By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Address correspondence and reprint requests to Stanley I. Rapaport, Brain Physiology and Metabolism Section, National Institute on Aging, National Institutes of Health, Building 9, Room 1S128, Bethesda, MD 20892–0947, USA. E-mail: email@example.com
Adult male unanesthetized rats, reared on a diet enriched in both α-linolenic acid (α-LNA) and docosahexaenoic acid (DHA), were infused intravenously for 5 min with [1-14C]α-LNA. Timed arterial samples were collected until the animals were killed at 5 min and the brain was removed after microwaving. Plasma and brain lipid concentrations and radioactivities were measured. Within plasma lipids, > 99% of radioactivity was in the form of unchanged [1-14C]α-LNA. Eighty-six per cent of brain radioactivity at 5 min was present as β-oxidation products, whereas the remainder was mainly in ‘stable’ phospholipid or triglyceride as α-LNA or DHA. Equations derived from kinetic modeling demonstrated that unesterified unlabeled α-LNA rapidly enters brain from plasma, but that its incorporation into brain phospholipid and triglyceride, as in the form of synthesized DHA, is ≤ 0.2% of the amount that enters the brain. Thus, in rats fed a diet containing large amounts of both α-LNA and DHA, the α-LNA that enters brain from plasma largely undergoes β-oxidation, and is not an appreciable source of DHA within brain phospholipids.
α-Linolenic acid (α-LNA; 18 : 3n-3) is a dietary essential n-3 polyunsaturated fatty acid (PUFA) that is a precursor for docosahexaenoic acid (DHA; 22 : 6n-3). Conversion of α-LNA to DHA occurs by a series of elongation, desaturation and β-oxidation steps, passing through eicosapentaenoic acid (EPA; 20 : 5n-3) and docosapentaenoic acid (DPA; 22 : 5n-3) intermediates (Sprecher 2000). DHA also can be directly obtained from dietary sources. Within brain, DHA is largely esterified at the stereospecifically numbered (sn)-2 position of phospholipids (Sastry 1985). It is thought to modulate membrane elasticity (Salem et al. 2001), activation of G-protein coupled receptors, ion channel flow, and neurotransmitter release (Innis 2003). Depletion of brain DHA by prolonged dietary deficiency of n-3 PUFAs can impair brain function in mammals (Innis 2000; Youdim et al. 2000).
Like other long-chain fatty acids (C16–C22) (Washizaki et al. 1994; Grange et al. 1995; Contreras et al. 2000), unesterified plasma α-LNA can rapidly diffuse from plasma into brain (Spector 2001). The extent, however, to which it is converted to DHA within brain is uncertain. Cultured neurons from fetal rat brain were reported to elongate α-LNA only to 20 : 3n-3 (homo-di-α-linolenic acid), suggesting an absence of desaturase activity, whereas cultured fetal rat astrocytes could synthesize the final DHA product as well as the EPA and DPA intermediates (Moore et al. 1991; Moore 2001; Williard et al. 2001). In the astrocytic cultures, about 50% of radiolabeled α-LNA was converted to DHA after 48 h of exposure. Likewise, it has also been noted that cultured rat C6 glioma cells, which are adult-derived but undifferentiated, can convert stable-isotope labeled α-LNA to DHA, with 3% of label in phospholipids as DHA after 16 h of incubation (Cook et al. 1991).
As most studies have focused on α-LNA metabolism in the immature brain, we thought it of importance to quantify the ability of the adult mammalian brain to synthesize DHA from plasma-derived α-LNA, when substantial DHA is included in the diet. We determined rates of uptake of α-LNA from plasma into the brain of unanesthetized adult male rats fed a diet enriched in DHA, as well as the extent to which α-LNA is incorporated into brain phospholipids per se or incorporated after being elongated to DHA inside the brain, as opposed to being lost through diversion to β-oxidation. To do this, we extended to α-LNA our in vivo fatty acid method for examining incorporation and turnover of circulating long chain fatty acids in brain phospholipids (Robinson et al. 1992; Rapoport et al. 2001). An abstract of part of this work has been presented (DeMar et al. 2004b).
Materials and methods
9,12,15[1-14C]α-linolenic acid ([1-14C]α-LNA) in 100% ethanol, specific activity = 54 mCi/mmol, was purchased from PerkinElmer Life Sciences, NEN Life Science Products (Boston, MA, USA). High performance liquid chromatography (HPLC) with scintillation counting (see below) confirmed that radioactive purity exceeded 98%. Di-heptadecanoate phosphatidylcholine (di-17 : 0 PC), unesterified heptadecanoic acid (17 : 0), heptadecanoyl-CoA (17 : 0-CoA), and thin layer chromatography (TLC) standards for cholesterol, triglycerides, and cholesterol esters were purchased from Sigma-Aldrich (St Louis, MO, USA). Standards for fatty acid methyl esters (FAMEs) and free fatty acids for GC and HPLC were from NuChek Prep (Elysian, MN, USA), whereas acyl-CoA standards for HPLC came from Sigma-Aldrich. 6-p-Toluidine-2-naphthalene sulfonic acid (TNS) was from Acros Organics (Fair Lawn, NJ, USA). Solvents were HPLC grade and were purchased from Fisher Scientific (Fair Lawn, NJ, USA) or EMD Chemicals (Gibbstown, NJ, USA). Other chemicals, unless noted, were from Fisher Scientific.
