Background Information

CII Preparation

CII was first extracted and purified from chicken cartilage by Miller and Matukas (1969) and was characterized as a homotrimer of a1(II) polypeptide chains with a molecular weight of 285 kD (Miller, 1971). Its biosynthetic precursor, type II procollagen, has a molecular weight of 450 kD and contains globular domains, at both the amino and carboxyl termini, which are termed propeptides. After secretion from the cell, but prior to the formation of CII fibrils in the cartilage, the amino and carboxyl propeptides are removed by the action of specific procollagen peptidases, leaving behind only short telopeptides attached to the a-helix. As is typical with most collagens, the amino acid sequence of the a-helical region is composed of repeating Gly-X-Y sequences in which X or Y is frequently a proline. The structure is also unusual in that many of the prolines and lysines in the Y positions are post-translationally hydroxylated. CII is restricted to hyaline cartilage and the vitreous body of the eye, and is normally absent in all other tissues.

Like other interstitial collagens, the long triple-helical segment of CII is highly resistant to digestion by enzymes , with the exception of collagenase. This property is advantageous in that it allows the use of enzymes--e.g., pepsin--to extract collagen from complex tissues. The pepsin digestion of proteoglycan-depleted cartilage greatly increases the yield of CII by freeing it from the remaining noncollagenous matrix proteins. CII isolated by enzyme digestion, however, differs slightly from CII found in situ in that pepsin also cleaves the short telopeptides found at the amino and carboxyl termini of the collagen molecule. Fortunately, this modification affects neither the stability of CII nor its ability to induce autoimmune arthritis. CII, like all proteins, is susceptible to denaturation, especially by temperatures near or above 38°C. Denaturation not only adversely affects the arthritogenic capacity of CII, but also renders the individual a chains of CII highly susceptible to rapid degradation by a host of proteases.

CIA Mouse Model

CIA in the mouse was first described by Courtenay et al. (1980), and the genetics controlling this susceptibility have been analyzed in detail by Wooley et al. (1981). The following are the major characteristics of the murine CIA model.

  1. Disease susceptibility is strongly linked to the class II molecules of the MHC, and specifically to the I-Aq and I-Ar alleles.
  2. CIA can be induced in susceptible mouse strains by immunization with a variety of type II collagens from different species; however, the susceptibility to these collagens differs between H-2q and H-2r strains of mice.
  3. Although stimulation of collagen-specific T cells is required to induce arthritis, high levels of circulating antibody to CII are strongly associated with the development of arthritis.
  4. CIA can be passively transferred to both syngeneic and allogeneic naive mice with purified polyclonal antibodies or a mixture of monoclonal antibodies that differ in epitope specificity for CII. Following administration, these antibodies can be found bound to CII in hyaline cartilage of inflamed joints along with deposits of the complement component C3.

Although several CIA-susceptible mouse strains have been described (Table 15.5.4), DBA/1 and B10.Q (H-2q strains), and B10.RIII (H-2r) are the most commonly used commercially available strains susceptible to CIA.

Table 15.5.4 Susceptiblity of Various H-2q and H-2r Strains to CIA

Strain I-A
DBA/1LacJ q - +++
DBA/1J q - ++
B10.RIII r r +++
B10.Q q - +++
B10.bQBr q - +++
BUB q - +++
SWR q - -
NFR/N q - ++
B10.CAS2 w17 - +

Susceptibility of H-2r mice differs significantly from H-2q in terms of the species of CII used for immunization. H-2r mice are susceptible to CIA when immunized with bovine or porcine CII but not with chick or human CII. Conversely, H-2q mice are susceptible to CIA when immunized with bovine, chick, and human, but not porcine CII. Unlike the rat model (discussed below), arthritis can also be induced by immunization with either purified a1(II) chains or large cyanogen bromide fragments of CII--e.g., CB11, CII(124-402) for H-2q (Terato et al., 1985; Brand et al., 1994) and CB8, CII 403-551 for H-2r (Myers et al., 1995). However, the incidence and severity of arthritis are lower in comparison to that induced by immunization with native CII, and the time of onset is delayed. Other less common H-2q strains have also been shown to be susceptible to CIA, with the exception of the SWR strain, which has a complement deficiency (Watson and Townes, 1985). Two of the CIA-susceptible H-2q strains in Table 15.5.4 (B10.bQBr and BUB) have large genomic deletions of T cell-receptor variable genes. Male mice are frequently preferred for CIA studies, as the incidence of arthritis is somewhat higher in male than in female mice. Although mice 8 weeks of age are generally preferred for CIA experiments, DBA/1 mice remain highly susceptible to arthritis at least through the age of 6 months.

