Macrophage Migration Inhibitory Factor in Psoroptes ovis: Molecular Characterization and Potential Role in Eosinophil Accumulation of Skin in Rabbit and Its Implication in the Host–Parasite Interaction

Psoroptes ovis, a common surface-living mite of domestic and wild animals worldwide, results in huge economic losses and serious welfare issues in the animal industry. P. ovis infestation rapidly causes massive eosinophil infiltration in skin lesions, and increasing research revealed that eosinophils might play an important role in the pathogenesis of P. ovis infestation. Intradermal injection of P. ovis antigen invoked massive eosinophil infiltration, suggesting that this mite should contain some relative molecules involved in eosinophil accumulation in the skin. However, these active molecules have not yet been identified. Herein, we identified macrophage migration inhibitor factor (MIF) in P. ovis (PsoMIF) using bioinformatics and molecular biology methods. Sequence analyses revealed that PsoMIF appeared with high similarity to the topology of monomer and trimer formation with host MIF (RMSD = 0.28 angstroms and 2.826 angstroms, respectively) but with differences in tautomerase and thiol-protein oxidoreductase active sites. Reverse transcription PCR analysis (qRT-PCR) results showed that PsoMIF was expressed throughout all the developmental stages of P. ovis, particularly with the highest expression in female mites. Immunolocalization revealed that MIF protein located in the ovary and oviduct of female mites and also localized throughout the stratum spinosum, stratum granulosum, and even basal layers of the epidermis in skin lesions caused by P. ovis. rPsoMIF significantly upregulated eosinophil-related gene expression both in vitro (PBMC: CCL5, CCL11; HaCaT: IL-3, IL-4, IL-5, CCL5, CCL11) and in vivo (rabbit: IL-5, CCL5, CCL11, P-selectin, ICAM-1). Moreover, rPsoMIF could induce cutaneous eosinophil accumulation in a rabbit model and increased the vascular permeability in a mouse model. Our findings indicated that PsoMIF served as one of the key molecules contributing to skin eosinophil accumulation in P. ovis infection of rabbits.


Introduction
Psoroptes ovis is a common ectoparasite of both domestic and wild animals worldwide and the causative agent of psoroptic mange, which causes an inflammatory skin disease correlated with significant morbidity and significant adverse effects on the economics of the animal industry [1]. This surface-living mite abrades host skin epidermis and deposits excretory-secretory proteins onto the host stratum corneum, resulting in skin Note: Highly conserved residues and motifs are highlighted in gray shade.
Highly conserved residues and motifs are highlighted in gray shade.

Figure 2.
Level of structural similarity between PsoMIF and host MIF. Schematic representation of PsoMIF and Oryctolagus cuniculus MIF (OcMIF) monomers (top) and trimers (bottom) with secondary structure elements shown in yellow (β-sheet), red (α-helix), and green (random coil). The right side shows the superposition of the backbone of monomeric and trimeric PsoMIF (blue) and OcMIF (red).

Western Blotting Analysis of rPsoMIF
Western blotting analysis in Figure 3 showed that rPsoMIF was specifically recognized by P. ovis-positive rabbit serum and the correspondent rat anti-PsoMIF IgG. However, P. ovis-negative rabbit serum and pre-immune rat serum could not be bound, which demonstrated the favorable reactivity of rPsoMIF. The purified rPsoMIF probed with serum from a rabbit naturally infected with P. ovis (Lane 1) and P. ovis-negative serum (Lane 2). (B) Lane M: protein molecular weight marker. The purified rPsoMIF probed with the rat anti-rPsoMIF-IgG (Lane 1) and negative rat serum before immunization (Lane 2). Figure 4 showed the PsoMIF transcription expression in all lifecycle stages mites and for "starved" as well as "fed" mites. Results revealed that PsoMIF mRNA was transcribed throughout the lifecycle of P. ovis, and female mites had the highest transcription level compared with male (p < 0.01), nymph (p < 0.01), and larval mites (p < 0.01). In addition, nymph and male mites showed a significantly higher transcription level than larval mites (p < 0.05, Figure 4A). However, there was no significant difference between nymph and Level of structural similarity between PsoMIF and host MIF. Schematic representation of PsoMIF and Oryctolagus cuniculus MIF (OcMIF) monomers (top) and trimers (bottom) with secondary structure elements shown in yellow (β-sheet), red (α-helix), and green (random coil). The right side shows the superposition of the backbone of monomeric and trimeric PsoMIF (blue) and OcMIF (red).

Western Blotting Analysis of rPsoMIF
Western blotting analysis in Figure 3 showed that rPsoMIF was specifically recognized by P. ovis-positive rabbit serum and the correspondent rat anti-PsoMIF IgG. However, P. ovisnegative rabbit serum and pre-immune rat serum could not be bound, which demonstrated the favorable reactivity of rPsoMIF.

Figure 2.
Level of structural similarity between PsoMIF and host MIF. Schematic representation of PsoMIF and Oryctolagus cuniculus MIF (OcMIF) monomers (top) and trimers (bottom) with secondary structure elements shown in yellow (β-sheet), red (α-helix), and green (random coil). The right side shows the superposition of the backbone of monomeric and trimeric PsoMIF (blue) and OcMIF (red).

