Canine epidermal lipid sampling by skin scrub revealed variations between different body sites and normal and atopic dogs
Mandy Angelbeck-Schulze1, Reinhard Mischke1, Karl Rohn2, Marion Hewicker-Trautwein3, Hassan Y. Naim4, Wolfgang Bäumer5
1Small Animal Clinic, University of Veterinary Medicine Hannover, Foundation, 30559 Hanover, Germany
2Department of Biometry, Epidemiology and Information Processing, University of Veterinary Medicine Hannover, Foundation, 30559 Hanover, Germany
3Department of Pathology, University of Veterinary Medicine Hannover, Foundation, 30559 Hanover, Germany
4Department of Biochemistry, University of Veterinary Medicine Hannover, Foundation, 30559 Hanover, Germany
5Department of Pharmacology, Toxicology and Pharmacy, University of Veterinary Medicine Hannover, Foundation, 30559 Hanover, Germany
Correspondence:
Reinhard Mischke, Small Animal Clinic, University of Veterinary Medicine Hannover, Foundation, Buenteweg 9, 30559 Hanover, Germany
Sources of Funding: This study was supported by "Gesellschaft zur Förderung Kynologischer Forschung e.V.".
5.1 Abstract
In a previous study, we evaluated a minimally-invasive epidermal lipid sampling method called skin scrub, which was highly reproducible and produced comparable results to tape stripping. The present study aimed the investigation of regional variations in canine epidermal lipid composition and the suitability testing of the skin scrub technique for sampling of dogs with certain skin diseases. Eight different body sites, five of which with high and three of which with low predisposition to atopic lesions, of eight control dogs as well as lesional and non-lesional skin of 12 atopic dogs and of four dogs with other skin diseases were sampled by the skin scrub technique. Lipid fractions were separated by high performance thin layer chromatography and analysed densitometrically. Contrary to the lipid composition, no significant differences in total lipid content were found between the body sites tested of the control dogs. The pinna, the flew and the caudal back contained significantly lower concentrations of ceramides, whereas the palmar metacarpus and the axillary region contained significantly higher amounts of ceramides and cholesterol than most other body sites. The amounts of total lipids and ceramides including all ceramide classes were significantly lower in both lesional and non-lesional skin of atopic dogs compared to normal skin, the reduction being more pronounced in lesional skin. The sampling by the skin scrub technique was relatively painless and caused only slight erythema at the sampled areas, but no oedema. In conclusion, the present study revealed certain regional variations in the epidermal lipid and ceramide composition of dogs without skin abnormalities, but no relations between the lipid composition and predilection sites for canine atopic dermatitis lesions were detected. The skin scrub technique turned out to be a practicable sampling method for canine epidermal lipids and revealed satisfying results regarding alterations of skin lipid composition in canine atopic dermatitis. This sampling method might be suitable for epidermal lipid investigations of further canine skin diseases.
5.2 Introduction
During the last years, research on canine epidermal lipids attracted increasing interest, especially concerning canine atopic dermatitis (AD). Since ceramides (CER) are a major stratum corneum (SC) lipid fraction and thought to be responsible for skin barrier homeostasis in mammals (CODERCH et al. 2003; HOLLERAN et al. 2006), the focus of most studies was on this group of sphingolipids. Eleven classes of CER were described in human and canine SC until now (MASUKAWA et al. 2008; YOON et al. 2011). Consistently, the different CER classes are denominated according to their chemical structure (MOTTA et al. 1993), a sphingoid base - sphingosine (S), dihydrosphingosine (dS), 6-hydroxysphingosine (H) or phytosphingosine (P) -, which is amide linked to a long-chained non- (N), α- (A) or ω-hydroxylated (O) fatty acid. In free ceramides, the ω-hydroxylated fatty acid is further esterified (E), most often with linoleic acid (NOVOTNY et al. 2010). The 11 classes are CER[EOS], CER[NdS], CER[NS], CER[EOP], CER[NP], CER[EOH], CER[AdS], CER[AS], CER[NH], CER[AP] and CER[AH] with increasing polarity.
