Electrophysiological and gene expression characterization of the ontogeny of nestin-expressing cells in the adult mouse midbrain

Stem Cell Research 23 (2017) 143–153

Contents lists available at ScienceDirect

Stem Cell Research journal homepage: www.elsevier.com/locate/scr

Electrophysiological and gene expression characterization of the ontogeny of nestin-expressing cells in the adult mouse midbrain Anupama Dey 1, Parisa Farzanehfar 1, Elena V. Gazina, Tim D. Aumann ⁎ Florey Institute of Neuroscience & Mental Health, The University of Melbourne, Parkville, Victoria 3010, Australia

a r t i c l e

i n f o

Article history: Received 23 December 2016 Received in revised form 19 May 2017 Accepted 1 July 2017 Available online 4 July 2017 Keywords: Substantia nigra Ventral tegmental area Dopamine Parkinson's disease Single-cell RT-qPCR

a b s t r a c t The birth of new neurons, or neurogenesis, in the adult midbrain is important for progressing dopamine cellreplacement therapies for Parkinson's disease. Most studies suggest newborn cells remain undifferentiated or differentiate into glia within the adult midbrain. However, some studies suggest nestin + neural precursor cells (NPCs) have a propensity to generate new neurons here. We sought to confirm this by administering tamoxifen to adult NesCreERT2/R26eYFP transgenic mice, which permanently labelled adult nestin-expressing cells and their progeny with enhanced yellow fluorescent protein (eYFP). eYFP+ midbrain cells were then characterized 1–32 weeks later in acutely prepared brain slices using whole-cell patch clamp electrophysiology combined with single-cell RT-qPCR. Most eYFP+ cells exhibited a mature neuronal phenotype with large amplitude fast action potentials (APs), spontaneous post-synaptic currents (sPSCs), and expression of ‘mature’ neuronal genes (NeuN, Gad1, Gad2 and/or VGLUT2). This was the case even at the earliest time-point following tamoxifen (i.e. 1 week). In comparison to neighboring eYFP− (control) cells, eYFP+ cells discharged more APs per unit current injection, and had faster AP time-to-peak, hyperpolarized resting membrane potential, smaller membrane capacitance and shorter duration sPSCs. eYFP+ cells were also differentiated from eYFP− cells by increased expression of ‘immature’ pro-neuronal genes (Pax6, Ngn2 and/or Msx1). However, further analyses failed to reveal evidence of a place of birth, neuronal differentiation, maturation and integration indicative of classical neurogenesis. Thus our findings do not support the notion that nestin + NPCs in the adult SNc and midbrain generate new neurons via classical neurogenesis. Rather, they raise the possibility that mature neurons express nestin under unknown circumstances, and that this is associated with altered physiology and gene expression. © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction The motor symptoms of Parkinson's disease (PD; e.g. tremor, bradykinesia, postural instability) are caused by loss of dopamine (DA) signaling in the caudate putamen (CPu) brought about by degeneration of DA neurons in the substantia nigra pars compacta (SNc). We know this because killing or interfering with SNc DA neurons in rodent and nonhuman primate models of PD produce similar motor dysfunctions (Gubellini and Kachidian, 2015) and administering drugs that elevate DA signaling or transplanting DA neurons into SNc or CPu normalizes movement in these models (Duty and Jenner, 2011; Bjorklund and Lindvall, 2017; Redmond et al., 2010). Indeed these same drugs are currently frontline treatments for PD where they effectively alleviate motor symptoms early in treatment but become less effective and produce debilitating side-effects, probably due to the unphysiological and

⁎ Corresponding author. E-mail addresses: anupa.dey@florey.edu.au (A. Dey), parisa.farzanehfar@florey.edu.au (P. Farzanehfar), elena.gazina@florey.edu.au (E.V. Gazina), timothy.aumann@florey.edu.au (T.D. Aumann). 1 Authors contributed equally to this work

untargeted DA signaling they induce (Barker et al., 2015). This has led many to believe the key to longer-lasting benefits with fewer sideeffects is replacing SNc DA neurons, either by cell transplantation or by stimulating endogenous DA neurogenesis (Barker et al., 2015). Both of these cell-replacement approaches would benefit from an adult midbrain microenvironment that is conducive for DA neurogenesis. Unfortunately this does not appear to be the case. In adult rodents, SNc cells rendered bromodeoxyuridine-positive (BrdU+; a marker of dividing cells) remain either undifferentiated or differentiate into glia, not neurons (Aponso et al., 2008; Chen et al., 2005; Cooper and Isacson, 2004; Lie et al., 2002; Shan et al., 2006; Klaissle et al., 2012; Worlitzer et al., 2013, but see Zhao et al., 2003).On the other hand there is evidence that some progenitor cells in the rodent midbrain have neurogenic capacity. Retinoic acidinduced differentiation of cells isolated from adult rat SNc and cultured in the presence of fibroblast growth factors (FGF2 or FGF8) generate βtubulin III+ cells (neurons) in vitro (Lie et al., 2002). Moreover, these same cells become NeuN + (neurons) following transplantation into the hippocampus, an established neurogenic niché, but not when transplanted back into SNc of adult rats (Lie et al., 2002). Lie et al. (Lie et al., 2002) speculated that these cells are nestin-expressing neural

http://dx.doi.org/10.1016/j.scr.2017.07.001 1873-5061/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

144

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

progenitor cells (NPCs) and Shan et al. (Shan et al., 2006) reported evidence that nestin + cells can indeed generate new neurons, including DA neurons, within the microenvironment of the adult mouse midbrain. If nestin + cells do generate new neurons and DA neurons within the microenvironment of the adult midbrain, knowledge about their ontogenesis will be crucial to identify signaling mechanisms regulating neurogenesis and DA neurogenesis here, which might help progress cell-replacement therapies for PD. Hence, the aims of this study were to: (1) assess whether cells derived from nestin + cells in the adult midbrain have a neuronal phenotype as defined by electrophysiology and gene expression; and (2) if so, assess whether they achieved this via classical neurogenesis.

2. Materials and methods All experimental procedures on animals were approved by the Howard Florey Institute Animal Ethics Committee and are in accordance with the National Health & Medical Research Council of Australia's published code of practice for the care and use of animals for scientific purposes, 7th edition 2004.

2.1. Mice NesCreERT2 (lines 5.1 & 4) C57BL/6 mice were obtained with permission from Professor Ryoichiro Kageyama and Kyoto University Institute for Virus Research (53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 6068507, Japan) (Imayoshi et al., 2006). These mice express a tamoxifeninducible form of Cre-recombinase (CreERT2) under the control of a 5.8-kb fragment of the promoter region and a 1.8-kb fragment of the second intron of the rat nestin gene. This second intron fragment contains a neural stem cell/progenitor-specific enhancer (Mignone et al., 2004; Zimmerman et al., 1994). Activation of CreERT2 is achieved by administering tamoxifen, an estrogen receptor agonist/antagonist, to the mice. NesCreERT2 mice were crossed with R26eYFP C57BL/6 reporter mice obtained with permission from Professor Frank Costantini (Columbia University Medical Center, New York) (Srinivas et al., 2001). R26eYFP reporter mice have a LoxP-flanked DNA STOP codon upstream of their reporter sequence (eYFP), which prevents reporter expression in the absence of active Cre-recombinase (i.e. absence of nestin promoter/ enhancer activity and absence of tamoxifen). If on the other hand the DNA STOP codon is removed by active Cre-recombinase (i.e. presence of nestin promoter/enhancer activity and presence of tamoxifen) eYFP expression is driven by the constitutively and ubiquitously active ROSA26 gene locus. All experimental NesCreERT2/R26eYFP mice used were F1 generation obtained via NesCreERT2 homozygous (male or female) and R26eYFP homozygous (male or female) crosses. The majority (85%) of cells included in the analyses were harvested from Line 5.1 NesCreERT2/R26eYFP mice. Line 5.1 is more specific (i.e. less tamoxofenindependent recombination CNS-wide) but less efficient at labelling nestin + cells in the adult hippocampal SGZ than line 4 (Imayoshi et al., 2006). Thus line 5.1 is the better to avoid false-positive labelling. Thus, to identify and lineage trace nestin + cells in the adult midbrain, adult (≥8-weeks old) NesCreERT2/R26eYFP male and female mice were administered a ‘pulse’ of tamoxifen (10 mg/day in 0.5 ml corn oil via oral gavage) for 3–4 consecutive days, which permanently labels cells with concurrent nestin promoter/enhancer activity (i.e. Nes gene expression) with the protein product of eYFP (i.e. enhanced yellow fluorescent protein or eYFP). Note that this tamoxifen dose is similar to that used in other studies on these mice (Sun et al., 2014; Sakamoto et al., 2011). Importantly, because the eYFP transgene is constitutively induced by a gene recombination event, eYFP protein will also be permanently expressed in any progeny of cells rendered eYFP+ at the time of tamoxifen administration (i.e. eYFP + cells that divide any time after tamoxifen).

