Lipin-1 expression is critical for keratinocyte differentiation[S]

Lipin-1 is an Mg2+-dependent phosphatidate phosphatase that facilitates the dephosphorylation of phosphatidic acid to generate diacylglycerol. Little is known about the expression and function of lipin-1 in normal human epidermal keratinocytes (NHEKs). Here, we demonstrate that lipin-1 is present in basal and spinous layers of the normal human epidermis, and lipin-1 expression is gradually downregulated during NHEK differentiation. Interestingly, lipin-1 knockdown (KD) inhibited keratinocyte differentiation and caused G1 arrest by upregulating p21 expression. Cell cycle arrest by p21 is required for commitment of keratinocytes to differentiation, but must be downregulated for the progress of keratinocyte differentiation. Therefore, reduced keratinocyte differentiation results from sustained upregulation of p21 by lipin-1 KD. Lipin-1 KD also decreased the phosphorylation/activation of protein kinase C (PKC)α, whereas lipin-1 overexpression increased PKCα phosphorylation. Treatment with PKCα inhibitors, like lipin-1 KD, stimulated p21 expression, while lipin-1 overexpression reduced p21 expression, implicating PKCα in lipin-1-induced regulation of p21 expression. Taken together, these results suggest that lipin-1-mediated downregulation of p21 is critical for the progress of keratinocyte differentiation after the initial commitment of keratinocytes to differentiation induced by p21, and that PKCα is involved in p21 expression regulation by lipin-1.

The cyclin-dependent kinase (CDK) inhibitor, p21, suppresses the CDKs, cyclin E/CDK2, and cyclin D/CDK4, thus inhibiting phosphorylation of retinoblastoma protein to control cell cycle arrest, cell differentiation, or senescence (23)(24)(25)(26)(27). The p21 controls the transition of cells from G1 to S phase of the cell cycle by inhibiting cyclin activity at the G1 checkpoint ( 28 ). Induction of p21 in early differentiation is necessary to initiate keratinocyte differentiation ( 29,30 ). However, because expression of terminal differentiation markers is attenuated by sustained overexpression of p21 independent of the cell cycle, p21 expression must be downregulated at later stages of differentiation ( 31 ). Therefore, tight regulation of p21 expression is critical for the process of keratinocyte differentiation ( 32,33 ).
In this study, we investigated the expression and function of lipin-1 in normal human epidermal keratinocytes (NHEKs). We show that lipin-1 is present in basal and spinous layers of normal human epidermis, and that lipin-1 expression is gradually downregulated during NHEK differentiation. We also demonstrate that lipin-1 knockdown (KD) causes G1 arrest associated with p21 upregulation, which is known to be necessary for initial commitment of keratinocytes to differentiation. However, interestingly, lipin-1 KD suppressed keratinocyte differentiation. We demonstrate that the reduced keratinocyte differentiation was due to the sustained expression of p21 driven by lipin-1 KD, mediated by inhibition of PKC ␣ activation.

Cell culture
NHEKs from neonatal foreskin were purchased from Lonza (Basel, Switzerland) and cultured in KBM-GOLD medium with KGM-GOLD growth supplements containing insulin, human epidermal growth factor, bovine pituitary extract, hydrocortisone, epinephrine, transferrin, and gentamicin/amphotericin B. The cells were serially passaged at 70-80% confl uence no more than three times.

Immunostaining of tissue section
Normal human skin purchased from Biochain (Hayward, CA) was incubated with polyclonal rabbit anti-lipin-1 antibody (Sigma, St. Louis, MO) for 40 min at room temperature. Anti-rabbit HRPconjugated antibody (ab6802; Abcam Inc., Cambridge, MA) was then applied as the secondary antibody for 20 min. The secondary antibody was detected with using chromogenic substrate 3,3 ′ -diaminobenzidine. Immunoreactivity was evaluated under a light microscope (BX-53; Olympus, Tokyo, Japan) and photomicrographs were taken using a digital camera (DP72; Olympus).

