Mouse Models of Psoriasis – A Comprehensive Review

Abstract

The use of preclinical animal models of psoriasis has significantly increased over the last three decades, with each model having unique strengths and limitations. Some models translate better to human disease, and many have provided unique insight into psoriasis disease pathogenesis. In this comprehensive review we present comparative description and discussion of genetic mouse models, xenograft approaches, and elicited methods using cytokine injections into, and topical imiquimod onto mice. We provide an inclusive list of genetically modified animals that have had imiquimod applied to, or cytokines injected into their skin and describe the outcomes of these manipulations. This review will provide a valuable resource for those interested in working with psoriasis mouse models.

INTRODUCTION

Since 2007, when we published our first Mouse Models of Psoriasis review (Gudjonsson et al., 2007) many new genetically innovative as well as easy-to-use models have been generated to study psoriasiform skin inflammation in mice (Figure 1, Figure S1, Table S1). Without repeating what was published then and continues to be applicable and useful in describing differences between mouse and human skin and immune systems, this review will modernize past information and will include a description and discussion of models since that time. New genetic mouse models of psoriasis are presented (Figures 1–2, Table 1, Tables S1–6) and compared with those previously described (Table 1, Table S1, Table S7). New cellular players identified in murine psoriasis, that may not be critical to human disease are discussed, and the advent of topical imiquimod as a heavily utilized acute model of psoriasis-like skin inflammation comprehensively reviewed (Figures 1–2, Figure S1, Table S2, (van der Fits et al., 2009)). We have attempted to summarize the vast literature encompassing the use of topical imiquimod, which currently stands at more-than 300 unique mouse strains (Table S2) and how the genetics of each strain impact psoriasiform inflammation. Moreover, intradermal injection of psoriasis-related cytokines, in particular IL-23 (Chan et al., 2006), has also become a regularly used approach for studying acute outcomes that model some aspects of psoriasis. Like imiquimod, many unique genetically modified animals have had IL-23 injected into their skin in order to elucidate IL-23 contributions to psoriasis (Figures 1–2, Table 1, Table S3).

Figure 1. Timeline of milestone discoveries in psoriasis pathogenesis and mouse models of psoriasis.

Chronological landmark discoveries made for human psoriasis pathogenesis are depicted in addition to the advent of individual mouse models of psoriasis. Each model has been placed into a broadly defined category and listed by date of publication and is called by the name used in the original paper. Mouse models that develop a psoriatic arthritis-like phenotype are listed in blue font. The timeline of therapeutic discoveries and FDA-approved drugs (brown background), including biologics (white font, dark brown background) are also presented. Discontinued drugs appear in a white background. See Table S1 for PMID and original publication citations for each animal model.

Figure 2. Summary of mouse models of psoriasis according to category and overlap with human psoriasis.

Graphical presentation of all mouse models of psoriasis listed in Table S1 according to total number of publications for each general category (a). Total publications for specific models within each category (b). Bubble size in a and b represent number of models. Total publications for each model classified by target, excluding the imiquimod model (c). Bubble size in c represents number of unique models for each target. (GF, growth factor). (d) Bioinformatic comparison of genes and outcomes manipulated in transgenic, topical imiquimod and IL-23 intradermal injections mouse models with changes in gene expression in psoriasis patient lesional skin. (e) Reactome fold enrichment analyses identifies top pathways for each model.

Table 1.

Summary comparison of animal models of psoriasis compared to human psoriasis categorized by type.

