To establish the role of posttranslational modification in modulating the immune response to collagen, recombinant human type II collagen (rCII) was produced using a yeast expression system (rCIIpic) and a baculovirus expression system (rCIIbac). The biosynthesis of CII requires extensive posttranslational modification including the hydroxylation of prolyl and lysyl residues and glycosylation of selected hydroxylysyl residues. Amino acid analyses indicated that the rCIIbac was adequately hydroxylated at prolyl residues but underhydroxylated at lysyl residues and underglycosylated compared with tissue-derived CII, whereas rCIIpic was adequately hydroxylated at prolyl residues but unhydroxylated at lysyl residues and had no glycosylation. When DBA/1 mice were immunized with rCII, rCIIpic induced a lower incidence of arthritis than tissue-derived CII, whereas rCIIbac induced an intermediate level of arthritis. The severity of the arthritis was significantly lower in mice immunized with rCIIpic compared with mice immunized with tissue-derived CII, whereas that of rCIIbac was intermediate. These data indicate that the degree of lysine hydroxylation and glycosylation plays a role in the induction of arthritis. The recombinant collagens were then compared with tissue-derived CII when given as i.v. or oral tolerogens to suppress arthritis. Both recombinant collagens were less potent than tissue-derived CII, and this decrease in arthritis was associated with a decrease in Ab response to CII. These data suggest that the degree of glysosylation affects the immune response to CII, so that underglycosylated CII is less effective in the induction of arthritis and in its ability to suppress collagen-induced arthritis.

The study of type II collagen (CII)3 antigenic determinants and their role in the induction of collagen-induced arthritis (CIA) has been hampered by several aspects of the molecular features of CII and the pathogenic mechanisms of CIA. Confounding factors include the inability to induce disease with small antigenic peptides, and the possible role of posttranslational modifications of CII in the induction of the autoimmune response. Although native CII is very effective at eliciting CIA, denatured CII and fragments of the α1(II) chain are progressively inferior (1, 2). These characteristics make it difficult to study the role of simple antigenic determinants in the pathogenesis of this model. Similarly, T cell recognition of posttranslation modifications of CII, including hydroxylation of prolines and lysines and the glycosylation of the hydroxylysines (Hyl), may also play an important role in the development of the T cell autoimmune response. The role of the carbohydrate is especially of interest, because carbohydrate-dependent CII-specific T cells have been described in the H-2q model of CIA (3, 4, 5).

Using recently developed expression systems, rCII can now be produced so that the proteins expressed will maintain the triple helical structure of the native CII molecule, yet have a variable level of glycosylation. Production of CII in yeast using Pichia pastoris results in CII with no detectable Hyl and Hyl glycosides, as P. pastoris is lacking in lysyl hydroxylase and Hyl glycosyl transferases. Production of CII in baculovirus results in collagen that contains some glycosylation, but significantly less than that of cartilage-derived CII. Collagens produced using these expression systems can be used to evaluate the importance of glycosylation and posttranslational modifications for both the induction of autoimmune arthritis and preventing the development of CIA.

Native CII was solubilized from either human cartilage obtained from individuals of <18 years of age, fetal calf articular cartilage, or murine sternal cartilage by limited pepsin digestion and purified as described earlier (6). The purified collagen was dissolved in cold 0.01 M acetic acid at 4 mg/ml and stored frozen at −70°C until used.

The baculovirus expression vector system was used to produce rCIIbac in insect cells essentially as described earlier (7). Briefly, Trichoplusia ni cells (High Five; Invitrogen, San Diego, CA) were grown in suspension and infected with recombinant baculoviruses encoding the human (h)CII and a second virus encoding the α and β subunits of human prolyl 4-hydroxylase. After 72-h incubation in the presence of l-ascorbic acid phosphate, rCII was isolated from the culture medium by limited proteolysis with pepsin followed by precipitation with ammonium sulfate, redissolved and purified by gel filtration and cation exchange chromatography, and dialyzed in dilute acetic acid.

Similarly, cDNA for the pro-α1 chains of CII were cloned into the expression vector pPICZB and transformed into a recombinant P. pastoris strain expressing human prolyl 4-hydroxylase (8). The recombinant collagen (rCIIpic) produced was purified by pepsin digestion and selective salt precipitation followed by gel filtration in a similar manner (8, 9).

