Remember me
Bartonella are gram-negative, facultative intracellular bacteria. Infection by Bartonella manifests as different clinical syndromes collectively known as bartonellosis. The well-known diseases caused by these bacteria are cat scratch disease (Bartonella henselae), trench fever (Bartonella quintana) and Carrion’s disease (Bartonella bacilliformis). Excluding B. bacilliformis, which is evolutionarily more distinct than the other species, Bartonella infections result in self-limiting disease that is often undiagnosed and untreated. However, individuals with compromised immune systems or other undefined conditions may experience clinical manifestations, which can become life-threatening and need to be treated with effective antibiotics. To date, there is no standard treatment course for these infections, and many doctors prescribe antibiotics based on limited case studies. Within the host, Bartonella can grow extracellularly, intracellularly, and in biofilms. To determine an effective antibiotic strategy, it is important to understand Bartonella susceptibility in each of these growth conditions. We hypothesized that combination antibiotic treatments are required to effectively eliminate Bartonella henselae growth, particularly in biofilm and intracellular environments. Our previous work has shown that B. henselae treatment with single antibiotics in different media, as well as in DH82 canine macrophages, was ineffective in eliminating bacteria. We expanded this work with different antibiotics supported by case reports, as well as double and triple combinations. The following antibiotics were tested: doxycycline, gentamicin, azithromycin, azlocillin, rifampin, and clarithromycin. We found that while monotherapy may inhibit growth extracellularly, it is ineffective when used against intracellular bacteria or pre-existing biofilms. Gentamicin in combination with either rifampin or azlocillin significantly reduced bacterial growth in multiple microenvironments. The effectiveness of combination therapy supports the notion that Bartonella species utilize host cells and biofilms as antibiotic evasion strategies.
1 IntroductionBacteria within the genus Bartonella are gram-negative coccobacilli in the class proteobacteria, with over 35 named species, of which at least seven have been described in human cases (Breitschwerdt, 2017; Tahmasebi Ashtiani et al., 2024). Bartonella spp. are facultative intracellular bacteria, with target cells including, but not limited to, endothelial cells, red blood cells, and macrophages. Other cell types have been infected in vitro, suggesting that there are more niches in the host for this pathogen that remain to be elucidated. Additionally, Bartonella spp. have been identified as biofilms in human hosts, namely around the heart in cases of blood culture-negative endocarditis (BCNE) (Meidrops et al., 2023; Xi et al., 2024). The three most prominent species which cause human disease are B. henselae (causative agent of cat-scratch disease), B. quintana (causative agent of trench fever), and B. bacilliformis (causative agent of Carrion’s disease). However, numerous case reports demonstrate that a variety of Bartonella species can also cause clinical disease in humans (Delaney et al., 2024).
Bartonellosis, the term used to collectively describe Bartonella spp. infections, encompasses a range of clinical signs and symptoms that vary from mild manifestations such as fever and malaise to more serious complications like endocarditis, bacillary angiomatosis, and various neurological manifestations. Bartonellosis is extremely difficult to diagnose clinically due to the broad disease presentation as well as the need for highly sensitive diagnostic tools. Current methods of detecting Bartonella spp. include serological testing, blood cultures, and PCR. Clinical diagnosis is weighted heavily with exposure to felines (CDC, 2023) or known risk factors such as infestation with body lice. In many patients, Bartonella spp. infection is mild and may not require antibiotic treatment. However, individuals with high risk factors for severe disease, such as prior tissue damage or a compromised immune system, can progress to more serious forms of the disease which require antibiotic therapy.
1.1 Treatment for infectionCurrent guidelines for the treatment of bartonellosis are based on disease presentations (Angelakis and Raoult, 2014) and patient-specific contraindications. For example, patients with prior renal impairments may be administered an alternative to gentamicin, which is known to cause kidney damage. Physicians and veterinarians thus rely heavily on case reports to determine the best course of treatment for their patients. Studies demonstrating antibiotic efficacy are limited to in vitro experiments, small clinical trials, or limited meta-analyses (Prutsky et al., 2013). Treatment failure has been documented in as high as 39% of patients (Pizzuti et al., 2024), with the majority experiencing adverse events, acute toxicity to the drugs, or the need to escalate therapy. A list of the antibiotics used in these experiments and their characteristics are summarized in Table 1.
