Early Gene Expression During Natural Spinal Cord Regeneration in the Salamander Ambystoma mexicanum
Journal: Journal of Neuroscience
Manuscript ID: JN-RM-2849-06
Manuscript Type: Regular Manuscript
Manuscript Section: Development Plasticity Repair
Date Submitted by the Author:
05-Jul-2006
Complete List of Authors: Monaghan, James; University of Kentucky, Biology & SCoBIRC Walker, John; University of Kentucky, Biology & SCoBIRC Page, Robert; University of Kentucky, Biology & SCoBIRC Putta, Srikrishna; University of Kentucky, Biology & SCoBIRC Beachy, Christopher; Minot State University, Biology Voss, Stephen; University of Kentucky, Biology & SCoBIRC
Keywords:salamander, microarray, Regeneration, In situ hybridization, Spinal cord, axolotl, ependymal, real time PCR
Themes & Topics:c. Regeneration: CNS < 7. Transplantation and Regeneration < Theme A: Development
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
1
Section: Development/Plasticity/Repair Senior Editor: Dr. Pat Levitt
Title: Early Gene Expression During Natural Spinal Cord Regeneration in the Salamander
Ambystoma mexicanum
James R. Monaghan1, John A. Walker1, Robert Page1, Srikrishna Putta1, Christopher K. Beachy2,
and S. Randal Voss1
1Department of Biology & Spinal Cord and Brain Injury Research Center, University of
Kentucky, Lexington, KY 40506
2Department of Biology, Minot State University, Minot SD 58707
Corresponding Author: S. Randal Voss
Department of Biology & Spinal Cord and Brain Injury Research Center, University of
Kentucky, Lexington, KY 40506 ([email protected])
Figures: 3
Tables: 6
Pages: 49
Keywords (8): Salamander, microarray, regeneration, in situ hybridization, spinal cord,
axolotl, ependymal, real time PCR
Acknowledgments: The project described was supported by the Kentucky Spinal Cord Injury
Research Trust and Grant Number 5-R24-RR016344-06 from the National Center for Research
Resources (NCRR), a component of the National Institutes of Health (NIH).
Page 1 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
2
Abstract
In contrast to mammals, salamanders have a remarkable ability to regenerate their spinal cord
and recover full movement and function after tail amputation. To identify genes that may be
associated with the greater regenerative ability of salamanders, we designed a custom
oligonucleotide microarray and profiled early gene expression during natural spinal cord
regeneration in Ambystoma mexicanum. We sampled tissue at five early time points after tail
amputation and identified genes that registered significant changes in mRNA abundance during
the first seven days of regeneration. A list of 1,036 statistically significant genes was identified.
Additional statistical and fold change criteria were applied to identify a smaller list of 360 genes
that were used to describe predominant expression patterns and gene functions. Real-time RT-
PCR was used to validate gene expression and in situ hybridization was used to localize mRNAs
of ten genes to specific tissues and cell populations of the spinal cord. Most gene expression
profiles consisted of a single, significant deflection from baseline followed by relatively constant
mRNA abundance. Our results show that a robust and diverse injury response is activated in
concert with extra-cellular matrix remodeling mechanisms during the early acute phase of natural
spinal cord regeneration. We also report gene expression similarities and differences between our
study and studies that have profiled gene expression after spinal cord injury in rat. Our study
illustrates the utility of a salamander model for identifying genes and gene functions that may
enhance regenerative ability in mammals.
Page 2 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
3
Introduction
Salamanders have a remarkable ability to regenerate complex body parts including the limb, tail,
lens, and central nervous system (CNS). Although salamander regeneration has been studied for
several hundred years (Spallanzani, 1768; Müller, 1864), molecular-level studies have been
limited to a relatively few important transcription factors and signaling molecules that are highly
conserved among vertebrates, and in some cases metazoans (e.g. Schnapp et al., 2005;
Christensen et al., 2002; Carlson et al., 2001; Caubit et al., 1997; Torok et al., 1999). Broader
assessments of gene expression during salamander regeneration may identify mechanisms that
can be exploited to enhance regenerative ability in humans.
Salamanders regenerate their spinal cords and regain full movement and function after
tail amputation. Within a few hours of amputation, injury responses are initiated to increase cell
survival and transform the tissue-damaged environment into one that is permissive for repair and
subsequent regeneration. It is possible that the unrivaled regenerative ability of salamanders is
due in part to this early injury response phase of regeneration, but very little is known about early
response genes and associated biological processes. Most attention has been directed to
understand cellular and developmental changes during the dramatic and conspicuous de-
differentiation and re-patterning phases of regeneration. During de-differentiation, cells of
mesodermal origin (muscle, dermal fibroblasts, and cartilage) re-enter the cell cycle and
proliferate to form a mass called the blastema (Hay and Fischman, 1961). Blastemal cells
subsequently re-differentiate into mesodermal tissues but apparently do not contribute to the
regenerating spinal cord. Epithelial cells (ependymoglia) of the ependymal lining that surrounds
the central canal of the spinal cord re-form neural tissues of the regenerating spinal cord
(Nordlander and Singer, 1978; Egar and Singer, 1972). The signals that initiate and maintain the
Page 3 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
4
proliferative response of ependymoglia are largely unknown, however recent studies implicate
some of the same highly conserved genes that are known to regulate the proliferation and
differentiation of neural stem cells among vertebrates (O’Hara and Chernoff, 1994; Zhang et al.,
2000, 2002; Schnapp et al., 2005; Chernoff et al., 2003). This suggests that some aspects of
salamander spinal cord regeneration may be shared with organisms that have little or no potential
for neural regeneration. Analyses of gene expression in salamanders may point to key similarities
and differences that are associated with regenerative ability (Alvarez-Buylla et al., 2000;
Stevenson and Yoon, 1981; Takahashi et al., 2003; Zhang et al., 2003).
We designed a custom Affymetrix GeneChip® and performed the first microarray
analysis of spinal cord regeneration in the Mexican axolotl (Ambystoma mexicanum). We
sampled regenerating spinal cord tissue at five early time points after amputation and identified
differentially expressed genes and temporal patterns of gene expression. We compared our lists
of significantly regulated genes to lists that have been similarly compiled from microarray
studies of spinal cord injury in rat. Our results highlight genes and gene expression patterns that
are associated with the salamander’s natural ability to regenerate spinal cord.
Materials and Methods
Animals, tissue collection, and RNA isolation
The handling and surgical manipulation of all salamanders was carried out according to Animal
Care and Use guidelines at the University of Kentucky (IACUC #00609L2003). The caudal 1/3
of the tail was amputated from 225 Mexican axolotl sibs (avg. 6.2 cm snout-vent length) from an
inbred strain generated by the Voss laboratory. Spinal cord tissue was collected approximately
1.0 mm rostral to the injury plane at 0, 1, 3, 5, and 7 days post amputation. An average of 1.7 µg
Page 4 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
5
total RNA was extracted from pools of nine tissues for each of five replicates that were collected
at each time point. Labeled probes were produced for the 25 RNA samples and each was
hybridized to a separate Affymetrix GeneChip® (Santa Clara, CA). Probe labeling, GeneChip®
hybridization, and scanning were performed by a single staff member of the University of
Kentucky Microarray Core Facility.
Development of a microarray platform
A custom Ambystoma Affymetrix GeneChip® was designed from curated expressed sequence tag
(EST) assemblies for A. mexicanum and A. t. tigrinum (http://salamander.uky.edu/ESTdb; Smith
et al., 2005). These ESTs are enriched for genes expressed in neural and regenerating tissues
(Putta et al. 2004). The array contains 4,844 total probe sets, 254 of which are controls or
replicate probe sets. Of the remaining 4,590 probe sets, all but 188 correspond to unique A.
mexicanum contigs, of which 2,960 are significantly identical in nucleotide composition (e-7;
BLASTX) to a human sequence in the non-redundant, RefSeq protein database. Significant
salamander-human blast hits were considered gene orthologs in our analyses and we assumed
that salamander-human orthologs have similar gene functions or ontologies.
Assessment of GeneChip® precision
The repeatability of probe set estimates of hybridization intensity was evaluated between
microarray GeneChip®s. We examined the correlation of hybridization intensities across all
probe sets among GeneChip®s that were replicated for each regeneration time point (average
across all replicates, r = 0.994; range r = 0.983-0.998). These results demonstrate that we were
able to obtain a high level of repeatability. With respect to general performance measures, a high
Page 5 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
6
percentage (84.7%) of probe sets scored for presence using MAS 5.0; this likely reflects the
biased selection of regeneration associated genes and high quality contigs for probe set design.
Robust Multiarray Averaging (RMA) showed that average background intensity (61.5) and noise
(4.22) were low, and mean and median hybridization intensities were typical of Affymetrix
GeneChip® performance (1,772.8 and 420, respectively). Raw data files can be obtained at
http://www.ambystoma.org and http://www.ncbi.nlm.nih.gov/geo/.
Quality Control and Low Level Analyses.
We used the Bioconductor package affy (www.bioconductor.org) that is available for the
statistical programming environment R (www.r-project.org) to perform a variety of quality
control and preprocessing procedures at the individual probe level (Bolstad et al., 2005a). These
procedures included: (1) generating matrices of M versus A plots for all replicate chips (n = 5
chips for five sampling times), (2) investigating measures of central tendency, measures of
dispersion, and the distributions of all 25 chips via boxplots and histograms, (3) viewing images
of the log2 (intensity) values for each chip to check for spatial artifacts, and (4) viewing an RNA
degradation plot (Bolstad et al., 2005b) that allows for visualization of the 3’ RNA labeling bias
across all chips simultaneously. In addition, we used ArrayAssist Lite software (Stratagene, La
Jolla, CA) to assess several quality control measures that are recommended by Affymetrix
(GeneChip® Expression Analysis: Data Analysis Fundamentals, 2002, www.affymetrix.com)
such as average background, scale factors, and percent present. Next, we processed our data
similarly to the methods of Choe et al. (2005) to determine a probe set intensity value. Briefly,
our processing method consisted of using the MAS 5.0 background correction algorithm, the
quantiles algorithm for probe-level normalization, the MAS 5.0 algorithm for perfect match
Page 6 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
7
(PM)/mismatch (MM) correction, the median polish algorithm for expression summary
generation, and a loess normalization at the probe set level using the GoldenSpike package for R
(Choe et al., 2005; www.ccr.buffalo.edu/halfon/spike/index.html).
Detection of differentially expressed genes and data filtration.
Microarray platforms may not accurately or precisely quantify genes with low intensity values
(Choe et al., 2005; Draghici et al., 2006). Because low intensity genes contribute to the multiple
testing problem that is inherent to all microarray studies, we filtered probe sets whose mean
expression values across all chips (n = 25 per gene) were smaller than or equal to the mean of the
lowest quartiles (25th percentiles) across all chips (n = 25, mean = 6.44, SD = 0.09; data
presented on a log2 scale). Upon performing this filtration step, 3,641 probe sets were available
for significance testing. Probe sets were tested for differential expression via a one-way fixed
effect linear model (intensity = day sampled) using the Fs test of Cui et al. (2005) and
J/MAANOVA software (http://www.jax.org/staff/churchill/labsite/software/anova/index.html).
Initially, we adjusted for multiple testing by setting the false discovery rate (FDR) to 0.01 using
the step-up algorithm of Benjamini and Hochberg (1995). As is shown in Figure 1a, upon
performing this FDR correction, 2771 probe sets of the 3,641 probe sets tested (76.11%) were
selected as differentially expressed. We then took a more conservative approach to our first pass
at selecting differentially expressed genes by setting the family-wise error rate (FWER) to 0.01.
Upon adjusting the FWER to 0.01, 1,273 of the 3,641 genes tested (34.96%) were selected as
differentially expressed (Figure 1b). In order to identify a smaller subset of probe sets to focus
our attention, we prioritized probe sets that were selected as differentially expressed (with the
FWER adjusted to 0.01) that exhibited: (1) a 2-fold (raw intensity scale) change at any time point
Page 7 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
8
relative to day zero and (2) Fs values that were in the upper 50% of these 1,273 genes (Fs >
28.36). Upon performing this final filtering step, 376 probe sets were identified. Three pairs of
probe sets were designed for the same contigs and yielded nearly identical hybridization
intensities. Thus, these intensities were averaged. Contigs that correspond to the same human
gene were combined, yielding a final short list of 360 unique genes (Supplemental Table 1).
Candidate gene lists may differ when different preprocessing algorithms are used to
identify statistically significant genes from oligonucleotide microarrays (Millenaar et al., 2006).
To address this concern, we compared the 376 candidate probe set list above to a 646 probe set
list that was generated using only the robust multi-array average (RMA) algorithm (Irizarry et
al., 2003), One-way analysis of variance (FDR = 0.01), and a 2-fold change (raw intensity scale)
criterion. Only 11 of the 376 candidate probe sets (2.9%) were unique, indicating that our
methodology for identifying candidate genes is largely concordant with other statistical
approaches.
