Early Dendritic Morphogenesis of Adult-Born Dentate Granule Cells Is Regulated by FHL2

Dentate granule cells (DGCs), the progeny of neural stem cells (NSCs) in the sub-granular zone of the dentate gyrus (DG), must develop and functionally integrate with the mature cohort of neurons in order to maintain critical hippocampal functions throughout adulthood. Dysregulation in the continuum of DGC development can result in aberrant morphology and disrupted functional maturation, impairing neuroplasticity of the network. Yet, the molecular underpinnings of the signaling involved in adult-born DGC maturation including dendritic growth, which correlates with functional integration, remains incompletely understood. Given the high metabolic activity in the dentate gyrus (DG) required to achieve continuous neurogenesis, we investigated the potential regulatory role of a cellular metabolism-linked gene recently implicated in NSC cycling and neuroblast migration, called Four and a half LIM domain 2 (FHL2). The FHL2 protein modulates numerous pathways related to proliferation, migration, survival and cytoskeletal rearrangement in peripheral tissues, interacting with the machinery of the sphingosine-1-phosphate pathway, also known to be highly active especially in the hippocampus. Yet, the potential relevance of FHL2 to adult-born DGC development remains unknown. To elucidate the role of FHL2 in DGC development in the adult brain, we first confirmed the endogenous expression of FHL2 in NSCs and new granule cells within the DG, then engineered viral vectors for genetic manipulation experiments, investigating morphological changes in early stages of DGC development. Overexpression of FHL2 during early DGC development resulted in marked sprouting and branching of dendrites, while silencing of FHL2 increased dendritic length. Together, these findings suggest a novel role of FHL2 in adult-born DGC morphological maturation, which may open up a new line of investigation regarding the relevance of this gene in physiology and pathologies of the hippocampus such as mesial temporal lobe epilepsy (MTLE).


INTRODUCTION
The discovery of mammalian adult hippocampal neurogenesis has been fundamental to our understanding of physiological homeostasis in the adult brain (Altman and Das, 1965). Most recently, studies of cellular metabolism and pharmacology have shown that lipid metabolism and signaling play an active role in adult neural stem cell cycling and are important for synaptic homeostasis and functional regulation of newborn neurons (Kanno and Nishizaki, 2011;Riganti et al., 2016;Beckervordersandforth et al., 2017;Stessin et al., 2017;Callihan et al., 2018). In particular, sphingosine-1-phosphate (S1P), well-known for its effects on cellular survival, proliferation, and motility in cancer biology and immunology, has recently been found to signal in the hippocampal dentate gyrus via various S1P receptors (S1PRs), controlling levels of neurogenesis (Matloubian et al., 2004;Milstien and Spiegel, 2006;Pyne and Pyne, 2010;Efstathopoulos et al., 2015;De Marco et al., 2017). However, the downstream effectors and intracellular machinery involved in the regulation of adult hippocampal neurogenesis remain largely unexplored.
Interestingly, a previous unbiased transcriptional analysis of human patients with mesial temporal lobe epilepsy (MTLE), a condition characterized by aberrant hippocampal discharge related to dysregulated neurogenesis, revealed a candidate gene linked to the S1P pathway that was markedly downregulated in hippocampi of patients with MTLE versus those without MTLE (Jamali et al., 2006;Parent and Murphy, 2008;Sierra et al., 2015). The gene in question is Four and a Half LIM domain protein 2 (FHL2), which codes for an adapter protein that mediates a variety of protein-protein interactions. FHL2 modulates numerous signaling pathways including cell proliferation and migration, with its conserved zinc finger motif allowing it to serve as a central hub for diverse signaling pathways to converge (Wixler et al., 2000). For instance, in cell culture S1P triggers RhoA GTPase activation followed by nuclear translocation of FHL2, where it acts as a transcriptional co-regulator for a number of genes including those involved in molecular memory and extracellular communication (Muller et al., 2002). Despite the finding of changes in FHL2 expression in subjects with MTLE (Jamali et al., 2006), surprisingly little follow-up on mechanistic characterization has taken place, though the authors speculated that the physical association of FHL2 to a potassium channel subunit may predispose these patients to epilepsy via a channelopathy. One recent study identified an alternative action of FHL2 expressed in adult neural stem cells (NSCs) that may help explain its clinical phenotype and physiological importance in the brain. Kim et al. reported that global FHL2 deletion in mice via transgenic constitutive knockout led to low self-renewal activity among NSCs, premature differentiation into astrocytes at the expense of neuronal differentiation and delayed neuroblast migration in the developing brain (Kim et al., 2019).
To determine the specific role of FHL2 in adult hippocampal neurogenesis while avoiding developmental effects from constitutive expression and non-specific effects from communicating adjacent cells, we thus aimed to carry out cell-autonomous and temporally controlled manipulation of FHL2 expression in adult-born dentate granule cells (DGCs). Through virally mediated genetic silencing and overexpression studies, we found that FHL2 is critical for dendrite growth in the early phase of newborn neuron development in the adult dentate gyrus.

