The Influence of the Proinflammatory Cytokine, Osteopontin, on Autoimmune Demyelinating Disease Dorothée Chabas, Sergio E. Baranzini, Dennis Mitchell, Claude C. A. Bernard, Susan R. Rittling, David T. Denhardt, Raymond A. Sobel, Christopher Lock, Marcela Karpuj, Rosetta Pedotti, Renu Heller, Jorge R. Oksenberg, and Lawrence Steinman

Supplemental Table 1. Genes differentially expressed in MS and normal libraries. Only genes corresponding to transcripts with an average fold difference (AFD) of 2 are listed. The first section of the table lists genes whose expression was statistically significant in both MS libraries when compared with the CTRL library (Fisher's exact test, P < 0.05). The second section contains genes with significant difference in expression in only one of the MS libraries and the CTRL. The last section includes genes with nonsignificant differences but AFD ±3.00. N/A, mapping position is not known. Asterisk indicates genomic regions that reached nominal criteria of linkage in genome-wide screenings.
Accession #	Gene description	Clone count			Average	Cellular function	Genomic location
		MS1	MS2	CTRL	fold difference
Section 1. Only genes corresponding to transcripts with an AFD of 2
M17885	Acidic ribosomal phosphoprotein P0	9	13	2	5.2	Gene/protein expression	12q24 *
X16869	Elongation factor 1-alpha	32	33	15	2.07	Gene/protein expression	6q14
M54927	Myelin proteolipid protein	41	16	64	-3	Cell structure/motility	Xq22 *
U66623	Small GTPase	1	1	7	-7.35	Cell signaling/cell communication	2q21.2
M26252	Piruvate kinase, muscle	2	1	10	-8.04	Energy metabolism	15q22
Section 2. Genes with significant difference in expression in only one of the MS libraries and the CTRL
M59828	Heat shock protein HSP70-1	2	14	1	7.29	Cell/organism defense	6p21.3 *
AF068846	Scaffold attachment factor A	6	1	1	2.49	Cell division	N/A
AF035283	Clone 23916	10	15	5	2.36	Unclassified	N/A
U46571	Tetratricopeptide repeat protein 2	10	4	3	2.29	Unclassified	17q11.2
AF131756	Clone 24912	2	10	10	-3.02	Unclassified	N/A
X92845	N-myc downstream regulated	1	7	7	-4	Unclassified	8q24.1
M97168	X (inactive)-specific transcript	1	13	10	-4.33	Unclassified	Xq13.2
AB023167	Neural membrane protein 35 (KIAA0950)	1	3	7	-4.74	Unclassified	12q13
AB002391	HERC2 (KIAA0393)	2	1	7	-5.63	Unclassified	15q13
AF055026	RaP2-interacting protein 8	1	1	6	-6.3	Unclassified	17
U89330	Microtubule-associated protein 2	1	1	6	-6.3	Cell structure/motility	2q34-q35
Section 3. Genes with nonsignificant differences but AFD ±3.00
M20020	Ribosomal protein S6	5	6	1	5.23	Gene/protein expression	9p21 *
X03747	Na/K-ATPase subunit	5	5	1	4.78	Energy metabolism	1q22-q25
V00572	Phosphoglycerate kinase	5	4	1	4.33	Metabolism	Xq13
M59488	S100 protein subunit	5	4	1	4.33	Cell signaling/cell communication	21q22.3
AB018271	KIAA0728 (brain)	4	4	1	3.83	Unclassified	6p11-11.2
AB020718	KIAA0911	4	3	1	3.38	Unclassified	1p36 *
AAD02202	CaM-KII inhibitory protein (Rattus norvegicus)	2	5	1	3.26	Unclassified	N/A
D67025	Proteasome 26S subunit (non-ATPase, 3)	1	1	3	-3.15	Cell/organism defense	17q21.1
D63424	Glycogen synthase kinase 3	1	1	3	-3.15	Metabolism	19q13.3-13 *
AF051976	Unconventional myosin XV	1	1	3	-3.15	Cell structure/motility	17p11.2
X13916	LDL receptor-related protein	1	1	3	-3.15	Metabolism	12q13.q14
AB028981	KIAA1058	1	1	3	-3.15	Unclassified	13
AF102846	N-Ethylmaleimide-sensitive factor	3	1	5	-3.61	Metabolism	17q21
L10284	Calnexin	3	1	5	-3.61	Cell/organism defense	5q35
D88435	Cyclin G-associated kinase	1	2	5	-3.85	Cell division	4p16
AF054987	Aldolase C	2	1	5	-4.02	Metabolism	17cen-q12
CAB01750	similar to Mitochondrial carrier proteins [Caenorhabditis elegans)	1	1	4	-4.2	Unclassified	N/A
AL137406	Clone DKFZp434M162	1	1	4	-4.2	Unclassified	N/A
AB032436	Brain-specific Na+-dependent inorganic phosphate cotransporter	1	1	4	-4.2	Metabolism	19q13 *
L10911	Splicing factor (CC1.3)	1	1	4	-4.2	Gene/protein expression	20
L77864	Amyloid -protein (A4) precursor-binding, family B, member 1 (Fe65)	1	4	7	-4.42	Unclassified	11p15
U64520	Synaptobrevin-3	1	1	5	-5.25	Unclassified	1p35-p36
D87465	KIAA0275 (brain)	1	1	5	-5.25	Unclassified	10

