Physiological and druggable skipping of immunoglobulin variable exons in plasma cells

Abstract

The error-prone V(D)J recombination process generates considerable amounts of nonproductive immunoglobulin (Ig) pre-mRNAs. We recently demonstrated that aberrant Ig chains lacking variable (V) domains can be produced after nonsense-associated altered splicing (NAS) events. Remarkably, the expression of these truncated Ig polypeptides heightens endoplasmic reticulum stress and shortens plasma cell (PC) lifespan. Many questions remain regarding the molecular mechanisms underlying this new truncated Ig exclusion (TIE-) checkpoint and its restriction to the ultimate stage of B-cell differentiation. To address these issues, we evaluated the extent of NAS of Ig pre-mRNAs using an Ig heavy chain (IgH) knock-in model that allows for uncoupling of V exon skipping from TIE-induced apoptosis. We found high levels of V exon skipping in PCs compared with B cells, and this skipping was correlated with a biallelic boost in IgH transcription during PC differentiation. Chromatin analysis further revealed that the skipped V exon turned into a pseudo-intron. Finally, we showed that hypertranscription of Ig genes facilitated V exon skipping upon passive administration of splice-switching antisense oligonucleotides (ASOs). Thus, V exon skipping is coupled to transcription and increases as PC differentiation proceeds, likely explaining the late occurrence of the TIE-checkpoint and opening new avenues for ASO-mediated strategies in PC disorders.

Introduction

During the V(D)J recombination process, combinatorial and junctional diversity ensures that the immunoglobulin (Ig) primary repertoire is adapted to the vast heterogeneity of antigens. However, random nucleotide additions or deletions generate numerous out-of-frame V(D)J junctions that lead to the appearance of premature termination codons (PTCs) in Ig mRNAs.^1,2 The nonsense-mediated mRNA decay (NMD) pathway is very active in B-lineage cells and ensures the rapid degradation of PTC-containing Ig mRNAs, limiting the synthesis of truncated Ig polypeptides.^3–6

Whereas NMD basically protects cells from truncated protein synthesis, the activation of another RNA surveillance pathway referred as nonsense-associated altered splicing (NAS) can exert opposite effects and yield internally deleted mRNAs and proteins.^7–9 We previously observed that the presence of nonsense codons within the variable (V) exon enhanced exon skipping and the production of Igκ light chains without V domains (ΔV-κLCs).^5,10 Remarkably, these truncated Ig chains exhibited toxic effects and induced endoplasmic reticulum (ER) stress-associated apoptosis in antibody-secreting cells.¹⁰ Hence, the production of ΔV-κLCs blunts plasma cell (PC) differentiation through activation of an antigen-independent process known as the truncated Ig Exclusion (TIE-) checkpoint. This novel PC checkpoint modifies the Ig repertoire by eliminating terminally differentiated cells harboring biallelic Igκ rearrangements with a PTC-containing V exon (V^PTC) on the nonproductive allele. Thus, aberrantly rearranged Ig alleles are frequent but not innocuous, and their translation into truncated Ig chains is often possible after skipping of V^PTC exons.

Because the TIE-checkpoint provokes PC death as a consequence of deregulated proteostasis, its activation per se precludes the analysis of V exon skipping. Accordingly, alternatively spliced mRNAs (alt-mRNAs) encoding structurally abnormal Ig chains are hardly detectable in PCs.¹⁰ To overcome this issue and decipher how exon skipping events are regulated during PC differentiation, we utilized a mouse model harboring an additional frameshift-inducing V exon (frVκ) at the Ig heavy (IgH) chain locus.^11,12 Using this dedicated model, skipping of the supplemental frVκ exon does not encode truncated Ig chains provided in-frame VDJ junctions. Therefore, by uncoupling Ig exon skipping to the activation of the TIE-checkpoint, we could depict the regulation of NAS of IgH pre-mRNAs in vivo. As a new strategy for specifically affecting the fate of PCs transcribing a given V exon, we also assessed whether V exon skipping was achievable with splice-switching antisense oligonucleotides (ASOs). The proof of the concept for ASO-mediated V exon skipping should thus open new avenues for “inducible-TIE” therapeutic approaches in PC dyscrasias.

