Abstract

   Parasites enhance their fitness by manipulating host dispersal.
   However, the strategies used by ectoparasites to influence host
   movement and the underlying mechanisms remain poorly understood. Here,
   we show that ectoparasites alter metabolic activity in specific brain
   regions of mice, with evidence pointing to a potential role for
   microglial activation in the prefrontal cortex. This activation appears
   to contribute to synaptic changes and altered neuronal differentiation,
   particularly in GABAergic neurons. Consequently, exploratory behavior
   decreases—an effect likely mediated through the skin–brain axis. In
   both indoor and field experiments with striped hamsters, ectoparasites
   reduce host exploration and modify their dispersal patterns. This
   behavioral shift ultimately restricts the host’s distribution, enabling
   parasites to avoid environmental pressures. Our findings reveal that
   ectoparasites limit host dispersal to improve their own fitness,
   offering key insights for parasite control strategies that promote
   health and preserve ecological stability within the One Health
   framework.

   Subject terms: Behavioural ecology, Animal behaviour, Parasitology
     __________________________________________________________________

   Ectoparasites activate prefrontal microglia via the skin–brain axis,
   damaging GABAergic neurons and reducing host exploration. This limits
   host distribution and may help ectoparasites avoid environmental
   pressures.

Introduction

   Host dispersal significantly influences the ecology and evolution of
   parasites. Most notably, it can facilitate the spread of parasites
   across large spatial scales^[50]1. However, host dispersal can also
   impact parasite survival^[51]2. Evidence indicate that hosts employ
   dispersal to effectively remove parasites^[52]3–[53]5. To survive, some
   parasites have evolved the ability to precisely and adaptively
   manipulate host dispersal^[54]6, such as altering the likelihood and
   timing of departure^[55]7 or even influencing the distance of
   transfer^[56]8.

   Although host manipulation by parasites has been documented in several
   hundred host-parasite associations across major phylogenetic groups,
   manipulation by ectoparasites appears to have been relatively
   understudied. Ectoparasites are organisms that live on a host’s
   surface, feeding on its blood or skin, including fleas, ticks, and
   mites. For ectoparasites, host dispersal can introduce greater stress,
   as ectoparasites are more vulnerable to environmental pressure^[57]9.
   Consequently, these ectoparasites may need to develop strategies to
   manipulate their hosts in order to enhance their own fitness. We
   hypothesize that ectoparasites may suppress their hosts’ exploratory
   behavior to improve the parasites’ survival, thereby reducing host
   dispersal and occupancy.

   The flea-rodent parasitic system provides a unique opportunity to
   deepen our understanding of how ectoparasites manipulate their hosts’
   dispersal. Fleas are significant vectors of zoonotic diseases,
   including plague, murine typhus, and rural epidemic typhus^[58]10. At
   the same time, wild rodents—the primary reservoirs of these
   pathogens—are almost always infested with fleas. Historically,
   outbreaks of vector-borne diseases have often been linked to changes at
   the human-animal interface^[59]11. Parasite-induced behavioral
   manipulation of hosts can significantly alter the likelihood of
   pathogen spillover from wildlife to humans. The One Health approach
   helps prevent and control outbreaks of vector-borne diseases by
   recognizing the interconnection between human, animal, and
   environmental health^[60]12,[61]13. Investigating how ectoparasites
   influence host behavior is therefore crucial for understanding pathogen
   spillover, monitoring shifts at the human-animal interface, and
   advancing One Health strategies.

   Here, we show that ectoparasites manipulate host dispersal to enhance
   their own survival using two rodent models. Mice infected with fleas
   exhibit reduced exploratory behavior and increased anxiety, along with
   metabolic and functional abnormalities in the prefrontal cortex,
   including microglial activation and GABAergic neuron damage. Using a
   flea–striped hamster (Cricetulus barabensis) model, we replicate these
   effects through controlled indoor infections and enclosed field
   experiments, confirming that flea bites limit host exploration.
   Simulations further reveal that fleas restrict host movement to
   influence host distribution, helping them avoid environmental
   heterogeneity (Fig. [62]1). This study expands our understanding of how
   parasites manipulate host behavior, identifies potential
   neurobiological mechanisms underlying this phenomenon, and provides
   insights into parasite-induced behavioral changes that inform
   intervention strategies within the One Health framework.

