Lyme disease, caused by Borrelia burgdorferi, is the most common vector-borne disease in the United States with an estimated 300,000 cases in 2013 (CDC, 2015a). The infection is transmitted to humans by tick vectors that feed upon rodents, reptiles, birds, and deer (Radolf et al., 2012). In the early stage of Lyme disease, patients often have localized erythema migrans skin lesions, but late stage Lyme disease is a disseminated multi-system disorder with signs and symptoms including arthritis, carditis, and neurologic impairment (CDC, 2015a). The majority of Lyme disease patients can resolve their symptoms if treated promptly with doxycycline, amoxicillin, or cefuroxime (Wormser et al., 2006). However, at least 10–20% of Lyme disease patients have lingering symptoms such as fatigue, muscular and joint pain, and neurologic impairment, 6 months after the standard 2–4 week antibiotic treatment, a collection of symptoms called Post-Treatment Lyme Disease Syndrome (PTLDS; CDC, 2015b).

While the cause of PTLDS is unknown, there are several theories, including co-infections (Swanson et al., 2006), autoimmune response (Steere et al., 2001), immune response to continued presence of antigenic debris (Bockenstedt et al., 2012), as well as B. burgdorferi persisters that are not killed by the current antibiotics (Hodzic et al., 2008, 2014; Embers et al., 2012). Using xenodiagnosis and quantitative PCR, various studies have found evidence of B. burgdorferi persistence in dogs (Straubinger et al., 1997), mice (Hodzic et al., 2008, 2014), monkeys (Embers et al., 2012), and humans (Marques et al., 2014) after antibiotic treatment, though no viable bacteria could be cultured.

Recently, B. burgdorferi has been shown to develop persisters in stationary phase cultures in vitro, which are tolerant to the antibiotics used to treat Lyme disease (Feng et al., 2014a, 2015a; Caskey and Embers, 2015; Sharma et al., 2015). These persister bacteria have an altered gene expression profile, which may underlie their phenotypic drug tolerance (Feng et al., 2015c). In log phase cultures (3–5 day old), B. burgdorferi is primarily in motile spirochetal form and is highly susceptible to current Lyme antibiotics doxycycline and amoxicillin (Feng et al., 2014a, 2015a). However, in stationary phase cultures (7–15 day old), increased numbers of atypical morphological variants such as round bodies and aggregated biofilm-like microcolonies develop (Feng et al., 2014a, 2015a). These atypical forms can be considered part of the heterogeneous persisters (Zhang, 2014) as they have high tolerance to doxycycline and amoxicillin compared to growing spirochetal forms (Feng et al., 2014a, 2015a). Therefore, stationary phase cultures (7–15 day old) which are enriched in persisters have been used as a surrogate persister model for high-throughput drug screens against persister populations (Feng et al., 2014a, 2015a,b,d) and have been shown to have overlapping persister populations with amoxicillin tolerant round body persisters (Feng et al., 2016a).

Drugs with high activity against the B. burgdorferi stationary phase persisters were identified through screens of FDA approved drug library and NCI compound library (Feng et al., 2014a, 2015a,b,d). Among them, daptomycin, a lipopeptide antibiotic targeting bacterial cell membranes used to treat MRSA, has the highest activity against B. burgdorferi persisters (Feng et al., 2014a). Although the anti-persister drugs such as daptomycin are more active than the current Lyme antibiotics such as doxycycline or amoxicillin against B. burgdorferi persisters (Feng et al., 2014a), they alone could not completely eradicate the more resistant biofilm-like microcolonies and a drug combination approach is required to do so (Feng et al., 2015a). The more effective drug combination approach to eradicate biofilm-like microcolonies is consistent with the drug combination principle for treatment of persistent infections like tuberculosis (Zhang et al., 2012; Zhang, 2014). In a recent study, ceftriaxone pulse dosing was shown to completely eradicate B. burgdorferi persisters using a relatively young 5 day old late log phase culture (Sharma et al., 2015), which is known to primarily consist of spirochetal form that is more susceptible to antibiotics. It is unclear whether the pulse dosing approach is able to eradicate the more resistant microcolony form of persisters (Feng et al., 2015a). In this study, we evaluated the pulse dosing approach with ceftriaxone and also other Lyme antibiotics and drug combinations using an older stationary phase culture (10 day old) enriched with more resistant biofilm-like microcolonies as well as a 5 day old late log phase culture as a control to determine their ability to eradicate B. burgdorferi persisters. We found that four rounds of pulse dosing treatment with ceftriaxone could eliminate mainly the planktonic spirochetal form of B. burgdorferi in the log phase and the stationary phase culture, but failed to eradicate the aggregated microcolony form of B. burgdorferi persisters in the stationary phase culture.

