Introduction

L-ornithine-L-aspartate (LOLA) is a compound used in the management of hepatic encephalopathy [1, 2] where it has shown significant efficacy [3]. It is ingested as a crystalline salt with the L-ornithine and L-aspartate hydrogen bond separating on dissolution or exposure to gastric acidity. Both of these amino acids enter the urea cycle and, among other reactions, help to produce glutamine and urea, reducing the levels of ambient ammonia. Much of this activity takes place in liver, kidney and muscle [4]. This makes it difficult to dissect the direct effects of LOLA itself on brain metabolism from peripherally induced effects associated with plasma ammonia reduction. In healthy brain, the impact of LOLA administration on brain metabolism remains largely unknown.

There is good evidence that orally administered L-ornithine enters the brain. It is taken up from plasma by sodium-independent cation transporters from system y+, which constitute a sub-family of solute carrier 7 (SLC7) [5, 6] and are present widely in different cells in the brain [7]. Ornithine can enter the mitochondria via the mitochondrial ornithine carrier (ORNT1; SLC25A15) [8]. In brain, the main metabolic fate for L-ornithine is likely conversion to glutamate and GABA via ornithine aminotransferase [9], with feedback inhibition by GABA [10]. In line with this, L-ornithine has been suggested to act as a sedative, as oral administration of L-ornithine produces sedative-like effects on mouse behavior [11].

The case for L-aspartate entering the brain is much less clear. A “low affinity” transporter (ASCT2; SLC1A5) which handles L-aspartate has been demonstrated in brain endothelial cells and could be responsible for control of L-aspartate at the blood brain barrier [12, 13]. ASCT2 has been reported to have very low affinity (KM > 60 mM) for L-aspartate [14] and [15] have suggested it becomes a meaningful substrate only at very low pH. Uptake of aspartate by brain has also been demonstrated in conditions where plasma levels are kept artificially high for relatively long periods, with resultant brain levels being higher than plasma levels [16]. This suggests that longer term supplementation with LOLA may result in aspartate entry into brain but others have suggested that the route of administration may be important. Orally administered aspartate or aspartate that interacts with mucosal cells is likely rapidly converted via transaminases to alanine and oxaloacetate, meaning that the resulting plasma levels of aspartate may be significantly reduced [17].

Metabolism of ornithine has been reported to be quantitatively significant in brain [10] with strong evidence that it is converted into GABA and glutamate, mostly via the ornithine aminotransferase catalyzed reaction [18] with GABA providing inhibitory feedback control [9, 10] (Fig. 1).

Fig. 1
figure 1

Scheme showing likely pathways for brain metabolism of L-ornithine and L-aspartate. Pathways for metabolism of ornithine (in green) and aspartate (red). Abbreviations: AGAT, arginine-glycine amidinotransferase; GAD, glutamate decarboxylase; GSSA, L-glutamic acid-γ-semialdehyde; GSSA deH, L-glutamic acid-γ-semialdehyde dehydrogenase; GDH, glutamate dehydrogenase; GOT, Glutamate-oxoglutarate aminotransferase; GS, glutamine synthetase; OAA, oxaloacetate; 2-OG, 2-oxoglutarate (Color figure online)

Full size image

L-aspartate is an “excitatory” amino acid which plays key roles in metabolism, as an amino donor and acceptor. It contributes to the synthesis of protein, of arginine and nitric oxide, asparagine, N-acetylaspartate and N-methyl-D-aspartate. Its major metabolic role in the brain is in the recycling of reducing equivalents (NADH and H+) between the cytoplasm and mitochondrial matrix as part of the malate-aspartate shuttle (Fig. 1). It also serves to help balance nitrogen levels between mitochondria and cytosol [19].

There is interest in the use of LOLA to treat other disorders beyond hepatic encephalopathy, such as in mutations of the aspartyl-tRNA synthetases (DARS; aspartyl-aminoacyl-RNA-synthetase), whose role is to accurately pair the appropriate t-RNA with aspartate and ensure exactitude in protein synthesis [20]. Homozygous and compound heterozygous DARS mutations result in the leukodystrophy Hypomyelination with Brainstem and Spinal cord involvement and Leg spasticity (HBSL) [21, 22].