Western Blotting Analysis of rPsoMIF
Western blotting analysis in Figure 3 showed that rPsoMIF was specifically recognized by P. ovis-positive rabbit serum and the correspondent rat anti-PsoMIF IgG. However, P. ovis-negative rabbit serum and pre-immune rat serum could not be bound, which demonstrated the favorable reactivity of rPsoMIF. The purified rPsoMIF probed with serum from a rabbit naturally infected with P. ovis (Lane 1) and P. ovis-negative serum (Lane 2). (B) Lane M: protein molecular weight marker. The purified rPsoMIF probed with the rat anti-rPsoMIF-IgG (Lane 1) and negative rat serum before immunization (Lane 2). Figure 4 showed the PsoMIF transcription expression in all lifecycle stages mites and for "starved" as well as "fed" mites. Results revealed that PsoMIF mRNA was transcribed throughout the lifecycle of P. ovis, and female mites had the highest transcription level compared with male (p < 0.01), nymph (p < 0.01), and larval mites (p < 0.01). In addition, nymph and male mites showed a significantly higher transcription level than larval mites (p < 0.05, Figure 4A). However, there was no significant difference between nymph and The purified rPsoMIF probed with serum from a rabbit naturally infected with P. ovis (Lane 1) and P. ovisnegative serum (Lane 2). (B) Lane M: protein molecular weight marker. The purified rPsoMIF probed with the rat anti-rPsoMIF-IgG (Lane 1) and negative rat serum before immunization (Lane 2). Figure 4 showed the PsoMIF transcription expression in all lifecycle stages mites and for "starved" as well as "fed" mites. Results revealed that PsoMIF mRNA was transcribed throughout the lifecycle of P. ovis, and female mites had the highest transcription level compared with male (p < 0.01), nymph (p < 0.01), and larval mites (p < 0.01). In addition, nymph and male mites showed a significantly higher transcription level than larval mites (p < 0.05, Figure 4A). However, there was no significant difference between nymph and male mites in PsoMIF expression (p > 0.05, Figure 4A). Additionally, the expression of PsoMIF in "starved" mites was 8.9 times higher than that in "fed" mites (p < 0.01, Figure 4B). male mites in PsoMIF expression (p > 0.05, Figure 4A). Additionally, the expression of Pso-MIF in "starved" mites was 8.9 times higher than that in "fed" mites (p < 0.01, Figure 4B).

Detection of Native PsoMIF in Adult Female Mites and P. ovis Skin Lesions
Immunofluorescence histochemistry was performed to detect the location of native PsoMIF using rat anti-rPsoMIF serum. Native PsoMIF was predominantly localized in the female mite ovary and oviduct with strong green fluorescence ( Figure 5A, a and b). In P. ovis skin lesions, native PsoMIF was mainly distributed in the stratum spinosum and stratum granulosum of the epidermis and even entered the basal layer in severe lesions (Figure 5B, c'), while no staining was observed in non-lesional skin ( Figure 5B, a'). The preimmune IgG showed no green fluorescence in female mites ( Figure 5A, N) and skin lesions ( Figure 5B

Detection of Native PsoMIF in Adult Female Mites and P. ovis Skin Lesions
Immunofluorescence histochemistry was performed to detect the location of native PsoMIF using rat anti-rPsoMIF serum. Native PsoMIF was predominantly localized in the female mite ovary and oviduct with strong green fluorescence ( Figure 5A(a,b)). In P. ovis skin lesions, native PsoMIF was mainly distributed in the stratum spinosum and stratum granulosum of the epidermis and even entered the basal layer in severe lesions ( Figure 5B(c')), while no staining was observed in non-lesional skin ( Figure 5B(a')). The pre-immune IgG showed no green fluorescence in female mites ( Figure 5A(N)) and skin lesions ( Figure 5B(b')).  Figure 4A). Additionally, the expression of Pso-MIF in "starved" mites was 8.9 times higher than that in "fed" mites (p < 0.01, Figure 4B).

Detection of Native PsoMIF in Adult Female Mites and P. ovis Skin Lesions
Immunofluorescence histochemistry was performed to detect the location of native PsoMIF using rat anti-rPsoMIF serum. Native PsoMIF was predominantly localized in the female mite ovary and oviduct with strong green fluorescence ( Figure 5A, a and b). In P. ovis skin lesions, native PsoMIF was mainly distributed in the stratum spinosum and stratum granulosum of the epidermis and even entered the basal layer in severe lesions (
To determine whether rPsoMIF had impacts on the eosinophil-related gene expression in allergic rabbit PBMC, qRT-PCR analysis was performed on cells. Results revealed that rPsoMIF significantly upregulated the relative expressions of chemokine CCL5 (RANTES) and CCL11 (eotaxin-1) compared with pET32a and blank medium control groups (p < 0.05 for CCL5; p < 0.001 for CCL11); however, rPsoMIF had only slight and insignificant upregulation of cytokines IL-4, IL-5, and IL-13 compared to those in control groups (p > 0.05) ( Figure 6A). 0.05, Figure S1). Thus, we chose 1 µg/mL rPsoMIF as the appropriate concentration for subsequent tests.
To determine whether rPsoMIF had impacts on the eosinophil-related gene expression in allergic rabbit PBMC, qRT-PCR analysis was performed on cells. Results revealed that rPsoMIF significantly upregulated the relative expressions of chemokine CCL5 (RANTES) and CCL11 (eotaxin-1) compared with pET32a and blank medium control groups (p < 0.05 for CCL5; p < 0.001 for CCL11); however, rPsoMIF had only slight and insignificant upregulation of cytokines IL-4, IL-5, and IL-13 compared to those in control groups (p > 0.05) ( Figure 6A).

