Survival of H. bakeri L₄ was less during coinfection (LCN or HCN) than in a H. bakeri-only infection regardless of the number of H. microstoma (F_2,21 = 8.54, P = 0.002; Figure 1A). In contrast, survival of adult H. bakeri worms did not vary among treatment groups (F_3,34 = 2.5, P = 0.073; Figure 1B). However, if only coinfected mice were considered (i.e., CN, NC, SI), then there were marginally fewer H. bakeri worms in CN mice compared to NC and SI mice (F_2,20 = 3.4, P = 0.052). As expected, there were 58% fewer H. microstoma worms in LCN than in HCN mice (T = 5.87, df = 8.3, P < 0.0001; Figure 2) but H. microstoma infection intensity did not differ among treatment groups (F_3,41 = 1.37, P = 0.265; Figure 2) when similar numbers of cysticercoids were used during inoculation in Experiment 2.

Site selection along the small intestine for L₄ that survived varied with treatment group (F_12,34 = 2.57, P = 0.015) because nematode only infected mice (N) had more L₄ in segments 1 (F_2,21 = 55.76, P < 0.0001) and 2 (F_2,21 = 63.09, P < 0.0001) but fewer in segment 6 (F_2,21 = 8.23, P = 0.009) than coinfected mice which did not differ between each other (Figure 1A). When examined based on percentage of total worms, the distribution of L₄ still varied among treatment groups (arcsine-square root transformed; F_12,34 = 3.18, P = 0.004) because LCN mice had a greater percentage of L₄ in Segment 4 (F_2,21 = 5.39, P = 0.013) and Segment 6 (F_2,21 = 8.46, P = 0.002) than mice in other treatment groups, with no differences in percent distribution for the other segments (Segment 1: F_2,21 = 2.91, P = 0.077; Segment 2: F_2,21 = 1.806, P = 0.189; Segment 3: F_2,21 = 0.99, P = 0.389; Segment 5: F_2,21 = 1.98, P = 0.164). The distribution of adult H. bakeri along the small intestine varied among treatment groups (reciprocal transformed; F_{18, 93} = 2.20, P = 0.008; Figure 1B); this was because for Segment 1 CN mice had fewer H. bakeri worms than the other treatment groups (F_3,34 = 24.63, P < 0.0001) and because of more modest differences in number of H. bakeri worms among treatment groups for segment 4 (F_3,34 = 3.17, P = 0.037; Figure 1B). When examined based on percentage of total worms, the distribution of adult worms still varied among treatment groups (arcsine-square root transformed; F_18,93 = 1.82, P = 0.033). Using Tahame’s post-hoc test (due to unequal variances among treatment groups for some segments), CN mice had a smaller percentage of total worms in Segment 1 compared to all other treatment groups (P < 0.05) which did not differ from each other. Percent distribution among Segments 2–6 did not vary among any treatment groups at P < 0.05.

Sex ratio of L₄ (F_2,21 = 0.39, P = 0.68) or adult (Log₁₀-transformed; F_3,34 = 1.09, P = 0.366) H. bakeri did not vary among treatment groups. While sex ratio did not differ from 1.0 for L₄ (LCN: 1.06 ± 0.16, t = 0.36, df = 7, P = 0.73; HCN: 1.27 ± 0.37, t = 0.75, df = 7, P = 0.48; N: 0.99 ± 0.10 t = 0.113, df = 7, P = 0.91), sex ratio of adult worms was female biased for SI (t = 3.36, df = 7, P = 0.012) and N (t = 4.62, df = 14, P < 0.0001) treatment groups but not for CN (t = 0.09, df = 7, P = 0.93) or NC (t = 2.09, df = 6, P = 0.08) treatment groups (back-transformed mean ± upper/lower standard error, CN: 1.030 ± 0.001/0.001; NC: 0.762 ± 0.002/0.001; SI: 0.674 ± 0.002/0.002; N: 0.758± 0.003/0.003).

