Abstract
Acute intermittent porphyria (AIP), an autosomal dominant metabolic disease (MIM #176000), is due to a deficiency of hydroxymethylbilane synthase (HMBS), which catalyzes the third step of the heme biosynthetic pathway. The clinical expression of the disease is mainly neurological, involving the autonomous, central and peripheral nervous systems. We explored mitochondrial oxidative phosphorylation (OXPHOS) in the brain and skeletal muscle of the Hmbs−/− mouse model first in the basal state (BS), and then after induction of the disease with phenobarbital and treatment with heme arginate (HA). The modification of the respiratory parameters, determined in mice in the BS, reflected a spontaneous metabolic energetic adaptation to HMBS deficiency. Phenobarbital induced a sharp alteration of the oxidative metabolism with a significant decrease of ATP production in skeletal muscle that was restored by treatment with HA. This OXPHOS defect was due to deficiencies in complexes I and II in the skeletal muscle whereas all four respiratory chain complexes were affected in the brain. To date, the pathogenesis of AIP has been mainly attributed to the neurotoxicity of aminolevulinic acid and heme deficiency. Our results show that mitochondrial energetic failure also plays an important role in the expression of the disease.
Introduction
Acute intermittent porphyria (AIP) is characterized by episodic neurovisceral attacks involving severe abdominal pain, peripheral neuropathy and psychiatric disturbances (1,2). Diffuse abdominal pain, mimicking the acute abdomen syndrome, is accompanied by other neurovisceral symptoms such as nausea, vomiting, constipation or diarrhoea. These symptoms, together with tachycardia, cardiac arrhythmia and high blood pressure, are thought to be due to the dysfunction of the sympathetic autonomic nervous system (3). The central nervous system is also involved in AIP. Porphyric attacks are often preceded by a prodromic phase with behavioural changes such as anxiety, restlessness and insomnia (4). In 20–30% of patients, signs of mental disturbance such as anxiety, insomnia, depression, disorientation, hallucinations, paranoia, agitation or confusion have been reported (5). Encephalopathy may also develop with altered consciousness and cortical lesions detected by brain Magnetic Resonance imaging (MRI) (6). Acute attacks can be life-threatening because of complications such as paralysis of the respiratory muscles. Indeed, unrecognized or untreated AIP is associated with a significant mortality of up to 10% (7). Muscle pain, weakness and paresis, beginning proximally in the arms and legs, also often develop during the crisis (5,8) as a consequence of peripheral neuropathy (2,3). Electromyography and muscle biopsies show features of denervation together with muscle atrophy (9,10). Pathological studies have shown widespread peripheral nerve degeneration and chromatolysis of the anterior horn cells (10). These neurophysiological features may resolve after the crisis, but the partial improvement and persistence of irreversible nerve damage and muscle atrophy may accompany the evolution of the disease especially after recurrent AIP attacks (11). Subtle anomalies of nerve conduction have also been identified in patients not in the acute phase of the disease, reflecting subclinical neuropathy between acute attacks (12).
The causes of this constellation of neurological features in AIP are far from being understood. AIP, an autosomal dominant metabolic disease (MIM #176000), is caused by the deficiency of hydroxymethylbilane synthase (HMBS, EC 2.5.1.61), which catalyzes the third step of heme biosynthetic pathway (13). Acute attacks are provoked by endocrine or environmental factors including malnutrition (14) and drugs such as phenobarbital (PB) that induce the gene expression of cytochrome P450 enzymes in the liver (15). AIP patients are unable to respond satisfactorily to the increased demand for heme. In addition, the relative lack of heme greatly activates hepatic aminolevulinate synthase-1 (ALAS-1), the first step of the heme biosynthetic pathway that catalyzes the synthesis of aminolevulinic acid (ALA). The increased hepatic ALAS-1 activity leads to the sharp accumulation of ALA and porphobilinogen (PBG), the heme precursors synthetized upstream of the HMBS blockage. Thus, during the AIP crisis, the liver is the main source of overproduction of ALA. The pre-eminence of the liver in the pathogenesis of AIP is supported by the reported remission of the disease after liver transplantation (16,17). The current treatment of AIP attacks associates the withdrawal of porphyrinogenic drugs, adequate intake of calories and carbohydrates and intravenous administration of heme arginate (HA) (Normosang®) that restores the negative heme feedback on ALAS-1.
