Skip to main content
PMC home page
HHS Author Manuscripts logoLink to HHS Author Manuscripts
. Author manuscript; available in PMC: 2019 Jun 7.
Published in final edited form as: J Med Entomol. 2016 Feb 3;53(3):674–680. doi: 10.1093/jme/tjw004

Infection Prevalence, Bacterial Loads, and Transmission Efficiency in Oropsylla montana (Siphonaptera: Ceratophyllidae) One Day After Exposure to Varying Concentrations of Yersinia pestis in Blood

Karen A Boegler 1,1, Christine B Graham 1, Tammi L Johnson 1, John A Montenieri 1, Rebecca J Eisen 1
PMCID: PMC6555412  NIHMSID: NIHMS872659  PMID: 26843450

Abstract

Unblocked fleas can transmit Yersinia pestis, the bacterium that causes plague, shortly (≤4d) after taking an infectious bloodmeal. Investigators have measured so-called early-phase transmission (EPT) efficiency in various fleas following infection with highly bacteremic blood (≥108cfu/ml). To date, no one has determined the lower limit of bacteremia required for fleas to acquire and transmit infection by EPT, though knowing this threshold is central to determining the length of time a host may be infectious to feeding fleas. Here, we evaluate the ability of Oropsylla montana (Baker) to acquire and transmit Y. pestis after feeding on blood containing 103 to 109cfu/ ml. We evaluated the resulting infection prevalence, bacterial loads, and transmission efficiency within the early-phase time period at 1d postinfection. Fleas acquired infection from bacteremic blood across a wide range of concentrations, but transmission was observed only when fleas ingested highly bacteremic blood.

Keywords: plague, early-phase transmission, Oropsylla montana, Yersinia pestis, flea

Plague, a primarily flea-borne zoonosis, is caused by the gram-negative bacterium, Yersinia pestis, and has been responsible for millions of human deaths throughout history (Gage and Kosoy 2005). The classical, biofilm-dependent regurgitative model of Y. pestis transmission, or the blocked flea model, was first described over a century ago (Bacot and Martin 1914, Eskey and Haas 1940, Burroughs 1947, Hinnebusch et al. 1996, Gage and Kosoy 2005, Hinnebusch 2012). According to this model, fleas develop a proventricular blockage composed of multiplying bacteria and biofilm. The blockage prevents the flea from successfully taking a bloodmeal, and through repeated attempts to feed, the flea transmits Y. pestis to naïve hosts typically 1wk or more postinfection (Bacot and Martin 1914, Hinnebusch et al. 1996, Jarrett et al. 2004).

Given the relatively long extrinsic incubation period associated with blockage formation and the short infectious window following complete blockage of the flea, mathematical models indicate that the blocked flea model is insufficient to explain the rapid rates of Y. pestis spread that are typically observed during plague epizootics (Lorange et al. 2005, Eisen et al. 2006, Webb et al. 2006). In contrast, transmission by unblocked fleas ≤4d following infection, termed early-phase transmission (EPT), has been proposed as a model to describe the rapid spread of Y. pestis observed during epizootics (Eisen et al. 2006, 2015; Eisen and Gage 2009; Buhnerkempe et al. 2011). The term EPT underscores the importance of the short extrinsic incubation period in the flea-borne spread of Y. pestis during epizootic periods. Although previous studies have demonstrated that the genes required for the blocked flea model are not required for EPT (Vetter et al. 2010, Johnson et al. 2014), the biological or biomechanical interactions that ultimately result in EPT remain poorly defined (for review of this subject, see Eisen et al. 2015). Under the EPT model, fleas become infected after feeding on highly bacteremic hosts, which quickly succumb to Y. pestis infection. Plague bacteria spread when infectious fleas leave a highly infectious dead or dying host to feed on new, susceptible, naïve hosts. In recent years, EPT has been evaluated using blood containing ≥108 cfu/ml to infect fleas, a level of bacteremia thought to be typical of those seen in the final hours before rodent hosts succumb to plague (Eisen et al. 2006, 2007a, 2008a,b; Wilder et al. 2008a, b). This design aims to replicate epizootic conditions and, owing to the consistent study design, allows for comparisons of transmission efficiency between flea species. However, to date, the range in concentration of Y. pestis in blood required to infect fleas such that they will reliably transmit the infection during the early phase has not been evaluated.

The concentration of Y. pestis in host blood required to generate a transmissible infection is important because it controls a number of ecologically relevant parameters (Eisen et al. 2015). One study suggested that bacteremia must be extraordinarily high (≥107 cfu/ ml) for fleas to reliably acquire and transmit plague bacteria through the blocked flea model (Engelthaler et al. 2000). This is compatible with earlier observations that bacterial acquisition rates are greatest during the terminal stage of infection when bacterial concentrations commonly exceed 106 cfu/ml (Douglas and Wheeler 1943, Burroughs 1947, Holdenried 1952, Engelthaler et al. 2000, Sebbane et al. 2005). Therefore, the period of time during which an animal is highly bacteremic and infectious to feeding fleas is thought to be short. However, there is some evidence to suggest that susceptible animals may also have lower bacterial concentrations in the blood during the hours or days leading up to terminal septicemia (Sebbane et al. 2005, Weller et al. 2012), or that resistant animals may exhibit transient bacteremia (Rust et al. 1971, Nichols et al. 2014). The length of the infectious window is an important factor when simulating the dynamics of Y. pestis spread under epizootic conditions (Eisen et al. 2006, Webb et al. 2006, Buhnerkempe et al. 2011); the longer the host is infectious, the greater the force of infection (Lorange et al. 2005). Recognizing the implications for epizootic transmission of a prolonged infectious period in the bacteremic host coupled with high transmission efficiency following a short extrinsic incubation period, we aimed to determine the minimum concentration of Y. pestis in blood required to infect feeding fleas and result in transmission during the early phase (≤4d, here measured at 1d postinfection [p.i.]) by Oropsylla montana (Baker), an important bridging vector of Y. pestis to humans in the United States.

