Gene expression in human lupus: bone marrow differentiates active from inactive patients and displays apoptosis and granulopoiesis signatures

Abstract

Objective

The cells of the immune system originate from the bone marrow (BM), where many of them also mature. To better understand the aberrant immune response in systemic lupus erythematosus (SLE), we examined the BM in lupus patients using DNA microarrays and compared it to the peripheral blood (PB).

Patients and Methods

Bone marrow mononuclear cells (BMMCs) from 20 SLE patients (11 with active disease and 9 with inactive disease) and peripheral blood mononuclear cells (PBMCs) from 27 patients (16 active/ 11 inactive); BMMCs and PBMCs from 7 healthy individuals and 3 osteoarthritis patients served as controls. Samples were analyzed on genome-scale microarrays with 21,329 genes represented.

Results

We found 102 differentially expressed genes between patients’ and controls’ BMMCs (unpaired student t-test), involved in various biologic processes; 53 of them are involved in major networks including cell death, growth, signaling and proliferation. Comparative analysis between BM and PB of patients identified 88 genes differentially expressed; 61 out of 88 participate in cell growth and differentiation, cellular movement and morphology, immune response and other hematopoietic cell functions. Unsupervised clustering of highly expressed genes revealed two major SLE patient clusters (active and inactive) in BM, but not in PB. The upregulated genes in the bone marrow of active patients included genes involved in cell death and granulopoiesis.

Conclusion

Microarray analysis of the bone marrow differentiates active from inactive lupus patients and provides further evidence for the role of apoptosis and granulocytes in the pathogenesis of the disease.

Systemic lupus erythematosus (SLE) is the prototypic systemic autoimmune disease characterized by the production of autoantibodies to components of the cell nucleus in association with diverse clinical manifestations encompassing almost all organ systems. Although its etiology is not established, much is known about the pathogenic pathways that lead to tissue injury (1, 2)

Bone marrow is a central lymphoid organ consisting of various types of hemopoietic and non- hemopoietic cells and the “stroma” that supports their growth, differentiation and function, collectively called as “the BM micro-environment” (3) In addition to production, maturation and activation of neutrophils, monocytes/macrophages and B cells, BM has a central role in regulating the immune response (4, 5). In lupus patients, bone marrow exhibits a variety of histopathologic findings including necrosis, stromal alterations and abnormal localization of immature precursors (3, 6)

Microarray analysis is a broad-based profiling method that permits the concomitant comparison of gene expression profiles among different study groups revealing active networks of interrelated genes within subpopulations under study (7–11). In SLE these studies have shown interferon-inducible and granulopoiesis signatures that correlate with both disease severity (12) and activity (13). Several IFN related genes were found highly overexpressed in the peripheral blood and kidney glomeruli of lupus patients (14). These data have provided important insights into genetic pathways underlying SLE, as well as the effector cells and molecules involved in its pathogenesis.

In view of the central role of BM in regulating the immune response, we sought to explore bone marrow gene expression profiles using the microarrays in lupus patients and healthy individuals and compare it to the peripheral blood.

PATIENTS AND METHODS

Patients and controls

Twenty seven patients with SLE - followed by the Rheumatology Department of the University Hospital of Crete, a tertiary referral center- were studied following written informed consent. All bone marrow samples were obtained from patients that provided peripheral blood. All patients met the 1982 American College of Rheumatology revised criteria for the classification of SLE (15). In order to capture patients with higher disease activity, we used a SLE Disease Activity Index score cut-off of (SLEDAI) ≥8. Clinical and laboratory characteristics of the patients included in the study are summarized in Table 1. Seven patients had active proliferative and/ or membranous nephritis, while six had active neuropsychiatric lupus with manifestations such as psychosis, major depression, myelitis and polyneuropathy. Patients had not received steroids for at least 24 hours before blood and bone marrow was obtained for study. Controls enrolled in the study included seven healthy individuals and three osteoarthritis patients (5 males and 5 females, age ranging from 35–55 years) from the Department of Transfusion Medicine, Hematology Clinic and Rheumatology Clinic of University Hospital of Crete.

Table 1.

Clinical and demographic characteristics of SLE patients ^*

Sex, female/male	26 /1
Age, mean ± SD	47.28 ± 17.2
Active / Inactive	16 (59%) / 11 (41%)
SLE duration, mean ± SD	6.7 ±5.6
SLEDAI (mean ± SD)
Active SLE	13.86 ± 4.99
Inactive SLE	4.22 ± 1.56
Nephritis
Active SLE	7 / 16
Inactive SLE	0 /11
Total	7 / 27
CNS
Active SLE	6 / 16
Inactive SLE	0 /11
Total	6 / 27
Cytotoxic therapy
Active SLE	4 / 16
Inactive SLE	1 / 11
Total	5 / 27
Steroids
Active SLE	9 / 16
Inactive SLE	3 / 11
Total	12 / 27
Hydroxychloroquine
Active SLE	10 / 16
Inactive SLE	3 / 11
Total	13 / 27

^*Active SLE was defined as an SLE Disease Activity Index score of ≥8 and inactive as SLEDAI < 8. Controls included 10 healthy bone marrow donors (5 men and 5 women), 35–55 years of age.

Processing of peripheral blood mononuclear cells

(PBMCs) and bone marrow mononuclear cells (BMMCs) and RNA extraction

PBMCs and BMMCs from lupus patients or healthy volunteers were isolated by Ficoll-Histopaque (Sigma-Aldrich, St. Louis, MO) density-gradient centrifugation of heparinized venous blood and bone marrow aspirates immediately after peripheral blood and bone marrow draw. Cells were immediately placed in Trizol (Invitrogen, Carlsbad, CA, USA) and processed for RNA extraction using the RNeasy kit (Qiagen) according to the manufacturer’s instructions. RNA integrity was assessed using capillary gel electrophoresis (Agilent 2100 BioAnalyzer, Agilent Technologies, Santa Clara CA) to determine the ratio of 28s:18s rRNA in each sample. A threshold of 1.0 was used to define samples of sufficient quality and only samples above this limit were used for microarray studies.

cDNA synthesis

cDNA was synthesized using Omniscript reverse transcriptase (Qiagen) with direct incorporation of Cy3-dUTP from 2 ug of RNA. Labeled cDNA was purified using a Montage 96-well vacuum system (Millipore). The cDNA was added to hybridization buffer containing human CoT-1 DNA (0.5 mg/ml final concentration), yeast tRNA (0.2 mg/ml), and poly(dA)_40–60 (0.4 mg/ml).

