DMEM (high glucose; Cat. No. 12100-046) and trypsin-EDTA were from Gibco (Carlsbad, CA, United States). FBS was from HyClone (Logan, UT, United States). Penicillin-streptomycin solution and penicillin-streptomycin-amphotericin B solution were from Biological Industries (Kibbutz Beit Haemek, Israel). Poly (I:C) was from InvivoGen (San Diego, CA, United States). TNF-α was from Biolegend (San Diego, CA, United States). PMA, PBS, mitomycin C, LPS, and H₂O₂ were from Sigma-Aldrich (St. Louis, MO, United States). Gefitinib was from Selleckchem (Houston, TX, United States). Recombinant human EGF was from PeproTech (Rocky Hill, NJ, United States). Blimp-1 (#9115), p-EGFR (Y1068, #2234) and Syk (#2712) antibodies were from Cell Signaling (Beverly, MA, United States). EGFR antibody (sc-03) was from Santa Cruz (Santa Cruz, CA, United States). p-Syk antibody (PK1010) was from Millipore (Burlington, Ma, United States).

Human immortalized HaCaT keratinocytes, oral SCC Cal-27 and SAS cells were cultured in high glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS and 1% penicillin-streptomycin-amphotericin B. NHEKs (normal human epidermal keratinocytes) were obtained from normal adult human foreskin and isolated as described previously (Chiu et al., 2021). The experiments were conducted according to the Declaration of Helsinki principles and approved by the Ethics Committee of Mackay Memorial Hospital (Institutional Review Board codes 19MMHIS173e). All cell lines were incubated at 37°C under a humidified atmosphere of 5% CO₂ in air. HaCaT and SCC cell lines were seeded in 6-well or 12-well tissue culture plates at a density of 8 × 10⁴ cells/well.

In construction of stable short hairpin RNA (shRNA) knockdown cell lines, lentiviral particles encoding shRNA targeting human PRDM1 (Sigma-Aldrich, St. Louis, MO, United States) were used for transfection. Lentivirus-containing supernatants were harvested 24 h after transfection, filtered using a 0.45 μm filter, and diluted with fresh culture media to transduce target cells in the presence of 8 μg/ml PolyBrene (hexadimethrinebromide). Transduced cells were selected with puromycin (3 μg/ml) (Thermo Fisher Scientific, Waltham, MA, United States) for 2 weeks to select successful transfection.

Cells were lysed by adding radioimmunoprecipitation assay (RIPA) buffer. The extracts were sonicated for 10-15 sec to complete cell lysis and shear DNA, and then centrifuged at 16,200 g, 4°C for 30 min. The protein concentrations of the supernatants were determined using the Bio-Rad protein assay. Equal amounts of the protein were loaded and electrophoresed on 8–15% SDS-PAGE, and then electro-transferred to Immobilon-P (0.45 μm PVDF; Millipore). After transfer, the membrane was blocked with Tris-buffered saline with Tween 20 (TBST) containing 5% (w/v) nonfat dry milk for 1 h at room temperature. After incubation with the primary antibodies (at the appropriate dilution as recommended in the product data sheet) with gentle agitation overnight at 4°C, the membranes were washed with TBST for three times and incubated with horseradish-peroxidase-linked secondary antibodies with gentle agitation for 1 h at room temperature. After washing with TBST for three times, the protein bands were detected on X-ray film with ECL reagents.

After stimulation, cells were harvested with TriPure isolation reagent (Roche Diagnostics, Indianapolis, IN, United States) and RNA was extracted according to the manufacturer’s procedure. Total RNA (1–2 μg) was converted into cDNA by reverse transcription system kit (Promega, Heidelberg, Germany). Q-PCR was performed using FastStart SYBR Green Master (Roche Diagnostics, Indianapolis, IN, United States) in 96-well plates, and determined using ABI Prism 7900 (Applied Biosystems, Oakland, CA, United States). The primers used for human RPDM1 were 5′-CGA​AAT​GCC​CTT​CTA​CCC​T-3′ and 5′-GCG​TTC​AAG​TAA​GCG​TAG​GA-3′ and the primers used for human β-actin were 5′-AGG​AAG​GCT​GGA​AGA​GTG​C-3′ and 5′-CGG​GGA​CCT​GAC​TGA​CTA​CC-3’.

Cells were seeded (4 × 10⁴ cm²) into 12-well culture-insert purchased from ibidi (Martinsried, Germany). After attachment, the culture-insert was gently removed from each well, and the well was washed three times with PBS to remove the suspended cells. Before adding the drugs, the cells were incubated with anti-proliferative agent mitomycin C (5 μg/ml) for 30 min. Finally, cells were incubated in medium in the absence or presence of PMA (30 nM) or TNF-α (10 ng/ml) in HaCaT cells, and TNF-α (10 ng/ml) or EGF (50 ng/ml) in SCC. The process was recorded by photographs, and cell migration was quantified.