Materials and Methods

All animal procedures were approved by the Centers for Disease Control and Prevention Division of Vector-Borne Diseases Animal Care and Use Committee, protocol number 12–023. Use of Y. pestis was approved by the Centers for Disease Control, Division of Select Agents and Toxins, under registration number C20130826–1489.

Flea Infection With Y. pestis CO96–3188

Colony-reared O. montana (Centers for Disease Control and Prevention, Fort Collins, CO) were starved for 5d, then infected using an artificial feeding system described in detail previously (Eisen et al. 2006, 2007a). To generate the spiked bloodmeals for each feeder, fully virulent Y. pestis CO96–3188 was grown overnight (16h) in Heart Infusion Broth (BD Biosciences, Franklin Lakes NJ; 28°C, 200 rpm) to a concentration of ∼109 cfu/ml, then pelleted and resuspended in Sprague Dawley defibrinated rat blood (Bioreclamation, Jericho, NY). Tenfold serial dilutions of this initial suspension yielded concentrations of 103 to 109 cfu/ml blood. We prepared three feeders for each concentration, and groups of 50 fleas were placed in each feeder. Fleas were allowed to feed for 1h, and those that successfully fed were pooled by concentration group, then held for 24h at 32°C and 85% relative humidity. The bacterial concentration within each feeder was verified by spread plating on 6% Sheep Blood Agar (SBA) immediately following the flea feed and counting colony-forming units after an ∼48-h incubation at 28C. To obtain enough fleas to evaluate transmission, artificial flea feeds were conducted on four separate dates.

Transmission of Y. pestis to SKH-1 Mice

Twenty-four hours after feeding on spiked blood, 10 surviving fleas from each concentration group were pooled and allowed to feed on individual anesthetized 6- to 8-wk-old naïve outbred SKH-1 mice (Centers for Disease Control and Prevention). The 24-h time point was selected as representative of transmission efficiency during the early phase because earlier work showed statistically similar transmission rates from 1–4d p.i. After one hour, fleas were removed from the mice and immediately stored at 80°C for later bacterial quantitation. Mice were allowed to recover from anesthesia and then housed individually in HEPA-filtered Isocages (Techniplast, Exton, PA) to prevent transmission between cagemates. Mice were monitored daily and euthanized if signs of infection (e.g., slow response to stimuli, weight loss, and hunched posture) became evident. Infection with Y. pestis was confirmed in liver and spleen tissues using a direct fluorescence assay with fluorescein-conjugated rabbit polyclonal antibodies targeting the F1 antigen (Centers for Disease Control and Prevention, Division of Vector-Borne Diseases) followed by culture and use of specific bacteriophage lysis on resulting isolates (Chu 2000). Mice that did not show signs of illness during 21d of observation were euthanized, and blood serum was evaluated for the presence of Y. pestis-specific antibodies using a passive hemagglutination and inhibition assay (Chu 2000). Specific titers greater than 1:16 were interpreted as evidence of flea-borne transmission. For each Y. pestis concentration group, transmission was evaluated in at least 10 mice.

Infection Status and Quantitation of Bacterial Loads in Infected Fleas

To determine if fleas were infected with Y. pestis at the time of placement on mice, each flea was triturated in 1.0 ml heart infusion broth containing 10% glycerol. Infection status was determined by spread plating 100μl of undiluted triturate to duplicate SBA plates; fleas were considered infected if at least one Y. pestis cfu was present on a single plate at this concentration. To determine the bacterial load within each infected flea, quantitative plating to duplicate SBA plates was performed, and plates were enumerated following 48-h incubation at 28°C. Fleas yielding ≥1 colony per plate (10cfu/flea) were included in quantitative bacterial estimates. Because the fleas used in transmission evaluations are not kept in a sterile environment, contaminant growth was present on a subset of the plates evaluated. When contaminant growth inhibited accurate plate counts but Y. pestis colonies were present, that flea was included in infection prevalence estimates but excluded from quantitative bacterial estimates.

Evaluation of Transmission Success and Statistical Analyses

The proportions of fleas that remained infected with Y. pestis 24h after taking a bloodmeal were compared across the seven target concentration groups (103–109 cfu/ml) using contingency table analysis with Bonferroni correction for multiple pairwise comparisons (JMP 11.2.1, SAS Institute 2014). All fleas that took a bloodmeal, survived 24h, and fed on a naïve mouse were included in the analysis.

The relationship between concentration of Y. pestis in the feeding reservoir and the resulting bacterial load in infected fleas was measured using Spearman’s q correlation. Analysis included all fleas that fed on infectious blood, were alive at 24h, and fed on naïve mice. Fleas were excluded that were found to be uninfected, did not take a bloodmeal from a naïve mouse at 24h, or for which accurate Y. pestis counts could not be obtained owing to nonspecific growth of contaminant organisms. Counts of Y. pestis in the blood (cfu/ml, averaged within concentration group for each artificial feed date) and in each flea (cfu/flea) were log10-transformed before analysis (JMP 11.2.1, SAS Institute 2014).