The diffusion of ALA from the liver to the rest of the body is thought to be the main factor in the neurological expression of AIP (18,19). The mechanism of ALA neurotoxicity may be linked to its pro-oxidant properties or to its structural homology with GABA (20–23). However, ALA neurotoxicity is probably not the unique cause underlying the neurological expression of AIP (2). First, no close correlation was found between ALA levels during an AIP attack and the clinical expression of the disease (24). Second, administration of ALA to healthy individuals did not produce symptoms of porphyria (25,26). Third, even in the absence of PB induction, Hmbs−/− mice spontaneously suffer from progressive motor dysfunction and axonal neuropathy due to axonal degeneration without any significant increase of ALA (27,28). The cause of this neurotoxicity in the absence of accumulation of ALA is unknown but the lack of heme might in itself contribute to the pathogenesis of the disorder (23).
The Hmbs−/− mouse model was generated in 1996 (27). These C57BL6 mice are compound heterozygotes of one Hmbs null allele and a functional Hmbs allele with a milder mutation. Hmbs−/− mice have 30% residual HMBS activity in the liver and present the typical biochemical characteristics of human AIP with increased ALAS-1 activity and massive urinary excretion of ALA and PBG after intraperitoneal administration of PB. We recently reported a failure of mitochondrial energetic metabolism in the liver of the Hmbs−/− mouse model, with the first demonstration of a direct pathogenic mechanism linking the alteration of heme synthesis with the tricarboxylic acid cycle and impairment of the respiratory chain (29). Since neurological and muscular symptoms are the major consequences of mitochondrial defects, we hypothesized that such defects might be involved in the pleiotropic neurological dysfunctions observed in AIP. We therefore investigated mitochondrial energetic metabolism in the brain and the skeletal muscle of the Hmbs−/− mouse model of AIP.
Results
Enzymatic activities of complexes I and II are altered in the muscle of Hmbs−/− mice treated with PB
We verified that PB alone did not induce any enzymatic defects in the gastrocnemius of WT mice. WT mice injected with phenobarbital (WT-PB) showed no significant impairment of the mitochondrial respiratory chain activity (Fig. 1A). Similarly, PB did not alter citrate synthase (CS) activity, making it possible to normalize the mitochondrial respiratory complex activity with respect to CS activity.
Enzymatic activities of the four respiratory chain complexes in the gastrocnemius. (A) WT mice (Hmbs+/+) with (WT-PB) or without (WT) treatment with phenobarbital. (B) WT: wild-type mice; BS: Basal state corresponding to Hmbs−/− mice without treatment; PB: Hmbs−/− mice treated with phenobarbital; PB-HA: Hmbs−/− mice treated with phenobarbital and heme arginate. Results normalized with respect to CS activity. Statistical significance: *P < 0.05.
Measurement of the enzymatic respiratory complex activity (Fig. 1B) showed a spontaneous increase of complex I activity in the Hmbs−/− mice in the basal state (BS) in comparison to WT mice (+ 86%, P < 0.05). Exposure to PB led to a significant decrease in the activity of complexes I and II, with a reduction of 40% (P < 0.05) for complex I and 30% (P < 0.05) for complex II. These activities were not significantly restored by HA treatment. The enzymatic activity of the heminic complexes III and IV was not significantly affected.