Results

Determining exon skipping of IgH transcripts in vivo

To evaluate the extent of nonsense-associated Ig exon skipping during PC differentiation, experiments were performed using heterozygous IgH^wt/frVκ knock-in mice.^11,12 In this model, the expression of VDJ-rearranged IgH^wt alleles drives B-cell development, whereas the supplemental frVκ exon leads to the appearance of PTCs that trigger NMD degradation of full-length IgH^frVκ mRNAs.⁶ By contrast, the activation of the NAS pathway can yield functional IgH chains by eliminating the frVκ exon from transcripts with in-frame VDJ junctions (Fig. 1a).

Fig. 1.

Assessment of exon skipping of nonproductive VDJ-recombined IgH transcripts in vivo. a Map of wild-type (b allotype) and targeted (a allotype) IgH loci. IgH^wt/frVκ knock-in mice were created by insertion of an additional frameshift-inducing Vκ exon (frVκ) between J_H4 and Eµ. Skipping of frVκ exon occurs after NAS of nonproductive IgH^frVκ pre-mRNAs, allowing for the expression of complete IgH^a chains upon in-frame VDJ rearrangement. Primers and probes used to detect specifically IgH^wt (b) and IgH^frVκ (a) alleles by qPCR are represented (black arrows and short rectangles). b-d To determine the amounts of total IgM (b), IgM^a (c), and IgM^b (d) allotypes, ELISA assays were performed in sera from 6 to 8-week-old IgH^wt/frVκ (wt^b/frVκ^a), B6/129 F1 (wt^b/wt^a) and B6 (wt^b/wt^b) mice (n = 3−8/group). c The negative threshold of IgM^a expression was obtained using sera from B6 mice (dotted line). Bars represent the mean expression ± SEM. Unpaired two-tailed Student’s t test was performed to determine significance (ns not significant; *P < 0.05; ***P < 0.001)

In IgH^wt/frVκ mice with a mixed B6/129 background, allotypic differences allowed for us to distinguish between IgM^b expression from the IgH^wt allele and IgM^a produced upon exon skipping of VDJ-rearranged IgH^frVκ pre-mRNAs. The secretion of total IgM, IgM^a, and IgM^b was assessed in the sera of B6, F1 (B6/129), and IgH^wt/frVκ mice using standard or allotype-specific ELISA assays. Sera from B6 mice served as negative controls for anti-IgM^a assays. As expected, similar serum concentrations of total IgM and IgM^b were found in B6 and IgH^wt/frVκ mice (Fig. 1b, d). Interestingly, we found significant amounts of IgM^a in sera from IgH^wt/frVκ mice, albeit diminished compared with F1 (B6/129) mice in which ∼50% of the B-lineage cells express a functional IgHa allotype (Fig. 1c). Thus, NAS is involved in the processing of nonproductive IgH pre-mRNAs, eliciting active translation of alternatively spliced (alt-) mRNAs in PCs harboring biallelically VDJ-rearranged IgH alleles.

High rate of NAS of IgH pre-mRNAs in PCs

To assess the regulation of NAS of IgH pre-mRNAs at the molecular level, alt-mRNA amounts were quantified in B and PCs. For this quantification, IgH^frVκ knock-in mice were crossed with DH-LMP2A mice harboring a replacement of JH segments by the Epstein−Barr virus LMP2A gene under the control of a DH promoter.¹³ The expression of LMP2A drives B-cell development and PC differentiation.^13,14 In heterozygous IgH^{DH-LMP2A/frVκ} mice, VDJ-rearranged IgH transcripts arise solely from the nonproductive IgH^frVκ allele, facilitating the analysis of NAS of IgH^frVκ pre-mRNAs. B cells (B220⁺/CD138⁻) and PCs (B220⁻/CD138⁺) were isolated from the spleens of IgH^{DH-LMP2A/frVκ} mice (Fig. 2a), and RT-PCR analysis revealed high alt-mRNA amounts in PCs whereas a faint band was detected in B cells (Fig. 2b). Next, the rate of NAS of IgH^frVκ transcripts was determined by qPCR and corresponded to alt-mRNA/pre-mRNA ratios. Likewise, this NAS rate was strongly increased in PCs compared with B cells (Fig. 2c). We then analyzed the VDJ repertoire of alt-mRNAs, as described.¹⁵ Highly diverse productive CDR3 junctions were obtained among the 11,506 clonotypes identified using IMGT/HighV-QUEST¹⁶ (Fig. 2d). In addition, VDJ junctions involved numerous VH and DH families and all the JH segments (Supplementary Fig. 1). The near absence (0.35%) of nonproductive CDR3 junctions reflected the strong degradation of PTC-containing Igµ mRNAs⁶ and the coupling of the NMD and NAS pathways.¹⁷ This repertoire analysis revealed that alt-mRNAs arose from polyclonal PCs rather than rare clones selected after splice site mutations. Hence, a marked upregulation of NAS of IgH pre-mRNAs occurs during PC differentiation.

