Background
The autonomous nervous system has been recently discovered to impact on cancer growth and progression in several solid and hematological cancers [
1]. In pancreatic cancer (PCa), surgical denervation via vagotomy or pharmacological suppression of the cholinergic signaling were shown to exert a cancer-promoting effect [
2,
3]. In genetically induced LSL-Kras
+/G12D;Pdx1-Cre (KC) mouse model of PCa, subdiaphragmatic vagotomy led to accelerated cancer growth, and treatment with direct muscarinic agonists restored normal KC phenotype [
3]. Here, cholinergic signaling was shown to suppress tumorigenesis and cancer stemness via muscarinic type 1 receptor (M
1R) signaling [
3].
However, activation of muscarinic receptors by its main direct agonist, i.e. acetylcholine (ACh), is a process that is not confined to the autonomous nervous system. In fact, non-neuronal cholinergic signaling is highly common in most cell types and has been shown to regulate basic cell functions, such as proliferation, differentiation and apoptosis [
4‐
6]. In this context, the role of non-neuronal acetylcholine as a local signaling molecule is often disregarded [
7,
8]. Together with its corresponding degrading and synthesizing enzymes (acetylcholine esterase/AChE and choline acetyltransferase/ChAT), it is expressed in many eukaryotic cell types and even in plants and primitive uni- and multicellular organisms with no autonomous nervous system [
8]. Depending on the muscarinic receptor subtype (M
1R – M
5R) to which ACh binds, muscarinic signaling can result in diverse cellular functions. The most potent and relevant ACh receptors that mediate cell proliferation and cell growth are muscarinic receptor type 1 (M
1R) and type 3 (M
3R); both widely expressed in most human tissues and especially in gastrointestinal tissues [
5,
9]. Local availability of ACh for autocrine and paracrine stimulation of muscarinic receptors regulates not only various physiological cell functions, but has also been shown to critically contribute to tumorigenesis [
4]. For instance, in colon, breast and liver cancer, muscarinic receptor activation increased cancer cell proliferation and contributed to cancer progression [
10‐
12]. This effect was mainly attributed to M
3R signaling [
13]. Interestingly, in PCa, M
1R signaling resulted in reduced tumor growth [
3]. This effect was mainly attributed to neuronal cholinergic input, e.g. from the vagus nerve [
3]. However, in the PCa microenvironment, there are though several other non-neuronal sources of acetylcholine, such as cancer-associated fibroblasts and pancreatic stellate cells (PSCs), which are thought to influence pancreatic exocrine function via ACh secretion [
14].
In this study, we demonstrate that PCa cell growth can also be decelerated by non-neuronal, indirect cholinergic signaling. This observation suggests that cancer cells, especially pancreatic cancer cells (PCCs), may be largely independent of the autonomous nervous system in their reaction to acetylcholine availability in the tumor microenvironment. Indeed, we demonstrate that human PCCs express high amounts of AChE and that inhibition of non-neuronal AChE suppressed PCC viability and invasion in vitro and in vivo. Notably, this effect was induced without surgical vagotomy, but only through non-neuronal, tumor cell intrinsic AChE inhibition. However, survival in a novel genetic, R0-resectable PCa mouse model was not influenced by AChE inhibition in the adjuvant setting. Accordingly the survival of resected PCa patients did not differ based on tumor AChE expression levels. These data imply that direct cholinergic stimulation, rather than indirect activation via AChE blockade, may be a more effective therapeutic strategy in PCa.
Methods
Cell culture
The human pancreatic cancer cell lines Panc-1 and SU86.86, the colon carcinoma cell lines SW620 and DLD-1, and the glioblastoma cell line LN229 were purchased from the American Type Culture Collection (ATCC). The T3M-4 cell line was a kind gift by Dr. R. Metzgar (Durham, NC, USA). Cells were kept and cultured in RPMI-1640 supplemented with 10% fetal calf serum (FCS), 100 U/ml penicillin and 100 μg/ml streptomycin (Gibco, Invitrogen, Karlsruhe, Germany) in a 5% CO2 humidified atmosphere at 37 °C.
