Abstract
Pyridoxine-dependent epilepsy (PDE), a rare autosomal recessively inherited metabolic disease, results from mutations in ALDH7A1, a gene crucial for lysine metabolism. Although early high-dose pyridoxine treatment can control seizures, ∼75% of PDE patients still have intellectual disabilities. In this study, we test the hypothesis of substrate reduction therapy for PDE by genetically perturbing lysine α-ketoglutarate reductase (LKR), an enzyme upstream of the defective ALDH7A1, in male and female laboratory mice. A homozygous mutation in LKR completely abolishes the accumulation of toxic lysine catabolism intermediates (α-aminoadipic-δ-semialdehyde and its cyclic form, δ-1-piperideine-6-carboxylate), ends the epileptic state, and restores the defective brain development and cognitive impairments in ALDH7A1-deficient mice. Therefore, these genetic data prove the concept of the effectiveness of substrate reduction therapy for PDE via LKR inhibition.
Significance Statement
ALDH7A1 encodes an aldehyde dehydrogenase also known as antiquitin, which functions in lysine and hydroxylysine degradation. Mutations in ALDH7A1 cause pyridoxine-dependent epilepsy (PDE), an autosomal recessive inborn error of metabolism. Although early high-dose pyridoxine treatment can control seizures, ∼75% of patients with PDE still have intellectual disabilities. In this study, we show that a homozygous mutation in lysine α-ketoglutarate reductase (LKR), an enzyme upstream of the defective ALDH7A1, completely eradicates the accumulation of toxic intermediates in lysine catabolism, ends the epileptic state, and rectifies the defective brain development and cognitive impairments in ALDH7A1-deficient mice. These findings prove the concept of the effectiveness of substrate reduction therapy for PDE by inhibiting LKR.
Introduction
Pyridoxine-dependent epilepsy (PDE, MIM #266100) is a rare autosomal recessively inherited metabolic disease. It is caused by mutations in ALDH7A1, a gene that plays a central role in lysine degradation (Coughlin et al., 2019). PDE is characterized by intractable and recurrent neonatal seizures that are resistant to conventional antiepileptic drugs but responsive to high dose of pyridoxine (Stockler et al., 2011; Coughlin et al., 2019). Deficiency in ALDH7A1 leads to accumulation of toxic intermediates in lysine catabolism, namely, α-aminoadipic-δ-semialdehyde (AASA) and its cyclic form, δ-1-piperideine-6-carboxylate (P6C). The latter inactivates pyridoxal 5′-phosphate, the active form of pyridoxine, by forming a Knoevenagel condensation product (Mills et al., 2006; Coughlin et al., 2019). Although early high-dose pyridoxine treatment can control seizures, ∼75% of PDE patients still experience intellectual disability and/or developmental delay (Coughlin et al., 2019). Despite discovery of ALDH7A1 deficiency as a genetic cause of PDE in 2006 (Mills et al., 2006), the pathophysiological mechanisms underlying the observed brain abnormalities and neurodevelopmental impairments have remained incompletely understood. Progress in this field has been hampered by two factors: difficulty in studying biochemical and neuropathological changes directly in patients’ tissues and the lack of an appropriate animal model for PDE. Recently, we have generated a PDE mouse model with a brain-specific deletion of Aldh7a1 (Yan et al., 2024; Wu et al., 2025), which exhibits cognitive dysfunction even when the seizure occurrence is well controlled. Therefore, this mouse model allows us to study the pathogenesis of PDE, as well as to test new therapeutic options.
To date, lysine reduction therapies (LRTs) have been proposed to reduce the accumulation of toxic lysine catabolism intermediates in order to improve developmental/cognitive outcomes (Coughlin et al., 2022). However, evidence for the benefit of LRTs is limited since only observational studies have been carried out (Coughlin et al., 2021). Although pyridoxine and LRTs are recommended for most PDE patients, not all patients have benefited from the addition of LRTs (Coughlin et al., 2022). Moreover, triple therapy with pyridoxine, ʟ-arginine supplementation, and dietary lysine restriction has been proposed to treat PDE (Coughlin et al., 2015). ʟ-Arginine has been known to compete for brain lysine influx and liver mitochondrial import, but elevated ʟ-arginine level leads to accelerated depletion of neural stem cell (NSC) pool in adult brain (Xu et al., 2023). Of note, given the existence of an endogenous (nondietary) source of lysine (such as through protein degradation), it prompts us to develop optimal therapeutic strategies for PDE.
