Small extracellular vesicles in plasma reveal molecular effects of modified Mediterranean-ketogenic diet in participants with mild cognitive impairment

Figures

• 

  Figure 1. Characterization of TE. TEs were isolated from the plasma of CN and MCI participants in both pre- and post-MMKD and AHAD conditions. Six random TE samples/groups were analysed for their size distribution and concentration by NTA. (A) A representative line graph is presented for each group depicting concentration and size distribution for each of the six samples by a unique colour, and the average size of TE is mentioned on the top of the graph. (B and C) The bar diagrams present the size and concentration (particles/ml) of TE as mean ± SEM of n = 6 samples. Protein concentration in TE for CN (n = 11) and MCI (n = 9) participants, both pre- and post-MMKD and AHAD is presented as mean ± SEM. Protein concentration per TE is presented in lower right panels (n = 6 each). Paired t-test was applied for comparing size, concentration and protein concentration per TE in pre- and post-measures and the mixed effects model with random intercept was applied for comparing TE protein concentration in pre- and post-measures.

• 

  Figure 2. Characterization of TE for EV and neuronal markers. (A) TE were characterized by Exo-Check array (n = 3) for EV biomarkers. A representative blot is shown. (B) TE were analysed for surface L1CAM expression by flow cytometry. TE with only CellBrite 488 membrane dye (FITC) but without any other fluorescent antibody (unlabelled) were used as control (left panel). Twelve TE samples from CN and MCI groups were randomly selected and labelled with PE-tagged L1CAM antibody. TE were diluted with 1:100 folds in 0.22 µn filtered PBS and acquired on Cytoflex for 60 s. TE in the gated regions represents L1CAM + vesicles. Representative flow panels are shown for CN (middle panel) and MCI (right panel).

• 

  Figure 3. Characterization of sEVL1CAM. (A) sEVL1CAM were characterized by Exo-Check (Neuro) array, and a representative blot is shown. (B) Confocal microscopy images of sEVL1CAM attached on agarose resin and labelled with membrane labelling dye CellBrite and L1CAM fluorescent antibody (upper panel). Agarose resin tagged with IgG-biotin antibody, incubated with TE, and imaged with L1CAM-PE antibody and membrane labelling dye (lower panel) served as control. Scale bar is 60 μm. (C) sEVL1CAMattached to agarose resin were labelled with L1CAM-PE and CD63-AF647 fluorescent antibodies (upper panel). Agarose resin attached with IgG-biotin antibody were labelled using similar conditions as above (lower panel). Scale bar is 60 μm. (D) sEVL1CAMand sEVNCAM were isolated using agarose beads tagged with L1CAM-biotin (n = 6) or NCAM-biotin (n = 6) antibodies, respectively. Intact sEVL1CAM or sEVNCAM (n = 6 each) (without lysis) were immobilized on CD63 antibody-coated ELISA plate in two sets. Next, one set was labelled with L1CAM-biotin antibody, and the other was with CD63-biotin and probed with streptavidin solution. The captured fluorescence from L1CAM-biotin wells was normalized with fluorescence…

• 

  Figure 4. Validation of purity of sEVL1CAM by flow cytometry. sEVL1CAM from CN (n = 5) and MCI (n = 5) were characterized by flow cytometry. Right shift in the fluorescence in the gated region represents the L1CAM, ENO2 (PE) and synaptophysin (AF647) positive sEVL1CAM.

• 

  Figure 5. The effect of MMKD and AHAD on the size, concentration and protein concentration of sEVL1CAM. (A) sEVL1CAM isolated from the TE of CN and MCI participants, both pre- and post-MMKD and AHAD (n = 3 each), were analysed for their size distribution and concentration by NTA. A representative line graph is shown for each group depicting concentration and size distribution for each of the three samples by a unique colour, and the average size of sEVL1CAM is mentioned on the top of the graph. The bar diagrams (B and C) present the size and concentration of sEVL1CAM as mean ± SEM of n = 3 samples. Protein concentration in sEVL1CAM for CN (n = 11) and MCI (n = 9) participants, (B) both pre- and post-MMKD and (C) AHAD is presented as mean ± SEM. Protein concentration per sEVL1CAM is presented in the right panels (n = 3 each). Paired t-test was applied to compare sEVL1CAMsize, concentration and protein concentration per sEVL1CAM in pre- and post-measures. The mixed effects model with random intercept was applied for comparing sEVL1CAM protein concentration in pre- and post-measures.

• 

  Figure 6. The effect of MMKD and AHAD on Aβ1–42 and p181-tau. sEVL1CAM from CN (n = 11) and MCI (n = 9) participants both pre- and post-MMKD and AHAD were analysed for various ADRD biomarkers by ELISA for (A) Aβ1–42, (B) p181-tau and (C) Aβ1–42/p181-tau. Each biomarker concentration (pg/ml) was presented as per mg sEVL1CAM protein concentration (mean ± SEM). The mixed effects model with random intercept was applied for comparing pre- and post-measures.

