Advances in metabolic reprogramming of pulmonary fibrosis

Abstract

Pulmonary fibrosis (PF) refers to a common pathological change that occurs in a series of lung diseases. It is manifested by fibroblast proliferation and excessive extracellular matrix (ECM) deposition, resulting in destruction of lung structure and deterioration of lung function. PF represents a widespread global health problem, particularly in developing countries, and poses a substantial threat to public health. Despite its clinical significance, effective and specific pharmacological therapies remain unavailable. Therefore, identifying novel therapeutic targets is urgently needed and holds considerable clinical importance. A clearer understanding of the underlying pathogenic mechanisms of PF is essential for improving both prevention and treatment strategies. In recent years, increasing evidence has indicated that metabolic reprogramming including alterations in glucose, lipid, and amino acid metabolism plays an important role in the initiation and progression of PF. In this review, we summarize current knowledge regarding metabolic reprogramming in PF and highlight the potential of metabolic pathways as targets for diagnosis and therapy, thereby offering new perspectives for future research.

Introduction

Pulmonary fibrosis (PF), a progressive and irreversible pathological state of multiple lung diseases, is characterized by fibroblast proliferation and excessive extracellular matrix (ECM) deposition, subsequently leading to structural and functional destruction of the lungs and eventually respiratory failure and serious complications.^1,2 Patients with PF experience dry cough, sputum, and progressive dyspnea.³ There are various diseases that can cause PF. As the underlying causes are primarily pathologically present in the lung interstitium, these diseases are referred to as interstitial lung diseases (ILDs).⁴ ILDs can be divided into five broad clinical categories: (1) ILDs related to distinct primary diseases; (2) ILDs related to environmental exposures (including pneumoconiosis and hypersensitivity pneumonitis); (3) ILDs induced by drugs or irradiation; (4) ILDs associated with connective tissue diseases (CTDs); (5) idiopathic interstitial pneumonias (including idiopathic pulmonary fibrosis [IPF], idiopathic nonspecific interstitial pneumonia and others).⁵ According to the Global Burden of Disease (GBD) study, the global burden of ILDs has increased substantially from 1990 and is projected to continue rising through 2050.^6,7 IPF is the most prevalent and fatal form, accounting for >30% of ILDs and with a median life expectancy of 2–3 years after diagnosis if untreated.^8,9 The incidence of IPF is in the range of 0.09 to 1.30 per 10,000 persons globally and the prevalence is in the range of 0.33 to 4.51 per 10,000 persons.¹⁰

In PF, disease-related triggers can initiate a series of inflammatory and fibrotic responses that ultimately promote excessive ECM deposition and subsequent remodeling of the lung tissue. However, the precise pathological mechanisms that drive PF remain incompletely understood and are likely multifactorial.¹¹ Moreover, effective therapeutic options for PF remain limited. Current antifibrotic drugs, which are frequently associated with notable adverse effects, primarily slow disease progression rather than reverse established fibrosis.¹² Consequently, there is an urgent need to further investigate the molecular mechanisms underlying PF and to identify novel therapeutic strategies that can more effectively target this disease.

Metabolic reprogramming was first described in cancer cells in the 1920s. It refers to adaptive metabolic alterations that occur when cells respond to various stresses in order to satisfy bioenergetic demands, support biosynthetic processes, and maintain redox balance.¹³ Initially, this concept primarily described the shift between glycolysis and oxidative phosphorylation. However, the definition of metabolic reprogramming has since expanded to include broader metabolic adaptations, particularly changes in the metabolism of nutrients such as lipids and amino acids. In addition, metabolic reprogramming has also been found to be associated with other diseases beyond cancer, such as organ fibrosis including cardiac, lung, liver, kidney, and skin fibrosis.¹⁴

In recent years, increasing research has shown that metabolic reprogramming plays a potential role in the occurrence and progression of PF.15, 16, 17, 18, 19 In this review, we summarize advances in the metabolic reprogramming of PF, which could help us to understand the pathogenesis of PF and explore feasible therapies for this disease. Furthermore, the potential diagnostic markers and therapeutic targets that are aimed at metabolic reprogramming of PF are discussed.

Metabolic reprogramming in PF

An increasing number of studies have found dysregulation of the three major metabolic processes in PF, including the metabolism of glucose, lipids, and amino acids.20, 21, 22, 23, 24 In brief, enhanced glycolysis appears to represent a common metabolic feature of PF triggered by different types of stress.^25,26 In particular, the expression of glucose transporters (GLUTs) and several glycolysis-related enzymes increases, thereby supplying the energy required for fibrotic processes, including fibroblast proliferation and migration. In addition, PF arising from different etiologies shows comparable alterations in amino acid metabolism, most notably elevated levels of glycine, arginine, and proline, which serve as key substrates for collagen synthesis.^27,28 Interestingly, lipid metabolic profiles may vary across different forms and stages of PF, possibly reflecting disease-specific characteristics such as inflammation or fibrosis.29, 30, 31, 32 For example, alterations in acylcarnitine and ceramide metabolism are mainly observed in IPF, whereas changes in arachidonic acid (AA) metabolism are more prominent in silicosis.^33,34 Together, these findings suggest that metabolic reprogramming of the three nutrients is closely related to PF and collectively promotes the incidence and progression of PF.

Glucose metabolism

Glucose metabolism plays an integral role in maintaining bodily functions not only by supplying energy but also by generating various precursors for the biosynthesis of amino acids, lipids, and nucleotides. After being transported into the cytoplasm via GLUTs, glucose undergoes a ten-step metabolic process involving multiple enzymes and is ultimately converted into pyruvate. This process simultaneously generates adenosine triphosphate (ATP) and is referred to as glycolysis. Subsequently, pyruvate enters the tricarboxylic acid (TCA) cycle leading to a massive production of ATP under aerobic conditions while it is converted to lactate without any additional ATP generation under anaerobic conditions. Researchers have demonstrated that abnormal glucose metabolism, especially dysregulation of glycolysis, is closely associated with the development of PF,35, 36, 37, 38, 39, 40 which is mainly manifested by alterations in the expression of GLUTs and glycolytic enzymes such as hexokinase 2 (HK2), 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3), and lactate dehydrogenase (LDH) (Fig. 1, Fig. 2).

Fig. 1.

Glucose metabolism reprogramming in PF. Various stimuli, including TGF-β1, LPS, and silica, lead to reprogramming of glucose metabolism in patients with PF, which is mainly manifested as enhanced glycolysis. Specifically: (1) TGF-β1 upregulates GLUT1 by activating Smad/PI3K/MEK or Smad/PI3K/mTOR signaling pathways and phosphorylation of STAT3. This in turn activates HK2 via SRC and circHIPK3, which elevates FOXK2 expression via sponging miR-30a-3p, and this activates PFK1 and PKM2. TGF-β1 is also able to stimulate PFKFB3 by promoting PCBP3 or inhibiting LRRN3. (2) LPS primarily activates PFKFB3 via the mTOR pathway downstream of PI3K/Akt and AMPK. (3) Silica enhances expression of GLUT1, HK2, and LDHA, which depends on generation of IL-11 and subsequent activation of ERK/HIF-1α. Due to the enhancement of these glycolytic steps, lactate accumulation induces p300-mediated histone lactylation in the promoters of profibrotic genes. This can also be achieved via stimulation of YAP/TAZ by HK2 directly. AMPK, Adenosine monophosphate-activated protein kinase; Akt, Protein kinase B; α-SMA, α-smooth muscle actin; circHIPK3, Circular RNA homeodomain interacting protein kinase 3; Col1, Collagen 1; ERK, Extracellular regulated protein kinases; FN, Fibronectin; FOXK2, Forkhead box k2; F1,6BP, Fructose-1,6-bisphosphate; F6P, Fructose-6-phosphate; G6P, Glucose 6-phosphate; GLUT1, Glucose transporter 1; H3K18la, Histone h3 lysine 18 lactylation; HIF-1α, Hypoxia-inducible factor-1 alpha; HK2, Hexokinase 2; IL-11, Interleukin-11; LDHA, Lactate dehydrogenase a; LPS, Lipopolysaccharide; LRRN3, Leucine rich repeat neuronal 3; MEK, Mitogen-activated protein kinase kinase; miR-30a-3p, MicroRNA-30a-3p; mTOR, Mammalian target of rapamycin; p300, E1a binding protein p300; PCBP3, Poly(rC) binding protein 3; PEP, Phosphoenolpyruvate; PF: Pulmonary fibrosis; PFK1, Phosphofructokinase 1; PFKFB3, 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase 3; PI3K, Phosphatidylinositol 3-kinase; PKM2, Pyruvate kinase m2 isoform; SRC, Sarcoma proto-oncogene; Smad3, Small mothers against decapentaplegic homolog 3; STAT3, Signal transducer and activator of transcription 3; TGF-β1, Transforming growth factor-beta 1; TGFβR, Transforming growth factor-beta receptor; YAP/TAZ, Yes-associated protein/transcriptional co-activator with PDZ-binding motif.

Fig. 2.

