Variant in NHLRC2 leads to increased hnRNP C2 in developing neurons and the hippocampus of a mouse model of FINCA disease

FINCA disease is a pediatric cerebropulmonary disease caused by variants in the NHL repeat-containing 2 (NHLRC2) gene. Neurological symptoms are among the first manifestations of FINCA disease, but the consequences of NHLRC2 deficiency in the central nervous system are currently unexplored. The orthologous mouse gene is essential for development, and its complete loss leads to early embryonic lethality. In the current study, we used CRISPR/Cas9 to generate an Nhlrc2 knockin (KI) mouse line, harboring the FINCA patient missense mutation (c.442G > T, p.Asp148Tyr). A FINCA mouse model, resembling the compound heterozygote genotype of FINCA patients, was obtained by crossing the KI and Nhlrc2 knockout mouse lines. To reveal NHLRC2-interacting proteins in developing neurons, we compared cortical neuronal precursor cells of E13.5 FINCA and wild-type mouse embryos by two-dimensional difference gel electrophoresis. Despite the significant decrease in NHLRC2, the mice did not develop severe early onset multiorgan disease in either sex. We discovered 19 altered proteins in FINCA neuronal precursor cells; several of which are involved in vesicular transport pathways and actin dynamics which have been previously reported in other cell types including human to have an association with dysfunctional NHLRC2. Interestingly, isoform C2 of hnRNP C1/C2 was significantly increased in both developing neurons and the hippocampus of adult female FINCA mice, connecting NHLRC2 dysfunction with accumulation of RNA binding protein. We describe here the first NHLRC2-deficient mouse model to overcome embryonic lethality, enabling further studies on predisposing and causative mechanisms behind FINCA disease. Our novel findings suggest that disrupted RNA metabolism may contribute to the neurodegeneration observed in FINCA patients.


Background
FINCA disease is a progressive cerebropulmonary disease (OMIM #618278) presenting with severe tissue fibrosis, neurodegeneration, and cerebral angiomatosis caused by pathogenic variants of the NHL Repeat Containing 2 (NHLRC2) gene (Uusimaa et al. 2018;Brodsky et al. 2020). Although FINCA patients have multi-organ manifestations including chronic hemolytic anemia, neurological symptoms were among the first to appear, by 2 months of age. Neuropathology of FINCA patients identified brain atrophy, vacuolar neurodegeneration, myelin loss with gliosis, cerebral angiomatosis, and neuronal depletion of the anterior horns of the spinal cord  (Uusimaa et al. 2018). NHLRC2 mRNA is present in several cell types and regions of the human and mouse brain (Zhang et al. ,2016. In addition, altered NHLRC2 and NHLRC2 mRNA levels have been detected in neurodegenerative diseases such as Parkinson's disease (PD) (Dijk et al. 2012) and Alzheimer's disease (AD) (Long et al. 2016). A wide range of neural tube-related developmental malformations called developmental duplications have been reported in Angus cattle that are homozygous for p.Val311Ala substitution in the beta-propeller domain of NHLRC2 (Denholm 2017). Despite accumulating evidence for the importance of NHLRC2 in the central nervous system, its function in neurons is currently unknown.

Molecular Medicine
NHLRC2 consists of an N-terminal thioredoxin (Trx)like domain, a six-bladed β-propeller domain, and a C-terminal β-stranded region (Biterova et al. 2018). Structural analysis of the protein has revealed a highly conserved cleft between the Trx-like and β-propeller domains that forms a possible binding site for currently unknown substrates or interaction partners (Biterova et al. 2018). Recent in vitro studies have shed some light on the possible functions of NHLRC2. FINCA patientderived and immortalized skin fibroblasts show enhanced differentiation to myofibroblasts and variants in NHLRC2 were found to affect the cytoskeleton organization and vesicle transport in normal human dermal fibroblasts . In macrophages, NHLRC2 was discovered to be a novel regulator of phagocytosis in two genome-wide knockout (KO) screens (Haney et al. 2018;Yeung et al. 2019), and has been proposed to affect phagocytosis via its effect on actin dynamics through RhoA-Rac1 signalling (Haney et al. 2018). In colon cancer cells, loss of NHLRC2 was found to increase the susceptibility of these cells to apoptosis induced by reactive oxygen species (ROS) (Nishi et al. 2017). However, the physiological function of NHLRC2 remains elusive.
