Eur. J. Immunol. 1991. 21: 2697-2701

Chemotaxis of germinal center B cells to C5a


Leo I. Kupp., Marie H. Kosco*, Harvey A. Schenkein. and John G. Tew.

Chemotaxis of germinal center B cells in response to C5a*

Department of Microbiology and Immunologyo, Medical College of V i i a Commonwealth University, Richmond and Basel Institute for Immunology*, Basel

An infiltrate of B cells and plasma cells is characteristic of certain chronic inflammatory lesions. However, mechasms involved in the local accumulation of these cells have not been established. Efforts to demonstrate that B cells from normal animals can migrate in response to inflammation-induced chemoattractants have been inconclusive. The objective of this study was to determine if murine germinal center (GC) B cells could respond chemotactically to a C5a gradient. On successive days after secondary immunization, draining lymph nodes were harvested and the activated GC B cells isolated. These GC B cells were placed in modified Boyden chambers, incubated for 3 h and the distance the leading front of cells migrated through the filters was determined. The results show that GC B cells migrated to factors in zymosan- and lipopolysaccharideactivated serum. The migratory response demonstrated distinct kinetics. Cells isolated between 2 to 4 days after secondary immunization migrated, whereas cells isolated at day 0 and beyond day 6 did not. Checkerboard analysis revealed that the migratory response was attributable to both chemokinesis and chemotaxis. Anti-C5 inhibited the migration of day-3 GC B cells implicating C5 in the migration mechanism. Studies using recombinant C5a established that this C5 fragment was chemotactically active. In conclusion, GC B cells generally were not chemotacticallyactive. However, at a particularstage of maturation B cells in the GC become responsive to C5a as a chemotactic agent. Thus, B cells from normal animals may respond chemotactically, and C5a may play a role in recruitment of recently activated B cells into inflammatory sites.

1 Introduction

(AFC; [16,17]). Furthermore, we found that day-3 GC B cells home to the BM when injected into naive mice [17]. In v i m studies indicate that T lymphocytes are capable of This raised tbe question of whether chemotactic agents chemotaxisand chemoattractantsmay serve to help localize might help recruit stimulated B cells from the circulation these cells in sites of inflammation (reviewed in [l, 21). In into inflammatory sites, and prompted this investigation tQ contrast, B lymphocyte chemotaxis is difficult to demon- determine if GC B cells are capable of chemotaxis. strate [3-51. Nevertheless, in chronicinflammatorydiseases such as periodontitis and Crohn’s disease, the lesions are CS-derived chemoattractants are known to be present in characterized by a dense B cell and plasma cell infiltrate chronic idammatory diseases such as periodontitis and [6-111. Furthermore, the antibody produced by these cells Crohn’s disease [18-211. Experiments in the present study may be specific for antigen administered at a distant site were designed to investigate whether GC B cells were [12,13]. However, the mechanisms involved in this local chemotactically responsive to C5-derived chemoattracaccumulation of B cells and plasma cells in chronic inflam- tants. The data reported here indicate that GC B cells at a specific stage of maturation (2-4 days after secondary matory lesions have not been established. immunization) can migrate in a C5a gradient and was It is known that activated B cells derived from draining LN attributable to both chemokinesis and chemotaxis. This are found in the lymph 2-4 days after secondary challenge may help explain the localization of B cells in certain [14]. These cells enter the blood via the thoracic duct and chronic inflammatory sites rich in C5a. are apparently in transit to the BM where most Ig are made [15]. We recently found that these activated Bcells in thoracic duct and blood are derived from germinal centers 2 Materials and methods (GC) and some are immature antibody-forming cells 2.1 Mice Eight- to twelve-week-old female BALBkByJ mice were purchased from The Jackson Mboratory (Bar Harbor, ME). The mice were housed and maintained in the * This work was supported by National Institutesof Health grants university animal facilities in standard shoebox cages and DE 08972 and DE 00151. kept in filtered cage racks. The mice were given food and water ad libitum. Correspondence: John G. Tew,Department of Microbiology/Immunology, PO. Box 678, MCVNCU, Richmond,VA 23298-0678, USA

Abbreviatiom: AFC: Antibody-forming cell G C Germinal center HRP: Horseradish peroxidase PNA: Peanut agglutinin

