Background: Equine influenza virus is a common cause of respiratory disease in equids. Few reports describe clinical presentation and disease progression in donkeys. Hypothesis/Objectives: Describe the clinical and diagnostic findings, outcome, and pathologic lesions associated with influenza pneumonia in donkeys. Animals: Thirteen unvaccinated donkeys ranging from 1 week to 12 years of age and sharing clinical signs and exposure history. Methods: Retrospective case series. Medical records from June to July 2020 at the Colorado State Veterinary Teaching Hospital and collaborating referring veterinary practices were reviewed. The diagnosis was confirmed by molecular testing, virus isolation, and partial genetic and phylogenetic analysis of the virus. Results: Survival in donkeys 1 year of age was 85.7% (6/7). Hemagglutinin gene sequencing and phylogenetic analysis confirmed a contemporary clade 1 Florida sublineage H3 virus as the causative agent. Conclusions and Clinical Importance: Clinical signs of equine influenza virus infection in donkeys are similar to those observed in horses. Prognosis for survival generally is good, but deaths have been observed especially in foals born to seronegative dams. This finding emphasizes the importance of prenatal vaccination protocols in all equids, including donkeys.

Mast cell tumors (MCTs) are an uncommon primary neoplasm of the nasal cavity in dogs for which there is a paucity of existing literature regarding their clinical behavior and molecular features. The objectives of this retrospective study were to examine the clinical findings, histopathologic and immunohistochemical features, and c-KIT mutation status of primary intranasal MCTs in dogs and identify potential prognostic factors. Canine biopsies submitted to a diagnostic laboratory in Colorado between 2010 and 2019 with intranasal neoplasms diagnosed as MCTs and no history of cutaneous or oral MCT were considered. Immunohistochemistry for CD117 and Ki67 and polymerase chain reaction (PCR) for internal tandem duplications at exons 8 and 11 of the c-KIT gene were performed. Twenty out of 1849 (1%) primary intranasal neoplasms were MCTs. Metastases were reported in 11/20 cases (55%), with the mandibular lymph node representing the most common site. One case had distant metastases to abdominal viscera. Of the cases with available outcome data, 6/14 (43%) died or were euthanized from MCT-related disease within 1 year of the onset of clinical signs. Only one case had a c-KIT mutation at exon 11. In our study, intranasal MCTs were prone to metastasize and had a generally poor prognosis, resembling the behavior of MCTs arising in other mucosal locations. While dogs with metastatic disease and survival times of <1 year tended to have atypical KIT localization, moderate to high Ki67 indices, and mitotic counts ≥8, definitive prognosticators could not be identified due to the limited number of cases with favorable clinical outcomes.

Background: Peripheral T-cell lymphoma (PTCL) refers to a heterogenous group of T-cell neoplasms with poor treatment responses and survival times. Canine PTCL clinically and immunophenotypically resembles the most common human subtype, PTCL-not otherwise specified (PTCL-NOS), leading to interest in this canine disease as a naturally occurring model for human PTCL. Gene expression profiling in human PTCL-NOS has helped characterize this ambiguous diagnosis into distinct subtypes, but similar gene expression profiling in canine PTCL is lacking. Methods: Bulk RNA-sequencing was performed on tumor samples from 33 dogs with either CD4+ (26/33), CD8+ (4/33), or CD4-CD8- (3/33) PTCL as diagnosed by flow cytometry, and sorted CD4+ and CD8+ lymphocytes from healthy control dogs. Following normalization of RNA-seq data, we performed differential gene expression and unsupervised clustering methods. Gene set enrichment analysis was performed to determine the enrichment of canine CD4+ PTCL for human PTCL-NOS, oncogenic pathways, and various stages of T-cell development gene signatures. We utilized gene set variation analysis to evaluate individual canine CD4+ PTCLs for various human and murine T-cell and thymocyte gene signatures. Cultured canine PTCL cells were treated with a pan-PI3K inhibitor, and cell survival and proliferation were compared to DMSO-treated controls. Expression of GATA3 and phosphorylated AKT was validated by immunohistochemistry. Results: While the canine CD4+ PTCL phenotype exhibited a consistent gene expression profile, the expression profiles of CD8+ and CD4-CD8- canine PTCLs were more heterogeneous. Canine CD4+ PTCL had increased expression of GATA3, upregulation of its target genes, enrichment for PI3K/AKT/mTOR signaling, and downregulation of PTEN, features consistent with the more aggressive GATA3-PTCL subtype of human PTCL-NOS. In vitro assays validated the reliance of canine CD4+ PTCL cells on PI3K/AKT/mTOR signaling for survival and proliferation. Canine CD4+ PTCL was enriched for thymic precursor gene signatures, exhibited increased expression of markers of immaturity (CD34, KIT, DNTT, and CCR9), and downregulated genes associated with the T-cell receptor, MHC class II associated genes (DLA-DQA1, DLA-DRA, HLA-DQB1, and HLA-DQB2), and CD25. Conclusions: Canine CD4+ PTCL most closely resembled the GATA3-PTCL subtype of PTCL-NOS and may originate from an earlier stage of T-cell development than the more conventionally posited mature T-helper cell origin.

