This review is divided into subsections according to the three major antibiotic mechanisms of action: protein biosynthesis inhibition, bacterial enzyme inhibition and cell wall disruption. Information is provided on antibiotic profiles, usage in osteomyelitis, antimicrobial resistance, bone concentration of antibiotic, intracellular penetration, extracellular minimum inhibitory concentration (MIC) and intracellular effectiveness. For some antibiotics, enhanced drug delivery systems have been tested in intracellular models, which are also discussed. Only a few studies with preclinical in vivo data could be included in this review at the end of each section, before a concluding statement about the antibiotic group. To our knowledge, there are no published clinical data showing the intracellular activity of antibiotics in human osteomyelitis patients. To compare different studies, all concentrations were adapted to units of micrograms per milliliter (μg·mL−1) or μg·g−1 if no volume percentage could be found, and the intracellular effectiveness was measured by logarithmic colony-forming unit (CFU) reduction.

Protein biosynthesis inhibitorsLincosamide: clindamycin

Clindamycin is commonly used for the treatment of osteomyelitis due to its activity spectrum against gram-positive cocci (especially Staphylococcus, including erythromycin-resistant Staphylococcus, and Streptococcus), gram-negative cocci and intracellular bacteria (Chlamydia and Rickettsia species) and its good bone penetration properties.45,46 With some exceptions (i.e., Bordetella pertussis, Campylobacter, Chlamydia, Helicobacter, and Legionella species), clindamycin is not effective against gram-negative bacilli.46,47,48

The antibacterial effect of lincosamides is achieved by the inhibition of elongation during protein synthesis by interfering with peptide bond formation between the A- and P-sites of the tRNA of the 50 S subunit of bacterial ribosomes.48,49 This common binding site shared by multiple drugs can lead to cross-resistances, most frequently macrolide, lincosamide and streptogramin B (MLSb) cross resistance (see 4.1.4 MLSb resistance).

In osteomyelitis patients, the bone penetration of clindamycin as a monotherapy is rapid and high, with a bone:plasma ratio between 0.2-0.6 leading to bone concentrations between 1.4-9.6 μg·mL−1. However, cotreatment with rifampicin can reduce the clindamycin bone concentration to a subtherapeutic level.40,41,50,51,52

In vitro, clindamycin treatments against S. aureus showed a high variation in intracellular activity between <1-log to 5-log reduction if cells were treated immediately after the infection.43,53,54,55,56,57,58 No measurable reductions were reported with delayed treatment after 12 h54 or 7 days.55 Clindamycin showed limited effects against S. aureus SCVs, reducing the number of intracellular bacteria by 1-log after 72 h and extracellular bacteria by 2-log after 24 h in THP-1 macrophages.17 Furthermore, clindamycin was also reported to induce SCV formation.55 The concentration of clindamycin used in the studies varied between 1.3–20 μg·mL−1, but most commonly 4 μg·mL−1 was used,17,43,53,54,55,56,57,58 which is a concentration that might realistically be reached in the bone. The MICs for various S. aureus strains were between 0.004–8 μg·mL−1 at pH 7. However, acidic pH can increase the MIC fourfold or more to >38 μg·mL−1 compared to 4–8 μg·mL−1 in the same strain at pH 7.43 This indicates that the treatment efficacy in patients depends on the pH of the compartment where the antibiotic and bacteria interact. Therefore, increasing the intracellular pH could be a strategy to improve treatment efficacy. As an example, combining clindamycin with hydroxychloroquine in vitro resulted in an intracellular CFU reduction of 2.7-fold.43 Moreover, clindamycin incorporated in a calcium phosphate powder showed increased antibacterial activity compared to that of the standard formulation after 24 h and 48 h, suggesting that smart drug delivery systems have the potential to improve treatment effectiveness against intracellular bacteria in osteomyelitis.57

From the current in vitro data, no conclusion can be made regarding the effectiveness of clindamycin in treating intracellular S. aureus in osteomyelitis. Most studies showed limited effectiveness that relied on immediate treatment after the infection, which is typically not possible in a clinical setting. Therefore, clindamycin is likely to be ineffective against intracellular S. aureus infections in osteomyelitis.