AntibioticClassMechanism of actionBactericidal/bacteriostaticDoxycycline (Parmar and Patel, 2025)TetracyclineBinds 30S ribosomal subunit, preventing protein synthesis and growthBacteriostaticGentamicin (Chaves and Tadi, 2023)AminoglycosideInhibits ribosome functionality, disrupting protein synthesisBactericidalAzithromycin (Patel and Hashmi, 2025)MacrolideInhibits 50S ribosomal subunit, preventing protein synthesis and growthBacteriostaticAzlocillin (Sanders, 1983)β-LactamPrevent synthesis of bacterial cell wall by binding to PBP during peptidoglycan cross-linkingBactericidalRifampin (Patel and Alan, 2023)AnsamycinInhibits DNA-dependent RNA polymerase, blocking RNA synthesisBactericidalClarithromycin (Patel and Hashmi, 2025)MacrolideInhibits 50S ribosomal subunit, preventing protein synthesis and growthBacteriostaticAntibiotics that were used for in vitro efficacy testing.
Treatment for bartonellosis is highly dependent on disease presentation and Bartonella species (Mikes et al., 2025). Treatment options can vary from no antibiotics in mild cat scratch disease to combination therapy delivered via intravenous infusions in complicated presentations such as nervous system involvement. One of the main antibiotics used for multiple presentations of bartonellosis is doxycycline (Prutsky et al., 2013), a semi-synthetic derivative in the tetracycline class with a broad range of antimicrobial activity. This antibiotic is often prescribed for other skin infections as well as in Lyme disease, caused by Borrelia burgdorferi, of which Bartonella spp. can be present as a co-infection (Parmar and Patel, 2025). Like other tetracyclines, doxycycline is a bacteriostatic antibiotic and confers high resistant rates in many species of bacteria (Rissardo and Caprara, 2019) although the exact mechanisms of Bartonella spp. resistance are not well characterized. It is thought that Bartonella spp. utilize their ability to enter host cells as a method of intrinsic antibiotic resistance rather than active resistance conferred by plasmids (Biswas and Rolain, 2010; Xi et al., 2024).
Other antibiotics are usually prescribed as part of a combination therapy cocktail during complicated bartonellosis, which can include but are not limited to gentamicin (Mikes et al., 2025), azithromycin (Ives et al., 2000; Margileth, 1992), clarithromycin (Ives et al., 2000), and rifampin (Margileth, 1992). Rifampin was found to have some activity against B. henselae biofilms (Zheng et al., 2020) after 6 days of treatment but was ultimately more effective when used in combination with other antibiotics. Rifampin was also shown to have activity against intra-erythrocytic B. quintana, although this activity was much less effective compared to the aminoglycoside gentamycin (Rolain et al., 2003) Azlocillin, also included in this study, has recently been shown to have antimicrobial activity against both Borrelia burgdorferi (Pothineni et al., 2020) and Bartonella henselae (Gadila and Embers, 2021), making it an attractive potential candidate for treatment in these co-infected patients. However, it is important to consider the relative toxicities to antibiotics when longer courses of antimicrobial cocktails are prescribed, which limits doctors’ options for extended therapy during chronic Bartonella spp. infections. A full table with each antibiotic and its concentrations, typical human doses, and the maximum serum concentrations (if reported) are listed in Table 2.
AntibioticConcentrations* tested for MIC/Singular MBCAntibiotic concentrations used in experiments.
*Concentrations were tested in a 1:1 ratio.
In addition to antimicrobials, additional treatments may be prescribed to patients with immune-mediated pathologies. The most common of these are steroids; however, this is controversial as they are also used to suppress the immune system, which in turn increases the opportunity for further Bartonella colonization in the host. Oral steroids with a tapering dose are often given to patients presenting with neuroretinitis. Another promising therapy, particularly with conditions related to immunosuppression or immune-mediated pathologies, is intravenous immunoglobulin infusions (IVIG) (Dietmann et al., 2022; Rissardo and Caprara, 2019). This treatment has been documented in several case reports of children infected with Bartonella that have developed neurological complications secondary to bartonellosis.
