Neuroimmunology of cancer and associated symptomology

Abstract

Evolutionarily, the nervous system and immune cells have evolved to communicate with each other to control inflammation and host responses against injury. Recent findings in neuroimmune communication demonstrate that these mechanisms extend to cancer initiation and progression. Lymphoid structures and tumors, which are often associated with inflammatory infiltrate, are highly innervated by multiple nerve types (e.g. sympathetic, parasympathetic, sensory). Recent preclinical and clinical studies demonstrate that targeting the nervous system could be a therapeutic strategy to promote antitumor immunity while simultaneously reducing cancer-associated neurological symptoms, such as chronic pain, fatigue and cognitive impairment. Sympathetic nerve activity is associated with physiological or psychological stress, which can be induced by tumor development and cancer diagnosis. Targeting the stress response through suppression of sympathetic activity or activation of parasympathetic activity has been shown to drive activation of effector T cells and inhibition of myeloid-derived suppressor cells within the tumor. In addition, there is emerging evidence that sensory nerves may regulate tumor growth and metastasis by promoting or inhibiting immunosuppression in a tumor-type specific manner. Because neural effects are often tumor-type specific, further study is required to optimize clinical therapeutic strategies. This review examines the emerging evidence that neuroimmune communication can regulate antitumor immunity as well as contribute to development of cancer-related neurological symptoms.

INTRODUCTION

Local and systemic neuroimmune interactions serve as key regulators of homeostasis. Humoral signaling enables the central nervous system to coordinate systemic immune responses while local innervation of lymphoid organs^1–3 and associated vasculature support more refined regulation.⁴ Given that neuroimmune crosstalk plays an important role in maintaining homeostasis, researchers have begun to examine how this communication impacts cancer. Many solid tumors are highly innervated (more so than the healthy organ)⁵ and are associated with a large inflammatory infiltrate.⁶ Anatomical studies indicate that within some solid tumors, immune cells aggregate in close proximity to nerves.⁷ Recently, the fields of neuroscience and immunology/oncology have made a concerted effort to elucidate the role of neuroimmune crosstalk in tumor development. Historically, investigations largely focused on the mechanistic underpinnings of neurological symptoms. Behavioral comorbidities including pain, sensorimotor deficits, depression, anxiety, fatigue and cognitive disturbances are prevalent in cancer patients.^8–17 This review summarizes findings about the role of neuroimmune communication in both tumorigenesis and the development of cancer-related changes in neurological processes (i.e. cancer symptoms).

NEUROIMMUNOLOGY OF CANCER

Models of tumor stroma often exclude nerve endings and little attention has been paid to the role of nerves in the development of tumors that originate outside of the nervous system. However, many solid tumors are highly innervated by multiple types of nerves (e.g. sympathetic, parasympathetic, sensory).⁵ The prevailing hypothesis underlying tumorigenesis is that cancer cells exploit immunological processes to induce an immunosuppressive environment.¹⁸ Specifically, tumors evade immune surveillance by subduing effector cells and promoting myeloid-derived suppressor cells (MDSCs). It has now been shown that nerve activity plays an important role in regulating tumor growth in a nerve-type specific manner,¹⁹ often through control of antitumor immunity (Figure 1).

Figure 1.

The cancer–neuroimmune network. The tumor microenvironment contains malignant cells, subsets of effector and regulatory immune cells and the peripheral nervous system composed of three different populations of peripheral nerves (e.g. sympathetic, parasympathetic and sensory). The representative schematic of communication between the peripheral nervous system and tumor-associated immune cells including myeloid-derived cells [i.e. myeloid-derived suppressor cell (MDSC), macrophage, monocyte, neutrophil] and lymphocytes [i.e. natural killer (NK) cell, CD8⁺ T cell] is based on recent findings. Selective manipulation of neurotransmission alters the number and/or activity of the indicated immune target in the direction indicated by the arrow. Cell communication can be either direct or mediated by cytokines, chemokines, growth factors and checkpoint proteins [i.e. tumor necrosis factor alpha (TNF-α), programmed cell death-1 (PD1)]. Intratumoral neurotransmitters and neuropeptides are known to regulate cancer progression by their direct effects on malignant cells and the modulation of tumor-infiltrating immune cell function. Activation/blockade of neurotransmission elicits protumor and antitumor immune responses in a nerve-type and tumor-type-dependent manner.