Fig. 1. Schematic diagrams of the experiment design in this study, created in
BioRender.

   [63]Fig. 1
   [64]Open in a new tab

   Liu, P. (2025) [65]https://BioRender.com/t7hic5v. a Three female mice
   were housed per cage and allowed to acclimate for one week. Flea
   infection was initiated by introducing 150 live fleas (X. cheopis) into
   each cage in the Flea+ group, while the Flea− group received the same
   number of heat-killed fleas as a control. Each week, remaining fleas
   were removed, and 150 new fleas were reintroduced to maintain
   infection. After four weeks of continuous infection, behavioral and
   molecular assays were conducted to test the hypothesis that
   ectoparasites reduce host exploratory behavior and to investigate the
   underlying neuro-molecular mechanisms. b Striped hamsters captured from
   the wild were acclimated to laboratory conditions for one month before
   flea infection. The fleas used for infection were derived from a wild
   population naturally infecting striped hamsters. Hamsters were housed
   individually at night and exposed to fleas in glass tanks during the
   day. After four weeks of infection, behavioral experiments were
   performed to assess changes in exploratory behavior. c Six 30 m × 30 m
   field enclosures were constructed to conduct three rounds of flea
   infection experiments, each lasting four weeks. In the first round,
   only striped hamsters were introduced. In the second round, fleas were
   added to a subset of enclosures, and in the third round, fleas were
   introduced to all enclosures. After each round, hamsters were
   recaptured, and their exploratory behavior was evaluated based on
   capture rates. d Using dispersal data from the experiments, combined
   with demographic data and habitat suitability maps, the RangeShiftR
   model was employed to simulate the impact of flea-induced reductions in
   exploratory behavior on the distribution and range dynamics of striped
   hamsters.

Results

Flea bites suppress exploratory behavior

   Fleas are blood-sucking arthropods, and different species exert varying
   levels of selective pressure on their hosts^[66]14. Meanwhile, the
   greater the pressure exerted by fleas, the lower the survival
   rate^[67]15. To establish a stable flea-mouse system, we compared the
   survival rates of two flea species, Ctenocephalides felis (C. felis)
   and Xenopsylla cheopis (X. cheopis). After one week, all C. felis had
   died, while approximately 50% of X. cheopis fleas survived
   (Supplementary Fig. [68]1). Thus, we chose X. cheopis for our infection
   experiments to ensure sustained chronic stress.

   After infecting for four weeks, we immediately conducted behavioral
   tests. We assessed dispersal behavior by measuring the time mice spent
   exploring the central area in the open field test (OFT) and the open
   arms in the elevated plus maze (EPM). Compared to the Flea− group, the
   Flea+ group spent less time in the central area of the open field and
   made fewer attempts to enter it (Fig. [69]2a). Moreover, in the EPM
   test, the Flea+ group spent less time in the open arms (Fig. [70]2b),
   indicating that flea bites increased anxiety-like behaviors/reduced
   exploratory behavior in the mice.

Fig. 2. Flea bites reduce mice’s exploratory behavior and alter brain
metabolic patterns.

