Neuronal Slc39a8 deletion leads to reduced Mn levels in the brain. To identify the role of SLC39A8 in neurons, we generated Slc39a8-NSKO mice by crossing Slc39a8-flox (Slc39a8fl/fl) mice with Synapsin I-Cre mice (Figure 1A). The quantitative PCR (qPCR) analysis verified lower Slc39a8 mRNA levels in the brains of Slc39a8-NSKO and Slc39a8 neuron-specific heterozygous (NSHet) mice (–44.5%, P < 0.001 and –33.3%, P < 0.05, respectively) than in control Slc39a8fl/fl mice but normal Slc39a8 expression in other tissues, including liver, heart, and kidney (Figure 1B). The Slc39a8-NSKO mice were born at expected Mendelian ratios and exhibited grossly normal appearance and body weight (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.168440DS1), indicating that neuron-specific deletion of Slc39a8 did not lead to lethality. We also compared the brain weights of Slc39a8-NSKO mice with those of the control mice and observed no substantial differences (Supplemental Figure 1B).

Loss of Slc39a8 in neurons results in Mn deficiency in the brain region. (A) Schematic representation of mice with deletion of Slc39a8 in the neurons. (B) qPCR analysis of Slc39a8 expression in 4-week-old control and Slc39a8-NSKO mice. (C−H) ICP-MS analysis of Mn levels in olfactory bulbs (OB) (C), prefrontal cortex (PFC) (D), cortex (CTX) (E), hippocampus (HPC) (F), midbrain (MB) (G), and cerebellum (CB) (H) of 4-week-old male and female control and Slc39a8-NSKO mice. Data are presented as individual values and represent the mean ± SEM. * P < 0.05, ** P < 0.01, and *** P < 0.001.

To examine whether the loss of Slc39a8 in neurons alters metal homeostasis, we used inductively coupled plasma mass spectroscopy (ICP-MS) to measure the concentrations of metals in the whole brain and brain regions from Slc39a8-NSKO and control mice. The Mn concentrations were lower in the whole brain of Slc39a8-NSKO mice than in the controls (–23.1%, P < 0.001) (Supplemental Figure 2A). To further determine whether the regulation of Mn by neuron-specific Slc39a8 was sex specific, we measured Mn levels in the same tissues of male and female mice. We observed reduced Mn levels in both female (–21.6%, P < 0.05) and male (–24.5, P < 0.01) Slc39a8-NSKO whole brains (Supplemental Figure 2A). We also found that although SLC39A8 can mediate the cellular uptake of zinc (12) and iron (13), the concentrations of these metals did not show any differences (Supplemental Figure 2, B and C). Furthermore, concentrations of Mn and other metals in the liver and whole blood did not differ between the 2 groups (Supplemental Figure 2, D–I).

We next determined whether Slc39a8-NSKO brains show region-specific reductions in Mn levels by quantifying the distribution of metals in different brain regions. The Mn concentrations were specifically reduced in the cerebellum (–17.9%, P < 0.001) of both female (–15.1%, P < 0.01) and male (–19.3%, P < 0.05) Slc39a8-NSKO mice (Figure 1, C–H). The zinc concentration in the female Slc39a8-NSKO hippocampus was reduced (–8.7%, P < 0.05; Supplemental Figure 3); however, the overall concentrations of zinc and iron in the brain regions did not differ between the 2 groups (Supplemental Figure 3). Taken together, our Slc39a8-NSKO mouse model results indicate that neuron-specific deletion of Slc39a8 leads to brain Mn deficiency, especially in the cerebellum, with little impact on other metal levels.

Slc39a8-NSKO mice display impaired brain uptake of 54Mn. The lower Mn concentrations in the whole brain and brain regions of the Slc39a8-NSKO mice (Figure 1 and Supplemental Figure 2) suggested that Slc39a8-NSKO mice might have a defect in brain Mn uptake. We tested this possibility by measuring radioactivity in the brain regions of control and Slc39a8-NSKO mice at 1 hour after intravenous injection of 54Mn. The 1-hour time point was chosen based on previous studies of brain uptake of Mn in mice (38, 39). We observed that brain 54Mn levels were substantially reduced in the hippocampus (–22%, P < 0.05) and cerebellum (–35%, P < 0.001) in the Slc39a8-NSKO mice compared with control mice (Figure 2). Notably, the amount of 54Mn in the cerebellum was reduced in both female (–21.2%, P < 0.01) and male (–11.1%, P < 0.05) Slc39a8-NSKO mice compared with control mice (Figure 2).