Identification of gene expression patterns
We used the following criteria to define temporal gene expression patterns for the 360 genes that
met statistical and fold-level criteria. For each gene we assigned a score to qualify the mRNA
abundance at each post-amputation sample time (d1, d3, d5, and d7). A gene received a score of
non-significant (N) for each sample time that mRNA abundance was < 2 fold deviant of the day
0 estimate. We refer to the day 0 estimate as the baseline estimate of mRNA abundance. A gene
received a score of up-regulated (U) or down-regulated (D) for the first post-amputation sample
time that mRNA abundance deviated by > 2 fold from baseline. For subsequent sample times,
each gene received one of three possible scores: constant, up, or down. A score of constant (C)
Page 8 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
9
was assigned if the fold level estimate was < 2 fold deviant of the previous U or D estimate (C
was never assigned after N), and > 2-fold deviant from baseline. A score of U or D was assigned
if mRNA abundance deviated again by > 2 fold. Using this scoring system, a gene received a
score of U, D, or N for day 1, and U, D, N, or C for each of the three subsequent time points
(Table 1). To annotate genes, we used multiple databases (GO, KEGG, IHOP, OMIM, etc.) and
searched the literature for information about the expression and functions of each gene that we
identified as significant in our study. We biased our annotations to emphasize possible gene
functions that have been described in regeneration and spinal cord injury research fields.
Identification of genes expressed differently between salamander regeneration and rat spinal
cord injury
A bioinformatics approach was used to identify gene orthologs that are expressed similarly or
differently after salamander tail amputation versus rat spinal cord injury. We used current (as of
May 2006) human Entrez Gene ID’s that were assigned to each annotated probe set on the
Ambystoma chip to identify all presumptive salamander orthologs on RatU34A, B, and C
GeneChip®s. To accomplish this cross-referencing task, we used Resourcerer (Tsai et al., 2001),
a database that allows orthologous genes to be identified among species-specific microarray
resources. This yielded a list of 1,036 probe sets between the Ambystoma and RatU34
GeneChip®s that presumably correspond to 662 unique, orthologous genes. We compared the
expression pattern of each of these genes using results from this study and published studies that
profiled gene expression after spinal cord injury in rat, using RatU34 GeneChip®s (Song et al.,
2001; Carmel et al., 2001; De Biase et al., 2005; Aimone et al., 2004; n = 1,571). For each gene,
we qualified gene expression as either significantly up, significantly down, or non-significant.
Page 9 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
10
We used statistical and fold-level criteria (FWER < 0.01; > 2-fold change) to score 522
salamander genes for these criteria. If a gene was reported as significantly regulated in the rat
studies, we recorded it as such; otherwise we recorded it as non-significant.
Quantitative Real-Time RT-PCR
Ten genes from the microarray experiment were selected for validation by real-time RT-PCR.
These genes yielded a broad range of relative fold change estimates by microarray analysis and
included both possible directions of differential expression. These genes also exhibited a range of
hybridization intensity values; for example, the average intensity value of hes1 ranked among the
bottom 37% of all probe sets while galectin 1 ranked among the top 95%. For each gene, three
reactions were performed using independent RNA samples from the microarray analysis. A
BioRad iScript cDNA synthesis kit (Hercules, CA) was used to synthesize cDNA templates
(poly-T and random hexamer priming). Primers were designed using Primer3 to amplify DNA
fragments from the same gene regions that were used to design corresponding GeneChip® probe
sets (Table 2; Rozen and Skaletsky, 2000). Reactions included cDNA that was synthesized from
10ng total RNA, 300nM primers, and iQ SYBR-Green real-time PCR mix (BioRad). The
reactions were performed on a BioRad I Cycler real-time RT-PCR system. PCR efficiencies
were calculated for each primer pair with a dilution series of 0, 0.1ng, 1ng, 5ng, 10ng, 20ng, and
40ng. The three replicates were normalized against a gene that showed no significant gene
expression change in the microarray experiment (glyceraldehyde-3-phosphate dehydrogenase,
MC01187). PCR efficiencies for each primer were incorporated into the relative fold change
calculations according to the Pfaffl method (Pfaffl, 2001). Student’s t-tests were performed using
the three normalized biological replicates for day 3 and day 0 samples.
Page 10 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
11
In situ hybridization
Digoxigenin (DIG)-labeled RNA probe production and in situ hybridization (ISH) were
performed as described by Hirota et al. (1992) with minor modifications. RNA probes were
synthesized by in vitro transcription using 300-350 base pair PCR products as template. PCR
primers were designed from the same gene region as corresponding GeneChip® probe sets and
included SP6 or T3 RNA polymerase promoters appended to the 5’ ends (Table 2). PCR
products were cleaned using a Qiagen PCR purification column before performing in vitro
transcription. In vitro transcription efficiency was evaluated by electrophoresis of DIG-probes on
1% agarose gels and visualized using ethidium bromide. Axolotl tissues were collected three
days after tail amputation and fixed at 4ºC in 1x PBS, 4% paraformaldehyde overnight. Bone
was decalcified by incubating the tissue in 500 mM EDTA (pH8.0), 1xPBS for at least two days.
Tissue was cryoprotected overnight in 30% sucrose and embedded in O.C.T. medium (Fisher
Sci., Pittsburgh, PA). Sections were cut to 16µm using a Microm 500HM cryostat, adhered to
Superfrost Plus microscopy slides (Fisher Sci.), and stored at -80ºC. Sections were taken within
2.5 mm from the end of the regenerating spinal cord. Hybridization, washing, and colorimetric
detection with NBT/BCIP were performed on a Tecan Genesis Workstation 200 liquid handling
robot with a Genepaint® hybridization station (Zurich, Switzerland). Sections were mounted in
Permount (Fisher Sci.) for microscopy using an Olympus AX70 microscope and image
acquisition using Magnafire 2.1 software.
Genes were classified as expressed in a particular cell population according to the locality
and morphology of the cell within multiple tissue sections. Ependymoglia were identified as cells
lining the central canal with radial morphology and processes extending outward toward the pia
Page 11 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
12
mater. Neurons were identified as large, round cells within the grey matter of the spinal cord.
Cells located in the white matter were grouped as a single category although this is likely a
heterogeneous population of oligodendrocytes, astrocytes, endogenous microglia, and infiltrating
leukocytes. Cells outside the spinal cord resembling blastemal cells or leukocytes were also
classified as a single category (Table 6).
Results
Identification of differentially expressed genes and gene expression patterns
We identified 360 probe sets as detecting significantly different mRNA abundances between day
0 and another time point (Day 1-7), using statistical and fold change thresholds (ANOVA p <
0.01; FWER of 0.01; Fs > 28.36; > 2-fold). More than half of these probe sets (n = 210)
correspond to salamander sequences (genes) that show high sequence identify to a presumptive
human protein-coding locus; the remainder correspond to anonymous EST contigs. In
comparison to day 0 (baseline) mRNA levels, most genes exhibited significantly different
mRNA abundances at two or more post-amputation time points. This temporal variability did not
yield an extensive list of gene expression patterns. Although a total of 100 different gene
expression patterns were possible under our scoring system, only 32 different patterns were
observed and over 85% of all genes were classified into 10 categories (Table 1; Table 3). Eight
of the top ten categories identified groups of genes in which mRNA abundance increased or
decreased at a particular time point, and afterward the level remained constant through day 7.
Transcript levels for a few genes did increase or decrease by > 2 fold among post-amputation
time points, however only three genes (UNND) yielded a temporal expression profile that
deviated significantly from baseline in both up- and down-regulated directions during the seven
Page 12 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
13
day period. Thus, the majority of the gene expression profiles that we examined consisted of a
single, significant deflection from base line levels followed by relatively constant mRNA
abundance. It is likely that many of the uniquely expressed genes at day 7 are regulated at later
time points because only 43 of the 360 genes exhibited transcript levels at day 7 that
approximated baseline. Clearly, we only sampled the initial phases of a continuous gene
expression program that extends beyond day 7. However, our experiment does precisely sample
discrete phases of gene expression variability during this temporal process (Figure 2). Below, we
describe major gene expression patterns in greater detail. We also highlight some of the genes
and gene functions that were found in each of the major gene expression categories. Finally, we
compare the expression of salamander genes to presumptive rat orthologs that have been
examined in microarray studies of spinal cord injury.
Gene expression patterns
Overall, a greater number of genes were up-regulated above baseline during the 7-day period (n
= 235). The majority (n = 134) were significantly up-regulated at the first sample time after
amputation (day 1) and half of these genes (UCCC: n = 64) registered constant mRNA
abundances above baseline at all subsequent post-amputation sample times (d3, d5, and d7). A
similar pattern of expression was observed for a small group of genes (UUCC: n = 12) that
maintained above baseline mRNA abundances after successive 2-fold increases at day 1 and 3.
The remainder of the day 1 up-regulated genes showed decreasing mRNA abundances at later
sample points. Some of these genes yielded mRNA abundances at day 3 (UCNN: n = 5) or day 5
(UCCN, UDNN: n = 5) that approximated day 0 levels, while others remained above baseline
(UDCC, UCDC, UNND, UDDC: n = 22). Up-regulated genes that were significantly down-
Page 13 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
14
regulated at later sample times were down-regulated below baseline level in only three cases
(UNND). The early group of up-regulated genes suggests that a diversity of regulatory pathways
and biological processes are activated within the first 24 hours after tail amputation. In addition
to genes that presumably function in wounding, stress, inflammation, and immunity, this group
includes genes that function in tissue remodeling, apoptosis, ion transport, cell-cell interactions,
cell migration, vitamin B economy, lipid metabolism, and cytoskeleton dynamics (Table 3).
Several different regulatory networks are implicated directly or indirectly among these day 1 up-
regulated responses, including MAPK, WNT, v-MYC, TNF, v-YES, RAS, and TGF-beta.
Other groups of genes were up-regulated for the first time at day 3, 5, and 7 (n = 119).
The majority of the day 3 and 5 genes maintained high, constant mRNA levels at subsequent
time points (NUCC: n = 24, NNUC: n = 44). Gene functions that were observed among day 1
up-regulated genes were also represented among day 3-7 up-regulated genes. However, the
distribution of genes among these functional categories was very different. In particular, fewer
injury response genes and a greater number of extracellular matrix (ECM) and cytoskeleton-
associated genes were observed compared to day 1 up-regulated genes. Also, a greater number of
cell cycle related genes were observed, especially at day 5 (NNUC; n = 14). In addition to genes
that regulate cell cycle entry, progression, and apoptosis, this later group includes genes that
presumably function in DNA replication, metabolism, chromatin assembly, and cytokinesis.
These results suggest that the regeneration gene expression program transitions during the first
seven days from an injury responsive phase to one that is defined primarily by the up-regulation
of genes that function in mitosis. Throughout both injury response and cell proliferation phases,
genes that function in tissue remodeling are significantly regulated.
Page 14 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
15
Relative to the total number of up-regulated genes, a much smaller number of genes (n =
126) were down-regulated significantly below baseline levels during regeneration. In contrast to
the up-regulated gene set, very few of these genes were down-regulated at day 1 (DCCC, DCNN,
DNNN, DCDC = 12). Also, the magnitude of the fold-level changes was generally lower than
those measured for significantly up-regulated genes (mean of maximum up-regulated fold
changes = 6.61; down-regulated = -3.07). The largest number of down-regulated genes was
observed at day 3 (n = 45) and this was followed by additional groups of down-regulated genes
at day 5 (n = 29) and day 7 (n = 35). In general, many genes with neural related functions were
down-regulated, including those that function in ion transport, glutamate metabolism, glutamate
binding, neuroprotection, neurotransmission, neurogenesis, and lipid metabolism. Some of the
same functional categories that were observed among up-regulated genes were also observed
among down-regulated genes, including apoptosis, cytoskeleton, ECM, signal transduction, and
heat shock (Table 3). The overall pattern indicates that fewer genes are down-regulated during
the first seven days of regeneration, and down-regulated genes show significantly lower mRNA
abundances at day 3, after the early up-regulation of genes at day 1.
Some gene expression patterns were more complicated than linear, directional responses,
involving changes in mRNA abundance that fluctuated both above and below the baseline. For
example, the mRNA abundances of 12 genes (UDCC) increased significantly at day 1, decreased
significantly at day 3 (but still above baseline), and then remained constant through day 7. A
similar pattern was observed for eight genes (UCDC) that increased significantly at day 1 and
decreased significantly at day 5. Some of these and other genes (NUCN, UDDC, NDCD,
UUCD) with complex expression patterns may function in the regulation of biological processes
during regeneration. These include genes that function in ECM remodeling (matrix
Page 15 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
16
metalloproteinase (mmp) 1, mmp13, tissue inhibitor of metalloproteinase 1), coagulation (tissue
factor pathway inhibitor 2), vitamin B transport (intrinsic factor, transcobalamin 1), cell
proliferation (v-Ha-ras viral oncogene, hypothetical protein FLJ20303), transcriptional
regulation (jun-b proto-oncogene), and cell signaling (latent TGF beta binding protein;
chromosome 8 orf 4; secreted frizzled-related protein 2 [sfrp2]).
Gene expression after spinal cord injury: salamander verses rat
To identify similarities and differences between the salamander and mammalian spinal cord
injury response, we compared our gene expression results to published results from the rat spinal
cord microarray literature. Specifically, we compared the expression of 662 presumptive rat –
salamander orthologous genes that are represented on both the Ambystoma and rat Affymetrix
GeneChip®s. The resulting list of gene orthologs represents an unbiased sampling of ~24,000
transcripts on the rat chips and 4,590 transcripts on the salamander chip. Although the majority
of gene orthologs were not significantly regulated (n = 553), we identified many similar and
dissimilar gene expression responses between these organisms (Table 4). Eleven genes are up-
regulated in both species, with no common genes down-regulated. There were 46 and 41
uniquely up-regulated genes in the salamander and rat, respectively. Among the 46 uniquely up-
regulated salamander genes, we identified ten that are either up-regulated during tissue
regeneration or are known to enhance tissue regeneration in other organisms. Furthermore, eight
cell cycle genes, a necessary process for true tissue regeneration, are present within these unique
salamander genes including cell division cycle 2, kinesin family member 11, and mitotic arrest
deficient-like 1.