Type	Target	Binary Approach	Promoter	Gene	Model Name	Reference	PMID	Year	Background Strain	Methods	Acanthosis	AlteredDifferentiation	VascularPhenotype	T Cells	Immune CellInfiltration	NeutrophilicMicro-abscesses	T cell Responsive	Responsive toOther Therapies	Phenotype Dependenton Immune Activation	Co-morbidities	Remarks	Additional Refs (PMIDs)
Human Psoriasis					Human Psoriasis		https://www.psoriasis.org/about-psoriasis/				Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
Spontaneous				Ttc7	Flaky skin; fsn/fsn	Sundberg, J. P., et al., J Invest Dermatol, 1990.	16788637	1990	A/J, now on BALB/cByJ & C57BL/6J	Spontaneous	Y	Y	Y	Y	Y	Y	NM	NM	N	NM	Renal glomerulopathy, testicular degeneration, fibrosis around portal triads, splenomegaly, granulomatous lymphadenitis	8406628, 8176263, 7650236,8752652, 9459497, 10651923,10771480, 11114969, 11298140,16179734, 16689863, 16470175,25533330, 29775636, 9459497
Xenograft				Prkdc	SCID-hu	Boehncke, W. H., et al., J Cutan Pathol, 1997.	9027626	1997	C.B17 SCID	Xenograft	Y	Y	NM	Y	NM	Y	NM	NM	Y	NM		11877475, 11286628, 16225604
				Ifnar1, Ifngr1t, Rag2	AGR 129	Boyman, O., et al., J Exp Med, 2004.	14981113	2004	ARG129	Xenograft	Y	Y	Y	Y	NM	NM	Y	NM	Y	NM	Muromonab-CD3 improves phenotype, dependent upon immune activation, and expansion of T cells	28063039, 26782974, 14981113
Global Gene Manipulation	MHC IAllele (OE)			HLA-B, B2M	B27/hβ 2m transgenic rat	Hammer, R. E., et al., Cell, 1990.	2257626	1990	RAT F344/Crli (Line 21–4H)	Spontaneous	Y	Y	Y	Y	Y	NM	Y	NM	?	Arthritis, neurological disorder	Diarrhea, multi-organ inflammation. Dermal neutrophil infiltrate. Rat specific. No phenotype observed in mice expressing HLA-B27 and hβ2m	8543835, 8293002, 17469106
	Adhesion Molecule (Reduced Expression)			Itgb2	CD18hypoPL/J	Bullard, D. C., et al., Proc Natl Acad Sci U S A, 1996.	8700894	1996	CD18–/–mice backcrossed to PL/J strain	Spontaneous	Y	Y	NM	Y	Y	Y	Y	Y	Y	NM	Background strain dependent, phenotype does not occur in C57BL/6, 129S/V, alopecia and crust formation, myeloid hyperplasia in spleen and bone marrow	8101543, 14634077, 16886059,17069006, 16982899, 18390736,18521187, 19752240, 19812597,12819024, 24190659, 23418628
	Receptor Antagonist(KO)			Il1rn	ll1rn–/–	Shepherd, J., et al., J InvestDermatol, 2004.	15086551	2004	BALB/c	Spontaneous	Y	Y	Y	Y	Y	Y	NM	NM	?	Arterial inflammation,arthritis	No disease in C57BL/6 mice	20610641, 28645307, 10637275,18060042
	Signaling Molecule (Mutant)			Card14	Card14 E138A/+ andCard14 ΔQ136/+	Wang, M., et al., Immunity, 2018.	29980436	2018	C57BL/6	Spontaneous	Y	Y	NM	Y	NM	Y	Y	?	Y	NM	CARMA2 blocks IMQ-induced psoriasis- like inflammatory circuit	31323190, 32343482, 32597759,33419765, 27071417, 29689250,27071417 (in vitro)
	Growth Factor (OE)		KRT14	Vegfa	K14-VEGF(Detmar homozygous) [Jax.org]	Detmar, M., et al., J Invest Dermatol, 1998.	9665379	1998	Mixed (FVB/N x C21) F1	Spontaneous	NM	NM	Y	NM	Y	?	NM	NM	NM	NM		20527041, 26220468, 24453242,24685902, 24895100, 27535613,25834349, 26383911, 21210341,18579711, 24485663
	Growth Factor (OE)		KRT14	Vegfa	K14-VEGF(homozygous) [Regeneron]	Xia, Y. P., et al., Blood, 2003.	12649136	2003	FVB	Wounding (young), spontaneous in old	Y	Y	Y	Y	Y	Y	NM	NM	?	N	Treatment with VEGF Trap improves phenotype	23732650, 19729876, 27527600,28414177, 26263167, 31024570,
	Growth Factor (OE)		KRT14	Vegfa	VEGF-TG(Detmar hemizygous)	Kunstfeld, R., et al., Blood, 2004.	15100155	2004	FVB	Elicited with Oxazolone	Y	Y	Y	Y	Y?	?	NM	NM	Y	NM		24695674, 25190364, 30518687,31406267,
	Growth Factor (OE)	tTAxTetoff	KRT5	Vegfa	KC-VEGF	Ward, N. L., et al., J DermatolSci, 2011.	21129919	2011	CD1	Spontaneous	N	N	Y	Y	N	N	NM	NM	NM	NM
	Growth Factor (OE)		KRT5	TGFB1	K5.TGFβ 1wt	Li, A. G., et al., EMBO J, 2004.	15057277	2004	Mixed (ICR X B6D-2)	Spontaneous	Y	Y	Y	Y	Y	Y	NM	NM	?	N	Alopecia	19357708, 21281802, 24714204,21483750, 23335955, 25132073,30423327, 21135506
	TF (OE)		KRT5	Stat3	K5.Stat3C	Sano, S., et al., Nat Med, 2005.	15592573	2004	FVB	Tape stripping (young), spontaneous in old	Y	Y	Y	Y	Y	Y	NM	Y	Y	N	Improved with TNF inhibitor, EGFR inhibitor, pre-treatment with STAT3 decoy (oligonucleotide)	21346238, 21483750, 20811392,23543761, 25384035, 25384035,25875168, 25268586, 33648938
Keratinocyte-specific Gene Mani	TF (OE)		KRT5	Stat3	K5.Stat3C:F759	Yamamoto, M., et al., J Invest Dermatol, 2015.	25268586	2015	FVB	Spontaneous	Y	Y	Y	Y	Y	Y	Y	Y	Y	Arthritis		10661409
	Receptor (OE)	tTAxTetoff	KRT5	Tek	KC-Tie2	Wolfram, J. A., et al., Am J Pathol, 2009.	19342373	2009	Original mice were on CD1 background, now on C57BL/6	Spontaneous	Y	Y	Y	Y	Y	Y	Y	Y	N	Spontaneous aortic root inflammation, shortened clotting times	Anti-IL-23, -IL-17A, -TNF or -IL-17RAmAb treatment improves skin phenotype and CVD; VEGF trap, clodronate liposomes, and cutaneous denervation improve skin phenotype; anakinra no effect, erlotinib worsens, IL-6 deletion no effect on skin but improves thrombosis	21070202, 21471984, 22572815,24005035, 25351201, 26223654,27942589, 28894119, 28951241,21483750
	Enzyme (OE)	tTAxTetoff	KRT5	Klk6	Klk6+	Billi, A. C., et al., J Clin Invest, 2020.	32155135	2020	C57BL/6J	Spontaneous	Y	Y	Y	Y	Y	Y	NM	NM	N	Dactylitis, enthesitis, arthritis, decreased bone mineral density, shortened clotting times	Skin transcriptome is more similar to PsA patients than those who do not have PsA, PAR1 dependent
	Signaling Molecule(OE)		KRT14	Rac1	Rac1 V12	Winge, M. C., etal., J Clin Invest, 2016.	27294528	2016	C57BL/6, BALB/cand CBA	Spontaneous	Y	Y	Y	Y	Y	Y	Y	Y	Y	Arthritis andvascular inflammation		29129599, 29321372
	TF (KO)	CreERxLoxP	KRT5	Junb-Jun	JunB𝚫ep*c-Jun𝚫ep*	Zenz, R., et al., Nature, 2005.	16163348	2005	(C57BL/6 X C3H)	Spontaneous	Y	Y	Y	Y	Y	Y	N, IL-17Fincreased	Y	N	PsA	Ciprofloxacin delayed phenotype onset and prevented PsA, mAbs targeting molecules within the IL-23 axis improve phenotype	19995970, 31870764,24714204,32484435, 25216727
	Non-native Cytokine (OE)	Cre-LoxP	KRT14	Il17a	K14-IL-17Aind/+	Croxford, A. L., et al., J Invest Dermatol, 2014.	24067382	2014	C57BL/6	Spontaneous	Y	Y	Y	Y	Y	Y	N	(Y)	N	Arthritis, uveitis, hypertension	Anti-IL-17A did not improve disease, no immunosuppression mentioned - but condition improved with anti-IL-6 mAb	25341795, 31622280, 31480330,30367871
	Cytokine (Inducible OE)	CreERxLoxP	KRT14	Il23	K23 mice; (K14CreERT2-p40-2A-p19fl/fl)	Chen, L., et al., Sci Rep, 2020.	32427877	2020	NM (most likely C57BL/6)	Spontaneous	Y	Y	Y	Y	Y	NM	NM	NM	NM	PsA, synovitis, enthesitis, dactylitis	Skin phenotype is IL-22 independent, loss of IL-22 worsens PsA-like phenotype
Elicited					Intradermal IL-23	Chan, et., al. 2006 J Exp med	17074928	2006	C57BL/6 and mixed	Intradermal IL-23 induced	Y	Y	Y	Y	Y	Y	N	Y	NM	NM	Anti-TNF antibody works partially, anti- IL-17A-antibody had no effect	See Supplementary Table S3
					Topical IMQ	van der Fits, L., et al., J Immunol, 2009.	19380832	2009	BALB/c, C57BL/6	IMQ induced	Y	Y	Y	Y	Y	Y	Y	NM	NM	NM	Splenomegaly, T cell depletion improves the phenotype, systemic effect, background specific differences	See Supplementary Table S2
					IL-23 minicircle	Adamopoulos et al., 2011 JI	21670317	2011	B10.RIII	Hydrodynamic delivery of IL-23 minicircle DNA	Y	Y	NM	Y	Y	NM	Y	NM	NM	Arthritis	Chronic arthritis, systemic bone loss, myelopoiesis in BM and spleen, hepatic necrosis, depletion of IL-17A-producingCD4+ T cells improves arthritis and bone loss. Neutrophil infiltration, skin inflammation is improved in CCR6–/– mice	See Supplementary Table S6
Other	T cells				Dsg3H1	Nishimoto, S., et al., J Immunol, 2013.	23956432	2013		CD4+ T cells recognize autoantigen, T celltransfer in Rag -/-background	Y	Y	N	Y	NM	Y	Y	NM	?	NM	Weight loss	21821914

For a complete list of all models, see Supplementary Table S1.