Respiration driven by complexes I and II is altered in the muscle of Hmbs−/− mice treated with PB
The ADP-stimulated respiration driven by complex I was spontaneously increased in the gastrocnemius of Hmbs−/− mice in the BS compared with that in WT mice (+60%, P < 0.05) (Fig. 2, EIII MP). The respiration driven by complexes I and II was significantly affected in the gastrocnemius by the administration of PB in Hmbs−/− mice compared with that in BS mice. The respiration driven by complex I was reduced by 64% (P < 0.01) (Fig. 2, EIII MP) in the phosphorylating condition. This respiratory defect was not significantly restored by HA treatment. The respiration driven by complex II was significantly altered by exposure to PB with a reduction of 68% (P < 0.01) in the phosphorylating condition with partial compensation after treatment with HA (+100%, P < 0.05) (Fig. 2, EIII SR).
Respiration in the gastrocnemius of the Hmbs−/− mouse. BS: Basal state corresponding to Hmbs−/− mice without treatment; PB: Phenobarbital; PB-HA: Phenobarbital + heme arginate. MP: malate-pyruvate substrates (complex I-driven respiration). SR: succinate-rotenone (complex II-driven respiration). EII: non-stimulated respiration (non-phosphorylating); EIII: ADP-stimulated respiration (phosphorylating). Statistical significance: *P < 0.05 and **P < 0.01.
ATP production is decreased in the muscle of Hmbs−/− mice treated with PB
ATP synthesis in the gastrocnemius of Hmbs−/− mice was significantly affected by the administration of PB. Phenobarbital induced a 48% decrease (P < 0.01) in ATP production that was completely restored after treatment with HA (+100%, P < 0.05) (Fig. 3). The ATP/O ratio was maintained in this muscle with a similar efficacy of ATP synthesis.
ATP synthesis in the gastrocnemius of the Hmbs−/− mouse. BS: Basal state corresponding to Hmbs−/− mice without treatment; PB: Phenobarbital; PB-HA: Phenobarbital + heme arginate. Statistical significance: *P < 0.05 and **P < 0.01.
Enzymatic activities of the four respiratory chain complexes are altered in the brain of Hmbs−/− mice treated with PB
We verified that PB alone did not induce any enzymatic defect in the brain of WT mice. WT mice injected with phenobarbital showed no significant impairment of respiratory complex activity (Fig. 4A). Similarly, PB did not alter CS activity, making it possible to normalize respiratory complex activity with respect to CS activity. Complex I activity was significantly reduced in the brain of Hmbs−/− mice after exposure to PB compared with BS mice (−45%, P < 0.05) (Fig. 4B). Complex I activity was restored by treatment with HA (+55%, P < 0.05). Complex II activity was spontaneously increased in the brain of Hmbs−/− mice in the BS compared with WT mice (+40%, P < 0.05), and the administration of PB induced a significant decrease of 30% (P < 0.05) that was not restored by HA. Complex III activity was significantly reduced by PB (−35%, P < 0.05) and was restored by HA (+86%, P < 0.05). Complex IV activity was affected by PB with a significant decrease of 30% (P < 0.05) that was not restored by HA. Respiration driven by complexes I and II (ADP-stimulated or not) were not significantly affected in the Hmbs−/− mice in the BS or after exposure to PB (Fig. 5).
Enzymatic activities of the four respiratory chain complexes in the brain of the Hmbs−/− mouse. (A) WT mice (Hmbs+/+) with (WT-PB) or without (WT) treatment with PB. (B) WT: wild-type mice; BS: Basic state corresponding to Hmbs−/− mice without treatment; PB: Phenobarbital; PB-HA: Phenobarbital + Heme arginate. Results normalized with respect to CS activity. Statistical significance: *P < 0.05 and **P < 0.01.
Respiration in the brain of the Hmbs−/− mouse. BS: Basal state corresponding to Hmbs−/− mice without treatment; PB: Phenobarbital; PB-HA: Phenobarbital + heme arginate. MP: Malate-pyruvate substrates (complex I-driven respiration); SR: succinate-rotenone (complex II-driven respiration). EII: non-stimulated respiration; EIII: ADP-stimulated respiration.
