The Mayo Clinic Florida Brain Bank provided post-mortem brain tissue from individuals with confirmed FTD with TDP-43 inclusions and individuals without neuropathological features. We obtained written informed consent from all participants or their family members. The Mayo Clinic Institutional Review Board and Ethics Committee approved all protocols associated with the collection of human tissues. The sample size was determined based on the availability of tissue. A description of patient demographic information for both control individuals and FTD cases can be found in Supplementary Table S1.

All animal procedures were conducted in accordance with the National Institutes of Health Guide for Care and Use of Experimental Animals and were approved by both the Mayo Clinic Institutional Animal Care and Use Committee (IACUC Protocol number A00005405-20) and the National Eye Institute ACUS committee (ASP number NEI-626).

The CRISPRi knockin mouse line used in this study was created at the Genetic Engineering Core, National Eye Institute, National Institutes of Health. The system uses a human codon-optimized dCas9 from Streptococcus pyogenes, fused with a nuclear localization sequence (NLS, 2X), an HA tag, and the coding sequence of blue fluorescent protein (BFP). The modified dCas9 is fused with the KRAB (Krüppel-associated box) domain of the human KOX1 gene at its C-terminus, which facilitates the recruitment of chromatin-modifying factors to the targeted gene locus (Margolin et al., 1994).

To create the CRISPRi knockin mouse line, a vector was designed to target the Rosa26 locus of the previously constructed mouse genome (Liu et al., 2019). The CRISPRi cassette containing the essential elements of the system was obtained by PCR amplification from pHF-SFFV-dCas9-BFP-KRAB, a gift from Stanley Qi and Jonathan Weissman (Addgene plasmid #46911) (Gilbert et al., 2014). The cassette was then inserted between the Cytomegalovirus early enhancer/chicken β actin (CAG) promoter and the rabbit beta-globin polyadenylation (RBG-PA) signal sequence in the mouse Rosa26 targeting vector, which contains a phosphoglycerate kinase-neomycin (PGK-Neo) cassette flanked by two FRT sites. To ensure insulation of the transgene, the CAG-CRISPRi-RBG-PA cassette was flanked by two copies of the 5′-HS4 chicken β-globin (cHS4) insulator (Chung et al., 1993).

To introduce the CRISPRi knockin allele at the Rosa26 locus, the targeting vector was linearized using AscI and then electroporated into mouse W4 embryonic stem (ES) cells as described (Bradley et al., 1984). Homologous recombination (HR) events were selected using neomycin (neo) selection followed by HR screening using a long genomic PCR assay. Two PCR assays were performed: one with forward primer 5′-CGG​CAC​TAC​TGT​GTT​GGC​GGA​CTG​GCG​GGA​C-3′ and reverse primer 5′-GAA​CAT​ACG​TCA​TTA​TTG​ACG​TCA​ATG-3′ to detect the left arm HR (2.85 kb PCR product) and another with forward primer 5′-GGT​GTG​GCG​GAC​CGC​TAT​CAG​GAC​ATA​GC-3′ and reverse primer 5′-CAT​GTT​TAG​TAA​TGG​CTG​CAG​ACT-3″ to detect the right arm HR (2.145 kb PCR product). ES clones carrying the CRISPRi knockin allele at the Rosa26 locus were then expanded and microinjected into blastocyst embryos obtained from C57BL6/J female donors. Male chimeric F0 pups with 90% or higher agouti coat color were crossed with wildtype C57Bl6/J females to achieve germline transmission of the knockin allele. F1 pups carrying the germline-transmitted knockin allele were bred with wild-type C57Bl6/J mice for several generations until they became congenic with the C57Bl6/J background. Finally, the neo cassette was removed by crossing the CRISPRi knockin mice with a germline flippase (FLP) recombination strain (actin-FLP, Jackson lab).

