Leeches Baicalobdella torquata feed on hemolymph but have a low effect on the cellular immune response of amphipod Eulimnogammarus verrucosus from Lake Baikal

Animal sampling and handling

All experimental procedures were conducted in accordance with the EU Directive 2010/63/EU for animal experiments and the Declaration of Helsinki; the protocol of the study was approved by the Animal Subjects Research Committee of the Institute of Biology at Irkutsk State University (protocol #2022/11) before the start of the experiments. Leech-free and leech-infected amphipods Eulimnogammarus verrucosus (Gerstfeldt, 1858) were collected by kick sampling with a hand net in Baikal littoral zone near Listvyanka village (51°52′05.5″N 104°49′47.1″E) at depths of 0–1.2 m (the animals belong to the W genetic lineage (Drozdova et al., 2022). Amphipods were acclimated to the laboratory conditions in well aerated 3-L plastic aquaria at 6°C in MIR-254 incubators (Sanyo, Osaka, Japan) for at least 3 days prior to any manipulations and experiments. All found leeches were attached to the gills of amphipods (Figs. 1A, 1B).

Identification of leech species

After samplings in October 2022, February 2023, and April 2023 and subsequent acclimation, some of the leeches were detached from amphipods and fixed in 96% ethanol for further species identification. Body width of fixed leeches was determined after photographing under a stereo microscope SPM0880 (Altami, Russia) in ImageJ software (Rueden et al., 2017).

Morphological analysis of fixed specimens was performed according to the standard keys (Bauer, 1987; Lukin, 1976). DNA extraction from the posterior sucker of leeches was performed using the S-sorb kit (Syntol, EX-516, Moscow, Russia). PCR amplification of the cytochrome c oxidase subunit I (COI) gene fragment was performed with a 5 × Screen Mix (Evrogen, Moscow, Russia), the Folmer primers (LCO1490/HCO2198 (Folmer et al., 1994)), and the following program: 94 °C for 1 min, 30 cycles of 94 °C for 20 s, 43 °C for 2 min, and 72 °C for 1 min.

The sequencing reactions were performed in both directions using the BigDye Terminator v3.1 Cycle Sequencing Kit (Life Technologies, Waltham, MA, USA) and analyzed with a Nanophor-05 Sanger sequencer (Syntol, Moscow, Russia). Sequencing reads were basecalled and converted with the programs Mutation Surveyor v5.1 and Chromas v2.6.6. Consensus sequences were compiled with UGENE v41.0 (Okonechnikov et al., 2012) using the sequence from Baicalobdella sp. (NCBI Genbank MN854834) as the reference COI fragment. The obtained sequences with a length of 559 bp were deposited in the NCBI GenBank database with accession numbers OR077511 –OR077525. Almost all COI sequences for Baicalobdella sp., as well as one sequence for a closely related genus Codonobdella, deposited in NCBI as of 5th February 2024 (Bolbat et al., 2021; Utevsky & Trontelj, 2004), were used (except for KM078844, KM078841, KM078820 and KM078810 due to their lengths shorter than 559 bp) to construct the phylogeny along with the obtained data. Several sequences of fish leeches of the genus Piscicola (KM095104, DQ414337, OX030972, and MH395321) were used as outgroups (Kaygorodova et al., 2014; Utevsky & Trontelj, 2004; Cichocka et al., 2018). Sequences were aligned with the MAFFT algorithm (Katoh & Standley, 2013) in the UGENE program (Okonechnikov et al., 2012); the alignment is available in Supplemental Information. The phylogeny was built with the IQ-Tree web server (http://www.iqtree.org/) using automatic model selection with Model Finder (Kalyaanamoorthy et al., 2017) and ultrafast bootstrap for assessment of the branch support values (Hoang et al., 2018). The resulting phylogenetic tree was visualized with iTOL (https://itol.embl.de/) (Letunic & Bork, 2021).

