The NR4A2/VGF pathway fuels inflammation-induced neurodegeneration via promoting neuronal glycolysis

A disturbed balance between excitation and inhibition (E/I balance) is increasingly recognized as a key driver of neurodegeneration in multiple sclerosis (MS), a chronic inflammatory disease of the central nervous system. To understand how chronic hyperexcitability contributes to neuronal loss in MS, we transcriptionally profiled neurons from mice lacking inhibitory metabotropic glutamate signaling with shifted E/I balance and increased vulnerability to inflammation-induced neurodegeneration. This revealed a prominent induction of the nuclear receptor NR4A2 in neurons. Mechanistically, NR4A2 increased susceptibility to excitotoxicity by stimulating continuous VGF secretion leading to glycolysis-dependent neuronal cell death. Extending these findings to people with MS (pwMS), we observed increased VGF levels in serum and brain biopsies. Notably, neuron-specific deletion of Vgf in a mouse model of MS ameliorated neurodegeneration. These findings underscore the detrimental effect of a persistent metabolic shift driven by excitatory activity as a fundamental mechanism in inflammation-induced neurodegeneration.


SUPPLEMENTAL METHODS
Calcium imaging.We seeded primary neuronal cultures on either Ibidi 60 µ-Dish Quad (cat.80411) or High (cat.81158) with a glass bottom.To measure cytosolic calcium changes, we transduced neuronal cultures with an AAV7 containing pAAV-Syn-GCamp6f-WPRE-SV4088 (Addgene 100837) at 8-12 d.i.v. with a 10,000-20,000 MOI.AAV particles were produced according to the standard procedures of the UKE vector facility.We acquired images with a confocal LSM 700 laser scanning confocal microscope (Zeiss) every 0.48 seconds with a 20× magnification in an imaging chamber maintaining 37 °C and 5% CO2.Infected cultures were imaged in the respective culture medium.In general, we recorded the first 5-10 minutes of baseline activity before applying the indicated chemicals.At the end of recording, we applied 10 µM ionomycin to induce maximum cellular calcium response that was used for normalization.Specific assay details and concentrations can be found in the respective figure legend.For data analysis, we measured mean fluorescence values of every cell using Fiji software (NIH) and normalized it to the maximal calcium response after ionomycin challenge (indicated as F/FMax).For each cell, we calculated maximal, minimal, mean and AUC of the calcium response using a custom R script.If not stated otherwise, AUC was used for statistical comparisons.
RNAscope in situ hybridization.We performed RNAscope fluorescent in situ hybridization (FISH) using the RNAscope Fluorescent Multiplex Kit V2 (Advanced Cell Diagnostics, cat.323100) according to the manufacturer's protocol.Probes against murine Mm-Snap25 (cat.516471) and Mm-Grm8-C2 (cat.521491-C2) were commercially available from Advanced Cell Diagnostics, Inc. RNAscope.We acquired single-plane confocal images with Zeiss LSM800 microscope, sampling spinal cord tissue at 63× magnification.We automatically stitched acquired tiles using ZEN imaging software (Zeiss).For each mouse Snap25+ neurons (n = 200-388) were evaluated for Grm8 signal.Grm8 + Snap25 + neurons were quantified in the ventral horns of cervical spinal cords by a blinded experimenter using Pannoramic Viewer software (3DHISTECH) and Fiji (NIH Image analysis), or with a custom-made script, which was based on Cognition Network Language (Definiens Cognition Network Technology; Definiens Developer XD software).
