Severe viral respiratory infections such as the severe acute respiratory syndrome and the H1N1 influenza virus pneumonia that have emerged in the past few years (1, 2) create strain to critical care services and are associated with a high mortality rate. For unknown reasons, some patients with H1N1 influenza virus pneumonia experienced a mild form of the disease, whereas others developed severe acute respiratory distress syndrome (ARDS) with mortality rates between 17% and 46% (2, 3).
ARDS is characterized by lung inflammation and hyperpermeability pulmonary edema. Currently, the diagnosis of ARDS is based on the presence of clinical, physiological, and radiological criteria (4–6). Unlike other clinical conditions, to date there are no specific molecular markers that help in the prognosis of this condition. Advancement in the understanding of the pathogenesis of ARDS is necessary for designing innovative and effective therapeutic approaches.
Metabolomic analysis has emerged in the last few years as a potentially useful tool for the discovery of novel lung disease biomarkers (7, 8). Nuclear magnetic resonance (NMR)-based metabolic approach has been used for the diagnosis of sepsis and acute lung injury (ALI) in experimental models (9–12) as well as to predict outcome in patients with trauma (13). Metabolomic analysis has also been used in childhood pneumonia to improve diagnosis (14). Finally, a metabolomic approach has been reported to detect metabolic alterations during influenza virus infection in mice (15, 16).
The aim of the current pilot study is the discovery of metabolic biomarkers of ARDS in patients with H1N1 influenza virus pneumonia using NMR spectroscopy. The identification of metabolic pathways involved in ARDS will improve our understanding of the pathogenesis of this condition. In addition, it will underpin the proof of concept that NMR spectroscopy could be a useful tool to define metabolomic patterns of prognostic significance in ARDS (e.g., patterns predictive of the development of the syndrome in at-risk patients).
We analyzed serum samples obtained during the 2009 pandemic H1N1 influenza virus pneumonia in Hospital Universitario de Getafe, Madrid, Spain (derivation set) and Hospital del Mar, Barcelona, Spain (validation set). Sample size for the derivation set was estimated using the MetSizeR method for metabolomic analysis (17), setting a False Discovery Rate of 0.05. Sample size for the validation set was calculated to assure less than 5% measurement uncertainty in the classification model. We obtained an optimal sample size of 28 for the derivation set and 20 for the validation set, aiming at a sample size of 30 and 26, respectively, to account for an approximate outlier rate of 10%.
Patients were admitted to the medical floor or to the intensive care unit (ICU), as per the attending physician judgment. Inclusion criteria were: age ≥ 18 years, and diagnosis of confirmed 2009 pneumonia by influenza A (H1N1) virus infection. Exclusion criteria were lack of informed consent.
A blood sample was obtained on the day after admission (either to the floor or to the ICU), that occurred always within 24 h of presentation to the Emergency Department, and serum was frozen at −80° until analysis. Clinical information was obtained by retrospective chart review, and data pertaining the requirement of mechanical ventilation, the diagnosis of ARDS, the Sequential Organ Failure Assessment (SOFA), and the Simplified Acute Physiologic Score-II (SAPS II) scores on admission (SAPS II was not available in the derivation set), the presence of renal or cardiovascular failure (SOFA score of the respective component > 2) (18) and status at hospital discharge (hospital mortality) were collected. The study was approved by the Ethics‘ Committee of the participating hospitals.
H1N1 influenza virus pneumonia was defined as per the World Health Organization (Centres for Disease 2009). H1N1 infection was confirmed by real-time reverse transcription-polymerase chain reaction or either nasopharyngeal swab samples or tracheal secretions. ARDS was diagnosed according to the American European Consensus Conference (19).
