Selenium unmasks protective iron armor: A possible defense against cutaneous inflammation and cancer

Abstract

Background: A link between selenium deficiency and inflammatory skin diseases have been noted by many, but this link is still not well understood. We have previously studied the efficacy of ceramide analogs, based on the fire ant venom Solenopsin A, against our psoriasis animal model. Treatment of animals with solenopsin analogs resulted in significantly improved skin as well as in a coordinate downregulation of selenoproteins, namely Glutathione Peroxidase 4 (GPX4). We thus hypothesize that ferroptosis may be a physiologic process that may protect the skin from both inflammatory and neoplastic processes.

Methods: We analyze and compare gene expression profiles in the GEO database from clinical skin samples taken from healthy patients and psoriasis patients (both involved and noninvolved skin lesions). We validated the gene expression results against a second, independent, cohort from the GEO database.

Results: Significant reduction in gene expression of GPX4, elevated expression of Nrf2 downstream targets, and expression profiles mirroring erastin-inhibition of Cystine/Glutamate Antiporter-System X_C activity in psoriatic skin lesions, compared to both noninvolved skin and healthy patient samples, suggest an innately inducible mechanism of ferroptosis.

Conclusions: We present data that may indicate selenoproteins, particularly GPX4, in resolving inflammation and skin cancer, including the novel hypothesis that the human organism may downregulate GPX4 and reactive oxygen (REDOX) regulating proteins in the skin as a way of resolving psoriasis and nonmelanoma skin cancer through increased reactive oxygen species. Further studies are needed to investigate ferroptosis as a possible physiologic mechanism for eliminating inflammatory and malignant tissues.

General significance: This study provides a fresh framework for understanding the seemingly contradictory effects of selenium supplementation. In addition, it offers a novel explanation of how physiologic upregulation of ferroptosis and downregulation of selenoprotein synthesis may mediate resolution of inflammation and carcinogenesis. This is of therapeutic significance.

Fig. 1.

Proposed Physiologic Ferroptosis in Psoriasis and SCC. We hypothesize the concept of physiologic ferroptosis as a mechanistic target that may be induced by ceramide analogs, such as solenopsin derivatives, which suppress GPX4 expression is psoriatic lesions in mouse models. We believe that further suppression of GPX4 expression seen in clinical settings may play a role to resolving skin disorders.

Fig. 2.

Lower GPX4 expression in Psoriatic Lesions. Significantly reduced GPX4 expression in clinical samples suggesting physiologic ferroptosis. Data from GEO data set GDS4602 [41] comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: Normal distribution, ANOVA test with bars indicating mean raw values, ± SEM. **** and #### indicating p < 0.0001; and ns as p > 0.05.

Fig. 3.

Expression of Selenoproteins and related family proteins in Psoriatic Lesions. Expression levels of major selenoproteins thioredoxin reductase1 (Txnrd1), and glutathione peroxidases 1, 2, 3, (GPX1, 2, 3), as well as, related proteins: thioredoxin (Trx1) and glutathione peroxidase 5 (GPX5) in healthy and psoriatic skin. Data from GEO data set GDS4602 comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: Normal distribution, ANOVA test with bars indicating mean raw values, ± SEM. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.

Fig. 4.

Evidence of Nrf2 Activation and Elevated Expression of Downstream Targets in Psoriatic Lesions. Significantly increased expression levels of downstream targets of Nrf2 suggesting oxidative stress in clinical samples from physiologic ferroptosis. Data from GEO data set GDS4602 comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: Normal distribution, ANOVA test with bars indicating mean raw values, ± SEM or nonparametric Kruskall-Wallis test with bars indicating median raw values, ± interquartile range. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.

Fig. 5.

Gene Expression Suggesting Halting of H2AX/Bcl-2 Apoptosis in Psoriatic Lesions. Significant upregulation of H2A histone family member X (H2AX), proapoptotic B-cell lymphoma 2 (Bcl-2), Bcl-2-like protein 4 (BAX), Bcl-2 homologous antagonist/killer (BAK), and Bcl-2 ovarian killer (BOK) suggesting apoptosis likely halted by significant downregulation of MIR24–2, an essential mediator of H2AX/Bcl-2 apoptotic pathway. Data from GEO data set GDS4602 comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: nonparametric Kruskall-Wallis test with bars indicating median raw values, ± interquartile range. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.

Fig. 6.

Gene Expression Suggesting Increase in Intracellular Iron in Psoriatic Lesions. Significant changes in gene expression of transferrins (TFRC or TFR1) and TFR2, metal regulatory transcription factor 1 (MTF1), specificity protein 1 (SP-1), glutaredoxin 3(GRX3), and ceruloplasmin (CP), involved cellular iron sensing suggest increases in intracellular iron in psoriatic skin. Data from GEO data set GDS4602 comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: Normal distribution, ANOVA test with bars indicating mean raw values, ± SEM. or nonparametric Kruskall-Wallis test with bars indicating median raw values, ± interquartile range. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.

Fig. 7.

Gene Expression in Psoriatic Skin lesions mirror Erastin-like System XC inhibition. Psoriatic lesions show gene expression pattern similar to erastin-like System XC inihibition. SLC7A11 and SLC3A2, were upregulated as compensatory response to similar erastin. Significant upregulation of Activating transcription factor 4 (ATF4), and ChaC Glutathione Specific Gamma-Glutamylcyclotransferase 1 (CHAC1) was also in agreement with an erastin-liked induced inhibition of System XC, the mechanism of erastin-mediated ferroptosis. Data from GEO data set GDS4602 comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: nonparametric Kruskall-Wallis test with bars indicating median raw values, ± interquartile range. ± SEM. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.

Fig. 8.

INFɣ suppresses GPX4 expression in Healthy skin and Psoriatic Lesions. Significantly reduced GPX4 expression in clinical samples from INFɣ treated patients suggesting effective therapies such cytokine-based therapies may act in part by physiologic ferroptosis. Data from GEO data set GDS4607 comparing healthy patients (n = 10) to Psoriasis Patients with samples from nonlesional healthy skin (n = 10) and psoriatic skin (n = 10). Statistical Analysis: non-parametric Kruskall-Wallis test with bars indicating median raw values, ± interquartile range. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.

Fig. 9.

Elevated NFKBIZ expression in Psoriatic Lesions. Significantly elevated expression NFKBIZ, a modulator of NFKB, in clinical samples linking psoriasis with senescence associated secretory phenomenon (SASP). Data from GEO data set GDS4602 comparing healthy patients (n = 64) to Psoriasis Patients with samples from nonlesional healthy skin (n = 58) and psoriatic skin (n = 58). Statistical Analysis: Normal distribution, ANOVA test with bars indicating mean raw values, ± SEM. **** and #### indicating p < 0.0001; *** or ### p < 0.001, * or # p < 0.05 and ns as p > 0.05.