Both B and T cell responses can be measured in mice following immunization with CII. As with most antigen systems, the T cell responses with CII peak around day 10 to 12, but can be detected for several weeks following immunization. Serum antibody production peaks as the incidence and severity of arthritis peaks, and, within a given strain, the levels of antibody correlate well with the presence or absence of arthritis. Among different susceptible and nonsusceptible strains, antibody levels can vary widely. Some nonsusceptible strains produce significant amounts of CII-specific antibody, yet fail to develop arthritis (Wooley et al., 1981).

CIA Rat Model

CIA as an experimental model of autoimmune arthritis was first described by Trentham et al. (1977) using an inbred rat strain. The following are the major characteristics of CIA in the rat.

  1. There is a larger number of highly inbred strains of rat that are susceptible to CIA (Table 15.5.5) than there are susceptible mouse strains. Many of these strains differ in their class II RT1 genes, thus providing a variety of genetics for studying the autoimmune response.
  2. Rats not only develop arthritis after immunization with CII, but also after immunization with type XI collagen, mycobacteria, and streptococcal cell walls. This feature provides useful controls and comparisons with the autoimmune arthritis induced by CII.
  3. Arthritis can be induced easily in rats with incomplete Freund's adjuvant (IFA), obviating the use of potent immunoadjuvants such as mycobacteria and pertussis which produce immune response to antigens other than CII.
  4. After immunization, high-responder rats develop arthritis more quickly (in 9 to 16 days) than mice (in 24 to 42 days), thereby shortening the length of the experiments and the amount of time the animals must be maintained.
  5. The rat's larger size makes many procedures easier (or possible) to perform--e.g., thoracic-duct canulation, implantation of long-standing venous access devices, recovery of cells from arthritic joints, and accurate, reproducible measurement of paw volume by Hg manometry (see Support Protocol 5).

The three most important variables in the induction of CIA in rats are: (1) strain selection, (2) animal health, and (3) the effectiveness of the CII used for immunization. Conveniently, a number of susceptible strains are available from commercial vendors that maintain required health standards. CIA has been reproducibly elicited in eight strains of inbred rats differing at RT1 (Table 15.5.5) and in two strains of commercially available outbred rats. The information summarized in Table 15.5.5 should be interpreted with the understanding that the strains shown were obtained from different sources, housed under different conditions for various lengths of time, immunized with different doses and preparations of CII, and in some instances given a single injection but in others boosted one to five weeks later. Although it is unlikely that these variables would affect the susceptibility of high-responder strains, they might have a significant impact on moderate- to low-responder strains. Notably, two inbred strains of rats, F344 and MAXX (RT1lv1 and RT1n, respectively), have proven highly resistant to CIA following immunization with different species sources of CII, thus they are not included in Table 15.5.5.

Like many animal models, bacterial and viral infections can adversely affect the incidence and severity of CIA. Hence, only specific pathogen-free rats (see UNIT 1.1) should be used to obtain reproducible results. The best documented infectious agent impairing CIA in the rat is Mycoplasma pneumoniae. High-responder LEW rats appearing deceptively healthy are quite resistant to CIA when infected with M. pneumoniae (Taurog et al., 1984). Likewise, active infection can impair lymphocyte proliferation in long and short-term in vitro studies. Regular monitoring of sentinel animal health and serology is mandatory when frequent or long-term studies of CIA are undertaken.

