The connection between innervation and metabolic rearrangements in pancreatic cancer through serine

Abstract

Pancreatic cancer is a kind of aggressive tumor famous for its lethality and intractability, and pancreatic ductal adenocarcinoma is the most common type. Patients with pancreatic cancer often suffer a rapid loss of weight and abdominal neuropathic pain in their early stages and then go through cachexia in the advanced stage. These features of patients are considered to be related to metabolic reprogramming of pancreatic cancer and abundant nerve innervation responsible for the pain. With increasing literature certifying the relationship between nerves and pancreatic ductal adenocarcinoma (PDAC), more evidence point out that innervation’s role is not limited to neuropathic pain but explore its anti/pro-tumor functions in PDAC, especially the neural–metabolic crosstalks. This review aims to unite pancreatic cancer’s innervation and metabolic rearrangements with terminated published articles. Hopefully, this article could explore the pathogenesis of PDAC and further promote promising detecting or therapeutic measurements for PDAC according to the lavish innervation in PDAC.

Background

Cancer is the most urgent public health problem worldwide, and it was estimated that 1,806,590 people will be diagnosed as cancer patients and 606,502 people will die of this deadly disease in America in 2020. Among these numbers, 57,600 cancer cases and 47,050 cancer deaths were attributed to pancreatic cancer in the United States (1). Moreover, pancreatic cancer has the lowest 5-year overall survival (OS) rate of about 9% to 10% of any other solid tumors. The percentage of incidence and mortality of pancreatic cancer is highest among people aged 65–74 for both sexes (2). The (age-adjusted) incidence and mortality rate of pancreatic cancer was 13.1 per 100,000 persons based on the database from 2013 to 2017 and 11.0 per 100,000 persons during 2014–2018. Statistically, men have a higher incidence than women (14.9 and 11.6 annually per 100,000 persons, respectively) as were the mortality rates (12.7 and 9.6 annually per 100,000 persons, respectively). Depending on the statistical analysis, the age-adjusted rates for new pancreatic cancer cases have been stable from 2008 to 2017; however, age-adjusted death rates have been increasing on average by 0.3% per year from 2009 to 2018 (2, 3). It is predicted that pancreatic cancer will be the second leading cause of cancer-related deaths by 2030 (4). Lately, it has exceeded breast cancer as the third leading cause of cancer death in all ages for both genders and the fourth leading cause of cancer death for males and females, respectively (1, 5).

Pancreatic cancer, with the malignant neoplasm, has many histologic types, beginning with neoplasms originating from ductal and non-ductal cells. The former includes pancreatic ductal adenocarcinoma (PDAC), accounting for approximately 90% of total types and representing typical pancreatic cancer (6). Other PDAC-related carcinomas in neoplasm of ductal origin include adenosquamous carcinoma, osteoclastic giant cell carcinoma, colloid carcinoma, and medullary carcinoma. The last kind of original ductal neoplasms arises in tumoral intraepithelial neoplasms consisting of intraductal papillary mucinous neoplasm (IPMN), intraductal tubulopapillary neoplasm, and mucinous cystic neoplasm with associated invasive carcinoma. Intriguingly, a comment points out that 18% IPMNs and PDAC are found to be co-occurring in the same pancreas, which supports the concept of “field effect” (7), but neither genetically related nor representing a similar oncogenetic development (8). Acinar cell carcinoma, pancreatoblastoma, pancreatic neuroendocrine neoplasm, and solid pseudopapillary neoplasm are the other four kinds belonging to neoplasms of non-ductal origin (9, 10). Known as the most aggressive cancer, PDAC always shows enormous difficulties in diagnosis and therapy. The pancreas is located within the upper abdomen, so PDAC is extremely hard to be detected in its early stage by image measurements and resected by surgery because of the surrounding encased vessels. PDAC shows an aggressive biological nature of invasion and early metastasis. With rapid proliferation and growth limiting nutrition and energy intake, patients often suffer cachexia and pancreatic dysfunction for exocrine and endocrine (11). Moreover, PDAC has a desmoplastic stromal and flexible elasticity, and plasticity exhibits apparent resistance to chemotherapy and radiotherapy and has little target molecule (12, 13). Cigarette smoking, severe obesity (body mass index known as >35) (14), type 2 diabetes mellitus, alcohol consumption, and pancreatitis increase the risk of PDAC, and cigarette smoking is the most common one (15, 16). According to the environmental risk factors listed, genetic alterations play a more significant role in the appearance of PDAC. Among the PDAC patients with a pathogenic germline mutation, BRCA2 (associated with breast and ovarian cancer) and ATM (related to ataxia–telangiectasia) are the top two genes (17, 18). Besides PLAB2 (correlated to breast cancer), BRCA1 (connected to ovarian and breast cancer), and p16/CDKN2A (associated with risk of melanoma and familial atypical multiple mole melanoma), DNA mismatch repair genes like hMSH2, hMLH1, hPMS1, hPMS2, and hMSH6/GTBP (Lynch syndrome), PRSS1 (causing rare inherited hereditary pancreatitis), and LKB1/STK11 (explaining about 80% of Peutz–Jeghers cases) also increased the estimated risk of PDAC (17, 19, 20). Notably, familial PDAC shows a remarkably rising risk of PDAC compared with the general population (21). Based on these altered genes and mutation-related syndromes, there is an excellent opportunity to quantify the risk of PDAC, which is essential for the clinical application of early detection and timely treatment of PDAC and screening for the risk of other malignant diseases ( Figure 1 ).

Figure 1.

Internal and external risks of pancreatic ductal adenocarcinoma (PDAC). Notably, familial PDAC shows a remarkably rising risk of PDAC compared with the general population. Based on these altered genes and mutation-related syndromes, there is an excellent opportunity to quantify the risk of PDAC, which is essential for the clinical application of early detection and timely treatment of PDAC and screening for the risk of other malignant diseases. ①, DNA mismatch repair genes include hMSH2, hMLH1, hPMS1, hPMS2, and hMSH6/GTBP.

PDAC presents few clinical symptoms before it evolves into an advanced stage; if any, those symptoms often lack specificity for one to make a diagnosis. Generally, these non-specific clinical symptoms include abdominal pain, jaundice, abnormal liver functions, diabetes, dyspepsia, nausea or vomiting, back pain, and weight loss (22). The most frequent feature of PDAC is abdominal pain related to neuropathic pain mainly caused by perineural invasion (PNI), defined as cancer cells diffusing at least 33% of the surrounding nerves, including the epineural, perineural, and endoneurial space of the nerve sheath (23). Invasion of PDAC cells impairs the neural sheath and reprograms neural density, causing neuropathic and inflammatory pain, whereby PNI suggests a metastatic route for PDAC metastasis and neuropathic pain. Interestingly, a study points out that PNI has the highest prevalence in PDAC within variant gastrointestinal malignancies and shows an impactful association with poor PDAC patient outcomes (24). The genetic analysis of PDAC also confirms frequent regulators of axon guidance (25), which is consistent with the correlation between neuronal supports and axonogenesis of PDAC.