Matrigel invasion assay
Five thousand PCCs (SU.86.86) were placed into each chemotaxis chamber insert of a 24-well plate (BD Falcon® 8 μm, Heidelberg, Germany) and incubated overnight. After 22 h, the inserts were removed, cleaned of non-migrating cells, fixed in 4% paraformaldehyde, and stained with Vybrant CFDA SE Cell Tracker Kit (Life Technologies, Darmstadt, Germany), and scanned via an automated digital epifluorescence microscope (Keyence BioRevo BZ-9000, Neu-Isenburg, Germany). The number of stained (migrated) cells was counted via ImageJ (version 1.44p, NIH, USA).
Heterotypic xenograft model
Athymic nude mice (NMRI-Foxn1nu/nu) of 4–5 weeks of age and weighing 15-20 g were kept under standard conditions in sterile cages and given food and water ad libitum. Mice were injected subcutaneously (s.c.) in their neck and dorsum with 4 × 105 cells (5000 cells/μl) of the PCa cell line T3M-4, and the animals were divided into two groups for the subsequent treatment: Group I was classified as the “prophylactic” group and treated starting with the day of tumor cell inoculation. Group II were not treated until the 1st week after tumor inoculation to allow the tumor to reach a palpable size. Mice were treated with subcutaneous injections of the AChE inhibitors physostigmine or pyridostigmine as indirect parasympathomimetic agents. Physostigmine, which can cross the blood-brain-barrier (BBB), was administered at 0.1 x LD50 and 0.3 x LD50, and Pyridostigmine, which cannot cross the BBB, at 0.2 x LD50 and 0.4 x LD50. Solvent (saline) was injected to the control group. The animals received treatment 5 times a week for a period of 4 and 3 weeks for group I and II respectively. Animals were sacrificed by neck dislocation, and tumor diameter (mm) and local invasive spread (visible cell spread into neighbouring organs) of cancer cells were assessed.
R0-resectable, electroporation induced transgenic mouse model of unilocular PCa
Current oncogenic
Kras-based mouse models of PCa develop multilocular tumors due to constitutive activation of the oncogene in the embryonic phase or due to its inducible activation in the adult age. This modality is not in harmony with human disease, which typically manifests as a single, i.e. unilocular, cancer in the pancreas. To address this discrepancy, Gürlevik et al. recently developed an R0-resectable, electroporation-induced genetic mouse model of unilocular PCa, which is induced upon injection and electroporation of plasmids containing the
Sleeping Beauty (SB) transposase SB13, a Kras-G12V encoding transposon, and the Cre recombinase into the pancreatic tail of p53floxed mice (
p53fl/fl) [
15]. Details on the plasmid constructs and the electroporation parameters have been described in the original publication [
15]. Upon activation of the Cre recombinase, tumor formation was initiated in a local fashion (the “Pfl” model), and 3 weeks after electroporation, the mice developed a unilocular tumor of the pancreatic tail that is amenable to surgical resection (Fig.
5a-b). The model allows real-life-like performance of neoadjuvant and adjuvant therapy trials with these mice, which exhibit a strongly similar phenotype to the classical, oncogenic Kras-based mouse models of PCa such as
Ptf1a-Cre;LSL-KrasG12D (KC) and
Ptf1a-Cre;LSL-KrasG12D;p53R172H (KPC) [
16,
17]. Using this model, we applied adjuvant chemotherapy combining gemcitabine with either physostigmine (at 0.2xLD
50, i.e. 160 μg/kg,
http://datasheets.scbt.com/sc-252784.pdf), or with pyridostigmine (0.2xLD
50, i.e. 520 μg/kg,
http://www.vetpharm.uzh.ch/reloader.htm?wir/00000015/5975_08.htm?wir/00000015/5975_00.htm), applied s.c. three times a week. Gemcitabine was administered as 6 repeats of weekly gemcitabine (100 mg/kg bodyweight diluted in physiological NaCl) intraperitoneally (i.p.), as shown previously [
15].
Multiplex enzyme-linked immunosorbent analysis (ELISA)
Protein levels of IL-6, IL-10 and TNFalpha were measured in mouse serum via the Luminex® MAGPIX® multiplex ELISA system (Merck Millipore, Darmstadt, Germany) according to the instructions of the manufacturer.