Length of adult female H. bakeri worms was shortest when worms were from CN mice and longest when from N and NC mice (F_{3, 34.2} = 7.39, P = 0.001; Figure 3), but length of male H. bakeri worms was similar among all treatment groups (F_3,35.1 = 2.33, P = 0.091; Figure 3). Average individual H. microstoma dry mass was greater for worms from LCN than HCN mice (F_1,14 = 15.36, P = 0.002; Figure 2) because dry mass was negatively correlated with infection intensity (R² = 0.78, y = −2.1x + 29.9; P < 0.0001). Despite the lack of difference in infection intensity of H. microstoma during Experiment 2, H. microstoma worms from C or CN mice weighed less than worms from SI mice (Log₁₀ transformed; F_3,41 = 5.89, P = 0.002; Figure 2) with worms from NC mice having intermediate mass.

H. bakeri via fecal egg output varied over time (Log₁₀ (egg output+10)–transformed; F_6.2,204.4 = 14.26, P < 0.0001) with a significant interaction between time and treatment group (F_18.6,204.4 = 5.31, P < 0.0001) because worms from CN mice did not show the increased egg output over time that occurred for worms of other treatment groups (Figure 4A). After accounting for H. bakeri infection intensity as a covariate (F_1,33 = 18.9, P < 0.0001), CN mice had fewer H. bakeri eggs in their feces than mice in the N, SI or NC treatment groups which did not differ from each other (F_3,33 = 12.65, P < 0.0001; Figure 4A). When individual H. bakeri worms were isolated to measure in vitro reproduction, we found that, for H. bakeri females that produced at least one egg, worms from CN mice produced the fewest eggs but worms from the other treatment groups did not differ from each other (F_3,34 = 10.07, P < 0.0001; Figure 5). The same pattern of egg output was seen if all worms, including those that produced zero eggs, were considered (data not shown). H. microstoma in vivo fecal egg output varied with time (Log₁₀ (egg output+10)–transformed; Greenhouse-Geisser F_7.4,216 = 36.95, P < 0.0001) similarly for all treatment groups (Greenhouse-Geisser F_14.9,216 = 0.77, P = 0.71; Figure 4B). Overall, H. microstoma from SI mice produced more eggs than H. microstoma from either C or CN mice which did not differ from each other (F_2,29 = 6.27, P = 0.005; Figure 4B).

Body mass of mice increased in both experiments (Experiment 1: F_1,21 = 9.52, P = 0.006, Experiment 2: F_1,54 = 349.22, P < 0.0001; Table 1), and the change in mass was similar across each treatment group (Experiment 1: F_2,21 = 0.11, P = 0.897, Experiment 2: F_4,54 = 2.12, P = 0.091). Overall, mouse body mass did not differ among treatment groups for Experiment 1 (F_2,21 = 0.90, P = 0.421) or Experiment 2 (F_4,54 = 0.98, P = 0.427; Table 1).

Small intestine mass did not vary among treatment groups when mice were examined with L₄ H. bakeri present (F_2,21 = 0.31, P = 0.74; Table 1). However, during an adult H. bakeri infection, mouse small intestine dry mass was heaviest for CN and SI mice and lightest for single-infected mice (F_4,53 = 17.21, P < 0.0001; after correcting for host body mass at dissection as a covariate, F_1,53 = 110.35, P < 0.0001; Table 1). Despite these differences in small intestine mass, small intestine length did not differ among treatment groups (F_4,53 = 2.09, P = 0.095) when corrected for host body mass at dissection as a covariate (F_1,53 = 6.50, P = 0.014; Table 1). In Experiment 1, the number of H. microstoma scolices did not affect bile duct mass (R² = 0.0004, slope = 0.0, P = 0.945) and bile duct mass also did not differ between LCN and HCN mice (F_1,14 = 0.05, P = 0.826). Therefore, LCN and HCN data were pooled and compared against nematode-only infected mice (N) which showed that nematode only mice had 83% lighter bile duct mass compared to coinfected mice (F_1,21 = 7.48, P = 0.012; Table 1). In Experiment 2, bile duct mass varied among treatment treatments (Log₁₀ transformed; F_4,55 = 74.21, P < 0.0001). Specifically, mice given only H. microstoma (C) or H. microstoma first (CN) had similar bile duct mass that was larger than the other two coinfected groups (NC or SI) which did not differ from each other and H. bakeri only infected mice (N) had the smallest bile ducts that were different from all other treatment groups (Table 1). The number of H. microstoma scolices did not affect bile duct mass for mice in Experiment 2 that had at least 1 scolex (R² = 0.046, y = 0.001x + 0.014, P = 0.159).