PDAC has a desmoplastic and abundantly innervated tumor microenvironment (TME). However, the dense stromal leads to a nutrient-poor environment, rendering PDAC a plastic metabolic remodeling for oxygen and metabolites (26). The metabolic rearrangements of PDAC allow the tumor to survive in the specific environment and, more importantly, increase and overgrow in human bodies. PDAC cells would choose non-classic metabolic pathways (aerobic glycolysis, hexosamine biosynthetic pathways, de novo lipogenesis, etc.) for their energy and nutrient production under some circumstances (12). Accordingly, other cellular components such as pancreatic satellite cells (“reverse Warburg effect”—pancreatic satellite cells promoting metabolites for PDAC cells to produce energy) and neurons (known to release amino acid-derived neurotransmitters like serine via peripheral axons) also provide the metabolic needs for PDAC cells under nutrient-depleted conditions (27, 28).

Nerve innervation in PDAC

Organs like the pancreas are innervated by the autonomic nervous system consisting of two parts: the sympathetic nervous system (SNS) and the parasympathetic nervous system (PSNS) (29). Sensory nerves, compose of dorsal root ganglia (DRG), are responsible for receiving all nerve impulses from the pancreas. Sympathetic, sensory, and parasympathetic nerves disperse the whole pancreas and play individual roles (30, 31). Classically, the plexus leaving from the coeliac plexus and then entering the pancreas head innervates the head of the pancreas, and the plexus derived from the splenic plexus innervates the body and the tail of the pancreas (32). It has been proved that not only PDAC cells but also PanIN cells and PDAC patients demonstrate highly increased nerve penetration and neurotropism by KPC in in vivo murine and in vitro co-culture models (33–35). The abundant nerve innervation of the pancreas is the foundation of PNI, besides the precondition for interactions between nerves and PDAC cells ( Figure 2 ).

Figure 2.

The innervation of the pancreas is complex, including the autonomic nervous system and sensory nerves, which is significant for tumorigenesis and the development of pancreatic ductal adenocarcinoma (PDAC). The sympathetic nervous and sensory systems promote the development of PDAC through neurotransmitters. Notably, neuropathic pain induced by PDAC activates sensory nerves in nerve impulses and then triggers transient receptor potential vanilloid 1 (TRPV1) to secrete substance P and calcitonin gene-related peptide. However, there is an extensive debate on the complicated role of parasympathetic nervous system in PDAC, and the entire signaling network remains to be discovered.

Sympathetic nerves

Sympathetic innervation in the pancreas comes from the sympathetic preganglionic neurons from the thoracic 5–10 segments of the spinal cord. They leave as sympathetic ganglia, become splanchnic nerves, and end in celiac ganglia (36). Sympathetic preganglionic neurons release acetylcholine (Ach) as functional neurotransmitters comprising cholinergic neurons. However, sympathetic postganglionic neurons releasing norepinephrine (NE) linking to adrenergic receptors on its downstream tissues as their functional neurotransmitters suggest that sympathetic postganglionic neurons should be classified as noradrenergic neurons (31, 37). It is widely accepted that catecholamines produced by sympathetic nerves as stress molecules could stimulate the proliferation of PDAC. An orthotopic mouse model, non-invasively tracked by in vivo optical imaging, is applied to evaluate the growth and proliferation of PDAC. This model greatly reappears the interlaced association of PDAC cells and microenvironment, and it indicates that pharmacological β-adrenergic is ample to reverse the activated β-adrenergic signaling caused by chronic stress responsible for the development of PDAC. After treating PDAC cells with isoproterenol, invasion-related genes like matrix metalloprotease 2 (MMP2) and MMP9 were assayed. These accelerating expressions collectively suggested that β-adrenergic signaling impacts the growth, progression, and invasion of PDAC in vivo (38). Another in vitro mouse model implanted a DRG nearby a colony of PDAC cells (including MIA PaCa-2 and BxPC-3 cell lines) in Matrigel, whose growth factors are depleted, and then applied with different concentrations of NE for 7 days (39). The model strongly shows that NE is critical in the PNI development of PDAC. Mechanistically, NE activates the STAT3 pathway through β-adrenergic receptors (ADRB) and its downstream PKA. After blocking the effect of STAT3 by phosphorylation inhibitor AG490, the expressions of STAT3 were attenuated, accompanied by the invasion-related protein levels of MMP9 and MMP2 and the secretion of neural growth factors (NGF) induced by NE. The in vitro experiments demonstrate that STAT3 blockade could inhibit NE-induced PDAC growth, migration, and PNI (39). It is well studied that sympathetic innervation impacts tumorigenesis and the growth of PDAC (mainly associated with NE produced by sympathetic postganglionic neurons) and breast cancers (associated with the expression of immune checkpoint molecules) (38–41). Additionally, a recent work aims to determine whether sympathetic neural signaling would affect the progression of PanIN (41). They utilized the genetically modified LSL-Kras+/LSL-G12D;Pdx1- [(KC) mice to simulate murine PanIN. After being treated by daily repeated immobilization, chronic restraint stress accelerated the systemic epinephrine levels and promoted PDAC carcinogenesis in the KC murine models. The common mediator of chronic stress is reported as β2-adrenergic receptor (ADRB2), which was also increasing in the pancreas of KC mice. Then, the LSL-Kras+/LSL-G12D;LSL-Trp53+/R172H;Pdx1-Cre (KPC) model, which could develop an advanced stage of PanIN, is established; these results claimed that ADRB2 blockade tremendously prolonged OS in KPC mice and that ADRB2 signaling is crucial for nerve growth in PDAC carcinogenesis. Mechanistically, adrenergic signaling related to catecholamine, specifically isoproterenol, induces PDAC development through acinar to ductal metaplasia (ADM), which is also described as the first step of PDAC carcinogenesis (42, 43), and promotes the secretion of neurotrophins (NT) through ADRB2 and the PKA/ERK pathways. Moreover, after assessing the expression of several NT, including NGF, brain-derived neurotrophic factor (BDNF), neurotrophic factor 3 (NTF3), NTF4, glial cell line-derived neurotrophic factor (GDNF), and Netrin-1 (encoded by NTN1), NGF demonstrated the highest expression in PDAC cell lines, KC mice, and KPC mice. Significantly, the blockade of the NGF/Trk pathway could inhibit proliferation, PNI, and invasion of PDAC, prolonging the OS for KPC mice (41, 44, 45). That hints us in utilizing β-blocker to suppress adrenergic signaling for highly innervated PDAC patients. Clinically speaking, treating postoperative PDAC patients of stages II and III with non-selective β-blocker reveals attenuated BDNF levels, decreased nerve density, and extended OS (41). In summary, sympathetic neuron-expressed NE promotes carcinogenesis and the development of PDAC while facilitating the secretion of NT, which conversely stimulates innervation in PDAC, followed by increased sympathetic related NE and PDAC growth. There is a feed-forward loop among sympathetic neurons, ADRB2 pathway, PDAC cells, and NT, and this loop could render PDAC sufficient neuroplasticity for innervation, proliferation, and growth. In that case, targeting this specific pathway gives us a novel strategy to treat the neoneurogenesis of PDAC. Coincidentally, this paper testified the possibility of targeting ADRB2 signaling to influence innervation in human PDAC to further improve the clinical outcomes of PDAC patients (41). Sahni and colleagues collected a cohort of PDAC patients who underwent surgical resection for their stage II and III tumors, and they were treated with non-selective β-blockers (NSBB), β-1 selective blocker (SB1B), and no β blockers (NBB). Then, the patients were retrospectively analyzed to compare their OS. Interestingly, patients with NSBB have no significant difference from those with SB1B in OS. However, they had a nearly double OS compared with patients with NBB. The immunohistochemistry analysis also confirms that there were fewer innervation (presented by neuronal marker S-100) (46) and decreased BDNF expressions in the NSBB group than in the NBB and SB1B groups. Notably, stages II and III PDAC patients who undergo surgical resection would benefit from NSBB and improve their clinical outcomes. Nevertheless, PDAC patients can barely benefit from β-blocker (NB) usage before diagnosis for their survival advantage. However, they acquire a clear survival advantage through constant β-blocker usage before and after diagnosis (n = 2,564) (47). Furthermore, another latest clinical research claims that NB usage is not correlated to accumulated PDAC risk but reduces the risk in long-term usage, especially NSBB usage (n = 4,113) (48). To date, these publications confirm the therapeutic effect of NSBB for the blockade of ADRB signaling associated with sympathetic relevant NE in PDAC and guarantee reliability and security in PDAC patients.

