Macrocystis pyrifera Lipids Reduce Cytokine-Induced Pro-Inflammatory Signalling and Barrier Dysfunction in Human Keratinocyte Models

Atopic dermatitis is a chronic condition where epidermal barrier dysfunction and cytokine production by infiltrating immune cells exacerbate skin inflammation and damage. A total lipid extract from Macrocystis pyrifera, a brown seaweed, was previously reported to suppress inflammatory responses in monocytes. Here, treatment of human HaCaT keratinocytes with M. pyrifera lipids inhibited tumour necrosis factor (TNF)-α induced TNF receptor-associated factor 2 and monocyte chemoattractant protein (MCP)-1 protein production. HaCaT cells stimulated with TNF-α, interleukin (IL)-4, and IL-13 showed loss of claudin-1 tight junctions, but little improvement was observed following lipid pre-treatment. Three-dimensional cultures of HaCaT cells differentiated at the air–liquid interface showed increased MCP-1 production, loss of claudin-1 tight junctions, and trans-epidermal leakage with TNF-α, IL-4, and IL-13 stimulation, with all parameters reduced by lipid pre-treatment. These findings suggest that M. pyrifera lipids have anti-inflammatory and barrier-protective effects on keratinocytes, which may be beneficial for the treatment of atopic dermatitis or other skin conditions.


Introduction
Atopic dermatitis is a common inflammatory skin condition that is prevalent in children (15-30%) and to a lesser extent in adults (1-3%) [1].Atopic dermatitis is characterised by a damaged skin barrier, recurrent infection, and a transition from a T helper (Th)2 to Th1 immune response during disease progression [2,3].
The epidermis of the skin is made of stratified keratinocytes that form a replicative basal layer and numerous differentiated suprabasal layers [4].Keratinocytes function as the main structural and barrier components of the epidermis [5] and play a critical role in the initiation and maintenance of skin inflammatory and wound healing responses [6].Keratin 5 and keratin 14 are expressed exclusively in basal keratinocytes, whereas keratin 1, keratin 10, involucrin, and filaggrin are expressed upon their differentiation [4,5].In the most superficial epidermal layer, flattened keratinocytes called corneocytes are embedded in a lipid matrix consisting of ceramides, cholesterol, and free fatty acids [7].Tight junctions between keratinocytes are critical to maintaining the epidermal barrier [8,9].Claudin-1, a tight junction protein that predominates in the skin [10], regulates keratinocyte proliferation and differentiation [11], with amounts inversely correlating to the extent of epidermal barrier leakage [12][13][14].Disruption to this epidermal barrier contributes to the development and severity of atopic dermatitis [15], with changes in lipid content correlating to transepidermal water loss and colonisation by Staphylococcus aureus in atopic children and dogs [16][17][18].
The acute and chronic phases of atopic dermatitis are characterised by Th2 and Th1 lymphocyte responses, respectively [19,20].Th cells maintain atopic immune responses in the skin through the secretion of specific effector cytokines [3,20].For example, tumour necrosis factor (TNF)-α and interferon (IFN)-γ are produced by Th1 cells, while Th2 cells produce interleukin (IL)-4, IL-5, and IL-13.Th2 cytokines are considered the most potent activators of keratinocytes, modulating tight junctions which control epidermal barrier function through the induction of signal transducer and activator of transcription (STAT) signalling [21,22].Th1 cytokines also activate nuclear factor kappa-light-chain-enhancer of activated B cell (NF-kB) signalling in keratinocytes, increasing their production of proinflammatory chemokines such as monocyte chemoattractant protein (MCP)-1, which exacerbates skin inflammation through recruitment of neutrophils and monocytes from the blood [23][24][25].
Treatments for atopic dermatitis fall into two categories: lipid-based emulsions that coat, protect, and hydrate the skin, and synthetic drugs that suppress inflammatory signalling within the skin [26].Topical corticosteroids, such as hydrocortisone, mediate their anti-inflammatory effects by binding with the glucocorticoid receptor suppressing NF-kB-mediated transcription of pro-inflammatory mediators.Emollients containing plant-derived lipids provide an artificial barrier against trans-epidermal water loss and environmental triggers [27].A systemic review supported that emollients alleviate atopic dermatitis symptoms, particularly when combined with corticosteroids [28].However, issues with efficacy, tolerability, and patient compliance with these treatment classes have led to the search for therapeutic alternatives [26,29].
Human keratinocyte and epidermal models are commonly used to assess the utility of topical therapeutics prior to testing in animal models [30][31][32].In general, keratinocytes isolated from human skin maintain a basal phenotype in the presence of low calcium concentrations and differentiate when exposed to high calcium (>0.1 mM) [33].But in vitro experimentation with primary keratinocytes has limitations, including a short lifespan and requirement for growth factor supplementation [34].As a result, immortalised human keratinocyte cell lines are often used for in vitro research owing to their capacity to avoid senescence and survive indefinitely in culture.HaCaT cells, a spontaneously transformed aneuploid immortal keratinocyte cell line derived from adult human skin, closely mimic the phenotype of primary human keratinocytes [33].Stimulation of HaCaT cells with IFN-γ, TNF-α, and/or IL-4 alter the expression of pro-inflammatory mediators and differentiation proteins, as reported in skin lesions of patients with atopic dermatitis [35].HaCaT cells also express tight junction proteins, such as claudin 1, similarly to human neonatal and adult keratinocytes [36].After culture at the air-liquid interface, HaCaT cells also form a squamous epithelium that expresses most differentiation markers [37].Additionally, when transplanted onto the subcutaneous tissue of nude mice, HaCaT cells developed a normal epidermal architecture [38].HaCaT cells have been employed to investigate epidermal biology, skin pathologies, and topical therapies [39][40][41].
Additionally, seaweed components have been reported to preserve epithelial barrier function in a manner that may be useful for the treatment of atopic dermatitis.An aqueous extract from Laminaia japonica reduced lipoteichoic acid-stimulated intestinal permeability in polarised Caco-2 epithelial cell monolayers by restoring expression of the tight junction protein, occludin [50].In addition, a sulphated polysaccharide from Cladosiphon okamuranus Tokida restored claudin-1 and claudin-2 protein abundance and prevented H 2 O 2 -induced permeability in the same intestinal model [51].As well, an extract containing a mixture of fucoidan and fucoxanthin from S. hemiphyllum enhanced occludin, claudin-1, and claudin-2 mRNA expression in Caco-2 cells challenged with lipopolysaccharide (LPS) [52].Further, lipid-rich extracts from Himantothallus grandifolius, Plocamium cartilagineum, Phaeurus antarcticus, and Kallymenia antarctica restored barrier integrity to Caco-2 and HaCaT monolayers exposed to LPS or ultraviolet (UV)-B light, respectively [53].
Lipid emulsions are used therapeutically to provide a protective skin barrier, while seaweed lipids have been shown to inhibit inflammation and damage induced by Th1 cytokines, pathogen-derived toll-like receptor agonists, and UV-B light.However, little is known as to whether lipid emulsions or seaweed extracts modulate skin inflammation and barrier dysfunction induced by Th2 cytokines.Therefore, this study investigated the effects of lipids extracted from the brown seaweed, M. pyrifera, on Th1 and Th2 cytokine-mediated inflammation and tight junction dysfunction in HaCaT cell monolayers and 3D epidermal constructs.This study aimed to determine if seaweed lipids could offer any therapeutic benefit for individuals suffering from atopic dermatitis.

