The platinum-based chemotherapeutic agents oxaliplatin and cisplatin, both of which are listed on the World Health Organization's List of Essential Medicines, are used for the treatment of various solid tumors. Oxaliplatin is used in combination with folinic acid and 5-fluorouracil as a part of the FOLFOX regimen for first-line and adjuvant colorectal cancer therapy, while cisplatin is one of the most effective treatments for solid tumors including small cell lung cancer, testicular, ovarian, brain, and bladder cancer. The intracellular concentration of platinum-based antineoplastics is maintained via several active transporters, including the copper transporter CTR1 which mediates drug uptake (Song et al., 2004) as well as copper-transporting ATPases that mediate efflux, such as ATP7A and ATP7B (Safaei and Howell, 2005).

Like other platinum-based compounds, oxaliplatin and cisplatin interfere with tumor cell proliferation via the formation of deoxyribonucleic acid (DNA)-platinum adducts. Cisplatin is converted to a strong electrophile after hydrolysis of its two chloride atoms, which are displaced by two molecules of water once the platinum complex enters the cell. Activated cisplatin then crosslinks two purine bases (adenine/guanine) of DNA by reacting with nitrogens at position seven of the purine rings, thus interfering with cell division and transcription of messenger ribonucleic acid (mRNA) (Dasari and Tchounwou, 2014). Similarly, the mono- and di-aquated oxaliplatin creates DNA intra-strand crosslinks by binding two guanine bases or a guanine-adenine pair of GC-rich regions of DNA (Faivre et al., 2003). Additionally, oxaliplatin causes inter-strand crosslinks and DNA-protein crosslinks which may contribute to the mode of action of oxaliplatin (Zwelling et al., 1979). These DNA mono- and bi-adducts in turn inhibit DNA replication and transcription, a fatal effect on cells in the S-phase of the cell cycle (when chromosomes are replicated; Alcindor and Beauger, 2011). Oxaliplatin also inhibits DNA transcription/mRNA production by interacting with transcription factors, inhibition of RNA polymerases and by creating nucleosomal DNA adducts leading to cell death (Todd and Lippard, 2009). In addition, activation of the immune system may be involved in the anticancer effects of oxaliplatin, with production of interferon γ by T-cells and resultant immunogenic cell death via TLR4 activation observed in both murine and human colon cancer cells (Tesniere et al., 2010). Likewise, cisplatin causes additional non-genomic effects including the production of reactive oxygen species resulting in altered mitochondrial function and activation of both intrinsic and extrinsic apoptosis pathways. Cisplatin also affects calcium signaling pathways and the function of several protein kinase families, including the MAPK (mitogen activated protein kinases), JNK (c-Jun N-terminal kinase o), PKC (protein kinase C) and AKT (serine/threonine kinases) leading to cell death (Dasari and Tchounwou, 2014).

A unique feature of oxaliplatin is its rapid non-enzymatic transformation to the reactive dichloro 1,2-diaminocyclohexyl-platinum complex and oxalate which occurs via replacement of the oxalate moiety with chloride ions in the blood plasma. Indeed, the generation of the oxalate metabolite is one of the few features distinguishing oxaliplatin from cisplatin, and has been proposed as one mechanism accounting for the differences in clinical presentation—most notably the presence of cold-induced neuropathy after treatment with oxaliplatin, but not cisplatin. However, a contribution of oxalate to oxaliplatin-induced cold pain has not been shown consistently (Deuis et al., 2013), and it is plausible that additional metabolites—including platinum complexes that bind to cellular proteins—may contribute to the development of CIPN.

Vincristine is one of the most important antineoplastic substances used for chemotherapy of several childhood and adult tumors, and is—administered by intravenous infusion—a component of several different chemotherapy regimens (e.g., MOPP, COPP and BEACOPP regimen for therapy of Hodgkin's disease). While the mechanisms of vincristine transport into the cell are still unclear, a carrier-mediated transport mechanism characterized by Michaelis-Menten kinetics, temperature dependence and competitive inhibition was demonstrated in different murine leukemia cells (Bleyer et al., 1975). The adenosine triphosphate (ATP) binding cassette (ABC) transporter family, ABCB1, ABCC1, ABCC2, and ABCC3 contribute to vincristine efflux from the cell cytosol and to the development of cancer cell resistance to the therapy (Huang et al., 2006). Additionally, a relationship between cholesterol and phospholipid levels and vincristine uptake into murine leukemia cells has been demonstrated, showing that increased levels of cholesterol and phospholipids in the cell membrane accounts for lower vincristine accumulation (Pallares-Trujillo et al., 1993).