   [71]Fig. 2
   [72]Open in a new tab

   a Open field test results shown as a boxplot: Flea− (red, n = 12) and
   Flea+ (blue, n = 12). One-sided t test was used. Flea+ mice exhibited
   reduced time in the central area (left) and fewer entries into the
   central area (right) compared to Flea− mice. Thick bars indicate the
   interquartile range (IQR) around the median, and whiskers represent 1.5
   times the interquartile range (maxima: Q3 + 1.5 × IQR, minima:
   Q1−1.5 × IQR). Each black dot represents an individual. Asterisks
   indicate the significance. b Elevated plus maze results shown as a
   boxplot: Flea− (red, n = 8) and Flea+ (blue, n = 8). One-sided t test
   was used. Flea+ mice spent less time in the open arms compared to Flea−
   mice. Thick bars indicate the interquartile range (IQR) around the
   median, and whiskers represent 1.5 times the interquartile range
   (maxima: Q3 + 1.5 × IQR, minima: Q1−1.5 × IQR). Each black dot
   represents an individual. Asterisks indicate the significance. c
   Spearman correlation of 18F-FDG uptake between brain regions in the
   Flea− group (n = 6) and Flea+ group (n = 3). Two-sided t test was used.
   d 3D rendering of SPM-based voxel-level statistical analysis overlaid
   on an MR template image. Blue indicates hypometabolic regions in the
   Flea+ group, while red indicates hypermetabolic regions. Statistically
   significant clusters are marked in the figure. e Bar graph showing
   brain regions with significant differences in^18F-FDG uptake between
   the Flea− (red, n = 6) and Flea+ (blue, n = 3) groups. Two-sided t test
   was used. Each black dot represents an individual. Asterisks indicate
   brain regions with significant differences. Data are presented as
   mean ± SEM. f–h Representative images of selected axial slices after
   spatial normalization. ARA standardized coordinates are indicated. PET
   brain activity is scaled to the cerebral global mean, with identical
   window/level settings applied to all animals. Source data are provided
   as a Source Data file. **p < 0.01, *p < 0.05.

Altered brain’s expression patterns

   To determine which brain regions are affected by flea bites, we mapped
   glucose uptake in the brain, using tail injections of
   2-deoxy-2-[fluorine-18] fluoro-D-glucose (18F-FDG)^[73]16, which is
   rapidly absorbed by active brain regions. We then aligned the metabolic
   images produced by PET-CT with the Allen Brain Atlas and calculated the
   standard uptake value for 18F-FDG in each brain region^[74]17. After
   flea bites, there was a significant increase in the correlation of
   activity across different brain regions (Fig. [75]2c), indicating that
   flea bites induce consistent changes in brain activity. Flea bites
   caused changes in glucose uptake in Basal Ganglia Dorsal Right (BG-DR),
   Hippocampal Right (HIP-R), Hypothalamus Left (HY-L), Hypothalamus Right
   (HY-R), Olfactory Left (OLF-L), Olfactory Right (OLF-R), Orbital Right
   (ORB-R), Somatosensory Right (SS-R), and Thalamus Right (TH-R)
   (Fig. [76]2d–h and Supplementary Fig. [77]2). Notably, many of these
   regions have also been linked to anxiety in previous
   studies^[78]18,[79]19. We concluded that flea bites lead to changes in
   multiple brain regions, including those involved in emotion
   regulation^[80]20–[81]22. Given the impact of flea bites on anxiety in
   mice, our subsequent research will focus on three key areas: the
   prefrontal cortex (PFC, representing orbital cortex), thalamus (TH),
   and hippocampus (HC).

   To uncover the molecular mechanisms underlying flea bite-induced
   anxiety in mice, we performed transcriptome sequencing on three brain
   regions from both groups. Cluster analysis showed significant
   differences in gene expression of flea bite in different brain regions
   (Fig. [82]3a), and high intergroup repeatability in brain regions
   (Supplementary Figs. [83]3a, [84]4a, and [85]5a). The total number of
   differentially expressed genes (DEGs) was similar across the three
   brain regions (Fig. [86]3b), indicating that they were affected to a
   comparable extent. However, the PFC exhibited a higher number of
   upregulated DEGs compared to the HC and TH (TH: 225; PFC: 277; HC: 179,
   Supplementary Figs. [87]3b, [88]4b, and [89]5b, and Supplementary
   data [90]1).

Fig. 3. Flea bites alter the expression patterns in the prefrontal cortex.

