Characteristics of cellular immune responses to collagen type I or collagen type II

CELLULAR

IMMUNOLOGY

la,314-330

(1986)

Characteristics of Cellular Immune Responses Collagen Type I or Collagen Type II

to

L. BUTLER, B. SIMMONS, J. ZIMMERMANN, P. DERISO, AND K. PHADKE Departments ofImmunology Research, Connective Tissue and Pulmonary Research, and Pathology< Lilly Research Laboratories, Indianapolis, Indiana 4628s Received October 7, 1985; acceptedMarch 2, I986 We have examined the murine cell-mediated immune (CMI) responseto collagens type I (CI) and type II (CII) as measured by in vivo delayed-type hypersensitivity responses.We have verified the histopathology and kinetics ofthe cell-mediated immune responses.Predominant cell-mediated responseswere obtained 7, 10,or 14 days following immunization. A presumed antibody-mediated reaction was observed at later times (e.g., >2 1 days following immunization). The CM1 responses to the collagens show a strain-dependent relationship. For CI, the CM1 responseprofile shows H2b > H-2’ = H-2q b H-2d. For bovine CII, the response profile is H-2d r H-2b = H-2’ = H-2‘$ the chick CR responseprofile is H-2q = H-2’ > H-Zb = H-2d, and in limited testing, only the H2q strain could generate murine CII-specific cell-mediated immune responses.The CII-specific CM1 responseis cross-reactivewith CII from several speciesof animals, but not with CI. Further, the collagen-specific CM1 response can be elicited with certain cyanogen-bromide fragments of bovine CII. Finally, our study also demonstrates that there is a non-H-2-linked locus(i) involved in the development of CII-induced arthritis. Q 1986 Academic dress,IK

INTRODUCTION In an effort to better understand the possible consequences of immune responses to connective tissue antigens, we have been studying murine cell-mediated immune responsesto rat collagen type I (R-CI),’ bovine collagen type II (B-CII), chick collagen type II (C-CII), or murine collagen type II (M-CII). We have extensively studied the CM1 response to collagen type II, as immunization with this antigen leads to the development of arthritis in selectedstrains of mice ( l-3). Although other investigators have examined the cell-mediated immune (CMI) response to B-CR in vitro (4) or in relationship to arthritic diseasesin rats and mice (3, 5-lo), these investigations have primarily focused on assayingthe CM1 responseat a time point during which arthritic changesare visible and have not studied the development, regulation, or characteristics of the cell-mediated response itself. It is to this extent that we have undertaken this study to investigate the cell-mediated immune response to these collagens. We have investigated strain relationships and routes of immunization, histopathologic changes following the elicitation of cell-mediated immune responses,kinetics of CM1 devel’ Abbreviations used: B-CII, bovine collagen type II; CI, collagen type I; CII, collagen type II; C-CII, chicken collagen type II; CMI, cell-mediated immunity; (IN-BCII, denatured bovine collagen type II; MCII, murine collagen type II; R-CI, rat collagen type I. 314 0008-8749/86 $3.00 Copyright 0 1986 by Academic Press,Inc. All rights of repmduction in any form reserved.

MURINE

COLLAGEN-SPECIFIC

CM1 RESPONSES

315

opment, cross-reactivity between various forms of collagen at the level of the CM1 response,and finally, the relationship between the capacity to respond to the collagens and the development of arthritis. MATERIALS

AND METHODS

Mice. The following strains were purchased from Jackson Laboratory (Bar Harbor Maine): DBA/l J, CBA/J, AKRIJ-nti TR, BlO.BR SgSn, C57Bl/lO, BlO.D2oSnJ, BlO.D2nSnJ, C57B1/6-bg, and SWR/J. C57/B16 and BALB/c mice were purchased from Charles River Laboratories (Portage, Mich.). CBA/Ca mice aged for at least 2 years were purchased through Charles River Laboratories. MRL/mplpr mice were purchased from Harlan Laboratories (Indianapolis, Ind.). All mice were maintained in our animal facilities in isolator cagesand given food and water ad Zibitum. Collagens.Rat collagen type I and bovine collagen type II were prepared aspreviously described (11, 12). Mouse collagen type II (M-CII), Lot 03402, and chick collagen type II (C-CII), Lot 10005, were purchased from Genzyme (Boston, Mass.). The prepared collagens were kept frozen prior to use. Some preparations were stored for a short period of time at 4°C. Collagen preparations used for sensitization were first diluted to 2 mg/ml in phosphate-buffered saline, pH 7.4. Cyanogen-bromide fragments of the B-C11molecule were prepared according to published procedures ( 13). Sensitization and elicitation of delayed-type hypersensitivity. For sensitization, mice were immunized with either a single or a double injection of collagen in the base of the tail. For the single injection, mice received 100 hg of R-C1 or B-C11emulsified in complete Freund’s adjuvant [(H37RA), Difco Laboratories, Detroit, Mich.)], in a total volume of 100 ~1on Day 0. For double immunization, mice received a secondinjection of the same material on Day 7. Control animals received adjuvant emulsified with phosphate-buffered saline. At various times following the immunization, mice were then challenged with the antigen solutions. For the challenge, 10 pg of the appropriate antigen was injected in a lo-p1 volume into the pinna of the ear or 30 pg of antigen was injected into the hind footpad. The antigens used for the challenge were dialyzed overnight against phosphate-buffered saline at pH 7.4 and kept on ice until use to prevent aggregation unless otherwise specified. Prior to injection, the ear thickness or footpad thickness of the animals was measured with a Mititoyu micrometer. Four or twenty-four hours following injection, the increase in ear thickness or footpad thickness was monitored. The expression of ear swelling is represented as the difference between the unchallenged ear thickness and the ear thickness monitored after 4 or 24 hr. Assessmentof arthritis. Animals were visually inspected on a weekly basis following immunization for signs of soft tissue swelling. Animals were scored with an arthritic index scale of 1 to 4; 1 indicating only one paw was involved, 4 indicating all paws were involved. Degreesof involvement on each paw were not recorded as the incidence of single digit swelling without later development into greater involvement was rare. The mean arthritic index was calculated by dividing the sum total arthritic index by the number of arthritic mice. The mean day of onset was calculated by dividing the sum total of the initial day of onset by the number of arthritic mice. Preparations of histologic sections. Respective sections from the ear or footpad were prepared after being fixed in 10%phosphate-buffered Formalin, dehydrated, embedded in paraplast, cut at 5 PM, and stained with hematoxylin and eosin (HE). Radiologic examination. Animals displaying obvious and persistent footpad swelling

