Ultrasonography of Osteosarcoma in Dogs
Sonography Atlas of Dogs

Abstract

Canine cancers

Although rigorous epidemiological data are not available for companion animals, current estimates suggest that approximately 2–4,000,000 dogs in the US are diagnosed with cancer annually [33, 34]. Overall, the best available data estimate that approximately one in four dogs in the US will be diagnosed with cancer, which (if confirmed) would translate to an incidence of canine cancers per year strikingly greater (5300 cases/ 100,000 dogs) than that of humans (500 cases/ 100,000 persons) [33, 34]. With growing cancer diagnoses in both human and canine populations, spontaneous cancers in dogs make them ideal for the study of cancer biology and immunotherapy, especially since pet owners are highly driven to seek out novel treatments for their companion animals. Moreover, with the decoding of the canine genome, important similarities between dog and human oncogenesis, including specific cancer-associated genes such as BRAFV600E, p53, Bcr-Abl, and c-kit have been recognized [35–37]. In osteosarcoma (OSA), for example, gene expression profiling has demonstrated remarkable homology between canine and human forms of the disease, reinforcing the shared biology between dog and human [38, 39].

In addition, cancers that develop in dogs show the same complex interplay of genetics, age, and environmental exposures as in humans [12, 33, 40, 41], and these similarities are stronger between humans and dogs than they are between humans and mice [30, 33, 42–44]. As with humans, cancer incidence in dogs is associated with increasing age, although certain cancers do display distinct epidemiological patterns between dogs and people [45, 46]. For example, in humans, OSA is markedly more common in children and adolescents, whereas the diagnosis peaks in middle age to older dogs between 7 and 9 years [47]. Interestingly, the risk of OSA in dogs also increases with increasing body weight and is specifically associated with large and giant breeds such as Great Danes, Saint Bernards and Irish wolfhounds [48]. Somewhat paradoxically, although outbred at the species level, the genetic diversity of dogs is also quite restricted at the breed level [44]. This artificial genetic selection has clearly placed different breeds at elevated risk for certain malignancies as shown in Table 2 [34, 49].

Table 2: Common Canine Cancers with Key Demographic Features. ^aData on breed predisposition of specific cancers are potentially subject to reporting bias given differences in breed popularity and differences in how owners may seek veterinary care.

Although dogs develop cancers from tissues throughout the body similar to humans, the incidence and prevalence of different tumor types show other notable differences from human counterparts. For example, while OSA, malignant mast cell tumors, and hemangiosarcoma are some of the most common malignancies in dogs, these cancers are uncommon in humans [34, 47]. Similarly, while non-Hodgkin’s lymphoma (NHL) occurs in humans with an estimated incidence rate of 19.6 per 100,000 people, NHL occurs with an even greater frequency in dogs (in 2014, approximately 250,000 cases of lymphoma were diagnosed in dogs versus approximately 71,000 cases in humans) [50–54]. Conversely, human colorectal, pancreatic and pulmonary carcinomas that cumulatively account for approximately 40% of estimated cancer deaths in the US, are much less common in dogs with an incidence of less than 1% for each tumor type [55]. Furthermore, there is a markedly lower incidence of canine breast cancer in the US than in other parts of the world or in humans, almost certainly due to the routine practice of spaying dogs in the US [56–59]. Nevertheless, it is important to recognize these differences in the incidence and prevalence of specific cancer types between humans and dogs because these differences impact the translational relevance of canine cancer studies to humans. However, such variations can also be advantageous for clinical translation of novel cancer therapies since the higher incidence of OSA and hemangiosarcoma in canines, for example, can be utilized to obtain clinical data more rapidly than what is achievable in humans with these rare tumors.

