Neural Circuitries between the Brain and Peripheral Solid Tumors

The nervous system, spread extensively throughout the body, maintains dynamic equilibrium across organs through peripheral neural modulation under brain control (1). Increasing evidence underscores that the nervous system plays a crucial role in the initiation and progression of tumors, with the majority of research focusing on the peripheral nervous system (PNS; refs. 2–4). Perineural invasion, a common indication of tumor metastasis, involves the spread of malignant tumor cells along nerve fibers and sheaths (5). Extensive clinical data show that perineural invasion is closely associated with prognosis in head and neck tumors, prostate cancer, pancreatic cancer, biliary tract cancer, and gastrointestinal (GI) tumors (6–9). Mechanistically, nerve cells, together with other nonmalignant cells such as immune cells and fibroblasts, actively regulate the tumor microenvironment (TME; ref. 10). The role of the local tumor neural microenvironment in peripheral solid tumors is being elucidated increasingly. In TME, interactions between tumor cells and nerve cells provide a supportive environment for tumor survival and invasion while promoting nerve infiltration (11). Tumor cells can nourish local peripheral nerves and encourage neurite growth by secreting neuroactive molecules, including neurotrophins and axon guidance molecules (12, 13). Meanwhile, peripheral nerves influence tumor development by secreting neurotransmitters that not only bind directly to corresponding receptors on tumor cells but can also affect immune cells to regulate the tumor immune microenvironment (TIME; refs. 14, 15).

The brain’s role in the development and progression of peripheral solid tumors has only recently been discovered. There also exists bidirectional communication between the brain and tumors. On the one hand, tumors can establish local autonomic and sensory neural networks that transmit signals to the central nervous system (CNS) via existing pathways, affecting brain activity (16, 17, bioRxiv 2023.10.18.562990). On the other hand, the CNS also participates in regulating the TME, primarily through controlling the peripheral nerves that innervate the organ (18, 19). This article reviews the neural circuitries between the brain and various peripheral solid tumors, discussing their roles in tumor occurrence and metastasis, along with the related molecular mechanisms. It also discusses the opportunities and challenges faced in repurposing psychiatric interventions, such as psychopharmacology and psychotherapy, in cancer treatment.

Each organ’s neural supply is unique, with its connections to the brain exhibiting distinct specificity. Therefore, a thorough understanding of the neuroanatomic information, such as neural connections between the CNS and peripheral organs, is crucial for mastering the mechanisms of brain–tumor cross-talk. Focusing on the organs implicated in the current research on peripheral solid tumors, we have, based on previous studies, mapped the neural connections between the CNS and these organs and marked the brain regions potentially involved in the tumor development process (Fig. 1; refs. 20–22).

Brain regions such as the hypothalamus, subventricular zone (SVZ), amygdala, and brainstem are closely linked to tumor occurrence and progression (16, 23–25). The hypothalamus seems to play a crucial role in the cross-talk between the brain and peripheral tumors, as it not only possesses a variety of neural groups but also receives extensive innervation controlled by neural projections from various brain regions (20, 26). The neural circuitries between these brain regions and tumors will be elaborated later in the text. The brain and spinal cord communicate with peripheral organs through the PNS, which includes cranial and spinal nerves. In humans, 12 pairs of cranial nerves (CN I–XII) originate from the brain or brainstem, whereas 31 pairs of spinal nerves arise from five spinal segments: 8 cervical nerves (C1–C8), 12 thoracic nerves (T1–T12), 5 lumbar nerves (L1–L5), 5 sacral nerves (S1–S5), and 1 coccygeal nerve (27, 28). The neuronal cell bodies of these nerves converge in clusters called ganglia, distributed at specific sites within the body (29, 30). In the cross-talk between CNS and solid tumors, communication is primarily achieved through the autonomic and sensory nervous systems within PNS.

The autonomic nervous system (ANS) comprises two divisions: the sympathetic and parasympathetic nervous systems (31). The sympathetic ganglia form two long chains parallel to the spine, known as the sympathetic chains (32). The head and neck receive innervation from the sympathetic nerves of the superior cervical ganglia (33). The sympathetic innervation of the breasts and lungs originates from T1 to T5 and T2 to T7, respectively (34, 35). In the abdomen, nerves from the thoracic and lumbar segments of the sympathetic chain form three ganglia: the celiac ganglion (CG), superior mesenteric ganglion (SMG), and inferior mesenteric ganglion (33). The GI tract is innervated by these three ganglia, whereas the ovaries receive sympathetic nerves from the CG and SMG (21, 35). The prostate is innervated by sympathetic nerves from T11 to L3 (35, 36). The parasympathetic nervous system primarily originates from CNs and three sacral nerves (S1–S3), forming ganglia located near the corresponding organs. The parasympathetic innervation of the head and neck comes from CN VII and CN IX (33, 35). The lungs and GI tract are both innervated by the vagus nerve (CN X; refs. 37, 38). Besides the vagus nerve, the GI tract also receives innervation from the pelvic nerves (38). Both the ovaries and the prostate are innervated by parasympathetic nerves from the pelvic nerves (21, 36).

