Highlighting zoonotic importance of synanthropic fleas through microbiome analysis

Abstract

The role of domestic and peridomestic animals in vector-borne diseases is critical as they share a common environment with people having the potential to extend the network of pathogen transmission to humans. In the present study, amplicon sequencing was employed to characterize the microbial communities associated with five flea species (Archaeopsylla erinacei, Ctenocephalides felis, Spilopsyllus cuniculi, Pulex irritans and Ctenocephalides canis) collected from dogs, cats, and hedgehogs in Andalusia (Spain). The analysis focused on identifying the presence and infection rate of pathogenic bacteria within these synanthropic flea populations. The higher relative abundance of the Phylum Pseudomonadota was primarily attributed to the presence of the endosymbiont Wolbachia, along with consistently elevated levels of the genera Rickettsia and Bartonella across all flea species. This study reports, for the first time, the detection of Babesia sp. in all tested flea species, with the highest abundance observed in S. cuniculi collected from cats, emphasizing the need for further investigation into its potential implications as vectors. Our results also demonstrate that the microbiota composition of fleas is largely influenced by the host they parasitize. The study of microbiota allowed for the ecological separation of flea species, with individuals from these five species clustering distinctly each other.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-21396-6.

Introduction

The Order Siphonaptera includes a variety of species and subspecies parasitizing different hosts. Many of these species and subspecies have been designated based on minimal and subtle morphological differences (presence and number of silks, male and female genitalia and quetotaxy, etc.). Consequently, some of these classifications may represent undiagnosed synonyms in the absence of molecular analyses. In this context, molecular taxonomy has proven instrumental in resolving long-standing enigmas and contradictions. Indeed, since the work published by Whiting et al.¹, numerous phylogenetic studies conducted by various authors^2–6 have successfully elucidated evolutionary relationships within specific flea groups.

Fleas play a prominent role as vectors of diverse pathogens. Dogs and cats are implicated in the changing epidemiology of pathogens of public health concern^7–10. Thus, Colella et al.⁹ investigated the presence of pathogens and ectoparasites in pet dogs and cats living in metropolitan areas near humans. These animals share a common environment with humans, which makes them likely key reservoirs for pathogens with the potential to infect people living in such areas and settings. Therefore, companion animals, particularly dogs, play a key role in the transmission dynamics of vector-borne diseases, as their close contact with humans facilitates the expansion of pathogen transmission networks. The most significant vector-borne infectious diseases that are shared by man, dogs and cats are leishmaniosis, borreliosis, bartonellosis, ehrlichiosis, rickettsiosis and anaplasmosis¹¹ with fleas acting as main vectors of pathogens in the transmission of some of these diseases^12,13.

One Health is the multisectoral approach to public health that recognizes the interactions and interdependencies between environmental, animal and human health to prevent and control the risks arising from such interfaces. The impact of environmental and lifestyle changes (e.g. climate change, deforestation and urbanization) and lifestyle change, including the increase in global human and animal movement¹⁴, has also become part of the One Health concept.

The potential role of companion animals, and particularly domesticated dogs and cats, in One Health, is often underestimated. In developed countries, pet ownership has reached unprecedented levels, and these animals play a significant role in family life¹¹. Over the preceding decades, small companion animals have increasingly spent a significant portion of their lifespan within the indoor domestic sphere, maintaining close physical proximity to their owners. However, beyond the perceived benefits of companionship, it is crucial to acknowledge the potential for zoonotic infectious disease transmission, which may occur through direct or indirect contact with these animals. Day¹¹ documented the principal flea-borne infectious diseases shared between companion animals and humans, with Bartonellosis, Rickettsiosis, and Yersiniosis identified as the most frequently reported. Bartonellosis warrants recognition as a significant potential emerging human infection. The most extensively characterized pathology is ‘cat scratch disease’, primarily attributed to Bartonella henselae, acquired through direct contact with infected felines and likely transmitted among cats via the cat flea, C. felis^15,16. Prior to the 1990 s, the genus Bartonella comprised a single species, Bartonella bacilliformis¹⁷. Subsequently, a marked exponential increase in both research investigations and clinical case reports concerning members of the genus Bartonella has been observed, culminating in the description of novel species isolated from diverse environmental sources¹⁸. Currently, clinic-pathological data in Europe are more limited, suggesting that bartonellosis may be an infrequent or underdiagnosed infectious disease in cats and dogs. Further investigation is warranted to definitively ascertain the role of Bartonella infection in the etiology of a range of feline and canine diseases. On a comparative medical basis, different clinical manifestations, such as periods of intermittent fever, granulomatous inflammation involving the heart, liver, lymph nodes and other tissues, endocarditis, bacillary angiomatosis, peliosis hepatis, uveitis and vasoproliferative tumors have been reported in both pets and humans¹⁹. Within the framework of human medicine, the One Health initiative, and pet ownership, B. henselae, Bartonella koehlerae, and Bartonella vinsonii berkhoffii are recognized as the three Bartonella species most implicated in diverse pathologies. Notably, cats serve as reservoir hosts for B. henselae and B. koehlerae, exhibiting the capacity for subclinical infection over extended periods, spanning months and potentially years²⁰.