316

BUTLER ET AL.

were randomly subjected to radiologic evaluation of bone involvement. These studies were conducted by Mr. Donald Benslay, Lilly Research Laboratories, Indianapolis, Indiana, and demonstrated that in those casesof observed swelling induced by collagen type II which were tested, evidence of bone damage was also present. Statistics.Statistical significance betweendifferencesof meansof experimental groups was determined through the use of Student’s t test. RESULTS

Ability to Elicit Collagen-Spec$cDelayed-typeHypersensitivity Responseby Using the Footpad or the Ear as the Site of Challenge We wished to examine whether we could utilize the footpads or ears as sites for elicitation of murine collagen-specific DTH responses.We chose to use two strains, one susceptible to arthritis induction (DBA/ 1J) and one resistant to arthritis induction (CBA/J). CBA/J or DBA/l J mice were immunized at the base of the tail with 100 pug of B-C11or R-C1 in complete Freund’s adjuvant and restedfor 7 to 14 days. Subsequent to the rest period, these animals were challenged with 1Opgof dialyzed collagen in the ear or in some caseswith 30 pg in the footpad. We chose to use the footpad because of its location in relation to joint tissue affected by arthritis and thus the site might serve as a suitable model for the study of immunoregulatory networks involved in the elicitation and/or maintenance of the disease. The antigen doses and routes of immunization were chosen following preliminary dose-responseexperiments. As in other systems(14, 15), it appears that subcutaneous or base of the tail injections but not ip injections are the best routes for inducing CM1 responses to collagens. At various times following the challenge, the appropriate tissue swelling was assessed.As is evident in Table 1, both strains were capable of generating collagen-specific cell-mediated immune responses.Although the data shown were obtained following a single immunization on Day 0, similar results were obtained with a double immunization on Days 0 and 7. Unless otherwise noted, only single immunizations were usedthroughout the remainder of the study.

Histopathology of Lesionsfollowing Elicitation of Cell-Mediated Immune Response We next studied the immunopathologic changes occurring in the ear at the site of challenge at various times following immunization with collagens at the base of the tail in an attempt to correlate these histopathology changeswith the ear swelling data that were obtained. The series of Figs. 1A through 1D demonstrate that subsequent to immunization at the base of the tail, the immunopathologic changes in the ear challenge site vary, depending upon time of challenge. This is also a coordinate in the patterns of swelling observed following ear challenge (Table 2). At Day 7, minimal 4hr ear swelling is detected and a peak ear swelling is seenat 24 hr. The cell types seen in the ear challenge lesion are primarily mononuclear and polymorphonuclear cell infiltrates, At Day 10, a definite 4-hr ear swelling is observed although maximum ear swelling is still seen at 24 hr. Again, the histopathology of the 24-hr lesion still shows a mononuclear cell and PMN involvement. The 14-day response shows an increased 4-hr ear swelling with a predominant peak at 24 hr, and once again histopathology shows a mononuclear cell and PMN infiltrate. In contrast by Day 2 1 and up to Day 60 or 90, there is a large 4-hr ear swelling responsewith a substantially decreased24-

317

MURINE COLLAGEN-SPECIFIC CM1 RESPONSES TABLE I

Collagen-Specific Delayed-Type Hypersensitivity ResponsesCan Be Elicited in the Ear or in the Footpad Antigen used for 24-hr ear swelling (X 10F4in. + SE)

Sensitization

Elicitation

Site of elicitation

DBA/J

None CII CI None CII CI

CII CII CII CI CI CI

E E E E E E

7.0 47.6 12.6 12.5 11.4 33.3

+ + f 2 + +

CBA/J

None CII CI None CII CI

CII CII CII CI CI CI

E E E E E E

20.2 76.4 17.5 13.9 22.9 38.8

It 1.1 + 2.4 zk0.9 + 0.6 + 1.1 + 2.7

Strain of resoonder

2.1 6.2 1.6 1.7 1.6 4.0

24-hr footpad swelling (mm + SE) CBA/J

None CII

CII CII

F F

0.035 + 0.01 1.03 + 0.08

Note. A representative experiment with five mice per group is shown. Mice were immunized at the base of the tail with 100 fig of the appropriate antigen emulsified in complete Freund’s adjuvant and rested for 14 days. The mice were then challenged with 10 pg of the appropriate antigen in the dorsum of the ear (E) or 30 pg of the appropriate antigen in the footpad (F). Differential increasesin ear or footpad swelling were determined 24 hr later.

hr ear swelling response.Greatly reduced numbers of mononuclear or polymophonuclear cells are seen in the lesion at 24 hr, a profile observed in antibody-mediated reactions (21). Thus, depending upon the time and the day of elicitation following immunization, various degrees of cell-mediated related immune responses can be observed. It is for this reason that we have chosen Day 7 and/or Day 14 to investigate the regulation of cell-mediated immune responsesin the collagen systems. Kinetics of CII Spec$c Ear Swelling Responses In an effort to determine whether the cellular immune responseprofiles for various strains were similar, we studied the kinetics of the ear swelling response following immunization with CII. These results (Fig. 2) demonstrate differences in the relative strengths of the ear swelling responses,as expected, but a similar overall pattern. The patterns indicate a curious bimodal response pattern with a decreased ear swelling response seen in the 21- to 2S-day time period. However, as noted previously, the CM1 responseitself is greatly complicated by an increasing antibody component subsequent to Day 14 after immunization. Thus, we have been careful to identify these data (post Day 2 1) as ear swelling and not specifically CM1 reactions.

