2023 article
Histopathological characterisation of trunk-dominant canine pemphigus foliaceus, and comparison with classic facial and insecticide-triggered forms
Gedon, N. K. Y., Bizikova, P., Olivry, T., Mendoza-Kuznetsova, E., Oberkirchner, U., Robertson, J. B., & Linder, K. E. (2023, June 14). VETERINARY DERMATOLOGY.
co-author countries:
United States of America 🇺🇸
author keywords: acantholysis; autoimmune diseases; dog diseases; pathological findings; pemphigus; skin diseases
onlinelibrary.wiley.com (publisher PDF)
Source: Web Of Science
Added: July 3, 2023
While the clinical features were described recently, the histopathological characterisation of trunk-dominant canine pemphigus foliaceus (PF) is lacking, and whether it differs from classic facial or insecticide-triggered PF is unknown. This study describes the histopathological findings of trunk-dominant PF, and compares the results to classic facial and insecticide-triggered PF. Skin biopsies from 103 dogs with clinically characterised trunk-dominant (n = 33), classic facial (n = 26) and insecticide-triggered PF (n = 44) were included. Histological sections, randomised and blinded, were scored for over 50 morphological parameters of pustules, epidermis, dermis, adnexa and crusts. Intact pustule area and width were measured by digital microscopy. In trunk-dominant PF, 77 intact pustules were predominantly subcorneal (0.0019–1.940 mm2 area, 0.0470–4.2532 mm wide), and contained from one to over 100 acantholytic keratinocytes. Pustules had boat acantholytic cells, corneocytes, perinuclear eosinophilic rings, neutrophil rosettes, acantholytic cell necrosis, rafts, cling-ons and/or eosinophils. Peripustular epidermal spongiosis, necrosis and lymphocyte exocytosis occurred, as did follicular pustules. Mixed dermal inflammation often contained eosinophils. Trunk-dominant PF did not differ from the other PF groups except for few parameters, such as having fewer rafts (p = 0.003). Additional autoimmune inflammatory patterns occurred in all PF groups. Trunk-dominant PF and other canine PF variants are histologically similar, which indicates shared pathomechanisms. The identification of common boat acantholytic cells and corneocyte separation has implications for the mechanisms of acantholysis. The diversity of histopathological and polyautoimmunity features support complicated immune mechanisms. Finally, results indicate that diagnostic biopsies cannot differentiate between these PF variants in dogs. 虽然最近对其临床特征进行了描述，但缺乏躯干显性犬落叶型天疱疮(PF)的组织病理学特征，也不知道它是否与经典的面部或杀虫剂引发的PF不同。 本研究描述了躯干显性PF的组织病理学发现，并将其结果与经典的面部和杀虫剂触发PF进行了比较。 包括103只具有临床特征的躯干显性(n=33)、典型面部(n=26)和杀虫剂触发PF(n=44)的犬的皮肤活检。 组织学切片，随机和盲法，对脓疱、表皮、真皮、毛囊附件和结痂的50多个形态学参数进行评分。通过数字显微镜测量完整的脓疱面积和宽度。 在躯干显性的PF中，77个完整的脓疱主要位于角质层下(面积0.0019–1.940 mm2，宽度0.0470–4.2532 mm)，包含1至100多个棘层松懈性角质细胞。脓疱有连片的棘层松懈细胞、角化细胞、核周嗜酸性粒细胞环、中性粒细胞花环、棘层松懈细胞坏死、筏状排列、紧靠和/或嗜酸性粒细胞。出现脓疱周围表皮海绵状水肿、坏死和淋巴细胞外排，毛囊性脓疱也是如此。混合性真皮炎症通常含有嗜酸性粒细胞。躯干显性PF与其他PF组没有差异，除了一些参数，例如筏状排列更少(p=0.003)。所有PF组都出现了额外的自身免疫炎症模式。 躯干显性PF和其他犬PF变体在组织学上相似，这表明共同的病理机制。常见连片棘层松懈细胞和分离角化细胞的鉴定，牵连出棘层松懈的机制。组织病理学和多重自身免疫特征的多样性支持复杂的免疫机制。最后，结果表明，诊断性活组织检查无法区分犬的这些PF变体。 Bien que les caractéristiques cliniques aient été décrites récemment, la caractérisation histopathologique du pemphigus foliacé (PF) canin à dominant tronculaire fait défaut, et on ne sait pas s'il diffère du PF facial classique ou induit par un insecticide. Cette étude décrit les résultats histopathologiques provenant de PF à dominante tronculaire et les compare à ceux de PF faciaux classiques et induits par un insecticide. les biopsies cutanées de 103 chiens présentant une PF à dominante tronculaire (n = 33), faciale classique (n = 26) et déclenchée par un insecticide (n = 44) sont étudiées. Des coupes histologiques, randomisées et en aveugle, sont évaluées pour plus de 50 paramètres