Osteoclasts, and in particular conditions also osteocytes, use acidity to solubilise the mineralised fraction of the bone matrix, allowing ionic calcium and phosphates to be released into the bloodstream [73]. The acidification process must be tightly controlled to avoid protons leaking towards the bone marrow and having ineffective bone resorption. Physiologically, osteoclasts create a tightly sealed area using podosomes, limiting proton efflux to a discreet area between the osteoclast itself and the bone matrix, called the resorption lacuna [30]. When local bone resorption is complete, osteoclasts release the seal and move to a close area or undergo apoptosis. In normal conditions this process involves little or no proton leakage, and the relatively small number of osteoclasts guarantees that the pH of the bone marrow is not altered by bone remodelling. Cancer-induced osteoclastogenesis creates an acidity excess in the bone marrow by increasing absolute osteoclast number. Additionally, tumour cells also contribute to the acidification of the microenvironment through the Warburg effect, especially considering that bone is not a normoxic tissue, but a hypoxic one (5% pO2) [74], which stimulates cancer cells to switch towards a glycolytic metabolism, and therefore produce lactate and protons. Acidity is therefore an important (and often overlooked) result of the vicious cycle especially on the standpoint of bone pain.

Although the role of acidity in nociception seems clear in fields such as gastroenterology, the molecular mechanisms underlying acidity-induced bone pain are not as clear to date. Recently, studies on mouse and rat models of cancer-induced bone pain (CIBP) revealed that there are two main families of receptors that are activated by acidity: transient receptor potential channel-vanilloid (TRPVs) subfamily members and acid-sensing ion channels (ASICs) [75,76,77,78,79,80]. TRPVs and ASICs are described as proton-activated, cationic current-generating receptors, and are present in periosteal as well as bone marrow nociceptive terminals. The main molecular players belonging to this family seem to be TPRV1 and ASIC3. Although similar in function, they have different features as far as pH-sensitivity and sensitisation/desensitisation go.

TRPV1 is expressed by calcitonin gene related peptide-positive (CGRP+) C-type neural sensory terminals. It is activated by pH < 6, and sensitised to other stimuli, such as capsaicin, heat and inflammatory mediators between pH 6 and 7 [81,82]. To confirm the role of TRPV1 in CIBP, recent preclinical studies showed that treating mouse models of CIBP with either an antagonist of TRPV1 or administering another TRPV1 antagonist along with morphine, caused reduction (or further reduction in the case of co-administration) of CIBP [75,83].

At variance with TRPV1, ASIC3 is activated by milder acidosis (pH 6.7–7.3), and different models of pain showed that its activation is sufficient to induce nociceptive behaviours [84,85]. Although its role in CIBP seems clear, researchers are still to demonstrate that inhibiting its signalling can reduce CIBP in solid tumours, although a study is present where inhibiting ASIC3 is able to reduce CIBP in multiple myeloma [86]. However, an inhibitor for this channel has been developed and has been used to reduce osteoarthritic pain, and even its progression [87].

The different pH sensitivity of these two receptors convinced investigators that they are used to discriminate between mild and severe extracellular acidification of the bone marrow [88].

Neurotrophins include a small family of molecules: nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3). Neurotrophins have a coreceptor—p75—that is common to all of them, and a specific receptor that is tropomyosin-related kinase (Trk)-A for NGF, TrkB for BDNF and TrkC for NT-3 [90]. These are present in both the CNS and the CGRP+ sensory terminals in the bone marrow. Usually, NGF is considered as the most potent pain inducer and blocking its binding to TrkA using a monoclonal antibody (mAb911, Rinat/Pfizer) is able to strongly reduce CIBP in mouse models [91,92,93]. Tanezumab, a humanised version of mAb911, is currently in a Phase 3 clinical trial in association with opioids in a randomised, double-blind, placebo-controlled, multicenter, parallel-group study (Clinical trial NCT02609828) and is probably the drug that currently holds the most promise for pain management in bone metastatic patients, in addition to anti-resorptive drugs (read below).