FIG. 1. DBA/lJ mice were immunized with 100 pg of B-C11emulsified in complete Freund’s adjuvant and rested. The animals were then challenged with 10 pg of B-C11in the dorsum of the ear and increasesin ear swelling were determined 24 hr later. Subsequent to this the ears were excised and saved for histology examination. (A), Challenge at Day 7, (B), challenge at Day 14, (C), challenge at Day 90, (D), challenge site in a nonimmunized animal.

318

MURINE COLLAGEN-SPECIFIC CM1 RESPONSES

FIG. I-Continued.

319

320

BUTLER ET AL. TABLE 2 Kinetics of Ear Swelling ResponseElicited at Different Times following Sensitization 24-hr ear swelling response X low4in. + SE

Day of challenge of DBA/l J mice following immunization with CII in CFA on Day 0

4 hr

24 hr

Day I Day 14 Day 90

13.0 f 3.1 38.1 f 3.1 142.2 + 4.1

30.2 + 2.1 48.1 + 2.6 61.2 k 3.3

Note. Mice were immunized with 100 Fg of B-C11emulsified in complete Freund’s adjuvant and rested for the appropriate length of time. The mice were challenged with 10 pg of B-C11in the dorsum of the ear and increasesin ear swelling were determined 24 hr later. Negative control values have been subtracted. The negative control values ranged from 15 to 25 X 10e4in. The results shown are from at least three experiments with five mice per group.

Strain Dependency of CM Responsesto CI and CII The results in Fig. 3 show a strain-dependent response pattern at Day 7 or 14 for CM1 responsesto R-CI. In this response pattern C57Bl/6, CBA/J, and DBA/l J mice are good responders, while BALB/c mice are poor responders. This H-Zrelated trend is further substantiated by responsesof H-2 congenic mice where BlO.BR (H-2k) and B10 (H-2b) mice are good responders and B 10.D2 (H-2d) mice are poor responders. The differences in the CI response patterns of Balb/c and BlO.D2 mice compared to their appropriate groups are statistically significant. As expected, AKR/J nuSTR/n$TR (athymic) mice did not show a cell-mediated immune response to R-CI. In Fig. 4, immune responses to B-C11at Day 7 or 14 are shown. In this case, BALB/c mice showed a good response; C57B1/6, CBA/J, DBA/l J, and SWR/J mice showed lower responses. The SWR/J strain was the poorest responder of this latter group. This strain-related responseis further demonstrated in the H-2 congenic mice where BlO.D2 (H-2d) mice generated good responsesand B10 (H-2b) and BlO.BR (H-2k) mice gave

89B $g40

CBA/J

_

BALBlC

C57B118

_

DBAIIJ

60-

ssm,d 0:

7 14 21 20 35 42 49 7 14 21 28 35 42 49 Days After lmmunlutlon

FIG. 2. Kinetics and strain dependence of ear swelling response following immunization with collagen type II. Mice were immunized on Day 0 with 100 pg of Bovine CII in CFA. At the times noted, selected cohorts were ear challenged and 24-hr response was measured. Mice immunized with CFA were used as controls. Responseshown is increment above CFA immunized control readings.

MURINE COLLAGEN-SPECIFIC CM1 RESPONSES

Strain

H-2

10

CBAIJ NUlNU DBAH BALBIC C57BlM

K K Q D B

C57lBLiO

B

BlO.BR BlO.DZ

K D

30

Day 7 50 70 90 10

30

321

Day 14 50 70 90

FIG. 3. Strain dependence of DTH responsesto collagen type 1; 24-hr ear swelling (mean + SE) X lob4 in. Mice were immunized with 100 ng of rat collagen type. I emulsified in complete Freund’s adjuvant and rested. The mice were then challenged with 10 pg of the appropriate antigen in the dorsum of the ear. Increase in ear swelling was determined 24 hr later. The responsesshown are the difference in the challenged ear swelling responsesof immunized mice minus the responsesof naive mice.

medium responses.Similar responseswere obtained with cyclophosphamide-pretreated responder mice (data not shown). The response patterns of the Balb/c and BlO.D2 mice differ significantly from their appropriate groups. Again, at Day 7 as well as Day 14 athymic AKR nuSTR/nnsTR mice did not demonstrate the capability to generate a B-CII-specific CM1 response. A different response pattern was obtained when mice were immunized with Chick-C11 or Murine-CII (Tables 3, 4). For Chick-CII, CBA/J, SWR/J and DBA/ 1J mice were good responders while C57B l/6 and Balb/c mice were poorer responders (Table 4). These results were verified using H-2 congenic mice and cyclophosphamide-pretreated responder mice (data not shown). Following immunization with Murine CII, DBA/ 1 mice responded with detectable CM1 reactions while CBA/J and Balb/c mice did not (Table 4).

CM Responsesof SelectedImmunodejicient Strains We further investigated the capacity of several reported immune-deficient strains (MRL/Zpr, aged CBA/Ca and C57/B 16 bgfbg and B 10.D20) to generate cell-mediated responsesto collagen type II. We chose the MRL/lpr strain becausethere have been Day 14

Day 7 Strain CBA/J NUfNU DBAll SWRlJ BALBlC C578116

10 “-*

34 50 70 SO 10

30

50 70 90

r

C57B1110 BlO.BR B10.02

FIG. 4. Strain dependence of DTH responseto bovine collagen type II; 24-hr ear swelling (mean t SE) X 10m4in. Mice were immunized with 100 M of bovine collagen type II emulsified in complete Freund’s adjuvant and rested. The mice were then challenged with 10 rg of the appropriate antigen in the dorsum of the ear. Increase in ear swelling was determined 24 hr later. The responsesshown are the difference in the challenged ear swelling responsesof immunized mice minus the responsesof naive mice.

