247
J. Photochem. Photobiol. B: Biol., 14 (1992) 247-259
UVA- and UVB-induced changes in collagen and fibronectin biosynthesis in the skin of hairless mice Benedicte
Boyera,
Annie
Fourtanierb,
Patrick
Kerr?
and Jacqueline
Labat-Robert”,+
aLaboratoire de Biologic du Tissu Conjonctif; UA CNRS 1460, Facultk de Mkdecine, Universite’ Paris XII, 8 rue du G&&al Sarrail, 94010 Criteil Cedar (France) bDkpatiernent de Biologie, Laboratoires de Recherche Fondamentale de I’Orkal, 93601 Aulnaysous-Bois Cedev (France) (Received
February
5, 1992; accepted
February
14, 1992)
Abstract The modifications induced in hairless mouse skin by chronic UV irradiation were investigated. Skin explant cultures were used to study UVA- and UVB-induced changes occurring in interstitial collagen (type I and type III) and fibronectin biosynthesis. To study the long-term effects, albino hairless mice were irradiated with UVA radiation alone from two sources with different spectral qualities or with UVB. UVA and UVB radiation produced a significant increase in the ratio of type III to type I collagen (more than 100% for WA-irradiated skin and about 60% for UVBirradiated skin) accompanied by a significantly increased fibronectin biosynthesis (50% or more in all irradiated groups). Irradiation with either UVA or UVB alone had no significant effect on the total collagen synthesis and resulted in only a slight decrease in the total collagen content of the skin determined as hydroxyproline. This decrease was significant only in the case of the group irradiated with UVA (xenon) (decrease of 25%, expressed as micrograms of hydroxyproline per milligram wet weight). A significant decrease in collagen hydroxylation (expressed as radioactive hydroxyproline/radioactive hydroxyproline plus proline in neosynthesized collagen) was observed of about 50% in skin irradiated with UVA (xenon) but not in UVB-treated skin. Several of the above modifications (increased fibronectin biosynthesis, increased collagen type III to type I ratio) correspond to the modifications observed during the aging of nonirradiated hairless mice. Therefore it appears that UV irradiation accelerates the modifications of extracellular matrix biosynthesis observed during aging.
Keywords:
UV light, collagens,
fibronectin,
biosynthesis.
1. Introduction Chronic exposure to sunlight and the use of artificial UV radiation for therapeutic treatment contribute to the photoaging of human skin [14]. It is probable that exposure to UV radiation can induce degenerative changes in the dermal connective tissue macromolecules, especially collagens [5-71. For example, it has been demonstrated that the ratio of type III to type I collagen increases during aging [8, 91 and as a ‘Author
to whom correspondence
loll-1344/92/$5.00
should be addressed.
0 1992 - Elsevier Sequoia. All rights reserved
248
result of UVB irradiation of skin [7]. In addition, we have previously demonstrated a parallel increase in fibronectin biosynthesis accompanying that of type III collagen [lo, 111. As fibronectin plays an important role in cell-matrix interactions, the observed modifications may well affect the structure and function of dermal connective tissue [12, 131. Therefore it appeared worthwile to investigate further the modifications produced by UVA and UVB irradiation in hairless mouse skin and to compare them with agedependent alterations. We focused our attention on fibrous collagen (types I and III) and fibronectin synthesis in the skin of hairless mice, used as a model for the study of UV-irradiation-induced tissue damage [5-71. In this paper, we report the effects of WA and UVB irradiation administered separately on interstitial collagen (types I and III) and fibronectin biosynthesis.
2. Materials
and methods
2.1. Animals and treatments The animals used in this study were MFl/hr OLAC hairless albino mice aged 8-12 weeks at the beginning of the experiments. They were housed individually during all the experiments. Each experimental group consisted of 18-20 females. 2.2. Radiation sources and schedules For WA irradiation, two different sources were used. The first was a xenon lamp (1000 W) mounted in a solar arc simulator. The collimated beam was passed through a water filter containing an MT0 IR filter (H 325a) and a 2 mm Schott WG 345 filter (50% cut-off at 345 nm). The second was a 3000 W medium pressure mercury vapour lamp containing argon and metal halides (UVASUN-MUTZAS, Miinich, Germany). At skin level, in these conditions, the irradiance between 340 and 400 nm corresponds to 94% of the total amount of UVA. For UVB, a bank of five Philips TL 20/12 tubes was used. The three different spectra of the above-mentioned lamps are given in Fig. 1. They were obtained with a spectroradiometer Macam (Livingston, UK) at skin level. The irradiation schedules are summarized in Table 1. 2.3. Investigations
At the end of the irradiation period, after clinical examination, the animals were killed by cervical dislocation and samples were taken from dorsal skin for histology and biochemical analysis, excluding areas with tumours from the investigated samples. Every animal (16 in most experiments) was treated individually. The results given represent the average of up to 16 determinations 5 the standard error of the mean (SEM). 2.4. Histology For each animal 6 mm biopsies were used. One was fixed in formalin, embedded in paraffin and stained with haematoylin-phloxine-saffron. Luna stain was used for elastin, and Mow~y stain combined with Van Gieson for glycosaminoglycans and collagen. 2.5. Determination of free amino acid pool size This was carried out for proline and methionine in control and UVB-irradiated mouse skin as follows. Excised skin samples were homogenized in 1% sodium do-
249
400
PHILIPS
280
300
320
340
360
360
Wavelength
(nm)
tubes TL SO/l2
400
Wavelength
( nm )
Fig. 1. Spectral irradiance at skin level of the three different UV sources used.
