Methods

Biopsies

Four-millimeter punch biopsies of clinically normal follicles from uninvolved skin and early inflamed lesions (papules of less than 6 h duration) were obtained, with local ethical committee approval, under local anaesthetic from the interscapular region of the backs of consenting adolescent acne patients. Approval had been given for the taking and handling of biopsies by The Research Ethics Committee of The Leeds Health Authority/United Leeds Teaching Hospitals (ref: 98/002) and biopsies were only taken with informed and written consent.

A mapping technique previously described (^{Norris and Cunliffe, 1988}) was used to obtain biopsies of inflamed lesions of known duration less than 6 h. Biopsies of uninvolved skin (n=20) and biopsies of early inflamed lesions (n=12) were obtained from acne patients, and control biopsies were taken of normal skin (n=10) from the interscapular region of the backs of adolescent subjects who had never suffered from acne. Different lesion biopsies were taken from separate subjects, i.e., biopsies of uninvolved skin were not taken from the same patients as the early inflamed lesions. The uninvolved skin was never closer than 1 cm to an inflamed lesion or any area of inflammation. All acne patients had been off treatment for more than 4 wk, had a mean age of 23 y (16–27 y), and a mean acne grade for their back of 2 (1–3) (Leeds revised acne grading system,^{O'Brien et al, 1998}). Control subjects had a mean age of 22 y (21–23 y). Biopsies were frozen and stored in liquid nitrogen. When required they were embedded in OCT and longitudinal serial sections (6 m) were cut through the whole sebaceous follicle or acne lesion. Sections were mounted on silane-treated glass slides, air dried for 12–24 h, fixed in 100% vol/vol acetone (10 min), dried (10 min), and stored at -70°C until used.

Primary antibodies for immunohistochemistry

Monoclonal antibodies UCHT1 (anti-CD3; specific for all T cells; dilution 1:50), MT310 (anti-CD4; helper/inducer T cells and some monocyte/macrophages; 1:8), DK25 (anti-CD8; cytotoxic/suppressor T cells; 1:25), 4KB5 (anti-CD45RA; naïve/resting T cells; 1:25), UCHL1 (anti-CD45RO; memory/effector T cells 1:50), NAI/34 (anti-CD1a; Langerhans cells; 1:100), DK22 (anti-HLA-DR; MHC class II; 1:25), Ki67 (anti-Ki67; nuclear proliferation antigen; 1:50), NP57 (antihuman neutrophil elastase/HRP EPOS), and an antilaminin rabbit polyclonal antibody (1:25) were purchased from Dako (Dako, Ely, UK). Monoclonal antibodies Ki-M6 (anti-CD68; monocytes/macrophages; 1:100) and NKI-GoH3 (anti-6 integrin; 1:50 and 1:10) were obtained from Serotec (Serotec, Kidlington, UK). HECA-45 (anti-CLA; cutaneous lymphocyte antigen; 1:25) was from BD Pharmingen (BD Pharmingen, Oxford, UK). Novacastra (Novacasta Newcastle-Upon-Tyne, UK) supplied monoclonal antibody LL025 (anti-K16; 1:20). Sanbio (Sanbio, Uden, The Netherlands) supplied the monoclonal antibody to EN4 (specific for endothelial cells; used at 1:50 dilution). Monoclonal antibodies to adhesion molecules were obtained from R & D Systems (R & D Systems, Abingdon, UK) and included BBIG-E4 (anti-E-selectin; 1:100), BBIG-I1 (anti-ICAM-1; 1:1000), and BBIG-V1 (anti-VCAM-1; 1:1000). Monoclonal antibody to P3G8 (anti-_v integrin; 1:750) was obtained from Chemicon (Chemicon, Harrow, UK). Anti-IL-1 rabbit polyclonal (1:50) and anti-IL-1 receptor type II (anti-IL-1RTII) monoclonal (1:50) antibodies were from Genzyme (now R & D Systems, UK). A rabbit polyclonal antihuman collagen type IV antibody was supplied by Biogenesis (Biogenesis, Poole, UK).

Secondary antibodies for immunofluorescence

Fluorescein isothiocyanate (FITC) conjugated goat antirabbit (1:75), biotinylated goat antirat (1:50), and Texas Red avidin DCS (1:100) were obtained from Vector Laboratories (Peterborough, UK).

