The Neuroimmune Axis in the Tumor Microenvironment

An important regulatory interface between the nervous system and the immune system during development, homeostasis, and disease has been well documented in both preclinical and clinical studies. The pathogenic role of the immune system in neuroinflammation, neurodegenerative disorders, and autoimmune neurologic disorders (autoimmune encephalopathy, dementia, epilepsy, and myelopathy) as well as the neuropathogenic pathways in autoimmune and autoinflammatory diseases, allergy, and immunodeficiencies is also well characterized (1–4). Moreover, aberrant bidirectional pathways between the CNS and the immune system have been implicated in several mood and developmental disorders (4–7). Overall, accumulating basic science and clinical studies reflect significant growth and advancement of neuroimmunology. Such efforts contribute to our understanding of the neuroimmune axis and its targetability in neurologic, psychiatric, and immunological diseases. However, less is known about the functions of the neuroimmune axis in the tumor milieu, and almost nothing is known about how the modulation of the neuroimmune cross-talk in cancer may alter tumor growth and progression.

Cancer is the cause of an estimated 10 million annual deaths worldwide, and it is the leading cause of death in the United States for those <75 y of age. Historically, cancer was thought to be initiated and promoted by the mutations of the key oncogenes. However, most genetic changes that are characteristic to epithelial cancers are already present in premalignant lesions, which rarely progress to cancer (8). The critical importance of the “cancerized” local tissue, which collaborates with the mutated epithelial cells to produce tumors and drive metastasis, is now widely recognized (9). The interaction between the malignant cells and the stromal and infiltrating cells forms the tumor microenvironment, which is a fundamental factor in tumorigenesis. Many cell types contribute to the generation of a cancer-supportive tumor microenvironment, including fibroblasts, endothelial cells, smooth muscle cells, epithelial cells, adipocytes, and cells of the immune system. The immunosuppressive and tolerogenic properties of the tumor microenvironment are strongly facilitated by cancer-associated fibroblasts, myeloid-derived suppressor cells (MDSC), T and B regulatory cells, regulatory dendritic cells, tumor-associated alternatively activated type 2 macrophages and type 2 neutrophils, and, potentially, a subset of mast cells (10). Nerves are also detected within solid tumors, but their impact on tumor growth and progression is not completely understood (11).

The first publication in this field was by Young (12), who, analyzing different tumors, noted that in 50% of tumors “the presence of nerves was positively demonstrated, and in some of them nerve fibers were present in considerable numbers.” He also concluded that detected intratumoral neurofilaments “represent nerves of normal structures which have been surrounded by the invading cancer cells” (12). Several follow-up studies concluded that tumors were innervated in the stroma, parenchyma, and along the blood supply and that the nerve supply to the tumor influenced its growth (13–16). Furthermore, the impact of the denervation on tumor growth was demonstrated more than half century ago (17). In contrast, others found that tumors lacked innervation or specific neuronal organization (18–20) and could not provide evidence of neurotrophic influence on neoplasms (21).

Recent data confirm the controversy of tumor innervation, revealing either a low, moderate, or abundant presence of neurons in solid tumors (11, 22, 23). Interestingly, several reports detected the formation of new nerve endings inside solid tumors, suggesting that tumors can stimulate their own innervation (neoneurogenesis or neoaxonogenesis) via the expression of neurotrophic factors, such as nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), semaphorins, netrins, and Slit proteins (24–26). Others provided strong evidence that solid tumors can injure and destroy nerves during expansion, activating neurodegenerative and nerve repair functions of glial cells at the tumor periphery (27). Tumor-induced reprogramming of neuroglia in this manner promoted cancer growth through the alterations in the extracellular matrix and the local immune microenvironment (27).

A large body of work describes the ability of malignant cells to spread along the nerves (perineural invasion) (28). Perineural, circumferential, or intraneural invasion is defined as the presence of cancer cells juxtaposed intimately along, around, or within a nerve. Nerve involvement in cancer is often used synonymously with neurotropism, and, traditionally, nerves were only considered as mechanical routes along which cancer cells can spread (28, 29). However, now the consensus is that the peripheral nervous system (PNS) is functionally relevant; it regulates a complex network of cellular and soluble factors related to the formation of the premalignant microenvironment and, consequently, tumor development, growth, and progression (24, 30, 31).

