Deficient Import of Acetyl-CoA into the ER Lumen Causes Neurodegeneration and Propensity to Infections, Inflammation, and Cancer

The S113R mutation does not affect the subcellular localization or the orientation of the C terminus of AT-1. A, Subcellular localization of AT-1 assessed in living H4 cells overexpressing a GFP-tagged version of WT (AT-1_WT) or mutant (AT-1_S113R) AT-1. In both cases, the GFP-tag was inserted at the C terminus so as not to disrupt the N-terminal signal anchor that ensures membrane insertion of AT-1. a, Phase-contrast; b, GFP fluorescence; c, ER tracker; d, merge of a, b, and c; e, phase-contrast; f, GFP fluorescence; g, ER tracker; and h, merge of e, f, and g. B, Predicted topology of WT (left) and SPG42-associated S113R mutant (right) AT-1. The following prediction models were used: a, SOSUI; b, PredictProtein; and c, UniProtKB. C, Schematic view of the experiment reported in D. The myc tag is shown in red. D, ER vesicles from AT-1_WT and AT-1_S113R expressing cells were incubated with an anti-myc antibody covalently attached to aldehyde-activated agarose beads for immunoprecipitation. After extensive washing, bound proteins were eluted by lowering the pH and analyzed by SDS-PAGE and immunoblotting. Vesicles were used under the sealed (no detergent) or open (with detergent) condition.

AT-1^S113R/+ mice are functionally haploinsufficient. A, Generation of AT-1^S113R/+ mice. A, SLC33a1-mini targeting vector fragment. B, SLC33a1 retrieval vector. C, SLC33a1 targeting vector. D, Murine chromosome 3. E, Targeted chromosome 3. F, Neo excised chromosome 3. B, SLC33a1 retrieval vector. C, SLC33a1 targeting vector. D, Murine chromosome 3. E, Targeted chromosome 3. F, Neo-excised chromosome 3. B, AT-1^S113R/+ mice are heterozygous for the mutation. C, Native ER vesicles from AT-1^S113R/+ mice display reduced acetyl-CoA influx. Two independent animals for each genotype were analyzed. Values are the mean ± SD (n = 3). D, Immunoblot showing reduced N^ε-lysine acetylation of ER proteins in AT-1^S113R/+ mice. Calreticulin was used as an ER loading marker.

AT-1^S113R/+ mice display increased propensity to infections, disseminated tissue inflammation, and cancer. A, Mice homozygous for the S113R mutation display developmental arrest. Embryos in the figure were collected at E10.5. B, Adult AT-1^S113R/+ mouse with conjunctivitis and severe dermatitis. C, Lifespan of WT and AT-1^S113R/+ mice housed in an open (top; n = 30) and pathogen-free (bottom; n = 50) facility. AT-1^S113R/+ mice display reduced survival in the open facility. D, Postmortem autopsy of the animal shown in B revealed inflammatory infiltration of different tissues/organs: a, skin (ulcerative dermatitis); b, stomach mucosa (gastritis); c, prostate parenchyma (prostatitis); d, bladder mucosa (cystitis); e, salivary gland (sialadenitis); and f, preputial gland (adenitis). E, Postmortem autopsy of the animal shown in B revealing different malignant foci: a, spindle cell sarcoma of the spine; b, spindle cell sarcoma invading the pleura; c, focal pulmonary adenoma; and d, sarcoma invading the bone structure of the knee.

Phenotypic characterization of the immune system. A, Dot plots showing frequency of B220⁺ B-cell precursors in bone marrow including: pre-pro B cells (CD43⁺CD19⁻), pro-B cells (CD43⁺CD19⁺), pre-B cells (CD43⁻CD19⁺CD2⁻B220^int), and mature B cells (CD43⁻CD19⁺CD2⁺B220^high). B, Representative histograms showing the percentage of cells expressing the activation marker LFA-1 on CD4⁺ and CD8⁺ cells isolated from lymph nodes (summarized in bar graphs). C, Representative dot plots showing frequency of T-cell precursors in thymus. CD4⁻CD8⁻ double-negative (DN) cells were subdivided into developmental stages DN1 (CD44⁺CD25⁻), DN2 (CD44⁺CD25⁺), and DN3 (CD44⁻CD25⁺). CD4⁺ and CD8⁺ single-positive populations are also shown, along with percentage of cells expressing the T-cell coreceptor CD3 (histogram). D, Frequency of granulocytic CD11b^highLy6g⁺SSC^high and monocytic CD11b⁺Ly6g⁻SSC^low myeloid cells in the spleen. E, T-helper lineage-specific intracellular cytokine production among lymphocytes stimulated with anti-CD3 and anti-CD28 for 4–5 h. Representative dot plots show IL-17 (Th17) and IFNγ (Th1); IL-4 (Th2) and IL-10 (T_reg) are not depicted. In all panels, values are the mean (n = 3) ± SD. *p < 0.05; **p < 0.005; ***p < 0.0005.

AT-1^S113R/+ mice display defective hind-leg clasping, reduced grip strength, and reduced sensitivity to pain. A, Normal clasping in WT mice. B, Pathological clasping in AT-1^S113R/+ mice. C, Pathological hind-leg clasping reflex in AT-1^S113R/+ animals. D, AT-1^S113R/+ mice have a reduced latency to fall on the inverted cage lid test. Values are the mean ± SD (n = 10). ***p < 0.0005. E, AT-1^S113R/+ mice display reduced sensitivity to pain. Values are the mean ± SD (n = 5). *p < 0.05. Animals were 10–12 months old when studied.