Sensory ganglia are primarily divided into cranial ganglia and dorsal root ganglia (DRG). The DRG, closely connected to but located outside the spinal cord, correspond to each pair of spinal nerves and transmit sensory information from peripheral organs to the brain (39). Sensory nerves from the trigeminal ganglion (CN V) and DRG C1 to C2 primarily innervate the head and neck (40, 41). The breasts are primarily innervated by sensory nerves from C3 to C4 and T3 to T6 (21, 34). The lungs are mainly served by cranial nodose ganglion (CN X) and the DRG T1 to T6 (42). The GI tract’s sensory innervation comes from the nodose ganglion and DRG spanning T8 to L1 and L4 to S1 (21, 40, 43, 44). The spinal segments controlling sensation in the ovaries and prostate are T10 to L1 and L5 to S2, respectively (21, 45).

Beyond neuroanatomic connections, neuropeptides and other bioactive substances serve as crucial mediators in CNS–peripheral effector cell communication. These substances form the neurochemical foundation for central–peripheral information cross-talk, coordinating to maintain the brain’s internal balance (46). Disruptions in specific brain regions or dysregulation of central neural circuitries, often caused by poor lifestyle choices or other factors, can upset this balance and impact tumor occurrence and metastasis. Notably, tumors themselves can also act as stressors affecting brain activity and modulating the tumor progression. We have summarized the specific brain–tumor circuits involved (Table 1).

Epidemiologic evidence indicates that negative stress, poor emotional health, and disruptions in circadian rhythms—attributes related to lifestyle habits and personal behaviors—may be risk factors for the onset and progression of tumors. Recent studies, integrating interdisciplinary approaches from cancer research and neuroscience, are gradually uncovering the specific circuits and related molecular mechanisms involved.

Stress events encompass both physical and psychologic stressors (such as major life events or factors associated with family, workplace, and social life). Research studies indicate that stress can induce adverse emotional states like anxiety, tension, and depression, altering immune, neurochemical, and endocrine functions, all of which are closely linked to the incidence and progression of tumors (47, 48). Epidemiologic studies have linked stressful life events, such as bankruptcy, job loss, or the death of a spouse or friend, to an increased risk of breast cancer in women (49). Furthermore, studies have found that major depression is associated with poorer survival rates among patients with cancer (50).