Substrate reduction therapy has been proposed for treating inborn errors of metabolism. This therapy aims to decrease the levels of toxic metabolites by inhibiting an enzyme upstream of the defective enzyme (Gambello and Li, 2018). Although it remains a matter of debate, recent studies suggest that the saccharopine pathway is the predominant degradative pathway for lysine in the brain (Pena et al., 2017; Leandro et al., 2020; Leandro and Houten, 2020; Guo et al., 2022; Yan et al., 2024). In the saccharopine pathway, the first two reactions are catalyzed by a bifunctional enzyme, α-aminoadipate semialdehyde synthase (AASS; Sacksteder et al., 2000). The N-terminal and C-terminal portions of AASS possess lysine α-ketoglutarate reductase (LKR) and saccharopine dehydrogenase (SDH) activity, respectively (Sacksteder et al., 2000). LKR condenses lysine and α-ketoglutarate to form saccharopine, and subsequently SDH oxidizes saccharopine to generate glutamate and AASA (Sacksteder et al., 2000). We have shown that mice carrying a homozygous R65Q mutation in LKR of AASS (AASS-R65Q) have hyperlysinemia without any detectable clinical phenotype (Zhou et al., 2019; Guo et al., 2022). Notably, patients with hyperlysinemia with LKR deficiency have limited clinical consequences and may not require any intervention (Dancis et al., 1983; Yeganeh et al., 2023). Therefore, we hypothesized that LKR could be an ideal target for a substrate reduction therapy to treat PDE.
Materials and Methods
Animal
The mice were housed at the Institute of Genetics and Developmental Biology (IGDB), Chinese Academy of Sciences, under a 12 h reverse light/dark cycle starting at 8 A.M. All procedures and animal care followed protocols approved by the Institutional Animal Care and Use Committee at IGDB. The mice were backcrossed to the C57BL/6N background for a minimum of six generations. The Aldh7a1-cKO mice and Aass-R65Q knock-in mice were generated in our previous study (Yan et al., 2024). The Aass-R65Q;Aldh7a1-cKO mice was generated by crossing Aass-R65Q mice with the Aldh7a1-cKO mouse. Littermate WT mice were used as controls.
Thymidine analog administration
For cell proliferation analysis, mice received EdU (200 mg/kg body weight, administered via intraperitoneal injection; Sigma-Aldrich, E9386) and were assessed either 2 h or 30 d postinjection. To evaluate cell cycle exit, mice were injected with EdU (200 mg/kg body weight, i.p. injection) and examined 24 h later.
α-15N-Lysine and ε-15N-lysine administration
Mice were fasted overnight and intraperitoneally injected with lysine solutions containing 20 mg of l ʟ-[α-15 N]lysine (Cambridge Isotope Laboratories, NLM-143-0) or ʟ-[ε-15 N]lysine (Cambridge Isotope Laboratories, NLM-631-0).
Isotopic labeling analysis
Isotopic labeling of intracellular metabolites was assessed by GC-MS based on previous study (Yan et al., 2024). Briefly, dried intracellular metabolites were dissolved in 50 μl of 2% methoxylamine hydrochloride in pyridine and incubated at 37°C for 90 min on a heating block. Subsequently, 80 μl of N-(tert-Butyldimethylsilyl)-N-methyl-trifluoroacetamide with 1% tertbutyldimethylchlorosilane (Thermo Fisher Scientific) was added, and the samples were incubated for 30 min at 60°C. After overnight incubation at room temperature, the derivatized samples were briefly centrifuged, and the resulting clear liquid was transferred into GC vials for subsequent analysis. GC-MS analysis was conducted using an Agilent 7890B GC system equipped with a DB-5MS capillary column (30 m length, 0.25 mm internal diameter, 0.25 μm phase thickness; Agilent Scientific), coupled to an Agilent 5977A mass spectrometer operating with electron impact ionization at 70 eV.