The mouse ortholog of human NHLRC2 has 84% protein sequence similarity (Uusimaa et al. 2018), and it has been proposed that the function of NHLRC2 is conserved across species (Biterova et al. 2018). The complete loss of Nhlrc2 leads to early embryonic lethality, highlighting its essential role in embryonic development (Uusimaa et al. 2018;Perez-Garcia et al. 2018;Delhotal 2016). X-gal staining has revealed widespread expression of Nhlrc2 in a variety of organs during embryonic development (Uusimaa et al. 2018). Nhlrc2 mRNA and NHLRC2 mRNA expression is especially high during early brain development in mice and humans respectively (Uusimaa et al. 2018;Cardoso-Moreira et al. 2019;Miller et al. 2014), and in situ hybridization of embryonic day 14.5 mice shows high expression in the ventricular layer of the telencephalon (Diez-Roux et al. 2011). Nhlrc2 has been detected in the transcriptional waves directing the differentiation of new born neurons in the neocortex (Telley et al. 2016). To date, molecular mechanisms related to Nhlrc2 during brain development remain unexplored.
We describe here a novel FINCA knockin (KI) mouse line, generated by editing the mouse endogenous Nhlrc2 gene to include the missense mutation identified in Finnish FINCA patients. This NHLRC2 deficient mouse escapes the embryonic lethality that has previously prevented further in vivo studies of a fully null KO mouse line. We compared the proteomes of FINCA and wildtype mouse embryonic neuronal precursor cells (NPCs), to elucidate the effect of NHLRC2 deficiency on developing neurons of the neocortex. The results obtained from our studies contribute to an understanding of the pathological mechanisms leading to neurodegeneration in FINCA disease.

Materials and methods
A detailed description of the materials and methods is found in the Additional File 1. Unedited full images of the immunoblots are presented in Additional File 2.

Results
The first mouse model for FINCA has significantly decreased NHLRC2 protein levels but normal tissue histology FINCA disease is caused by pathogenic variants in Nhlrc2. The mutation site and its adjacent area are highly conserved between human and mouse (Fig. 1a). We generated a mouse line with the FINCA patient variant c.442G > T by editing the mouse endogenous Nhlrc2 gene using the CRISPR/Cas9 method (Cong et al. 2013;Mali et al. 2013;Inui et al. 2014) (Fig. 1a, Additional File 1: Fig. S1). The heterozygous FINCA (hereafter Nhlrc2 FINCA/+ ) and heterozygous Nhlrc2 KO, C57BL/6N-A tm1Brd Nhlrc2 tm1a(KOMP)Wtsi /WtsiOulu (hereafter Nhlrc2 −/+ ) (Skarnes et al. 2011) mice were crossed to obtain mice (Nhlrc2 FINCA/− ) mimicking the genotype of compound heterozygote FINCA patients with the missense variant and a frameshifting nonsense variant on the other allele (Uusimaa et al. 2018).
Compound heterozygous Nhlrc2 FINCA/− mice (6 males and 6 females) and their wild-type Nhlrc2 +/+ litter mates (6 males and 5 females) were observed for 32 weeks. Nhlrc2 FINCA/− mice appeared normal at birth, gained weight comparably to wild-type litter mates (Fig. 1b), and reproduced normally (Additional File 1: Table S3). Nhlrc2 FINCA/− mice did not develop clinical signs during the observation period, and they were euthanized for tissue-based evaluation of FINCA disease-like phenotype. There was no brain atrophy or abnormal histology identified in hippocampal and cerebellar sections of Nhlrc2 FINCA/− mice (Fig. 3). Lung and liver fibrosis and Fig. 1 Schematic of the generated FINCA allele and body weight curve of Nhlrc2 FINCA/− and Nhlrc2 +/+ mice. a Modified region of mouse and human amino acid (aa) and nucleotide sequence (nt) of the modified region. Silent mutation leading to removal of the PAM site and resulting in the addition of a TatI restriction site (green) was included in the ssODN. FINCA variant c.442G > T (red) and amino acid substitution p.Asp148Tyr (yellow) are shown. b Body weight curve showing no significant difference in weight gain for either sex between Nhlrc2 FINCA/− and Nhlrc2 +/+ mice during the 32-week observation period (two-way ANOVA). Nhlrc2 FINCA/− male (N = 6), Nhlrc2 +/+ male (N = 6) Nhlrc2 FINCA/− female (N = 6) Nhlrc2 +/+ female (N = 5). Mean and standard deviation (SD) shown hemolytic anemia are also common manifestations of FINCA disease (Uusimaa et al. 2018;Brodsky et al. 2020). There were no observable abnormalities in lung and liver sections of Nhlrc2 FINCA/− mice (Additional File 1: Fig.  S4), nor were there significant changes in hemoglobin or hematocrit values (Additional File 1: Table S4).