In the CNS, it is the corticotropin-releasing hormone (CRH) and catecholaminergic (CA)/sympathetic neurons that are the core components of the stress response. These components regulate peripheral activity through the hypothalamic–pituitary–adrenal (HPA) axis and the sympathetic nervous system (SNS), respectively (51). The hypothalamus is crucial in regulating stress and emotions in the brain, modulating the secretion of adrenocortical hormones (glucocorticoids) through the HPA axis and adrenal medullary hormones (catecholamines) through the sympathetic–adrenal–medullary system (SAS; refs. 52, 53). Exposure to stressors can significantly activate the HPA axis or the SAS (54). The hypothalamic paraventricular nucleus (PVN), known as the brain’s stress integration center, receives direct and indirect projections from key areas involved in stress response, such as the hippocampus, prefrontal cortex, and amygdala (52). Additionally, catecholamines released from the solitary nucleus and the locus coeruleus–noradrenergic system can also activate the PVN-regulated HPA axis (55–57). Animal models have provided compelling evidence of the impact of stress on tumor development and progression, with research showing that exposing rats with chemically induced early hepatocellular carcinoma to restraint stress promotes tumor growth (58). Chronic stress promotes the progression of colorectal cancer in genetically engineered mouse models (59). The tumorigenic effects of these stressors may be related to the prolonged activation of the HPA axis and SAS during chronic stress responses, which impair the immune responses by suppressing both nonspecific and specific immune components. Stress hormones such as cortisol and catecholamines act on glucocorticoid receptors and adrenergic receptors of immune cells, leading to alterations in various components of TIME. Specifically, this results in enhanced immunosuppressive functions of regulatory T cells and myeloid-derived suppressor cells (MDSC), inhibited maturation and antigen-presenting functions of dendritic cells, reduced cytotoxicity of CD8⁺ T cells and NK cells, polarization of macrophages toward an M2 phenotype, altered differentiation of CD4⁺ T cells, and variations in inflammatory cytokines such as IL12, IFNγ, and TNFα (60). Consequently, key effectors in the tumor immune response are weakened, diminishing the body’s immunosurveillance against tumors. Animal experiments have shown that preexposing mice to chronic stress increases circulating tumor cell colonization in the lungs. This process is mediated by β-adrenergic signaling, which elevates the expression of C–C chemokine receptor type 2 in macrophages and monocytes, leading to macrophage recruitment and infiltration into premetastatic lungs, thereby altering the premetastatic niche to favor tumor colonization (61). Another research found that the surgical stress–induced reduction in NK cells is closely related to survival rates of spontaneous postoperative metastasis in mouse models (62). Besides influencing immune responses, stress-induced adrenergic activation significantly contributes to the production of tumor-related bioactive substances, including VEGF. In an ovarian cancer mouse model, adrenergic stimulation through β2-adrenergic receptors activates the cyclic adenosine monophosphate-protein kinase A signaling pathway in tumor cells, enhancing the expression of VEGF, matrix metalloproteinase 2, and matrix metalloproteinase 9, significantly increasing tumor vascularization, leading to a greater tumor burden and more aggressive ovarian cancer cell growth (63). However, complete adrenalectomy, which eliminates the HPA axis or SAS, fails to prevent stress-induced tumor development. This suggests that these neuroendocrine pathways are not the only mechanisms through which the brain influences cancer regulation during stress (64). Studies have found that some effects of stress may have an epigenetic basis. Sister chromatid exchange, considered one of the preclinical markers of cancer risk, reflects cytogenetic damage and genomic instability (65). Although different stressors induce varying levels of sister chromatid exchange in rats, the induction is universal, providing evidence that stress may affect other tumor-related biological processes (66). Additionally, stress-induced alterations in DNA damage, somatic mutations, and DNA repair mechanisms may contribute to the development and progression of certain cancers (67).

It is crucial to acknowledge that not all stress is detrimental. Some research indicates that chronic, mild activation of the HPA axis by positive stress can exhibit antitumor properties (68, 69). The concept of allostasis describes how organisms adaptively respond to external challenges (26). The immune system suppression mentioned previously is due to allostatic overload, in which the axis is damaged by negative stress, leading to inadequate responses to external stressors, resulting in maladaptation (70). In contrast, exposure to mild, nonaversive challenges can form a more adaptive hypothalamic axis, which can buffer responses to subsequent external stressors (71, 72). Additionally, besides positive stress, positive emotions also have antitumor effects. The brain’s reward system, essential for regulating emotions, also plays a role in triggering antitumor immune responses. Central to this neural network are the dopaminergic neurons located in the ventral tegmental area (VTA), which influence positive emotions, expectations, and motivations (73–75). These neurons project to the limbic system, directly linking the perception of rewards and motivational behavior (76, 77). New research has found that chemical activation of the reward system weakens the immunosuppressive properties of MDSCs. This reduction is evidenced by the heightened expression of granzyme B in CD8⁺ T cells in the TIME, which inhibits tumor growth (18). However, this effect is unlikely to be direct, as dopamine does not cross the blood–brain barrier, raising questions about the transmission of signals from the brain to the tumor. Subsequent research revealed that VTA activation leads to decreased norepinephrine levels in the bone marrow (BM), without affecting levels in the spleen or tumor sites. Additionally, MDSCs developing in the BM are regulated by noradrenergic signals (18). Changes in the environment of MDSCs developing in the BM affect their subsequent function in other locations. Although MDSCs in the BM are not the sole cell group influenced by the VTA, changes in these cells are sufficient to impact the TIME. Considering the reward system’s pivotal role in regulating positive emotions, these findings indicate that the patient’s psychologic state may influence antitumor immunity and potentially affect cancer progression.