8 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991

2.2 Immonizations

Animals were immunized with OVA (X A5-503, Sigma, St. Louis, MO) as previously described [161. The antigen was 0014-2980/91/1111-2697$3.50+ .25/0


L. I. Kupp, M.H. Kosco, H. A. Schenkein and J. G.%w

prepared in PBS (pH 7.2) at a concentration of 20 mg/ml. Mice were primed by injection of 100 pg of antigen in CFA behind the neck and challenged 2 weeks later with 100 pg of the same antigen preparation distributed among the four footpadsand 100 pgbehind the neck.Three weeks later, the mice were either used without further challenge (i.e. day 0 GC B cells) or given a third set of injections behind the neck and footpads with antigen dissolved in Dulbecco's PBS (approxhately 10 pghjection site).

Eur.J. Immunol. 1991.22: 2697-2701

2.6 Antibody preparations The IgG fraction of goat anti-human C5 Ab used in the study was purchased from Atlantic Antibodies (Scarborough, ME) and dialyzed extensively against PBS. mAb B220 :RA3-6B2 [26] was generously supplied by D. Conrad, Med. College of Virginia, Richmond,VA. The B220 and F4/80 [27] mAb were biotinylated according to the method described by Guesdon et al. [28] and stored at -70 "C.

2.3 Isolation of GC B cells 2.7 Light microscopic immunopemxidase labeling The procedure for GC B e l l isolation was reported previously [221and recently modified [23]. Mice were killed by cervicai dislocation and the draining popliteal, paraaortic, brachial, axillary and submandibular LN were excised. A peanut agglutinin (PNA)-horseradish peroxidase (HRP)conjugate (Sigma, L-0881) bound better to the plastic plates than PNA alone, and was used in the panning procedure. The excised LN were enzymatically digested and the cells placed onto a continuous Percoll gradient [23]. Cells comprising the second band from the bottom (e = 1.050-1.60 kg/l) were collected and washed in HBSS. The low-density lymphocytes were further separated on the basis of their ability to bind to PNAHR.Paated plates. Tissue culture-grade petri dishes (Falcon 3003, Oxnard, CA) were coated with 125 pg PNAHRP in 10 ml of a sodium carbonate-hydrogencarbonate coating buffer solution (pH 9.5) for 24 h at 4°C Nonbinding cells (PNA'O) were removed and the PNAhifraction was then eluted from the plates by incubating with 0.2M D-galactwe. The PNA+ cells which were eluted by the galactose were then washed twice and stored on ice. Cell viability was measured by trypan blue exclusion and was consistently >95%. The isolation procedure yielded a 10-fold enrichment for PNAhi cells relative to the starting population with a small population of contaminating IgD+ B cells and Tcells.


2 4 Generation of C-derived ehemotactic factors

Chemotactic activity in mouse or human sera was generated by the addition of zymosan or LPS (5 mg/ml of serum) and incubated at 37 "C in 5% C02 for 30 min. The serum was centrifuged at 400 x g and the SN was used in the Boyden chambers as the chemoattractant. rC5a was purchased from Sigma (# C-5788).

2.5 Mlgmtion of GC B cells The experiments were carried out using modified Boyden chambers [24] with 8-pn nitrocellulose filters (Sartorius, Gottingen, FRG) separating the GC Bcells and the chemoattractant. The chemoattractant was generally placed in the lower well of the chambers, but the concentrations were varied in both chambers for checkerboard analysis 1251. The Boyden chambers were prepared in triplicate and incubated for 3 h at 37°C in 5% C02. The filters were then fixed in formalin, stained, cleared in xyiene, and mounted on slides. Nine random high-power fields were counted per filter by the leading front technique 1251.