Vincristine binds to the β-subunit of tubulin and inhibits microtubule formation. The microtubules are cytoskeletal proteins that are involved in several important cell functions, for instance the regulation of cell shape, mitosis, chromosome segregation, cell division and retrograde as well as anterograde cellular transport. The proper function of microtubules depends on a balance between permanent aggregation and disaggregation of the α- and β-tubulin subunits. Therefore, the disruption of microtubule aggregation by bound vincristine can lead to mitotic arrest and cell death (Gan et al., 2010). The structure of the catharnine ring of the vinca alkaloids seems to be important for binding to axonal and cytoskeletal microtubules. Vinorelbine, composed of an eight-membered catharnine ring, shows a preference in binding to mitotic spindles over axonal microtubules, resulting in decreased neurotoxicity compared to vincristine, which is composed of a nine-membered catharnine ring (Gregory and Smith, 2000).

Paclitaxel and docetaxel belong to the family of taxanes; chemotherapeutic agents used in the treatment of breast, prostate, lung, pancreatic, gynecological and other solid tumors that act by inhibiting disassembly of tubulin from the microtubule polymer. However, despite a similar mechanism of action, subtle differences in their molecular pharmacology, pharmacokinetics and pharmacodynamics result in distinct clinical effects. Docetaxel binds to tubulin with greater affinity than paclitaxel, and causes cell cycle arrest in the S rather than G2-M phase junction as is the case for paclitaxel (Dorr, 1997). Both compounds do not cross the blood-brain-barrier (BBB), although paclitaxel is accumulated in dorsal root ganglion neurons—which lie outside the BBB—via largely unknown mechanisms (Cavaletti et al., 1997, 2000; ten Tije et al., 2004; Wozniak et al., 2016). Both paclitaxel and docetaxel are extensively metabolized, via cytochrome P 2C8 and 3A4, respectively, albeit pharmacological activity of these metabolites is modest at best and their contribution to the development of CIPN is unknown.

Chemotherapy-induced neuropathy is a common, dose-dependent side effect of many antineoplastics that not only leads to dose reduction or discontinuation of treatment, but also severely reduces the quality of life of patients (Bodurka-Bevers et al., 2000; Grisold et al., 2012). While the incidence of CIPN is compound-specific (Seretny et al., 2014), it also depends on patient co-morbidities and can, presumably, increase with concomitant treatment with other neurotoxic drugs (Johnson et al., 2015), although this effect has not been systematically investigated. Invariably, the percentage of patients suffering from CIPN as well as the severity of the condition increases with the (cumulative) dose (Dougherty et al., 2004; Seretny et al., 2014). Accordingly, the incidence of CIPN approaches nearly 100% for some agents at higher doses (Seretny et al., 2014), although differences in the definition, evaluation and reporting of peripheral neuronal deficits can lead to large variability in the reported occurrence. Notably, many studies rely on detailed clinician-administered grading scales or patient-reported outcomes using a range of questionnaires developed to evaluate the development of CIPN, while more objective assessments such as nerve conduction studies, nerve excitability studies, or quantitative sensory testing are used more rarely (Park S. B. et al., 2013). Importantly, CIPN is viewed by many clinicians as an unavoidable side effect of cancer chemotherapy, and one that is acceptable in light of the often greatly extended life-span offered by these chemotherapeutic agents (Cavaletti et al., 2011). In contrast, patients often view CIPN as a particularly troublesome side effect of cancer treatment that interferes significantly with their quality of life (Jones et al., 2015).

