IL-6 produced by immune complex-activated follicular dendritic cells promotes germinal center reactions, IgG responses and somatic hypermutation

Abstract

Reports that follicular dendritic cells (FDCs) produce IL-6 prompted the hypotheses that immune complexes (ICs) induce FDCs to produce IL-6 and that FDC–IL-6 promotes germinal center (GC) reactions, somatic hypermutation (SHM) and IgG production. FDCs were activated in vitro by addition of ICs and FDC–IL-6 production was determined. Wild-type (WT) and IL-6 knockout (KO) mice, as well as chimeras with WT and IL-6 KO cells, were immunized with (4-hydroxy-3-nitrophenyl)-acetyl (NP)–chicken gamma globulin (CGG) and used to study anti-(4-hydroxy-3-iodo-5-nitrophenyl) acetyl (NIP) responses, GC formation and SHM in the VH186.2 gene segment in Ig-gamma. FDC–IL-6 increased when FDCs encountered ICs. At low immunogen dose, 1 μg NP–CGG per mouse, the IgG anti-NIP response in IL-6 KO mice was low and immunohistochemistry revealed a reduction in both the number and size of GCs. The physiological relevance of FDC–IL-6 was apparent in the chimeric mice where total splenocytes from WT mice were unable to provide the IL-6 needed for normal IgG and GC responses in IL-6 KO animals with IL-6-defective FDCs. Moreover, the rate of mutation decreased from 18 to 8.9 mutations per 1000 bases (P < 0.001) in WT versus IL-6 KO mice. Addition of anti-IL-6 to GC reactions in vitro reduced antibody levels and SHM from 3.5 to 0.65 mutations per 1000 bases (P < 0.02). Thus, the absence of FDC–IL-6 correlated with a reduction in SHM that coincided with the reduction in GCs and specific anti-NIP. This is the first study to document that ICs induce FDC–IL-6 and that FDC-derived IL-6 is physiologically relevant in generating optimal GC reactions, SHM and IgG levels.

Introduction

Follicular dendritic cells (FDCs) are prominent in germinal centers (GCs) because their numerous long slender dendrites intertwine and create extensive FDC networks or reticula. These FDC networks are fixed in the follicles while T and B cells are free to circulate. Nevertheless, FDCs release chemokines that attract recirculating lymphocytes that help organize the follicle and participate in the GC reaction (1, 2). FDCs, bearing specific antigen in the form of immune complexes (ICs), are requisite for full development of GCs (3–5) and interaction of FDCs with ICs leads to activation and the expression of high levels of FDC-FcγRIIB, -intercellular adhesion molecule (ICAM)-1 and -vascular cell adhesion molecule (VCAM)-1 (6). FDC-ICs also deliver a late antigenic signal that promotes somatic hypermutation (SHM) (7). Important events include (i) induction of specific antibody that binds immunogen and forms ICs, (ii) IC trapping by FDCs, (iii) GC B cell stimulation by antigen in these FDC-ICs and (iv) promotion of SHM by this late second antigen signal delivered by IC-bearing FDCs (7). However, FDCs provide signals, beyond specific antigen, that promote B cell maturation and we sought to determine if FDC–IL-6 promotes SHM.

IL-6, known in the early literature as B cell stimulatory factor-2, is recognized for its role in regulating terminal B cell differentiation [reviewed in (8)]. Somatic hypermutation (SHM) is a major event in terminal B cell differentiation that occurs in GCs in the second week after primary immunization [reviewed in (9–12)]. It has also been reported that murine FDCs are the source of IL-6 in GCs (13) and that ICs activate FDCs (6). This prompted the hypothesis that IC-activated FDCs produce elevated levels of IL-6 and that FDC-derived IL-6 is important not only for optimal antibody responses but also for GC reactions and SHM. We postulated that elevated production of IL-6 by IC-activated FDCs, 1–2 weeks after primary immunization, would correlate with enhanced GC reactions, IgG responses and SHM.

While there are numerous reports supporting the hypothesis that FDCs produce IL-6 (14–17), there are also studies reporting the inability to detect IL-6 in FDCs (18–20). To further test the hypothesis that FDCs produce IL-6 and that IC-activated FDCs produce elevated levels of IL-6, ICs were added to purified FDCs and 2–6 days later, IL-6 production had clearly increased. Specific antibody is induced 4–7 days after primary immunization and IC formation would follow soon thereafter. Thus, 6–13 days after immunization, a burst in FDC–IL-6 production would be expected in developing GCs. To test the postulate that IL-6 promotes GC reactions, specific IgG responses and SHM, these responses were studied in IL-6 knockout (KO) mice and in GC reactions in vitro where IL-6 was specifically inhibited by anti-IL-6. The present study, including both in vivo and in vitro experiments, confirms earlier results indicating that optimal GC reactions and IgG anti-(4-hydroxy-3-iodo-5-nitrophenyl) acetyl (NIP) responses require IL-6 and that FDCs are the only cells in GC reactions making IL-6 (13). In addition, we found that IL-6 was not detectable in in vitro GC reactions with IL-6 KO FDCs with T and B cells from wild-type (WT) mice. In contrast, IL-6 production was normal in GC reactions with WT FDCs with T and B cells from IL-6 KO mice. The absence of IL-6 in cultures lacking WT FDCs resulted in marked reduction in the rate of SHM that coincided with the reduction in specific anti-NIP. Moreover, GCs were abundant in irradiated WT mice reconstituted with spleen cells from IL-6 KO mice while GCs were virtually undetectable in irradiated IL-6 KO mice reconstituted with normal spleen cells. These data provide strong support for the physiological relevance of FDC–IL-6 in GC reactions in vitro and in vivo and report for the first time that FDC–IL-6 is inducible by ICs and is involved not only in influencing the amount but also the mutations that are known to enhance the affinity of specific IgG produced.

