Distribution of bioactive lipid mediators in human skin
Alexandra C. Kendall1, Suzanne M. Pilkington2, Karen A. Massey3, Gary Sassano4,
Lesley E. Rhodes2, Anna Nicolaou1*
1
Manchester Pharmacy School and 2Dermatology Centre, Institute of Inflammation and
Repair, Faculty of Medical and Human Sciences, The University of Manchester, Manchester,
UK, 3School of Pharmacy and Centre for Skin Sciences, School of Life Sciences, University of
Bradford, Bradford, UK, 4Safety and Environmental Assurance Centre, Unilever, Sharnbrook,
MK44 1LQ, UK.
*Corresponding author: Professor Anna Nicolaou, Manchester Pharmacy School, The University
of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, UK. Tel: +44 (0) 161
2752374; Email: [email protected]
The work was carried out in Bradford, West Yorkshire, UK.
Short title: Cutaneous bioactive lipids
Abbreviations:
AA, arachidonic acid; AEA, N-arachidonoyl ethanolamide; 2-AG, 2-arachidonoyl glycerol;
ALAE, N-alpha-linolenoyl ethanolamide; CB, cannabinoid receptor; COX, cyclooxygenase;
CYP, cytochrome P450; DGLA, dihomo-gamma-linolenic acid; DHA, docosahexaenoic acid;
1
DHEA, N-docosahexaenoyl ethanolamide; EPA, eicosapentaenoic acid; EPEA: Neicosapentaenoyl ethanolamide; GC, gas chromatography; HDHA, hydroxydocosahexaenoic
acid; HEPE, hydroxyeicosapentaenoic acid; HETE, hydroxyeicosatetraenoic acid; HETrE,
hydroxyeicosatrienoic acid; HFA, hydroxy fatty acid; HODE, hydroxyoctadecadienoic acid;
HPLC, high performance liquid chromatography; LA, linoleic acid; LC-MS/MS, liquid
chromatography tandem mass spectrometry; LEA, N-linoleoyl ethanolamide; LOX,
lipoxygenase; LT, leukotriene; NAE, N-acyl ethanolamide; OA, oleic acid; OEA, N-oleoyl
ethanolamide; PA, palmitic acid; PEA, N-palmitoyl ethanolamide; PLA2, phospholipase A2;
PLD, phospholipase D; PPAR, peroxisome proliferator-activated receptor; PUFA,
polyunsaturated fatty acid; Rv, resolvin; SEA, N-stearoyl ethanolamide; S1P, sphingosine-1phosphate; S1P1, sphingosine-1-phosphate receptor 1; SPE, solid phase extraction; TP,
thromboxane receptor; TX, thromboxane.
2
Abstract
Skin produces bioactive lipids that participate in physiological and pathological states, including
homeostasis, induction, propagation and resolution of inflammation. However, comprehension of
the cutaneous lipid complement, and contribution to differing roles of the epidermal and dermal
compartments, remains incomplete. We assessed the profiles of eicosanoids, endocannabinoids,
N-acyl ethanolamides and sphingolipids, in human dermis, epidermis, and suction blister fluid.
We identified 18 prostanoids, 12 hydroxy-fatty acids, 9 endocannabinoids and N-acyl
ethanolamides, 21 non-hydroxylated ceramides and sphingoid bases, several demonstrating
significantly different expression in the tissues assayed. The array of dermal and epidermal fatty
acids were reflected in the lipid mediators produced, while similarities between lipid profiles in
blister fluid and epidermis indicated a primarily epidermal origin of suction blister fluid.
Supplementation with omega-3 fatty acids ex vivo showed that their action is mediated through
perturbation of existing species and formation of other anti-inflammatory lipids. These findings
demonstrate the diversity of lipid mediators involved in maintaining tissue homeostasis in resting
skin, and hint at their contribution to signalling, cross-support and functions of different skin
compartments. Profiling lipid mediators in biopsies and suction blister fluid can support studies
investigating cutaneous inflammatory responses, dietary manipulation, and skin diseases lacking
biomarkers and therapeutic targets.
3
INTRODUCTION
Skin is rich in lipids that not only contribute to formation and maintenance of the epidermal
barrier, but also perform critical roles in membrane structure and the functions of cutaneous
cells. Bioactive lipid mediators are derived from complex membrane lipids and are produced
upon request and in response to environmental and signalling stimuli. They contribute to both
skin physiology and pathology, with evidence of involvement in psoriasis, dermatitis, acne,
wound healing and UV responses (reviewed in (Kendall and Nicolaou, 2013)). Dermal and
epidermal compartments have diverse roles whilst supporting each other with nutrients and
transcellular signals (Edmondson et al, 2003; Yamaguchi et al, 2005; Yu et al, 2011), and
exhibit considerable inter-family cross-talk (Fig.1). Thus the lipid complement of human skin
and its compartments demand definition, to assist understanding of skin biology, and to facilitate
diverse applications including assessment of novel treatment targets in skin disorders.
Of particular relevance because of their wide range of potential activities in the skin are
the eicosanoid, endocannabinoid, N-acyl ethanolamide, and sphingolipid families. The
eicosanoids are oxygenated metabolites of the 20-carbon polyunsaturated fatty acids (PUFA)
arachidonic acid (AA, 20:4n-6), eicosapentaenoic acid (EPA; 20:5n-3) and dihomogamma
linolenic acid (DGLA; 20:3n-6). This family comprises the cyclooxygenase (COX)-derived
prostaglandins (PG), thromboxanes (TX) and prostacyclin (PGI2), the lipoxygenase (LOX)derived leukotrienes (LT), lipoxins, E-series resolvins and hydroxy fatty acids (HFA) including
hydroxyeicosatetraenoic acids (HETE) and hydroxyeicosapentaenoic acids (HEPE), derived
from LOX and cytochrome P450 (CYP)-mediated reactions. Oxygenation of linoleic acid (LA,
18:2n-6) gives rise to octadecanoids such as hydroxyoctadecanoic acids (HODE), while
docosahexaenoic acid (DHA, 22:6n-3) is the precursor of docosanoids such as D-series resolvins
4
and protectins and hydroxydocosahexaenoic acids (HDHA) (reviewed in (Massey and Nicolaou,
2011; Nicolaou, 2013)). A number of PG, HETE, HODE and LT have been identified in human
skin and contribute to keratinocyte proliferation, melanocyte dendricity, photocarcinogenesis,
allergy and inflammation (Honma et al, 2005; Kendall and Nicolaou, 2013; Rhodes et al, 2009;
Satoh et al, 2006; Scott et al, 2004; Ziboh et al, 2000).
The endocannabinoids N-arachidonoyl ethanolamide (anandamide; AEA) and 2arachidonoyl glycerol (2-AG) are also derivatives of AA and function as endogenous lipid
ligands of the two cannabinoid receptors (CB), CB1 and CB2 that are expressed throughout the
skin in keratinocytes, melanocytes, fibroblasts, sebocytes and hair follicles (Dobrosi et al, 2008;
McPartland, 2008; Pucci et al, 2012; Stander et al, 2005). Cell and organ culture systems
indicate endocannabinoids are released by, and alter the functions of, various skin cells (Czifra et
al, 2012; Dobrosi et al., 2008; Pucci et al., 2012; Sugawara et al, 2012; Toth et al, 2011), and
they have been implicated in several cutaneous pathologies (Kupczyk et al, 2009). A range of Nacyl ethanolamides (NAE) can derive from membrane PUFA. Their relevance to cutaneous
biology is illustrated by the activity of N-palmitoyl ethanolamide (PEA), which has been
implicated in suppression of mast cell degranulation and inflammatory cytokine release, and
explored for the treatment of atopic eczema (De Filippis et al, 2011; Eberlein et al, 2008;
Petrosino et al, 2010).
Sphingolipids are amides of sphingoid bases and an array of complex species can be
derived through addition of fatty acids and head groups. Species present in the skin include free
sphingoid bases and hundreds of ceramide species, as well as their phosphorylated versions
(Masukawa et al, 2008; Rabionet et al, 2014; t'Kindt et al, 2012; van Smeden et al, 2011). The
long chain omega-esterified ceramides found in the stratum corneum are considered structural
5
lipids pivotal for the integrity of the epidermal barrier, and recent studies have revealed their
diversity of structure and physiological functions (Breiden and Sandhoff, 2014; Iwai et al, 2012;
Janssens et al, 2012; Masukawa et al., 2008; van Smeden et al., 2011). In contrast, nonhydroxylated ceramides and sphingoid bases are important mediators of immune cell regulation,
with roles in homeostasis and inflammation (Chalfant and Spiegel, 2005; Uchida, 2014), and
require further investigation to elucidate their contributions in the skin.
