Comparison of classification methods that combine clinical data and high-dimensional mass spectrometry data.

BACKGROUND: The identification of new diagnostic or prognostic biomarkers is one of the main aims of clinical cancer research. Technologies like mass spectrometry are commonly being used in proteomic research. Mass spectrometry signals show the proteomic profiles of the individuals under study at a given time. These profiles correspond to the recording of a large number of proteins, much larger than the number of individuals. These variables come in addition to or to complete classical clinical variables. The objective of this study is to evaluate and compare the predictive ability of new and existing models combining mass spectrometry data and classical clinical variables. This study was conducted in the context of binary prediction. RESULTS: To achieve this goal, simulated data as well as a real dataset dedicated to the selection of proteomic markers of steatosis were used to evaluate the methods. The proposed methods meet the challenge of high-dimensional data and the selection of predictive markers by using penalization methods (Ridge, Lasso) and dimension reduction techniques (PLS), as well as a combination of both strategies through sparse PLS in the context of a binary class prediction. The methods were compared in terms of mean classification rate and their ability to select the true predictive values. These comparisons were done on clinical-only models, mass-spectrometry-only models and combined models. CONCLUSIONS: It was shown that models which combine both types of data can be more efficient than models that use only clinical or mass spectrometry data when the sample size of the dataset is large enough.

Low preoperative cholesterol level is a risk factor of sepsis and poor clinical outcome in patients undergoing cardiac surgery with cardiopulmonary bypass.

OBJECTIVES: Systemic inflammatory response syndrome and sepsis frequently occur after cardiac surgery with cardiopulmonary bypass. The aim of the present study was to investigate whether preoperative cholesterol levels can predict sepsis onset and postoperative complications in patients undergoing cardiac surgery with cardiopulmonary bypass. DESIGN: Prospective observational study. SETTING: Surgical ICU of a French university hospital. PATIENTS: Two hundred and seventeen consecutive patients older than 18 years admitted for planned cardiac surgery with cardiopulmonary bypass. INTERVENTIONS: Measurements of plasma blood lipids and inflammation markers before anesthesia induction (baseline), at cardiopulmonary bypass start, at cardiopulmonary bypass end, and 3 and 24 hours after cardiac surgery. Outcomes were compared in systemic inflammatory response syndrome patients with sepsis (n = 15), systemic inflammatory response syndrome patients without sepsis (n = 95), and non-systemic inflammatory response syndrome patients (n = 107). MEASUREMENTS AND MAIN RESULTS: A gradual decrease in plasma cholesterol concentration occurred during surgery with cardiopulmonary bypass but was no longer present after correction for hemodilution. Corrected cholesterol levels were significantly lower at baseline in sepsis patients than in other subgroups, and it remained lower in the sepsis group during and after cardiopulmonary bypass. With regard to sepsis, the discriminatory power of baseline cholesterol was fairly good as indicated by receiver operating characteristic curve analysis (area under the curve, 0.78; 95% CI, 0.72-0.84). The frequency of sepsis progressively decreased with increasing baseline cholesterol level quintiles (18.6% and 0% in the bottom and top quintiles, respectively, p = 0.005). In multivariate analysis, baseline cholesterol levels and cardiopulmonary bypass duration were significant and independent determinants of the 3-hour postcardiopulmonary bypass increase in concentrations of procalcitonin and interleukin-8, but not of interleukin-6. CONCLUSIONS: Low cholesterol levels before elective cardiac surgery with cardiopulmonary bypass may be a simple biomarker for the early identification of patients with a high risk of sepsis.

Constitutive inhibition of plasma CETP by apolipoprotein C1 is blunted in dyslipidemic patients with coronary artery disease.

Plasma cholesteryl ester transfer protein (CETP) promotes the cholesterol enrichment of apoB-containing lipoproteins (VLDL and LDL) at the expense of HDL. Recent studies demonstrated that apoC1 is a potent CETP inhibitor in plasma of healthy, normolipidemic subjects. Our goal was to establish whether the modulation of CETP activity by apoC1 is influenced by dyslipidemia in patients with documented coronary artery disease (CAD). In the total CAD population studied (n = 240), apoC1 levels correlated negatively with CETP activity, independently of apoE-epsilon, CETP-Taq1B, and apoC1-Hpa1 genotypes. In multivariate analysis, the negative relationship was observed only in normolipidemic patients, not in those with hypercholesterolemia, hypertriglyceridemia, or combined hyperlipidemia. In the normolipidemic subjects, apoC1 levels were positively associated with higher HDL- to LDL-cholesterol ratio (r = 0.359, P < 0.001). It is concluded that apoC1 as a CETP inhibitor no longer operates on cholesterol redistribution in high-risk patients with dyslipidemia, probably due to increasing amounts of VLDL-bound apoC1, which is inactive as a CETP inhibitor. Patients with dyslipidemia could experience major benefits from treatment with pharmacological CETP inhibitors, which might compensate for blunted endogenous inhibition.

Constitutive androstane receptor activation stimulates faecal bile acid excretion and reverse cholesterol transport in mice.