CIA first appears as early as 8 days after immunization in high-responder BB and WF rats and 12 to 16 days after immunization in outbred Wistar rats. Onset, however, can occur as late as 30 to 60 days after immunization in low- to moderate-responder strains. In general, gender makes little difference in arthritis susceptibility in rats, with possible exception of the BUF strain, in which males might be more susceptible than females (Cremer et al., 1994a). Young adult rats weighing 125 to 200 g provide the most consistent results. High-responder rats such as BB often remain arthritis-susceptible up to 3 or 4 months of age. The same can not be said with confidence regarding other inbred strains. Environmental conditions are important both for animal well-being and ensuring consistent results. Exposure to regular, intense stress can adversely affect CIA in rats (Rogers et al., 1980); however, gentle daily handling and weekly collections of £1 ml of blood do not noticeably affect the incidence or course of arthritis.

Arthritis has been induced in high-responder strains of rats with CII purified from the cartilage of a number of species--e.g., human, chick, bovine, porcine, rabbit, guinea pig, and deer. With the unique exception of the DA rat (Griffiths et al., 1993), most strains respond poorly to rat CII compared to other mammalian and avian products (Griffiths, 1988; Griffiths et al., 1992; Cremer et al., 1994a). Arthritic responses to chick and bovine CII are often equivalent, with the notable exception of ACI and BN rats where disparate responses are seen. Studies by Griffiths et al. (1992) have also shown that the arthritis-producing potential of porcine CII is essentially identical to that of bovine and human CII.

Whereas the species of collagen plays a relatively minor role in determining arthritogenicity in the rat, the tertiary structure of CII used for immunization is of critical importance. In sharp contrast to the mouse, where arthritis can be induced at a respectable incidence with highly purified a1(II) chains or cyanogen bromide fragments of the a1(II) chain, CII must be maintained in its native triple-helical conformation during preparation in order to induce arthritis in the rat. For this reason, CII solutions must be kept chilled at temperatures below 4°C to minimize the risk of denaturation.

The amount of CII used to induce arthritis varies broadly among investigators and to some degree depends on the strain being immunized. Doses ranging between 75 and 1000 ug per rat have been described as effective, with intermediate doses of 150 to 300 ug being used most commonly for high-responder inbred strains and outbred strains such as Wistar and Sprague-Dawley. Higher doses are unnecessary in susceptible strains and often ineffective in resistant ones. The use of a booster injection given 7 days after priming is often helpful in producing a more uniform time of onset and may also increase the incidence of disease in moderate- and poor-responder strains.

In the mouse, the H-2q (DBA/1LacJ, CIA-susceptible) proliferative response is dominated by recognition of the CII(260-270) determinant (Brand et al., 1994), whereas IFN-g production can be detected to both CII(260-270) and CII(181-210) (Myers et al., 1989, 1992, 1993). The dominant response in CIA-susceptible H-2r (B10.RIII) mice is directed to CII(610-618), although a second T cell determinant is also present in CB8, CII(442-456) (Miyahara et al., 1995; Myers et al., 1995). In the rat, T cell proliferative responses to native and denatured CII, as well as to collagen peptides (Cremer et al., 1994b), can also be measured. Collagen-specific antibody levels to both the immunogen and the autoantigen correlate very well with the presence or absence of arthritis within susceptible strains, and appear to be a major contributor to the overall pathology of the experimental disease.

Critical Parameters and Troubleshooting

Collagen Preparation

Generally, the younger the animal from which fresh cartilage is obtained, the higher the yield of CII. However, when using fetal bovine cartilage as a source of CII, the fetus size should be at least 7 kg. Long-term freezing of cartilage prior to collagen extraction should be avoided as this will result in diminished yield. Extraction and solubilization of CII can be greatly facilitated by rendering the animals lathyritic, thereby preventing collagen cross-linking (Trelstad et al., 1970). Lathyrism is induced by administration of b-aminopropionitrile. Although this procedure requires added expense and time and is not practical for large animals, it is helpful when preparing collagen from rat or mouse cartilage. In addition, the use of a lathyrogen is essential in preparing CII from the Swarm rat chondrosarcoma (Smith et al., 1975; Miller and Rhodes, 1982). For animal use, 0.3% b-aminopropionitrile fumarate or maleate is simply added to the food or drinking water for 7 to 10 days. Higher doses or longer exposure offer little advantage and increase animal mortality and morbidity.