Relevant flea-borne diseases in Europe include plague (caused by Yersinia pestis), murine typhus (caused by Rickettsia typhi), flea-borne spotted fever (Rickettsia felis), and cat scratch disease (B. henselae). Fleas may also transmit other animal diseases: myxomatosis and Trypanosoma nabiasi in rabbits (by S. cuniculi), Bartonella quintana (by C. felis) or caprine mycoplasmal polyarthritis (via ‘goat fleas’)²¹. Additionally, fleas can serve as intermediate hosts for tapeworms. Dipylidium caninum (a typical cestode of dogs, cats and, in some accidental cases, humans) is transmitted by ingesting fleas infected with the larval stages of Ctenocephalides spp. Similarly, Hymenolepis nana and Hymenolepis diminuta (both species reported in rodents and humans) develop in flea intermediate hosts (e.g., Nosopsyllus sp., Xenopsylla sp.). Further, the hedgehog flea (A. erinacei) has been cited for its possible potential vector capacity in the transmission of certain microorganisms including C. burnetii, Rickettsia spp., Wolbachia spp., Mycobacterium spp. and various Bartonella species for which this synanthropic mammal could act as a reservoir²². Recently, molecular techniques, especially the improved metagenomic Next-Generation Sequencing (mNGS), have provided alternatives for rapidly identifying and distinguishing microorganisms^23–25. In fact, the introduction of deep-sequencing technologies applied to microbiome investigation is pertinent in the progression of vector-borne pathogen detection and surveillance in public health²⁶.

In the present study, amplicon sequencing was applied to detect and understand the composition of the microbial communities of adults of distinct species of fleas collected on dogs, cats and hedgehogs from Andalusia (Spain) to estimate the infection rate of pathogenic bacteria among fleas of these hosts in Andalusia.

Materials and methods

Sample collection

We obtained flea samples from dogs, cats and one hedgehog. To collect fleas from the hosts mentioned, we contacted various veterinary clinics, veterinary hospitals, pet shelters and some pet owners. Out of these, 18 centers agreed to collaborate in the collection of samples (see Acknowledgements). They all participated in this sampling voluntarily. The informed consent of pet owners, veterinary clinics and dog kennels chairman was obtained for the inclusion of the hosts in the study. The sample collection period encompassed late June 2021 to January 2023, and the collection procedure was described previously by Zurita et al.²⁷. Adult flea counts on hosts were conducted as described in the World Association for the Advancement of Veterinary Parasitology guidelines²⁸. Catching and handling procedures of animals and experimental protocols in this study were approved by the Institutional Animal Care of the University of Seville. All methods were performed in accordance with the guidelines presented in RD 53/2013 of January 1 st, 2013 (BOE number 34 of February 8th, 2013) of the Spanish government. No anesthesia substances or euthanasia methods were applied for animal handling procedures in this study.

A total of fifty-two fleas were selected for this study, including male and female representative specimens and different hosts (hedgehogs, dogs and cats), whenever possible. Fleas were sexed and identified to species using a CX21 microscope (Olympus, Tokyo, Japan). Diagnostic morphological characters of all the samples were studied by comparison with figures, keys and descriptions reported by Beaucournu & Launay²⁹. The fleas were also identified by amplification and sequencing of two different molecular markers: the partial cytochrome c oxidase subunit 1 (cox1) and Internal Transcribed Spacer 2 (ITS2), assuming a reliable specific identification when the sequence reached at least a 99% similarity with the correspondent NCBI sequence. To obtain the percentage of nucleotide similarity among sequences, we used the BLAST algorithm (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

DNA extraction, high-throughput sequencing (HTS) and data processing of bacterial communities

For DNA extraction, the abdomen of each flea specimen under study was dissected using previously sterilized scalpels. Total DNA was then obtained from each flea´s abdomen using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA, U.S.A.) following the manufacturer’s specific protocol for Purification of Total DNA from Animal Tissues (Spin-Column Protocol). DNA concentration and purity was checked using an electrophoresis in 0.8% agarose gel electrophoresis infused with SYBR Safe. The DNA was stored at − 20 °C until molecular analysis. For molecular identification, both molecular markers (ITS2 rDNA and cox1 mtDNA) were amplified by a polymerase chain reaction (PCR) using a thermal cycler (Eppendorf AG; Eppendorf, Hamburg, Germany). PCR mix, PCR conditions and PCR primers are summarized in the Supporting information provided in Zurita et al.⁵. Negative controls were added for both DNA extraction and DNA amplification. As negative controls we used 5 µl of sterile distilled water replacing the DNA extract.