   [91]Fig. 3
   [92]Open in a new tab

   a Principal component analysis of all transcriptome data showing tight
   clustering of brain regions. b Differential gene analysis revealed the
   number of differentially expressed genes in the three tested brain
   regions, using one-way ANOVA. c GSEA analysis of transcriptomic data
   from three brain regions showed that the PFC was enriched with the most
   neural function-related pathways, supporting that flea bites have the
   greatest impact on PFC function. d KEGG pathway enrichment analysis of
   differential neurotransmitters in the PFC. Fisher’s Exact Test was
   used. e Neurotransmitters related to synthesizing dopamine,
   norepinephrine, and serotonin all showed intergroup differences based
   on PET-CT for Flea− (n = 6) and Flea+ (n = 3) groups; Thick bars
   indicate the interquartile range (IQR) around the median, and whiskers
   represent 1.5 times the interquartile range (maxima: Q3 + 1.5 × IQR,
   minima: Q1 − 1.5 × IQR). Data are presented as mean ± SEM. f
   Cell-specific enrichment analysis based on DEGs from the transcriptomic
   data of three brain regions: p-value < 0.05 in red. Fisher’s Exact Test
   was used. g Schematic of microglial cells in resting and activated
   states, created in BioRender. Liu, P. (2025)
   [93]https://BioRender.com/lrddnmc. h Representative flow cytometry
   plots of the PFC from the Flea− and Flea+ group. i Comparison of
   activated microglial cells between groups in the PFC using quantitative
   flow cytometry. Two-sided t test was used. Red represents the Flea−
   group (n = 6), and blue represents the Flea+ group (n = 6). Each black
   dot represents an individual. Asterisks indicates brain regions with
   significant differences. Data are presented as mean ± SEM. j
   Representative images of brain sections stained with IBA1 from Flea−
   and Flea+ mice; Scale 50 µm. k Comparison of IBA1 protein fluorescence
   intensity in brain sections between the Flea− (red, n = 5) and Flea+
   (blue, n = 5) groups of mice. Two-sided t test was used. Each black dot
   represents an individual. Asterisks indicates brain regions with
   significant differences. Data are presented as mean ± SEM. l Western
   blot images of IBA1 protein expression in the PFC of mice from the
   Flea− and Flea+ groups. Source data are provided as a Source Data file.
   **p < 0.01, *p < 0.05.

   We conducted functional enrichment analysis on the upregulated and
   downregulated DEGs in each brain region (Supplementary data [94]2).
   Some functions that are enriched with upregulated genes in all three
   regions include hemoglobin complexes, oxygen transport, and
   detoxification (Supplementary Figs. [95]3c, [96]4c, and [97]5c). This
   indicates a coordinated response across the regions, potentially
   enhancing oxygen transport and detoxification to protect against
   oxidative stress. In the PFC, additional upregulated functions included
   immune responses, metal ion regulation, erythrocyte development, and
   extracellular space functions. The downregulated genes in the PFC,
   associated with cell fate and neuronal differentiation, suggest
   significant alterations that may indicate long-term damage from flea
   bites (Supplementary Fig. [98]3d). In contrast, the HC and TH had
   downregulated genes linked to neuropeptide signaling and stress
   regulation (Supplementary Figs. [99]4d and [100]5d). Gene set
   enrichment analysis (GSEA) revealed that flea bites primarily affected
   PFC function, as a greater number of neurophysiological processes were
   suppressed in the PFC of the Flea+ group (PFC: 35, TH: 14, HC: 0;
   Fig. [101]3c). After flea bites, the PFC exhibited activated metabolic
   processes, energy production, immune responses, and inflammation, while
   functions related to synaptic plasticity, signal transduction, and
   neuron development were suppressed (Supplementary Figs. [102]3e–g,
   [103]4e–g, and [104]5e–g). This suppression may correlate with the
   decreased exploratory behavior observed in the mice^[105]23.

   We also conducted neurotransmitter metabolomics analysis on the three
   brain regions mentioned above. Both PCA and PLS-DA clustering analyses
   demonstrated high reproducibility of the experimental treatment across
   all brain regions (Supplementary Fig. [106]6a, b, d, e, g, h). The PFC
   exhibited 14 differential neurotransmitter metabolites (Supplementary
   Fig. [107]5c), more than 6 identified in the HC (Supplementary
   Fig. [108]6f) and 3 in the TH (Supplementary Fig. [109]6i). We
   performed KEGG pathway analysis on these differential metabolites. In
   the PFC, significant enrichment was observed in Tyrosine and Tryptophan
   metabolism pathways (Fig. [110]3d), which are crucial for synthesizing
   dopamine, norepinephrine, and serotonin. These neurotransmitters
   regulate mood, cognition, and pain perception, with pathway
   abnormalities potentially impacting neural excitability and signal
   transmission^[111]24,[112]25. Notably, we observed a decrease in
   tryptophan and serotonin levels in the Flea+ group (Fig. [113]3e). In
   the HC, enrichment of the Histidine metabolism pathway (Supplementary
   Fig. [114]6j), which influences histamine production—a neurotransmitter
   involved in wakefulness, attention, and appetite—was noted, though its
   role in emotion regulation is less clear^[115]26.