DBA/1 LacJ mice (I-Aq) were purchased from The Jackson Laboratory (Bar Harbor, ME). All were female and 5–7 wk of age at the time of purchase, and they were maintained in a specific pathogen-free environment. They were fed standard mouse chow and water ad libitum. All animals were kept until the age of 8–10 wk before being used in these experiments.

CII was solubilized in 0.01 M acetic acid at a concentration of 4 mg/ml and emulsified with an equal volume of CFA containing 4 mg/ml Mycobacterium tuberculosis (Ministry of Agriculture, Fisheries, and Food, Weybridge Surrey, U.K.) (9). Each mouse received 25 μg of rCII or CII emulsified in CFA intradermally at the base of the tail and a second dose of 25 μg of rCII or CII emulsified with IFA 3 wk later, for a total of 50 μg.

DBA/1 mice were tolerized by either i.v. or oral administration of rCII or CII. For i.v. tolerization, CII and rCII were solubilized in 0.01 M acetic acid and dialyzed against several changes of PBS in the cold. Mice were then injected i.v. with either PBS (negative control), tissue-derived hCII (positive control), rCIIpic, or rCIIbac. A total of 25 μg of protein was administered i.v. in three divided doses. These mice were immunized with CII 1 wk after tolerization and observed for arthritis.

In the oral tolerization experiments, groups of 12 DBA/1 mice were administered either 0.01 M acetic acid, CII (hCII), rCIIpic, or rCIIbac by oral gavage using ball-tipped gavage needles (Popper and Sons, New Hyde Park, NY). The collagens were dissolved in 0.01 M acetic acid and administered four times per week for 2 wk for a total of eight doses. Each dose administered was 10 μg of protein, and the total dose was 80 μg of protein. Mice were immunized with CII 3 days after the last dose and observed for the development of arthritis.

The incidence and severity of arthritis were determined by visually examining each forepaw and hindpaw and scoring them on a scale of 0–4 as described previously (9). Scoring was conducted by two examiners, one of whom was unaware of the identity of the treatment groups. Each mouse was scored thrice weekly beginning 3 wk postimmunization and continuing for 8 wk. The incidence of arthritis (number of animals with one or more arthritic limbs) was recorded at each time point. The incidence and severity values shown represent data analyzed 8 wk after immunization.

Mice were bled at 6 and 8 wk after immunization, and sera were analyzed for Abs reactive with native CII and rCII using a modification of an ELISA previously described (9). Serial dilutions of a standard serum were added to each plate. From these values, a standard curve was derived by computer analysis using a four-parameter logistic curve. Results are reported as units of activity, derived by comparison of test sera with the curve derived from the standard serum, which was arbitrarily defined as having 50 U of activity. Reactivity to CII and rCII was not detected in sera obtained from normal mice.

Ten days after immunization, draining lymph nodes were removed, disassociated, and washed in HL-1 (BioWhittaker, Walkersville, MD). Lymphocytes were cultured in 96-well plates at 4 × 105/well in 300 μl of HL-1 medium supplemented with 50 μM 2-ME, 2 mM glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin, and 0.1% BSA (fraction V; IgG free, low endotoxin; Sigma-Aldrich, St. Louis, MO) at 37°C and 5% humidified CO2 for 4 days. Eighteen hours before the termination of culture, 1 μCi of [3H]thymidine (New England Nuclear, Boston, MA) was added to each well. Cells were harvested onto glass fiber filters, washed, and counted using a Matrix 96 direct ionization beta counter (Packard Instrument, Meridian CT). Proliferation assays using rCII and CII were performed in triplicate at concentrations of 25, 50, and 100 μg. Results were confirmed by replicate experiments, and all data are expressed as disintegrations per minute.

In some experiments, quantitative measurement of murine IFN-γ, and IL-4, was done using a solid-phase ELISA based on the sandwich principle. Commercially available kits were used (IFN-γ; Life Technologies, Gaithersburg, MD; IL-4; Endogen, Boston, MA). Each sample was tested with duplicate wells.

The incidence of arthritis in various groups of mice was compared using Fisher’s exact test. Mean severity scores and Ab levels were compared using Student’s t test.