Plasmid construction
The human full-length lipin-1 cDNA (isoform 1 ␣ , accession number NM_145963) cloned into the pCMV6-AC-green fl uorescent protein (GFP) vector was obtained from Origene (Rockville, MD). To generate lipin-1 ␤ isoform (accession number NM_001261428), lipin-1 fragment 1 (F1) was amplifi ed by PCR calcium KGM-Gold medium up to 90% confl uence (day 0). To study the regulation of lipin-1 protein expression by keratinocyte differentiation, NHEKs were cultured for an additional 7 days (day 7) under either low calcium (50 M) or high calcium (1.2 mM) conditions. Western blot analysis showed that lipin-1 was strongly downregulated regardless of calcium concentration ( Fig. 1A, B ), corresponding to an increase in expression of involucrin, a marker for differentiation ( 29,31 ). A previous report demonstrated that PKC activation mediates human keratinocyte differentiation at high cell densities, independent of changes in extracellular calcium ( 36 ), which may be associated with similar patterns of lipin-1 expression in low and high calcium conditions. The data indicate that lipin-1 expression is gradually decreased during keratinocyte differentiation. To assess the expression pattern and localization of lipin-1 in human epidermis, immunostaining was performed for lipin-1 in human foreskin sections. Lipin-1 was largely detected in both the basal layer and the spinous layer ( Fig. 1C ). Lipin-1 expression appeared to be purplish brown as a result of brown 3,3 ′ -diaminobenzidine chromogen with blue hematoxylin. These data suggest that lipin-1 may play a role in keratinocyte proliferation or early differentiation.

Lipin-1 regulates cell cycle progression by regulating p53/p21
Because lipin-1 is predominantly expressed in proliferating keratinocytes, its effect on proliferation was examined by cell cycle analysis using propidium iodide staining. After transfection with control siRNA or lipin-1 siRNA, NHEKs were subjected to serum deprivation for 16 h and DNA content was analyzed by fl uorescence-activated cell sorting. The DNA histogram revealed that the proportion of G1 phase was increased in a serum-free medium compared with control (43.03 ± 0.34% vs. 56.07 ± 0.24%) (see supplementary Fig. 1). Fluorescence-activated cell sorting analysis revealed that lipin-1 siRNA induced G1 cell-cycle arrest (56 ± 0.38%) 24 h after transfection compared with control siRNA (50.77 ± 0.57%; Fig. 2A ). Analysis of proteins regulating the progression through the G1 phase of the cell cycle showed a strong increase in the levels of p21 and p53, but not p16 ( Fig. 2C ). Activation of p21 signaling is an important mechanism to prevent G1 cells from entering the S phase. The p21 is activated by p53-dependent or p53-independent mechanisms ( 37 ). Thus, we predict that lipin-1 KD results in an increased p53 level, which in turn induces its downstream target p21 expression and G1 arrest in NHEKs. Taken together, lipin-1 may be required for the progression of G1 phase in the cell cycle through regulating p53/p21 expression.

DAG content and PKC ␣ activity are regulated by lipin-1 in NHEKs
To confi rm DAG formation via dephosphorylation of PA by lipin-1 in keratinocytes, like in hepatocytes and adipocytes ( 1,8,9 ), PA and DAG contents were assayed in NHEKs transfected with lipin-1 or control siRNA. As expected, lipin-1 KD resulted in elevated PA and reduced (#D-001810-10) were purchased from Dharmacon (Lafayette, CO). NHEKs were plated 24 h before transfection and then transfected by lipofection using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA) and OPTI-MEM (Invitrogen) with siRNA at a fi nal concentration of 25 nM for 6 h. The medium was then changed to KGM-Gold containing all appropriate supplements. For lipin-1 overexpression, NHEKs were seeded 24 h before transfection and then transfected with 1 g/ml of plasmid using the X-tremeGENE HP DNA transfection reagent (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions. We used a 1.5:1 ratio of microliters of X-tremeGENE HP DNA transfection reagent to micrograms of pCMV6-AC-GFP vector as a transfection control.

