Upregulation of RET induces perineurial invasion of pancreatic adenocarcinoma

M Amit1,2,3, S Na’ara2,3, L Leider-Trejo4, Y Binenbaum2, N Kulish2, E Fridman2,3, A Shabtai-Orbach2, RJ Wong5 and Z Gil2,3

Abstract

Tumor spread along nerves, a phenomenon known as perineurial invasion, is common in various cancers including pancreatic ductal adenocarcinoma (PDAC). Neural invasion is associated with poor outcome, yet its mechanism remains unclear. Using the transgenic Pdx1-Cre/KrasG12D /p53R172H (KPC) mouse model, we investigated the mechanism of neural invasion in PDAC. To detect tissue-specific factors that influence neural invasion by cancer cells, we characterized the perineurial microenvironment using a series of bone marrow transplantation (BMT) experiments in transgenic mice expressing single mutations in the Cx3cr1, GDNF and CCR2 genes. Immunolabeling of tumors in KPC mice of different ages and analysis of human cancer specimens revealed that RET expression is upregulated during PDAC tumorigenesis. BMT experiments revealed that BM-derived macrophages expressing the RET ligand GDNF are highly abundant around nerves invaded by cancer. Inhibition of perineurial macrophage recruitment, using the CSF-1R antagonist GW2580 or BMT from CCR2-deficient donors, reduced perineurial invasion. Deletion of GDNF expression by perineurial macrophages, or inhibition of RET with shRNA or a small-molecule inhibitor, reduced perineurial invasion in KPC mice with PDAC. Taken together, our findings show that RET is upregulated during pancreas tumorigenesis and its activation induces cancer perineurial invasion. Trafficking of BM-derived macrophages to the perineurial microenvironment and secretion of GDNF are essential for pancreatic cancer neural spread.

INTRODUCTION

Perineurial invasion (PNI) by cancer is associated with poor prognosis in patients with carcinomas of the gastrointestinal tract, head and neck, pancreas and prostate.1 Most patients with pancreatic ductal adenocarcinoma (PDAC) undergo palliative treatment rather than curative surgery, due to distant metastases or neural spread along extra pancreatic nerves. Most importantly, these patients tend to suffer from debilitating neuropathic pain and poor quality of life.2
The prevalence of PNI varies considerably among cancer types, and reaches 80–100% in pancreatic cancer.3,4 Recent studies have suggested that the tumor microenvironment has a role in cancer dissemination along nerves.5,6 A prominent inflammatory infiltration is present around preinvasive pancreatic lesions and persists through invasive cancer.7 Of relevance, there is evidence that the neural microenvironment of invaded nerves has a unique inflammatory profile.8,9
Recently, we and others described the presence of immunocytes in the perineurial niche.8–10 Clinical and experimental studies have demonstrated a strong association between macrophage density and cancer cell metastasis in PDAC. Furthermore, targeting tumor-associated macrophages by inhibiting either the colonystimulating factor-1 receptor (CSF1R) or chemokine (C-C motif) receptor 2 (CCR2) improves chemotherapeutic efficacy and inhibits metastasis.11–14 In the normal nerve, macrophages that can be derived from either resident immune cells or from circulating monocytes participate in regeneration of peripheral nerves.9,15 However, in the nerve-cancer microenvironment, the origin and polarization of these macrophages, as well as their role in PNI, is not well defined.
It has been suggested that glial-derived neurotrophic factor (GDNF) can also promote PNI.16–18 GDNF expression in human samples was also associated with PNI and reduced survival.19 Immunofluorescent imaging revealed expression of the GDNF receptor, GDNF family receptor α-1 (GFRα1) and its co-receptor RET by pancreatic cancer cells.9 Although previous clinical and experimental data implicated the involvement of RET in PNI, no direct evidence links RET activity with PNI in transgenic animal models or in patients with PDAC. To detect tissue-specific factors influencing neural invasion by cancer cells, we characterized the perineurial environment in human samples and in a series of transgenic mice models expressing mutations in the Kras, p53, GDNF and CCR2 genes.
The current study provides direct evidence that upregulation and activation of RET by perineurial macrophages induce perineurial spread of PDAC.

