LETTER TO EDITOR

Using dPCR to detect minimal residual disease in myeloma by identifying FGFR3 up-regulation

Simon McAuliffe1, 2, Ross Brown1, Alberto Catalano1, P Joy Ho1, Najah Nassif2, Narelle Woodland2,Derek Hart3, Claire Weatherburn1, Shihong Yang1, Hayley Suen1, 2, Cheryl Paul1,Douglas Joshua1and John Gibson1

1: Institute of Haematology, Royal Prince Alfred Hospital, Sydney

2: School of Medical and Molecular Biosciences, University of Technology, Sydney

3: Dendritic Cell Biology and Therapeutics Group, ANZAC Research Institute, Concord

As new therapies increase the incidence of stringent complete remission (sCR) andoverallsurvival (OS) of patients withmultiple myeloma (MM), there is a growing demand for a sufficiently sensitive monitoring method to detect low levels of residual disease. Traditional monitoring using bone marrow (BM) morphology and M-protein requires a substantial tumour mass to be present before a positive signal can be detected.Such methods, although important for defining response, are becoming increasingly irrelevant in the context of an emerging need to detect lower levels of minimal residual disease (MRD).

MRD in MM has been detected using allele-specific oligonucleotide polymerase chain reaction (ASO-PCR), multiparametric flow cytometry (MFC) and deep sequencing using IGH-VDJH, IGH-DJH, and IGK assays[1-4]. These studies have shown that amongst the patients who achieved conventional CR, those patients withno detectable tumor cells in the bone marrow when assessed by PCR or MFC, have a significantly longer median progression free survival (PFS) and OS compared with those with detectable tumor cells[1, 2, 4]. Although highly sensitive, these methods are not widely used for routine MM monitoring. ASO-PCR requires the design and construction of individual probes for each MM patient andis not applicable to all patients[4]. MFC is still expensive, laborious, subjective and not universally applicable [1, 2, 5].

The t(4;14)(p16;q32) translocation is a genetic event that occurs in 15% of MM patients[6, 7]. The translocation affects the terminal ends of Ch4p and Ch14q, forming an IGH-MMSET fusion gene on the derivative 4 chromosome[8]. The FGFR3 locus and other downstream loci, originally on chromosome 4, are translocated to the derivative chromosome 14 (der14), and, consequently, FGFR3 expression is up-regulated by the upstream IGH promoters on der14[8]. Thus FGFR3 up-regulation may be used as a surrogate marker of t(4;14), and providesa tumor specific marker for MM unless there is an associated deletion of der14 (-der14), which would prevent the up-regulation of FGFR3[6].

Digital PCR (dPCR) technology potentially offers a greater level of sensitivity and convenience compared to other MM monitoring methods. The increased sensitivity of dPCR is due to the separation of individual DNA copies to determine their fluorescence, instead of measuring the bulk fluorescence of a sample[9]. This allows the detection of mutations on single DNA strands, offering a greater sensitivity in allele detection.Published gene expression data demonstrates that most patients would have either unregulated fibroblast growth factor receptor 3 (FGFR3), cyclin D1, cyclin D3 or c-maf expression[7]. We performed a pilot study which provided evidence that the detection of up-regulatedFGFR3 can be used to detect MRD in both BM and peripheral blood (PB) samples from patients with MM and that all-on-chip style dPCR[10]is an appropriate platform to detect MRD.

RNA was isolated from mononuclear cells in PB (n=44) and BM (n=23) samples from MM patients using the Maxwell 16 Simply Blood RNA kit (Promega, Madison, USA). Patients were previously screened at diagnosis or relapse for expression of FGFR3 antigen in malignant PCs by flow cytometry.Synthesized cDNA was analysed for FGFR3up-regulation by dPCRusing a QuantStudio 3D dPCR chip (Life Technologies) and real time quantitative PCR (RT-qPCR) (RotorGene 6000), with results recorded as an expression ratio of FGFR3(target gene) to ABL(reference gene) (FGFR3:ABL). Positivity of PB and BM was determined by setting a threshold limit of +2SD over the mean of theFGFR3- group.

Determination of the CV of replicate measurements of FGFR3+ expression in patient samples revealed that at most levels, dPCR was less variable than RT-PCR (Table 1). Tenfold dilutions of OPM2 cells (known FGFR3+) in U266 cells (known FGFR3-) demonstrated that FGFR3 detection by both dPCR and RT-PCR assays were linear to 10-4.

Both methods successfully detectedFGFR3:ABL in MM BM samples, but dPCR did so more consistently than RT-qPCR. dPCR detectedFGFR3:ABLpositivity in 15 of 17 knownFGFR3+ MM BM samples. Those undetected include a patient with stable disease (SD) and another with –der14 (Figure 1A). RT-qPCRdetectedFGFR3:ABL in 11 of 17 samples. Of those not detectedby RT-qPCR, there were 3 patients at diagnosis (Figure 1B).

