Brain-Derived Neurotrophic Factor Deficiency Restricts Proliferation of Oligodendrocyte Progenitors Following Cuprizone-Induced Demyelination

Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family of growth factors that through its neurotrophic tyrosine kinase, receptor, type 2 (TrkB) receptor, increases 5-bromo-2-deoxyuridine incorporation in oligodendrocyte progenitor cells (OPCs) in culture. Roles in vivo are less well understood; however, increases in numbers of OPCs are restricted in BDNF+/− mice following cuprizone-elicited demyelination. Here, we investigate whether these blunted increases in OPCs are associated with changes in proliferation. BDNF+/+ and BDNF+/− mice were fed cuprizone-containing or control feed. To assess effects on OPC numbers, platelet-derived growth factor receptor alpha (PDGFRα)+ or NG2+ cells were counted. To monitor DNA synthesis, 5-ethynyl-2′-deoxyuridine (EdU) was injected intraperitoneally and colocalized with PDGFRα+ cells. Alternatively, proliferating cell nuclear antigen (PCNA) was colocalized with PDGFRα or NG2. Labeling indices were determined in the BDNF+/+ and BDNF+/− animals. After 4 or 5 weeks of control feed, BDNF+/− mice exhibit similar numbers of OPCs compared with BDNF+/+ animals. The labeling indices for EdU and PCNA also were not significantly different, suggesting that neither the DNA synthesis phase (S phase) nor the proliferative pool size was different between genotypes. In contrast, when mice were challenged by cuprizone for 4 or 5 weeks, increases in OPCs observed in BDNF+/+ mice were reduced in the BDNF+/− mice. This difference in elevations in cell number was accompanied by decreases in EdU labeling and PCNA labeling without changes in cell death, indicating a reduction in the DNA synthesis and the proliferative pool. Therefore, levels of BDNF influence the proliferation of OPCs resulting from a demyelinating lesion.


Introduction
In adult animals subjected to a demyelinating lesion, oligodendrocyte progenitor cells (OPCs) derived from cells present in the subventricular zone and the lesion site are able to proliferate and differentiate into myelinating oligodendrocytes (OLGs) (Gensert and Goldman, 1997;Woodruff and Franklin, 1997;Di Bello et al., 1999;Nait-Oumesmar et al., 1999). Similarly, in multiple sclerosis (MS), progenitors increase in areas of active demyelination and remyelination (Wilson et al., 2006). To be in position to influence these OPCs, and potentially enhance the opportunity for remyelination, it is important to identify factors that increase numbers of these cells and define molecular events that regulate this process.
One factor of particular interest is brain-derived neurotrophic factor (BDNF), which is known to stimulate and control OLG development in vivo and in culture.
For example, BDNF increases numbers of basal forebrain (BF) OPCs in culture ( Van't Veer et al., 2009), and when there is reduced BDNF as occurs in BDNFþ/À mice, there are decreases in NG2þ OPCs in the BF. These effects of BDNF extend to differentiation, as BDNFþ/À mice exhibit marked reduction in levels of the myelin proteins (including myelin basic protein, myelin-associated glycoprotein, and proteolipid protein in postnatal and adult mice; VonDran et al., 2010). Moreover, BDNFÀ/À or BDNFþ/À mice exhibit decreased numbers of myelinated axons in the postnatal optic nerve, and myelin proteins are decreased throughout the brain of BDNFÀ/À mice during postnatal development (Cellerino et al., 1997;Djalali et al., 2005;Xiao et al., 2010). At least some of these defects are reversible. For instance, BDNF injection into the ventricles of postnatal day 10 (P10) and P12 mice increases proteolipid protein mRNA in the hippocampus at P14 (Cellerino et al., 1997).
In the present study, we extend observations that indicate that BDNF increases progenitor cell number following demyelination to explore underlying mechanisms. To do so, we took advantage of the cuprizone model. Cuprizone administration into the mouse diet induces OLG cell death and demyelination. However, it also results in a recovery from the demyelination process (Blakemore, 1973a(Blakemore, , 1973bMason et al., 2001;Matsushima and Morell, 2001), which is associated with increases in OPCs. Previously, it was determined that mice with reduced levels of BDNF exhibit a blunted increase in NG2þ OPCs (VonDran et al., 2011). In the present study, using both platelet-derived growth factor receptor alpha (PDGFRa) as well as NG2þ as markers of OPCs, we explore whether this blunted increase of OPCs is associated with alterations in DNA synthesis and proliferation. Our results indicate that BDNF not only impacts the numbers of OPCs that respond to a demyelinating insult but also influences both the DNA synthesis phase (S phase) and the proliferation of OLG progenitors. The studies indicate that BDNF levels are important for maintenance of the OPC pool that may impact myelin repair.

