Introduction

Bartonellae are facultative intracellular, fastidious, gram-negative bacteria that are transmitted to humans and other mammals by bloodsucking arthropod vectors such as fleas, mites, sand fleas and ticks1,2. The most important vectors for the maintenance and transmission routes of the Bartonella species are fleas, that feed on mammalian hosts such as cats, dogs, rodents, insectivores and rabbits3. There are currently 35 Bartonella species and three subspecies with standing in the Taxonomy Database of the National Center for Biotechnology Information (http://www.bacterio.net/Bartonella.html). However, the number of species in genus Bartonella is growing, as not all Bartonella species have been validated. Currently, 45 official and candidate Bartonella species have been detected in vertebrates, and at least fifteen of them have been related to human diseases4. During their evolution, Bartonella spp. have adapted to a variety of reservoir hosts and have become pathogenic when introduced into an incidental host5,6. With more than 20 Bartonella species associated with rodents, this group of mammals represents an important group of potential reservoirs for many Bartonella infections that have been reported worldwide7. Human pathogenic Bartonella species such as B. elizabethae, B. tribocorum, B. grahamii, B. rochalimae, B. vinsonii and B. washoensis have been isolated from various rodent species3,5,7. More than one Bartonella species can circulate in rodent communities, and multiple Bartonella genotypes in the same rodent host have been reported8,9.

The yellow-necked mouse (Apodemus flavicollis), wood mouse (Apodemus sylvaticus), striped field mouse (Apodemus agrarius), bank vole (Myodes glareolus), common vole (Microtis arvalis) field vole (Microtus agrestis) and root vole (Microtus oeconomus) are rodent species that play important roles in the maintenance and circulation of Bartonella infections in Europe5,8. Bartonella pathogens have been reported in several rodent populations in Poland, Germany, Denmark and Sweden (reviewed by Gutierrez et al.8). However, Bartonella epidemiology and host-pathogen associations in Europe, including in the Baltic region, are insufficiently characterized.

The objectives of this study were to investigate the prevalence of Bartonella species in different species of rodents collected from coastal and continental areas of Lithuania using real-time PCR targeting the ssrA gene, and to characterize the genetic diversity of Bartonella strains by sequence analysis of two housekeeping genes (rpoB, groEL) and the 16S-23S rRNA intergenic species region (ITS).

Results

Prevalence of Bartonella species in rodents

Based on real-time PCR analysis, a total of 318 of the rodent DNA samples (54.8% of the 580 samples analysed) were found positive for Bartonella spp., and out of these, 127 samples had Ct value below 30.

The highest prevalence of Bartonella spp. was found in A. flavicollis and M. agrestis (almost 80%); in M. minutus, M. glareolus and M. oeconomus the prevalence was in the range of 30–50%, while in other rodents, the prevalence was lower (Table 1).

Table 1 Prevalence of Bartonella spp. in different small rodent species from Lithuania, 2015–2016.
Full size table

Bartonella-infected A. flavicollis, A. agrarius and M. agrestis were found in all locations where these rodents were captured with the prevalence of infection ranging in different locations from 33.3 to 100%, from 7.7 to 71.4%, and from 50 to 100% respectively (Table 1). Bartonella-infected M. glareolus were found in nine of the 10 sampling locations: with the infection prevalence ranging from 0 to 70.4%. M. minutus and M. oeconomus were trapped only in the coastal area of the western part of the country, with the overall prevalence of infection estimated at 58.3% on the Curonian Spit and 50% in the Nemunas River Delta in M. minutus and 6.3% and 71.4% for the respective locations in M. oeconomus (Table 1). One Bartonella-infected M. musculus specimen was found in a mixed forest-meadow ecotone in the continental area (site 9) (Table 1).

Comparing A. flavicollis across the country, significantly higher overall Bartonella infection rates (85.7%; 144/168) were detected in those trapped on the Curonian Spit (odds ratio, 1.9; 95% confidence interval, 2.845–14.301; p < 0.000). Likewise, among all investigated areas, a significantly higher overall prevalence of Bartonella was observed on the Curonian Spit (71.1%, 182/256; odds ratio, 1.2; 95% confidence interval, 2.346–4.720; p < 0.000) with the highest prevalence of Bartonella detected in sampling site 2 (77.8%, 91/117; odds ratio, 1.3; 95% confidence interval, 2.265–5.825; p < 0.000).