M. pyrifera Lipids Inhibit Pro-Inflammatory Signalling in Keratinocyte Monolayers Stimulated by the Th1 Cytokine TNF-α
HaCaT cell monolayers were treated with a total lipid extract from M. pyrifera prepared previously [44], with cytotoxicity assessed using a crystal violet assay.At concentrations of ≥0.5 mg•mL −1 , cells showed reduced viability relative to the vehicle-only control (48-84%, Figure 1a, p ≤ 0.05).For subsequent testing, an extract concentration of 0.13 mg•mL −1 was used, as this suppressed inflammatory signalling in monocytes [44].
HaCaT cell monolayers were stimulated with Th1 cytokine TNF-α and treated with M. pyrifera extract, with effects on the production of the chemokine MCP-1 determined using an enzyme-linked immunosorbent assay (ELISA).The concentration of TNF-α utilised was guided by prior reports [33,35,49,54,55].The TNF-α-induced increase in MCP-1 production was significantly reduced by 2, 6, and 24 h pre-treatment and co-treatment, and 2 and 6 h post-treatment (Figure 1b, p ≤ 0.05), with 2 h extract pre-treatment reducing levels closest to that of unstimulated cells.Thus, for subsequent analyses, treatments were applied 2 h before cytokine stimulation.Hydrocortisone was investigated as a positive control due to its application as a topical treatment for skin inflammation.Hydrocortisone was non-toxic to HaCaT cell monolayers at concentrations ≤ 10 mg•mL −1 (Figure 1c).Hydrocortisone pre-treatment at the non-toxic concentration of 10 mg•mL −1 significantly suppressed MCP-1 production stimulated by TNF-α to below that of unstimulated cells (Figure 1d, p ≤ 0.05).These findings indicate that M. pyrifera lipids reduce TNF-α-induced chemokine production in keratinocyte monolayers to a similar extent as hydrocortisone.The main cell surface receptor responsible for TNF-α signalling within keratinocytes is TNF receptor 1 (TNFR1), which upon ligand binding is translocated to lipid rafts where its intracellular domain interacts with adaptor proteins that recruit TNF receptor associated factor 2 (TRAF2), leading to downstream activation of NF-κB and its transcription of pro-inflammatory cytokines [23,25].Immunohistochemistry was therefore used to assess TNFR1 signalling in the HaCaT cell monolayer after stimulation with TNF-α and treatment with M. pyrifera extract.Cell surface staining for TNFR1 was localised at the junction of unstimulated cells, with punctate staining spread across cells after 2 h stimulation with TNF-α (Figure 2a).TNFR1 staining at cell-cell junctions was also observed after 2 h pretreatment with the extract; however, after TNF-α stimulation for 2 h, the staining appeared to redistribute to cell poles (Figure 2a).Intracellular staining for TRAF2 was observed near the nuclear membrane of unstimulated cells, with staining increasing and redistributing throughout TNF-α-stimulated cells within 2 h (Figure 2b).In the extract pre-treated cells, TRAF2 staining was distributed throughout the cell, but the staining intensity appeared to be reduced relative to unstimulated and TNF-α-stimulated cells (Figure 2b).Values are expressed relative to the vehicle-only control (0 mg•mL −1 ).Cells were stimulated with TNF-α (20 ng•mL −1 ) with or without (b) pre-, co-, or post-treatment with extract (0.13 mg•mL −1 ) or (d) 2 h pre-treatment with HC (10 mg•mL −1 ).After 24 h incubation, conditioned medium was collected and assayed for MCP-1 production by ELISA.(−) represents unstimulated cells.(+) represents cells stimulated with TNF-α.Values are expressed as percentages of the (a,c) vehicle or (b,d) (+) controls and presented as means ± SD (n = 3), with those that differ significantly from those controls identified by one-way ANOVA followed by a Dunnett's test (* p ≤ 0.05).
The main cell surface receptor responsible for TNF-α signalling within keratinocytes is TNF receptor 1 (TNFR1), which upon ligand binding is translocated to lipid rafts where its intracellular domain interacts with adaptor proteins that recruit TNF receptor associated factor 2 (TRAF2), leading to downstream activation of NF-κB and its transcription of pro-inflammatory cytokines [23,25].Immunohistochemistry was therefore used to assess TNFR1 signalling in the HaCaT cell monolayer after stimulation with TNF-α and treatment with M. pyrifera extract.Cell surface staining for TNFR1 was localised at the junction of unstimulated cells, with punctate staining spread across cells after 2 h stimulation with TNF-α (Figure 2a).TNFR1 staining at cell-cell junctions was also observed after 2 h pretreatment with the extract; however, after TNF-α stimulation for 2 h, the staining appeared to redistribute to cell poles (Figure 2a).Intracellular staining for TRAF2 was observed near the nuclear membrane of unstimulated cells, with staining increasing and redistributing throughout TNF-α-stimulated cells within 2 h (Figure 2b).In the extract pre-treated cells, TRAF2 staining was distributed throughout the cell, but the staining intensity appeared to be reduced relative to unstimulated and TNF-α-stimulated cells (Figure 2b). ) with or without 2 h pre-treatment with M. pyrifera extract (0.13 mg•mL −1 ).After 0 and 2 h stimulation with TNF-α, cells were (a) fixed for extracellular staining or (b) fixed and permeabilised for intracellular staining, using rabbit (a) anti-TNFR1 or (b) anti-TRAF2 and anti-rabbit Alexa Fluor ® 488 antibodies (green).Nuclei were stained using DAPI (blue).Representative images are shown, with yellow and white arrows indicating cell junction and pole localisation, respectively.Scale bar, 50 µm.After 2 or 4 h stimulation with TNF-α, Western blot analyses were performed on cell lysates using rabbit anti-GAPDH, (c) anti-TNFR1, or (d) anti-TRAF2, and anti-rabbit HRP antibodies.Band densities were determined from Western blots detecting (e) TNFR1 and (f) TRAF2.(−) represents unstimulated cells.(+) represents cells stimulated with TNF-α.Values are expressed relative to the (+) control and loading control (GAPDH) and presented as mean ± SD (n = 3), with those that differ significantly from the (+) control identified by one-way ANOVA followed by a Dunnett's test (* p ≤ 0.05).
Western blot analyses were used to quantify TNFR1 and TRAF2 abundance in cytokine-stimulated and extract-treated monolayers.Bands were observed as 55 and 56 kDa corresponding to TNFR1 and TRAF2, respectively (Figure 2c,d).Two bands of 36 and 38 kDa were observed for the loading control, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), which may reflect post-translational modifications [56,57].Differences were noted in loading control density across samples, but these were not consistent across the TNFR1 and TRAF2 Western blots (Figure 2c,d).When normalised to the loading control Western blot analyses were used to quantify TNFR1 and TRAF2 abundance in cytokinestimulated and extract-treated monolayers.Bands were observed as 55 and 56 kDa corresponding to TNFR1 and TRAF2, respectively (Figure 2c,d).Two bands of 36 and 38 kDa were observed for the loading control, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), which may reflect post-translational modifications [56,57].Differences were noted in loading control density across samples, but these were not consistent across the TNFR1 and TRAF2 Western blots (Figure 2c,d).When normalised to the loading control bands, no changes in TNFR1 abundance were detected (Figure 2e); however, the increase in TRAF2 density with TNF-α stimulation was suppressed by extract pre-treatment (Figure 2f, p ≤ 0.05).These findings suggest that M. pyrifera lipids may limit MCP-1 production in keratinocyte monolayers by modulating TNFR1 localisation and TRAF2 production.
HaCaT cell monolayers were treated with a fresh M. pyrifera total lipid extraction, with cytotoxicity reassessed by crystal violet staining.Similar to the first batch of extract, at concentrations of ≥0.63 mg•mL −1 , cell viability was reduced (Figure S1a, p ≤ 0.05).TNFα-stimulated HaCaT cell monolayers were pre-treated with this extract (at an equivalent concentration of 0.16 mg•mL −1 ), with MCP-1 production reassessed by ELISA.MCP-1 production was suppressed to levels observed in unstimulated cells (Figure S1b, p ≤ 0.05).
These findings indicate that the M. pyrifera lipid extracts were equivalent in activity and that an extract concentration of 0.16 mg•mL −1 was suitable for subsequent testing.