Real-time RT-PCR
Page 16 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
17
Using real-time RT-PCR, we estimated fold change between day 0 and day 3 for ten genes from
the microarray experiment (Table 5). All of the transcripts that met statistical and fold level
criteria from the microarray analysis registered significant differences in mRNA abundance by
real-time RT-PCR (6/10). Overall, we were able to verify nine of the ten gene changes in the
correct direction with close agreement in most cases. We only failed to replicate the microarray
estimate for sox3, which was not significant by real-time RT-PCR and registered such a low fold
change that it was excluded from the short list of microarray gene candidates. Thus, for all genes
that met our stringent statistical and fold level criteria, and three genes that did not, real-time RT-
PCR validated microarray estimates of gene expression with very good precision.
Spatial analysis of mRNAs using in situ hybridization
Ten genes that were significantly regulated during spinal cord regeneration were examined
further by ISH. We probed tissues that were collected three days after tail amputation to localize
expression among cell types that were within 2.5 mm from the end of the regenerating spinal
cord. This spatial range included degenerating, regenerating, and nearly intact tissue, thus
allowing analysis of normal and early-regenerating tissue all within the same animal. Diverse
patterns of spatial expression were observed with hybridization found in the ependymoglia,
neurons, and white matter, as well as among presumptive blastema cells outside the spinal cord
(Table 6; Figure 3). Four of the ten transcripts assayed were expressed primarily in
ependymoglia including annexin A1, mmp9, cytokeratin 18 (ck18), and fibroblast growth factor
binding protein 1. Sfrp2 and thioredoxin were expressed in multiple cell types. Four genes were
highly expressed within inflammatory-like cells in the white matter including apoliprotein E
(apoE), galectin 1, galectin 3, and ferritin heavy polypeptide 1 (Table 6; Figure 3). These results
Page 17 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
18
show that the general increase or decrease in mRNA abundance as determined from microarray
analysis can be replicated and localized to specific cell populations and tissues of the spinal cord
using ISH.
Discussion
We built a custom Affymetrix GeneChip® and profiled gene expression during the early phases
of natural spinal cord regeneration in a salamander model (Ambystoma mexicanum). Our results
show that regeneration involves significant changes in mRNA abundance for many genes that
are represented on the chip. The overall list of 1,273 genes that met a very stringent statistical
criterion is available as a new resource for regeneration and spinal cord injury research fields
(www.ambystoma.org). The large number of genes on this list, which were identified using a
custom microarray with enriched gene content, shows that thousands of genes are significantly
regulated during the first few days of natural spinal cord regeneration. We used additional
statistical and fold change criteria to sample a smaller sub-group of candidate genes to describe
gene expression patterns and biological functions. The presumptive functions of this smaller list
of genes suggest the operation of many biological processes that change temporarily during
spinal cord regeneration. Below we discuss up- and down-regulated genes and gene functions
that may be important in the regenerative response. In particular, we compare our gene
expression results to several studies that have examined rat spinal cord injuries using
microarrays.
Up-regulated gene responses
Page 18 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
19
Similar gene expression changes are often observed after tissue injury, regardless of the type of
injury or specific tissue type examined. With respect to CNS tissue injury in mammals, an early
acute phase is characterized in part by the expression of transcription factors and immune
response genes (Bareyre and Schwab, 2003). Our study identified several genes that change by
day 1 in salamander spinal cord that are also expressed during the mammalian CNS acute injury
response (Vazquez-Chona et al., 2005). These include jun-B proto-oncogene, interferon
regulatory factor 1, heme oxygenase 1, and apoE. Overall, many of our day 1 and day 3 up-
regulated genes encode proteins that participate in immune response functions, including
lymphocyte, platelet and monocyte activation, macrophage differentiation and migration, cell
adhesion, thrombosis, coagulation, inflammation, oxidative and metabolic stress, and apoptosis.
In addition to immune response genes, we also observed the up-regulation of genes that function
in transport and binding of vitamin B and lipids, and ECM remodeling. While processes like
vitamin B homeostasis have received little attention in regeneration and injury fields (Bauer,
1998), the association of lipid turnover and MMP activity is well documented to be associated
with regeneration (Vance et al., 2000; Poirier, 1994; Yang et al., 1999; Vinarsky et al., 2005). In
A. mexicanum, MMP2, MMP9, and MMP1 activity is associated with proliferating ependymal
cells after 2-3 weeks of regeneration (Chernoff, 2000). Our study shows that mmps 1, 3, 9, 13,
27, and tissue inhibitor of metalloproteinase 1 are all highly up-regulated by 24 hours after
injury, which is the first association of mmps 13 and 27 with regeneration in urodeles. Overall,
our results show that a robust and diverse injury response is activated in concert with ECM
remodeling mechanisms during the acute phase of natural spinal cord regeneration.
The majority of the early-activated genes were up-regulated throughout the first seven
days, extending into a subsequent phase of cell cycle-related gene expression at day 5. The
Page 19 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
20
accumulation of mRNAs that increase during the first seven days of regeneration suggests a
temporal change toward biological processes that are associated with cell division. Many genes
up-regulated at day 5 (NNUC) are associated with mitotic cell cycle regulation including four
genes involved in the G2/M transition and six associated with mitosis (Table 3). These gene
expression changes maybe associated with the early proliferation of blastema and ependymal cell
populations, which are known to expand after the first week of regeneration (Lo et al., 1993;
Zhang et al., 2003). Cell cycle-related genes are also up-regulated early after rat spinal cord
injury, but the functions of these genes are associated primarily with S-phase and DNA repair
and expressed in damaged or apoptotic neurons, not proliferating cells (Di Giovanni et al., 2003).
Thus, within a few days after spinal cord injury, cell-cycle gene expression is biased towards cell
death pathways in mammals but cell survival and proliferation pathways in salamanders.
Down-regulated gene responses
In comparison to up-regulated genes, there were fewer down-regulated genes and most showed
gradual changes over time. Multiple genes were down-regulated whose products are associated
with neural functions, including axon guidance, ion transport, glutamate metabolism,
neuroprotection, and neurotransmitter signaling. Changes in neural-related gene expression
patterns may reflect the damage or loss of neural cell types verses the survival, infiltration, and
proliferation of other cell types. This explanation has been advanced to explain the down-
regulation of genes after mammalian spinal cord injury, where there can be extensive tissue
damage and cell loss (Profyris et al., 2004; Bareyre and Schwab, 2003). Indeed, even in the
regenerating salamander, there is local spinal cord tissue loss after injury (Zhang et al., 2003;
Stensaas, 1983). Thus, in both mammals and salamanders, many of the down-regulated gene
Page 20 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
21
expression patterns may reflect the variability and stochastic nature of cell survival at the injury
site. However, we did observe significant down-regulation of several genes that are associated
with glutamate metabolism and transport, that are up-regulated after CNS injury in mammals.
This suggests the possibility that some genes are actively and uniquely repressed during
salamander regeneration.
Identification of genes expressed differently between salamander regeneration and rat
spinal cord injury
We compared genes that changed during early salamander spinal cord regeneration to gene lists
that were compiled from microarray studies of spinal cord injury in rats. We acknowledge that
this comparison is potentially confounded by several sources of variance including experimental,
technical, statistical, tissue, and organismal differences. However, as we described above, some
of the same genes that are expressed early after mammalian CNS injury are also up-regulated
during spinal cord regeneration in salamander. If similar gene expression programs underlie
homologous tissues, then comparisons of homologous tissues among distantly related organisms
may filter conserved gene expression responses and help identify uniquely regulated genes.
Some of the uniquely regulated genes from salamander are associated with regeneration in other
organisms and tissues including amphibian limb regeneration (ck18, Corcoran and Ferretti, 1997;
msx1, Beck et al., 2003; msx2, Koshiba et al., 1998; Carlson et al., 1998; and mmp9, Yang et al.,
1999), fish tailfin regeneration (ck18 and periostin, Padhi et al., 2004), and annelid epimorphic
regeneration (phosphoribosylaminoimidazole carboxylase, Myohara et al., 2006). Several other
up-regulated genes are associated with mammalian liver regeneration, including follistatin
(Borgnon et al., 2005), cystathionase (Teshigawara et al., 1995), laminin alpha 1 (Kikkawa et al.,
Page 21 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
22
2005), transglutaminase 1 (Ohtake et al., 2006), and uncoupling protein 2 (Horimoto et al.,
2004). Up-regulation of the same gene orthologs across multiple regeneration paradigms
suggests that regeneration is definable across taxa and tissues by distinct gene expression
patterns. Further studies are needed to determine if a conserved group of genes function in
molecular pathways that are required for regeneration.
Molecules that regulate morphogenic signaling
Morphogenic molecules such as sonic hedgehog (Shh), bone morphogenic proteins (BMPs),
WNT factors, and fibroblast growth factors (FGFs) have been associated with regeneration
because they establish positional identity, control cell proliferation, and regulate cell fate during
development (Vergara et al., 2005; Whitehead et al., 2005; Zhang et al., 2000; Schnapp et al.,
2005; Beck et al., 2003; Niemann, 2006). In this study, we identified changed genes participating
in BMP, WNT, and FGF signaling. Follistatin, a protein that regulates dorsal-ventral patterning
of the developing vertebrate nervous system through BMP inhibition, is up-regulated during the
first week of regeneration (Table 5; NUCC; Lee and Jessel, 1999). mRNAs for sfrp2, a secreted
WNT antagonist that blocks ligand binding to frizzled receptors (Kawano and Kypta, 2003), is
highly up-regulated in both the early regenerating spinal cord and tail (Table 5; UUCD).
Furthermore, wnt5A demonstrates a 3.86 fold increase in expression (UCCC), suggesting a
network of pro- and negative WNT signaling during the regeneration. Lastly, fibroblast growth
factor binding protein 1, a secreted molecule that sequesters FGF ligands from the ECM (Tassi
et al., 2001), is expressed in ependymoglia and neurons of the uninjured spinal cord (Table 6)
and down-regulated 6.42 fold at 24 hours (DCCC). These large gene expression changes suggest
that BMP, WNT, and FGF signaling pathways are all altered during early spinal cord
Page 22 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
23
regeneration. Further studies with each of these molecules and their corresponding binding
substrates will be needed in order to assess their possible roles during regeneration.
Conclusion
The salamander’s unique ability to regenerate complex body parts has long been recognized as
an important model in developmental biology, however salamanders have received relatively
little attention from researchers of mammalian spinal cord injury. Our study shows that genomic
and bioinformatics resources are now available to associate gene expression changes with
cellular and molecular aspects of natural spinal cord regeneration. The emerging salamander
perspective on regeneration promises to extend existing research paradigms and may suggest
novel therapies for CNS injury in humans.
Acknowledgements
The project described was supported by the Kentucky Spinal Cord Injury Research Trust and
Grant Number 5-R24-RR016344-06 from the National Center for Research Resources (NCRR),
a component of the National Institutes of Health (NIH). Its contents are solely the responsibility
of the authors and do not necessarily represent the official views of NCRR or NIH.
References
Aimone JB, Leasure JL, Perreau VM, Thallmair M, The Christopher Reeve Paralysis Foundation
Research Consortium (2004) Spatial and temporal gene expression profiling of the contused rat
spinal cord. Experimental Neurology 189:204.
Page 23 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
24
Alvarez-Buylla A, Herrera D, Wichterle H (2000) The subventricular zone: source of neuronal
precursors for brain repair. Prog Brain Res 127:1-11.
Bareyre FM, Schwab ME (2003) Inflammation, degeneration and regeneration in the injured
spinal cord: insights from DNA microarrays. Trends in Neurosciences 26:555.
Bauer JA (1998) Hydroxocobalamins as biologically compatible donors of nitric oxide
implicated in the acceleration of wound healing. Medical Hypotheses 51:65.
Beck CW, Christen B, Slack JM (2003) Molecular Pathways Needed for Regeneration of Spinal
Cord and Muscle in a Vertebrate. Developmental Cell 5:429.
Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: a practical and powerful
approach to multiple testing. J Royal Stat Soc Series B 57:289-300.
Bolstad B, Irizarry R, Gautier L, Wu Z. (2005a) Preprocessing high-density oligonucleotide
arrays. In: Bioinformatics and Computational Biology Solutions Using R and Bioconductor
(Gentleman R, Carey V, Huber W, Irizarry R, Dudoit S, ed), pp13-32. New York: Springer.
Bolstad B, Collin F, Brettschneider J, Simpson K, Cope L, Irizarry RA, and Speed TP (2005b)
Quality assessment of Affymetrix GeneChip® data. In: Bioinformatics and Computational
Biology Solutions Using R and Bioconductor (Gentleman R, Carey V, Huber W, Irizarry R, and
Dudoit S, ed), pp33-47. New York: Springer.
Borgnon J, Djamouri F, Lorand I, Rico VD, Loux N, Pages JC, Franco D, Capron F, Weber A
(2005) Follistatin allows efficient retroviral-mediated gene transfer into rat liver. Biochemical
and Biophysical Research Communications 328:937.
Page 24 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
25
Caubit X, Nicolas S, Le Parco Y (1997) Possible roles for Wnt genes in growth and axial
patterning during regeneration of the tail in urodele amphibians. Dev Dyn 210:1-10.
Carlson MR, Bryant SV, Gardiner DM (1998) Expression of Msx-2 during development,
regeneration, and wound healing in axolotl limbs. The Journal of Experimental Zoology
282:715-723.
Carlson MR, Komine Y, Bryant SV, Gardiner DM (2001) Expression of Hoxb13 and Hoxc10 in
developing and regenerating axolotl limbs and tails. Dev Biol 229:396.
