Conceived and designed the experiments: MRH. Performed the experiments: HM LC SMR WL SJ AMC AW OAA PY. Analyzed the data: PD SML CG. Wrote the paper: HM TL MRH.
The authors have declared that no competing interests exist.
To determine whether optic nerve head (ONH) astrocytes, a key cellular component of glaucomatous neuropathy, exhibit differential gene expression in primary cultures of astrocytes from normal African American (AA) donors compared to astrocytes from normal Caucasian American (CA) donors.
We used oligonucleotide Affymetrix microarray (HG U133A & HG U133A 2.0 chips) to compare gene expression levels in cultured ONH astrocytes from twelve CA and twelve AA normal age matched donor eyes. Chips were normalized with Robust Microarray Analysis (RMA) in R using Bioconductor. Significant differential gene expression levels were detected using mixed effects modeling and Statistical Analysis of Microarray (SAM). Functional analysis and Gene Ontology were used to classify differentially expressed genes. Differential gene expression was validated by quantitative real time RT-PCR. Protein levels were detected by Western blots and ELISA. Cell adhesion and migration assays tested physiological responses. Glutathione (GSH) assay detected levels of intracellular GSH.
Multiple analyses selected 87 genes differentially expressed between normal AA and CA (P<0.01). The most relevant genes expressed in AA were categorized by function, including: signal transduction, response to stress, ECM genes, migration and cell adhesion.
These data show that normal astrocytes from AA and CA normal donors display distinct expression profiles that impact astrocyte functions in the ONH. Our data suggests that differences in gene expression in ONH astrocytes may be specific to the development and/or progression of glaucoma in AA.
Primary open angle glaucoma (POAG), the most common form of glaucoma, is a blinding disease that affects older adults
Astrocytes, the major glial cell type in the ONH in humans, provide cellular support function to the axons while interfacing between connective tissue surfaces and surrounding blood vessels
In this study, we referred to African American individuals (AA) as Black Americans of African ancestry and to Caucasian Americans individuals (CA) as White Americans of Western European ancestry. We used the race, gender and age identification provided with the anonymous donor history according to guidelines published in
We have investigated possible differences in ONH astrocytes from populations with different genetic backgrounds, using well characterized primary human astrocytes cultures. We have found that astrocytes derived from normal AA donors exhibit differential gene expression profiles compared to astrocytes derived from normal age-matched CA donors. Our data shows that genes associated with oxidative stress, astrocyte motility, ECM structure, immune responses and the reactive astrocyte phenotype are differentially expressed in normal astrocytes from these different populations. These results demonstrate baseline differences in ONH astrocytes from human populations with different genetic backgrounds and provide a molecular framework for future analyses from normal and glaucomatous astrocytes from different populations.
We established primary cultures of ONH astrocytes from 16 normal African American donors (age 60±11) and 21 normal Caucasian American donors (age 62±12) as described in
There were no significantly different demographic variables between populations of the 24 astrocyte lines used in microarray (
To better demonstrate the process of identifying significant genes,