After molecular identification, the remaining total DNA was submitted to a commercial service provider for analysis of sequence variation in the 16 S ribosomal RNA (rRNA) bacterial gene region V4 (Novogene; Beijing, China). Regarding the amplicon generation, the bacterial communities were characterized by Illumina Sequencing for the 16 S rRNA V4 gene region. DNA was amplified for this hypervariable region with specific primers and further reamplified in a limited-cycle PCR reaction to add sequencing adaptor and dual indexes. The forward primer used was 515 F 5′-GTGCCAGCMGCCGCGGTAA-3′ and the reverse primer was 806R 5′-GGACTACHVGGGTWTCTAAT-3′³⁰. First, PCR reactions were carried out with 15 µL of Phusion High-Fidelity PCR Master Mix (New England Biolabs); 2 µM of forward and reverse primers, and about 10 ng template DNA. Thermal cycling consisted of initial denaturation at 98℃ for 1 min, followed by 30 cycles of denaturation at 98℃ for 10 s, annealing at 50℃ for 30 s, and elongation at 72℃ for 30 s, and a final extension step at 72 C for 5 min. Negative controls (without DNA), which were run through the extraction kit, were also included for all amplification reactions. These PCR products were stored at 4 °C.

PCR products quantification, qualification, mixing and purification

The electrophoresis of the PCR products was undertaken on a 2% (w/v) agarose gel for detection using 1X loading buffer with SYBR Green. Samples with the bright main strip between 400 and 450 base pairs (bp) were chosen for further experiments.

PCR products were mixed in equidensity ratios. Then, mixture PCR products were purified with Qiagen Gel Extraction Kit (Qiagen, Germany), following manufacturer’s instructions.

Library preparation

Samples passing quality control analysis were used for library preparation. Sequencing libraries were generated using TruSeq DNA PCR- Free Sample Preparation Kit (Illumina, USA) following manufacturer’s recommendations and index codes were added. The library quality was assessed on the Qubit 2.0 Fluorometer (Thermo Scientific, USA) and the Bioanalyzer 2100 system (Agilent, USA). At last, the library was sequenced on an Illumina NovaSeq platform and 250 bp paired-end reads were generated.

Data analysis

Paired- end reads were assigned to samples based on their unique barcode and truncated by cutting off the barcode and primer sequence. The whole process was performed through Python (V3.6.13) and adaptors were removed through cutadapt (V3.3)³¹.

Paired-end reads were merged using FLASH V1.2.7³², and the splicing sequences were called raw tags. Quality filtering on the raw tags was performed under specific filtering conditions to obtain the high-quality clean tags³³ according to the QIIME V1.9. quality-controlled process³⁰. Chimeric merged reads were detected and removed using UCHIME³⁴ against the Silva database³⁵. Sequences were clustered into operational taxonomic units (OTU) with a 97% similarity threshold. Sequence was picked as a representative for each OTU and the Silva database³⁶ was used based on Mothur algorithm to annotate taxonomic information. Multiple sequence alignment was conducted using MUSCLE software (Version 3.8.31)³⁷ to study the phylogenetic relationship of different OTUs and the differences in the dominant species across the samples.

The OTU was used for alpha diversity analysis, including Chao1 and Shannon indexes. Beta diversity was explored by Unweighted Unifrac distance, which allowed us to calculate the distance between samples by using the evolutionary information between microbial sequences in each sample. Then, the Weighted UniFrac distance (a variant of the original Unweighted unifrac) was further constructed using the abundance information of the feature sequence³⁸. At the same time, we used a hierarchical clustering method called Unweighted Pair Group Method with Arithmetic averages (UPGMA) to cluster the community samples³⁸. These analyses were conducted by QIIME software (V1.9. 1) and displayed with R software (Version 2.15.3).