2.2. Electrophysiology At different times (1–32 weeks) following administration of the tamoxifen ‘pulse’ NesCreERT2/R26eYFP mice were anesthetized with isofluorane in air then decapitated. The brain was rapidly (b 1 min) removed and placed in ice-cold (0 °C) “cutting mix” containing 125 mM NaCl, 25 mM NaHCO3, 3 mM KCl, 1.25 mM NaH2PO4·H2O, 1 mM CaCl2, 6 mM MgCl2, 25 mM glucose, bubbled with 95%O2 and 5%CO2, pH 7.4. Slices (300 μm thick) were cut in the coronal plane with a vibratome through the midbrain and transferred into 35 °C artificial cerebrospinal fluid (ACSF) containing 125 mM NaCl, 25 mM NaHCO3, 3 mM KCl, 1.25 mM NaH2PO4·H2O, 2 mM CaCl2, 1 mM MgCl2, 25 mM glucose, bubbled with 95%O2 and 5%CO2, pH 7.4. A slice was transferred into a bath perfused with 30 °C ACSF and individual eYFP+ and neighboring eYFP− (control) cells were viewed at high power (×63 objective) using fluorescence and infrared differential interference contrast (IR-DIC) microscopy. Whole-cell recordings were made with glass micropipettes (~1 μm tip diameter, ~6–10 MΩ resistance) containing ~ 6 μl 144 mM K-Gluconate, 3 mM MgCl2.6H2O, 10 mM HEPES, 0.5 mM EGTA, pH 7.2, osmolarity 290. Membrane current and voltage were recorded and controlled using an Axoclamp 2B amplifier (3 kHz bandwidth) and Clampex 9.0 software (Molecular Devices LLC, CA, USA). Data were digitized at 10 kHz using an Axon Digidata 1550 (Molecular Devices LLC, CA, USA). 2.3. Single-cell RT-qPCR 2.3.1. Cell samples Following electrophysiological characterization and with the glass micropipette still tightly attached (i.e. GΩ resistance) to the cell's membrane, a b 1 femtolitre sample of its cytoplasm was aspirated into the tip of the pipette using negative pressure and under visual guidance. The pipette was removed from the slice and its outer surface cleared of any cellular debris by rapid removal from the bath solution (the tip was re-examined under the microscope following this to ensure no cellular debris remained). The broken micropipette tip and its entire contents (~6 μl of internal pipette solution plus cell sample) were collected in a 200 μl eppendorf tube and immediately placed at − 80 °C for later RT, pre-amplification and qPCR. Care was taken to avoid contamination of cell samples with extraneous RNA and DNA. The capillary glass used to pull micropipettes was autoclaved; the micropipette holder and wire were sterilized under ultraviolet light and bleach solution, respectively; the internal pipette solution was autoclaved; the syringe used to fill the pipette was sterilized under ultraviolet light; the eppendorf tube into which the aspirate was placed was sterile; and the electrophysiologist wore gloves and used sterile technique throughout the cell-sampling procedures. Cell-negative controls were collected in exactly the same way as described above except the micropipette tip was not sealed onto the cell membrane, no whole-cell electrophysiology was performed, and no cell cytoplasm was aspirated. Rather, the micropipette tip was placed near a cell and left there for the same amount of time it takes to seal, record and aspirate (approximately 10 min) before being withdrawn. All cell samples collected on days where cell-negative controls returned positive gene expression readings were excluded from analyses. RTqPCR-negative and RT-qPCR-positive controls were performed on aliquots of RT buffer-only samples and RNA harvested from tissue dissected from developing mouse midbrain, respectively. 2.3.2. Reverse transcription To perform first-strand cDNA synthesis the eppendorf tube containing a cell sample was thawed, briefly centrifuged and kept on ice until the RT reaction (~30 min.). First, the volume of the aspirate was measured and made up to 7.5 μl with 66.67 mM Tris-HCl, 100 mM KCl, 4 mM MgCl2, pH 8.3. Using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen™, catalog #18080-051), 0.5 μl of random

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

145

hexamers (50 ng/μl) was added and incubated 5 min at 65 °C then 1 min on ice. Next, 2 μl of RT mix [0.5 μl RNaseOUT (40 U/μl), 0.5 μl of 10 mM dNTP mix, 0.5 μl of 0.1 M DTT and 0.5 μl of superscript III RT (200 U/μl)] was added and the total 10 μl reaction volume was incubated for 10 min at 25 °C, 60 min at 42 °C and 5 min at 85 °C.

process extending from one or both poles. The long axis of these cells was usually aligned ventrolaterally. In the PAG, VTA and SNc the morphology of eYFP+ cells was variable with soma sizes ranging in diameter from 10 to 40 μm and shapes from round/oval with no processes to multipolar with several long processes (e.g. Fig. 1E-H).

2.3.3. Pre-amplification We used Single Cell-to-CT Kit (Ambion®, catalog #4458236). First, 0.2xpooled TaqMan Gene Expression Assays for our target mRNAs of interest (see Supplementary Table 1 for catalog numbers of the probes used for each gene) were prepared using 1x TE buffer, pH 8.0. Next, 3.22 μl of PreAmp Mix and 3.87 μl of 0.2x pooled TaqMan Gene Expression Assays were added to each RT sample then placed in a thermal cycler where it was held at 95 °C for 10 min, cycled (20 cycles) at 95 °C for 15 s, then 60 °C for 4 min and finally held at 99 °C for 10 min.

3.2. Electrophysiology

2.3.4. qPCR Single Cell-to-CT Kit (Ambion®, catalog #4458236) was again used. First, a 1:20 dilution of pre-amplified products was prepared using 1x TE buffer, pH 8.0. Next, 50 μl of mix (25 μl of 2x TaqMan Gene Expression Master Mix, 2.5 μl of 20x TaqMan Gene Expression Assay, 12.5 μl of Nuclease-free water and 10 μl of 1:20 diluted pre-amplified products) was loaded into a Micro Amp Fast Optical 96-well (0.1 ml) reaction plate with barcode and placed into a ViiA7 PCR machine (Applied Biosystems). The reaction was held at 50 °C for 2 min, then 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s then 60 °C for 1 min. 2.3.5. Data analysis Electrophysiological signals were analyzed using Clampfit (version 10.1; Molecular Devices LLC, CA, USA). Statistical analyses included ttests (for comparisons of two independent groups), ANOVAs (for comparisons of more than two independent groups), Chi-square (for comparisons of two or more independent proportions) and principle component analysis (PCA). Prior to PCA, each datapoint was normalized by dividing by its population standard deviation. This ensured similar variability across all parameters and therefore equal likelihood of contributing to the outcome. PCA was performed using XLSTAT® (Addinsoft SARL, http://www.xlstat.com/). Data are reported as mean ± SEM unless otherwise indicated. Graphical representations were performed using GraphPad prism (GraphPad Software Inc., La Jolla, CA, USA). 3. Results 3.1. eYFP+ cell location and morphology within midbrain At all times (1–32 weeks) following administration of a ‘pulse’ of tamoxifen (10 mg/day in 0.5 ml corn oil via oral gavage for 3–4 consecutive days) to adult (≥8-week old) NesCreERT2/R26eYFP mice eYFP+ cells were most prominent in the lining of the cerebroventricular aqueduct (Aq) and in the ventral midline of the midbrain between the Aq and the interpeduncular nuclei; specifically in the rostral linear nucleus of the raphae (RLi), the interfascicular nucleus (IF) and the paranigral nucleus (PN). Much lower densities of eYFP+ cells were present throughout the remainder of the midbrain including in the periaqueductal gray (PAG), the ventral tegmental nucleus (VTA) and SNc. Note that approximately equal numbers of male and female mice were used in these experiments and we saw no sex differences in cell location, morphology, electrophysiology or gene expression. Thus data from males and females has been combined throughout the manuscript. The morphology of eYFP+ cells varied according to anatomical location in midbrain. In the Aq lining it was uniform, comprising a simple oval or columnar shape with no processes apart from several cilia protruding into the ventricular space (e.g. Fig. 1A,B). In the slice preparation these cilia could be seen beating rapidly. In the ventral midline eYFP+ cells also exhibited a relatively uniform morphology (e.g. Fig. 1C,D), which was often fusiform or bipolar shaped, sometimes with a long

