Clinical, pathologic and molecular findings in 2 Rottweiler littermates with appendicular osteosarcoma

Abstract

Appendicular osteosarcoma was diagnosed and treated in a pair of littermate Rottweiler dogs, resulting in distinctly different clinical outcomes despite similar therapy within the context of a prospective, randomized clinical trial (NCI-COTC021/022). Histopathology, immunohistochemistry, mRNA sequencing, and targeted DNA hotspot sequencing techniques were applied to both dogs' tumors to define factors that could underpin their differential response to treatment. We describe the comparison of their clinical, histologic and molecular features, as well as those from a companion cohort of Rottweiler dogs, providing new insight into potential prognostic biomarkers for canine osteosarcoma.

Keywords: Osteosarcoma; Rottweiler; canine; genomics.

Figure 1. Primary tumor histopathology of dog 1410 and 1411.

Representative images of tumor tissue from dogs 1410 and 1411 stained with hematoxylin & eosin (H&E). Scale bar = 50 μm.

Figure 2. Radiographic images from dog 1410 and 1411.

Right hindlimb radiographs of stifle joint of dog 1410 (poor outcome) taken at the time of diagnosis (panel A: anterior-posterior projection, B: lateral projection). In the distal metaphysis of the right femur, there is a mild moth-eaten bone lysis and marked sclerosis extending into the distal diaphysis and epiphysis. Radiographic images from dog 1411 (elite outcome), also from the right hindlimb but highlighting the tarsal joint (panel C: anterior-posterior projection, panel D: lateral projection). At the distal metaphysis and epiphysis of the tibia there is moth eaten lysis. The cranial and caudal margins of the cortex are smooth but thinned. Within the mid diaphysis, the medullary cavity has a mottled appearance. Circumferentially to the tarsus and within the tibiotarsal joint, there is a severe (more severe dorsomedially) amount of soft tissue swelling.

Figure 3. Somatic Mutations Detected by SearchLight DNA in Tumors from Dogs 1411 and 1410.

Mutations that are shared by both dogs are shown followed by those detected only in 1411 (bottom of left column) or 1410 (right column). These mutations represent primarily somatic SNVs and CNVs detected in 120 genes via the SearchLight DNA panel based on sequencing matched tumor and normal tissue for each dog. One germline sequence variant was detected in NF2 in Dog 1411, but was not shared by Dog 1410. Six genes bore similar somatic mutations in both cases whereas 27 genes bore mutations in only a single sibling’s tumor.

Figure 4. Copy number and gene expression relationships in tumor tissue from dogs 1410 and 1411.

A. mRNA expression vs log 2-fold change copy number variation for patient 1411 (Spearman correlation 0.4, p=0.14) B. mRNA expression vs log 2-fold change copy number variation for patient 1410 (Spearman correlation 0.54, p=0.0083). C. mRNA expression for both 1411 and 1410 of 33 genes exhibiting a copy number variation in either dog (Spearman correlation −0.0137, p=0.94).

Figure 5. Differentially expressed genes (DEGs) between Rottweilers with disparate clinical outcomes.

Ward Clusters of 10 Rottweilers in the DOG² Cohort over 97 differentially expressed genes between Elite responders (Blue) and Poor Responders (Red). Patient ID numbers are listed along the x axis. Dogs 1410 and 1411 are indicated by arrows. A complete gene list is also provided in Supplemental Table 2.