A previous study showed that ceftriaxone could eliminate all cells including persisters in a 5 day old B. burgdorferi culture after four rounds of pulse dosing treatment (Sharma et al., 2015). In this study, we were able to confirm this finding using the same protocol with a relatively young 5 day old late log phase B. burgdorferi culture. However, we found that although four rounds of ceftriaxone pulse dosing treatment could eliminate nearly all planktonic spirochetes in the 5 day old culture and even those in a 10 day old stationary phase B. burgdorferi culture, it failed to eradicate the more resistant aggregated biofilm-like microcolonies enriched in stationary phase cultures (Figure 1). Our previous studies demonstrated that a 5 day old culture is quite young and could be considered late log phase (Feng et al., 2014a) rather than “stationary phase” as used in a previous study (Sharma et al., 2015) and is mainly consisting of spirochetal organisms that are more susceptible to antibiotics. In contrast, 7–10 day old cultures of B. burgdorferi can be considered true stationary phase and contain higher proportions of aggregated microcolony forms, which are much more tolerant to antibiotics than the spirochetal forms (Feng et al., 2015a). Thus, it is not surprising that ceftriaxone failed to eradicate the more resistant biofilm-like microcolonies of B. burgdorferi as shown in this study (Figure 1), as heterogeneous microcolonies have previously been shown to be highly tolerant to antibiotics including single drugs as well as many drug combinations except daptomycin/doxycycline/cefoperazone (Feng et al., 2015a). In addition, a short 1 day recovery in fresh medium used in this study per the previous pulse dosing protocol (Sharma et al., 2015) may not be sufficiently long to allow all the dormant forms in aggregated microcolonies to wake up and become susceptible to ceftriaxone. Indeed, the time interval in between pulse dosing treatment needed to allow the persisters to revive and become susceptible to antibiotics is hard to determine not only for in vitro persisters but also for in vivo persisters in patients, considering the heterogeneity of the bacteria and the disease and individual differences in immune control, which could pose significant challenges to the use of pulse dosing in clinical settings (Zhang, 2014).

An important observation of this study is that pulse dosing may not always work as a principle against persisters but rather its effect depends on the particular drugs being used and that not all drugs are suitable for pulse dosing. We found that daptomycin, which is the most effective antibiotic against B. burgdorferi persisters (Feng et al., 2014a, 2016a), fared poorly in pulse dosing experiments either singly (Figure 2A) or in drug combinations (Figure 3A), as it had even less activity than the ceftriaxone and doxycycline in the pulse dosing treatment (Figure 2A). Despite its excellent activity against non-growing B. burgdorferi persisters (Feng et al., 2014a), daptomycin is known to have limited activity against growing spirochetal forms as shown by a relatively high MIC (12.5–25 μg/ml) compared to ceftriaxone and doxycycline (Feng et al., 2014a). When used at a concentration under its MIC (5 μg/ml) as in this study, daptomycin is expected to have little activity against growing B. burgdorferi thus explaining its poor activity in the pulse dosing treatment. However, we do find better activity when daptomycin was combined with antibiotics such as ceftriaxone and doxycycline which act on actively growing bacteria even in the pulse dosing experiment (Figure 2A). The basic idea of pulse dosing treatment is to allow non-growing antibiotic tolerant persisters formed after drug treatment to recover and become growing spirochetes so they become susceptible to drugs again. Thus, pulse dosing seems to work well with cidal antibiotics as shown with ceftriaxone (Sharma et al., 2015) (Figure 1) and cefuroxime (Figure 2A) but not with static drugs like doxycycline (Caskey and Embers, 2015) (Figure 2Ac) and persister drugs such as daptomycin as shown in this study (Figure 2A). Our findings suggest that only bactericidal antibiotics like cephalosporins may be suitable for pulse dosing but not bacteriostatic agents or persister drugs. These observations may have important implications on clinical use of pulse dosing for treatment of persistent infections like Lyme disease.