To study the role of hydroxylation and glycosylation in the arthritogenicity of CII, rCII expression systems were used to express proteins having differing degrees of these posttranslational modifications while maintaining the triple helical structure of the native CII molecule. Samples from the resulting rCIIbac and rCIIpic were analyzed using a Beckman 6300 amino acid analyzer to determine the content of hydroxyproline, proline, and total hydoxylysine. As shown in Table I, the total numbers of galactosylhydroxyline (GH) and glucosylgalactosylhydroxyline (GGH) for rCIIbac are 1 and 4 per thousand residues, respectively (Table I), compared with a total of 4.2 GH and 1.9 GGH residues per thousand for cartilage-derived collagen, respectively. In contrast, rhCII produced in P. pastoris does not contain any Hyl or its glycosides, because the yeast cells lack the necessary enzymes (Table I). The denaturation temperature of the various collagens was determined using an enzyme susceptibility assay. rhCIIbac and rhCIIpic each were shown to have a melting temperature of 40°C, which is consistent with that found by Nokelainen et al. (7) by circular dichroism and is comparable with that of cartilage-derived CII (Fig. 1). Taken together, these data confirm that rhCIIbac is only partially glycosylated, whereas rCIIpic has no Hyl glycosides. Both forms of recombinant collagen maintain the triple helical structure in a manner comparable to the cartilage-derived CII.

Table I.

Quantity of selected amino acids and carbohydrates in human CII obtained from cartilage or produced recombinantlya

rCIIpicrCIIbacCIIb
Hyl 9.2 
GH 4.2 
GGH 1.9 
Total (Hyl + GH + GGH) 10 15.3 
    
Hydroxyproline 87 89 99 
Proline 116 124 120 
rCIIpicrCIIbacCIIb
Hyl 9.2 
GH 4.2 
GGH 1.9 
Total (Hyl + GH + GGH) 10 15.3 
    
Hydroxyproline 87 89 99 
Proline 116 124 120 
a

The content of hydroxyproline, proline, and total Hyl was determined on collagen samples after hydrolysis in 6 N HCl using a Beckman 6300 amino acid analyzer (15 ). Hyl glycoside contents were analyzed after hydrolysis of collagen samples in 2 N KOH. Values are expressed as number of residues per thousand. rCIIbac, rCII produced in the baculovirus expression system; CII, cartilage-derived CII; rCIIpic, rCII produced in the yeast cell expression system.

b

Data adapted from Ref. 16 .

FIGURE 1.

Thermal stability of cartilage-derived hCII, rCIIbac, and rCIIpic. Aliquots of collagen samples (1 mg) dissolved in 0.1 M Tris/0.4 M NaCl/10 mM CaCl2 (pH 7.4) were treated with a mixture of trypsin (5 μg) and chymotrypsin (12.5 μg) for 1 min at the temperatures indicated. At the end of the incubation, soybean trypsin inhibitor (15 μg) was added. The samples were boiled in SDS sample buffer and analyzed by SDS-PAGE. As can be seen, the thermal stability of rCIIpic and rCIIbac was comparable and consistent with that of tissue-derived CII as well as with the values reported previously (7 ).

FIGURE 1.

Thermal stability of cartilage-derived hCII, rCIIbac, and rCIIpic. Aliquots of collagen samples (1 mg) dissolved in 0.1 M Tris/0.4 M NaCl/10 mM CaCl2 (pH 7.4) were treated with a mixture of trypsin (5 μg) and chymotrypsin (12.5 μg) for 1 min at the temperatures indicated. At the end of the incubation, soybean trypsin inhibitor (15 μg) was added. The samples were boiled in SDS sample buffer and analyzed by SDS-PAGE. As can be seen, the thermal stability of rCIIpic and rCIIbac was comparable and consistent with that of tissue-derived CII as well as with the values reported previously (7 ).

Close modal

The efficacy of the rhCII in inducing arthritis was evaluated by immunizing CIA-susceptible mice with native CII, rCIIbac, or rCIIpic, and observing the mice for arthritis. Both the incidence and severity of arthritis were recorded on a thrice weekly basis. As shown in Table II and Fig. 2, both the rCIIbac and rCIIpic elicited a lower incidence of arthritis (40 and 30%, respectively) compared with that induced using tissue-derived CII (90%). Similarly, the severity of the arthritis varied among the three groups. At 51 days after immunization, mean severity scores were 3.7 ± 1.5 for mice immunized with CII, 0.7 ± 0.3 for mice immunized with rCIIpic, and 1.7 ± 1 for mice immunized with rCIIbac. The mice immunized with CII differed from those immunized with rCIIpic (p ≤ 0.005), whereas mice immunized with rCIIbac had intermediate scores (Fig. 2). Similarly, Ab titers to human and murine CII correlated with the severity (Table II). Thus, following immunization with the rhCII, the incidence and severity of the arthritis observed varied, and correlated with the degree of glycosylation, especially with the GH content, of the collagen used for immunization.