A. Volcano plot indicates the size of biological effect (fold change) versus the statistical significance of the result (statistical p-value). Volcano plot represents the total number of genes used in the analysis after removing ‘absent’ genes and redundant probes (10504) on the Affymetrix Human Genome HG U133A Chip. Each point represents a gene plotted as a function of fold change (Log2 (fold change), x-axis) and statistical significance (−Log 10 (p-value), y-axis). Vertical dotted lines represent fold changes of ±1.3, respectively. The horizontal dotted line represent FDR = 0.05 (p-value is 0.00086 for this data). The red dots represent 239 selected differentially expressed genes with FDR<0.05 and fold-change >1.3. B. Estimate of the proportion of genes differentially expressed among populations. The p-value distribution of AA-CA comparison shows that a number of genes have very small p-values, which are significant even after considering the effect of multiple testing through FDR adjustment. C. Changes in gene expression in major categories in AA astrocytes, compared to CA astrocytes. The x-axis is the selected categories: signal transduction, adhesion, motility, ECM related, oxidative stress and growth factors. The y-axis is the number of genes under the category from the differentiated gene list (
Gene Symbol | Gene Title | Fold Change | P-value | Chromosome location |
PSPH | phosphoserine phosphatase | 2.70 | 0.0000 | 7q11.2 |
PDE4DIP | phosphodiesterase 4D interacting protein | 2.50 | 0.0000 | 1q12 |
RGS5 | regulator of G-protein signalling 5 | 2.35 | 0.0007 | 1q23.1 |
SOS1 | son of sevenless homolog 1 | 1.67 | 0.0001 | 2p22-p21 |
RAB3B | Member RAS oncogene family | 1.64 | 0.0006 | 1p32-p31 |
GPR56 | G protein-coupled receptor 56 | 1.62 | 0.0051 | 16q12.2-q21 |
PLA2G4C | phospholipase A2, group IVC | 1.60 | 0.0016 | 19q13.3 |
PPP1R12B | protein phosphatase 1, regulatory (inhibitor) subunit 12B, (MYPT2) | 1.53 | 0.0026 | 1q32.1 |
MYLK | myosin, light polypeptide kinase | 1.47 | 0.0435 | 3q21 |
CENTG2 | centaurin, gamma 2 | 1.35 | 0.0103 | 2p24.3-p24.1 |
NPR3 | natriuretic peptide receptor C | 1.35 | 0.0404 | 5p14-p13 |
SYDE1 | synapse defective 1, Rho GTPase, homolog 1 | 1.34 | 0.0012 | 19p13.12 |
PTK2 | protein tyrosine kinase 2 | 1.33 | 0.0393 | 8q24-qter |
ADCY3 | adenylate cyclase 3 | 1.30 | 0.0300 | 2p23.3 |
ADCY9 | adenylate cyclase 9 | 1.21 | 0.0216 | 16p13.3 |
TEK | TEK tyrosine kinase | −2.08 | 0.0006 | 9p21 |
AK3L1 (AK3) | adenylate kinase 3-like 1 | −1.81 | 0.0055 | 1p31.3 |
STAC | SH3 and cysteine rich domain | −1.71 | 0.0067 | 3p22.3 |
FZD7 | frizzled homolog 7 | −1.45 | 0.0041 | 2q33 |
ADRBK2 | adrenergic, beta, receptor kinase 2 | −1.31 | 0.001 | 22q12.1 |
WISP2 | WNT1 inducible signaling pathway protein 2 | 2.00 | 0.0001 | 20q12-q13.1 |
EFNB2 | ephrin-B2 | 1.95 | 0.0038 | 13q33 |
NLGN1 | neuroligin 1 | 1.73 | 0.0045 | 3q26.31 |
EPB41L3 | erythrocyte membrane protein band 4.1-like 3 | 1.66 | 0.0020 | 18p11.32 |
ITGA6 | integrin, alpha 6 | −1.64 | 0.0055 | 2q31.1 |
JUP | junction plakoglobin | −1.55 | 0.0024 | 17q21 |
ST3GAL5 | ST3 beta-galactoside alpha-2,3-sialyltransferase 5 | −1.50 | 0.0005 | 2p11.2 |
ANTXR1 | anthrax toxin receptor 1 | −2.30 | 0.0000 | 2p13.1 |
AMFR | autocrine motility factor receptor | 2.79 | 0.0000 | 16q21 |
MYLK | myosin, light polypeptide kinase | 1.47 | 0.0435 | 3q21 |
PPP1R12B | protein phosphatase 1, regulatory (inhibitor) subunit 12B | 1.53 | 0.0026 | 1q32.1 |
ELN | elastin | 2.20 | 0.0023 | 7q11.23 |
COL18A1 | collagen type XVIII, alpha 1 | 1.41 | 0.0169 | 21q22.3 |
LTBP1 | latent transforming growth factor beta binding protein 1 | 1.54 | 0.0239 | 2p22-p21 |
MFAP2 | microfibrillar-associated protein 2 | −1.51 | 0.0016 | 1p36.1-p35 |
