Histochemical Journal, 11 (1979), 73-82

Specific immunohistochemical localization of Type I collagen in porcine periodontal tissues using the peroxidase-labelled antibody technique L. G. RAO, H. M. WANG, R. K A L L I E C H A R A N , J. N. M. HEERSCHE a n d J. SODEK MRC Group in Periodontal Physiology, Faculty of Dentistry, University of Toronto, Toronto, Ontario M5S 1AS, Canada

Received 23 February 1978 and in revised form 2 June 1978

Synopsis. Antibody against Type I coUagen was raised in rabbits and purified by immunoadsorption on Sepharose-conjugated Types I and III collagen. The crossreactivity of purified antibody to Type III collagen was found to be less than 0.5% by passive haemagglutination and less than 1.5% by radioimmunoassay. When paraffin sections of fixed and decalcified pig molars were incubated with purified antibody to Type I collagen, varying degrees of staining were observed in the ligament, gingiva, bone and cementum. The periodontal ligament adjacent to bone was more widely stained than that adjacent to cementum in some regions, whereas in others, no difference in staining could be discerned between the two halves of the ligament. The lamina propria of gingiva was stained, and this appeared to be most intense in the vicinity of the overlying epithelium. The fibrous component in the endosteal spaces, the dentine and the extracellular coronal elements in the pulp were generally stained. The impression obtained from the staining pattern is that Type I collagen is not restricted to particular regions of the periodontal ligament or the lamina propria of the gingiva.

Introduction

The pedodontium which supports the teeth is an organ composed of two soft connective tissues, the periodontal ligament and gingiva, and two mineralized connective tissues, the alveolar bone and cementum. The major collagen in all of these tissues is Type I, but significant amounts of Type III collagen have been found in periodontal ligament (Butler et al., 1975) and in gingiva (Ballard and Butler, 1974). However, nothing is known about the distribution of the two different types of collagen in periodontal ligament and gingiva. In the present communication, we report the preparation and purification of an antibody specific for Type I collagen and its use 9 1979 Chapman and Hall Ltd. Printed in Great Britain.

74

Rao, Wang, Kalliecharan, Heersche and Sodek

to localize Type I collagen in periodontal tissues and some dental tissues using the peroxidase-labelled antibody technique of Nakane et al. (1968). Materials and methods

Preparation and purification of Type 1collagen Type I collagen was purified from acetic acid-soluble protein from porcine gingiva. The pig mandibles were obtained from a local slaughter house (Quality Packers, Toronto, Ontario, Canada) and stored on ice until use. Attached gingiva associated with primary first, second and third mandibular molars was carefully dissected and as much of the epithelium as possible removed with a scalpel. Pooled tissue was homogenized in 0.45 M sodium chloride solution using a Polytron homogenizer (Eppendorf, Brinkman Instruments, Toronto, Ontario, Canada). The tissue suspension was kept in an ice bath and subjected to short high-speed bursts on the homogenizer until the tissue was well dispersed. All subsequent procedures involving collagen purification were also performed at a temperature below 40C. The homogenate was extracted three times with a 10-fold volume of 0.45 M sodium chloride solution to remove neutral salt soluble collagen and other non-collagenous proteins. After each extraction the tissue residue was separated from the extracting solution by centrifugation at 50 000 g for 1 h in an IEC B60 preparative ultracentrifuge (Damon International Equipment Co., Needham Heights, Mass., U.S.A.). To obtain acid-soluble collagen the tissue residue was extracted with 0.5 M acetic acid (Piez et al., 1963). A 10-fold volume of extraction solution was used and the products of three separate extractions pooled. The collagenous proteins were precipitated from this solution by dialysis against 0.02 M disodium monohydrogen phosphate solution. The solubilization-precipitation procedure was repeated until pure collagen was obtained (usually after 3 - 4 cycles) as judged by amino acid and sodium dodecyl sulphate(SDS)-polyacrylamide gel electro~ phoresis analysis. Preparation and purification of Type 111collagen Type III collagen was prepared from the residue of gingival tissue remaining after the neutral salt and acetic acid extraction described above. The purification involved a preliminary pepsin digestion followed by differential salt fractionation of native collagen molecules as described by Chung & Miller (1974). Final purification was achieved by gel filtration on Biogel A 15 M (Biorad Labs., Richmond, California,