MTT viability assay
To assess human PCa cell line growth, the MTT (3-(4, 5-methylthiazol-2-yl)-2, 5-diphenyl-tetrazolium bromide) assay was used. Cells were seeded at a density of 2000 cells/well in a 96-well plate in serum-free RPMI-1640 medium. Treatment of cells with acetylcholine (Sigma-Aldrich, Taufkirchen, Germany), physostigmine, pyridostigmine or carbachol (all three provided by the internal pharmacy of the Klinikum rechts der Isar) began 12 h after seeding at the concentration of 10, 20, 50, 100 or 300 ng per well (in 100 μl) for physostigmine and pyridostigmine, at 100, 500 and 1000 μM for acetylcholine and at 1 μM, 10 μM, 100 μM, and 1 mM for carbachol. The viability was measured at 0 h, 24 h, 48 h and 72 h after adding the MTT to each well (50 μg/well) and allowed to incubate for 4 h. The formazan products were solubilized with 100 μl of propan-2-ol and the optical density was measured using a photometer at 570 nm.
Cell cycle analysis
Upon reaching 90% of confluence, T3M-4 cells were treated with ACh at a concentration of 1000 μM, physostigmine and pyridostigmine at 30 ng/μl each and the combined agents at 30 + 30 ng/ μl. The PCC were then harvested, centrifuged at 200×g for 5 min and washed 2 times using phosphate-buffered saline (PBS). They were then resuspended in 1 ml of ice-cold PBS and added dropwise afterwards to ice-cold absolute ethanol for cell fixation. Cells were fixated for 24 h at 4 °C, recentrifuged and resuspended in 500 μl Triton X-100 (Sigma) in PBS with added 100 μg of DNAse-free RNAse A (Sigma) propidium iodide (PI) at a concentration of 20 μg/ml. Cells were then incubated for 15 min at 37 °C and pipetted afterwards into 96-well plates protected from light for data acquisition. Forward and side scatter was measured using a Guava® easyCyte HT Sampling Flow Cytometer.
Immunoblot analysis
At 90% confluence, PCa cell lines were lysed and 30 μg of protein was separated, electroblotted and the membrane was exposed to monoclonal and polyclonal antibodies (Table
1) at 4 °C overnight. The equal loading of AChE-Blots was assured by re-probing with alpha-Tubulin antibody. The densitometric analysis of the Western Blot was performed via ImageJ (version 1.44p, NIH, USA).
Table 1
Primary antibodies. IF: immunofluorescence, WB: Western blot, IHC: immunohistochemistry
p44/42 MAPK (Erk1/2) | 137F5 | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
Phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) | D13.14.4E XP® | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
Src | 36D10 | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
Phospho-Src (Ser17) | D7F2Q | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
AMPKα | D5A2 | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
Phospho-AMPKα (Thr172) | 40H9 | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
p38α MAPK | 7D6 | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
Phospho-p38 MAPK (Thr180/Tyr182) | D3F9 XP® | Rabbit | Cell Signaling, Leiden, The Netherlands | 1:1000 |
Anti-Acetylcholine-esterase (AChE) | | Rabbit pAb | Prestige Antibodies®, Sigma-Aldrich, Taufkirchen. Germany | 1:200 (IF) 1:1000 (WB) 1:400 (IHC) |
Anti-alpha-Tubulin | Ab11034 | Rabbit | Abcam, Cambridge, UK | 1:10.000 |
F4/80 | Ab6640 | Rat | Abcam, Cambridge, UK | 1:75 |
CD45 | Ab10558 | Rabbit | Abcam, Cambridge, UK | 1:100 |
ChAT | Polyclonal | Rabbit | Kindly provided by Prof. M. Schemann, TU Munich | 1:1000 |
Phospho-kinase profiling
The Proteome Profiler Human Phospho-Kinase Array Kit (R&D Systems, Minneapolis, MN, USA) was used to obtain a semiquantitative comparison of the phosphorylation of 43 different human kinases in T3M4 cells that were either treated with 30 ng/μl physostigmine for 5 min or untreated, according to the instructions of the manufacturer.
Patients and human tissue
Tissue samples from patients undergoing pylorus-preserving Whipple’s procedure for PCa of the pancreatic head were collected and processed as described before [
18]. A total of 39 patients were included for the survival analyses (for patient characteristics, please see Table
2). A total of 20 patients were used for the correlation analysis between ChAT content in nerves and Union for International Cancer Control (UICC) stage (median age: 61, 11 male, 9 female, UICC stage distribution: IIA: 4 patients, IIB: 15 patients, III: 1 patient). All patients were informed, and written consent was obtained for tissue collection.
