Introduction

Depression is a prevalent affective disorder, characterized by persistent low mood, slowed thinking, reduced volitional activity, cognitive impairments, and somatic symptoms1,2,3,4. It has been ranked as the third cause of global burden of disease by the World Health Organization5. Despite extensive research, the precise etiology of depression remains unclear, it is believed to involve a complex interplay of psychological, social, and biological factors6. Most studies exploring the mechanisms of depression have relied on chronic external stress models in rodents7. However, it has long been observed that somatic diseases are often accompanied by affective, cognitive, and motivational changes. Individuals with chronic physical conditions have a two- to three-fold higher risk of developing depression compared to the general population8. At present, the comorbidity mechanism of somatic diseases and depression is poorly understood in the field of neuroscience. The existing studies in this field mainly focus on the mechanism of chronic pain and its accompanying chronic inflammation promoting depression9. Psychiatric symptoms, including depression, anxiety, substance abuse, physical disorders, and stress reactions, are common in cancer patients10, with up to 20–30% of cancer patients affected by depression11. The mechanisms underlying the comorbidity of depression in cancer patients warrant further investigation. Colorectal cancer (CRC) is the third most prevalent cancer globally, with an estimated two million new cases and nearly one million deaths annually. The overall 5-year survival rate in patients with CRC is 64%12. Clinically significant depressive and anxiety symptoms are common among CRC patients13, with rates of depression and anxiety ~10% higher than those in physically healthy individuals14. Research suggests that depression and anxiety may contribute to increased mortality risk in CRC patients. For instance, a one standard deviation increases in anxiety or depressive symptoms in one sample corresponded to a 16% higher mortality risk15,16. Numerous studies have demonstrated that the gut and brain communicate in a bidirectional manner, known as the gut-brain axis17. Recent research has highlighted the relationship between depression and gut microbiota, as well as irritable bowel syndrome18,19. Importantly, tumor-induced dysbiosis can disrupt immune homeostasis and alter central nervous system signaling via microbial metabolites and immune mediators. For instance, intestinal tumors may modulate cytokine production and peripheral inflammation, which can cross the blood-brain barrier and affect mood-regulating brain regions20,21,22,23. Despite this, crosstalk between intestinal tumors and the brain remains poorly understood. Previous studies have shown that chronic stress not only induces depression by altering the activity of brain regions critical for emotional regulation but also promotes tumor progression24,25. Thus, in recent years, there have been examples of central nervous system modulation of tumors. Activation of the VTA enhances both innate and adaptive immunity, reduces sympathetic release of norepinephrine from the bone marrow, and promotes T cell activation, all of which contribute to reduced tumor growth26. It will be interesting to investigate whether emotional brain regions can influence colorectal cancer comorbid depression. In this study, we established model of mice with orthotopic colorectal cancer (CRC mice) and observed depression-like behaviors in the CRC mice. Through whole-brain c-FOS mapping, functional connectivity analysis, gut-brain inverse mapping, and correlation analysis, combined with chemogenetic manipulation, we identified the anterior cingulate cortex (ACC) as a key hub in the depression-related brain network. To investigate the underlying mechanisms, we performed whole transcriptome sequencing of key neural nuclei, and identified molecular characteristic changes in comorbidity of CRC and depression. Furthermore, chemogenetic modulation of neuronal excitability not only alleviated depression-like behavior but also reduced CRC progression in these mice. Importantly, our findings suggest that targeting the ACC, in addition to conventional colorectal cancer therapies, may offer therapeutic and quality-of-life benefits for patients with both CRC and depression.

Results

CRC mice exhibited depression-like behaviors

Clinical studies have reported a high prevalence of comorbid depression in patients with CRC13. To investigate the mechanisms underlying the comorbidity of depression and CRC, we first examined whether CRC mice exhibited depression-like behaviors. Towards that end, MC38 cells, was originally derived in 1975 from a colon adenocarcinoma induced by subcutaneous injection of dimethylhydrazine in a female C57BL/6 mouse27, were mixed 1:1 with Matrigel and implanted into the rectum of mice (Fig. 1A). Sham-operated control mice were implanted with Matrigel only. Hematoxylin and eosin (HE) staining showed the formation of dense and clumped tumor nests, loss of glandular architectures in the rectum tissue sections collected from mice implanted with MC38 cells (Fig. 1B). The average colorectal tumor size was 83 mm3, 230.4 mm3, and 443.4 mm3 at 1, 2, and 3 weeks after implantation, respectively. (Fig. 1C, D). Concurrently, the mice bearing colorectal tumor exhibited depression-like behaviors, which manifested as decreased pain threshold in the von Frey test, increased the latency to eating in the novelty-suppressed feeding test, decreased grooming duration in the splash test, increased immobility in the tail suspension test and impaired social preference (Fig. 1E–N).

Fig. 1: CRC mice exhibited depressive-like behavior.
Fig. 1: CRC mice exhibited depressive-like behavior.
Full size image

A The experimental timeline for establishment of CRC mice, behavioral tests, and tissue collection (Sac). B Representative images of hematoxylin-eosin (HE) staining of colorectum tissues, comparing sham-operated (Sham) mice and mice implanted with MC38 cells in the colorectum (CRC), Scale bars: 100 µm (n  =  7). C Representative images of tumor size at 1, 2, and 3 weeks after MC38 implantation. D Quantification of tumor volume at 1, 2, and 3 weeks (n  =  8,7,8 per group). E Changes in mechanical pain threshold measured by the von Frey test (VFT) (n  = 15 per group). F, G Schematic diagram and quantification of the latency to eat in the novelty-suppressed feeding test (NSF) (n  = 15 per group). H, I Schematic diagram and quantification of grooming duration in the splash test (ST) (n  = 15 per group). J, K Schematic diagram and quantification of the immobile time in tail suspension test (TS) (n  = 15,14 per group). L–N Schematic diagram, trajectory heat map and quantification of the total exploration time and differentiation index (DI) in the social interaction (SI) (n  =  15 per group). All data are presented as mean ± s.e.m. and analyzed by unpaired Student’s t tests (E, I, K, N) or Mann–Whitney tests (G). The source data are provided as Supplementary Data 1 in the Figshare repository.

The activity of the central nervous system is altered in CRC mice

The neural circuit mechanisms underlying depression can vary depending on its cause, and appropriate interventions must be tailored accordingly. However, the neural mechanism of depression in CRC are unclear. Based on our findings that mice exhibited depressive-like phenotypes at 2 weeks after tumor implantation. c-FOS staining was performed in 27 brain regions (Supplementary Table 1) related to digestive system, pain and emotion by immunofluorescence to screen the brain regions with changes in central nervous system nuclear activity during the progression of colorectal cancer (Fig. 2A). Our results revealed that, compared with the sham-operated mice, the activity of brain regions did not change significantly 7 days after tumor implantation (Supplementary Fig. 1A). However, 14 days after tumor implantation, the active brain regions in CRC mice included infralimbic cortex (IL), ACC, nucleus accumbens (NAc), insular cortex (IC), retrosplenial granular cortex (RSG), medial hypothalamic nucleus (MH), cornu ammonis 1 (CA1), dentate gyrus (DG) (Fig. 2B, Supplementary Fig. 1B, C). Only ACC, IL and RSG were still significantly increased at day 21 (Fig. 2C). Importantly, the ACC and IL brain regions showed a time-dependent increase in activity (Fig. 2D–G), but not RSG brain regions (Supplementary Fig. 1C, D). To exclude increased activity in brain regions caused by foreign body sensation, silicone was injected into the colon of mice, and no significant increase was found in the brain regions with increased activity on day 14 (Supplementary Fig. 1E–H).