Injection of fluorescent latex beads into amphipods and further visualization

We analyzed the ability of leeches Baicalobdella sp. to consume amphipod hemolymph after sampling in July 2023. For this, leeches were detached from amphipod gills with tweezers and kept in aquaria separately from hosts for ∼24 h. Next, 10 non-infected individuals of E. verrucosus were immobilized in an incised wet polyurethane sponge at the acclimation temperature and injected with 1 µl of saline containing about 3  × 10⁶ latex microbeads (L3030, Sigma-Aldrich, St. Louis, MO, USA) using an IM-9B microinjector (Narishige, Tokyo, Japan). Right after the injection, the amphipods were placed in aquaria with free leeches, which attached to the new hosts within 30 min.

Four hours post-injection, we anesthetized the amphipods in clove oil suspension (50 µL of clove oil per 50 mL of Baikal water) and detached leeches and two pieces of gills from each individual for further observation under an inverted fluorescent microscope Celena S (Logos Biosystems, Republic of Korea). Prior to the visualization, the leeches were placed into sterile 1.5-mL microtubes and homogenized with 50 µL of phosphate buffered saline using a plastic pestle.

Hemolymph extraction and characterization of hemocytes

In all experiments before the hemolymph extraction, the dorsal side of the amphipod pereon surface was always sterilized with 70% ethanol. The central hemolymph vessel was punctured with a sterile needle, and hemolymph was collected with a sterile glass capillary. The obtained hemolymph was immediately mixed 1:1 with isotonic anticoagulant solution (150 mM NaCl, 5 mM Na₂HPO₄, 30 mM sodium citrate, 10 mM EDTA, pH 8.0; filtered through a 0.45 µm syringe filter) on ice to avoid degranulation of granulocytes (Shchapova et al., 2019). Amphipod hemolymph was always extracted before the detachment of leeches.

Hemocytes were visualized using the Celena S inverted microscope (Logos Biosystems, Republic of Korea) or the Mikmed-2 upright microscope (LOMO, Russia) with an attached EOS 1200D camera (Canon, Tokyo, Japan). Total hemocyte count (THC; i.e., hemocyte concentration in a certain volume) and granulocyte percentage were estimated in glass hemocytometers or disposable hemocytometers (Aptaca, Italy).

Characterization of hemocyte types was performed with a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, USA). The hemolymph of 8 non-infected amphipods E. verrucosus was extracted and measured for forward (allows for the discrimination of cells by size) and side scatter (gives the information about cell complexity).

Biochemical measurements of glycogen content and phenoloxidase activity

Along with the estimation of THC, some of infected and non-infected animals collected in October 2022, February 2023 or April 2023 were used for glycogen content measurements. Glycogen along with lipids and protein content are the main resources depleting under energy demand in crustaceans (Sánchez-Paz et al., 2006; Sacristán et al., 2017). Glycogen extraction was performed as described previously (Vereshchagina et al., 2016) with modifications. Frozen amphipod tissues (after hemolymph extraction) were ground into a powder, mixed with the solution (0.5 mL per 100 mg of wet weight) containing 0.6 M HClO₄, and further homogenized in a Potter-Elvehjem tissue grinder until no visible particles remained. Next, 20 µL of the homogenate were mixed with 75 µL of 1% amyloglucosidase (10115-5G-F, Sigma-Aldrich, Germany; 5,250 U/µL) in a 0.2 M acetic acid buffer (acetic acid/sodium acetate; pH 4.8). The mix was incubated at 40 °C for two hours and then 62.5 µL of 0.6 M HClO₄ and 100 µL of 1 M KHCO₃ were added. The supernatant was centrifuged at 16000 ×g for 15 min. Glycogen concentration was measured with the “Glucose-Vital” kit (Vital Development, Russia): 40 µL of experimental sample was added to 190 µL of “Glucose-Vital” monoreagent and incubated at 25 ° C for 15 min. Light absorption was measured at 510 nm with a CLARIOstar Plus microplate reader (BMG Labtech, Ortenberg, Germany).