Cell cycle analysis
After cells were transfected with siRNAs, complete medium was replaced with serum-free medium for 16 h. The cells were centrifuged, washed in PBS, and fi xed in 70% ethanol overnight at Ϫ 20°C. After washing twice in PBS, the cells were stained in 0.5 ml of propidium iodide/RNase solution (BD Bioscience, San Jose, CA) for 15 min. Cell cycle distribution was analyzed using FACS Calibur (BD Bioscience). Ten thousand events were counted during data collection. The percentage of cells in G1, S, G2/M phase was determined using ModFit LT curve fi tting software (Verity Software, Topsham, ME).

Lipid analysis
Lipids were extracted by a modifi cation of the Bligh and Dyer method ( 34 ). Briefl y, keratinocyte cells were collected by centrifugation at 1,500 g for 5 min and suspended in 0.5 ml of cold PBS buffer (pH 7.4) followed by sonication. Pellet was extracted with 1.5 ml of methanol, 2.25 ml of 1 M of sodium chloride, and 2.5 ml of chloroform, and the phase was separated by centrifugation at 1,500 g for 5 min. The lower phase was dried and redissolved in 500 l of 1% Triton X-100 for PA analysis or 100 l of chloroform for DAG analysis. PA was measured with the Total Phosphatidic Acid Fluorometric Assay kit (Cayman Chemical, Ann Arbor, MI) ( 10 ). For the DAG quantifi cation, extracted lipids in chloroform were spotted on silica 60 TLC plates (Sigma) ( 35 ). DAG was migrated with diethylether/heptane/acetic acid (75:25:1 v:v:v) mixture, the plates were dried and stained with 0.003% Coomassie brilliant blue in 30% methanol and 10 mM sodium chloride for 30 min. The plates were destained for 5 min in dye-free solution and the band density was calculated using ImageQuant TL software (GE Healthcare Life Sciences).The amount of DAG was calculated using a standard curve of 1,2-dioleyl-sn -diacylglycerol (Avanti Polar Lipids, Alabaster, AL)

Statistical analysis
Statistical comparisons were performed using Student's t -test between two groups or one-way ANOVA test within multiple groups, followed by Turkey's post hoc test. All measurements were obtained from at least three independent experiments and values are expressed as the mean ± SD.

Lipin-1 is downregulated during epidermal differentiation
To understand the physiological role of lipin-1 in keratinocytes, the lipin-1 protein level was measured during proliferation and differentiation. NHEKs were cultured in low cells, but the increase was less pronounced compared with Go6976 treatment. These data suggest that PKC ␣ inhibition is associated with the stimulation of p21 expression in NHEKs. Therefore, PKC ␣ inactivation by lipin-1 siRNA may be associated with elevated p21 in NHEKs. To further confi rm the effect of lipin-1 on p21 expression, NHEKs transfected with lipin-1 vector or pCMV-AC empty vector were analyzed by Western blot. Lipin-1 overexpression induced PKC ␣ phosphorylation/activation ( Fig. 4E ) and reduced p21 expression in both Go6976-treated and untreated cells ( Fig. 5C ). Collectively, these results suggest that lipin-1-induced PKC ␣ activation inhibits p21 expression in NHEKs.