Here, in order to better understand the impact of LOLA administration on brain metabolism, we studied the metabolism of LOLA, L-aspartate and L-ornithine in the guinea pig cortical tissue slice preparation. This represents a reductionist model of brain metabolism without the interference of the blood brain barrier or peripheral metabolism [23]. Additionally, we studied the impact of intraperitoneal (i.p.) administration of LOLA on brain metabolite levels in vivo in a mouse model where the blood brain barrier and peripheral metabolism are intact.

Materials and Methods

Experiments were carried out according to the guidelines of the National Health and Medical Research Council of Australia and were approved by the institutional (UNSW) Animal Care Ethics Committee. Guinea pigs (Dunkin-Hartley), both male and female (obtained from Pipers Farm, NSW, Australia and Flinders University, Australia respectively) and weighing 400–800 g, were fed ad libitum on standard Guinea pig/rabbit pellets, with fresh carrots and hay roughage, and were maintained on a 12 h light/dark cycle. 12-week-old, female C57Bl6J mice were obtained from ABR Moss Vale. Mice were housed in groups of 4–5 and fed ad libitum with standard chow diet.

Sodium [1,2-13C]acetate and [1-13C]D-glucose, and sodium [13C]formate were purchased from Cambridge Isotope Laboratories Inc (Andover, MA, USA). L-aspartate, L-ornithine, and LOLA (L-ornithine-L-aspartate) were purchased from Sigma-Aldrich (St Louis, MO, USA). All other reagents were of Analytical Reagent grade.

Effects of LOLA, Ornithine and Aspartate on Guinea pig Brain Cortical Tissue Slice Metabolism

Cortical tissue slices were prepared from guinea pigs as described previously [23]. The following experiments were performed, each using four guinea pig brains:

Incubation of 0.5 mM sodium [1,2-13C]acetate and 5.0 mM [1-13C] D-glucose (control) with 20 or 100 µmol/L L-ornithine.

Incubation of 0.5 mM sodium [1,2-13C]acetate and 5.0 mM [1-13C] D-glucose (control) with 20 or 100 µmol/L L-aspartate.

Incubation of 0.5 mM sodium [1,2-13C]acetate and 5.0 mM [1-13C]D-glucose (control) with 20 or 100 µmol/L LOLA.

Incubation of 0.5 mM sodium [1,2-13C]acetate and 5.0 mM [1-13C] D-glucose (control) with 20 or 100 µmol/L L-ornithine and L-aspartate. This latter experiment was conducted as the commercially available preparation of LOLA was found to contain significant amounts of methanol (~ 4% of LOLA concentration, measured by NMR spectroscopy) and ethanol (10% of LOLA concentration). Ethanol, even at very low concentrations, is known to have significant effects on brain metabolism [24] although the lowest concentration of ethanol previously tested by us was 100 µmol/L.

After 90 min the slices were removed from the buffer by rapid filtration and were frozen in liquid nitrogen. To extract metabolites, pulverised frozen tissue was extracted using chloroform/methanol [25]. The aqueous phase was lyophilised and reconstituted in 2H2O containing 2 mM sodium [13C]formate as an internal intensity reference, and 6 mM EDTA as a chelating agent to remove paramagnetic species [26].

Uptake of LOLA, L-Asp and L-Ornithine Administered i.p. by Mouse Brain

12-week-old, female C57Bl6J mice (N = 24) were divided into four groups (N = 6 each). Each group received an i.p. injection of 200 uL of phosphate buffered saline (PBS) vehicle (control) containing either 100 mmol/L L-aspartate, 100 mmol/L L-ornithine or 100 mmol/L LOLA. The mice were injected in sequential order (control, L-asp, L-ornithine, LOLA) each at five-minute intervals to minimise order effects. One hour after injection mice were euthanised by cervical dislocation and the brain, kidney and liver rapidly removed. The brain was rapidly dissected and the brain stem (pons and medulla) and cortices were weighed and snap frozen in liquid nitrogen as was the left kidney and an identical section of liver.