OE, Overexpression; KO, knockout; TF, transcription factor; Y, yes; (Y), yes but no difference from the control; N, no/not present; NM, not mentioned in the paper; ? Unknown; Immune cells, APCs/macrophages/monocytes/DCs; Other therapies, CsA/Methotrexate/steroids/UV/etc.; i.d., intradermal; ET, S. aureus superantigen exfoliative toxin (ET); PsA, psoriatic arthritis.

Unlike our current understanding of psoriasis pathogenesis, up until 1980, psoriasis was primarily considered as a keratinocyte disease due to the characteristic keratinocyte hyperproliferation and hyperplasia (Figure 1). Following the serendipitous discovery that cyclosporine cleared psoriasis plaques (Mueller and Herrmann, 1979), a paradigm shift occurred such that the focus of study switched to understanding contributions of the immune system to disease pathogenesis. Since then, research and clinical practice has focused on elucidating the pathogenic contributions of immune cell subsets and their derived factors, and how these interact with keratinocytes to promote disease. It’s likely not one cell type or another that is responsible for psoriatic disease, but a well-orchestrated series of events that are tightly coordinated amongst many cell types that reside in, and traffic into, the skin. Newly developed biologic therapies target these interactions and have changed clinical practice for treating psoriasis patients (Figure 1 (Brownstone et al., 2021)). Many of these clinical advances have been supported by preclinical animal model findings that have helped elucidate and confirm the role of individual immune cells and their derived factors and have dissected out the cellular and molecular mechanisms by which immune cells and keratinocytes interact to promote the development of psoriasiform skin inflammation (Figure 1).

Mouse models of psoriasis

Historically, a lack of appropriate preclinical psoriasis animal models hampered new insight into the pathogenesis of psoriasis and psoriasis comorbidities. As new genetic, cellular, and molecular knowledge emerged about human psoriasis pathogenesis the number of novel preclinical psoriasiform mouse models being developed dramatically increased. Since 1963, when the first animal model of “psoriasis-like” disease was published (Pearson, 1963, Pearson and Wood, 1963) almost 1,600 citations referring to “psoriasis animal model” have been published (Figures 1–2, Figure S1). These approaches model individual or sometimes multifactorial components of human psoriasis, but none-to-date replicate entirely the pathogenesis or complexity of the disease. The components of human psoriasis used here to define a psoriasis-like or psoriasiform phenotype in mice includes the presence of acanthosis (epidermal thickening), altered differentiation/proliferation of keratinocytes, increased angiogenesis, immune cell infiltrate that includes T cells (Th1 and Th17), myeloid cells (antigen presenting cells and macrophages), and Munro-like neutrophilic microabscesses (Figure 3). Psoriasiform models should be T cell and IL-23/IL-17A/TNF-dependent, respond to broad immune suppression, and have a phenotype that is dependent on immune cell activation (Table 1, Table S1, Table S4). For organizational purposes we have categorized each individual psoriasis model into one of five model types: spontaneous, xenograft, transgenic/knockout, intradermal injection of cytokines, and topical application of imiquimod. The PMID of the original paper describing the model, the year of publication, and additional publications using the same model in future years (PMIDs) are all presented in Table 1 and Tables S1–S7. Due to space limitations, most of the models mentioned within the text will not be cited beyond the Table PMID and the Supplemental References.

Figure 3. Comparison of psoriasis patient lesional skin with mouse psoriasiform skin.

H&E staining of lesional psoriasis skin and dorsal skin from Klk6⁺, KC-Tie2, and imiquimod (IMQ) models. Human skin is thicker than mouse skin. Acanthosis (epidermal thickening) is a key characteristic present in psoriasis patient and mouse model skin. Rete pegs occur only in psoriasis patient plaque (black stars), not mouse skin. Mouse skin has fur, more hair follicles, and hair follicle epidermis thickens in animal models of psoriasis (e.g., white stars). Mouse skin has a muscle layer called the panniculus carnosus (black arrow) not found in human skin. Psoriasis plaque and psoriasiform mouse skin contain dense immune cell infiltrate comprised of T cells, myeloid cells (not shown), and Munro-like neutrophilic microabscesses (grey arrows). Neutrophils are primarily contained to the microabscesses in plaque psoriasis and can be seen exiting the dermal vasculature (*). Psoriasiform mouse skin also contains neutrophilic microabscesses, however imiquimod skin has an abundance of neutrophils in the dermis. Scale bar = 100μm for each mouse image; scale bar = 50μm for each human image.

1. Spontaneous mouse models

As described in detail in our prior review (Gudjonsson et al., 2007) spontaneous mutations can occur naturally in mice resulting in a phenotype that models certain aspects of psoriasis and are described in detail in Table 1 and Table S1 (and see Figures 1–2). Briefly, homozygous Asebia (ab/ab; Scd1^ab/Scd1^ab) mutant mice develop scaly skin with alopecia, and histologically present with a thickened dermis, increased dermal angiogenesis, and prominent fibroblasts (Gates and Karasek, 1965). The dermal infiltrate contains mast cells and macrophages, but lacks T cells and neutrophils (Brown and Hardy, 1988). The Flaky skin mutant model (fsn/fsn; Ttc^fsn/Ttc^fsn; Table 1) (Sundberg et al., 1994) contains an autosomal recessive mutation on mouse chromosome 17 that leads to some characteristics of psoriasis, including increased inflammation, hyperkeratosis of the epithelium, and a positive Köebner reaction. Skin lesions contain neutrophils and are vascularized. The Sharpin^cpdm/Sharpin^cpdm model (called chronic proliferative dermatitis, cpd/cpd) is an autosomal recessive spontaneous mutation identified in inbred C57BL/KaLawRij animals (HogenEsch et al., 1993) and skin lesions are characterized by epidermal hyperplasia, hyper- and parakeratosis, focal necrosis of keratinocytes and a chronic persistent inflammatory presence including infiltration of eosinophils, mast cells, macrophages, but very few T cells. These models, although historically important, are not highly used for studying psoriasis pathogenesis, likely due to the lack of T cell infiltration in the skin.