It was pertinent that several PB samples had a positive FGFR3 signal, suggesting that monitoring of PB for a tumour specific marker could be achieved in MM patients. dPCRdetected a positive FGFR3:ABL in more PB samples than RT-qPCR. dPCR detected a positive FGFR3:ABL in 6 patient PB samples, including 2 patients with a CR (Figure 1C), one of whom had a rising FLC level, further indicating a potential relapse. RT-qPCR was similarly able to detect FGFR3:ABL in 4FGFR3+ PB samples, but also falsely identified positive FGFR3:ABL in a known FGFR3- patient sample (Figure 1D). Thus the FGFR3:ABLratio is a sensitive means of detectingthis tumour specific marker in the PB of MM patients, and dPCR does so with greater consistency than RT-qPCR.This pilot study suggests that FGFR3up-regulationcan be used to detect MRD in FGFR3+ MM.

FGFR3 is an appropriate biomarker to monitor MRD in patients who have been previously identified to be appropriate for MRD monitoring. It was evident that dPCR was more suitable, as it was able to detect FGFR3:ABL above a result based threshold more consistently than RT-qPCRand with greater precision. Each MM patient will need to be evaluated at diagnosis for suitability before this method can be applied. The addition of several other potential MM biomarkers including cyclin D1, cyclin D3 and c-maf would allow monitoring of MRD in most MM patients[7]. This pilot study has shown that dPCR possesses sufficient sensitivity and specificityfor the detection of MRDin MM patients in remission, even in PB samples.

References:

1.Martinez-Lopez, J., J.J. Lahuerta, F. Pepin, et al., Prognostic value of deep sequencing method for minimal residual disease detection in multiple myeloma. Blood, 2014. 123(20): p. 3073-9.

2.Paiva, B., M.B. Vidriales, J. Cervero, et al., Multiparameter flow cytometric remission is the most relevant prognostic factor for multiple myeloma patients who undergo autologous stem cell transplantation. Blood, 2008. 112(10): p. 4017-23.

3.Puig, N., M.E. Sarasquete, A. Balanzategui, et al., Critical evaluation of ASO RQ-PCR for minimal residual disease evaluation in multiple myeloma. A comparative analysis with flow cytometry. Leukemia, 2014. 28(2): p. 391-7.

4.Silvennoinen, R., T. Lundan, V. Kairisto, et al., Comparative analysis of minimal residual disease detection by multiparameter flow cytometry and enhanced ASO RQ-PCR in multiple myeloma. Blood Cancer Journal, 2014. 4: p. e250-7.

5.Puig, N., M. Sarasquete, A. Balanzategui, et al., Critical evaluation of ASO RQ-PCR for minimal residual disease evaluation in multiple myeloma. A comparative analysis with flow cytometry. Leukemia, 2014. 28(2): p. 391-8.

6.Prideaux, S.M., C.E. O'Brien, and T.J. Chevassut, The genetic architecture of multiple myeloma. Advanced Hematology, 2014. 2014(5): p. 335-48.

7.Chesi, M. and P.L. Bergsagel, Molecular pathogenesis of multiple myeloma: basic and clinical updates. International Journal of Hematology, 2013. 97(3): p. 313-22.

8.Walker, B., P. Leone, L. Chiecchio, et al., Acompendium of myeloma-associated chromosomal copy number abnormalities and their prognostic value. Blood, 2010. 116(15): p. 56-65.

9.Pohl, G. and M. Shih, Principle and applications of digital PCR. Expert Review: Molecular Diagnostics, 2004. 4(1): p. 41-7.

10.Zhu, Q., L. Qiu, B. Yu, et al., Digital PCR on an integrated self-priming compartmentalization chip. Lab Chip, 2013. 14(6): p. 1176-85.

Figure 1: Detection of FGFR3 up-regulation in the BM and PB samples of MM patients

FGFR3:ABL in the BM by A) dPCR and B) RT-qPCR and PB by C) dPCR and D) RT-qPCR are able to identify MRD in patients with FGFR3+ MM with differing consistencies.D: diagnostic sample, PD: progressive disease, CR: complete response, -der14: loss of derivative chromosome 14.

Table 1: The CV of FGFR3: ABL in patient samples in dPCR and RT-qPCR

CV of FGFR3: ABL in FGFR3+ MM patient BM samples demonstrates that dPCR is consistently less variable than RT-qPCR with a minimal number of replicates.

%BM PC / N / CV (%)
dPCR / RT-qPCR
1 / 3 / 27.98 / 55.65
4 / 3 / 17.44 / 29.08
20 / 4 / 17.08 / 104.88
30 / 3 / 17.63 / 135.00
40 / 3 / 13.65 / 45.53
55 / 4 / 20.24 / 86.60
80 / 3 / 6.96 / 16.04
90 / 3 / 60.25 / 18.53