Carmel JB, Galante A, Soteropoulos P, Tolias P, Recce M, Young W, Hart RP (2001) Gene
expression profiling of acute spinal cord injury reveals spreading inflammatory signals and
neuron loss. Physiol Genomics 7:201-213.
Chernoff EA, O'Hara CM, Bauerle D, Bowling M (2000) Matrix metalloproteinase production in
regenerating axolotl spinal cord. Wound Repair and Regeneration 8:282-291.
Chernoff E, Stocum D, Nye H, Cameron J (2003) Urodele spinal cord regeneration and related
processes. Dev Dyn 226:295-307.
Christensen RN, Weinstein M, Tassava RA (2002) Expression of fibroblast growth factors 4, 8,
and 10 in limbs, flanks, and blastemas of Ambystoma. Dev Dyn 223:193-203.
Choe SE, Boutros M, Michelson AM, Church GM, Halfon MS (2005) Preferred analysis
methods for Affymetrix GeneChip®s revealed by a wholly defined control dataset. Genome Biol
6:R16.
Page 25 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
26
Corcoran JP, Ferretti P (1997) Keratin 8 and 18 expression in mesenchymal progenitor cells of
regenerating limbs is associated with cell proliferation and differentiation. Developmental
Dynamics 210:355-370.
Cui X, Hwang JT, Qiu J, Blades NJ, Churchill GA (2005) Improved statistical tests for
differential gene expression by shrinking variance components estimates. Biostat 6:59-75.
De Biase A, Knoblach SM, Di Giovanni S, Fan C, Molon A, Hoffman EP, Faden AI (2005)
Gene expression profiling of experimental traumatic spinal cord injury as a function of distance
from impact site and injury severity. Physiol Genomics 22:368-381.
Di Giovanni S, Knoblach SM, Brandoli C, Aden SA, Hoffman EP, Faden AI (2003) Gene
profiling in spinal cord injury shows role of cell cycle in neuronal death. Annals of Neurology
53:454-468.
Draghici S, Khatri P, Eklund AC, Szallasi Z (2006) Reliability and reproducibility issues in
DNA microarray measurements. Trends in Genetics 22:101.
Egar M, Singer M (1972) The role of ependyma in spinal cord regeneration in the urodele,
Triturus. Exp Neurol 37:422-430.
Hay E, Fischman D (1961) Origin of the blastema in regenerating limbs of the newt Triturus
viridescens. An autoradiographic study using tritiated thymidine to follow cell proliferation and
migration. Dev Biol 3:26-59.
Horimoto M, Fulop P, Derdak Z, Wands JR, Baffy G (2004) Uncoupling protein-2 deficiency
promotes oxidant stress and delays liver regeneration in mice. Hepatology 39:386-392.
Page 26 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
27
Hirota S, Ito A, Morii E, Wanaka A, Tohyama M, Kitamura Y, Nomura S (1992) Localization of
mRNA for c-kit receptor and its ligand in the brain of adult rats: an analysis using in situ
hybridization histochemistry. Brain Res Mol Brain Res 1:47-54.
Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP (2003)
Exploration, normalization, and summaries of high density oligonucleotide array probe level
data. Biostat 4:249-264.
Kawano Y, Kypta R (2003) Secreted antagonists of the Wnt signaling pathway. J Cell Sci
116:2627-2634.
Kikkawa Y, Mochizuki Y, Miner JH, Mitaka T (2005) Transient expression of laminin alpha1
chain in regenerating murine liver: restricted localization of laminin chains and nidogen-1. Exp
Cell Res 305:99-109.
Koshiba K, Kuroiwa A, Yamamoto H, Tamura K, Ide H (1998) Expression of Msx genes in
regenerating and developing limbs of axolotl. The Journal of Experimental Zoology 282:703-
714.
Lee KJ, Jessell TM (1999) The specification of dorsal cell fates in the vertebrate central nervous
system. Annu Rev of Neurosci 22:261-294.
Lo D, Allen F, Brockes J (1993) Reversal of muscle differentiation during urodele limb
regeneration. PNAS 90:7230-7234.
Page 27 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
28
Millenaar F, Okyere J, May S, van Zanten M, Voesenek L, Peeters A (2006) How to decide?
Different methods of calculating gene expression from short oligonucleotide array data will give
different results. BMC Bioinformatics 7:137.
Müller H (1864) Ueber regeneration der wirbelsäule und des rückenmarks bei Tritonen und
Eidechsen. Abhandl. d. Senckenbergischen nature-forsch. Gesellsch 5:113-136.
Myohara M, Niva CC, Lee JM (2006) Molecular approach to annelid regeneration: cDNA
subtraction cloning reveals various novel genes that are upregulated during the large-scale
regeneration of the oligochaete, Enchytraeus japonensis. Developmental Dynamics In Press..
Niemann C (2006) Controlling the stem cell niche: right time, right place, right strength.
BioEssays 28:1-5.
Nordlander R, Singer M (1978) The Role of Ependyma in Regeneration of the Spinal Cord in the
Urodele Amphibian Tail. J Comp Neur 18037:349-374.
O’Hara C, Chernoff E (1994) Growth factor modulation of injury-reactive ependymal cell
proliferation and migration. Tissue Cell 26:599-611.
Ohtake Y, Maruko A, Abe S, Fukumoto M, Ohkubo Y (2006) Effect of retinoic acid-induced
transglutaminase on cell growth in regenerating liver. Biomed Res 27:75-80.
Padhi BK, Joly L, Tellis P, Smith A, Nanjappa P, Chevrette M, Ekker M, Akimenko MA (2004)
Screen for genes differentially expressed during regeneration of the zebrafish caudal fin.
Developmental Dynamics 231:527-541.
Page 28 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
29
Pfaffl MW (2001) A new mathematical model for relative quantification in real-time RT-PCR.
Nucl Acids Res 29:e45-.
Poirier J (1994) Apolipoprotein E in animal models of CNS injury and in Alzheimer's disease.
Trends Neurosci 17:525-530.
Profyris C, Cheema SS, Zang D, Azari MF, Boyle K, Petratos S (2004) Degenerative and
regenerative mechanisms governing spinal cord injury. Neurobiology of Disease 15:415.
Putta S, Smith J, Walker J, Rondet M, Weisrock D, Monaghan J, Samuels A, Kump K, King D,
Maness N, Habermann B, Tanaka E, Bryant S, Gardiner D, Parichy D, Voss SR (2004) From
biomedicine to natural history research: EST resources for ambystomatid salamanders. BMC
Genomics 5:54.
Rozen S, Skaletsky H (2000) Primer3 on the WWW for general users and for biologist
programmers. In: Krawetz S, Misener S (eds) Bioinformatics Methods and Protocols: Methods
in Molecular Biology. pp 365-386 Totowa, NJ:Humana Press.
Schnapp E, Kragl M, Rubin L, Tanaka EM (2005) Hedgehog signaling controls dorsoventral
patterning, blastema cell proliferation and cartilage induction during axolotl tail regeneration.
Development 132:3243-3253.
Smith J, Putta S, Walker J, Kump D, Samuels A, Monaghan J, Weisrock D, Staben C, Voss S
(2005) Sal-Site: Integrating new and existing ambystomatid salamander research and
informational resources. BMC Genomics 6:181.
Page 29 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
30
Song G, Cechvala C, Resnick DK, Dempsey RJ, Rao VL (2001) GeneChip® analysis after acute
spinal cord injury in rat. Journal of Neurochemistry 79:804-815.
Spallanzani L (1768) In: Prodromo di un opera da imprimersi sopra le riproduzioni animali dato
in luce G. Montanari, Modena.
Stensaas LJ (1983) Regeneration in the Spinal Cord of the Newt Notopthalmus (Titurus)
Pyrrhogaster. In: Spinal Cord Reconstruction (Kao CC, Reier PJ ed), pp121-150. New York:
Raven Press.
Stevenson J, Yoon M (1981) Mitosis of radial glial cells in the optic tectum of adult goldfish. J
Neurosci 1:862-875.
Takahashi M, Arai Y, Kurosawa H, Sueyoshi N, Shirai S (2003) Ependymal cell reactions in
spinal cord segments after compression injury in adult rat. J Neuropathol Exp Neurol 62:185-
194.
Tassi E, Al-Attar A, Aigner A, Swift MR, McDonnell K, Karavanov A, Wellstein A (2001)
Enhancement of Fibroblast Growth Factor (FGF) Activity by an FGF-binding Protein. J Biol
Chem 276:40247-40253.
Teshigawara M MS, Tsuboi S, Ohmori S (1995) Changes in levels of glutathione and related
compounds and activities of glutathione-related enzymes during rat liver regeneration. Res Exp
Med (Berl) 195:55-60.
Torok MA, Gardiner DM, Izpisua-Belmonte JC, Bryant SV (1999) Sonic Hedgehog (shh)
expression in developing and regenerating axolotl limbs. J Exp Zool 284:197-206.
Page 30 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
31
Tsai J, Sultana R, Lee Y, Pertea G, Karamycheva S, Antonescu V, Cho J, Parvizi B, Cheung F,
Quackenbush J (2001) RESOURCERER: a database for annotating and linking microarray
resources within and across species. Genome Biology 2:software0002.
Vance JE, Campenot RB, Vance DE (2000) The synthesis and transport of lipids for axonal
growth and nerve regeneration. Biochimica et Biophysica Acta (BBA) - Molecular and Cell
Biology of Lipids 1486:84.
Vazquez-Chona FR KA, Chan CK, Moore AN, Dash PK, Hernandez MR, Lu L, Chesler EJ,
Manly KF, Williams RW, Geisert EE Jr (2005) Genetic networks controlling retinal injury. Mol
Vis 11:958-970.
Vergara MN, Arsenijevic Y, Del Rio-Tsonis K (2005) CNS regeneration: A morphogen's tale.
Journal of Neurobiology 64:491-507.
Vinarsky V, Atkinson DL, Stevenson TJ, Keating MT, Odelberg SJ (2005) Normal newt limb
regeneration requires matrix metalloproteinase function. Developmental Biology 279:86.
Whitehead GG, Makino S, Lien CL, Keating MT (2005) fgf20 Is Essential for Initiating
Zebrafish Fin Regeneration. Science 310:1957-1960.
Yang EV, Gardiner DM, Carlson MR, Nugas CA, Bryant SV (1999) Expression of Mmp-9 and
related matrix metalloproteinase genes during axolotl limb regeneration. Developmental
Dynamics 216:2-9.
Zhang F, Clarke J, Ferretti P (2000) FGF-2 Up-regulation and proliferation of neural progenitors
in the regenerating amphibian spinal cord in vivo. Dev Biol 225:381.
Page 31 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
32
Zhang F, Clarke J, Santos-Ruiz L, Ferretti P (2002) Differential regulation of fibroblast growth
factor receptors in the regenerating amphibian spinal cord in vivo. Neuroscience 114:837.
Zhang F, Ferretti P, Clarke J (2003) Recruitment of postmitotic neurons into the regenerating
spinal cord of urodeles. Developmental Dynamics 226:341-348.
Page 32 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
33
Table 1. Distribution of gene expression patterns for 360 significantly regulated
salamander genes during spinal cord regeneration
Pattern N Description
UCCC 64 up D1, above baseline D1-7
NNUC 44 baseline D1-3, up D5, above baseline D5-7
NDCC 42 baseline D1, down D3, below baseline D3-7
NNND 35 baseline D1-5, down D7
NNNU 28 baseline D1-5, up D7
NNDC 27 baseline D1-D3, down D5, below baseline D5-7
NUCC 24 baseline D1, up D3, above baseline D3-7
UNNN 20 up D1, baseline D3-D7
UUCC 12 up D1 and D3, above baseline D1-7
UDCC 11 up D1, down D3, above baseline D1-7
UCDC 8 up D1, down D3, above baseline D1-7
DCCC 6 down D1, below baseline to D1-7
UCNN 5 up D1, above baseline D1-3, baseline D5-7
UCCN 4 up D1, above baseline D1-5, baseline D7
DCNN 4 down D1, below baseline D1-3, baseline D5-D7
UNND 3 up D1, baseline D3-D5, below baseline D7
UNNU 3 up D1 and D7, baseline D3-D5
NUCN 3 baseline D1, up D3, above baseline D5, baseline D7
NUNN 2 baseline D1, up D3, baseline D5-7
NNUU 2 baseline D1-D3, up D5 and D7
UNUC 2 up D1 and D5, baseline D3, above baseline D5-7
DNNN 1 down D1, baseline D3-D7
NDCD 1 baseline D1, down D3 and D5, below baseline D3-7
NNDN 1 baseline D1-D3, down D5, baseline D7
NNDD 1 baseline D1-D3, down D5 and D7
Page 33 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
34
NDNN 1 baseline D1, down D3, baseline D5-7
NDCN 1 baseline D1, down D3, below baseline D5, baseline D7
UUCD 1 up D1 and D3, down D7, above baseline D1-7
UDNN 1 up D1, down D3, above baseline D1-3, baseline D5-7
UDDC 1 up D1, down D3 and D5, above baseline D7
NUCU 1 baseline D1, up D3 and D7, above baseline D3-7
DCDC 1 down D1 and D5, below baseline D1-7
The 32 gene expression patterns are described in materials and methods. N = total number of
genes in each category.