Whole-cell patch clamp recordings were made from eYFP + and neighboring eYFP − (control) cells in midbrain slices cut from NesCreERT2/R26eYFP mice 1–32 weeks after they had received a ‘pulse’ of tamoxifen (10 mg/day in 0.5 ml corn oil via oral gavage for 3–4 consecutive days) when they were ≥8-weeks of age. Recorded cells were located in the Aq lining, PAG, ventral midline, VTA and SNc. Twohundred and six eYFP+ and 222 eYFP− cells were included in the following analyses based on a N 1GΩ seal, access resistance b20 MΩ, low holding current at −70 mV (−60pA on average) and stability of these parameters throughout the recording period (~10 min.). 3.2.1. eYFP+ cells are predominantly neuronal We first asked to what extent are eYFP+ cells neuronal. Of the n = 206 eYFP+ midbrain cells analyzed 165 (80%) were ‘neuronal’ in that they exhibited action potentials (APs). Of these, 69 (42%) also exhibited spontaneous post-synaptic currents (sPSCs). Thirty-one/206 (15%) eYFP+ midbrain cells exhibited a ‘glial-like’ electrophysiology comprising relatively hyperpolarized resting membrane potential (RMP), low membrane capacitance (Cm), low membrane resistance (Rm), no voltage-sensitive dynamic membrane currents (such as APs) and no sPSCs. The majority of these glial-like cells were located in the Aq lining; indeed almost every (26/27, 96%) eYFP+ cell in the Aq lining was glial-like; the one exception exhibited a hybrid ‘glial-neuronal’ phenotype comprising hyperpolarized RMP, low Cm and Rm, but low amplitude and long duration APs (no sPSCs). The remaining 5/206 (2%) glial-like cells were located in the midbrain parenchyma. A small number of eYFP+ cells in the midbrain parenchyma (which includes some of the n = 165 ‘neuronal’ cells above with APs) exhibited hybrid ‘glial-neuronal’ phenotypes such as no APs but depolarized RMP (N − 56 mV; 6/206, 3%), or low amplitude (b 30.3 mV) and long duration (N 2.4 ms) APs (3/206, 1%). 3.2.2. Comparison of midbrain eYFP+ and eYFP− cell electrophysiology We next asked whether eYFP+ cells differ from neighboring eYFP− (control) cells. It is important to recognize that the eYFP− population could include cells that expressed nestin in adulthood, just not during the short 3–4 day period of tamoxifen administration. They nevertheless represent a good point of comparison against eYFP+ cells because they were recorded at the same time and under the same conditions. First we removed all Aq cells (n = 27 eYFP+ and n = 6 control) from these comparisons because they would have introduced a sampling bias [Aq cells are almost exclusively ‘glial-like’ and most (27/33, 82%) of the recorded population were eYFP+]. Comparisons between the remaining n = 179 eYFP+ and n = 216 eYFP− midbrain parenchymal cells revealed the following. The most remarkable differences were: (1) a 3-fold increase in input-output slope (number of APs per unit current injection); and (2) a 2-fold decrease in AP time-to-peak in eYFP+ cells (Table 1). Other significant differences were: (1) more negative RMP; (2) smaller Cm; and (3) shorter sPSC duration (including shorter sPSC decay tau) (Table 1) in eYFP+ cells. All other parameters were not significantly different, i.e. Rm, % of cells with APs, AP threshold, AP amplitude, AP duration, % cells with sPSCs, sPSC frequency and sPSC amplitude (Table 1). These data were also analyzed using principal components analysis (PCA). Input parameters were: RMP, Cm, Rm, input-output slope, AP amplitude, AP duration, sPSC frequency, sPSC amplitude and sPSC duration. In Fig. 2 the first two principal components account for 43% of the data variability. eYFP+ cells are plotted as red-colored squares and control cells as blue-colored crosses. To aid interpretation of this plot

146

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

principal component 1 (PC1, the X-axis) is best correlated (positively) with AP amplitude and principal component 2 (PC2, Y-axis) is best correlated (positively) with input-output slope. Thus moving left-to-right in

the X-dimension can be thought of as increasing AP amplitude and moving up-to-down in the Y-dimension can be thought of as decreasing input-output slope. In Fig. 2, three main clusters of cells are evident:

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

147

Table 1 Comparisons of the electrophysiology of eYFP+ and eYFP− (control) cells in the midbrain parenchyma of adult mice (Aq cells excluded). Where appropriate, values given are 25th percentile b median b 75th percentile, mean ± SD or percentages. Significant differences (p b 0.05; unpaired t-test, Mann-Whitney rank sum test, or chi-square test) are highlighted in red font with p-values in parentheses.

(1) A spatially localized cluster of n = 28 eYFP+ and n = 5 eYFP− cells circumscribed by the small red-colored circle in the lower left quadrant. This represents the ‘glial-like’ phenotype and mostly comprises Aq cells [unlike the comparison in Table 1 Aq cells were included in this PCA because it nicely illustrated the difference in electrophysiology between Aq (glial-like) and parenchymal (mostly neuronal) cells]. (2) An adjacent ‘smear’ of n = 106 eYFP+ and n = 114 eYFP− cells circumscribed by the long narrow gray-colored ellipse passing from the lower left quadrant through the upper left quadrant and finishing in the upper right quadrant. (3) A third cluster of n = 68 eYFP + and n = 91 eYFP − cells circumscribed by the blue-colored shape and tracking back down from the upper right quadrant to the lower right quadrant. Analysis of the extent to which eYFP+ and eYFP− cells overlap in Fig. 2 revealed a tendency for eYFP− cells to distribute more to the right side of the plot than eYFP+ cells. However, while there is a statistically significant (χ2 = 19.5, p b 0.01, 2d.f.) difference in the proportion of the

two different cell-types across the 3 different clusters (i.e. a higher proportion of eYFP− cells as one moves from clusters 1 to 3), there is no significant difference if one excludes the ‘glial-like’ cluster from the analysis (i.e. compares the proportions of the different cells in clusters 2 & 3 only). 3.3. Single-cell RT-qPCR One hundred and four (n = 68 eYFP + & n = 36 eYFP −) cells characterized electrophysiologically were analyzed also for their gene expression on a cell-by-cell basis. Inclusion criteria were: (1) cellnegative controls (see Materials and Methods) collected on the same day were negative for gene expression; and (2) cells were positive for housekeeping gene expression (ACTB). The genes examined are listed along the top of Table 2 (grouped by function) and the catalog numbers of the probes used for each gene are listed in Supplementary Table 1. Positive gene expression was defined using the Auto-Ct function in

Fig. 1. eYFP+ cell morphology in acutely prepared living midbrain slices. Each row shows the same region imaged with infrared differential interference contrast (IR-DIC) optics (left) and eYFP fluorescence microscopy (right). A–B eYFP+ cells lining the cerebroventricular aqueduct. Note the relatively small size of these cells and their simple oval or columnar morphology with no obvious processes apart from multiple cilia (arrows) protruding into the ventricular space. C–D eYFP+ cell in the ventral midline. Note again the relatively small size and simple round morphology with no obvious processes. Note also the micropipette (p) attached to this and the cells in E–H below, which was used to make electrophysiological recordings and to sample some of the cell's cytoplasm for single-cell qPCR. E–F eYFP+ cell in substantia nigra pars compacta (SNc). This cell is relatively small with simple morphology and no obvious processes, which was common for eYFP+ cells here and in the adjacent ventral tegmental area (VTA). G–H Example of another commonly observed eYFP+ cell-type in VTA and SNc, which was much larger with more complex morphology reminiscent of a mature neuron. Note the large process (dendrite) extending upward from this cell (arrow).