Loss of Slc39a8 in neurons results in impaired brain Mn uptake. Control and Slc39a8-NSKO mice at 4 weeks of age were administered 0.1 μCi [54Mn]MnCl2 per gram body weight via tail vein injection. Brain regions were collected at 1 hour, and 54Mn uptake was determined with a gamma counter (counts per minute [cpm]). Levels of 54Mn in olfactory bulbs (OB) (A), prefrontal cortex (PFC) (B), cortex (CTX) (C), hippocampus (HPC) (D), midbrain (MB) (E), and cerebellum (CB) (F) from 4-week-old male and female control and Slc39a8-NSKO mice. Data are presented as individual values and represent the mean ± SEM. * P < 0.05, ** P < 0.01, and *** P < 0.001.

Animals acquire Mn primarily through diet under normal conditions. We also measured radioactivity in the brain regions of control and Slc39a8-NSKO mice 1 hour after intragastric gavage of 54Mn. The amount of 54Mn in the cerebellum was reduced by 35.4% (P < 0.01) in Slc39a8-NSKO mice compared with control mice (Supplemental Figure 4, A–F). The amount of 54Mn in the blood was similar between control and Slc39a8-NSKO mice 1 hour after administration of 54Mn via intravenous injection or intragastric gavage (Supplemental Figure 4, G and H). The total amount of 54Mn was also measured similarly between control and Slc39a8-NSKO mice, verifying uniform exposure conditions (Supplemental Figure 4, I and J). The residual 54Mn levels observed in the Slc39a8-NSKO cerebellum suggest an alternative entry route for Mn into neurons. To address this, we examined the expression of other metal ion transporters that might compensate for Mn uptake. Notably, the cerebellum of Slc39a8-NSKO male mice exhibited an upregulation in Slc39a14/ZIP14 transcript levels, although this difference did not reach statistical significance (Supplemental Figure 5, P = 0.0649).

No substantial changes were observed in other tested metal ion transporters, including Slc30a10/ZnT10, Slc11a2/DMT1 (divalent metal transporter 1), and Slc40a1/FPN (ferroportin-1) transcript levels (Supplemental Figure 5). These findings suggest that Slc39a14 may compensate for the loss of SLC39A8 expression. Overall, these data indicate that neuronal Slc39a8 deficiency impairs brain 54Mn uptake.

Cerebellar morphological defects in Slc39a8-NSKO mice. We next sought to determine whether the loss of Slc39a8 in neurons leads to alterations in overall brain architecture. The appearance of the brain of Slc39a8-NSKO mice revealed no gross abnormalities. However, systematic histological analysis revealed a striking difference in the foliation and fissuration of the cerebellum in Slc39a8-NSKO mice at P28 (Figure 3A). In control mice, the intercrural fissure normally separates lobules VI and VII, but this fissure was absent in the Slc39a8-NSKO cerebellum (Figure 3A, indicated by red arrows and red outlines). The observed cerebellar morphological defects were indeed present in multiple Slc39a8-NSKO mice at P28, albeit with incomplete penetrance (Supplemental Figure 6). These results suggest a critical role for Slc39a8 in cerebellar development.

Morphological defects, reduced neurogenesis, and accelerated apoptosis in Slc39a8-NSKO cerebellum. (A) H&E-stained sagittal sections of paraffin-embedded mouse brains from 4-week-old control and Slc39a8-NSKO mice. Numerals indicate the lobules, and the fissures between lobules VI and VII of cerebellums are outlined with red lines and indicated by red arrows. Scale bars: 1,000 μm. Upper original magnification, 2×, and lower original magnification, 10×. (B) Immunohistochemical staining of cerebellar Purkinje cells (PCs) with anti-calbindin antibody. PC morphologies of control and Slc39a8-NSKO mice at P14. EGL, external granule cell layer; ML, molecular layer; PCL, Purkinje cell layer; IGL, internal granule cell layer. Scale bars: 25 μm. (C) Numbers of PCs at P14. (D) BrdU staining on the cerebellar sections at P14. Scale bars: 100 μm. (E) The number of BrdU+ cells per unit size of sagittal sections from the entire cerebellum was quantified. (F) TUNEL staining on the cerebellar sections at P14. Scale bars: 200 μm. Insets original magnification, 20×. (G) Quantification of the number of TUNEL-positive cells per field of cerebellar folia. At least 9 sections from 3 animals per genotype were quantified for all panels. Data are presented as individual values and represent the mean ± SEM. * P < 0.05, ** P < 0.01, and *** P < 0.001.

