vibrio cholerae, the causative agent of cholera collected from many geographic locations to examine their genetic relationships. Preliminary sequencing studies of V. cholerae collected at a depth of 2,000 m at a site located off the coast of Oregon indicate that those isolates may represent ancestral strains; interestingly, one strain studied in detail has genes in common with other Vibrio pathogens, as well, including Vibrio vulnificus and Vibrio parahaemolyticus, the latter being the most common food-borne pathogen in Asian countries, where raw seafood is consumed.


The Ecology of Cholera

My laboratory accomplished the first isolation of Vibrio cholerae from the Chesapeake Bay more than two decades ago, and we now know that this bacterium is found in estuaries of similar salinity, (ca. 15 parts per thousand), where the water temperature rises seasonally to 15°C or higher and where an influx of nutrients encourages plankton blooms (Colwell, 1996). Other species of Vibrio, including V. parahaemolyticus and V. vulnificus, also thrive under these conditions. One of my current graduate students, Brad Haley, has just returned from Iceland, where he was able to isolate V. cholerae at locations where geothermal effluent flows into bays. Clearly, water temperature is critical to the growth of this pathogen.

Vibrio cholerae also has a dormant state, which it assumes between epidemics and during which it cannot be cultured but can be detected with probes (fluorescent antibodies and gene signature sequences). Only during the peak of the zooplankton bloom, in the spring and the fall, is V. cholerae easily culturable. We were able to show that by adding nalidixic acid and nutrient (yeast extract) to water containing the quiescent bacterium, we can stimulate cell elongation and metabolism.

Another important discovery was that cholera is transmitted by plankton. Thus, it is not enough to say that its growth correlates with sea surface temperature and salinity; it is important to recognize the ecological interactions that produce these correlations. There is a commensal relationship—which may prove to be symbiosis—between Vibrio bacteria and zooplankton. Vibrios are chitinolytic (i.e., capable of breaking down chitin, the material that forms the carapaces of zooplankton and crustaceans (e.g., crabs, shrimp). V. cholerae also produces a powerful proteolytic enzyme that the bacterium apparently uses to perform an additional function for zooplankton: breaking down its egg sac, enabling the eggs to disperse into the water column. We are discovering that interactions between V. cholerae and various zooplankton species are quite intricate; for example, certain strains of the bacterium attach preferentially to certain species of zooplankton (Rawlings et al., 2007). All of this leads to the conclusion that V. cholerae is integral to marine ecosystems, and therefore cannot be eradicated.

The Epidemiology of Cholera

We have determined in earlier studies that between 10,000 and 50,000 Vibrio cholerae bacteria may be attached to an individual copepod (the zooplankton favored by V. cholerae). A liter of water drawn by a villager from a pond in Bangladesh between epidemics may contain 10 copepods. However, during a zooplankton bloom, that concentration can increase a hundredfold or more per liter, carrying a dose of cholera bacteria sufficient to cause cholera. The severity of the disease is dose dependent: a low concentration of bacterial cells will pro-


duce mild diarrhea; hospitalized cases—which represent about 25 percent of all infections—require more since one million bacteria per milliliter has been shown to be required to produce the disease. Thus, it has been estimated that only 25 percent of those with cholera end up in hospitals and many more may have been infected (Colwell and Huq, 2004).

Cholera is a disease with rapid onset. Within 24 to 48 hours, the typical patient can lose up to 18 liters of fluid. If that fluid can be replenished quickly, either intravenously or through oral rehydration (using a simple mixture of bicarbonate of soda, table salt, and sugar), recovery is fairly rapid.

From years of study in Bangladesh, we have determined several factors that interact and are associated with the massive annual biennial (spring and fall) cholera epidemics, so that we can predict the onset and severity of epidemics. Our recent research focuses on the communities of Mathbaria and Bakerganj, which are located in the barrier islands region of the Ganges delta (see Figure 2-1). Mangrove-based ecosystems are abundant in copepods. Thus, the Vibrio population is also abundant, and during the zooplankton/Vibrio bloom, cholera results from drinking untreated water.

