A Twisted-Integrated Multifunctional Fiber Sensor for Real-Time Metabolic Monitoring and Management of Sepsis

Abstract

Sepsis-induced systemic metabolic dysregulation involves complex interactions between acid–base imbalance and glucose metabolism abnormalities. Traditional metabolic monitoring methods, which rely on intermittent blood sampling, lack sufficient spatiotemporal resolution and fail to capture the dynamic pathological changes in detail. To address this, we present a minimally invasive multifunctional fiber sensor (PGFs) with a twisted integration structure for real-time, simultaneous monitoring of pH and glucose concentrations in sepsis. PGFs integrate pH and glucose fiber sensors along with their reference electrodes through a twisted design, offering excellent flexibility, rapid response, and high sensitivity. Additionally, the twisted structure enhances the stability of the bioelectrode–organic interface and improves biocompatibility. Long-term monitoring using PGFs in a sepsis animal model allowed us to construct a temporal metabolic profile of sepsis. Furthermore, metabolic management with PGFs significantly improved the survival rate of septic mice and alleviated sepsis-induced organ damage. Mechanistic studies revealed that PGFs-based combined intervention effectively disrupted the vicious cycle between acidosis and glucose dysregulation, reducing sepsis-induced inflammation and immune responses, improving the metabolic microenvironment, and restoring energy homeostasis. In conclusion, this study provides a platform for metabolic monitoring and management in sepsis using PGFs, offering valuable insights for clinical therapeutic strategies.

Introduction

Sepsis is a systemic inflammatory response syndrome caused by infection, characterized by an abnormal immune response that can lead to systemic inflammation, organ dysfunction, and even organ failure, with a fatality rate that remains alarmingly high. , According to data from the World Health Organization, sepsis is responsible for approximately 11 million deaths annually. Despite advancements in early identification and treatment strategies, the mortality rate of sepsis continues to be persistently high due to its complex pathogenesis, with metabolic disorders being among the most lethal mechanisms. Patients with sepsis frequently present with widespread metabolic disturbances, including glucose metabolism abnormalities and acid–base imbalances. These metabolic disturbances not only worsen the clinical manifestations of sepsis but also activate the immune response, promote inflammation and oxidative stress, and contribute to the impairment of multiple organ systems.

Sepsis is often accompanied by significant metabolic dysregulation during its pathophysiological progression. pH disturbances primarily manifest as metabolic acidosis, particularly lactic acidosis, which is caused by factors such as lactate accumulation due to tissue hypoperfusion, impaired lactate metabolism due to renal dysfunction, and respiratory acidosis. The accumulation of lactate leads to a decrease in pH, further exacerbating the progression of sepsis. Glucose metabolism disorders typically present as hyperglycemia, hypoglycemia, or glucose fluctuations, primarily driven by factors such as the stress response triggered by sepsis, insulin resistance, tissue hypoxia, and liver dysfunction. − pH dysregulation and glucose metabolism disorders do not exist in isolation but interact through various mechanisms, forming a complex metabolic network. Therefore, integrated management of pH and glucose dysregulation is crucial. However, in clinical practice, these metabolic abnormalities are often detected only when obvious clinical symptoms appear, by which time metabolic imbalance has already reached an advanced stage. Currently, delayed diagnostic methods relying on intermittent blood sampling hinder timely intervention, further exacerbating organ damage and reducing patient survival rates. , Therefore, real-time monitoring of pH and glucose levels, along with early and continuous metabolic intervention based on the monitoring results, plays a pivotal role in improving clinical outcomes for sepsis patients.

In this study, we introduce an injectable and implantable dual-functional flexible sensor for simultaneous pH and glucose monitoring (PGFs). , This sensor can be implanted into the body via injection, enabling real-time, high-resolution, and highly selective monitoring of both pH and glucose changes. First, highly conductive carbon nanotube fibers are synthesized through dry spinning technology to serve as the substrate, providing flexibility and excellent electrical conductivity. A highly sensitive pH sensor is then fabricated by electrodepositing polyaniline onto the carbon nanotube fibers. Meanwhile, a glucose sensor is created by coating the fibers with a glucose-responsive layer and an anti-interference layer. This glucose sensor exhibits a wide detection range, high sensitivity, and excellent selectivity. The pH sensor, glucose sensor, and an Ag/AgCl reference electrode are coaxially twisted to form a multifunctional electrode, allowing for simultaneous acquisition of pH and glucose data from the same site, thereby minimizing spatiotemporal discrepancies that would arise from separate measurements. , The twisted structure also endows the sensor with natural tissue-like stability in complex and dynamic physiological environments through a triple-synergistic mechanism of mechanical adaptation, interfacial interlocking, and biological integration. Biosafety assessments using hematoxylin and eosin (HE) staining and immunofluorescence confirmed the sensor’s biocompatibility. In a sepsis animal model, dynamic pH and glucose changes showed distinct temporal correlations. Combined pH-glucose management improved survival time (49.0h to 56.5h), increased blood oxygen saturation (93% to 94%), and reduced lung tissue damage (30% reduction in alveolar septal thickening, 2-fold decrease in lung injury score). Continuous monitoring and intervention alleviated inflammation, enhanced the metabolic microenvironment, and maintained energy homeostasis. Combined intervention also reduced p-P38 and p-ERK levels, inhibited the MAPK pathway, and improved inflammatory and metabolic disorders. This study demonstrates that combined pH-glucose management is more effective than single-parameter interventions in restoring energy metabolism during sepsis, advancing intelligent sensing systems for complex diseases.