RESULTS

Macrophages are a prominent component of the perineurial microenvironment

Head and Neck Surgery Department, MD Anderson Cancer Center University of Texas, Houston, TX, USA; The Laboratory for Applied Cancer Research, Clinical Research Institute at Rambam, Rappaport Institute of Medicine and Research, The Technion, Israel Institute of Technology, Haifa, Israel; 3Department of Otolaryngology Head and Neck Surgery, The Head and Neck Center, Rambam Healthcare Campus, Clinical Research Institute at Rambam, Rappaport Institute of Medicine and Research, Rambam Medical Center, The Technion, Israel Institute of Technology, Haifa, Israel; 4Department of Pathology, Tel Aviv Medical Center, Tel Aviv, Israel and 5Department of Surgery Memorial Sloan Kettering Cancer Center, New York, NY, USA. Correspondence: Professor Z Gil, Department of Otolaryngology Head and Neck Surgery, The Head and Neck Center, Rambam Healthcare Campus, Clinical Research Institute at Rambam, Rappaport Institute of Medicine and Research, Rambam Medical Center, The Technion, Israel Institute of Technology, 8 Ha’Aliya Street, POB 9602, Haifa 31096, Israel.
The patterns of inflammatory response secondary to pancreatic tumorigenesis are distinct from those of chronic pancreatitis.8,11 Furthermore, an immune cell profile is stage-dependent in many types of cancer.20–22 To investigate the involvement of immunocytes in the perineurial microenvironment during tumorigenesis, we evaluated pancreata excised from 2-, 3- and 6-month-old KPC mice with normal pancreas, PanIN and PDAC, respectively.
Immunofluorescent analysis revealed significantly greater infiltration of lymphocytes around nerves in PDAC and PanINs than around nerves in normal pancreas (Po0.01, n = 10 per group, Figures 1a and b). Macrophage infiltration was more prominent around nerves invaded by cancer than around nerves in PanIN lesions or normal pancreas (Po0.001, Figures 1a and b). Hence, recruitment of immunocytes to the neural niche occurs during pancreatic carcinogenesis.
To validate our findings, we compared the patterns of infiltration of lymphocytes and macrophages around nerves in human specimens of normal pancreas (resected serous cystadenoma), PanIN lesions and PDACs (Supplementary Table 1). The number of infiltrating lymphocytes around nerves in PanIN and PDAC was significantly higher than around nerves in normal pancreas (P = 0.02, n = 10 per group, Figures 1c and d). Similarly, immunostaining of macrophages revealed increased infiltration around nerves with cancerous PNI compared with nerves within PanIN or normal pancreas (Po0.001).

Characterization of perineurial macrophages in the neural niche

Infiltrating tumor-associated macrophages are abundant in PDAC7 and are associated with nerves invaded by cancer.9 We further characterized the subtypes of infiltrating macrophages around nerves in human specimens of PDAC. We used immunolabeling with CD86 for the M1 macrophages (classically activated) and antiCD163 Ab for the M2 (alternatively activated) macrophages.23 In the perineurial microenvironment of PDAC, we found greater expression of CD163-positive macrophages (M2) than CD86positive (M1) cells (Po0.01, n = 10, Figure 2a). Accordingly, in 6-month-old KPC mice, infiltrating CD163-positive macrophages around nerves were significantly more abundant than CD86positive macrophages (P = 0.01, n = 10, Figures 2b and c). These findings demonstrate that macrophage polarization toward the M2 phenotype occurs near nerves invaded by cancer.

The origin of perineurial macrophages

Macrophages at the perineurial niche can arise from the infiltration of bone marrow-derived monocytes or from resident perineurial macrophages. To identify the origin of perineurial macrophages, we performed bone marrow transplantation (BMT) from Cx3cr1GFP/+ donor mice. The insertion of a reporter gene encoding enhanced green fluorescence protein (eGFP) to the CX3CR1 locus, permits tracking of in vivo-labeled BM-derived F4/80-positive macrophages. Flow cytometry analysis of whole blood from recipient mice, 3 weeks after BMT, revealed full reconstitution of the bone marrow in KPC mice by the donor Cx3cr1GFP/+ grafts. Using mouse spleen as control, we were able to detect the populations of both resident (Cx3cr1BM+/+) and bone marrow-derived (Cx3cr1BMGFP/+) macrophages. Immunofluorescence analysis of the F4/80+ cells around nerves invaded by cancer in BMT-recipient KPC mice revealed that infiltrating macrophages were mainly Cx3cr1BMGFP/+ (Figures 3a and b, Po0.01, n = 9 per group), demonstrating that nerve-associated macrophages are recruited from the bone marrow.

Neural invasion of pancreatic ductal adenocarcinoma is mediated by bone marrow-derived macrophages