Experimental Animals
Breeding pairs of BDNF Tm1Jae mice on a 129/BalbC/C57 background were previously purchased from Jackson Laboratories (Bar Harbor, ME) and maintained in the Robert Wood Johnson Medical School Animal Facility, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. Animal maintenance, husbandry, and housing are in compliance with the Laboratory Animal Welfare Act . Breeding pairs were maintained by crossing wild-type (WT) and heterozygous animals. The heterozygous mice exhibit approximately 50% of normal levels of BDNF but appear normal (VonDran et al., 2010). The mouse genotype was determined by polymerase chain reaction analysis of ear-or tail-derived DNA as described elsewhere (Ernfors et al., 1994). Mice were housed in a temperature and humiditycontrolled environment with a 12-hr light-dark cycle and maintained on standard mouse chow with water ad libitum prior to cuprizone treatment.
Cuprizone Treatment and 5-Ethynyl-2 0 -Deoxyuridine Injection At 8 to 10 weeks of age, two BDNFþ/þ (WT) and two BDNFþ/À littermates were randomly selected for each experiment. One WT or BDNFþ/À mouse was fed cuprizone-containing feed, while one WT or BDNFþ/À mouse was fed control feed for 4 or 5 weeks. The cuprizone feed consisted of 0.2% cuprizone milled into mouse feed (Harlan Teklad, Madison, WI). The control feed was the same, but without cuprizone supplementation (Harlan Teklad). Food containing cuprizone or control food was changed every 2 to 3 days. In some cases, 5-ethynyl-2 0 -deoxyuridine (EdU), or its vehicle, saline, was injected intraperitoneally 2 hr before sacrifice.

Immunohistochemistry
Mice were euthanized by cardiac perfusion with saline followed by 4% paraformaldehyde. The brains were fixed in 4% paraformaldehyde, cryoprotected in 30% sucrose, embedded in Optimal Cutting Temperature compound (OCT), and frozen at À80 C for immunohistochemistry. Serial 20 mm frozen coronal sections, were prepared and used for analysis.
To identify progenitor cells, sections were stained using an antibody against PDGFRa (Santa Cruz Biotechnology, Dallas, TX) or an antibody against NG2 (EMD Millipore, Temecula, CA) and costained with anti-proliferating cell nuclear antigen (PCNA, a marker for cell proliferation; Santa Cruz Biotechnology) or, in the case of PDGFRa, anti-EdU (Invitrogen, Grand Island, NY).
For PDGFRa and PCNA or EdU costaining, sections were first heated at 95 C in 0.01 M citrate buffer, blocked in 5% goat serum/0.5% Triton/phosphate buffered saline (PBS) for 1 hr, and then incubated with rabbit PDGFRa primary antibody (1:100, 1 hr at room temperature [RT]). Sections were then washed three times and incubated with anti-rabbit Alexa Fluor 488 antibody (1 hr at RT; Invitrogen). After washing with PBS, sections were incubated with mouse PCNA antibody (1:1000, overnight, 4 C) followed by washing with PBS and incubation with either anti-mouse Alexa Fluor 594 (1:500) or Cy3 antibody (1:300; Invitrogen) for 1 hr at RT. For double immunostaining for PDGFRa and EdU, sections were stained for EdU, according to the manufacturer's instruction (Click-iT Edu Imaging Kit, Invitrogen).
For double staining for PDGFRa or CC1 (a marker for mature OLGs) and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL), sections were treated with Cytonin (Invitrogen) for 40 min at RT, and apoptotic cells were determined by assay using the TUNEL Kit, according to the manufacturer's instruction (TACS 2 TdT-Fluor In Situ Apoptosis Detection Kit, Trevigen, Gaithersburg, MD). Thereafter, sections were stained for PDGFRa as described earlier, while for CC1, sections were blocked in 5% goat serum/0.5% Triton/ PBS for 1 hr, and then incubated with anti-mouse CC1 antibody (1:200; Calbiochem, Darmstadt, Germany) for 48 hr at 4 C, followed by washing with PBS and incubation with Cy3 antibody (1:300, 1 hr at RT).
Images were collected using a Leica inverted fluorescence microscope equipped with an Olympus MagnaFire digital camera and ImagePro Plus 7.0 image analysis software. Contrast of whole images was enhanced using Adobe Photoshop.

