Metabolic dysfunctions in multiple sclerosis: implications as to causation, early detection, and treatment, a case control study

Primary findings in this study are three-fold. Firstly, significant elevations of multiple species of VLCFA-PtdEtn were observed in SPMS patients, relative to age-matched healthy controls. Similar increases were observed in RRMS patients also, but a comparatively lesser number of VLCFA-PtdEtns were involved. However, RRMS patients with longer disease duration (>13 y) had more VLCFA-PtdEtn species elevated than RRMS patients with lesser disease duration (<13 y). Secondly, PlsEtn, especially those containing DHA, were significantly increased in SPMS compared to controls. PlsEtn levels did not differ in RRMS relative to controls, irrespective of the disease duration. Thirdly, anti-inflammatory molecules (GTAs) were significantly reduced in SPMS relative to the controls. In contrast, RRMS patients had significantly elevated levels of these molecules, compared to controls and levels tended to decrease with advancing disease duration.

Our study population included a cohort of RRMS, SPMS, and PPMS patients varying in disease duration and age. The control group consisted of age and gender matched individuals with no known morbidity at the time of blood collection. Therefore, metabolic differences observed in MS clinical entities compared to the controls represent underlying metabolic abnormalities. Although the study is cross sectional, which is a limitation, study subjects represented disease duration varying from 0–34 years in RRMS and from 2–26 years in SPMS, enabling us to correlate the metabolic profiles with the natural history of the disease.

VLCFA are fatty acids having carbon chains longer than 22 carbons. Circulating levels of these fatty acids are elevated in a number of disorders, for example in X-linked adrenoleukodystrophy (X-ALD) [18, 19] and Zellweger spectrum disorders [20, 21]. ATP-binding cassette sub-family D member 1 (adrenoleukodystrophy protein), a very long chain acyl-CoA transporter in the peroxisomal membrane, is deficient in X-ALD [19], whereas in Zellweger spectrum disorders, multiple peroxisomal proteins are deficient [22]. PlsEtn, which are produced solely by the peroxisomes, are characteristically reduced in all these diseases [23, 24] as a result of the defective peroxisomes. However, compared to controls, PlsEtn levels were not deficient either in RRMS or in SPMS patients in our study, implying that peroxisomal deficits, as found in the aforementioned diseases, are not the cause of VLCFA elevation in MS patients in our study.

Long chain fatty acids, such as palmitate (16:0) and oleic (18:1) are β-oxidized in mitochondria whereas VLCFA are β-oxidized in peroxisomes to medium chain fatty acids (MCFA) [25]. MCFA are then transported to mitochondria for further oxidation [25]. The acetyl-CoA units produced as a result of peroxisomal β-oxidation are either utilized for anabolic reactions within the peroxisomes such as for plasmalogen synthesis [24, 17], shuttled out to cytosol for chain elongation of fatty acids (the process of synthesizing VLCFA) [26], or become substrates for other synthetic reactions, for example cholesterol synthesis.

Significance of this peroxisomal-mitochondrial interaction in VLCFA metabolism is elaborated in following studies. If fatty acid trafficking to the mitochondria is blocked by inhibiting Carnitine Palmitoyl Transferase I (CPT I), the oxidation of long chain fatty acids (for example 16:0 or 18:0) in peroxisomes increases in isolated rat hepatocytes [27]. Diversion of polyunsaturated fatty acids to peroxisomes for metabolism has been shown when mitochondrial β-oxidation is inhibited in rats [28]. Furthermore, drug induced inhibition of mitochondrial function in rats enhanced peroxisomal β-oxidation gene expression [29]. However, since peroxisomal metabolism has an anabolic outcome, the acetyl-CoA produced is used for chain elongation of fatty acid [26]. Elongation of fatty acids is a major source of circulating VLCFA [30]. Therefore, elevation of circulating VLCFA is a consequence of mitochondrial β-oxidation deficiency.

We have previously shown that glutamate can dose dependently inhibit β-oxidation of palmitate [9]. Potential mechanisms included reduced processing of acetyl-CoA via the tricarboxylic acid (TCA) cycle due to the inhibition of citrate formation and/or changes in the NADH/NAD⁺ ratio due to efflux of aspartate [31]. Both mechanisms lead to inhibition of β-oxidation [9] increasing the accumulation of acyl-CoA in the cytoplasm. Carnitine is used both for the removal of excess acyl-CoA from mitochondria as well as for fatty acid transport into mitochondria, so the shifting of this equilibrium towards acetyl-carnitine could possibly reduce the availability of carnitine [32] for CPT I-dependent fatty acid transport across outer mitochondrial membrane, further reducing fatty acid oxidation. Cytosolic acetyl-CoA is also used for malonyl-CoA production [33, 34], which provides additional impetus to inhibit mitochondrial fatty acid oxidation [35].

VLCFA are high affinity ligands for Peroxisome Proliferator-Activated Receptor α (PPAR-α) [36]. Increased peroxisomal activity along with increased availability of substrate (acetyl-CoA produced during peroxisomal β-oxidation) will drive PlsEtn synthesis [9], elevating their levels. Although not observed in RRMS patients where VLCFA were not upregulated as much as in SPMS, elevated PlsEtn levels were seen in SPMS patients. Therefore, chronic elevation of certain VLCFA during the RRMS stage as shown in this study can be expected to lead to an eventual increase in PlsEtn in SPMS.

