Critical methodological factors in diagnosing minimal residual disease in hematological malignancies using quantitative PCR

Abstract

Hematological malignancies are a heterogeneous group of cancers with respect to both presentation and prognosis, and many subtypes are nowadays associated with aberrations that make up excellent molecular targets for the quantification of minimal residual disease. The quantitative PCR methodology is outstanding in terms of sensitivity, specificity and reproducibility and thus an excellent choice for minimal residual disease assessment. However, the methodology still has pitfalls that should be carefully considered when the technique is integrated in a clinical setting.

Keywords:

• AML
• CML
• hematology
• MCL
• minimal residual disease
• qPCR

Hematological cancer is a heterogeneous group of disease entities with respect to both presentation and prognosis [1,2], but have been extensively characterized by genetic and epigenetic alterations.

Thus, a sizable part of leukemias and lymphomas are associated with clonal aberrations that drive the transformation of the affected cell population toward malignancy but which also make excellent molecular targets for the quantification of minimal residual disease (MRD). Multicenter studies evaluating the prognostic effect of measuring MRD for treatment stratification in leukemia, lymphoma and multiple myeloma have been performed [3–7]. Among the approaches currently available, the quantitative PCR (qPCR) methodology stands out because of its preclinical validation and sensitivity. Consequently, the use of qPCR for MRD assessment is now being increasingly integrated in the diagnostic routine even outside clinical trials [8,9]. However, there are still outstanding issues, which need to be resolved prior to widespread routine use of the MRD concept.

What are the optimal targets?

The targets for MRD quantification by qPCR can either be genomic DNA or mRNA and comprise different types of aberrations in the neoplastic cell. MRD quantification based on genomic DNA reflects directly the fraction of malignant cells present in the sample, whereas MRD calculated from assays using mRNA reflects the fraction of RNA transcripts. RNA-based assay is often the only choice when targeting chromosomal translocations as the breakpoints normally are widely scattered in the intermediate intron sequences, hence ruling out the possibility of short qPCR amplicons.

Chromosomal translocations and variations on single gene level such as indels or point mutations often influence cell differentiation or proliferation and are often abundantly present in the cancer cells at diagnosis. In addition to chromosomal translocations, lymphoid malignancies are also characterized by clonal immune receptor rearrangements, which while not driving the disease by themselves, are highly valuable as they can be used for the majority of clonal lymphoid neoplasms as MRD target.

A third category of targets is genes that are aberrantly highly expressed in the neoplastic cells, such as the WT1 gene in acute myeloid leukemia (AML) [10] and the SOX11 gene in mantle cell lymphoma (MCL) [11].

Although genetic alterations have been demonstrated for a major part of the hematopoietic neoplasms, knowledge of the heterogeneity underlying these genetic alterations is an ongoing process with the realization that the malignant clone can contain subclones, which can hamper the diagnosis, as well as the detection of a forthcoming relapse. Some of the aberrations are rare and altogether these issues greatly challenge the repertoire and design of MRD assays.

For the ideal MRD assay, its genetic target should be stable (i.e., present at both diagnosis and relapse) to reflect the kinetics of the malignant clone during treatment. Here, fusion transcripts are usually stable because these often are regarded as mutations impairing hematopoietic differentiation and apoptosis and therefore a primary abnormality and an initiating lesion in the leukemogenic development in AML while mutations conferring a proliferative advantage typically occur later and are regarded as a secondary abnormality driving the disease process [12]. This can be exemplified by the FLT3-ITD mutation that is not necessarily conserved from diagnosis to relapse as a considerable number of patients loose or gain the aberration at relapse compared to the diagnostic sample [13], while the frequent NPM1 indel mutation is highly stable throughout the disease course [14] and consequently the latter should be selected if both are present. In B-cell neoplasms, clonal evolution must be considered when using clonal IgH rearrangements as target, as the V gene can be changed during clonal evolution [15]. However, this can to a certain extent be circumvented by careful design of the patient-specific primer to cover the D-N-J junction. These types of DNA-based assays are generally resource demanding as DNA sequencing and individual primer design is needed and the sensitivity is often hampered by background signals. In conclusion, the choice of MRD assay is important and great care should be exerted in introducing such assays.

Sample quality is fundamental

Several international multicenter trials have been performed to pinpoint optimal circumstances in identifying time points, shipment and processing of samples intended for MRD analysis targeting mRNA [16] and protocols for optimal cDNA synthesis and primer/probe choice has been recommended [17–19]. Especially for mRNA-based assays, material from blood and bone marrow samples should be processed preferentially within 24 h from drawing the sample and a substantial number of cells should be obtained to achieve high-quality nonfragmented nucleic acids and the desired sensitivity of the assays. If long transportation time cannot be avoided, a bedside stabilizing agent such as Paxgene can be advantageous in ensuring high mRNA quality [20]. Another important step is to avoid PCR inhibition caused by, for example, too high concentration of nucleic acids or remnants of hemoglobin or heparin in the sample, which, in particular, is an issue for DNA-based assays [21].

