- Email: [email protected]
Nonimmune lymphocyte-macrophage interaction
CELLULAR
92,265-276 (1985)
IMMUNOLOGY
Nonimmune
Lymphocyte-Macrophage
1. Quantification
by an Automated
Interaction
Calorimetric
Assay
ANITA S.-F. CHONG AND CHRISTOPHERR. PARISH Department of Microbiology, John Curtin School of Medical Research, Australian National University, Canberra, ACT 2601, Australia Received November 23, 1984; accepted December 31, 1984 Previous studies have demonstrated a spontaneous, nonimmune interaction between lymphocytes and macrophages. This paper describes an automated calorimetric assay based on the dye, rose bengal, to quantify thii interaction. The procedure entails allowing lymphocytes to adhere to preformed macrophage monolayers in the wells of microplates and then staining bound lymphocytes with rose bengal. Dye uptake and the consequent number of lymphocytes bound were quantified using an automated spectrophotometer developed for reading microplates. This procedure was used to confirm and extend the basic pammeters of the system. The interaction was found to be temperature dependent but the kinetics and percentage of cells binding varied with the source of lymphocytes. However, all lymphocyte populations tested, namely, mature and immature thymocytes, T and B lymphocytes, and a range of thymoma cell lines, bound to macrophages. Furthermore, all macrophage populations examined had the ability to bind lymphocytes. The interaction also showed no strain specificity and generally lacked species specificity. It is proposed that the interaction is a highly dynamic process that enables lymphocytes to scan the surface of macrophages for self and/or foreign antigens.
0 1985 Academic
FWs, Inc.
INTRODUCTION Macrophages have been shown to be very important for the generation of many immune functions in vitro, such as T-lymphocyte activation by alloantigens, soluble antigens, and mitogens, as well as activation of B lymphocytes. The mechanism by which macrophages function in these interactions is not well understood. Numerous observations, both in vivo and in vitro, in a number of animal species, have demonstrated a spontaneous, antigen-independent, lymphocyte-macrophage interaction (l-9). The possibility that this interaction may mediate the initial physical association between lymphocytes and macrophages was postulated by Lipsky and Rosenthal (3, 4, 10). However, studies of this interaction have been hampered by the slow, tedious, and subjective microscopic assay used. The procedure entails forming macrophage monolayers on microscope slides, adding lymphocytes and after lymphocyte adhesion washing, fixing, and staining the slides. Lymphocyte binding is then assessedunder the light microscope. The procedure has the additional disadvantage that the vigorous washing procedures used tend to dissociate the less avid interactions. We have developed an automated, calorimetric assay based on the dye, rose 265 000%8749/85 $3.00 Copyright 0 1985 by Academic Ftw, Inc. All rights of reproduction in any form rimed.
266
CHONG AND PARISH
bengal, to quantify this interaction which overcomes these problems. Using this assay, we have confirmed, to a large extent, previous work and have used the method to further characterize the interaction. MATERIALS
AND METHODS
Animals. All mice, rats, and guinea pigs were bred at the John Curtin School of Medical Research. Mice were used as donors of lymphocytes and macrophages from 4 to 12 weeks of age. Fetal mice were obtained from pregnant mice 16 and 19 days after the appearance of a plug. Newborn mice were obtained from litters at the specified number of days after birth. Preparation of macrophage and lymphocyte suspensions. Thioglycolate-induced peritoneal exudates were produced by intraperitoneal (ip) injections of 2.5 ml of 3% (w/v) thioglycolate (Difco, Detroit, Mich.) solution 4-14 days previously. The thioglycolate was aged at room temperature for l-2 months before use. Cells were harvested from thioglycolate-induced or normal mice by the ip injection of 10 ml of ice-cold Puck’s saline. Macrophages were pelleted twice by centrifugation at 4°C and resuspendedin Eagle’sminimum essential medium (F15; Grand Island Biological Co., Grand Island, N.Y.) containing 1% fetal calf serum (FCS) and buffered with 2.5 m&f Hepes (ULTROL, Calbiochem-Behring Corp., La Jolla, Calif.). For all experiments, unless otherwise stated, data presentedwere obtained using thioglycolateinduced macrophages. Single-cell suspensions of different lymphoid organs were prepared in F15/ 1% FCS by forcing organ fragments through a fine wire mesh, followed by centrifugation on a cushion of Isopaque-Ficoll to remove red and dead cells (11). Following three washes in Fl5/1% FCS cells were resuspended to the required concentration in Hepes-buffered medium. Fractionation of lymphocyte populations. Thymocytes were fractionated into peanut agglutinin (PNA)-positive and -negative cells by the method described by Draber and Kisielow (12). Thymocytes (2 X 10’) resuspended in 0.25 ml of medium were mixed with an equal volume of a solution of 1 mg/ml PNA. After 10 min coincubation at room temperature, the suspension was layered onto 14 ml of F15/ 20% FCS in a 15-ml round-bottomed glass tube. After 20 min at room temperature, the PNA-positive (bottom layer) and -negative (top layer) cells were collected with a Pasteur pipet, washed once in 0.15 M r+galactose in phosphate-buffered saline (PBS), and then incubated in 0.15 A4 Dgalactose/PBS for 20 min at room temperature before washing twice in medium. Hydrocortisone-resistant thymocytes were obtained from mice injected ip with 0.2 ml/mouse of 25 mg/ml hydrocortisone acetate (Rousel, England) 24 hr prior to the collection of the thymus. Spleen cell suspensionswere separatedinto Ig-negative and Ig-positive populations by rosetting surface Ig-bearing cells with sheep anti-mouse Ig-coupled erythrocytes and separating rosetting and nonrosetting cells on Isopaque-Ficoll (13). The Ignegative cells were collected as a band at the Isopaque-Ficoll interface. The Igpositive cells sedimented to the bottom of the gradient and were recovered by lysing the erythrocytes with 40% isotonic, Hanks’ balanced salt solution (HBSS), bringing the mixture to isotonicity with lo-fold concentrated HBSS and washing the cells once with F15/1% FCS. Viable cells were collected by an additional centrifugation on Isopaque-Ficoll. Spleen cells were depleted of Thy-l+ cells by being treated with monoclonal rat anti-Thy- 1.2 antibody (clone 30H12) and guinea pig complement (14) and the
LYMPHOCYTE-MACROPHAGE
INTERACTION
267
resulting dead cells being removed by Isopaque-Ficoll centrifugation. Nylon woolnonadherent cells were obtained by incubating cells with nylon wool in Fl5/ 10% FCS medium for 30 min at 37’C. Nonadherent cells were eluted by flushing the column with 10 ml of F15/10% FCS at 37°C under unit gravity. Cell lines. The macrophage cell lines PU-5-1.8 and J-774 were cultured until confluent in Dulbecco’s minimum essential medium (H 16; Grand Island Biological Co.) containing 10% FCS and supplemented with 0.1 mM I--asparagine. The continuous fibroblast cell line, L929, was cultured in Eagle’s modified minimum essential medium (Autopow, Flow Laboratories, Australasia Pty. Ltd., Cat. No. 1l110-24) supplemented with 10% bovine serum, 0.036% NaHCO~,O.O03M NaOH, 0.14% L-Glutamine, and 0.023% Tris buffer (pH 7.6). After washing off excess medium with PBS the monolayers were trypsinized with 0.0025% (w/v) trypsin in HBSS at 37°C to obtain single-cell suspensions. Cells were washed and resuspended in Fl5/1% FCS and viability assessedby trypan blue exclusion. Thymoma cell lines Rl (TL+), Rl (TL-) and BW5147 were cultured in Fl5/10% FCS, whereas the EL4 thymoma line was cultured in H16/10% FCS. All cells were grown in 75-cm2 sterile tissue culture flasks (Coming Glassware, Coming, N.Y.). Calorimetric assay for quantifying binding of lymphocytes to macrophages. Macrophage monolayers were prepared in the wells of 96-well round-bottomed tissue culture microplates (Catalog No. 76-013-05; Flow Laboratories, McLean, Va.) by adding 100 &well of a suspension of macrophages (106/ml) in Hepes-buffered F15/1% FCS. In addition, 10% heat-denatured FCS was added to the medium to prevent nonspecific sticking of lymphocytes to plastic. Heat-denatured FCS was prepared by incubating lo-ml aliquots of FCS for 15 min at 80°C. This preparation was stored at 4°C and used within one week as aggregatesappeared in the solution after a week and tended to displace macrophages from plastic. The microplates were incubated overnight at 37°C in a sealed container. Nonadherent cells were then flicked off and fresh medium was added to each well with a multichannel pipette. Different numbers of lymphocytes were added to the macrophage monolayers ( 100 pi/well) and left to incubate at 37°C for 90 min unless otherwise stated. All treatments were done in triplicate. Nonadherent lymphocytes were then thrown off and the adherent cells present in the wells stained with 0.25% (w/v) rose bengal (C.I. 45440, Hopkin and Williams, Chadwell Heath, Essex, England) in PBS (100 &well) for 3 min at room temperature. Excess dye was thrown off and the microplates washed twice by immersion in a PBS bath at room temperature. After a brief draining of the tray face downward, stain was released from intact cells by the addition of 200 &well of 50% ethanol in PBS. Following dispersion of the dye in each well with a multichannel pipette, dye uptake by cells was measured by absorbance (OD) at 570 nm using a microplate reader (Model MR600; Dynatech Laboratories, Alexandria, Va.) with the reference wave length set at 630 nm. For each tray the machine was blanked against a well that contained medium alone and was treated with dye. The OD values obtained from the microplate reader represented the amount of dye taken up by both macrophages and lymphocytes in each well. The number of lymphocytes bound per well (X 105) was calculated according to the formula (T - C)/X, where T = mean OD of wells containing macrophages and lymphocytes, C = mean OD of wells containing macrophages and X = OD of IO5 lymphocytes (usually 0.20-0.25 OD units). The X value was determined for each experiment as follows: Serial dilutions (100 ~1) of lymphocytes (initial concentration 5 X 106/ml)
