Methodological considerations for the high sensitivity detection of multiple myeloma measurable residual disease

Abstract

Background: Recent advances in therapeutic interventions have dramatically improved complete response rates in patients with multiple myeloma (MM). The ability to identify residual myeloma cells (e.g., measurable residual disease [MRD]) can provide valuable information pertaining to patient's depth of response to therapy and risk of relapse. Multiparametric flow cytometry is an excellent technique to monitor MRD and has been demonstrated to correlate with patient outcome post-treatment. To achieve the high sensitivity (one abnormal cell in 10⁵ -10⁶ cells) required for MRD evaluation, millions of cells have to be acquired and conventional immunophenotyping protocols are unable to attain these numbers, indicating the needs for alternative flow cytometric staining procedures. A bulk, "Pre-lysis" method is the consensus approach for staining large number of cells, requires two red blood cell lysis steps, and can adversely affect epitope density. In this study, we tested the "Pooled-tube" and "Dextran Sedimentation" staining procedures and correlated them with the "Pre-lysis" method as potential alternative approaches.

Methods: A total of 22 bone marrow aspirates from patients with plasma cell (PC) dyscrasia were processed in parallel using the "Pre-lysis," "Pooled-tube," and "Dextran Sedimentation" techniques. Stain indices were calculated and compared to assess their impacts on staining performance for each antibody used in the consensus panel. The recovery of normal and abnormal PCs, mast cells, and B cell precursors was enumerated and compared after their counts were normalized using fluorescent beads. The limit of blank, limit of detection, and lower limit of quantification were established using serial dilution experiments.

Results: The staining performances of CD19 PECy7, CD27 BV510, CD81 APCH7, and CD138 BV421 were improved using the "Pooled-tube" method when compared to "Pre-lysis." "Pre-lysis" was better at resolving CD56 using clone C5.9 but our results demonstrated similar improvement can also be achieved by "Pooled-tube" when alternative CD56 PE clones were used. "Dextran sedimentation" yielded similar staining results when compared to "Pre-lysis" for all the markers analyzed. The "Pooled-tube" method, when normalized to "Pre-lysis," recovered higher numbers of total PCs (1.2 ± 0.2 times higher; p = .049), normal PCs (1.4 ± 0.26; p = .007), mast cells (1.46 ± 0.27; p = .003), and B cell precursors (1.42 ± 0.3; p = .011), but not abnormal PCs (1.09 ± 0.2; p = .352). There was no evidence that the recovery of cells was different between "Pre-lysis" versus "Dextran Sedimentation." All three flow cytometric assays achieved a minimum sensitivity of 10^-5 and approached that of 10^-6 for detecting rare events.

Conclusion: Both "Pooled-tube" and "Dextran Sedimentation" staining procedures were comparable to the "Pre-lysis" method and are suitable high sensitivity flow cytometric approaches that can be used to process bone marrow samples for MM MRD testing.

Keywords: dextran sedimentation; flow cytometry; high sensitivity; measurable residual disease; multiple myeloma; pooled-tube; pre-lysis.

Figure 1.. Correlation of the recovery of various bone marrow subpopulations after samples were stained using ‘Pooled-tube’ and ‘Dextran Sedimentation’, and ‘Pre-lysis’.

Bone marrow samples of patients with myeloma were processed using the ‘Pooled-tube’ (n = 22) and ‘Dextran Sedimentation’ (n = 10) and compared to samples stained in parallel using ‘Pre-lysis’. A fixed amount of fluorescent PKH26 Reference Microbeads was recorded during flow cytometric acquisition and used as an internal counting control such that any reduction in cell recovery can be attributed to cell loss during the staining procedure. Regression analyses were performed to correlate the recovery of cells with regards to the (1A) absolute cell count and (1B) percentages of cells. Percentages were calculated as a function of the number of detected events divided by total leukocyte events (defined as CD45+ and/or CD38+) x 100%. Dotted line represents equal performance. Black line represents the simple regression correlating the techniques. R² > 0.9: strong; R² > 0.8 – 0.9: moderate; and R² < 0.8: weak agreement.

