Inhibition of vascular endothelial growth factor‐A downregulates angiogenesis in psoriasis: A pilot study

Abstract Background Vascular Endothelial Growth Factor (VEGF)‐A‐mediated angiogenesis participates in the pathogenesis of psoriasis, thus inviting the hypothesis that anti‐VEGF‐A therapy could be beneficial in psoriasis. While anti‐angiogenic agents are used in oncology and ophthalmology, these therapeutic strategies remain unexplored for the management of psoriasis. Objective Our objective was to investigate ex vivo how VEGF‐A blockade impacts blood vessels, epidermis and immune cells in organ‐cultured plaque and non‐lesional skin from patients with psoriasis. Methods Skin biopsies from patients with psoriasis (n = 6; plaque and non‐lesional skin) and healthy controls (n = 6) were incubated with anti‐VEGF‐A monoclonal antibody (bevacizumab, Avastin®) or a human IgG1 isotype control for 72‐h in serum‐free organ culture. CD31/LYVE‐1, Ki‐67, and mast cell tryptase expression were assessed by quantitative immunohistomorphometry. VEGF‐A levels in plasma, PBMCs and skin culture supernatants were measured. Results Inhibition of VEGF‐A blocked all free VEGF‐A ex vivo, reduced blood vessel area and the number of blood vessel endothelial cells in plaques of psoriasis (*p < 0.05). The treatment effect correlated significantly with levels of VEGF‐A in organ culture supernatants (r = 0.94; *p < 0.05) from plaque skin and with plasma levels of VEGF‐A from patients with psoriasis (r = 0.943; *p = 0.017). Conclusions These ex vivo data are the first studies to objectively investigate the potential of VEGF‐A inhibition as a novel adjuvant treatment strategy for psoriasis. Taken together, our data encourage further investigation by clinical trial to explore whether downregulating pathological angiogenesis has clinical utility, especially in patients with severe psoriasis or those with elevated levels of VEGF‐A in plasma and/or skin.

in the development of lesions of psoriasis 10,11 and the clearance of lesions of psoriasis during therapy is accompanied by vascular normalisation. 12Vascular abnormalities are macroscopically visible and represent the Auspitz phenomenon, a characteristic feature of psoriasis. 13utaneous angiogenesis in patients with psoriasis is characterised by pronounced vascular dilation and elongation, in conjunction with enhanced vascular permeability. 14,15Vascular endothelial growth factor-A (VEGF-A), a potent pro-angiogenic cytokine, is highly expressed in the skin and plasma of patients with psoriasis, 16,17 promoting endothelial cell (EC) proliferation, migration, survival, increased vasodilation and permeability 9,18 ; as well as keratinocyte proliferation and differentiation. 19,20Two common single nucleotide polymorphisms (SNPs) of the VEGF-A gene, at −460 (rs833061) and +405 (rs2010963), determine VEGF-A production, both in health and disease. 21,22Moreover, individuals who are 'high VEGF-A producers', are significantly more likely to develop psoriasis and to manifest a severe psoriasis phenotype. 21espite preclinical evidence [23][24][25][26][27][28] and substantial evidence from case reports of resolution of psoriasis whilst on treatment with anti-VEGF-A therapies, [29][30][31][32][33][34][35] the effects of VEGF-A inhibition remain unexplored in psoriasis, thus raising the question of whether anti-VEGF-A therapy could be integrated into psoriasis management by repurposing licenced anti-VEGF-A biologicals.Therefore, we postulated that VEGF-A inhibition represents a novel adjuvant treatment strategy for psoriasis, particularly in patients who are 'high VEGF-A producers'. 36e interrogated this hypothesis by exposing organ-cultured plaque and non-lesional psoriatic skin to treatment with bevacizumab (Avastin®, Roche, Basel, Switzerland), a humanised monoclonal antibody against VEGF-A or isotype control IgG 1 ex vivo. 15,37he effects of bevacizumab ex vivo were assessed by quantitative immunohistomorphometry on blood vessels using double immunofluorescence staining CD31/ LYVE-1 (CD31 is an endothelial cell marker and LYVE-1 is a lymphatic-specific endothelial cell marker) to visualise the blood capillary network (CD31 + LYVE-1 -). 38,39The effects of VEGF-A inhibition in the epidermis were assessed using the proliferation marker Ki-67 and measuring keratin 6 expression, which is upregulated in psoriatic epidermis. 40Lastly, we quantified the levels of mast cell activation, which contributes to psoriasis pathogenesis and constitutes an important source of VEGF-A 20,41,42 ; and key immune mediators in the pathogenesis of psoriasis, CD4 + and CD8 + cells.This showed reduced blood vessel area and EC numbers in plaques of psoriasis following VEGF-A inhibition, which was especially pronounced in patients with high VEGF-A levels in plasma and/or in organ culture supernatant, thus supporting our working hypothesis.

