Neuronal Phenotype of col4a1 and col25a1: An Intriguing Hypothesis in Vertebrates Brain Aging

Collagens are the most abundant proteins in vertebrates and constitute the major components of the extracellular matrix. Collagens play an important and multifaceted role in the development and functioning of the nervous system and undergo structural remodeling and quantitative modifications during aging. Here, we investigated the age-dependent regulation of col4a1 and col25a1 in the brain of the short-lived vertebrate Nothobranchius furzeri, a powerful model organism for aging research due to its natural fast-aging process and further characterized typical hallmarks of brain aging in this species. We showed that col4a1 and col25a1 are relatively well conserved during vertebrate evolution, and their expression significantly increases in the brain of N. furzeri upon aging. Noteworthy, we report that both col4a1 and col25a1 are expressed in cells with a neuronal phenotype, unlike what has already been documented in mammalian brain, in which only col25a1 is considered a neuronal marker, whereas col4a1 seems to be expressed only in endothelial cells. Overall, our findings encourage further investigation on the role of col4a1 and col25a1 in the biology of the vertebrate brain as well as the onset of aging and neurodegenerative diseases.


Introduction
Aging is a progressive and irreversible process associated with physical and functional decline, and therefore considered the primary risk factor in the onset and exacerbation of neurodegenerative diseases [1,2].
According to Wyss-Coray (2016), it is possible that normal brain aging almost physiologically forms a continuum with neurodegeneration and disease given the high incidence of neurodegenerative diseases in the elderly and the rarity of disease-free brains with advancing age [3]. However, this author recognized that considering neurodegeneration, the natural expression of hastened aging is unhelpful as it is critical to understand how stochastic factors interact in defining a specific disease. Therefore, to enhance our understanding of the aging contribution to disease, we need to analyze how environmental and genetic factors cause the onset of a particular disease rather than another, and recognize the relevance of these processes in the disease itself [3]. From this perspective, interesting findings stemmed from Podtelezhnikov et al. (2011), who performed an extended analysis of age-and disease-related changes in gene expression in the brain of people affected by Alzheimer's disease (AD) and those disease-free. In this work, the authors showed signatures of accelerated aging in a neuronal-gene expression module identifying four transcriptional biomarkers including numerous collagen genes [4].

Identification and Evolutionary History of col4a1 and col25a1
The branching pattern shows that col4a1 of N. furzeri is closely related to most teleost fish species ( Figure 1A). Interestingly, in this species, col4a1 displays a common ancestor with human COL4A1 in comparison to what happens in mice and zebrafish. Intriguingly, a clearly divergent evolutionary event occurred in the case of the col25a1 branching pattern ( Figure 1B), where col25a1 in N. furzeri is phylogenetically distant from all the other selected species, being on a separated branch. Exploiting our knowledge of N. furzeri and its above-mentioned points of strength [29], in this work, we report for the first time that (a) COL4A1 and COL25A1 show a discrete degree of evolutionary conservation among vertebrates; (b) col4a1 and col25a1 expression displays age-related level changes coherently with a wide neuroanatomical pattern of expression; (c) col4a1 and col25a1 mRNAs are expressed in neurons.

Identification and Evolutionary History of col4a1 and col25a1
The branching pattern shows that col4a1 of N. furzeri is closely related to most teleost fish species ( Figure 1A). Interestingly, in this species, col4a1 displays a common ancestor with human COL4A1 in comparison to what happens in mice and zebrafish. Intriguingly, a clearly divergent evolutionary event occurred in the case of the col25a1 branching pattern ( Figure 1B), where col25a1 in N. furzeri is phylogenetically distant from all the other selected species, being on a separated branch.
We also analyzed the degree of evolutionary conservation of COL4A1 and COL25A1 proteins by aligning sequences of human, mouse, zebrafish, and killifish for both proteins (Supplementary Figure S1). The aminoacidic sequence of N. furzeri col4a1 displayed 57.53% of identity with zebrafish, and 61.84% and 58.90% with human and mouse sequences, respectively. N. furzeri col25a1, instead, showed 72.62%, 58.17%, and 71.13% of identity with zebrafish, human, and mouse sequences, respectively.