148

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

Fig. 2. Principal component analysis of the electrophysiology of eYFP+ (red squares) and control (eYFP−, blue crosses) midbrain cells reveals substantial overlap between the two populations (i.e. their electrophysiology is similar). Input parameters were: RMP, Cm, Rm, input-output slope, AP amplitude, AP duration, sPSC frequency, sPSC amplitude and sPSC duration (see Materials and Methods for further details). Plotted are the first two principal components (PC1 & PC2), which account for 43% of the variability. PC1 is best correlated (positively) with AP amplitude and PC2 is best correlated (positively) with input-output slope. Three distinct clusters of cells have been highlighted with outlines and colored according to the dominant cell-type within (the precise number of eYFP+ (red font) and control (blue font) cells in each cluster are also shown). The small round cluster circumscribed by the red-colored circle in the lower left quadrant comprises cells with a ‘glial-like’ electrophysiology. These are mostly located in the ependymal lining of the Aq but some are located in the midbrain parenchyma. The remaining cells are predominantly neuronal with a mixture of immature and mature electrophysiological phenotypes. There is a tendency as one moves from the ‘glial-like’ cluster up and to the right through the cells circumscribed by the gray-colored elipse then down and further to the right through the blue-colored shape for AP amplitude to increase and for cell excitability (i.e. input-output slope) to increase then decrease. Chi-squared analysis detected statistical differences between clusters (see text).

the qPCR software (i.e. 50–75% up the amplification profile in exponential plot view) and was detected at 28.19 ± 3.89 qPCR cycles (mean ± SD all positive genes except housekeeping). Note this was following 20 cycles of cDNA pre-amplification. By comparison positive housekeeping gene (ACTB) expression was detected at 25.53 ± 2.64 qPCR cycles. Negative gene expression was defined as not reaching Auto-Ct threshold by ≥40 qPCR cycles. 3.3.1. Comparison of eYFP+ and eYFP− midbrain cell gene expression Qualitative gene expression data for these n = 68 eYFP+ and n = 36 eYFP− cells are summarized in Table 2 (red squares = genes that were expressed, white squares = genes that were not expressed, black squares = genes that were not examined). Cell identities are listed in the left-hand column with eYFP + (yellow) cells grouped in the top half and eYFP− cells grouped in the bottom half. The percentage of cells in each population that were positive for each gene is detailed below these. The key findings are: (1) Ki67, a marker of cell proliferation, was being expressed (at the time of investigation) by the same proportion of cells in both populations [4/46 (9%) eYFP +, 1/36 (3%) eYFP −; χ2 = 1.2, p N 0.10, 1d.f.]. (2) Nes was being expressed (at the time of investigation) by the same proportion of cells in both populations [4/46 (9%) eYFP +, 3/36 (8%) eYFP −; χ2 = 0.003, p N 0.10, 1d.f.]. (3) More eYFP + than eYFP − cells expressed the ‘immature’ pro-neuronal genes Pax6, Ngn2 & Msx1. For example a significantly higher proportion of eYFP + cells expressed one or more of Pax6, Ngn2 or Msx1 [36/68 (53%) eYFP +, 6/36 (17%) eYFP −; χ2 = 12.9, p b 0.01, 1d.f.]. (4) The majority of cells in both populations expressed the ‘mature’ neuronal genes NeuN [32/61 (52%) eYFP +, 22/36 (61%) eYFP −; χ 2 = 0.7, p N 0.1, 1d.f.], Gad1 [58/61 (95%) eYFP+, 30/36 (83%) eYFP−; χ2 = 3.7, p N 0.05, 1d.f.], Gad2 [36/61 (59%) eYFP+, 16/36 (44%) eYFP−; χ2 = 1.9, p N 0.10, 1d.f.], and/or VGLUT2 [43/61 (70%) eYFP+, 24/36 (67%) eYFP−; χ2 = 0.2, p N 0.10, 1d.f.]. (5) Most eYFP+ but very few eYFP− cells co-expressed mature neuronal and pro-neuronal genes [31/61 (51%) eYFP+, 6/36 (17%)

eYFP−; χ2 = 11.2, p b 0.01, 1d.f.]. (6) There was a notable paucity of expression of pro-DA genes (Lmx1a, Lmx1b, Pitx3, En1, En2) with the exception of Nurr1 in both populations (eYFP+ and eYFP−). For example, only 3/68 (4%) eYFP+ and 5/36 (14%) eYFP− cells expressed Lmx1b [χ2 = 3.0, p N 0.05, 1d.f.], whereas 33/68 (49%) eYFP+ and 14/36 (39%) eYFP− cells expressed Nurr1 [χ2 = 0.9, p N 0.10, 1d.f.]. (7) Many cells in both populations expressed genes normally associated with mature DA neurons (TH and D2R). For example, 26/46 (57%) eYFP+ and 17/36 (47%) eYFP− cells expressed TH and/or D2R [χ2 = 0.7, p N 0.10, 1d.f.] Quantitative analyses of gene expression (relative to ACTB housekeeping gene) revealed broadly three levels of expression based on three peaks in the distribution of medians across all genes examined. We designated these high, medium and low. High levels of expression ranged between 0 and 4 qPCR cycles below the level of ACTB expression (e.g. Gad1 in Fig. 5A), medium were 5–9 qPCR cycles below ACTB (e.g. NeuN in Fig. 5A), and low were 10–13 qPCR cycles below ACTB (e.g. Pax6 in Fig. 5A). Highly expressed genes included Gad1, Gad2, VGLUT2, DCX, NCAM2, TH and D2R; medium included NeuN, Pax6, Lmx1b, Pitx3, En1, En2 and Nurr1; low included Ngn2 and Msx1. The remaining genes (Nes, Ki67, Lmx1a) were expressed in too few cells to reliably categorize. Remarkably the level of expression of each gene was almost identical in the eYFP+ and eYFP− cell populations; the only differences being En2 (high in eYFP+ and medium in eYFP−) and Nurr1 (medium in eYFP+ and high in eYFP−). Within the eYFP+ population, where we could correlate levels of gene expression with both location and time after tamoxifen, only 1/13 genes were significantly correlated with location (levels of NeuN were significantly higher in cells harvested progressively from Aq, PAG, ventral midline, VTA and SNc; p b 0.05, r2 = 0.09, n = 51) whereas 6/13 genes were significantly correlated with time. Four of these were ‘mature’ neuronal genes [Gad1 (p b 0.01, r2 = 0.16, n = 59), Gad2 (p b 0.01, r2 = 0.25, n = 42), NeuN (p b 0.05, r2 = 0.08, n = 51), and VGLUT2 (p b 0.05, r2 = 0.10, n = 44)] all of which increased with time, and 2 were ‘immature’ neuronal genes [Pax6 (p b 0.05, r2 = 0.20, n = 22) and DCX (p b 0.05, r2 = 0.24, n = 21)] both of which also increased with time. Accordingly we did not detect any relationships between the levels of expression of different genes that would be anticipated to signify classical neurogenesis (e.g. high expression of mature genes when immature genes were low and/or vice versa). These data were also analyzed using PCA. Here we examined separation between the eYFP+ and eYFP− cell populations when: (1) gene expression alone was used (Fig. 3A); (2) electrophysiology alone was used (Fig. 3B); and (3) electrophysiology and gene expression combined were used (Fig. 3C). A subset of 91 cells were included (n = 58 eYFP+ & n = 33 control) from which we measured the same gene expression and electrophysiological parameters. Input parameters for each of the corresponding PCAs were: (1) Fig. 3A: level of expression (ΔCt relative to ACTB housekeeping gene) of Pax6, Ngn2, Msx1, DCX, NCAM2, Lmx1b, Nurr1, NeuN, Gad1, Gad2 & VGLUT2; (2) Fig. 3B: RMP, Cm, Rm, input-output slope, AP amplitude, AP duration, sPSC frequency, sPSC amplitude & sPSC duration; and (3) Fig. 3C: all of the above. In the gene expression-only PCA (Fig. 3A) there is almost complete separation of eYFP + and eYFP − cells. In contrast in the electrophysiology-only PCA (Fig. 3B) the two cell-types overlap substantially (similar to the other electrophysiology-only PCA in Fig. 2). When gene expression and electrophysiology were combined (Fig. 3C) there was again almost complete separation of eYFP+ and eYFP− cells. Thus gene expression more powerfully distinguishes eYFP+ from eYFP− cells than electrophysiology. In Fig. 3C PC1 is best correlated (positively) with the level of Ngn2 expression and PC2 is best correlated (positively) with inputoutput slope (a measure of cell excitability). Thus eYFP + cells have higher expression of Ngn2 and higher excitability than eYFP− cells. 3.3.2. Electrophysiology and gene expression of eYFP+ cells is not related to their location in midbrain or their age The above data establish clear differences between eYFP + and eYFP− cells that are consistent with eYFP+ cells being more immature