Clinically, CIPN presents as deficits in sensory, motor, and autonomic function which develop in a compound-specific manner (Park S. B. et al., 2013). Sensory symptoms usually develop first in the feet and hands—reflective of the axon length, with longer neurites affected first—and manifest as numbness, tingling, paresthesias and dysesthesias induced by touch, warm or cool temperatures, impaired vibration and altered touch sensations. In addition, painful sensations including non-evoked burning, shooting or electric shock-like pain as well as mechanical or thermal allodynia or hyperalgesia frequently occur (Postma et al., 1993; Sahenk et al., 1994; Becouarn et al., 1998; Bernhardson et al., 2007). In severe cases, these symptoms can progress to loss of sensory perception, which can be disabling (Strumberg et al., 2002). Motor symptoms include distal weakness, gait and balance disturbances and impaired fine movements. While rarer than sensory symptoms, motor impairment can progress to paralysis and cause significant functional disruption (Hile et al., 2010; Mols et al., 2016). In contrast, autonomic symptoms occur less frequently and may involve orthostatic hypotension, constipation and altered sexual or urinary function (Mols et al., 2016). With the exception of paclitaxel and oxaliplatin, which cause an acute neuropathy that emerges either during or shortly after infusion (Loprinzi et al., 2011; Argyriou et al., 2013b), the onset of CIPN is usually delayed and appears to depend on the total cumulative dose (Maestri et al., 2005). Some chemotherapeutics cause “coasting” of symptoms, the progressive worsening of neuropathy after cessation of treatment, and in severe cases CIPN can develop into an irreversible sensory neuron deficit.

Cisplatin causes a wide range of side effects including ototoxicity (hearing loss and tinnitus), nephrotoxicity, myelotoxicity and neuropathy. One of the most dose-limiting of these is peripheral neuropathy, which occurs in a dose- and time-dependent manner (Ozols and Young, 1985). The onset of cisplatin-induced neuropathy is variable, with some patients reporting the first appearance of symptoms after the first dose, and others after 12 cycles of therapy (Cersosimo, 1989; Vanderhoop et al., 1990). Generally, cisplatin-induced neuropathy develops after cumulative doses above 350 mg/m², with approximately 92% of patients developing neurotoxic symptoms—characterized by tingling, numbness and mechanical and thermal hyperalgesia—at cumulative doses of 500–600 mg/m² cisplatin (Roelofs et al., 1984; Thompson et al., 1984; Krarup-Hansen et al., 2007). Cisplatin-induced CIPN affects mostly the lower and upper limbs and includes mixed sensory and motor effects, including loss of vibration sense and taste, paresthesia, weakness, and tremor (Lomonaco et al., 1992; Amptoulach and Tsavaris, 2011). Motor nerve conduction velocity measurements in patients with decreased vibration sensitivity, loss of ankle jerks and paresthesias showed loss of sural nerve responses (Thompson et al., 1984), with emergence of reduced sensory nerve action potentials and loss of large myelinated fibers at higher cisplatin doses (Krarup-Hansen et al., 2007). The symptoms of cisplatin-induced neuropathy may persist for several months and can progressively worsen over time, a phenomenon called “coasting” (Siegal and Haim, 1990). With higher cumulative doses and longer times of exposure to cisplatin, the severity of CIPN increases, as does the likelihood of development of a chronic, irreversible neuropathy (Cersosimo, 1989; Gregg et al., 1992). Chronic cisplatin-induced neuropathy—the development of which appears to be independent of pre-treatment, vibration perception threshold, age, sex, tumor type or co-treatment with other chemotherapeutics—has been reported in approximately 5–20% of patients at 12 months (Hilkens et al., 1994; Hoskins et al., 1994; Bogliun et al., 1997; Schmoll et al., 2003; Park S. B. et al., 2013).

Several mechanisms contributing to cisplatin-induced neurotoxicity have been suggested (Figure 2), including the loss of peripheral sensory neurons, changes in cell signaling cascades, changes to calcium homeostasis and signaling, oxidative stress, mitochondrial dysfunction and induction of apoptosis as a result of DNA platination (Meijer et al., 1999).

Treatment with oxaliplatin induces various side effects including myelotoxicity and peripheral neuropathy, albeit it lacks the ototoxic and nephrotoxic effects commonly observed after treatment with cisplatin. The neurotoxic effects of oxaliplatin include the development of an acute, transient neuropathy that occurs in almost 90% of patients within hours of infusion and is characterized by dysesthesias and paresthesias of the hands, feet and the perioral region. These symptoms are often induced by exposure to cool temperatures and are a key feature of oxaliplatin-induced cold allodynia (Argyriou et al., 2013a). In addition, motor symptoms including tetanic spasms, fasciculations, and prolonged muscular contractions commonly occur (Saif and Reardon, 2005). Acute oxaliplatin-induced neuropathy generally subsides between treatment cycles (Extra et al., 1998), although continued exposure to oxaliplatin can lead to the development of a severe chronic neuropathy. The incidence of chronic peripheral neuropathy following oxaliplatin treatment has been estimated as approximately 70%, with development of the condition usually occurring at cumulative doses exceeding 540 mg/m² (Cersosimo, 2005; Argyriou et al., 2013a). Clinically, the symptoms of chronic oxaliplatin-induced neuropathy closely resemble those of the acute condition and include temperature-insensitive paresthesias, hypoesthesias and dysesthesias of the hands and feet. Additionally, changes in proprioception, which may affect normal daily activities requiring fine motor coordination, reportedly occur at cumulative doses exceeding 780 mg/m² (Cersosimo, 2005; Saif and Reardon, 2005).

