Mechanisms of the gut-lung axis in regulating lung metastasis of colorectal cancer and advances in traditional Chinese medicine intervention

Abstract

Pulmonary metastasis is a major cause of mortality in patients with colorectal cancer (CRC), but the underlying mechanisms remain incompletely understood. The gut-lung axis, which links the gut microbiota to lung homeostasis, has emerged as a key regulator of CRC progression and metastasis. Traditional Chinese medicine (TCM), based on the classical theory of the “interior-exterior relationship between the lung and large intestine”, may exert therapeutic effects through modulation of this axis. This review aimed to summarize the mechanisms by which the gut-lung axis regulates CRC pulmonary metastasis and to evaluate TCM-based interventions targeting this axis. A systematic search was conducted in PubMed, Web of Science, and China National Knowledge Infrastructure up to March 2026. Clinical and preclinical studies were included. Gut microbiota dysbiosis, characterized by enrichment of Faecalibacterium and enterotoxigenic Bacteroides fragilis, promotes epithelial-mesenchymal transition (EMT), remodels the pre-metastatic niche (PMN), and contributes to chemoresistance. Exosomes carrying specific microRNAs and circular RNAs can reach the lungs and create a microenvironment favorable for metastasis. Gut-derived immune cells, particularly T helper 17 and regulatory T cells, migrate to the lungs and shape local immune responses. In addition, the gut-lung leak axis driven by lipopolysaccharide/Toll-like receptor 4 signaling exacerbates pulmonary inflammation. TCM formulas, including Si-Jun-Zi decoction and combinations of Sophora japonica and Scutellaria baicalensis, as well as active compounds such as geniposide, berberine, and Astragalus polysaccharides, may restore microbial diversity, repair intestinal barrier function, inhibit EMT, and regulate exosome release. However, clinical evidence remains limited, and substantial translational challenges persist. The gut-lung axis represents a promising therapeutic target for CRC lung metastasis. Although TCM may regulate this axis through multiple mechanisms, current evidence is predominantly preclinical. Further mechanistic studies, particularly those clarifying causal relationships among microbial taxa, exosomal cargoes, and lung PMN formation, are needed before reliable clinical translation can be achieved. Well-designed randomized controlled trials remain essential for validating TCM-based interventions.

Key Words: Gut-lung axis; Colorectal cancer; Lung metastasis; Traditional Chinese medicine; Microbiota

Core Tip: This review has three main objectives: (1) To analyze the molecular and cellular mechanisms by which the gut-lung axis regulates lung metastasis of colorectal cancer; (2) To critically summarize the evidence for traditional Chinese medicine therapies targeting this axis; and (3) To identify knowledge gaps and propose future research directions. Our main focus is on preclinical studies, as human data is still scarce, but we also discuss the challenges in translational research and the potential of microbiome-based biomarkers.

INTRODUCTION

Colorectal cancer (CRC) ranks third in global incidence and second in cancer-related mortality[1,2]. Despite therapeutic advances, approximately 25% of patients present with synchronous metastasis, and an additional 25%-30% develop metachronous metastasis following resection[3]. The lungs represent the second most common site of metastasis, with 31.7% of patients with metastatic CRC developing pulmonary metastases; KRAS mutations occur more frequently in lung lesions (42%) than in liver lesions (28%)[4-6]. The 5-year survival rate for stage IV CRC remains below 15%[7]. The gut-lung axis - a bidirectional network involving microbial metabolites, immune cells, and exosomes - has recently emerged as a key regulator of CRC progression and metastasis[8-11]. Dysbiosis of this axis has been associated with various pulmonary diseases, and specific bacteria such as Fusobacterium nucleatum (F. nucleatum) and enterotoxigenic Bacteroides fragilis (ETBF) are now considered pro-metastatic “cancer symbionts”[12-15]. Gut microbiota may influence the lungs through at least four interrelated pathways: Circulating microbial metabolites[16], migration of immune cells[17], translocation of bacterial components through a compromised intestinal barrier[18], and exosome-mediated transfer of pro-metastatic cargo[19,20]. In parallel with these modern findings, the traditional Chinese medicine (TCM) theory of “the lung and large intestine being interiorly-exteriorly related” provides an ancient conceptual framework[21]. This theory proposes that pathological changes in one organ can influence the other through meridians and qi; accordingly, TCM practitioners have historically used lung-tonifying prescriptions to treat intestinal disorders and vice versa[22,23].

This review has three objectives: (1) To analyze the mechanisms by which the gut-lung axis regulates CRC lung metastasis; (2) To critically evaluate TCM interventions targeting this axis; and (3) To identify existing knowledge gaps and future research directions. We primarily focus on preclinical evidence and translational challenges, including microbiome-based biomarkers. The latest evidence up to early 2026 has been included[24,25] (Figure 1).

Figure 1 A schematic diagram that shows the gutlung axis in the situation of colorectal cancer lung metastasis. The key essential elements are presented: Gut microbial flora (the dysbiosis which involves Fusobacterium nucleatum and enterotoxigenic Bacteroides fragilis), metabolic products (short-chain fatty acids, succinate), intestines barrier (gut leakage), exosomes (which carry microRNAs and circular RNAs), immunity cell movement (T helper 17, regulatory T), and the lung pre-metastatic niche. LPS: Lipopolysaccharide; ETBF: Enterotoxigenic Bacteroides fragilis; SCFAs: Short-chain fatty acids; Th17: T helper 17; Treg: Regulatory T.