322

BUTLER ET AL. TABLE 3 Strain-Dependent Ability to Elicit DTH Reactions following Immunization and Challenge with Heterologous Collagens Day 14” 24-hr ear swelling response X lo-“ in. (*SE) following challenge with Strain

H-2

Collagen used’ in immunization

DBA/lJ’

q

None B-C11 c-c11

SWR/JC

q

CBA/J

k

Balb/c

d

C57Bl/6

b

None B-C11 c-c11 None B-CII c-c11 None B-C11 c-c11 None B-C11 c-c11

B-C11

c-c11

9.5 (1.0) 50.8 (6.1) 19.9 (2.0) 29.5 (1.2) 77.7 (5.6) 56.9 (1.7) 7.5 (0.54) 57.3 (3.1) 64.9 (1.3) 14.6 (0.6) 84.5 (6.0) 22.7 (0.6) 13.2 (0.7) 51.5 (1.3) 21.4 (1.0)

6.3 (1.6) 33.5 (5.8) 33.3 (3.3) 27.7 (1.4) 84.8 (4.5) 70.2 (2.8) 7.2 (0.70) 60.3 (4.5) 49.8 (1.5) 18.1 (1.0) 34.8 (1.0) 26.7 (1.1) 9.8 (1.2) 22.3 (0.5) 19.4 (0.7)

a Mice were immunized with 100pg of the appropriate collagen emulsified in complete Freund’s adjuvant at the base of the tail. The mice were rested 14 days and challenged. Similar results were obtained when only 7 days rest was utilized. bA representative experiment is shown with five mice per group. Similar patterns were obtained in a minimum of three separate experiments. ‘The incidence of arthritis in optimally immunized DBA/lJ mice after 90 days was 100% (5/5) with B-C11and 80% (4/5) with C-CII. The incidence of arthritis in the SWR/J strain was 0% (O/5) with B-C11 and 0% (O/5) with C-CII. Similar results were obtained in two separate experiments.

reports in the literature that this strain spontaneously develops an arthritic condition which appears as the lymphoproliferative state increases (16). We chose the >2-yearold CBA/Ca mouse to determine whether in the aged state this strain would show any changes in its capacity to respond to collagen or to develop arthritis. We also chose the C57/B 16-bg mice since this strain has a reported deficiency of PMNs ( 17) among other immune defects. Finally, we chose the BlO.D20 strain because of the reported complement deficiency (18). As demonstrated in Table 5, the diseased 4-month-old MRL/lpr strain demonstrated a decreased capacity to respond to the collagens as compared to their younger controls. This observation is not peculiar to the collagen% as the same strain shows a generalized decreasedcell-mediated immune response capacity. Further, the MRL/lpr strain did not show a change in the frequency of arthritis following immunization with B-C11(seebelow). The aged CBA/Ca mice also displayed a significantly reduced capacity to respond to CII. The BlO.D2 and the C57B1/6-bg mice displayed normal capacities to generate CII-specific CM1 responses.

MURINE COLLAGEN-SPECIFIC CM1 RESPONSES

323

TABLE 4 Strain-Dependent Ability to Generate DTH Reactions following Immunization and/or Challenge with Murine Collagen Type II (M-CII)

Strain

H-2

DBA/lJ

9

CBA/J

k

Balb/c

d

Collagen used in immunization” B-C11b M-C11b B-C11 M-C11 B-C11 M-C11

Day 14 24-hr ear swelling response X 10m4in. (*SE) following challenge with M-C11 10.3 (0.9) 29.5 (0.5) 31.4 (0.4) 27.2 (1.0) 41.4 (0.5) 20.9 (0.6) 18.8 (0.6) 18.2 (1.2) 13.9 (0.7)

’ Mice were immunized with 100 pg of the appropriate collagen emulsified in complete Freund’s adjuvant and rested. The mice were then challenged with the appropriate antigen in the pinna of the ear and increases in ear swelling were determined 24 hr later. Results shown are from two separateexperiments with five mice per group. bThe incidence of arthritis in optimally sensitized DBA/l mice after a 90-day observation period was 100%(4/4) for B-C11and 0% (O/5) for M-CII. Arthritis induction following M-C11immunization was tested in a single experiment.

TABLE 5 Cell-Mediated Responsesof Selected Strains of Immunodeficient Mice

Strain of responder 2-month MRL/mp-lpr” 4-month MRLlmplpr” 2-month CBA/Ca 2-year CBA/Ca C57B1/6 C57B1/6-6g BIO.D20SnJ BlO.DZnSoJ

24-hr ear swelling response X 10e4in. (*SE) 28.9 (4.9) 11.0 (1.0) 23.5 (4.3) 12.5 (3.4) 32.7 (2.3) 45.8 (3.1) 37.8 (1.9) 34.6 (3.9)

[email protected] representative experiment with five mice per group is shown. Mice were immunized with 100 Fg of B-C11emulsified in complete Freund’s adjuvant and rested for 14 days. The mice were challenged with 10 pegof B-C11in the dorsum of the ear and increases in ear swelling were determined 24 hr later. Values shown have the negative control values subtracted. These negative control values ranged from 10 to 25 X 1O-4in. a The incidence of arthritis following optimal sensitization and a 90day observation period in the 2-month MRL/lpr mice was 0% (O/5) and in the 4-month-old MRL/lpr mice was 0% (O/5). The incidence in DBA/J immunized mice was 100%(5/5). Similar results were obtained in other experiments.

324

BUTLER ET AL.

Cross-Reactivity of Collagens at the DTH Level We next determined whether mice immune to CII from different animal species show cross-reactive CM1 responses. This approach was taken as part of a study to investigate potential strain-dependent response patterns of arthritis-susceptible and arthritis-resistant strains. The results of this approach are shown in Tables 4, 5, and 6. The results show that DBA/ 1, SWR/J, and CBA/J mice showed cross-reactive CM1 responsesto B-C11 and C-CII. In contrast, C57B1/6 and Balb/c mice were low responders to C-C11either as an immunizing or a challenge antigen (Table 4). The results of Table 5 show that DBA/l and CBA/J mice show cross-reactivity between M-C11 and Bovine CII. However, in the case of the CBA/J mice, M-C11 could elicit a CM1 response but not immunize for a CM1 response. Balb/c mice were not responsive to M-C11 at all. BecauseDBA/l mice are susceptible to arthritis induction following immunization with CII, we more extensively investigated the cross-reactivity pattern of this strain. DBA/l mice were immunized with bovine CII (B-CII), chick CII (C-CII), or mouse CII (M-CII) and then challenged with each of the collagens or in certain caseswith denatured B-CII. The results show that all the CII species can generate and elicit a CM1 response and that there appears to be a general cross-reactivity between the various speciesof collagen although some cross-reactivity patterns are more restricted than others. The order of magnitude of responseselicited in B-CII-immunized mice was denatured (dN)-B-C11 > C-C11> B-C11= M-CII. For mice immunized with CCII, the order was C-C119 B-C11> M-CII. Finally, for mice immunized with M-CII, the order was M-C11 > B-C11= C-CII. DBA/l mice optimally immunized (100 pg in CFA at the base of the tail on Days 0 and 7, i.e., 2X) with B-C11or C-C11developed arthritis (at 80% or greater incidence) during a 3-month observation period while