Animals and surgery
The Animal Care and Use Committee of the National Institute on Child Health and Human Development approved the protocol, which followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication no. 80–23). Adult (2 month-old) male Fischer-344 (CDF) rats were purchased from Charles River Laboratories (Portage, MI, USA) and were housed in an animal facility for 4 weeks before study, with regulated temperature, humidity, and a 12/12 h light/dark cycle. The rats were provided ad libitum water and rodent chow, formulation NIH-31 18–4 (Zeigler Bros. Inc., Gardners, PA, USA). The chow contained soybean oil and fishmeal, and had 4% by weight, crude fat. By GC (see below), saturated and monounsaturated fatty acids contributed to 20.1 and 22.5%, respectively, of its total fatty acid content, whereas the n-3 PUFAs α-LNA, EPA, and DHA contributed 5.1, 2.0 and 2.3%, respectively, and the n-6 PUFAs LA and ARA were 47.9 and 0.02%, respectively. The 9 : 1 LA to α-LNA concentration ratio is within the appropriate balance for optimal synthesis of ARA and DHA, respectively; additionally, the DHA content is ∼3-fold greater than required by mammals on a daily basis (Bourre et al. 1989; Van Aerde and Clandinin 1993).
Rats (n = 6) weighing 300 ± 39 (SD) g were anesthetized with 1–3% halothane. Polyethylene catheters (PE 50, Intramedic™, Clay Adams™, Becton Dickinson, Sparks, MD, USA) filled with heparinized saline (100 IU/mL) were implanted into the right femoral artery and vein as reported (Grange et al. 1995; Chang et al. 2001). Surgery lasted about 20 min and was performed between 10.00 h and noon. After surgery, an animal was allowed to recover from anesthesia for 3–4 h, with its hindquarters loosely wrapped in a fast-setting plaster cast that was taped to a wooden block. Body temperature was maintained at 36–38°C using a rectal probe and a feedback-heating element (YSI Indicating Temperature Controller; Yellow Springs Instrument Co., Yellow Springs, OH, USA). Prior to surgery, the rats were allowed free access to food, thus were not in a fasted state entering the experiment.
Each rat was infused with 500 μCi/kg of [1-14C]α-LNA. An aliquot of [1-14C]α-LNA, based on body weight, was evaporated to dryness by nitrogen gas. To the residue was added 1.3 mL of 5 mm HEPES buffer (pH 7.4), containing 50 mg/mL fatty acid-free bovine serum albumin. With an average body weight of 300 g, this gave an infusate concentration of 115 μCi/mL and a dose per rat of 150 μCi. The mixture was sonicated at 40°C for 20 min and mixed by vortexing. The rat was infused via the femoral vein catheter for 5 min, using a computer-controlled variable speed pump (No. 22; Harvard Apparatus, South Natick, MA, USA). The infusion rate, 0.223 (1–e−1.92t) ml/min (t is infusion time in min) (Washizaki et al. 1994), produces a constant plasma radioactivity within 1 min (Robinson et al. 1992; Chang et al. 1999). Arterial blood was collected at 0, 0.25, 0.5, 0.75, 1.5, 3, 4, and 5 min after beginning infusion. At 5 min, the rat was rapidly killed with an overdose of sodium pentobarbital (100 mg/mL, i.v.) and its head immediately subjected to focused-microwave irradiation (5.5 kW, 3.4 s) (Model S6F, Cober Electronics; Stamford, CT, USA) to halt brain metabolism. The brain was removed, cut sagittally in half, and stored at −80°C. Plasma was frozen at −80°C after removal from whole blood samples that had been microcentrifuged at 15 000 g for 5 min.
Isolation of brain and plasma lipids
Total lipids from one brain hemisphere and from plasma were extracted with chloroform/methanol/0.5 m KCl (2 : 1 : 0.75, v/v/v) (Folch et al. 1957). The aqueous phase was washed once with an equal volume of chloroform to remove residual lipids. The lipid extracts were separated into cholesterol, cholesterol esters, phospholipids, triglycerides, and unesterified fatty acids by neutral lipid TLC on Silica gel 60 plates (EM Separation Technologies; Gibbstown, NJ, USA), using heptane/diethyl ether/glacial acetic acid (60 : 40 : 3, v/v/v) (Skipski et al. 1968). The TLC plates were sprayed with 0.03% TNS in 50 mm Tris buffer (pH 7.4) (w/v), then the lipid bands were visualized under UV light. Appropriate standards were run to identify the lipids.
Quantification of lipid radioactivity and fatty acid concentrations
Aliquots of total lipid extracts and corresponding aqueous fractions from brain and plasma were subjected to liquid scintillation counting. Radioactivities in the neutral lipid bands scraped from the TLC plates also were determined. The endogenous fatty acid content of TLC scrapes containing cholesterol ester, phospholipid, triglycerides, and unesterified fatty acids were determined by converting their associated fatty acids to FAMEs, which were separated and identified on a GC with a flame ionization detector (DeMar et al. 2004a). Di-17 : 0 PC was added as an internal standard to lipids containing esterified fatty acids, and unesterified 17 : 0 was added as a standard to the unesterified fatty acids. Fatty acid concentrations (nmol/g brain or nmol/mL plasma) were calculated by proportional comparison of GC peak areas to the area of the 17 : 0 internal standard. Unesterified fatty acid concentrations in plasma were determined more precisely by pooling equal amounts of plasma from all blood samples in a given rat, and subjecting the pool to total lipid extraction, TLC, methylation, and GC as described above; unesterified 17 : 0 was added as an internal standard during the extraction. Radioactivities of different brain lipids were corrected for the blood contribution by subtracting the product of the cerebral blood volume (0.020 mL/g brain) multiplied by the radioactivity of the lipid of interest (Grange et al. 1995; Chang et al. 1999, 2001).
Fatty acid phenacyl ester preparation and HPLC analysis
Equal quantities of total lipid extracts from brain and plasma were pooled from each animal (n = 6). Brain phospholipids were isolated by TLC and converted to FAMEs, as discussed above. The FAMEs from brain phospholipids or triglycerides and the plasma total lipids were saponified with 2% KOH/EtOH (w/v) at 100°C for 45 min, then acidified with HCl and extracted with hexane. The fatty acid extracts were converted to fatty acid phenacyl esters (FAPEs) and separated on HPLC (DeMar et al. 2004a). Radioactivity profiles were obtained by collecting fractions every 0.5–1 min during HPLC elution and subjecting them to liquid scintillation counting. Peaks were identified from retention times of FAPEs prepared from unlabeled fatty acid standards, and from [1-14C]α-LNA and [1-14C]DHA (PerkinElmer Life Sciences, NEN Life Science Products; Boston, MA, USA).