Microglia activation

   We infer that flea bites primarily impact the PFC, which may be linked
   to reduced exploratory behavior in mice. Cell-specific enrichment
   analysis revealed that neuron-specific expression decreased across all
   three brain regions, while microglia and endothelial cell expression
   increased in the PFC (Fig. [116]3f). Microglia, the brain’s resident
   macrophages, are responsible for clearing damaged neurons and promoting
   neuronal differentiation and development^[117]27. Genes related to
   microglial activation, such as DAP12, FCER1, were upregulated in the
   Flea+ group (Supplementary Fig. [118]7). To test the hypothesis that
   flea bites promote microglial activation (Fig. [119]3h), we conducted
   flow cytometry, which revealed a significant increase in CD45+ and
   CD11b+ cells in the PFC of Flea+ mice (Fig. [120]3h, i). This finding
   was further supported by immunofluorescence (Fig. [121]3j, k) and
   Western blot (WB) analyses (Fig. [122]3l and Supplementary
   Fig. [123]8a), confirming that flea bites preferentially induce
   microglial activation in the PFC.

Reduction in GABAergic neurons

   We observed a decreased expression of the mature neuron (NeuN),
   alongside an apoptotic (TUNEL) increase in the PFC (Fig. [124]4a–c). To
   determine whether damage to neuronal synapse was responsible for
   impaired neuronal differentiation and subsequent reduction in neuronal
   expression, we observed a decrease in PSD95 expression in the Flea+
   group, accompanied by an increase in the synaptic apoptosis marker
   Homer1, indicating damage to synaptic terminals (Fig. [125]4d–f).
   Furthermore, neuronal synaptic damage was accompanied by a reduction in
   synaptic vesicles (Fig. [126]4d, g). Transmission electron microscopy
   revealed that the synaptic terminals in the Flea+ group were more
   fragmented (Fig. [127]4h).

Fig. 4. Flea bites cause neuronal damage in the PFC and a reduction in
GABAergic neurons.

   [128]Fig. 4
   [129]Open in a new tab

   a Representative images of brain sections from Flea− and Flea+ mice;
   NeuN (red) and TUNEL (green). Scale 50 µm. b Comparison of mature
   neurons in brain sections between the Flea− (red, n = 5) and Flea+
   (blue, n = 5) groups of mice. Two-sided t test was used. Asterisks
   indicates brain regions with significant differences. Data are
   presented as mean ± SEM. c Comparison of cell apoptosis in brain
   sections between the Flea− (red, n = 5) and Flea+ (blue, n = 5) groups
   of mice. Two-sided t test was used. Asterisks indicates brain regions
   with significant differences. Data are presented as mean ± SEM. d
   Representative images of brain sections from Flea− and Flea+ mice;
   PSD95 (left), Homer1 (mid), and Syn (right). Scale 50 µm. e–g
   Comparison of synaptic structure: PSD95 (e), Homer1 (f), and Syn (g)
   between the Flea− (red, n = 5) and Flea+ (blue, n = 5) groups of mice.
   Two-sided t test was used. Asterisks indicates brain regions with
   significant differences. Data are presented as mean ± SEM; Scale 50 µm.
   h Representative transmission electron microscopy images.; Scale 500 nm
   (left) and 200 nm (right). Synaptic gaps between neurons are indicated
   by red arrows. i Western blot images of GAD65/67 expression in the PFC
   of mice from the Flea− and Flea+ groups. j Representative images of
   brain sections from Flea− and Flea+ mice; GAD65/67 (left), GABAaRγ2
   (mid) and VGLuT1(right). Scale 50 µm. k–m Comparison of synaptic
   structure: GAD65/67 (j), GABAaRγ2 (k), and VGLuT1 (l) between the Flea−
   (red, n = 5) and Flea+ (blue, n = 5) groups of mice. Two-sided t test
   was used. Each black dot represents an individual. Asterisks indicates
   brain regions with significant differences. Data are presented as
   mean ± SEM. n Differences in GABAergic neuron expression levels between
   Flea− and Flea+ after controlling for glutamatergic neuron expression
   levels (Flea− n = 5, Flea+ n = 5). Two-sided t test was used. Asterisks
   indicates brain regions with significant differences. Data are
   presented as mean ± SEM. Source data are provided as a Source Data
   file. ***p < 0.001, **p < 0.01, *p < 0.05.