EEG recording
EEG recordings were conducted on freely moving animals using an RPC-1 receiver (DSI) with Ponemah software based on previous study (Yan et al., 2024). Briefly, mice were anesthetized and secured in a stereotaxic frame. Following exposure of the skull, three holes were drilled over the mouse cortices at the following bregma coordinates: AP = +1.2 mm, ML = +1.8 mm; AP = −2.3 mm, ML = ±1.7 mm. Four electrodes (DSI), divided into two channels, were implanted into the skull for recording EEG signals. A subcutaneous radio transmitter (HD-X02) was implanted on the back. After a 14 d recovery period postsurgery, EEG recordings were conducted on freely moving animals using an RPC-1 receiver (DSI) with Ponemah software. EEG data were subsequently analyzed using NeuroScore software. Spontaneous seizures were identified by spiking activity persisting for at least 15 s with spike amplitudes exceeding two times the background amplitude.
Tissue preparation and immunohistochemistry
The immunohistochemistry was performed as previously described (Tang et al., 2019; Wang et al., 2019, 2021; Xu et al., 2023). Mice were killed by intraperitoneal injection of Avertin and then transcardially perfused with saline followed by 4% PFA. Brains were dissected out, postfixed overnight in 4% PFA, and then equilibrated in 30% sucrose. Then, 40 µm brain sections were generated using a sliding microtone and stored in a −20°C freezer as floating sections in 96-well plates filled with cryoprotectant solution (glycerol, ethylene glycol, and 0.1 M phosphate buffer, pH 7.4, 1:1:2 by volume).
The tissue sections were preblocked with TBS++ (TBS containing 3% goat or donkey serum and 0.3% Triton X-100) for 1 h at room temperature. Subsequently, they were incubated with primary antibodies at 4°C overnight. Primary antibodies utilized in this study included the following: mouse anti-NeuN (Abcam, ab104224), mouse anti-Sox2 (Abcam, ab79351, diluted 1:1,000), rabbit anti-Sox2 (Abcam, ab97959, diluted 1:1,000), and goat anti-GFAP (Abcam, ab53554, diluted 1:1,000). After washing three times, secondary antibodies (Invitrogen, 1:500) were incubated 1 h at room temperature. All sections were counterstained with a nuclear counter stain, DAPI (4′,6-diamidino-2-phenylindole). Primary antibodies were visualized with Alexa-conjugated secondary antibodies. After staining, sections were mounted, coverslipped, and maintained at 4°C in the dark until analysis.
EdU analysis
EdU staining was followed according to the manufacturer’s protocol (Beyotime, C0081L).
Dendritic spine analysis
The mice (three per group) were deeply anesthetized with chloral hydrate (1 mg·kg−1). The brains were promptly removed from the skull with utmost care to avoid tissue damage or compression. The tissue surface was rinsed quickly with double distilled or Milli-Q water to remove any blood. For staining, the FD Rapid GolgiStain Kit (FD Neurotechnologies, Consulting & Services) was used according to the manufacturer's instructions.
Following Golgi staining, observation and image acquisition were performed using an upright microscope (Leica, DM500). Dendritic spine counting was conducted in the CA1 region of the hippocampus. Spines on secondary and tertiary branches of basal dendrites were randomly observed under a 100× objective. Thirty neurons were examined per group, and images were captured accordingly. Neurons were included in the analysis under the following criteria: dendrites were intact without fragmentation, dendrites exhibited continuous Golgi staining, and measured dendrites were spaced sufficiently apart from neighboring dendrites to ensure measurements were not interfered with. Spines were categorized morphologically into thin and filopodia-like and mushroom-shaped types. Thin and filamentous pseudopodia types were considered immature, while mushroom-shaped type was classified as mature. Counts were then conducted separately for mature and immature spine types. ImageJ software was utilized for spine counting analysis.
Quantification and fate mapping of EdU+ cells in the DG
For quantification of EdU+ cells and phenotype of EdU+ cells (double labeling with either GFAP, Sox2, or NeuN), 1 in 12 serial sections starting at the beginning of the hippocampus (relative to bregma, −1.5 mm) to the end of the hippocampus (relative to bregma, −3.5 mm) were used. All indicated cells in the brain sections (4–6 sections per mouse) were counted inside the section center between 5 μm guard zones of the section surfaces under a Nikon-ECLIPSE 80i microscope with NIS-Elements, BR. 3.00 software. The data were presented as number of cells in a cubic millimeter of DG.