MTCBP-1 | membrane-type 1 matrix metalloproteinase cytoplasmic tail binding protein-1 | −1.43 | 0.0012 | 2p25.2 |
NID2 | nidogen 2 | −1.41 | 0.0001 | 14q21-q22 |
PLOD2 | procollagen-lysine, 2-oxoglutarate 5-dioxygenase 2 | −1.60 | 0.0054 | 3q23-q24 |
GSTT2 | glutathione S-transferase theta 2 | 2.82 | 0.0000 | 22q11.2; 22q11.23 |
GGT1 | gamma-glutamyltransferase 1 | 1.62 | 0.0004 | 22q11.22 |
GGT2 | gamma-glutamyltransferase 2 | 1.54 | 0.0019 | 22q11.1 |
GGTLA4 | gamma-glutamyltransferase-like activity 4 | 1.44 | 0.0091 | 20p11.1 |
GSTM4 | glutathione S-transferase M4 | 1.42 | 0.0245 | 1p13.3 |
GSTM1 | glutathione S-transferase M1 | 1.38 | 0.0010 | 1p13.3 |
GSTM3 | glutathione S-transferase M3 | 1.31 | 0.0145 | 1p13.3 |
GSTM2 | glutathione S-transferase M2 | 1.30 | 0.0021 | 1p13.3 |
IGFBP3 | insulin-like growth factor binding protein 3 | 1.74 | 0.0488 | 7p13-p12 |
FGF9 | fibroblast growth factor 9 (glia-activating factor) | 1.51 | 0.0268 | 13q11-q12 |
GFRA1 | GDNF family receptor alpha 1 | 1.49 | 0.0060 | 10q26 |
CX3CL1 | chemokine (C-X3-C motif) ligand 1 | 1.48 | 0.0498 | 16q13 |
IGFBP5 | insulin-like growth factor binding protein 5 | 1.46 | 0.0113 | 2q33-q36 |
VEGF | vascular endothelial growth factor | −1.48 | 0.0051 | 6p12 |
HBEGF | heparin-binding EGF-like growth factor | −1.57 | 0.0044 | 5q23 |
To classify differentially expressed genes, we separated the genes from RMA-SAM analysis manually by function using available data from public databases (UniGene, OMIM and Entrez PubMed). The complete list of the functional categories is shown in
RNA from 13 AA and 17 CA age-matched normal donors was used to validate microarray data. Twenty-six genes were selected from 10 functional groups from the RMA-SAM analysis and from the GO lists. 18S was used to normalize the expression value. Fold change obtained by qRT-PCR correlated well with the direction of fold change obtained from the normalized intensity data, confirming the validity of the microarray gene expression patterns (
Genes that are associated with intracellular signaling were differentially expressed in AA astrocytes, including cAMP signaling, intracellular vesicular transport, G protein regulation and protein phosphatases (
Expression of the RGS5 gene was upregulated in AA astrocytes by microarray and confirmed by qRT-PCR (
A. Confirmation of six differentially expressed signaling genes by qRT-PCR in human normal ONH astrocytes: RGS5, ADCY3, ADCY9, AK3L1, SOS1 and ADRBK2. Genes were normalized to 18S. Graphical representation of the relative mRNA levels in normal AA and CA astrocytes (n = 8, respectively, two-tailed t test was used. **indicates p<0.01 and * indicates p<0.05). B. Cellular localization of the Regulator of G protein signaling 5 (RGS5) in primary cultures of ONH astrocytes. Immunofluorecent staining of RGS5 (red) demonstrated higher levels of RGS5 protein in AA astrocytes, compared to CA astrocytes. Nuclei stained with DAPI (blue). Note that RGS5 localizes to the cytoplasm and in the nucleus of astrocytes (arrows). Magnification bar: 25 µm. C. Representative Western blots of astrocyte cell lysates with human RGS5 antibody and β-actin used as a loading control. Note that AA1-4 normal donors express more RGS5 than CA1-4 donors. D. Representative double immunofluorescent staining of RGS5 (red) astrocyte marker GFAP (green) in sections of human ONH from an AA donor (71 year-old male) and a CA donor (75 year-old male). Nuclei stained with DAPI (blue). Note strong granular staining of RGS5 in astrocytes (arrows) in the cribriform plates in the lamina cribrosa. Fewer astrocytes stain for RGS5 in the lamina cribrosa of a CA donor. V: blood vessel, NB: nerve bundle, Magnification bar 55 µm.