Microarrays

A commercially available genome-scale oligonucleotide library containing gene-specific 70 mer oligonucleotides representing 21,329 human genes was used for microarray production. The library includes 16 replicate spots of 12 random negative controls that are designed to have no significant homology to known human DNA sequences (Qiagen Inc., Valencia, CA, USA). Oligonucleotides were spotted onto Corning UltraGAPS™ amino-silane coated slides, which were then rehydrated with water vapor and then, snap dried at 90°C. Oligonucleotide DNAs were covalently fixed to the surface of the glass using 300 mJ of ultraviolet radiation at a 254nm wavelength. Unbound, free amines on the glass surface were blocked for 15 min with moderate agitation in a solution of 143mM succinic anhydride dissolved in 1-methyl-2-pyrolidinone, 20mM sodium borate, pH 8.0. Slides were rinsed for 2 min in distilled water, immersed for 1 min in 95% ethanol and dried with a stream of nitrogen gas. Hybridization was performed in an automated liquid delivery, air-vortexed, hybridization station for 9 h at 58°C under an oil based cover slip (Ventana Medical Systems Inc., Tucson, AZ, USA). Microarrays were washed at a final stringency of 0.1X SSC. Microarrays were scanned using a simultaneous dual-colour, 48-slide scanner (Agilent Technologies). Background substracted luorescent intensity values were determined using Kaodarray software (Koada Technology, Stirling, UK)

Statistical analysis

a)Normalization: The R/Bioconductor Package “Affy” was used to perform quantile normalization to adjust the marginal distribution of each sample. b) Filtering: Genes that had an average background adjusted fluorescent intensity value > 50 across all arrays were retained in the analysis. Additionally, the variance across all genes was calculated. Genes that have a variance below the median variance are unlikely to be differentially expressed and are therefore removed from further analysis. c) Class Comparison: Genes that are differentially expressed between two classes were identified through an unpaired Student’s t test. A 10% false discovery rate p-value multiplicity adjustment was used. The false discovery rate is the proportion of the list of genes claimed to be differentially expressed that are false positives. Only statistically significant differentially expressed genes with greater than a 2- fold change in expression between groups were retained. d) SLEDAI Modeling: The SLEDAI score was modelled as a continuous variable according to the Generalized Linear Model (GLM) equation: Log₂ (Expression) = B1*SLEDAI + B2*Disease_State + B3*SLEDAI * Disease_State + Intercept, where Disease_State is a categorical indicator variable for Active or Inactive Disease, and B# are beta coefficients. A 10% false discovery rate was used on the SLEDAI term to determine statistical significnce. All analysis was performed in JMP Genomics 6.0.3 (Cary, NC). e) SLEDAI Modeling (PBMC): The SLEDAI score was modeled as a continuous variable according to the GLM equation: SLEDAI = B1*Gene X + B2 * Gene Y + … + BN * Gene N, where B1 through BN are beta coefficients and Gene X through Gene N are log base 2 normalized expression values. Multivariate models were created using only genes identified to be significantly associated to SLEDAI in the BM comparison f) Granulopoiesis Score: The score was created by taking genes known to be associated with granulopoiesis from the literature (LYZ, CD63, DEFA4, ELA2, S100A8, S100A12,S100P, CD24, NCF4) and adding their log expression values together. This resulted in a score that was used to correlate this plurality of genes vs SLEDAI through a GLM.

Real-time PCR validation

Nine genes were selected out of the set related to granulopoeisis for confirmation by PCR: S100A8, DEFA4, S100A12, ELA2, S100P, CD24, CD63, LYZ, NCF4. Eight patients were selected, four that represented patients in the "active" group and four that represented patients in the "inactive" group according to SLEDAI.

Reverse Transcription: cDNA was generated from 1.0 ug of total RNA per sample according to the OmniScript Reverse Transcriptase (Qiagen, Valencia, CA) manual, with the replacement of the RT primer mix with for 500ng anchored oligo dT(dT20VN). cDNA was purified with the Montage PCR Cleanup kit (Millipore, Billerica, MA) according to manufacturer's instructions. cDNA was diluted 1:20 in water and stored at −20° C.

Quantitative PCR: Gene-specific primers for the human genes S100A8, DEFA4, S100A12, ELA2, S100P, CD24, CD63, LYZ, NCF4 were designed with a melting temperature close to 60°C length of 19–25 bp for PCR products with a length of 110–150 bp, using Applied Biosystems Inc.(ABI) Primer Express 1.5 software. PCR was run with 2 ul cDNA template in 15ul reactions in triplicate on an ABI SDS 7700 using the ABI SYBR Green I Master Mix and gene specific primers at a concentration of 1uM each. The temperature profile consisted of an initial 95° C step for 10 minutes (for Taq activation), followed by 40 cycles of 95° C for 15 sec, 60° C for 1 min, and then a final melting curve analysis with a ramp from 60° C to 95° C over 20 min. Gene-specific amplification was confirmed by a single peak in the ABI Dissociation Curve software. No template controls were run for each primer pair and no RT controls were run for each sample to detect nonspecific amplification or primer dimers. Average Ct values for B-Actin (run in parallel reactions to the gene of interest) were used to normalize average Ct values of the gene of interest. These values were used to calculate the average group (active vs inactive) and the relative change in Ct was used to calculate fold change between the two groups.

RESULTS

Differentially expressed genes in the bone marrow of SLE patients vs controls

A total of 102 genes were found to have differential levels of expression between the SLE patients and the control subjects using unpaired student t-test. Of the 102 differentially expressed genes, 53 genes are involved in major networks including cell death, differentiation, cell signaling and cellular growth and proliferation (Table 2A). Data mining was performed to identify genes that were expressed in various subpopulations of BMMCs including B and T cells, monocytes and neutrophils. Of the 102 differentially expressed genes, 37 were up-regulated in the bone marrow of patients relative to controls including: TNFR17 (Tumor necrosis factor receptor superfamily, member 17) usually expressed in mature B lymphocytes and may be important for B cell development and autoimmune response. B cell involvement was highlighted by the presence of genes involved in the antigen presentation pathway such as, HLA-F (major histocompatibility complex, class I, F), and IGHG3 (Immunoglobulin heavy constant gamma 3). We found ITPR1, belonging to the family of Inositol 1,4,5-trisphosphate receptors (IP3Rs) which are expressed in most hematopoietic cells, including B cells, upregulated in the bone marrow of patients as well as a transcriptional co-activator BCL3 (B-cell CLL/lymphoma 3) reported to be upregulated in polyclonal plasmablastic cells(16). In mouse bone marrow, transgenic human BCL3 protein increases accumulation of mature B lymphocytes (17).