We also examined the morphological abnormalities in the cerebellum of Slc39a8-NSKO mice by staining the Purkinje cells (PCs) with an anti–calbindin D antibody, a specific marker for this cell type, at P8 (Supplemental Figure 7A) and P14 (Figure 3B). The cerebellar cortex in the control and Slc39a8-NSKO mice displayed the same orderly, 4-layered structural organization of the EGL, ML, PCL, and IGL (40) (Supplemental Figure 7A and Figure 3B). However, the numbers of PCs were noticeably reduced in Slc39a8-NSKO mice at both P8 and P14, with lobules IV and VII exhibiting the most significant reductions (Supplemental Figure 7B and Figure 3C). These results indicate that abnormal cerebellar histogenesis occurs during early postnatal development in Slc39a8-NSKO mice.

Reduced neurogenesis and accelerated apoptosis in the EGL in Slc39a8-NSKO cerebellum. The mouse cerebrum reaches near maturity at birth, whereas the cerebellum continues to grow postnatally (41, 42). During postnatal development, the granule cell precursors (GCPs) proliferate, differentiate into mature granule cells, and migrate to form the IGL (41). Recognizing that cerebellar morphogenesis largely depends on GCP proliferation (41), we sought to determine whether the morphological defects in the Slc39a8-NSKO cerebellum can be explained mechanistically through decreased cell proliferation or increased apoptosis. We performed a BrdU staining to examine whether GCPs are altered in the developing cerebellum of Slc39a8-NSKO mice at P14. We observed a reduction in the number of BrdU-positive cells in Slc39a8-NSKO mice compared with controls (Figure 3, D and E). We then performed a terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay to examine whether neuronal cell death is altered in the developing cerebellum of Slc39a8-NSKO mice at P14. The number of TUNEL-positive cells was higher in the EGL of Slc39a8-NSKO mice than in the controls (Figure 3, F and G). These results indicate that deletion of Slc39a8 in neurons impairs neurogenesis and accelerates apoptotic cell death in the EGL of the cerebellum, suggesting an association with the defects in cerebellar morphogenesis of Slc39a8-NSKO cerebellum.

Loss of Slc39a8 in neurons results in defects in dendritic arborization and spine morphology in PCs. Abnormal dendritic arborization and spine morphology have been suggested as a cellular basis of neurodevelopmental disorders (43). The role of Slc39a8 in dendritic growth and spine morphogenesis is not known. Thus, we sought to determine whether dendritic growth in the cerebellum is affected by Slc39a8 deficiency. Cerebellum impregnated by Golgi staining clearly exhibited a severe reduction in the extent of the spiny dendritic arborization of individual Slc39a8-NSKO PCs compared with controls (Figure 4A). Many dendritic spines on the Slc39a8-NSKO neurons were remarkably thinner than those of control neurons and lacked the mature mushroom-like morphology (Figure 4A). Golgi staining followed by Sholl analysis showed a shorter total dendritic length in Slc39a8-NSKO mice than in the controls (–11.3%, P < 0.01, Figure 4B). The Slc39a8-NSKO mice also exhibited lower spine density (–25.9%, P < 0.01, Figure 4C). Moreover, the reduced spine density in the Slc39a8-NSKO mice was evident from the proximal to distal segments of dendritic branches (Figure 4D). These results suggest that developmental defects in the PCs of Slc39a8-NSKO mice affect the fine structure of dendritic arborization.

Impaired dendritic arborization of PCs in Slc39a8-NSKO cerebellum. (A) Representative images of Golgi-stained PCs from control and Slc39a8-NSKO mice cerebellum at P28. Scale bars: 50 μm (left) and 5 μm (right). (B−D) Quantification of total dendritic lengths (B), density (C), and spine number (D) in male and female control and Slc39a8-NSKO mice. At least 6 sections from 3 animals per genotype were quantified for all panels. Data are presented as individual values and represent the mean ± SEM. * P < 0.05, ** P < 0.01.