Liver function did not vary among treatment groups as measured by ALP after 27d of H. microstoma infection in Experiment 1 (F_2,19 = 0.58, P = 0.572; Figure 6). However, after a longer H. microstoma infection in Experiment 2, ALP levels varied among treatment groups (Log₁₀ transformed; F_4,55 = 10.59, P < 0.0001; Figure 6) because ALP for mice infected with H. microstoma only (C) was greater than for mice infected with H. bakeri only (N) or H. microstoma followed by H. bakeri (CN), with mice from other treatment groups (NC and SI) having intermediate values. ALP levels did not vary with number of H. microstoma scolices for mice with at least one scolex (R² = 0.002, y = 0.63x + 59.2, P = 0.765). There was no difference in ALT production among treatment groups (F_4,55 = 1.84, P = 0.135) or when only mice with at least one H. microstoma scolex were included (F_3,41 = 1.37, P = 0.267; Figure 6). ALT levels were not related to the number of scolices (R² = 0.008, y = −5.01x + 206.2, P = 0.569) for mice with at least one scolex.

When parasites share a host, there are numerous ways in which they may affect each other. Having another parasite species in the host can alter the parasite’s environment indirectly (e.g., via bile concentrations along the small intestine) or directly (e.g., occupation of the same host space where physical contact is possible). Data from our experiments show that order of coinfection influences life history traits of both H. bakeri and H. microstoma and that it is not always advantageous to be the first, or even the only, parasite in the host.

For each experiment, there was one independent variable of “treatment group” and numerous dependent variables to assess parasite life history and host hepatobiliary and small intestine responses. The Institutional Animal Care and Use Committee (IACUC) of California State University San Marcos approved all procedures.

Male Swiss Webster mice (Harlan, Indianapolis, IN, USA) 5.5–7 weeks old were housed in groups of three to four, maintained on a 14:10 light:dark cycle, and provided water and food (Purina Mouse Chow #5001) ad libitum. Our original stock of H. microstoma was graciously supplied by Dr. Jerzy Behnke at the University of Nottingham and was maintained in our lab by passage through confused flour beetles (Tribolium confusum) and Swiss Webster laboratory mice. Using a dissecting microscope, we dissected H. microstoma cysticercoids from T. confusum that had been placed in 0.9% saline. To inoculate mice, cysticercoids, along with 20–25 µL saline, were taken into a 0–200 µL pipette tip and then dispensed into the mouse’s mouth. Following each inoculation, the pipette tip was flushed with saline, contents were inspected using a dissecting microscope, and any remaining cysticercoids were re-administered as described above. For both experiments, flour beetles were infected with H. microstoma eggs between 39-80 days prior to cysticercoid dissection. H. bakeri was cultured using standard methods [58] and maintained in non-experimental Swiss Webster mice. Within each experiment, all L₃ were from a single culture where L₃ age was less than three months. To inoculate experimental mice with H. bakeri, L₃ (suspended in tap water) were dispensed into the mouse’s mouth via a 0–200 µL pipette; nematode control mice received an equal volume of tap water.

Mouse body mass was measured just prior to infection and on day of dissection. For Experiment 2, in vivo parasite egg output was counted from 1–3 fecal pellets that were collected daily from each mouse (starting at 8d post- H. bakeri inoculation or 12d post- H. microstoma inoculation). Number of eggs was counted using floatation with saturated sucrose and egg output per gram of feces was calculated separately for each parasite by the following equation: ([volume of water + volume of sucrose]/feces mass) * (# eggs per grid/volume of grid). For each mouse sample, two grids of a McMaster slide were counted and the average was used as the eggs per gram of feces for each parasite. Just prior to euthanasia, 200 µL of blood was collected from the ocular sinus of anesthetized (via inhaled isoflurane) mice using 70 µL heparinized capillary tubes. Blood was centrifuged for 15 minutes at 9,500 × g and resulting plasma was aliquoted and frozen at −80 °C until used for ALT or ALP kinetic colorimetric assays. ALT (units per liter, U/L) was measured on a microplate reader at 340nm for a 15 µL plasma sample after 5 minutes of reaction time according to manufacturer instructions (ALT reagent C164-01, calibrator C1200-10, Catachem Inc., Bridgeport¸ CT, USA; assay was verified using manufacturer controls of Catatrol I C1200-11 and Catatrol II C1200-12, Catachem Inc., Bridgeport¸ CT, USA). ALP (international units per liter, IU/L) was measured on a microplate reader at 405 nm for a 10 µL plasma sample after 4 minutes of reaction time according to manufacturer instructions (DALP-250, BioAssay Systems, Hayward, CA, USA).