| Volunteers
Twelve individuals (6 with psoriasis and 6 without psoriasis) were recruited.Potential participants were excluded if they had used topical treatment on the biopsied area within 4 weeks or systemic treatments within 12 weeks of enrolment in the study; or had inflammatory arthropathy.The study was approved by the UK Health Research Authority (15/NW/0585, 09/H/101143) and adhered to the Declaration of Helsinki Guidelines.All subjects gave written, informed consent.

| Skin sampling
Skin punch biopsies (3 mm) were collected from all donors.Those with psoriasis had two biopsies taken from plaques of psoriasis and two additional biopsies from non-lesional skin (at least 5 cm away from plaques).Anatomical site-to-site variations were minimised by biopsy of the same body site (buttock).A detailed description of skin sampling is provided in Supplementary Files S1.

What is already known about this topic?
� Vascular endothelial growth factor-A (VEGF-A)-mediated angiogenesis participates in the pathogenesis of psoriasis.

What does this study add?
� Vascular endothelial growth factor-A (VEGF-A) inhibition downregulates blood vessel area in psoriasis plaque skin ex vivo.� Downregulation of blood vessel area is more abrupt in patients with high levels of VEGF-A and severe disease.

What is the translational message?
� This raises the question whether patients with severe psoriasis and/or high levels of VEGF-A in skin/plasma could benefit most from VEGF-A blocking therapy. 1 mL of supplemented EpiLife media (MEPI500CA, Gibco, ThermoFisher Scientific) for 72 h.To stimulate T cells and restore their inflammatory activity in skin organ culture, a T-cell activation mix containing 1 μg/ml anti-CD3 (clone ICHT1, cat: MAB100), 1 μg/ml anti-CD28 (clone 37407, Cat: MAB342) and 50 ng/mL rh IL-23 (Cat:1290-IL; all from R&D Systems, Bio-techne, Minnesota, USA) was added to the culture as previously described. 43A detailed description of skin organ culture and treatment is provided in Supplementary Files S1.Concentration of bevacizumab and duration of culture was iterated from our optimising study as previously described. 44Biopsies were incubated with 0.8 mg/mL of bevacizumab or 0.8 mg/mL of IgG 1 isotype control (BioXCell, 2BScientific Ltd, Oxford, UK) for 72 h at 37°C and 5% CO 2 .Culture media was changed after 48 h.After 72 h, biopsies were embedded in optimal cutting temperature gel, snap frozen in liquid nitrogen and stored at −80°C until cryosectioning.

| Peripheral blood mononuclear cells culture
Venous blood was collected into EDTA-tubes and peripheral blood mononuclear cells (PBMCs) were isolated using density gradient media -Lymphoprep ™ (STEM-CELL Technologies Ltd, Cambridge, UK).Cells were incubated with lipopolysaccharide (LPS; L4391, Sigma-Aldrich Inc, Munich, Germany) at 1 ng/mL or 5 ng/mL, as previously described. 45Cell culture supernatant were harvested and stored at −20°C before VEGF-A assay.A detailed description of PBMC culture is provided in Supplementary Files S1.

| Quantification of VEGF-A levels
Levels of VEGF-A protein in PBMCs culture supernatant and in organ culture supernatant were measured using an enzyme-linked immunosorbent assay (ELISA; Duo set ELISA kit, R&D Systems Europe Ltd., Abingdon, UK).Plasma was separated from a portion of venous blood.A detailed description of plasma extraction is provided in Supplementary Files S1.Levels of VEGF-A in plasma were measured using an ELISA (Quantikine ELISA, R&D Systems, Abingdon, UK) designed to assay both VEGF-A 165 and VEGF-A 121 isoforms.Optical density was determined using a microplate reader (CLARIOstar® plus, BMGLabtech).