Human histopathology.The specimen used for histopathology were obtained from brain biopsies of pwMS, where MS-typical changes were described by board-certified histopathologists, and the patients had a MS diagnosis as determined by board-certified neurologists.The specimens were anonymized (cohort characteristics are provided in Supplemental Table 4).The use of patient specimens for research upon anonymization was in accordance with local ethical standards and regulations at the University Medical Center Hamburg-Eppendorf.We deparaffinized the sections using a standard ethanol/xylol dilution.After rinsing the deparaffinized sections in water we performed antigen retrieval using a sodium citrate buffer (pH = 6.5).After permeabilization with 0.1% Triton and blocking with 10% NDS immunolabeling was performed with antibodies against NeuN (1:200; Millipore, ABN91), VGF (Santa Cruz; 1:100, cat.sc-365397), and DAPI.Images were acquired using a confocal LSM 700 laser scanning confocal microscope (Zeiss).

Estimation of mitochondrial and glycolytic respiration.
We used an assay adapted from (2) to measure mitochondrial and glycolytic respiration.We measured ATP levels using the CellTiter-Glo Luminescent Cell Viability Assay (Promega, cat.G7570) which was performed according to manufacturer's protocol.We recorded luminescence with a Spark 10M multimode microplate reader (Tecan).Per plate we measured ATP levels in neurons without stimulation, stimulation with 10 µM oligomycin and 0.5 µM rotenone, or stimulation with 10 µM oligomycin and 0.5 µM rotenone and 5 mM 2-DG.The mitochondrial respiration was defined as difference between baseline ATP and ATP levels after respiratory chain blockade.The glycolytic respiration was defined as difference between ATP levels after respiratory chain blockade and ATP levels after respiratory chain and glycolysis blockade.At least four technical replicates were used per condition which were averaged per biological replicate for statistical analysis.

Chromatin immunoprecipitation (ChIP).
ChIP was performed using the SimpleChIP Enzymatic Chromatin IP Kit (Cell Signaling, cat.9003) according to the manufacturer's protocol with modifications.All buffers were prepared according to the manufacturer's protocol.We sacrificed three mice that were immediately perfused with 10 mL ice-cold PBS and the cortices were removed and stored on ice.The tissue was chopped to approximately 1 mm sized pieces and incubated with 1% formaldehyde and proteinase inhibitor cocktail (PIC) for 10 minutes on a thermomixer at 400 rpm.The cross-linking was stopped by adding glycine.Subsequently, the cross-linked tissue was washed twice with ice-cold PBS and PIC (centrifugation 1500 × g, 5 minutes, 4 °C) before being disintegrated into a single-cell suspension by 20 strokes of a dounce homogenizer.The nuclei were isolated by centrifugation with 1500 × g at 4 °C.The pellet was resuspended in buffer A (containing dithiothreitol and PIC) and was constantly inverted while being incubated for 10 minutes on ice.Subsequently, the suspension was centrifuged at 3000 × g for 5 minutes before 2 wash steps with buffer B (containing dithiothreitol and PIC).Subsequently, 0.5 µL micrococcal nuclease was added to the cell suspension and incubated for 20 minutes at 37 °C on a thermomixer shaking with 200 rpm until the digest was terminated by adding 10 µL 0.5 M EDTA.Subsequently, the nuclei were pelleted by centrifuging with 13,000 rpm for minute at 4°C.The pellet was resuspended in 100 µL ChIP buffer and incubated for 10 minutes on ice.Subsequently, the sample was sonicated using the Bioruptor 2000 (intensity: H, interval: 0.5, time: 8 minutes) before clarifying the lysate by centrifugation with 10,000 rpm for 10 minutes at 4 °C.For immunoprecipitation we added 5 µL antibody against NR4A2 (Merck; 1:500, cat.ABE1455) and as negative control 2 µg of the provided IgG and incubated the samples rotating overnight at 4 °C.Subsequently, the immunoprecipitated proteins were isolated using magnetic protein G beads according to the manufacturer's protocol.To remove the beads, the final pellets were solved in 150 µL ChIP elution buffer and incubated for 30 minutes on 65 °C on a thermomixer shaking with 400 rpm.Using a magnetic rack, the supernatant was removed and 6 µL 5 M NaCl and 2 µL proteinase K were added before incubating the sample for 2 hours at 65 °C to reverse the cross-linking.The remaining lysate was used to isolate DNA according to the manufacturer's protocol.We designed primers that amplified the flanking 125 bp of the consensus nuclear receptor binding motif AGGTCA (3) 1757 base pairs upstream of the transcription starting site of Vgf using the following primers: fwd -TACAGAATTTGCTGGAGATGGAATCC, rev -TTCCCCATGGACAAGAAGTGAAACA.Additionally, we performed PCR with primers that amplify a 125 bp segment in the 3' untranslated region of Vgf 157 bp downstream of the stop codon where NR4A2 binding is not expected, using the following primers: fwd -ACAAGGCCCCGCCTCGGG, rev -TCGCTAGAGAAACTAGATGGGGAC. We performed PCR from the respective immunoprecipitated DNA and a 5% input control.Abbreviations: Nt/Ct = N-terminus or C-terminus of the corresponding antigen sequence; h = human; m = mouse; CV1 = inter-assay coefficient of variability; CV2 = intra-assay coefficient of variability

Figure 1 .