Results

Fleas collected from dogs, cats and one hedgehog were identified as A. erinacei, C. felis, S. cuniculi, P. irritans and C. canis (Table 1) based on morphological and morphometric features and molecular analysis. To facilitate the interpretation of the results, all specimens used in this study were grouped based on their species name, gender, and host, resulting in the identification of 13 main groups summarized in Table 1: CFP1 (males of C. felis collected from dogs), CFP2 (females of C. felis collected from dogs), CFG1 (males of C. felis collected from cats), CFG2 (females of C. felis collected from cats), AEP1 (males of A. erinacei collected from dogs), AEP2 (females of A. erinacei collected from dogs), AEE1 (males of A. erinacei collected from one hedgehog), AEE2 (females of A. erinacei collected from one hedgehog), CCP1 (males of C. canis collected from dogs), PIP1 (males of P. irritans collected from dogs), PIP2 (females of P. irritans collected from dogs), SCG1 (males of S. cuniculi collected from cats) and SCG2 (females of S. cuniculi collected from cats). These groups were not homogeneous in terms of sample size and were conditioned by the greater specificity of certain flea species with respect to their hosts (dogs and cats).

Table 1.

Flea specimens collected in this study, grouped by species, gender, and hosts.

Flea group	Number of fleas/species/gender	Hosts/habitat	Geographical origin
CFP1	2/C. felis/males	Dog/rural farm	Córdoba/Spain
		Dog/domestic-rural	Córdoba/Spain
CFP2	16/C. felis/females	Dog/domestic-urban	Seville/Spain
		Dog/stray dog	Seville/Spain
		Dog/rural house	Seville/Spain
		Dog/rural farm	Córdoba/Spain
		Dog/rural farm	Córdoba/Spain
		Dog/domestic-urban	Huelva/Spain
		Dog/shelter dog	Huelva/Spain
		Dog/rural farm	Huelva/Spain
		Dog/pet shelter	Huelva/Spain
		Dog/domestic-urban	Cádiz/Spain
		Dog/pet shelter	Cádiz/Spain
		Dog/domestic-urban	Cádiz/Spain
CFG1	5/C. felis/males	Cat/domestic-urban	Córdoba/Spain
		Cat/domestic-rural	Cádiz/Spain
		Cat/stray cat	Seville/Spain
		Cat/stray cat	Seville/Spain
		Cat/stray cat	Seville/Spain
CFG2	12/C. felis/females	Cat/stray cat	Seville/Spain
		Cat/stray cat	Seville/Spain
		Cat/rural farm	Seville/Spain
		Cat/domestic-rural	Seville/Spain
		Cat/domestic-rural	Seville/Spain
		Cat/rural farm	Córdoba/Spain
		Cat/rural farm	Córdoba/Spain
		Cat/domestic-rural	Huelva/Spain
		Cat/domestic-urban	Huelva/Spain
		Cat/stray cat	Cádiz/Spain
		Cat/domestic-urban	Cádiz/Spain
AEP1	1/A. erinacei/male	Dog/domestic-urban	Seville/Spain
AEP2	3/A. erinacei/females	Dog/domestic-rural	Córdoba/Spain
		Dog/domestic-urban	Seville/Spain
AEE1	2/A. erinacei/males	Hedgehog/domestic-rural	Seville/Spain
AEE2	2/A. erinacei/females	Hedgehog/domestic-rural	Seville/Spain
CCP1	2/C. canis/males	Dog/domestic-urban	Huelva/Spain
		Dog/rural farm	Seville/Spain
PIP1	1/P. irritans/male	Dog/domestic-rural	Huelva/Spain
PIP2	2/P. irritans/females	Dog/domestic-rural	Córdoba/Spain
		Dog/rural farm	Seville/Spain
SCG1	1/S. cuniculi/male	Cat/rural farm	Córdoba/Spain
SCG2	3/S. cuniculi/females	Cat/rural farm	Córdoba/Spain

The top ten taxa relative abundance in each taxonomic rank was calculated including Phylum, Order, family and genus. A total of forty-three phyla, 710 genera and 371 species were detected. The ten most abundant phyla were (showed in descending order): Pseudomonadota (Proteobacteria), Bacillota (Firmicutes), Actinomycetota (Actinobacteriota), Bacteroidota, Euryarchaeota, Cyanobacteriota, Coprothermobacterota, Patescibacteria, Myxococcota and Spirochaetota (Fig. 1).

Fig. 1.

Phylogenetic tree of bacterial phyla identified with 16 S sequences showing the relative abundance for each flea group identified in this study (see material and methods section).

