Med. Microbiol. Immunol. 165,191--201 (1978)
9 by Springer-Verlag 1978
Cell-Mediated Immunity to Bacterial Flagellin as Assessed by Leucocyte Adherence Inhibition J.G. Aaskov 2 and W.J. Halliday 1 1Department of Microbiology, University of Queensland, St. Lucia, Brisbane, Australia 2Queensland Institute of Medical Research, Herston, 4006, Brisbane, Australia
Abstract. Leucocyte adherence inhibition (LAI) was used to detect cellmediated immunity of mice to Salmonella adelaide polymeric flagellin and its monomeric derivative. In the direct LAI technique, antigen inhibited the in vitro adherence to glass of peritoneal cells (PC) from antigen-primed mice which were capable of exhibiting in vivo delayed hypersensitivity reactions to the same antigen. In the indirect technique, primed PC exposed to antigen in vitro released a soluble factor, which inhibited the adherence of normal PC. Production of the factor was prevented by prior treatment of primed PC with anti-8 serum, indicating the participation of T-lymphocytes. The LAI reaction could be blocked by serum from mice which had been re-injected with antigen 72 h after a priming injection. Features of the production and biological properties of serum blocking activity suggest that it may be attributed to antigen-antibody complexes. Introduction The leucocyte adherence inhibition (LAI) assay of Halliday and Miller (1972) has been improved and modified (Halliday et al., 1974; Maluish and Halliday, 1975; Halliday, 1976) to a stage where it can be used to study a number of aspects of cell-mediated immunity (CMI). One of the problems besetting the study of CMI to tumours and its modulation by serum factors has been the difficulty of obtaining tumour antigen in even a partially purified form. Since Cooper (1972a, b, and c) had described CMI to Salmonella flagellin in the mouse, it was proposed to investigate this system using the LAI test. In this way the nature of CMI and the properties of serum blocking factors were studied in vitro with a defined antigen. Materials and Methods Animals. Inbred CBA mice, 6 - 1 2 weeks old, were used throughout.
0300-8584/78/0165/0191/~ 02.20
192
J.G. Aaskov and W.J. Halliday
Antigens. Polymeric flageUin (POL) from Salmonella adelaide (SW 1338) was kindly provided b y D r . C.R. Parish, Australian National University, Canberra. It was dissolved in saline. The monomeric form of flagellin (MON) was prepared by acid hydrolysis of POL (Cooper, 1972a). Ovalbumin was a product of K & K Laboratories, Hollywood, California.
Immunization. Unless otherwise stated, mice were primed with a subcutaneous (SC) injection into the flank of 25/ag MON or POL in 0.05 ml saline. Sera to be tested for blocking factors were prepared by priming groups of four to five mice and then injecting each mouse intraperitoneally (IP) with antigen 72 h later. Mice were then bled at various intervals following the IP injection. Blood collected in this manner was allowed to cl0t and the serum assayed for blocking activity in an LAI test.
Footpad Assay. The delayed-type hypersensitivity (DTH) reaction to POL was measured using the footpad assay as described by Cooper (1972a). Footpad swelling was calculated by substracting the 24 h swelling after saline injection (right foot) from that after I 0 /ag POL (left foot).
LAI Tecbnique. The direct LAI technique used has been described elsewhere (Halliday et al., 1974). Briefly, mixtures of peritoneal cells (PC; from four to six mice; 106 cells in 0.1 ml of culture medium) and normal mouse serum (0.05 ml) were prepared with antigen (dissolved in 0.05 ml saline) or without antigen (0.05 ml saline), and the adherence of the PC to glass was determined after a preliminary incubation of 30 rain in plastic tubes, followed by 60 min in haemocytometer counting chambers. Blocking activity of serum was assayed by replacing the normal serum by 0.05 ml of the test serum. If the adherence of PC in mixtures with normal serum was inhibited by antigen, but in mixtures with test serum was not inhibited by antigen, then the test serum was recorded as blocking. The indirect LAI technique is a modification of that described by Maluisk and Halliday (1975) and depends on the formation of a soluble mediator by lymphocytes stimulated with antigen. Supernates to be assayed for this leucocyte adherence inhibition factor (LAIF) were recovered by centrifuging the PC-antigen mixtures after the 30 min incubation step. Normal PC from untreated mice were pelleted in a plastic tube, the culture medium was drained off and the supernate to be assayed was used to resuspend the PC. This mixture was incubated for 30 min with occasional shaking and then run into haemocytometers as in the direct LAI technique. A full description of these LAI procedures appears in a recent publication (Halliday, 1976).
