Introduction

Bacterial infections pose a global health threat, driving widespread antibiotic use [1]. Under antibiotic pressure, bacteria have evolved intricate survival mechanisms, leading to treatment failures [2, 3]. While antibiotic-resistant mutants are well-known causes of treatment failure [4], a small subpopulation (0.001–1%) of bacteria, even when sensitive to antibiotics, transiently evade antibiotic action, known as persisters [5, 6].

First described in 1944 by Joseph Bigger [7], who noted the survival of a subset of Staphylococcus under lethal penicillin doses, persistence refers to bacteria’s ability to survive exposure to bactericidal drugs [8]. These bacteria do not grow in the antibiotic’s presence but resume growth once the stress is removed [9, 10]. Studies found recurrent infections like Salmonella and Streptococcus pyogenes were often caused by the same genovar, suggesting persistent bacteria drive these infections, leading to increased antibiotic use [11, 12]. Bacterial persistence is a significant public health concern as infections from persisters lead to antibiotic failure [13] and contribute to antibiotic resistance [11, 14]. This poses a real problem; especially as current trends are to reduce the use of antibiotics. The study of bacterial persistence is therefore of great importance to public health.

In vitro, bacteria exhibit various phenotypes and destruction kinetics under antibiotic treatment, such as resistant, persistent, tolerant, and susceptible bacteria. Persistent bacteria exhibit a characteristic biphasic killing curve, where most susceptible cells are rapidly eradicated by a high antibiotic concentration, while a small proportion survives longer due to the presence of a mixed bacterial population [15, 16]. To form persisters, bacteria utilize various mechanisms, including toxin-antitoxin (TA) systems [17]. There are currently eight types (I to VIII) of TA systems found in nearly all bacterial strains [18]. TA systems consist of a toxin that inhibits growth and an antitoxin that neutralizes the toxin’s effects [19]. Among these, Type II is extensively studied, widely distributed, and heavily implicated in bacterial persistence [20, 21].

Unfortunately, despite the clinical significance of persistent cells in bacterial infections, there’s a scarcity of studies on this phenomenon among clinical bacterial isolates, especially in Africa. Most research has focused on laboratory strains. Hence, this study was initiated to evaluate antibiotic susceptibility and assess persister cell formation in clinical isolates of Pseudomonas aeruginosa, Staphylococcus aureus, and Salmonella sp. after exposure to ciprofloxacin and gentamicin. Additionally, due to the potential involvement of TA systems in bacterial persistence, we investigated four type II TA system genes (mazF, hipA, relE1, relE2) in the studied isolates.

Materials and methods

Ethical considerations

Approval for the present study protocol was obtained from the Institutional Ethics Committee of the CERBA/LABIOGENE in its deliberation N° 2022-25/09–015 of 5 September 2022. All study participants or their guardians provided their free and informed consent in accordance with the Helsinki Declaration.

Type and study period

This study was a descriptive cross-sectional investigation focusing on bacteria collected from the laboratories of Hopital Saint Camille de Ouagadougou (HOSCO) and Centre de Recherche Biomoléculaire Pietro Annigoni (CERBA) in Ouagadougou. The collection of bacterial strains occurred between October 2022 and February 2023.

Bacterial strains

Eighty (80) bacterial isolates, comprising 41 Staphylococcus aureus, 32 Pseudomonas aeruginosa, and 07 Salmonella sp., were obtained from diverse human clinical samples. Bacterial suspensions of these isolates in Luria-Bertani (LB) broth supplemented with 20% glycerol were frozen and stored at the Laboratoire de Biologie et de Génétique Moléculaire (LABIOGENE), Université Joseph KI-ZERBO, for subsequent analysis.

Antibiotic susceptibility testing

Antibiotic susceptibility testing was conducted following the manual of procedures for performing antibiograms in Burkina Faso [22], which aligns with the 2015 recommendations of the Antibiogram Committee of the French Microbiology Society and the European Society of Clinical Microbiology and Infectious Diseases (CA-SFM/EUCAST-2015) [23]. The disk diffusion method (Kirby-Bauer) using Mueller-Hinton (MH) agar was employed. A bacterial suspension with a turbidity matching that of the 0.5 McFarland standard was prepared from a bacterial culture obtained after 24 h incubation. MH agar plates were inoculated, and gentamicin and ciprofloxacin antibiotic disks were applied. The plates were then incubated at 37 °C, and readings were taken after 24 h. Inhibition diameters were measured using vernier calipers, and the results were used to classify isolates as susceptible or resistant based on CA-SFM/EUCAST-2015 recommendations [23]. For sensitive strains, the higher critical concentrations of these antibiotics were considered.