268
CHONG AND PARISH
were aliquoted into the wells of a V-bottomed !&well tray, pelleted by centrifugation at 3OOgfor 2 min and the medium was thrown off. Pellets were resuspended in 50 &well of 0.25% rose bengal, left for 3 min at room temperature and, after addition of 100 &well of PBS, were again sedimented by centrifugation. Excess dye was removed by washing the cells twice more with 200 &well of PBS before releasing bound dye with 50% ethanol/PBS and reading OD in the microplate reader. RESULTS Calorimetric Assay for Quantifying Lymphocyte-Macrophage Interaction An automated calorimetric assay was developed for rapidly measuring the spontaneous interaction of lymphocytes with macrophages. The assaywas based on the ability of the dye, rose bengal, to rapidly and strongly stain lymphocytes bound to preformed macrophage monolayers in the wells of microplates. Dye uptake and consequently number of lymphocytes bound was quantified using an automated spectrophotometer developed for reading microplates. Although macrophages also stained with rose bengal, the assay was sufficiently sensitive and reproducible to detect the binding of comparatively small numbers of lymphocytes, i.e., changes of 0.1 OD units or approximately 5 X lo4 lymphocytes bound. Furthermore, staining of the macrophages allowed for a continual check of the stability of the monolayers, a particularly important point when metabolic inhibitors were being used. Figure 1 shows the results of a typical binding experiment. The data are presented as either OD values or transformed into the number and percentage of thymocytes bound. The results reveal an almost linear relationship between the number of thymocytes bound and the number of thymocytes added per well, suggesting that macrophages were not limiting the interaction. The percentage of thymocytes bound confirmed this point. However, subsequent experiments revealed that greater than
-5.0
100
u) ‘0 ;
--LO
SO
4
--3.0
iz Y
e 50
z -
--2.0
40
Y E s
--1.0
20
E
0
-0 I
0
1
,625
1.25
2.5
1
5.0 THYC/WELL
0
I
.625
1.25
25
5.0
Xd
FIG. 1. Binding of varying numbers of BALB/c thymocytes to syngeneic thioglycolate-induced macrophages (lO’/well). Varying numbers of viable thymocytes were added to 24-hr macrophage monolayers and incubated for 90 min at either 20°C (left) or 37“C (right). Data are presented as OD (W), number of thymocytes bound (A) and percentage of thymocytes bound (0). A control of nonspecific binding of thymocytes to plastic is shown (0). Each point represents the mean of three wells with vertical bars indicating the standard errors.
LYMPHOCYTE-MACROPHAGE
INTEXACTION
269
5 X lo5 thymocytes/well resulted in crowding effects and a rapid decline in the proportion of bound cells (data not shown). The specificity of the interaction was clearly evident from the inability of thymocytes to bind to plastic. Such a control with thymocytes in wells without macrophages was included in all experiments and when nonspecific binding to plastic exceeded lo%, the experiment was discarded. Finally, the data in Fig. 1 indicate that the percentage of thymocytes bound at 20°C (approximately 80%) was consistently higher than at 37°C (approximately 60%). The temperature dependence of the interaction is considered in more detail in a later section. Important Assay Variables At this point four important assay variables should be highlighted. (1) The type of plastic microplate used to form the macrophage monolayers was found to be of critical importance. The most stable monolayers were obtained when roundbottomed, 96-well polystyrene tissue culture trays (Catalog No. 76-013-05) from Flow Laboratories were used. However, even different batches of trays from this manufacturer varied somewhat in their adherence properties. (2) All experiments were carried out in tissue culture medium buffered with Hepes rather than sodium bicarbonate. Preliminary experiments established that Hepes buffering eliminated the edge effect observed in some microplates where macrophages and lymphocytes adhered less effectively to the outermost wells. Such edge effects have been observed previously in enzyme-linked immunosorbent assays(ELISA) using plastic microplates ( 15, 16). (3) Preliminary experiments also demonstrated that lo5 macrophages/well resulted in the most sensitive assay for lymphocyte binding and more stable macrophage monolayer-s were obtained when they were incubated overnight before use. (4) Finally, the nonspecific binding of lymphocytes (particularly splenic B cells) to plastic posed a problem but was overcome by forming macrophage monolayers in the presence of 10% heat-denatured FCS. Temperature Dependence and Kinetics of the Lymphocyte-Macrophage Interaction Figure 2 depicts the effect of temperature on the kinetics of the lymphocytemacrophage interaction, as determined by the automated calorimetric assay. In these experiments, macrophage monolayers were preformed overnight at 37”C, fresh medium was added, and the monolayers were allowed to equilibrate at the experimental temperatures for 30 min before addition of the lymphocytes. As observed in Fig. 1, more thymocytes bound to macrophages at 20°C than at 37°C after the first hour of coincubation and by 4 hr at 37°C virtually all thymocytes had detached from the monolayer. This contrasted with the kinetics observed at 20°C suggesting that the interaction observed involved both a binding and release of cells. It must be noted, however, that the maximum percentage of cells bound and the net rate of release of thymocytes varied somewhat from experiment to experiment, although little variation was observed within each experiment. Spleen cells gave a pattern of interaction different from that of thymocytes, with fewer cells binding, the interaction being slower, and there being little or no difference between binding at 20 and 37°C. In addition, unlike thymocytes, there
270
CHONG AND PARISH loo-
e z Y id "
2o"
80-
1
+j\
37O
so
lYJ-
60
40-
40
20-
20
"k,
J'
x
o-
i I 0
J /
J I 1
t \-
I/:;;. 0 I 2
I 3
I 4
4"
J I 0
B I I 1
I 2
I 3
I 4
TIME-HRS FIG. 2. The kinetics of the interaction between BALB/c lymphocytes and syngeneic macrophages. Viable thymocytes (0) or splenocytes (0) at 4 X lO’/well were incubated with macrophage monolayers at either 20 or 37°C. Each point represents the mean of three wells and the vertical bars the standard errors.
was no net dissociation of spleen cells from the macrophage monolayer after 4 hr incubation at 37°C. On the basis of these results, to optimize lymphocyte binding it was decided to coincubate macrophages and lymphocytes for 90 min in all subsequent experiments. Nature of Cells Participating in Lymphocyte-Macrophage Interaction Lymphocytes from primary (thymus, bone marrow) and secondary (spleen, lymph nodes) lymphoid organs were tested for their ability to bind to macrophages (Table 1). It was found that lymphocytes from all lymphoid organs could bind to macrophages, but thymus had a consistently higher percentage of binding cells (maximum of 87.5%), compared with other lymphoid populations (32.5-63.6s). Furthermore, when the thymus was separated into subsets by PNA agglutination and hydrocortisone acetate treatment, it was shown that all subsets bound to macrophages, although hydrocortisone-resistant subsets bound with a somewhat lower efficiency. It is possible, however, to ascribe this difference to the nonspecific effect of hydrocortisone on the thymocytes. A proportion of both T and B cells separated from spleen by Ig rosetting, nylonwool passage,and anti-Thy- 1.2 antibody and complement treatment were capable of binding to macrophages, although the B-cell populations showed a slightly higher percentage of binding cells than the T-cell preparations when the experiment was performed at 37°C. Another feature of Table 1 is that all spleen cell populations bound to macrophages equally well at 20 and 37°C whereas thymus, lymph node, and bone marrow cells bound less at 37°C. Next, a variety of cell lines were tested for their ability to participate in the interaction (Table 2). It was found that all four different thymoma lines bound to macrophages, although their reactivity varied somewhat, EL-4 being the most reactive (72.4% bound) and BW5147 the least (57.2%). The Rl (TL-) cell line, which was reported by Hyman and Stallings (17) to be lacking in both TL and
LYMPHOCYTE-MACROPHAGE
271
INTERACTION
TABLE 1 Comparison of Binding Capacity of Different Lymphocyte Populations to Syngeneic Macrophages Lymphocytes bound (96) Lymphocytes
4°C
20°C
37°C
BALB/c thymus Unfiactionated Hydrocortisone resistant PNAPNA+
31.6 zk 3.8 N.D. N.D. N.D.