The possibility of tumor regulation by the CNS was carefully examined for the last four decades, and psychological or psychosocial stress has been emerging as one of the main factors associated with cancer initiation, growth, and spreading (32, 33). Numerous clinical and epidemiological studies provided considerable contribution to the understanding of the effect of psychological stress on cancer, with particular emphasis on cancer-associated suppression of immune responses to malignancies and stress-induced alterations in the tumor microenvironment (34–37). For instance, prevalence of depression and mood-related disorders in cancer patients exceeds that detected in the general population, and depression is associated with a poorer prognosis in cancer patients (38, 39). The increased prevalence is not solely explained by the psychosocial stress associated with the diagnosis. Although depressive spectrum disorders are closely linked to stress and induce hypothalamic–pituitary–adrenal axis and sympathetic nervous system activation, downstream suppression of immune cell surveillance, chronic inflammation, and alterations of immune cell generation, function, and homing are likely responsible for accelerated cancer burden (40). Hypothalamic–pituitary–adrenal axis was reported to play a role in tumor initiation, development, and survival by altering apoptosis of lymphocytes, expression of chemotherapy-resistant and survival genes, and controlling immune responses to tumors (41). The stress-induced and depression-induced changes in the sympathetic nervous system were also shown to lead to significant immune and prolonged anxiety-like behavior changes and could be reversed by the blockade of sympathetic signaling prior to stressor exposure (42, 43).

The PNS acts as a crucial part of the cancer microenvironment, and the peripheral nerves have been shown to support cancer progression (44–46). For example, chemical denervation of the pancreas by the Botulinum neurotoxin affected both tumor size and apoptosis of the cancer cells in the orthotopic animal model. Tumors that were coinjected with the neurotoxin were smaller, and the apoptotic rate was increased in the tumor cells (47). In mouse xenograft models, inhibiting the ability of carcinoma cells to generate sympathetic and parasympathetic neurons interferes with cancer progression (48). The authors of this study concluded that nerves, arising from cancer stem cells, are critical for cancer development and progression. In another study, coadministration of neuronal cells with mouse melanoma cells or the presence of neurons in a Matrigel prior to the addition of tumor cells resulted in a significant acceleration of tumor growth (49). This effect was associated with an amplified production of chemokines by tumor-activated neurons and the attraction of MDSC (49).

PNS-derived catecholamines, dopamine, acetylcholine, serotonin, aspartate, glycine, glutamate, γ-aminobutyric acid, bradykinin, gastrin, cholecystokinin, neuropeptide Y, vasoactive intestinal polypeptide, adrenomedullin, neurotensin, calcitonin gene-related peptide, opioid peptides, substance P, and other factors have been reported to bind to specific receptors on malignant cells and have been implicated in cancer progression (30, 50, 51). Neurotransmitters can directly affect carcinogenesis: β-adrenergic signaling can inhibit DNA damage repair, downregulate p53-associated apoptosis, modulate a wide variety of growth and survival pathways, and activate a range of mesenchymal cell types present in tumor stroma (such as fibroblasts, pericytes, mesenchymal stem cells) (52). Adrenergic signals by the autonomic nerves in the cancer microenvironment could “fuel” tumor growth by altering the metabolism and suppressing oxidative phosphorylation in the endothelial cells and activating an angiogenic switch (53). The activation of β-adrenoceptors can promote metastases in animal models of breast, pancreatic, colon, neuroblastoma, ovarian, and prostate cancers (44, 51, 54–56). In humans, increased adrenergic stimulation may drive ovarian tumor growth and metastasis (57) and can impair response to chemotherapy (58). Furthermore, β-blockers can prevent metastasis, suggesting a potentially novel therapeutic approach (59).

