Myeloid cells Limit Production of Antibody Secreting cells After Immunization in the Lymph Node

Abstract

Antibody secreting cell (ASC) expansion and survival are important processes in optimizing vaccines and controlling autoimmunity. The microenvironment of the medullary cords is positioned to control these key processes. Previously, we imaged and characterized ASC differentiation and migration by intravital microscopy in the lymph node (LN) by transferring and activating B cells expressing YFP only in the ASC compartment. In this study, we observed that YFP+ ASCs in the medullary cords migrated along myelomonocytic cells and arrested in contact with them. Acute ablation of myeloid cells using the human diphtheria receptor system (DTR) expressed in Lysmd1-cre positive cells increased ASCs and antibody production by 2 fold. Increases in ASC numbers were associated with cell proliferation based on Ki-67 staining, rather than reduced apoptosis, or changes in egress from the LN. Using DTR-mediated ablation targeted to Ccr2 expressing myeloid cells also generated increases in ASCs. In contrast, neither the depletion of Gr-1-positive cells with an antibody nor the ablation of cells using a cd11c- DTR resulted in any change in ASCs. IL-6 cytokine signaling can enhance ASC production and has been implicated in dampening ASCs in lupus mouse models through myeloid cells. Using mixed BM chimeras, we observed that IL-6 enhances ASC production, but IL-6 production was not required by myeloid cells to dampen ASCs in the LN. Inhibition of ASCs by these myeloid cells in the LN provides a new regulatory mechanism with implications for tuning antibody responses.

Introduction

Effective and long-lived antibody production in the body is dependent on the generation and survival of plasma cells (PCs). This process involves a multi-step differentiation program, cell-cell interactions and migration through several anatomical niches that facilitate each stage (1, 2). In the lymph node (LN), plasmablasts (PBs), precursors of the PCs, are produced following antigen exposure, in the first few days at extrafollicular regions and within germinal centers at later stages of the response. PBs migrate and accumulate in the medullary cords of the LN, with some PBs exiting and homing to other tissues such as the spleen and bone marrow (BM). Although reductions in motility and proliferation tend to correlate with differentiation of PB to PC, this transition is not sharply defined but rather a continuum. Both a PC and a PB can be described as an antibody secreting cell (ASC) that express high levels of Blimp1 protein, encoded by the Prdm1 gene.

Previously, we characterized ASC migration in the LN by tracking cells expressing yellow fluorescent protein (YFP) under the control of Prdm1 promoter by imaging using two photon microscopy and observed that both early and late forming cells migrate rapidly, in a non-directed but highly linear “random sprint” eventually arresting in the medullary cords (3). We detected a correlation between reduced cell migration and the differentiation state of the ASC in vivo and in vitro on an ICAM-1 coated substrate. We concluded that there was a cell autonomous component to arrest in the medullary cords. However, we did not assess the role of other medullary cord cells on ASC arrest or any functional role for ASC physiology. These auxiliary cells are often referred to as niche cells, and seem to vary in a tissue-specific manner (1).

Many cell types have been implicated in ASC differentiation and survival that are tissue and species specific. For example, within the BM, stromal cells, megakaryocytes, eosinophils, dendritic cells (DCs), neutrophils, and other cells types have all been assigned a functional role, many based on colocalization studies (1). In the LN, MacLennan and colleagues used immunohistochemistry to identify and catalogue cells that neighbor ASCs during their migration and differentiation in the mouse LN (4). They detected ASCs juxtaposed to DCs in the T cell zone, and with neutrophils, monocytes, and macrophages in the medullary cords, as well as subcapsular sinus macrophages. Based on the high expression of IL-6 and APRIL transcripts in these myeloid cells, they proposed that these cells may provide a niche for ASC differentiation and survival. These correlative studies provide hints at important cell niches, but call attention to the need for direct studies to test these hypotheses. It can be difficult to distinguish which cell contacts are important based on thin section histology of lymphoid tissues, due to a crowded micro-environment full of an assortment of cell types. Some cells are dynamic, and may only contact plasma cells briefly in passing.