Acyl-CoA isolation and HPLC analysis
Long chain acyl-CoAs were extracted from the brain hemisphere, not subjected to total lipid extraction, using a modified affinity chromatography method (Deutsch et al. 1994). Tissue was homogenized, using a probe sonicator, in isopropanol/25 mm KH2PO4/acetonitrile (1 : 1 : 2 v/v/v), to which heptadecanoyl-CoA (17 : 0-CoA; 10 nmol) had been added as an internal standard. A small volume (∼3% of total) of saturated (NH4)2SO4 solution was added to the homogenate to precipitate protein, which was removed by centrifugation. The supernatant was diluted with a 1.25-fold volume of 25 mm KH2PO4 and passed three times through an oligonucleotide purification cartridge (ABI Masterpiece™, OPC®, Applied Biosystems; Foster City, CA, USA). After washing the cartridge with 25 mm KH2PO4, bound acyl-CoA was eluted with isopropanol/1 mm glacial acetic acid (75 : 25 v/v).
Acyl-CoA species were separated on HPLC (System Gold® model 125, Beckman; Fullerton, CA, USA), using a 25 cm × 4.6 mm i.d., 5-μm, C18 reverse-phase column (Symmetry®, Waters Corp; Milford, MA, USA). The species were eluted (1 mL/min) by a linear gradient of 75 mm KH2PO4/acetonitrile, initiated at 56 : 44 (v/v), decreased to 51 : 49 (v/v) in 25 min, decreased to 32 : 68 (v/v) in 10 min, held at 32 : 68 (v/v) for 4 min, returned to 56 : 44 (v/v) in 6 min, and held for another 6 min (52 min total run time). Elution was monitored at 260 nm on a UV/VIS detector (System Gold® model 168, Beckman), and peaks were identified from retention times of acyl-CoA standards. The standards for acyl-CoA derivatives of α-LNA, EPA, DPA, and DHA are not commercially available and were prepared by reacting purchased fatty acids with free CoA (Hajra and Bishop 1986). Brain acyl-CoA concentrations (nmol/g brain) were calculated by proportional comparison with the peak area of the added 17 : 0-CoA internal standard. After the peaks were collected, their radioactivity was determined by liquid scintillation counting and, after correcting for per cent recovery of the 17 : 0-CoA, expressed as nCi/g brain.
Under our HPLC conditions, 14 : 0-CoA, α-LNA-CoA, and EPA-CoA co-eluted as a single unresolved peak. This peak was collected and its associated fatty acids were released by saponification, converted to FAPEs, and separated on HPLC, as previously described above. Radioactivity was measured by scintillation counting of the collected HPLC fractions (0.5 mL/min). The concentrations of the acyl-CoA species were estimated by monitoring the eluting FAPE absorbance at 242 nm and making direct proportional comparison (extinction coefficient adjusted) of the resulting peaks (Chen and Anderson 1992). The FAPE derivatives of 14 : 0, α-LNA, and EPA were fully resolved on HPLC.
We can write for the incorporation of intravenously injected radiolabeled α-LNA into brain phospholipid or triglyceride (‘stable’ lipid) i,
where = radioactivity in i (nCi/g brain) because of α-LNA, = plasma radioactivity of α-LNA (nCi/mL plasma), t = time, and = the incorporation coefficient of α-LNA in units of ml plasma/s/g brain (or 1/t, as 1 g brain ≈ 1 mL plasma). Integrating to time of death gives,
where is brain radioactivity at T.
To determine the extent to which α-LNA is esterified into ‘stable’ lipid i in the form of its DHA product, we can write an equation similar to Equation 2,
where = the conversion-incorporation coefficient and = brain radioactivity as a result of DHA at time T.
Because the incorporation coefficients and apply to both radiolabeled and unlabeled α-LNA (there is no significant kinetic isotope effect with a single 14C labeling), we can use them to calculate rates of incorporation of unlabeled unesterified plasma α-LNA into brain phospholipid or triglyceride i, in the forms of unchanged α-LNA or of synthesized DHA, respectively,
and these rates are in units of nmol/s/g brain.
As shown in our general model (Robinson et al. 1992; Rapoport et al. 2001), we can convert a rate of incorporation of the fatty acid from plasma into ‘stable’ brain lipid compartment i, into a rate of incorporation from the brain acyl-CoA (precursor) pool, by correcting it against the steady-state ratio of the specific activity of that pool over the specific activity of the infused fatty acid in plasma. The ratio of the two steady-state specific activities is defined as a dilution factor λα–LNA–CoA,
where the numerator and denominator are the specific activities of brain α-LNA-CoA and plasma α-LNA, respectively. During programmed infusion of a labeled fatty acid, steady-state specific activities for both brain and plasma fatty acid pools are established within 1 min (Washizaki et al. 1994; Grange et al. 1995).
It has been shown elsewhere that λα–LNA–CoA represents the ratio of the steady-state flux of unesterified plasma fatty acid into the brain acyl-CoA pool, divided by the sum of fluxes into that pool from unesterified and esterified plasma fatty acid, recycling, and de novo synthesis. For long chain fatty acids, within the time frame of a 5-min infusion period, fluxes due to de novo synthesis and esterified plasma fatty acid can be ignored (Robinson et al. 1992; Purdon et al. 1997; Rapoport et al. 2001).