Hemolymph phenoloxidase activity was measured for amphipods sampled in May 2023. For this, hemolymph was collected as described above, mixed 1:1 with a buffer solution (150 mM NaCl, 10 mM Na₂HPO₄, 7 mg/mL phenylmethanesulfonyl fluoride, pH 8.0), and frozen at −80 °C. The samples were thawed at 4 °C and centrifuged for 10 min at 500 g and 4 °C to precipitate the cellular pellets. 10 µL of hemolymph extract were mixed with 40 µL of buffer solution, 280 µL of distilled water, and 40 µL of 4 mg/mL 3,4-dihydroxy-L-phenylalanine. Measurements were performed with the CLARIOstar Plus microplate reader at 490 nm (absorbance) for 40 min. The activity of phenoloxidase was calculated as the slope of the reaction curve during the linear phase and expressed in arbitrary units (Shchapova et al., 2019).

Assessing the encapsulation of Sephadex beads by amphipod hemocytes in primary culture

The encapsulation reaction of hemocytes extracted from leech-infected and non-infected amphipods was quantitatively assessed using primary cell culture with added Sephadex beads as model foreign bodies (Wu et al., 2014). This in vitro approach allowed us to maintain similar concentrations of hemocytes and the beads in different replicates of the experiment, which would hardly be possible with in vivo experiments due to the highly variable hemocyte concentration in hemolymph.

Sephadex microbeads (G100120-50G, Sigma-Aldrich, St. Louis, MO, USA) were washed with 5 mg/mL streptomycin and 5,000 U/mL penicillin solution (1.3.18, Biolot, St. Petersburg, Russia). The bead suspension was pipetted into a sterile 96-well plate (GT204-0096DV, Minimed, Northridge, CA, USA) in a laminar flow box. Then, 100 µL of complete medium L-15 (Leibovitz medium with L-glutamine, L4386-10X1L, Sigma-Aldrich, St. Louis, MO, USA) containing 15% fetal bovine serum (FBS-HI-11A, Capricorn Scientific, Germany) was added (Shchapova et al., 2019) and the beads were imaged under a Mikmed-2 microscope and counted using the CountThings application (https://countthings.com). Since initially the amount of microbeads per well varied substantially, for further tests we used only the wells with approximately the same number of beads (on average 230 ± 70 beads per well).

Hemolymph was extracted from 10 leech-free and 12 leech-infected amphipods (collected in April 2023) as described above and pooled within each group. 10-µL aliquotes were collected from the pools to estimate the hemocyte concentrations. Then, each pool of hemolymph was divided into the selected wells with microbeads with control for the equal amounts of hemocytes for the leech-free (15 wells) and leech-infected (11 wells) groups (on average, 1 ± 0,4 × 10⁵ cells per well).

After cell sedimentation to the well bottom, the upper layer of the suspension was collected, and 100 µL of fresh L-15 medium with L-glutamine and 15% fetal bovine serum were added; cells were kept at 6 °C (Shchapova et al., 2019). The hemocyte response to the Sephadex microbeads was analyzed after 24 h of incubation, and the number of microbeads with hemocyte aggregates was counted (Mastore et al., 2015; Wu et al., 2014; Ling & Yu, 2006) under the Celena S inverted fluorescent microscope. We categorized four stages of the encapsulation reaction: no reaction, low reaction, medium reaction, the stage showing partially encapsulated beads, and the intense reaction showing fully covered beads (Figs. 1C, 1D). The hemocyte nuclei were stained with 10 µg/mL 4′,6-diamidino-2-phenylindole (DAPI, A4099, AppliChem, Darmstadt, Germany) to visually contrast the encapsulation reaction (Fig. 1D). Cell viability was assessed by staining with 1 µg/mL propidium iodide (81845-100MG, Sigma-Adrich, St. Louis, MO, USA).