87.5 + 5.9 63.2 f 6.3 71.7 +: 4.1 75.4 zk 2.8
62.3 rt 49.5 + 59.2 2 59.8 +
Spleen Unfractionated Nylon wool nonadherent Ig negative Ig positive Thy- 1.2 negative
23.9 + 1.1 N.D. N.D. N.D. N.D.
44.9 k 34.5 iz 44.9 + 45.5 + 47.3 +
48.6 f 6.0 37.0 + 1.9 37.1 312.7 43.8 2 4.9 32.5 + 1.2
Lymph node
21.2 f 4.4
63.6 -+ 12.3
43.0 + 10.3
Bone marrow
22.1 + 2.0
55.0 2 10.3
38.3 + 6.3
4.4 0.9 5.9 5.6 8.8
5.4 4.6 7.4 9.4
Note.Serial dilutions of lymphocytes (0.63-5.0 X 10s/well) were added to 24hr preformed monolayers of thioglycolate-induced BALB/c macrophages for 90 min at 4,20, or 37°C. The results are expressed as the mean percentages of lymphocytes bound for each titration + standard errors. N.D. = not done.
H-2 antigen expression, showed comparable binding ability (63.1% bound) to its Rl (TL+) counterpart (68.9%). Similarly, it was found that there was no significant difference in the binding capacity of resident (69.1%) or thioglycolate-elicited (63.4%) macrophages. Both the macrophage cell line, J-774 (54.3%) and PU-5-1.8 (46.6%) could also bind thymocytes, although with a lower efficiency. The fibroblast cell line L929, however, could not bind thymocytes, suggesting that only cells of the macrophage lineage could interact with lymphocytes. TABLE 2 Comparison of the Ability of Different Cell Lines to Participate in the Lymphocyte-Macrophage Interaction Cell population Lymphocyte
Macrophage
EL-4 thymoma cell line RI (TL+) thymoma cell line RI (TL-) thymoma cell line BW5 147 thymoma cell line
BALB/c BALB/c BALB/c BALB/c
BALB/c thymocytes
Normal resident peritoneal macrophage Thioglycolateelicited peritoneal macrophage PU-5-1.8 macrophage cell line J-774 macrophage cell line L929 fibroblast cell line
BALB/c thymocytes BALB/c thymocytes BALB/c thymocytes BALB/c thymocytes
macrophage monolayer macrophage monolayer macrophage monolayer macrophage monolayer
Thymccytes bound VJ) 72.4 + 68.9 + 63.1 + 57.2 f
5.9 11.2 15.4 6.8
69.1 + 8.5 63.4 f 8.5 46.6 + 2.6 54.3 -t 5.1 1.9 f 0.13
272
CHONG AND PARISH
Strain and Species Specijicity of Lymphocyte-Macrophage Interaction The ability of thymocytes from eight different mouse strains to interact with CBA/H macrophages is presented in Table 3. There was no significant difference in the percentage of thymocytes bound for all the strains tested, suggesting no strain specificity in the interaction. Next, we investigated the species specificity of the interaction. A complex but highly reproducible pattern of reactivity was observed (Table 4). The most striking feature of these experiments was the high degree of cross-reactive binding between mouse, rat, and guinea pig cells. There was, however, one clearcut case of species specificity. Mouse thymocytes showed a hierarchy of reactivity with macrophages from the three species, reacting most strongly with mouse, less effectively with rat, and weakly with guinea pig macrophages. In contrast, this hierarchy of reactivity was not observed with mouse spleen cells. Surprisingly, there were two casesof lymphocytes binding more strongly to xenogeneic than isogeneic macrophages, namely rat thymocytes and guinea pig spleen cells binding to mouse macrophages. Ontogeny of Lymphocytes Involved in the Lymphocyte-Macrophage Interaction To follow the development of lymphocytes capable of binding to macrophages, lymphocytes from the liver, thymus, and spleen of fetal and newborn mice were assayed(Table 5). Fetal liver lymphocytes could bind macrophages as early as Day 16 gestation (37.0% bound) and the percentage of binding cells remained apparently constant until birth. Similarly, fetal thymus contained a proportion of binding cells and resembled adult thymus. In the spleen, however, there was a very low percentage of binding cells in newborn mice, but the proportion rapidly increased by Day 2, and approached adult levels by Day 6 after birth. DISCUSSION Although there have been several attempts in the past to analyze and attach a functional significance to the nonimmune interaction between lymphocytes and TABLE 3 Strain Specificity of Thymocyte-Macrophage Interaction Strain of origin of thymocytes CBA/H BALB/c DBA/ 1 DBA/2 SJL/J
A/J C57Bl/iOJ BlO.BR C3H.OH
Thymocytes bound (S) 72.8 + 66.8 k 60.5 + 67.5 f 73.3 f 78.9 f 70.2 + 74.8 + 65.2 f
2.8 5.0 4.5 6.3 4.8 6.0 4.5 5.6 4.3
Nofe. Serial dilutions of lymphocytes (0.63-5.0 X 105/well) were added to 24-hr preformed CBA/H macrophage monolayers for 90 min at 37OC. Results are expressed as mean percentages of lymphocytes bound for each titration + standard errors.
LYMPHOCYTE-MACROPHAGE
273
INTERACTION
TABLE 4 Species Specificity of Thymocyte-Macrophage Interaction Lymphocytes bound (96)
Lymphocyt=
Mouse macrophages
Rat macrophages
Guinea pig macrophages
Thymocytes Mouse Rat Guinea pig
16.4 + 4. I 83.9 + 1.1 87.2 + 4.7
46.6 + 3.3 47.7 + 0.5 N.D.
14.4 f 0.7 N.D. 95.4 +- 4.3
Splenocytes Mouse Rat Guinea pig
48.8 xk 7.5 49.1 f 4.7 77.7 * 4.5
46.6 rf: 1.6 39.7 + 0.6 N.D.
36.2 + 2.5 N.D. 48.3 -c 3.7
Nofe. Serial dilutions of lymphocytes (0.63-5.0 X 105/well) were added to 24-hr preformed monolayers of thioglycolate-induced BALB/c macrophages for 90 min at 37°C. The results are expressedas the mean percentages of lymphocytes bound for each titration + standard errors. N.D. = not done.
macrophages, these studies have been hampered by the lack of a fast and objective means of measuring the interaction. This paper describes an automated calorimetric assay for measuring this interaction and is based on the ability of the dye, rose bengal, to stain lymphocytes bound to macrophage monolayers in the wells of microplates. Initial experiments using the calorimetric method to investigate the kinetics, temperature dependence, and percentage of lymphocytes binding (Table 1, Figs. 1 and 2) gave comparable results to those previously reported by Seigal (1, 2) Lipsky and Rosenthal (3, 4, 10) and Lopez et al., (5, 6) using a microscopic procedure. Further experiments (Table 3) contirmed previous reports that this interaction is not strain specific (6, 18). The data presented in Table 1 indicate that both cortical and medullary thymocytes, as well as peripheral T and B cells (spleen and lymph node) can bind to macrophages, TABLE 5 Comparison of Ability of Lymphocytes from Fetal and Newborn BALB/c Mice to Participate in the Lymphocyte-Macrophage Interaction Lymphocytes bound (%y Age (days)
Fetal liver
Thymus
Spleen
16’ 18” Ob 2b 46 6b
37.0 f 4.6 34.7 f 4.6 43.9 f 6.1 N.D. N.D. N.D.
N.D. 69.6 -t 4.4 71.1 + 6.4 N.D. N.D. 66.0 f 4.7
N.D. N.D. 13.6 z!c2.9 38.1 f 5.0 38.4 iz 4.0 45.5 +- 6.2
DAge of fetus calculated from day of appearance of plug. ’ Age of newborn mice calculated from day of birth. ‘Serial dilutions of lymphocytes (0.63-5.0 X lO’/well) were added to 24-hr preformed monolayers of thioglycolate-induced BALB/c macrophages for 90 min at 37°C. The results am expressed as the mean percentages of lymphocytes bound for each titration rt standard errors. N.D. = not done.
