Dcc Regulates Asymmetric Outgrowth of Forebrain Neurons in Zebrafish

The guidance receptor DCC (deleted in colorectal cancer) ortholog UNC-40 regulates neuronal asymmetry development in Caenorhabditis elegans, but it is not known whether DCC plays a role in the specification of neuronal polarity in vertebrates. To examine the roles of DCC in neuronal asymmetry regulation in vertebrates, we studied zebrafish anterior dorsal telencephalon (ADt) neuronal axons. We generated transgenic zebrafish animals expressing the photo-convertible fluorescent protein Kaede in ADt neurons and then photo-converted Kaede to label specifically the ADt neuron axons. We found that ADt axons normally project ventrally. Knock down of Dcc function by injecting antisense morpholino oligonucleotides caused the ADt neurons to project axons dorsally. To examine the axon projection pattern of individual ADt neurons, we labeled single ADt neurons using a forebrain-specific promoter to drive fluorescent protein expression. We found that individual ADt neurons projected axons dorsally or formed multiple processes after morpholino knock down of Dcc function. We further found that knock down of the Dcc ligand, Netrin1, also caused ADt neurons to project axons dorsally. Knockdown of Neogenin1, a guidance receptor closely related to Dcc, enhanced the formation of aberrant dorsal axons in embryos injected with Dcc morpholino. These experiments provide the first evidence that Dcc regulates polarized axon initiation and asymmetric outgrowth of forebrain neurons in vertebrates.


Introduction
Neurons are polarized cells. The acquisition of correct polarity is a critical step in neuronal differentiation and subsequent circuit formation. Intracellular kinases and polarity proteins regulate cytoskeletal dynamics and intrinsic cellular programs that determine neuronal polarity [1][2][3][4][5][6][7][8]. In vivo, developing neurons encounter extracellular cues that promote asymmetric growth and direct the formation of correct polarity [9][10][11][12][13][14]. In Caenorhabditis elegans, the guidance receptor UNC-40 (DCC, deleted in colorectal cancer) is localized to the ventral side of the hermaphrodite specific neurons (HSN) before they extend axons ventrally towards the ventral nerve cord, where the UNC-6 (netrin) guidance cue is secreted [11]. Mutations in UNC-40 may result in randomly oriented neuronal processes. In unc-40 loss of function mutants or in unc-40 missense plus unc-6 double mutants, HSN axons extend anteriorly, posteriorly, or dorsally, instead of towards the correct ventral position [15]. These studies suggest that the guidance receptor DCC is a key regulator of neuronal asymmetry in C. elegans. The functions of DCC in axon guidance and cell migration are conserved in bilateria [16][17][18][19], but few studies have examined the roles of DCC in neuronal polarity regulation in vertebrates.
We studied Dcc function in the zebrafish forebrain and found that it regulates asymmetric neuronal growth. Utilizing transgenic animals expressing fluorescent proteins, we determined that anterior dorsal telencephalic (ADt) neurons project axons ventrally. ADt neurons express Dcc before the axons emerge, and the Dcc ligand, Netrin1, is expressed in ventrally located cells. We found that knock down of Dcc function by antisense morpholino oligonucleotides causes ADt neurons to project aberrant axons dorsally. We further examined the axon projection patterns of individual ADt neurons and found that ADt neurons project axons dorsally or form multiple processes after knockdown of Dcc function. In addition, we found that knock down of the Dcc ligand, Netrin1, also caused the ADt neurons to project axons dorsally. Knockdown of Neogenin1, a guidance receptor closely related to Dcc, enhanced formation of aberrant dorsal axons in embryos injected with Dcc morpholino. These data show that Dcc regulates polarized axon initiation and asymmetric outgrowth of neurons in a vertebrate forebrain.