Amongst genes differentially regulated in AA astrocytes were several genes that impact upon cAMP signaling. β-adrenergic receptor kinase (ADRBK2) is downregulated in AA normal astrocytes (
Comparing AA to CA astrocytes, differentially expressed genes that are associated with cell adhesion were ephrin B2 and GPR56, which were both upregulated, and ITGA6, which was downregulated (
A. Cellular localization of G protein-coupled receptor 56 (GPR56), ephrin-B2 (EFNB2) and integrin α 6 (ITGA6) in primary cultures of ONH astrocytes. Nuclei stained with DAPI (blue). Magnification bar: 25 µm. Upper: Double immunofluorescence for GFAP (green), an intermediate filament characteristic of astrocytes and GPR56 (red). Note granular staining for GPR56 (red) is more abundant in the cytoplasm of AA astrocytes compared to CA astrocytes. Middle: Immunofluorescence showed that EFNB2 is more abundant in the cytoplasm of AA astrocytes compared to CA astrocytes. Lower: Immunofluorescence showed that Integrin α 6 is less abundant in the cytoplasm of AA astrocytes compared to CA astrocytes. B. Confirmation of three differentially expressed adhesion genes by qRT-PCR in human normal ONH astrocytes: GPR56, EFNB2 and ITGA6. Genes were normalized to 18S. Graphical representation of the relative mRNA levels in AA and CA astrocytes (n = 8, respectively, *indicates p<0.05 in two-tailed t-test). C. Representative Western blots of astrocyte cell lysates with GPR56, EFNB2 and ITGA6 antibodies. β-actin was used as a loading control. Note that AA1-4 donors express more GPR56 and EFNB2, less ITGA6 than CA1-4 donors. D. Representative immunohistochemistry showed more abundant granular staining of GPR56 (red) in astrocytes in the lamina cribrosa from AA donors compared to CA donors. Note that GPR56 is also localized in astrocyte processes in the nerve bundles (NB). CP: cribriform plates, Magnification bar: 25 µm. E. AA astrocytes adhered to collagen IV 26.5% less than CA astrocytes did (* indicates p<0.05 in two-tailed t-test). Values represent mean optical density (OD)±standard deviation of triplicate experiments using primary astrocyte cultures of four AA donors and six CA donors.
The molecular changes described above suggested to us decreased cell adhesion of AA astrocytes. An assay was used to compare AA astrocytes to CA astrocytes for adhesion to collagen type IV. We found decreased attachment to collagen type IV of AA astrocytes compared to CA astrocytes (
mRNA levels of the autocrine motility factor receptor (AMFR), myosin light chain kinase (MYLK) a calcium/calmodulin dependent kinase and PPP1R12B (also refered as MYPT2), a myosin phosphatase were upregulated in AA astrocytes compared to CA astrocytes (
A. Confirmation of three differentially expressed motility genes by qRT-PCR in human normal ONH astrocytes: autocrine motility factor receptor (AMFR), myosin light chain kinase (MYLK) and myosin phosphatase target subunit 2 (MYPT2). Genes were normalized to 18S. Graphical representation of the relative mRNA levels in AA and CA astrocytes (n = 8, respectively, * indicates p<0.05 in two-tailed t-test). B. Representative Western blots of astrocyte cell lysates with MYLK antibody. β-actin was used as a loading control. Note that AA1-4 donors express more MYLK 130 kDa than CA1-4 donors. No difference was detected at the levels of 210 kDa isoforms. C. Densitometry analysis of MYLK western blots. β-actin was used as loading control. Astrocytes derived from 7 AA and 10 CA were used in this experiment. The level of the 130 kDa isoform was significantly higher in AA astrocytes, compared to CA astrocytes. D. Cell migration assay shows that AA astrocytes migrate significantly faster than CA astrocytes. The assay was performed as described in the
Western blots detected two isoforms of MYLK: 210 kDa and 130 kDa. The 130 kDa isoform was the predominant form in both AA and CA normal astrocytes. The protein level of MYLK 130 kDa was significantly higher in AA astrocytes compared to CA astrocytes (
Based on the gene expression and protein data for MYLK, we compared migration in AA and CA astrocytes. Experiments using a chemotaxis model indicated that AA astrocytes migrated faster than CA (
Microarray analysis indicated a significant upregulation of elastin (ELN) mRNA which was confirmed by qRT-PCR and immunoblot in AA astrocytes (