Oxaliplatin-induced central neuropathy is rare and is characterized by Lhermitte's sign (an electric sensation experienced with flexing of the neck), proprioception deficiencies and urinary retention (Cersosimo, 2005). Risk factors for the development of both acute and chronic neuropathy include the cumulative dose, low body weight, a body-surface area >2.0, young age, persistent neuropathy in a past cycle, and variations in genes such as glutathione-S-transferase genes P1 [GSTP1] and glutathione-S-transferase genes M1 [GSTM1], as discussed below (Saif and Reardon, 2005; Alejandro et al., 2013). Furthermore, oxaliplatin induced neuropathy may be exacerbated by surgery (Gornet et al., 2002).

The mechanisms contributing to development of oxaliplatin-induced CIPN (Figure 2) include alteration in axonal excitability due to ion channel dysfunction, dysregulation of calcium homeostasis and altered function of transient receptor potential channels. Additionally, oxidative stress leading to neuronal and glial cell dysfunction and cell death induced by caspases, mitogen-activated protein kinases and protein kinase C may contribute to development of oxaliplatin-induced CIPN (Carozzi et al., 2015).

Despite widespread use of vincristine for the treatment of pediatric cancers including acute lymphoblastic leukemia, sarcoma, medulloblastoma and neuroblastoma as well as a range of tumors in adults, the epidemiology of vincristine-induced peripheral neuropathy (VIPN) is relatively poorly defined. While the reported incidence rates vary considerably—depending on the severity of symptoms, diagnostic criteria and concomitant use of other neurotoxic agents—most patients receiving vincristine develop some degree of sensory peripheral neuropathy at a cumulative dose of >4 mg/m² (Ramchandren et al., 2009; Toopchizadeh et al., 2009; Seretny et al., 2014; Lavoie Smith et al., 2015; Tay et al., 2017).

Symptoms of neuropathy can be present soon after initiation of vincristine therapy but usually develop within several weeks, and generally worsen with increasing cumulative doses of the drug. In addition, genetic factors can enhance the susceptibility to develop vincristine-induced neuropathy. An increased severity of VIPN was also observed in older compared with younger children, suggesting that age can be a risk factor for development of VIPN (Lavoie Smith et al., 2015).

VIPN can be divided into sensory, motoric and autonomic components, with tendon reflexes, vibration sensitivity and strength most affected in the first year of treatment (Lavoie Smith et al., 2015). Vincristine-induced sensory neuropathy is characterized by numbness, tingling and neuropathic pain in the upper and lower extremities. In addition, patients receiving vincristine experience loss of sensory discrimination, specifically an inability to detect light touch, pinprick sensations or vibration, and an inability to differentiate between hot and cold temperatures (Barton et al., 2011). Vincristine-induced motor neuropathy is characterized by weakness in the upper and lower extremities and the development of wrist- or foot-drop due to impaired dorsiflexion that arises from damage to peripheral nerves. This is accompanied by gait abnormalities, cramps, and loss of or reduction in deep tendon reflexes which can be severe (Mora et al., 2016). Typical symptoms of autonomic neuropathy are constipation, urinary retention, and orthostatic hypotension. As would be expected, these symptoms can significantly reduce the quality of life of these patients (Lavoie Smith et al., 2013).

The mechanisms contributing to development of VIPN include disruption of calcium homeostasis, activation of the immune system and subsequent neuroinflammation, membrane remodeling of peripheral neurons and loss of large myelinated fibers (Figure 3) (Devor, 2006; Boehmerle et al., 2014; Carozzi et al., 2015).

