B Cell Activation and Plasma Cell Differentiation Are Promoted by IFN-λ in Systemic Lupus Erythematosus

This article is featured in Top Reads, p.

Production of a wide variety of autoantibodies is a hallmark of systemic lupus erythematosus (SLE). Thus, therapies that directly target B cells have long been sought for the treatment of lupus, with belimumab Food and Drug Administration approved (1) and recent promising results in a phase II clinical trial of obinutuzumab (anti-CD20) (2). An alternative approach is to target factors that lead to B cell dysfunction in lupus. There is strong evidence for a role of IFNs in the disease pathogenesis and the B cell dysfunction characteristic of the disease. However, most work has focused on type I IFN (reviewed in Refs. 3, 4).

The effects of type I IFN on B cells in lupus are particularly important (5–9). Type I IFN sensitizes naive B cells to RNA-associated Ags through the upregulation of RNA-binding TLR7 (5). Type I IFN also upregulates costimulatory molecules, such as CD69, CD86, and MHC class II, thus enhancing B cell ability to present Ag, receive T cell help, and generate germinal centers (reviewed in Ref. 10). Moreover, type I IFN promotes the differentiation of B cells to plasma cells (PCs) and the production of Abs (5, 8, 9). In addition, SLE is associated with perturbations in B cell homeostasis with increases in B cell subsets normally seen during viral infection (11–13), including IgD⁻CD27⁻ double-negative (DN), memory, and plasmablast populations. These B cell abnormalities increase during disease flare (14, 15).

Recent studies have also highlighted the potential role of type II IFN (IFN-γ) in SLE. IFN-γ promotes age- or autoimmunity-associated B cells (ABCs), a B cell subset first described in aged mice (16–18) and subsequently in SLE (17, 19, 20). Phenotypic markers for ABCs have varied between studies but include CD11c⁺, CD21⁻, CXCR5⁻, and the transcription factor T-bet (17, 21, 22). Studies in murine and human lupus indicate that ABCs are poised for PC differentiation and enriched in autospecificities (18). CD11c⁺T-bet⁺ B cells in human patients with SLE are overrepresented in the IgD⁻CD27⁻ (DN) compartment, where they are designated DN2 (22). In lupus, DN2/ABCs are expanded in both the blood and tissues (22), where they are speculated to contribute to local production of autoantibodies that result in end-organ damage. Thus, ABCs are of interest, as they may be directly pathogenic in patients with lupus given that autoantibody-producing PCs can be derived from DN2/ABC, DN2/ABCs are found in lupus patients with high disease activity (22), and certain autoantibodies are directly responsible for disease manifestations (reviewed in Ref. 23).

Although type I and type II IFN attracted considerable attention as drivers of B cell aberrations in SLE and infection models, type III IFN remains poorly studied. For many years, the ability to respond to type III IFN (IFN-λ, previously called IL-28 and IL-29) was thought to be restricted to primarily epithelial-derived, plasmacytoid dendritic, and monocyte-derived dendritic cells (24, 25). In responsive cell types, a gene signature that was very similar, if not identical, to that of type I IFN is triggered by IFN-λ (26). More recently, reports suggest human B lymphocytes also respond to IFN-λ (22, 24, 27–31). This is in contrast to murine B cells that cannot be stimulated by type III IFN (28). We hypothesized that the so-called type I IFN gene signature, which is a hallmark of SLE-derived cells, might in part be derived from type III IFN. This has important clinical implications, given that type I IFN receptor–neutralizing agents under development for the treatment of SLE, such as anifrolumab, have no effect on type III IFN signaling, as type III IFNs use a different receptor (32).

In this study, we examine the impact of IFN-λ1 on human B cells and its ability to accelerate PC differentiation. We demonstrate that DN B cells in patients with SLE have an IFN gene signature. Further, B cells transcribe IFN-regulated transcripts historically associated with type I IFN in response to type III IFN. Notably, IFN-λ1 was elevated in SLE serum and correlates with the frequency of DN2/ABCs in peripheral blood. We show that CD11c⁺ B cells robustly phosphorylate STAT1 signaling proteins in response to IFN- λ1 and that IFN-λ1 augments T-bet–expressing B cell and plasmablast formation in the setting of TLR7/8 human B cell activation. Thus, our results demonstrate that IFN-λ1 can have effects on human B cells, with critical implications for both viral infection and autoimmune disease.

