Molecular Targeted Therapy for the Bone Loss Secondary to Pyogenic Spondylodiscitis Using Medications for Osteoporosis: A Literature Review

Pyogenic spondylodiscitis can cause severe osteolytic and destructive lesions in the spine. Elderly or immunocompromised individuals are particularly susceptible to infectious diseases; specifically, infections in the spine can impair the ability of the spine to support the trunk, causing patients to be bedridden, which can also severely affect the physical condition of patients. Although treatments for osteoporosis have been well studied, treatments for bone loss secondary to infection remain to be elucidated because they have pathological manifestations that are similar to but distinct from those of osteoporosis. Recently, we encountered a patient with severely osteolytic pyogenic spondylodiscitis who was treated with romosozumab and exhibited enhanced bone formation. Romosozumab stimulated canonical Wnt/β-catenin signaling, causing robust bone formation and the inhibition of bone resorption, which exceeded the bone loss secondary to infection. Bone loss due to infections involves the suppression of osteoblastogenesis by osteoblast apoptosis, which is induced by the nuclear factor-κB and mitogen-activated protein kinase pathways, and osteoclastogenesis with the receptor activator of the nuclear factor-κB ligand-receptor combination and subsequent activation of the nuclear factor of activated T cells cytoplasmic 1 and c-Fos. In this study, we review and discuss the molecular mechanisms of bone loss secondary to infection and analyze the efficacy of the medications for osteoporosis, focusing on romosozumab, teriparatide, denosumab, and bisphosphonates, in treating this pathological condition.


Introduction
Pyogenic spondylodiscitis is an infection that occurs at an intervertebral disc and the adjacent vertebrae, and it can lead to bone loss and subsequent instability of the spine. Elderly and immunocompromised people are particularly susceptible to spondylodiscitis, have a high risk of being bedridden, and experience a deterioration of their physical condition [1]. Moreover, the prevalence of spondylodiscitis has increased over the last 20 years [1]. The prevalence of spondylodiscitis increased from 2.2/100,000 to 5.8/100,000 within 14 years in Denmark [2] and from 5.3/100,000 to 7.4/100,000 within 3 years in Japan [1,3]. Regarding the treatment, antibiotics can be effective in fighting microbes; however, they have no reparative effects on destroyed bone lesions. Notably, bone loss may have fewer treatment options than infection and tends to become problematic. A possible therapeutic measure is the administration of medications for osteoporosis. These fewer treatment options than infection and tends to become problematic. A possible therapeutic measure is the administration of medications for osteoporosis. These medications modulate bone formation and/or resorption, and the effects vary by the type and characteristics of the medication. However, currently, there is no consensus on which agent is appropriate for these patients.
Recently, we managed the case of a patient with postoperative pyogenic osteolytic spondylodiscitis who was treated with romosozumab, a monoclonal antibody against sclerostin that stimulates canonical Wnt/β-catenin signaling and exhibited enhanced bone formation following treatment. Briefly, the elderly woman with diffuse idiopathic skeletal hyperostosis and untreated severe osteoporosis had a lumbar vertebral fracture at the L3 level. She was administered teriparatide, calcium, and alfacalcidol, but these medications were stopped after a short period because exanthema occurred. She underwent surgical fixation of the spine from the T12 to the L5 level using a percutaneous pedicle screw system but developed a surgical-site infection in the L3 vertebral body and L2/3 and L3/4 intervertebral discs six weeks after the surgery. Severe osteolysis was evident at the surgical site on the CT images ( Figure 1). We started the administration of romosozumab immediately after the implant removal and thorough irrigation and continued it in parallel with the treatment of infection using antibiotics. Although serological inflammation persisted for six weeks after surgery, the CT images revealed robust bone formation in the destroyed bone lesions at the L3 vertebra and bony spur bridging at the L2/3 and L3/4 disc levels at two months after operation ( Figure 1).  Although our treatment was successful in this case, a decision regarding treatment for bone loss secondary to pyogenic spondylodiscitis is still difficult and controversial. Therefore, we review the molecular signaling pathways involved in bone loss secondary to infection and discuss the potential role of the currently available osteoporosis medications, including romosozumab, teriparatide, denosumab, and bisphosphonates (BPs), in the treatment of this intractable pathological condition.

