Bisphosphonates for delivering drugs to bone

1.1. Bortezomib

Bortezomib (Btz, Velcade^®) is a drug used to treat patients with MM and mantle cell lymphoma (Jin et al., 2018; Manasanch & Orlowski, 2017; Robak et al., 2015), and its ability to treat other types of cancer is being tested in clinical trials (clinicaltrials.gov, 2019). Btz is a very specific 26S proteasome inhibitor, which reversibly but strongly binds the threonine proteases of the 20S subunit, enabling inhibition of the hydrolyzing catalytic activities of the 26S proteasome (Painuly & Kumar, 2013). Mechanistically, Btz directly inhibits proteasomal degradation of the large amounts of immunoglobulin produced by myeloma cells. This allows the accumulation of these useless/ineffective proteins, leading to endoplasmic reticulum stress, activation of the unfolded protein response, and eventually to myeloma cell apoptosis (Brewer & Diehl, 2000; Manasanch & Orlowski, 2017; Obeng, Carlson, Gutman, Harrington, Lee & Boise, 2006; Zinszner et al., 1998). Btz may also inhibit NF-κB signaling by protecting IκBα from degradation, thus preventing its pro-survival activity and promoting myeloma cell apoptosis (Roy, Sarkar & Basak, 2018; Vrabel, Pour & Sevcikova, 2019). However, other studies have reported that instead of inhibiting IκBα degradation, Btz can decrease the expression of IκBα, resulting in NF-κB activation in various tumor cell lines, including MM (Hideshima et al., 2009; Li et al., 2010). Btz also inhibits osteoclast formation by modulating p38, activator protein-1, and NF-κB signaling pathways (von Metzler et al., 2007; Zavrski et al., 2005), and promotes osteoblast differentiation by stabilizing the key osteoblast transcription factor, Runx2. Therefore, Btz also suppresses bone resorption and stimulates bone formation, which can ameliorate the abnormal bone homeostasis caused by MM (Giuliani et al., 2007; Pennisi, Li, Ling, Khan, Zangari & Yaccoby, 2009). These complementary functions of Btz on myeloma cells and bone cells place it as an ideal drug for treating patients with MM. Btz is administrated intravenously or subcutaneously. Similar to other anti-cancer drugs given by systemic administration, the off-target effects of Btz on non-skeletal tissues, mainly peripheral neuropathy (Cavaletti et al., 2007) and thrombocytopenia (Lonial et al., 2005; Murai et al., 2014; Shi et al., 2014), have restricted the amounts of the drug that can be administered. More importantly, patients develop inherent or acquired Btz resistance (BR), which limits its clinical efficacy (Dispenzieri, Jacobus, Vesole, Callandar, Fonseca & Greipp, 2010; Kuhn et al., 2012; Murray, Auger & Bowles, 2014; Richardson et al., 2003; Richardson et al., 2005; Ruschak, Slassi, Kay & Schimmer, 2011). The development of BR involves multiple mechanisms, such as activation of chemo-resistance pathways, immunoproteasomes, and mutations in proteasome subunits (Niewerth, Jansen, Assaraf, Zweegman, Kaspers & Cloos, 2015). Agents targeting these molecular pathways have been tested in preclinical studies using BR myeloma cell lines and in a set of CD138⁺ cells obtained from BM of patients with relapsed/refractory myeloma (Chauhan et al., 2010; Richardson et al., 2014; Zangari & Suva, 2016).

1.2. Autophagy and its inhibitors, Chloroquine and Hydroxychloroquine, in treatment of multiple myeloma

Recent studies report that autophagy may play a role in drug resistance in cancer patients (Sui et al., 2013) including MM (Abdel Malek et al., 2015; Hamouda et al., 2014; Zhang et al., 2015). Autophagy is a lysosome-dependent protein degradation process. Autophagy degrades unnecessary or dysfunctional cellular components and releases nucleotides, amino acids or fatty acids from them (Yun et al., 2017) as a source of energy to ensure cell survival under critical conditions, such as hypoxia and starvation (Verbaanderd et al., 2017). However, extensive autophagy may also lead to cell death (Tsujimoto & Shimizu, 2005). In myeloma cells, massive amounts of immunoglobulin that are not degraded because of Btz-mediated proteasome inhibition are proposed to serve as an energy source via autophagy, thereby exerting a pro-survival effect under drug treatment (Attar-Schneider, Drucker, Zismanov, Tartakover-Matalon, Rashid & Lishner, 2012; Auner & Cenci, 2015; Dykstra, Allen, Born, Tong & Holstein, 2015; Hoang, Benavides, Shi, Frost & Lichtenstein, 2009; Zeng, Chen, Zhao & Cui, 2012). Thus, inhibition of autophagy may kill drug-resistant myeloma cells by limiting energy supply.