Immunohistochemistry

Frozen sections were stained by standard immunoperoxidase techniques using monoclonal antibodies and Vectastain Elite ABC kits (Vector Laboratories). As all sections had been numbered consecutively after cutting it was possible to stain three sections, i.e., from the beginning, center, and end of each follicle/lesion, with each antibody. Briefly the technique involved the rehydration of sections in phosphate-buffered saline (PBS) pH 7.4 for 10 min, followed by addition of blocking serum (Vectastain Elite kit) for 15 min, and then incubation in primary antibody (45 min). Sections were washed in PBS (3 min 3). The appropriate biotinylated secondary antibody was added (30 min) followed by washing in PBS (3 min 3) and then the addition of the appropriate ABC peroxidase complex (30 min). After further washes in PBS, the remaining peroxidase activity was visualized using hydrogen peroxide (120 l) and diaminobenzidine (DAB, 1 ml in 400 ml PBS) as the substrate. Sections were counterstained with methyl green (10 min) and dehydrated in 100% vol/vol ethanol (4 min 2) and xylene (4 min 2) before mounting in Eukitt (Merck Leicester, UK). For the IL-1 labeling, the protocol was the same apart from (1) after initial thawing the sections were fixed in freshly made 4% paraformaldehyde in PBS pH 7.4 at 4°C for 8 min and then every wash step and reagent was with or made up in PBS/0.1% Saponin and (2) the primary antibody was diluted in PBS/bovine serum albumin 0.1%/Saponin 0.1%.

Immunofluorescence

Laminin, collagen IV, and ₆ integrin were visualized via immunofluorescence. Briefly the technique involved rehydration and blocking of non-specific binding using normal goat serum in PBS (15 min). Endogenous avidin and biotin were blocked (Vector Laboratories) (15 min each) before the primary antisera was added (1 h). Sections were washed in PBS (3 min 3) before addition of the secondary antibody, which was FITC conjugated, Texas Red (TxRD) conjugated, or biotinylated (30 min). Sections were washed in PBS (3 min 3). As the TxRD labeling required further amplification sections were incubated with a biotinylated tertiary antibody (30 min) followed by a PBS wash (3 min 3) and TxRD avidin DCS (30 min). The slides were kept in the dark at 4°C after being mounted with Vectashield^® (Vector Laboratories) and sealed with clear nail varnish.

Quantitative assessment

Skin sections were examined using a Zeiss Axioplan microscope (200 magnification) and photographed using a digital camera (Nikon Coolpix 950). The total number of positively immunolabeled cells and blood vessels were counted by two observers independently for each section. As three sections per follicle/lesion were stained with each antibody, a mean count of immunolabeled cells and blood vessels was determined per follicle/lesion. Computer-aided image analysis was used to measure section area. A black and white image of each section area was captured via a CCD camera attached to a microscope (Leitz Laborlux S) using 2.5 magnification. Image analysis software (Image Pro+), which was calibrated to measure in square millimeters, was then used to obtain section areas. Areas of three different sections, through each follicle/lesion, were determined and the mean of the three values was taken as the section area for that follicle/lesion. Positive immunolabeled cells and blood vessels were finally quantified as counts per mm² per follicle/lesion.

₆ integrin expression was examined using both immunofluorescence and immunoperoxidase techniques with suprabasal cell expression being assessed using a 0 to 3 scale (0,no suprabasal expression; 1, odd suprabasal cell expression; 2, patches of suprabasal expression; 3, large numbers of suprabasal cell expression). IL-1 epidermal labeling was assessed according to the location and the intensity of labeling as follows: 0, no labeling; 1, weak labeling; 2, moderate labeling; 3, strong labeling. Basal and suprabasal cell assessment of the epidermis was carried out separately and the results were subdivided into further areas as follows: interfollicular, perifollicular, and the follicle, which was divided into acroinfundibulum and infrainfundibulum. Both IL-1 dermal cell labeling and IL-1RTII cell labeling were quantified as counts per mm² per follicle/lesion.

Ki67 labeling was expressed as the follicular growth fraction. The follicular growth fraction is the proportion of proliferative basal cells down the follicle wall. This is therefore calculated as the number of positively labeled (cycling) basal cells divided by the total number of basal cells. K16 labeling, of cells of the sebaceous gland and suprabasal keratinocytes, was assessed according to three patterns observed by^{Aldana et al (1998)}: stage 1, K16 expression was found in cells at the sebaceous gland/follicular junction; stage 2, K16 expression was found in cells at the sebaceous gland/follicular junction and up the follicle wall; stage 3, K16 expression was found in cells at the sebaceous gland/follicular junction, up the follicle wall, and in the interfollicular epidermis.

Statistical analysis

The data are represented as the meanSEM. As sample numbers in each group were small, the data were analysed by the nonparametric Mann-Whitney U test. p-values of less than 0.05 were considered significant.

Results

Proliferation and differentiation

The follicular growth fraction in the epithelium of follicles from uninvolved skin was significantly lower than proliferative values found in early inflamed lesions (p=0.0014) and was comparable to the values in control follicles (Figure 1).

Figure 1.

The proliferative activity (Ki67) and differentiation status of keratinocytes in the follicular epithelium of PSUs. Frozen sections of PSUs from normal control skin, uninvolved acne skin, and papules of less than 6 h were labeled with antibodies to Ki67 and K16. The percentage of Ki67⁺ basal keratinocytes (growth fraction, GF) was calculated. K16 expression was graded from 0–3. Values represent the meanSEM per follicle or lesion.