Pseudomonadota was present, with a high percentage, in all the species of fleas analyzed (Figs. 1 and 2A). The highest Relative Abundance (RA) (Table S1), in this Phylum, was observed in AEE2 (0.969098) and PIP2 (0.956493), while C. canis showed the lowest one (RA: 0.674990). Further, Phylum Bacillota was present in all species of fleas but with values of relative abundance that ranged from 0.205137 (CCP1) to 0.016228 (AEE2) (Table S1). Among families (Table S2), Anaplasmataceae appeared as the most common one, which was present in all flea species and showed the highest abundance values of both PIP1 and PIP2 and AEP1, while the lowest values were observed in CCP1, AEP2 and CFG2. In these latter two groups, the highest abundance was reflected in the families Rickettsiaceae and Rhizobiaceae, respectively (Table S2). On the other hand, the high abundance of the Neisseriaceae in CFP2 group was noteworthy compared to the rest of the species (Table S2). The relative abundance at the genus level resulted in the presence of nine main genera present in all flea species listed below in decreasing order: Wolbachia, Bartonella, Rickettsia, Ricketsiella, Spiroplasma, Staphylococcus, Pseudomonas, Streptococcus, and Ralstonia. Among these genera, Wolbachia stood out with the highest abundance in PIP1 and PIP2, whereas Rickettsia sp. appeared as the most common genus within the microbiota of both AEP2 and AEE2 groups (Fig. 2B). Despite that, the microbiota defined by genus observed in the individuals of each group of species appeared inconsistent without any defined pattern of abundance, varying among individuals of the same genus. Thus, the highest abundance of the genus Wolbachia (RA > 0.75) was observed in 3 females of C. felis and one female of P. irritans, both isolated from dogs. In general, the highest relative abundance of the genus Bartonella was observed in C. felis isolated from cats. Further, Rickettsia sp. was present in all individuals of each flea species to a greater or lesser extent. It should also be highlighted the relative abundance of Staphylococcus, Streptococcus, Pseudomonas and Ralstonia in CCP1 (Figs. 2B and 3B) and the status of Babesia sp. as the species with the highest relative abundance observed in all flea species, especially in SCG2 (Fig. 2C). Since the sample size varied between the different groups observed, the results obtained are not conclusive, but indicative of the abundance values reflected by genus.

Fig. 2.

Barplot showing top ten most abundant bacterial phyla (A), genera (B) and species (C) for each flea group identified in this study (see material and methods section).

Fig. 3.

Heatmap showing the taxonomic abundance of bacterial phyla (A) and genera (B) for each flea group identified in this study (see material and methods section).

According to the abundance information of the top 35 genus of all samples, the heatmap was drawn to check whether the samples with similar processing are clustered or not, while the similarity and differences of samples can also be observed. The result is shown in Fig. 3. Thus, the heatmap observed at the Phyla level showed a downward resolution for the Phylum Pseudomonadota in CCP1, while this flea species showed an upward resolution for the Phyla Bacillota, Myxococcota, Actinobacteriota and Cyanobacteria (Fig. 3A). Further, in Fig. 3B we can observe an upward resolution for the genus Bacillus, Enterococcus, and other genera in SCG1 as well as an upward resolution for those mentioned above, Staphylococcus, Pseudomonas, Corynebacterium and Ralstonia for CCP1for genus.

To investigate the existence of unique and shared 16 S OTUs between groups we first studied the differences observed among individuals of the same species. Thus, we detected that females of A. erinacei from hedgehog (AEE2) had slightly more unique 16 S OTUs than males (AEE1), although this pattern of higher genetic abundance in females compared to males was only observed in A. erinacei. On the other hand, specimens belonging to AEP1 showed a high increase in the number of unique 16 S OTUs (n = 664) (Fig. 4A). Comparative analysis revealed a decrease in the number of unique 16 S OTUs present in the AEP1 group relative to CFP1, suggesting the existence of shared 16 S OTUs with C. felis, which exhibited a high number of unique 16 S OTUs (Fig. 4B). Although the CFP1 group maintained an elevated number of unique 16 S OTUs when compared to PIP1 and AEP1 (Fig. 4C), this number decreased in comparison with CCP1 and CFG1, indicating a greater overlap of OTUs (n = 723) among C. felis specimens collected from different hosts (dogs and cats) (Fig. 4D).

Fig. 4.

Venn diagram showing the number of both unique and shared genes among different flea groups identified in this study (see material and methods section).

Boxplots diagram was used to analyze the differences in Alpha Diversity indices between groups. The highest alpha diversity calculated using the Shannon index (Table S3) was observed in CCP1 and SCG2, while the lowest value was observed in PIP2 and AEP2 (Fig. 5A, table S3). Nevertheless, the Chao 1 index, based on the number of rare species in the sample, showed the highest diversity in CFP1, CFG1 and CFG2 (Fig. 5B, table S3).

Fig. 5.

Boxplots indicating alpha diversity indices: Shannon (A) and Chao (B) among different flea groups identified in this study (see material and methods section).

