Diesel Particulate Extract Accelerates Premature Skin Aging in Human Fibroblasts via Ceramide-1-Phosphate-Mediated Signaling Pathway

Both intrinsic (i.e., an individual’s body clock) and extrinsic factors (i.e., air pollutants and ultraviolet irradiation) accelerate premature aging. Epidemiological studies have shown a correlation between pollutant levels and aging skin symptoms. Diesel particle matter in particular leads to some diseases, including in the skin. Our recent study demonstrates that diesel particulate extract (DPE) increases apoptosis via increases in an anti-mitogenic/pro-apoptotic lipid mediator, ceramide in epidermal keratinocytes. Here, we investigated whether and how DPE accelerates premature skin aging using cultured normal human dermal fibroblasts (HDF). We first demonstrated that DPE increases cell senescence marker β-galactosidase activity in HDF. We then found increases in mRNA and protein levels, along with activity of matrix metalloprotease (MMP)-1 and MMP-3, which are associated with skin aging following DPE exposure. We confirmed increases in collagen degradation in HDF treated with DPE. Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) is activated by DPE and results in increased ceramide production by sphingomyelinase activation in HDF. We identified that ceramide-1-phosphate (C1P) (produced from ceramide by ceramide kinase activation) activates MMP-1 and MMP-3 through activation of arachidonate cascade, followed by STAT 1- and STAT 3-dependent transcriptional activation.


Introduction
Both intrinsic and extrinsic factors (i.e., air pollutants and ultraviolet irradiation) drive premature aging [1,2]. Aging is unpreventable in human organs, but premature aging can be slowed down by minimizing the effects of environmental factors on cells/tissues. Indeed, topical agents containing antioxidants have shown some efficacy in attenuating development of visible skin aging symptoms, such as excessive melanogenesis, melanin deposition, and wrinkle formation [3,4].
Diesel particle matter, produced from diesel engines and composed of a various mixture of volatile components, including aldehydes, benzene, polyaromatic hydrocarbons and their derivatives, are major components of air pollutants [2,5]. Polyaromatic hydrocarbons and their derivatives contribute to the development of some diseases, including skin diseases such as contact hypersensitivity and dermatitis, and to the impairment of epidermal keratinocyte function [6,7]. Diesel particle matter also impairs epidermal permeability barrier integrity [8].

DPE Increases Cell Senescence
We recently found that DPE induces apoptosis in keratinocyte [14]. In this current study, we investigated the effects of DPE exposure on skin aging. We first optimized concentrations of DPE which do not show cell toxicity. Since up to 20 µg/mL of DPE did not affect cell viability ( Figure 1A), concentrations of 0-20 µg/mL of DPE were used in this study. We next examined whether DPE accelerates cell senescence in HDF. βgalactosidase activity, a marker of cell senescence, was increased in HDF following DPE exposure ( Figure 1B). These results suggest that DPE accelerates cell senescence in HDF.

DPE Activates MMP-1 and MMP-3
Next we investigated levels of MMP-1 and MMP-3, which are involved in skin aging development by alteration of the extracellular matrix structure [16], in HDF. mRNA expression of MMP-1 and MMP-3 were significantly increased in HDF incubated with DPE in a dose-dependent fashion (Figure 2A,B). We then found dose-dependent significant increases in MMP-1 and MMP-3 protein levels ( Figure 2C) and their activities ( Figure 2D,E) in cultured medium. These results suggest that DPE could accelerate skin aging by activation of MMP-1 and MMP-3.

Figure 1.
Cell viability and β-galactosidase activity. HDF were incubated with DPE for 24 h. Cell viability was measured by water-soluble tetrazolium salt cell quantification method (A). Cells positive for β-galactosidase activity were stained with blue-green (B). All values are mean ± SD (n = 3). Statistical significance was calculated using the unpaired Student's t-test, and significance was defined as * p < 0.01 vs. vehicle control. See details in the Materials and Methods section.

DPE Activates MMP-1 and MMP-3
Next we investigated levels of MMP-1 and MMP-3, which are involved in skin aging development by alteration of the extracellular matrix structure [16], in HDF. mRNA expression of MMP-1 and MMP-3 were significantly increased in HDF incubated with DPE in a dose-dependent fashion (Figure 2A,B). We then found dose-dependent significant increases in MMP-1 and MMP-3 protein levels ( Figure 2C) and their activities ( Figure  2D,E) in cultured medium. These results suggest that DPE could accelerate skin aging by activation of MMP-1 and MMP-3.