274
CHONG AND PARISH
although thymus yielded the highest percentage of binding cells. Thus, both immature and mature T cells appear to interact with macrophages, although there have been earlier reports that mature T cells, namely cortisone-resistant thymoqtes (6) and splenic T cells (18) cannot bind to macrophages. However, these reports contradict work showing that glass or nylon wool-passaged splenocytes (5) and lymph node cells (3, 4, 10) can interact with macrophages. Furthermore, a comparison of four thymoma cell lines (Table 2) demonstrated that binding ability did not correlate with the maturation status of the tumour cell. Thus EL-4, which expresses the characteristics of a mature thymocyte (i.e., low Thy-l expression and hydrocortisone resistance) bound macrophages as well as the BW5 147 and Rl lines that exhibit the immature thymocyte phenotype (19). On the other hand, Wu and Thomas (9) recently reported that EL-4 does not react with macrophages. A likely explanation for this discrepancy is that the calorimetric assay described in this paper does not involve vigorous washing procedure, whereas the microscopic method used by others entailed extensive washing and fixation. Thus, the old procedure may not detect lower avidity interactions between lymphocytes and macrophages. The interaction was shown to be independent of H-2 and Tla antigen expression by the thymocytes, since the RI (TL-) variant line bound as well as its parental Rl (TL+) line, and also independent of Ia expression on the macrophages, since Ianegative thioglycolate-induced macrophages and PU-5- 1.8 and J-774 cell lines could bind thymocytes (Table 2). However, the fibroblast cell line L929 could not bind thymocytes, suggesting a specificity for cells of the macrophage lineage similar to that reported by Lipsky and Rosenthal (3). The noninvolvement of Ia antigens was further confirmed when polyclonal anti-Ia antisera and monoclonal anti-Ia. did not block the interaction (data not shown). However, results obtained by Agrwal and Thomas (20) indicate that thymocyte binding is, in part, dependent upon macrophage Ia expression. The main cause of this discrepancy could lie in differences in the assay methods used, the binding assay of Agrwal and Thomas (20) only detecting the most avid interactions as their procedure employed a short incubation time (30 min at 37”C), serumfree medium, fixation of the conjugates, and vigorous washing. Furthermore, in the following paper (21), we provide evidence that this interaction is, in fact, mediated by a lymphocyte receptor specifically recognizing a sulfated polysaccharide on the macrophage surface. On the other hand, it is interesting to note that Sant et al. (22) recently reported that sulfated polysaccharides are tightly associated with Ia antigens. The studies described in this paper demonstrated a major difference between thymocytes and spleen cells in the kinetics of their interaction with macrophages. Thymocytes bound reversibly to the macrophage monolayers at 37”C, whereas spleen cells appeared to react more slowly and no net dissociation occurred after prolonged culture (Fig. 2). A possible explanation for this difference is that spleen cells bind irreversibly to macrophages, although this is unlikely as time lapse cinematography revealed reversible binding between splenocytes and macrophages (A. S.-F. Chong, unpublished observations). A more likely explanation is that spleen cells can bind and dissociate from macrophages on multiple occasions, whereas the majority of thymocytes are only capable of interacting with macrophages on a single occasion.
LYMPHOCYTE-MACROPHAGE
INTERACTION
275
An important implication of the above discussion is that the cell-binding assay is only measuring the net association between lymphocytes and macrophages at any one time point, the interaction representing a dynamic equilibrium between binding and release events as suggested by Lipsky and Rosenthal (3). Thus, the lower percentage of binding cells detected in some lymphocyte populations could simply reflect reduced binding avidity and shorter binding time, rather than a lower proportion of cells being capable of binding to macrophages. In fact, the data presented in this paper indicate that lymphocytes at all stagesof development can interact with macrophages, ranging from fetal liver and bone marrow cells to peripheral T and B lymphocytes. The only exception was newborn spleen cells (Table 5), an interesting point that warrants further investigation. Analysis of the species specificity of the lymphocyte-macrophage interaction revealed substantial cross-reactivity across species barriers (Table 4), with the only clearcut species specificity being the reaction of mouse thymocytes with different macrophages. Similar findings have been reported by Siegel (2) and Lopez et al. (5), although the strong reaction of guinea pig thymocytes with mouSe macrophages was not demonstrated. However, strong speciesspecificity exhibited by mouse thymocytes was not observed in the spleen-macrophage interaction, thus defining another difference between the thymus-macrophage and spleen-macrophage interactions. It has been postulated that the thymocyte-macrophage interaction results in the maturation of immature thymocytes. This hypothesis was based on the observation that immature thymocytes that bound to macrophages lost their ability to rebind (6), became immunocompetent, and acquired the surface phenotype of mature T cells (23-25). However, limiting dilution experiments by Chen et al. (26) and Ceredig et al. (27) showed that thymocytes with the immature phenotype are unable to mature and function. Similarly, thymocytes that were coincubated with macrophagesfor 2-24 hr showed no change in their H-2 and Thy- 1 surfaceantigen expression (A. S.-F. Chong, unpublished data), again demonstrating no maturation. A possible explanation for the apparent maturation of thymocytes during coincubation experiments is that mature thymocytes are responsible for the effects, macrophages increasing the survival and proliferative potential of these cells. Recent data reported by Young (28), showing that macrophages can substantially augment the response of thymocytes to concanavalin A, support this concept. Furthermore, we have found that any increase in phytomitogen responsiveness of thymocytes cultured with macrophage monolayers could be explained by contamination of the thymocytes with macrophages (A. S.-F. Chong, unpublished data). In contrast to the maturation concept discussed above, we postulate that all classes of lymphocytes can bind to macrophages and the interaction is a process which allows lymphocytes to search the surface of macrophages for antigen. In the case of immature lymphocytes, such as the majority of thymocytes, recognition of antigen results in tolerance induction and hence loss of ability to rebind to macrophages. On the other hand, mature lymphocytes could use the interaction to recognize and respond to foreign antigens on the macrophage surface as postulated initially by Lipsky and Rosenthal (3, 4, 10). REFERENCES 1. Siegal, I., J. Allergy 44, 190, 1969. 2. Siegal, I., J. Immunol. 105, 879, 1970.
276
CHONG AND PARISH
3. Lipsky, P. E., and Rosenthal, A. S., .Z Exp. Med. 138, 900, 1973. 4. Lipsky, P. E., and Rosenthal, A. S., J. Exp. Med. 141, 138, 1975. 5. Lopez, L R., Johansen, K. S., Radovich, J., and Talmage, D. W., J. Allergy Clin. Zrnmunol. 53, 336, 1974. 6. Lopez, L. R., Vatter, A. E., and Talmage, D. W., J. Zmmunol. 119, 1668, 1977. 7. Fan, P. T., Yu, D. T. Y., Clements, P. J., and Bluestone, R., Clinical Zmmunol. Zmmunopathol. 9, 363, 1978. 8. Kyewski, B. A., Rouse, R. U., and Kaplan, H. S., Proc. Natl. Acad. Sci. USA 79, 5646, 1982. 9. Wu, S. K., and Thomas, D. W., J. Zmmunol. 131,2110, 1983. 10. Lipsky, P. E., and Rosenthal, A. S., J. Zmmunol. 115, 440, 1975. 11. Davidson, W. F., and Parish, C. R., J Zmmunol. Methods 7,291, 1975. 12. Draber, P., and Kisielow, P., Eur. J Zmmunol. 11, 1, 1981. 13. Parish, C. R., Kirov, S. M., Bowem, N., and Blanden, R. V., Eur. J. Zmmunol. 4, 808, 1974. 14. McKenzie, I. F. C., and Parish, C. R., .Z.Exp. Med. 144, 847, 1976. 15. Denmark, J. R., and Chessurn, B. S., Lancet 1, 161, 1978. 16. Burt, S. M., Carter, T. J. N., and Kricka, L. J., J. Zmmunol. Methods 31, 231, 1979. 17. Hyman, R., and Stallings, V., Immunogenetics 3, 75, 1976. 18. G’Toole, M. M., and Wortis, H. H., J. Zmmunol. 124,2010, 1980. 19. Ralph, P., J. Zmmunol. 110, 1470, 1973. 20. Agrwal, N., and Thomas, D. W., Cell. Zmmunol. &I, 352, 1984. 21. Chong, A. S.-F., and Parish, C. R., Cell. Zmmunol. 92, 277, 1985. 22. Sant, A. J., Schwartz, B. D., and Cullen, S. E., Z. Exp. Med. 158, 1979, 1983. 23. Mosier, D. E., and Pierce, C. W., J. Exp. Med. 136, 1484, 1972. 24. van den Tweel, J. G., and Walker, W. S., Immunology 33, 817, 1977. 25. Beller, D. I., and Unanue, E. R., J. Zmmunol. 121, 1861, 1978. 26. Chen, W. F., Scollay, R., and Shortman, K., J. Zmmunol. 129, 18, 1982. 27. Ceredig, R., Glasebrook, A. L., and MacDonald, H. R., .Z.Exp. Med. 155, 358, 1982. 28. Young, B., Immunology 46, 31, 1981.