Moreover, DHA containing PtdEtn and PlsEtn were also elevated in SPMS patients, but not in RRMS patients. Conversion of 24:6 to 22:6 by β-oxidation, which is the final step in DHA synthesis, occurs in peroxisomes [37]. DHA synthesis is regulated by PlsEtn, as demonstrated by the reduction of DHA levels in PlsEtn deficient cells [38]. When PlsEtn was stimulated in these cells by adding a precursor, DHA synthesis also increased [38]. Therefore, upregulation of PlsEtn synthesis, along with global enhancement of fatty acid elongation, can be expected to trigger DHA synthesis, elevating its levels as seen in SPMS.

Consequently, all lipid derangements observed in MS patients can be explained by the reduction of fatty acid β-oxidation due to subtle mitochondrial dysfunction (Fig. 3). Evidence supporting this hypothesis come from our previous in vitro studies using HepG2 cells that showed dose dependent reduction of palmitate oxidation by glutamate [9]. These cells also had increased levels of VLCFA and DHA containing PlsEtn [9]. Interestingly, L-carnitine supplementation has been reported to ameliorate fatigue in MS patients [39, 40], lending further support to this hypothesis. Moreover, a patient presenting with a relapsing and remitting demyelinating disorder similar to multiple sclerosis was found to have defects in mitochondrial β-oxidation and accumulated acylcarnitines in blood [41]. The patient improved with the supplementation of L-carnitine and riboflavin in the diet along with a low-fat, high-carbohydrate diet [41]. These clinical examples also provide indirect evidence for a likely reduction of mitochondrial fatty acid β-oxidation in MS.

Altered mitochondrial structure, molecular and biochemical abnormalities, impaired Complex I activity and impaired bioenergetics has been demonstrated in MS patients, as reviewed by Morris and Berk et al. [42]. Their review provide ample evidence that mitochondrial dysfunction is a frequent finding in MS. Mitochondrial dysfunction could be secondary to oxidative damage of this organelle. Oxidized macromolecules have been demonstrated in active MS lesions and are associated with apoptotic oligodendrocytes and areas of neurodegeneration in MS, as reviewed by Haider [43]. Oxidative stress in mitochondria could be induced by reactive oxygen and nitrogen species from activated microglia [13, 43]. Oxidative stress in turn could also induce pro-inflammatory genes, initiating a self-perpetuating cycle [42]. Therefore, an initial inflammatory trigger or incipient mitochondrial dysfunction triggered by environmental factors could initiate this vicious cycle.

VLCFA are known to cause inflammation and demyelination [44]. Chronic elevated levels of VLCFA can induce inflammation in the brain [24, 45]. Indeed, VLCFA infusion has induced lypooxygenases in the brain in experimental animals [46]. Demyelination is a hallmark of X-ALD, wherein deleterious effects of abnormally high VLCFA levels are well established. Sphingolipids that contain VLCFA are a major component of myelin, along with cholesterol and phospholipids [47]. They also exist in outer plasma membrane mostly in lipid rafts, in conjunction with cholesterol [48]. Elevation of VLCFA can therefore be expected to change sphingolipid composition. Perturbations in sphingolipid metabolism are known in MS [49]. Consequently, it can be speculated that there is a link between VLCFA, sphingolipids, and inflammation in MS.

Potential sources of elevated VLCFA are the liver and CNS. The liver is a source of many circulating lipids. As described above, myelin in the CNS is enriched with sphingolipids containing VLCFA. Demyelination can therefore possibly cause a local elevation of VLCFA, although it is unclear whether peripheral levels would be elevated. Also, activated microglia that emerge during demyelination produce copious amounts of glutamate [50] which can induce mitochondrial dysfunction. Mitochondrial dysfunction can lead to VLCFA accumulation as described elsewhere, possibly initiating and perpetuating a vicious cycle.

All three clinical forms of MS in this study had higher PtdEtn 16:0/28:0 than controls (p < 0.05, student’s t-test). However, an increase in the number of elevated VLCFA-PtdEtn species in patients with longer disease duration in RRMS, and an even higher number of elevated species in SPMS, may indicate that deranged VLCFA metabolism is an accumulating deficit in MS. In contrast, the opposite behavior of GTAs in these two patient categories is thought to be due to changes in the anti-inflammatory response. GTAs are novel fatty acids known to have anti-inflammatory effects in vitro [7] and are associated with a protective phenotype in various malignant tumors [24]. Higher GTA levels (p < 0.05 vs. controls) in early RRMS (<13 y disease duration) and low GTA levels (p < 0.05 vs. controls) observed in SPMS is suggestive of a protective response in early relapsing remitting phase, that eventually diminish in the progressive phase. Although unconfirmed, reduction in GTAs in the latter stages of RRMS could be ascribed to pathophysiological inability to sustain this protective response.

The reversal of the ratio of PtdEtn 16:0/28:0 to GTA 465/403 at about 13 years of disease duration in the RRMS group (data not shown) indicates that RRMS is comprised of two metabolically distinct sub-groups with the disease duration as the key discriminator between the subgroups (<13 y and > = 13 y). This is further confirmed by the differences in the total number of elevated VLCFA-PtdEtn and by significant differences in the levels of PtdEtn 16:0/28:0 and in GTA 465/403, observed between these two sub-groups. Although highly speculative without a longitudinal study, the presence  of metabolically distinct subpopulations in RRMS, with the subgroup with a disease duration > = 13 y being closer to SPMS (in terms of sum of elevated VLCFA-PtdEtn, PtdEtn 16:0/28:0 and GTA 465/403), indicate that transition from RRMS to SPMS can be monitored using these metabolic changes.

Longitudinal studies will be needed to overcome the limitations posed by cross-sectional nature of the RRMS samples. Smaller sample sizes in SPMS and PPMS patient groups limits the generalization of the results. Larger and longitudinal studies are planned to overcome these limitations.