Table 2.

A. Selected up- and down-regulated genes in the bone marrow of SLE patients relative to controls. B. Selected upregulated genes in BMMCs relative to PBMCs of lupus patients

A.
Name	Description upregulated genes in BM	Genbank	Normalized ratio	Location	Family
BCL3	B-cell CLL/lymphoma 3	NM_005178	4,24	Nucleus	transcription regulator
ITPR1	inositol 1,4,5-triphosphate receptor, type 1	NM_002222	3,388	Cytoplasm	ion channel
RPL32	ribosomal protein L32	NM_000994	2,883	Cytoplasm	other
PHACTR1	phosphatase and actin regulator 1	AB051520	2,311	Cytoplasm	other
RPS2	ribosomal protein S2	NM_002952	2,258	Cytoplasm	other
RPS13	ribosomal protein S13	NM_001017	2,094	Cytoplasm	other
ABCG2	ATP-binding cassette, sub-family G (WHITE), member 2	NM_004827	2,062	Plasma Membrane	transporter
EPS15L1	epidermal growth factor receptor pathway substrate 15-like 1	AK023744	2,056	Plasma Membrane	other
ABLIM1	actin binding LIM protein 1	NM_002313	1,987	Cytoplasm	other
HLA-F	major histocompatibility complex, class I, F	NM_018950	1,904	Plasma Membrane	transmembrane receptor
NPTX1	neuronal pentraxin I	NM_002522	1,85	Extracellular Space	other
TNFRSF17	tumor necrosis factor receptor superfamily, member 17	NM_001192	1,837	Plasma Membrane	other
ADD3	adducin 3 (gamma)	U92992	1,781	Cytoplasm	other
TEAD2	TEA domain family member 2	BC007556	1,735	Nucleus	transcription regulator
BAIAP3	BAI1-associated protein 3	NM_003933	1,726	Unknown	other
PRKAA2	protein kinase, AMP-activated, alpha 2 catalytic subunit	NM_006252	1,641	Cytoplasm	kinase
GMFG	glia maturation factor, gamma	NM_004877	1,613	Cytoplasm	growth factor
BACE1	beta-site APP-cleaving enzyme 1	NM_012104	1,505	Cytoplasm	peptidase
TMEPAI	transmembrane, prostate androgen induced RNA	AF305616	1,481	Plasma Membrane	other
K-ALPHA-1	alpha tubulin	NM_006082	1,165	Cytoplasm	other
	downregulated in BM
PRKD1	protein kinase D1	NM_002742	4,724	Cytoplasm	kinase
CCR5	chemokine (C-C motif) receptor 5	NM_000579	3,877	Plasma Membrane	G-protein coupled receptor
CRHR1	corticotropin releasing hormone receptor 1	X72304	3,387	Plasma Membrane	G-protein coupled receptor
GJB3	gap junction protein, beta 3, 31kDa (connexin 31)	NM_024009	3,376	Plasma Membrane	transporter
GUCY2D	guanylate cyclase 2D, membrane (retina-specific)	NM_000180	3,32	Plasma Membrane	kinase
MPHOSPH1	M-phase phosphoprotein 1	NM_016195	3,023	Nucleus	enzyme
GAP43	growth associated protein 43	NM_002045	2,938	Plasma Membrane	other
MBP	myelin basic protein	NM_002385	2,901	Extracellular Space	other
PCLO	piccolo (presynaptic cytomatrix protein)	AB011131	2,895	Cytoplasm	transporter
MYH10	myosin, heavy polypeptide 10, non-muscle	AK026977	2,786	Cytoplasm	other
AEBP2	AE binding protein 2	BC015624	2,7	Nucleus	transcription regulator
PAX6	paired box gene 6 (aniridia, keratitis)	NM_001604	2,646	Nucleus	transcription regulator
MAG	myelin associated glycoprotein	NM_002361	2,541	Plasma Membrane	other
USP33	ubiquitin specific peptidase 33	AB029020	2,407	Cytoplasm	peptidase
ACE	angiotensin I converting enzyme (peptidyl-dipeptidase A) 1	NM_000789	2,385	Plasma Membrane	peptidase
CX3CR1	chemokine (C-X3-C motif) receptor 1	U20350	2,334	Plasma Membrane	G-protein coupled receptor
OTP	orthopedia homolog (Drosophila)	NM_032109	2,27	Nucleus	transcription regulator
SRGAP1	SLIT-ROBO Rho GTPase activating protein 1	AB037725	2,227	Unknown	other
RGS11	regulator of G-protein signalling 11	NM_003834	2,176	Plasma Membrane	enzyme
HOXB3	homeobox B3	U59298	2,171	Nucleus	transcription regulator
SLAMF6	SLAM family member 6	NM_052931	2,164	Plasma Membrane	transmembrane receptor
SSX2IP	synovial sarcoma, X breakpoint 2 interacting protein	NM_014021	2,076	Unknown	other
CDH2	cadherin 2, type 1, N-cadherin (neuronal)	NM_001792	2,064	Plasma Membrane	other
DUSP4	dual specificity phosphatase 4	NM_057158	2,015	Nucleus	phosphatase
GRB2	growth factor receptor-bound protein 2	NM_002086	1,938	Plasma Membrane	other
EIF4EBP2	eukaryotic translation initiation factor 4E binding protein 2	AK057643	1,82	Cytoplasm	other
CD276	CD276 molecule	NM_025240	1,813	Plasma Membrane	other
PRKCG	protein kinase C, gamma	NM_002739	1,768	Cytoplasm	kinase
C10ORF10	chromosome 10 open reading frame 10	NM_007021	1,662	Unknown	other
CTNNAL1	catenin (cadherin-associated protein), alpha-like 1	NM_003798	1,65	Plasma Membrane	other
MAGI1	membrane associated guanylate kinase, WW and PDZ domain containing 1	AK023358	1,635	Plasma Membrane	kinase
CDKN3	cyclin-dependent kinase inhibitor 3 (CDK2-associated dual specificity phosphatase)	NM_005192	1,426	Nucleus	phosphatase
SNX2	sorting nexin 2	NM_003100	1,306	Cytoplasm	transporter