Behavioral tests
All the mice used for behavioral tests were housed in groups with mixed genotypes. Experimenters were blind to the genotype when behavioral tests were carried out.
Novel object location test
Mice were habituated to an empty white chamber by allowing them to freely explore for 15 min. After 24 h, every mouse was rehabituated to the empty chamber for 1 min and then placed in a holding cage while two identical objects (big clips) were placed in the corners of arena 7 cm from the walls. Mice were returned to the chamber for training and allowed to freely explore until they accumulated a total of 30 s exploring the objects (exploration recorded when the front paws or nose contacted the object). Mice were then removed from the chamber, immediately infused, and returned to their home cage. After 24 h, object recognition was tested by substituting novel object location. Time spent with each object was recorded. Measure the time spent exploring each object for 10 min in the novel location tests by watching recorded behavior (via digital video). Scoring can be done by using two stopwatches and collecting exploratory time for each object simultaneously. Novel location preference is expressed as a percentage of time exploring the novel location among the cumulative time spent exploring both objects.
Novel object recognition test
Mice have an innate preference for novelty; if the mouse recognizes a familiar object, it will spend most of its time on the new object. As a result, the New Object Recognition Test (NOR) became a commonly used behavioral analysis method for studying learning and memory in mice. Mice were habituated to an arena without objects for 10 min on Day 1. On Day 2, mice were habituated to the arena for 15 min with two identical objects placed in two corners of the arena. On Day 3, the mice are tested by using the same two objects, one object placed in a familiar position, and the second object placed in a new “novel” corner of the arena. Mice were recorded for 10 min during the test phase. The exploratory preference was calculated as the percentage of time spent investigating the object in the new location in total time spent exploring objects.
Fear conditioning test
In this task, mice learned to associate the environmental context (fear conditioning chamber) with an aversive stimulus (mild footshock; unconditioned stimulus) enabling testing for hippocampal-dependent contextual fear conditioning. To assess amygdala-dependent cued fear conditioning, the mild footshock was paired with a light and tone cue (conditioned stimulus). Freezing behavior was used as a readout of conditioned fear. First, mice were placed into a shock chamber and allowed to explore for 2 min. Then, on the training day (Day 1), specific training parameters were as follows: a white noise tone (87 dB) sounded for 30 s (conditional stimulus or “CS”). During the last 1.5 s of the tone, mice received a mild footshock (0.5 mA) (unconditioned stimulus or “US”). The same tone-footshock (CS-US) combination was delivered again 2 min later. This cycle was presented a total of three times with a 60 s interval.
The context test was performed 24 h after the training. On the testing day (Day 2), each mouse was first placed into the fear conditioning chamber containing the same context, but with no CS or US and monitored by the overhead camera for 6 min. The cue test was performed 2 h after the context test, in which colored Plexiglass inserts were placed into the training chamber to hide the shock grid and to change the “context” of the chamber. Mice were then placed in the chamber and monitored by the overhead camera for 6 min, during which two CS (spaced the same way as in the training session) were given. The events in the fear conditioning test were programmed and data recorded through the Startle and Fear conditioning system (Panlab) and Packwin software (V2.0.05).
Y-maze test
At first, the mice were subjected into the behavioral room and adapt for at least 30 min. Wipe the maze with 95% alcohol to keep it clean and dry. Then the three walls of the maze were labeled “A/B/C,” and each time the mice were placed near the center of one of the walls. Each mouse was allowed to explore freely for 8 min, and the order in which the mice entered the ABC wall was recorded. (Mice do not fully explore until all their limbs are in one wall.) If the rodent remained in the same spot for >60 s and did not exhibit exploratory behavior, the rodent was moved toward the center of the Y-maze and the experiment continued. Calculate spatial cognition by spontaneous alternation: Number of successful alternations / (total number of entries − 2). Define a successful alternations when the rodent moves to three different locations (ABC, BCA, CAB, etc.). If the rodents entered at least five branches of the maze within 8 min, the Y-shaped maze was considered a success.
