Lyme disease is the most rapidly growing tick borne zoonotic disease of the Northern Hemisphere and is among the 10 most commonly reported nationally notifiable diseases in the United States.1 Clinical presentations include erythema migrans, fever, chills, muscle and joint pain.2, 3 Though these symptoms tend to fade away even without therapeutic intervention, a significant number of untreated patients develop arthritis and persistent myalgia following exposure to Borrelia burgdorferi.4 Furthermore, 10–20% of patients treated for Lyme disease develop symptoms considered typical, or even exaggerated, including muscle, joint pain and generalized fatigue5, 6. This condition is referred as post-treatment lyme disease syndrome (PTLDS). Though existence of PTLDS is debatable, some researchers consider the presence of persister forms of B. burgdorferi and/or the continuous presence of antigenic debris to be the underlying cause of PTLDS.7, 8, 9, 10

Like other pathogens, B. burgdorferi protects itself from the immune system and from drug treatment.11, 12 B. burgdorferi evades immune response by antigenic variation of its surface proteins13, 14, 15. In a recent study, researchers identified B. burgdorferi persisters in in vitro cultures.12 They found the killing of B. burgdorferi by antibiotics is biphasic, with a small subpopulation of surviving persisters.12 The surviving antibiotic tolerant cells are not resistant mutants upon regrowth, the population bifurcates into new antibiotic-susceptible and new persister subpopulations.12 Currently prescribed drugs for treating Lyme disease, including amoxicillin, ceftriaxone and doxycycline were unable to completely eliminate the B. burgdorferi.12, 16, 17, 18 So, efforts to identify new, potent drug candidates for Lyme disease are on the rise.

Many researchers are performing high-throughput screening of drugs against B. burgdorferi persisters to identify molecules that can eliminate complete Borrelial infection.7, 17, 18, 19, 20 Screening of chemical compound libraries serves to test a large number of structurally and functionally diverse molecules against pathogenic agents. Among the many chemical libraries currently available, the Developmental Therapeutics Program of the National Cancer Institute, National Institute of Health, provides a unique yet diverse array of chemical compounds for screening purpose. Four sets of compounds within the National Cancer Institute-Developmental Therapeutics Program (NCI-DTP) library (http://dtp.nci.nih.gov/) tend to represent a wide variety of structural and functional diversity.

This study aimed to identify new, effective drugs for Lyme disease. We used a well-established, highly efficient, BacTiter-Glo assay (Promega Corporation, Fitchburg, WI, USA) which can detect as few as 7 × 103 Borrelial cells in BSK-II medium.21, 22 The NCI-DTP compound library of four diverse sets containing 3084 chemical compounds was screened using this assay. We identified 101 unique compounds which inhibited Borrelia growth by more than 85% at or below a concentration of 25 μM. From these 101 compounds we selected 12 molecules and studied their MIC and MBC. The lead compounds identified in the current study can be further evaluated for their therapeutic potential in pre-clinical and clinical studies. Moreover, the outcomes of the study could provide a deeper insight into treatment strategies for Lyme disease.

We have developed a one-step, straightforward, highly sensitive BacTiter-Glo (Promega Corporation) Assay to screen drugs in high-throughput format. We optimized a BacTiter-Glo (Promega Corporation) Assay in high-throughput screening format as reported in our previous papers.21, 22 The BacTiter-Glo Assay (Promega Corporation) assesses bacterial viability by measuring ATP in the sample. This sensitive assay can reliably detect as few as 10 Borrelia cells in phosphate buffered saline or 7 × 103 Borrelia cells in BSK-II medium21, 22. By using this BacTiter-Glo (Promega Corporation) Assay we have screened NCI-DTP library containing 3084 chemical compounds.22 The NCI-DTP compound library we have screened contains a total of 3084 chemical compounds from four highly divergent sets viz, Structural diversity set (1974 compounds), Mechanistic diversity set (827 compounds), Natural Products set (230 compounds) and Challenge set (53 compounds). The NCI-DTP library was obtained from High-Throughput Bioscience Center (HTBC), Stanford University. The NCI-DTP library stocks were maintained in dimethyl sulfoxide (DMSO) solutions at 10 mM compound concentrations.

