- Email: [email protected]
Immunofluorescent microscopic investigation of the distal arrector pili: A demonstration of the spatial relationship between α5β1 integrin and fibronectin
REPORTS
Immunofluorescent microscopic investigation of the distal arrector pili: A demonstration of the spatial relationship between α5β1 integrin and fibronectin Missy M. Clifton,a,b Jerri K. Mendelson,a Bruce Mendelson,e Donna Montague,c Charleata Carter,d Bruce R. Smoller,a,b and Thomas D. Horn Little Rock and Conway, Arkansas Currently there is limited knowledge regarding the anatomy of the distal arrector pili (AP) muscle. A previous study implicated fibronectin and α5β1 integrin binding as the anchor between the AP and the extracellular matrix (ECM). The purpose of this study was to strengthen this hypothesis. Serial frozen sections of human scalp skin were double-labeled via immunofluorescent staining for α5β1 with fluorescein and fibronectin with rhodamine, followed by fluorescent microscopy. Granular staining for α5β1 with fluorescein and smooth staining for fibronectin with rhodamine were seen at the periphery of the AP muscle bundles and along the distal fibers. Precise co-localization of α5β1 and fibronectin was observed at the AP-ECM interface by means of a dual filter. Analysis of variance was used on the relative density of staining for each epitope. Staining for both epitopes was significantly brighter at the distal fibers than at the middle or proximal portions of the muscle. A computerized three-dimensional reconstruction provides a detailed picture of the microanatomy of the distal AP, which allows mathematical evaluation of the forces of contraction. The anatomic co-localization between α5β1 and fibronectin strengthens our hypothesis that interaction of these epitopes mediates the attachment of the distal AP to the ECM. (J Am Acad Dermatol 2000;43:19-23.)
T
he proximal attachment of the arrector pili (AP) muscle to the pilosebaceous apparatus has been studied in detail. There is general agreement that the proximal fibers of the AP attach, via elastic tendons,1 to the hair follicle at the bulge area, which consists, in part, of a pool of undifferentiated keratinocytes.2 Three bulge areas on the hair follicle can be identified in embryogenesis. The uppermost bulge involutes or forms an apocrine gland, the middle bulge forms the sebaceous gland, and the lowermost bulge forms the attachment for the AP muscle.3 It has also been shown that fibers of
From the Departments of Dermatology,a Pathology,b Orthopedics,c and Obstetrics/Gynecology,d University of Arkansas for Medical Sciences, Little Rock, and the Department of Physical Therapy, University of Central Arkansas, Conway.e Supported by a student fellowship grant from the American Dermatological Association (to M. M. C.). Accepted for publication Dec 9, 1999. Reprint requests: Thomas D. Horn, Chairman, Department of Dermatology, University of Arkansas for Medical Sciences, 4301 W Markham, Slot 576, Little Rock, AR 72205-7199. Copyright © 2000 by the American Academy of Dermatology, Inc. 0190-9622/2000/$12.00 + 0 16/1/105159 doi:10.1067/mjd.2000.105159
some AP muscles can be found admixed with the connective tissue sheath around the entire circumference of the hair follicle.4 The distal anatomy of the muscle and its attachments to the extracellular matrix (ECM) are less well characterized. Many texts schematically depict the distal AP as attaching as a single fascicle to the epidermal basement membrane. Narisawa, Hashimoto, and Kohda5 showed that most of the distal ends of the AP were actually situated in the upper dermis, whereas some fibers seemed to be in close contact with the basal layer of the epidermis. The mechanisms of attachment of the muscle to the ECM allowing hair erection upon contraction remain uncertain. Using immunohistochemistry6 Mendelson et al documented expression of integrin subunits α1, α5, and β1 by the distal AP muscle. Fibronectin, a known ligand of α5β1, was distributed in a pattern that roughly coincided with α5 distribution at the distal AP-ECM interface. Interaction between α5β1 and fibronectin occurs in other areas of the body and allows signal transduction and mechanoreception in these sites.7,8 From these results we hypothesized that the interaction between fibronectin and α5β1 integrin 19
20 Clifton et al
J AM ACAD DERMATOL JULY 2000
A
B
C Fig 1. Views of a distal arrector pili muscle bundle in the fluorescein wavelength (A), in the rhodamine wavelength (B), and using a dual filter (C) demonstrate co-localization of α5β1 and fibronectin at areas of contact between the AP and ECM.
mediates the attachment of the AP muscle to the ECM. The purpose of our study was to strengthen this hypothesis and better delineate the attachments and anatomy of this muscle to shed light on its contractile physiology. The study involved immunofluorescent double labeling for fibronectin and α5β1 integrin to better demonstrate the distal anatomy of the AP and its attachments to the epidermis and ECM.