Beta diversity data, on the other hand, showed significant differences among all assessed flea species groups, whereas Weighted and Unweighted UniFrac were used to compare microbial communities by incorporating phylogenetic relationships. Since the weighted (quantitative) Unifrac accounts for abundance of observed organisms diminish the impact of low-abundance features, we observed low differences values (Fig. 6A) in community structure, while Unweighted (qualitative) Unifrac based on their presence or absence, is more sensitive to differences in low-abundance features. Thus, the lowest beta diversity value (< 0.5) was observed among all different groups of C. felis each other as well as between CCP1 and all the C. felis groups defined in this study (Fig. 6A). However, clear differences, with higher values, were observed for S. cuniculi, P. irritans and A. erinacei when compared to each other and with Ctenocephalides species (Fig. 6A).

Fig. 6.

Beta diversity Weighted (up data) & unweighted (down data) heatmap (A) and UPGMA Unweighted Unifrac group tree (B) for all the flea groups identified in this study (see material and methods section).

Clustering analysis together with its derived tree were displayed with the integration of clustering results and the relative abundance of each group by Phylum in Fig. 6B. Based on these results, we observed a clade that included both congeneric species, C. felis and C. canis irrespective of host, whereas P. irritans was positioned as the most distant species. In contrast, A. erinacei demonstrated a clear influence of host type on its clustering. Specimens segregated into two distinct clades, correlated with the host from which they were collected (hedgehogs or dogs).

Discussion

In Europe, flea-borne diseases receive comparatively less attention than other vector-borne diseases regarding targeted surveillance, research, and control. A major knowledge gap exists concerning the risk of pathogen transmission from fleas to humans and pets, primarily due to the paucity of epidemiological data for cosmopolitan and promiscuous fleas, such as P. irritans (house fleas) and X. cheopis (rat flea). Furthermore, prevalence data for most flea-borne pathogens, specifically relating to vector species and vertebrate reservoirs, including humans, are limited. Consequently, focused and detailed studies are imperative to generate precise information regarding this critical medical and veterinary concern³⁹. The findings from this type of research can provide valuable insights for epidemiological surveillance, the assessment of public and animal health risks, and the development of more effective intervention strategies.

In the present work, we carried out a study of the microbiota of fifty-two flea specimens (males and females) belonging to five different synanthropic species: C. felis, C. canis, P. irritans, A. erinacei and S. cuniculi. It is worth noting that all the species identified in our study belong to the family Pulicidae, which once again highlights the epidemiological and medical importance of this family. As demonstrated in this study and corroborated by previous findings²⁷, Pulicidae encompasses most flea species with frequent human contact, thereby representing a key family in the potential transmission of infectious agents. To the best of our knowledge, this study offers one of the first comparative overviews of the microbiota of up to five flea species examined concurrently, which are particularly significant due to their shared domestic and peridomestic habitats. Additionally, as far as we know, detailed microbiota data for the species S. cuniculi and A. erinacei are presented here for the first time.

The Phylum Pseudomonadota is frequently reported as the dominant bacterial group within the microbiota of various flea species, including C. felis, C. canis, P. irritans^40,41, and Echidnophaga ambulans ambulans⁴². This high relative abundance is largely attributed to the presence of the endosymbiont Wolbachia, prevalent across arthropods⁴³, and to genera such as Rickettsia and Bartonella, which are recognized as significant pathogens transmitted by synanthropic flea species like C. felis^40,44. Despite that, Wu et al.⁴¹ reported the absence of Bartonella species in C. felis isolated from cats in China, although they cited the existence of genera such as Cupriavidus, Puccinia, Massilia and Lysobacter that were absent both, in our results and in those published by other authors^40,45. These discrepant results regarding the common presence of the genus Bartonella in the microbiota of the species C. felis suggest that there might be a certain geographic influence on the occurrence of this bacterial genus in this flea species. Additionally, it has been demonstrated that the cat flea (C. felis) species is composed of multiple mtDNA haplotypes with different global distributions³. Although the capacity of different C. felis haplotypes to act as reliable vectors of Rickettsia and Bartonella spp. is currently not addressed^42,46, the existence of different haplotypes could have an influence in the vector capacity of C. felis. Manvell et al.⁴⁶ suggested that their inability to detect or recognize above cited genera within C. felis microbiomes might be due to the absence of these genera in the assessed flea samples or to misidentification as contaminants.

