ESR1 and PIK3CA mutational status in serum and plasma from metastatic breast cancer patients: A comparative study

Abstract

PURPOSE:

Plasma and serum cell-free DNA (cfDNA) are useful sources of tumor DNA, but comparative investigations of the tumor mutational status between them are rare.

METHODS:

we performed droplet digital PCR assay for representative hotspot mutations in metastatic breast cancer (MBC) (ESR1 and PIK3CA) in serum and plasma cfDNA concurrently extracted from the blood of 33 estrogen receptor-positive MBC patients.

RESULTS:

ESR1 mutations in plasma cfDNA were found in 7 of the 33 patients; ESR1 mutations in serum cfDNA were detected in only one out of 7 patients with ESR1 mutations in plasma cfDNA. PIK3CA exon 9 and exon 20 mutations in plasma cfDNA were found in 3 and 7 out of the 33 patients, respectively; PIK3CA exon 9 mutations in serum cfDNA were detected in 2 out of 3 patients with PIK3CA exon 9 mutations in plasma cfDNA; PIK3CA exon 20 mutations in serum cfDNA were detected in 2 out of 7 patients with PIK3CA exon 20 mutations in plasma cfDNA.

CONCLUSIONS:

Here we show the higher frequency of ESR1 and PIK3CA mutations in the plasma than in the serum in 33 MBC patients; therefore, serum samples should not be considered the preferred source of cfDNA.

1. Introduction

Analysis of cell-free DNA (cfDNA) circulating in serum and plasma has been shown as a potential alternative procedure for studying tumor tissue DNA in metastatic breast cancer (MBC). The detection of cfDNA is not easy but the introduction of new techniques, such as digital PCR (dPCR), has overcome many challenges. In addition, qualitative and quantitative changes of tumor mutational status in circulating cfDNA have been shown in various therapeutic conditions [1]. Therefore, analysis of the tumor mutational status in cfDNA from MBC patients can be used not only as a monitoring tool of disease progression, but also as a predictor of tumor response to therapy [2].

In alteration frequency in metastatic versus primary tumors in the BOLERO-2 cohort, Hortobagyi et al. demonstrated that PIK3CA mutations had the highest frequency in metastatic and primary breast tumors, and that ESR1 mutations had higher frequency in metastatic breast tumors than in primary breast tumors [3]. Digital PCR (dPCR) assays on plasma cfDNA of two phase III cohorts recently demonstrated that the difference of clinical features between the hotspot mutations in MBC, ESR1 and PIK3CA mutations. The BOLERO-2 cohort recently revealed that progression-free survival (PFS) benefit of everolimus was maintained irrespective of PIK3CA mutations, but it decreased according to the presence of ESR1 mutations [4, 5]. On the other hand, the PALOMA-3 cohort revealed that PIK3CA and ESR1 status did not affect PFS benefit of palbociclib [6, 7]. Plasma is currently recommended as a source of cfDNA, however, serum samples still remain the preferred candidate molecular sources especially in retrospective studies using biobanked samples [8].

To verify whether serum and plasma could show similar tumor mutational results, we assayed the representative hotspot mutations in MBC, ESR1 and PIK3CA [3], in serum and plasma cfDNA concurrently extracted from the same blood of 33 estrogen receptor (ER)-positive MBC patients.

2. Materials and methods

A total of 33 MBC patients, treated at Kumamoto University Hospital between 2007 and 2016, were enrolled in this study. Cases were selected if archival paired plasma and serum were available. Informed consent was obtained from all patients before biopsy or surgery. The Ethics Committee of Kumamoto University Graduate School of Medicine (Kumamoto, Japan) approved the study protocol. The treatment of MBC patients was performed in accordance with the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology [9]. Basically, MBC patients were assessed monthly for clinical response at the Kumamoto University Hospital.

Blood collected in EDTA K2 tubes and serum separating tubes was processed as soon as possible and was centrifuged at 1,467 $\times$ g for 10 min, with plasma and serum stored in freezers until DNA extraction. cfDNAs were isolated from 200 $\mu$ L of plasma and same volume of paired serum using a PureLink ${}^{\mathrm{\circledR }}$ Viral RNA/DNA Mini Kit (ThermoFisher scientific, Waltham, MA, USA) as per manufacturer’s instructions. All DNAs were quantified and qualified using both DNA 1000 kit and RNA 6000 Pico kit, and an Agilent 2100 Bioanalyzer equipped with Expert 2100 software according to the manufacturer’s instructions (Agilent Technologies Inc., Tokyo, Japan). The ddPCR assay was carried out in the same sample twice using the QX200™ Droplet Digital™ PCR System (Bio-Rad Laboratories, Hercules, CA, USA) as described previously [10]. LBx ${}^{\mathrm{\circledR }}$ Probe ESR1 Multi (A082) was used for the detection of ESR1 Y537S/Y537N; D538G, LBx ${}^{\mathrm{\circledR }}$ Probe PIK3CA Screen1 (A087) was used for the detection of PIK3CA E542K/V and E545V/G/A/Q/K, Q546L/R/P/E/K; and LBx ${}^{\mathrm{\circledR }}$ Probe PIK3CA Screen2 (A088) was used for the detection of PIK3CA H1047L/R/Y and G1049R/S (Riken Genesis, Tokyo, Japan). PCR data were quantified using QuantaSoft™ software (Bio-Rad Laboratories) and results are expressed as a percentage of mutant to total (mutant $+$ wild type) for each plasma and serum sample. Our ddPCR method had been optimized by comparative analysis of a dilution series of each synthetic mutant oligonucleotide as reported previously [11, 12]. A mutation was considered positive with more than three ESR1 and PIK3CA mutant droplets, because the assay was able to detect as few as three copies of the mutant allele in wild-type DNA. Immunohistochemical staining for ER alpha, progesterone receptor, human epidermal growth factor receptor 2, and Ki67 was carried out as previously described [13].

