2.1 TYPES OF pain. ACute and CHRONIC PAIN
2.1.1 Neuronal pathways involved in chronic pain
The nociceptive system includes a complicated network of peripheral and central neurons.
Signals from the periphery detected by nociceptors are first generated in the ‘free nerve endings’ of the primary afferent fibers (PAFs, Fig 1). The free nerve endings of nociceptors terminate in skin, tendons, bones, joints, and in other body organs. It has been assumed that nociceptors of different tissues share most of their properties. Nevertheless, it has been claimed that there might be differences in the detailed morphological and functional properties of nociceptors in individual tissues (Schaible, 2006). For example, there is a difference in the mechanical threshold for activation in neurons supplying skin and bone, as demonstrated in Fig 2 (Castañeda-Corral et al., 2011). Aδ- and C- fibers are the typical nociceptors that preferably transduce noxious stimuli. In contrast, Aβ-fibers preferentially
transduce innocuous stimuli such as vibration, touch and pressure, but not pain (Millan, 1999). PAFs are often classified based on their main characteristics such as diameter, structure and conduction velocity. Thus, they are: (i) thin C-fibers that are 0.4-1.2 μm in diameter, they are unmyelinated and are slowly propagating at a rate of 0.5-2.0 m/s; (ii), medium diameter Aδ-fibers have a diameter of 2-6 μm, they are myelinated to some extent, with a conduction rate of 12-30 m/s; (iii) Aβ-fibers have the large diameter (≥10 μm), they are heavily myelinated, with very fast conduction rate up to 30-100 m/s.
Figure 1. Schematic representation of input from PAFs into the spinal cord of dorsal horn laminas.
Aδ fibers preferably innervate lamina I, while the principal target of the C fibre input is lamina II0. Unmyelinated C fibers projecting from viscera, joints and muscle appear to preferentially innervate laminae I /IV/V/IV rather than lamina II. Sensory neurons (NS - nociceptive-specific, WDR – wide dynamic range and NON-N – non-nociceptive) are indicated. Adapted from Millan, 1999.
Most of the nociceptive neurons are multimodal as they can sense mechanical, thermal and chemical stimuli. Therefore, they are able to detect cutaneous, somatic and visceral pain (Belmonte and Cervero, 1996). Information from the peripheral nociceptors is transmitted to the neuronal bodies of the sensory afferents, which are located in the dorsal root ganglia (DRG, 31 pairs in humans) and its cranial analogue, the trigeminal ganglion (TG, one pair in humans). Further, DRG axons project to the spinal cord and form the synapses mostly in the substantia gelatinosa (lamina I-III) of the dorsal horn in the spinal cord as illustrated in Fig 1. Second order nociceptive neurons in the spinal cord form several ascending pathways to the higher pain centers in the CNS. Some of them project to the thalamo-cortical system that produces the conscious sensation of pain, whereas others, such as dorsal and ventral spinocerebellar tracts, can control motor functions (Willis, 2007). The other subset of sensory neurons projects to the motor nucleus of the spinal cord, to be involved in withdrawal reflexes as well in more complex pain avoiding behaviour (Schaible, 2006).
The macroscale morphology of the nociceptive system has been relatively well studied.
However, many fine mechanisms, especially those at the molecular level, are far from clear due to disease-specific pathophysiology diversities. Disease-specific diversity could be based on tissue specific prevalence of innervation by C-, Aδ- or Aβ-fibers, profile of pro-nociceptive
endogenous agents and variable contribution of neuronal damage or peripheral inflammation (Goff et al., 1998).
2.2 DISTINCTIVE CHARACTERISTICS OF BONE PAIN 2.2.1 Bone innervation
The innervation of bone has been studied for over a century with a variety of methods, especially the now widely used technique of immunohistochemistry (Hurrell, 1937; Duncan and Shim, 1977). It is evident that bones are innervated by the sympathetic and somatic sensory nerves (Hukkanen et al., 1992; Serre et al., 1999; García-Castellano et al., 2000). Nerve fibers accompany the blood vessels in the bone which travel through all its tissue layers (Mach et al., 2002). Of all the bone tissue types, the periosteum is the most densely innervated.