Artificial infection of amphipods with leeches and bacteria

In order to evaluate the potential synergistic effects of infection with bacteria and leeches on amphipod immune system, we performed two 3-day-long experiments. The first (auxillary) experiment was intended to check for the possible influence of injection (sham treatment) on the studied parameters, THC and granulocyte percentage. The experiment included the initial control group and amphipods (sampling of leech-free animals in February 2024) after injection of 2.5 µL of buffered saline (150 mM NaCl, 10 mM Na₂HPO₄) into the central hemolymph vessel between the 5th and 6th segments with an IM-9B microinjector (Narishige, Tokyo, Japan). Animals were immobilized in an incised wet polyurethane sponge at the acclimation temperature during all injections. After 1.5 h, 1 and 3 days hemolymph was extracted from amphipods and mixed 1:1 with adjusted anticoagulant solution (150 mM NaCl, 5 mM Na₂HPO₄, 30 mM sodium citrate, 10 mM EDTA, 50 mM EDTA-Na₂, pH 8.0). This adjusted anticoagulant solution allows to fix hemocytes in the state when nuclei and granules are visible more clearly (Skafar & Shumeiko, 2022) and was applied to later visually distinguish granulocytes among all hemocytes. THC and granulocyte proportion were estimated under the Mikmed-2 microscope in a glass hemocytometer.

For the second (main) experiment (animal sampling in July 2023), we injected the bacterial strain Pseudomonas sp. H5-2 (belongs to the P. fluorescens species group) that was previously extracted from the hemolymph of E. verrucosus collected in the same location (Shchapova et al., 2021). For the cultivation, we used the tryptic soy broth (TSB) medium (casein peptone, dipotassium hydrogen phosphate, glucose, NaCl, and soy peptone) as suggested previously (Robach, 2006; (Murali, Bhargava & Wright, 2018). For injection into amphipods, Pseudomonas sp. cells were washed by centrifugation and resuspended in physiological solution in order to achieve a concentration of 10⁵ Pseudomonas sp. cells per 1 µL (i.e., 2.5 × 10⁵ cells per animal). After ∼15–30 min, the amphipods with and without the bacterial injection were infected by leeches as described above with a 1:1 parasite-to-host ratio.

All experimental groups of the main experiment for 1.5 h and 1 day time points included 10 animals per group and showed no mortality both with and without bacterial infection (since hemolymph samplings always failed for some animals, the number of analyzed hemolymph samples had to be reduced down to seven for some groups). The first round for the 3-day time point also included 10 animals per experimental group but showed high mortality specifically for animals injected with bacteria (60% for leech-free and 50% for leech-infected), with no mortality for amphipods without injection. Since this high mortality could be an artifact of the specific injection procedure, we performed the second round of the experiment with bacterial injection into 9 animals per experimental group. Both leech-free and leech-infected animals showed no mortality during 3 days post injection, and their hemolymph was used for the tests along with the hemolymph of the animals from the first round.

Statistical analysis

All data analyses were performed in R v.4.3.1 using built-in functions (R Core Team, 2022). Statistically significant differences between experimental groups were always estimated using the Mann–Whitney U test with Holm’s correction for multiple comparisons. The differences were considered statistically significant at p < 0.05. The p-values to linear regression coefficients for the relation between THC and summarized leech width per host were obtained with the summary() function.

Specifically for the experiment with artificial infections with bacteria and leeches, we applied a generalized linear model (GLM) for the analysis of factor effects. The model was fitted using the glm() function with Gaussian distribution to three independent factors (time as numeric variable, absence or presence of leech, and injected bacteria) and all of their interactions. The assumptions for GLM were mostly met for the dataset: the outcome with time was acceptably linear (slightly violated specifically for THC), the residuals were always homoscedastic, and the normality assumption was slightly violated only for THC.

Infection rates and identification of leeches

We collected leech-infected amphipods E. verrucosus at the same site in Lake Baikal but at different times of the year. The infection rate was not estimated precisely, but it clearly varied greatly: in October 2022 94 individuals out of ∼120 examined (∼78%) were infected, in February 2023 11 out of ∼130 (∼9%) and in April 2023 12 out of ∼100 (∼12%).