MATERIALS AND METHODS
All experimental and surgical procedures were approved by the Stony Brook University Animal Use Committee and followed the guidelines of the National Institutes of Health (NIH).

Animals
Eight-week-old male and female C57BL/6 wild type mice (Charles River Laboratories) were used in the protocol as approved by the Institutional Care and Animal Use Committee of Stony Brook University. All mice were housed in a cage on a 12-h light/dark cycle and were provided ad libitum access to food and water.

Doxycycline Induction
Transgenes were activated with the administration of doxycycline in drinking water. The drinking water included 400 mg doxycycline dissolved in 200 mL distilled water plus 30 g of sucrose.

Surgeries
All surgeries were performed aseptically. Before surgery mice were first weighed and anesthetized with 200 mg/kg of ketamine/xylazine administered intraperitoneally. Retrovirus was infused, or lentivirus and retrovirus (lenti-GFAP-Cre (retro-Flex-reverse-eGFP or retro-Flex-reverse-shFHL2) were co-infused into the DG (0.5 µL/injection site) at stereotactic coordinates -2.0 mm from bregma, ±1.6 mm lateral, -0.25 mm ventral, and -3.0 mm from bregma, ±2.6 mm lateral, -0.32 mm ventral. Mice were monitored during the intraoperative and postoperative duration and a heating pad (37 C) was provided for 2 h during the recovery phase. Also, mice were given 0.05 mg/kg of buprenorphine HCl intraperitoneally as postoperative recovery analgesic.

Perfusion and Tissue Processing
Mice were anesthetized with urethane (200 µg/g) and were perfused transcardially with Phosphate Buffered Saline (PBS) and then 4% Paraformaldehyde (PFA). Brains were removed and fixed in 4% PFA for 24 h at 4 0 C and were transferred following day to 30% (w/v) sucrose solution. The brains were sectioned coronally at 60 µm on Leica microtome. Three-fourths of the sections were stored in cryopreservative (30% sucrose, 30% glycerol by weight in deionized water) until future staining and a quarter of the sections were immediately washed 3x with PBS before immunostaining.

Imaging and Quantification
All coronal sections were imaged on an Olympus FLV1000 confocal microscope. The Z stack confocal images of the different regions of the brain including DG, CA3, CA2, and CA1 regions were collected and the fluorescence intensity was measured using Image J software. The fluorescence intensity of FHL2 of the whole cell was measured by taking the difference between the fluorescent intensity of the cell and background signal from the same image. The morphometric profile used for the neuron selection for analysis includes bipolar feature of the anatomical phase of the cell, the elliptical cell body and approximately 10 µm size of the cell (Stanfield and Cowan, 1979;Claiborne et al., 1990;Danzer et al., 2008) and immunolabeling with Prox1 marker. Dendritic tracings were constructed using Imaris software (Oxford Instruments). Sholl analysis was performed using ImageJ software, dendritic branches were measured from soma at 5 µm interval.