decylsulphate (SDS) in an Ultra-turrax homogenizer and centrifuged at 10 000 Xg for 15 min. Aliquots of the supematants (including some SDS-soluble proteins) were hydrolysed and derivatized with phenylisothiocyanate, analysed by reverse-phase high performance liquid chromatography (HPLC) (Waters PICOTAG system) and finally evaluated for proline and methionine concentrations.
“Initial dose (0.5 MED) increased
320-400 340-400 280-350
(nm)
Wavelength
30 12 a
Daily (min) 3 3 5
Weekly (days) 52 52 12
Total (weeks)
Periodic&y of irradiation
of hairless mice
by 20% every week. 1 MED =40 mJ cm-*.
Xenon lamp + WG 345 UVASUN 3000 Philips tubes
UVA low dose UVA high dose UVB
equipment
Irradiation
of irradiation
Group
Sources of UV light and schedules
TABLE 1
20 50 0.34
Irradiance at skin level (mW cm-‘)
5460 5460 4
Total dose of UV received (J cm-‘)
14-15 14-15 5-6
Age at the end of irradiation (months)
251
2.6. Collagen biosynthesis Weighed samples of intact skin were hydrolysed directly in 6 N HCl at 105 “C determination using a for 24 h to evaluate total skin collagen by hydroxyproline calorimetric method 1141. The weighed tissues were cut into small pieces in sterile conditions, put in 10 cm diameter Petri dishes and incubated in 15 ml of Dulbecco’s modified Eagle’s medium supplemented with 10% foetal calf serum, streptomycin (100 pg ml-‘), penicillin (100 U ml-‘) and ascorbic acid (50 pg ml-‘) and labelled with ~-[4,5-~H] proline (10 @Zi ml-‘; 49 Ci mmol-‘, Commissariat ?I 1’Energie Atomique, Saclay, France) [S, 15, 161. Incubations were carried out in a 37 “C incubator in air-CO2 (19:1, v/v) for 24 h. The incubation was terminated by cooling the mixture quickly at 4 “C. The medium did not contain measurable amounts of collagen and was discarded and the tissues were extensively washed with phosphate-buffered saline (PBS). 2.6.1. Extraction procedure The tissues were homogenized in an ultra-Turrax homogenizer and submitted to limited pepsin digestion. Briefly, pepsin (E.C. 3.4.23.1, Sigma, 3200 U per mg protein) was added to the tissue suspension in 0.5 M acetic acid with a pepsin to collagen ratio of 1:lO. The mixture was stirred for 2-3 days at 4 “C. Solubilized material was submitted to salt precipitation at acid pH by dialysis against 0.5 M acetic acid containing 0.7 M NaCl. The precipitate was dialysed and lyophilised [8, 151. Total collagen biosynthesis was compared with protein synthesis using the ratio of radioactivity incorporated in pepsin-solubilized collagen (up to 90% of skin collagen was solubilized by this method [15, 171) to the total radioactivity present in the tissue before pepsin treatment. Aliquots of pepsin-solubilized collagen were hydrolysed to determine the quantities and specific activities of proline and hydroxyproline by a modification of the technique of Rojkind in order to give an estimation of prolyl hydroxylation [14, 18, 191. No attempt was made to correct for the loss of isotope during oxidation by chloramine T, and thus the values for prolyl hydroxylation are relative. 2.6.2. Determination of collagen phenotype Electrophoretic separation of collagen (Ychains was performed according to Laemmli on 7.5% SDS polyacrylamide gels using interrupted electrophoresis (SDS PAGE) [20, 211. After staining with Coomassie blue, the gel was analysed by densitometry. In some cases pepsin-solubilized material was treated by bacterial collagenase, protease free (Calbiochem, USA) before electrophoresis 1221. 2.7. Fibronectin biosynthesis Dorsal skin was minced into pieces of about 1 mm3 and weighed fresh skin samples were incubated at 37 “C for 24 h in 5 ml of Dulbecco’s modified Eagle’s medium containing methionine at 10% of its normal level, bovine serum albumin at 0.2% and antibiotics as described for collagen. Tissues were labelled with L-[~?S] methionine (50 &i ml-‘; specific activity, 1426 Ci mmol-‘; Amersham, France). After incubation, tissues were collected and washed extensively, and homogenized in boiling 1% SDS. After centrifugation at 10 OOOXg for 15 min, the supernatants were dialysed against PBS containing protease inhibitors: phenylmethylsulphonyl fluoride, sodium iodoacetate, N-ethylmaleimide and ethylenediaminetetraacetic acid (EDTA) at 2mM. The SDS extracts were submitted to radioactivity determinations. Fibronectin was then immunoprecipitated from the SDS extracts as described previously [lo, 11,231. Radiolabelling of immunoprecipitated fibronectin was expressed as a percentage of the total radioactivity present in SDS extracts.