Formation of persisters

For persistence testing, 5 bacterial isolates per species were selected. Only isolates sensitive to both gentamicin and ciprofloxacin were included. A random selection of 5 isolates per species was made to identify those used for this purpose. To assess the presence of persisters, bacterial strains were reactivated and plated on MH agar. Existing protocols were adjusted and customized for this study [24,25,26,27,28,29]. Bacterial suspensions in LB broth were incubated at 37 °C on a shaker-incubator (New Brunswick Innova® 44) set at 200 rpm until the optical density at 600 nm, measured using Biomate 3 spectrophotometer (Thermo Fisher Scientific, Waltham, MA), reached a range between 0.2 and 0.3. Two tubes of bacterial suspension were prepared for each isolate. Gentamicin was added to one tube and ciprofloxacin to the other, both at very high concentrations. The entire set was then incubated at 37 °C. The antibiotics used for the study were Gentamicin 80 mg/2 mL (Panpharma, France) and Ciprofloxacin (Cipronat® 200 IV) 200 mg/100 mL (Dafra Pharma GmbH, Switzerland). Concentrations used were 100 times the upper critical concentration as follows: Ciprofloxacin at 0.1 mg/mL and Gentamicin at 0.1 mg/mL for Staphylococcus aureus, Ciprofloxacin at 0.1 mg/mL and Gentamicin at 0.4 mg/mL for Pseudomonas aeruginosa and Ciprofloxacin at 0.1 mg/mL and Gentamicin at 0.4 mg/mL for Salmonella sp.

At 1, 3-, 5-, 7-, and 20-hours post-antibiotic addition, the culture underwent two washes with 0.85% sterile saline to eliminate the antibiotic. Washing involved removing 100 µL of the antibiotic-containing bacterial suspension, adding 500 µL of sterile 0.85% saline, centrifuging and discarding the supernatant. Two serial dilutions (1/100) were then performed, with the final dilution being inoculated onto MH agar plates. The plates were incubated at 37 °C for 24 h. Subsequently, colonies on the agar plates were counted using the APD Colony App Lite cell phone application [30]. Time-kill curves for each isolate were plotted to analyze the typical biphasic curve. An antibiotic-free bacterial culture of each species served as a control.

Detection of bacterial persistence genes

Bacterial DNA extraction was conducted via heat shock following a protocol previously outlined [31]. Bacterial colonies were suspended in 1 mL of distilled water and subsequently boiled in a water bath for 10 min. Centrifugation was carried out at 1000 rpm for 5 min, and the resulting supernatant was collected in Eppendorf tubes. The quantity and purity of DNA were assessed using a BioDrop spectrophotometer (BioDrop, Cambridge, UK).

PCR was used to detect the presence of TA type II system genes (mazF, hipA, relE1, and relE2) using specific primers. Amplification was conducted in GeneAmp® PCR System 9700 thermal cycler (Applied Biosystems, USA) with a program comprising an initial denaturation step at 94 °C for 4 min, followed by 35 cycles of denaturation at 95 °C for 45 s, annealing at 60 °C for 45 s, and extension at 72 °C for 30 s. A final extension step at 72 °C for 5 min was included. Each PCR reaction had a final volume of 25 µl, containing 4 µl of master mix (FIREPol® Master Mix Ready to Load, 5X, Solis BioDyne), 0.5 µl of each primer (10 µM), 1 µl of DNA extract, and sterile PCR water to reach the final reaction volume.

PCR products were separated by electrophoresis (Mupid® One) on a 1.5% agarose gel (Cleaver Scientific, UK) containing ethidium bromide. Electrophoresis was conducted in 1X Tris-Acetate-EDTA buffer for 40 min at 100 volts alongside a 100 bp molecular weight marker (100 bp Ladder Ready to Load, Solis BioDyne). After migration, amplicons were visualized under UV light using the Trans-illuminator E-BOX and photographed. The primers used for the genes are listed in Table 1.