rPsoMIF Significantly Upregulates Eosinophil-related Gene expression in HaCaT cells
A CCK-8 assay was carried out to determine the appropriate concentration of rPsoMIF in HaCaT cells. Four different concentrations of rPsoMIF (0.05, 0.1, 0.5, and 1 µg/mL) processed for 24, 48, and 72 h, respectively, all promoted cell proliferative rates. There was a significant proliferation rate in 0.1, 0.5, 1 µg/mL rPsoMIF (p < 0.001, Figure  S2A), and the highest proliferation rate occurred after 24 h. Therefore, 0.1 µg/mL rPsoMIF and 24 h were chosen in subsequent tests.
qRT-PCR analysis was used to evaluate whether rPsoMIF had impacts on the eosinophil-related gene expression in human immortalized-keratinocyte-line HaCaT cells. Compared to pET32a and blank medium control groups, rPsoMIF showed a significantly upregulated the relative expression level of chemokine CCL11, CCL5, and cytokines IL-3, IL-4, as well as IL-5 (p < 0.05 for IL-3; p < 0.001 for other target genes, Figure 6B). Among them, CCL11 showed the highest-level expression in rPsoMIF treatment group, and its expression was 27 times higher than in the blank medium control group. However, cytokine IL-13 showed no significant difference between rPsoMIF treatment and pET32a as well as the blank medium control treatments (p > 0.05, Figure 6B).

rPsoMIF Significantly Upregulates Eosinophil-Related Gene Expression in HaCaT Cells
A CCK-8 assay was carried out to determine the appropriate concentration of rPsoMIF in HaCaT cells. Four different concentrations of rPsoMIF (0.05, 0.1, 0.5, and 1 µg/mL) processed for 24, 48, and 72 h, respectively, all promoted cell proliferative rates. There was a significant proliferation rate in 0.1, 0.5, 1 µg/mL rPsoMIF (p < 0.001, Figure S2A), and the highest proliferation rate occurred after 24 h. Therefore, 0.1 µg/mL rPsoMIF and 24 h were chosen in subsequent tests.
qRT-PCR analysis was used to evaluate whether rPsoMIF had impacts on the eosinophilrelated gene expression in human immortalized-keratinocyte-line HaCaT cells. Compared to pET32a and blank medium control groups, rPsoMIF showed a significantly upregulated the relative expression level of chemokine CCL11, CCL5, and cytokines IL-3, IL-4, as well as IL-5 (p < 0.05 for IL-3; p < 0.001 for other target genes, Figure 6B). Among them, CCL11 showed the highest-level expression in rPsoMIF treatment group, and its expression was 27 times higher than in the blank medium control group. However, cytokine IL-13 showed no significant difference between rPsoMIF treatment and pET32a as well as the blank medium control treatments (p > 0.05, Figure 6B).