Screening of NCI-DTP library was performed in 384-well plate format according to the procedure we have reported earlier.22 The NCI-DTP library screening was performed on stationary phase Borrelial cultures grown in BSK-II medium for 7–10 days. The NCI-DTP library compounds from the 10 μM DMSO stocks were pinned into 384-well plates and then 106 ml–1 B. burgdorferi stationary phase cultures with BSK-II medium was added. Each compound was tested at seven different concentrations ranging from 25 to 0.45 μM in a seven-point titration (that is, 25, 12.5, 7.25, 3.625, 1.81, 0.9, 0.45 μM). Then, these 384-well culture plates were incubated at 33 °C for 96 h in a 5% CO2 incubator. After 96 h of incubation, the BacTiter-Glo (Promega Corporation) reagent was added to the plates and luminescence was measured in relative luminescence units on a Flex Station 3 Microplate Reader (Molecular Devices LLC, Sunnyvale, CA, USA). The data was analyzed and interpreted by MDL Assay Explorer (Elsevier MDL, San Ramon, CA, USA). Compounds were designated as hits if the luciferase signal was decreased significantly in comparison with the control compounds.22

From the primary screening, we have identified 101 hit molecules which inhibited bacterial growth ⩾85% compared with control. Out of these 101 hit molecules, 88% are FDA approved compounds. We selected top 12 molecules which were shown in Table 1, based on their ability to inhibit >95% of Borrelia growth in primary screening. This level of inhibition is higher than the currently prescribed drugs, doxycycline (94.14%) and ceftriaxone sodium (95.26%) which is shown in Table 1. We also confirmed the activity of these 12 compounds (Table 1) by a secondary screening with BacTiter-Glo (Promega Corporation) Assay. In addition, we provided data on 31 novel compounds belonging to different classes of molecules, which inhibited more than 95% of Borrelial viability (Supplementary Table S1).

Table 1 Structure and activity of novel Top 12 hits against B. burgdorferi
Full size table

After the primary screening the hits were validated again by secondary screening with BacTiter-Glo (Promega Corporation) cell viability assay.21 The BacTiter-Glo (Promega Corporation) cell viability assay was performed as per manufacturer’s instructions.21 All the compounds tested in secondary screening were graciously provided by National Cancer Institute Development Therapeutics Program, National Institute of Health, USA. The compounds were tested in 96-well plates at different drug concentrations ranging from 250 to 0.31 μM. The efficacy of some compounds that were confirmed by secondary screening using BacTiter-Glo (Promega Corporation) Assay is shown in Figure 1. The compounds chloro-triphenyl-stannane showed Borrelia viability at 1.25 μM, cinerubin B at 0.31 μM and mikamycin B at 0.31 μM (Figure 1a). In Figure 1b, trichopolyn-B shows Borrelial inhibition at 20 μM, whereas zinc pyrithione shows inhibition at 0.31 μM. The vehicle DMSO (Control) did not show any significant effect on the Borrelia growth. The positive control doxycycline (94.14%) showed viability at 5 μM. Except trichopolyn-B, all other drugs showed better viability than prescribed drug doxycycline. On the basis of bacterial inhibition observed at low concentrations from the secondary screening, the MIC and MBC values were determined for selected potential candidates.

Figure 1
figure 1

Bac-titer Glo Inhibition assay of drugs on CA8 strain. Effect of drugs on Borrelia cell viability was studied with drugs (a) Chloro-triphenyl-stannane, Cinerubin B and Mikamycin B. (b) Trichopolyn-B and Zinc pyrithione. The control has no drugs. The results represent mean±s.d.