Full figure and legend (13K)

K16, the marker of keratinocyte hyperproliferation and abnormal differentiation, exhibited a similar trend to the follicular growth fraction, with significantly lower expression in follicles from uninvolved skin compared to early inflamed lesions (p<0.0001). The majority of uninvolved follicles exhibited stage 1 K16 expression, which was akin to that found in follicles from control skin (Figure 1).

T cells and neutrophils

CD3⁺(pan T cells), CD4⁺(helper T cells), CD8⁺(cytotoxic T cells), and neutrophils

To identify the nature of the cellular infiltrate within the skin surrounding pilosebaceous follicles, key T cell markers were examined as well as a marker for neutrophils. In uninvolved acne skin, numbers of CD3⁺ T cells were elevated in the perifollicular and papillary dermis compared to controls (p<0.0001), although levels were not equivalent to those in early papules in which cell numbers were further increased (p<0.0001) (Figure 2a–c).

Figure 2.

Differences in the number of T cells present in and around pilosebaceous follicles. Frozen sections were obtained from (a) normal control skin, (b) uninvolved acne skin, and (c) papules of less than 6 h. Immunohistochemical staining was carried out using a CD3 antibody and visualized by a standard streptavidin biotin horseradish peroxidase technique. Scale bar: 100 m.

Full figure and legend (255K)

CD4⁺ T cells were found to be the predominant subset present, with increased numbers around uninvolved follicles compared to controls (p<0.0001), although not to levels seen in papules of less than 6 h (p=0.0027). There was no significant increase in CD8⁺ T cells or neutrophils in uninvolved skin compared to controls (p>0.3455), although the number of neutrophils did increase in papules of less than 6 h (p=0.0009) (Figure 3).

Figure 3.

Numbers of helper T cells (CD4), cytotoxic T cells (CD8), and neutrophils in and around pilosebaceous follicles. Immunostaining was carried out on biopsies from normal control skin, uninvolved acne skin, and papules of less than 6 h. The numbers of positive cells per mm² were determined for CD4, CD8, and neutrophil elastase. Data are shown as meanSEM per follicle or lesion.

Full figure and legend (14K)

CD45RA⁺ (naïve T cells), CD45RO⁺ (memory/effector T cells), and CLA⁺ (skin homing T cells)

In order to further characterize the T cell population, naïve (CD45RA⁺) and memory/effector (CD45RO⁺) subsets were investigated. Increased numbers of both CD45RA⁺ (p=0.0278) and CD45RO⁺ (p<0.0001) subsets were found around follicles in uninvolved skin compared to controls and these were at similar levels to those observed in papules of less than 6 h (p=0.5093) (Figure 4). Around control follicles the ratio of memory/effector T cells to naïve T cells was 2.5:1. The predominance of memory/effector T cells was further increased around the follicles from uninvolved skin and in early papules with a ratio of 6:1. The number of skin homing (CLA⁺)T cells also increased around the follicles from uninvolved skin compared to controls (p<0.0001), reaching similar levels to those found in papules of less than 6 h (p=0.8337) (Figure 4). Although the cells were not double labeled it can be inferred that the majority of the memory/effector cell population possessed a skin homing phenotype. In all of the skin samples examined (controls, uninvolved, and papules of less than 6 h) summation of naïve and memory/effector T cell subsets resulted in fewer cell numbers than the total CD4⁺ T cell count and therefore the remainder were unclassified (CD3⁺, CD4⁺, CD45RO^-, CD45RA^-, CLA^-).

Figure 4.

Numbers of T helper cell subsets including naïve (CD45RA), memory/effector (CD45RO), and skin homing (CLA) in and around pilosebaceous follicles. Immunostaining was carried out on biopsies from normal control skin, uninvolved acne skin, and papules of less than 6 h. The numbers of positive cells per mm² were determined for CD45RA, CD45RO, and CLA. Data are shown as meanSEM per follicle or lesion.

Full figure and legend (13K)

Macrophages (CD68⁺)

There were a significantly large number of macrophages (CD68⁺ cells) in the perifollicular and papillary dermis around follicles from uninvolved acne skin compared to control skin (p<0.0001) and these were found to be at similar numbers to those exhibited in early papules (p=0.9379) (Figure 5).

Figure 5.

Numbers of macrophages (CD68) present around pilosebaceous follicles. Immunostaining was carried out on biopsies from normal control skin, uninvolved acne skin, and papules of less than 6 h. The numbers of positive cells per mm² were determined for CD68. Data are shown as meanSEM per follicle or lesion.

Full figure and legend (12K)

HLA-DR⁺ cells

Despite an increase in the number of macrophages and T cells the number of HLA-DR⁺ cells did not change within the cellular infiltrate in the perifollicular and papillary dermis around uninvolved follicles compared with controls. A 2- to 3-fold increase of HLA-DR⁺ cells was observed in papules of less than 6 h, however (p<0.0062) (data not shown). No keratinocyte expression was observed.

