2012 SOUTHEASTERN NATURALIST 11(2):173–182 Temporal Aspects of Leprosy Infection in a Wild Population of Nine-Banded Armadillos Andrew J. Williams1 and William J. Loughry1,* Abstract - Although Dasypus novemcinctus (Nine-banded Armadillo) is the only vertebrate other than humans to exhibit naturally occurring infections with Mycobacterium leprae, the causative agent in producing leprosy, little is known about patterns of infection in wild populations. Here we provide data on some temporal aspects of infection, obtained from sampling a population of armadillos in western Mississippi from 2005–2010. Annual prevalence of infection varied between 4.5–15%. Incidence density estimates calculated over progressively longer time intervals generated values ranging from 0.11–0.61 new cases of infection/1000 animal days. Of 77 animals that tested seropositive over the course of the study, 14 (18.2%) were seropositive in two consecutive years. Four of these animals were seropositive in three consecutive years, but no armadillos tested positive in >3 straight years. Finally, the proportion of seropositive animals increased with the number of years individuals were enrolled in the study. Together, these data indicate a substantial potential for transmission of infection within this population and confirm the view of leprosy as a slowacting disease that is largely manifested in older individuals. Introduction Aside from humans, Dasypus novemcinctus L. (Nine-banded Armadillo) is the only free-ranging vertebrate known to exhibit naturally occurring infections of Mycobacterium leprae, the causative agent in producing leprosy (Truman 2005, 2008). Although many aspects of the disease have been studied in captive armadillos (review in Truman 2008), knowledge of infection dynamics in wild populations remains sparse, with most studies limited to prevalence surveys (e.g., Loughry et al. 2009). An important feature in understanding any disease concerns temporal aspects of infection (Rohani and King 2010). For example, it is vital to know the probability of an individual becoming infected. One standard way of assessing this is by calculating an incidence density estimate (Page et al. 1995). Paige et al. (2002) have provided the only such estimate for armadillos. Working with a population in Louisiana with a high prevalence of M. leprae infection (≈18%), they calculated an incidence density estimate of 3.5 new cases/1000 animal days. This rate was considered quite high and likely to promote widespread transmission of infection (Paige et al. 2002). Even so, estimates from other locales are useful to determine the extent of variation in incidence density between sites. In addition, the estimate of Paige et al. (2002) was derived from a relatively small number of animals (n = 23) recaptured over a single summer. Thus, it seems worthwhile to calculate an estimate from a larger number of animals sampled over a longer time frame. 1Department of Biology, Valdosta State University, Valdosta, GA 31698-0015. *Corresponding author - jloughry@valdosta.edu. 174 Southeastern Naturalist Vol. 11, No. 2 Three other important temporal features of M. leprae infection concern (1) the consistency of infection prevalence from year to year, (2) how infection status changes with age, and (3) how long infected animals persist in a population. Although many diseases exhibit cyclical fluctuations in prevalence (Rohani and King 2010), the limited data available suggest this may not occur with M. leprae infection in armadillos. In the only study of this issue to date, Truman et al. (1991) found no differences in infection prevalence for a population of armadillos in Louisiana that was sampled in 1960–1964 and again annually between 1984–1989. However, it is still necessary to determine the generality of this result with data from other locales. A standard finding in the epidemiology of any disease is that seroprevalence increases with age. This is certainly the case with leprosy, where comparisons between the broad age categories of juvenile, yearling, and adult have only found infection in the adult cohort (Morgan and Loughry 2009, Truman et al. 1991). However, it remains unknown how the likelihood of infection changes as adults continue to age. The expectation is that the proportion of infected individuals should increase as they become older, but no data have been available to test this hypothesis. Regarding persistence, armadillos infected with M. leprae usually die as leprosy spreads to many of the internal organs, with mortality often resulting from secondary infections that occur due to compromised organ functioning (Truman 2008). However, because leprosy is a relatively slow-acting disease, it seems likely infected individuals might survive in a population for extended periods of time. At present, no