Fig. 2.

Comparative analysis of NAS of IgH pre-mRNAs in B cells and PCs. a B cells and PCs were isolated from the spleens of IgH^{DH-LMP2A/frVκ} mice after staining with anti-B220 and anti-CD138 mAbs. Representative dot plot and gates used for cell sorting are depicted. The purity of sorted populations was always above 90%. b To identify alternatively spliced mRNAs (alt-mRNAs), RT-PCR were performed using VH₇₁₈₃F/CµR primers. One representative experiment out of three performed is shown. c Alt-mRNA and IgH^frVκ pre-mRNA levels were assessed by qPCR using VH₇₁₈₃F/CµR primers and the probe a respectively, as described in Fig. 1a. NAS rates corresponding to the relative alt-mRNA/pre-mRNA ratios were determined in sorted B cells (empty bar) and PCs (gray bar) (n = 5). d VDJ repertoire analysis of alt-mRNAs showing polyclonal CDR3 junctions. RACE-PCR, high throughput sequencing and CDR3 length analysis were performed as described in methods and in Fig. S1. Unpaired two-tailed Student’s t test was performed to determine significance (*P < 0.05)

Elevated expression of VDJ-rearranged IgH alleles and RNA splicing genes during PC differentiation

Many studies have reported that the rate of transcription by RNA polymerase II (RNAPII) can regulate alternative splicing.^18,19 To address this question, the transcription of nonproductive IgH^frVκ alleles was evaluated by quantifying pre-mRNA levels in B cells (B220⁺/GL7⁻/FAS⁻/CD138⁻), germinal center (GC-) B cells (B220⁺/GL7⁺/FAS⁺/CD138⁻), and PCs (B220⁻/CD138⁺) isolated from spleens of SRBC-immunized IgH^wt/frVκ mice (Fig. 3a). Consistent with the increase in Ig gene transcription during PC differentiation,²⁰ we found approximately threefold more IgH^frVκ pre-mRNA levels in PCs than in B cells (Fig. 3b). Next, we performed a comparative analysis of RNAPII binding on productive and nonproductive IgH alleles in LPS-stimulated B cells isolated from IgH^wt/frVκ mice. Supporting a biallelic IgH transcription pattern, chromatin immunoprecipitation (ChIP) experiments revealed similar phosphorylated serine 2 (Ser2P) and 5 (Ser5P) RNAPII levels on both IgH alleles (Fig. 3c, d). The nuclear location of IgH alleles was further analyzed in B cells, LPS-stimulated B cells, GC-B cells, and PCs by DNA-FISH using IgH-specific probes (Fig. 3e). We observed that asymmetric heterochromatin recruitment of one IgH allele was extremely rare (<20%) in the vast majority of B-lineage cells, including terminally differentiated PCs.

Fig. 3.

Biallelic transcription of IgH alleles during PC differentiation. a B cells, germinal center (GC-) B cells, and PCs were isolated from the spleens of SRBC-immunized IgH^wt/frVκ mice after staining with anti-B220, anti-GL7, anti-FAS, and anti-CD138 mAbs. Representative dot plots and gates used for cell sorting are depicted. The purity of each sorted population was always above 90%. b Relative IgH^frVκ pre-mRNA levels were determined by qPCR as described in Fig. 1a (probe a) after normalization to Gapdh mRNA expression. The value obtained for B cells served as a reference and was set to 1 in each FACS sorting experiment (n = 3). c, d Resting B cells were isolated from the spleens of IgH^wt/frVκ mice and stimulated with LPS (1 µg/ml) for 3 days. ChIP experiments were performed using phospho-specific antibodies against phosphorylated Ser2 (Ser2P) and Ser5 (Ser5P) RNAPII (n = 4). Relative RNAPII enrichments (percent input) were analyzed on productive and nonproductive IgH alleles by qPCR, as described in Fig. 1a (probes b and a to detect IgH^wt and IgH^frVκ alleles, respectively). Background signals from mock samples with irrelevant antibody staining were subtracted. e Map of the IgH locus detailing genomic regions covered by BAC used as FISH probes. Representative 3D-FISH images of B-cell nuclei from B6 mice, reporting no allele or 1 or 2 alleles within heterochromatin (scale bars: 5 µm). IgH alleles are indicated by white arrows. Bars represent the percentage of IgH alleles within the heterochromatin, determined in resting B cells (n = 137), LPS-stimulated B cells (n = 97), GC-B cells (n = 52) and PC (n = 53) nuclei. For each cell type, data were obtained from three individual mice. Unpaired two-tailed Student’s t test was performed to determine significance (ns not significant; *P < 0.05)