In this study, we extend these observations using intravital imaging to visualize the duration of cell-cell interactions. This technology provides the ability to distinguish transient from stable interactions as well as observe cell contacts in an intact volume, which provides important contextual information that is obscured in thin sections. We also used a variety of depletion techniques to target different myeloid subsets to directly assess what functional roles they play in ASC differentiation and antibody production.

Materials and Methods

Mice, Immunizations, Treatments

For most experiments, C57BL/6 (B6) or congenic CD45.1+ (so called B6.SJL) mice were used as recipients (from Taconic or Charles River). CCR2-DTR mice were provided by Eric Pamer, LysM-GFP+ mice were a gift from Tomas Graf. LysM-cre, iDTR, CFP, tdTomato, CD11c-DTR, Blimp1-YFP, IL-6−/− mouse strains are available from Jackson Labs.

To generate antigen-specific ASCs, recipient mice were immunized by i.p. injection with ovalbumin (50µg) emulsified in alum (Pierce) to generate abundant T cell help. After 2–4 weeks, mice received i.v. adoptive transfer of approximately 3×10⁶ naive B18-high+/− Blimp1-YFP+ B cells that were purified by negative selection using CD43-depletion kit (Miltenyi Biotec). The following day, mice were boosted with 50µg/mouse of nitrophenyl-conjugated ovalbumin (NP-OVA) (Biosearch Tech) by s.c. injections distributed into the footpads, handpads, and base of the tail to target draining LNs. Mice were sacrificed on day 7 for flow cytometry analysis of the draining LNs (popliteal, inguinal, axillary, and brachial), spleen, and BM from hind leg bones.

For DTR depletion experiments, mice were treated with an i.v. injection of Diphtheria toxin (1µg in 100µL of PBS) on day 4 and 6 after boost. For antibody depletion of monocytes and neutrophils with anti-Gr-1, mice received high dose RB6-8c5 antibody (i.v. injection with 300µg) on days 4 and 6 after boost.

For imaging experiments, mice received s.c. injection of NP-OVA antigen (50µg/mouse in 50 µL) into the footpads to target the popliteal LN. Mice were imaged on day 7 after boost using techniques described, previously (3). On day 6, fluorescent (CFP+ or td-Tomato+) polyclonal purified naive B cells were transferred as controls for migration or cell contacts.

For experiments with chimeric animals, recipient C57Bl/6 mice were irradiated once with 900 Rad. BM mononuclear cells were harvested from donor mice, RBCs were removed, and washed in PBS and transferred intravenously (approximately 300,000 cells/mouse) into recipients. Mice were used after eight weeks in experiments, following similar immunization and transfer protocols as described for other recipients. All experiments were approved by Institutional Animal Care and Use Committee.

Flow cytometry, Antibody staining and ELISA

Standard commercially available antibodies from Ebioscience, Biolegend, or BD Biosciences for flow cytometry were titrated for concentration (typically at 50 ng/mL final), and used for cell staining, in PBS with 0.5% BSA and 0.5mM EDTA for 30 minutes at 4°C. Flow cytometry data were collected on LSRII or FACSCalibur in the NYU cytometry core facility and analyzed using Flowjo software. For ELISA measurements, NP-KLH was titrated and used as capture antigen at 2µg/mL. Serum samples were diluted to 1:1000 and further diluted by serial dilutions on the plate. Purified NP-specific antibody, 9T13, was used as control and to generate a standard curve for concentration. Anti-mouse IgG HRP (Biorad) was used for detection (at 1:5000 dilutions) in conjunction with TMB substrate, and measured using an absorbance plate reader.