Transmission efficiency was estimated for fleas fed at seven different concentrations of Y. pestis in blood. Because fleas within each concentration group were pooled before placement on naïve mice, it was impossible to determine which individual flea (or if multiple fleas) transmitted plague bacteria to an individual mouse. Therefore, transmission efficiency of individual fleas within each concentration group was estimated using likelihood methods for pooled samples (PooledInfRate Excel Add-In 4.0, Biggerstaff 2009). To reflect uncertainty in point estimates of transmission efficiency that are based on the flea pool size, 95% confidence intervals are reported for each point estimate. To estimate differences of transmission efficiency between concentration groups, confidence intervals for the differences of proportions for pooled samples (Biggerstaff 2008) were computed using the binGroup package in R, version 3.0.1 (R Foundation for Statistical Computing, Vienna, Austria). The Bonferroni correction was used to adjust for multiple comparisons. For both analyses, only fleas that both fed on transmission mice and were infected (as confirmed by plating) were included in the calculation. Likewise, only mice that were exposed to at least one infected flea were included in the analyses.

To determine if transmission to mice was dependent on infected flea bacterial loads, bacterial counts for each flea were log10- transformed, and the median, minimum, and maximum bacterial load calculated for the flea pool associated with each mouse. Fleas that were not infected, did not feed on the mouse, or for which accurate plate counts could not be obtained were excluded from the analysis. A logistic regression model was then fitted to the data to determine if the median, minimum, or maximum bacterial load per flea pool was predictive of transmission outcome (JMP 11.2.1, SAS Institute 2014). For all analyses, statistical significance was measured at the α = 0.05 level.

Results

Fleas that fed on blood containing a wide range of Y. pestis concentrations (103–109 cfu/ml) acquired and maintained infection for at least 24h p.i. (Table 1). At the lowest concentration of Y. pestis evaluated (103 cfu/ml), infection prevalence in fed fleas (28.0%) was significantly lower than in fleas fed at all higher concentrations, 104– 109 cfu/ml (range: 85.2–100.0%; χ2 ≥ 73.5; P<0.0001 for all pairwise comparisons). Additionally, fleas fed at 108 or 109 cfu/ml were significantly more likely to be infected than fleas fed at 105 cfu/ml (χ2 ≥ 11.9; P≤0.0006). All other pairwise differences among concentration groups were not significant when measured against the Bonferroni-corrected P-value of 0.002 (χ2 ≤ 8.7; P ≥ 0.003). The concentration of Y. pestis in the blood contained in the feeder was positively correlated with the resultant bacterial loads observed in infected fleas (ρs =0.81, P<0.0001).

Table 1.

Flea infection status, observed transmission events, and estimated transmission efficiency of individual fleas 24 h after feeding on blood containing a target concentration of 103–109 cfu/ml Y. pestis

Target concentration of Y. pestis Concentration of Y. pestis (cfu/ml) used in flea feed, range Median bacterial load (range) of infected fleas that fed on naïve mice (cfu/flea)a % of infected fleas that fed (no. of infected/no. of fed) No. of mice infected (total no. of exposed) Estimated transmission efficiency (%) (95% CI)b
1×103 1.6×103–3.9×103 28 (10–2.5×102), n=20 28.0 (23/82) 0 (9) 0.0 (0.0, 13.7)
1×104 2.0×104–3.3×104 2.0×102 (10–3.5×103), n=82 91.7 (88/96) 0 (10) 0.0 (0.0, 3.6)
1×105 2.3×105–3.0×105 9.2×102 (15–2.1×104), n=71 85.2 (75/88) 0 (10) 0.0 (0.0, 4.2)
1×106 2.2×106–2.7×106 2.0×103 (10–1.8×105), n=84 93.7 (89/95) 0 (10) 0.0 (0.0, 3.6)
1×107 2.0×107–4.1×107 1.1×104 (20–1.4×106), n=89 96.8 (90/93) 1 (10) 1.1 (0.1, 5.4)
1×108 2.3×108–3.4×108 2.9×105 (40–3.1×106), n=99 100.0 (100/100) 5 (10) 6.3 (2.4, 14.2)
1×109 1.7×109–3.4×109 1.0×106 (1.9×103–1.0×107), n=92 98.9 (92/93) 9 (10) 18.5 (9.6, 41.1)
Open in a new tab
a

Calculated from fleas with≥10 cfu. Individual fleas were excluded if nonspecific growth impeded enumeration of plates.

b

Calculated using maximum likelihood estimation (MLE) method. Results expressed as maximum likelihood of transmission per infected flea.

Transmission to naïve mice at 24h was observed for groups of fleas fed on blood containing concentrations of Y. pestis at ≥107 cfu/ml (Table 1). Ten, 50, and 90% of mice exposed to fleas fed at 107, 108, and 109 cfu/ml, respectively, became infected, while transmission was not observed in mice exposed to fleas fed on lower concentrations of Y. pestis-infected blood. For concentration groups from which transmission was observed, per-flea transmission efficiency estimates ranged from 1.1% at 107 cfu/ml (95% CI: 0.07– 5.4%), to 18.5% for those fed at 109 cfu/ml (95% CI: 9.6–41.1%). Though transmission was not observed from flea pools fed at concentrations below 107 cfu/ml, transmission efficiency could have been as high as 3.6–4.2% for fleas fed on 104–106 cfu/ml, and as high as 13.7% for fleas fed at 103 cfu/ml. Notably, the confidence interval is the largest for the lowest concentration group because fewer fleas were infected, resulting in a small sample size of infected fleas feeding on naïve mice.