Rates of incorporation of unlabeled α-LNA and its DHA product from the brain acyl-CoA pool (α-LNA-CoA or DHA-CoA) into ‘stable’ brain lipid compartment i are given as,
The equations described above (1–7) can be summed over all compartments i to give equivalent equations for net incorporation and turnover in brain total lipids.
Summarized and defined in Table 1 are the critical parameters measured during our study regarding the metabolism of [1-14C]-α-LNA in the brain. These include for plasma and/or brain: lipid fatty acid concentrations, lipid radioactivities, and the kinetic parameters k*, Jin,λacyl-CoA, JFA, FFA, and t1/2.
Table 1. Various parameters measured for study of brain metabolism of[l-14C]-α-LNA.
Plasma aqueous and lipid fraction radioactivities (nCi/ml)3
Brain aqueous and lipid fraction radioactivites (nCi/g)4
Net rates for uptake of plasma unesterified α-LNA and synthesized DHA into brain lipids5:
k*; incorporation coefficient (ml/sec/g)
Jin; incorporation rate (nmol/sec/g)
Direct incorporation and recycling rates for α-LNA in brain phospholipids6:
λα-LNA-CoA; fractional dilution of plasma unesterified α-LNA into brain acyl-CoA (no units)
JFA; incorporation rate of α-LNA-CoA into brain phospholipids (nmol/sec/g)
FFA; turnover rate for PLA2 release of brain phospholipid α-LNA and recycling into acyl-CoA (%/h)
t1/2; half-life for release / recycling of brain phospholipid α-LNA (h)
Concentrations and radioactivities are expressed as means ± SD (n = 6). However, HPLC profiles in Figs 2(b), 3(c), and 4(a) were from pooled samples from six animals, respectively, and do not have standard deviations. Statistical comparisons between means were performed using a two-tailed t-test. Differences were considered statistically significant at p < 0.05.
As shown in Table 2, unlabeled endogenous concentrations of esterified and unesterified fatty acids in plasma phospholipids, triglycerides, and cholesterol esters were determined in each rat from six pooled arterial samples collected during the [1-14C]α-LNA infusion. The concentration of plasma unesterified α-LNA equaled 41 ± 13 nmol/mL, which is about twofold greater than that of esterified α-LNA. The unesterified concentrations of EPA, DPA, and DHA equaled 12.6 ± 8.7, 16.4 ± 10.5, and 26.0 ± 12.2 nmol/mL plasma, respectively. These concentrations were much less than their corresponding net esterified concentrations. The concentrations in Table 2 are consistent with published values (Spector 2001).
Table 2. Concentrations of fatty acids in different lipid compartments of arterial plasma
Esterified fatty acids
Unesterified fatty acids
Data are means ± S.D. (n = 6).
4.6 ± 0.6
20.5 ± 7.5
23.6 ± 12.7
41.1 ± 5.1
430 ± 79
471 ± 196
123 ± 22
243 ± 117
521 ± 73
139 ± 90
70.3 ± 14.6
86.5 ± 13.7
62.4 ± 13.4
510 ± 202
255 ± 46
214 ± 136
310 ± 72
457 ± 245
166 ± 30
289 ± 185
1.1 ± 0.4
23.3 ± 14.5
41.3 ± 13.2
354 ± 65
68.0 ± 27.1
338 ± 62
24.5 ± 13.0
17.9 ± 4.6
62.2 ± 19.1
27.6 ± 7.7
12.6 ± 8.7
24.5 ± 5.4
42.2 ± 19.1
0.83 ± 0.21
16.4 ± 10.5
103 ± 16
90.1 ± 41.1
13.1 ± 2.8
26.0 ± 12.2
1828 ± 314
1884 ± 754
1018 ± 168
994 ± 493
In Fig. 1, radioactivities in the total lipid and aqueous fractions of arterial plasma are shown plotted against time over the [1-14C]α-LNA infusion. A constant total lipid radioactivity was established by 0.5 min and was maintained to the end of infusion, when it represented 88% of plasma radioactivity. Over the 5-min period, the integral of total lipid radioactivity equaled 571 871 ± 129 023 nCi/mL plasma/s.
TLC (Fig. 2a) showed that 97% of plasma lipid radioactivity at 5 min was in the unesterified fatty acid fraction. HPLC (Fig. 2b) showed that ∼ 99% of this fraction represented [1-14C]α-LNA, ≤ 0.2% was labeled DHA, and labeled EPA and DPA were not detected. The inconsequential level of labeled DHA means that any labeled DHA found inside the brain at 5 min was not derived from labeled DHA formed by the liver and secreted into the blood during the infusion. Furthermore, the integrated plasma lipid radioactivity (see above) is completely representative of cumulative [1-14C]α-LNA radioactivity, , the input function used to calculate incorporation coefficients by Equations 2 and 3.
Brain lipid composition
Given in Table 3 are the unlabeled endogenous fatty acid concentrations in different brain lipid compartments at 5 min. Unesterified α-LNA and EPA were not detected. α-LNA-CoA equaled 0.077 ± 0.012 nmol/g brain, which is 94% less than the concentration of DHA-CoA, at 1.3 ± 0.2 nmol/g brain. Low concentrations of EPA-CoA and DPA-CoA were detected, at 0.057 ± 0.011 and 0.36 ± 0.10 nmol/g brain, respectively. In phospholipid, the α-LNA concentration equaled 9.1 ± 1.9 nmol/g brain, compared with 36.2 ± 6.3, 328 ± 43, and 17227 ± 2391 nmol/g brain for EPA, DPA, and DHA, respectively. Overall, the percentages of α-LNA in brain unesterified fatty acid, acyl-CoA, and phospholipid were < 0.001, 0.2, and 0.006%, respectively; whereas, DHA was found at 1, 3, and 12%, respectively, in these same lipids. All of our above values are comparable with previously reported values by our group in rats fed NIH-31 chow (Chang et al. 1999, 2001).
Table 3. Concentrations of fatty acids in different brain lipid compartments.