Methods

Mice, antigen and immunization

C57BL/6 mice (6–8 weeks old) were purchased from the National Cancer Institute and female IL-6 KO mice (B6.129S2-116tm1Kopf/J) of the same age were obtained from the Jackson Laboratory. The mice were housed in standard shoebox cages and given food and water ad libitum. Animals were handled in compliance with the Virginia Commonwealth University Institutional Animal Care and Use Committee guidelines. Mice were immunized by injecting 0.25 or 25 μg of alum-precipitated [(4-hydroxy-3-nitrophenyl)-acetyl (NP)]₃₆–chicken gamma globulin (CGG) with ∼1.5 × 10⁷ heat-killed Bordetella pertussis subcutaneously in each front leg and hind foot in a 50 μl volume to give a total of 1 or 100 μg of (NP)₃₆–CGG per animal. Fourteen days later, these mice were bled, serum collected and draining lymph nodes from each group were pooled to isolate lambda light-chain-positive B cells (λ⁺ B cells) for extracting RNA. The serum was used to determine NIP-specific antibody levels and the RNA was used to determine mutations per 1000 bases in the VH186.2 gene segment. In vitro GC reactions were set up using memory T cells specific for CGG [T_(CGG) cells] isolated from mice 1 month after immunization with 100 μg CGG as described above. To get NP-specific λ⁺ B cells, WT or IL-6 KO mice were immunized with 100 μg (NP)₃₆–CGG plus heat-killed B. pertussis as explained above and the λ⁺ B cells were isolated 6 days later.

Establishment of IL-6 KO/WT chimeras, immunization and immunohistochemistry

Two days after irradiation with 600 rads, IL-6 KO (B6.129S2-116tm1Kopf/J) mice were reconstituted with 10⁸ WT or IL-6 KO splenocytes injected subcutaneously behind the neck. Similarly, WT C57B/6 mice were reconstituted with 10⁸ IL-6 KO or WT splenocytes and 48 h later, mice were immunized with 1 μg (NP)₃₆–CGG per animal. Fourteen days later, these mice were bled, sacrificed and the spleens were frozen in optimal cutting temperature medium. Sera were used for determination of the anti-NIP levels, and 10 μm cryostat spleen sections were prepared and fixed in absolute acetone. The mid-sagittal spleen sections were labeled for GC B cells with FITC-conjugated anti-B cell activation marker GL-7 and for the FDC-ICs with Rhodamine-Red-X anti-mouse IgG. Resting B cells were labeled with PerCP–cyanine 5.5 (Cy5.5) anti-mouse B220 and the number of GL-7⁺ GCs were counted in the mid-sagittal spleen sections.

Antibodies and reagents

Functional grade, azide-free, sterile-filtered, purified anti-mouse IL-6 (Cat# 16-7061, Clone MP5-20F3) and rat IgG isotype control antibody were obtained from eBioscience. Rat anti-mouse FDC (FDC-M1), biotin mouse anti-rat kappa (MRK-1), anti-mouse CD21/CD35 (Clone 7G6) and anti-mouse CD32/CD16 (Clone 2.4G2) were purchased from PharMingen (San Diego, CA, USA). Mouse CD45R (B220) MicroBeads, mouse CD90 (Thy1.2) MicroBeads, anti-Biotin MicroBeads and MACS LS columns were purchased from Miltenyi Biotec GmbH (Auburn, CA, USA). Biotin-labeled rat anti-mouse kappa was purchased from Zymed (San Francisco, CA, USA). Alkaline phosphatase-labeled goat anti-mouse IgG (H + L) and p-nitrophenylphosphate substrate were obtained from Kirkegaard & Perry Laboratories (Gaithersburg, MD, USA). NIP-ovalbumin (OVA) (NIP conjugated to OVA) and (NP)₃₆–CGG were obtained from Biosearch Technologies (Novato, CA, USA). Anti-CGG was obtained from hyperimmunized mice with serum anti-CGG IgG levels in excess of 1 mg ml⁻¹.