For most qPCR assays, reference genes have been carefully selected to act as indicators of sample quality because of their constant expression within the subset of malignancy that is in question. The Europe against Cancer concerted action has recommended a number of genes to be used in leukemia [18] and the Minimal Information for Publication of Quantitative Real-Time PCR experiments (the MIQE guidelines) published in 2009 recommend inclusion of more than one reference gene for normalization purpose unless there is a clear evidence for invariant expression [22].

The sensitivity & specificity of the assays

To accommodate the clinical demand of MRD assessment in clinical treatment protocols, targets with an inherent high sensitivity and specificity are preferred. Thus, a sensitivity level of at least 10⁻⁴ to 10⁻⁵ can be achieved for most of the cancer-specific aberrations such as fusions transcripts, where no background amplification should be present if the assay is carefully designed. However, in IgH-based MRD assays, the time of sampling and the event of B-cell regeneration can influence MRD levels and should be considered [23].

Contrasting this, the WT1 gene, which has been used during the past two decades for this purpose in AML [24,25], and the overexpression of the SOX11 gene described as MRD target in MCL [11] have sensitivities, which are clearly inferior to assays quantifying fusion transcripts. On the other hand, the advantage of such targets is a high coverage of patients using a consensus assay while a downside can be a heterogeneous level of overexpression at diagnosis within the subgroup of neoplasm, implying that not all patients reach a sufficient sensitivity of more than 10⁻⁴.

Finally, with respect to sensitivity of MRD assays, it should be noted, that in the majority of clinical situations, the sensitivity of even the WT1 assay would suffice in answering questions of lack of efficacy or even impending relapse.

What are the critical factors in the MRD assay?

To ensure optimal reaction conditions during qPCR, it is paramount to include sufficient controls. These encompass nontemplate controls ensuring that no cross-contamination have taken place, positive controls including a sample with high MRD content and a sample with low MRD content, as well as a normal sample without the target of interest ensuring that the primer/probe batch is not contaminated. These controls should be included in each run while dilutions curves, unraveling the efficiency of the assay, can be satisfied with a periodical run if the assay is stable and robust.

Certified plasmid materials have recently been recommended for use as calibrators for quantification of MRD in chronic myeloid leukemia [26] and are generally preferred for absolute quantification of consensus targets. If targets are patient-specific, diagnostic material can be used for dilution series. Alternatively, cell lines or diagnostic material can be used for relative quantification or the expression can be normalized to the number of reference gene copies in the sample. It is, naturally, mandatory only to report MRD in the linear area of the assay where the quantifiable range can be determined by serial dilution of positive controls.

Some genomic sequences challenge the efficiency of the assays as the design of assays is greatly hampered by, for example, an extremely high or low guanine-cytosine content. This can to a certain extent be overcome by introducing nucleotide analogs such as locked or unlocked nucleic acid in the primer/probe design.

Handing over the MRD results to the clinicians

A lot of efforts has been put into multicenter studies standardizing assays specifically within the hematological field [17,18,26–28]. In addition to these initiatives, general recommendations defining the MIQE guidelines were published in 2009 [22].

The need for a standardized and comprehensible report format is high when MRD results are handed over from the laboratory scientists to the clinicians because the MRD results often have a direct influence on the treatment strategy of the patients. A report should preferentially include a graphic presentation of the data showing the increase or decrease of MRD during treatment together with absolute or relative MRD values. Under the auspice of European LeukemiaNet, a software customized for this use was developed to facilitate the comprehension of the data [29].

Challenges in MRD quantification: where are we now & where are we going?

MRD has become a key instrument in monitoring treatment response in chronic myeloid leukemia and acute lymphoblastic leukemia. AML protocols are right now assessing the value of MRD for the patients and MCL studies have shown that pre-emptive treatment using MRD assessment is highly valuable [30]. The future prospects of MRD quantification imply highly sensitive assays. An excellent pioneer example is chronic myeloid leukemia where treatment with tyrosine kinase inhibitors right now is discontinued in protocols such as EUROSKI to identify patients where tyrosine kinase inhibitors can be discontinued without the risk of an upcoming relapse. A methodology, with an excellent sensitivity for MRD quantification is required to evaluate very low or nondetectable levels of the BCR-ABL1 fusion transcript. MRD monitoring is a benchmark that places a high responsibility on the quality of MRD measurements.

The future-refined PCR techniques, such as digital PCR, will probably be a step forward enabling more precise quantification without the need for perfectly optimized assays and exclusive of compromising the sensitivity of the assays [31]. Finally, next-generation sequencing is seen as a promising tool for other targets in both leukemia and lymphoma and reports on sensitivities show that this technique is highly competitive and in some cases superior to the present qPCR technology [32,33]. However, the next-generation sequencing is not yet standardized enough within the MRD territory, which is required, when patients, as matters stand, are stratified to specific treatment protocols. Thus, the qPCR technology will persist as an important player until novel joint initiatives are commenced and fulfilled just as the Europe against Cancer and European LeukemiaNet have set examples through the past three decades. Irrespective of the methodology, we are now at a stage, where practically all patients with hematological malignancies are amenable to MRD monitoring. The implementation of these assays should within the next decade be the new gold standard for following our patients in a biologically relevant fashion.

Acknowledgements

The author would like to thank A Aggerholm and P Hokland for careful reading of the manuscript.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.