Diversity of Bartonella species in rodents

Eighty-seven Bartonella-positive PCR products of partial rpoB, groEL genes and ITS region derived from 56 different rodent specimens of eight species were subjected to sequence analysis (Table 2). A total of 73 good-quality sequences of rpoB (n = 43), groEL (n = 9) genes, and ITS region (n = 21) were obtained and analyzed.

Table 2 Distribution of Bartonella species in the rodent communities in Lithuania.
Full size table

Analysis of 43 partial rpoB gene sequences showed that the Bartonella species circulating among the investigated rodents were heterogenic and belonged to B. grahamii, B. taylorii, B. tribocorum, B. coopersplainsensis and B. rochalimae genogroup (Fig. 1). Bartonella strains derived from six rodent species A. flavicollis (n = 2), A. agrarius (n = 3), harvest mouse Micromys minutus (n = 3), common house mouse Mus musculus (n = 1), M. glareolus (n = 5) and M. arvalis (n = 1) were 95–100% similar to B. grahamii sequences deposited in the GenBank (CP001562, JN810824, JN647927) (Fig. 1). Three B. grahamii genotypes (differing at 5 nucleotides positions) were identified (Fig. 1). Bartonella rpoB sequences derived from four species of rodents A. flavicollis (n = 7), M. glareolus (n = 11), M. agrestis (n = 4) and M. oeconomus (n = 1) were 97–100% identical (differing at 48 nucleotide positions) to B. taylorii sequences obtained in the GenBank (AF165995; JX984664) (Fig. 1). Twelve different B. taylorii genotypes among Lithuanian isolates were detected (Fig. 1): seven genotypes in M. glareolus (differing at 29 nucleotide positions); four in A. flavicollis (differing at 28 nucleotide positions) and two in M. agrestis (differing at 10 nucleotide positions). Two Bartonella rpoB sequences derived from A. agrarius (sites 6, 8) shared 99–100% identity to the B. tribocorum detected in A. agrarius from South Korea (GenBank: JN81812) (Table 2; Fig. 1). Bartonella sequences derived from M. glareolus (site 2), M. oeconomus (site 7) and A. agrarius (site 6) were 98–99% identical to B. rochalimae (GenBank: DQ676489), Candidatus Bartonella rudakovii (GenBank: EF682088) and B. coopersplainsensis (detected in rats from Thailand, GenBank: MF105907) respectively (Fig. 1).

Figure 1
figure 1

Maximum-likelihood phylogenetic tree for the partial rpoB gene of Bartonella spp. The phylogenetic tree was created using the Tamura-Nei model and bootstrap analysis of 1000 replicates. Identification source (rodent host) is given after the code of sample. Sequences MH547313, MH547314 and MH547315 are representative of three other samples sequenced in this study (all derived from A. flavicollis); Sequence MH547320 is representative of one other sample derived from M. glareolus; Sequence MH547321 is representative of two other samples (all derived from M. agrestis); Sequence MH547327 is representative of other sample from M. glareolus; Sequence MH547329 is representative of other sample from M. minutus; Sequence MH547328 are representative of two other samples sequenced in this study (all derived from M. glareolus). Samples sequenced in the present study are marked. Abbreviations: A. fla – Apodemus flavicollis; M. min – Micromys minutus; M. gla – Myodes glareolus; A. agr – Apodemus agrarius; M. mus – Mus musculus; M. oec. – Microtus oeconomus; M. agr – Microtus agrestis.

Full size image

The sequence analysis of the partial ITS region of 21 samples revealed the presence of Bartonella strains from the B. grahamii, B. taylorii, B. tribocorum, B. coopersplainsensis and B. doshiae genogroup (Fig. 2). Ten Bartonella strains derived from six rodent species A. flavicollis (n = 2), A. agrarius (n = 2), M. minutus (n = 3), M. arvalis (n = 1), M. oeconomus (n = 1) and M. glareolus (n = 1) were 97–100% identical (differing at 17 nucleotide positions) to B. grahamii sequences deposited in the GenBank (AJ269785; AJ269789). Four B. grahamii genotypes were identified (Fig. 2). Eight Bartonella ITS region sequences derived from four rodent species A. flavicollis (n = 4), M. oeconomus (n = 2), M. agrestis (n = 1) and M. glareolus (n = 1) shared 97–99% similarity with B. taylorii sequences deposited in the GenBank (JN10860; AJ269784). Obtained sequences were heterogenic (differing at 33 nucleotide positions), and eight B. taylorii genotypes were detected (Fig. 2). The Bartonella ITS region sequence derived from A. agrarius (site 6) was 99% identical (differing at nine nucleotide positions) to B. coopersplainsensis detected in rats from the Australia (GenBank: EU111770). Two sequences, one derived from A. agrarius and another from M. agrestis (site 6) were 98–99% identical to B. tribocorum (GenBank: JN810856) and B. doshiae (GenBank: AF442954) respectively (Fig. 2).