Fig. 2: Whole-brain and region-specific dynamics of c-FOS expression during CRC progression.
Fig. 2: Whole-brain and region-specific dynamics of c-FOS expression during CRC progression.
Full size image

A The coronal atlas, with the regions of interest highlighted in blue. See Supplementary Table 1 for the full names of brain region. B Quantification of c-FOS positive neurons at two weeks after establishment of CRC mice, comparing Sham mice and CRC mice (n  =  10 to 17 per group, source data provide the number of animals specific to each group). C Quantification of c-FOS positive neurons at three weeks after establishment of CRC mice, comparing Sham mice and CRC mice (n  =  5 to 8 per group, source data provide the number of animals specific to each group). D Representative immunofluorescence images of c-FOS (Green) expression in the anterior cingulate cortex (ACC) at 1, 2, and 3 weeks after MC38 implantation and Sham. E Quantification of c-FOS–positive cells in the ACC over time (n  =  22,8,11,8 per group). F Representative immunofluorescence images of c-FOS (Green) expression in the infralimbic cortex (IL) at 1, 2, and 3 weeks after MC38 implantation and Sham. G Quantification of c-FOS–positive cells in the IL over time (n  =  21,7,15,7 per group). Scale bars: 100 μm. Sham: sham-operated mice, CRC: mice implanted with MC38 cells in the colorectum. Data are presented as mean ± s.e.m. and analyzed by t test with Welch’s correction (B, C). The source data are provided as Supplementary Data 1 in the Figshare repository.

Characteristics of brain functional network in the progression of colorectal cancer

To identify the key brain regions and their functional connectivity associated with depression-like behavior in colorectal cancer mice, we investigated the functional communication between the brain regions of sham-operated and CRC mice on day 14(a representative phase of CRC progression with established depressive-like behaviors but minimal cachexia). Previous studies have suggested that functional communication between brain regions can be evaluated through correlation analysis of c-FOS expression across multiple regions28. Pearson correlation coefficient analysis was performed to evaluate the relationship of c-FOS expression in each brain region for both groups, and the results were presented as heat maps (Fig. 3A, D). Based on this, we identified brain regions with strong correlations (Pearson’s r > 0.6, p < 0.05) in each group, and organized their functional connectivity into neural network maps (Fig. 3B, E). Each brain region was represented as a node in the network, and functional connectivity between regions was shown as lines, with red lines indicating positive correlations and line thickness corresponding to the strength of the correlation. The results indicated that the correlation between c-FOS expression in multiple brain regions was weakened in the CRC mice compared to the sham-operated mice group, suggesting a global reduction in functional connectivity across brain regions (Fig. 3B, E). In this type of neural network, core nodes exhibit greater connectivity and have a more significant impact on overall network function29. We used the node deletion to simulate the effect of changes in core brain areas on global network efficiency (GE)30. By comparing the changes in global efficiency (ΔGE) before and after removing core nodes, we identified the five brain regions with the most significant effects on GE in each group. The results revealed different core brain regions for the two groups. The results showed that the core brain regions of the two groups were different. In the sham-operated group, the top ranked were solitary nucleus (NTS), paraventricular nucleus of hypothalamus (PVN), paraventricular thalamic nucleus (PVT), RSG, IC. In CRC mice, the top ranked were dorsal raphe (DR), NAC, IC, cortical amygdaloid nucleus (COA), prelimbic area (PL) (Fig. 3C, F). To more directly investigate the upstream regulatory centers in CRC mice, we injected a transsynaptic retrograde tracer virus (pseudorabies virus, PRV) into the colorectum (Fig. 3G). The results showed that PRV-labeled upstream centers of the colon and rectum included ACC, PL, lateral septum (LS), medial septum (MS), IC, bed nucleus of stria terminalis (BNST), medial preoptic area (MPO), paraventricular nucleus of hypothalamus (PVN), supplemental somatosensory (SS), MH, lateral hypothalamus (LH), ventral tegmental area (VTA), periaqueductal grey (PAG), DR, dorsal motor nucleus of vagus (DMV), NTS and area postrema (AP) (Fig. 3H). These brain regions could also be traced by injecting trans-synaptic retrograde tracking virus (PRV) into the tumor of CRC mice (Fig. 3I, J). Notably, the activity of ACC, IC and MH brain regions was abnormally increased in CRC mice, and PL and DR Brain regions played key roles in the global efficiency of the network in CRC mice. A background review of these brain regions revealed that the anterior cingulate cortex (ACC) is the most dorsal subregion of the medial prefrontal cortex and critical for the affective component of pain perception and pain-related mood disorders31,32. Previous studies have highlighted the involvement of regions associated with emotional processing, such as the thalamus, amygdala, ACC, and anterior insula, in the emotional aspects of pain and other gastrointestinal symptoms, including anxiety and depression33,34,35.

Fig. 3: Characteristics of changes in the activity of brain regions and their functional network of CRC mice and the reverse tracing of colorectum.
Fig. 3: Characteristics of changes in the activity of brain regions and their functional network of CRC mice and the reverse tracing of colorectum.
Full size image

A, D The inter-regional correlations of c-FOS expression between brain regions. Color scale represents Pearson’s r value. B, E Network graphs derived from inter-regional correlation analysis of c-FOS, with nodes representing each brain region, and the edges representing significant correlations (the red line represents the positive correlation, the thickness represents the high correlation). C, F The top 5 brain regions that contribute to network global efficiency. ΔGE was calculated as changes in the global efficiency of the networks after isolating each node one-by-one. A–C: The Ctrl group. D–F The CRC group. G Schematic of trans-synaptic reverse tracer virus PRV injection into the colorectum and tissue collection (Sac). H Representative of brain regions labeled by PRV. I Schematic illustration of the experimental design: two weeks after orthotopic implantation of MC38 cells, PRV was injected into the colorectum, and tissue collection (Sac) (J) Representative images of PRV-labeled brain regions following trans-synaptic retrograde tracing from the Colorectal tumor. Scale bars: 200 μm. See Supplementary Table 1 for the full names of brain region. Sham: sham-operated mice, CRC: mice implanted with MC38 cells in the colorectum.