Macrophages recruitment to the tumor microenvironment is mediated by CCL2 binding to its receptor, CCR2.24 To assess the role of BM-derived macrophages in neural invasion, we performed BMT using CCR2− / −donor mice (n = 7 per group) and KPC recipients. Wild-type c57/bl (CCR2+/+) donors were used as controls. After full reconstitution of the BM, recipient KPC; CCR2BM − / − or KPC;CCR2BM+/+ mice were grown to the age of 6 months and their tumors were harvested and analyzed for neural invasion indices. F4/80 labeling showed a decreased infiltration of perineurial macrophages in KPC;CCR2BM − / − mice compared with KPC;CCR2BM+/+ mice (P = 0.02, Figures 3d and e). To evaluate the effect of macrophage reduction on PNI, cancer neural invasion indices were measured in a blind manner by two independent pathologists using immunohistochemical staining with anti-S100 ab as a neuron marker. The rate of neural invasion index was significantly lower, and the nerve severity score reduced, in KPC;CCR2BM − / − mice compared with KPC;CCR2BM+/+ controls, suggesting that inhibition of macrophages trafficking to the perineurial niche, reduces PNI (Figures 3g and h). Reduction in perineurial macrophage infiltration had no effect on tumor size (472 ± 128 and 426 ± 211 mm3 for KPC;CCR2BM+/+ and KPC; CCR2BM − / − mice, respectively, P = 0.82, n = 7 per group).
To confirm the role of perineurial macrophages on PNI, we also blocked the recruitment of circulating monocyte-derived macrophages with the colony-stimulating factor 1 (CSF-1R) antagonist GW2580.25 Tumor bearing (with a measurable 3–6 mm tumor) KPC mice were assigned to treatment and matched using small animal ultrasound.26 Treatment groups included vehicle (DMSO 0.1%) or GW2580 (80 mg/kg in DMSO 0.1%). After randomization using simple randomization sheet, animals were treated twice daily and tumor growth was monitored weekly by ultrasound for 4 weeks. The ultrasound technician was blinded to the group allocation. Reduction in the number of recruited monocytes was confirmed by F4/80 staining (P = 0.01, Figure 3f, n = 6 group). Although tumor size did not differ significantly between vehicle and GW2580 groups (484 ± 234 and 520 ± 196 mm3, respectively, P = 0.6), the neural invasion index and invasion severity scores were significantly higher in the control group than in GW2580 treated animals (mean neural invasion index P = 0.03 and mean invasion severity P = 0.002, Figures 3i and j). These findings indicate that perineurial macrophages positively regulate nerve invasion by PDAC.

GDNF expression by perineurial macrophages induces nerve invasion

GDNF is expressed by macrophages and participates in nerve growth and repair. Pancreatic cancer cells express several types of GFRα.17 To assess the role of perineurial macrophages in neural invasion, we first analyzed GDNF expression by CD68+ perineurial macrophages around nerves in 10 patients with PDAC. Figure 4a shows significantly higher expression of GDNF by perineurial macrophages in the vicinity of invaded nerves than in that of noninvaded nerves (Po0.01, n = 10, see Figure 4c for analysis). GDNF expression by nerve-associated macrophages was also significantly greater around nerves invaded by cancer than around noninvaded nerves in KPC mice (P = 0.02, n = 10, Figures 4b and d).
To further assess the effect of GDNF on nerve invasion, KPC mice underwent BMT from GDNF-deficient donors (KPC; GDNFBM+/ −) or controls (KPC;GDNFBM+/+) at the age of 2 months before PDAC evolved (n = 8 per group). Figure 4e (i) shows a decreased GDNF expression in S100 positive nerve derived from GDNF+/ − mice compared with wild-type mice. The numbers of infiltrating macrophages around invaded nerves in both KPC; GDNFBM+/ − and KPC;GDNFBM+/+ were similar (Figure 4f). However, the neural invasion index in KPC;GDNFBM+/− mice was significantly smaller than in KPC;GDNFBM+/+ controls (P = 0.001, Figure 4g). The neural invasion severity in the KPC;GDNFBM+/+ controls was also significantly greater than in KPC;GDNFBM+/− (Po0.001, Figure 4h).

Upregulation of the RET proto-oncogene during carcinogenesis is involved in PDAC nerve invasion

Finally, we sought to investigate the involvement of the GDNF tyrosine kinase receptor RET in the development of PNI. We first analyzed the patterns of RET expression at the perineurial niche in 30 pathological specimens excised from patients with PDAC. We performed double immunofluorescence with RET and a cytokeratin-19 (CK19) to ascertain that RET-expressing cells are of pancreatic origin (Figure 5a). Figure 5b shows significantly higher expression of RET in PDAC specimens than in PanINs or in normal pancreata (Po0.01, n = 10 per group). Immunofluorescent analysis of KPC mice during PDAC tumorigenesis revealed significantly enhanced RET expression in PDAC than in PanINs or normal pancreas (Po0.01, n = 8 per group, Figure 5b). The expression of the RET co-receptor GFRα1 remained stable throughout tumorigenesis (Figures 5c and d).
We previously demonstrated the role of RET in cancer cell migration toward endoneurial macrophage-conditioned media.9 To further assess the role of RET in neural invasion, KPC4139 murine PDAC cells were transfected with short hairpin RNA (shRNA) oligonucleotides directed against RET (the knockdown yield of three shRET constructs is shown in Supplementary Figure 1). Orthotopic tumors were induced by orthotopic injection of 5 × 105 cells to the pancreata of WT c57/bl mice (n = 10 per group). Stable RET knockdown was verified by immunostaining and western blotting (Figures 5e and f, respectively). There was no difference in tumor size between RET-deficient and control PDAC tumors (P = 0.3, Figure 5g). Neural invasion analysis27 revealed that RET-deficient tumors exhibited significantly lower rates of neural invasion index and invasion severity score, compared with controls (P = 0.008 and P = 0.03, respectively Figure 5h).