B.
Symbol	Short Description	Normalized ratio	Location	Family
CTSG	Cathepsin G	6,65	cytoplasm	peptidase
HBD	Hemoglobin, delta	5,84	cytoplasm	transporter
ELA2	Elastase 2, neutrophil	4,49	extracellular space	peptidase
MPO	Myeloperoxidase	3,3	cytoplasm	enzyme
S100A9	S100 calcium binding protein A9 (calgranulin B)	2,83	cytoplasm	other
NR2C2	Nuclear receptor subfamily 2, group C, member 2	2,58	nucleus	ligand-dependent nuclear receptor
PRDX2	Peroxiredoxin 2	2,41	cytoplasm	enzyme
S100A12	S100 calcium binding protein A12 (calgranulin C)	2,41	cytoplasm	other
PPIB	Peptidylprolyl isomerase B (cyclophilin B)	2,34	cytoplasm	enzyme
ITPR1	Inositol 1,4,5-triphosphate receptor, type 1	2,23	cytoplasm	ion channel
STAG3	Stromal antigen 3	2,2	nucleus	other
CRMP1	Collapsin response mediator protein 1	2,18	cytoplasm	enzyme
HP	Haptoglobin	2,15	extracellular space	peptidase
SDHA	Succinate dehydrogenase complex, subunit A, flavoprotein (Fp)	2,14	cytoplasm	enzyme
PSCD1	Pleckstrin homology, Sec7 and coiled/coil domains 1(cytohesin 1)	2,02	cytoplasm	other
K-ALPHA-1	Tubulin, alpha, ubiquitous	2,01	cytoplasm	other
PGLYRP	Peptidoglycan recognition protein	1,96	plasma membrane	transmembrane receptor
S100P	S100 calcium binding protein P	1,95	cytoplasm	other
LTF	Lactotransferrin	1,91	extracellular space	peptidase
ATP7A	ATPase, Cu++ transporting, alpha polypeptide (Menkes syndrome)	1,86	plasma membrane	transporter
LCN2	Lipocalin 2 (oncogene 24p3)	1,78	extracellular space	transporter
SNCA	Synuclein, alpha (non A4 component of amyloid precursor)	1,77	cytoplasm	other
DUSP4	Dual specificity phosphatase 4	1,74	nucleus	phosphatase
CSPG4	Chondroitin sulfate proteoglycan 4 (melanoma-associated)	1,7	plasma membrane	other
PHKA1	Phosphorylase kinase, alpha 1 (muscle)	1,64	cytoplasm	kinase
DLC1	Deleted in liver cancer 1	1,56	cytoplasm	other
UQCRC2	Ubiquinol-cytochrome c reductase core protein II	1,54	cytoplasm	enzyme
CD24	CD24 antigen (small cell lung carcinoma cluster 4 antigen)	1,41	plasma membrane	other
CDKN3	Cyclin-dependent kinase inhibitor 3 (CDK2-associated dual specificity phosphatase)	1,34	nucleus	phosphatase

Sixty five (65 genes) were expressed at lower levels in patients than controls including the chemokine receptors CX3CR1 and CCR5, the latter is normally expressed by T cells and macrophages. Also downregulated were CDH2 (Cadherin 2), CTNNAL1 (Catenin,cadherin-associated protein), CDKN (Cyclin-dependent kinase inhibitor 3) and KALI (Activating NK receptor) whose protein is expressed on Natural Killer, T and B lymphocytes and lung. CD276 (B7-H3) a costimulatory molecule for T cell activation and IFN-gamma production was also found downregulated in the bone marrow of patients.

Bone marrow genes associated with SLE disease activity

By the use of multiple regression analysis, as outlined in Materials and Methods, seven genes were statistically associated with SLEDAI. KIAA1674 (GenBank AB051461) (r² =0.82), NY-REN-25 antigen, an ankyrin repeat domain (GenBank AF155103) (r²=0.81), cDNA FLJ32586 fis (GenBank AK057148) (r² = 0.79), the hypothetical protein FLJ10254 (GenBank NM_018041) (r² = 0.83) and the coiled coil domain CCDC91 (GenBank NM_018318) (r² = 0.84), CENPH (Centromere protein H, GenBank NM_022909) (r² =0.79) and EBI3 (Epstein-Barr virus induced gene 3, GenBank NM_005755)) (r² = 0.77) a subunit of IL-27 which may play an important role in initiation of Th1 responses. The expression of these genes was highly correlated with one another producing Pearson correlation coefficients of 0.89 to 0.97. Due to co-linearity concerns we were not able to combine these terms into a single multivariate model.

In order to test if these 7 SLEDAI-associated bone marrow expressed genes were also associated with SLEDAI in the peripheral blood, we re-analyzed the data and used the genes selected in the BM and refitted the terms to the PBMC data to predict SLEDAI. Only 2 genes, NY-REN-25 antigen, an ankyrin repeat domain (GenBank AF155103) and coiled coil domain CCDC91 (GenBank NM_018318) associated with SLEDAI in the active PBMC patients (r²=0.37, p=0.0108) (Figure 1).

Figure 1. SLEDAI-associated bone marrow expressed genes were also associated with SLEDAI in the periphery.

Two genes, NY-REN-25 antigen, an ankyrin repeat domain (GenBank AF155103) and coiled coil domain CCDC91 (GenBank NM_018318) associated with SLEDAI in active PBMC patients thorugh a GLM (r²=0.37, P=0.0108). The Y Axis represents the predicted SLEDAI from our model of 2 genes and the X axis is the observed SLEDAI from the patient’s medical records.

Differentially expressed genes in BMMCs relative to PBMCs of SLE patients

We next compared bone marrow derived mononuclear cells with peripheral blood mononuclear cells in the lupus cohort. Eighty eight (88) genes were differentially expressed, 41 out of 88 were up-regulated in the bone marrow of lupus patients relative to the peripheral blood (Table 2B) while the remaining 47 were up-regulated in the peripheral blood. Among the lupus bone marrow up-regulated genes, the highest overexpression was found in granulopoiesis- related genes. These genes include major components of neutrophils such as myeloperoxidase MPO, ELA2 (elastase 2) responsible for hydrolyzing proteins within granules, CTSG (cathepsin G), DEFA4 (defensin), LTF (lactotransferrin) and CD24 found on mature granulocytes. Three small abundant proteins found in human neutrophil cytosol S100A9, S100A12 and S100P were upregulated in the bone marrow of SLE patients. Finally, in the peripheral blood of lupus patients we identified a number of chemokines such as CCR5, CXCL3L1, CXCL2 and CXCL3 that are overexpressed relative to the bone marrow. These genes participate in processes such as chemotaxis and migration of leukocytes. Comparison of bone marrow versus peripheral blood in the control cohort revealed that most of the neutrophil related genes found in the previous comparison, were also overexpressed in the bone marrow of control subjects relative to the peripheral blood with minor differences in gene expression level (data not shown). For example, cathepsin CTSG was overexpressed by 7.3-fold in bone marrow of controls relative to peripheral blood and by 6.6-fold in the bone marrow of lupus patients when compared to the peripheral blood. Thus many of these genes are tissue associated differences rather than disease-associated differences.

