Generation and cyclic remodeling of the hair follicle immune system in mice.

In this immunohistomorphometric study, we have defined basic characteristics of the hair follicle (HF) immune system during follicle morphogenesis and cycling in C57BL/6 mice, in relation to the skin immune system. Langerhans cells and gammadelta T cell receptor immunoreactive lymphocytes were the predominant intraepithelial hematopoietic cells in neonatal mouse skin. After their numeric increase in the epidermis, these cells migrated into the HF, although only when follicle morphogenesis was almost completed. In contrast to Langerhans cells, gammadelta T cell receptor immunoreactive lymphocytes entered the HF only via the epidermis. Throughout HF morphogenesis and cycling, both cell types remained strikingly restricted to the distal outer root sheath. On extremely rare occasions, CD4+ or CD8+ alphabetaTC were detected within the HF epithelium or the sebaceous gland. Major histocompatibility complex class II+, MAC-1+ cells of macrophage phenotype and numerous mast cells appeared very early on during HF development in the perifollicular dermis, and the percentage of degranulated mast cells significantly increased during the initiation of synchronized HF cycling (first catagen). During both depilation- and cyclosporine A-induced HF cycling, the numbers of intrafollicular Langerhans cells, gammadelta T cell receptor immunoreactive lymphocytes, and perifollicular dermal macrophages fluctuated significantly. Yet, no numeric increase of perifollicular macrophages was detectable during HF regression, questioning their proposed role in catagen induction. In summary, the HF immune system is generated fairly late during follicle development, shows striking differences to the extrafollicular skin immune system, and undergoes substantial hair cycle-associated remodeling. In addition, synchronized HF cycling is accompanied by profound alterations of the skin immune system.

MHC class I expression in murine skin: developmentally controlled and strikingly restricted intraepithelial expression during hair follicle morphogenesis and cycling, and response to cytokine treatment in vivo.

Hair bulb keratinocytes generate one of the few "immune privileged" tissue compartments of the mammalian organism by suppressing classical MHC class I (MHC Ia) antigens. Expression of non-classical MHC class I (MHC Ib) antigens in the follicle has been found, but only in its distal epithelium. Here, we have defined when during murine hair follicle morphogenesis these peculiar MHC Ia and Ib expression patterns are established, how they change during the murine hair cycle, and how different MHC I modulatory agents alter follicular MHC Ia and Ib expression in vivo. During neonatal hair follicle morphogenesis in C57BL/6 mice, distal follicle keratinocytes began to express MHC Ia (H2b) only late in development. The MHC Ib antigens, Qa-1 and Qa-2, did not become visible until the initiation of follicle cycling, with Qa-1 expression being more widespread than that of Qa-2. H2b, Qa-1, and TAP-1 immunoreactivity on previously negative keratinocytes of the proximal anagen hair bulb was upregulated by intradermal injection of the proinflammatory cytokine interferon-gamma, but not by tumor necrosis factor-alpha or interleukin-1beta. Injection of the reportedly MHC class I downregulating agents interleukin-10, insulin-like growth factor-1, transforming growth factor-beta, alpha-melanocyte stimulating hormone, or dexamethasone, however, all failed to downregulate constitutive or interferon-gamma-induced follicular MHC Ia expression. This shows that the hair follicle is a previously unrecognized site of Qa-1 expression and that interferon-gamma is a key regulator of follicular MHC I expression in vivo. It also suggests that the developmental and immunologic controls of MHC I expression by follicle keratinocytes differ from those of other epithelial cells.

Towards defining the pathogenesis of the hairless phenotype.

Mutation of the hairless (hr) gene in mice causes severe abnormalities during the first hair follicle regression (catagen), resulting in complete baldness. Here, we further characterize how hairlessness develops in HRS/J hairless mouse skin (hr) by histology, histochemistry, immunohistology, and in situ hybridization. We show that, in hr skin, only two defined epithelial cell populations in the distal outer root sheath (ORS) retain their integrity, whereas the rest of the ORS disintegrates. The surviving distal ORS forms the characteristic utriculi, whereas the remnants of the bulge get isolated from other epithelial compartments, but retain the capacity to proliferate and to produce either columnar epithelial outgrowths or selected dermal cysts. Normal dermal papilla structures get lost during the development of hairlessness. Based on the patterns of keratin 17 mRNA and neural cell adhesion molecule antigen expression, and on the distribution of alkaline phosphatase activity, we propose that dermal cysts in hr skin arise from (i) the central ORS, (ii) bulge-derived cells, or (iii) the disintegrating proximal ORS under the influence of dermal papilla remnants. The hr mutation seems to disrupt the integrity of key functional tissue units in the hair follicle, possibly due to a dysregulation of normal, catagen-associated apoptosis and/or an impairment of cell adhesion, whereas the distal follicle epithelium (including its stem cell region) seems to be largely protected from this. Thus, hairless mice offer a unique model for dissecting the as yet obscure functional properties of the hr gene product in maintaining follicle integrity during normal catagen.