Langerhans cells (CD1a⁺)

A network of CD1a⁺ cells (considered to be Langerhans cells) was seen throughout the epidermis and follicle wall in all of the biopsies investigated. The density of these cells was decreased in the epidermis around follicles from uninvolved skin compared to controls (p<0.0001), whereas the numbers in the epidermis surrounding papules of less than 6 h were equivalent to those in controls (p=0.7248). The number of CD1a⁺ cells was also decreased in the dermis around uninvolved follicles compared to control skin (p=0.0387) whereas numbers in papules of less than 6 h were similar to those found in controls (p=0.3600) (data not shown).

Vascular markers

There was no change in the number of blood vessels (EN4⁺) found around follicles from control skin and uninvolved skin, although there was an increase in papules of less than 6 h (Table I, Figure 6).

Figure 6.

Vascular marker expression on the dermal blood vessels around pilosebaceous follicles. Frozen sections from biopsies of normal control skin, uninvolved acne skin, and papules of less than 6 h were labeled with antibodies to the endothelial cell marker EN4, ICAM-1, E-selectin, HLA-DR, and VCAM-1. Values represent mean numberSEM of positive staining blood vessels per mm² per follicle or lesion.

Full figure and legend (13K)

Table I - Statistical p-values for the vascular markers.

Full table

Vascular intercellular adhesion molecule 1 (ICAM-1) expression appeared to increase in uninvolved skin compared to controls despite not being statistically significant, although a significant increase was found in papules of less than 6 h (Table I, Figure 6). When the number of blood vessels expressing ICAM-1 was represented as a percentage of the total number of blood vessels, however, an increase was found from 50% in control skin to 66% in uninvolved skin. This was equivalent to the 65% expression observed in early papules.

Vascular E-selectin expression was also found to be increased in papules of less than 6 h compared to controls, whereas the expression around follicles in uninvolved skin was increased to levels not dissimilar to those in papules of less than 6 h (Table I, Figure 6). The results tell a different story, however, when shown as a percentage of blood vessels. The controls and early papules were at similar levels at 44% and 42% respectively, whereas the level of expression in the uninvolved skin was increased at 62%.

Vascular HLA-DR expression was significantly increased around uninvolved follicles compared to control skin, with a further increase in expression observed in papules of less than 6 h (Table I, Figure 6). As a percentage of the total number of blood vessels, HLA-DR⁺ vascular expression was at similar levels in uninvolved skin as in early inflamed lesions at 33% and 30% respectively. This was double the 16% observed in control skin.

Vascular cell adhesion molecule 1 (VCAM-1) expression was upregulated on the vasculature in uninvolved skin compared to control skin, although it had not reached levels found in papules of less than 6 h (Table I, Figure 6). The same trend of expression was observed when plotted as a percentage of the total number of blood vessels, at 1% in control skin, 10% in uninvolved skin, and 29% in papules of less than 6 h.

Vascular _v integrin expression was increased in the uninvolved skin compared to normal control skin, with a further increase found in papules of less than 6 h compared to the expression in uninvolved skin (Table I, Figure 7). When determined as a percentage of the total number of blood vessels the increase in _v expression in uninvolved skin was further emphasized with 9% expression in controls and 44% in both the uninvolved skin and papules of less than 6 h.

Figure 7.

_v integrin expression on the dermal vasculature around pilosebaceous follicles. Frozen sections from biopsies of normal control skin, uninvolved acne skin, and papules of less than 6 h were labeled with an _v integrin antibody. Values represent mean numberSEM of positive staining blood vessels per mm² per follicle or lesion.

Full figure and legend (11K)

Laminin and collagen IV

Both laminin and collagen IV were restricted to the basement membrane and labeling was unbroken along the epidermis and down the follicle wall in all of the samples investigated. In the dermis both were located on blood vessels and collagen IV was also significantly present in the papillary dermis.

₆ integrin

₆ integrin expression was predominantly found on basal keratinocytes of the epidermis and follicle wall, with the strongest labeling observed at the basal pole of these cells. Using both immunofluorescence and immunoperoxidase techniques all samples investigated exhibited strong basal cell labeling in the epidermis and down the follicle wall. Suprabasal cell expression was also found in all groups although only a small number of suprabasal cells expressed ₆ in normal controls, whereas around uninvolved follicles and inflamed lesions an increase in suprabasal expression was observed in both the interfollicular (p<0.0393) and perifollicular (p<0.018) epidermis. This increase was greatest in the perifollicular epidermis. Down the follicle wall, suprabasal cell expression was absent in all of the samples using immunofluorescence. In contrast, using immunoperoxidase, suprabasal cell expression was present in inflamed lesions but not in normal control or uninvolved follicles.