TABLE 6 Cross-Reactivity of Collagens at the DTH Level in DBA/l J Mice 24-hr ear swelling responseat Day 7 X 10e4in. with a challenge antigen” Immunizing antigen of DBA/lJ mice

B-C11

Part A None B-C11 c-c11 M-C11

10.0 + 28.0 + 33.0 f 23.7 +

PartB None B-C11

11.0 + 0.8 56.8 -t 1.4

dNB-CII

1.3 0.8 1.3 0.9

c-c11

M-C11

20.4 f 2.4 38.0 -t 0.3 45.4 f 1.6 24.5 -+ 0.9

11.0 f 0.5 26.7 _+3.6 20.0 + 0.8 29.4 + 2.3

16.3 12.0 67.6 2 1.6

Note. Mice were immunized with the appropriate collagen type II antigen emulsified in complete Freund’s adjuvant and rested for 7 days. The mice were then challenged with 10 pg of the appropriate antigen in the dorsum of the ear and increasesin ear swelling were determined 24 hr later. The data shown (mean f SE) are from at least two experiments with five mice per group. a B-CII, bovine collagen type II; C-CII, chick collagen type II; dNB-CII, denatured (56°C for 1 hr) bovine collagen type II; M-CII, murine collagen type II.

325

MURINE COLLAGEN-SPECIFIC CM1 RESPONSES

DBA/ 1 mice immunized with M-C11 did not. A single immunization ( 1X) with B-C11 or C-C11resulted in a 20-40% incidence of arthritis. These general trends are further substantiated by the representative experimental data regarding arthritis assessments described in the following paragraph. All the data were obtained during a 90day observation period [incidence: B-C11 (2X), 80%; B-C11(1 X), 37%; C-C11(2X), 83%; C-C11(1 X), 40%; M-C11 (2X), 0%. Mean arthritic index: B-C11(2X), 2.4; B-C11(1 X), 1.7; C-C11(2X), 2.8; C-C11(1 X), 1S. Mean day of onset: B-C11(2X), 63.8 days; B-C11(lx), 74.7 days; C-C11(2X), 48.6 days; CCII (1 X), 42.0 days.] Similar profiles were obtained in other experiments. Ability of B-CII Immune Mice to Respond to CNBr Peptides of Collagen Type II To further examine the CM1 responseto CII and in an attempt to identify a potential immunodominant determinant on bovine collagen type II, we examined the ability of B-C11immunized mice to respond to ear challenge with cyanogen-bromide fragments of bovine collagen type II. Initial studies with DBA/l mice failed to reveal strong CM1 responsesto the isolated peptides although mixtures of the isolated peptides did elicit strong CM1 reactions. Therefore, we chose the Balb/c strain becauseof their higher responder status. Balb/c mice were immunized with B-C11and challenged with various CNBr fragments of B-CII. The results are shown in Table 7. A mixture of the B-C11fragments (CB7-9, CB- 10, CB- 11, and CB- 12) or peptide CB- 11 or CB- 12 was capable of eliciting a CM1 response while peptide CB-7-9 or CB-10 was not effective for eliciting a CM1 response. DISCUSSION We have investigated the capacity of mice to develop cell-mediated immune responsesto collagens in vivo. Our initial studies examined routes of immunization and sites of elicitation for collagen-specific CM1 responses.The results demonstrated that TABLE I B-C11Immune Mice Can Respond to CNBr” Peptides of the B-C11Molecule Immunization of Balb/C mice with B-C11

+

+

24-hr ear swelling response X 10m4in. with a challenge antigen Day challenge

B-C11

Pep-7-9

Pep-10

Pep- 11

Pep-12

7 7

25.5 rt 0.9 45.9 + 2.3

16.7 + 1.3 21.5 + 1.6

21.9 + 0.3 20.9 -+ 1.3

17.3 5 0.9 33.5 f 1.4

16.7 + 0.6 44.3 f 0.9

14 14

B-C11

Pep-7-9,10,11-12

Pep-7-9,10

Pep-l l-12

30.4 2 0.7 80.8 + 1.4

21.1 + 0.9 56.1 + 1.1

28.4 + 1.6 30.2 zk 1.7

18.6 + 0.6 46.8 + 1.0

Note. Mice were immunized with 100 @ of B-C11emulsified in complete Freund’s adjuvant and rested. The mice were then challenged with the appropriate antigen in the dorsum of the ear and increasesin ear swelling were determined 24 hr later. Negative control values have been subtracted. These negative control values ranged from 5 to 25 X lo-’ in. The data shown (mean 2 SE) are from three experiments with five mice per group. ’ CNBr, cyanogen bromide.

326

BUTLER ET AL.

