Immunomodulation with monoclonal antibodies

Eur. J. Immunol. 1990. 20: 1457-1461

Siri Mjaaland and Sigbjern Fossum Anatomical Institute, University of Oslo, Oslo

1457

Modulation of immune responses with monoclonal antibodies I. Effects on regional lymph node morphology and on anti-hapten responses to haptenized monoclonal antibodies Repeated injections of monoclonal antibody (mAb) culture supernatants into rat footpads increased the weights of the draining lymph nodes. Immunostained freeze sections showed that injection of MRC 0 x 2 , a mAb reacting with rat follicular dendritic cells and MRC OX7 (anti-Thy-l.l), led to gross hypertrophy primarily of the follicular areas, whereas MRC OX6 (anti-rat major histocompatibility complex class I1 molecules) resulted in selective stimulation of the paracortex. These findings indicate that mAb, when conjugated to certain antigens,would modulate the immune response to these antigens. Consequencly, the mAb were conjugated with fluorescein isothiocyanate (FITC) and the humoral response against the hapten measured. The primary anti-FITC antibody response was tenfold stronger than after stimulation with FITC conjugated to a conventional carrier such as ovalbumin, and had some characteristics of a secondary response: a fast increase of IgG level to very high titers and a long duration without further amplification at later antigen challenges.

1 Introduction In vivo injections of antibodies are used to suppress the development of cell lines [l,21, delete cells or interfere with cell functions [3,4].We wanted to try to block retention of immune complexes or disrupt follicles by local injections of the mAb MRC OX2 (anti-follicular dendritic cell FDC [ 5 ] ) into the rat foot pads. Unexpectedly, we found a strong stimulation of the draining LN and particularly of germinal center formation. The specificity of this reaction was tested by injection of mAb against other non-lymphoid cells: MRC OX6 against rat MHC class I1 molecules [5]. MRC OX7 against rat Thy-1 molecules [6] and two control mAb not reactive with rat lymphoid tissues (W6/32; [7], and 7.6; Funderud, S., pers. communication). Here we report that footpad injection of mAb against different accessory cells of the LN leads to hypertrophy of the draining LN, with different mAb stimulating different regions of the nodes. We studied putative functional consequences of this stimulation by measuring primary and secondary anti-hapten (FITC) antibody formation after injecting haptenized mAb.

2 Materials and methods

following mAb were used (target specificities in brackets): MRC OX6 (rat MHC class I1 molecules; [S]), MRC OX7 (rat Thy-1 molecules; [6]) and MRC OX2 (rat FDC; [S]), W6/32 (human HLA class I molecules; [7]) and 7.6 (human HLA class I1 molecules; S. Funderud, personal communication). 2.2 FITC conjugation The mAb used for immunization were affinity purified on a protein A column. FITC (isomer 1; Sigma, St. Louis, MO) was conjugated to the four different mAb and to OVA (Sigma) as described by Brandtzaeg [8] with an IgG concentration of 14 mg/ml and addition of 8-16 pg FITC/mg protein (stirred for 1h at pH 9,5 at 25°C). Free fluorochrome was separated from conjugated by gel filtration (Sephadex G-50, Pharmacia, Uppsala, Sweden, equilibrated with 0,1 M Tris buffer / 0.3 M NaCl, pH 7.3), dialyzed for 40 h with PBS (pH 7.4,4"C) and the degree of conjugation was analyzed by measuring absorbance at 280 and 492 nm.

2.3 Injections

2.1 Animals and antibodies PVG (RT1") and (PVG x DA)FI (RTldaV1)hybrid rats of either sex were bred in conventional animal houses. The

[I 83311

*

This work was supported by the Norwegian Research Council for Science and the Humanities (NAVF), and Bergljot og Sigurd Skaugens Fond ti1 kreftens bekjempelse.

Correspondence: Siri Mjaaland, Anatomical Institute, Karl Johansgt. 47, N-0162 Oslo 1, Norway

Abbreviations: FDC: Follicular dendritic cell 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990

For morphological studies, the animals were injected into the hind footpads with 200 p1 tissue culture SN containing 3 mg/ml of protein. Each animal received the test antibody on one side and the control antibody (W6/32 or 7.6) on the other, the contralateral LN serving as internal controls.The animals were injected on day 1 , 3 , 5 and 7. On day 8 the rats were killed by ether overdosage. The draining popliteal LN were removed, weighed and immediately frozen in isopentane cooled in liquid nitrogen. For studies of the humoral responses, rats were immunized with the FITC-conjugated antibodies (IgG concentration: 30 pglml, fluorochrome/ protein ratio: 1.5; 200 pl in the hindfoot and 100 p1 in the forefoot) on both sides of the animal, with booster doses 0014-2980/90/0707-1457$3.50+ .25/0

1458

S. Mjaaland and S. Fossum

repeated every third week. Control animals received FITCOVA conjugates or unconjugated FITC dissolved in PBS.