The suprachiasmatic nucleus (SCN), located in the anterior hypothalamus just above the optic chiasm, is a vital neural structure (78). It synchronizes external light and dark changes by processing light signals received from the retina. The SCN is a major component of the body’s biological clock, responsible for maintaining and regulating central and peripheral circadian rhythms. This includes regulating body temperature, ANS functions, endocrine system activity, and sleep–wake cycles (79). Light signals received by the retina are transmitted to the SCN primarily through excitatory glutamatergic axon fibers of the retinohypothalamic tract. This activates rhythmic cyclic adenosine monophosphate signaling, essential for maintaining the transcriptional cycles of core clock genes in the SCN (80).

Recent research has identified circadian rhythm disruption as an independent cancer risk factor and classified it as a carcinogen. It plays a crucial role in various cancer processes including cell proliferation, cell death, DNA repair, and metabolic changes (81). Epidemiologic studies consistently demonstrate that shift work significantly increases the risk of cancers such as breast, prostate, colorectal, and lung (82–85). The underlying mechanisms are gradually being revealed, focusing mainly on circadian rhythm genes. For example, high expression of circadian rhythm genes such as circadian locomotor output cycles kaput, period circadian regulator 1 (PER1), PER2, PER3, cryptochrome 2, neuronal PAS domain protein 2, and RAR-related orphan receptor C is closely associated with a longer metastasis-free survival in patients with breast cancer (86). Similarly, genetic engineering experiments in mice have shown that lung cancer is closely related to core circadian rhythm genes, with the absence of PER2 and brain and muscle ARNT-like 1, leading to increased expression of the oncogene c-Myc (85). Additionally, prolonged night-time light exposure may inhibit the pineal gland’s nocturnal secretion of melatonin, potentially affecting estrogen levels and increasing breast cancer risk (87). Notably, the activity of the SCN is not isolated; it can also modulate hypothalamic activity via neural and humoral signals, affecting the HPA axis (88). These findings indicate that disruptions in the circadian expression of immune or metabolic genes, triggered by night-time light exposure, may be linked to tumorigenesis.

Orexinergic neurons are primarily located in the hypothalamic nuclei, specifically the arcuate nucleus (ARC) and the lateral hypothalamic area. They are integral to a complex network that includes neural connections and neurochemical signaling with the PVN, the ventromedial hypothalamus (VMH), and the dorsomedial hypothalamus (DMH). These networks collectively regulate appetite, hunger, satiety, and energy balance (89, 90). Evidence indicates that activation of orexigenic neurons modulates the HPA axis’s response to stress, significantly influenced by leptin (91). Recent reports link elevated serum leptin levels to various cancers, such as pancreatic, breast, and endometrial cancers (92–94). Studies have demonstrated that leptin release inhibits the hypothalamic release of orexin (95). For pancreatic cancer cells, hypocretin-1/orexin A and almorexant can induce apoptosis and inhibit growth through orexin receptor type 1 activation (96). Orexin and almorexant may be potent candidate drugs for treating pancreatic cancer. Furthermore, in mouse models, mice housed in enriched environments (EE) showed reduced tumor growth and enhanced mitigation effects. Further molecular analysis showed that EE selectively upregulates the expression of brain-derived neurotrophic factor (BDNF) in the hypothalamus ARC and VMH/DMH nuclei, which modulates the HAS axis and SNS to decrease leptin production in adipocytes and increase adiponectin levels. As leptin is a mitogenic factor and adiponectin is an antimitogenic factor, the expression of BDNF in the ARC and VMH/DMH nuclei can lead to decreased proliferative capacity and increased apoptosis in tumor cells (26). The gene or environmental activation of the BDNF/leptin axis between the brain and periphery may have therapeutic implications for cancer.

The brain communicates bidirectionally with other organs of the body through specific communication systems involving the interaction of multiple systems, which are part of the complex biological feedback and regulatory mechanisms within the human body. Known pathways such as pathways like the brain–gut and brain–lung axes act as bidirectional conduits between the CNS and peripheral tissues, integrating neuroanatomic, immunologic, endocrine, and microbial interactions (46). These axes significantly influence the development and progression of tumors. Similarly, recent research has uncovered a specific tumor-promoting circuit between the brain and the breast, further illustrating the intricate connections facilitating tumor development.

Introduced in the early 20th century, the brain–gut axis encompasses the CNS, the ANS, and the GI microbiome (97). The brain–gut axis represents brain-to-gut and gut-to-brain bidirectional relationship: the brain regulates intestinal functions via the HPA axis and the ANS, whereas the gut affects CNS functions through microbial metabolites, neuroactive substances, and gut hormones that reach the brain through the enteric nervous system, the vagus nerve, the circulatory system, or the immune system (98). Stress is closely linked to the brain–gut axis, with significant implications for inflammation and cancer. Stress responses can disrupt the gut microbiome and alter the TIME, whereas inflammatory responses in the gut can significantly affect brain homeostasis, impair hippocampal neurogenesis, and GI microbiome even potentially inhibit the development of brain cancer (99–101).

