How cancer hijacks the body’s homeostasis through the neuroendocrine system

Abstract

During oncogenesis, cancer not only escapes the body’s regulatory mechanisms, but also gains capabilities to affect local and systemic homeostasis. Specifically, tumors produce cytokines, immune mediators, classical neurotransmitters, hypothalamic and pituitary hormones, biogenic amines, melatonin and glucocorticoids, as demonstrated in human and animal models of cancer. The tumor, through the release of these neurohormonal and immune mediators, can control the main neuroendocrine centers such as the hypothalamus, pituitary, adrenals and thyroid to modulate body homeostasis through central regulatory axes. We hypothesize that the tumor-derived catecholamines, serotonin, melatonin, neuropeptides, and other neurotransmitters can affect body and brain functions. The bidirectional communication between local autonomic and sensory nerves and the tumor, with putative effects on the brain, are also envisioned. Overall, we propose that cancers can take control of the central neuroendocrine and immune systems to reset the body homeostasis in a mode favoring its expansion at the expense of the host.

Cancer and its progression in the context of local and global homeostasis

The process of carcinogenesis and tumor progression involves several steps that are affected by genetic, epigenetic, and environmental factors[1-7]. During this process, a malignant tumor escapes local or systemic regulatory mechanisms and becomes resistant to and/or evades immune surveillance, which can result in the death of the patient. The process is multifactorial, and involves the recruitment of surrounding normal cells (fibroblasts, endothelial cells, macrophages, etc.), which interact with the tumor cells to generate a permissive environment for cancer growth and to autoregulate the tumor’s phenotype and behavior[6,8-12]. Moreover, growing tumors exert homeostatic stress through both passive processes (necrosis and effusion of metabolites) or active release of metabolites, lipids, necrotic factors, growth factors, cytokines, and chemokines acting at local and systemic levels, of which cachexia is the extreme form[11,13-15]. Cachexia, defined as extreme weight loss and muscle wasting, affects about 80% of metastatic cancer patients, and is an example of how cancer changes body homeostasis and metabolism through passive or active release of multiple factors by the tumor[13,16,17].

The tumor-induced homeostatic distress described above might be sensed by the brain through tumor innervation as inferred by the role of autonomic and sensory nerves in the tumor development[18-20]. Humoral or metabolic factors released by the tumor or tumor-primed immune cells can act on central coordinating centers to control systemic homeostasis[15,21,22]. The access to the regulatory brain centers including hypothalamus is augmented by the disrupted blood-brain barrier (BBB) in advanced cancers[23]. The role of the hypothalamus, the hypothalamic-pituitary-adrenal (HPA) axis[15] and the autonomic nervous system[15,24] in restoring body homeostasis disrupted by advanced cancer with their concomitant attempts to control the neoplasm was recently proposed. Furthermore, it was proposed that the inability to restore global homeostasis due to aberrant metabolic and cytokine/chemokines profile can lead to disrupted physiological functions, examples of which are blunted glucocorticoid responses and sleep/wake disruption[15].

Therefore, following prior studies from our group, we decided to investigate interactions that can lead to a control of body homeostasis by the cancer through regulation of the central neuroendocrine axes and through production and release of classical neuroendocrine mediators that can benefit cancer itself. The instructive models for such mechanisms of action are provided by skin cancers, especially melanomas, which produce classical regulators or mediators of the HPA axis such as corticotropin releasing hormone (CRH), related urocortins, POMC-peptides including ACTH, MSH and endorphins or glucocorticoids, other pituitary and hypothalamic hormones, enkephalins, biogenic amines including catecholamines, serotonin, N-acetylserotonin (NAS), melatonin and acetylcholine as examples[25-32]. These factors, when released into circulation or acting on nerve endings, could reset body homeostasis to benefit the tumor at the expense of the host.

Below, we will evaluate some of the hypotheses outlined so far by investigating the evolution of the neuroimmunoendocrine tumor-host-tumor interactions from the local to the central level, with a particular emphasis on the mechanisms by which advanced cancers can control the brain and/or central endocrine systems. We also propose extending the current theories of tumor progression by the addition of neuroimmunoendocrine regulatory elements employed by cancer cells during tumor evolution.

Neuroendocrinology of peripheral organs.

Instructive for the presented hypothesis is the production by several peripheral tissues of classical neuroendocrine mediators and local expression of signaling networks mirroring those operating in the brain and central endocrine organs. These include the immune system[33-35], barrier organs such as the skin[36,37], gastrointestinal tract (GI)[38-40] and lungs[41], and others [42]. These organs, depending on the environmental stressor, can communicate with the central neuroendocrine system via activation of the hypothalamus and/or pituitary or through release of biogenic amines or neuropeptides to counteract the stress and rebuild local homeostasis as illustrated in the skin[29,37,43,44], GI[38-40], lungs[45] and immune system[33-35].