Results and Discussion

Necessity of PGFs in Sepsis

Sepsis-induced metabolic disorders are fundamentally a synergistic dysregulation of acid–base balance and glucose homeostasis. Clinical observations have shown that pathogen infections disrupt the internal homeostasis of patients through multiple pathways. For instance, metabolic acidosis results from the accumulation of acidic metabolites due to renal dysfunction, while respiratory dysfunction exacerbates carbon dioxide retention, leading to respiratory acidosis. Importantly, these acid–base imbalances inhibit key glycolytic enzymes, worsen insulin resistance, and disrupt hepatic gluconeogenesis. Additionally, systemic stress, tissue hypoxia, and endocrine dysfunction further destabilize glucose metabolism. This creates a self-reinforcing vicious cycle between pH and glucose fluctuations (Figure a). To break this cycle, we developed a twisted pH/glucose dual-functional sensor (PGFs). Its spatially colocalized design enables real-time monitoring of the dynamic correlation between pH and glucose concentrations in the subcutaneous tissue microenvironment (Figure b). Predictions from animal models suggest that, compared with traditional single-parameter management, the dual-parameter synergistic regulation approach can significantly accelerate metabolic recovery (Figure c). This offers a novel strategy for the precise monitoring and treatment of sepsis.

1.

Sepsis-induced metabolic disruptions and intervention strategies. (a) Pathogenesis of pH and glucose metabolic disorders in sepsis patients. Sepsis triggers a systemic response, including metabolic and respiratory acidosis, renal dysfunction, tissue hypoxia, and insulin resistance, leading to dynamic changes in pH and glucose levels. These disturbances form a self-perpetuating vicious cycle, thereby exacerbating disease progression. (b) A fiber sensor (PGFs) capable of simultaneously and in real-time detecting pH and glucose in the subcutaneous tissue. (c) Comparison of treatment outcomes under different management approaches. Compared to single-parameter interventions, integrated management targeting both glucose and pH results in a higher cure rate and faster recovery.

Performance of PGFs

The twisted structure successfully integrated and decoupled the dual-functional modules (Figures a–c). The glucose sensor exhibited response performance comparable to standalone devices, demonstrating an excellent linear relationship within a concentration range of 25 mmol/L (R² = 0.995). Moreover, its response time met the real-time monitoring requirements for clinical applications (Figures d–e). Selectivity experiments demonstrated that the glucose sensor, modified with a functional membrane, exhibited negligible responses to common interfering substances, consistent with the specific capture of glucose by its molecular recognition sites (Figure f). Meanwhile, the pH sensor module maintained its sensitivity at the theoretical Nernstian slope (∼59 mV/pH) and showed good selectivity (Figures g–i). Notably, the integration process did not affect the performance of either the glucose or the pH sensor, as both retained comparable selectivity before and after integration (Figures f, i). Notably, the integration process did not affect the performance of either the glucose or the pH sensor, as both retained comparable selectivity before and after integration (Figures f, i).

2.

Structural and electrochemical performance characterization of PGFs. (a–c) Scanning electron microscope (SEM) images of the PGFs (a), the surface of the glucose sensor (b), and the surface of the pH sensor (c). (d–e) Amperometric responses of the glucose sensor before and after integration (d), along with its calibration curve (e). (f) Selectivity test of the glucose sensor before and after integration against common interferents. (g–h) Open-circuit potential response (g) and sensitivity (h) of the pH sensor. (i) Selectivity test of the pH sensor before and after integration. (j–k) Stability of the glucose (j) and pH (k) sensors over a 10-day period. (l) Time-dependent changes in the sensitivity retention rates of both sensors over a 14-day period.