Recent advances in animal models and neural tracing technologies have begun to unveil the specific mechanisms by which certain brain nuclei within the brain–gut axis influence tumor development. Studies in colorectal cancer–bearing mice reveal that CA neurons in the ventrolateral medulla (VLM) become activated and their activity levels significantly influence tumor growth, underscoring the critical role of the brainstem in regulating tumor progression (19). VLM CA neurons can influence the microbiome via the brain–gut axis, potentially modulating the immune system to regulate tumor growth (19, 102). Additionally, hypothalamic PVN has neurons that produce oxytocin, a hormone crucial in regulating anxiety, stress, and depression (103). In a colorectal cancer mouse model, knockout of oxytocin-producing neurons (Oxt neuron) in the brain increased anxiety behaviors, enhanced tumor cell proliferation, and reduced apoptosis. Activating these neurons not only increased serum oxytocin levels but also reversed the mice’s anxiety behaviors and tumor progression. Moreover, serum ACTH and corticosterone levels decreased, and an increase in CD8⁺ T cells within the TME was observed, suggesting that hypothalamic Oxt neurons inhibit colorectal cancer growth potentially by altering the TIME and possibly through regulation via the HPA axis. These effects depend on the role of Oxt neurons in the CNS and are regulated through the Oxt^PVN neuron → CG–SMG neuron pathway (23).

The lungs and brain form an integrated physiologic whole, where damage to one can affect the other and vice versa. These effects are mediated through a complex network of signals, including neural, inflammatory, immune, and neuroendocrine pathways (104). A 2022 study in Nature disclosed the brain–lung axis, suggesting that imbalances in the pulmonary microbiome significantly affect the CNS’s autoimmune responses (105). The brain–lung axis features a bidirectional network where neural anatomic pathways are innervated by three types of peripheral nerves, mainly under the control of thoracic spinal sympathetic and parasympathetic nerves, with the vagus nerve playing a crucial role (37). When external stimuli activate sensory neurons on the lung surface, vagal afferent nerves transmit this information to the CNS. The brain integrates these data and sends feedback to peripheral tissues through vagal efferent nerves, enabling cross-talk between the brain and lungs. For example, the medullary nuclei can receive glutamatergic neurotransmitter information from pulmonary vagal afferent nerves, and the CNS influences acetylcholine release via the vagal pathway, interacting with the α7 nicotinic acetylcholine receptor on lung immune cells, thus modulating the TIME (46, 106, 107). Recent research using an in situ lung cancer mouse model and neural analysis techniques has provided a comprehensive view of the central neural networks that directly or indirectly innervate the lungs. Neural tracing technologies have mapped neural nuclei in the medulla (NTS, IRt, LRt, SP5, RVL, ROb, RPa, and GiV), pons (LC, PAG, Subc, Su5, and EW), midbrain (VTA), hypothalamus (PVN), amygdala (Ce), and cerebral cortex (M1 and M2) directly connected to the lungs. Subsequent tests on neuronal activity in medullary and pontine nuclei demonstrated increased activity in the GiV, ROb, and RVL regions, alongside decreased activity in the RPa regions of lung cancer mice. These findings provide fundamental insights into the interactions between brain and lung tumors, where the ventral medulla’s RVL, ROb, RPa, and GiV regions play a crucial role in regulating lung tumors associated with the CNS. The ventral medulla may represent a critical area in the brain–lung axis for tumor regulation (25).

The amygdala is a critical region in the brain for emotional processing (108). The central medial amygdala (CeM), a key output nucleus of the amygdalar complex, integrates cortical and intra-amygdalar inputs, controlling anxiety via projections to brainstem nuclei like the lateral paragigantocellular nucleus (LPGi; ref. 109). A newly identified brain–tumor neural circuit closely related to the CeM has been discovered in a breast cancer mouse model. Ablating CeM^CRH neurons and inhibiting the CeM^CRH → LPGi pathway both led to decreased norepinephrine levels in tumors, significantly reducing anxiety behaviors in mice and concurrently decreasing tumor size. Additionally, the research has shown that this effect is linked to the influence of HPA axis and SNS on TIME by mediating norepinephrine levels in tumor tissue. Ablation of CeM^CRH neurons markedly increased the presence of CD45⁺ T cells, including CD4⁺ and CD8⁺ subpopulations, within tumors. This change was accompanied by an increase in M1 macrophages and a reduction in regulatory T cells (16). These research findings provide theoretical support for previous epidemiologic observations that patients with breast cancer are closely associated with anxiety (110, 111). Additionally, the discovery provides compelling evidence for the role of brain–tumor neural circuits in regulating tumor growth and its immune microenvironment, offering new insights for the treatment of breast cancer. Most importantly, due to the presence of specific connections between various organs and the brain, the discovery of this new circuit in breast cancer implies the potential existence of unexplored specialized brain–tumor circuits in other types of tumors.