It has been demonstrated that all bone tissues (mineralized bone, bone marrow and periosteum) are highly innervated by the primary nociceptive fibers (Mach et al., 2002; Falk et al., 2014; Mantyh, 2014). However, as the total volume of the periosteum is less than that of the mineralized bone and the bone marrow, the total number of sensory and sympathetic fibers in the latter is also high (Castañeda-Corral et al., 2011; Mantyh, 2014).
Figure 2. A schematic showing the approximate percentage of PAFs that innervate the skin and the bone. The skin is innervated by Aβ fibers - , thiny mielinated Aδ - and , peptide-rich C fibers - , and peptide-poor C fibers ( ). In contrast, the bone has been shown to be predominantly innervated by thinly myelinated Aδ, and , and peptide-rich C fibers ( ). In both, skin and bone, there is also a small proportion (< 5% of the total) of non-myelinated C-fibers (CGRP+, calcitonin gene related peptide positive; TrkA−, tropomyosin receptor kinase A negative fibers). (NF 200+, neurofilament 200 positive fibers) Adapted from Castañeda-Corral et al., 2011.
Notably, it has been shown that the bone is primarily innervated by the mildly myelinated Aδ- and unmyelinated C–fibers, which transmit noxious stimuli to the dorsal horn. These fibers express various neuropeptides such as neuropeptide Y (NPY), substance P (SP), vasoactive intestinal peptide (VIP) and calcitonin gene related peptide (CGRP) and some of these peptides are clearly pro-nociceptive. These fibers also display the major pro-nociceptive ATP-gated P2X3 receptors, acid sensing ion channels (ASIC) and several types of vanilloid receptors (TRPV1-4) (Bjurholm et al., n.d.; Fischer et al., 1996; Mach et al., 2002; Castañeda-Corral et al., 2011; Falk et al., 2014; Mantyh, 2014). However, most of the sensory fibers that
innervate bone are mainly by thinly myelinated TrkA+ (Tropomyosin Receptor Kinase A) sensory nerve fibers (̴ 80%, belong to A- fibers), i.e. they are very different from the sensory fibers in skin (̴ 30% of Trka+, Fig 2)(Castañeda-Corral et al., 2011). Taken together, these data suggest that there is an abundance of the main pain transducing receptors in the bone and this has clear implications with respect to chronic bone pain.
2.2.2 Bone pain at a glance
Bone pain represents one of main chronic pain conditions. Bone pain shares many of the common characteristics of both inflammatory and neuropathic pain but it also has unique components due to the specific profile of cells in this tissue (Falk and Dickenson, 2014). The variety of bone disorders including malignant and non-malignant processes, contributes to the complex pathophysiology of bone pain. Pain associated with malignant bone processes is a complex condition involving spontaneous (background) pain and movement-evoked pain (Portenoy and Hagen, 1990; William and Macleod, 2008). Background pain is specified as dull and continuous pain, which increases with the progression of the main disease.
Usually, it can be reasonably well treated with traditional analgesics. Evoked pain is usually described as a breakthrough feeling, being less sensitive to common analgesics which effectively decrease background pain (William and Macleod, 2008). The treatment of this type of pain requires the combination of analgesics with the cancer- and bone-targeting specific therapies. This is especially important for patients with breast, prostate, kidney and lung cancers which preferably metastasize to the bone (Makhoul et al., 2015).
Until recently, we had very limited knowledge of tumor-induced bone pain. The popular view was that this kind of pain was caused by the compression of peripheral nerves or by vascular occlusion in the bone. In 1998, the first tumor-induced bone pain model was developed and published by Schwei et al. (Schwei et al., 1999). This animal model of femoral osteosarcoma provided the opportunity for researchers to study mechanisms underlying tumor-induced bone pain (Falk et al., 2014).
Notably, bone resorption and construction is permanent and very important processes. This processes of bone remodeling, bone formation and resorption are tightly coupled and directly influenced by interactions between osteoblasts and osteoclasts as illustrated in Fig 3 (John P. Bilezikian, Lawrence G. Raisz, 1996; Crockett JC1, Rogers MJ, Coxon FP, Hocking LJ et al., 2011). It is essential that there is an equilibrium in bone remodeling between resorption and formation in order to ensure the integrity of the bone structure. Therefore, an increase of the resorption by osteoclasts leads to osteoporosis, in contrast, excessive bone formation should result in disruption of calcium homeostasis. All of these conditions potentially could lead to bone pain, which is one of the most excruciating types of chronic pain.