Image analysis

In order to calculate cells in contact, (Figure 1B, 2D), time lapse movies were analyzed by Volocity (Perkin-Elmer) at start, end and middle time points and the results averaged. Fluorescently-labeled cells (CFP+, GFP+, and YFP+) were distinguished on the basis of colocalization between channels based on two channel raw data using CFP and YFP filter settings. By this method of cell detection, cell-cell overlaps are eliminated. The cell volume for plasma cells and naïve B cells was automatically and equally increased using the dilate function, to evaluate neighboring cells in contact. B cells in contacts with GFP+ were quantified as a frequency of total B cell population.

Figure 1. ASCs Contact Myelomonocytic cells in the Medullary Cords.

A. Diagrammatic representation of experimental setup for imaging ASCs in LysM-GFP+ recipients. B. Still image taken from intravital time lapse imaging showing NP-specific Blimp-1-YFP+ ASCs (red) and polyclonal CFP+ naïve B cells (green) in contact with LysM-GFP+ (yellow) cells in the medullary cords of the popliteal LN. C. Based on image analysis (see methods), the fraction of ASCs in contact with LysM-GFP+ cells is higher than the fraction of naïve B cells in contact LysM-GFP+ cells. D. The distance between naïve B cells or ASCs with the nearest LysM-GFP+ cell is plotted as a cumulative frequency histogram. Cell quantitation is based on a single time-lapse movie, but analysis was repeated three times with additional mice. Error bars reflect standard deviation. * indicates p value < 0.05. Scale bar is 20 microns.

Figure 2. Various Myeloid Subsets are LysM-GFP+ in the LN.

A. Tiled image of LysM-GFP+ popliteal LN by two-photon with GFP cells throughout the organ, with medullary cords (MC), T cell zone (TZ), and B cell follicle (BF) labeled. B. Dot plot of cells from the LN, highlighting several distinct populations of GFP+ cells based on intensity (I, II, III, IV). Gating strategy for DCs (CD11c+), Macrophages (F4/80+ CD11b+), Neutrophils (Ly-6G^high CD11b+) and monocytes (CD11b+ Ly6c^high) in the LN (C). These populations in the LN of LysM-GFP were classified by their GFP expression pattern based on the gates in B, quantified in D (n = 3 mice).

Calculations for the distance between ASCs and naive B cells with LysM-GFP+ cells were conducted using Imaris (Bitplane). Briefly, the GFP+ cells were converted into a single surface using the surface feature, naive and ASCs were captured using the ball feature to determine the cell centroid. The distance between the GFP+ surface and cell centroids was converted into a distance channel using a matlab plugin function, and used to compute a cumulative frequency distribution for the cells to the nearest GFP+ neighbor.

Statistical Analysis

For depletion experiments, cohorts of 4 or more mice were used per condition in multiple experiments. In some cases, to improve the power of statistical comparisons, independent experiments were pooled together by normalizing the values to the average for the control condition and the values were scaled accordingly.

Results

We wanted to determine what cells are in direct and stable contact with ASCs to assess their potential role in regulating ASC migration and differentiation. Previous studies using fixed tissues cannot resolve which contacts are stable and which are transient. Therefore, we used time-lapse intravital two-photon microscopy to visualize ASCs in the LN medullary cords of live anesthetized mice. As before (3), we induced antigen-specific ASCs by transferring nitrophenyl (NP)-specific gene targeted naive B cells (B18-high) from mice also containing a bacterial artificial chromosome containing a transgene in which YFP is expressed using Prdm1 regulatory sequences (Blimp1-YFP), and immunizing the mice with NP-conjugated to ovalbumin (NP-OVA) antigen targeted to the popliteal LNs and imaging on day 7 at the peak of the ASC response (Figure 1A, details in methods). To visualize and assess potential niche cells in the medullary cords, we tested a variety of congenic fluorescent reporter strain mice as recipient mice. The most promising of these was a mouse stain in which green fluorescent protein (GFP) was produced in cells based on the Lysmd1 promoter expression (LysM-GFP) (5). These mice express GFP in cells of the myelomonocytic lineage including neutrophils, monocytes, macrophages and DCs. In the medullary cords, Blimp1-YFP+ ASCs were in close contact with LysM-GFP+ cells while naive cyan fluorescent protein (CFP) expressing B cells were not interacting as closely or as stably with GFP+ cells (Figure 1B–D). Migrating ASCs moved along a GFP+ cell network and arrested ASCs were always found to be in contact with GFP+ cells (Supplemental Movie 1). In some cases we observed GFP+ and YFP+ cells tugging on each other, actively. We did not observe a similar interaction between polyclonal CFP+ naive B cells with GFP+ myelomonocytic cells, which migrate rapidly in the medullary cords, as previously described (3).