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

Mature DA

11 3

14 11 8

11 39 61 83 44 67 42 28

33 61 3

TH

8

D2R

3

Gad2

20 49 52 95 59 70 43 37

NeuN

9

Gad1

En1

En2

7

Nurr1

Pitx3

VGLUT2

Mature neuronal

Midbrain development & Pro-DA Lmx1b 4

DCX

26 25 21 29 54 4

Msx1

9

8

Lmx1a

NCAM2

Migration/ plasticity

Pro-neuronal

9

Ngn2

Pax6

Precursor Nes

Proliferation

186 187 103 110 140 184 201 100 101 138 169 221 170 127 204 253 259 404 406 411 346 388 119 111 299 423 369 67

Ki67

Mature DA D2R

TH

VGLUT2

Gad2

Mature neuronal Gad1

NeuN

Nurr1

En2

En1

Pitx3

Midbrain development & Pro-DA Lmx1b

Lmx1a

Migration/ plasticity NCAM2

DCX

Pro-neuronal

Msx1

Precursor

Pax6

Ngn2

Proliferation

Nes

Table 2 (continued)

% positive % positive eYFP+ eYFP-

438 439 440 441 443 445 361 398 401 385 337 340 348 217 205 216 220 413 289 288 298 297 276 427 277 227 279 225 262 270 284 283 282 22 31 41 44 25 19 16 2 3 54 55 62 99 98 84 92 93 124 112 113 117 118 120 134 160 161 162 178 182 181 179 164 171 172 173 68 64 52 89 115 108 82 102

Ki67

Table 2 Comparisons of qualitative gene expression of eYFP+ and eYFP− (control) midbrain parenchymal cells in adult mice. Individual cells are in rows and genes in columns (grouped by putative function). eYFP+ cells (highlighted yellow) are grouped at the top of the table, eYFP− (control) cells next, and the proportion of each cell population positive for each gene is given at the bottom of the table. Red indicates the gene was expressed, white indicates the gene was not expressed, black indicates the gene was not assayed.

149

(see Discussion). To this extent eYFP+ cells could derive from nestin + cells following a ‘classical’ neurogenic trajectory. To investigate this possibility further we next analyzed the electrophysiology and gene expression of eYFP+ cells with respect to their location in midbrain and with respect to time following tamoxifen administration (an approximation of the cell's age). Note this is only an approximation of the cell's age because if an eYFP + cell divides at a later time its progeny will also be eYFP+. If eYFP+ cells are following a classical neurogenic trajectory one might expect their electrophysiology and gene expression to differ as a function of location and/or time because there ought to be: (1) different sites of birth, migration, maturation and/or integration in the midbrain; and/or (2) a temporal sequence of birth, migration, maturation and integration. However we could find no evidence this was the case. Rather, the distributions of phenotypes exhibited by eYFP + cells in different midbrain locations or times following tamoxifen overlapped considerably. An example of this is illustrated in Fig. 4, which is the same PCA as in Fig. 3C (i.e. using combined electrophysiology and gene expression data) but with all control cells removed and all eYFP+ cells color-coded according to midbrain location (Fig. 4A) or time after tamoxifen (Fig. 4B). Other PCAs (i.e. electrophysiologyonly and gene expression-only) parsed by location or time produced the same result (data not shown). 4. Discussion The question of whether or not neurogenesis occurs in the adult SNc has been addressed many times, motivated by the prospect that replacing degenerated SNc DA neurons in PD will provide much longer motor symptom relief and fewer side effects than current treatments. Although the weight of evidence indicates there is little or no neurogenesis in adult SNc there remains cause for optimism because many cells in the adult

150

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

Fig. 4. Parsing eYFP+ cells by their location in midbrain (A) or time after tamoxifen (B, an approximation of the cell's age) reveals no differences in electrophysiology or gene expression profiles, which is inconsistent with a ‘classical’ neurogenesis trajectory for nestin-expressing cells in the adult midbrain. A & B Principal component analyses (PCA) of a subset of n = 91 (n = 58 eYFP+ & n = 33 control) midbrain cells for which a full complement of electrophysiological and gene expression data were obtained. Input data were: (1) electrophysiology, RMP, Cm, Rm, input-output slope, AP amplitude, AP duration, sPSC frequency, sPSC amplitude & sPSC duration; and (2) gene expression ΔCt (relative to housekeeping gene) of Pax6, Ngn2, Msx1, DCX, NCAM2, Lmx1b, Nurr1, NeuN, Gad1, Gad2 & VGLUT2 (see Materials and methods for further details). This is the same PCA illustrated in Fig. 3C but with all eYFP− cells removed and all eYFP+ cells colorcoded according to midbrain location (A) or time after tamoxifen (B). PC1 = principal component 1; PC2 = principal component 2. Note the spatial distributions of eYFP + cells (i.e. phenotype distributions) grouped by midbrain location or by time following tamoxifen overlap considerably. Fig. 3. Gene expression more powerfully distinguishes eYFP+ from eYFP− (control) midbrain cells than electrophysiology A–C Three different principal component analyses (PCA) of a subset of n = 91 (n = 58 eYFP+ & n = 33 control) midbrain cells for which a full complement of gene expression and electrophysiological data were obtained. Individual cells are indicated as symbols (eYFP+ red squares and control blue crosses) and the red and blue colored outlines represent clusters of eYFP+ and control cells, respectively. PC1 = principal component 1; PC2 = principal component 2. A PCA using gene expression only. Input data were ΔCt (relative to housekeeping gene) of Pax6, Ngn2, Msx1, DCX, NCAM2, Lmx1b, Nurr1, NeuN, Gad1, Gad2 & VGLUT2 (see Materials and methods for further details). Note the almost complete separation of eYFP+ and control cells. B PCA using electrophysiology only. Input data were RMP, Cm, Rm, input-output slope, AP amplitude, AP duration, sPSC frequency, sPSC amplitude & sPSC duration. Note the almost complete overlap of eYFP+ and control cells. C PCA using gene expression and electrophysiology. Input data were ΔCt (relative to housekeeping gene) of Pax6, Ngn2, Msx1, DCX, NCAM2, Lmx1b, Nurr1, NeuN, Gad1, Gad2, VGLUT2, RMP, Cm, Rm, input-output slope, AP amplitude, AP duration, sPSC frequency, sPSC amplitude & sPSC duration. Note the almost complete separation of eYFP+ and control cells.

rodent SNc incorporate the cell division marker BrdU (Aponso et al., 2008; Chen et al., 2005; Lie et al., 2002; Zhao et al., 2003; Frielingsdorf et al., 2004; Peng et al., 2008) but see Hermann et al., 2009) and some SNc cells appear capable of generating neurons in the presence of appropriate environmental cues such as FGF2 or FGF8 and retinoic acid in vitro or when transplanted into an adult neurogenic niché (the hippocampus (Lie et al., 2002)). Lie et al. (Lie et al., 2002) speculated that these are multi-potent nestin + NPCs. Indeed even within the adult rodent SNc

there is evidence of very low levels of neurogenesis from nestin + NPCs (Shan et al., 2006). Given the importance for progressing better therapies for PD we sought to confirm whether or not nestin + cells in the adult SNc and midbrain generate new neurons and DA neurons in situ using NesCreERT2/R26eYFP transgenic mice and a combination of electrophysiology combined with single-cell gene expression. We found most cells labelled with eYFP by prior (1–32 weeks) administration of a ‘pulse’ of tamoxifen to adult (≥8-weeks old) NesCreERT2 /R26eYFP mice (i.e. adult nestin + cells and their progeny and ontogeny) do indeed exhibit a mature neuronal phenotype with large amplitude fast action potentials (APs), spontaneous post-synaptic currents (sPSCs) and expression of ‘mature’ neuronal genes (NeuN, Gad1, Gad2 and/or VGLUT2). These findings imply a strong association between nestin expression and neurons in the adult midbrain. However the more important question is whether this association is through classical neurogenesis. During rodent brain development ventral mesencephalic DA neurons derive from nestin + stem cells or NPCs in the ventricular (germinal) zone of the Aq (Ling et al., 1998; Schiff et al., 2009) from where they undergo radial migration ventrolaterally to populate the VTA and SNc, differentiating into DA neurons along the way (Prakash and Wurst, 2006). In the present study the morphology and spatial distribution of eYFP + cells suggested a similar ontogenesis