IL-1 and IL-1RTII

In the epidermis of all the samples investigated, IL-1 immunolabeling was observed as a diffuse pattern within the suprabasal layers, but there was stronger reactivity in the basal layer and occasionally in the granular layer and/or the stratum corneum. An increase in IL-1 labeling was observed in all layers of both the interfollicular and perifollicular epidermis (p<0.0064) and down the follicle wall (p<0.0278) in uninvolved skin compared to controls. Significant increases in labeling were also observed in the epidermis around inflamed lesions compared to uninvolved skin (p<0.0207). Down the follicle wall, however, only labeling in the infrainfundibular basal cells increased (p=0.0019), whereas labeling in the suprabasal layers of the infrainfundibulum and all layers of the acroinfundibulum did not differ in intensity in uninvolved skin and early inflamed lesions (p>0.2235) (data not shown). As a trend, IL-1 labeling intensity was found to decrease from the acroinfundibulum to the infrainfundibulum.

IL-1RTII labeling was predominantly found as discrete basal cell reactivity. Within the dermis both IL-1 and IL-1RTII labeling presented as discrete cell associated immunoreactivity, with the majority of positive cells located perifollicularly. In normal control skin, only a small number of dermal cells expressed IL-1 whereas IL-1RTII expression was absent or at low levels in both the dermis and the basal keratinocytes of the epidermis. Around follicles in uninvolved skin, both dermal IL-1 and dermal/epidermal IL-1RTII levels were significantly elevated, 3-fold and 30-fold respectively, compared to normal controls (p<0.0002) (Figure 8). In skin around early inflamed lesions a further 3-fold increase in IL-1 was observed, together with a doubling in IL-1RTII expression, compared to that exhibited in uninvolved skin (IL-1, p=0.003; IL-1RTII, p=0.0243) (Figure 8).

Figure 8.

Dermal cell IL-1 and total dermal/epidermal cell IL-1RTII expression around pilosebaceous follicles. Frozen sections from biopsies of normal control skin, uninvolved acne skin, and papules of less than 6 h were labeled with IL-1 and IL-1RTII antibodies. The numbers of positive cells per mm² were determined. Data are shown as meanSEM per follicle or lesion.

Full figure and legend (15K)

Discussion

This study provides novel evidence for the involvement of inflammatory events in the very earliest stages of acne lesion development. Quantitative assessment has identified significant inflammatory factors around clinically normal pilosebaceous follicles from uninvolved skin from acne patients prior to hyperproliferation of the follicular epithelium.

Around uninvolved follicles there were large numbers of CD4⁺ T cells over and above the constitutive level of surveillance T cells in normal skin, accompanied by a large macrophage presence, which was equivalent to that observed in clinically apparent early (<6h) inflamed lesions. In addition, there was a notable absence of neutrophils and a low number of CD8⁺ T cells within the cellular infiltrate, similar to that in normal skin. The decrease in dermal and especially epidermal Langerhans cell numbers in the skin around these follicles suggests migration of these cells to the skin draining lymph nodes. Within the CD4⁺ T cell population, the majority were memory/effectors, with a similar proportion exhibiting a skin homing phenotype. This would suggest the start of a specific inflammatory response from the adaptive immune system. Despite the increase in cellular infiltrate, however, the absence of an increase in the cellular expression of HLA-DR within the infiltrate would indicate a lack of activation of these cells.

In association with the type of lymphocytic infiltrate present, an increase in the expression of vascular adhesion molecules VCAM-1, E-selectin (ligand for the CLA skin homing molecule), and ICAM-1 was observed, which would facilitate such cellular recruitment. An increase in endothelial expression of HLA-DR also provides evidence of vascular activation, and despite no obvious increase in the number of blood vessels around these follicles the upregulation of vascular _v integrin expression suggests that an initiation of angiogenic events has already taken place.

The uninvolved follicles from acne-prone areas were investigated to search for the factors involved in the initiation of acne and to determine whether keratinocyte hyperproliferation comes before or after an inflammatory initiation. At the present time it is not possible to identify "acne-prone" follicles. Thus, only clinically normal follicles from uninvolved skin from acne patients can be biopsied, which is clearly a random process. Therefore, in this study the uninvolved follicles were analysed as one group. From this, the proliferative and differentiation state of these follicles was found to be similar to that of normal control follicles. This is a significant result, as it shows that subclinical inflammatory events are occurring in the skin around pilosebaceous follicles prior to hyperproliferative or abnormal differentiation events.

The absence of hyperproliferation was also confirmed by the fact that there were no microcomedonal features, such as distension of the duct wall, present in the majority of the uninvolved follicles investigated. The follicular duct appeared unbroken with no evidence even of partial rupture and this was verified by immunolabeling of laminin and collagen IV. These are located along the dermal–epidermal junction and were evident as unbroken linear labeling in all of the biopsies investigated, which suggests an intact basement membrane.