Open field test
Mice were placed in an unfamiliar arena with clear side walls (10 × 10 × 16 inch; RWD Life Science) and allowed to freely explore the arena for 20 min. They were returned to their home cages after the test. Their locomotor activity was tracked by photo beams preinstalled to the arena and then analyzed by Panlab SMART 3.0 Software.
Elevated plus maze test
Mice were placed in an elevated plus maze for 5 min and tracked using Panlab SMART 3.0 Software. The amount of time spent in the open arms was measured.
Statistical analysis
All experiments were performed and analyzed by the same experimenter, blind to the animals’ genotype or group treatment under assessment. All data were shown as means with standard error of mean (means ± SEM). Statistical analysis was performed using one-way analysis of variance (ANOVA) with Tukey's post hoc test with the aid of SPSS (version 22). Probabilities of p < 0.05 were considered significant.
Results
Genetic perturbation of LKR eliminates the accumulation of toxic lysine catabolism intermediates in brain of Aldh7a1-deficient mice
Lysine in mammals is catabolized via the saccharopine and pipecolate pathways (Fig. 1A). In this process, the nitrogen epsilon (ε-N) and nitrogen alpha (α-N) of lysine are respectively deaminated (Struys and Jakobs, 2010; Pena et al., 2017). To evaluate lysine catabolism in the brain, we carried out liquid chromatography-mass spectrometry on mice injected with lysine labeled either at α-15N or ε-15N (Fig. 1A). Given that saccharopine contains both α- and ε-N derived from lysine, the incorporation of α-15N and ε-15N into saccharopine was detected in the brain of wild-type (WT) mice (Fig. 1B). However, we scarcely detected 15N-P6C and 15N-AASA in the brains of WT mice injected with α-15N-lysine and ε-15N-lysine (Fig. 1C,D), indicating a rapid turnover rate of these two toxic lysine catabolism intermediates in the brain. Pipecolate can form by converting P6C via P5C reductase (Struys and Jakobs, 2010). Although 15N-pipecolic acid was found in the brains of mice injected with both α-15N- lysine and ε-15N-lysine (Fig. 1E), 15N-α-aminoadipic acid (AAA) was mainly detected in the brains of WT mice injected with α-15N-lysine, but not ε-15N-lysine (Fig. 1F), which corroborates the idea that the saccharopine pathway is mainly responsible for lysine catabolism in the adult brain (Yan et al., 2024).
Genetic perturbation of LKR eliminates the toxic metabolites (AASA/P6C) in the brain due of ALDH7A1-deficient mice. A, Schematic illustration of the saccharopine or pipecolic acid pathways of nitrogen deamination from lysine by tracing α-15N-lysine or ε-15N-lysine. B–G, Mass spectrometry analysis of 15N-labeled saccharopine, pipecolic acid, AASA, P6C, AAA, and lysine from the catabolism of α-15N-lysine or ε-15N-lysine in the brain of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice. Values represent means ± SEM (n = 3 mice). *p < 0.05, **p < 0.01 and ***p < 0.001; one-way ANOVA with Tukey's post hoc test.
In line with our previous study (Yan et al., 2024), we observed that, in comparison with WT mice, Aldh7a1-cKO mice had an accumulation of saccharopine, AASA, P6C, pipecolic acid, and AAA derived from α-15N-lysine, but not from ε-15N-lysine (Fig. 1B–F). However, the accumulation of α-15N-saccharopine, α-15N-AASA, α-15N-P6C, and α-15N- pipecolic acid was entirely eliminated in the brains of both Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice (Fig. 1B–E). Although the incorporation of α-15N from lysine into α-15N-AAA was reduced, it could still be detected in both Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice. This result lent support to the previous assertion that there is probably an alternative route for AAA production from lysine in the brain, independent of the saccharopine and pipecolate pathways. Nevertheless, α-15N-lysine and ε-15N-lysine were significantly accumulated in the brains of Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice (Fig. 1G). Overall, the perturbation of LKR is enough to eliminate the accumulation of toxic metabolites (AASA and P6C) in the brain resulting from Aldh7a1 deficiency.
