Early label-free analysis of mitochondrial redox states by Raman spectroscopy predicts septic outcomes

Abstract

Background: Sepsis remains an unacceptably high mortality due to the lack of biomarkers for predicting septic outcomes in the early period. Mitochondrial redox states play a pivotal role in this condition and are disturbed early in the development of sepsis. Here, we hypothesized that visualizing mitochondrial redox states via resonance Raman spectroscopy (RRS) could identify septic outcomes at an early time point. Sepsis was induced by cecal ligation and puncture (CLP). We applied RRS analysis at baseline and 30 min, 1 h, 2 h, 4 h, and 6 h after CLP, and the mitochondrial redox states were identified. The levels of blood lactate as a predictor in sepsis were assessed. Our study is the first to reveal the possibility of in vivo detection of the mitochondrial redox state by using RRS in septic mice. The peak area for the Raman reduced mitochondrial fraction, the indicator of mitochondrial redox states, fluctuated significantly at 2 h after CLP. This fluctuation occurred earlier than the change in lactate level. Moreover, this fluctuation had higher prognostic accuracy for mortality than the lactate level during sepsis and could be a novel diagnostic marker for predicting septic outcomes according to the cutoff value of 1.059, which had a sensitivity of 80% and a specificity of 90%.

Objectives: To explore an effective indicator concerning mitochondrial redox states in the early stage of sepsis and to predict septic outcomes accurately in vivo using non-invasive and label-free Resonance Raman spectroscopy (RRS) analysis.

Methods: Mitochondria, primary skeletal muscle cells andex-vivo muscles harvested from gastrocnemius were detected mitochondrial redox states respectively by using RRS. Sepsis was induced by cecal ligation and puncture (CLP). We applied RRS analysis at baseline and 30 min, 1 h, 2 h, 4 h, and 6 h after CLP, and the mitochondrial redox states were identified. The levels of blood lactate as a predictor in sepsis were assessed. The predictive correlation of mitochondrial redox states on mortality, inflammation and organ dysfunction was further assessed.

Results: Mitochondrial redox states were clearly recognized in ex-vivo gastrocnemius muscles as well as purified mitochondria and primary skeletal muscle cells by using RRS. The peak area for the Raman reduced mitochondrial fraction, the indicator of mitochondrial redox states, fluctuated significantly at 2 h after CLP. This fluctuation occurred earlier than the change in lactate level. Moreover, this fluctuation had higher prognostic accuracy for mortality than the lactate level during sepsis and could be a novel diagnostic marker for predicting septic outcomes according to the cutoff value of 1.059, which had a sensitivity of 80% and a specificity of 90%.

Conclusions: This study demonstrated that monitoring mitochondrial redox states using RRS as early as 2 h could indicate outcomes in septic mice. These data may contribute to developing a non-invasive clinical device concerning mitochondrial redox states by using bedside-RRS.

Keywords: Inflammatory response; Label-free; Mitochondrial redox state; Multiple organ failure; Resonance Raman spectrum; Sepsis.

Graphical abstract

Fig. 1

Resonance Raman spectra during manipulation of the mitochondrial redox state, (A) Raman spectrum of (1) mitochondria, (2) primary skeletal muscle cells and (3) ex vivo gastrocnemius muscles before and after adding the strong reductant sodium dithionate (SDT). Blue lines represent reduced species (SDT group), and red lines represent normal species (control group). (B) Effect of redox reaction on the intensity of the mitochondrial peaks at 750, 1128 and 1585 cm⁻¹ in (1) mitochondria, (2) primary skeletal muscle cells and (3) ex vivo gastrocnemius muscles. n = 3 per group. (C) Comparisons of the NAD⁺ and NADH levels and the NAD⁺/NADH (n = 5 per group) and GSH to GSSG ratios (n = 3 per group) in mitochondria between the control and SDT groups. Data are expressed as the mean ± SEM and analyzed by Student’s t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared with the control group. NAD⁺ = oxidized nicotinamide adenine dinucleotide, NADH = reduced nicotinamide adenine dinucleotide, GSSG = oxidized glutathione, GSH = reduced glutathione, ns = no significance.