Page 34 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
35
Table 2. Primer sequences for Q-RT-PCR and in situ hybridization analysis
Sal ID Forward Primer Reverse Primer
MC03237 5’-GGCAAACTGCCTCTCCTT-3’ 5’-CCTGTGGTTTTCCCATGA-3’
MC00341 5’-TTTCTGGACAGCCACTGC-3’ 5’-TTTCTGGACAGCCACTGC-3’
MC01765 5’-CCTGATGGGGATCATCG-3’ 5’-GTGCAGCCGGTACTTGTC-3’
MC03278 5’-TGCATCAAAGCCAAGTCC-3’ 5’-CCTCGGTTCACCTTGAAA-3’
MC01583 5’-ATCCCGGAGAACAAGAGC-3’ 5’-CATCCTTGAGCCAGAGCA-3’
MC02459 5’-CAACGAGTGCATGAACGA-3’ 5’-GCCAGACAGGTGGCCTA-3’
MC01067 5’-TGAGACCAATGCCTTTGC-3’ 5’- GAGCCCCAGAAGCAGAGT-3’
MC01275 5’-GAGGCCAGAAAACCCAGA-3’ 5’-CCGGTTTGGAAATTTCATC-3’
MC01620 5’-TGGCCTGACCAGTAACGA-3’ 5’-AAGTCCCATTCAGCACCA-3’
MC02501 5’-TCCATCCATGTCCTCTGC-3’ 5’-CTGTTTGCGATTGCATGA-3’
MC01187 5’-CCAGGCGGCAGGTCAAGTCAAC-3’ 5’-GTCGGCAAGGTCATCCCAGAGC-3’
in situ hybridization primers
MC02287 5’-GCGCACGATGTCTTTCTGTA-3’ 5’-GCGGTGGTACTCCAACTCAT-3’
MC00365 5’- CGGCTTAGCCAGAAAATGAG-3’ 5’- GTGGTTTCAGCAAAGCCAAT-3’
MC00145 5’- GGCAAGATCATGGCTGAAAT-3’ 5’- CAAGCCGCAGTATCATGTTG -3’
MC02339 5’- TTTGGGAGCGAAGAAGAAAA-3’ 5’- GAACGCTTGGTCTGGTAAGC-3’
MC01583 5’-AAATTCCAATGCAAGTTGGG-3’ 5’-ACGCCGTTCAGCTTGTAGAT-3’
MC01706 5’-ACAACTCAGCCTGGACCATC-3’ 5’- GTCTTTCACCCATTGCAGGT-3’
MC00018 5’-ACATGGAGGACACCAAAAGC-3’ 5’-AGCTCTGGAGTTCAGCTGGT-3’
MC01275 5’-TATTAAGGGCCACGTTCCAG-3’ 5’-TATGGAAGCCCTCCACAGAC-3’
MC01182 5’-GTGCTGCAGGATGTCAAGAA-3’ 5’-GCTGGTGTCTTCCTCTCTGG-3’
MC01277 5’- AATCCAGCCACATCCTTCAC-3’ 5’- GGCGCACCACATACACATAC-3’
Page 35 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
36
Forward ISH primers were appended with an SP6 RNA polymerase promoter and reverse
primers with a T3 promoter at the 5’ end (MC02287 and MC01277 RNA polymerase promoters
are appended to the opposite primers). RNA polymerase promoter sequences are SP6: 5’-
ATTTAGGTGACACTATAGAAGAG-3’ and T3: 5’-AATTAACCCTCACTAAAGGGAGA-3’.
Page 36 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
37
Table 3. Gene symbols and functions for 360 changed salamander genes during the first
week of spinal cord regeneration
Symbol Function Symbol Function
UCCC = 64 Unk; n = 30 PAICS purine metabolism
MYO1B actin binding CTPS pyrimidine biosynthesis
TAGLN actin binding SSB mRNA processing
ADFP lipid metabolism EXOSC2 rRNA processing
TNFAIP8 anti-apoptotic RANBP1 signal transduction
ETHE1 anti-apoptotic TPRT ubiquinone biosynthesis
LGALS1 apoptosis C6ORF115 unknown
GADD45G apoptosis CCDC58 unknown
GADD45B apoptosis NNDC = 27 Unk; n = 8
EFHD2 calcium binding CDKN1C negative cell proliferation
AGC1 unknown MLF1 cell differentiation
MYC cell cycle KRT6L cytoskeleton
CTSL cysteine protease ABLIM1 cytoskeleton
CTSK cysteine protease GLUD1 glutamate catabolism
TUBB6 cytoskeleton ATP1B3 ion transport
FBP1 glycolysis FXYD3 ion transport
B3GNT5 glycosylation PRKAG2 lipid metabolism
TGFB1 growth factor COX4I2 metabolism
TYROBP immune response BHMT metabolism
SLC11A1 immune response PCBD metabolism
Page 37 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
38
MPEG1 immune response CKMT1A metabolism
CXCR4 immune response SERPINI1 neurogenesis
LGALS3BP immune response ABAT neurotransmission
CYBB immune response TM4SF2 protein biosynthesis
ANKRD1 injury response GNB5 signal transduction
FTH1 iron homeostasis POLR2L transcription
ATP6V0D1 proton transport C6ORF110 unknown
C1ORF33 ribosomal FAM79A unknown
MAP2K3 signal transduction NUCC = 24 Unk; n = 6
MMP9 tissue remodeling TAGLN actin binding
MMP1 tissue remodeling ASAH1 apoptosis
LYN tyrosine kinase UHRF1 cell cycle; S
TMEM49 unknown CDK4 cell cycle; G1/S
FLJ2262 unknown KIAA0101 cell cycle; S
WNT5A WNT signaling PPGB cellular transport
NNUC = 44 Unk; n = 14 LGMN cysteine protease
GLRX antioxidant CSTB protease inhibitor
LGALS3 carbohydrate binding RRM2 DNA metabolism
KIF11 cell cycle; M ANXA1 inflammation
CCNB3 cell cycle; G2/M THBS2 ECM component
PLK1 cell cycle; G1/S: G2/M GLB1 metabolism
RFC2 cell cycle; S ATP6V0E proton transport
PCNA cell cycle; S RPL31 ribosomal
MCM7 cell cycle; S APOE lipid metabolism
STK6 cell cycle; M METTL2 ubiquinone biosynthesis
Page 38 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
39
CCNA2 cell cycle; G1/S: G2/M CCDC43 unknown
CDC2 cell cycle; G1/S: G2/M ANGPTL2 growth factor
CHEK1 cell cycle; DNA damage UNNN = 20 Unk; n = 7
MAD2L1 cell cycle; M BYSL cell adhesion
CDC20 cell cycle; M GLN3 cell cycle; G1/S
AURKB cell cycle; M DUSP1 heat shock
CDCA8 cell cycle; M IRF1 immune response
KRT18 cytoskeleton HSPA5 injury response
CALD1 cytoskeleton SLC30A1 ion transport
FEN1 DNA metabolism AGXT2L1 metabolism
SLBP mRNA processing DKC1 ribosomal
CKAP4 inflammation RPS6KA1 signal transduction
KPNA2 intracellular transport PUS1 tRNA processing
VRK1 kinase FLJ36031 unknown
COL12A1 ECM component TGM1 injury response
UCP2 neuroprotection IFIH1 immune response
CTSK protease UUCC = 12 Unk; n = 8
OLFML2B signal transduction F13A1 wound healing
CNIH4 unknown MARCO immune response
CCDC82 unknown FABP2 lipid metabolism
RPL38 ribosomal GPNMB negative cell proliferation
NDCC = 42 Unk; n = 22 UDCC = 10 Unk; n = 6
KRT7 cytoskeleton TFPI2 coagulation
KIF21A cytoskeleton MMP1 tissue remodeling
SLC1A2 glutamate transport GIF vitamin B transport
Page 39 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
40
SLC1A3 glutamate transport C8ORF4 WNT signaling
HSPA8 heat shock UCDC = 8 Unk; n = 3
TTR hormone transport HMOX1 heat shock
SLC12A2 ion transport MMP13 tissue remodeling
KCTD3 ion transport MMP1 tissue remodeling
FDPS lipid metabolism TIMP1 tissue remodeling
FAAH lipid metabolism JUNB transcription
COL8A1 ECM component DCCC = 6 Unk; n = 5
FBN2 ECM component CYP2A13 metabolism
GSTM4 metabolism UCCN = 4
GRM3 neurotransmission LTB4DH antioxidant
PADI3 protein metabolism TXNDC2 antioxidant
APCDD1 signal transduction TXN antioxidant
TJP1 signal transduction BTBD3 protein binding
APC signal transduction DCNN = 4 Unk; n = 2
MAPRE3 structural protein RGMA axon guidance
CRYAB structural protein FHL1 protein binding
NNND = 35 Unk; n = 17 UCNN = 5
PDCD4 apoptosis CES1 neuroprotection
APP apoptosis USP2 protein breakdown
CCNI cell cycle RAP2B signal transduction
SPTAN1 cytoskeleton SAT polyamine homeostasis
SPTBN1 cytoskeleton CD63 signal transduction
PYGM glycogen metabolism UNNU = 3 Unk; n = 2
FABP7 lipid metabolism NOL5A ribosomal
Page 40 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
41
CHPT1 lipid metabolism NUCN = 3
SLC25A4 mitochondrial transport HRAS cell proliferation
GPM6B neurogenesis NSUN2 cell proliferation
AHNAK neurogenesis LTBP1 TGFbeta signalling
GSTM1 neuroprotection UNND = 3 Unk; n = 3
GSTP1 neuroprotection NNUU = 2
GABARAPL2 neurotransmission COL11A1 ECM component
PBP protease inhibitor POSTN skeletal development
CALCA signal transduction NUNN = 2 Unk; n = 2
TRAPPC6B transport UNUC = 2 Unk = 2
C11ORF74 unknown UDDC = 1
NNNU = 28 Unk; n = 7 TCN1 vitamin B transport
KIFC1 cell cycle; M NDCD = 1
CHC1 cell cycle; G2/M AGR2 cell survival
RPA2 cell cycle; DNA damage NUCU = 1
MCM6 cell cycle; S FLJ1472 unknown
CTH cysteine synthesis UUCD = 1
CALD1 cytoskeleton SFRP2 WNT signaling
NASP histone transport UDNN = 1
COL2A1 ECM component IL8RB immune response
LAMA1 ECM component DCDC = 1
DAG1 ECM component KRT5 cytoskeleton
P2RY2 neuronal differentiation NDCC; DNNN; NDCN; NDNN; NNDD; NNDN
NUP107 nuclear transport Unk; n = 1
Page 41 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
42
ANP32E phosphotase inhibitor
Each column contains highlighted categories that describes gene expression patterns on days 1,
3, 5, and 7 days post amputation compared to basal gene expression (day 0). Gene symbols are
found under each category.
Page 42 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
43
Table 4. Gene list for changed genes found on both the Ambystoma and rat U34 Affymetrix
GeneChip®s
Sal ID Symbol Sal ID Symbol Sal ID Symbol
Up sal. and rat; n=11 MC04538 VKORC1L1 MC00251 AK1
MC02474 LGALS3BP MC01709 TYROBP MC05114 C11orf74
MC01211 HMOX1 MC05477 OLFML2B MC00841 FHL1
MC01277 LGALS3 MC07609 MPS1 MC01620 SOX2
MC01006 ATP1B2 MC03420 ANKRD1 Up rat, N/C sal.; n=41
MC01230 IRF1 MC02287 MMP9 MC04731 HNRPD
MC03756 NIP7 MC02182 ATP6V0D1 MC04672 SEC13L1
MC01354 GPNMB gi|40809686 MSX2 MC01058 CBR1
TC01086 TIMP1 MC02833 POSTN gi|5199142 LDHA
MC03344 CNIH4 MC00043 F13A1 MC01272 LDHB
MC04256 C8ORF4 D. rat, N/C sal.; n=18 MC03456 C9ORF10
MC00018 APOE MC01716 UBE2G1 MC01484 PSMD8
Up sal, N/C rat; n=46 MC02022 RNF14 MC04823 ELOF1
MC00197 TGM1 MC02955 TMED2 MC02100 HNRPAB
MC04407 SLC30A1 NP_002133 HOXA9 MC00811 EIF4A1
MC00663 ARG2 MC00031 COL1A2 MC01697 TPT1
MC02499 SGK MC01720 UBE2L3 MC00919 HSPB1
MC03162 LTB4DH MC03866 LOC51255 MC00523 RPL28
MC00337 LTBP1 MC00030 COL1A1 MC04974 HTRA3
MC01521 RAC2 MC01092 COL2A1 MC03977 NOLA2
MC00770 DKC1 MC00212 COL5A2 MC04373 RPL36A
Page 43 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
44
TC00995 RPA2 MC05202 YWHAZ MC03396 CORO1C
MC02792 NOL5A MC00423 IGF1R MC03842 HSPC152
MC03278 FST MC01940 ATIC MC02234 GSTO1
MC01728 UCP2 MC00862 GSTA1 MC00196 TGFBI
MC01639 SSB MC04981 FKBP1A MC01786 DUSP11
gi|1139526 MSX1 MC03940 OTUB1 MC05214 ANXA11
MC01194 GLRX MC01364 OCLN MC03882 FKBP11
MC04782 WDR75 MC04214 DDIT4 MC04970 LRRC42
MC02823 PAICS D. rat, up sal; n=1 MC01722 UBE2N
MC01405 PIM1 MC02177 RAB11A MC00545 RPS3
MC02564 MPHOSPH10 D. sal., N/C rat; n=21 MC02237 EIF4E2
MC01526 RANBP1 MC00837 FAAH MC05309 EMP3
MC04306 NUP107 MC00399 CYP2A13 MC01868 TRADD
MC00145 KRT18 MC01099 COL8A1 MC01933 ANXA7
MC00535 RPL38 MC02628 FXYD3 MC01392 PGD
MC00222 GLB1 MC00345 ADH1A MC01095 COL4A1
MC01541 RGS10 MC02307 SERPINI1 MC00332 GPX1
MC02123 KIF11 TC00019 APC MC00274 HEXB
MC01119 CTH MC03555 MAST3 MC01711 RPS27A
MC02344 GIF MC01037 BHMT MC03904 CHRAC1
MC02468 LAMA1 MC00741 CSRP3 MC03898 SMN1,SMN2
MC01297 MAD2L1 MC01503 PTPRR MC02876 HSPA8
MC02065 DAG1 MC04857 COX4I2 gi|49473435 NPY
MC04570 CRELD2 MC04494 MLF1 MC02061 CSPG3
MC01751 VRK1 MC01631 SPTBN1 MC00282 TP53
Page 44 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
45
MC02308 PLK1 TC00539 FABP7 Up rat, D. sal.; n=1
MC01068 CDC2 MC05220 GSTM5 TC00219 APP
MC00594 RRM2 MC04424 MID1IP1
Genes were determined to be up-regulated, down-regulated, or not changed (N/C) according to
the criteria set by each rat microarray study.