Although Nhlrc2 FINCA/− mice did not recapitulate the tissue manifestations of FINCA disease, the decrease in NHLRC2 protein at the tissue level is striking. Thus, our Nhlrc2 FINCA/− mouse model enables studies of the changes at the molecular and cellular level resulting from decreased NHLRC2. This may reveal the affected pathways that predispose and ultimately lead to a severe disease phenotype.
The LacZ cassette of the KO allele, Nhlrc2 tm1a(KOMP)Wtsi , resides in the intron between exons 4 and 5 of Nhlrc2, where it leads to the termination of transcription. Since KO first conditional-ready tma1 alleles can skip over the LacZ cassette and restore the original gene expression to some extent (White et al. 2013), we compared the expression prior to and after the LacZ site by quantitative polymerase chain reaction (qPCR) using primers preceding and primers overlapping the intronic area with the cassette. qPCR preceding the cassette resulted in 70% Nhlrc2 mRNA in Nhlrc2 FINCA/− NPCs compared to expression levels in wild-type NPCs (p = 0.004) (Fig. 4c). The expression of the full-length Nhlrc2 mRNA was 50.2% in the Nhlrc2 FINCA/− NPCs compared to wildtype NPCs (p = 0.0005) (Fig. 4b). Similarly, the full-length Nhlrc2 mRNA was decreased to 50% in immortalized FINCA patient fibroblast expressing only the missense variant (Uusimaa et al. 2018). Collectively, this suggests that Nhlrc2 mRNA encoding p.Asp148Tyr is stable both in humans and in mice. The KO allele of Nhlrc2 tm1a(KOMP) Wtsi retains 20% expression of the truncated protein with LacZ without a sign of leakage of the full-length Nhlrc2.

2D-DIGE revealed 19 proteins affected by Nhlrc2 FINCA/− genotype in embryonic cortical NPCs
To further study the effect of NHLRC2 deficiency on developing neurons, we compared Nhlrc2 FINCA/− and Nhlrc2 +/+ NPCs with two-dimensional difference gel electrophoresis (2D-DIGE). 2D-DIGE revealed 21 spots with significantly changed intensities between the two genotypes (Additional File 1: Fig. S6). Further analysis using mass spectrometry (MS) identified 19 unique proteins (Table 1, Additional File 1: Table S5). Among them, only the transitional endoplasmic reticulum ATPase (VCP) was decreased in Nhlrc2 FINCA/− NPCs, whereas all other proteins were increased in the mutant in comparison to Nhlrc2 +/+ NPCs.

STRING protein-protein interaction analysis
Protein-protein interaction analysis of the changed mouse proteins observed in 2DE-DIGE, and their human orthologs, showed significant enrichment of interaction for the identified proteins (mouse p = 0.00056 and human p = 0.00204) (Additional File 1: Fig. S7). Interestingly, VCP was found to be a central node in both networks; however, our further qPCR, SDS-PAGE, and immunoblotting experiments failed to show significant differences between Nhlrc2 FINCA/− and Nhlrc2 +/+ NPCs (Additional File 1: Table S6, Fig. S8). Collectively, this suggests that the observed change for VCP is the result of an additional posttranslational modification which, in turn, might hinder detection with the antibodies.