A critical involvement of ACC in comorbid depression in CRC mice

We then examined the correlations of c-FOS+ cell density in areas with altered activity and tumor size with depression-like behaviors in CRC mice. Out of the eight brain regions examined, only c-FOS expression in the ACC showed a significant and positive correlation with most of the scores of depressive-like behaviors (Fig. 4A–D; Supplementary Fig. 2). Tumor size was positively correlated with several depression-like behavioral measures, including reduced grooming activity, increased immobility in the tail suspension test, and decreased social interaction. Although the correlation with latency in the NSF test did not reach statistical significance, it exhibited a similar trend (Fig. 4E–H). To further investigate the molecular mechanisms underlying the abnormal functional activity in the ACC during colorectal cancer progression, we performed RNA sequencing (RNA-seq) analysis of the ACC in CRC mice (Fig. 4I). Differentially expressed gene (DEG) analysis was carried out using the DESeq2 package in R, with criteria set to p < 0.05 and |log2 fold change| > 0.585. This analysis identified 798 DEGs, of which 660 were up-regulated and 138 were down-regulated in the CRC group. The heatmap of the top 50 differential genes showed the greatest changes in gene expression (Fig. 4J). Notably, the volcano map showed that the most significant up-regulated genes in the CRC group were Col6a1, AI480526 and Neat1 while the down-regulated genes included Dusp4 and Rasd1 (Fig. 4K). Previous studies have suggested that AI480526 has been proposed as a potential genetic and epigenetic biomarker in microbiologically associated depression involved in hippocampal neurodevelopmental dysfunction in mice with dysbiosis36. These findings provide insight into the potential mechanisms underlying comorbid depression in intestinal diseases, including CRC. Additionally, KEGG (Kyoto Encyclopedia of Genes and Genomes) and GO (Gene Ontology) analyses revealed that the most significantly enriched gene sets in the ACC of CRC mice were related to extracellular matrix organization, plasma membrane organization, and secretory granules, etc. (Fig. 4L). Conversely, pathways related to developmental growth, secretion, and hormone response were significantly de-enriched (Fig. 4L). Notably, de-enrichment in secretion and hormone response pathways may impact neurotransmitter release37,38, potentially contributing to the abnormal activation of ACC regions and the development of depression in CRC mice.

Fig. 4: RNA sequencing revealed transcriptomic alterations in ACC of CRC mice.
Fig. 4: RNA sequencing revealed transcriptomic alterations in ACC of CRC mice.
Full size image

A–D Correlation of the density of c-FOS positive cell in anterior cingulate cortex (ACC) brain regions with the depression-like behavior score (ST: splash test, NSF: novelty-suppressed feeding test, TS: tail suspension test, SI: social interaction) (n  =  14 per group). E–H Correlation of the tumor size with the depression-like behavior scores (ST: splash test, NSF: novelty-suppressed feeding test, TS: tail suspension test, SI: social interaction) (n = 11 per group). I Schematics for the establishment of CRC mice. ACC tissue collection (Sac), RNAseq and data analyses. In situ implantation of MC38 cells in the mouse colon, and samples were collected two weeks later. J Heatmap showing relative expression levels of differentially expressed genes (DEGs, p < 0.05 by Deseq2) in the ACC, comparing the CRC group to the Sham group. K Volcano plot showing gene expression changes. Red and blue dots represent upregulated and downregulated DEGs, respectively. L Top the Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway and Gene ontology (GO) gene sets. Sham: sham-operated mice, CRC: mice implanted with MC38 cells in the colorectum. The source data are provided as Supplementary Data 1 in the Figshare repository.

Inhibition of ACC activity ameliorates depression-like behaviors in CRC mice

To investigate whether ACC hyperactivity contributes to depression-like behavior in CRC mice, we injected AAVs expressing hM4Di-EGFP into the ACC to inhibit ACC neuronal activity, while control mice received AAVs expressing EGFP alone (Fig. 5B). Three weeks after AAV infusion, CRC mice were established and the chemogenetic ligand C21 was injected intraperitoneally for 14 consecutive days to inhibit ACC neuronal excitability. Sensitivity to mechanical pain and depression-like behaviors were assessed at the specified time points shown in Fig. 5A. Results showed that inhibition of ACC neuronal activity significantly reduced depression-like behaviors in CRC mice, which manifested as increased pain threshold in the von Frey test (Fig. 5C), decreased the latency to eating in the novelty-suppressed feeding test(Fig. 5D), increased grooming duration in the splash test (Fig. 5E), decreased immobility in the tail suspension test (Fig. 5F) and recovery of impaired social preference (Fig. 5G). Conversely, CRC mice injected with AAVs expressing hM3Dq-EGFP into the ACC exhibited exacerbated depression-like behaviors following chemogenetic activation, while mechanical pain sensitivity remained unaffected. (Fig. 5B–G). To rule out potential behavioral effects of C21, CRC mice not expressing DREADDs were subjected to a single intraperitoneal injection of C21, followed by behavioral testing. The results indicated that C21 alone does not affect depressive-like behaviors (Supplementary Fig. 3A–F). Furthermore, to validate the behavioral effects observed using chemogenetic inhibition, we employed an independent approach by expressing tetanus toxin light chain (Tettoxlc) to block synaptic transmission from ACC neurons. Consistent with the results from hM4Di-mediated chemogenetic inhibition, CRC mice expressing tettoclc-EGFP showed remission of depression-like behavior (Supplementary Fig. 3G–M). To explore the molecular mechanisms underlying the improvement in depression-like behavior following ACC inhibition, we performed RNA-seq analysis on the ACC regions of these mice (Fig. 5H). DEG analysis was conducted using the DESeq2 package in R, with filtering criteria set to p < 0.05 and |log2 fold change| > 0.585. A total of 304 DEGs were identified, of which 251 genes were up-regulated and 53 genes were down-regulated after ACC inhibition. The heatmap of the top 50 differential genes showed the greatest changes in gene expression (Fig. 5I). Notably, genes such as Sgk1, Plekhf1, Plin4, and Nlrp6 were significantly up-regulated, while Btg2 was among the most down-regulated genes (Fig. 5J). Among them, Sgk1, Plekhf1 and Nlrp6 genes have been reported to be associated with antidepressive-like behaviors39,40,41. Additionally, we examined the genes that were up-regulated in the ACC of CRC mice compared to sham-operated controls and down-regulated after ACC regulation (Fig. 5K). Among these, Rspo4, Cd70, Igkv6-20, and Fhl4 are related to immune function. The interplay between gut disease, immune response, and central nervous system dysfunction is well-documented42,43,44. Hlx and Alox12 are involved in nervous system function, with Hlx specifically playing a role in enteric nervous system development45,46. These findings raise the question: how are enteric nerves affected in CRC mice? Are alterations in the enteric nervous system contributing to the depression-like behaviors observed in CRC?

Fig. 5: Effect of inhibition of ACC activity on depressive-like behavior and transcriptome.
Fig. 5: Effect of inhibition of ACC activity on depressive-like behavior and transcriptome.
Full size image

A The experimental timeline for the AAV injection into the ACC, establishment of CRC mice, the C21 injection, behavioral tests and tissue collection (Sac). In brief, first, the virus is injected into the bilateral ACC brain regions, three weeks later, MC38 cells are implanted in situ in the colon of the mice, then, C21 is administered intraperitoneally for two consecutive weeks, after behavioral tests, samples are collected. B Representative images of ACC virus infection. AAV-EGFP (Green) and DAPI (blue). Scale bars: 300 μm. C Changes in mechanical pain threshold measured by the von Frey test (VFT). D Quantification of the latency to eat in the novelty-suppressed feeding test (NSF). E Quantification of grooming duration in the splash test (ST). F Quantification of the immobile time in tail suspension test (TS). G Quantification of the total exploration time and differentiation index (DI) in the social interaction. (C–G: n  = 11,12,8 per group). All data are presented as mean ± s.e.m of the fold change of the CRC+Ctrl-ACC group and analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test (C–G). H The experimental timeline for the AAV injection into the ACC in both CRC+Ctrl-ACC and CRC+Gi-ACC group, establishment of CRC mice, the C21 injection, ACC tissue collection, RNAseq and data analyses. In brief, first, viruses were injected into the bilateral ACC brain regions, three weeks later, MC38 cells were implanted in the colon of mice in situ, then, C21 was intraperitoneally injected for two consecutive weeks, finally, samples are collected. I Heatmap showing relative expression levels of differentially expressed genes (DEGs, p < 0.05 by Deseq2) in the ACC, comparing the CRC+Gi-ACC group to the CRC+Ctrl-ACC group. J Volcano plot showing gene expression changes. red and blue dots represent upregulated and downregulated DEGs, respectively. K The heat map shows the relative expression levels of differentially expressed genes in the CRC+Gi-ACC vs CRC+Ctrl-ACC vs Sham+Ctrl-ACC groups. Sham+Ctrl-ACC: Sham-operated mice received bilateral injection of AAV9-hSyn-EGFP into the ACC, without MC38 tumor implantation, CRC+Ctrl-ACC: Mice received bilateral injection of AAV9-hSyn-EGFP into the ACC, followed by orthotopic implantation of MC38 cells into the colorectum three weeks later, CRC+Gi-ACC: Mice received bilateral injection of AAV9-hSyn-hM4Di(Gi)-EGFP into the ACC to chemogenetically suppress neuronal activity, followed by MC38 implantation three weeks later, CRC+Gq-ACC: Mice received bilateral injection of AAV9-hSyn-hM3Dq(Gq)-EGFP into the ACC to chemogenetically activate neuronal activity, followed by MC38 implantation three weeks later. The source data are provided as Supplementary Data 1 in the Figshare repository.