DISCUSSION

Perineurial invasion is one of the hallmarks of pancreatic adenocarcinoma and a significant source of morbidity and mortality. While the tendency of cancer to track along nerves is well recognized in the clinical literature, the mechanism of perineurial invasion is not clear.28 Cancer perineurial invasion can be a source of distant tumor dissemination and, for some tumors, may be the major route of metastatic spread.29 PNI is also emerging as an important clinical problem for cancer patients who suffer from severe morbidity due to nerve deficits and neuropathic pain induced by cancer invasion.30 Using in vitro and in vivo animal models we previously showed that PDAC has a unique tendency to track unidirectionally along nerves toward the soma.31
The nerve is a dense and highly oxygenated environment that is rich in immune cells, including perineurial macrophages. These macrophages enrich the nerve niche with growth factors required for neural sprouting and regeneration.32,33 Using four distinct transgenic models, we provided clinically significant novel evidence regarding the role of these perineurial macrophages in PNI. First, we demonstrated for the first time that perineurial macrophages are recruited from the bone marrow to the neural niche. Second, we showed in mice models and in human samples that bone marrow-derived macrophages are polarized toward the M2 phenotype at the perineurial microenvironment. These perineurial macrophages are actively involved in mediating perineurial invasion by secretion of GDNF and activation of the RET proto-oncogene on cancer cells. Our data also demonstrated the upregulation of RET expression during pancreatic carcinogenesis. We also showed that RET inhibition with shRNA can block cancer invasion along nerves.
RET is an oncogene known to promote medullary thyroid, breast, lung and pancreatic carcinomas. Expression of the GDNF family of ligands and their receptors in PDACs is associated with poor outcome.34–38 As a result of RET phosphorylation by the GDNF-GFRα1 complex, ERK and AKT are activated, inducing invasion and proliferation of PDAC cells.39–42 Inflammatory response might augment the effect of GDNF.18,34
We are aware of several limitations that constrain our conclusions. First, CX3CR1 is not specifically expressed by macrophages; hence, it is possible that some of the GFP coexpressing cells represent dendritic, natural killers or T cells. However, this lack of specificity has been minimized by using F4/80 as a co-marker specific for macrophages and adjusting for background nonspecific staining using the spleen as a control. Second, CCR2 is expressed by other CD4 T cells, which might obscure our results. Still, our findings show no specificity of T-cell infiltration around invaded nerves. Bone marrow implanted KPCBMCCR2 − / − did not show a difference in CD3+ cell populations around nerves, compared with KPCBMCCR2+/+. Furthermore, we used different approaches to target bone marrow-derived monocytes, specifically the use of GW2580.25 Finally, the use of immunocompetent animals for the RET knockdown experiments limited our ability to test human cell lines. However, we used KRAS-driven tumors that reliably recapitulate the human spectrum of the PDAC phenotype.43
We showed that inhibiting macrophage recruitment to the tumor microenvironment, by blocking CSF1R, reduces the number of invaded nerves (that is, invasion index); this finding suggests that PNI is highly dynamic and can be reversible. However, whether this contributes to an interruption in the migration process or to signals that induce cancer cell resistance to apoptosis at the perineural niche remains to be determined.
Nevertheless, the study suggests two strategies to target PNI: inhibition of macrophage trafficking by blocking CSF1R, and inhibition of RET by small-molecule inhibitors. A CSF1R antagonist, PLX3397, is under investigation in clinical trials for patients with advanced solid tumors.44 RET inhibitors are also used as standard of care in thyroid and kidney cancers. Adjuvant therapy directed against neural invasion could theoretically prevent the progression of cancer along the nervous system and reduce cancer related neuropathy.

MATERIALS AND METHODS

Cell cultures

The murine pancreatic carcinoma cell line KPC4139 was derived from 6-month-old KPC mice bearing advanced PDAC. Tumor tissue from KPC mice was dissociated in collagenase V (Sigma-Aldrich, Rehovot, Israel) for 30 min at 37 °C; after incubating overnight at 37 °C in 5% CO2, the medium was replaced with fresh medium. After 10 passages, cells were cultured in Dulbecco’s Modified Eagle’s Media (DMEM) + 10%FBS for four passages.43 The cell line was maintained in continuous exponential growth by twice weekly passage in DMEM (Life Technologies, Inc., Gaithesburg, MD, USA) supplemented with 10% fetal bovine serum, 100 units/ml penicillin and 10 mg/ml streptomycin in a humidified incubator containing 5% CO2 in air at 37 °C. No cell line authentication was performed, as all cell lines used in this article were obtained from primary cultures of tumors derived from KPC mice.

