Cutting Edge: Intracellular IFN-β and Distinct Type I IFN Expression Patterns in Circulating Systemic Lupus Erythematosus B Cells

Activation of the type I IFNs, consisting of 13 IFN-α subtypes and one high-affinity IFN-β subtype, is highly associated with the development of systemic lupus erythematosus (SLE) as well as clinical disease manifestations (1). Type I IFNs can be produced by most cell types, although their activity in SLE is most often measured indirectly using the presence of specific type I IFN–inducible transcripts, termed the type I IFN signature (1, 2). Previous studies have identified unique type I IFN signatures among different immune cell populations (3–7), but cell-specific expression patterns and roles of distinct type I IFNs in SLE, especially high-affinity IFN-β, remain elusive (8), largely because of their low levels of transcription and circulation (9).

In autoimmune mice, IFNαR1 deficiency ameliorates germinal center and autoantibody development (10, 11). Although a previous study in NZB mice reported no difference in anti-chromatin Abs or renal disease in Ifnb^–⁄– mice (12), these findings have been challenged by reports that IFN-β is elevated and dysregulated in SLE (13, 14) as well as the identification of distinct IFN signatures not restricted to IFN-α (15). Serum detection of IFN-β was recently associated with disease flares, particularly in African American (AA) patients, a population with increased disease prevalence, severity, and robust type I IFN dysregulation (14, 16). Autocrine IFN-β signaling has been identified as a mechanism of type I IFN dysregulation in SLE mesenchymal stem cells (13), and B cells have also been shown to produce type I IFNs in SLE and other diseases (9, 17, 18). In this study, we examined expression patterns of type I IFNs and their target type I IFN–stimulated genes (ISGs) in circulating B cells and identified a potential role for B cell–associated IFN-β in SLE.

All SLE subjects met the American College of Rheumatology 1997 revised criteria for SLE (19) and were recruited from the University of Alabama at Birmingham (UAB) Lupus Clinic. PBMCs were isolated by density gradient centrifugation (Lymphoprep/SepMate; STEMCELL Technologies). Clinical data were determined by the UAB clinical laboratory and attending physician. All flow cytometry and bulk cell gene expression data were collected in a double-blinded manner.

These studies were conducted in compliance with the Helsinki Declaration and approved by the institutional review board at UAB. All participants provided informed consent.

Human Abs included BioLegend BV510–anti-CD24 (ML5), PE–anti-CD303 (clone 201A), BV510–anti-IgM (clone MHM-88), Pacific Blue–anti-CD4 (clone RPA-T4), PE-Cy7–anti-CD10 (clone HI10a), BV650–anti-CD27 (O323), Pacific Blue–anti-CD19 (HIB19), PE-Cy7–anti-CD38 (HB-7); SouthernBiotech PE-IgD (IADB6); and PBL Assay Science FITC–anti–IFN-β (MMHB-3). Dead cells were excluded from analysis with Fixable Viability Dye eFluor 780 (eBioscience). For intracellular staining, cells were stained with eFluor 780 viability dye, followed by fixation in 2% PFA and 70% ice-cold methanol permeabilization prior to staining. Purity was validated by postsort analysis of FACS-sorted cells to verify that >99% of cells fell into the sort gate after resorting. FACS data were acquired with an LSR II FACS analyzer (BD Biosciences) and analyzed with FlowJo software (Tree Star, Ashland, OR).

Negative selection–purified B cells (STEMCELL Technologies) were fixed in 2% paraformaldehyde and permeabilized with Triton X-100, followed by blocking with 2% BSA. Cells were stained with AF647 goat anti-human IgM (SouthernBiotech), FITC mouse anti-human IFN-β (clone MMHB-3; PBL Assay Science), or polyclonal rabbit anti-human IFN-β (Abcam), followed by anti-rabbit IgG AF488 secondary Ab. DAPI nuclear stain was included for nucleus determination (200 ng/ml).

Superresolution imaging was carried out using the Nikon N-SIM E superresolution microscope (resolution capable of 120 nm x–y and 300 nm z). Images were acquired with a 100× 1.49 numerical aperture objective; ORCA-Flash 4.0 v2 scientific complementary metal oxide semiconductor camera (Hamamatsu); 488-, 561-, and 640-nm laser excitation; a multipass dichroic cube; and specific emission filters. DAPI images were acquired with a widefield excitation and are at a conventional resolution. Images were acquired and reconstructed with Nikon Elements software. Staining intensity and localization of IFN-β were carried out using Fiji/ImageJ.

RNA isolation, cDNA synthesis, and real-time PCR reactions were carried out as described previously (20, 21). Primers for GAPDH are as follows: forward, 5′-GCACTCACTGGAATGACCTC-3′; backward, 5′-TTCTTCGCACTGACACACTG-3′. All primers used for single-cell analysis are described in Supplemental Table I.