M. pyrifera Lipids Do Not Prevent Disruption of Claudin-1 Tight Junctions in Keratinocyte Monolayers Stimulated by Th1 and Th2 Cytokines
HaCaT cell monolayers were stimulated with the Th1 cytokine TNF-α or the Th2 cytokines IL-4 and IL-13, with immunocytochemistry used to determine their effects on claudin-1 tight junctions.The concentrations of TNF-α, IL-4, and IL-13 utilised were guided by prior reports [35,54,55].In unstimulated monolayers, claudin-1 staining appeared as a continuous and distinct line at cell-cell junctions ( Figures 3 and S2).In TNF-α-stimulated monolayers, claudin-1 remained at the cell-cell junctions, although increased content was observed within the cell (Figure S2).Following IL-4 and IL-13 stimulation, claudin-1 staining was observed at cell-cell junctions, but in less cells than the number observed for unstimulated monolayers (Figure S2).Greater disruption of claudin-1 staining from cell-cell junctions was observed in monolayers stimulated concurrently with TNF-α, IL-4, and IL-13 (Figures 3 and S2).
Int. J. Mol.Sci.2023, 24, x FOR PEER REVIEW 6 of 20 bands, no changes in TNFR1 abundance were detected (Figure 2e); however, the increase in TRAF2 density with TNF-α stimulation was suppressed by extract pre-treatment (Figure 2f, p ≤ 0.05).These findings suggest that M. pyrifera lipids may limit MCP-1 production in keratinocyte monolayers by modulating TNFR1 localisation and TRAF2 production.
HaCaT cell monolayers were treated with a fresh M. pyrifera total lipid extraction, with cytotoxicity reassessed by crystal violet staining.Similar to the first batch of extract, at concentrations of ≥0.63 mg•mL −1 , cell viability was reduced (Figure S1a, p ≤ 0.05).TNFα-stimulated HaCaT cell monolayers were pre-treated with this extract (at an equivalent concentration of 0.16 mg•mL −1 ), with MCP-1 production reassessed by ELISA.MCP-1 production was suppressed to levels observed in unstimulated cells (Figure S1b, p ≤ 0.05).
These findings indicate that the M. pyrifera lipid extracts were equivalent in activity and that an extract concentration of 0.16 mg•mL −1 was suitable for subsequent testing.