Age-Related Expression of Brain Cell Lineage Markers and col4a1 and col25a1
Terzibasi Tozzini and colleagues [26] characterized, at the molecular level, the brain aging phenotype in N. furzeri and documented an increased level of some markers (S100β, gfap, and pcna) in 32-week old specimens compared to 25-week old ones. Therefore, we We also analyzed the degree of evolutionary conservation of COL4A1 and COL25A1 proteins by aligning sequences of human, mouse, zebrafish, and killifish for both proteins (Supplementary Figure S1). The aminoacidic sequence of N. furzeri col4a1 displayed 57.53% of identity with zebrafish, and 61.84% and 58.90% with human and mouse sequences, respectively. N. furzeri col25a1, instead, showed 72.62%, 58.17%, and 71.13% of identity with zebrafish, human, and mouse sequences, respectively.

Age-Related Expression of Brain Cell Lineage Markers and col4a1 and col25a1
Terzibasi Tozzini and colleagues [26] characterized, at the molecular level, the brain aging phenotype in N. furzeri and documented an increased level of some markers (S100β, gfap, and pcna) in 32-week old specimens compared to 25-week old ones. Therefore, we decided to characterize the brain phenotype in 27-week old N. furzeri to test whether brain aging is already well established at this age by measuring S100β, gfap and pcna [27], dcx, and sox2 [30,31].
In compliance with Terzibasi Tozzini and colleagues (2012), our quantitative analyses showed significant upregulation for gfap p ≤ 0.0001, s100β (p ≤ 0.0001), and pcna (p ≤ 0.0001) over aging. Furthermore, we also observed a strong increase of sox2 (p ≤ 0.0001) expression while a slight upregulation was detected for dcx (p ≤ 0.5) (Figure 2A). decided to characterize the brain phenotype in 27-week old N. furzeri to test whether brain aging is already well established at this age by measuring S100β, gfap and pcna [27], dcx, and sox2 [30,31].

Neuroanatomical
Localization of col4a1 and col25a1 mRNAs in the Brain of Young and Old Animals 2.3.1. col4a1 and col25a1 mRNAs in the Brain of Young Animals In the brain of young animals, col4a1 mRNAs appeared poorly expressed and restricted only to a few cells in the periventricular gray zone (PGZ) of the optic tectum (OT) and around the longitudinal tori (Tl) ( Figure 3A). Conversely, col25a1 mRNA was abundantly expressed at the margin between caudal diencephalon and cranial midbrain, specifically in the glomerular nucleus (NG), dorsal hypothalamus (Hd), nucleus of posterior recess (NRP), and posterior tuberal nuclei (TNp) ( Figure 3B,B'). In the midbrain, numerous positive cells were observed in the PGZ of the OT ( Figure 3B). In the brain of young animals, col4a1 mRNAs appeared poorly expressed and r stricted only to a few cells in the periventricular gray zone (PGZ) of the optic tectum (O and around the longitudinal tori (Tl) ( Figure 3A). Conversely, col25a1 mRNA was abu dantly expressed at the margin between caudal diencephalon and cranial midbrain, sp cifically in the glomerular nucleus (NG), dorsal hypothalamus (Hd), nucleus of posteri recess (NRP), and posterior tuberal nuclei (TNp) ( Figure 3B,B'). In the midbrain, nume ous positive cells were observed in the PGZ of the OT ( Figure 3B).

col4a1 mRNA in the Brain of Old Animals
The expression pattern of col4a1 mRNA in the brain of old animals was widely o served in all brain regions.
Forebrain col4a1 mRNA labeling was detected in dense and positive cells in the external (EC and internal cellular layer (ICL) of the olfactory bulbs ( Figure 4A,A'). In the medial zon of dorsal telencephalon, few packed cells were observed in cell groups 1 and 2 (Dm1-( Figure 4A-A'') and caudally in 4 ( Figure 4B). Numerous cells were detected in the dors lateral zone (Dld) ( Figure 4A). Few cells were labeled along the dorsal zone (Dd) and