Table II.

Arthritogenicity of rCIIpic and rCIIbac in DBA/1 mice

ImmunogenaArthritic MicebArthritic LimbsbAbs to hCIIcAbs to Murine CIIc
rCIIbac 4/10 (40%)d 8/40 (20%)e 43 ± 22 20 ± 8 
rCIIpic 3/10 (30%)d 4/40 (10%)e 38 ± 14 18 ± 7 
CII 9/10 (90%) 21/40 (53%) 57 ± 31 34 ± 10 
ImmunogenaArthritic MicebArthritic LimbsbAbs to hCIIcAbs to Murine CIIc
rCIIbac 4/10 (40%)d 8/40 (20%)e 43 ± 22 20 ± 8 
rCIIpic 3/10 (30%)d 4/40 (10%)e 38 ± 14 18 ± 7 
CII 9/10 (90%) 21/40 (53%) 57 ± 31 34 ± 10 
a

DBA/1 mice (10 per group) were immunized with rCIIbac, rCIIpic, or tissue-derived CII. Each mouse received 25 μg of collagen emulsified in CFA s.c. at the base of the tail and a second dose of 25 μg of protein emulsified with IFA 3 wk after the first dose, for a total of 50 μg.

b

The incidence of arthritis and number of arthritic limbs is reported at 8 wk postimmunization with statistical analysis done using Fisher’s exact test.

c

Abs represent the mean units per group, using sera collected 8 wk after immunization. ELISAs were performed using hCII or murine CII, and results are reported as units of activity, derived by comparison of test sera with the standard serum, which was arbitrarily defined as having 50 U of activity. Sera were analyzed individually, and results shown represent the mean ± SD for each group of animals with statistical analysis done using Student’s t test.

d

, p ≤ 0.05 compared to CII.

e

, p ≤ 0.005 compared to CII.

FIGURE 2.

DBA/1 mice (10 per group) were immunized with rCIIbac (▴), rCIIpic (▪), or tissue-derived CII (•). Each mouse received 25 μg of collagen emulsified in CFA given s.c. at the base of the tail and a second dose of 25 μg of protein emulsified with IFA given 3 wk after the first dose, for a total of 50 μg. The data are expressed as mean severity scores over time. At 51 days after immunization, mean severity scores were 3.7 ± 1.5 for mice immunized with CII, 0.7 ± 0.3 for mice immunized with rCIIpic, and 1.7 ± 1 for mice immunized with rCIIbac. Scores for the mice immunized with CII differed from those immunized with rCIIpic (p ≤ 0.005).

FIGURE 2.

DBA/1 mice (10 per group) were immunized with rCIIbac (▴), rCIIpic (▪), or tissue-derived CII (•). Each mouse received 25 μg of collagen emulsified in CFA given s.c. at the base of the tail and a second dose of 25 μg of protein emulsified with IFA given 3 wk after the first dose, for a total of 50 μg. The data are expressed as mean severity scores over time. At 51 days after immunization, mean severity scores were 3.7 ± 1.5 for mice immunized with CII, 0.7 ± 0.3 for mice immunized with rCIIpic, and 1.7 ± 1 for mice immunized with rCIIbac. Scores for the mice immunized with CII differed from those immunized with rCIIpic (p ≤ 0.005).

Close modal

T cells from mice immunized with either cartilage-derived CII or each of the recombinant collagens were cultured with varying doses of α1(II) chains of CII, and the resulting proliferative responses were analyzed. The greatest responses to α1(II) were detected in T cells taken from mice immunized with tissue-derived CII. T cells from mice immunized with rCIIpic had the lowest proliferative responses, whereas cells from mice immunized with rCIIbac gave intermediate responses (Fig. 3). These titers correlate with both the glycosylation, especially the GH content, of the immunizing Ag and the severity of the arthritis. Similarly, when T cells from mice immunized with the various recombinant collagens were cultured with CII and the supernatants were evaluated to determine the amounts of cytokines secreted, both Th1 and Th2 cytokines secreted correlated with the degree of glycosylation (IFN-γ responses were 647 pg/ml from rCIIpic, 699 pg/ml from rCIIbac, and 3000 pg/ml from CIItd; IL-4 responses were 3 pg/ml from rCIIpic, 6 pg/ml from rCIIbac, and 8 pg/ml from CIItd). Taken together, these data reveal that, although both types of recombinant collagen are immunologically competent, the degree of glycosylation affects the severity of the T cell immune response to CII.