Next, the expression of genes involved in RNA splicing (Gene ontology ID GO:0008380, n = 360) was assessed in purified follicular (FO-) B cells, GC-B cells, plasmablasts (PBs), and PCs using available microarray dataset GSE26408 and GenomicScape platform (www.genomicscape.com).²¹ Multiclass analysis revealed that the expression profile of RNA splicing regulators is highly modified during PC differentiation, with 249 (∼70%) genes differentially expressed, and among them, 106 (>40%) were significantly overexpressed in PCs (fold change > 1.5 and P < 0.05) (Fig. 4; Supplementary Table 1). These data underline specific gene expression patterns related to RNA splicing at the different stages of B-cell maturation with a significant enrichment of genes related to the spliceosomal complex in PCs as identified by gene set expression analysis.

Fig. 4.

Analysis of RNA splicing gene expression profiles during PC differentiation. Clustergram of GO RNA splicing gene list significantly differentially expressed between follicular B cells (FO B cells, n = 3), germinal center B cells (GC B cells, n = 5), plasmablasts (PBs, n = 4), and plasma cells (PCs, n = 3) (fold change > 1.5 and FDR < 0.05)

Thus, biallelic hypertranscription of IgH alleles together with huge modifications of the RNA splicing environment are triggered during PC differentiation, and this was correlated with the reinforcement of NAS of IgH pre-mRNAs.

Delineating the chromatin profile of the alternative frVκ exon

To evaluate the interplay between NAS and chromatin structure, a comparative analysis of specific histone H3 marks was conducted among the alternative frVκ exon, constitutive CH1µ exon, and intronic “JH-Eµ” sequence (Fig. 5a). ChIP experiments were performed using anti-H3K4me3, H3K9me3, H3K9ac, and H3K36me3 antibodies in LPS-stimulated B cells isolated from IgH^{wt/ frVκ} mice. H3K4me3, H3K9ac, and H3K36me3 marks are associated with transcriptional activation, whereas H3K9me3 is linked to transcriptional repression.^22,23 Remarkably, the alternative frVκ exon and the “JH-Eµ” intron exhibited highly similar chromatin profiles for all epigenetic marks investigated (Fig. 5b−e). By contrast, alternative frVκ and constitutive CH1µ exons displayed striking differences. The levels of H3K4me3 and H3K9ac increased (Fig. 5b, c), whereas H3K9me3 amounts decreased in the alternative frVκ exon compared with the constitutive CH1µ exon (Fig. 5e). The comparative analysis of H3K36me3 levels did not reveal any significant changes (Fig. 5d). Supporting the idea that epigenetic histone marks can influence the pattern of alternative splicing by favoring exclusion or inclusion of alternative exons,^22,24–26 skipped frVκ exons exhibited an “intron-like” chromatin signature.

Fig. 5.

Analysis of the alternative frVκ exon chromatin profile. a Chromatin marks were analyzed comparatively at the JH4-Eµ intron, the alternative frVκ exon, and the constitutive CH1µ exon. Primers and probes used in qPCR are depicted. b−e ChIP assays were performed in splenic B cells isolated from IgH^wt/frVκ mice 3 days after stimulation with LPS (1 µg/ml). Immunoprecipitation experiments were conducted using anti-H3K4me3 (b), H3K9ac (c), H3K36me3 (d), and H3K9me3 (e) mAbs. Background signals from mock IP using irrelevant antibodies were subtracted, and relative enrichments were expressed after normalization to total input DNA (percent of input). Data are the mean ± SEM (n = 5), and unpaired two-tailed Student’s t test was performed to determine significance (ns not significant; **P < 0.01; ***P < 0.001)