Chloroquine (CQ) and Hydroxychloroquine (HCQ) are FDA-approved autophagy inhibitors and have been used to treat patients with malarial infections (CQ), systemic lupus erythematosus and rheumatoid arthritis (HCQ) for decades. Recent reports indicate that both CQ and HCQ reduce bone resorption by inhibiting osteoclast formation (Both et al., 2018; Xiu et al., 2014). They have been extensively studied both in vitro and in vivo in various cancer types. Results show that CQ or HCQ effectively increase the efficacy and limit drug resistance of standard anti-cancer therapies (Verbaanderd et al., 2017). Recently, 4 clinical studies (ClinicalTrials.gov Identifier: NCT01438177; NCT00568880; NCT01396200; NCT01689987 (Verbaanderd et al., 2017)) have investigated the use of CQ or HCQ in combination with other anti-MM drugs in patients with relapsed/refractory myeloma. A phase I clinical trial (NCT00568880) led by Dr. Dan Vogl (Vogl et al., 2014) and a phase II trial (NCT01438177) led by Dr. Amitabha Mazumder (Montanari, Lu, Marcus, Saran, Malankar & Mazumder, 2014) examined the effects of a combination of HCQ and Btz in 25 cases with refractory or relapsed MM, or CQ, Btz, and Cyclophosphamide in 11 similar patients. Both trials report promising results and demonstrated that CQ or HCQ was able to partially restore Btz sensitivity (Montanari, Lu, Marcus, Saran, Malankar & Mazumder, 2014; Vogl et al., 2014). However, some serious adverse events, including hypoglycemia (Goyal & Bordia, 1995; Prabha, 1996), BM suppression (Nagaratnam, Chetiyawardana & Rajiyah, 1978), cardiomyopathy (Costedoat-Chalumeau et al., 2007) and irreversible retinal toxicity (Browning, 2014) have been associated with long-term treatment with CQ or HCQ. Some subjects suffered from the toxic side effects of the combined regimen (Montanari, Lu, Marcus, Saran, Malankar & Mazumder, 2014) that led some of them to stop therapy (Vogl et al., 2014). Low doses of CQ or HCQ in combination with Btz and/or other drugs limited by the systemic adverse effects have been proposed to be responsible for hindering achievement of more robust clinical responses (Scott et al., 2017; Vogl et al., 2014).

1.3. Preclinical studies using bone-targeted strategies to deliver Bortezomib or other drugs to bone marrow in mouse models of MM

Given that MM primarily originates in the BM and destroys bone, selective delivery of effective anticancer drugs such as Btz specifically to bone has been a focus of myeloma research. Bone targeting should reduce the toxic side effects arising from systemic distribution of Btz and make dose escalation possible in order to improve therapeutic outcomes. In 2010, Reinholz et al. linked the chemotherapeutic drug arabinocytidine-5′-phosphate to a bisphosphonate (BP) drug, etidronate, and generated a series of bone-targeted anti-cancer compounds, of which the leading candidate was named MBC-11. MBC-11 was designed to target both osteoclasts (etidronate) and tumor cells (arabinocytidine-5′-phosphate), and it reduced bone metastases and bone tumor burden in mice-bearing 4T1/luc mouse breast cancer cells following mammary gland inoculation. It inhibited bone destruction and increased overall survival in mice bearing KAS-6/1-MIP1α human myeloma cells following tail vein injection (Reinholz et al., 2010). Subsequently, MBC-11 advanced to a phase I oncology clinical trial (NCT02673060) in 2014 for patients with cancer-induced bone disease and achieved partial efficacy (Zinnen, Rowinsky, Alexandrov, Plekhova, Roudas & Karpeisky, 2017; Zinnen, Karpeisky, Von Hoff, Plekhova & Alexandrov, 2019). In 2013, Agyin et al. synthesized a series of BP (alendronate)-proteasome inhibitor conjugates and showed that they had similar effects as their non-bone-targeted counterparts to kill MM cells in vitro (Agyin, Santhamma & Roy, 2013). In 2014, Swami et al. engineered bone-homing polymeric nanoparticles (NPs), BP (alendronate)-conjugated polymer PLGA-b-PEG-Ald, to load Btz and demonstrated that the bone-targeted NPs could bind to bone fragments in vitro and retain and accumulate in bone in vivo. Pretreatment of Btz loaded into bone-targeted NPs enhanced survival and decreased tumor burden in mice bearing Luc+/GFP+ MM1S human myeloma cells (Swami et al., 2014). In 2018, 2 studies reported the synthesis of bone-targeted nanoparticles (NPs), loaded Btz with a pH-sensitive aryl boric acid ester linkage that offers a pH-responsive drug release mechanism (Wang, Cai, et al., 2018; Zhu et al., 2018). In both studies, in vitro release assays showed that only 20-30% of Btz was released from the conjugated drugs in neutral buffer (pH 7.4), while under acidic conditions of around pH 5.0, the release rate of Btz was faster, delivering as much as 60-70% of the drug load. In mice with intra-tibial administration of MDA-MB-231 cells, bone-targeted Btz significantly reduced tumor burden and bone destruction relative to free Btz-treated mice.

