- Email: [email protected]
Collagen type III in rotator cuff tears: An immunohistochemical study
Collagen type III in rotator cuH tears: An immunohistochemical study J. Kumagai, MO, H. K. Uhthoff, MO, K. Sarkar, MO, and
J. P. Murnaghan, MO, Ottawa, Ontario, Canada The presence and distribution of collagen type III, which represents newly formed fibers of early repair, were investigated in four surgically removed en bloc specimens of complete rotator cuH tears. The patients were two men and two women who ranged in age from 63 to 77 years. They had received preoperative, conservative treatment for a period varying from 3 months to 3 years. For the identification of the collagen immunohistochemical staining was performed by the peroxidase-antiperoxidase method with the use of monoclonal antibody against human collagen type III. Collagen type III was discernible throughout the tendon proper in all specimens. It often occurred in association with proliferating fibroblastic cells. Collagen type III was abundant only in the perivascular spaces of the much thickened bursal wall that extended over the margins of the tear. We concluded that the presence of collagen type III indicates the need for a continued attempt for repair at the site of the tear, including both bursal and tendinous tissues. (J SHOULDER ELBOW SURG 1992;1: 187-92.]
In
most elderly patients with a rotator cuff tear that develops either spontaneously or after minimal indirect trauma, an indefinite period of conservative management is usually followed. A considerable number of patients show satisfactory improvement/ and surgery is undertaken in only those who have persistent pain accompanied by a notable loss of function, especially of arm abduction. IS Thus the period of time elapsing between the diagnosis of the tear and the operation on the tear may vary widely from patient to patient and may often extend from months to years . However, preoperative duration of symptoms and the extent of the tear appear to be of little significance in the surgical outcome of cuff tecrs.' We have shown that repair reaction, which is represented by granulation tissue and apparently originates both in the overlying subacromial bursa and to a lesser degree in the torn tendon margins themselves, From the Division of O rthopaed ics, O ttowa General Hosp ital and Ottowa Civic Hosp ital, and the Departments of Surgery and Pathology, University of Ottowa. Support ed by the Medical Research Coun cil of Ca nada. Reprint requests: Hans K. Uhthoff, MD, FRCS(CL Professor and Cha irman, Deportment of Surgery, 501 Smyth Rd., Ottowa, Ontario Kl H 8L6.
32/1140100
is evident in all cases regardless of the duration of conservative moncqement." It has been demonstrated in experiments thot granulation tissue in the initial stages of wound heal ing is mostly comprised ofcollagen type III. This is the case even in tissues such as the skin and the tendon, which have collagen type I as the predominant component." 8. 14 It has been shown in the healing medial collateral ligament of the rabbit that the amount of collagen type III reaches its peak around the end of the second week of wound healing and then begins to decrease with the simultaneous rise of collagen type 1. 1 The collagen type III is replaced by collagen type I as the scar motures.'? This study was undertaken to determine whether variations in the length of preoperative periods of conservative management before the surgical procedure for complete rotator cuff tear would significantly alter the distribution of collagen type III at the margins of the tear and in the vicinity of the tear.
MATERIAL AND METHODS En bloc specimens of the supraspinatus tendon containing the complete tear were obtained during surgery from two men and two women. All ruptures occurred without a history of severe
187
188
Kumagai et al.
J. Shoulder Elbow Surg. July/August 1992
p
•
"'-"'· -M
l
A
mm
II rill' I Il'
II
1111 I
Figure 1 Art icular surface of typ ica l en bloc specimen of supraspinatus tendon w ith central tear . Sections for immuno histochemica l staining were token along dotted lines: A, anterior; P, posterior; M, medial; L, lateral. GT ind icates area of greater tuberosity.