Although sympathetic preganglionic neurons are concluded into cholinergic neurons, sympathetic nerves have typically been considered to exert NE/ADRB signaling in PDAC. There is little literature to investigate sympathetic Ach in PDAC and other cancers to date. Ach, as a chemical messenger secreted from the ganglion synapse, directly binds to nicotinic acetylcholine receptors (nAchRs) at postganglionic neurons to induce the release of NE from postganglionic neurons. We believe that the cholinergic parts of sympathetic nerves have a little direct impact on the oncogenesis and neoneurogenesis of PDAC. Hypothetically, we deem that the secretion of sympathetic preganglionic Ach is straightly binding to nAchRs at postganglionic neurons, and superfluous Ach is sabotaged and devitalized by cholinesterase which barely leaks into the microenvironment nor has direct connections with cancer cells in PDAC or other cancers.

Sensory nerves

Afferent sensory (afferent) nerves transmit sensory information from the pancreas to the central nervous system (CNS) and company with splanchnic nerves and vagi, whose neural bodies lie in DRG, the spinal afferents and nodose ganglia (NG), vagal afferents. Sensory nerves consist of unmyelinated fibers and have sensory and secretory functions (49, 50). It was observed that there is augmented axon density in PanIN lesions in vivo. Moreover, a microfluidic device was utilized to co-culture DRG neurons with PDAC cells and separate the axons (33). This co-cultured system allows axonal interactions and simulates intrapancreatic sensory innervation in human PDAC. Undoubtedly, accelerated sensory axons were recruited through PDAC cells, and an increased number of axons expressed synapsin proteins, suggesting functional axons with enhanced neurotransmitter release and transport. Intriguingly, the genetic and functional aberration of axons has a strong association with the outcome of PDAC patients and cancer progression in murine PDAC models (51). It was found that semaphorin exhibits an augmented expression, which is correlated to the poor OS of PDAC patients on univariate analysis. The same change occurred in early pancreatic tumorigenesis modeling by in vivo pancreatic injury and in vitro ADM, in contrast to normal pancreas (25). Taken together, utilizing antagonists against aberration of axon guidance genes and expression would significantly inhibit the initiation and progress of PDAC. DRG and NG contain sensory neurons sensitive to capsaicin, and capsaicin activates TRPV1 on the unmyelinated sensory fibers (52). Activated TRPV1 has a calcium preference, mainly leading to the release of substance P (SP) and calcitonin gene-related peptide (CGRP) (53). SP belongs to tachykinins encoded by TAC-1, also known as pre-pro-tachykinin-A (Ppt-a) (54). Neurokinin receptor 1 (NK-1R), encoded by TACR1, is one of the families of G-protein coupled receptors (GPCR) combined with SP, which then promotes several signaling pathways for the progress of cancer (55)—for instance, SP, along with its high-affinity target NK-1R, was augmented and activated HER2 in breast cancer cell lines (HER2+) and primary cell culture from breast cancer patients (HER2+). SP acted as a pro-inflammatory factor to modulate the expression of EGFR and HER2 for tumor malignancy. This transactivation is significantly reduced by the inhibition of NK-1R RNA expression, a chemical inhibitor of NK-1R, and inhibition of GPCR-induced signaling (56). In conclusion, activated sensory nerves in the pancreas release SP, and the SP–NK-1R axis could stimulate many pathways like JAK/STAT, MAPK/ERK, and HER2/EGFR (mentioned before) signaling for oncogenesis and evolution of PDAC (56–59). While CGRP (encoded by the calcitonin gene) is a relatively novel neuropeptide first isolated in 1982, human-type CGRP is isolated from medullary thyroid carcinoma (60, 61). The release of CGRP by sensory neurons regulated by NGF via a cAMP/ras manner and activation of mitogen-activated protein kinase (MAPK) binds it to its target receptors CGRP1 and CGRP2, facilitating the concentration of cAMP followed by the activation of protein kinase A (PKA), which then unfolds K+ channels leading to the relaxation of smooth muscles in the vascular endothelium (62–64). Therefore, blood flow in PDAC is accelerated by the generation of CGRP, which is important for transporting trophic substances and modulation of metabolic plasticity. It has been published that CGRP has chemotherapy resistance in acute myeloid leukemia (AML) cell lines via the CGRP/CALCRL axis and promotes prostate tumor growth in murine models, possibly via ERKs/STAT3 signaling (65, 66). By the end of the article, there are few experimental articles about the specific impacts of CGRP and its receptors in PDAC. Hence, we promote a possible mechanism of sensory released CGRP based on the existing theories: secreted CGRP binds to CGRP1/CGRP2, activates cAMP/PKA signaling, K+ channel, and expansionary blood flow potentially accompanied with another bypass, which renders PDAC abundant nutrition, metabolic materials, and neural/metabolic plasticity. Hopefully, this article could provoke more related experiments to unveil the secretory characteristics of CGRP in PDAC. Taken together, neuropathic pain caused by PDAC induces sensory nerve impulses and then evokes TRPV1 to release neuropeptides, including CGRP and SP, which directly or indirectly activates individual signaling pathways for the tumorigenesis and evolution of PDAC. Recent research claims that the neuropathic pain of PDAC results from the abnormally expressed sonic hedgehog (sHH) signaling and is mediated by SP and CGRP in DRGs in an NGF-dependent manner (67). Furthermore, PanIN organoid colonies are directly promoted by sensory neurons via SP/NK-1R signaling and phosphorylated STAT3 (33, 68). Similarly, the NK-1R+ neuroendocrine cells in PanIN shows trophic effects to potentiate global organoid growth and provide developmental signals for NK-1R‐ cells to accelerate the growth rate of organoid colonies. Inversely, sensory denervation in KPCPdx1 mice greatly reduced the growth and progression of PanIN. Meanwhile, the subpopulation of NK-1R+ cells possibly correlated to impaired STAT3 phosphorylation and neuroendocrine cell maintenance (33). Interestingly, capsaicin purified from the pepper plant has a dose-dependent effect on its target TRPV1; the appropriate dose of capsaicin binds to TRPV1 and activates sensory nerves with the release of SP and CGRP, whereas an accumulating dose of capsaicin could achieve capsaicin desensitization which reversibly or permanently silences whole neurons for the treatment proposed (69, 70). According to these mechanisms, the application of TRPV1 antagonist (AG1529 has a partially inhibitory effect) has been approved to prevent the transition from acute pancreatitis to chronic pancreatitis (CP), along with the development of CP and inflammation-related pain in a murine model (71, 72). Similar treatments targeting the sensory innervation signaling in PDAC should be established only after they have been well tested to guarantee safety and validity.