252
2.8. Statistical analysis All statistical analyses were performed
using the Student
t test.
3. Results 3.1. Clinical data 3.1.1. WA irradiation At the end of the 12 months of UVA irradiation, the group irradiated with the xenon and UVASUN 3000 lamps showed similar changes: wrinkling, loss of skin colouration and sagging on the neck and along the flanks. The only difference between the two groups was the degree of scaling. In the group irradiated with the xenon lamp scaling was more severe. In each irradiated group, four animals of the 16 alive at the end of the irradiations showed small papillomas. The control animals, aged 15 months, presented a slight thickening of the skin and loss of colouration. No wrinkling or scaling and no lesions were observed. 3.1.2. UK!3 irradiation At the end of the 3 months of WB irradiation, a mild to moderate erythema was observed with scaling and palpable keratosis. Several, but not all, of the animals developed visible wrinkling. No tumours (1 mm diameter or greater) were recorded. The control animals presented a fine purple-pink skin. 3.2. Histological data 3.2.1. WA irradiation Only very slight changes were observed in both groups irradiated with WA. In addition, they were much more variable than those occurring with UVB exposure. In the dermis, they consisted of a slight loss of the fine perpendicular sub-epidermal fibres of elastin, focal deposition of glycosaminoglycans, mostly at the dermo-epidermal junction and an increase in dermal cyst size and number. No tinctorial changes in the staining of collagen were observed in WA-exposed animals. The epidermis did not show any acanthosis. The histopathology of lesions confirmed the clinical assessment. All the lesions, except one, were squamous papillomas or acantho papillomas. One of the lesions of a mouse irradiated with xenon was invasive squamous cell carcinoma. 3.2.2. UYB irradiation After 12 weeks of UVB irradiation, the epidermis was thick (hyperplasia) and the dermis showed evident modifications. The elastic fibres, visualized with Luna staining, were more numerous, thicker and curled; they proliferated throughout the upper dermis. There was some loss of staining of collagen and a definite increase in glycosaminoglycan deposition (Mowry’s stain) mainly near the dermo-epidermal junction. Haematoxylin and eosin-stained sections showed a hyperproliferation of dermal cysts. 3.3. Biochemical data 3.3.1. Amino acid pool size determination The pool size of proline and methionine was determined in control and UVBirradiated mouse skin. For proline, a pool size of 358.59 X lo-” mol (g wet weight)-’ (fl2.61) was found in the controls and a pool size of 338.35X1O-‘o mol (g wet weight)-’ (-f 23.22) was found in UVB-irradiated mouse skin. For methionine, a pool size of 121.7~ lo-lo mol (g wet weight)-’ (k5.57) was found in the controls and a
253
pool size of 103.95 x lo-” mol (g wet weight)-’ (* 3.91) was found in UVB-irradiated mouse skin. None of these variations was significant. 3.3.2. Collagen content and biosynthesis The effect of UV irradiation on the total collagen of hairless mouse skin, as determined from the hydroxyproline content of intact skin, is shown in Fig. 2. The heights of the columns indicate hydroxyproline contents expressed as micrograms per milligram wet weight. Both UVA and UVB produced a decrease which was significant for UVA irradiation with the xenon lamp and for UVB irradiation. Pepsin treatment of skin samples solubilized up to 90% of the total collagen in all the experimental groups. The effect of UV irradiation on the incorporation of [3H]-proline in pepsin-soluble collagens (compared with the total amount of radiolabclled protein) in mouse skin is shown in Table 2. No significant modifications of the percentage of collagen biosynthesis were observed with UVA or UVI3 irradiation. 3.3.3. Collagen hydroxylation and typing The pepsin-soluble collagens were precipitated at 0.7 M NaCl and analysed by SDS PAGE. Figure 3 shows the scanning of typical gels. Under reducing conditions the typical patterns of type I and type III collagens were obtained. Furthermore, all detectable pepsin-solubilized material was collagenase sensitive. The effect of UV irradiation on the hydroxylation of collagens is shown in Fig. 4. Both UVA sources produced a decrease in proline hydroxylation which was significant only with the xenon lamp. UVB irradiation had no such effect. The effect of UV irradiation on the ratio of collagen type III to collagen types I plus III is shown in Fig. 5. The type III to type I plus type III ratio increased significantly both with UVA and UVB, by 120%, 115% and 60% for the xenon lamp, UVASUN 3000 and UVB irradiation respectively. 3.3.4. Fibronectin biosynthesis The effect of UV irradiation on fibronectin biosynthesis is shown in Fig. 6. Both UVA and UVB produced a significant increase in [35S]-methionine incorporation in
UVA
UVB
8 6
**
4
2
0I
Fig. 2. Effect of irradiation with UVA or UVB on the total collagen content of hairless mouse skin. Total skin collagen was estimated by hydroxyproline determination as described in Section 2. The results are the mean ( f SEM) of 16 determinations. *significance, p < 0.01; **significance p
6.15 ( + 0.30)
7.34 ( f 0.28)
7.71 ( * 0.21)
A (c.p.m.