Calculation of Results. Cells from each mixture were counted (before and after washing to remove adherent cells) in both chambers of each of two haemocytometers. The count for each chamber was the total of the counts on five squares (0.2 mm x 0.2 ram) in the red cell area. The % adherence for each of the four chambers was calculated as count after washin$ . x 100 count before washing
Leucocyte Adherence Inhibition with Flagellin
193
and the mean % adherence was determined from these four values. Standard error (SE) of the mean was obtained by treating each chamber as a separate determination and the P values were calculated using Student's t-test.
Anti-O Serum. Anti-0 serum was produced in AKR mice injected with CBA thymocytes by the method of Reif and Allen (1966). This serum preparation (in the presence of guinea pig complement) killed 100% of CBA thymocytes, 14- 17% of CBA PC and 37 49% of CBA spleen cells, as judged by trypan blue exclusion. Fractionation o f Serum on a Graded Porosity Filter. Serum was diluted 20 times in saline and added to the chamber of an Amicon A-75 filter (Amicon Corp., Lexington, Mass.) which retains molecules larger than 75,000 daltons. When the volume had been reduced to one-twentieth (to the same volume as serum before dilution), the fluid remaining in the chamber was recovered. This was repeated four times with each serum treated. Treatment o f Mouse Serum with Rabbit-Antimouse Antiserum. Rabbit-antimouse IgG antiserum was prepared by immunizing rabbits with three injections of mouse IgG (prepared by chromatography on DEAE cellulose) given at two-week intervals in Freund's complete adjuvant. Normal mouse serum or blocking serum (0.4 ml) was mixed with normal rabbit serum or rabbit-antimouse lgG antiserum (0.1 ml) and stored overnight at 4oc. Serum mixtures were then centrifuged at 1500 g and the supernates recovered. The effect of the supernates on direct LAI was tested by adding 0.05 ml to 0.1 ml of primed PC in the presence or absence of antigen. Results
Characteristics o f the Response to Flagellin, as Detected by Skin Test in vivo and LAI in vitro PC recovered from mice three to six days after a single SC injection of 25 gtg of POL were reactive with the antigen in vitro, as demonstrated by direct LAI (Table 1). In some experiments, reactivity could be detected as early as 48 h after injection. Priming doses of 10 and 50 pg of POL gave similar results. It is important to note that the normal level of adherence of PC from different mice varied over a wide range; comparisons should thus be made only between treatments applied to the same cell population (as in Table 1). Mice also gave a positive footpad DTH reaction three days after a priming injection, indicating in vivo reactivity to the same antigen at the same time. Groups of five mice, primed SC with 25 #g of POL and re-injected into the footpad three days later with 10 gtg, exhibited 24 h swellings of 0.20 -+ 0.06 mm (mean -+ SD); controls injected into the footpad without priming gave swellings of 0.09 -+ 0.03 mm (P < 0.01). The in vitro reactivity of MON- or POL-primed PC, as related to the dose of antigen used in the LAI reaction, is shown in Fig 1. The adherence of PC from unprimed animals was not affected by the range of antigen concentrations tested. POL-primed PC failed to respond to an unrelated antigen (ovalbumin). MON or POL at a final concentration of 125 gtg/ml were used in all subsequent LAI experiments.
194
J.G. A a s k o v a n d W.J. Halliday
Table 1. Time course of development of CMI to POL as detected in vitro with the direct LAI test Time after injection a (days)
Controlb
Test c
0 1 2 3
64 75 64 75
4 2 6 2
61 73 58 58
4
81 •
2
70 • 5
5
90 •
6
74 + 3
6
72 + 10
50 • 4
Mean % leucocyte adherence (• SE) Result
+ + +•
+5 • 7 +6 • 4
No inhibition No inhibition No inhibition Inhibition (P < o.o01) Inhibition (P < 0.005) Inhibition (P < o.oo5) Inhibition (e < 0.005)
a25/zg POL SC in 0.05 ml saline bcontrol mixtures contained PC harvested at times indicated, normal mouse serum, and saline CTest mixtures contained, in addition, 125/~g/ml POL
90
~o c
so 70 0
~
~..,0
0
~
6o
"
50
E 40
!