Statistical Analyses
Data were analyzed with independent samples t-tests and one-way ANOVA followed by post hoc Tukey HSD test.
Two-tailed values of α < 0.05 were considered the cutoff for statistical significance. All data are represented as mean ± SEM. N represents the number of animals unless otherwise specified.

Endogenous Expression of FHL2 in the Adult Hippocampus
In order to study the regulatory role of FHL2 in the adult brain, we first mapped the expression of FHL2 in the subfields of the adult hippocampus [dentate gyrus (DG), CA3, CA2, and CA1] (Figures 1A-D). We observed wholecell fluorescence intensity of FHL2 throughout the neurons of these regions, with no significant differences between regions CA3, CA2 and CA1 (p > 0.05), however, significant differential expression was observed between the sub-granular zone (SGZ) and granular cell layer (GCL) (p < 0.001), CA3 and GCL (p = 0.019), CA2 and SGZ (p = 0.001), CA1 and GCL (p = 0.016) ( Figure 1E). Thus, FHL2 is expressed throughout the adult hippocampus, consistent with previous observations in the embryonic central nervous system (CNS) (Kudo et al., 2007).

Expression of FHL2 in Neural Stem Cells, Proliferative Cells and Mature Granule Cells
Next, we sought to determine whether the expression of FHL2 changes as a function of developmental time for maturing cells in the DG. We measured the number of FHL2 positive cells in the population of cells expressing Hes5, a marker of neural stem cells (NSCs), Ki67, a proliferative marker, and NeuN, a marker of differentiated neurons in the granule cell layer of the DG (Figure 2A). The majority of Hes5 + cells (73.5 ± 0.75%) and Ki67 + cells (79.6 ± 1.13%) positive cells express FHL2 under basal conditions, while the expression of FHL2 is significantly lower in NeuN + neurons (26.7 ± 0.79%) ( Figure 2B). Next, we birthdated the dividing cells by injecting the thymidine analog BrdU intraperitoneally and euthanized the animals at 1, 5, and 14 days after injection and quantified the number of cells positive for FHL2 expression over three-time points (Figures 2C,D). We found that the percentage of FHL2 positive cells declined over time, day 1 (71.1 ± 2.6%), 5 days (67.0 ± 1.97%), and 14 days (37.5 ± 0.65%) ( Figure 2E). Thus, our data suggest based on the expression pattern of FHL2 in the majority of stem cells and proliferative cells and decreased expression in neurons as they mature that FHL2 may play a regulatory role in modulating the early phase of the cell programming.

Overexpression of FHL2 Expands Dendritic Outgrowth in Adult-Born Dentate Granule Cells
Dendritic elaboration in adult-born DGCs begins at approximately 1 week after birth, with this phase in DGC maturation accelerating by 2 weeks (Zhao et al., 2006;Goncalves et al., 2016;Wang et al., 2019), at a developmental time when we observed endogenous FHL2 expression (Figure 3). This led us to hypothesize that FHL2 may play a role in dendrite outgrowth during a critical period in the initial development of nascent DGCs. To determine whether FHL2 affects the dendritic growth of developing DGCs, we induced overexpression of FHL2 in the dividing hippocampal cells at either 3 or 5dpi as illustrated in Figure 4A and sacrificed mice at 7 and 14 dpi. As a control, a retroviral vector carrying only an RFP tag was injected, and mice were treated with doxycycline from day 3 and sacrificed at 7 or 14 dpi (Figures 4A,B). Among cells overexpressing FHL2 from 3 dpi, such overexpression expanded dendrite length and branching at 7 dpi, an effect that persisted at 14 dpi (Figures 4C-F). Interestingly, in cells that had been induced at 5dpi, there was a similar increase in total dendrite length and branching at 7dpi, however, this effect was short-lived, reverting to the wildtype phenotype by 14dpi (Figures 4C-F). Sholl analysis revealed increased dendritic complexity in the experimental groups at 7 and 14 dpi as compared to the control group (Figures 4G,H). In addition, we wondered whether FHL2 overexpression would affect axonal length at 7 and 14 dpi, but found no significant differences at any time point compared to control (Figure 4I). Representative cell traces in the various conditions are shown in Figure 4J. Together, these data demonstrate that FHL2 overexpression expands dendritic outgrowth in newborn DGCs at an early developmental stage.