2. Xenograft models

Xenotransplantation approaches began in the 1980’s (Figures 1–2, Table 1, Table S1), are uniquely translational and come closest to incorporating the complete genetic, phenotypic, and immunopathogenic processes of psoriasis (Boyman et al., 2004, Wrone-Smith and Nickoloff, 1996). This approach transplants uninvolved (non-lesional) or lesional plaque skin onto severely immunodeficient mice (Tables S1 and S7 and for review (Gudjonsson et al., 2007)). Xenotransplantation models demonstrated that psoriasis non-lesional skin was diseased and predisposed for development into lesional skin (Fraki et al., 1982, Krueger et al., 1981). For example, non-lesional psoriasis patient skin grafted onto the backs of AGR129 mice developed into lesional skin in a human skin-resident T cell dependent manner, identifying the importance of T cells in mediating plaque formation (Boyman et al., 2004). Similarly, injecting autologous superantigen-stimulated blood derived immune cells underneath grafted non-lesional skin on SCID mice also leads to the development of lesional skin (Wrone-Smith and Nickoloff, 1996). Xenograft approaches readily translate to human psoriasis (Table 1, Table S1), have provided mechanistic insight into contributions of skin resident vs. circulating immune cells to development of psoriasis, and are ideal for testing new drugs preclinically. However, their use is underutilized as they are technically challenging, require access to psoriasis patients willing to donate large keratome sheet of skin, or multiple punch biopsies and/or provide peripheral blood, and the collection of tissues and grafting need to occur quickly to prevent graft ischemia and rejection.

3. Genetically engineered mouse models

With the discovery of recombinant DNA technology and the advent of genetic engineering came new discoveries resulting from the development and study of genetically modified mice (Figure 1; (Brinster et al., 1981, Costantini and Lacy, 1981, Gordon and Ruddle, 1981). Paradigm shifting technological advances in genetic engineering, molecular genetics, Cre-LoxP (Orban et al., 1992, Sternberg and Hamilton, 1981), tTA-Tet^os (Furth et al., 1994, Gossen and Bujard, 1992) and CRISPR/Cas9 genome editing technologies (Jinek et al., 2012, Mojica et al., 1993) has resulted in the generation of thousands of novel transgenic animals targeting specific genes critical to human disease. In the field of psoriasis, these mice number in the hundreds (Figures 1–2, Table 1, Figure S1, Tables S1–S7). These innovative psoriasis models have been designed to overexpress (gain-of-function) or eliminate (loss-of-function) a gene of interest either globally or in a cell- or lineage-specific manner, often temporally controlled with exposure to tetracycline, tamoxifen or RU486. Other genetic approaches modulate known mutant alleles. Each of the resulting models, recapitulates some components of human psoriasis, although none of them thus far appear to capture the entirety of human disease. An inclusive list of all genetically modified animals that develop psoriasis-like phenotypes is presented in Table S1 and dates back to the first reported transgenic rat model (global overexpression of HLA-B27 and human β2m; B27/hβ2m transgenic rat) (Hammer et al., 1990). Studying genetically engineered psoriasis models are ideal for elucidating the direct effects of a single gene modulation and can be used to further refine signaling pathways and cell-cell interactions by backcrossing these models with other genetic knockout models in order to provide critical insight into disease pathogenesis.

3.1. Global gene overexpression and knockout

The ability to express or knockout a gene of interest in an animal has resulted in ~100 models that spontaneously develop a psoriasis-like phenotype (Table 1, Table S1, Figures 1–2). There are few global gene knockin or knockout models that develop a psoriasis-like phenotype, but those that do include: gain-of-function models, Pla2g2f^tg/+, LynΔN and the B27/hβ2m transgenic rat; and global knockout models, Act1^−/− (Traf3ip2), Integrin α_E^−/− (α_E^−/−, Cd103), IKK1^−/− (Chuk), Ikba^Δ/Δ (Nfkbia), ll1rn−/−, IL-17rc^−/− (Il17rc), TPP (−/−) (Zfp36) and IRF-2^−/− (Irf2) )(n.b. all models are formatted and written using the name given to them by the originating authors). The concept that genetically overexpressing a transgene in a mouse observed to be increased in psoriasis patient lesional (or non-lesional skin) may lead to the spontaneous development of skin inflammation intrinsically makes sense and demonstrates a proinflammatory role for these molecules. Similarly, development of psoriasiform skin inflammation following the elimination of a gene observed to be decreased in psoriasis patient provides evidence for a suppressive role for these genes and pathways. However, overexpression or elimination of any gene has limitations including possible non-physiologic expression levels and lack of natural counter regulatory mechanisms. Additional constraints include the inability to determine cause and effect between skin inflammation and development of distant organ comorbidities (i.e., arthritis, vascular inflammation) and can also be challenging at distinguishing cell-specific contributions to the phenotype, as the gene of interest has been deleted ubiquitously. This is evident in HLA-B27.hβ2m and Il1rn−/− psoriasis models which develop arthritis and arterial inflammation, two comorbidities associated with psoriasis.

To circumvent the limitations of global knockout or knockin approaches and to better determine the impact of a specific gene in a specific cell, or how the gene of interest within a specific cell type interacts with other cells within a single model, scientists began to engineer mice using cellspecific promoters to eliminate or express genes of interest. These methods allowed for identifying changes in molecules and cell types of interest in psoriasis patient skin, then modeling either their increased or decreased expression.

3.2. Cell specific over-expressing models

The ability to overexpress a gene of interest in a cell-specific manner to mimic the increased expression observed in psoriasis patient skin is an approach used by many investigators to model human psoriasis in mice. Traditional cloning methods allow for a cell-specific promoter to be cloned upstream of a gene of interest and gene expression begins at the time the endogenous promoter is expressed. Psoriasis models taking advantage of this approach include those using keratinocyte-specific promoters (K5, K10, K6, K14, involucrin and loricrin) to drive overexpression of cytokines (Il6, TNFA, Il1a, Ifng, Il12b, Il36a, Il36a/Il36rn, IL20), adhesion molecules (ITGB1, ITGA2-IGAB1, ITGA5-ITGAB1), growth factors (KGF, TGFB1, Tgfb1, BMP6, AREG, Vegfa, Tgfa, Angptl6), signaling molecules and transcription factors (MAP2K1, Stat3c, Rac1^v12, Mek1, Stat3c) and leptin (Figures 1–2, Table 1, Table S1). Each of these models develops a psoriasis-like phenotype containing various components of human psoriasis. However, each is limited in that if gene expression modulates the phenotype of the mouse significantly enough during embryonic development, embryonic or perinatal lethality can occur, limiting the utility of the model. Similarly, phenotypes can vary between different models expressing the same gene of interest depending on the number of transgene gene copies in the mouse genome or the site of integration, impacting gene expression levels. This phenomenon is observed in the VEGF and TGFβ overexpressing psoriasis models. For example, VEGF-TG (KRT14-Vegfa) hemizygous mice (Detmar et al., 1998, Kunstfeld et al., 2004) develop psoriasiform skin inflammation in response to topical oxazalone. K14-VEGF homozygous mice engineered by Regeneron (Thurston et al., 1999, Xia et al., 2003) develop skin inflammation in response to wounding in young animals, but spontaneously develop a psoriasis-like skin phenotype by 6 months of age. A more recent K14-VEGF mouse engineered by Cyagen also develops spontaneous skin inflammation in older mice, but requires imiquimod to elicit a response in young animals (Wang et al., 2015). Our group used a binary Tet^off approach (see below) to overexpress Vegfa under the control of the K5 promoter (Ward et al., 2011a). No double transgenic mice were born in the absence of doxycycline, demonstrating embryonic lethality. Therefore, gene expression was turned on postnatally, and transgenic KC-VEGF mice developed robust dermal angiogenesis, an influx of immune cells into the dermis and epidermis, but an absence of acanthosis and failed to thrive. One K14-VEGF model (Detmar et al., 1998, Kunstfeld et al., 2004) is now distributed through The Jackson Labs and is being used as a homozygous model. The variability in skin phenotypes, despite overexpressing the same gene may reflect gene expression levels, such that lower skin VEGF models require a skin stimulus to develop inflammation, or time to increase skin VEGF expression levels (age), at which time an inflammatory phenotype will develop, whereas higher skin VEGF expressing models die.