ADt Neurons Project Axons Ventrally
In the developing zebrafish forebrain, a cluster of cells are specified as neurons in the dorsal rostral region of the neural tube. By the pharyngula period, these telencephalic neurons have extended axons along a commissural tract (the anterior commissure, AC) and a descending tract (the supraoptic tract, SOT) [20]. In the anterior dorsal telencephalon, ADt neurons express lhx5 (Fig. 1A), which encodes a LIM homeodomain protein that regulates the antagonism of Wnt signaling, thus promoting forebrain development [21]. To visualize the lhx5 expressing ADt neurons and their axons in live intact animals, we generated transgenic zebrafish expressing the photo-convertible fluorescent protein Kaede [22] using a bacteria artificial chromosome (BAC) homologous recombination method [23,24]. The BAC clone carried approximately 200 kb of zebrafish genomic sequence, of which approximately 110 kb was upstream of the transcriptional start site of the lhx5 gene (Fig. 1B). The modified BAC construct contained Kaede coding sequence that replaced the first exon of the lhx5 gene. Two lines of Tg(lhx5BAC:Kaede) stable transgenic fish were obtained after screening 80 of the modified BAC injected fish. The expression patterns of the Kaede protein were identical between the two lines. We maintained one of the Tg(lhx5BAC:Kaede) lines (referred to as lhx5:Kaede hereafter).
By the mid-pharyngula period at 36 hpf, fluorescent signal from transgenic Kaede expression was observed in the somata of the ADt neurons and axon tracts in the forebrain (Fig. 1C). To ascertain which axon tracts originated from the ADt neurons, we photo-converted Kaede protein in the ADt neuronal somata from the green light emission form to the red light emission form. Diffusion of the photo-converted, red colored Kaede from the ADt somata into the axons showed that the ADt neurons project axons ventrally into the AC and the SOT (Fig. 1C).

ADt Neurons Express dcc
The guidance receptor Dcc and its ligand Netrin1 are expressed in the central nervous system during zebrafish embryonic development [25][26][27][28]. During the pharyngula period (at 24 hpf and 36 hpf), dcc is expressed in dorsal telencephalic regions ( Fig. 2A). Double labeling by fluorescent in situ hybridization demonstrated that dcc is co-expressed with lhx5 in the ADt neurons at 20 hpf, a stage when the lhx5:Kaede labeled ADt neurons are still migrating and lack axonal processes (Fig. 2B). During these periods of development, netrin1a and netrin1b, the duplicated genes orthologous to the mammalian Netrin1 gene, are expressed in the ventral midline region of the forebrain and ventral regions along the SOT ( [27,28], and data not shown). These expression patterns indicate that the guidance receptor Dcc is expressed in the right place at the right time to direct the asymmetric growth of the ADt neurons.

DCC is Required For Correct Asymmetric Growth of the ADt Neurons
To inhibit Dcc function in zebrafish, we injected a morpholino antisense oligonucleotide (dcc-MO) that blocked Dcc protein synthesis [29]. In embryos injected with dcc-MO (6 ng per embryo), the expression of Dcc protein was reduced to levels below the detection limits of our Western blot analyses (Fig. 3A). Photo-conversion of Kaede in dcc-MO injected lhx5:Kaede animals revealed that ADt neurons sent long processes dorsally, in addition to the normal ventrally projecting axons (Fig. 3B). These dorsally projecting processes sometimes occupied lateral positions and extended posteriorly, or they crossed the midline and extended into the contralateral side of the telencephalon (Fig. 3B). Injection of dcc-MO did not affect ADt neuron migration from the medial region to the lateral region of the telencephalon. The dorsally projecting processes were rarely observed in animals injected with a standard control morpholino. To summarize and compare our results, we analyzed each confocal data file and quantified the pixel intensity values of aberrant dorsal processes using custom written ImageJ macro and MATLAB scripts (Fig. S1). To present the quantified results, we assigned phenotypic scores (Grade 0-3) to each embryo based on the pixel intensity value (Fig. S1). Embryos with more pronounced dorsal processes were assigned higher scores (Fig. 3C). Average phenotypic scores (a.p.s.) were calculated for each treatment group where a higher a.p.s. reflected more severe aberrant dorsal projections. For statistical analyses, we used the non-parametric Mann-Whitney U method to test whether samples of two independent observations had equally severe phenotypic values. The rank based Mann-Whitney U test had the advantage that the statistical conclusions were independent of the specific phenotypic score matrix we employed. Based on this analysis, the dcc-MO injected embryos (a.p.s = 0.983) were significantly different from the standard control morpholino injected animals (a.p.s = 0.063) (Fig. 3C, p = 7.01610 27 ). Similar statistical results were obtained using the original quantification data and the ANOVA method (Fig. S1). Injection with a lower dosage of dcc-MO (3 ng per embryo) reduced the severity (a.p.s. = 0.500) of the aberrant dorsal axons. Injection with a higher dosage of dcc-MO (12 ng per embryo) caused early death in about 30% of the injected embryos (52 out of 156). The intensity of the abnormal dorsal axon was increased in the remaining embryos, but the differences were not statistically significant (a.p.s = 1.13, p = 0.425 versus embryos injected with 6 ng of dcc-MO). Embryos were injected with 6 ng of dcc-MO in following experiments.
We used two methods to validate the specificity of the dcc-MO knockdowns. We examined phenotypes caused by injection of a mismatch control morpholino, and we tested for rescue of the dcc knockdown phenotypes by coinjection of a dcc-MO resistant mRNA (dcc-R-mRNA, Fig. 3C). We designed a mismatch control morpholino that contained 5 mismatch bases. In embryos injected with the mismatch control morpholino at the same dose as the dcc-MO, we rarely observed dorsally projecting processes (a.p.s. = 0.065, p = 1.31610 26 versus dcc-MO injected group). To construct the dcc mRNA whose translation was resistant to the blocking effects of the dcc-MO, we introduced 7 silent mutations into the dcc mRNA sequence. After co-injection of the dcc-MO and the dcc mRNA with silent mutations, the occurrence and severity of the dorsal processes were significantly reduced (a.p.s. = 0.281, p = 4.62610 24 versus dcc-MO injected embryos). Similar statistical results were obtained using the original quantification data and the ANOVA method (Fig. S1). These results confirmed that aberrant ADt axon formation was due to specific inhibition of the function of Dcc.