In human and canine AD, there is evidence of a defective cutaneous permeability barrier (CHOI and MAIBACH 2005; OLIVRY 2011; NISHIFUJI and YOON 2013), partially related to decreased amounts of CER (IMOKAWA 2001; REITER et al.
2009; SHIMADA et al. 2009; POPA et al. 2011; YOON et al. 2011). Structural differences in the SC of atopic dogs do not only exist in affected body areas but also in unaffected skin (MARSELLA et al. 2011). Despite the impaired barrier function seems to be an overall feature in patients suffering from AD, both human and canine AD are characterised by a classic distribution of lesions (MARSELLA and SAMUELSON 2009). Differences in lipid composition of specific body sites of healthy individuals may be an explanation for site predispositions for atopic lesions. In humans, regional variations of skin lipid composition and barrier function were described (LAMPE et al. 1983b; YOSHIKAWA et al. 1994; NORLEN et al. 1999;
KLEESZ et al. 2012), but there was no correlation between lipid composition and barrier properties (NORLEN et al. 1999). Furthermore, no relations between lipid composition and body sites predisposed for atopic lesions were detected
(YOSHIKAWA et al. 1994). In dogs, regional variations of permeability barrier properties were described as well (HIGHTOWER et al. 2010), but there are no studies investigating the lipid composition of more than two different body regions until now. With regard to existing site predispositions for canine AD (JAEGER et al.
2010), the investigation of the lipid composition of these predisposed sites versus usually unaffected sites in normal dogs might reveal interesting facts about the nature of site predispositions.
In a previous study, we were able to show that a minimally-invasive method designated as skin scrub is a suitable method for sampling canine epidermal lipids (ANGELBECK-SCHULZE et al. 2013). Furthermore, we detected some differences in epidermal lipid composition between the inguinal region und the caudal back normal canine skin. The present study aimed at the determination of regional variations in lipid and particularly ceramide composition of eight different body sites in normal dogs. Another important purpose of this study was to test the suitability of the skin scrub technique concerning the detection of deviations in epidermal lipid composition of certain skin diseases.
5.3 Material and methods Study design
Samples from eight different body sites (concave pinna, flew, palmar metacarpus, axillary, lateral thorax, lateral abdomen, inguinal, caudal back) were taken by skin scrub from each of eight dogs with normal skin post mortem. Additionally, skin scrubs were taken from variable lesional and non-lesional areas of 12 atopic dogs as well as of four dogs with selected skin disorders other than atopic dermatitis. Lipids were analysed using high performance thin layer chromatography. The Lower Saxony State Office for Consumer Protection and Food Safety was informed about the study protocol in advance (Reference number 09A665).
Animals
All dogs participating in this study were client-owned. Client consent for study performance was obtained.
The eight dogs with normal skin were euthanized for reasons not related to this study and of different breeds (Labrador retriever, Border collie, Dalmatian, Cairn terrier, four mixes) with a median age of 12.3 years (4.2–13.3 years). None of the dogs had a history of skin disease or any skin lesions. Dogs with administrations of corticosteroids less than eight weeks or any topical treatments less than one week prior to the sampling were ruled out. The integrity of the skin was confirmed by histological examinations of skin biopsies.
The 12 dogs with AD were of different breeds (Jack Russell terrier, German shepherd, Golden retriever, two French bulldogs, American pit bull terrier, Rhodesian ridgeback, Magyar Vizsla, Dalmatian, three mixes) with a median age of 3.6 years (1.0–10.5 years). The diagnosis of canine AD was based on appropriate history and clinical criteria according to FAVROT et al. (2010). Flea bite hypersensitivity, ectoparasite infestation, hormonal imbalances and secondary bacterial or yeast overgrowth were ruled out by the common tests or trial therapies. Each dog had to perform a food elimination diet for at least eight weeks. Finally, the 12 atopic dogs included four dogs with solely food-induced AD, four dogs with non-food-induced AD and four dogs with partly food-induced AD.