Proposed pathophysiological model of AIP. Phenobarbital administration sharply increases heme demand in the liver for the synthesis of cytosolic cytochrome-containing detoxification enzymes. In Hmbs−/− mice, loss of the negative control exerted by heme increases ALAS1 activity, leading to the accumulation of heme precursors such as ALA. ALA diffuses in the body, inducing an inhibitory effect on the respiratory chain. An AIP crisis alters the activity of the non-heminic complexes I and II, as well as the respiration driven by these complexes in the muscle, leading to an impairment of ATP production. In the brain, the enzymatic activities of the four respiratory chain complexes are altered whereas the respiratory rates are conserved. The nervous system and skeletal muscle, having the highest energy demands in the organism, are particularly vulnerable to bioenergetic defects leading, at least in part, to the neuromuscular expression of AIP. This pathophysiological cascade can be reversed by HA, which restores the negative retrocontrol on ALAS1 in the liver.
Discussion
Alterations of mitochondrial energy production are involved in many inherited diseases and common disorders. Although the clinical expression of inherited mitochondrial disorders is highly heterogeneous involving most of the organs, neurological and muscular impairments are most frequently observed. The sensitivity of the nervous system and muscle to mitochondrial dysfunction is related to their particularly high energy demand. In cases of inherited mitochondrial dysfunction, the neuromuscular expression of the disorder is highly heterogeneous and pleiotropic, affecting the central, peripheral and autonomous nervous systems. The neuromuscular alterations observed in AIP together with the close relationship between the biosynthesis of heme and mitochondrial energetic metabolism led us to investigate the possible contribution of a mitochondrial energetic defect to the clinical expression of AIP.
We found that Hmbs−/− mice have a bioenergetic phenotype distinct from that of WT mice with the same nuclear background. The mitochondrial energetic metabolism of Hmbs−/− mice was found to be spontaneously modified in the brain and muscle tissues in comparison to controls. The respiration driven by complex I and the enzymatic activity of complex I were spontaneously higher in the muscle of the Hmbs−/− mice, whereas complex II activity alone was spontaneously higher in the brain than in controls. In a previous study on the spontaneous adaptation of complexes I–III in the liver of Hmbs−/− mice, we used a cocktail of antibodies against complexes I–IV to show that the alteration of the mitochondrial respiratory activity was not due to a modification of protein expression (29). These modifications, probably due to the action of adaptive mechanisms, reflect the establishment of a re-equilibrium of the mitochondrial energetic metabolism with regard to HMBS deficiency. The spontaneously modified activities of the respiratory complexes in the brain and the muscles may result from the allosteric regulation occurring in response to the imbalance between the three interdependent mitochondrial metabolite functions, i.e. heme biosynthesis, the Krebs cycle and the respiratory chain.
When the AIP phenotype was exacerbated by PB induction, the mitochondrial energy metabolism was severely affected in the brain as well as in the muscle tissues. In the brain, the compensatory increase of complex II activity observed in Hmbs−/− mice disappeared after administration of PB, and the enzymatic activities of complexes I–IV were significantly altered without affecting the respiration. In the muscle, the increase in complex I activity was reversed after administration of PB, with a reduction of respiration and a decrease in the enzymatic activities of complexes I and II, finally leading to the collapse of ATP production. In addition, we found that treatment with HA was highly effective in restoring the capacity of ATP production in the muscle and rectifying several of the PB -induced OXPHOS alterations.