Supplemental Figure 1. (A) Frequency distribution of libraries. The libraries were made in collaboration with Incyte Genomics. Libraries were constructed using 1.5 mg of poly(A)⁺ RNA from each sample. cDNA synthesis was initiated using a Not I-anchored oligo(dT) primer. Double-stranded cDNA was blunted, ligated to Eco RI adaptors, digested with Not I, size-selected, and cloned into the Not I and Eco RI sites of the pINCY vector (Incyte, Palo Alto, CA). About 4000 clones from each library were sequenced in ABI automatic DNA Sequencer (Applied Biosciences, Foster City, CA). Annotated data were extracted from the Incyte database LifeSeq Gold and incorporated into MS Access 2000 and MS Excel 2000 for further analysis. All queries were designed and performed in MS Access; charts and tables were generated with MS Excel. Cellular roles were assigned after consulting the Expressed Gene Anatomy Database (EGAD, The Institute for Genomic Research, www.tigr.org/egad). Genomic location was included according to NCBI's MapView and Genemap'98 (National Institute for Biotechnology Information, www.ncbi.nlm.nih.gov). Comparisons of gene frequencies between each MS library and the CTRL were performed, and the average fold change was calculated. Differences in gene expression were subjected to Fisher's exact test, and a P value of 0.05 or lower was selected as the criterion for inclusion in each comparison.

We sequenced 3678, 4174, and 3740 clones from MS1, MS2, and control libraries, respectively. Each of the libraries had a substantial number of clones with no match to the GenBank database, and were thus considered novel. Clones in library MS1 could be assigned to 2387 different cDNA species from which 331 corresponded to novel genes. MS2 and CTRL yielded 2727 (546 novel) and 2352 (511 novel) species respectively. Analysis of frequency distribution revealed a similar pattern for all three libraries, with the most abundant transcripts being represented by few species including two myelin genes, myelin basic protein (MBP) and proteolipid protein (PLP), and the astrocyte-specific transcript, glial fibrillary acidic protein (GFAP). Similarly, there was an exponentially decreasing frequency observed for less frequent cDNA species in all three libraries. Taken together, the data reveal that the composition and complexity of the three libraries were similar, and that there were no obvious biases, therefore enabling comparative analysis.

Clones with a count higher than six were organized in decreasing order according to their frequency. Arrows indicate the three most common ESTs in each library. MBP was the most highly expressed gene in all three libraries. GFAP and PLP were the next most abundant species in the MS libraries, but their frequency order was reversed in the control library. Unidentified ESTs are shown in lighter color than known, annotated clones.

(B) Category distribution. Clones were distributed into one of the following categories: RN, redundant novel; RK, redundant known; HA, high abundance; SN, solitary novel; and SK, solitary known. The relative contribution to each category is shown in a pie chart for all libraries.

(C) Intersectional queries. All possible comparisons were performed among the three libraries. Clones were counted and distributed into their corresponding intersection on the Venn diagrams. The total number of sequenced clones is shown for each library. The number of different mRNA species for each library is also shown along with the number of unknown genes in parentheses. The number of RNA species that were specific for each library or intersection of libraries is displayed underlined, along with the number of unknown genes in parentheses.