To produce AAV expressing sgRNA and sgRNA plus a C9ORF72 G₄G₂ repeat, the coding sequence for the mouse U6 promoter, including modified stem loop 1, 2 and guide sequence site from pU6-sgRNA EF1Alpha-puro-T2A-BFP, a gift from Jonathan Weismann and Martin Kampmann (Addgene plasmid #60955) (Gilbert et al., 2014), were cloned into the pAAV CBA EGFP backbone and pAAV CBA C9ORF72 with 2 or 149 G₄C₂ repeats (referred to as 2R and 149R, respectively) (Chew et al., 2015). To generate AAV sgStmn2 we selected four sgRNA sequences against Stmn2, as described previously (Horlbeck et al., 2016). Top and bottom oligonucleotides were annealed and ligated into BstXI and BlpI restriction sites within the AAV sgRNA or sgRNA-2R/149R backbone. Sequences for sgRNA are as follow sgStmn2#1: 5′-GGC​CGC​GGA​GAG​AGC​GAG​AG-3′, sgStmn2#2: 5′-GGG​AGT​GGT​GGC​GAA​GGC​AA-3′, sgStmn2#3: 5′-GGG​GAT​GTG​CAC​GCA​CGG​AG-3′, sgStmn2#4: 5′-GCG​AGA​GAG​GAG​TCT​GAG​TA-3′ and sgCtrl: 5′-GGACTAAGCGCAAGCACCTA-3′.

To produce AAV9 virus we followed a previously described protocol (Chew et al., 2015). Briefly, AAV vectors were co-transfected with helper plasmids in HEK293T cells. After 72 h, cells were harvested, treated with 50 Units/mL Benzonase (Sigma-Aldrich), and lysed by several cycles of freeze-thaw. The virus was purified using a discontinuous iodixanol gradient and diluted in PBS.

Intracerebroventricular injections (ICV) of AAV were performed as previously described with some minor modifications (Chew et al., 2015). In a few words, a 30-gauge needle attached to a 10 μL syringe (Hamilton Company) was inserted into the lateral ventricle of cryo anesthetized CRISPRi pups at postnatal day 0 (P0). Two microliters (1.5E10 genomes/μL) of AAV9 virus were manually injected into each ventricle. Following injection, pups were placed on a heated pad until they recovered from anesthesia, at which time they were placed back into their home cage. Approximately three-to-four-week later pups were weaned, DNA was extracted from tails and genotyped for the presence of CRISPRi transgene using the following primer set: dCas9 forward: 5′- ATC​AAC​CGG​CTG​TCC​GAC​TA-3′, dCas9 reverse: 5′-CCT​CTG​GGT​AAT​CAG​CTT​GG-3′.

Mice were euthanized by CO₂ and brains were harvested and cut sagittally across the midline. Sagittal half brains were fixed in 10% formalin, embedded in paraffin, sectioned, and mounted on glass slides. The other half brain was dissected and frozen on dry ice. Frozen mouse forebrain tissues were homogenized in ice-cold Tris-EDTA (TE; 50 mM Tris, pH 8.0, 50 mM NaCl, 1 mM EDTA), with 2× protease and phosphatase inhibitors.

To extract total RNA, brain homogenates were mixed with Trizol LS (Thermo Fisher Scientific) and frozen. The following day, RNA was extracted using the Direct-zol RNA MiniPrep kit (Zymo Research). cDNA was synthesized from 250 ng of RNA using the High-Capacity cDNA transcription kit (Applied Biosystems). Quantitative reverse transcription-PCR (qRT-PCR) was performed using the SYBR GreenER qPCR SuperMix (Invitrogen, Thermo Fisher Scientific). Each sample was run in triplicate and processed on a Prism 7900HT Fast Real-Time system (Applied Biosystems). Samples were analyzed using the ΔΔCt method and normalized to the endogenous controls (Rplp0 and Gapdh). The following primers and their sequences were used:

Stmn1 forward: 5′-TCT​GTC​CCC​GAT​TTC​CCC​C-3′, Stmn1 reverse: 5′-AGC​TGC​TTC​AAG​ACT​TCC​GC-3’.

Stmn2 forward: 5′-GCA​ATG​GCC​TAC​AAG​GAA​AA-3′, Stmn2 reverse: 5′-GGT​GGC​TTC​AAG​ATC​AGC​TC-3′.

Stmn3 forward: 5′-CCA​GCA​CCG​TAT​CTG​CCT​AC-3′, Stmn3 reverse 5′- CAG​AGG​CCC​GCT​TAT​CCA​G-3’.

Stmn4 forward: 5′-GCG​TCA​TCT​CTG​ATA​TGG​AAG​TC-3′, Stmn4 reverse: 5′- CTC​TTC​TAA​TGA​TGG​GTC​TCG​C-3’.