The Peculiarity of Bone Loss Secondary to Pyogenic Spondylodiscitis
The peculiarity of the spine versus other regions regarding the locus of infection is a high propensity of patients losing ambulatory competencies. Bone loss in pyogenic spondylodiscitis patients is usually rapidly progressive and highly destructive [4]; therefore, patients tend to become bedridden and cannot recover ambulatory function without restoring the support and stability of the spine. Moreover, major bone loss is due to not only the causal bacteria but also the disuse of the spine owing to bed rest, brace immobilization, and limited sunlight exposure from long hospitalization for intravenous antibiotic administration [5]. Another factor that may lead to abrogating ambulatory competence is neurological deficits. Due to its function as a container of the spinal cord and cauda equina, the disruption of the spinal column by bone loss may result in the impairment of neurological function, which is highly relevant to the patient's ambulatory competence. The prevalence of neurological deficits, including radiculopathy, paralysis involving the trunk, limb, bladder, and rectum, and sensational disorder, is reported to be 10-50% depending on the study, and although its incidence is less than 10%, there is a chance of their progression to permanent deficits [6][7][8][9]. Fast-acting medications for bone loss are ideal for restoring the structure and function of the spine before the general body condition and neurological function deteriorates.

Molecular Mechanisms of Bone Resorption Secondary to Infection
Bone resorption driven by osteoclastogenesis is the primary factor responsible for bone loss secondary to pyogenic osteomyelitis. Osteoclastogenesis is initiated by proinflammatory cytokines secreted by lymphocytes and macrophages, such as tumor necrosis factor (TNF)-α, interleukin (IL)-1β, and IL-6, inducing the binding of the receptor activator of nuclear factor-κB (NF-κB) ligand (RANKL) to RANK [10]. TNF-α promotes the development of osteoclast precursors to express RANK and the development of stromal cells, osteoblasts, and activated T cells to express RANKL [10][11][12]. IL-1β also facilitates the expression of RANKL and causes osteoclast precursors to differentiate into mature osteoclasts [10,13,14]. IL-6 induces the expression of RANKL in osteoblasts and stromal cells [10,15]. Subsequently, the RANK-RANKL combination causes TNF receptor-associated factor 6 (TRAF-6) to activate NF-κB and MAPKs, followed by the activation of the transcription factors (TFs) nuclear factor of activated T cells cytoplasmic 1 (NFATc1) and c-Fos in osteoclast precursors [10,16]. These factors upregulate tartrate-resistant acid phosphatase (TRAP), matrix metalloproteinase 9 (MMP9), calcitonin receptor (CTR), and cathepsin K (CTSK), which differentiate the multinucleated osteoclasts with the assistance of macrophage-colonystimulating factor, which stimulates the proliferation and survival of osteoclast precursors, leading to bone resorption [10,14,16]. Another important transcription factor that induces proinflammatory cytokines is signal transducer and activator of transcription (STAT) [10]. STAT3 has been demonstrated to play an important role in the production of IL-1β, IL-6, and nitrogen oxide and the subsequent differentiation of cultured preosteoclasts to mature osteoclasts induced by lipopolysaccharide (LPS) [10,[17][18][19]. Although the patient in our study had an infection caused by methicillin-resistant coagulase-negative staphylococci, Gram-negative bacteria also commonly cause pyogenic spondylodiscitis. LPS is a component of the outer membrane of Gram-negative bacteria and has been shown to play important roles in bone resorption secondary to pyogenic osteomyelitis [20]. LPS activates Toll-like receptor 4 (TLR4) and stimulates various signaling pathways, including the NF-κB, MAPK, and STAT3 pathways [10,21]. The effects of LPS on bone loss include not only the stimulation of bone resorption by osteoclasts but also the inhibition of osteoblast differentiation [22,23]. LPS suppresses the expression of runt-related transcription factor (Runx) 2, osterix, and activating transcription factor 4 (ATF4), consequently