NR4A2 is predominantly expressed in neurons in the CNS.(A) Representative gating scheme to isolate NeuN + nuclei by flow cytometry.(B) Principal component analysis of our RNA-sequencing results from cortical neuronal nuclei of WT and Grm8 -/-mice (n = 3 per group).(C) NR4A2 MFI in NeuN -and NeuN + cortical nuclei (n = 4 per group).(D) NR4A2 MFI in NeuN -, NeuN -NR4A2 high , and NeuN + cortical nuclei.Additionally, fluorescence minus one control (FMO) where the primary antibody against NR4A2 was not added were measured for NeuN -and NeuN + nuclei (n = 6 per group).(E) Nr4a2 mRNA expression of neuronal cultures that overexpress mScarlet (control) or Nr4a2 (n = 3 per group) in relative units (RU).(F) Immunoblot analysis of NR4A2 protein levels in neuronal cultures that overexpress mScarlet (control) or Nr4a2 (n = 3 per group).(G) Cumulative cytosolic calcium levels in mScarlet and Nr4a2 overexpressing neurons (n = 4).Points represent individual experiments, additionally mean is shown.If not stated otherwise, unpaired t-test with FDR-correction for multiple comparisons was used.*P < 0.05, ***P<0.001,****P < 0.0001.Supplemental Figure 2. VGF deficiency protects from glutamate excitotoxicity.(A) Enrichment analysis of inflamed neuronal signature (ranked gene list retrieved from (4)) in identified modules that contain Nr4a2.The normalized enrichment score (NES) is shown.All P values for enrichment for depicted modules < 0.05.(B) Representative images of mScarlet (controls), NR4A1-, and NR4A3-overexpressing neurons.Shown are VGF, NeuN and DAPI.Scale bar shows 20 µm.(C) Representative images of WT and Vgf -/-neurons.Scale bar shows 30 µm.(D) Baseline cell viability of WT and Vgf-deficient neurons (n = 5 per group).(E) Relative cell viability of WT and Vgfdeficient neurons after glutamate exposure (n = 5 per group) in relative units (RU).Points represent individual experiments, additionally mean is shown.If not stated otherwise, unpaired t-test with FDR-correction for multiple comparisons was used.**P < 0.01.Supplemental Figure 3. VGF peptides in the plasma of EAE animals and pwMS.(A) Absolute levels of VGF peptides GGGE Nt (pmol/mL), TLQP (pmol/mL), and AQEE (pmol/mL) in plasma of EAE healthy mice and EAE animals 10 days (onset), 15 days (acute), and 30 days (chronic) after immunization (n = 6 per group; only GGGE acute n = 5).(B) Relative fold change of VGF peptides N-terminus of VGF, GGEE N-terminus, NAPP, TLQP, and AQEE in sera of healthy individuals and pwRRMS (n = 20 per group).(C) Relative fold change of VGF peptides AQEE, NAPP, TLQP, and GGEE N-terminus, GAED C-terminus, TLQP, in sera of pwRRMS during stable disease and relapse (n = 9 per group).Paired t-test was used for statistical comparisons.(D) Relative fold change of VGF peptides AQEE, NAPP, TLQP, and GGEE N-terminus, GAED C-terminus, TLQP, in sera of pwRRMS before and after treatment with natalizumab (n = 6 per group).Paired t-test was used for statistical comparisons.Points represent individual experiments, additionally mean is shown.If not stated otherwise, unpaired t-test with FDR-correction for multiple comparisons was used.