We have reported deficiencies of complexes I–III in the liver of the Hmbs−/− mouse model (29). This OXPHOS impairment was thought to be produced by a mechanism of tricarboxylic acid cycle cataplerosis due to the hyperactivation of ALAS-1, leading to a decreased supply of reduced cofactors to the mitochondrial respiratory chain. Since the hyperactivation of ALAS-1 during the AIP crises is mainly liver-specific, a similar mechanism cannot be evoked in the case of brain and muscle tissues. The neurotoxicity of ALA, overproduced by the liver, remains the most plausible hypothesis for explaining the AIP-associated OXPHOS defect in the brain and muscle. Working with HepG2 cells, we recently reported that, in addition to inducing an oxidative stress, ALA also directly impaired mitochondrial respiration and decreased ATP content (30). We cannot exclude a direct effect of heme deprivation on respiratory chain in the brain where all the four complexes are affected. However, in the muscle, only the non-heminic complexes are affected. In addition, HA is not known either to release free heme or to be incorporated in hemoproteins. OXPHOS impairment may also be related to the tight interdependency of the mitochondrial metabolic pathways. For instance, the heme synthesis begins in the mitochondria by the condensation of glycine, with succinyl-CoA being supplied by the tricarboxylic acid cycle, which itself is closely linked to the respiratory chain (31).
Although OXPHOS is severely impacted by the exacerbation of the AIP phenotype by PB in the liver, the brain and the muscle of Hmbs−/− mice, the respiratory dysfunction observed is not identical in the three tissues. The differences are probably due to the tissue diversity of the mitochondrial and OXPHOS proteomes. Indeed, approximately half of the mitochondrial proteins have a tissue-specific expression (32). Interestingly, it was shown that the mitochondrial proteome was sharply altered by heme depletion in cultured mouse hepatocyte (33). Such heme-depletion elicited down-regulation of the mitochondrial proteome specifically included several ATP synthase subunits. A decrease in ATP synthase expression could worsen the ATP synthesis deficiency that was observed upon PB -mediated increased heme demand. The various tissues consume different physiological energy substrates and, at the functional level, there is a great diversity in the kinetic parameters of the four respiratory complexes (34). These differences may also reflect the presence of the brain–blood barrier that considerably limits the delivery of ALA and HA to the brain.
Our results suggest that the level of OXPHOS alteration is compatible with the neurological and muscular clinical expression of AIP (Fig. 6). Such mitochondrial energy deficits incurred for the purposes of concurrent ALA generation could synergize its neurotoxicity and contribute to the clinically observed neurovisceral manifestations of AIP attacks. The muscular clinical expression of the disease, mainly muscle weakness, is usually attributed to peripheral neuropathy and muscle denervation that have been evidenced in human patients and the mouse model. Our results strongly argue for the participation of primary muscular dysfunction due to bioenergetic failure. The implication of the mitochondrial respiratory dysfunction in AIP highlights the importance, already well known, of providing patients with appropriate energetic and carbohydrate supplies for the management of acute attacks of the disease (14). Moreover, the prospect of a subcutaneous RNA interference (RNAi) treatment targeting ALAS-1 underscores the necessity for a better characterization of the relationship between ALAS-1, the tricarboxylic acid cycle, the mitochondrial respiratory complex and OXPHOS in extra-hepatic tissues (35). The investigation of the impaired mitochondrial energetic metabolism of muscle and brain in the AIP mouse model, throwing new light on the pathophysiology of the disease, constitutes a prerequisite step to the introduction of novel ALAS-1 biotherapy in AIP. The further knowledge of the mitochondrial involvement in AIP may be expected to lead to a better understanding of the factors influencing the penetrance, expressivity and treatment of the disease.
Materials and Methods
Animals
The murine Hmbs−/− model of AIP used in this study was generated by Lindberg et al. (27,28). These mice are compound heterozygotes of two different disruptions: one null allele of the Hmbs gene from the T2 strain C57BL/6-Hmbstm2(neo)Uam, and a functional allele with a milder mutation from the T1 strain C57BL/6-Hmbstm1(neo)Uam.
The specific impact of PB on mitochondria was assessed by comparing the enzymatic activities of the mitochondrial respiratory complexes I–IV and of CS in the gastrocnemius and the brain tissues of five WT mice subjected to intraperitoneal administration of the drug with five untreated control WT mice, under the same conditions as imposed on the Hmbs−/− mice. The effect of heme arginate on respiratory chain was also tested by complexes I–IV enzymatic activities measurements in isolated brain mitochondria. We did not observe any significant impact of heme arginate on respiratory chain complexes (data not shown).