Figure 1.
Cell viability and β-galactosidase activity. HDF were incubated with DPE for 24 h. Cell viability was measured by water-soluble tetrazolium salt cell quantification method (A). Cells positive for β-galactosidase activity were stained with blue-green (B). All values are mean ± SD (n = 3). Statistical significance was calculated using the unpaired Student's t-test, and significance was defined as * p < 0.01 vs. vehicle control. See details in Section 4.

DPE Activates NOX
NOX activity was significantly increased in HDF following DPE treatment in a dosedependent fashion (Figure 3).

DPE Activates NOX
NOX activity was significantly increased in HDF following DPE treatment in a dosedependent fashion (Figure 3).

DPE Activates NOX
NOX activity was significantly increased in HDF following DPE treatment in a dependent fashion (Figure 3).

Changes in Ceramide and Its Metabolite Profile
We next assessed ceramide and its metabolites (C1P and S1P) levels in HDF t with DPE. Ceramide levels are significantly increased in HDF following DPE ex ( Figure 4A), in parallel with activation of acidic and neutral SMase ( Figure 5A). Cer

Changes in Ceramide and Its Metabolite Profile
We next assessed ceramide and its metabolites (C1P and S1P) levels in HDF treated with DPE. Ceramide levels are significantly increased in HDF following DPE exposure ( Figure 4A), in parallel with activation of acidic and neutral SMase ( Figure 5A). Ceramide species containing shorter chains of amide-linked fatty acid (C16) amounts were significantly increased in HDF, while longer chain ceramide species were modestly decreased (C24≥) ( Figure 4B). Analysis of ceramide metabolites revealed that C1P containing C16 fatty acid levels are significantly increased ( Figure 4C,D), while S1P is not changed following DPE treatment ( Figure 4E). Ceramide kinase was significantly activated by DPE ( Figure 5B). Two isoforms of sphingosine kinase 1 and 2 activity were not changed ( Figure 5C,D). These results suggest that DPE increases ceramide production through SMase activation, followed by increased conversion from ceramide to C1P by ceramide kinase activation.

C1P Modulates MMP-1 and MMP-3 Activation
Because both ceramide and C1P are lipid mediators that modulate cellular functions [17,18], we next investigated whether ceramide and/or C1P is (are) responsible for activation of MMP-1 and MMP-3. First, inhibition of acidic and neural SMase by a specific pharmacological inhibitor, imipramine (for acidic SMase) and GW4869 (for neutral SMase), respectively, suppressed DPE-mediated upregulation of both MMP-1 and MMP-3 mRNA expression (Table 1) and protein production ( Figure 6). These results suggest that ceramide and/or its metabolite(s) increase(s) MMP-1 and MMP-3 expression.
Next, blocking of C1P synthesis by a specific inhibitor of ceramide kinase, NVP-231 reduced MMP-1 and MMP-3 mRNA expression (Table 1) and protein production ( Figure 6). Thus, increases in C1P rather than in ceramide stimulates MMP-1 and MMP-3 production. We further investigated whether activation of MMP-1 and MMP-3 increases collagen degradation. Collagen levels were decreased in HDF exposed to PDE ( Figure 7A,B). When SMase and ceramide kinase were inhibited, decreases in collagen were suppressed ( Figure 7A,B). These results appear to confirm that DPE-increased C1P simulates MMP-1 and MMP-3, leading to collagen hydrolysis. fatty acid levels are significantly increased ( Figure 4C,D), while S1P is not changed following DPE treatment ( Figure 4E). Ceramide kinase was significantly activated by DPE ( Figure 5B). Two isoforms of sphingosine kinase 1 and 2 activity were not changed ( Figure  5C,D). These results suggest that DPE increases ceramide production through SMase activation, followed by increased conversion from ceramide to C1P by ceramide kinase activation.
Inhibition of cyclooxygenase by acetylsalicylic acid (aspirin) diminished increases in MMP-1 and MMP-3 protein ( Figure 6) and decreased in collagen levels in HDF exposed to DPE (Figure 7). STAT1 and STAT3 phosphorylation were also evident in HDF exposed to DPE, while inhibition of sphingomyelinases and ceramide kinase, as well as cyclooxygenase-2, suppressed STAT1 and STAT3 phosphorylation (Figure 8). Thus, C1P stimulates MMP-1 and MMP-3 production through a cPLA2-mediated arachidonate pathway of STAT1 and STAT3 activation.