Transmission efficiency estimates were compared between concentration groups, and results are summarized in Table 2. Individual infected fleas fed at 109 cfu/ml were significantly more likely to transmit Y. pestis than those fed at 104–107 cfu/ml, but were not significantly more likely to transmit than those fed at 108 or 103 cfu/ml. All other differences of transmission efficiency between concentration groups were not statistically significant. Low infection prevalence among fleas fed at 103 cfu/ml yielded a small pool size and correspondingly wide confidence intervals, likely explaining why a significant pairwise difference in transmission efficiency could not be detected when these fleas were compared against the fleas fed at 109 cfu/ml.

Table 2.

Pairwise comparisons of transmission efficiency estimates by concentration group

Concentration groups compared Difference (95% CI) between transmission efficiency estimates
109 vs. 108 12.2 (−7.1, 83.9)
109 vs. 107 a 17.4 (2.7, 86.6)
109 vs. 106 a 18.5 (5.0, 47.1)
109 vs. 105 a 18.5 (4.5, 47.1)
109 vs. 104 a 18.5 (4.9, 47.1)
108 vs. 107 5.2 (−4.9, 19.6)
108 vs. 106 6.3 (−2.2, 25.0)
108 vs. 105 6.3 (−3.2, 25.0)
108 vs. 104 6.3 (−2.3, 25.0)
108 vs. 103 6.3 (−17.8, 25.0)
107 vs. 106 1.1 (−6.1, 17.8)
107 vs. 105 1.1 (−7.2, 17.8)
107 vs. 104 1.1 (−6.1, 17.8)
107 vs. 103 1.1 (−22.7, 17.8)
106 vs. 105 0.0 (−11.0, 9.4)
106 vs. 104 0.0 (−9.5, 9.4)
106 vs. 103 0.0 (−28.6, 9.4)
105 vs. 104 0.0 (−9.5, 11.0)
105 vs. 103 0.0 (−28.6, 11.0)
104 vs. 103 0.0 (−28.6, 9.5)
Open in a new tab
a

Transmission efficiency between groups is significantly different when CI does not include zero (α=0.05 with Bonferroni correction).

Transmission outcome could be predicted by the bacterial load present in the fleas that took a bloodmeal on sentinel mice (Table 4). When considering all concentration groups from 103–109 cfu/ml, the log odds of a mouse becoming infected was positively related to the bacterial loads of individual fleas within the pool, regardless of whether minimum, median, or maximum cfu per flea was used to predict transmission outcome (Tables 3 and 4). With each increase of log10 bacterial load in the flea, the odds of transmission by the pool increased by as much as 8.5 times. Overall, the models fit the data significantly better than the null models (χ2 ≥ 34.1; P < 0.0001 for all models). Finally, each model performed well, with area under the receiver operator curve >0.92, indicating that at least 92% of values were correctly classified. When considering only the concentration groups tested in previous studies (≥108 cfu/ml; Eisen et al. 2007a, Wilder et al. 2008a, Schotthoefer et al. 2011), a significant relationship between bacterial loads in individual fleas and transmission outcome was not observed (χ2 ≤ 2.4; P > 0.1 for all models).

Table 4.

Logistic regression analysis of transmission of Y. pestis by infected flea pools to 68 individual mice

Predictor estimates
 Overall model evaluation
Single predictora β (SE) Wald’s χ2 P eβ (odds ratio) χ2 df P R2 AUC
Minimum cfu 1.74 (0.42) 17.04 <0.0001 5.69 42.43 1 <0.0001 0.60 0.95
Median cfu 1.68 (0.46) 13.08 0.0003 5.37 34.05 1 <0.0001 0.48 0.92
Maximum cfu 2.14 (0.66) 10.42 0.0012 8.52 34.88 1 <0.0001 0.49 0.93
Open in a new tab
a

Minimum and maximum cfu—counts from single flea in a pool of 1–10 infected fleas per mouse. Median cfu—calculated from all fleas in a flea pool of 1–10 infected fleas per mouse. All values were log10-transformed before analysis.

Table 3.

Bacterial loads of infected fleas by individual flea pool and concentration group where transmission was observed

Target concentration of Y. pestis in artificial feeder (cfu/ml) Mouse no. Median (range) bacterial load of infected fleas within each flea pool (cfu/flea)a No. of fleas infected (total fed on naïve mouse) Transmission from flea pool to mouse
1×107 1 5.5×103 (3.9×102–1.2×105), n=10 10 (10) N
2 4.0×103 (75–8.0×104), n=10 10 (10) Y
3 1.7×105 (2.3×103–9.3×105), n=9 9 (9) N
4 2.1×104 (6.5×102–1.4×106), n=10 10 (10) N
5 1.9×104 (4.9×102–5.0×105), n=10 10 (10) N
6 2.6×105 (35–3.9×105), n=7 8 (8) N
7 2.1×104 (70–1.8×105), n=9 9 (9) N
8 3.4×102 (20–3.2×104), n=8 9 (9) N
9 2.8×103 (1.4×102–7.0×104), n=10 10 (10) N
10 3.5×104 (6.2×102–1.1×106), n=6 6 (9) N
1×108 1 1.2×106 (1.1×102–1.9×106), n=10 10 (10) N
2 5.9×105 (40–3.1×106), n=10 10 (10) N
3 3.8×105 (2.0×104–1.7×106), n=10 10 (10) N
4 2.2×105 (5.1×104–1.0×106), n=10 10 (10) Y
5 3.1×105 (85–8.8×105), n=10 10 (10) Y
6 8.6×105 (5.1×104–1.6×106), n=9 10 (10) N
7 2.6×105 (7.8×104–9.8×105), n=10 10 (10) Y
8 1.6×105 (4.7×104–4.4×105), n=10 10 (10) Y
9 1.4×105 (1.9×104–6.6×105), n=10 10 (10) Y
10 2.7×105 (8.7×103–6.0×105), n=10 10 (10) N
1×109 1 3.5×106 (6.1×105–9.8×106), n=9 9 (10) Y
2 2.5×106 (9.5×105–5.7×106), n=7 7 (7) Y
3 1.3×106 (2.2×105–3.6×106), n=10 10 (10) N
4 1.2×106 (1.9×103–8.0×106), n=10 10 (10) Y
5 1.4×106 (1.5×105–5.5×106), n=9 9 (9) Y
6 1.3×106 (6.3×105–2.8×106), n=9 9 (9) Y
7 5.7×105 (5.8×104–1.0×107), n=10 10 (10) Y
8 7.4×105 (4.8×103–3.6×106), n=10 10 (10) Y
9 8.6×105 (6.0×104–2.1×106), n=9 9 (9) Y
10 4.1××105 (3.8×104–3.0×106), n= 9 9 (9) Y
Open in a new tab
a