U.S.A.). Preparation of antisera Rabbits were immunized with Type I collagen using the procedure of Vaitukatis et al., 1971). A 1 mg]ml solution of Type I collagen in 0.05% acetic acid (Pontz etal., 1970) was emulsified directly with an equal volume of Freund's complete adjuvant containing 5 mg dried tubercle bacilli (Difco Laboratories, Detroit, Michigan, U.S.A.). The emulsion was injected intradermally at approximately 30 sites on the shaved lower back of the rabbits (1 ml per rabbit). At a separate site, 0.5 ml of crude BorteUa pertussis vaccine (Connaught Laboratories, Ltd, Toronto, Ontario, Canada) was administered subcutaneously. Seven and nine weeks after initial injections, intradermal

Localization of collagen in the periodontium

75

booster injections with emulsion (1 mg/ml collagen in 0.05% acetic acid plus an equal volume of Freund's incomplete adjuvant) were administered. Blood was collected and antisera prepared at 10 weeks and 12 weeks after the initial immunization.

Preparation of immunoabsorbent column Collagen was coupled to CNBr-Sepharose 4B (Pharmacia, Uppsala, Sweden) following the manufacturer's instructions except that Tris-HC1 buffered saline at pH 7.4 was used as the solvent for the coupling reaction. The acid-soluble collagen was dissolved in 0.5 M acetic acid and dialyzed overnight at 4~ against the Tris-HCl-saline buffer, while the Type III collagen was dissolved directly in the buffer. The reaction was allowed to continue for 18h at 4~ The amount of collagen conjugated to the Sepharose 4B was estimated by anaino acid analysis in a Beckrna~ 121 M amino acid analyzer (Beckman Instruments, Palo Alto, California, U.S.A.). Purification of antibody to Type I collagen by affinity chromatography All subsequent steps were carried out in the cold. The antiserum against acid-soluble Type I collagen was passed through inter-connected columns of Type III collagen. Sepharose 4B and Type 1 collagen-Sepharose 4B and washed with phosphate-buffered saline (PBS) until the optical absorbance of the effluent at 280 nm was less than 0.01. The specific anti-Type I collagen antibody was then eluted from the Type I collagenSepharose 4B column with a 0.1 M acetic acid solution containing 1 M sodium chloride. The elution profile was determined by measuring the optical density at 280 nm. The fractions containing the antibody were pooled, immediately adjusted to pH 7.0, and then dialyzed overnight against 10 times diluted PBS, pH 7.4 (Dulbecco & Vogt, 1954). The final volume was reduced to one-tenth of the original volume of antiserum loaded on the column by using Amicon ultraffltration apparatus (Amicon Co, Lexington, Mass, U.S.A.) and stored at -20~ in 0.1-0.2 ml portions until needed. Normal rabbit serum (NRS) was fractionated in the same manner to serve as control for immunohlstochemical staining. The material eluted from the Type I collagen affinity column with the 1 M sodium chloride solution in 0.1 M acetic acid was designated 'NRS fraction'. Passive haemagglutination and haemagglutination-inhib ition assay The titres of antisera and purified antibodies were determined using the passive haemagglutination assay of Campbell et al. (1970). The cro~s-reactivity of antiserum against Type III collagen was determined by haemagglutination inhibition test (Steffen et al., 1968). For inhibition, 50 pg of Type I or Type III collagen dissolved at a concentration of 0.12% in 0.1% of acetic acid solution and neutralized with 0.9 M NaCI, 0.15 M Na2 HPO4 buffer solution (final cone. 0.1%) was pre-incubated with each tube of the progressively two-fold diluted antiserum or antibody for 1 h before the Type I collagen coated sheep red blood cells were added. A 2% solution of NRS was used as a control and diluent in these assays. Radioimmunossay of Type I collagen Type I collagen was radioiodinated by using lactoperoxidase-sepharose 4B as catalyst in the presence of hydrogen peroxide (H202) as described by David (I972). The

76

Rao, Wang, Kalliecharan, Heersche and Sodek

radioactivity of the preparation obtained was 1 x 106 Cipm/mg. The radioimmunoassay procedure was performed exactly as described by Adelmann et al. (1973).

