Letter to the editor

Minimal residual disease (MRD) in multiple myeloma (MM) may be studied by allele-specific oligonucleotide real-time quantitative PCR (ASO-RQPCR), first applied in acute lymphoblastic leukemia (ALL) using patient-specific ASO forward primers derived from complementarity-determining region 3 (CDR3) sequences of immunoglobulin (Ig) genes [1], combined with consensus reverse germline primers and TaqMan probes in the JH region [2]. While applicable in >90 % ALL [3], the consensus primers/probes are only applicable in 42–75 % MM due to somatic hypermutations [4, 5].

Herein, in 13 consecutive MM (Additional file 1: Table S1), clonality and sequence of CDR3 were identified in diagnostic BM of all 13 (100 %) cases by a staged PCR approach, involving sequential PCR of IgH VDJ, IgH DJ, and IgK VJ rearrangements (Additional file 2: Figure S1; Table S2), superior to studies using BIOMED-2 multiplex PCR [4, 5]. Clonality was identified in three (23.1 %) cases by FR3 IgH VDJ PCR, seven (53.8 %) by FR1 IgH VDJ PCR, two (15.4 %) by IgH DJ PCR, and one (7.7 %) by IgK VJ PCR. Similar to previous MM studies [5–7], VH1, VH3, and VH4 families were mostly used (12~19 %, 47~63 %, and 16~20 %, respectively) (Additional file 2: Table S3).

ASO-RQPCR was first performed using ASO forward primers and consensus primers/probes, according to the EuroMRD guidelines (Additional file 2: Figure S2; Table S2) [2, 8, 9]. By standard curves serially diluted from diagnostic BM, ASO-RQPCR was applicable in only three (23.1 %) cases, with a sensitivity of 5 × 10−4–10–4 and a quantitative range (QR) of 10−3–10−4 (Table 1), suggestive of MRD quantification in follow-up samples with one myeloma cell in a background of 104 normal cells [10]. Causes of failures in other cases included absence of amplification (n = 5) or non-specific amplification in normal control (n = 1) during preliminary qualitative PCR, suboptimal QR of 10−2 (n = 3), or shortage of diagnostic BM for standard curve (n = 1).

Table 1 Performance of RQPCR by consensus germline or entirely patient-specific primers/probes
Full size table

To account for a low applicability using consensus primers/probes, a systematic search for sequences mismatches against consensus primers/probes was performed. Sequencing data revealed mismatches in five (50 %) cases against consensus primers, six (50 %) against probes, and two (20 %) against both sites, but not in two IgH cases (M7 and M10) or one IgK case (M9) (Table 2). While a previous MM study also showed a high frequency of mismatches against consensus primers/probes in patients with IgH VDJ rearrangements (herein 80 vs. 75 %), the number of mismatches per case was higher in our study (five vs. three) [11]. Given that mismatches of ≥3 bases would preclude a successful ASO-RQPCR [11], the use of patient-specific primers/probes is inevitable in Chinese MM with profound mismatches than Caucasian patients.

Table 2 Mismatches against consensus germline reverse primers and/or probes
Full size table

Therefore, ASO-RQPCR using entirely patient-specific primers/probes was performed (Additional file 2: Figure S2; Table S4). Except one case (M2) amplified by IgH DJ PCR, where the absence of an N region precluded design of a specific ASO primer, patient-specific primers/probes ASO-RQPCR was successful in eight (61.5 %) cases (Table 1).

Moreover, as diagnostic BM was limited, we attempted to construct standard curves by plasmid DNA cloned with patient-specific CDR3, whereby ASO-RQPCR was applicable in all 12 cases, with a sensitivity of ≤10−4 in all 12 (100 %) cases and a QR ≤5 × 10−4 in ten (83.3 %) cases (Table 1). Collectively, using CDR3-cloned plasmid DNA, more cases achieved a sensitivity of ≤10−4 than using diagnostic BM (12 vs. 5 cases: p = 0.0046; Table 1), and applicability was increased from 61.5 to 92.3 %. Moreover, plasmid DNA provides an unlimited supply for standard curve construction in multiple MRD analysis. Furthermore, the use of plasmid DNA is more accurate for cases in which the percentage of plasma cells was not specified in the diagnostic BM.

Finally, in a patient (Dx53) who achieved serological complete response (CR) after four cycles of VTD induction, by ASO-RQPCR with a sensitivity of 10−5 and QR of 10−4, MRD was present in both BM samples on days 120 and 16 prior to autologous stem cell transplantation (ASCT), but turned negative on day 14 after ASCT, testifying further reduction of tumor load below clinical detection limit by ASCT (Additional file 2: Figure S3).

In summary, a high applicability and sensitivity of MRD detection by ASO-RQPCR can be achieved by clonality detection using stepwise PCR approach, mismatch analysis, and hence entirely patient-specific primers/probes in addition to standard curve construction using CDR3-cloned plasmid DNA. Moreover, the use of ASCT in a patient achieving serological CR after induction was justified by further reduction of tumor load illustrated by ASO-RQPCR. Finally, the frequency and intensity of mismatches against consensus primers/probes between different populations warrants further study in an expanded cohort.