Medium version | Full size version

Methods and Experimental Details
Details of immunohistochemical staining. After perfusion with 100 ml of 10% formalin, brains were removed at seven different time points between day 11 and day 28 after immunization. OPN was detected with the OPN-specific antibody MPIIIB10₁, dilution 1/50. We used the secondary biotinylated rat adsorbed anti-mouse IgG (H + L) antibody from Vector Laboratories (catalog no. BA 2001). The staining was evaluated by an observer blinded to the study design, according to a semiquantitative scale (3 grades).

Details of induction of EAE. We induced EAE with MOG 35-55 in CFA in 129/C57BL/6 OPN ^+/+mice, and 129/C57BL/6 OPN^+/+ controls. Here we slightly modified the protocol: we injected 100 g of MOG 35-55 emulsion subcutaneously into the flanks of each female at day 0, and 400 ng of pertussis toxin at day 0 and day 2.

Seven OPN^+/+ and 6 OPN^-/- mice were examined on days 28, 48, and 72 after immunization for histopathology. These unpublished data show similar numbers and appearances of inflammatory foci within the central nervous system in the two groups.

Details of the transcriptional profiling with custom microarrays. Spinal cord was homogenized in TRIzol reagent (Gibco BRL) using a Polytron homogenizer (Kinematica AG, Switzerland), and total RNA was prepared according to the recommended protocol. mRNA was purified by two rounds of selection using oligo(dT) resin (Oligotex, Qiagen). mRNA (2 g) was used to prepare double-stranded cDNA (Superscript, Gibco BRL). The primer for cDNA synthesis contained a T7 RNA polymerase promoter site. cDNA (1 g) was used for an in vitro transcription reaction (Ambion T7 Megascript) with biotinylated CTP and UTP (Enzo Diagnostics, Inc.). The labeling procedure amplifies the mRNA population ~60-fold. Microarray chips (GeneChip⁽ System, Affymetrix) were hybridized for 16 hours in a 45°C incubator with constant rotation at 60 rpm. Chips were washed and stained on a fluidics station, and scanned using a laser confocal microscope. Affymetrix provided the procedures for sample preparation, fluidics station, scanner, and computer workstation. Chips were analyzed with GeneChip v3.1 software, and scaled to a value of 150. The software determines whether a particular RNA transcript is present or absent, based on the intensity of the signal. Fold change was calculated by divided the intensity of the average difference change in the experimental sample by the intensity of the average difference change in the control.

Further details of the initial experiment with custom microarrays described in (11).
Custom microarrays were designed that allowed large-scale profiling of mRNA for 517 components of the inflammatory response, including cytokines, chemokines, various adhesion molecules, and matrix metalloproteases. Profiles of mRNA transcripts from the spinal cord of six Lewis rats with EAE were analyzed. Rats were immunized with 400 g of guinea pig spinal cord homogenate (GPSCH) and monitored for EAE as previously described (12). mRNA was isolated from the brain and spinal cord of three rats with hind limb paralysis (mean EAE score 2.7, indicating severe paraplegia), 15 days after immunization with GPSCH, and from three rats treated with a metalloprotease inhibitor after the initiation of EAE. It is established that matrix metalloprotease inhibitors can reverse EAE (13), and rats treated with the metalloprotease inhibitor (RS110379) displayed no clinical disease (mean EAE score 0.2). Spinal cord from two other normal rats served as controls.

OPN transcripts were increased 3.4-fold in the spinal cord of rats with EAE and paralysis, compared with controls without EAE (average difference change for intensity of OPN transcripts was 16,609 fluorescent units in untreated rats with EAE versus an average difference change for intensity in OPN transcripts of 4846 in rats without EAE) (14). After treatment with RS110379, levels of OPN mRNA were no different than control rats without EAE. Thus, there was a 1.1-fold change between the intensity of OPN transcripts in EAE rats treated with MMP inhibitor and rats without EAE (the average difference of the change in intensity of OPN transcripts on the custom microarrays was 5176 units in rats treated with RS110379 versus an average difference intensity of 4846 units in rats without EAE) (14).