Exposure of the skin to different environmental stressors and signals including physical, chemical, and biological and expression of local neuroendocrine networks, represents an excellent testing model for such neuroimmunoendocrine connections. Specifically, all skin cellular components can produce neuroendocrine mediators in a cell type- and context-specific fashion and can communicate with local neural and vascular networks[36,44,46,47]. Because of these properties, the skin has been referred to as “the brain on the outside”[48]. The skin can also be considered as an example of a “diffuse neuroendocrine system”[42] that would also communicate with the immune system[37]. It also has been proposed that the body's main stress response system, the HPA axis, may have developed initially in the integument[49]. These mechanisms will be analyzed next in the context of the neuroendocrinology of human cancer.

Neuroendocrinology of cancer in a nutshell

It has been proposed that the hypothalamus can sense cancers [15], and that biogenic amines can affect tumor behavior[24,50-54]. The neuroendocrine activity of each tumor is determined by its embryonic lineage and the functional evolution of its tumor cells. For example, melanoma develops from neural crest-derived melanocytes[29], whereas squamous and basal cell carcinomas develop from keratinocytes (squamous epithelium) that are of ectodermal origin[36]. It is therefore not surprising perhaps that these tumors can produce classical neuromediators[25,36,46,55-58](Table 1). Similar considerations may apply for the tumors of immune/hematopoietic lineage, which are of mesenchymal origin. The diverse neuroendocrine capabilities of the immune system have been discussed in depth previously, and include communication with the CNS under both physiological and pathological conditions[33,34]. The line of argumentation that malignant tumors can regulate the CNS, endocrine system, and central homeostasis using classical hormones, neurohormones and neurotransmitters is discussed below and main regulatory elements are presented in Figure 1.

Table 1.

Examples of neurohormones, neurotransmitters, hormones and neuroregulatory factors produced by malignant tumors.

Category of neuroendocrine factors	Precursors and final molecules produced by tumor cells	References	Mechanism of action
Hypothalamic hormones	CRH, GHRH TRH	[60-63] and reviewed in [26,44,81] Reviewed in [73,74] [68-71] and reviewed in [44]	Mechanisms include: 1) respective stimulations of pituitary to produce and release of POMC peptides, TSH and GH; 2) direct action on peripheral and endocrine organs and immune system via corresponding membrane bound receptors; 3) actions on the brain after crossing the BBB in circumventricular organs (CVOs) or when the BBB is disrupted in advanced cancer or during cancer therapy
Pituitary hormones	POMC and derived peptides including ACTH, α-MSH, β-MSH, γ-MSH and β-endorphin TSH	[61,75,77-79,82-85,87,139] and reviewed in [25,44,81] [68] and reviewed in [44,56]	Mechanisms include: 1) activation of cortisol/corticosterone production in adrenal cortex or peripheral organs via ACTH-MC2 axis; 2) POMC peptides action on MC and opioid receptors in the periphery and at the central level if crossed BBB; 3) uncontrolled release of ACTH into circulation may disrupt circadian rhythm; 4) TSH action on the thyroid or in the periphery via activation of TSH receptors to produce or activate thyroid hormones
Catecholamines and precursors	L-DOPA, dopamine, epinephrine and norepinephrine	[107,108,111] and reviewed in [44,57,81,110]	Mechanisms include: 1) action in the brain when L-DOPA crosses the BBB, whereas catecholamines will cross the BBB only when it is disrupted in advanced cancers; 2) pleiotropic actions in the periphery through activation of dopaminergic and adrenergic receptors, or supplying catecholamines producing cells/organs with L-DOPA; 3) actions as neurotransmitters in the sympathetic system
Serotonin and derivatives	Serotonin NAS	[52,107,117-119,121,122] and reviewed in [30,50,53,57] [118-120] and reviewed in [30]	Serotonin can act: 1) in the brain after active transport through the BBB; 2) in the periphery and on the immune system through activation of membrane bound serotonin receptors. NAS can cross the BBB or will enter pineal gland to serve as a precursor to melatonin
Melatonin and derivatives	Melatonin and its indolic or kynuric metabolites	[30,118-120] and reviewed in	Melatonin will cross the BBB to affect brain functions and will disrupt circadian rhythm and/or functions of peripheral organs through action on MT1 and MT2 receptors. Melatonin metabolites such as indolic compounds or AFMK and AMK may affect body functions, however, their precise mechanism of action is undefined.
Steroids	Cortisol and corticosterone, mineralocorticoides, sex hormones, precursors to steroids	[92,95-98] and reviewed in [44,91,94,100]	Steroids will inhibit immune system and affect body metabolic status through action on corresponding nuclear receptors and they will affect brain functions if crossed the BBB. Cortisol/corticosterone will inhibit HPA axis. Precursors to steroids including pregnenolone and progesterone would increase velocity of steroids production in peripheral organs or brain
Cytokines affecting HPA axis	Leptin IL1 and IL6 and TNFα	[125,126] [67] and reviewed in [21,97]	Cytokines can activate HPA axis with downstream regulation of homeostasis, and will act on peripheral organs and immune system through corresponding receptors
Cholinergic mediator	Acetylcholine	Reviewed in [47,131-133]	Acetylcholine will act through cholinergic receptors as a nonneuronal mediator in the periphery or as a neurotransmitter in the brain or parasympathetic system
Opioids	PENK derived met- and leu-enkephalins	[31,130] and reviewed in [44,57]	Opioids can regulate the brain, immune system and peripheral tissues through action on opioid receptors
Peptides affecting body metabolism	Leptin NPY	[125,126] Reviewed in [129]	Leptin and NPY can regulate CRH and POMC signaling at the central and peripheral levels in a context dependent fashion, in addition to regulating the immune system and peripheral tissues through corresponding receptors
CRH related peptides	Urocortin 1-3	[140] and reviewed in [26,55,57]	Urocortin 1 can activate the HPA axis through action on CRHR1. Urocortin 1-3 can regulate peripheral organs and immune system through activation of CRH-R1 and/or CRH-R2
Peptide hormones regulating body calcium metabolism	PTH, PTHrP and calcitonin	[70,84,139] and reviewed in [36,58,84,128]	These peptide hormones can have pleiotropic effects through activation of the corresponding membrane bound receptors in target tissues