Prior to the recent growth in companion animal clinical trials, the laboratory research beagle represented a more standardized way to proceed with experiments in canines. International requirements, particularly in the United Kingdom and in Europe, require toxicology and pharmacology studies in at least two animal species, a rodent and non-rodent, prior to human clinical trials. The non-rodent model is frequently the research beagle, which has been chosen because of its relatively small size as well as its passive nature and affable personality. A single breed also minimizes breed variability that might otherwise exist among studies. Choi et al. and Ikeuchi et al. have provided reference values for hematological, serum biochemical and urological, as well as organ weight parameters to establish a standardized set of normal values, and minimize the use of laboratory canines for baseline studies [60, 61]. However, breed-related variations in hematologic and serum biochemical values have been reported, signaling caution in applying immunologic findings from beagle studies to other dog breeds [62, 63]. Curiously, laboratory beagles anecdotally appear to be less sensitivity to toxicity than most pet dogs [64]. In fact, it has been recommended that when undertaking a Phase I clinical trial of a chemotherapy agent in client-owned dogs, the first dose administered should be 50% of the maximum tolerated dose observed in laboratory beagles because of their apparent favorable toxicity profile. Although the reason for this remains unknown, it does reinforce the concept that data derived from laboratory beagles may not be easily compared to that obtained from client-owned dogs and other breeds.

Canine immune assays

Although there are many advantages to the canine model, currently a key barrier to detailed mechanistic/correlative studies (outside of the inherent variability and cost associated with clinical monitoring in large numbers) in canine models and clinical trials is the paucity of widely available, standardized, and validated canine reagents for laboratory use. For example, although the fundamental components of the dog immune system have been examined to date, characterization of specific components has been much less detailed. Neonatal and post-natal studies of dogs suggest that canines resemble humans and differ from rodents in that dogs appear to be immunologically competent at, or before, birth [72]. Moreover, similar to humans, the phenotype of lymphocytes in the peripheral blood and tumor microenvironment of dogs with cancer has been linked with prognosis. For example, Estrela-Lima et al. observed that both increased tumor-infiltrating lymphocytes based on phenotypic analysis of single cell suspension of tumor tissue by flow cytometry and increased blood CD4/CD8 ratios were correlated with worse survival in canine mammary cancers [73]. Similarly, elevated Tregs, tumor-associated macrophages, and myeloid-derived suppressor cells, respectively, have been associated with adverse outcome in canine B cell lymphoma and mammary tumors [74–76]. Although these studies reinforce the impression of important homology between dog and human immunobiology, especially in cancer, they also highlight the correlative nature of many canine studies with a notable absence of carefully controlled and functional experiments to satisfy high levels of evidence regarding causation and mechanism. Interestingly, there is evidence for breed effects on immune function which likely relate to the inheritance of particular haplotypes of major histocompatibility complex (MHC) genes and further reinforce the paradox that dogs are a highly outbred species which nevertheless manifest significant effects of genetic inbreeding [77–79].

To address the increasing focus on canine models, researchers and vendors have recently focused on the development and dissemination of commercially available, canine-specific antibodies for basic and translational research, as researchers often rely on human and mouse antibodies which have been validated to be cross-reactive for canine markers. Table 3 shows some cell surface markers that are used to phenotype the various canine immune subsets. Mixed-lymphocyte reactions, co-culture killing assays (Chromium release and/or flow cytometry based), IFN ELISPOT, intracellular cytokine staining, and phagocytic activity of dendritic cells using fluorescent-labeled latex beads are all standard immune functional assays used in canine models [80].

Table 3: Phenotype of Canine Immune Subsets. ^aPossibly unique to canines, CD4 is expressed in granulocytes. Similar to other species, CD4 is expressed in a subset of monocyte-derived cells.