Flow cytometric analysis of B cells including GC (GL-7+) B cells

GC B cells were identified with flow cytometry using FITC-labeled anti-GL-7 (553666; BD PharMingen) and anti-CD19 [phycoerythrin (PE) labeled, 553786; BD PharMingen]. GL-7 is an activation marker and activated T cells may label. We found that GL-7 data were easier to interpret if Thy-1-positive cells were removed using mouse CD90 (Thy1.2) MicroBeads. Viable B cells persisting in cultures were identified by double labeling with B220–FITC (553088; BD PharMingen) and propidium iodide (PI) exclusion. Data were collected using FC500 Flow Cytometer Beckman Coulter, analyzed using Cytomics RXP analysis software and plotted using WinMDI software (Scripps Research Institute).

FDC isolation, activation with ICs in vitro and assessment of IL-6 production

FDCs were isolated as described previously (21). Briefly, lymph nodes from lethally irradiated mice were gently digested and the cell suspension incubated with rat anti-mouse FDC-M1, biotin mouse anti-rat kappa (Clone MRK-1) and anti-biotin magnetic microbeads sequentially. Finally, the FDCs were positively selected using a magnetic column (MACS). Approximately 85–95% of these cells expressed the FDC phenotype, FDC-M1⁺, CD32⁺, CD21/35⁺, CD40⁺ and ICAM⁺ (21). Purified FDCs were activated by incubating 1 × 10⁶ cells with 100 ng ml⁻¹ OVA-ICs made of OVA/rabbit anti-OVA at a ratio of 1:6. Four independent experiments with triplicates of IC-stimulated FDCs were set up and culture supernatant fluids were sampled after 2, 4 and 6 days. In some experiments, culture supernatant fluids were collected after 3, 6 and 10 days. Controls of B220-purified B cells, Thy1.2-purified T cells, IL-6 KO FDCs, WT B and T cells, WT B and T cells with IL-6 KO FDCs with or without ICs, WT FDCs incubated with IL-6 KO B and T cells with or without ICs, WT FDCs incubated with soluble 100 ng ml⁻¹ OVA, WT FDCs pre-treated with 10 μg ml⁻¹ azide-free FcγR or CR blockers (mAb 2.4G2 and 7G6, respectively) or their isotype controls prior to IC loading and low-Tox complement were included. IL-6 levels in the fluids were assessed using a Bio-Plex Mouse IL-6 assay (Bio-Rad, Hercules, CA, USA; 171-G10738), data were collected on a Bio-Plex array reader (#LX10004042104) and analyzed using Bio-Plex Manager Software 4.1.

Isolation of λ⁺ B cells and memory T_(CGG) cells

Single-cell suspensions were prepared from draining lymph nodes after priming with (NP)₃₆–CGG or CGG as explained above. After depleting kappa light-chain-positive B cells with biotin-anti-κ and anti-biotin microbeads with a MACS column, the persisting λ⁺ B cells were isolated with anti-B220 microbeads using a MACS column. These preparations of isolated cells were >98% λ⁺ B cells. Memory T_(CGG) cells from CGG-primed mice were incubated with 40 μl of anti-CD90.2 microbeads and the T cells were positively selected using a MACS column.

In vitro GC reactions and the anti-NIP antibody response

GC reactions were initiated in vitro by culturing 1 × 10⁶ unmutated but 6-day primed λ⁺ B cells, 0.5 × 10⁶ T cells and 0.5 × 10⁶ FDCs in the presence of 100 ng of (NP)₃₆–CGG as free antigen or ICs. The ICs were made by incubating NP–CGG with anti-CGG at a ratio of 6 ng ml⁻¹ anti-CGG to 1 ng of NP–CGG for 2 h and then adding IC containing 100 ng NP–CGG to appropriate wells in 24-well culture plates. Supernatant fluids were harvested 7 days later and were assayed for NIP-specific IgG antibody by ELISA using NIP–OVA to coat the plate.

Analysis of mutations in the V region

Anti-NP Igs predominantly use the VH186.2 gene (22, 23) and it was analyzed in this study. Cells were collected on day 7 of culture and RNA was isolated using Trizol (Invitrogen, Carlsbad, CA, USA). The RNA was then reverse transcribed to cDNA with GeneAmp gold RNA PCR Core kit (PE Biosystems, Foster City, CA, USA) using oligo dt. The VH186.2 gene followed by D and J regions and a proximal segment of the C gamma gene segment was amplified by two rounds of nested PCR with high-fidelity Pfu-Ultra DNA polymerase (Stratagene, La Jolla, CA, USA). The forward and reverse primers for the first-round PCR were 5′-catgctcttcttggcagcaacagc-3′ and 5′-gtgcacaccgctggacagggatcc-3′ and for second-round PCR were 5′-caggtccaactgcagccag-3′ and 5′-agtttgggcagcaga-3. The PCR was carried out as follows: 95°C 1 min, 56°C 1 min, 72°C 1 min for 30 cycles followed by 72°C for 10 min. The expected 400-bp PCR product was then electrophoresed on 1.2% agarose gel, purified by gel extraction using QIAQuick Kit (Qiagen, Valencia, CA, USA) and cloned using Zero Blunt® TOPO PCR Cloning Kit (Invitrogen). The plasmids were transformed into one shot TOP10 chemically competent cells (Invitrogen) and plated onto Luria-Bertani (LB) media containing 50 μg ml⁻¹ ampicillin. Multiple discrete colonies were picked the following day and grown for 24 h in LB broth containing 50 μg ml⁻¹ ampicillin in a 37°C shaker. The plasmids were isolated using a QIAprep Spin Miniprep Kit (Qiagen) and the insert was sequenced using an ABI sequencer. The sequences were analyzed using GCG Wisconsin Package (Accelrys, San Diego, CA, USA), clustalW (24, 25) and BioEDIT (http://www.mbio.ncsu.edu/BioEdit/bioedit.html) software packages. Redundant sequences were excluded from our calculation of mutations per 1000 bases to avoid multiplying the effect of a single mutation. To avoid including N-terminal additions, changes within five bases of the V–D junction were also excluded from our calculations of mutations per 1000 bases.