Samples were collected from donors with SLE and healthy donors (HDs) via venipuncture at the University of Rochester Medical Center or affiliated outpatient clinics after informed consent in accordance with approved Institutional Review Board protocol. For RNA-sequencing (RNA-seq) analysis, blood was obtained from eight patients with SLE and five HDs (Supplemental Table I). For ex vivo B cell subset flow cytometry phenotyping studies, blood was obtained from 26 patients with SLE and 6 HDs (Table II). All patients fulfilled the American College of Rheumatology (ACR) SLE classification criteria (33). Disease activity was assessed by the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) (34). All patients were on pharmacologic therapy for lupus (Table II). All patients had a positive anti-nuclear Ab (ANA; indirect immunofluorescence with Hep-2 cells) of at least 1:80. Anti-dsDNA titers were determined by the clinical laboratory. Anticoagulated venous blood from patients with SLE and HDs was subjected to density gradient centrifugation to obtain PBMCs. RNA-seq, cultures, and PCR experiments used fresh cells without freezing. Samples for ex vivo phenotyping and phosphorylation flow cytometry were stored in liquid nitrogen and rapidly thawed in batches for analysis. Serum was stored at −80°C.

Fresh PBMCs were FACS sorted into a Qiagen RLT lysis buffer. Total RNA was isolated using the RNeasy Plus Micro Kit (Qiagen, Valencia, CA) per the manufacturer’s recommendations. RNA concentration was determined with the NanoDrop 1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE) and RNA quality assessed with the Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA). Total RNA (1 ng) was preamplified with the SMARTer Ultra Low Input Kit v4 (Clontech, Mountain View, CA) per the manufacturer’s recommendations. The quantity and quality of the subsequent cDNA were determined using the Qubit Fluorimeter (Life Technologies, Carlsbad, CA) and the Agilent Bioanalyzer 2100. Among the DN B cell samples, one sample for the SLE group and one from the healthy controls were excluded based upon poor RNA quality. cDNA (150 pg) was used to generate Illumina-compatible sequencing libraries with the Nextera XT Library Prep Kit (Illumina, San Diego, CA) per the manufacturer’s protocols. All samples were sequenced as one batch to eliminate batch effects. The amplified libraries were hybridized to the Illumina single-end flow cell and amplified using the cBot (Illumina). Single-end reads of 100 nt were generated by the HiSeq2500 (Illumina). Raw reads generated from the Illumina basecalls were demultiplexed using bcl2fastq version 2.19.1. Quality filtering and adapter removal are performed using FastP version 0.20.0 with the following parameters: “- -length_required 35 - -cut_front_window_size 1 - -cut_front_mean_quality 13 - -cut_front - -cut_tail_window_size 1 - -cut_tail_mean_quality 13 - -cut_tail -y -r”. Processed/cleaned reads were then mapped to the Homo sapiens reference genome (GRCh38 + Gencode-28 Annotation) using STAR_2.7.0f with the following parameters: “- -twopass Mode Basic - -runMode alignReads - -outSAMtype BAM SortedByCoordinate - -outSAMstrandField intronMotif - -outFilterIntronMotifs RemoveNoncanonical - -outReads UnmappedFastx” (35, 36). Gene-level read quantification was derived using the subread-1.6.4 package (featureCounts) with a GTF annotation file (Gencode-28) and the following parameters: “-s 0 -t exon -g gene_name” (37). Differential expression analysis was performed using DESeq2-1.22.1 with a p value threshold of 0.05 at a false discovery rate of 0.05 within R version 3.5.1 (https://www.R-project.org/) (38). Counts were normalized using DESeq2, which uses a median of ratio methods where counts are divided by sample-specific size factors determined by median ratio of gene counts relative to the geometric mean per gene. Heat maps were generated using the ComplexHeatmap package with variance-stabilized transformation expression values. Gene ontology analyses were performed using the EnrichR and clusterProfiler packages (39). Data were deposited in the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE175913.

For cell culture experiments, up to 120 cc peripheral blood was obtained from HDs per Institutional Review Board protocol. After density centrifugation with Lymphocyte Separation Media, mononuclear cells were positively sorted by CD19-positive magnetic microbead selection using LS columns (Miltenyi Biotec). All cell cultures were performed in 96-well round-bottom plates with freshly isolated cells in 200 μl RPMI 1640 with 10% FCS supplemented with penicillin/streptomycin at 250,000 cells/well. Stimulated cultures contained combinations of BAFF (100 ng/ml), R848 (1 μg/ml), IL-21 (10 ng/ml), IFN-γ (20 ng/ml), anti-IgG/A/M F(ab)2 fragments (10 μg/ml), IFN-λ1 (500 U/ml), or IFN-α2 (500 U/ml). As no CD27^hiCD38⁺ cells were detectable when BCR stimulation was included for the duration of the 7-d culture in agreement with previous studies (20), cultures to generate PCs did not contain anti-IgG/A/M.