Both fluorescent and peroxidase immunolabeling of ₆ integrin in normal skin was similar to that demonstrated by^{Cavani et al (1993)} with labeling in the basal cells, focused at the dermal pole. Cavani et al also showed an increase in intensity of labeling and suprabasal ₆ expression in wounded epidermis, whereas^{Carroll et al (1995)} demonstrated that transgenic mice with suprabasal integrin expression exhibited keratinocyte hyperproliferation and skin inflammation. The increased suprabasal expression of ₆ integrin seen in the uninvolved skin with even greater expression in inflamed lesions could be indicative of a response to inflammation (or wound healing) and/or be related to hyperproliferation, giving rise to the abnormal differentiation characteristic in acne. The aberrant integrin expression appears to be present alongside the inflammatory events and this would therefore mean that it precedes hyperproliferation, although whether it is responsible for it remains to be elucidated.

This therefore raises the question of the identity of the stimulus for the increased cellular presence and inflammatory vascular marker expression. Within the T cell infiltrate, the greatest population increase was those with a memory/effector phenotype compared to the naïve T cell population, with an increased presence of 6:1 in uninvolved skin compared to 2.5:1 in normal skin. As previously mentioned, the predominance of a CD4⁺"memory/effector skin homing" T cell phenotype and the lack of a neutrophilic presence would suggest a specific inflammatory response such as an antigenic response, as opposed to a non-specific response driven by the innate immune system. This may be a consequence, however, of the upregulation of the vascular marker E-selectin, which is the receptor for the skin homing T cell, and this would therefore cause the selective migration of cells of this phenotype into the local tissue. An upregulation of E-selectin and other vascular adhesion molecules could be caused by cytokine stimulation of the endothelium. This scenario may also account for the lack of activated (HLA-DR⁺) cells present. In addition, memory/effector T cells have been demonstrated to be the predominant T cell population within the skin-derived afferent lymph of humans (^{Yawalkar et al, 2000}). Therefore these cells would be likely to be the primary cell type to migrate to the site of an inflammatory stimulus within the skin.

Apart from the lymphocytic infiltrate, of possible relevance is the large macrophage presence, which was found to be at a similar level to that found in the early inflamed lesions and over double the number of those seen in normal skin. This suggests a prominent role for macrophages in the early phase of acne lesion development, although, as stated previously, the apparent absence of an increase in HLA-DR expression within the cellular infiltrate points to a lack of cellular activation. The reasons for this remain unclear. These findings of a CD4⁺ lymphocyte/macrophage led response, i.e., cell-mediated response, however, would tie in with the concept previously introduced from a study of timed inflamed acne lesions (^{Layton et al, 1998}), which pointed to a type IV delayed hypersensitivity response and would therefore implicate a soluble antigen stimulus.

The cytokines produced in a delayed-type hypersensitivity response could activate local endothelial cells causing an upregulation of the inflammatory vascular markers E-selectin, VCAM-1, ICAM-1, and HLA-DR, exhibited on the vasculature around the pilosebaceous follicles in the uninvolved skin from this study. In addition, the increased vascular _v integrin expression observed could be an early sign of angiogenesis. During angiogenesis _v3 integrin is expressed on the vasculature (^{Brooks et al, 1994}) and an increase in _v expression has been noted in the vasculature in psoriatic skin compared to normal skin (^{Creamer et al, 1995}).

Low levels of IL-1 and negligible IL-1RTII expression were observed in normal control skin. An upregulation of IL-1RTII is indicative of a corresponding increase in IL-1 activation (^{Colotta et al, 1993;Groves et al, 1994}). Therefore, the 3-fold increase in IL-1 expression and a 30-fold increase in IL-1RTII expression would suggest an increase in IL-1 activity in the skin around the uninvolved follicles compared to normal control skin. It is possible that this increase in IL-1 activity may have initiated the upregulation of vascular adhesion molecule expression (ICAM-1, E-selectin, and VCAM-1) and cellular (CD4⁺ T cell and macrophage) migration. ICAM-1 and E-selectin expression have been shown to be upregulated by tumor necrosis factor , interferon- and, more importantly, IL-1 (^{Dustin et al, 1986;1988;Groves et al, 1991;1992;1995}). In addition, IL-1 can induce the expression of vascular endothelial growth factor (^{Kozlowska et al, 1998}) and can therefore indirectly contribute to angiogenesis and induction of selected adhesion molecules involved in leukocyte migration such as E-selectin and VCAM-1 (^{Detmar et al, 1998}).

If IL-1 is the stimulus for a non-specific inflammatory response, then what factors could be responsible for its own increased expression and release? It might be a consequence of the overall inflammation present, or we could hypothesize that, as an essential fatty acid (linoleic acid) deficiency exists in acne (^{Downing et al, 1986}) in the keratinocytes of the follicular wall (^{Perisho et al, 1988}) caused by the dilution effect of high sebum production, this would lead to perturbation of the barrier function (^{Elias et al, 1980}) within individual follicles and induce the release of pro-inflammatory cytokines, i.e., IL-1, tumor necrosis factor , from keratinocytes into the dermis, stimulating an inflammatory cascade (^{Wood et al, 1994}). Thus, the upregulation of IL-1 demonstrated around uninvolved follicles could be an initiating factor and be responsible for the inflammatory events observed around acne lesions.