Discussion
We demonstrated here that DPE exposure activates MMP-1 and MMP-3, and decreases type 1 collagen levels in HDF. We identified that MMP-1 and MMP-3 activation was initiated by C1P through increases in ceramide production by activation of both acidic and neutral SMase, followed by increased conversion from ceramide to C1P by ceramide kinase activation. We further characterized that C1P-initiated activation of arachidonate pathway followed by STAT1 and STAT3 activation is a downstream mechanism of increased MMP-1 and MMP-3 synthesis (Figure 9) Prior studies have demonstrated that C1P modulates cellular functions; i.e., adipogenesis, cell migration, and autophagy [17,18]. We also have reported that endoplasmic reticulum (ER) stress-driven increases in C1P promote major epidermal antimicrobial peptide production (human β-defensin [hBD]2 and hBD3) in normal human keratnocyte [22]. Our current study reveals that DPE-increased C1P could accelerate premature skin aging. NOX (which lead to increasing in ceramide and then C1P production) are activated by not only DPE, but also other external and internal stressors, such as ultraviolet irradiation, cigarette smoke, inflammatory cytokines, and epidermal permeability barrier perturbation [9,11,23,24]. Therefore, C1P could play a critical role in skin aging in response to diverse oxidative stressors.
Blocking of sphingomyelinase to reduce C1P production suppresses C1P-mediated MMP-1 and MMP-3 production. Yet, because ceramide is needed to maintain normal epidermal functions, including epidermal permeability barrier formation, specific suppression of C1P rather than ceramide production should be an appropriate therapeutic target to suppress MMP-1 and MMP-3 activation. Inhibition of cPLA2 (which is activated by C1P) to block arachidonate cascade is a potent therapeutic approach to treat inflammatory diseases [25]. However, cPLA2 is activated by multiple pathways and it has not only pathological, but also physiological roles in cells [25,26]. Thus, the inhibition of cPLA2 could lead to unpredictable adverse consequences. Hence, specific modulation of C1P production rather than inhibition of cPLA2, i.e., suppression of C1P production and/or its activity, should be a safer approach to suppressing premature skin aging.
Our prior studies have shown that C16 ceramide is a major backbone constituent of sphingomyelin in skin [27,28]. We found that: (1) C16 ceramide and C16 C1P are significantly increased in HDF following DPE treatment in parallel with SMase activation, and (2) SMase inhibition decreases MMP-1 and MMP-3 activation. Therefore, ceramide derived sphingomyelin, not de novo ceramide synthesis or glycosylceramide (major backbone ceramide containing longer fatty acids) hydrolysis, is the source of ceramide that generates C1P for upregulation of MMP-1 and MMP-3.

