­Glial and stem cell expression of Fibroblast Growth Factor Receptor 1 in the embryonic and perinatal nervous system

12 Background. Fibroblast growth factors (FGFs) and their receptors (FGFRs) are involved in the 13 development and function of multiple organs and organ systems, including the central nervous 14 system (CNS). FGF signaling via FGFR1, one of the three FGFRs expressed in the CNS, 15 stimulates proliferation of stem cells during prenatal and postnatal neurogenesis and participates 16 in regulating cell-type ratios in many developing regions of the brain. Anomalies in FGFR1 17 signaling have been implicated in certain neuropsychiatric disorders. Fgfr1 expression has been 18 shown, via in situ hybridization, to vary spatially and temporally throughout embryonic and 19 postnatal development of the brain. However, in situ hybridization lacks sufficient resolution to 20 identify which cell-types directly participate in FGF signaling. Furthermore, because antibodies 21 raised against FGFR1 commonly cross-react with other members of the FGFR family, 22 immunocytochemistry is not alone sufficient to accurately document Fgfr1 expression. Here, we 23 elucidate the identity of Fgfr1 expressing cells in both the embryonic and perinatal mouse brain. 24 Methods. To do this, we utilized a tgFGFR1-EGFPGP338Gsat BAC line (tgFgfr1-EGFP+) 25 obtained from the GENSAT project. The tgFgfr1-EGFP+ line expresses EGFP under the control 26 of a Fgfr1 promoter, thereby causing cells endogenously expressing Fgfr1 to also present a 27 positive GFP signal. Through simple immunostaining using GFP antibodies and cell-type 28 specific antibodies, we were able to accurately determine the cell-type of Fgfr1 expressing cells. 29 Results. This technique revealed Fgfr1 expression in proliferative zones containing BLBP+ 30 radial glial stem cells, such as the cortical and hippocampal ventricular zones, and cerebellar 31 anlage of E14.5 mice, in addition to DCX+ neuroblasts. Furthermore, our data reveal Fgfr1 32 expression in proliferative zones containing BLBP+ cells of the anterior midline, hippocampus, 33 cortex, hypothalamus, and cerebellum of P0.5 mice, in addition to the early-formed GFAP+ 34 astrocytes of the anterior midline. 35 Discussion. Understanding when during development and where Fgfr1 is expressed is critical to 36 improving our understanding of its function during neurodevelopment as well as in the mature 37 CNS. This information may one day provide an avenue of discovery towards understanding the 38 involvement of aberrant FGF signaling in neuropsychiatric disorders. 39 40

previously showed that PV+ interneurons did not colocalize with GFP+ cells, suggesting the loss 128 of PV+ interneurons in mutants with inactivated Fgfr1 occurs in a non-cell-autonomous manner 129 (Smith et al., 2014). In the present study, we aimed to validate the sites of Fgfr1 expression 130 reported in previous studies that utilized in situ hybridization in embryonic mice, and then to 131 extend those studies by determining which cell types in the developing mouse CNS express 132 Fgfr1. These data were obtained by using immunocytochemistry to detect GFP using reliable 133 anti-GFP antibodies. Furthermore, we aimed to establish the utility of the tgFgfr1-EGFP+ line 134 as a tool for future investigations of FGFR1 during embryonic development. To do this research, 135 we utilized embryonic day 14.5 (E14.5) and postnatal day 0.5 (P0.5) mice from the tgFgfr1-136 EGFP+ line, along with immunostaining and fluorescence microscopy. 137 Presently, we show that in E14.5 tgFgfr1-EGFP+ mice there is co-localization of GFP and brain 138 lipid-binding protein (BLBP) in the radial glial cells of the VZ, cortical midline, lateral 139 ganglionic eminence (LGE), lateral pallial-subpallial boundary, and cerebellar anlage. 140 Furthermore, we also find co-localization of GFP and doublecortin (DCX) in neuroblasts of the 141 developing cortex and hippocampal primordium of E14.5 tgFgfr1-EGFP+ mice. We also 142 investigated cell type expression of Fgfr1 in P0.5 tgFgfr1-EGFP+ mice. Our studies indicate co-143 localization of GFP with BLBP+ radial glia in the VZ of the cortex, hippocampus, 144 hypothalamus, and glial wedge, and also throughout the cerebellum. Co-localization with GFP 145 can also be seen in astrocytes positive for glial fibrillary acidic protein (GFAP) within the corpus 146 callosum, glial wedge, and indusium griseum, and in NeuN+ neurons within the anterior 147 cingulate cortex. 148 This construct allowed us to map and identify the cells expressing Fgfr1 in the developing CNS. 155