morphologiques relatifs aux pustules, à l'épiderme, au derme, aux annexes et aux croûtes. La surface et la largeur des pustules intactes sont mesurées par microscopie numérique. Concernant les PF à dominante tronculaire, 77 pustules intactes sont principalement sous-cornéennes (surface de 0,0019 à 1,940 mm2, largeur de 0,0470 à 4,2532 mm) et contiennent de un à plus de 100 kératinocytes acantholytiques. Les pustules contiennent des cellules acantholytiques en radeau, des cornéocytes, des halos éosinophiles périnucléaires, des rosettes de neutrophiles, une nécrose des cellules acantholytiques, des radeaux, des amas et/ou des éosinophiles. Une spongiose épidermique péripustuleuse, une nécrose et une exocytose lymphocytaire sont observées, ainsi que des pustules folliculaires. L'inflammation cutanée mixte contient souvent des éosinophiles. Le PF à dominante tronculaire ne différe pas des autres groupes de PF, à l’exception de quelques paramètres, comme le fait de présenter moins de radeaux (p = 0,003). D’autres patterns inflammatoires auto-immuns ont été observés dans tous les groupes de PF. le PF à dominante tronculaire et les autres variants canins de PF sont histologiquement similaires, ce qui indique des mécanismes pathologiques communs. L'identification commune de cellules acantholytiques en radeau associée à la disruption des cornéocytes a des implications en termes de mécanisme acantholytique. La diversité des caractéristiques histopathologiques et de polyauto-immunité argue en faveur de mécanismes immunitaires complexes. Enfin, les résultats indiquent que l’examen des biopsies ne permet pas de différencier ces variants de PF chez le chien. Während die klinischen Merkmale jüngst beschrieben wurden, fehlt eine histopathologische Charakterisierung des Rumpf-dominierten Pemphigus foliaceus (PF) des Hundes, ebenso ist die Unterscheidung zum klassischen Gesichts- oder Insektizid-ausgelöstem PF unbekannt. Diese Studie beschreibt die histopathologischen Befunde des Rumpf-dominierten PF, und vergleicht die Ergebnisse zum klassischen Gesichts- und Insektizid-ausgelösten PF. Hautbiopsien von 103 Hunden mit klinisch beschriebenem Rumpf-dominierten PF (n = 33), klassischem Gesichts- (n = 26) und Insektizid-ausgelöstem PF (n = 44) wurden inkludiert. Histologische Schnitte, zufällig ausgewählt und geblindet, wurden auf über 50 morphologische Parameter wie Pusteln, Epidermis, Dermis, Adnexe und Krusten untersucht. Bei intakten Pusteln wurde die Stelle und der Durchmesser mittels Digitalmikroskopie gemessen. Beim Rumpf-dominierten PF traten intakte Pusteln vor allem subcorneal auf (0,0019-1,949 mm2; 0,0470-4,2532 mm Durchmesser) und enthielten zwischen einem und über 100 akantholytische Keratinozyten. Die Pusteln hatten Boot-ähnliche akantholytische Zellen, Korneozyten, perinukleäre eosinophile Ringe, neutrophile Rosetten, akantholytische Zellnekrosen, zusammenhängende Zellen, einzeln anhängende Zellen und/oder Eosinophile. Es trat eine peripustulöse epidermale Spongiose, Nekrose und Lymphozyten Exozytose auf, sowie follikuläre Pusteln. In der gemischten dermalen Entzündung traten auch häufig Eosinophile auf. Der Rumpf-dominierte PF unterschied sich nicht vom PF der anderen Gruppen außer in einigen wenigen Parametern, wie weniger zusammenhängende Zellen (p = 0,003). Zusätzliche autoimmune entzündliche Merkmale traten in allen PF-Gruppen auf. Der Rumpf-dominierte PF und andere PF-Varianten des Hundes sind histologisch ähnlich, was einen gemeinsamen Pathomechanismus deutlich macht. Die Identifizierung sogenannter akantholytischer Zellen in Boot-Form und eine Korneozyten Separierung hat auch eine Bedeutung im Bezug auf die Mechanismen der Akantholyse. Die Vielfalt der histopathologischen und polyautoimmunen Merkmale weist auf komplizierte immune Mechanismen hin. Letztendlich zeigen diese Ergebnisse, dass die diagnostischen Biopsien nicht zwischen den PF-Varianten des Hundes unterscheiden können. 