Droplet digital PCR is a sensitive tool for the detection of <i>TP53</i> deletions and point mutations in chronic lymphocytic leukaemia

Abstract

TP53 gene defects encompassing del(17p) and mutations are associated with adverse disease outcomes in chronic lymphocytic leukaemia (CLL). TP53 disruption is a well-established prognostic marker in CLL with clinical and therapeutic implications, therefore its assessment is recommended prior to initiation of the first and every subsequent line of treatment (Pospisilova et al., 2012). Approximately 90% of del(17p) patients also carry TP53 mutations in the remaining allele and mutations in the absence of del(17p) occur in a significant proportion of CLL patients (c. 5% in first-line treatment situation) (Zenz et al., 2010; Pospisilova et al., 2012; Malcikova et al., 2018). Del(17p) assessment is addressed by means of fluorescence in situ hybridization (FISH). FISH is an informative technique, albeit laborious and time-consuming. Hence, there is the need for a faster and more efficient method to quantitate the percentage of the deletion starting from a limited amount of sample. Droplet digital PCR (ddPCR) is a very sensitive and reproducible technique and is mostly used for research purposes (Del Re et al., 2019; Della Starza et al., 2019; Hamfjord et al., 2019; Kjaer et al., 2019). Here we present an approach based on ddPCR aimed at detecting and measuring the deletions and point mutations within the TP53 gene. The most recent guidelines suggest the evaluation of this gene’s abnormalities for intervention and monitoring purposes (Ladetto et al., 2016; Malcikova et al., 2018). Minimal residual disease (MRD) is an excellent prognostic tool in CLL patients and a surrogate for progression-free survival (PFS) (Ladetto et al., 2016). The achievement of MRD response is associated with prolonged PFS and overall survival and is a desirable goal of CLL therapies (Seymour et al., 2018; Kater et al., 2019). Here we propose the use of ddPCR to assess and monitor MRD in CLL patients. All samples were tested by FISH for del(17p) and divided into two groups, one positive and the other negative for the deletion. Thus, we designed ddPCR probes for exons 5, 6 and 7 and assessed a total of n = 47 patients. Copy number variation (CNV) values demonstrate how each probe is able to identify patients with no deletions or positive for del(17p) (Fig 1A and Figure S1). Receiver Operating Characteristic (ROC) curve analysis was performed for exons 5, 6 and 7 (Fig 1C). The three areas under the curve (AUC) were between 0·91 and 0·98, indicating very high sensitivities and specificities of the deletion assessment by ddPCR. The diagnostic thresholds were 1·906, 1·881 and 1·833 for exons 5, 6 and 7 respectively (Fig 1D). Diagnostic thresholds indicate that, for instance, CNV values below 1·8 refer to a patient bearing a deletion of exon 7 in TP53. Patient no. 146 had CNV values of 1·902, −2·020 and −2·020 while patient no. 200 had CNV values of 1·912, −1·870 and −1·780 for exons 5, 6 and 7 respectively. Therefore, these two patients were wild-type by ddPCR. However, further probing for exons 1, 2 and 4 shows how the CNV values were compatible with a deletion for patient no. 200 (1·79, −1·65, −1·805 respectively) but not patient no. 146 (1·92, −1·99, −1·92 respectively; Fig 1B). In order to establish the lowest fraction of DNA with del(17p) or point mutations which can be detected by ddPCR, we titrated a sample with a known percentage of TP53 deletion and a sample bearing a homozygous NC_000017.11: c.517G>C p.V173L mutation. The patient CL bore a chromosome 17 monosomy; therefore the CNV was 1·0 (Fig 1E). Titration of CL within a context of undeleted DNA (sample no. 348) shows the increase in CNV value up to the physiological value of 2·0 (1:1000 dilution) and the minimal quantity of measurable deleted sample is therefore 0·5 ng (1:100 dilution). Patient no. 163 bore a homozygous V173L point mutation, as detected by Sanger sequencing. The titration of no. 163 within a context of non-mutated DNA (sample no. 348) shows that the mutation is detectable in up to a 1:100 dilution, corresponding to 0·5 ng of starting target DNA. At this dilution, the mean number of positive droplets for V173L is 2·44 on a background of 477·3 total negative droplets (corresponding to 5 × 10−5 mutant copies; Fig 1E). Five CLL patients with known point mutations were assessed by ddPCR with specific probes, and their mutant variant allele frequencies (VAFs) were compared to the percentage of TP53 deletions by FISH (Fig 2A). Interestingly, patient no. 163 was homozygous for V173L in exon 5, and ddPCR returned a 95·2% VAF, meaning that all the alleles harbour the mutant sequence. Patient no. 200 showed 11% positive nuclei for del(17p) and the mutant VAF of NC_000017.11: c.797G>A p.G266E was 11·2%, consistent with the fact that deleted nuclei also bear a point mutation on the other allele. No clear correlations were possible for the other three patients. Next we assessed the ability of ddPCR to monitor a patient’s mutation during the course of disease and therapy. Thus, we selected a patient diagnosed in 2013 and re-evaluated him until March 2019. A point mutation located in exon 8 was identified by Sanger sequencing and was missense, leading to the deleterious substitution NC_000017.11: c.817C>T p.R273C (IARC TP53 database; Research, 2019). We assessed the five progressive samples by ddPCR (Fig 2B,C), showing that the point mutation was already detectable in 2015 and the VAF grew until September 2018, when the patient resulted in progression and started ibrutinib in October 2018. Re-evaluation after five months shows how the VAF dropped significantly paralleling the response to ibrutinib therapy. TP53 deletions and point mutations are widely recognized as being among the most relevant drivers of relapsed CLL (Amin et al., 2016; Ladetto et al., 2016). Our evidence suggests that at least six ddPCR probes within TP53 should be used to test del(17p) with high confidence. Concordance between FISH and ddPCR is high for both non-deleted and deleted patients (93·1% and 90·0% respectively). Specificity and sensitivity come out very high with this approach and the simultaneous probing for different exons ensures that a higher percentage of deleted patients can be detected. The time required for the whole ddPCR procedure finalized to the quantitation of TP53 deletions or point mutations is considerably shorter compared to the FISH procedure (Table SI). This is relevant since the time required impacts the costs of the diagnostic procedure. We therefore suggest an algorithm aimed at improving the diagnostic workflow for CLL at diagnosis, including an initial TP53 deletion assessment by ddPCR with six probes. The resulting wild-type patients should be further tested by Sanger sequencing for point mutations. The patients shown by ddPCR to have one or more exons deleted could proceed towards the therapeutic options for del(17p), namely ibrutinib, idelalisib or venetoclax. We also show that point mutations detected by ddPCR may be used to monitor the mutant VAF during follow-up in CLL (Fig 2). Point mutations may serve as effective tools for therapy response and disease monitoring, suggesting its application to the MRD of CLL patients after each cycle of therapy. Figure S1. (A) Representative copy number variation plots for exons no. 5, 6 and 7. Red dots: patients resulted del(17p)-negative by FISH. Blue dots: patients resulted del(17p)-positive by FISH. (B) Fluorescence ddPCR dot plots representative of three patients. Table SI. Time required for the whole procedure of ddPCR on a given set of samples tested for 3–6 exons or point mutations. Table SII. Detailed description of the CNV values obtained by ddPCR for exons no. 5-6-7 for each patient. The data obtained by FISH and sanger sequencing, when performed, are also reported for each patient. Data S1. Materials and Methods. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.