Figure 1.. Correlation of the recovery of various bone marrow subpopulations after samples were stained using ‘Pooled-tube’ and ‘Dextran Sedimentation’, and ‘Pre-lysis’.

Bone marrow samples of patients with myeloma were processed using the ‘Pooled-tube’ (n = 22) and ‘Dextran Sedimentation’ (n = 10) and compared to samples stained in parallel using ‘Pre-lysis’. A fixed amount of fluorescent PKH26 Reference Microbeads was recorded during flow cytometric acquisition and used as an internal counting control such that any reduction in cell recovery can be attributed to cell loss during the staining procedure. Regression analyses were performed to correlate the recovery of cells with regards to the (1A) absolute cell count and (1B) percentages of cells. Percentages were calculated as a function of the number of detected events divided by total leukocyte events (defined as CD45+ and/or CD38+) x 100%. Dotted line represents equal performance. Black line represents the simple regression correlating the techniques. R² > 0.9: strong; R² > 0.8 – 0.9: moderate; and R² < 0.8: weak agreement.

Figure 2.. Antigen expression patterns and scatter characteristics of cellular events after treatment using different staining procedures.

Bone marrow aspirate from a patient with PC dyscrasia was processed in parallel using (A) ‘Pooled-tube’, (B) ‘Pre-lysis’, and (C) ‘Dextran Sedimentation’. Bivariate plots displaying representative combinations of all markers tested and light scatter characteristics are shown. All PCs were first identified based on their CD38+, CD138br, and CD45 expression. These PCs were further stratified into normal (blue dots) and abnormal (red dots) subpopulations based on their unique expression profiles. Even though the calculated Stain Indices indicated significance difference for some markers between assays (see Table 5), it did not compromise the ability of each assay to discern and distinguish between normal and abnormal PCs.

Figure 3.. Impact of ‘Pre-lysis’ and ‘Pooled-tube’ staining procedures on the ability of different CD56 clones of CD56 PE reagents to detect and resolve NK cells.

Bone marrow samples submitted for lymphoma staging were processed using the ‘Pooled-tube’ (gray bars) and ‘Pre-lysis’ (white bars) methods as described above. The samples were stained using a panel consisting of CD45 PCPCy5.5, CD19 PECy7, and CD56 PE. The ability of CD56 PE reagent clones HCD56, 5.1H11, MEM-188, QA17A16 (all from BioLegend) and B159, MY31, R19–760, NCAM16.2 (all from BD), C5.9 (Cytognos), and N901/NKH-1 (Beckman Coulter) to resolve NK cells were tested. Stain indices were calculated by subtracting the median fluorescence intensity of a negative population (B cells; characterized as CD19+, CD56-, CD45br) from a positive population (NK cells; characterized as CD19-, CD56+, CD45br, SSClo), and dividing that number by 2 times the robust standard deviation of the negative population. The experiment was repeated 3 times. Error bars indicate standard deviation. *: p < 0.05; ***: p < 0.001

Figure 4.. Empirical determination of the ‘Pooled-tube’, ‘Pre-lysis’, and ‘Dextran Sedimentation’ sensitivities.

To determine LLOQ, a bone marrow aspirate from a patient with myeloma was surface-stained with the antibodies described in Tube 1 of Table 3. This fully-stained sample was serially diluted 10-fold in replicate with cells that were stained with all surface antibodies except CD38 and CD138. The LLOQ was defined as the highest dilution to have a CV of <30%. To determine LOB, a similar strategy was used (e.g. patient sample was stained with all surface antibodies except CD38 and CD138). A targeted 5 million events were acquired for all tubes and the sensitivity of the assay was calculated by dividing the number of events falling into the region used to define PCs by total leukocytes (calculated as CD45+ and/or CD38+ events). Error bars represent the standard deviation of the repeated measurements. Solid line: actual sensitivity of the assay; dotted line: theoretical sensitivity; dashed line: LOB.