Calculated from fleas with ≥10 cfu. Individual fleas were excluded if nonspecific growth impeded enumeration of plates.

Discussion

For O. montana, a competent vector of Y. pestis that rarely forms proventricular blockage (Eisen et al. 2009), transmission during the early phase of infection (≤4d, measured here at 1d p.i.) requires that the flea feed on a highly bacteremic host. In this study, despite fleas acquiring infection across the concentration gradient from 103– 109 cfu/ml of blood, transmission was only observed for flea pools that fed on blood containing at least 107 cfu/ml and was significantly more likely to occur at 109 cfu/ml, the highest concentration we evaluated. At lower concentrations of bacteria (<107 cfu/ml), transmission was not observed.

Fleas acquired infection across all tested bacterial concentrations, although at a significantly lower rate in the lowest concentration group (103 cfu/ml) compared with other concentration groups. Bacterial loads in the fleas were positively associated with the bacterial concentration in the feeders, and in general, were roughly 3–4 logs lower than the feeder concentration; however, variation within pools was considerable (Tables 1 and 3). Taking into account that bacterial loads can increase by up to a log during the first 24h after blood feeding (Eisen et al. 2006), this finding is consistent with the range in bloodmeal volumes reported for O. montana and Xenopsylla cheopis Rothschild (0.03–0.5μl; Ledingham 1907, Douglas and Wheeler 1943, Perry and Fetherston 1997, Hinnebusch 2005, Oyston and Isherwood 2005). Some of the lower bacterial loads observed among flea pools may be attributable to ingestion of smaller blood volumes or partial feeding. Notably, fewer fleas took bloodmeals in the 105 cfu/ml concentration group compared with those in the 104, and 106–109 cfu/ml concentration groups. It is possible that those in the 105 cfu/ml concentration group also fed less fully, resulting in a lower infection prevalence for that group compared with others.

Despite the high prevalence of Y. pestis infection among pools of fleas exposed to blood containing at least 104 cfu/ml of blood, transmission was observed only in fleas exposed to at least 107 cfu/ml. Furthermore, the likelihood of transmission in the lower concentration groups was significantly lower than those exposed to 109 cfu/ml. Based on the sample sizes used in this study, we cannot rule out the possibility that fleas exposed to bacteremia at levels <107 cfu/ml are capable of transmitting plague bacteria 1 d p.i. Nonetheless, we similar to results observed in studies of transmission by blocked fleas (Engelthaler et al. 2000), a high bacteremia (≥107 cfu/ml) is required for efficient transmission within the early phase (≤4d). However, it is important to note that our findings were restricted to an extrinsic incubation period of 1d p.i. and evaluations at later time points up to 4d would be needed to extrapolate across the full EPT time period.

Previous studies evaluating transmission efficiency during the early phase did not show a significant association between bacterial load and transmission outcome (Eisen 2007a, Wilder et al. 2008a, Schotthoefer et al. 2011). The perceived discrepancy in these results is likely an artifact of study design, as these earlier studies used blood containing ≥108 cfu/ml to infect feeding fleas. Within this narrow range of bacteremia, the resulting bacterial loads of infected fleas were likely too similar to produce significant differences in transmission efficiency. Indeed, when we reexamined this relationship with our data and restricted analysis to fleas fed at ≥108 cfu/ml, we did not identify a significant association between bacterial loads and transmission outcome. This suggests that beyond a critical threshold, increases in bacterial concentration in the blood (and thus bacterial loads in fleas) may not translate to increases in transmission efficiency.

Findings from a number of studies indicate that this rough threshold for reliable flea-borne transmission (between 107–108 cfu/ ml) is met or exceeded only during the terminal stage of Y. pestis infection (Eskey and Haas 1940, Douglas and Wheeler 1943, Holdenried 1952, Engelthaler et al. 2000). While few studies have explored how long a transmissible bacteremia can be sustained in a susceptible host, terminal end point studies show that the life expectancy after reaching this threshold is short. For example, Lorange et al. (2005) reported that once Brown Norway rats achieved bloodborne bacterial loads of ≥106 cfu/ml, they survived only 2d. Similarly, Engelthaler et al. (2000) observed that mice with ≥107 cfu of Y. pestis/ml of blood appeared “severely ill” and often died during the time it took for fleas to take a bloodmeal. Therefore, given that reliable flea-borne transmission of Y. pestis at 1d p.i. depends on a high host bacteremia (e.g., ≥107 cfu/ml), it may also ultimately depend upon terminal illness of the host. Our findings support the concept that low transmission efficiency by fleas, particularly those fed on low concentrations of Y. pestis in the blood, exerts selection pressure for the bacterium to maintain a high level of virulence (Hinnebusch 2005, Lorange et al. 2005, Oyston and Isherwood 2005).