Genetic perturbation of LKR terminates the pyridoxine-dependent epileptic status of Aldh7a1-deficient mice
As per the National Research Council in 1995, a diet containing 1 ppm of pyridoxine is sufficient for the growth and maintenance of adult mice. Aldh7a1-cKO mice showed spontaneous electrographic seizures [100% of seizure occurrences (9/9 mice), with a seizure frequency of 5.22 ± 1.23 times per day and a seizure duration of 30.01 ± 2.03 s] when fed with a special diet (SD, 1.6 ppm pyridoxine), but not when fed with a regular diet (RD, 13 ppm pyridoxine; Fig. 2A,B; Table 1). These results were in line with our previous study (Yan et al., 2024), which indicated that specific deletion of Aldh7a1 in the brain leads to PDE. However, Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice did not show spontaneous electrographic seizures when fed either with SD or RD (Fig. 2B, Table 1). On the other hand, Aldh7a1-cKO mice began to die on Day 2 and only ∼60% of Aldh7a1-cKO mice survived when the SD trial was followed up for 30 d (Yan et al., 2024). Remarkably, all Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice were able to survive when fed with SD (Fig. 2C). Consequently, perturbation of LKR is enough to stop the pyridoxine-dependent seizures in Aldh7a1-deficient mice.
Genetic perturbation of LKR restores the epileptic status of ALDH7A1-deficient mice independently of pyridoxine. A, Experimental scheme for multi-channel EEG recording. B, Representative EEG for WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice feeding with RD or SD. C, Survival curves of WT, Aldh7a1-cKO mice, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice when following up SD trial for 30 d. Values represent mean ± SEM (n = 12 mice).
Seizure status of the mice
Genetic perturbation of LKR rescues abnormal adult neurogenesis and defective dendritic spine development in ALDH7A1-deficient mice
ALDH7A1-deficient mice with seizure control exhibit abnormal adult hippocampal neurogenesis and defective spine development (Yan et al., 2024; Wu et al., 2025). To determine whether genetic perturbation of LKR can rescue defective cell proliferation in Aldh7a1-deficient mice, we injected 5-ethynyl-2′-deoxyuridine (EdU) into the mice at postnatal day 60 (P60) and collected their brains for analysis 2 h later (Fig. 3A). Quantification revealed a significant reduction in the numbers of EdU+GFAP+Sox2+ NSCs and EdU+GFA−Sox2+ progenitors in Aldh7a1-cKO mice compared with WT controls, which is consistent with our previous study (Yan et al., 2024). However, Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice had numbers of EdU+GFAP+Sox2+ NSCs and EdU+GFAP−Sox2+ progenitors comparable with those in WT mice (Fig. 3B,C).
Genetic perturbation of LKR rescues abnormal adult neurogenesis and defective dendritic spine development in ALDH7A1-deficient mice. A, Experimental scheme for assessing cell proliferation by a 2 h EdU pulse-chase in P60 mice. B, C, Quantification of GFAP+ Sox2+ EdU+ NSCs and GFAP−Sox2+ EdU+ progenitors in the adult DG of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice. Values represent means ± SEM (n = 6 mice). D, Experimental scheme for assessing adult-born neurons by a 30 d EdU pulse-chase in P60 mice. E, Quantification of NeuN+EdU+ adult-born neurons in the adult DG of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice. Values represent means ± SEM (n = 6 mice). F, Representative images of Golgi-stained dendritic spines on pyramidal neurons in the CA1 region of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice at P42. Scale bars, 5 μm. G, Quantification of the density of dendritic spines in CA1 pyramidal neurons. Values represent mean ± SEM (30 neurons from 3 mice). H, Quantification of the proportion of mature and immature spines in CA1 pyramidal neurons of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice at P42. Values represent mean ± SEM (30 neurons from 3 mice). *p < 0.05, **p < 0.01; ***p < 0.001; Student's t test.