Fig. 2

Changes in mitochondrial redox states were detected early using resonance Raman spectroscopy in vivo, (A) The survival rate was monitored between the CLP group and the sham group. n = 16–19 mice per group. (B) Levels of serum lactate at baseline (1 h before CLP) and 2 h, 6 h, 12 h and 24 h after CLP. n = 7 mice per group. (C) Peak area for the Raman reduced mitochondrial fraction (PA-2RMF) at the different time points (baseline, 30 min, 1 h, 2 h, 4 h, 6 h after CLP). n = 10 mice per group, 0.5 h = 30 min. (D) Skeletal muscle position with respect to the objective and laser light for Raman spectrum measurement in CLP-induced sepsis in vivo. The mouse was fixed in a warmed black pad (37 °C). The Raman spectra were collected at baseline (1 h before CLP) and 30 min, 1 h, 2 h, 4 h and 6 h after CLP. Light field (scale bars, 10 μm). (E)-(G) Correlation between the ratio of PA-2RMF from 2 h after CLP to baseline (RPA-2RMF_2h) and the NAD⁺, NAD⁺/NADH or GSH/GSSG levels in the gastrocnemius muscle of septic mice. n = 16 mice per group. Data are expressed as the mean ± SEM. Comparisons between groups were analyzed by one-way ANOVA with Dunnett's multiple comparison post-test. The Pearson correlation coefficient was used to assess the correlation between PA-2RMG_2h and mitochondrial redox status parameters. *p < 0.05, **p < 0.01 compared with the baseline group. CLP = cecal ligation and puncture, NAD⁺ = oxidized nicotinamide adenine dinucleotide, NADH = reduced nicotinamide adenine dinucleotide, GSSG = oxidized glutathione, GSH = reduced glutathione.

Fig. 3

Performance of mitochondrial redox states using resonance Raman spectroscopy for septic mortality, (A)-(D) Correlation between the ratio of PA-2RMF from 2 h after CLP to baseline (RPA-2RMF_2h) and the parameters of sepsis including blood serum lactate, high mobility group box 1 (HMGB1), procalcitonin (PCT), and C-reactive protein (CRP) at 24 h post CLP. n = 23–29 mice per group. Pearson or Spearman correlation coefficients were used to assess the correlation between PA-2RMF_2h and these parameters of sepsis. (E) Receiver operating characteristic curve of RPA-2RMF_2h (solid red, AUC = 0.88, n = 45) as a diagnostic test for predicting mortality in sepsis. (F) Sensitivity (dotted red) and specificity (solid black) of the RPA-2RMF_2h values for predicting mortality in sepsis within 7 days after CLP. CLP = cecal ligation and puncture, AUC = area under the curve.

Fig. 4

Measurement of inflammatory cytokines and bacterial loading associated with mitochondrial redox states in septic mice, (A)-(C) Changes in the body weight, rectal temperature and blood systolic pressure in the H group and the L group. (D) Colony-forming units (CFUs) in the peritoneal lavage were counted 24 h post CLP in mice from the H group and the L group. n = 3 or 5 per group. (E) Concentrations of the inflammatory cytokines interleukin (IL)-6, IL-10 and tumor necrosis factor (TNF)-α in the serum samples at baseline (1 h before CLP) and 24 h post CLP in the H group and the L group. n = 5–6 per group. Data are expressed as the mean ± SEM and were analyzed by Student’s t test or two-way ANOVA with multiple-comparison testing. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared with the baseline group, ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 compared with the H group at 24 h post CLP. H group = the ratio of PA-2RMF from 2 h after CLP to baseline (RPA-2RMF_2h) was at or exceeded 1.0059, L group = RPA-2RMF_2h was below 1. 0059, ns = no significance, CLP = cecal ligation and puncture.