Knocking down Dcc Function Causes ADt Neurons to Project Axons Dorsally or to form Multiple Processes
Because our Kaede photo-conversion results could not discriminate whether the ADt neurons extended single aberrant or multiple processes in the dcc-MO injected embryos, we labeled single ADt neurons with a transient transgenic method (Fig. 4A). Plasmids injected into early stage zebrafish embryos are mosaically distributed among the dividing cells, and mosaic expression of transgenes encoded by the injected plasmids may result in random labeling of single or small numbers of cells in the injected embryos [30,31]. We coinjected a plasmid carrying emx3:Gal4FF and a plasmid carrying UAS:tdTomato, with and without dcc-MO into lhx5:Kaede embryos at the one cell stage. The emx3:Gal4FF plasmid carries a Gal4 transcriptional activator variant [32] under control of a forebrain specific enhancer element from the emx3 gene [33]. The UAS:tdTomato plasmid carries tdTomato, a red fluorescent protein gene [34] under control of 5 copies of the Gal4 recognition element UAS. By screening the injected embryos for axons that can be traced and assigned to single labeled ADt neurons, we obtained ADt neuronal projection patterns resolvable at single cell resolution (Fig. 4B). Among the labeled ADt neurons in control animals, none had dorsally projecting axons (Fig. 4B, n = 80). In contrast, about 40% (19 out of 50) of the labeled ADt neurons in dcc-MO injected animals had either dorsally projecting axons or multiple processes (Fig. 4B). The dorsally projecting axons extended laterally or crossed the midline to extend into the contralateral telencephalon (n = 7). The axons with multiple processes often projected both dorsally and ventrally (n = 12). We further examined all single confocal slices to determine the origins of the axons on the surface of the labeled neurons (Fig. 4B). In control embryos, the origins of the axons were found overwhelmingly on the ventral side of the cell bodies (92%, 74 out of 80). The other 6 labeled neurons either had the axon origins on the dorsal side (4%, n = 3) or close to the middle (4%, n = 3) of the cell bodies. In contrast, for those labeled neurons that had dorsally projecting axons in dcc-MO injected embryos, the origins of the axons were found on the dorsal side of the cell bodies (84%, 16 out of 19). Thus, the single cell labeling results mirrored our results obtained from group labeling, and together these results suggest that Dcc function is required for the asymmetric outgrowth of the ADt axons. Furthermore, because the labeled ADt neurons always extend single axons in control animals, the formation of ADt neuron axons with multiple processes in dcc morpholino injected embryos ( Figure 4C) suggest that Dcc function is required for polarized axon initiation of the forebrain neurons in vivo.