Table 2
Patient characteristics. UICC: Union for International Cancer Control
| 39 |
Sex |
Male | 22 (56.4%) |
Female | 17 (43.6%) |
Age (median; min-max) | 67.4 (31–83) |
UICC |
UICC Ia | 3 (7.7%) |
UICC Ib | 8 (20.5%) |
UICC IIa | 9 (23.1%) |
UICC IIb | 7 (17.9%) |
UICC III | 10 (25.7%) |
UICC IV | 2 (5.1%) |
Immunohistochemistry, immunofluorescence, semiquantitative analysis
Consecutive 3 μm sections from formalin-fixed human tissues were incubated with the corresponding antibodies (Table
1) overnight in a humid chamber at 4 °C. AChE IHC was detected with the Avidin Biotin Complex Peroxidase Standard Staining Kit (Thermo-Fischer, Waltham, USA). Histopathological analysis was performed by two independent observers (PLP, MJ) followed by resolution of any differences by joint review and consultation with a third observer (IED). Scores of 0–3 were given in 0.5 steps according to the amount of immunoreactivity in each tissue samples. For immunofluorescence staining, Alexa® Fluor 488 and 594 antibodies (Invitrogen, Germany, 1:250 concentration) in combination with 4′,6-diamidino-2-phenylindol (DAPI) nuclear stain were utilized.
AChE activity assay
For comparison of the AChE activity between the human PCa cell lines SU86.86 and T3M4, a colorimetric AChE activity assay kit (Sigma-Aldrich, Taufkirchen, Germany), which is based on the Ellman method, was applied according to the instructions of the manufacturer.
Ethics approval
All animal studies were conducted according to the national regulations and approved by the Regierung von Oberbayern (approval nr. ROB-55.2-2532.Vet_02–16-165 and 55.2-1-54-2531-36-08), and Hannover (15/1949). The study has been approved by the ethics committee of the Technische Universität München, Munich (approval nr. 154/20).
Statistical analysis
Results are expressed as mean ± SD. Two-group analyses were performed using the unpaired t-test for continuous values and with the Mann–Whitney U test for scores and indices. Linear regression was used for correlating tissue expression of AChE or ChAT with the UICC stages of PCa. Survival analyses were performed with the log-rank test and depicted as Kaplan-Meier curves. All tests were two-sided, and a p value < 0.05 was considered to indicate statistical significance. All authors had access to the study data and had reviewed and approved the final manuscript.
Discussion
The present study suggested an anti-proliferative and anti-invasive effect of non-neuronal cholinergic signaling in pancreatic cancer. Inhibition of endogenous, non-neuronal AChE decelerated PCC growth and invasiveness in vitro & in vivo, which was linked to intracellularly reduced MAPK phosphorylation and reduced downstream phosphorylation of ERK1/2 and p38. Furthermore, prophylactic cholinergic activation in PCa mouse models with intact vagal innervation reduced both tumor invasiveness in vivo and immune cell infiltration by tumor-associated macrophages. However, administration of parasympathomimetic agents as co-adjuvant therapy together with gemcitabine did not influence the overall survival of mice in a resectable, transgenic mouse model of unilocular genetic PCa. Accordingly, AChE did not correlate to survival in human PCa and was actually suppressed in parallel with ChAT in higher grade tumors. Therefore, our study suggests that for targeting PCa, direct cholinergic stimulation of the muscarinic signaling, rather than indirect activation via AChE blockade, may be a more effective therapeutic strategy.
Various studies have previously reported a cancer-promoting effect of the vagus nerve. In a mouse model of gastric cancer, surgical vagotomy decreased gastric mucosal thickness and cellular proliferation [
26,
27]. The effect was thought to be mediated via muscarinic receptor type 3 (M
3R) signaling, since knock-out of the M
3R suppressed gastric cancer. These studies led to the conclusion that vagal innervation promotes gastric cancer via muscarinic M3 receptor in a Wnt mediated pathway [
28]. However, in PCa, Renz et al. demonstrated that ablation of the vagal nerve actually accelerated cancer progression [
3]. Treatment with the muscarinic receptor agonist, bethanechol was able to reverse the accelerated cancer progression due to vagal ablation. Overall, this study along with others led to the general hypothesis that vagal innervation has a cancer-attenuating effect in the pancreas. In a wide-scale analysis of neural fiber quality in PCa specimens, we previously found a low parasympathetic fiber content of nerves that were invaded by pancreatic cancer cells [
29]. In line with these previous studies, in the current study, we were able to demonstrate a cancer-cell-suppressive effect of AChE inhibition and thus indirect cholinergic activation in vitro and in vivo. However, this effect was obtained without directly interfering with the autonomous nervous system, and yet did also not translate into an improved clinical outcome, i.e. survival, in mouse PCa. These findings are of major importance for all studies related the role of cholinergic / parasympathetic nervous system in cancer, since all components of the cholinergic system (ACh, acetylcholinesterase, muscarinic acetylcholine receptors, acetylcholine transferase) are not exclusively expressed by neurons but ubiquitously present in almost all mammalian cells, including non-neuronal cells [
8].