Percentage of collagen biosynthesis (A/B x 100)
35.5
36.4
35.5
44.1
3.32
43.1
c.p.m., counts per minute. immunoprecipitable fibronectin of the order of 60%-70%, 50% and 50% for the xenon lamp, UVASUN 3000 and UVB irradiations respectively.
4. Discussion Histologically, the modifications observed in chronically irradiated skins are more clear cut with WB-treated mice than with UVA-treated mice, whatever the source used. This confirms the results of Kligman and Kligman [24] and Bissett et al. [25] who reported similar findings. More detailed information on the UVA- and UVBinduced changes in dermal extracellular matrix was obtained however, by the investigation of the composition and biosynthesis of extracellular matrix macromolecules such as collagen type I and type III and fibronectin. Several conflicting results have been published on the modification of the composition of the extracellular matrix of hairless mouse skin after W irradiation. Kligman et al. [26] found an important increase in the collagen type III to I plus III ratio after 24 weeks of irradiation with a somewhat less dramatic increase at 8 weeks. Plastow et al. [7] also found an increase in collagen type III after 12-36 weeks of irradiation. Schwartz et al. [27] recently studied the aminopropeptide distribution of collagen type III and found a reduction by immunoblot after S-10 weeks of irradiation. In earlier work, Schwartz [28] also found an increased fibronectin content in irradiated mouse skin extracts. Because of the difficulties of interpretation of static data and the variability of extraction procedures, we preferred to use the study of the biosynthesis of matrix macromolecules using radioactive precursors. In this study, W-irradiated hairless albino mice developed alterations of the extracellular matrix of the skin. More specifically, irradiation with both UVA and WB significantly increased the proportion of type III to type I plus III collagen. This modification was accompanied by a parallel increase in fibronectin biosynthesis. Similar effects have been observed with hairless mice exposed to UVB and with sun-exposed human dermis [7, 291. Such parallel alterations of both type III collagen and fibronectin biosynthesis have been described during the aging of connective tissues and in some
255 b
/
Ji
a
L MIGRATION
Fig. 3. Typical densitometric scans of SDS PAGE. Pepsin digests of skin were submitted to SDS PAGE as described in Section 2: A, controls; B, xenon lamp + WG 345; C, UVASUN 3000 lamp. Migrations of collagen P chains were as follows: 1, a2 (I); 2, al (I); 3, al (III).
pathologies, for example, diabetes type II [S, 11, 151, which imitates an accelerated aging process of connective tissues. UV-induced modifications seem to follow a similar pattern of extracellular matrix alterations. In the present study, a moderate decrease in total collagen was noted in irradiated mice, but this was only significant using the xenon lamp and WB irradiation. For skin treated with the xenon lamp, this decrease in hydroxyproline content could be correlated with an alteration of prolyl hydroxylation (Fig. 4 and discussion below). This effect was not observed with the shortest period of irradiation described by Johnston et al. [S] (12 treatments, 1 J cm-* in 6.25 min, on alternate days with eight Sylvania fluorescent tubes F40BLB Lifeline). An intriguing observation was the decrease in hydroxylation of proline in WA-irradiated animals contrasting with the absence of such an effect with UVB irradiation. Johnston et al. [S], who measured the activity of proline hydroxylase, also indicated a decrease in the activity of this enzyme in WA-treated mouse skin without alterations with UVB irradiation. These workers used long-wavelength WA. Despite the differences in the spectra and irradiance of our UVA sources, we found a decrease in collagen hydroxylation with both WA
UVA UVB
Fig. 4. Effect of irradiation with UVA or UVB on the hydroxylation of collagen in hairless mouse skin. Incorporation of [3H]-proline into hydroxyproline and proline of pepsin-solubilized collagen was determined as described in Section 2. The results are expressed as the percentage of radioactivity incorporated in hydroxy-[3H]-proline relative to the total incorporated radioactivity (hydroxy-[3H]-proline + 13H]-proline). No attempt was made to correct data for the loss of isotope during the oxidation process in the Rojkind method. Data are the mean (+SEM) of 16 determinations. *Significance, p < 0.02 in comparison with unirradiated age-matched controls.