rPsoMIF Induces Skin Inflammation and Eosinophil Recruitment In Vivo
Given the finding in in vitro experiments that rPsoMIF can upregulate eosinophilrelated gene expression in PBMC and HaCaT cells, we further investigated whether the rPsoMIF is also important in skin eosinophil recruitment in rabbits.
We found that both rPsoMIF and P. ovis WE (whole extract) caused skin inflammation at the injection site, including obvious redness and swelling in a time-dependent manner, while PBS and pET32a treatments did not cause obvious skin reactions (Figure 7). Interestingly, P. ovis WE protein even became a scab at the injection site at 10 h after treatment. Similar effects were observed in histologic examination with obvious dermal telangiectasia and bleeding in rPsoMIF and WE treatments, while no obvious pathological changes in PBS and pET32a treatments ( Figure 8A) were observed. Next, we found increased recruitment of eosinophil to the skin in chromotrope-2R staining after rPsoMIF and WE treatments compared with PBS and pET32a treatments, and there was statistical significance ( Figure 8B). Surprisingly, the number of eosinophils at the injection skin site was significantly higher in rPsoMIF treatment than in WE treatment (p < 0.05, Figure 8B). Next, we investigated the expression of mRNAs for eosinophil-related genes in injection skin specimens from rPsoMIF, WE, PBS, and pET32a treatments. The expression levels of cytokine IL-5, chemokines CCL5, CCL11, and adhesion receptors P-selectin and ICAM-1 were significantly increased in rPsoMIF treatment compared to PBS and pET32a treatments (p < 0.05; p < 0.001, Figure 8C). In contrast, no significant difference in cytokines IL-4 and IL-13 between rPsoMIF treatment and PBS as well as pET32a treatments were detected (p > 0.05, Figure 8C). WE treatment had a statistically significant higher-level expression in all target genes in comparison with PBS and pET32a treatments (p < 0.001, Figure 8C). Given the finding in in vitro experiments that rPsoMIF can upregulate eosinophilrelated gene expression in PBMC and HaCaT cells, we further investigated whether the rPsoMIF is also important in skin eosinophil recruitment in rabbits.
We found that both rPsoMIF and P. ovis WE (whole extract) caused skin inflammation at the injection site, including obvious redness and swelling in a time-dependent manner, while PBS and pET32a treatments did not cause obvious skin reactions (Figure 7). Interestingly, P. ovis WE protein even became a scab at the injection site at 10 h after treatment. Similar effects were observed in histologic examination with obvious dermal telangiectasia and bleeding in rPsoMIF and WE treatments, while no obvious pathological changes in PBS and pET32a treatments ( Figure 8A) were observed. Next, we found increased recruitment of eosinophil to the skin in chromotrope-2R staining after rPsoMIF and WE treatments compared with PBS and pET32a treatments, and there was statistical significance ( Figure 8B). Surprisingly, the number of eosinophils at the injection skin site was significantly higher in rPsoMIF treatment than in WE treatment (p < 0.05, Figure 8B). Next, we investigated the expression of mRNAs for eosinophil-related genes in injection skin specimens from rPsoMIF, WE, PBS, and pET32a treatments. The expression levels of cytokine IL-5, chemokines CCL5, CCL11, and adhesion receptors P-selectin and ICAM-1 were significantly increased in rPsoMIF treatment compared to PBS and pET32a treatments (p < 0.05; p < 0.001, Figure 8C). In contrast, no significant difference in cytokines IL-4 and IL-13 between rPsoMIF treatment and PBS as well as pET32a treatments were detected (p > 0.05, Figure 8C). WE treatment had a statistically significant higher-level expression in all target genes in comparison with PBS and pET32a treatments (p < 0.001, Figure 8C).

rPsoMIF Enhances Vascular Permeability
Additionally, we evaluated whether rPsoMIF could affect the vascular permeability using the Miles assay by observing vascular leakage of Evans Blue dye in mouse skin. Figure 9 showed that rPsoMIF significantly increased the vascular permeability into skin subcutaneous tissue in comparison with 0.3% formalin, pET32a, and PBS (p < 0.05).

Discussion
Despite the increasing evidence that mammalian MIF contributes to the pathology in many inflammatory diseases [10], the role of P. ovis homolog of MIF in psoroptic mange pathology is poorly described. Here, we conducted a sequence comparison of PsoMIF and other organism MIFs and evaluated the transcriptional profile of PsoMIF and its tissue localization in Psoroptes mites and infected rabbit lesional skin. Particularly, we investigated the potential roles of rPsoMIF in skin eosinophil accumulation both in vitro and in vivo. Our results showed that PsoMIF appeared a high similarity in the topology of monomer and trimer formation with host MIF but with differences in tautomerase and hiolprotein oxidoreductase active sites. PsoMIF was transcribed by all developmental stages, and its native protein was mainly localized in the ovary and oviduct of female mites as well as in the epidermis of lesional rabbit skin. Furthermore, rPsoMIF enhanced the vascular permeability and significantly induced eosinophil accumulation in the skin. To our knowledge, this study is the first report of antigen molecules of Psoroptes mite inducing eosinophil accumulation in the skin lesions of psoroptic mange.
Despite limited homology, PsoMIF exhibited the same secondary structure distribution of two α-helices and five β-strand sequences (β1α1β2β3β4α2β5) with host MIF, and it appeared a similar topology of monomer and trimer formation with host MIF ( Figure  1A). However, there were differences in catalytic active sites between PsoMIF and host MIFs. PsoMIF possessed five out of six amino acid residues known as the conserved active site for MIF's tautomerase activity [19], suggesting that PsoMIF was accessible for interaction with tautomerase substrates. This speculation was supported by the result of enzyme activity assays in our previous study, revealing that rPsoMIF possessed tautomerase ac-