Differentially expressed genes in the peripheral blood of SLE patients vs controls

Among the SLE upregulated genes in the PBMCs two IFN-inducible genes were overexpressed in lupus patients: IL6R whose expression is regulated by IFNα and PRKCG (protein kinase C, gamma) that is involved in antiviral response of IFNγ and its signaling (data not shown). In total 35 genes were upregulated in SLE peripheral blood including a number of regulatory molecules such as: TCF7 (transcription factor 7, T-cell specific), CYC1 (cytochrome c-1), UBTF (upstream binding transcription factor), HDAC10 (histone deacetylase 10) involved in the acetylation status of histone tails, a ubiquitin-conjugating enzyme UBE2D3 and U2AF1 ( U2(RNU2) small nuclear RNA auxillary factor 1) belonging to the splicing factor SR family of genes. The expression of 18 genes was downregulated in the PBMCs of SLE patients compared to controls including: BACE (beta-site APP-cleaving enzyme), HOXD13 (homeo box D13), K-ALPHA-1 (tubulin), PHKA1 (phosphorylase kinase,alpha) and PRKAA2 (protein kinase AMP-activated).

Hierarchical clustering reveals patient subgroups in the bone marrow

In order to group individuals with similar expression profiles in their peripheral blood mononuclear cells (PBMCs) and bone marrow mononuclear cells (BMMCs), we used unsupervised hierarchical clustering of the top 25% of expressed genes (n= 2652) in each group separately. Lupus samples derived from peripheral blood did not form distinct, well-characterized clusters but scattered across the graph regardless of disease activity (Figure 2A). In contrast, hierarchical clustering in the bone marrow demonstrated that lupus patients fell into two groups that displayed different pattern of expression for these 2652 genes (Figure 2B) and the controls formed another distinct cluster. Of interest, these two groups were primarily active patients in one cluster and inactive patients in the other with 20% and 30% misclassification respectively (p=0.07, Fisher’s Exact test). These clusters were not associated with disease manifestations or drug treatment. Taken together, these findings suggest that bone marrow in SLE provides supplemental information to that obtained from peripheral blood in the context of disease activity.

Figure 2. Hierarchical clustering of the top 25% (2652) of highly expressed genes in the peripheral blood and bone marrow of SLE patients.

Each column represents a gene and each row shows the expression for the top 25% genes expressed by each individual. (A). Peripheral blood samples from SLE patients scattered across the graph regardless of disease activity. (B). Hierarchical clustering distinguish SLE active patients from inactive in the bone marrow. The first big cluster corresponds to the active patients and the second one to the inactive

Granulopoiesis and apoptosis signature in the bone marrow of active patients

To explore any gene signatures that characterize the two patient subgroups in the bone marrow, we further analyzed on a whole-genome scale the differentially expressed genes between patients in these two clusters. We found 245 differentially expressed genes that were up-regulated in the active patient cluster as compared to the inactive (Table 3). No genes were identified as significantly repressed in the active relative to inactive group. Genes involved in antigen presentation such as HLA-A, HLA-C and CD74 were expressed in higher levels in the bone marrow of active SLE patients. Among the up-regulated genes we noticed granulopoiesis-related genes such as ELA2 (elastase) which is usually transcribed within the earliest granulocytes, LYZ (lysozyme) a component of azurophil and specific granules which was overexpressed in active patients, DEFA (defensin), CD63 a marker of azurophil granules and CD24 expressed on mature granulocytes. Three family members of S100 calcium-binding proteins, S100A6, S100A8 and S100P were also upregulated in the active group. Among these three genes, S100A8 a leukocyte chemoattractant protein was overexpressed 10-fold compared to the inactive group. All genes that were upregulated on the microarrays were also upregulated when analyzed by quantitative real-time PCR. The expression values of these granulopoiesis genes correlated highly with each other. Additional up-regulated genes included genes involved in processes such as cell death, immune response and cellular movement. The genes implicated in the cell death of leukocytes include ANXA1, CD24, CST3, CXCR4, ELA2, FOS, FOXO3A, IER3, ITGB2, LCN2, LYN, and PCBP2. We also identified the genes involved in the apoptosis of granulocytes which include ANXA1 (annexin 1), chemokine CXCR4, FOXO3A, ITGB2 (integrin) and LYN.

Table 3.

Selected up-regulated genes in the bone marrow of active patients relative to inactive patients after unsupervised hierarchical clustering. Shown highlighted are selected genes involved in granulopoiesis, apoptosis and antigen presentation.