Sympathetic nervous system

The sympathetic nervous system and hypothalamic–pituitary–adrenal axis are responsible for regulating the body’s physical response to stressful stimuli. The primary sympathetic neurotransmitters epinephrine and norepinephrine bind to neuronal adrenergic receptors driving the “flight or fight” response. Adrenergic receptors are also expressed by immune cells enabling direct as well as indirect regulation of both neurons and immune cells.²⁰ Adrenergic inhibitors can be used for the treatment of psychological (e.g. anxiety) and physiological (e.g. hypertension) stress.^21–23 The utility of adrenergic inhibitors as adjuvant cancer therapy has been a hot topic because epidemiological studies suggest that beta-adrenergic receptor blocker use is associated with prolonged survival; the greatest increases in cancer-specific survival rates have been observed with melanoma, pancreatic, ovarian and breast cancers.^24–26 The degree of improvement varies with not only tumor type, but also the receptor subtype that is targeted. For instance, type 2, but not type 1, beta-adrenergic receptor antagonists enhance responsiveness to immunotherapy.²⁶

Sympathetic nerve ablation or pharmacological (e.g. beta-blockers) inhibition of adrenergic signaling reduces stress signaling and tumor growth. In some cases, this loss of sympathetic nerve activity has been linked directly to changes in immune responses. Systemic administration of 6-hydroxydopamine was used to ablate sympathetic nerves in Wistar rats prior to inoculation with hepatoma cells; while chemical sympathectomy did not alter tumor incidence or survival, it was associated with a systemic (i.e. blood) shift in the neutrophil-to-lymphocyte ratio as well as increased tumor necrosis factor-alpha (TNF-α), suggesting increased extratumoral mobilization of innate immune cells.²⁷ By contrast, treatment with a beta-blocker, propranolol, was sufficient to slow tumor growth and metastasis in multiple models of melanoma, breast and pancreatic cancers.^28–30 Here, the authors analyzed tumor-infiltrating leukocytes (TILs) and found that blockade of beta adrenergic signaling was associated with a significant decrease in infiltration of myeloid cells (e.g. MDSCs), and a concomitant increase in the cytotoxic lymphocytes.^28–30 Given the shift in TIL profile, it may not be surprising that adrenergic signaling blockade exhibits synergy with immunotherapy. Propranolol reduces the expression of programmed cell death-1 (PD1) on CD8⁺ T cells, which results in improved responses to anti-PD1 immunotherapy.²⁹ The monoclonal antibody trastuzumab is used to treat HER2-positive breast cancers, but treatment resistance is common.³¹ Liu et al.³¹ found that propranolol increases antitumor activity by resensitizing tumor cells to trastuzumab. Beta-blockade has also been tested in models for hemopoietic malignancies. The stress neurohormones epinephrine, prostaglandin and corticosterone are thought to promote tumor growth via their suppression of natural killer (NK) cell activity.³² In the CRNK-16 rat model of leukemia, individual or combination therapy with nadolol, a nonspecific beta-blocker, and indomethacin, a COX inhibitor, significantly prolonged survival.³² Thus, suppression of sympathetic nerve activity promotes local antitumor immunity in multiple tumor models and exhibits potential for synergistic combination therapy regimens. Given that both adrenergic receptors are expressed on both neurons and immune cells,²⁰ future studies will be required to determine the exact cell types governing adrenergic-dependent changes in tumor immunity.

Several clinical trials have utilized pharmacological approaches to establish the effect of beta blockers on response to conventional therapies in breast, ovarian and colorectal cancers. In addition to psychological stress associated with a cancer diagnosis, primary interventions such as surgery or chemotherapy can drive physiological stress that induces secretion of prostaglandins and catecholamines such as epinephrine/norepinephrine, all of which promote immunosuppression and metastases as well as reduce survival.^33–36 In a phase II randomized clinical trial, women with early stage breast cancer undergoing surgical resection were given 11 days of perioperative placebo or propranolol combined with the COX-2 inhibitor etodolac.^37,38 Transcriptomic analyses of the resected tumor indicated that the drug treatment was associated with reduced circulating levels of proinflammatory mediators [e.g. interleukin (IL)-6, C-reactive protein, interferon-γ] and increased anti-inflammatory mediators (e.g. IL-10).^37,38 This dual drug treatment was associated with shifts in the TIL profile, including decreased monocytes, increased NK cell activity and an increase in B cells.³⁸ Interestingly, similar shifts in monocytes and NK cells were observed in colorectal cancer patients given perioperative propranolol plus etodolac, but there was a decrease in B-cell infiltration.³⁹ The effect of propranolol alone has been tested in the adjuvant setting for patients with epithelial ovarian cancer undergoing either chemotherapy or surgery.⁴⁰ Propranolol was associated with improved mood as well as significant reductions in proinflammatory IL-6 and IL-8 which correlated with response to therapy.⁴⁰ Propranolol has also been assessed for effects on biomarkers in the treatment of hemopoietic malignancies. In a phase II trial of patients receiving hematopoietic cell transplantation, transcriptomic analyses suggest that adjuvant/neoadjuvant propranolol results in greater downregulation of transcripts associated with monocyte activation and a shift toward a CD34 stem/progenitor cell profile.⁴¹ While these small clinical trials show promising results, additional studies are needed to parse out the impact of sympathetic nerve inhibition at various disease stages and across different tumor types.

