Flow cytometry (FC) is widely used to diagnose and characterize lymphoma and leukemia in dogs. It allows the differentiation of reactive from neoplastic conditions by identifying a predominant homogeneous population or immunophenotypic aberrancies (1). Conversely, its use is still limited in cats because of the low availability of species-specific monoclonal antibodies and the high frequency of samples unsuitable for processing (2).

One of the most important aspects for differentiating a reactive versus a neoplastic process is the identification of cell clonality. During their development, the genes encoding the T-cell receptor gamma (TCRγ) and Immunoglobulin heavy chain receptor (IgH) undergo rearrangement. Thus, virtually each rearrangement is unique to individual lymphocytes. In a reactive population, a polyclonal expansion would be observed, while a single or predominant rearrangement indicates expansion of a population with the same antigen receptor gene. In cats, identification of clonality of TCRγ or IgH is currently performed by PCR for Antigen Receptor Rearrangement (PARR) (3–5). However, while all neoplasms are clonal in origin, not all clonal lymphocyte expansions are neoplastic and can be secondary to infectious or inflammatory diseases.

In this scenario, FC analysis of lymphoid population could be a support in differentiating truly neoplastic samples. Unfortunately, information about immunophenotypic characteristics of lymphoid subsets using FC in cats is limited, with few reports carried out in peripheral blood (PB) (6–8) and in lymph nodes (LN) (9, 10). In particular, currently there is no reference information on lymphoid subsets in PB of healthy cats, other than the percentage of CD4 + and CD8 + cells (6, 7). A deep knowledge of the composition and FC features of lymphoid circulating cells in normal and pathological conditions other than lymphoid neoplasms is mandatory, for a better interpretation of the data obtained on routine diagnostic samples, ultimately leading to a more accurate diagnosis.

The aim of this study is to describe FC features of lymphoid subsets in PB from healthy and diseased cats, in terms of percentage of different populations, cell size and degree of antigen expression, in order to provide reference information possibly useful for differentiating between reactive and neoplastic processes.

Overall, 20 feline PB samples were included in the present study, including 9 (45%) healthy cats, and 11 (55%) diseased cats (6 in Group1 and 5 in Group2).

Based on 7-AAD staining, all samples had ≥97.0% vital cells.

To the authors knowledge, this is the first study describing FC features of circulating lymphoid subsets in healthy cats using a 10 antibody panel. Data on diseased cats with and without lymphoid neoplasia are also presented.

Flow cytometry is only at its debut in feline oncology. In the past, indeed, it was mentioned in the diagnostic workup of occasional case reports (14, 15), while studies specifically focused on FC analysis of oncological samples date back to less than one decade ago (2, 8, 9, 16, 17). Importantly, data on non-neoplastic cells in cats are limited to two recent studies (10, 12). Herein, we applied a large antibody panel and report a number of FC features for each lymphoid cell subset in healthy cats, including absolute count, cell size (expressed as nFSC) and degree of antigen expression (expressed as MFIratio).

While designing the antibody panel, we encountered some unexpected findings with CD9 antibody, which prevented us from confirming the cross-reactivity with feline antigen reported by the manufacturer. Still, we retained the antibody in the panel since we believe that our findings could help improve FC analysis of feline samples, irrespective of the precise antigen detected by the antibody. According to the published literature, CD9 should be expressed on all leukocytes and platelets (18–20). However, we detected no CD9 + leukocyte. Unfortunately, CD9 expression has never been tested on feline fresh PB samples. Thus, we could not define whether our findings are linked to the antibody clone or to minimal to null CD9 expression on circulating feline WBC. However, we found a variable amount of CD9 + events among samples and among tubes within single samples. Better separation of WBC subclasses was obtained in the morphological scattergram when excluding those CD9 + events (Supplementary Figure 2). This fact, together with the peculiar distribution of CD9 + events in the morphological scattergram, and concomitant negative staining for the panleukocyte markers CD44 and CD18, prompted us to speculate that those events could represent platelet clumps. Further studies are needed to confirm this hypothesis.

Overall, three main populations were identifiable in healthy subjects and Group1 diseased cats, namely CD5 + CD45R-, CD21 + CD45R + and CD5 + CD45R+, in decreasing order of prevalence. CD45R has been used in the literature to identify B-cells in cats (21, 22). However, its expression on effector T-cytotoxic cells and neoplastic T-cell has also been reported (23–25). Our results suggest that CD45R expression is not limited to B-cells even in normal conditions. Concomitant use of additional B- and T-cell markers is necessary to accurately identify lymphoid subsets. Additionally, analysis of CD45R- and MHCII-MFIratios could be of aid in discriminating among the 3 subclasses, based on the significant differences we reported, in both healthy and Group1 diseased cats. Unfortunately, we did not couple CD45R with further T-cell markers, such as CD4 and CD8. Thus, the full antigenic pattern of CD5 + CD45R + cells and their the biological role still have to be defined.