col4a1 mRNA in the Brain of Old Animals
The expression pattern of col4a1 mRNA in the brain of old animals was widely observed in all brain regions.
Forebrain col4a1 mRNA labeling was detected in dense and positive cells in the external (ECL) and internal cellular layer (ICL) of the olfactory bulbs ( Figure 4A,A'). In the medial zone of dorsal telencephalon, few packed cells were observed in cell groups 1 and 2 (Dm1-2) ( Figure 4A-A") and caudally in 4 ( Figure 4B). Numerous cells were detected in the dorsolateral zone (Dld) ( Figure 4A). Few cells were labeled along the dorsal zone (Dd) and in the zones of ventral telelencephalon (Vd, Vv, Vp, and Vs) ( Figure 4B). Strong probe signal was displayed along the telencephalic ventricles.
In the diencephalon, an intense col4a1 mRNA signal was detected in cells of the an terior preoptic (PPa) ( Figure 4B) and posterior preoptic (PPp) nuclei, magnocellular pre optic (PM), ventro-medial (VM), and anterior (A) thalamic nuclei ( Figure 4C) and habenu lar nuclei (Ha) ( Figure 4C,D). More caudally, numerous intensely labeled cells in the dor sal hypothalamus and a few diffused cells in the inferior lobe of the hypothalamus wer detected ( Figure 5A,A''').  In the diencephalon, an intense col4a1 mRNA signal was detected in cells of the anterior preoptic (PPa) ( Figure 4B) and posterior preoptic (PPp) nuclei, magnocellular preoptic (PM), ventro-medial (VM), and anterior (A) thalamic nuclei ( Figure 4C) and habenular nuclei (Ha) ( Figure 4C,D). More caudally, numerous intensely labeled cells in the dorsal hypothalamus and a few diffused cells in the inferior lobe of the hypothalamus were detected ( Figure 5A,A"').

Midbrain
In the tegmentum, dense positive cells were observed nearby the nucleus of the lateral valvula (LV) ( Figure 5A) and nucleus of the lateral longitudinal fascicle (Nllf) ( Figure 5A,A'). col4a1 mRNA was detected in the PGZ ( Figure 5A-A"). An intense signal was observed in the Tl at the margin with the PGZ. Strong col4a1 mRNA signal was also displayed caudally in the layer of the semicircular tori. observed in the Tl at the margin with the PGZ. Strong col4a1 mRNA signal was also displayed caudally in the layer of the semicircular tori.

Hindbrain
In the cerebellum, strong signal was observed in the valvula and in the granular and molecular layers of the body and, caudally, in the granular eminentiae. Several cells containing col4a1 mRNA were seen in the central griseum, at the dorsal margin with the medulla oblongata.

Hindbrain
In the cerebellum, strong signal was observed in the valvula and in the granular and molecular layers of the body and, caudally, in the granular eminentiae. Several cells containing col4a1 mRNA were seen in the central griseum, at the dorsal margin with the medulla oblongata.

col25a1 mRNA in the Brain of Old Animals Forebrain
In the telencephalon, col25a1 mRNA was detected in numerous positive cells in the internal cellular layer (ICL) of olfactory bulbs, mostly along the ventricular region and in the external cellular layer (ECL) ( Figure 6A). Densely packed cells were observed in layers 1-2 of the medial zone of dorsal telencephalon ( Figure 6A,B). Additionally, in the dorsal (Dd), lateral-dorsal (Dld), and latero-ventral (Dlv) ( Figure 6A,C,D) zones of the dorsal telencephalon, numerous intensely positive cells were seen. In the ventral telencephalon, intense staining was detected in densely packed cells of the ventral (Vv) and supracommissural (Vs) zones.
internal cellular layer (ICL) of olfactory bulbs, mostly along the ventricular region and in the external cellular layer (ECL) ( Figure 6A). Densely packed cells were observed in layers 1-2 of the medial zone of dorsal telencephalon ( Figure 6A,B). Additionally, in the dorsal (Dd), lateral-dorsal (Dld), and latero-ventral (Dlv) ( Figure 6A,C,D) zones of the dorsal telencephalon, numerous intensely positive cells were seen. In the ventral telencephalon, intense staining was detected in densely packed cells of the ventral (Vv) and supracommissural (Vs) zones.

Hindbrain
In the hindbrain, col25a1 mRNA is widely diffused in the granular emin in the granular and molecular layers of the body of the cerebellum (CCe), and i tral griseum (gc).

