Comparative reservoir competence of Peromyscus leucopus, C57BL/6J, and C3H/HeN for Borrelia burgdorferi B31

Abstract

Borrelia burgdorferi, a Lyme disease spirochete, causes a range of acute and chronic maladies in humans. However, a primary vertebrate reservoir in the United States, the white-footed deermouse Peromyscus leucopus, is reported not to have reduced fitness following infection. Although laboratory strains of Mus musculus mice have successfully been leveraged to model acute human Lyme disease, the ability of these rodents to model B. burgdorferi-P. leucopus interactions remains understudied. Here, we compared infection of P. leucopus with B. burgdorferi B31 with infection of the traditional B. burgdorferi murine models-C57BL/6J and C3H/HeN Mus musculus, which develop signs of inflammation akin to human disease. We find that B. burgdorferi was able to reach much higher burdens (10- to 30-times higher) in multiple M. musculus skin sites and that the overall dynamics of infection differed between the two rodent species. We also found that P. leucopus remained transmissive to larval Ixodes scapularis for a far shorter period than either M. musculus strain. In line with these observations, we found that P. leucopus does launch a modest but sustained inflammatory response against B. burgdorferi in the skin, which we hypothesize leads to reduced bacterial viability and rodent-to-tick transmission in these hosts. Similarly, we also observe evidence of inflammation in infected P. leucopus hearts. These observations provide new insight into reservoir species and the B. burgdorferi enzootic cycle.IMPORTANCEA Lyme disease-causing bacteria, Borrelia burgdorferi, must alternate between infecting a vertebrate host-usually rodents or birds-and ticks. In order to be successful in that endeavor, the bacteria must avoid being killed by the vertebrate host before it can infect a new larval tick. In this work, we examine how B. burgdorferi and one of its primary vertebrate reservoirs, Peromyscus leucopus, interact during an experimental infection. We find that B. burgdorferi appears to colonize its natural host less successfully than conventional laboratory mouse models, which aligns with a sustained seemingly anti-bacterial response by P. leucopus against the microbe. These data enhance our understanding of P. leucopus host-pathogen interactions and could potentially serve as a foundation to uncover ways to disrupt the spread of B. burgdorferi in nature.

Keywords: Borrelia burgdorferi; Borreliella burgdorferi; Ixodes scapularis; Lyme disease; Peromyscus leucopus; enzootic cycle; reservoir competency; tick-borne disease; vector-borne disease.

Fig 1

B. burgdorferi B31 expands to greater numbers in M. musculus than in P. leucopus following tick bite. (A) B. burgdorferi can be cultured from needle injection sites from P. leucopus and M. musculus (n = 6) 1 week post-injection. (B) B. burgdorferi are present at substantially lower numbers at the needle injection site in P. leucopus than C3H/HeN mice by ddPCR. For panels A and B, rodents were injected with 100,000 B. burgdorferi via subcutaneous injection. (C) B. burgdorferi can be cultured from tick placement sites 1 week post-infestation. (D) B. burgdorferi are present at substantially lower numbers at the tick placement site in P. leucopus (n = 4) than C57BL/6J (n = 6) or C3H/HeN (n = 6) 1 week post-infestation by ddPCR. For panels C and D, infected nymphal I. scapularis were used to infect rodents. The tick placement region was harvested 1 week post-infestation. For panels A and C, cultures were monitored for spirochetes for 3 weeks. For panels B and D, B. burgdorferi ospA DNA copies are normalized to copies of the mammalian Rpp30 gene, and P-values are generated by one-way ANOVA on log-transformed data with Dunnett’s multiple comparison test. For panels (A–D), all rodents were between 8 and 12 weeks old at the time of infection, and both sexes were present. For panels A and D, bars represent the mean, and error bars the standard error of the mean. Individual rodents are plotted as dots (circle = female and square = male). (E) B. burgdorferi has different population dynamics in ear tissue from male P. leucopus or M. musculus (n = 3) following tick bite. Nymphal I. scapularis were used to infect rodents and ear punches were taken weekly starting at 1.5 weeks post-tick placement for ddPCR analysis. P-value was generated by two-way ANOVA on log-transformed data. Individual rodents were plotted as dots, and lines intercept the mean. For all panels, data are pooled from two independent experiments.