Western Blot
To detect PDGFRa protein levels by Western blot, the midline of the corpus callosum was dissected and lyzed. Protein concentration was quantified as previously described (VonDran et al., 2011). Lysates were run on 3% to 8% Tris-acetate gels (Invitrogen) and then transferred to a polyvinylidene difluoride (PVDF) membrane (EMD Millipore). Membranes were blocked in 4% BSA/ TBS-T for 1 hr followed by overnight incubation with rabbit anti-PDGFRa (1:750; Santa Cruz Technology) at 4 C. Membranes were then incubated in horseradish peroxidase-linked anti-rabbit antibody (1:3000; GE Healthcare, Little Chalfont, Buckinghamshire, UK). Bands were visualized with a chemiluminescence system (Perkin Elmer, Inc, Waltham, MA). Membranes were stripped and reprobed with anti-b-tubulin (1:2000; Sigma, St Louis, MO) as loading control.

Quantification and Statistical Analysis
The number of single-and double-positive cells was counted in serial sections taken from the midline of the corpus callosum through the whole extent of the area over the fimbria-fornix, from bregma À0.22 to 0.94 (Paxinos and Franklin, 1997). Magnification for the quantitation used a 40Â objective lens. The midline of the corpus callosum (a width of 240 mm) was chosen because of its significant demyelination, as well as its astrogliosis and microglial responses compared with the other regions of the corpus callosum following cuprizone (Schmidt et al., 2013). In total, 36 serial sections were counted for PDGFRaþ cells; of these, 24 sections were assessed for EdU/PDGFRaþ cells from each mouse, and 12 sections from each mouse were assessed for PCNA/ PDGFRaþcells. Twelve serial sections were assessed for PCNA/NG2þ cells. Six sections per mouse were used to identify PDGFRa or CC1 cell death by the TUNEL assay. Double counting was eliminated using the Abercrombie correction (Abercrombie, 1946). Statistical analysis was performed with StatView software, and the data were presented as mean AE SEM and analyzed using analysis of variance followed by Fisher's protected least significant difference post hoc test or the Matched Pair Student's t test as appropriate, and p < .05 was considered significant.
For each experiment, one BDNFþ/þ mouse fed control feed was compared with one BDNFþ/þ mouse fed cuprizone-containing food, one BDNFþ/À mouse fed control food, and one BDNFþ/À mouse fed cuprizonecontaining feed. The labeling index (LI) was defined by the ratio of the numbers of EdU/PDGFRaþ doublelabeled cells or numbers of PCNA/PDGFRaþ doublelabeled cells over the numbers of total PDGFRaþ cells. Alternatively, the LI was defined by the ratio of the numbers of PCNA/NG2þ double-labeled cells over the numbers of total NG2þ cells. For each individual experiment, the LI of the BDNFþ/À mouse fed control feed, the BDNFþ/À mouse fed cuprizone, and the BDNFþ/þ mouse fed cuprizone was then expressed as percent of the control LI determined for BDNFþ/þ mice fed control feed. Within each experiment, then, each group was compared with its own control. Four-week studies were repeated in seven independent experiments. Five-week studies examining PDGFRa þ cells were repeated in six independent experiments. Five-week studies examining NG2þ cells were repeated in three independent experiments. TUNEL staining was assessed in four BDNFþ/À and four BDNFþ/þ mice.
For Western blot studies, BDNFþ/þ and BDNFþ/À mice fed cuprizone were compared with their own controls (mice fed control feed) for the same time points. Differences between controls and cuprizone-fed mice were then evaluated. Western blots were analyzed using Quantity One V 4.2.1 software (Bio-Rad, Hercules, CA). Statistical analysis was performed with StatView software, and the data were presented as mean AE SEM. Statistical differences were determined using the Student's t test, and p < .05 was considered significant.