1.2 ImportanceThe goal of this study was to systematically measure the efficacy of candidate antibiotics against B. henselae when grown extracellularly, intracellularly, and as a biofilm. While many studies examine antibiotic efficacy through minimum inhibitory concentrations (MICs) and minimum bactericidal concentrations (MBCs), we used both measurements as well as bacterial killing in different growth conditions that better model the different microenvironments this pathogen can occupy in the human host. We hypothesized that combinations of antibiotics are required to fully eliminate Bartonella growth in these different growth conditions.
2 Materials and methods2.1 Bartonella culturingBartonella henselae strain San Antonio 2 was used in these experiments. Bacterial stocks were grown in both Schneider’s Drosophila Medium (Thermofisher, Cat No. 21720024) and Grace’s Insect Medium (Gibco, 11605–094) and frozen in media supplemented with 50% glycerol. For use in assays, a frozen stock of B. henselae was streaked onto blood agar plates (Remel, R01198) and incubated at 37°C with 5% CO2 for 7–10 days. After incubation, isolated colonies were used to inoculate either Grace’s media or Schneider’s media. After 5–8 days in growth media, cultures were harvested. The culture was diluted to an OD600 of 0.1 using the SmartSpec 300 (Bio-Rad, 170–2501) before use in assays, for an approximate count of 3 × 107 colony-forming units (CFU)/mL.
2.2 Preparation of antibioticsThe following antibiotics were used in these assays: doxycycline (Sigma, D9891-1G) gentamicin (FisherSci, BP918-1) azithromycin (TCI, 117772-70-0), azlocillin (provided by Flightpath Biosciences, Inc.), rifampin (Acros, 45562), and clarithromycin (Acros, 45522). Antibiotic stocks were created by dissolving singular antibiotics in 100% molecular-grade ethanol (rifampin) at 1 mg/ml or sterile cell-culture grade water (all others). Stock solutions were stored at −20°C until use. All antibiotics were diluted fresh in the respective media on the day of their use in the assays.
2.3 Extracellular antibiotic testingOn the day of the assay, Bartonella species were diluted in their respective medias as previously described to an OD of 0.1. In each well of a 96 well plate, 50 μL of OD 0.1 bacteria was added to 50 μL of media with the respective dilution of antibiotics for a total approximate concentration of 1.5 × 107 CFU/mL. Table 2 shows a full list of dilutions for each antibiotic. Antibiotic concentrations used in the experiments are listed on the center column and typical adult human doses for mild or systemic infections are given on the right column. The duration of antibiotic administration in humans is highly dependent on infection type and severity. For complicated Bartonella infections, multiple antibiotics can be prescribed up to months at a time (Biswas and Rolain, 2010; Prutsky et al., 2013; Raoult et al., 2003). Wells containing only bacteria or only media served as the positive and negative controls, respectively. Plates were incubated at 37°C in 5% CO2 for 4 days. After incubation, the OD600 of each plate was read using the SmartSpec 300 (Bio-Rad, 170–2501). The MIC was determined by comparing the control wells to experimental conditions using a cut-off with the negative control for growth. This assay was repeated three times with all six antibiotics.
Contents of the wells from the MIC assays that showed no bacterial growth were used for the MBC assay. To collect all remaining bacteria possible, these wells were vigorously washed and scraped with a pipette tip, and the contents of all replicates were collected into 1.5 mL tubes. Tubes were centrifuged at 6,000 × g for 10 min. The supernatant was removed, and the pellet was resuspended with 500 μL of 1X PBS. This was repeated for a total of 2 washes. After the second wash, the pellet was then resuspended with 100 μL of PBS and plated onto blood agar plates. The pellets from the control tubes were resuspended with 500 μL of PBS and serially diluted to determine quantification of bacterial load. The plates were examined for bacterial growth after incubating for 14 days at 37°C in 5% CO2. This assay was repeated 3 times with each antibiotic. The methods for these assays are highlighted in Figure 1.