inhibiting the differentiation of cultured osteoblasts [22]. The NF-κB and MAPK pathways also contribute to osteoblast apoptosis [23]. Furthermore, noncoding microRNAs (miRNAs) are single-stranded RNAs that have 18-24 nucleotides and a hairpin structure [24]. miRNAs have been confirmed to function as gene regulators posttranscriptionally [24]. miRNA-34c is upregulated by LPS stimulation, and its overexpression has been demonstrated to suppress the expression of osteogenic gene markers, such as alkaline phosphatase, Runx2, osteopontin (OPN), and bone morphogenetic protein (BMP) 2 [23] (see Figure 2). LPS on bone loss include not only the stimulation of bone resorption by osteoclasts but also the inhibition of osteoblast differentiation [22,23]. LPS suppresses the expression of runt-related transcription factor (Runx) 2, osterix, and activating transcription factor 4 (ATF4), consequently inhibiting the differentiation of cultured osteoblasts [22]. The NF-κB and MAPK pathways also contribute to osteoblast apoptosis [23]. Furthermore, noncoding microRNAs (miRNAs) are single-stranded RNAs that have 18-24 nucleotides and a hairpin structure [24]. miRNAs have been confirmed to function as gene regulators posttranscriptionally [24]. miRNA-34c is upregulated by LPS stimulation, and its overexpression has been demonstrated to suppress the expression of osteogenic gene markers, such as alkaline phosphatase, Runx2, osteopontin (OPN), and bone morphogenetic protein (BMP) 2 [23] (see Figure 2).

Osteoporosis Medications as a Treatment for Bone Loss Secondary to Infection
Some medications used for osteoporosis are considered potential therapeutics for bone loss secondary to infection because these medications are used to treat bone loss in other pathological conditions. Increased osteoclastic activity is evident in many osteopenic disorders, including Paget's disease, lytic bone metastases, and rheumatoid arthritis, causing increased bone resorption and destruction [25]. Osteoporosis medications comprise anabolic drugs and antiresorptive drugs.
Major options for anabolic medications are teriparatide and romosozumab. Teriparatide has been reported to be effective in the treatment of medication-related osteonecrosis

Osteoporosis Medications as a Treatment for Bone Loss Secondary to Infection
Some medications used for osteoporosis are considered potential therapeutics for bone loss secondary to infection because these medications are used to treat bone loss in other pathological conditions. Increased osteoclastic activity is evident in many osteopenic disorders, including Paget's disease, lytic bone metastases, and rheumatoid arthritis, causing increased bone resorption and destruction [25]. Osteoporosis medications comprise anabolic drugs and antiresorptive drugs.
Major options for anabolic medications are teriparatide and romosozumab. Teriparatide has been reported to be effective in the treatment of medication-related osteonecro-sis of the jaw (MRONJ) [26], which has a pathological manifestation similar to that of pyogenic osteomyelitis. The pathological manifestation of MRONJ is considered to involve the antiresorptive effect of BPs or denosumab, a reduction in macrophages, and an increase in monocytes by the effect of BPs, local bacterial infection, inflammation, and necrosis [26,27]. However, there is increasing evidence that infection can be a histological hallmark of MRONJ [28][29][30][31][32]. Bacteria form biofilms to protect themselves from the host immune system and antibiotics [32]. Teriparatide has been suggested to promote osseous wound healing in cases of MRONJ [26,33]. When considering the use of teriparatide in the treatment of bone loss secondary to an infection, one concern is that teriparatide is known to promote bone resorption concomitant with bone formation, as shown by an elevation in the levels of serum N-telopeptide of type I collagen or collagen cross-linked C-telopeptide, although an anabolic window exists in cases of usual osteoporosis [33][34][35][36]. Romosozumab, a monoclonal antibody against sclerostin, seems to be a better choice because it has been reported to increase bone formation and decrease bone resorption [37][38][39][40][41].