BDNF could also have a role in sensitising central neurons to pain response [94], therefore general Trk inhibitors are also being developed [95]. Intriguingly, metastatic breast and prostate cancer cells express high levels of NGF and BDNF [96,97], which not only can stimulate nociceptors directly, but may also induce macrophages to secrete TNF-α, IL-6, IL-1β and prostaglandin (PG)E2 [98]. These inflammatory mediators can further activate pain responses, fuel the vicious cycle, sensitise nociceptors and activate other molecular mediators of pain, such as acidity [99,100,101]. A further link between NGF and acidity is that this neurotrophin can sensitise TRPV1 to protons and increase its expression, causing acidity-directed allodynia and hyperalgesia, which further worsens CIBP [81].

As mentioned above, ILs can increase osteoclastogenesis and tumour survival, but also induce bone pain. As a matter of fact, IL-1β, arising from cancer-related inflammation, increases macrophage expression of cyclooxygenase (COX)-2, eventually leading to an increased production of prostaglandins, which bind prostanoid receptors on sensory terminals, resulting in CIBP [102,103]. Supporting the conclusion that IL-1β is involved in CIBP, an in vivo mouse model of osteosarcoma showed that this interleukin is not only increased in the tumour area, but also in the spinal cord, and inhibiting its receptor reduces mechanical and thermal hyperalgesia [104].

TNF-α is another notable hyperalgesia-inducing substance: it has been suggested that this happens through the sensitisation of TRPV1 channels, linking back to acidity [105]. Monocyte chemoattractant protein (MCP)-1 has also been directly implicated in CIBP and mechanical allodynia and, as demonstrated by Hu and colleagues, intrathecal administration of an MCP-1 neutralising antibody reduces CIBP in a breast cancer model of bone metastasis [106]. Although other cytokines and chemokines have been correlated to bone pain [61] the link with CIBP is most likely indirect and is dependent on osteoclastic bone resorption.

Although the aforementioned factors would be sufficient to explain most of CIBP (if every hypothesis proves correct), there is still a sizeable side of it that still needs to be explained. As a matter of fact, many other molecular mediators emerged in the last few years.

One of the most important examples is arguably ATP. This molecule is present in every single living cell, and it is never found in the extracellular environment, unless there is cell or nerve damage, which may culminate in necrosis. Extracellular ATP can be sensed by the purinergic receptors P2X (ionotropic) and P2Y (metabotropic) [107]. A member of the former family, P2X3, has been studied extensively in CIBP, since it is expressed quite specifically in small diameter nociceptive fibres throughout the bone marrow and periosteum [108]. Interestingly, treating rats with antagonists for this receptor reduces pain-related behaviours in CIBP models, and the same has been shown in mouse models [108,109]. Nowadays, second generation P2X3 receptor antagonists are being developed to overcome some of the limitations encountered by the first-generation antagonists, such as dysgeusia (i.e., alteration of taste) and hypogeusia (i.e., reduction of taste) (Clinical trial NCT03449134). Nevertheless, the P2X3 antagonist Gefapixant has gone through several Phase 2 clinical trials with promising results for chronic endometriosis-related pain (ongoing, NCT03654326) and bladder pain syndrome (completed, NCT01569438), but still not CIBP.

TGF-β1 and IGF-1 are also two strong candidates for the induction of CIBP. They are highly represented in the organic bone matrix [110] and are therefore released in the microenvironment during tumour-induced bone resorption. Intriguingly, they are already notorious vicious cycle-fuelling factors [111]. A recent report showed that TGF-β1 signalling is crucial for CIBP onset and development in a preclinical study on rats [112]. This makes TGF-β1 an even more interesting molecular target for cancer bone metastases, although the ubiquitous nature of this protein makes inhibition a difficult path to walk. Nevertheless, a number of clinical trials have been completed or are ongoing, and many of them show promising results. As for IGF1, it has been shown to sensitise the TRPV1 receptor, promoting acidity-induced nociception in the context of bone metastases [113]. Another important mediator of cancer induced bone pain is osteocyte-derived Sclerostin (SOST) [114]. Osteocytes, in addition to mediating pain in response to microfractures, which are often cancer-related [76], release SOST as a consequence of multiple myeloma (MM). SOST is a potent inhibitor of the Wnt/β-Catenin pathway, which in this case appears to be a driver of MM-induced bone disease. In fact, genetically ablating SOST or blocking it with an antibody can strongly reduce osteolysis and consequent acidity-mediated pain [114].

Finally, serotonin which seems to be able to increase sensitivity of ASIC3 in the dorsal root ganglion (DRG) of rats intratibially inoculated with cancer cells [115] and to modulate TRPV1 activity as well [116].