体幹優位型犬落葉状天疱瘡(PF)の病理組織学的特徴は最近報告されているが、古典的な顔面型や駆虫薬誘発型PFと異なるかどうかは不明である。 本研究では、体幹優位型PFの病理組織学的所見を述べ、古典的顔面型PFおよび駆虫薬誘発型PFと比較することであった。 体幹優位型PF(n = 33)、古典的顔面型PF(n = 26)、駆虫薬誘発型PF(n = 44)の臨床的特徴を有する犬103頭の皮膚生検が対象となった。 組織切片を無作為に盲検化し、膿疱、表皮、真皮、付属器、痂皮の50以上の形態学的パラメータについてスコア化した。膿疱の面積と幅は、デジタル顕微鏡で測定した。 体幹優位型PFでは、77個の膿疱は主に角層下(面積0.0019-1.940mm2、幅0.0470-4.2532mm)で、1~100個以上の角化細胞が含まれていた。膿疱は、ボート状の棘融解細胞、角質細胞、核周囲の好酸性リング、好中球ロゼット、棘融解細胞壊死、ラフト、クリングオン、好酸球を有していた。毛包の膿疱と同様に、膿疱周囲表皮の海綿状変化、壊死、リンパ球のエクソサイトーシスが起こった。真皮の混合性炎症はしばしば好酸球を含んでいた。体幹優位型PFは、ラフトが少ない(p = 0.003)といったいくつかのパラメータを除いて、他のPFグループと差がなかった。自己免疫性炎症パターンは、すべてのPF群で発生した。 体幹優位型PFと他の犬PF変種は組織学的に類似しており、これは病態メカニズムを共有していることを示した。共通のボート状棘融解細胞および角質細胞の分離の特定は、棘融解のメカニズムを示唆していた。病理組織学的および多自己免疫学的特徴の多様性は、複雑な免疫機構を支持した。最後に、診断用生検では、犬におけるこれらのPF変種を区別することができないことが示された。 Enquanto as características clínicas foram descritas recentemente, ainda falta a caracterização histopatológica do pênfigo foliáceo (PF) canino predominante no tronco e não se sabe se ele difere do PF foliáceo facial clássico e do desencadeado por inseticida. Este estudo descreve os achados histopatológicos do PF predominante no tronco e compara os resultados com o PF clássico e com o desencadeado por inseticidas. Foram incluídas biópsias cutâneas de 103 cães com PF caracterizado clinicamente como predominante no tronco (n = 33), facial clássico (n = 26) e desencadeado por inseticidas (n = 44). Cortes histológicos randomizados e cegos foram classificados por mais de 50 parâmetros morfológicos de pústulas, epiderme, derme, anexos e crostas. A área e a largura das pústulas intactas foram mensuradas por microscopia digital. No PF predominante no tronco, as pústulas intactas foram principalmente subcórneas (0,0019–1,940 mm2 de área, 0,0470–4,2532 mm de largura), e continham de um a mais de cem queratinócitos acantolíticos. As pústulas apresentavam células acantolíticas em formato de barco, corneócitos, anéis eosinofílicos perinucleares, rosetas neutrofílicas, necrose de células acantolíticas, aglomerados de células acantolíticas, grupos de células do estrato granuloso aderidas ao estrato córneo (cling-on) e/ou eosinófilos. Ocorreram espongiose epidérmica peripustular, necrose e exocitose de linfócitos, bem como pústulas foliculares. No infiltrado inflamatório dérmico misto frequentemente continha eosinófilos. O PF predominante no tronco não diferiu dos outros grupos de PF exceto para poucos parâmetros, como por exemplo apresentar menos aglomerados de células acantolíticas (p = 0.003). Os outros padrões inflamatórios autoimunes ocorreram em todos os grupos de PF. O PF predominante no tronco e as outras variantes do PF canino são histologicamente similares, o que indica a existência de patomecanismos comuns. A identificação frequente de células acantolíticas em formato de barco e separação de corneócitos tem implicações nos mecanismos de acantólise. A diversidade as características histopatológicas e poliautoimunes corroboram com os mecanismos imunológicos complexos. Finalmente, os resultados indicam que as biópsias diagnósticas não são capazes de diferenciar essas variantes de PF em cães. si bien las características clínicas se describieron recientemente, falta la caracterización histopatológica del pénfigo foliáceo (PF) canino dominante en el tronco, y se desconoce si difiere del PF facial clásico o desencadenado por insecticidas. Este estudio describe los hallazgos histopatológicos de la PF predominante en el tronco y compara los resultados con PF facial clásico y el PF desencadenado por insecticidas. Se incluyeron biopsias de piel de 103 perros con PF predominante en el tronco clínicamente caracterizado (n = 33), PF facial clásico (n = 