FIGURE 3.

DBA/1 mice were immunized with either rCIIpic, rCIIbac, or CII emulsified in CFA. Ten days later, draining lymph node cells were harvested, pooled, and cultured with α1(II) chains or purified protein derivative as indicated. Ags were tested at 100, 50, or 25 μg/ml. Eighteen hours before the termination of the cultures, 1 μCi of [3H]thymidine was added to each well. Cells were harvested onto glass fiber filters, and radioactivity was counted. Data are expressed as disintegration per minute.

FIGURE 3.

DBA/1 mice were immunized with either rCIIpic, rCIIbac, or CII emulsified in CFA. Ten days later, draining lymph node cells were harvested, pooled, and cultured with α1(II) chains or purified protein derivative as indicated. Ags were tested at 100, 50, or 25 μg/ml. Eighteen hours before the termination of the cultures, 1 μCi of [3H]thymidine was added to each well. Cells were harvested onto glass fiber filters, and radioactivity was counted. Data are expressed as disintegration per minute.

Close modal

In another approach, the efficacy of rCII in inducing tolerance in mice was evaluated by i.v. administration of either rCIIbac, rCIIpic, or tissue-derived CII into mice before immunization with CII. Mice were administered a total of 25 μg of each CII, immunized with CII, and observed for the development of arthritis. As expected, tissue-derived CII was very effective as a tolerogen, totally preventing the induction of arthritis (Fig. 4). The efficacy of the rhCII in inducing tolerance and preventing CIA correlated with their degree of glycosylation, with CIIbac inducing a strong tolerance as evaluated both by incidence and severity of arthritis in comparison with the weak tolerance induced by rCIIpic, which lacks glycosylation (Fig. 4). The effect on the Ab responses correlated with their effect in preventing arthritis development (Table III).

FIGURE 4.

DBA/1 mice (10 per group) were injected i.v. with either 50 mM acetic acid (♦; vehicle control), tissue-derived CII (•), rCIIbac (▴), or rCIIpic (▪). A total of 25 μg of protein was administered i.v. in three equal doses. Mice were immunized with tissue-derived CII emulsified in CFA 7 days after the last i.v. dose. The upper panel reports incidence of arthritis, and the lower panel shows mean severity scores at various time points after immunization. At 51 days after immunization, severity scores of mice treated with vehicle control (severity score of 5.8 ± 2) differed significantly from those of all other treatment groups: rCIIpic (3.3 ± 2; p ≤ 0.05), rCIIbac (0.6 ± 0.3; p ≤ 0.0005), and CII (0). Similarly, mice treated with rCIIpic differed from mice treated with either CII (p ≤ 0.0005) or rCIIbac (p < 0.01). The severity scores of mice treated with rCIIbac differed significantly from that of mice treated with rCIIpic (p < 0.01), but did not differ statistically from mice treated with tissue-derived CII.

FIGURE 4.

DBA/1 mice (10 per group) were injected i.v. with either 50 mM acetic acid (♦; vehicle control), tissue-derived CII (•), rCIIbac (▴), or rCIIpic (▪). A total of 25 μg of protein was administered i.v. in three equal doses. Mice were immunized with tissue-derived CII emulsified in CFA 7 days after the last i.v. dose. The upper panel reports incidence of arthritis, and the lower panel shows mean severity scores at various time points after immunization. At 51 days after immunization, severity scores of mice treated with vehicle control (severity score of 5.8 ± 2) differed significantly from those of all other treatment groups: rCIIpic (3.3 ± 2; p ≤ 0.05), rCIIbac (0.6 ± 0.3; p ≤ 0.0005), and CII (0). Similarly, mice treated with rCIIpic differed from mice treated with either CII (p ≤ 0.0005) or rCIIbac (p < 0.01). The severity scores of mice treated with rCIIbac differed significantly from that of mice treated with rCIIpic (p < 0.01), but did not differ statistically from mice treated with tissue-derived CII.

Close modal
Table III.