Inhibition of Netrin1 Function Affects the Asymmetric Growth of the ADt Neurons
To provide further evidence for the role of Dcc in the asymmetric growth of ADt axons, we examined the effect of knocking down Netrin1, a Dcc ligand. In lhx5:Kaede embryos coinjected with ntn1a and ntn1b morpholinos (referred to as ntn1-MO, 6 ng each per embryo), ADt neurons developed aberrant dorsal projections similar to neurons in dcc-MO injected embryos (Fig. 5A, a.p.s. = 0.857, p = 0.820 versus dcc-MO injected embryos). Injection of either ntn1a-MO or ntn1b-MO alone didn't cause ADt neurons to project axons dorsally. Co-injection of ntn1a and ntn1b mismatch control morpholinos didn't cause ADt neurons to    Fig. 5B and Fig. S2). A lower dose of ntn1-MO (3 ng each versus 6 ng each per embryo) was used in the Dcc-Netrin1 double knockdown experiment due to concerns over potential toxicity of high dose morpholino injections.
To test the genetic interaction between the Dcc and Netrin1 knockdowns further, we examined the synergistic effect of Dcc-Netrin1 morpholino injections at sub-threshold concentrations that did not cause significant aberrant axon phenotype when injected alone [35,36]. We first generated dose response curves for the Dcc and Netrin1 single knockdowns (Fig. S3). When a subthreshold dose of dcc-MO (1.5 ng per embryo) was injected, 9.

Neogenin1 Knockdown Enhances the Aberrant ADt Axons in dcc-MO Injected Embryos
The guidance receptor Neogenin is closely related to DCC [37,38]. Neogenin1 is expressed in the central nervous system and the mesoderm throughout early development in zebrafish embryos [39,40]. Previous studies showed that knock down of Neogenin1 function by morpholino injection caused abnormal development of the somite and the central nervous system in zebrafish embryos [40,41]. We examined whether inhibition of Neogenin1 function also affected the asymmetric growth of ADt neurons. In embryos injected with neogenin1 morpholino (neo-ATGMO, 2 ng per embryos [40]), the aberrant dorsal ADt neuronal projections (a.p.s. = 0.091) were rarely observed and were not significantly different from those in the standard control morpholino injected embryos (p = 1.000). Injection of higher dosages of neo-ATGMO (4 ng per embryos) caused reductions of the AC or both the AC and the SOT (Fig. 5A), whilst it didn't result in significant aberrant dorsal projections (a.p.s. = 0.053, p = 1.000 versus standard control morpholino injected embryos). Similar to previous report [40], increasing the dosage of the neo-ATGMO further (8 ng per embryo) caused malformation of the neural tube and the ADt neurons (data not shown). When we co-injected both the neo-ATGMO (2 ng per embryo) and the dcc-MO (6 ng per embryo) into the lhx5:Kaede embryos (Fig. 5A), the aberrant ADt neuronal projections were enhanced (Fig. 5B, a.p.s. = 1.800, Fig. 5B and

DCC Regulates Polarized Initiation and Asymmetric Outgrowth of Forebrain Neuronal Axons
Our studies demonstrate that Dcc signaling plays a central role in asymmetric outgrowth of ADt neuronal axons in the zebrafish forebrain. Dcc and its ligand, Netrin1, are expressed prior to the time that ADt neurons extend their axons ventrally (Fig. 2). Knock down of Dcc function caused ADt axons to extend aberrantly (Fig. 3). These incorrect axonal projections are not likely due to abnormal specification of the ADt neurons, because the neurons migrated to their correct locations and differentiated along their normal time course. Single cell labeling showed that ADt neurons projected axons dorsally or formed multiple axons (Fig. 4). Because the labeled ADt neurons always extend single axons in control animals, the formation of ADt neuron axons with multiple processes in dcc morpholino injected embryos suggests that Dcc signaling is required for polarized axon initiation of the forebrain neurons in vivo. Previous studies in C. elegans showed that the correct asymmetry of HSN neurons was lost in unc-40 (dcc) mutants [11] and HSN axons deviated from their correct ventral trajectory to extend dorsally [15]. Thus, our results in zebrafish mirror those observed in C. elegans, suggesting that the functions of Dcc in asymmetric outgrowth of neurons are likely conserved in metazoans.
Previous studies have shown that DCC signaling may promote axonal branching in cultured cortical pyramidal neurons [42], and that disrupting DCC function leads to a decrease in the number of axon terminals in cultured dopaminergic neurons [43]. In Dcc knockout mice, ventral mitral cell axons deviated from their normal path and extended towards more dorsal locations [44]. A recent study showed that cultured hippocampal neurons developed multiple axons when DCC expression was reduced in the context of Bcl11A knockdown [45]. These observations, together with our results, suggest that DCC's function in neuronal polarity regulation may depend on the particular environment experienced by developing neurons.
Our studies show that DCC, Netrin1, and Neogenin1 play overlapping and differential roles in the asymmetric outgrowth of ADt neuronal axons. Inhibition of the function of either Dcc or Netrin1 caused ADt neuron to project axons dorsally, whereas inhibition of Neogenin function didn't result in significant dorsal ADt axons (Fig. 5). These results suggest that Netrin1-Dcc signaling plays a primary role in directing the asymmetric outgrowth of forebrain neuronal axons. Interestingly, knock down of Dcc and Neogenin1 together enhanced the formation of aberrant dorsal ADt axon. This result suggests that Neogenin1 may act in parallel with Dcc and that Neogenin may partially compensate for loss of Dcc function in dcc-MO injected embryos. Because knock down of Dcc and Neogenin1 together caused more aberrant dorsal ADt axons than knock down of Dcc and Netrin1 together, it is likely that other cues in addition to Netrin1 mediate asymmetric outgrowth of the ADt axons. Netrin2 and Netrin4, two Netrin family member proteins that may interact with Dcc in zebrafish, are not expressed in the forebrain prior to 36 hpf [46] and thus are unlikely to act with Netrin1 as ADt guidance factors. On the other hand, RGMa, Draxin and MADD-4 (ADAMTL1 and ADAMTL3) are known to interact genetically or physically with Dcc or Neogenin1 [41,[47][48][49]. Further studies are needed to determine whether these additional factors are involved in the asymmetric outgrowth of the ADt neurons.