Filters after chemotaxis assays using 30% activated serum were fixed in formalin and incubated with PBS and 5% BSA for 2 h to inhibit nonspecific binding of Ab.The filters were then incubated with biotinylated mAb B220 or F480 (1 :100) for 1h. After washing the filters, avidin-HRP (Vector, Burlingame, CA; # A-2004) was added (1: 100) and allowed to incubate at ambient temperature for 1h with gentle rocking. The filters were then washed twice in PBS with 1%BSA followed with 0.1 M sodium cacodylate buffer (pH 7.4) wash. The filters were then submerged in 0.03% 3,3'-diaminobenzidine (Sigma, # D-5637) solution in 0.1 M cacodylate buffer for 30 min with 0.01% H202 for development as previously described [29]. After a 0.5% methyl green counterstain for 1min, the filters were dehydrated in a graded series of ethanols, cleared by xylene and mounted using Permount (Fisher Scientific Co.,Pittsburgh, PA) for observation at the light microscopiclevel. In addition, several filters were labeled with PNA-HRP and developed as described above. Biotinylated mAb MK-1 was used as an isotype control for

3220 (IgG2a). MK-1 is a rat anti-mouse mAb produced in our laboratory which reacts with follicular dendritic cells (K.Maeda, M. H. KOSCO,A. K. Szakal and J. G. Tew, unpublished observations). Several filters were examined for any reactivities with peroxidase-conjugated avidin D alone and for endogenous peroxidase activity in the absence of mAb or avidin. Photomicrographs were taken with an Olympus BH-2 microscope.

3 Results 3.1 Migration of GC B cells to LPS-and zymosan-activated serum

Initial experiments investigated whether GC B cells isolated 3 days after secondary immunization would migrate in response to activated serum (Fig. 1).Day 3 was selected since this is the time when GC are edematous [30] and the GC B cells are found in the thoracic duct lymph and blood [17]. By day 6 the GC are no longer edematous and GC B cells are no longer in the blood or thoracic duct and was included for a comparison. Zymosan and LPS were chosen as the activators because both are known to activate the alternative complement pathway and elicit production of chemotactic agents, and both of these activators are of microbial origin and may be found in inflammatory lesions. As indicated in Fig. 1, zymosan- and LPS-activated sera contain factors that induce the migration of day-3 GC B cells, but not day-6 cells. Zymosan-activated serum

Eur. J. Immunol. 1991. 21: 2697-2701

Chemotaxis of germinal center B cells to C5a


fronts, cells in the filters were examined for the GC B cell phenotype (B220+ and PNAh-). The photomicrograph shows typical cells with a motile morphology labeled with mAb B220 (Fig. 2). Analysis indicated that > 95% of the cells in the leading fronts were B220+ and PNAh- GC B cells. The isotype control showed no labeling.

3.3 Kinetics of the migration of GC B cells to zymosan-activated serum Control




Figure 1. Migration of day-3 and -6 GC B cells to activated serum. Control was buffered media. Mouse serum was activated either with zymosan (ZAS) or LPS (LAS) and diluted in medium to 30%. Results were calculated as the mean of triplicate filters k SD, *p < 0.01, Student’s t-test.

These experiments investigated whether the migration was associated with an early phase when GC B cells are in the thoracic duct, blood, and are in transit to BM to become plasma cells [17]. GC B cells were isolated at days 0,2,3,4, 5, 6 and 9 after secondary immunization and allowed to migrate towards factors in mouse activated serum (Fig. 3). The migration of GC B cells was only apparent between 2-4 days after secondary antigen challenge with a peak at day 3.

1 80



60 40


0 0







Dam after HcopullzJT immruriutfon Figure 3. Migration of GC B cells isolated at indicated days to activated serum. Mouse activated serum was used as the chemoattractant and was diluted to 30% in medium. GC B cells were obtained at days indicated after secondary immunization. Results were calculated as the mean of triplicate filters k SD, * p < 0.01, Student’s t-test.

Figure2. Photomicrograph of cells in the leading front labeled with the Bcell marker B220. Cells in a micropore filter were incubated with biotinylated mAb B220 followed with the addition of avidin-=. Development using diaminobenzidineverified that cells in the leading front were B cells. The cells also labeled with PNA-HRP indicating that these cells were PNAE B cells or GC B cells. Note the locomotor morphology of the labeled lymphocytes.The leading edge of the cell contains the nucleus followed by the cytoplasmic tail. Magnification x 1550. nhsQ

demonstrated consistently higher values and consequently was used for further study. 3.2 Light microscopy

The eqriched day-3 GC B cell preparations included cantaminating T cells and a few M@. To be certain GC B cells were responding to the chemoattractantsand in the leading

2 4 6 0 Aatlvatdmr~~rn


wm- ~n







Figure 4. Inhibition of day-3 GC B cell migration to activated serum and rC5a by anti-C5 antibody. nhs; nonactivated human semm: m; medium. Cells were obtained at day 3 after secondary immunization. Activated serum (30%) and rC5a (lo-* M) were used as attractants. The addition of anti-C5 antibody reduced the migration of the B cells to control levels in activated serum and rC5a. There were no statistically significant differences between the values for nonactivated serum and medium in this and other experiments. Results were calculated as the mean of triplicate filters rt SD.