In this study, to the best of our knowledge, we detected Babesia sp. in all species assessed for the first time, especially with higher abundance in S. cuniculi collected from cats living on a rural farm. It is important to note that the species S. cuniculi is a parasite primarily associated with lagomorph hosts, specifically Oryctolagus cuniculi (rabbits). However, it has occasionally been reported to parasite secondary hosts, including certain mammals such as cats, dogs, pigs, and even humans²⁸. Therefore, it is crucial to determine whether domestic animals could serve as true reservoirs for this protozoan or if the flea infection might result from a prior bite on a rabbit, which has previously been demonstrated to function as a reservoir for babesiosis⁴⁷. On the other hand, the presence of B. microti has also been detected in C. felis collected from dogs and cats in Polonia⁴⁸ and even in the rodent flea genus Ctenophthalmus sp. in Slovakia⁴⁹. This fact does not demonstrate the vectorial role of fleas on Babesia sp., but it does provide evidence of the ability of fleas to maintain this pathogen as reservoirs. Consequently, the mere detection of B. microti in fleas does not unequivocally establish their role in transmission cycles. Nonetheless, the presence of zoonotic Babesia species within flea populations warrants careful consideration. Continuous surveillance of ectoparasites from both free-ranging and domestic animals is crucial for the identification of novel and emerging infectious agents and for elucidating their involvement in the transmission dynamics of diverse vector-borne pathogens⁴⁹.

Domestic cats and dogs may serve as crucial bridging hosts for fleas, facilitating the transmission of parasites among diverse animal populations, including wild animals, domestic animals, and humans. Their ranging behavior increases their exposure to various hosts, thus promoting the acquisition of fleas from distinct species^50,51. For instance, the phoresy of the rabbit flea, S. cuniculi, by cats and dogs could lead to the transfer of fleas to rabbits (including pets) living in the same habitat, considering that this species is an important vector of myxomatosis⁵².

The comparative analysis simultaneously addressed both the presence of specific microorganisms of sanitary interest and highlight the absence of others that have been previously reported. Thus, although we did not detect Anaplasma phagocytophilum or C. burnetii in this study, both species have been widely detected in fleas^53,54 using species-specific quantitative PCR (qPCR) technique. This fact could be due to 16 S resolution of the microbiota fails to provide specific identification in a considerable proportion of genera of medical importance to humans such as Staphylococcus, Streptococcus or Rickettsia and Bartonella detected in our flea samples. On the other hand, in our samples, previous quantitative PCR-based studies¹⁰ detected R. felis and R. asembonensis in C. felis and (A) erinacei, as well as (B) henselae and Bartonella clarridgeiae in (C) felis. In addition, in the same work, simultaneous detection of Rickettsia and Bartonella species within individual flea specimens was not observed. In spite of that, the current investigation revealed multiple instances of co-infection with both bacterial genera, albeit with inversely proportional relative abundances (Fig. 2B, Table S3). It is well established that vector microbiome constituents can exert both positive and negative effects on the arthropod host and other microbial populations⁴⁶. For example, studies on Oropsylla spp., a rodent-infesting flea genus, have demonstrated a negative correlation between Bartonella and Rickettsia spp⁵⁵. Conversely, other reports indicate that co-infection with multiple Rickettsia species is possible in ticks and fleas⁴⁶.

In previous studies, among the fifty-two flea specimens analyzed, sequences of Bartonella spp. (B. henselae and B. clarridgeae) were detected by qPCR in 50% (26/52) of the fleas¹⁰ while NGS identified Bartonella spp. in 100% of the infected specimens. Similarly, for Rickettsia species, qPCR detected Rickettsia spp. (R. felis and R. asembonensis) in 40.4% (21/52) of the samples, whereas NGS confirmed the presence of Rickettsia spp. in 100% of the infected fleas. This analysis highlights the higher sensitivity of NGS compared to qPCR for identifying genera but not species of Bartonella and Rickettsia. Consequently, while NGS offers enhanced sensitivity, it demonstrates lower specificity than qPCR for species-level identification.

Currently, as vaccines are not available to prevent Bartonella infection, flea control is the only successful measures to prevent these vector-borne infections in healthy animals⁵⁶ to decrease the dispersion of these bacteria among canine and feline populations, and to decrease the risk of zoonotic pathogen transmission to humans⁵⁷. Year-round protection of cats and dogs populations from flea infestations is recommended through the consistent application of flea control products, administered via collars, topical solutions, sprays, or oral formulations⁵⁸. Additionally, both human and animal subjects should minimize contact with stray cats and dogs. Within a One Health framework, prior research advocates for the development of vaccines designed to confer protection against B. henselae and B. vinsonii berkhoffii infections in companion animals, thereby mitigating reservoir potential and reducing zoonotic transmission risks¹⁹.

