Monitoring PD-1-Blocking Antibodies Bound to T Cells Derived from a Drop of Peripheral Blood

We developed a simple flow cytometry assay for evaluating the binding of PD-1–blocking antibodies to T cells, requiring only a drop of peripheral blood from cancer patients.

Immune checkpoint inhibitors, including PD-1–blocking antibodies, have significantly improved treatment outcomes in various types of cancer. The pharmacological efficacy of these immunotherapies is long lasting, extending even beyond the discontinuation of their injections, due to persistent blood concentrations. Here we developed a simple flow cytometry assay to evaluate the T cell binding status of the PD-1–blocking antibodies nivolumab and pembrolizumab. Like a glucose test, this assay requires just a single drop of peripheral blood. Visualizing antibody binding on T cells is more reliable than measuring antibody blood concentrations. In addition, if necessary, we can potentially analyze many distinctive immune-related markers on T cells bound to PD-1–blocking antibodies. Thus, this is a simple and minimally invasive strategy to analyze the pharmacological effect of PD-1–blocking antibodies in cancer patients.

PD-1–blocking antibodies have become the standard choice for treatment of various types of cancer, including non-small cell lung cancer (NSCLC)^1,2,3,4. They show a remarkable therapeutic effect in a subset of cancer patients who have not responded to conventional cytotoxic chemotherapies. However, immune checkpoint inhibitors (ICIs), which include PD-1–blocking antibodies, can cause a unique and distinct spectrum of adverse events, termed immune-related adverse events (irAEs)⁵. Although irAEs can affect almost all tissues, they are most commonly observed in the gastrointestinal tract, endocrine glands, skin, and liver, and they can cause pruritus, rash, nausea, diarrhea, and thyroid disorders^6,7. In general, most irAEs appear within 1 to 2 months after the initiation of ICIs. However, in some cases, they can occur later than 1 year after the beginning of treatment or even after treatment cessation^6,7. They also cause various symptoms that may be difficult to discriminate from other pathologies. Thus, it can be challenging to promptly diagnose irAEs and treat them appropriately. irAEs can affect all tissues, and their onset is strongly influenced by circulating immune cells, especially T cells bound to PD-1–blocking antibodies. Therefore, a straightforward and minimally invasive method to monitor antibody-targeted T cells is important in clinical settings.

Here, we developed a simple method to assess the binding of PD-1–blocking antibodies to T cells using a drop of peripheral whole blood from cancer patients who received nivolumab or pembrolizumab. Using this approach, we were able to monitor each of the following: 1) the duration of antibody binding to T cells, 2) the occupancy of T cell PD-1 molecules by therapeutic antibodies, and 3) the activation status and immunological features of T cells. This method is a modification of a previously reported technique⁸. The amount of blood required is almost the same as that needed for a glucose test, and the approach does not require mononuclear cell enrichment or co-culturing with PD-1–blocking antibodies. We confirmed that this method can also be performed using frozen samples, including peripheral blood mononuclear cells (PBMCs) and cells from pleural effusion, pericardial effusion, bronchoalveolar lavage fluid, and cerebrospinal fluid, suggesting that this strategy may be useful in the context of a multicenter study. This method may facilitate the early diagnosis of irAEs, and also help to determine the appropriate immunosuppressive treatments to control their symptoms and to identify the optimal times to initiate subsequent therapies after PD-1 inhibitors.

Sampling was performed during routine clinical procedures. All human samples were obtained after informed consent was provided by the subjects, in accordance with the Declaration of Helsinki and with the approval of the ethical review board of the Graduate School of Medicine, Osaka University, Japan (15383 and 752).

1. Whole Blood Sample Preparation and Staining

1. Collect whole blood samples into blood collection tubes containing ethylene diamine tetra-acetic acid (EDTA).
   NOTE: Blood collection can be performed using either a regular needle or a blood lancet.
2. Transfer 20 µL of whole blood samples to 5 mL round-bottom polystyrene flow cytometry tubes.
   NOTE: To reduce non-specific binding of cells to tubes, 1 mL of 2% fetal bovine serum (FBS) in phosphate buffered saline (PBS) is added into tubes and vortexed for 10 s before application to samples.
3. Add 20 µL of 2% FBS in PBS.
4. Add 10 µL of human-specific FcR blocking reagent. Mix well and incubate for 15 min at room temperature.
5. Add 500 µL of red blood cell lysis buffer. Mix well and incubate for 10 min at room temperature.
6. Add 4 mL of 2% FBS in PBS and spin down cells at 400 x g (1,500 rpm) for 5 min at 4 °C. Remove supernatant by aspiration.
7. Repeat the wash and aspiration process described in step 1.6.
8. Re-suspend cells in 100 µL of 2% FBS in PBS and divide into two tubes of 50 µL each.
9. Add surface marker antibodies (Table 1). Mix well and incubate for 20 min at room temperature in the dark.
   NOTE: When profiling the T cell immune status, the number of markers can be increased based on the flow cytometry machine quality.
10. Wash samples 2x as described in step 1.6.
11. Resuspend cells in 200 µL of 2% FBS in PBS.

2. Flow Cytometric Analysis

1. Insert tubes into the flow cytometer and acquire cells, basically following the recommended protocol⁹.
2. Record 10,000 events as the lymphocyte gate (Figure 1A) and export flow data as .fcs files for analysis.
3. Open files in the analysis software. Visualize the cells on a forward scatter (FSC) (A) vs. side scatter (SSC) (A) plot and gate lymphocytes (Figure 1A).
4. After selecting single cells using FSC (H) vs. FSC (W) and SSC (H) vs. SSC (W) (Figure 1B) and displaying them on a CD3 vs. CD8 or CD3 vs. CD4 plot, gate the CD8 T cells and CD4 T cells, respectively (Figure 1C).
5. After selecting the gated cells and displaying them on a PD-1 vs. human IgG4 plot, identify PD-1–blocking antibody–bound CD8 and CD4 T cells based on isotype control (Figure 1D).