BACKGROUND & AIMS: The constitutive androstane receptor (CAR) is a nuclear receptor expressed in the liver and involved in xenobiotic metabolism. The aim of this study was to assess whether pharmacological CAR activation could affect neutral sterol and bile acid elimination under conditions of cholesterol overload. METHODS: Wild type, Car-/-, ApoE-/-, and low-density lipoprotein receptor (Ldlr)-/- mice fed a western-type diet were treated with the CAR agonist TCPOBOP. RESULTS: CAR activation was associated with a decrease in faecal cholesterol output related to the repression of the Abcg5/g8 cholesterol transporters. In contrast, TCPOBOP treatment induced a marked increase (up to three fold, p<0.01) in the elimination of faecal bile acids. In the liver, it was related to the coordinated induction of genes involved in synthesis, sulfo-conjugation, and excretion of bile acids as well as the repression of the ileal apical sodium-dependent bile acid transporter. Importantly, cholesterol accumulation was reduced in the liver of TCPOBOP-treated animals. In all cases, TCPOBOP had no effect in Car-/- mice. To determine directly whether CAR activation could affect the elimination of endogenous cholesterol, kinetic studies were performed with high-density lipoproteins (HDL) labelled with (3)H-cholesteryl esters. We observed that TCPOBOP-treated mice excreted more HDL cholesterol-derived bile acids in their faeces. Finally, long-term CAR activation was associated with decreases in cholesterol content of the whole body and atherosclerosis susceptibility. CONCLUSIONS: CAR is involved in the control of cholesterol and bile acid homeostasis, increasing reverse cholesterol transport under hyperlipidemic conditions.

Multi-omics profiling reveals that eating difficulties developed consecutively to artificial nutrition in the neonatal period are associated to specific saliva composition.

Prolonged enteral or parenteral nutrition in neonatal periods sometimes results in eating difficulties persisting for years, with reduced food intake through the oral route and thereby reduced stimulation of the oral cavity. Aiming at describing the consequences on oral physiology, saliva of 21 children with eating difficulties (ED) was compared to that of 23 healthy controls, using various omics and targeted methods. Overall, despite heterogeneity within the groups (age, medication etc.), the three spectral methods (MALDI-TOF, SELDI-TOF, (1)H NMR) allowed discriminating ED and controls, confirming that oral stimulation by food intake plays a role in shaping the composition of saliva. Saliva of ED patients exhibited a lower antioxidant status and lower levels of the salivary protease inhibitors cystatins. Other discriminant features (IgA1, dimethylamine) may relate to modified oral and/or intestinal microbial ecology. Finally, salivary profiles of ED patients were partly comparable to those of subjects with exacerbated gustatory sensitivities, in particular with reduced abundance of cystatin SN and higher abundance of zinc-alpha-2-glycoprotein. Whether this translates taste hypersensitivity and contributes to the eating difficulties deserves further attention.

Human infant saliva peptidome is modified with age and diet transition.

In order to describe developmental changes in human salivary peptidome, whole saliva was obtained from 98 infants followed longitudinally at 3 and 6months of age. Data on teeth eruption and diet at the age of 6months were also recorded. Salivary peptide extracts were characterised by label-free MALDI-MS. Peptides differentially expressed between the two ages, and those significantly affected by teeth eruption or introduction of solid foods were identified by MALDI TOF-TOF and LC ESI MS-MS. Out of 81 peaks retained for statistical analysis, 26 were overexpressed at the age of 6months. Exposure to solid foods had a more pronounced effect on profiles (overexpression of nine peaks) than teeth eruption (overexpression of one peak). Differential peaks corresponded to fragments of acidic and basic PRPs, statherin and histatin. Comparison with existing knowledge on adult saliva peptidome revealed that proteolytic processing of salivary proteins is qualitatively quite comparable in infants and in adults. However, age and diet are modulators of salivary peptidome in human infants.

Increased apolipoprotein AI production rate and redistribution of high-density lipoprotein size induced by estrogen plus progestin as oral contraceptive.

CONTEXT: The impact of estrogen plus progestin as an oral contraceptive on high density lipoprotein (HDL) apolipoprotein (apo) AI metabolism in humans is poorly understood. OBJECTIVES: This study was designed to measure the in vivo effect of Moneva (30 microg ethinylestradiol, 75 microg gestodene) on HDL apoAI production rate and fractional catabolic rate. DESIGN: Using (13)C-leucine, we performed two kinetic studies in the fed state in 10 normolipidemic young women, before and 3 months after beginning Moneva. RESULTS: On Moneva, serum triglycerides increased by 12% (P = 0.03) in the fed state, whereas low-density lipoprotein and HDL cholesterol remained unchanged. HDL apoAI pool size and production rate were increased by 9.2% (67.3 +/- 7.1 vs. 61.6 +/- 6.7 mg x kg(-1); P = 0.05) and 26.5% (14.3 +/- 2.7 vs. 11.3 +/- 2.2 mg x kg(-1) x d(-1); P = 0.02), respectively. HDL apoAI fractional catabolic rate was not significantly modified. Three-month treatment by Moneva induced a shift of HDL size distribution from HDL2 toward HDL3 (HDL3 = 51.5 +/- 8.1 vs. 46.5 +/- 9.2% of total HDL; P = 0.02) and an increase in the proportion of apoAI among HDL components (38.8 +/- 4.3 vs. 34.4 +/- 2.8%; P = 0.01). CONCLUSION: Oral contraception by estrogen plus progestin induces changes in HDL apoAI metabolism characterized by an increase in production rate and pool size, with a higher proportion of HDL3 particles. Whether or not these changes are beneficial to prevent atherosclerosis has to be explored further.