Compromised epidermal barrier stimulates Harderian gland activity and hypertrophy in ACBP-/- mice.

Acyl-CoA binding protein (ACBP) is a small, ubiquitously expressed intracellular protein that binds C14-C22 acyl-CoA esters with very high affinity and specificity. We have recently shown that targeted disruption of the Acbp gene leads to a compromised epidermal barrier and that this causes delayed adaptation to weaning, including the induction of the hepatic lipogenic and cholesterogenic gene programs. Here we show that ACBP is highly expressed in the Harderian gland, a gland that is located behind the eyeball of rodents and involved in the production of fur lipids and lipids used for lubrication of the eye lid. We show that disruption of the Acbp gene leads to a significant enlargement of this gland with hypertrophy of the acinar cells and increased de novo synthesis of monoalkyl diacylglycerol, the main lipid species produced by the gland. Mice with conditional targeting of the Acbp gene in the epidermis recapitulate this phenotype, whereas generation of an artificial epidermal barrier during gland development reverses the phenotype. Our findings indicate that the Harderian gland is activated by the compromised epidermal barrier as an adaptive and protective mechanism to overcome the barrier defect.


Hematoxylin and eosin staining
Paraffi n sections (5 µm) were mounted on Super Frost Plus ® microscopic slides. Before staining, sections were dried for 60 min at 60°C, paraffi n was removed with xylene, and sections were rehydrated with graded ethanol. Slides were then incubated in Mayer's hematoxylin (1 mg/ml in water) for 10 min and quickly in 1% HCl in 70% ethanol and rinsed for 10 min in fl owing tap water before staining with eosin (2 min in 2% eosin in water). After staining, the slides were rinsed in water, dehydrated (2 × 1 min in 96% ethanol, 3 × 1 min in 99% ethanol, 2 × 5 min in xylene), and subsequently mounted.

Osmium tetroxide and toluidine blue staining
Adult mice were submitted to perfusion fi xation with 2.5% glutaraldehyde in PBS (pH 7.4). Biopsies were postfi xed in 1% osmium tetroxide (pH 7.2) for 2 h at 4°C, dehydrated in ethanol, and embedded in Araldite. Semithin sections (0.5 µm) were stained with toluidin blue for light microscopy.

Morphometry
The area of acinar cells, the number of vesicles per cell and per µm 2 , and the diameter of the vesicles were measured in 0.5 µm sections stained with osmium tetroxide and toluidin blue using the Leica IM500 measurement application (100× magnifi cation). Sections of glands from 3 month old ACBP Ϫ / Ϫ (n = 6) and ACBP +/+ (n = 8) mice were evaluated. For each animal, the size and number of vesicles in a minimum of 10 acinar cells were determined.

De novo lipid synthesis and total lipid quantifi cation in vivo
For determination of de novo lipid synthesis in the Harderian gland, adult mice (n = 8 for each genotype and condition) were injected i.p. with 0.04 µCi/g mouse 14 C-acetic acid ([1- 14 C] acetic acid, sodium salt; Amersham Biosciences) in 0.9% saline and euthanized 1 h later by cervical dislocation. The Harderian glands were removed, snap-frozen in liquid nitrogen, and stored at Ϫ 80°C. Total lipids were extracted from one Harderian gland using standard Bligh and Dyer extraction ( 29 ). Lipids from equal amounts of tissue were loaded onto silica plates (Merck High Performance Thin-Layer Chromatography [HPTLC]) and left to air dry for 15 min. The plates were developed in a horizontal development chamber in a solvent mixture of hexane-diethyl etheracetic acid (v/v/v) (80:30:1) for the separation of neutral lipids. Band intensities were determined by densitometry. The visualization of 14 C-labeled lipids and the total lipids was carried out as previously described ( 13 ). of triacylglycerol and cholesteryl esters ( 13 ). Intriguingly, we recently showed that K14-ACBP Ϫ / Ϫ mice recapitulate the delayed hepatic adaptation to weaning, demonstrating that lack of ACBP in the epidermis leads to suppression of the lipogenic and cholesterogenic gene programs in the liver. Importantly, the hepatic phenotype could be rescued by establishing an artifi cial barrier using either Vaseline or latex, thereby for the fi rst time establishing a physiological link between the epidermal barrier function and hepatic gene expression and metabolism ( 18 ). The Harderian gland, originally discovered by Harder in 1694, is located behind the eyeball of most terrestrial vertebrates. The gland contains more than 20% lipids per wet weight, where neutral lipids constitute 55-57% of total lipids ( 19 ). The main lipid species in the murine Harderian gland is monoalkyl diacylglycerol (MADAG) ( 19,20 ), which is secreted in an exocrine fashion onto the external surface of the cornea ( 21 ). Although the function(s) of the Harderian gland is still not completely understood, it is well established that lipids produced in the Harderian gland lubricate and protect the cornea and are used for grooming of the fur ( 22 ). In some species, lipids from the Harderian gland have been shown to provide thermoregulation (23)(24)(25). Thus, harderianectomy leads to a decrease in the ability of the muskrat and Mongolian gerbil to defend their core body temperature in cold water ( 23,24 ). It was shown that the majority of the fur lipids originate from the Harderian gland in Mongolian gerbils; however, similar fi ndings could not be observed in rats, hamsters, or mice ( 23 ). Thus, the function of the Harderian gland in thermo-protection in other rodents is unclear. In addition, the gland has been implicated in pheromone production and sexual behavior in some rodents (26)(27)(28).
Here we show that lack of ACBP in mice stimulates the activity and hypertrophy of the Harderian gland during weaning, and we present evidence indicating that this occurs as a defense mechanism in response to the impaired epidermal barrier of these mice.