*P < 0.05.Supplemental Figure 4. Neurotrophic factors are not differently expressed in VGF-overexpressing neurons.(A) Baseline viability of control (GFP-) and VGF-overexpressing neurons (n = 6 per group).(B) Relative cell viability of EGFP (control) and Vgf-overexpressing neurons that were exposed to 50 µM glutamate for 15 hours.If indicated the medium of EGFP-overexpressing neurons was replaced with the medium of Vgf-overexpressing neurons and vice versa every other day (n = 8 per group).(C and D) Normalized expression counts in relative units (RU) of Bdnf, Ngf (C), Apoe, Cst3, and Clu (D) from our RNA-sequencing experiment of control (GFP-) and VGF-overexpressing neurons (n = 6 per group).Statistics were taken from differential expression analysis with FDR-correction for multiple comparisons.(E) Baseline ATP levels and mitochondrial respiration in NR4A2-overexpressing WT and Vgf-deficient neurons (n = 5 per group).(F and G) Mitochondrial respiration and glycolytic index in mScarlet-(control) and NR4A1-(F; n = 5 per group) or NR4A3-overexpressing neurons (G; n = 5 per group).(H) Baseline ATP levels and mitochondrial respiration in WT and Vgf-deficient neurons (n = 5 per group).(I) Relative cell viability (relative units = RU) of WT neurons that were exposed to glutamate and were pretreated with vehicle or 5 mM 2-DG (n = 5 per group).Points represent individual experiments, additionally mean is shown.If not stated otherwise, unpaired t-test with FDRcorrection for multiple comparisons was used.*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.Supplemental Figure 5. Microglia activation is not different in Vgf-deficient and control EAE mice.(A) Clinical score at 40 days after immunization of WT (n = 11) and Vgf-cKO (n = 10) mice that underwent EAE.Kolmogorov-Smirnov test was used for statistical comparisons.(B to E) Relative fold change of blood levels of total VGF (B), GGEE (C), TLQP (D), and AQEE (E) in WT (n = 10), and Vgf-cKO (n = 10) EAE mice 40 day after immunization.Data were normalized to the average value of the WT animals.(F) Immunoblot analysis of brains of WT (n = 7), and Vgf-cKO (n = 7) EAE mice 40 days after immunization.The VGF protein levels were normalized to vinculin as loading control.(G) Number of NfL DegenoTag positive axonal puncta per 100 µm 2 in the dorsal columns of the cervical spinal cords of WT (n = 5) and Vgf-cKO (n = 8) EAE mice 40 days after immunization.Scale bar shows 10 µm.(H) Area covered by GFAP in Vgf-cKO (n = 9) and control mice (n = 10) undergoing EAE 40 days after immunization.Representative images of GFAP are shown in Figure 5F.(I) IBA1 + covered area in the ventral horns of cervical spinal cords of WT (n = 6) and Vgf-cKO (n = 10) EAE mice 40 days after immunization.Scale bar shows 50 µm.(J) IBA1 + covered area in the white matter of cervical spinal cords of WT (n = 8) and Vgf-cKO (n = 10) EAE mice 40 days after immunization.Scale bar shows 50 µm.Points represent individual experiments, additionally mean is shown.If not stated otherwise unpaired ttest was used for statistical comparisons.*P < 0.05.