Dermis and epidermis exhibit distinct cell populations, enzyme profiles and roles, and
consequently lipid mediator expression could differ between the compartments. While epidermis
comprises keratinocytes, Langerhans cells, melanocytes and Merkel cells, dermis contains
fibroblasts, immune cells, hair follicles, sweat glands, blood vessels and sensory nerves, and
provides skin with elasticity and resistance to mechanical stress. Historically, epidermis has been
considered more important in terms of lipid biology because of the lipid-rich nature of the
epidermal barrier. However, dermis also displays considerable activity and provides biochemical
support to the epidermis, with mediator cross-talk between compartments: e.g. the epidermis
exhibits impaired chain elongation of essential fatty acids and depends on the dermis for local
production of long chain PUFA including AA, EPA and DHA (Chapkin et al, 1986), while
dermally-derived 15-HETE impairs epidermal 12-LOX activity and dermally-produced PGE2
stimulates keratinocyte proliferation (Conconi et al, 1996; Kragballe et al, 1986).
Despite increased awareness of the role of lipid mediators in cutaneous physiology and
disease, there is insufficient information on individual lipid mediator species and their variation
through the skin. To address this, we undertook comprehensive analysis of (i) eicosanoids and
related species, (ii) endocannabinoids, (iii) N-acyl ethanolamides, and (iv) non-hydroxylated
ceramides and free sphingoid bases, in human skin tissue and blister fluid, using mass
6
spectrometry-based targeted mediator lipidomics assays. Our aim was to assess the production,
range and variation of cutaneous lipid metabolites, including the contributions of dermal and
epidermal compartments, in humans. This valuable information can provide insights into the
mode of action of bioactive lipids, including the anti-inflammatory omega-3 fatty acids, and
assist understanding of cutaneous biology, with potential for wide application in skin disease and
its treatment.
RESULTS
Eicosanoids in human dermis, epidermis and blister fluid
Eighteen species of prostanoids and 12 HFA were identified and quantified (Fig.2A; fold
changes from dermal to epidermal expression are presented in Supplementary Table S1). The
AA-derived prostaglandins PGE2, PGD2, PGF2 and PGI2 (as its stable metabolite 6-keto PGF1α)
were found in all tissues tested, with prevalence significantly higher in the epidermis (P=0.0003,
P=0.003, P=0.003, P=0.00004 respectively, compared with dermis); TXA2 (measured as its
stable metabolite TXB2) and PGJ2 were found predominantly in the epidermis. The EPA-derived
PGE3, PGD3, PGF3 and DGLA-derived PGE1 and PGD1, although identified in both
compartments were higher in epidermis. Higher epidermal prevalence was also observed for all
keto- and dihydro-keto PG (e.g. 15-keto-PGE2 and 13,14-dihydro-15-keto-PGE2); these derive
from tissue specific catabolic reactions reducing the bioactivity of primary PG (Tai et al, 2002).
A large number of LOX-derived HFA products of DGLA, AA, EPA, LA and DHA were
found in both dermis and epidermis (HETrE, HETE, HEPE, HODE and HDHA, respectively).
Interestingly, the LA-derived 9-HODE and 13-HODE were the predominant species in both
7
compartments, although found at higher levels in the epidermis (P=0.015, P=0.045, respectively)
with 9-HODE at mean concentrations of 326.9 ± 88.2 pg/mg protein in dermis and 2728.5 ±
548.8 pg/mg protein in epidermis, and 13-HODE at 350.6 ± 83.5 pg/mg protein in dermis and
1416.7 ± 201.5 pg/mg protein in epidermis. While 13-HODE is considered a 15-LOX product, 9HODE can be produced by a partially-completed COX-2 reaction (Laneuville et al, 1995), its
levels reflecting the high COX activity observed in both skin compartments. Concentrations of
the 15-LOX products 15-HETrE and 15-HETE were not significantly different between dermis
and epidermis (Fig.2A; P=0.75, P=0.38, respectively). Although two more 15-LOX products
were identified (EPA-derived 15-HEPE and DHA-derived 17-HDHA), these were not found in
all skin samples and their levels were close to the assay detection limit (20 pg and 10 pg on the
column, respectively) (Massey and Nicolaou, 2013). The 5-LOX-derived product of AA, 5HETE, was found at a very low concentration (≤34.7 pg/mg protein) in both epidermis and
dermis. The 12-LOX products 12-HETE, 12-HEPE, 13-HDHA and 14-HDHA were significantly
higher in epidermis (P=0.002, P=0.027, P=0.012, P=0.04 respectively) with 12-HETE being one
of the predominant mediators found in both skin compartments (127.3 ± 10.5 pg/mg protein in
dermis and 570.4 ± 54.3 pg/mg protein in epidermis). Epidermis had higher concentrations of 8and 11-HETE, products of either oxidation or cutaneous CYP isoforms (Rhodes et al., 2009).
The profile and level of eicosanoids found in suction blister fluid was similar to that in
epidermis, though with differences in some species (Fig.2A): 15-deoxy-Δ12,14 PGJ2 was found
only in blister fluid while 13,14 dihydro PGE1, PGJ2, PGD3 and PGF2α, all found in the dermis
and epidermis, appeared to be minor species of blister fluid with concentrations close to the
method detection limit (0.5-5 pg on the column) (Massey and Nicolaou, 2013).
8
Endocannabinoids and N-acyl ethanolamides in human dermis, epidermis and blister fluid
Nine endocannabinoids and NAE species derivatives of palmitic acid (PA; C16:0), alphalinolenic acid (ALA; C18:3n-3), linoleic acid (LA; C18:2n-6), oleic acid (OA; C18:1n-9), stearic
acid (SA; C18:0), EPA, AA and DHA were quantified. Although a larger number of NAE were
detected, only those accurately identified and quantified based on the availability of commercial
standards are reported. Both dermis and epidermis expressed all species measured, mostly with
comparable expression (Fig.2A). Derivatives of AA, EPA and DHA, namely: AEA,
eicosapentaenoyl ethanolamide (EPEA) and docosahexaenoyl ethanolamide (DHEA), were
significantly higher in epidermis than dermis (P=0.016, P=0.004, P=0.004, respectively) (Fig.2A
and Supplementary Table S1). Finally, all these metabolites were also present in blister fluid, in
ratios similar to the epidermis.
Ceramides and sphingoid bases in the dermis, epidermis and blister fluid
A huge diversity of cutaneous ceramides derives from the different sphingoid bases and acyl
chains (Janssens et al., 2012; Mizutani et al, 2009; Rabionet et al., 2014; t'Kindt et al., 2012; van
Smeden et al., 2011). Initial screening for ceramides for which information on specific transition
ions suitable for mass spectrometric analysis was available, was based on the data reported by
Masukawa et al (2008). Analysis of whole skin lipid extracts, indicated the presence of
ceramides with sphingosine, dihydrosphingosine, phytosphingosine, and 6-hydroxy sphingosine
bases, and non-hydroxy, alpha-hydroxy or ester-linked omega hydroxy fatty acids, belonging to
11 of the 15 ceramide families currently identified (Rabionet et al., 2014), namely: CER[NS],
CER[AS], CER[NH], CER[AH], CER[NP], CER[AP], CER[NDS], CER[ADS], CER[EOP],
CER[EOS] and CER[EOH]. Lack of commercially available appropriate synthetic standards did
not permit identification of individual species. We therefore focused on non-hydroxylated
9
ceramides of the CER[NS] family and phosphorylated ceramides, sphingoid bases and
phosphorylated base species that could be analysed using 17-carbon-containing ceramides as
internal standards (Fig.2B and Supplementary Table S1). Further to their role in epidermal
barrier function, the CER[NS] family includes bioactive medium-chain (40-48-carbon) nonhydroxylated ceramides involved in cutaneous inflammation (Uchida, 2014).
Overall, we identified two sphingoid bases, 18-carbon sphingosine (C18 S) and 18carbon dihydrosphingosine (C18 DS), and their phosphorylated forms 18-carbon sphingosine-1phosphate (C18 S1P) and 18-carbon dihydrosphingosine-1-phosphate (C18 DS1P). We also
found 13 CER[NS] ceramides and four phosphorylated ceramides with long-chain non-hydroxy
fatty acids (22-26 carbons) and sphingoid bases with 16-24 carbons. With the exception of
phosphorylated ceramides that were expressed at similar levels, the epidermis showed greater
expression of all these sphingolipids than the dermis. However, high inter-individual variability
meant there was limited statistical significance, with only two ceramide species,
CER[N(26)S(16)] and CER[N(24)S(20)] (P=0.01, P=0.005, respectively), two phosphorylated
bases (C18S1P, C18DS1P, P=0.024 and P=0.018, respectively) and one sphingoid base (C18DS,
P=0.003) being statistically significantly higher in epidermis (Fig.2B). Of the sphingolipid
species found in blister fluid, the majority were in similar proportions to the dermis and
epidermis (Fig.2B). However, some species were at noticeably low levels in the blister fluid,
including C18S and the largest ceramides measured (CER[N(26)S(16)], CER[N(24)S(21)],
CER[N(25)S(22)] and CER[N(24)S(24)]).