Unesterified fatty acids nmol/g brain
Esterified fatty acids
Data are means ± S.D. (n = 6). ND, not detected, < 0.001 nmol/g brain.
1.8 ± 0.5
0.59 ± 0.09
237 ± 29
49.3 ± 8.0
11.3 ± 1.7
47061 ± 3733
42.5 ± 6.8
7.8 ± 0.9
32897 ± 3162
10.8 ± 6.3
17.5 ± 1.0
35449 ± 3834
1.2 ± 0.3
1235 ± 139
0.077 ± 0.012
9.5 ± 2.0
1.7 ± 0.7
1.1 ± 0.2
13731 ± 1742
0.057 ± 0.011
38.0 ± 6.7
0.23 ± 0.15
0.36 ± 0.10
338 ± 51
1.3 ± 0.7
1.3 ± 0.2
17632 ± 2779
107 ± 18
41.4 ± 2.2
148628 ± 13620
Figure 3 (a and b) and the first column of Table 4 both give the radioactivity in the different brain compartments (aqueous and lipid fractions) following 5 min of infusion of 500 μCi/kg [1-14C]α-LNA. Radioactivity accumulated in the brain aqueous fraction was 105 ± 29 nCi/g brain, whereas the total lipid radioactivity consisted of 4.0 ± 1.0, 6.5 ± 5.9, and 33.7 ± 12.0 nCi/g brain in the triglyceride, cholesterol, and phospholipid fractions, respectively. HPLC (Fig. 3c) was used to determine the exact nature of fatty acid radioactivity in phospholipids from the six pooled brains, or more specifically, what percentage of the phospholipid radioactivity was found as α-LNA and DHA. [1-14C]α-LNA represented 45% of this radioactivity (15 ± 5 nCi/g brain), whereas [14C]DHA represented ≤ 1% (0.32 ± 0.14 nCi/g brain) (Table 4). There were no radioactive peaks that matched with certainty the anticipated HPLC retention times for EPA or DPA (∼18 and ∼26 min, respectively); thus, we did not quantify these two fatty acids. However, it is possible some minor unknown peaks observed in the HPLC profile could have represented small amounts of [14C]EPA or [14C]DPA.
Table 4. Radioactivity, incorporation coefficients k*i and incorporation rates Jin,i in different brain lipid compartments.
Radioactivity nCi/g brain
ki* ml/sec/g brain × 10−4
Jin,i nmol/sec/g brain × 10−4
Data are means ± S.D. (n = 6) following 5 min of i.v. infusion of 500 μCi/kg [1-l4C]α-LNA. Values for ki* and Jin,i are given only for ‘stable’ lipid compartments (Robinson, 1992). α-LNA, α-linolenic acid (18:3n-3); EPA, eicosapentaenoic acid (20:5n-3); DPA, docosapentaenoic acid (22:5n-3); DHA, docosahexaenoic acid (22:6n-3).
α-LNA - CoA
3.3 ± 0.9
α-LNA - phospholipid
15.2 ± 5.3
0.26 ± 0.03
10.4 ± 2.5
2.8 ± 0.7
0.049 ± 0.008
2.0 ± 0.7
EPA and DPA-CoA
0.33 ± 0.05
DHA - CoA
0.45 ± 0.20
DHA - phospholipid
0.32 ± 0.14
0.0055 ± 0.0010
0.22 ± 0.06
0.023 ± 0.006
0.00040 ± 0.00006
0.016 ± 0.006
105 ± 29
6.5 ± 5.9
Fatty acid - CoA (recycled)
2.9 ± 2.2
Fatty acid - phospholipid (recycled)
18.2 ± 6.5
Fatty acid- triglycerides (recycled)
1.2 ± 0.3
157 ± 44
Peaks for 14 : 0, 16 : 0, and 18 : 0 fatty acids (recycled [1-14C]acetate) equaled 9% of the total radioactivity. The remaining radioactivity (∼45%) was in unidentifiable peaks throughout the chromatogram, but most likely represent some form of recycled β-oxidation products.
We also analyzed by HPLC the exact nature of the fatty acid radioactivity in the triglycerides from the six pooled brains (HPLC trace not shown). In brain triglycerides, 69% of the radioactivity was as [1-14C]α-LNA (2.8 ± 0.7 nCi/g brain), whereas < 0.6% was as [14C]DHA (0.023 ± 0.006 nCi/g brain). Similar to phospholipids, the remaining 29% of the radioactivity in triglycerides was only in saturated fatty acids and unidentifiable species, likely representing recycled β-oxidation products.
As illustrated by the representative chromatogram in Fig. 4(a), HPLC separation of brain acyl-CoA extracts from six pooled rat brains yielded resolved UV absorbance peaks for the CoAs of 16 : 0, 17 : 0 (internal standard), 18 : 0, 18 : 1n-9, 18 : 2n-6 (LA), 20 : 4n-6 (ARA), 22 : 5n-3 (DPA), and 22 : 6n-3 (DHA) (Fig. 4a). The identities of acyl-CoA species were determined by comparing their retention times with those of authentic acyl-CoA standards (profiles not shown). The CoAs of 14 : 0, 18 : 3n-3 (α-LNA), and 20 : 5n-3 (EPA) co-eluted as a common peak with a retention time of ∼17 min. FAPE derivates prepared for this peak were subjected to HPLC with UV absorbance detection and scintillation counting, which gave the resolved concentrations and radioactivities of these three CoA-coupled fatty acids. As shown in Fig. 4(b) and Table 4, α-LNA-CoA and DHA-CoA radioactivities were 3.3 ± 0.9 and 0.45 ± 0.22 nCi/g brain, respectively (at a ratio of 7 : 1). EPA-CoA and DPA-CoA radioactivities equaled 0.14 ± 0.05 and 0.18 ± 0.03 nCi/g brain, respectively, whereas radioactivity in the CoA species of saturated fatty acids altogether equaled 1.7 ± 0.6 nCi/g brain. Radiolabeled saturated acyl-CoAs most likely arose from recycling of β-oxidation products.