Name	Description	Genbank	Normalized ratio	Location	Family
HSPB1	heat shock 27kDa protein 1	NM_001540	19,969	Cytoplasm	other
RAP1B	RAP1B, member of RAS oncogene family	NM_015646	17,784	Cytoplasm	enzyme
ITGB2	integrin, beta 2 (complement component 3 receptor 3 and 4 subunit)	NM_000211	16,509	Plasma Membrane	other
H3F3A	H3 histone, family 3A	M11354	14,347	Nucleus	other
G0S2	G0/G1switch 2	NM_015714	14,192	Unknown	other
MRLC2	myosin regulatory light chain MRLC2	NM_033546	14,025	Cytoplasm	other
MRCL3	myosin regulatory light chain MRCL3	NM_006471	11,738	Unknown	other
S100A8	S100 calcium binding protein A8 (calgranulin A)	NM_002964	10,686	Cytoplasm	other
GRN	granulin	NM_002087	10,351	Extracellular Space	growth factor
LSP1	lymphocyte-specific protein 1	AK056576	8,902	Cytoplasm	other
PRG1 (includes EG:55	proteoglycan 1, secretory granule	NM_002727	8,773	Extracellular Space	other
CD74	CD74 molecule, major histocompatibility complex, class II invariant chain	NM_004355	7,637	Plasma Membrane	transmembrane receptor
DDX5	DEAD (Asp-Glu-Ala-Asp) box polypeptide 5	NM_004396	7,573	Nucleus	enzyme
TMSB10	thymosin, beta 10	NM_021103	6,974	Cytoplasm	other
AMPH	amphiphysin (Stiff-Man syndrome with breast cancer 128kDa autoantigen)	NM_001635	6,745	Plasma Membrane	other
HNRPA1	heterogeneous nuclear ribonucleoprotein A1	NM_031157	6,67	Nucleus	other
TPT1	tumor protein, translationally-controlled 1	NM_003295	6,586	Cytoplasm	other
S100A6	S100 calcium binding protein A6 (calcyclin)	NM_014624	6,251	Cytoplasm	other
ATP5G2	ATP synthase, H+ transporting, mitochondrial F0 complex, subunit C2 (subunit	NM_005176	6,065	Cytoplasm	transporter
UBE2D3	ubiquitin-conjugating enzyme E2D 3 (UBC4/5 homolog, yeast)	NM_003340	5,866	Cytoplasm	enzyme
TMC6	transmembrane channel-like 6	NM_007267	5,324	Unknown	transporter
SF3A3	splicing factor 3a, subunit 3, 60kDa	NM_006802	5,26	Nucleus	other
SFRS2	splicing factor, arginine/serine-rich 2	NM_003016	5,252	Nucleus	other
PAPOLA	poly(A) polymerase alpha	NM_032632	5,208	Nucleus	enzyme
ELA2	elastase 2, neutrophil	NM_001972	5,171	Extracellular Space	peptidase
PFN1	profilin 1	NM_005022	5,146	Cytoplasm	other
PMPCB	peptidase (mitochondrial processing) beta	NM_004279	5,119	Cytoplasm	peptidase
SRP14	signal recognition particle 14kDa (homologous Alu RNA binding protein)	NM_003134	5,034	Cytoplasm	other
COTL1	coactosin-like 1 (Dictyostelium)	BC010039	4,898	Cytoplasm	other
ERAF	erythroid associated factor	NM_016633	4,881	Cytoplasm	other
CST7	cystatin F (leukocystatin)	NM_003650	4,701	Extracellular Space	other
HLA-A	major histocompatibility complex, class I, A	NM_002116	4,612	Plasma Membrane	transmembrane receptor
PCBP2	poly(rC) binding protein 2	AK023529	4,557	Nucleus	other
PKM2	pyruvate kinase, muscle	NM_002654	4,47	Cytoplasm	kinase
ATP7A	ATPase, Cu++ transporting, alpha polypeptide (Menkes syndrome)	NM_000052	4,343	Plasma Membrane	transporter
ANXA1	annexin A1	NM_000700	4,273	Plasma Membrane	other
TBCA	tubulin-specific chaperone a	NM_004607	4,088	Cytoplasm	other
ALOX5	arachidonate 5-lipoxygenase	NM_000698	4,046	Cytoplasm	enzyme
HP	haptoglobin	AK055872	4,03	Extracellular Space	peptidase
LCN2	lipocalin 2 (oncogene 24p3)	NM_005564	4,029	Extracellular Space	transporter
C6ORF115	chromosome 6 open reading frame 115	AF116682	4,004	Unknown	other
ARD1A	ARD1 homolog A, N-acetyltransferase (S. cerevisiae)	NM_003491	3,983	Nucleus	enzyme
TMSB4X	thymosin, beta 4, X-linked	AK055976	3,942	Cytoplasm	other
PPAP2B	phosphatidic acid phosphatase type 2B	NM_003713	3,872	Plasma Membrane	phosphatase
CD63	CD63 molecule	NM_001780	3,861	Plasma Membrane	other
YWHAB	tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, protein, beta p	NM_003404	3,841	Cytoplasm	other
EXOSC6	exosome component 6	NM_058219	3,784	Nucleus	other
HLA-C	major histocompatibility complex, class I, C	M12679	3,754	Plasma Membrane	transmembrane receptor
SNCA	synuclein, alpha (non A4 component of amyloid precursor)	NM_000345	3,689	Cytoplasm	other
PDHA1	pyruvate dehydrogenase (lipoamide) alpha 1	NM_000284	3,652	Cytoplasm	enzyme
HSPE1	heat shock 10kDa protein 1 (chaperonin 10)	NM_002157	3,605	Cytoplasm	other
SNRPD2	small nuclear ribonucleoprotein D2 polypeptide 16.5kDa	NM_004597	3,57	Nucleus	other
ACTG1	actin, gamma 1	NM_001614	3,52	Cytoplasm	other
SYNGR2	synaptogyrin 2	NM_004710	3,493	Plasma Membrane	other
NDUFA13	NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 13	NM_015965	3,449	Cytoplasm	enzyme
CXCR4	chemokine (C-X-C motif) receptor 4	NM_003467	3,411	Plasma Membrane	G-protein coupled receptor
AURKAIP1	aurora kinase A interacting protein 1	NM_017900	3,407	Nucleus	other
CYC1	cytochrome c-1	NM_001916	3,386	Cytoplasm	enzyme
MAFK	v-maf musculoaponeurotic fibrosarcoma oncogene homolog K (avian)	AK056767	3,38	Nucleus	transcription regulator
BLOC1S1	biogenesis of lysosome-related organelles complex-1, subunit 1	NM_001487	3,375	Cytoplasm	other
BPI	bactericidal/permeability-increasing protein	NM_001725	3,368	Plasma Membrane	other
FOS	v-fos FBJ murine osteosarcoma viral oncogene homolog	NM_005252	3,295	Nucleus	transcription regulator
S100P	S100 calcium binding protein P	NM_005980	3,204	Cytoplasm	other
PSMB1	proteasome (prosome, macropain) subunit, beta type, 1	AK023290	3,174	Cytoplasm	peptidase
APOB	apolipoprotein B (including Ag(x) antigen)	NM_000384	3,131	Extracellular Space	transporter
GTF3C5	general transcription factor IIIC, polypeptide 5, 63kDa	NM_012087	3,117	Nucleus	transcription regulator
SF3A2	splicing factor 3a, subunit 2, 66kDa	NM_007165	3,11	Nucleus	other
MAP3K11	mitogen-activated protein kinase kinase kinase 11	NM_002419	3,078	Cytoplasm	kinase
