Plasma BDNF is a more reliable biomarker than erythrocyte omega-3 index for the omega-3 fatty acid enrichment of brain

Abstract

Enriching brain DHA is believed to be beneficial for the prevention and treatment of several neurological diseases, including Alzheimer's disease. An impediment in assessing the effectiveness of the treatments is the lack of a reliable biomarker for brain DHA. The commonly used erythrocyte omega-3 index is not suitable for brain because of the involvement of unique transporter at the blood brain barrier (BBB). We recently showed that dietary lysophosphatidylcholine (LPC)-DHA significantly increases brain DHA, which results in increase of brain BDNF. Since there is bidirectional transport of BDNF through the BBB, we tested the hypothesis that plasma BDNF may be used as biomarker for brain DHA enrichment. We altered the brain DHA in rats and mice over a wide range using different dietary carriers of DHA, and the correlations between the increase in brain omega-3 index with the increases in plasma BDNF and the erythrocyte index were determined. Whereas the increase in brain omega-3 index positively correlated with the increase in plasma BDNF, it negatively correlated with the erythrocyte index. These results show that the plasma BDNF is more reliable than the erythrocyte index as biomarker for assessing the effectiveness of omega-3 supplements in improving brain function.

Figure 1

Correlation of plasma BDNF levels with brain BDNF in rats. Two month old rats were gavaged daily with the indicated DHA-compounds (40 mg DHA/kg body weight) for one month, and the BDNF levels in plasma and brain regions were determined by ELISA. Two doses of LPC-DHA (5 mg and 10 mg) equivalent to 20 mg DHA and 40 mg DHA/kg body weight respectively were used. The insets show the absolute values (mean ± SD, n = 10 rats/group) of BDNF in the control (untreated) and DHA-treated groups. Bars of the same color without common superscripts are significantly different from each other (one-way ANOVA, with Tukey multiple comparison correction). The increase in BDNF by DHA treatment was calculated by subtracting the average of the control values from the individual samples of the treated groups. Pearson correlation was calculated between the increase in plasma BDNF vs the increase in cortex or hippocampus (Graphpad, Prism 8.0).

Figure 2

Correlation of omega-3 FA levels of brain with plasma BDNF and with erythrocyte omega-3 index in rats. The increase in cortex and hippocampus omega-3 FA (DHA + EPA) by the dietary treatment with various DHA carriers is plotted against the increase in plasma BDNF (A and B respectively) or against the erythrocyte omega-3 index (C and D respectively). The increase in brain omega-3 index was positively correlated with plasma BDNF, but negatively correlated with the increase in the erythrocyte index. The insets show the absolute values (mean ± SD, n = 10 rats/group), of BDNF and omega-3 FA (EPA + DHA), including those of the controls. Bars of same color without common superscripts are significantly different from each other by one-way ANOVA, with Tukey post-hoc correction.

Figure 3

Correlation of liver omega-3 FA enrichment with plasma BDNF and erythrocyte omega-3 index in rats. The increase in omega-3 FA levels in the liver correlated positively with the increase in plasma BDNF (top), but correlated negatively with the increase in erythrocyte omega-3 index (bottom).

Figure 4

Correlation of brain omega-3 changes in adipose tissue (top) and heart (bottom) with changes in plasma BDNF or erythrocyte omega-3 index in rats. The increase in omega-3 index of peri-gonadal adipose tissue (A) and of heart (C) correlated negatively with the increase in plasma BDNF, whereas the they were correlated positively with the increase in erythrocyte omega-3 index (B, D).

Figure 5

Correlation of brain omega-3 index with plasma BDNF and erythrocyte omega-3 index in normal mice. Normal male mice were gavaged with 40 mg DHA/kg body weight in the form of free DHA, sn-1 acyl LPC-DHA, or sn-2 acyl LPC-DHA for 30 days. The omega-3 FA content (DHA + EPA) of the brain regions and erythrocytes was measured by GC/MS, and the plasma BDNF levels were measured by ELISA. The insets show the absolute values (mean ± SD, n = 8 mice/group) for the % of omega-3 fatty acids (DHA + EPA) and the plasma BDNF values (pg/ml) in all groups, including controls (no treatment). In the inserts, bars of same color without common superscripts are significantly different from each other by one-way ANOVA, with Tukey post-hoc correction. The increases in omega-3 index and plasma BDNF levels above the control values are plotted. The increases in omega-3 FA of cortex as well as hippocampus correlated positively with the increases in plasma BDNF (A, B), whereas they correlated negatively with the omega-3 index of the erythrocytes (C, D).

Figure 6

Correlation of brain omega-3 index with plasma BDNF or erythrocyte omega-3 index in mice fed EPA. Normal male mice were gavaged with 40 mg EPA/kg body weight in the form of either free (unesterified) EPA or LPC-EPA for 15 days, and the plasma BDNF as well as omega-3 indexes were measured. The increases in omega-3 index of the brain (over the averages of control values) are plotted against the increases in plasma BDNF (top) or erythrocyte omega-3 index (bottom). The absolute values of omega indexes (EPA + DHA) and plasma BDNF (pg/ml) for all groups including the control are shown in the insets (mean ± SD, n = 6 mice/group). In the inserts, bars of same color without common superscripts are significantly different from each other by one-way ANOVA.

Figure 7

Correlation of brain omega-3 index with plasma BDNF and erythrocyte omega-3 index in mice fed krill oil or fish oil. Natural or lipase-treated fish oil and krill oil were incorporated into AIN-93G diet to provide 2.64 g of EPA + DHA per kg diet. These diets were fed to normal male mice for 30 days, and the tissue FA composition as well as plasma BDNF contents were measured. The top 2 panels (A, B) show the correlation of the increase in omega-3 indexes of cortex and hippocampus with the increase in plasma BDNF levels, whereas the bottom 2 panels (C, D) show the correlation of the increases in cortex and hippocampus omega-3 indexes with that of erythrocytes. The insets show the absolute values (mean ± SD, n = 5 mice/group) of omega-3 indexes (% of EPA + DHA) and the plasma BDNF levels (pg/ml) for all groups including the controls (which were fed unsupplemented AIN-93G diet). Bars of same color with different superscripts are significantly different from each other by one-way ANOVA.