Glucose metabolism reprogramming in different cells and the interactions between them. GLUT1 and glycolytic enzymes HK2, PFKFB3, PKM2, and LDHA are activated in macrophages of PF patients, resulting in enhanced glycolysis. This subsequently induces the M2 polarization phenotype, which is accompanied by cytokine release, histone lactylation, ER stress, and inflammasome activation. All cytokines induced by macrophages, including TGF-β, VEGFA, IL-1β, IL-18, and TNF-α, can activate fibroblasts and epithelial cells. Lactate is accumulated due to upregulated generation and inhibition of transportation toward the extracellular space in epithelial cells, activating EMT and NLRP3 inflammasomes. IL-1β can also be induced by epithelial cells to stimulate fibroblasts. Analogously, silica-induced tPA acts on fibroblasts through extracellular vesicles via AMPK/GLUT1 and miR-21 induced by NaAsO2, which activates HK2 and thus promotes proliferation, migration, and differentiation of fibroblasts. AIM2, Absent in melanoma 2; AMPK, Adenosine monophosphate-activated protein kinase; Akt, Protein kinase B; Akt2, Protein kinase B2; BLM, Bleomycin; caspase1, Cysteinyl aspartate specific proteinase 1; EMT, Epithelial–mesenchymal transition; ER: Endoplasmic reticulum; ERK, Extracellular regulated protein kinases; GLUT1, Glucose transporter 1; HK2, Hexokinase 2; IGF2BP1, Insulin-like growth factor 2 mRNA binding protein 1; IL-1β, Interleukin-1 beta; IL-18, Interleukin-18; JNK, C-Jun N-terminal kinase; LDHA, Lactate dehydrogenase A; LPS, Lipopolysaccharide; MCT1, Monocarboxylate transporter 1; MEK, Mitogen-activated protein kinase kinase; miR-21, MicroRNA-21; NLRP3, Nod-like receptor pyrin domain-containing 3; PDK1, 3-phosphoinositide-dependent protein kinase 1; PFKFB3, 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase 3; PKM2, Pyruvate kinase M2 isoform; PTEN, Phosphatase and tensin homolog; Smad2/3, Small mothers against decapentaplegic homolog 2/3; t-PA, Tissue-type plasminogen activator; TGF-β1, Transforming growth factor-beta 1; TGFBR: Transforming growth factor-β receptor; THBS1, Thrombospondin 1; TLR4, Toll-like receptor 4; TNF-α, Tumor necrosis factor-alpha; VEGFA, Vascular endothelial growth factor A.

GLUT1 is the most conserved and widely distributed glucose transporter in mammalian cells.⁴¹ The expression of GLUT1 was found to be increased in the lung tissues of patients with IPF and mice treated with bleomycin (BLM). Notably, GLUT1 expression exhibits significant cell-type-specific differences. Specifically, the level of GLUT1 was much higher in the human lung fibroblasts from IPF patients compared to the normal ones as well as human and murine fibroblasts stimulated with transforming growth factor-β1 (TGF-β1).^42,43 TGF-β1 could activate the suppressor of mother against decapentaplegic 2/3 (Smad2/3), which subsequently promoted activation of the mitogen-activated protein kinase kinase/extracellular signal-regulated kinase (MEK/ERK) and mammalian target of rapamycin complex 2 (mTORC2) pathways downstream of phosphoinositide 3-kinase (PI3K) and finally upregulated GLUT1.⁴² Additionally, inhibiting the phosphorylation of signal transducer and activator of transcription 3 (STAT3) could downregulate GLUT1, which alleviated IPF via suppressing the differentiation, migration, and proliferation of fibroblasts.⁴³ Furthermore, the expression of GLUT1 was significantly higher in acute exacerbation of IPF (AE-IPF).³⁶ Cho et al³⁶ demonstrated that superimposed Streptococcus pneumoniae (S. pneumoniae) infection induced AE-IPF through absent in melanoma 2 (AIM2) inflammasome activation depending on augmentation of GLUT1. Both AIM2 deficiency and macrophage-specific Glut1 knockout alleviated fibrosis in BLM-treated lung after S. pneumoniae infection.³⁶ PF is more prevalent in aging populations. Consistent with this, the protein expression of GLUT1 in whole lung tissues from aged mice was significantly increased compared with young mice, not only at baseline but also with BLM treatment.⁴⁴ Conversely, BLM led to a significant decrease in expression of GLUT1 and induced a dose-dependent suppression of glycolysis in alveolar epithelial cells, which led to diminished mitochondrial function contributing to the loss of epithelial repair.⁴⁵ Li et al⁴⁶ also found genetic loss of GLUT1 reduced the proliferation of alveolar type II epithelial cells (AECⅡs). In a word, elevated GLUT1 expression in pulmonary fibroblasts promotes their proliferation, migration and myofibroblast differentiation; upregulated GLUT1 in alveolar macrophages drives their pro-fibrotic activation; whereas GLUT1 is downregulated in AECⅡs, which impairs the regenerative capacity of the alveolar epithelium. All these cellular events collectively contribute to the development and progression of PF,¹¹ which resolves the apparent contradictory expression pattern.

After being transported into the cytoplasm by GLUTs, glucose is metabolized by a sequence of enzyme-mediated processes to function. Dysregulation of these glycolytic enzymes, especially HK2, PFKFB3, and lactate dehydrogenase A (LDHA), plays a crucial role in the development of PF and targeting these enzymes shows potential therapeutic effect.47, 48, 49 HK2 is the enzyme that catalyzes the first step to convert glucose into glucose-6-phosphate (G6P). Yin et al⁵⁰ found that the expression, abundance, and activity of HK2 were increased in TGF-β-stimulated fibroblasts. HK2 could promote the induction of profibrotic markers such as collagen Ι (Col Ι), fibronectin (FN), and α-smooth muscle actin (α-SMA), as well as the proliferation and migration of fibroblasts through activating the transcriptional regulators yes-associated protein/transcriptional coactivator with a PDZ-binding domain (YAP/TAZ). Furthermore, inhibition of HK2 exhibited an anti-fibrotic effect,⁵⁰ highlighting the importance of HK2 in the development of PF. Another study found that the 3’ untranslated region (3’UTR) of HK2 interacted with microRNA-29b-2-5p (miR-29b-2-5p) and increased miR-29b-2-5p could reduce the expression of HK2, affecting the proliferation, migration, and invasion of lung fibroblasts.⁵¹ Increased HK2 was also observed in alveolar macrophages from fibrotic lungs and administration of 2-deoxy-d-glucose (2-DG), a potent inhibitor of HK2, effectively reversed the profibrotic M2 phenotype of alveolar macrophages.⁵² PFKFB3 is able to enhance glycolysis in the cells significantly. PFKFB3 catalyzes the conversion of fructose-6-phosphate (F6P) into fructose-2,6-bisphosphate (F2,6BP) and F2,6BP is the strongest allosteric activator of phosphofructokinase 1 (PFK1), which catalyzes the generation of F1,6BP and serves as the most important enzyme of glycolysis. Increase of PFKFB3 at the protein level was detected in human and murine lung fibroblasts treated with TGF-β1 as well as lung tissues of BLM-treated mice.53, 54, 55 Mechanistically, TGF-β1 treatment facilitated fibroblasts undergoing DNA hypermethylation via Smad3 signaling and downregulated the expression of leucine-rich repeat neuronal protein 3 (LRRN3). Decreased LRRN3 enhanced PFKFB3 expression and glycolysis, which subsequently induced the differentiation, proliferation, and migration of fibroblasts. Administration of albendazole could reverse the downregulation of LRRN3 and ameliorate the established PF.⁵³ While at the mRNA level, it was controversial. In TGF-β1 (5 ng/mL)-treated NIH3T3 cells and BLM-treated mice lungs, the mRNA level of PFKFB3 was upregulated. And peimine inhibited glycolysis by repressing PI3K/protein kinase B (PKB or Akt)/PFKFB3 signaling, ultimately suppressing fibroblast activation and PF.⁵⁴ But the increased PFKFB3 mRNA level was not confirmed in primary mouse lung fibroblasts treated with 10 ng/mL TGF-β1. Further study revealed that poly(rC)-binding protein 3 (PCBP3) increased PFKFB3 expression by promoting its translation, activating fibroblasts. Anlotinib could target PCBP3 and then inhibit PFKFB3-driven glycolysis in fibrotic lungs, accelerating the resolution of inflammation.⁵⁵ In alveolar macrophages from mice treated with BLM, the expression of PFKFB3 was also upregulated, which promoted the profibrotic M2 phenotype of macrophages. 3-(3-pyridinyl)-1-(4-pyridinyl)-2-propen-1-one (3PO), a specific inhibitor of PFKFB3, attenuated the expression of profibrotic markers.⁵² Importantly, epigenetic changes in genes involved in glucose metabolism such as PFKFB3 were related to clinical features of IPF severity.⁵⁶ While in AECⅡs, activation of AMP-activated protein kinase (AMPK)-PFKFB2 signaling upregulated glycolysis, which was required for alveolar regeneration.⁵⁷ And interestingly, F1,6BP, an intermediate of the glycolytic pathway, was able to inhibit PF development and fibroblast proliferation.^58,59