Midbrain
Numerous positive cells were seen in the layers of optic tectum, mostly in the periventricular gray zone and in the superficial white zone ( Figure 7C,D). Few cells expressing col25a1 mRNA were detected in the longitudinal tori (Tl). In the tegmentum, densely packed positive cells were localized in the semicircular tori (TS), in the nucleus of lateral longitudinal fascicle (Nllf), and in the nucleus of lateral valvula (LV) ( Figure 7E).

Hindbrain
In the hindbrain, col25a1 mRNA is widely diffused in the granular eminentiae EG, in the granular and molecular layers of the body of the cerebellum (CCe), and in the central griseum (gc).

Characterization of col4a1 and col25a1 Expressing Cells in the Brain of Old Animals
To characterize the cellular phenotypes expressing both col4a1 and col25a1 mRNAs in the brain of old N. furzeri, we conducted experiments of in situ hybridization and immunofluorescence via anti-S100β as a marker of glial cell population [32]. We further identified the neuronal phenotype by specifically labeling neuronal processes, neural RNAbinding proteins, and mature neurons via anti-DCX, anti-HuC/HuD, and anti-MAP2, respectively [33].
2.4.1. S100β col4a1/s100β was faintly co-localized only along the diencephalic ventricle ( Figure 8A). In contrast, in the same area, we observed strong co-localization of col25a1/s100β and in few sparse thalamic glial cells ( Figure 8B,C). To characterize the cellular phenotypes expressing both col4a1 and col25a1 mRNA in the brain of old N. furzeri, we conducted experiments of in situ hybridization and im munofluorescence via anti-S100β as a marker of glial cell population [32]. We further iden tified the neuronal phenotype by specifically labeling neuronal processes, neural RNA binding proteins, and mature neurons via anti-DCX, anti-HuC/HuD, and anti-MAP2, re spectively [33].

S100β
col4a1/s100β was faintly co-localized only along the diencephalic ventricle (Figur 8A). In contrast, in the same area, we observed strong co-localization of col25a1/s100β and in few sparse thalamic glial cells ( Figure 8B,C).

DCX, HuC/HuD, MAP2
In the forebrain, col4a1 and col25a1 were faintly co-localized with DCX along the tel encephalic ventricle ( Figure 9A,B), whereas col25a1/DCX also appeared co-localized in some neurons of layer 4 of the central zone of dorsal telencephalon ( Figure 9B). No co staining was observed in the mid-nor in the hindbrain.

DCX, HuC/HuD, MAP2
In the forebrain, col4a1 and col25a1 were faintly co-localized with DCX along the telencephalic ventricle ( Figure 9A,B), whereas col25a1/DCX also appeared co-localized in some neurons of layer 4 of the central zone of dorsal telencephalon ( Figure 9B). No co-staining was observed in the mid-nor in the hindbrain.  Conversely, col4a1 and col25a1 were mainly co-localized with HuC/HuD in the midand hindbrain. Weak col4a1/HuC/HuD co-localization was observed in the optic tectum, in the PGZ (Figure 9C), in the body of cerebellum and granular eminentiae (EG) ( Figure 9D). In very few neurons col25a1/HuC/HuD were co-localized in the PM of caudal diencephalon ( Figure 9E), in the body of the cerebellum ( Figure 9F) as well as in the TS ( Figure 9G) and granular eminentiae (EG) ( Figure 9G).
Only in a few neurons of the PGZ were col4a1 and col25a1 co-localized in MAP2 immunoreactive cells (Figure 9H,I).