Inhibition of ACC activity partially improves local nerve density and inhibits colorectal cancer progression

To investigate changes in the enteric nervous system in CRC mice, we examined the parasympathetic, sympathetic, sensory and peptidergic nerves. The results showed that the densities of VAChT-labeled parasympathetic nerves (Fig. 6A, C), TH-labeled sympathetic nerves (Fig. 6B, D), and CGRP-labeled sensory nerves (Fig. 6E, G) were significantly decreased in the CRC group compared to the sham-operated control group. However, there was no significant change in peptidergic nerve density (Fig. 6F, H). Notably, inhibition of ACC activity in CRC mice led to a significant recovery of parasympathetic nerve density (Fig. 6A, C). These findings suggest that changes in parasympathetic nerve density may play a role in mediating depression-like behaviors in CRC mice. We further investigated the effects of ACC activity modulation on colorectal tumor progression. Inhibition of ACC activity (CRC+Gi-ACC group) significantly reduced tumor volume (Fig. 6), suppressed tumor cell proliferation—as indicated by decreased Ki67 expression in immunofluorescence (a marker of proliferating cells; Fig. 6I, K) and a downward trend in both Ki67 and PCNA levels (another proliferation marker) in western blot analysis (Fig. 6M–O). In parallel, tumor apoptosis was enhanced, as evidenced by a significant increase in cleaved caspase-3 expression (a marker of apoptosis; Fig. 6J, L, M, P).

Fig. 6: The effect of inhibition of ACC activity on nerve density and tumor progression in CRC mice.
Fig. 6: The effect of inhibition of ACC activity on nerve density and tumor progression in CRC mice.
Full size image

A Representative images of parasympathetic nerve markers VAChT staining in the colorectal tissue. VAChT (Green) and DAPI (blue). B Representative images of sympathetic nerve markers TH staining in the colorectal tissue. TH (Green) and DAPI (blue). Scale bar: 100 μm. C Quantitative of the density of VAChT-labeled parasympathetic nerve in the colorectal tissue across groups (n  =  5,8,12,6 per group). D Quantitative of the density of TH-labeled sympathetic nerve in the colorectal tissue across groups (n  =  5,10,12,6 per group). E Representative images of sensory nerve markers CGRP staining in the colorectal tissue. CGRP (Green) and DAPI (blue). F Representative images of peptiderergic nerve markers VIP staining in the colorectal tissue. VIP (Green) and DAPI (blue). Scale bar: 100 μm. G Quantitative of the density of CGRP-labeled sensory nerve in the colorectal tissue across groups (n  =  4,10,8,6 per group). H Quantitative of the density of VIP-labelled peptiderergic nerve in the colorectal tissue across groups (n  =  5,8,11,6 per group). I Representative images of cell proliferation markers Ki67 staining in the colorectal tumor tissue. Ki67 (Green) and DAPI (blue). Scale bar: 100 μm. J Representative images of apoptosis markers Cleaved-caspase3 staining in the colorectal tumor tissue. Cleaved-caspase3 (Green) and DAPI (blue). Scale bar: 100 μm. K Quantitative of the fluorescence intensity of the Ki67 in the colorectal tumor tissue across groups (n  = 10,10,6 per group). L Quantitative of the fluorescence intensity of the Cleaved-caspase3 in the colorectal tumor tissue across groups (n  = 11,10,5 per group). M Representative Western blot images showing the expression of Ki-67, PCNA, and cleaved caspase-3 in tumor tissues from CRC+Ctrl-ACC (Ctrl), CRC+Gi-ACC (Gi), and CRC+Gq-ACC (Gq) mice. N–P Quantification of Ki-67, PCNA, and cleaved caspase-3 protein levels normalized to β-Tublin in each group, (n = 5,5,6 per group; n = 5,5,6 per group, n = 5,6,6 per group). Q Representative images of tumor size in CRC+Ctrl-ACC, CRC+Gi-ACC and CRC+Gq-ACC groups. R Quantification of tumor volume in CRC+Ctrl-ACC, CRC+Gi-ACC and CRC+Gq-ACC groups, (n = 5,6,6 per group). All data are presented as mean ± s.e.m of the fold change of the Sham+Ctrl-ACC group and analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test (C, D, G, H). All data are presented as mean ± s.e.m of the fold change of the CRC+Ctrl-ACC group and analyzed by one-way ANOVA followed by Tukey’s multiple comparisons test (K, L, N–P, R). The source data are provided as Supplementary Data 1 in the Figshare repository.

Conversely, chemogenetic activation of ACC neurons via hM3Dq (CRC+Gq-ACC group) led to a significant increase in tumor volume (Fig. 6Q, R). Although local nerve density was not substantially altered, a slight decrease in parasympathetic (VAChT+) nerve fibers was observed (Fig. 6A, C). Immunofluorescence staining showed a mild, non-significant reduction in cleaved caspase-3 levels (Fig. 6J, L). Western blot analysis further demonstrated significantly elevated PCNA levels and decreased cleaved caspase-3 in the CRC+Gq-ACC group (Fig. 6M, O–P), suggesting enhanced tumor proliferation and reduced apoptosis in response to ACC hyperactivation. To validate the findings from hM4Di-mediated inhibition, we employed a parallel approach by expressing tetanus toxin light chain (Tettoxlc) to suppress ACC output. Consistent with the CRC+Gi-ACC group, the CRC+TeTx-ACC group exhibited a similar reduction in tumor (Supplementary Fig. 4Q, R), as indicated by significantly decreased Ki67 expression and significantly elevated cleaved caspase-3 levels in immunofluorescence (Supplementary Fig. 4I–L), as well as significantly decreased Ki67 and PCNA levels(Supplementary Fig. 4M–O) and a trend toward increased cleaved caspase-3 in western blot analysis(Supplementary Fig. 4M, P). Moreover, parasympathetic (VAChT + ) nerve density was also restored in the CRC+TeTx-ACC group (Supplementary Fig. 4A,C), whereas no obvious recovery was observed in sympathetic (TH+) or sensory (CGRP+) nerve fibers (Supplementary Fig. 4B, D, E, G) and there was also no significant change in peptidergic nerve density (Supplementary Fig. 4F, H), further supporting the role of ACC activity in modulating both tumor progression and autonomic nerve alterations in the tumor microenvironment. Together, these results suggest that ACC activity modulates both tumor progression and parasympathetic innervation in CRC mice, potentially linking emotion and tumor-related processes. While our data support a modulatory role of ACC in CRC development, further studies are needed to fully elucidate the underlying mechanisms.