After thaw, patient serum was centrifuged and supernatant ran on ELISA per the manufacturer’s instructions. IFN-α level was measured by an all-α-subtypes ELISA. IFN-β was measured by using a high-sensitivity ELISA kit. Both kits were labeled for use in serum (PBL Assay Science, Piscataway, NJ). To measure IFN-λ1, the IL-29 ELISA kit was used (Invitrogen, Carlsbad, CA). To measure IgM or IgG, B cell cultures were spun down and supernatants frozen at −20°C until time of ELISA (Bethyl Laboratories).

Cell viability was assessed by LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (Invitrogen) following the manufacturer’s protocol. Samples were incubated with Abs for 30 min at 4°C. Intracellular staining was performed using the eBioscience FoxP3/Transcription Factor Staining Buffer Set (Invitrogen) following the manufacturer’s recommendations. Cells were fixed for 20 min in 1% methanol-free formaldehyde and then diluted in 1% BSA in PBS in a 2:1 ratio. Cells were run on an LSR II Fortessa flow cytometer (BD Biosciences). Data analysis was performed using FlowJo software (Tree Star). The following anti-human mAbs were used for B cell phenotyping: CD19-allophycocyanin-CY7 (clone SJ25C1, 557791; BD Biosciences), IGD-FITC (clone IA62, 555778; BD Biosciences), CD11c-BUV395 (clone B-ly6, 563787; BD Biosciences), CD27-BV605 (clone O323, 302830; BioLegend), CD24-PE610 (clone SN3, MHCD2422; Life Technologies), CD38-PerCP-5.5 (clone HIT2, 551400; BD Biosciences), CD95-allophycocyanin (clone DX2, 558814; BD Biosciences), CD20-ALX700 (clone 2H7, 302322; BioLegend), CXCR3-PE (clone 1C6, 550633; BD Biosciences), CD21-PE-CY5 (clone B-LY4, 551064; BD Biosciences), CD3-BV421 (clone UCHT1, 562426; BD Biosciences), and T-bet–PECY7 (clone 4B10, 644824; BioLegend). Gating is shown in each figure. B cell subsets were defined as: naive, IgD⁺CD27⁻; DN, IgD⁻CD27⁻; unswitched memory, IgD⁺CD27⁺; switched memory, IgD⁻CD27⁺; PC, IgD⁻CD27^hiCD38⁺; preplasmablast, CD27⁻CD38⁺; and DN2, IgD⁻CD27⁻CD21⁻CD24⁻.

B cell pellets were suspended in RLT plus lysis buffer (Qiagen) with 10 μl/ml 2-ME and frozen at −80°C until RNA extraction. RNA was purified using Quick-RNA Prep Kits (Zymo Research). RNA was transcribed to cDNA using qScript SuperMix (Quantabio). Quantitative real-time PCR was performed on the QuantStudio 3 System (Applied Biosystems) using ROX as a reference dye in PerfeCTa FastMix II (Quantabio) and TaqMan-specific probes (Hs02758991-g1 GAPDH, Hs01911452_s1 IFIT1, Hs01921425_s1 ISG15, Hs01086370_m1 IFI27, and Hs00276441_m1 USP18). GAPDH was used as an endogenous control. Fold change was calculated using the 2^−ΔΔCt method.