CIPN is also a dose-limiting adverse effect of treatment with taxanes, particularly paclitaxel, and occurs in a dose- and treatment duration-dependent manner. The threshold dose for development of taxane-induced peripheral neuropathy lies close to standard doses used in a range of chemotherapy regimens at approximately 300 mg/m² for paclitaxel, and 100 mg/m² for docetaxel (Forsyth et al., 1997; Winer et al., 2004; Park S. B. et al., 2013). Accordingly, this symptom is very common, occurring in as many as 90% of patients, although it usually remains relatively mild until cumulative paclitaxel doses exceed 1,400 mg/m² (Lipton et al., 1989; van Gerven et al., 1994; Pace et al., 2007). Overall, while paclitaxel is slightly less potent than docetaxel, it is more commonly associated with the development of CIPN (Hilkens et al., 1997; Chon et al., 2009). Consistent with CIPN being a direct consequence of the pharmacological effects of the taxanes on sensory neurons, higher single as well as higher cumulative doses are associated with both a greater risk for development of this side effects, as well as increased severity of symptoms (Pace et al., 2007; Baldwin et al., 2012; Ghoreishi et al., 2012). In addition, concomitant treatment with other neurotoxic chemotherapeutic agents, or pre-existing neuropathies, may additionally increase the risk of developing CIPN (Chaudhry et al., 2003). While symptoms usually develop within several weeks of treatment initiation, paclitaxel and docetaxel can also induce an acute painful neuropathy—peaking approximately 3 days after infusion—that is characterized by pain, numbness and tingling, and which can be a predictor for the development of chronic neuropathy (Loprinzi et al., 2007, 2011; Reeves et al., 2012; Tanabe et al., 2013; Fernandes et al., 2016).

Symptoms of taxane-induced neuropathy are consistent with a predominantly sensory neuropathy, although motor effects (particularly distal weakness, muscle cramps, and muscle aches) and autonomic dysfunction (including arrhythmias and orthostatic hypotension) can occur with higher doses. Specifically, numbness, tingling, mechanical allodynia, and neuropathic pain developing symmetrically in the digits and extending to the extremities usually predominate; cold allodynia, loss of pinprick sensation as well as altered reflexes can also occur (Dougherty et al., 2004; Park et al., 2011; Tofthagen et al., 2013).

Mechanisms that contribute to paclitaxel-induced neuropathy include immune-mediated processes, loss of peripheral fibers, demyelination and axon degeneration, altered retrograde and anterograde transport as well as mitochondrial dysfunction (Figure 4).

Mitochondria are small organelles involved in many important cellular processes including energy production, storage of intracellular calcium and calcium signaling, apoptosis, regulation of membrane potential and cell metabolism. The main function of mitochondria is to produce adenosine triphosphate (ATP) via aerobic respiration. In healthy tissues, mitochondria produce small amounts of reactive oxygen species (ROS) such as peroxide, superoxide, hydroxyl radicals and singlet oxygen as a by-product of oxygen metabolism. These radicals carry out important functions in cell signaling.

Most chemotherapeutic agents cause damage to neuronal and non-neuronal mitochondria, leading to increased production of ROS and thus to increased oxidative stress (Sangeetha et al., 1990; Look and Musch, 1994; Weijl et al., 1998; McDonald and Windebank, 2002). The pathological increase in ROS production in turn can cause damage to intracellular biomolecules such as enzymes, proteins and lipid molecules (Slater, 1984; Stadtman et al., 2003; Valko et al., 2005), which in turn leads to demyelination and disruption of the cytoskeleton of peripheral nerves as well as sensitization of signal transduction processes (Anderson et al., 1992; Zheng et al., 2011). Furthermore, ROS can cause the activation of apoptotic pathways (Cashman and Hoke, 2015) and increase production of pro-inflammatory mediators (Bulua et al., 2011). These processes can cause further damage to mitochondria, amplifying the production of ROS and pathological processes of oxidative stress (Areti et al., 2014).

Oxaliplatin and cisplatin bind to mitochondrial DNA (mDNA) and form mDNA adducts which cannot be repaired since mitochondria do not express DNA repair systems. These Pt-mDNA adducts impair mitochondrial DNA replication and transcription, leading to altered protein synthesis and functional errors of the respiratory chain (Canta et al., 2015). Similarly, vincristine causes dysregulation and structural modification of neuronal mitochondria, leading to activation of apoptotic pathways, alteration in neuronal excitability and dysfunction of glial cells. In contrast to platinum-induced mitochondrial dysfunction, the effects of vincristine on mitochondria likely involves altered mitochondrial Ca²⁺ signaling (Canta et al., 2015; Carozzi et al., 2015). This hypothesis is supported by the observation that potentiation of vincristine-induced cytotoxicity occurs in the presence of verapamil, a calcium channel blocker (Kaba et al., 1985).