Figure 2
figure 2

Maximum-likelihood phylogenetic tree for the partial ITS region of Bartonella spp. The phylogenetic tree was created using the Tamura-Nei model and bootstrap analysis of 1000 replicates. Identification source (rodent host) is given after the code of sample. Sequence MH547346 is representative of three other samples sequenced in this study (all derived from M. minutus); Sequence MH547348 is representative of one other sample derived from A. flavicollis; Samples sequenced in the present study are marked. Abbreviations: A. fla – Apodemus flavicollis; M. min – Micromys minutus; M. gla – Myodes glareolus; A. agr – Apodemus agrarius; M. mus – Mus musculus; M. oec. – Microtus oeconomus; M. agr – Microtus agrestis.

Full size image

The groEL gene sequences showed 97–100% similarity to B. grahamii sequences deposited in the GenBank (AB426677; AF014833; GU559869) (Fig. 3). Phylogenetic analysis demonstrated the presence of three B. grahamii genotypes in M. glareolus, (n = 3), M. minutus (n = 2), A. agrarius (n = 1), M. arvalis (n = 1) and M. agrestis (n = 1). One Bartonella groEL sequence derived from M. agrestis (site 6) was 100% identical to B. doshiae (GenBank: AF014832) (Fig. 3).

Figure 3
figure 3

Maximum-likelihood phylogenetic tree for the partial groEL gene of Bartonella spp. The phylogenetic tree was created using the Tamura-Nei model and bootstrap analysis of 1000 replicates. Identification source (rodent host) is given after the code of sample. Samples sequenced in the present study are marked. Sequence MH547353 is representative of other sample from M. minutus. Abbreviations: M. min – Micromys minutus; M. gla – Myodes glareolus; A. agr – Apodemus agrarius; M. arv – Microtus arvalis; M. agr – Microtus agrestis.

Full size image

Discussion

The present study represents the prevalence and molecular characterization of the Bartonella strains circulating in rodent communities in the continental and coastal areas of Lithuania. Phylogenetic analysis based on two housekeeping genes and ITS region demonstrated that small rodents in Lithuania harbor multiple Bartonella species belonging to six genogroups, specifically B. grahamii, B. taylorii, B. tribocorum, B. coopersplainsensis, B. doshiae and B. rochalimae (Table 2).

Worldwide, the prevalence of Bartonella spp. is ranging from 25 to 80%. The high rate of Bartonella infection in rodent communities suggests a reciprocal adaptation between the bacteria and their reservoirs8. We found the overall prevalence of Bartonella species in different rodent species in Lithuania ranging from 8.3 to 80% with the highest Bartonella-infection rates in M. agrestis (80%) and A. flavicollis (79.6%) (Table 1). Our findings are similar to those reported in A. flavicollis from the eastern Germany (84.4%10) but higher than those from Slovakia (63.0%; 244/38711), Slovenia (62.7%; 27/4312), Denmark (53.3%; 8/1513), Poland (42.2; 68/16114). In M. agrestis, lower infection rates were found in Sweden (33.3%; 1/315), Denmark (33.3%; 5/1512) and France (11%; 1/95). In Lithuania, a high Bartonella prevalence was also detected in M. minutus (57.5%), M. oeconomus (53.4%) and M. glareolus (42.4%). In other European countries, reported prevalence of Bartonella infection in M. glareolus include 69.0% (160/232) in Slovakia, 56.4% (252/447) in France, 52.8% (21/36) in Germany and 15.0% (9/60) in Sweden10,11,15,16. Bartonella-infected M. oeconomus have been found in Poland with an 11.1% (14/128) prevalence of infection14, a figure that s five times lower than that obtained in our study. The lowest overall prevalences of Bartonella spp. in the present study were found in A. agrarius (15.8%) and M. musculus (8.3%). In comparison, studies conducted in Slovenia and Slovakia detected 26.6% (8/30) and 9% (31/344) prevalence of Bartonella spp. in A. agrarius11,12, while Holmberg et al.15 reported a 5.6% (1/18) prevalence of infection in M. musculus from Sweden.