Fig 2

Population dynamics of B. burgdorferi B31 in P. leucopus and M. musculus over 13 weeks. (A) Schematic of the experimental design used for Fig. 2 to 4. Rodents were infested with five nymphal ticks (infected or uninfected), and rodents were harvested at the listed terminal time points. Rodents were subjected to a maximum of one intermediate time point. All rodents were between 6 and 12 weeks old at the time of infection. Each time point included one male and one female rodent per species per experiment where possible. All time points represent between three and eight samples depending on whether the time point combines terminal and intermediate values (e.g., week 2; n = 8, week 8; n = 4) and/or whether samples needed to be excluded for technical reasons (e.g., rodent death, failed DNA extraction). (B, C) B. burgdorferi can be cultured from P. leucopus ear punches late in infection (B), whereas M. musculus can be cultured from ear punches starting at 2 weeks post-nymph placement (C). (D, E) The population dynamics of B. burgdorferi in ear tissue of P. leucopus (D) and M. musculus (E) ears are measured by ddPCR. The B. burgdorferi ospA gene copies are normalized to copies of the mammalian Rpp30 gene. Each dot represents a single mouse (circles = female; squares = male), and the connecting line intercepts the median value of each time point. Samples that did not have any copies of ospA are plotted at 0.0001 copies per 100 Rpp30 copies.

Fig 3

P. leucopus and M. musculus B. burgdorferi transmission to larval ticks over time. (A) P. leucopus has reduced transmission of B. burgdorferi to larval ticks over time. Rodents were infested with approximately 50 larval I. scapularis at 1, 2, 4, and 10 weeks post-infestation with B31-infected nymphs. Fed larvae were crushed in BSK-II and monitored for growth to determine the rate of transmission from the rodents to I. scapularis. (B) In P. leucopus, B. burgdorferi burden in the ear (measured by ddPCR) does not correlate with transmission. Ear punches were taken just before infestation with larval ticks. (C, D) C57BL/6 J mice remain highly transmissive over time (C) and burden in the ear (measured by ddPCR) does not correlate with transmission (D). (E, F) C3H/HeN mice remain highly transmissive over time (E) and burden in the ear (measured by ddPCR) does correlate with transmission (F). For panels A, C, and D, the transmission was calculated as the number of fed larval ticks that yielded culturable spirochetes divided by the total number of fed larval ticks collected (5–30 fed larval ticks per rodent were collected, average = 20.5 ticks/rodent). Rodents with fewer than five fed larval ticks collected were excluded. Each dot is an individual rodent (circles = female, squares = male), and bars mark the arithmetic mean. Reported P-value is from a one-way ANOVA on the log-transformed data with Šidák’s multiple comparison test comparing 2 weeks with 10 weeks to test for reduced infectivity over time. For panels B, D, and F, each dot is an individual rodent, the error bars mark the 95% CI of the slope, and the P-value is from a Pearson r correlation.

Fig 4

P. leucopus have a subtle but prolonged immune response to B. burgdorferi B31 in skin. (A–D) Differentially expressed genes in P. leucopus skin at 1 (A), 2 (B), 4 (C), and 8 weeks (D) post-infestation compared with uninfected P. leucopus. Genes with P < 0.05 are colored green or pink if down- or up-regulated, accordingly. (E) More genes are upregulated than downregulated across time points using a P < 0.05 threshold. (F) Many upregulated genes overlap across time points. Overlap between gene sets was determined using Biovenn software (33). (G–I) Use of QIAGEN IPA software (34) reveals activation or suppression of (G) biological pathways, (H) pathways regulated by cytokines, and (I) pathways regulated by transcription factors in B. burgdorferi-infected P. leucopus. For G–I, activation of pathways was determined using a Z score cutoff of 2, and repression was determined using a Z score cutoff of −2. RNA was derived from 4 P. leucopus rodents (two males, two females) at each time point.