Human samples

Tissue samples were collected from patients operated at Rambam Medical Center tissue bank, Israeli Institute of Technology (Institutional Review Board No. 0238-13-RMB), following pancreatectomy for PDAC or high-risk lesion (for example, suspicious cyst). Supplementary Table 1 presents patients’ demographic and clinical data. Hematoxylin and eosin (H&E)stained sections were collected and previewed to confirm pathological diagnoses and to identify specimens also containing pancreatic intraepithelial neoplasia (PanIN) and/or histologically normal pancreatic exocrine tissue. Only specimens containing adjacent regions of histologically normal pancreatic tissue were included in this study.

Neural invasion quantification

Neural invasion indices were measured as we and others previously described using standard neural invasion assessment.4,30,33,45,46 For neural invasion indices analysis, slides were stained with the nerve tissue marker, S100, to demonstrate intrapancreatic nerves. Stained slides and their matching H&E slides were scanned using 3D-Histech scanner (Sysmex AG, Haifa, Israel), in a 0.465 μm/pixel resolution (Supplementary Figure 2). Next, nerve invasion was assessed by two independent pathologists (L.L.T. and I.N.) using Panoramic Viewer (3D-Histech, Sysmex AG). For analysis of neural invasion indices, nerves were blindly categorized as non-invaded (0), epineural association (ENA) (1), perineurial invasion (PNI) (2) or endoneural invasion (ENI) (3), as shown in Supplementary Figure 2. A neural invasion severity score was generated using the following standardized formula: n(ENA) × 1+n(PNI) × 2+n(ENI) × 3.4,9,30,31,33,45,46 The neural invasion index was calculated as the ratio between the number of invaded nerves (PNI or ENI) and the total number of nerves.9,31 For each experiment, the values represent an average between two observers.

Immunohistochemistry

We used the following primary antibodies: rabbit anti-mouse cytokeratin (CK) 19 antibody (1:200; ab15463 Abcam Cambridge, UK), rat anti-mouse/human F4/80 (1:250; AbDSerotec, Kidlington, UK), rat anti-mouse/human CD3e (1:50; AbDSerotec, Kidlington, UK), mouse monoclonal anti-CD4 antibody (1:250; clone ab51312 Abcam), rabbit polyclonal anti-CD8 antibody (1:200; clone ab4055 Abcam), rabbit polyclonal anti-GDNF (1:200 Santa Cruz Biotech, Santa Cruz, CA, USA), rabbit monoclonal anti-CD68 antibody (EPR1392Y; 1:100; clone ab76308, RabMab, Abcam), rabbit monoclonal anti-Ret (EPR2871; 1:200; clone ab134100 Abcam), rabbit monoclonal CD163, rat monoclonal CD86 antibody (1:200 R&D Systems, Minneapolis, MN, USA) and rat anti-mouse CD31 antibody (1:100; Dianova, Hamburg, Germany).

Animals and bone marrow transplantation

All experimental procedures were performed in accordance with the Technion, the Israeli Institute of Technology Animal Care and Use Committee and the Department of Agriculture regulations.
Wild-type C57BL/6 female breeders, and CCR2−/− mice (strain B6.129S4CCR2tm1Ifc/J),9,48 were obtained from Jackson Laboratories (Bar Harbor, ME, USA). For GDNF-depleted myeloid cell experiments, we used bone marrow from mice deficient of GDNF (a generous gift of Dr M Saarma, nstitute of Biotechnology, University of Helsinki).49,50 Homozygous GDNF-null mice die at birth; hence we used heterozygous GDNF mice. We performed immunofluorescent staining using anti-S100 and anti-GDNF Abs to demonstrate the decrease in GDNF expression in nerves derived from GDNF−/+ mice, compared with wild types (Figure 4e(i)). Western blotting analysis from whole brain lysate of wild type and GDNF-deficient mice is shown in Figure 4e(ii).
CX3CL1 signaling through CX3CR1 has been shown to be a key factor in recruitment of macrophages to tissue lesions or sites of inflammation.51 To investigate the origin of nerve associated macrophages, we performed experiments on BMT using Cx3cr1GFP/+ donor mice (Jax Labs), which express GFP reporter gene.52
To study the role of macrophages, we used 6- to 8-week-old CCR2− / − donor mice (The Jackson Laboratory) that lacked the receptor for CCL2.
In these mice, recruitment of macrophages to the tumor microenvironment is impaired, and infiltration and activation of perineurial macrophages are reduced.53Finally, we used BM from mice deficient in GDNF (generously provided byDrs M Saarma and JF Bertram). Briefly, KPC recipients were lethally irradiated (9.5G) at 5 weeks of age, a time point that was selected to precede the development of the earliest preneoplastic stage. Recipient mice were then injected intravenously with BM cells harvested from the femur and tibiae of naive donor mice. BM cells and blood samples of all recipient animals were analyzed by flow cytometry, to ensure that the grafts had been fully reconstituted and had not interfered with the process of hematopoiesis.Recipient KPC;Cx3cr1BMGFP/+, KPC;CCR2BM − / − and KPC;GDNFBM − /+ mice were then grown to the age of 3 and 6 months, and their tumors were analyzed to reveal the presence of macrophages and neural invasion in normal and neoplastic sites (that is, perineurial area, pancreas and spleen).