| DNA extraction and genotyping
Genomic DNA was isolated using QIAamp DNA blood midi kit (QIAGEN®, Crawley, UK).Genotyping of the −460 (rs833061) and +405 (rs2010963) VEGF-A SNPs 15,16,37 was performed using the SNP genotyping assay (Applied Biosystems, Foster City).A detailed description of DNA extraction and genotyping is provided in Supplementary Files S1.

| Statistical analysis
Data analysis was performed using GraphPad Prism v8.1.2.Non-normal distribution was assumed for continuous variables and results reported as median (interquartile range [IQR]), unless otherwise specified.The Wilcoxon matched pairs signed rank test was used to evaluate differences between control and treated samples for healthy, non-lesional and plaque skin.Independent two-sided t or Mann-Whitney U tests were used to compare continuous variables between both groups.All tests were two-sided and *p-value < 0.05 was considered statistically significant.Non-parametric Spearman's correlation was used to determine relationships between variables.

| Inhibition of VEGF-A with bevacizumab blocks free VEGF-A in organ culture
Skin samples were taken from 5 women and 1 man with psoriasis, with median age of 46.5 years (inter-quartile range ([IQR] 27.75), median age of psoriasis onset 24 years (IQR 17) and median PASI (psoriasis area and severity index) of 12.65 (IQR 9.8).50% (n = 3) had a family history of psoriasis.Skin samples were taken from 6 healthy controls with median age of 42.5 years (IQR 14.5; Table 1).Psoriasis plaque skin maintained its psoriatic phenotype ex vivo, including increased rete ridge length, epidermal area, epidermal hyperproliferation and keratin 6 expression compared to nonlesional and healthy skin (Figure S1).
First, we investigated whether bevacizumab effectively blocked VEGF-A, showing that VEGF-A was undetectable at 12 and 48 h in culture supernatant of healthy, non-lesional and plaque skin incubated with 0.8 mg/mL of bevacizumab (Figure 1a,b).At 12 h, there were significantly higher levels of VEGF-A in culture media of isotype-control treated plaque compared to isotype control-treated non-lesional (**p < 0.01) and healthy skin (**p < 0.01, Figure 1a).Therefore, bevacizumab effectively blocks free VEGF-A in human skin organ culture.

| VEGF-A inhibition does not significantly change epidermal read-out parameters
We assessed the effects of bevacizumab on the epidermis, finding that VEGF-A inhibition did not affect rete ridge length, epidermal length, epidermal area or keratin 6 expression in plaques ex vivo (Figure S1).After 3 days organ culture, the number of proliferating keratinocytes in the stratum basale (SB) was significantly higher in isotype control-treated plaque (16.74 [21.79]) compared to isotype control-treated non-lesional (0 [0.61]; **p < 0.01) and isotype control-treated healthy skin (0.68 [1.25]; **p < 0.01; Figure 3a,b).VEGF-A inhibition did not affect TUNEL expression in the epidermal SB (Figure S1).This suggests that short-term anti-VEGF-A therapy hardly impacts on key epidermal parameters, at least ex vivo.