data are available on how long infected armadillos might persist in a population. Data on this issue are important because such animals may represent an important source of infection transmission during the time they remain active within a population while still relatively healthy. As part of a long-term project on the ecology of leprosy, we collected data on the prevalence of infection in a wild population of armadillos located in western Mississippi. Sampling occurred over multiple years, with many animals captured in more than one year. Consequently, these data allowed us to examine the temporal features of M. leprae infection described above. Specifically, we had four goals in this study: (1) to describe changes in the annual prevalence of infection in the population, (2) to calculate an incidence density estimate from a large number of animals sampled over multiple years, (3) to determine how long infected animals persisted in the population, and (4) to test whether the likelihood of infection increased with age. Field-Site Description Sampling was conducted at the Yazoo National Wildlife Refuge (33º05'N, 90º59'W) near Hollandale, MS. Habitats there are diverse, including bottomland hardwoods that normally flood on an annual basis, shallow cypress swamps, and oxbow lakes that support Nyssa sylvatica (Tupelo Gum), Callitris intratropica (Cypress), and Cephalanthus occidentalis (Buttonbush). There are also many agricultural fields planted primarily with corn and soybeans. 2012 A.J. Williams and W.J. Loughry 175 Methods Study animal Nine-banded Armadillos are medium-sized (adult body weight averages ≈4 kg) burrowing mammals (McBee and Baker 1982). They are primarily nocturnal and feed on various soil invertebrates (McDonough and Loughry 2008). Dasypus novemcinctus has the largest geographic distribution of any species of armadillo, extending from northern Argentina to the southern United States (Abba and Superina 2010). In the US, this range is continuing to expand northwards at an estimated rate of 4 km/yr (Taulman and Robbins 1996). Molecular data indicate armadillos initially acquired leprosy from humans subsequent to the colonization of the Americas by Europeans and their African slaves (Monot et al. 2005). However, the biogeographic distribution of infection remains unclear, with a mix of infected and uninfected populations occurring over the vast range of the species (Loughry and McDonough, in press; Loughry et al. 2009; Truman 2008). Field methods Data were collected from 2005–2010. Sampling was limited in 2005 and 2006, consisting of 2–3 weeks of fieldwork in May of each year. In the remaining four years, sampling was far more extensive, lasting from mid-May to late July (50–55 days in the field/year). In all years, we attempted to capture all armadillos observed during nightly censuses that lasted from about 16:00–24:00 each day. Field procedures followed those described previously by Loughry et al. (2009). Briefly, animals were captured with dip nets, and marked with PIT tags for permanent identification and with various shapes and colors of reflective tape glued to the carapace for longrange, temporary identification. In addition to marking, captured armadillos were measured and weighed, reproductive condition was noted, a piece of ear tissue was taken for genetic analyses, and a blood sample was collected to screen for M. leprae infection. Animals were then released at the point of capture. Although many individuals were resighted and recaptured during a field season, blood samples were only taken at the first capture of each year. Data collection and analysis Blood samples were collected onto Nobuto blood strips (Advantec, Dublin, CA) and allowed to completely air dry. The samples were then screened for the presence of immunoglobulin M antibodies to the M. leprae-specific PGL-I antigen using an enzyme-linked immunosorbent assay (ELISA), following the same protocols originally described by Truman et al. (1986; see also Loughry et al. 2009). All sample analyses were performed at the National Hansen’s Disease Center in Baton Rouge, LA. All samples were run at least twice to confirm consistency. A minimum antibody titer of 0.70 optical density was required to designate an animal as seropositive. We calculated incidence density estimates from these data following the same procedure as Paige et al. (2002). Beginning with 2006, we calculated a series of estimates that progressively added one more year of sampling to the dataset. 