ASO-mediated exon skipping of Ig transcripts

In recent years, the considerable improvement in terms of stability and cellular penetration of ASOs has driven the development of these molecules for medical purposes in various genetic diseases and, in particular, for the treatment of neuromuscular disorders.^27–29 By simply monitoring the production of IgM^a following exon skipping events, the IgH^wt/frVκ mouse model represents a unique tool facilitating the analysis of ASO-mediated Ig exon skipping approaches in primary B and PCs. To address this issue, LPS-stimulated B cells were treated with a specific “vivo-morpholino” ASO targeting the frVκ exon donor splice site (ASO-frVκ_dss) (Fig. 6a) or with an irrelevant ASO as a control (ASO-ctrl). Remarkably, the passive administration of ASO-frVκ_dss strongly increased the secretion of IgM^a in culture supernatants (Fig. 6b). We also observed that both the frequency of IgM^a-positive cells (Fig. 6c, d) and the mean fluorescence intensity (MFI) of IgM^a (Fig. 6e) were increased after ASO-frVκ_dss treatment compared with the ASO-ctrl. Based on previous observations showing that biallelic VDJ-rearrangements are retrieved in ∼40% of B-lineage cells³⁰ and that ∼22% of VDJ-rearranged IgH^frVκ transcripts exhibited in-frame CDR3 junctions,⁶ we estimated a theoretical maximum of IgM^a-expressing cells of ∼8.8%. Interestingly, the frequency of PB-expressing IgM^a reached ∼7.5% after ASO-frVκ_dss treatment, indicating that passive administration of ASOs was highly efficient (∼85% of the theoretical maximum) and targeted nearly all antibody-secreting cells (Fig. 6d). Likewise, we found that ASO-frVκ_dss treatment drastically increased the rate of frVκ exon skipping by monitoring alt-mRNA/pre-mRNA ratios (Fig. 6f). Of note, we observed similar proliferative responses (Supplementary Fig. 2a) and Igκ light chain production (Supplementary Fig. 2b) in LPS-stimulated B cells treated or not with ASOs, indicating that ASO treatment did not induce adverse toxic effects. Hence, passive administration of ASO targeting V exon donor splice sites strongly induced exon skipping in PCs.

Fig. 6.

Significant IgM^a production upon passive administration of ASOs. a An ASO sequence targeting the donor splice site of the frVκ exon (AON-frVκ_dss) was designed and synthetized as “vivo-morpholino ASO” (Gene Tools, LLC). b Splenic B cells isolated from IgH^wt/frVκ mice were stimulated with LPS (1 μg/ml) for 4 days and treated with ASO-frVκ_dss or irrelevant ASO (ASO-Ctrl) during the last 2 days, as described in the Methods. At day 4, supernatants were harvested, and IgM^a levels were determined. c Cells were labeled with anti-B220, anti-CD138, anti-IgM^b, and anti-IgM^a mAbs. Representative dot plots and gates used for FACS analysis are depicted. d The left y-axis represents the percentage of IgM^a-expressing cells in IgM^b-positive B cells and plasmablasts treated with ASO-frVκ_dss (gray bars) or ASO-Ctrl (empty bars). The right y-axis shows ASOs’ efficiency as percentage of theoretical maximum of IgM^a-positive cells. e The mean fluorescence intensity (MFI) of IgM^a was determined on IgM^b-positive B cells and PBs. f Alt-mRNA/pre-mRNA ratios were determined in IgH^{DHLMP2A/frVκ} LPS-stimulated B cells, as described in Fig. 2c. Passive ASO administration was performed as described in panel (b). Data are representative of five independent experiments. Bars represent the mean expression ± SEM. Unpaired two-tailed Student’s t test was performed to determine significance (*P < 0.05; **P < 0.01 ****P < 0.0001)

Discussion

Alternative splicing regulates gene expression and affects more than 90% of multiexonic pre-mRNAs in humans.^31–33 Likewise, ASO-mediated splicing modulation can be exploited for numerous therapeutic purposes.³⁴ In this study, we provide evidence that the biallelic boost of Ig gene transcription accompanying PC differentiation promotes NAS of VDJ-rearranged IgH pre-mRNAs and facilitates ASO-mediated skipping of V exons.

Several studies including ours have shown that the pattern of IgH gene transcription is mostly biallelic in B cells and PCs.^6,20,35–37 Accordingly, we found that ∼80% of mature B cells and PCs exhibited no IgH allele in heterochromatin. This observed frequency is consistent with previous 3D RNA FISH data showing that >80% of IgH and Igκ genes are biallelically transcribed in PCs.²⁰ ChIP experiments further revealed similar Ser2P and Ser5P RNAPII binding on both IgH alleles in LPS-stimulated B cells. In agreement with a drastic increase in Ig gene transcription during PC differentiation, we found that the amount of nonproductive IgH pre-mRNAs was ∼3-fold higher in PCs compared with B cells. Altogether, biallelic hypertranscription of Ig genes generates high levels of nonproductively V(D)J-rearranged Ig pre-mRNAs in antibody-secreting cells.