Although some advances were achieved by using the above bone-targeted agents to treat MM or metastatic bone tumors, the possibly reduced systemic adverse effects of these drugs, especially the Btz derived compositions, have not been formally investigated. Furthermore, most studies have used an antiresorptive BP, such as etidronate (Reinholz et al., 2010; Zinnen, Rowinsky, Alexandrov, Plekhova, Roudas & Karpeisky, 2017) or alendronate (Agyin, Santhamma & Roy, 2013; Swami et al., 2014; Zhu et al., 2018). With this approach, it can be difficult to distinguish the pharmacological effect of the BP component from that of released Btz. Recently, we linked Btz to a bisphosphonate (BP) (Figure 1B) that binds avidly to bone, but is not antiresorptive, using a novel covalently bonded linker to generate a BP-linked Btz (named BP-Btz) conjugate and demonstrated that BP-Btz, but not Btz, bound to bone slices and inhibited the growth of myeloma cells in vitro. BP-Btz more effectively reduced tumor burden and bone loss than Btz in the 5TGM1 mouse model of MM (Wang, Xiao et al., 2018). In particular, we also demonstrated that BP-Btz generated significantly less systemic adverse effects, such as thrombocytopenia (Wang, Xiao, et al., 2018) and peripheral neuropathy (Wang et al., 2020), compared with Btz alone.

BPs bind preferentially, but not solely, to skeletal sites with high bone turnover (Russell, Watts, Ebetino & Rogers, 2008). Thus, some BP-conjugates which might be delivered to other skeletal sites could cause bone side effects. For example, BP-conjugates that utilize BPs with significant antiresorptive function (Agyin, Santhamma & Roy, 2013; Reinholz et al., 2010; Zinnen, Karpeisky, Von Hoff, Plekhova & Alexandrov, 2019) could in theory exhibit adverse effects that have been associated with other bisphosphonates, including osteonecrosis of the jaw, severe suppression of bone turnover, and atypical subtrochanteric femoral fractures (Abrahamsen, 2010; Keller, 2014). Therefore, in our conjugate approach, we designed BP moieties, which could bind avidly to the bone, but lack significant antiresorptive activity, to limit any chance of side effects and to clarify the source of pharmacological effect in our studies. Furthermore, even the effects of delivery to lower bone turnover skeletal sites would likely be relatively minimal, since less turnover involving osteoclast activity would likely lead to a slower release of the active warhead at those sites. Regarding the free drug, Btz promotes osteoblast differentiation, and reduces osteoclast formation (Giuliani et al., 2007; Uy et al., 2007), while chloroquine (Xiu et al., 2014) or hydroxychloroquine (Both et al., 2018) reduces osteoclastogenesis, which are features that may be beneficial to MM therapy, since a drastic loss of bone is observed in patients with this cancer. Although 99m-technetium bone scans indicate a high preference for bisphosphonates to localize at sites of higher turnover where their primary effects are likely to occur, we do not currently have data to indicate how much BP or free drug will be delivered to bone sites with lower bone turnover and what side effects this might lead to. Isotope labeling and related pharmacokinetic and pharmacodynamic studies will help to evaluate these important questions.