Blimp-(wt) and Blimp-(AP1 mt) report constructs (human Blimp-1 promoter in pGL3-basic) were gifts from Alexander Dent (Addgene plasmid # 40340 and #40341). Following the commercial standard protocol, HaCaT cells were transfected with Blimp-1 reporter plasmid and β-galactosidase expression vector by using Lipofectamine 2000 reagent (Invitrogen) and then EGF (50 ng/ml) was treated for 24 h. After harvest, the luciferase activity was determined by luciferase assay system kit (Promega, Heidelberg, Germany), followed by microplate luminometer. Luciferase activity was normalized with activity of β-galactosidase, and expressed as fold of control without stimulus treatment.

HaCaT, Cal-27, and SAS cells in full serum DMEM medium were fixed with 4% paraformaldehyde and permeabilized with 0.2% Triton X-100 in PBS for 20 min. After this process, the samples were blocked with 4% BSA for 1 h and incubated with primary antibody for 2 h at room temperature or overnight at 4°C after aspiration of blocking solution. The primary antibody was then discarded and cells were washed three times with PBS. The samples were incubated with fluorochrome-conjugated secondary antibody for 1 h in the dark afterwards. Following immunostaining process, the coverslip was counterstained with 4′6-diamidino-2-phenylindole (DAPI), and mounted on microscope slides in dark. Samples were analyzed by LSM 880 confocal microscope (Zeiss).

Values were expressed as the mean ± S.E.M. of at least three independent experiments. Student’s t-test or one-way ANOVA was performed to analyze the statistical significance of the differences, and the P value <0.05 was considered statistically significant.

To understand the regulation of Blimp-1 expression in keratinocytes, we tested several agents including tumor promoter PMA, TLR4 ligand lipopolysaccharide (LPS), cytokine TNF-α, ROS stressors H₂O₂ and UVB in human HaCaT keratinocytes. We found that LPS (1 μg/ml) (Figure 1A), PMA (30 nM), TNF-α (10 ng/ml) and H₂O₂ (200 μM) (Figure 1B) can increase Blimp-1 protein expression. Compared to other stimuli, the effect of H₂O₂ is much weaker. Nevertheless, polyIC (TLR3 ligand) (50 μg/ml) had no significant effect Blimp-1 expression in HaCaT cells (data not shown). On the other hands, UVB (50 mJ/cm²) could also induce Blimp-1 expression within 5–7 h (Figure 1C). Consistently the data from the Q-PCR study indicated that PMA, TNF-α, LPS and UVB can increase Blimp-1 mRNA level (Figure 1D).

Because previously EGF was shown to upregulate Blimp-1 expression in keratinocytes (Chang et al., 2018) and EGFR can be transactivated via extracellular and intracellular manners in various cell types including keratinocytes (Nanba et al., 2013; Wu et al., 2016), we interested to explore if Blimp-1 inducers mentioned above might exert actions related to EGFR. To this end, we determined the effects of these stimuli on EGFR expression and activation. We found that LPS (Figure 1A), PMA, TNF-α, H₂O₂ (Figure 1B) and UVB (Figure 1C) could activate EGFR in HaCaT cells. Strengthening these observations not only in HaCaT cells, we found that LPS and TNF-α also can induce Blimp-1 expression in NHEKs (Figure 1E). In our experimental conditions and time intervals, we ruled out the death effect of all tested agents including H₂O₂ and UVB in keratinocytes and SCC.

Besides keratinocytes, we chose two SCC cell lines Cal-27 and SAS to understand the Blimp-1 regulation. In the same manners as seen in HaCaT cells, PMA, TNF-α, LPS, H₂O₂ and UVB could time-dependently increase Blimp-1 protein expression in Cal-27 cells (Figures 2A,B). Unlike HaCaT cells where polyIC failed to increase Blimp-1 protein, we found it can exert this action in Cal-27 cells (Figure 2A). In Q-PCR study, our data revealed the abilities of PMA, LPS, TNF-α, H₂O₂ and UVB to upregulate Blimp-1 gene expression (Figure 2C). Moreover, in SAS cells PMA, TNF-α, LPS, polyIC, H₂O₂ and UVB all increased Blimp-1 protein expression (Figures 2D,E), and UVB also increased Blimp-1 mRNA level (Figure 2F). All these data suggest that Blimp-1 can be induced by PMA, TNF-α, LPS, H₂O₂ and UVB in both keratinocytes and SCC. Similarly, we also determined EGFR activation in SCC cells. We found that PMA, TNF-α, LPS, polyIC, H₂O₂ and UVB all activated EGFR in Cal-27 cells (Figures 2A,B). In SAS cells, the EGFR was also activated by PMA, TNF-α, LPS, polyIC, H₂O₂ and UVB (Figures 2D,E). All these findings indicate that transactivation of EGFR is induced by all Blimp-1 regulators, suggesting the role of EGFR activation in Blimp-1 gene expression as we previously reported in the case of exogenous EGF (Chang et al., 2018).

