*3.2.1 Selection of CAR detection antibodies*

CAR-positive cells can be detected by MFC using direct or indirect fluoresceinlabeled antibodies against their extracellular domain (ECD). Initial clinical studies used sheep anti-mice IgG (polyclonal anti-IgG antibodies), which was not suitable for humanized CAR-T. Anti-idiotype monoclonal antibodies, antigen-Fc, protein L-based assays, and anti-linkers antibodies are commonly used CAR detection methods as well. Each method has different properties and shortcomings. For example, the anti-IgG antibodies and protein L have higher reagent stability but lower specificity to CAR, and antigen-Fc and anti-idiotype antibodies can detect CARs with very high specificity [36, 37, 69–75].

At present, most CAR protein detection antibodies are customized by CAR-T companies resulting in a lack of standardization in the assay. Therefore, an extremely strict quality control is needed for MFC. Firstly, since CAR is an unknown antigen and CAR-positive cells may have a high background because CAR-T cells are often large activated cells, the effect of compensation and fluorescence spillover should be eliminated by fluorescence minus one (FMO) in panel design. Secondly, fluorescence with high brightness should also be selected to reduce the possible false negative results caused by dim fluorescence. Third, in addition to the traditional isotype negative control, a group of cells processed with the same method but without transduction should be added as biological control, while successfully constructed CAR-T cells as a positive control, especially for those with low fluorescence intensity or without a gap between negative and positive cells. Fourth, the exclusion of dead cells and nonspecific binding are carried out by different methods, such as using dead/living cell dyes to exclude dead cells, CD14 to exclude monocytes, and Fc receptors blocking reagent or serum/IgG to eliminate nonspecific binding of IgG1 and IgG2a to Fc fragments. Fifth, the performance of new lots/shipments of antibodies and reagents should be compared with old ones to minimize inter-lot and even inter-shipment differences, which is more important for polyclonal antibodies. Sixth, in addition to strict quality control of MFC, it should also be compared with the qPCR method at the beginning of the study. Last but not the least, it is necessary to use the same antibody panel in one study [21–27, 36, 37, 69–75].

#### *3.2.2 Cellular kinetics at different time points*

After infusion, CAR-T kinetic detection is an important indicator to evaluate its effectiveness and safety. Although a big variety exists in different CAR-T and different studies, regular changes in CAR-T kinetics can be observed. For example, CAR-T cells begin to proliferate *in vivo* 4 days after CD19-CAR-T treatment, peak at 7–19 days, and most of them recover around 28–60 days [1–4, 21, 22]. The indicators reflecting CAR-T cell kinetics include direct (the proportion and number of CARpositive cells, and concentration of corresponding cytokines) and indirect one (the recovery kinetics of target antigens on cells mentioned in 2.3 above).

At present, data from many clinical trials show that CAR-T proliferation *in vivo* is significantly related to the therapy effect. Compared with patients with ineffective treatment, patients with effective treatment have a much higher CAR-T cell proliferation peak and the area under the curve (AUC) within one month of CAR-T infusion [32–37]. CAR-T < LOD is associated with B-cell recovery, and the consistent result of MFC and PCR has been verified in many studies [21–26, 33–37].

#### *3.2.3 Composition of CAR-positive cells*

MFC can detect the subsets of CAR-T cells (CD4 or CD8), including differentiation (naive, memory, and effector), activation (expression of activation markers), and inhibitory receptors (PD1, Tim3, LAG3, CTLA-4, and TIGIT). These markers

*Recent Developments in Application of Multiparametric Flow Cytometry in CAR-T… DOI: http://dx.doi.org/10.5772/intechopen.108836*

may correlate with clinical responses. However, the results of testing the proportion and number of CAR-positive cells in patients at different time points vary greatly. Therefore, the panel varies according to them.