Full size image

After confirming the bacterial inhibition by secondary screening, MIC and MBC of the drugs were evaluated to determine the amount of drug needed to kill bacteria. The standard microdilution methods were used to study MIC and MBC of the drugs.16, 22 For the determination of MIC, 106 ml–1 of B. burgdorferi was cultured in BSK-II medium with different concentrations (0.3–160 μM) of test compounds for 72 h at 33 °C. The evaluated MIC and MBC for top 12 hit compounds confirmed after secondary screening were shown in Table 1. The MIC values of cinerubicin B hydrochloride, mikamycin B, zinc pyrithione and nigericin is 0.31 μM. The compound lankacidin C inhibited Borrelia below ⩾1 μM which are shown in Table 1. For the drugs NSC 121145, NSC 75140, chloro-triphenyl-stannane, valinomycin, gliotoxin and michellamine B the MIC values are ⩾3 μM. The MIC value of trichopolyn-B showed 20 μM which is higher compared with other drugs tested. The MBC was determined by sub-culturing 20 μl of the Borrelia cultures grown at different drug concentrations in fresh BSK-II medium for 21 days. The MBC was determined when no motile spirochete was observed microscopically in the subculture.16, 23 Among the selected 12 compounds, cinerubicin B hydrochloride and mikamycin B had the lowest MBC of 1.25 μM. The MBC values for, zinc pyrithione and chloro-triphenyl-stannane are ⩾6 μM. For the remaining other compounds MBC values are higher than 6 μM. The MBC value of trichopolyn-BT, NSC 121145 and NSC 75140 are very high which is more than 40 μM (Table 1).

In this study cinerubin B and the mikamycin B are the drugs which inhibited and killed Borrelia at very low concentrations. The cinerubin B is an anthracycline antibiotic inhibited Borrelial growth by 98% and gave a MIC value of 0.31 μM, thereby making it one of the most potent drug molecules studied. Mikamycin B, a macrolide antibiotic, was similarly potent, inhibiting growth by 98% and having a MIC value of 0.31 μM. These drugs target the DNA replication and translation machinery, respectively.24

In parallel to our work, Feng et al.17 have also screened a NCI compound collection containing 2526 compounds against stationary phase B. burgdorferi. Out of their reported 30 active hits, 11 compounds (Table 2) match with our 101 hit compounds of our primary screen. These 11 compounds in Table 2 are different from the 12 compounds (Table 1) we selected in this study. The 11 compounds (Table 2) which are already reported by Feng et al.17 are NSC267229, NSC267461, NSC345647, NSC637578, NSC82151, NSC143491, NSC70845, NSC258812, NSC311153, NSC3053 and NSC136044. All these compounds showed more than 90% inhibition of Borrelia growth in our screening as shown in Table 2. We have screened 558 more chemical compounds compared with the study published by Feng et al. Of those, 381 compounds were from the Structural Diversity Set, 11 from the Mechanistic Diversity Set, 113 from the Natural Products Set and we are the first to report screening the 53 compounds from the Challenge Set. Finding the same compounds validates both screens. The 12 selected compounds in our study have not been reported previously. We identified novel compounds using our highly sensitive BacTiter-Glo (Promega Corporation) Assay.

Table 2 Percentage inhibition Borrelia growth of our compounds identified from our screening similar to reported list of compounds earlier
Full size table

This study identified compounds with diverse structural features and functional activity. Reports on the antimicrobial activity of these compounds against other bacteria, viruses and fungi are available which can help in determining and establishing the therapeutic potential of these agents. Moreover, these compounds could serve as leads for the design of new derivatives and serve to guide medicinal chemistry efforts. In addition, the scope of targets exploited by these agents is very broad. Though the exact mode of action is not known in Borrelia, for many of these compounds, their mechanism is well-established in other microbes. They act by targeting cellular membrane, cellular organelles, metabolic processes and genetic machinery among other means.12, 23, 24 The current study, therefore, provides a means to identify novel therapeutic combinations for the treatment of the disease. Since the candidate molecules have been identified by a highly sensitive and efficient three-step screening method against both actively dividing and stationary phases of the spirochete, the identified molecules could be directly advanced for further in vivo analysis. In conclusion, the current report could inform the design and development of newer, improved therapies against Lyme disease.