NHLRC2 affects proteins enriched in vesicular compartments
Gene ontology (GO) enrichment analysis of human orthologs for cellular components revealed that the majority of changed proteins were linked to extracellular exosomes, vesicles, and protein-containing complexes, as well as cytosol (Table 2). No enrichment was found in certain molecular functions or biological processes in the GO term analysis. Similarly, NHLRC2 has previously been linked to an exceptionally large variety of biological (See figure on next page.) Fig. 2 NHLRC2 is significantly decreased in Nhlrc2 FINCA/− mouse tissues. SDS-PAGE immunoblotting showing NHLRC2 in hippocampus, cerebellum, and brainstem of 13-week-old and lung and liver of 32-week-old Nhlrc2 FINCA/− and Nhlrc2 +/+ mouse tissue lysates. The amount of NHLRC2 decreased to 1.1% in the hippocampus (p < 0.0001), 3.3% in the cerebellum (p < 0.0001), 6.2% (p = 0.0004) in the brainstem, 3.2% (p < 0.0001) in the liver, and 1.7% (p = 0.011) in the lung of Nhlrc2 FINCA/− mice compared to wild-type mice (Student's t-test). The expected size of full length NHLRC2 is 78.4 kDa. Protein amounts are relative to one of the wild-type samples, and GAPDH was used for normalization. Dot blots show individual data points, group mean and standard error of the mean (SEM). ****p < 0.0001, ***p < 0.001 Hiltunen et al. Mol Med (2020) 26:123 processes Haney et al. 2018;Yeung et al. 2019).

Heterogeneous nuclear ribonucleoprotein C1/C2 is increased in NPCs and in the hippocampus of adult
We next performed qPCR to study whether the changes observed in 2D-DIGE could be explained by changes in transcription. Only three of the identified proteins, annexin A6 (ANXA6), ferritin light chain 1 (FTL1), and heterogeneous nuclear ribonucleoproteins C1/ C2 (hnRNP C1/C2), had significantly changed transcriptional levels in Nhlrc2 FINCA/− NPCs ( Fig. 5a and Additional File 1: Table S6). Contrary to the increased intensity observed in 2D-DIGE, the expression of these genes was decreased. Anxa6 had 75% (p = 0.023), Ftl1 80% (p = 0.035), and Hnrnpc 70% (p = 0.012) expression in mutant NPCs compared to the wild-type NPCs. Hence, the changes in 2D-DIGE are likely to be differences in the protein turnover rates or in the chemical and physical properties of the proteins, caused by, for example, changes in their post-translational modifications (PTMs), rather than differences in their expression. HnRNP C1/C2, isoform C2 showed elevated protein levels in Nhlrc2 FINCA/− compared to Nhlrc2 +/+ NPCs in the 2D-DIGE analysis (Table 1), whereas qPCR indicated decreased mRNA levels ( Fig. 5a and Additional File 1: Table S6). SDS-PAGE and immunoblotting confirmed the observed changes at the protein level with a 65% increase in Nhlrc2 FINCA/− NPCs (p = 0.045) (Fig. 5b). Immunocytochemistry (ICC) staining indicated normal localization of hnRNP C1/C2 in both mutant and wildtype cells (Additional File 1: Fig. S9).
In situ hybridization (ISH) of Nhlrc2 showed ubiquitous expression throughout the adult brain of 32-weekold male mice, with the most prominent expression in cerebellar granule cells, followed by granule cells in the dentate gyrus, and then by pyramidal cells in hippocampal CA1 layer and layer 2 of the piriform cortex (Fig. 6). ISH of Nhlrc2 FINCA/− brain revealed a similar expression pattern of the mutated Nhlrc2 mRNA to that of the wildtype (Additional File 1: Fig. S11). Since the hippocampus, cerebellum, and brainstem are affected in FINCA disease, we performed SDS-PAGE and immunoblotting analyses of hnRNP C2 in these tissues from 13-week-old female Nhlrc2 FINCA/− and Nhlrc2 +/+ mice. Interestingly, hnRNP C2 (820%, p = 0.046) was observed to be significantly increased in the hippocampus of Nhlrc2 FINCA/− in comparison to Nhlrc2 +/+ (Fig. 5c), but not in the cerebellum or brainstem of the same animals (Additional File 1: Fig.  S10). Furthermore, hnRNP C2 has been found to affect transcription of myelination related genes in human neuroblastoma cells (Iwata et al. 2011). Immunoblotting of 5-week-old mouse hippocampus lysates revealed a significant decrease in the amount of myelin-associated glycoprotein (MAG) in Nhlrc2 FINCA/− mice compared to wild type mice (69%, p = 0.027) (Fig. 5d). Myelin basic protein (MBP) also showed a decrease in the amount of protein but was not statistically significant (52%, p = 0.092) (Fig. 5d). These findings associate hnRNP C2 with the FINCA disease pathology and suggest that NHLRC2 plays an important role especially in the hippocampus.