The development of tumors bears many similarities to embryonic development processes, including the requirement of neurogenesis. Similarly, the initiation of primary tumor formation and metastasis also necessitates the development of neural networks (112). In adult mammals, there are two neurogenic zones within the CNS: the SVZ of the lateral ventricle and the dentate gyrus of the hippocampus (113). Doublecortin (DCX) is a classic marker for neural precursor cells located in these neurogenic regions of the CNS (114, 115). It has been suggested that the increased neural density in tumors is not due to the generation of new neurons (as the number of neurons remains constant) but rather through axonal growth promoted by neurotrophic factors (116). However, recent research using a prostate cancer mouse model has discovered that DCX⁺ neural progenitor cells in the SVZ can migrate to the periphery, enhancing tumor initiation and metastasis. Specifically, these DCX⁺ progenitor cells breach the blood–brain barrier to enter the circulatory system, selectively migrate to, and colonize primary and metastatic sites, where they further differentiate into adrenergic neurons. This supports an adapt circumstance for the early stages of tumor development and metastasis (24). Similar experimental findings were observed in a breast cancer mouse model, suggesting that this may be a more general characteristic of cancer development (24). These findings challenge traditional thought and reveal a unique interplay between the CNS and tumors, pointing to new neural targets for cancer treatment.

Patients with cancer frequently experience a range of behavioral changes such as depression, sleep disturbances, and cognitive impairments. These behavioral comorbidities are prominent throughout the cancer diagnosis and treatment process and may persist during the survival period even after complete remission (117). Increasingly, studies suggest that these changes are closely related to the transmission of signals from tumor-infiltrated nerves to the brain, although other factors such as damage from chemotherapy cannot be ruled out (Table 1).

Patients with cancer exhibit a higher prevalence of emotional dysregulations, including anxiety, depression, and suicidal tendencies, compared with the general population (118). The tumor itself can also serve as a chronic stressor, triggering various tense emotions from physical, economic, and interpersonal aspects, thereby increasing the risk of deteriorating mental health (119). Patients with cancer frequently experience psychologic changes, which may be partly due to the tumor influencing the brain via neural circuitries. PET scans have shown reduced metabolism in the limbic systems of patients with cancer, supporting the notion that the psychologic deficits in patients with cancer are associated with abnormal brain metabolism in limbic areas (120). Furthermore, studies have found that long-term cancer survivors have a higher incidence of depression compared with healthy individuals. These survivors, though showing no signs of disease recurrence after cancer treatment, require long-term mental health care (121). It may be due to the initial nonspecific symptoms not being promptly addressed. This delay can allow the tumor–brain circuit to cause irreversible changes at the transcriptional and functional levels in brain neurons, effects that may persist even after the tumor’s diagnosis and treatment (bioRxiv 2023.10.18.562990). Importantly, these psychologic disorders not only degrade patients’ quality of life but also may hasten tumor progression (122). Research using mouse models has found that mice with in situ breast cancer exhibit severe tumor-associated anxiety compared with normal mice. In these models, levels of norepinephrine were correlated with the tumor size and weight, indicating a positive correlation between cancer-induced anxiety, sympathetic nervous activity, and the progression of breast cancer (16). Therefore, in clinical treatment, addressing only the physical aspects is insufficient; the psychologic health of the patient is also crucial to the efficacy of cancer treatment.

Although evidence has clearly established that patients with cancer receiving chemotherapy may experience chemotherapy-related cognitive impairment because of long-term neurotoxic side effects, clinical data indicate that patients with untreated breast cancer and those with colorectal cancer also exhibit cognitive decline (123, 124). Possible reasons include the following: First, tumors may attract DCX⁺ neural progenitor cells from the SVZ to migrate out of the brain to support their own development, thereby depleting neural precursors in the brain, which is considered one of the causes of cognitive impairment (24, 125). Second, the hippocampus, prefrontal cortex, and amygdala, regions that play roles in the regulation of cognitive functions, may have their activity disrupted by the stress induced by tumors (126).