The addition of behavioral stress paradigms to preclinical models has been used to mimic the psychological stress associated with the diagnosis of cancer. Psychological stress is commonly induced through behavioral stressors (e.g. chronic restraint stress) which elevate circulating levels of epinephrine, norepinephrine and glucocorticoids as well as promote anxiety- and depression-like behaviors.^{32,42,43,44,45,46,47,48,49} Multiple studies have shown that these behavioral stressors are sufficient to accelerate tumor progression, dissemination and metastasis of a diverse range of cancers, including melanoma,⁴³ glioma,⁴⁸ leukemia,^32,46 pancreatic,^45,47 breast⁴² and prostate cancer.⁴⁴ In many cases, treatment with beta blockers was sufficient to reverse the effects of behavioral stressors by preventing immunosuppression. Using a chronic unpredictable stress model, Chen et al.⁵⁰ showed that chronic unpredictable stress or isoproterenol (beta adrenergic agonist) promotes more severe colonization of the lung by breast cancer cells. Stress-induced lung metastasis caused upregulation of the monocyte chemoattractant CCL2, resulting in an accumulation of monocytes and macrophages; depletion of monocytes/macrophages or propranolol prevented stress-induced metastasis.⁵⁰ Isolation or overcrowding can also cause stress in rodents. Both stressors are associated with significant decreases in the size of lymphoid organs such as thymus in melanoma-bearing mice; the shift in lymphoid organ size was prevented by administration of propanolol.⁴³ Initially, propranolol also retarded tumor growth in the overcrowded mice.⁴³ Social or group housing partially abrogates tumor-induced depressive behaviors in tumor-bearing mice.⁴⁹ Furthermore, when a social hierarchy is established in the group housing setting, submissive mice exhibit elevated anxiety and stress. Submissive melanoma-bearing mice have increased metastases to lung, decreased NK cell cytotoxicity and decreased basal oxidative burst in neutrophils and monocytes compared with their dominant cagemates⁵¹; this is correlative and a direct link between the elevated stress response, sympathetic nervous system (SNS) activation and the immune system has yet to be established in this context.

Physiological environment has also been implicated in regulation of the stress response. For example, a deviation from thermoneutral ambient temperature is sufficient to drive stress as measured by elevated plasma norepinephrine.^29,30 Thermally induced stress accelerates melanoma, pancreatic and mammary tumor growth.^29,30,52 Maintaining mice at thermoneutral temperatures reduces plasma epinephrine as well as lowers stress and anxiety behaviors. Housing mice at thermoneutral temperatures was sufficient to increase intratumoral effector T cells (e.g. CD8⁺) and reduce the number of immunosuppressive MDSCs and regulatory T cells.^29,52 Furthermore, intratumoral effector cells exhibited increased markers of activation and reduced expression of exhaustion markers (e.g. PD1).^29,52 The effects of temperature were occluded in beta-2-adrenergic receptor knockout mice, implicating the sympathetic signaling pathway in the regulation of adaptive tumor immunity. Thermal treatments (e.g. weekly hyperthermia, localized heat source) for mice housed at subthermoneutral temperatures were sufficient to reduce the accumulation of MDSCs to a similar extent as propranolol.⁵³ Given the important role of physiological stress and sympathetic nerve activity in antitumor immunity, it may be possible to leverage this system to improve therapeutic responsiveness in patients.