PDGFR Is Decreased When BDNFþ/À Mice Are Compared With BDNFþ/þ Mice After Cuprizone
The number of OPCs and their maturation into myelinproducing cells is important for tissue repair and remyelination following demyelination. Because BDNF may be involved in supporting the expansion of oligodendrocyte progenitors, whether reduced levels of BDNF impact the OPC response was examined. To do so in previous work (VonDran et al., 2011), BDNFþ/þ and BDNFþ/À mice were fed normal or cuprizone-containing feed, and NG2 protein levels were assessed by Western blotting and NG2þ cell numbers by counting. Increases in NG2 protein as well as NG2þ cells were evident at 4 weeks and 5 weeks of cuprizone, but these changes were blunted in the BDNFþ/À mice, suggesting that BDNF impacts the response of OPCs to injury. Interestingly, at 6 weeks when increases in OPCs are no longer evident, there was no difference in the BDNFþ/þ versus BDNFþ/À mice, suggesting that BDNF's actions are primarily on the cuprizone-elicited increases in OPCs.
To extend these studies and begin defining mechanisms underlying the cuprizone-elicited increases of OPCs, initial studies evaluated PDGFRa, a second marker of progenitor cells. In the absence of cuprizone, numbers of PDGFRa þ OPCs in BDNFþ/þ mice are not different from those in BDNFþ/À animals (Figure 1(a)), suggesting that baseline precursor pools were unaffected by diminished levels of BDNF. However, there is a marked contrast in number of PDGFRa þ OPCs following 4 or 5 weeks of cuprizone feed. After cuprizone, the numbers of PDGFRa þ OPCs in BDNFþ/þ mice increased. While BDNFþ/À mice also exhibited an increase in PDGFRa þ OPCs after cuprizone, their OPC response is half that of the WT mice. Similarly, levels of PDGFRa protein that increased greater than threefold Figure 1. BDNFþ/À mice exhibit a blunted increase in PDGFRa þ progenitors following cuprizone compared with BDNFþ/þ mice. (a) BDNF deficiency does not impact numbers of PDGFRaþ cells under control conditions at 4 or 5 weeks of cuprizone treatment. However, in response to cuprizone, increases in PDGFRaþ cells that occur in the BDNFþ/þ mice are blunted in BDNFþ/À animals. *Significantly different from control (WT mouse, no cuprizone or BDNFþ/À mouse, no cuprizone), p < .05, # significantly different from BDNFþ/þ mice (fed cuprizone), p < .05. The data represent the total number of PDGFRaþ cells quantitated in serial sections taken from the corpus callosum (see ''Methods'' section). Data were obtained from seven (4 weeks) or six (5 weeks) individual experiments, each containing littermates, either WT or BDNFþ/À mice, fed cuprizone or control feed. (b) Western blots of PDGFRa in control mice and animals fed cuprizone for 4 or 5 weeks. Anti-b-tubulin was used to normalize to total proteins levels. (c) The graphs represent densitometric analyses of the blots at 4 or 5 weeks of treatment. Each result is shown as a percentage of its own control (no cuprizone). *Significantly different from control (WT mouse, no cuprizone or BDNFþ/À mouse, no cuprizone), p < .01, #significantly different from BDNFþ/þ mice (fed cuprizone), p < .05. Insert represents densitometric analysis of PDGFRa blots of 4-and 5-week-old BDNFþ/þ and BDNFþ/À unlesioned corpus callosum. Data were analyzed by Student's t test (N > 3). BDNF ¼ brain-derived neurotrophic factor; PDGFR¼plateletderived growth factor alpha; WT ¼ wild type.
in the BDNFþ/þ mice in response to cuprizone were blunted in the BDNFþ/À mice at 4 and 5 weeks (Figure 1(b) and (c)). The blunted responses described here in PDGFRa cells are highly parallel to the previous studies of NG2-labeled cells (VonDran et al., 2011). Consistent with these findings, we detect extensive colocalization of NG2 and PDGFRa in cells within the lesion site ( Figure 2). Approximately 80% of the NG2þ population colocalizes with PDGFRaþ cells, while all PDGFRaþ cells are NG2þ, suggesting that the markers are identifying essentially the same population of progenitors.