Previous studies investigating early inflammatory events in acne lesions were carried out after clinical presentation (^{Norris and Cunliffe, 1988;Layton et al, 1998}). These provided strong evidence for an initial CD4⁺ lymphocytic infiltrate as a primary inflammatory event. This is supported by the results from this study, with similar observations but of much earlier events in the sequence of lesion development. Previously, inflammation has always been considered a secondary event, preceded by hypercornification of the follicular duct via hyperproliferation (^{Webster, 1995;Cunliffe et al, 2000}). Therefore, the results from this study suggest that inflammatory events occur prior to and act as possible causal factors in the hyperproliferative changes observed in acne lesions, as opposed to secondary consequential events. The demonstration that an increase in IL-1 activity occurs prior to that of hyperproliferation around uninvolved follicles also coincides with the "keratinocyte activation cycle" proposed by^{Freedberget al (2001)}, in which keratinocytes can be activated via the release of IL-1, with activation markers K6 and K16 being expressed as a subsequent event. Thus, this would give weight to the argument that acne vulgaris should be classified as an inflammatory skin disease as opposed to a keratinocyte/hyperproliferative disorder. This study also provides good evidence to support the treatment of uninvolved skin in acne patients and not just the visible lesions alone and may validate the topical use of anti-inflammatory based therapies for this skin disorder.

Notes

¹ Aldana OL, Holland DB, Cunliffe WJ: Precomedonal events in acne. J Invest Dermatol 107:488, 1996 (abstr.)

² Holland DB, Aldana OL, Cunliffe WJ: Abnormal integrin expression in acne. J Invest Dermatol 110:559, 1998 (abstr.)

³ Aldana OL, Holland DB, Cunliffe WJ: A role for interleukin-1 in comedogenesis. J Invest Dermatol 110:558, 1998 (abstr.)