To assess the stability of the integrated PGFs system under metabolically complex conditions, we introduced typical sepsis-related metabolites including urea and bilirubin during electrochemical measurements. Both the pH and glucose sensors maintained stable responses without noticeable drift (Figures S1–S2). Additionally, the influence of hydrogen peroxide generated by the glucose sensing reaction was evaluated, and no significant interference with the adjacent pH sensor was observed (Figure S2). These results support the electrochemical independence and anti-interference robustness of the integrated system. Additionally, no cross-talk was observed between the two monitoring functions. The glucose sensor signal remained stable even when the pH fluctuated, and vice versa (Figure S3). Under mechanical changes, such as repeated stretching and bending, the sensitivity change of PGFs was less than 3% (Figures S4–S5). Long-term stability tests showed that the sensor’s sensitivity decay was less than 5% after being immersed in physiological saline for 10 days, demonstrating its durability for clinical applications (Figures j–l), and meeting the long-term implantation requirements.

Efficacy and Biocompatibility In Vivo

To validate the reliability of the implanted device, we implanted PGFs into the subcutaneous tissue of mice, establishing a dynamic monitoring model (Figure a). In a glucose-insulin intervention experiment, the sensor accurately captured in vivo glucose concentration changes, with a 98% matching rate between the sensor response curve and data from glucose test strips (Figure b). Similarly, subcutaneous pH changes induced by intraperitoneal acidic liquid injection were monitored in real-time, with the signal returning to baseline after neutralization, confirming the system’s dynamic response capability. The twisted structure disperses stress and reduces local strain concentration in individual fibers, improving mechanical stability. Consequently, the impedance change rate was less than 1% under various movement states of the animals. Notably, the sensor retained 96% of its initial sensitivity 14 days after implantation, while pH and glucose sensors without the twisted structure retained only 85% of their initial sensitivity (Figure e, Figures S6–S7). This suggests that the twisted structure can effectively disperse external pressure through the interactions between fibers and the untwisting buffering effect. Furthermore, it enhances the sensor’s mechanical strength, improves wear resistance, stabilizes the contact between the sensor and surrounding tissues, and increases adaptability to the internal environment, thereby significantly improving stability in vivo.

3.

Performance and biocompatibility of PGFs in vivo. (a) Photographs of PGFs implanted in mice. (b–c) Real-time monitoring of in vivo glucose (b) and pH (c) changes, compared with clinical detection results. (d) Mechanical stability of the sensor under different motion states (sitting, walking, standing, running, and stopping). (e) Sensitivity retention of the glucose and pH sensors over a 14-day implantation period. (f–g) Blood routine indicators (f) and biochemical indicators (g) demonstrate the good biosafety of the implanted sensor. (h) hematoxylin and eosin (H&E) staining of skin tissue in the control group and 14 days after PGFs implantation (red: cytoplasm; blue: nucleus). (i) IL-1β immunofluorescence staining of skin tissue in the control group and 14 days after PGFs implantation (green: IL-1β; blue: nucleus).

Systematic biosafety assessments showed that routine blood indicators (e.g., white blood cell, red blood cell, and platelet counts) and liver and kidney function parameters in the experimental group were within normal ranges, with no significant differences compared to the control group (Figures f–g). Histopathological analysis further confirmed that there was no significant inflammatory cell infiltration or fibrotic hyperplasia at the implantation site, and quantitative analysis of IL-1β fluorescence intensity revealed no significant difference between the implanted and sham-operated groups (p > 0.05, Figures h–i; Figure S8). These findings collectively support the excellent in vivo biocompatibility of PGFs.

To further evaluate the advantages of PGFs, we compared them with clinically available blood glucose and pH sensors in septic mice. The survival rates of mice treated with clinical sensors were similar to those obtained with individual glucose or pH sensors, but remained significantly lower than those achieved with PGFs (Figure S8). This superiority of PGFs can be attributed to their integrated monitoring of pH and glucose, which provides more comprehensive metabolic regulation, as well as their lower Young’s modulus (Figure S9), which minimizes tissue mismatch and inflammatory responses. In contrast, commercial sensors induced more evident local inflammation at the implantation site, limiting their therapeutic efficacy. It should be noted that the operational lifetime of pH and glucose sensors, as well as commercial implantable continuous pH and glucose monitoring devices, is typically limited to 7–14 days, and long-term stability remains a common challenge in this field. Although the lifetime of our flexible sensor is comparable to current devices, its lower Young’ s modulus (Figure S9) enables better conformal contact with soft tissues, thereby reducing inflammation responses caused by mechanical mismatch (Figure S8). This advantage highlights its potential for improved safety and comfort in long-term clinical applications.

Dynamic Metabolic Dysregulation Profile in Sepsis Monitored by PGFs

Sepsis-induced metabolic dysregulation involves complex interactions between acid–base imbalances, energy metabolism dysfunction, and inflammatory responses. To gain deeper insights into the real-time metabolic state during sepsis, we utilized PGFs for continuous monitoring in an LPS-induced sepsis mouse model. The PGFs were connected to an external PCB for signal collection, and the data were transmitted wirelessly via Bluetooth for real-time monitoring (Video S1, Figures S10–S11). The monitoring results revealed a significant decline in pH, from 7.39 to 7.21 within 6 h, and further dropping to the lowest value of 7.05 at 72 h, with no signs of recovery (Figures a–b). This suggests that metabolic acidosis induced by LPS infection persists throughout the progression of sepsis, with the body’s natural regulatory mechanisms failing to restore balance.