MATERIAL AND METHODS Tissue samples Sections of normal fresh scalp skin were obtained from the occipital area of cadavers (because good examples of AP muscles in serial sections are hard to find, tissue from this study was only taken from the occipital area; AP muscles are more plentiful in this region than any other area of the body). The tissue was immediately cut into 1- × 3-mm strips and embedded in OCT. The embedded tissue was then frozen at –70°C. The blocks were cut into serial 5-µm vertical sections by cryostat and mounted on charged glass slides. The slides were stored frozen until immunofluorescent studies were performed.
Immunofluorescent technique Standard immunofluorescent techniques were used to achieve double-labeling. Serial incubations were performed with anti-α5β1 (at 1:20 for 90 minutes, DAKO, Carpinteria, Calif) followed by antifibronectin (at 1:500 for 90 minutes, DAKO) with separate steps to tag the primary antibodies with fluorescent labels between these incubations. The slides were then lightly counterstained in hematoxylin and Scott’s water blueing agent. Coverslips were applied with the use of mounting media containing a fixative to preserve fluorescence. Microscopy and reconstruction A dual light and fluorescent microscope mounted with filters for viewing fluorescein and rhodamine and a dual filter for viewing both the fluorescein and rhodamine wavelengths simultaneously were used to examine the sections. Photomicrographs demonstrating the staining of α5β1 with fluorescein, fibronectin with rhodamine, and areas of overlapping staining were obtained from the serial 5-µm sections. Photomicrographs demonstrating overlapping staining of fibronectin and α5β1 were used as templates to reconstruct a computerized
Clifton et al 21
J AM ACAD DERMATOL VOLUME 43, NUMBER 1, PART 1
Fig 2. This three-dimensional representation of the anatomy of the distal AP muscle was created by means of serial sections through a single muscle as templates. Figure shows the arrector pili muscle (AP), the hair follicle (HF), the sebaceous gland (SG), and the epidermis (Epi). (Courtesy of Jones Productions Inc, Little Rock, Ark.)
three-dimensional picture of the distal AP and its attachments. Statistical analysis To determine whether the immunofluorescent staining of α5β1 exhibited significant differences in relative signal density at the distal AP-ECM interface compared with the middle and proximal portions of the muscle, the signal density was measured at the distal, middle, and proximal portions of the muscle and corresponding areas of the ECM. These measurements were also taken in the same regions for fibronectin. With an area of 50 µm2, signal measurements from representative areas at each portion of the muscle and representative areas from each corresponding portion of the ECM were obtained and digitized. The digitized densities obtained from each part of the muscle were averaged. The ECM densities for each area were also averaged and used as control values for that corresponding area of the muscle. A percent change over control was determined for the distal, middle, and proximal portions of the muscle for both α5β1 and fibronectin. Specific interregional
comparisons were performed by means of one-way analysis of variance followed by the Scheffé test. The above measurements and calculations were done on serial sections of 3 separate AP muscles.
RESULTS Examination of the serial sections with light and fluorescent microscopy showed muscle fibers splaying into multiple branches, which terminated in areas of the papillary and reticular dermis as previously described. Some fibers could be seen approaching the epidermal basement membrane. Fluorescein staining, indicating expression of α5β1 integrin, was granular and more pronounced at the periphery of the AP bundles and on the distal fibers of the muscle (Fig 1, A). Rhodamine staining, indicating expression of fibronectin was seen in a smooth pattern around muscle bundles and at the interface of the distal fibers and the ECM and basement membrane (Fig 1, B). A dual filter allowed simultaneous viewing of areas of colocalization of staining for α5β1 and fibronectin. Dual staining, in orange, was seen at all points of contact between the AP muscle and the ECM (Fig 1, C).
22 Clifton et al
A
J AM ACAD DERMATOL JULY 2000
B Fig 3. Demonstration of the results of ANOVA analysis followed by the Scheffé test for α5β1 (A) and fibronectin (B). The y-axis represents the percent increase in staining intensity of the AP versus the ECM in all 3 areas measured. Staining for both epitopes at the distal AP fiber–ECM interface was statistically brighter than staining at either the middle or proximal portion of the muscle (P = .0001).