Generation and maintenance of mice
Generation of mice with constitutive ( 13 ) and conditional ( 18 ) targeting of the Acbp gene have previously been described. The mice were housed under standard conditions at ‫ف‬ 55% relative humidity at 22 ± 3°C as described elsewhere ( 13 ). Cold exposure was carried out in a 4°C refrigerated room with normal light cycles and ad libitum access to food. Vaseline was applied as described previously ( 18 ). In brief, Vaseline was applied on the core body of the mice twice a day from 7 days of age to 28 days of age. The mice were weaned at 21 days of age as described ( 13 ). Breeding of transgenic mice and all animal experiments were approved by the Danish Animal Experiment Inspectorate.

Real-time PCR
For determination of mRNA levels, tissues were processed, and real-time PCR was carried out as previously described ( 13 ). Primers used are: Acbp fwd: 5 ′ TTTCGGCATCCGTATCACCT; Acbp rew: 5 ′

Ex vivo secretion of de novo synthesized MADAG
Harderian glands were isolated from adult ACBP Ϫ / Ϫ (n = 7) and ACBP +/+ (n = 7) mice and washed in Krebs ringer buffer (Sigma) containing 0.5% fatty acid-free BSA (Sigma). Immediately thereafter, similar-sized explants (approximately half a gland) were incubated for 2 h in Krebs ringer buffer with 1% fatty acid-free BSA containing 0.5 µCi/ml 14 C-acetic acid ([1- 14 C] acetic acid, sodium salt; Amersham Biosciences) at 37°C. Explants were removed and washed, and the weight was determined. Lipids were extracted from the media as described above. The extracted lipids were diluted relative to the explant weight (equal volume per tissue weight) and analyzed by HPTLC as described above.

Choline level in the Harderian gland
The free level of choline in the Harderian gland was measured using a Choline/Acetylcholine Quantifi cation Kit (MAK056-1KT; Sigma). Harderian glands from adult mice were homogenized in supplied assay buffer, and the level of choline was determined according to the manufacturer by fl uorometry using a FLOUstar Omega fl uorescence plate reader (BMG Labtech) and normalized to protein content as determined by a Pierce TM BCA Protein Assay kit (23227; Thermo Scientifi c).

Plasma triiodothyronine
Plasma from fed adult mice were collected and separated as described elsewhere ( 13 ). The triiodothyronine (T3) level was analyzed using a FLOUstar Omega fl uorescence plate reader (BMG Labtech) and a mouse/rat T3 ELISA kit (T3043T-100, Calbiotech).

