The prebiotic effects of omega-3 fatty acid supplementation: A six-week randomised intervention trial

Study population

Study subjects were enrolled from the TwinsUK registry, a national register of adult twins recruited as volunteers without selecting for any particular disease or trait traits. A total of 69 subjects were enrolled into the study and randomized into either the omega 3 or fiber arm.

Study design and intervention

Participant eligibility included those aged >18 y who had a body mass index (BMI) between 20 and 39.9 kg/m2 and had a low habitual fiber consumption of less than 15 g/d. The following exclusion criteria were considered: ongoing or planned regular use of other omega-3 PUFA or cod liver oil supplements; seafood allergy; concomitant use of non-steroidal anti-inflammatory medications, including aspirin; current treatment for any chronic inflammatory condition or malignancy; previous colonic or small bowel resection; current smoker (minimum 6 months smoking cessation) and pregnancy.

If an individual was eligible, he/she was consented and booked in for their baseline clinical visit at the clinical research facility at St Thomas’ Hospital, London, UK. Participants were randomized to take either 20 g of inulin fiber or 500 mg of omega-3 supplements daily (165 mg of EPA, 110 mg DHA, in gelatin capsules) for a period of 6 weeks.

Neither participants nor researchers were blinded to the interventions and hence allocation order. The participants were booked in for a follow-up visit at the end of the 6-week intervention period. Randomization was performed using an online software (www.sealedenvelope.co.uk). All participants provided written informed consent. The trial was approved by the West Midlands Black Country Research Ethics Committee (18/WM/0066) and is registered under the clinicaltrials.gov database (NCT03442348).

Sample collection

Blood, stool, and anthropometric measures (height, weight, blood pressure, body composition) were collected at both the baseline and follow-up visits. Blood samples were collected from participants between 8:30am and 10am during each visit. Participants were instructed to come in fasted state at least since 9 pm the night before (i.e. minimum fasting time was 11.5 hours). Blood samples were collected using Serum Separator Tubes (SST) and were processed within 2–3 hours of collection for separating serum and aliquoted for storage at −20 C until the end of the intervention period.

Diet and lifestyle patterns were measured at baseline, mid-intervention (i.e. 3 weeks into the intervention), and at follow-up using a set of validated questionnaires including the EPIC-Norfolk: Food Frequency Questionnaire;(25) the Bristol stool form scale(26), and the SF-12 quality of life questionnaire.(27) Fecal samples were provided at study visits and immediately frozen at −80°C until DNA extraction, which occurred as soon as the study was completed.

Microbiota analysis

The stool DNA extraction is detailed in Goodrich et al.28 of 100 mg were taken from the sample and used for extraction. There was no homogenization prior to this step. Fecal samples were collected and the composition of the gut microbiome was determined by 16 S rRNA gene sequencing carried out as previously described.(29,30)

Briefly, the V4 region of the 16S rRNA gene was amplified using universal primers 355 F (CCAGACTCCTACGGGAGGCAGC) and 806 R (GGACTACHVGGGTWTCTAAT). Amplified DNA was sequenced on the MiSeq platform (Illumina). Read filtering and clustering were carried out using the MYcrobiota pipeline.31 Chimeric sequences were filtered using the VSEARCH algorithm within Mothur, and reads were clustered into OTUs using closed-reference clustering against the SILVA database v132 based on a 97% similarity. Diversity metrics (Shannon index, observed OTUs, and Unweighted UniFrac) were calculated by rarefying the OTU table down to 7000 sequences per sample 50 times and taking the average. These analyses were carried out in QIIME 2 (v2018.11).

Metabolite analysis

Serum short and branched chain fatty acids:

The method employed for the serum SCFA and BCFA was based on the in-situ pentafluorbenzylation of the free acid species, followed by GC-NCI-MS determination of the resulting derivatives. Measures were only obtained from serum and not feces given recent reports indicating that circulating levels but not fecal levels, in much larger sample sizes, correlate with clinical traits.(32) 

All reagents and primary standards for acetic, propionic, iso-butyric, butyric, isovaleric, valeric acids including the internal standard (d4 acetic acid) were purchased from Sigma Aldrich.100 µl of serum were dispensed into an Eppendorf tube, followed by the addition of 100 µl of acetonitrile containing internal standard (d4 acetic acid) at 6 µmol/l. After vortexing for 30s, the mixture was treated with 5 µl of neat pentafluorobenzyl bromide, followed by 3 µl of neat diisopropylethylamine and then incubated for 30 min at 60°C to effect derivatization. On cooling, 100 µl of heptane were added and the mixture vortexed briefly to facilitate analyte transfer. Upon a brief centrifugation, the supernatant (containing the pentafluorbezyl derivatives) was transferred to a glass insert vial for GC-MS analysis.

The analysis was performed on an Agilent Technologies 6890 gas chromatograph interfaced with a 5973 mass spectral detector operated in negative chemical ionization mode. The system inlet was operated in splitless mode with the injection point temperature set at 220°C. The short-chain fatty acid derivatives were separated on a DB-5 MS capillary column of dimensions 30 m × 250 µm ×0.25 µm using temperature programming and a constant carrier flow rate. Mass spectral data were acquired in selected ion monitoring mode with masses corresponding to the carboxylate anion chosen for both qualification and quantitation. Five level calibration curves were generated by preparing standards spanning the concentration range of interest and running under identical conditions. Briefly, a binary stock solution containing acetic acid and propionic was prepared at concentrations of 1.5 mg/ml and 0.15 mg/ml, respectively. Similarly, a stock solution containing iso-Butyric, Butyric, iso-Valeric, and Valeric was prepared with concentrations of 0.15 mg/ml, 0.27 mg/ml, 0.04 mg.ml and 0.15 mg/ml, respectively. It was necessary to have acetic/propionic calibrators independent of the other species to prevent traces of the other species present contributing to the lower levels of butyric onwards. These two independent stock mixtures were then diluted to form working calibrators which in turn were serially diluted to form the basis of the calibration. Hundred-microliter aliquots of the resulting calibrants were prepared as per serum.

Lipids and gut-derived metabolites:

Circulating levels of cholesterol and triglyceride fractions from fasting serum samples were measured using the high-throughput 1 H-NMR metabolomics platform (Nightingale Health Ltd., Helsinki, Finland; nightingalehealth.com/).(33) In addition, circulating levels of Docosahexaenoic acid (DHA) and Total omega 3 were also measured using the same platform.

Certain gut-derived metabolites such as TMAO and IPA were measured using tandem mass spectrometry with the Biocrates MxP Qaunt 500 kit (Biocrates Life Science AG, Innsbruck, Austria).(34)

Statistical analysis

OTUs with a relative abundance of <0.1% in every sample were removed, and zero inflated relative OTU abundances were inverse normal transformed before further analysis. We investigated the different effects of fiber and omega-3 interventions on changes in OTU abundances (genus level) by running general linear models, with change in OTU abundance as the outcome and fiber/omega-3 intervention as the exposure. We adjusted for age, gender, and BMI and multiple testing using false discovery rate (FDR<0.05). Linear regressions were employed to investigate the association between OTUs and serum metabolites adjusting for covariates such as age, sex, BMI, and multiple testing (FDR < 0.05). All statistical analyses were carried out in R v3.5.2.