Fig. 1. How malignant tumors can hijack body homeostasis.

Malignant tumors can putatively affect body homeostasis through the production of neurohormonal modulators entering the circulation, neurotransmitters activating receptors on nerve ending, or through priming circulating immune cells to regulate other organ functions. For example, tumor-derived classical hypothalamic or related neuropeptides such as corticotropin releasing hormone (CRH), urocortins (URC), growth hormone releasing hormone (GHRH) or thyroid releasing hormone (TRH) may control the anterior pituitary gland via production of corresponding POMC-peptides, growth hormone (GH) and thyroid stimulating hormone (TSH). These will then affect body homeostasis directly through corresponding receptors or indirectly through activation of the adrenal glands, liver, or thyroid gland with the production of cortisol, IGF1/2 or thyroid hormones, respectively as proposed previously for various pathological conditions[26,37,73,74]. Tumors can also release biogenic amines, melatonin, acetylcholine or glucocorticoids (see Table 1) that could regulate body functions either directly or indirectly through actions on endocrine organs, brain, the parasympathetic and sympathetic systems, or through nerves reflexes. Additional regulatory elements relevant in this context are IL1, IL6 and TNFα[26,33,90] that would dysregulate the anterior pituitary or hypothalamus, or other brain centers when blood-brain-barrier (BBB) is disrupted. Figure 1 was created with BioRender.

Production of hypothalamic and pituitary factors by tumor cells

Instructive for the line of argumentation below is cancer’s potential to affect the hypothalamus and pituitary (Fig. 1). This is well illustrated in the case of the HPA axis (Box 1).

Box 1. Central and peripheral hypothalamus-pituitary-adrenal (HPA) axis.

The central HPA axis plays an important role in regulating the body’s responses to stress. At the central level, stress induces corticotropin releasing hormone (CRH) production in the hypothalamus, which is released into the portal circulation to activate CRH receptor type 1 (CRHR1) in the anterior pituitary, leading to increased production and release of adrenocorticotropin hormone (ACTH), which in the adrenal gland activates the melanocortin receptor type 2 (MC2) leading to production and release of cortisol. Cortisol will counteract stress, would attenuate the production of CRH, and would inhibit immune activity. Proinflammatory cytokines including IL1, IL6 and TNFα can also activate production of CRH and proopiomelanocortin (POMC) peptides[26,65]. An important role in this process is played by CRHR1, which after activation is coupled to cAMP signal transduction leading to production and release of ACTH.