To determine if these interactions had any physiological effect on ASCs in the LN, we attempted to ablate various subpopulations of LysM-GFP+ cells. Various cell populations are GFP⁺ throughout the LN particularly in the medullary cords (Figure 2A), and at differing expression levels (Figure 2B). Based on our imaging experiments, it was clear that the cells in question had an intermediate GFP expression level (Figure 2C–D) corresponding to various myeloid cell types including monocytes, macrophages and DCs.

To target LysM-GFP+ cells for depletion, we bred mice to express a LysM-driven cre recombinase transgene allele with two LoxP cleavage sites flanking a stop cassette followed by a human diphtheria toxin receptor, so called (inducible-DTR), to generate a DTR receptor on LysM-expressing cells, or simply, a LysM-DTR mouse. This approach has been used to deplete macrophages and monocytes, previously (6). Treatment of these mice with diphtheria toxin (DT) eliminated a variety of cell types, including F4/80+ macrophages, and Ly-6c+ monocytes in the LN and spleen (Figure 3A). After B cell transfer and boost to generate ASCs targeted to the draining LNs (as described in methods), depletion of LysM-cre+ cells by DT treatment on day 4 and 6 increased the number of NP-specific YFP+ ASCs in the LN by 2 fold compared to untreated mice (Figure 3B–D) on day 7. We measured NP-specific antibody production by indirect ELISA, with or without DT treatment and found increased antibody titers in treated mice compared to untreated mice (Figure 3E). This was consistent with increased numbers of NP-specific ASCs. To see how long-lasting the effects of acute myeloid depletion were, we measured ASC numbers and serum titers at 14 and 28 days after immunization. At day 14 days, the mice treated with DT had increased ASCs numbers in the spleen and anti-NP serum, but by day 28, we saw no difference in treated vs untreated mice (Figure 3F–G), possibly due to repopulation of the myeloid compartment or limited capacity in the ASC niche. By four weeks, we detected very few ASCs remaining in the LN, which is not a niche for long-lived ASCs (1, 2). To determine if the depleted cells were the same GFP+ cells engaged with ASCs in the medullary cords, we crossed the LysM-DTR allele combination with the LysM-GFP allele. Indeed, after DT treatment, fewer GFP+ cells were present in the LN and ASCs were less associated with GFP+ cells in the medullary cords after depletion (Figure 3H–J).

Figure 3. Depletion of LysM-cre+ Expressing cells Increases ASCs in the LN.

A. LysM-DTR (CD45.2+) recipient mice were preimmunized with OVA, transferred B1-8high+/− Blimp-1YFP+ (CD45.1+) naïve B cells and boosted with NP-OVA by s.c. injection. Bar graphs showing % of monocytes, macrophages, DCs, and neutrophils in LysM-DTR LN and spleen from (n=3 mice) on day 7 after boost, with or without DT treatment on day 4,6. Gates are based on Figure 2C gates for myeloid subsets. B. A representative dot plot of YFP+ ASC in the LN with or without DT treatment. Cell numbers were pooled and quantified from multiple experiments as a percentage of the transferred population (CD45.1+) of cells in the LN (C) or as the total number of YFP+ ASCs in the LN (D), with WT recipients with DT shown as additional controls. In E, NP-specific antibody was measured in serum on day 7 after boost with or without DT, based on two experiments, n=6/condition total. Antibody concentration was calculated based on standard curve using purified monoclonal NP-specific antibody (9T13). In F-G, wild-type mice lethally-irradiated and reconstituted with LysM-DTR BM for 8 weeks were used as recipients similar to experiments shown in B-E, treated with or without DT on day 4 and 6 after boost, n=10 mice/condition, pooled from two independent experiments. Total number of YFP+ ASCs on day 14 and 28 in the LN and spleen are shown (F). The NP-specific serum titer was measured on day 14 and 28 for mice treated with or without DT (G). As in Figure 1, intravital imaging of YFP+ ASCs (red), naive B cells (green) and LysM-GFP+ cells (yellow) in the medullary cords of LysM-GFP+ LysM-DTR+ recipients, with or without DT treatment in H. The fraction of GFP+ cells in the LN +/− DT treatment, measured by flow cytometry in I. Quantification of contacts between ASCs or naïve with GFP+ cells is shown in J. * indicates a p value < 0.05 for t-tests. Imaging experiments were conducted twice. Errors bars indicate standard error.

