Ultrasonography of Osteosarcoma in Dogs
Sonography Atlas of Dogs

Abstract

Table 2: Characteristics of healthy dogs in the study.

Table S3 is a summary of samples assayed in each figure. Clinical data were obtained from medical records. Control dogs were determined to be healthy based on physical examination, owner observations and complete blood count exams. For dogs with cancer, the diagnosis and tumor staging were based on complete physical examinations, histopathology of tumor biopsy specimens, blood work and specialized imaging tests, such as CT scans, ultrasound or radiographs, to assess tumor location and size, as well as the presence of metastatic disease. Dogs with large, necrotic or multiple masses, lytic or severe bone destruction (with osteosarcoma) or presence of metastasis, were placed into the advanced stage/metastatic group. Animals presenting with small masses or no metastatic nodules were placed into the early stage non-metastatic group. Tables S1 and S2 also list specifics about any treatment that dogs with cancer had received prior to or at the time of blood collection for this study.

Blood samples from both cancer and healthy control dogs were obtained specifically for this study. Samples were collected in heparinized tubes by the Oncology and Community Practice Services of the Veterinary Medical Center at the University of Minnesota according to Institutional Animal Care and Use Committee guidelines. The samples were drawn after the owners signed the client consent form. The Institutional Animal Care and Use Committee (IACUC) reviewed and approved the study entitled as “Flow Cytometric Immunophenotyping of Peripheral Blood Cells in Dogs” via designated member review under the code number 0912A75493. Unless explicitly stated otherwise, the cells being analyzed for this manuscript co-purified with peripheral blood mononuclear cells (PBMCs) of dogs with cancer or age-matched healthy controls that were isolated using Ficoll (Sigma) gradient centrifugation as follows. Heparinized peripheral blood was diluted 1∶3 with sterile PBS (Invitrogen) and layered over Ficoll-Histopaque (Sigma). Samples were centrifuged at 400-× g for 30 min. The PBMCs collected at the interface were transferred to a fresh tube, washed twice with PBS, and resuspended with freezing solution consisting of 90% fetal bovine serum (Invitrogen) 10% Dimethyl sulfoxide (DMSO) (Sigma) and then frozen at −80°C. Lastly, PBMCs were thawed for 2 minutes in a 37°C water bath before staining and analysis. For analysis of fresh samples, PBMCs were isolated as above, resuspended in FACS buffer, stained with antibodies and immediately analyzed by flow cytometry or FACS as indicated.

Dogs with advanced or metastatic cancer had a significantly greater fraction of putative MDSCs (36.04±2.542, mean ± SEM) compared to dogs with early stage non metastatic tumors (9.40±0.953, mean ± SEM) and healthy control dogs (10.24±1.412, mean ± SEM) (Fig. 2A). Moreover, this elevation in the CD11b⁺CD14⁻MHCII⁻ fraction did not appear to be restricted to a specific tumor type. The differences were statistically significant in dogs with sarcomas, carcinomas, and mast cell tumors compared with healthy controls (Fig. 2B). Conversely, the percentage of CD11b⁺MHCII⁻ cells that did express CD14 was not significantly different among any group. Therefore, the frequency of CD11b⁺CD14⁻MHCII⁻ cells that co-purify PBMCs correlates with tumor burden. This finding is in agreement with previously published data regarding MDSC levels and tumor burden in mice and humans [17], [18].

Figure 1: Percentages of circulating CD11b⁺CD14⁻MHCII⁻ cells in dogs with correlates with clinical tumor stage.

(A) Analysis of average CD11⁺CD14⁻MHCII⁻ population frequency in dogs with advanced stage or metastatic tumors (n = 30) compared with early stage non-metastatic tumors (n = 15) and control dogs (n = 18). There was a significantly higher percentage of CD11b⁺CD14⁻MHCII⁻ cells in dogs with advanced cancer versus early stage non-metastatic tumors and healthy dogs (36.04% vs. 9.40% and10.24%, respectively. B) Average CD11b⁺CD14⁻MHCII⁻ population frequency in the major cancer subtypes: advanced stage or metastatic sarcomas (n = 18), early stage non-metastatic sarcomas (n = 6), advanced stage or metastatic carcinomas (n = 7) early stage non-metastatic carcinomas (n = 7), advanced stage or metastatic mast cell tumors (n = 5) and early stage non-metastatic mast cell tumors (n = 2) compared with control dogs (n = 18). Significantly elevated percentages were detected in all advanced tumors subtypes relative to early stage tumors and healthy dogs. Percentages of CD11b⁺CD14⁺MHCII⁻ cells were not significant between groups (* indicates P<0.001). Mean ± SEM are shown.