A. Confirmation of three differentially expressed adhesion genes by qRT-PCR in human normal ONH astrocytes: Elastin (ELN), microfibrillar-associated protein 2 (MFAP2) and latent transforming growth factor beta binding protein 1 (LTBP1). Genes were normalized to 18S. Graphical representation of the relative mRNA levels in normal AA and CA astrocytes (n = 8, respectively, two-tailed t-test was used. ** indicates p<0.01 and * indicate p<0.05). B. Representative Western blots of astrocyte cell lysates with LTBP1, ELN and collagen type XVIII antibodies. β-actin was used as a loading control. Note that AA1-4 donors express more LTBP1, ELN and collagen type XVIII than CA1-4 donors. C, D. Immunocytochemistry of LTBP1 in AA and CA astrocytes. LTBP1 (red) is more abundant in the cytoplasm and extracellular space in AA astrocytes compared with CA astrocytes. Nuclei stained with DAPI (blue). Magnification Bar: 25 µm. E, F. Double immunostaining of ELN (red) and GFAP (green) in representative cross sectional of the ONH tissue in AA and CA donors. ELN is located in the cribriform plates and not in the nerve bundles. Astrocytes cell bodies are located in the cribriform plates (CP) and extend processes into the nerve bundle (NB). Note that there are no apparent differences in ELN staining between AA and CA samples. E is from a 75 year-old AA male donor and F is from a 74 year-old CA female donor. Colocalization between ELN and GFAP in the AA tissue is microscopic effect showing the overlap between ELN and astrocytes processes. Magnification bar: 25 µm. G, H. Double immunostaining of collagen type IV (red) and GFAP (green) in representative sagittal sections of ONH tissues from AA and CA donors. Collagen type IV and GFAP follow the lamellar structure of the astrocytic basement membranes in the human lamina cribrosa. Note that staining for collagen type IV is more intense and abundant in the CA donor than in the AA donor. G is from a 65 year-old AA male donor and H is from a 57 year-old CA male donor. V: blood vessel, Magnification bar: 35 µm.
Several genes associated with basement membranes were downregulated in AA astrocytes (
Genes involved in GSH metabolism, including glutathione S-transferases (GSTs) and gamma-glutamyltransferases (GGTs), are upregulated in AA astrocytes (
A. Confirmation of two differentially expressed glutathione metabolic enzyme genes by qRT-PCR in human normal ONH astrocytes: glutathione S-transferase theta 2 (GSTT2) and gamma-glutamyltransferase 1 (GGT1). Genes were normalized to 18S. Graphical representation of the relative mRNA levels in AA and CA astrocytes (n = 8, respectively, * indicate p<0.05 in two-tailed t-test). B. AA astrocytes have significantly lower level of intracellular GSH in vitro, compared to CA astrocytes (n = 10, ** indicated p<0.01 in two-tailed t-test). GSH content is normalized by the amount of the protein. C. Representative Western blots of astrocyte cell lysates with heat shock 70 kDa protein (HSP70) antibody. β-actin was used as a loading control. HSP70 levels are higher in AA astrocytes compared to CA astrocytes. D. Immunofluorecent staining for HSP70 shows that AA astrocytes exhibit more abundant intracellular HSP70 staining compared to CA astrocytes. Magnification bar: 25 µm.
Several chaperones were upregulated in AA astrocytes compared to CA astrocytes, including heat shock protein 70 protein 2 (HSPA2), alpha-crystallin-related heat shock protein B6 (HSPB6), and crystallin-β B2 (CRYBB2) (
Among the upregulated genes by microarray in AA astrocytes compared to CA astrocytes were two members of the IGFBP family, IGFBP3 and IGFBP5, the GDNF family receptor α1, the receptor for glial derived growth factor-1 and neurturin (
A. Confirmation of two differentially expressed growth factor genes by qRT-PCR in human normal ONH astrocytes: insulin-like growth factor binding protein 5 (IGFBP5) and heparin-binding EGF-like growth factor (HBEGF). Genes were normalized to 18S. Graphical representation of the relative mRNA levels in normal AA and CA astrocytes (n = 8, respectively, two-tailed t-test was used. * indicate p<0.05). B, C. Seven CA samples and six AA samples were used in ELISA experiment. B. The total amount of secreted and intracellular HBEGF is significantly lower in AA astrocytes compared to CA astrocytes (** indicated p<0.01 in two-tailed t-test) C. The level of secreted IL-6 is significantly lower in AA astrocytes compared to CA astrocytes (* indicated p<0.05 in two-tailed t-test).