26) y el desencadenado por insecticida (n = 44). Las secciones histológicas se puntuaron para más de 50 parámetros morfológicos al azar y de forma ciega para las características de pústulas, epidermis, dermis, anejos y costras. El área y el ancho de la pústula intacta se midieron mediante microscopía digital. En la PF dominante de tronco 77 pústulas intactas fueron predominantemente subcorneales (0,0019–1,940 mm2 de área, 0,0470–4,2532 mm de ancho) y contenían de uno a más de 100 queratinocitos acantolíticos. Las pústulas tenían células acantolíticas agrupadas, corneocitos, anillos eosinófilos perinucleares, rosetas de neutrófilos, necrosis de células acantolíticas, acantolisis en línea, células acantoliticsas adheridas y/o eosinófilos. Se produjo espongiosis epidérmica peripustulosa, necrosis y exocitosis de linfocitos, así como pústulas foliculares. La inflamación dérmica mixta a menudo contenía eosinófilos. La PF de tronco dominante no se diferenció de los otros grupos de PF excepto por algunos parámetros, como tener menos células en línea (p = 0,003). Se produjeron patrones inflamatorios autoinmunitarios adicionales en todos los grupos de PF. El PF dominante en el tronco y otras variantes del PF canino son histológicamente similares, lo que indica mecanismos patogénicos compartidos. La identificación de células acantolíticas en grupos y la separación de corneocitos tiene implicaciones para los mecanismos de acantólisis. La diversidad de características histopatológicas y de poliautoinmunidad respalda mecanismos inmunitarios complicados. Finalmente, los resultados indican que las biopsias de diagnóstico no pueden diferenciar entre estas variantes de PF en perros. Trunk-dominant pemphigus foliaceus (tPF) is a clinical variant of canine pemphigus foliaceus (PF) that presents without dorsal muzzle/nasal planum involvement and, sometimes, without footpad involvement typically seen in classic facial PF (fPF).1 The trunk-dominant patterning of skin lesions and the acantholytic pustular reactions seen on cytological and histological evaluation also occur in canine superficial bacterial pyoderma, thus making this variant of PF more difficult, time-consuming and expensive to diagnose.1 Although the clinical and immunological features of tPF were recently described and compared with fPF,1 little is known about its histological features. Additionally, it is unclear if there are histopathological differences between tPF and the other clinical PF variants recognised in dogs that could suggest potential pathomechanisms of the lesion formation, explain the clinical variation, or help in differentiating these conditions. Therefore, the purpose of this study was to (i) describe the histopathological features of canine tPF and (ii) compare the histopathological features of tPF with those of classic fPF and insecticide-triggered PF (iPF). Skin biopsy specimens from client-owned dogs with active tPF and fPF collected for standard-of-care, routine diagnostic assessment between January 2004 and March 2022 were selected for the study (North Carolina State University and Cummings School of Veterinary Medicine at Tufts University). Client consent to use the diagnostic samples for study purposes was obtained. Case inclusion criteria and the majority of PF patients were clinically described in a previously published study.1 Briefly, patients had a pustular skin disease with cytological confirmation of keratinocyte acantholysis and lesion distributions of tPF or fPF.1 No potential drug or other trigger was identified. For fPF, symmetrical lesions affected the dorsal muzzle/nasal planum, with or without other body areas involved. For tPF, skin lesions affected the body, and not the dorsal muzzle/nasal planum. Dogs had a negative bacterial culture and/or clinical