The frozen tissue samples were stored at − 80 °C. Frozen tissue samples were subsequently extracted for 1H NMR analysis as described below. Lyophilised extracts were resuspended in 2H2O containing 2 mM [13C]formate as an internal concentration reference and dispensed into 3 mm Norell NMR tubes for analysis.

NMR Analysis

1H, {13C}-decoupled 1H and {1H}-decoupled 13C NMR spectra were acquired from the Guinea pig slice extracts using a Bruker AVANCE III HD 600 NMR spectrometer equipped with a TCI cryoprobe and refrigerated sample changer, as described previously [27, 28]. Peaks areas were compared to the area of the [13C]- formate peak in the case of the 13C spectra and adjusted for saturation and nuclear Overhauser effect by reference to a standardised fully relaxed (TR = 90 s) 13C spectrum acquired with 1H decoupling only during the acquisition time. Peaks in the fully relaxed (TR = 30 s) 1H spectrum were referenced to the area of the 13C satellites from the known concentration of [13C]-formate. Values were expressed as µmol of metabolite per 100 mg protein where protein concentration was measured in the pellet obtained after extraction using the method of Lowry [29] as described previously [30].

Isotopomers were inferred to have originated from either [1-13C]D-glucose or [1,2-13C]acetate or likely to have come from either as discussed previously [28], according to the scheme shown in Fig. 2.

Fig. 2
figure 2

Scheme showing metabolism and distribution of 13C from [1-13C]glucose and [1,2-13C]acetate. [1-13C]glucose (blue) generates two molecules pyruvate, one unlabeled and one molecule of [3-13C]pyruvate. 3-13C]Pyruvate can enter the Krebs cycle via pyruvate dehydrogenase (PDH) and subsequently citrate synthase (CS) to form [2-13C]citrate. It can also be subject to carboxylation by pyruvate carboxylase (PC) and form oxaloacetate and subsequently [3-13C]aspartate and [3-13C]glutamate. [1,2-13C]Acetate (red) enters the Krebs cycle via acetyl-CoA synthetase (AceCS) and subsequently CS to form [1,2-13C]citrate. In the first turn of the Krebs cycle, [1-13C]glucose exclusively labels [4-13C]glutamate, [2-13C]GABA and [4-13C]glutamine. [1,2-13C]Acetate exclusively labels [4,5-13C]glutamate, [1,2-13C]GABA and [4,5-13C]glutamine

Full size image

Pattern Recognition of the Data

In order to better detect similarities and differences among the metabolic profiles resulting from the addition of L-ornithine, L-aspartate, LOLA, and L-ornithine and L-aspartate the data were subjected to principal components analysis using SIMCA P+ (v11.5; Umetrics, Umeå, Sweden) with each variable expressed as a ratio relative to the control mean (N = 4) as described previously [31]. Data were subjected to unit variance scaling to ensure that each variable contributed equally to the model. A goodness of fit algorithm was used to assess the robustness of the resultant model (generally accepted as R2 > 0.6) which is then cross validated by calculating a predicted residual sum of squares (Q2) [32]. Q2 has a maximum value of 1 (for 100% accurate prediction), while a negative value implies no prediction.

Statistical Analyses

All statistical analysis was done in SPSS (IBM Statistics, 24). The slice experiment is designed as a population with repeated sampling. All data with more than two groups were imported and analysed by repeated measures ANOVA followed by Scheffé’s post hoc test using SPSS (v22), as described previously [28]. Only those variables which were determined to be significantly different via the Scheffé’s post hoc test were then subject to the non-parametric Mann-Whitney U test to calculate a P-value. Only those comparisons which passed both of these tests were reported as significant. Data are presented as mean ± SD and are all N = 4 unless otherwise indicated.