10
Contribution of cutaneous fatty acids
The fatty acid precursors of the eicosanoids and NAE reported here comprise 77.1 % and 61.9 %
of fatty acids in dermis and epidermis, respectively (Fig.3D). Other, non-precursor, fatty acids
make up the remainder (Supplementary Table S2). Despite similar proportions of LA in dermis
and epidermis (10.7 % and 9.6 % of total fatty acids, respectively), the epidermis demonstrated
higher levels of the LA-derived 9- and 13-HODE compared to dermis (69.7 % and 50.3 % of
total HFA detected in the dermis and epidermis, respectively) and this was reflected in the
composition of blister fluid (65.2 % of total HFA detected in blister fluid) (Fig.3B). The
eicosanoid precursors AA and EPA contributed in all classes of mediators detected. AA was
found at a higher concentration in epidermis than dermis (2.7 % versus 0.7 % of total,
respectively) and this was directly reflected in the higher abundance of AEA in epidermis and
blister fluid (13.1 % and 11.2 %, respectively) compared with dermis (8.3 %). In all cases AEA
and EPEA levels were lower than DHEA (Fig.3C). Although cutaneous DHA was detected at
very low levels in epidermis and dermis (0.2 % and 0.5 %, respectively) and was only a minor
contributor of epidermal, but not dermal, HFA (Fig.3B), it was detected as DHEA in dermis,
epidermis and blister fluid (11.7 % ,16.5 % and 14.6 %, respectively) (Fig.3C). The high levels
of dermal OA (44.8 % of total fatty acids) were reflected in the higher abundance of OEA (18.3
% of total dermal NAE), while PA, which was the most abundant fatty acid in skin, contributed
as PEA to almost 40% of the NAE detected in dermis, epidermis and blister fluid (Fig.3C).
Manipulation of lipid mediators by omega-3 PUFA supplementation ex vivo
Exogenous provision of EPA did not have a statistically significant effect on the levels of COXderived PGE2 and LOX-derived 12-HETE in dermis or epidermis, although it induced the
formation of PGE3 and 12-HEPE, two less inflammatory eicosanoids that can attenuate PGE211
and 12-HETE mediated activities (Fig. 4A-D). Conversely, DHA appeared to inhibit PGE2 with
concomitant stimulation of 12-HETE production, suggesting the diversion of cutaneous AA from
COX to LOX-mediated metabolism. Both EPA and DHA stimulated the production of AEA,
EPEA and DHEA in epidermis and dermis, indicating that their anti-inflammatory activities may
be mediated, at least in part, via the endocannabinoid system (Fig.4E-F). Finally, the formation
of 18-HEPE, 17-HDHA and 14-HDHA post-supplementation, shows that human skin can
produce the biochemical precursors of the anti-inflammatory and protective lipids resolvins (RvE
and RvD) and maresins (Fig.4C-D) (Serhan, 2014).
DISCUSSION
Our findings demonstrate the ability of both dermis and epidermis to produce an array of
bioactive lipids that may act locally or be transported and contribute to cross-talk between these
skin compartments. It is noteworthy that despite differences in their physiology, dermis and
epidermis produce the same mediators, although significant differences in the levels and ratios of
certain species highlight their varying requirements.
Higher production of AA, DGLA and EPA-derived epidermal prostanoids can be
attributed to increased expression and/or activity of COX isoforms and prostanoid synthases, and
agrees with a higher epidermal concentration of AA (Fig.3) as well as reports of higher
expression of the constitutive COX-1 in the epidermis (Kragballe et al., 1986). The presence of
deactivated prostanoids such as 15-keto-PGE2 and 13,14-dihydro-15-keto PGE2, shows that the
levels of potent proinflammatory PG such as PGE2 are under active control, a step crucial for
successful resolution of inflammation. Although the prostanoids found in blister fluid matched
12
the epidermal profile (Fig.2A), some minor epidermal species were not detected suggesting that
skin biopsy samples are more appropriate when investigating low abundance lipid mediators.
Interestingly, PGJ2 was found in the epidermis, but not blister fluid, whilst its metabolite 15deoxy-∆12,14 PGJ2, was found at high levels in blister fluid but not in the epidermis or dermis.
Since 15-deoxy-∆12,14 PGJ2 is believed to act as a PPARγ agonist involved in anti-inflammatory
proresolving signalling, its presence in blister fluid could potentially be linked to a cutaneous
healing response to the trauma involved in raising suction blisters (Surh et al, 2011).
The epidermal expression of HFA was also higher, with LA- and AA-derived species
being the predominant mediators detected (Fig.2A). High production of 12-HETE suggests
increased expression/activity of 12-LOX and agrees with reports of three different isozymes
highly expressed in the epidermis (Boeglin et al, 1998; Muller et al, 2002; Takahashi et al,
1993). 12-HETE has proinflammatory and neutrophil chemotactic properties indicating the
potential involvement of 12-LOX in inflammatory skin disease (e.g. psoriasis) (Baer et al, 1995;
Fogh et al, 1993). While the LA-derived 15-LOX product 13-HODE was found at higher levels
in the epidermis, the AA-derived 15-HETE and DGLA-derived 15-HETrE were present at
similar levels in both compartments, albeit at lower concentrations, in accordance with the much
lower levels of their precursor fatty acids (Fig.3). 15-LOX products have anti-inflammatory
activities and may have a specific role in the cross-talk of dermis and epidermis, as demonstrated
by the regulation exhibited by dermal 15-HETE on epidermal 12-LOX activity (Kragballe et al.,
1986; Yoo et al, 2008). Additionally, hair follicles, found in the dermis, are known to express
high levels of a 15S-LOX (Brash et al, 1997) that preferentially metabolises AA over LA.
Although the abdominal skin used in this study was not densely populated with hair, the presence
of hair follicles in the dermis could possibly explain why 15-HETE was present at comparable
13
levels in the dermis and epidermis but 13-HODE was not. Finally, the 5-LOX product 5-HETE
was found at very low levels in both skin compartments, suggesting a low cutaneous expression
of 5-LOX. This supports current understanding that 5-LOX-derived leukotrienes, found in
inflammatory skin disease such as psoriasis, are derived from infiltrating immune cells rather
than resident cells (Ford-Hutchinson, 1993; Sadik et al, 2014).
The significantly higher epidermal levels of the endocannabinoid AEA, and the NAE
species EPEA and DHEA, are of interest, indicating their potential roles in homeostatic
processes in resting skin. AEA has been reported to regulate keratinocyte proliferation, sebum
production and melanogenesis (Conconi et al., 1996; Dobrosi et al., 2008; Pucci et al., 2012). It
is noteworthy that although all HFA derivatives of DHA contributed less than 2% of total class
species, DHEA was one of the main NAE detected (11–16% of total; Fig.3C). DHEA is an
endocannabinoid-like molecule exhibiting weak affinities for cannabinoid receptors (Kim and
Spector,
2013).
EPEA
and
DHEA
have
anti-inflammatory
properties,
reducing
lipopolysaccharide (LPS)-induced interleukin (IL)-6 and monocyte chemotactic protein (MCP)-1
production in adipocytes, and LPS-induced ˙NO production in macrophages (Satoh et al., 2006;
Uchida, 2014), and may operate similar anti-inflammatory effects in the skin. The saturated NAE
species, PEA and OAE, derivatives of palmitic (PA) and oleic (OA) acid, respectively, were
found at equally high level in both dermis and epidermis. PEA is known to suppress cutaneous
mast cell activity, and may confer anti-inflammatory activity (Mizutani et al., 2009; Wang and
Ueda, 2009). OEA is known to affect food intake and the sleep-wake cycle, but its role in skin is
unknown (Honma et al., 2005). Supplementation with exogenous EPA and DHA showed that
their cutaneous metabolism alters the profile of dermal and epidermal lipids, stimulating
production of anti-inflammatory species including COX- and LOX-derived products, and NAE
14
(Fig.4). This provides information on how skin can respond to nutritional interventions and
suggests that n-3PUFA can also create a protective and pro-resolving environment, as supported
by the identification of biochemical precursors for protectins and resolvins in the skin.
Ceramides are components of the stratum corneum involved in epidermal barrier
function, particularly the omega-esterified species with long acyl chains, with more than 300
species identified to date (Behne et al, 2000; Iwai et al., 2012; Rabionet et al., 2014; van Smeden
et al., 2011). In this study, we focus our attention particularly on the family of non-hydroxylated
medium-chain ceramides that can be found in cellular membranes, and notably, we demonstrate
the presence of 21 species in epidermis as well as in dermis and blister fluid. This suggests their
wider involvement in epidermal and dermal function, where they may mediate activities
including regulation of apoptosis and inflammation (Maceyka and Spiegel, 2014; Stiban et al,
2008). Blister fluid showed lower expression of the longest ceramides measured, possibly
indicating impaired release of these species during the blister formation. Further work is
indicated to elucidate the prevalence of these non-hydroxylated species in epidermal and dermal
cells, and explain the intriguing finding of their comparable abundance in both skin
compartments.
Phosphorylated ceramides were also present in dermis and epidermis at equally high
concentrations, suggesting that they originate from cellular membranes and are not restricted to
the differentiated epidermis. This is indicative of an active signalling role in both compartments
that may include the ceramide-1-phosphate and sphingosine-1-phosphate mediated activation of
cutaneous PLA2 (Chalfant and Spiegel, 2005). Ceramide-1-phosphate has also been recently
reported to be involved in wound healing through acting in concert with eicosanoid production
(Wijesinghe et al, 2014), and sphingosine-1-phosphate is a modulator of cutaneous immunity
15
(Herzinger et al, 2007). In this study we focused on the more common 18-carbon sphingosine
and dihydrosphingosine bases, and found that the free sphingoid base C18DS, known to promote
keratinocyte differentiation (Paragh et al, 2008), demonstrated increased expression in the
epidermis (Fig.2B). Future investigations could include analysis of the less abundant
sphingosines and dihydrosphingosines with other carbon chain-lengths, as well as the
phytosphingosine and 6-hydroxy sphingosine bases.