Discussion
We demonstrated here that DPE exposure activates MMP-1 and MMP-3, and decreases type 1 collagen levels in HDF. We identified that MMP-1 and MMP-3 activation was initiated by C1P through increases in ceramide production by activation of both acidic and neutral SMase, followed by increased conversion from ceramide to C1P by ceramide kinase activation. We further characterized that C1P-initiated activation of arachidonate pathway followed by STAT1 and STAT3 activation is a downstream mechanism of increased MMP-1 and MMP-3 synthesis (Figure 9). Longer chain ceramide species are decreased in HDF following DPE exposure. DPE may affect either/both synthesis and/or activity of ceramide synthase(s) (CerS2 and CerS3, which are synthesized longer chain ceramides) and/or ELOVL1, which synthesizes long chain fatty acids (C24) [29], and results in decreases in longer chain ceramide. Prostaglandin E2 promotes INF-γ production [30]. INF-γ suppresses CerS3 expression [31]. Therefore, DPE could change Cer2, CerS3 and/or ELOVL1 expression in cells, and results in decreases in longer chain ceramide. S1P is another ceramide metabolite that regulates diverse cell functions, such as cell proliferation, differentiation, apoptosis, angiogenesis and cell motility [32][33][34]. However, S1P levels were decreased in HDF following DPE exposure. Our prior studies showed that Prior studies have demonstrated that C1P modulates cellular functions; i.e., adipogenesis, cell migration, and autophagy [17,18]. We also have reported that endoplasmic reticulum (ER) stress-driven increases in C1P promote major epidermal antimicrobial peptide production (human β-defensin [hBD]2 and hBD3) in normal human keratnocyte [22]. Our current study reveals that DPE-increased C1P could accelerate premature skin aging. NOX (which lead to increasing in ceramide and then C1P production) are activated by not only DPE, but also other external and internal stressors, such as ultraviolet irradiation, cigarette smoke, inflammatory cytokines, and epidermal permeability barrier perturbation [9,11,23,24]. Therefore, C1P could play a critical role in skin aging in response to diverse oxidative stressors.
Blocking of sphingomyelinase to reduce C1P production suppresses C1P-mediated MMP-1 and MMP-3 production. Yet, because ceramide is needed to maintain normal epidermal functions, including epidermal permeability barrier formation, specific suppression of C1P rather than ceramide production should be an appropriate therapeutic target to suppress MMP-1 and MMP-3 activation. Inhibition of cPLA2 (which is activated by C1P) to block arachidonate cascade is a potent therapeutic approach to treat inflammatory diseases [25]. However, cPLA2 is activated by multiple pathways and it has not only pathological, but also physiological roles in cells [25,26]. Thus, the inhibition of cPLA2 could lead to unpredictable adverse consequences. Hence, specific modulation of C1P production rather than inhibition of cPLA2, i.e., suppression of C1P production and/or its activity, should be a safer approach to suppressing premature skin aging.
Our prior studies have shown that C16 ceramide is a major backbone constituent of sphingomyelin in skin [27,28]. We found that: (1) C16 ceramide and C16 C1P are significantly increased in HDF following DPE treatment in parallel with SMase activation, and (2) SMase inhibition decreases MMP-1 and MMP-3 activation. Therefore, ceramide derived sphingomyelin, not de novo ceramide synthesis or glycosylceramide (major backbone ceramide containing longer fatty acids) hydrolysis, is the source of ceramide that generates C1P for upregulation of MMP-1 and MMP-3.
Longer chain ceramide species are decreased in HDF following DPE exposure. DPE may affect either/both synthesis and/or activity of ceramide synthase(s) (CerS2 and CerS3, which are synthesized longer chain ceramides) and/or ELOVL1, which synthesizes long chain fatty acids (C24) [29], and results in decreases in longer chain ceramide. Prostaglandin E2 promotes INF-γ production [30]. INF-γ suppresses CerS3 expression [31]. Therefore, DPE could change Cer2, CerS3 and/or ELOVL1 expression in cells, and results in decreases in longer chain ceramide. S1P is another ceramide metabolite that regulates diverse cell functions, such as cell proliferation, differentiation, apoptosis, angiogenesis and cell motility [32][33][34]. However, S1P levels were decreased in HDF following DPE exposure. Our prior studies showed that ceramide levels were increased, but S1P levels decreased in normal human keratinocyte following DPE treatment [14]. We previously found that both acidic and neutral ceramidase activities are decreased in normal human keratinocyte after UVB irradiation [35]. These results suggest that, under oxidative stress, hydrolysis of ceramide by ceramidase to sphingosine followed by S1P production by sphingosine kinase is likely attenuated. Thus, ceramide could be preferentially metabolized to C1P. It is currently unknown whether and why ceramidase is sensitive to oxidative stress.
Chronic exposure of low levels of DPE may increase in ceramide and its metabolites that are at sub-apoptotic levels, affecting cell proliferation and differentiation. Yet, the mechanism that is responsible for thinner epidermis is still not characterized very well. In addition, the effect of C1P on keratinocyte proliferation and differentiation is unknown. Thus, it is not clear whether C1P contributes to epidermal thinning and regeneration in response to air pollutants and other external stressors. However, we here characterized that C1P activates an arachidonate pathway, leading to stimulation of MMP-1 and MMP-3 mediated collagen hydrolysis. Activation of this arachidonate pathway promotes inflammatory cytokine/chemokine production. Increased secretion of certain inflammatory cytokine/chemokines leads to a senescence-associated secretory phenotype (SASP). C1P may be a driver of SASP.
Penetration of pollutants into skin is not completely elucidated yet either, but a recent study showed that topical diesel engine exhaust penetrated into the dermis and activated MMP-9 in ex vivo cultured human skin [36]. Thus, MMP-1 and MMP-3 could be activated in skin following exposure to DPE.
In conclusion, C1P should be a driver to accelerate premature skin aging through activation of MMP-1 and MMP-3, leading to collagen degradation in dermal fibroblasts following DPE exposure.