Because the TIE-checkpoint eliminates PC-expressing truncated Ig chains,¹⁰ it remains very challenging to evaluate the extent of exon skipping during splicing of Ig pre-mRNAs. Using our dedicated IgH^frVκ model allowing for skipping of V exon without production of shortened Ig polypeptides, we provided evidence that NAS of IgH pre-mRNAs is greatly enhanced in PCs compared with B cells. These data strongly suggest that NAS and conventional alternative splicing are governed by similar rules and controlled by the transcription elongation rate.^{19,26,38–40} The transcriptomic analysis in Fig. 4 also indicates that PC differentiation entails major changes with regard to the expression of genes involved in RNA splicing. Further investigations are necessary to define the role of differentially expressed RNA splicing regulators. Therefore, increasing Ig gene expression in PCs with biallelic V(D)J-rearrangements favors antibody production but detrimentally induces activation of the TIE-checkpoint after nonsense-associated exon skipping events.

Several studies have identified a specific histone mark signature around alternative exons.^22,24–26 Interestingly, trimethylation and acetylation of H3K9 have been associated with inclusion or exclusion of alternative exons, respectively.^19,25 High levels of H3K9me3 promotes heterochromatin protein 1γ (HP1γ) recruitment and provokes a local slowdown of RNAPII, supporting the inclusion of CD44 variant exons.²⁵ By contrast, a local increase in H3K9ac is associated with the exclusion of NCAM (neural cell adhesion molecule) exon 18 during the depolarization of human neuronal cells.¹⁹ Remarkably, we observed a massive enrichment in H3K9ac and an opposite decrease in H3K9me3 at the skipped frVκ exon compared with the constitutive CH1µ exon. Altogether, this chromatin profile revealed that the alternative frVκ exon is turned into a pseudo-intron.

This study provides evidence that NAS of Ig pre-mRNAs is coupled to transcription and increases as PC differentiation proceeds. We propose that the late occurrence of TIE-induced apoptosis eliminating PCs with biallelic Ig gene rearrangements results from the combined effects of enhanced NAS activation during splicing of nonproductive Ig pre-mRNAs and the intrinsic sensitivity of antibody-secreting cells towards prolonged ER stress. Altogether, these data reinforce the idea that nonproductively V(D)J-recombined alleles can be considered as drivers rather than passengers at late steps of B-cell differentiation.

Finally, we validated that ASO-mediated Ig exon skipping can be easily achieved by targeting the donor splice site of V exons. Thus, the use of ASOs to induce massive production of detrimental truncated Ig chains represents an attractive therapeutic approach that should be of great interest in patients with monoclonal gammopathies.

Materials and methods

Mice

IgH^frVκ mice harboring an extra frameshift-inducing Vκ exon between J_H4 and Eµ have been described elsewhere previously.⁶ Heterozygous IgH^wt/frVκ mice were backcrossed to C57BL/6 (B6) background to distinguish between IgH^wt (b allotype) and IgH^frvκ (a allotype) alleles. F1 (B6^b/129^a) mice were used as a control. IgH^DH-LMP2A mice harbor a replacement of JH segments by the Epstein–Barr virus LMP2A gene under the control of a DH promoter.¹³ The expression of LPM2A mimics the BCR tonic signal, allowing for B-cell development and PC differentiation.^13,14 Two- to three-month-old mice were used in all experiments and maintained in our animal facilities, at 21–23 °C with a 12-h light/dark cycle. Experiments were performed according to the guidelines of our institutional review board for animal experimentation (No. CREEAL 6-07-2012).

ASO treatments

Vivo-morpholino ASOs (ASO-frVκ_dss: 5′-ATGCTCGAGACTTACCCGTTTGATT-3′) and an irrelevant ASO (ASO-Ctrl: 5′-CCTCTTACCTCAGTTACAATTTATA-3′) were designed and purchased at Gene Tools, LLC. Splenic B cells isolated from IgH^wt/frVκ mice were purified by negative selection using anti-CD43 magnetic beads (Miltenyi Biotec) and stimulated (0.8×10⁶ cells/ml) with 1 μg/ml lipopolysaccharide (LPS) (LPS-EB Ultrapure; InvivoGen) in RPMI 1640 with 10% fetal calf serum. At day 2, cells were harvested and incubated with 10 μM AON (4 h in PBS). After incubation, the cells were stimulated with LPS for 2 days in the presence of AON (1 µM) in culture media.