sc or intratail injections were the most efficient routes for CM1 sensitization while the ip route was much less efficient. These results are similar to those reported in other antigen systems (14, 15). Further, both the ears and footpads, standard sites for CM1 elicitation in the mouse, could serve as sites for detection of collagen-specific CM1 reactions. Having optimized the protocols for sensitization and elicitation of collagenspecific CM1 responses,we next examined the immunologic specificity of the collagenspecific CM1 response as well as the immunopathology profile. Our results show that the CM1 response is specific for either collagen type II or collagen type I. These collagen-specific cell-mediated immune response profiles vary with time relative to the number of days following immunization. Seven to fourteen days subsequent to the immunization of mice with collagen type II or collagen type I, strong cell-mediated immune responsescan be detected at 24 hr (but not at 4 hr) following challenge either in the ear or in the footpad. The kinetics of development of theseresponsesfollowing challenge resemblethe profile seenin other antigen systems (19-22). An increasing involvement of antibody and 4-hr ear swelling is seeninitially beginning at Day 10, becoming increasingly clear at Day 14, and actually becoming the predominant immune response elicited in animals that are 21 days or older following immunization. Histopathology studies confirmed our ear swelling data. The reactions at Day 7, 10, or 14 showed a marked cellular infiltrate containing neutrophils and monocytes, the latter usually predominating. This histopathologic profile has also been observed in CM1 responsesto other antigens in mice (19-24). In mice, it appears that neutrophils play a larger role in CM1 reactions (20). It is noteworthy that the C57B 1/6&g mice, with a documented defect in granulocyte chemotoxis (17), as well as BlO.D20 with a documented complement deficiency (18), displayed a normal capacity to generate CM1 reactions. Only a minimal cellular infiltrate was observed in Day 2 1 through Day 90 reactions. This latter histopathology profile is similar to the results obtained by Titus and Chiller (21) when only serum was used to elicit the immune response.Our results indicate that one must be careful when assessingapparent CM1 responses,as the immunopathologic profile will differ depending upon the time course of the elicitation of the response relative to the sensitization phase. We feel that our results showing shifting immunopathology profiles, in part, might provide an explanation for some of the discordant results reported by other researchersin assessing the relationship between cell-mediated immune responses and the development of arthritis. Becauseof these differences in the type of immune reaction elicited following challenge at various times after immunization, we studied the ear swelling responsesof several strains for a consecutive 42-day period. The results revealed a bimodal response pattern with a peak at Day 14, followed by a decreaseat Day 2 l-28, and then a gradual increase. Stuart et al. (3) observed a similar pattern when they monitored in vitro collagen-specific proliferation. Several possibilities can explain this pattern including the development of regulatory anti-idiotype antibodies (25, 26), negative regulatory cells (27, 28), or switch-over to a predominate Arthus component due to increased IgG antibody levels (3, present report). These possibilities are currently under investigation. During our study of the kinetics of the histopathological changes associated with these responses,we found that strain-dependent differences were apparent in the capacity to generate CM1 responsesto CI and CII (29). These differences appear to be

MURINE

COLLAGEN-SPECIFIC

CM1 RESPONSES

327

related to H-2 genes since the same strain relationships seen in noncongenic strains were also apparent using the BlO series, for CI, H-2b, and H-2k strains are good CM1 responders while H-2d strains are poorer responders. This CM1 response pattern is similar to that described for the CI-specific antibody response(30) except that BlO.BR mice show better CI-specific CM1 responses.Indeed, if the CI immunization dose was lowered to 5kg in the Ab study (30), the response pattern looked very similar to our CM1 profile. For CII-specific CM1 responses,the response patterns varied, depending on the origin of the collagen. Of particular note is the ability of the arthritis-susceptible DBA/J strain to respond to M-CII. Our in vivo CM1 response patterns are different from those reported for CII-specific murine T proliferative responses in vitro (4) as well as for CII-specific antibody production (2). However, as our data indicate, the source of collagen type II is a critical factor for these response pattern studies. Additionally, the reason for the differences in the former case may involve differing responding T-cell populations (3 1, 32) while in the latter situation both differing responding T-cell populations aswell as different B-cell repertoires (in non-H-2 congenic mice) may be responsible.We have observedthat one can elicit CM1 responsesfollowing pretreatment with cyclophosphamide to reduce potential antibody involvement as well as negative regulatory T cells (33, 34). The latter point is especially relevant as some response patterns change with differing antigen amounts (30, 33, 35, 36) and suppressor cells generated by antigen overload are cyclophosphamide sensitive (33). With this kind of an approach, we have been able to verify strain-related differences similar to those observed without the use of cyclophosphamide (data not shown). The immunodeficient autoimmune 4-month-old MRL/lpv strain showed decreased collagen-specificCM1 responsesalthough the MRL/lpr strain has been reported to develop arthritic changes spontaneously in a low percentage of mice ( 16). The immunization of these strains with CII did not cause a change in the time of onset or frequency of arthritis observed during the 3-month observation period. It is particularly interesting to note that the strain survey shown indicates that SWR/J mice, an H-2q strain and thus presumably susceptible to arthritis induction (2) does not develop arthritis following our immunization protocol either with a single injection or with a double injection of either B-C11or C-CII. A previous report has also indicated a lower susceptibility for SWR/J (1). These results suggesta role for non-H-2-linked genesin the development of CII-related murine arthritis, a situation also recently described for the CII-induced arthritis in the rat (37). We have begun to addressthe CM1 response profile to different speciesof collagen type II or CNBr fragments of B-CII. Our results indicate that in vivo CM1 responses in mice show a great degree of cross-reactivity between different species of collagen type II, but not with other antigens, even collagen type I. This capacity for crossreactivity is strain dependent. The relationship between CM1 responsesto mouse and chick collagens appear to be more restricted than the other relationships. It is especially noteworthy that the arthritis-susceptible strain DBA/ 1J responded well to M-C11while two other strains were more restricted in their capacity to respond to M-CII. DBA/ 1 mice immunized with M-C11 did not develop noticeable arthritis during a 90-day observation period. A similar observation has recently been reported by Holmdahl et al. (39). These authors further reported that a small percentage of male DBA/l mice immunized with M-C11 would develop arthritis if left for a longer observation period. The fact that denatured B-C11is very effective in eliciting a CII-specific CM1 response

328

BUTLER

ET AL.