Photomicrographs of 25 serial 5-µm sections through the same muscle dually stained were used as templates to create a computerized three-dimensional picture of the distal anatomy of the AP and highlight the co-localized staining of fibronectin and α5β1 (Fig 2). Statistical analysis of the digitized densities showed significant increased density in staining for both α5β1 and fibronectin in areas of AP-ECM contact than in areas within the ECM. Density of staining for α5β1 at the distal fibers of the AP was statistically more dense than at either the middle or proximal areas of the muscle (P = .0001) (Fig 3, A). Density of staining for fibronectin at the distal AP-ECM interface was also statistically more dense than staining at the middle or proximal portion of the muscle (P = .0001) (Fig 3, B). There was no statistically significant difference in density of staining between the middle and proximal portions of the muscle for either α5β1 or fibronectin. With our three-dimensional reconstruction of the AP, measurements of each of the branches were taken and the surface area of each branch was determined by assuming that each branch was cylindrical in shape (using the equation for the surface area of a sphere, excluding the base: πr2 + 2πrh). A total surface area for the 7 branches was then obtained. This total surface area was compared with the surface area of a theoretical single fascicle AP. Measurements for this model were obtained by using an average value for the branches of our model. With this approach, the total surface area for our model was 1761.98 µm2, whereas the surface area of the single fascicle model was 254.34 µm2. Our model had 6.93 times more surface area contacting the ECM than the single fascicle model. This increased surface area roughly cor-
responded to the number of branches of the AP. The actual increase in surface area is likely much greater than calculated above because such calculation does not take into account the continued splaying of the branches into terminal fibers. We also wished to calculate the area over which the contractile force is spread. Because of AP muscle splaying based on the three-dimensional model, the area over which the force of contraction would be spread was circular and had an area of 2552 µm2 (using the formula for the area of a circle: πr2; with the radius of the area determined from the threedimensional reconstruction). The area over which force would be dispersed for a single fascicle model was estimated to be 7.1 µm2. With this strategy the AP transduces the force of contraction over 359 times more area.
DISCUSSION With immunofluorescent double-labeling, our results document a precise co-localization of α5β1 integrin and fibronectin at the AP-ECM interface. Furthermore, these epitopes concentrate along the distal splayed portion of the muscle when compared with more proximal segments. These results further support our hypothesis that the interaction of these two epitopes mediates anchorage of the AP to the ECM and the basement membrane. Serial sectioning and use of computer graphics provide a detailed representation of the distal anatomy of the AP. Based on many studies, fibronectin and α5β1 are known to be involved in a myriad of biologic processes, including mediating cell attachment and cell stabilization.9-13 Such an interaction can be found between α5β1 and fibronectin in various tis-
J AM ACAD DERMATOL VOLUME 43, NUMBER 1, PART 1
sues including cartilage7 and endometrium.14 Their interaction and expression can be up-regulated in transformed cells, including those of Rous sarcoma virus–induced tumors15 and colon carcinoma cells.16 Although fibronectin binds at least 5 other integrins, α5β1 integrin is thought to be the primary fibronectin receptor because of its abundance and high affinity compared with other integrins.13 The integrin binds fibronectin at the peptide sequence known as the RGD (Arg-Gly-Asp) sequence.17-19 The fibronectin binding sites involve both subunits of the integrin.13 The cytoplasmic domain of the β1 subunit binds to talin, vinculin, actin, and fibulin, which make up portions of the cytoskeleton.20 In studies of cultured chondrocytes the normal cellular hyperpolarization response to pressure was ablated when the α5β1 receptor was blocked.7 When mechanical stress was applied to capillary endothelial cells, focal adhesion involving α5β1 occurred, as did an increase in cytoskeletal stiffness.21,22 Our data suggest that the interaction between fibronectin and α5β1 integrin plays an important role in mechanoreception, signal transduction, and stabilization of the cytoskeleton. Such an interaction would be important at the interface between the AP and the ECM and might allow the force transduction and stabilization necessary for muscle contraction and subsequent piloerection. The co-localization of α5β1 and fibronectin was seen at virtually every point connecting the AP with the ECM or the basement membrane. This finding suggests that rather than a single attachment to the basement membrane, the muscle actually interacts with its surroundings via these epitopes at many points in the papillary and reticular dermis, and at the basement membrane. Knowing the precise anatomy of the distal AP via computerized modeling allowed further assessment of the AP. The significant splaying of the AP and the increased surface area of the distal fibers for contact with the ECM (×7) is logical considering the equation for stress: Stress = Force/Area. The splayed anatomy of the AP muscle spreads the force of contraction over a much larger surface area (×359). Therefore the anatomic stress of contraction would be greatly lessened with the same force of contraction. To further support our hypothesis that fibronectin and α5β1 integrin interact to allow force transduction necessary for AP muscle contraction, a model of inhibition of the interaction between these molecules would be ideal to determine whether piloerection is abolished. We thank Jones Productions of Little Rock for assistance with the three-dimensional reconstruction.