ACBP is expressed at a high level in the Harderian gland ACBP
Ϫ / Ϫ mice consistently display swollen eyelids and increased secretion from the eye cavity ( Fig. 1A ). This phenotype is visible from around weaning and persists throughout adulthood. Interestingly, when we examined the eye cavity of adult mice, we found that the lipid-producing Harderian gland is approximately twice as large in ACBP Ϫ / Ϫ mice compared with ACBP +/+ mice ( Fig. 1B ). Examination of the expression of ACBP in this gland showed that, similar to other lipogenic tissues ( Fig. 1E ) ( 4,30,31 ), ACBP is expressed at very high levels in Harderian gland of ACBP +/+ mice. In contrast, ACBP is completely absent in the Harderian gland of ACBP Ϫ / Ϫ mice ( Fig. 1C, D ). In mice, the Harderian gland is not fully developed at birth. It becomes functional from around 14 days of age when the pups open their eyes; however, the gland continues to grow and mature until adulthood ( 32 ). We found that the size of the gland is similar in ACBP Ϫ / Ϫ and ACBP +/+ mice until just before weaning at 21 days of age ( Fig. 1F ). However, in the week after weaning, the size of the Harderian glands of ACBP Ϫ / Ϫ mice doubles, whereas gland size in ACBP +/+ mice remains relatively constant. This is true both when gland size is normalized to body weight ( Fig. 1F ) and when it is expressed only as gland weight (supplementary Fig. 1). This indicates that weaning initiates systemic and/or endogenous signals in the gland that leads to a major and rapid hypertrophy of the gland, which is maintained in adulthood.

month old ACBP
Ϫ / Ϫ and ACBP +/+ male mice. RNA was quantifi ed using real-time PCR (n = 8). Error bars represent SEM. *** P < 0.001. Protein expression was visualized by Western blotting of protein extracts pooled from fi ve animals. E: Comparison of ACBP expression in Harderian glands relative to other tissues with high expression of ACBP. ACBP expression was visualized by Western blotting of protein extract (pooled from four mice). F: Harderian glands are enlarged in ACBP Ϫ / Ϫ mice from weaning at 21 days of age. Harderian glands were isolated from mice (n = 3) at the indicated age, and the weight was determined and normalized to bodyweight. Error bars indicate mean ± SEM. * P < 0.05; ** P < 0.01. Signifi cance was determined using unpaired parametric Student's t -test. BAT, brown adipose tissue; HG, Harderian gland; WAT, white adipose tissue (inguinal). a decrease in the amount of MADAG, the main lipid product of the Harderian gland ( 19,20 ), in glands from ACBP Ϫ / Ϫ mice compared with the glands from ACBP +/+ mice ( Fig. 3A and data not shown). The fatty acids used for synthesis of MADAG originate primarily from de novo fatty acid synthesis in the gland ( 34,35 ). In vivo pulselabeling with 14 C-acetic acid showed that de novo synthesis of MADAG in Harderian glands is signifi cantly higher in ACBP Ϫ / Ϫ mice compared with ACBP +/+ mice ( Fig. 3B ). However, this increase occurred in the absence of changes in the mRNA expression of key enzymes involved in synthesis of MADAG ( 35 ), such as dihydroxyacetone phosphate acyltransferase 1-alkyl-dihydroxyacetone phosphate synthase (supplementary Fig. 2A, B). Similarly, there was no increase in the expression of genes encoding key enzymes in the synthesis of fatty acids, such as fatty acid synthase or acetyl-CoA carboxylase (supplementary Fig. 2C, D). This indicates that the increase in the synthesis of MADAG is not driven by increased expression of enzymes involved in synthesis of fatty acids but rather by increased activity of The Harderian gland consists of a simple tubuloalveolar system with acinar cells arranged in acini ( 33 ). This overall structure is maintained in the ACBP Ϫ / Ϫ mice ( Fig. 2A ); however, the acinar cells from ACBP Ϫ / Ϫ mice are larger ( Fig. 2C ) and contain fewer ( Fig. 2A , E, F) but signifi cantly larger vesicles ( Fig. 2D ) compared with glands from ACBP +/+ mice. Furthermore, lipid staining with osmium tetroxide demonstrated that the vesicles and the lumen contained more lipid in the ACBP Ϫ / Ϫ glands compared with ACBP +/+ glands ( Fig. 2B ). These data indicate that lack of ACBP causes hypertrophy and increased activity of the Harderian gland.

Depletion of ACBP leads to increased secretion of MADAG from the Harderian gland
The increased size and lipid fi lling of the vesicles in the Harderian gland of ACBP Ϫ / Ϫ mice prompted us to investigate the lipid content of the gland. We therefore carefully dissected the gland and extracted the lipids. Using HPTLC, we found that the major overall change in lipid classes was the gland is lower in ACBP Ϫ / Ϫ mice than in ACBP +/+ mice, suggesting that secretion of MADAG from the gland exceeds the increase in synthesis rate. To further investigate this, we isolated Harderian glands and exposed similar-sized explants to 2 h ex vivo pulse-labeling with 14 C-acetic acid. Analysis of the explant medium showed that secretion of MADAG is signifi cantly higher for ACBP Ϫ / Ϫ compared with ACBP +/+ explants ( Fig. 3C ).