Frozen healthy PBMCs were thawed and rested at 37°C for 1.5 h in RPMI 1640 with 10% FCS and 1× penicillin/streptomycin (Life Technologies) media. Experiments were performed to establish kinetics and concentration of maximal STAT1 phosphorylation for IFN-α2 and IFN-λ1. Cells were washed and then treated with media or media with 500 U/ml IFN-α2 or 2000 U/ml IFN-λ1 for 25 min at 37°C. For IFN receptor neutralization conditions, anti-IFNAR2 (50 μg/ml; MAB1155; Millipore Sigma) or anti-IFNLR1 (1 μg/ml; 21885-1; PBL Assay Science) were incubated with cells 1 h prior to stimulation. After stimulation, cells were fixed with 1% formalin for 10 min at room temperature and surface stained with Abs for 30 min at 4°C, followed by LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (Invitrogen) for dead cell exclusion. Surface stain panel included: CD3-PE-Cy5 (clone UCHT1, 555334; BD Biosciences), CD11c-BUV395 (clone B-ly6, 563787; BD Biosciences), CD14-ALX700 (clone M5E2, 557923; BD Biosciences), CD19-BV605 (clone SJ25C1, 363024; BioLegend), CD20-allophycocyanin-Cy7 (clone 2H7, 302314; BioLegend), CD27-BUV737 (clone L128, 612829; BioLegend), CD38-BB700 (clone HIT2, 566445; BD Biosciences), and IgD-BV421 (clone IA6-2, 348226; BioLegend) Abs. Cells were then fixed using BD Cytofix for 10 min at room temperature followed by 90% BD Phosflow Perm Buffer III for 20 min at 4°C. Two washes were performed prior to intracellular fluorescent Ab staining of STAT1-PE (clone 1/STAT1, 558537; BD Biosciences) and p-STAT1–Alexa 488 (clone pY701, 612596; BD Biosciences) for 45 min at 4°C. Cells were then washed and fixed in 1% formalin for 15 min at room temperature. An additional 2:1 of 1% BSA/PBS was added and samples stored at 4°C overnight until flow cytometry. Fold change was calculated by dividing median fluorescence intensity of p-STAT1 for the treated sample by that of the unstimulated sample.

Data were plotted with GraphPad Prism or R Studio with ggplot2 (40) and ggpubr (41) packages. Statistical tests were performed as described in the figure legends. Statistical significance is indicated with *p < 0.05, **p < 0.01, and ***p < 0.001.

To identify key signals mediating B cell activation in lupus, we performed transcriptomic analysis on sorted naive and DN B cells. Human peripheral blood from HDs (n = 5) or patients with SLE (n = 8) meeting ACR criteria was flow sorted into CD14⁺ (monocytes), IgD⁻CD27⁻CD19⁺ (DN B cells), or IgD⁺CD27⁻CD19⁺ (naive/transitional B cells) populations and RNA isolated for bulk RNA-seq transcriptomic analysis (Fig. 1A). SLEDAI on the lupus samples ranged from 0 to 12 (Supplemental Table I). Hierarchical clustering analysis demonstrated that lupus samples clustered separately from controls for each cell type (Fig. 1B). Differential gene expression analysis between lupus and healthy showed 179 differentially expressed genes for naive B cells, 196 for DN, and 511 for monocytes (p adjusted <0.05) (Supplemental Table II). Notably, IFN-stimulated genes (ISGs) were prevalent among the top 30 upregulated genes for each cell type (volcano plots in (Fig. 1C). Naive and DN B cells had no differentially downregulated genes in common (Fig. 1D). Among the upregulated genes, 56 were shared between naive and DN B cells and 47 shared among all three cell types (Fig. 1D). These shared upregulated genes included ISGs, such as IFI27, USP18, RSAD2, IFI44, IFI44L, ISG15, IFIT1, IFIT3, STAT1, MX1, MX2, and OAS2 (Table I). The pathways with the greatest number of genes upregulated in SLE that were shared among all three cell types (Fig. 1E) were the cytokine-mediated signaling pathway (GO:0019221), cellular response to type I IFN (GO:0071357), and type I IFN signaling pathway (GO:00600337). Gene set enrichment analysis independently performed for the DN cells identified type I IFN signaling pathway (GO:0060337 and GO:0071357), regulation of viral genome replication (GO:0045069, GO:0045071, and GO:1903901), and dsRNA binding genes as among the top statistically significant upregulated pathways in SLE (Fig. 2A, Supplemental Fig. 1B). These pathways were also upregulated in the SLE naive and monocytes (Supplemental Fig. 1). Because type I and type III IFN receptors share many components of their signaling pathways, these data cannot distinguish which IFN is inducing these ISGs.