UVA 40 8
UVB
1
30 -I
L
3
10
0
Fig. 5. Effect of irradiation with UVA or UVB on the proportion of collagen type III in hairless mouse skin. The relative proportion of collagen type III was determined according to densitometric analysis of SDS polyacrylamide gels as described in Section 2. The data are the mean (&SEM) of 16 determinations. *Significance, p
J. Photochem. Photobiol. B: Biol., 14 (1992) 247-259
UVA- and UVB-induced changes in collagen and fibronectin biosynthesis in the skin of hairless mice Benedicte
Boyera,
Annie
Fourtanierb,
Patrick
Kerr?
and Jacqueline
Labat-Robert”,+
aLaboratoire de Biologic du Tissu Conjonctif; UA CNRS 1460, Facultk de Mkdecine, Universite’ Paris XII, 8 rue du G&&al Sarrail, 94010 Criteil Cedar (France) bDkpatiernent de Biologie, Laboratoires de Recherche Fondamentale de I’Orkal, 93601 Aulnaysous-Bois Cedev (France) (Received
February
5, 1992; accepted
February
14, 1992)
Abstract The modifications induced in hairless mouse skin by chronic UV irradiation were investigated. Skin explant cultures were used to study UVA- and UVB-induced changes occurring in interstitial collagen (type I and type III) and fibronectin biosynthesis. To study the long-term effects, albino hairless mice were irradiated with UVA radiation alone from two sources with different spectral qualities or with UVB. UVA and UVB radiation produced a significant increase in the ratio of type III to type I collagen (more than 100% for WA-irradiated skin and about 60% for UVBirradiated skin) accompanied by a significantly increased fibronectin biosynthesis (50% or more in all irradiated groups). Irradiation with either UVA or UVB alone had no significant effect on the total collagen synthesis and resulted in only a slight decrease in the total collagen content of the skin determined as hydroxyproline. This decrease was significant only in the case of the group irradiated with UVA (xenon) (decrease of 25%, expressed as micrograms of hydroxyproline per milligram wet weight). A significant decrease in collagen hydroxylation (expressed as radioactive hydroxyproline/radioactive hydroxyproline plus proline in neosynthesized collagen) was observed of about 50% in skin irradiated with UVA (xenon) but not in UVB-treated skin. Several of the above modifications (increased fibronectin biosynthesis, increased collagen type III to type I ratio) correspond to the modifications observed during the aging of nonirradiated hairless mice. Therefore it appears that UV irradiation accelerates the modifications of extracellular matrix biosynthesis observed during aging.
Keywords:
UV light, collagens,
fibronectin,
biosynthesis.
1. Introduction Chronic exposure to sunlight and the use of artificial UV radiation for therapeutic treatment contribute to the photoaging of human skin [14]. It is probable that exposure to UV radiation can induce degenerative changes in the dermal connective tissue macromolecules, especially collagens [5-71. For example, it has been demonstrated that the ratio of type III to type I collagen increases during aging [8, 91 and as a ‘Author
to whom correspondence
loll-1344/92/$5.00
should be addressed.
0 1992 - Elsevier Sequoia. All rights reserved
248
result of UVB irradiation of skin [7]. In addition, we have previously demonstrated a parallel increase in fibronectin biosynthesis accompanying that of type III collagen [lo, 111. As fibronectin plays an important role in cell-matrix interactions, the observed modifications may well affect the structure and function of dermal connective tissue [12, 131. Therefore it appeared worthwile to investigate further the modifications produced by UVA and UVB irradiation in hairless mouse skin and to compare them with agedependent alterations. We focused our attention on fibrous collagen (types I and III) and fibronectin synthesis in the skin of hairless mice, used as a model for the study of UV-irradiation-induced tissue damage [5-71. In this paper, we report the effects of WA and UVB irradiation administered separately on interstitial collagen (types I and III) and fibronectin biosynthesis.
2. Materials
and methods
2.1. Animals and treatments The animals used in this study were MFl/hr OLAC hairless albino mice aged 8-12 weeks at the beginning of the experiments. They were housed individually during all the experiments. Each experimental group consisted of 18-20 females. 2.2. Radiation sources and schedules For WA irradiation, two different sources were used. The first was a xenon lamp (1000 W) mounted in a solar arc simulator. The collimated beam was passed through a water filter containing an MT0 IR filter (H 325a) and a 2 mm Schott WG 345 filter (50% cut-off at 345 nm). The second was a 3000 W medium pressure mercury vapour lamp containing argon and metal halides (UVASUN-MUTZAS, Miinich, Germany). At skin level, in these conditions, the irradiance between 340 and 400 nm corresponds to 94% of the total amount of UVA. For UVB, a bank of five Philips TL 20/12 tubes was used. The three different spectra of the above-mentioned lamps are given in Fig. 1. They were obtained with a spectroradiometer Macam (Livingston, UK) at skin level. The irradiation schedules are summarized in Table 1. 2.3. Investigations
At the end of the irradiation period, after clinical examination, the animals were killed by cervical dislocation and samples were taken from dorsal skin for histology and biochemical analysis, excluding areas with tumours from the investigated samples. Every animal (16 in most experiments) was treated individually. The results given represent the average of up to 16 determinations 5 the standard error of the mean (SEM). 2.4. Histology For each animal 6 mm biopsies were used. One was fixed in formalin, embedded in paraffin and stained with haematoylin-phloxine-saffron. Luna stain was used for elastin, and Mow~y stain combined with Van Gieson for glycosaminoglycans and collagen. 2.5. Determination of free amino acid pool size This was carried out for proline and methionine in control and UVB-irradiated mouse skin as follows. Excised skin samples were homogenized in 1% sodium do-
249
400
PHILIPS
280
300
320
340
360
360
Wavelength
(nm)
tubes TL SO/l2
400
Wavelength
( nm )
Fig. 1. Spectral irradiance at skin level of the three different UV sources used.