The gating strategy and flow cytometry analysis (Figure 1) can detect PD-1–blocking antibody binding to T cells obtained from a drop of NSCLC patient peripheral blood. Before PD-1–blocking antibody is administered, no human IgG4-positive CD8 or CD4 T cells are present, and PD-1 expression can be confirmed by a PD-1–detecting antibody (EH12.1) (Figure 2A). After nivolumab or pembrolizumab administration, IgG4 (nivolumab, pembrolizumab) can be detected on T cells by anti-IgG4 antibody (HP6025) whereas the PD-1–detecting antibody EH12.1 does not recognize any PD-1 on T cells because therapeutic PD-1 antibodies disturb EH12.1 binding. This means that we are indirectly measuring the therapeutic binding of PD-1–blocking antibody based on the lack of binding of PD-1–detecting antibody. Representative data show the different binding statuses of PD-1–blocking antibody (Figure 2A). Nivolumab and pembrolizumab binding and occupancy of PD-1 on T cells decrease over time¹⁰, and there is partial binding (PB), which is shown in the double-positive area, and finally complete loss of binding (LB) (Figure 2B).

Figure 1: Representative gating strategy to evaluate PD-1–blocking antibody binding to T cells from a drop of peripheral blood. (A) FSC (A) vs. SSC (A) plot and gating of lymphocytes. (B) Doublets are excluded by drawing gates around the main cell population on plots of FSC (H) vs. FSC (W) and SSC (H) vs. SSC (W). (C) CD3 vs. CD8 plot (upper) and CD3 vs. CD4 plot (lower) and gating of CD8 T cells and CD4 T cells, respectively. (D) PD-1 vs. human IgG4 plot and detection of the binding of nivolumab (a PD-1–blocking antibody) to CD8 and CD4 T cells. Orange dots and black dots indicate anti-IgG4 antibody and isotype control staining, respectively. Please click here to view a larger version of this figure.

Figure 2: Representative flow cytometry analysis indicating the change in binding status of PD-1–blocking antibodies. (A) Staining of PD-1 and human IgG4 in CD8 T cells from ICI-pretreated patient blood was evaluated by flow cytometry (left). Ten microliters of pretreated blood was treated with serial dilution of nivolumab for 15 min, and steps 1.4 to 1.10 of the protocol were completed. Complete binding (CB) (red), partial binding (PB) (blue), and loss of binding (LB) (green) are defined by the indicated gates. (B) The status of PD-1–blocking antibody binding to CD8 T cells was analyzed at the follow-up time points, as indicated, in NSCLC patients who discontinued nivolumab and pembrolizumab. Please click here to view a larger version of this figure.

Table 1: Antibodies used in flow cytometric analysis.

In this article, we report a method using a flow cytometer to detect PD-1–blocking antibodies bound to T cells derived from a drop of peripheral blood, which we originally developed for nivolumab detection¹⁰. Although this technique is very simple and easy to perform, two important points should be noted in order to obtain accurate results. One is that to detect PD-1 molecules, an appropriate antibody that competes with nivolumab and pembrolizumab should be used. This issue was evaluated in a previous study¹¹. The other is that RBC lysis should be performed thoroughly before surface staining. As long as an isotype control is established for each assay to determine the gate of the IgG4-positive cluster, the frequency is unaffected. However, with this protocol, when RBC lysis is insufficient there may be reductions of both the T cell count in the lymphocyte gate and of the intensity of each surface marker. In addition, the RBC lysis step must be performed before surface staining, otherwise the anti-IgG4 antibody (HP6025) does not function properly.

The limitation of this method is that the population of T cells bound to PD-1–blocking antibodies include specific T cell clones responsible for therapeutic effects and irAEs; however, the frequencies of these specific clones among the antibody-bound population are quite low. Therefore, we still need to enrich the specific target using certain markers, for instance CD39¹². Another limitation is that the fluorescence intensity of IgG4 is not high in some cases, which makes it difficult to determine binding status (i.e., PB and LB).

Other studies have monitored therapeutic PD-1 antibodies in the blood to evaluate pharmacokinetics. Measuring the plasma concentration of nivolumab or pembrolizumab is essential to determine how the residual amount of these antibodies in blood correlates with time. However, we previously reported that the concentration of nivolumab does not completely correlate with residual binding to T cells¹⁰. Compared to the measurement of plasma antibody concentration¹³, our method may be more appropriate for evaluating the residual binding of anti–PD-1 antibodies to T cells. The original method of monitoring nivolumab binding to T cells in blood required incubation with fluorescence-labeled nivolumab⁸. We simplified this approach and succeeded in substantially reducing the assay time and the amount of blood needed for analysis.

Frozen cells can also be analyzed using this assay, suggesting that this method is a good fit for multicenter studies. The usefulness of this approach will be further enhanced by combining monitoring binding status with other immunological markers, such as Ki-67, of T cells bound to PD-1–blocking antibodies¹⁰. Future studies should use our method in conjunction with single-cell RNA sequencing and T cell receptor sequencing to try to identify and characterize the specific subsets of T cells that are bound to PD-1–blocking antibodies and are responsible for irAEs.

The authors have nothing to disclose.

This work was supported by grants to S.K. from the Japan Society for the Promotion of Science KAKENHI (JP17K16045) and the Japan Agency for Medical Research and Development (JP18cm0106335 and 19cm0106310).