Conflicting reports exist within the literature regarding whether B cells express functional type III IFN receptor (29, 42). To determine which IFNs could contribute to the ISG signature in lupus B cells, we compared IFN receptor RNA transcript expression in the healthy versus SLE naive and DN B cells. For the type I IFN receptor, there was no difference in the IFNAR2 normalized transcript counts between SLE and HDs for either naive or DN B cells (Fig. 2B). For the type II receptor, more IFNGR1 transcript was detected among the naive cells compared with the DN for both HDs and SLE (Fig. 2B). DN B cells expressed more transcript for the unique IFNLR1 subunit of the type III IFN receptor than did naive for both HDs and donors with SLE (Fig. 2B). For IFNLR1, a significant interaction (p < 0.02) was found between B cell subset and disease status. Among the SLE DN B cells, there was more heterogeneity in IFNLR1 expression compared with the naive B cell subset (Fig. 2B, Supplemental Fig. 2A).

To examine which IFNs contribute to the DN B cell transcriptomic heterogeneity, we evaluated the relationship between CD11c⁺CD21⁻ B cell abundance as determined by flow cytometry and the expression of each of the IFN receptor transcripts (Fig. 2D for IFNLR1, Supplemental Fig. 2B). IFNLR1 was highly and positively correlated (Pearson r = 0.97) with percentage of CD11c⁺CD21⁻ B cells (Fig. 2D). High numbers of IFNLR1 transcripts in DN from patients with SLE positively associated with expression of known DN2/ABC markers, including FCRL3, FCRL5, and ZEB2, and negatively associated with CXCR5 and CD27 transcript (Fig. 2E). This association of the IFNLR1 transcript with DN2/ABCs suggests that IFN-λ could influence the extrafollicular B cell differentiation pathways in SLE.

Given that transcriptomic gene set enrichment analysis identified a strong IFN gene signature and both type I and type III IFN receptors in SLE B cells, we next sought to determine which IFNs are present in SLE serum. We obtained serum and peripheral blood cells from patients with SLE (n = 26) and HDs (n = 6) (Table II) and measured both type I IFN (all subtype IFN-α as well as IFN-β) and type III IFN (IFN-λ1). All three IFNs were detectable at increased levels in the patients with SLE compared with healthy controls (Fig. 3A). B cell phenotype from these donors was determined by flow cytometry (Fig. 3B). A statistically significant expansion in phenotypes associated with extrafollicular pathway PC precursors, including CD19⁺CD11c⁺CD21⁻, CD19⁺IgD⁻CD27⁻CD21⁻CD24⁻ (DN2), and CD19⁺T-bet⁺CD11c⁺ DN2 B cell subsets, was found in patients with SLE compared with healthy controls (Fig. 3C). The highest percentage of CD11c⁺CD21⁻ B cells was found in the DN compartment (Fig. 3D, HDs, 58.9 ± 2.7%; SLE, 55.1 ± 3.2% of CD11c⁺CD21⁻ B cells, mean ± SEM). In SLE, there is an expansion of the CD11c⁺CD21⁻ DN cells (Fig. 3D, mean for HDs or SLE shown as pie graph) among total B cells. CD11c⁺CD21⁻ B cells showed a positive correlation with serum levels of IFN-λ1 (Fig. 3D). IFN-α (all subtypes) and IFN-β showed weaker positive correlation with these subsets and did not achieve statistical significance (Table III). These data raised the possibility that IFN-λ could be playing a role in induction of atypical B cells in lupus.

As a first step to understand the influence of IFN-λ on B cells, we asked if B cells directly respond to type III IFN stimulation. Time course experiments demonstrated expression of IFIT1 at 2, 4, and 6 h for both IFN-λ1 and IFN-α2 treatment (Fig. 4A, n = 4 donors, independent experiments, log₂fold change, mean ± SEM). We measured select ISGs identified in our SLE B cell RNA-seq transcriptomic analysis. Relative expression of IFIT1, IFI27, ISG15, and USP18 was quantitated for IFN-λ1 and IFN-α2 stimulation at concentrations from 1 to 10,000 U/ml (Fig. 4B, n = 4 donors, independent experiments, log₂fold change, mean ± SEM). Both IFN-λ1 and IFN-α2 induced expression of these ISGs, but higher concentrations of IFN-λ1 were required, and the magnitude of mRNA induction was greater with IFN-α2. IFN-λ1 and IFN-α2 treatment did not induce B cell expression of IFNA2, IFNB1, or IFNL1, which were undetectable by quantitative RT-PCR (data not shown). Thus, we confirmed based on induction of ISG mRNA that human B cells are directly responsive to type III IFN, but also found IFN-α2 had a more potent effect.