decylsulphate (SDS) in an Ultra-turrax homogenizer and centrifuged at 10 000 Xg for 15 min. Aliquots of the supematants (including some SDS-soluble proteins) were hydrolysed and derivatized with phenylisothiocyanate, analysed by reverse-phase high performance liquid chromatography (HPLC) (Waters PICOTAG system) and finally evaluated for proline and methionine concentrations.
“Initial dose (0.5 MED) increased
320-400 340-400 280-350
(nm)
Wavelength
30 12 a
Daily (min) 3 3 5
Weekly (days) 52 52 12
Total (weeks)
Periodic&y of irradiation
of hairless mice
by 20% every week. 1 MED =40 mJ cm-*.
Xenon lamp + WG 345 UVASUN 3000 Philips tubes
UVA low dose UVA high dose UVB
equipment
Irradiation
of irradiation
Group
Sources of UV light and schedules
TABLE 1
20 50 0.34
Irradiance at skin level (mW cm-‘)
5460 5460 4
Total dose of UV received (J cm-‘)
14-15 14-15 5-6
Age at the end of irradiation (months)
251
2.6. Collagen biosynthesis Weighed samples of intact skin were hydrolysed directly in 6 N HCl at 105 “C determination using a for 24 h to evaluate total skin collagen by hydroxyproline calorimetric method 1141. The weighed tissues were cut into small pieces in sterile conditions, put in 10 cm diameter Petri dishes and incubated in 15 ml of Dulbecco’s modified Eagle’s medium supplemented with 10% foetal calf serum, streptomycin (100 pg ml-‘), penicillin (100 U ml-‘) and ascorbic acid (50 pg ml-‘) and labelled with ~-[4,5-~H] proline (10 @Zi ml-‘; 49 Ci mmol-‘, Commissariat ?I 1’Energie Atomique, Saclay, France) [S, 15, 161. Incubations were carried out in a 37 “C incubator in air-CO2 (19:1, v/v) for 24 h. The incubation was terminated by cooling the mixture quickly at 4 “C. The medium did not contain measurable amounts of collagen and was discarded and the tissues were extensively washed with phosphate-buffered saline (PBS). 2.6.1. Extraction procedure The tissues were homogenized in an ultra-Turrax homogenizer and submitted to limited pepsin digestion. Briefly, pepsin (E.C. 3.4.23.1, Sigma, 3200 U per mg protein) was added to the tissue suspension in 0.5 M acetic acid with a pepsin to collagen ratio of 1:lO. The mixture was stirred for 2-3 days at 4 “C. Solubilized material was submitted to salt precipitation at acid pH by dialysis against 0.5 M acetic acid containing 0.7 M NaCl. The precipitate was dialysed and lyophilised [8, 151. Total collagen biosynthesis was compared with protein synthesis using the ratio of radioactivity incorporated in pepsin-solubilized collagen (up to 90% of skin collagen was solubilized by this method [15, 171) to the total radioactivity present in the tissue before pepsin treatment. Aliquots of pepsin-solubilized collagen were hydrolysed to determine the quantities and specific activities of proline and hydroxyproline by a modification of the technique of Rojkind in order to give an estimation of prolyl hydroxylation [14, 18, 191. No attempt was made to correct for the loss of isotope during oxidation by chloramine T, and thus the values for prolyl hydroxylation are relative. 2.6.2. Determination of collagen phenotype Electrophoretic separation of collagen (Ychains was performed according to Laemmli on 7.5% SDS polyacrylamide gels using interrupted electrophoresis (SDS PAGE) [20, 211. After staining with Coomassie blue, the gel was analysed by densitometry. In some cases pepsin-solubilized material was treated by bacterial collagenase, protease free (Calbiochem, USA) before electrophoresis 1221. 2.7. Fibronectin biosynthesis Dorsal skin was minced into pieces of about 1 mm3 and weighed fresh skin samples were incubated at 37 “C for 24 h in 5 ml of Dulbecco’s modified Eagle’s medium containing methionine at 10% of its normal level, bovine serum albumin at 0.2% and antibiotics as described for collagen. Tissues were labelled with L-[~?S] methionine (50 &i ml-‘; specific activity, 1426 Ci mmol-‘; Amersham, France). After incubation, tissues were collected and washed extensively, and homogenized in boiling 1% SDS. After centrifugation at 10 OOOXg for 15 min, the supernatants were dialysed against PBS containing protease inhibitors: phenylmethylsulphonyl fluoride, sodium iodoacetate, N-ethylmaleimide and ethylenediaminetetraacetic acid (EDTA) at 2mM. The SDS extracts were submitted to radioactivity determinations. Fibronectin was then immunoprecipitated from the SDS extracts as described previously [lo, 11,231. Radiolabelling of immunoprecipitated fibronectin was expressed as a percentage of the total radioactivity present in SDS extracts.