Flow cytometry and cell sorting

Erythrocyte-depleted spleen cells were labeled with anti-mouse CD138 (281-2; BD Pharmingen), B220 (RA3-6B2; BioLegend), anti-mouse- and B-cell Activating Ag (GL7; BD Pharmingen) and anti-mouse CD95 (Jo2) mAbs (BD Pharmingen). Intracellular staining was performed using anti-mouse IgM^a (DS-1; BD Pharmingen) and IgM^b (AF6-78; BD Pharmingen) mAbs, according to the manufacturer’s instructions. Cells were analyzed on an LSR II Fortessa apparatus using FACS-Diva software (BD Pharmingen). B cells, GC-B cells, and PCs were sorted from the spleens of B6 or IgH^wt/frVκ mice 7 days after intraperitoneal (i.p.) injection of sheep red blood cells (SRBCs, BioMérieux^®SA-France). Cell sorting was performed using a FACS-Aria III (BD Biosciences), and the gates used are depicted in dot plots (purity was above 90%).

ELISA assays

Sera from IgH^wt/frVκ (wt^b/frVκ^a), B6 (wt^b/wt^b), and B6/129 F1 (wt^b/wt^a) mice were collected and analyzed using total IgM or allotype-specific IgM^a and IgM^b ELISA assays. Ig titers were determined in polycarbonate 96-multiwell plates (Maxisorp, Nunc) using unlabeled IgM (Southern Biotech), IgM^a (DS-1; BD Pharmingen), and IgM^b (AF6-78; BioLegend) Abs and alkaline phosphatase-conjugated goat anti-mouse IgM Abs (Southern Biotech), as described.^41,42 Ig amounts were revealed by the addition of p-nitrophenyl phosphate (Sigma-Aldrich) and blocked with the addition of 3 M NaOH; optic density was measured at 405 nm.

Chromatin immunoprecipitation experiments

ChIP experiments were performed using anti-H3K4me3 (Millipore, 07-473), anti-H3K9ac (Millipore, 06-942), anti-H3K9me3 (Millipore, 05-1242), anti-H3K36me3 (Abcam, ab9050), anti-RNA Pol II ser2P (Abcam, ab5095), and anti-RNA Pol II ser5P (Abcam, ab5131), as previously described.⁴³ In brief, 1×10⁷ LPS-stimulated B cells from IgH^wt/frVκ mice were harvested at day 3, washed twice in PBS and cross-linked at 37 °C for 15 min in 15 ml of PBS with 1% formaldehyde. The reaction was quenched with 0.125 M glycine. After lysis, chromatin was sonicated to 0.5–1 kb using a Vibracell 75043 (Thermo Fisher Scientific). After dilution in ChIP buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH 8.1, and 167 mM NaCl), chromatin was precleared by rotating for 2 h at 4 °C with 100 μl of 50% protein A/G slurry (0.2 mg/ml sheared salmon sperm DNA, 0.5 mg/ml BSA, and 50% protein A/G; Sigma), 0.3−0.5×10⁶ cell equivalents were saved as input, and 3−5×10⁶ cell equivalents were incubated overnight with specific or control antibodies. Immune complexes were precipitated by the addition of protein A/G. Cross-linking was reversed by overnight incubation (70 °C) in TE buffer with 0.02% SDS, and genomic DNA was obtained after phenol/chloroform extraction.

PCR and RT-PCR

Total RNA was prepared using Tri-reagent (Invitrogen) procedures. RT-PCR was carried out on 1−3 µg of DNase I (Invitrogen)-treated RNA using Superscript III (Invitrogen). Priming for reverse transcription was done with random hexamers. Quantitative PCR was performed on cDNA samples equivalent to 10 ng of RNA per reaction, using TaqMan or SYBR-Green Universal MasterMix (Applied Biosystems) on a StepOnePlus Real-Time PCR system (Applied Biosystems). Transcripts were quantified according to the standard 2^−ΔΔCt method after normalization to Gapdh (Mm99999915_g1). Primers and probes used for Q-PCR are listed in Supplementary Table 2.