Slc39a8-NSKO mice exhibit impaired motor coordination. The morphological defects observed in the Slc39a8-NSKO cerebellum prompted us to investigate the impact on cerebellar motor function. To assess motor function comprehensively, we applied 4 behavioral tests: the hind limb clasping test, the rotarod test, the balance beam test, and the open field test. The hind limb clasping test measures the occurrence of limb clasping during tail suspension, serving as a functional test for corticospinal deficits (44). Reduced hind limb reflexes were observed in Slc39a8-NSKO mice compared with controls (Figure 5, A and B, P < 0.001). This indicates a potential deficit in corticospinal function. Next, we evaluated motor coordination using the rotarod test. In this test, mice were placed on a rotating rod with gradually accelerating speed, and the time they stayed on the rod before falling off (retention time) was used as an indicator of overall motor coordination (45). Slc39a8-NSKO mice exhibited a shorter latency on the rod compared with controls (Figure 5C, P < 0.001), suggesting impaired motor coordination. The balance beam test assesses fine motor skills and balance by requiring mice to walk across an elevated narrow beam to reach a safe platform (46). The overall performance on the balance beam was reduced in Slc39a8-NSKO mice, though this difference did not reach statistical significance (Figure 5D, P = 0.0794). However, when the data were analyzed by sex, the female Slc39a8-NSKO mice showed a significant reduction in the ability to traverse the balance beam (Figure 5D, P < 0.05). This highlights a sex-dependent effect on fine motor skills and balance. The open-field assay involved placing the mice in a chamber equipped with sensors to monitor their movement for 30 minutes (47). The overall distance traveled and center zone duration (%) were similar between the genotypes across sexes (Figure 5, E and F). Notably, Slc39a8-NSKO mice exhibited a longer duration (%) in the intermediate zone (Figure 5G, P < 0.01) and the outer zone (Figure 5H, P < 0.05), with a more pronounced effect in male Slc39a8-NSKO mice (Figure 5, G and H). These results suggest altered exploratory behavior and anxiety-like behavior, particularly in males. In summary, our integrated analysis of the neurobehavioral tests reveals that Slc39a8-NSKO mice exhibit a range of motor function deficits, including impaired corticospinal function, motor coordination, and fine motor skills. Additionally, there are sex-dependent sensitivities, with female mice showing significant impairments in balance and male mice exhibiting altered anxiety-like behavior. These findings collectively underscore the broad and sex-biased impact of Slc39a8 deficiency on motor function and behavior.

Motor dysfunction in Slc39a8-NSKO mice. (A) Representative images of a 4-week-old control and Slc39a8-NSKO mouse during tail suspension in the clasping test. (B) Quantitative analysis of the hind limb clasping scores. (C) Rotarod performance of 4-week-old male and female control and Slc39a8-NSKO mice. (D) Balance performance of 4-week-old male and female control and Slc39a8-NSKO mice. (E–H) Open field test of 4-week-old male and female control and Slc39a8-NSKO mice. Distance traveled (E), center zone duration (%) (F), intermediate zone duration (%) (G), and outer zone duration (%) (H). Data are presented as individual values and represent the mean ± SEM. * P < 0.05, ** P < 0.01, and *** P < 0.001.

Slc39a8-NSKO cerebellum shows altered transcriptome profiles. To interrogate the molecular mechanisms contributing to impaired cerebellar development, we determined the impact of Slc39a8 neuronal deficiency on the cerebellum transcriptome. We focused on the cerebellum because Slc39a8-NSKO cerebellum showed a pronounced reduction in Mn levels (Figure 1), impaired brain Mn uptake (Figure 2), abnormal development (Figure 3), and defects in spine morphogenesis (Figure 4). We used male mice to include Y-linked genes along with all other chromosomes. A principal component analysis (PCA) plot showed that the 3 replicates clustered together and were segregated into 2 genotype groups, indicating that Slc39a8 neuronal deficiency triggered transcriptomic alterations (Figure 6A). Slc39a8 was downregulated in the Slc39a8-NSKO cerebellum, verifying the validity of the RNA-Seq approach (Figure 6B). Differentially expressed genes in Slc39a8-NSKO and control mice were determined by DESeq2 (48). A total of 36 genes were differentially regulated in the cerebellum of Slc39a8-NSKO mice, with an adjusted P value (Padj) < 0.05 (Supplemental Table 1). Of these, 7 genes were upregulated (19.4%), and 29 genes were downregulated (80.6%) (Figure 6, C and D). The downregulated genes with the 15 largest log2 fold-changes are presented in Table 1. A UCSC Genome Browser shot of the Nr4a2 locus shows a decreased read density in the Slc39a8-NSKO mice compared with the controls (Figure 6E). Transcripts that were reduced in the Slc39a8-NSKO cerebellum included genes related to CNS development, neuronal differentiation, neurite outgrowth, neuronal survival, circadian rhythms, and CNS myelination (Table 1). The transcripts that increased in the Slc39a8-NSKO cerebellum included genes involved in muscle contraction, proteolysis, and transcription regulation (Supplemental Table 1). A qPCR analysis using an independent set of RNA samples verified the downregulation of Nr4a2, Nr4a3, Apold1, Per1, and Fosb and the upregulation of Myh6 (Figure 6F). These results indicate that Slc39a8 neuronal deficiency leads to the misregulation of specific genes in the cerebellum.