NMR data acquisition
Serum samples (40 μL of serum) were examined using high-resolution magic angle spinning (HR-MAS) NMR operating at 4°C to reduce metabolic degradation. HR-MAS NMR was performed at 500.13 MHz using a Bruker AMX500 spectrometer 11.7 T. Samples were placed into a 50 μL zirconium oxide rotor using a rinsed with a cylindrical insert, together with 15 μL of 0.1 mM solution of trimethylsilylpropanoic acid in deuterium water, and spun at 4,000 Hz spinning rate to remove the effects of spinning side bands from the acquired spectra. A number of bidimensional homonuclear and heteronuclear experiments such as standard gradient-enhanced correlation spectroscopy, 1H–1H total correlated spectroscopy, and gradient-selected heteronuclear single quantum correlation protocols were performed to carry out component assignments. Between consecutive two-dimensional (2D) spectra, a control 1H NMR spectrum was always measured. No gross degradation was noted in the signals of multiple spectra acquired under the same conditions. Standard solvent-suppressed spectra were grouped into 32,000 data points, averaged over 256 acquisitions. The data acquisition lasted in total 13 min using a sequence based on the first increment of the nuclear Overhauser effect spectroscopy pulse sequence to effect suppression of the water. Sample acquisitions were performed using spectral width of 8333.33 Hz prior to Fourier transformation, and the free induction decay signals were multiplied by an exponential weight function corresponding to line broadening of 0.3 Hz. Spectra were referenced to the trimethylsilylpropanoic acid singlet at 0 ppm chemical shift.
NMR data were processed as previously described (10). Briefly, NMR spectra were data-reduced to equal length integral segments (δ = 0.01 ppm) and they were normalized to total sum of the spectral regions.
Quantitative and qualitative variables were compared by the Student t test or the chi square test, respectively. A P value less than 0.05 was considered statistically significant. The statistical package SPSS IBM Statistics 19.0 was used for the analysis. Descriptive data are presented as mean (SD) for continuous variables, and as percentage for discrete variables.
Principal components analysis (PCA) (20) was applied to H1N1 derivation set to extract the most discriminative spectral subset from the total pool of metabolites. PCA is the fundamental method used in chemometrics, where the metabolic data collected on a set of samples are resolved into principal components. The first principal component is defined by the spectral profile (loading) in the data that describes most of the variation; the second principal component, orthogonal to the first, is the second-best profile describing the variation, and so on. The principal components are composed of so-called scores and loadings. Loadings contain information about the variables (NMR chemical shifts) in the dataset, and scores hold information on samples (concentrations) in the dataset. The data obtained from this analysis were centered and scaled. The potential biomarkers selected from the PCA loading matrix were confirmed by the Hotteling T2 test (21).
Multivariate statistical computing was performed with the Metabonomic package (rel.3.3.1) (22). In addition, the resonances identified as significantly different by the Hotteling T2 test were individually integrated for metabolic quantification using the Global Spectral Deconvolution algorithm of MestRenova v. 8.1 (Mestrelab Research S.L., Santiago de Compostela, Spain). For the metabolic quantification, statistical significance was determined using a Bonferroni corrected Student t test (23) assuming unequal variance with P < 0.05 considered significant.
PLS discriminant analysis (PLS-DA) was developed as classificatory model. PLS-DA models are commonly used classification methods for analyzing high-dimensional data. We have used the algorithm proposed by Ding and Gentleman (24). The classification functions derived from the probability of belonging to the ARDS or to the non-ARDS group were trained with the respective patients with H1N1 influenza virus pneumonia from the Hospital Universitario de Getafe. The number of PLS components used was chosen by the percentage of variance explained, the R2, and the mean squared error of cross-validation graphics. The classification functions were used afterward to classify cases (ARDS vs. non-ARDS) with H1N1 influenza virus pneumonia from Hospital del Mar. The percentages of correct classification were calculated as a measure of the model performance. PLS classification functions were also correlated with classical ICU‘s predictors such as SOFA, SAPS II, Acute Physiology and Chronic Health Disease Classification System II (APACHE II), and P02/FI02 ratio
Characteristics of study patients and laboratory findings on admission
Patients with H1N1 influenza virus pneumonia in the validation and the derivation sets were comparable in age, prevalence of ARDS, SPAS II score, and renal and cardiovascular failures (Table 1).