1.4. Effects of BP-Btz and BP-CQ or BP-HCQ on Bortezomib-resistant MM

As mentioned above, previous studies suggest that Btz resistance might be attributed to autophagy (Abdel Malek et al., 2015; Hamouda et al., 2014; Zhang et al., 2015), and the autophagy inhibitors, CQ or HCQ, were able to improve sensitivity to Btz or other anti-MM drugs to relapsed/refractory MM in clinical studies (Montanari, Lu, Marcus, Saran, Malankar & Mazumder, 2014; Vogl et al., 2014). However, the efficacy was limited due to several systemic adverse effects, hindering the use of higher doses of CQ or HCQ or Btz (Vogl et al., 2014). In view of the findings that targeting Btz to bone successfully improved its efficacy and ameliorated off-target adverse effects (Wang, Xiao et al., 2018), we also generated a bone-targeted CQ (BP-CQ) and a bone-targeted HCQ (BP-HCQ) by conjugating them to a bisphosphonate using linker chemistry similar to that for BP-Btz (Figure 1B) (Boeckman, Boyce, Xiao, YAO & Ebetino, 2017; Xing et al., 2020; Yao et al., 2016). Our rationale is that delivering Bortezomib or chloroquine conjugates to bone using a BP with high bone affinity, but no significant antiresorptive effects, could result in high and persistent local drug concentrations at sites of bone diseases, thereby killing more myeloma cells and preventing bone loss, with fewer adverse effects.

In our initial study, we examined if BP-Btz plus BP-CQ or BP-HCQ could more effectively kill BR myeloma cells than individual drug alone in vitro. We used a BR U266 human myeloma cell line (Mitra et al., 2016) and found that they were resistant to both Btz and BP-Btz to a similar extent (Figure 1C&D). Treatment of the BR U266 human myeloma cell line with 6 nM Btz or 30 μM CQ for 72 hrs killed 40% of the cells, while Btz+CQ killed 70%, and BP-Btz and BP-CQ had similar effects in vitro (Figure 1E&F). We also demonstrated that the equimolar concentration of the BP, including the linker structure that is likely to be generated on release of Btz, had no greater effect on myeloma cells than on vehicle-treated cells (Figure 1F). We then tested the effects of BP-Btz plus BP-CQ on primary myeloma cells that were isolated from BM of 3 patients with relapsed/refractory MM (Figure 1G). We used the same doses of Btz (6 nM) and CQ (30 μM) for cells from patient #1 and #2 and found that BP-Btz or BP-CQ alone killed 20-30% of myeloma cells, while a combination of these killed 50%. To determine if increasing the dose of Btz or CQ could achieve a better effect, we treated cells from patient #3 with high doses of Btz (12 nM) and CQ (40 μM) and found that they had a similar killing effect as the lower doses. We also examined the effects of BP-Btz and BP-CQ on CD138⁺ myeloma cells from 2 newly diagnosed MM patients. Cells from one patient responded well to Btz, with 6 nM killing 65%, while 30 μM BP-CQ killed 70%, and the combination killed 90% of the patient’s myeloma cells (Figure 1H). However, cells from another patient responded poorly to BP-Btz, with 6 nM BP-Btz killing 15%, while 20 μM BP-CQ or BP-HCQ killed 25%, and the combination killed 50% of the patient’s myeloma cells (Figure 1H), suggesting that myeloma cells from the second case may already be Btz-resistant. These data that have been presented in several international conferences (Boyce, 2019; Tao et al., 2018) and discussed in our recent review paper (Xing et al., 2020) suggest that combined BP-Btz and BP-CQ or BP-HCQ therapy may overcome BR myeloma and represent a new treatment for relapsed/refractory MM patients or newly diagnosed MM patients who are inherently BR.

Studies in the literature have estimated the local pH under the tight sealing zone of the active osteoclast on the bone surface. Mechanistically, it is accepted that the pH in the resorption lacunae is acidic to account for rapid dissolution of bone mineral mediated by the cell. Silver et al (Silver, Murrills & Etherington, 1988) reported that the local pH under the active osteoclast sealing zone is as low as 4.5. A more recent paper reported a pH range of 4-6 in the resorption pit (Kowada et al., 2011). Thus, we designed the linkage in BP-Btz to be acid sensitive. The in vitro experiments were performed at physiological pH (around 7.2), and we did not observe a significant difference between the effect of free drug and bone-targeted drug on inhibition of MM cell growth. Since free drug appears to be released primarily at an acidic pH, we speculate that the free drug is released under the osteoclast or within lysosomes or other cell organelles where the pH is low after the entire conjugate enters the cell. These release mechanisms could be explored further using fluorescent- or isotope-labeled BPs and BP-Btz.