To further define the B cell subsets targeted by IFN-λ, peripheral blood B cells were stimulated with IFN-λ1 or IFN-α2 and p-STAT1 measured by flow cytometric analysis. STAT1 is a protein that becomes phosphorylated during signal transduction for both IFN-α2 and IFN-λ1. A time course experiment was performed that determined B cells phosphorylate STAT1 faster after IFN-α2 stimulation compared with IFN-λ1 (Supplemental Fig. 2C). A dose-response experiment demonstrated maximal STAT1 phosphorylation with 500 U/ml IFN-α2 and 2000 U/ml IFN-λ1 (data not shown). These concentrations were used for subsequent STAT1 phosphorylation experiments at a time of 25 min. Total CD3⁺ T lymphocytes and CD14⁺ monocytes demonstrated an increased median fluorescence intensity for p-STAT1 protein after IFN-α2 treatment, but not for IFN-λ1 stimulation (Fig. 5A), in agreement with recent reports (30). p-STAT1 protein increased after IFN-λ1 stimulation in total human B cells, though to a lesser degree than with IFN-α2 (Fig. 5B). Further phenotyping of B cells showed that naive (IgD⁺CD27⁻), DN (IgD⁻CD27⁻), switched memory (IgD⁻CD27⁺), and unswitched memory (IgD⁺CD27⁺) B cells all increased phosphorylation of STAT1 with IFN-λ1 and IFN-α2 stimulation (Fig. 5B). Of these, naive cells had the highest fold change. In each population, IFN-α2 had a larger effect than IFN-λ1. For donors with a clearly defined CD11c⁺ DN population, a robust phosphorylation of STAT1 was evident in response to IFN-λ1 that approached that of IFN-α2 (Fig. 5C). CD11c⁺ B cells had a higher baseline p-STAT1 compared with their CD11c⁻ counterparts and higher levels of p-STAT1 in response to IFN-λ1 in all B cell subsets (Fig. 5D, 5E). CD11c⁺ DN B cells (mean ± SD, 3.03 ± 0.43; n = 9) and CD11c⁺ naive B cells (mean ± SD, 2.97 ± 0.41; n = 9) had the highest fold changes with IFN-λ1 stimulation. These results are in agreement with previous reports that DN2, which are CD11c⁺, are hyperresponsive to IFN-λ (22). IFN receptor neutralization experiments using blocking Ab to IFNAR2 (type I IFN) or IFNLR1 (type III IFN) confirmed the specificity of the effect for both type I and type III IFN (Fig. 5F, 5G, respectively; n = 3 HDs; each point indicates a donor, each run as independent experiments, mean ± 95% confidence interval shown for median p-STAT1 for CD19⁺ B cells). As expected, these results demonstrate that type I IFN (IFN-α2) broadly stimulates multiple types of peripheral human blood cells, including monocytes, T cells, and B cells. In contrast, type III IFN (IFN-λ1) is specific for B cells and is particularly effective for CD11c⁺-expressing B cells, such as the DN2/ABC subset.

Ab-secreting cell precursors DN2 or ABCs have been reported to be generated from human B cell cultures containing anti-Ig BCR stimulation, TLR7 stimulation, IL-21, BAFF, and IFN-γ (type II IFN) (20, 43). As our data demonstrate that ABC phenotype and serum IFN-λ1 (type III IFN) correlate in patients with SLE in vivo (Fig. 3D), we assessed whether type III IFN-λ can induce DN2, similar to IFN-γ in vitro. Human CD19⁺ B cells from HDs (n = 6) were BCR and TLR7/8 stimulated with or without IFN-λ1 or IFN-γ in the presence of BAFF and IL-21 for 7 d. T-bet⁺CD11c⁺ B cells developed under these culture conditions (Fig. 6A). BCR and TLR7/8 stimulation upregulated CD11c expression in the TLR7/8-activated cultures regardless of whether IFN was included, but the percentage of T-bet⁺ B cells increased only in conditions with either IFN (Fig. 6A, n = 6). Increased frequencies of T-bet⁺ B cells were found among B cells differentiated with BCR and TLR7/8 stimulation in cultures containing IFN-λ1 compared with cultures without IFN (Fig. 6C), and, as expected, cultures containing IFN-γ generated T-bet⁺ B cells at high frequency (Fig. 6C, Wilcoxon signed-rank test). Additionally, T-bet protein levels measured by geometric mean fluorescence intensity (MFI) increased in the IFN-λ1–containing conditions compared with those without IFN, but to a lesser degree than in IFN-γ–containing cultures (Fig. 6B).