Transgenic Labeling of Forebrain Axons
Few previous studies have examined the roles of DCC in neuronal polarity regulation in vertebrates. We took advantage of transgenic labeling methods to investigate function of Dcc in ADt axon development in zebrafish. The Kaede photo-conversion method allowed specific labeling of ADt axons. Such specific labeling should facilitate future studies of gain and loss of function of other genes that may be involved in ADt axon development. Due to diffraction and scattering of the laser beam in live intact animals, other photo-convertible fluorescent proteins such as KikGR [50], in combination with two-photon conversion may be required to study cell population located in deeper regions.
Our study also utilized single cell labeling by generating mosaic transgenic animals. Single cell labeling provides increased specificity and should be useful in studies of cell migration and neural circuit formation. The mosaic labeling method requires injection and screening of many animals; in our studies, fewer than 10% of the injected embryos yielded useful information and screening by confocal microscopy was time consuming. Although such analyses are much more efficient in C. elegans, where stereotyped neurons can be observed routinely at single cell resolution, further analyses of single cells in vertebrate nervous systems are warranted due to the large numbers of different cells types and neuronal connections.

Fish Maintenance
Zebrafish were maintained in a recirculating water system according to standard protocol [51]. Recirculated water was dosed to 500mF salinity with artificial ocean salt mix and buffered to pH7.2 with NaHCO 3 . Embryos were obtained by breeding adult fish and raised at 28.5uC as described [51]. Embryos were staged by hours post fertilization (hpf) as described [52]. The Fudan University Institutional Animal Care and Use Committee approved all work with zebrafish animals.

Synthetic mRNA and Morpholinos
Capped mRNA was synthesized with mMESSAGE mMA-CHINE kit (Ambion). The dosage for mRNA injections was 350 pg of dcc-R-mRNA per embryo.
Generation of Tg(lhx5BAC:Kaede) Transgenic Line A zebrafish bacteria artificial chromosome clone (BAC) containing the lhx5 coding regions (cloned id: CH211-230H6) was identified using a genomic database search (http://www.ensembl. org/Danio_rerio/Info/Index) and fingerprinting analysis (http:// www.sanger.ac.uk/Projects/D_rerio/WebFPC/zebrafish/large. shtml). Targeted modification of the BAC with the photoconvertible fluorescent protein Kaede [22] was based on homologous recombination [23], and microinjection of BAC DNA was previously described [24]. To generate the targeting fragments for homologous recombination, plasmid DNA containing Kaede and a kanamycin resistant cassette flanked by FRT-sites were used as templates with oligonucleotide primers kaede-F (59-CCCAATCC-GAGGCGCGGTGCCCGTTTGGGAAGGTCG GCTCTCGGTCCCC TCCAGGGCGGACACCATGAGTCTGATTAAACCAG) and kaede-R (59-GAGGATTAGTCCAGAAGTTTGACGAGGTGATG-CAGCCAGAATGCAAGCAGCTTCGC GTCTTCTATTCCA GAAGTAGTGAGGAG). For these primers, nucleotides in italics are homologous to the targeted lhx5 gene, and those underlined are homologous to the plasmid DNA templates. The kanamycin resistant marker was removed by induction of Flpe expression [23]. Homologous insertion of the transgenes and the removal of the kanamycin marker were verified by PCR analysis. Integrity of the modified BAC was checked by fingerprinting analysis. The Tg(lhx5BAC:Kaede) line was maintained by outcross and incross protocols in our fish facility.