Discussion
KI mouse models can be considered precision disease models that recapitulate the single pathogenic variant created into the endogenous target gene, omitting any artificial overexpression of gene products seen in conventional transgenic mouse models. Here, we generated a KI mouse line with the FINCA patient variant c. 442G < T, p.Asp148Tyr, in the endogenous mouse Nhlrc2 gene. By crossing this mouse line with the Nhlrc2 KO mouse line, we produced a similar compound heterozygous genotype to that identified in FINCA patients. Compound heterozygous Nhlrc2 FINCA/− mice overcame the embryonic lethality of Nhlrc2 KO mice and developed normally in utero. Since the missense mutation does not reside in the most evolutionarily conserved area of NHLRC2 that is predicted to be the active core of the protein (Biterova et al. 2018), the p.Asp148Tyr NHLRC2 may have retained some of its physiological activity, allowing normal embryonic development both in humans and mice. Nhl-rc2 FINCA/− mice retained 5% residual NHLRC2 throughout the body, including NPCs.

Table 1 List of significantly changed proteins from 2D-DIGE identified by MS (p < 0.05, minimum 1.5-fold change)
The ratio indicates the change of the normalized spot volumes in Nhlrc2 FINCA/− in comparison to Nhlrc2 +/+ NPCs  diseases are often characterized by protein aggregates, and, not surprisingly, mutations in endo-and autolysosomal pathway genes have been associated with several neurodegenerative disorders, such as AD, PD, and lysosomal storage disease (Wang et al. 2018;Menzies et al. 2015). Neurons aim to remove the accumulated proteins through an endosomal pathway by lysosomal degradation or by releasing them into the extracellular space via multivesicular bodies and exosomes (Kalani et al. 2014).
In addition to the late-onset storage diseases mentioned above, there are neurodevelopmental disorders with multisystemic involvement where autophagy impairment has been implicated, such as Vici syndrome (Hori et al. 2017) and Rett syndrome (Sbardella et al. 2017), which partly resemble FINCA disease. Collectively, the dysfunction in vesicular trafficking may create a predisposition to neurodegeneration in FINCA disease. The proteomic approach we used resulted not only in the identification of changes in the amount of proteins in vitro, but it also provided further information about the characteristics of these proteins revealing the presence of altered isoforms (hnRNP C2, RBM8A, PEA15) as well as fragments (UBA1, FTL1). This suggests that additional regulatory mechanisms apart from transcription and translation, such as alternative splicing, PTM, or different turnover rates may play a role in FINCA disease pathology. Whereas isoforms can have different functions, PTMs regulate protein folding or activity, subcellular targeting, and interaction with ligands or other proteins (Burkle 2001), among other functions. In this study, VCP was the only spot that showed decreased intensity in Nhlrc2 FINCA/− NPCs by proteomic analysis, and it was a central node in the protein-protein interaction analysis. Although the interaction analysis may include theoretical interactions that have not been confirmed by experimental evidence, these results made VCP an interesting hit. However, several VCP antibodies failed to recognize the spot identified by MS. This may suggest we identified a change in a specific VCP form, possibly resulting from a PTM, which changes the motility of VCP in 2D electrophoresis and prevents its recognition by antibodies. VCP is known to be regulated by a large number of PTMs (Hornbeck et al. 2014). Loss of the Nhlrc2 plant ortholog, suppressor of quenching 1, has been found to affect the electrophoretic mobility of plastid lipocalin, due to an unidentified protein modification (Malnoë et al. 2018). This raises the question whether VCP could be a target for such a protein modification. On the other hand, Nhlrc2 FINCA/− NPCs showed a striking decrease in the amount of NHLRC2, in contrast to a more modest decrease in Nhlrc2 mRNA expression. The introduction of the Tyr residue has been proposed to disrupt hydrogen bonding which may destabilize the conformation of the mutated NHLRC2 (Biterova et al. 2018). Mutated NHLRC2 seems to be unstable in vivo and would need to be degraded. VCP is a well-known player in the ubiquitin-proteasome pathway (Ye et al. 2005), and the decrease in VCP spot intensity could result from changes in PTMs of VCP required for the degradation of the mutated NHLRC2. Mutations in VCP have been connected to neurodegenerative multisystemic proteinopathies in humans, such as inclusion body myopathy with Paget disease of bone and frontotemporal dementia (IBMPBFD) (Kimonis et al. 2000), amyotrophic lateral sclerosis (ALS) (Johnson et al. 2010), and Charcot-Marie-Tooth disease type 2 (CMT2) (Gonzalez et al. 2014). VCP is involved in a large number of cellular processes, but it is also involved in endo- (Ritz et al. 2011;Ramanathan and Ye 2012;Pleasure et al. 1993) and autophagic pathways (Ju and Weihl 2010). Interestingly, VCP deficiency leads to accumulation of immature autophagic vesicles (Ju et al. 2009;Tresse et al. 2010), and IBMPFD patient myoblasts accumulate large LAMP-1 and LAMP-2 positive vacuoles and LC3-II (Tresse et al. 2010). LAMP-1 positive multilamellar bodies have also been detected in electron microscopy images of immortalized FINCA patient fibroblasts .