Traditional methods for engineering promoter-specific overexpressors has been largely replaced with the availability of binary mouse methods: tetracycline transactivator (tTA or rtTA) × tetracycline responder mice (Tet^os or pTRE; gene expression off or on, respectively, with exposure to doxycycline) or CreER (estrogen receptor)-floxed approach combined with tamoxifen. These approaches circumnavigate embryonic or early neonatal lethality and allow for the study of gene overexpression after development occurs. The strengths of the Tet-on/off approach is that gene expression can be controlled throughout an animal’s lifespan (turning the gene on or off at critical time points, after disease is achieved, only in adulthood, or after birth) and is reversible and can be done repeatedly. The K5-IL-17C, K5-TGFβ1, KC-VEGF, Tie2-Tek, KC-Tie2, tTA/mS100a7a15, and Klk6+ models use this approach (Figures 1–3, Table 1, Table S1,). Similarly, the tamoxifen inducible CreER-floxed “gene of interest” approach also allows for timing specific overexpression, however once Cre excision occurs, it is not reversible. Models using this approach include, K23 mice (K14-Cre^ERT2xp40–2A-p19^fl/f), which overexpress IL-23, and TGFβ1^ΔKC mice (K14-Cre^ERT2TGFβ1^fl/f) overexpressing TGFβ1. For these last two models, each has a “sister-model” using the more traditional approach of making transgenics (i.e., K14/p40, HK1.TGF-β1) and the phenotypes developed are not overlapping in either case, much like the VEGF model differences observed described above, demonstrating that the timing of gene expression, and level of gene expression contributes to the specific phenotypic traits (Table 1, Table S1). In contrast, keratinocyte overexpressing PLA2G2F transgenic mice (K14cre⁺Pla2g2f^tg/+) develop a similar phenotype as the global overexpressing PLA2G2F model (Pla2g2f ^tg/+) supporting a critical role for keratinocyte derived Pla2g2t in this system.

Two widely used psoriasis models use keratinocyte promoters to express genes not normally found in keratinocytes. The KC-Tie2 (K5 promoter) and the K14-IL-17A^ind/+ models (Table 1 and Table S1) each develop a psoriasiform phenotype that phenocopies human psoriasis. Although neither Tie2 nor IL-17A are normally expressed by keratinocytes, their ectopic expression in keratinocytes is sufficient to elicit a spontaneous psoriasiform phenotype. K14-IL-17A^ind/+ mice have increases in skin and circulating IL-17A, known to be critical for human psoriasis pathogenesis. These mice develop chronic skin inflammation, accompanied by uveitis, arthritis and vascular inflammation modeling, known comorbidities of psoriasis patients (Croxford et al., 2014, Karbach et al., 2014). However, unlike psoriasis patients, K14-IL-17A^ind/+ mice do not respond to targeted blockade of IL-17A, likely due to the genetic overexpression of large amounts of Il17a using the K14 promoter. K14-IL-17A^ind/+ animals do improve with IL-6 blockade, however IL-6 inhibition does not work in psoriasis patients. Mice engineered to ectopically express Il17a in dendritic cells (DC-IL-17A^ind/+) also develop a psoriasis-like phenotype (Table S1). Together, these IL-17A models demonstrate that increasing skin IL-17A levels, using non-T cell specific promoters leads to the development of skin inflammation. The increases in circulating IL-17A in K14-IL-17A^ind/+ mice also support a role for IL-17A in promoting vascular inflammation, arthritis, and inflammation of the eye. The KC-Tie2 model ectopically expresses the membrane bound tyrosine kinase receptor Tie2 in keratinocytes ((Wolfram et al., 2009); Table 1, Table S1, Figure 3). KC-Tie2 animals respond to TNFα, IL-23p40, IL-23p19, IL-17A, and IL-17RA inhibition with phenotype improvement (Li et al., 2018, Ward et al., 2011b), and do not improve following IL-6-elimination (Wang et al., 2016) or anakinra (recombinant IL-1 receptor antagonist (IL-1Ra)) treatment, and worsen with erlotinib (anti-EGFR) treatment (Ward et al., 2015), modeling what occurs in psoriasis patients, making this a translatable model. Old KC-Tie2 mice spontaneously develop aortic root vascular inflammation and form blood clots more quickly in an experimental thrombosis model and these comorbidities are skin inflammation dependent, as repressing skin inflammation (using doxycycline, (Wang et al., 2012)), removing IL-6 (Wang et al., 2016) or blocking IL-23/IL-17A cytokines (Li et al., 2018), attenuates these outcomes. Together, these models demonstrate chronic keratinocyte activation can elicit distant vessel injury likely by changing the circulating inflammatory cellular and systemic milieu. Other keratinocyte-specific overexpressing psoriasis models including Klk6+, K23 (K14Cre^ERT2-p40–2A-p19^fl/fl), Rac1^v12, K14-AREG, and K5-Stat3C:F759 also develop distant organ damage in their joints, tendons, and bones (i.e., PsA-like dactylitis, enthesitis, arthritis) subsequent to skin inflammation (Table 1, Table S1). In these mouse models, skin-derived inflammation, resulting from a secreted protein (Klk6, IL-23, AREGF, IL-17A) or an intrinsically activated keratinocyte (Stat3C) leads to systemic changes that result in tendon and joint inflammation and eventual bone erosions. These models are uniquely suited for exploring the mechanisms underlying the link between cutaneous psoriasis and PsA.