Experimental schematic depicting the minimum inhibitory/bactericidal concentration (MIC/MBC) assay for extracellular Bartonella. B. henselae grown in either Grace’s or Schneider’s Media was diluted to an OD600 of 0.1 for the extracellular MIC/MBC assays. Antibiotics diluted in the same media were added in a 1:1 ratio with bacterial slurry. After 96 h of incubation, inhibition of growth (MIC) was determined by comparing absorbence values between respective antibiotic concentrations with control wells containing only media and wells with only bacteria. Wells containing antibiotics that successfully inhibited bacterial growth were plated on a single blood agar plate to determine bactericidal activity (MBC) and were reported as either positive or negative for growth. Combination therapies were tested for bactericidal activity only, and bacteria remaining after 96 h of antibiotic exposure were plated in triplicate for quantification. For all assays, the wells for each respective condition were combined and antibiotics were removed from solution by washing with PBS three times before plating. Only the control wells containing B. henselae without antibiotics were serial diluted for enumeration. Colonies were counted after 14 days of incubation at 37°C in 5% CO2. Significant reduction of bacterial counts (CFU) from combination therapy was determined as the difference between the antibiotic-free control and the antibiotic-treated bacterial count at each specified concentration.
For combination MBC assays, the same concentrations of antibiotics were diluted in either Schneider’s or Grace’s media, and the assays were set up as previously described for determining MIC and singular MBC. The pellets were resuspended in 500 μL instead of 100μL, and 100μL of this was plated onto each blood agar plate in triplicate for quantification. Plates were counted after 14 days of incubation at 37°C in 5% CO2.
2.4 DH82 cell culturing and antibiotic testing of intracellular B. henselaeDH82 cells (ATCC, CRL-10389) were thawed from frozen stocks and centrifuged at 125 × g for 6 min. Cells were washed with Eagle’s Minimum Essential Medium (EMEM) (ATCC, 30-2003) supplemented with 15% FBS (EMEM-15). Cells were then resuspended and allowed to grow for 3–5 days in T-75 flasks before passaging. DH82 cells used for in vitro assay testing were between passage 3 and 12.
For assays (see Figure 2), 24 h prior to infection, cells were seeded at a concentration of 1 × 105 cells/ml in a 24-well cell culture plate. On the day of infection, the OD600 of the Bartonella liquid culture was measured. The amount of culture needed to inoculate each well with an OD600 of 0.05 was removed and centrifuged at 6,000 × g for 10 min. The culture was then washed in 1X PBS and centrifuged again at 6,000 × g for 10 min. The pellet was resuspended in the appropriate amount of EMEM-15 for an approximate multiplicity of infection (MOI) of 1:50. The media was removed from the DH82 cells and replaced with 500 μL of EMEM-15 with an OD of 0.05 of B. henselae. These plates were centrifuged 1200x g for 8 min and then allowed to incubate at 37°C in 5% CO2 for 2 h.
Experimental MBC assay for intracellular B. henselae. DH82 cells were seeded in 24-well plates at 5 × 104 cells/ml. After 24 h, media was replaced with EMEM-15 containing B. henselae at a target MOI of 50:1 and plates were centrifuged to enhance bacteria: cell contact. Cells and bacteria were incubated for 2 h, after which remaining extracellular bacteria were removed with three PBS washes. Approximately 36–48 h later, the media in the wells was removed and replaced with media containing antibiotics at multiple concentrations. Samples were treated for 96 h, followed by washing with PBSbefore osmotic lysis using ice-cold water. The lysed contents of each well were combined and plated onto blood agar for bacterial detection. For monotherapy, the bacteria were plated on a single blood agar plate per treatment and designated as positive or negative based on growth. The bacteria were enumerated after combination antibiotic treatment by plating the undiluted DH82 lysate from each well/treatment in triplicate on blood agar plates. Wells containing only bacteria and no bacteria were used as positive and negative controls, respectively. Control wells with untreated infected cells were serially diluted before plating to obtain accurate counts. All incubations during the assay were at 37°C in 5% CO2. For combination treatments, statistically significant bacterial reductions were determined by comparing bacteria counts of each respective condition to the antibiotic-free control.