Anxa6, Ftl1, and Hnrnpc showed significant but opposite changes in mRNA levels out of the 19 proteins identified by proteomics analysis. This kind of discrepancy between mRNA and protein levels has been observed in numerous studies (Abdulghani et al. 2019;Poverennaya et al. 2017) and reflects the independent but most often compensatory regulatory mechanisms of transcription  5 qPCR and immunoblotting showed changes in the hnRNPC C2 level in NPCs and hippocampus of Nhlrc2 FINCA/− mouse. a qPCR results of the three protein transcripts; Ftl1, Anxa6, and Hnrnpc, showed significantly changed RNA levels in Nhlrc2 FINCA/− (square) compared to Nhlrc2 +/+ (round) NPCs, 80% (p = 0.035), 75%, (p = 0.023), and 70% (p = 0.012) respectively (Student's t-test). Expression was normalized against one wild-type sample. Individual data points, group mean, and SEM are shown. b SDS-PAGE immunoblot and measured protein levels of three Nhlrc2 −/− and Nhlrc2 FINCA/− NPC lysates revealed increased hnRNP C2 (165% p = 0.045). c SDS-PAGE immunoblot and measured protein levels from 13-week-old female Nhlrc2 FINCA/− and wild-type hippocampal tissue lysates showed increased hnRNP C2 (820%, p = 0.046). hnRNP C2-specific antibody was used and the expected size of hnRNP C2 is 34.4 kDa. d SDS-PAGE immunoblot and measured protein levels from 5-week-old male Nhlrc2 FINCA/− and wild-type hippocampal tissue lysates for myelin basic protein (MBP) and myelin associated glycoprotein (MAG). MAG was significantly decreased in Nhlrc2 FINCA/− (69%, p = 0,027). Protein amounts are relative to one of the wild-type samples and GAPDH was used for normalization. Individual data points, group mean, and SEM are shown. Statistical analysis was done using Student's t-test *p < 0.05 (See figure on next page.) Hiltunen et al. Mol Med (2020) 26:123 and translation. For Hnrnpc, the lower mRNA level could also be explained by autoregulation of its own translation, which has been described for several hnRNP family members (Wollerton et al. 2004;McGlincy et al. 2010;Buratti and Baralle 2011;Rossbach et al. 2009;Müller-McNicoll et al. 2019). The increase in the hnRNP C2 protein in Nhlrc2 FINCA/− embryonic NPCs and adult mouse hippocampus suggests a perturbation in overall RNA metabolism, which in turn may contribute to the neurodegeneration seen in FINCA disease. HnRNP C1/ C2 is part of the hnRNP family, which is a large RNAbinding family contributing to multiple aspects of nucleic acid metabolism. Dysregulation of RNA homeostasis has been suggested as a common feature of neurodegenerative diseases, where RNA-binding proteins (RBPs) play a crucial role (Wolozin and Ivanov 2019;Conlon and Manley 2017). RBPs often have low-complexity domains, which are thought to facilitate stress and P body formation, but also make RBPs prone to self-aggregate (March et al. 2016). VCP, in contrast, has been shown to function in the clearance of stress and P bodies (Buchan et al. 2013). hnRNPs have been recognized in several common late-onset neurodegenerative diseases, such as spinal muscular atrophy (SMA), ALS, AD, and frontotemporal dementia (Wolozin and Ivanov 2019;Geuens et al. 2016;Purice and Taylor 2018). Interestingly, an increase in hnRNP C2 was detected in the hippocampus of adult Nhlrc2 FINCA/− mice, but not in the cerebellum or brainstem. The hippocampal CA1 pyramidal cell layer was also one of the brain regions of adult mice with the highest expression of Nhlrc2 according to our ISH data. Altogether, this data suggests that NHLRC2 may play an important role in hippocampal pyramidal cells, which were also affected by neurodegeneration in FINCA patients (Uusimaa et al. 2018). KI models often show a milder phenotype than human patients (Dawson et al. 2018). Here, the Nhl-rc2 FINCA/− mice showed normal growth compared to their wild-type litter mates and did not develop a severe disease phenotype as described in FINCA patients. Environmental factors such as pathogens