The abundance of models using keratinocyte promoters suggest that in mouse models at least, the keratinocyte plays an integral function to the development of skin inflammation. However, psoriasiform models also exist where T cells have been modulated to recognize autoantigen (Th17-Dsg3H1) and when transferred into an immunodeficient Rag2^−/− mouse develop a skin phenotype (Dsg3H1 mice), as do transgenic mice engineered to express dual MHC restricted TCR (TCR MM14.4) or to express Cd1b and CD1b/cand CD1b-autoreactive TCR (HJ1Tg) on an ApoE^−/− background (Table S1).

3.2. Cell-specific deletion

Cre-LoxP engineering strategies are most useful for the directed elimination of genes of interest. This approach allows for the refinement of understanding cell-specific contributions to observations identified in whole genome knockout animals i.e., Ikba^Δ/Δ (Nfkbia^−/−), TPP (−/−) (Zfp36^−/−), and Act1^−/− (Traf3ip2^−/−) mice (Table 1, Table S1). For example, keratinocyte-specific Zfp36 knockout mice (Zfp36^fl/flK14-Cre, Zfp36^ΔEP), but not dendritic cell- (Zfp36^ΔDC) or myeloid cell-specific knockout (Zfp36^ΔM) animals, phenocopy the global deficient Zfp36^−/− mice, including the development of psoriatic arthritis-like disease (Andrianne et al., 2017). Similarly, keratinocyte-specific Nfkbia knockout animals (Ikba ^K5Δ/K5Δ), but not T cell-specific (Ikba^LCKΔ/LCKΔ) phenocopy global deficient Ikba^Δ/Δ mice, except for the presence of neutrophilic microabscesses, suggesting a critical role for keratinocyte (and not T cell) IκBα signaling in promoting inflammation (Rebholz et al., 2007). Traf3ip2^−/− mice (called Act1^−/−) mice also develop spontaneous skin inflammation that is phenocopied following T cell-specific Act1 deletion (Lck-Cre⁺Act1^fl/−) consistent with T cell-Act1 signaling having a suppressive role (Wang et al., 2013). The same group produced a K5cre⁺ Act1^fl/− mouse; however, no data was presented on whether spontaneous epidermal hyperplasia occurred in this model (Wu et al., 2015). In contrast, Claudio et al., (Claudio et al., 2009) engineered a separate Traf3ip2-deficient animals (called CIKS^−/− used by Ha et al., (Ha et al., 2014) that did not develop spontaneous skin inflammation and showed less severe response to topical imiquimod (largely driven by keratinocyte response; Table S2), indicative of a proinflammatory role for Act1 rather than a suppressive role. Other cell-specific deletion models that develop psoriasis-like skin inflammation include SPT-c^−/− (KRT5) and Arpc4 e^−/− mice (KRT14; Table S1).

The ability to delete a cell-specific gene using an inducible approach (conditional knockout) following exposure to tamoxifen or RU486 is also used to circumvent lethality. Models taking advantage of this approach include the JunB^Δep*c-Jun^Δep* mouse, the K14-Cre/Ikk2^FL/FL mouse, and the Srf mutant model, (Table 1, Table S1). This approach provides greater control in deciding when to delete the gene of interest.

3.4. Other genetic models of psoriasis

Two other psoriasis mouse models that spontaneously develop psoriasiform skin inflammation are worth discussing and including the Card14 mutant mouse and the CD18 hypomorph model (Table 1, Table S1, Figure 1). Mutations in the Card14 gene have been detected in psoriasis and psoriatic arthritis patients (Jordan et al., 2012). The gain-of-function mutation Glu138Ala (Card14 ^E138A/+) and the Gln136-deleted mutation (Card14 ^ΔQ136/+, accidently obtained by the investigators) each develops spontaneous psoriasis-like skin inflammation mediated via enhanced activation of IL-23/IL-17A cytokine production (Mellett et al., 2018, Wang et al., 2018). These reports indicate that hyperactivation of Card14 is sufficient to drive the Th17-mediated psoriasis skin inflammation. CD18^hypoPL/J animals also develop spontaneous skin inflammation because of decreased CD18 expression. Interestingly, the phenotype only occurs on the PL/J background, and not in C57BL/6 mice.

4. Intradermal Cytokine injections

Beginning in 2006, intradermal injection of various cytokines identified as increased in lesional human psoriasis skin were determined to also elicit acute inflammation if injected into mouse skin (Figure 1). To a various degree, injection of IL-23, IL-12, IL-21, IL-36, IL-17C, IL-22, IL-17E, combination of cytokines, oncostatin-M and IL-33 have been used to elicit various characteristics of psoriasiform skin inflammation (Table 1, Figures 1–2, Table S4). These models are easy to use, can be combined with various knockout models, do not require a lot of time to complete, and are relatively inexpensive. However, the scope of phenotype is limited, the model itself is acute in nature, and the phenotype resolves on its own once the injections are stopped. Of the various cytokine injection models, IL-23 is the most used, with more than 50 different genetically engineered mice being injected with this cytokine (Figure 2). A summary of the impact of global gene deficiency, depletion of specific subsets of immune cells using diphtheria toxin inducible models, as well as cell-specific gene deletion and their effects on the skin phenotype are presented in Table S3. Of note are reports of different groups using the same gene deficient mice and observing different outcomes (i.e., Ccr6). These differences may reflect different methods, including the use of different IL-23 concentrations, different vendors for recombinant protein, different dosing regimens (i.e., daily vs. every other day), length of time mice are injected, as well as the time point in which tissue is examined. Differences in results may also reflect differences in the microenvironments of each of the animal facilities, and the microbiome environment of each vivarium, as C57BL/6 germ-free mice show improved phenotype compared to regular specific pathogen-free (SPF) animals (Zhu et al., 2017). The acute IL-23 model has provided insight into the contributions of unique genes, individual cells, and signaling pathways and how they coordinate to result in IL-23-mediated skin inflammation. In addition to injecting proinflammatory cytokines, the introduction of TWEAK and topical anisomycin have also been used to initiate a psoriasis-like inflammatory skin phenotype (Table S5).