Additionally to the 12 atopic dogs, four dogs with other skin diseases participated in this study: A Chow Chow with histologically confirmed sebaceous adenitis, a poodle-mix with contact dermatitis as well as a Small Munsterlander and a Landseer with flea bite hypersensitivity. The flea bite hypersensitivity was diagnosed by compatible clinical signs, the finding of fleas or their excrements on the coat and the complete vanishing of clinical signs after a suitable ectoparasite treatment.
None of the atopic and non-atopic dogs with skin disorders was treated with anti-inflammatory medications for at least eight weeks or shampoos for one week prior to the sampling. One exception was the Landseer with flea bite hypersensitivity, which was treated with a depot methylprednisolone four weeks prior to the sampling. Thus, the sampling was repeated another four weeks later.
Lipid sampling and extraction
Samples of the dead dogs were taken within three hours post mortem from the body sites as described above. The atopic dogs were sampled from a lesional - in nine cases axillary, in two cases inguinal, one abdominal - and a non-lesional - always lateral thorax - area. The remaining four dogs were also sampled from a lesional and a non-lesional area, the Chow-Chow both from the caudal back, the poodle-mix from the ventro-lateral abdomen and the lateral thorax and the Small Munsterlander and the Landseer from inguinal and the lateral thorax, respectively. Each body site was sampled once by skin scrub, as described previously (ANGELBECK-SCHULZE et al.
2013). Briefly, a metal cylinder of 21 mm inner diameter was placed tightly on the skin after clipping the coat cautiously. The epidermal lipids were collected by stirring one millilitre of n-hexane and ethanol 2:1 (v/v) with a roughened glass rod with slight pressure for 30 sec, followed by aspirating the extract into a glass tube. This procedure was done twice on the same area and the samples were pooled.
Immediately after the sampling, the concerned skin areas were washed and a skin lipid complex (Allerderm; Virbac, Glattbrugg, Switzerland) was applied. The specimens were dried in a SpeedVac (Concentrator Plus; Eppendorf, Hamburg, Germany) at 60°C. Lipid extraction was performed according to BLIGH and DYER (1959) by homogenisation of the specimens in methanol, chloroform and distilled water 2:1:0.8 (v/v/v). Adding of one millilitre of chloroform and distilled water, respectively, led to phase separation. The upper hydrophilic phase was discarded, the lower lipophilic phase was dried and stored at -20°C until analysis. All solvents used were of HPLC grade (Carl Roth, Karlsruhe, Germany).
High performance thin layer chromatography (HPTLC)
For chromatographical analysis, HTPLC-plates (silica gel 60, 20 x 10 cm, Merck;
Darmstadt, Germany) were pre-cleaned in chloroform and methanol 1:1 (v/v) and activated in an oven at 110°C for 10 min. The specimens were dissolved in 500 µl of chloroform and methanol 1:1 (v/v), respectively. Ten, five and three microlitres of each specimen as well as ten microlitres of each of six standard mixtures containing
increasing concentrations of corresponding lipid standards were applied by a microlitre syringe (N 701, Hamilton; Bonaduz, Switzerland) in one centimetre distance to the bottom of the plate. The following lipid standards were used:
phosphatidylcholine, phosphatidylethanolamine, sodium cholesteryl sulphate, galactocerebrosides, α-hydroxy fatty acid ceramide (CER[AS]), non-hydroxy fatty acid ceramide (CER[NS]), oleic acid and glyceryl trioleat (triolein) from Sigma-Aldrich (Steinheim, Germany), Ceramide VI (CER[AP]), Ceramide III (CER[NP]) and Ceramide I (CER[EOS]) from Evonik Industries AG (Essen, Germany), cholesterol and cholesteryl stearate from Fluka Chemie AG (Buchs, Switzerland). Adapting the method described by STAHL et al. (2009) to our HPTLC-System, the lipids were separated consecutively in four different development chambers (CAMAG; Berlin, Germany) filled with 50 ml of the following solution systems: 1) chloroform, methanol and acetic acid 80:18:2 (v/v/v) to 4 cm, 2) chloroform, methanol and acetic acid 91.4:4.3:4.3 (v/v/v) to 8 cm, 3) n-hexane, diethylether and acetic acid 72.7:18.2:9.1 (v/v/v) to 9.5 cm, 4) n-hexane to the top. All solvents were of HPLC grade and purchased from Carl Roth and Sigma-Aldrich. The plates were dipped in an aquaeous solution of copper sulphate (75 g/L; Merck) and phosphoric acid (8.5%;
Sigma-Aldrich) for 5 sec, charred in an oven at 170°C for 15 min and scanned (Scanjet G3010, Hewlett-Packard Company; Palo Alto, CA, USA) to allow densitometric analysis using the software program ImageJ (http://rsb.info.nih.gov/ij).
