Purpose: Neural invasion (NI) is a histopathologic feature of colon cancer that receives little consideration. Therefore, we conducted a morphologic and functional characterization of NI in colon cancer.Experimental Design: NI was investigated in 673 patients with colon cancer. Localization and severity of NI was determined and related to patient's prognosis and survival. The neuro-affinity of colon cancer cells (HT29, HCT-116, SW620, and DLD-1) was compared with pancreatic cancer (T3M4 and SU86.86) and rectal cancer cells (CMT-93) in the in vitro three-dimensional (3D)-neural-migration assay and analyzed via live-cell imaging. Immunoreactivity of the neuroplasticity marker GAP-43, and the neurotrophic-chemoattractant factors Artemin and nerve growth factor (NGF), was quantified in colon cancer and pancreatic cancer nerves. Dorsal root ganglia of newborn rats were exposed to supernatants of colon cancer, rectal cancer, and pancreatic cancer cells and neurite density was determined.Results: NI was detected in 210 of 673 patients (31.2%). Although increasing NI severity scores were associated with a significantly poorer survival, presence of NI was not an independent prognostic factor in colon cancer. In the 3D migration assay, colon cancer and rectal cancer cells showed much less neuritetargeted migration when compared with pancreatic cancer cells. Supernatants of pancreatic cancer and rectal cancer cells induced a much higher neurite density than those of colon cancer cells. Accordingly, NGF, Artemin, and GAP-43 were much more pronounced in nerves in pancreatic cancer than in colon cancer.Conclusion: NI is not an independent prognostic factor in colon cancer. The lack of a considerable biologic affinity between colon cancer cells and neurons, the low expression profile of colonic nerves for chemoattractant molecules, and the absence of a major neuroplasticity in colon cancer may explain the low prevalence and impact of NI in colon cancer.

Pancreatic ductal adenocarcinoma (PDAC) cells (PCC) have an exceptional propensity to metastasize early into intratumoral, chemokine-secreting nerves. However, we hypothesized the opposite process, that precancerous pancreatic cells secrete chemokines that chemoattract Schwann cells (SC) of nerves and thus induce ready-to-use routes of dissemination in early carcinogenesis. Here we show a peculiar role for the chemokine CXCL12 secreted in early PDAC and for its receptors CXCR4/CXCR7 on SC in the initiation of neural invasion in the cancer precursor stage and the resulting delay in the onset of PDAC-associated pain. SC exhibited cancer-or hypoxia-induced CXCR4/CXCR7 expression in vivo and in vitro and migrated toward CXCL12-expressing PCC. Glia-specific depletion of CXCR4/CXCR7 in mice abrogated the chemoattraction of SC to PCC. PDAC mice with pancreas-specific CXCL12 depletion exhibited diminished SC chemoattraction to pancreatic intraepithelial neoplasia and increased abdominal hypersensitivity caused by augmented spinal astroglial and microglial activity. In PDAC patients, reduced CXCR4/CXCR7 expression in nerves correlated with increased pain. Mechanistically, upon CXCL12 exposure, SC down-regulated the expression of several pain-associated targets. Therefore, PDAC-derived CXCL12 seems to induce tumor infiltration by SC during early carcinogenesis and to attenuate pain, possibly resulting in delayed diagnosis in PDAC.Schwann cells | pancreatic cancer | CXCL12 | CXCR4 | CXCR7

PCa is characterized by intrapancreatic neuroplasticity and NI. Here, we show that PCC produce the neurotrophic factor NRTN, which reinforces their biological properties, triggers neuroplastic alterations, NI and influences pain sensation via the GFRα-2 receptor.

No abstract

No abstract

<p>PDF file - 62K, Supplementary Table with patient characteristics</p>

<p>PDF file - 62K, Supplementary Table with patient characteristics</p>

<div>Abstract<p><b>Purpose:</b> Neural invasion (NI) is a histopathologic feature of colon cancer that receives little consideration. Therefore, we conducted a morphologic and functional characterization of NI in colon cancer.</p><p><b>Experimental Design:</b> NI was investigated in 673 patients with colon cancer. Localization and severity of NI was determined and related to patient's prognosis and survival. The neuro-affinity of colon cancer cells (HT29, HCT-116, SW620, and DLD-1) was compared with pancreatic cancer (T3M4 and SU86.86) and rectal cancer cells (CMT-93) in the <i>in vitro</i> three-dimensional (3D)–neural-migration assay and analyzed via live-cell imaging. Immunoreactivity of the neuroplasticity marker GAP-43, and the neurotrophic-chemoattractant factors Artemin and nerve growth factor (NGF), was quantified in colon cancer and pancreatic cancer nerves. Dorsal root ganglia of newborn rats were exposed to supernatants of colon cancer, rectal cancer, and pancreatic cancer cells and neurite density was determined.</p><p><b>Results:</b> NI was detected in 210 of 673 patients (31.2%). Although increasing NI <i>severity</i> scores were associated with a significantly poorer survival, <i>presence</i> of NI was not an independent prognostic factor in colon cancer. In the 3D migration assay, colon cancer and rectal cancer cells showed much less neurite-targeted migration when compared with pancreatic cancer cells. Supernatants of pancreatic cancer and rectal cancer cells induced a much higher neurite density than those of colon cancer cells. Accordingly, NGF, Artemin, and GAP-43 were much more pronounced in nerves in pancreatic cancer than in colon cancer.</p><p><b>Conclusion:</b> NI is not an independent prognostic factor in colon cancer. The lack of a considerable biologic affinity between colon cancer cells and neurons, the low expression profile of colonic nerves for chemoattractant molecules, and the absence of a major neuroplasticity in colon cancer may explain the low prevalence and impact of NI in colon cancer. <i>Clin Cancer Res; 19(1); 50–61. ©2012 AACR</i>.</p></div>