The Proliferation of OPCs Is Decreased in BDNFþ/À Mice Compared With BDNF+/+ Mice After Cuprizone
To begin defining the mechanisms underlying the differential increases in progenitors in BDNFþ/À and BDNFþ/þ mice, DNA synthesis was assessed by the intraperitoneal injection of EdU 2 hr before sacrifice followed by colabeling with anti-PDGFRa. Incorporation of EdU into DNA identifies the S phase of the cell cycle. Alternatively, PCNA signal was colocalized with PDGFRaþ cells. PCNA is expressed in all phases of the cell cycle and is a marker of proliferation (von Bohlen and Halbach, 2011). Labeling indices (the ratio of double-positive cells to total PDGFRaþ cells) were also assessed.
We speculated that the diminished oligodendrocyte progenitor response after cuprizone in the BDNFþ/À mouse might depend on deficits in DNA synthesis and/ or proliferation. To examine this issue, we first examined S phase entry in the different genotypes at 4 and 5 weeks of cuprizone (Figure 3). At both time points, a subset of PDGFRaþ cells was labeled with EdU in both the there is no difference when BDNFþ/þ and BDNFþ/À mice fed control feed are compared. In contrast, following cuprizone, there are significant differences in the increases in LI in BDNFþ/þ mice versus BDNFþ/À mice. The LI of the control group (BDNFþ/þ mice fed control feed) of the seven independent experiments ranges between 5% and 6%. Results for each experimental animal are presented as a percent relative to its own control for that experiment (see ''Methods'' section). *Significantly different from control mice (no cuprizone), p < .05, #-significantly different from BDNFþ/þ mice (fed cuprizone), p < .05. At 5 weeks, there is also no difference when BDNFþ/þ and BDNFþ/À mice fed control feed are compared. Following cuprizone, there are significant increases in the LI of BDNFþ/þ and BDNFþ/À mice relative to control. *Significantly different from control mice (no cuprizone), analysis of variance, Fisher's post hoc test. Insert represents the LI of the cuprizonetreated BDNFþ/À cells as percent control cuprizone-fed BDNFþ/þ cells. #Significantly different from BDNFþ/þ mice fed cuprizone, p < .05, using the matched pair Student's t test. The LI of the control group (BDNFþ/þ mice fed control feed) of the six independent experiments ranges between 5% and 7%. Results for each experimental animal are presented as a percent relative to its own control for that experiment (see ''Methods'' section). Scale bar ¼ 20 mm. BDNF ¼ brain-derived neurotrophic factor; PDGFRa¼plateletderived growth factor alpha; OPCs ¼ oligodendrocyte progenitor cells; EdUþ ¼ 5-ethynyl-2 0 -deoxyuridine; LI ¼ labeling index. BDNFþ/þ and þ/À mice, fed control or cuprizonecontaining food (Figure 3(a)). When the LI was assessed in BDNFþ/þ and BDNFþ/À mice fed control feed, there was no difference in baseline LI, suggesting that altered BDNF levels do not impact the overall rate of DNA synthesis in the PDGFRaþ precursor pool under normal conditions (Figure 3(b)). However, when challenged by cuprizone, the BDNFþ/þ mice exhibited a greater than twofold increase in mitotic LI relative to animals fed control feed, indicative of a robust increase in cells synthesizing DNA. This suggests the observed increase in total precursors depended on enhanced proliferation that followed cell cycle progression. In contrast, in the BDNFþ/À mice, there was a blunted increase in the mitotic LI, compared with the BDNFþ/þ response, indicating that reduced BDNF limited significantly the ability of OPCs to enter the cell cycle.
One possible mechanism underlying reduced S phase entry would be a diminished overall proliferative precursor pool in the BDNFþ/À mice. To define this population, we performed double immunolabeling, identifying PDGFRaþ precursors that coexpress PCNA ( Figure 4). As was the case in PDGFRa/EdU labeling, a subset of PDGFRaþ cells expresses PCNA at 4 and 5 weeks of cuprizone feeding (Figure 4(b)). Moreover, under conditions of normal feed, the baseline PDGFRa/PCNA double-labeled cells were not different between genotypes, a result consistent with no genotype differences in the mitotic LI (Figure 3). However, after cuprizone treatment, BDNFþ/þ mice exhibited a significant increase in PDGFRa/PCNA double-labeled cells compared with controls, whereas BDNFþ/À mice exhibited a significant blunting of the effect (Figure 4(b)). In aggregate, the data suggest that when BDNF is limited, the oligodendrocyte precursor pool and its proliferation are reduced.