Generally, a relatively detailed panel is chosen when the quality of the product is evaluated before infusion and when the proportion of CAR-positive cells is high in the early post-CAR-T infusion period. However, the basic panel may be chosen for the consideration of price, sample size, and the low concentration of CAR-positive cells. Taking CD19-CAR-T as an example, the basic panel may include CD4, CD8, CD3, CD19, CD16 + CD56, CD45, CAR, and CD14, to evaluate the common lymphocyte composition, CD4/CD8 ratio, residual CD19-positive cells, and recovery kinetics, in addition to accurate detection of CAR-positive cells. Further assays include CD25/ CD127 for CD25dim/CD127+ regulatory T cells (Treg) and CD25 high-activated cells. The different effector and memory T subsets are evaluated by using CCR7 (CD197) and CD45 RA, such as naive T cells (TN, CD197+/CD45RA+), central memory T cells (TCM, CD197+/CD45RA−), effector memory T cells (TEM, CD197−/CD45RA−), and effector T cells (TEFF, CD197−/CD45RA+).

If further evaluated, CD38/HLA-DR assay for activated T cells will be added, and CD38+ or HLA-DR + or double-positive (DP) activated subsets can be acquired. Stem cell memory-like T cells (TSCM) expressing markers, such as CD45RA, CCR7 (optional CD62L), CD95, CD27, CD28, CD127, CD11adim, and lacking CD45RO, these cells can be detected by simply adding CD95 to CD197/CD45RA panel. Studies have shown that TSCM has the ability to self-renew and differentiate [22–27, 78]. The immune composition of CAR-T products is associated with antitumor efficacy, and CAR-T cells with TN, TSCM, and TCM phenotypes have been found to have longer *in vivo* persistence and higher antitumor efficacy [32–37]. After infusion, CAR-positive cells are mainly TEM in the expansion phase, which will last in a long term, and TN begins to appear later [22].

It has been found that specific populations of the donor T cells identified by MFC can predict the prognosis, especially T subsets that co-express certain suppressive signals. Finney [36] found that increased frequency of LAG-3+/TNF-αlow CD8+ T cells in PB apheresis product was related to relapse of pediatric B-ALL patients treated with anti-CD19 CAR-T.

High expression of target antigens by tumor cells can effectively stimulate CAR-T cell proliferation [36, 37], but high tumor load is in turn a poor prognostic factor for CAR-T [40, 41], the paradox may be caused by the expansion of certain CAR-T subpopulations expressing inhibitory signals hindering CAR-T cell expansion *in vivo* [36, 37, 77, 78].

#### *3.2.4 Detection of other cells*

At present most CAR-T cells are derived from the patient's own immune cells, and although most patients' tumor cells will die during *in vitro* culture, there will be some cases in which the tumor cells remain alive or even survive off-target [75–77]. Therefore, a rigorous MRD test of tumor cells should be performed in the quality evaluation of the product. The MRD panel and data analysis should be performed according to the method described in 3.2 above.

#### *3.2.5 Absolute counts of CAR-positive cells*

Absolute counting of CAR-T cells can be performed using either a single- or dual-platform method, where the single-platform can be done using the volumetric method or the absolute counting microbeads method [79].

The single-platform method is considered to be more accurate than the dualplatform method and requires less sample volume. However, since the single platform method cannot be washed, there may be a high background signal or failure to correctly detect some antibodies or fluorescent dyes. Therefore, any method can be chosen, but use the same method in a total clinical study, including multicenter studies.

The quality control of CAR-T cell enumeration is referenced to that of CD34+ hematopoietic stem cells [15, 79], with a minimum collection of 100 positive cells, and the LOD and lower limit of quantitation (LLOQ ) for MRD assays should be verified. Mostly more than 1,000,000 events are recommended to acquire for detailed analysis of CAR-T cells when the percentages are high, and for accurate enumeration at later time points when the percentages may be down to less than 10−4 [15, 16, 19, 20]. In the lymphocyte ablation phase, as many as possible cells are acquired, and a minimum of 100,000 cells is recommended [15, 22–26].

#### **3.3 Immune cell subsets**

Immune cell subset detection has many similarities and overlaps with CAR-T cell detection. Therefore CAR-positive cells can be detected along with the immune cell subset detection before infusion and early days after CAR-T treatment. When CAR expression cells cannot be detected twice in a row, immune subset detection will last for a longer time without CAR antibody [15, 22–26].

Because CAR-T is a kind of immunotherapy, including the specific killing of target antigen-positive cells and nonspecific killing of CAR-negative cells, there is a positive and negative regulation balance between efficacy and side effects. Nowadays 8 or more colors panel are recommended to analyze detailed subsets in CAR positive and negative parts with a similar panel [15, 22–26].