Increases in ASCs in the LN after myeloid cell depletion could be the result of many potential mechanisms. The most simple and direct possibilities we examined were that ASCs numbers could be higher due to more differentiation or proliferation, less cell death, or less egress from the LN. Since ASCs numbers were also increased in the spleen after DT treatment, we ruled out changes in migration as a possibility (Figure 4A). ASCs, particularly those generated early in the response, are known to undergo apoptosis in the LN, on days 7 and onward (7). Annexin V staining for apoptotic YFP+ ASCs showed no decrease in response to myeloid cell depletion (Figure 4B). The frequency of germinal center B cells in the transferred cell population was similar in treated and untreated conditions (Figure 4C), suggesting the precursors of ASC differentiation were unaffected. However, we detected more cell proliferation in YFP+ ASCs in the DT treated mice versus controls (Figure 4D–E) based on Ki-67 staining (8).

Figure 4. Increases in ASCs after LysM-DTR depletion are largely due to Enhanced Proliferation rather than other factors.

A. Experimental setup as in Figure 2A–E. The population of YFP+ ASCs in LysM-DTR mice is higher in LN and spleen but not BM as compared with C57BL/6, when treated with DT. B. Quantification of apoptotic marker, Annexin V, on the surface of YFP+ ASCs in the LN was similar in LysM-DTR mice with or without DT. B. Assessment of Fas+GL7+ germinal center cells, gated on CD45.1+ CD19+ transferred B cells in LysM-DTR mice with or without DT treatment D. Results are from a single experiment that were repeated twice with n=3–4 per condition in each experiment. Measuring proliferation based on mitotic marker, Ki-67, on YFP+ ASCs with or without depletion. Two experiments shown separately in D, and pooled and normalized to control mice (without depletion) in E. * indicates a p value < 0.05 for standard student t-test. ** indicates a p value < 0.05 for a one-sided t-test Welch’s test. Error bars reflect standard deviation.

To confirm our results in the LysM-DTR model, we attempted to deplete similar myeloid cells using a CCR2-DTR transgenic mouse model (9). Myeloid cells that are recruited from the spleen or BM express CCR2, which include monocytes, macrophages and DCs. We generated chimeric mice using CCR2-DTR BM as donors cells (Figure 5A). After reconstitution, we followed a similar transfer, immunization, and depletion strategy as before for LysM-DTR system. Treatment with DT depleted various CCR2+ cell types including Ly6c+ CD11b+ monocytes, F4/80+ CD11b+ macrophage populations and other myeloid populations (Figure 5B), while increasing frequency of ASCs in the transferred cell population in the LN and spleen as compared with controls (Figure 5C–D). In addition, more ASCs in the LN expressed elevated Ki-67+ levels compared to controls (Figure 5E). These results were similar to those obtained using the LysM-DTR model.

Figure 5. Depletion of Myeloid cells by CCR2-DTR Increases ASC numbers in the LN.