trauma to the supraspinatus tendon, although in one 67-year-old woman the clinical evidence of tear was preceded by shoulder dislocation. The men were 63 and 66 years old, and the preoperative durations of symptoms were 9 and 6 months, respectively. The two women were 67 and 71 years old; their durations of symptoms were 3 months and 3 years, respectively. The en bloc specimens were fixed in 10% neutral buffered formalin . After fixation was performed sections were taken from anterior (toword the subscapularis tendon), posterior (toward the infraspinatus tendon), medial (toward the supraspinatus muscle), and lateral (toward the bony insertion) margins of the tear for paraffin embedding (Fig . 1). Serial sections 5 urn thick were placed on slides coated with a 2% solution of 3-aminopropyltriethoxysilane to prevent detachment of the tissue section. 11 Four age-matched controls were then obtained from cadavers that had no history of shoulder ailment, and these tissues were treated in the same manner. The preparation of the anti-human collagen type III monoclonal antibody that we used has been described in detail by Sakakibara et 01. 12 We obtained the antibody from Professor Akira Ooshima of Wakayama Prefectural Med ical
College of Japan. In this method collagen type III is extracted from human placenta and is digested with trypsin. The digest, combined with Freund's complete adjuvant, is injected into BALBI c mice . Splenic lymphocytes of the sensitized mice are then fused with mouse myeloma cells. The immunoglobulin G-producing cells are then cloned by limited dilution assay, and the immunoglobulin G reaction with the a-cha in of type III collagen is determined by immunoblotting . In our laboratory serial sections were first de paraffinized with xylene and rehydrated in graded ethanol. They were then treated with 0.25% trypsin (GIBCO, Life Technologies, Inc., Gaithersburg, Md.) that was dissolved in the 0.1 molll tris-HCI buffer (TBS) at pH 7.7 at 3]0 C for 60 minutes and washed in TBS at room temperature to stop all enzyme activity. For immunohistochemical staining we adopted the technique of Yulis and Lederis." All antisera were diluted in TBS conta ining 0.1 % Triton X-l00 (Sigma Diagnostics, St. louis, Mo.) and 0.7% carrageenan (Sigma) . The procedure was performed at room temperature. The sections were incubated overnight w ith the primary antibody acting against collagen type III in a dilution of 1 : 16,000 (0.71 j.Lgl ml) . Sections were
Volume 1 Number 4
Collagen type III in rotator cuff tears
189
-
Figure 2 Immunohistochemical staining of supraspinatus tendon from 67year-old man used as control shows fibrillar collagen type III in endotenon, which conta ins vascular channels . (Original magn ification x 200 .)
Figure 3 Immunohistochemical staining of bursal tissue at site of tear shows collagen type III mostly in relation to proliferating vascular channels. (Original magnification x 100.) then treated with the second antiserum, rabbit anti-mouse immunoglobulin (DAKO Corp., Carpenteria, Calif.) diluted 1 :25, applied for 60 minutes. Finally mouse peroxidase-antiperoxidase complex (DAKO), diluted 1 :250, was ap-
plied to the sections for 60 minutes. Between each incubation the slides were washed three times with TBS for 5 minutes. After the sections were incubated with peroxidase-antiperoxidase, they were treated for 20 minutes with
190 Kumagai et
01.
J. Shoulder Elbow Surg. July/August 7992
" ,
Figure 4 Immunoh istochemical staining from lateral aspect of rotator cuff tendon tear shows diffuse distribution of collagen type III. (Original magnification x 200.) 0.25% solution of diaminobenzidine tetrachloride in TBS containing 0.05% H2 0 2 • Controls included substitution of the primary antibody with TBS. The sections were counterstained with hematoxylin for improved visualization of nuclei. Routine sections were stained with hematoxylin-eosin.
RESULTS Control sections of the supraspinatus tendon from elderly individuals showed thinning of the fascicles and broadening of the endotenon space, which contained a number of prominent vascular channels. Collagen type III staining was found in the endotenon but was absent in the adjacent fascicles (Fig. 2). In all of the specimens of cuff tearing the usual architecture of the tendon, which is characterized by fascicles in parallel arrangement with the intervening endotenon, could not be recognized in the stump area. The torn margins showed fronds of tendinous tissue in close proximity to bursal tissue. There was a significant degree of fibrovascular proliferation in the thickened bursal wall. Cellular as well as vascular proliferation occurred in the tendon tissue, not only at the region of the tear but also close to the margins of surgical resection . Most of the proliferating cells in the tendon substance appeared to be plump mesenchymal cells. Inflammatory cells were not present. Cellular and vas-
eular proliferation was conspicuous in all of the representative sections taken from the medial, lateral, anterior, and posterior segments of the specimens . Collagen type III staining was predominant in areas around the vessels and contiguous with the mesenchymal cells. The fibrillary staining often lacked specific orientation, although its major concentration around the vessels in the bursal tissue gave it a concentric appearance (Fig. 3). In the tendon substance, vascular proliferation was comparatively much less than in the bursa, and the presence of collagen type III was more visible in relation to proliferating cells (Fig. 4). Staining in connection with the cells could also be seen in the fronds at the tear margins, which were devoid of vascular channels (Fig. 5).