When comparing healthy and Group1 diseased cats, an increase in the cell size of all T-cell subsets was noted. This could be attributed to their activation, although not all cats had inflammatory/infectious diseases. Irrespective of the underlying mechanisms, this difference is relevant from a diagnostic point of view, since it remarks that moderate increase in T-cell size alone should not be considered a criterium for diagnosing lymphoid neoplasia in cats.

When considering CD4 and CD8 expression, herein we confirmed the higher prevalence of CD4 + CD8- cells already reported in healthy cats in literature (6, 26). More interestingly, we reported a small amount of CD4 + CD8 + cells in >80% samples, without differences between healthy and Group1 diseased cats. It seems reasonable to argue that in the two remaining cats (1 healthy and 1 diseased cat) CD4 + CD8 + cells count was lower than the detection limit of the technique, rather than being truly absent. In the literature, low percentages of CD4 + CD8 + cells have been found in normal or reactive lymph nodes in cats (10), whereas higher proportions can be found in mediastinal masses. This is due to the fact that CD4 + CD8 + cells represent a transient stage during T-cell maturation (27). The percentage of CD4 + CD8 + cells in mediastinal masses is a leading criterium for FC differential between mediastinal lymphoma and thymoma in dogs, but the same is not valid in cats (16, 28). Altogether, our results and the data already published suggest that CD4 + CD8 + phenotype should not be considered a hallmark of neoplasia in cats.

We also found a low number of CD134 + cells in circulation without differences between healthy and Group1 diseased cats. CD134, also named OX40, is a member of the tumor necrosis factor (TNF) receptor superfamily, and works as a co-stimulator of lymphocyte activation, preventing apoptosis and enhancing proliferation (29, 30). In cats, this molecule is considered a key-driver of the progression of FIV infection, although almost all studies on this topic report in-vitro results (31–33). All cats included in the present study tested negative for FIV antibodies within the last month prior to sampling. Thus, in the future, it could be interesting to test the concentration of CD134 + cells in FIV-infected cats with different stage of disease, in order to assess its possible clinical relevance in the management of such patients. In addition, 3 out of 5 neoplastic samples stained positive for CD134. To the authors’ knowledge, this is the first report of CD134 expression in neoplastic cells in cats. Thus, it is still unknown whether it reflects a different stage of maturation of the cells, or has any prognostic implication.

Concerning the data from 5 cats diagnosed with probable T-CLL we present herein, T-cell lineage was confirmed in all cases by PARR TCRγ monoclonal results and FC expression of CD4. Our results agree with a recent study, which reported a high prevalence of CD4 + neoplastic lymphocytosis in cats, whereas B-cell lymphocytosis was generally not clonal, and indeed not represented in our caseload (8). Of notice, Rout and colleagues also identified a less common “CD5low” subgroup, bearing a worse prognosis. In contrast, the majority of the cases in the present study (3 out of 4 CD5+) had lower CD5-MFIratio than normal T-cells, and one case was even CD5-. Unfortunately, we lack follow-up data for our cohort, thus preventing comparison with the survival data for the “CD5low” disease.

Neoplastic cells were similar to those of the non-neoplastic counterparts, either in healthy or Group1 diseased cats, with the only exception of CD5-MFIratio (Figure 3). In particular, nFSC values overlapped in Group1 and Group2 diseased cats (Figure 3). Once more, this fact suggests that cell size alone should not be used to differentiate between neoplastic and non-neoplastic cells in cats. Another interesting finding is the level of expression of CD18. We already reported that this integrin is expressed at low levels in circulating lymphocytes in healthy cats (12). Based on our present results, CD18 seems to be overexpressed on neoplastic cells. On one side, this could play a role in the spread of the disease, since this molecule is involved in leukocyte rolling and extravasation (34). On the other hand, the differential degree of expression could serve as a laboratory tool to be included in the diagnostic process for suspected feline leukemias. Further studies are needed to support both hypotheses.

The low number of samples included represent the main limitation of the present study. This is particularly true for diseased cats, where subjects with different diseases were included in Group1, and all Group2 cats had a T-cell neoplasia. Also, the cats included had different breed, sex and age possibly biasing our results. Because of this, they should not be interpreted as reference interval for healthy cats or laboratory parameters to differentiate among clinical conditions. Rather, they represent a basis to improve FC analysis of feline samples in a diagnostic setting. Finally, the lack of follow-up data for cats with probable CLL prevented us from making any hypothesis about possible prognostic role of the different phenotypes we recorded.

In conclusion, this is the first study reporting FC features of lymphocyte subsets in cats, applying a large antibody panel, and highlighting some differences between healthy and diseased cats, either with neoplastic or non-neoplastic conditions. The antibody panel we present is composed of fluorochrome-conjugated antibodies available in the market, and as such can be easily reproduced in any veterinary FC facility. However, it is important to note that some of the antibody clones we used are reported not to work on preserved samples (35). Thus, the panel is suitable only to analyze fresh samples, which is overall generally recommended in FC practice. Future studies should address the biological and diagnostic relevance of FC data in cats.