We performed a morphological analysis of 35 leeches obtained from amphipods that were further used for estimation of hemocyte concentration (five leeches in October 2022, 15 leeches in February 2023 and 15 leeches in April 2023). All 35 analyzed leeches belonged to the same genus Baicalobdella, with most of them being representatives of the morphospecies B. torquata. Four leeches (sampled in February 2023) were identified as potentially belonging to the morphospecies B. cottidarum, but recent data indicate that B. torquata may have significant morphological variability (Matveenko & Kaygorodova, 2020; Matveenko, 2023), so identification of these specimens remained uncertain. In order to clarify the diversity of the leeches, we performed sequencing of the COI gene fragment in 15 specimens in total; all samples with ambiguous morphological identification and from October 2022 were included in the analysis, and the rest morphologically identified as B. torquata were chosen randomly from two samplings.

The phylogenetic tree clearly showed that all 15 leeches belonged to the same species B. torquata (Fig. 2); their COI fragments also showed low pairwise differences of no more than 1.8% (i.e., 10 mutations per 559 bp). Since E. verrucosus was found to be infected with only one species of Baicalobdella locally, we had the possibility to test the physiological influence of these leeches on the amphipods.

Leeches B. torquata consume amphipod hemolymph

The assumption that the leeches attached to the gills of amphipods also feed on their hemolymph is obvious, but these ectosymbionts may simply be in phoretic relationships with specifically these hosts. In order to test this assumption, we injected fluorescent microbeads into the hemolymph of E. verrucosus and tracked their distribution. Five hours post-injection, the microbeads were easily observable in amphipod gills and also inside some ciliates that were found to be attached to gills (Figs. 3A, 3B). The homogenates of five out of 10 tested leech bodies also contained these fluorescent microbeads (Fig. 3C). Since the leech oral apparatus is not suitable for consumption of ciliates (Bauer, 1987; Nesemann & Neubert, 1999; Sawyer, 1986), our data unambiguously confirms that leeches B. torquata can indeed feed on the hemolymph of E. verrucosus.

Characterization of amphipod hemocytes

Since hemocytes are an important component of the crustacean immune system, before further analysis we investigated their possible subdivision into populations. Flow cytometry clearly differentiated the hemocytes of E. verrucosus into two main groups, one with a smaller cell size and lower internal complexity and the other with a larger cell size and higher internal complexity (Fig. 4). The groups are usually called hyalinocytes and granulocytes, respectively, (Rowley, 2016) and can also be differentiated with conventional phase contrast microscopy. In particular, the larger size of granulocytes is evident right after the sample is placed under the microscope, while the higher amount of vesicular structures in granulocytes is better visualized after attachment to the surface (Fig. 4). Additionally, we observed hemocytes with intermediate internal complexity and size between granulocytes and hyalinocytes, i.e., semi-granulocytes, but their proportion was only ∼10%.

Influence of leeches on hemocyte concentration and other parameters of amphipods in different seasons

The consumption of hemolymph by leeches may directly reduce the hemocyte concentration and phenoloxidase content in the hemolymph and indirectly reduce the available energy resources such as glycogen due to the compensation of the tissue loss. We used the amphipods collected in October 2022, February 2023 and April 2023 to discriminate between the effects of leech infection on two of these parameters of E. verrucosus in different seasons.

The median total hemocyte count (THC) of non-infected animals gradually increased from October to April by 2.8 times (Fig. 5A), but the difference between seasons was not statistically significant (all three p-values > 0.12). There were also no significant differences in THC between leech-infected and non-infected amphipods in these months (all three p-values > 0.42). However, THC for infected animals in February was significantly higher than in October (p < 0.01) and by median 1.5 times higher than in respective non-infected animals. This coincides with over 1.6 larger median width of leeches in February (Figs. 5B, 5C) in comparison to both October and April (both p-values < 0.005), while the size in the latter two months was effectively identical (p = 0.57). With a larger leech size the hemocyte concentration would be expected to be the lowest, but the obtained data suggested no such relation or even the opposite tendency. Since after acclimation most amphipods were infected with 2–4 leeches (11 out of 19 infected), we could not check the correlation between THC and leech size directly, but the dependence between THC and summarized leech width per host was practically absent with Spearman’s correlation coefficient of 0.36 (Fig. 5D).