The main players among antiresorptive drugs are a group of BPs and denosumab. Although the area of bone loss is systemic in the established stage, rheumatoid arthritis (RA) causes local periarticular bone loss in the preclinical stage [42], which leads to mild but similar features to those of bone infection. Many inflammatory cytokines involved in RA have osteoclastogenic effects [43], and they commonly accompany the bone loss secondary to an infection, as described in the previous section. In patients with RA, bone loss is treated with BPs or denosumab, which has been shown to significantly increase the bone mass index at the lumbar spine and hip [42]. These drugs have also been used in bone loss secondary to tumorigenesis, such as cancer metastasis or osteolytic primary bone tumors, for example, a giant-cell tumor of bone [44][45][46]. Although these medications reduce bone resorption, they are also known to suppress bone formation, resulting in decreased bone turnover [32]. Based on these facts, we will further introduce the currently available evidence for the individual drugs.

Anabolic Drug: Romosozumab
Romosozumab is a humanized monoclonal sclerostin antibody that sabotages sclerostin and inhibits the suppression of canonical Wnt/β-catenin (cWnt) signaling, resulting in osteoblastogenesis, bone formation, and the suppression of osteoclastogenesis [47][48][49]. Romosozumab prevents sclerostin from binding to the low-density lipoprotein receptorrelated protein (LRP)-5 or LRP-6, the receptors of Wnt ligands, increasing bone formation and decreasing bone resorption [50]. The Wnt ligands, without restriction by sclerostin, bind to a specific Frizzled family receptor and an LRP-5/6 coreceptor on the surface of the osteoblasts and activate Disheveled, which prevents glycogen synthase kinase 3β (GSK3β) from phosphorylating β-catenin, forming a complex with adenomatous polyposis coli (APC) and Axin [51]. Cytoplasmic β-catenin accumulates, escapes ubiquitination and degradation by proteasomes, translocates into the nucleus, and binds to the transcription factor T cell factor/lymphoid enhancer factor (TCF/LEF), upregulating the target gene expression [51,52]. cWnt signals are involved in the self-renewal, proliferation, and epithelial-mesenchymal transition of target cells [53] and affect fibroblast growth factor, Notch, and transforming growth factor-β signals [49,[54][55][56]. Specifically, in bone, cWnt signals promote the Runx2-dependent osteoblastic differentiation of mesenchymal stem cells or progenitor cells and have an anabolic effect [49,57,58]. The activation of cWnt signaling induced by Wnt3a or the overexpression of β-catenin/TCF4 also elevates BMP2 promoter activity and its mRNA levels [59]. This mechanism has been validated by confirming the existence of TCF/LEF response elements in the BMP2 promoter region using a site-directed mutagenesis approach [59]. Furthermore, cWnt signaling upregulates BMP2 and then upregulates Wnt7A/10B through BMP receptor type IA, resulting in the synergistic enhancement of osteoblastogenesis and bone formation [47,49,59]. Moreover, cWnt signaling causes antiresorptive effects by suppressing osteoclast differentiation at the precursor level [52]. cWnt signaling in osteoblasts elevates the expression of osteoprotegerin (OPG) and suppresses osteoclast differentiation [60,61]. In contrast, osteoclast precursor-specific β-catenin knockout mice showed enhanced osteoclast differentiation and osteopenia [52]. This result demonstrated that cWnt signaling also suppresses osteoclastogenesis at the precursor level in an OPG-independent manner [52]. In another study, early treatment with Wnt3a inactivated the crucial transcription factor NFATc1 in osteoclast precursors [60]. Although Wnt3a is known to activate cWnt signaling, that study revealed the involvement of noncanonical Wnt signaling in the Wnt3a inactivation of NFATc1 by demonstrating that the deletion of β-catenin does not block this inactivation [60]. In brief, Wnt3a activates cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA) signaling and suppresses osteoclast differentiation by phosphorylating NFATc1 and suppressing its nuclear localization [60]; thus, romosozumab may also affect the noncanonical signals induced by Wnt3a. In addition, PKA has also been shown to negatively regulate NF-kB nuclear localization; therefore, romosozumab may contribute to inhibiting apoptosis of osteoblasts (see Figure 2).