Data from mice injected with saline, L-ornithine, L-aspartate or LOLA was similarly analysed by repeated measures ANOVA followed by Scheffé’s post hoc test using SPSS (v22), then Mann-Whitney U test to calculate a P-value where significance had previously been indicated by Scheffé’s post hoc test.

Results

Metabolism of [1-13C]D-Glucose and [1,2-13C]Acetate in the Presence of 20 and 100 µmol/L L-Ornithine

Incubation of cortical tissue slices with 20 µmol/L L-ornithine resulted in significant increases in the total metabolite pool sizes of lactate, GABA and alanine (Fig. 3) with significant reductions in glutamate, aspartate and glutamine. Increasing the concentration of L-ornithine to 100 µmol/L resulted in decreased pool sizes of all metabolites measured (Fig. 3). Resonances from the C5 CH2 “triplet” of ornithine were detected in the 1H NMR spectra at δ = 3.04 ppm but it was not possible to quantitate the amount with any accuracy due to overlap with the nearby creatine (N-CH3) and glutathione Cys βCH2 resonances [33]. Flux of 13C was decreased into all isotopomers apart from into lactate C3 and Ala C3, which were both significantly increased (Fig. 3). The decrease in net flux was seen in isotopomers made from both [1-13C]D-glucose and [1,2-13C]acetate (Fig. 3).

Fig. 3
figure 3

Distribution of 13C label following incubation of brain cortical tissue slices with 5 mmol/L [1-13C]D-glucose and 0.4 mmol/L sodium [1,2-13C]acetate (control) and 20 or 100 µmol/L L-ornithine. Bars show mean and error bars show standard deviations, N = 4. *P < .05 as shown with statistical significance between pairs indicated by the line below the asterisk

Full size image

Taken together, this suggests that L-ornithine resulted in significant slowing of flux into the Krebs cycle, with resultant slower pyruvate clearance leading to increased lactate and alanine levels. The higher concentration of L-ornithine produced further Krebs cycle and glycolytic slowing resulting in smaller pool sizes and reduced net flux into the Krebs cycle.

Metabolism of [1-13 C]D-Glucose and [1,2-13 C]Acetate in the Presence of 20 and 100 µmol/L L-Aspartate

Administration of L-aspartate to cortical slices gave contrasting results to L-ornithine, with 20 µmol/L L-aspartate resulting in significant increases in the metabolite pool sizes of all measured metabolites (Fig. 4). Increasing the concentration to 100 µmol/L L-aspartate resulted in further increases in metabolite concentrations apart from that of lactate, which decreased compared to the lower concentration of L-aspartate but which was still significantly higher than the control.

Fig. 4
figure 4

Distribution of 13C label following incubation of brain cortical tissue slices with 5 mmol/L [1-13C]D-glucose and 0.4 mmol/L sodium [1,2-13C]acetate (control) and 20 or 100 µmol/L L-aspartate. Bars show mean and error bars show standard deviations, N = 4. * P < .05 as shown with statistical significance between pairs indicated by the line below the asterisk

Full size image

Net flux of 13C was increased into all isotopomers in the presence of 20 µmol/L L-aspartate apart from Asp C2, Ala C3 and Cit C2/4. Increasing the concentration of L-aspartate to 100 µmol/L resulted in further significant increases in net flux of 13C into all isotopomers measured (Fig. 4). The increase in flux into glutamate C4 and GABA C2 was largely driven by increased incorporation of 13C from [1,2-13C]acetate (Fig. 4).

Overall, L-aspartate had strong, concentration dependent, activating effects on metabolism with the lower concentration having more relative impact on astrocytic metabolism as indicated by significant effects on net flux of 13C derived from [1,2-13C]acetate.

Metabolism of [1-13 C]D-Glucose and [1,2-13 C]Acetate in the Presence of 20 and 100 µmol/L L-Ornithine-L-Aspartate

Addition of 20 µmol/L LOLA resulted in significant increases in all measured metabolites (Fig. 5) along with significant increases in net flux of 13C into Glu C4, GABA C2, Asp C3 and Ala C3 and lactate C3 (Fig. 5). Net flux from [1-13C]D-glucose was increased into Glu C4, GABA C2, Asp C3, Ala C3, citrate C2 and lactate C3 (Fig. 5). Net flux 13C derived from [1,2-13C]acetate was significantly increased into Glu C4,5, GABA C1,2 and citrate C1,2 (Fig. 5).