Naturally occurring canine acute leukemia (cAL) is a promising spontaneous model of human acute leukemia, but limited understanding of its molecular heterogeneity has hindered classification and cross-species translation. We performed bulk RNA sequencing on peripheral blood leukocytes from 110 dogs with acute leukemia and normal bone marrow and peripheral blood controls. Unsupervised transcriptomic analyses were used to define molecular subgroups. We also generated a single-cell RNA-sequencing atlas of normal canine thymus and integrated it with a published canine bone marrow atlas to establish developmental reference signatures. Transcriptomic profiling identified major myeloid and lymphoid groups strongly associated with flow cytometric lineage assignment (χ² = 44.62, P < 0.0001). Myeloid leukemias were enriched for canonical myeloid and inflammatory programs, whereas lymphoid leukemias showed T-lineage developmental enrichment and suppression of B-lineage genes. Both groups retained progenitor- and leukemic stem cell-associated programs. Within each lineage, three distinct molecular subgroups revealed previously unrecognized transcriptional heterogeneity associated with developmental cell-state enrichment. Together, these findings establish a molecular framework for cAL by defining conserved developmental programs and lineage-specific heterogeneity, strengthening its translational relevance as a comparative model and providing a foundation for future cross-species studies of acute leukemia biology and therapeutic development.

Nonclinical toxicity studies are essential components of the drug development process and are based on labor-intensive visual examination of large numbers of glass slides under the microscope by pathologists. In this work, we present an artificial intelligence (AI)-based solution using an unsupervised representation learning model with a Bidirectional Generative Adversarial Network (BiGAN) for triaging normal and abnormal tissues and a machine learning-based severity grade classifier (SGC) for severity grade estimation on whole-slide images (WSIs) of rat liver. The BiGAN model was trained solely on vehicle control WSIs to learn normal histology and computed a tile-level similarity error, representing the deviation from normalcy, which was used to identify tile- and slide-level abnormalities. Our BiGAN model demonstrated an abnormality discrimination of receiver operating characteristic (ROC) area under the curve (AUC) of 0.77 with highest sensitivity for high-grade abnormalities. We utilized the similarity error data from BiGAN to train the SGC model and predicted slide-level severity grades with ROC AUC values ranging from 0.68 to 0.96 (minimal to marked) with most errors within ±1 grade, reflecting real-world pathologist grading variability. In addition, our heatmap visualizations of tile-level abnormalities revealed good to fair agreement with ground truth abnormalities in 70% of evaluated WSIs. By training the BiGAN on normal tissue slides, including minimal grade background abnormalities from vehicle control animals, our workflow eliminates annotation requirements and enables rapid adaptation to new tissues and species, paving the way for scalable AI models that streamline nonclinical pathology with minimal pathologist input.