signs did not respond to oral antibiotics, complete remission was achieved with immunosuppressive therapy, and 3-month follow-up information was available. The type and dose of immunomodulatory drugs received during the 2-week period before biopsy collection were recorded and statistically assessed for impact on results. Previously published cases of iPF were included.2-4 Haematoxylin and eosin-stained biopsy sections were pre-screened to establish histological descriptive criteria and a scoring system, before being randomised (https://www.random.org), blinded and digitally scanned (Aperio AT2 microscope slide scanner; Leica Biosystems). Whole-slide image files were viewed with Aperio ImageScope (v.12.4.6; Leica Biosystems) and glass slides were viewed with an Olympus BX40 microscope. Pustules had to contain acantholytic cells, while granulocyte accumulations in the epidermis without acantholytic cells were classified as micropustules. Pustules with more than 50% of healing at the base by parakeratosis were classified as crusts. Using image files, each histological section and every pustule was individually identified and numbered for each case, and pustules were recorded as intact or ruptured (more than 15% disrupted or partially dried). Pustule descriptors were tracked to each pustule uniquely. The cross-sectional area of each intact pustule was calculated from manual morphometric tracings and the greatest pustule width was measured using Aperio ImageScope. Morphological descriptors are provided in Table 1, including the related scoring scheme used for each criterion. For parameters scored per case, the histological section with the most developed lesion was selected. Boat acantholytic cells were defined as separated keratinocytes in pustules that retained a flattened or ellipsoidal shape. These cells were further subcategorised into granulated boat cells and boat-like cells without keratohyalin granules. All slides were reviewed by two authors (Natalie Gedon and Keith Linder). Hyperplasia Hyperkeratosis Perivascular fibrina Fibrosis severitya Statistical analyses were carried out using Prism 9.5.0 (GraphPad). All multiple association statistics were performed in R Core Team (A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, 2022). Statistical details are provided in Appendix S1 in the Supporting information. The histopathological description of 33 tPF cases is provided below and comparison to the most relevant findings of 26 fPF and 44 iPF cases is summarised in Table 2. Additional descriptive data and statistical analysis of the three groups can be found in Tables S1–S10 and Appendix S2. Thirty-three cases of tPF had 232 unique histological sections with a median number of seven unique sections per case (range: 2–12); 226 were from haired skin and six from footpad/margin. A total of 147 intact and ruptured pustules were evaluated in the epidermis. Seventy-seven intact pustules (Figure 1a) in the epidermis had a median width of 0.37 mm (range: 0.05–4.25 mm) and a median cross-sectional area of 0.04 mm2 (range: 0.002–1.94 mm2). While controlling for area of the intact pustules, there were no significant differences between groups in any of the pustule descriptive parameters (p = 0.37). Ruptured pustules spanned significantly higher number of hair follicles (median: 2, range: 0–12) than intact pustules (median: 0, range: 0–4; p < 0.0001; Wilcoxon–Mann–Whitney U-test); however, the follicle density varied greatly in biopsy samples. The epidermal level of pustules was identifiable in 130 pustules and was most often located from the stratum corneum (SC) to the stratum spinosum (SS; 94; Figure 1a); a few pustules extended from the SC to the stratum granulosum (SG) (four), the SG to the SS (five) or were within the SS (12). In some