M. pyrifera Lipids Do Not Prevent Disruption of Claudin-1 Tight Junctions in Keratinocyte Monolayers Stimulated by Th1 and Th2 Cytokines
HaCaT cell monolayers were stimulated with the Th1 cytokine TNF-α or the Th2 cytokines IL-4 and IL-13, with immunocytochemistry used to determine their effects on claudin-1 tight junctions.The concentrations of TNF-α, IL-4, and IL-13 utilised were guided by prior reports [35,54,55].In unstimulated monolayers, claudin-1 staining appeared as a continuous and distinct line at cell-cell junctions (Figures 3 and S2).In TNF-α-stimulated monolayers, claudin-1 remained at the cell-cell junctions, although increased content was observed within the cell (Figure S2).Following IL-4 and IL-13 stimulation, claudin-1 staining was observed at cell-cell junctions, but in less cells than the number observed for unstimulated monolayers (Figure S2).Greater disruption of claudin-1 staining from cell-cell junctions was observed in monolayers stimulated concurrently with TNF-α, IL-4, and IL-13 (Figures 3 and S2).Claudin-1 tight junctions in TNF-α-, IL-4-, and IL-13-stimulated HaCaT cell monolayers were assessed after pre-treatment with the fresh M. pyrifera extract.A similar claudin-1 staining pattern was observed relative to cytokine-stimulated cells but with a greater number of cells observed with light staining at cell-cell junctions (Figure 3).This was not observed in cytokine-stimulated cells following hydrocortisone pre-treatment, where every cell showed complete disruption of claudin-1 tight junctions (Figure 3).
The abundance of claudin-1 in HaCaT cell monolayers was also quantified using Western blotting.Claudin-1 was observed as a single band at ≈20 kDa (Figure S3a), while the loading control β-tubulin was identified at ≈55 kDa.Neither cytokine stimulation nor extract pre-treatment altered the claudin-1 band density relative to unstimulated cells (Figure S3b, p > 0.05).To more accurately mimic the human epidermis and facilitate measures of barrier functionality, HaCaT cells were cultured in Transwell ® inserts for 2 days, then lifted to the air-liquid interface and cultured with differentiation stimuli for 6 days [53,58,59].The differentiation state of the constructs was evaluated at day 2 and 6 of the differentiation protocol using whole-mount immunocytochemistry and confocal microscopy.Optical slices from confocal Z-stacks revealed increased expression of markers of keratinocyte differentiation, namely, keratin-10 and filaggrin, from day 2 to day 6 (Figure 4a,b).Keratin-10 staining was distributed throughout the cell layers at day 6 (Figure S4a), while filaggrin staining was localised to the apical side of the cell layers (Figure S4b).These observations suggest that 3D constructs were produced which could be used to assess barrier function.
HaCaT cell constructs were stimulated with the Th1 cytokine TNF-α or the Th2 cytokines IL-4 and IL-13, individually and in combination, with effects on trans-epidermal leakage assessed using a dye exclusion assay.The TNF-α, IL-4, and IL-13 concentrations were as used previously, as they were consistent with previous studies in 3D skin equivalents [55].A dye exclusion assay was used to assess barrier function as prior reports indicated HaCaT cells do not support measures of trans-epithelial electrical resistance [53,60] but do limit permeation of topically applied fluorescent dyes such as lucifer yellow [53,55,61,62].With unstimulated controls, ≈35% of the lucifer yellow passed from the apical to the basal side of the 3D constructs within 60 min (Figure 4c).While constructs stimulated with IL-4 and IL-13 showed negligible change in dye leakage relative to that of unstimulated controls, TNF-α-stimulated constructs showed a cumulative increase in lucifer yellow leakage, with ≈55% passing through the construct after 60 min (Figure 4c).The greatest increase in lucifer yellow leakage was, however, observed following stimulation with TNF-α, IL-4, and IL-13, with ≈80% passing through the construct after 60 min (Figure 4c), culminating in a ≈3.3-fold increase in the rate of dye leakage than unstimulated constructs (Figure 4d).In subsequent experiments, TNF-α, IL-4, and IL-13 cytokines were used in combination to induce epidermal barrier dysfunction.
HaCaT cell constructs were pre-treated with the extract for 2 h, then stimulated with TNF-α, IL-4, and IL-13, with trans-epidermal leakage determined using the dye exclusion assay.3D constructs stimulated with cytokines showed ≈78% lucifer yellow dye leakage at 60 min, while leakage in the extract pre-treated constructs was similar to that from the unstimulated controls (58% versus 55%, respectively, Figure 4e).The rate of dye leakage increased ≈1.6-fold following cytokine-stimulation (Figure 4f) but remained similar to that of unstimulated constructs following extract pre-treatment.These findings suggest that the M. pyrifera extract may limit cytokine-induced trans-epidermal leakage of lucifer yellow dye.The viability of the 3D epidermal constructs after cytokine stimulation and extract pre-treatment was assessed using a lactate dehydrogenase (LDH) assay.No significant differences in LDH levels were detected between constructs (Figure 5a).MCP-1 production by the 3D epidermal constructs after cytokine stimulation and extract pre-treatment was assessed by ELISA.Stimulation with TNF-α, IL-4, and IL-13 increased MCP-1 production relative to that of unstimulated constructs, while pre-treatment with M. pyrifera extract significantly suppressed this response (Figure 5b, p ≤ 0.05).These findings indicate that the M. pyrifera extract limits cytokine-induced chemokine production in epidermal constructs without compromising viability.
Int. J. Mol.Sci.2023, 24, x FOR PEER REVIEW 9 of 20 The viability of the 3D epidermal constructs after cytokine stimulation and extract pre-treatment was assessed using a lactate dehydrogenase (LDH) assay.No significant differences in LDH levels were detected between constructs (Figure 5a).MCP-1 production by the 3D epidermal constructs after cytokine stimulation and extract pre-treatment was assessed by ELISA.Stimulation with TNF-α, IL-4, and IL-13 increased MCP-1 production relative to that of unstimulated constructs, while pre-treatment with M. pyrifera extract significantly suppressed this response (Figure 5b, p ≤ 0.05).These findings indicate that the M. pyrifera extract limits cytokine-induced chemokine production in epidermal constructs without compromising viability.Claudin-1 tight junctions were assessed in TNF-α-, IL-4-, and IL-13-stimulated and M. pyrifera extract-pre-treated 3D epidermal constructs using whole-mount immunocytochemistry and confocal microscopy (Figure 6a).Claudin-1 staining was observed throughout the unstimulated constructs, localising at cell-cell boundaries.Following cytokine stimulation, there was a substantial loss of claudin-1 staining, with punctate staining evident across the construct (Figure 6a).Following extract pre-treatment, claudin-1 staining was increased relative to the stimulated constructs, with both punctate and junctional staining observed (Figure 6a).
Western blot analyses were used to quantify claudin-1 abundance in cytokine-stimulated and extract-treated 3D epidermal constructs.Bands corresponding to claudin-1 and β-tubulin were detected for all constructs, at ≈20 and ≈55 kDa, respectively (Figure 6b).Stimulation with cytokines significantly decreased the band density for claudin-1 by ≈58% (Figure 6c, p ≤ 0.05), while a partial restoration in claudin-1 protein levels was observed in constructs pre-treated with the extract.These findings suggest that the M. pyrifera lipid extract may protect claudin-1 tight junctions in the 3D epidermal constructs from cytokine-induced disruption to a greater extent than that observed in HaCaT cell monolayers.Claudin-1 tight junctions were assessed in TNF-α-, IL-4-, and IL-13-stimulated and M. pyrifera extract-pre-treated 3D epidermal constructs using whole-mount immunocytochemistry and confocal microscopy (Figure 6a).Claudin-1 staining was observed throughout the unstimulated constructs, localising at cell-cell boundaries.Following cytokine stimulation, there was a substantial loss of claudin-1 staining, with punctate staining evident across the construct (Figure 6a).Following extract pre-treatment, claudin-1 staining was increased relative to the stimulated constructs, with both punctate and junctional staining observed (Figure 6a).
Western blot analyses were used to quantify claudin-1 abundance in cytokine-stimulated and extract-treated 3D epidermal constructs.Bands corresponding to claudin-1 and β-tubulin were detected for all constructs, at ≈20 and ≈55 kDa, respectively (Figure 6b).Stimulation with cytokines significantly decreased the band density for claudin-1 by ≈58% (Figure 6c, p ≤ 0.05), while a partial restoration in claudin-1 protein levels was observed in constructs pre-treated with the extract.These findings suggest that the M. pyrifera lipid extract may protect claudin-1 tight junctions in the 3D epidermal constructs from cytokine-induced disruption to a greater extent than that observed in HaCaT cell monolayers.