Page 45 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
46
Table 5. Comparison of microarray versus real-time RT-PCR estimates of fold change for
ten genes that were quantified on day 0 and day 3
Sal ID Gene Name Microarray Real Time
MC01620 SOX2 -1.7 -1.86**
MC02459 HES1 -2.15 -1.35
MC02501 SOX3 -1.92 1.14
MC03278 FST 3.09 3.45**
MC01765 CXCR4 3.17* 4.19**
MC01067 CD63 2.17* 4.77**
MC00341 TGFB1 3.69* 5.43**
MC01275 LGALS1 2.55* 10.5**
MC01583 SFRP2 17.95* 23.8**
MC03237 AGC1 56.84* 42.9**
*In the list of 360 genes that met statistical and fold level criteria.
** Significant fold change difference between day 0 and day 3 according to real-time RT-PCR
(student’s unpaired t-test, P < 0.05).
Page 46 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
47
Table 6. In situ hybridization results using tails collected three days after injury
Spatial expression pattern
Sal ID Gene Name Neurons Ependymoglia White matter Outside SC
MC00365 ANXA1 ↑ + ++ + +
MC02287 MMP9 ↑ + +++ + ++
MC00145 CK18 ↑ ++ +++ 0 0
MC02339 FGFBP1 ↓ ++ +++ 0 0
MC01583 SFRP2 ↑ ++ ++ ++ +++
MC01706 TXN ↑ ++ ++ 0 0
MC00018 APOE ↑ 0 0 +++ +
MC01275 LGALS1 ↑ + + +++ +++
MC01182 FTH1 ↑ + + +++ +++
MC01277 LGALS3 ↑ + 0 ++ +++
Gene expression was qualitatively ranked as 0, +, ++, or +++ according to the intensity of stain
and number of cells expressing the transcript. Arrows indicate if the transcript was up-regulated
or down-regulated after injury.
Figure 1. Volcano plots showing the number of genes selected as differentially expressed with
the FDR set to 0.01 (a) and the FWER set to 0.01 (b). Genes selected by each of these respective
criteria are gray and non-selected genes are black.
Page 47 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
48
Figure 2. Two-dimensional plot of a principal component analysis (PCA) showing the
relatedness of each GeneChip®. JMP statistical software was used to perform PCA on 25
GeneChip®s. A PCC matrix was made for 25 Genechip®s using intensity values for 376 changed
genes. Principal component 1 (PC1; 80.35% of the variation; eigenvalue 20.09) is displayed on
the x-axis and Principal component 2 (PC2; 13.54% of the variation; eigenvalue 3.34) is
displayed on the y-axis. The cumulative variation accounted for by PC1 and PC2 is 93.89%.
Twenty-five principal components account for 100% of the variation in the dataset. Five
biological replicate chips used for each of the five time points are enclosed by an oval to
illustrate their close proximity. ♦: Day 0, n = 5; ■: Day 1, n = 5; ▲: Day 3, n = 5; ●: Day 5, n =
5; ▬: Day 7, n = 5.
Figure 3. In situ hybridization of axolotl spinal cord using RNA probes that correspond to
significantly regulated genes from the microarray analysis. Rows show three representative
cross-sections of tissue samples collected three days after tail amputation. Photos in column 1
show a more caudal section than photos in column 2. Anti-sense probes are represented in
columns 1 and 2, and sense control probes in column 3. (A) Mmp9 expression can be seen in the
ependymoglia cell population (arrowhead) lining the central canal. Other cells found outside the
spinal cord are also positive for Mmp9 expression (arrow and data not shown). (B) Mmp9 is
expressed in few cells caudal to the injured spinal cord. Strong expression can be seen in a single
cell (arrowhead). (D) ck18 mRNA is strongly expressed exclusively in ependymoglial cells near
the end of the regenerating spinal cord. (E) ck18 is expressed in both ependymoglial cells and
subpopulations of neurons rostral to the injury. (G) Cross-section just rostral to the injury plane
Page 48 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
49
shows thioredoxin mRNA expression in both ependymoglia and some neurons. (H) Rostrally,
thioredoxin is expressed more exclusively in cells resembling neurons rather than ependymoglia.
(J) Many cells in the white matter are expressing apoE with little or no expression in neurons or
ependymoglia. Cells are often closely associated with the ependymoglia (arrowhead). Several
positive cells have processes (arrows; Fig J and K) resembling microglia or other phagocytic
cells. (K) Rostrally, few cells are expressing ApoE (arrow). Bar = 100 µm.
Supplemental Table 1. Genes and associated functions for the list of 360 significantly regulated
genes that met all statistical and fold level criteria. Average fold level changes between Day 0
and each of the post-amputation time points (Day1, 3, 5, or 7). Empty rows represent probe sets
that have another probe set designed for the same gene. One set of intensity values is represented
for probe set pair.
Page 49 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
Volcano plots showing the number of genes selected as differentially expressed with the FDR set to 0.01 (a) and the FWER set to 0.01 (b). Genes selected by each of these
respective criteria are gray and non-selected genes are black.
Page 50 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
Two-dimensional plot of a principal component analysis (PCA) showing the relatedness of each GeneChip®. JMP statistical software was used to perform PCA on 25 GeneChip®s. A PCC matrix was made for 25 Genechip®s using intensity values for 376 changed genes. Principal component 1 (PC1; 80.35% of the variation; eigenvalue 20.09) is displayed on the x-axis and Principal component 2 (PC2; 13.54% of the variation; eigenvalue 3.34) is
displayed on the y-axis. The cumulative variation accounted for by PC1 and PC2 is 93.89%. Twenty-five principal components account for 100% of the variation in the
dataset. Five biological replicate chips used for each of the five time points are enclosed by an oval to illustrate their close proximity. ♦: Day 0, n = 5; ■: Day 1, n = 5; ▲: Day 3, n =
5; ●: Day 5, n = 5; ▬: Day 7, n = 5.
Page 51 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
In situ hybridization of axolotl spinal cord using RNA probes that correspond to significantly regulated genes from the microarray analysis. Rows show three
representative cross-sections of tissue samples collected three days after tail amputation. Photos in column 1 show a more caudal section than photos in column 2. Anti-sense
probes are represented in columns 1 and 2, and sense control probes in column 3. (A) Mmp9 expression can be seen in the ependymoglia cell population (arrowhead) lining the
central canal. Other cells found outside the spinal cord are also positive for Mmp9 expression (arrow and data not shown). (B) Mmp9 is expressed in few cells caudal to the injured spinal cord. Strong expression can be seen in a single cell (arrowhead). (D) ck18
mRNA is strongly expressed exclusively in ependymoglial cells near the end of the regenerating spinal cord. (E) ck18 is expressed in both ependymoglial cells and
subpopulations of neurons rostral to the injury. (G) Cross-section just rostral to the injury plane shows thioredoxin mRNA expression in both ependymoglia and some
neurons. (H) Rostrally, thioredoxin is expressed more exclusively in cells resembling neurons rather than ependymoglia. (J) Many cells in the white matter are expressing
Page 52 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
apoE with little or no expression in neurons or ependymoglia. Cells are often closely associated with the ependymoglia (arrowhead). Several positive cells have processes
(arrows; Fig J and K) resembling microglia or other phagocytic cells. (K) Rostrally, few cells are expressing ApoE (arrow). Bar = 100 µm.
Page 53 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
Pattern/ID Gene Name Function Day1 Day3 Day5 Day7
UCCC = 64
MC02410 myosin IB actin binding 2.47 4.37 4.19 4.03MC01655 Transgelin actin binding 2.15 3.67 5.39 10.18MC00632 adipose differentiation-related protein lipid metabolism 19.99 11.86 7.4 4.04MC03405 TNF, alpha-induced protein 8 anti-apoptosis 2.76 2.64 2.23 2.18MC03384 ethylmalonic encephalopathy 1 anti-apoptotic 3.67 3.15 3.11 2.77MC01275 galectin 1 apoptosis 2.01 2.55 2.58 2.57MC02911 growth arrest and DNA-damage-inducible, gamma apoptosis 7.03 5.98 4.06 2.34MC03641 growth arrest and DNA-damage-inducible, beta apoptosis 11.59 6.42 4.24 3.73MC04572 EF hand domain containing 2 calcium binding protein 2.81 2.65 2.22 1.98MC03237 c-type lectin unknown 34.29 56.84 47.01 31.02MC01330 v-myc viral oncogene cell cycle 5.33 4.41 2.34 2.02MC05219
TC03644
TC00185 cathepsin K cysteine protease 2.58 3.21 2.57 2.16MC04847 tubulin beta cytoskeleton 4.7 4.67 3.59 3.35MC00259 fructose-1,6-bisphosphatase 1 glycolysis 4.95 4.49 3.02 2.82MC04771 UDP-GlcNAc glycosylation 5.19 2.63 3.06 3.39MC00341 transforming growth factor, beta 1 growth factor 2.65 3.69 3.37 2.94MC01709 TYRO protein tyrosine kinase binding protein immune response 2.93 3.79 3.32 3.27MC00330 solute carrier family 11, member 1 immune response 3.93 7.46 6.57 5.14MC07609 macrophage expressed gene 1 immune response 3.47 6.91 5.92 4.61MC01765 chemokine (C-X-C motif) receptor 4 immune response 3.59 3.17 3.02 3.03MC02474 galectin 3 binding protein immune response 2.94 5.12 4.4 3.81MC00218
TC00186
MC03420 ankyrin repeat domain 1 (cardiac muscle) injury response 6.78 5.13 5.42 5.65MC01182 ferritin, heavy polypeptide 1 iron homeostasis 3.44 4.88 4.56 4.8MC02182 ATPase, H+ transporting, V0 subunit d isoform 1 proton transport 3.07 5.74 5.73 5.97MC03786 chromosome 1 open reading frame 33 ribosomal 2.82 3.07 2.29 2.56MC05180 mitogen-activated protein kinase kinase 3 signal transduction 2.08 3.17 3.4 2.62MC02287 matrix metalloproteinase 9 tissue remodeling 7.44 5.58 5.29 5.83MC01311 matrix metalloproteinase 1 tissue remodeling 13.1 6.45 6.24 6.97TC00829 v-yes-1 viral related oncogene homolog tyrosine kinase 3.43 3.5 2.71 2.6MC04702 vacuole membrane protein 1 unknown 8.46 7.71 8.2 7.23MC10230 similar to CG12279-PA unknown 2.21 3.3 5.13 5.74MC04619 hypothetical protein FLJ22662 unknown 30.97 41.64 30.2 23.85gi|62426 wingless-type MMTV, member 5A WNT signaling 2.37 3.86 4.55 4.71MC05999 unknown unknown 12.23 8.36 4.4 2.46MC10145 unknown unknown 33.58 20.9 22.76 21.76MC09114 unknown unknown 9.48 6.61 5.18 3.33MC01847 unknown unknown 9.12 4.62 3.93 4.56MC10221 unknown unknown 5.26 7.55 6.17 4.14MC08530 unknown unknown 2.5 4.57 6.18 6.4MC10258 unknown unknown 7.18 4.48 4.94 4.68MC10136 unknown unknown 4.34 7.18 7.99 7.51MC07184 unknown unknown 6.69 5.5 3.07 2.46MC09299 unknown unknown 7.46 11.28 11.7 6.91MC00640 unknown unknown 5.89 4.66 3.91 2.98MC10333 unknown unknown 9.5 6.16 5.29 4.93