Indirectly, neurotransmitters and neuropeptides accelerate cancer progression by suppressing the immune response (52, 60). This is not surprising because the majority of immune cells, such as T cells, dendritic cells, NK cells, macrophages, neutrophils, and MDSC, express cell-surface neurotransmitter and neuropeptide receptors (61, 62). For example, tachykinins (e.g., substance P and neurokinin A) derived from peptidergic fibers can modulate the immune response (63). Many studies indicate that dopamine can influence the growth and progression of tumors by affecting the functions of the immune cells (64). Catecholamines may induce apoptosis of lymphocytes, suppress CD8⁺ T cell and NK cell activities, and promote tumor infiltration by the macrophages, leading to tumor escape from the immune surveillance (54, 65, 66). β-Adrenergic signaling influences the secretion of proinflammatory IL-1, IL-6, and IL-8 cytokines; vascular endothelial growth factor (VEGF); and matrix metalloproteinases, facilitating angiogenesis and tissue invasion (50, 67, 68). Acetylcholine may stimulate the secretion of anti-inflammatory immunosuppressive IL-10 and TGF-β cytokines and inhibit the production of proinflammatory IL-1β, TNF-α, and IL-12 cytokines (61, 69). A recent analysis of prostate cancer specimens demonstrated that the density of sympathetic and parasympathetic nerve fibers in tumor tissue was associated with a poor clinical outcome (44). Furthermore, chemical and surgical sympathectomy or the deletion of β2- and β3-adrenergic receptors prevented an early phase of prostate cancer development in an animal model, whereas a pharmacological blockade or the genetic disruption of type 1 muscarinic receptor inhibited tumor invasion and dissemination (44).

These and other data suggest that a persistent release of neurotransmitters in the tumor milieu from the nerve terminals may promote tumor growth and metastasis by modulating the immune system and altering the cancerous cells directly. At the same time, it is important to realize that acute immunomodulatory effects of neurotransmitters may cause an opposite effect on tumor progression, which depends on a specific model system. For instance, tumor-bearing mice on voluntary wheel-running showed >60% reduction in tumor incidence and growth, which was due to NK cell mobilization by epinephrine, as blockade of β-adrenergic signaling blunted training-dependent tumor inhibition (70). Given that results from numerous observational and preclinical studies indicate that physical activity/exercise paradigms regulate intratumoral vascular maturity and perfusion, hypoxia, and metabolism and may augment the antitumor immune response (71), there is an accepted belief that exercise may not just be “healthy” but may, in fact, be therapeutic (72).

Schwann cells are the most abundant neuroglia of PNSthat closely associate with peripheral axons and consist of myelin-forming and non–myelin-forming cells (73–75). Myelin-forming Schwann cells produce a multilayered myelin sheath that insulates nerves, accelerating the conduction of the nerve impulse. Emerging data highlight the involvement of Schwann cells in the immune regulation, nerve maintenance and repair, neuropathic pain, and the repair of nonneuronal tissue (76–81). Recently, the role of Schwann cells in cancer progression and the formation of the tumor immunoenvironment has also been demonstrated (82–84).

One fascinating aspect of Schwann cells is their ability to dedifferentiate and re-enter the cell cycle in response to nerve injury. When the peripheral nerves are damaged and the injured axons disintegrate [i.e., Wallerian degeneration (85)], Schwann cells respond by adopting a repair phenotype: denervating, dedifferentiating, proliferating, clearing myelin and axonal debris, and promoting the growth of a new axon (78, 86). During nerve repair, Schwann cells also secrete proinflammatory cytokines, chemokines, and growth factors that control Wallerian degeneration and nerve regeneration. At the same time, Schwann cells produce factors that counterbalance proinflammatory cytokines, including IL-10 and erythropoietin (77–79). A similar phenotypic switch by Schwann cells occurs during injury and repair of nonneuronal tissues (87, 88).

Injury-induced reprogramming of Schwann cells to a repair phenotype has recently been demonstrated in the context of cutaneous melanoma (27). Specifically, genetic, phenotypic, and functional similarities between tumor-reactive Schwann cells (repair-like Schwann cells) in the tumor microenvironment and nerve injury–induced repair Schwann cells were revealed. Furthermore, ex vivo and resident skin repair Schwann cells accelerated tumor growth and the formation of metastases in vivo. These results, as well as an almost ubiquitous presence of Schwann cells in the human body, suggest that Schwann cells may be involved in the formation of the tumor microenvironment as well as in the establishment of premalignant niche in various tissues (84). Although the reprogramming of Schwann cells to a pro‐repair progenitor‐like state is a weakly understood pathway, increasing evidence suggests that Schwann cell plasticity and multipotency in tissue repair can contribute to the formation of unique pathogenic features of the tumor microenvironment (89). New findings implicate that both genetic changes and the microenvironment can synergize to increase the plasticity of normally lineage‐restricted Schwann cells. Unsurprisingly, therefore, innervation is emerging as an important constituent of the tumor microenvironment; the regenerative field has clearly demonstrated that innervation is an imperative part of the local environment required to make new tissue and thus would seem likely to contribute to tumor development (89).