An unsupervised classification study with PCA was carried out in patients with H1N1 influenza virus pneumonia (n = 30) to analyze the metabolic differences between ARDS (n = 12) and non-ARDS (n = 18). PCA (Fig. 1) provided a nearly perfect discrimination between the two groups of subjects. PLS regression coefficient analyses (Fig. 2) were carried out to identify the biomarker candidates. Metabolic differences elicited by our method between ARDS and non-ARDS subjects are highlighted in representative serum spectra (Fig. 3). The resonances were identified according to the Human Metabolome Database (25), and characteristic cross-peaks from 2D spectra to help in unequivocal assignation of these metabolites. Patients with H1N1 influenza virus pneumonia and ARDS showed, as compared with H1N1 patients without ARDS, lower serum concentrations of glucose (−34%, P < 0.05), alanine (−40%, P < 0.001), glutamine (−21%, P < 0.05), methylhistidine (−50%, P < 0.05), and fatty acids (−20%, P < 0.05), whereas phenylalanine (336%, P < 0.01) and methylguanidine (188%, P < 0.05) concentrations increased.
Based on the significantly different metabolic profile in patients with H1N1 influenza virus pneumonia between ARDS and non-ARDS, we developed a predictive model for the identification of ARDS. The classification functions derived from the probability of belonging to the ARDS or to the non-ARDS group were trained with the derivation set. Four latent variables (classification functions) were selected to build the PLS-DA model based on model robustness parameters, e.g., R2, mean squared error of prediction in cross-validation and variance explained (Fig. 4). These classification functions showed a good correlation with the SOFA score (R = 0.74, P < 0.0001) and the P02/FiO2 ratio (R = 0.41, P = 0.03), but lower correlation with other variables such as APACHE II score (R = 0.32, P = 0.09) or SAPS II score (R = 0.20, P = 0.29) (Fig. 5). PLS-DA was validated with the second set of H1N1 patients (validation set, n = 26). Three non-ARDS samples were identified as outliers in PCA analysis and were removed from the analysis. The PLS-DA classification model predicted the diagnosis of ARDS in the validation set (n = 23) of patients with H1N1 influenza virus pneumonia with a success rate of 92% (sensitivity 100% and specificity 91%) (Fig. 6).
In the present study we report for the first time the metabolomic profile of ARDS by NMR–spectroscopy using untargeted multivariate statistical analysis in patients with H1N1 influenza virus pneumonia. The metabolomic profile was validated in an external set of patients with H1N1 influenza virus pneumonia, in which the ARDS predictive model had a high accuracy of 92%. These findings are helpful for the understanding of the pathogenesis of the syndrome. Specifically, the metabolomic profile of ARDS in these patients suggests alterations in energy pathways and inflammatory response. Our findings of decreased levels of glucose, alanine, glutamine, and fatty acids in serum from ARDS patients with H1N1 influenza virus pneumonia resulted presumably from enhanced energetic metabolism in the lung. During lung injury, lung epithelial cells as well as other cells are likely to be challenged by energetic stress (9, 26), and the ability to induce energy generating pathways rapidly and effectively may be critical for cell survival (27). Glucose and fatty acids are the primary substrates for energy. Glucose and alanine are the precursors for pyruvate metabolism. The ability to increase lung glutamine may be also critical for lung cells survival. Glutamine is the precursor for glutamate and subsequently alfa-ketoglutarate. Glutamine is also oxidized for added energy in alveolar epithelial cells. Alternatively, amino acid and fatty acid catabolism also generate substrates for the tricarboxylic acid cycle and energy production. The decrease in the concentration of these metabolites in the serum form patients with ARDS may implicate an increased utilization of aerobic metabolism via the Krebs cycle (Fig. 7) in ARDS.