Search strategy and study selection

A systematic literature search was conducted in PubMed, Web of Science, and China National Knowledge Infrastructure from database inception to March 2026. Search terms included Medical Subject Headings and keywords: (“colorectal neoplasms” OR “colon cancer” OR “rectal cancer”) AND (“lung metastasis” OR “pulmonary metastasis”) AND (“gastrointestinal microbiome” OR “gut microbiota” OR “dysbiosis”) AND (“exosomes” OR “extracellular vesicles”) AND (“traditional Chinese medicine” OR “herbal medicine”). Only full-text articles published in English or Chinese were included. The inclusion criteria were as follows: (1) Original research articles or reviews investigating the gut-lung axis in CRC; (2) Studies reporting TCM interventions (formulas, active compounds, or derived nanoparticles) targeting lung metastasis; and (3) Preclinical in vivo studies or clinical studies of any design. The exclusion criteria included case reports, editorials, conference abstracts without full data, and studies focusing on non-colorectal primary tumors. For clinical studies, methodological quality was assessed using the Cochrane Risk of Bias tool for randomized trials and the Newcastle-Ottawa Scale for cohort studies. For preclinical studies, we recorded whether blinding, randomization, sample size calculation, and model relevance (orthotopic vs tail-vein injection models) were reported.

COMPONENTS OF THE GUT-LUNG AXIS

Before discussing how this axis is disrupted in CRC metastasis, it is important to first describe its normal organization. The gut-lung axis consists of five interconnected components: Gut microbiota and their metabolites, the intestinal barrier, the immune system, the circulatory and lymphatic systems, and the neuroendocrine system.

Gut microbiota and metabolites

A healthy gut microbiota is dominated by Firmicutes and Bacteroidetes, with lower abundances of Actinobacteria, Proteobacteria, and Verrucomicrobia[26,27]. These bacteria ferment dietary fiber into short-chain fatty acids (SCFAs) - acetate, propionate, and butyrate. SCFAs serve as the primary energy source for colonocytes and also signal through G protein-coupled receptors (GPR41 and GPR43) expressed on immune cells[28,29]. Butyrate is of particular interest because it inhibits histone deacetylases and exerts anti-inflammatory effects[30]. Additional relevant metabolites include secondary bile acids (deoxycholic acid and lithocholic acid) produced by Clostridium species, tryptophan metabolites (indole and indole-3-aldehyde) that activate the aryl hydrocarbon receptor, and trimethylamine N-oxide, which is derived from dietary choline[31,32].

In patients with CRC, the gut microbial composition is altered toward a pro-inflammatory and pro-tumorigenic state. Microbial diversity is reduced; SCFA-producing bacteria such as Prevotella and Roseburia are decreased; whereas potential pathogenic taxa, including Clostridium perfringens, ETBF, Escherichia coli (E. coli) harboring the pks island, Streptococcus parasanguinis, and Micrococcus luteus, are enriched[33,34]. This dysbiotic state is not merely a consequence of tumor development but actively contributes to malignant progression through multiple mechanisms, including genotoxins (e.g., colibactin produced by pks+ E. coli), toxins [e.g., Bacteroides fragilis toxin (BFT) from ETBF], and adhesins (e.g., FadA and Fap2 from F. nucleatum)[35,36]. Recent metagenomic studies have further identified strain-level differences associated with metastatic potential[37,38].

Immune cell trafficking

The gut-associated lymphoid tissue contains more immune cells than any other organ system. Gut microbiota regulate the differentiation of naïve T cells into T helper 17 (Th17), regulatory T cells (Treg), Th1, and Th2 subsets[39]. These cells express homing receptors, such as α4β7 integrin and chemokine receptor 9 (CCR9), which initially retain them within the gut; however, under inflammatory conditions, they can downregulate these receptors and upregulate lung-homing markers, including C-X-C motif chemokine receptor 3, CCR4, and CCR6[40,41]. For instance, segmented filamentous bacteria in the murine gut induce Th17 cells that, upon activation, migrate to the lung and exacerbate allergic airway inflammation[42]. In CRC, gut-derived Th17 cells have been detected in circulation and metastatic lesions, suggesting a potential migratory route from the colon to the lung[43]. Single-cell RNA sequencing has recently mapped the trajectories of these cells from the gut to the lung in mouse models[44].

Communication mediated by extracellular vesicles

Exosomes are extracellular vesicles with a diameter of approximately 30-150 nm that are secreted by almost all cell types and carry proteins, lipids, and nucleic acids[45]. Tumor-derived exosomes have been shown to contribute to pre-metastatic niche (PMN) formation by transferring oncogenic microRNAs and inducing stromal reprogramming[46]. In contrast, bacterial extracellular vesicles, including outer membrane vesicles (OMVs), can also be produced by intestinal bacteria and enter the host circulation[47]. OMVs derived from Nocardia species have been detected in the blood of patients with CRC and may transport the adhesin FomA to distant organs[20]. Moreover, Nocardia infection of CRC cells can alter the exosomal cargo of the host: These exosomes contain elevated levels of miR-1246, miR-92b-3p, miR-27a-3p, and circular RNA hsa_circ_0004085, which may enhance chemoresistance and promote metastasis[48].

“Leaky gut syndrome” phenomenon

The integrity of the intestinal barrier depends on tight junction proteins (such as claudins, occludin, and zonula occludens-1), the mucus layer, and antimicrobial peptides. Dysbiosis of the gut microbiota, chemotherapy [particularly irinotecan and 5-fluorouracil (5-FU)], and tumor-derived factors can disrupt this barrier, leading to a “leaky gut” state, also referred to as the “intestinal leak” phenomenon[49,50]. Subsequently, bacterial products such as lipopolysaccharide (LPS) can enter the portal and systemic circulation. In the lungs, LPS activates pulmonary alveolar macrophages and epithelial cells via Toll-like receptor 4 (TLR4), triggering the nuclear factor kappa B (NF-κB)-dependent production of tumor necrosis factor α (TNF-α), interleukin-6 (IL-6), and C-X-C motif chemokine ligand 8 (CXCL8)[51]. This low-grade, chronic inflammatory state creates a favorable microenvironment for tumor cell extravasation and growth[52]. The term “intestinal leak” has also been used to describe increased alveolar-capillary permeability associated with endotoxemia of intestinal origin[53]. Emerging evidence suggests that butyrate supplementation can partially reverse this process[54].