Clifton et al 23
REFERENCES 1. Rodrigo GF, Cotta-Pereira G, David-Ferreira JF.The fine structure of the elastic tendons of the arrector pili muscle. Br J Dermatol 1975;93:631-6. 2. Akiyama M, Dale BA, Sun T, Holbrook KA. Characterization of hair follicle bulge in human fetal skin: the human fetal bulge is a pool of undifferentiated keratinocytes. J Invest Dermatol 1995;105:844-50. 3. Elder D, Elenitsas R, Jaworsky C, Johnson B Jr. Histology of the skin. In: Lever’s Histopathology of the skin. 8th edition. Philadelphia: Lippincott-Raven; 1997. p. 19-23. 4. Narisawa Y, Kohda H. Arrector pili muscles surround human facial vellus hair follicles. Br J Dermatol 1993;129:138-9. 5. Narisawa Y, Hashimoto K, Kohda H. Merkel cells participate in the induction and alignment of arrector pili muscles of human fetal skin. Br J Dermatol 1996;134:494-8. 6. Mendelson JK, Smoller BR, Mendelson B, Horn TD. The microanatomy of the distal arrector pili: possible role for alpha1 beta1 and alpha5 beta1 integrins mediating cell-cell adhesion and anchorage to the extracellular matrix. J Cutan Pathol 2000;27:61-6. 7. Clark EA, Brugge JS. Integrins and signal transduction pathways: the road taken. Science 1995;268:233-9. 8. Wright MO, Nishida K, Bavington C, Godolphin JL, Dunne E, Walmsley S, et al. Hyperpolarisation of cultured human chondrocytes following cyclical pressure-induced strain: evidence of a role for α5β1 Integrin as a chondroctye mechanoreceptor. J Orthop Res 1997;15:742-7. 9. Hay ED, editor. Cell biology of extracellular matrix. 2nd ed. New York: Plenum Press; 1991. 10. Ruoslahti E, Pierschbacher MD. New perspectives in cell adhesion: RGD and integrins. Science 1987;238:491-7. 11. Mosher DF, editor. Fibronectin. New York: Academic Press; 1989. 12. Hynes RO. Fibronectins. New York: Springer-Verlag; 1989. 13. Yamada KM. Fibronectins: structure functions and receptors. Curr Opin Cell Biol 1990;1:956-63. 14. Albright CD, Carter CA, Kaufman DG. Tamoxifen alters the localization of F-actin and α5/β1-integrin fibronectin receptors in human endometrial stomal cells and carcinoma cells. Pathobiology 1997;65:177-83. 15. Saga S, Chen W, Yamada K. Enhanced fibronectin receptor expression in Rous sarcoma virus induced tumors. Cancer Res 1988;48:5510-3. 16. Varner JA, Emerson DA, Juliano RL. Integrin α5β1 expression negatively regulates cell growth: reversal by attachment to fibronectin. Mol Biol Cell 1995;6:725-40. 17. Pierschbacher MD, Rouslahti E. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature 1984;309:30-3. 18. Pierschbacher MD, Rouslahti E. Variants of the cell recognition site of fibronectin that retain attachment promoting activity. Proc Natl Acad Sci U S A 1984;81:5985-8. 19. Yamada KM, Kennedy DW. Dualistic nature of adhesive protein function: fibronectin and its biologically active peptide fragments can auto-inhibit fibronectin function. J Cell Biol 1984; 99:29-36. 20. Horwitz A, Duggan K, Buck C, Beckerle MC, Burridge K. Interaction of plasma membrane fibronectin receptor with talin: a transmembrane linkage. Nature 1986;320:531-2. 21. Akiyama SK, Yamada SS, Chen WT, Yamada KM. Analysis of fibronectin receptor function with monoclonal antibodies: roles in cell adhesion, migration, matrix assembly, and cytoskeletal organization. J Cell Biol 1989;109:863-75. 22. Wang N, Butler JP, Ingber DE. Mechanoreception across the cell surface and through the cytoskeleton. Science 1993;260:11247.