These results are in line with the findings recently reported by Banoei et al. (28), who pointed out that disruption of amino acid metabolism and gluconeogenesis pathways may be determinant of a poor prognosis in patients with H1N1 infection. Similar results were found in animal models of ALI by us and others. Chen et al. (15) reported energy disorders in H1N1-induced pneumonia in mouse serum. Our previous study of ventilator-induced ALI showed alterations in energy metabolism in lung tissue and serum samples. Similarly, Fabisiak et al. (27) found decreased levels of glucose, alanine, and glutamine in lung tissue in a model of chlorine-induced ALI in mice. Serkova et al. (9) reported decreased high-energy phosphates, energy balance, and energy charge, as well as an increased lactate-to-glucose ratio in a model of inflammation-induced ALI in mice. Hu et al. (26) also reported increased levels of glucose, lactate, and creatine in lung tissue in a model of silica-induced lung inflammation. Although these studies agree on the increased energy requirement during lung injury, the apparently discrepant results reported may be due to differences in the specific lung insult (intratracheal administration of cytokines (9), administration of an irritant (26, 29), or pneumonia (28) [present study]) and the species studied (rodents (9, 15, 26, 27, 29) and humans (28) [present study]).
Methylhistidine, a marker of inflammation (30), was found to decrease in ARDS patients, probably indicating increased utilization by inflammatory cells. Recently, Maltesen et al. (31) reported low levels of methylhistidine as an early marker of ALI after cardiac surgery.
We detected increased phenylalanine concentration in cases with ARDS. This result is in line with previous studies that reported significant increases in human serum phenylalanine during bacterial and viral infection (32). A similar increase was also observed during experimentally induced infections in dogs (33), rodents (34), and monkeys (32). This increase is observed regardless of the aetiology of the infectious disease and is characteristic of the inflammatory process per se(32).
Methylguanidine, an index of hydroxyl radical formation (35), was also increased in ARDS. This finding is in line with previous reports indicating that methylguanidine is a serum marker of lung injury in rat models of sepsis-induced ALI (36) and in patients with leptospirosis (37). Methylguanidine, as an inhibitor of nitric oxide synthase, has been also proposed as anti-inflammatory treatment of acute inflammation (38, 39).
The present design does not allow concluding as to the prognostic value of the above-described metabolic changes, as patients had already been diagnosed when serum samples were obtained, nor does it allow conclusions as to the specific metabolomic profile in other forms of ARDS. However, our results help build the proof of concept that NMR spectroscopy could be a useful tool for biomarker discovery in ARDS, and help identify in patients with predisposing conditions those at high risk of developing ARDS. The novel findings reported in this investigation have to be interpreted considering several limitations. First, the relationship between the serum metabolic profile and the lung metabolic reprograming is limited. Disruption in energy metabolism is a common metabolomics finding in critical illness; thus serum metabolome could reflect ARDS-induced changes in other organs such as the heart or the kidney. A more rigorous interpretation of the mechanisms involved in the pathogenesis of syndrome would require the study of lung tissue to obtain a correlation between serum biomarkers and lung metabolome. This multisample approach was used in our previous studies of lung dysfunction in sepsis and ventilator-induced lung injury in animal models (10, 11), where serum metabolic profile was correlated with changes detected in lung tissue and bronchoalveolar lavage, but unfortunately in vivo lung tissue NMR spectroscopy is not yet available in clinical practice. Second, we show differences in the metabolomic profile in patients with and without ARDS, but the design of the present pilot study did not allow us to conclude as to the predictive value of metabolomic changes in at-risk patients (i.e., before they develop the conditions of interest). Also, we cannot determine whether the reported changes are specific for ARDS induced by H1N1 influenza virus pneumonia, or whether other types of acute lung injury share the same metabolomic profile. Although these are limitations of the present study, we think that the results presented are original in this area and help build the proof of concept for future studies on ARDS biomarkers based on metabolomic analysis. In addition, the characteristics of these changes as disease biomarkers (e.g., correlation with disease severity, time course, relationship to response to treatment, hypoxia dependence) remain to be determined. Third, the limited sample size precludes further analyses to correlate the metabolomic changes with other outcomes such as mortality. Finally, as we studied only patients with H1N1 influenza virus pneumonia, it remains to be determined whether the herein-reported metabolomic profile (or a different one) is also useful for the identification of ARDS in the context of other infections.
