In the adult mouse brain, neurogenesis occurs continuously in the subventricular zone (SVZ) of the lateral ventricle (Altman, 1963; Doetsch et al., 1997) and the subgranular zone of the hippocampal dentate gyrus (Seki and Arai, 1993; Eriksson et al., 1998). In the adult SVZ, type B stem cells give rise to type C transit-amplifying cells, which, in turn, produce type A neuroblasts (Doetsch, 2003). These neuroblasts migrate toward the olfactory bulb along the rostral migratory stream, where they mature into GABAergic interneurons (Lois and Alvarez-Buylla, 1994; Alvarez-Buylla et al., 2002; Kriegstein and Alvarez-Buylla, 2009). The complex microenvironment that supports this series of events is commonly referred to as the neurogenic niche. This niche consists of various cell types that surround neural stem cells (NSCs), such as neural stem and progenitor cells, ependymocytes, mature and immature neurons, and astrocytes, as well as the vasculature. The extracellular matrix (ECM) is another critical component of this niche. Notably, NSCs have been shown to contact the basement membrane (BM) of the vasculature at sites lacking astrocyte endfeet and pericyte coverage (Tavazoie et al., 2008). In addition, vascular BM NSCs also contact a local ECM structure called fractones (Figures 1A,B).

Fractones are extravascular ECM structures that are localized along the ventricular wall. These structures were initially observed through laminin immunostaining as small punctate structures of 2–5 μm, located behind the ependyma (arrow, Figures 1A–C). However, transmission electron microscopy revealed that fractones are electron-dense structures with branched morphology that allow them to contact numerous surrounding cells, including ependymocytes, astrocytes, NSC, and progenitor cells (Mercier et al., 2002, 2003). The ependymal wall contains interstitial clefts that allow the diffusion of signaling molecules from the cerebrospinal fluid (Brightman, 2002). Fractones are located at the end of these narrow channels and are ideally placed to receive growth factors and cytokines produced by the choroid plexus (Kerever et al., 2007; Mercier, 2016). Fractones first appear around postnatal day 5 and are composed of a ubiquitous BM component (Kerever et al., 2007; Nascimento et al., 2018; Sato et al., 2019). The presence of fractones rich in BM protein may participate in increasing the tissue stiffness of the neurogenic niche (Kjell et al., 2020). Recent studies have proposed ependymocytes (Nascimento et al., 2018) and GFAP-expressing cells (Sato et al., 2019) as cells that produce fractones. This suggests that the formation of fractones results from the contribution of various cells in the niche.

While fractone protein composition closely resembles that of the vascular BM, the fractone heparan sulfate (HS) composition is unique. HS chains belong to the Glycosaminoglycans (GAGs) family. GAGs are long, unbranched, hydrophilic, highly charged chains composed of repeating disaccharide units that can be classified into four groups based on their core disaccharide structure: keratan, hyaluronan, chondroitin sulfate/dermatan sulfate, and HS. Only HS chains can be found in vascular BM and fractones. N-sulfated HS chains recognized by 10E4 epitope immunoreactivity suggests that fractones HS present higher levels of sulfation than HS from the vascular BM (Figure 1C; Kerever et al., 2007).

Fractones are composed of ubiquitous BM components. Collagen type IV, the most abundant component of the BM, forms a network-like structure and is linked to a network of laminins with the help of nidogen/entactin (Pozzi et al., 2017). In addition, fractones contain two major types of heparan sulfate proteoglycans (HSPGs), perlecan and agrin (Figure 1D). The other main member of the BM type of HSPG, collagen XVIII, remains undetected in either fractones or vascular BM in the SVZ (Kerever et al., 2007).

Fractones may play various roles in the neurogenic niche through laminin-integrin interactions (Shen et al., 2008; Nascimento et al., 2018; Sato et al., 2019) by regulating heparin-binding ligand availability (Kerever et al., 2007) in the niche and promoting growth factor signaling (Douet et al., 2012; Kerever et al., 2014; Mercier and Douet, 2014). NSC interaction with laminin through α 6 β 1 integrin expressed on its cell surface is essential to maintain NSC quiescence (Shen et al., 2008). Laminins are heterotrimeric glycoproteins composed of 1 α, 1 β, and 1 γ chains. The β and γ chains coil around the α chain to form a cross-like structure with three short and one long arms. Short arms are responsible for self-polymerization and interactions with other BM molecules (Hohenester and Yurchenco, 2013). The longer arm most notably interacts with integrin and dystroglycan on the cell surface, leading to cytoskeleton rearrangement and impacting cell behavior. Various laminin isoforms can be found in fractones and the vascular BM in the neurogenic niche. While laminin α1 and γ2 are absent from both fractones and the vascular BM, laminin α5, β1/2, and γ1 are present in both. In addition, laminin α2 and α4 are present only in the vascular BM, but α3 is only present in fractones (Kerever et al., 2007; Nascimento et al., 2018; Sato et al., 2019). Other glycoproteins such as secreted modular calcium-binding protein 1 and 2 (SMOC1/2) and the laminin-related molecule netrin 4, which shares homology with the N-terminal portion of laminin β1, (Sun et al., 2011) have also been detected in fractones (Sato et al., 2019).

The capacity of fractones to specifically capture heparin-binding growth factors from the extracellular milieu highlights a critical role for its HSPG perlecan and agrin. Perlecan, also referred to as HSPG 2, is a major BM type of HSPG, and deficiency of perlecan causes perinatal lethality in mice and humans (Arikawa-Hirasawa et al., 1999, 2001). The core protein is composed of five distinct domains and interacts with a variety of molecules from the ECM (laminin, nidogen, collagen IV; Figure 1D). Through its protein core and its HS chains, perlecan is involved in numerous biological processes, including embryonic development, tissue homeostasis, and pathology (Gubbiotti et al., 2017; Yu et al., 2017). In the neurogenic niche, perlecan is present in both vascular BM and fractones, and we previously reported that the presence of perlecan in fractones through its HS chains promoted FGF-2 stimulation of neurogenesis (Kerever et al., 2014).