NG2þ Cells Act Similarly to the PDGFRaþ Population
NG2 and PDGFRa are largely colocalized within the lesion site (Figure 2), suggesting that they are labeling essentially the same population of progenitors. However, to confirm that the NG2þ cells respond, as do the PDGFRa progenitors, with a reduction in proliferation in mice with reduced levels of BDNF, we performed studies to evaluate the NG2þ/PCNAþ population at 5 weeks of cuprizone. As was the case with PDGFRa, a subset of NG2þ cells expresses PCNA in mice fed control feed (Figure 5(a)). However, after cuprizone treatment, increases in NG2þ/PCNAþ cells seen in BDNFþ/þ mice were reduced in the BDNFþ/À mice (Figure 5(b)). The data indicate that reduced levels of BDNF impact both NG2þ and PDGFRaþ progenitors in parallel fashion. . BDNF deficiency at 4 or 5 weeks reduces enhanced proliferation in OPCs following cuprizone, while no differences are noted in mice fed control feed. (a) PDGFRa/PCNAþ cells are indicated by the arrows in sections taken from mice fed control or cuprizone feed for 5 weeks. (b) When the labeling indices within individual experiments are expressed as percent control, there is no difference when BDNFþ/þ and BDNFþ/À mice are compared at either time point. In contrast following cuprizone, there are significant differences in the increases in LI in BDNFþ/þ mice versus BDNFþ/À mice. The LI of the control groups (BDNFþ/þ mice fed control feed) for the seven independent experiments ranges between 11% and 17% at 4 weeks and for the six independent experiments ranged between 21% and 25% at 5 weeks. Each group value is expressed as a percent relative to own control for that experiment (see ''Methods'' section). *Significantly different from control (no cuprizone) mice, p < .05. #-Significantly different from BDNFþ/þ (fed cuprizone) mice, p < .05. Scale bar ¼ 20mm. BDNF ¼ brain-derived neurotrophic factor; OPCs ¼ oligodendrocyte progenitor cells; PDGFRa ¼ platelet-derived growth factor alpha; LI ¼ labeling index; PCNA ¼ proliferating cell nuclear antigen.

Cuprizone Is not Toxic for OPCs but Is Toxic for Mature OLGs
While the foregoing data suggest that OLG proliferation is diminished when BDNF is limited, an alternative mechanism may also involve programmed cell death.
To determine whether the reduction in numbers of progenitors in the BDNFþ/À mice may be affiliated with cell death, we examined apoptosis using PDGFRa/TUNEL double immunostaining in BDNFþ/þ and BDNFþ/À mice following 4 weeks cuprizone treatment. No colocalization of PDGFRaþ cells with TUNEL was detected (Figure 6(a)). This suggests that proliferation differences were not due to cell death at this time because TUNEL labeling of PDGFRa cells was not seen in either genotype. On the other hand, examination of mature CC1þ OLGs, known to die during this time period (i.e., Arnett et al., 2001;Be´nardais et al., 2013), did colabel with TUNEL ( Figure 6(b)), indicating the sensitivity of the assay.