A. Experimental setup described. In B, systemic myeloid depletion in CCR2-DTR mice after DT treatment was assessed in the spleen, compared with untreated CCR2-DTR chimeric mice and C57BL/6 mice as controls. Quantitation of % of macrophages (F4/80+ CD11b+) and monocytes (CD11b+ Ly6c+) subsets, as in Figure 2C, ((n=4 or more mice/ condition). C. Assessment of YFP+ ASCs in the LN spleen with or without DT treatment, pooled from two independent experiments (n=8 mice/condition), also shown as total number of YFP+ ASCs in the LN for two independent experiments. D. Proliferation of YFP+ ASCs was assessed by Ki-67+ staining.

By using these depletion techniques we could deplete myeloid cells and increase ASC responses. However, these strategies are not specific for monocyte, macrophage or DCs, which is a problem with most approaches (explored in Discussion). Nevertheless, we attempted additional myeloid cell depletion strategies to try to resolve which cells are regulating ASCs. To deplete a subset of monocytes (and neutrophils), we treated mice with a regiment of high doses of Gr-1-specific antibody (Figure 6A). After depletion, we saw a marked reduction in neutrophils (GFP^high, Gr-1^high) and monocytes(GFP^int Gr-1^int) in the mouse, however we saw no effect on Blimp1-YFP+ ASC numbers in the LN, or in ASC migration to spleen and BM compartments (Figure 6B).

Figure 6. Depletion of Gr-1+ Myelomonocytic cells or CD11c+ DCs does not alter ASC population.

A. LysM-GFP+ mice were treated with anti-Gr-1 or control (PBS), and the depletion of monocytes (green gate) and neutrophils (red gate) was assessed in the LN and spleen, in representative dot plots. B. Using transfer/boost model as in Figure 1A, recipients were treated with IV anti-Gr-1 (300µg) or PBS on day 4 and 6 after boost. The % of YFP+ ASCs in the LN, BM and Spleen was assessed in control vs anti-Gr-1 treated mice. Single representative experiment is shown in B. In C, CD11c-DTR-ires-GFP BM was used to reconstitute mice as in Figure 4A. Images of popliteal LN analyzed from CD11c-DTR recipients to assess depletion of GFP+ DCs (in yellow) with or without DT treatment, with YFP+ plasma cells (red). D. Quantification of plasma cells in the LN, spleen, and BM, with or without depletion, pooled. In order to correct for variability between experiments, values were normalized as a ratio of the raw % of ASCs to the average for the control in each experiment, thereby allowing pooling from four independent experiments with different BM chimeras, n=16 mice per condition total.

Next, we used mice expressing human diphtheria-toxin receptor under the CD11c promoter with a GFP reporter (CD11c-DTR), which can deplete DCs and other CD11c+ cell types upon the introduction of DT intravenously. To avoid depletion of non-hematopoietic CD11c+ cells (such as endothelial cells), which can lead to early mortality (10), we irradiated WT mice and reconstituted their hematopoietic compartment with CD11c-DTR-derived BM (chimeric CD11c-DTR) to use as recipient mice as in Figure 5. After reconstitution, B cell transfer, and immunization, mice were treated with DT on days 4 and 6 and analyzed on day 7. We chose to deplete late so as not to interfere in the initial DC-T and B-T interactions, which are critical for PC production. We did not detect any defect in B18-high germinal center B cell numbers as a result of late depletion (data not shown). Upon DT treatment, all of the CD11c-GFP+ were efficiently depleted in the LN, however, the YFP+ cells were still present (Figure 6C) and their numbers were not different as compared to untreated mice in a statistically significant level (Figure 6D). We detected a decrease in YFP+ plasma cell numbers in the spleen and BM. Since antigen is targeted to the draining LNs, any decreases in YFP+ plasma cells in the spleen or BM are likely due to either migration or survival defects, that occur downstream of plasma cell egress from the LNs. DCs have been implicated in survival of ASCs in the BM and spleen (11).