Stress reduces cancer patients’ quality of life and may independently influence prognosis.⁵⁴ While the studies discussed earlier suggest that pharmacological interventions may be useful in the adjuvant setting, it is important to note that nonmedical interventions such as mindfulness, meditation, yoga and other psychosocial strategies also reportedly improve quality of life and prognosis.⁵⁵ The effects of such lifestyle changes on immune responses in cancer, however, have not been characterized. It is known that exercise can mobilize cytotoxic T and NK cells.⁵⁶ Thus, in addition to pharmacological manipulations, future studies should consider nonmedical interventions that reduce sympathetic nerve activity as a potential means to promote antitumor immunity.

Parasympathetic nervous system

The primary neurotransmitter of parasympathetic neurons is acetylcholine, which activates muscles, glands and other organs. The most common approach to study parasympathetic activity is through manipulation of the vagus nerve (e.g. transection, electrical stimulation) or by targeting cholinergic signaling pharmacologically. Vagotomy has been shown to inhibit prostate⁵⁷ and stomach⁵⁸ tumor growth and metastasis. Conversely, vagotomy has been shown to promote pancreatic ductal adenocarcinoma^59,60 as well as colon^61,62 and breast cancer.⁶³ While the effects of manipulating the vagus nerve have been attributed to changes in parasympathetic activity, it is well known that parasympathetic axons make up only 15–20% of the axons traveling in the vagus nerve; the majority are sensory fibers.^64–66 In transplant, chemical and transgenic models of breast cancer, viral expression of a bacterial voltage-gated sodium channel or diphtheria toxin subunit A was used to drive activation or ablate parasympathetic neurons, respectively.⁶³ Bacterial voltage-gated sodium channel-dependent increases in parasympathetic activity reduced metastasis while diphtheria toxin subunit A-mediated ablation accelerated tumor growth and metastasis.⁶³ Given that parasympathetic activity appears to exert opposing effects depending on the tumor type, it is important to determine whether the effects of vagotomy and/or vagus nerve stimulation on tumor responses are due purely to the small percentage of parasympathetic axons or whether there is also a sensory nerve component.

In addition to modulating primary tumor growth, parasympathetic activity has been implicated in the process of tumor immunity. One study demonstrated that vagal nerve stimulation upregulated splenic Tff2, a memory T-cell derived anti-inflammatory peptide, in a CXCR4-dependent manner; Tff2-dependent blockade of immunosuppression has been shown to slow colorectal tumorigenesis.⁶¹ It should be noted that it is unclear whether the stimulation parameters were selectively activating only parasympathetic fibers. In mammary tumors, genetically induced parasympathetic activation reduced the expression of the immunosuppressive markers PD1 and Foxp3 on CD4⁺ T cells and PD1 on CD8⁺ T cells.⁶³ Furthermore, activation of the vagal parasympathetic neurons via electroacupuncture in mammary tumors reduced proinflammatory cytokines and infiltration of MDSCs as well as increased proportions of CD8⁺ T cells and NK cells.⁶⁷ In pancreatic cancer models, vagotomy was associated with significantly elevated levels of TNF-α and increased numbers of tumor-associated macrophages.⁶⁰ Treatment with bethanechol, a selective muscarinic acetylcholine receptor agonist, suppressed infiltration of myeloid cells and TNF-α levels, promoting antitumor immunity and reducing liver metastases.⁵⁹ However, another recent study on pancreatic cancer showed that the nonselective agonist acetylcholine inhibited CD8⁺ T-cell recruitment, interferon-γ production and promoted Th2 responses, thereby fostering immunosuppression; in this case vagotomy promoted recruitment of CD8⁺ T cells and improved survival.⁶⁸ Despite these conflicting findings, bethanechol is currently in clinical trials as adjuvant therapy for pancreatic cancer (NCT03572283). Currently, cholinergic agonists are Food and Drug Administration approved for treatment of elevated ocular pressure and urinary retention.

Sensory nervous system

The sensory nervous system has been considered largely in the context of cancer symptoms such as paraneoplastic syndromes and sensorimotor deficits.^8,9 Sensory nerves innervating the body below the neck arise from spinal dorsal root ganglia and the nodose/jugular ganglia (vagal). Afferent signals transduce information to the central nervous system, but sensory neurons also have an important efferent component. For example, release of neurotransmitters and neuropeptides such as substance P (SP) and calcitonin gene related peptide (CGRP) significantly contribute to neurogenic inflammation and pain.⁶⁹ Several denervation studies have provided evidence that sensory nerves promote pancreatic ductal adenocarcinoma and basal cell carcinoma growth (reviewed in Saloman et al.¹⁹). Sensory denervation prior to pancreatic tumor development significantly reduced the size of the inflammatory stromal compartment, which includes infiltrating leukocytes.⁷⁰ Unfortunately, no direct analysis of immune cells was done in this study to examine the phenotype of TILs. However, it has been shown that SP increases cytotoxic immune responses to colorectal cancer cells in vitro.⁷¹ With only few studies done so far, it is difficult to conclude whether neuroimmune crosstalk is responsible for the effects of denervation on primary tumor growth and whether this is a targetable pathway for therapeutic interventions.