Immunofluorescence and confocal microscopy

Draining lymph nodes (popliteal brachial and axillary) were harvested from WT or IL-6 KO mice 14 days after immunizing with 1 μg of (NP)₃₆–CGG and frozen on dry ice in Optimum Temperature Compound (Tissue-Tec, Torrance, CA, USA). Serial 10-μm thin sections were cut from the frozen blocks using a Jung Frigocut 2800E Cryostat, fixed in absolute acetone and air-dried. The sections were blocked with 10% BSA in PBS and then washed and incubated for 1 h with 2 μg ml⁻¹ FITC-conjugated anti-GL-7 and Cy5.5-conjugated anti-B220 (BD PharMingen). Sections were washed, mounted with anti-fade mounting medium, Vectashield, coverslipped and examined with a Leica TCS-SP2 AOBS confocal laser scanning microscope fitted with an oil plan-Apochromat ×40 objective. Two lasers were used, Argon (488 nm) for FITC and HeNe (633 nm) for Cy5.5 (shown as pseudocolor magenta). Parameters were adjusted to scan at 512 × 512 pixel density and 8-bit pixel depth. Emissions were recorded in two separate channels. Digital images were captured, overlaid and processed with Leica Confocal and LCS Lite softwares. BIOQUANT NOVA Advanced Image Analysis software (R&M Biometrics, Nashville, TN, USA) was used to analyze the immunohistochemistry as described previously (26). After setting an arbitrary threshold, the system allowed us to determine number, position, area and density of labeling that best defined FDC reticula. This predefined threshold was used throughout to enable reliable comparisons.

Statistical analysis

The number of mutations per sequence was compared using a Poisson regression analysis (SAS Proc Genmod version 9.1, SAS Institute Inc.) An offset of log(306) was used to normalize the fitted group means to the number of nucleotides. Significant chi-square values for SHM data were determined at P < 0.05. Analysis of IL-6 production, GC numbers, area and IgG production were done using a two-tailed Student’s t-test and significance was accepted at P < 0.05.

Results

FDC-derived IL-6 and activation via IC–FcR interaction enhanced FDC–IL-6

To begin testing the hypothesis that FDCs produce IL-6 and that IC-activated FDCs produce elevated levels of IL-6, ICs were added to purified FDCs or purified FDCs with T and B cells. As shown in Fig. 1, all cultures containing WT FDCs produced IL-6 at all time points, 3, 6 and 10 days. Moreover, IL-6 production was enhanced by IC-mediated FDC activation. In marked contrast, cultures of WT B cells and T cells, separately or in combination with IL-6 KO FDCs, did not produce detectable levels of IL-6 even when ICs were added. FDCs isolated from normal mice with FDC-M1 have detectable levels of ICs on their surfaces but are capable of trapping more IC (21) and of being further activated in vitro via engagement of FcγRIIB (27). To further examine the need for ICs and FcγRs, FDCs were isolated from normal animals, ICs were added as an in vitro activation signal and supernatant fluids were collected on days 2, 4 and 6 (Fig. 2). IL-6 was apparent in the control lacking added ICs and it continued to accumulate over the 6-day period. FDCs may constitutively express IL-6 or this background IL-6 response may occur as a consequence of FDCs encountering ICs in vivo. In any case, addition of ICs enhanced FDC–IL-6 production within 48 h (P < 0.0001) and IL-6 production continued to increase over the 6-day period. Inhibition of IC trapping via FcγR with the FcR blocker (mAb clone 2.4G2) significantly inhibited FDC–IL-6 production whereas the isotype control had no inhibitory effect. Moreover, addition of soluble antigen (OVA) to the FDCs, rather than OVA in ICs, did not stimulate IL-6 production above background. We did not see enhancement of FDC–IL-6 production by the including low-Tox complement and complement receptors do not activate FDCs whereas engagement of FDC–FcγRIIB does (6).

Fig. 1.

FDCs but not B or T cells produce IL-6 and its production is enhanced by ICs. WT FDCs incubated with IL-6 KO B and T cells with or without ICs together with controls of WT B cells, T cells, IL-6 KO FDCs, WT B and T cells and WT B and T cells with IL-6 KO FDCs with or without ICs were set up and culture supernatant fluids were sampled after 3, 6 and 10 days. IL-6 levels were assessed with Bio-Plex. All cultures containing WT FDCs produced IL-6 at all time points, 3, 6 and 10 days, and IL-6 production was enhanced by activating FDCs with ICs. In marked contrast, cultures of WT B cells and T cells, separately or in addition to IL-6 KO FDCs, in the presence or absence of ICs did not produce detectable levels of IL-6 at any time point. Data are representative of the mean of one of four independent experiments with triplicates ± standard deviation.