In summary, we report for the first time the serum metabolomic profile of patients with H1N1 influenza virus pneumonia that develop ARDS. The identified changes, indicating derangement of the energetic metabolism, have translational implications, as they may be useful as biomarkers of the syndrome and may shed light on the mechanisms involved in the pathogenesis of the syndrome. Our findings support the role of NMR spectroscopy for biomarker discovery in ARDS. Future studies on the role of metabolomics for the prediction of ARDS or other outcomes in viral pneumonia as well as in other ARDS predisposing conditions are warranted.
The authors thank Palmira Villa of the NMR Centre of Complutense University of Madrid for NMR spectra acquisition.
1. Booth CM, Stewart TE. Severe acute respiratory syndrome and critical care medicine: the Toronto experience. Crit Care Med
33 (1 suppl):S53–60, 2005.
2. Perez-Padilla R, de la Rosa-Zamboni D, Ponce de Leon S, Hernandez M, Quinones-Falconi F, Bautista E, Ramirez-Venegas A, Rojas-Serrano J, Ormsby CE, Corrales A, et al. Pneumonia and respiratory failure from swine-origin influenza A (H1N1) in Mexico. N Engl J Med
361 7:680–689, 2009.
3. Nin N, Soto L, Hurtado J, Lorente JA, Buroni M, Arancibia F, Ugarte S, Bagnulo H, Cardinal P, Bugedo G, et al. Clinical characteristics and outcomes of patients with 2009 influenza A(H1N1) virus infection with respiratory failure requiring mechanical ventilation. J Crit Care
26 2:186–192, 2011.
4. Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, Fan E, Camporota L, Slutsky AS. ARDS Definition Task Force. Acute respiratory distress syndrome
: the Berlin Definition. JAMA
307 23:2526–2533, 2012.
5. Lorente JA, Cardinal-Fernandez P, Munoz D, Frutos-Vivar F, Thille AW, Jaramillo C, Ballen-Barragan A, Rodriguez JM, Penuelas O, Ortiz G, et al. Acute respiratory distress syndrome
in patients with and without diffuse alveolar damage: an autopsy study. Intensive Care Med
41 11:1921–1930, 2015.
6. Thille AW, Esteban A, Fernandez-Segoviano P, Rodriguez JM, Aramburu JA, Penuelas O, Cortes-Puch I, Cardinal-Fernandez P, Lorente JA, Frutos-Vivar F. Comparison of the Berlin definition for acute respiratory distress syndrome
with autopsy. Am J Respir Crit Care Med
187 7:761–767, 2013.
7. Stringer KA, McKay RT, Karnovsky A, Quemerais B, Lacy P. Metabolomics
and its application to acute lung diseases. Front Immunol
8. Bos LD, Weda H, Wang Y, Knobel HH, Nijsen TM, Vink TJ, Zwinderman AH, Sterk PJ, Schultz MJ. Exhaled breath metabolomics
as a noninvasive diagnostic tool for acute respiratory distress syndrome
. Eur Respir J
44 1:188–197, 2014.
9. Serkova NJ, Van Rheen Z, Tobias M, Pitzer JE, Wilkinson JE, Stringer KA. Utility of magnetic resonance imaging and nuclear magnetic resonance-based metabolomics
for quantification of inflammatory lung injury. Am J Physiol Lung Cell Mol Physiol
295 1:L152–161, 2008.