4.

PGFs real-time monitoring and intervention efficacy evaluation in sepsis. (a) Sepsis was induced by LPS (lipopolysaccharide) injection, and PGFs were used to monitor 24-h pH and glucose level changes in response to the sepsis model. (b) Real-time monitoring of pH and glucose level changes in septic mice via PGFs, with the red and green curves representing the respective variations in glucose concentrations and pH over time. (c) Kaplan–Meier survival curves comparing survival rates among untreated mice, single-intervention groups, and the PGF treatment group (n = 10; **p < 0.0001, data are presented as mean ± s.d.). (d) Blood oxygen saturation (SpO₂) levels in untreated mice, single-intervention groups, and the PGF treatment group (n = 6; ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; data are presented as mean ± s.d.). (e) Lung tissue pathological injury scores assessed using the Smith scoring system in untreated mice, single-intervention groups, and the PGF treatment group (n = 6; ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; data are presented as mean ± s.d.). (f–g) H&E staining of lung tissue (f) and KIM1 immunofluorescence staining of kidney tissue (g) from untreated group, single-intervention group, and the PGF treatment group (n = 6; ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; data are presented as mean ± s.d.).

Simultaneously, PGFs detected glucose level fluctuations: within the first 12 h, blood glucose levels rose from 5.5 mmol/L to 15.2 mmol/L, followed by a gradual decline to 1.7 mmol/L, until the mice succumbed to the disease (Figure b). This pattern indicates that early inflammation-induced hyperglycemia was followed by hypoglycemia, likely due to diminished vitality, impaired gluconeogenesis, and abnormal glucose uptake. Taken together, these metabolic abnormalities highlight the complexity of sepsis-related metabolic dysregulation, with the existing regulatory mechanisms unable to timely restore metabolic balance.

In the LPS-induced sepsis model, our sensor effectively monitored disease progression and revealed that combined intervention markedly alleviated sepsis-associated organ injury. To further validate its clinical utility, we extended the evaluation to a cecal ligation and puncture (CLP) model, which mimics polymicrobial infection and is considered the gold standard for studying clinical sepsis. In this model, application of the sensor demonstrated that combined intervention significantly improved hepatic, pulmonary, and renal tissue damage, with reduced neutrophil infiltration as evidenced by myeloperoxidase (MPO) staining (Figure S12a). Quantitative analyses further confirmed reduced lung injury scores (Figure S12b), decreased MPO-positive cell infiltration in lung tissue (Figure S12c), and improved blood oxygen saturation (Figure S12d). Taken together, results from both the LPS and CLP models highlight the broad applicability of our sensor in monitoring and guiding therapeutic interventions across different sepsis settings, thereby underscoring its translational potential for clinical application.

PGF-Based Therapy Ameliorates Sepsis-Induced Organ Damage

PGFs not only provide real-time metabolic state monitoring but also serve as a tool to guide combined drug therapy and real-time metabolic management for sepsis. By using PGFs for combined intervention in sepsis mice, we assessed the impact of this therapeutic approach on disease progression. The results demonstrated that PGFs-based combination therapy significantly improved survival, increasing the survival time (49.0h to 56.5h) (Figure c). This further validates the potential of PGFs in sepsis treatment.

Organ damage is a major manifestation of sepsis, particularly in the lungs and kidneys. To further assess the therapeutic effects of PGFs-based combined intervention, we evaluated lung, kidney and liver damage in the mice (Figures S13–S14). Regarding lung injury, blood oxygen saturation, a key indicator of lung function, was 92.38 ± 0.66% in the untreated sepsis group, while the PGFs intervention group showed an improvement to 93.68 ± 0.33%, indicating improved lung function (Figure d). This improvement suggests that PGFs-based therapy alleviates lung injury and enhances oxygen exchange. Histological examination of lung tissue via HE staining confirmed this, with the PGFs intervention group showing significantly reduced inflammatory cell infiltration, alleviated alveolar wall thickening, and diminished inflammatory exudate. Furthermore, the composite lung injury score was notably lower in the PGFs-treated group compared to the untreated group (Figures e–f). These results indicate that PGFs-based therapy not only mitigates lung injury but also improves overall lung function.