Glucose is finally converted into pyruvate which could be further catalyzed by LDH to generate lactate and lactate has been reported to promote PF.⁶⁰ Expression of LDHA was found to be significantly higher in AECⅡs from IPF lungs compared to controls and TGF-β treated fibroblasts.^61,62 Accumulation of lactate could promote PF by stimulating epithelial-mesenchymal transition (EMT) of alveolar epithelial cells and trans-differentiation of lung fibroblasts.^62,63 A serum proteomics study showed the levels of serum LDHA and chaperonin-containing tailless complex polypeptide subunit 6A (CCT6A), which were associated with glucose metabolic reprogramming, were correlated with high serum lactate content in patients with IPF.⁶⁴ Similarly, the level of lactate in BLM-treated mice was markedly increased in the serum, while CCT6A attenuated the production of lactate via the ubiquitin-dependent degradation of hypoxia-inducible factor-1 alpha (HIF-1α), alleviating PF.⁶⁵ Hypoxia is a sentinel feature of IPF which enhances cellular lactate production in a HIF-1α dependent pathway.65, 66, 67, 68 HIF-1α binds to the lactate receptor G-protein-coupled receptor 81 (GPR81) promoter and increases the expression of GPR81; this leads to elevated lactate levels that augment the fibrogenicity of mesenchymal progenitor cells (MPCs) and drive fibrosis in the lungs of patients with IPF,⁶⁹ indicating the important crosstalk between hypoxia and lactate in PF. Inhibition of both HIF-1α and LDH5 or LDHA inhibited TGF-β-induced myofibroblast differentiation.70, 71, 72 However, another study found LDH5 inhibition had no influence on fibroblast transformation to myofibroblast, though it could prevent metabolic shifts in fibroblasts.⁷³ In addition, lactate accumulation could be induced by higher pyruvate dehydrogenase kinase 1 (PDK1) expression which inhibits pyruvate dehydrogenase (PDH) to convert pyruvate into acetyl-CoA.⁶³ Furthermore, lactate contributes to PF development via histone lactylation,^74,75 an important form of epigenetic modifications. TGF-β treatment increased HK2 and lactate in fibroblasts via SRC, which led to lactylation of histone H3 at lysine 18 (H3K18la) and finally promoted fibrosis.⁷⁶ Furthermore, Cui et al⁷⁴ found that increased lactate in the TGF-β1-treated fibroblasts could be secreted into the extracellular environment and then induced p300-mediated histone lactylation in the promoters of the profibrotic genes in macrophages, suggesting a potential interplay between metabolic reprogramming and epigenetic modifications in PF. Correspondingly, the N6-methyladenosine (m6A) reader insulin-like growth factor 2 mRNA-binding protein 1 (IGF2BP1) enhanced macrophage M2 polarization through upregulation of thrombospondin-1 (THBS1)-mediated glycolysis to promote PF.⁷⁷

In addition to IPF, the upregulated glycolysis was also involved in PF induced by environmental exposures, lipopolysaccharide (LPS), mechanical ventilation (MV), and ionizing radiation.^26,78, 79, 80, 81, 82 Lu et al⁸³ found the expression of GLUT1 was increased in both the mice lung tissues and alveolar macrophages after being exposed to silica. Moreover, the glycolysis inhibitor 2-DG reduced the level of GLUT1, further inhibited HIF-1α expression and finally reduced the expression of interleukin-1β (IL-1β), thereby alleviating inflammatory injury in silicosis. Similarly, Xie et al⁸⁴ and Li et al⁸⁵ observed increased mRNA and protein levels of GLUT1 in silica-induced fibrotic lung tissues and fibroblasts. Mechanistically, silica-induced IL-11 promoted the process of fibroblast to myofibroblast transition by augmenting glycolysis via the ERK-HIF-1α axis.⁸⁴ And silica-induced ferroptosis in airway epithelium led to heightened secretion of tissue-type plasminogen activator (tPA) and the secreted tPA subsequently induced GLUT1 accumulation via AMPK phosphorylation, thereby activating the fibroblasts.⁸⁵ Glycolytic enzymes including HK2, PFK, PDK1, LDHA, and pyruvate kinase M2 (PKM2) were also increased in silicosis.86, 87, 88, 89 In silica-induced PF, the expression of circHIPK3 was increased in fibroblasts and consequently enhanced forkhead box K2 (FOXK2) expression via sponging miR-30a-3p, which induced the augmented glycolysis and finally the activation and proliferation of fibroblasts.⁸⁷ While in macrophages, lactate concentration was increased, which subsequently promoted endoplasmic reticulum (ER) stress or NOD-like receptor protein 3 (NLRP3) activation via lactylation modifications, leading to macrophage polarization toward a proinflammatory phenotype and secretion of inflammatory cytokines.^88,89 Oxamate, an amide isostere of pyruvate, serves as a competitive inhibitor of LDHA and was found to attenuate the augmentation of glycolysis and ER stress for treating silicosis.⁸⁸ Besides, long-term exposure to PM_2.5 is closely associated with an increased risk of PF. Compared to healthy lungs, PM_2.5-exposed lungs also exhibited significantly elevated mRNA expression level of GLUT1 and glycolytic enzymes such as HK2, PKM2, pyruvate dehydrogenase kinase 2 (PDK2), and LDHA. These events increased lactate contents, which induced histone lactylation at promoters of pro-fibrotic genes in macrophages, together with the secretion of pro-fibrotic cytokines. And then these cytokines promoted the EMT through activating TGF-β/Smad2/3 and vascular endothelial growth factor A (VEGFA)/MEK/ERK pathways, leading to PF.⁷⁸ Exposure to arsenic can induce PF as well. Wang et al⁷⁹ found dysregulation of glycolysis in arsenic-induced PF, manifested by elevated HK2 and lactate. Further study revealed arsenic-treated epithelial cells transferred miR-21 into lung fibroblasts through extracellular vesicles (EVs), which promoted glucose uptake and lactate secretion of lung fibroblasts by regulating the phosphatase and tensin homolog (PTEN)/Akt pathway, leading to the differentiation of fibroblasts.

An increase in PFKFB3 was also detected in human lung fibroblasts treated with LPS and the lung tissues of LPS-induced PF mice via activation of mammalian target of rapamycin (mTOR).^90,91 Moreover, Xu et al⁸⁰ found that LPS promoted endogenous tumor necrosis factor alpha (TNF-α) secretion in pulmonary macrophages, which could facilitate lung fibroblast aerobic glycolysis through PFKFB3 activation, suggesting an intrinsic interaction between fibroblasts and macrophages in the pathogenesis of PF. LPS could also inhibit the expression of monocarboxylate transporter-1 (MCT1) in mouse alveolar epithelial cells and cause lactate transport disorder, which led to lactate accumulation, ultimately promoting the process of EMT.⁹² PF additionally occurs as a common complication during some therapies, such as thoracic radiotherapy and MV. Luo et al⁸² discovered radiation promoted increased glycolysis levels in alveolar macrophages, fibroblasts, and lung tissues. Furthermore, the pro-inflammatory and pro-fibrotic factors in alveolar macrophages were increased after radiation. Besides, the study found that radiation induced the myofibroblast-like transformation of fibroblast and epithelial cells. Radiation also activated NLRP3 inflammasome in lung epithelial cells, relying on glycolysis related reactive oxygen species (ROS) accumulation. Consequently, activation of NLRP3 inflammasome triggered fibroblast maturation and migration through IL‐1β.⁹³ MV is the most important supportive therapy for acute respiratory distress syndrome (ARDS) patients, while MV might induce PF and aggravate sepsis-associated pulmonary fibrosis, termed as MV-associated pulmonary fibrosis (MVPF) and sepsis and MV-associated pulmonary fibrosis (S-MVPF), respectively.^81,94 Mei et al^81,94 found the glycolytic markers including lactate, LDHA, and PKM2, were increased in MVPF and S-MVPF. The enhanced glycolysis was mediated by integrin β3–PKM2 pathway, which could become a therapeutic target for MVPF and S-MVPF. Table 1 lists the changes of metabolic enzymes and metabolites in glucose metabolism in PF.^35,36,43,44,49, 50, 51, 52, 53, 54, 55^{,62,63,65,71,72,75,76,78,82,83,86,88,91}

Table 1.

Changes of metabolic enzymes and metabolites in glucose metabolism in PF.