Table 1 TA system genes to search for and their primers
Full size table

Statistical analysis

Data were entered into Excel 2019. Statistical analyses and the elaboration of figures were carried out using STATA software version 14.

Results

Our study population consisted of 80 bacterial isolates, including 32 Pseudomonas aeruginosa, 41 Staphylococcus aureus and 7 Salmonella sp. The bacteria were isolated from diverse biological samples, with most isolates coming from pus (48/80) as indicated in Table 2.

Table 2 Distribution of bacterial species by sample type
Full size table

Susceptibility of bacterial strains to the antibiotics tested

Our results revealed high sensitivities of clinical isolates to the two antibiotics used: ciprofloxacin (72.5%) and gentamicin (82.5%). These findings are detailed in Table 3. It’s noteworthy that 53 isolates, specifically 27 P. aeruginosa, 21 S. aureus, and 5 Salmonella sp., exhibited sensitivity to both ciprofloxacin and gentamicin.

Table 3 Sensitivity of bacterial strains to ciprofloxacin and gentamicin
Full size table

Bacterial persistence under antibiotic stress

In Pseudomonas aeruginosa, no persistent phenotype was observed (Fig. 1). Among the isolates, P.a_54 and P.a_60 were sensitive to ciprofloxacin, whereas P.a_13, P.a_47, and P.a_79 displayed distinctive biphasic curves, indicating initial destruction followed by a gradual increase in Colony Forming Unit (CFU) (Fig. 1A). Regarding gentamicin, only P.a_54 was sensitive, while the remaining four isolates exhibited resistance (Fig. 1B).

Fig. 1
figure 1

Survival curve of P. aeruginosa strains as a function of time in the presence of ciprofloxacin (A) and gentamicin (B)

Legend : P.a: Pseudomonas aeruginosa; Numbers correspond to the identification number given to Pseudomonas aeruginosa; QC: Quality Control strain.

Full size image

In S. aureus, we present the results of testing on four strains. At the end of the manipulations, we realized that the fifth had been contaminated and the results were not usable. Persistent cells were observed in the presence of gentamicin, as evidenced by the biphasic curves (S.a_9 and S.a_84) in Fig. 2B, indicating initial destruction followed by a consistent or gradually decreasing number of CFUs. Additionally, two isolates (S.a_9 in Fig. 2A and S.a_77 in Fig. 2B) displayed distinctive biphasic curves similar to those described for P. aeruginosa. The remaining isolates exhibited a susceptible phenotype.

Fig. 2
figure 2

Survival curve of S. aureus strains as a function of time in the presence of ciprofloxacin (A) and gentamicin (B)

Legend: S.a: Staphylococcus aureus; Numbers correspond to the identification number given to Staphylococcus aureus; QC: Quality Control strain.

Full size image

In Salmonella sp., persistent cells were observed in the presence of ciprofloxacin, as evidenced by the biphasic curve (Sal_20) in Fig. 3A, showing initial destruction followed by a gradual decrease in CFUs. The remaining four isolates displayed a susceptible phenotype. Additionally, all five isolates were sensitive to gentamicin, as shown in Fig. 3B.

Fig. 3
figure 3

Survival curve of Salmonella strains as a function of time in the presence of ciprofloxacin (A) and gentamicin (B)

Legend: Sal: Salmonella sp.; Numbers correspond to the identification number given to Salmonella sp; QC: Quality Control strain.

Full size image

Persistence genes

The outcomes of the persistence gene search are compiled in Table 4. None of the bacteria harbored more than one of the targeted genes. The gel images displaying the amplicons of persistence genes are depicted in Fig. 4.

Table 4 Frequency of bacterial persistence genes
Full size table
Fig. 4
figure 4

Electrophoresis of persistence gene PCR products on 1.5% agarose gel. (A: hipA (156 bp); B: mazF (110 bp))

Legend: PM: 100 pb DNA ladder ; E51, E53, E57, E59: bacterial strain + identification number. Full length gels are presented in supplementary maz-hipa_1–3.