   Given the transcriptomic evidence of downregulated GABAergic neuron
   differentiation, we investigated whether the damage and reduced
   expression of GABAergic neurons contributed to the observed effects.
   GABAergic neurons, responsible for releasing γ-aminobutyric acid
   (GABA), are the main inhibitory neurons in the central nervous system
   (CNS)^[130]28. When these neurons malfunction, mice remain in a
   heightened state of arousal, leading to increased anxiety-like
   behaviors^[131]29,[132]30. The presynaptic marker GAD65/67 and the
   postsynaptic marker GABARγ2 both showed decreases after flea bites
   (Fig. [133]4i–l and Supplementary Fig. [134]8b, c). Considering the
   broader impact on pan-neuronal expression, we also examined the
   glutamatergic neuron marker VGLUT1, which similarly showed a decrease
   in expression (Fig. [135]4j, m). Furthermore, even after controlling
   for glutamatergic neurons, the reduction in GABAergic neurons in the
   Flea+ group remained significant and was independent of any decrease in
   the number of mature neurons (Fig. [136]4n), indicating that flea bites
   primarily cause structural damage to GABAergic neurons, potentially
   linked to reduced exploratory behavior.

Systemic inflammation and PFC damage

   Flea bites may manipulate the host’s CNS through the skin-brain axis,
   which refers to the bidirectional communication between the skin and
   the CNS, influencing brain function and behavior^[137]31. When fleas
   feed on blood, they inject substances like anesthetics and
   anticoagulants into the host, which the host’s immune system recognizes
   as antigens, initiating an immune response. The skin transcriptome
   showed high reproducibility between the Flea− and Flea+ groups
   (Supplementary Fig. [138]9a). We analyzed the skin transcriptome data,
   identifying 103 upregulated and 71 downregulated DEGs (Supplementary
   Fig. [139]9b). The upregulated DEGs were associated with numerous
   immune-related pathways (Fig. [140]5a and Supplementary Fig. [141]9c),
   including inflammation and oxidative damage, while no significant
   functions were enriched in the downregulated genes (Supplementary
   Fig. [142]9d). GSEA and the protein-protein interaction (PPI) network
   analysis highlighted that flea bites provoke an immune defense response
   in mice and cause vascular dilation (Supplementary Fig. [143]9e, f). We
   measured serum IgE levels, as previous research suggests that small
   antigens are first recognized by IgE, which activates mast cells to
   release histamine, initiating an immune response via IgG binding. In
   the Flea+ group, we observed an increase in serum IgE and skin IgG
   levels, accompanied by elevated expression of inflammatory cytokines
   that promote inflammation and oxidative stress (Fig. [144]5b). However,
   HE staining did not reveal significant skin damage (Supplementary
   Fig. [145]9g), indicating that flea bites may cause only mild
   inflammation in mice. This observation aligns with the finding that the
   downregulated gene set in the transcriptome did not show significant
   functional enrichment.

Fig. 5. Flea bites affect neural function through the skin-brain axis.