Materials and Methods
The GENSAT project has the aim of creating a library of transgenic BAC lines in order to map 156 the most important genes in the CNS (Heintz, 2004). Control mice (tgFgfr1-EGFP-) lacked the 157 EGFP transgene and were thus EGFP negative. All animals used in this study were treated 158 humanely and ethically in accordance to the recommendations from The Guide for the Care and 159 Use of Laboratory Animals of the National Institutes of Health. All euthanasia was conducted 160 ethically and humanely to minimize suffering, as outlined under the University of Louisiana at 161 Lafayette IACUC committee APS number 2013-8717-053. 162

Genotyping 163
In order to determine if the mice contained the tgFGR1-EGFPGP338Gsat gene construct, 164 we conducted polymerase chain reactions (PCR) for EGFP or by screening with goggles that 165 contained a GFP filter (BLS LTD). The following procedure for PCR based genotyping was 166 used: tails of mice were collected and DNA was extracted from the tail using 50mM sodium 167 hydroxide (95°C for 30 minutes), followed by neutralization with 1M TRIS (pH 7.6). Master mix 168 for 1 reaction of PCR for amplifying EGFP was created using 2.5µl of 10x PCR buffer, 0. Embryos were obtained from timed pregnancies based on identification of a vaginal mucus plug 176 (Embryonic Date=0.5 on morning found). Pregnant dams were euthanized by CO 2 inhalation at 177 E14.5, the uterine horn was then removed and placed in cold Hank's balanced salts solution phosphate buffered saline (PBS) overnight at 4°C, followed by cryoprotection in 20% 180 sucrose/1X PBS at 4°C overnight. P0.5 mice were anesthetized by being placed on a metal plate 181 on ice for 10-15 minutes to induce hypothermia followed by decapitation. P0.5 brains were 182 dissected and fixed as described for embryos. E14.5 and P0.5 mouse brains were cryopreserved 183 in OCT within cryomolds and placed in a dry ice/ethanol bath. We then stored all samples at -184 80°C. Embryonic and P0.5 brains were thin sectioned (20µm onto superfrost slides) in a cryostat 185 (Microm, HM 505 E) and stored at -80°C until immunostaining. 186 Slides were removed from -80°C and allowed to equilibrate to room temperature. We then 187 removed any excess OCT from the slides. Using an ImmEdge™ pen (Vector Laboratories, Inc.) 188 we circumscribed each tissue section and allowed it to dry. We then blocked the tissue by adding 189 a droplet of 10% normal goat serum (NGS) in 1xPBS with 0.01% Tween (Sigma Aldrich) and 190 0.02% TritonX (Sigma Aldrich). The appropriate primary antibodies in 5% NGS in 1xPBS with 191 0.01% Tween (Sigma Aldrich) and 0.02% TritonX (Sigma Aldrich) ( Table 1) were added and 192 allowed to incubate overnight at 4°C. The sections were washed, and primary antibodies detected 193 with Alexa conjugated secondary antibodies (Jackson labs and Abcam) in 5% NGS in 1xPBS 194 with 0.01% Tween and 0.02% TritonX and allowed to incubate for 2 hours, followed by a wash 195 with 1xPBS. Coverslips were placed on the slides with VECTASHIELD DAPI as the mounting 196 medium. 197

Tissue Clearing 198
After sacrificing the pups to obtain brain samples, the paws were processed for tissue clearing 199 Axioimager M2 microscope equipped with an ApoTome.2. Tissue cleared using the CLARITY 204 technique was imaged as described above. In order to determine the presence of double labeled 205 cells in immunostained tissue, single and z-stack images were obtained and observed as separate 206 channels as well as in composite images. Fluorescence microscopy using a Nikon SMZ18 stereo 207 microscope was used to image paws at low magnification. 208