Fig. 2.

ICs induce FDC–IL-6 production by engaging FDC–FcγRs. Purified WT FDCs were activated by incubating 1 × 10⁶ FDCs with 100 ng ml⁻¹ OVA-ICs made of OVA/rabbit anti-OVA at a ratio of 1:6. Controls of WT FDCs incubated with soluble 100 ng ml⁻¹ OVA and WT FDCs pre-treated with 10 μg ml⁻¹ azide-free FcγR (mAb 2.4G2) or its isotype control prior to IC loading were included. IL-6 levels in the fluids were assessed using a Bio-Plex Mouse IL-6 assay after 2, 4 and 6 days. As shown, ICs enhanced FDC–IL-6 production within 48 h (P < 0.0001). The IL-6 surge apparent at day 6 after IC loading precedes the late phases of the GC reaction (starting day 9–10 after primary challenge) when the presence of IL-6 appears to promote SHM. Inhibiting IC trapping via FcγRs with the FcR blocker (mAb Clone 2.4G2) significantly inhibited FDC–IL-6 production whereas the isotype control had no inhibitory effect. Moreover, addition of soluble antigen (OVA) to the FDCs did not stimulate IL-6 production above background. We did not see enhanced FDC–IL-6 production by the addition of low-Tox complement and complement receptors do not activate FDCs whereas engagement of FcγRIIB does (6). Data represent the mean of triplicate cultures ± SD and the results are representative of four independent experiments of this type

Anti-IL-6 inhibited specific antibody production in vitro

GC reactions were set up using purified FDCs, CGG-primed memory T cells and λ⁺ B cells isolated from WT mice. The λ⁺ B cells were obtained 6 days after immunization with 100 μg of NP–CGG per mouse. The 6 days allowed NP-specific λ⁺ B cells to begin clonal expansion and begin producing IgG anti-NIP in vivo. However, as previously shown (23, 28, 29) and confirmed in our studies (7), λ⁺ B cells taken 6 days after primary immunization have begun class switching and IgG production is increasing rapidly, but SHM has not occurred. The amount of NIP-specific antibody in the supernatant fluids after 7 days of culture was consistent with previous results with cells from WT mice (Fig. 3) (7). Without FDCs, the anti-NIP IgG levels were low (Fig. 3, columns 1 and 2). B cells do not survive well in the absence of FDCs and without FDCs only ∼20% of our B cells survive for a week compared with 80–90% in the presence of FDCs (30). Low antibody levels in cultures lacking FDCs are, at least in part, a reflection of a loss in B cells. Addition of FDCs to the B and T cells resulted in a dramatic increase in anti-NIP production (Fig. 3, column 3). To test the importance of IL-6 in promoting this IgG response, neutralizing anti-IL-6 (10 μg ml⁻¹) or isotype control were added to the B cell, T cells and FDC cultures where optimal anti-NIP responses are obtained (7). Column 4 of Fig. 3 illustrates that anti-IL-6 reduced the anti-NIP response by nearly 80%. To determine if the IL-6 served as a survival signal, trypan blue exclusion data indicated that 91.7 ± 3% of the cells at the start of the culture were recovered from cultures containing FDCs and isotype control after 7 days (1.8 ± 0.06 × 10⁶ viable cells recovered). Flow cytometry after double labeling with B220–FITC and PI revealed that 45.7 ± 1.6% of these cells were B220⁺/PI⁻. In cultures containing FDCs and anti-IL-6, there were 1.7 ± 0.06 × 10⁶ total viable cells persisting or 88.3 ± 3% of the cells at the start of the culture. Flow cytometry of these cultures revealed that 45.0 ± 1% of these cells were B220⁺/PI⁻ (P > 0.05). These results suggest that IL-6 is promoting B cell maturation rather than survival.

Fig. 3.

Anti-IL-6 mediated inhibition of anti-NIP IgG production in GC reactions in vitro. GC reactions were set up using cells isolated from WT mice. Culture contents are indicated beneath each column (BTAg: B cells, T cells and antigen; BTIC: B cells, T cells and ICs; BTFDCIC ISO: B cells, T cells, FDCs and isotype control antibody; BTFDCIC α-IL-6: B cells, T cells, FDCs and anti-IL-6). After 7 days of culture, supernatant fluids were collected and assayed for NIP-specific IgG by ELISA. The values represent the mean ± SEM from triplicate independent experiments. Either isotype control or anti-IL-6 antibody (at 10 μg ml⁻¹) was added to B, T and FDC cultures and the difference between isotype control and anti-IL-6 was statistically significant (P < 0.05).