Discussion

Clinical studies have demonstrated a high comorbidity between colorectal cancer (CRC) and depression13. In our orthotopic mouse model of CRC, we observed depression-like behaviors, which were associated with elevated activity in the anterior cingulate cortex (ACC). Notably, inhibiting ACC activity not only alleviated the depression-like behaviors but also slowed tumor progression in CRC mice. These findings collectively suggest that the ACC plays a pivotal role in regulating the comorbidity of CRC and depression. Previous studies on the comorbidity of colorectal cancer (CRC) and depression are increasing. It has been suggested that CRC-induced depression is associated with pain, fatigue, and inflammation47. Among colorectal cancer patients awaiting chemotherapy treatment, depressive and anxiety symptoms were positively associated with increased levels of interleukin-1 beta (IL-1β), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-ɑ) were detected48,49,50. However, mostly similar studies focus on correlations, limiting insights into the underlying causal mechanisms, which hampers the development of targeted intervention strategies. Few studies have specifically addressed the mechanisms in comorbidity of CRC and depression. Wang et. al demonstrated that the traditional Chinese medicine formula XCHT ameliorates depressive symptoms by modulating gut microbiota composition, inhibiting the TLR4/MyD88/NF-κB signaling pathway, and regulating inflammatory cytokine levels, ultimately yielding significant antitumoral effects in vivo51. Flores’ study identified miR-218 as a repressor of DCC (deleted in colorectal cancer) gene expression and found a significant reduction in miR-218 expression in the prefrontal cortex (PFC) of adult mice that are susceptible to chronic stress-induced social avoidance and adult non-medicated human with depression who died by suicide. In fact, the expression of miR-218 and DCC in the PFC are negatively correlated in depression. However, it has remained uninvestigated whether manipulations of the brain region could indeed affect CRC comorbid depression. Moreover, depression is regulated by multiple brain regions, with distinct subtypes characterized by specific neural features and varying responsiveness to treatments52. Hence, identifying the key brain region mediating CRC comorbid depression is important for the development of effective treatment. Previous study has shown that stimulation of hypothalamic oxytocin neurons can inhibit colorectal cancer progression53. However, they predominantly employed mouse models of colitis-associated cancer and targeted the hypothalamus, a brain region implicated in emotional regulation. In contrast, our study utilized a multifaceted approach to investigate CRC-associated alterations in central nervous system activity. We first conducted a comprehensive analysis of c-FOS expression in 27 brain regions involved in digestive, nociceptive, and emotional processing. To capture a representative phase of CRC progression with established depressive-like behaviors but minimal cachexia, we focused our main analysis on day 14 after tumor implantation. At this time point, c-FOS activity was significantly increased in multiple regions, including the IL, ACC, NAC, IC, RSG, MH, CA1 and DG. Notably, several of these regions—including the ACC, NAc, RSG, and CA1—are functionally interconnected components of the limbic system and are deeply involved in emotional and visceral integration54,55,56,57,58. The insular cortex (IC), while not a core limbic structure, is intimately involved in interoceptive awareness and decision-making processes related to taste, smell, and aversive stimuli. These regions collectively project to the hypothalamus and other subcortical structures, forming a distributed network that integrates emotional, cognitive, and visceral inputs59,60. Further combining gut-brain inverse mapping and network analysis, we concluded that, the activity of ACC, IC and MH brain regions was abnormally increased in CRC mice, and PL and DR brain regions played key roles in the global efficiency of the network in CRC mice. A time-course analysis was conducted to delineate the dynamics of neuronal activation across CRC progression, revealing a peak at day 14 and persistent activity in ACC and IL through day 21. Among them, ACC and IL showed a clear time-dependent increase in activity, whereas the RSG did not. These findings suggest that a subset of brain regions—especially the ACC—undergo progressive hyperactivation during CRC progression, which may underlie the emotional symptoms that emerge as the disease advances. From a systems neuroscience perspective, the consistent excitability changes observed across these interconnected regions in the CRC model may reflect a coordinated network-level dysfunction underlying the emotional and somatic features of depression. Previous studies have highlighted the involvement of regions associated with emotional processing, such as the thalamus, amygdala, ACC, and anterior insula, in the emotional aspects of pain and gastrointestinal symptoms including anxiety and depression33,34,35. Notably, the ACC has been identified as one of the most sensitive regions to chronic gut inflammation and visceral pain, with robust evidence indicating its role in mediating pain-related affective disturbances and mood disorders61. Consistent with these findings, our study revealed that among multiple brain regions screened, only c-FOS expression in the ACC showed a significant and positive correlation with the depression-like behavior score in CRC mice. Furthermore, chemogenetic inhibition of ACC activity not only alleviated depression-like behaviors but also modulated gene expression pathways associated with neuroinflammation and mood regulation. These results underscore the central role of ACC hyperactivation in CRC comorbid depression and support its involvement as a neural hub linking chronic gut pathology with affective dysregulation.