A similar regulatory system can operate in the periphery, including barrier organs, to coordinate and regulate responses to stress, with some similarities to and differences from the central HPA axis. In fact, it has been proposed that the organizational structure of the HPA axis has evolved from the primordial stress response system that has developed in the integument for the organismal defense against physicochemical and biological environmental stressors[49]. It is composed of proinflammatory cytokines, CRH and related peptides, POMC peptides and corticosteroids with corresponding receptors acting in organized fashions. The peripheral HPA axis can follow the central hierarchy: CRH stimulates CRHR1, which stimulates POMC expression and processing to ACTH, which stimulates production and secretion of cortisol and corticosterone. However, all elements of the peripheral HPA axis are in close cellular contact, as opposite to the central HPA axis, which is separated by space and anatomical organization. Therefore, the local HPA axis can act in diffuse manner with close collaboration with immunederived cytokines. Thus, neuropeptide messengers and their receptors can act in para-, auto- and intracrine fashions, leading to various phenotypic effects[26], having an impact on tumor development and progression. As proposed previously[26,37], there are likely departures from the central organization leading to the truncated organizations: 1) cytokines stimulating CRH/urocortins that would activate CRHR1/2, 2) cytokines stimulating POMC expression with processing to ACTH that would stimulate cortisol and corticosterone production, 3) cytokines stimulating POMC peptides production, 4) CRH stimulating CRHR1 leading to POMC peptides production; 5) cytokines stimulating POMC peptides production, 6) cytokines directly stimulating cortisol and corticosterone production. Self-amplifying production of CRH and POMC peptides through activation or CRH and MC receptors is also envisioned in this local HPA axis. These interactions would be defined by cell lineage and histoarchitecture and nature of the stressor. In the context of cancer, the organizational structure of the HPA axis would be broken, allowing intra-, auto- and paracrine mechanisms to take a center stage. This can disorganize the peripheral HPA axis, allowing the tumor to use its different elements by the tumor to control local and systemic environment. Since ACTH and cortisol levels follow the central circadian rhythm, this can also lead to circadian rhythm deregulation.

Hypothalamic factors:

As it relates to skin cancers, normal and malignant melanocytes and keratinocytes, melanomas and squamous and basal cell carcinomas, produce and secrete upper regulatory elements of the HPA axis such as CRH and urocortin (URC)[25,26,55,59-62], with their expression positively correlated with tumor progression and more malignant behavior[55,59,60]. Furthermore, tumors of other lineages such as ovarian, endometrial, breast, lung and GI cancers can produce CRH[26,63,64](Table 1).

CRH and URC released by the cancer into circulation would enter the anterior pituitary and interact with CRH-R1 leading to the activation of the pituitary-adrenal axis [26]. These molecules can directly activate CRH-R1 and CRH-R2 expressed in almost all organs and tissues with dysregulation of their functions and metabolic effects[26]. Furthermore, cytokines such as IL1, IL6 or TNFα, produced by melanomas and majority of tumors will affect central HPA axis[65], and its peripheral equivalents[26,66]. An example of this phenomenon is the activation of the central HPA axis by IL6 produced by transplantable tumors in mice[67].

Thyroid releasing hormone (TRH), representing a hypothalamic factor which stimulates the hypothalamus-pituitary thyroid axis, is also produced by normal and malignant melanocytes[56,68-70] and some non-melanoma cancer cells[71]. Increased TRH expression in melanomas correlated with tumor progression and pathological activation of TRH-TSH axis stimulated malignancy[69,72]. Growth hormone-releasing hormone (GHRH), another hypothalamic factor, is produced by tumors and can stimulate tumor growth through activation of pituitary GH and consequently the downstream stimulation of insulin growth factor 1 and 2 (IGF1 and 2) in the liver and tumor itself[73,74].

The forementioned “hypothalamic factors” can regulate pituitary secretory functions through interactions with corresponding receptors, in addition to their direct action in the periphery.

Pituitary factors:

In addition to the pituitary and brain, POMC is expressed and processed in several peripheral tissues[25] and tumors of different lineages (Table 1). Skin-derived normal and malignant cells, including melanomas and cancers of epidermal origin express POMC and produce and release the POMC derived ACTH, MSH and β-endorphin peptides constitutively and in regulated fashion[25,29,59-61,75-79]. These peptides have diverse and distant homeostatic, metabolic and behavioral effects through action on melanocortin receptors widely distributed in the body, including MC1 to MC5 and endorphin receptors. These properties are in addition to their ability to stimulate melanogenesis, affect the behavior of cancer cells, and inhibit immune activities. These outcomes are context-dependent, e.g., local and systemic (circulation) releases, activation of nerve endings, and activation of the corresponding receptors on the local, systemic, and central levels. An important homeostatic input is represented by stimulation with tumor derived ACTH of local[80,81] and systemic corticosterone and cortisol production and release, with corresponding phenotypic effects. It should be noted that advanced melanomas predominantly express POMC and there is a positive correlation between human melanoma progression and intratumoral POMC expression/activity[57,59-61,75,77], and circulation levels of α-MSH correlate with advanced stages of melanoma[82]. A similar relationship is found for non-melanoma epidermal cancers[59,75,79]. Interestingly, the expression of neuroactive mediators such as β-endorphin and ACTH in giant basal cell carcinomas might lead to psychological disturbances[31].