Although we did not observe transmission at concentrations <107 cfu/ml, it is interesting to note that when fleas were fed blood containing 104–106 cfu/ml, between 85.2 and 93.7% became infected and maintained infection for at least 24h. Further, for fleas fed on these lower concentrations of Y. pestis in blood, we estimated by maximum likelihood that transmission efficiency (while not observed) could have been as high as 4.2%. Given these findings, we cannot rule out the possibility that animals with mild bacteremia may contribute to infectious spread, either by extending the length of time the host is infectious (before reaching fulminant bacteremia), or by providing an infectious bloodmeal during a transient, lowlevel bacteremia, as in the case of animals that ultimately recover from infection.

There is some evidence to suggest that low-level bacteremia may occur in animals that are considered “resistant” to Y. pestis infection. For example, SEG strain mice, when compared with susceptible B6 strain mice, are much more likely to survive Y. pestis challenge (10 and 95% mortality, respectively; Blanchet et al. 2011), and in at least one study, Demeure et al. (2012) demonstrated low-level bacteremias (mean levels roughly 101–103 cfu/ml) on days 2 and 3p.i. However, as groups of mice in the study were euthanized at 24-h intervals, neither recovery from infection nor duration of low-level bacteremia could be determined for individual mice. Similarly, Goldenberg et al. (1964) reported a bacteremia of 104 cfu/ ml in an otherwise asymptomatic wild-captured California vole (Microtus californicus), though it is unclear how long this bacteremia could have been sustained, as the culture was obtained at necropsy. Finally, both Rust et al. (1971) and Nichols et al. (2014) described cases of domestic dogs having quantifiable Y. pestis bacteremia before recovery from infection, though neither reported the bacterial concentrations they measured. While these studies point to the possibility of fleas acquiring infection from hosts that recover from plague infections, it is difficult to conceive a situation under which they could contribute meaningfully to epizootic transmission of Y. pestis in a natural system. Even if fleas were able to acquire infection from animals during a phase of transient, low-level bacteremia, subsequent bloodmeals would be noninfectious once the host animal recovered. Previous work has suggested that for O. montana, taking a noninfectious secondary bloodmeal may reduce EPT efficiency (Eisen et al. 2007b). Further, fleas would likely remain associated with hosts that survived infection, reducing the likelihood of transmission to new individuals. Indeed, to impact the spread of Y. pestis, fleas must not only become infected but also transmit infection to new hosts.

If O. montana could become infectious after feeding at very low bacterial concentrations, the infectious window provided by the host to feeding fleas would probably only be extended minimally. For example, in a rat model, blood-borne concentrations of Y. pestis were below the limit of detection (<101.3 cfu/ml) 24h postneedle inoculation. By 48 and 72h postinfection, bacteremia was as high as 106.8 and 109 cfu/ml, respectively (Sebbane et al. 2005). In a mouse model, bacteremia was undetectable 24h postneedle inoculation, but reached levels up to 106 cfu/ml by 48–72h p.i. (Weller et al. 2012). In both studies, animals were humanely euthanized before the terminal stage of infection, so the duration of time that lapsed between achieving an infectious bacteremia and host death could not be determined. However, it seems that low-level bacteremia before fulminant sepsis is short-lived and likely contributes little to the total length of time a host might be infectious to feeding fleas.

Although fleas fed on weakly bacteremic hosts are unlikely to contribute significantly to EPT and epizootic spread of Y. pestis, it could be argued that they could occasionally acquire and maintain infection for sufficient lengths of time to become partially or completely blocked and thereby contribute later to enzootic maintenance of infection. Indeed, authors have discussed the potential for animals with low-level or transient bacteremia to contribute to plague maintenance between epizootics (Pollitzer 1954, Goldenberg et al. 1964). Although we did not observe transmission from fleas fed at the lowest concentrations of Y. pestis during the early phase (1d p.i.), we did not measure transmission at later time points; therefore, we cannot speculate as to whether or not they could contribute to enzootic maintenance via a blocked-flea or late-phase mechanism. Therefore, the contribution of weakly infected hosts to enzootic maintenance of Y. pestis by O. montana, while an interesting speculation, remains outside the scope of our current study and has not yet been demonstrated.

In conclusion, transmission of Y. pestis following a short extrinsic incubation time (1d p.i.) likely depends on the host achieving a high bacteremia (≥107 cfu/ml) within the short period of time before death. Unblocked O. montana were able to acquire and maintain Y. pestis for 24h postinfection at rates exceeding 96% when feeding on ≥107 cfu/ml of blood and transmit the plague bacterium efficiently following an extrinsic incubation period of only 1d. These findings support claims that EPT, defined as transmission by unblocked fleas within 4d of infection, could be an important driver of epizootic spread (Eisen et al. 2006, Eisen and Gage 2009, Buhnerkempe et al. 2011) and support theories that low vector efficiency, particularly among fleas exposed to low bacteremia, imposes selective pressure that favors the maintenance of virulence in Y. pestis (Hinnebusch 2005, Lorange et al. 2005, Oyston and Isherwood 2005).

Acknowledgments

We thank Brad Biggerstaff (CDC, Fort Collins, CO) for his valuable input and assistance with statistical analyses as well as interpretation of results obtained using the PooledInfRate Excel Add-In of his design. We also thank Rebecca J. Clark (CDC, Fort Collins) for her assistance with statistical analyses. This research was supported by the Oak Ridge Institute for Science and Education (ORISE) fellowship program administered by Oak Ridge Associated Universities (ORAU), and funded by the Centers for Disease Control and Prevention (CDC).