In order to understand non-neuronal cholinergic signaling and its involvement in basic cellular functions, such as proliferation and differentiation [
6], one has to consider the different subtypes of muscarinic receptors that initiate these diverse cellular outcomes. There are 5 different muscarinic receptor subtypes (M1R – M5R), all of which are G protein-coupled receptors but may lead to different intracellular cascades in order to exert different extracellular outcomes. Upon activation, odd-numbered muscarinic receptors couple to G proteins that activate phospholipase C-β to initiate the phosphatidylinositol trisphosphate cascade, whereas even numbered muscarinic receptors couple to G proteins that inhibit adenylyl cyclase activity [
9]. This complexity explains in part why, for instance, activation of the M
3R subtype has been shown to promote cancer cell proliferation in gastric cancer, whereas activation of M
1R subtype has been shown to attenuated pancreatic cancer proliferation. Although the role of muscarinic receptors in colon cancer has been previously characterized, most studies only focused on one of the 5 different receptors [
9,
13]. Even though there has been extensive research about the tissue-specific expression of muscarinic receptors, a comprehensive overview about the role of muscarinic receptors and its ligands in different cancer entities is still missing.
The lack of basic research on non-neuronal cholinergic signaling is even more evident for PCa. Very little is known about the role of the different muscarinic receptor subtypes as well as other components of the cholinergic-signaling-machinery, such as ChAT, AChE or ACh expression in PCa.
Therefore, our study contributes to the attempts to understand the non-neuronal AChE in PCa. Here, we demonstrated mild to weak staining in premalignant lesions, with increasing staining in overt pancreatic cancer. Mammals express 3 different classes of AChE, which differ with regard to their subunits. Each type of AChE has a different 3′ RNA sequence with a corresponding C-terminal sequence, which encodes the respective subunit. The AChE
H subunit contains a hydrophobic C-terminal sequence forming amphilic monomers and dimers and incorporates a GPI [
6]. It is therefore often found closely spaced to the cell membrane. This would explain the perimembranous staining found in our study.
Based on our findings,
indirect activation of cholinergic signaling via AChE inhibition is not sufficient to achieve a survival benefit in PCa, although it resulted in a prominent suppression of the tumor-associated inflammation in the tumor, and a drop of serum cytokine levels. This conclusion, which is based on our findings from a translational mouse PCa model and from human PCa, underlines that increasing the cholinergic input for attenuating PCa progression and for improving patient survival will probably not be possible via administration of two widely used clinical drugs, i.e. physostigmine and pyridostigmine. In contrast, Renz et al. made use of a direct activator of muscarinic cholinergic signaling. i.e. the bethanechol, which did result in improved survival in the KPC model of PCa [
3]. In the present study, we combined the indirect cholinergics with an older chemotherapeutic, i.e. gemcitabine, in the adjuvant and palliative treatment. It is imaginable that a combination with a more current regimen such as gemcitabine and nab-paclitaxel or with FOLFIRINOX may yield even more potent results. Nonetheless, we observed immunosuppressive effects of indirect cholinergic stimulation in our study. It is conceivable that in a more humanized model, this immunmodulatory effect of indirect cholinergic stimulation may have been much greater. The NMRI-Foxn1
nu/nu model is deficient with regard to T cell function due to a thymus abnormality. For this reason, we primarily assessed macrophage distribution upon treatment with indirect cholinergic agents. As a recent report showed, cholinergic activity of the vagus nerve inhibits macrophage-derived tumor necrosis factor-α secretion via T-cell derived acetylcholine in the spleen [
30]. Our data suggest that there may also be a T-cell-independent, immunosuppressive effect of cholinergic activation on macrophages.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.