Once type I collagen contaminates a CII preparation, it is very difficult to remove it. All steps in this procedure must be performed at £4°C unless otherwise indicated. This enhances the solubility of the collagen, ensures the retention of native conformation in the solubilized collagen, and reduces the risk of bacterial growth.

Collagen solubilized in 10 mM acetic acid may be lyophilized and stored in a desiccator at -20° or 4°C. Lyophilized collagen tends to become progressively insoluble with time, for reasons that are not well understood. Care should be taken to prevent ambient moisture from condensing on collagen each time the desiccator is removed from the refrigerator. It is important to allow the desiccator to equilibrate to room temperature before opening it. Lyophilized collagen is best dissolved in 100 or 10 mM acetic acid by stirring overnight in the cold. It should be emphasized that the acetic acid solution must be prechilled to 4°C before use to dissolve the collagen. When it is necessary to work with collagen in a neutral buffer, the collagen should be first dissolved in acetic acid and then dialyzed into the desired buffer in the cold.

CII solubilized in acetic acid may be stored frozen at -70°C. Under these conditions, CII remains stable for periods up to 3 years and possibly longer. This mode of storage has the advantage of avoiding the time required for solubilizing lyophilized collagen. CII stored in this manner retains arthritogenicity for long periods (years).

High concentrations of native CII--up to 6 mg/ml--are achievable, but at the expense of a highly viscous solution. Acetic acid concentrations >100 mM should be avoided as they only cause pain on injection and irritation at the injection site. If for some reason CII must be used in a neutral buffer, such as PBS or TBS, it should be first dissolved in acetic acid and then diluted or dialyzed in to the desired buffer.

Induction of CIA

As with most immunizations, complete emulsification of CII in either CFA or IFA is essential to establishing a high incidence of autoimmune arthritis. Additionally, care must be taken to ensure that the CII is kept cold throughout the emulsification procedure, and once emulsified, that it is kept cold until used. The use of Mycobacteria-containing adjuvants is not recommended in the rat model because complete Freund's adjuvant alone can induce an "adjuvant arthritis" in the rat (Pearson, 1963). Once arthritis has developed in either model, it is helpful to add food to the bottom of the cages as severely arthritic mice have difficulty feeding from the cage top.

Anticipated Results

CII preparation

Although yields of CII will vary widely with the species and tissue source, generally 1% to 3% of the initial cartilage weight can be recovered as native CII.

Mouse and Rat CIA

The incidence of arthritis in both the rat and mouse model following immunization with native CII should be _80%. In the mouse model, an incidence of 25% to 50% is expected when using either fragments of CII--e.g., CB11 or CB8 or a1(II) chains to induce arthritis. The absence of arthritis or the induction of only mild disease at a low incidence can usually be attributed either to inadvertent denaturation of the collagen, a poor quality emulsion, or poor injection technique. Booster injections do little to correct these problems.

Serum collected at the peak of arthritis--21 days after immunization in the rat model and 35 to 42 days after immunization in the mouse model--should yield positive results by ELISA at minimally 1:1000 dilution. High-responder rat strains (BB and WF) and mouse strains (DBA/1 and B10.RIII) frequently yield positive ELISA results at serum dilutions ranging from 1:10,000 to 1:100,000.

Time Considerations

Extraction and purification of CII from 100 to 400 g of cartilage will take minimally 4 weeks. Smaller quantities of starting material will not appreciably shorten this time.

Center for Excellence Diseases of Connective Tissues Logo

Contact Us


Arnold Postlethwaite, Director
Phone: 901-448-4979

Principal Investigators:

David Brand, PhD
Monica L. Brown, DO
Hongbo Chi, PhD
Weikuan Gu, PhD
Karen Hasty, PhD
Andrew H. Kang, MD
Linda K. Myers, MD
Eugene Pinkhassik, PhD
Arnold Postlethwaite, MD
Edward Rosloniec, PhD
Andrzej Slominski, MD, PhD
John Stuart, MD
Ae-Kyeung Yi, PhD