Loss of axonal fibers in the sciatic nerve of AT-1^S113R/+ mice. A, Toluidine staining. Left, Representative images. Right, Computer-assisted quantitation (n = 3 animals; 7 nerve sections per animal). Computer-assisted quantitation was done with entire/intact serial sections. Error bars indicate SD. ***p < 0.0005. B, C, Immunolabeling with anti-neurofilament (NFL) antibody. Both transverse (B) and longitudinal (C) sections are shown. In B, high-magnification of indicated areas is also shown. D, Computer-assisted quantitation of NFL staining per 100 μm² section-areas (n = 3 animals; 5 nerve sections per animal). Error bars indicate SD. ***p < 0.0005. E, Electron microscopy of sciatic nerve. a, WT; b and c, AT-1^S113R/+ (two different animals are shown). AT-1^S113R/+ mice show several features of myelin and axonal degeneration. Selected degenerating features are shown in d–h. They include “onion bulbs” with degenerating and regenerating myelin, myelin outfolds, and axonal degeneration. R, Remak bundle. Animals were 10–12 months old when studied.

Neuronal loss and microglia activation in the spinal cords of AT-1^S113R/+ mice. A, NeuN staining in the ventral horn of the spinal cord showing neuronal loss in AT-1^S113R/+ mice. Representative images of the ventral horns of the lumbar section are shown. B, Computer-assisted quantification of NeuN(⁺) neurons per 100 μm² section areas (n = 3 animals; 7 sections per animal). Error bars indicate SD. **p < 0.005. C, Nissl staining showing neuronal loss in AT-1^S113R/+ mice. D, GFAP staining showing reactive astrocytes in AT-1^S113R/+ mice. E, Iba1 staining showing microglia activation in AT-1^S113R/+ mice. High-magnification insets show typical features of resting microglia in WT (left) and activated microglia in AT-1^S113R/+ (right) mice. F, Electron microscopy showing myelin and axonal degeneration in the white matter of AT-1^S113R/+ mice. Representative images of the lateral columns (lumbar section) are shown. a, WT; b–i, AT-1^S113R/+. Selected degenerating features are shown in b–i, including vacuolar myelinopathy, myelin outfolds, degenerating and regenerating myelin, and axonal degeneration. Animals were 10–12 months old when studied.

Loss of association fibers, gliosis, and microglia activation in the brains of AT-1^S113R/+ mice. A, B, MRI imaging showing reduced FA of the corpus callosum in AT-1^S113R/+ mice. Representative images are shown in A and FA changes are shown in B. Results are expressed as percentage of WT ± SD (n = 3 animals; 14 contiguous sections per animal). **p < 0.005. C, Thickness of the corpus callosum determined after Kluver–Barrera staining of consecutive coronal sections. Results are the average ± SD (n = 5 animals; 7 contiguous sections per animal). #p < 0.0005. D, Cresyl violet, H&E, and Iba1 staining showing gliosis and microglia activation in the dentate gyrus of AT-1^S113R/+ mice. E, [¹¹C]PBR28 PET imaging showing widespread microglia activation and inflammation of the brain in AT-1^S113R/+ mice. Values are the mean ± SD (n = 4). **p < 0.005. F, No overall difference can be seen in the clearance of [¹¹C]PBR28 from different brain regions. Animals were 10–12 months old when studied.

AT-1^S113R/+ mice display abnormal activation of autophagy-neuronal studies. A, EM of spinal cord neurons shows alterations that are consistent with abnormal autophagy. Asterisks indicate vacuolar material resembling lipofuscin; arrowhead indicates a large structure resembling a complex autophagosome. N, Nucleus. B, C, Similar EM alterations were observed in the cortex (B) and in culture (C). D, LC3β-positive neurons can be observed in the spinal cord of AT-1^S113R/+ mice. Arrowheads in the NeuN staining indicate two different neurons: one displays LC3β puncta and one does not. No LC3β-positive neurons were observed in WT mice. E, LC3β puncta in AT-1^S113R/+ cultured neurons.

AT-1^S113R/+ mice display abnormal activation of autophagy-MEF studies. A, EM of MEFs. a, MEFs from WT animals. b–f, MEFs from AT-1^S113R/+ mice. Different features that are consistent with abnormal activation of autophagy are evident in AT-1^S113R/+ MEF cells. A large autolysosome is evident in d, whereas cells undergoing autophagic (type II) cell death are shown in e and f. B, Immunoblots showing increased levels of autophagy markers beclin-1 and LC3β. Error bars indicate SD of n = 3 MEF lines. *p < 0.05. C, Immunoblots showing decreased levels of the autophagy “cargo” protein p62. Error bars indicate SD of n = 3 MEF lines. *p < 0.05. D, Immunoblots showing the autophagy flux in WT and AT-1^S113R/+ MEFs. BPE, bafilomycin (500 nm), pepstatin A (10 μg/ml), and E64 (10 μg/ml). E, LC3β-GFP activation and redistribution in AT-1^S113R/+ MEF cells. The increased induction of autophagy observed in AT-1^S113R/+ MEFs is rescued by expressing the dominant Atg9A_Gln mutant. Error bars indicate SD of n = 3 MEF lines. ***p < 0.0005. F, Immunoblots showing rescue of the autophagy phenotype after expression of Atg9A_Gln mutant. Lane 1, WT; lane 2, AT-1^S113R/+; and lane 3, AT-1^S113R/+ + Atg9A_Gln. G, Colocalization of Atg9A and LC3β in AT-1^S113R/+ MEFs.