| Anti-angiogenic response to bevacizumab correlates with levels of VEGF-A in the plasma and plaque skin
Finally, we asked whether differences in the antiangiogenic response to VEGF-A inhibition were determined by levels of VEGF-A in plasma or skin, disease severity, VEGF-A polymorphisms or PBMC production of VEGF-A.There were no differences in VEGF-A genotype frequencies at the +405 or −460 sites between our donors when comparing healthy volunteers with patients with psoriasis or between volunteers with severe psoriasis and volunteers with mild to moderate psoriasis (Tables S1 and S2).Thus, no association between genotype and response to anti-VEGF-A treatment could be identified.The VEGF-A plasma levels between all patients with psoriasis and healthy volunteers, overall, showed no significant differences (Figure 4a).However, patients with severe disease (PASI>10: n = 3), tended towards exhibiting higher VEGF-A plasma levels than patients with moderate disease (PASI<10; n = 3), although this did not reach significance (p = 0.1; Figure 4b).VEGF-A production by PBMCs did not differ between healthy volunteers and psoriasis patients (Figure S5), and VEGF-A production by PBMCs did not correlate with response to anti-VEGF-A treatment (Figure 4c).VEGF-A plasma levels of patients with psoriasis correlated well with VEGF-A levels measured in culture supernatant of plaques at 12 h (r = 0.94, *p < 0.05; Figure 4c).
Importantly, we noted a positive correlation between bevacizumab-induced blood vessel area downregulation in 3-day organ-cultured psoriasis plaques ex vivo with the VEGF-A plasma levels of psoriasis patients (r = 0.94, *p < 0.05; Figure 4c).Bevacizumab-induced blood vessel area downregulation in plaque also correlated with the levels of VEGF-A in culture supernatant of isotype control-treated plaque at 12 h (r = 0.89, *p < 0.05; Figure 4c).Despite the relatively low numbers that could be examined, these ex vivo/in vivo correlations suggest that the anti-angiogenic response to bevacizumab is greater in those with higher levels of VEGF-A in the skin and/or plasma.