Parasympathetic nerves

Parasympathetic innervation in the pancreas originates from the dorsal motor nucleus of the vagus nerve as parasympathetic preganglionic fibers. It substitutes neurons with parasympathetic postganglionic neurons in intrapancreatic ganglia, which innervate the pancreas (36, 73–75). The parasympathetic nervous system is sorted into cholinergic nerves because its preganglionic and postganglionic neurons release Ach as neurotransmitters. When the vagal nerve is activated, the parasympathetic preganglionic nerve secrets Ach to activate parasympathetic postganglionic nerves via nAchRs about the ligand-gated ion channel family, and then postganglionic neurons also release Ach which binds to muscarinic acetylcholine receptors (mAchRs) at target tissues (76). There are five subtypes of mAchRs on target tissue cells, differing from M1 to M5 (encoded by CHRM1 to CHRM5), and all five belong to the GPCR family (77, 78). It is illustrated that Ach, along with its acetylcholine receptor ligands, overexpresses several human cancers like colon, gastric, lung, and PDAC (44, 79–82). Cholinergic signaling induces an immune-suppressive microenvironment for the neoplastic part in PDAC to support tumor growth. Mechanistically, the Ach/nAchRs pathway impairs CD8+ T cell accumulation, attenuates IFN-γ generation in CD8+ T cells, and switches Th1/Th2 balance to Th2 guiding phenotype through both in vivo orthotopic PDAC model and in vitro high-performance liquid chromatography (83). The proximate discovery about cholinergic signaling reveals an aspect of the secret why PDAC is considered a “cold tumor” and lacks reactions to immunotherapy. In other words, diffuse cholinergic signaling renders PDAC a privilege to escape the immune system and indirectly mediate the microenvironment of PDAC to promote its growth. In addition, nicotine, the agonist of nAchRs, simultaneously suppresses GABA signaling and accumulates ARDB signaling to activate the SHH pathway collectively and then directly induces the self-renewal of pancreatic cancer stem cells, a mass of which initializes oncogenesis and the progression of PDAC (84). Moreover, clinical statistics display that overexpressed M3 is associated with pejorative development and unfavorable prognosis of PDAC patients (n = 58), in whom the cells are usually located at invasive tumor budding cells, metastatic lymph nodes, and parasympathetic nerve fibers (79). This article suggests that M3 should be a novel marker to predict the prognosis of PDAC. However, it seldom excavates the molecular mechanisms of the Ach/M3 axis in PDAC, and a small specimen leads to its limitation for clinical usage. Intriguingly, increasing publications also show that PSNS has recently decelerated cancer evolvement (40, 82, 85). Experiments in breast cancers reveal the innervation of SNS and PSNS, illustrating opposite effects in the tumor microenvironment. The denervation of sympathetic nerves and the neurostimulation of parasympathetic nerves in breast tumor decelerate the expression of immune checkpoint molecules probably mediated by M1 cholinergic mechanisms including PD-1, PD-L1, and FOXP3, which significantly suppress the anti-cancer immune response and become a potential treatment for breast cancer (40, 86, 87). However, whether parasympathetic nervous neurostimulation (M1-associated) also decumulates the immune checkpoint molecules in PDAC (like breast cancer) is not discovered up to the present, and the relational theoretical blank remains as both mystery and opportunity for researchers. As mentioned before, PSNS is also indicated to impair immune response by regulating immune cells to induce an immune-suppressive microenvironment for PDAC growth (83). Meanwhile, Renz and colleagues investigated the parasympathetic system’s specific effects and the possible mechanisms in PDAC. They performed subdiaphragmatic vagotomy on genetically engineered KC mice to disclose parasympathetic denervation in PDAC, which appeared to accelerate PDAC tumorigenesis and augment the expression of M1 under KRAS mutation. Along with vagotomy, utilizing the antagonist (pilocarpine) and the agonist (scopolamine) of M1 receptors has a similar and opposite result for cell viability and expression of M1. Mechanistically, cholinergic signaling suppresses PDAC oncogenesis and development in primary PDAC and metastatic lesions through the activation of M1-related signaling, then reduces the CD44+ CD133+-marked CSC subpopulation, and suppresses the downstream EGFR/MAPK and PI3K/AKT pathway, which indirectly and directly inhibits PDAC development, respectively (85). Overall, this article reveals that PSNS, the cholinergic nerves, suppress the initial PDAC. Still it did not clarify the process of cholinergic signaling activation and what form of M1-related signaling stimulates it downstream for restraining the development of PDAC. Given the existing evidence of neurotransmitters mediating the tumorous functions of automatic innervation in PDAC, we consider that activated cholinergic signaling secretes Ach through parasympathetic postganglionic neurons binding to M1 receptors on the membrane of target tissue cells, followed by the inhibition of EGFR/MAPK and AKT/PI3K signaling and the reduction of CSC subpopulation stimulated by Ach/M1 receptor axis. Furthermore, treatment with pilocarpine reveals a lower M1 expression, but scopolamine shows a higher M1 expression consistent with vagotomy, which suggests that the expression of M1 muscarinic receptors has a potential feedback loop with M1-related signaling activation.

To date, there is an extensive debate on the complicated role of the parasympathetic nervous system during the tumorigenesis of PDAC. Abundant innervations in the pancreas and complex cholinergic neurotransmitters and receptors result in the externally inconsistent conclusion of PSNS in PDAC. Here we propose a possible unified explanation for the two distinct opposite mechanisms of cholinergic signaling: tumorous plasticity of PDAC. PDAC, the most aggressive cancer, has incomparable adaption and high plasticity for its aggressive proliferation, growth, invasion, and metastasis. Take metabolic plasticity, for instance, PDAC could transfer its respiration manner from oxidative phosphorylation (OXPHOS)-dependent manner into aerobic glycolysis, leading to a style under a hypoxic microenvironment (emerging of the Warburg effect) (88); glycolysis and OXPHOS can also co-exist in the same lesions (89). Moreover, glycolytic PDAC cells even transform pyruvate in the mitochondria for OXPHOS when glycolysis is inhibited to promote energy and metabolites for proliferation and invasion (90). Taken together, we deem that the only famous principle about PDAC is that such aggressive cancer cells would take whatever it takes to invade and proliferate. Besides metabolic plasticity, angiogenesis, neurogenesis, immune suppression, systematic cell death (like apoptosis, autophagy, pyroptosis, and ferroptosis), chemo/radiotherapy resistance, etc., generally confirm the proposed principle. The flexible plasticity weaves a complex and gigantic signaling network for tumor initialization and development. It is pretty arbitrary to consider that there is only one independent signaling pathway like PSNS for PDAC progression, which has a dichotomous “all or none” impact without any other possibility. We forecast that PSNS might have a plastic impact on PDAC, while the entire signaling network remains obscure and needs to be thoroughly excavated in PDAC. Moreover, targeting or suppressing associated signaling should be affirmatively verified and experimented with for clinical application in PDAC patients because the deformability and plasticity might switch the PDAC cells into more invasive types.