To verify the expression of growth factors and cytokines, we used ELISA to measure the quantity of both secreted (medium) and intracellular (cell lysate) forms in 6 AA and 7 CA age matched astrocyte lines. Consistent with lower mRNA levels in AA astrocytes (
Astrocytes are the major cell type that controls the homeostasis and microenvironment of the retinal ganglion cells axons as they traverse the optic nerve head. The complex interplay of astrocytes with RGC axons involves cell-cell signaling, synthesis of ECM, control of ions and pH, inter- and intracellular transport, immune surveillance, synthesis of growth factors and cytokines, and many other interactions. This study presents data that demonstrate population based differences (AA compared to CA) that may affect important astrocyte functions in the normal ONH. Based on data from gene expression, protein levels and functional assays
Previous work in our laboratory demonstrated that ONH astrocytes in culture retain many of the phenotypic characteristics that ONH astrocytes display
Amongst genes differentially regulated in AA astrocytes were several genes that impact cAMP signaling. β-adrenergic receptor kinase (ADRBK2), which is downregulated in AA normal astrocytes, specifically phosphorylates the agonist-occupied form of the β-adrenergic receptor and promotes its desensitization and internalization
AA astrocytes exhibited differential expression of genes associated with cell adhesion and cell migration. Cell adhesion genes which were downregulated in AA astrocytes included: integrin α-6 (ITGA6), which connects these cells with laminin family members
Upregulated genes in AA astrocytes that are involved in cell adhesion of astrocytes included: GPR56, that activates pathways that inhibit cell adhesion
Cell migration involves changes in adhesion and cytoskeletal proteins and changes in cell shape during movement via reorganization of actin filament networks in the cell periphery
Using microarray analysis, we have found changes in the expression of genes that suggest the machinery for migration is enhanced in AA astrocytes. In AA astrocytes, these changes included: upregulation of MYLK, a calcium/calmodulin dependent enzyme which phosphorylates myosin regulatory light chains to facilitate myosin interaction with actin filaments, producing contractile activity
MYLK genetic variants confer increased risk of sepsis and sepsis-associated with acute lung injury and a more severe asthma phenotype in individuals of African ancestry
Our microarray data suggests that alterations in elastic fibers and associated microfibrils, a major component of the ECM in the lamina cribrosa, may be a susceptibility factor predisposing the remodeling of the ONH in response to elevated IOP in glaucoma in the AA population. Microarray analysis indicated a significant upregulation in AA astrocytes of elastin (ELN), LTBP1, a member of the elastin microfibrils that binds TGF-β, and COL18A1, a collagen with strong anti-angiogenic properties. MFAP2, the gene encoding for MAGP1, a component of the elastin associated microfibrils, and the gene encoding for LOXL2, a lysyl oxidase that participates in maturation of elastic fiber, were downregulated in AA astrocytes. We previously reported that normal AA astrocytes expressed high levels of elastin mRNA and protein, and decreased levels of LOXL2
LTBP1 is an ECM glycoprotein that plays a major role in storage of latent TGF-β in the ECM and regulate its availability
Additional ECM related genes that were downregulated included: nidogen 2 (NID2), a linker protein that joins laminin and collagen IV networks in basement membranes
Taken together, population based differences in important components of the ECM that provide elasticity and resiliency to the lamina cribrosa, and in basement membranes that participate in tissue homeostasis and adhesion of astrocytes in the ONH, may render the tissue more susceptible to the elevation of IOP that occurs in glaucoma.