DISCUSSION This study clearly showed that although collagen type III in control sections was segregated as fine fibrillary sheaths in the endotenon, forming a minor component of the tendinous tissue, it was greatly increased in the specimens of torn rotator cuff tendons. The normal tendon is primarily composed of collagen type I, and thus it is similar to skin and bone in collagen composition. The distribution and the relative quantity of collagen type III in normal tendons have been established. Its preferential presence in
Collagen type III in rotator cuff tears
Volume 7 Number 4
191
Figure 5 Tendinous fragments apparently devoid of vessels at site of tear show presence of collagen type III. Acellular fragment without staining on right is probably necrotic. (Original magn ification x 400.) the endotenon has been demonstrated by immunofluorescent staining in bovine tendons: Williams et ol." estimated that collagen type III constituted less than 1% of collagenous component in normal equine tendons. It is well known that collagen type III increases significantly during the healing of a wound." In the previously mentioned articles on equine tendon, Williams et ol." found that scar formation raised collagen type III content of the tendon to 20% to 30%. Significantly, the rise in type III collagen is most apparent in the early phase of wound healing. This was demonstrated by Gay et ol.," who implanted viscose cellulose sponge in surgical wounds of children and also cultured the granulation tissue of wounds from similar patients. From immunohistological and biochemical studies they concluded that primitive mesenchymal cells that are present in .the initial period of wound healing produce collagen type III, whereas fibroblasts, appearing at a later stage, are responsible for the production of type I. The initial prevalence of collagen type III has also been shown in rat dermal
wounds.' In this study the presence of collagen type III appeared to be dependent on both interstitial and vascular cells. The abundance of collagen type III around the vessels, especially in the encroaching bursal tissue overlying the torn tendon margins, was remarkable. In the vicinity of
the tear the tendinous tissue that had lost both its normal configuration of fascicles and its intervening endotenon showed collagen type III both in the tendon substance and around the vessels. The vascular proliferation in the tendon substance, however, was much less prevalent than in the bursal tissue, and collagen type III staining in the tendon was more conspicuous in relation to cells. The association of collagen type III staining with proliferating cells was even demonstrable at the cuff tear margins in the fragmented tissues, some of which were devoid of vessels (Fig. 5). The principal aim in surgical repairs of rotator cuff tears is to reestablish continuity so that the forces for movement can be transmitted from the muscles to the bones. During surgery an extensive debridement is often practiced not just for adequate visualization but because of the belief that debridement will remove apparently dead tissue and will ensure sufficient vascularity at the suture line to induce new matrix formation . We have shown that regardless of the elapsed time from the onset of symptoms to surgery, granulation tissue in the area of tear is adequate, and tr imming for the sake of improving vascularity is counterproductive.v ln the present study the presence of collagen type III, which represents the new matrix, clearly indicates that the attempt to heal at the margins of a tear persists over time. It appears that ol-
192 Kumagai et
01.