Preparation of tissue sections Mandibles from six-month-old pigs were transported to the laboratory on ice to avoid tissue degradation. The primary molars surrounded by periodontium and part of the jaw bone were dissected with the aid of a dental air turbine drill (Midwest American Co) and immediately fixed in situ in 10% buffered formalin (pH 7.2) for a period of one week at room temperature. After washing under running tap water to remove the fixative, the tissues were transferred for decalcification to a 250 ml beaker containing 12.5% EDTA (Na2) adjusted to pH7.4 with sodium hydroxide. The decalcifying solution was stirred continuously at room temperature and changed daily for a period of one month. After decalcification, the tissues were processed according to standard histological procedures to prepare paraffin sections about 7/am thick. Immunoperoxidase method Tissue sections on microscope slides were deparaffmized in two changes of xylene, hydrated in decreasing concentrations of ethanol, rinsed in running tap water and stored in PBS. To remove non-specific staining, the tissues were successively incubated in the following solutions: (a) 0.3% hydrogen peroxide in methanol for 30 min to remove endogenous peroxidase (Mason & Taylor, 1975); (b) 11 mM Tris-saline buffer pH 9.4, twice for 15 min, in order to alter tissue charges introduced as a result of formalin fixation (Cawley, t976); and (c) a few drops of normal goat serum (NGS) for 30 rain (Moriarty, 1973). Paraffin sections of porcine ear elastic cartilage were processed by the same method described above with one additional step between steps (b) and (c): the sections were incubated with a few drops of either 2% hyaluronidase (from ovine testes; Sigma Chemical Co, St. Louis, Montana) in PBS or in PBS alone for 20 min. Afterwards, the sections were washed twice for 10 min in PBS. Thereafter, the tissues were stained according to the indirect immunoperoxidase method of Nakane (1968) as follows. The NGS was drained off and the tissues were incubated for 30 min in a few drops of undiluted purified anti-Type I collagen antibody, or in antiserum to Type I collagen diluted 1:10 with NGS to avoid further non-specific tissue staining, and further diluted 1:20 with PBS. Controls were incubated with either undiluted NRS fraction or NRS diluted in the same manner as the antiserum. After incubation, the tissues were washed twice for 10 min with PBS and the incubated for 30 rain in goat anti-rabbit immunoglob~tin-peroxidase (GARP) diluted 1 : 10 with NGS and finally 1:100 with PBS. The tissues were washed overnight with PBS. All these incubations were carried out in humidified glass petri dishes provided with slide holders. After overnight washing, the slides were then transferred to co_plin jars containing 27 ml of a solution of 20 mg 3,3'-diaminobenzidine (DAB) (Sigma)in 100 ml 0.05 M ammonium acetatecitrate buffer, pH 5.0, prepared by stirring for 4 h in the dark (Weir et al., 1974). After 2 min, 3 ml 0.05% hydrogen peroxide was added with shaking and the reaction allowed to proceed in the dark for another 7 min. The solution was decanted, whereafter the slides were washed in running tap water for 5 min, dehydrated and mounted in Permount. Stained sections were observed microscopically with a Reichert Univar microscope using phase contrast or Nomarksi interference optics.

Localization of collagen in the periodontium

77

Results

Purification of antigen and antibodies Amino acid analysis and sodium dodecyl'sulphate potyacrylamide gel electrophoresis showed that the Type I collagen prepared from acetic acid soluble porcine gingival tissue was pure. The proportions of hydroxylysine, 4-hydroxyproline and proline were similar to values expected for Type I collagen (Piez et al., 1963), while cysteine, which is present in Type III collagen, could not be detected. Sodium dodecylsulphate polyacrylamide gel electrophoresis analysis showed the expected a t , a2, 13 and 3' chains~ but no other protein. Antibodies to this collagen preparation were successfully raised in two of three rabbits immunized. The antisera from these rabbits gave titres of 1/640 and 1/80 in the passive haemagglutination assay and showed titres with Type III collagen of less than 1/2. When the antibody with a titre of 1/640 was purified by immunoabsorption on Type III and Type I collagen affinity columns and concentrated by Diaflo, the purified antibody had a titre of 1/512 and the titae to Type tII collagen in the passive haemagglutination assay was again less than 1/2. In haemagglutination inhibition tests, the anti-type I collagen antibody (titre 1/512) was completely inhibited by 50/~g of Type I collagen, however, the titre was only reduced to 1/256 by same amount of Type III collagen. The cross-reactivity of antibody to Type I collagen against Type III collagen was, therefore, calculated to be less than 0.5%. The purified anti-Type I collagen antibody was also tested for binding to 12 s I-labelled Type I collagen and the cross-reactivity with Type III collagen determined by radioimmunoassay. The results, given in Fig. I, demonstrate that the inhibition of labelled Type I collagen by unlabelled Type III collagen is less than 1.5%. Localization of collagen Type I in periodontal tissues The results of using the unpurified anti-serum containing antibody to Type I collagen and the purified antibody to Type I collagen were similar. We wiil describe the results 100 90