Specific antibody responses and GC development impaired in IL-6 KO mice

To establish the impact of IL-6 on specific antibody and GC development, WT and IL-6 KO mice were immunized with NP–CGG. At 100 μg of NP–CGG per mouse, the IgG anti-NIP levels were not significantly depressed in IL-6 KO mice 14 days after immunization (data not shown). However, when immunogen was limited to 1 μg NP–CGG per mouse, the IgG anti-NIP response 14 days after primary immunization was clearly suppressed in IL-6 KO mice (Fig. 4A). We reasoned that ICs would not form as rapidly in IL-6 KO animals with reduced specific antibody and that FDC trapping and activation would be delayed and reduced. Consequently, B cells would not respond as vigorously resulting in fewer and less robust GCs in IL-6 KO mice. To test this, the number and sizes of GCs were determined 14 days after primary immunization with 1 μg of NP–CGG. As illustrated in Fig. 4(B and C), GCs were apparent in both WT and IL-6 KO mice. However, GCs in IL-6 KO mice appeared to be scattered and smaller and this was confirmed using morphometric analysis which indicated that the number and size of GCs in IL-6 KO mice were about half the WT level (Fig. 4D and E). Moreover, splenic leukocyte numbers in IL-6 KO mice were 94.6 × 10⁶ ± 11.5 leukocytes per spleen in WT versus 61.7 × 10⁶ ± 12.6 in KO (P < 0.01), and the GL-7⁺ (GC) B cells represented 3 ± 0.2 and 0.6 ± 0.3% of the splenic leukocytes in the WT and the IL-6 KO groups, respectively (P <0.01). Thus, the GC response in IL-6 KO mice was significantly impaired and that is consistent with the reduced anti-NP response.

Fig. 4.

Reduced NIP-specific serum IgG and a similar reduction in GCs in draining axillary lymph nodes from IL-6 KO mice 14 days after primary immunization. Panel (A) WT or IL-6 KO mice were injected subcutaneously with 1 μg of alum-precipitated (NP)₃₆–CGG using heat-inactivated Bordetella pertussis as adjuvant. Two weeks after primary immunization, serum was collected and assayed for IgG anti-NIP. The data are representative of two independent experiments with six mice in each group. Panel (B) illustrates labeling with FITC-conjugated anti-GL-7 (B1) that labels GC B cells and defines a GC within the cortex of a lymph node from a WT mouse. (B2) illustrates labeling of the same follicle with Cy5.5-conjugated anti-B220 that labels all B cells. (B3) is an overlay of panels (B1) and (B2) indicating the extent of GC B cells relative to B cells generally. Panels (C1, C2 and C3) are the same reagents with an IL-6 KO mouse. Panel (D) illustrates the number and panel (E) illustrates the size of GL-7⁺ GCs per mid-sagittal axillary lymph node section.

Irradiated WT mice reconstituted with IL-6 KO splenocytes but not IL-6 KO mice reconstituted with WT splenocytes developed GCs and anti-NIP IgG in response to 1 μg NP–CGG.

FDCs are irradiation resistant and to determine if FDC-derived IL-6 is important under physiological conditions in vivo, we set up irradiation chimeras reasoning that WT FDCs would provide IL-6 and induce GCs and IgG responses with IL-6 KO splenocytes. As illustrated in Fig. 5(A), anti-NIP levels were significantly lower in the IL-6 KO mice (P < 0.02) as compared with the WT mice regardless of the source of donor splenocytes. GCs in the WT mice reconstituted with IL-6 KO splenocytes were readily apparent and labeled intensely with GL-7 (Fig. 5B) and, as expected, these GCs developed in association with the IC-retaining FDC network (Fig. 5B2). In marked contrast, GC in IL-6 KO mice were poorly developed and very difficult to find even with WT splenocytes which could serve as a source of IL-6 (data not shown). Moreover, the number of GCs per mid-sagittal splenic section from IL-6 KO mice reconstituted with WT splenocytes averaged less than two as compared with 16 GL-7⁺ GCs per section in WT mice reconstituted with IL-6 KO splenocytes (P < 0.0005).

Fig. 5.

WT mice reconstituted with IL-6 KO splenocytes, but not IL-6 KO mice reconstituted with WT splenocytes, developed GCs and anti-NIP IgG. Chimeras were generated by reconstituting irradiated WT C57B/6 mice with 10⁸ IL-6 KO splenocytes and vice versa. One day later, these mice with irradiation resistant FDCs and splenocytes were challenged with 1 μg NP–CGG and the NIP-specific IgG was assessed after 14 days when the mice were sacrificed. Cryosections were prepared from each spleen, and GCs in the mid-sagittal sections were labeled with GL-7 and counted. As shown, the anti-NIP levels were significantly lower in the IL-6 KO mice (P < 0.02) as compared with the WT mice. GC development in the irradiated IL-6 KO mice was markedly impaired (averaged less than two as compared with 16 GL-7⁺ GCs counted in the reconstituted WT mice) and GC development was not rescued by reconstitution with WT splenocytes capable of producing IL-6.