Immunofluorescent microscopic investigation of the distal arrector pili: A demonstration of the spatial relationship between α5β1 integrin and fibronectin Missy M. Clifton,a,b Jerri K. Mendelson,a Bruce Mendelson,e Donna Montague,c Charleata Carter,d Bruce R. Smoller,a,b and Thomas D. Horn Little Rock and Conway, Arkansas Currently there is limited knowledge regarding the anatomy of the distal arrector pili (AP) muscle. A previous study implicated fibronectin and α5β1 integrin binding as the anchor between the AP and the extracellular matrix (ECM). The purpose of this study was to strengthen this hypothesis. Serial frozen sections of human scalp skin were double-labeled via immunofluorescent staining for α5β1 with fluorescein and fibronectin with rhodamine, followed by fluorescent microscopy. Granular staining for α5β1 with fluorescein and smooth staining for fibronectin with rhodamine were seen at the periphery of the AP muscle bundles and along the distal fibers. Precise co-localization of α5β1 and fibronectin was observed at the AP-ECM interface by means of a dual filter. Analysis of variance was used on the relative density of staining for each epitope. Staining for both epitopes was significantly brighter at the distal fibers than at the middle or proximal portions of the muscle. A computerized three-dimensional reconstruction provides a detailed picture of the microanatomy of the distal AP, which allows mathematical evaluation of the forces of contraction. The anatomic co-localization between α5β1 and fibronectin strengthens our hypothesis that interaction of these epitopes mediates the attachment of the distal AP to the ECM. (J Am Acad Dermatol 2000;43:19-23.)
T
he proximal attachment of the arrector pili (AP) muscle to the pilosebaceous apparatus has been studied in detail. There is general agreement that the proximal fibers of the AP attach, via elastic tendons,1 to the hair follicle at the bulge area, which consists, in part, of a pool of undifferentiated keratinocytes.2 Three bulge areas on the hair follicle can be identified in embryogenesis. The uppermost bulge involutes or forms an apocrine gland, the middle bulge forms the sebaceous gland, and the lowermost bulge forms the attachment for the AP muscle.3 It has also been shown that fibers of
From the Departments of Dermatology,a Pathology,b Orthopedics,c and Obstetrics/Gynecology,d University of Arkansas for Medical Sciences, Little Rock, and the Department of Physical Therapy, University of Central Arkansas, Conway.e Supported by a student fellowship grant from the American Dermatological Association (to M. M. C.). Accepted for publication Dec 9, 1999. Reprint requests: Thomas D. Horn, Chairman, Department of Dermatology, University of Arkansas for Medical Sciences, 4301 W Markham, Slot 576, Little Rock, AR 72205-7199. Copyright © 2000 by the American Academy of Dermatology, Inc. 0190-9622/2000/$12.00 + 0 16/1/105159 doi:10.1067/mjd.2000.105159
some AP muscles can be found admixed with the connective tissue sheath around the entire circumference of the hair follicle.4 The distal anatomy of the muscle and its attachments to the extracellular matrix (ECM) are less well characterized. Many texts schematically depict the distal AP as attaching as a single fascicle to the epidermal basement membrane. Narisawa, Hashimoto, and Kohda5 showed that most of the distal ends of the AP were actually situated in the upper dermis, whereas some fibers seemed to be in close contact with the basal layer of the epidermis. The mechanisms of attachment of the muscle to the ECM allowing hair erection upon contraction remain uncertain. Using immunohistochemistry6 Mendelson et al documented expression of integrin subunits α1, α5, and β1 by the distal AP muscle. Fibronectin, a known ligand of α5β1, was distributed in a pattern that roughly coincided with α5 distribution at the distal AP-ECM interface. Interaction between α5β1 and fibronectin occurs in other areas of the body and allows signal transduction and mechanoreception in these sites.7,8 From these results we hypothesized that the interaction between fibronectin and α5β1 integrin 19
20 Clifton et al
J AM ACAD DERMATOL JULY 2000
A
B
C Fig 1. Views of a distal arrector pili muscle bundle in the fluorescein wavelength (A), in the rhodamine wavelength (B), and using a dual filter (C) demonstrate co-localization of α5β1 and fibronectin at areas of contact between the AP and ECM.
mediates the attachment of the AP muscle to the ECM. The purpose of our study was to strengthen this hypothesis and better delineate the attachments and anatomy of this muscle to shed light on its contractile physiology. The study involved immunofluorescent double labeling for fibronectin and α5β1 integrin to better demonstrate the distal anatomy of the AP and its attachments to the epidermis and ECM.