Kidney injury is another critical manifestation of sepsis, primarily reflected by elevated serum creatinine and blood urea nitrogen (BUN) levels. Our findings show that PGFs-based intervention effectively lowered serum creatinine and BUN levels in mice (Figures S15–S16), suggesting significant improvement in sepsis-induced kidney damage. Further analysis of renal tissue using PAS staining and KIM1 immunofluorescence revealed that PGFs-based intervention significantly alleviated glomerular and tubular damage induced by sepsis (Figure g, Figure S17). These findings suggest that PGFs-based therapy can attenuate sepsis-induced kidney damage in an LPS-induced model by modulating metabolic pathways. Together, the restoration of endothelial cell function, suppression of inflammatory mediator release, and reduction of apoptosis indicate that PGFs-based therapy may be an effective strategy for mitigating organ damage caused by sepsis. −

Mechanism of PGFs-Based Therapy in Alleviating Sepsis-Induced Damage

Immune cell infiltration and the resulting inflammatory storm are widely recognized as central contributors to the progression of sepsis. Recent studies employing multifunctional nanozyme-based and nucleic acid–based strategies, such as multienzyme-active Au/CeO₂ nanozymes, antioxidant catalytic nanoplatforms, DNA nanostructure–mediated miRNA delivery systems, and MBL-functionalized Au nanoparticles, have demonstrated that regulating oxidative stress and inflammatory cytokine networks can effectively mitigate sepsis-induced immune dysfunction. − To further elucidate the underlying mechanisms of PGFs-based therapy, we systematically investigated changes in immune cell dynamics and inflammatory signaling pathways. Immunofluorescence staining of lung tissues revealed a significant reduction in the infiltration of macrophages (F4/80-positive) and neutrophils (MPO-positive) in the PGFs intervention group compared with untreated and monotherapy groups (Figure a, Figures S18–S19). This suggests that PGFs-based therapy effectively inhibits immune cell infiltration, particularly in the lungs, one of the most affected organs in sepsis.

5.

PGFs intervention shows better inhibition of the MAPK pathway compared to individual interventions. (a) Myeloperoxidase (MPO) and F4/80 immunofluorescence demonstrated decreased neutrophil infiltration in lung tissues in PGFs-based intervention group (red: MPO or F4/80; blue: nucleus). (b) Western blot results demonstrate that combined intervention effectively inhibits the expression of MAPK pathway downstream markers pP38 and pERK. (c) mRNA quantitative analysis shows that combined intervention significantly outperforms individual interventions in suppressing the downstream signaling of the MAPK pathway (n = 6, mean ± s.d.; ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Activation of inflammatory pathways plays a pivotal role in the inflammatory storm induced by sepsis and can further enhance the immune response. − Among them, the MAPK inflammatory signaling pathways, including P38 and ERK, are core pathways in the pathogenesis of sepsis. Metabolic acidosis and glucose dysregulation are both can independently activate MAPK signaling cascades through distinct mechanisms. Acidic pH can alter intracellular enzyme activity and redox homeostasis, thereby promoting MAPK phosphorylation and proinflammatory gene expression. Similarly, glucose fluctuation has been shown to activate the MAPK pathway through effects on glycolysis and stress response signaling. MAPK signalingparticularly the P38 and ERK axesrepresents a central regulator of immune cell activation and inflammatory cytokine release in the context of sepsis. , Given that immune cell accumulation is a well-known upstream driver of MAPK pathway activation, we hypothesized that PGFs intervention alleviates sepsis-induced inflammation by suppressing immune cell infiltration and downregulating MAPK signaling. To test this, we assessed the phosphorylation status of ERK and P38, along with the expression levels of downstream inflammatory mediators such as IL-6, TNF-α, JUNB, and JUND. Western blot analysis revealed that PGFs treatment led to significant reductions in both p-ERK and pP38 levels, as well as decreased expression of inflammatory cytokines (Figures b–c).

To further clarify whether the anti-inflammatory effects were due to individual or synergistic contributions of pH normalization and glucose supplementation, we performed parallel experiments with separate intervention groups. Notably, while both individual treatments elicited moderate reductions in MAPK activation and cytokine production, the PGFs combined intervention exhibited a significantly enhanced suppressive effect that exceeded the theoretical additive sum of the individual interventions. These results suggest a synergistic interaction between pH and glucose regulation in dampening inflammatory signaling, potentially due to the interconnected nature of acid–base and glucose metabolism in sepsis. By concurrently correcting both disturbances, PGFs disrupts the positive feedback loop linking metabolic stress and inflammatory amplification.

In conclusion, PGFs-based therapy alleviates sepsis-induced inflammation by suppressing immune cell infiltration and modulating key inflammatory signaling pathways, especially the P38/ERK axis. The observed synergistic effect between pH and glucose regulation highlights the therapeutic potential of integrated metabolic interventions. PGFs-based modulation thus offers a promising strategy to break the vicious cycle of metabolic and inflammatory dysfunction in sepsis.