Discussion
In this study, we analyzed the age-dependent regulation of the collagen genes col4a1 and col25a1 in the brain of the short-lived teleost Nothobranchius furzeri. We showed that both col4a1 and col25a1 undergo age-dependent upregulation and that in N. furzeri, are expressed in brain region homologous to that of mammals. Furthermore, we showed that both col4a1 and col25a1 are expressed in neurons rather than in glial cells.
Collagen genes have been detected in short-and long-lived species [34,35] and their expression was shown to correlate with the lifespan of 33 diverse mammalian species [35]. Ewald and coworkers (2015) confirmed that (i) experimental conditions extending lifespan upregulate collagen genes, and (ii) lifespan extension by genetic reduction in insulin/IGF-1 signaling requires the expression of collagen genes [36]. Based on these results, collagen (and the extracellular matrix in general) undergoes age-dependent modifications and plays an important role in the aging process.
The phylogenetic analysis of col4a1 and col25a1 displays a different evolution of the two genes among vertebrate species, with col4a1 showing a high degree of conservation among vertebrates, even between the short-lived vertebrate N. furzeri and humans (61.84% homology). Conversely, the nucleotide sequence of col25a1 occupies a very distant position compared to the selected actinopterygians or mammalian species.
In agreement with our findings, it has been previously reported that the expression of collagen type IV increases during aging in the human brain, where it accumulates in the basal lamina of cerebral microvessels, therefore, it is considered as a major risk factor for the development of stroke and vascular dementia [9]. Notably, in mouse brain, expression levels of collagen type IV α6 chain followed the expression of Sox2, a key regulator of neurogenesis [37,38]. In the brain of N. furzeri, we also observed a parallel age-related increase in col4a1 and sox2, showing that the functional relationship of these two genes might be conserved in this model organism, in which they could possibly be implicated in neuronal physiology (e.g., axon guidance and neurite outgrowth), as already shown in zebrafish [12].
With regard to col25a1 mRNA, this is the first time that col25a1 has been quantitatively measured in the brain of a vertebrate. A previous study has documented that the expression levels of col25a1 mRNA in the mammalian brain increased from developmental to adult stages [16], and the expression levels remained stable from adult to old stages.
The age-increased levels of the col4a1 and col25a1 in the N. furzeri brain were confirmed by the wide expression observed via in situ hybridization. In the brain of five week post hatching (wph) animals, col4a1 and col4a1 mRNAs localization was restricted to very few brain areas of the fore-and midbrain, with col4a1 being expressed only in the optic tectum. The wide neuroanatomical expression patterns of col4a1 and col25a1 in old animals appear in several areas that are homologous to the adult mammalian brain [6]. In more detail, in N. furzeri col4a1 mRNA is expressed in the (a) olfactory bulbs, the brain region where the olfactory stimulus is integrated first in vertebrates brain; (b) dorsal telencephalic areas homologous to the mammalian hippocampus; and (c) brain areas responsible for integrating visual stimuli such as optic lobes.
A more detailed description of col25a1 mRNA in the adult murine brain reports that it is expressed in the retina and appears specifically enriched in retino-recipient nuclei within the brain (including the suprachiasmatic nucleus, lateral geniculate complex, olivary pretec-tal nucleus, and superior colliculus) [6]. Notably, in N. furzeri, we observed strong col25a1 labeling in brain nuclei involved in visual processing (i.e., preoptic nuclei, longitudinal tori) and the highest expression in the optic tectum, where the visual stimulus is processed and integrated in fish. These observations contribute to further corroborate the hypothesis that col25a1 is evolutionary conserved in the visual processing systems of vertebrates.
To gain more insights into the cell phenotype expressing col4a1 and col25a1, we employed S100β to identify glial cells [27] and DCX, HuC/HuD, and MAP2 to identify neurons [39] by combined in situ hybridization and immunofluorescence experiments. Very interestingly, col4a1 and col25a1 were both expressed in neuronal cells and not in glial cells. Our observations are of valuable interest in the role and expression of collagen genes in the brains of vertebrates, demonstrating the neuronal expression of both genes. However, some differences were observed in the neuronal expression of col4a1 and col25a1: for instance, in the same brain region as caudal telencephalon, we documented the expression of col25a1 in some DCX immunopositive neurons and very few in the case of col4a1/DCX co-labeled neurons. When analyzing col4a1/HuC/HuD and col25a1/HuC/HuD as well as col4a1/MAP2 and col25a1/MAP2 co-stainings, very few neurons of mid-and hindbrain appeared co-localized. Interestingly, RNA-Seq analysis coupled with qPCR and in situ hybridization carried out on the cerebral cortex of wild type mouse models, identified different collagen enriched genes and splicing isoforms [40], and clearly confirmed that col25a1 is synthesized by mature neurons. In contrast, col4a1 is not expressed in neuronal cell types but only in pericytes [40].
Finally, our observations on the age-dependent changes of col4a1 and col25a1 are corroborated by the further characterization of the aging phenotype of the N. fuzeri brain. To this aim, we analyzed the age-dependent expression of the major brain cell lineage markers: pcna, dcx, gfap, s100β, and sox2. We observed no significant differences in the expression of pcna between 5 wph and 27 wph animals. Significant upregulation of pcna has been previously reported by Terzibasi Tozzini et al. (2012) between 32 wph and 25 wph animals. Our data suggest that this gene might undergo age-related upregulation during later stages, probably after 25-27 wph. We confirmed the upregulation of gfap and s100β in the brain of 27 wph animals. Both gfap and s100β are markers of gliosis, and their upregulation during aging has been previously reported in the same model at 32 wph by Terzibasi Tozzini et al. [26]. Our results suggest that the brain aging phenotype already appears around the age of 27 wph. Noteworthy, S100 proteins are reported to be upregulated during aging and/or in the course of brain injuries and neurodegeneration [41]. With regard to neuronal markers, our experiments showed a significant increase in sox2 (i.e., the sex determining region Y-box 2) in the brain of N. furzeri at 27 wph, whereas it is reported to diminish with aging in mouse and human [31]. Unchanged levels of dcx expression upon aging was observed when comparing the brain of young (5 wph) and old (27 wph) specimens, unlike what has been reported in the human hippocampus [39]. Altogether, these data open intriguing avenues for future studies on the timing and identification of potential hallmarks of aging in the brain of this model species.
However, we have to consider that the impact of our results could have provided more exhaustive information by including in the study the comparison with another model species such as zebrafish. In addition, the evaluation of the expression of the two target genes at a further time-point, specifically an intermediate age between five and 27 weeks would have provided interesting insights to identify the moment of transition between the non-expression and the expression of the two target genes over aging.
In any case, our findings represent robust evidence that col4a1 and col25a1 are evolutionary well conserved in vertebrates, further confirmed by the neuronal expression of col25a1, and shed light on the potential neuronal role of col4a1. Thus, the age-related increase in the two genes may help to gain more insights into the crucial role of collagen in brain aging onset.