The observed differences in the prevalence of Bartonella spp. might be related to habitats and to the abundance of the rodent species16. In this study, extremely high Bartonella-infection rates (56.4–77.8%) were detected in rodent communities on the Curonian Spit (Table 1), a region characterized by high diversity of natural habitats this resulting in diversity in animal and plant species. The dominant rodent species in all habitats on the Curonian Spit was A. flavicollis, this rodent species characterized by high Bartonella-infection rates (B. grahamii and B. taylorii most frequent) (Tables 1, 2).

In Europe, B. grahamii and B. taylorii have been detected in A. flavicollis, A. agrarius, A. sylvaticus, herb field mouse (Apodemus uralensis), steppe field mouse (Apodemus witherbyi), M. minutus, M. glareolus, M. arvalis, M. agrestis and M. musculus; while B. rochalimae has been detected in A. flavicollis, M. glareolus and M. arvalis (reviewed by Špitalska et al.11), and B. doshiae, in A. flavicollis, A. agrarius, M. agrestis and M. glareolus (reviewed by Buffet et al.17).

In this study, the 52.9% of sequences derived from small rodents were ascribable to B. taylorii. In general, B. taylorii strains demonstrate high diversity and are frequently found in mice, as well as in Myodes and Microtus voles, inhabiting boreal forests of the Eurasian continent (reviewed by Buffet et al.5). In line with this, the present study showed high diversity of B. taylorii strains in Lithuanian rodents. Phylogenetic analysis of rpoB gene revealed the presence of twelve B. taylorii genotypes associated with A. flavicollis mice (four out of twelve) and three species of voles M. glareolus, M. agrestis and M. oeconomus (eight out of twelve) (Fig. 1). The high variability of B. taylorii strains in rodents could be explained by a potential accelerated evolution of the Bartonella genus in small rodents as a result of frequent recombination events, horizontal gene acquisitions, and accumulation of mutations (reviewed by Gutierrez et al.8).

In this study, human pathogenic B. grahamii strains were detected in 39.2% (20 out of 51) of rodent specimens and showed lower sequence diversity in comparison to B. taylorii. B. grahamii was detected in four species of mice (A. flavicollis, A. agrarius, M. musculus, M. minutus), and four species of voles (M. glareolus, M. oeconomus, M. arvalis and M. agrestis) (Table 2). Similar results have been observed in some European countries, while higher polymorphism of B. grahamii has been observed in Asia17. The low diversity of B. grahamii has been explained by the spread of these bacteria from Asia to Europe by the introduction of its hosts and/ or a severe bottleneck relatively with too little time having elapsed for polymorphisms to reaccumulate17.

B. rochalimae is typically associated with carnivores (cats, dogs, foxes and raccoons), and Eremeeva et al.18 reported clinical case of bacteremia, fever, and splenomegaly in a patient who traveled to Peru. In Europe, Bartonella spp. related to B. rochalimae group have been detected in A. agrarius, A. flavicollis and M. arvalis from Slovakia11,19. We report Bartonella spp. from B. rochalimae group in M. oeconomus for the first time (Fig. 1).

The main vectors for the maintenance and transmission of B. grahamii, B. taylorii and B. rochalimae among populations of small mammals are fleas5, these also having been identified as a risk factor for the transmission of Bartonella pathogens to humans. In Lithuania, strains identical or similar to B. grahamii, B. taylorii and B. rochalimae were detected in five flea species with an overall prevalence of 29.1%. B. grahamii was also detected in Ixodes ricinus ticks20.

This study is the first report of B. doshiae infection in M. agrestis from Lithuania. In Europe, this species has been found with very low prevalence and genetic diversity in Apodemus spp., M. glareolus and M. agrestis5.