It has been long known that POMC is expressed by a variety of squamous cell carcinomas, adenocarcinomas and tumors of non- or of neuroendocrine phenotype[83-87], with the release of POMC-derived peptides into circulation (Table 1). Increased POMC expression correlates with poorer prognosis and tumor progression[87]. Secretion of POMC-derived ACTH from the tumors could lead to Cushing’s Syndrome[88], which is an example of systemic effects in addition to those discussed above. Because immune cells of different lineages can produce POMC-derived peptides and CRH, it seems likely that this capability may be conserved in lymphomas and leukemias [33,34,89,90].

Thus, tumor-derived CRH, urocortin and POMC-derived ACTH, α-MSH, and β-endorphin can affect body functions and homeostasis through activation of other endocrine glands. For example, CRH or URC may activate the pituitary or pancreas, while ACTH will act on the adrenals. The forementioned neuropeptides can also act on corresponding receptors in peripheral tissues, or centrally if passed or transported through BBB. Depending on the amount released to the circulation and site of action, they could affect behavior, energy expenditure and body metabolism, circadian rhythm, stress responses, nociception and addiction, or cardiovascular and immune systems. This is inferred from the recognized regulatory functions of these neuropeptides.

Production of steroids by tumor cells

Adrenals represent the effector arm of the HPA axis. It is recognized that aside from adrenals, peripheral organs produce glucocorticoids in a highly organized fashion as part of local physiological responses to different stimuli[89,91-94]. This capability extends to cancers[95-98](Table 1). It is best illustrated in the human skin, where normal, immortalized and malignant melanocytes and keratinocytes express crucial steroidogenic enzymes and produce and release corticosterone and cortisol[80,95-97,99-101]. In these cells, cortisol can either be generated from cholesterol, with the first step catalyzed by CYP11A1[97], or from cortisone through the action of HSD11B1[96,100,102]. The local production of glucocorticoids can have tumor suppressive or tumor-promoting effects depending on the context and stage of tumor development[81,92,96,98,100]. This pathway has been proposed to serve as a part of local autoregulatory mechanisms. Systemic effects, however, are also likely if sufficient amounts of corticosterone and cortisol are released into the circulation by the tumor[81,94]. Importantly, CYP11A1-driven steroidogenesis in tumor cells indicates that tumor cells may also use other steroids (e.g., estrogens, androgens, neurosteroids) to regulate its local environment after the metabolic transformation from pregnenolone. If released to systemic circulation, these steroids could exert systemic regulatory activities. Lastly, the ability to produce glucocorticoids by lymphocytes and other immune cells starting from cholesterol[89,92] indicates that such production may occur in lymphomas and leukemias. In fact, in a study from our group, we detected the expression of CYP11B1, CYP17A1, CYP21A2 and HSD11B1 in all leukemias tested in the study[101].

Production of catecholamines by tumor cells

It is recognized that cells of different lineages, including immune cells[103,104], keratinocytes[105], and melanocytes[28,106] can produce L-DOPA and catecholamines in a selective manner. Melanoma cells produce catecholamines[107,108], and melanocytes can produce dopamine and norepinephrine[28,106] and express tyrosine hydroxylase (TH)[109]. However, in melanoma, L-DOPA is produced mainly through the action of tyrosinase, serving both as the precursor to dopamine and melanin, and is also considered a neurohormone by itself[110,111]. Tyrosinase-derived L-DOPA after release from melanoma cells can induce generalized melanosis or can be decarboxylated to dopamine and metabolized to norepinephrine and epinephrine at the local and systemic levels, as indicated by studies on TH deficient mice[112].