SHM impaired in IL-6 KO mice

IL-6 helps regulate late events in B cell differentiation and we reasoned that IL-6 may influence SHM, which occurs after the initial induction of IgG anti-NIP late in the first week after primary immunization (7). The antibody response to the hapten NP coupled to the T cell-dependent carrier CGG is dominated in C57BL/6 mice by lambda light-chain-bearing antibodies (31–34) expressing the VH186.2 gene (35). This response has been well characterized and consequently was adopted in the current study. To examine SHM, the draining lymph nodes from the mice immunized with 1 μg of NP–CGG and studied in Fig. 4 were collected at day 14, the λ⁺ B cells were harvested and SHM was determined. Consistent with the IgG response and GC reaction data in Fig. 4, mutations were also apparent but reduced in the IL-6 KO mice. The rate of mutation in the VH186.2 gene segment switched to Ig-gamma decreased from 18 to 8.9 mutations per 1000 bp (P < 0.001) in WT versus IL-6 KO mice. The mutations are illustrated in Table 1 which shows the DNA sequence alignment of the VH186.2 gene. The corresponding amino acid sequence alignments were done and as expected, most mutations were in the complementarity-determining regions (CDRs) 1 and 2 (Table 2).

Table 1.

Vh186.2 sequence alignments from 1 μg (NP)36–CGG-immunized WT versus IL6 KO

The sequences shown in Table 1 are for lambda⁺ B cells obtained from WT or IL-6 KO mice whose anti-NIP levels are shown in Fig. 4(A). The anti-NIP levels as well as sequence analysis were performed 14 days after challenge of these mice with 1 μg of NP-CGG.

Table 2.

Vh186.2 amino acid sequence alignment from 1 μg (NP)₃₆–CGG-immunized WT versus IL-6 KO mice

Data represent amino acid translation of DNA sequences shown in Table 1, for lambda⁺ B cells obtained from WT or IL-6 KO mice whose anti-NIP levels are shown in Fig. 4(A).

Anti-IL-6 inhibited SHM in GC reactions in vitro

Provided that FDCs bearing the appropriate ICs are present (7), SHM can be studied in GC reactions in vitro. This prompted experiments to determine if anti-IL-6 will inhibit SHM under controlled conditions in culture to confirm the results in vivo. GC reactions were set up in vitro using purified FDCs, CGG-primed memory T cells and λ⁺ B cells isolated from WT mice. The λ⁺ B cells were obtained 6 days after immunization with 100 μg of NP–CGG per mouse. The 6 days allowed NP-specific λ⁺ B cells to clonally expand and begin producing IgG anti-NIP in vivo. However, as previously shown (23, 28, 29) and confirmed in our studies (7), λ⁺ B cells taken 6 days after primary immunization are not mutated. However, SHM would be one of the next events in a GC reaction. These 6-day B cells were cultured for seven additional days in vitro and the production of specific anti-NIP antibody and mutations in the VH186.2 gene segment were determined. The anti-NIP response of these mice is illustrated in Fig. 3 where anti-IL-6 reduced the IgG response by nearly 80%. Similarly, the mutation rate in these anti-IL-6-treated cultures was reduced (3.50 versus 0.65 mutations per 1000 bases in control versus anti-IL-6-treated cultures; P < 0.02). The data in Table 3 illustrate the DNA sequence alignment of VH186.2 gene obtained in cultures with the isotype control and cultures treated with anti-IL-6. Note that most sequences from cultures treated with the isotype control contained mutations whereas only two sequences from the anti-IL-6-treated cultures contained mutations. Of the 12 sequences analyzed from isotype control cultures, there were a total of 13 mutations. Eight of these were replacement mutations and five were silent. Six of the eight replacement mutations were in the CDR2 region. These data from cultures with the isotype control are compatible with the much larger published data sets (7) where mutations in vitro with IC-loaded FDCs ranged from 3 to 12 mutations per 1000 bases with >60% of these mutations in CDR1 or CDR2. Results from the anti-IL-6-treated cultures were similar; in that, two of the three mutations were replacement mutations.

Table 3.

Comparison of Vh186.2 sequence alignments from isotype and anti-IL-6-treated cultures

This experiment was repeated three times with similar results.

Discussion

The contribution of IL-6 to terminal B cell differentiation has long been recognized (8). However, the physiological relevance of FDC-IL-6 to GC B cell differentiation is unclear given that IL-6 can be produced by many cells and IL-6 production by FDCs has not been consistently found (18–20). In the present study, we confirmed that murine FDCs produce IL-6 and provide strong support for the physiological relevance of FDC–IL-6 in GC reactions. IL-6 was not detectable in in vitro GC reactions with IL-6 KO FDCs even with T and B cells from WT mice that could serve as a source of IL-6. In contrast, IL-6 production was normal in GC reactions with WT FDCs with T and B cells from IL-6 KO mice. Moreover, FDCs are irradiation resistant and GC reactions were abundant in irradiated WT mice reconstituted with spleen cells from IL-6 KO mice while GCs were virtually undetectable in irradiated IL-6 KO mice reconstituted with normal spleen cells. Moreover, like ICAM-1, VCAM-1 and FcγRIIB (6), activating FDCs with ICs up-regulated FDC–IL-6 production. ICs would be generated in vivo ∼3 to 7 days after primary immunization suggesting that a burst in FDC–IL-6 production would occur in the follicles ∼5 days after immunization and continue for a week or more. Although, class switching and low-affinity antibody production may occur before a GC reaction (36, 37), the IgG response is amplified as GCs develop over the next week. Moreover, SHM is a late event apparent in GCs after the first week (7, 23, 28, 29). Thus, SHM would coincide with optimal production of IL-6 by IC-activated FDCs. Both in vivo and in vitro data indicated that anti-NIP responses and SHM can take place in the absence of IL-6, but IL-6 clearly promoted both responses. The novel findings reported here are that FDC–IL-6 levels are up-regulated by ICs and that FDC–IL-6 is high in the second week after immunization concurrent with SHM and that IL-6 is required for optimal SHM. Thus, FDC–IL-6 appears to influence not only GC development and the level of IgG produced but also the mutations that promote enhanced IgG affinity.