Like vincristine, paclitaxel does not directly affect mitochondrial DNA, but nonetheless induces swollen and vacuolated mitochondria in both myelinated and unmyelinated sensory axons in the saphenous nerve (Flatters and Bennett, 2006). These changes in turn are accompanied by increased ROS production in sensory neurons and spinal cord, although the mechanisms leading to altered mitochondrial function are relatively poorly understood (Doyle et al., 2012; Areti et al., 2014; Griffiths and Flatters, 2015; Duggett et al., 2016).

Calcium (Ca²⁺) is a key regulatory ion in many cellular and physiological processes. Its free intracellular concentration, which is usually maintained at nanomolar levels, is tightly regulated by various transport and sequestration mechanisms, including extracellular influx and release from internal stores, as well as efflux via plasma membrane pumps and uptake into the endoplasmic reticulum and mitochondria (Siau and Bennett, 2006). Changes in the intracellular Ca²⁺ concentration influence membrane excitability, neurotransmitter release and gene expression of neuronal and glial cells (Carozzi et al., 2015). Accordingly, dysregulation of Ca²⁺ homeostasis and Ca²⁺ signaling has been suggested to contribute to the development of oxaliplatin-, cisplatin-, vincristine-, and paclitaxel-induced CIPN.

The oxaliplatin metabolite oxalate is a well-known Ca²⁺ chelator that has been proposed to contribute to the development of oxaliplatin-induced CIPN. Indeed, local injection of oxalate into the footpad of mice induces spontaneous nocifensive behavior as well as mechanical allodynia, albeit the dose required to observe this effect is considerably higher than the dose of oxaliplatin required to induce neuropathic pain (Deuis et al., 2013). In addition, the phenotype of pain behaviors induced by oxalate differs significantly from oxaliplatin-induced neuropathy, which is characterized by cold allodynia and a lack of spontaneous nocifensive behavior (Deuis et al., 2013). Oxalate-induced pain likely arises as a consequence of chelation of extracellular Ca²⁺ ions, which in turn leads to an increase in Na⁺ conductance and a decrease of threshold potential and membrane resistance (Deuis et al., 2013). In contrast, an increase in extracellular Ca²⁺ concentration increases the probability of Na⁺ channel closure and results in decreased excitability of peripheral neurons (Armstrong and Cota, 1999). Accordingly, the administration of Ca²⁺/Mg²⁺ prior to oxaliplatin infusion has been evaluated in several clinical trials as a strategy to prevent development of CIPN, although a consistent clinical benefit is unfortunately not apparent (Jordan et al., 2016).

In contrast, although effects on intracellular Ca²⁺ levels have been reported both in sensory neurons, renal tubular cells and cancer cells, the contribution of Ca²⁺ to cisplatin-induced neuropathy is relatively poorly understood. In sensory neurons, cisplatin increased the expression of the N-type voltage-gated Ca²⁺ channels (Ca_V2.2), although it differentially affects the function of specific Ca_V channel subtypes (Tomaszewski and Busselberg, 2007; Leo et al., 2017). However, the contribution of these effects to the clinical presentation of cisplatin-induced neuropathy remains unclear.

Contributions of Ca²⁺ signaling to the pathology of CIPN have also been reported for paclitaxel, which causes rapid mitochondrial depolarization and Ca²⁺ release in both neuronal and non-neuronal cells, possibly via activation of the mitochondrial permeability transition pore (mPTP) (Kidd et al., 2002; Mironov et al., 2005). In addition, lower concentrations of paclitaxel induce Ca²⁺ oscillations downstream of neuronal calcium sensor 1 (NCS-1), a protein that regulates G protein-coupled receptor phosphorylation in a calcium-dependent manner (Boehmerle et al., 2006).

Several studies in humans and animals have demonstrated axon degeneration after long-term administration of chemotherapeutic agents, including the loss of large myelinated, small unmyelinated (more rarely), and intra-epidermal nerve fibers (IENF) which may be connected to the development of sensory-motoric peripheral neuropathy (Bradley et al., 1970; Cavaletti et al., 1992; Sahenk et al., 1994; Bennett et al., 2011; Boyette-Davis et al., 2011, 2013; Boehmerle et al., 2014). Intra-epidermal nerve fibers are unmyelinated or thinly myelinated nociceptors located in the dermis and are necessary for the sensation of pain arising from the periphery. Additionally, loss of myelin and changes to the axonal cytoskeleton likely alters the structure and function of peripheral nerves, which in turn may contribute to development of altered perception. However, while correlation of nerve fiber loss with the degree of neuropathy has been attempted in several conditions, and functional deficits arising from axon degeneration appear intuitive, the contribution of demyelination and peripheral nerve degeneration to the pathobiology of CIPN is not entirely clear.