In our experience with this patient, the effects of romosozumab against bone loss secondary to infection presumably addressed the osteoclastogenesis with regard to the parts driven by NFATc1 activation, the role of OPG, and the role of β-catenin in suppressing osteoclast differentiation. However, romosozumab was also effective in modulating osteoblastogenesis and consequent bone formation in concert with BMP2 signaling, as well as by hindering the osteoblast apoptosis induced by the NF-κB pathways (see Figure 2). The patient exhibited enhanced bone formation at a severely osteolytic infection site as early as two months after implant removal and thorough irrigation. Importantly, serological inflammation was present for six weeks after the surgery, meaning that romosozumab strongly promoted bone formation, even in the presence of osteoclastogenesis and the suppression of osteoblastogenesis induced by infection. However, there is a lack of evidence to explicitly demonstrate the efficacy and possible disadvantages of romosozumab, and well-designed studies are necessary.

Anabolic Drug: Teriparatide
Teriparatide is a recombinant human parathyroid hormone (PTH) that contains the 1st to 34th amino acids of the N-terminus of intact PTH [62]. When PTH was administered at a low dose intermittently, it had anabolic effects on the skeleton [35]. Most PTH signals are transmitted through the PTH-1 receptor, a G protein-coupled protein, activated by peptide sequences in the N-terminal region [35]. PTH regulates osteoblast function through the activation of the cAMP-dependent PKA and calcium-dependent protein kinase C signaling pathways [35,63]. The other mechanisms that are involved are the MAP kinase and phospholipase A and D pathways [35]. In addition, the expression of the Wnt antagonist sclerostin was downregulated by chronic and excess administration of PTH, which explains the anabolic effects of PTH, although intermittent daily administration failed to suppress sclerostin [64]. PTH acts directly on osteoblast lineage cells through an antiapoptotic action by Runx2 in intermittent administration and indirectly affects those cells through growth factors, such as IGF-I, and growth factor antagonists, such as sclerostin [35,63,64]. IGF-I synthesized in osteoblasts results in potent anabolic effects in concert with the direct effects of PTH on cancellous bone [35].
Despite the accumulation of knowledge about teriparatide, there are still a limited number of studies reporting the effects of teriparatide on bone loss secondary to infection. There are a few case reports of the effect of teriparatide on septic arthritis with and without arthroplasty, pyogenic osteomyelitis and nonunion of the tibia, and pyogenic spondylitis, all of which concluded that teriparatide was effective on this pathological condition without the aggravation of infection [65][66][67][68]. Although not simulating the exact situation of pyogenic osteomyelitis, regarding the molecular mechanism of how teriparatide works on the bone when the host is undergoing sepsis, Terashima et al. studied its effect using cecal ligation and puncture-treated mice [69]. Previous studies have demonstrated that IL-7 stimulates the proliferation and activation of lymphoid cells and improves host survival in sepsis [69][70][71]. In the study of Terashima et al., sepsis resulted in a rapid ablation of osteoblasts and subsequent reduction in common lymphoid progenitors via the effects of granulocyte colony-stimulating factor, leading to lymphopenia-associated immunodeficiency, but sepsis caused no change in osteoclasts [69]. Interestingly, the use of teriparatide not only stimulated the proliferation of osteoblasts but also improved the lymphopenia induced by sepsis through an increase in IL-7 [69]. The occurrence of consequent augmentation of T and B cells was not limited to peripheral blood but was also evident in the bone marrow [69]. Considering these findings, teriparatide may positively affect bone loss secondary to pyogenic myelitis by bone formation, boosting the proliferation of osteoblasts and strengthening the immune system by increasing the number of lymphocytes through IL-7 signaling. However, further studies are needed to rule out adverse side effects, such as paradoxical bone resorption or hypercalcemia, when using teriparatide in an acute phase of bone infection.