Lipin-1 downregulation inhibits keratinocyte differentiation by sustaining p21 induction
We investigated whether the cell cycle arrest linked to upregulation of p21 by lipin-1 KD affects keratinocyte differentiation. NHEKs were transfected with lipin-1 or control siRNA for 24 h (day 1) and differentiation marker expression was assessed 4 days posttransfection by Western blot. Inhibition of lipin-1 expression was observed in lipin-1 KD cells on day 4, indicating that lipin-1 siRNA was effective at least 4 days after transfection. Surprisingly, the differentiation markers, keratin 1 and keratin 10, were reduced in lipin-1 KD cells compared with control ( Fig. 6A ). Previous studies demonstrated that elevated expression of p21 is observed at the induction stage of differentiation, while sustained upregulation of p21 expression inhibits keratinocyte differentiation. Therefore, elevated p21 expression must be downregulated at a later differentiation stage ( 31 ). Thus, we examined whether lipin-1 KD affects tight regulation of p21 during early differentiation (days 0, 1, 2, 3, and 4). On the transfection day (day 0), cell confl uency was approximately 60% to conduct siRNA transfection. Lipin-1 expression was increased (day 2) before reaching confl uence, but suppressed when the cells began to differentiate ( Fig. 6C ). The data refl ect lipin-1 expression DAG levels compared with control ( Fig. 3A ). The data indicated that lipin-1 is involved in DAG formation in NHEKs (supplementary Fig. 2). PKC activation by DAG controls the proliferation of keratinocytes ( 38,39 ). To assess whether altered DAG contents led to impaired PKC activation, we assessed the phosphorylation of PKC substrates. Phosphorylation of PKC substrates and MARCKS, a major PKC substrate, was inhibited by lipin-1 KD ( Fig. 3B-E ). Phosphorylation of fi ve PKC isoforms expressed in keratinocytes ( 11 ) was also analyzed by Western blot. Most PKC isoforms exhibited no signifi cant change in phosphorylation levels in lipin-1 KD cells compared with control, but PKC ␣ phosphorylation at Ser657 was inhibited by lipin-1 KD ( Fig. 4A ). To confi rm the inhibition of PKC ␣ activation, we demonstrated that lipin-1 KD decreased phosphorylation at another PKC ␣ autophosphorylation site (Thr638) ( Fig. 4C ). We transfected vectors containing lipin-1 cDNA or pCMV-AC empty vector as a negative control to assess the effect of exogenous expression of lipin-1. Western blot analysis indicated that overexpression of lipin-1 enhances PKC ␣ phosphorylation ( Fig. 4E ). These results indicate that lipin-1-catalyzed DAG formation is associated with the activation of PKC, specifi cally PKC ␣ , in NHEKs.