Clinical and electrophysiological studies have shown that oxaliplatin causes moderate sensory-motor axon degeneration and loss of intra-epidermal nerve fibers (Boyette-Davis and Dougherty, 2011). Similarly, electron microscopy of peripheral nerves from cisplatin-treated patients revealed axonal degeneration of large myelinated fibers as well as secondary myelin breakdown associated with loss of ankle jerks and decreased vibration sensitivity (Thompson et al., 1984). In mice, cisplatin damaged myelinated fibers of the sciatic nerve, diminished the action potential amplitude and reduced the nerve conduction velocity of the caudal sensory nerve (Boehmerle et al., 2014). Degeneration of distal sensory axons, secondary demyelination and nerve fiber loss have also been reported in vincristine- and paclitaxel-induced neuropathy (Sahenk et al., 1994; Topp et al., 2000; Argyriou et al., 2008; Boehmerle et al., 2014).

The molecular mechanisms leading to these observed changes in axonal function and structure remain unclear. Chemotherapeutics causing CIPN likely have direct toxic effects on axons—evidenced by a lack of effect on axon integrity when applied to neuronal cell bodies under compartmentalized culturing conditions (Yang et al., 2009)—although indirect effects due to altered gene expression cannot be ruled out. Altered microtubule function can also impair the anterograde and the retrograde axonal transport of synaptic vesicles loaded with lipids, proteins and ion channels. This in turn causes a length-dependent degeneration of axonal distal segments (Wallerian degeneration) and axonal membrane remodeling of peripheral nerves. However, these effects cannot adequately explain axon degeneration induced by cisplatin and oxaliplatin, which target cellular DNA. In addition, the different vinca alkaloids used in chemotherapy display variable efficacies and side effects but equally block tubulin polymerization, suggesting that additional mechanisms contribute to the direct damage of peripheral nerves (Himes et al., 1976).

In addition to axonal degeneration, chemotherapeutic agents also cause changes to peripheral nerve excitability that contribute to the development of sensory peripheral neuropathy. These are likely caused by altered expression and function of a range of ion channels—including voltage-gated sodium (Na_V), voltage-gated potassium (K_V) and transient receptor potential (TRP) channels.

Changes in sensory nerve excitability in patients treated with oxaliplatin, including a significant increase in the duration of the relative refractory period, have been attributed to effects on Na_V channels expressed at the nodes of Ranvier (Krishnan et al., 2005). Altered Na_V channel function was also observed in rodent peripheral axons as well as dorsal root ganglion neurons, where oxaliplatin causes an increase in Na⁺ current, inhibition of maximal amplitude, and the emergence of enhanced resurgent and persistent current amplitudes (Adelsberger et al., 2000; Sittl et al., 2012). Specifically, the Na_V channel isoform Na_V1.6 appears to be involved in the development of oxaliplatin-induced cold allodynia, with cooling-induced bursts of action potential firing abolished in neurons from Scn8a(med/med) mice lacking functional Na_V1.6 (Sittl et al., 2012). Furthermore, acute oxaliplatin-induced cold pain behaviors were abolished by treatment with a selective Na_V1.6 inhibitor, as was cisplatin-induced mechanical allodynia, albeit cisplatin does not directly affect the gating properties of Na_V1.6 (Deuis et al., 2013, 2014).

Effects on neuronal potassium (K⁺) channels—of which four major groups [voltage-gated K⁺ channels (K_V), Ca²⁺-activated K⁺ channels (K_Ca), two-pore K⁺ channels (K_2P) and inwardly-rectifying K⁺ channels (K_ir)] are expressed in peripheral nerves—further exacerbate altered neuronal excitability in CIPN. Specifically, decreased expression of several K⁺ channels, including the K_2P channels TREK1 and TRAAK, were found after oxaliplatin and paclitaxel treatment in rodent dorsal root ganglion neurons (Descoeur et al., 2011; Zhang and Dougherty, 2014). In addition, broadening of the repolarization phase, repetitive firing and after-hyperpolarization, consistent with effects on K_V channels, were induced by treatment with oxaliplatin in the isolated sciatic nerve of the adult rat (Kagiava et al., 2008).