Antiresorptive Drug: Denosumab
Denosumab is a human monoclonal antibody against RANKL that inhibits the differentiation of osteoclast precursors into mature osteoclasts [72]. However, because its receptor, RANK, is also expressed by monocytes/macrophages and dendritic cells, and RANKL, RANK, and OPG deficiency in murine models underscores the importance of this pathway in the development and maturation of the immune system [25,[73][74][75][76], a concern arose regarding potential infectious events when RANKL was inhibited [72]. Initially, the results from a clinical trial, fracture reduction evaluation of denosumab in osteoporosis every 6 months (FREEDOM), showed that there was no increase in the risk of serious infections [77]. However, a more recent meta-analysis concluded that there was a significant increase in the risk of serious infections in osteoporosis in females or osteopenia patients treated with denosumab [77,78]. This result was supported by newer data from the FREEDOM trial showing concordant results, although the risk increase fell into borderline significance when nonmetastatic breast cancer was excluded [78,79]. Watts et al. also evaluated infectious events in the FREEDOM trial, reporting a more frequent incidence of erysipelas and cellulitis and infections in other regions, such as the gastrointestinal system, renal and urinary system, ear, and endocardium [80]. However, the number of each event was small, and there was no relationship between infectious events and the timing of the administration or duration of denosumab [80], which leads to skepticism regarding causality between the administration of denosumab and the occurrence of infectious events. In summary, denosumab might be associated with an increase in serious infections [81] based on the underlying evidence in the area of molecular science. Denosumab may, therefore, cause the deterioration of infection when used in pyogenic osteomyelitis patients, even if it was effective in suppressing osteolysis. Likewise, further studies are necessary to clarify the efficacy of denosumab in the treatment of this pathological condition.

Antiresorptive Drug: BPs
BPs are a group of pyrophosphate analogs that bind to hydroxyapatite on the bone surface at the sites where remodeling is active and are internalized into osteoclasts through endocytosis, leading to their apoptosis or the suppression of bone-resorbing activity [82]. BPs comprise two types of chemical structures specific to the side-chain groups: nonnitrogencontaining BPs (non-NBPs) and nitrogen-containing BPs (NBPs), which exhibit more powerful effects than the prior and are widely used [83]. Non-NBPs exert their activities by conversion into cytotoxic ATP analogs, interfering with mitochondrial function in osteoclasts and leading to apoptosis [84,85]. NBPs bind to key enzymes of the intracellular mevalonate pathway and inhibit them, preventing the prenylation and activation of small GTPases that are essential for the bone-resorbing activity and survival of osteoclasts [82,[86][87][88][89]. The mevalonate pathway is responsible for the production of cholesterol and isoprenoid lipids [82,90]. Some of the isoprenoid lipids are indispensable for the prenylation and activation of GTPases [82,[90][91][92]. These GTPases play an important role in regulating osteoclast morphology, cytoskeleton arrangement, membrane ruffling, trafficking, and cell survival [82,93,94].
In the context of using BPs as a treatment for bone loss secondary to pyogenic osteomyelitis, the majority of studies have highlighted the downsides of NBPs. First, NBPs have inflammatory side effects leading to osteomyelitis [95]. Deng et al. pretreated macrophages that were infected with Porphyromonas gingivalis and Tannerella forsythia with alendronate, an NBP, and demonstrated augmented production of IL-1β through caspase-1 activation [95]. Furthermore, intraperitoneal injection of NBPs induced histamine-forming enzyme histidine decarboxylase (HDC) in tissues, such as the liver, lungs, spleen, and bone marrow, through IL-1 signaling in murine models [96]. HDC is induced by NBPs, LPS, IL-1, and TNF, and histamine is an inflammatory mediator and a regulator of immune responses, including Th1/Th2 balance and hematopoiesis [97]. Pretreatment with alendronate augmented LPS-stimulated IL-1 production and HDC induction; conversely, pretreatment with LPS augmented alendronate-induced HDC elevation [96]. Moreover, macrophages activate human γδ T cells when treated with NBPs, and it is suggested that macrophages present NBPs to γδ T cells [96,98]. Another study reported that NBPs stimulated human γδ T cells to release TNFα and/or interferon-γ [99] through the inhibition of the mevalonate pathway [100].