5. Topical imiquimod

Imiquimod is a TLR7 agonist and since the seminal paper in 2009 by Van Der Fits et al., is by far the most heavily utilized acute mouse model of psoriasis-like skin inflammation (Figures 1–3, Table 1, Figures S1, Table S2). Topical imiquimod onto mouse ear or back skin results in a phenotype that overlaps with many human psoriasis characteristics (Table 1, Figure 3). We have previously reviewed the snowballing literature using imiquimod as a model of psoriasis (Hawkes et al., 2017) and summarized the strengths and weaknesses of this model. Briefly, the strengths include the ease of use, the acute nature of the inflammatory response, its convenience, and low cost. Limitations include unintended systemic consequence of topical treatment, its overuse with limited validation studies, unclear mechanisms of action, potential observer bias, histological misinterpretation, and that it recapitulates limited aspects of human psoriasis. We (Swindell et al., 2017) and others (Badanthadka et al., 2021) have also demonstrated strain specific differences in gene responses. Table S2 contains a list of more than 300 individual mouse lines (and their background strains) that have had imiquimod put onto their skin, and describes how outcomes change (improved, worsened or no change). However, it’s important to note that despite statements of “improvement” in skin inflammation, significant acanthosis and inflammation is often still present in histological images of skin. This is observed in signature pathogenic psoriasis cytokine knockout mice, including Il23a^−/−, Il12^−/−, Il22^−/−, Il17a^−/−, Tnf^−/− animals (Johansen et al., 2015, Kulig et al., 2016, Pantelyushin et al., 2012, Raaby et al., 2015, Tortola et al., 2012, Van Belle et al., 2012, van der Fits et al., 2009, Vinter et al., 2016, Yoshiki et al., 2014, Zheng et al., 2018), indicating a degree of caution in interpreting imiquimod findings. Indeed, Tlr7^−/− mice develop inflammation in response to topical imiquimod (Kataoka et al., 2018, Novoszel et al., 2021, Pantelyushin et al., 2012, Tanaka et al., 2018, Ueyama et al., 2014) demonstrating non-TLR7-mediated effects of topical imiquimod (Table S2). Inconsistencies are also observed in outcomes between different groups using similar genetic models, including, as examples, Ccr6^−/−, Mrp14 (S100A9)^−/−, Tlr7^−/−, Tcrd^−/−, Apoe^−/−, Il33^−/−, Ahr^−/−, Il1f10^−/−, Ltb4^−/− and langerin-DTR animals. This may reflect difference in mouse vivariums and microbiomes between institutes, different vendors from which imiquimod is obtained, and how outcomes are reported. Differences may also reflect the different time points at which acanthosis, gene expression, immune cell activation and infiltration are reported. Attention to these details is necessary for proper comparison and interpretation.

Although both intradermal IL-23 and topical imiquimod psoriasis models are each dependent to some degree on IL-23 signaling (Kulig et al., 2016, van der Fits et al., 2009, Zhu et al., 2017) it is interesting to note that not all mice respond similarly to these different stimuli. P2×7r^−/− mice have improved skin inflammation in response to IL-23 (vs. control mice) but not imiquimod (Diaz-Perez et al., 2018). Similarly, Mapk8^−/− and K14creMAPK8^fl/fl mice as well as Tlr7^−/−/Tlr9^−/− mice show improvement (vs. controls) following topical imiquimod but not intradermal IL-23 (Le et al., 2020, Tanaka et al., 2018)(Table S2–S3). Le and colleagues also demonstrate that T cell (LsyMcre) and dendritic cell (Itgxcre) Mapk8^fl/fl mouse skin does not change, demonstrating a critical role for keratinocytes and not immune cells in this model. The differing levels of improvement (or lack thereof) in similar genetic mice between IL-23 and imiquimod elicited approaches likely reflects gene-specific responses to TLRs and IL-23 stimuli.

As reviewed previously (Hawkes et al., 2017), some debate has occurred about imiquimod-mediated inflammation with respect to the effects of the vehicle alone (Walter et al., 2013) and whether ingestion of imiquimod contributes to disease (Grine et al., 2016). Moreover, whether the inflammatory skin phenotype elicited by imiquimod actually represents psoriasis has also been debated, as topical imiquimod onto non-lesional skin of psoriasis patients results in a transcriptionally distinct skin lesion compared to lesional plaque psoriasis (Vinter et al., 2015). Moreover, imiquimod-mediated inflammation is strongly dependent on the presence of γδT cells, and pathogenic T cells in psoriasis patient plaque have been shown to be αβT cells (Matos et al., 2017).

Thus, imiquimod offers a quick, easy, inexpensive approach for studying acute skin inflammation that models some aspects of human psoriasis, but also has distinct differences. Undoubtedly scientists will continue to utilize this model for studying the role of specific genes, cells, and cellspecific genes in modulating psoriasis-like skin inflammation, however some level of attention to interpretation and translation to human psoriasis should be considered when using this approach.

Combining approaches

The ability to backcross genetic models with other knockout models to define contributions of specific cell types (i.e., T cells, dendritic cells, macrophages) in addition to mating these mice with cell-specific diphtheria toxin mediated depletion have provided enormous insight into the roles of immune cells to the pathogenesis of murine skin inflammation. For example, clodronate liposomes have been used to deplete phagocytic cells in KC-Tie2 (Ward et al., 2011b) and K14-Cre/Ikk2^FL/FL (Stratis et al., 2006) models and have demonstrated phenotype improvement. Ikba^Δ/Δ (Rebholz et al., 2007) mice backcrossed to Rag2^−/−mice also develop a less severe phenotype, as do Rag2^−/− mice treated with imiquimod and Rag1^−/− mice intradermally injected with IL-21 (Table S2, Table S4), supporting a critical role for T cells in mouse pathogenesis. In contrast, Rag2 deficiency has no effect on the phenotypes of K14/IL-1F6, IL1F5^−/−mice (Blumberg et al., 2007) or in the IL-23 injection model (Hedrick et al., 2009). γδ T cell specific DTR mice used with topical imiquimod demonstrate the critical role for this cell type in mediating the skin phenotype (Sandrock et al., 2018). Similarly, CD11c-DTR mice treated with either imiquimod (Tortola et al., 2012) or intradermal IL-23 (Singh et al., 2016) each develop less severe inflammation. Backcrossing genetic models or injecting cytokine or topical imiquimod into mice deficient in Il23, Il17a, Tnf has confirmed the contributions of these cytokines in driving skin hyperplasia (Table S2, Table S4), as has treating chronic models with function blocking antibodies targeting these proinflammatory proteins (refer to remarks in Table 1 and Table S1). The ability to compare outcomes following imiquimod application onto the backs of various whole knockout and cellspecific cre-floxed (knockout) mice (i.e., Il17ra, (Moos et al., 2019)) also provides unique insight into individual cell-gene contributions in a model (Table S2). Other combination approaches take advantage of bone marrow chimera approaches and adoptive transfer of immune cells into acute or chronic psoriasis models.

Recently, the introduction of topical application of siRNA targeting newly identified genes of interest onto mouse models of psoriasis has begun. This approach can be combined with acute or chronic models of psoriasis, and is rapidly evolving, quick to accomplish, and although the costs of purchasing sufficient siRNA can be costly, the approach is still more efficient in terms of time and cost compared to engineering a new line of mice (Gao et al., 2021, Hsieh et al., 2015, Li et al., 2019, Li et al., 2016, Luan et al., 2019, Qu et al., 2021, Xiao et al., 2020, Xue et al., 2020, Yan et al., 2019, Yang et al., 2021, Zhang et al., 2018)(Table S2).