Previous studies strongly supported the association of B. tribocorum with rats of the genus Rattus5. However, Ko et al.21 detected B. tribocorum in A. agrarius in South Korea. In the current study, the human pathogenic B. tribocorum was identified in A. agrarius in coastal (site 6) and continental (site 8) areas with 100% similarity to the B. tribocorum isolate from South Korea (Fig. 1). Kraljik et al.19 reported 94.8% similarity of Bartonella isolates derived from A. agrarius to the B. elizabethae/B. tribocorum clade. Our study is the first detection of B. coopersplainsensis in A. agrarius in the coastal area of the Nemunas River Delta. This result is unexpected, as B. coopersplainsensis is distributed in Asia and Australia, and the species spread in Europe is associated with rats. The detection of B. tribocorum and B. coopersplainsensis in A. agrarius could be explained through the accidental contact between A. agrarius and rats in agricultural habitats.

The highest Bartonella species diversity was detected in the specific habitats of flooded meadow and flooded forest within the Nemunas River Delta: five rodent species trapped in this area harbored six different Bartonella species (Table 2). The high Bartonella species diversity detected in the Nemunas River Delta could be explained by another particular features of this area, namely that it is the most important stopover area for migratory birds in Lithuania, with about 200 species also breeding22. Migratory birds are important carriers for a variety of ectoparasites with a great potential to assist in their spread. An alternative explanation for the high Bartonella species diversity in the Nemunas Delta is that the area is also characterized by a high intensity of local shipping, thereby also providing a suitable environment for rats and consequently for the transmission of rat-related Bartonella spp.

In conclusion, the present study demonstrates that small rodents from both coastal and continental areas of Lithuania are frequently infected with Bartonella spp. Our findings provide evidence of a high diversity of Bartonella species and the presence of strains identical or closely related to the humans pathogenic B. grahamii, B. tribocorum, and B. rochalimae. This study provides the first detection of B. coopersplainsensis, B. doshiae and B. tribocorum in Lithuania. To our knowledge, it is also the first report of Bartonella spp. in the agriculture-related rodent species A. agrarius and indoor rodent species such as M. musculus in Lithuania.

Material and Methods

Ethical statement

Rodents sampling was conducted with permission from the Environmental Protection Agency (EPA) and approved by the Ministry of Environment of the Republic of Lithuania, licenses No. 22 (2015-04-10) and No. 12 (2016-03-30) in accordance with Lithuanian (the Republic of Lithuania Law on Welfare and Protection of Animals No. XI-2271) and European legislation (Directive 2010/63/EU) on the protection of animals. Live traps were checked at least twice a day. Captured animals were anesthetized with CO2 exposure and killed humanely according to the Lithuanian Animal Protection Act.

Rodents trapping

A total of 580 rodents representing eight species – A. flavicollis, A. agrarius, M. minutus, M. musculus, M. glareolus, M. oeconomus, M. agrestis and M. arvalis were trapped with live-traps or snap traps in 12 locations situated in western (coastal area; sites 1–7) and eastern (continental area; sites 8–12) parts of Lithuania during 2015–2016 (Fig. 4; Table 1).

Figure 4
figure 4

Small rodent trapping sites on the Curonian Spit (sites 1–5), the Nemunas River Delta (sites 6–7) and continental (sites 8–12) areas of Lithuania, 2015–2016.

Full size image

In the western part of the country rodents were trapped on the Curonian Spit (sites 1–5, close to the Baltic Sea) and in the Nemunas River Delta (sites 6–7). The Curonian Spit is a narrow sand peninsula (2 km wide and 98 km long, half of which is in Lithuania) in the southeastern coast of the Baltic Sea separating the Curonian Lagoon from the Baltic Sea. Rodent sampling on the Curonian Spit was conducted in coastal meadows and ecotone between meadow and mixed forest. In the Nemunas River Delta, trapping was conducted in two specific habitats: in a flood meadow in Rusnė (site 6; meadows overgrown by reeds with the main vegetation being Poaceae and Cyperaceae plants) and in Žalgiriai forest (site 7; the main habitat spring-flooded block alder stands) (Table 1). Eastern Lithuania represented continental habitats (sites 8–12, Fig. 4). Site 8 was located in deciduous forest on a peninsula in the northern part of Lake Lukstas; site 9 in mixed forest-meadow ecotone; sites 10–11 in mixed forest and site 12 in deciduous forest on an island in Elektrėnai Reservoir (Table 1).