Protocols
The protocols for animal care and use were approved by the appropriate committee at the University of Naples Federico II (2015/0023947). All experimental procedures involving animals were carried out in accordance with the European Parliament and the Council of the European Union Directive of 22 September 2010 (2010/63/UE) and Italian law (D. lgs. 26/2014).

Phylogenetic Analysis and Protein Alignment
Putative col4a1 (GAIB01100805.1) and col25a1 (GAIB01090796.1) coding sequences were retrieved from GeneBank-NCBI. Available from: https://www.ncbi.nlm.nih.gov/, accessed on 9 January 2022). The evolutionary history was inferred by using the maximum likelihood (ML) method and Tamura-Nei model [42]. The bootstrap consensus tree inferred from 100 replicates [43] was considered to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates were collapsed. Initial tree(s) for the heuristic search were obtained automatically by applying neighbor-join and BioNJ algorithms to a matrix of pairwise distances estimated using the maximum composite likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. This analysis involved 28 nucleotide sequences. Codon positions included were 1st + 2nd + 3rd + Noncoding. All positions containing gaps and missing data were eliminated (complete deletion option). There was a total of 2338 positions in the final dataset. Evolutionary analyses were conducted in MEGA X [44].

Animals and Tissue Sampling
All experiments were performed on group-housed N. furzeri, wild-type MZM-04/10 strain, of both sexes. Maintenance was performed as previously described [46]. In order to avoid the effect of circadian rhythms and feeding, respectively, animals were suppressed around 10 a.m. and in the fasted state. Young [5 weeks post hatching (wph)] and old (27 wph) animals were anesthetized by immersion in tricaine methanesulfonate (MS-222, Sigma-Aldrich, 300 mg/L of in aqueous solution). For RNA extraction, fish were decapitated, brains were rapidly dissected, kept in sterile tubes (Eppendorf) with 500 µL of RNA later (Qiagen), and stored at 4 • C until RNA extraction. For fluorescence in situ hybridization (FISH) and combined FISH with immunofluorescence (IF), fish were decapitated, heads were rapidly excised and fixed in 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS) treated with diethylpyrocarbonate (DEPC) for 24 h at 4 • C. Brains were then embedded for cryostat sectioning by successive incubation in 30% sucrose solution and 20% sucrose solution at 4 • C overnight (ON). After embedding, samples were frozen at −80 • C. Serial transversal sections of 14 µm thickness were cut with a Leica cryostat (Deerfield, IL).