Neuronal–metabolic interactions in PDAC

The latest research suggests the interaction between neuronal innervation and metabolic remodeling in PDAC (28, 91). Notably, metabolic plasticity renders PDAC cells to adapt to nutrient-deprived microenvironment via increasing scavenging pathways like autophagy, apoptosis, and reverse Warburg effects associated with pancreatic stellate cells (PSCs) or cancer-associated fibroblasts (CAFs) (92–94). Highly innervated PDAC has a strong interaction between PDAC cells and innervation from the microenvironment, and neurons also release amino acids (AA) as neurotransmitters, like D-serine, for long-term potentiation at the synapses or neurodevelopment like synapse maturation and axonal stabilization (95, 96). Given that serine is a non-essential AA and the little evidence about the metabolic association with tumor growth, the authors investigated the possibility of whether neurons have metabolic support to PDAC cells under a nutrient-deprived condition. In summary, this paper innovatively proves that peripheral axons secrete serine to metabolically sustain PDAC growth under a nutrient-deprived microenvironment (28).

Metabolic reprogramming

Aggressive cancer cells experience extensive metabolic rewiring during initial development. Dysregulated metabolism of glucose, lipid, and amino acid is confirmed in both primary and metastatic PDAC; those products and intermediates from those processes are essential for energy generation and material syntheses like cellular membranes, nucleotide, extracellular matrix (ECM), and cell cytoskeletons, which are fundamental for proliferation, invasion, and metastasis (12, 97–100). In PDAC, cancer cells uptake augmented glucose for metabolism and energy generation, which is utilized in positron emission tomography-computed tomography (101, 102). Glucose transfers into the cytoplasm through glucose transporter 1 (GLUT1) in the cell membrane, also highly expressed in PDAC associated with worse prognosis, and hexokinase catalyzes glucose into glucose-6-phosphate (G-6-P) where it divides into a bypass pentose phosphate pathway (PPP) contributing to the synthesis of DNA or RNA (103, 104). Another side path destination of G-6-P is synthesizing glycogen. During the course, glycogen synthase kinase 3 (GSK-3), an important kinase for glycogen synthesis, is reported to induce chemo-resistance for PDAC cells by sustaining the TopBP1/ATR/Chk1 DNA damage response pathway (105). Then, G-6-P is converted into fructose-6-phosphate (F-6-P) through the catalysis of phosphohexose isomerase, under the name of autocrine motility factor, phosphoglucose isomerase, and glucose-6-phosphate isomerase, which leads to another route, the hexosamine biosynthetic pathway (HBP). The end product of HBP is uridine diphosphate N-acetyl glucosamine (UDP-GlcNAc), generated by glucose, glutamine, and glucosamine and utilized in O-GlcNAcylation which is a post-translational modification, mainly glycosylation of serine and threonine (106). Moreover, HBP promotes the progression and survival pathways in PDAC and induces hyaluronan synthesis in the ECM. Furthermore, a study points out that they target glutamine-utilizing enzyme glutamine-fructose aminotransferase 1 via a small molecular glutamine analog, 6-diazo-5-oxo-I-norleucine, demonstrating reduced metastasis and self-renewal potential, along with decreased hyaluronan and collagen in the ECM, and sensitivity to immune checkpoint therapy (anti-PD1 therapy), thus resulting in better prognosis (107). HBP is significantly interlinked together with other metabolic pathways, including glucose, lipid, protein, and nucleotide, which exceedingly support the survival and progression of PDAC. Next, F-6-P is turned into 3-phosphoglycerate (3-PG) through the catalysis of a series of glycolytic enzymes, including phosphofructokinase (PFK), triose phosphate isomerase (TPI), etc. 3-PG directs to a shunt pathway, the serine biosynthesis pathway (SBP), employing 3-PG as a substrate for serine production. Such suggests that glycolytic intermediates could link to SBP from 3-PG and vice versa; serine can be catabolized into pyruvate. To start, 3-PG is transformed into 3-phosphohydroxypyruvate by phosphoglycerate dehydrogenase (PHGDH), which is the first enzyme in SBP, followed by phosphoserine aminotransferase 1 (PSAT1), thus inducing the transition to phosphoserine (p-serine). Subsequently, phosphoserine phosphatase (PSPH) mediates the generation of serine from p-serine, which is the last step of SBP. Serine, the second most abundant amino acid in human proteins, participates in many biosynthetic pathways, such as lipid synthesis, protein synthesis, and the one-carbon cycle, which supports nucleotide synthesis, methylation reactions, and antioxidant defense (108, 109). Interestingly, PHGDH, PSAT1, and PSPH overexpression is found in PDAC and correlated to shorter OS (110–112). When serine is limiting, augmented SBP, especially PHGDH, provides a growth advantage for tumor proliferation via sufficient serine in breast cancer and melanoma compared with normal environmental serine levels (113). Besides the metabolic aspect, SBP enzymes promote PDAC development by regulating molecular signaling. Specifically, PHGDH interlinks with translation initiation factors eIF4A1 and eIF4E while assembling translation initiation complex eIF4F to directly regulate protein synthesis, thus further promoting the progression and development of PDAC (110). In addition, overexpressed enzymes like PHGDH and PSAT1 promote the MYC/miR-494/enhancer of zeste homolog 2 (EZH2) feed-forward loop and are at least partially driven by MYC/ATF4 signaling in Burkitt lymphoma cells to sustain metabolic reprogramming. Notably, one carbon unit (including folate and methionine) produced from serine induces the production of S-adenosylmethionine, which facilitates the reaction of histone H3 trimethylation (H3K27me3), subsequently repressing miR-494 expression and thus decreasing the expression of MYC. MYC indirectly induces the methylation of H3K27me3 through regulating EZH2; the regulators combine into a feed-forward loop to sustain metabolic rearrangements (114). Clinically, the expression of PHGDH has been linked with lymph node metastasis in PDAC patients (115). Similarly, the selective loss of PSAT1 also abrogates migration, invasion, and metastasis in triple-negative breast cancer, suggesting that PSAT1 induces the migratory potential for metastasis without synthesizing serine (116). PSPH, the last enzyme in TSBP, also shows serine-expectant function via suppressing 2-hydroxyglutarate to liberate DNA 5-hydroxymethylcytosine and nuclear receptor NR4A1 expression to facilitate pro-oncogenic gene expression for melanoma growth and metastasis (117). Serine catabolism in the mitochondria is indicated by serine hydroxymethyltransferase 2 (SHMT2) to generate glycine and formate. Inversely, SHMT1 can also promote the cytosol transition from glycine to serine (118). Chemical pyrrolo[3,2-d]pyrimidine compounds targeting SHMT2 inhibitors have been investigated for serine catabolism. Small molecular compounds AGF291, AGF320, and AGF347 were established for antitumor efficiency toward lung, colon, and PDAC in vitro, respectively. AGF327 also demonstrates a potent effect in vivo models exerting a therapeutic potential for SHMT2 suppression for PDAC (119). Afterward, glycine is used to synthesize glutathione (GSH); high levels of GSH have an anti-apoptosis function for chemotherapy resistance in PDAC, and GSH/GSSG ratios regulate reactive oxygen species damage (120). Following on the heels of 3-PG, pyruvate is produced from a three-step catalytic reaction based on 3-PG, which is the initial alanine synthesis. The reversible reaction of converting pyruvate to alanine catalyzes alanine aminotransferase 1, also known as glutamate pyruvic transaminase, in the cytoplasm and vice versa. Finally, pyruvate enters the mitochondria for the tricarboxylic acid (TCA) cycle or is directly catalyzed by lactate dehydrogenase for glycolysis in the cytoplasm, severally. The TCA cycle in the mitochondria unifies the urea cycle, de novo lipogenesis, and amino acid metabolism via its intermediates. Among the intermediates, citrate is transformed into acetyl-coenzyme A via ATP-citrate lyase and finally produces fatty acid for proliferative cellular membranes and energy production signaling molecules’ second messenger generation. In addition, fumarate, as an intermediate from the TCA and urea cycles, acts as the bridge between the two cycles. Moreover, another nutrient fuel for biosynthesis other than glucose glutamine, a dispensable amino acid, enters cells through the alanine/serine/cysteine-preferring transporter 2 (ASCT2) and turns into glutamate via glutaminase, which is subsequently transferred into α-ketoglutarate (α-KG) via ALT2 as carbon donors (121). Notably, glutamine also provides nitrogen for other non-essential amino acid synthesis like serine, glycine, alanine, etc. ( Figure 3 ).