Twenty-four female mice (8–10 weeks old) were allocated into four groups, each comprising six mice. The first group (WT), consisting of six wild-type Hmbs+/+ mice with the same C57BL/6 genetic background, served as controls. The second group (BS), consisted of six Hmbs−/− mice in the BS. The third group (PB) consisted of six Hmbs−/− mice subjected to intraperitoneal administration of PB (Gardenal®) at 100 mg/Kg for three consecutive days as described elsewhere (36). These mice were sacrificed 15 h after the last injection. The fourth group (PB-HA) consisted of six Hmbs−/− mice exposed to PB and treated by an intraperitoneal injection of HA (Normosang®) at 8 mg/kg for two consecutive days. These mice were sacrificed 2 h after the last injection. The metabolic profiles of urinary ALA and PBG, analysed in the four groups of mice confirmed the specific and reversible accumulation of ALA and PBG during exposure to PB (29).
All the animal experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals, 8th Edition, 2011, National Research Council, National Academies Press, Washington, DC, USA. The mice were anaesthetized with isoflurane, in accordance with European Community guidelines (directive 86/609/CEE), before decapitation and dissection. The whole brain was dissected immediately before homogenization and preparation of brain mitochondria for oxygraphy and enzymology. The gastrocnemius was dissected immediately at the level of its white section before dissection and permeabilization of the muscle fibres for oxygraphy, and homogenization of muscle mitochondria for enzymology.
Mitochondrial enzymatic activities
The activities of the mitochondrial respiratory complexes I–IV and CS) were measured on the gastrocnemius and brain homogenates at 37°C with a Beckman DU-640B spectrophotometer (Beckman Coulter) using standard methods (37).
Mitochondrial respiratory rates
The mitochondrial respiration rates were measured in the gastrocnemius muscle fibres (38) and in brain (39), using a two-channel high-resolution oxygraph respirometer (Oroboros, Innsbruck, Austria). State II respiration was initiated in the presence of either complex I substrates (5 mm malate and 2.5 mm pyruvate) or complex II substrates (10 mm succinate supplemented with 10 µm rotenone in order to inhibit complex I). The active respiratory states (state III, coupled-respiration) were initiated by saturating the ADP concentration to 0.5 mm in isolated brain mitochondria or to 1.5 mm in the muscle fibres. Cytochrome c (8 μm) was added to check the integrity of the mitochondrial outer membrane (result not shown).
ATP synthesis
The ATP synthesized in situ in permeabilized muscle fibres was measured as described (40). Aliquots were sampled at 30-sintervals, quenched with equal volumes of 7% perchloric acid and preserved at 80°C. The ATP content was measured from neutralized supernatants by an assay based on a tandem enzyme reaction driven by hexokinase (0.9 U/ml) and glucose-6-phosphate dehydrogenase (0.1 U/ml), which, in the presence of ATP and glucose (1 mm), converts NADP (0.5 mm) to NADPH at an equimolar ratio (41).
Statistical analysis
Statistical comparisons were made with the Mann–Whitney U-test. Differences were considered statistically significant at P < 0.05. All analyses were performed using GraphPad Prism 5 software (GraphPad Software, La Jolla, CA, USA). All results are expressed as mean ± SEM.
Funding
This work was supported by grants from the following patients' foundations: ‘Association contre les Maladies Mitochondriales’, ‘Ouvrir les Yeux’, ‘Retina France’, ‘Union Nationale des Aveugles et Déficients Visuels’ and from the Laboratory of Excellence GR-Ex, reference ANR-11. LABX-0051 is funded by the program ‘Investissements d'avenir’ of the French National Research Agency, reference ANR-11-IDEX-0005-02.
Acknowledgement
We are grateful to Kanaya Malkani for critical reading and comments on the manuscript.
Conflict of Interest statement. None declared.
References
Author notes
Authors contributed equally to this work