Characterization of canine immunoglobulins dates to the work of Johnson et al. in 1967 [16]. While this body of work, and studies that followed, demonstrated that canine IgGs consist of four subclasses, the diverse functions and interactions of canine immunoglobulins with other immune effector cells have remained less characterized [81]. Nevertheless, there has been a longstanding interest in canine specific mAbs, including canine-CD20 targeted antibodies, for therapeutic and diagnostic use [66, 68, 69]. Important for cross-species translational studies, Bergeron et al. demonstrated that canine Fc gamma receptors bind to dog, human, and mouse IgGs, suggesting that a human therapeutic antibody could be effective at stimulating ADCC in a canine therapeutic model, although species differences may result in significant differences in activity as well as eventual neutralization by the host [81]. Speciated antibodies in a dog IgG framework are now routinely developed using the hypervariable regions of the variable antigen-binding domain (Fv) derived from mouse antibodies [81]. Important for immunotherapy studies, expression of checkpoint molecules, including PD-L1, has been observed on several canine tumors including mastocytoma, melanoma, and renal cell carcinoma [82], and elevated CTLA-4 expression using mouse anti-human antibodies has been observed in dog histiocytic sarcoma patients compared to healthy controls [83]. Unfortunately, studies to explore immune checkpoint blockade in dogs will have to wait as reagents against canine PD1, PDL-1, and CTLA-4 are not yet commercially available, nor do they exist in formats that are suitable for clinical translation.

Overall, the major immune subsets have been characterized in dogs, and significant homology with humans has been demonstrated, but notable differences have also been observed. In 1994, an international Canine Leukocyte Antigen Workshop was held, establishing important canine homologues for key leukocyte populations such as CD4, CD8, and CD90 [84]. Subsequently, homologues of CD45R, CD45RA, CD11, and CD62L were also identified. However, despite these advances, characterizing naïve, activated, and memory subsets for T cells and other immune effector cells has remained limited. For example, Isotani et al. characterized canine dendritic cells with morphology and phagocytic function comparable to mouse and human DCs [80]. In addition, the DCs demonstrated expression of MHC class II, CD11c, CD80, and CD86, and these markers have been used to identify canine DCs in other studies [80]. In another important study, Hartley et al. used multiple cross-reactive antibodies including rat anti-human CCR7 and mouse anti-human CD62L to show downregulation of these surface molecules on activated T cells. Based on these data, the authors proposed a schema for distinguishing canine central memory T cells (CCR7⁺CD62L^hiCTL2.58⁻) from activated T cells (CCR7⁻CD62L^loCTL2.58⁺) [85].

Conversely, dog NK cells have proved more difficult to characterize as dogs do not express CD56 and marker systems such as CD3-CD5^dim have been used to describe NK cell activity. Overall, although a clear consensus has not yet emerged [86–88], recent work of Foltz et al. and Grondahl-Rosado et al. have independently demonstrated that canine NKp46 expression appears to identify a canine CD3⁻ lymphocyte population with characteristics and cytotoxicity of NK cells [89–91].

Advances in canine cancer immunotherapy

Paralleling the rapid adoption of immunotherapy in human clinical medicine, immunotherapy in canine veterinary medicine is gaining increasing utilization for both approved and investigational indications. For example, we reported a canine clinical trial in metastatic sarcoma and melanoma testing a novel immunotherapy combination including local radiotherapy (RT), intratumoral CpG oligodeoxynucleotides (immune stimulatory toll-like receptor 9 agonists), and systemic administration of indolamine-2,3 dioxygenase (IDO) blockade with 1-Methyl-Tryptophan (to circumvent immune suppressive pathways) [98]. The premise of this approach was to assess combination therapies involving conventional treatments which are immunostimulatory with strategies to inhibit immunosuppressive pathways [99]. This canine trial was paired with detailed mechanistic studies in murine models which demonstrated that the triple therapy of local RT, intratumoral CpG, and systemic IDO not only reduced intratumoral immune suppression/IDO blockade, but also induced robust systemic anti-tumor effects and tumor regression in five dogs with metastatic melanoma and sarcoma (Fig. 1) [98]. Importantly, a biomarker as shown by reduction in circulating and tumor Tregs was observed in responding but not non-responding dogs. The lack of toxicities associated with the regimen and promising clinical results is leading to human clinical extrapolation.

