BROAD-SPECTRUM SUNSCREENS 131 Hillside, NJ) was fitted with a 1-mm WG 320 Schott filter to provide the summer noontime UV spectrum for 70 ø solar altitude. The collimated beam was passed through a water filter before impinging on a 45 ø UV-reflecting dichroic mirror. Irradiance at dorsal skin level (16 inches from source) was approximately .042 m/Wcm 2 UVB and --5.84 mW/cm 2 UVA as measured with an IL 700 research radiometer (International Light, Inc., Newburyport, MA). The UVB sensor had a peak sensitivity at --290 nm, while that of the UVA sensor was at --360 nm. Spectral distribution was measured with an Optronic 742 spectral radiometer at the beginning and end of the study. Mice were irradiated twice weekly: half of each group on Monday and Thursday, the other half on Tuesday and Friday. Over the first two weeks, exposures of all groups were raised by •/2- MED increments from 5 to 7 MEDs to avoid erythema in the group treated with the lowest SPF sunscreen. One MED, under the experimental conditions, required approx- imately .04 J/cm 2 as measured with the UVB sensor and was delivered in approximately 18 minutes. This corresponds closely to an MED for a type II human under the same experimental conditions. To accumulate the 10 and 16 MPDs at this rate required 18 and 30 weeks, respectively, with approximately two-hour exposures twice weekly. Ambient temperature, at mouse level, was maintained at 27øC with the use of two small fans. SUNSCREENS Each group of 18 mice was treated with its assigned sunscreen just prior to irradiation. Neither unprotected nor vehicle controls were included because the 7-MED dose could not be tolerated by unprotected mice. However, our laboratory has a large histologic data base of skin from unirradiated mice, long-term UV-irradiated at tolerated high doses, and various vehicle controls that provide relevant comparisons. Sunscreens were applied as 100 }xl to cover the entire exposed dorsal surface of the mice, including tail and feet, yielding an amount that was equivalent to --2.5 •l/cm 2. The active ingre- dients of the three sunscreens were as follows: 1: octyl methoxycinnamate, mainly a UVB absorber (290-320 nm) with an SPF of 7 2) octyl methoxycinnamate and oxy- benzone with an SPF of 16, providing absorption of UVB and UVA II (290-340 nm) 3) octyl methoxycinnamate, oxybenzone, and avobenzone (butyl methoxy dibenzoyl- methane Parsol 1789, Givaudan SA, Vernier, Switzerland). The latter had an SPF of 18, providing full-spectrum absorption of UVB, UVA II, and UVA I (290-400 nm). The SPF of each sunscreen was determined on human subjects according to the US Food and Drug Administration's proposed sunscreen monograph (12), using a xenon arc solar simulator fitted with a 1-mm WG-320 filter and with a spectrum similar to the one used in this study. HISTOLOGY Eight of the animals in each group were sacrificed at a cumulative dose of 10 MPD (18 weeks), with the surviving remainder (8-9) at a cumulative dose of 16 MPD (30 weeks). Dorsal skin biopsies (2 x 0.5 cm) were fixed in 10% buffered formalin and processed for light microscopy. The histochemical stains were hernatoxylin and eosin for general

132 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS morphology, Masson's trichrome for collagen, Luna's aidehyde fuchsin for elastic fibers (13), and Mowry's colloidal iron for glycosaminoglycans. SKIN-THICKNESS MEASUREMENTS Full-thickness skin was measured microscopically at 100X magnification on H&E stained sections using an eyepice micrometer. Measurements were taken from six animals per group at six sites along the --2-cm-long specimens at the 10- and 16-MPD dose points. TUMOR DATA Animals surviving at the 16-MPD dose point were examined for skin tumors prior to sacrifice. RESULTS Ten Minimal Photoaging Doses of Solar-Simulating Radiation HEMATOXYLIN AND EOSIN STAIN: GENERAL MORPHOLOGY SPF-7 sunscreen. The epidermis was hyperplastic with 6-8 cell layers compared to the normal 3-4 (10). There was mild loss of order, little or no atypia aside from an occasional sunburn cell, and a focally mild-to-moderate lymphocytic infiltrate (Figure 1A). Inflammation in the upper dermis was moderate, with occasional severe foci, and consisted mainly of lymphocytes with a few eosinophils and even fewer neutrophils. Sebaceous glands were enlarged, with --25 sebocytes compared to the normal 11-13 (14), as were the keratinizing cysts normally present in a single row in the lower dermis (10). The latter had proliferated to three rows in most specimens. The granulomatous reaction, evoked by rupturing cysts, was severe in most specimens (Figure lB). S?F-16 sunscreen. The general histology was similar to that seen with the SPF-7 sun- screen except that the epidermal lymphocytic infiltrate was more mild (Figure 1C). Although the skin was significantly thickened, cyst proliferation was mild, with most specimens having only 1-2 rows. The accompanying granulomatous reaction was mild to moderate (Figure 1D). S?F-18 sunscreen. Epidermal hyperplasia was slightly reduced, with most specimens having 5-7 cell layers. There was little or no loss of order or atypia aside from a rare sunburn cell (Figure 1E). The lymphocytic infiltrate was mild and consisted of scattered single cells. Upper dermal inflammation was similar to that in the other two groups. Sebaceous glands were less enlarged (--20 cells), and dermal cysts were not proliferated in most specimens. The granulomatous reaction was usually mild (Figure IF). LUNA'S STAIN: ELASTIC FIBERS S?F-7 sunscreen. Elastic fiber hyperplasia was mainly moderate, with some severe foci that approached elastosis (Figure 2A). The fibers were slightly thickened. Most speci-