1.5. Conclusion

Despite current clinical treatments for MM, including Btz, showing remarkable efficacy, drug resistance and off-target effects have restricted the use of Btz clinically. Several clinical studies have also shown some promising results of the autophagy inhibitors, CQ and HCQ, to overcome Btz resistance. However, the side effects of CQ or HCQ limit the use of more effective doses, hindering achievement of more robust clinical responses. Development of novel bone-targeted therapeutic agents in recent years has significantly improved efficacy in preclinical studies in the treatment of MM. Although some critical information, such as drug release kinetics, possible synergies of BP-drug payload interactions, mechanisms of action, tissue distribution, and toxicity, still need to be gathered, the BP conjugates represent a promising therapeutic strategy.

2.1. Current antimicrobial treatment for osteomyelitis

A number of Gram-positive and Gram-negative bacteria, as well as fungi and mycobacteria, can cause osteomyelitis, but by far the most common organism implicated in bone and joint infections is Staphylococcus aureus (SA), both methicillin-susceptible (MSSA) and methicillin–resistant (MRSA) (Dubost et al., 2014; Ferrand et al., 2016). For jawbone infections, Aggregatibacter actinomycetemcomitans (Aa) is a key organism (Stoodley et al., 2011). The standard of care for bone and joint infections usually requires systemic administration of antibiotics, typically vancomycin for MRSA and multidrug-resistant MRSA strains, and fluoroquinolones (ciprofloxacin, moxifloxacin etc.) for gram-negative pathogens (Fraimow, 2009). For acute infections, intravenous antibiotics are generally prescribed for 2 – 6 weeks. Prolonged courses of oral antibiotics may follow for chronic infections, or infections associated with retained implanted hardware. Both for acute and chronic infections, these extended courses of therapy can lead to drug-related adverse events in a significant percentage of patients – 15% in one estimate for a cohort treated for infections with MSSA. This is particularly an issue with vancomycin where nephrotoxicity can occur in as many as 43% of patients (Carreno, Kenney & Lomaestro, 2014). In general, these antibiotic treatments have poor bone pharmacokinetics and limited osseous absorption in vivo (Fong, Ledbetter, Vandenbroucke, Simbul & Rahm, 1986; Kim, Kim & Oh, 2014). Systemic toxicity or adverse effects, and resistance, are clinical issues associated with prolonged dosing regimens.

The inadequate efficacy of current antimicrobial treatments for osteomyelitis has been ascribed to the limited access of systemically administered antibiotics to sites where causative bacteria can reside as biofilms on bone surfaces, even surfaces deep within the osteocytic canalicular network (de Mesy Bently et al., 2017). Sedghizadeh et al. recently demonstrated that osteomyelitis pathogens can invade bone, establish chronic biofilms on bone surfaces, and directly destroy and resorb bone without participation of host immunity or osteoclastogenesis (Junka et al., 2017). These observations add to the difficulty in treating osteomyelitis biofilms, and to the urgency of resolving an unmet medical need in this indication. Therefore, any improvement in bone bioavailability of therapeutic antibiotics would be a significant advance in treating osteomyelitis.

Local antibiotic delivery to treat bone and joint infections can be attempted at the time of surgical debridement using antibiotic-impregnated beads or antibiotic-impregnated cement (Uskokovic, 2015). However, currently marketed beads are not bio-absorbable; therefore, a second procedure is required to remove them. They also tend to release antibiotics in an initial burst pattern that quickly depletes the bulk of the drug from the carrier beads, followed by a slow release of lower concentrations that may not be adequate to control infection and may foster development of resistance. While there have been recent advances in the use of bio-stimuli responsive release strategies (Lavrador, Gaspar & Mano, 2018; Lin, Caldwell, Bhaduri, Goel & Agarwal, 2017), these concerns limit the usefulness of this approach in the majority of bone and joint infections (Bozhkova, Novokshonova & Konev, 2015; Uskokovic, 2015). In the realm of prosthetic joint infection (PJI), antibiotic-impregnated cement is used commonly at the time of first debridement of an infected implant to improve control of the infection. A recent study showed that this can be efficacious in the control of the infection in two-stage hip revision surgeries (Staats et al., 2017). Different cement substrates are available and several antibiotics are used in these treatments, with vancomycin being the most common. Variable anti-bacterial loading and cement characteristics can lead to changes in mechanical properties (Slane, Gietman & Squire, 2018) of the cement implant as well as varying anti-microbial elution rates (Boelch et al., 2018). Concerns about prolonged sub-therapeutic antibiotic concentrations and selection of resistant organisms also apply to cement (Corona, Espinal, Rodriguez-Pardo, Pigrau, Larrosa & Flores, 2014; Walker, Baker, Holleyman & Deehan, 2017).