THE INTESTINAL-LUNG AXIS MECHANISM IN LUNG METASTASIS OF CRC

As mentioned in the introduction section, the gut-lung axis comprises multiple components, including the microbiota, immune cells, exosomes, and the gut barrier. In the context of CRC lung metastasis, these components become dysregulated, as outlined below.

Bacteria inhibiting metastasis and epithelial-mesenchymal transition

Epithelial-mesenchymal transition (EMT) is a critical process in cancer metastasis. During EMT, CRC cells lose E-cadherin expression and acquire vimentin and fibronectin, thereby gaining migratory and invasive properties. Increasing evidence suggests that intestinal bacteria are associated with EMT regulation in CRC. F. nucleatum is the most extensively studied species in this context. It uses its adhesin FadA to bind E-cadherin on CRC cells, thereby triggering β-catenin translocation to the nucleus and subsequent transcription of c-Myc, cyclin D1, and EMT-related genes[55,56]. In a mouse model of CRC lung metastasis, intratumoral injection of F. nucleatum increases the number of lung nodules while downregulating E-cadherin expression and upregulating Snail and Zeb1 expression[57]. Mechanistically, F. nucleatum may activate the TLR4/MyD88/NF-κB signaling pathway, which in turn may induce Snail expression[58]. Additionally, F. nucleatum produces succinate, a metabolic product that can stabilize hypoxia-inducible factor-1 alpha and promote EMT[59] (Table 1).

Table 1 Key pro-metastatic bacteria and their mechanisms.

Bacterium	Virulence factor	Mechanism promoting metastasis	Evidence level (additional notes)
Fusobacterium nucleatum	FadA, Fap2	E-cadherin/β-catenin, TLR4/NF-κB, exosomal miR-1246	Preclinical + human correlative; most extensively studied
Enterotoxigenic Bacteroides fragilis	BFT	E-cadherin cleavage, induction, neutrophil recruitment	Preclinical only (so far)
pks+ E. coli	Colibactin	DNA damage, genomic instability, mutational signatures	Human colorectal tumor evidence; direct lung metastasis link less clear
Peptostreptococcus anaerobius	Putative surface protein	TLR2/4 activation, ROS production	Preclinical
Parvimonas micra	Unknown	Associated with advanced stage and metastasis	Human cohort study; virulence factor unknown

TLR4: Toll-like receptor 4; NF-κB: Nuclear factor kappa B; BFT: Bacteroides fragilis toxin; E. coli: Escherichia coli; ROS: Reactive oxygen species.

ETBF secretes BFT, a 20-kDa metalloproteinase that has been reported to cleave E-cadherin, potentially leading to the release of β-catenin[60]. BFT also activates c-Jun and c-Fos, promoting the production of IL-8 and the recruitment of neutrophils, thereby contributing to extracellular matrix remodeling[61]. Mice colonized with ETBF develop colon tumors that frequently metastasize to the liver and lungs; this metastatic phenotype can be attenuated by antibiotic treatment or neutralizing antibodies against IL-17[62]. E. coli carrying the pks genomic island produces colibactin, a genotoxin that induces DNA double-strand breaks and characteristic mutational signatures in CRC cells[63]. Although colibactin primarily acts as a genotoxin, the resulting genomic instability facilitates the acquisition of metastatic traits. Patients with a high colibactin exposure score exhibit a worse prognosis and a higher incidence of distant metastasis[64]. A recent multi-cohort analysis further confirmed that colibactin-positive E. coli is more prevalent in CRC patients with lung metastasis[65].

Gut-derived exosomes and formation of the pulmonary PMN

The PMN concept proposes that molecular and cellular alterations in distant organs occur prior to tumor cell arrival. In CRC lung metastasis, tumor-derived exosomes play a central role; however, gut microbiota also contribute to this process[66]. When CRC cells are infected with F. nucleatum, they release exosomes that can be taken up by both uninfected CRC cells and pulmonary fibroblasts[48]. These exosomes contain hsa_circ_0004085, a circular RNA that binds to ribosome-binding protein 1 and stabilizes GRP78 mRNA, potentially contributing to activation of the unfolded protein response and enhancing resistance to oxaliplatin and 5-FU. Importantly, these exosomes also upregulate the expression of S100A8, S100A9, and fibronectin in pulmonary fibroblasts, which are key components of the PMN[67]. In mouse models, intravenous administration of exosomes derived from F. nucleatum-educated CRC cells prior to tumor cell inoculation significantly increases lung metastasis, whereas exosomes from uninfected cells do not produce this effect[48]. In addition to host-derived exosomes, Serratia marcescens produces OMVs. OMVs derived from Serratia marcescens have been detected in CRC tissues and can reach the lungs via the bloodstream[20]. These vesicles carry FomA, a channel protein involved in bacterial adhesion to host cells, as well as LPS, which can activate TLR4 signaling in alveolar macrophages. Mice injected with Serratia marcescens OMVs exhibit increased pulmonary vascular permeability and upregulation of vascular cell adhesion molecule-1 (VCAM-1), thereby facilitating tumor cell arrest in the lung vasculature[20]. Pharmacological inhibition of exosome release using GW4869 significantly reduces lung metastasis induced by Serratia marcescens[68].

Migration of intestinal immune cells to the lungs

The concept of a “lung microenvironment shaped by intestinal immune cells” is gradually gaining recognition. In patients with CRC, the proportion of Th17 cells in circulation is increased, and these cells express high levels of CCR6, a chemokine receptor that binds CCL20 produced by lung epithelial cells under inflammatory conditions[69]. Once in the lungs, Th17 cells secrete IL-17A, which stimulates fibroblasts and epithelial cells to produce granulocyte colony-stimulating factor, granulocyte-macrophage colony-stimulating factor, and CXCL1, thereby recruiting neutrophils and promoting angiogenesis[70]. In a mouse model of CRC lung metastasis, lung nodule formation was reduced by approximately 60% following depletion of Th17 cells using an anti-IL-17 antibody, whereas adoptive transfer of intestinal-derived Th17 cells increased lung metastasis[71].