MATERIAL AND METHODS Tissue samples Sections of normal fresh scalp skin were obtained from the occipital area of cadavers (because good examples of AP muscles in serial sections are hard to find, tissue from this study was only taken from the occipital area; AP muscles are more plentiful in this region than any other area of the body). The tissue was immediately cut into 1- × 3-mm strips and embedded in OCT. The embedded tissue was then frozen at –70°C. The blocks were cut into serial 5-µm vertical sections by cryostat and mounted on charged glass slides. The slides were stored frozen until immunofluorescent studies were performed.
Immunofluorescent technique Standard immunofluorescent techniques were used to achieve double-labeling. Serial incubations were performed with anti-α5β1 (at 1:20 for 90 minutes, DAKO, Carpinteria, Calif) followed by antifibronectin (at 1:500 for 90 minutes, DAKO) with separate steps to tag the primary antibodies with fluorescent labels between these incubations. The slides were then lightly counterstained in hematoxylin and Scott’s water blueing agent. Coverslips were applied with the use of mounting media containing a fixative to preserve fluorescence. Microscopy and reconstruction A dual light and fluorescent microscope mounted with filters for viewing fluorescein and rhodamine and a dual filter for viewing both the fluorescein and rhodamine wavelengths simultaneously were used to examine the sections. Photomicrographs demonstrating the staining of α5β1 with fluorescein, fibronectin with rhodamine, and areas of overlapping staining were obtained from the serial 5-µm sections. Photomicrographs demonstrating overlapping staining of fibronectin and α5β1 were used as templates to reconstruct a computerized
Clifton et al 21
J AM ACAD DERMATOL VOLUME 43, NUMBER 1, PART 1
Fig 2. This three-dimensional representation of the anatomy of the distal AP muscle was created by means of serial sections through a single muscle as templates. Figure shows the arrector pili muscle (AP), the hair follicle (HF), the sebaceous gland (SG), and the epidermis (Epi). (Courtesy of Jones Productions Inc, Little Rock, Ark.)
three-dimensional picture of the distal AP and its attachments. Statistical analysis To determine whether the immunofluorescent staining of α5β1 exhibited significant differences in relative signal density at the distal AP-ECM interface compared with the middle and proximal portions of the muscle, the signal density was measured at the distal, middle, and proximal portions of the muscle and corresponding areas of the ECM. These measurements were also taken in the same regions for fibronectin. With an area of 50 µm2, signal measurements from representative areas at each portion of the muscle and representative areas from each corresponding portion of the ECM were obtained and digitized. The digitized densities obtained from each part of the muscle were averaged. The ECM densities for each area were also averaged and used as control values for that corresponding area of the muscle. A percent change over control was determined for the distal, middle, and proximal portions of the muscle for both α5β1 and fibronectin. Specific interregional
comparisons were performed by means of one-way analysis of variance followed by the Scheffé test. The above measurements and calculations were done on serial sections of 3 separate AP muscles.
RESULTS Examination of the serial sections with light and fluorescent microscopy showed muscle fibers splaying into multiple branches, which terminated in areas of the papillary and reticular dermis as previously described. Some fibers could be seen approaching the epidermal basement membrane. Fluorescein staining, indicating expression of α5β1 integrin, was granular and more pronounced at the periphery of the AP bundles and on the distal fibers of the muscle (Fig 1, A). Rhodamine staining, indicating expression of fibronectin was seen in a smooth pattern around muscle bundles and at the interface of the distal fibers and the ECM and basement membrane (Fig 1, B). A dual filter allowed simultaneous viewing of areas of colocalization of staining for α5β1 and fibronectin. Dual staining, in orange, was seen at all points of contact between the AP muscle and the ECM (Fig 1, C).
22 Clifton et al
A
J AM ACAD DERMATOL JULY 2000
B Fig 3. Demonstration of the results of ANOVA analysis followed by the Scheffé test for α5β1 (A) and fibronectin (B). The y-axis represents the percent increase in staining intensity of the AP versus the ECM in all 3 areas measured. Staining for both epitopes at the distal AP fiber–ECM interface was statistically brighter than staining at either the middle or proximal portion of the muscle (P = .0001).