Given T-bet expression increased in our TLR7/8-activated cultures exposed to IFN-λ1 and T-bet promotes Ab-secreting cell differentiation in mice (44), we sought to determine the effects of IFN-λ stimulation on the differentiation of plasmablasts. Human B cells from HDs were cultured for 7 d with the TLR7/8 agonist with or without IFN-λ1 and plasmablasts enumerated based on high CD27 and CD38 expression. We found a statistically significant enhancement in CD27^negCD38⁺ preplasmablasts and CD27^hiCD38⁺ plasmablasts with the addition of IFN-λ1 (Fig. 7A–C, No IFN versus +α or +λ). We also directly compared the effects of IFN-λ1 to type I IFN-α2. The addition of either type I (IFN-α2, n = 9) or type III (IFN-λ1, n = 10) to the cultures caused a similar statistically significant increase in CD27^negCD38⁺ preplasmablasts and CD27^hi CD38⁺ plasmablasts (Fig. 7B, 7C, No IFN versus +α or +λ).

To further define the effects of IFN-λ1 on PC differentiation, we measured IgG and IgM secretion in the cultures. R848, BAFF, and IL-21 stimulation resulted in IgG and IgM secretion (Fig. 7D, 7E, No IFN). Inclusion of IFN-λ1 in the R848-, BAFF-, and IL-21–stimulated cultures significantly enhanced IgM, but not IgG, production (Fig. 7D, +λ). IFN-λ enhances B cell differentiation into plasmablasts and increases secretion of IgM.

Type I IFN has long been associated with SLE. However, the type I IFN gene transcription profile is indistinguishable from that produced by the type III IFN (IFN-λ) in cells that express IFN-λ receptor (IFNLR1). In this study, we have found that type III IFN (IFN-λ1), in addition to type I IFN (IFN-α and IFN-β), is increased in SLE serum. IFN-λ1 serum levels, but not IFN-α and IFN-β levels, correlate with B cell phenotypes associated with DN2/ABC (increased CD11c expression and downregulation of CD21). We show that human B cells express transcripts for the receptors of all three types of IFN and that the level of IFNLR1 mRNA transcripts in DN B cells correlates with the frequency of DN2/ABCs. By measuring STAT1 phosphorylation, we were able to show that the IFN-λ receptor is functional in all major B cell subsets and that levels of STAT1 phosphorylation after IFN-λ stimulation are greatest in B cells with cell surface markers for DN2/ABCs. Moreover, in all major B cell subsets, CD11c⁺ B cells responded better to IFN-λ than CD11c⁻ B cells. In marked contrast, peripheral blood T cells and monocytes are not responsive to IFN-λ. Functionality of the IFN-λ receptor on B cells was further demonstrated by induction of ISG transcripts upon stimulation with IFN-λ1. We also demonstrate that IFN-λ1 promotes TLR7/8-mediated differentiation of human B cells to Ab-secreting PCs. As CD11c⁺ B cells are associated with PC differentiation (45), we also establish that IFN-λ1 increases the numbers of T-bet⁺ B cells and plasmablasts among TLR7-activated B cell cultures. Thus, our data confirm recent reports that IFN-λ stimulates human B cells (22, 24, 27–31) and is likely a driver of the prominent IFN signature and B cell dysregulation in SLE.

Even though type I IFN induces a similar transcriptional profile on a broader distribution of cell types, nonredundant roles for IFN-λ are described. In a murine TLR7-driven model of lupus, IFNLR1^−/− mice had reduced skin and renal inflammation compared with wild-type mice (28). In this setting, IFN-λ effects were mediated via chemokine expression in tissue, leading to recruitment of inflammatory cells (28). However, murine models cannot completely recapitulate human SLE, given murine B cells are not directly responsive to IFN-λ (28). In epithelial cell viral infections, IFN-λ1 is produced at greater concentrations and with more prolonged duration of action than IFN-α (46, 47). If similar IFN-λ1 kinetics are also found in SLE, we speculate there may be points in the disease course at which IFN-λ becomes more dominant in its effects on human B cells.