Streptogramins: quinupristin/dalfopristin

Quinupristin/dalfopristin is a 30:70 combination of two Streptomyces-derived streptogramin antibiotics (streptogramin A and streptogramin B), which have a synergistic bactericidal effect. The combination is used against staphylococci, streptococci and Enterococcus faecium59,60 and inhibits protein biosynthesis at the 50 S subunit of bacterial ribosomes. Dalfopristin (streptogramin A) binds to the A- and P-sites of the peptidyl transfer center of the ribosome and changes the ribosomal conformation, enhancing the binding of quinupristin (streptogramin B) approximately 100-fold.59,61 Furthermore, dalfopristin hinders peptidyl transfer elongation by inhibiting tRNA binding and translation. Similar to other MLSB antibiotics, quinupristin obstructs the elongation and particularly the translocation and extension of proteins and causes the release of incomplete protein chains.61,62,63

Currently, quinupristin/dalfopristin (market name Synercid) has been approved only for skin and soft tissue infections,64 but it has been used off-label in osteomyelitis patients. A case report described the successful treatment of a patient with a vancomycin-resistant enterococcus (VRE) infection,65 and a phase I clinical trial concluded with the successful treatment of 32 out of 40 patients, defined as clinical cure or improvement.66

Another drug combination is pristinamycin, composed of pristinamycin IA (streptogramin B) and pristinamycin IIA (streptogramin A). It is effective against MRSA, erythromycin-resistant staphylococci and streptococci and has been used effectively against osteoarticular infections and osteomyelitis.67,68

In our search, only one in vitro study on the effect of quinupristin/dalfopristin against intracellular S. aureus SCVs was found, which reported an intermediate effectiveness of a 2-log reduction against intracellular SCVs compared to a 3-log reduction against extracellular SCVs. Remarkably, a sudden onset of the antibacterial effect after 72 h intracellularly and after 5 h extracellularly was noted, which did not increase further at later time points. Even though the MICs against all bacterial phenotypes were the same (0.5 μg·mL−1), the antibiotics were more effective against the WT and the revertant strains than against the SCV form in an experimental infection in THP-1 macrophages.17 Although quinupristin/dalfopristin is not commonly used to treat osteomyelitis and there are only limited data about its intracellular effectiveness, it seems to be a promising treatment option; however, further studies are needed.

Macrolides: azithromycin, telithromycin, erythromycin, spiramycin

Macrolides are antibiotics produced by gram-positive Actinomycetes and were discovered in the 1950s.69 A large number of macrolide antibiotics are now available for standard medical care and specifically as an alternative therapy for penicillin-allergic patients.70

Similar to penicillins, macrolides are mainly effective against gram-positive bacteria (e.g., staphylococci and streptococci) but are also effective against some gram-negative species (Neisseria gonorrhoeae, Haemophilus influenzae, Bordetella pertussis and Neisseria meningitis) and mycoplasma species.71,72

Macrolides inhibit bacterial protein synthesis by hindering tRNA attachment to the peptidyl transferase of the 50 S subunit of bacterial ribosomes, thus disrupting polypeptide chain elongation. The shared target with linezolid and streptogramins B can lead to MLSB resistance.49,72,73

Macrolides penetrate well into macrophages, with an intra- to extracellular ratio between 8.6–50.74,75 The bone concentration has a high variation, with bone:serum ratios between <0.05 and >3 and a median of 0.91.40 Azithromycin and telithromycin penetrate the bone best, with bone:plasma ratios of 2.5–6.3 and 1.5–2.6, while erythromycin and spiramycin show low bone penetration, with bone:plasma ratios of 0.18–0.28 and 0.047.40 The MICs of macrolides are pH sensitive, with up to a > 100-fold increase at pH 5 (0.5 μg·mL−1 for azithromycin, 0.06 μg·mL−1 for telithromycin) compared to those at pH 7 (512 μg·mL−1 for azithromycin, 4 μg·mL−1 for telithromycin), and their minimum bactericidal concentrations (MBCs) are much greater than their MICs, up to >30-fold higher (MBC: 2 μg·mL−1 compared to MIC: 0.06 μg·mL−1),74,75,76 leading to a high risk of inducing the development of persister phenotypes. Against intracellular S. aureus in macrophages, azithromycin and telithromycin showed less than a 1-log reduction to no measurable effect and only at very high dosages, although both antibiotics exhibited a dose-dependent extracellular effect.74,75 In osteoblasts, erythromycin showed some effectiveness when administered immediately but not after a 12 h delay.54 Since different antibiotics have been used in experiments with macrophages than those with osteoblasts and with the low number of studies reported, the difference in susceptibility of intracellular S. aureus to macrolides may be related to the specific drug tested rather than the cell type.