A limitation of the present study is the method used to separate dermis from epidermis,
i.e. physical separation. The commonly-used methods for their separation along the dermoepidermal junction rely on long incubation with salts or enzymes, or rapid transition from high to
low temperatures, to degrade the basement membrane, and are inappropriate for this study’s
purposes as they can induce oxidation or thermal degradation of the lipid mediators (Oakford et
al, 2011; Zhang et al, 2003). Although physical separation of the dermis and epidermis resulted
in slight dermal contamination of the epidermis, care was taken that the less cellular dermis
would remain completely free of any epidermal component. Following analysis, it was found
that differences in lipid expression between the two compartments were reliably identified.
Overall, this study shows that dermis produces a wide range of lipid mediators and this
production may be important biochemical support for the epidermis, acting as an intermediary
between the epidermal interface with the external environment and the subcutaneous circulatory
and lymphatic systems. To our knowledge, this is the most complete analysis of bioactive lipid
mediators in human skin to be reported. Profiles of lipid mediators can support the
characterisation of COX, LOX and CYP isoforms, whose activity and pattern of expression
depends on post translational modifications and may contribute to cutaneous disorders (Aguiar et
al, 2005; Kuhn and O'Donnell, 2006; Mbonye and Song, 2009). We have also shown an ability
16
to manipulate the balance of bioactive lipids in favour of the protective omega-3-derived species,
with implications for their mode of action in skin health and disease. Finally, direct comparison
of lipid mediators found in suction blister fluid, a widely used technique for sampling cutaneous
intercellular fluid, with skin lipid profiles, provides confirmatory information on study design to
explore the potency and importance of lipid mediators in skin biology.
MATERIALS AND METHODS
Tissue Samples
Biobank skin: Skin samples were obtained from a biobank (Ethical Tissue, University of
Bradford, Bradford, UK) with full ethical approval (Leeds East Research Ethics Committee
reference 07/H1306/98+5). Skin was provided by 8 healthy donors (6 female, 2 male; 30-60
years; white Caucasian), undergoing elective abdominoplastic surgery, delivered to the biobank
within 1 hour of the operation. The tissue was washed and the adipose layer removed. Punch
biopsies (3 mm) were cut, snap frozen and stored at -80ºC, or were cultured ex vivo. Ex vivo
culture was performed in DMEM containing 100 U/ml penicillin, 100 µg/ml streptomycin and
1.4 mM Ca2+(Promocell, Heidelberg, Germany), with or without the addition of EPA or DHA
(50 µM) (Sigma Aldrich, Poole, UK), for 3 days (Tavakkol et al, 1999). For lipidomics assays
three punch biopsies per donor were used. Prior to analysis, skin was divided into dermis and
epidermis (on ice, by scalpel, with the aid of visual inspection at 40X magnification). Whilst
epidermal samples demonstrated minor contamination with dermal tissue, care was taken that
dermal tissue was not contaminated with epidermal tissue.
17
Skin suction blister fluid and punch biopsies: Healthy human volunteers (n=8; all females; 28-56
years; white Caucasian) were recruited by the Photobiology Unit, Dermatology Centre, Salford
Royal Hospital, Manchester, UK. Ethical approval was obtained from the North Manchester
Research Ethics Committee (reference 08/H1006/79). Written informed consent was obtained
from participants and the study adhered to Declaration of Helsinki principles. Volunteers
provided 5mm skin punch biopsies and suction blisters from buttock skin for these analyses.
Punch biopsy dermis and epidermis were separated as described above, and samples used for
total fatty acid analysis. Suction blistering was performed using suction cups with a 1 cm central
aperture as described previously (Rhodes et al., 2009). Skin blister fluid was aspirated with a 23gauge needle, snap-frozen in liquid nitrogen and stored at -80ºC.
Eicosanoid extraction and analysis
Skin samples (30-60 mg) and blister fluid samples (40-90 µl) were extracted using ice-cold 15%
(v/v) methanol solution. 12-HETE-d8 and PGB2-d4 (Cayman Chemicals, Ann Arbor, MI, USA)
were used as internal standards, as published before (Masoodi et al, 2008; Masoodi and
Nicolaou, 2006; Massey and Nicolaou, 2013).+The extracts were semi-purified by SPE
cartridges (C18-E; Phenomenex, Macclesfield, UK); the assay recovery was estimated at 9698%. LC/ESI-MS/MS was performed on an HPLC pump (Waters Alliance 2695) coupled to an
electrospray ionisation triple quadrupole mass spectrometer (Quattro Ultima, Waters, Elstree,
Hertfordshire, UK). Results are expressed as pg/mg protein (skin) or pg/ml (blister fluid),
estimated using calibration lines covering the range 0.02-20 ng/ml. Detailed description of the
experimental protocol is provided as Supplementary Information (SM1).
18
Endocannabinoid and NAE extraction and analysis
Skin samples (30-60 mg) and blister fluid samples (40-90 µl) were extracted using ice-cold 2:1
(v/v) chloroform/methanol. Arachidonoyl EA-d8 and 2-arachidonoyl glycerol-d8 (Cayman
Chemicals) were used as internal standards. Lipid extracts were analysed by LC/ESI-MS/MS.
Recovery was estimated at 91%. Results are expressed as pg/mg protein (skin) or pg/ml (blister
fluid), estimated using calibration lines covering the range 0.02-50 ng/ml. Detailed description of
the experimental protocol is provided as Supplementary Information (SM2).
Sphingolipid extraction and analysis
Skin samples (30-60 mg) and blister fluid samples (40-90 µl) were extracted using ice-cold
isopropanol:water:ethyl acetate (30:10:60; v/v/v). The following internal standards were used:
C17 S (used for C18 S), C17 DS (used for C18 DS), C17 S1P (used for C18 S1P), C17 DS1P
(used for C18 DS1P), d18:1/12:0 C1P (used for all C1P species) and C25 Cer (used for all
CER[NS] species) (Ceramide/Sphingoid Internal Standard Mixture I, Avanti Polar Lipids,
Alabaster, Alabama, USA). The resulting lipid extracts were analysed by LC/ESI-MS/MS
(Bielawski et al, 2006; Kelly et al, 2011). The assay recovery was estimated at 75 -79%. Results
are expressed as pmol/g protein (skin) or pmol/l (blister fluid), relative to the appropriate
internal standard (covering the range 0.005-8.0 nmol/ml). Detailed description of the
experimental protocol is provided as Supplementary Information (SM3).
Fatty acid analysis
Fatty acids were analysed by gas chromatography (GC) in dermis and epidermis as previously
described (Pilkington et al, 2014). Results are expressed as percentage of total fatty acids.
19
Detailed description of the experimental protocol is provided as Supplementary Information
(SM4).
Protein content
During lipid extractions, protein pellets were retained for analysis of protein content using a
standard Bradford protein assay kit (Bio-Rad, Hemel Hempstead, UK) (Bradford, 1976).
Proteins were extracted using 1 M NaOH and analysed within the linear range of the assay to
ensure accuracy.
Statistical analysis
Statistical analyses of lipid mediator expression were performed using repeated measures
ANOVAs with Greenhouse-Geisser corrections and Bonferroni post-hoc tests. Analyses were
conducted using SPSS 20 software and P<0.05 was considered significant.
CONFLICT OF INTEREST
The authors state no conflict of interest.
ACKNOWLEDGEMENTS
This work was funded by Unilever as part of its ongoing program developing novel non-animal
approaches for assessing consumer safety. We thank Andrew Healey, Analytical Centre,
University of Bradford and Wayne Burrill, Ethical Tissue, University of Bradford, for excellent
technical support.
20
REFERENCES
Aguiar M, Masse R, Gibbs BF (2005). Regulation of cytochrome P450 by posttranslational
modification. Drug Metab Rev 37: 379-404.
Baer AN, Klaus MV, Green FA (1995). Epidermal fatty acid oxygenases are activated in nonpsoriatic dermatoses. J Invest Dermatol 104: 251-5.
Behne M, Uchida Y, Seki T, et al (2000). Omega-hydroxyceramides are required for corneocyte
lipid envelope (CLE) formation and normal epidermal permeability barrier function. The Journal
of investigative dermatology 114: 185-92.
Bielawski J, Szulc ZM, Hannun YA, et al (2006). Simultaneous quantitative analysis of
bioactive sphingolipids by high-performance liquid chromatography-tandem mass spectrometry.
Methods 39: 82-91.
Boeglin WE, Kim RB, Brash AR (1998). A 12R-lipoxygenase in human skin: mechanistic
evidence, molecular cloning, and expression. Proc Natl Acad Sci U S A 95: 6744-9.