RNA Isolation and cDNA Synthesis
For RNA extraction, tissues were taken out of RNA later and the excess reagent was removed by means of sterile pipettes. N. furzeri total RNA was isolated from 10 animals for each time point (5 wph and 27 wph) with QIAzol (Qiagen), as previously described [44]. Homogenization was performed using a TissueLyzer II (Qiagen) at 20 Hz for 2-3x 1 min. Total RNA was then quantitated with Eppendorf BioPhotometer and 500 ng was retrotranscribed to cDNA using the QuantiTect ® Reverse Transcription Kit (Qiagen) and following the supplier's protocol. The newly synthetized cDNA was then diluted to a final volume of 200 µL with ultra-pure sterile nuclease free water to an approximate final cDNA concentration of 40 ng/µL. was included as the housekeeping gene. Experiments were run on a 7300 Real-Time PCR System (Applied Biosystems).

Statistical Analysis
Expression levels of col4a1, col25a1, pcna, dcx, sox2, s100β, and gfap mRNAs were analyzed by the ∆∆Ct method and normalized to the expression of the reference gene TBP ± SEM. Differences in expression levels for each time point (5 wph and 27 wph) were measured through fold change (FC) values to TBP (log2 base scale). The results in output were subsequently employed to build the relative ∆∆Ct. Graphic was built with the ggplot2 package [48] for R programming language (R Development Core Team, 2008).

Riboprobe Synthesis
Riboprobes to identify N. furzeri col4a1 and col25a1 mRNA were synthesized by in vitro transcription (IVT) using the MAXIscript ™ SP6/T7 IVT Kit (Invitrogen by Thermo Fisher Scientific-Catalogue number AM1312) following the manufacturer's instructions. One µg of DNA template was transcribed to RNA in 20 µL volume reaction including 5 µL of 10× DIG RNA labeling mix (Roche, cat. 11 277 073 910) containing 3.5 mM of digoxigenin labeled uracil (DIG-11-UTP). T7 RNA polymerase was employed. Therefore, the T7 promoter sequence (5 -GGTAATACGACTCACTATAGG-3 ) was associated upstream to the reverse primers of col4a1 and col25a1. All components were briefly centrifuged and incubated at 37 • C for 1 h. Then, 1 µL of turbo DNase 1 was added, the sample was mixed well, and incubated for 15 min at 37 • C. One µL of EDTA 0.5 M, pH 8.8, was added to stop the reaction. Reaction products were analyzed by gel electrophoresis and quantified. To validate and determine the riboprobe concentration, the dot blot technique was employed. Riboprobe's fresh dilution buffer was prepared mixing DEPC H2O, saline sodium citrate buffer (SSC) 20×, and formaldehyde 37% (5:3:2). Probes were then diluted in the dilution buffer, as indicated in Table 2, in a final volume of 20 µL. A negative control was included for each probe. One µL of each probe dilution plus negative controls were then pipetted onto a nitrocellulose membrane, forming little separated spots (Table 1). The membrane was then fixed in the Stratalinker ® UV Crosslinker for 30 s, gently washed in 1 × PBS and incubated for 30 min with in situ blocking solution (BS) (1 × PBS, 10% of sheep serum heat inactivated, 0.5% Roche blocking reagent, 0.1% Tween-20, DEPC H 2 O to final volume). Riboprobes were then incubated in anti-digoxigenin-AP, Fab fragments from sheep (Roche, Germany), 1:2000 in BS. Subsequently, the membrane was washed twice for 15 min in 1 × PBS and incubated with a BM purple chromogenic substrate for alkaline phosphatase (AP) (Merck, Darmstadt, Germany) until detection of the mRNA's spots. A final concentration of 500 pg/µL was taken into account.