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

151

Fig. 5. Distributions of the expression levels of individual (A) or functionally related (B) genes (relative to housekeeping gene ACTB) in eYFP+ (left column) and eYFP− (control; right column) cells. Three distinct levels of expression were observed in the distribution of medians across all genes examined. These are typified in A by Gad1 (high), NeuN (medium) and Pax6 (low). Note that the data in A and B reflect the overall conclusions of our study in that the levels of expression of these genes and most others (see text) are not different in eYFP+ and eYFP− (control) cells. Also, significantly more eYFP+ cells express pro-neuronal genes (Pax6, Ngn2, Msx1) than eYFP− (control) cells, whereas both cell-types express mature neuronal genes equally (NeuN, Gad1, Gad2, VGLUT2).

might be occurring in the adult midbrain. eYFP + cells lining the Aq were oval- or columnar-shaped and had several cilia protruding into and beating within the ventricular space. In these respects they are classical ependymal glial cells, which could be stem cells or latent stem cells (Alvarez-Buylla et al., 2001). We did not investigate glia-related genes in the present study so are unable to confirm at the transcriptional level the glial phenotype of cells lining the Aq. Note that this also introduces a bias away from glial cells more broadly in the present study. Immediately below the Aq in the ventral midline eYFP+ cells had a fusiform or bipolar shape often with a long process extending from one or both poles, which is reminiscent of migrating neuroblasts or type-3 cells. Also the predominant alignment of the long axis of these cells ventrolaterally suggested radial migration between Aq and VTA/SNc. Lastly, eYFP+ cells in the VTA and SNc had various sizes and shapes including large multipolar cells reminiscent of mature neurons. Shan et al. (Shan et al., 2006) reported decreasing proliferation of nestin + cells along this same anatomical trajectory (Aq to ventral midline to VTA/SNc) in mice, which is also consistent with birth in Aq followed by migration through ventral midline and maturation/integration in VTA and SNc. However, notwithstanding the above, we found very little electrophysiological and gene expression evidence that stem cells in the

midbrain Aq give rise to new cells that migrate ventrolaterally to become VTA and SNc neurons. Although the electrophysiology of eYFP+ Aq cells was glial cell- or stem cell-like (i.e. hyperpolarized RMP, no APs and no sPSCs) whereas those in the ventral midline, VTA and SNc were predominantly neuronal, we were unable to detect any electrophysiological or gene expression differences between ventral midline and VTA or SNc cells that would signify differentiating and migrating cells. Furthermore, parsing our data by time after tamoxifen revealed no differences in eYFP+ cell electrophysiology or gene expression, which is also inconsistent with a maturational trajectory indicative of classical neurogenesis. It may be that differences in time were masked by the fact that, in our mice, any later progeny of eYFP+ cells would also be eYFP+, meaning that every time bin could have comprised cells ranging in age from newborn to time after tamoxifen. We did not employ proliferation tools in this study to more definitively mark cell age. Furthermore, while time after tamoxifen may be reflective of age, in some cases cycling time may differ in different brain regions and in a cell-specific manner, therefore we cannot exclude that longer cycling time may occur in the adult midbrain. Nevertheless, even at the earliest times after tamoxifen (e.g. 1 week) eYFP+ cells were already present throughout Aq, ventral midline, VTA and SNc. It would be surprising if this was enough time for a cell to migrate from

152

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

the Aq to SNc in the adult brain. Thus, overall, our data suggest that nestin + cells lining the Aq do not give rise to new neurons in the VTA and SNc via migration through the ventral midline in the adult midbrain, as occurs during development. Alternatively nestin + cells could give rise to new neurons locally in SNc, VTA and ventral midline without migration. However the present data do not support this either because as early as 1 week after tamoxifen many eYFP + cells in these locations were large, had neural processes, mature neuronal electrophysiology, mature synaptic inputs and expressed mature neuronal genes. It is highly unlikely that these cells derived via differentiation, maturation and integration of local nestin + NPCs because this process takes much longer than 1 week in the permissive microenvironment of the adult mouse hippocampus (Zhao et al., 2006; Espósito et al., 2005) (contrasting with the inhibitory environment of the adult midbrain (Lie et al., 2002)). Thus, overall, our findings support the conclusions of the majority of studies on adult rodent midbrain neurogenesis (Aponso et al., 2008; Chen et al., 2005; Cooper and Isacson, 2004; Lie et al., 2002; Shan et al., 2006; Klaissle et al., 2012; Worlitzer et al., 2013), but see Zhao et al., 2003), namely that it does not occur, even from nestin + NPCs. Note that DeCarolis et al. (DeCarolis et al., 2013) used a different NesCreERT2 mouse line to the one used in the present study to show evidence that GLASTlineage radial glia-like cells (RGCs) but not nestin-lineage RGCs give rise to neurogenesis in the adult mouse hippocampus. This raises the question of the origin of our eYFP+ cells? We do not believe they are an artefact of leaky transgene expression (i.e. falsepositives) because: (1) eYFP+ cells in the established neurogenic nichés of our NesCreERT2/R26eYFP mice (i.e. sub-granular zone of the hippocampus and sub-ventricular zone of the lateral ventricles) bear all the hallmarks expected of adult nestin + NPCs here (see Supplementary Fig. 1A); (2) nestin gene expression is apparent in the adult mouse midbrain including in what appear to be large SNc neurons in both the Allen Brain Atlas (e.g. see http://mouse.brain-map.org/experiment/siv?id= 1387&imageId=101306479&initImage=ish&coordSystem=pixel&x= 3264&y=2388&z=1) and the GENSAT (e.g. see http://www.gensat.org/ imagenavigator.jsp?imageID=2886); (3) as far as we have investigated the expression of eYFP is entirely dependent on the presence of the NesCreERT2 transgene and is also controlled by tamoxifen in the majority (≥70%) of eYFP+ cells (see Supplementary Fig. 1B); and (4) most compellingly our eYFP+ cells were electrophysiologically and transcriptionally different from neigbouring eYFP − (control) cells and in directions consistent with a more immature neuronal phenotype, which concords with the classical neurogenic function of nestin expression. Specifically, eYFP+ cells were more excitable, had faster AP time-to-peak, hyperpolarized RMP, smaller Cm, shorter duration sPSCs and exhibited elevated expression of pro-neuronal genes (Pax6, Ngn2 & Msx1), all of which are associated with immature neurons in the established adult rodent neurogenic nichés (Belluzzi et al., 2003; Carleton et al., 2003; Filippov et al., 2003; Fukuda et al., 2003; Lai et al., 2010; Marin-Burgin et al., 2012; Schmidt-Hieber et al., 2004; Spampanato et al., 2012). Alternatively, and our data support this as the most likely explanation, eYFP + cells may arise from nestin expression by extant mature neurons. Nestin is an intermediate-sized cytoskeletal protein that is important for cell remodeling, particularly in developing and regenerating tissues. In the rodent nervous system it is expressed by the majority of mitotically active progenitors (Cattaneo and McKay, 1990; Lendahl et al., 1990) but it is down-regulated upon differentiation and replaced by other cell-specific intermediate filament proteins (e.g. neurofilament in neurons and glial fibrillary acidic protein or GFAP in astrocytes) (Steinert and Liem, 1990). However nestin expression persists in NPCs and vascular endothelial cells in accepted and emerging adult rodent neurogenic niches (Palmer et al., 2000; Migaud et al., 2010; Horner et al., 2000) and is re-expressed more widely throughout the CNS in reactive astrocytes, vascular endothelial cells and ependymal cells in response to inflammation, cellular stress or injury (Michalczyk and Ziman, 2005; Mokry and Nemecek, 1999; Frisén et al., 1995). Moreover