10. Izquierdo-Garcia JL, Nin N, Ruiz-Cabello J, Rojas Y, de Paula M, Lopez-Cuenca S, Morales L, Martinez-Caro L, Fernandez-Segoviano P, Esteban A, et al. A metabolomic approach for diagnosis
of experimental sepsis. Intensive Care Med
37 12:2023–2032, 2011.
11. Izquierdo-Garcia JL, Naz S, Nin N, Rojas Y, Erazo M, Martinez-Caro L, Garcia A, de Paula M, Fernandez-Segoviano P, Casals C, et al. A metabolomic approach to the pathogenesis of ventilator-induced lung injury. Anesthesiology
120 3:694–702, 2014.
12. Stringer KA, Serkova NJ, Karnovsky A, Guire K, Paine R 3rd, Standiford TJ. Metabolic consequences of sepsis-induced acute lung injury revealed by plasma (1)H-nuclear magnetic resonance quantitative metabolomics
and computational analysis. Am J Physiol Lung Cell Mol Physiol
300 1:L4–L11, 2011.
13. Mao H, Wang H, Wang B, Liu X, Gao H, Xu M, Zhao H, Deng X, Lin D. Systemic metabolic changes of traumatic critically ill patients revealed by an NMR
-based metabonomic approach. J Proteome Res
8 12:5423–5430, 2009.
14. Laiakis EC, Morris GA, Fornace AJ, Howie SR. Metabolomic analysis in severe childhood pneumonia in the Gambia, West Africa: findings from a pilot study. PLoS One
5 9:e12655, 2010.
15. Chen L, Fan J, Li Y, Shi X, Ju D, Yan Q, Yan X, Han L, Zhu H. Modified Jiu Wei Qiang Huo decoction improves dysfunctional metabolomics
in influenza A pneumonia-infected mice. Biomed Chromatogr
28 4:468–474, 2014.
16. Milner JJ, Wang J, Sheridan PA, Ebbels T, Beck MA, Saric J. 1H NMR
-based profiling reveals differential immune-metabolic networks during influenza virus infection in obese mice. PLoS One
9 5:e97238, 2014.
17. Nyamundanda G, Gormley IC, Fan Y, Gallagher WM, Brennan L. MetSizeR: selecting the optimal sample size for metabolomic studies using an analysis based approach. BMC Bioinformatics
18. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med
22 7:707–710, 1996.
19. Artigas A, Bernard GR, Carlet J, Dreyfuss D, Gattinoni L, Hudson L, Lamy M, Marini JJ, Matthay MA, Pinsky MR, et al. The American-European Consensus Conference on ARDS, part 2: Ventilatory, pharmacologic, supportive therapy, study design strategies, and issues related to recovery and remodeling. Acute respiratory distress syndrome
. Am J Respir Crit Care Med
157 (4 pt 1):1332–1347, 1998.
20. Hotelling H. Analysis of a complex of statistical variables into principal components. J Educ Psychol
24 6:417, 1933.
21. Hotelling H. The generalization of Student's ratio. Ann Math Statist
2 3:360–378, 1931.
22. Izquierdo-Garcia JL, Rodriguez I, Kyriazis A, Villa P, Barreiro P, Desco M, Ruiz-Cabello J. A novel R-package graphic user interface for the analysis of metabonomic profiles. BMC Bioinformatics
23. Vinaixa M, Samino S, Saez I, Duran J, Guinovart JJ, Yanes O. A guideline to univariate statistical analysis for LC/MS-based untargeted metabolomics
-derived data. Metabolites
2 4:775–795, 2012.
24. Ding B, Gentleman R. Classification using generalized partial least squares. J Comput Graph Statist
14 2:280–298, 2005.