Clusters of perifollicular macrophages in normal murine skin: physiological degeneration of selected hair follicles by programmed organ deletion.

In back skin sections from adolescent C57BL/6 mice, regularly distributed, perifollicular inflammatory cell clusters (PICC) were found located around the distal noncycling portion of about 2% of all hair follicles examined. The PICC and the affected hair follicles were characterized during spontaneously developed or induced hair cycle stages, using antibodies against MHC Class II, F4/80, ER-MP23, NLDC 145, CD4, CD8, gammadeltaTCR, IL-1 receptor, and ICAM-1. PICC consisted predominantly of macrophages (MAC), accompanied by a few CD4+ cells, whereas gammadeltaTCR+ and CD8+ cells were absent. During anagen and catagen, some of the PICC+ hair follicles showed variable degenerative phenomena reminiscent of scarring alopecia: thickened basement membrane, ectopic MHC II expression, MAC infiltration into the follicle epithelium, and signs of keratinocyte apoptosis. Loss of distal outer root sheath keratinocytes was detected in 10% of PICC+ hair follicles (0.2% of all hair follicles). Because PICC were located in the vicinity of the bulge region, MAC-dependent damage to follicle stem cells might eventually lead to follicle degeneration. These perifollicular MAC clusters around selected hair follicles may indicate the existence of a physiological program of MAC-dependent controlled follicle degeneration by which damaged or malfunctioning follicles are removed by programmed organ deletion (POD).

Analysis of apoptosis during hair follicle regression (catagen)

Keratinocyte apoptosis is a central element in the regulation of hair follicle regression (catagen), yet the exact location and the control of follicular keratinocyte apoptosis remain obscure. To generate an "apoptomap" of the hair follicle, we have studied selected apoptosis-associated parameters in the C57BL/6 mouse model for hair research during normal and pharmacologically manipulated, pathological catagen development. As assessed by terminal deoxynucleotide transferase dUTP fluorescein nick end-labeling (TUNEL) stain, apoptotic cells not only appeared in the regressing proximal follicle epithelium but, surprisingly, were also seen in the central inner root sheath, in the bulge/isthmus region, and in the secondary germ, but never in the dermal papilla. These apoptosis hot spots during catagen development correlated largely with a down-regulation of the Bcl-2/Bax ratio but only poorly with the expression patterns of interleukin-1beta converting enzyme, p55TNFR, and Fas/Apo-1 immunoreactivity. Instead, a higher correlation was found with p75NTR expression. During cyclophosphamide-induced follicle dystrophy and alopecia, massive keratinocyte apoptosis occurred in the entire proximal hair bulb, except in the dermal papilla, despite a strong up-regulation of Bax and p75NTR immunoreactivity. Selected receptors of the tumor necrosis factor/nerve growth factor family and members of the Bcl-2 family may also play a key role in the control of follicular keratinocyte apoptosis in situ.

Toxicity of polyunsaturated fatty acid esters for human monocyte-macrophages: the anomalous behaviour of cholesteryl linolenate.

We have investigated the toxicity to human monocytemacrophages, and susceptibility to oxidation, of different individual dietary fatty acids in cholesterol esters and triglycerides, added to the cell cultures as coacervates with bovine serum albumin. Toxicity was assessed using release of radioactivity from cells preloaded with tritiated adenine. Lipid oxidation was measured by gas chromatography (GC). The triglycerides showed a direct relationship between toxicity and increasing unsaturation, which in turn correlated with increasing susceptibility to oxidation. Triolein (18:1; omega-9) and trilinolein (18:2; omega-6) were non-toxic. Trilinolenin (18:3; omega-3) was toxic only after prolonged incubation. Triarachidonin (20:4; omega-6), trieicosapentaenoin (20:5; omega-3) and tridocosahexaenoin (22:6; omega-3) were profoundly and rapidly toxic. There was a similar relationship between toxicity and increasing unsaturation for most of the cholesterol esters, but cholesteryl linolenate was apparently anomalous, being non-toxic in spite of possessing three double bonds and being extensively oxidised. Probucol and DL-alpha-tocopherol conferred protection against the toxicity of cholesteryl arachidonate and triarachidonin. The oxidation in these experiments was largely independent of the presence of cells. GC indicated that formation of 7-oxysterols might contribute to the toxicity of cholesteryl linoleate. The toxicity of triglycerides suggests that polyunsaturated fatty acid peroxidation products are also toxic. Possible mechanisms of cytotoxicity and relevance to atherosclerosis are discussed.

The carotenoids beta-carotene, canthaxanthin and zeaxanthin inhibit macrophage-mediated LDL oxidation.

Human monocyte-macrophages were incubated for 24 h in Ham's F-10 medium with human low-density lipoprotein (LDL) in the presence or absence of beta-carotene, canthaxanthin or zeaxanthin, at final concentrations of 2.5, 12.5 and 25 mg/l. LDL oxidation, measured by agarose gel electrophoresis, the thiobarbituric acid assay and gas chromatography, was inhibited by each of the carotenoids in a concentration-dependent manner. Canthaxanthin was more effective when incorporated into LDL before addition to the cultures whereas beta-carotene and zeaxanthin were more effective when added simultaneously with LDL. The results suggest that dietary carotenoids might help slow atherosclerosis progression.