Previous studies indicated that AP-1 is involved in BCR-mediated and TGF-β-induced Blimp-1 gene expression in B lymphocytes (Calame, 2008) and breast cancer cells (Romagnoli et al., 2012), respectively. We also found that PMA-induced Blimp-1 is dependent on EGFR transactivation in keratinocytes, because EGFR inhibitor gefitinib can reduce this effect (Chang et al., 2018). Here we used reporter assay to check if AP-1 is required for Blimp-1 expression caused by various agents in keratinocytes. As shown in Figure 3A, PMA, LPS, TNF-α and EGF treatment could increase the luciferase activity of Blimp-1 in HaCaT cells with the highest effect of EGF, and AP-1 mutation abolished these effects of different agents (Figure 3A). In addition, we also observed the ability of gefitinib to reduce PMA-induced Blimp-1 gene expression in Cal-27 cells (Figure 3B). To clarify if gefitinib might have off-target effect on PMA signaling, we used PKCδ phosphorylation as the index of PKC activation. We found that gefitinib did not change PMA-induced PKCδ phosphorylation in Cal-27 cells (Figure 3C).

Moreover, previously we found that Syk is not only an upstream signaling molecule of EGFR in keratinocytes (Wu et al., 2016) but also can be activated by PKC in monocytes (Chang et al., 2012). Therefore, we wonder if Syk is involved in the action of PMA for Blimp-1 expression. First, as our previous study showing the effect of PMA on Syk activation in monocytes (Chang et al., 2012), we found that in HaCaT cells, PMA could also increase active phospho-Syk level (Figure 3D, left panel) and Syk inhibitor GS-9973 could block PMA-induced Blimp-1 upregulation (Figure 3E). Likewise, TNF-α also can activate Syk (Figure 3D, right panel). In Cal-27 cells, GS-9973 was also found to inhibit PMA-induced Blimp-1 gene expression (Figure 3F) as well as EGFR activation (Figure 3C). These findings suggest the involvement of PKC-Syk axis in PMA-induced EGFR transactivation and subsequent Blimp-1 gene expression.

After observing the relationship between Blimp-1 expression and EGFR activity, we determined the subcellular localization of Blimp-1 and EGFR in keratinocytes and SCC. Data of immunoblotting analysis of subcellular fractions revealed that Blimp-1 is located in the nuclei and EGFR is present in both nuclei and plasma membrane in HaCaT cells (Figure 4A). Data of confocal microscopy revealed that Blimp-1 and EGFR are mainly present in the nuclei and plasma membrane, respectively (Figure 4B). Similar intracellular distributions of Blimp-1 and EGFR were also observed in NHEK. In addition, PMA and TNF-α treatment for 6 h cannot alter both proteins’ localization in HaCaT cells (data not shown). Besides we checked the subcellular location of Blimp-1 and EGFR in Cal-27 and SAS cells before and after stimuli treatment. We found that Blimp-1 is localized in the nuclei of both cancer cell lines (Figure 4B) and is still kept in the nuclei after PMA, TNF-α or UVB stimulation (Supplementary Figure S1).

Next to know the role of Blimp-1 in cell migration, we knocked down Blimp-1 in HaCaT cells using shRNA. We found silencing of Blimp-1 increases wound closure percentages as compared to control group in SAS (Figure 5B) and Cal-27 cells (Figure 5C), but not in HaCaT cells (Figure 5A). PMA slightly inhibited the cell migration of HaCaT keratinocytes. The inhibitory effect of PMA is similar to previous finding and possibly is due to the keratinocyte differentiation action of PMA (Ando et al., 1993). In both SCC, EGF but not TNF-α can promote cell migration. shBlimp-1 induced a significant enhancement on cell migration in both types of SCC, and this effect was still observed upon TNF-α treatment and was non-additive to the stimulating response of EGF (Figures 5B,C). These effects indicate that Blimp-1 is a negative regulator of cell migration in SCC.