252
2.8. Statistical analysis All statistical analyses were performed
using the Student
t test.
3. Results 3.1. Clinical data 3.1.1. WA irradiation At the end of the 12 months of UVA irradiation, the group irradiated with the xenon and UVASUN 3000 lamps showed similar changes: wrinkling, loss of skin colouration and sagging on the neck and along the flanks. The only difference between the two groups was the degree of scaling. In the group irradiated with the xenon lamp scaling was more severe. In each irradiated group, four animals of the 16 alive at the end of the irradiations showed small papillomas. The control animals, aged 15 months, presented a slight thickening of the skin and loss of colouration. No wrinkling or scaling and no lesions were observed. 3.1.2. UK!3 irradiation At the end of the 3 months of WB irradiation, a mild to moderate erythema was observed with scaling and palpable keratosis. Several, but not all, of the animals developed visible wrinkling. No tumours (1 mm diameter or greater) were recorded. The control animals presented a fine purple-pink skin. 3.2. Histological data 3.2.1. WA irradiation Only very slight changes were observed in both groups irradiated with WA. In addition, they were much more variable than those occurring with UVB exposure. In the dermis, they consisted of a slight loss of the fine perpendicular sub-epidermal fibres of elastin, focal deposition of glycosaminoglycans, mostly at the dermo-epidermal junction and an increase in dermal cyst size and number. No tinctorial changes in the staining of collagen were observed in WA-exposed animals. The epidermis did not show any acanthosis. The histopathology of lesions confirmed the clinical assessment. All the lesions, except one, were squamous papillomas or acantho papillomas. One of the lesions of a mouse irradiated with xenon was invasive squamous cell carcinoma. 3.2.2. UYB irradiation After 12 weeks of UVB irradiation, the epidermis was thick (hyperplasia) and the dermis showed evident modifications. The elastic fibres, visualized with Luna staining, were more numerous, thicker and curled; they proliferated throughout the upper dermis. There was some loss of staining of collagen and a definite increase in glycosaminoglycan deposition (Mowry’s stain) mainly near the dermo-epidermal junction. Haematoxylin and eosin-stained sections showed a hyperproliferation of dermal cysts. 3.3. Biochemical data 3.3.1. Amino acid pool size determination The pool size of proline and methionine was determined in control and UVBirradiated mouse skin. For proline, a pool size of 358.59 X lo-” mol (g wet weight)-’ (fl2.61) was found in the controls and a pool size of 338.35X1O-‘o mol (g wet weight)-’ (-f 23.22) was found in UVB-irradiated mouse skin. For methionine, a pool size of 121.7~ lo-lo mol (g wet weight)-’ (k5.57) was found in the controls and a
253
pool size of 103.95 x lo-” mol (g wet weight)-’ (* 3.91) was found in UVB-irradiated mouse skin. None of these variations was significant. 3.3.2. Collagen content and biosynthesis The effect of UV irradiation on the total collagen of hairless mouse skin, as determined from the hydroxyproline content of intact skin, is shown in Fig. 2. The heights of the columns indicate hydroxyproline contents expressed as micrograms per milligram wet weight. Both UVA and UVB produced a decrease which was significant for UVA irradiation with the xenon lamp and for UVB irradiation. Pepsin treatment of skin samples solubilized up to 90% of the total collagen in all the experimental groups. The effect of UV irradiation on the incorporation of [3H]-proline in pepsin-soluble collagens (compared with the total amount of radiolabclled protein) in mouse skin is shown in Table 2. No significant modifications of the percentage of collagen biosynthesis were observed with UVA or UVI3 irradiation. 3.3.3. Collagen hydroxylation and typing The pepsin-soluble collagens were precipitated at 0.7 M NaCl and analysed by SDS PAGE. Figure 3 shows the scanning of typical gels. Under reducing conditions the typical patterns of type I and type III collagens were obtained. Furthermore, all detectable pepsin-solubilized material was collagenase sensitive. The effect of UV irradiation on the hydroxylation of collagens is shown in Fig. 4. Both UVA sources produced a decrease in proline hydroxylation which was significant only with the xenon lamp. UVB irradiation had no such effect. The effect of UV irradiation on the ratio of collagen type III to collagen types I plus III is shown in Fig. 5. The type III to type I plus type III ratio increased significantly both with UVA and UVB, by 120%, 115% and 60% for the xenon lamp, UVASUN 3000 and UVB irradiation respectively. 3.3.4. Fibronectin biosynthesis The effect of UV irradiation on fibronectin biosynthesis is shown in Fig. 6. Both UVA and UVB produced a significant increase in [35S]-methionine incorporation in
UVA
UVB
8 6
**
4
2
0I
Fig. 2. Effect of irradiation with UVA or UVB on the total collagen content of hairless mouse skin. Total skin collagen was estimated by hydroxyproline determination as described in Section 2. The results are the mean ( f SEM) of 16 determinations. *significance, p < 0.01; **significance p
6.15 ( + 0.30)
7.34 ( f 0.28)
7.71 ( * 0.21)
A (c.p.m.