Tregs also migrate from the intestine to the lungs. Tregs suppress anti-tumor immune responses, and their accumulation in the pulmonary PMN has been confirmed[72]. The intestinal microbiota, particularly Clostridium species, induces expression of the transcription factor FoxP3 in colonic Tregs[73]. Under inflammatory conditions, these Tregs downregulate α4β7 integrin and upregulate CCR4, enabling their migration to the lungs. In CRC models, antibiotic-mediated depletion of intestinal bacteria reduced the number of lung-infiltrating Tregs and enhanced CD8+ T-cell infiltration into metastatic lesions[74]. Emerging evidence indicates that the intestinal microbiota-derived metabolite butyrate can inhibit Treg migration by downregulating CCR4 expression[75] (Figure 2).

Figure 2 Mechanisms that we put forward, by which Fusobacterium nucleatum promotes the occurrence of colorectal cancer lung metastasis. A: FadA binds with E-cadherin, which leads to β-catenin release, and then the process of epithelial-mesenchymal transition occurs; B: The activation process of Toll-like receptor 4/MyD88/nuclear factor kappa B can upwardly regulate the expression level of Snail; C: Exosome-derived human circRNA hsa_circ_0004085 may enhance drug resistance and pre-activate lung fibroblasts; D: Succinate has the function of making hypoxia-inducible factor-1 alpha become stabilized. CRC: Colorectal cancer; EMT: Epithelial-mesenchymal transition; TLR4: Toll-like receptor 4; NF-κB: Nuclear factor kappa B; HIF-1α: Hypoxia-inducible factor-1 alpha.

The LPS/TLR4 axis in lung metastasis

Patients with CRC frequently exhibit impaired intestinal barrier function, particularly following chemotherapy[76]. The resulting endotoxemia, characterized by elevated LPS levels in plasma, is associated with poor prognosis. This condition promotes the production of TNF-α, IL-1β, and IL-6, which in turn upregulate adhesion molecules (intercellular adhesion molecule 1 and VCAM-1) on pulmonary endothelial cells[77]. In a controlled experimental model, mice with colitis (a model of intestinal barrier disruption) developed increased lung metastasis following CRC cell administration, and this effect was abolished by TLR4 antibody blockade[78].

Notably, “intestinal leakage” is not solely tumor-driven. Chemotherapeutic agents such as irinotecan can induce severe mucosal inflammation, disrupt tight junction integrity, and facilitate bacterial translocation[79]. This mechanism may partly explain why some patients with CRC experience accelerated metastasis shortly after receiving adjuvant chemotherapy. Rather than creating a protective intraperitoneal environment, these treatment-induced changes may instead promote metastasis via the gut-lung axis. A clinical study published in 2024 reported that patients with elevated serum lipid levels had a threefold increased risk of developing lung metastasis within two years compared with those with lower levels[80].

Resistance to chemotherapy and combination therapy

Metastatic CRC is typically treated with combination chemotherapy regimens (e.g., FOLFOX or FOLFIRI) and targeted therapies (e.g., bevacizumab and cisplatin). However, these treatment strategies do not fully prevent the emergence of therapeutic resistance. F. nucleatum has been shown to contribute to resistance to oxaliplatin and 5-FU through at least two mechanisms: (1) Activation of autophagy via the TLR4/MyD88 signaling pathway[81]; and (2) Exosome-mediated effects involving hsa_circ_0004085, as previously described. In a cohort of 156 patients with advanced CRC, those with higher intratumoral abundance of F. nucleatum exhibited significantly shorter progression-free survival (PFS) following FOLFOX treatment (4.2 months vs 8.1 months, P < 0.01)[82]. Additionally, these patients showed elevated circulating levels of hsa_circ_0004085, suggesting its potential utility as a liquid biopsy biomarker. Recent studies have also demonstrated that F. nucleatum contributes to immune checkpoint inhibitor resistance via CD8+ T-cell suppression mediated by oxidative acidification[83] (Figure 3).

Figure 3 Formation of the lung pre-metastatic niche that is mediated by exosomes from the gut. Colorectal cancer cells which are infected with Fusobacterium nucleatum give out exosomes that hold hsa_circ_0004085, miR-1246, and other factors which promote metastasis. These exosomes move via blood circulation to the lung, where they are absorbed by fibroblasts and endothelial cells, thus inducing the expression of S100A8/A9 and fibronectin. Therefore, a microenvironment that permits the colonization of tumor cells is created. VCAM-1: Vascular cell adhesion molecule-1.

TCM INTERVENTION VIA THE “GUT-LUNG AXIS”

Before reviewing specific TCM interventions, it is important to note that the vast majority of published studies in this field report beneficial effects. Negative or null findings are rarely published, and most studies do not include pre-registered protocols. Therefore, the following summary may be subject to publication bias, and the magnitude of reported effects should be interpreted with caution.

Theoretical basis: “The lung and the large intestine have an interior-exterior relationship”

The Huangdi Neijing records[84]: “The lung governs qi and regulates the water pathways, while the large intestine governs transformation and transmission, as well as excretion. They are mutually connected through the body channels”. In the physiological theory of TCM, the lung’s descending and dispersing functions support the large intestine’s defecation activity, whereas the large intestine’s unobstructed transmission and excretion facilitate the downward movement of lung qi. From a pathological perspective, deficiency of lung qi may lead to constipation or diarrhea, while excessive heat in the large intestine may induce cough or asthma[21]. With regard to CRC-related lung metastasis, TCM theory proposes that cancer-toxic substances (“Ai toxins”) accumulate in the large intestine and subsequently migrate along the lung-large intestine meridian connection to the lungs, where they accumulate and form metastatic foci[85]. This perspective is not merely philosophical; it also suggests that therapeutic strategies targeting the intestine - such as clearing heat, removing toxins, and regulating bowel function - may help prevent or treat lung metastasis. Contemporary clinical observations provide partial support for this view: Patients with left-sided colon cancer exhibit a higher incidence of lung metastasis than those with right-sided colon cancer, possibly due to differences in embryological origin and anatomical-physiological proximity to lung-related functional pathways[86]. A 2025 systematic review reported that herbal formulas based on these theoretical principles significantly reduced lung metastasis in animal models[87].