NRG1	neuregulin 1	NM_013957	3,069	Extracellular Space	growth factor
CST3	cystatin C (amyloid angiopathy and cerebral hemorrhage)	NM_000099	3,046	Extracellular Space	other
PSCD1	pleckstrin homology, Sec7 and coiled-coil domains 1(cytohesin 1)	NM_004762	3,038	Cytoplasm	other
COPE	coatomer protein complex, subunit epsilon	NM_007263	3,014	Cytoplasm	transporter
TALDO1	transaldolase 1	NM_006755	3,003	Cytoplasm	enzyme
YBX1	Y box binding protein 1	NM_004559	3	Nucleus	transcription regulator
RNASE3	ribonuclease, RNase A family, 3 (eosinophil cationic protein)	NM_002935	2,889	Extracellular Space	enzyme
SERPINB1	serpin peptidase inhibitor, clade B (ovalbumin), member 1	NM_030666	2,886	Cytoplasm	other
SAT	spermidine/spermine N1-acetyltransferase 1	NM_002970	2,881	Cytoplasm	enzyme
ARHGAP4	Rho GTPase activating protein 4	NM_001666	2,863	Cytoplasm	other
LAPTM5	lysosomal associated multispanning membrane protein 5	NM_006762	2,861	Plasma Membrane	other
ABCA7	ATP-binding cassette, sub-family A (ABC1), member 7	NM_019112	2,859	Plasma Membrane	transporter
CALCR	calcitonin receptor	NM_001742	2,844	Plasma Membrane	protein coupled receptor
HNRPM	heterogeneous nuclear ribonucleoprotein M	AK024911	2,794	Plasma Membrane	transmembrane receptor
CD24	CD24 molecule	NM_013230	2,77	Plasma Membrane	other
TM7SF3	transmembrane 7 superfamily member 3	NM_016551	2,734	Plasma Membrane	other
GYPC	glycophorin C (Gerbich blood group)	NM_002101	2,705	Plasma Membrane	other
ATP6V1F	ATPase, H+ transporting, lysosomal 14kDa, V1 subunit F	NM_004231	2,702	Cytoplasm	transporter
HMGB2	high-mobility group box 2	NM_002129	2,661	Nucleus	transcription regulator
EIF4B	eukaryotic translation initiation factor 4B	NM_001417	2,569	Cytoplasm	translation regulator
NDUFA1	NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 1, 7.5kDa	NM_004541	2,568	Cytoplasm	enzyme
WBSCR1	Williams-Beuren syndrome chromosome region 1	NM_022170	2,562	Cytoplasm	translation regulator
NDUFS5	NADH dehydrogenase (ubiquinone) Fe-S protein 5, 15kDa (NADH-coenzyme Q red	NM_004552	2,498	Cytoplasm	enzyme
BCCIP	BRCA2 and CDKN1A interacting protein	NM_078469	2,459	Nucleus	other
SNIP1	Smad nuclear interacting protein 1	NM_024700	2,456	Nucleus	other
UBA52	ubiquitin A-52 residue ribosomal protein fusion product 1	NM_003333	2,376	Cytoplasm	transcription regulator
USP7	ubiquitin specific peptidase 7 (herpes virus-associated)	NM_003470	2,374	Nucleus	peptidase
NR6A1	nuclear receptor subfamily 6, group A, member 1	NM_033334	2,36	Nucleus	ligand-dependent nuclear receptor
ZNF9	CCHC-type zinc finger, nucleic acid binding protein	NM_003418	2,334	Nucleus	transcription regulator
GPSM3	G-protein signalling modulator 3 (AGS3-like, C. elegans)	NM_022107	2,332	Unknown	other
IDI1	isopentenyl-diphosphate delta isomerase 1	NM_004508	2,324	Cytoplasm	enzyme
KLF6	Kruppel-like factor 6	NM_001300	2,311	Nucleus	transcription regulator
UQCRC2	ubiquinol-cytochrome c reductase core protein II	NM_003366	2,308	Cytoplasm	enzyme
RAN	RAN, member RAS oncogene family	NM_006325	2,289	Nucleus	enzyme
SUB1	SUB1 homolog (S. cerevisiae)	NM_006713	2,286	Nucleus	transcription regulator
TGOLN2	trans-golgi network protein 2	AK025557	2,282	Cytoplasm	other
BZRP	translocator protein (18kDa)	NM_000714	2,273	Cytoplasm	transmembrane receptor
SCPEP1	serine carboxypeptidase 1	NM_021626	2,249	Cytoplasm	peptidase
FOXO3A	forkhead box O3A	AK024103	2,241	Nucleus	transcription regulator
BLVRB	biliverdin reductase B (flavin reductase (NADPH))	NM_000713	2,233	Cytoplasm	enzyme
C1D	nuclear DNA-binding protein	NM_006333	2,212	Nucleus	transcription regulator
TAF9	TAF9 RNA polymerase II, TATA box binding protein (TBP)-associated factor, 32kDa	NM_003187	2,177	Nucleus	transcription regulator
RAD23A	RAD23 homolog A (S. cerevisiae)	NM_005053	2,175	Nucleus	other
UGCG	UDP-glucose ceramide glucosyltransferase	AJ420423	2,165	Cytoplasm	enzyme
NFKBIZ	nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, zeta	NM_031419	2,157	Nucleus	transcription regulator
ATP6AP2	ATPase, H+ transporting, lysosomal accessory protein 2	NM_005765	2,155	Cytoplasm	transporter
CTSW	cathepsin W (lymphopain)	NM_001335	2,153	Cytoplasm	peptidase
CCNI	cyclin I	NM_006835	2,138	Unknown	other
UBE1	ubiquitin-activating enzyme E1 (A1S9T and BN75 temperature sensitivity complem	NM_003334	2,136	Cytoplasm	enzyme
ARPC3	actin related protein 2/3 complex, subunit 3, 21kDa	NM_005719	2,13	Cytoplasm	other
DEK	DEK oncogene (DNA binding)	NM_003472	2,129	Nucleus	transcription regulator
SSBP3	single stranded DNA binding protein 3	NM_018070	2,107	Nucleus	transcription regulator
CAMLG	calcium modulating ligand	NM_001745	2,102	Cytoplasm	other
CHAF1A	chromatin assembly factor 1, subunit A (p150)	NM_005483	2,102	Nucleus	other
F11R	F11 receptor	AF172398	2,093	Plasma Membrane	other
NCF4	neutrophil cytosolic factor 4, 40kDa	NM_013416	2,092	Cytoplasm	enzyme
COX7A2	cytochrome c oxidase subunit VIIa polypeptide 2 (liver)	NM_001865	2,089	Cytoplasm	enzyme
H1FX	H1 histone family, member X	NM_006026	2,083	Nucleus	other
FOSL1	FOS-like antigen 1	NM_005438	2,077	Nucleus	transcription regulator
ACTB	actin, beta	NM_001101	2,075	Cytoplasm	other
PNN	pinin, desmosome associated protein	NM_002687	2,068	Plasma Membrane	other
LYN	v-yes-1 Yamaguchi sarcoma viral related oncogene homolog	NM_002350	2,054	Cytoplasm	kinase
TNC	tenascin C (hexabrachion)	AK024586	2,049	Extracellular Space	other
SUMO2	SMT3 suppressor of mif two 3 homolog 2 (S. cerevisiae)	NM_006937	2,045	Nucleus	other
CKAP4	cytoskeleton-associated protein 4	NM_006825	2,043	Cytoplasm	other
SLC44A2	solute carrier family 44, member 2	AK027519	2,036	Extracellular Space	other