Percentage of collagen biosynthesis (A/B x 100)
35.5
36.4
35.5
44.1
3.32
43.1
c.p.m., counts per minute. immunoprecipitable fibronectin of the order of 60%-70%, 50% and 50% for the xenon lamp, UVASUN 3000 and UVB irradiations respectively.
4. Discussion Histologically, the modifications observed in chronically irradiated skins are more clear cut with WB-treated mice than with UVA-treated mice, whatever the source used. This confirms the results of Kligman and Kligman [24] and Bissett et al. [25] who reported similar findings. More detailed information on the UVA- and UVBinduced changes in dermal extracellular matrix was obtained however, by the investigation of the composition and biosynthesis of extracellular matrix macromolecules such as collagen type I and type III and fibronectin. Several conflicting results have been published on the modification of the composition of the extracellular matrix of hairless mouse skin after W irradiation. Kligman et al. [26] found an important increase in the collagen type III to I plus III ratio after 24 weeks of irradiation with a somewhat less dramatic increase at 8 weeks. Plastow et al. [7] also found an increase in collagen type III after 12-36 weeks of irradiation. Schwartz et al. [27] recently studied the aminopropeptide distribution of collagen type III and found a reduction by immunoblot after S-10 weeks of irradiation. In earlier work, Schwartz [28] also found an increased fibronectin content in irradiated mouse skin extracts. Because of the difficulties of interpretation of static data and the variability of extraction procedures, we preferred to use the study of the biosynthesis of matrix macromolecules using radioactive precursors. In this study, W-irradiated hairless albino mice developed alterations of the extracellular matrix of the skin. More specifically, irradiation with both UVA and WB significantly increased the proportion of type III to type I plus III collagen. This modification was accompanied by a parallel increase in fibronectin biosynthesis. Similar effects have been observed with hairless mice exposed to UVB and with sun-exposed human dermis [7, 291. Such parallel alterations of both type III collagen and fibronectin biosynthesis have been described during the aging of connective tissues and in some
255 b
/
Ji
a
L MIGRATION
Fig. 3. Typical densitometric scans of SDS PAGE. Pepsin digests of skin were submitted to SDS PAGE as described in Section 2: A, controls; B, xenon lamp + WG 345; C, UVASUN 3000 lamp. Migrations of collagen P chains were as follows: 1, a2 (I); 2, al (I); 3, al (III).
pathologies, for example, diabetes type II [S, 11, 151, which imitates an accelerated aging process of connective tissues. UV-induced modifications seem to follow a similar pattern of extracellular matrix alterations. In the present study, a moderate decrease in total collagen was noted in irradiated mice, but this was only significant using the xenon lamp and WB irradiation. For skin treated with the xenon lamp, this decrease in hydroxyproline content could be correlated with an alteration of prolyl hydroxylation (Fig. 4 and discussion below). This effect was not observed with the shortest period of irradiation described by Johnston et al. [S] (12 treatments, 1 J cm-* in 6.25 min, on alternate days with eight Sylvania fluorescent tubes F40BLB Lifeline). An intriguing observation was the decrease in hydroxylation of proline in WA-irradiated animals contrasting with the absence of such an effect with UVB irradiation. Johnston et al. [S], who measured the activity of proline hydroxylase, also indicated a decrease in the activity of this enzyme in WA-treated mouse skin without alterations with UVB irradiation. These workers used long-wavelength WA. Despite the differences in the spectra and irradiance of our UVA sources, we found a decrease in collagen hydroxylation with both WA
UVA UVB
Fig. 4. Effect of irradiation with UVA or UVB on the hydroxylation of collagen in hairless mouse skin. Incorporation of [3H]-proline into hydroxyproline and proline of pepsin-solubilized collagen was determined as described in Section 2. The results are expressed as the percentage of radioactivity incorporated in hydroxy-[3H]-proline relative to the total incorporated radioactivity (hydroxy-[3H]-proline + 13H]-proline). No attempt was made to correct data for the loss of isotope during the oxidation process in the Rojkind method. Data are the mean (+SEM) of 16 determinations. *Significance, p < 0.02 in comparison with unirradiated age-matched controls.
UVA 40 8
UVB
1
30 -I
L
3
10
0
Fig. 5. Effect of irradiation with UVA or UVB on the proportion of collagen type III in hairless mouse skin. The relative proportion of collagen type III was determined according to densitometric analysis of SDS polyacrylamide gels as described in Section 2. The data are the mean (&SEM) of 16 determinations. *Significance, p