During lupus flare, there is an increase in the IgD⁻CD27⁻ B cell compartment due to expansion of CD11c⁺CD21⁻ DN2/ABCs (48, 49). Our data show a correlation between DN2/ABC frequency and IFN-λ1 serum levels. This subset is of significant interest, as it is poised for PC differentiation and enriched in autospecificities (18). In SLE, these DN2/ABCs share transcriptomic and epigenetic features with the naive compartment (20, 50, 51). Trajectory analysis supports a resting naive to activated naive (CD11c⁺T-bet⁺) to DN2 to Ab-secreting cell differentiation pathway (20, 50, 51). IFN-λ3 augmentation of IgM plasmablast differentiation from naive B cells using a BCR-activated system was recently described in which effects were mediated via the mTORC1 pathway (30). Our data also show that IFN-λ1 augments production of IgM in connection with increased T-bet⁺ B cell and PC differentiation using TLR7 for activation. Our experiments do not directly address the effects of IFN-λ on class switching and IgG PC differentiation in part because of the lack of CD40 stimulation in the cultures.

TLR7 stimulation generates DN2/ABCs via an extrafollicular pathway in viral and autoimmune disease (22). In infection, the extrafollicular response provides IgM-producing plasmablasts early in the infection course (10, 52–54). In lupus, DN2/ABCs are expanded in both the blood and tissues (51), such as the renal microenvironment, where they are speculated to contribute to local production of autoantibodies that result in end-organ damage. Plasmacytoid dendritic cells, activated T cells, and cells of epithelial origin are capable of type III IFN production. Thus, RNA-containing immune complexes can stimulate plasmacytoid dendritic cells to produce IFN-λ1 (55). Immunohistochemistry staining of renal lupus nephritis biopsies identified IFN-λ protein (56, 57). We previously reported that primary renal tubule epithelial cells produce IFN-λ in vitro in response to TLR3 stimulation (58). Thus, within the renal microenvironment, we posit that renal tubule epithelial cells could promote T-bet–expressing DN2/ABCs and subsequent differentiation toward plasmablasts. In other autoimmune diseases, similar networks might contribute to the generation of autoimmunity in other epithelial tissues. The relative contribution of type III versus other classes of IFN to autoimmunity remains to be parsed apart.

These findings are of critical clinical importance, given type I IFN remains a promising SLE therapeutic target of interest culminating in clinical trials for type I IFN receptor blockade with the mAb anifrolumab (59). However, type I IFN mAb blockade would not neutralize IFN-λ, given type I and type III IFN signal through unique receptors. As type III IFN induces similar genes as type I IFN in cell types capable of expressing the receptor, type III IFN might also drive disease pathogenesis. Alternatively, escape of IFN-λ from mechanisms targeting type I IFN might serve as a safety net to prevent severe immunosuppression that might occur if all IFNs were universally blocked therapeutically or via viral proteins. In contrast to IFN blockade, pegylated IFN-λ was developed as an antiviral therapeutic for hepatitis C (60), is in trial for use in the treatment of severe acute respiratory syndrome coronavirus 2, and may be less inflammatory than type I IFN treatment (61). Our phosphorylation data suggest this is likely true given minimal responsiveness of monocytes and resting T cells to IFN-λ, but we would anticipate this therapy to still have effects on human B cells.

The use of clinically relevant human and patient with SLE samples is a strength of our study. Previous studies relying upon Western blot for the measurement of STAT1 phosphorylation events suggested B cells were poorly responsive to IFN-λ1 (42, 62). The use of single-cell measurements of phosphorylation events via flow cytometry allowed phenotyping of responsive B cells in our study. Weaknesses of our study include the use of bulk transcriptomic analysis as opposed to single-cell transcriptomic analysis and a reliance on in vitro cultures given the limitations associated with working with human samples.

In conclusion, our studies highlight that IFN-λ directly targets human B cells to promote differentiation to plasmablasts. IFN-λ preferentially targets CD11c⁺ B cells, the majority of which are found in the DN compartment and expanded in SLE. Therefore, IFN-λ may contribute to autoantibody formation via the TLR7-stimulated extrafollicular B cell activation pathway from naive to DN2/ABCs to Ab-secreting PCs. In conclusion, IFN-λ is likely an underappreciated player in promoting Ab responses for autoimmune diseases like SLE and may have implications in viral infection.

We thank Tyler Cavin, Dr. Lin Gao, Amanda Howell, Dr. Neha Nandedkar-Kulkarni, Mary O’Connell, and Dr. Javier Rangel-Moreno for technical assistance. We also thank the staff of the University of Rochester Genomics Core for performing RNA-seq for this study.

The authors have no financial conflicts of interest.