Bradford MM (1976). A rapid and sensitive method for the quantitation of microgram quantities
of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248-54.
Brash AR, Boeglin WE, Chang MS (1997). Discovery of a second 15S-lipoxygenase in humans.
Proc Natl Acad Sci U S A 94: 6148-52.
Breiden B, Sandhoff K (2014). The role of sphingolipid metabolism in cutaneous permeability
barrier formation. Biochim Biophys Acta 1841: 441-52.
Chalfant CE, Spiegel S (2005). Sphingosine 1-phosphate and ceramide 1-phosphate: expanding
roles in cell signaling. J Cell Sci 118: 4605-12.
Chapkin RS, Ziboh VA, Marcelo CL, et al (1986). Metabolism of essential fatty acids by human
epidermal enzyme preparations: evidence of chain elongation. J Lipid Res 27: 945-54.
Conconi MT, Bruno P, Bonali A, et al (1996). Relationship between the proliferation of
keratinocytes cultured in vitro and prostaglandin E2. Ann Anat 178: 229-36.
Czifra G, Szollosi AG, Toth BI, et al (2012). Endocannabinoids regulate growth and survival of
human eccrine sweat gland-derived epithelial cells. J Invest Dermatol 132: 1967-76.
De Filippis D, Luongo L, Cipriano M, et al (2011). Palmitoylethanolamide reduces granulomainduced hyperalgesia by modulation of mast cell activation in rats. Mol Pain 7: 3.
Dobrosi N, Toth BI, Nagy G, et al (2008). Endocannabinoids enhance lipid synthesis and
apoptosis of human sebocytes via cannabinoid receptor-2-mediated signaling. FASEB J 22:
3685-95.
21
Eberlein B, Eicke C, Reinhardt HW, et al (2008). Adjuvant treatment of atopic eczema:
assessment of an emollient containing N-palmitoylethanolamine (ATOPA study). J Eur Acad
Dermatol Venereol 22: 73-82.
Edmondson SR, Thumiger SP, Werther GA, et al (2003). Epidermal homeostasis: the role of the
growth hormone and insulin-like growth factor systems. Endocr Rev 24: 737-64.
Fogh K, Iversen L, Herlin T, et al (1993). Modulation of eicosanoid formation by lesional skin of
psoriasis: an ex vivo skin model. Acta Derm Venereol 73: 191-3.
Ford-Hutchinson AW (1993). 5-Lipoxygenase activation in psoriasis: a dead issue? Skin
Pharmacol 6: 292-7.
Herzinger T, Kleuser B, Schafer-Korting M, et al (2007). Sphingosine-1-phosphate signaling and
the skin. Am J Clin Dermatol 8: 329-36.
Honma Y, Arai I, Hashimoto Y, et al (2005). Prostaglandin D2 and prostaglandin E2 accelerate
the recovery of cutaneous barrier disruption induced by mechanical scratching in mice. Eur J
Pharmacol 518: 56-62.
Iwai I, Han H, den Hollander L, et al (2012). The human skin barrier is organized as stacked
bilayers of fully extended ceramides with cholesterol molecules associated with the ceramide
sphingoid moiety. The Journal of investigative dermatology 132: 2215-25.
Janssens M, van Smeden J, Gooris GS, et al (2012). Increase in short-chain ceramides correlates
with an altered lipid organization and decreased barrier function in atopic eczema patients. J
Lipid Res 53: 2755-66.
Kelly L, Grehan B, Chiesa AD, et al (2011). The polyunsaturated fatty acids, EPA and DPA
exert a protective effect in the hippocampus of the aged rat. Neurobiol Aging 32: 2318 e1-15.
Kendall AC, Nicolaou A (2013). Bioactive lipid mediators in skin inflammation and immunity.
Prog Lipid Res 52: 141-64.
Kim HY, Spector AA (2013). Synaptamide, endocannabinoid-like derivative of docosahexaenoic
acid with cannabinoid-independent function. Prostaglandins Leukot Essent Fatty Acids 88: 1215.
Kragballe K, Pinnamaneni G, Desjarlais L, et al (1986). Dermis-derived 15-hydroxyeicosatetraenoic acid inhibits epidermal 12-lipoxygenase activity. J Invest Dermatol 87: 494-8.
Kuhn H, O'Donnell VB (2006). Inflammation and immune regulation by 12/15-lipoxygenases.
Prog Lipid Res 45: 334-56.
22
Kupczyk P, Reich A, Szepietowski JC (2009). Cannabinoid system in the skin – a possible target
for future therapies in dermatology. Exp Dermatol 18: 669-79.
Laneuville O, Breuer DK, Xu N, et al (1995). Fatty acid substrate specificities of human
prostaglandin-endoperoxide H synthase-1 and -2. Formation of 12-hydroxy-(9Z, 13E/Z, 15Z)octadecatrienoic acids from alpha-linolenic acid. J Biol Chem 270: 19330-6.
Maceyka M, Spiegel S (2014). Sphingolipid metabolites in inflammatory disease. Nature 510:
58-67.
Masoodi M, Mir AA, Petasis NA, et al (2008). Simultaneous lipidomic analysis of three families
of bioactive lipid mediators leukotrienes, resolvins, protectins and related hydroxy-fatty acids by
liquid chromatography/electrospray ionisation tandem mass spectrometry. Rapid Commun Mass
Spectrom 22: 75-83.
Masoodi M, Nicolaou A (2006). Lipidomic analysis of twenty-seven prostanoids and
isoprostanes by liquid chromatography/electrospray tandem mass spectrometry. Rapid Commun
Mass Spectrom 20: 3023-9.
Massey KA, Nicolaou A (2011). Lipidomics of polyunsaturated-fatty-acid-derived oxygenated
metabolites. Biochem Soc Trans 39: 1240-6.
Massey KA, Nicolaou A (2013). Lipidomics of oxidized polyunsaturated fatty acids. Free Radic
Biol Med 59: 45-55.
Masukawa Y, Narita H, Shimizu E, et al (2008). Characterization of overall ceramide species in
human stratum corneum. J Lipid Res 49: 1466-76.
Mbonye UR, Song I (2009). Posttranscriptional and posttranslational determinants of
cyclooxygenase expression. BMB Rep 42: 552-60.
McPartland JM (2008). Expression of the endocannabinoid system in fibroblasts and myofascial
tissues. J Bodyw Mov Ther 12: 169-82.
Mizutani Y, Mitsutake S, Tsuji K, et al (2009). Ceramide biosynthesis in keratinocyte and its
role in skin function. Biochimie 91: 784-90.
Muller K, Siebert M, Heidt M, et al (2002). Modulation of epidermal tumor development caused
by targeted overexpression of epidermis-type 12S-lipoxygenase. Cancer Res 62: 4610-6.
Nicolaou A (2013). Eicosanoids in skin inflammation. Prostaglandins Leukot Essent Fatty Acids
88: 131-8.
Oakford ME, Dixon SV, August S, et al (2011). Migration of immunocytes across the basement
membrane in skin: The role of basement membrane pores. J Invest Dermatol 131: 1950-3.
23
Paragh G, Schling P, Ugocsai P, et al (2008). Novel sphingolipid derivatives promote
keratinocyte differentiation. Exp Dermatol 17: 1004-16.
Petrosino S, Cristino L, Karsak M, et al (2010). Protective role of palmitoylethanolamide in
contact allergic dermatitis. Allergy 65: 698-711.
Pilkington SM, Rhodes LE, Al-Aasswad NM, et al (2014). Impact of EPA ingestion on COXand LOX-mediated eicosanoid synthesis in skin with and without a pro-inflammatory UVR
challenge  report of a randomised controlled study in humans. Mol Nutr Food Res 58: 580-90.
Pucci M, Pasquariello N, Battista N, et al (2012). Endocannabinoids stimulate human
melanogenesis via type-1 cannabinoid receptor. J Biol Chem 287: 15466-78.
Rabionet M, Gorgas K, Sandhoff R (2014). Ceramide synthesis in the epidermis. Biochim
Biophys Acta 1841: 422-34.
Rhodes LE, Gledhill K, Masoodi M, et al (2009). The sunburn response in human skin is
characterized by sequential eicosanoid profiles that may mediate its early and late phases.
FASEB J 23: 3947-56.
Sadik CD, Sezin T, Kim ND (2014). Leukotrienes orchestrating allergic skin inflammation. Exp
Dermatol 22: 705-9.
Satoh T, Moroi R, Aritake K, et al (2006). Prostaglandin D2 plays an essential role in chronic
allergic inflammation of the skin via CRTH2 receptor. J Immunol 177: 2621-9.
Scott G, Leopardi S, Printup S, et al (2004). Proteinase-activated receptor-2 stimulates
prostaglandin production in keratinocytes: analysis of prostaglandin receptors on human
melanocytes and effects of PGE2 and PGF2 on melanocyte dendricity. J Invest Dermatol 122:
1214-24.
Serhan CN (2014). Pro-resolving lipid mediators are leads for resolution physiology. Nature 510:
92-101.
Stander S, Schmelz M, Metze D, et al (2005). Distribution of cannabinoid receptor 1 (CB1) and
2 (CB2) on sensory nerve fibers and adnexal structures in human skin. J Dermatol Sci 38: 17788.