Cell Viability
Cell viability was measured by the water-soluble tetrazolium salt (WST) method using the cell counting kit-8 (CCK-8) assay kit (Dojindo, Kumamoto, Japan), in accordance with the manufacturer's instructions.

MMP-1 and MMP-3
MMP-1 and MMP-3 activity were assessed by the fluorometric-based method using the MMP-1 or MMP-3 activity assay kit (abcam, Cambridge, MA, USA) according to the manufacturer's instructions.

NADPH Oxidases
Activity of NADPH oxidase (NOX) was measured by the lucigenin chemiluminescence assay kit using N,N -Dimethyl-9,9 -bicridium dinitrate (Sigma-Aldrich), in accordance with the manufacturer's instructions. NADPH oxidases activity was measured by the ratio of NADP+ to NADPH using LC-ESI-MS/MS (API 3200 QTRAP mass, AB/SCIEX, Framingham, MA, USA) by multiple reaction monitoring mode, as we described previously [14].

Sphingomyelinase
Activities of acidic or neutral sphingomyelinase were measured as we described previously [14]. Briefly, cells suspended in assay buffers (acidic sphingomyelinase: 250 mM sodium-acetate, 0.2% Triton X-100, pH 4.5, and neutral sphingomyelinase: 20 mM HEPES, 0.2% Triton X-100, pH 7.4) were incubated with 5 nmol of C12-sphingomyeline for 20 min at 37 • C. The reaction was stopped by the addition of CHCl 3 :CH 3 OH (2:1, v/v). The organic phases were dried and were resolved in MeOH, and then analyzed by LC-MS/MS. The activities of both sphingomyelinases are expressed as pmol (C12-ceramide) per mg protein per min.

Ceramide Kinase
Ceramide kinase activity was determined as we described previously [39] with modification. Briefly, cell lysates were incubated with 20 µM of C8-ceramide in 20 mM MOPS, pH 7.4, 137 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl 2 , 0.34 mM Na 2 HPO 4 , 5.6 mM glucose, 0.44 mM KH 2 PO 4 , 4.2 mM NaHCO 3 , 1.9 mM CaCl 2 , and 0.1% BSA with 1 mM ATP at 37 • C for 30 min. Enzyme reactions were terminated by the addition of cold MeOH containing 30 pmoles of d17:1/C18:0 ceramide as the internal standard and the activity was quantified by LC-MS/MS. The activity of ceramide kinase is expressed as pmol (C8-ceramide) per mg protein per min.

Sphingosine Kinase
Sphingosine kinase 1 and Sphingosine kinase 2 activities were assessed, as we described previously [14,40]. Briefly, cell lysates were incubated with 200 µM C17-Sphingosine as a substrate. To assay each isoform of sphingosine kinase activity, 0.5% Triton X-100 or 1 M KCl for sphingosine kinase 1 and sphingosine kinase 2, respectively, were added into assay buffer and then incubated at 37 • C for 30 min. Enzyme reactions were terminated by the addition of CHCl 3 :MeOH:HCl (8:4:3, v/v/v). C17-dihydrosphingosine-1-phosphate (100 pmol) was added as an internal standard. The organic phage was separated by addition of CHCl 3 , dried and redissolved in MeOH, and then analyzed by LC-MS/MS. The activity of Sphingosine kinase was expressed as C17-S1P pmol per mg protein per min.

Detection of Cellular Reactive Oxygen Species
Production of cellular reactive oxygen species (ROS), including superoxide (O 2 − ) and hydrogen peroxide (H 2 O 2 ) was detected by the oxidant-sensing probe 2,7dichlorodihydrofluorescein diacetate (DCFH-DA) (abcam), as described previously [38]. ROS production was analyzed using a fluorescence microscopy (Eclipse Ti-U; Nikon Corporation, Tokyo, Japan) and fluorospectrophotometer (Molecular devices M2e, Molecular Devices, Sunnyvale, CA, USA) with 485 nm of excitation and 520 nm of emission filters and was expressed as a fluorescence intensity (a.u.).