PF type	Changed substances	Location	Treatment	Mechanism	References
IPF	GLUT1↑	Lung/Macrophage	Macrophage knockout of GLUT1/AIM2	Decreases IL-1β and IL-18	36
		Lung/Fibroblast	Dihydromyricetin	Inhibits GLUT1 by STAT3/p-STAT3 signaling pathway and inhibits the proliferation, migration, and differentiation of fibroblast	43
		Lung/Fibroblast	Phloretin	Inhibits GLUT1 and the differentiation of fibroblast	44
	HK2↑	Fibroblast	2-DG/3-bromopyruvate	Inhibits HK2 and the differentiation of fibroblast	35
		Lung/Fibroblast	Alamandine	Inhibits HK2 and the differentiation of fibroblast	49
		Lung/Fibroblast	Lonidamine	Inhibits HK2 and the differentiation of fibroblast	50
		Fibroblast	Dahuang Zhechong Pill	Inhibits HK2 by miR-29b-2-5p and inhibits the proliferation and migration of fibroblast	51
		Macrophage	2-DG	Inhibits HK2 and the M2 polarization of macrophage	52
		Fibroblast	Jinshui Huanxian Formula (ECC-JHF)	Inhibits HK2 by SRC, and the lactylation in fibroblast	76
	PFKFB3↑	Fibroblast	3PO	Inhibits PFKFB3 and the differentiation of fibroblast	35,75
		Lung/Fibroblast	Alamandine	Inhibits PFKFB3 and the differentiation of fibroblast	49
		Macrophage	3PO	Inhibits PFKFB3 and the M2 polarization of macrophage	52
		Fibroblast	Albendazole	An agonist of LRRN3 inhibits PFKFB3 and the differentiation of fibroblast	53
	PFKFB3↑	Lung/Fibroblast	Peimine	Inhibits PFKFB3 by suppressing PI3K/Akt and inhibits the differentiation of fibroblast	54
		Fibroblast	Anlotinib	Inhibits PCBP3 to inhibit PFKFB3 and inhibits the differentiation of fibroblast	55
		Fibroblast	GLP-1R agonist	Activates the GLP-1R and inhibits PFKFB3 and the differentiation of fibroblast	75
	PDK1↑	Lung/Fibroblast/AEC	CCT128930/AZD7545	Inhibits PDK1 and the differentiation of fibroblast, as well as the EMT of AEC	63
	LDHA↑	Fibroblast	Dapper homolog 2 (DACT2)	Promotes LDHA degradation and inhibits the differentiation of fibroblast	62
		Lung/Fibroblast	Gossypol	Inhibits LDHA and the differentiation of fibroblast	71,72
	Lactate↑	AEC	CCT6A	Decreases lactate by inhibiting HIF-1α	65
PM2.5 induced PF	LDH↑	Lung/macrophage	GNE-140	Inhibits LDHA and the lactylation in macrophage	78
RIPF	GLUT1, HK2, PFKFB3↑	Lung/macrophage	IR-780	Inhibits glycolysis	82
Silicosis	HK2↑	Lung/macrophage	2-DG	Inhibits HK2 and the M2 polarization of macrophage	83
		Lung	Triiodothyronine	Inhibits HK2	86
	LDHA↑	Lung/macrophage	Oxamate	Inhibits LDHA to reduce ER stress in macrophage	88
LPS-induced PF	PFKFB3↑	Fibroblast	Metformin	Inhibits PFKFB3 by suppressing AMPK/mTOR and inhibits the differentiation of fibroblast	91

AEC, Alveolar epithelial cell; AIM2, Absent in melanoma 2; AMPK, AMP-activated protein kinase; CCT6A, Chaperonin-containing tailless complex polypeptide subunit 6A; EMT, Epithelial-mesenchymal transition; ER, Endoplasmic reticulum; GLP-1R, Glucagon-like peptide-1 receptor; GLUT1, Glucose transporter type 1; HIF-1α, Hypoxia-inducible factor-1 alpha; HK2, Hexokinase 2; IL, Interleukin; IPF, Idiopathic pulmonary fibrosis; LDHA, Lactate dehydrogenase A; LPS, Lipopolysaccharide; LRRN3, Leucine-rich repeat neuronal protein 3; miR, MicroRNA; mTOR, Mammalian target of rapamycin; PCBP3, Poly(rC)-binding protein 3; PDK1, Pyruvate dehydrogenase kinase 1; PF, Pulmonary fibrosis; PFKFB3, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3; PI3K, Phosphoinositide 3-kinase; PM_2.5, Fine particulate matter; p-STAT3, Phosphorylated signal transducer and activator of transcription 3; RIPF, Radiation-induced pulmonary fibrosis; STAT3, Signal transducer and activator of transcription 3; 2-DG, 2-deoxy-d-glucose; 3PO, 3-(3-pyridinyl)-1-(4-pyridinyl)-2-propen-1-one.

Together, in PF caused by various defined or unknown etiologies, the glycolytic pathway undergoes metabolic reprogramming, resulting in a shift in cellular energy metabolism from oxidative phosphorylation toward glycolysis. Increased expression of GLUT1 and several glycolysis-associated enzymes enables cells to produce lactate and generate sufficient energy to support their metabolic demands. In addition, dysregulated activation of glycolysis influences the physiological state of several pulmonary cell populations, including epithelial cells, fibroblasts, and macrophages.^63,77,95 Overall, diverse stimuli activate multiple signaling pathways that elevate the expression of GLUT1 and glycolysis-related enzymes in fibroblasts and macrophages. This metabolic shift enhances glycolytic activity, which promotes fibroblast proliferation, migration, and differentiation, and drives macrophage polarization toward the M2 phenotype. In contrast, AECⅡs exhibit a different pattern: GLUT1 expression tends to decrease, whereas glycolysis-associated enzymes such as LDHA show increased expression. This apparent inconsistency likely reflects their involvement in distinct physiological processes within AECⅡs. Specifically, reduced GLUT1 expression impairs the regenerative capacity of AECⅡs, while increased LDHA expression contributes to the EMT. Both impaired regeneration and enhanced EMT of AECⅡs are key drivers of PF initiation and progression. Taken together, reprogramming of glucose metabolism promotes pro-fibrotic activities in multiple cell types and thereby contributes to the development of PF. Importantly, alterations in cellular glucose metabolism influence not only the affected cells but also neighboring cells within the lung microenvironment. Glucose metabolic reprogramming in macrophages and AECⅡs modifies their functional states and stimulates the release of cytokines, which can directly regulate fibroblast proliferation, migration, and differentiation, or indirectly affect these processes by altering fibroblast glucose metabolism (Fig. 2). Therefore, targeting metabolic regulatory factors that contribute to dysfunction in these specific cell populations may provide promising therapeutic strategies for PF.

Lipid metabolism

Lipids are a heterogeneous group of hydrophobic molecules that consist of glycerides (including phospholipids), fatty acids, non-glyceride lipids (including cholesterol and sphingolipids), and complex lipids (lipoproteins).⁹⁶ Lipids are essential biological molecules for producing substantial amounts of ATP through the β-oxidation of fatty acids as well as for maintaining the structure and function of cell membranes. Additionally, lipids are important components of alveolar surfactant, which plays a crucial role in maintaining the normal alveolar surface tension.⁹⁷ Lipid metabolism, encompassing the synthesis, transport, breakdown, and utilization of lipids, maintains an active state in the lungs due to the abundance of lipids in the organ.⁹⁶ Studies have identified dysregulation of lipid metabolism as one of the major metabolic alterations in PF.98, 99, 100, 101, 102, 103 Several lipids and the expression of related enzymes were found to be altered in PF (Fig. 3).

Fig. 3.

Lipid metabolism reprogramming in PF. Several lipids are changed in the occurrence and progression of PF. Red: upregulated, blue: downregulated. AA: Arachidonic acid; ACC, Acetyl-CoA carboxylase; ACLY, ATP-citrate lyase; ATGL, Adipose triglyceride lipase; ATX, Autotaxin; CPT1/2, Carnitine palmitoyltransferase 1/2; CACT, Carnitine-acylcarnitine translocase; DG, Diacylglycerol; FA, Fatty acid; FASN, Fatty acid synthase; HDL, High-density lipoprotein; HMGCR, 3‑hydroxy-3-methylglutaryl-CoA reductase; LCAT, Lecithin cholesterol acyltransferase; LDL, Low-density lipoprotein; LPA, Lysophosphatidic acid; LysoPC, Lysophosphatidylcholine; NEFA, Non-esterified fatty acids; PA, Phosphatidic acid; PC, Phosphatidylcholine; PDH, Pyruvate dehydrogenase; PGD/E/F/G/H/I2, Prostaglandin D/E/F/G/H/I2; PLA1/2, Phospholipase A1/2; PLD, Phospholipase D; Sphk, Sphingosine kinase; TCA, Tricarboxylic acid; TG, Triglycerides; TXA2, Thromboxane A2.

Metabolomics refers to the systematic identification and quantification of the small molecule metabolic products (the metabolome) of a biological system (cell, tissue, organ, biological fluid, or organism) at a specific time point, serving as a promising tool to study the development of diseases. A targeted metabolomics study involving 300 patients with IPF and individuals without known lung disease revealed that the levels of 64 metabolites were significantly different between patients with IPF and controls.¹⁰⁰ Among these differential metabolites, multiple long-chain acylcarnitines, non-esterified fatty acids, and ceramides exhibited the most prominent differences, which were higher in IPF patients.³³ Consistently, IPF lung tissues showed elevated levels of long-chain acylcarnitines and two members of the hexosylceramide lipid family in another study.¹⁰⁴ Carnitine plays an indispensable role in the mitochondrial oxidation of long-chain fatty acids through regulating acyltransferases. Long-chain acylcarnitines, a class of lipids released by mitochondria when metabolism is inhibited, co-localize with pulmonary surfactant (PS) and inhibit the surface adsorption of PS as well as its ability to reduce alveolar surface tension, therefore representing a risk factor for lung injury in people with dysfunctional fatty acid oxidation.¹⁰⁵ However, levels of some medium-chain acylcarnitines decreased in IPF lung,¹⁰⁶ possibly due to the competitive oxidation between medium-chain acyl-Coenzyme A (acyl-CoAs) and long-chain acyl-CoAs.