92,265-276 (1985)
IMMUNOLOGY
Nonimmune
Lymphocyte-Macrophage
1. Quantification
by an Automated
Interaction
Calorimetric
Assay
ANITA S.-F. CHONG AND CHRISTOPHERR. PARISH Department of Microbiology, John Curtin School of Medical Research, Australian National University, Canberra, ACT 2601, Australia Received November 23, 1984; accepted December 31, 1984 Previous studies have demonstrated a spontaneous, nonimmune interaction between lymphocytes and macrophages. This paper describes an automated calorimetric assay based on the dye, rose bengal, to quantify thii interaction. The procedure entails allowing lymphocytes to adhere to preformed macrophage monolayers in the wells of microplates and then staining bound lymphocytes with rose bengal. Dye uptake and the consequent number of lymphocytes bound were quantified using an automated spectrophotometer developed for reading microplates. This procedure was used to confirm and extend the basic pammeters of the system. The interaction was found to be temperature dependent but the kinetics and percentage of cells binding varied with the source of lymphocytes. However, all lymphocyte populations tested, namely, mature and immature thymocytes, T and B lymphocytes, and a range of thymoma cell lines, bound to macrophages. Furthermore, all macrophage populations examined had the ability to bind lymphocytes. The interaction also showed no strain specificity and generally lacked species specificity. It is proposed that the interaction is a highly dynamic process that enables lymphocytes to scan the surface of macrophages for self and/or foreign antigens.
0 1985 Academic
FWs, Inc.
INTRODUCTION Macrophages have been shown to be very important for the generation of many immune functions in vitro, such as T-lymphocyte activation by alloantigens, soluble antigens, and mitogens, as well as activation of B lymphocytes. The mechanism by which macrophages function in these interactions is not well understood. Numerous observations, both in vivo and in vitro, in a number of animal species, have demonstrated a spontaneous, antigen-independent, lymphocyte-macrophage interaction (l-9). The possibility that this interaction may mediate the initial physical association between lymphocytes and macrophages was postulated by Lipsky and Rosenthal (3, 4, 10). However, studies of this interaction have been hampered by the slow, tedious, and subjective microscopic assay used. The procedure entails forming macrophage monolayers on microscope slides, adding lymphocytes and after lymphocyte adhesion washing, fixing, and staining the slides. Lymphocyte binding is then assessedunder the light microscope. The procedure has the additional disadvantage that the vigorous washing procedures used tend to dissociate the less avid interactions. We have developed an automated, calorimetric assay based on the dye, rose 265 000%8749/85 $3.00 Copyright 0 1985 by Academic Ftw, Inc. All rights of reproduction in any form rimed.
266
CHONG AND PARISH
bengal, to quantify this interaction which overcomes these problems. Using this assay, we have confirmed, to a large extent, previous work and have used the method to further characterize the interaction. MATERIALS
AND METHODS
Animals. All mice, rats, and guinea pigs were bred at the John Curtin School of Medical Research. Mice were used as donors of lymphocytes and macrophages from 4 to 12 weeks of age. Fetal mice were obtained from pregnant mice 16 and 19 days after the appearance of a plug. Newborn mice were obtained from litters at the specified number of days after birth. Preparation of macrophage and lymphocyte suspensions. Thioglycolate-induced peritoneal exudates were produced by intraperitoneal (ip) injections of 2.5 ml of 3% (w/v) thioglycolate (Difco, Detroit, Mich.) solution 4-14 days previously. The thioglycolate was aged at room temperature for l-2 months before use. Cells were harvested from thioglycolate-induced or normal mice by the ip injection of 10 ml of ice-cold Puck’s saline. Macrophages were pelleted twice by centrifugation at 4°C and resuspendedin Eagle’sminimum essential medium (F15; Grand Island Biological Co., Grand Island, N.Y.) containing 1% fetal calf serum (FCS) and buffered with 2.5 m&f Hepes (ULTROL, Calbiochem-Behring Corp., La Jolla, Calif.). For all experiments, unless otherwise stated, data presentedwere obtained using thioglycolateinduced macrophages. Single-cell suspensions of different lymphoid organs were prepared in F15/ 1% FCS by forcing organ fragments through a fine wire mesh, followed by centrifugation on a cushion of Isopaque-Ficoll to remove red and dead cells (11). Following three washes in Fl5/1% FCS cells were resuspended to the required concentration in Hepes-buffered medium. Fractionation of lymphocyte populations. Thymocytes were fractionated into peanut agglutinin (PNA)-positive and -negative cells by the method described by Draber and Kisielow (12). Thymocytes (2 X 10’) resuspended in 0.25 ml of medium were mixed with an equal volume of a solution of 1 mg/ml PNA. After 10 min coincubation at room temperature, the suspension was layered onto 14 ml of F15/ 20% FCS in a 15-ml round-bottomed glass tube. After 20 min at room temperature, the PNA-positive (bottom layer) and -negative (top layer) cells were collected with a Pasteur pipet, washed once in 0.15 M r+galactose in phosphate-buffered saline (PBS), and then incubated in 0.15 A4 Dgalactose/PBS for 20 min at room temperature before washing twice in medium. Hydrocortisone-resistant thymocytes were obtained from mice injected ip with 0.2 ml/mouse of 25 mg/ml hydrocortisone acetate (Rousel, England) 24 hr prior to the collection of the thymus. Spleen cell suspensionswere separatedinto Ig-negative and Ig-positive populations by rosetting surface Ig-bearing cells with sheep anti-mouse Ig-coupled erythrocytes and separating rosetting and nonrosetting cells on Isopaque-Ficoll (13). The Ignegative cells were collected as a band at the Isopaque-Ficoll interface. The Igpositive cells sedimented to the bottom of the gradient and were recovered by lysing the erythrocytes with 40% isotonic, Hanks’ balanced salt solution (HBSS), bringing the mixture to isotonicity with lo-fold concentrated HBSS and washing the cells once with F15/1% FCS. Viable cells were collected by an additional centrifugation on Isopaque-Ficoll. Spleen cells were depleted of Thy-l+ cells by being treated with monoclonal rat anti-Thy- 1.2 antibody (clone 30H12) and guinea pig complement (14) and the
LYMPHOCYTE-MACROPHAGE
INTERACTION
267
resulting dead cells being removed by Isopaque-Ficoll centrifugation. Nylon woolnonadherent cells were obtained by incubating cells with nylon wool in Fl5/ 10% FCS medium for 30 min at 37’C. Nonadherent cells were eluted by flushing the column with 10 ml of F15/10% FCS at 37°C under unit gravity. Cell lines. The macrophage cell lines PU-5-1.8 and J-774 were cultured until confluent in Dulbecco’s minimum essential medium (H 16; Grand Island Biological Co.) containing 10% FCS and supplemented with 0.1 mM I--asparagine. The continuous fibroblast cell line, L929, was cultured in Eagle’s modified minimum essential medium (Autopow, Flow Laboratories, Australasia Pty. Ltd., Cat. No. 1l110-24) supplemented with 10% bovine serum, 0.036% NaHCO~,O.O03M NaOH, 0.14% L-Glutamine, and 0.023% Tris buffer (pH 7.6). After washing off excess medium with PBS the monolayers were trypsinized with 0.0025% (w/v) trypsin in HBSS at 37°C to obtain single-cell suspensions. Cells were washed and resuspended in Fl5/1% FCS and viability assessedby trypan blue exclusion. Thymoma cell lines Rl (TL+), Rl (TL-) and BW5147 were cultured in Fl5/10% FCS, whereas the EL4 thymoma line was cultured in H16/10% FCS. All cells were grown in 75-cm2 sterile tissue culture flasks (Coming Glassware, Coming, N.Y.). Calorimetric assay for quantifying binding of lymphocytes to macrophages. Macrophage monolayers were prepared in the wells of 96-well round-bottomed tissue culture microplates (Catalog No. 76-013-05; Flow Laboratories, McLean, Va.) by adding 100 &well of a suspension of macrophages (106/ml) in Hepes-buffered F15/1% FCS. In addition, 10% heat-denatured FCS was added to the medium to prevent nonspecific sticking of lymphocytes to plastic. Heat-denatured FCS was prepared by incubating lo-ml aliquots of FCS for 15 min at 80°C. This preparation was stored at 4°C and used within one week as aggregatesappeared in the solution after a week and tended to displace macrophages from plastic. The microplates were incubated overnight at 37°C in a sealed container. Nonadherent cells were then flicked off and fresh medium was added to each well with a multichannel pipette. Different numbers of lymphocytes were added to the macrophage monolayers ( 100 pi/well) and left to incubate at 37°C for 90 min unless otherwise stated. All treatments were done in triplicate. Nonadherent lymphocytes were then thrown off and the adherent cells present in the wells stained with 0.25% (w/v) rose bengal (C.I. 45440, Hopkin and Williams, Chadwell Heath, Essex, England) in PBS (100 &well) for 3 min at room temperature. Excess dye was thrown off and the microplates washed twice by immersion in a PBS bath at room temperature. After a brief draining of the tray face downward, stain was released from intact cells by the addition of 200 &well of 50% ethanol in PBS. Following dispersion of the dye in each well with a multichannel pipette, dye uptake by cells was measured by absorbance (OD) at 570 nm using a microplate reader (Model MR600; Dynatech Laboratories, Alexandria, Va.) with the reference wave length set at 630 nm. For each tray the machine was blanked against a well that contained medium alone and was treated with dye. The OD values obtained from the microplate reader represented the amount of dye taken up by both macrophages and lymphocytes in each well. The number of lymphocytes bound per well (X 105) was calculated according to the formula (T - C)/X, where T = mean OD of wells containing macrophages and lymphocytes, C = mean OD of wells containing macrophages and X = OD of IO5 lymphocytes (usually 0.20-0.25 OD units). The X value was determined for each experiment as follows: Serial dilutions (100 ~1) of lymphocytes (initial concentration 5 X 106/ml)