Discussion
The present study indicates that following a cuprizoneelicited demyelinating lesion, WT BDNFþ/þ mice exhibit an increase in PDGFRaþ progenitors that is due to increased DNA synthesis and cell proliferation. On the other hand, when BDNF is limited as occurs in BDNFþ/À Figure 5. BDNF deficiency at 5 weeks reduces enhanced proliferation in NG2þ OPCs following cuprizone, while no differences are noted in mice fed control feed. (a) NG2/PCNAþ cells are indicated by the arrows in sections taken from mice fed control or cuprizone feed for 5 weeks. (b) When the labeling indices within individual experiments are expressed as percent control, there is no difference when BDNFþ/þ and BDNFþ/À mice are compared. In contrast following cuprizone, there are significant differences in the increases in LI in BDNFþ/þ mice versus BDNFþ/À mice. The LI of the control groups (BDNFþ/þ mice fed control feed) for the three independent experiments ranges between 15% and 21%. Each group value is expressed as a percent relative to own control for that experiment (see ''Methods'' section). *Significantly different from control (no cuprizone) mice, p < .05. #-Significantly different from BDNFþ/þ (fed cuprizone) mice, p < .05, n ¼ 3 independent experiments. Scale bar ¼ 20mm. BDNF ¼ brain-derived neurotrophic factor; OPCs ¼ oligodendrocyte progenitor cells; PCNA ¼ proliferating cell nuclear antigen; LI ¼ labeling index. Figure 6. PDGFRaþ cells do not undergo apoptosis after cuprizone toxicity in either BDNFþ/þ or BDNFþ/À mice as detected by PDGFRa/TUNEL assay. Oligodendrocyte apoptosis is detected in mice following cuprizone toxicity. TUNEL staining indicated in green and is not associated with PDGFRaþ cells, indicated in red (a). TUNEL staining is associated with CC1þ cells indicated in red (b). Arrows indicate double labeling. Six sections per mouse were used to identify PDGFRa or CC1 cell death by the TUNEL assay. Four BDNFþ/þ mice and BDNFþ/À mice were sampled. Scale bar ¼ 20mm. PDGFRa ¼ platelet-derived growth factor alpha; BDNF ¼ brain-derived neurotrophic factor; TUNEL ¼ terminal deoxynucleotidyl transferase dUTP nick end labeling. mice,the increase in PDGFRaþ progenitors is reduced, a deficiency that depends on reductions in both DNA synthesis and subsequent cell proliferation. Similar effects are noted when NG2þ cells are evaluated. This effect is not due to differential actions on cell death.
This new work confirms and extends previous observations that indicated that in culture BF OPCs respond to BDNF by increasing DNA synthesis ( Van't Veer et al., 2009) and differentiation (Du et al., 2006). The increase in DNA synthesis observed in culture, suggesting levels of BDNF regulate cell proliferation, is consistent with in vivo studies in which BDNFþ/À mice exhibit decreased numbers of NG2þ progenitor cells at postnatal stages and in adults (VonDran et al., 2010) and after a demyelinating lesion elicited by cuprizone treatment (VonDran et al., 2011). The present work now proposes a mediating mechanism that underlies the previous in vivo work.

Effects of BDNF on OPCs Are no Longer Evident after 6 Weeks of Cuprizone Treatment
In previous work, it was found that increases in progenitors, evident at 4 and 5 weeks of cuprizone, are no longer observed at 6 weeks. Moreover, differences in the progenitors evident in the BDNFþ/À and BDNFþ/þ populations are also gone. Interestingly, the absence of differences in cell numbers also pertains to the postmitotic OLG lineage cells evident in the lesion site, as the numbers of CC1þ OLGs are not different in the BDNFþ/þ and þ/À mice (VonDran et al., 2011). This suggests that the differential effects on proliferation do not impact the final number of cells that go on to express mature traits. (However, note that although the numbers of OLGs are similar, the cells maturing in an environment deficient in BDNF are limited in their ability to produce normal levels of myelin proteins; VonDran et al., 2011.) These results are reminiscent of previous studies examining the BF during development and in the adult in which BDNFþ/À mice exhibit reduced numbers of NG2þ cells in the BF, but similar numbers of postmitotic CC1þ cells that were reduced in their ability to produce normal levels of myelin proteins (VonDran et al., 2010).
It was suggested in these previous studies that it may only be a subpopulation of the progenitors that go on to mature into postmitotic OLG populations, and the other progenitors might die. In the present work, we evaluated TUNEL staining at 4 weeks, but we cannot rule out the possibility that differential death of the progenitors is occurring between 5 and 6 weeks following cuprizone. Such an observation would be consistent with studies, indicating that 50% of OLG lineage cells undergo cell death (Barres et al., 1992) during development and in particular during early myelination (Trapp et al., 1997).
Alternatively, it has been reported that following a lesion, a subset of OPCs, known to predominantly mature into OLGs (Zuo and Nishiyama, 2013), may, surprisingly, mature into Schwann cells (Zawadzka et al., 2010). Based on these reports, it is possible that BDNF impacts either the death or the fate of a subset of OPCs that does not go on to become mature OLGs, subjects for future investigation.
The Significance of the Differential Increase in Progenitors in BDNFþ/þ Versus BDNFþ/À Mice If differences in numbers of OLG lineage cells are gone by 6 weeks, what then is the importance of the effects of BDNF on progenitor cell numbers and can it apply to other progenitor populations? The literature suggests that the NG2 population of cells may serve roles other than being progenitors that mature into OLGs. Thus, effects on final OLG cell numbers may not be the only important outcome. For example, NG2þ cells within the hippocampus, cortex, cerebellum, optic nerve, and corpus callosum are known to receive glutamatergic synaptic contacts (Lin and Bergles, 2004;Lin et al., 2005;Kukley et al., 2007;Ziskin et al., 2007) and respond to glutamate by induction of calcium-dependent phosphorylation of extracellular signal-regulated kinase (ERK1/2) (Hamilton et al., 2009). Moreover, in other work, NG2þ cells are known to receive Gamma-Amino Butyric Acid (GABA) ergic synapses and respond to GABA stimulation by increasing BDNF (Tanaka et al., 2009). These cells are also reported to exhibit processes that are directed between axon terminals (Ong and Levine, 1999) and extend to nodes of Ranvier (Butt et al., 1999). The possible participation of NG2þ cells in a synaptic network may allow them to modulate neural signals within the recovering brain and potentially influence the lesion environment. It is intriguing to hypothesize that BDNF may influence these NG2þ cell properties to influence the environment of a demyelinating lesion, a direction to pursue in other studies.