What are the potential mechanisms underlying CRC-induced hyperactivation of the ACC? First, by comparing the Sham versus the CRC mice, our RNAseq analysis of the ACC revealed transcriptomic changes hormone response impairment. Phillips’ study found that, local administration of corticosterone into the prefrontal cortex increased prefrontal dopamine efflux. By blocking prefrontal glucocorticoid receptors prevented deficits in cognitive function observed following exposure to acute stress37. These data suggest that impaired hormonal responses may render the body unable to cope with changes in stress caused by physical disease stimuli. Second, we also found that the expression of genes related to extracellular matrix organization processes was enriched (Fig. 4H). Stress and corticosterone have also been linked to increased deposition of perineuronal net proteins in rodent models related to depression62, perineuronal net are a form of extracellular matrix (ECM). A recent study identified an increase in hippocampal ECM to underlie hippocampus-dependent memory deficits at the sustained depressive-like state62. In exploring the mechanism by which chemogenetic inhibition of ACC activity improves depression-like behavior as described previously, we also compared the transcriptomic changes of ACC in CRC +Ctrl-ACC and CRC+Gi-ACC groups. We were excited to find that some antidepressant related genes (such as Sgk1, Plekhf1, Nlrp6) were significantly up-regulated (Fig. 5K). Previous studies have shown that Sgk1, a serine/threonine kinase, regulates ion channels, membrane transporters, and transcription factors, and its loss of expression is associated with depression39. Plekhf1, a glucocorticoid receptor pathway gene, has been linked to antidepressant behavior in studies of gut microbiota regulation in mice40. Nlrp6 is involved in maintaining hippocampal neurogenesis, which counters stress-induced depression41. These findings provide mechanistic insights into the remission of depression-like behaviors in CRC mice following modulation of ACC activity. In addition, we examined the genes that were up-regulated in the ACC of CRC mice compared to sham-operated controls and down-regulated after ACC regulation (Fig. 5L). It is worth noting that Hlx playing a role in enteric nervous system development63. Hlx regulates ENS development through transcriptional regulation of one or more mesenchymally-expressed genes that direct these processes45. Based on these data, it would be interesting to examine how enteric neurodevelopment changes during the co-development of CRC and depression. Nerves have long been recognized as tissues affected by cancer cells, and cancer signals induce nerve growth and innervation through a variety of mechanisms. Ayala confirmed cancer-related axonogenesis (the enlargement of nerves or increase in nerve density) and demonstrated an association between prostate cancer and neurogenesis (an increased number of neurons)64. During the development of human pancreatic cancer, the pancreatic nerve changes significantly, these include intrapancreatic nerve hypertrophy and increased nerve density65. In contrast, this study focused on the changes in the nerve density of autonomic, sensory and peptidoergic nerves during the occurrence and development of CRC. Our findings show that both autonomic and sensory nerves are disrupted in CRC model, as shown by decreased nerve density (Fig. 6A-E, G), this seriously destroys gut functions and intestinal homeostasis. Previous studies have also revealed the relationship between the regulation of the nervous system and intestinal tumors. For example, activation of the autonomic nervous system (ANS) and the sympathetic-adrenal axis promotes gastrointestinal tumorigenesis, and chemical sympathectomy with 6-hydroxydopamine can reduce the incidence of CRC in rats66. Vagotomy and atropine administration of parasympathetic denervation significantly reduced tumor incidence, cell proliferation, tumor volume and weight, as well as angiogenesis mediated by NGF, β2 adrenergic, and muscarinic M3 receptors67. Furthermore, there is substantial evidence supporting the role of neurosignaling molecules, including neurotransmitters (e.g., dopamine, gamma-aminobutyric acid, acetylcholine, serotonin, norepinephrine, and glutamate) and neurotrophic factors in CRC development68. While previous studies investigating the nervous system’s role in tumor progression have primarily focused on neurotransmitters or relied on long-term, irreversible nerve ablation models, our study explored the modulation of ACC activity as a potential approach. We demonstrated that ACC activity influences nerve density within intestinal tumors, with ACC blockade reversing local parasympathetic nerve disruption in CRC mice. This effect may be linked to transcriptomic changes associated with neural development within the ACC. Unexpectedly, in addition to enhancing local parasympathetic nerve density, we observed that reducing ACC activity suppressed tumor proliferation and promoted apoptosis in orthotopic colorectal cancer models. These effects were supported by molecular and histological evidence indicating decreased tumor growth and enhanced cell death following ACC inhibition. In contrast, chemogenetic activation of ACC neurons exacerbated tumor progression and was associated with diminished parasympathetic signaling. Together, these findings suggest that the ACC exerts a bidirectional influence on tumor biology. Inhibiting its activity not only alleviates CRC-associated depressive-like behaviors but may also restrain tumor development, potentially via restoration of vagal-parasympathetic integrity and modulation of neuroimmune interactions. These results highlight an underappreciated brain-tumor regulatory axis, wherein central modulation of emotional brain circuits can impact peripheral tumor behavior through autonomic pathways. In conclusion, our study highlights the critical contribution of elevated ACC activity to the co-development of CRC and depression. These findings provide valuable insights and an experimental basis for the development of combined brain–tumor intervention strategies in clinical practice. Several limitations should be acknowledged: although we employed a well-established MC38-based CRC mouse model combined with validated behavioral assays to assess depressive-like phenotypes, rodent models may not fully recapitulate the complexity and heterogeneity of CRC–depression comorbidity observed in humans. Therefore, future work should focus on developing preclinical models that more accurately reflect human disease features, thereby enhancing the translational relevance of our results.

Methods

Animals

Eight-week-old female and male C57BL/6J mice were purchased from the Institute of Experimental Animals of Sun Yat-sen University. The mice were housed in groups of five in a specific pathogen-free (SPF) facility under a 12-h light/dark cycle, with ad libitum access to food and water. All experiments were conducted during the light phase, and animals were randomly assigned to the various experimental groups. Both male and female mice were used but their sex was not recorded and so cannot be reported. All animal studies were reviewed and approved by the Institutional Animal Care and Use Committee of Sun Yat-sen University (approval number: SYSU-IACUC-2024-000134), we have complied with all relevant ethical regulations for animal use. For surgical tumor implantation, mice were anesthetized with a mixture of teletamine, zolazepam (85 mg/kg, ip, Zoletil 50®, BAILAIYUAN, China) and acepromazine (10 mg/kg, ip, MCE, USA) and given meloxicam (5 mg/kg, subcutaneously) for perioperative analgesia. Post-surgical monitoring was conducted daily. Mice were euthanized using CO2 followed by cervical dislocation. Tumor volume was calculated using the formula: volume =  0.5 × length × width × height69. Mice were euthanized when tumor volume exceeded 1.5 cm3 or signs of distress (weight loss >15%, lethargy) were observed, in accordance with humane endpoints.

Establishment of mice with orthotopic colorectal cancer

MC38 cells (ATCC) at 2–3 passages, reaching 70%–80% confluence, were used for orthotopic tumor implantation. Mice were anesthetized as described in the “Animal” section of the Methods. A suspension of 106 MC38 cells in PBS was mixed 1:1 with ice-cold Matrigel (Corning, USA; Cat#354248) to a final volume of 20 µL. Under a stereomicroscope, the colorectal mucosa approximately 1 cm from the anus was gently grasped with blunt forceps, retracted to expose the area, and the cell mixture was injected into the submucosa using an insulin syringe. The needle was held in place for an additional 30 seconds to ensure proper tissue absorption before removal.

Behavioral tests

To evaluate depression-like behaviors, we employed four widely used behavioral paradigms: the novelty-suppressed feeding test (latency to feed), splash test (grooming time), tail suspension test (immobility time), and social interaction test (social preference index). These tests are well-validated indicators of behavioral despair, anhedonia, and social withdrawal in rodent models of depression70,71. Each behavior was analyzed independently to provide a multi-dimensional profile of depression-like behavior in CRC mice. Behavioral tests were conducted 14 days after the establishment of CRC in mice. Prior to testing, the mice were allowed to acclimate to the testing room for at least 2 h. On the first day, the novelty-suppressed feeding test was performed in the morning, followed by the grooming test in the afternoon. On the second day, social interaction testing was conducted, and pain threshold testing was performed on the third day in the morning, with the tail suspension test conducted in the afternoon. The behavioral testing protocols followed our previous study72.

Novelty-suppressed feeding

After food restriction for 24 h, mice were placed into one of the corners of a white plastic box (40 × 40 × 40 cm) with a single food pellet present in the center. The latency to begin food consumption in a 10-m test was measured.

Splash test

The mice were placed in the cage for free movement for 1 min, and then 10% sucrose water was sprayed on the buttock and back of the mice, and the grooming time of the mice within 5 min was recorded.

Social Interaction test

The social interaction test was conducted in a (60 cm × 40 cm × 20 cm) cuboid field, which was divided into three equal-sized rooms by two transparent partitions. In each partition, a square opening (5 cm × 5 cm) was centrally located, and mice had access to each chamber. Two small round iron cages were placed in the left and right ventricles. An unfamiliar mouse was habituated to the room for 10 min before a day of testing to minimize excessive stress. On the day of testing, a test mouse was first explored for 5 min in a field holding an empty wire cage. The stranger mice were then housed in a wire cage, while the other cage contained a toy mouse. The test mice were placed into the field for 5 min. The social interaction index (DI%) was calculated as: [(time spent exploring the unfamiliar mouse - time spent exploring the toy mouse) / total time spent exploring both] ×100.

von Frey test

Mice were placed in a metal mesh cage in a quiet environment and allowed to acclimate for 10–15 min. Von Frey filaments, with forces ranging from 0.008 g to 4 g, were then used to apply pressure to the soles of the mice. A positive response was recorded when the mouse exhibited a nociceptive reaction, such as withdrawal or flinching. The pressure value of the filament that elicited the response was recorded.