   [146]Fig. 5
   [147]Open in a new tab

   a GSEA result on upregulated genes from the mice skin transcriptome.
   Functions related to immunoglobulin and hemoglobin genes are
   significantly enriched. b Cytokine levels in mice skin presented in
   boxplot where red is Flea− (n = 6/12) and blue is Flea+ (n = 6/12).
   Two-sided t test was used. Asterisks indicate the significance. All
   p-values were below 0.001. c Cytokine levels in mice serum presented in
   boxplot where red is Flea− (n = 12) and blue is Flea+ (n = 12).
   Two-sided t test was used. Thick bars indicate the interquartile range
   (IQR) around the median, and whiskers represent 1.5 times the
   interquartile range (maxima: Q3 + 1.5 × IQR, minima: Q1 − 1.5 × IQR).
   Asterisks indicate the significance. All p-values were below 0.001. d
   Cytokine levels in mice PFC presented in boxplot where red is Flea−
   (n = 10) and blue is Flea+ (n = 10). Two-sided t test was used.
   Asterisks indicate the significance. All p-values were below 0.001. e
   Representative HE plots of the PFC. In the Flea+ group, cortical tissue
   shows evident cellular damage, with minor mononuclear cell infiltration
   (yellow arrow) at the injury site. Additionally, widespread nuclear
   pyknosis and condensation (red arrow) are observed in the parenchyma.
   The infected group exhibits a larger area of damage; Scale 50 µm. f
   Western blot analysis of Claudin1 markers of tight junction protein
   from the PFC brain region (n = 5 samples each group). g Comparison of
   Claudin1 in WB between the Flea− (red, n = 5) and Flea+ (blue, n = 5)
   groups of mice. Two-sided t test was used. Each black dot represents an
   individual. Asterisks indicates brain regions with significant
   differences. Data are presented as mean ± SEM. Source data are provided
   as a Source Data file. h Validation by qPCR of select BBB genes altered
   between Flea− (n = 3) and Flea+ (n = 3) group. Two-sided t test was
   used. Thick bars indicate the interquartile range (IQR) around the
   median, and whiskers represent 1.5 times the interquartile range
   (maxima: Q3 + 1.5 × IQR, minima: Q1 − 1.5 × IQR). Each black dot
   represents pooled sample of 3 decapitated individuals. Asterisks
   indicate the significance. ***p < 0.001, **p < 0.01, *p < 0.05.

   To explore whether flea bites can trigger systemic inflammation, we
   assessed serum cytokine levels and observed increased stress,
   inflammation, and oxidative stress in mice (Fig. [148]5c). Further
   ELISA analysis of the PFC reveals elevated levels of inflammatory and
   oxidative markers (Fig. [149]5d). Consistently, HE staining shows
   noticeable cellular damage in the PFC of the Flea+ group, with a small
   number of infiltrating mononuclear cells at the injury site and
   extensive nuclear pyknosis and condensation leading to deep staining in
   the parenchyma. Additionally, the Flea+ group exhibits a larger area of
   damage (Fig. [150]5e). WB analysis showed a decrease (~50%) in Claudin1
   protein expression in the Flea+ group (Fig. [151]5f, g). Although this
   was not supported by the qPCR results (Fig. [152]5h), we still believe
   that systemic inflammation triggered PFC inflammation by disrupting the
   blood-brain barrier (BBB). In summary, flea bites trigger inflammation
   and oxidative stress in the skin, elevate stress levels, and are
   associated with systemic inflammation, which may contribute to PFC BBB
   alterations. This may lead to inflammation and oxidative changes in the
   PFC, potentially contributing to microglial activation.

Reduced exploratory behavior in Flea+ striped hamster

   We wanted to determine whether the observed reduction in exploratory
   behavior due to ectoparasite infestation in hosts could be generalized
   to other parasite-host pairs in natural settings. To test this, we
   conducted a laboratory infection experiment using field-captured
   striped hamsters and their associated fleas (Fig. [153]6a and
   Supplementary Fig. [154]10a–c). After 4 weeks of infection, we assessed
   exploratory levels in the hamsters using an OFT. Compared to the Flea−
   group, female striped hamsters in the Flea+ group spent less time in
   the central area (Fig. [155]6b); however, there was no difference in
   the time spent in the central area between the Flea− and Flea+ groups
   of male striped hamsters (Supplementary Fig. [156]10d), similar to the
   findings in mice. Additionally, heightened immune responses, increased
   inflammation, and oxidative stress were detected in both the skin and
   serum of the infected hamsters based on elisa results (Fig. [157]6c,
   d).

Fig. 6. Striped hamsters exposed to flea bites exhibited reduced exploratory
behavior in both indoor infection and outdoor enclosure experiments.