Discussion
Despite the increasing evidence that mammalian MIF contributes to the pathology in many inflammatory diseases [10], the role of P. ovis homolog of MIF in psoroptic mange pathology is poorly described. Here, we conducted a sequence comparison of PsoMIF and other organism MIFs and evaluated the transcriptional profile of PsoMIF and its tissue localization in Psoroptes mites and infected rabbit lesional skin. Particularly, we investigated the potential roles of rPsoMIF in skin eosinophil accumulation both in vitro and in vivo. Our results showed that PsoMIF appeared a high similarity in the topology of monomer and trimer formation with host MIF but with differences in tautomerase and hiol-protein oxidoreductase active sites. PsoMIF was transcribed by all developmental stages, and its native protein was mainly localized in the ovary and oviduct of female mites as well as in the epidermis of lesional rabbit skin. Furthermore, rPsoMIF enhanced the vascular permeability and significantly induced eosinophil accumulation in the skin. To our knowledge, this study is the first report of antigen molecules of Psoroptes mite inducing eosinophil accumulation in the skin lesions of psoroptic mange.
Despite limited homology, PsoMIF exhibited the same secondary structure distribution of two α-helices and five β-strand sequences (β1α1β2β3β4α2β5) with host MIF, and it appeared a similar topology of monomer and trimer formation with host MIF ( Figure 1A). However, there were differences in catalytic active sites between PsoMIF and host MIFs. PsoMIF possessed five out of six amino acid residues known as the conserved active site for MIF's tautomerase activity [19], suggesting that PsoMIF was accessible for interaction with tautomerase substrates. This speculation was supported by the result of enzyme activity assays in our previous study, revealing that rPsoMIF possessed tautomerase activity [17]. However, single-amino-acid mutations in active sites might significantly affect MIF enzymatic activity [20]. PsoMIF appeared a single-amino-acid difference at position 98 from mammalian MIFs, and this change could be related to lower levels of tautomerase activity in PsoMIF than that of human MIF [18] and might also account for mammalian MIF-specific antagonist Z950 only having a slight inhibitory effect on PsoMIF activity [18]. This dissimilarity in tautomerase active sites might allow for the development of specific inhibitors for parasite rather than host MIF, and this inhibitor targeting PsoMIF might act as a novel agent for therapeutic intervention [21]. Additionally, three out of four amino acid residues (CXXC motif) for MIF's thiol-protein oxidoreductase activity [22] were mutated in PsoMIF, possibly explained the lack of oxidoreductase activity of rPsoMIF in vitro [18].
Previous studies documented that parasite-derived MIF, including Plasmodium, Toxoplasma, and Ostertagia ostertagi, had an important role in their life cycle stage development [23][24][25]. Our transcriptional analysis data in this study supported this opinion. In the present study, we found that PsoMIF was expressed in all developmental stages of P. ovis. Interestingly, both gene transcript and native protein of PsoMIF were in high abundance at the female stage mites, particularly this native protein was localized to the ovary and oviduct of female mite, indicating that this MIF might be associated with female mite reproduction physiology. Additionally, MIF has been reported to enhance Myzus percicae and Acyrthosiphon pisum feeding [26], so we suggested that the high abundance expression of MIF in female mites might be related to this stage need for feeding a large amount of food to acquire nutrients before oviposition [26]. Additionally, MIF appeared with a higher transcript expression in starved mites, suggesting that PsoMIF may be beneficial for enhancing P. ovis survival under malnourished conditions [23,27] or seeking host or initiating feeding behavior [28], and that this similar function of MIF could increase parasite survival under adverse nutrient stress condition has been confirmed in Trichomonas vaginalis [27]. Several studies reported that MIF was clearly detectable in excretory/secretory products of the parasite [25,29], and the MIF was secreted by the parasite into host tissue and blood during infection [30]. We found a very similar outcome in this study, where PsoMIF was released by Psorotpes mites during infection, and this view was supported by the following evidence in our study: (i) anti-PsoMIF antibodies were detected in P. ovis-positive serum but not in P. ovis-negative serum; (ii) native PsoMIF protein was detected throughout the stratum spinosum, stratum granulosum, and even basal layers of the epidermis in lesional skin. Interestingly, the predominant cell type in the epidermis is the keratinocyte [31], and this cell was the first contact point between Psoroptes mites and host immune response [32]. So, we hypothesized that P. ovis secreted native PsoMIF to contact with keratinocyte to influence the cytokine secretion and thus induce a cutaneous inflammatory response leading to the host's skin lesions. As expected, the injection skin showed redness and swelling within 5 min after rPsoMIF injection (Figure 7), and these skin symptoms were similar to the early signs of P. ovis infection [33]. Additionally, these results suggested that P. ovis might secrete native PsoMIF to inhibit host macrophage migration into skin infection sites [13], thus allowing this mites escape detection by host's innate immune system.
Under normal conditions, eosinophils are present almost exclusively in the gastrointestinal mucosa but not in most other healthy tissues [34]; however, under certain disease conditions, eosinophil scan selectively accumulates in any tissue of the body [34]. Eosinophils develop in the bone marrow from CD34+ pluripotent progenitor stem cells and differentiate into unique eosinophil-lineage-committed progenitors (EoPs), and then these cells develop into mature eosinophils in the bone marrow [35]. Mature eosinophils are released from the bone marrow and mobilize to the peripheral blood, then traffic rapidly from blood to tissue (including roll, adhere, and trans-endothelial migration) [36]. This multistage process of the abnormal accumulation of eosinophils in tissue is exquisitely regulated by cytokines IL-3, GM-CSF, and IL-5 [37], adhesion molecules P-selectin and intercellular adhesion molecule 1 (ICAM-1), and vascular-cell adhesion molecule 1 (VCAM-1) [38], as well as chemoattractants IL-4, IL-5, IL-13, CCL5, and CCL11 [39,40]. So, in this study, we investigated these relative molecules in vitro and in vivo. In in vitro study, the significantly increased expression of IL-5 was found in HaCaT cells but not in PBMC after exposure to rPsoMIF, and this cytokine is the most specific factor in all aspects of eosinophil proliferation, differentiation, migration, activation, and survival in tissues [38]. In addition to IL-5, IL-3 was also significantly upregulated in rPsoMIF-treatment of HaCaT cells, and this cytokine, combined with IL-5, has an additive effect on eosinophil colony production from the bone marrow [39,40]. Meanwhile, IL-4 showed a significant upregulation in rPsoMIF-treatment HaCaT cells, and this cytokine in cooperation with IL-5 promotes eosinophil differentiation and survival [36]. Many studies showed that the increased expression of IL-4 and IL-5 can induce the increased production of C-C chemokines (CCL5, CCL11, MCPs), and these chemokines are implicated in eosinophil trafficking from blood to tissue [39,40]. As expected, cells both in vitro (PBMC and HaCaT cells) and in vivo (rabbits) after rPsoMIF-treatment exhibited significant upregulation of CCL5 and CCL11. CCL5, along with IL-5, are specific selective factors for eosinophils and cooperatively promote recruitment of eosinophils into the local tissue. CCL11 involved in early recruitment of eosinophils into the tissue. Additionally, rPsoMIF can upregulate the adhesion receptor P-selectin and ICAM-1 in rabbits, and these molecules mediated eosinophil transendothelial migration [41]. However, it is noteworthy that P-selectin and ICAM-1 are adhesion molecules not only for eosinophil but also for neutrophils [42] and T lymphocytes [43]. In particular, rPsoMIF induced higher expression levels of all target genes in HaCaT cells than in PBMCs cells, which suggested that native MIF secreted by Psoroptes mites mainly affects the cytokines and chemokine secretions of keratinocytes, leading to the accumulation of cutaneous eosinophils at the site of infection. Interestingly, rPsoMIF caused a stronger accumulation of eosinophils than in WE proteins of P. ovis, illustrating that PsoMIF was one of the key proteins involved in promoting eosinophil accumulation in P. ovis infestation.
It is well documented that increased vascular permeability leads to edema and favors leukocyte transendothelial migration [44]. Thus, we evaluated the vascular permeability of rPsoMIF by the Miles assay in a mouse model. Our study revealed that rPsoMIF evoked increased vascular permeability leading to the serous and lymph exudations into the skin, and these exudations supply sufficient food for P. ovis [45]. Meanwhile, the increased vascular permeability by rPsoMIF potentially leads to increased eosinophil transendothelial migration from the blood vessel to skin [44], and this speculation has been confirmed in vivo in rabbits. An earlier study reported the detection eosinophils within the midgut of P. ovis [46], so we suggested eosinophils in the skin may serve as a food source for driving the Psoroptes mite population growing at the infected site. This similar beneficial role of eosinophils in the parasite survival has been characterized in Trichinella spiralis, revealing that eosinophils was necessary for Trichinella spiralis larvae growth in skeletal muscle [47]. On the other hand, eosinophils are also demonstrated to attack and kill parasites [48]. Taken together, eosinophils have a paradoxical effect on parasites, so further research is needed to elucidate the potential role of eosinophils in the pathogenesis of Psoroptes mite infection.

