Characterization of hyaluronan and TSG-6 in skin scarring: differential distribution in keloid scars, normal scars and unscarred skin

Abstract

Background: Hyaluronan (HA) is a major component of the extracellular matrix (ECM) with increased synthesis during tissue repair. Tumour necrosis factor-stimulated gene-6 (TSG-6) is known to catalyze the covalent transfer of heavy chains (HC1 and HC2) from inter-α-inhibitor (IαI) onto HA, and resultant HC•HA complexes have been implicated in physiological and pathological processes related to remodelling and inflammation.

Objective: The aims of this study were to determine the expression of HA, TSG-6 and the IαI polypeptides in unscarred skin, normal scars and keloid scars.

Methods: Formalin-fixed paraffin-embedded sections of unscarred skin, normal scars and keloid scars were prepared from patient samples collected during scar revision surgery. Haematoxylin and eosin, as well as immunofluorescent staining for HA, TSG-6 and the three polypeptide chains of IαI (i.e. HC1, HC2 and bikunin) were performed.

Results: All skin types stained positive for TSG-6, HC1, HC2 and bikunin, associated with keratinocytes, fibroblasts and skin appendages all in close proximity to HA. Keloid lesions showed altered HA organization patterns compared with unscarred skin and normal scars. TSG-6 staining was significantly more intense in the epidermis compared with the dermis of all sample types. There was a significant reduction in TSG-6 levels within keloid lesions compared with the dermis of unscarred skin (P=0.017).

Conclusion: TSG-6 is expressed in unscarred skin, where its close association with HA and IαI could give rise to TSG-6-mediated HC•HA formation within this tissue. A reduction in the beneficial effects of TSG-6, caused by diminished protein levels in keloid lesions, could contribute to this abnormal scarring process.

Figure 1

Haematoxylin and eosin (H&E) staining. Unscarred skin (a) with an undulating dermal-epidermal junction (white arrow) and haphazardly arranged dermal collagen bundles (black arrow). A normal scar (b) with a flattened epidermis (white arrow) and thin strand-like collagen bundles (black arrow) parallel to the epidermis. A keloid scar (c) with flattened epidermis (white arrow) and thin, parallel strand-like collagen bundles (black arrow) in the superficial dermis. Emigrant lymphocytes were seen in the dermis around vascular structures in all three tissue types (blue arrows). Hair follicles (d) consist of an internal (black arrow) and external root sheath (white arrow); a sebaceous gland (e) containing sebocytes (black arrows) filled with abundant lipid droplets and thin strands of cytoplasm; and an eccrine gland (f) with several cross sections of a coiled duct (black arrows) in unscarred skin. A low magnification view of a keloid scar (g) showing a flattened epidermis, thin strand-like collagen bundles in the SD and thick bundles of collagen arranged in a haphazard orientation within the actual KL in the reticular dermis. The central region of a KL (h) shows a high density of cells interspersed between the thick bundles of collagen (black arrows). The central region of another KL (I) shows thinner bundles of collagen arranged in a reticular pattern (black arrows). (Ep: epidermis; D: dermis). Scale bars: 100 μm (Original magnification a–f,h,i: 20×; g: 10×).

Figure 2

Immunofluorescent staining of unscarred skin. A combined image (a) is shown for HA (green), TSG-6 (red) and cell nuclei (blue) together with single filter images for TSG-6 (c), and HA (d). The negative control slide (b) was treated with the RAH-1 pre-immune serum. There was pericellular localization of HA in the epidermis and diffuse HA staining in the dermis. TSG-6 staining (a) was associated with keratinocytes in the epidermis and with fibroblasts (orange arrow), endothelial cells (white arrow) and perivascular lymphocytic cells (yellow arrow) in the dermis. Scale bars: 100 μm (Original magnification: 20×).