Another fact of relevance in this article was the presence of the Neisseriaceae family in all flea species studied. This could be justified since bacteria of the genus Neisseria are commensals of the mucous membranes of healthy dogs and cats and may also be isolated in bite wounds⁵⁹. According to our results, we could not explain why this bacterial group was more abundant in C. felis isolated from dogs compared to specimens of the same species collected from cats, or even compared to other flea species that also infest dogs. However, our findings align with previous studies where both the family Neisseriaceae and the genus Neisseria were identified in the microbiome of C. felis collected from dogs and cats in Greece⁴⁰, USA⁶⁰ and Australia⁴². This suggests that the presence of this bacterial family is well-established in the microbiome of C. felis across diverse geographic regions. This pattern might be associated with ecological or host-specific factors influencing the microbiome composition of fleas. Nonetheless, further research is needed to understand the mechanisms underlying these differences.

Complementarily, the study of the existence of unique and shared 16 S OTUs among the five flea species analyzed revealed interesting data with significant differences in the number of these OTUs depending on the flea species compared. Therefore, analysis of unique 16 S OTUs between male and female of A. erinacei collected from hedgehogs by Venn diagrams, demonstrated a slight increase in them within females. This fact has been cited by Dougas et al.⁴⁰, indicating that, although both males and females need to feed on blood, females use to be more aggressive in their consumption compared to males. Despite that, males of A. erinacei collected from dogs had many more unique 16 S OTUs than females, indicating the importance of the host in the flea microbiota. This fact was reported by Cohen et al.⁶¹, who supported the idea that the microbiome of hematophagous insects was strongly influenced by the host on which they fed. Indeed, when different species of flea isolated from dogs were compared each other (A. erinacei males versus C. felis males and A. erinacei males versus P. irritans males) the number of unique OTUs in A. erinacei decreased indicating that this species shared a greater number of them with both species. On the other hand, C. felis isolated from dogs showed the highest number of unique 16 S OTUs (twelve times more than P. irritans and five times more than A. erinacei). This could be explained because C. felis is a promiscuous flea species highly adapted to different hosts (dog and cat), whereas the other two ones could be accidentally found in dogs because of arbitrary contact with hedgehogs or other different hosts. According to Cohen et al.⁴¹, C. felis could feed on the blood of multiple hosts and would therefore be exposed to greater microbial diversity⁶². Finally, upon analyzing the unique 16 S OTUs present in the microbiota of congeneric species (C. felis versus C. canis), a reduction was observed, accompanied by an increase in the number of them shared between these species. This higher number of OTUs in C. felis was corroborated by calculating alpha diversity using Chao’s index based on the number of rare species in the sample, i.e., it looks that all taxa in a sample with counts of 2 or 1 estimating the taxa that really exist in the sample.

Lastly, analyzing the beta diversity results obtained in our study, we observed a certain correlation between the results and the general phylogeny of the Pulicidae family. Therefore, if we examine a UPGMA dendrogram based on Unweighted UniFrac distances, we can observe that the congeneric species C. felis and C. canis share the same clade, showing a more similar microbiome. Additionally, the genus Ctenocephalides appears to be more closely related to the genus Archaeopsylla, sharing a more similar microbiome compared to S. cuniculi and P. irritans. These results align with the phylogeny of these genera assessed by some authors within the Pulicidae family^2,5, where the genus Ctenocephalides was closely related to Archaeopsylla, while S. cuniculi and P. irritans appeared more phylogenetically distant.

Conclusions

This study provides initial insights into the microbiota of five flea species parasitizing pets (dogs and cats) and hedgehogs in Andalusia (Spain). Notably, C. felis exhibited the highest number of unique OTUs, potentially linked to its broad host range.

Across all flea species analyzed, the Phylum Pseudomonadota was dominant, followed by Bacillota, with the genus Wolbachia consistently showing the highest relative abundance. The detection of genera with known clinical relevance, such as Staphylococcus, Streptococcus and Pseudomonas, underscores the need for further investigation into their potential public health implications.

Additionally, the first detection of Babesia sp. in some flea species highlights a possible vectorial role that warrants further study. Finally, microbiota-based clustering revealed clear ecological separation among flea species, contributing to a better understanding of their epidemiological significance.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Author contributionsC.C. contributed to funding acquisition, project administration, methodology, resources, conceptualization, investigation, data analysis, data interpretation, writing – original draft, review & editing. A.M.G-S. contributed to methodology, data analysis, writing-review & editing, data curation. I.T. contributed to experimental analysis, methodology and validation. A.Z. contributed to data curation, data analysis, validation, writing- review & editing, methodology, supervision. All authors reviewed the manuscript.

Funding

This publication is part of the project PID2023-147663NB-I00, funded by MICIU/AEI/10.13039/501100011033 and by ERDF/EU.

Data availability

Sequence data that support the findings of this study have been deposited in the European Nucleotide Archive with the primary accession code GSE284667.

Declarations

Competing interests

The authors declare no competing interests.