Sequence Comparison of Orthologous MIF from Psoroptes ovis and Other Organisms
Orthologous MIF genes from P. ovis and six other mites and three species mammals were extracted from GenBank and performed multiple sequence alignment by Jalview version 2.11.1 (Dundee, UK) and the online BLASTp tool (https://blast.ncbi.nlm.nih.gov/ Blast.cgi, accessed on 1 June 2022). Phylogenetic tree was inferred using the neighborjoining (NJ) method with MEGA version 7.0 through 1000 bootstrap replicates and Poisson correction method. The protein secondary and third structure of PsoMIF were predicted using Jpred 4.0 (http://www.compbio.dundee.ac.uk/jpred/, accessed on 20 June 2022) and SWISS Model (https://swissmodel.expasy.org/interactive, accessed on 25 June 2022), respectively. The predicted three dimensional structures were furtherly evaluated their ve-racity by SAVES v6.0 (https://servicesn.mbi.ucla.edu/SAVES/, accessed on 28 June 2022). Predicted tertiary structures were visualized and aligned using PyMol software.

Mite Collection, Preparation of Recombinant PsoMIF (rPsoMIF), and Whole Mite Protein
A rabbit infestation with Psoroptes ovis var. cuniculi was maintained at the Department of Parasitology, Sichuan Agricultural University (Chengdu, China). The collection of Psoroptes mites from ear scabs was performed as described previously by Zhang et al. [49]. Mites from the different life-stage (larva, nymph, adult male, and female mites) were distinguished according to the morphological characteristics described [50]; then, the mites of each lifecycle stage and the mixed population pooled with each life-stage mite were respectively harvested. The "fed" and "starved" P. ovis mites were collected essentially according to the protocol of McNair [28]. Briefly, the mixed-stage "fed" mites were immediately harvested from the freshly ear scabs of rabbit. The freshly mites were placed in a cardboard sprayed with sterilized saline and incubated at 25 • C for 4 days under 80-90% relative humidity, then the live mites were harvested as the "starved" mites.
Each life-cycle stage mite and "fed" and "starved" mites were used to extract total RNA using the EASYspin RNA micro kit with DNase I (Aidlab Biotechnologies Co., Ltd., Beijing, China) for the following qRT-PCR analysis. The whole mite protein (WE) was extracted from 100 mg stage-mixed P. ovis with BBproExtra™ total protein extraction kit (BestBio, Shanghai, China) for in vivo study on New Zealand rabbit. Additionally, the adult female mites were also involved in immunolocalization analysis.
The purification of rPsoMIF was obtained as previously described by Zheng et al. [18]. rPsoMIF was treated using Endotoxin Removal Kit (Smart-Life Sciences Biotechnology Co., Ltd., Changzhou, China) to remove endotoxin.