Conclusion

This study presents the development of a twisted structure-integrated multifunctional fiber sensor (PGFs) that combines pH and glucose sensing capabilities. This sensor enables minimally invasive, real-time, and high spatiotemporal resolution simultaneous detection of pH and glucose levels. The twisted integration structure of PGFs enhances the stability and safety of long-term in vivo monitoring through fiber interactions. Real-time monitoring using PGFs in a sepsis animal model revealed dynamic fluctuations in pH and glucose, as well as their spatiotemporal correlation. Our findings demonstrate that the dual management of pH and glucose with PGFs significantly improves sepsis prognosis by inhibiting the activation of inflammatory signaling pathways. Employing a “structure-empowered functionality” design paradigm, this work provides a multifunctional platform for dynamic multiparameter analysis in complex pathological environments. It also offers valuable insights for sepsis monitoring and treatment. Looking ahead, the integration of wireless power supply and closed-loop feedback systems holds promise for the autonomous management of sepsis-induced metabolic dysregulation, advancing critical care medicine toward a ″dynamic precision″ therapeutic strategy.

Materials and Methods

Materials and Chemicals

All chemicals and reagents were of analytical grade and used without further purification. Potassium hexachloroplatinate (K₂PtCl₆, 98.0%), aniline (≥99.5%), chitosan (200–400 mPa·s), and silver nitrate (AgNO₃, 99.8%) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Glucose oxidase (GOx, 30.0 U·mg^–1), D-(+)-glucose (99.5%), and Nafion (5 wt % in a mixture of lower aliphatic alcohols and water) were obtained from Sigma-Aldrich. Phosphate-buffered saline (PBS, pH 7.4), potassium chloride (KCl, ≥ 99.5%), potassium nitrate (KNO₃, ≥ 99.0%), and ethanol (≥99.0%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Acetic acid (AR) and glutaraldehyde (25.0–28.0% in water) were obtained from Shanghai Meryer Chemical Technology Co., Ltd. Sulfuric acid (H₂SO₄, 95.0%–98.0%) was purchased from Yonghua Chemical Co., Ltd. Hydrogen chloride (HCl, AR) was purchased from Nanjing Chemical Reagent Co., Ltd. Polydimethylsiloxane (PDMS, ELASTOSIL RT 601 A/B) was obtained from WACKER Chemie AG (Munich, Germany). Multiwalled carbon nanotubes (MWCNT, 100253) and Au nanoparticle solution (0.1 mg·mL^–1) were purchased from Nanjing XFNANO Materials Tech Co., Ltd. Ag/AgCl ink was obtained from Guangzhou City Silver Well Trading Co., Ltd. Sodium bicarbonate (NaHCO₃) was purchased from Sigma-Aldrich (S8875).

Instruments

A scanning electron microscope (SEM, Zeiss Ultra 55) was used to characterize the structure and morphology of the samples. Electrochemical properties were analyzed using a CHI660e electrochemical workstation (CH Instruments Inc.). Images of stained tissue sections were captured using a Nikon Alphaphot-2 microscope, coupled with a NIKON ECLIPSE C1 imaging system and a NIKON DS-U3 scanner.

Device Fabrication

Synthesis of CNT Fibers

Carbon nanotube (CNT) fibers were synthesized using a floating catalyst method, employing thiophene and ferrocene as catalysts, with ethanol serving as the carbon source. CNT aerogels were continuously produced at a reaction temperature of 1200 °C in the hot zone of a tubular furnace. The aerogels were collected by pulling them through a cold water bath and then densified through sequential washing with water and ethanol to form a ribbon-like structure. The CNT ribbons were further densified by washing with acetone, dried, twisted, and wound onto a spool to produce CNT fibers.

Preparation of pH Fiber Sensors

pH fiber sensors were fabricated by electrodepositing polyaniline (PANI) onto modified CNT fibers using cyclic voltammetry in a 0.1 M aniline/0.1 M H₂SO₄ solution. The deposition was conducted with a scan rate of 100 mV·s^–1, cycling between – 0.2 and 1 V for 25 consecutive cycles.

Preparation of Glucose Sensing Fibers

In a three-electrode system, PANI was deposited onto CNT fibers via enhanced electrochemical polymerization at 0.75 V for 20 s in a 0.5 M aniline/1 M H₂SO₄ solution. The fibers were then subjected to 50 cycles of potential steps in a 0.1 M KCl/1 mM K₂PtCl₆ aqueous solution, with each cycle consisting of 10 s at 0.5 V and 10 s at – 0.7 V. Chitosan was dissolved in a 2 wt % acetic acid aqueous solution and magnetically stirred for 1 h to form a 1 wt % chitosan solution. Glucose oxidase (40 mg·mL^–1) and MWCNTs (2 mg·mL^–1) were dispersed in the chitosan solution under ultrasonic treatment for 30 min. A 4 μL mixture of MWCNTs/chitosan/glucose oxidase was then dip-coated onto the fibers. Finally, a 2 μL Nafion (0.5 wt %) coating was applied, and the fibers were immersed in a 2 wt % glutaraldehyde solution for 30 min before being dried overnight at 4 °C.