References

References

1.	Aldana OL, Holland DB & Cunliffe WJ. Variation in pilosebaceous duct keratinocyte proliferation in acne patients. Dermatology (1998) 196: 98–99. | Article | PubMed | ISI | ChemPort |
2.	Brooks PC, Clark RAF & Cheresh DA. Requirement of vascular integrin v3 for angiogenesis. Science (1994) 264: 569–571. | PubMed | ISI | ChemPort |
3.	Carroll JM, Romero R & Watt FM. Suprabasal integrin expression in the epidermis of transgenic mice results in developmental defects and a phenotype resembling psoriasis. Cell (1995) 83: 957–968. | Article | PubMed | ISI | ChemPort |
4.	Cavani A, Zambruno G, Marconi A, Manca V, Marchetti M & Giannetti A. Distinctive integrin expression in the newly forming epidermis during wound healing in humans. J Invest Dermatol (1993) 101: 600–604. | Article | PubMed | ISI | ChemPort |
5.	Colotta F, Re F & Muzio IM. et al Interleukin-1 type II receptor: A decoy target for IL-1 that is regulated by IL-4. Science (1993) 261: 472–475. | PubMed | ISI | ChemPort |
6.	Creamer D, Allen M, Sonsa A, Poston R & Barker J. Altered vascular endothelium integrin expression in psoriasis. Am J Pathol (1995) 147: 1661–1667. | PubMed | ISI | ChemPort |
7.	Cunliffe WJ, Holland DB, Clark SM & Stables GI. Comedogenesis: Some new aetiological, clinical and therapeutic strategies. Br J Dermatol (2000) 142: 1084–1091. | Article | PubMed | ISI | ChemPort |
8.	Detmar M, Brown L & Schon MP. et al Increased microvascular density and enhanced leukocyte rolling and adhesion in the skin of VEGF transgenic mice. J Invest Dermatol (1998) 111: 1–6. | Article | PubMed | ISI | ChemPort |
9.	Downing DT, Stewart ME, Wertz PW & Strauss JS. Essential fatty acids and acne. J Am Acad Dermatol (1986) 14: 221–225. | PubMed | ISI | ChemPort |
10.	Dustin ML, Rothlein R, Bhan AK, Dinarello CA & Springer TA. Induction by IL-1 and IFN, tissue distribution, biochemistry and function of a natural adherence molecule (ICAM-1). J Immunol (1986) 137: 245–254. | PubMed | ISI | ChemPort |
11.	Dustin ML, Singer KH, Tuck DT & Springer TA. Adhesion of T-lymphoblasts to epidermal keratinocytes is regulated by IFN and is mediated by ICAM-1. J Exp Med (1988) 167: 1323–1340. | Article | PubMed | ISI | ChemPort |
12.	Elias PM, Brown BE & Ziboh VA. The permeability barrier in essential fatty acid deficiency: Evidence for a direct role for linoleic acid in barrier function. J Invest Dermatol (1980) 74: 230–233. | Article | PubMed | ISI | ChemPort |
13.	Freedberg IM, Tomic-Canic M, Komine M & Blumenberg M. Keratins and the keratinocyte activation cycle. J Invest Dermatol (2001) 116: 633–640. | Article | PubMed | ISI | ChemPort |
14.	Groves RW, Allen MH, Barker JNWN, Haskard DO & MacDonald DM. Endothelial leucocyte adhesion molecule-1 (ELAM-1) expression in cutaneous inflammation. Br J Dermatol (1991) 124: 117–123. | PubMed | ISI | ChemPort |
15.	Groves RW, Rauschmayr T, Nakamura K, Sarkar S, Williams IR & Kupper TS. Effect of in vivo interleukin-1 on adhesion molecule expression in normal human skin. J Invest Dermatol (1992) 98: 384–387. | Article | PubMed | ISI | ChemPort |
16.	Groves RW, Sherman L, Mizutani H, Dower SK & Kupper TS. Detection of interleukin 1 receptors in human epidermis (induction of the type II receptor after organ culture and in psoriasis). Am J Pathol (1994) 145: 1048–1056. | PubMed | ISI | ChemPort |
17.	Groves RW, Allen MH, Ross EL, Barker JNWN & MacDonald DM. Tumour necrosis factor is pro-inflammatory in normal human skin and modulates cutaneous adhesion molecule expression. Br J Dermatol (1995) 132: 345–352. | PubMed | ISI | ChemPort |
18.	Guy R & Kealey T. The effects of inflammatory cytokines on the isolated human sebaceous infundibulum. J Invest Dermatol (1998) 110: 410–415. | Article | PubMed | ISI | ChemPort |
19.	Guy R, Green MR & Kealey T. Modeling acne in vitro. J Invest Dermatol (1996) 106: 176–182. | Article | PubMed | ISI | ChemPort |
20.	Ingham E, Eady A, Goodwin CE, Cove JH & Cunliffe WJ. Pro-inflammatory levels of IL-1 -like bioactivity are present in the majority of open comedones in acne vulgaris. J Invest Dermatol (1992) 98: 895–901. | Article | PubMed | ISI | ChemPort |
21.	Knaggs HE, Holland DB, Morris C, Wood EJ & Cunliffe WJ. Quantification of cellular proliferation in acne using the monoclonal antibody Ki67. J Invest Dermatol (1994) 102: 89–92. | Article | PubMed | ISI | ChemPort |
22.	Knutson DD. Ultrastructural observations in acne vulgaris: The normal sebaceous follicle and acne lesions. J Invest Dermatol (1974) 62: 288–307. | Article | PubMed | ISI | ChemPort |
23.	Kozlowska U, Blume-Peytavi U, Kodelja V, Sommer C, Goerdt S, Jablonska S & Orfanos CE. Vascular endothelial growth factor expression induced by pro-inflammatory cytokines (Interleukin 1, ) in cells of the human pilosebaceous unit. Dermatology (1998) 196: 89–92. | Article | PubMed | ISI | ChemPort |
24.	Layton AM, Morris C, Cunliffe WJ & Ingham E. Immunohistochemical investigation of evolving inflammation in lesions of acne vulgaris. Exp Dermatol (1998) 7: 191–197. | PubMed | ISI | ChemPort |
25.	Norris JFB & Cunliffe WJ. A histological and immunocytochemical study of early acne lesions. Br J Dermatol (1988) 118: 651–659. | PubMed | ISI | ChemPort |
26.	O'Brien SC, Lewis JB & Cunliffe WJ. The Leeds revised acne grading system. J Dermatological Treatment (1998) 9: 215–220.
27.	Orentreich N & Durr NP. The natural evolution of comedones into inflammatory papules and pustules. J Invest Dermatol (1974) 62: 316–320. | Article | ISI |
28.	Perisho K, Wertz PW, Madison KC, Stewart ME & Downing DT. Fatty acids of acylceramides from comedones and from the skin surface of acne patients and control subjects. J Invest Dermatol (1988) 90: 350–353. | Article | PubMed | ISI | ChemPort |
29.	Plewig G, Fulton JE & Kligman AM. Cellular dynamics of comedo formation in acne vulgaris. Arch Dermatol Forsch (1971) 242: 12–29. | Article | PubMed | ISI | ChemPort |
30.	Webster GF. Inflammation in acne vulgaris. J Am Acad Dermatol (1995) 33: 247–253. | Article | PubMed | ISI | ChemPort |
31.	Wood LC, Feingold KR, Sequeira-Martin SM, Elias PM & Grunfeld C. Barrier function coordinately regulates epidermal IL-1 and IL-1 receptor antagonist mRNA levels. Exp Dermatol (1994) 3: 56–60. | PubMed | ChemPort |
32.	Yawalkar N, Hunger RE, Pichler WJ, Braathen LR & Brand CU. Human afferent lymph from normal skin contains an increased number of mainly memory/effector CD4+ T cells expressing activation, adhesion and co-stimulatory molecules. Eur J Immunol (2000) 30: 491–497. | Article | PubMed | ISI | ChemPort |