Fatty Acid Composition of M. pyrifera Lipid Extracts
The fatty acid composition of the M. pyrifera extract prepared for this study was evaluated using gas chromatography-mass spectrometry (GC-MS) (Figure S5), with individual fatty acids quantified as percentages of the total (Table S1).In this extract, saturated fatty acids were most abundant, at ≈30%, with palmitic acid being the largest constituent (C16:0, ≈21%).Omega-6 fatty acids represented ≈22%, with the most abundant being

Discussion
This study provides the first evidence of the protective anti-inflammatory and barrierenhancing effects of seaweed-derived lipids in keratinocyte models of atopic dermatitis.Treatment with M. pyrifera lipids reduced inflammatory signalling and chemokine production in keratinocyte monolayers stimulated by the Th1 cytokine TNF-α.Lipid pre-treatment did not protect keratinocyte monolayers against claudin-1 tight junction loss induced by concurrent stimulation with Th1 and Th2 cytokines, TNF-α, IL-4, and IL-13.Epidermal constructs stimulated with these cytokines showed improved barrier functionality after lipid pre-treatment, with partial reductions in chemokine production, claudin-1 tight junction loss, and trans-epidermal leakage observed.
The acute and chronic phases of atopic dermatitis are characterised by Th2 and Th1 cell responses, respectively [19,20].Seaweed lipids have been reported to attenuate proinflammatory cytokine production in Th1 cytokine-stimulated keratinocytes.In HaCaT cells stimulated with TNF-α and IFN-γ, an ethanolic extract of S. horneri suppressed IL-1β, IL-4, IL-6, IL-13, IFN-γ, and TNF-α [63], while a preparation of Capsosiphone fulvescens inhibited TNF-α and IL-6 production [49].Here, M. pyrifera lipids inhibited MCP-1 production in TNF-α-stimulated HaCaT cells to a similar extent as 1% hydrocortisone, a standard treatment for atopic dermatitis.The anti-inflammatory action of M. pyrifera lipids was also observed in 3D epidermal constructs stimulated simultaneously with TNF-α, IL-4, and IL-13, a model that more closely replicates atopic skin.Further, lipid pre-treatment altered the cell surface distribution of TNFR1 and inhibited the production of TRAF2, which suggests the lipids may prevent activation of the TNF-α signalling cascade.This is consistent with our prior study, where M. pyrifera lipid treatment reduced monocyte production of the cytosolic adapter protein, myeloid differentiation factor 88, preventing NF-kB-mediated inflammatory chemokine production in response to toll-like receptor activation [44].Together, these findings suggest a common mechanism whereby the antiinflammatory effects of the M. pyrifera lipids are mediated at the cell surface (Figure 7).Indeed, omega-3 fatty acid supplementation has been shown to increase the size and fluidity of lipid microdomains in cell membranes, affecting the assembly of receptors, adapters, and signalling molecules, modulating downstream inflammatory signalling [64][65][66].Therefore, the M. pyrifera lipids could be beneficial for Th1-mediated skin inflammation, offering alternative or combination therapies that could reduce corticosteroid usage and side effects.
Epidermal barrier dysfunction in atopic dermatitis is associated with disruptions to the claudin-1 tight junctions found between keratinocytes [10].Prior studies suggest that seaweed lipids protect keratinocytes from damage to this barrier.In HaCaT cells exposed to fine dust, an ethanol extract of S. horneri increased the abundance of various tight junction proteins, including claudin-1 [67].Further, various seaweed lipid extracts prevent barrier loss in HaCaT cells induced by UV-B light [53].In this study, the M. pyrifera lipids did not prevent the loss of claudin-1 protein at tight junctions in TNF-α-, IL-4-, and IL-13-stimulated HaCaT cell monolayers, although minor improvements were observed relative to 1% hydrocortisone treatment.In 3D epidermal constructs stimulated simultaneously with these Th1 and Th2 cytokines, treatment with the lipid extract partially restored claudin-1 protein density and tight junction localisation whilst reducing transepidermal leakage of lucifer yellow.These findings are in contrast with a previous report, where omega-3 fatty acids increased claudin-1 expression and localisation in intestinal cells treated with 2,4,6-trinitrobenzenesulfonic acid or TNF-α [68,69].Together, these findings indicate that seaweed lipids may primarily exert barrier protective effects in the context of Th1-mediated inflammation.Indeed, the protective effect of the M. pyrifera lipids was greater in the 3D epidermal construct than for the keratinocyte monolayer, where TNF-α, as opposed to IL-4 and IL-13, was the primary mediator of damage.TNF-α-mediated NF-kB signalling directly regulates the expression of tight junction proteins, including various claudins, occludin, and zonula occludens (ZO)-1, and indirectly controls their cellular localisation through the modulation of myosin light chain kinase and intracellular and extracellular protease activity [23,[70][71][72].By contrast, Th2 cytokine-mediated STAT signalling can induce ubiquitination and proteasomal degradation of tight junction proteins, including various claudins and ZO-1 [73].Omega-3 fatty acids are thought to influence tight junction localisation by altering cell membrane lipid microdomains [74], but the findings here suggest an alternative mechanism via the suppression of the TNFR1-TRAF2-NF-κB signalling pathway (Figure 7).However, investigations