Preparation of Fiber-Based Ag/AgCl Reference Electrodes

Fiber-based Ag/AgCl reference electrodes were fabricated using cyclic voltammetry. The CNT fibers were immersed in a mixed aqueous solution of 5 mM AgNO₃ and 1 M KNO₃ and scanned from – 0.9 to 0.9 V for 7 cycles at a scan rate of 0.1 V·s^–1, using a two-electrode system. The CNT fiber served as the working electrode, and a commercial silver electrode was used as the counter and reference electrode. Subsequently, the silver/CNT fiber was immersed in a mixed aqueous solution of 0.1 M KCl and 0.01 M HCl and scanned from – 0.15 to 1.05 V for 2 cycles at a scan rate of 0.05 V·s^–1. The fiber was then dried to obtain the silver/AgCl reference fiber electrode.

Fabrication of PGFs

The pH and glucose fiber sensors were integrated with Ag/AgCl fiber electrodes to fabricate the PGFs. Initially, the fiber sensors and Ag/AgCl reference electrodes were coated with PDMS (prepared by mixing the base and cross-linker in a 10:1 weight ratio) as an insulating layer, and cured at 80 °C for 30 min. The ends of all fiber electrochemical sensors were aligned and placed parallel to each other. One end was fixed to a rotating motor shaft, while the other end was secured with tape. The motor was operated at 200 rad·min^–1 to form a helical structure. Afterward, the helical structure was stabilized by applying an additional layer of PDMS, resulting in PGFs with multiple detection functions at the same location.

Characterization and Performance of PGFs

The morphology of the PGFs was characterized using a scanning electron microscope (SEM, Hitachi FE-SEM S-4800). The electrochemical sensing performance of the PGFs was evaluated in phosphate-buffered saline (PBS) by studying their response to different analyte solutions. The concentration of the analyte solution was varied by rapidly adding the analyte to the stirred solution. For glucose sensing, amperometric I-t curves were recorded at 0.65 V. For pH sensing, real-time open-circuit potential measurements were taken. The obtained current and potential values were calibrated by subtracting the baseline values measured in a blank solution. The selectivity of the sensors was assessed by measuring the current or potential response of the target electrolyte in the presence of various interfering substances. All electrochemical measurements were performed using a CHI660e electrochemical workstation.

Animal Experiments

Animals

Male C57BL/6 mice (8–10 weeks old) were obtained from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China; Certificate number: SYXK-Hu-2022–0012). All experimental procedures were approved by the Animal Ethics Committee of Central South University (Approval number: CSU-2023–0034) and conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Mice were housed under specific pathogen-free (SPF) conditions with a 12-h light/dark cycle, a controlled temperature of 22 ± 1 °C, and ad libitum access to standard chow and water.

Animal Model Induction

Lipopolysaccharide (LPS, derived from Escherichia coli O111:B4 strain, Sigma catalog number L4391) was dissolved in physiological saline to a concentration of 10 mg/kg for the LPS model and 30 mg/kg for the survival analysis model. Prior to the experiment, the animals were fasted for 12 h to minimize external interference. For the LPS model, the solution was administered via intraperitoneal injection (i.p.) at a dose of 10 mg/kg. After injection, the animals were monitored for 24 h, with physiological changes such as body temperature, behavior, and overall health status recorded at regular intervals. At the end of the 24-h observation period, blood and organ samples were collected for subsequent analysis.

To construct the cecal ligation and puncture (CLP) model, the mice were first anesthetized using either isoflurane or pentobarbital sodium. Prior to surgery, the animal’s abdominal area was disinfected with iodine solution. A 1–2 cm incision was made in the abdominal wall to expose the abdominal cavity. A fine needle was used to puncture the cecum twice, creating small perforations to simulate bacterial peritonitis. After the punctures, the abdominal cavity was quickly sutured, ensuring sterile techniques to prevent infection.

Histopathological Analysis of Lung and Kidney Tissues

Mice were euthanized 24 h post-LPS administration, followed by systemic perfusion with 10–20 mL of ice-cold phosphate-buffered saline (PBS). Lung and kidney tissues were immediately collected and fixed in 4% paraformaldehyde (pH 7.4) for 24 h. Lung tissues were fixed without inflation. The left kidney was sliced into 1–2 mm transverse sections. Samples were dehydrated in a graded ethanol series (70% → 80% → 95% → 100%), cleared in xylene, embedded in paraffin, and sectioned at 4 μm thickness. Sections were mounted on adhesive glass slides, baked at 60 °C for 2 h, and stained with hematoxylin and eosin (H&E). The staining procedure included xylene dewaxing, ethanol rehydration, hematoxylin nuclear staining (5 min), eosin cytoplasmic staining (2 min), dehydration, and mounting with neutral resin.