After 2 h, the plates were washed 3X with 1X PBS. EMEM-15 was then added to each well, and the plates were allowed to incubate for 48–72 h at 37°C in 5% CO2 as previously published (Gadila and Embers, 2021). After this time, the media in the wells was replaced with EMEM-15 supplemented with the designated concentration of antibiotic(s). Wells that had no antibiotics or wells that were uninfected with Bartonella served as the controls. The cells were allowed to incubate for 4 days at 37°C in 5% CO2. The wells were then washed 3X with 1X PBS, and 100 μL (singular antibiotics) or 300 μL (combinations) of ice-cold sterile water was added to each well and set on ice for 10 min to lyse. A pipette tip was used to scrape the bottom of the well and vigorously pipette the water in the plate. The replicates were combined, and the total volume (singular antibiotics) was plated on a blood agar plate or 100 μL was plated in triplicate (combination antibiotics) for quantification. Plates were checked after 10–14 days when visible colonies were present.
2.5 Antibiotic testing on Bartonella biofilmsTo set up biofilm growth, Collagen I (Corning, 354231) was diluted to 7.58 μg/cm2 in 0.01 M HCl. Using a culture treated 96-Well plate, 50 μL of Collagen I solution was added to each well. These plates were incubated at room temperature for 1 h, then were rinsed 2 times with sterile 1X PBS. Plates were sterilized with UV light for 15 min before use or storage at 4°C for up to 7 days.
On the initial day of the assay (Figure 3), B. henselae was grown as described in bacterial culturing. The 0.1 OD600 inoculum was further diluted 1:100 for biofilms at an approximate concentration of 3 × 105 CFU/ml. In each well, 200 μL of diluted bacteria was added and allowed to incubate for 3 days at 37°C in 5% CO2. Wells containing only Schneider’s media served as a negative control. After incubation, the supernatant in each well was aspirated, with care to not disturb the bottom or sides of the wells. The media was replaced with fresh Schneider’s medium with the appropriate dilution of antibiotics. Schneider’s media without antibiotics was added to the control wells. The plates were then incubated for an additional 4 days at 37°C in 5% CO2. After incubation, the plates were washed three times with 1X PBS, with care taken to not disturb the bottom or sides of the wells. After washing, 150 μL of sterile PBS was used to vigorously wash the bottom and sides of the well. The bottom and sides were also scraped with a pipette tip to ensure maximum recovery of the biofilms. The replicates were combined into a single tube and 100 μL was plated onto blood agar plates in triplicate. Plates were quantified after 10–14 days of incubation at 37°C in 5% CO2. Biofilm formation was optimized in preliminary experiments using crystal violet staining (Supplementary Figure 1).
Schematic of B. henselae biofilm MBC assay. Biofilm matrices were generated by coating plates with Bovine Collagen I. Approximately 200 μL of Schneider’s Media with 3 × 105 CFU/mL of B. henselae was added to the plates and allowed to form biofilms for 72 h before the media in the wells was replaced with media containing antibiotics. After 96 h of treatment, the films were washed with PBS three times and biofilms were disrupted with vigorous pipetting and scraping. The remaining bacteria were combined and plated onto blood agar plates. For monotherapy, MBC was determined by plating the entirety of the remaining bacteria on one plate and designated as positive or negative based on any bacterial growth. In combination experiments, remaining bacteria were plated in triplicate for enumeration. Significance in the MBC of antibiotic combinations was determined by comparing the number of remaining bacteria to the counts in the antibiotic-free control.
3 Results3.1 Monotherapies only inhibit growth of extracellular Bartonella henselaeThe MICs appeared similar for doxycycline, azlocillin, and rifampin against B. henselae grown in either Grace’s or Schneider’s Insect Media (Figure 4 and Table 3). However, B. henselae had higher MICs to gentamicin, azithromycin, and clarithromycin when grown in Grace’s Insect Media compared to Schneider’s. The lowest MIC against B. henselae was observed in media supplemented with rifampin, followed by azlocillin. The highest MIC in both medias was gentamicin. Concentrations of antibiotics tested are reflective of the concentration of antibiotic typically achieved in patient serum (National Academies of Sciences Engineering and Medicine, 2020) with standard doses. Data points are reflective of the higher range interval which showed no growth in three separate iterations of these experiments. For example, if gentamicin-treated B. henselae grew at 4 μg/mL but not at 8μg/mL, then the point is set to 8 μg/mL on the graph. It is important to note that there was variation in results between each iteration of the experiment, particularly when B. henselae was grown in Grace’s Media (Figure 4A). This is perhaps reflective of the sensitive and variable nature in which Bartonella grows in different culture medias, and there is no accepted standard for growing these bacteria.