Prevention of CIA by induction of i.v. tolerance with rCIIbac and rCIIpic in micea

TolerogenAbs to CIIb
hCIIMurine CII
PBS 57 ± 12 26 ± 12 
rCIIbac 10 ± 5c <1.0c 
rCIIpic 13 ± 7c 2.7 ± 5c 
CII 2 ± 2c <1.0c 
TolerogenAbs to CIIb
hCIIMurine CII
PBS 57 ± 12 26 ± 12 
rCIIbac 10 ± 5c <1.0c 
rCIIpic 13 ± 7c 2.7 ± 5c 
CII 2 ± 2c <1.0c 
a

DBA/1 mice (10 per group) were injected i.v. with either PBS (negative control), tissue-derived CII, rCIIbac, or rCIIpic. A total of 25 μg of protein was administered i.v. in divided doses on 3 consecutive days. Mice were immunized with CII emulsified in CFA 7 days after the last i.v. dose.

b

Abs represent the mean units per group using sera collected 6 and 8 wk after immunization. ELISAs were performed, and results are reported as units of activity, derived by comparison of test sera with standard, which was arbitrarily defined as having 50 U of activity. Sera were analyzed individually, and results shown represent the mean ± SD for each group of animals with statistical analysis done using Student’s t test comparing each group to the PBS control group.

c

, p ≤ 0.005, using Fisher’s exact test.

Similarly, when the various types of rhCII were administered orally before immunization, tissue-derived CII was very effective as a tolerogen (Fig. 5). rCIIbac was very similar to tissue-derived CII in reducing the incidence and severity of arthritis; rCIIpic, which lacks Hyl glycosides, was slightly less effective (Fig. 5). The effect on the Ab responses correlated with their effect in preventing arthritis development (Table IV).

FIGURE 5.

Oral tolerance. DBA/1 mice (10 per group) were treated orally with either 50 mM acetic acid (♦; vehicle control), tissue-derived hCII (•), rhCIIbac (▴), or rhCIIpic (▪). A total of 80 μg of protein was administered orally in eight equal doses over 2 wk before immunization. Mice were immunized with hCII emulsified in CFA after the last dose. The upper panel reports incidence of arthritis, and the lower panel shows mean severity scores at various time points after immunization. At 53 days after immunization, the severity scores of mice treated with vehicle control (severity score, 5.2 ± 2) differed significantly from severity scores of all other treatment groups: rCIIpic (2.6 ± 1.5; p ≤ 0.005), rCIIbac (1.3 ± 1; p ≤ 0.005), and CII (0.4 ± 0.2; p ≤ 0.005). Similarly, severity scores of mice treated with rCIIpic differed from those of mice treated with either CII (p ≤ 0.005) or rCIIbac (p < 0.05). The severity scores of mice treated with rCIIbac differed significantly from that of mice treated with rCIIpic (p < 0.05), but did not differ statistically from the severity scores of mice treated with tissue-derived CII.

FIGURE 5.

Oral tolerance. DBA/1 mice (10 per group) were treated orally with either 50 mM acetic acid (♦; vehicle control), tissue-derived hCII (•), rhCIIbac (▴), or rhCIIpic (▪). A total of 80 μg of protein was administered orally in eight equal doses over 2 wk before immunization. Mice were immunized with hCII emulsified in CFA after the last dose. The upper panel reports incidence of arthritis, and the lower panel shows mean severity scores at various time points after immunization. At 53 days after immunization, the severity scores of mice treated with vehicle control (severity score, 5.2 ± 2) differed significantly from severity scores of all other treatment groups: rCIIpic (2.6 ± 1.5; p ≤ 0.005), rCIIbac (1.3 ± 1; p ≤ 0.005), and CII (0.4 ± 0.2; p ≤ 0.005). Similarly, severity scores of mice treated with rCIIpic differed from those of mice treated with either CII (p ≤ 0.005) or rCIIbac (p < 0.05). The severity scores of mice treated with rCIIbac differed significantly from that of mice treated with rCIIpic (p < 0.05), but did not differ statistically from the severity scores of mice treated with tissue-derived CII.

Close modal
Table IV.

Ab levels in DBA/1 mice after oral tolerancea

TolerogenAbs to hCIIAbs to Murine CII
Vehicle control 55 ± 18 32 ± 9 
CII 15 ± 10b 9 ± 3b 
rCIIbac 19 ± 11b 11 ± 4b 
rCIIpic 19 ± 12b 4 ± 1b 
TolerogenAbs to hCIIAbs to Murine CII
Vehicle control 55 ± 18 32 ± 9 
CII 15 ± 10b 9 ± 3b 
rCIIbac 19 ± 11b 11 ± 4b 
rCIIpic 19 ± 12b 4 ± 1b 
a

Groups of 12 mice were administered orally either 0.01 M acetic acid, human tissue-derived CII, rCIIbac, or rCIIpic. The collagens were dissolved in 0.01 M acetic acid and administered four times per week for 2 wk for a total of eight doses. Ten micrograms was administered with each dose so that they received a total of 80 μg of protein. Mice were immunized with CII 3 days after the last dose. The incidence of arthritis is reported at 6 and 8 wk after immunization, and statistics are reported using Fisher’s exact test. Abs represent the mean levels per group (expressed in units) against native hCII using sera collected 6 or 8 wk after immunization. Statistics are reported using Student’s t test comparing each group to the control group.

b

, p ≤ 0.05.