Catecholamines produced by melanoma and non-melanoma cancers[113-115] can have diverse systemic effects acting on dopaminergic and adrenergic receptors in target organs (Table 1). Such activation will induce metabolic changes leading to increased levels of blood glucose and free fatty acids. L-DOPA, produced by tumors (Table 1), can potentially cross the BBB and after transformation to catecholamines impact the CNS. Catechols and catecholamines released by tumors can disrupt functions of the CNS and the autonomous nervous system, through diverse mechanisms, including wide activation of dopaminergic and adrenergic receptors at different tissue levels. It has been reported that in head and neck cancers increased catecholamines in the serum is associated with significant behavioral changes[116]. While the sympathetic system can induce both protumorigenic and antitumorigenic effects[22,24,51], tumor-derived signals could switch such regulation favoring tumor survival and progression at the expense of the host[81]. Biogenic amines are discussed in Box 2.

Box 2. Biogenic amines and their precursors produced by tumors.

L-DOPA, a precursor to melanin and dopamine and a hormone-like bioregulator on its own[110,111], is a product of L-tyrosine hydroxylation by tyrosine hydroxylase[24] or by tyrosinase[110]. In the latter case, L-DOPA (L-3,4-dihydroxyphenylalanine) is oxidated to L-Dopaquinone, which is transformed to melanin in a process called melanogenesis[29]. It can also be generated from L-tyrosine non-enzymatically through the action of reactive oxygen species (ROS) such as *OH, O₂⁻ or H₂O₂, produced from photolysis of H₂O under an aerobic condition. In the catecholaminergic pathway, L-DOPA is decarboxylated by the aromatic L-amino acid decarboxylase (AAD) to dopamine, an enzyme that is ubiquitously distributed through the body tissues. Dopamine is converted by dopamine-hydroxylase to norepinephrine with following demethylation by phenylethanolamine N-methyltransferase to produce epinephrine. These reactions predominantly occur in the central and peripheral nervous systems including the sympathetic system, in the adrenal medulla, especially for epinephrine, in the APUD cells[24,51] and other nonneuronal cells such melanocytes, keratinocytes and immune cells[46,104,105]. Dopamine acts not only as precursor for norepinephrine but also as a neurotransmitter involved in reward processing, motivation, addiction, and regulation of body movement at the central level. It can also act as pleiotropic regulator in the periphery acting through dopaminergic receptors. Norepinephrine and epinephrine, through activation of α and β-adrenergic receptors, act both as neurotransmitters, regulating brain functions and the sympathetic system, and as stress hormones in the flight-or-fight responses. They also have diverse phenotypic effects on various body organs and tissues. Examples include organs of the cardiovascular, skeletal, respiratory, gastrointestinal, urinary, immune, cutaneous and adipose systems. Epinephrine also regulates conversion of serotonin to melatonin in the pineal gland.

In serotoninergic and melatoninergic pathways, L-tryptophan is hydroxylated either by the ubiquitously expressed tryptophan hydroxylase type 1 (TPH1) or by TPH type 2 (TPH2), expressed in neuronal cells, to produce 5-hydroxytryptophan[28,30,123,124]. 5-Hydroxytryptophan can also be produced non-enzymatically through the action of ROS (see above). It is decarboxylated by AAD to form serotonin, which acts as a neurotransmitter that can regulate mood, feeding and sexual behavior. Through serotoninergic receptors it also regulates or modifies functions of peripheral organs and tissues. It is further acetylated by arylalkylamine-N-acetyltransferase (AANAT) or arylamine N-acetyltransferase (NAT) to N-acetyl-serotonin (NAS) with its final methylation by hydroxyindole-O-methyltransferase (HIOMT) resulting in the formation of melatonin. Melatonin acts as a neurotransmitter in the brain and as a major regulator of circadian rhythm, and also exerts pleiotropic effects in the body through interaction with MT receptors or through non-receptor mediated mechanisms.

Evolutionarily, DOPA, catecholamines, serotonin and melatonin are more than two billion years old molecules produced by diverse complex and simple organisms including plants and bacteria to improve the probability of survival in hostile environments [123,124]. Therefore, it seems reasonable to hypothesize that tumor cells will utilize these messengers to affect local and systemic body homeostasis including circadian rhythms to survive and expand at the cost of the host.

Production of serotonin, N-acetylserotonin and melatonin by tumor cells

The cancers of different lineages not only express tryptophan hydroxylase (TPH) and crucial enzymes for serotonin transformation to melatonin but also produce and metabolize these neurohormones[30,50,107,117-122](Table 1). The serotonin production by tumor cells implicates its auto and paracrine regulation of tumor behavior and surrounding tissues through interactions with the corresponding receptors[50]. There is also a possibility of its systemic effects[50] that could also include modification of behavior and cognition, if serotonin crosses the damaged BBB.