176 Southeastern Naturalist Vol. 11, No. 2 Data consisted of the total number of days between date of first capture and the date infected individuals first became seropositive, and the total number of days between first capture and last recapture for those animals that never became infected. Regarding the latter, some animals tested negative in one year, were not caught in one or more succeeding years, but then were recaptured and still tested negative at some later date. Because infection status did not change over time for these individuals, we included them in the data for those years they were not caught, using the last day of sampling for that year as the recapture date. For example, an animal might have tested negative in 2007, not been caught in 2008 or 2009, but tested negative upon recapture in 2010. If so, then this animal was classifi ed as seronegative in 2008 and 2009, using the last date of sampling in each of these years as the “recapture date”. However, data were not added from individuals that, after an interval of one or more years, tested positive upon recapture, because we could not know when their infection status had changed. Likewise, because we only took a blood sample from each animal at the first capture each year, we used that date, rather than any subsequent resighting in that same year, for all animals that were not caught again in a subsequent year, in case their infection status might have changed over the course of the summer. Over each time interval, incidence density was calculated as the number of individuals that became seropositive divided by the sum of the cumulative animal-days for animals that never became seropositive plus one-half of the cumulative animal-days for animals that did become leprous. For comparison with the value from Paige et al. (2002), these values were then adjusted to an estimate per 1000 animal days. To examine how the likelihood of infection changed with age, we performed a Spearman rank-order correlation of the proportion of infected individuals versus the number of years each individual was enrolled in the study. Because no juveniles or yearlings were seropositive (see Morgan and Loughry 2009), we only used data from adults. We chose years enrolled in the study for this analysis because it is not possible to accurately age wild adult armadillos. The age categories of juvenile, yearling, and adult are discernible by differences in body size and weight (Loughry and McDonough 1996, McDonough et al. 1998), but at present there is no way to know the precise age of an animal first caught as an adult. Even so, our analysis is linked with age because, regardless of their true age (in years), each individual armadillo was older with each successive year it was enrolled in the study. Finally, the persistence of infected individuals in the population was determined as the number of intervening years between when an animal first tested positive for infection with M. leprae and the last time it was seen or recaptured. Results The prevalence of infected individuals varied in the Yazoo population from a low of 4.50% in 2005 to a high of 14.92% in 2010 (Fig. 1). Mean prevalence across all years was 9.29% (95% C.I. = 5.50–13.07%). Most differences in prevalence between years were not significantly different from one another (Fisher exact tests, all P > 0.10), except for 2005 versus 2010 (P = 0.004) and the 2012 A.J. Williams and W.J. Loughry 177 marginal cases of 2005 versus 2008 (P = 0.055), 2007 versus 2010 (P = 0.054), and 2009 versus 2010 (P = 0.06). Prevalence data were not entirely independent because some animals were seropositive in more than one year (Table 1). However, although our data spanned a 6-year time frame, no seropositive individual persisted in the population for more than 3 years (Table 1). Of 77 animals that tested seropositive over the course of our study, 14 (18.2%) were positive in 2 consecutive years (Table 1). Four of these animals tested positive in one additional year. There were no significant differences between males and females in the proportion of infected individuals that persisted for 2 or 3 consecutive years (Fisher exact tests: both P > 0.25). Beginning with 2006, incidence density estimates were calculated that included data from each succeeding year of the study (Table 2). These estimates varied from year to year, but all were substantially lower than that reported by Paige et al. (2002). In some years, a few of the seronegative data came from juveniles Table 1. Persistence of adult male and female nine-banded armadillos that were seropositive for infection with Mycobacterium leprae in more than one consecutive year at Yazoo National Wildlife Refuge. Numbers in parentheses are the total number of individuals for which infection status was known over each time frame. 