Effects of BDNF on Increases in Cell Number Are not Limited to OPCs
These results indicating that BDNF can increase numbers of OPCs are consistent with studies of McTigue et al. (1998) who found that BDNF elevates numbers of proliferating OLG lineage cells after spinal cord injury and are reminiscent of BDNF effects on neuronal progenitors. For example, BDNF administration into the lateral ventricle of adult rats results in significant increases in numbers of newly generated neurons in the adult olfactory bulb, the striatum, septum, thalamus, and hypothalamus (Zigova et al., 1998;Pencea et al., 2001). When BDNF is injected into the hippocampus, increases in proliferating granule cells are observed (Scharfman et al., 2005). BDNF associated with human bone marrow-derived mesenchymal stem cells enhances numbers of subventricular neural stem cells in a stroke model (Jeong et al., 2014). These increases in progenitors may be due to enhanced DNA synthesis and cell proliferation. Thus, VGF, a factor that acts downstream of BDNF, enhances both of these processes in hippocampal progenitor cells in vitro and in vivo (Thakker-Varia et al., 2007). Taken together, the studies suggest that BDNF may play a general role in multiple progenitor populations in the mature brain and after injury.

Roles of Other Growth Factors on Progenitors
Our study that confirms effects of BDNF on dividing cells following a demyelinating lesion adds BDNF to a growing list of growth factors that can influence the progenitor population. For example, following cuprizone-elicited demyelination, proliferation is inhibited in mice lacking insulin-like growth factor1 receptor (IGFR1), or tumor necrosis factor-alpha, or tumor necrosis factor receptor-2 (Arnett et al., 2001;Mason et al., 2003). Similarly, mice with decreased levels of PDGFRa exhibit decreases in DNA synthesis following cuprizone (Murtie et al., 2005). Consistent with these reports of decreases in factors inhibiting proliferation are those indicating that addition of growth factors may increase the proliferative population. Thus, intracranial administration of a combination of the growth factors PDGF-AA, basic fibroblast growth factor, neurotrophic factor 3, and insulin-like growth factor1 resulted in increases of NG2 cells and numbers of NG2/5-bromo-2-deoxyuridine (BrdU)þ cells in the corpus callosum (Kumar et al., 2007). We suggest that BDNF in combination with these factors may further enhance numbers of progenitor cells and that it may do so by enhancing proliferation and DNA synthesis.

Significance of Increases in BDNF to Disease
In MS, there is a correlation between BDNF and the severity of MS symptoms. MS patients exhibit decreased levels of BDNF in serum and cerebrospinal fluid during disease activity. Moreover, after treatment with interferon beta, glatiramer acetate, or laquinimod, all agents used to reduce the severity of MS, BDNF levels increase (Chen and Dhib-Jalbut, 2003;Caggiula et al., 2006;Bruck and Wegner, 2011). The evidence suggests that BDNF may have a neuroprotective role. Our results contribute to this possibility and suggest that it does so by enhancing the proliferation of OPCs that may influence the recovery from a demyelinating lesion.

Summary
In response to demyelination, there are increases in oligodendrocyte progenitors that are reduced in mice deficient in BDNF. The present work indicates that this reduction is due to effects on DNA synthesis and subsequent cell proliferation but not differential actions on cell death.