Stiban J, Caputo L, Colombini M (2008). Ceramide synthesis in the endoplasmic reticulum can
permeabilize mitochondria to proapoptotic proteins. J Lipid Res 49: 625-34.
Sugawara K, Biro T, Tsuruta D, et al (2012). Endocannabinoids limit excessive mast cell
maturation and activation in human skin. J Allergy Clin Immunol 129: 726-38 e8.
24
Surh YJ, Na HK, Park JM, et al (2011). 15-Deoxy-Delta(1)(2),(1)(4)-prostaglandin J(2), an
electrophilic lipid mediator of anti-inflammatory and pro-resolving signaling. Biochem
Pharmacol 82: 1335-51.
t'Kindt R, Jorge L, Dumont E, et al (2012). Profiling and characterizing skin ceramides using
reversed-phase liquid chromatography-quadrupole time-of-flight mass spectrometry. Anal Chem
84: 403-11.
Tai HH, Ensor CM, Tong M, et al (2002). Prostaglandin catabolizing enzymes. Prostaglandins
Other Lipid Mediat 68-69: 483-93.
Takahashi Y, Reddy GR, Ueda N, et al (1993). Arachidonate 12-lipoxygenase of platelet-type in
human epidermal cells. J Biol Chem 268: 16443-8.
Tavakkol A, Varani J, Elder JT, et al (1999). Maintenance of human skin in organ culture: role
for insulin-like growth factor-1 receptor and epidermal growth factor receptor. Arch Dermatol
Res 291: 643-51.
Toth BI, Dobrosi N, Dajnoki A, et al (2011). Endocannabinoids modulate human epidermal
keratinocyte proliferation and survival via the sequential engagement of cannabinoid receptor-1
and transient receptor potential vanilloid-1. J Invest Dermatol 131: 1095-104.
Uchida Y (2014). Ceramide signaling in mammalian epidermis. Biochim Biophys Acta 1841:
453-62.
van Smeden J, Hoppel L, van der Heijden R, et al (2011). LC/MS analysis of stratum corneum
lipids: ceramide profiling and discovery. J Lipid Res 52: 1211-21.
Wang J, Ueda N (2009). Biology of endocannabinoid synthesis system. Prostaglandins Other
Lipid Mediat 89: 112-9.
Wijesinghe DS, Brentnall M, Mietla JA, et al (2014). Ceramide kinase is required for a normal
eicosanoid response and the subsequent orderly migration of fibroblasts. J Lipid Res 55: 1298309.
Yamaguchi Y, Hearing VJ, Itami S, et al (2005). Mesenchymal-epithelial interactions in the
skin: aiming for site-specific tissue regeneration. J Dermatol Sci 40: 1-9.
Yoo H, Jeon B, Jeon MS, et al (2008). Reciprocal regulation of 12- and 15-lipoxygenases by
UV-irradiation in human keratinocytes. FEBS Lett 582: 3249-53.
Yu C, Fedoric B, Anderson PH, et al (2011). Vitamin D(3) signalling to mast cells: A new
regulatory axis. The international journal of biochemistry & cell biology 43: 41-6.
Zhang X-J, Chinkes DL, Wolfe RR (2003). Measurement of protein metabolism in epidermis
and dermis. Am J Physiol Endocrinol Metab 284: E1191-E201.
25
Ziboh VA, Miller CC, Cho Y (2000). Metabolism of polyunsaturated fatty acids by skin
epidermal enzymes: generation of antiinflammatory and antiproliferative metabolites. Am J Clin
Nutr 71: 361s-6s.
26
Figure legends
Figure 1. Schematic outline of bioactive lipid mediator production and related signalling
events. AA, arachidonic acid; AEA, arachidonoyl ethanolamide; AP-1, activator protein 1; BLT,
leukotriene B4 receptor; CB, cannabinoid receptor; COX, cyclooxygenase; C1P, ceramide-1phosphate; DHA, docosahexaenoic acid; DP, prostaglandin D2 receptor; EP, prostaglandin E2
receptor; EPA, eicosapentaenoic acid; FA, fatty acid; FA-EA, fatty acid ethanolamide; G2A, G
protein-coupled
receptor
132;
GPR55,
G
protein-coupled
receptor
55;
HDHA,
hydroxydocosahexaenoic acid; HETE, hydroxyeicosatetraenoic acid; IP, prostacyclin receptor;
LOX, lipoxygenase; LT, leukotriene; LX, lipoxin; NF-κB, nuclear factor of kappa-light-chainenhancer in B-cells; PD, protectin; PEA, palmitoyl ethanolamide; PG, prostaglandin; PLA2,
phospholipase A2; PLD, phospholipase D; PPAR, peroxisome proliferator-activated receptor;
PUFA, polyunsaturated fatty acid; Rv, resolvin; S1P, sphingosine-1-phosphate; S1P1,
sphingosine-1-phosphate receptor 1; TP, thromboxane receptor; TRPV, transient receptor
potential vanilloid; TX, thromboxane.
Figure 2. Expression of eicosanoids, endocannabinoids and N-acyl ethanolamides (a), and
ceramides, phosphorylated ceramides and sphingoid bases (b), in human dermis, epidermis
and blister fluid. All lipid mediators were analysed by LC-MS/MS. The arachidonic acidderived PGE2, 12-HETE, AEA and 2-AG, and C18 S and its derivatives C18 S1P,
CER[N(22)S(18)] and d18:1/16:0 C1P are provided as example structures of the different classes
of lipid mediators presented in this figure. Data for prostaglandins (PG), prostacyclin (PGI2)
(measured as the stable derivative 5-keto PGF1), thromboxanes (TX), hydroxy fatty acids,
endocannabinoids and N-acyl ethanolamides (NAE) are expressed as pg/mg tissue protein or
27
pg/ml blister fluid. Data for ceramides, phosphorylated ceramides and sphingoid bases are
expressed as pmol/g protein or pmol/l blister fluid (dermis and epidermis; n=8 donors; labelled
1-8) and pg/ml blister fluid (n=3 donors, labelled 9-11). *P<0.05, **P<0.01 and ***P<0.001
when comparing dermis to epidermis.
C18S: 18-carbon sphingosine (S); C18DS: 18-carbon dihydrosphingosine (DS); C18S1P: 18carbon
sphingosine-1-phosphate;
C18DS1P:
18-carbon
dihydrosphingosine-1-phosphate;
Ceramides derivatives of sphingosine (S) with a non-hydroxy fatty acid (N) are named according
to the number of carbons of the base (e.g. 16, 18 etc) and fatty acid (e.g. 22, 24, etc).
Phosphorylated ceramides are denoted by the base (d18:1 representing sphingosine and d18:0
representing dihydrosphingosine) and the fatty acid (e.g. 16:0 represents palmitic acid).
Figure 3. Contribution of cutaneous fatty acid precursors of lipid mediator populations
found in dermis, epidermis and suction blister fluid. Prostanoids (a), hydroxy fatty acids (b)
and N-acyl ethanolamides (c) (as quantified in Figure 2) are shown as % of total mediators
detected in each tissue, together with the % abundance of their precursor fatty acid (d) in dermis
and epidermis. Data are expressed as a % of total lipid mediators detected in the dermis and
epidermis (n=8 donors), % of total lipid mediators detected in the blister fluid (n=3 donors) or as
% of total fatty acids detected in dermis and epidermis (n=5 donors). AA, arachidonic acid;
ALA, α-linolenic acid; DGLA, dihomo-gamma-linolenic acid; DHA, docosahexaenoic acid;
EPA, eicosapentaenoic acid; LA, linoleic acid; OA, oleic acid; PA, palmitic acid; SA, stearic
acid.
28
Figure 4. Effect of omega-3 PUFA supplementation on the production of cutaneous lipid
mediators ex vivo. Skin was treated for 3 days with EPA or DHA (50 M). PGE2 and PGE3
extracted from dermis (a) and epidermis (b), 12-HETE, 12-HEPE, 18-HEPE, 17-HDHA and 14HDHA extracted from dermis (c) and epidermis (d), AEA, EPEA and DHEA extracted from
dermis (e) and epidermis (f), and analysed by LC-MS/MS. Data are expressed as pg/mg tissue
protein (n=3 biopsies). *P<0.05, **P<0.01 and ***P<0.001 when comparing all data to control.
PG, prostaglandin; HETE, hydroxyeicosatetraenoic acid; HEPE, hydroxyeicosapentaenoic acid;
HDHA, hydroxydocosahexaenoic acid; EPA, eicosapentaenoic acid; DHA, docosahexaenoic
acid; AEA, N-arachidonoyl ethanolamide; EPEA, N-eicosapentaenoyl ethanolamide; DHEA, Ndocosahexaenoyl ethanolamide.