we know of two reports of evidence that mature neurons express nestin. The first demonstrates nestin immunoreactivity in cells with mature neuronal morphology and immunoreactivity in the basal forebrain and closely related areas in adult rat and human brain (Hendrickson et al., 2011). The second indicates a short isoform of nestin (Nes\\S) exerts a cytoprotective function in mature sensory and motor neurons in adult rats (Su et al., 2013). Nestin gene expression is regulated through enhancer elements in the second intron, which are present in the NesCreERT2 transgene employed here (Imayoshi et al., 2006). Interestingly one of these enhancer elements is midbrain specific and is thought to be regulated by nuclear hormone receptors such as Nurr1 (reviewed in (Michalczyk and Ziman, 2005)). A large proportion (49%) of our eYFP+ cells, which were verified to be mature neurons by electrophysiology and gene expression, co-expressed Nurr1 and 9% coexpressed Nes at the time they were investigated. Furthermore the fact that eYFP + cells were immature in other respects indicates that putative nestin expression by mature midbrain neurons is functionally relevant and, if not in the classical neurogenesis sense, then perhaps in association with another form(s) of cellular plasticity such as inflammation or cellular stress mentioned above, or neurite remodeling. Prominent expression of DCX & NCAM2 by our otherwise mature eYFP + neurons provides further evidence of plasticity here. DCX is a marker of NPC migration but its role may be more specifically in growth of neural processes (Friocourt et al., 2003). NCAM also regulates neurite outgrowth through cell adhesion. Although DCX protein can be detected (at low levels) in proliferating SNc cells following nigrostriatal degeneration (Worlitzer et al., 2013, but see Klaissle et al., 2012; Frielingsdorf et al., 2004; Van Kampen and Eckman, 2006) or FGF2 administration (Peng et al., 2008), these cells do not differentiate into neurons and so are unlikely to correspond to our eYFP+ cells. On the other hand polysialic acid (PSA, a binding partner for NCAM) and PSA-NCAM immunoreactivity have been reported in midbrain cells with mature neuronal morphology and TH immunoreactivity in rodents, monkeys and humans (Yoshimi et al., 2005; Nomura et al., 2000). Thus the DCX & NCAM2 gene expression we describe here might signify neurite growth and/or synaptic plasticity by mature midbrain neurons rather than neurogenesis. Neurite growth and/or synaptic plasticity in mature midbrain neurons is also indicated by persistent expression of the axon guidance ligand Netrin (e.g. see http://mouse.brain-map.org/experiment/siv?id=74511838&imageId= 74409159&initImage=ish&coordSystem=pixel&x=5376.5&y=3752. 5&z=2) and its receptor deleted in colorectal cancer or DCC (Osborne et al., 2005) in the ventral midbrain of adult rodents. See also (Frisén et al., 1995; Hendrickson et al., 2011; Su et al., 2013) for discussion on potential non-neurogenic functional roles of nestin.

5. Conclusions Our findings do not support the notion that nestin + NPCs in the adult SNc and midbrain generate new neurons via classical neurogenesis. Rather, they point to a capacity for expression of nestin along with other immature genes such as Pax6, Ngn2, & Msx1 by already mature midbrain neurons. This may be in response to alternative forms of plasticity such as neurite remodeling, inflammation or cellular stress. However given that a single mouse transgenic line was used in the present study, and the known variability in specificity and efficiency of reporter expression amongst nestin-CreERT2 mouse lines (e.g Sun et al., 2014), further evidence of nestin expression by already mature midbrain neurons gathered using alternative methods is needed to support this possibility. For example it will be important to determine whether the protein products of the Nes, Pax6, Ngn2 and Msx1 genes are present in mature midbrain neurons and what functional consequences they have. If they do signify adult neurogenesis or, more likely, other forms of plasticity further study of nestin + cells in the adult midbrain may provide important insights for progressing cell-replacement therapies for PD.

A. Dey et al. / Stem Cell Research 23 (2017) 143–153

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scr.2017.07.001. Conflicts of interest The authors declare no conflicts of interest. Acknowledgments This project was supported by grants from the National Health and Medical Research Council of Australia (Project grant 1022839), and the Bethlehem Griffiths Research Foundation (BGRF 1302). The Florey Institute of Neuroscience and Mental Health acknowledges support from the Victorian Government's Operational Infrastructure Support Grant. References Alvarez-Buylla, A., Garcia-Verdugo, J.M., Tramontin, A.D., 2001. A unified hypothesis on the lineage of neural stem cells. Nat. Rev. Neurosci. 2, 287–293 (Apr). Aponso, P.M., Faull, R.L., Connor, B., 2008. Increased progenitor cell proliferation and astrogenesis in the partial progressive 6-hydroxydopamine model of Parkinson's disease. Neuroscience 151, 1142–1153 (Feb 19). Barker, R.A., Drouin-Ouellet, J., Parmar, M., 2015. Cell-based therapies for Parkinson disease-past insights and future potential. Nat. Rev. Neurol. 11, 492–503 (Sep). Belluzzi, O., Benedusi, M., Ackman, J., LoTurco, J.J., 2003. Electrophysiological differentiation of new neurons in the olfactory bulb. J. Neurosci. 23, 10411–10418 (Nov 12). Bjorklund, A., Lindvall, O., 2017. Replacing dopamine neurons in Parkinson's disease: how did it happen? J. Parkinsons Dis. 7, S23–S33. Carleton, A., Petreanu, L.T., Lansford, R., Alvarez-Buylla, A., Lledo, P.M., 2003. Becoming a new neuron in the adult olfactory bulb. Nat. Neurosci. 6, 507–518 (May). Cattaneo, E., McKay, R., 1990. Proliferation and differentiation of neuronal stem cells regulated by nerve growth factor. Nature 347, 762–765 (Oct 25). Chen, Y., Ai, Y., Slevin, J.R., Maley, B.E., Gash, D.M., 2005. Progenitor proliferation in the adult hippocampus and substantia nigra induced by glial cell line-derived neurotrophic factor. Exp. Neurol. 196, 87–95 (Nov). Cooper, O., Isacson, O., 2004. Intrastriatal transforming growth factor alpha delivery to a model of Parkinson's disease induces proliferation and migration of endogenous adult neural progenitor cells without differentiation into dopaminergic neurons. J. Neurosci. 24, 8924–8931 (Oct 13). DeCarolis, N.A., Mechanic, M., Petrik, D., Carlton, A., Ables, J.L., Malhotra, S., Bachoo, R., Gotz, M., Lagace, D.C., Eisch, A.J., 2013. In vivo contribution of nestin- and GLASTlineage cells to adult hippocampal neurogenesis. Hippocampus 23, 708–719 (Aug). Duty, S., Jenner, P., 2011. Animal models of Parkinson's disease: a source of novel treatments and clues to the cause of the disease. Br. J. Pharmacol. 164, 1357–1391 (Oct). Espósito, M.S., Piatti, V.C., Laplagne, D.A., Morgenstern, N.A., Ferrari, C.C., Pitossi, F.J., Schinder, A.F., 2005. Neuronal differentiation in the adult hippocampus recapitulates embryonic development. J. Neurosci. 25, 10074–10086. Filippov, V., Kronenberg, G., Pivneva, T., Reuter, K., Steiner, B., Wang, L.P., Yamaguchi, M., Kettenmann, H., Kempermann, G., 2003. Subpopulation of nestin-expressing progenitor cells in the adult murine hippocampus shows electrophysiological and morphological characteristics of astrocytes. Mol. Cell. Neurosci. 23, 373–382 (July). Frielingsdorf, H., Schwarz, K., Brundin, P., Mohapel, P., 2004. No evidence for new dopaminergic neurons in the adult mammalian substantia nigra. Proc. Natl. Acad. Sci. U. S. A. 101, 10177–10182 (Jul 6). Friocourt, G., Koulakoff, A., Chafey, P., Boucher, D., Fauchereau, F., Chelly, J., Francis, F., Jun 2003. Doublecortin functions at the extremities of growing neuronal processes. Cereb. Cortex 13, 620–626. Frisén, J., Johansson, C.B., Török, C., Risling, M., Lendahl, U., 1995. Rapid, widespread, and longlasting induction of nestin contributes to the generation of glial scar tissue after CNS injury. J. Cell Biol. 131, 453–464 (Oct). Fukuda, S., Kato, F., Tozuka, Y., Yamaguchi, M., Miyamoto, Y., Hisatsune, T., 2003. Two distinct subpopulations of nestin-positive cells in adult mouse dentate gyrus. J. Neurosci. 23, 9357–9366 (Oct 15). Gubellini, P., Kachidian, P., 2015. Animal models of Parkinson's disease: an updated overview. Rev. Neurol. (Paris) 171, 750–761 (Nov). Hendrickson, M.L., Rao, A.J., Demerdash, O.N., Kalil, R.E., 2011. Expression of nestin by neural cells in the adult rat and human brain. PLoS One 6, e18535 (Apr 07). Hermann, A., Suess, C., Fauser, M., Kanzler, S., Witt, M., Fabel, K., Schwarz, J., Hoglinger, G.U., Storch, A., 2009. Rostro-caudal gradual loss of cellular diversity within the periventricular regions of the ventricular system. Stem Cells 27, 928–941 (Apr). Horner, P.J., Power, A.E., Kempermann, G., Kuhn, H.G., Palmer, T.D., Winkler, J., Thal, L.J., Gage, F.H., 2000. Proliferation and differentiation of progenitor cells throughout the intact adult rat spinal cord. J. Neurosci. 20, 2218–2228 (Mar 15). Imayoshi, I., Ohtsuka, T., Metzger, D., Chambon, P., Kageyama, R., 2006. Temporal regulation of Cre recombinase activity in neural stem cells. Genesis 44, 233–238 (May). Klaissle, P., Lesemann, A., Huehnchen, P., Hermann, A., Storch, A., Steiner, B., 2012. Physical activity and environmental enrichment regulate the generation of neural precursors in the adult mouse substantia nigra in a dopamine-dependent manner. BMC Neurosci. 13, 132.