It is well established that the MHC molecules bind small peptide sequences for the purpose of stimulating T cells in an Ag-specific manner, and the vast majority of studies of the structure-function relationship between these peptides and the MHC and TCR have focused on the peptide’s primary amino acid sequence. Yet most proteins expressed by eukaryotes have a significant amount of posttranslational modification, e.g., the addition of hydroxyl and glycosyl groups that occurs during the biosynthesis of CII. To enable us to study the role of posttranslational modifications of CII, we have established expressions systems for the production of native CII with varying degrees of hydroxylation and glycosylation. rhCII can now be produced in several different expression systems by simultaneous coexpression with recombinant human prolyl 4-hydroxylase, so that the proteins expressed will maintain the triple helical structure of the native CII molecule, but have differing degrees of other posttranslational modifications. CII produced by the baculovirus system in insect cells, rhCIIbac, is partially glycosylated, because the insect cells contain some but not adequate levels of lysyl hydroxylase and glycosyl transferases. In contrast, rhCII produced in P. pastoris, rhCIIpic, does not contain any Hyl or its glycosides, because the yeast cell is totally lacking the enzyme system. Our data reveal that both the arthritogenicity of the immunizing collagen and the potency of the suppression induced when collagen is used as a tolerogen correlate with the content of Hyl glycosides of the collagen used.

The in vitro T cell proliferative and cytokine responses also correlate with the degree of posttranslational modification present on the immunizing collagen. Although it has been established that many TCR are capable of recognizing synthetic peptides lacking posttranslational modifications, recently it has also become clear that TCR specific for modified or glycosylated peptides do exist and may play important roles in shaping the immune response. DBA/1 mice immunized with periodate-treated CII had a significant decrease in the incidence and severity of arthritis (3), possibly due to the periodate destruction of the sugar residues. Corthay et al. (5), using a panel of synthetic peptides containing various posttranslational modifications (e.g., hydroxylation of Lys264 or Lys270 and glycosylation of OH-Lys264 or OH-Lys270) of the immunodominant T cell determinant of collagen, demonstrated that a subset of T cells recognized specifically the OH-Lys264 glycosylated form of the collagen peptide in vitro. Interestingly, glycosylation of OH-Lys270 had no effect on T cell recognition of the determinant. These data suggest that glycosylation of CII at specific lysine residues might be important in the immune response to CII (5, 10).

Posttranslational modifications of proteins other than collagen have been reported to play a significant role in determining antigenicity. Jensen et al. (11) reported that T cells from CBA/J mice recognize a hemoglobin-derived decapeptide Hb67–76 only when it contains the tumor-associated carbohydrate at position 72, wheras the peptide remained nonimmunogenic if unglycosylated. Reitter et al. (12) have recently described conserved N-linked glycosylation sites of the HIV envelope glycoprotein that limit the resulting immune response to the virus. Similarly, Abs against African trypanosomes appear to be directed primarily against the variable surface glycoprotein portion of the parasite, and the glycosylation sites appear to play a major role in host defense (13). Dimitroff et al. (14) report that a specialized glycoform of P-selectin glycoprotein ligand confers the specificity of human T cells to enter dermal tissue and modulates lymphocyte migration to skin.

It has long been known that native CII molecules are immunologically more potent than chemically synthesized peptides. We now show that the potency of the immune response against the CII molecule correlates with the extent of the lysine hydroxylation and glycosylation. These modifications, which are present on some amino acids in the native state, are not present in the synthetic peptides ordinarily developed to test immune recognition. A better understanding of the importance of glycosylation and posttranslational modifications for both the induction of autoimmune arthritis and preventing the development of CIA will help us in developing therapeutic reagents for autoimmune arthritis.