though the strongest reaction for collagen type III is in the vicinity of the tear, its presence extends to the edge of the surgical resection. There were no differences in staining among the medial, lateral, anterior, and posterior aspects of the specimens. The present study also supports our previous contention that the duration of time between the onset of symptoms and the eventual surgery is of little significance because the formation of new matrix collagen was unaffected by time. The ongoing presence of collagen type III probably indicates that little remodeling is occurring in the wound. Whether the persistence of collagen type III for months or even years at the margins of a rotator cuff tear is beneficial or detrimental to eventual repair when the tendon is reconstructed is a matter of speculation. The presence of collagen type III is believed to indicate tissue immaturity. There is a preponderance of collagen type III in fetal skin, but as the skin grows postnatally, collagen type III becomes sparse." In healing wounds the decline in collagen type III has been associated with successful remodeling, and in instances such as hypertrophic scars where remodeling is deficient, excess collagen type III is found.' It appears, then, that the wound in cuff tearing remains in an early stage of attempted repair without significant progression. Ehrlich" has shown that matrix consisting mostly of collagen type III contracts faster and to a greater degree than does that consisting predominantly of type I. No morphologic evidence of contracture in the torn cuff tendons could be recognized with certainty in our specimens, although the torn edges appeared to be retracted during surgery. We conclude that collagen type III is abundant in rotator cuff tears, which indicates a need for continued attempts for repair at the site. The duration of the symptomatic preoperative period has no discernible influence on the collagen formation. It can be postulated that the transformation of collagen type III to collagen type I might occur in patients who have surgical procedures. One can also hypothesize that bridging of the wound is necessary for the replacement of type III collagen by type I collagen; type I collagen is essential for the development of tensile strength in the tendon. We sincerely thank Professor A. Ooshima of First Department of Pathology, Wakayama Medical College, Wakayama, and Dr. K. Iwata of Fuji Chemical Industries, Toyama, Japan, for supplying the anti-
1. Shoulder Elbow Surg. July/ August 7992
body. Dr. Y. Okawara, Department of Anatomy, Faculty of Medicine of the University of Ottawa, and Dr. T. Takahara, Third Department of Internal Medicine, Toyama Medical and Pharmaceutical University, gave us valuable suggestions. Expert technical assistance was provided by Mrs. Brigitte Eiselt and Mrs. Julia Katnick-Fournier.
REFERENCES 1. Amiel D, Frank CB, Harwood FL, Akeson WH, Kleiner JB. Collagen alteration in medial collateral ligament healing in a rabbit model. Connect Tissue Res 1987;16:357-66. 2. Bailey AJ, Bazin S, Sims TJ, Le Lous M, Nicoletis C, Delaunay A. Characterization of the collagen of human hypertrophic and normal scars. Biochim Biophys Acta 1975;405:412-21. 3. Clore IN, Cohen IK, Diegelmann RF. Quantitation of collagen types I and III during wound healing in rat skin. Proc Soc Exp Bioi Med 1979;161 :337-40. 4. Duance VC, Restall DJ, Beard H, Bourne FJ, Bailey AJ. The location of three collagen types in skeletal muscle. FEBS Lett 1977;79:248-52. 5. Ehrlich HP. Wound closure: evidence of cooperation between fibroblasts and collagen matrix. Eye 1988;2: 14957. 6. Epstein EH Jr. [a 1 (1I11h human skin collagen: release by pepsin digestion and preponderance in fetal life. J Bioi Chem 1974;249:3225-31. 7. EssmanJA, Bell RH, Askew M. Full-thickness rotator-cuff tear: an analysis of results. Clin Orthop 1991;265: 170-7. 8. Gay S,Viljanto J, Raekallio J, Penttinen R. Collagen types in early phases of wound healing in children. Acta Chir Scand 1978; 144:205-11. 9. Itoi E, Tabata S. Conservative treatment of rotator cuff tears. C1in Orthop 1992;275: 165-73. 10. McPherson JM, Piez KA. Collagen in dermal wound repair. In: Clark RAF, Henson PM, eds. The molecular and cellular biology of wound repair. New York: Plenum Press, 1988:471-96. 11. Rentrop M, Knapp B, Winter H, Schweizer J. Aminoalkylsilane-treoted glass slides as support for in situ hybridization of keratin cDNA to frozen tissue sections under varying fixation and pretreatment conditions. Histochem J 1986; 18:271-6. 12. Sakakibara K, Ooshima A, Igarashi S, Sakakibara J. Immunolocalization of type III collagen and procollagen in cirrhotic human liver using monoclonal antibodies. Virchows Arch A Pathol Anat Histopathol 1986;409:3746. 13. Uhthoff HK, Sarkar K. Surgical repair of rotator cuff ruptures: the importance of the subacromial bursa. J Bone Joint Surg [Br] 1991;73:399-401. 14. Willioms IF, Heaton A, McCullagh KG. Cell morphology and collagen types in equine tendon scar. Res Vet Sci 1980;28:302-10. 15. Wulker N, Melzer C, Wirth CJ. Shoulder surgery for rotator cuff tears: ultrasonographic 3-year follow-up of 97 cases. Acta Orthop Scand 1991;62: 142-7. 16. Yulis CR, Lederis K. Co-localization of the immunoreactivities of corticotropin-releasing factor and arginine vasotocin in the brain and pituitary system of the teleost Catostomus commersoni. Cell Tissue Res 1987;247:26773.