Photomicrographs of 25 serial 5-µm sections through the same muscle dually stained were used as templates to create a computerized three-dimensional picture of the distal anatomy of the AP and highlight the co-localized staining of fibronectin and α5β1 (Fig 2). Statistical analysis of the digitized densities showed significant increased density in staining for both α5β1 and fibronectin in areas of AP-ECM contact than in areas within the ECM. Density of staining for α5β1 at the distal fibers of the AP was statistically more dense than at either the middle or proximal areas of the muscle (P = .0001) (Fig 3, A). Density of staining for fibronectin at the distal AP-ECM interface was also statistically more dense than staining at the middle or proximal portion of the muscle (P = .0001) (Fig 3, B). There was no statistically significant difference in density of staining between the middle and proximal portions of the muscle for either α5β1 or fibronectin. With our three-dimensional reconstruction of the AP, measurements of each of the branches were taken and the surface area of each branch was determined by assuming that each branch was cylindrical in shape (using the equation for the surface area of a sphere, excluding the base: πr2 + 2πrh). A total surface area for the 7 branches was then obtained. This total surface area was compared with the surface area of a theoretical single fascicle AP. Measurements for this model were obtained by using an average value for the branches of our model. With this approach, the total surface area for our model was 1761.98 µm2, whereas the surface area of the single fascicle model was 254.34 µm2. Our model had 6.93 times more surface area contacting the ECM than the single fascicle model. This increased surface area roughly cor-
responded to the number of branches of the AP. The actual increase in surface area is likely much greater than calculated above because such calculation does not take into account the continued splaying of the branches into terminal fibers. We also wished to calculate the area over which the contractile force is spread. Because of AP muscle splaying based on the three-dimensional model, the area over which the force of contraction would be spread was circular and had an area of 2552 µm2 (using the formula for the area of a circle: πr2; with the radius of the area determined from the threedimensional reconstruction). The area over which force would be dispersed for a single fascicle model was estimated to be 7.1 µm2. With this strategy the AP transduces the force of contraction over 359 times more area.
DISCUSSION With immunofluorescent double-labeling, our results document a precise co-localization of α5β1 integrin and fibronectin at the AP-ECM interface. Furthermore, these epitopes concentrate along the distal splayed portion of the muscle when compared with more proximal segments. These results further support our hypothesis that the interaction of these two epitopes mediates anchorage of the AP to the ECM and the basement membrane. Serial sectioning and use of computer graphics provide a detailed representation of the distal anatomy of the AP. Based on many studies, fibronectin and α5β1 are known to be involved in a myriad of biologic processes, including mediating cell attachment and cell stabilization.9-13 Such an interaction can be found between α5β1 and fibronectin in various tis-
J AM ACAD DERMATOL VOLUME 43, NUMBER 1, PART 1
sues including cartilage7 and endometrium.14 Their interaction and expression can be up-regulated in transformed cells, including those of Rous sarcoma virus–induced tumors15 and colon carcinoma cells.16 Although fibronectin binds at least 5 other integrins, α5β1 integrin is thought to be the primary fibronectin receptor because of its abundance and high affinity compared with other integrins.13 The integrin binds fibronectin at the peptide sequence known as the RGD (Arg-Gly-Asp) sequence.17-19 The fibronectin binding sites involve both subunits of the integrin.13 The cytoplasmic domain of the β1 subunit binds to talin, vinculin, actin, and fibulin, which make up portions of the cytoskeleton.20 In studies of cultured chondrocytes the normal cellular hyperpolarization response to pressure was ablated when the α5β1 receptor was blocked.7 When mechanical stress was applied to capillary endothelial cells, focal adhesion involving α5β1 occurred, as did an increase in cytoskeletal stiffness.21,22 Our data suggest that the interaction between fibronectin and α5β1 integrin plays an important role in mechanoreception, signal transduction, and stabilization of the cytoskeleton. Such an interaction would be important at the interface between the AP and the ECM and might allow the force transduction and stabilization necessary for muscle contraction and subsequent piloerection. The co-localization of α5β1 and fibronectin was seen at virtually every point connecting the AP with the ECM or the basement membrane. This finding suggests that rather than a single attachment to the basement membrane, the muscle actually interacts with its surroundings via these epitopes at many points in the papillary and reticular dermis, and at the basement membrane. Knowing the precise anatomy of the distal AP via computerized modeling allowed further assessment of the AP. The significant splaying of the AP and the increased surface area of the distal fibers for contact with the ECM (×7) is logical considering the equation for stress: Stress = Force/Area. The splayed anatomy of the AP muscle spreads the force of contraction over a much larger surface area (×359). Therefore the anatomic stress of contraction would be greatly lessened with the same force of contraction. To further support our hypothesis that fibronectin and α5β1 integrin interact to allow force transduction necessary for AP muscle contraction, a model of inhibition of the interaction between these molecules would be ideal to determine whether piloerection is abolished. We thank Jones Productions of Little Rock for assistance with the three-dimensional reconstruction.