Polyclonal Antibody Production and Western Blotting Analysis
Anti-serum against rPsoMIF was raised in rats by subcutaneously injecting with 300 µg rPsoMIF emulsified in an equal volume of Freund's complete adjuvant (Sigma, St. Louis, MO, USA). Following the first inoculation, three 300 µg boosters of rPsoMIF emulsified in Freund's incomplete adjuvant (Sigma, USA) were applied at 7-day intervals. Serum was collected before inoculation and 4 days after the forth inoculation. Anti-rPsoMIF IgG was purified using Protein G affinity chromatography (Protein G Resin FF Prepacked Column, L00681, GenScript, Nanjing, China).
The purified rPsoMIF was subjected to 10% SDS-PAGE and eletrophoretically transferred onto a polyvinylidene difluoride (PDVF) membrane for Western blotting analysis. The membranes were blocked with 5% (w/v) skimmed milk in phosphate-buffered saline (PBS) for 2 h at 37 • C, then washed three times (5 min/wash) in TBST (0.02 M Tris-HCl, pH 7.6, 0.15 M NaCl, 0.05% Tween-20) and incubated overnight with the primary antibody (P. ovis-positive rabbit serum, P. ovis-negative rabbit serum, rat anti-rPsoMIF-IgG, and pre-immune serum) at 1:200 dilution in PBS at 4 • C, respectively. After washing three times in TBST, the membrane was incubated with horseradish-peroxidase (HRP)-conjugated goat anti-rabbit IgG or goat anti-rat IgG (1:3000 dilution in PBS; ABclonal Technology Co., Ltd., Wuhan, China) for 2 h at room temperature. Following four washes with TBST, the signal was detected using an Enhanced HRP-DAB Chromogenic Substrate Kit (Tiangen Biotech Co., Ltd., Beijing, China). For immunolocalization analysis, a protocol was performed as previously described from Gu et al. [51]. In brief, the adult female mites were fixed in 0.8% molten. The 8 mm punch biopsy was harvested using a biopsy punch (HEAD boitechnology Co., Ltd., Beijing, China) from the advancing border of the ear lesional skin of a rabbit naturally infested by P. ovis var. cuniculi following the administration of an anesthetic (JiaBo pharmaceutical Co., Ltd., Qingyuan, China). Another 8 mm punch biopsy as the control was collected from the ear skin of a healthy rabbit. Both mites and skin biopsies were fixed in 4% polyformaldehyde, then embedded in paraffin and sliced into 5 µm sections. The sections were incubated with rat anti-rPsoMIF IgG or pre-immune IgG (1:200 dilution in PBS) at 4 • C overnight. After four washes with PBS, sections were incubated with fluorescein isothiocyanate (FITC)-conjugated goat anti-rat IgG (ABclonal Technology Co., Ltd., Wuhan, China; 1:2000 dilution in PBS) at 37 • C 30 min. Then, the sections were visualized by a fluorescent microscope (BX53; Olympus Corp., Tokyo, Japan).

In Vitro Treatment of Rabbit PBMC
The rabbit PBMC was isolated from the blood of four allergic rabbits with the Ficoll-Hypaque centrifugation method (Solarbio Co., Ltd., Beijing, China). The allergic rabbits was performed according to previously described procedures [52]. Briefly, rabbits severely infested with P. ovis var. cuniculi were subcutaneously injected with ivermectin. Thirty days later, these treated rabbits were for clinical trials. The freshly isolated rabbit PBMC was re-suspended to a density of 1 × 10 4 cells/mL in RPMI 1640 medium containing 1% penicillin/streptomycin solution and 10% fetal bovine serum (FBS) and seeded in a 96-well culture plate (Corning, New York, NY, USA). Cells were treated with endotoxin-free rPsoMIF protein (0, 0.01, 0.05, 0.1 and 1µg/mL) for 24 h at 37 • C in a 5% CO 2 atmosphere. 10 µL/well of cell counting kit-8 (CCK-8) solution (Meilunbio, Dalian, China) were added for 1 h, and the optical density (OD) at 450 nm was detected using a microplate reader (Multiskan Spectrum, Thermo Fisher scientific, Waltham, MA, USA).
In subsequent experiments, the freshly isolated PBMC (1 × 10 6 cells/well) was seeded in 12-well plates and treated with the optimum concentration of endotoxin-free rPsoMIF by CCK-8 assay. The pET32a protein or blank medium were set as control groups. Three biological replicates were applied to each group. After incubation for 24 h, cells were harvested for qRT-PCR analysis
For investigation the effect of rPsoMIF on cell viability, we plated HaCaT in 96-well plates at 5 × 10 3 cells per well in DMEM-10 and cultured them for 1 day until reaching approximately 50% confluence, when the media was replaced with DMEM-0.5. The following day, the cells were stimulated with 5 µg/mL LPS for 24 h, and then the different concentrations of rPsoMIF (0, 0.05, 0.1, 0.5, and 1 µg/mL) were added. After incubation for 24 h at 37 • C in a 5% CO 2 atmosphere, CCK-8 solution (10 µL/well) (Meilunbi) was added for 1 h to measure cell viability. The optical density (OD) at 450 nm was measured using a microplate reader (Multiskan Spectrum).
After analysis of HaCaT viability, we seeded HaCaT into a 24-well plate at the density of 1 × 10 4 cells/well. After stimulation with LPS (5 µg/mL) for 24 h, the optimum concentration of endotoxin-free rPsoMIF (0.1 µg/mL), which was screened by CCK-8 analysis, was added to each well for another 24 h. An equivalent concentration of pET32a protein or an equal volume of medium was both as the control groups. Three biological replicates were performed for each group. The cells were collected for qRT-PCR analysis.