Tail suspension test

The mouse was suspended by attaching one end of a 10 cm tape to its tail and the other end securely to a hanging frame in a quiet environment. The mouse was suspended for 5 min, during which the total immobility time, defined as the duration in which the mouse ceased struggling, was recorded.

Colorectal tracer injection in mice

Mice were anesthetized as described in the “Animal” section of the Methods. Vaseline was applied to their eyes to prevent dryness. Under a body microscope, the colorectal mucosa ~1 cm from the anus was gently lifted using blunt forceps, exposed to the field of view, and the tracer was drawn into a Hamilton syringe connected to a syringe pump. The needle was inserted into the colorectal mucosa or inject into the tumor that has been implanted with MC38 for two weeks at a 45° angle, and 1 µL of PRV (PRV-CAG-EGFP, 4.00E + 8 PFU/mL, BrainVTA, China) was injected at three sites. Following the injection, the needle was left in situ for an additional 2 min, and the mice were allowed to recover on a heated pad.

Stereotaxic injection into the brain

Mice were anesthetized as described in the “Animal” section of the Methods. Vaseline was applied to their eyes to prevent dryness. The mice’s head fur was shaved, and the area was sterilized using iodine followed by 75% ethanol. The scalp was incised with surgical scissors, and the head was secured in a stereotaxic apparatus. A heating pad was placed beneath the head to maintain the mouse’s body temperature. The anterior fontanel was used as the reference point, and the coordinates were set to AP: 0.98 mm, ML: ±0.25 mm, DV: −2 mm. Virus was injected through a Hamilton steel needle at a rate of 20 nL/min, with a total volume of 100 nL per side, and the needle was held in place for 5 min to allow for complete infusion. For chemogenetic inhibition of ACC neurons, AAV9-hSyn-hM4Di(Gi)-EGFP (BrainVTA, China) and AAV8-hSyn-hM4Di(Gi)-mCherry (BrainVTA, China) was used, for chemogenetic activation of ACC neurons, AAV9-hSyn-hM3Dq(Gq)-EGFP (BrainVTA, China) was used, to suppress ACC neurons output, AAV9-hSyn-Tettoxlc- EGFP (BrainVTA, China) was used, while the control group received AAV9-hSyn-mCherry (BrainVTA, China) and AAV9-hSyn- EGFP (BrainVTA, China).

Chemogenetic inhibition of ACC neurons

To reduce the excitability of ACC neurons, mice expressing the inhibitory chemogenetic receptor hM4Di in the ACC were injected i.p. with 1 mg/kg compound 21 (C21, Tocris Bioscience, cat# 6422), once per day, starting from the establishment of CRC mice.

Hematoxylin-eosin staining

Mice were euthanized using CO2 followed by cervical dislocation. Open their abdominal cavities, and cut out the colonic and intestinal tissues. Wash the tissues with normal saline, and immediately immerse them in neutral formalin fixative for fixation for 30–50 min. After fixation, rinse with running water for several hours. The tissues are then dehydrated successively with 70%, 80%, and 90% ethanol solutions for 30 min each, followed by two more times of 95% and 100% solutions, each for 20 min. Then, immerse them in a 15-min mixture of pure alcohol and xylene, xylene for 15 min each (until transparent), mixture of xylene and paraffin for 15 min each, and finally paraffin I and paraffin II for dewaxing for 50–60 min each. Then, proceed with embedding and sectioning. Prepare 3–5 µm paraffin sections for staining, and bake the slides in an oven at 65 °C for 2 h. Perform the following steps sequentially: Deparaffinization: Place the sections on a slide rack and immerse them in xylene I for 10 min, xylene II for 10 min, and xylene III for 5 min. Alcohol Hydration: Gradually rehydrate the sections using absolute ethanol I (10 min), absolute ethanol II (10 min), 95% ethanol (5 min), and 70% ethanol (3 min). Hematoxylin Staining: Stain the nuclei with hematoxylin for 3 min, then rinse thoroughly with distilled water to remove excess stain. Differentiation: Differentiate the sections in 1% hydrochloric acid alcohol differentiation solution for 30 s, followed by rinsing under tap water I for 3 min and tap water II for another 3 min. Eosin Staining: Stain the cytoplasm with eosin for 2 min, then rinse with distilled water to remove excess stain. Dehydration: Sequentially dehydrate the sections in 75% ethanol, 85% ethanol, 95% ethanol, and absolute ethanol for 1 min each. Clearing: Immerse the sections in xylene I for 1 min, followed by xylene II for 1 min. Mounting: Apply a suitable amount of neutral resin to the center of the tissue, then gently lower a coverslip onto the slide.

Preparation of immunofluorescence staining samples

After acclimatization in a quiet environment for more than 2 h, mice were anesthetized with a mixture of teletamine, zolazepam (85 mg/kg, ip, Zoletil 50®, BAILAIYUAN, China) and acepromazine (10 mg/kg, ip, MCE, USA) and the hearts were exposed by opening the thoracic cavity using surgical scissors. A needle was used to infuse 1×PBS for 2 min (10 mL /min) through the mouse heart, followed by 4% PFA for 5 min (10 mL /min). The colorectum and brain were removed and placed in 4% PFA at 4 °C overnight. Then they were transferred to 30% sucrose solution and stored at 4 °C until sedimentation. The colorectum and brain was removed from the sugar water and the surface sugar water was carefully blotted dry. They were then embedded in optimal cutting temperature (OCT) embedding agent and flash-frozen and embedded in a microtome at −40 °C. For handling colorectal tissues, the embedded tissue was placed on the sample stage of a freezing microtome with a thickness of 20 μm, and the sections were directly attached to slides and stored in a refrigerator at −80 °C until use. For handling brain tissue, the brain was sectioned continuously at a thickness of 30 μm, placed in PBS for storage, and then prepared for staining.

Immunofluorescence staining of brain slices

The brain slices,30 μm thick, were placed into 24-well plates containing blocking solution and incubated for 2 h at room temperature with blocking solution. After that, the brain slices were incubated in primary antibody (diluted in the blocking buffer) for c-FOS (Cell Signaling Technology, USA; cat#2250S; 1:500) for 48 h at 4 °C and rinsed with 0.4%PBST three times at room temperature for 10 min each time. They were then placed in secondary antibodies (diluted in the blocking buffer) for Anti-rabbit IgG (H + L) Alexa Fluor® 488(Cell Signaling Technology, USA; cat# 4412; 1:500) and incubated at room temperature in the dark for 2 h. Rinse with 0.4%PBST three times at room temperature in the dark for 10 min each time. Finally, the cells were rinsed in 1×PBS for 10 min in the dark. Brain slices were laid flat on slides, allowed to dry, and then sealed with drops of anti-fluorescence quench for observation. NIKON Ti2-E fluorescence microscope (Nikon, Japan) was used.