Additionally, a study involving two IPF cohorts (n = 143) and one age-matched healthy cohort (n = 16) demonstrated several upregulated triglycerides (TG) and downregulated phosphatidylcholines (PC) in both IPF cohorts compared to the healthy cohort. Also, a few PCs (40:6, 38:6, 36:3) were found to be increased.¹⁰⁷ Lipid accumulation was found in IPF biopsy specimens and lipidomic analysis revealed that lipid accumulation was particularly related to TG deposition, which was significantly increased in IPF specimens compared to controls.¹⁰⁸ Further study found that the levels of three TGs (54:5, 54:6, 53:7) and two PCs (40:6, 36:3) were higher in the progressive IPF.¹⁰⁹ Similarly, certain TGs and PCs were also found to be elevated in mice treated with BLM.110, 111, 112 Fatty acid synthase (FASN), an enzyme that initiates de novo lipogenesis, was increased in fibroblasts from the lung tissues of patients with IPF and BLM-treated mice. It was reported that inhibition of FASN expression in fibroblasts effectively suppressed the proliferation, migration, and differentiation of lung fibroblasts.^113,114 Adipose triglyceride lipase (ATGL) is the first and rate-limiting enzyme of intracellular lipolysis. Kanti et al¹¹⁵ found ATGL deficiency caused TG accumulation in bronchiolar club cells as well as enhanced airway resistance and reduced airway compliance. Furthermore, lack of ATGL reduced the regeneration potential of airway epithelium in mice, which might contribute to the development of PF. In addition, α/β-hydrolase domain-containing protein 5 (ABHD5), which encodes the comparative gene identification-58 (CGI-58) protein, an intracellular lipolytic activator acting as a co-factor for ATGL, was low in the lung tissues of IPF patients.¹¹⁶ Pyroptosis of alveolar macrophages has been found to play an essential role in PF. Disturbance of lipid metabolism, such as accumulation of TG, was detected within the pyroptotic cells, which might trigger an NLRP3 inflammasome-related inflammatory cascade reaction. And inhibiting lipid accumulation was able to rescue pyroptotic macrophages, alleviating PF.¹¹⁷ PS, which is produced by AECⅡs, is essential for reducing alveolar surface tension, maintaining alveolar stability and improving pulmonary compliance. Surfactant protein C (SP-C) deficiency and mutation of SP-C gene (SFTPC) were found in patients suffering IPF.^97,118 PS is composed of 90% lipids and 10% proteins, and the lipid components mainly include phospholipids and cholesterol.¹¹⁹ Romero et al¹¹¹ found the synthesis of surfactant lipids was inhibited in AECⅡs isolated from BLM-treated mice while their levels were elevated in bronchoalveolar lavage fluid (BALF). This abnormal pattern reflects injury of AECⅡs. FASN, associated with synthesis of phospholipids and cholesterol, was decreased in AECⅡs.^111,120,121 And FASN deficiency in AECⅡs exacerbated BLM-induced lung fibrosis,¹²² probably because lipid deficiency contributed to impaired regeneration of AECⅡs.¹²¹ Lipid replenishment was able to promote AECⅡs’ renewal.¹²¹ And overexpression of FASN reduced BLM-induced loss of mitochondrial membrane potential and the production of mitochondrial ROS, therefore significantly inhibiting death of AECⅡs.¹²³

PC, one kind of phospholipids, is a major constituent of PS and the upregulated level of PC might be partially associated with greater alveolar epithelial injury in the progressive group compared to the stable group.⁹⁷ Interestingly, phospholipids and cholesterol increased in alveolar macrophages of BLM-treated mice due to uptake of lipids, which led to the formation of lipid-laden macrophages or foam cells. These foam cells exhibited an M2 macrophage phenotype. Specifically, PCs phagocytosed by macrophages were oxidized, which further promoted M2 polarization. The M2-polarized macrophages induced the production of TGF-β1,¹¹¹ subsequently activating lung fibroblasts to develop PF. Mechanistically, CD36, the scavenger receptor for oxidized phospholipids (oxPL), mediated TGF-β production by macrophages in response to oxPL, depending on Lyn kinase activation.¹²⁴ Furthermore, during glycerophospholipid metabolism, PCs can be hydrolyzed to lysophosphatidylcholines (lysoPCs). A serum metabolomic study found the levels of several LysoPCs, including LysoPC (16:0), LysoPC (18:0), and LysoPC [20:4 (5Z 8Z 11Z 14Z)], were significantly increased in BLM-exposed rats.¹²⁵ And it was reported that the lysophospholipid mediator lysophosphatidic acid (LPA) could induce fibrosis by mediating fibroblast recruitment as well as epithelial cell apoptosis through its receptor lysophosphatidic acid receptor 1 (LPA1).^126,127 Five LPAs (LPA16:0, 16:1, 18:1, 18:2, 20:4) were significantly higher in IPF patients, associating with the progression of disease.¹²⁸

Activated sphingolipid metabolism was also observed in BLM-treated mice and IPF patients and could be sensed by Trem2⁺ macrophages in mice fibrotic lungs (corresponding to glycoprotein non-metastatic melanoma protein B positive [GPNMB⁺] macrophages identified in human IPF patients) and then induced PF. Trem2 deficiency regulated the sphingolipid metabolism and alleviated BLM-induced lung fibrosis.¹²⁹ Sphingosine-1-phosphate (S1P) is an important metabolite of sphingolipid metabolism, converted from sphingosine by sphingosine kinase 1 (Sphk1). S1P and Sphk1 were found to be upregulated in IPF patients,^130,131 and elevated S1P then induced EMT via sphingosine-1-phosphate receptor 2 (S1P2) and partially by increasing TGF-β1 expression and secretion.^131,132 Deletion of Sphk1 or S1P2 as well as administration of a Sphk inhibitor or S1P2 antagonist alleviated BLM-induced lung injury and fibrosis.^130,132 Qiu et al¹³³ also observed increased expression of S1P3 in the lungs of BLM-treated mice and specific overexpression of S1P3 in the infiltrated M2 macrophages. And inhibition of S1P3 ameliorated BLM-induced PF by reducing macrophage M2 polarization via the PI3K/Akt–signal transducer and activator of transcription 3 (Stat3) signaling pathway. Furthermore, adeno-associated virus 9 (AAV9)-mediated overexpression of S1P lyase (SPL), the enzyme that degrades S1P, showed an anti-fibrotic effect in BLM-treated mice.¹³⁴ In cigarette smoke (CS)-exposed mouse models, pulmonary expression of Sphk2 and S1P secretion was significantly upregulated. Deletion of Sphk2 attenuated CS-induced PF.¹³⁵ Increased S1P content was also observed in lungs of rat silicosis models. Desipramine could reduce the level of S1P as well as inflammatory and fibrotic markers in silicosis.¹³⁶ Conversely, endothelial cell-specific deletion of S1P1 resulted in increased lung vascular permeability and exacerbated BLM-induced lung fibrosis.¹³⁷

Apolipoprotein A-I (ApoA-I) concentrations were lower in BALF from IPF patients and BLM-treated mice.¹³⁸ Furthermore, among the serum lipid profiles, the level of ApoA-I was significantly lower in non-survivors than in survivors. And the low ApoA-I level was independently associated with the risk of mortality via the multivariate Cox analysis.⁹⁸ Excessive accumulation of lipids such as cholesterol could result in lung inflammation and fibrosis by inducing the expression of collagens and multiple cytokines.¹³⁹ ApoA-I is able to maintain lipid homeostasis by transporting excess cholesterol and phospholipids to the liver for removal, therefore reducing the inflammatory and fibrotic responses in PF. Intranasal treatment with ApoA-I protein and overexpression of ApoA-I were able to alleviate PF,^138,140 partially by inhibiting EMT of alveolar epithelial cells.¹⁴¹ Mechanistically, cholesterol is transferred to ApoA-I via ABCA1 and forms the nascent high-density lipoprotein (HDL) particles on the cell membrane to remove cholesterol,¹⁴² while low-density lipoprotein (LDL) governs the uptake of cholesterol packaged with ApoB from the blood.¹⁴³ Consistently, the plasma levels of LDL were increased in IPF patients while HDL levels were decreased,¹⁴⁴ and a high level of HDL was associated with a lower incidence of death in patients with IPF.¹⁴⁵ Besides, ApoE, which is produced abundantly by monocyte-derived alveolar macrophages in the lungs, directly bound to collagen I and mediated phagocytosis of collagen I via LDL receptor-related protein 1 (LRP1), promoting the resolution of PF.¹⁴⁶ Table 2 lists the changes in some specific lipids in PF.^{24,28,33,34,98,99,101,102,104,106,}107, 108, 109, 110, 111, 112^,126, 127, 128^{,131,136,138,144}

Table 2.

Changes of metabolites in lipid metabolism in PF.

PF type	Changed substances	Location	References
IPF	Medium-chain acylcarnitines↓	Plasma/Lung	28,106
	Long-chain acylcarnitines, hexosylceramide↑	Plasma/Lung	33,104
	Triglyceride↑	Serum/Plasma/Lung/BALF	107, 108, 109,111,112
	Cholesterol, phospholipid↑	BALF/macrophage	111
	Phosphatidylcholine↑	Lung	110,112
	Lysophosphatidic acid↑	BALF	126, 127, 128
	Sphingosine-1-phosphate↑	Serum/Lung/BALF	131
	Monounsaturated fatty acid↓	Plasma/AEC	102,106
	Sphingolipid↓	Plasma/Lung	101,104
	ApoA-I↓	Serum/BALF	98,138
	LDL↑, HDL↓	Plasma	144
Silicosis	Prostaglandin D2/E2, thromboxane A2↑	Lung	34
	Triglyceride, cholesterol, phospholipid↑	BALF	111
	Sphingosine-1-phosphate↑	Lung	99,136
RIPF	Cholesterol, long-chain acylcarnitine↑	Plasma/Lung	24

AEC, Alveolar epithelial cell; ApoA-I, Apolipoprotein A-I, BALF, Bronchoalveolar lavage fluid; HDL, High-density lipoprotein; IPF, Idiopathic pulmonary fibrosis; LDL, Low-density lipoprotein; PF, Pulmonary fibrosis; RIPF, Radiation-induced pulmonary fibrosis.