suggeststhat at least some of the collagen-specific CM1 response may be directed to denatured CII determinants. This might explain the large degree of cross-reactivity. Further, our results using CNBr fragments of the B-C11molecule indicate that peptides CB-1 I and/or CB-12 are sufficient to elicit a murine CM1 response. This result is in agreement with the recent report of Terato et al. (40) who have studied antibody responsesand arthritogenic potential of chick-C11 CNBr fragments in DBA/ 1J mice. They find a predominant antibody responseto CB 11. However, since B-C11and Balb/ c mice were used in our study and since these mice are not susceptible to arthritis induction, the relationship between responsesto this determinant and the arthritogenic potential is still unclear. The relationship between the capacity to mount DTH reactions to CII and arthritis susceptibility in this animal model is not simple. DTH reactions involve at least two T-cell subsets (41), yet we are aware that there could be a better direct correlation between other T-cell subset response profiles (e.g., T helper function for antibody) and susceptibility to arthritis. The data presented in Tables 3-5 and Figs. 2 and 4 indicate there is not a direct correlation between the capacity to generate a CII-specific DTH response and susceptibility to arthritis, as only DBA/ 1J mice were susceptible to arthritis induction, yet most of the mouse strains tested could generate CII-specific DTH responsesto at least one type of CII. The DTH profile of the arthritis-susceptible DBA/l J strain is noteworthy, however, becausethese mice exhibited a high degreeof DTH responsivenessto all three types of CII and also displayed a highly cross-reactive DTH profile. These combined characteristics distinguish the DTH profile of DBA/ 1J mice from the other strains with the exception of the other H-2q strain, SWR/J. Additional evidence demonstrating the highly cross-reactive immune responseprofile of T cells from CII-immunized DBA/ 1J mice has recently been obtained from analysis of CII-specific T-cell lines in vitro (42,43). Suggestiveinformation identifying a critical role for T cells in the diseaseprocess came from major histocompatibility complexassociated immune response patterns (2, 4) and recently, most convincingly when a report indicated that the arthritic state could be transferred using CII specific T-cell lines (43). This latter result suggestsa strong role for T-cell involvement in the pathogenesisof arthritis in this animal model. However, there is also compelling evidence that humoral immunity is intimately involved in the pathogenesis process (44, 45). Resolution of this problem could involve both types of immune response, perhaps in kinetically different phasesof the diseaseprocess. The pathogenesis of human rheumatoid arthritis (RA) as well as the relationship of RA to immune responsesto collagen type II remains unsolved (reviewed in Ref. 46). Evidence exists for the involvement of both the cellular and humoral immune systemsjust as it does in the mouse arthritis model. The association of certain major histocompatibility gene haplotypes with RA (47-49) suggesteda possible link to Ir gene-controlled antigen-specific T-cell responses.Indeed, Yu et al. (50) described activated T cells in RA patients and Trentham et al. (5 1) reported cellular reactivity to collagens in RA patients. Later, Solinger et al. (52) reported a relationship between MHC haplotypes associated with RA and cellular responsesto collagen type II. Additional reports supplemented these observations, including identification of activated T cells in synovium and further cellular immune reactivity to collagens (53-59). Although the association of RA with certain MHC haplotypes continues to be observed (60,6 l), the absolute relationship is unclear (6 1). Additionally, the MHC-linked cellular

MURINE COLLAGEN-SPECIFIC CM1 RESPONSES

329

and humoral immune responses to collagen have also become less clear (62, 63). Perhaps new relationships between RA and MHC-controlled gene functions can be identified using newer cellular-defined MHC typing procedures (64, 65). Additional information is neededto evaluate the actual role of the cell-mediated immune response in the RA diseaseprocess. In this study we have attempted to develop information regarding in vivo murine collagen-specific CM1 responses.We have demonstrated time-dependent changes in the type of murine immune response elicited with these connective tissue antigens and have begun to develop CM1 response profiles for strain-related differences in reactivity to different speciesof CII and peptide fragments of B-CII. Currently, we are investigating the relationship of arthritis susceptibility with collagen-specific CM1 crossreactive profiles. We hope to use these approaches to better understand the immunoregulatory networks involved in potential disease-producing immune responsesto these connective tissue antigens. ACKNOWLEDGMENTS The authors acknowledge helpful discussionsearly in the project with Dr. John Stuart and Dr. Joanne Horn, the radiologic evaluations by Mr. Donald Benslay, and the expert secretarial assistanceof Ms. Mary M. Bolton.

REFERENCES I. Courtenay, J. S., Dallman, M. J., Dgan, A. D., Martin, A., and Mosedale, B.,Nature (London) 283,665. 1980. 2. Wooley, P. H., Luthra, H. S., Stuart, J. M., and David, C. S., J. Exp. Med. 154,688, 1981. 3. Stuart, J. M., Townes, A. S., and Kang, A. H., J. Clin. Invest. 69,673, 1982. 4. Rosenwasser,L. F., Bhatnagar, R. S., and Stobo, J. D., J. Immunol. 124, 2854, 1980. 5. Trentham, D. E., Townes, A. S., and Kang, A. H., J. Exp. Med. 146,857, 1977. 6. Trentham, D. E., Townes, A. S., Kang, A. H., and David, J. R., J. C/in. Invest. 61, 89, 1978. 7. Schoen, R. T., Mehiman, H., Trentham, D. E., Perry, L., Greene, M. I., and David, J. R., J. Immunol. 127,2275, 1981. 8. Stuart, J. M., Cremer, M. A., Kang, A. H., and Townes, A. S., Arthritis Rheum. 22, 1344, 1979. 9. Griffiths, M. M., Eichwald, E. J., Martin, J. H., Smith, C. B., and Dewitt, C., Arthritis Rheum. 24,78 I.

1981. 10. Ofosu-Appiah, W. A., Morgan, K., and Holt, P. J., Ann. Rheum. Dis. 42, 331, 1983. 1I. Deshmukh, K., and Nimni, M. E., Eiochem. J. 112, 397, 1969. 12. Strawich, E., and Nimni, M. E., Biochemistry 10, 3905, 1971. 13. Miller, E. J., and Lunde, L. A., Biochemistry 12, 3153, 1973. 14. Greene, M. I., Sugimuto, M., and Benacerraf, B., J. fmmunol. 120, 1604, 1978. 15. Miller, S. D., Wetzig, R. P., and Claman, H. N., J. Exp. Med. 149, 758, 1979. 16. Hang, H., Theofilopoulos, A. N., and Dixon, F. J., J. Exp. Med. 155, 1690, 1982. 17. Gallin, J. I., Bujak, J. S., Patten, E., and Wolf, S. M., Blood 43, 201, 1974. 18. Nilsson, U., and Miiller-Eberhard, H. J., J. Exp. Med. 125, 1, 1967. 19. Asherson, G. L., and Ptak, W. Immunology 15, 405, 1968. 20. Crowle, A., In “Advances in Immunology” (K. Moldave and L. Grossman, Eds.), Vol. 20, p. 197, 1975. 21. Titus, R. G., and Chiller, J. M., J. Immunol. Methods 45, 65, 1981. 22. Van Loveren, H., and Askenase, P. W., J. Immunol. 133,2397, 1984. 23. Schultz, L. D., and Bailey, D. W., Immunogenetics 1, 570, 1975. 24. Cohen, S., Hum. Pathol. 7, 249, 1976. 25. Sy, M.-S., Moorhead, J. W., and Claman, H. N., J. Immunol. 123, 2593, 1979. 26. Asherson, G. L., Zembala, M., Gautam, S., and Watkins, M. C., Cell Immunol. 70, 160, 1982. 27, Sy, M.-S., Bach, B. A., Brown, A., Nisonoff, A., Benacerraf, B., and Greene, M. I., J. Exp. Med. 150, 1229, 1979.