Standard curves obtained by the six standard mixtures, which were applied together with the specimens on each plate as described above, were used for the quantification of the respective lipid fractions.
Statistical analysis
Data-distribution was determined by Kolmogorov-Smirnov test and visual assessment of Q-Q plots. Since part of the data did not show standard normal distribution, all data were calculated with non-parametrical tests.
Comparisons of the different body sites were done by Friedman test. In case of a significant result, pairwise comparisons were performed by Wilcoxon signed rank test. Due to the fact that the application of an alpha-adjustment did not result in any
significant differences, which was inappropriate regarding the results of the global test, its usage was waived.
For the comparison between atopic and normal dogs, Wilcoxon two sample test was used. The epidermal lipids of the lesional area of the atopic dogs (mostly axillary) were compared to the epidermal lipids of the axillary region of the dead dogs with normal skin, whereas the epidermal lipids of the non-lesional area (thoracical) were compared to the epidermal lipids of the lateral thorax of the normal dogs.
Comparisons of lesional and non-lesional areas of the atopic dogs were done by Wilcoxon signed rank test.
All analyses were performed using SAS 9.2 (SAS Institute Inc.; Cary, NC, USA).
Significance level was set at P < 0.05.
5.4 Results
Comparison of different body sites in normal skin
No significant differences were detectable for total lipid content between the eight tested body sites (Table 1). In contrast, significant differences of the absolute amounts of CER were found among six of the eight body sites tested (P < 0.01, Friedman test) with higher amounts metacarpal, axillary and thoracical and the lower amounts at the pinna, the flew and the caudal back, in each case with significant differences compared to most other body sites (Table 1). The ceramide classes also significantly differed in content between the body sites except for CER[NP] (Table 1).
Significant differences in ceramide classes occurred between canine AD predilection sites, such as pinnae and metacarpus, as well as between non-predilection sites, such as the lateral thorax and the caudal back. No significant differences in ceramide classes were found between the palmar metacarpus, the axillary region and the lateral thorax or between the axillary and inguinal region, for example (Table 1). Most differences were found for CER[AS+NH] (P < 0,001) with significantly lower levels at the pinna, the flew and the caudal back compared to most other sites. The levels of CER[EOS] were significantly lower at the pinna and the flew when compared to the
metacarpus, the axillary and the inguinal region, whereas the levels of CER[EOP]
were significantly lower at the caudal back than at most other body sites (Table 1).
Among the remaining lipid fractions, no significant differences were found between the body sites in the concentrations of glucosylceramides and free fatty acids. The absolute values of phospholipids, phosphatidylethanolamin, cholesterol sulphate, cholesterol and cholesteryl esters significantly differed between the tested body sites (P < 0.01, Friedman test) as well as the amounts of triglycerides did (P < 0.0001, Friedman test; Table 1). The pinna contained significantly lower amounts of cholesterol sulphate, cholesterol and cholesteryl esters but significantly higher amounts of triglycerides than most other body sites. For the flew region, significantly higher concentrations of phospholipids including phosphatidylethanolamin and cholesterylesters were found compared to most other sites. The skin at the palmar metacarpus was characterised by significantly lower amounts of phospholipids including phosphatidylethanolamin and significantly higher amounts of cholesterol sulphate and cholesterol compared to most other sites. Axillary and thoracic skin contained significantly higher concentrations of cholesterol and phosphatidylethanolamin than most other body sites, respectively. We found significantly lower amounts of cholesterol sulphate and triglycerides at both sites abdominal and inguinal compared to most other sites, whereas the inguinal region also contained significantly lower amounts of cholesteryl esters than most other body regions. At the caudal back, the concentrations of phosphatidylethanolamin and cholesteryl esters were significantly higher compared to most other body sites.