Lipin-1-induced PKC ␣ activation inhibits p21 expression
It has been reported that PKC downregulation-mediated growth arrest is associated with p21 induction in fi broblasts ( 40 ). Thus, to confi rm that inhibition of PKC ␣ is associated with upregulation of p21 expression in NHEKs, cells transfected with lipin-1 or control siRNA were treated with Go6976 (PKC ␣ / ␤ inhibitor) or safi ngol (PKC ␣ inhibitor/ sphingosine kinase 1 inhibitor). Go6976 treatment stimulated p21 expression in both lipin-1 siRNA and control siRNA transfected cells ( Fig. 5A ). Although safi ngol lacks specifi city to PKC ␣ due to its activity of sphingosine kinase 1 inhibitor ( 41 ), safi ngol treatment also increased p21 expression in both lipin-1 siRNA and control siRNA transfected  2 mM) (B) were subjected to immunoblot analysis to assess lipin-1 expression levels. Involucrin was also analyzed as a differentiation marker. C: Immunoperoxidase staining of normal human foreskin was performed to visualize lipin-1 localization. Negative control was done by using normal rabbit IgG. Scale bar is 40 m. These data are representative of three independent experiments.
To verify the effect of lipin-1 KD-induced p21 expression on keratinocyte differentiation, NHEKs were knocked down for p21, lipin-1, or both genes and incubated for 4 days. The p21 KD showed more pronounced inhibition of keratin 1 and keratin 10 expression compared with lipin-1 KD ( Fig. 6E ). A previous study demonstrated that p21-null keratinocytes show a lower expression level of keratin 1 compared with wild-type keratinocytes ( 29 ). While lipin-1 KD inhibited the expression level of keratin 1 and keratin 10, as shown in previous data ( Fig. 6A-D ), simultaneous KD of both p21 and lipin-1 restored not only the levels of p21, but also the levels of keratin 1 and keratin 10, to almost peaks when the cell reaches full confl uence and is gradually decreased by cell differentiation, as observed in Fig. 1A . Our data showed that p21 expression was induced on day 1, maintained until day 3, and then downregulated on day 4 in both control and lipin-1 KD cells. Our data also showed that the expression of p21 was sustained at a higher level in lipin-1 KD cells compared with control. In addition, keratin 1 and keratin 10 began to express later in lipin-1 KD cells compared with control. These results imply that sustained p21 upregulation resulting from lipin-1 KD may lead to the inhibition of the differentiation markers, keratin 1 and keratin 10 . Lipin-1 KD induces G1 arrest via p53/p21 induction. A, B: NHEKs were transfected with lipin-1 siRNA or control siRNA, and cell cycle distributions were analyzed at 24 h posttransfection following serum starvation for 16 h using fl ow cytometry. Curve fi tting analysis was performed to compare the distribution of lipin-1 KD or control NHEKs in G1, S, and G2/M phases. C, D: NHEK cell lysates prepared after lipin-1 siRNA or control siRNA transfection for 24 h were subjected to Western blot analysis for p53, p21, and p16. Protein levels of p53, p21, and p16 were normalized to GAPDH. All data (mean ± SD) represent three independent experiments. * P < 0.05, ** P < 0.01, *** P < 0.001 versus control. phospholipids phosphatidylcholine, phosphatidylethanolamine, and phosphatidylserine ( 9,42 ). In addition, both substrates and products have functions in lipid signaling. PA is involved in transcription, activation of cell growth, membrane proliferation, secretion, and vesicular traffi cking, while DAG is known to be involved in the activation of PKC ( 9 ). In keratinocytes, TAG is a minor lipid species in lamellar bodies ( 43,44 ) and thus, epidermal TAG metabolism has not been focused in the past ( 45 ). However, DAG acylransferase-2-defi cient mice lacking the enzyme catalyzing the same level as control, indicating that lipin-1 KD-induced p21 contributes to the inhibition of keratin 1 and keratin 10 expression.