The dominant rodent species trapped in the western part of the country was A. flavicollis (n = 173), while in the eastern part of Lithuania, it was M. glareolus (n = 99). The indoor rodent species M. musculus was captured only in one sampling location (site 9) in the mixed forest-meadow ecotone in the continental area (Table 1). The highest rodent species diversity (six rodent species) was detected in the flooded forest (site 7) in the coastal area of the Nemunas River Delta. In contrast, only one rodent species, M. glareolus was detected in the deciduous forest on the island in Elektrėnai Reservoir (continental area; site 12) (Table 1).

Spleens from rodents were collected and stored in 70% ethanol for DNA extraction.

Molecular analyses

DNA was extracted from each rodent spleen using a Genomic DNA Purification Kit, (Thermo Fisher Scientific, Lithuania), according to the manufacturer’s instructions. Screening for the presence of Bartonella DNA in rodents was conducted through the amplification of the 124-bp fragments of ssrA gene by TaqMan real-time PCR with ssrA-F1 and ssrA-R1 primers and ssrA-P1 probe (Table 3). The ssrA-F1 primer was designed in this study, while ssrA-R1 primer and ssrA-P1 hydrolyzation probe were obtained after modification of the reverse primer and hydrolyzation probe published in Diaz et al.23 on the basis of NCBI (National Center for Biotechnology Information) GenBank database sequences, NCBI BLAST® blastn suite applet for alignment and FastPCR online (http://primerdigital.com/tools/pcr.html) java applet for primers test. The qPCR reactions were carried out with SensiMix™ II Probe Kit (Bioline Reagents Ltd, UK) in a total volume of 15 μL using 1 μL of the isolated DNA sample, primers at a concentration of 200 nM each and probe at a concentration of 100 nM. Real-time PCR assays were performed employing a real-time thermocycler Rotor-Gene Q 5plex model with software version 1.7 (Qiagen GmbH, Germany) under the following conditions: denaturation at 95 °C for 10 min (1 cycle); followed by 50 cycles of denaturation at 95 °C for 20 s, annealing at 50 °C for 1 min and extension at 72 °C for 10 s. In each real-time PCR run, positive (DNA of Bartonella-infected rodents, confirmed by sequencing) and negative controls (which consisted of sterile, double-distilled water added to the PCR mix rather than DNA) were used. The results that satisfy amplification cutoffs below 40 Ct (cycle threshold) when threshold was 0.10101 indicated positive samples. Samples with Ct value below 30 were chosen for further molecular characterization of bacteria strains. For the amplifications of the 795 bp fragment of the RNA polymerase β-subunit (rpoB) gene and 336 bp of 60 kDa heat-shock protein (groEL), conventional PCR was used and for the amplification of 0.9–1.6 kb fragment of the 16S-23S rRNA gene intergenic species region (ITS), nested PCR was used (Table 3). A selected number of Bartonella-positive samples for two housekeeping genes (rpoB, groEL) and the ITS region (from different rodent species and different regions) were chosen for sequencing (Macrogen Europe, Netherlands). PCR products were extracted from agarose gel and purified using a GenJET PCR purification kit (Thermo Fisher Scientific, Lithuania). The obtained partial rpoB, groEL genes and ITS region sequences were analyzed using the Mega software package, version 6.05, and compared with the sequence data available from the NCBI GenBank database using the NCBI BLAST® blastn suite applet. Phylogenetic trees were constructed using the maximum-likelihood (ML) method with Tamura-Nei model.

Table 3 Oligonucleotide primers used for conventional, nested-PCR and real-time PCR of ssrA, rpoB, groEL genes and ITS region.
Full size table

Partial rpoB, groEL genes and ITS region sequences for representative samples were submitted to the GenBank under the accession numbers MH547308 to MH547335 and MH687373 for the rpoB gene, MH547336 to MH547350 and MH687377-MH687379 for the ITS region, and MH547351 to MH547354 and MH687374-MH687376 for the groEL gene.

Statistical analysis

Differences in the prevalence of Bartonella spp. infection between different species of small rodents, sampling locations, regions were assessed by Fisher’s exact test, supplemented with the Mantel-Haenszel common odds ratio estimate and 95% confidence intervals using SPSS software version 22 (IBM SPSS, Chicago, IL, USA). p < 0.05 was considered significant.