Even though macrolides seem to penetrate the bone and bone cells well, the little available evidence regarding intracellular S. aureus makes it difficult to discern if they are an effective treatment against intracellular S. aureus in an osteomyelitis setting.

MLSB resistance

MLSB resistance defines a cross-resistance between macrolides, lincosamides and streptogramin B due to a similar mode of action, targeting the bacterial 50 S subunit of the ribosomes, more specifically the 23 S ribosomal RNA segment at the peptidyl transferase center.77

Streptococci, enterococci and staphylococci are the main carriers of MLSB resistance determinants, but they can also be expressed by other gram-positive and gram-negative species.47 Resistance can be constitutive (cMLSB) (especially in staphylococci) or induced (iMLSB).78 There are three main mechanisms of resistance: (1) methylation (esp. streptococci and enterococci) or mutation (esp. E. coli) that prevents target binding to most MLSB antibiotics,47,79 (2) expression of efflux pumps to remove the antibiotic from the cell, which can be drug-specific,80,81 and 3) drug inactivation by esterases and phosphotransferases, which is a drug-specific mechanism that does not always lead to cross-resistance.82

Oxazolidinones: linezolid, tedizolid, radezolid

Oxazolidinones are a group of synthetic drugs with excellent oral bioavailability, making them easy for patients to use.

These bind to the 50 S subunit of the bacterial ribosome and inhibit the initiation phase of protein biosynthesis by preventing the formation of the initiation complex.83 Even though the binding site is closely related to MLSB cross-resistant antibiotics, to date, no cross-resistance has been observed.84,85 Resistances determinants are rare but include a point mutation (G2576T) in staphylococci and E. faecium83 and G244 methylation, which reduces antibiotic binding, as well as an ABC transporter in Streptococcus pneumoniae.86

Oxazolidinones are mostly used against gram-positive bacteria (esp. streptococci, enterococci, and S. aureus), particularly for vancomycin-resistant strains and MRSA, including in osteomyelitis.45,87,88,89,90,91,92,93

Linezolid penetrates well into the bone, with a bone:plasma ratio of 0.2–0.6, leading to bone concentrations of 4–9 μg·mL−1.40,41,53,94 The intracellular uptake varies among the compounds, with an intra- to extracellular ratio of 0.5 in THP-1 macrophages for linezolid and between 9.3–14.4 in other cell types (9.8 for osteoblasts) for radezolid.43,74,95

Radezolid is a newer compound that passed phase 2 clinical trials in 2008 and 2009 but has not been used to date for treating osteomyelitis. However, it has already been tested for its intracellular activity against S. aureus in one study in multiple cell types (including osteoblasts), compared to that of linezolid.95,96 The superiority of radezolid showed an up to 8-fold reduced MIC against S. aureus (0.25–2 μg·mL−1 for radezolid compared to 1–16 μg·mL−1 for linezolid), as well as an approximately 10-fold higher intracellular accumulation. Similar to linezolid, it shows a reduced intra- to extracellular activity ratio, which compared to that of linezolid is higher when measured as drug concentration but is comparable when dosages similar to the MIC are compared, with an overall bacterial reduction of approximately 1-log.95

The effectiveness of oxazolidinones seems to be dependent on the specific formulation, the host cell type examined, and the bacterial strain being targeted. For linezolid, two studies with THP-1 macrophages reported a maximal intracellular reduction of S. aureus of 1-log and only slightly greater effectiveness against extracellular bacteria, independent of the growth phenotype.17,74,95,97 In contrast, linezolid in osteoblasts caused an up to 3-log reduction in intracellular S. aureus levels, even when the treatment was delayed until 7 days after the infection. However, this effect was observed only for certain strains, and some results were statistically significant but showed only 1-2-log reductions.43,53,55,95,98,99 No difference in the induction of or effectiveness against SCVs has been reported for linezolid.17,53,55 The MICs of the observed strains were between 1–4 μg·mL−1, with higher values in an acidic environment and for the WT compared to a those with pH of 7 and against SCVs.17,31,43,