The effect of thermoregulation on the activity of the Harderian glands
We have previously shown that ACBP Ϫ / Ϫ pups tolerate weaning less well and need to be housed with littermates and soaked food in the bottom of the cage to survive ( 13 ). We therefore speculated that the hypertrophy and increased activity of the Harderian gland may be caused by cold perception. To investigate whether cold exposure per se could induce Harderian gland activity, we exposed adult ACBP +/+ mice to cold (4°C) for 15 days. Interestingly, this resulted in a modest but signifi cant increase in Harderian gland size ( Fig. 4A ) and in a reduction in total MADAG level ( Fig. 4B ) when compared with Harderian glands of ACBP +/+ mice kept at room temperature. However, de novo MADAG synthesis, as determined by in vivo pulse-labeling with 14 C-acetic acid, was not affected by cold exposure ( Fig. 4C ). This indicates that cold perception can indeed lead to an increase in gland size but that the increase in size and activity is much less dramatic than what is observed in the ACBP Ϫ / Ϫ mice around weaning.

Endogenous effects of ACBP depletion on the Harderian gland
We recently reported that mice with conditional targeting of the Acbp gene in keratinocytes (K14-ACBP Ϫ / Ϫ ), similar to ACBP Ϫ / Ϫ mice, display a compromised epidermal barrier and an increased transepidermal water loss ( 14,18 ). Furthermore, the K14-ACBP Ϫ / Ϫ mice recapitulate the delayed hepatic adaption during weaning previously observed in ACBP Ϫ / Ϫ mice ( 13,18 ). Importantly, we were able to rescue the hepatic phenotype by creating an artificial epidermal barrier using Vaseline or latex, thereby demonstrating that it is the impaired barrier per se, possibly through an increased heat loss, that leads to changes in hepatic gene expression and metabolism during weaning.
To investigate if the imperfect epidermal barrier may also cause Harderian gland hypertrophy and increased activity at weaning, we took advantage of the K14-ACBP ؊ / ؊ mice ( 18 ). Careful examination of the K14-ACBP Ϫ / Ϫ mice revealed that ACBP was not only depleted in the keratinocytes but also in the Harderian glands ( Fig. 5A, B ). This is consistent with a previous report from Ntambi and coworkers reporting that the K14 promoter drives CRE-mediated disruption of the Scd1 locus in the Harderian gland in addition to the epidermis ( 41 ). Interestingly, examination of Harderian glands from K14-ACBP ؊ / ؊ mice revealed that these are also signifi cantly larger than those from control mice ( Fig. 5C ). Furthermore, similar to Harderian glands from ACBP Ϫ / Ϫ mice, glands from K14-ACBP ؊ / ؊ these enzymes and/or increased substrate fl ux. Choline and T3 have been shown to stimulate secretion from the Harderian gland in rats (36)(37)(38)(39)(40). However, neither plasma levels of T3 nor glandular choline levels were affected by ACBP depletion at 28 days of age or in adult mice ( Fig. 3D, E ). This indicates that the observed increase in Harderian gland activity in ACBP-depleted mice is not caused by changes in choline or T3 levels.
Taken together, these data show that the enlarged Harderian glands in ACBP Ϫ / Ϫ mice synthesize more MADAG per weight gland compared with glands from ACBP +/+ mice. Despite this, the total MADAG content of Harderian glands of ACBP +/+ and ACBP Ϫ / Ϫ mice (n = 8 per genotype). Lipids were extracted from similar amounts of Harderian glands, and samples were diluted to obtain equal volume per tissue weight in all samples. Neutral lipids were analyzed by HPTLC, and total MADAG was determined by densitometry and plotted relative to the ACBP +/+ levels. Data are presented as mean ± SEM. ** P < 0.01. B: De novo synthesis of MADAG in adult male mice (n = 8 per genotype) was determined by in vivo pulse-labeling with 14 C-acetic acid for 2 h. Lipids were extracted from similar amounts of Harderian gland, and samples were diluted to obtain equal volume per tissue weight. De novo MADAG synthesis was determined by quantifi cation of radioactivity in MADAG bands and expressed relative to the mean value in Harderian glands of ACBP +/+ mice. Data are presented as mean ± SEM. *** P < 0.001. C: Secretion of de novo synthesized MADAG from Harderian gland explants from adult male mice (n = 7 per genotype) was determined by ex vivo pulselabeling with 14 C-acetic acid for 2 h. Lipid were extracted from the media and quantifi ed by HPTLC as above. MADAG secretion from Harderian gland explants is expressed relative to the mean normalized secretion from ACBP +/+ explants. Data are presented as mean ± SEM. ** P < 0.01. D: The level of choline in the Harderian glands was determined in adult (n = 9) and 28 day old (n = 5 Ϫ 6) ACBP +/+
response that stimulates the gland to enlarge and to produce more protective lipid for secretion on the fur. Consistent with that, we have shown that there is signifi cantly more total lipid, including MADAG, on the fur of ACBP Ϫ / Ϫ mice than on the fur of ACBP +/+ mice ( 14 ).