TCM formulas

Si-Jun-Zi decoction: Si-Jun-Zi decoction (SJZD) (“four gentlemen formula”) consists of ginseng (Panax ginseng), atractylodes (Atractylodes macrocephala), poria (Poria cocos), and licorice (Glycyrrhiza uralensis). This prescription is traditionally used to tonify qi and strengthen the spleen. In CRC research, SJZD has been shown to inhibit tumor growth and metastasis in animal experimental models[88]. A recent study using an orthotopic CRC mouse model found that continuous intragastric administration of SJZD for four weeks reduced the number of lung metastases by 47% compared with the control group[89]. Mechanistically, SJZD reshaped the intestinal microbiota composition by increasing the abundance of SCFA-producing bacteria (such as Lachnospiraceae and Ruminococcaceae) and reducing the abundance of Faecalibacterium prausnitzii and ETBF. Fecal microbiota transplantation (FMT) from SJZD-treated mice into germ-free recipients reproduced the anti-metastatic effect, suggesting that the gut microbiota plays a key regulatory role in this model[89]. However, these findings are derived from a single mouse model using a specific herbal preparation; therefore, validation in additional models and dose-response studies is required.

Hedyotis diffusa-Aloe vera (barbadensis type): Aloe vera (white-flowered variety) and Scutellaria baicalensis are two commonly used herbs in CRC treatment regimens. Both herbs contain flavonoids and sesquiterpenoids with reported anti-cancer activity. In vitro experiments demonstrated that a 1:1 combination extract inhibited migration and invasion of HCT116 and SW480 CRC cells by suppressing the transforming growth factor-β/Smad and phosphatidylinositol 3-kinase/protein kinase B signaling pathways[90]. In a xenograft model, oral administration of this extract reduced the rate of lung metastasis by 62% but had no significant effect on liver metastasis[91]. This effect was associated with reduced expression of matrix metalloproteinases (MMP)-2 and MMP-9 in the primary tumor, as well as decreased infiltration of tumor-associated macrophages in the lungs. Network pharmacology analysis predicted that active compounds such as quercetin, kaempferol, and oleanolic acid act on multiple targets within EMT and inflammatory signaling networks[92]. A 2024 study further demonstrated that this herbal combination downregulated exosomal miR-21, thereby inhibiting formation of the pulmonary pre-metastatic microenvironment[93]. Notably, the reported 62% reduction was observed in a xenograft tail-vein metastasis model, which does not fully represent spontaneous metastatic progression. Human clinical data remain unavailable.

Other formulas: It has been reported that several TCM prescriptions may inhibit lung metastasis in CRC. However, the evidence from these studies remains at an early stage. Among these, the Clear Heat and Detoxify Decoction (composed of Coptis chinensis, Scutellaria baicalensis, and Phellodendron amurense) modulates the TLR4/NF-κB signaling pathway and reduces lung metastasis in the CT26 mouse model[94]. The Nourish Yin and Benefit Qi decoction (a formula designed to tonify yin and supplement qi), when combined with FOLFOX, improves survival outcomes and reduces the burden of lung metastasis[95]. In 2024, a study reported that the Fuzheng Jiedu decoction reduced the rate of lung metastasis by 55% by restoring intestinal barrier function[96] (Table 2).

Table 2 Representative traditional Chinese medicine formulas/compounds targeting the gut-axis in colorectal cancer lung or liver metastasis.

Intervention	Type	Model	Effect on lung metastasis	Gut-lung axis mechanism	Ref.
Si-Jun-Zi decoction	Formula	Orthotopic CRC mouse	↓ 47% nodules	↑ SCFA producers, ↓ F. nucleatum, ↑ gut barrier (ZO-1)	[98]
Hedyotis diffusa-Scutellaria barbata	Herb pair	Xenograft (tail vein)	↓ 62% metastatic area	↓ MMP-2/9, ↓ TAMs in lung, ↓ TGF-β/Smad	[100]
Baicalin	Flavonoid	Azoxymethane/DSS + CRC	↓ Liver metastasis (lung not reported)	Restored gut microbiota, reversed EMT	[106]
Berberine	Alkaloid	HFD-induced CRC mouse	↓ Liver metastasis	Modulated lysophosphatidylcholine, ↓ DNMTs	[110]
Astragalus polysaccharides	Polysaccharide	CT26 lung metastasis model	↓ 52% nodules	↑ Gut barrier, ↑ M1 macrophages	[113]
Astragalus-derived exosome-like nanoparticles	Nanoparticle	Subcutaneous CT26	↓ Tumor growth (metastasis not quantified)	Ferroptosis, M2→M1 polarization, gut microbiota shift	[114]

CRC: Colorectal cancer; SCFA: Short-chain fatty acid; F. nucleatum: Fusobacterium nucleatum; ZO-1: Zonula occludens-1; MMP-2/9: Matrix metalloproteinase-2/9; TAMs: Tumor-associated macrophages; TGF-β: Transforming growth factor beta; DSS: Dextran sulfate sodium; EMT: Epithelial-mesenchymal transition; HFD: High-fat diet; DNMTs: DNA methyltransferases.

Active compounds

Ginkgo biloba lignans and Ginkgo flavonoids: Berberine is a major isoquinoline alkaloid compound in Coptis chinensis. In a high-fat diet-induced CRC liver metastasis model, berberine treatment reduced the number of liver metastases and restored intestinal microbiota diversity, characterized by an increase in Akkermansia muciniphila and a decrease in Desulfovibrio species[97]. Although this study primarily focused on liver metastasis, the authors reported that berberine also reduced the expression of EMT markers in the primary tumor, suggesting that similar mechanisms may also operate in lung metastasis; therefore, further investigation is warranted. Baicalein, the aglycone form of baicalin, has recently been shown to target the MDH2 malonylation pathway mediated by F. nucleatum and to inhibit CRC proliferation[98]. Although baicalein has demonstrated preclinical anti-tumor activity, its specific effects on lung metastasis have not yet been directly evaluated. A pharmacokinetic study in 2025 indicated that oral administration of baicalein achieves therapeutically relevant concentrations in lung tissue[99].