Given the role of pulse dosing with cidal antibiotics, its effect as a single drug on the more resistant microcolonies is limited and could not eradicate aggregated biofilm-like structures (Figure 1). In our previous studies, we found that drug combinations containing persister drug daptomycin are more effective against B. burgdorferi persisters than single drug treatment (Feng et al., 2015a, 2016b). Consistent with this, we found that the drug combination pulse dosing treatment also showed more activity against stationary phase B. burgdorferi than single drug pulse dosing (Figures 3A and 2A). This is supported by microscopic observations that the amount and the size of the aggregated microcolonies were significantly reduced after the drug combination pulse dosing treatment compared with the single drug pulse dosing treatments (Figure 3A). In addition, three drug combination pulse dosing treatment was more effective than two drug combination pulse dosing treatment (Figure 3A). However, either the two drug combination or even three drug combination pulse dosing were still unable to eradicate the more resistant microcolony forms, and the residual microcolonies grew up again in the subculture test without antibiotics (Figure 3B).

We previously showed that daptomycin/doxycycline combined with cefoperazone or cefuroxime as a triple drug combination has the best activity against stationary phase B. burgdorferi cultures and could completely eradicate all bacteria including the most resistant aggregated microcolony form of persisters (Feng et al., 2015a, 2016b). In this study, we found the ceftriaxone or cefuroxime could also replace cefoperazone in the daptomycin/doxycycline combination to kill all the stationary phase B. burgdorferi cells (Figure 3). Unexpectedly, however, the triple drug combination daptomycin/doxycycline/ceftriaxone pulse dosing treatment turned out to be less effective than the continuous triple drug combination treatment as shown by SYBR Green I/PI assay as well as the subculture test (Figure 3). Thus pulse dosing treatment compromised the activity of daptomycin-containing drug combinations through reducing the activity of daptomycin as shown in Figure 2Ad. Further studies are required to validate the findings of these in vitro studies in vivo and compare the utility of the pulse dosing and the drug combination treatment in eradication of borrelia persistence phenomenon in animal models.

We found that four rounds of ceftriaxone pulse dosing could eradicate planktonic log phase B. burgdorferi spirochetal forms and round body forms but not the more resistant aggregated biofilm-like microcolony forms in the stationary phase culture. In addition, we showed that not all drugs are appropriate for pulse dosing, with ceftriaxone and cefuroxime being more suitable for pulse dosing than bacteriostatic drug doxycycline and persister drug daptomycin. This study also demonstrated that drug combination pulse dosing treatment is more effective than single drug pulse dosing. Importantly, we found that pulse dosing could impair the activity of the persister drug daptomycin and its drug combination on B. burgdorferi persisters and that the more effective way to kill the aggregated biofilm-like microcolony form of persisters is the triple daptomycin drug combination without pulse dosing. Future studies are needed to validate these findings in animal studies and clinical trials.

YZ conceived the experiments; JF, SZ, and WS performed the experiments; JF and YZ analyzed the data; JF and YZ wrote the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer NBDC and handling Editor declared their shared affiliation and the handling Editor states that the process nevertheless met the standards of a fair and objective review.