Granulopoiesis signature in the bone marrow as a marker for SLE activity

Nine genes selected a priori were used to investigate associations between a granulopoeisis signature and SLEDAI. To facilitate these comparisons we created a granulopoeisis “score” for both PBMCs and BMMCs, as described above in Materials and Methods and found a correlation between granulopoiesis score and SLEDAI (r = 0.33) in the periphery of the SLE patient subset that had also provided bone marrow (Figure 3A). Linear regression analysis showed that granulopoiesis signature significantly correlated with SLEDAI in the bone marrow (r = 0.55, p = 0.013), (Figure 3B) and the granulopoiesis score from bone marrow active group of patients was higher versus the inactive (p =0.004), (Figure 3C)

Figure 3. Granulopoiesis signature in the bone marrow as a marker for SLE activity.

Nine genes selected a priori were used to investigate associations between a granulopoeisis signature and SLEDAI. These genes resulted in a statistically significant model of SLEDAI. (A) A numerical score was calculated by using the normalized expression levels of 9 granulopoiesis-related genes that comprise the granulopoiesis signature. Linear regression analysis demonstrates a correlation between granulopoiesis score and SLEDAI (r = 0.33) in the peripheral blood of patient subset that have also provided BM. (B) Granulopoiesis signature of the patients analyzed on Panel A, correlates stronger with SLEDAI in the bone marrow (r = 0.55, p = 0.013). C. The granulopoiesis score from bone marrow was higher in the active group of SLE patients vs the inactive (p =0.004)

DISCUSSION

This study sought to shed additional light on the pathogenesis of SLE by analyzing gene expression in the bone marrow, an organ with a central role not only in hemopoiesis but also in the immune response. Herein, we report that bone marrow analysis discriminates active from inactive lupus, and displays apoptosis and granulopoiesis signatures

Our microarray analysis has several strong points. First, we only selected the statistically significant genes with a change in expression of at least 2-fold when comparing the means of the two groups; the increased stringency increases the specificity of the results. Second, hierarchical clustering revealed two overlapping groups: active and inactive patients and we determined the differentially expressed genes between these two clusters. By doing so, the analysis was not biased by any arbitrary clustering. Third, our clinical data on organ involvement were obtained at the time on the blood drawn and were not based on historical data on involvement of a particular organ at some time in the course of the disease. Lastly, in addition to the peripheral blood, our analysis included the bone marrow an important organ in the biology of lymphocytes and neutrophils. Using this analysis, we were able to duplicate-only in part- data suggesting an interferon signature in SLE by finding up-regulation of only two IFN regulated genes: IL6R whose expression is regulated by IFNα and PRKCG which is involved in IFNγ signaling. This may be due to differences in the analysis between our study and those of Baechler et al (12) and Bennet et al (13) or-more likely-to genetic differences in the populations studied ( North American vs Mediterranean). The later is supported by data both in murine models (18) and in humans whereby the relative contribution of IFN in the pathogenesis of the disease may depend upon the genetic background (19).

Ficol-Histopaque density gradient preparations of peripheral blood mononuclear cells of patients with lupus contain high numbers of low buoyant density activated neutrophils (20). We found increased expression of early neutrophil genes; most of them encode products of immature granulocytes and their expression is regulated during myeloid cell differentiation. These proteins include components of the neutrophil granules (such as defensin A4, cathepsin, myeloperoxidase, lactoferrin and elastase 2) that may be released by neutrophils and initiate the production of autoantibodies directed against constituents of neutrophil cytoplasm (ANCA) described in patients with systemic vasculitis and autoimmune diseases such as lupus (21–31).

It is of particular interest that we found a significant increase in the levels of β2-integrin and other genes involved in the integrin signaling pathway such as MAP3K11, ACTB, ACTG1, ARPC3, MRCL2 and MRCL3 in the bone marrow of active patients relative to inactive. The integrin beta 2, LFA-1 (leukocyte function-associated antigen-1) and its ligand ICAM-1, has a key role in tissue injury in lupus (32) and other inflammatory diseases (33–36). On the other hand, neutrophils are important effector cells in a variety of acute and chronic inflammatory states including lupus (37). Our data corroborate and expand those from Bennet et al (13). Although we did not observe the granulopoiesis signature in the peripheral blood, we found it in the BM when comparing active vs inactive and BM vs peripheral blood. Common genes overexpressed in our and their study include myeloperoxidase, elastase, cathepsin, CD24, S100P and defensin. These results suggest that the granulopoiesis signatures reported by others in SLE peripheral blood may in fact originate from bone marrow cells and may persist or expand in the blood of certain SLE patients.

In our study we also identified genes involved in the apoptosis of granulocytes in the bone marrow of active SLE, such as FOXO3A which has been described as one of the genes induced after phagocytosis of pathogens (38) as well as annexin 1 reported to have a proapoptotic role in human neutrophils (39) and CXCR4 usually expressed on senescent neutrophils. Accelerated apoptosis of lymphocytes in the peripheral blood of lupus patients and impaired clearance of apoptotic cells due to the decreased phagocytic ability of macrophages, monocytes and neutrophils has a pathogenic role in SLE (40–47). This has been speculated to be the result of unidentified serum factors (48). In addition to the necrotic changes in the bone marrow reported in SLE by Voulgarelis et al. (3) apoptotic bodies were also recently observed in the bone marrow of 8 of 10 SLE patients, many of whom had cytopenias(49). Both alive and apoptotic neutrophils have been implicated in tissue injury in patients with Wegener granulomatosis (50). We speculate that neutrophil apoptosis is probably a result of the excessive activation of polymorphonuclear cells.

In our analysis, we found seven genes correlating with SLEDAI in the bone marrow. Although these genes are highly correlated, they share no known functional relationships. Out of these 7 genes associated with SLEDAI in the bone marrow, only two genes, NY-REN-25 antigen and the coiled-coil domain CCDC91 associated with SLEDAI in active PBMC patients. For both genes there is considerable variation between predicted and actual SLEDAI probably because these genes were selected within the bone marrow comparison.

In summary, these data support the use of microarray analysis to uncover novel immunopathologic pathways in the disease and have shown that the bone marrow, distinguishes active from inactive lupus patients. These data provide additional credence to the role of bone marrow and neutrophils in the pathogenesis of the disease and suggest additional pathways for potential therapeutic modulation targeted at the effector cells to minimize tissue injury.