Perineural invasion (PNI) by tumor cells as well as hypertrophy and axonogenesis of intratumoral nerve fibers has been documented in peripheral nerves; both are associated with reduced quality of life, increased metastasis and poorer overall prognosis.⁷² Anatomical studies from resected pancreatic tumors show that these intratumoral nerves are prominently surrounded and infiltrated by immune cells.^7,73 Furthermore, transcriptomic analyses of resected pancreatic tumors show that the degree of PNI is associated with reduced expression for cytotoxic T cells and increased expression of Th2 signatures.⁶⁸ With respect to sensory neuron–immune interactions, in vitro studies have demonstrated that PNI is dependent on immune cells. While most of the clinical work on PNI does not identify what “type” of nerve is involved, many preclinical studies use sensory nerves as the model. These studies incorporated dorsal root ganglia or sciatic nerve into the classic tumor cell migration assay. Prior to the formation of pancreatic tumor cell–sensory neuron contacts, endoneurial macrophages are recruited and release glial cell derived neurotrophic factor (GDNF) that acts as a chemoattractant for the tumor cells to migrate to the sciatic nerve.⁷³ Neonatal ablation of sensory afferents systemically reduces primary pancreatic tumor incidence (metastasis was not directly studied)^19,70; however, systemic ablation and selective ablation of vagal sensory afferents result in increased breast cancer metastasis.⁷⁴ This suggests that vagal versus spinal sensory afferents may have differing effects on tumor immunity. Alternatively, the role of sensory afferents could be tumor specific, as has been reported for the vagus nerve. For example, blocking SP promoted antitumor immunity in mice bearing tumors derived from a brain metastasis by upregulating CD8⁺ and CD4⁺ T cells.⁷⁵ However, SP blockade in mice bearing tumors derived from a liver metastasis had a significant increase in MDSCs promoting an immunosuppressive environment.⁷⁵ Regardless, sensory-derived SP clearly regulates metastasis through modulation of TILs; thus, sensory neuropeptides may be targets for adjuvant therapy.

NEUROIMMUNE REGULATION OF GLIAL TUMORS

Immune interactions with peripheral nerves, including neurons and glia (e.g. satellite, Schwann), are an important focus area for tumors of non-neural origin; however, there is also emerging evidence that neuroimmune crosstalk regulates glial-derived tumors. Neurons induced from human pluripotent stem cells harboring mutations in Nf1 were found to secrete the chemokine midkine, which directly activates CD8⁺ T cells.⁷⁶ Canonically, activation of CD8⁺ T cells is associated with increased cytotoxic capacity, but here T-cell-derived CCL4 induced microglia to make growth factors that promote stem cell survival and glioma growth. Tumor cell expression of ligands that activate inhibitory immune checkpoint proteins is one of the major pathways thought to contribute to tumor-induced immunosuppression. However, cerebellar and cortical neuron expression of the canonical immune checkpoint protein PD-L1 is associated with increased survival in glioblastoma patients, suggesting that neuronal expression of canonical immune checkpoint proteins could contribute to regulation of antitumor immunity and tumor growth.⁷⁷ More studies are needed to assess the impact of nonimmune cell expression of “immune” proteins to elucidate whether this is a potential new therapeutic target.

NEUROIMMUNOLOGY OF CANCER-RELATED PAIN

Almost half of all cancer patients struggle with moderate to severe cancer-related pain,⁷⁸ which can arise throughout the disease process⁷⁹; the pain incidence and intensity can increase with the progression of the cancer.⁷⁸ While the etiology of cancer-related pain is debated, the common consensus is that cancer pain arises from processes involving crosstalk between tumor cells, the immune system and peripheral and central nervous system.^80–82