After mice were euthanized, the small intestine and the combined cystic duct and common bile duct (hereafter called “bile duct”) were removed. The bile duct was opened longitudinally and inspected using a dissection microscope and all H. microstoma scolices and worm tissue were removed to a separate Petri dish. The small intestine was cut into 6 equal-length segments where segment #1 began immediately after the stomach at the pyloric sphincter and segment #6 ended immediately proximal to the cecum at the ileocecal valve. Each segment was measured to the nearest 1mm (for Experiment 2), placed in 0.9% saline, and opened longitudinally. Total small intestine length was calculated as the sum of the six segments. Pieces of H. microstoma were removed, added to H. microstoma tissue taken from the bile duct, and all scolices were counted. Adult and L₄ H. bakeri were removed from each segment using watchmaker forceps (L₄ were removed from the serosal side and adults from the luminal side) under a dissecting microscope. All H. bakeri worms were counted and sexed. H. bakeri sex ratio was calculated as the number of males divided by the number of females. In Experiment 2, length of a sub-sample of H. bakeri worms (10 male, 10 female) from each mouse was measured to the nearest 0.01 mm using an ocular micrometer fitted to a dissecting microscope. Another set of 12 adult H. bakeri female worms per mouse was used to measure in vitro egg output. Each worm was individually placed in a well of a 48-well plate that contained 500 µL of RPMI media that was supplemented with 1X penicillin/streptomycin and 0.08% amphotericin B. The plates were incubated in 5% CO2:95% N for 24 h at 37 °C. After 24 h, female worms were removed, well contents were mixed by pipetting and eggs were counted in a 350µL sample from each well using flotation with an equal volume of saturated sucrose. Egg output was calculated by the following equation: # eggs per grid/volume of grid * 2 (a dilution factor to account for volume of saturated sucrose) * volume of media. One grid of a McMaster slide was counted per worm. The bile duct, H. microstoma worms, and each intestinal segment were dried at 58 °C for 48 h to a constant mass and were then weighed to the nearest 0.0001 g to determine dry mass. Average individual H. microstoma worm dry mass was calculated as the dry mass of all worms combined divided by the number of scolices per mouse. Total small intestine dry mass was calculated as the sum of dry masses of the six segments.

We tested each variable for homogeneity of variance and transformed data as necessary. When appropriate, covariates were examined and only retained in the statistical model if significant. Repeated measures ANOVA was used to test if host body mass at infection and at dissection varied among treatment groups, and for in vivo fecal egg output over time. For Hymenolepis microstoma worms, host in vivo fecal egg output was only evaluated starting at day 16 post-infection for C, CN and SI mice for 34 d of infection to allow for a repeated measures model without missing values. Total infection intensity for each parasite, sex ratio of H. bakeri, ALT, ALP, bile duct mass, small intestine mass, small intestine length, and average individual H. microstoma mass were analyzed by ANOVA. A one-sample T test was used to determine if sex ratio differed from 1.0 (i.e., an equal number of male and female worms). Infection intensity of H. microstoma in the high versus low infection treatments in Experiment 1 was tested with an adjusted Student’s T test without assumption of equal variances (because transformations would not yield homogeneity of variances). The distribution of H. bakeri worms in the six small intestine segments was analyzed with a one-factor multivariate analysis of variance (MANOVA) with segment number as the within subjects variable and the number of worms as the between subjects variable. Length of H. bakeri worms and in vitro egg output were analyzed using nested ANOVA (nested by mouse ID) because individual H. bakeri worms from each mouse could not be considered independent. Linear regression was used to determine if bile duct mass, ALP and ALT levels were affected by the number of H. microstoma scolices. When appropriate, post-hoc analyses used Tukey’s least significant difference or a comparison of 95% confidence intervals (unless otherwise stated). All data were analyzed with SPSS/PASW and considered significant at P < 0.05.