Pain is the most common chief complaint leading to emergency department visits among patients with cancer (127). Tumor infiltration of sensory nerves alters the transmission of sensory signals to the brain, exhibiting different patterns at early and late stages of tumor development; early stages are marked by suppressed sensory transmission, whereas later stages see enhanced transmission. In the initial stages of tumor development, due to dysregulation of sensory transmission between the tumor and the CNS, patients often do not experience significant symptoms, such as pain, leading to delayed diagnosis (120, 128). In contrast, in the later stages of cancer, pain becomes a typical clinical manifestation in patients, not only due to more severe tumor nerve infiltration but also due to the amplification of sensory signals. Hematoxylin and eosin staining and IHC reveal that high-grade serous ovarian carcinoma predominantly has sensory nerves, unlike the sympathetic nerves found in normal fallopian tubes and ovaries, suggesting that sensory nerve acquisition is a consequence of the disease (17). Further studies indicate that in cancers such as breast, prostate, pancreatic, lung, liver, and colorectal, nerve infiltrates within tumors are composed of individual fibers, not nerve bundles, suggesting that these fibers are recruited to the TME rather than being preexisting. Neurotracing has revealed that signals from these nerve fibers are transmitted to the spinal DRG, playing a role in the transmission of tumor pain signals (17). Additionally, cross-excitation between sensory neurons and sympathetic–sensory coupling in DRGs can cause abnormal sensory signal input, such as signal amplification (129–131). Research using a mouse model indicates that TRPV1-expressing nociceptor neurons infiltrating head and neck squamous cell carcinomas converge on the ipsilateral V3 branch of the trigeminal ganglion and brain areas such as the spinal nucleus of the trigeminal, the parabrachial nucleus, and the central amygdala. These areas are integral to a preexisting pain regulatory circuit. In genetically modified mice lacking nociceptive sensory neurons, decreased neuronal activity in the parabrachial nucleus and spinal nucleus of the trigeminal is observed, accompanied by decreased anhedonia, reduced tumor growth, and improved survival rates (bioRxiv 2023.10.18.562990). It suggests that the connection of sensory nerves within the tumor to preexisting circuits, including brain regions, not only transmits sensory signals but also impacts tumor onset and survival prognosis.

Additionally, cancer-related fatigue (CRF) is a prevalent, persistent, and challenging symptom experienced by patients with cancer and survivors (132). Fatigue may be heightened before treatment begins and can increase further during treatment (133). The biological and genetic mechanisms that trigger CRF include cytokine dysregulation, HPA axis dysfunction, 5-hydroxytryptamine neurotransmitter imbalance, disruptions in circadian rhythms, changes in ATP and muscle metabolism, and activation of vagal afferents (134). The tumor itself, acting as a stressor, contributes to these mechanisms. Although there is no gold standard for managing CRF, interventions are possible. According to the current level of evidence, exercise seems to be the most effective way to prevent or alleviate CRF both before and after treatment (135).

The brain–tumor neural circuitries exert bidirectional regulation between the CNS and tumors. Although the specific pathways and mechanisms have not been fully elucidated, existing evidence suggests that therapies targeting brain–tumor cross-talk could offer clinical benefits and create opportunities for new clinical cancer treatment paradigms that integrate biological behavior perspectives and combined psychiatric interventions.