Figure 3

Immunofluorescent staining of skin appendages in unscarred skin. Combined images are shown for HA (green; a–l) with TSG-6 (red; a–c), HC1 (red; d–f), HC2 (red; g–i) or bikunin (red; j–l) and cell nuclei (blue). The negative control slides (m–o) were treated with the RAH-1 pre-immune serum; identical results were obtained for slides treated with the pre-immune sera corresponding to the antisera specific for HC1, HC2 and bikunin (not shown). The highest levels of HA staining (white arrows in a–c) were associated with the external aspects of hair follicle, intercellular space of external root sheaths and inner aspect of internal root sheaths (a), the external aspects and central parts of sebaceous glands (b) and the external aspects of eccrine sweat glands (c). TSG-6 staining (yellow arrows) was associated with the hair follicular cells (a), sebocytes (b) and secretory epithelial cells of eccrine sweat glands (c). Similar staining patterns were seen for HC1 (d–f), HC2 (g–i) and bikunin (j–l). Scale bars: 100 μm (Original magnification: 20×).

Figure 5

Immunofluorescent staining of keloid lesions within the reticular dermis. Combined images are shown for HA (green; a–d) with TSG-6 (red; a), HC1 (red; b), HC2 (red; c) or bikunin (red; d) and cell nuclei (blue). Reticular HA staining was seen between the thick bundles of collagen. Staining for TSG-6 (a), HC1 (b), HC2 (c) and bikunin (d) were localised to keloid fibroblasts (orange arrows). Scale bars: 100 μm (Original magnification: 20×).

Figure 4

Immunofluorescent staining of unscarred skin, normal scars and keloid scars. Combined images are shown for HA (green; a–l) with TSG-6 (red; a–c), HC1 (red; d–f), HC2 (red; g–i) or bikunin (red; j–l) and cell nuclei (blue). The negative control slides (m–o) were treated with the RAH-1 pre-immune serum; identical results were obtained for slides treated with the pre-immune sera corresponding to the antisera specific for HC1, HC2 and bikunin (not shown). In unscarred skin (a,d,g,j) there was pericellular localisation of HA in the epidermis and diffuse HA staining in the dermis. In normal scars (b,e,h,k), and keloid scars (c,f,i,l) there was pericellular epidermal and strand-like dermal staining for HA. TSG-6 staining was associated with keratinocytes in the epidermis and with fibroblasts (orange arrows) and endothelial cells (white arrows) in unscarred skin (a) and normal scars (b) and keloid scars (c) as well as with perivascular lymphocytic cells (yellow arrows) in unscarred skin (a) and normal scars (b). Similar staining patterns were seen for HC1 (d–f), HC2 (g–i) and bikunin (j–l), except that there was increased staining in the keratin layers of unscarred skin and normal scars (cyan arrows). Scale bars: 100 μm (Original magnification: 20×).

Figure 6

Quantification of TSG-6 staining intensity in unscarred skin (n = 3), normal scars and (n = 6) keloid scars (n = 7). Data are represented as mean staining intensity ± 2 x S.E.M. In all skin types there was significantly less TSG-6 detectable in the dermis compared to the epidermis (*); P = 0.029, P < 0.001 and P = 0.004 for unscarred skin, normal scars and keloid scars, respectively. There were no significant differences between the intensities of TSG-6 immunostaining either in the epidermis or in the dermis (immediately above lesions) when comparing unscarred skin, normal scars and keloid scars.

Figure 7

Quantification of TSG-6 staining intensity in unscarred skin dermis (n = 3), and in the centres of normal scars (n = 6) and keloid scars (n = 7). Data are represented as mean staining intensity ± 2 x S.E.M. TSG-6 immunostaining was significantly reduced in keloid lesions (P = 0.017), but not in normal scar lesions (P = 0.08), compared to unscarred skin dermis. * indicates P < 0.05.

Figure 8

Quantification of HA staining intensity in unscarred skin (n = 3), normal scars (n = 6) and keloid scars (n = 7). Data are represented as mean staining intensity ± 2 x S.E.M. There was significantly less HA (*) detectable in the dermis compared to the epidermis of normal scars (P = 0.004). The reduced HA staining in unscarred skin epidermis compared with unscarred skin dermis staining was not significant (P = 0.086). There was no significant difference between HA staining intensities of keloid epidermis and keloid dermis. Keloid dermis showed significantly reduced staining for HA (*) compared with unscarred skin (P = 0.04) and normal scar dermis (P = 0.003).