Eur. J. Immunol. 1990. 20: 1457-1461

3 Results 3.1 Changes in morphology after injection of mAb

Repeated injections of culture SN of a mAb against F'DC (0x2)resulted in a threefold increase in weight of the draining LN (Table 1). An equivalent increase in weight Six 8-pm thick freeze serial sections taken from three was seen after injections with the control antibodies; different levels in each LN were stained with the monoculture SN of hybridoma W6/32 (Table 1) or affinityclonal peroxidase-antiperoxidasemethod after being fixed purified antibody 7.6 (data not shown) had both the same in acetone, washed four times in PBS and preincubated for 2 effect on the draining LN. Immunostained freeze sections min with 10% normal rabbit serum. The sections were then incubated overnight at 4 "C with the following primary revealed, however, marked differences in the morphology antibodies: OX2 (anti-rat FDC [5]), OX22 [anti-rat B cells of the LN. OX2 injections stimulated mainly the follicular (CD 45)] [9], OX6 (anti-rat MHC class I1 molecules [ 5 ] ) , areas, with development of numerous secondary follicles OX7 (anti-rat Thy-1 [6]), W3/13 [anti-rat Tcells (leukosia- with germinal centers (Fig. 2), whereas the follicular and lin) [lo]] and as a negative control W6/32 (anti-human the paracortical areas appeared to have increased approxMHC class I molecules [7]). The secondary antibody was a imately equally in size after injections of the control mAb rabbit anti-mouse Ig (Dakopatts, Copenhagen, Denmark) (W6/32 or 7.6; Fig. 2). These results prompted us to inject diluted 1/10, preabsorbed with rat LN cells, and to which mAb against other accessory cells. Injections of OX6 2% rat serum was added (incubation time 2 h at 25 "C). The against h4HC class I1 molecules constitutively present on third step consisted of adding a monoclonal mouse perox- interdigitating cells (IDC) resulted in a fivefold increase in idase-antiperoxidase complex (Clono-PAP, Sternberger weight, but here sectioning of the LN revealed a gross Meyer, Jarrettsville Pike, MD; 1/100 dilution, 1h, 25°C). selective hypertrophy of the paracortex. The follicular areas Peroxidase activity was demonstrated using a solution of seemed relatively unstimulated with few secondary follicles 0,1% 3,3-diaminobenzidine tetrachloride (Sigma) in and small germinal centers (Figs. 1 and 2). Finally, injec0,05 M Tris buffer (pH 7,6) and 0.01% H202 (Ferak tions of 0 x 7 , against Thy-1, gave a fourfold increase in weight. With this mAb we obtained an even stronger Laborat, GmbH, West Berlin) for 8 min. stimulation of the follicular areas with the formation of numerous large germinal centers (Fig. 1 and 2). 2.5 Morphometry

2.4 Immunohistochemistry

By the aid of a drawing tube fitted to the microscope, a composite drawing was made of the different areas in each of the three different levels in every LN based on the six immunohistochemically stained sections from each level. These areas were then measured by a digitalizing board (Calcomp type 9000) and computerized by a suited data program (T. Blackstad, personal communication). The percentages of T and B areas relative to the total LN area were then estimated for each LN.The presented values are median values from five to seven different LN obtained from four to seven different animals in each group.

2.6 Measurement of specific antibody in an ELISA Blood samples were taken on day 3,5,7,10 and 14 after the first and second immunization and tested for the presence of anti-FITC antibody by ELISA [ l l ] , with FITC-BSA (5 pg/ml) as antigen, adsorbed to the bottom of 96-well polyethylene terephthalate glycol assay plates (Cat.no. 6596, Costar Europe Ltd., Badhoevedorp, the Netherlands). Goat anti-rat IgG antibody (Sigma; 1/1000dilution) was used as secondary antibody and peroxidase-conjugated rabbit anti-goat Ig (Sigma; 1/1000 dilution) as the tertiary antibody. All incubations were performed at 25°C on a shaking platform for 1h. Between each incubation the wells were washed three times with PBS containing 0.05% Tween 20 (Ferak), whereas the final three washes were in 0.1 M citrate buffer, pH 5. Peroxidase activity was demonstrated using a solution with 0.1% o-phenylenediamine dihydrochloride (Sigma) in 0.1 M citrate buffer, pH 5 and 0.01% H202, and measured in a Titertek Multiscan Plus (Flow Laboratories Ltd., Rickmansworth, GB) at 492 nm.

3.2 Quantitative estimates of T and B cell areas The T and B cell areas were estimated on the immunostained serial freeze sections (Table 1).The relative changes in B vs. Tcell areas caused by the mAb against accessory cells were statistically significant (Wilcoxon-van Elteren test). The quantitative measurements (summarized in Fig. 3) confirmed that the considerable hypertrophy obtained with OX6 injections was caused mainly by increased paracortical volume, whereas OX7 and OX2 injections preferentially stimulated the follicles. OX2 gave a smaller increase in total weight than 0 x 7 , but a larger relative increase in the B cell areas than in the T cell areas.