Within tumoral tissue serotonin is acetylated to NAS with further transformation to melatonin (see Box 2). In addition, NAS after entering circulation will be transformed to melatonin at the distant sites expressing hydroxyindole methyl transferase[28,30]. This would lead to increased systemic levels of melatonin, which could disrupt body metabolism and circadian rhythm. The effect on brain functions is also likely, since NAS and melatonin can cross the BBB[123]. In addition, due to the anti-oxidative properties of melatonin[123,124], its endogenous production by tumors could increase the tumor resistance to chemo- or radiotherapy. This could be an unexpected side effect, since beneficial effects of melatonin in oncology are widely recognized.

Production of other neuromodulators by tumor cells

Selected neuromodulators and neurohormones produced by cancer cells are listed in Table 1. One of them that may affect not only tumor growth in autocrine manner[125,126] but also the hypothalamus and pituitary and consequently body homeostasis is leptin[127]. A classic example of this type of deregulation is cachexia[17]. In addition, there are several neuromodulators and neurohormones produced by cancer cells as best illustrated by skin cancers[31,36,44,56-58,70,126,128-130]. These neuromodulators can affect the central neuroendocrine system through activation of corresponding receptors after their delivery through circulation or possibly via activation of ascending nerve endings. The latter is inferred by phenomena of brain sensing inflammation and tissue damage or cutaneous itch[33,37,90]. This can be further substantiated by recently reviewed data showing that the brain can sense ultraviolet radiation through the skin[43].

It is generally accepted that tumor cells, directly or indirectly, regulate immune activity. Therefore, it is highly feasible that immune cells primed by the tumor can affect neuroendocrine organs, not only by the release of cytokines, but also by release of neurotransmitters, neuromodulators and hormones, as discussed by others[33,34,90].

Nonneuronal functions of acetylcholine have been appreciated for about two decades[131,132]. Acetylcholine is produced and released by normal and cancer cells, which also express the types of cholinergic receptors that promote cell growth and affect tumor behavior[32,46,47,131-133]. In addition, it can act as a tumor-derived neurotransmitter activating the parasympathetic system, as discussed elsewhere [32,131].

There are also several other neuromodulators produced by cancer cells including enkephalins, parathyroid hormone (PTH), PTH related peptide (PTHrP), calcitonin, neuropeptide Y (NPY), and indolic and kynuric metabolites of melatonin with potential local and systemic homeostatic activity (Table 1). In addition, neurotrophins, calcitonin related-gene peptide, substance P, pituitary adenylate cyclase-activating polypeptide, neurotensin, bombesin, glutamate can be produced by the tumor cells[19,21,44,57,58,84,128]. Additional factors such as prolactin, oxytocin and dynorphins can also potentially be produced by cancer cells since they are produced by epidermal and adnexal skin cells[37]. These examples are included to indicate that the known ways in which cancers affect the neuroendocrine system can be just a small part of a much broader set of 'disruption' pathways.

Concluding remarks and future perspectives

The status and the function of physiological tissues are shaped by the tissues’ architecture, as well as the nature and the integrity of the dynamic interactions between their cellular and extracellular compartments. Since the characteristics of these interactions can be affected by local neuroendocrine networks, composed of nerve endings, neuroendocrine mediators and expression and activities of the corresponding receptors, disruption of such neuroendocrine communication could be a factor involved in carcinogenesis, starting from its earliest stages.

Cancer cells at the phase of tumor promotion can release neuroendocrine mediators in a stochastic manner to protect the tumor from the host and/or recruit surrounding cells to be involved in a neoplastic process. At the stage of tumor progression and/or metastatic spread, cancer cells may produce neuromediators, defined by the tumor lineage and selection process at the early stages of carcinogenesis. Examples of such cancer mediators are POMC-peptides, CRH, GHRH, other hypothalamic and pituitary hormones, biogenic amines, melatonin, acetylcholine, TRH, leptin, glucocorticoids, pro-inflammatory cytokines activating HP and other neuroendocrine factors (Table 1 and Fig. 1). These factors can regulate the local environment and disrupt global homeostasis in a mode favoring tumor growth, expansion, and metastatic spread and protecting cancer from host responses that include but are not limited to immune responses. These cancer-derived allostatic regulatory mechanisms would complement classical principles of neoplastic transformation and tumor progression[1-7]. Since immune cells can produce several neuromediators, neurotransmitters and hormones[33,34,65,90], a challenging question is how tumors can instruct immune cells[11] to pathologically produce such neuroendocrine factors and serve as cellular messengers entering different body organs to change their neuroendocrine and metabolic activities.