In Vivo Treatment of New Zealand Rabbits
Four allergic rabbits with a history of psoroptic mange were prepared for the present experiment as previously described [52]. The back of four allergic rabbits was shaved, and four injection locations were marked on the shaved back section. The distance between two injection sites were spaced at least 5 cm apart. Each site was injected intradermally with a volume of 0.1 mL of endotoxin-free rPsoMIF (1 mg/mL) or pET32a (1 mg/mL) or the whole mite extraction (1 mg/mL) or 0.01 M PBS (pH 7.0). The representative photographs of the skin injection site were recorded at four different time points (0 min, 5 min, 2 h, and 10 h). Rabbits were sacrificed at 10 h after injection; then, 8 mm-diameter skin tissue centered on the injection site was harvested for qRT-PCR and histological observation.

qRT-PCR
Total RNA was extracted from PBMC or HaCaT cells or rabbit skin samples by using RNA extraction kit (Foregene, Chengdu, China). qRT-PCR was performed for total RNA from each lifecycle stage mites or "fed" mites or "starved" mites or PBMC or HaCaT cells or rabbit skin samples. Primer sequences used in this study are listed in Table S1. Gene transcription was assessed by LightCycler ® 96 System (Roche, Basel, Switzerland) using the Real Time PCR EasyTM-SYBR Green I (Foregene, Chengdu, China). Each sample was performed in triplicate. Relative gene expression levels were examined by the 2 −∆∆CT method. Relative mRNA levels were normalized to the housekeeping gene β-actin for mite, skin, and PBMC analysis [53], and the YWHAZ gene was normalized for HaCaT analysis [53].

Skin Histopathology
All skin tissues were fixed in 4% paraformaldehyde and then buried in paraffin blocks. The tissues were sliced into 5 µm sections for pathological analysis with hematoxylin-eosin staining and for eosinophils counts by Chromotrope-2R staining [54]. The stained slides were scanned with Olympus VS120 S6 Slide scanner (OLYMPUS, Tokyo, Japan). The measurements were carried out using ImagePro Plus 6.0. The number of eosinophils in scanned files was counted from 9 consecutive square areas (each at 0.09 mm 2 area) for each sample [54].

Vascular Permeability Assay
Vascular permeability assay in 7-week-old BABL/c mice (n = 6) was performed as previously described [55]. Under anesthesia, mice were injected 100 µL Evans Blue dye (0.5% w/v in PBS) into the tail vein. After 30 min, mice were injected intradermally on the back marked sites (20 µL final volume) with (i) PBS (0.01 M, pH 7.0), (ii) formalin 0.3% in PBS, (iii) 20 µg of endotoxin-free rPsoMIF in PBS, or (iv) 20 µg of endotoxin-free pET32a in PBS. After 30 min, animals were euthanized and the area of skin that contained the entire injection region were removed and photographed to document leakage of the Evans Blue dye. The Evans Blue dye in the skin was extracted with 250 µL of 50% formamide overnight at 55 • C, and absorbance was measured at 620 nm after centrifugation at 4000 rpm for 10 min.

Statistical Analysis
All data were recorded as mean ± standard error (SE). All statistical analyses were performed with SPSS Statistics 27.0. Transcriptional analysis and eosinophil counts were evaluated by one-way analysis of variance (ANOVA) with Tukey-Kramer test. Vascular permeability was calculated by the fold difference of readings from permeability-inducing agent versus vehicle control. p-value < 0.05 was considered as statistically significant.

Conclusions
In conclusion, PsoMIF was expressed throughout all the developmental stages of P. ovis, particularly with the highest expression in female mites. Native PsoMIF was located on the ovary and oviduct of female mites and secreted into skin lesions. rPsoMIF significantly enhanced eosinophil-related gene expression both in vitro and in vivo, and this protein could induce cutaneous eosinophil accumulation in rabbits and increased the vascular permeability in mice. These findings suggested that PsoMIF served as one of the key molecules contributing to skin eosinophils accumulation during P. ovis infection in rabbits.