209
Fgfr1 is expressed in the developing E14.5 and perinatal P0.5 mouse brain 210 Fgfr1 expression at E14.5, evident by a positive GFP signal, can be seen in the VZ and 211 upper layers of the developing dorsal and cingulate cortex, the VZ and intermediate zone of the 212 hippocampal primordium, the choroid plexus, and the hypothalamus (Fig. 1A). By using the 213 tgFgfr1-EGFP+ mouse line, Fgfr1 expression at P0.5, as apparent by a positive GFP signal in 214 immunostained tgFgfr1-EGFP+ mice, can be seen in the VZ surrounding the lateral ventricles, 215 anterior cingulate cortex, indusium griseum, glial wedge, and in the lateral pallial-subpallial 216 boundary (Fig. 1B). Fgfr1 expression can also be observed in regions other than the brain, such 217 as the postnatal growth plates of the distal limbs (Figs. 1C and 1D) and in the apparent Müller 218 glia of the eyes as well as in the lens (Fig. 1E). Since we expected to see GFP in the growth 219 plates, and a GFP signal was observed in all four paws when we imaged the pups under the 220 fluorescence dissection microscope (Fig. 1C), we performed optical clearing of the dissected 221 limbs according to the protocol by Chen et al, 2013 and imaged the cleared tissue to observe 222 expression (Fig. 1D). GFP in the glial wedge (Fig. 6A), but not in the indusium griseum (Fig. 6B). Immunostaining for 283 GFAP+ astrocytes and Fgfr1 expressing GFP+ cells revealed co-localization of GFAP+ 284 astrocytes with GFP in the glial wedge of tgFgfr1-EGFP+ mice (Fig. 6D). There was also co-285 localization of GFAP+ astrocytes with GFP in the corpus callosum and indusium griseum (Fig.  286   6E) of P0.5 tgFgfr1-EGFP+ mice immunostained for GFAP and GFP. All control mice 287 (tgFgfr1-EGFP-) presented little to no GFP+ signal (Figs. 6C and 6F). 288 Immunostaining for NeuN+ neurons and Fgfr1 expressing GFP+ cells of P0.5 tgFgfr1-EGFP+ 289 mice, revealed co-localization of NeuN and GFP within cells of the anterior cingulate cortex 290 (Figs. 7A-7C). In contrast to what was observed with GFAP staining, when NeuN and GFP 291 immunostaining was examined in the indusium griseum of P0.5 tgFgfr1-EGFP+ mice, we found little to no indication of NeuN+ neurons within the indusium, however Fgfr1 expressing GFP+ 293 cells were found to be present (Figs. 7D-7F). 294

Fgfr1 expression in the P0.5 perinatal hippocampus and cortex 295
We examined Fgfr1 expression in the P0.5 hippocampi of tgFgfr1-EGFP+ mice immunostained 296 for GFAP+ astrocytes and BLBP+ radial glia. At P0.5, GFAP positive immunostaining has not 297 attained the same level as adult brains, and most GFAP signal is observed in the midline glial 298 structures (Figs. 6D-6F). GFAP and GFP immunostaining revealed that GFAP+ astrocytes were 299 not found in the hippocampus; notably, however, there were Fgfr1 expressing GFP+ cells in the 300 hippocampus and cortex (Figs. 8A-8C). Immunostaining for BLBP and GFP in the hippocampus 301 showed co-localization of BLBP+ radial glia with GFP in the VZ and cornu ammonis (CA) 302 region of P0.5 tgFgfr1-EGFP+ mice (Figs. 8E-8G); however, there was no observable BLBP 303 signal detected in the dentate gyrus at this age (DG not shown in figure). BLBP and GFP 304 immunostaining of P0.5 tgFgfr1-EGFP+ mice revealed a high occurrence of co-localization of 305 GFP within BLBP+ radial glia throughout the layers of the cortex (Figs. 8I-8K). All control 306 mice (tgFgfr1-EGFP-) presented little to no GFP+ signal (Figs. 8D, 8H, and 8L). 307 Fgfr1 is expressed in BLBP+ radial glia of the hypothalamus, cerebellum, and lateral 308 pallial-subpallial boundary at P0.5 309 To further our investigation of Fgfr1 expression in BLBP+ cells of the perinatal mouse brain, we 310 examined BLBP+ glial cells and Fgfr1 expressing GFP+ cells in the hypothalamus, cerebellum, 311 and lateral pallial-subpallial boundary of P0.5 tgFgfr1-EGFP+ mice. Here we show that BLBP+ 312 radial glia along the hypothalamic VZ co-localized with GFP, as evident by the immunostaining 313 for BLBP and GFP in P0.5 tgFgfr1-EGFP+ mice (Figs. 9A-9C). Furthermore, we also show co-314 localization of BLBP and GFP within the perinatal cerebellum (Figs. 9D-9F)  callosum were GFP positive ( Fig. 6B and 6E). These may represent the fibers of neurons in the 416 anterior cingulate that express Fgfr1 (Fig. 7). Previous studies found that the pioneer axons of 417 the corpus callosum arise from the anterior cingulate (Brian G. Rash & Richards, 2001). Thus, it 418 is reasonable to postulate that FGFR1 signaling is also present within the axons that pioneer the 419 anterior midline; however, Synapsin Cre mediated inactivation of Fgfr1 did not result in axon 420 guidance defects (Smith et al., 2006). 421 Here, we find   (GFP) and red (BLBP) channels, respectively), and in the lateral pallial-subpallial boundary (G-6 I, with G and H showing separated green (GFP) and red (BLBP) channels, respectively). All 7 scale bars = 50µm. 8