Importance of chronicity, background strain and environment

Psoriasis is a chronic skin condition and having an animal model that reproduces this protracted component of disease is important. A model that develops a sustainable phenotype provides some advantages for study, including the ability to study the mechanisms of development, the ability to interfere with cellular or molecular targets in a mouse with a sustained established phenotype and look for improvement, as well as the likelihood that mouse models with chronic skin inflammation are more likely to develop systemic and distant organ comorbid conditions (Golden et al., 2015). Almost all mouse models of psoriasis that develop arthritis, uveitis or cardiovascular disease are models where skin inflammation is chronic (Table 1, Table S1). Acute elicited models, like intradermal cytokine injections simply don’t achieve a sustained level of inflammation to cause distant disease, and although easy to use, and model some aspects of some pathways, they are short lived, and not reversible, although prevention of, and interference with, the development of the phenotype can be examined. Il-23 minicircle approaches (Adamopoulos et al., 2011) circumvent the limitations of acute IL-23 intradermal injection approaches and provide a convenient, method of long-term expression of IL-23 that not only leads to psoriasiform skin inflammation (Table 1, Table S6) but a robust model for studying the spondyloarthropathies related to PsA, including enthesitis, aortitis, uveitis, arthritis, and bone erosion (Reinhardt et al., 2016). To our knowledge, the acute imiquimod model does not result in cardiovascular comorbidities (Golden et al., 2015) or arthritis. Attempts to lengthen the treatment period of imiquimod out to 8 weeks by treating mice less frequently with imiquimod and covering less body surface area resulted in a model that resembles more closely lupus, rather than psoriasis (Yokogawa et al., 2014).

Mouse background strain has a clear impact on phenotype (Badanthadka et al., 2021, Swindell et al., 2017). The background strain of each model has been included within each table to reflect the differences in phenotype that can arise because of this. Obvious examples are CD18^hypoPL/J mice which only develop a phenotype on the PL/J background (Bullard et al., 1996), α_E^−/− mice (Integrin α_E^−/− (Schon et al., 2000) which lose their phenotype when crossed to BALB/C animals, and IL1rn−/− mice where a phenotype is also only seen in BALB/C and not C57Bl/6 background strains (Shepherd et al., 2004). Even different strains of C57Bl/6 (6N vs. 6J) mice affect imiquimod-elicited skin inflammation (Bezdek et al., 2017), such that C57Bl/6N mice develop less severe skin inflammation. Similarly, we now understand that microenvironment also can contribute to response to imiquimod. It is likely that different microbes and pathogens between different vivariums and between different rooms within the same vivarium can influence immune response and thus skin inflammation. Experimental depletion of the microbiome using topical antibiotics (Zanvit et al., 2015) or systemic oral antibiotics and metronidazole impacts the severity of phenotype that develops following imiquimod application (Stehlikova et al., 2019, Zakostelska et al., 2016, Zanvit et al., 2015)(Table S2) or western diet-elicited inflammation (Shi et al., 2020)(Table S5). Similarly, mice raised under germ-free conditions also develop less-severe skin inflammation in response to imiquimod (Zakostelska et al., 2016) and intradermal IL-23 injections (Zhu et al., 2017). However, others have reported that treating neonate mice with antibiotics made them more susceptible to imiquimod and IL-23 inflammation in adulthood (Zanvit et al., 2015). The differences observed because of timing of antibiotics (neonatally vs. in adulthood) highlight differential contributions of beneficial vs. harmful microbiota. Il1rn−/− mice develop less severe arthritis under germ-free conditions (Abdollahi-Roodsaz et al., 2008) and aberrant microbiota increases Th17 cell differentiation, demonstrating cross-talk between the immune systems and microbial communities of the gut and skin (Rogier et al., 2017). Microbiome contributions to psoriatic disease models is complex, likely involves the skin and gut microenvironments, and is an emerging area of research.

Translating mouse models of psoriasis – considerations

The intent of studying and developing mouse models of psoriasis and psoriasis comorbidities is to better understand the pathogenesis of human psoriasis. Therefore, the models must translate to human disease. However there are species-specific differences in anatomy and biology that need to be accounted for and include anatomical differences exist between mouse and human skin and include thickness of the skin, thickness of the epidermis (human skin and epidermis are much thicker than mouse skin), the lack of rete peg development in mice, the presence of the panniculus carnosus (a subcutaneous muscle layer in the skin of mice present only in rudimentary forms in humans) and the presence of fur (hence differences in hair follicle distribution) (Figure 3; for review, (Gudjonsson et al., 2007)). The lifespan of mice is significantly less than that of humans, limiting long term studies of aging and impact of chronic inflammation on comorbid conditions. Mouse metabolism also varies significantly from human metabolism (Perlman, 2016) making metabolic studies of skin cells or immune cells challenging to translate to psoriasis patients. Psoriasis patients represent a heterogenous group of people, while mice for the most part are inbred (and still show strain variability). There are critical differences between the immune systems of mice and humans that are directly impacted by many of the preclinical psoriasis mouse model approaches. For example, γδT cells are required for imiquimod- (Table S2), IL-23- (Tables S3 and S6), and western diet- (Table S5) psoriasiform skin inflammation and much work has focused on these cell types in mouse models, their biology, and immune functions in driving inflammatory responses. However, they are not critical IL-17A producing cells in human psoriasis and are unlikely to contribute to human psoriasis pathogenesis (Matos et al., 2017). Dendritic epidermal T cells (DETCs) are another immune cell only found in mice (Nielsen et al., 2017). Finally, as new microbiome studies evolve, it will become important to ensure that this work is recapitulated in psoriasis patients. As the intent in studying mouse models of psoriasis is to add new information to the field of human psoriasis and human immunology, it’s important to ensure and validate that the observations and interpretation made whilst studying an animal model of psoriasis are upheld in human disease. It is only when this occurs, does the field move forward in a meaningful manner, paradigm shifts can occur, and the long-term goal of disease remission and/or cure be achieved.

Looking forward

The future of psoriasis models looks positive. The ongoing study of existing models that develop comorbidities (arthritis and cardiovascular disease) that have gene expression limited to the skin should contribute new pathogenic insight into why comorbidities occur and whether skin inflammation causes distant disease. The number of innovative genetic approaches to delete, deplete or track tagged cells in vivo offer new methods for identifying the molecular and cellular mechanisms linking skin-initiated inflammation and distant organ damage. As new models continue to be developed around genetic loci identified from GWAS studies (Tsoi et al., 2015, Tsoi et al., 2017), new knowledge about specific gene mutations will be gained. Systems biology approaches combined with novel sequence analyses (bulk RNASeq, spatial RNASeq, scRNAseq, ATACseq), CyTOF, and new applied “omics” approaches will provide new targets for the development of new mouse models. The ability to integrate sequence and compare data from different murine models (skin, blood, lymph nodes, stool) with specific phenotypes, and perform comparative analyses with human psoriasis patient tissues (skin, blood, PBMCs, stool), may identify new pathways of pathogenesis and new biomarkers of disease, comorbidities, and drug responsiveness.