Figure 1: Efficacy of Radiation + CpG + 1MT in a Canine Clinical Trial. Canines with metastatic melanoma and sarcoma were accrued to a pilot clinical trial at the UC Davis School of Veterinary Medicine [98]. a. Baseline computed tomography of the thorax demonstrates untreated metastatic lesions in a dog with buccal melanoma. b. One month after local RT and intra-lesional CPG to the primary tumor combined with systemic indolamine-2,3 dioxygenase (IDO) inhibition, there is complete regression of some lesions and partial regression of others. Arrows denote index lesions.

An example of cell therapy studies under active investigation in canine cancers include the recent CD20-targeting RNA CAR T cells by Panjwani et al. [100]. These authors demonstrated for the first time in dogs that autologous RNA-transfected CAR T cells could be generated, expanded, and administered to a dog with relapsed B cell lymphoma. They observed that treatment was not only well tolerated, but also associated with a reduction in CD20+ B cells in target lymph nodes at 72 h, providing proof-in-concept that CAR therapies can be successfully applied in canine patients with the caveat that more meaningful clinical affects will be contingent on stable CAR expression as long-term engraftment may be an issue. Similarly, Mata et al. tested the ability of genetically modified canine T cells to express a chimeric human HER2-canine TCR CAR T cell [101]. The authors demonstrated successful expansion and activation of the engineered cells which also effectively and selectively killed HER2-positive target cells using in vitro assays. Although not formally tested in a dog trial as yet, the pre-clinical work of Mata et al. on CAR T cells illustrates both the promise and potential barriers of using the canine model. Unique challenges such as reliance on better characterized, more widely available human or mouse-based proteins, cytokines, and transgenes risks the potential for the consequences of xeno-antibody formation [101, 102]. However, given the risk of severe, even life-threatening, adverse events with CAR T cell and other strong immunotherapy regimens, especially when given systemically, clinical trials of these novel therapies in dogs should be helpful to answer key questions about toxicity and efficacy [103].

NK cell immunotherapy approaches are also being assessed in canine models. At our institution, an ongoing phase 2 canine clinical trial is evaluating the intratumoral injection of autologous activated NK cells following palliative RT for appendicular OSA. Treatment consists of palliative RT weekly for 1 month, and following RT, dogs receive two intra-lesional injections of autologous canine NK cells isolated, expanded, and activated ex vivo, supplemented with clinical grade rhIL-2 for in vivo cytokine support. Another issue in dog immunotherapy studies is the difficulty in obtaining cost-effective amounts of recombinant canine cytokines to be given in vivo for such trials. Administration of human cytokines will be eventually neutralized with repeated use. To date, we have accrued eight patients, and preliminary results have been promising showing minimal toxicities, supporting the use of the canine model for the testing of this novel NK approach (manuscript in preparation). Thus, the canine model is well suited for adoptive cellular immunotherapy evaluation.

Conclusions

There is a growing body of evidence that the spontaneous cancers in dogs represent attractive translational models that bridge mechanistic studies in mice to the heterogeneous human situation where clinical trials are time and resource intensive. Particularly in the burgeoning field of immunotherapy, as a complement to murine studies and human clinical trials, dogs offer an innovative model for translational research, as they present many of the same challenges faced in “scaling up” a therapeutic system dependent on complex interactions between multiple cell types yet under more controlled settings. They also allow for long-term assessment on efficacy and toxicities. Canine clinical trials offer unique access to a rich source of spontaneously occurring, genetically and immunologically diverse cancers with the benefits of reduced time, expense, and regulatory hurdles of a human trial.

Yet, it is important to recognize that there are disadvantages to the canine model, in particular the currently limited canine-specific/cross-reactive reagents and characterized epitopes available for use. Ultimately, as future of cancer therapy appears to increasingly point to immunotherapy, canine clinical/co-clinical trials represent an ideal format for the rapid and clinically relevant translation of novel and high impact immune therapies and immune combination therapies.

Acknowledgements

Share this article / Teilen Sie diesen Artikel