268
CHONG AND PARISH
were aliquoted into the wells of a V-bottomed !&well tray, pelleted by centrifugation at 3OOgfor 2 min and the medium was thrown off. Pellets were resuspended in 50 &well of 0.25% rose bengal, left for 3 min at room temperature and, after addition of 100 &well of PBS, were again sedimented by centrifugation. Excess dye was removed by washing the cells twice more with 200 &well of PBS before releasing bound dye with 50% ethanol/PBS and reading OD in the microplate reader. RESULTS Calorimetric Assay for Quantifying Lymphocyte-Macrophage Interaction An automated calorimetric assay was developed for rapidly measuring the spontaneous interaction of lymphocytes with macrophages. The assaywas based on the ability of the dye, rose bengal, to rapidly and strongly stain lymphocytes bound to preformed macrophage monolayers in the wells of microplates. Dye uptake and consequently number of lymphocytes bound was quantified using an automated spectrophotometer developed for reading microplates. Although macrophages also stained with rose bengal, the assay was sufficiently sensitive and reproducible to detect the binding of comparatively small numbers of lymphocytes, i.e., changes of 0.1 OD units or approximately 5 X lo4 lymphocytes bound. Furthermore, staining of the macrophages allowed for a continual check of the stability of the monolayers, a particularly important point when metabolic inhibitors were being used. Figure 1 shows the results of a typical binding experiment. The data are presented as either OD values or transformed into the number and percentage of thymocytes bound. The results reveal an almost linear relationship between the number of thymocytes bound and the number of thymocytes added per well, suggesting that macrophages were not limiting the interaction. The percentage of thymocytes bound confirmed this point. However, subsequent experiments revealed that greater than
-5.0
100
u) ‘0 ;
--LO
SO
4
--3.0
iz Y
e 50
z -
--2.0
40
Y E s
--1.0
20
E
0
-0 I
0
1
,625
1.25
2.5
1
5.0 THYC/WELL
0
I
.625
1.25
25
5.0
Xd
FIG. 1. Binding of varying numbers of BALB/c thymocytes to syngeneic thioglycolate-induced macrophages (lO’/well). Varying numbers of viable thymocytes were added to 24-hr macrophage monolayers and incubated for 90 min at either 20°C (left) or 37“C (right). Data are presented as OD (W), number of thymocytes bound (A) and percentage of thymocytes bound (0). A control of nonspecific binding of thymocytes to plastic is shown (0). Each point represents the mean of three wells with vertical bars indicating the standard errors.
LYMPHOCYTE-MACROPHAGE
INTEXACTION
269
5 X lo5 thymocytes/well resulted in crowding effects and a rapid decline in the proportion of bound cells (data not shown). The specificity of the interaction was clearly evident from the inability of thymocytes to bind to plastic. Such a control with thymocytes in wells without macrophages was included in all experiments and when nonspecific binding to plastic exceeded lo%, the experiment was discarded. Finally, the data in Fig. 1 indicate that the percentage of thymocytes bound at 20°C (approximately 80%) was consistently higher than at 37°C (approximately 60%). The temperature dependence of the interaction is considered in more detail in a later section. Important Assay Variables At this point four important assay variables should be highlighted. (1) The type of plastic microplate used to form the macrophage monolayers was found to be of critical importance. The most stable monolayers were obtained when roundbottomed, 96-well polystyrene tissue culture trays (Catalog No. 76-013-05) from Flow Laboratories were used. However, even different batches of trays from this manufacturer varied somewhat in their adherence properties. (2) All experiments were carried out in tissue culture medium buffered with Hepes rather than sodium bicarbonate. Preliminary experiments established that Hepes buffering eliminated the edge effect observed in some microplates where macrophages and lymphocytes adhered less effectively to the outermost wells. Such edge effects have been observed previously in enzyme-linked immunosorbent assays(ELISA) using plastic microplates ( 15, 16). (3) Preliminary experiments also demonstrated that lo5 macrophages/well resulted in the most sensitive assay for lymphocyte binding and more stable macrophage monolayer-s were obtained when they were incubated overnight before use. (4) Finally, the nonspecific binding of lymphocytes (particularly splenic B cells) to plastic posed a problem but was overcome by forming macrophage monolayers in the presence of 10% heat-denatured FCS. Temperature Dependence and Kinetics of the Lymphocyte-Macrophage Interaction Figure 2 depicts the effect of temperature on the kinetics of the lymphocytemacrophage interaction, as determined by the automated calorimetric assay. In these experiments, macrophage monolayers were preformed overnight at 37”C, fresh medium was added, and the monolayers were allowed to equilibrate at the experimental temperatures for 30 min before addition of the lymphocytes. As observed in Fig. 1, more thymocytes bound to macrophages at 20°C than at 37°C after the first hour of coincubation and by 4 hr at 37°C virtually all thymocytes had detached from the monolayer. This contrasted with the kinetics observed at 20°C suggesting that the interaction observed involved both a binding and release of cells. It must be noted, however, that the maximum percentage of cells bound and the net rate of release of thymocytes varied somewhat from experiment to experiment, although little variation was observed within each experiment. Spleen cells gave a pattern of interaction different from that of thymocytes, with fewer cells binding, the interaction being slower, and there being little or no difference between binding at 20 and 37°C. In addition, unlike thymocytes, there
270
CHONG AND PARISH loo-
e z Y id "
2o"
80-
1
+j\
37O
so
lYJ-
60
40-
40
20-
20
"k,
J'
x
o-
i I 0
J /
J I 1
t \-
I/:;;. 0 I 2
I 3
I 4
4"
J I 0
B I I 1
I 2
I 3
I 4
TIME-HRS FIG. 2. The kinetics of the interaction between BALB/c lymphocytes and syngeneic macrophages. Viable thymocytes (0) or splenocytes (0) at 4 X lO’/well were incubated with macrophage monolayers at either 20 or 37°C. Each point represents the mean of three wells and the vertical bars the standard errors.
was no net dissociation of spleen cells from the macrophage monolayer after 4 hr incubation at 37°C. On the basis of these results, to optimize lymphocyte binding it was decided to coincubate macrophages and lymphocytes for 90 min in all subsequent experiments. Nature of Cells Participating in Lymphocyte-Macrophage Interaction Lymphocytes from primary (thymus, bone marrow) and secondary (spleen, lymph nodes) lymphoid organs were tested for their ability to bind to macrophages (Table 1). It was found that lymphocytes from all lymphoid organs could bind to macrophages, but thymus had a consistently higher percentage of binding cells (maximum of 87.5%), compared with other lymphoid populations (32.5-63.6s). Furthermore, when the thymus was separated into subsets by PNA agglutination and hydrocortisone acetate treatment, it was shown that all subsets bound to macrophages, although hydrocortisone-resistant subsets bound with a somewhat lower efficiency. It is possible, however, to ascribe this difference to the nonspecific effect of hydrocortisone on the thymocytes. A proportion of both T and B cells separated from spleen by Ig rosetting, nylonwool passage,and anti-Thy- 1.2 antibody and complement treatment were capable of binding to macrophages, although the B-cell populations showed a slightly higher percentage of binding cells than the T-cell preparations when the experiment was performed at 37°C. Another feature of Table 1 is that all spleen cell populations bound to macrophages equally well at 20 and 37°C whereas thymus, lymph node, and bone marrow cells bound less at 37°C. Next, a variety of cell lines were tested for their ability to participate in the interaction (Table 2). It was found that all four different thymoma lines bound to macrophages, although their reactivity varied somewhat, EL-4 being the most reactive (72.4% bound) and BW5147 the least (57.2%). The Rl (TL-) cell line, which was reported by Hyman and Stallings (17) to be lacking in both TL and
LYMPHOCYTE-MACROPHAGE
271
INTERACTION
TABLE 1 Comparison of Binding Capacity of Different Lymphocyte Populations to Syngeneic Macrophages Lymphocytes bound (96) Lymphocytes
4°C
20°C
37°C
BALB/c thymus Unfiactionated Hydrocortisone resistant PNAPNA+
31.6 zk 3.8 N.D. N.D. N.D.