Characterization of the perineurial microenvironment

The perineurial niches were defined using four-micron sections stained with H&E. The perineurial niche was delineated ‘free-hand’ 500 μm from the nerve margin (perineurium) encircling the nerve 360°.

Statistics

All statistical analyses were performed using Origin statistical package (OriginLab Corporation, Northampton, MA, USA). Power calculations were performed for animal studies, and the number of mice reflects the number needed to have sufficient power (80%) to measure the expected difference (⩾ 20%) in the incidences of tumor formation at Po0.05. On the basis of our preliminary studies, the numbers of animals are included. Welch’s t-test was used for comparisons between two mouse groups of average tumor volumes and continuous variables with similar estimate variation. Means are presented with s.e.m. Fisher’s exact test was used to assess differences in the frequencies of categorical variables with a limited number of clinical samples. All statistical tests were two sided, and a P-value o0.05 was considered statistically significant. All experiments were repeated at least three times.

REFERENCES

1 Kelly K, Brader P, Rein A, Shah JP, Wong RJ, Fong Y et al. Attenuated multimutated herpes simplex virus-1 effectively treats prostate carcinomas with neural invasion while preserving nerve function. FASEB J 2008; 22: 1839–1848.
2 Ceyhan GO, Schafer KH, Kerscher AG, Rauch U, Demir IE, Kadihasanoglu M et al. Nerve growth factor and artemin are paracrine mediators of pancreatic neuropathy in pancreatic adenocarcinoma. Ann Surg 251: 923–931.
3 Gil Z, Carlson DL, Gupta A, Lee N, Hoppe B, Shah JP et al. Patterns and incidence of neural invasion in patients with cancers of the paranasal sinuses. Arch Otolaryngol Head Neck Surg 2009; 135: 173–179.
4 Liebl F, Demir IE, Mayer K, Schuster T, D’Haese JG, Becker K et al. The impact of neural invasion severity in gastrointestinal malignancies: a clinicopathological study. Ann Surg 2014; 260: 900–907.
5 Wang K, Demir IE, D’Haese JG, Tieftrunk E, Kujundzic K, Schorn S et al. The neurotrophic factor neurturin contributes toward an aggressive cancer cell phenotype, neuropathic pain and neuronal plasticity in pancreatic cancer. Carcinogenesis 2014; 35: 103–113.
6 Kim-Fuchs C, Le CP, Pimentel MA, Shackleford D, Ferrari D, Angst E et al. Chronic stress accelerates pancreatic cancer growth and invasion: a critical role for betaadrenergic signaling in the pancreatic microenvironment. Brain Behav Immun 2014; 40: 40–47.
7 Clark CE, Hingorani SR, Mick R, Combs C, Tuveson DA, Vonderheide RH. Dynamics of the immune reaction to pancreatic cancer from inception to invasion. Cancer Res 2007; 67: 9518–9527.
8 Zeng L, Guo Y, Liang J, Chen S, Peng P, Zhang Q et al. Perineural invasion and TAMs in pancreatic ductal adenocarcinomas: review of the original pathology reports using immunohistochemical enhancement and relationships with clinicopathological features. J Cancer 2014; 5: 754–760.
9 Cavel O, Shomron O, Shabtay A, Vital J, Trejo-Leider L, Weizman N et al. Endoneurial macrophages induce perineural invasion of pancreatic cancer cells by secretion of GDNF and activation of RET tyrosine kinase receptor. Cancer Res 2012; 72: 5733–5743.
10 Demir IE, Schorn S, Schremmer-Danninger E, Wang K, Kehl T, Giese NA et al. Perineural mast cells are specifically enriched in pancreatic neuritis and neuropathic pain in pancreatic cancer and chronic pancreatitis. PLoS One 2013; 8: e60529.
11 Yoshikawa K, Mitsunaga S, Kinoshita T, Konishi M, Takahashi S, Gotohda N et al. Impact of tumor-associated macrophages on invasive ductal carcinoma of the pancreas head. Cancer Sci 2012; 103: 2012–2020.
12 Jamieson NB, Mohamed M, Oien KA, Foulis AK, Dickson EJ, Imrie CW et al. The relationship between tumor inflammatory cell infiltrate and outcome in patients with pancreatic ductal adenocarcinoma. Ann Surg Oncol 2012; 19: 3581–3590.