The titration of all monoclonal antibodies is highly recommended before performing actual experiments. An isotype control should be used at the same concentration of the antibody of interest. DAPI or 7AAD or other dye to distinguish live or dead cells may be added. In addition, with the development of immunology, the different configuration of instruments, and the intersection of various antigens, the antibody combination to define the same functional subgroups in different studies maybe not the same, and the detection panels are also very heterogeneous [15, 22–26]. Therefore, it is necessary to adopt a consistent panel in a study, especially a multicenter study.

#### **3.4 Cytokine detection**

#### *3.4.1 Cytokines*

Due to different processes and cell sources of CAR-T, different cytokines may be produced. For example, CAR-T from PB CD3+ T cells may lead to the formation of multiple cytokines. CD4+ T cell-related factors are IL-2, IL-4, IL-5, IL-10, IL −13, and IL-17, while CAR-T from CD8+ T cells mainly produces IFN-γ, TNF-α, perforin, and granzyme B. CAR-T proliferation and efficacy are related to most important cytokines, so cytokine detection is an important assay for quality evaluation of the product in development process and CAR-T efficacy evaluation after the immunotherapy [26–30, 36, 37, 80, 81].

However, the toxicity of CAR-T is always along with its effectiveness. The most common toxic side effects are cytokine release syndrome (CRS). CRS is caused by the

#### *Recent Developments in Application of Multiparametric Flow Cytometry in CAR-T… DOI: http://dx.doi.org/10.5772/intechopen.108836*

release of a large number of inflammatory factors by activated immune cells. IL-6, IL-1, and IFN-γ are all related to CRS. Any CAR-T or other immune cells that cause IFN-γ to elevate will aggravate CRS, which is more obvious in CRS level ≥ 3 [26, 27, 36, 37, 41]. Each CAR-T study uses cytokines to evaluate the activation characteristics of T cells, and basically includes IL-6, IFN-γ, TNF-α, and IL-2. The selection of other cytokines varies from study to study, including IL-1RA, IL-1β, IL-4, IL-5, IL-8, IL-10, IL-12, IL-13, IL-15, IL-17, IL-18, IL-21, IL-22, IL-23, IL-31, IL-36, monocyte chemoattractant protein (MCP)-1, perforin, granzyme B, erythropoietin, granulocyte/macrophage colony-stimulating factor (GM-CSF), soluble CD25, (sCD25), ferritin, CCL20, REG3a, ST-2, TNFRI, and elafin [1, 4, 26–30, 36, 37, 80, 81].

#### *3.4.2 Cytokine detection method*

Although some studies used enzyme-linked immune sorbent assay (ELISA) [31, 34] or ELISpot assay [82], recently MFC has been used in CAR-T cytokine detection with requirements for more cytokines, or subsets that secrete cytokines. Cytometric bead array (CBA) [28] or couple intracellular cytokines staining [29, 30, 36, 37] are two main kinds of cytokines assay methods by MFC, and each has its own advantages and disadvantages. The advantages of CBA are fast, simple, sensitive, repeatable, flexible, and high throughput, which can detect dozens of cytokines in a short time with rare samples [28]. However, the unique advantage of intracellular cytokines staining lies in the simultaneous detection of cellular immunophenotyping and intracellular cytokines, and it is the only one to allocate cytokines to subsets without the help of cell isolation [29, 30, 36, 37].

Clinical studies can choose one of them, but a complete study, especially a multicenter study, should use the same method from beginning to end because there is a lack of comparability between different methods.

#### **3.5 Tumor microenvironment and immune checkpoint detection**

TME is a complex network of local immune cells, stromal cells, signaling molecules and cytokines secreted by these cells. In the study of solid tumors, signal networks represented by PD1 and PDL1 have achieved remarkable results in mechanism research and immunotherapy [13, 14, 36, 37, 54]. The immune microenvironment of hematologic malignancies is more complicated. MFC can detect a variety of immune cells and immune signals, so it has become the main research tool in this field in recent years [22–27, 36, 37].