330

BUTLER ET AL.

28. Miller, S. D., Butler, L. D., and Claman, H. N., J. Immunol. 129, 461, 1982. 29. Butler, L. D., Simmons, B., Fouts, R., DeRiso, P., Parrish, J., and Phadke, K., Fed. Proc. 43, 1992, 1984. 30. Hedrick, S. M., and Watson, J. D., J. Immunogenet. (Oxford) 7, 271, 1980. 31. Miller, S. D., and Butler, L. D., J. Immunol. 131, 77, 1983. 32. Van Loveren, H., Kate, K., Meade, R., Green, D. R., Horowitz, M., Ptak, W., and Askenase, P. W., J. Immunol. 133,2402, 1984. 33. Sy, M.-S., Miller, S. D., and Claman, H. N., J. Immunol. 119, 240, 1977. 34. Turk, J. L., Parker, D., and Poulter, L. W., Immunology 23, 493, 1972. 35. LaGrange, P. H., Mackaness, G. B., and Miller, T. E., J. Exp. Med. 139, 1529, 1974. 36. Askenase, P. W., Hayden, B. J., and Gershon, R. K., J. Exp. Med. 141,697, 1974. 37. Griffiths, M. M., and Dewitt, C. W., J. Immunol. 133, 3043, 1984. 38. Watson, W. C., and Townes, A. S., Fed. Proc. 43, 1992, 1984. [Abstract] 39. Holmdahl, R., Jansson, L., Gullberg, D., Rubin, K., For&erg, P. O., and Klareskog, L., Clin. Exp. Immunol. 62, 639, 1985. 40. Terato, K., Hastey, K. A., Cremer, M. A., Stuart, J. M., Townes, A. S., and Kang, A. H., J. Exp. Med. 162,637, 1985.

41. Miller, S. D., and Butler, L. D., J. Immunol. 131, 77, 1983. 42. Horn, J., Stuart, J. M., and Chiller, J. M., J. Immunol. 136, 769, 1986. 43. Holmdahl, R., Klareskog, L., Rubin, K., Larsson, E., and Wigzell, H., Scund. J. Zmmunol. 22, 295, 1985. 44. Watson, W., and Townes, A. S., J. Exp. Med. 162, 1878, 1985. 45. Watson, W., Townes, A. S., Brown, P. S., and Pitcock, J. A., Clin. Res. 33, 794A, 1985. 46. Trentham, D., Proc. Sot. Exp. Biol. Med. 176, 95, 1984. 47. Gibovsky, A., Winchester, R., Hansen, J., Patarroyo, M., DuPont, B., Paget, S., Lahita, R., Halper, J., Fotino, M., Yunis, E., and Kunkel, H., Arthritis Rheum. 21, S134, 1978. 48. Stastny, P., Arthritis Rheum. 21, S139, 1978. 49. Gibovsky, A., Winchester, R. J., Patarroyo, M., Fotino, M., and Kunkel, H. G., J. Exp. Med. 148, 1728, 1978. 50. Yu, D., Winchester, R. J., Fu, S., Gibovsky, A., Ko, H., and Kunkel, H. G., J. Exp. Med. 151, 91, 1980. 51. Trentham, D., Dynesius, R., Rocklin, R., and David, J. R., N. Engl. J. Med. 299, 327, 1978. 52. Solinger, A. M., Bhatnagar, R., and Stobo, J. D., Proc. N&l. Acad. Sci. USA 78,3877, 1981. 53. Stuart, J. M., Postlethwaite, A. E., Townes, A. S., and Kang, A. H., Amer. J. Med. 69, 13, 1980. 54. Robinson, A., and Muirden, K., Ann. Rheum. Dis. 39, 539, 1980. 55. Burmester, G., Yu, D., Irani, A-M., Kunkel, H. G., and Winchester, R., Arthritis Rheum. 24, 1370, 1980. 56. Trentham, D., Kammer, G., McCune, W., and David, J. R., Arthritis Rheum. 24, 1363, 1981. 57. Klareskog, L., Forsum, U., Scheynius, A., Kabelitz, D., and Wigzell, H., Proc. Natl. Acud. Sci. USA 79, 3632, 1982. 58. Golds, E., Stephen, I,, Esdaile, J., Strawczynski, H., and Poole, A. R., Cell. Immunol. 82, 196, 1983. 59. Pincus, S., Clegg, D., and Ward, J. R., Arthritis Rheum. 28, 8, 1985. 60. Sasrzuki, T., J. Rheum. (SuppI.) 10, 54, 1983. 61. Walker, D., Grifiths, M., Dewar, P., Coates, E., Dick, W., Thompson, M., and Griffiths, I. D., Ann Rheum. Dis. 44, 519, 1985. 62. Collier, D., Kerwar, S., Garovog, M., Fye, K., and Stobo, J. D., Arthritis Rheum. 27, 1201, 1984. 63. Kammer, G., and Trentham, D., Arthritis Rheum. 27,489, 1984. 64. Weyland, C. M., Goronzy, J., and Fathman, C. G., Clin. Rex 33, 5 13A, 1985. 65. Goronzy, J., Weyland, C. M., and Fathman, C. G., Clin. Res. 33, 377A, 1985.