References Cited

  1. Bacot AW, and Martin CJ 1914. Observations on the mechanism of the transmission of plague by fleas. J. Hyg 13: 423–439. [PMC free article] [PubMed] [Google Scholar]
  2. Biggerstaff BJ 2008. Confidence intervals for the difference of two proportions estimated from pooled samples. J. Agric. Biol. Envir. Stat 13: 478–496. [Google Scholar]
  3. Biggerstaff BJ 2009. PooledInfRate, version 4.0: A Microsoft Excel Add-In to compute prevalence estimates from pooled samples. Centers for Disease Control and Prevention, Fort Collins, CO. [Google Scholar]
  4. Blanchet C, Jaubert J, Carniel E, Fayolle C, Milon G, Szatanik M, Panthier JJ, and Montagutelli X 2011. Mus spretus SEG/Pas mice resist virulent Yersinia pestis, under multigenic control. Genes Immun. 12: 23–30. [DOI] [PubMed] [Google Scholar]
  5. Buhnerkempe MG, Eisen RJ, Goodell B, Gage KL, Antolin MF, and Webb CT 2011. Transmission shifts underlie variability in population responses to Yersinia pestis infection. PLoS ONE 6: e22498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Burroughs AL 1947. Sylvatic plague studies: The vector efficiency of nine species of fleas compared with Xenopsylla cheopis. J. Hyg 45: 371–396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chu MC 2000. Laboratory manual of plague diagnostic tests. Centers for Disease Control and Prevention/WHO, Geneva, Switzerland. [Google Scholar]
  8. Demeure CE, Blanchet C, Fitting C, Fayolle C, Khun H, Szatanik M, Milon G, Panthier JJ, Jaubert J, Montagutelli X, et al. 2012. Early systemic bacterial dissemination and a rapid innate immune response characterize genetic resistance to plague of SEG mice. J. Infect. Dis 205: 134–143. [DOI] [PubMed] [Google Scholar]
  9. Douglas JR, and Wheeler CM 1943. Sylvatic plague studies: The fate of Pasteurella pestis in the flea. J. Infect. Dis 72: 18–30. [Google Scholar]
  10. Eisen RJ, and Gage KL 2009. Adaptive strategies of Yersinia pestis to persist during inter-epizootic and epizootic periods. Vet. Res 40: 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Eisen RJ, Bearden SW, Wilder AP, Montenieri JA, Antolin MF, and Gage KL 2006. Early-phase transmission of Yersinia pestis by unblocked fleas as a mechanism explaining rapidly spreading plague epizootics. Proc. Natl. Acad. Sci. USA 103: 15380–15385. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Eisen RJ, Wilder AP, Bearden SW, Montenieri JA, and Gage KL 2007a. Early-phase transmission of Yersinia pestis by unblocked Xenopsylla cheopis (Siphonaptera: Pulicidae) is as efficient as transmission by blocked fleas. J. Med. Entomol 44: 678–682. [DOI] [PubMed] [Google Scholar]
  13. Eisen RJ, Lowell JL, Montenieri JA, Bearden SW, and Gage KL 2007b. Temporal dynamics of early-phase transmission of Yersinia pestis by unblocked fleas: Secondary infectious feeds prolong efficient transmission by Oropsylla montana (Siphonaptera: Ceratophyllidae). J. Med. Entomol 44: 672–677. [DOI] [PubMed] [Google Scholar]
  14. Eisen RJ, Holmes JL, Schotthoefer AM, Vetter SM, Montenieri JA, and Gage KL 2008a. Demonstration of early-phase transmission of Yersinia pestis by the mouse flea, Aetheca wagneri (Siphonaptera: Ceratophylidae), and implications for the role of deer mice as enzootic reservoirs. J. Med. Entomol 45: 1160–1164. [DOI] [PubMed] [Google Scholar]
  15. Eisen RJ, Borchert JN, Holmes JL, Amatre G, Van Wyk K, Enscore RE, Babi N, Atiku LA, Wilder AP, Vetter SM, et al. 2008b. Early-phase transmission of Yersinia pestis by cat fleas (Ctenocephalides felis) and their potential role as vectors in a plague-endemic region of Uganda. Am. J. Trop. Med. Hyg 78: 949–956. [PubMed] [Google Scholar]
  16. Eisen RJ, Eisen L, and Gage KL 2009. Studies of vector competency and efficiency of North American fleas for Yersinia pestis: State of the field and future research needs. J. Med. Entomol 46: 737–744. [DOI] [PubMed] [Google Scholar]
  17. Eisen RJ, Dennis DT, and Gage KL 2015. The role of early-phase transmission in the spread of Yersinia pestis. J. Med. Entomol 52: 1183–1192. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Engelthaler DM, Hinnebusch BJ, Rittner CM, and Gage KL 2000. Quantitative competitive PCR as a technique for exploring flea-Yersina pestis dynamics. Am. J. Trop. Med. Hyg 62: 552–560. [DOI] [PubMed] [Google Scholar]
  19. Eskey CR, and Haas GE 1940. Plague in the western part of the United States. Public Health Bull. 254: 1–83. [Google Scholar]
  20. Gage KL, and Kosoy MY 2005. Natural history of plague: Perspectives from more than a century of research. Annu. Rev. Entomol 50: 505–528. [DOI] [PubMed] [Google Scholar]
  21. Goldenberg MI, Quan SF, and Hudson BW 1964. The detection of inapparent infections with Pasteurella pestis in a Microtus californicus population in the San Francisco bay area. Zoonoses Res. 65: 1–13. [PubMed] [Google Scholar]
