Proton magnetic resonance spectroscopy detects cerebral metabolic derangement in a mouse model of brain coenzyme a deficiency

Pantothenate kinase (PanK) is the first and rate-controlling step in the only pathway for coenzyme A (CoA) biosynthesis [1]. CoA is essential for hundreds of metabolic reactions including the tricarboxylic acid (TCA) cycle, fatty acid oxidation and synthesis, amino acid metabolism, and neurotransmitter synthesis [1, 2]. A rare, life-threatening neurological disorder known as pantothenate kinase-associated neurodegeneration (PKAN) arises from mutations in the human PANK2 gene leading to a prominent extrapyramidal movement disorder and a characteristic deposition of iron in the basal ganglia [3]. CoA deficiency is thought to be the cause of the movement disorder and neurodegeneration in PKAN [4], but probes are not available to monitor CoA levels in the brain. Figure 1a depicts the effects of loss of CoA synthesis due to disruption of murine Pank1 and Pank2 genes in neurons. Pyruvate produced from glycolysis requires CoA to form acetyl-CoA, a substrate for the mitochondrial TCA cycle, and CoA limitation can disrupt TCA cycling [5] which, in turn, affects glutamate metabolism thereby causing cell death [6, 7] (Fig. 1a). Recent reviews discuss the various molecular consequences of CoA insufficiency in preclinical PKAN models [8, 9]. There are no disease-modifying treatments for PKAN, and therapeutics are desperately needed. A new potential PKAN therapeutic approach uses Pantazine small molecules that activate the PanK isoforms and thereby stimulate CoA production [10]. Pantazine PZ-2891 was previously reported to restore cerebral CoA levels in the SynCre⁺ Pank1,2 neuronal dKO mice [10]. In this study, we used proton magnetic resonance spectroscopy (¹H MRS) to evaluate the brain metabolic derangements that are associated with neuronal CoA deficiency in this mouse model, with and without treatment with a newly developed Pantazine, BBP-671. ¹H MRS is particularly advantageous for this study because it permits the noninvasive examination of the most abundant cerebral metabolites in vivo [11]. This Pantazine allosterically activates the alternate PanK3 isoform that is expressed in murine neurons and a proposed mechanism for the neurometabolic effect of CoA restoration by BBP-671 is shown in Fig. 1b.

¹H MRS is a powerful and non-invasive technique which can provide information about neurometabolites in vivo [11, 12]. ¹H MRS can identify biomarkers and be applied to evaluate clinical neurodegeneration and therapeutic strategies [13, 14]. In the limited clinical studies demonstrating the application of ¹H MRS in studying PKAN, a reduced N-acetyl aspartate (NAA) level was commonly observed [15‐18]. This is the first report using ¹H MRS for monitoring PKAN therapeutics in a mouse model of the disease. In this study, we treated SynCre⁺Pank1,2 neuronal dKO mice with BBP-671 and used ¹H MRS to identify the prominent neurochemical metabolites and evaluate treatment response (Fig. 2). We found that the cerebral metabolite levels, including glutamate + glutamine (Glx), NAA and lactate, were altered in the mid-brain of the animal model. Based on our observations, we show that Glx levels may be an indicator of restored CoA levels in the response to BBP-671 treatment.

The SynCre⁺Pank1,2 neuronal dKO mice show clear physiological symptoms of CoA deficiency such as reduced growth rate and impaired locomotor activity [10]. Figure 3 represents the ¹H MRS spectra acquired from the midbrain of representative mice in each group. There is a clear reduction in the Glx/tCr, NAA/tCr and lactate/tCr ratios in the midbrains of the SynCre⁺Pank1,2 neuronal dKO mice compared to WT (Wildtype). Glx is the summed group of Glu and Gln, where Glu is the most abundant excitatory neurotransmitter in brain and Gln is the main precursor for Glu [13, 24]. Reduction of Glx points toward impaired excitatory neuronal metabolism in the dKO mice and Glx fully recovered following treatment with BBP-671 (Fig. 4, Additional file 1: Table S1). Reduction of NAA indicated loss of neuronal integrity and function, and the NAA/tCr ratio trended toward improvement but was not significant following BBP-671 treatment. Reduction of NAA in the globus pallidus is consistently reported in patients with PKAN [15‐18]. The lactate/tCr ratio was lower in the dKO mice representing the net balance between lactate production and consumption [13]. Reduced cerebral glycolysis and/or enhanced lactate oxidation to pyruvate to maintain the redox state in the neurons is suggested from the data. The lactate/tCr ratio trended toward improvement with treatment but without statistical significance. The metabolic recovery following therapy, together with a strong signal strength in the ¹H MRS points to Glx as a promising biomarker. NAA and lactate are also candidate markers but require further study. While no changes were observed, additional metabolites—inositol, choline, and taurine—are shown in Additional file 1: Fig. S1.

