Promoter Methylation Leads to Decreased ZFP36 Expression and Deregulated NLRP3 Inflammasome Activation in Psoriatic Fibroblasts

Abstract

The mRNA-destabilizing protein tristetraprolin (TTP), encoded by the ZFP36 gene, is known to be able to end inflammatory responses by directly targeting and destabilizing mRNAs encoding pro-inflammatory cytokines. We analyzed its role in psoriasis, a disease characterized by chronic inflammation. We observed that TTP is downregulated in fibroblasts deriving from psoriasis patients compared to those deriving from healthy individuals and that psoriatic fibroblasts exhibit abnormal inflammasome activity compared to their physiological counterpart. This phenomenon depends on TTP downregulation. In fact, following restoration, TTP is capable of directly targeting for degradation NLRP3 mRNA, thereby drastically decreasing inflammasome activation. Moreover, we provide evidence that ZFP36 undergoes methylation in psoriasis, by virtue of the presence of long stretches of CpG dinucleotides both in the promoter and the coding region. Besides confirming that a perturbation of TTP expression might underlie the pathogenesis of psoriasis, we suggest that deregulated inflammasome activity might play a role in the disease alongside deregulated cytokine expression.

Keywords: NLRP3; ZFP36; inflammasome; methylation; psoriasis.

Figure 1

(A–D) Relative expression of selected genes, obtained from a publicly available database. Expression of each gene is shown in healthy skin and in lesional psoriatic skin. Boxplots report ±SEM (***p < 0.001). (A) ZFP36 expression in psoriatic skin samples is significantly reduced, compared to healthy skin samples [Healthy donor: n = 21, median 10.83, range 8.59–12.85; Lesional psoriatic fibroblasts (PSO): n = 33, median 9.23, range 8.79–9.78]. (B) NLRP3 levels show no significant variation among the two groups of healthy and psoriatic skin (Healthy donor: n = 21, median 2.73, range 2.58–3.13; Lesional PSO: n = 33, median 2.80, range 2.56–3.13). (C) CASP1 expression results increased in lesional psoriatic samples compared to normal skin samples (Healthy donor: n = 21, median 6.88, range 5.13–8.52; Lesional PSO: n = 33, median 8.00, range 5.45–8.77). (D) IL1B is significantly more expressed in lesional skin samples from psoriatic patients than skin from healthy donors (Healthy donor: n = 21, median 4.04, range 3.69–6.67; Lesional PSO: n = 33, median 5.05, range 4.06–7.03). (E,F) Relative quantity of mRNA measured by real-time (RT) PCR in healthy and psoriatic fibroblasts samples. (E) Differential expression of TTP mRNA in psoriatic and healthy fibroblasts, measured through qRT-PCR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as endogenous control. Results are represented as means of three experiments (±SEM) (*p < 0.05). (F) IL1β mRNA levels in psoriatic and healthy fibroblasts, measured through qRT-PCR. GAPDH was used as endogenous control. Results are represented as the means of three experiments (±SEM) (**p < 0.01).

Figure 2

(A) Tristetraprolin (TTP) protein levels in healthy and psoriatic fibroblasts. Two different healthy samples and two different psoriatic samples were divided into two series. In “Fibroblasts II series,” psoriatic cell samples from the same donor are distinguished in lesional and non-lesional. Vinculin was used as loading control. (B) NLRP3, pro-IL1β, and IL1β (p17) protein levels in healthy and psoriatic fibroblasts. (C) Pro-CASP1 and active CASP1 (p20) protein levels in healthy and psoriatic fibroblasts. Vinculin was used in (B,C) as a loading control.

Figure 3

(A) qRT-PCR shows a decrease in NLRP3 mRNA in HaCaT cells transfected with a TTP-expressing vector (pcDNA3.1-ZFP36), compared to control transfection (pcDNA3.1). GAPDH was used as endogenous control. Results are represented as the means of three experiments (±SEM) (**p < 0.01). (B) Luciferase reporter assay shows a decrease in luciferase activity in HEK293T cell line cotransfected with luciferase reporter vector and the TTP-expressing vector, compared to the control cotransfection with the empty vector. Measured luciferase activity was normalized over β-gal signals. Results are represented as the means of three experiments (±SEM) (*p < 0.05). (C) Human NLRP3 3′UTR displaying TTP binding site (highlighted). (D) NLRP3 mRNA levels measured by qRT-PCR in HaCaT cells transfected at time 0 with a TTP-overexpressing vector or an empty vector. After transfection (48 h), cells were treated with Act.D to block transcription. NLRP3 mRNA levels were recorded at the moment of ActD treatment (Ctr), after 1 h and after 3 h. Results are represented as the means of three experiments (±SEM). Statistical analysis was performed comparing ZFP36-transfected and empty vector-transfected cells at each timepoint (*p < 0.05). GAPDH was used as endogenous control.

Figure 4

(A) TTP silencing achieved by lentiviral infection (sample labeled shZFP36) was verified by Western blot, in comparison to endogenous TTP levels measured after infection with a vector expressing an aspecific scrambled short hairpin (labeled EV). Actin was used as loading control. (B) Western blot displaying the expression levels of inflammasome components. Inflammasome interactor levels vary after lentiviral infection with a short hairpin directed against TTP mRNA. Vinculin is included as loading control. (C) NLRP3, IL1β, and TTP mRNA levels have been analyzed by qRT-PCR in both cell samples. GAPDH was used as endogenous control. Results are represented as the means of three experiments (±SEM).

Figure 5

(A) NLRP3 inflammasome protein levels are displayed by Western blot in three different fibroblast cell populations: healthy fibroblasts (Healthy), psoriatic fibroblasts (PSO), and psoriatic fibroblasts infected with a pRRL-TTP-overexpressing vector (PSO TTP). Tubulin was used as loading control. (B) TTP overexpression is shown in psoriatic fibroblast sample. Actin was used as loading control. (C) Inflammasome components variation pattern already seen in primary fibroblasts is maintained in the HaCaT keratinocyte cell line. Actin was used as loading control. (D) pRRL-empty vector was used to obtain the overexpression vector pRRL-TTP, by substitution of ΔNGFR sequence with TTP cDNA.

Figure 6

(A) TTP mRNA levels in healthy and psoriatic fibroblasts treated with 5-aza, measured by qRT-PCR. GAPDH was used as endogenous control, and results are represented as the means of three experiments (±SEM) (**p < 0.01). Healthy fibroblasts ctr were used as internal calibrator. (B) TTP protein levels in healthy and psoriatic fibroblasts after treatment with 5-aza, measured by Western blot. Vinculin was used as loading control.

Figure 7

Methylation profile of ZFP36 promoter in dermal fibroblasts from skin biopsies. CpG sites are represented by red vertical bars along the upstream region of ZFP36 transcription start site, which was taken as “+1” position of base counting. The “F1” subregion, between −1,000 and −500, was amplified by PCR after bisulfite conversion and then sequenced to obtain the presented methylation profile. Two fibroblast samples from healthy donors and two lesional fibroblast samples from psoriasis donors were analyzed. White circle, unmethylated CpG; black circle, methylated CpG.