Gene expression analysis of single transitional (Tr) B cells (CD24⁺CD38⁺IgD⁺CD27⁻) obtained from SLE PBMCs was performed using the Fluidigm single-cell capture and BioMark RT-PCR analysis system (Fluidigm, South San Francisco, CA) as described in detail previously (22). Single-cell gene expression clustering analysis was carried out using the ClustVis online web tool (23). Single-cell gene expression data can be retrieved at https://biit.cs.ut.ee/clustvis/?s=zlbtSCZKfAcPrUg.

B cell production of functional type I IFNs was analyzed using coculture of purified primary B cells (5.0 × 10⁵ cells per well) with HEK-Blue IFN-α/β cells that expressed an inducible secreted embryonic alkaline phosphatase (SEAP) (InvivoGen, San Diego, CA). B cells purified using the Human Pan B Cell Isolation Kit (Miltenyi Biotec) were unstimulated or stimulated for 10 min with an F(ab')2 anti-hIg (IgM + IgG) Ab (Thermo Fisher Scientific) (10 μg/ml) plus 5 μg/ml poly(I:C) or 5 μg/ml CL264 (InvivoGen) before HEK-Blue IFN-α/β cells (2.5 × 10⁴ cells per well) were added into the medium. Supernatants were collected at the 24-h time point and were incubated for 1 h with the QUANTI-Blue colorimetric enzyme assay reagent for determination of absorbance at OD₆₅₀.

Results are mean ± SD or mean ± SEM as described in the figure legends. The p values <0.05 were considered significant. Unless otherwise indicated, all analyses were performed using GraphPad Prism software (La Jolla, CA).

We previously showed that type I IFN expression is a prominent feature of transitional stage 1 B cell development in BXD2 autoimmune mice and that T1 B cell IFN-β acts in an autocrine priming mechanism to promote Ifna and ISGs expression (22). Analysis of type I IFN gene expression from SLE patients revealed a significant increase in the expression of IFNB, IFNA1, IFNA14, IFNA17, and MX1 in Tr B cells from AA patients with SLE (Supplemental Fig. 1A). To determine whether distinct type I IFN and type I ISG gene expression patterns were present in SLE B cells, Tr B cells from three female AA SLE subjects as described in Supplemental Fig. 1B were FACS sorted and analyzed for expression of IFNB, IFNA, and ISGs (Fig. 1A). Hierarchical clustering analysis revealed three prominent clusters with distinct gene signatures, including a mixed IFN and ISG producer/responder signature (IFN_P/R), an IFN-responder signature (IFN_R), and an IFN-producing signature (IFN_P) (Fig. 1A, 1B, Supplemental Table I). Cells from all three patients were equally represented in each cluster, revealing the presence of these major clusters in different SLE patients (Fig. 1C).

Cells within the IFN_P/R cluster expressed the highest levels of Tr B cell marker CD24 (Fig. 1D, upper left). Cells within the IFN_P/R cluster also expressed higher levels of IFNB, IFIT1, IRF7, IRF9, ZBP1, IFNA1, IFNA7, and CCND1 compared with the IFN_R or IFN_P cluster or both (Fig. 1D, Supplemental Table I). This is consistent with our previous observations in BXD2 mice in which IFNB expression was upregulated in early T1 B cells (22). Cells within the IFN_R cluster expressed higher levels of IFNAR1, MX1, IFIT2, PKR, RIG1, and CCND2 (Fig. 1E). The IFN_P cluster was characterized by higher levels of IFNA4, IFNA5, IFNA10, IFNA14, IFNA16, and IFNA17 (Fig. 1F). Expression of TLR3 and TLR7 but not TLR9 was different among the groups (Supplemental Fig. 1C). Single-cell gene expression and cell identities were validated by analysis of CD20, CD3, and CD303 expression (Supplemental Fig. 1D). Together, the results reveal heterogeneous type I IFN–producing and –responding signatures in circulating Tr B cells, suggesting that B cells are not only type I IFN targets but also producers of type I IFNs in SLE.