Photo-conversion of Kaede
Photo-conversion of Kaede was performed with a confocal microscope (Olympus FV1000) equipped with SIM module and 405 nm laser. Live embryos were mounted with 1.2% low melting temperature Agarose and the ADt region on one hemisphere of the forebrain was marked with the polygon selection tool of the FluoView software (Olympus). Ten pulses of focused 405 nm laser light were then applied to the selected region. Each pulse lasted 500 msec and utilized 60% of the light output of the 405 nm laser. Confocal image stacks were acquired 30 min later after the converted Kaede diffused into the neuronal processes.

Western Blot
Embryos were deyolked in cold Ringer's solution supplemented with 1 mM EDTA, 0.3 mM PMSF and protease inhibitor cocktail (Roche). Deyolked embryos were lysed in SDS-sample buffer and heated at 95uC to extract and denature protein. Protein samples were resolved on SDS-PAGE gel then transferred to nitrocellulose membrane (Whatman). Rabbit zebrafish Dcc antibodies (Ana-Spec) were used at 1:50. The specificity of the Dcc antibodies were verified by cross checking with antibodies against Myc (AbMart) and GFP (Invitrogen) for heterologous expression of Myc-and GFP-tagged Dcc. Endogenous Dcc protein was detected as a band of approximately 170 kb.

Genetic Single Cell Labeling
A mixture of two plasmids 25.2emx3:Gal4FF and 5xUAS:tdTomato (10 ng/ml each) was injected into one-cell stage embryos. Injected embryos were screened by fluorescent dissection microscopy at 24 hpf and then by confocal microscopy at 36 hpf to identify animals with axons that can be traced and assigned to labeled single cells. To label cells in dcc morpholino injected animals, dcc morpholino was injected at one-to four-cell stage after the plasmid injection.

Microscopy and Image Analysis
Whole mount in situ images were acquired on a dissection microscope with a CCD camera (Leica M205FA). For confocal microscopy, live or fixed embryos were mounted with 1.2% low melting temperature Agarose and imaged on an Olympus FV1000 confocal laser scanning system with 40x water immersion objective. The captured image stacks were converted by ImageJ (http://rsbweb.nih.gov/ij/) for reconstruction with FluoRender software (http://www.sci.utah.edu/software/46-documentation/ 137-fluorender.html). Reconstructed three dimensional images were projected to a standardized view for figure presentation. To quantify the aberrant projection phenotypes, confocal data files were automatically processed by custom written ImageJ macro and MATLAB scripts (Fig. S1, Text S1, and Text S2). For statistical analysis, non-parametric Mann-Whitney U tests and ANOVA tests were performed with SPSS software. Two-tailed exact p values were reported for the Mann-Whitney U tests and significance P values were reported for the ANOVA tests (SPSS Inc.). Figure S1 Quantification of aberrant dorsal axons. (A) Procedure of quantification. The ImageJ marco extracts image stacks, selects the converted Kaede channel, uses maximum intensity method to project the image stacks along the Z axis, and saves the projections as gray scaled images. The processed images were manually cropped to rectangles containing aberrant axons and the cell bodies were masked with GIMP software. The MATLAB scripts use an adaptive thresholding method to index the axon images then calculate the sum of the pixel intensity values corresponding to the indexed axonal positions. One example of the indexed axon image and its pixel intensity value are shown below the rendered confocal image. The indexed axon positions are in white. (B) Quantification and statistics test results. Increased shading corresponds to Grade 0-3. Grade 0: pixel intensity value less than 5000; Grade 1: less than 25000; Grade 2: less than 45000; Grade 3: greater than 45000. Statistic test results (F and P values) with the quantification data and the ANOVA methods are given at the bottom. Post Hoc test results between dcc-MO injected and other experimental conditions are given on the header lines.