CD11b⁺CD14⁻MHCII⁻ cells are functionally defined as MDSCs

To test whether the CD11b⁺CD14⁻MHCII⁻ subset was able to inhibit T cell function, we conducted a series of co-culture experiments. Purified CD11b⁺CD14⁻MHCII⁻ cells from three different subtypes of cancer were co-cultured with autologous or healthy responder PBMCs. In all cases, CD11b⁺CD14⁻MHCII⁻ cells exhibited a potent ability to suppress proliferative responses in a dose-dependent manner. Representative examples of proliferative suppression are shown using samples from a dog with tonsillar squamous cell carcinoma (Fig. 3A) and prostatic adenocarcinoma (Fig. 3B). In order to determine if suppression was an artifact of using responders from tumor-bearing dogs, we assayed for proliferative suppression using normal responders. The addition of CD11b⁺CD14⁻MHCII⁻ cells, but not normal PMNs, impaired the proliferation of PBMCs from healthy dogs (Fig. 3C). Moreover, the amount of IFN-γ secretion was assessed in the conditioned medium from these co-cultures, revealing that CD11b⁺CD14⁻MHCII⁻ cells, but not normal PMNs, suppressed the secretion of IFN-γ (Fig. 3D).

Figure 3: CD11b⁺CD14⁻MHCII⁻ cells suppress T cell proliferation and cytokine elaboration.

CD11b⁺CD14⁻MHCII⁻ cells were sorted from peripheral blood sample of dogs with cancer and then co-cultured with autologous PBMCs (A, B) or healthy dog PBMCs (C) in the presence of mitogen for 72 hs. Representative examples from a total of eight dogs are shown. The graphs represent proliferative responses after addition of CD11b⁺CD14⁻MHCII⁻ isolated from a single dog with squamous cell carcinoma (3A), prostatic adenocarcinoma (3B) and osteosarcoma (3C). Non-stimulated PBMCs were used as negative control and PBMCs stimulated in absence of CD11b⁺CD14⁻MHCII⁻ cells were used as positive control for proliferation. PBMCs were also co-incubated with PMNs, to control for presence of additional cells (3C, 3D). Proliferative responses were measured by ³H-thymidine incorporation. CPM, counts per minute. Amount of IFN-γ secretion in the co-culture was determined using canine specific IFN-γ ELISA assay (3D). All experiments were performed in triplicate. Mean ± SEM are shown.

MDSCs suppress both CD4⁺ and CD8⁺ T cells

To further interrogate the direct effect on T lymphocytes, purified CD11b⁺CD14⁻MHCII⁻ cells from a dog with osteosarcoma were co-cultured with purified CD4⁺ and CD8⁺ T cells from a healthy dog for 72 h. Non-stimulated cells and CD4⁺ and CD8⁺ cells co-incubated with healthy PBMCs were used as controls. As expected, CD11b⁺CD14⁻MHCII⁻ cells inhibited the proliferation of CD8⁺ (Fig. 4A) and CD4⁺ T cells (Fig. 4B) while PBMCs from a normal dog did not. Taken together, these data demonstrate that CD11b⁺CD14⁻MHCII⁻ cells are indeed functionally defined as canine MDSCs.

Figure 4: CD11b⁺CD14⁻MHCII⁻ cells suppress T cell proliferation.

Facs sorted CD11b⁺CD14⁻MHCII⁻ cells isolated from a dog with osteosarcoma or healthy PBMCs were co-incubated with mitogen-stimulated CD4⁺ and CD8⁺ T cells isolated from a healthy dog for 72 hs. No stimulated cells were used as negative control. Proliferative responses were measured by ³H-thymidine incorporation from experiments performed in triplicate. CPM, counts per minute. Mean ± SEM are shown.