Loss of FHL2 Increases Early DGC Dendritic Length
Given that FHL2 expression, especially during an early critical window immediately after fate determination, is important for dendritic arborization, we next sought to understand the implications of its loss on the maturation in DGCs, which may ultimately influence their functional integration (Kuipers et al., 2009;Ma et al., 2009;Krueppel et al., 2011;Bergami et al., 2015). To this end, we first co-injected lenti-GFAP-Cre and retro-shFHL2-DF-rGFP in the hippocampus to target active radial glia-like (RGL) cells for subsequent FHL2 downregulation ( Figure 5A). We induced expression at 3 dpi, as we had found this early time point to be critical for dendritic elaboration (Figure 4). Loss of FHL2 in newborn DGCs did not affect dendritic branching, but significantly increased dendritic length (Figures 5B-F). These results suggest that loss of FHL2 in newborn neurons has a distinct yet overlapping effect as gain. Of note, we examined the effect of FHL2 downregulation at 14 dpi as well, but were unable to establish a clear, reproducible phenotype (data not shown). Taken as a whole, early expression of FHL2 in adult-born DGCs appears to regulate appropriate early morphological growth, ensuring that dendrites do not "overshoot", with implications for these cells' proper integration.

DISCUSSION
Dendritic development, a process essential for DGC maturation and plasticity of the hippocampal network, is tightly regulated by molecular machinery during the course of dentate granule cell growth. Several mechanisms regulate the initiation, formation, and maintenance of dendritic arborization such as microtubule nucleation and microtubule-associated motor protein (Sharp et al., 1997;Delandre et al., 2016), secretory pathways involving the Golgi apparatus (Rao et al., 2018) and endoplasmic reticulum (Cui-Wang et al., 2012), and synaptic scaffolding proteins (Vessey and Karra, 2007) suggesting that the sculpting and maintenance of the dendritic tree has many layers of complexity. Although the dysregulation of dendritic development is reported in MTLE (Curia et al., 2014) the precise mechanism underlying the pathophysiology of MTLE remains uncertain. Interestingly FHL2, a gene linked to cellular metabolism has been found to be dysregulated in MTLE, although the reason for this is unknown (Tran et al., 2016). In this study, we found that manipulation of FHL2 disrupts dendritic modeling in newborn DGCs in the hippocampus by causing dendritic hypertrophy, a phenotypic recapitulation of neurons of patients with MTLE which points to a potential underlying mechanism (Ryufuku et al., 2011).
FHL2, a LIM domain protein has a unique affinity for various intracellular proteins and receptors such as in the context of regulation of bone marrow-derived dendritic cell migration through its Sphingosine 1-phosphate receptor (S1PR1) interaction (König et al., 2010). FHL2 also interacts with a microtubule-associated protein 1 light chain 3 (LC3) regulating the development of skeletal muscle cells (Liu et al., 2019), and with β-catenin to upregulate its transactivation activity in cancer cells (Wei et al., 2003). Thus, FHL2 participates in various protein complexes regulating cell growth, differentiation and cytoskeletal remodeling as reported in immune cells and in the pathogenesis of numerous cancers (Canault et al., 2006;Wixler et al., 2007;Li et al., 2008;Qian et al., 2009;Kurakula et al., 2014). Although several studies have shed light on the relevance of FHL2 in terms of CNS-related pathology, including epilepsy and glioblastoma (Jamali et al., 2006;Kleiber et al., 2007;Li et al., 2008;Okamoto et al., 2013), the understanding of the regulatory role of FHL2 is yet to be fully appreciated. Recently, Kim et al. (2019) reported that FHL2 is expressed by NSCs and plays a role in both fate determination of cells as well as the migratory capabilities of their progeny cells (Kim et al., 2019). To assess the physiological implication of FHL2 on the dynamic process of dendritic arborization, we report for the first time that FHL2 is expressed endogenously in adult-born hippocampal neurons and interestingly that there is a significant decline in its expression observed as these neurons mature, suggesting the participation of FHL2 during the early development of adult-born neurons. We used viral vector-mediated manipulation of FHL2 expression at the selected time point to avoid any confounding effect on fate determination (Kumamoto et al., 2012) and successfully targeted the cells committed to the neuronal lineage (Figure 3).