3D-DNA FISH experiments

A DNA probe specific to the IgH locus was prepared with the RP23-109B20 BAC construct (Life Technologies) labeled by random priming with digoxigenin-11-dUTP (Roche). A DNA probe specific to heterochromatin compartment was prepared with the pγSat plasmid (kindly provided by Dr. Nial Dillon) containing multiple copies of the 234 bp major satellite (γ-satellite) repeat labeled by random priming with Alexa Fluor 488-dUTP (Life Technologies). Before hybridization, the probes were denatured in 2× SSC/50% formamide/10% dextran sulfate hybridization buffer at 95 °C for 5 min. Resting B cells, LPS-stimulated B cells, GC-B cells, and PCs were dropped onto poly-l-lysine slides and fixed with PBS/4% paraformaldehyde for 10 min at room temperature (RT). After washing steps with PBS, cells were permeabilized using pepsin 0.02%/HCl 10 mM for 15 min at RT. Cells were washed with PBS, post-fixed with PBS/1% paraformaldehyde for 5 min at RT, washed with PBS/0.2 M MgCl₂, dehydrated by successive incubations in ethanol solutions (70, 90, 100%, 2 min each), denatured for 5 min in 2× SSC/70% formamide at 72 °C, dehydrated by successive incubations in cold ethanol solutions (70, 90, 100%, 2 min each) and hybridized overnight at 37 °C using 200 ng of each labeled probe in hybridization buffer in a dark humid chamber under coverslips. The next day, the slides were washed three times with 1× SSC, and a blocking step was performed by incubating slides in 4× SSC/3% BSA for 30 min followed by incubation with rhodamine-conjugated anti-digoxigenin secondary antibody (Roche) diluted 1/200 in 4× SSC/3% BSA for 60 min. Slides were washed four times in 2× SSC and mounted under coverslips using ProLong Gold antifade reagent containing DAPI (Life Technologies). Images were acquired along the z-axis with an epifluorescence microscope (Nikon Eclipe Ni-E) with a ×100 oil objective. Fifty-one optical sections separated by 0.2 µm were collected, and stacked images were deconvoluated using Huygens software (SVI) and analyzed using Volocity 3D Image Analysis software (PerkinElmer). Separation of alleles was measured in 3D from the center of mass of each signal. Volumetric pixel size was 0.064 μm in the xy and 0.2 μm in the z-direction.

Bioinformatics analysis

Gene expression profiles of purified mouse follicular (FO-) B cells, GC-B cells, PBs, and PCs were obtained from the publicly available dataset GSE 26408 (Gene Expression Omnibus database).⁴⁴ Affymetrix Mouse Genome 430A 2.0 Array gene expression data were analyzed with the GenomicScape bioinformatics platform (www.genomicscape.com).²¹ Clustering was performed and visualized with Cluster and TreeView.⁴⁵ Differentially expressed genes between cell populations were identified with the significance analysis of microarray statistical method (fold change ≥ 1.5, false discovery rate ≤ 0.05).⁴⁶

High throughput sequencing of the IgH repertoire

Repertoire sequencing was performed as previously described.⁴⁷ In brief, RNA (500 ng) was extracted from sorted PCs of IgH^{DH-LMP2A/frVκ} mice, and RACE-PCR (rapid amplification of cDNA-ends by polymerase chain reaction) was performed using a CH1µ reverse primer.⁴⁸ Sequencing adapter sequences were added by primer extension, and amplicons were sequenced on an MiSeq sequencing apparatus (Illumina). Repertoire was analyzed using IMGT/High-V-Quest⁴⁹ and the bcRep R package⁵⁰ available on the IMGT Web site (https://www.imgt.org).

Statistical analysis

The results are expressed as the mean ± standard error of the mean (SEM), and overall differences between variables were evaluated by two-tailed unpaired Student’s t test using Prism GraphPad software (San Diego, CA).

Author contributions

M.O.A. and N.S. designed and performed experiments, analyzed data and wrote the paper. J.-M.L. and A.M. contributed data to Fig. 6. O.M., S.L.N., and E.P. contributed 3D DNA-FISH data in Fig. 3. J.M. performed gene expression analysis in Fig. 4. M.V.A., J.S., C.S., and M.C. helped with the experiments and interpretation of the data. L.D. conceived the project, designed experiments, analyzed data and wrote the paper.

Competing interests

The authors declare no competing interests.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Electronic supplementary material

The online version of this article (10.1038/s41423-018-0160-6) contains supplementary material.