Molecular biology strategies to detect residual disease

Abstract

The prognostic significance of minimal residual disease (MRD) has been demonstrated for a variety of hematologic malignancies. PCR based assays are among the most important methods for identifying MRD. They are aimed at detecting genetic abnormalities of residual leukemic cells with high specificity and sensitivity and represent an important diagnostic tool to assess the quality of therapeutic response, for clinical risk assessment, and for clinical management. In the present review technical aspects of different MRD detection methods are discussed which depend on the available targets regularly present in the respective leukemia type and subtype. As such fusion transcripts, gene mutations, and clonal rearrangements of antigen-receptor genes may be available for detection. Emphasis is given on discussing benefits and limitations of MRD detection and quantification in CML, AML and ALL.

Keywords:

• Leukemia
• Molecular markers
• Detection
• Quantificatio

When acute leukemia is diagnosed in a patient, the total number of leukemia cells may reach up to 10¹³. In general, patients reach complete remission after about 4 weeks of chemotherapy, a level defined by the sensitivity of traditional techniques of morphology and cytogenetics (approx. 1–5%). Yet, this level still corresponds to about 10¹⁰ malignant cells in the patient and is associated with high relapse rates. Minimal residual disease (MRD) monitoring by more sensitive techniques allows a longer follow-up of the tumor burden during chemotherapy and thus, improves assessment of treatment response, enables the identification of particularly patients at high risk of relapse, and allows choosing finally an MRD driven risk oriented therapy. Flow cytometric immunophenotyping may attain a sensitivity of aberrant cell detection of 10⁻⁴, while sensitivity of molecular techniques may reach further down to 10⁻⁵. Minimum residual disease assays have to distinguish between leukemic and normal cells and rely on the detection of leukemia cell specific markers which are basically of two sorts: disease, patient, or cell-specific immunophenotypes detected by flow cytometric analysis or disease, patient, or cell-specific genetic markers detected by PCR-based methods.1,2

In the latter case, the analysis can be directed to three types of genetic targets: recurrent leukemia-related chromosome aberrations, leukemia-associated point mutations, and if none of the aforementioned lesions have been detected at diagnosis, antigen-receptor gene rearrangements may be used if present. The assays can be designed as qualitative (classical end-point PCR) or quantitative methods (quantitative real-time PCR or RQ-PCR). The kinetics of therapy induced tumor load reduction as well as the increase of the tumor cell pool at the time of relapse can be readily monitored by RQ-PCR analysis of bone marrow samples. In several studies, paired bone marrow and peripheral blood samples have been tested for MRD quantification by molecular methods yielding generally a rather moderate correlation, but the less accurate and less sensitive determination of MRD in peripheral blood may be outweighed by the benefit of a less traumatic sampling for the regular analysis of MRD.2

MRD Detection in CML

When setting up a molecular test for MRD determination, several questions have to be addressed: which are the targets to be looked for? Is quantification useful or even required? What are the desired sensitivity levels? In the case of chronic myeloid leukemia (CML), these questions are easy to answer and have already led to the development of an international standard scale for reporting and further standardization efforts. The target to be chosen is of course the BCR/ABL1 fusion transcript derived from the Philadelphia chromosome or cryptic t(9;22) translocations, present in almost 100% of the CML cases. Fusion transcripts are a preferred target as it is a rather simple task to achieve the high sensitivities (below 10^–4) desired for a meaningful follow-up and monitoring of the patients under or after treatment. Quantification of MRD is useful as it is an outstanding prognostic factor for relapse with or without allogeneic hematopoietic stem cell transplantation. Furthermore, quantification of MRD allows following up very closely patients under treatment with kinase inhibitors such as imatinib or dasatinib in order to determine their individual response kinetics which is of prognostic value for long-lasting treatment response. An optimal goal is defined by the European Leukemia Net as attaining a major molecular response after 18 months of therapy, i.e. a 10³-fold reduction relative to the standardized baseline (defined as 100% of the international standard BCR-ABL1^IS).3,4