Fig. 5
figure 5

Distribution of 13C label following incubation of brain cortical tissue slices with 5 mmol/L [1-13C]D-glucose and 0.4 mmol/L sodium [1,2-13C]acetate (control) and 20 or 100 µmol/L L-ornithine-L-aspartate. Bars show mean and error bars show standard deviations, N = 4. * P < .05 as shown with statistical significance between pairs indicated by the line below the asterisk

Full size image

Increasing the concentration of LOLA to 100 µmol/L further increased the pool size of aspartate but the pools of glutamate, lactate and alanine, although higher than control, were not increased to the same extent as with 20 µmol/L LOLA (Fig. 5). Resonances from L-ornithine were detected in the 1H NMR spectrum. The higher concentration of LOLA further increased levels of Glu C4, Asp C2 and C3, citrate C2 and Ala C3 but levels of GABA C2 and lactate C3 were similar to control (Fig. 5). Label derived from [1-13C]D-glucose was incorporated less than with 20 µmol/L LOLA, being increased in Glu C4, Asp C3 and Ala C3, while the only increase in labelling from [1,2-13C]acetate was in Glu C4,5, where it was still significantly less than with 20 µmol/L LOLA (Fig. 5).

Taken together, the results suggest that the additive effects of L-ornithine and L-aspartate are concentration dependent. Ornithine tends to decrease fluxes and pool sizes while aspartate increases them. Together, they have different effects on individual metabolic compartments, making the net outcomes highly concentration dependent.

Metabolism of [1-13 C]D-Glucose and [1,2-13 C]Acetate in the Presence of 20 and 100 µmol/L L-Ornithine and L-Aspartate

The metabolic profile generated by separate addition of L-ornithine and L-aspartate rather than the LOLA salt (Fig. 6) was largely similar to that of LOLA (Fig. 5) although some differences were noted (see below). The aspartate and ornithine preparations did not contain additional ethanol and methanol like the LOLA salt. Resonances from L-ornithine were detected in the 1H NMR spectrum as in samples from experiments with 100 µmol/L LOLA and µmol/L L-ornithine.

Fig. 6
figure 6

Distribution of 13C label following incubation of brain cortical tissue slices with 5 mmol/L [1-13C]D-glucose and 0.4 mmol/L sodium [1,2-13C]acetate (control) and 20 or 100 µmol/L L-ornithine and L-aspartate. Bars show mean and error bars show standard deviations, N = 4. * P < .05 as shown with statistical significance between pairs indicated by the line below the asterisk

Full size image

Pattern Recognition of the Data

A plot showing variable distribution under the first two components of a principal component (PC) model using data from all four slice experiments is shown in Fig. 7. The model generated eight components accounting for 99% of the variance in the data (Q2 = 0.94) of which the first four components accounted for 95% of the variance (Q2 = 0.88; PC1 = 0.63, PC2 = 0.20, PC3 = 0.07; PC4 = 0.05). The first principal component is dominated by increased levels of all metabolites and increased fluxes into all variables apart from the lactate pool and lactate C3 (Fig. 7), both of which have negligible contributions to this principal component. The second principal component, the loading of which are shown in Fig. 7, is dominated by the net fluxes into lactate and Ala and the pool sizes thereof, with lesser contributions from decreases in net fluxes into Asp isotopomers and the pool sizes of Glu and GABA.