J. P. Murnaghan, MO, Ottawa, Ontario, Canada The presence and distribution of collagen type III, which represents newly formed fibers of early repair, were investigated in four surgically removed en bloc specimens of complete rotator cuH tears. The patients were two men and two women who ranged in age from 63 to 77 years. They had received preoperative, conservative treatment for a period varying from 3 months to 3 years. For the identification of the collagen immunohistochemical staining was performed by the peroxidase-antiperoxidase method with the use of monoclonal antibody against human collagen type III. Collagen type III was discernible throughout the tendon proper in all specimens. It often occurred in association with proliferating fibroblastic cells. Collagen type III was abundant only in the perivascular spaces of the much thickened bursal wall that extended over the margins of the tear. We concluded that the presence of collagen type III indicates the need for a continued attempt for repair at the site of the tear, including both bursal and tendinous tissues. (J SHOULDER ELBOW SURG 1992;1: 187-92.]
In
most elderly patients with a rotator cuff tear that develops either spontaneously or after minimal indirect trauma, an indefinite period of conservative management is usually followed. A considerable number of patients show satisfactory improvement/ and surgery is undertaken in only those who have persistent pain accompanied by a notable loss of function, especially of arm abduction. IS Thus the period of time elapsing between the diagnosis of the tear and the operation on the tear may vary widely from patient to patient and may often extend from months to years . However, preoperative duration of symptoms and the extent of the tear appear to be of little significance in the surgical outcome of cuff tecrs.' We have shown that repair reaction, which is represented by granulation tissue and apparently originates both in the overlying subacromial bursa and to a lesser degree in the torn tendon margins themselves, From the Division of O rthopaed ics, O ttowa General Hosp ital and Ottowa Civic Hosp ital, and the Departments of Surgery and Pathology, University of Ottowa. Support ed by the Medical Research Coun cil of Ca nada. Reprint requests: Hans K. Uhthoff, MD, FRCS(CL Professor and Cha irman, Deportment of Surgery, 501 Smyth Rd., Ottowa, Ontario Kl H 8L6.
32/1140100
is evident in all cases regardless of the duration of conservative moncqement." It has been demonstrated in experiments thot granulation tissue in the initial stages of wound heal ing is mostly comprised ofcollagen type III. This is the case even in tissues such as the skin and the tendon, which have collagen type I as the predominant component." 8. 14 It has been shown in the healing medial collateral ligament of the rabbit that the amount of collagen type III reaches its peak around the end of the second week of wound healing and then begins to decrease with the simultaneous rise of collagen type 1. 1 The collagen type III is replaced by collagen type I as the scar motures.'? This study was undertaken to determine whether variations in the length of preoperative periods of conservative management before the surgical procedure for complete rotator cuff tear would significantly alter the distribution of collagen type III at the margins of the tear and in the vicinity of the tear.
MATERIAL AND METHODS En bloc specimens of the supraspinatus tendon containing the complete tear were obtained during surgery from two men and two women. All ruptures occurred without a history of severe
187
188
Kumagai et al.
J. Shoulder Elbow Surg. July/August 1992
p
•
"'-"'· -M
l
A
mm
II rill' I Il'
II
1111 I
Figure 1 Art icular surface of typ ica l en bloc specimen of supraspinatus tendon w ith central tear . Sections for immuno histochemica l staining were token along dotted lines: A, anterior; P, posterior; M, medial; L, lateral. GT ind icates area of greater tuberosity.