or the specific genetic background may play an additional role in triggering a more severe outcome of the disease in mice (Doetschman 2009;Keane et al. 2011;Qosa and Kaddoumi 2016). Interestingly, NHLRC2 has recently been connected to immunological responses via initiation of phagocytosis in human macrophages, possibly through its effect on RhoA/Rac1 signalling, which controls actin polymerization and filopodia formation (Haney et al. 2018;Yeung et al. 2019). In addition to phagocytosis, actin cytoskeleton dynamics are important for many other cellular functions requiring remodelling of plasma membrane, or biogenesis and transport of vesicular cargo, such as cell motility, cytokinesis, endocytosis, and autophagy (Anitei and Hoflack 2012). Our proteomic analysis revealed changes in several proteins connected to actin dynamics in Nhlrc2 FINCA/− NPCs, including BRICK1 (Eden et al. 2002;Gautreau et al. 2004), costars family protein ABRACL (Pang et al. 2010), plastin-3 (Bretscher 1981, prefoldin subunit 2 (Tsao et al. 2006;Martin-Benito et al. 2002), and VCP (Nalbandian et al. 2012;Chan et al. 2012;Shah and Beverly 2015). Finally, Nhlrc2 FINCA/− mice, housed in specific pathogen-free conditions, may have lacked an immunological component important for triggering the FINCA disease onset, where recurrent infections were present in human patients.

Conclusions
Even though mouse models of neurodegeneration seldom show the whole spectrum of the human disease, they have proven to be invaluable tools for studying the molecular pathogenesis of neurodegeneration (Dawson et al. 2018;Leung and Jia 2016). Our results suggest for the first time that the Nhlrc2 FINCA/− genotype results in dysregulation of RNA homeostasis in mouse neurons, which could ultimately contribute to the pathophysiology of neurodegeneration in FINCA disease. Further studies on human neurons are still required, but our findings presented here open future possibilities in the search for treatment and diagnosis tools for FINCA and other relevant, more common neurodegenerative diseases (Samie and Xu 2014;Spilman et al. 2010;Caccamo et al. 2014;Kim et al. 2014;Cheng et al. 2015;Lugli et al. 2015). Considering the broad spectrum of manifestations of FINCA disease, including lung and liver fibrosis, chronic hemolytic anemia, and cerebral angiomatosis, the FINCA mouse model described here offers an interesting and biologically relevant research tool for a variety of future studies.
(See figure on next page.) Fig. 6 Representative in situ hybridization of Nhlrc2 in the brain of a 32-week-old male C57BL/6NCrl mouse. a Overview image of ISH with Nhlrc2 probe (red). Red boxes mark the magnified areas: a CA1 layer of hippocampus, b dentate gyrus of hippocampus, c piriform cortex, d cerebellar hemisphere. b Nhlrc2 expression in areas a-d. A strong signal was observed from the pyramidal cell layer of the hippocampus and piriform cortex as well as dentate and cerebellar granule cells. c Negative control from areas a-d Hiltunen et al. Mol Med (2020) 26:123 Supplementary information Supplementary information accompanies this paper at https ://doi. org/10.1186/s1002 0-020-00245 -4.
Additional file 1: Detailed description of the materials and methods, supplementary figures S1-S11, and supplementary tables S1-S6. Materials and methods include generation of FINCA mouse, used animals, genotyping, Sanger equencing, histology, blood analysis, ISH, NPC culture, 2DE-DIGE, mass spectrometry, qPCR, immunoblotting, ICC, data analysis, and statistical considerations. Fig. S1 Genotyping and Sanger sequencing of three founders obtained from microinjections of Cas9 ribonucleoprotein and ssODN into mouse zygotes. Fig. S2 SDS-PAGE immunoblotting showing decrease of NHLRC2 in different brain regions and different tissues of Nhlrc2 FINCA/− mice compared to wild type mice.  Additional file 2: Unedited full images of the immunoblots for Fig. 2, Fig. 4, Fig. S2, Fig. S3, Fig S8,