Figure 3. Bone remodeling and involved cell types (adapted from Weilbaecher 2011). Osteoclasts are mane players in resorption after they are maturation, they are multinucleated, polarized cells that adhere to the bone surface. RANKL – , RANK – , HSC – , MSC – , M-CSF – , specified in the abbreviation list.
Further improved models for bone pain helped to reveal the role of bone resorptive processes that were derived from osteoclast over-activation in bone malignancy disorders. Osteoclasts-induced resorption promotes a highly acidic environment in the resorption region. Osteoclast resorb the mineral bone matrix by secreting collagenases and proteases that demineralize and degrade proteins such as type I collagen of the bone (Weilbaecher et al., 2011; Florencio-Silva et al., 2015). During this resorption osteoclast release H+ (hydrogen ions) through the action of carbonic anhydrase through the ruffled border into the resorptive cavity what strongly increasing acidity in the actively resorptive regions, acidifying and aiding dissolution of hydroxyapatite mineral to Ca2+, H3PO4 and H2CO3 (Currey, 2006). While normal remodeling this will not be noted, however during osteoclast over activated resorption, while osteoblast are in minor, this events directly lead to acidosis and first local and further humoral hypercalcemia (Fig 3 and 4). This directly activates the acid-sensitive TRPV1 and ASIC receptors expressed in nerve terminals innervating bone (Luger et al., 2005;
Nagae et al., 2007). Notably, it has been demonstrated that high Ca2+ concentrations can inhibit certain P2X receptors (Virginio et al., 1998). Intriguingly, detailed studies of the modulation by different cations of the P2X3 receptor subunit, revealed that unlike other P2X receptors, it is strongly facilitated by extracellular Ca2+ (Fig 5) that acts via specific sites in the ectodomain located next to the ATP binding pocket (Giniatullin et al., 2003; Petrenko et al., 2011; Giniatullin and Nistri, 2013). Furthermore, acidosis dramatically increases the expression of ASIC1b, ASIC1a and ASIC3 receptors in the DRG ganglion, resulting in peripheral sensitization and hyperalgesia, as presented in Fig 3 (Nagae et al., 2006, 2007). In addition, osteoclast induced hypercalcemia influences many processes in the whole body including the sensory system. Calcium is known as a strong modulator of nociceptive ATP-gated P2X3 receptors in sensory neurons (Giniatullin et al., 2003) being a co-activator of these P2X3 receptors (Fig 3). It can also affect the inhibitory activity of P2X3 antagonists (Giniatullin et al., 2003; Ishchenko et al., 2017). Notably, it has been shown in both vitro and in vivo models that bone cancer is associated with enhanced expression of P2X3 receptors in CGRP-positive sensory nerves, which are extensively present in the bone (Gilchrist et al., 2005; Wu et al., 2012, 2016). Thus acidosis and hypercalcemia, can lead to peripheral nerve sensitization
and further, to central sensitization, the two main features of bone pain (Urch, 2004; Luger et al., 2005; Mantyh, 2014)
Figure 4. Scheme of the pathophysiology of malignant bone pain biology. (A) Representation of the peripheral and central sensitization. (B) Molecular mechanisms implicated into development of the bone pain and receptors involved in its transduction. Represented receptor and chemicals (P2X3, ASICs, TRPV3, TRPA1 and H+, IPP, ApppI, ATP, BPs, RANK, RANKL) see in main abbreviation list.
As sensory nerve fibers are distributed over all of the bone tissues, tumor proliferation associated with the secretion of the various pro-nociceptive compounds can strongly activate sensory nerve fibers in the bone (Luger et al., 2005). In addition, tumors are able to destabilize and reorganize the peripheral nerves, leading to either pathological sprouting or the destruction of distal sensory and sympathetic nerve fibers (Mach et al., 2002; Mantyh, 2014). Tumors and associated stromal cells can release high amounts of multiple algogens (i.e. endogenous pain-evoking compounds) such as NGF, pro-nociceptive cytokines, including tumor necrosis factor (TNF-alpha) along with extracellular ATP, which is the endogenous agonist of the P2X3 receptors, (Fig 4) (Urch, 2004; Goblirsch et al., 2006).