trauma to the supraspinatus tendon, although in one 67-year-old woman the clinical evidence of tear was preceded by shoulder dislocation. The men were 63 and 66 years old, and the preoperative durations of symptoms were 9 and 6 months, respectively. The two women were 67 and 71 years old; their durations of symptoms were 3 months and 3 years, respectively. The en bloc specimens were fixed in 10% neutral buffered formalin . After fixation was performed sections were taken from anterior (toword the subscapularis tendon), posterior (toward the infraspinatus tendon), medial (toward the supraspinatus muscle), and lateral (toward the bony insertion) margins of the tear for paraffin embedding (Fig . 1). Serial sections 5 urn thick were placed on slides coated with a 2% solution of 3-aminopropyltriethoxysilane to prevent detachment of the tissue section. 11 Four age-matched controls were then obtained from cadavers that had no history of shoulder ailment, and these tissues were treated in the same manner. The preparation of the anti-human collagen type III monoclonal antibody that we used has been described in detail by Sakakibara et 01. 12 We obtained the antibody from Professor Akira Ooshima of Wakayama Prefectural Med ical
College of Japan. In this method collagen type III is extracted from human placenta and is digested with trypsin. The digest, combined with Freund's complete adjuvant, is injected into BALBI c mice . Splenic lymphocytes of the sensitized mice are then fused with mouse myeloma cells. The immunoglobulin G-producing cells are then cloned by limited dilution assay, and the immunoglobulin G reaction with the a-cha in of type III collagen is determined by immunoblotting . In our laboratory serial sections were first de paraffinized with xylene and rehydrated in graded ethanol. They were then treated with 0.25% trypsin (GIBCO, Life Technologies, Inc., Gaithersburg, Md.) that was dissolved in the 0.1 molll tris-HCI buffer (TBS) at pH 7.7 at 3]0 C for 60 minutes and washed in TBS at room temperature to stop all enzyme activity. For immunohistochemical staining we adopted the technique of Yulis and Lederis." All antisera were diluted in TBS conta ining 0.1 % Triton X-l00 (Sigma Diagnostics, St. louis, Mo.) and 0.7% carrageenan (Sigma) . The procedure was performed at room temperature. The sections were incubated overnight w ith the primary antibody acting against collagen type III in a dilution of 1 : 16,000 (0.71 j.Lgl ml) . Sections were
Volume 1 Number 4
Collagen type III in rotator cuff tears
189
-
Figure 2 Immunohistochemical staining of supraspinatus tendon from 67year-old man used as control shows fibrillar collagen type III in endotenon, which conta ins vascular channels . (Original magn ification x 200 .)
Figure 3 Immunohistochemical staining of bursal tissue at site of tear shows collagen type III mostly in relation to proliferating vascular channels. (Original magnification x 100.) then treated with the second antiserum, rabbit anti-mouse immunoglobulin (DAKO Corp., Carpenteria, Calif.) diluted 1 :25, applied for 60 minutes. Finally mouse peroxidase-antiperoxidase complex (DAKO), diluted 1 :250, was ap-
plied to the sections for 60 minutes. Between each incubation the slides were washed three times with TBS for 5 minutes. After the sections were incubated with peroxidase-antiperoxidase, they were treated for 20 minutes with
190 Kumagai et
01.
J. Shoulder Elbow Surg. July/August 7992
" ,
Figure 4 Immunoh istochemical staining from lateral aspect of rotator cuff tendon tear shows diffuse distribution of collagen type III. (Original magnification x 200.) 0.25% solution of diaminobenzidine tetrachloride in TBS containing 0.05% H2 0 2 • Controls included substitution of the primary antibody with TBS. The sections were counterstained with hematoxylin for improved visualization of nuclei. Routine sections were stained with hematoxylin-eosin.
RESULTS Control sections of the supraspinatus tendon from elderly individuals showed thinning of the fascicles and broadening of the endotenon space, which contained a number of prominent vascular channels. Collagen type III staining was found in the endotenon but was absent in the adjacent fascicles (Fig. 2). In all of the specimens of cuff tearing the usual architecture of the tendon, which is characterized by fascicles in parallel arrangement with the intervening endotenon, could not be recognized in the stump area. The torn margins showed fronds of tendinous tissue in close proximity to bursal tissue. There was a significant degree of fibrovascular proliferation in the thickened bursal wall. Cellular as well as vascular proliferation occurred in the tendon tissue, not only at the region of the tear but also close to the margins of surgical resection . Most of the proliferating cells in the tendon substance appeared to be plump mesenchymal cells. Inflammatory cells were not present. Cellular and vas-
eular proliferation was conspicuous in all of the representative sections taken from the medial, lateral, anterior, and posterior segments of the specimens . Collagen type III staining was predominant in areas around the vessels and contiguous with the mesenchymal cells. The fibrillary staining often lacked specific orientation, although its major concentration around the vessels in the bursal tissue gave it a concentric appearance (Fig. 3). In the tendon substance, vascular proliferation was comparatively much less than in the bursa, and the presence of collagen type III was more visible in relation to proliferating cells (Fig. 4). Staining in connection with the cells could also be seen in the fronds at the tear margins, which were devoid of vascular channels (Fig. 5).
