Intensity-dependent lipidomic dynamic regulation following acute swimming train

Study design and swimming train protocol

This research utilized a randomized managed trial design. Participant inclusion and exclusion standards have been described beforehand¹⁹. All individuals had been wholesome, recreationally energetic college college students with no power illnesses, cardiovascular situations, or metabolic problems. None had been aggressive athletes or engaged in structured coaching. All individuals might full 50-meter freestyle swimming. Baseline traits are offered in Table 1. Forty-two individuals had been randomly assigned to the HIIT (n = 21) or MICT (n = 21) group. Both teams accomplished a one-week adaptation interval earlier than present process the formal swimming check in a fasted state (10–12 h in a single day quick) on the third morning.

After a 10-minute warm-up, the HIIT group carried out repeated cycles of 50-meter maximal sprints with 2-minute passive restoration for 30 min, whereas the MICT group accomplished 30 min of moderate-intensity swimming (70–80% HRmax). Heart price was repeatedly monitored utilizing the BHT-TEAM telemetry system (BoHaoTong, China), and RPE was assessed post-exercise utilizing the Borg 6–20 scale.

The research was permitted by BSU-IRB (protocol: 2022188 H) in accordance with the Declaration of Helsinki and was potential registered within the Chinese Clinical Trial Registry (ChiCTR2400089036, first registered 30/08/2024). All individuals supplied written knowledgeable consent. Swimming efficiency knowledge are proven in Table 2.

Blood assortment and serum preparation

Participants supplied fasting morning blood samples from the cubital vein on the preliminary timepoint. Following every swimming session, further venous samples had been obtained inside 2 min of pool exit, with subsequent collections at 15-minute and 30-minute restoration intervals. All collections utilized BD Vacutainer tubes (red-top, anticoagulant-free). Samples underwent coagulation for 45 min below ambient situations previous to centrifugation (1200 g, 10 min, 4℃) utilizing a refrigerated centrifuge. Resulting serum was aliquoted into pre-labeled 1.5mL cryovials and maintained at -80℃ pending evaluation. All pattern underwent just one freeze-thaw cycle.

Lipid and metabolite extraction

For lipid profiling, 50-µL serum aliquots underwent biphasic extraction utilizing a modified Bligh-Dyer protocol tailored from established methodology²⁰. The extraction process commenced with the addition of 750 µL chloroform: methanol answer (1:2, v/v) to every pattern, adopted by orbital agitation (1500 rpm) for 30 min at 4℃ to facilitate lipid solubilization. Phase partitioning was then induced by sequential addition of deionized water (350 µL) and chloroform (250 µL). Following centrifugal separation, the decrease lipid-rich natural layer was rigorously aspirated and transferred to recent microcentrifuge tubes. To maximize lipid restoration, the residual aqueous part underwent a second extraction cycle with an extra 450 µL of chloroform. The mixed natural extracts had been consolidated in a single tube and concentrated to dryness utilizing a SpeedVac vacuum concentrator (OH mode setting). Dried lipid residues had been maintained at -80℃ till instrumental evaluation.

For metabolite profiling, frozen serum samples underwent managed thawing at 4 °C with mild vortex mixing, adopted by clarification centrifugation to take away particulate matter. A 100-µL aliquot of clarified serum was then mixed with 400 µL of ice-chilled methanol containing 0.28 mM phenylhydrazine to facilitate chemical derivatization of α-ketoacid species²¹. Following vigorous vortex mixing, samples had been incubated at -20 °C for 30 min to permit full derivatization. The response combination was then subjected to high-speed centrifugation (12,000 rpm, 10 min, 4 °C) to pellet any precipitated proteins. The clarified supernatant was quantitatively transferred to wash 1.5 mL microcentrifuge tubes and concentrated to dryness by way of SpeedVac vacuum centrifugation. For instrumental evaluation, dried metabolite extracts had been reconstituted in 100 µL of 5% acetonitrile answer with thorough vortex mixing, adopted by a remaining clarification spin (12,000 rpm, 10 min, 4 °C) to take away any remaining particulates.