29
Figure 1
30
Figure 2
31
Figure 3
32
Figure 4
33
Distribution of bioactive lipid mediators in human skin
Alexandra C. Kendall1, Suzanne M. Pilkington2, Karen A. Massey3, Gary Sassano4,
Lesley E. Rhodes2, Anna Nicolaou1*
Supplementary Information
SM: Supplementary Methods- Extraction and analysis of lipids
SM1. Eicosanoids
Skin samples (30-60 mg) were homogenised using a blade homogeniser in ice-cold 15% (v/v)
methanol solution; blister fluid samples (40-90 µl) were mixed with ice-cold 15% (v/v) methanol
solution (3 ml per sample). Internal standards (40 ng each of 12-HETE-d8 and PGB2-d4;
Cayman Chemicals, Ann Arbor, MI, USA) were added, and samples incubated on ice for 90 min.
Protein precipitates were removed by centrifugation (1500xg, 4ºC, 10 min), and supernatants
acidified to pH 3 using 0.1M HCl. Solid phase extraction was performed using C18 cartridges
(C18-E; 500 mg, 6 ml; Phenomenex, Macclesfield, UK) pre-conditioned with 20 ml methanol
(HPLC grade) followed by 20 ml H2O under a low vacuum. The acidified sample was added to
the SPE cartridges and allowed to pass through without vacuum. The cartridges were then
washed sequentially with 20 ml each of 15 % (v/v) methanol solution, H2O, and hexane (HPLC
grade). Finally, lipid mediators were eluted into glass tubes using 12 ml methyl formate (reagent
grade solvent – 97% purity). The solvent was evaporated off under a stream of nitrogen and the
lipid extracts were reconstituted in 100 µl 70 % (v/v) ethanol (HPLC grade). LC/ESI-MS/MS
1
was performed on an HPLC pump (Waters Alliance 2695) coupled to an electrospray ionisation
triple quadrupole mass spectrometer (Quattro Ultima, Waters, Elstree, Hertfordshire, UK).
Prostanoids were analysed on a C18 column (Luna, 5 µm, 150 x 2 mm; Phenomenex),
using a gradient of solvent A (acetonitrile:water:glacial acetic acid; 45:55:0.02; v/v/v) and
solvent B (acetonitrile:water:glacial acetic acid; 90:10:0.02; v/v/v), as follows: 0% B (0-8.0 min),
0-50% B (8.0-8.1min), 50% B (8.1-12 min), 50-70 % B (12.0-12.1 min), 70% B (12.1-20.0 min),
70-0% B (20-21 min), 0% B (21-30 min), at a flow rate 0.2 ml/min. Compounds were
fragmented using argon as collision gas. The analytes were monitored in the negative ion mode
by multiple reaction monitoring (MRM; collision energy in parentheses): PGE2 mass to charge
ratio (m/z) 351>271 (17 eV); PGD2 m/z 351>271 (17 eV); 15-keto PGE2 m/z 349>113 (25 eV);
13,14-dihydro-15-keto PGE2 m/z 351>333 (12 eV); PGF2α m/z 353>193 (25 eV); 13,14 dihydro15-keto PGF2α m/z 353>113 (25 eV); PGJ2 m/z 333>271 (15 eV); 15-deoxy-∆12,14 PGJ2 m/z
315>271 (12 eV); TXB2 m/z 369>169 (17 eV); PGE3 m/z 349>269 (15 eV); PGD3 m/z 349>269
(15 eV); PGF3α m/z 351>193 (25 eV); PGE1 m/z 353>317 (15 eV); PGD1 m/z 353>317 (15 eV);
13,14-dihydro-PGE1 m/z 355>337 (15 eV); 13,14-dihydro-15-keto-PGE1 m/z 353>335 (12 eV);
PGF1α m/z 355>311 (23 eV); 6-keto-PGF1α m/z 369>163 (23 eV).
HFA were analysed on a C18 column (Kinetex, 2.6 µm, 100 x 2.1 mm; Phenomenex),
using a gradient of solvent A (acetonitrile:water:glacial acetic acid; 45:55:0.02; v/v/v) and
solvent B (methanol:water:glacial acetic acid; 80:20:0.02; v/v/v) as follows: 30% B (0-1 min),
30-83% B (1-1.1 min), 83% B (1.1-25 min), 83-100% B (25-25.1 min), 100% B (25.1-28 min),
100-30% B (28-28.1 min) and 30% B (28.1-35 min), at a flow rate 0.2 ml/min. Compounds were
fragmented using argon as collision gas. The analytes were monitored in the negative ion mode
by MRM (collision energy in parentheses): 9-HODE m/z 295>171 (25 eV); 13-HODE m/z
2
295>195 (25 eV); 11-HEPE m/z 317>167 (20 eV); 12-HEPE m/z 317>179 (17 eV); 5-HETE m/z
319>115 (20 eV); 8-HETE m/z 319>155 (20 eV); 11-HETE m/z 319>167 (20 eV); 15-HETE m/z
319>175 (18 eV); 12-HETE m/z 319>179 (17 eV); 15-HETrE m/z 321>303 (15 eV); 14-HDHA
m/z 343>161 (18 eV); 13-HDHA m/z 343>193 (18 eV); 18-HEPE m/z 317>215 (18 eV); 17HDHA m/z 343>201 (18 eV). The internal standard 12-HETE-d8 was used to control for
variation between injections.
Calibration lines were constructed using commercially-available standards to cover a
range of 0.02-20 ng/ml, which showed a linear response. All samples were analysed within this
linear range prior to normalisation against protein content. The assay recovery was estimated at
96-98%.
SM2. Endocannabinoids
Skin samples (30-60 mg) were homogenised in ice-cold 2:1 (v/v) chloroform/methanol; blister
fluid samples (40-90 µl) were added to ice-cold 2:1 (v/v) chloroform/methanol (3 ml per
sample). Internal standards (40 ng each of arachidonoyl EA-d8 and 2-arachidonoyl glycerol-d8)
(Cayman Chemicals) were added, and samples were incubated on ice for 90 min. Water was
added (500 µl per sample), and organic extracts separated by centrifugation (1500xg, 4ºC, 5
min). The solvent was dried under a stream of nitrogen and the lipid extracts were reconstituted
in 100 µl ethanol (HPLC grade) and LC/ESI-MS/MS was performed on an HPLC pump (Waters
Alliance 2695) coupled to an electrospray ionisation triple quadrupole mass spectrometer
(Quattro Ultima, Waters).
Analytes were separated on a C18 column (Luna, 5 µm, 150 x 2 mm; Phenomenex) using
a gradient of solvent A (acetonitrile:water:glacial acetic acid; 2:98:0.5; v/v/v) and solvent B
3
(acetonitrile:water:glacial acetic acid; 98:2:0.5; v/v/v) as follows: 30-90 % B (0-40 min), 90 % B
(40-55 min) 90-30 % B (55-56 min) and 30 % B (56-69 min), at a flow rate 0.2 ml/min.
Compounds were fragmented using argon as collision gas. Compounds were monitored in the
positive ion mode by MRM (collision energy in parentheses): PEA m/z 300>62 (13 eV); ALEA
m/z 322>62 (14 eV); LEA m/z 324>62 (15 eV); OEA m/z 324>62 (16 eV); SEA m/z 328>62 (15
eV); EPEA m/z 346>62 (15 eV); AEA m/z 348>62 (15 eV); DHEA m/z 372>62 (15 eV); 2-AG
m/z 379>287 (18 eV). The internal standard AEA-d8 was used to control for variation between
injections. The assay recovery was estimated at 91%.
Calibration lines were constructed using commercially-available standards to cover a
range of 0.02-50 ng/ml, which showed a linear response. All samples were analysed within this
linear range prior to normalisation against protein content.
SM3. Sphingolipids
Skin samples (30-60 mg) were homogenised in ice-cold isopropanol:water:ethyl acetate
(30:10:60;
v/v/v);
blister
fluid
samples
(40-90
µl)
were
mixed
with
ice-cold
isopropanol:water:ethyl acetate (30:10:60; v/v/v; 4 ml per sample). The following internal
standards were used: C17 S (used for C18 S), C17 DS (used for C18 DS), C17 S1P (used for
C18 S1P), C17 DS1P (used for C18 DS1P), d18:1/12:0 C1P (used for all C1P species) and C25
Cer (used for all CER[NS] species) (Ceramide/Sphingoid Internal Standard Mixture I, Avanti
Polar Lipids, Alabaster, Alabama, USA). Internal standards (200 pmol per standard) were added,
and samples incubated on ice for 90 min. Protein precipitates were removed by centrifugation
(1500xg, 4ºC, 10 min) and the clear supernatant was dried under a stream of nitrogen. The
samples were reconstituted in 150 µl methanol containing 0.2 % formic acid and LC/ESI4
MS/MS was performed on an HPLC pump (Waters Alliance 2695) coupled to an electrospray
ionisation triple quadrupole mass spectrometer (Quattro Ultima, Waters).