The analysis of glycogen content (Fig. 5F) included uniformly selected samples from October, February, and April and showed identical median values between leech-infected and non-infected amphipods (p = 0.51). Finally, phenoloxidase activity was measured for the separate set of E. verrucosus sampled in May 2023 (Fig. 5E) and indicated no statistically significant differences between infected and non-infected amphipods (p = 0.9).

Cellular immune response of infected and non-infected amphipods estimated in vitro

Despite we did not find substantial effects of leech infection on the amounts of immune components in amphipod hemolymph, they might modulate the intensity of the host immune response through bioactive components in their saliva. For preliminary testing of this hypothesis, we chose the primary culture of amphipod hemocytes as a convenient model system and Sephadex microbeads (consisting of specifically processed dextran) as model foreign bodies. The primary hemocyte culture allows for observing the behavior of these immune cells and quantitative estimation of their reactions such as aggregation and further encapsulation of foreign bodies.

In particular, we measured the fraction of Sephadex beads encapsulated by hemocytes that were originally extracted from leech-infected and non-infected amphipods 24 h after contact with the beads. This time point was previously shown to be enough for the development of a strong immune reaction even to artificial, non-microbial foreign bodies (Shchapova et al., 2019). We found no difference in the intensity of the immune reaction between the experimental groups since the proportions of fully encapsulated (∼6%) and partially encapsulated microbeads (∼93%) were equal (both p-values > 0.07) for hemocytes from infected and non-infected amphipods (Fig. 6A). Some of the beads were not encapsulated at all, and there was a high mortality of hemocytes around Sephadex microbeads in contrast to free hemocytes, as indicated by propidium iodide staining (Fig. 6B).

Changes in hemocyte concentration and composition after injection of bacteria and artificial leech infection

Finally, in order to evaluate the potential synergistic effects of leeches and other immunity-related factors we experimentally tested the influence of leeches on the ability of amphipods to deal with bacterial infection. First, we examined the potential effects of the injection procedure on the chosen parameters. Both THC and the fraction of granulocytes among all hemocytes (Figs. 7A, 7B) demonstrated no statistically significant changes during the 3 days after injection of physiological solution in comparison to amphipods without any injections (all six p-values > 0.32 in comparisons to the respective control groups).

Next, we performed the experiment with (i) an artificial infection of leech-free amphipods with the Pseudomonas sp. strain originally extracted from hemolymph of the same species and (ii) an artificial infection with leeches. The order of the procedures was motivated by the high infection rate of E. verrucosus with Pseudomonas (Shchapova et al., 2021) and the variability in infection rate with leeches (see above). The number of injected bacterial cells was comparable to the number of circulating hemocytes in the animal hemolymph to model a significant microbial infection. Some of the amphipods were then infected with one leech per animal (Fig. 7; Table 1).

The mortality during the three-day experiment was mostly low, and it was never higher for leech-infected animals than for leech-free ones. The generalized linear model indicated that infection with leeches itself and time after the infections had no statistically significant effects on both the concentration of hemocytes in amphipod hemolymph and the fraction of granulocytes among them, while the injection of bacteria clearly leads to a statistically significant decrease in hemocyte concentration by ∼2,800 cells per µl on average and an increase in granulocyte proportion by ∼16% (Table 1). Interestingly, the interaction of bacterial injection and leech infection, in contrast, resulted in a statistically significant decrease in the proportion of granulocytes by 12% but caused no statistically significant changes in hemocyte concentration (Table 1). Other interactions between factors, even being statistically significant in the case of granulocyte percentage, did not exceed 0.5% in absolute value in the estimated effect. However, pairwise comparisons between leech-free and leech-infected animals showed no statistically significant differences between any experimental groups not only in hemocyte concentration but also in granulocyte fraction during the whole experiment (all twelve p-values > 0.09; Fig. 7).