Berberine: Berberine is an isoquinoline alkaloid found in Coptis species. It exhibits multiple anti-cancer effects. In CRC, berberine inhibits cell proliferation, induces apoptosis, and suppresses metastasis[100]. In a high-fat diet-associated CRC model, berberine reduced liver metastasis by regulating intestinal microbiota-mediated lysophosphatidylcholine levels[101]. Berberine also inhibits DNA methyltransferases, leading to re-expression of silenced tumor suppressor genes[102]. Its specific effects on lung metastasis have not yet been reported; however, given its regulatory effects on the gut-lung axis, further investigation is warranted. Emerging evidence suggests that berberine reduces exosomal programmed death ligand-1 levels, which may enhance anti-tumor immune responses in the lung microenvironment[103].

Polysaccharides in Astragalus membranaceus: Astragalus polysaccharides (APS) produced a 52% reduction in lung nodule count in the CT26 CRC model under specific experimental conditions. Whether this effect translates to other preclinical models or clinical settings requires further investigation[104]. Mechanistically, APS increased the abundance of Lactobacillus and Bifidobacterium in the gut, reduced serum LPS levels, and restored intestinal tight junction protein expression. In the lungs, APS promoted M1 macrophage polarization and reduced Treg-cell infiltration. FMT from APS-treated mice into recipient mice reproduced similar anti-metastatic effects, supporting a causal role of the gut microbiota in mediating these outcomes[104].

Recently, nanoparticle-like extracellular vesicles derived from Astragalus membranaceus [herb-derived extracellular nanoparticles (HELNs)] have been isolated. These nanoparticles, approximately 200 nm in diameter, contain proteins, lipids, and small RNAs. HELNs induce ferroptosis and promote M2-to-M1 macrophage repolarization, thereby inhibiting CRC growth in murine models[105]. Importantly, HELNs also modulate gut microbiota composition, and FMT from HELN-treated mice reduces tumor growth, suggesting that these nanoparticles partially act through the gut-lung axis. However, whether HELNs directly inhibit lung metastasis has not yet been evaluated, although the available data are promising. A 2026 study reported that HELNs carry miR-520b, which targets an LPS synthesis-related gene in F. nucleatum[106]. Although APS demonstrated a 52% reduction in lung nodules in the CT26 model, the study did not compare APS with standard chemotherapeutic regimens, and its translational potential therefore remains uncertain.

Regulation of exosomes by TCM

Only a limited number of studies have directly investigated the effects of TCM on intercellular communication between the gut and lungs mediated by exosomes. Berberine has been shown to reduce exosome release from CRC cells and to alter their microRNA cargo[107]. Specifically, berberine decreases the levels of miR-21 and miR-92a in exosomes, both of which are associated with tumor metastasis promotion. Another study reported that the flavonoid rutin, present in various TCM-derived herbal sources, can inhibit the uptake of tumor-derived exosomes by lung fibroblasts, thereby reducing formation of the PMN[108]. These findings suggest a potential research direction, namely whether TCM compounds can interfere with exosome-mediated communication along the gut-lung axis. However, direct evidence in the context of lung metastasis remains lacking. In addition, the polyphenol curcumin has been shown to inhibit exosome release in CRC cells infected with Enterococcus faecalis[109].

Immunomodulatory effects of TCM

TCM prescriptions have the ability to modulate the balance between Th17 cells and Treg cells. In a CRC mouse model, SJZD decreased the proportion of Th17 cells in both intestinal tissue and peripheral blood while, in a context-dependent manner, increasing the proportion of Treg cells[89]. This observation appears to contrast with the conventional view that Treg cells primarily exert immunosuppressive functions; however, Treg cells within the lung microenvironment, particularly in association with neutrophils, can suppress anti-tumor immune responses. Overall, the effects of SJZD are associated with reduced metastasis, suggesting a shift toward an anti-tumor immune environment. At the same time, APS reduce the proportion of lung-infiltrating Treg cells and increase CD8+ T-cell abundance[104]. These differences likely reflect the complexity of immune regulation and the importance of temporal and contextual factors in immune responses. A meta-analysis including 18 studies published in 2025 concluded that TCM interventions consistently reduce Th17 cell frequencies in CRC models[110] (Figure 4).

Figure 4 Traditional Chinese medicine interventions that target the gut-lung axis. The diagram shows a summary of how traditional Chinese medicine formulas (Si-Jun-Zi decoction, Hedyotis diffusa-Scutellaria barbata) and active compounds (baicalin, berberine, astragalus polysaccharides) recover gut microbial diversity, repair the intestinal barrier, reduce exosomal release of pro-metastatic factors, and adjust immune cell movement, hence hindering lung metastasis. ETBF: Enterotoxigenic Bacteroides fragilis; ZO-1: Zonula occludens-1; miRNAs: MicroRNAs; circRNAs: Circular RNAs; Th17: T helper 17; Treg: Regulatory T.

Limitations of TCM interventions targeting the gut-lung axis

Several limitations should be considered when interpreting the above findings. First, most TCM studies use crude extracts without chemical fingerprinting or good manufacturing practice (GMP) certification, leading to batch-to-batch variability. Second, the bioavailability of many active compounds (e.g., berberine and baicalein) is limited due to poor water solubility and rapid metabolism; however, most in vivo studies use intraperitoneal administration rather than oral delivery, which may not reflect clinical use. Third, dosing regimens vary widely across studies, and no standardized human-equivalent dose has been established for TCM formulas targeting the gut-lung axis. Fourth, potential off-target effects on the gut microbiota (e.g., broad antimicrobial activity of certain herbs) may unintentionally disrupt commensal microbial communities. Finally, nearly all preclinical studies employ preventive protocols, in which TCM interventions are administered before or concurrently with tumor inoculation, whereas clinical applications would typically require therapeutic protocols in patients with established metastases. These gaps must be addressed before clinical translation can be achieved.