Homeostasis of brain [1-14C]α-LNA
Figure 5 , as determined from the brain radioactivities listed in Table 4, illustrates per cent net brain radioactivity in the compartments arranged according to known pathways of α-LNA diffusion and metabolism. Of net brain radioactivity, 67% was in the aqueous fraction representing β-oxidation metabolites of [1-14C]α-LNA, of which the majority was likely [1-14C]acetate, whereas an additional 4% in cholesterol, 2% in acyl-CoA, 1% in triglycerides, and 12% as phospholipid bound fatty acids represented recycled synthesis products of these metabolites. As we did not measure volatile [14C]CO2, a terminal product of β-oxidation and Kreb's cycle metabolism of [1-14C]α-LNA, we concluded ∼ 86% of the [1-14C]α-LNA that entered brain was subjected to β-oxidation. This value was estimated by summing the percentage of radiolabel present in the aqueous fraction and that recycled into cholesterol plus saturated fatty acids in lipids.
The acyl-CoA pool contained 2, 0.1, 0.1, and 0.3% of the radioactivity within the CoA-derivatives of α-LNA, EPA, DPA, and DHA, respectively. At least 10 and 2% of [1-14C]α-LNA was directly esterified into phospholipids and triglycerides, respectively, while ≤ 0.2% and ≤ 0.01% was esterified into phospholipid and triglycerides, respectively, in the form of synthesized [14C]DHA.
α-LNA incorporation into brain and turnover in brain phospholipid
The second and third data columns in Table 4 list the calculated incorporation coefficients ki* of α-LNA (Equation 2) and of DHA synthesized from α-LNA (Equation 3) into ‘stable’ phospholipids and triglycerides, as well as the corresponding values for incorporation rates from plasma, Jin,i (Equation 4). These rates were obtained by multiplying incorporation coefficients by the plasma concentration of unesterified unlabeled α-LNA (Table 2 and 41 nmol/mL). Although radiolabeled EPA-CoA and DPA-CoA were detected in brain, neither radiolabeled EPA nor DPA could be discerned in the phospholipids or triglycerides, so we could not determine their incorporation coefficients. We did not calculate incorporation coefficients or Jin,i of α-LNA as recycled carbon into cholesterol or into fatty acids within phospholipids and triglycerides, as these values would not represent direct delivery from the α-LNA-CoA pool as required by our model (Robinson et al. 1992).
As shown in Table 5, the calculated steady-state dilution coefficient of α-LNA, λα–LNA–CoA (Equation 5), equaled 0.77. This value represents the flux into the acyl-CoA pool of unesterified α-LNA from plasma, divided by the flux of α-LNA into acyl-CoA as derived from brain phospholipids (see above) (Robinson et al. 1992; Rapoport et al. 2001). Thus, a value of 0.77 indicates that enzymatic cleavage (PLA2) of α-LNA from phospholipids contributes about 23% of the brain α-LNA-CoA, whereas 77% originates from uptake of unesterified α-LNA found in plasma. Table 5 also shows that the rate of incorporation of α-LNA into brain phospholipid from the brain precursor α-LNA-CoA pool, JFA,i (Equation 6a), is 14.3 nmol/s/g brain, that the turnover rate, FFA, of α-LNA in phospholipid equals 60% per h (Equation 7a), and that it has a half-life t1/2 of 1.3 h (Equation 7b).
Table 5. Parameters for incorporation and recycling of α-linolenic acid in rat brain total phospholipids.
JFA nmol/sec/g brain × 10−4
FFA % per h
Data are means ± S.D. (n = 6).
0.77 ± 0.23
14.3 ± 2.3
55.9 ± 12.3
1.3 ± 0.3
This study demonstrates that unesterified radiolabeled α-LNA, circulating in the blood, rapidly enters the brain of unanesthetized rats fed high amounts of DHA, as has been reported for plasma unesterified radiolabeled DHA, ARA (arachidonic acid; 20 : 4n-6) and palmitic (16 : 0) acids (DeGeorge et al. 1991; Washizaki et al. 1994; Grange et al. 1995). Most (∼86%) of the radiolabeled α-LNA that enters the brain rapidly undergoes β-oxidation, whereas only a small fraction is incorporated unchanged into phospholipid (10%) and triglycerides (2%) or as newly synthesized DHA (≤ 0.2%) within these lipids (Table 4, Figs 3 and 5).
Our intravenous infusion conditions followed standard tracer techniques, and rapidly produced a steady-state level of [1-14C]α-LNA in the plasma. The endogenous plasma unesterified α-LNA (41 ± 13 nmol/mL) was not significantly perturbed by infusion of [1-14C]α-LNA, which provided 2.8 ± 0.4 nmol total unlabeled α-LNA. The quantity, if distributed only in plasma, would have raised the plasma unesterified α-LNA concentration by 0.9 ± 0.4%; and thus did not produce a non-tracer effect. The values above were calculated using the amount of [1-14C]α-LNA infused per rat (0.5 mCi/kg), specific radioactivity of the [1-14C]α-LNA (54 mCi/mmol), and estimated whole body plasma volume of rats (29 ± 3 mL/kg) (NIH Guide for the Care and Use of Laboratory Animals; Pub. no. 80–23).