Here we found the paradoxical effects of upregulation and downregulation of FHL2, which both increased the dendritic growth during the early developmental phase of DGCs. One plausible reason could be that levels of FHL2 required for proper dendritic outgrowth follow a "U shaped curve", with higher or lower levels both leading to loss of inhibition of dendritic expansion, although this remains to be tested. We also did not observe a consistent phenotype of downregulation of FHL2 on dendritic morphogenesis at the later developmental time point of 14 dpi (data not shown), which may relate to effects of low levels of FHL2 on cell viability at the early survival/synaptic integration stage. Nevertheless, our findings suggest a contribution of early expression of FHL2 in DGC morphogenesis, which may be explained by various signaling partner interactions.
One candidate pathway by which FHL2 may participate in DGC dendritic morphogenesis is the sphingolipid signaling pathway. Based on FHL2 interactome data (Tran et al., 2016), Sphingosine Kinase (SphK) is one of the metabolism-linked interacting partners of FHL2 and regulates the availability of Sphingosine 1 Phosphate (S1P), a bioactive lipid metabolite, reported to be involved in cell proliferation, survival, migration and cytoskeletal remodeling (Spiegel and Milstien, 2003). S1P exerts its downstream effects by binding to one of the substrate receptors, S1PR1, highly expressed on new adult-born DGCs to modulate neurite growth (Toman et al., 2004;Hayashi et al., 2009;Yang et al., 2020). Yet, whether FHL2 participates in sphingolipid signaling in this context will require further molecular characterization. Alternatively, the influence of FHL2 on dendritic modeling could be explained through interactions with the Wnt/β-catenin signaling pathway (Martin et al., 2002), which has been shown to refine dendrites in newborn DGCs (Kumamoto et al., 2012) and is also dysregulated in some forms of epilepsy (Hodges and Lugo, 2018).
In this study, we focused on the acute effects of FHL2 expression on early DGC development, however, there may be differential effects of chronic upregulation and blockade of FHL2 expression on dendritic growth in the adult brain, which will require further exploration. Based on the diverse interaction profile of FHL2 and its affinity to form proteinprotein complexes as part of a signal interaction node, there remains much to be gleaned regarding the mechanisms governing dendritic modeling as well as dendritic hypertrophy and synaptic perturbation reported in MTLE (Ryufuku et al., 2011;Han et al., 2016;Janz et al., 2018). While we limited the scope of the present study to focus on dendritic alteration, FHL2 may play roles in other aspects of DGC morphogenesis and physiology. In summary, we have shown that FHL2 is an important regulator of early dendritic morphogenesis in adultborn hippocampal neurons. This study serves as a new line of evidence demonstrating the biological relevance of FHL2 in the central nervous system.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by Stony Brook University Animal Use Committee.

AUTHOR CONTRIBUTIONS
JW and SG conceived of the idea. JW engineered the vectors. AA, JW, and GK conducted the experiments. AA and GK wrote the initial draft. All authors analyzed the data and agreed with the final version of the manuscript.

FUNDING
This work was supported by the National Institutes of Health (Grants NS089770, AG046875, and NS104858 to SG and 1F30MH110103 to GK).