I
I
I
0
25
50
75
I 100
! 125
Fig. 1. Effect of antigen dose on direct LAI activity of PC from normal and primed mice (three days after 25 gg of MON or POL SC). o - - o : Normal PC challenged with POle., * - - * : POLprimed PC challenged with POL, o - - Q : POL-primed PC challenged with ovalbumin, LX---~: Normal PC challenged with MON, " - - ' ; MON-primed PC challenged with MON
Antigen Challenge Dose(,LIg/ml)
Effect of A n t i ~ Serum on L A I F Production T r e a t m e n t o f P O L - p r i m e d PC w i t h anti-0 Serum a n d c o m p l e m e n t d e s t r o y e d t h e i r capacity t o p r o d u c e L A I F (Table 2). Similar t r e a t m e n t o f p r i m e d PC w i t h n o r m a l A K R m o u s e s e r u m a n d c o m p l e m e n t had n o e f f e c t .
L e u c o c y t e A d h e r e n c e I n h i b i t i o n w i t h FlageUin
195
Table 2. Effect of treatment of mouse PC with anti-0 serum and complement on LAIF production (indirect LAI test) Type of PC
Treatment of PC
Mean % leucocyte adherence (+ SE) Control a
Test b
Result
Normal
Normal mouse serum + complement
65 + 3
63 9 8
No inhibition (supernate inactive)
Normal
Anti-0 serum + complement
73 • 1
67 • 7
No inhibition (supernate inactive)
primed c
Normal mouse serum + .complement
76 + 3
51 + 6
Inhibition P < .005 (supernate active)
Primed
Anti-0 serum + complement
73 • 2
69 + 3
No inhibition (supernate inactive)
aControl mixtures contained fresh normal PC and supernate (from original PC treated with serum indicated and incubated without antigen) bTest mixtures contained fresh normal PC and supernate (from original PC treated with serum indicated and incubated with 125 tag/m1POL) CMice were injected SC with 25 tag POL three days before harvesting PC
601
"0
>.
40
D
cO
20
t
-~ o Z
I
0
I
24
I
48
I
72
Time of Serum Collection (hours)
Fig. 2. Kinetics of production of serum blocking factors following challenge with 10 ~tg MON IP, into mice primed 72 h before with 10 gg MON SC. Serum was pooled from six mice. Open circles: MON-primed PC cultured with serum but without antigen. Closed circles: MON-primed PC cultured with serum and antigen
196
J.G. Aaskov and W.J. Halliday
Blocking Factor Production MON- or POL-primed mice produced circulating inhibitors of LAI (blocking factors) if challenged with an 1P injection of the priming antigen three days after priming. An example of this is shown in Figure 2. Addition of MON to MON-primed PC, in the presence of normal serum (or serum obtained at time of challenge), promoted LAI. However, serum taken 8 h and 24 h after IP challenge contained material which blocked this LAI reaction. Similar observations were made in mice primed with 1 0 - 2 5 / a g POL SC and challenged with 1 0 - 2 5 / a g POL IP 72 h later; blocking factor was then detected only 48 h after challenge. In all cases the presence of the blocking factor was only transient.