2.2. Bone-targeted antimicrobial therapy for osteomyelitis

BP drugs have a high affinity for bone and preferentially accumulate at sites of active bone disease or biofilm infection, resorption, and remodeling (Cheong et al., 2014; Ebetino et al., 2011). BPs also penetrate the canalicular network to osteocytes and osteocytic lacunae where no blood flow exists and where S. aureus biofilm organisms are known to embed (de Mesy Bentley et al., 2017). In addition, the BP class of drugs has a long track-record of clinical use and a good safety profile relative to most drug classes, and have thus been utilized for pharmacotherapy in several bone diseases, such as osteoporosis, multiple myeloma, metastatic cancer to bone, and Paget’s disease of bone. To exploit this BP affinity for bone, we are investigating a “target and release” chemistry approach for treating bone infections which involves delivery of antibiotics to the hydroxyapatite (HA) mineral of the skeleton via conjugation to BPs, utilizing serum-stable drug-BP linkers that metabolize and release the antibiotic at bone surfaces and within the canalicular network. Here, we have also utilized pharmacologically “inert” BPs for this purpose to avoid confounding antiresorptive effects from the carrier BP.

We recently reported impressive in vivo efficacy in animal studies with a novel bisphosphonate-carbamate-ciprofloxacin (BCC) conjugate, BV600022 (Sedghizadeh et al., 2017b) (Figure 2A). Quantitative determination of the colony forming units (CFU) of bacteria from resected peri-prosthetic osteomyelitis tissue showed that a single dose of 10 mg/kg of conjugate gave ~2 log reduction of CFU or 99% bacterial killing efficacy and nearly an order of magnitude greater activity than ciprofloxacin alone given in multiple doses (10 mg/kg X 3) (Figure 2B). Furthermore, the effects of a single dose of 10 mg/kg of conjugate in vivo were significantly greater (p=0.0005, unpaired t-test) than the untreated control arm. Also, we synthesized and tested a non-cleavable (non-antibiotic releasing) amide-linked BP conjugate (BAC) BV600026 (Figure 2A), which was found to be nearly inactive against osteomyelitis biofilms grown on HA surfaces in vitro (Sedghizadeh et al., 2017b).

Additional in vitro assays of S. aureus and Aa (jawbone) osteomyelitis pathogens in both planktonic and biofilm assays have been performed with BCC treatment as well as with a few other novel BP-antibiotic conjugates (Figure 2A). A new fluoroquinolone carbamate conjugate, bisphosphonate-carbamate-moxifloxacin (BCX, BV600052), based on the later generation and more potent moxifloxacin fluoroquinolone, was compared to BCC in these assays (Sedghizadeh et al., 2017a; Sedghizadeh et al., 2019; Sedghizadeh et al., 2017b; Sedghizadeh et al., 2020). In addition, some of the most important oxazolidinones are currently used against gram-positive pathogens, including superbugs such as MRSA. These antibiotics are considered clinically as choices of last resort when other antibiotics have failed. We therefore also designed, synthesized, and tested two new BP-oxazolidinone conjugates, including a novel BP-carbamate-tedizolid conjugate (BCT, BV600037), and a related more rapidly cleavable BP-ester-tedizolid conjugate (BET, BV600039), for activity against S. aureus. Tedizolid is a recent FDA-approved oxazolidinone antibiotic indicated for the treatment of skin infections, particularly those caused by both methicillin-susceptible S. aureus and methicillin-resistant S. aureus pathogens, which cause the majority of clinical cases of skin infection (Ebetino, Sun, McKenna, Sadrerafi & Cherian, 2020). Since most osteomyelitis cases are also caused by similar Staphylococcal species, the application of tedizolid to bone infections could prove efficacious. An ester-linked conjugate is expected to have more rapid linker release kinetics than a carbamate-linked conjugate, and we projected it could be a good comparison for the identification of the optimal linkage. We also tested various unconjugated antibiotics, including vancomycin (V) and minocycline (M), which are used clinically in the setting of infectious bone disease, in addition to the parent drugs of these conjugates (ciprofloxacin (C), moxifloxacin (X) and tedizolid (T), Figure 2A) for direct comparison with our conjugates in preventing or eradicating planktonic and biofilm osteomyelitis pathogens in vitro and ex vivo.