Fluorescence In Situ Hybridization
Fluorescence in situ hybridization (FISH) experiments were conducted on cryostat sections by means of sterile solutions and materials. DEPC was added to PBS and water 1 mL/L to inactivate RNase enzymes; solutions were shaken vigorously and autoclaved.
Sections were dried for 2 h at room temperature (RT), well washed in 1 × DEPC/PBS and treated with 10 µg/µL Proteinase K (Sigma-Aldrich, St. Louis, MO, USA) 1:200 in 1 × PBS for 10 min. Proteinase K action was then inactivated by two washes in 2 mg/mL glycine, 5 min each. Sections were post fixed in 4% PFA for 20 min and well washed in 1 × PBS at RT. A pre-hybridization step was carried out in the hybridization solution (HB) (50% formamide, 25% 20 × SSC, 50 µg/mL heparin, 10 µg/mL yeast RNA, 0.1% Tween 20, 0.92% citric acid) at 55 • C for 1 h. Riboprobes were then denatured for 10 min at 80 • C, diluted to a final concentration of 500 pg/µL in HB and employed to hybridize sections, ON at 52 • C. Post-hybridization washes were carried out at 55 • C as follows: 2 × 20 min in 1 × SSC, 2 × 10 min in 0.5 × SSC, 1 × 5 min 2 × PBS at RT. Sections were blocked in BS for 1 h at RT. Afterward, sections were incubated in anti-digoxigenin-AP, Fab fragments from sheep (Roche, Germany), 1:2000 in BS for 2 h at RT. Sections were well washed in 1 × PBS. Chromogenic reaction was carried out by using Fast Red tablets (Sigma-Aldrich, St. Louis, MO, USA) in Tris buffer and incubating the slides at RT, in the dark. Signal was observed every 20 min until signal detection. After signal developing, sections were washed in 1 × PBS at RT and mounted with Fluoreshield Mounting Medium with DAPI as counterstaining for the nuclei.

Combined FISH with IF
After the detection of FISH chromogenic reaction, sections were well washed in 1 × PBS and incubated at RT for 1 h with blocking serum (normal goat serum 1:5 in 1 × PBS containing 0.1% Triton X-100 from Sigma, St. Louis, MO, USA) and subsequently with primary antisera against different neural markers (Table 3)

Image Acquisition and Processing
FISH reactions were observed and analyzed with a Zeiss AxioScope AX 1.0 microscope (Carl Zeiss, Jena, Germany) with AxioCam MC5 and AxioVision software and Leica-DM6B (Leica, Wetzlar, Germany) software. Double FISH/IHC reactions were observed and analyzed with Leica-DM6B (Leica, Wetzlar, Germany) and LasX software. Digital raw images were optimized for image resolution, contrast, evenness of illumination, and background using Adobe Photoshop CC 2018 (Adobe Systems, San Jose, CA, USA). Anatomical structures were identified according to the adult N. furzeri brain atlas [49].

Conclusions
This manuscript reports the age-related analysis of two collagen genes, col4a1 and col25a1, in the brain of the short-lived teleost N. furzeri. We discovered that age affects both the quantitative expression and neuroanatomical distribution of the genes, with a higher and wider expression in old brains. In depth analysis of the phenotype of cells expressing col4a1 and col25a1 mRNAs in the brains of old animals showed that they were preferentially expressed in the neuronal lineage (DCX, HuC/HuD, MAP2 immunoreactive cells).
Our work may stimulate further investigation on the role of col4a1 and col25a1 in the onset of neurodegenerative diseases, and more in general in the aging process. Additionally, by further characterizing the aging phenotype of killifish brain, we contribute by adding new features in terms of similarities among killifish and mammals (e.g., structural conservation of genes and protein sequences; conservation of gene regulation, function and distribution), thus enhancing the reliability of N. furzeri as a translational model.
Future studies may address the investigation of col4a1 and col25a1 protein distribution and molecular pathways to generate Alzheimer's or Parkinson's disease models. These may be extremely useful in various research fields, from drug discovery to nanomedicine, engaged in the development of therapies for the treatment of neurodegenerative diseases.