Many studies have examined α-LNA metabolism in mammals by giving the tracer orally, which is the normal route for its entry into the body (Su et al. 1999; Cunnane 2001; Pawlosky et al. 2001; Lefkowitz et al. 2005). We chose a rapid 5-min intravenous infusion to deliver the [1-14C]α-LNA, because an oral route is not able to bypass liver metabolism and leads to a marked delay in brain uptake (Purdon et al. 1997). With an oral dose, it would be very difficult to distinguish synthesized DHA contributed by the liver, via plasma, to the brain from any DHA synthesized within the brain itself. In our study, [14C-DHA] was essentially absent from the plasma at the end of the 5-min infusion (Fig. 2). The albumin-bound unesterified [1-14C]α-LNA that we infused is the normal form that the brain is exposed to for fastest uptake under physiological conditions where brain lipid metabolism is at a constant flow (Rapoport 2001). Additionally, determination of the tracer kinetics of [1-14C]α-LNA uptake and metabolism within a specific organ alone (brain) requires fast delivery of the tracer and tracee at steady-state levels to that compartment (Fig. 1).
Equations derived from our published kinetic incorporation model (Robinson et al. 1992; Rapoport et al. 1997, 2001), when applied to our current experimental data, provided values for incorporation coefficients ki* and incorporation rates Jin,i of plasma-derived unesterified α-LNA into brain phospholipids and triglycerides, and of DHA synthesized within brain from plasma-derived α-LNA into brain phospholipid and triglycerides. The latter equations were valid because [14C]DHA was essentially absent in plasma during the 5-min infusion period. Our equations also provided values for incorporation rates of unlabeled α-LNA from the brain precursor α-LNA-CoA pool, JFA,i, and for the turnover rate and half-life of α-LNA in brain phospholipid, at steady-state unlabeled concentrations.
As illustrated in Table 4, the sum (over i) of Jin,i(α-LNA) for α-LNA in brain phospholipid and trigylceride equaled 12.4 nmol/s/g brain × 10−4, whereas Jin,i(α–LNA→DHA) for DHA synthesized from plasma-derived α-LNA equaled 0.24 nmol/s/g brain × 10−4. The 50 : 1 ratio of these incorporation rates indicates that plasma-derived α-LNA is incorporated into ‘stable’ brain lipids at a much greater rate than is the DHA synthesized from it, in DHA-fed rats. The rate of incorporation of α-LNA and synthesized DHA, however, represents ∼12% of the rate of entry of α-LNA into brain, as the remainder mostly undergoes β-oxidation. A high fractional β-oxidation of α-LNA in brain is consistent with a reported high percentage of α-LNA β-oxidation in the body as a whole (Menard et al. 1998; Cunnane et al. 1999; Cunnane 2001).
The incorporation rate of unesterified plasmaα-LNA in the form of DHA, Jin,i(α–LNA→DHA), into brain phospholipids, 0.22 nmol/s/g brain × 10−4, is 1% of the published incorporation rate of unesterified plasma DHA into phospholipids, which is 17.4 nmol/s/g brain × 10−4 (Chang et al. 1999). This difference arises, not because α-LNA enters brain more slowly than does DHA, but because most (∼86%) of the α-LNA that enters undergoes β-oxidation, compared with only 10% of the DHA that enters (Rapoport 2001). It remains to be seen, however, whether shorter-chain n-3 PUFA precursors of DHA, including both α-LNA and EPA, can be efficiently converted to DHA when using them to treat brain disorders in which DHA may be deficient (Pawlosky and Salem 1999; Conquer et al. 2000; Tanskanen et al. 2001; Morris et al. 2003; Noaghiul and Hibbeln 2003; Kack et al. 2004).
Our rats were fed a diet with a high DHA content, and DHA is reported to directly bind to and activate transcription factors that increase expression of β-oxidation enzyme, and decrease expression of the desaturases (Δ-5 and Δ-6) that elongate it from α-LNA (Cho et al. 1999; Price et al. 2000; Clarke 2001; Nakamura and Nara 2003). In neuronal tumor cell cultures (retinoblastoma), when the cellular content of DHA in phospholipids was increased, the transcription factor PPAR-δ (peroxisome proliferator-activated receptor, subtype-δ) up-regulates the expression of oxidative enzymes (Langelier et al. 2003). Mouse brain expression of PPAR-γ (γ-subtype), but not of other PPARs, is elevated in vivo by a diet containing fish oil rich in DHA (Puskas et al. 2004). In contrast, expression of Δ6-desaturase has been shown in the liver to be increased by PPAR-α (α-subtype) activation (Tang et al. 2003). Instead, Δ-5 and Δ-6 desaturase expression in liver are positively controlled by the transcription factors SREBP-1 and NF-Y (sterol regulatory element binding protein-1 and nuclear factor-Y), which are both turned off when DHA is abundant (Matsuzaka et al. 2002; Nara et al. 2002). Thus, if the diet instead contained α-LNA alone without DHA, it is likely that the rate of conversion of α-LNA to DHA in both brain and liver would have been elevated and its rate of β-oxidation reduced. The likelihood for this is predicted from a number of studies where active DHA synthesis is detected in the absence of DHA (Dwyer and Bernsohn 1979; Emken et al. 1999; Moore 2001; Williard et al. 2001). We could test whether conversion of α-LNA to DHA within brain or liver is up-regulated in animals fed α-LNA but not other n-3 PUFAs with the method of this paper. DHA synthesis might also be induced in the brain and liver of rats by subjecting them to prolonged n-3 PUFA deprivation over a single generation (DeMar et al. 2004a) and then abruptly feeding them α-LNA.
We have hypothesized (Rapoport et al. 2001; Rapoport 2003) that the rate of metabolic loss of DHA in brain should equal the rate of its replacement by plasma-derived unesterified DHA plus α-LNA, as n-3 PUFAs cannot be completely synthesized de novo from acetyl-CoA in vertebrate tissues. Now that we have shown, under DHA enriched dietary conditions, that the sum (over i) Jin,i(DHA) >> Jin,i(α–LNA→DHA), we have proven that the sum of Jin,i(DHA) (over i) is identical to the rate of DHA loss from brain. This conclusion is supported by evidence that DHA loss from rat brain, as calculated from its measured half-life of disappearance, approximates Jin,i(DHA) (Contreras et al. 2000; DeMar et al. 2004b).