   [158]Fig. 6
   [159]Open in a new tab

   a Schematic diagram of laboratory infection in striped hamsters,
   created in BioRender. Liu, P. (2025)
   [160]https://BioRender.com/ps1cm2y. b Open field test for striped
   hamster, n = 5 each group, tested using a Generalized Linear Model
   (GLM). Thick bars indicate the interquartile range (IQR) around the
   median, and whiskers represent 1.5 times the interquartile range
   (maxima: Q3 + 1.5 × IQR, minima: Q1 − 1.5 × IQR). Each black dot
   represents an individual. Asterisks indicate the significance.
   (P[female] = 0.03, P[male] = 0.26). c Cytokine levels in striped
   hamster skin presented in boxplot where red is Flea− (n = 3–7) and blue
   is Flea+ (n = 7–9). Two-sided t test was used. Thick bars indicate the
   interquartile range (IQR) around the median, and whiskers represent 1.5
   times the interquartile range (maxima: Q3 + 1.5 × IQR, minima:
   Q1 − 1.5 × IQR). Each black dot represents an individual. Asterisks
   indicate the significance. d Cytokine levels in striped hamster serum
   presented in boxplot where red is Flea− (n = 5) and blue is Flea+
   (n = 5). Two-sided t test was used. Thick bars indicate the
   interquartile range (IQR) around the median, and whiskers represent 1.5
   times the interquartile range (maxima: Q3 + 1.5 × IQR, minima:
   Q1 − 1.5 × IQR). Each black dot represents an individual. Asterisks
   indicate the significance. e Schematic diagram of the three rounds of
   enclosure field experiments under natural conditions. f Bar graph of
   the capture rate of striped hamsters released into the enclosures. g
   Forest plot illustrating that Flea+ (n = 72) striped hamsters had a
   significantly higher escape rate than Flea− (n = 62) hamsters, tested
   using a generalized linear mixed model (GLMM). The x-axis represents
   the coefficients of various predictor variables. Solid circles indicate
   statistically significant results, while hollow circles represent
   non-significant results. Data are presented as mean values ± SEM. h
   Field sampling sites for striped hamsters, selected based on
   temperature and precipitation gradients. Created using the open-source
   R packages ggplot2 (MIT License). i The number of striped hamsters
   captured during each sampling event at each site, with no capture
   records for striped hamsters at site N5. Source data are provided as a
   Source Data file. ***p < 0.001, **p < 0.01, *p < 0.05.

   To further evaluate whether this behavioral phenotype persisted in
   natural conditions, we conducted a field enclosure experiment, a
   larger-scale OFT (Supplementary Fig. [161]10e). We hypothesized that
   flea-bitten striped hamsters would spend more time near the borders of
   the enclosure and be more likely to escape, resulting in a higher
   non-capture rate. On the contrary, a hamster that explores more will
   end up in the center of the enclosure and is more likely to be
   recaptured. Before starting the experiment, we removed any existing
   rodents and ectoparasites from the enclosures and confirmed that there
   were no significant differences in plant diversity between the
   enclosures (Supplementary Fig. [162]10f). We conducted three rounds of
   enclosure infection experiments with striped hamsters: no fleas were
   introduced in any enclosures-R1, fleas were introduced in some
   enclosures-R2, and fleas were introduced in all enclosures-R3
   (Fig. [163]6e and Supplementary Fig. [164]10g). In the first round,
   most hamsters remained in their respective enclosures, with few
   escaping. In the second round, the flea-infested hamsters exhibited a
   higher non-capture rate. In the third round, we observed a significant
   increase in non-capture rates compared to the first round, but no
   difference between the enclosures (Fig. [165]6f). Flea bites
   significantly increased the non-capture rate of striped hamsters.
   However, other factors, particularly host sex and enclosure density,
   did not affect the likelihood of escaping the enclosure (Fig. [166]6g
   and Supplementary Fig. [167]10h). Overall, the semi-natural enclosure
   experiment demonstrated that flea bites reduce host exploratory
   behavior, thereby decreasing the likelihood of the host leaving its
   natal habitat.

Reduced occupancy in Flea+ host

   We used individual-based mechanistic models that integrate host habitat
   preferences, demographic indicators, and migration patterns to simulate