Lung injury was evaluated using the Smith scoring system, assessing four parameters: pulmonary edema, alveolar/interstitial inflammation, hemorrhage, and atelectasis/hyaline membrane formation. Each parameter was scored from 0 (no lesion) to 4 (>75% of the field affected), for a total score range of 0–16. Renal injury was scored semiquantitatively based on tubular necrosis, luminal dilation, and inflammatory infiltration (each 0–4, total score 0–12). All assessments were performed independently by two pathologists blinded to group allocation.

Survival Rate Monitoring

Following LPS challenge, mice were observed for 72 h. Clinical signs (activity level, respiratory distress, body weight) were recorded hourly during the first 24 h and every 3 h thereafter. Survival rates were calculated per group (n = 6 mice/group) and analyzed using Kaplan–Meier survival curves with Log-rank test for intergroup comparison. All protocols adhered to institutional ethical standards and the 3R principles (Replacement, Reduction, Refinement).

Immunofluorescence Staining

After anesthesia with 5% isoflurane, mice were euthanized, and tissues were harvested, fixed in 4% paraformaldehyde for 24 h, paraffin-embedded, and sectioned at 4 μm thickness. Sections were processed for immunofluorescence staining using specific primary antibodies and appropriate fluorescent secondary antibodies. Nuclei were counterstained with DAPI. Fluorescence images were acquired using a fluorescence microscope, and marker expression levels were quantified using image analysis software.

Periodic Acid-Schiff (PAS) Staining

Kidney tissues were fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned at 4 μm. Sections were deparaffinized in xylene and rehydrated through graded ethanol (80%, 95%, and 100%). Oxidation was carried out with 0.5% periodic acid (10 min), followed by staining with PAS solution (20 min). Slides were rinsed in distilled water, counterstained with hematoxylin (15 s), differentiated in 1% acid alcohol (3 dips), and blued in ammonia–water. After dehydration and xylene clearing, sections were mounted with neutral resin. Glomerular basement membrane thickness and tubular damage were assessed under a light microscope (Olympus BX53) at 20× magnification.

Serum Urea Nitrogen and Creatinine Assay

Blood was collected via orbital puncture, and serum was isolated by centrifugation at 3,000 × g for 10 min and stored at – 80 °C. Urea nitrogen levels were determined using a urease–glutamate dehydrogenase method (kit C013–2, Nanjing Jiancheng Bioengineering Institute), based on NADH absorbance at 340 nm. Creatinine levels were measured using an enzymatic colorimetric assay (kit C011–2, Nanjing Jiancheng), relying on the Jaffe reaction with absorbance at 510 nm. All assays were conducted using a Hitachi 7150 automatic biochemical analyzer.

Murine SpO₂ Monitoring

Noninvasive oxygen saturation (SpO₂) was monitored using a dedicated murine pulse oximeter with miniaturized sensors. Mice were acclimated for ≥ 30 min to minimize stress-related hypoxia. When necessary, 1.5%–2% isoflurane was used to maintain stable respiration. Sensor placement was adjusted based on consciousness, targeting hairless areas (e.g., tail, hindlimb, cervical region). A sampling frequency ≥ 300 Hz was used to detect transient fluctuations. Calibration was performed using reference gas mixtures (95% O₂/5% CO₂), and accuracy was validated against arterial blood gas results, yielding an r² > 0.9.

Statistical Analysis

The statistical analysis was conducted with Origin and Excel. The results are presented as the mean ± s.d..

The data sets supported by this study are available from the corresponding authors upon request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.5c14776.

• Selectivity of PGFs glucose sensors and PGFs pH sensors; normalized impedance under stretching and bending; in vivo sensitivity stability over 14 days; inflammatory response, survival outcomes, and mechanical properties of PGFs; circuit schematic and list of electronic components of the readout system; H&E staining of liver and kidney tissues; serum biochemistry (BUN and creatinine); PAS staining of glomeruli; fluorescence analysis of F4/80, MPO, PERK, and pP38 in lung tissues (PDF)

• Real-time in vivo monitoring of the PGFs system implanted in mice (MP4)

∇.

S.Y. and Y.J. contributed equally to this work. S.Y.: conceptualization, methodology, investigation, writing–original draft, writing–review and editing. Y.J.: methodology, investigation, funding acquisition, writing–review and editing. Y.L.: investigation, methodology, conceptualization, investigation. Z.Z.: supervision, conceptualization, methodology, writing–review and editing. T.L.: supervision, funding acquisition, writing–review and editing.

The authors declare no competing financial interest.