Neuron–tumor communication as a target for neuromodulatory drug delivery in cancer

The presence of nerves within tumors was largely unexplored in the early 19 th and 20 th century, with contemporary pathology texts generally considering tumors to be devoid of nerves. However, observations of tumor-associated pain and intratumoral nerve fibers challenged this view, prompting questions about their role in neoplastic tissue. As early as 1897, Young Y. reported nerves in various non-neural tumors, including breast and cervical carcinomas, sarcomas, myxomas, and metastatic carcinomas, providing early histological evidence of neural elements within tumors (1). This seminal study raised an important question of whether intratumoral nerves represented pre-existing neural structures incorporated during tumor growth or reflected active neural changes associated with neoplastic development.To address this question, investigations in the late 1920s and early 1930s began to explore the origin and biological significance of these intratumoral nerves. In 1928, Herzog E. investigated the presence of nerve fibers in 100 neoplasms and proposed that intratumoral nerves represented pre-existing nerves incorporated into infiltrating neoplastic tissue. His observations suggested that the distribution of nerves was related to the pattern of tumor invasion, supporting the view that their presence could result from the incorporation of normal neural structures during tumor growth (2). In 1929, Oertel H. proposed a different interpretation, proposing that human carcinomas possessed an intrinsic neural organization, with nerve fibers closely associated with the tumor stroma, blood vessels, and parenchymal elements. This raised the possibility that the relationship between nerves and tumors might extend beyond passive entrapment (3). In 1930, Martynow W. further challenged the concept of simple nerve incorporation by suggesting that nerve fibers within carcinomas could be newly formed. He also described degenerative changes in these fibers, indicating that intratumoral nerves might undergo active remodeling in response to the tumor microenvironment (4).In 1931, Oertel H., Nye H., and Thomlinson B. provided additional observations on nerve fibers within human tumors, further supporting the presence of an organized neural component within neoplastic tissue (5). In 1933, Ryrie synthesized these observations in his original investigation of the anatomical and physiological relationship between nerve fibers and neoplasms. He considered several possibilities, including the incorporation of pre-existing nerves into tumors, the formation of new nerve fibers within neoplastic tissue, and degenerative and regenerative changes resembling those observed after peripheral nerve injury. These observations raised the possibility of an active interaction between nerves and malignant tissue. Nevertheless, Ryrie G. concluded that the available evidence did not justify assuming a trophic influence of nerves on malignant cells (6).Two studies published in 1982 and 1985 investigating neuroendocrine biomarkers also provided early biochemical evidence linking neural biology and cancer, demonstrating significantly elevated serum neuron-specific enolase (NSE) levels in patients with small-cell lung cancer (SCLC) (7,8). In 1982, Carney et al. first reported that serum NSE was elevated in 69% of newly diagnosed patients with SCLC (n = 94) compared with controls, with a higher prevalence in extensive-stage disease (87%) than limited-stage disease (39%). Mean NSE levels were significantly higher in extensive-stage disease (59 ng/ml vs 13.8 ng/ml), increased further with metastatic burden (84% of patients with 1-2 sites and 100% of those with ≥3 sites), and showed a strong correlation with chemotherapy response. SCLC cell lines derived from patients also consistently expressed high levels of NSE (7). In a comparable study, Akoun et al. found that serum NSE levels were elevated in 65.1% of patients diagnosed with SCLC (n = 43), with significantly higher mean levels in extensive-stage disease (50 ng/ml) than in limited-stage disease (16 ng/ml), while levels were much lower in non-SCLC (11 ng/ml) and noncancerous lung disease (7 ng/ml). NSE closely correlated with tumor burden and showed dynamic changes during chemotherapy, with higher early levels (>100 ng/ml) associated with better treatment response (8). This mechanism was subsequently revealed in a 2022 study (Figure 1), which demonstrated that NSE interacts with and suppresses neuroblastoma suppressor of tumorigenicity 1 (NBL1), thereby relieving NBL1-mediated inhibition of BMP2 signaling. This interaction enhances BMP2-BMPR1A signaling and activates the BMP2/Smad/ID1 pathway, ultimately promoting malignant characteristics of SCLC (9). Keeping track of these conceptual advances, a growing body of experimental evidence (detailed in section 2) have provided evidence of functional neuron-tumor interactions across multiple cancer types. Studies in human specimens have identified such interactions in squamous cell carcinoma (10), breast cancer (11), gastric cancer (12), pancreatic ductal adenocarcinoma (13), lung adenocarcinoma (14), prostate cancer (15), and diffuse midline glioma (16,17). SCLC has been investigated using human specimens, ex vivo electrophysiological and co-culture models (18), and in vivo models (19), while glioblastoma (20,21), glioma (22), diffuse midline glioma (16), gastric cancer (23), and breast-to-brain metastasis (24), have been studied primarily using in vitro and in vivo models. Evidence from these studies shows that neural signaling influences tumor growth, invasion, metastasis, and proliferation, while inhibition through pharmacological or genetic targeting of CHRM3, AMPARs, NMDARs/GRIN2B, GRIN2D, GABA receptors, NKCC1, CGRP, β-adrenergic signaling, and sensory or sympathetic neuron ablation has demonstrated therapeutic potential (12,18,22,23,25).Several recent reviews have synthesized the rapidly expanding field of cancer neuroscience, encompassing