As mentioned above, lipid accumulation is common in PF. Yang et al¹⁴⁷ reported that injured AECⅡs represented the primary source of lipid accumulation in PF. In particular, lysoPCs (14:0 and 16:0) released from damaged AECⅡs activated lung fibroblasts and thereby accelerated PF progression. Mechanistically, the expression of 3‑hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2) was markedly reduced in AECⅡs, whereas high expression level of HMGCS2 can significantly attenuate PF progression. This protective effect occurred through modulation of lipid degradation in AECⅡs, as HMGCS2 interacts with peroxisome proliferator-activated receptor α (PPARα) and promotes the expression of carnitine palmitoyltransferase 1A (CPT1A) and carnitine palmitoyltransferase 2 (CPT2); overexpression of CPT1A exerts a protective effect against PF by markedly reducing lipid peroxide accumulation and inhibiting fibroblast proliferation.¹⁴⁸ In contrast, Wang et al¹⁴⁹ showed that acetyl-CoA synthetase short-chain family member 3 (ACSS3) reduced fatty acid oxidation by causing CPT1A deficiency in AECⅡs. This change led to the accumulation of lipid droplets but, unexpectedly, alleviated PF. These seemingly contradictory findings may reflect the complex and sometimes opposing roles of lipids during PF progression. Overall, certain lipids promote AECⅡ injury and apoptosis, thereby enhancing fibrotic responses, whereas others support AECⅡ regeneration and may contribute to PF resolution. The two enzymes described above may regulate the oxidation of different lipid species in AECⅡs, which could explain their opposite effects. Therefore, further studies are urgently required to clarify the regulatory mechanisms governing specific lipid species in PF.

Lipid peroxidation, which is related to ferroptosis, is another characteristic of PF.150, 151, 152, 153 Ferroptosis is a newly discovered type of regulated cell death, characterized by the iron-dependent accumulation of lipid ROS generated from lipid peroxidation.¹⁵⁴ Shao et al¹⁵⁵ found that BLM-induced increase in iron level in AECⅡs was accompanied by pathological changes and collagen deposition, indicating iron deposition-induced ferroptosis promoted the development of PF.¹⁵⁴ More concretely, a marked iron accumulation within the mitochondria of AECⅡs in PF tissues was discovered, along with a considerable increase in the protein level of mitoferrin 2 (MFRN2), a pivotal mediator of iron transport into the mitochondria. Mitochondrial iron deposition contributed to AECⅡs injury and attenuating mitochondrial iron deposition in AECⅡs via activating E-type prostanoid receptor 4 (EP4) could mitigate injury following BLM exposure.¹⁵⁵ Additionally, in BLM-induced and particle-induced mouse modesl, significant iron deposition and lipid peroxidation in pulmonary macrophages were observed.^156,157 Furthermore, iron overload in macrophages triggered lipid peroxidation and TGF-β1 elevation, which subsequently activated lung fibroblasts and promoted myofibroblast differentiation,^156,158 highlighting the potential of targeting iron overload and lipid peroxidation for treating PF. Sirtuin 6 was found to be effective in alleviating PF. Mechanistically, it could ameliorate BLM-induced ectopic lipotoxicity by enhancing lipid degradation, thereby increasing the energy supply and reducing the levels of lipid peroxidation.¹⁵⁹ In addition, lipid peroxidation triggered ferroptotic angiocrine signaling in endothelial cells (ECs) to aberrantly activate fibroblasts. However, dopaminylation of triosephosphate isomerase 1 (TPI1) directed ether phospholipid synthesis to glucose metabolism in regenerating lung ECs, which attenuates lipid peroxidation and blocks ferroptosis,¹⁶⁰ indicating the important roles of lipid metabolism as well as its balance with glucose metabolism in PF. In silica-treated mouse lungs and alveolar macrophages of BLM-treated mice, the metabolic status also displayed a transformation from lipid metabolism to glucose metabolism.^161,162

Metabolomics studies also demonstrated various differential metabolites between mice with silicosis and healthy controls.^34,163,164 The Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed significant enrichment of AA metabolism.^34,163 Fatty acids, usually unsaturated fatty acids such as AA, are indispensable components of glycerophospholipids alongside glycerol, choline, and ethanolamine. Consistently, another study also discovered significantly altered metabolic pathways of glycerophospholipids in the lungs of silica-treated mice.¹⁶⁴ Further targeted metabolomics of AA metabolism revealed increased levels of prostaglandin D2 (PGD2), prostaglandin E2 (PGE2), and thromboxane A2 (TXA2) in silicosis mouse lungs.³⁴ PGD2 and TXA2 levels have been demonstrated to promote the production of proinflammatory cytokines and profibrotic cytokines while the role of PGE2 in inflammation and fibrogenesis remains controversial.^165,166 Chen et al¹⁶⁷ conducted a population-based case-control study, detecting the serum metabolites of 150 coal workers’ pneumoconiosis (CWP) cases and 120 healthy controls. They identified 131 different lipid metabolites between the two groups, among which acyl-hexosylceramide, diacylglycerol, and dimethyl-phosphatidylethanolamine demonstrated good sensitivity and specificity. However, the underlying mechanisms of these lipids involved in CWP need to be further studied.

Lipid composition and metabolism are highly complex. In PF, lipid accumulation and lipid peroxidation represent prominent metabolic features that influence alveolar surfactant biosynthesis and alter the physiological state of multiple lung cell types (Fig. 4). In general, fibroblasts and club cells primarily exhibit lipid accumulation. Increased lipid synthesis promotes fibroblast proliferation, migration, and differentiation, while also impairing the regenerative capacity of club cells, thereby contributing to PF progression. In contrast, lipid peroxidation occurs more prominently in macrophages. Peroxidized lipids drive macrophages toward a pro-fibrotic phenotype, leading to the release of inflammatory cytokines that subsequently activate fibroblasts. The effects of lipids in AECⅡs are more complex, as different lipid species exert opposing roles. Certain lipids, such as LPA and S1P, promote apoptosis or EMT in AECⅡs and thus exert pro-fibrotic effects. In contrast, other lipids, including surfactant-associated lipids, support AECⅡ regeneration, and display anti-fibrotic properties. Therefore, both regulation of lipid synthesis and precise control of specific lipid species that drive disease progression are essential. Some lipid mediators may influence multiple cell types and thus represent potential therapeutic targets. For example, S1P can induce EMT in AECⅡs and promote M2 polarization of macrophages through the S1P2 and S1P3 receptors, respectively. However, the molecular mechanisms that regulate lipid metabolism and determine how altered lipid species contribute to PF remain poorly understood. Moreover, lipid alterations show less consistency across different types of PF compared with glucose metabolism. For instance, AA metabolism, rather than TG or PC metabolism, represents the primary lipid metabolic alteration in silicosis. This difference may arise because inflammation serves as the principal upstream driver of fibrosis in silicosis, and AA metabolism closely links to inflammatory responses. Therefore, further studies are needed to clarify the roles of lipid metabolism in specific cell types during PF and to identify precise therapeutic targets and strategies.

Fig. 4.

Lipid metabolism reprogramming of different cells in PF. (A) TGF-β1 upregulates FASN expression via β-catenin and PI3K/Akt/mTOR, inducing proliferation, migration, and differentiation of fibroblasts. BLM results in fibroblast recruitment by increasing LPA and its receptor LPA1. (B) BLM leads to endocytosis of oxPL in macrophages by the scavenger receptor CD36, which subsequently releases TGF-β to activate fibroblasts. BLM also promotes interaction of S1P and its receptor S1P3, resulting in M2 polarization of macrophages via PI3K/Akt/STAT3. TG stimulation induced by LPS and Nigericin acts on fibroblasts by activating NLRP3 inflammasomes. (C) Alveolar surface tension is damaged by downregulation of FASN and HMGCR by BLM, which is important for synthesis of alveoli surfactant lipids. However, BLM increases LPA, which induces apoptosis of epithelial cells. Sphk1 is also activated to promote EMT in epithelial cells of PF via S1P/S1P2. (D) ABHD5 encodes the CGI-58 protein, a co-factor of ATGL, which could break down TG. While in PF club cells, ABHD5 is decreased, leading to TG accumulation and obstruction of cell regeneration. ABHD5, Abhydrolase domain-containing protein 5; Akt, Protein kinase B; ATGL, Adipose triglyceride lipase; BLM, Bleomycin; CD36, Cluster of differentiation 36; CGI-58, Comparative gene identification-58; EMT, Epithelial-mesenchymal transition; FASN, Fatty acid synthase; GSDMD, Gasdermin D; HMGCR, 3‑hydroxy-3-methylglutaryl-coa reductase; IL-1β, Interleukin-1 beta; LPA, Lysophosphatidic acid; LPA1, Lysophosphatidic acid receptor 1; LPS, Lipopolysaccharide; Lyn, Lyn proto-oncogene tyrosine-protein kinase; mTOR, Mammalian target of rapamycin; N-GSDMD, N-terminal gasdermin D; NLRP3, Nod-like receptor pyrin domain-containing 3; oxPL, Oxidized phospholipids; PI3K, Phosphatidylinositol 3-kinase; S1P, Sphingosine-1-phosphate; S1P2/3, Sphingosine-1-phosphate receptor 2/3; Sphk1, Sphingosine kinase 1; STAT3, Signal transducer and activator of transcription 3; TGF-β1, Transforming growth factor-beta 1; TG, Triglyceride; β-catenin, Beta-catenin; caspase1, Cysteinyl aspartate specific proteinase 1.