pustules, there was extension from the SC to the stratum basale (SB; 11), from the SG to the SB (three) or from the SS to the SB (one). Ruptured pustules contained a significantly higher number of acantholytic cells (median score 3) than intact pustules (median 1; p < 0.0001; Wilcoxon–Mann–Whitney U-test). Acantholytic cells contained keratohyalin granules (Figure 2d) in 59% of pustules (82 of 138). Rafts were present (Figure 3a) in 25% of pustules (36 of 143) and their size varied from three to 74 keratinocytes (median 5). Perinuclear eosinophilic rings (Figure 3a) occurred in 57% of pustules (81 of 142). These occurred specifically in acantholytic cells (72 of 142; 51%), in nondetached peripustular keratinocytes in the epidermal margin of pustules (57 of 142; 40%) and/or in rafts. In nondetached keratinocytes with perinuclear rings, cells were not always rounded or separated (Figure 3a) and occurred with or without adjacent neutrophil exocytosis. Boat acantholytic cells (Figures 2 and 3) occurred in most pustules (122 of 145; 84%) and were granulated boat cells (85 of 145; 59%) and/or nongranulated boat-like cells (112 of 145; 77%). Flatter boat cells with granules did not have perinuclear rings, or collapse of keratohyalin granules to around the nucleus, and formed just below the SC (Figures 2a–c and 3b). Wider boat cells sometimes had collapse of perinuclear rings only on the faces of the cell and not from the tapered periphery, granules were not collapsed towards the nucleus from the tapered periphery, and these cells formed deeper in the SG (Figure 2a–c). Corneocytes occurred often in pustules (99 of 145; 68%; Figure 2a,e) and were significantly more common in ruptured pustules (p < 0.0001; Fisher's exact test). Corneocytes delaminated from the inner surface of the stratum compactum (Figure 2a), sometimes with segmental complete loss of the stratum compactum and additional delamination of cells from the stratum disjunctum. Roof cling-on acantholytic cells were present in 30 of 105 (29%) pustules. Pustules always contained neutrophils and, often, eosinophils (92 of 146 [63%]; median severity score 1, range: 0–3). In some instances, eosinophils outnumbered neutrophils (19 of 146; 13%). Macrophages were common in pustules (126 of 146; 86%) and were usually in very low numbers (<5 macrophages); occasional cases had pustules with higher numbers of macrophages (Figure 2d). Degenerate granulocytes occurred in intact pustules (37 of 77; 48%), although the score was usually low (median 0.5; range: 1–3). Neutrophil rosettes of at least two acantholytic cells were present in around half of the pustules (74 of 144; 51%; Figure 3b). Neutrophil rosettes also occurred on free or partly detached boat acantholytic cells and corneocytes (Figures 2e and 3b). Lymphocytic exocytosis into the epidermis was always present, usually mild (median severity score 1, range: 1–3), and more often peripustular (28 of 32; 88%) with lymphocytes below or next to pustules (Table S3). A few micropustules, usually one to three, were present in the epidermis (Figure 1d) in the majority of cases (27 of 33; 82%), were peripustular and/or nonperipustular and were more often in the SS (24 of 27; 89%) than the subcorneal location (11 of 27; 41%). Micropustules contained neutrophils and often eosinophils (14 of 27; 52%) and macrophages (15 of 27; 56%). Eosinophils outnumbered neutrophils in micropustules in nine of 27 (33%) cases. Macrophage microaggregates, one to three, also occurred in the epidermis, yet were uncommon (5 of 33; 15%). Extension of epidermal pustules into hair follicle infundibula occurred in 48 of 134 (36%) pustules. Pustules in hair follicle infundibula (Figures 1c and 2d) occurred in 15 of 33 (45%) cases, although the numbers were low (median: 0; range: 