Primary tumor-related pain

An often-cited hypothesis regarding primary tumor-related pain is that tumor cells and immune cells in the cancer microenvironment produce mediators that activate and sensitize pain-sensing sensory neurons (i.e. nociceptors).⁸² TNF-α, a proinflammatory mediator released by both cancer and immune cells, has been implicated in nociceptive signaling in the presence of cancer. TNF-α antagonist, given systemically in a bone cancer pain mouse model, only partially reduced cancer-mediated mechanical hyperalgesia, indicating that TNF-α present in the tumor microenvironment, not circulating TNF-α, may contribute to tumor-related nociception. In an oral cancer mouse model, inhibition of TNF-α signaling reduced orofacial nociception induced by tongue cancer and decreased the number of CD3⁺ T cells infiltrating into the tongue⁸³; clinical trials using a monoclonal antibody against TNF-α (i.e. infliximab) demonstrated a similar decrease in epidermal CD3⁺ T cells in patients with moderate to severe psoriasis vulgaris,⁸⁴ but it is yet to be investigated clinically for cancer-related pain. Proteases, released from both primary tumor cells and immune cells (e.g. neutrophils and mast cells) in the tumor, bind and activate protease-activated receptor 2 (PAR2) on sensory neurons which also drives cancer pain.^85,86 In a murine breast cancer model,⁸⁷ proteases released after cell activation and degranulation evoked SP release in peripheral sensory nerves via PAR2 signaling, which subsequently contributed to sustained neuroinflammation and pain.⁸⁸ Lastly, nerve growth factor (NGF), which promotes the local growth and survival of neurons, is secreted by both cancer and cancer-infiltrating immune cells.^89,90 NGF can induce cancer pain via its association with PNI⁹¹; NGF has been demonstrated to promote PNI in mouse models of pancreatic⁹² and oral⁹⁰ cancer. In addition, NGF can drive axonal sprouting in sensory and sympathetic nerves within the tumor microenvironment and intratumoral space; anti-NGF significantly decreased bone cancer-mediated nociceptive behavior, as well as sensory and sympathetic nerve sprouting in tumor-bearing mice.⁹³

Inflammation within the peripheral nerve, called neuritis, arises from perineural immune cell infiltration into the nerve bundle. Most neuritis episodes in cancer-related studies have been studied in pancreatic cancer.⁹⁴ Nerve presence within the tumor microenvironment positively correlates with the presence and severity of pancreatic neuritis as well as with the degree of pain in patients.⁹⁴ Demir et al.⁷ have demonstrated that spatial interaction of peripheral nerves with mast cells may be at the root of the inflammatory response and hypersensitivity within the nerve. Furthermore, perineural mast cell numbers, but not macrophages or cytotoxic T cells, were increased in pancreatic patients with severe pain compared with those without pain.⁷

Metastasis-related pain

Many cancers, most commonly breast and prostate, will metastasize from their original site to the bone. Bone cancer is thought to be the most common cause of cancer-related pain, with the resulting pain thought to be a distinct type of persistent pain; the neurochemical changes documented in peripheral sensory neurons as well as the spinal cord are unique compared to the response to peripheral inflammation or nerve injury.⁹⁵ Cancer-induced injury to sensory nerve fiber terminals, release of inflammatory mediators and increased bone degradation all impact the development of bone cancer pain.⁹⁶ ATP, secreted from both cancer and immune cells, and P2X7 receptor upregulation in spinal microglia have been implicated in preclinical models of bone cancer pain; spinal P2X7 inhibition reduced cancer-related nociceptive behavior and blocked hyperactivity in the spinal cord.⁹⁷ Sigma-1 receptor is a modulator of calcium signaling on the endoplasmic reticulum; in a rat model of bone cancer pain, administration of sigma-1 antagonist, BD-1047, resulted in a decrease in the number of activated microglia in the spinal cord dorsal horn.⁹⁸ However, these studies were done prior to the establishment of pain and neuroinflammation and additional investigation is needed to determine the impact of sigma-1 antagonism after the onset of cancer-induce pain and neuroinflammation.