In addition to acute effects on sensory neuron excitability, exposure to oxaliplatin and cisplatin has also been reported to alter expression of several thermo- and mechano-sensitive TRP channels, including TRPV1, TRPA1, and TRPM8 (Gauchan et al., 2009; Ta et al., 2010; Descoeur et al., 2011; Nassini et al., 2011; Schulze et al., 2011; Yamamoto et al., 2015; Mizoguchi et al., 2016). The reported behavioral contributions of these channels to CIPN, however, remain seemingly at odds, with studies attributing functional effects to all, some, or none of these transducer ion channels. Whether these discrepancies arise from insufficient subtype selectivity of compounds with activity at TRP channels, compensatory expression changes in knockout animals, or difference between animal models, remains to be determined. Paclitaxel- and vincristine-induced neuropathy has also been attributed to activation of TRPA1 and TRPV4 via the generation of reactive oxygen species (Alessandri-Haber et al., 2008; Materazzi et al., 2012; Old et al., 2014), albeit disruption of TRP channel function due to altered association with microtubules may be an additional contributing mechanism (Goswami, 2012).

Chemotherapy agents are well known to cause profound effects on the immune system, most notably a transient immunosuppression due to inhibition of myeloproliferation. However, activation of the immune system by chemotherapeutics has increasingly received attention as an effect that is thought to support the destruction of tumor cells (Zitvogel et al., 2008; Westbom et al., 2015), but which may also lead to neuroinflammation and thus contribute to the development of CIPN. Specifically, effects of chemotherapeutics on the innate (Kiguchi et al., 2008; Liu et al., 2014; Li et al., 2016) and adaptive immune system (Zhang et al., 2008; Zhu et al., 2011; Krukowski et al., 2016), as well as effects on peripheral and central neuronal accessory cells—including satellite glial cells (Peters et al., 2007; Kiya et al., 2011; Warwick and Hanani, 2013), Schwann cells (Cavaletti et al., 1995), astrocytes (Zhang et al., 2012; Robinson et al., 2014) and microglia (Burgos et al., 2012; Ruiz-Medina et al., 2013)—have been observed, although most of these studies were carried out using rodent models (Lees et al., 2017).

Consistent with a contribution of the immune system to CIPN, oxaliplatin- and paclitaxel-induced mechanical hyperalgesia and epidermal nerve fiber loss was prevented by the tetracycline minocycline, an antibiotic known to inhibit macrophages/monocytes and microglia (Raghavendra et al., 2003; Liu et al., 2010; Boyette-Davis and Dougherty, 2011; Di Cesare Mannelli et al., 2014). Both chemotherapeutic agents also lead to increased levels of proinflammatory cytokines (IL-6, IL-8, IL1β, TNF-α), which can lead to sensitization of nociceptors (Ledeboer et al., 2007; Loprinzi et al., 2007, 2011; Callizot et al., 2008; Doyle et al., 2012; Wang et al., 2012; Zhang et al., 2012, 2016; Pevida et al., 2013; Janes et al., 2014a,b, 2015; Li et al., 2015; Makker et al., 2017). In addition, oxaliplatin increases the levels of circulating CD4+ and CD8+ lymphocytes in mice and down-regulates regulatory T (T-reg) cells (Makker et al., 2017).

Vincristine induces the expression of integrins (immune markers) on the surface of endothelial cells which allows macrophages expressing the CX3CR receptor to adhere to the endothelium and to migrate into nervous tissue. Activation of monocyte-macrophages by the chemokine CX3CL1 also lead to production of ROS and subsequent activation of TRPA1 (Old et al., 2014). Additionally, vincristine and paclitaxel enhanced the binding of the STAT3 (Signal Transducer and Activator of Transcription 3) to the CXCL12 gene promotor (Xu et al., 2017), causing up-regulation of C-X-C Motif Chemokine Ligand 12 in dorsal horn ganglia. CXCL12 is a member of the integrin family and acts as a ligand of the CXCR4 (CD184, C-X-C chemokine receptor type 4) and as an attractant for T-lymphocytes and monocytes. The activation of CXCR4 receptors in turn leads to an increase in intracellular Ca²⁺ and chemotaxis of immune cells to the site of inflammation (Chatterjee et al., 2014). Consequently, activation of the immune system, recruitment of immune cells and neuroinflammation should be considered as a putative mechanism contributing to the development of vincristine-induced neuropathy specifically, and CIPN more broadly.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.