402 . / 30

20

0.2

0.4

0,6 0.81.0

2

AMOUNT OF INHIBITOR

4

6

8 10

20

(PICOMOLES)

Figure 1. Inhibition of the reaction of [1 2s I] type i collagen and its antibody by unlabelled type I (o) and type III (m) collagen. The specific activity of the type I collagen was 1 x 106 cpm/mg. Antibody dilution used was 1:200. Each point represents the average of duplicate determinations.

78

Rao, Wang, Kalliecharan, Heersehe and Sodek

obtained using the purified antibody. Bone, ligament, cementum (Fig. 2) and dentine were not stained when tissues were incubated as controls with the NRS fraction. However, erythrocytes were lightly stained in spite of the removal of endogenous peroxidase b y incubation in hydrogen peroxide-methanol solution. When the tissue was incubated with purified antibody to Type I collagen, varying degrees of staining were observed in the ligament, gingiva, bone, cementum and dentine. The periodontal ligament adjacent to bone was often more widely stained than that adjacent to cementum, but the extent of the difference between the two halves of the ligament varied (Figs. 3, 4 & 5). In other instances, no difference could

Figure 2. Control section from the mid-root region of the mesial aspect of the mesial root of a first mandibular primary molar, incubated in the NRS fraction and GARP (goat anti-rabbit immunoglobulin peroxidase) and stained with the DAB reaction. Note the absence of staining; the colour is the result of microscope iUumination. B = bone; P = periodontal ligament; C = cementum. Nomarski optics, x 770 Figure 3. The mid-root region of the mesial aspect of the mesial root of a first mandibular primary molar incubated in purified Type I collagen antibody and GARP and stained with the DAB reaction. The ligament on the bone side is a little more intensely stained than that on the cementum side. P = periodontal ligament; C = cementum; B = alveolar bone; E = endosteal space; R = erythrocytes. Nomarski optics. x 770 Figure 4. The mid-root region of the distal aspect of the distal root of a mandibular first primary molar treated as described in Fig. 3. Note that the periodontal ligament (P), cementum (C), dentine (D) and alveolar bone (B) are all stained, but that little Type I collagen has been stained in the ligament on the cementum side. Some sites in the bone are less intensely stained than elsewhere (arrow). Nomarski optics. X 330 Figure 5. The periodontal ligament on the mesial aspect of the mesial root in the mid-root region of a mandibular first primary molar treated as described in Fig. 3. Note that collagen fibres in the bone half of the periodontal ligament (arrow) are more widely stained than are those in the cementum half. P = periodontal ligament; C = cementum. Nomarsld optics, x 1200 Figure 6. The apical region of the mesial root of mandibular first primary molar treated as described in Fig. 3. Note that the staining of Type I collagen fibres is fairly uniform throughout the periodontal ligament (P). C = cementum; B = alveolar bone; V = blood vessel, Nomarski optics. X 1200 Figure 7. The bifurcation in a first mandibular primary molar treated as described in Fig. 3. Note that the staining of Type I collagen fibres is fairly uniform throughout the periodontal ligament (P). C = cementum; B = alveolar bone; D = dentine; V = blood vessel Nomarski optics, x 770 Figure 8. An endosteal space (E) in alveolar bone (B) treated as described in Fig. 3. Note stained Type I collagen fibres and also the blood vessels (V) in the space. Nomarski optics, x 1200 Figure 9. Attached gingiva from the buccal surface of alveolar bone treated as described in Fig. 3. Note the staining of Type I collagen in the lamina propria (L), and its increased intensity in the vicinity of the epithelium (1"). The epithelium (E), particularly in the keratinized layers, is also stained (