2 years 3 years ≥4 years Males 5 (61) 1 (21) 0 (9) Females 9 (74) 3 (28) 0 (13) Figure 1. Yearly prevalence of Mycobacterium leprae infection in a population of wild Nine-banded Armadillos at Yazoo National Wildlife Refuge. Numbers in parentheses under each point are the total number of animals sampled each year. 178 Southeastern Naturalist Vol. 11, No. 2 recaptured as yearlings. Inclusion of these individuals in the calculation of incidence density estimates is debatable because no infection has ever been reported in such young animals (see above). Consequently, we recalculated our estimates excluding data from yearlings (Table 2). Doing so increased the values slightly, Table 2. Annual incidence density estimates (new cases/1000 animal days) at Yazoo National Wildlife Refuge. Each value was generated using data from all preceding years. Number Number of of new Total number seronegative Total number Incidence Time interval seropositves of days recaptures of days density All data 2005–2006 3 1093 14 4993 0.54 2005–2007 1 742 9 6898 0.14 2005–2008 9 4788 50 24,861 0.31 2005–2009 4 2703 57 34,322 0.11 2005–2010 37 27,387 116 66,380 0.46 Yearlings excluded 2005–2006 3 1093 12 4355 0.61 2005–2007 1 742 9 6898 0.14 2005–2008 9 4788 45 23,028 0.35 2005–2009 4 2703 53 32,845 0.12 2005–2010 37 27,387 101 61,150 0.49 Figure 2. The proportion of adult Nine-banded Armadillos at Yazoo National Wildlife Refuge that tested positive for infection with Mycobacterium leprae as a function of the number of years individuals were enrolled in the study. Data were pooled across males and females (see Table 2). Numbers in parentheses under each point are the total number of animals enrolled. 2012 A.J. Williams and W.J. Loughry 179 but all estimates still remained low, falling within a range from 0.12–0.61 new cases per 1000 animal days (Table 2). The proportion of infected individuals increased with the number of years enrolled in the study (rs = 0.91, P = 0.04), with approximately half of all animals enrolled for ≥4 years being seropositive (Fig. 2). A significantly higher proportion of females enrolled for 2 years were seropositive than were males (Fisher exact test: P = 0.038), but otherwise there were no significant differences in the proportion of seropositive males and females enrolled for the same length of time (Fisher exact tests: all P > 0.40; Table 3). Discussion This study provides important details regarding several temporal aspects of M. leprae infection in wild armadillos. As such, it contributes to a growing body of work in which long-term ecological monitoring helps to illuminate epidemiological patterns of wildlife diseases (review in Rohani and King 2010). Notably, we found that (1) infection prevalence fluctuated significantly between some years; (2) a higher proportion of older animals were infected; (3) some infected animals were able to persist in the population for considerable lengths of time; and (4) the likelihood of acquiring infection (i.e., the incidence density estimates) was substantially lower than that reported from another population (Paige et al. 2002). We treat each of these findings in turn below. Truman et al. (1991) sampled a population of armadillos in Louisiana sporadically over a 20-year period and showed that infection prevalence remained relatively stable. Within a year, they did find significant seasonal changes, with prevalence higher in summer than in spring. This was attributed to the limited availability of adult females for sampling in the spring because they were giving birth and nursing young (Truman et al. 1991). Nonetheless, comparisons of prevalence across years, but within the same season, revealed very little variation. We cannot address seasonal changes in infection prevalence because all our data were collected in the summer. However, in contrast to Truman et al. (1991), our data indicated annual fluctuations can occur. For example, infection prevalence in 2010 was over three times higher than that observed in 2005 (Fig. 1). Unfortunately, at present there are no data available that might provide a plausible explanation for such changes. In addition, a longer time series will be necessary to determine if fluctuations in infection prevalence exhibit regular patterns, as has been observed in some other pathogens (Rohani and King 2010). Table 3. The number of adult male and female Nine-banded Armadillos that were seropositive for infection with Mycobacterium leprae as a function of the number of years individuals were enrolled in the study. The total number of individuals enrolled is given parenthetically. Number of years enrolled 1 2 3 4 5 6 Male 20 (244) 7 (53) 5 (19) 7 (12) 1 (2) 0 (2) Female 19 (225) 17 (56) 3 (21) 8 (19) 1 (2) 3 (4) 180 Southeastern Naturalist Vol. 11, No. 2 Collectively, our data support the conventional view of leprosy as a disease that largely affects older individuals (Morgan and Loughry 2009; Truman 2005, 2008). To date, no study has found evidence of M. leprae infection in juvenile or yearling armadillos (Morgan and Loughry 2009, Truman et al. 1991), which suggests these individuals were either free of infection