GATM Mutations Cause a Dominant Fibrillar Conformational Disease in Mitochondria—When Eternity Kills

More than 60 years ago, Luder and Sheldon¹ and then, other nephrologists reported on a rare form of autosomal dominant renal Fanconi syndrome (RFS) that appeared in childhood and slowly evolved toward kidney insufficiency. These clinical features were thus quite distinct from nephropathic cystinosis, the most frequent cause of hereditary (recessive) RFS.² Over the years, additional families were described, and the affected locus was identified; however, the causative gene and disease mechanism remained elusive. In this issue of the Journal of the American Society of Nephrology, the paper by Reichold et al.³ finally defines the mutated gene as GATM with identification of another family with de novo mutation, explains the dominant mode of transmission by an attractive multimerization-based fibrillar molecular model, and further proposes a rational dietary treatment that remains to be tested.

GATM encodes for a nuclear-encoded mitochondrial matrix dimeric enzyme, glycine amidinotransferase (hence GATM; also known as l-arginine:glycine amidinotransferase), catalyzing the penultimate reaction of the creatine biosynthesis pathway. Fortunately, much is known about GATM, including information from crystallographic studies. All patients with GATM disease exhibited a single heterozygous missense mutation at evolutionary conserved proline or threonine residues clustered in a small central stretch representing <5% of the protein. As predicted from dominant transmission, disease-causing GATM mutations did not abolish enzymatic activity. Furthermore, GATM haploinsufficient mice had no significant kidney phenotype, opening the possibility of suppressive intervention.

In kidneys, GATM expression is restricted to proximal tubular cells (PTCs), which house one of the largest mitochondrial compartments in the human body to support the high energy demand for their titanic transport activity. This explained the specific PTC damage related to GATM disease (thus RFS). Remarkably, GATM immunogold labeling of a patient kidney biopsy decorated fibrillary aggregates in mitochondrial PTCs. To address the mechanism of fibril formation by a cell biologic approach, the authors generated stable transfectants in proximal tubular LLC-PK1 cells for tetracyclin-inducible overexpression of wild-type human GATM and each of its four known point mutants. Wild-type GATM overexpression had no effect on mitochondria, but all mutants caused a dramatic structural mitochondrial deformation, resembling sickle erythrocytes. Such mitochondria reproduced the immunogold pattern of the patient biopsy. Transfectants elegantly served to show irreversibility of the deformation by aggregates on removal of the tetracyclin inducer, much beyond the normal mitochondrial turnover time. Longitudinal fibrillary aggregation was found to prevent mitochondrial dynamics (fission), trigger reactive oxygen species (ROS) production, and activate the NLRP3 inflammasome.

The paper culminates in proposing an explanation for the propensity for aggregation by molecular dynamic modeling. Normal GATM acts as a homodimer (A:A). All mutations were predicted in silico to generate an additional pathogenic β-sheet–interacting interface (A/A) supporting alternate bead-like elongation, thus converting normal dimers into longitudinal multimers (A:A/A:A/A...; i.e., fibrillary aggregates). Incidentally, it is remarkable that, among the 1500 predicted mitochondrial proteins, GATM mutations are so far unique in this behavior. Abnormal β-sheet formation is well known to favor bead-like fibrillar polymerization in α1-antitrypsin affected by homozygous Z mutation (PIZZ) to cause liver cirrhosis and a variety of brain neurodegenerative diseases causing dementia.⁴ In PIZZ hepatocytes, bead-like polymers and entangled polymeric aggregates accrue slowly over months/years. Interestingly, unequal exposure to inflammatory episodes explains in part the variable age of cirrhosis onset due to changes in the rate of PIZZ synthesis in the acute-phase response. This indicates that fibrillary disease evolution may be influenced by intervention on monomer synthesis. Dominant mutations of the microtubule-associated τ-protein induce neurofibrillary lesions, which cause frontotemporal dementia and are identical to those prevalent in acquired Alzheimer disease. Most recently, atomic resolution of τ-fibrils, achieved by cryoelectron microscopy, revealed a typical cross–β-structure.⁵ Thus, the predicted propensity to aggregation being dominant readily explains the autosomal dominant transmission and the slow progression of GATM disease; it places this disease in the family of fibrillar conformational diseases and suggests the possibility of beneficial intervention by decreasing synthesis of the specific protein culprit. Validation of molecular modeling of GATM fibrils by the recently available high-resolution cryoelectron microscopy might be considered.