Gapdh forward: 5′-CATGGCCTTCCGTGTTCCTA-3′, Gapdh reverse: 5′-CCTGCTTCACCACCTTCTTGAT-3′.

Rplp0 forward: 5′-ACTGGTCTAGGACCCGAGAAG-3′, Rplp0 reverse: 5′-CTCCCACCTTGTCTCCAGTC-3′.

Iba1/Aif1 forward: 5′-GGA​TTT​GCA​GGG​AGG​AAA​AG-3′, Iba1/Aif1 reverse: 5′-TGG​GAT​CAT​CGA​GGA​ATT​G-3′.

Cd68 forward: 5′-GAC​ACT​TCG​GGC​CAT​GTT-3′, Cd68 reverse: 5′-GAG​GAG​GAC​CAG​GCC​AAT-3.’

Gfap forward: 5′-GGA​GAG​GGA​CAA​CTT​TGC​AC-3′, Gfap reverse: 5′-AGC​CTC​AGG​TTG​GTT​TCA​TC-3′.

C9orf72 G ₄ C ₂ forward: 5′-TAGTACTCGCTGAGGGTGAAC-3′, C9orf72 G ₄ C ₂ reverse: 5′-CTA​CAG​GCT​GCG​GTT​GTT​TC-3′.

Tardbp forward: 5′-GGG​GCA​ATC​TGG​TAT​ATG​TTG-3′, Tardbp reverse: 5′-TGG​ACT​GCT​CTT​TTC​ACT​TTC​A-3’.

To prepare protein lysates from mouse brains, Triton X-100 and Sodium dodecyl sulfate (SDS) were added to brain homogenates at final concentrations of 1% and 2%, respectively. Homogenates were sonicated on ice and centrifuged at ×16,000 g for 20 min. The supernatants were collected, and protein concentrations were determined by a bicinchoninic acid (BCA) assay (ThermoFisher Scientific). For human post-mortem frontal cortex samples, 50 mg of tissue was homogenized in Radioimmunoprecipitation buffer (RIPA, 25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% sodium deoxycholate, 1% Nonidet P-40, 0.1% sodium dodecyl sulfate, protease, and phosphatase inhibitors). Tissue homogenates were sonicated on ice and centrifuged at ×100,000 g for 30 min at 4°C. The supernatant was collected and is referred to as the RIPA-soluble fraction, with protein concentration determined by BCA assay. The RIPA-insoluble pellet was extracted using 7 M urea buffer, sonicated, and then centrifuged for 30 min at ×100, 000 g at room temperature. Protein concentrations of the urea-soluble fractions were determined by Bradford assay.

Protein lysates were diluted in Tris-Glycine SDS sample buffer (ThermoFisher Scientific) and 5% beta-mercaptoethanol (Sigma Aldrich). Protein samples were heat denatured for 5 min at 95°C and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) on 4%–20% Tris-Glycine gels (Life Technologies). Protein was subsequently transferred to Polyvinylidene (PVDF) membrane (Millipore). Membranes were first blocked with 5% non-fat dry milk in Tris-buffered saline with Triton (TBS-T, 100 mM Tris-HCl pH 7.5, 140 mM NaCl, 0.1% Triton X-100) and incubated overnight at 4°C with mouse anti-STMN2 (1:1000 MAB6930, R&D Systems), rabbit anti-TDP-43 (1:5,000 12892-AP ProteinTech) and mouse monoclonal anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (1:30,000, H86504M, Meridian Life Sciences), diluted in TBS-T plus 5% nonfat dry milk. Membranes were then incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies (1:5,000; Jackson ImmunoResearch) and signal was detected by Enhanced chemiluminescence (ECL, PerkinElmer). Quantitative densitometry was performed using ImageJ, and Stmn2 levels were normalized to Gapdh expression.