DISCUSSION
Lipin-1 catalyzes the formation of DAG by dephosphorylating PA. The DAG generated from the lipin-1-mediated reaction is utilized for the formation of TAG, as well as the Fig. 3. Lipin-1 KD inhibits DAG production and phosphorylation of PKC substrates. A: After transfection of lipin-1 siRNA or control siRNA in NHEKs for 24 h, total lipids were extracted from the cells and analyzed biochemically to quantify PA and DAG. B, C: Cell lysates after transfection of lipin-1 siRNA or control siRNA in NHEKs for 24 h were immunoblotted to assess phosphorylation of PKC substrates using anti-phosphoserine PKC substrate antibody and the level of phosphorylation of PKC substrates normalized to GAPDH was expressed as mean ± SD for three independent experiments. D, E: MARCKS was also detected using antiphospho MARCKS and anti-MARCKs antibodies and the level of phospho-MARCKS normalized to MARCKS was expressed as mean ± SD for three independent experiments. * P < 0.05, ** P < 0.001 versus control. decreases during differentiation of NHEKs ( Fig. 1A, B ). Immunostaining of skin tissue showed that lipin-l localizes in basal and spinous layers of human epidermis ( Fig. 1C ). The localization of lipin-1 was somewhat surprising because TAG synthesis reportedly increases during keratinocyte differentiation ( 47 ). These results suggest that lipin-1 functions in proliferating NHEKs rather than increasing TAG synthesis during differentiation. In this study, we show that PKC ␣ activation is regulated by lipin-1 in NHEKs. PKC ␣ is expressed and activated in the fi rst basal or lower spinous layers in human epidermis ( 39 ). These results suggest that lipin-1 may be involved in PKC ␣ activation in the basal and spinous layers the fi nal reaction in TAG synthesis die shortly after birth due to abnormal energy metabolism and impaired skin permeability barrier function ( 46 ). Mice defi cient in comparative gene identifi cation-58 (CGI-58), a lipid dropletassociated protein that controls intracellular TAG levels by activating adipose triglyceride lipase (ATGL), develop a severe permeability barrier defect ( 43 ). These reports demonstrate that effi cient inter-conversion between TAG and DAG is required for permeability barrier homeostasis in the epidermis. Given the critical role of lipin-1 in formation of DAG and subsequent synthesis of TAG and signaling lipids, in this study, we investigated the role of lipin-1 in the epidermis. We found that lipin-1 expression gradually Fig. 4. Lipin-1 regulates PKC ␣ activity. A, B: Cell lysates after transfection of lipin-1 siRNA or control siRNA in NHEKs for 24 h were immunoblotted to assess phosphorylation levels of PKC isoforms with the indicated antibodies to PKC isoforms. Protein levels of PKC isoforms normalized to GAPDH were expressed as mean ± SD for three independent experiments. C, D: The cell lysates were also detected by phospho (P)-PKC ␣ (Thr638) or PKC ␣ antibodies. The level of phospho-PKC ␣ (Thr638) normalized to PKC ␣ were expressed as mean ± SD for three independent experiments. E, F: NHEK cell lysates transfected overnight with vector containing lipin-1 cDNA or pCMV-AC empty vector were subjected to Western blot analysis for phospho-PKC ␣ (Ser657) and total PKC ␣ . The levels of phospho-PKC ␣ (Ser657) normalized to PKC ␣ were expressed as mean ± SD for three independent experiments. * P < 0.05, ** P < 0.01 versus control. cell types. For example, activation of the PKC -Vps34 signaling cascade is dependent on lipin-1 in skeletal muscle ( 10 ). In addition, PKC activity requires lipin-1 in liver ( 48 ). We observed that DAG formation and PKC ␣ activation were inhibited by lipin-1 KD in NHEKs ( Fig. 3 ). PKC activation with 12-O -tetradecanoylphorbol-13-acetate (TPA), a PKC agonist, triggers an irreversible growth arrest, to control proliferation or induce early differentiation of NHEKs.
Because lipin-1 KD caused G1 arrest linked to p21 induction, the relationship between lipin-1 and p21 was investigated by analyzing DAG content and PKC activation. Several studies demonstrated that activation of PKC isoforms is regulated by DAG converted by lipin-1 in various Cells were lysed and analyzed for p21 expression. The p21 levels were normalized to GAPDH. C, D: NHEKs transfected overnight with vector containing lipin-1 cDNA or pCMV-AC empty vector were treated with Go6976 (500 nM) for 4 h. Cell lysates were immunoblotted to assess p21 expression. The p21 levels were normalized to GAPDH. All data (mean ± SD) represent three independent experiments. * P < 0.05, ** P < 0.01, *** P < 0.001 versus control. linked to activation of the extracellular signaling cascade, retinoblastoma hypophosphorylation, induction of CDK inhibitors p21/p27, and inhibition of