Fig. 7
figure 7

Plot of first two principal components from principal components analysis of labelled and total metabolite levels from brain cortical slice experiments with L-aspartate, L-ornithine, LOLA, and L-aspartate together with L-ornithine. This plot demonstrates the differences between these data sets as it includes the changes in each metabolite or labelling level relative to the control mean for that experiment. Clear symbols represent lower (20 µmol/L) concentrations, filled symbols 100 µmol/L concentrations. Black squares, L-ornithine; red ellipses, L-aspartate; Blue diamonds, LOLA and green triangles, L-ornithine plus L-aspartate. The large outer ellipse represents the 95% confidence interval (Hotellings circle). The relative loadings of each variable on principal component 1 (PC1) and 2 (PC2) are shown as bar plots

Full size image

The experiments with L-ornithine can be seen in the left hand quadrants of the Hotelling’s circle, reflecting the overall reduction in net fluxes into most variables with L-ornithine apart from lactate C3 and underlining the general sedative effect of L-ornithine on brain energy metabolism. In keeping with its stimulatory effect on brain metabolism, the two concentrations of aspartate fall in the bottom right hand quadrant.

Aspartate (20 µmol/L), Asp + Ornithine (100 µmol/L) and LOLA (100 µmol/L) clustered relatively closely with Asp + Ornithine (20 µmol/L) showing less weighting on PC1. LOLA (20 µmol/L) while similarly weighted on PC1 was separated by PC2 (Fig. 7), which reflects the higher concentrations of lactate and alanine seen in these samples.

Mouse Brain Metabolites Following i.p. Injection of LOLA, L-Ornithine or L-Asp

There were no significant changes in brain metabolite levels one hour following i.p. injection of L-aspartate, L-ornithine or LOLA (Fig. 8) apart from a decrease in brain glutamate following injection of L-ornithine (P = .041). There were no significant changes in the levels of measured liver lobe metabolites but there was a significant increase in kidney Asp and Ala following injection of LOLA and a significant increase in Asp following injection of L-ornithine.

Fig. 8
figure 8

Metabolite levels in brain cortex, brain stem, liver and kidney following intraperitoneal injection of PBS, L-aspartate, L-ornithine, or LOLA. 12-week-old, female C57Bl6J mice were intraperitoneally injected with 200 µL of phosphate buffered saline (PBS) vehicle (control; white bars) containing either 100 mmol/L L-aspartate (grey bars), 100 mmol/L L-ornithine (blue bars) or 100 mmol/L LOLA (red bars). Displayed are total metabolite pool sizes. Bars show mean and error bars indicate standard deviations, N = 6. * P < .05 as shown with statistical significance between pairs indicated by the line below the asterisk (Color figure online)

Full size image

Discussion

Metabolism of Asp, Ornithine and LOLA in Brain Cortical Tissue Slices

L-Aspartate has actions at synaptic receptors and may or may not be a true neurotransmitter. It is known as an “excitatory” amino acid, along with glutamate and, metabolically, the actions of exogenous aspartate support that conclusion with increased pool sizes and, generally, increased net fluxes into metabolites from [1-13C]glucose (Fig. 5). The situation with metabolism of [1,2-13C]acetate is more nuanced, with increased net flux into the main metabolites normally synthesized from acetate (citrate C1,2 and Gln C4,5; Fig. 5) along with decreased net flux into Glu C4,5 indicating that, as the metabolic rate increases, the slices favour oxidation of glucose over acetate. This outcome is similar to that previously reported, where increasing metabolic workload has been shown to favour oxidation of glucose over acetate in both neurons and glia [28]. The metabolic consequences of increasing the concentration of exogenous aspartate are also similar to those seen with increasing concentrations of glutamate [34]. In the case of increasing glutamate concentrations it is known that the system will reach a point of “metabolic exhaustion” through over stimulation once exogenous glutamate levels are too high [35]. It is not known if this is also the case with aspartate but it is clear that this point has not been reached at the highest exogenous concentration used here, 100 µmol/L as decline of metabolic pool sizes and net fluxes, which would be the expected outcome if over stimulation was happening [36] was not seen (Fig. 4).

The metabolic effects of L-ornithine in the cortical tissue slice contrasted sharply with those of L-aspartate with ornithine having a largely inhibitory effect on metabolism. Previous work has shown that ornithine can be metabolized in brain to produce GABA via the ornithine aminotransferase route (Fig. 1) [9] with subsequent GABA-mediated feedback inhibition [10]; ornithine administration has also been reported to have a “sedative” effect [11].