or the bacterium was present but not abundant or active enough to trigger a detectable immunological response. Our data extend these findings by showing that the proportion of seropositive adult armadillos increased as enrolled animals became older. Indeed, by the time animals had been enrolled for 4 years, the proportion of infected individuals was nearly 50% (Fig. 2). We recognize that the small sample sizes in the older enrollment classes must temper any conclusions about the relationship between age and infection status. Nonetheless, it appears that adult armadillos are at considerable risk of acquiring infection with M. leprae as they get older. It would be interesting to compare survivorship, age structure, and other parameters between a leprosy-free population of armadillos and an infected one in order to explore the demographic consequences of this association between infection status and age. The incidence density estimates we calculated varied from year to year (Table 2). Even so, all the estimates were substantially lower than that obtained by Paige et al. (2002) for a population of armadillos in Louisiana. Presumably, our final estimate (2005–2010) is reasonably precise because of the long time interval and large number of animals included. However, it remains difficult to evaluate this value against those generated over shorter time spans or at other locations. Thus, at present, we can offer no complete explanation for annual or between-sites variability. Undoubtedly, differences in infection prevalence between years and between sites are at least partially responsible. However, the extent to which other factors, such as sampling effort, population density, environmental conditions, and so on, may have contributed to this variation is still unknown. Nonetheless, we would point out that, even though lower, the incidence density estimates for the Yazoo population still indicate a substantial potential for transmission of infection, particularly given the large, dense population of armadillos that is present there (Loughry et al. 2009). The persistence of infected armadillos within a population may represent an important source of infection for other individuals. Most likely this would occur as infected animals shed M. leprae from the respiratory tract, thus potentially contaminating soil at foraging sites and within burrows (see Lavania et al. 2008). Although considered solitary (McBee and Baker 1982), armadillos do exhibit considerable overlap in space use, including using some of the same burrows (albeit not normally at the same time; McDonough and Loughry 2008). Thus, it seems likely an infected armadillo, particularly one that remained active in the population for an extended period of time, could play a significant role in transmitting M. leprae to non-infected individuals. However, arguing against this possibility is the failure to find any evidence of the spatial clustering of infection (Morgan and Loughry 2009, Paige et al. 2002), which suggests infected animals do not represent “hotspots” of disease transmission. Perhaps social network 2012 A.J. Williams and W.J. Loughry 181 models (e.g., Fenner et al. 2011, Perkins et al. 2009) may prove a useful tool in further analyzing the consequences of infected armadillos persisting in populations over multiple years. Given the foregoing, one might expect relatively rapid and widespread transmission of infection within populations of armadillos (see also Truman et al. 2011). This, coupled with the high vagility of the animals (Taulman and Robbins 1996) would seem to lead to the prediction that M. leprae infection should be sweeping across the broad distribution of the species in the southern United States (and, possibly, elsewhere). However, prevalence surveys do not support this hypothesis, with most seropositive populations occurring in a narrow band along the west side of the Mississippi River and the Gulf of Mexico coast from Louisiana to southern Texas (Loughry et al. 2009; Truman 2005, 2008). Truman (1996) proposed that certain environmental conditions (e.g., wet, humid soils), coupled with high densities of armadillos, may be necessary to permit longterm maintenance of M. leprae infection. Thus, it seems plausible that leprosy has been widely disseminated throughout wild populations of armadillos in the US, but rapidly disappeared from those areas unable to sustain the bacterium. Targeted prevalence surveys, based on predictive models of the occurrence of leprosy (Truman et al. 2011), will be necessary to further our understanding of the biogeographical distribution of this disease. We conclude by noting that our results must be considered preliminary. Armadillos may reach ages of 8–12 years old in the wild (Loughry and McDonough, in press) and our time series only extended over 6 