Altered transcription profiles in 4-week-old male Slc39a8-NSKO cerebellum. (A) Principal component analysis (PCA) plot showing the clustering of each of the samples with technical triplicates along 2 principal components (PC1 — 39% variance; PC2 — 24% variance). Each technical triplicate clusters with the other. (B) Relative read counts of Slc39a8. (C) A total of 29 genes (80.6%) are downregulated and 7 genes (19.4%) are upregulated. (D) Volcano plot profiles of –log10Padj value and log2 fold-change of gene expression between control and Slc39a8-NSKO cerebellum samples. (E) UCSC Genome Browser shot of the Nr4a2 locus. (F) qPCR validation of Nr4a2, Nr4a3, Apold1, Per1, Fosb, and Myh6, 6 genes that were shown to be dysregulated in the RNA-Seq analysis of the cerebellum of control and Slc39a8-NSKO mice. (G) A luciferase construct with 3 cAMP-responsive elements (CRE: TGACGTCA) inserted upstream of the HSV–thymidine kinase (HSV-TK) promoter. SH-SY5Y cells were cotransfected with the luciferase construct and expression plasmids for SLC39A8 or scramble siRNAs. cAMP signaling was elicited by the treatment of forskolin, a cAMP analog, and the luciferase activity was measured. (H) The mutated 2 critical nucleotides in the CRE sequence (mCRE: TGATATCA) were used as a negative control. The same experiment was performed using the mutated (mCRE) reporter. Data are presented as individual values and represent the mean ± SEM (n = 3 samples/group). * P < 0.05, ** P < 0.01.

The key downregulated genes in the Slc39a8-NSKO cerebellum

Slc39a8 is required for optimal transcriptional response to the cAMP-mediated signaling pathway. Global transcriptome analysis revealed that cAMP-related genes were among the top downregulated genes (Table 1). The cAMP signaling pathways play crucial roles in diverse physiological processes, such as neurodevelopment, synaptic plasticity, and neuroprotection (49). The cAMP signaling elicited by extracellular stimuli leads to phosphorylation of the transcription factor, cAMP response element binding protein (CREB), and the expression of CREB target genes, resulting in multiple physiological functions (50). For example, CREB regulates cell proliferation, differentiation, and survival in the developing brain while it participates in neuronal plasticity, learning, and memory in the adult brain (51). Notably, the Slc39a8-NSKO cerebellum showed downregulation of key genes whose products mediate cAMP signaling (Table 1). Examples of downregulated genes included the NR4A family genes Nr4a1, Nr4a2, and Nr4a3; the FOS families Fos, Fosb, and Fosl2; and the circadian clock gene Per1, which all have well-characterized roles as CREB-responsive inducible genes (52, 53).

The observation of downregulation of cAMP signaling mediators and CREB target genes in Slc39a8-NSKO cerebellum (Table 1) prompted us to test whether Slc39a8 deficiency alters the induction of cAMP-mediated gene transcription. We used a reporter plasmid with 3 cAMP-responsive elements (CRE: TGACGTCA) upstream of the firefly luciferase gene (54), as well as a reporter plasmid with mutations in 2 critical nucleotides in the CRE sequence (mCRE: TGATATCA) as a negative control (54). These luciferase constructs and SLC39A8 or scramble siRNA were transfected into human neuroblastoma SH-SY5Y cells, and the treatment with forskolin, a cAMP analog, stimulated cAMP signaling. The luciferase reporter assay revealed no changes in the basal activity of the CRE-luciferase construct by SLC39A8 knockdown (Figure 6G). However, following forskolin stimulation, cells with SLC39A8 siRNA could not induce as high an expression of CRE-luciferase as was observed in the scramble siRNA–treated cells (Figure 6H). The construct with the mutated CRE sequence did not respond to forskolin or change upon SLC39A8 knockdown, indicating that this effect is cAMP/CREB dependent. These results demonstrate that SLC39A8 is required for the transcriptional response to cAMP-mediated signaling.