Cell Sorting Enables iFISH Detection Of Low BCR-ABL Producing Stem Cells In CML Patients Beyond Deep Molecular Remission

The exact disease state of chronic myeloid leukemia (CML) patients in deep molecular remission is unknown, since even the most sensitive quantitative polymerase chain reaction (qPCR) methods cannot identify the patients who are prone to relapse after treatment withdrawal. Consequently, new strategies are needed to unravel the true disease state in these patients. Through evaluation of FACS sorted stem- and progenitor cells with standard qPCR and interphase fluorescence in situ hybridization (iFISH) assays, we here demonstrate that residual disease can be assessed at the stem cell level in patients with 4-log reduction in BCR-ABL expression (MR⁴).

Fresh bone marrow aspirates were acquired from 17 CML patients treated with tyrosine kinase inhibitors (TKIs) for a period of 12-119 months. Samples were enriched for CD34+ cells, stained with fluorescent monoclonal antibodies and FACS sorted. In the sort setup we included the human myeloid inhibitory C-type lectin-like receptor (hMICL) antigen (van Rhenen et al Blood 2007, Larsen et al Cytometry 2012), which identifies granulocyte-macrophage committed progenitors. Sorted subsets were: Hematopoietic stem cells (HSCs, CD34+CD38-), hMICL+ progenitors (MpP; CD34+CD38+hMICL+), and hMICL- progenitors (MnP; CD34+CD38+hMICL-). Sorted cells were evaluated for their content of Philadelphia positive (Ph+) cells using both iFISH and a sensitive qPCR assay optimized to small cell samples.

Eleven of the 17 studied patients had attained MR⁴ for a median of 30 months (range: 11 – 88), with the remainder having achieved major molecular response (MMR). In the sorted CD34+ subsets (fig 1A) our sensitive qPCR assay only detected BCR-ABL expression in 10/48 (21%) subsets (fig 1B), despite a median input of 85,000 cells (range: 600 – 370,000). Considering these data, we were surprised to find that iFISH detected Ph+ cells in 14/34 (41%) subsets in frequencies ranging from 2% to 14%, and most frequently in the HSC fraction (fig 1C). Taken together, iFISH proved significantly better than qPCR in detecting residual Ph+ cells in the sorted fractions (p=0.038), especially within the HSC fraction (p=0.005) (fig 1D). Moreover, in subsets where both methods were applied, all qPCR positive subsets were also iFISH positive (4/33) while 9/33 samples were qPCR-/iFISH+, including all positive samples from MR⁴ patients (fig 1E).

Figure 1

View largeDownload slide

Detection of residual Ph+ cells in FACS sorted CD34+ cell subpopulations in CML patients in MMR and MR⁴. (A) FACS sorting of HSCs, MpPs, and MnPs. (B) Expression of BCR-ABL mRNA (on the International Scale (IS)) in the sorted CD34+ subsets. Error bars denote the median upper boundary of BCR-ABL expression (± 95% confidence interval). (C) Residual Ph+ cells as detected by iFISH in the sorted CD34+ subsets. (D) Detection of residual Ph+ cells in sorted subsets by iFISH and mRNA-based qPCR. Error bars denote SEM. (E) Combined qPCR and iFISH status of samples analyzed by both modalities.

Figure 1

View largeDownload slide

Detection of residual Ph+ cells in FACS sorted CD34+ cell subpopulations in CML patients in MMR and MR⁴. (A) FACS sorting of HSCs, MpPs, and MnPs. (B) Expression of BCR-ABL mRNA (on the International Scale (IS)) in the sorted CD34+ subsets. Error bars denote the median upper boundary of BCR-ABL expression (± 95% confidence interval). (C) Residual Ph+ cells as detected by iFISH in the sorted CD34+ subsets. (D) Detection of residual Ph+ cells in sorted subsets by iFISH and mRNA-based qPCR. Error bars denote SEM. (E) Combined qPCR and iFISH status of samples analyzed by both modalities.

Close modal

Combining the data from all sorted subsets within the individual patients (fig 2), iFISH detected Ph+ cells in at least one CD34+ fraction in 9/10 patients, while qPCR only found 5/15 positive (p=0.013). Moreover, qPCR positivity was significantly less frequent in MR⁴ patients (1/10) compared to MMR patients (4/5, p=0.017), while iFISH positivity was equally frequent (p=1.0), thus indicating that low BCR-ABL producing Ph+ cells may be even more predominant in the MR⁴ patients.

Figure 2

View largeDownload slide

Detection of residual disease in the CD34- and in at least one of the analyzed CD34+ fractions of the individual patients as assessed by both iFISH and mRNA-based qPCR.

Figure 2

View largeDownload slide

Detection of residual disease in the CD34- and in at least one of the analyzed CD34+ fractions of the individual patients as assessed by both iFISH and mRNA-based qPCR.

Close modal

To our knowledge, we are the first to directly demonstrate the discrepancy between DNA- and mRNA-based methods when assessing persisting Ph+ stem- and progenitor cells in patients with MR⁴. Pre-selection of CD34+ stem- and progenitor cells unmasked residual disease, and enabled detection of Ph+ cells by iFISH in almost all analyzed patients, including 5/6 of patients with MR⁴. Our results show that the residual Ph+ cells in CML patients with deep molecular response are low BCR-ABL producers, and that DNA-based methods are required to define the next level of response beyond the achievement of undetectable BCR-ABL mRNA. This approach demonstrates that residual disease can be assessed at the stem cell level in CML patients with MR⁴, and may provide us with a tool for early and safe identification of candidates for TKI withdrawal.