LC-MS-Based focused lipidomics and metabolomics profiling of serum

Statistical evaluation and knowledge visualization

Statistical energy evaluation was carried out utilizing G*Power software program (model 3.1.9.7; https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower), which decided {that a} minimal of 18 individuals per group can be required to detect a medium impact dimension (Cohen’s d = 0.6) with 80% statistical energy and α = 0.05. Raw lipidomic and metabolomic datasets had been processed utilizing R statistical surroundings (v4.4.1), with remaining visualization generated in Adobe Illustrator 2024. Data preprocessing included quantile-based normalization (preprocessCore bundle, v1.66.0) and z-score standardization (scales bundle, v1.3.0) to attenuate technical variability.

The analytical technique employed a multi-tiered strategy centered on log₂-fold change (log₂FC) values. For every lipid and metabolite species, log₂FC was calculated at every post-exercise timepoint (0, 15, and 30 min) relative to baseline measurements. Baseline demographic and physiological traits had been in contrast between train depth teams utilizing unbiased samples t-tests (Tables 1 and 2).

To visualize international metabolic response patterns, principal element evaluation (PCA) was utilized to the whole log₂FC dataset, enabling evaluation of intensity-related segregation in lipid perturbation profiles (Fig. 1). Exercise-induced modifications at particular person timepoints had been evaluated utilizing paired t-tests evaluating post-exercise values towards baseline. Lipid species exhibiting P < 0.05 with |log₂FC| > 0.2 had been labeled as considerably perturbed (Fig. 2).

To determine intensity-dependent lipid signatures, two-way repeated-measures evaluation of variance (ANOVA) was carried out on log₂FC datasets utilizing the automotive R bundle (v3.1-2), with train depth, time, and their interplay as elements (Fig. 3)²³. Partial least squares discriminant evaluation (PLS-DA; mixOmics v6.24.0) was carried out with log2FC values of 608 lipid species as predictor variables (X) and group project (HIIT vs. MICT) as response variable (Y), with Variable Importance in Projection (VIP) scores used to rank discriminatory energy. Mean space below the curve (AUC) was calculated to quantify cumulative perturbation magnitude (Fig. 3).

Fuzzy c-means clustering was applied by way of the Mfuzz bundle (v2.62.0) to determine shared and distinct temporal response patterns throughout the whole log₂FC dataset. The optimum clustering configuration (n = 3 clusters, fuzzification parameter m = 2.0) was chosen primarily based on cluster validity indices. Lipids assembly the importance standards (P < 0.05, |log₂FC| > 0.2) had been included within the remaining evaluation (Fig. 4).

Lipid structural traits, together with fatty acid chain size and diploma of unsaturation, had been analyzed to determine intensity-dependent structural preferences. For TAGs, structural distributions had been visualized utilizing scatter plots (carbon quantity on x-axis, double bond rely on y-axis), with every level representing a considerably regulated lipid species (P < 0.05, |log₂FC| > 0.2). For phospholipids and sphingolipids, structural traits had been displayed as radar charts, the place every axis represents a lipid subclass and the radial distance signifies the imply carbon quantity (Fig. 4F) or imply double bond rely (Fig. 4G). This strategy enabled identification of intensity-specific structural selectivity throughout lipid courses (Fig. 4E-G).

To discover lipid-metabolite regulatory relationships, correlation evaluation was carried out between the utmost log₂FC values of intensity-dependent metabolites and all lipid species utilizing Spearman’s rank correlation coefficient. P-values had been adjusted for a number of speculation testing utilizing the Benjamini-Hochberg false discovery price (FDR) technique, with q < 0.05 thought of statistically important (Fig. 5).

Baseline characterization and acute swimming train outcomes

No statistical variations had been noticed between HIIT and MICT teams in morphometric indices together with peak, weight, physique mass index (BMI), physique fats share, and waist and hip measurements (P > 0.1). Similarly, age (P = 0.732), resting coronary heart price (P = 0.194), and metabolic equivalents (P = 0.649) confirmed no important variations between teams. Total swimming length (P = 0.053) and vitality expenditure (P = 0.204) had been related between teams, however HIIT had considerably shorter web swimming time (P < 0.001), increased maximal coronary heart price (P < 0.001), and better imply swimming velocity (P < 0.001).