Results presented here confirm and extend the report of Kopf et al. (13) who first demonstrated that IL-6 mRNA is produced by cells within murine GC clusters and determined that it was the FDC-enriched populations (>80% pure) rather than the lymphocytes that produced the IL-6 mRNA. Immunohistochemistry performed by these authors on cryosections from immune WT mice further supported their results revealing that cells with dendritic morphology within FDC network produced the IL-6. This production was limited to active GCs and not detected in primary follicles, which may help explain why some studies have failed to detect IL-6 in FDCs. By use of an IL-6 bioassay, Kopf et al. found that IL-6 was present only in cultures containing either the FDC lymphocyte clusters or the FDC-irradiated population but not in the sorted T or B lymphocytes. However, IL-6 was not quantified and our data extend this report by quantification of IL-6 production in highly purified FDCs in vitro and demonstrated IC-mediated induction.

The extent of in vivo suppression of the immune response in the IL-6 KO mice in the present study appears to be less than that observed by Kopf et al. (13). Our use of B. pertussis as adjuvant may have helped compensate for the lack of IL-6 in vivo. Alternatively, CGG may be more stimulatory in the absence of IL-6 than the OVA they used.

A previous study of GC reactions in vitro indicated that FDCs promote IgG production by B cells taken 6 days after primary immunization. However, in the absence of a second late exposure to antigen, SHM was not detected (7). This model was used in the present study which suggests that optimal SHM also requires IL-6. Typically, SHM occurs in the second week after primary immunization; however, if memory T cells are given along with immunogen in the form of IgG-ICs, extensive GC development and high levels of SHM can be obtained in the first week in vivo (7). The importance of ICs is indicated by the observation that replacing ICs with antigen, which will not load on FDCs, does not induce SHM in the first week even though memory T cells are present (7). Moreover, it is important for the ICs be made with IgG to promote SHM (38) and IgG-ICs are arranged on FDCs with a periodicity known to optimally engage B cell receptors for antigen and stimulate B cells (7, 39). In short, the combined data provide support for the concept that the unique ability of FDCs to trap and retain IgG-ICs for months provides them with the antigen necessary to deliver late specific and non-specific signals like IL-6 needed to promote SHM. Moreover, FDCs bearing the appropriate ICs promote activation-induced cytidine deaminase production and affinity maturation by NP-specific B cells (40).

In the present study, the in vivo reductions in specific antibody and SHM in IL-6 KO mice were most apparent at low antigen dose. At our usual immunization dose of 100 μg of NP–CGG per mouse, neither the reductions in IgG anti-NIP levels nor SHM was statistically significant 14 days after immunization (data not shown). However, defects in antibody production and SHM were readily apparent when the dose was reduced to 1 μg of NP–CGG per mouse. A similar antigen dose–effect relationship has been reported in lymphotoxin alpha KO mice. Mature FDC reticula and GCs do not develop in these mice; nevertheless, if the antigen dose is high enough, SHM is apparent in these FDC reticula-defective animals (41). Most infections start with only a few organisms, and if specific antibody is significant in immunity, it is important that immunity develops before the infection gets widespread. The results reported here indicate that under low immunogen dose conditions, FDC–IL-6 is likely to be important in generating rapid production of high-affinity-specific IgG. We look forward to studies testing the relationship between FDC–IL-6, selection of high-affinity clones with SHM and affinity maturation both in vivo and in vitro.

Funding

National Institutes of Health (AI-17142) to J.G.T.

Glossary

Abbreviations

CDR

complementarity-determining region

CGG

chicken gamma globulin

Cy5.5

cyanine 5.5

FDC

follicular dendritic cell

GC

germinal center

IC

immune complex

ICAM

intercellular adhesion molecule

KO

knockout

λ⁺ B cell

lambda light-chain-positive B cell

LB

Luria-Bertani

NIP

(4-hydroxy-3-iodo-5-nitrophenyl) acetyl

NP

(4-hydroxy-3-nitrophenyl) acetyl

OVA

ovalbumin

PE

phycoerythrin

PI

propidium iodide

SHM

somatic hypermutation

T_(CGG) cell

T cell specific for CGG

VCAM

vascular cell adhesion molecule

WT

wild type