DISCUSSION
In this study, we show that targeted disruption of ACBP leads to hypertrophy and increased activity of the Harderian gland, and we demonstrate that this occurs as a result of the impaired epidermal barrier of the ACBP Ϫ / Ϫ mice. The hypertrophy of the Harderian gland is induced around and immediately subsequent to weaning, indicating that it occurs as a result of the challenges faced by the ACBP Ϫ / Ϫ mice at weaning.
Morphometric measurements indicated that the increased gland size found in the ACBP Ϫ / Ϫ Harderian gland is the result of marked hypertrophy and possibly also of modest hyperplasia of the acinar cells. Furthermore, there are several indications that ACBP Ϫ / Ϫ glands are more active and secrete more lipids. First, although total content of MADAG was decreased in glands from ACBP Ϫ / Ϫ mice, we show that there is a signifi cant increase in the de novo synthesis of MADAG in these glands. Second, vesicles in ACBP Ϫ / Ϫ glands are larger and contain more lipids compared with vesicles in ACBP +/+ mice, a phenomenon also observed in choline-stimulated Harderian glands ( 21,39,40 ). An estimate of the percentage of the gland volume taken up by vesicles indicates that the relative vesicle volume is slightly increased in glands from ACBP Ϫ / Ϫ mice relative to ACBP +/+ mice. It is possible that MADAG is more mice contain lower levels of MADAG ( Fig. 5D ) and display a signifi cantly increased de novo synthesis of MADAG ( Fig.  5E ). Collectively, these data indicate that the phenotype of the Harderian gland in ACBP Ϫ / Ϫ mice is caused primarily by the lack of ACBP in the epidermis and/or in the Harderian gland itself.

Impairment of the epidermal barrier activates the Harderian gland
To investigate whether the impaired epidermal barrier in ACBP Ϫ / Ϫ and K14-ACBP Ϫ / Ϫ mice was causally related to the phenotype of the Harderian gland, we examined the effect of Vaseline on gland size. Vaseline was applied to the core body of the mice twice a day from 7 to 28 days of age to generate an artifi cial epidermal barrier during the time when the glands of ACBP Ϫ / Ϫ mice experience the most dramatic increase in gland size ( Fig. 1F ). Interestingly, Vaseline occlusion of the epidermis eliminates the size difference between Harderian glands from 28 day old ACBP +/+ and ACBP Ϫ / Ϫ mice ( Fig. 6A ), indicating that it is the compromised epidermal barrier that stimulates the hypertrophy of the gland in ACBP Ϫ / Ϫ mice around weaning. Importantly, Vaseline also normalizes total MADAG content ( Fig. 6B ) and de novo synthesis of MADAG in these mice ( Fig. 6C ). These results show for the fi rst time that an impaired epidermal barrier can stimulate hypertrophy and activity of the Harderian gland and fl ux of MADAG from the Harderian gland. The exact signaling pathways and molecular mechanisms underlying this dramatic effect on the gland remains unclear; however, it is likely that water and/or heat loss across the epidermis initiates a protective systemic  14 C-acetic acid. Lipids were extracted from Harderian glands, and samples were diluted to obtain equal volume per tissue weight in all samples. De novo MADAG synthesis was determined by quantifi cation of radioactivity in MADAG bands and expressed relative to the mean value in Harderian glands of ACBP +/+ mice maintained in standard housing conditions. Data are presented as mean ± SEM. ** P < 0.01. Signifi cance was determined using unpaired parametric Student's t -test (n.s., nonsignifi cant).
of ACBP +/+ mice ( 14 ). Finally, we show that glandular explants from ACBP Ϫ / Ϫ mice secrete signifi cantly more MADAG than explants from ACBP +/+ mice. Taken together, these fi ndings strongly indicate that Harderian glands from ACBP Ϫ / Ϫ mice are not only larger but also more active than glands from ACBP +/+ mice. Thus, although the C-acetic acid. Lipids were extracted from Harderian glands, and samples were diluted to obtain equal volume per tissue weight in all samples. De novo MADAG synthesis was determined by quantifi cation of radioactivity in MADAG bands and expressed relative to the mean value in Harderian glands of control mice. De novo MADAG synthesis is displayed relative to mean values in control mice. Data are presented as mean ± SEM. *** P < 0.01. Signifi cance was determined using unpaired parametric Student's t -test.  was determined by pulse-labeling with 14 C-acetic acid. Lipids were extracted from Harderian glands, and samples were diluted to obtain equal volume per tissue weight in all samples. De novo MADAG synthesis was determined by quantifi cation of radioactivity in MADAG bands and expressed relative to the mean value in Harderian glands of control ACBP +/+ mice. Data are presented as mean ± SEM. * P < 0.05. Signifi cance was determined using unpaired parametric Student's t -test.