0–17). Micropustules were present in follicles in 14 of 32 (44%) cases. The number of micropustules per case was generally low, one to two only, and thus micropustules were not enumerated per case. Crusts were present in infundibula (evidence of resolved pustule) in 12 of 33 (36%) cases. Pustules with hair follicle extension, follicular pustules and micropustules had similar morphological features to those in the epidermis, including the presence of eosinophils, and occasional macrophages. The total number of cases with evidence of any of the four features of follicular involvement was 27 of 33 (82%), indicating that some form of follicular involvement was common in tPF. Epidermal hyperplasia occurred in all cases (Figure 1a,b), ranged from mild to marked and was regular or irregular. Neither pseudocarcinomatous, papillary nor psoriasiform-like hyperplasia patterns occurred. Hyperkeratosis away from pustules and attached crusts was uncommon (six of 33; 18%), always mild, orthokeratotic and basket-weave to laminated. One or several ulcers occurred in a third of the cases. Ulcers were generally small (less than the width of 50 keratinocytes) and usually were associated with the base or margin of a pustule (92% of ulcers). Most cases (31 of 33; 94%) exhibited moderate epidermal spongiosis (median: 2, range: 0–3), which was predominantly peripustular (Figure 1b). Peripustular spongiosis did not affect all pustules in cases where it occurred. Spongiosis occurred along the base of pustules, below acantholysis and not above it. Spongiotic pustules (no acantholytic cells) were uncommon (four of 33; 12%), small and few in number (one to two per case). Mild laminar hydropic change of the epidermis was present in a few cases. Necrotic acantholytic keratinocytes (Figure 3a,c) with retained shape and nuclear profiles occurred in 39 of 75 (52%) of intact pustules. They were hypereosinophilic and lacked differential staining (coagulative necrosis). Nuclei were sometimes pyknotic. Low numbers of necrotic acantholytic keratinocytes were usually observed (median: 1, range: 0–3). Necrosis was occasionally observed in nondetached peripustular keratinocytes (13 of 75; 17%) within intact pustules. A few cases had similar, rounded, individual necrotic keratinocytes in the epidermis next to pustules and were satellited by neutrophils. Moderate numbers (11–25) of apoptotic keratinocytes were seen in the epidermis of most cases (31 of 33; 94%; median: 2, range: 0–3). They were more often away from pustules (29 of 30; 97%) than near pustules (14 of 30; 47%). Basal layer apoptosis (29 of 31; 94%) occurred more commonly (Figure 3d) than suprabasal apoptosis (20 of 31;65%) and basal layer apoptosis was usually more numerous (26 of 31; 84%) than suprabasal (five of 31; 16%) apoptosis. The basal layer of hair follicle infundibula often had apoptosis, which was not associated with pustules. Lymphocytic satellitosis of apoptotic keratinocytes was seen in nine of 31 (29%) cases and usually only on rare cells in the epidermis. Additional inflammatory patterns occurred in all three PF clinical phenotypes. Lymphocytic bulbitis (tPF; Figure 4a), subepidermal vesicular dermatitis (iPF; Figure S1b) and suprabasal clefting (iPF; Figure S1a) were all seen in one case each. Lymphocytic interface dermatitis was mild-to-marked in four fPF cases (Figure S1c), and mild in one iPF and in one tPF case. Vasculitis in the panniculus was neutrophilic, eosinophilic and/or fibrinous and moderate in one tPF case (Figure 4b), and mild in one iPF case (Figure S1d). Multifocal, neutrophilic to pyogranulomatous sebaceous gland (or duct) adenitis was mild in five tPF cases (Figure 4c), mild-to-marked in two fPF cases (Figure S1e) and mild in four iPF cases. Furunculosis was mild to marked and found