DISCUSSION This study clearly showed that although collagen type III in control sections was segregated as fine fibrillary sheaths in the endotenon, forming a minor component of the tendinous tissue, it was greatly increased in the specimens of torn rotator cuff tendons. The normal tendon is primarily composed of collagen type I, and thus it is similar to skin and bone in collagen composition. The distribution and the relative quantity of collagen type III in normal tendons have been established. Its preferential presence in
Collagen type III in rotator cuff tears
Volume 7 Number 4
191
Figure 5 Tendinous fragments apparently devoid of vessels at site of tear show presence of collagen type III. Acellular fragment without staining on right is probably necrotic. (Original magn ification x 400.) the endotenon has been demonstrated by immunofluorescent staining in bovine tendons: Williams et ol." estimated that collagen type III constituted less than 1% of collagenous component in normal equine tendons. It is well known that collagen type III increases significantly during the healing of a wound." In the previously mentioned articles on equine tendon, Williams et ol." found that scar formation raised collagen type III content of the tendon to 20% to 30%. Significantly, the rise in type III collagen is most apparent in the early phase of wound healing. This was demonstrated by Gay et ol.," who implanted viscose cellulose sponge in surgical wounds of children and also cultured the granulation tissue of wounds from similar patients. From immunohistological and biochemical studies they concluded that primitive mesenchymal cells that are present in .the initial period of wound healing produce collagen type III, whereas fibroblasts, appearing at a later stage, are responsible for the production of type I. The initial prevalence of collagen type III has also been shown in rat dermal
wounds.' In this study the presence of collagen type III appeared to be dependent on both interstitial and vascular cells. The abundance of collagen type III around the vessels, especially in the encroaching bursal tissue overlying the torn tendon margins, was remarkable. In the vicinity of
the tear the tendinous tissue that had lost both its normal configuration of fascicles and its intervening endotenon showed collagen type III both in the tendon substance and around the vessels. The vascular proliferation in the tendon substance, however, was much less prevalent than in the bursal tissue, and collagen type III staining in the tendon was more conspicuous in relation to cells. The association of collagen type III staining with proliferating cells was even demonstrable at the cuff tear margins in the fragmented tissues, some of which were devoid of vessels (Fig. 5). The principal aim in surgical repairs of rotator cuff tears is to reestablish continuity so that the forces for movement can be transmitted from the muscles to the bones. During surgery an extensive debridement is often practiced not just for adequate visualization but because of the belief that debridement will remove apparently dead tissue and will ensure sufficient vascularity at the suture line to induce new matrix formation . We have shown that regardless of the elapsed time from the onset of symptoms to surgery, granulation tissue in the area of tear is adequate, and tr imming for the sake of improving vascularity is counterproductive.v ln the present study the presence of collagen type III, which represents the new matrix, clearly indicates that the attempt to heal at the margins of a tear persists over time. It appears that ol-
192 Kumagai et
01.