rapidly packaged into vesicles in ACBP
Ϫ / Ϫ mice. Third, we have previously shown that there is three times more MADAG on the fur of ACBP Ϫ / Ϫ mice compared with that relative contribution of lipids from the Harderian gland to fur lipid in mice is unclear ( 23,42 ), our results indicate that stimulation of the Harderian gland could indeed contribute to fur lipids. We show that, similar to other lipogenic cells ( 4,5,30,31,43 ), the acinar cells of the Harderian gland express very high levels of ACBP. We therefore initially speculated that the Harderian gland phenotype was caused directly by failure to express ACBP in these cells. This hypothesis was consistent with the fi nding that the K14 promoter is also active in the acinar cells and that K14-ACBP Ϫ / Ϫ mice recapitulate the Harderian gland phenotype. Importantly, however, the Harderian gland phenotype with increased gland size and activity could be rescued by Vaseline occlusion of the epidermis before and during weaning. Thus, although we cannot exclude a contribution from endogenous lack of ACBP in the Harderian gland, our results show for the fi rst time that hypertrophy and increased activity of the Harderian gland is caused primarily by the compromised epidermal barrier. It is currently unclear how the imperfect barrier signals to the Harderian gland to increase activity. Adrenergic nerve fi bers and cholinergic fi bers are associated with blood vessels and secretory and myoepithelial cells in the murine Harderian gland, respectively ( 44 ). Increased levels of both choline and the thyroid hormone T3 has been shown to lead to increased lipid secretion from the Harderian gland in rats ( 21,36,39 ). However, ACBP Ϫ / Ϫ mice neither presented increased glandular levels of choline nor increased plasma levels of T3 ( Fig. 3D, E ), indicating that these pathways are not responsible for the increased activity of the gland in ACBP Ϫ / Ϫ mice. Thus, it appears that the epidermis signals to promote gland activity and hypertrophy through alternative mechanisms. Whatever the signals are, our results indicate that the potential to induce gland hypertrophy and activity is greater in the days after weaning compared with in adult mice where prolonged cold exposure only led to a minor increase in gland size but no increase in de novo MADAG synthesis. A possible reason for this could be that the Harderian gland is more sensitive to activating signals in pups than in adult mice. Alternatively, ACBP Ϫ / Ϫ and K14-ACBP Ϫ / Ϫ pups perceive a stronger cold signal at room temperature than adult ACBP +/+ mice do at 4°C.
In conclusion, we have demonstrated that the Harderian gland of ACBP Ϫ / Ϫ mice displays signifi cant hypertrophy and increased activity, which is instigated around weaning and is caused by the impaired epidermal barrier of these mice. This represents the fi rst direct demonstration that impairment of the epidermal barrier can stimulate the activity of this gland. We propose that this occurs as a protective mechanism, in particular in pups, to increase the production of fur lipids and thereby overcome the barrier defect.
The authors thank laboratory technicians Yoanna Vladimirova and Brigita Meskauskaite for technical assistance.