Paraffin-embedded brains were deparaffinized in xylene and rehydrated through a series of ethanol solutions. Antigen retrieval was performed by steaming in sodium citrate buffer (10 mM sodium citrate, 0.05% Tween-20, pH 6), distilled water, or pH 9 Tris-EDTA for 30 min. Tissues were immunostained with polyclonal antibodies against poly(GP) [1:10,000, Rb5823 (Gendron et al., 2013)] and Gfap (1:2,500, BioGenex). For phosphorylated Tdp-43 (pTdp-43) staining, tissue sections were incubated with Dual Endogenous Enzyme Block (DAKO) and blocked with 2% normal goat serum. Slides were incubated with our novel MA1290/1291AP in-house antibody (1:100), an anti-rabbit antibody for both mouse and human TDP-43 phosphorylated at serine 409/410, overnight at 4°C. The recognized antigen is amino acid residues 404 to 414 of full-length TDP-43 with phosphorylation of S409/410 (CSMDSK[pS][pS]GWGM), and equivalent to our previously described Rb3655 antibody (Todd et al., 2020; Wu et al., 2021). The following day, slides were washed and incubated with biotinylated goat anti-rabbit secondary (1:200, Vectastain^® ABC-HRP Kit, Peroxidase (Rabbit IgG) PK-4001) for 2 h at room temperature. Slides were incubated with avidin-biotin complex solution for 30 min. Slides were then incubated with 3,3′-diaminobenzidine (DAB, Acros Organics) activated with hydrogen peroxide, and the reaction was stopped by rinsing slides in distilled water. Sections were counterstained with hematoxylin, dehydrated through a series of ethanol and xylene washes, and coverslipped with Cytoseal mounting media (Thermo Fisher Scientific).

The burden of GP inclusions and Gfap immunoreactivity was quantified using digital pathology technology by Leica Biosystems. We used a ScanScope^® AT2 (Leica Biosystems) to obtain high-resolution digitized images of immunostained slides. ImageScope^® software (v12.1; Leica Biosystems) was used to annotate the cortex on mid-sagittal serial sections stained for GP of Gfap for each mouse. A custom-designed Positive Pixel Count algorithm was used to quantify either poly-GP burden or Gfap immunoreactivity in the cortex. pTdp-43 inclusions in the cortex of each mouse sample were counted manually in a blinded fashion, with the number of pTdp-43 inclusions per mm² determined for each slide.

Previously characterized Meso Scale Discovery (MSD) sandwich immunoassays were used to quantify poly (GP) and poly (GA) (Gendron et al., 2015) from mouse brains. The abundance of pTDP-43 in the urea-soluble fraction from the frontal cortex of human cases and controls had been measured in a previous study by sandwich Meso Scale Discovery (MSD) immunoassay (Prudencio et al., 2020). Lysates were diluted in Tris buffered saline (TBS) and an equal amount of protein was loaded in duplicate wells. Response values corresponding to the intensity of emitted light upon electrochemical stimulation of the assay plate using the MSD QUICKPLEX SQ120 were acquired.

GraphPad Prism 9 (GraphPad software) was used for all statistical analyses. The type of statistical analysis used, the number of subjects, and the p values used are indicated in the figure and/or legend. To compare STMN2 protein abundance in FTD and controls, the data was analyzed with a single-variable (unadjusted) and a multivariable linear regression model (adjusted), considering age at death and sex. p-values less than 0.0125 were considered statistically significant, after comparing the four groups, all FTD combined and each genotype separately. To compare pTDP-43 abundance between FTD groups, that data was analyzed using both single-variable (unadjusted) and multivariable linear regression (adjusted) models, factoring in age at death and sex. p-values less than 0.025 were considered statistically significant after multiple comparisons. Associations between STMN2 protein and TDP-43 or pTDP-43 protein levels in FTD cases were determined using a single-variable and multivariable linear regression model, the latter was adjusted for age at death and sex. p-values less than 0.025 were considered statistically significant after multiple comparisons. Due to their skewed distribution, STMN2, TDP-43 and pTDP-43 protein levels were plotted on a base 2 logarithmic scale.