The major determinant of the diffusion rate into brain of a circulating unesterified long chain fatty acid is its rate of disassociation from plasma albumin, whereas a major factor that governs its retention in brain is its esterification to Coenzyme A (acyl-CoA) by long chain acyl-CoA synthetase (Marcel and Suzue 1972; Robinson and Rapoport 1986; Robinson et al. 1992; Waku 1992; Rapoport 2003). From the acyl-CoA pool, the fatty acid can be trans-esterified into ‘stable’ triglycerides and phospholipids (Yamashita et al. 1997), or be shuttled into mitochondria after transfer to carnitine by carnitine O-palmitoyl transferase (CPT) and undergo β-oxidation (Fig. 5). The catalytic efficiency of CPT for α-LNA-CoA (Vmax/Km), as bound to albumin, was found to be threefold higher than that for DHA-CoA (Gavino and Gavino 1991). This selectively may provide ‘an enzymatic rationale’ for the relatively low content of α-LNA in esterified lipid, and for the high fractional rate of β-oxidation of α-LNA in brain compared with that of DHA (∼9-fold higher than that of DHA). As in brain, the body as a whole diverts most ingested α-LNA (> 80%) towards β-oxidation (Yang and Cunnane 1994; Cunnane 2001).
Reported turnover rates of DHA and ARA in net brain phospholipid are 0.9 and 3.6% per h, respectively (Rapoport 2001). They correspond to half-lives of 77 and 19 h, although some turnover rates in individual phospholipids, such as transitory molecular species of PI, are higher with corresponding half-lives of a few hours (Washizaki et al. 1994; Shetty et al. 1996; Chang et al. 1999; Contreras et al. 2000, 2001). Reported values for λacyl–CoA (Equation 5) of DHA and ARA are 0.02–0.04, suggesting, according to our model (Robinson et al. 1992; Rapoport 2001), that 96–98% of brain ARA-CoA or DHA-CoA comes from fatty acid released from phospholipid and triglycerides, compared with 2–4% from plasma. In contrast, the half-life for α-LNA in net brain phospholipid equaled 1.3 h, and λα–LNA–CoA equaled 0.77 (Table 5), implying that plasma contributes to about 77% of brain α-LNA-CoA, compared with 23% from hydrolysis of phospholipids.
Rat brain phospholipids contain only a trace of esterified α-LNA (Table 3), in agreement with the literature (Sastry 1985). Nevertheless, 10% of radioactivity was in brain phospholipid after 5 min of constant [1-14C]α-LNA infusion (Fig. 5). These observations, and our finding that the half-life of unlabeled α-LNA in phospholipid equals 1.3 h (Table 5), show that α-LNA in brain phospholipids is rapidly recycled and, when released, undergoes extensive β-oxidation similar to what we found for plasma-derived α-LNA. Recycling (deacylation and reacylation) of polyunsaturated fatty acids ARA and DHA in brain phospholipids is an ongoing and active process (Lands and Crawford 1976; Rapoport 2001), but unlike α-LNA they are much less susceptible to β-oxidation and thus are conserved in the phospholipids by the recycling process. In pregnant rhesus monkeys, DHA undergoes ∼80% less β-oxidation than α-LNA, as determined by recycling of β-oxidation products into maternal plasma and fetal brain fatty acids (Sheaff Greiner et al. 1996). Physiological differences in the potential for β-oxidation thus may largely account for the maintained abundance of ARA and DHA, and the near absence of α-LNA, in neuronal membranes. Enrichment of specific polyunsaturated fatty acids in brain does not appear to be as a result of any discriminatory uptake by PUFA transporters at the blood–brain barrier (Edmond 2001).
Below normal blood levels of DHA or reduced dietary n-3 PUFA intake has been reported to be associated with bipolar disorder, Alzheimer disease, depression, chronic alcoholism, and age-related cognitive dysfunction (Pawlosky and Salem 1999; Conquer et al. 2000; Tanskanen et al. 2001; Morris et al. 2003; Noaghiul and Hibbeln 2003; Whalley et al. 2004). Furthermore, although some clinical trials have suggested that feeding long chain n-3 PUFAs, such as EPA and DHA, is therapeutically effective in bipolar disorder (Stoll et al. 1999), a randomized control trial with dietary EPA did not show a beneficial outcome (Kack et al. 2004). Rigorously designed trials are lacking and it is not certain whether DHA itself, one of its more immediate n-3 precursors such as EPA, or both, should be used in these trials (Kidd 2004; Marangell et al. 2004). In this regard, the results of our study suggest α-LNA does not have great potential to effectively serve as a precursor for synthesizing DHA within the adult brain, when substantial DHA is found in the diet.
In summary, we have demonstrated quantitatively that α-LNA circulating in the plasma can enter the brain of unanesthetized adult rats fed a DHA-enriched diet. However, the amount that enters is largely lost to β-oxidation products and little if any is used for synthesizing DHA found esterified into phospholipids. Because of this, the rate of influx of plasma unesterified DHA across the blood–brain barrier and into brain phospholipids, Jin,i(DHA), alone represents the rate of loss by metabolism of brain DHA.
Additionally, these observations suggest that dietary supplementation with n-3 PUFAs to increase brain DHA content might consist of only DHA and not have to include its precursors, when DHA is already present in the diet. It should be recognized, however, that consumption of large amounts of α-LNA is capable of supporting nearly normal levels of brain DHA, when DHA is completely absent from the diet (Abedin et al. 1999; DeMar et al. 2004a; Lefkowitz et al. 2005). In this case, it is possible that the β-oxidation of α-LNA is greatly suppressed and DHA synthesis strongly up-regulated. Our study also did not directly address the degree to which the brain oxidizes, stores, and elongates EPA or DPA, two more immediate dietary precursors to DHA.