Figure 3.

Metabolic rearrangements in pancreatic ductal adenocarcinoma (PDAC). In both primary and metastatic PDAC, metabolic rearrangements render PDAC extraordinary plasticity for the metabolites deprived. Among these rearrangements, mitochondrial respiration unifies several important bioenergetics; it produces precursors for lipid, amino acid, and nucleotide biosynthesizing and contributing to glutamine catabolism. Moreover, serine biosynthesis pathway is derived from glycolysis to produce serine for lipid synthesis, protein synthesis, and one-carbon cycle supporting nucleotide synthesis, methylation reactions, and antioxidant defense. Serine is catalyzed into glycine by serine hydroxymethyltransferase, which is converted to glutathione subsequently.

Until recently, increasing evidence supports the dynamically metabolic transition in metastatic cancers during the changing microenvironment. Interestingly, an article claims that aggressive cancer adjusts its metabolism to adapt to a shifty microenvironment and support growth at every stage during its metastatic cascade (98). Moreover, the authors define two novel concepts about metabolic adjustment: metabolic plasticity and flexibility. Notably, metabolic plasticity refers to the same metabolite used by metastatic cells during its different stages of metastasis, and metabolic flexibility (established on nutrient flexibility) describes different metabolites that can all fulfill the exact metabolic requirement of metastatic cascade in metastatic cells (98). Notably, PDAC, the most aggressive cancer, undergoes complicated metabolic rearrangements. Although metastasis in PDAC is a rare process, distant metastasis and cachexia of PDAC are significant causes of the death of cancer patients. Understanding the metabolic reprogramming of PDAC cells in every stage during its metastasis may promote measurements to destroy its metabolic process via targeting the metabolic vulnerabilities for potential therapy ( Table 1 ).

Table 1.

Clinical evidence of metabolic expression for detection in pancreatic ductal adenocarcinoma (PDAC).

Molecules	Expression	Source	References	Possible pathways
Glycolysis
Glucose	Up	Temporal fasting blood glucose (FBG) profiles before diagnosis (n = 219), for resected volume and grade (n = 526), and for long-term FBG data (n = 103)	(122)	Participating in glycolysis, oxidative phosphorylation (OXPHOS), and tumorigenesis (103, 123)
Pyruvate	Up	RNA sequencing from laser-captured microdissected human PDAC compared with normal pancreatic tissue (n = 21)	(124)	Participating in glycolysis, OXPHOS, and alanine (125)
Lactate	Up	Pancreatic juice from PDAC patients (n = 79) and non-PDAC patients (n = 27)	(126)	Form acidified environment for PDAC invasion and epigenomic reprogramming of cancer-associated fibroblasts (127, 128)
Glut	Up	Immunohistochemical staining for PDAC resected samples compared with adjacent normal pancreatic tissue (n = 39)	(104)	Transporting glucose into the cytoplasm, associated with tumor grade and high density of PD-1+ T cells, and predicting poor prognosis (104, 126)
HK	Up	Gene dataset of resected tumors from primary and metastatic sites of PDAC patients (n = 143)	(129)	Catalyzing glucose into G-6-P and regulating lactate production to promote the growth of PDAC (129, 130)
GPI	Up	Immunohistochemical staining of paraffin sections from paired PDAC and control patients (n = 13)	(131)	Promoting the catalysis of G-6-P into F-6-P and enhancing the metastasis of PDAC (131)
PFK	Up	GeneChip hybridization of paired normal and tumor specimens from PDAC patients (n = 36)	(132)	Catalyzing F-6-P into F-1,6-BP and involved in localized ATP supply at the plasma membrane for poor prognosis (126)
LDH	Up	Immunohistochemical staining of LDHB on resected PDAC tumors (n = 59)	(133)	Catalyzing pyruvate into lactate and demonstrating resistance to therapy with gemcitabine (134)
TPI	Up	Gene dataset of primary PDAC tumors compared with normal pancreas and in metastatic PDAC compared with primary tumors (n = 143)	(129)	Inducing catalysis between DHAP and GA3P but showing no correlation with PDAC progression (135)
Pyruvate to alanine
Alanine	Down	Individual amino acid profiles of PDAC patients (n = 12) and healthy control subjects (n = 12)	(136)	Participating in protein synthesis and maintaining compartmentalized pyruvate homeostasis (137)
ALT	Up	Blood samples obtained from every PDAC patient in 10 years (n = 1787)	(138)	Inducing the catalysis of alanine from pyruvate and predicting the clinical outcome of PDAC patients (139)
Asparagine synthetase	Up	Immunohistochemistry of tumor samples in oral squamous cell carcinoma (OSCC) patients (n = 86); also confirmed by The Cancer Genome Atlas OSCC cohort (n = 279)	(140)	Remarkably associating with PNI+ in cancer cells (140)
SBP
Serine	Undifferentiated	Individual amino acid profiles of PDAC patients (n = 12) and healthy control subjects (n = 12)	(136)	Participating in protein synthesis and generation of glycine (28)
PHGDH	Up	Immunohistochemical analysis of PDAC patients (n = 24) compared with normal adjacent tissues (n = 4)	(110)	Catalyzing 3-PG into p-PYR and interacting with eIF4A1 and eIF4E for the development of PDAC (110, 112, 141)
PSAT	Up	Genome expression datasets of PDAC patients (n = 119) compared with normal subjects (n = 199)	(111)	Catalyzing p-PYR into p-serine and involved in the diagnosis and prognosis of PDAC (111, 112)
PSPH	Up	Mass spectrometry for protein abundance from four cell lines, but blank in PDAC patients (BxPC3M1, BxPC3, Panc1, and MiaPaca2)	(112)	Catalyzing p-serine into serine and promote cancer progression (112, 142)
Serine to glycine
SHMT	Up	Immunohistochemical analysis of PDAC patients (n = 24) compared with normal adjacent tissues (n = 4)	(110)	Catalyzing the reversible interconversion of serine to glycine and promoting the proliferation and colony formation of PDAC (110)
Glycine	Undifferentiated	Individual amino acid profiles of PDAC patients (n = 12) and healthy control subjects (n = 12)	(136)	Participating in protein synthesis and generation of GSH (28)