25. Wishart DS, Jewison T, Guo AC, Wilson M, Knox C, Liu Y, Djoumbou Y, Mandal R, Aziat F, Dong E, et al. HMDB 3.0—The Human Metabolome Database in 2013. Nucleic Acids Res
41 (Database issue):D801–D807, 2013.
26. Hu JZ, Rommereim DN, Minard KR, Woodstock A, Harrer BJ, Wind RA, Phipps RP, Sime PJ. Metabolomics
in lung inflammation:a high-resolution (1)h NMR
study of mice exposedto silica dust. Toxicol Mech Methods
18 5:385–398, 2008.
27. Fabisiak JP, Medvedovic M, Alexander DC, McDunn JE, Concel VJ, Bein K, Jang AS, Berndt A, Vuga LJ, Brant KA, et al. Integrative metabolome and transcriptome profiling reveals discordant energetic stress between mouse strains with differential sensitivity to acrolein-induced acute lung injury. Mol Nutr Food Res
55 9:1423–1434, 2011.
28. Banoei MM, Vogel HJ, Weljie AM, Kumar A, Yende S, Angus DC, Winston BW. Canadian Critical Care Translational Biology Group (CCCTBG). Plasma metabolomics
for the diagnosis
and prognosis of H1N1 influenza pneumonia. Crit Care
21 1:97, 2017.
29. Leikauf GD, Pope-Varsalona H, Concel VJ, Liu P, Bein K, Berndt A, Martin TM, Ganguly K, Jang AS, Brant KA, et al. Integrative assessment of chlorine-induced acute lung injury in mice. Am J Respir Cell Mol Biol
47 2:234–244, 2012.
30. Peng J, Zeng J, Cai B, Yang H, Cohen MJ, Chen W, Sun MW, Lu CD, Jiang H. Establishment of quantitative severity evaluation model for spinal cord injury by metabolomic fingerprinting. PLoS One
9 4:e93736, 2014.
31. Maltesen RG, Rasmussen BS, Pedersen S, Hanifa MA, Kucheryavskiy S, Kristensen SR, Wimmer R. Metabotyping patients’ journeys reveals early predisposition to lung injury after cardiac surgery. Sci Rep
32. Wannemacher RW Jr, Klainer AS, Dinterman RE, Beisel WR. The significance and mechanism of an increased serum phenylalanine-tyrosine ratio during infection. Am J Clin Nutr
29 9:997–1006, 1976.
33. Newberne PM. Overnutrition on resistance of dogs to distemper virus. Fed Proc
25 6:1701–1710, 1966.
34. Wannemacher RW Jr, Powanda MC, Pekarek RS, Beisel WR. Tissue amino acid flux after exposure of rats to Diplococcus pneumoniae. Infect Immun
4 5:556–562, 1971.
35. Nakamura K, Ienaga K, Yokozawa T, Fujitsuka N, Oura H. Production of methylguanidine from creatinine via creatol by active oxygen species: analyses of the catabolism in vitro. Nephron
58 1:42–46, 1991.
36. Kao SJ, Su CF, Liu DD, Chen HI. Endotoxin-induced acute lung injury and organ dysfunction are attenuated by pentobarbital anaesthesia. Clin Exp Pharmacol Physiol
34 (5–6):480–487, 2007.
37. Chen HI, Kao SJ, Hsu YH. Pathophysiological mechanism of lung injury in patients with leptospirosis. Pathology
39 3:339–344, 2007.
38. Marzocco S, Di Paola R, Genovese T, Sorrentino R, Britti D, Scollo G, Pinto A, Cuzzocrea S, Autore G. Methylguanidine reduces the development of non septic shock induced by zymosan in mice. Life Sci
75 12:1417–1433, 2004.
39. Marzocco S, Di Paola R, Serraino I, Sorrentino R, Meli R, Mattaceraso G, Cuzzocrea S, Pinto A, Autore G. Effect of methylguanidine in carrageenan-induced acute inflammation in the rats. Eur J Pharmacol
484 (2–3):341–350, 2004.