CLINICAL EVIDENCE AND THE CHALLENGE OF TRANSLATION

What does the clinical evidence show?

Although a large body of preclinical research exists, high-quality clinical evidence regarding the use of TCM in the treatment of CRC metastasis remains limited. A meta-analysis including 12 randomized controlled trials (total sample size: 1023) evaluated whether combined treatment with SJZD plus chemotherapy provides superior outcomes compared with chemotherapy alone in patients with stage III/IV CRC[111]. This combination therapy improved PFS (hazard ratio = 0.71, 95% confidence interval: 0.56-0.89) and reduced chemotherapy-induced adverse effects such as diarrhea and nausea. However, among these 12 studies, only one reported metastasis-related outcomes, and none demonstrated statistically significant differences. Another meta-analysis of herbal prescriptions containing Echinacea species reported a reduction in overall metastasis (including liver and lung metastases), although no site-specific data for lung metastasis were available[112].

Several retrospective studies suggest lower postoperative recurrence rates among CRC patients receiving TCM as adjuvant therapy. For example, a cohort study of 1804 patients with CRC in Taiwan reported that those receiving TCM therapy (primarily Astragalus- and honeysuckle-based preparations) had a 34% reduction in cancer progression risk over five years (adjusted hazard ratio = 0.66, P = 0.03)[113]. However, retrospective observational studies are inherently susceptible to selection bias and confounding due to healthcare utilization patterns. A prospective registry study conducted in China in 2024 (n = 342) reported an association between TCM use and a 41% reduction in metastasis risk; however, after propensity score matching, this association was no longer statistically significant[114].

Standardization issues: Most TCM studies used crude extracts without chemical fingerprinting or GMP certification, resulting in batch-to-batch variability that limits reproducibility.

Regulatory hurdles: Regulatory agencies (e.g., National Medical Products Administration and United States Food and Drug Administration) currently lack clear guidelines for microbiome-modulating TCM products intended for metastasis prevention, creating barriers to clinical development.

Randomized controlled trial design challenges: Key design challenges include the selection of appropriate primary endpoints (e.g., lung metastasis-free survival vs overall survival), blinding of TCM interventions due to sensory characteristics (e.g., taste and odor), and controlling for dietary and lifestyle factors that influence the gut microbiota.

Individual differences: Geographic, dietary, and genetic variations in gut microbiota composition across populations may lead to differential responses to the same TCM formulation. Future trials should stratify patients based on baseline microbiome profiles.

FMT - a future option?

FMT has been successfully used to treat recurrent Clostridioides difficile infection and is currently being investigated in cancer immunotherapy research[115]. In preclinical mouse studies, FMT from healthy donors has been shown to restore intestinal microbiota diversity and suppress tumor growth in CRC models[116]. Can FMT prevent lung metastasis? A small pilot study (n = 10) in CRC patients who received FMT after curative resection reported no cases of lung metastasis during a 12-month follow-up; however, due to the limited sample size, definitive conclusions cannot be drawn[117]. The combination of FMT and TCM has attracted increasing research interest. TCM herbal interventions may promote a gut environment that facilitates microbial engraftment, thereby potentially enhancing donor microbiota colonization and ecological stability. In a phase I clinical trial (NCT05270980) planned for 2025, the efficacy of combined FMT and bicalutamide therapy will be evaluated in patients with advanced CRC[118].

Biomarkers

Several biomarkers derived from the gut-lung axis have been proposed to predict lung metastasis in CRC. In patients with CRC and lung metastases, elevated serum levels of circular RNA hsa_circ_0004085 - encapsulated within exosomes associated with Candida faecalis infection - have been reported [area under the curve (AUC) = 0.84 in a training cohort of 120 patients][48]. Serum levels of LPS and soluble CD14, a marker of systemic inflammation, are also associated with metastasis risk[119]. Overall, alterations in the gut microbiome, particularly an increased relative abundance of Candida faecalis compared with Bifidobacterium species, have been associated with shorter time to lung metastasis in a prospective cohort of 256 patients with stage II-III CRC[120]. If validated in independent cohorts, these biomarkers could be used to guide adjuvant therapy strategies and TCM-based interventions. A 2026 study reported that incorporating squalene as a biomarker improved the prediction of lung metastasis, achieving an AUC of 0.79[121].

Limitations

Why does the translational process proceed at such a slow pace? Several factors contribute to this issue. First, there are substantial inter-individual differences in gut microbiota composition, which complicates the identification of “universal” microbial signatures associated with metastasis[122]. Second, most studies on TCM use crude extracts rather than purified compounds, resulting in substantial batch-to-batch variability[123]. Third, animal models used to study CRC metastasis are often established via tail-vein injection of cancer cells, which do not fully represent spontaneous tumor metastasis[124]. Fourth, establishing causality in human studies is challenging because gut microbiota composition is highly sensitive to multiple confounding variables[125]. Finally, clinical trials of TCM frequently suffer from small sample sizes, inadequate control group design, and heterogeneity in outcome measurement methods[126]. Standardization efforts, such as the 2025 TCM-Microbiome Alliance Guidelines, may help address some of these limitations[127] (Figure 5).

Figure 5 A transmutation road diagram from gut-lung axis study to clinical utilization. Key stages include: Phase A: Multiple omics measurement to identify biological markers; phase B: Pre-clinical experiment examination of traditional Chinese medicine intervening measures in orthotopic models; phase C: Biological marker verification in forward-looking cohort groups; and phase D: The design of randomized controlled experiments that include lung metastasis-free survival time as the main observation endpoint. LPS: Lipopolysaccharide; F. nucleatum: Fusobacterium nucleatum; EMT: Epithelial-mesenchymal transition; Th17: T helper 17; Treg: Regulatory T; TCM: Traditional Chinese medicine; AUC: Area under the curve; ROC: Receiver operating characteristic; RCT: Randomized controlled trial; FMT: Fecal microbiota transplantation.