Full size image

Discussion

Despite the public health problem caused by bacterial persistence, few studies have addressed this phenomenon among clinical bacterial isolates, particularly in Africa. Our objective was to assess the susceptibility profile of bacteria in our study population to two commonly used antibiotics in Burkina Faso. Subsequently, we aimed to demonstrate the formation of bacterial persistence after exposing these isolates to the antibiotics. Lastly, we sought to identify four genes (mazF, hipA, relE1, relE2) of the type II TA system associated with this phenomenon in these clinical isolates.

More than half of the 80 bacterial strains in our study originated from pus samples. This is consistent with the fact that our study population primarily comprised P. aeruginosa and S. aureus, which are pyogenic bacteria commonly found in such samples. This observation is also reflected in the various laboratory antimicrobial resistance surveillance reports in Burkina Faso, which have been published annually since 2018 [22, 32,33,34]. Ciprofloxacin and gentamicin exhibited favorable activity against all bacterial strains in the current study, with rates of 72.5% and 82.5%, respectively. These antibiotics are recognized as broad-spectrum agents renowned for their potent bactericidal effects on both Gram-negative and Gram-positive bacteria. Consequently, they are extensively utilized in Burkina Faso for treating bacterial infections.

Our findings regarding Salmonella sensitivity (71.4% to ciprofloxacin and 85.7% to gentamicin) are similar to national data. The latest 2022 report on national surveillance of antimicrobial resistance in Burkina Faso [34] revealed sensitivities of 75.76% and 87.10% to ciprofloxacin and gentamicin, respectively. Similarly, national sensitivities for gentamicin in P. aeruginosa (96.7%) and S. aureus (70.7%) closely mirrored our results, with national figures at 71.79% for P. aeruginosa and 78.20% for S. aureus. However, notable differences were observed in ciprofloxacin sensitivities. National data indicated low sensitivities for P. aeruginosa (19.38%) and S. aureus (18.21%), whereas our study found much higher sensitivities at 84.4% for P. aeruginosa and 63.4% for S. aureus. Other studies have also reported low sensitivities of P. aeruginosa, such as 28.57% to gentamicin and 33.57% to ciprofloxacin [35]. These findings highlight the presence of still sensitive strains and emphasize the need to intensify efforts to curb the spread of resistance genes. The high antibiotic sensitivities observed in bacterial isolates during our study led to a significant number of strains being sensitive to both ciprofloxacin and gentamicin (53 strains in total, comprising 27 P. aeruginosa, 21 S. aureus, and 5 Salmonella sp.). This enabled us to employ random sampling to select five strains of each species for investigating bacterial persistence.

P. aeruginosa, S. aureus, Salmonella and many other bacteria have been recognized as harboring persistent cells [6]. To examine biphasic killing curves, we counted CFUs up to 20 h after antibiotic exposure in exponential growth phases. Out of the 5 Pseudomonas aeruginosa isolates tested with ciprofloxacin, 2 demonstrated susceptibility. The remaining 3 displayed distinctive biphasic curves, indicating rapid destruction followed by a progressively increasing number of CFUs, suggesting the proliferation of resistant bacteria. This phenomenon could be indicative of heteroresistance, characterized by a heterogeneous behavior where a subset of bacterial cells can multiply in the presence of antibiotic concentrations lethal to most of the population. Since the population primarily consists of susceptible bacteria, the Minimum Inhibitory Concentration (MIC) of the overall population is comparable to that of a susceptible strain, which might not have been detected by the initial susceptibility test. These curves exhibited similar characteristics to those developed by Gollan et al [36]. , outlining the heterogeneous behavior of a resistant bacterial subpopulation surviving in the presence of antibiotics at concentrations fatal to the rest of the population. Considering the experiments conducted by Mlynarcik and Kolar [37], this could also indicate acquired resistance to ciprofloxacin, as they also observed this phenomenon while testing P. aeruginosa with tobramycin. On the contrary, in gentamicin tests, only one strain exhibited a sensitive phenotype, while the remaining four displayed a resistant phenotype, despite their initial sensitivity. This may suggest resistance developed through exposure to antibiotics or acquired resistance via resistance genes. In summary, our findings suggest that ciprofloxacin and gentamicin did not induce persistence formation in the tested P. aeruginosa isolates. However, other studies have shown persistence formation in P. aeruginosa in the presence of ciprofloxacin [17]. A larger-scale study would have provided more insight into identifying persistent cells.