CD11b⁺CD14⁻MHCII⁻ cells express hallmark MDSC-derived immunosuppressive factors

It has been shown that MDSCs can inhibit T cell function by the production of soluble factors such as arginase-1, reactive oxygen species, nitric oxide and TGF-β (8–10). In order to assess whether CD11b⁺CD14⁻MHCII⁻ cells from dogs with cancer could possibly utilize these mechanisms to mediate T cell suppression, we evaluated the expression of ARG1 and iNOS2, as well as the immunosuppressive cytokines TGF-β and IL-10, within this cell population and from PMNs isolated from peripheral blood of healthy dogs. PCR analysis of RNA extracted from FACS isolated CD11b⁺CD14⁻MHCII⁻cells confirmed the expression of ARG-1, iNOS2 enzymes and immunosuppressive cytokines TGF-β and IL-10 mRNA (Fig. 5A). In contrast, normal dog PMNs did not express ARG1, although iNOS, TGF-β and IL-10 mRNA were detectable (Fig. 5B). Because mRNA for ARG-1, iNOS2, TGF- β and IL-10 were all found, we conclude that these factors could play a role in the inhibition of T cell proliferation and effector function. However, since PMNs isolated from healthy dogs did not express detectable ARG-1 mRNA or impair T cell function, suggesting that ARG-1 may be a tumor-induced mechanism that MDSCs could employ for T cell suppression. This finding was not unexpected and has been previously documented in human MDSC studies [19].

Figure 5: CD11b⁺CD14⁻MHCII⁻ cells express MDSC-derived immunosuppressive factors.

RT-PCR analysis of FACS purified CD11b⁺CD14⁻MHCII⁻ cells detected expression of ARG1 and iNOS2, as well TGF-β and IL-10 immunosuppressive cytokines. ARG-1 expression was not detected in normal PMNs. CD11b⁺CD14⁻MHCII⁻ cells were isolated from the peripheral blood of a dog with osteosarcoma and PMNs were isolated from a healthy dog. NRT, RNA template in the absence of reverse transcriptase. Results are representative three experiments.

Discussion

The field of comparative oncology shows great promise to advance the development of novel therapeutics for pet dogs and human patients alike. However, the paucity of reagents and poorly defined immunophenotype of canine leukocytes has restrained our ability to understand tumor immunology in dogs with naturally occurring cancer. Our data demonstrates the existence of MDSCs in the peripheral blood of dogs, which are elevated in all types of advanced or metastatic cancer analyzed compared to early stage non-metastatic cancer and healthy controls. With this basic foundation of knowledge in place, it will now be possible to prospectively monitor MDSC burden in dogs treated with experimental drugs and immunotherapy. The CD11b⁺CD14⁻MHCII⁻ cell population that we defined as MDSC co-purified with PBMCs, had polymorphonuclear granulocytic morphology, suppressed T cell proliferation and effector function, expressed hallmark suppressive factors of human MDSC, and positively correlated with tumor burden. Proliferation assays revealed relatively weak proliferation in PBMCs from tumor-bearing dogs (Fig. 3A,B) compared to normal responders (Fig. 3C) in the absence of exogenous MDSC. This likely reflects elevated levels of endogenous (not experimentally added) MDSCs and regulatory T cells in the PBMCs from dogs with cancer. Furthermore, it is crucial to note that a second subset of MDSC that is more monocytic in nature is widely appreciated in murine and human tumor immunology. We found no evidence for selective expansion of a CD14⁺ monocyte-like cell in the blood of dogs with cancer. However, CD11b⁺MHCII⁻ cells that were purified from dogs with advanced cancer that were also CD14⁺ potently inhibited T cell proliferation (Figure S2), revealing that although monocytic MDSC are not a dominant population in dogs with cancer, they are indeed present. This finding of preferential expansion of granulocytic MDSC is not surprising and is in agreement with similar studies carried out in murine tumor models [15]. Overall, our data are consistent with a global state of immune suppression in dogs with advanced cancer that is likely attributable to several mechanisms.

The practical deliverable of this study is a simple three marker surface immunophenotype that can be used to prospectively monitor MDSC burden in dogs. We have performed pilot studies to look for additional markers. Specific preliminary results that are worth noting are as follows. We have been unable to demonstrate successful staining using anti-human CD66b antibodies. CD66b is an activation marker expressed on some human MDSC [19]. The most widely used marker for MDSC in the mouse is Gr-1, and an antibody against mouse Gr-1 cross-reacts nicely with canine cells, as does anti-mouse CD11b (Figure S1). Further studies will be required to determine if canine cells that are identified by anti-mouse Gr-1 and CD11b antibodies are indeed MDSCs.

