Can pharmaceutical pollution alter the spread of infectious disease? A case study using fluoxetine

Abstract

Human activity is changing global environments at an unprecedented rate, imposing new ecological and evolutionary ramifications on wildlife dynamics, including host-parasite interactions. Here we investigate how an emerging concern of modern human activity, pharmaceutical pollution, influences the spread of disease in a population, using the water flea Daphnia magna and the bacterial pathogen Pasteuria ramosa as a model system. We found that exposure to different concentrations of fluoxetine-a widely prescribed psychoactive drug and widespread contaminant of aquatic ecosystems-affected the severity of disease experienced by an individual in a non-monotonic manner. The direction and magnitude of any effect, however, varied with both the infection outcome measured and the genotype of the pathogen. By contrast, the characteristics of unexposed animals, and thus the growth and density of susceptible hosts, were robust to fluoxetine. Using our data to parameterize an epidemiological model, we show that fluoxetine is unlikely to lead to a net increase or decrease in the likelihood of an infectious disease outbreak, as measured by a pathogen's transmission rate or basic reproductive number. Instead, any given pathogen genotype may experience a twofold change in likely fitness, but often in opposing directions. Our study demonstrates that changes in pharmaceutical pollution give rise to complex genotype-by-environment interactions in its influence of disease dynamics, with repercussions on pathogen genetic diversity and evolution. This article is part of the theme issue 'Infectious disease ecology and evolution in a changing world'.

Keywords: R0; epidemic; global change; host–parasite interactions; parasite reproduction.

Figure 1.

The effect of fluoxetine exposure on the supply of susceptible hosts, as measured by (a) early fecundity (as measured for the first 12 weeks), (b) lifetime fecundity, (c) lifespan, and (d) intrinsic growth rates (r) of Daphnia magna that have not been exposed to Pasteuria ramosa. Freshwater control (0 ng l⁻¹), low (measured concentration: 12 ng l⁻¹), medium (114 ng l⁻¹) and high fluoxetine (959 ng l⁻¹) treatments are denoted as C, L, M and H, respectively. Points represent treatment means (± s.e.). Lowercase letters indicate significant groupings by post hoc comparisons, where shared letters indicate that groups are not significantly different from each other (p < 0.05). (Online version in colour.)

Figure 2.

The effect of fluoxetine exposure on (a) the probability of successful infection, (b) mature spore loads per infected individual, and (c) lifespan of infected hosts for six different pathogen genotypes (C18, C24, C20, C10, C14 and C1). Freshwater control (0 ng l⁻¹), low (measured concentration: 12 ng l⁻¹), medium (114 ng l⁻¹) and high fluoxetine (959 ng⁻¹) treatments are denoted as C, L, M and H, respectively. Points represent treatment means (± s.e.). Lowercase letters indicate significant groupings by post hoc comparisons conducted separately for each pathogen genotype, where shared letters indicate that groups are not significantly different from each other (p < 0.05). (Online version in colour.)

Figure 3.

The effect on fluoxetine exposure on (a) transmission rate σ and (b) the basic reproductive number (R₀) for six pathogen genotypes (C18, C24, C20, C10, C14 and C1). Freshwater control (0 ng l⁻¹), low (measured concentration: 12 ng l⁻¹), medium (114 ng l⁻¹) and high fluoxetine (959 ng l⁻¹) treatments are denoted as C, L, M and H, respectively. Pairwise reaction norms between control and low, and control and high, fluoxetine treatments are shown for (c) σ and (d) R₀. Data displayed include the mean and 90% uncertainty intervals for the derived metrics and are shown on a log₂ scale to emphasis fold changes in the metrics of pathogen fitness.