C-FOS quantification and network analysis

To evaluate neuronal activity and interregional functional connectivity across the brain, we first quantified c-FOS+ neurons in multiple brain regions using ImageJ software (NIH). Regions of interest (ROIs) were manually delineated on fluorescence images according to anatomical landmarks from the Allen Mouse Brain Atlas. The density of C-FOS+ cells was calculated as the number of c-FOS+ nuclei per mm² within each ROI. For each group, we included bilateral counts from 2 to 3 representative coronal sections per region per animal. Subsequently, interregional relationships were assessed by computing the Pearson correlation coefficients (Pearson’s r) between c-FOS+ cell densities of each pair of regions across all animals within each experimental group. These correlation matrices were visualized as heatmaps to provide an overview of pairwise associations. To construct functional brain networks, we selected region pairs showing statistically significant and strong correlations (Pearson’s r > 0.6, p < 0.05) to define edges in the network. Each brain region was represented as a node, and the strength of interregional correlations defined the edges. Graph-theoretical analysis was performed to calculate the global efficiency (GE) of each network, which reflects the average inverse shortest path length among all nodes and is defined as:

$${GE}=\frac{1}{N\left(N-1\right)}{\sum}_{i\ne j}\frac{1}{{d}_{{ij}}}$$

where N is the number of nodes and d(ij) is the shortest path length between node i and node j. To identify the most influential regions, we implemented a node-deletion strategy: for each node, we recalculated the GE after its removal and computed the change (ΔGE). The top five nodes causing the greatest reductions in GE were defined as core nodes in the network, representing functionally central regions likely critical for maintaining network integration.

Immunofluorescence staining of colorectal frozen sections

Take out the colorectal frozen sections and let them dry at room temperature for 0.5 h (away from light). Then put the sections into a jar with 1×PBS and wash for 5 min to wash out OCT. Dry the edges of the sections with absorbent paper, draw closed circles along the edges of the slides with an immunohistochemical pen, place them flat in a wet box, and add 200 µL blocking solution. After sealing at room temperature for 2 h, tip off the blocking solution, drain the excess water on the filter paper pad, place it flat in the wet box, add 200 µL primary antibody (diluted in the blocking buffer) for VAChT (Synaptic Systems, Germany; cat#139103; 1:500), TH (Cell Signaling Technology, USA; cat#E2L6M; 1:500), CGRP (Cell Signaling Technology, USA; cat#14959; 1:500), VIP (Cell Signaling Technology, USA; cat#63269; 1:500), Ki67 (Cell Signaling Technology, USA; cat#9129; 1:500) and cleaved caspase-3 (Cell Signaling Technology, USA; cat# 9664; 1:500) and incubate at 4 °C for 12 h, and then incubate at room temperature for 4 h. On the second day, the colorectal frozen sections were rinsed with 0.4% PBST three times at room temperature for 10 min each time. Add 200 µL secondary antibody (diluted in the blocking buffer) for Anti-rabbit IgG (H + L) Alexa Fluor® 488 (Cell Signaling Technology, USA; cat# 4412; 1:500) and apply for 2 h at room temperature in the dark, and rinse with 0.4% PBST by shaking for 3 times, 10 min each time. Excess water was removed, DAPI-containing sealed tablets were sealed, NIKON Ti2-E fluorescence microscope (Nikon, Japan) was used.

Protein extraction, and western blot analysis

Tumor samples were immediately transferred to a mortar pre-filled with liquid nitrogen and ground into a fine powder using a pestle in the presence of liquid nitrogen. Subsequently, lysis buffer (Cell Signaling Technology, USA, cat#9803S)) was added to the powdered tissue, and homogenization was continued. The homogenate was centrifuged at 12,000 × g for 20 min at 4 °C, and the supernatant was carefully aspirated and transferred to a new tube. Protein concentration was determined using the BCA assay (Beyotisme, Chian, cat#P0012). For Western blot analysis,30 μg of total protein extract per lane was separated on a denaturing SDS-PAGE gel and subsequently transferred onto a 0.22 µm PVDF membrane using electroblotting. The following primary antibodies were used: Cleaved Caspase-3 Rabbit Antibody (1:1000, Cell Signaling Technology, USA, cat#9664S), Ki-67 Polyclonal antibody (1:500, Proteintech, USA, cat#27309-1-AP), PCNA Polyclonal antibody (1:5000, Proteintech, USA, cat#10205-2-AP), Beta Tublin Polyclonal antibody (1:5000, Proteintech, USA, cat#10094-1-AP) for loading normalization. The secondary antibody, anti-rabbit IgG (1:4000, Cell Signaling Technology, USA, cat#7074S), was used for detection. After incubation with the antibodies, the Immobilon-FL membranes were washed three times with TBST for 5 min each. The membranes were developed using enhanced chemiluminescence (ECL) solution and imaged with a chemiluminescence imaging system.

RNA sequencing

Sequencing samples were obtained from sham-operated control group (Sham+Ctrl-ACC), mice with orthotopic colorectal cancer group (CRC+Ctrl-ACC) and colorectal cancer chemogenetically suppressed ACC neuronal activity group (CRC+Gi-ACC), with 4–7 samples from each female and male group. RNA amplification library building and sequencing were performed by Nuohezhiyuan Inc. One μg of total RNA with a RIN value higher than 7 was used for the next step of amplification and library construction. The NEBNext® Ultra™ RNA Library Prep Kit for Illumina® was used for RNA amplification and cDNA library construction. Libraries were loaded onto an Illumina HiSeq instrument 6000 (USA) for 150 bp double-end sequencing. Raw sequencing reads were filtered to remove adaptor sequences, N-containing reads, and bases of low read quality. The quality of reads was assessed with the use of Fastp, version 0.19.7. The processed sequences were then aligned to the mm9 mouse reference genome (UCSC) using Hisat2 (version 2.0.1). Genes with counts <5 reads in more than 70% of the samples (low expression genes) were excluded from the analysis. DESeq2 package (version 1.40.2) of R was used to standardize the readings and perform the analysis of differentially expressed genes (DEGs). The screening conditions of differentially expressed genes were p < 0.05. Differential gene results were visualized as heat maps and volcano maps using the pheatmap, ggplot2 and ggpubr packages. Gene Set Enrichment Analysis (GSEA) software was used to analyze the differentially expressed Genes in the Kyoto Encyclopedia of Genes and Genomes (GSEA) database. KEGG pathway analysis (c2.cp.kegg.v2023.1.symbols.gmt), molecular function of Gene ontology (MF, c5.go.mf.v2023.1.symbols.gmt), biological process (BP, c5.go.bp.v2023.1.symbols.gmt) and cellular component (CC, c5.go.cc.v2023.1.symbols.gmt).

Statistics and reproducibility

Fluorescence image analysis was performed using image J software (Version 1.53j). GraphPad Prism 8.0 software was used for data analysis and expressed as mean ± standard error (mean ± s.e.m). We conducted the Shapiro–Wilk test to evaluate the normality of the datasets. For datasets which were normally distributed, outliers were identified as being >2 standard deviations from the mean and excluded. For two-group comparisons, two-tailed unpaired Student’s t test was used for normally-distributed datasets, and the Mann–Whitney test was used for non-normally distributed datasets. For multiple-group comparisons, one-way ANOVA with Tukey’s post hoc tests was used, and p < 0.05 was considered as statistically significant. Images were processed using Adobe Illustrator CC software (Version 19, 23.0.2). Group sizes were determined based on prior work and the literature. Each data point represents an independent biological sample, and the number of biological replicates (n) for each experiment is indicated in the figure legends. For all animal studies, each experimental group contained at least [n = 4] independent animals. The mice used in the experiments were randomly assigned to the different experimental groups.