Previous studies demonstrated a reduction of STMN2 RNA in FTD and ALS cases (Klim et al., 2019; Melamed et al., 2019; Prudencio et al., 2020) and a decrease in STMN2 immunoreactivity in motor neurons from ALS cases compared to controls (Klim et al., 2019). However, STMN2 protein levels in FTD cases have yet to be evaluated. To do so, we assembled an FTD cohort with a neuropathological diagnosis of frontotemporal lobar degeneration with TDP-43 pathology (FTLD-TDP), with an approximately equal number of cases having an expanded G₄C₂ repeat in the C9ORF72 gene, GRN mutations or no known FTD-causative mutation (Supplementary Table S1). We then measured detergent-soluble STMN2 protein in the frontal cortex by Western blot, which revealed a significant or nominally significant reduction of STMN2 protein in all FTLD-TDP cases compared to controls in unadjusted analysis and adjusted analysis when controlling for age and sex, respectively (Figure 1A; Supplementary Table S2). Lower levels of STMN2 protein in the FTLD-TDP group were largely driven by carriers of a C9ORF72 expanded repeat and GRN mutations, as individuals without a mutation had similar levels of STMN2 protein compared to controls. We speculated that cases harboring mutations may have greater reductions in STMN2 protein due to lower levels of soluble TDP-43 protein. To that end, we measured soluble TDP-43 levels in our cohort of controls and FTLD-TDP cases. Compared to controls, we observed a significant or nominally significant reduction in TDP-43 levels in unadjusted and analysis adjusted for age and sex, respectively (Figure 1B; Supplementary Table S2). Like STMN2 levels, the reduction in soluble TDP-43 protein was more pronounced in mutation carriers. Additionally, lower levels of TDP-43 protein were significantly associated with reduced levels of soluble STMN2 in FTLD-TDP cases in both unadjusted and adjusted analyses controlled for age and sex (Supplementary Table S3). We further suspected that levels of pTDP-43 pathology may also be greater in mutation carriers. To evaluate this possibility, we used pTDP-43 protein levels we had measured in these same cases from a previous study (Prudencio et al., 2020) and compared the pTDP-43 burden in sporadic and familial FTLD-TDP cases. Indeed, mutation carriers had significantly elevated levels of pTDP-43 than those without, in both unadjusted and adjusted analysis (Figure 1C; Supplementary Table S4), in line with our previous findings (Prudencio et al., 2017). Higher levels of pTDP-43 were found to have a significant or nominally significant association with lower levels of STMN2 protein in FTLD-TDP cases (Supplementary Table S3). Taken together, we provide evidence of decreased STMN2 and TDP-43 abundance in FTLD-TDP cases harboring C9ORF72 and GRN mutations. These findings suggest that loss of STMN2 protein is a feature of familial FTD and may contribute to pathogenesis.

To model Stmn2 loss of function in vivo and assess its relevance in FTD, we first generated a novel knockin CRISPRi mouse model using a system that previously demonstrated effective silencing of target genes in vitro (Gilbert et al., 2014). The CRISPRi machinery was integrated into the Rosa26 locus and expresses a catalytically dead form of Cas9 (dCas9) fused to a transcriptional repressor domain, KRAB (Figure 2A). AAV was used to deliver sgRNA to direct CRISPRi machinery to the Stmn2 transcription start site (Figure 2B). We tested four sgRNAs targeted against Stmn2 from a CRISPRi library of the whole mouse genome (Horlbeck et al., 2016). To deliver guides to the mouse brain, ICV injections of AAV sgStmn2 (#1–4) were performed at postnatal day zero, and mice were harvested at 2 months of age (Figure 2B). Notably, all four guides were able to significantly reduce Stmn2 RNA (Figure 2C) and protein (Figures 2D, E) compared to CRISPRi mice injected with enhanced Green Fluorescent Protein (eGFP) or non-transgenic mice (NT, lacking the CRISPRi machinery). While Stmn2 protein knockdown ranged from 40% to 75%, mice injected with sgStmn2#1 displayed the largest reduction in Stmn2 protein (Figures 2D, E). We therefore selected sgStmn2#1 for further experiments and will refer to it as sgStmn2 hereafter. We were also able to modulate the levels of Stmn2 knockdown by titrating the amount of virus delivered, with the highest levels of virus showing the largest decrease in both Stmn2 RNA (Supplementary Figure S1A) and protein levels (Supplementary Figures S1B, C). Taken together, we were able to achieve robust knockdown of Stmn2 in the brain of CRISPRi mice.

Our group has previously shown that in comparison to mice expressing a non-pathological number of 2 G₄C₂ repeats (2R), mice expressing 149 G₄C₂ repeats (149R) recapitulate C9FTD relevant disease features including pTdp-43 inclusions, accumulation of RNA foci and DPR proteins from both sense and antisense transcripts, neurodegeneration and behavioral deficits (Chew et al., 2015; Chew et al., 2019). To understand if Stmn2 deficiency would impact C9ORF72-associated pathology we used AAV to simultaneously express either sgStmn2 or a non-targeting guide RNA (sgCtrl) along with either 2R or 149R.