13 Mitchem JB, Brennan DJ, Knolhoff BL, Belt BA, Zhu Y, Sanford DE et al. Targeting tumor-infiltrating macrophages decreases tumor-initiating cells, relieves immunosuppression, and improves chemotherapeutic responses. Cancer Res 2013; 73: 1128–1141.
14 Beatty GL, Winograd R, Evans RA, Long KB, Luque SL, Lee JW et al. Exclusion of T cells from pancreatic carcinomas in mice is regulated by Ly6C(low) F4/80(+) extratumoral macrophages. Gastroenterology 2015; 149: 201–210.
15 Kieseier BC, Hartung HP, Wiendl H. Immune circuitry in the peripheral nervous system. Curr Opin Neurol 2006; 19: 437–445.
16 Kelly DJ, Chanty A, Gow RM, Zhang Y, Gilbert RE. Protein kinase Cbeta inhibition attenuates osteopontin expression, macrophage recruitment, and tubulointerstitial injury in advanced experimental diabetic nephropathy. J Am Soc Nephrol 2005; 16: 1654–1660.
17 Veit C, Genze F, Menke A, Hoeffert S, Gress TM, Gierschik P et al. Activation of phosphatidylinositol 3-kinase and extracellular signal-regulated kinase is required for glial cell line-derived neurotrophic factor-induced migration and invasion of pancreatic carcinoma cells. Cancer Res 2004; 64: 5291–5300.
18 Sawai H, Okada Y, Kazanjian K, Kim J, Hasan S, Hines OJ et al. The G691S RET polymorphism increases glial cell line-derived neurotrophic factor-induced pancreatic cancer cell invasion by amplifying mitogen-activated protein kinase signaling. Cancer Res 2005; 65: 11536–11544.
19 Ben QW, Wang JC, Liu J, Zhu Y, Yuan F, Yao WY et al. Positive expression of L1CAM is associated with perineural invasion and poor outcome in pancreatic ductal adenocarcinoma. Ann Surg Oncol 2010; 17: 2213–2221.
20 Greten FR, Eckmann L, Greten TF, Park JM, Li ZW, Egan LJ et al. IKKbeta links inflammation and tumorigenesis in a mouse model of colitis-associated cancer. Cell 2004; 118: 285–296.
21 Qian B, Deng Y, Im JH, Muschel RJ, Zou Y, Li J et al. A distinct macrophage population mediates metastatic breast cancer cell extravasation, establishment and growth. PLoS One 2009; 4: e6562.
22 Ruffell B, Affara NI, Coussens LM. Differential macrophage programming in the tumor microenvironment. Trends Immunol 2012; 33: 119–126.
23 Mantovani A, Sozzani S, Locati M, Allavena P, Sica A. Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol 2002; 23: 549–555.
24 Zhang J, Lu Y, Pienta KJ. Multiple roles of chemokine (C-C motif) ligand 2 in promoting prostate cancer growth. J Natl Cancer Inst 2010; 102: 522–528.
25 Conway JG, McDonald B, Parham J, Keith B, Rusnak DW, Shaw E et al. Inhibition of colony-stimulating-factor-1 signaling in vivo with the orally bioavailable cFMS kinase inhibitor GW2580. Proc Natl Acad Sci USA 2005; 102: 16078–16083.
26 Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science 2009; 324: 1457–1461.
27 Yamaguchi R, Nagino M, Oda K, Kamiya J, Uesaka K, Nimura Y. Perineural invasion has a negative impact on survival of patients with gallbladder carcinoma. Br J Surg 2002; 89: 1130–1136.
28 Amit M, Na’ara S, Gil Z. Mechanisms of cancer dissemination along nerves. Nat Rev Cancer 2016; 16: 399–408.
29 Demir IE, Ceyhan GO, Liebl F, D’Haese JG, Maak M, Friess H. Neural invasion in pancreatic cancer: the past, present and future. Cancers 2010; 2: 1513–1527.
30 Ceyhan GO, Bergmann F, Kadihasanoglu M, Altintas B, Demir IE, Hinz U et al. Pancreatic neuropathy and neuropathic pain–a comprehensive pathomorphological study of 546 cases. Gastroenterology 2009; 136: 177–86 e1.
31 Gil Z, Cavel O, Kelly K, Brader P, Rein A, Gao SP et al. Paracrine regulation of pancreatic cancer cell invasion by peripheral nerves. J Natl Cancer Inst 2010; 102: 107–118.
32 Demir IE, Ceyhan GO, Rauch U, Altintas B, Klotz M, Muller MW et al. The microenvironment in chronic pancreatitis and pancreatic cancer induces neuronal plasticity. Neurogastroenterol Motil 2010; 22: 480–490; e112–113.
33 Demir IE, Friess H, Ceyhan GO. Neural plasticity in pancreatitis and pancreatic cancer. Nat Rev Gastroenterol Hepatol 2015; 12: 649–659.
34 Esseghir S, Todd SK, Hunt T, Poulsom R, Plaza-Menacho I, Reis-Filho JS et al. A role for glial cell derived neurotrophic factor induced expression by inflammatory cytokines and RET/GFR alpha 1 receptor up-regulation in breast cancer. CancerRes 2007; 67: 11732–11741.