Fig. 5

Levels of high mobility group box 1 (HMGB1), C-reactive protein (CRP) and procalcitonin (PCT) at baseline and 24 h post CLP in the H group and the L group. n = 7 or 8 per group. Data are expressed as the mean ± SEM and were analyzed by Student’s t test or two-way ANOVA with multiple-comparison testing. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared with the baseline group, ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 compared with the H group at 24 h post CLP. H group = the ratio of PA-2RMF from 2 h after CLP to baseline (RPA-2RMF_2h) was at or exceeded 1.0059, L group = RPA-2RMF_2h was below 1. 0059, ns = no significance, CLP = cecal ligation and puncture.

Fig. 6

Organ function assessment associated with mitochondrial redox states in septic mice, (A) Representative hematoxylin-eosin sections of the lung, heart, kidney and liver in the H group and L group are shown at ×200 original magnification, and the scale bars are 20 μm. In lung histology, the red arrow indicates alveolar congestion and infiltrated inflammatory cells in the alveoli, the blue arrow indicates alveolar wall thickening, and the black arrow indicates enlarged interstitial space. In heart sections, the red arrow indicates coagulative necrosis, the blue arrow indicates interstitial lymphocyte infiltration, and the black arrow indicates lytic necrosis. The kidney tissues demonstrated leukocyte infiltration in glomeruli (black arrow), glomerular destruction (red arrow) and tubular proteinaceous casts (blue arrow). Histologic analysis of the liver showed disappearance of the sinus hepaticus (red arrow), hepatocyte vacuolization (black arrow) and cell necrosis (blue arrow). (B) Quantification of the lung, heart, kidney and hepatic injury scores are presented. n = 3 per group. (C) Lung wet-to-dry weight ratio analysis. (D), (E) Serum marker levels of organ injury at baseline and 24 h post CLP in the H group and the L group were analyzed with the following: (D) cardiac troponin I (cTnI); (E) blood urea nitrogen (BUN), creatinine (Cre), alanine aminotransferase (ALT) and aspartate aminotransferase (AST). n = 5–7 per group. Data are expressed as the mean ± SEM and were analyzed by Student’s t test or two-way ANOVA with multiple-comparison testing. *p < 0.05, **p < 0.01, ***p < 0.001 compared with the baseline group, ^#p < 0.05, ^##p < 0.01 compared with the H group at 24 h post CLP. H group = the ratio of PA-2RMF from 2 h after CLP to baseline (RPA-2RMF_2h) was at or exceeded 1.0059, L group = RPA-2RMF_2h was below 1. 0059, ns = no significance, CLP = cecal ligation and puncture. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Supplementary Fig. 1

Original resonance Raman spectra. (A) Original Raman spectra of (1) mitochondria, (2) primary skeletal muscle cells and (3) ex vivo gastrocnemius muscles before and after adding the strong reductant sodium dithionate (SDT). Blue lines represent reduced species (SDT group), and red lines represent normal species (control group). (B) Original Raman spectra were collected at baseline (1 hour before CLP) and 30 minutes, 1 hour, 2 hours, 4 hours and 6 hours after CLP insult. CLP = cecal ligation and puncture.

Supplementary Fig. 2

Serum lactate was collected at 2 hours after cecal ligation and puncture (CLP) for septic outcomes. In (A) to (D), correlations between the serum lactate level at 2 hours after CLP (blood lactate_2h) and parameters of sepsis, including serum lactate, high mobility group box 1 (HMGB1), procalcitonin (PCT), and C-reactive protein (CRP), at 24 hours post CLP. n = 10 mice per group. The Pearson correlation coefficient was used to assess the correlation between the blood lactate_2h value and the parameters of sepsis. (E) Receiver operating characteristic curve of blood lactate_2h (dotted red, AUC 0.621, n = 45) as a diagnostic test for predicting mortality in sepsis. AUC = area under the curve.