3.3 Humoral responses The above results prompted the question whether the mAb could be exploited in modulating the immune response to a hapten conjugated to the mAb. We therefore haptenized the mAb by FITC conjugation and measured the production of anti-FITC serum antibody after footpad injections of the conjugates. Both the experimental animals (Fig. 4.A-C) and the controls (data not shown) gave normal primary IgM responses with a maximum antibody level on day 10 to 14. In contrast, the primary IgG responses were enhanced when the hapten was conjugated to the mAb against accessory cells. In particular, FITC-OX6 gave a vigorous primary IgG response (Fig. 4): as early as 4 days after immunization specific IgG antibody could be detected, and by day 10 the titer was tenfold higher than the maximum titers achieved in the secondary response of the controls (Fig. 4D). The antibody levels subsequently

Eur. J. Immunol. 1990. 20: 1457-1461

Immunomodulation with monoclonal antibodies

1459

Figure 1. Serial freeze sections from popliteal LN after footpad injections of the mAb OX6 (antirat MHC class I1 molecules, A and B) and OX7 (anti-rat Thy-1 molecules; C and D). The sections were stained by immunohistochemistry to visualize theTand B areas, using the pan-T antibody W3/13 (A and C), and OX6 against B cells (B and D) as primary antibodies.

0.89

0.59

1 L3

189

n

7 t

W6/32

OX 6

ox 7

ox

2

Figure 3. Relative volumes of whole LN, paracortex and follicles in test compared to normal (unstimulated) LN. Note the increase of paracortex after injections of OX6 and the increase of follicular areas after OX7 and OX2 injections. The numbers give the ratios of follicular to paracortical areas. LN = lymph node,T = paracortex, B = follicles.

Figure 2. Drawings of typical sections of LN from animals injected every second day for 1 week with hybridoma SN containing one of the following mAb: (A) 0 x 6 , (B) OX7 (C) OX2 and (D) W6/32. (E) illustrates a LN section from a non-injected animal. Note the large size of the experimental LN (A-C) compared with the controls (D, E), and the differences in sizes of paracortical area (A) compared with follicular areas (B, C). Note also the appearance of numerous secondary follicles with prominent germinal centers in (B). F = follicle, GC = germinal center, I = interfollicular area, M = medulla, T = T area (paracortex).

decreased, but were restored upon new immunizations.The secondary responses lasted longer than the primary, but the titers did not increase beyond the levels obtained in the primary response. Immunization with FITC-OX7 led to even higher antibody titers than with FITC-OX6 although it took somewhat longer, approximately 14 days, to reach maximum levels (Fig. 4B). As with FITC-OX6 the secondary responses lasted longer, but were no stronger than the primary response. The results from the FITC-OX2 immunization were less clear-cut: two of the four animals immunized showed a vigorous primary response, similar to that seen for the other two conjugates. The other two animals, however, showed a poor primary responses (Fig. 4C). To visualize this variability the results were split into two groups, one for the high response and one for the low.

4 Discussion The FCS proteins and the mouse Ig in the hybridoma culture SN caused antigenic stimulation with hypertrophy

1460

Eur. J. Immunol. 1990. 20: 1457-1461

S. Mjaaland and S. Fossum

Figure 4. Primary and secondary

(IgM

-0-

and IgG

-0-)

anti-

body responses after immunization

20

10

D a y s a f t e r immunization

with different FITC-mAb conjugates: (A) FITC-0x6, (B) FITC0 x 7 , (C) FITC-0x2, (D) Shows the controls FITC-W6/32 and FITC-OVA (--0--). where only the IgG response is noted. 1"and 2" indicate first and second immunization. Symbols represent medians and ranges from four rats in each group. The antibody response to FITC alone was not detectable.

Table 1. Changes in weight and morphology in the draining popliteal LN after repetitive local injections of different mAba)

mAb OX6 OX7 0x2 W6l32 No injection

% Tarea

% B arca

of total

of total

64.06 48.74h) 44.04h) 59.62

14.18 26.08c) 31.12c) 20.4

46.87

17.62

Weight of Weight of Weight of LN (mg) Tarea (mgk) B area (mg) 21 .o

16.6 11.4 11.2

13.45 8.09 5.02 6.67

2.97 4.32 3.54 2.24

4.1

1.92

0.72

of the popliteal LN when injected into the footpads. When OX6 and OX7 antibodies were used instead of nonspecific mAb, however, the responses were strongly augmented, e.g., OX6 caused a further twofold increase in LN weight. The presence of these antibodies did not merely give a nonspecific stimulatory effect, but biased the response with a gross hypertrophy of the paracortex in the case of OX6 and of the follicles in the case of 0 x 7 . Finally, when purified and haptenized, these two mAb gave a tenfold increase in anti-hapten antibody titer compared with the controls. Thus, the primary responses obtained with these conjugates had several of the characteristics of a traditional secondary response: a high titer and prolonged response, consisting predominantly of IgG antibody, and without further increase in titers upon subsequent stimulation. A s explanation for the adjuvant effects of OX6 and OX7 we postulate that the molecules act by increasing the efficiency of antigen presentation. These m A b bind to surface

a) Numbers indicate an average of estimates from three different levels in each node used, and the median value of these averages from five to seven different LN taken from four to seven different animals. The p values were estimated using Wilcoxon-van Elteren test, comparing experimental (0x6, 0 x 7 , 0 x 2 ) LN with control (W6/32): b) p