Third, the existence of NBPs on the surface of the bone can significantly increase the number of bacteria attached to the bone [104]. When pamidronate was used to coat a hydroxyapatite (HA) material, the number of adherent bacteria was 60-fold greater than that when the HA was uncoated; therefore, NBPs presumably increase the bacterial load at the infection site and exacerbate the infection [104].
Fourth, the insufficient efficacy of BPs on bone loss secondary to pyogenic osteomyelitis can also be pointed out. Kim et al. retrospectively analyzed the efficacy of BPs in pyogenic vertebral osteomyelitis patients, subgrouping as follows: group A, patients who received BPs within 6 weeks after diagnosis; group B, patients who received BPs between 6 weeks and 3 months after diagnosis; and group C, patients who received no treatment for osteoporosis [5]. Although the hazard ratios for the recurrence of infection were not significantly different among the three groups, bone mineral densities measured by dualenergy X-ray absorptiometry decreased by 0.7% in group A and 1.7% in group B at the lumbar spine one year after the diagnosis [5]. In another study, the administration of an NBP aggravated the infection. This study characterized the bone changes resulting from Staphylococcus epidermidis infection in a rodent orthopedic device-related infection model and further evaluated whether ovariectomy (OVX) or BP treatment influenced the infection [105]. As a result, treatment with zoledronic acid did not have bone-protective effects on OVX-infected animals; moreover, it significantly increased the bacterial load, suggesting that osteoclasts might be important in the control of the infection [105]. Supporting this theory, there is a study reporting osteoclasts as immune-competent cells that can internalize and present bacterial antigens to T cells [106]. Several studies have reported that NBPs cause patients to be more susceptible to infection [107,108]. Although one study reported a protective role of zoledronic acid on healing tooth extraction wounds and bone loss in a mouse model of pyogenic osteomyelitis of the jaw [109], this may be limited to the oral region environment. The majority of the studies reported the disadvantages of NBPs in treating bone loss secondary to pyogenic osteomyelitis (see Figure 3).  Evidence for each drug regarding the therapeutic effects on pyogenic osteomyelitis is recapitulated in Table 1.  Evidence for each drug regarding the therapeutic effects on pyogenic osteomyelitis is recapitulated in Table 1. Insufficient efficacy on the bone loss * Romosozumab lacks evidence from previous studies; therefore, we discussed based on the known efficacy of romosozumab. BMP2, bone morphogenetic protein 2; HDC, histamine-forming enzyme histidine decarboxylase; IFNγ, interferon-γ; IL, interleukin; LPS, lipopolysaccharide; MCP-1, monocyte chemoattractant protein-1; MIP-1α, macrophage inflammatory protein-1α; NBPs, nitrogen-containing BPs; NFATc1, nuclear factor of activated T cells cytoplasmic 1; NF-κB, nuclear factor-κB; OPG, osteoprotegerin; RANKL, receptor activator of NF-κB ligand; TLR, Toll-like receptor; TNFα, tumor necrosis factor α.

Conclusions
Considering the underlying characteristics of bone loss in pyogenic spondylodiscitis and the relatively older population of patients with this condition, the use of anabolic medications may be desirable in these patients, given the potential advantages reported in previous studies. Further accumulation of evidence will help in the treatment of patients with this intractable disease, which is increasing in prevalence and impacts society.
Funding: This research did not receive any specific grants from any funding agencies in the public, commercial, or not-for-profit sectors.