Signaling through the IL-6 cytokine pathway can enhance antibody secretion and plasma cell differentiation (12), but IL-6 may also dampen B cell responses in lupus-prone mice (13, 14). In these conditions, DCs and macrophages have been shown to secrete IL-6, which reduces B cell responses. We wanted to test if myeloid cell inhibition of ASCs proliferation in the LN involved IL-6 signaling. However, IL-6 plays an important role in various aspects of B cell activation, and many cell types secrete IL-6 including B cells, myeloid cells, and non-hematopoietic follicular DCs (12, 15). To circumvent this requirement partly, we generated BM chimeras using IL-6 deficient, LysM-DTR, or WT BM into WT CD45.1+ mice to be used as recipients for our transfer and immunization protocol (Figure 7A). ASC production was reduced in recipients with IL-6−/− hematopoietic cells as compared with WT or LysM-DTR BM indicating that IL-6 enhanced ASC production (Figure 7B). To see if IL-6 signaling was required in myeloid cell inhibition of ASCs, we generated chimeric mice with 50/50 mixtures of donor BM cells from LysM-DTR mice and IL-6−/− mice. In these mice, upon treatment with DT, IL-6-sufficient myeloid cells would be depleted while IL-6-deficient myeloid cells would remain. As controls, we generated 50/50 BM donor cell mixtures of IL-6−/− and WT, or LysM-DTR and WT. We could assess chimerism in the myeloid compartment on the basis of CD45 congenic markers and GFP expression (Figure 7C). All chimeric mice were treated with DT, but only those that had LysM-DTR BM had reduced levels of Gr-1+ monocytes, as expected (Figure 7D). Mice with mixtures of LysM-DTR BM with either WT or IL-6−/− deficient BM had similar ASC production after DT treatment, indicating that the non-depleted population of myeloid cells (LysM-DTR-negative subset) does not require IL-6 production to inhibit ASC generation (Figure 7E). In contrast, recipients with mixed IL-6−/− and WT BM had impaired ASC production that was similar to mice with IL-6-deficient only BM, but lower ASC production than in WT mice or mice with mixed IL-6−/− and LysM-DTR+ BM. This suggests that the level of IL6 production in recipients with mixed WT and IL-6−/− BM is not sufficient to restore WT levels of ASC production. Furthermore, the capacity for myeloid cell depletion to increase ASC numbers can still occur under these reduced IL-6 levels in mixed LysM-DTR, IL-6−/− BM chimera as compared with IL6−/− WT mixed chimeras. Taken together, IL-6 plays a positive role on ASC production while myeloid cells play a negative role in ASCs. These two effects are unlikely to be involved in the same pathway since we could identify contributions of each pathways when they are combined in the mixed chimera strategies and we can conclude that IL-6 signaling does not contribute to myeloid cell inhibition of ASCs.

Figure 7. IL-6 Production is Dispensable for Myeloid cell Regulation of ASCs.

A. Experimental setup diagrammed. B. Quantification of ASCs numbers as a percent of transferred cells, n=4 per condition, measured once. C. Chimerism was assessed by CD45.1/CD45.2 staining of Gr-1+ cells for mixed BM conditions. D. Depletion of Gr-1+ cells was measured in spleens on day 7, for three mixed BM conditions, shown for a single experiment for C-D, but conducted three times. E. Quantification of ASCs numbers as a percent of transferred cells. Values are pooled from three experiments n=12 per condition in total. To correct for variance between independent experiments with different batches of BM chimeric mice, values in a given experiment were normalized to the average value for mice with BM containing 50:50 LysM-DTR: IL-6−/− (first column), shown in F. p values are for unpaired t-test comparison, error bars reflect standard error.

Discussion

The plasma cell niche in the medullary cords of LN is a transient microenvironment that is poorly understood. At steady state, ASCs in the LN are very rare, less than 0.001%. However, during an immune response in the LN, ASCs are 100-fold more abundant, accumulated mainly in the medullary cords. Based on intravital imaging showing a strong cell-cell interaction between ASCs with LysM-GFP+ cells, we expected that these cell-cell contacts promoted ASC survival and function, based on the niche model, previously proposed (4). However, we discovered that macrophages and monocytes serve to dampen ASC proliferation, which ultimately dampen antibody responses.