Amino acid metabolism

Amino acid metabolism refers to the synthesis, degradation, and utilization of amino acids in organisms. This process serves as a fundamental basis for several biochemical processes and physiological functions within the body, most notably the biosynthesis of proteins, given that amino acids are the fundamental building blocks of proteins. It is also involved in the synthesis of various other biomolecules, such as lipids, nucleic acids, and redox substances. Moreover, amino acids can be oxidized to generate ATP, acting as an important energy source when glycogen and fatty acids are in short supply.

Moreover, amino acids are indispensable components of collagen, which is the most abundant component of the ECM. Collagen is a fibrous protein consisting of three parallel polypeptide strands forming triple helical structure known as tropocollagen. The polypeptide strands are composed of a repeating glycine (Gly)-X-Y motif, where X and Y can represent any amino acid but are mostly proline in X and hydroxyproline in Y. Therefore, amino acids on their own and their metabolism play an important role in the regulation of IPF. Recently, several amino acids such as glycine, arginine, proline, glutamine, and methionine have been shown to have effects on PF (Fig. 5).^27,168, 169, 170, 171, 172, 173

Fig. 5.

Amino acid metabolism reprogramming in PF. Certain amino acids and related enzymes are changed in PF. Red: upregulate; blue: downregulate. 3-PG, 3-phosphoglycerate; 3-PHP, 3-phosphohydroxypyruvate; 3-PS, 3-phosphoserine; α-KG, Alpha-ketoglutarate; Arg, Arginine; ASCT2, Alanine, serine, cysteine transporter 2; ATF4, Activating transcription factor 4; G6P, Glucose 6-phosphate; GDH, Glutamate dehydrogenase; GLS, Glutaminase; GLUT1, Glucose transporter 1; GSA, Glutamate semialdehyde; mRNA: Messenger RNA; mTORC1, Mammalian target of rapamycin complex 1; OAT, Ornithine aminotransferase; P5C, Delta-1-pyrroline-5-carboxylate; P5CR, Pyrroline-5-carboxylate reductase; P5CS, Pyrroline-5-carboxylate synthase; PHGDH, Phosphoglycerate dehydrogenase; PSAT1, Phosphoserine aminotransferase 1; PSPH, Phosphoserine phosphatase; SHMT2, Serine hydroxymethyltransferase 2; Smad3, Small mothers against decapentaplegic homolog 3; TCA, Tricarboxylic acid cycle; TGF-β1, Transforming growth factor-beta 1.

Glycine, proline, and hydroxyproline are the major amino acids forming collagen. Studies have demonstrated that enzymes involved in the de novo serine‑glycine biosynthetic pathway were upregulated in fibroblasts stimulated with TGF-β1,^174,175 including phosphoglycerate dehydrogenase (PHGDH), phosphoserine aminotransferase 1 (PSAT1), phosphoserine phosphatase (PSPH), and serine hydroxy methyltransferase (SHMT). TGF-β1 induced a Smad3-dependent increase in activating transcription factor 4 (ATF4) mRNA abundance and mTORC1 activation, then mTORC1 promoted ATF4 protein production, subsequently promoting the transcription of key serine‑glycine pathway genes and GLUT1. Together, they promoted glucose-derived glycine biosynthesis to support enhanced collagen synthesis rates in activated myofibroblasts.¹⁷⁴ Inhibition of PHGDH decreased TGF-β-induced collagen synthesis in murine and human lung fibroblasts as well as lung fibrosis in BLM-treated mice.¹⁷⁶ In the airway epithelium, aberrant activation of YAP/TAZ induced mTORC1–ATF4 activity, enabling the uptake of essential amino acids. It is required for the fate conversion of secretory cells, which coincided with severe subepithelial fibrosis.¹⁷⁷ l-proline, a precursor of collagen, is synthesized from l-arginine by the action of arginase (Arg) and subsequent catalyzed by ornithine aminotransferase (OAT). High levels of arginine metabolites, including creatine, putrescine, spermidine, 4-hydroxyproline, and proline-hydroxyproline dipeptide, were detected in IPF patients.²⁸ And Arg-1 and OAT were also highly expressed in IPF patients and BLM-treated mice,178, 179, 180, 181, 182 while the arginase inhibitor as well as the traditional Chinese herbs trichodelphinine A and Pueraria lobata could regulate Arg-1 and OAT to affect arginine and proline metabolism, impacting the production of collagen.182, 183, 184 Additionally, de novo synthesis of proline was also found to be crucial for TGF-β-induced collagen production and for organ fibrosis.^185,186 The urea cycle (UC), which provides adequate levels of proline, is needed for biosynthesis of collagen; carbamoyl phosphate synthetase 1 (CPS1) controls the first reaction of the UC. In arsenite-induced PF, high amounts of ammonia in the lung microenvironment promoted the expression levels of CPS1 and the proline metabolism. Reduction of ammonia and CPS1 ablation could inhibit collagen synthesis and ameliorate PF induced by arsenite.¹⁸⁷

Glutamine (Gln), which is the most abundant amino acid in the human circulation and intracellular pools, is involved in many cellular processes such as regulation of immune cell proliferation and function, production of proline, which is vital for collagen synthesis, and generation of energy through glutaminolysis.^188,189 Glutaminolysis is the metabolic process of glutamine, which involves the conversion of glutamine to glutamate by glutaminase (GLS) and constitutes a key anaplerotic reaction that replenishes the TCA cycle via α-ketoglutarate (α-KG) generation.¹⁸⁹ A targeted metabolomics analysis of amino acid-related metabolites using the human plasma samples found that glutamate increased in IPF and it showed an increasing linear trend across different ILD phenotypes, reflecting progression of PF.¹⁹⁰ The enhanced glutaminolysis accelerates IPF fibroblast proliferation, migration, differentiation, and collagen production, which are some of the characteristics of pro-fibrotic fibroblasts in IPF,^191,192 indicating that glutamine metabolism is critical for supporting lung fibrosis. Similarly, increased GLS1 was observed in silica-induced PF, which resulted from activation of the mechanotransduction regulator YAP1 in the nucleus and led to fibroblast activation via regulating glutaminolysis.¹⁹³ Moreover, enhanced glutaminolysis was found in differentiated myofibroblasts compared with controls, whereas both depletion of extracellular Gln and silencing of GLS1 expression prevented myofibroblast differentiation.^194,195 The glutaminolysis metabolite α-KG affected Jumonji domain-containing protein-3 (JMJD3), which further regulated trimethylation at lysine 27 of histone H3 (H3K27me3) levels and the expression of X-linked inhibitor of apoptosis (XIAP) in an α-KG-dependent pathway, leading to fibroblasts’ acquisition of an apoptosis-resistant phenotype which contributed to the accumulation of myofibroblasts.¹⁹⁶ And augmentation of glutaminolysis in myofibroblasts enhanced the production of collagen by activating the mTORC1 and promoting hydroxylation of proline, which could be abrogated by inhibiting GLS1.¹⁸⁹ Inhibition of GLS1 could reverse glutamine metabolic reprogramming, reducing mitochondrial energy expenditure and inhibiting collagen matrix synthesis, and inhibit the transformation of fibroblasts toward myofibroblasts.^197,198 Moreover, inhibition of glutamine transporter alanine‑serine-cysteine transporter 2 (ASCT2, known as SLC1A5) showed an anti-PF effect,^199,200 which could decrease collagen production and hydroxyproline levels as well as inflammatory cytokines.²⁰⁰ Interestingly, inhibition of glutamine metabolism in AECⅡs was observed during BLM-induced lung fibrosis in mice. Further analysis showed that the expression levels of catalytic enzymes responsible for glutamine catabolism, such as GLS1 and glutamic-pyruvic transaminase-2 (GPT2), were downregulated in AECⅡs of patients with IPF, which resulted in reduced proliferation and differentiation of AECⅡs,²⁰¹ suggesting the importance of glutamine metabolism for alveolar regeneration. Glutamine treatment could restore mitochondrial respiration and reduce cytotoxicity in lung epithelial cells.⁴⁵ TGF-β1 decreased intracellular methionine and S-adenosylmethionine (SAM) levels, while maintaining s-adenosylhomocysteine (SAH) levels in fibroblasts. Futhermore, methionine and SAM were found to inhibit the fibroblast-to-myofibroblast transition, but the underlying mechanism remained unclear.¹⁷⁰ Table 3 lists the changes of metabolic enzymes and metabolites in amino acid metabolism in PF.^{20,21,23,24,27,28,31,106,125,168,172,173,178,}180, 181, 182, 183^{,187,189,190,193,195,202}

Table 3.

Changes of metabolic enzymes and metabolites in amino acid metabolism in PF.