cyclin D1 ( 51,52 ). In addition, sustained extracellular signal-regulated kinase activation by phorbol 12-myristate 13-acetate-induced PKC/ PKC ␣ stimulation results in growth arrest and differentiation, while transient activation of extracellular signalregulated kinase cascades with serum leads to cell cycle progression in these cells ( 53 ). On the other hand, similar to our results, PKC ␣ and PKC suppression induces p21 expression and growth arrest in Swiss3T3 fi broblast cells ( 40 ). In our study, inhibition of lipin-1 expression by targeting which is accompanied by the induction of p21 in normal keratinocytes ( 49,50 ). Consistent with TPA treatment, overexpression of PKC ␦ and -␣ in keratinocytes prevents proliferation by a mechanism associated with elevated p21 expression ( 38,39 ). On the contrary, we found that lipin-1 KD-induced suppression and lipin-1-induced activation of PKC ␣ are associated with induction and inhibition of p21 expression, respectively, in NHEKs ( Fig. 5 ). The effect of PKC ␣ activation on cell cycle progression is somewhat controversial, and depends on cell type and strength/duration of stimuli. In intestinal epithelial cells, PKC ␣ overexpression and TPA treatment-induced G1 arrest are Fig. 6. Lipin-1 downregulation inhibits keratinocyte differentiation by sustaining p21 induction. NHEKs transfected with lipin-1 siRNA or control siRNA for 24 h were incubated in KGM-Gold medium. A, B: Cells were harvested after 4 days and cell extracts were analyzed by immunoblotting with antibodies specifi c for the indicated differentiation markers. Protein levels of keratin 5, keratin 14, keratin 1, keratin 10, aquaporin 3 [glycosylated (upper band), and unglycosylated (lower band) forms], and involucrin were normalized to GAPDH. C, D: Cells were harvested at the indicated time points (days 0, 1, 2, 3, and 4). Cell lysates were subjected to immunoblot analysis for lipin-1, keratin 1, keratin 10, and p21. The p21 level was normalized to GAPDH. E, F: NHEKs transfected with either control siRNA, p21 siRNA, lipin-1 siRNA, or both siRNAs were incubated in KGM-Gold medium for 4 days. Cells were lysed and analyzed for p21, keratin 1, and keratin 10. Protein levels of p21, keratin 1, and keratin 10 were normalized to GAPDH. All data (mean ± SD) represent three independent experiments. * P < 0.05, ** P < 0.01, *** P < 0.001 versus control. to differentiation induced by p21. While further studies are required to verify involvement of lipin-1 in various PKC ␣ -downstream signaling pathways in keratinocytes, our study provides fundamental insights into the function of lipin-1 and the importance of lipin-1-mediated downregulation of p21 in differentiation of NHEKs.
lipin-1 siRNA and consequent upregulation of p21 lasted for at least 4 days after transfection ( Fig. 6 ), while PKC ␣ overexpression transiently induced PKC ␣ and p21 expression ( 39 ). Therefore, the discrepancy of the effect of PKC ␣ activation/inhibition on p21 may be attributed to differences in duration/magnitude of stimuli. Also, difference in phosphorylation status of PKC ␣ might affect the level of p21 expression.
Earlier work has demonstrated the complexity of p21 function in keratinocyte differentiation. Induction of p21 expression is one of the earliest cell cycle regulatory events contributing to differentiation-associated growth arrest ( 29,30 ). In addition, increased p21 expression must be downregulated by proteasome-mediated degradation for the later stages of differentiation; otherwise, sustained p21 expression blocks the expression of differentiation markers, including keratin 1, loricrin, involucrin, and fi laggrin ( 31 ). In our study, p21 induction caused by lipin-1 KD was sustained during keratinocyte differentiation stages, resulting in delayed and insuffi cient downregulation of p21 and consequently reduced expression of early differentiation markers compared with control. A study of atopic eczema (AE) accompanied with impaired epidermal differentiation indicates lipin-1 downregulation in AE skin, suggesting lipin-1 inhibition may play a role in pathogenesis of AE ( 54 ). The relationship of lipin-1 downregulation on abnormal differentiation in AE skin should be further investigated.
In summary, lipin-1 is required for production of DAG and activation of PKC ␣ signaling, which is associated with downregulation of p21 expression in NHEKs ( Fig. 7 ). Lipin-1 KD caused G1 arrest associated with p21 induction, which is necessary to initiate keratinocyte differentiation. Lipin-1 KD also suppressed the expression of keratin 1 and keratin 10 due to delayed downregulation of p21, which is critical for proper progress of keratinocyte differentiation after the initial commitment of keratinocytes