It seems likely that solid tumor innervation might recapitulate normal nerve degeneration and regeneration processes that take place during tissue repair. Therefore, Schwann cells are likely to play a significant role in both initiating tumor innervation and then contributing to the microenvironment that can promote and maintain tumor development. In fact, it was reported that although the main function of Schwann cells is to maintain axonal integrity, they can direct malignant cell migration toward nerves, stimulate pancreatic and prostate cancer cell invasion in an integrin-dependent manner, and promote perineural invasion via neural cell adhesion molecule 1 (NCAM1) signaling (90, 91). Schwann cells demonstrate strong affinity toward malignant cells. For example, the presence of Schwann cells in the premalignant lesions of pancreatic and colon cancers implies that their migration may be an early sign and a driver of early stages of tumor development (92). Other reports also found that Schwann cells can accelerate metastases in various cancer models (83, 84). The interaction between Schwann cells and oral squamous cell carcinoma was shown to promote proliferation, migration, and invasion via an increased production of adenosine and IL-6 (93). Importantly, IL-6 is also a pronociceptive mediator, suggesting that Schwann cells are not only important promoters of cancer progression but can also serve as mediators of cancer-associated pain syndrome.

In addition to their direct effect on the malignant cells, Schwann cells are able to modulate the immune cells in the tumor microenvironment. For example, tumor-activated Schwann cells upregulate the expression of various chemokines and attracted MDSC to the tumor environment (84). Furthermore, Schwann cells augment the immunosuppressive potential of the attracted MDSC, measured by the MDSC’s ability to suppress the proliferation of preactivated T lymphocytes. This is the first evidence of Schwann cell modulation of the tumor-immune microenvironment, which affects cancer progression (84). Similarly, in animal models of melanoma, Schwann cells can attract and polarize dendritic cells to a regulatory phenotype (80). These results explain why administering Schwann cells with melanoma cells in mice results in a significantly faster tumor growth compared with tumors composed of the melanoma cells without Schwann cells (27). Specifically, as Schwann cells do not increase the proliferation of melanoma cells in vitro, it is possible that attraction and polarization of myeloid regulatory cells by tumor-activated Schwann cells result in an accelerated tumor growth of this mouse melanoma model. Our new unpublished data further reveal that melanoma-activated Schwann cells can attract and facilitate the exhaustion of T cells (G.V. Shurin, Y. Bunimovich, and M.R. Shurin, manuscript in preparation). Taken together, these results lead to the following conclusions: 1) malignant cells influence Schwann cell functionality by reprogramming them to a repair-like phenotype and 2) tumor-activated Schwann cells are potent promoters of the tumor-supportive microenvironment (Fig. 1). Therefore, the interaction between neuronal and immune components of the tumor microenvironment is an important factor in cancer development and progression.

Over the past two decades, it has become evident that development and progression of cancer are supported by many nonneoplastic cell types that comprise tumor stroma and include immune cells, endothelial cells, cancer-associated fibroblasts, and nerves (94). We are now beginning to discover how the physiological macroenvironment of the body interacts with the local tumor microenvironment and influences tumor growth and metastasis (41, 95). The nervous system can modulate the immune response systemically and locally at the site of the tumor and, therefore, represents an important regulatory component of the tumor micro- and macroenvironment. The tumor-associated neuroimmune network plays a critical role in malignant cell dissemination, being tightly regulated by both immunological and neuronal cues within the tumor microenvironment (Fig. 1). Although many of the neuroimmune interactions described in tumor models recapitulate the dysregulated neuroimmune responses observed in other diseases, such as autoimmune, autoinflammatory, neurodegenerative, and neuroinflammatory disorders, many functional alterations of the peripheral nerves within tumor tissue remain unknown. Furthermore, much remains to be discovered about the cellular and molecular pathways of “Nerve-Driven Immunity” (96) in cancer (i.e., the immune responses dictated by the nervous system in the presence of agonistic and antagonistic factors derived from malignant and stromal cells). Targeting aberrant neuronal functions within the tumor milieu may prove to be an effective therapeutic strategy against cancer.

The authors have no financial conflicts of interest.