In Bakerganj and Mathbaria, copepods comprise the majority of zooplankton species. We now have evidence that the severity of a given local cholera epidemic can be determined by copepod population dynamics, with intense epidemics occurring during times of abundance of those copepod species to which epidemic strains of V. cholera preferentially attach. We are currently conducting a seasonal study of zooplankton species in an attempt to determine which species carry V. cholera and to identify factors that influence population size; we will use this information, with other environmental data, to build a predictive capacity for cholera epidemics.

We are also using our knowledge of cholera epidemiology to help the people of Bangladesh to avoid contracting cholera. In one study, for example, we found that by simply educating women to filter drinking water through several layers of sari cloth, we were able to reduce cholera incidence by 50 percent. This result supported our hypothesis that plankton and particulates—to which the bacteria are attracted—transmit cholera and when removed by simple filtration, the incidence of the disease is significantly reduced.

Predictive Models of Cholera

Currently, the spring bloom of phytoplankton in the Bay of Bengal can be measured by satellite sensors that measure chlorophyll intensity and, therefore, the phytoplankton population. Phytoplankton blooms are followed by zooplankton blooms, but the latter cannot yet be measured directly by satellite sensors. However, the zooplankton peak can be inferred using a series of calculations from measurements of the phytoplankton populations that precede the zooplankton


population peak. This information taken together with salinity, temperature, and other environmental factors, provides a more complete picture.

We have also gathered ground truth data over the past 10 years in the Bak-erganj area, including conductivity of the water, presence of inorganic nutrients, temperature, and salinity. With these data, we are able to improve our prediction of the timing and, possibly, the severity of cholera epidemics.

In our original work, we were able to use satellite imagery to measure sea surface temperature and sea surface height in the Bay of Bengal. As shown in Figure 2-2, the correlation of chlorophyll and temperature data, measured by satellite sensors, provides a predictive capacity for conditions conducive to cholera outbreaks. We are currently working on a predictive model that takes into account ocean currents to monitor the movement of plankton into the Bay of Bengal estuaries from the southern tip of India. This could provide as much as a 3-month warning prior to an impending cholera outbreak.

In Latin America, the 1991-1992 El Niño event corresponded with a cholera epidemic that was initially attributed to the dumping of ballast water by a ship in the harbor of Lima, Peru (Gil et al., 2004). We were able to disprove this hypothesis by demonstrating that cholera outbreaks had occurred in three different cities along the coast of Peru, starting before the peak of the 1991-1992 El Niño event. The epidemic more likely resulted from the effect of increased sea surface temperatures on existing plankton and V. cholerae populations.

Our most sophisticated predictive model for cholera incorporates chlorophyll, sea surface height, temperature, and extensive ground truth data. Within a few years, the National Oceanic and Atmospheric Administration (NOAA) will launch a satellite that may provide salinity data. We are also refining our model, based on the 40 years of data accumulated on cholera in Bangladesh and in India, which we are presently analyzing. Nevertheless, with the analyses we have performed to date—sea surface temperature and sea surface height from satellite sensors; measurements of chlorophyll intensity (corrected for the time lag from chlorophyll-phytoplankton bloom to the zooplankton bloom that feeds on the phytoplankton); and measurements of vibrio dispersion in the water—we are able to determine significant correlations and, thus, a foundation from which to predict cholera epidemics.


Climate change is likely to increase the burden of cholera in Bangladesh, but even greater suffering will occur if sea levels rise to predicted levels, displacing millions of people. However, our interdisciplinary, international (as demonstrated by our large number of collaborators from many countries), and biocomplex-ity approach to studying cholera extends well beyond Bangladesh and even beyond the disease itself. By gaining an understanding of the complex interactions between infectious disease, ecology, and the physical environment, we can

Was this article helpful?

0 0

Post a comment