Nature of Blocking Factors I t was n o t possible to block the LAI reaction in vitro by pre-incubating primed PC for 30 rain with smaU quantities of antigen, before testing the ceils in the normal manner with antigen in a direct LAI test. Amounts of MON or POL ranging between 5 x 10 .5/ag and 5 x 10 -1/ag were mixed with the PC but had no effect on their subsequent adherence inhibition when challenged with 125/lg/ml and MON or POL. Filtration of blocking serum from MON-primed and challenged mice through an Amicon A-75 filter did n o t reduce the in vitro blocking activity of the retained components (Table 3). Molecules less than 75,000 daltons are removed by this treatment (MON is 40,000 daltons). Passage.of normal mouse serum through a similar filter did not produce a serum which interfered with an LAI test. Blocking activity in serum could be removed by pre-treating the serum with antiserum to mouse IgG. Results and appropriate controls are given in Table 4. Antigen specificity of serum blocking was demonstrated by showing that LAI of primed PC (with MON in the presence of normal mouse serum) was blocked by related Table 3. Leucocyte reactivity (direct LAI test) in the presence of serum freed Qf substances with molecular weight less than 75,000
Serum
Normal
Treatment
Mean % leucocyte adherence (• SE)
Result
Control a
Test b
-
88 + 6
49 • 4
Leucocyte active (e < o.o01)
Blockingc
-
64 • 6
64 •
3
Reaction blocked
Normal
Amicon filtrationd
70 + 3
61 •
3
Leucocytes active (P < 0.005)
Blocking
Amicon filtration
71 • 4
73 • 13
Reaction blocked
aControl mixtures contained primed PC (three days after priming injection of MON), serum as indicated, and saline bTest mixtures contained, in addition, 125 gg/ml MON CBlockingserum from mice primed and challenged in vivo with MON dsee Materials and Methods
Leucocyte Adherence Inhibition with Flagellin
197
Table 4. Removal of blocking activity from mouse serum by treatment with rabbit-antimouse IgG antiserum Mouse Serum
Treatment
Mean % leucocyte adherence (-+ SE) Control a
Test b
Result
Normal
--
73 :~ 2
51 +- 3
Inhibition (P < 0.001)
Blocking c
-
71 • 6
75 • 2
No inhibition; serum blocking
Normal
Normal rabbit serum
71 ~ 8
53 • 4
Inhibition (P < 0.005)
Normal
Rabbit-antimouse lgG
60 + 3
34 + 5
Inhibition (P < 0.001)
Blocking
Normal rabbit serum
54 +-8
52 + 4
No inhibition; serum blocking
Blocking
Rabbit-antimouse IgG
69 9 8
35 • 5
Inhibition (P < 0.001); blocking abrogated
aControl mixtures contained primed PC, serum as indicated, and saline bTest mixtures contained, in addition, 125/~g/ml POL CBloeking serum from mice primed and challenged in vivo with POL
blocking serum as above, but not by sera which were effective in blocking CMI to other antigens. These latter sera were from mice bearing syngeneic tumours (Halliday et al., 1974) and from mice made tolerant to dinitrochlorobenzene (Halliday and Noonan, 1978).
Other Factors Affecting LAI When primed PC from POL-injected mice were mixed with various proportions of normal PC, LAI activity in the direct test diluted out when the cell ratios were 1 : 1 (Figure 3a). The same primed PC were used to prepare an active supernate by treatment with antigen, and this supernate was diluted and tested for LAI activity by the indirect method. LAIF activity in undiluted supernate was significantly greater than in the control (P < 0.02), but when diluted 3:1, 1:1, or 1:3, no significant L A I F activity could be detected (P > 0.05). Over a sixfold range of cell concentrations the number of ceils which failed to adhere to the hemocytometers was proportional to the total number o f cells used (Fig. 4). This linear relationship was found for normal PC incubated with both test and control culture supernates, although, as expected, test supernates (containing LAIF) gave lower adherences. Figure 4 shows the result of a single experiment in which supernates prepared from the one batch of primed PC were assayed using a pool of normal PC. Identical cell numbers were not used to assay test and control supernates, but as many points as possible were obtained with the culture supernate available.