These destructive events and active pro-nociceptive compounds can induce peripheral and central sensitization and lead to pathological changes in the nociceptive system (Luger et al., 2005; Mantyh, 2006). This can contribute to osteoclast-induced resorption, mediating the release of the bone stored growth factors, further stimulating tumor proliferation and tumor growth (Fig 4) (Yoneda, 2013). The growing scientific interest in this field and the improved animal models have widened our understanding of the bone pain associated with malignant and non-malignant processes. However, the complexity of the bone pain processes still requires clarification.
2.2.3 Calcium metabolism and hypercalcemia
It is well known that the skeletal system is crucially involved in calcium metabolism. Both aging and a range of bone disorders are able to alter the calcium levels in the extracellular milieu and in blood, leading to either hypo- or hyper-calcemia. Recent evidence has revealed that the cancer-related bone pain is often associated with osteoclast activation and hypercalcemia as represented on Fig 4 (Luger et al., 2005; Nagae et al., 2006). In some pathological conditions, such as in parathyroid hormone-related crisis, the serum calcium levels are elevated from the normal ~2 mM to become hypercalcemic at 4.8 mM (Rahil and Khan, 2012). In addition, hypercalcemia has been reported to be a frequent (20 - 30 % of cases) complication in patients with cancer (Basso et al., 2011; Hu et al., 2013; Goldner, 2016; Tagiyev
et al., 2016). Hypercalcemia, to even a mild degree, can dramatically worsening the patient’s condition, leading to a progressive mental impairment, as well as to renal failure and it carries a high risk of mortality. Even moderate hypercalcemia frequently results in marked neurologic dysfunction (Stewart, 2009). The two most frequent types of hypercalcemia are humoral hypercalcemia of malignancy (HHM) and osteolytic hypercalcemia. The HHM is often caused by the systemic secretion by tumor cells of the parathyroid hormone related protein (PTHrP). PTHrP promotes renal tubular calcium reabsorption, which elevates the serum calcium level (Ratcliffe et al., 1992). The second type of calcium growth is associated with a local osteolytic activity due to an increase in osteoclastic bone resorption in the areas surrounding the malignant cells. When compensatory mechanisms are exceeded, the serum calcium level rises causing hypercalcemia (Clines and Guise, 2005). There are also different, more specific cases of malignant hypercalcemia such as the production of ectopic calcitriol by malignant lymphocytes in multiple myelomas (Roodman, 1997) or direct PTH production ectopically by tumor cells (Inzucchi, 2004).
Notably, the neurological impairments in hypercalcemia often include pain and migraine (Malangone and Campen, 2015; Yin et al., 2016). Cancer patients diagnosed with hypercalcemia often suffer from headaches. Recent, clinical research based on 23,285 migraine patients and 95,425 controls revealed a direct association between the elevated serum calcium level and an increased risk of migraine (Yin et al., 2016). The current therapy of hypercalcemia, which aims to reduce the serum calcium concentration is focused on increased calciuresis, decreased bone resorption, and reduced intestinal absorption of calcium. It is important to understand the pathogenesis and treatment options for hypercalcemia associated with malignancy, in order that prompt treatment interventions can be administered.
However, there is still very little known about the molecular (receptor) mechanisms underlying malignant and non-malignant bone pain. However, data from our laboratory and from others already suggested some clues in elucidating the mechanism. Indeed, it has been demonstrated that extracellular calcium plays an important role in the signaling of P2X receptors, accelerating re-sensitization of the P2X3 receptor currents and making these pro-nociceptive receptors functionally more active (Cook and McCleskey, 1997; Cook et al., 1998) (Cook et al., 1998; Giniatullin et al., 2003). On the other hand, high calcium and magnesium levels can inhibit the activity of P2X7 receptors (Coddou, S. Stojilkovic, et al., 2011).
Thus, the role of hypercalcemia on development of pain should be clarified in order to identify new targets and develop analgesics able to combat these debilitating types of pain such as bone pain and migraine.