One potential limitation of this study that many of the samples we analyzed were frozen, the thawed before analysis, which could have influenced cell viability. However, freeze-thaw did not significantly affect cell viability of either granulocytic or monocytic MDSC (Figure S3). We consider this a positive finding because canine MDSCs could be frozen from multiple time points in future prospective studies, then thawed and analyzed simultaneously to limit batch to batch variability. A second limitation is that the RT-PCR analysis of immunosuppressive molecules was qualitative, was performed on a small number of dogs (Table S3), and was not a direct comparison to matched healthy cells. We were not able to obtain adequate viable CD11b⁺CD14⁻MHCII⁻ cells from healthy dogs by FACS to directly compare to the same population from dogs with cancer due to their low frequency and apparently high rate of cell death following FACS. For this reason, normal PMNs isolated by gradient centrifugation were used for comparison in our studies. Quantitative mechanistic studies should be conducted to dissect which of the candidate molecules studied herein mediate T cell suppression. Additionally, some of the dogs had received treatment for their cancer. This is relevant because MDSC levels in human cancer patients have been shown to be influenced by prior therapy. It is also known that tumor burden and inflammation significantly affect circulating MDSC levels. Studies in mice have shown that accumulation and suppressive activity of MDSCs are regulated by the inflammatory milieu [20]. Thus treatment, such as surgical excision of the tumor, chemotherapy, and nonsteroidal anti-inflammatory drug (NSAID) administration, can alter the levels of these cells in the peripheral blood. Evaluation of the medical records of dogs in our study revealed that many dogs received some therapy prior to blood sample collection, which could have affected the levels of MDSCs in these samples (see Tables S1 and S2). However, Figure S4 demonstrates that treatment of dogs with advanced cancer did not significantly alter MDSC burden relative to dogs that had not been previously treated. Therefore, our study provides evidence that expanded MDSCs are likely a robust, general feature of cancer in canines despite genetic heterogeneity and a range of previous treatments (or lack of previous treatment).

In summary, we have identified a granulocytic subset of cells with immunosuppressive function that are elevated in dogs with advanced cancer that can be characterized as MDSCs. Canine MDSCs may be a potential target for therapeutic interventions in dogs with cancer. Furthermore, the study of MDSCs in dogs treated with experimental therapies should reveal unique insights into what might be expected in human patients. This cross-species comparison provides an attractive opportunity to move the field of translational medicine forward.

Figure S1

Mouse anti-CD11b and Gr-1 antibodies cross-react with canine samples. Fresh PBMCs from healthy dog and cancer patients were isolated by Ficoll, stained with anti-mouse CD11b and anti-mouse Gr-1 antibodies.

Figure S2

CD11b⁺CD14⁺MHCII⁻ cells demonstrate ability to suppressive T cell proliferation. (A) CD11b⁺CD14⁺MHCII⁻ cells were sorted from peripheral blood sample of an osteosarcoma dog (B) and co-cultured with healthy dog PBMCs in the presence of mitogen for 72 hs. Non-stimulated PBMCs were used as negative control and PBMCs co-cultured with healthy PMNs were used to control for the effect of adding cells to the assay. Proliferative responses were measured by ³H-thymidine incorporation. CPM, counts per minute. The experiment was performed in triplicate. Mean ± SEM are shown.

Figure S3

Frequency of MDSCs measured was not significantly altered by cryopreservation. MDSC percentages in fresh and frozen samples were assessed for comparison. Mean ±SEM are shown.

Figure S4

No significant effect of pretreatment on MDSC burden. Analysis of the average CD11b⁺CD14⁻MHCII⁻ population frequency in treated (n = 17) or untreated dogs with advanced stage or metastatic tumors (n = 13) compared to control dogs (n = 18). There was a significantly higher percentage of CD11b⁺CD14⁻MHCII⁻ cells in dogs with advanced cancer treated or untreated compared to healthy dogs (32.69±3.24%, 40.42±3.86% vs. 10.24±1.412%, respectively). N.S., not statistically significant (there was no significant difference between samples that had been treated compared to those from untreated samples). Mean ± SEM are shown (* indicates P<0.0001).

Table S1

Summary data for dogs with advanced stage or metastatic tumors.

Table S2

Summary data for dogs with early stage non-metastatic tumors.

Table S3

Table of cancer patient samples and the experiment in which the PBMCs were used.

Table S4

Primer sequences for genes evaluated by semi-quantitative PCR.

Acknowledgments

Share this article / Teilen Sie diesen Artikel