FACS staining (Fig. 2A) and confocal imaging (Fig. 2B) of IFN-β revealed increased levels of IFN-β in B cells from SLE patients compared with healthy controls (HCs). Quantification of IFN-β mean fluorescence intensity (MFI) revealed that although Tr and naive B cells from SLE patients exhibited significantly increased IFN-β compared with HCs, in CD4 T cells and CD303⁺CD4^low plasmacytoid dendritic cells (24), IFN-β levels were not significantly increased in SLE compared with HCs (Fig. 2A). Staining of IFN-β was specifically inhibited by preincubation with human IFN-β but not mouse IFN-α (Supplemental Fig. 2A). We next determined the subcellular localization of IFN-β in SLE B cells. Two anti–IFN-β Abs detected a mainly cytoplasmic distribution of IFN-β in IgM⁺ B cells (Fig. 2C, Supplemental Fig. 2B). Although significantly less prominent compared with the cytoplasmic localization, IFN-β was also detected in the nucleus (Fig. 2C) as previously reported for other cell types (25, 26). Western blot analysis of cytoplasmic extracts from isolated SLE B cells ex vivo further confirmed the presence of a 25- and 50-kDa band (Supplemental Fig. 2C) consistent with the predicted molecular mass of human IFN-β monomer and dimer (27, 28). Together, these results confirm the presence of cytoplasmic IFN-β in SLE B cells.

To determine if B cells produced biologically active type I IFNs, an HEK IFN-α/β reporter cell line assay was carried out. Stimulation of the human B cell lymphoma cell line Ramos with TLR3 ligand poly(I:C) induced the highest IFN-β response compared with a TLR7 ligand (CL264) and a TLR9 ligand (ODN-2006) as measured by both intracellular FACS for IFN-β (Supplemental Fig. 2D, left) and by the HEK reporter assay (Supplemental Fig. 2D, right). Stimulation of purified SLE B cells with anti-Ig plus TLR3 induced an increased HEK reporter response compared with B cells derived from HCs (Fig. 2D). As an additional control, levels of IFN-β measured in the FACS assay were correlated with IFN-β protein secretion as measured by the HEK IFN-α/β reporter assay (Fig. 2E). These results are consistent with previous findings by Gram et al. (29), which reported type I IFN secretion by human B cells upon poly(I:C) stimulation and further suggest the importance of further evaluation of in vivo TLR3 ligands, including U1 RNA that may be associated with B cell secretion of functional type I IFN in SLE patients (30).

The higher expression and production of IFN-β from SLE B cells suggest that B cell intracellular IFN-β may be an important factor associated with SLE pathogenesis. We identified that patients who were seropositive for anti-dsDNA at the time of specimen collection exhibited significantly increased levels of IFN-β (MFI) in Tr and CD27⁺ memory B cells compared with SLE patients who were seronegative for anti-dsDNA at the time of collection (Fig. 3A). Subjects who were positive for anti-Sm at any time during their disease course exhibited a significant increase in IFN-β levels in Tr and CD27⁺ memory B cells, whereas subjects who were seropositive for anti-SSA exhibited increased IFN-β expression in Tr, naive, and CD27⁺ B cells (Fig. 3B, 3C). These data suggest an association between IFN-β expression and enhanced survival of B cells exhibiting reactivity with nucleic acid/protein complexes able to coactivate BCR and TLR signaling (31).

SLE subjects with a history of renal disease also exhibited a significant increase in IFN-β in Tr and naive B cells compared with SLE patients without a history of renal disease (Fig. 3D). Interestingly, intracellular IFN-β in Tr and naive B cells was significantly higher in AA compared with non-AA patients (Fig. 3E). Levels of IFN-β in SLE B cells were generally significantly higher compared with HCs, even in comparisons between autoantibody-negative or otherwise low-severity patients and HCs (Fig. 3A–E). B cell endogenous IFN-β levels were not significantly different in SLE subjects with low complement or other clinical parameters, including the SLE Disease Activity Index or hydroxychloroquine treatment, but were significantly higher compared with HCs (Fig. 3F–H).

These results suggest that B cell IFN-β is most strongly associated with increased autoantibodies and renal disease and that polymorphisms in the IFN-β enhanceosome genes or other upstream genes may predispose some patients to the development of type I IFN dysregulation and autoimmune disease (32). This notion is supported by recent population-level studies that identified the IFNB locus as a trans-regulatory hotspot that controlled antiviral networks enriched in genes differentially expressed in AA versus European American healthy volunteers (33). Together, the present findings support the importance of cell-specific analyses of both IFNs and IFN response genes in patients of defined ancestral backgrounds, as type I IFN expression may not be highlighted in analyses of patient groups with diverse genetic ancestry. It is important to note that the present PCR-based targeted single-cell gene expression analysis approach was selected in order to detect IFN pathway genes, which exhibit a broad expression range spanning from lower-expressed type I IFN genes to higher-expressed ISGs. The detection of type I IFN genes may be more challenging in conventional RNA-sequencing analyses as detection of these genes can be limited by read depth. The present single-cell analyses reveal a new level of understanding in type I IFN dysregulation, as the proper regulation of these type I IFN–producing and –responding populations in early B cells may influence functional cell trajectories.

The authors have no financial conflicts of interest.