Here, we see that this effect is dose dependent, with large reductions in net flux into Krebs cycle intermediates with decreased pyruvate clearance leading to increased labelling and pools of lactate and Ala, in the presence of 10 µmol/L L-ornithine. Increasing L-ornithine to 100 µmol/L lead to further reductions in net flux from both [1-13C]glucose and, to a lesser extent, [1,2-13C]acetate but relatively large (~ 40%) reductions in metabolite pool sizes indicative of a strong sedative effect of L-ornithine (Fig. 3). The presence of small resonances from L-ornithine in 1H spectra from samples incubated with 100 µmol/L L-ornithine suggests that slice uptake is not rate-limiting to ornithine metabolism and that increasing L-ornithine concentration further is unlikely to result in further proportionate metabolic effects.

The impact of ornithine on net flux is not due to dilution of label by unlabeled ornithine; inspection of isotopomer fractional enrichment levels shows them remaining largely stable, with the main impact on labelling of isotopomers in the second turn of the Krebs cycle, such as Glu C3,4, GABA C2,3 and Asp C2,3 (data not shown), indicating that flux through the Krebs cycle is slowing with less label penetrating to the second turn of the cycle.

Having established that L-ornithine and L-aspartate have opposite metabolic effects, with one being sedative-like and one being excitatory, the metabolic impact of incubation with LOLA, or with L-ornithine and L-aspartate together, can then be explained as a competition between these two effects. At the lower concentration of LOLA, pool sizes are maintained or slightly increased by L-ornithine, and increased by L-aspartate but at 100 µmol/L LOLA pool sizes are reduced relative to 20 µmol/L LOLA as the L-ornithine “sedative” effect is more pronounced. The pool size of Asp is an obvious exception to this, due to uptake of the exogenous L-aspartate. The stimulatory effect of the L-aspartate in LOLA on net flux into Krebs cycle intermediates is also apparent at the higher LOLA concentration.

Incubation of slices with separate preparations of L-ornithine and L-aspartate together, as opposed to the commercial LOLA preparation, yielded similar outcomes with some minor differences which are more apparent when viewing the PCA plot (Fig. 7) which is very sensitive to differences. It is not possible here to determine which of these differences are due to the significant contaminants in the LOLA preparation, which were mostly ethanol and methanol, or experimental variation. Regardless of these differences, the general conclusions from these experiments remain.

The effects of LOLA on healthy brain metabolism therefore derive from competing effects of L-ornithine and L-aspartate and, as a consequence, are highly dose dependent.

Metabolic Impact of Asp, Ornithine and LOLA Administration in Mouse in vivo

Here, i. p. injection of Asp, Orn or LOLA had no significant impact on the levels of measured metabolites in cortex or brain stem one hour after injection apart from a decrease in glutamate levels following injection of L-ornithine. Previous work has examined uptake on a longer time scale (5–6 h of infusion; [4, 37, 38]) and, most often, under conditions of hyperammonaemia or liver failure.

A study that infused LOLA, L-ornithine, and L-aspartate for four hours into portacaval-shunted rats treated with ammonia to model hyperammonemia-induced encephalopathy [38] collected brain dialysates at conclusion of four hours of infusion and assayed for aspartate, and glutamate. Dialysate aspartate levels were increased in LOLA and L-Asp treated rats but not in L-ornithine treated rats compared to control (Ctl 2.42 ± 0.39; N = 9: LOLA 30.5 ± 1.2; N = 5: L-Ornithine 1.79 ± 0.39; N = 6: L-Asp 22.7 ± 2.0; N = 4; mean ± s.e.m. µmol/L ). The investigators also reported higher dialysate glutamate in the LOLA and L-Asp groups but not in the L-ornithine group compared to control.

However, these results conflict somewhat with the measurements made in the same rats (LOLA, L-ornithine and control rats only) using 1H MRS in vivo, where no differences in glutamate concentration was reported between the three groups. Glutamine and lactate levels were lower in both LOLA and L-ornithine groups compared to control, but not significantly different from one another [38].