are warranted to determine if M. pyrifera lipids regulate tight junctions through the turnover pathways differentially regulated by Th1 and Th2 cytokines.Further, these studies should be extended beyond claudin-1 to the constituents of all cellular junctions that control epidermal permeability [8,9].leakage of lucifer yellow.These findings are in contrast with a previous report, wher omega-3 fatty acids increased claudin-1 expression and localisation in intestinal cell treated with 2,4,6-trinitrobenzenesulfonic acid or TNF-α [68,69].Together, these finding indicate that seaweed lipids may primarily exert barrier protective effects in the contex of Th1-mediated inflammation.Indeed, the protective effect of the M. pyrifera lipids wa greater in the 3D epidermal construct than for the keratinocyte monolayer, where TNF-α as opposed to IL-4 and IL-13, was the primary mediator of damage.TNF-α-mediated NF kB signalling directly regulates the expression of tight junction proteins, including variou claudins, occludin, and zonula occludens (ZO)-1, and indirectly controls their cellular lo calisation through the modulation of myosin light chain kinase and intracellular and ex tracellular protease activity [23,[70][71][72].By contrast, Th2 cytokine-mediated STAT signal ling can induce ubiquitination and proteasomal degradation of tight junction proteins including various claudins and ZO-1 [73].Omega-3 fatty acids are thought to influenc tight junction localisation by altering cell membrane lipid microdomains [74], but the find ings here suggest an alternative mechanism via the suppression of the TNFR1-TRAF2-NF κB signalling pathway (Figure 7).However, investigations are warranted to determine i M. pyrifera lipids regulate tight junctions through the turnover pathways differentiall regulated by Th1 and Th2 cytokines.Further, these studies should be extended beyond claudin-1 to the constituents of all cellular junctions that control epidermal permeabilit [8,9].The anti-inflammatory and barrier-protective effects of the M. pyrifera lipid extract should be investigated beyond cultured cells in models that better mimic atopic dermatitis in humans.Although HaCaT cells were successfully used in this study to simulate pathological processes involved in atopic dermatitis, the models used here have limitations.For example, HaCaT cells differ from primary human keratinocytes in their expression of cornified envelope-associated proteins [75], phospholipid content, and cholesterol metabolism [76].This was evident in the 3D epidermal construct employed here, where only limited layers of keratinocytes were formed, with partial filaggrin coverage and trans-epidermal leakage observed in the absence of cytokine stimulation.It would be interesting in future studies to determine whether treatment with the M. pyrifera lipids alters the differentiation status and lipid content or metabolism within the 3D epidermal constructs.Further, in vitro primary human keratinocyte models [30,31], or genetic or chemically induced mouse models of atopic dermatitis [32], could be used to validate these findings using alternative measures of epidermal permeability [8,9] and to assess dosing regimens of topical formulations containing M. pyrifera lipids.Gil and colleagues illustrated that a Capsosiphone fulvescens extract, which actively suppresses pro-inflammatory cytokine production in stimulated HaCaT cells, inhibited the development of erythema, haemorrhage, edema, scarring, and inflammatory cell infiltration in a NC/Nga mouse model of house dust mite-induced atopic dermatitis [49].This suggests that the effects of the M. pyrifera extract may be replicated in more biologically relevant models.
This study utilised two independent lipid extracts from M. pyrifera.The original extract had previously been shown to suppress pro-inflammatory cytokine production in THP-1 monocytes [44], with bioactivity-directed fractionation indicating that myristic acid, palmitoleic acid, and α-linolenic acid mediated this activity without compromising cell viability.Here, both the original and freshly prepared lipid extract suppressed pro-inflammatory chemokine production in HaCaT keratinocytes, at equivalent non-toxic concentrations.This was despite differences being observed in fatty acid composition.In seaweed, the main carriers of fatty acids are complex polar lipids such glycoglycerolipids and phospholipids [77,78].Esterified polar lipids, such as sulfoquinovosyl diacylglycerols, phosphatidylglycerols, phosphatidylcholines from Palmaria palmata, and monogalactosyldiacylglycerols and digalactosyldiacylglycerols from Chondrus crispus, showed greater anti-inflammatory properties than isolated polyunsaturated fatty acids [58,79].This indicates that the polar lipid class, rather than fatty acid composition, may have a greater influence on extract bioactivity.Although further compositional analyses of the M. pyrifera extracts would be beneficial, it appears that any differences in their lipid or fatty acid content did not greatly affect their biological activity.
In summary, lipids isolated from M. pyrifera reduced inflammatory signalling and barrier dysfunction in keratinocyte monolayers and epidermal constructs, mediated primarily via the Th1 cytokine TNF-α (Figure 7).Further compositional analyses, mechanistic studies, and pre-clinical testing will provide further insight into the reparative effects of M. pyrifera lipids for individuals suffering from atopic dermatitis or other skin conditions.