| DISCUSSION
In this ex vivo pilot study, we demonstrate that VEGF-A inhibition by bevacizumab exerts significant antiangiogenic activity in plaques of psoriasis.In agreement with previous studies, 16,49 patients with severe psoriasis (PASI >10) had higher levels of VEGF-A in plasma and plaques of psoriasis.Despite the limited number of skin biopsies and donors that were available for study, our preclinical data encourage one to systematically explore clinically whether blocking VEGF-A is an effective adjunct management strategy for downregulating psoriasis-associated pathological angiogenesis.We therefore hypothesise that this treatment strategy may be most effective in patients with high levels of VEGF-A in the skin and/or plasma.
2][63] To the best of our knowledge, we describe here the first studies to explore a potential therapeutic effect in inflammatory skin disease, demonstrating that VEGF-A inhibition by bevacizumab exerts significant anti-angiogenic activity in skin organ culture of psoriasis.Manifest by significant diminution in blood vessel area and blood vessel ECs in plaques of psoriasis, F I G U R E 1 Bevacizumab blocks free VEGF-A in organ culture.VEGF-A levels were measured in culture supernatant of organ-cultured healthy, non-lesional and plaque skin after 12 and 48 h using an ELISA.VEGF-A was undetectable in organ culture supernatant of healthy, non-lesional and psoriasis plaque skin 12 and 48 h after incubation with bevacizumab.(a) At 12 h, the levels of VEGF-A in organ culture supernatant of isotype control-treated plaque were higher than those of isotype control-treated non-lesional (**p < 0.01) and isotype controltreated healthy skin (**p < 0.01).(b) At 48 h, the levels of VEGF-A in organ culture supernatant of isotype control-treated non-lesional skin (1726 [320]) were higher than those of isotype control-treated plaque (1338 [779.3];*p < 0.05).Number of independent experiments n = 36 (1 punch biopsy per patient and per treatment group).Data were presented as median and were analysed with two-tailed Wilcoxon matchedpairs signed rank test (paired analysis) or two-tailed Mann Whitney test (unpaired analysis).CTR, isotype control; H, healthy; NL, non-lesional; PLQ, plaque; T, treated with anti-VEGF-A monoclonal antibody.response to bevacizumab was significantly greater in patients with high levels of VEGF-A in plasma/plaque skin and appeared to offer most treatment benefit to those with severe psoriasis.These preclinical data encourage further exploration of whether downregulation of pathological angiogenesis represents an adjuvant management strategy for patients with psoriasis.Although we observed no histological resolution of psoriasis, this was expected, given the short incubation time (72 h).However, early knockdown of the blood vasculature and prolonged VEGF-A inhibition could induce downregulation of epidermal hyperplasia due to keratinocyte starvation.While further research is required to test this hypothesis, psoriasis management could be optimised in the future through downregulation of angiogenesis in conjunction with the simultaneous inhibition of key regulatory pathways in psoriasis such as IL-17/IL-23.
Long-term inhibition of VEGF-A may lead to undesired age-promoting effects 64,65 and, therefore, anti-VEGF-A therapy for the management of psoriasis may best be administered as repetitive pulse therapy.In addition, nanotechnology-based angiogenesistargeting strategies could be used to minimise the potential systemic complications associated with tissue ageing effects of blocking VEGF-A, as these can also F I G U R E 4 Blood vessel area downregulation correlated with the levels of VEGF-A in plasma and VEGF-A production in psoriasis plaque skin (a) The median level of VEGF-A in plasma of patients with psoriasis and healthy volunteers was 25.38 (range 13-81) pg/ml and 14.13 (3-71) pg/ml, respectively.(b) The levels of VEGF-A in plasma in volunteers with psoriasis with PASI<10 were not significantly different to those with PASI>10 (p = 0.10).(c) There was a positive correlation between blood vessel area downregulation in plaque skin and VEGF-A levels in culture supernatant of plaque skin at 12 h (r = 0.89, p < 0.05).Blood vessel area downregulation also correlated positively with VEGF-A levels in plasma of patients with psoriasis (r = 0.94; p < 0.05).There was a positive correlation between VEGF-A levels in culture supernatant of plaque skin at 12 h and VEGF-A levels in plasma of patients with psoriasis (r = 0.94; p < 0.05).Data were analysed using two-tailed Mann-Whitney test.Correlations were analysed using Spearman correlation coefficient (r).PBMCs, peripheral blood mononuclear cells.lead to renal and cardiovascular impairment. 66,67owever, anti-VEGF-A therapy could be beneficial for patients with psoriasis through a direct effect on both the disease and its systemic sequelae, [68][69][70] perhaps reducing the risk of psoriasis-associated vascular events (myocardial infarction and stroke), and cardiometabolic diseases (hypertension, diabetes mellitus and metabolic syndrome). 71,72e also show that VEGF-A inhibition increases the number of tryptase + -MCs in plaques.Since our skin organ cultures had no access to circulating MC progenitors, this could reflect a promoting effect on the differentiation of mature MCs from resident mast cell progenitor cells that we have previously shown to be present in human skin mesenchyme under organ culture conditions. 73,74Tryptase, a major secretory product of human MCs, is a marker of MC activation 75 and a mitogen for dermal ECs, 76,77 and the release of proinflammatory and pro-angiogenic mediators from MCs can decrease the efficacy of anti-angiogenic therapy in tumours. 78,79Therefore, it deserves further exploration whether the efficacy of an anti-angiogenic psoriasis management strategy can be further enhanced by combining it with inhibitors of MC degranulation.
A recent prospective cohort study identified a correlation between the degree of nail fold capillary (NFC) changes and the severity of psoriatic arthritis (PsA), suggesting that NFC abnormalities may be an indicator of PsA progression. 80Anti-VEGF therapy may not only be effective against psoriasis skin lesions, but may also inhibit the progression to PsA.
The study had limitations.First, a limited number of biopsies and donors were available and further investigations in a larger group of patients are required to identify associations between VEGF-A genetic signatures and response to treatment.However, our experimental protocol was carefully and uniformly performed from various viewpoints, thus informing future investigations.Second, the ex vivo nature of the study and lack of tissue perfusion prevented investigation of histological changes over a prolonged period of time.Nevertheless, skin organ culture allows instructive study the effects of VEGF-A inhibition on the vasculature, including all skin compartments and resident skin cells.Finally, skin organ culture may be intrinsically affected by a wound healing response at 48 h, and we observed levels of VEGF-A increasing, in all groups in keeping with this. 81However, despite this, bevacizumab effectively blocked all free VEGF-A in culture supernatant of healthy, non-lesional and plaque skin ex vivo after 72 h.
In conclusion, this pilot study provides proof-ofprinciple for the investigation of VEGF-A inhibition as an adjuvant management strategy to selectively target vascular pathology in psoriasis.This approach could be especially beneficial for patients who have high levels of VEGF-A, offering an opportunity to personalise management and complement current anti-cytokine strategies and other standard-of-care psoriasis therapeutics.
Demographic details and clinical characteristics of donors.