L. I. Kupp, M. H. Kosco, H. A. Schenkein and J. G. Tew

Eur. J. Immunol. 1991.21: 2697-2701

3.4 C5a as a chemoattractant

4 Discussion

To test whether a C5-derived peptide was responsible for the migration observed, activated human serum or rC5a was placed into the lower compartment along with increasing amounts of anti-C5 antibody and allowed to incubate for 15 min before the assembly of the Boyden chambers (Fig. 4). As the graph indicates, the migration of day-3 B cell was inhibited in a dose-dependent fashion with goat polyclonal antLC5 antibody, implicating a C5-derived peptide in the migration mechanism. Furthermore, 10-8M rC5a was clearly an active chemoattractant and could similarly be blocked by the anti-C5 antibody. Normal goat IgG was used as a control at the 8-10 vg level and did not significantly alter migration (data not shown).

The results reported here indicate that GC B cells taken from a healthy animal are capable of chemotaxis in a C5a gradient. However, chemotactically active B cells were restricted to an early phase of the GC reaction (2-4 days after secondary immunization)when some cells are becoming AFC [16]. An anti-C5 antibody inhibited migration of day-3 GC B cells in response to zymosan-activated serum indicating that a C5-derived peptide was involved in the migration of these cells. The ability of a C5-derived chemoattractant to induce a chemotactic response in such cells was directly demonstrated using rC5a. Checkerboard analysis established that C5a could elicit both chemotactic and chemokinetic responses by day-3 GC B cells. Chemoattractants such as C5a in sites of chronic inflammation may help explain the mechanism by which recently activated B cells localize in such sites.

3.5 Checkerboard analysis of day3 GC B cells demonstrating chemokinetic and chemotactic responses

Interestingly, the time interval of greatest cell migration (2-4 days after secondary immunization) coincides with AFC localizing in BM [31], antibody formation by GC B cells [16], and GC B cells in the thoracic duct lymph and blood [17]. Benner et al. [31] demonstrated that the migration of B cells from the spleen into the BM takes place within the first few days after booster immunization. They discovered that a splenectomy 4 or more days after boost does not influence the BM antibody response. However, splenectomyon day 2 can completelyprevent the BM antibody production. DiLosa et al. [17] found lymphoblasts bearing the GC B phenotype in both the thoracic duct lymph and the peripheral blood 3 days following booster immunization.These day-3 lymphocytes produced specific antibody without further stimulation while those isolated on days 0, 8 or 11did not [17]. DiLosa et al. [17] injected GC B cells, isolated 3 days after booster immunization, into the tail veins of naive recipients and demonstrated that these donor-specific plasma cells localized in the BM of the recipients. These data support the concept that at day 3 some GC B cells leave the secondary lymphoid tissues, enter the circulation, and localize in the BM.

Chemotaxis is defined as directional migration requiring a positive gradient of attractant, whereas chemokinesis is defined as an undirected migratory response that leads to an increased cell migration in either a negative gradient or in the absence of a gradient. Chemotaxis can be distinguished from chemokinesis by varying the concentrations of the attractant both above and below the filter in the Boyden chambers [25]. Results of a checkerboard assay are presented in Table 1.The values along the diagonal, where there is no gradient, are clearly higher than the background level of migration [24 pm) demonstrating that the zymosanactivated serum stimulated a chemokinetic response. However, the GC B cells migrated further into the filters when the concentration of the chemoattractantwas greater below the filter (a positive gradient) than when the concentration was greater above the filter (solid arrows vs. dashed arrows). A highly sighcant positive slope (i.e. slope greater than zero) was apparent using the values beside the solid arrows in a regression analysis (p

Chemotaxis of germinal center B cells in response to C5a.

An infiltrate of B cells and plasma cells is characteristic of certain chronic inflammatory lesions. However, mechanisms involved in the local accumul...
640KB Sizes 0 Downloads 0 Views