Principal element evaluation was carried out on the whole lipidomic dataset comprising 608 quantified lipid molecular species to evaluate international metabolic response patterns and group separation. Unsupervised PCA evaluation confirmed no clear separation between teams or time factors (Fig. 1B-E). Baseline traits had been constant between teams, and depth parameters had been well-differentiated (Tables 1 and 2). Therefore, we employed extra analytical methods in subsequent research and analyses to isolate and determine the true organic indicators pushed by train depth.

Panorama of serum lipidomic modifications after acute swimming train of various intensities

We examined lipidome modifications at three time factors (0-, 15-, and 30-minutes post-exercise) after MICT and HIIT, with considerably altered lipids outlined as these displaying adjusted P < 0.05 and |log₂FC| > 0.2 at any time level post-exercise relative to baseline.

The variety of considerably down-regulated lipid molecules differed between HIIT and MICT throughout all post-exercise time factors (Fig. 2A-C). At rapid, 15-minute, and 30-minute post-exercise intervals, HIIT down-regulated 1.49-fold, 2.10-fold, and a couple of.87-fold extra lipid molecules than MICT, respectively. Venn diagram evaluation revealed that at 30 min post-exercise, HIIT down-regulated 15-fold extra distinctive lipid molecules in comparison with MICT.

At the lipid subclass stage (Fig. 2D), HIIT down-regulated a higher variety of lipid subclasses than MICT, and this elevated through the post-exercise restoration interval. Both teams exhibited differential up-regulation throughout subclasses. HIIT-induced up-regulation was highest instantly post-exercise and decreased thereafter, whereas MICT exhibited steady up-regulation all through the restoration interval. TAG confirmed the biggest variety of regulated species amongst all lipid subclasses. Additionally, CE and SL had been solely regulated by HIIT in any respect time factors, whereas no MICT-specific subclasses had been recognized.

Intensity-dependent differential lipid screening

To determine train intensity-dependent lipid, we employed a multi-strategy strategy combining three statistical strategies. First, partial least squares discriminant evaluation (PLS-DA) recognized lipids with the best depth discriminatory capability at every time level. Second, imply space below the curve (AUC) evaluation quantified the cumulative modifications of particular person lipids throughout restoration and recognized lipids with the biggest cumulative variations between intensities. Finally, two-way ANOVA detected intensity-by-time interplay results, figuring out lipids with completely different dynamic patterns between intensities.

PLS-DA with VIP scoring recognized the 20 lipids with highest VIP scores (2.1 to 2.8) at every time level (Fig. 3A-C). LPA18:2, PC32:2, PI36:4 (18:2_18:2), TAG58:8 (22:6), TAG46:3 (18:2), TAG46:2 (16:0), TAG46:2 (18:2), TAG48:4 (18:2), and TAG48:3 (18:1) confirmed excessive VIP scores in any respect three time factors, with TAGs comprising 70% of those lipids.

Mean AUC evaluation revealed distinct lipid profiles between teams (Fig. 3D-F), aside from shared up-regulated GM3 d18:1/18:1. Further AUC evaluation recognized the highest 20 lipids with the best depth variations in cumulative response (-log₁₀ P 2.0 to eight.0), with PC32:2 exhibiting probably the most pronounced intensity-dependent modifications.

Five intensity-dependent lipids had been recognized by integrating PLS-DA with VIP scoring (Fig. 3A-C), imply AUC evaluation (Fig. 3D-F), and two-way ANOVA (Fig. 3G). These particular species, which included PC32:2, LPA18:2, TAG46:3 (18:2), TAG46:2 (18:2), and TAG48:4 (18:2), persistently emerged as important in every analytical strategy.