Uptake of aspartate by brain has been reported to be minimal, with no transporter facilitating significant transport [16] and would not have been expected to take place under these conditions. The majority of liver uptake of aspartate has been reported to be restricted to a small population of perivenous cells which contain the majority of liver glutamine synthetase and act as “gatekeepers” for dealing with elevated ammonia [39] which may explain why little change was seen in liver lobe samples from mice injected with aspartate in this experiment. In kidney, aspartate uptake has been reported to be fractionally higher than in liver [40] but inhibited strongly by equivalent amounts of glutamate and confined, similarly to the case in liver, mostly to peritubular cell membranes [41]. Conversion of aspartate to other metabolites by aspartate aminotransferase is also a confounding factor [17].

Ornithine administered orally (3 mmol/10 ml/kg, which we take to be approximate to 300 mM concentration of solution) has previously been shown to increase cortical ornithine concentrations rapidly, with the highest recorded concentration 30 min after administration, the earliest time point measured [11]. Brain arginine and citrulline levels were not impacted at this point, or at 60 min post administration suggesting that metabolism of the ornithine was slower, although the authors did not measure levels of glutamate or GABA. They reported behavioral effects with only one of the three concentrations of ornithine used (0.75 mmol/10 ml/kg; i.e. 75 mM to attempt to equate it to the concentration used here). Inspection of 1H NMR spectra from the cerebral cortex in our experiment showed no detectable resonances from L-ornithine although the highest concentration reported by Kurata et al. at 30 min in the cortex was 16 nmol/g wet weight, which is unlikely to be detected by NMR. The decrease in brain glutamate concentration seen in the cortex (Fig. 8) is consistent with the decrease in glutamate pool size seen in cortical slices when incubated with L-ornithine (Fig. 3).

In humans with cirrhosis (22), including a small number (7) with minimal hepatic encephalopathy, treated for 28 days with 9 g/day oral LOLA no significant differences compared to baseline were reported in brain metabolite levels, including glutamate/glutamine, with no changes on brain diffusion, magnetization transfer or brain volumes (edema) [42]. This could be because the route of administration of the LOLA was sub-optimal, or because LOLA simply does not penetrate brain to any degree under these circumstances, or because the magnetic resonance spectroscopy method was not sufficiently sensitive to detect a change. In this latter case, the authors reported a small coefficient of variation for their measures which should have been sufficient to detect a change of the order reported here in slices with LOLA suggesting that LOLA penetration of the brain via this delivery route is poor.

The only change in any non-brain tissue metabolite one hour after i.p. injection in this experiment was recorded in the kidney, where administration of L-ornithine, and LOLA resulted in a significant increase in kidney aspartate levels, while administering LOLA also resulted in increased kidney alanine (Fig. 8). Infusion of L-ornithine has previously been reported to reduce kidney blood flow by competing with L-arginine uptake and reducing production of NO [43], which may be inferred to impact kidney urea cycle activity and result in increased aspartate levels.

The presence of unknown factors in liver disease has been speculated to impact brain uptake of ornithine via modulation of system y + in the absence of symptoms [44] and via modulation of the blood brain barrier in more severe disease [45]. Here, in healthy mice with an intact blood brain barrier, LOLA, ornithine and aspartate appear not to have any significant impact on brain metabolism an hour after i.p. administration. This suggests that chronic administration of LOLA is indicated if significant brain penetration is a desired outcome.

Summary

In conclusion, peripheral administration of LOLA is highly route-dependent. For best brain penetration an intravenous route is most likely optimal, with a chronic infusion (hours) needed. Once inside the blood brain barrier, L-aspartate stimulates brain metabolism in a dose-dependent manner, while L-ornithine, particularly at 100 µmol/L, is strongly sedative, most likely through conversion to GABA via the ornithine aminotransferase pathway. The metabolic outcomes of LOLA administration in brain are therefore highly dependent on the balance between the two constituent amino acids.