By employing an overlapping set of depletion techniques, we were able to narrow down the potential myeloid cell types that are responsible. We can rule out conventional DCs from any functional role at late stages based on Figure 6D results. Using the LysM-DTR and CCR2-DTR models, we were able to target monocytes and monocyte-derived macrophages and DCs, however these techniques are not specific for any given myeloid subset. In contrast, depletion with high-dose anti-Gr-1, which targets Gr-1+ monocytes and neutrophils, had no effect. This left macrophages and Gr-1-negative monocytes, such as CX3CR1+ subset. Since the myeloid cells were providing a dampening effect on ASC numbers and antibody titers, we hypothesize that we are potentially targeting a myeloid-derived suppressor cell (MSDC). MDSCs are poorly understood, comprised of monocytes, macrophage, and DC subtypes that can inhibit other immune cells (16). Based on our depletion schemes, anti-Gr-1 treatment would spare Gr-1^low CX3CR1+ myeloid-derived suppressor cells, which can secrete inhibitory cytokines such as IL-10 and may act to inhibit ASCs. However, further characterization with depletion techniques that target these cells types specifically would be needed to validate this possibility.

Regardless the cell, the striking engagement of ASCs by myeloid cells in the medullary cords suggests a new regulatory mechanism that modulates antibody production. Myeloid cells are ideally positioned and concentrated in the medullary cords to provide further regulation and input into the ASC humoral response. Typically monocytes are recruited in the lymph node during an inflammatory response to provide innate help during an infection. However, in our experiments, which model vaccination with no adjuvant, monocytes may dampen ASC responses in an effort to qwell ASC responses that may be unnecessary or potentially directed towards self antigens. These effects seem to be ameliorated over time as new myeloid cells repopulate the lymphoid tissues, based on our time course, and seems to implicate short-lived plasmablasts as the target of regulation. This is consistent with studies that show in mouse models of lupus, macrophages and DCs can inhibit ASC responses (13, 14). However, in those models, IL-6 played a role in signaling to the ASCs. When we tried to block test if myeloid cells required IL-6 to inhibit ASCs using mixed BM chimeras, we saw no role of IL-6 in these cells. This difference in mechanism may be due to differences in our vaccination model which lacks the inflammation common to the lupus system. In lupus patients, conditions of chronic inflammation produced by myeloid cells leads to plasmablasts are overrepresented in the blood and in peripheral tissues including kidney (17, 18). It would be interesting to repeat these experiments using strong adjuvants or an infection model where inflammation may change the role that myeloid subsets play regulating ASC responses.

Based on the tight cell-cell interactions we observed in the LN, there is also the potential that myeloid cells may also help regulate the balance of ASC egress versus retention in the LN. Although, we were not able to measure compensatory changes in cell numbers between LN and spleen, this does not exclude the possibility that myeloid cells trap ASCs in the LN. Myeloid cells increase and decrease in the LN with kinetics that match ASC numbers in the LN. In studies with CD18-deficient mice, with defective LFA-1, demonstrated increased ASCs numbers in the LN due to failure to egress (19). Interestingly, these mice also have elevated IL-6 levels in the serum (20). These results may indicate a bidirectional signaling that may occur between ASCs and myeloid cells. Interfering with these contacts may lead to increased inflammation, which could explain the increases in ASC function after myeloid cell depletion. The interplay between these cell types may provide new ways to modulate immune response in both immunization and autoimmune contexts.

Abbreviations used

ASC

antibody secreting cell

BM

bone marrow

CFP

cyan fluorescent protein

DC

Dendritic cell

GFP

green fluorescent protein

LN

lymph node

NP-OVA

nitrophenyl-coupled ovalbumin

PB

Plasmablast

PC

Plasma cell

YFP

yellow fluorescent protein