2.4 BISPHOSPHONATES AND CURRENT TREATMENTS OF BONE PAIN 2.4.1 Non-specific treatments for bone cancer
Majority of treatments applied for pain treatment in malignant and non-malignant bone disorders are similar to other pain disorders and their main differences are determined primarily by the severity of the disease. Common (non-specific) chronic pain treatment includes non-steroid anti-inflammatory drugs (NSAIDS) (Fu et al., 2015) and different opioids (Yaldo et al., 2016). Their efficacy is still limited by various adverse side effects and
NSAIDS often provide insufficient pain relief. Thus, in many cases, bone pain cannot be resolved by common current therapies what strongly decrease quality of life (van den Beuken-van Everdingen et al., 2007, 2016; Coleman et al., 2014). In fact, currently, pain management of malignant bone disorders is considered as not satisfactory (Smith and Mohsin, 2013). Currently 85-95% of patients with bone cancer have significant malignancy-induced pain and 45% of these patients collide with an inadequate pain control therapy or unmanaged pain (Smith and Mohsin, 2013). Also, many NSAIDs treatments are not sufficient to decrease skeletal pain or have not strong enough effect in consideration of the personal side effect risks (Rasmussen-Barr et al., 2017). In addition, it is well known that NSAIDs have side effects including the gastrointestinal and cardiovascular complications (Sostres et al., 2010). In particular, they injure the upper and lower gut by depleting COX-1 derived prostaglandins causing the peptic ulcer (Sostres et al., 2010). Also, recent clinical meta-analysis found a direct link of NSAIDs treatment to the heart failure (Arfè et al., 2016).
Another major issue to consider is that some typical NSAIDs such as ibuprofen and Cox-2 inhibitors slow down fracture healing in animal models of bone fracture (O’Connor et al., 2009; Barry, 2010).
Therefore, new approaches to the chronic bone pain control are strongly needed. Ideally, if the analgesic effect is associated with promotion of the bone formation and with skeletal healing or, at least, the anti-nociception develops in the absence of bone healing inhibition.
2.4.2 Beneficial effects of the BPs for bone pain treatment
Currently, bisphosphonates (BPs) are the safest treatments for bone disorders associated with bone lesions. BPs are highly specific to the hard bone tissue containing hydroxyapatite. BPs are divided into two types: i) compounds lacking a nitrogen group (non-NBP) (Frith et al., 2001; Räikkönen et al., 2009; Rogers et al., 2011) and ii) nitrogen-containing compounds (NBP).
Interestingly, both type of BPs can induce the formation of endogenous ATP-analogues as represented on Fig 5. Thus, non-NBPs promote the synthesis of compounds with ApCp-groups, whereas NBPs induce IPP/DMAPP (Isopentenyl pyrophosphate/dimethylallyl pyrophosphate) and the ApppI (1-adenosin-5’-yl ester 3-(3-methylbut-3-enyl) ester). Since these are relatively recently discovered compounds, their roles in health and disease are still far from clear.
Figure 5. Scheme of the cellular mechanism of BPs (top). Induction of the ATP analogues by BPs (bottom). Adopted from Russell, 2011.
Both non-NBPs and NBPs have anti-osteoclast and anti-cancer activity (Fig 5 and 6). They can interfere with mitochondrial ATP production or inhibit the mevalonate pathway and activate caspases (Fromigue et al., 2000; Oades et al., 2003; Green, 2004; Koshimune et al., 2007; Tanaka et al., 2013). Recently, evidence emerging from clinical studies has suggested that BPs, in combination with other treatment modalities, can diminish pain in cancer patients (Body et al., 2004; Tagiyev et al., 2016). Other clinical research findings have indicated that BPs can successfully reduce the level of persistent pain not only in bone disorders, but also in patients with chronic low back pain and complex regional pain syndrome type I (Haslbauer and Fiegl, 2009; Abe et al., 2011; Pappagallo et al., 2014).
In general, the anti-nociceptive effect of BPs can be either direct or indirect, mediated by production of endogenous ATP-analogues. For instance, there is a strong correlation between BPs’ anti-resorptive and anti-tumor effects and their ability to induce the formation of endogenous ATP-analogues (ApppI, IPP, AppCClp) in vivo (Ramanlal Chaudhari et al., 2012). Additionally, it has been proposed that BP-induced formation of ATP derivatives can participate in pain relief (Fromigue et al., 2000; Hadji et al., 2016). In particular, NBPs represent the most effective group, pointing to a possible involvement of ApppI and IPP in analgesia.
Notably, BPs induced analgesia could involve even more complicated pathways including modification of the immune response. Thus, two NBPs, zoledronate and residronate, activated human Vγ9Vδ2 T-cells, which have potent anti-tumor properties (Benzaïd et al.,
2011, 2012). NBP pre-treated monocytes accumulate ApppI and IPP, contributing to the
2011, 2012). NBP pre-treated monocytes accumulate ApppI and IPP, contributing to the