Agrin is another major HSPG component of the BM, and plays a critical role in the hematopoietic stem cell niche (Mazzon et al., 2011; Pozzi et al., 2017). Agrin is also a key component of the microenvironment that regulates synapse differentiation at the neuromuscular junction (Gautam et al., 1996) and in neurons of the hippocampus (Böse et al., 2000), as well as in newborn neurons of the olfactory bulb (Burk et al., 2012). In the SVZ, agrin is found in both the vascular BM and fractones (Kerever et al., 2014).

In addition to the cell/ECM interaction that plays a role in maintaining NSC quiescence through laminin/integrin signaling, fractones also play a role in regulating growth factor signaling. Both perlecan and agrin bear HS chains, and might therefore contribute to the regulation of heparin-binding growth factor signaling in the SVZ.

The building blocks of HS are glucuronic acid (GlcA) and N-acetylglucosamine (GlcNAc). They are alternatively transferred to a linker composed of one xylose residue, two galactose residues, and one GlcA residue. This polysaccharide subsequently undergoes extensive modification in the Golgi apparatus, which is catalyzed by a series of enzymes. First, N-deacetylase/N-sulfotransferase (NDST) acts on a subset of GlcNAc residues to produce N-sulfated glucosamine (GlcNS). This enzyme also generates a small number of N-unsubstituted glucosamine residues due to incomplete N-sulfation. Then, a glucuronyl C5-epimerase (GLCE) acts on the GlcA residue to create Iduronic acid (IdoA), followed by the action of HS 2-sulfotransferases that catalyze the transfer of a sulfate to the C2-position of uronic acid residues. Subsequently, HS 6-sulfotransferases (HS6ST) catalyze the transfer of sulfate onto the C6 position of the glucosamine residue in HS. Finally, HS 3-sulfotransferases (HS3ST) can transfer sulfate to the 3-OH position of the glucosamine residues of HS (Sugahara and Kitagawa, 2002; Kreuger and Kjellén, 2012). Upon release into the extracellular space, secreted endosulfatase (Sulf1 and Sulf2) can catalyze the removal of a subset of 6O sulfated group from the HS chains. The modification reactions in heparan sulfate biosynthesis occur in clusters along the polysaccharide, resulting in a highly sulfated region (S domain) separated by regions devoid of sulfate (NA domain) (Figure 2). Together, these steps contribute to the formation of a sulfated polysaccharide with tremendous chemical heterogeneity, allowing HS chains to specifically interact with a wide range of molecules (Annaval et al., 2020).

Binding of HS to a ligand may impact signaling in various ways, thereby key stem cell function (Ravikumar et al., 2020). In addition, extracellular modification of HS sulfation by endosulfatase can actively modulate growth factor signaling.

Fractones have been shown to be altered under some pathological conditions. In the SVZ of BTBR t + tf/J mice, a mouse model of autism spectrum disorder, fractones were reported to be drastically reduced in size and number (Mercier et al., 2011, 2012). In contrast, the size of fractones has been reported to increase following long-term hydrocephalus in adult mice (Campos-Ordoñez et al., 2014). In addition, we previously reported that the structure and composition of fractones were altered in aged mouse SVZ (Kerever et al., 2015; Yamada et al., 2017). With aging, neurogenesis declines, (Maslov et al., 2004) and the neurogenic niche undergoes structural and functional remodeling (Luo et al., 2006; Rojas-Vázquez et al., 2021). Ependymal cells present altered morphology, their number decreases while the number of astrocytes interposed between ependymocytes increases (Luo et al., 2008; Capilla-Gonzalez et al., 2014). Additionally, the blood brain barrier in SVZ capillaries is compromised, leading circulating pro-inflammatory molecules to potentially affect the niche (Obermeier et al., 2013; Segarra et al., 2021).

Fractone size gradually increases with age (Kerever et al., 2015; Nascimento et al., 2018). In addition, the HS composition of fractones was also modified in the aged SVZ. We previously reported on disaccharide analysis of the young and aged SVZ that total 6O-sulfation decreased in the aged SVZ. This loss of 6O-sulfation was accompanied by impaired FGF-2 signaling (Yamada et al., 2017). The aged fractones also displayed reduced immunoreactivity for the N-sulfated epitope 10E4 (Kerever et al., 2015) and reduced AO4B08 immunoreactivity (Kerever et al., 2021). This suggests dramatic changes in the sulfation of fractones HS, and these changes may affect growth activity and participate in the age-related decline of neurogenesis.

As we have described, minute modification of sulfation along the HS chains leads to dramatic changes in HS regulation of growth factor signaling. Therefore, it is critical to identify strategies to reveal precise HS sequences. The biochemical approach, which consists of breaking apart the HS chains into disaccharide units, is helpful to obtain broad information pertaining to HS composition, but it fails to elucidate the actual organization within the HS chain and cannot provide information on HS heterogeneity in the microenvironment. As we previously reported in the context of the SVZ, immunoreactivity for 10E4 shows that HS displays dramatically different sulfation signatures in the vascular BM than in fractones. The development of specific anti-HS antibodies is a great tool for deciphering the HS code (van Kuppevelt et al., 1998; Dennissen et al., 2002; Thompson et al., 2009), and studying the influence of local changes in HS composition on cell signaling in health and pathological conditions.

AK and EA-H conceptualized the manuscript. Both authors drafted the manuscript and have reviewed and agreed with the publication of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.