87.5 + 5.9 63.2 f 6.3 71.7 +: 4.1 75.4 zk 2.8
62.3 rt 49.5 + 59.2 2 59.8 +
Spleen Unfractionated Nylon wool nonadherent Ig negative Ig positive Thy- 1.2 negative
23.9 + 1.1 N.D. N.D. N.D. N.D.
44.9 k 34.5 iz 44.9 + 45.5 + 47.3 +
48.6 f 6.0 37.0 + 1.9 37.1 312.7 43.8 2 4.9 32.5 + 1.2
Lymph node
21.2 f 4.4
63.6 -+ 12.3
43.0 + 10.3
Bone marrow
22.1 + 2.0
55.0 2 10.3
38.3 + 6.3
4.4 0.9 5.9 5.6 8.8
5.4 4.6 7.4 9.4
Note.Serial dilutions of lymphocytes (0.63-5.0 X 10s/well) were added to 24hr preformed monolayers of thioglycolate-induced BALB/c macrophages for 90 min at 4,20, or 37°C. The results are expressed as the mean percentages of lymphocytes bound for each titration + standard errors. N.D. = not done.
H-2 antigen expression, showed comparable binding ability (63.1% bound) to its Rl (TL+) counterpart (68.9%). Similarly, it was found that there was no significant difference in the binding capacity of resident (69.1%) or thioglycolate-elicited (63.4%) macrophages. Both the macrophage cell line, J-774 (54.3%) and PU-5-1.8 (46.6%) could also bind thymocytes, although with a lower efficiency. The fibroblast cell line L929, however, could not bind thymocytes, suggesting that only cells of the macrophage lineage could interact with lymphocytes. TABLE 2 Comparison of the Ability of Different Cell Lines to Participate in the Lymphocyte-Macrophage Interaction Cell population Lymphocyte
Macrophage
EL-4 thymoma cell line RI (TL+) thymoma cell line RI (TL-) thymoma cell line BW5 147 thymoma cell line
BALB/c BALB/c BALB/c BALB/c
BALB/c thymocytes
Normal resident peritoneal macrophage Thioglycolateelicited peritoneal macrophage PU-5-1.8 macrophage cell line J-774 macrophage cell line L929 fibroblast cell line
BALB/c thymocytes BALB/c thymocytes BALB/c thymocytes BALB/c thymocytes
macrophage monolayer macrophage monolayer macrophage monolayer macrophage monolayer
Thymccytes bound VJ) 72.4 + 68.9 + 63.1 + 57.2 f
5.9 11.2 15.4 6.8
69.1 + 8.5 63.4 f 8.5 46.6 + 2.6 54.3 -t 5.1 1.9 f 0.13
272
CHONG AND PARISH
Strain and Species Specijicity of Lymphocyte-Macrophage Interaction The ability of thymocytes from eight different mouse strains to interact with CBA/H macrophages is presented in Table 3. There was no significant difference in the percentage of thymocytes bound for all the strains tested, suggesting no strain specificity in the interaction. Next, we investigated the species specificity of the interaction. A complex but highly reproducible pattern of reactivity was observed (Table 4). The most striking feature of these experiments was the high degree of cross-reactive binding between mouse, rat, and guinea pig cells. There was, however, one clearcut case of species specificity. Mouse thymocytes showed a hierarchy of reactivity with macrophages from the three species, reacting most strongly with mouse, less effectively with rat, and weakly with guinea pig macrophages. In contrast, this hierarchy of reactivity was not observed with mouse spleen cells. Surprisingly, there were two casesof lymphocytes binding more strongly to xenogeneic than isogeneic macrophages, namely rat thymocytes and guinea pig spleen cells binding to mouse macrophages. Ontogeny of Lymphocytes Involved in the Lymphocyte-Macrophage Interaction To follow the development of lymphocytes capable of binding to macrophages, lymphocytes from the liver, thymus, and spleen of fetal and newborn mice were assayed(Table 5). Fetal liver lymphocytes could bind macrophages as early as Day 16 gestation (37.0% bound) and the percentage of binding cells remained apparently constant until birth. Similarly, fetal thymus contained a proportion of binding cells and resembled adult thymus. In the spleen, however, there was a very low percentage of binding cells in newborn mice, but the proportion rapidly increased by Day 2, and approached adult levels by Day 6 after birth. DISCUSSION Although there have been several attempts in the past to analyze and attach a functional significance to the nonimmune interaction between lymphocytes and TABLE 3 Strain Specificity of Thymocyte-Macrophage Interaction Strain of origin of thymocytes CBA/H BALB/c DBA/ 1 DBA/2 SJL/J
A/J C57Bl/iOJ BlO.BR C3H.OH
Thymocytes bound (S) 72.8 + 66.8 k 60.5 + 67.5 f 73.3 f 78.9 f 70.2 + 74.8 + 65.2 f
2.8 5.0 4.5 6.3 4.8 6.0 4.5 5.6 4.3
Nofe. Serial dilutions of lymphocytes (0.63-5.0 X 105/well) were added to 24-hr preformed CBA/H macrophage monolayers for 90 min at 37OC. Results are expressed as mean percentages of lymphocytes bound for each titration + standard errors.
LYMPHOCYTE-MACROPHAGE
273
INTERACTION
TABLE 4 Species Specificity of Thymocyte-Macrophage Interaction Lymphocytes bound (96)
Lymphocyt=
Mouse macrophages
Rat macrophages
Guinea pig macrophages
Thymocytes Mouse Rat Guinea pig
16.4 + 4. I 83.9 + 1.1 87.2 + 4.7
46.6 + 3.3 47.7 + 0.5 N.D.
14.4 f 0.7 N.D. 95.4 +- 4.3
Splenocytes Mouse Rat Guinea pig
48.8 xk 7.5 49.1 f 4.7 77.7 * 4.5
46.6 rf: 1.6 39.7 + 0.6 N.D.
36.2 + 2.5 N.D. 48.3 -c 3.7
Nofe. Serial dilutions of lymphocytes (0.63-5.0 X 105/well) were added to 24-hr preformed monolayers of thioglycolate-induced BALB/c macrophages for 90 min at 37°C. The results are expressedas the mean percentages of lymphocytes bound for each titration + standard errors. N.D. = not done.
macrophages, these studies have been hampered by the lack of a fast and objective means of measuring the interaction. This paper describes an automated calorimetric assay for measuring this interaction and is based on the ability of the dye, rose bengal, to stain lymphocytes bound to macrophage monolayers in the wells of microplates. Initial experiments using the calorimetric method to investigate the kinetics, temperature dependence, and percentage of lymphocytes binding (Table 1, Figs. 1 and 2) gave comparable results to those previously reported by Seigal (1, 2) Lipsky and Rosenthal (3, 4, 10) and Lopez et al., (5, 6) using a microscopic procedure. Further experiments (Table 3) contirmed previous reports that this interaction is not strain specific (6, 18). The data presented in Table 1 indicate that both cortical and medullary thymocytes, as well as peripheral T and B cells (spleen and lymph node) can bind to macrophages, TABLE 5 Comparison of Ability of Lymphocytes from Fetal and Newborn BALB/c Mice to Participate in the Lymphocyte-Macrophage Interaction Lymphocytes bound (%y Age (days)
Fetal liver
Thymus
Spleen
16’ 18” Ob 2b 46 6b
37.0 f 4.6 34.7 f 4.6 43.9 f 6.1 N.D. N.D. N.D.
N.D. 69.6 -t 4.4 71.1 + 6.4 N.D. N.D. 66.0 f 4.7
N.D. N.D. 13.6 z!c2.9 38.1 f 5.0 38.4 iz 4.0 45.5 +- 6.2
DAge of fetus calculated from day of appearance of plug. ’ Age of newborn mice calculated from day of birth. ‘Serial dilutions of lymphocytes (0.63-5.0 X lO’/well) were added to 24-hr preformed monolayers of thioglycolate-induced BALB/c macrophages for 90 min at 37°C. The results am expressed as the mean percentages of lymphocytes bound for each titration rt standard errors. N.D. = not done.