PF type	Changed substances	Location	References
IPF	Proline↑	Lung/Breath	20,21,168,172,202
	Hydroxyproline↑	Breath	168
	Glycine↑	Lung	20,21,172
	Serine↑	Lung	20,23
	Leucine↑	Serum	125,172
	Arginine↑	Lung	20
	Glutamine↓	Plasma	106,189
	Glutamate↑	Plasma/Lung	20,21,172,189,190
	Arg↑	Lung	31,178,180, 181, 182, 183
	OAT↑	Lung	21,28,183
	GLS1↑	Fibroblast	189,195
Silicosis	Arginine↑	Plasma	27,173
	GLS1↑	Fibroblast	193
Asbestosis	Arginine↑	Plasma	27
RIPF	Leucine, arginine, proline↓	Plasma/Lung	24
Arsenite-induced PF	Proline↑	Fibroblast	187

Arg, Arginase; GLS1, Glutaminase 1; IPF, Idiopathic pulmonary fibrosis; OAT, Ornithine aminotransferase; PF, Pulmonary fibrosis; RIPF, Radiation-induced pulmonary fibrosis.

Disruption of amino acid metabolism in PF mainly relates to the synthesis of collagen, a key component of the ECM. Increased production of collagen-associated amino acids, including glycine, arginine, and proline, supports excessive ECM deposition during fibrosis. Inhibition of the generation of these amino acids has shown antifibrotic effects, suggesting a potential therapeutic strategy for PF. Glutaminolysis, however, exerts complex and bidirectional effects during PF development. On the one hand, it promotes the pro-fibrotic phenotype of fibroblasts. On the other hand, glutaminolysis supports alveolar regeneration, which contributes to tissue repair and recovery. Therefore, these dual effects must be carefully considered when targeting glutaminolysis as a therapeutic approach for PF.

Overview of metabolic reprogramming in patients with PF

PF is progressive and irreversible, while currently there are no specific drugs or strategies to reverse it. Thus, the exploration of new drug targets is of great importance. Accumulating evidence indicates that metabolic reprogramming, including glucose metabolism, lipid metabolism, and amino acid metabolism, acts as a key mediator during the pathogenesis of PF. In this review, we conclude that alterations of various metabolites and related enzymes could influence the occurrence and development of PF. Specifically, metabolic alterations in glucose and lipid pathways mainly influence fibroblast differentiation into myofibroblasts, macrophage polarization toward the M2 phenotype, and the regeneration and EMT of AECⅡs. Meanwhile, several amino acids that serve as fundamental components of collagen show increased synthesis in PF. In addition, these metabolites and metabolic genes have been proposed as potential biomarkers for PF diagnosis and prognosis.202, 203, 204, 205, 206, 207 Despite these advances, several fundamental challenges remain to be addressed before metabolic interventions can be translated into clinical practice.

First, the causal relationship between metabolic reprogramming and PF pathogenesis remains incompletely defined. Most current evidence is derived from correlative analyses of clinical biospecimens or end-stage animal models, whereas the temporal dynamics of metabolic alterations across disease initiation, progression, and resolution phases remain poorly characterized. Therefore, it remains unclear whether these metabolic dysregulation acts as an upstream causal driver of fibrogenesis, a compensatory adaptive response to tissue injury, or merely a secondary consequence of progressive tissue remodeling. Although several studies have further demonstrated that genetic deletion or pharmacological inhibition of specific metabolic regulators can alleviate the severity of PF,^35,36,52 more research is needed to define whether and how metabolic reprogramming drives PF so that therapies can target disease mechanisms rather than only alleviating symptoms.

Second, cellular heterogeneity of the lung poses a major barrier to non-selective metabolic intervention. The lung comprises dozens of functionally distinct cell types, and metabolic pathways often exert cell-type-specific, even opposing, effects in PF pathogenesis. The same metabolic pathway may be profibrotic in fibroblasts or macrophages but antifibrotic or reparative in AECⅡs. This functional duality means that systemic metabolic modulation may simultaneously suppress pathogenic programs in one cell type while impairing endogenous repair in another. This functional divergence makes cell-specific interventions essential. Given the cellular heterogeneity of the lung, effective therapies must act selectively on the relevant cell populations. Emerging single-cell metabolomics can pinpoint the metabolically vulnerable cellular subpopulations in diseased lungs, and targeted delivery strategies—for example, decorating nanocarriers with cell-specific ligands or using cell type-restricted promoters—should enable precise modulation of pathogenic metabolic programs.

Third, heterogeneity among fibrotic lung diseases further complicates therapeutic design. Distinct metabolic profiles have been observed across different fibrotic lung diseases, for example, lipid profiles in BALF and serum differ between IPF and hypersensitivity pneumonitis,²⁰⁸ two ILDs that share histologic features but arise from distinct processes: hypersensitivity pneumonitis is primarily a post-inflammatory, immune-driven disorder, whereas IPF is dominated by aberrant tissue remodeling and repair. These differences argue for disease- and sample-specific metabolic studies. Broadly, systematic analyses across diverse clinical specimens and ILD subtypes will be required to define actionable, cell-selective metabolic targets for precision therapy.

Fourth, the interconnected nature of metabolic networks and systemic off-target toxicity limit therapeutic efficacy and safety. The three major metabolic pathways involved in PF do not operate independently; instead, they are tightly interconnected and influence one another. Effective regulation of pulmonary regeneration and fibrotic remodeling requires a balanced interaction between glucose and lipid metabolism.¹⁶⁰ In addition, intermediates generated from glucose metabolism provide essential substrates for the synthesis of multiple lipids and amino acids.^174,175 Therefore, therapeutic strategies that target key mediators or upstream regulators controlling two or more metabolic pathways may offer greater efficacy. Several signaling molecules, including PI3K, Akt, and mTOR, simultaneously regulate glucose and lipid metabolism and may therefore represent important therapeutic targets. For example, TGF-β or LPS can enhance glycolysis primarily by stimulating GLUT1 and PFKFB3 through activation of the PI3K/Akt/mTOR signaling pathway, which in turn promotes fibroblast activation.^42,54,90 At the same time, the PI3K/Akt/mTOR pathway can activate FASN, thereby driving fibroblast proliferation, migration, and differentiation.¹¹⁴ Consequently, components of this pathway may represent promising upstream targets for PF treatment. However, targeting metabolic pathways also raises concerns regarding potential toxic side effects. Because metabolic processes are systemic and most drugs are administered systemically, metabolic interventions may produce off-target effects in other organs, including the liver and kidneys, which depend on the same metabolic pathways. To mitigate these risks, future studies should investigate localized pulmonary delivery strategies, such as nebulized inhalation, or develop prodrugs that are selectively activated by lung-specific enzymes to improve treatment specificity and safety.

The convergence of emerging biotechnologies and computational advances opens new avenues to address these challenges and advance metabolic-targeted therapy for PF. First, single-cell and spatial multi-omics, coupled with artificial intelligence (AI), will enable high-resolution dissection of metabolic reprogramming and precision target discovery. Single-cell metabolomics, combined with single-cell transcriptomics and proteomics, can pinpoint cell-subpopulation-specific metabolic signatures in damaged lungs and identify metabolically vulnerable pathogenic cell subsets. Spatial metabolomics further maps metabolic alterations onto tissue architecture, linking metabolic dysregulation directly to pathological lesions such as fibroblastic foci. Second, precision targeted delivery strategies will improve the therapeutic efficacy of metabolic interventions. To overcome cellular heterogeneity and systemic toxicity, next-generation therapies must achieve cell-type-selective and lung-localized modulation of metabolic pathways. Nanomedicine-based delivery systems—for example, nanocarriers functionalized with ligands that recognize cell-specific surface markers—can deliver metabolic drugs selectively to pathogenic cell populations. For gene-based metabolic modulation, cell type-restricted promoters can be used to restrict transgene expression to target cell types. Local pulmonary delivery via nebulized inhalation represents another straightforward strategy to increase drug concentration in the lung while minimizing systemic exposure. The development of lung-specific prodrugs, which are pharmacologically inert in the circulation but selectively activated by enzymes enriched in fibrotic lung tissue, will further enhance therapeutic specificity and safety. Third, precision nutritional and dietary interventions offer a low-risk adjunctive therapeutic approach. As the major substrates of cellular metabolism—glucose, lipids, and amino acids—are primarily obtained from the diet,209, 210, 211 dietary patterns can profoundly shape the pulmonary metabolic microenvironment and influence fibrotic progression. Future research should evaluate the therapeutic potential of tailored dietary strategies, such as ketogenic diet, specific amino acid restriction, or bioactive lipid supplementation, in preclinical models and clinical cohorts. In the long term, metabolomics-guided precision nutrition, which tailors dietary recommendations to individual metabolic signatures, may be integrated into standard PF management to slow disease progression and improve patient quality of life.

In conclusion, metabolic reprogramming has emerged as a key hallmark of PF. It actively contributes to the initiation and progression of the disease and has also been proposed as a potential therapeutic target. Interventions that modulate metabolic pathways therefore hold significant clinical promise for PF treatment. Although many studies have investigated metabolic alterations in PF, substantial gaps remain before these findings can be translated into clinical practice. Future research should further clarify the mechanisms and heterogeneity of metabolic reprogramming in PF and identify more specific therapeutic targets. Such advances may provide new opportunities for effective PF treatment.

Funding

This study was funded by National Natural Science Foundation of China (No. 82300098), CAMS Innovation Fund for Medical Sciences (CIFMS) (No. 2023-I2M-2-001), and the State Key Laboratory Special Fund (No. 2060204).

CRediT authorship contribution statement

Huan Zhou: Writing – review & editing, Writing – original draft. Tiantian Zhang: Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Edited by: Xiangxiang Pan and Peifang Wei