The role of tumor innervation in regulation of tumor behavior and reciprocal signaling to the brain represents a new frontier in oncology and neuroscience[18-20]. While there is evidence that nociceptive ablation impacts tumor progression, it remains to be tested whether afferent fibers relay information from developing cancer to the brain. Since such a regulatory mechanism seems likely, as inferred from similar transmission from inflamed or pathologically damaged tissues, proper investigations are awaiting using animal models of cancer and proper neurophysiological tools including electrophysiology to define the nature of such communication.

The interactions between tumor and the human host is analogous to a “tug-of-war” and can be modeled using “game theoretic” approach[134]. In the advanced stage, it is conceivable that cancer would hijack central neuroendocrine immune regulatory centers and/or networks or brain directly to counteract the body's anti-tumor activity, to disrupt global homeostasis and biological rhythms to promote tumor growth. In this context, the use of antagonists of GHRH in cancer therapy has already been proposed[73,74] raising a quest for similar strategies to be tested for other “hypothalamic” factors. An additional challenge worth noting is how to restore circadian and other biological rhythms in patients with advanced cancers using proper pharmacological and neurophysiological tools.

Considering the variety of neuroendocrine factors and hormones produced by tumors, and their diverse structures and functions, it is challenging to monitor changes in body homeostasis and to define the direct regulatory factors being involved. Network medicine is a new approach that focuses on applying graph network techniques to the characterization of diseased cells, tissues, and organs and their interactions as a whole, from mechanism understandings to treatment [135]. Furthermore, computational systems biology methods have emerged recently to model both intra-cellular and inter-cellular molecular interaction networks[136], represent pathway-level responses to external stimuli[137], and characterize complex multi-scale processes from cells to the whole body[136]. In this framework, first, bulk RNA-sequencing or single-cell RNA-sequencing data will be collected and analyzed within the context of local tissues or a population of cells. Then, new tools that map, predict, and model cell-to-cell communication networks[138] shall be developed to clarify the varying coordinated effects of tissue-specific or cell-type specific gene expressions. Future research to deconvolute multi-omics whole-body scale neuroendocrine signals and tumor sites responses may reveal a refined map of this new mechanism with help from artificial intelligence (AI).

In conclusion, a full understanding of neuroimmunoendocrine mechanisms by which progressing cancer can affect global homeostasis is required for future evidence-based and holistic therapeutic approaches. This exciting perspective should stimulate comprehensive studies on proper animal models of cancer and educated use of AI in multifactorial monitoring of human cancer progression and responses to therapy at different levels.

Outstanding Questions Box.

Multi-omics data and current network biology tools have primarily focused on signaling cascades and pathways within local tissues or a population of cells, with predictive tools still in their infancy. This raises challenging questions as to how big multi-omics data would reflect changes at the scale of whole-body neuroendocrine systems and their target tissue homeostasis. Further, what could be effective strategies for analysis using artificial intelligence (AI)?

Circadian and other biological rhythms are often disrupted or deregulated in patients with advanced cancers, which putatively, may lead to acceleration of the disease. Which pharmacological and neurophysiological tools should be used to restore such rhythms, and in which patients?

The use of antagonists of GHRH in cancer therapy has been proposed. Can a similar strategy be used for other “hypothalamic” factors? In addition, which other pharmacological tools could be used to thwart tumor cells from hijacking the neuroendocrine system?

Immune cells produce several neuromediators, neurotransmitters and hormones, that can be harnessed by tumors for their own advantage. Can tumors instruct immune cells to pathologically produce such neuro-endocrine factors and serve as cellular messengers entering different body organs to change their neuroendocrine and metabolic activities?

In advanced cancers, the BBB can be disrupted by the damaging action of tumor-derived factors or therapy. Can new strategies be developed to restore damaged BBB and/or to prevent the tumor-derived factors from crossing the BBB?

The role of tumor innervation in modulation of tumor behavior represents a new frontier in oncology. Among the challenges for future work is to decipher whether tumor-induced homeostatic distress is sensed by the brain through tumor innervation. And if so, can this process be leveraged for developing new therapeutic strategies?

Highlights.

• Cancer cells can produce and release biogenic amines, melatonin, neurotransmitters, neuropeptides, neurohormones, and glucocorticoids

• Cancer-derived mediators can stimulate main neuroendocrine centers such as the hypothalamus, pituitary, adrenals, and thyroid to regulate body homeostasis

• Cancer-derived neurotransmitters and neurohormones can affect brain functions, disrupt biological rhythms and systemic autoregulatory networks

• We hypothesize that tumors use neuroendocrine and immune systems to set the body in a homeostatic mode favoring the tumor’s expansion