years. Thus, continued surveillance of this population (and, ideally, others) will be necessary to confirm and extend the findings reported here. Acknowledgments We thank the staff of the Yazoo National Wildlife Refuge for all their help and support of our work. Thanks also to K. Ancona, M. Ard, M. Bernhardt, W. Burnett, C. Gibson, R. Hooks, R. Morgan, and B. Spychalski for assistance collecting blood samples. We are extremely grateful to R.W. Truman and his lab for screening all of those samples. This work was financially supported in part by grants from the National Geographic Society, and Faculty Research Awards from Valdosta State University. Thanks to J.M. Lockhart, C.M. McDonough, D. Scholl, M. Superina, and R.W. Truman for comments on earlier versions of this paper. Literature Cited Abba, A.M., and M. Superina. 2010. The 2009/2010 armadillo Red List assessment. Edentata 11:135–184. Fenner, A.L., S.S. Godfrey, and C.M. Bull. 2011. Using social networks to deduce whether residents or dispersers spread parasites in a lizard population. Journal of Animal Ecology 80:835–843. Lavania, M., K. Katoch, V.M. Katoch, A.K. Gupta, D.S. Chauhan, R. Sharma, R. Gandhi, V. Chauhan, G. Bansal, P. Sachan, S. Sachan, V.S. Yadav, and R. Jadhav. 2008. Detection of viable Mycobacterium leprae in soil samples: Insights into possible sources of transmission of leprosy. Infection, Genetics and Evolution 8:627–631. 182 Southeastern Naturalist Vol. 11, No. 2 Loughry, W.J., and C.M. McDonough. 1996. Are road kills valid indicators of armadillo population structure? American Midland Naturalist 135:53–59. Loughry, W.J., and C.M. McDonough. In press. The Nine-Banded Armadillo. University of Oklahoma Press, Norman, OK. Loughry, W.J., R.W. Truman, C.M. McDonough, M-K. Tilak, S Garnier, and F. Delsuc. 2009. Is leprosy spreading among Nine-banded Armadillos in the southeastern United States? Journal of Wildlife Diseases 45:144–152. McBee, K., and R.J. Baker. 1982. Dasypus novemcinctus. Mammalian Species 162:1–9. McDonough, C.M., and W.J. Loughry. 2008. Behavioral ecology of armadillos. Pp. 281–293, In S.F. Vizcaíno and W.J. Loughry (Eds.). The Biology of the Xenarthra. University Press of Florida, Gainesville, FL. 370 pp. McDonough, C.M., S.A. McPhee, and W.J. Loughry. 1998. Growth rates of juvenile Nine-banded Armadillos. Southwestern Naturalist 43:462–468. Monot, M., et al. 2005. On the origin of leprosy. Science 308:1040–1042. Morgan, R.E., and W.J. Loughry. 2009. Consequences of exposure to leprosy in a population of wild Nine-banded Armadillos. Journal of Mammalogy 90:1363–1369. Page, R.M, G.E. Cole, and T.C. Timmreck. 1995. Basic epidemiological methods and biostatistics. Jones and Bartlett, Sudbury, MA. 433 pp. Paige, C.F., D.T. Scholl, and R.W. Truman. 2002. Prevalence and incidence density of Mycobacterium leprae and Trypanosoma cruzi infections within a population of wild Nine-banded Armadillos. American Journal of Tropical Medicine and Hygiene 67:528–532. Perkins, S.E., F. Cagnacci, A. Stradiotto, D. Arnoldi, and P.J. Hudson. 2009. Comparison of social networks derived from ecological data: Implications for inferring infectious disease dynamics. Journal of Animal Ecology 78:1015–1022. Rohani, P., and A.A. King. 2010. Never mind the length, feel the quality: The impact of long-term epidemiological data sets on theory, application, and policy. Trends in Ecology and Evolution 25:611–618. Taulman, J.F., and L.W. Robbins. 1996. Recent range expansion and distributional limits of the Nine-banded Armadillo (Dasypus novemcinctus) in the United States. Journal of Biogeography 23:635–648. Truman, R.W. 1996. Environmental associations for Mycobacterium leprae. Pp. 437–449, In S.K. Majumdar, E.W. Miller, and F.J. Brenner (Eds.). Environmental Contaminants, Ecosystems and Human Health. Pennsylvania Academy of Sciences, Philadelphia, PA. 507 pp. Truman, R.W. 2005. Leprosy in wild armadillos. Leprosy Review 76:198–208. Truman, R.W. 2008. Leprosy. Pp. 111–119, In S.F. Vizcaíno and W.J. Loughry (Eds.). The Biology of the Xenarthra. University Press of Florida, Gainesville, FL. 370 pp. Truman, R.W., E.J. Shannon, H.V. Hagstad, M.E. Hugh-Jones, A. Wolff, and R.C. Hastings. 1986. Evaluation of the origin of Mycobacterium leprae infections in the wild armadillo, Dasypus novemcinctus. American Journal of Tropical Medicine and Hygiene 35:588–593. Truman, R.W., J.A. Kumaresan, C.M. McDonough, C.K. Job, and R.C. Hastings. 1991. Seasonal and spatial trends in the detectability of leprosy in wild armadillos. Epidemiology and Infection 106:549–560. Truman, R.W., P. Singh, R.Sharma, P. Busso, J. Rougemont, A. Paniz-Mondolfi, A. Kapopoulou, S. Brisse, D.M. Scollard, T.P. Gillis, and S.T. Cole. 2011. Probable zoonotic leprosy in the southern United States. New England Journal of Medicine 364:1626–1633.