Oxidation of low-density lipoprotein by human monocyte-macrophages results in toxicity to the oxidising culture.

Human monocyte-macrophage cultures were exposed to native low density lipoprotein (LDL) for up to 24 h in Ham's F10 medium and the extent of cell-mediated LDL oxidation was determined by measurement of electrophoretic mobility on agarose gels and measurement of lipids and oxidised lipids (including 7 beta-hydroxycholesterol) by GC. After an initial lag phase, which varied from 2-8 h, there was a steady increase in oxidation over 24 h. No-cell control incubations showed minimal increases in oxidation over 24 h. Significant toxicity, measured as release of radioactivity from macrophages pre-loaded with tritiated adenine, was observed in the cells when they oxidised LDL and the extent of radioactivity release correlated closely with the extent of LDL oxidation. Inhibition of oxidation using alpha-tocopherol or probucol reduced toxicity within the oxidising culture. This self-inflicted toxicity may help to explain the origin and enlargement of the lipid core of advanced atherosclerotic lesions.

Macrophages, lipid oxidation, ceroid accumulation and alpha-tocopherol depletion in human atherosclerotic lesions.

Necropsy samples of atherosclerotic lesions of different histological stages have been analysed. Ceroid was present in all the lesions, within lipid-laden macrophage foam cells and extracellularly in the atheromatous core of advanced lesions. Mean levels of 7 beta-hydroxycholesterol, 26-hydroxycholesterol and hydroxyoctadecadienoic acids were all significantly greater in lesions than in normal intima. Levels of hydroxycholesterols were very low or undetectable in normal intima. Fatty streaks showed the highest ratio of 7 beta-hydroxycholesterol to cholesterol, and the lowest ratio of linoleate to oleate, suggesting that this type of lesion experiences the greatest free radical activity. Levels of 26-hydroxycholesterol, a product of the cytochrome P-450 enzyme sterol 26-hydroxylase, and the ratio of 26-hydroxycholesterol to cholesterol were significantly higher in advanced lesions than in intermediate lesions or fatty streaks. The ratio of alpha-tocopherol to cholesterol levels varied widely in normal intima but was consistently low in lesions, especially those rich in macrophage foam cells, suggesting that oxidative activity in the lesion may lead to significant oxidation of the lesion constituents only after alpha-tocopherol has been depleted. Macrophage death was a characteristic feature of advanced lesions, with apoptotic bodies present, and occasionally, intact apoptotic cells were seen in lesions. These striking correlations between macrophages, lipid oxidation, alpha-tocopherol depletion, ceroid accumulation, and macrophage death in advanced lesions, strongly support a role for oxidative damage in atherosclerosis, and lend credence to the idea that alpha-tocopherol dietary supplementation may slow the progression of atherosclerosis in humans.

Oxidized low-density lipoprotein is cytotoxic to human monocyte-macrophages: protection with lipophilic antioxidants.

Human monocyte-macrophages were incubated for 24 h with low-density lipoprotein (LDL) which had been previously oxidized for varying periods up to 24 h with copper ions, in the presence or absence of DL-alpha-tocopherol or probucol. The release of radioactivity from cells preloaded with tritiated adenine was used as an assay of toxicity. Toxicity of oxidized LDL increased with duration of copper oxidation and with increasing evidence of lipid oxidation, measured by assay of thiobarbituric acid-reactive substances and by gas chromatography. Oxidation and toxicity were inhibited by DL-alpha-tocopherol (200 microM) and probucol (50 microM).

Inhibition of mitochondrial protein synthesis influences the glucocorticoid sensitivity of lymphoid cells.

Inhibition of mitochondrial protein synthesis impairs the formation of the 13 polypeptides encoded on the mitochondrial genome. These polypeptides are part of enzyme complexes involved in oxidative phosphorylation. Prolonged inhibition of mitochondrial protein synthesis thus reduces the oxidative phosphorylation capacity which ultimately results in impairment of energy-requiring processes. Via a different mechanism glucocorticoid hormones also decrease the oxidative phosphorylation capacity of, e.g., lymphoid cells. The present study shows that inhibition of mitochondrial protein synthesis influences glucocorticoid-induced responses of lymphoid cells in two opposing manners. (a) It is enhanced after induction in cells with a reduced oxidative phosphorylation capacity resulting from preceding inhibition of mitochondrial protein synthesis. This can be explained by the synergistic effects of glucocorticoids and prolonged inhibition of mitochondrial protein synthesis on energy-producing processes. (b) It is counteracted when mitochondrial protein synthesis is impaired during induction of the response. The latter observation suggests that mitochondrial protein synthesis is involved in the generation of glucocorticoid-induced effects on lymphoid cells.