• 

  Figure 7. The effect of MMKD and AHAD on tau and NfL. sEVL1CAM from CN (n = 11) and MCI (n = 9) participants both pre- and post-MMKD and AHAD were further analysed for (A) t-tau, (B) p181-tau/t-tau and (C) NfL. Each biomarker concentration (pg/ml) was presented as per mg sEVL1CAM protein concentration (mean ± SEM). The mixed effects model with random intercept was applied for comparing pre- and post-measures.

• 

  Figure 8. The effect of MMKD and AHAD on glutamate. sEVL1CAM from CN (n = 11) and MCI (n = 9) participants both pre- and post-MMKD and AHAD, were analysed for glutamate and glutamate receptors. (A) sEVL1CAM were lysed and analysed for glutamate levels. Total glutamate concentration was normalized with sEVL1CAM protein concentration and represented as µM glutamate per mg of sEVL1CAM (mean ± SEM). The mixed effects model with random intercept was applied for comparing pre- and post-measures. (B) Glutamate concentration in CSF samples presented as µM/ml of CSF (mean ± SEM). Repeated measure analysis of variance was used to calculate the statistical difference in CN and MCI groups. For CN, F = 0.10, d.f. = (2,30), P = 0.91 and for MCI, F = 0.86, d.f. = (2,22), P = 0.44). (C) Correlation of sEVL1CAM glutamate with CSF neurogranin.

• 

  Figure 9. The effect of MMKD and AHAD on glutamate receptors. sEVL1CAM from CN (n = 11) and MCI (n = 9) participants both pre- and post-MMKD and AHAD, were analysed for various glutamate receptors. (A–D) sEVL1CAM were lysed with RIPA buffer and analysed for the concentration of glutamate receptors (GRIN1, GRIN2A, GRIN2B and GRIA1) by ELISA and presented as pg/ml per mg of sEVL1CAM (mean ± SEM). The mixed effects model with random intercept was applied for comparing pre- and post-measures. (E) Pearson correlation coefficients between changes in sEVL1CAM (MMKD) from CN and MCI with corresponding changes in CSF measures are presented.

• 

  Figure 10. The effect of MMKD on inflammatory response and predictability of MCT2 expression on sEVL1CAMfor MMKD effect. sEVL1CAM from CN (n = 11) and MCI (n = 9) participants both pre- and post-MMKD and AHAD diet were analysed for the following: (A) The concentration of AGEs was analysed by ELISA assay, and the final concentration was calculated from the standards and normalized with sEVL1CAM protein concentration. The final concentration of AGEs was represented as ng/ml per mg sEVL1CAM (mean ± SEM). The mixed effects model with random intercept was applied for comparing pre- and post-measures. (B) NF-κB activation assay in THP-1 lucia monocyte cells was performed as described in the methods. sEVL1CAM from both CN and MCI participants in both pre- and post-MMKD and -AHAD conditions were pooled randomly to make five replicates/group. THP-1 lucia cells were incubated with sEVL1CAM and analysed for NF-κB activation. Graphs were plotted from five independent replicates (mean ± SEM). Paired t-test was applied to compare pre- and post-measures. (C) The expression of MCT2 was analysed on sEVL1CAM surface by flow cytometry. sEVL1CAM were isolated using biotin-labelled L1CAM antibody tagged on…

• 

  Figure 11. Characterization of NEE isolated using different surface biomarkers. NEE (sEVSYPand sEVNCAM) were isolated from TE from both CN (n = 5) and MCI (n = 5) participants in pre- and post-MMKD conditions using biotin-labelled synaptophysin or NCAM antibodies. (A) A representative line graph is presented for each group (Top panel: sEVSYP; bottom panel: sEVNCAM) depicting concentration and size distribution for each of the three samples by a unique colour and the average size of NEE is mentioned on the top of the graph. (B) sEVSYPand sEVNCAM were analysed for L1CAM, ENO2 and synaptophysin by flow cytometry.

• 

  Figure 12. Validation of results in plasma NEE. (A) sEVSYPand sEVNCAM were analysed for Aβ1–42 and NfL in both pre- and post-MMKD by ELISA. Final concentrations are presented as pg/ml per mg of sEVSYPor sEVNCAM. Paired t-test was applied to compare pre- and post-measures. (B) The matrix of correlation plots with smoothed regression lines for NfL and Aβ1–42. Pearson correlations with the corresponding significance levels are presented (*P < 0.05, **P < 0.01, ***P < 0.001).