274
CHONG AND PARISH
although thymus yielded the highest percentage of binding cells. Thus, both immature and mature T cells appear to interact with macrophages, although there have been earlier reports that mature T cells, namely cortisone-resistant thymoqtes (6) and splenic T cells (18) cannot bind to macrophages. However, these reports contradict work showing that glass or nylon wool-passaged splenocytes (5) and lymph node cells (3, 4, 10) can interact with macrophages. Furthermore, a comparison of four thymoma cell lines (Table 2) demonstrated that binding ability did not correlate with the maturation status of the tumour cell. Thus EL-4, which expresses the characteristics of a mature thymocyte (i.e., low Thy-l expression and hydrocortisone resistance) bound macrophages as well as the BW5 147 and Rl lines that exhibit the immature thymocyte phenotype (19). On the other hand, Wu and Thomas (9) recently reported that EL-4 does not react with macrophages. A likely explanation for this discrepancy is that the calorimetric assay described in this paper does not involve vigorous washing procedure, whereas the microscopic method used by others entailed extensive washing and fixation. Thus, the old procedure may not detect lower avidity interactions between lymphocytes and macrophages. The interaction was shown to be independent of H-2 and Tla antigen expression by the thymocytes, since the RI (TL-) variant line bound as well as its parental Rl (TL+) line, and also independent of Ia expression on the macrophages, since Ianegative thioglycolate-induced macrophages and PU-5- 1.8 and J-774 cell lines could bind thymocytes (Table 2). However, the fibroblast cell line L929 could not bind thymocytes, suggesting a specificity for cells of the macrophage lineage similar to that reported by Lipsky and Rosenthal (3). The noninvolvement of Ia antigens was further confirmed when polyclonal anti-Ia antisera and monoclonal anti-Ia. did not block the interaction (data not shown). However, results obtained by Agrwal and Thomas (20) indicate that thymocyte binding is, in part, dependent upon macrophage Ia expression. The main cause of this discrepancy could lie in differences in the assay methods used, the binding assay of Agrwal and Thomas (20) only detecting the most avid interactions as their procedure employed a short incubation time (30 min at 37”C), serumfree medium, fixation of the conjugates, and vigorous washing. Furthermore, in the following paper (21), we provide evidence that this interaction is, in fact, mediated by a lymphocyte receptor specifically recognizing a sulfated polysaccharide on the macrophage surface. On the other hand, it is interesting to note that Sant et al. (22) recently reported that sulfated polysaccharides are tightly associated with Ia antigens. The studies described in this paper demonstrated a major difference between thymocytes and spleen cells in the kinetics of their interaction with macrophages. Thymocytes bound reversibly to the macrophage monolayers at 37”C, whereas spleen cells appeared to react more slowly and no net dissociation occurred after prolonged culture (Fig. 2). A possible explanation for this difference is that spleen cells bind irreversibly to macrophages, although this is unlikely as time lapse cinematography revealed reversible binding between splenocytes and macrophages (A. S.-F. Chong, unpublished observations). A more likely explanation is that spleen cells can bind and dissociate from macrophages on multiple occasions, whereas the majority of thymocytes are only capable of interacting with macrophages on a single occasion.
LYMPHOCYTE-MACROPHAGE
INTERACTION
275
An important implication of the above discussion is that the cell-binding assay is only measuring the net association between lymphocytes and macrophages at any one time point, the interaction representing a dynamic equilibrium between binding and release events as suggested by Lipsky and Rosenthal (3). Thus, the lower percentage of binding cells detected in some lymphocyte populations could simply reflect reduced binding avidity and shorter binding time, rather than a lower proportion of cells being capable of binding to macrophages. In fact, the data presented in this paper indicate that lymphocytes at all stagesof development can interact with macrophages, ranging from fetal liver and bone marrow cells to peripheral T and B lymphocytes. The only exception was newborn spleen cells (Table 5), an interesting point that warrants further investigation. Analysis of the species specificity of the lymphocyte-macrophage interaction revealed substantial cross-reactivity across species barriers (Table 4), with the only clearcut species specificity being the reaction of mouse thymocytes with different macrophages. Similar findings have been reported by Siegel (2) and Lopez et al. (5), although the strong reaction of guinea pig thymocytes with mouSe macrophages was not demonstrated. However, strong speciesspecificity exhibited by mouse thymocytes was not observed in the spleen-macrophage interaction, thus defining another difference between the thymus-macrophage and spleen-macrophage interactions. It has been postulated that the thymocyte-macrophage interaction results in the maturation of immature thymocytes. This hypothesis was based on the observation that immature thymocytes that bound to macrophages lost their ability to rebind (6), became immunocompetent, and acquired the surface phenotype of mature T cells (23-25). However, limiting dilution experiments by Chen et al. (26) and Ceredig et al. (27) showed that thymocytes with the immature phenotype are unable to mature and function. Similarly, thymocytes that were coincubated with macrophagesfor 2-24 hr showed no change in their H-2 and Thy- 1 surfaceantigen expression (A. S.-F. Chong, unpublished data), again demonstrating no maturation. A possible explanation for the apparent maturation of thymocytes during coincubation experiments is that mature thymocytes are responsible for the effects, macrophages increasing the survival and proliferative potential of these cells. Recent data reported by Young (28), showing that macrophages can substantially augment the response of thymocytes to concanavalin A, support this concept. Furthermore, we have found that any increase in phytomitogen responsiveness of thymocytes cultured with macrophage monolayers could be explained by contamination of the thymocytes with macrophages (A. S.-F. Chong, unpublished data). In contrast to the maturation concept discussed above, we postulate that all classes of lymphocytes can bind to macrophages and the interaction is a process which allows lymphocytes to search the surface of macrophages for antigen. In the case of immature lymphocytes, such as the majority of thymocytes, recognition of antigen results in tolerance induction and hence loss of ability to rebind to macrophages. On the other hand, mature lymphocytes could use the interaction to recognize and respond to foreign antigens on the macrophage surface as postulated initially by Lipsky and Rosenthal (3, 4, 10). REFERENCES 1. Siegal, I., J. Allergy 44, 190, 1969. 2. Siegal, I., J. Immunol. 105, 879, 1970.
276
CHONG AND PARISH
3. Lipsky, P. E., and Rosenthal, A. S., .Z Exp. Med. 138, 900, 1973. 4. Lipsky, P. E., and Rosenthal, A. S., J. Exp. Med. 141, 138, 1975. 5. Lopez, L R., Johansen, K. S., Radovich, J., and Talmage, D. W., J. Allergy Clin. Zrnmunol. 53, 336, 1974. 6. Lopez, L. R., Vatter, A. E., and Talmage, D. W., J. Zmmunol. 119, 1668, 1977. 7. Fan, P. T., Yu, D. T. Y., Clements, P. J., and Bluestone, R., Clinical Zmmunol. Zmmunopathol. 9, 363, 1978. 8. Kyewski, B. A., Rouse, R. U., and Kaplan, H. S., Proc. Natl. Acad. Sci. USA 79, 5646, 1982. 9. Wu, S. K., and Thomas, D. W., J. Zmmunol. 131,2110, 1983. 10. Lipsky, P. E., and Rosenthal, A. S., J. Zmmunol. 115, 440, 1975. 11. Davidson, W. F., and Parish, C. R., J Zmmunol. Methods 7,291, 1975. 12. Draber, P., and Kisielow, P., Eur. J Zmmunol. 11, 1, 1981. 13. Parish, C. R., Kirov, S. M., Bowem, N., and Blanden, R. V., Eur. J. Zmmunol. 4, 808, 1974. 14. McKenzie, I. F. C., and Parish, C. R., .Z.Exp. Med. 144, 847, 1976. 15. Denmark, J. R., and Chessurn, B. S., Lancet 1, 161, 1978. 16. Burt, S. M., Carter, T. J. N., and Kricka, L. J., J. Zmmunol. Methods 31, 231, 1979. 17. Hyman, R., and Stallings, V., Immunogenetics 3, 75, 1976. 18. G’Toole, M. M., and Wortis, H. H., J. Zmmunol. 124,2010, 1980. 19. Ralph, P., J. Zmmunol. 110, 1470, 1973. 20. Agrwal, N., and Thomas, D. W., Cell. Zmmunol. &I, 352, 1984. 21. Chong, A. S.-F., and Parish, C. R., Cell. Zmmunol. 92, 277, 1985. 22. Sant, A. J., Schwartz, B. D., and Cullen, S. E., Z. Exp. Med. 158, 1979, 1983. 23. Mosier, D. E., and Pierce, C. W., J. Exp. Med. 136, 1484, 1972. 24. van den Tweel, J. G., and Walker, W. S., Immunology 33, 817, 1977. 25. Beller, D. I., and Unanue, E. R., J. Zmmunol. 121, 1861, 1978. 26. Chen, W. F., Scollay, R., and Shortman, K., J. Zmmunol. 129, 18, 1982. 27. Ceredig, R., Glasebrook, A. L., and MacDonald, H. R., .Z.Exp. Med. 155, 358, 1982. 28. Young, B., Immunology 46, 31, 1981.