Each AAV (sgCtrl-2R, sgCtrl-149R, sgStmn2-2R, and sgStmn2-149R) was introduced to the mouse CNS via ICV injection at postnatal day 0 (Figure 3A). Animals were harvested at 6 months of age, a time point when we previously observed robust pathology and gliosis in the 149R-AAV model (Chew et al., 2019). We first confirmed the expression of the G₄C₂ repeat transcript, with similar levels observed between all four groups (Figure 3B). Moreover, as anticipated, Stmn2 RNA levels were significantly reduced in mice injected with sgStmn2 compared to mice injected with the non-targeting sgCtrl (Figure 3C). Stmn2 protein levels were also significantly reduced in sgStmn2-2R and sgStmn2-149R mice (Figures 3D, E). Of note, we did not observe any differences in Stmn2 knockdown efficiency as a function of the G₄C₂ repeat number. We also measured transcripts of other stathmin family genes to both verify the specificity of the guide RNA for Stmn2 and exclude any possible compensatory upregulation mechanisms (Gagliardi et al., 2022). Analysis of Stmn1, 3, or 4 transcript levels revealed no significant differences among any of the four groups (Supplementary Figures S2A–C).

To validate whether our CRISPRi dual model replicated C9ORF72 expanded repeat-associated pathology, we assessed DPR protein expression and pTdp-43 accumulation. First, we measured the expression of DPR proteins in the cortex using established immunoassays (Chew et al., 2019). As anticipated, we detected robust expression of poly(GP) (Figure 4A) and poly(GA) (Figure 4B) in sgCtrl-149R and sgStmn2-149R mice compared to animals not expressing the expanded repeat. Reduction of Stmn2 protein levels did not impact either poly(GA) or poly (GP) protein levels. Immunohistochemical analysis using a poly(GP)-specific antibody revealed both diffuse nuclear staining in the cortex as well as cytoplasmic inclusions that were restricted to animals expressing the expanded repeat (149R) (Figures 4C, D). The number of GP inclusions was similar between sgCtrl-149R and sgStmn2-149R animals consistent with our immunoassay results.

We next evaluated pTdp-43 inclusions and observed small cytoplasmic inclusions in the cortex of sgCtrl-149R mice that were absent in sgCtrl-2R mice (Figures 4E, F). Unexpectedly, we found fewer pTdp-43 inclusions in sgStmn2-149R mice compared to sgCtrl-149R mice (Figure 4F) in the absence of changes in Tardbp RNA levels (Supplementary Figure S3A).

We also assessed gliosis in the cortex and observed a significant increase in Gfap immunoreactivity, a marker of activated astrocytes, (Figures 4G, H), in mice expressing the C9ORF72 expanded repeat, in line with our previous mouse model (Chew et al., 2019). Levels of the Gfap transcript were also significantly elevated in the 149R groups (Supplementary Figure S4A). Additionally, RNA levels of markers of activated microglia, (Iba1/Aif1, Supplementary Figure S4B), and phagocytic microglia (Cd68, Supplementary Figure S4C) were significantly increased in groups expressing the expanded repeat, suggestive of microgliosis, although this was not confirmed at the protein level. Reduction of Stmn2 protein did not consistently induce or exacerbate gliosis. However, we did observe a significant increase in Cd68 transcript levels in the sgStmn2-2R group compared to sgCtrl-2R (Supplementary Figure S4C). A significant increase of Gfap RNA levels was observed in Stmn2 depleted mice (sgStmn2-149R) compared to controls (sgCtrl-149R) in animals expressing the expanded repeat (Supplementary Figure S4A), although this was not preserved at the protein level (Figure 4H). Together, these data provide proof-of-concept that we were able to successfully deliver guide RNA to knock down a target of interest in CRISPRi mice in a disease model (in this case mice expressing the expanded C9ORF72 repeat). Overall, loss of Stmn2 in the presence of the C9ORF72 expanded repeat does not exacerbate C9-mediated pathology at 6 months of age in the current model.