Clifton et al 23
REFERENCES 1. Rodrigo GF, Cotta-Pereira G, David-Ferreira JF.The fine structure of the elastic tendons of the arrector pili muscle. Br J Dermatol 1975;93:631-6. 2. Akiyama M, Dale BA, Sun T, Holbrook KA. Characterization of hair follicle bulge in human fetal skin: the human fetal bulge is a pool of undifferentiated keratinocytes. J Invest Dermatol 1995;105:844-50. 3. Elder D, Elenitsas R, Jaworsky C, Johnson B Jr. Histology of the skin. In: Lever’s Histopathology of the skin. 8th edition. Philadelphia: Lippincott-Raven; 1997. p. 19-23. 4. Narisawa Y, Kohda H. Arrector pili muscles surround human facial vellus hair follicles. Br J Dermatol 1993;129:138-9. 5. Narisawa Y, Hashimoto K, Kohda H. Merkel cells participate in the induction and alignment of arrector pili muscles of human fetal skin. Br J Dermatol 1996;134:494-8. 6. Mendelson JK, Smoller BR, Mendelson B, Horn TD. The microanatomy of the distal arrector pili: possible role for alpha1 beta1 and alpha5 beta1 integrins mediating cell-cell adhesion and anchorage to the extracellular matrix. J Cutan Pathol 2000;27:61-6. 7. Clark EA, Brugge JS. Integrins and signal transduction pathways: the road taken. Science 1995;268:233-9. 8. Wright MO, Nishida K, Bavington C, Godolphin JL, Dunne E, Walmsley S, et al. Hyperpolarisation of cultured human chondrocytes following cyclical pressure-induced strain: evidence of a role for α5β1 Integrin as a chondroctye mechanoreceptor. J Orthop Res 1997;15:742-7. 9. Hay ED, editor. Cell biology of extracellular matrix. 2nd ed. New York: Plenum Press; 1991. 10. Ruoslahti E, Pierschbacher MD. New perspectives in cell adhesion: RGD and integrins. Science 1987;238:491-7. 11. Mosher DF, editor. Fibronectin. New York: Academic Press; 1989. 12. Hynes RO. Fibronectins. New York: Springer-Verlag; 1989. 13. Yamada KM. Fibronectins: structure functions and receptors. Curr Opin Cell Biol 1990;1:956-63. 14. Albright CD, Carter CA, Kaufman DG. Tamoxifen alters the localization of F-actin and α5/β1-integrin fibronectin receptors in human endometrial stomal cells and carcinoma cells. Pathobiology 1997;65:177-83. 15. Saga S, Chen W, Yamada K. Enhanced fibronectin receptor expression in Rous sarcoma virus induced tumors. Cancer Res 1988;48:5510-3. 16. Varner JA, Emerson DA, Juliano RL. Integrin α5β1 expression negatively regulates cell growth: reversal by attachment to fibronectin. Mol Biol Cell 1995;6:725-40. 17. Pierschbacher MD, Rouslahti E. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature 1984;309:30-3. 18. Pierschbacher MD, Rouslahti E. Variants of the cell recognition site of fibronectin that retain attachment promoting activity. Proc Natl Acad Sci U S A 1984;81:5985-8. 19. Yamada KM, Kennedy DW. Dualistic nature of adhesive protein function: fibronectin and its biologically active peptide fragments can auto-inhibit fibronectin function. J Cell Biol 1984; 99:29-36. 20. Horwitz A, Duggan K, Buck C, Beckerle MC, Burridge K. Interaction of plasma membrane fibronectin receptor with talin: a transmembrane linkage. Nature 1986;320:531-2. 21. Akiyama SK, Yamada SS, Chen WT, Yamada KM. Analysis of fibronectin receptor function with monoclonal antibodies: roles in cell adhesion, migration, matrix assembly, and cytoskeletal organization. J Cell Biol 1989;109:863-75. 22. Wang N, Butler JP, Ingber DE. Mechanoreception across the cell surface and through the cytoskeleton. Science 1993;260:11247.