Ceramides, phosphorylated ceramides, phosphorylated bases and free sphingoid bases
were separated on a C8 column (Luna, 5 µm, 150 x 2 mm; Phenomenex) using a gradient of
solvent A (methanol:water:formic acid; 99:1:0.2; v/v/v) and solvent B (methanol:water:formic
acid; 1:99:0.2; v/v/v) as follows: 70-97% A (0-2 min), 97-99% A (2-20 min), 97-99% A (20-21
min), 99% B (21-55 min), 99-70% A (55-56 min) and 70% A (56-60 min), at a flow rate 0.2
ml/min. Compounds were fragmented using argon as collision gas. The analytes were monitored
in the positive ion mode by MRM (collision energy in parentheses): C18 S m/z 300>282 (21 eV);
C18 DS m/z 302>284 (21 eV); C18 S1P m/z 380>264 (25 eV); C18 DS1P m/z 382>266 (25 eV);
d18:1/14:0 C1P m/z 590>264 (43 eV); d18:1/16:0 C1P m/z 619>264 (43 eV); d18:0/16:0 C1P
m/z 621>266 (46 eV); d18:1/18:0 C1P m/z 647>264 (48 eV); CER[N(24)S(16)] m/z 622>236
(40 eV); CER[N(22)S(18)] m/z 622>264 (40 eV); CER[N(24)S(17)] m/z 636>250 (40 eV);
CER[N(23)S(18)] m/z 636>264 (40 eV); CER[N(26)S(16)] m/z 650>236 (40 eV);
CER[N(24)S(18)] m/z 650>264 (40 eV); CER[N(22)S(20)] m/z 650>292 (40 eV);
CER[N(24)S(19)] m/z 664>278 (40 eV); CER[N(24)S(20)] m/z 678>292 (40 eV);
CER[N(25)S(20)] m/z 692>292 (40 eV); CER[N(24)S(21)] m/z 692>306 (40 eV);
CER[N(25)S(22)] m/z 720>320 (40 eV); CER[N(24)S(24)] m/z 734>348 (40 eV).
Reliable standards for all of the sphingolipid species of interest are not yet commercially
available, and so we employed a relative quantification approach using class-specific internal
standards. Since the internal standards are provided with molar concentration rather than massbased concentration, the sphingolipid data are reported in a molar fashion, covering the range
5
0.005-8.0 nmol/ml. Recoveries for all sphingoid bases and ceramides assayed here were based on
the internal standards, and ranged from 75-79%.
SM4. Fatty acid analysis
Fatty acids were analysed by gas chromatography (GC) in dermis and epidermis as previously
described (Pilkington et al., 2014). Briefly, skin samples were homogenised and lipids extracted
using chloroform:methanol (2:1; v/v) containing butylated hydroxy toluene (BHT; 0.01 % w/v).
Fatty acid methyl esters were prepared using BF3 in methanol (Sigma Aldrich, Poole, UK) and
analysed by gas chromatography with flame ionisation detector and a BPX70 GC capillary
column (length 60 m, internal diameter 0.25 mm, film thickness 0.25 µm; SGE Europe, Milton
Keynes, UK). Heneicosaenoic acid (Sigma Aldrich) was used as the internal standard with a 37
fatty acid methyl ester mixed standard (Supelco, Belleforte, PA, USA), the reference for
identification of fatty acids. Results are expressed as percentage of total fatty acids.
6
Supplementary Table S1. Mean fold change in expression of bioactive lipids from dermis to
epidermis (supplementary to Figure 2).
Lipid Mediator
Prostanoids
Hydroxy fatty acids
Endocannabinoids
and NAE
Compound
PGE1**
PGD1**
13,14 dihydro-15-keto PGE1*
PGF1α**
13,14 dihydro PGE1**
PGE2***
PGD2**
13,14 dihydro-15-keto PGE2*
TXB2
PGJ2*
15-keto PGE2**
PGF2α**
13,14 dihydro-15-keto PGF2α
6-keto PGF1a***
PGE3***
PGD3
PGF3α
9-HODE*
13-HODE*
15-HETrE
5-HETE
8-HETE*
11-HETE**
15-HETE
12-HETE**
11-HEPE**
12-HEPE
14-HDHA
13-HDHA*
PEA
ALEA
LEA
OEA
SEA
EPEA*
AEA*
Fold change in
epidermis (mean ± SE)
15.51 ± 4.98
#
†
32.13 ± 12.39
13.97 ± 4.00
54.73 ± 23.70
37.45 ± 14.95
14.48 ± 4.77
43.17 ± 16.44
†
†
9.62 ± 2.83
†
2.51 ± 0.72
†
19.18 ± 6.96
9.11± 2.75
17.20 ± 6.38
7.55 ± 2.92
2.79 ± 0.86
†
†
1.81 ± 0.63
5.74 ± 1.74
7.75 ± 1.95
6.64 ± 1.36
7.48 ± 3.59
†
†
2.94 ± 1.03
2.87 ± 0.79
8.18 ± 3.59
4.25 ± 1.72
4.03 ± 1.68
2.81 ± 0.50
2.52 ± 0.57
7
Free Sphingoid Bases
Phosphorylated Bases
Phosphorylated Ceramides
Non-hydroxylated Ceramides
DHEA*
2-AG
C18 S
C 18 DS**
C18 S1P*
C18 DS1P*
d18:1/14:0
d18:1/16:0
d18:0/16:0
d18:1/18:0
N(24)S(16)
N(22)S(18)
N(24)S(17)
N(23)S(18)
N(26)S(16)**
N(24)S(18)
N(22)S(20)
N(24)S(19)
N(24)S(20)**
N(25)S(20)
N(24)S(21)
N(25)S(22)
N(24)S(24)
2.66 ± 0.49
10.89 ± 5.32
10.38 ± 4.28
13.62 ± 2.32
2.34 ± 0.37
2.28 ± 0.40
3.03 ± 0.66
1.07 ± 0.23
0.69 ± 0.17
4.88 ± 2.82
4.80 ± 1.01
129.49 ± 92.48
8.04 ± 3.04
7.33 ± 3.86
25.45 ± 4.97
233.88 ± 169.37
13.93 ± 4.84
17.87 ± 9.96
7.75 ± 0.84
26.47 ± 12.83
92.72 ± 62.08
185.24 ± 133.07
130.25 ± 81.91
# Fold-change from dermal to epidermal expression not possible as not expressed in epidermis
† Fold-change from dermal to epidermal expression not possible as not expressed in dermis
*P<0.05, **P<0.01 and ***P<0.001 for epidermis vs dermis
8
Supplementary Table S2. Fatty acid composition of dermal and epidermal tissue samples
(supplementary to Figure 3).
Common name
lauric acid
myristic acid
myristoleic acid
pentadecylic acid
palmitic acid
palmitoleic acid
margaric acid
heptadecenoic acid
stearic acid
elaidic acid
oleic acid
vaccenic acid
linolelaidic acid
linoleic acid
γ-linolenic acid
α-linolenic acid
arachidic acid
eicosenoic acid
cis-11,14-eicosadienoic acid
dihomo-γ-linolenic acid
arachidonic acid
eicosatrienoic acid
behenic acid
erucic acid
eicosapentaenoic acid
tricosylic acid
cis-13,16,docosadienoic acid
lignoceric acid
nervonic acid
docosapentaenoic acid
docosahexaenoic acid
Shorthand notation
C12:0
C14:0
C14:1
C15:0
C16:0
C16:1
C17:0
C17:1
C18:0
C18:1n-9t
C18:1n-9c
C18:1n-7
C18:2n-6t
C18:2n-6c
C18:3n-6
C18:3n-3
C20:0
C20:1n-9
C20:2
C20:3n-6
C20:4n-6
C20:3n-3
C22:0
C22:1n-9
C20:5n-3
C23:0
C22:2
C24:0
C24:1
C22:5n-3
C22:6n-3
Percentage of total fatty acids (mean ± SE)
Dermis
Epidermis
0.18 ± 0.02
0.35 ± 0.11
1.87 ± 0.20
2.70 ± 0.44
0.68 ± 0.16
0.19 ± 0.09
0.25 ± 0.02
0.89 ± 0.24
19.89 ± 3.03
23.91 ± 2.21
11.46 ± 1.67
4.08 ± 1.45
0.16 ± 0.05
0.78 ± 0.18
0.23 ± 0.07
0.36 ± 0.14
2.85 ± 0.61
22.08 ± 7.49
0.40± 0.11
0.13 ± 0.03
44.78 ± 3.53
24.25 ± 6.79
3.19 ± 0.31
1.98 ± 0.49
0.12 ± 0.04
0.31 ± 0.14
10.73 ± 0.93
9.58 ± 1.31
0.09 ± 0.00
0.15 ± 0.06
0.73 ± 0.13
0.52 ± 0.06
0.18 ± 0.04
0.17 ± 0.07
0.41 ± 0.02
0.42 ± 0.04
0.26 ± 0.11
0.22 ± 0.05
0.18 ± 0.03
0.59 ± 0.10
0.69 ± 0.06
2.72 ± 0.54
0.03 ± 0.01
0.53 ± 0.12
0.02 ± 0.00
0.00 ± 0.00
0.04 ± 0.01
0.44 ± 0.12
0.07 ± 0.01
0.50 ± 0.32
0.01 ± 0.00
0.17 ± 0.04
0.01 ± 0.00
0.16 ± 0.06
0.04 ± 0.01
0.65 ± 0.34
0.03 ± 0.01
0.26 ± 0.08
0.18 ± 0.02
0.45± 0.11
0.19 ± 0.03
0.45 ± 0.10
9