Osteoclasts, the bone’s innate immune cells, contribute to pain via inflammatory cytokine and chemokine production to create an acidic extracellular environment and increase bone resorption; this process ultimately leads to sensory and sympathetic nerve compression as well as bone fractures. Osteoclasts can be involved in several processes that lead to bone metastasis-related cancer pain, such as tumor growth and angiogenesis.⁹⁹ Recent preclinical and clinical studies have demonstrated that inhibition of osteoclast signaling via src, a nonreceptor protein tyrosine kinase linked to spinal glutamate excitatory neurotransmission, may be used to treat cancer-related bone resorption and metastasis.^100–102 In bone cancer pain rat model, src inhibitors reduced nociceptive behavior via reduced N-methyl-d-aspartate activity as well as inhibition of bone resorption in the tumor.¹⁰³ However, a recent randomized controlled trial of src inhibitor, saracatinib, did not show a significant effect on self-reported bone pain scores in patients with metastatic bone pain.¹⁰⁴

Immune-mediated endogenous analgesia

Endogenous opioid peptides consist of beta-endorphin, met- and leu-enkephalin, dynorphin and nociceptin¹⁰⁵; endogenous opioid peptides and their receptors are distributed in the central nervous system and peripheral nervous system and their contribution to antinociception has been extensively characterized.¹⁰⁶ Outside of the nervous system, endogenous opioids have been identified in non-neuronal peripheral tissues (e.g. leukocytes, keratinocytes) and found to impact nociceptive signaling.¹⁰⁷ For example, endothelin-1, an endogenous vasoconstrictor secreted by endothelin cells, contributes to endogenous analgesia by driving opioid secretion from keratinocytes.¹⁰⁸

While the immune system is best known for its contribution to inflammatory pain, there are instances in which leukocytes provide an endogenous opioid-based mechanism of pain control.¹⁰⁹ Lymphocytes, macrophages, neutrophils and mast cells have been found to contain beta-endorphin and met-enkephalin; minimal dynorphin has been detected in many immune subtypes.^110,111 The current understanding is that when stressed or when activated by specific inflammatory mediators (e.g. IL-1β, corticotropic factor), immune cells secrete opioids, which activate opioid receptors on sensory neurons and promote analgesia.¹¹² Cancer-recruited neutrophils make beta-endorphin and met-enkephalin, which are antinociceptive, while neutrophil depletion exacerbated nociception in a mouse oral cancer model.¹¹³ Endogenous opioid analgesia has a therapeutic benefit over exogenous opioids as it is restricted to the inflammatory site; thus, there are no side effects of opioid receptor activation in the central nervous system.¹¹⁴ These preclinical studies suggest that tumor-infiltrating immune cells can release opioids in the tumor microenvironment to inhibit cancer-related pain.^113,115 Further research is needed to investigate whether cancer-fighting immune cells evoked by recently developed immunotherapies for cancer treatment¹¹⁶ increase the endogenous opioid signaling within the tumor as well. However, a recent clinical study involving multiple different solid tumor cases found that severe pain correlated with high amounts of circulating beta-endorphin, and that oxycontin treatment reduced pain as well as beta-endorphin levels.¹¹⁷ Therefore, additional preclinical and clinical studies are needed to determine whether immune-mediated analgesia can be leveraged for the treatment of cancer-related pain.

NEUROIMMUNE–CANCER COMMUNICATION IN CANCER-RELATED FATIGUE AND OTHER SYMPTOMS

Tumor-associated biological processes have a direct impact on affective and cognitive symptoms. Preclinical and clinical studies have provided evidence that cancer-induced immune-secreted cytokines can induce behavioral comorbidities, including depression, fatigue, impaired sleep and cognitive dysfunction.¹¹⁸ Furthermore, activation of the SNS as a result of the psychological stress from cancer diagnosis, treatment and survivorship can lead to alterations in the immune system such as decreased NK cell cytotoxicity and decreased basal oxidative burst in neutrophils and monocytes.⁵¹

Fatigue and sleep

Cancer patients often report cancer-related fatigue (CRF), a multifaceted condition coupled with decreased muscle mass, weakness and depression. CRF can present at any time throughout the disease process; it can decrease quality of life, limit functionality and decrease overall survival in patients. Recent research suggests that activation of proinflammatory cytokines in the periphery can impact the brain, leading to fatigue and other behavioral changes.^119,120 Several clinical studies have shown an association between inflammatory markers and fatigue in cancer patients.¹¹⁹ For example, cytokine gene polymorphisms¹²¹ and human papilloma virus infection associated with head and neck cancer can influence inflammatory activity and fatigue.¹²² Immune cells recruited to the tumor microenvironment can produce proinflammatory cytokines. Shifts in circulating immune cell subtypes have also been noted in relation to CRF. There were more circulating T lymphocytes, specifically CD4⁺ and CD56⁺ effector T cells, in fatigued breast cancer survivors, compared with nonfatigued patients.¹²³ However, because of the fact that most fatigue-related studies focus on cancer survivors after treatment, it is uncertain whether the neuroimmune communication associated with fatigue is driven by cancer treatment (i.e. chemotherapy) or the cancer itself. Preclinical models have begun to demonstrate that changes in peripheral immune cells can influence behavioral responses in mice,¹²⁴ but not yet in the context of cancer.