In normal AA astrocytes the GDNF family receptor α 1 (GDRA1) was upregulated. GDRA1 is the receptor for Glial -derived neurotrophic factor (GDNF) and neurturin (NTN) which are two potent neurotrophic factors that play key roles in the control of neuron survival and differentiation
Two important members of the IGF axis, the insulin like growth factor binding protein 5 and 3 (IGFBP5 and IGFBP3) were upregulated in normal AA astrocytes. IGFBP5 interacts with heparin containing glycosaminoglycans (GAGs) in the ECM and facilitates migration, an important differentiated function of normal AA astrocytes compared to CA
The membrane associated heparin binding EGF (HBEGF) was downregulated in AA astrocytes. Several ligands and receptors of the EGF family are known to be expressed in ONH astrocytes including HBEGF
AA astrocytes exhibit downregulation of expression of IL-6, a key cytokine upregulated in reactive astrocytes that promotes glial scar formation and is an impediment to axon regeneration and neuronal survival in the CNS
Major histocompatibility complex (MHC) Class I genes (HLA-A, HLA-B, HLA-F and HLA-G) and other molecules in the Class I antigen presentation pathway, such the immunoproteasome (PSMB9, PSMB8), were marginally up-regulated in AA astrocytes. HLA-A and HLA-B belong to the classical class I genes and are expressed in astrocytes in the CNS in disease
Astrocytes in the CNS participate in the innate immune response by modulating local reactions to endogenous or exogenous antigens, and by modulating astrogliosis through release of cytokines and by isolating areas of inflammation
The antioxidant glutathione (GSH) is vital for cellular defense against oxidative stress in astrocytes and neurons
Upregulation of GSTs in AA astrocytes may indicate active detoxification activity by GSH metabolizing enzymes, resulting lower levels of GSH as we demonstrated
Our study raises many questions that will require future investigations. Our data indicates that the microenvironment supported by astrocytes in the normal ONH in AA has important differences that may impact susceptibility to glaucomatous optic neuropathy. An important question that remains is whether the observed differential gene expressions in AA astrocytes compared to CA astrocytes are the result of continued exposure to stress signals
Twenty one human eyes from 21 normal age-matched Caucasian American (CA) donors (age 62±12) and 16 human eyes from 16 normal African Americans (AA) (age 60±11) were used in this study to generate primary cultures of optic nerve head (ONH) astrocytes (
Donors did not have history of eye disease, diabetes, or chronic CNS disease. Eyes were obtained from the local eye banks and from the National Disease Research Interchange (NDRI) (
Cultures of human ONH astrocytes were generated as previously described
Normal human eyes used for microarray were from 12 CA normal donors (age 58±11) and 12 AA normal donors (age 58±12) (
The first step in the analysis of the microarray data was to determine which genes were to be considered “present” or “absent.” We estimated the probe-set present/absent calls by using the Wilcoxon signed rank-based algorithm
We used RMA and Bioconductor in R
To assess the biological significance of the gene list derived above we classified differentially expressed genes in normal AA and CA by manual separation by function using available data from public databases such as UniGene, OMIM and Entrez PubMed and the gene lists from RMA/SAM.
In addition, we used GOstats Bioconductor package
Independent confirmation of differential expression was conducted using 17 astrocyte cultures from CA and 13 astrocyte cultures from AA age-matched donors obtained as described earlier in
Human ONH astrocytes were grown until reaching 90% confluence and Western blot analysis was carried out as previously described
ELISA was used to determine the content of HB-EGF and IL-6, using ELISA kits (R & D Systems) specific for each protein. Both cell lysate and culture medium were measured. When cells reached confluence, they were washed with PBS before changed into serum-free DMEM/F-12 medium. After 24 hours of incubation, medium was collected, clarified by centrifugation, and concentrated using Centriprep YM-10 (Millipore) with a molecular weight cut-off of 10,000 kDa. As a control, fresh medium was collected and concentrated in the same fashion. Cells were collected and lysed in 100 µl of 50 mM phosphate buffer with sonication. Five µl of the supernatant was used to measure protein concentration using the Pierce Protein Assay Kit (BCA method). Assay was carried out according to manufacture's protocol. The content for each sample was calculated based on the standard curve, and results were expressed as content (nmol) per mg protein. Six normal AA and seven CA samples were used. The means of the content were considered significantly different if p<0.05 (unpaired t-test).
