Unraveling the Interplay Between FGF and NRF2 in Inflammatory Skin Diseases

Abstract

The skin is the largest and outermost organ in mammals and serves a large variety of functions, including temperature regulation, sense of mechanical stimuli, and most importantly formation of a protective barrier. It can be subdivided into the epidermis, in which keratinocytes are the primary cell type, the underlying dermis, and the lowest layer - the hypodermis. The skin can be affected by various diseases, which often cause a severe burden for the patients and strongly impact the quality of life. The most common chronic inflammatory skin disease is atopic dermatitis (AD), which affects at least 230 million people worldwide. It is a relapsing disease characterized by recurrent eczematous lesions, erythroderma, intense itch and discomfort. The pathophysiology of AD is complex, multifactorial and highly heterogeneous, and involves genetic predisposition as well as environmental factors. Epidermal barrier dysfunction and immunological alterations promoting a heightened type 2 response have been recognized as crucial drivers of the disease. Due to its complex nature, AD pathogenesis is still incompletely understood. Until today, the disease cannot be cured, and the available treatments only ameliorate the symptoms. Genetic studies indicate a pivotal role of defects in keratinocytes in the pathogenesis of AD. However, the global changes of the proteome of the entire epidermis remained unknown. Our laboratory previously used a pressure-cycling technology and data-independent acquisition approach to generate the first quantitative proteomics dataset of epidermis from healthy volunteers and from lesional and non-lesional skin of AD patients. Validation was performed using targeted proteomics with parallel reaction monitoring mass spectrometry. In my thesis project, I analyzed these proteomics data, and I verified some of the hit proteins be immunofluorescence staining. This analysis identified impaired activation of the NRF2-antioxidant pathway and a reduced abundance of mitochondrial proteins involved in key metabolic pathways in the affected epidermis. Knock-down of NRF2 in primary human keratinocytes by small interfering RNAs revealed that NRF2 controls the expression of a small subset of mitochondrial genes, and reduction of NRF2 activity reduced the concentrations of intracellular ATP. To study the role of NRF2 in the context of skin inflammation and to determine if the down-regulation of NRF2 activity in the epidermis of AD patients is functionally relevant, I made use of a mouse model that exhibits various features resembling AD in humans. These mice, designated K5-R1/R2 mice, lack fibroblast growth factor receptors 1 and 2 (Fgfr1 and Fgfr2) in keratinocytes. I found an age-dependent down-regulation of Nrf2 activity in the epidermis of these mice, which correlated with disease progression. This was associated with increased DNA damage and senescence. Additional loss of Nrf2 in keratinocytes of these mice, which was achieved by breeding with conditional Nrf2 knockout mice, mildly aggravated the phenotype. Surprisingly, long-term genetic activation of Nrf2 in keratinocytes of K5-R1/R2 mice caused severe hyperkeratosis, keratinocyte hyperproliferation, epidermal thickening, as well as increased keratinocyte apoptosis, DNA damage and senescence, which were at least in part a consequence of excessive upregulation of certain cornified envelope proteins. However, time-limited pharmacological activation of Nrf2 in the skin of adult K5-R1/R2 mice promoted skin barrier function and protected from DNA damage. This protective activity of Nrf2-activating compounds was even more pronounced in Spink5 knockout mice, a mouse model for Netherton syndrome. Previous studies from our laboratory showed that this treatment significantly alleviates the cutaneous phenotype of Spink5 knockout mice and promotes attachment of the stratum corneum and concomitant epidermal barrier function. I contributed to this work by showing that Nrf2 activation induces overexpression of secretory leukocyte protease inhibitor (SLPI), a known inhibitor of kallikrein 7 and elastase 2, in human keratinocytes in vitro. In the Spink5-deficient epidermis, upregulation of Slpi is likely to promote stabilization of corneodesmosomes, thereby preventing premature desquamation. Besides NRF2, analysis of published transcriptome data revealed downregulation of FGFR2 expression in the epidermis of lesional skin of AD patients. This is of likely importance for the disease phenotype, because mice lacking Fgfr1 and Fgfr2 in keratinocytes develop AD-like symptoms. To study the relevance of FGFR signaling for skin inflammation in humans, we established a genetic in vitro model. Through CRISPR/Cas9-mediated FGFR2 knockout in the human HaCaT keratinocyte cell line, we abrogated the activity of FGF7 and FGF10 to induce cell signaling, migration and proliferation of keratinocytes, without triggering compensatory upregulation of other FGF receptors. Rather, loss of FGFR2 suppressed the expression of FGFR3. Most importantly, the knockout of FGFR2 promoted the expression of interferon-stimulated genes under homeostatic conditions and in response to inflammatory mediators. This activity is likely to further aggravate the inflammatory phenotype in an in vivo setting. Taken together, the results obtained in this thesis provide insight into the molecular mechanisms underlying the pathogenesis of AD and identify NRF2 and FGFR2 as potential targets for future therapeutic applications. In particular, we propose restricted activation of NRF2 as a possible therapy for this common inflammatory skin disease.