To examine the production of adult-born mature neurons, we injected P60 mice with EdU four times at 12 h intervals and collected brain tissues 30 d later (Fig. 3D). Quantification showed a significant decrease in the number of EdU+NeuN+ neurons in Aldh7a1-cKO mice compared with WT controls, which is consistent with our previous study (Yan et al., 2024). Notably, Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice had the number of EdU+NeuN+ adult-born mature neurons similar to that in WT mice (Fig. 3E).
To evaluate whether genetic perturbation of LKR can rescue defective spine development in ALDH7A1-deficient mice, we carried out a Golgi staining assay and examined the pyramidal neurons in the hippocampal CA1 region of adult mice. There was a significant increase in the density of dendritic spines in Aldh7a1-cKO mice, but not in Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice (Fig. 3F,G). In particular, the number of immature spines (filopodia-like) was increased in Aldh7a1-cKO mice, accompanied by a decrease in the number of mature spines (mushroom-like; Fig. 3H). Notably, Aass-R65Q and Aass-R65Q;Aldh7a1-cKO mice had numbers of immature and mature spines comparable with those in WT mice (Fig. 3H).
Collectively, these findings indicate that genetic perturbation of LKR restores defective adult neurogenesis and abnormal dendritic spine development in Aldh7a1-cKO mice.
Genetic perturbation of LKR restores defective cognitive functions in Aldh7a1-deficient mice
Although the seizure control is achieved when fed with RD, Aldh7a1-deficient mice displayed impaired cognitive function (Yan et al., 2024; Wu et al., 2025). This was manifested by a reduction in alternation rates during the Y-maze test (Fig. 4A,B), a decrease in the exploration preference for the novel location of the same object in the novel object localization (NOL) assessment (Fig. 4C,D), or for the novel object in the novel object recognition (NOR) test (Fig. 4E,F), as well as a reduction in freezing behaviors in the fear conditioning contextual and cue tests (Fig. 4G–I). These results were consistent with our previous study (Yan et al., 2024). However, analysis revealed no differences in cognitive performances between WT and Aass-KO mice in the Y-maze test (Fig. 4A,B), NOL test (Fig. 4C,D), NOR test (Fig. 4E,F) and the fear conditioning contextual and cue test (Fig. 4G–I). Notably, Aass-R65Q;Aldh7a1-cKO mice showed performances similar to those of WT mice in the Y-maze test (Fig. 4A,B), the NOL test (Fig. 4C,D), the NOR test (Fig. 4E,F), and the fear conditioning contextual and cue test (Fig. 4G–I). Nonetheless, Aldh7a1-cKO, Aass-R65Q, Aass-R65Q;Aldh7a1-cKO mice behaved normally like WT mice in the open field (Fig. 4J–L) and elevated plus maze tests (Fig. 4M,N), suggesting that these mice do not exhibit abnormal locomotion and anxiety-like behaviors. Altogether, perturbation of LKR is sufficient to rescue the cognitive impairment of Aldh7a1-deficient mice.
Genetic perturbation of LKR recovers the cognitive impairment in Aldh7a1-cKO mice. A, Schematic diagram of Y-maze test. B, Quantification of spontaneous alternations of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q; Aldh7a1-cKO mice in Y-maze test. C, Schematic diagram of novel object location (NOL) test. D, Quantification of exploring time of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice in NOL test. E, Schematic diagram of novel object recognition (NOR) test. F, Quantification of exploring time of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice in NOR test. G, Schematic diagram of fear conditioning test. H, Quantification of freezing behaviors of WT, Aldh7a1-cKO mice, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice in the contextual fear conditioning test. I, Quantification of freezing behaviors of WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice in the cue fear conditioning test. J, Representative traces of movement for WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice during open field test. K, L, Moving distance and time spend in the center of the arena or in the margin of the arena for WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice during open field test. M, Representative traces of movement for WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice in elevated plus maze test. N, Total time spend in open arms for WT, Aldh7a1-cKO, Aass-R65Q, and Aass-R65Q;Aldh7a1-cKO mice during elevated plus maze test. Values represent mean ± SEM (n = 10 mice). *p < 0.05, **p < 0.01, and ***p < 0.001; one-way ANOVA with Tukey's post hoc test.