Total Lipid Extracts
Initial experiments utilised a previously reported total lipid extract prepared from M. pyrifera samples collected at Wellers Rock, Otago Harbour, New Zealand (45.79 • S, 170.71 • E), in October 2018 [44].Due to insufficient stocks, another total lipid extract was prepared from fresh M. pyrifera samples for the remaining experiments.Mature thalli were collected during low tide at the subtidal rocky shore from the same location, in September 2020, with a dried and pressed specimen deposited in the University of Otago Herbarium, Dunedin, New Zealand (identifier: OTA72991).This extraction was performed according to methods by Bligh and Dyer [80].Rinsed and dried seaweed (40 g) was powdered and homogenised (Ultra-Turrax) in chloroform: methanol: water [EMSURE ® , analytical grade, 400 mL, 1:2:0.8(v/v/v)].The macerate was filtered with Whatman filter paper no. 4 and subjected to phase separation with a concentrated saline solution (0.9%, w/v) overnight at room temperature in a separating funnel.The chloroform layer at the bottom of the flask was collected and evaporated using a vacuum rotary evaporator at 50-55 • C. Extracts were dried under a stream of nitrogen and stored at −20 • C. Extract yields were ≈1.5% and ≈2% of total lipids/gram of dry weight, respectively.
To stimulate epidermal inflammation and barrier dysfunction, 3D constructs were stimulated with TNF-α (20 ng•mL −1 ), IL-4 (50 ng•mL −1 ), and IL-13 (50 ng•mL −1 ).Where indicated, constructs were pre-treated with the extract or vehicle control (1% ethanol) for 2 h, prior to the addition of cytokines.The extract and stimulants were added to the differentiation media below the construct.
For whole-mount immunocytochemistry, constructs were incubated for 24 h, then fixed with ice-cold methanol for 5 min, and processed immediately.For the dye exclusion assay, constructs incubated with cytokines for 24 h were rinsed with PBS, then processed immediately.For chemokine and viability analyses, the conditioned medium was collected by centrifugation after 24 h of incubation, then stored at −80 • C. For Western blot analysis, constructs were incubated for 24 h, rinsed with PBS, scrapped, and then collected by centrifugation.Proteins were extracted using sample buffer and processed immediately.

Cytotoxicity Assays
A crystal violet assay was used to measure the biomass of HaCaT cell monolayers.The medium was removed, then each well was washed with PBS and the cells fixed with ice-cold methanol for 5 min.Then, cells were stained with crystal violet solution (0.25% (v/v), Sigma) for 10 min at room temperature and rinsed with PBS, and the crystals were solubilised with 33% (v/v) glacial acetic acid (Merck).Absorbance was read at 595 nm, with percentages of viable cells calculated relative to those of the vehicle-only control.
A CyQuant TM kit (Thermo Fisher Scientific) was used to measure LDH levels in the conditioned media from 3D constructs according to the manufacturer's instructions.Briefly, sample medium (50 µL) was transferred to a 96-well plate, followed by 50 µL reaction mixture (assay buffer and substrate stock), with incubation for 30 min in the dark.The LDH positive control was also measured.A stop solution (50 µL) was added to all reactions and absorbance was measured at 490 nm and 680 nm (background).The release of LDH was calculated with the formula (E − C)/(T − C) × 100, where E is the absorbance of the conditioned media, C is the absorbance of differentiation media used as the control, and T is the absorbance of the LDH positive control.The percentage of LDH release was normalised to that of cytokine-stimulated cells.