Neuronal–metabolic interaction through serine

It has been accepted that PDAC has a strong tendency for innervation and abundant PNI. These features indicate the worse prognosis of PDAC (143). Early major research about neuron–tumor interlinks points tumor-inducing neurons in PDAC (144). Neurons in PDAC express abundant neurotrophic factors, including NGF, GDNF, artemin, and neuronal chemokines, promoting the malignancy of PDAC via signaling activation (145). As we described earlier, the New York research team identified the unique group of neoplastic neuroendocrine cells in PanINs of the mouse and human models. Neuropeptide receptor NK1-R is detected in the cells, and NK-1R⁺ cells functionally promote PanIN nutrition utilization, thus assisting PanIN organoid growth. Furthermore, denervation of sensory neurons in murine PDAC reduces NK-1R⁺ cell number and retards PanIN-PDAC progression (33). In this publication, they believe that sensory neurons facilitate PDAC formation and metabolic signaling via the SP–NK1R–STAT3 pathways. To date, other direct shreds of evidence of neurons and cancer development are discovered in PDAC from the metabolic aspect (28, 146). Although PSCs and macrophages in TME can supply PDAC with some metabolites or nutrition, it is insufficient for PDAC to get growth requirements with such metabolites (147). Thus, 5-HT (serotonin), a neuro-regulator, grows in PDAC tissues and cell lines compared with normal pancreas. Incubation of PDAC with serotonin or activator of HTR2B renders the proliferative capacity and inhibition of apoptosis. Moreover, serotonin activation leads to a complex of HTR2B-LYN-p85 that activates PI3K–Akt– mTOR signaling and then enhances the Warburg effect (glycolysis) via overexpressed MYC and HIF-1α. The neurotransmitter serotonin released by axons affects metabolic rearrangements, including glycolytic flux, PPP, and HBP, through the expression of enzymes (146). This work suggests that neurotransmitters enable signaling pathways in PDAC and then incite metabolic rearrangements, influencing the malignancy or survival of PDAC afterward. To sum up, the previous research connects neural–metabolic crosstalk in PDAC through chemokines, neurotrophins, and neurotransmitters to investigate the role of innervation during PDAC development. Innovatively, Hindson and colleagues hypothesized that neurons could directly promote metabolic support for PDAC cells under nutrient deprivation in PDAC (91). To assess, micro-fluidic devices that could isolate axons from the neuronal body of DRGs were applied to stimulate an in vivo situation where PDAC cells in nutrient-deprived TME are only exposed to peripheral axons, but not the neuronal bodies (148). Surprisingly, the detection of the conditional axonal media exhibits AA levels, including serine and glycine, suggesting that axons release AA into a nutrient-poor environment. There is a heterogeneity of human PDAC about their dependence on exogenous serine. Approximately 40% of human PDAC cell lines concretely depend on exogenous serine for proliferation due to a lack of expression or ability to augment PHGDH and PSAT1, while mouse PDAC cells can promote serine synthesis and are insensitive to serine starvation. However, the repressed growth of PDAC could be rescued by a co-culture with axons in the microfluidic devices (149). Mechanistically, they noticed reduced mitochondrial activity in serine-dependent PDAC under serine deprivation and a rapidly increased oxygen consumption rate (OCR). This change is similar to utilizing an inhibitor of protein synthesis, cycloheximide (150), which indicates that serine starvation selectively suppresses the mRNA translation rates for protein synthesis, causing decreased mitochondrial activity and increased OCR. More specifically, the authors explored the concrete effects through destabilized green fluorescent proteins (151), which replace all serine codons with one of the six serine codons. Among them, the translation efficiency of TCT and TCC is repressed because of ribosome stalling as revealed by ribosome profiling (152). Interestingly, NGF is intensely expressed by serine-dependent PDAC cells under serine deprivation because the selective translation allows the synthesis of TCC and TCT-low coded proteins, exhibiting enhanced nerve infiltration for serine and glycine secretion to restore PDAC growth. These observations are consistent with in vivo models. Exogenous serine-dependent PDAC under serine-starved conditions shows a significantly smaller tumor mass, accelerating NGF secretion and accumulating sympathetic and sensory innervation. In treating mice with LOXO-101, the NGF receptor inhibitor adversely reduces tumor mass and innervation under a serine-free diet compared with the serine-free-diet mice or control diet with LOXO-101 mice (153). Moreover, PDAC tumors with augmented levels of PHGDH typically have attenuated NGF, tumor innervation, and shorter OS, which is also consistent with other results (28). In summary, this article successfully testifies the metabolic assistance of neurons in PDAC, rendering great metabolic plasticity via neural innervation, which also hints us to concentrate on other aspects in TME for metabolic arrangements. However, as the first published document implicating the function of serine and innervation, there are still unclear aspects of this article, mainly about the mechanisms allowing neurons to release amino acid to support PDAC cells metabolically, mechanisms about the selective translation efficiency of codons, and why PDAC would lose the ability to synthesize serine? More importantly, clinical transformation based on the observed phenomenon would promote novel therapeutic potential targeting PDAC metabolic and neural crosstalk for PDAC patients ( Figure 4 ).

Figure 4.

Neural and metabolic interactions in pancreatic ductal adenocarcinoma (PDAC). A large fraction of human PDAC cells lack SBP enzymes (PHGDH and PSAT-1) and cannot generate serine, and other pathways like autophagy and pancreatic stellate cells cannot generate sufficient levels of serine for growth in PDAC cells under serine deprivation. Therefore, these PDAC cells which must import serine for growth are exogenous serine-dependent PDAC. Moreover, those with SBP enzymes (PHGDH and PSAT-1) for the generation of serine from glucose are exogenous serine-independent PDAC. To date, neurons could directly promote metabolic support (releasing serine) for PDAC cells under the serine-deprived conditions in PDAC.