  22. Hinnebusch BJ 2005. The evolution of flea-borne transmission in Yersinia pestis. Curr. Issues Mol. Biol 7: 197–212. [PubMed] [Google Scholar]
  23. Hinnebusch BJ 2012. Biofilm-dependent and biofilm-independent mechanisms of transmission of Yersinia pestis by fleas. Adv. Exp. Med. Biol 954: 237–243. [DOI] [PubMed] [Google Scholar]
  24. Hinnebusch BJ, Perry RD, and Schwan TG 1996. Role of the Yersinia pestis hemin storage (hms) locus in the transmission of plague by fleas. Science 273: 367–370. [DOI] [PubMed] [Google Scholar]
  25. Holdenried R 1952. Sylvatic plague studies: Plague transmission potentials of the fleas Diamanus montanus and Polygenis gwyni compared with Xenopsylla cheopis. J. Infect. Dis 90: 131–140. [DOI] [PubMed] [Google Scholar]
  26. Jarrett CO, Deak E, Isherwood KE, Oyston PC, Fischer ER, Whitney AR, Kobayashi SD, DeLeo FR, and Hinnebusch BJ 2004. Transmission of Yersinia pestis from an infectious biofilm in the flea vector. J. Infect. Dis 190: 783–792. [DOI] [PubMed] [Google Scholar]
  27. Johnson TL, Hinnebusch BJ, Boegler KA, Graham CB, MacMillan K, Montenieri JA, Bearden SW, Gage KL, Eisen RJ 2014. Yersinia murine toxin is not required for early-phase transmission of Yersinia pestis by Oropsylla montana (Siphonaptera: Ceratophyllidae) or Xenopsylla cheopis (Siphonaptera: Pulicidae). Microbiology 160: 2517–2525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Ledingham JCG 1907. Reports on plague investigations in India. J. Hyg 7: 323–476. [Google Scholar]
  29. Lorange EA, Race BL, Sebbane F, and Joseph Hinnebusch B 2005. Poor vector competence of fleas and the evolution of hypervirulence in Yersinia pestis. J. Infect. Dis 191: 1907–1912. [DOI] [PubMed] [Google Scholar]
  30. Nichols MC, Ettestad PJ, Vinhatton ES, Melman SD, Onischuk L, Pierce EA, and Aragon AS 2014. Yersinia pestis infection in dogs: 62 cases (2003–2011). J. Am. Vet. Med. Assoc 244: 1176–1180. [DOI] [PubMed] [Google Scholar]
  31. Oyston PC, and Isherwood KE 2005. The many and varied niches occupied by Yersinia pestis as an arthropod-vectored zoonotic pathogen. Antonie van Leeuwenhoek 87: 171–177. [DOI] [PubMed] [Google Scholar]
  32. Perry RD, and Fetherston JD 1997. Yersinia pestis-etiologic agent of plague. Clin. Microbiol. Rev 10: 35–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pollitzer R 1954. World Health Organization Monograph Series No. 22. World Health Organization, Geneva, Switzerland. [Google Scholar]
  34. Rust JH Jr., Cavanaugh DC, O’Shita R, and Marshall JD Jr. 1971. The role of domestic animals in the epidemiology of plague: Experimental infection of dogs and cats. J. Infect. Dis 124: 522–526. [DOI] [PubMed] [Google Scholar]
  35. SAS Institute. 2014. Using JMP 11. SAS Institute, Cary, NC. [Google Scholar]
  36. Schotthoefer AM, Bearden SW, Holmes JL, Vetter SM, Montenieri JA, Williams SK, Graham CB, Woods ME, Eisen RJ, and Gage KL 2011. Effects of temperature on the transmission of Yersinia pestis by the flea, Xenopsylla cheopis, in the late phase period. Parasit. Vectors 4: 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Sebbane F, Gardner D, Long D, Gowen BB, and Hinnebusch BJ 2005. Kinetics of disease progression and host response in a rat model of bubonic plague. Am. J. Pathol 166: 1427–1439. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Vetter SM, Eisen RJ, Schotthoefer AM, Montenieri JA, Holmes JL, Bobrov AG, Bearden SW, Perry RD, and Gage KL 2010. Biofilm formation is not required for early-phase transmission of Yersinia pestis. Microbiology 156: 2216–2225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Webb CT, Brooks CP, Gage KL, and Antolin MF 2006. Classic fleaborne transmission does not drive plague epizootics in prairie dogs. Proc. Natl. Acad. Sci. USA 103: 6236–6241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Weller SA, Cox V, Essex-Lopresti A, Hartley MG, Parsons TM, Rachwal PA, Stapleton HL, and Lukaszewski RA 2012. Evaluation of two multiplex real-time PCR screening capabilities for the detection of Bacillus anthracis, Francisella tularensis and Yersinia pestis in blood samples generated from murine infection models. J. Med. Microbiol 61: 1546–1555. [DOI] [PubMed] [Google Scholar]
  41. Wilder AP, Eisen RJ, Bearden SW, Montenieri JA, Gage KL, and Antolin MF 2008a. Oropsylla hirsuta (Siphonaptera: Ceratophyllidae) can support plague epizootics in black-tailed prairie dogs (Cynomys ludovicianus) by early-phase transmission of Yersinia pestis. Vector Borne Zoonotic Dis. 8: 359–367. [DOI] [PubMed] [Google Scholar]
  42. Wilder AP, Eisen RJ, Bearden SW, Montenieri JA, Tripp DW, Brinkerhoff RJ, Gage KL, and Antolin MF 2008b. Transmission efficiency of two flea species (Oropsylla tuberculata cynomuris and Oropsylla hirsuta) involved in plague epizootics among prairie dogs. EcoHealth 5: 205–212. [DOI] [PubMed] [Google Scholar]

RESOURCES