198
J.G. Aaskov and W.J. Halliday (a)
100
90
80
|
7O
60
I Primed
! 3:1
I 1:1
11:3
Fig. 3. a A d h e r e n c e o f POL-primed PC diluted with normal PC, as measured b y direct LAI a--e: PC cultured with 125 ~ug/ml POL o - - o : PC cultured in absence o f POL. b Adherence o f normal PC in an indirect LAI test: LAIF activity in culture superhates from POL-primed PC cultured with 125/zg/ml POL, diluted in supernate f r o m primed PC cultured in absence of POL (e--e). Open circle: adherence in supernate f r o m primed PC cultured in absence o f antigen. Bars show SE o f determinations
Norlmal
Dilution of Primed PC with Normal PC
(b)
.,~ 1001
t
80
70
6o
L
Undiluted
| 3~1
I 1:1
I 1:3
I
Control
Dilution of Test Supernate
E g
E
~
5
o
4
3
9
U
0
0
9
2
4 Initial
6
8
iO
count (millions/ml)
Fig. 4. Relationship b e t w e e n total n u m b e r of PC used for the indirect LAI test and the n u m b e r o f PC failing to adhere, o o: Normal PC in s u p e m a t e from POL-primed PC cultured in absence o f antigen. 9 e: Normal PC in supernate f r o m POL-primed PC cultured in presence of POL. Lines calculated b y m e t h o d o f least squares
Discussion
CMI to the defined bacterial antigen flagellin has been demonstrated using the LAI technique. Activity could be detected in primed PC cultured in the presence of flagellin (but not in primed PC cultured without antigen, primed PC cultured with unrelated antigen, or in normal PC cultured with antigen). In confirmation of previous
Leucocyte Adherence Inhibition with Flagellin
199
work with other antigens (Maluish and Halliday, 1975; Holt et al., 1975), LAIF activity was found in the culture supernates from primed PC challenged in vitro with the antigen (but not in the culture supernates from primed PC cultured in the absence of antigen or in the culture supernates from normal PC cultured with antigen). LAI reactivity was related to the number of primed ceils challenged in both direct and indirect tests (Fig. 3). Hold~ et al. (1974) were unable to detect LAIF activity in culture supernates from immune lymph node cells challenged with antigen in vitro, although cellular reactivity appeared in a direct LAI test. The same authors presented data on cell mixtures, analogous to the results in Figure 3a of this paper, and interpreted the findings as evidence for absence of a soluble mediator. However, as we have shown, supernates do contain LAIF but this factor is very susceptible to dilution. The factor may be produced in larger amounts in some other systems, for example from leucocytes sensitized to tumour antigens, where it was found to be active even when considerably diluted (Maluish and Halliday, 1975). As described also by Cooper (1972a), POL-primed mice gave positive footpad DTH reactions to challenge three days after the priming injection. LAI-reactive cells could be recovered from the peritoneal cavity at this time. Treatment of such cells with anti0 serum and complement abolished reactivity, demonstrating that T-lymphocytes are required at some stage in the production of LAIF; this confirms the observations of Holt et al. (1975) with other antigens. Attempts to investigate cellular interactions in the production of LAIF, by challenging decreasing numbers of primed cells with a constant concentration of antigen, did not provide any conclusive results (Fig. 3). Activity of cells decreased too rapidly for the slope of the curve to be determined. It was of interest to note that, over a wide range of cell concentrations, the number of cells which failed to adhere to hemocytometers bore a linear relationship to the number of cells employed in the test (Fig. 4). This was equally true in the presence or absence of LAIF, and would suggest that, provided the number o f indicator cells employed in an indirect LAI test fell between about 1 x 106 and 6 x 106 per ml, similar values for % LAI should be obtained. These data are also compatible with the concept of varying numbers of LAIF receptors on different cells or populations of cells. That the lines obtained in Figure 4 do not pass through the origin, and also that the slope of the control line is not zero, indicate the existence of several cell populations; some fail to adhere even in the absence of LAIF and some remain adherent irrespective of the concentration of LAIF. It could be postulated that ceils apparently resistant to LAIF carry few or no receptors for it. Using either POL or MON and appropriately primed mice, it was possible to generate serum factors which blocked the in vitro reaction of primed cells with antigen. This serum blocking activity was a transient phenomenon, appearing as early as 8 h after an IP challenge with MON, or 48 h after an IP challenge with POL, and being detectable for less than 24 h. No attempt was made to skin test animals during the short time they carried blocking factors. Paranjpe and Boone (1975) found that IP injection of a tumour homogenate into tumour-immune mice (but not into normal mice) elicited phenomena analogous to those observed here: specific serum blocking factors appeared and were detected by their effect on CMI in vitro. The transient nature of the blocking activity is in keeping with data from tumour studies, which show the prompt disappearance of