Recently, some documents gave an overview on serine and other metabolites in neural–cancer progress (154). Serine is considered to not only facilitate cancer progress but also maintain neurites’ formation—for example, NGF induces the phosphorylation of RhoA on serine 188 and reduces RhoA–Rho-related kinase for neurite growth (155); another in vitro experiment declaims that NGF-associated phosphorylated STAT3 at serine (727) is involved in neurite growth (156). Apart from serine, additional metabolites are associated with neural innervation in cancers. For a start, solute carrier family 2 member 3 (SLC2A3), asparagine, and asparagine synthetase are all remarkably expressed in PNI+ cancer. Overexpressed SLC2A3, increased glucose, and reduced vitamin C indicate a diminished OS in colorectal cancer patients and AML patients (157, 158). However, the concrete part of SLC2A3 in neural–metabolic crosstalk is not excavated. The immunohistochemistry of OSCC clinical study (n = 86), together with The Cancer Genome Atlas database (n = 267), presents that asparagine synthetase is notably added in PNI+ tumors, and L-asparagine is the only amino acid able to indicate PNI with satisfactory sensitivity and specificity (140). Regrettably, these studies fail to trench the concrete mechanisms of the metabolic–neural functions during cancer development. Banh and colleagues introduce an excellent method to excavate metabolic–neural interactions in cancer through microfluidic devices to isolate neurons and cancer cells in nutrition-deprived conditions to stimulate an in vivo environment, and we believe that utilizing the unique method could elaborate the specific mechanisms of metabolic–neural interlinks during cancer evolution, especially metabolites like serine and glycine.

Conclusion

To sum up, we first summarized the epidemiology of PDAC, especially risk factors and clinical symptoms, and then introduced a classification for PDAC and related genomic alterations. Notably, the pancreas is a highly innervated organ, so we also discussed three different nerve innervation and pro/anti-oncogenic functions in PDAC. Lastly, we mainly reviewed metabolic rearrangements in PDAC and commented on a recent article about neural and metabolic crosstalk. This review concludes the relationship between innervation and PDAC and also recommends a novel method to excavate metabolic–neural interactions in cancer through microfluidic devices to isolate neurons and cancer cells introduced by Banh and colleagues. We believe that a more direct proof of metabolic–neural crosstalk in PDAC remains unclear, especially metabolites like amino acids. Hopefully, this article could further explore the pathogenic mechanisms of PDAC and then induce the transformation of detection and treatment for PDAC considering the innervation.

Author contributions

RC and MD contributed to the conception and design of the study. LC wrote the first draft of the manuscript. YX and MD wrote sections of the manuscript and provided the critical revisions. All authors contributed to the article and approved the submitted version.

Funding

This article is funded by following projects: Special Medical and Health Personnel Program of Jilin Province Finance Department (2019SCZT015) and Health Technology Innovative Program of Jilin Province (2018J053).

Conflict of interest

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

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Glossary

2-HG	2-hydroxyglutarate
3-PG	3-phosphoglycerate
3-pPYR	3-phosphohydroxypyruvate
5hmC	5-hydroxymethylcytosine
AA	amino acid
ACC	acinar cell carcinoma
Ach	acetylcholine
ACLY	ATP-citrate lyase
ADM	acinar to ductal metaplasia
ADRB	β-adrenergic receptors
ALT1	Alanine aminotransferase 1
AML	acute myeloid leukemia
ANS	autonomic nervous system
AP	acute pancreatitis
ASCT2	alanine/serine/cysteine-preferring transporter 2
BDNF	brain derived neurotrophic factor
BL	Burkitt lymphoma
CAFs	cancer-associated fibroblasts
CGRP	calcitonin gene-related peptide
CHX	cycloheximide
CNS	central nerve system
CoA	acetyl-coenzyme A
CP	chronic pancreatitis
DON	6-diazo-5-oxo-I-norleucine
DRG	dorsal root ganglia
ECM	extracellular matrix
EZH2	enhancer of zeste homolog 2
F-6-P	fructose-6-phosphate
G-6-P	glucose-6-phosphate
GDNF	glial cell line-derived neurotrophic factor
GFAT1	glutamine-fructose aminotransferase 1
GFPs	green fluorescent proteins
GLUT1	glucose transporter 1
GPCR	G-protein coupled receptors
GPT	glutamate pyruvic transaminase;
GSH	glutathione
GSK-3	glycogen synthase kinase 3
H3K27me3	H3 trimethylation;
HBP	hexosamine biosynthetic pathway
HK	hexokinase
HPLC	high-performance liquid chromatography
IPMN	intraductal papillary mucinous neoplasm
ISO	isoproterenol
ITPN	intraductal tubulopapillary neoplasm
KC	LSL-Kras +/LSL-G12D;Pdx1-Cre
KPC	LSL-Kras +/LSL-G12D;LSL-Trp53 +/R172H;Pdx1-Cre
LDH	lactate dehydrogenase
LGICs	ligand-gated ion channels
mAchRs	muscarinic acetylcholine receptors
MAPK	mitogen-activated protein kinase
MCN	mucinous cystic neoplasm
MMP2	matrix metalloprotease 2
nAchRs	nicotinic acetylcholine receptors
NB	β-blocker
NBB	no β blockers
NE	norepinephrine
NG	nodose ganglia
NGF	neural growth factors
NK-1R	neurokinin receptor 1
NSBB	non-selective β-blockers
NT	neurotrophins
NTF3	neurotrophic factor 3
OCR	oxygen consumption rate
OS	overall survival
OXPHOS	oxidative phosphorylation
PanNEN	pancreatoblastoma, pancreatic neuroendocrine neoplasm
PCSCs	pancreatic cancer stem cells
PDAC	pancreatic ductal adenocarcinoma
PET-CT	positron emission tomography computed tomography
PFK	phosphofructokinase
PHGDH	phosphoglycerate dehydrogenase
PHI	phosphohexose isomerase
PKA	protein kinase A
PNI	perineural invasion
PPP	pentose phosphate pathway
Ppt-a	pre-pro-tachykinin-A
PSAT1	phosphoserine aminotransferase 1
PSCs	pancreatic satellite cells
p-serine	phosphoserine
PSNS	parasympathetic nervous system;
PSPH	phosphoserine phosphatase
PTM	post-translational modification
ROS	reactive oxygen species
SAM	S-adenosylmethionine
SB1B	β-1 selective blocker
SBP	serine biosynthesis pathway
sHH	sonic hedgehog
SHMT2	serine hydroxymethyltransferase 2
SP	substance P
SPN	solid pseudopapillary neoplasm
TCA	tricarboxylic acid
TME	tumor microenvironment
TNBC	triple negative breast cancer
TPI	triose phosphate isomerase
TRPV1	transient receptor potential vanilloid 1
UDP-GlcNAc	uridine diphosphate N-acetyl glucosamine
α-KG	α-ketoglutarate