Psychiatric interventions not only guarantee the improvement of the overall treatment experience and quality of life for patients but may also provide new avenues for cancer treatment. Randomized clinical trials have shown that nonpharmacologic treatments such as psychotherapy and mind–body therapies can alleviate depressive symptoms in patients with cancer (136). Individualized psychotherapy can significantly improve spiritual well-being and quality of life of patients with cancer in the short term (137). Alprazolam and progressive muscle relaxation both can reduce cancer-related anxiety and depression (138). Proper management of stress and negative emotions in patients with cancer may beneficially affect the neuroendocrine regulation of tumor occurrence, growth, metastasis, and the cancer immunoediting process (139). Psychiatric interventions that alter negative emotions seem to regulate HPA axis hormone activity (140). Stress management interventions that mitigate chronic stress–related physiologic changes may aid the “recovery” of the immune system, thus enhancing immune surveillance during active cancer treatment (141). Recent speculations and related mechanisms have also been confirmed in animal models. For example, triptolide, a pentacyclic triterpenoid compound extracted from the roots of thunder god vine, although not a psychotropic drug, can regulate leptin’s effect on the hypothalamus (142). In a colorectal cancer mouse model, intraventricular administration of triptolide stimulated the activity of Oxt^PVN neurons, increased oxytocin secretion, acted on the brain’s oxytocin receptors, and inhibited CG–SMG sympathetic neurons, thereby inhibiting tumor cell proliferation, promoting apoptosis, and suppressing tumor progression (23). Alprazolam, one of the most widely used benzodiazepines for treating generalized anxiety disorder and panic disorder, inhibited the activity of CeM^CRH and LPGi neurons and reduced cancer-induced anxiety and sympathetic activity in a mouse model. This significantly boosts antitumor immunity, thereby slowing the progression of in situ and spontaneous breast cancer in tumor-bearing mice (16, 143). Additionally, psychotropic drugs may exert antitumor effects through other mechanisms. For instance, the tricyclic antidepressant imipramine blue may interact with the proto-oncogene FoxM1 and inhibit the activity of FoxM1 and related signaling pathways, thereby inhibiting breast cancer cell growth and metastasis by impairing the homologous recombination repair of DNA strand breaks (144).

It is important to note that although evidence suggests that these psychiatric interventions may have potential antitumor effects, they should be considered complementary to, not substitutes for, standard therapies (145). The additional antitumor effects of therapies targeting brain–tumor circuitries have been further confirmed in animal models; for example, ablating VLM CA neurons boosts the therapeutic efficacy of paclitaxel (19). Therefore, the best practice would be to integrate psychotherapy and psychotropic drugs with conventional therapies to maximize treatment efficacy.

Additionally, the choice of analgesic regimen should be made cautiously. Most patients with cancer suffer from tumor-induced pain, severely affecting their quality of life, and inevitably require pain medication (146). Dexmedetomidine provides sedation, analgesia, and opioid-sparing effects, making it suitable for both short-term and long-term sedation in intensive care (147). Unfortunately, a breast cancer mouse model found that dexmedetomidine might encourage tumor development and metastasis (148). Furthermore, in a head and neck cancer mouse model, although the pain relievers carprofen and buprenorphine alleviated issues related to pain, they did not improve tumor growth. Notably, buprenorphine increased tumor growth compared with carprofen (bioRxiv 2023.10.18.562990). These findings indicate that the choice of analgesics in cancer-related medication plans may potentially adversely affect tumor growth. The proper selection of analgesic regimens is also crucial for improving the quality of life and prognosis for patients.

In this review, we have highlighted the pivotal role and specific mechanisms of the neural circuitries between the brain and peripheral solid tumors in tumor initiation and progression. The bidirectional communication between different peripheral solid tumors and the brain demonstrates clear specificity. Brain–tumor cross-talk can occur based on preexisting neural structures or through new local autonomic and sensory nerve networks formed by tumor induction. Tumors themselves can act as stimulants affecting brain activity, which in turn can shape the local TME favorable for tumor growth through the PNS, forming a positive feedback loop that promotes tumor development. Targeted treatment of brain–tumor cross-talk could potentially disrupt this cycle.

Although the specific pathways and mechanisms are not yet fully elucidated, incorporating clinical treatments that combine biological behavioral perspectives and psychiatric interventions as supplementary approaches in cancer treatment is becoming feasible. However, despite research findings on the potential antitumor effects of psychotropic drugs, evidence also suggests that some psychotropic drugs may promote tumor growth. For instance, the antidepressant desipramine has been demonstrated to enhance the growth of 4T1 tumors in a breast cancer mouse model (148). Therefore, more experiments are needed to classify psychotropic drugs based on their effects on tumors to better repurpose those with antitumor properties for clinical adjunctive use. It is worth noting that prevention is always better than cure. Lifestyle habits and behavioral patterns are closely linked to brain activity and tumors. Thus, adopting healthy lifestyles will help maintain brain health and reduce the risk of certain types of cancer, including a balanced diet, regular exercise, good sleep quality, and effective stress management.

Given the critical role of brain–tumor interactions in malignant tumor research, we anticipate more extensive and in-depth studies in the future. A deeper understanding of these mechanisms will lay a solid theoretical foundation for further clinical research and may open new avenues for the reutilization of psychiatric interventions in tumor treatment.

No disclosures were reported.

This study was funded by grants from the National Natural Science Foundation of China (grant numbers 82073324 and 82372927). The authors thank editors for editing grammar, spelling, and other errors.