153

Lai, B., Mao, X.O., Xie, L., Chang, S.Y., Xiong, Z.G., Jin, K., Greenberg, D.A., 2010. Electrophysiological properties of subventricular zone cells in adult mouse brain. Brain Res. 1340, 96–105 (Jun 22). Lendahl, U., Zimmerman, L.B., McKay, R.D., 1990. CNS stem cells express a new class of intermediate filament protein. Cell 60, 585–595 (Feb 23). Lie, D.C., Dziewczapolski, G., Willhoite, A.R., Kaspar, B.K., Shults, C.W., Gage, F.H., 2002. The adult substantia nigra contains progenitor cells with neurogenic potential. J. Neurosci. 22, 6639–6649 (Aug 1). Ling, Z.D., Potter, E.D., Lipton, J.W., Carvey, P.M., 1998. Differentiation of mesencephalic progenitor cells into dopaminergic neurons by cytokines. Exp. Neurol. 149, 411–423 (Feb). Marin-Burgin, A., Mongiat, L.A., Pardi, M.B., Schinder, A.F., 2012. Unique processing during a period of high excitation/inhibition balance in adult-born neurons. Science 335, 1238–1242 (Mar 9). Michalczyk, K., Ziman, M., 2005. Nestin structure and predicted function in cellular cytoskeletal organisation. Histol. Histopathol. 20, 665–671 (Apr). Migaud, M., Batailler, M., Segura, S., Duittoz, A., Franceschini, I., Pillon, D., 2010. Emerging new sites for adult neurogenesis in the mammalian brain: a comparative study between the hypothalamus and the classical neurogenic zones. Eur. J. Neurosci. 32, 2042–2052 (Dec). Mignone, J.L., Kukekov, V., Chiang, A.S., Steindler, D., Enikolopov, G., 2004. Neural stem and progenitor cells in nestin-GFP transgenic mice. J. Comp. Neurol. 469, 311–324 (Feb 9). Mokry, J., Nemecek, S., 1999. Cerebral angiogenesis shows nestin expression in endothelial cells. Gen. Physiol. Biophys. 18 (Suppl. 1), 25–29 (Dec). Nomura, T., Yabe, T., Rosenthal, E.S., Krzan, M., Schwartz, J.P., 2000. PSA-NCAM distinguishes reactive astrocytes in 6-OHDA-lesioned substantia nigra from those in the striatal terminal fields. J. Neurosci. Res. 61, 588–596 (Sep 15). Osborne, P.B., Halliday, G.M., Cooper, H.M., Keast, J.R., 2005. Localization of immunoreactivity for deleted in colorectal cancer (DCC), the receptor for the guidance factor netrin-1, in ventral tier dopamine projection pathways in adult rodents. Neuroscience 131, 671–681. Palmer, T.D., Willhoite, A.R., Gage, F.H., 2000. Vascular niche for adult hippocampal neurogenesis. J. Comp. Neurol. 425, 479–494 (Oct 2). Peng, J., Xie, L., Jin, K., Greenberg, D.A., Andersen, J.K., 2008. Fibroblast growth factor 2 enhances striatal and nigral neurogenesis in the acute 1-methyl-4-phenyl-1,2,3,6tetrahydropyridine model of Parkinson's disease. Neuroscience 153, 664–670 (May 15). Prakash, N., Wurst, W., 2006. Development of dopaminergic neurons in the mammalian brain. Cell. Mol. Life Sci. 63, 187–206 (Jan). Redmond Jr., D.E., Weiss, S., Elsworth, J.D., Roth, R.H., Wakeman, D.R., Bjugstad, K.B., Collier, T.J., Blanchard, B.C., Teng, Y.D., Synder, E.Y., Sladek Jr., J.R., 2010. Cellular repair in the parkinsonian nonhuman primate brain. Rejuvenation Res. 13, 188–194 (Apr-Jun). Sakamoto, M., Imayoshi, I., Ohtsuka, T., Yamaguchi, M., Mori, K., Kageyama, R., 2011. Continuous neurogenesis in the adult forebrain is required for innate olfactory responses. Proc. Natl. Acad. Sci. U. S. A. 108, 8479–8484 (May 17). Schiff, M., Weinhold, B., Grothe, C., Hildebrandt, H., 2009. NCAM and polysialyltransferase profiles match dopaminergic marker gene expression but polysialic acid is dispensable for development of the midbrain dopamine system. J. Neurochem. 110, 1661–1673 (Sep). Schmidt-Hieber, C., Jonas, P., Bischofberger, J., 2004. Enhanced synaptic plasticity in newly generated granule cells of the adult hippocampus. Nature 429, 184–187 (May 13). Shan, X., Chi, L., Bishop, M., Luo, C., Lien, L., Zhang, Z., Liu, R., 2006. Enhanced de novo neurogenesis and dopaminergic neurogenesis in the substantia nigra of 1-methyl4-phenyl-1,2,3,6-tetrahydropyridine-induced Parkinson's disease-like mice. Stem Cells 24, 1280–1287 (May). Spampanato, J., Sullivan, R.K., Turpin, F.R., Bartlett, P.F., Sah, P., 2012. Properties of doublecortin expressing neurons in the adult mouse dentate gyrus. PLoS One 7, e41029. Srinivas, S., Watanabe, T., Lin, C.S., William, C.M., Tanabe, Y., Jessell, T.M., Costantini, F., 2001. Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev. Biol. 1, 4. Steinert, P.M., Liem, R.K., 1990. Intermediate filament dynamics. Cell 60, 521–523 (Feb 23). Su, P.H., Chen, C.C., Chang, Y.F., Wong, Z.R., Chang, K.W., Huang, B.M., Yang, H.Y., 2013. Identification and cytoprotective function of a novel nestin isoform, Nes-S, in dorsal root ganglia neurons. J. Biol. Chem. 288, 8391–8404 (Mar 22). Sun, M.Y., Yetman, M.J., Lee, T.C., Chen, Y., Jankowsky, J.L., 2014. Specificity and efficiency of reporter expression in adult neural progenitors vary substantially among nestinCreER(T2) lines. J. Comp. Neurol. 522, 1191–1208 (Apr 01). Van Kampen, J.M., Eckman, C.B., 2006. Dopamine D3 receptor agonist delivery to a model of Parkinson's disease restores the nigrostriatal pathway and improves locomotor behavior. J. Neurosci. 26, 7272–7280 (Jul 5). Worlitzer, M.M., Viel, T., Jacobs, A.H., Schwamborn, J.C., 2013. The majority of newly generated cells in the adult mouse substantia nigra express low levels of Doublecortin, but their proliferation is unaffected by 6-OHDA-induced nigral lesion or Minocyclinemediated inhibition of neuroinflammation. Eur. J. Neurosci. 38, 2684–2692 (Sep). Yoshimi, K., Ren, Y.R., Seki, T., Yamada, M., Ooizumi, H., Onodera, M., Saito, Y., Murayama, S., Okano, H., Mizuno, Y., Mochizuki, H., 2005. Possibility for neurogenesis in substantia nigra of parkinsonian brain. Ann. Neurol. 58, 31–40 (Jul). Zhao, M., Momma, S., Delfani, K., Carlen, M., Cassidy, R.M., Johansson, C.B., Brismar, H., Shupliakov, O., Frisen, J., Janson, A.M., Jun 24 2003. Evidence for neurogenesis in the adult mammalian substantia nigra. Proc. Natl. Acad. Sci. U. S. A. 100, 7925–7930. Zhao, C., Teng, E.M., Summers Jr., R.G., Ming, G.L., Gage, F.H., 2006. Distinct morphological stages of dentate granule neuron maturation in the adult mouse hippocampus. J. Neurosci. 26, 3–11 (Jan 4). Zimmerman, L., Parr, B., Lendahl, U., Cunningham, M., McKay, R., Gavin, B., Mann, J., Vassileva, G., McMahon, A., 1994. Independent regulatory elements in the nestin gene direct transgene expression to neural stem cells or muscle precursors. Neuron 12, 11–24 (Jan).