Minimum Inhibitory Concentrations for extracellular B. henselae. Monotherapies of doxycycline, gentamicin, azithromycin, azlocillin, rifampin, and clarithromycin were tested against extracellular B. henselae grown in either Grace’s (A) or Schneiders (B) Media. Concentrations of antibiotics shown on the y-axis represent the upper limit of the range in which bacterial growth was no longer detected through absorbance value determinations. These experiments were repeated three times and values for each test are represented by one symbol. The starting inoculation dose and antibiotic-free control for each assay was approximately the same for each replicate (Figures 5K, 6M).
Monotherapy efficacy against B. henselaeMinimum inhibitory concentrationBartonella henselae minimum bactericidal concentrations.
Most singular antibiotic treatments against B. henselae were ineffective at killing the bacteria in lower concentrations as seen in Table 3, although higher concentrations of the aminoglycoside gentamicin were effective at killing bacteria. This is surprising because rifampin and azlocillin are bactericidal as well. It would be expected that the bactericidal antibiotics killed B. henselae at higher concentrations while the bacteriostatic antibiotics did not. Monotherapy was also ineffective at treating B. henselae residing in DH82 canine macrophages and when B. henselae was grown as a biofilm, as is shown in Table 3. Indeed, the majority of blood agar plates from the biofilm assays contained too many colonies to count (approximately at least 300 colonies), regardless of concentration of antibiotic used (not shown).
Extracellular bactericidal activity was measured in both Schneider’s (Figure 5) and Grace’s (Figure 6) insect medias as both reagents are used in the literature. The most effective combinations were azithromycin and azlocillin (Figures 5E, 6D) as well as gentamicin in combination with either azlocillin (Figures 5H, 6J) or rifampin (Figures 5J, 6L), as these combinations reduced B. henselae loads significantly in both medias. Efficacy in this context is evidenced by significant bacterial reductions at lower concentrations of antibiotics when compared to saline controls. Gentamicin, azlocillin, and rifampin are all bactericidal antibiotics, which may explain the increased efficacy in these combinations when compared to doxycycline combinations. Azithromycin and azlocillin, the combination that our lab has previously shown to be effective against B. henselae (Gadila and Embers, 2021), was recapitulated in these experiments (Figures 5E, 6D), with similar results.
Combination therapy against extracellular B. henselae grown in Schneider’s Media. Dual combination therapy against extracellular B. henselae grown in Schneider’s Media yielded countable results in 9 different combinations at various concentrations: (A) Doxycycline and gentamicin, (B) doxycycline and azlocillin, (C) doxycycline and azithromycin, (D) doxycycline and rifampin, (E) azithromycin and azlocillin, (F) azlocillin and rifampin, (G) azlocillin and clarithromycin, (H) gentamicin and azlocillin, (I) gentamicin and clarithromycin, and (J) gentamicin and rifampin. Concentrations of antibiotics were a 1:1 ratio of the concentration listed on the x-axis. Effective bactericidal activity is shown as significant p-values generated from Kruskal-Wallace tests with Dunn’s corrections for multiple comparisons and show comparison between the saline control group and the respective concentration. Concentrations of 0.05 μg/mL not listed (A,B,J) are indicative of results that were unable to be enumerated. The inoculation dose and the saline controls are represented in (K), with the dotted line on the graph representing the target inoculation dose.
Combination therapy against extracellular B. henselae grown in Grace’s Media: (A) Doxycycline and gentamicin, (B) doxycycline and azlocillin, (C) doxycycline and rifampin, (D) azithromycin and azlocillin, (E) azithromycin and clarithromycin, (F) azithromycin and rifampin, (G) azlocillin and rifampin, (H) azlocillin and clarithromycin, (I) rifampin and clarithromycin, (J) gentamicin and azlocillin, (K) gentamicin and clarithromycin, and (L) gentamicin and rifampin. A total of 12 different dual therapies yielded countable reductions in B. henselae burden when grown extracellularly in Grace’s Media. Concentrations of antibiotics were a 1:1 ratio of the concentration listed on the x-axis. Specific combinations and concentrations are indicated by a statistically significant reduction in bacteria compa
Comments (0)