8.65 7.41
cathepsin L cysteine protease
cytochrome b-245, beta polypeptide immune response 11.11 11.7
3.84 5.95 3.26 2.6
Page 54 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
MC08352 unknown unknown 14.29 5.98 5.65 5.77MC06325 unknown unknown 2.1 3.42 4.09 3.39MC09544 unknown unknown 2.5 3.35 3.27 3.2MC09298 unknown unknown 7.09 7.71 9.38 6MC09381 unknown unknown 3.54 2.81 2.96 3.05MC07259 unknown unknown 2.78 4.86 4.37 3.64MC09443 unknown unknown 4.74 4.92 5.26 5.64MC09474 unknown unknown 4.25 4.3 4.41 5.29MC10140 unknown unknown 3.9 3.52 2.95 2.3MC09930 unknown unknown 3.61 3.07 2.77 2.27MC07617 unknown unknown 3.46 4.54 4.08 3.91MC10140 unknown unknown 3.9 3.52 2.95 2.3MC07874 unknown unknown 6.22 6.41 5.98 6.11MC10062 unknown unknown 4.4 4.74 5.09 3.89MC08348 unknown unknown 3.21 2.1 1.92 2MC05353 unknown unknown 2.04 3.43 2.71 2.26MC10093 unknown unknown 6.42 3.25 2.59 2.42
NNUC = 44
MC01194 glutaredoxin antioxidant 1.49 1.97 2.05 2.06MC01277 galectin 3 carbohydrate binding 1.42 1.97 2.77 3.32MC02123 kinesin family member 11 cell cycle; M -1.22 1.08 2.17 2.31MC04914 -1.2 -1 2.23 2.91MC04954 -1.17 -1.01 2.29 2.76MC02308 polo-like kinase 1 (Drosophila) cell cycle; G1/S:G2/M -1.2 1.38 2.27 2.57MC05457 replication factor C 2 cell cycle; S 1.15 1.74 2.33 2.8MC05488 proliferating cell nuclear antigen cell cycle; S -1.07 1.95 2.18 2.18MC02615 minichromosome maintenance deficient 7 cell cycle; S 1.05 1.68 2.02 2.07MC01795 serine/threonine kinase 6 cell cycle; M -1.08 1.58 2.96 3.51MC00687 cyclin A2 cell cycle; G1/S: G2/M -1.08 1.29 2.42 2.65MC01068 cell division cycle 2 cell cycle; G1/S: G2/M -1.21 1.44 2.58 2.61MC00703 CHK1 checkpoint homolog cell cycle; DNA damage -1.26 1.71 2.14 2.24MC01297 MAD2 mitotic arrest deficient-like 1 cell cycle; M -1.21 1.81 2.18 2.46MC00694 cell division cycle 20 homolog cell cycle; M -1.18 1.4 2.53 3.25MC02001
MC02003
MC04043 cell division cycle associated 8 cell cycle; M -1.18 1.07 2.01 2.41MC00145 cytokeratin 18 cytoskeleton component 1.63 1.92 2.08 2.19MC04928 caldesmon 1 cytoskeleton component 1.02 1.52 2.24 3.14MC01972 flap structure-specific endonuclease 1 DNA metabolism -1 1.9 2.47 2.92MC02851
MC02852
MC02963 cytoskeleton-associated protein 4 inflammatory response -1.12 1.97 3.5 5.93MC01242 karyopherin alpha 2 intracellular transport 1.68 1.79 2.27 2.58MC01751 vaccinia related kinase 1 kinase 1.02 1.3 2.21 2.56MC05045 collagen, type XII, alpha 1 ECM component 1.24 1.79 2.66 4.47MC01728 uncoupling protein 2 neuroprotection 1.44 1.87 2.05 2.01MC00215 cathepsin K protease -1.52 1.75 3.6 6.26MC05477 olfactomedin-like 2B signal transduction -1.27 1.06 2.36 3.36MC03344 HSPC163 protein signal transduction 1.59 1.99 2.09 2.36MC04608 hypothetical protein FLJ23518 unknown 1.19 1.81 2.88 3.22MC00535 ribosomal protein L38 ribosomal 1.54 1.97 2.08 2.2
2.12 2.47stem-loop (histone) binding protein histone mRNA processing -1.03 1.83
-1.1 1.19 2.16 2.44
cyclin B3 cell cycle; G2/M
aurora kinase B cell cycle; M
Page 55 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
MC04760 unknown unknown -1.48 1.09 2.65 3.51MC10004 unknown unknown 1.07 1.37 2.14 2.96MC09962 unknown unknown 1.11 1.34 2.12 2.72MC06887 unknown unknown -1.15 1.46 2.63 3.19MC00214 unknown unknown -1.03 1.63 3.94 6.6MC08928 unknown unknown 1.3 1.8 3.53 5.1MC07152 unknown unknown 1.01 1.48 2.15 2.17MC06483 unknown unknown -1.07 1.53 2.19 3.01MC09462 unknown unknown 1.19 1.9 2.05 2.28MC05554 unknown unknown 1.53 1.82 2.82 3.33MC08286 unknown unknown 1.77 1.83 2.28 2.4MC09310 unknown unknown -1.15 1.61 2.13 2.35MC07493 unknown unknown -1.07 1.33 2.15 2.57MC06189 unknown unknown 1.3 1.45 2.17 3.04
NDCC = 42
TC01675 keratin 7 cytoskeletal element -1.83 -2.19 -4.89 -9.13MC03935 kinesin family member 21A cytoskeleton component -1.56 -2.27 -2.46 -3.16
gi|2655016 solute carrier family 1, member 2 glutamate transport -1.61 -5.22 -5.98 -7.32 gi|2655014 solute carrier family 1, member 3 glutamate transport -1.8 -2.71 -2.71 -3.02MC02877 heat shock 70 kDa protein 8 heat shock 1.6 -2.16 -2.05 -2.52MC00207 transthyretin hormone transport -1.12 -2.11 -2.43 -2.77MC00599
TC00434
MC03761 potassium channel KCTD3 ion transport -1.42 -2.41 -3.16 -4.47MC01168 farnesyl diphosphate synthase lipid metabolism -1.27 -2.98 -3.01 -2.4MC00837 fatty acid amide hydrolase lipid metabolism -1.44 -2.14 -2.63 -4.47MC01099 collagen, type VIII, alpha 1 ECM component -1.41 -4.34 -4.66 -4.34MC00050 fibrillin 2 ECM component -1.64 -2.45 -2.86 -2.44MC00415 glutathione S-transferase M4 metabolism -1.39 -2.01 -2.27 -2.93MC00413 glutamate receptor, metabotropic 3 neurotransmission -1.61 -3.28 -2.91 -3.34MC03121 peptidyl arginine deiminase, type III protein metabolism -1.27 -2.04 -3.18 -4.82MC05295 adenomatosis polyposis coli down-reg 1 signal transduction -1.67 -2.67 -2.45 -2.33MC01681 tight junction protein 1 signal transduction -1.34 -2.2 -2.26 -2.36TC00019 adenomatosis polyposis coli signal transduction -1.4 -2.12 -2.24 -2.97MC03199 microtubule-associated protein MAPRE3 structural protein -1.61 -2.01 -2.89 -4.48MC01113 crystallin, alpha B structural protein -1.26 -2.9 -3.37 -4.5MC07514 unknown unknown -1.18 -2.01 -2.02 -2.22MC10135 unknown unknown -1.95 -2.94 -2.55 -2.91MC09228 unknown unknown 1.68 -2.04 -2.14 -2.58MC08604 unknown unknown -1.39 -2.67 -2.71 -2.82MC10167 unknown unknown -1.38 -2.14 -2.43 -2.99MC07622 unknown unknown -1.45 -2.2 -2.85 -3.83MC07141 unknown unknown -1.24 -2.12 -2.79 -4MC09886 unknown unknown -1.51 -3.05 -3.95 -5.79MC07029 unknown unknown -1.74 -3.61 -3.41 -4.55MC07816 unknown unknown -1.45 -2.05 -2.52 -3.42MC05592 unknown unknown -1.88 -2.58 -3.54 -4.59MC10052 unknown unknown -1.41 -2.01 -2.25 -2.67MC05593 unknown unknown -1.64 -2.17 -2.67 -3.61MC08388 unknown unknown -1.22 -2.12 -1.94 -2.28MC07961 unknown unknown -1.3 -2.49 -2.44 -2.66
-2.91 -3.7solute carrier family 12, member 2 ion transport -1.63 -2.07
Page 56 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
MC07497 unknown unknown -1.31 -2.86 -3.37 -3.47MC07444 unknown unknown -1.32 -2.09 -1.91 -2.07MC09586 unknown unknown -1.36 -2.09 -2.15 -2.77MC10106 unknown unknown -1.42 -2.04 -2.68 -3.42MC10355 unknown unknown -1.29 -2.01 -2.08 -3.2MC09589 unknown unknown -1.57 -2.61 -3.09 -3.65MC07565 unknown unknown -1.89 -2.13 -1.92 -2.35
NNND = 35
MC03428 programmed cell death 4 apoptosis 1.07 -1.8 -1.85 -2.13TC00219 amyloid beta (A4) precursor protein apoptosis -1.38 -1.41 -1.78 -2.24MC02974 cyclin I cell cycle -1.24 -1.6 -1.93 -2.25MC01630 spectrin, alpha, non-erythrocytic 1 cytoskeleton component -1.34 -1.72 -1.91 -2.38MC01631
MC01632
MC02490 phosphorylase, glycogen glycogen metabolism -1.39 -1.53 -1.96 -2.43TC00539 fatty acid binding protein 7, brain lipid metabolism -1.24 -1.37 -1.68 -2.34MC04289 choline phosphotransferase 1 lipid metabolism -1.17 -1.97 -1.81 -2.16MC00987 solute carrier family 25, member 4 mitochondrial transport -1.39 -1.36 -1.67 -2.1MC02382 glycoprotein M6B neurogenesis -1.34 -1.75 -1.85 -2.13MC10102 AHNAK nucleoprotein (desmoyokin) neurogenesis -1.19 -1.66 -1.85 -2.03MC00326
MC00327
MC00419 glutathione S-transferase pi neuroprotection -1.29 -1.74 -1.92 -2.52MC03106 GABA receptor-associated GABARAPL2 neurotransmission -1.36 -1.67 -1.91 -2.54MC01369 prostatic binding protein protease inhibitor -1.31 -1.38 -1.57 -2.03MC01046 calcitonin-related polypeptide, alpha signal transduction 1.15 -1.15 -1.63 -2.84MC05420 trafficking protein particle complex 6B transport -1.41 -1.66 -1.91 -2.18MC05114 hypothetical protein BC009561 unknown -1.26 -1.67 -1.78 -2.06MC09850 unknown unknown -1.3 -1.71 -1.92 -2.86MC09217 unknown unknown -1.21 -1.38 -1.67 -2.39MC07208 unknown unknown -1.21 -1.59 -1.65 -2.11MC07303 unknown unknown 1.06 -1.24 -1.67 -2.11MC06742 unknown unknown 1.04 -1.55 -1.69 -2.09MC10155 unknown unknown 1.11 -1.43 -1.89 -2.09MC09850 unknown unknown -1.3 -1.71 -1.92 -2.86MC07685 unknown unknown -1.26 -1.69 -1.85 -2.34MC07838 unknown unknown -1.56 -1.91 -1.89 -2.41MC09395 unknown unknown -1.22 -1.78 -1.91 -2.28MC09898 unknown unknown 1.1 -1.45 -1.71 -2.16MC09918 unknown unknown 1.21 -1.45 -1.63 -2MC10156 unknown unknown -1.17 -1.57 -1.7 -2.15MC09554 unknown unknown -1.56 -1.77 -1.95 -2.17MC09092 unknown unknown -1.24 -1.31 -1.58 -2.19MC07421 unknown unknown -1.32 -1.69 -1.78 -2.22MC10165 unknown unknown -1.28 -1.45 -1.66 -2.07
NNNU = 28
MC02466 kinesin family member C1 cell cycle; M -1.83 -1.08 1.8 2.03MC00700 chromosome condensation 1 cell cycle; G2/M 1.16 1.67 1.82 2.29TC00995 replication protein A2, 32kDa cell cycle; DNA damage -1.06 1.28 1.74 2MC02613 minichromosome maintenance deficient 6 cell cycle; S -1.05 1.64 1.99 2.14
-1.73 -2.18
spectrin, beta, non-erythrocytic 1 cytoskeleton component
glutathione S-transferase M1 neuroprotection -1.02 -1.58
-1.38 -1.75 -1.92 -2.65
Page 57 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
MC01119 cystathionase cysteine synthesis -1.02 1.95 1.86 2.36MC02037 caldesmon 1 cytoskeleton -1.26 1.51 1.98 2.95MC05347 nuclear autoantigenic sperm protein histone transport 1.31 1.62 1.93 2.03MC01093 collagen, type II, alpha 1 ECM component -1.18 -1.99 1.05 2.56MC02468 laminin, alpha 1 ECM component -1.3 1.09 1.73 2.43MC02066 dystroglycan 1 ECM component -1.31 1.29 1.77 2.44MC01368 purinergic receptor P2Y neuronal differentiation -1.25 1.22 1.75 2.32MC04306 nucleoporin nuclear transport 1.02 1.61 1.75 2.14MC04694 acidic nuclear phosphoprotein ANP32E phosphotase inhibitor 1.09 1.47 1.65 2.16MC02823
MC02824