The ratio of ceramides to cholesterol was significantly lower at the caudal back compared to the inguinal region and the pinnae (data not shown).
Differences between atopic and normal dogs
The total lipid amount was significantly lower in both lesional and non-lesional areas of atopic dogs compared to normal canine skin (Figure S1a). In a similar manner, the contents of most lipid fractions (except for free fatty acids, triglycerides and cholesteryl esters) were significantly reduced in atopic dogs. Samples of the lesional regions contained significantly lower amounts of all lipid fractions compared to the
samples of the non-lesional regions, except for phospholipids and free fatty acids (Figure S1a). In contrast, we only rarely found significant differences in the percentage lipid composition between atopic and control dogs (Figure S1b). No differences were detectable between the ratios of ceramides to cholesterol of atopic and normal skin (data not shown).
The skin of atopic dogs contained significantly lower amounts of all ceramide classes than normal canine skin (Figure 1a). Additionally, the contents of CER[AH], CER[AP], CER[AS+NH] and CER[NP] were significantly reduced in lesional skin compared to non-lesional skin. The differences between lesional and normal skin were more pronounced than between non-lesional and normal skin (Figure 1a). In contrast, there were just a few differences regarding the ceramide profile (Figure 1b).
A visual comparison of graphical presentation revealed no differences in lipid composition between the three subgroups of canine AD divided as defined above (data not shown).
Lipid composition of selected further skin disorders
Total lipid content and the amounts of all lipid fractions and ceramide classes were markedly reduced in the lesional skin of the dog with contact dermatitis compared to its non-lesional skin or to normal skin, whereas only the total ceramides and all ceramide classes were reduced in its non-lesional skin compared to normal skin (Table S1).
In the two dogs with flea bite hypersensitivity, the contents of total ceramides, all ceramide classes and cholesterol were reduced in both lesional and non-lesional skin compared to normal skin, without differences being detected between lesional and non-lesional skin (Table S1).
The lesional skin of the dog with sebaceous adenitis contained lower amounts of total lipids as well as lower amounts of free fatty acids, triglycerides and cholesteryl esters than its non-lesional skin or normal skin, the amount of cholesteryl esters in the lesional skin being even merely one tenth of that in the non-lesional skin of this dog.
Total ceramides and ceramide classes did not show any differences to normal skin (Table S1).
Additionally there were decreased levels of total ceramides, all ceramide classes and cholesterol in both lesional and non-lesional skin of the Landseer with flea bite hypersensitivity eight weeks after the systemic administration of a time-released corticosteroid compared to four weeks after the administration. At the latter time of sampling, the quantities of all lipid fractions were comparable to these of normal skin (Tab. C5, Anhang).
Tolerance of the skin scrub technique
The sampling was viable without sedation and with relatively low restraint of the dogs, excessive struggling did not occur. After the sampling, the skin of the dogs tested in our study was erythematous but no oedema was visible. A few dogs were itching at the sampled areas for one to two days. In some cases the sampled area of non-lesional skin developed a crusty circle that vanished a few days after.
5.5 Discussion
Canine AD is characterised by a classic distribution of lesions (MARSELLA and SAMUELSON 2009). Typical affected body sites in canine AD are the face, pinnae, paws, axillary, inguinal and flexural regions (FAVROT et al. 2010; JAEGER et al.
2010). To determine a possible predisposition of such sites for the development of
2010). To determine a possible predisposition of such sites for the development of