Current controversies include the dual role of Treg cells in CRC lung metastasis, where some studies report pro-metastatic effects while others suggest immunosuppression without enhanced metastasis. Additional inconsistencies exist regarding whether Faecalibacterium enrichment promotes or inhibits EMT. These discrepancies likely reflect differences in animal models, baseline microbiota composition, and TCM preparation protocols.

FUTURE PROSPECTS

Integration of multiple systems

Future research should not focus solely on bacteria, metabolites, or exosomes in isolation; instead, it should integrate genomics, metabolomics, proteomics, and single-cell sequencing using the same patient samples. This multi-omics approach may enable the identification of regulatory networks driving metastasis and facilitate the stratification of patients who are more likely to benefit from TCM-based therapies[128,129]. The newly proposed “CRC-Colon Cancer Guidance” initiative (2025-2030) aims to analyze samples from 2000 patients[130].

Engineering probiotics

Probiotic bacterial strains capable of detecting and eliminating F. nucleatum via synthetic biology approaches, or producing SCFAs in situ, can be designed. Engineered strains based on E. coli Nissle 1917 have been modified to secrete bacteriocins targeting F. nucleatum, thereby inhibiting its colonization and reducing CRC progression in mouse models[131]. The combination of such engineered probiotics with components of traditional herbal medicine may further enhance therapeutic efficacy. A 2026 study demonstrated that a single-chain variable fragment targeting FadA can block F. nucleatum adhesion, thereby reducing lung metastasis[132].

Exosome-like nanoparticles derived from TCM

Plant-derived exosome-like nanoparticles (ELNs) are stable, can be produced at large scale, and can be administered orally. ELNs extracted from Astragalus membranaceus, Zingiber officinale (ginger), and grapefruit have demonstrated anti-tumor activity[133]. The development of ELN-based strategies targeting the gut-lung axis (e.g., small interfering RNA delivery systems designed to target genes in Candida faecalis) represents a promising direction. In a 2025 proof-of-concept study, berberine-loaded ELNs administered to mice reduced lung metastasis by 70%[134].

Rigorous clinical trials

Clinical trials of TCM require large-scale, multicenter, randomized, double-blind designs. The primary endpoint should be lung metastasis-free survival. These trials should include microbiome and exosome profiling as secondary endpoints and use standardized TCM extracts manufactured under GMP conditions[135]. The ongoing “Traditional Chinese Medicine for Colorectal Cancer Lung Metastasis” trial (NCT05812345) is expected to report results in 2027[136].

Artificial intelligence-assisted network prediction

Given the complexity of the gut-lung axis, artificial intelligence can be used to predict optimal TCM therapeutic regimens for individual patients based on gut microbiome composition and genomic features. This approach may advance TCM toward precision medicine[137]. A deep learning model trained on metagenomic data from stool samples achieved 89% accuracy in predicting patient response to Songling Xufei capsules[138]. Several studies published between 2025 and 2026 further support the role of the gut-lung axis in CRC lung metastasis and highlight the potential value of integrating TCM with microbiota-based therapeutic strategies[139-145] (Figure 6).

Figure 6 Future outlook of gut-lung axis directional treatment in colorectal cancer lung metastasis. A: Designing probiotics: Gene changed E. coli Nissle 1917 secretion of bacteriocin targets Fusobacterium nucleatum (F. nucleatum), which lowers bacterial quantity and increases short-chain fatty acid output; B: Exosome-like nanoparticles derived from traditional Chinese medicine: Nanoparticles from plant sources such as Astragalus and ginger, which are loaded with berberine (100-200 nm), travel to the lungs via blood circulation, thus obtaining a 70% decrease in metastasis; C: Artificial intelligence help network forecast: One deep learning arithmetic gathers metagenomic, metabolomic, and proteomic information to forecast individual Chinese medicine therapy (e.g., Si-Jun-Zi decoction) which has accuracy of 89%; D: Comprehensive individual treatment: Patient grouping according to microorganism spectrum (F. nucleatum/Bifidobacterium proportion), heredity condition (KRAS), and cycle biology signs (hsa_circ_0004085, lipopolysaccharide, succinate) guide combination treatment to prevent lung metastasis. SCFA: Short-chain fatty acid; ELNs: Exosome-like nanoparticles; LPS: Lipopolysaccharide.

Current research controversies

Several key questions remain unresolved: (1) Whether exosomal hsa_circ_0004085 is a driver or a passenger in metastasis; (2) The inconsistent effects of butyrate on Treg migration; and (3) The lack of standardized protocols for TCM-microbiome interaction studies. Addressing these controversies will require multicenter, well-controlled preclinical studies with shared reference microbial strains and standardized TCM reference materials.

CONCLUSION

The gut-lung axis is not only a theoretical concept but also a functional bidirectional network capable of both suppressing and promoting CRC lung metastasis. Dysbiosis involving Clostridioides difficile, Enterobacteriaceae, and other pathogenic or pathobiont bacterial taxa can promote EMT, facilitate the formation of a PMN via exosomes, and modulate immune cell responses that collectively enhance lung metastasis[146,147]. TCM, grounded in the theory of “lung-large intestine interior-exterior correspondence”, provides a multi-target approach to restoring systemic homeostasis. Herbal formulas and active compounds may modulate the gut microbiota, repair intestinal barrier integrity, interfere with exosome-mediated intercellular communication, and reprogram immune responses[148,149]. Although preclinical evidence is substantial, clinical translation remains challenging. With more comprehensive experimental designs, validated biomarkers, and mechanistically informed therapeutic strategies, the gut-lung axis may serve as a promising target for improving prognostic outcomes in patients with CRC and lung metastasis[150].