The treatment response to BBP-671 was associated with recovery of cerebral Glx. BBP-671 treatment activates the PANK3 isoform to increase CoA production that, in turn, has a direct role on TCA cycle metabolism (Fig. 1) [5] and hence Glx production. In contrast to our findings, a study performed on 3 human PKAN patients reported an increase of Glx in the white matter [25]. However, the clinical study was performed in patients with chronic disease progression and potential involvement of multiple brain cell types while this preclinical study was performed in a mouse model with rapidly progressing disease that is specific to neurons. Although neurons have been identified as a focal target of disease in PKAN [26], the role of glial cells in disease progression, for example, is not understood. A clinical MRS study reported elevated myo-inositol (m-Ins) levels in white matter due to gliosis and glial proliferation in patients with PKAN [16]. However, we did not find any significant changes in the m-Ins levels in the SynCre⁺Pank1,2 neuronal dKO mice.

The SynCre⁺Pank1,2 neuronal dKO mouse model represents only selected molecular effects of PKAN, and the study has limitations. Alternative mouse models were not appropriate for evaluation of cerebral MRS either due to the lack of disease-related movement dysfunction and normal lifespan [27] or early postnatal death [4, 28]. Analysis of human PKAN brains identified gene expression signatures that indicated neurons as a focal target of disease [26] and development of the SynCre⁺Pank1,2 neuronal dKO mouse model resulted in a longer-lived animal with consistent and measurable movement dysfunction [13] that was tractable for MRS. Previous studies reported white matter pathology in the patients of PKAN [29]. The previous ¹H MRS studies in human patients were often performed in white matter [18], which is not easily detected in the mouse model since the mouse brain consists of less than 12% white matter [30], as compared to 43% white matter in humans [31]. Mouse models are recognized as not effective for studying white matter pathologies [32]. In addition, there was no observable iron accumulation in the model, in contrast with the iron accumulation often observed in the basal ganglia of PKAN patients. This study may be limited to detection of metabolic changes in the gray matter in the Pank1,2 neuronal dKO mice.

PKAN is a life-threatening neurological disorder that can be caused by a variety of PANK2 mutations that affect protein expression, enzyme activity and/or stability [33, 34], resulting in loss or reduction of kinase activity [33]. PKAN diagnosis is based on characteristics of the movement dysfunction, exon sequencing and MRI evidence for iron accumulation in the basal ganglia [35] with a characteristic “eye-of-the-tiger” pattern in T₂-weighted images [25, 36]. Our study showed no qualitative differences among the three groups in the T₂-weighted images. There are currently no approved clinical therapies for this genetic disorder [37]. Progress with newly developing PKAN therapeutics has been made by evaluating mouse models with deactivated Pank genes in which brain CoA biosynthesis has been disrupted [9, 10, 38]. One of the current challenges is to reliably identify molecular events associated with PKAN dysfunction in addition to monitoring therapeutic responses [39]. Thus, a technique which can quantitatively, non-invasively, and reproducibly analyze neurometabolism is needed and has potential value.

In summary, we have successfully applied ¹H MRS to investigate neuronal chemical alterations in a mouse model of brain CoA deficiency, thought to be an underlying cause of PKAN. The effects of a potential PKAN therapeutic were evaluated by comparing the cerebral metabolic derangements among three mouse groups, where reductions of important metabolites (Glx, NAA, and Lactate) were observed in the model of brain CoA deficiency and recoveries of the same metabolites were found following treatment with a newly developed Pantazine, BBP-671. The most promising biomarker for this potential PKAN therapeutic was the recovery of the Glx/tCr ratio. This study shows that ¹H MRS can be a powerful tool in evaluating therapeutics for metabolic responses of neurological disorders.