How do fibrils escape clearance to cause disease? Protein aggregation is a central issue in biology and disease. Appropriate formation of intra- and extracellular fibrils has been remarkably tamed by evolution (e.g., to generate cystoskeletal and collagen fibers). In the cytosol, labile actin-based microfilaments and tubulin-based microtubules undergo permanent remodeling by assembly at the “plus” end and depolymerization at the “minus” end (unless stabilized by drugs, such as taxol). Also, in the secretory apparatus, premature intracellular collagen fibril formation is prevented by the addition of transient, extracellularly cleavable globular termini at the two ends of the newly synthesized triple helix. Unwanted aggregation of cytosolic proteins into irreversible fibrils is normally prevented by chaperone proteins, such as heat-shock proteins, and aggregates are normally cleared by the proteasome, selective macroautophagy, or chaperone-mediated autophagy. In the endoplasmic reticulum of PIZZ hepatocytes, fibrils preserve the folding of individual monomers, which prevents recognition by chaperones and thus, evades proteolytic clearance.⁴ Nuclear-encoded mitochondrial matrix proteins are first assembled by free cytosolic ribosomes and remain unfolded thanks to specific cytosolic chaperones until they cross the double mitochondrial membrane by unidirectional transport systems; only after that do they fold (and can dimerize). As in their bacterial ancestors, mitochondria contain folding chaperones (e.g., mtHSP70) and proteases competent to clear misfolded (e.g., aged) proteins.⁶ Alternatively, misfolded ubiquitinated mitochondrial matrix proteins can be retrotranslocated into the cytosol for degradation by the proteasome.⁷ It can, however, be suspected that, like in PIZZ proteins, preservation of mutated GATM monomer structure in bead-like polymers prevents their recognition by chaperone and proteolytic clearance, allowing for fibril extension and aggregation. Furthermore, large mitochondrial fibrils should obviously not be eligible for retrotranslocation into the cytosol. As a final quality control, mitochondria constantly rejuvenate by fission/fusion,⁸ and defective mitochondria are disposed of by mitophagy. However, when GATM mutant mitochondria are rigidified by transversal fibril bundles, fission is prevented. By analogy to “frustrated phagocytosis,” which occurs when extracellular preys are too large to be engulfed,⁹ huge mitochondria deformed by longitudinal GATM fibers could predictably no longer be wrapped by the autophagosome membrane, causing “frustrated autophagy.” Consequences easily follow. Healthy mitochondria keep ROS production under control by their complex electron transfer system. Conversely, aging mitochondria release abundant ROS, which trigger the inflammasome as documented here.

What are new perspectives? The paper by Reichold et al.³ stresses how important in-depth genetic analysis of unexplained causes of RFS is, including identification of de novo mutations. This will allow, in particular, a better estimate of GATM disease frequency, which is especially justified if effective treatment can indeed be proposed. As is the case for other human enzymes in biosynthetic pathways downregulated by the end product, this paper further shows that creatine supplementation to wild-type mice at equivalent acceptable doses for humans significantly decreases GATM expression and thus, might slow down disease progression (analogous to the control of inflammatory status in PIZZ patients). Whether creatine would prevent fibril formation in the transgenic LLC-PK1 cells could not be tested, because the endogenous promotor was swapped with the tetracycline-inducible promotor. A knock-in animal model (e.g., by CRISPR/Cas9 technology) is eagerly awaited to not only further investigate pathogeny in vivo (in particular, natural course and disease adaptations that might open unexpected avenues)¹⁰ but also, test effectiveness of creatine supplementation.

Disclosures

None.