35 Qiao S, Iwashita T, Ichihara M, Murakumo Y, Yamaguchi A, Isogai M et al. Increased expression of glial cell line-derived neurotrophic factor and neurturin in a case of colon adenocarcinoma associated with diffuse ganglioneuromatosis. Clin Neuropathol 2009; 28: 105–112.
36 Edstrom E, Frisk T, Farnebo F, Hoog A, Backdahl M, Larsson C. Expression analysis of RET and the GDNF/GFRalpha-1 and NTN/GFRalpha-2 ligand complexes in pheochromocytomas and paragangliomas. Int J Mol Med 2000; 6: 469–474.
37 Frisk T, Farnebo F, Zedenius J, Grimelius L, Hoog A, Wallin G et al. Expression of RET and its ligand complexes, GDNF/GFRalpha-1 and NTN/GFRalpha-2, in medullary thyroid carcinomas. Eur J Endocrinol 2000; 142: 643–649.
38 Kang J, Perry JK, Pandey V, Fielder GC, Mei B, Qian PX et al. Artemin is oncogenic for human mammary carcinoma cells. Oncogene 2009; 28: 2034–2045.
39 Tang MJ, Worley D, Sanicola M, Dressler GR. The RET-glial cell-derived neurotrophic factor (GDNF) pathway stimulates migration and chemoattraction of epithelial cells. J Cell Biol 1998; 142: 1337–1345.
40 Fukuda T, Kiuchi K, Takahashi M. Novel mechanism of regulation of Rac activity and lamellipodia formation by RET tyrosine kinase. J Biol Chem 2002; 277: 19114–19121.
41 Stahle M, Veit C, Bachfischer U, Schierling K, Skripczynski B, Hall A et al. Mechanisms in LPA-induced tumor cell migration: critical role of phosphorylated ERK. J Cell Sci 2003; 116(Pt 18): 3835–3846.
42 Giehl K, Skripczynski B, Mansard A, Menke A, Gierschik P. Growth factordependent activation of the Ras-Raf-MEK-MAPK pathway in the human pancreatic carcinoma cell line PANC-1 carrying activated K-ras: implications for cell proliferation and cell migration. Oncogene 2000; 19: 2930–2942.
43 Torres MP, Rachagani S, Souchek JJ, Mallya K, Johansson SL, Batra SK. Novel pancreatic cancer cell lines derived from genetically engineered mouse models of spontaneous pancreatic adenocarcinoma: applications in diagnosis and therapy. PLoS One 2013; 8: e80580.
44 Patwardhan PP, Surriga O, Beckman MJ, de Stanchina E, Dematteo RP, Tap WD et al. Sustained inhibition of receptor tyrosine kinases and macrophage depletion by PLX3397 and rapamycin as a potential new approach for the treatment of MPNSTs. Clin Cancer Res 2014; 20: 3146–3158.
45 Liebl F, Demir IE, Rosenberg R, Boldis A, Yildiz E, Kujundzic K et al. The severity of neural invasion is associated with shortened survival in colon cancer. Clin Cancer Res 2013; 19: 50–61.
46 Demir IE, Tieftrunk E, Schorn S, Saricaoglu OC, Pfitzinger PL, Teller S et al. Activated Schwann cells in pancreatic cancer are linked to analgesia via suppression of spinal astroglia and microglia. Gut 2016; 65: 1001–1014.
47 Amit M, Laider-Trejo L, Shalom V, Shabtay-Orbach A, Krelin Y, Gil Z. Characterization of OSI-930 the melanoma brain metastatic niche in mice and humans. Cancer Med 2013; 2: 155–163.
48 He S, He S, Chen CH, Deborde S, Bakst RL, Chernichenko N et al. The chemokine (CCL2-CCR2) signaling axis mediates perineural invasion. Mol Cancer Res 2015; 13: 380–390.
49 Pichel JG, Shen L, Sheng HZ, Granholm AC, Drago J, Grinberg A et al. Defects in enteric innervation and kidney development in mice lacking GDNF. Nature 1996; 382: 73–76.
50 Cullen-McEwen LA, Drago J, Bertram JF. Nephron endowment in glial cell linederived neurotrophic factor (GDNF) heterozygous mice. Kidney Int 2001; 60: 31–36.
51 Luster AD. Chemokines—chemotactic cytokines that mediate inflammation. N Engl J Med 1998; 338: 436–445.
52 Jung S, Aliberti J, Graemmel P, Sunshine MJ, Kreutzberg GW, Sher A et al. Analysis of fractalkine receptor CX(3)CR1 function by targeted deletion and green fluorescent protein reporter gene insertion. Mol Cell Biol 2000; 20: 4106–4114.
53 Pahler JC, Tazzyman S, Erez N, Chen YY, Murdoch C, Nozawa H et al. Plasticity in tumor-promoting inflammation: impairment of macrophage recruitment evokes a compensatory neutrophil response. Neoplasia 2008; 10: 329–340.