Adhesion was measured in six CA samples and four AA samples using CytoSelect™ Cell Adhesion Assay (Cell Biolabs). Cells were harvested and resuspended in serum-free DMEM/F-12 medium. For each donor, four Collagen IV coated wells and one BSA coated well were used. Fifty thousand cells were added to each well in a volume of 150 µl. For each donor, one Collagen IV well was treated with sterile serum-free DMEM F12 as a negative control. The plate was then incubated at 37°C, 5% CO2 for 90 minutes. The cells were then washed 4 times with 250 µl of PBS. 200 µl of cell stain solution (Cell Biolabs) was then added to each well and incubated for 10 minutes at room temperature. The stain was then removed and the cells were washed four times with 500 µl deionized water. After allowing the cells to air dry for 25 minutes, 200 µl of extraction solution (Cell Biolabs) was added to each well and placed on an orbital shaker for 10 minutes. Extracted samples' absorbance was measured at 560 nm by a Thermo Multiskan Spectrum plate reader. Results were corrected by subtracting absorbance of the negative controls. The means from AA and CA groups were considered significantly different if p<0.05 (unpaired t-test).
Migratory properties of five CA samples and six AA samples were measured by CytoSelect™ 24-well cell migration assay (Cell BioLabs). Assay was performed according to the manufacture's protocol. Five hundred µl of media supplemented with 10% fetal bovine serum was added to the bottom wells. Fifty thousand cells resuspended in 300 µl serum-free medium were added to the inserts. The plate was then incubated at 37°C, 5% CO2 for 24 hours. After removing the media from the inserts, non-migratory cells from the interior of the inserts were removed with cotton-tipped swabs. The inserts were stained with 400 µl of cell staining solution and washed three times with water before transferred to a clean well with 200 µl of extraction solution. The absorbance of the extracted samples was measured at 560 nm by a Thermo Multiskan Spectrum plate reader. The means from AA and CA groups were considered significantly different if p<0.05 (unpaired t-test). To inhibit MYLK, 10 µM of ML-7 (EMD Biosciences)
Total GSH content was quantified in astrocyte cell lysates based on previously described methods using the Glutathione Assay Kit from Cayman Chemical (Ann Arbor, MI)
Six eyes from normal Caucasian donors (CA) and six eyes from normal age-matched African American (AA) donors were used. Detail methods are described in
Primary ONH astrocytes from six normal AA donors and from six normal CA donors were used. Detail methods are described in
Demographic information of normal donors. Demographic information of CA and AA normal donor eyes used to generate primary cultures of ONH astrocytes.
(0.09 MB DOC)
Genes differentially expressed in ONH astrocytes from normal AA donors compared to their caucasian counterpart using RMA SAM. List of genes differentially expressed in AA ONH astrocytes compared to CA ONH astrocytes, obtained by RMA-SAM analysis.
(0.11 MB DOC)
Genes differentially expressed in normal AA ONH astrocytes compared to their CA counterpart using Lima. List of genes differentially expressed in AA ONH astrocytes compared to CA ONH astrocytes, obtained by using Limma package in Bioconductor.
(0.06 MB DOC)
Functional analysis of genes differentially expressed in AA vs. CA. Functional classification of genes differentially expressed in AA ONH astrocytes compared to and CA ONH astrocytes.
(0.06 MB DOC)
Selected Gene ontology for AA-CA comparison. Gene Ontology for genes differentially expressed in AA ONH astrocytes compared to and CA ONH astrocytes.
(0.04 MB DOC)
Real-Time PCR validation of microarray expression analysis of normal Caucasian American and African American ONH astrocytes. Real-Time PCR validation of microarray expression analysis of normal CA and AA ONH astrocytes.
(0.08 MB DOC)
Quantitative RT-PCR primer information. RT-PCR primer infortion
(0.10 MB DOC)
Primary Antibodies Used in this study. Information on the primary antibodies used in this study.
(0.04 MB DOC)
Supplemental Methods. Detailed methods for Western Blots, immunohistochemistry and immunocytochemistry.
(0.05 MB DOC)
We gratefully acknowledge Jose Bongolan for immunohistochemistry staining and Marina Vracar-Grabar for cell culture. We also thank Genechip Core Facility of Washington University for microarray experiments.