Out of the planned 5 isolates of Staphylococcus aureus, only 4 were tested. Among these, 3 isolates exhibited a phenotypically sensitive response to ciprofloxacin, indicating the efficacy of this antibiotic against S. aureus strains. However, one isolate displayed a phenotype suggestive of heteroresistance when exposed to ciprofloxacin, similar to the observed behavior in P. aeruginosa. Interestingly, ciprofloxacin did not induce persistence in S. aureus. Conversely, in the presence of gentamicin, 2 isolates demonstrated persistence phenotypes. Our findings revealed that gentamicin could induce bacterial persistence in S. aureus. This aligns with other studies that have also reported the formation of persistence in S. aureus when exposed to gentamicin [38]. Our findings are inconsistent with certain studies that have highlighted antibiotic-dependent persistence. For instance, a study examining 10 clinical isolates of Staphylococcus revealed the formation of persisters in over half of the isolates when exposed to vancomycin, oxacillin, ciprofloxacin, and penicillin. However, fewer isolates displayed persistence in the presence of gentamicin [39].

All 5 strains of Salmonella exhibited a gentamicin-sensitive phenotype. However, in the presence of ciprofloxacin, only one strain displayed a biphasic curve indicative of a bacterial persistence phenotype, while the remaining four showed sensitive phenotypes. Our findings highlighted the superior activity of gentamicin against the tested Salmonella strains. Some authors suggest that the persistence of these bacteria may occur within spleen and liver macrophages [40, 41]. Contrary to expectations, our results demonstrated that ciprofloxacin was capable of inducing persistence in Salmonella. Interestingly, ciprofloxacin is commonly prescribed as the first-line treatment for salmonellosis in Sub-Saharan Africa, including Burkina Faso, where these infections are prevalent. This observation might contribute to the high frequency of recurrent infections with these bacteria and the alarming rise in Salmonella resistance to fluoroquinolones, a concerning global trend.

The mazEF system is a well-studied type II TA module found in numerous bacteria [6, 42, 43]. However, in our study, the mazF gene was not detected in Salmonella sp., possibly due to the limited number (7) of isolates examined. We did identify the mazF gene in a small percentage of P. aeruginosa (28.13%) and S. aureus (2.44%) isolates. Our findings contrast with previous literature, which often reports a higher prevalence of the mazF gene in clinical isolates. For instance, a study in Iran [35] found the gene in 85.71% of 140 clinical P. aeruginosa strains, while in Turkey, it was present in 89.1% of 148 Staphylococcus isolates [44]. Additionally, another study reported universal mazF gene presence across all strains examined (78 S. aureus and 42 P. aeruginosa) [45]. A study reported a very low frequency (1.42%) of the hipA gene in 140 clinical strains of P. aeruginosa [35], a finding that aligns with our results as we did not detect this gene in any P. aeruginosa isolates. With a larger sample size, we might have identified at least one isolate with the hipA gene. Similarly, although some studies have identified the hipA gene in Salmonella [21], we only detected it in 2 Salmonella isolates in our study. Despite being the first gene associated with bacterial persistence [46], hipA does not appear to be prevalent in clinical isolates.

In our study, we did not detect the presence of relE1 or relE2 genes. However, these genes are recognized to play a role in bacterial persistence, particularly in clinical strains of P. aeruginosa as noted by Fernández-García et al. [42]. Additionally, studies have reported the presence of these genes in clinical isolates of S. aureus [38] and Klebsiella pneumoniae [47].

Conclusion

The threat of antibiotic resistance is well-documented in Burkina Faso, but bacterial persistence remains less explored. Our study aimed to fill this gap by investigating persistence in clinical isolates. Our findings demonstrate that these isolates can form persister cells when exposed to commonly used antibiotics like ciprofloxacin and gentamicin, and they possess type II TA system genes associated with persistence. Understanding bacterial persistence is crucial for addressing treatment failures, particularly in the context of recurrent infections. Further research is warranted to comprehensively understand and tackle this phenomenon in bacterial infections.