ELISAs
MCP-1 abundance within the conditioned medium from HaCaT cell monolayers and 3D constructs was quantified using a BD-OptEIA™ ELISA kit (San Diego, CA, USA, #2624KI) in accordance with the manufacturer's instructions.Chemokine abundance was normalised to that of cytokine-stimulated monolayers or constructs.

Lucifer Yellow Dye Exclusion Assay
An exclusion assay using lucifer yellow dye was used to determine epidermal barrier function following the methods described by [55].Lucifer yellow (100 µg•mL −1 , 100 µL, Thermo Fisher Scientific) was administered to the apical side of the cell construct, with PBS (400 µL) added to the basal side.At various time points, half of the PBS (200 µL) was withdrawn from the basal region and replaced with an equivalent volume of PBS.Once the samples had been collected, their relative fluorescence units (RFU) were detected (excitation 485 nm; emission 538 nm) using the SpectraMax Gemini fluorescence plate reader (Molecular Devices, San Jose, CA, USA).The percentage of lucifer yellow leakage was calculated from the RFU of the basal solution relative to that of the RFU positive control (amount of lucifer yellow added to each construct).

GC-MS
The fatty acid methyl esters (FAMEs) were analysed using a Thermo Fisher Scientific TRACE 1300 GC equipped with an ISQ 7000 single quadrupole MS equipped with a Restek Rxi ® -5Sil MS column.The parameters used for the analysis of FAME were as previously described [44].The obtained mass spectra were further evaluated by employing the NIST database (MS Search; NIST, MSS Ltd., Manchester, England).The proportion of each fatty acid is reported as a percentage of total integrated fatty acid peak areas for each chromatogram.

Figure 1 .
Figure 1.Macrocystis pyrifera lipid pre-treatment reduces MCP-1 production in TNF-α-stimulated HaCaT cell monolayers to a similar extent as hydrocortisone.Cells were incubated for 24 h with (a) M. pyrifera extract or (c) hydrocortisone (HC), with biomass assessed by crystal violet staining.Values are expressed relative to the vehicle-only control (0 mg•mL −1 ).Cells were stimulated with TNFα (20 ng•mL −1 ) with or without (b) pre-, co-, or post-treatment with extract (0.13 mg•mL −1 ) or (d) 2 h pre-treatment with HC (10 mg•mL −1 ).After 24 h incubation, conditioned medium was collected and assayed for MCP-1 production by ELISA.(−) represents unstimulated cells.(+) represents cells stimulated with TNF-α.Values are expressed as percentages of the (a,c) vehicle or (b,d) (+) controls and presented as means ± SD (n = 3), with those that differ significantly from those controls identified by one-way ANOVA followed by a Dunnett's test (* p ≤ 0.05).

Figure 1 .
Figure 1.Macrocystis pyrifera lipid pre-treatment reduces MCP-1 production in TNF-α-stimulated HaCaT cell monolayers to a similar extent as hydrocortisone.Cells were incubated for 24 h with (a) M. pyrifera extract or (c) hydrocortisone (HC), with biomass assessed by crystal violet staining.Values are expressed relative to the vehicle-only control (0 mg•mL −1 ).Cells were stimulated with TNF-α (20 ng•mL −1 ) with or without (b) pre-, co-, or post-treatment with extract (0.13 mg•mL −1 ) or (d) 2 h pre-treatment with HC (10 mg•mL −1 ).After 24 h incubation, conditioned medium was collected and assayed for MCP-1 production by ELISA.(−) represents unstimulated cells.(+) represents cells stimulated with TNF-α.Values are expressed as percentages of the (a,c) vehicle or (b,d) (+) controls and presented as means ± SD (n = 3), with those that differ significantly from those controls identified by one-way ANOVA followed by a Dunnett's test (* p ≤ 0.05).

Figure 7 .
Figure 7. Schematic representation of the anti-inflammatory and barrier protective effects of M. pyr ifera extract on Th1 and Th2 cytokine-mediated signalling in human keratinocytes.This fatty acid rich extract reduced MCP-1 production from TNF-α-stimulated HaCaT cells through decrease translocation of TNFR1 and abundance of intracellular TRAF2.The extract also, in part, reduce trans-epidermal leakage and claudin-1 tight junction turnover, primarily mediated by TNF-α.Re arrows depict biological processes and cytokine signalling pathways in keratinocytes suppresse by the M. pyrifera lipid extract (arrow density indicates extent of suppression shown in this study IκB, inhibitor of NFκB; MLCK, myosin light chain kinase; TRADD, tumour necrosis factor recepto type 1-associated death domain.Created with BioRender.com.

Figure 7 .
Figure 7. Schematic representation of the anti-inflammatory and barrier protective effects of M. pyrifera extract on Th1 and Th2 cytokine-mediated signalling in human keratinocytes.This fatty acid-rich extract reduced MCP-1 production from TNF-α-stimulated HaCaT cells through decreased translocation of TNFR1 and abundance of intracellular TRAF2.The extract also, in part, reduced trans-epidermal leakage and claudin-1 tight junction turnover, primarily mediated by TNF-α.Red arrows depict biological processes and cytokine signalling pathways in keratinocytes suppressed by the M. pyrifera lipid extract (arrow density indicates extent of suppression shown in this study).IκB, inhibitor of NFκB; MLCK, myosin light chain kinase; TRADD, tumour necrosis factor receptor type 1-associated death domain.Created with BioRender.com.