Discussion
Lysine is broken down in mammals via the pipecolate and saccharopine pathways. Traditionally, the former is considered to be the main degradative route for lysine in brain tissue (Chang, 1976), while the latter mainly occurs in the liver and kidneys (Blemings et al., 1994). However, it should be noted that recent studies also provide evidence for the existence of an active saccharopine pathway in the brain (Pena et al., 2017; Crowther et al., 2019; Leandro et al., 2020; Leandro and Houten, 2020; Guo et al., 2022). By tracing 15N-labeled lysine at its α- or ε-N-atoms in the brain, we discovered that the accumulation of toxic lysine catabolism intermediates (AASA/P6C) resulting from ALDH7A1 deficiency was entirely eliminated in Aass-R65Q;Aldh7a1-cKO mice. These genetic evidences strongly suggest that the saccharopine pathway is the dominant degradative pathway for lysine in the brain.
Mutations in AASS, a gene initiating the saccharopine pathway, are known to lead to hyperlysinemia, an autosomal recessive inborn error of lysine degradation, which can be divided into two subtypes. Hyperlysinemia-I is caused by mutations in LKR and affected individuals have elevated lysine levels but are likely to be without clinical symptoms (Dancis et al., 1983; Houten et al., 2013; Yeganeh et al., 2023). Hyperlysinemia-II (also known as saccharopinuria), on the other hand, is caused by mutations that mainly affect SDH activity. Patients with hyperlysinemia-II have a significant elevation in both lysine and saccharopine levels and are mostly accompanied by neurological damage and intellectual disability (Houten et al., 2013). In our previous studies, we created two Aass knock-in mouse models (Zhou et al., 2019; Guo et al., 2022). The first model contains a mutation in the LKR domain (Aass-R65Q), and the second one has a mutation in the SDH domain (Aass-G489E), which results in impaired SDH activity while maintaining LKR function (Zhou et al., 2019; Guo et al., 2022). The Aass-R65Q mice show no detectable clinical phenotype except for an elevation in lysine levels, whereas the Aass-G489E mice exhibit mitochondrial damage and functional impairment in the liver and brain, leading to liver hypertrophy, defective brain development, postnatal developmental retardation, and death, probably due to the accumulation of saccharopine (Zhou et al., 2019; Guo et al., 2022). In this study, we demonstrated that Aass-R65Q mice have normal brain development, intact cognitive functions, and no susceptibility to seizures. Moreover, Aass-R65Q;Aldh7a1 double KO mice showed normal cognitive functions without seizure susceptibility. In conclusion, our data prove the concept of the effectiveness of substrate reduction therapy for inborn errors of lysine metabolism, including PDE, through the inhibition of LKR.
Aldh7a1 and Aass are highly expressed in the adult mouse liver and kidney and are expressed to a lesser extent in the brain (Pena et al., 2017; Johal et al., 2024). It is believed that the saccharopine pathway in the liver and kidney is more highly active than that in the brain. Using a brain-specific Aldh7a1 knock-out mouse model, we have demonstrated here that inhibition of LKR can reverse the brain dysfunction caused by the specific loss of Aldh7a1 in the brain. However, in patients with PDE who have an ALDH7A1 mutation, AASA/P6C is expected to accumulate in organs other than the brain. In future studies, it is worthwhile to investigate whether these toxic metabolites produced in peripheral tissues are delivered to the brain, thereby causing brain dysfunction. Notably, Aass-R65Q mice lose LKR activity in both cerebral and peripheral tissues (Zhou et al., 2019; Guo et al., 2022), which enables us to test whether inhibition of LKR activity in peripheral tissues can ameliorate the pathogenesis of PDE.
Footnotes
This work was supported by grants from the National Science Foundation of China (82271202 and 32394030 to W.G.) and STI2030-Major Projects (2021ZD0202302 to W.G.). The General Program of Guangdong Natural Science Foundation, Guangdong province, China (2022A1515010654 to X.Z.) and the cultivation project for National Natural Science Foundation of China from The Third Affiliated Hospital of Sun Yat-sen University (2023GZRPYMS02 to X.Z.).
↵*Z.L. and J.W. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Xiaofen Zhong at zhongxf29{at}mail.sysu.edu.cn or Weixiang Guo at wxguo{at}genetics.ac.cn.