200
J.G. Aaskov and W.J. Halliday
similar activity once the tumour (the source of antigen) is removed or regresses (Hellstr6r and Hellstr6m, 1970; Halliday et al., 1974; Prather and Lausch, 1976). LAI reactivity could not be blocked by pre-incubating primed PC in a range of substimulatory doses of antigen, and it was not possible to remove blocking activity from serum by filtering from it substances with a molecular weight similar to that of free antigen (MON, MW 40,000). Blocking activity could be abolished by treating the mouse serum with antimouse IgG antiserum. These observations, along with rapid appearance, short in vivo life, and the need for injecting previously primed mice with antigen, all support the concept that the serum blocking factors are not free antigen but antigenantibody complexes. There is thus a further similarity to reactions of tumour immunity (Sj6gren et al., 1971; Baldwin et al., 1972). With the LAI technique, we detected an in vitro correlate of CMI against a defined antigen in experimental animals. In addition, serum blocking factors were readily produced. By analogy with previous work on tumours (Maluish and Halliday, 1975), this investigation could be extended to follow the kinetics of development of CMI to defined antigens in individual animals. Acknowledgements. This work was supported by the National Health and Medical Research Council. The authors thank Mrs. Anne Visona and Miss Barbara Morris for their assistance, and Dr. C.R. Parish for generous gifts of POL. References
Baldwin, R.W., Price, M.R., Robins, R.A.: Blocking of lymphocyte-mediated cytotoxicity for rat hepatoma cells by tumour-specific antigen-antibody complexes. Nature (New Biol.) 238,185--186 (1972) Cooper, M.G.: Delayed hypersensitivity in the mouse. I. Induction and elicitation by Salmonella adelaide flagellin and its derivatives. Scand. J. Immunol.. 1 , 1 6 7 178 (1972a) Cooper, M.G.: Delayed-type hypersensitivity in the mouse. II. Transfer by the thymusderived (T) cells. Scand. J. Immunol. 1 , 2 3 7 - 2 4 5 (1972b) Cooper, M.G.: Delayed-type hypersensitivity in the mouse. III. Inactivation of the thymus-derived effector cells and their precursors. Scand. J. Immunol. 1 , 2 4 7 253 (1972c) Halliday, W.J.: Leukocyte-adherence inhibition test and blocking factors in cancer. In Vitro Methods in Cell-Mediated and Tumor Immunity. B.R. Bloom and J.S. David (eds.) New York: Academic Press 1976 Halliday, w.J., Maluish, A., Miller, S.: Blocking and unblocking of cell-mediated antitumor immunity in mice, as detected by the leukocyte adherence inhibition test. Cell. Immunol. 1 0 , 4 6 7 - 4 7 5 (1974) Halliday, W.J., Miller, S.: Leukocyte adherence inhibition: a simple test for cellmediated tumour immunity and serum blocking factors. Int. J. Cancer 9 , 4 7 7 483 (1972) Halliday, W.J., Noonan, F.P.: Studies of contact hypersensitivity and tolerance in vivo and in vitro. II. Regulation by serum factors. Int. Archs. Allergy Appl. Immunol. 56, 533-542 (1978)
Leucocyte Adherence Inhibition with Flagellin
201
Hellstr6m, I., Hellstr6m, K.E.: Colony inhibition studies on blocking and non-blocking serum effects on cellular immunity to Moloney sarcomas. Int. J. Cancer 5, 195-201 (1970) Hol~, V., Ha~ek, M., Buhenik, J., Chutn~i, J.: Antigen-mediated macrophage adherence inhibition. Cell. lmmunol. 13,107-116 (1974) Holt, P.G., Roberts, L.M., Fimmel, P.J., Keast, D.: The LAI microtest: a rapid and sensitive procedure for the demonstration of cell-mediated immunity in vitro. J. lmmunol. Methods 8,277--288 (1975) Maluish, A.E., Halliday, W.J.." Quantitation of anti-tumor cell-mediated immunity by a lymphokine-dependent reaction using small volumes of blood. Cell. Immunol. 17,131-140 (1975) Paranjpe, M.S., Boone, C.W.: Specific depression of the anti-tumor cellular immune response with autologous tumor homogenate. Cancer Res. 35, 1205-1209 (1975) Prather, S.O., Lauseh, R.N.: Kinetics of serum factors mediating blocking, unblocking and antibody-dependent cellular cytotoxicity in hamsters given isografts of para7 tumor cells. Int. J. Cancer 17,380-388 (1976) Reif, A.E., Allen, J.M.: Mouse thymic iso-antigens. Nature (Lond.) 209,523 (1966) Sj6gren, H.O., Hellstr6m, I., Bansal, S.C., Hellstr6m, K.E.: Suggestive evidence that the "blocking antibodies" of tumor-bearing individuals may be antigen-antibody complexes. Proc. Natl. Acad. Sci. USA 68, 1372-1375 (1971) Receioed May 2, 1978