Temporal dynamics of serum lipids and structural characterization of chosen lipid courses

We recognized three distinct lipid dynamic patterns by C-mean fuzzy cluster evaluation to analyze post-swimming modifications in serum lipids (Fig. 4A-C). Cluster 1 confirmed persistent decline (MICT n = 50, HIIT n = 65, Fig. 4A), whereas clusters 2 and three (restoration clusters) confirmed decline-recovery (MICT n = 69, HIIT n = 137, Fig. 4B) and rise-recovery patterns (MICT n = 27, HIIT n = 70, Fig. 4C), respectively. HIIT regulated extra lipids than MICT in all clusters. In clusters 2 and three (Fig. 4B-C), HIIT regulated 2.27 and a couple of.59 instances extra lipids, respectively. No co-regulated lipids between HIIT and MICT had been noticed in cluster 3.

At the subclass stage (Fig. 4D), HIIT regulated a higher variety of lipids in practically all subclasses throughout clusters, whereas MICT regulated extra lipids solely in particular subclasses corresponding to FFA in cluster 3. TAG, probably the most responsive subclass to swimming, was predominantly regulated by HIIT in clusters 1 and a couple of however by MICT in cluster 3. SL, PS, and HexCer had been HIIT-exclusive, with no MICT-exclusive subclasses recognized. MICT confirmed cluster-specific regulation of PA, PC, PI and their hydrolysis merchandise (LPA, LPC, LPI), with the previous showing solely in cluster 1 and the latter solely in cluster 2.

To look at whether or not train depth selectively mobilizes lipids with particular structural options, we analyzed the distribution of carbon chain size and double bond rely amongst considerably regulated lipid species inside every temporal cluster (Fig. 4E-G). For TAG (Fig. 4E), persistently declining TAGs in cluster 1 had been predominantly shorter-chain and extra saturated in HIIT, however longer-chain and extra unsaturated in MICT. In cluster 2, the place regulatory results had been stronger, HIIT-regulated TAGs had longer chains and broader double bond rely ranges than MICT-regulated TAGs. In cluster 3, each intensities confirmed notably weaker results, with HIIT-regulated TAGs having longer chains and better unsaturation.

For phospholipids (Fig. 4F), HIIT and MICT differed primarily in subclass composition and unsaturation in cluster (1) These variations expanded to incorporate chain size in cluster (2) cluster 3 confirmed depth specificity, with solely HIIT regulating phospholipids.

For sphingolipids (Fig. 4G), HIIT and MICT differed in subclass, chain size, and unsaturation in cluster 1. In clusters 2 and three, solely HIIT regulated sphingolipids, with no MICT-regulated species detected.

Association of exercise-induced lipidome modifications with intensity-dependent vitality metabolism intermediates

Previous analysis has proven that the metabolites together with TCA cycle and glycolytic intermediates, in addition to sure ketones, exhibit intensity-dependent patterns^19,24. Targeted metabolomics recognized eight intensity-dependent metabolites (Fig. 5A). To additional elucidate the differential regulatory results of swimming train at various intensities on serum lipids, we carried out Pearson correlation evaluation between the maximal log₂FC of all lipid molecules and the maximal log₂FC of those eight metabolites and screened the highest 5 lipid molecules exhibiting the best correlation with every metabolite (Fig. 5B).

75% of lipids had been negatively correlated with intensity-dependent metabolites, with TAG and FFA because the predominant subclasses (Fig. 5B-C). FFA confirmed robust destructive correlations with L-lactic acid and pyruvate, sharing 4 similar molecules (Fig. 5B). The prime 5 lipids most strongly correlated with (R)-3-hydroxybutyric acid and L-malic acid had been all TAGs (Fig. 5B), with two TAGs shared between the 2 metabolites. Fumaric acid, as a precursor of L-malic acid, shared three negatively correlated TAGs with L-malic acid (Fig. 5B). Analysis of the fatty acid chain composition of negatively correlated lipids proven that 18:2, 16:0, and 18:1 had been the foremost fatty acid chain sorts, with 18:2 exhibiting the best frequency (n = 17) (Fig. 5D).