The biggest problem of CAR-T is resistance and relapse [1–7], involving a variety of complex mechanisms [8–14, 34, 35, 40, 41, 66, 67, 72–78], among which the study of immune-suppressive signals and immune microenvironment has been the focus of attention in recent years: (1) T-cell exhaustion, effector T-cell reduction, and the increased expression of inhibiting receptors. The high expression of LAG-3 and PD-1 and low expression of TNF-α in CD8+ T cells are associated with CAR-T loss of function, which will reduce the antitumor ability of CAR-T cells and lead to CD19-positive relapse. Target expression cell recovery in PB is a signal of the weakening of CAR-T cell function. CAR-T exhaustion is one of the factors. In some studies, exhaustion-related signals, such as PD1 (CD279), LAG-3(CD223), CTLA-4(CD152), and Tim3 (CD366), are included in the CAR-T detection panel, hoping to find the role of immune checkpoints, and trying to relieve the inhibitory signals by targeted drugs, such as PD1 monoclonal antibody or PDL1-CAR-T, to acquire long-term OS [13, 14, 54, 72–78]. (2) Immune aging and age-related T-cell quality. With the introduction of theories about immune age and immunosenescence, as well as the discovery of age-related CAR-T cell phenotypes, studies begin to include immunosenescence-related markers. Senescent T cells exhibit some phenotypes, including downregulation of CD27, CD28, and upregulation of CD57, KLRG-1, Tim-3, TIGIT, and CD45RA [83]. The emergence of these phenotypes is a signal of the weakening of CAR-T cells [22–27]. (3) Other signaling-related studies on the antagonization of the CAR-T function. When CD19-negative tumor recurrences, tumor cells may have high expression of CD123, and CD123-CAR-T cell therapy may be effective; it may be related to the increased expression of Bcl-2. Thus, monitoring Bcl-2 in tumor cells and Bcl-2 antagonist treatment in Bcl-2 highly expressed patients may be effective [72–78]. (4) Inhibitive BM microenvironment. Myeloid-derived suppressor cells (MDSC), TAMs, and Treg inhibit CAR-T cell proliferation and function. Detecting these inhibitory signals through MFC contributes to the CAR-T mechanism research. Blocking these signals can restore lymphocyte function. Combined CD30-CAR-T and anti-PD-1 therapy have showed promising results in CD30-positive lymphomas [54].

Similar to immune subsets, there are also some differences in the panels of TME and immune checkpoints. CD15/CD33/CD11b/HLA-DR/CD16/CD4/CD14/ CD45 is one recommended panel to detect MDSC subgroups, and CD27, CD28, CD57, and PD1 (optional LAG-3, CTLA-4, and Tim3) may be simple supplements to common immune subsets panel for immune checkpoints and immunosenescence [22–27]. CD123 and Bcl-2 should be added to MRD panel. Similar to the previous requirements, it is necessary to perform quality control, stick to same panel, and keep a high degree of standardization, stability, and good reproducibility.

### **4. Advancement in MFC promote CAR-T study**

As technology advances, MFC evolves toward more and more channels, of which mass cytometry and full spectral flow cytometry are two major trends [27, 84]. The traditional MFC is limited by fluorescence channels, so the tumorrelated and immune-related assays are basically carried out separately. In future, with the introduction of more than 20 or even 40 multiparameter MFC, it will realize one complicated panel to simultaneously finish the above-related assays, saving costs and samples, and more importantly, obtaining geometric growth of big data information [27, 84].

Other MFC-related latest advances, such as single-cell sequencing, high-dimensional data analysis, and artificial intelligence, will also enter the field of CAR-T research with the application of MFC in CAR-T. These new advances will certainly promote the realization of MFC-assisted CAR-T efficacy-related factor analysis and obtain standardized treatment formula.

Therefore, with the improvement of more clinical information and more detailed MFC data, it is possible for us to obtain a formula for the best performance of CAR-T. We can obtain the prediction of each patient by bringing the number of malignant cells in different patients with different tumors, the sum of all tumor antigen expressions, immunosuppressive signals, and the immune-stimulative and immunesuppressive components of CAR-T subsets into the formula. If the corrective strategies of various inhibitory factors are added to the formula, it is expected that in future we will provide a standardized prediction of prognosis and treatment guidance for obtaining the best curative effect of CAR-T therapy.

*Recent Developments in Application of Multiparametric Flow Cytometry in CAR-T… DOI: http://dx.doi.org/10.5772/intechopen.108836*