Central nervous system-derived immune and inflammatory activity plays an important role in CRF. Tumors can influence neuroinflammatory processes, including alterations in brain proinflammatory cytokines.¹²⁵ Norden et al.¹²⁶ found that CRF and depression-like behaviors were correlated with increased IL-1β messenger RNA in the cortex and hippocampus. Furthermore, microgliosis has been implicated in the pathogenesis of CRF.¹²⁵ A microglial inhibitor (minocycline) reduced tumor-induced expression of IL-1β in the brain as well as CRF and depression-like behavior.¹²⁶ Although circulating inflammatory markers related to CRF have been extensively studied, additional data integrating in neural processes are needed to allow for a more thorough evaluation of biobehavioral contributors.

Sickness behavior, affective behaviors and cognitive impairment

Sickness behavior is characterized as lethargy, sleepiness, hyperalgesia and depressed mood, in response to an illness.¹²⁷ The driving force behind sickness behavior is thought to be multiple proinflammatory incidences in response to cancer, resulting in a flood of proinflammatory cytokines such as TNF-α, IL-1β and IL-6 in circulation¹²⁸; these peripheral cytokines then act on the brain through the vagus nerve or by crossing the blood–brain barrier, causing neuroinflammation.¹²⁹ In a mouse model of human papilloma virus-positive head and neck cancer, increased IL-6, IL-1β and TNF-α expression in the liver and increased IL-1β expression in the brain were found in tumor-bearing animals as well as deficits in burrowing behaviors; minocycline was able to attenuate tumor-induced inflammation and depression-like behaviors, suggesting that microglial activation and associated neuroinflammation play a significant role.¹³⁰

Dysregulated cytokine production associated with peripheral tumors can induce affective and somatic symptoms that can present months prior to and at diagnosis of cancer,^131,132 which is often consistent with the tumor development timeframe.^133,134 Prior to diagnosis, pancreatic cancer patients commonly report experiencing affective symptoms (e.g. crying, hopelessness) in addition to somatic symptoms (e.g. loss of appetite, insomnia).¹³⁵ A study examining medical records of ovarian cancer patients found that, in the 6 months prior to diagnosis, these patients are significantly more likely to report affective and somatic symptoms compared with controls.¹³⁶ Increased basal levels of brain proinflammatory cytokines were found mainly in the cortex and hippocampus in tumor-bearing rats with depression-like behavior and select cognitive impairments.¹²⁶ In addition, inflammatory insult, such as lipopolysaccharide, in a mammary tumor-bearing rat exacerbated the neuroinflammatory IL-1β signaling and microglial activation across various brain regions.¹³¹ However, increased peripheral inflammatory markers linked to somatic and affective symptoms appear to be tumor-type specific. In prostate cancer patients with nonmetastatic disease, elevated serum TNF-α is associated with fatigue, depression and anxiety.¹³⁷ By contrast, in a study of colorectal cancer patients, levels of TNF-α and C-reactive protein in plasma were not associated with baseline depressive symptoms.¹³⁸ Additional studies are needed to understand changes in subtypes of circulating immune cells from cancer patients prior to treatment which will allow for a broader view of the neuroimmune interaction beyond systemic inflammatory markers.

CONCLUSIONS/FUTURE DIRECTIONS

Until recently, it was thought that the nervous system’s role in cancer was to react to tumor growth and immunity. It is now evident that neuroimmune communication can modulate cancer progression. Manipulation of the nervous system is associated with stark changes in TILs and circulating inflammatory cytokines, which directly affect tumor growth and neuroplasticity as well as contribute to the development of the neurological symptoms reported by many cancer patients. While the field of neuroimmunology in cancer is still in its infancy, there is sufficient evidence to suggest that the points of interaction between these systems could potentially reveal novel therapeutic targets and strategies. Furthermore, inflammatory signaling molecules currently being studied as targets for the treatment of cancer (e.g. checkpoint proteins, type I interferons) should be considered in the context of neurological signaling prior to translation to clinic. Additional interdisciplinary collaboration is needed to identify the cell type of origin and the cell type targeted by individual signaling pathways. We can then take advantage of the biology to refine therapeutic interventions and minimize undesired side effects.