in six tPF cases (Figure 4d and Figure S1f) and in one case each of fPF and iPF. Nodular pyogranulomatous dermatitis was mild-to-marked and occurred in two cases of tPF and one case each of fPF and iPF, which were likely to have been secondary to furunculosis in some cases. The histomorphological features of canine tPF are characterised for the first time and are similar to fPF and iPF. Some morphological features were statistically different, yet their incidence was relatively high in all PF groups and histological features were not identified that could confidently differentiate PF phenotypes. The shared morphological features support common molecular mechanisms in canine PF phenotypes, despite—presumably—different disease triggers. Indeed, PF phenotypes are thought to share desmocollin-1(DSC1) as a major autoantigen.1, 3, 5 Beyond this commonality, in-depth investigations of the molecular mechanisms of canine PF are lacking and the reason for clinical variations is still unknown. Likewise, clinical variations in human PF and pemphigus vulgaris (PV) are well-recognised.6 Differences in circulating autoantibodies (anti-DSG1 versus anti-DSG3) have been associated with clinical lesion location and epidermal depth of acantholysis.6, 7 In dogs, anti-desmoglein-1(DSG1) antibodies were demonstrated in only 5%–7% of fPF and tPF patients, and antibodies to other DSGs and DSCs have not been investigated.1, 5, 8, 9 In proteomics-based and protein-array investigations of human PF and PV, a much larger number of autoantibody targets and autoreactive B cell clones have been identified; which also might relate to clinical phenotypes.10 Potentially, the clinical differences in tPF versus fPF are partly to the result of differences in target tissues, nasal planum and footpad being more similar anatomically and less targeted in tPF than haired skin.1 In fact, some canine PF patients present only with nasal planum or footpad lesions.11, 12 Based on keratinocyte morphology, different mechanisms of keratinocyte acantholysis are likely to occur in canine PF variants. Perinuclear eosinophilic rings occurred in acantholytic and peripustular keratinocytes, and these rings are thought to indicate active mechanisms of cell separation.13, 14 Ultrastructurally, desmosomes are reduced in number and size, and keratin intermediate filaments retract towards the nucleus to form the ring.14 Mechanistically, it is thought that autoantibodies induce cell separation by depletion of the membrane pool of cadherins, thus limiting desmosome formation, and by desmosome dismantling and internalisation, mediated partly by active cell signalling via calcium, protein kinase-C and phospholipase-C.15, 16 In this study, pustule margin keratinocytes formed rings before cell detachment, with or without adjacent neutrophils, suggesting neutrophil proximity is not needed to initiate this process. A role for secreted neutrophil enzymes/mediators from distant neutrophils in pustules is not excluded, however. Our study identified boat acantholytic cells and individualised corneocytes in pustules which indicate that nonactive mechanisms of cell separation also occur. Active acantholysis cannot separate corneocytes because of terminal differentiation and covalent cross-linking of corneodesmosomes and intermediate filaments by transglutaminases.17 Boat acantholytic cells formed in the SG and have intermediate features. These cells remain partially flattened, retain tapered edges, contain a nucleus and fail to fully round when separated, suggesting that the cytoskeleton is partly stabilised, possibly by transglutaminase activity and/or by interactions with tight junctions that occur at the cell periphery in SG2 and SG1 layer cells.17, 18 Tight junction interactions could explain collapse of inte