1

This work was supported, in part, by U.S. Public Health Service Grant AR-39166, U.S. Public Health Service Grants AR-43589 and AR-45987, grants from the Academy of Finland, and program-directed funds from Department of Veterans Affairs and the Arthritis Foundation.

3

Abbreviations used in this paper: CII, type II collagen; CIA, collagen-induced arthritis; Hyl, hydroxylysine; GH, galactosylhydroxyline; GGH, glucosylgalactosylhydroxyline; h, human; bac, baculovirus; pic, Pichia pastoris; td, tissue derived.

1
Terato, K., K. A. Hasty, R. A. Reife, M. A. Cremer, A. H. Kang, J. M. Stuart.
1992
. Induction of arthritis with monoclonal antibodies to collagen.
J. Immunol.
148
:
2103
.
2
Stuart, J. M., F. J. Dixon.
1983
. Serum transfer of collagen-induced arthritis in mice.
J. Exp. Med.
158
:
378
.
3
Michaelsson, E., V. Malmstrom, S. Reis, A. Engstrom, H. Burkhardt, R. Holmdahl.
1994
. T cell recognition of carbohydrates on type II collagen.
J. Exp. Med.
180
:
745
.
4
Axford, J..
1997
. Proceedings of the 4th Jenner International Glycoimmunology Meeting.
Immunol. Today
18
:
511
.
5
Corthay, A., J. Backlund, J. Broddefalk, E. Michaelsson, T. J. Goldschmidt, J. Kihlberg, R. Holmdahl.
1998
. Epitope glycosylation plays a critical role for T cell recognition of type II collagen in collagen-induced arthritis.
Eur. J. Immunol.
28
:
2580
.
6
Stuart, J. M., M. A. Cremer, S. N. Dixit, A. H. Kang, A. S. Townes.
1979
. Collagen-induced arthritis in rats: comparison of vitreous and cartilage-derived collagens.
Arthritis Rheum.
22
:
347
.
7
Nokelainen, M., T. Helaakoski, J. Myllyharju, H. Notbohm, T. Pihlajaniemi, P. P. Fietzek, K. I. Kivirikko.
1998
. Expression and characterization of recombinant human type II collagens with low and high contents of hydroxylysine and its glycosylated forms.
Matrix Biol.
16
:
329
.
8
Myllyharju, J., M. Nokelainen, A. Vuorela, K. I. Kivirikko.
2000
. Expression of recombinant human type I-III collagens in the yeast Pichia pastoris.
Biochem. Soc. Trans.
28
:
353
.
9
Rosloniec, E. F., A. H. Kang, L. K. Myers, M. A. Cremer.
1997
. Collagen-induced arthritis. R. Coico, and E. Shevach, eds.
Current Protocols in Immunology
15.5.1
. Wiley, New York.
10
Backlund, J., S. Carlsen, T. Hoger, B. Holm, L. Fugger, J. Kihlberg, H. Burkhardt, R. Holmdahl.
2002
. Predominant selection of T cells specific for the glycosylated collagen type II epitope (263–270) in humanized transgenic mice and in rheumatoid arthritis.
Proc. Natl. Acad. Sci. USA
99
:
9960
.
11
Jensen, T., L. Galli-Stampino, S. Mouritsen, K. Frische, S. Peters, M. Meldal, O. Werdelin.
1996
. T cell recognition of Tn-glycosylated peptide antigens.
Eur. J. Immunol.
26
:
1342
.
12
Reitter, J. N., R. E. Means, R. C. Desrosiers.
1998
. A role for carbohydrates in immune evasion in AIDS.
Nat. Med.
4
:
679
.
13
Muller, N., J. M. Mansfield, T. Seebeck.
1996
. Trypanosome variant surface glycoproteins are recognized by self-reactive antibodies in uninfected hosts.
Infect. Immun.
64
:
4593
.
14
Dimitroff, C. J., R. J. Bernacki, R. Sackstein.
2003
. Glycosylation-dependent inhibition of cutaneous lymphocyte-associated antigen expression: implications in modulating lymphocyte migration to skin.
Blood
101
:
602
.
15
Miller, E. J., S. Gay.
1982
. Collagen: an overview.
Methods Enzymol.
82
:
3
.
16
Yang, C. L., H. Rui, S. Mosler, H. Notbohm, A. Sawaryn, P. K. Muller.
1993
. Collagen II from articular cartilage and annulus fibrosus: structural and functional implication of tissue specific posttranslational modifications of collagen molecules.
Eur. J. Biochem.
213
:
1297
.