MRD Detection in AML

For MRD monitoring in patients with acute myeloid leukemia (AML), there are three types of disease specific targets available: translocation-derived fusion genes (preferred) like PML/RARA [t(15;17)(q24;q21)], RUNX1/RUNX1T1 [t(8;21)(q22;q22)], CBFB/MYH11 [inv(16)(p13·1q22) or t(16;16) (p13·1;q22)], and MLL/AF9 [t(9;11)(p22;q23)], overexpressed genes like WT1, as well as point mutations of the FLT3, NPM1, and CEBPA genes. Which one may be used depends on its detection at diagnosis at an individual level. In general, 30–35% of the AML cases are positive for one of the mentioned recurrent translocations. A further 30% are positive for mutations at either FLT3 (internal tandem duplication, FLT3-ITD) or NPM1, leaving about 40% of the AML patients without a suitable marker for MRD detection. The recent description of WT1 overexpression as a nearly panleukemic marker closes this gap further: nearly 50% of the AML patients exhibit WT1 expression levels at diagnosis sufficiently increased as to discriminate a ⩾2-log reduction during therapy. Although modest in its sensitivity, the early reduction kinetics of WT1 expression has been shown to be nevertheless highly predictive of relapse.5

The time point during the treatment course, when MRD should be determined for optimal risk assessment, is another critical question which has been addressed by several studies. As optimal time point, it is now generally recommended to determine MRD after consolidation therapy in case of acute promyelocytic leukemia (APL) and about 7–10 days after induction therapy in other non-APL cases.6

MRD Detection in ALL

In case of acute lymphoblastic leukemia (ALL), MRD methods should fulfill at least two requirements: first of all, they should reach sensitivity levels of at least 10^–3, preferably 10^–4. Furthermore, they should be applicable to the majority of patients. Currently, three methods fulfill the above requirements: flow cytometric immunophenotyping, PCR-based detection of translocation-derived fusion RNAs, and in its absence PCR-based detection of junctional regions of rearranged immunoglobulin and T-cell receptor (TCR) genes. Fusion gene transcripts may be detected in 30–40% of pediatric and up to 45% of adult precursor-B-ALL cases, particularly ETV6/RUNX1 [t(12;21)(p13;q22)], but also BCR/ABL1, MLL/AF4 [t(4;11)(q21;q23)], and E2A/PBX1 [t(1;19)(q23;p13)], however, only in about 10–15% of T-ALL cases. If none of these translocations can be detected at diagnosis, clone-specific immunoglobulin and TCR gene rearrangements may be used in more than 90–95% of the cases, although at the cost of a higher level of technical complexity. The random insertion and deletion of nucleotides at the junction sites of V, (D), and J gene segments convert the junctional regions to unique clone-specific sequences, which are different in each lymphocyte and thus also in ALL. Therefore, junctional regions are useful as tumor-specific targets for MRD-PCR analysis. In precursor-B-ALL, IGH (more than 95%, mainly VH-JH), IGK (about 65%, mainly Kde), IGL (15–20%), TCRB (about 35%), TCRG (about 55%), TCRD (about 40%), and V2-J29 (40–45%) rearrangements may be detected, whereas in T-ALL, the following rearrangements may be detectable: TCRB (over 90%), TCRG (about 95%), TCRD (about 55%), and IGH (20–25%, mainly DH-JH). However, it is considered as a mayor drawback that these clone-specific rearrangements are not disease-specific. Thus, continuing rearrangements may lead to the loss of the molecular target and therefore to potentially false-negative results.7,8

Combining the results of currently available studies, two time points might be needed for identification of both high-risk and low-risk patients: at the end of induction therapy and a second time prior to consolidation therapy. This combined information on MRD levels seems to be significantly better than a single time point measurement in predicting relapse-free survival. There is, however, still considerable variation in the definition of cutoff levels for MRD-based risk stratification and debate about the optimal time point for MRD measurement. Thus, further research has to be awaited to clarify these issues.9