though the strongest reaction for collagen type III is in the vicinity of the tear, its presence extends to the edge of the surgical resection. There were no differences in staining among the medial, lateral, anterior, and posterior aspects of the specimens. The present study also supports our previous contention that the duration of time between the onset of symptoms and the eventual surgery is of little significance because the formation of new matrix collagen was unaffected by time. The ongoing presence of collagen type III probably indicates that little remodeling is occurring in the wound. Whether the persistence of collagen type III for months or even years at the margins of a rotator cuff tear is beneficial or detrimental to eventual repair when the tendon is reconstructed is a matter of speculation. The presence of collagen type III is believed to indicate tissue immaturity. There is a preponderance of collagen type III in fetal skin, but as the skin grows postnatally, collagen type III becomes sparse." In healing wounds the decline in collagen type III has been associated with successful remodeling, and in instances such as hypertrophic scars where remodeling is deficient, excess collagen type III is found.' It appears, then, that the wound in cuff tearing remains in an early stage of attempted repair without significant progression. Ehrlich" has shown that matrix consisting mostly of collagen type III contracts faster and to a greater degree than does that consisting predominantly of type I. No morphologic evidence of contracture in the torn cuff tendons could be recognized with certainty in our specimens, although the torn edges appeared to be retracted during surgery. We conclude that collagen type III is abundant in rotator cuff tears, which indicates a need for continued attempts for repair at the site. The duration of the symptomatic preoperative period has no discernible influence on the collagen formation. It can be postulated that the transformation of collagen type III to collagen type I might occur in patients who have surgical procedures. One can also hypothesize that bridging of the wound is necessary for the replacement of type III collagen by type I collagen; type I collagen is essential for the development of tensile strength in the tendon. We sincerely thank Professor A. Ooshima of First Department of Pathology, Wakayama Medical College, Wakayama, and Dr. K. Iwata of Fuji Chemical Industries, Toyama, Japan, for supplying the anti-
1. Shoulder Elbow Surg. July/ August 7992
body. Dr. Y. Okawara, Department of Anatomy, Faculty of Medicine of the University of Ottawa, and Dr. T. Takahara, Third Department of Internal Medicine, Toyama Medical and Pharmaceutical University, gave us valuable suggestions. Expert technical assistance was provided by Mrs. Brigitte Eiselt and Mrs. Julia Katnick-Fournier.
REFERENCES 1. Amiel D, Frank CB, Harwood FL, Akeson WH, Kleiner JB. Collagen alteration in medial collateral ligament healing in a rabbit model. Connect Tissue Res 1987;16:357-66. 2. Bailey AJ, Bazin S, Sims TJ, Le Lous M, Nicoletis C, Delaunay A. Characterization of the collagen of human hypertrophic and normal scars. Biochim Biophys Acta 1975;405:412-21. 3. Clore IN, Cohen IK, Diegelmann RF. Quantitation of collagen types I and III during wound healing in rat skin. Proc Soc Exp Bioi Med 1979;161 :337-40. 4. Duance VC, Restall DJ, Beard H, Bourne FJ, Bailey AJ. The location of three collagen types in skeletal muscle. FEBS Lett 1977;79:248-52. 5. Ehrlich HP. Wound closure: evidence of cooperation between fibroblasts and collagen matrix. Eye 1988;2: 14957. 6. Epstein EH Jr. [a 1 (1I11h human skin collagen: release by pepsin digestion and preponderance in fetal life. J Bioi Chem 1974;249:3225-31. 7. EssmanJA, Bell RH, Askew M. Full-thickness rotator-cuff tear: an analysis of results. Clin Orthop 1991;265: 170-7. 8. Gay S,Viljanto J, Raekallio J, Penttinen R. Collagen types in early phases of wound healing in children. Acta Chir Scand 1978; 144:205-11. 9. Itoi E, Tabata S. Conservative treatment of rotator cuff tears. C1in Orthop 1992;275: 165-73. 10. McPherson JM, Piez KA. Collagen in dermal wound repair. In: Clark RAF, Henson PM, eds. The molecular and cellular biology of wound repair. New York: Plenum Press, 1988:471-96. 11. Rentrop M, Knapp B, Winter H, Schweizer J. Aminoalkylsilane-treoted glass slides as support for in situ hybridization of keratin cDNA to frozen tissue sections under varying fixation and pretreatment conditions. Histochem J 1986; 18:271-6. 12. Sakakibara K, Ooshima A, Igarashi S, Sakakibara J. Immunolocalization of type III collagen and procollagen in cirrhotic human liver using monoclonal antibodies. Virchows Arch A Pathol Anat Histopathol 1986;409:3746. 13. Uhthoff HK, Sarkar K. Surgical repair of rotator cuff ruptures: the importance of the subacromial bursa. J Bone Joint Surg [Br] 1991;73:399-401. 14. Willioms IF, Heaton A, McCullagh KG. Cell morphology and collagen types in equine tendon scar. Res Vet Sci 1980;28:302-10. 15. Wulker N, Melzer C, Wirth CJ. Shoulder surgery for rotator cuff tears: ultrasonographic 3-year follow-up of 97 cases. Acta Orthop Scand 1991;62: 142-7. 16. Yulis CR, Lederis K. Co-localization of the immunoreactivities of corticotropin-releasing factor and arginine vasotocin in the brain and pituitary system of the teleost Catostomus commersoni. Cell Tissue Res 1987;247:26773.