MC01129 CTP synthase pyrimidine biosynthesis 1 1.56 1.77 2.03MC01639 Sjogren syndrome antigen B RNA processing 1.4 1.96 1.94 2.03MC03377 exosome component 2 rRNA processing 1.28 1.82 1.83 2.09MC01526 RAN binding protein 1 signal transduction 1.36 1.82 1.78 2.09MC03393 trans-prenyltransferase ubiquinone biosynthesis 1.26 1.87 1.77 2.28MC05825 chromosome 6 orf 115 unknown 1.39 1.81 1.94 2.23MC07600 hypothetical LOC131unknown76 unknown 1.31 1.6 1.77 2.08MC05568 unknown unknown -1.29 1.15 1.89 2.67MC09570 unknown unknown -1.14 1.1 1.87 2.58MC07466 unknown unknown 1.58 1.68 1.81 2MC10295 unknown unknown -1.11 1.04 1.64 2.26MC10029 unknown unknown -1.33 1.1 1.53 2.19MC09210 unknown unknown 1.49 1.42 1.85 3.05MC05927 unknown unknown 1.87 1.93 1.95 2.18
NNDC = 27
MC00028 cyclin-dependent kinase inhibitor 1C negative cell proliferation -1.45 -1.72 -2.23 -3.29MC04494 myeloid leukemia factor 1 cell differentiation -1.4 -1.72 -2.09 -2.45MC00232 keratin 6L cytoskeleton component -1.12 -1.6 -2.53 -5.03MC02916 actin binding LIM protein 1 cytoskeleton component -1.37 -1.88 -2.1 -2.66MC02377 glutamate dehydrogenase 1 glutamate catabolism -1.52 -1.64 -2.01 -2.64MC01008 ATPase, Na+/K+ transporting, beta 3 ion transport -1.39 -1.66 -2.01 -2.25MC02628 FXYD ion transport regulator 3 ion transport -1.26 -1.66 -2.04 -3.85MC03790 AMP-activated protein kinase, gamma 2 lipid metabolism -1.56 -1.8 -2.03 -2.54MC04857 cytochrome c oxidase COX4I2 metabolism -1.29 -1.87 -2.14 -2.62MC01037 betaine-homocysteine methyltransferase metabolism -1.55 -1.65 -2.2 -2.8MC00170 6-pyruvoyl-tetrahydropterin synthase metabolism -1.48 -1.87 -2.23 -2.65MC04363 creatine kinase, mitochondrial 1A metabolism -1.62 -1.56 -2.15 -2.74MC02307 serine proteinase inhibitor SERPINI1 neurogenesis -1.58 -1.73 -2.29 -3.21MC00344 4-aminobutyrate aminotransferase neurotransmission -1.6 -1.98 -2.82 -4MC02165 transmembrane 4 superfamily member 2 protein biosynthesis -1.53 -1.87 -2.54 -3.6
gi|14029843 guanine nucleotide binding GNB5 signal transduction -1.21 -1.99 -2.11 -2.26MC04397 polymerase (RNA) II polypeptide L transcription -1.31 -1.9 -2.16 -3.18MC06347 chromosome 6 orf 110 unknown -1.33 -1.7 -2.32 -2.98MC05509 hypothetical protein LOC127262 unknown -1.42 -1.95 -2.29 -2.63MC07786 unknown unknown -1.26 -1.6 -2.85 -4.95MC08476 unknown unknown -1.41 -1.94 -2.03 -2.52MC08958 unknown unknown -1.32 -1.85 -2.19 -2.59MC08418 unknown unknown -1.42 -1.98 -2.26 -3.12TC07363 unknown unknown -1.3 -1.55 -2.11 -2.26MC09779 unknown unknown -1.5 -1.69 -2.11 -2.64
1.86 2.14phosphoribosylaminoimidazole carboxylase purine metabolism 1.07 1.63
Page 58 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
MC10283 unknown unknown -1.44 -1.64 -2.04 -2.52MC09671 unknown unknown -1.3 -1.46 -2.06 -2.82
NUCC = 24
TC01065 transgelin actin binding 1.99 2.66 3.47 5.7MC02029 N-acylsphingosine amidohydrolase 1 apoptosis 1.58 2.69 2.77 2.44MC03249 ubiquitin-like 1 cell cycle; S -1.23 2.71 2.49 2.71MC06890 cyclin-dependent kinase 4 cell cycle; G1/S 1.17 2.28 2.18 2.66MC03476 KIAA0101 gene product cell cycle; S -1.33 2.86 3.24 3.56MC00182 protective protein for beta-galactosidase cellular transport 1.4 3.31 2.6 2.31MC02488
MC02489
MC00038 cystatin B (stefin B) protease inhibitor 1.71 2.49 2.54 2.4MC00594 ribonucleotide reductase M2 polypeptide DNA metabolism -1.42 2.27 2.29 2.62MC00365 annexin A1 inflammatory response 1.67 2.1 2.09 2.08MC01673 thrombospondin 2 ECM component 1.65 2.43 2.26 3.35MC00222 galactosidase, beta 1 metabolism 1.42 2.79 2.91 2.22MC01923 ATPase, H+ transporting, lysosomal 9kDa proton transport 1.7 2.25 2.15 2.27MC00526 ribosomal protein L31 ribosomal 1.44 2.11 2.43 2.88MC00018 apolipoprotein E lipid metabolism 1.32 3.76 4.27 3.7MC04130 methyltransferase like 2 ubiquinone biosynthesis 1.34 2.03 1.83 2MC05155 hypothetical protein FLJ31795 unknown 1.6 2.04 1.84 2.21MC03138 angiopoietin-like 2 vascular growth factor 1.77 5.5 4.97 5.68MC10381 unknown unknown 1.26 2.18 2.34 2.2MC07325 unknown unknown 1.43 3.67 3.08 2.19MC02207 unknown unknown 1.71 2.52 2.18 2.27MC10141 unknown unknown 1.91 3.36 3.43 3.33MC10204 unknown unknown 1.36 2.59 2.22 2.27MC09797 unknown unknown 1.68 1.81 1.62 1.88
UNNN = 20
MC01947 bystin-like cell adhesion 2.45 1.9 1.48 1.44MC03408 nucleostemin cell cycle; G1/S 2.19 1.81 1.6 1.65MC02084 dual specificity phosphatase 1 heat shock 4.09 1.94 1.57 1.71MC01230
MC01231
MC02404 heat shock 70 kDa protein 5 injury response 3.89 1.88 1.8 1.93MC04407 solute carrier family 30, member 1 ion transport 2.06 1.36 1.15 -1.02MC04719 glyoxylate aminotransferase 2-like 1 metabolism 3.95 1.23 1.17 1.26MC00770 dyskeratosis congenita 1, dyskerin ribosomal 2.52 1.81 1.8 1.99MC01559 ribosomal protein S6 kinase signal transduction 2.52 1.71 1.14 1.17MC04660 pseudouridylate synthase 1 tRNA processing 2.82 1.81 1.66 1.8MC05410 hypothetical protein FLJ36031 unknown 2.37 1.4 1.79 1.87MC00197 transglutaminase 1 injury response 2.24 1.04 -1.34 -1.99MC04487 melanoma differentiation associated 5 immune response 2.98 1.05 1.09 1.11MC05886 unknown unknown 2.13 1.25 1.12 1.07MC07124 unknown unknown 2.46 1.48 1.57 1.58MC09976 unknown unknown 2.28 1.62 1.37 1.17MC09774 unknown unknown 4.36 1.1 1.23 1.28MC09902 unknown unknown 2.07 1.65 1.45 1.5MC07587 unknown unknown 3.21 1.15 1.07 1.24MC09869 unknown unknown 2.3 1.26 1.47 1.42
interferon regulatory factor 1 immune response 3.82 1.62
3.08 2.69
1.31 1.22
legumain protease 1.45 3.45
Page 59 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
UUCC = 12
MC00043 coagulation factor XIII, A1 polypeptide wound healing 5.25 11.56 10.03 9.58MC02942 macrophage receptor with collagenous structure immune response 3.49 12.45 11.39 7.7MC00047 fatty acid binding protein 2, intestinal lipid metabolism 6.54 21.27 21.75 15.3MC01354
MC01355
MC10310 unknown unknown 9.86 24.32 33.76 31.76MC07205 unknown unknown 9.51 29.58 36.73 35.52MC10047 unknown unknown 3.01 6.8 6.62 4.3MC09973 unknown unknown 2.48 5.14 4.36 3.36MC09209 unknown unknown 2.08 6.18 5.7 3.43MC09990 unknown unknown 3.8 11.74 15.51 13.67TC02494 unknown unknown 2.38 7.37 5.29 3.65MC03347 unknown unknown 2.4 6.44 6.58 9.98
UDCC = 10
MC02853 tissue factor pathway inhibitor 2 coagulation 4.62 1.89 1.98 2.29TC00840 matrix metalloproteinase 1 tissue remodeling 18.97 5.6 5.4 5.76MC02344 intrinsic factor vitamin B transport 14.23 5.01 2.83 2.43MC04256 chromosome 8 orf 4 WNT signaling 14.21 3.99 2.32 2.49MC07780 unknown unknown 63.75 26.12 23.1 21.7MC09473 unknown unknown 21.19 5.79 3.33 3.21MC10351 unknown unknown 9.08 3.36 2.54 2.18MC09159 unknown unknown 14.23 4.11 2.62 2.12MC09527 unknown unknown 16.37 6.79 8.32 5.85MC07978 unknown unknown 6.88 2.21 1.94 2.17
UCDC = 8
MC01211 heme oxygenase 1 heat shock 18.02 19.45 6.88 6.84TC00846 matrix metalloproteinase 13 tissue remodeling 18.49 32.2 10.16 5.76MC01312 matrix metalloproteinase 1 tissue remodeling 18.53 27.54 11.47 6.97TC01086 tissue inhibitor of metalloproteinase 1 tissue remodeling 29.63 16.72 5 2.71MC01239 jun-B proto-oncogene transcription 24.14 12.15 10.8 11.45MC09943 unknown unknown 18.07 9.67 3.55 2.08MC07614 unknown unknown 12.92 8.72 2.98 2.51MC09763 unknown unknown 49.69 43.99 16.7 17.47
DCCC = 6
MC00399 cytochrome P450 (CYP2A13) metabolism -2 -2.03 -3.7 -4.35MC06562 unknown unknown -2.03 -3.92 -3.6 -3.9MC05696 unknown unknown -3.95 -2.91 -2.23 -2.7MC09604 unknown unknown -2.25 -2.9 -2.78 -2.89MC08782 unknown unknown -2.06 -2.67 -2.74 -2.53MC07692 unknown unknown -2.52 -2.15 -2.33 -3.03
UCCN = 4
MC03162 leukotriene B4 12-hydroxydehydrogenase antioxidant 2.66 2.74 2.03 1.7MC04794 thioredoxin domain containing 2 antioxidant 2.53 2.41 2.16 1.94MC01706 thioredoxin antioxidant 2.33 2.34 2.09 1.64MC04937 BTB (POZ) domain containing 3 protein binding 3.96 2.54 2.09 1.69
26.95 22.3glycoprotein nmb negative cell proliferation 3.48 17.61
Page 60 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
DCNN = 4
MC04283 RGM domain family, member A axon guidance -2.29 -2.37 -1.92 -1.93MC00841 four and a half LIM domains 1 protein binding -2.04 -2.33 -1.8 -1.59MC06550 unknown unknown -2.1 -2.31 -1.42 -1.47MC09691 unknown unknown -3.17 -2.2 -1.33 -1.21
UCNN = 5
MC01890 carboxylesterase 1 neuroprotection 3.29 2.97 1.55 1.15MC05339
MC05340
TC00979 RAP2B, member of RAS oncogene family signal transduction 2.43 2.23 1.5 1.2MC01571 spermidine N1-acetyltransferase polyamine homeostasis 2.75 2.03 1.72 1.55MC01067 CD63 antigen (melanoma 1 antigen) signal transduction 2.16 2.17 1.84 1.61
UNNU = 3
MC02792 nucleolar protein 5A ribosomal 2.15 1.51 1.79 2MC07460 unknown 2.03 1.47 1.97 2.38MC09319 unknown 3.68 1.67 1.92 2.16
NUCN = 3
TC00879 v-Ha-ras viral oncogene cell proliferation 1.65 2.55 2.05 1.97MC03953 hypothetical protein FLJ20303 cell proliferation -1.08 2.09 2.12 1.89MC00337 latent TGF beta binding protein 1 TGFbeta signalling 1.72 2.84 2.21 1.7
UNND = 3
MC05564 unknown unknown 2.61 1.3 -1.66 -2.17MC05563 unknown unknown 2.75 1.37 -1.8 -2.16MC07520 unknown unknown 2.27 -1.05 -1.4 -2.14
NNUU = 2
MC05041 collagen, type XI, alpha 1 ECM component -1.01 -1.22 2.13 4.38MC02833 periostin, osteoblast specific factor skeletal development -1.04 1.64 4.05 9.56
NUNN = 2
MC10107 unknown unknown -1.04 2.39 1.86 1.75MC09645 unknown unknown 1.4 2.3 1.96 1.71
UNUC = 2
MC09345 unknown unknown 5.11 1.91 3.03 4.08MC06118 unknown unknown 2.13 1.56 2.08 2.48
UDDC = 1
MC00601 transcobalamin I vitamin B transport 19.97 8.71 3.89 3
NDCD = 1
MC02800 anterior gradient 2 homolog cell survival 1.01 -2.57 -4.34 -9.82
NDCC = 1
MC09373 unknown unknown -1.16 -2.7 -2.97 -4.17
1.6 1.39ubiquitin specific protease 2 protein breakdown 3.45 2.14
Page 61 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
NUCU = 1
MC05482 hypothetical protein FLJ14712 unknown 1.15 2.16 4.11 8.65
UUCD = 1
MC01583 secreted frizzled-related protein 2 WNT signaling 8.16 17.95 17.42 7.55
UDNN = 1
MC00339 interleukin 8 receptor, beta immune response 12.11 2.66 1.22 1.38
DCDC = 1
TC00198 keratin 4 cytoskeleton component -2.01 -2.1 -5.21 -9.37
DNNN = 1
MC07468 unknown unknown -2.49 -1.5 -1.53 -1.89
NDCN = 1
MC07147 unknown unknown -1.08 -2.96 -2 -1.82
NDNN = 1
MC08249 unknown unknown -1.37 -2.04 -1.8 -1.95
NNDD = 1
MC10174 unknown unknown -1.47 -1.95 -2.62 -5.21
NNDN = 1
MC08155 unknown unknown -1.46 -1.71 -2.19 -1.88
Page 62 of 61
ScholarOne, 375 Greenbrier Drive, Charlottesville, VA, 22901
The Journal of NeuroscienceFor Peer Review Only
Top Related