The nonparametric Mann-Whitney U test and the chi-square test or Fisher’s exact test were adopted for statistical analysis of the associations of the presence of cfDNA ESR1 and PIK3CA mutations mutations between plasma and paired serum. Differences were considered significant when a $P$ -value $<$ 0.05 was obtained. All statistical analyses were two-sided and were performed using JMP software version 10.0.1 for Windows (SAS Institute Japan, Tokyo, Japan).

Table 1

Patients’ characteristics

	No. of patients (%)
Variables	( $N=$ 33)
Age at the blood draw
Median (range)	58 (31–84)
Histological type
Invasive ductal	30 (90.9)
Invasive lobular	3 (9.1)
Histological grade
1	13 (39.4)
2	9 (27.3)
3	6 (18.2)
Lobular	2 (6.1)
Unknown	3 (1.0)
The percentage of HR median (25%, 75%)
ER $\alpha$	90 (80–95)
PgR	10 (1–80)
HER2
Negative	30 (90.9)
Positive	3 (9.1)
Visceral involvement
No	15 (45.5)
Yes	18 (54.5)
Bone involvement
No	18 (54.5)
Yes	15 (45.5)
The number of metastasis lesion
1	11 (33.3)
2	6 (18.2)
3 $⩽$	16 (48.5)
The treatment line at the blood draw
1st	6 (18.2)
2nd	7 (21.2)
3rd	3 (9.1)
4th	3 (9.1)
5th	1 (3.0)
More than 6th	13 (39.4)
Prior endocrine therapy
Total	29 (87.9)
AI	26 (78.8)
SERM	23 (69.7)
Outcome
Death	14 (42.4)

Abbreviations: HR, hormone receptor; ER $\alpha$ , estrogen receptor alpha; PgR, progesterone receptor; HER2, human epidermal growth factor receptor 2, AI, aromatase inhibitor; SERM, selective estrogen receptor modulator.

3. Results

A total of 33 MBC patients were enrolled in this study. All tumor plasma and serum samples contained sufficient DNA for this study. The demographics and baseline characteristics of MBC patients are shown in Table 1. The median age of the patients at blood biopsy was 58 years (range, 31–84). A total of 29 (87.9%) out of 33 patients had been treated with hormonal therapy, and a total of 26 (78.8%) out of 33 patients had been treated with aromatase inhibitors before the blood draw. The median duration of follow-up was 146 months (range, 15–284 months). Figure 1A shows the concentrations of cfDNA of paired plasma and serum. There were no statistically significant differences between the concentrations of cfDNA in the plasma (median: 117 pg/ $\mu$ L, range: 14–1018) and in the serum (median: 73 pg/ $\mu$ L, range: 14–1375). There was also no statistically significant correlation of total allele concentration in either the plasma or serum with clinicopathological factors (data not shown). Figure 1B shows the peak fragment size of paired plasma and serum cfDNA. The peak fragment size of serum cfDNA was consistently larger than that of paired plasma cfDNA (median difference: 41 bp, range of difference: $-$ 21–121 bp). Figure 1C shows the presence of cfDNA ESR1 and PIK3CA, and their fractions in plasma and serum, respectively. ESR1 mutations were found in 7 (21.2%) of the 33 patients tested for their plasma cfDNA. ESR1 mutations in serum cfDNA were detected only in one out of 7 patients with ESR1 mutations in the plasma cfDNA (kappa 0.208; CI $-$ 0.14–0.56). PIK3CA exon 9 and exon 20 mutations were found in 3 (9.1%) and 7 (21.2%) out of the 33 patients tested for their plasma cfDNA, respectively. PIK3CA exon 9 mutations in serum cfDNA were detected in 2 out of 3 patients with PIK3CA exon 9 mutations in plasma cfDNA (kappa 0.48; CI $-$ 0.12–1.07), and PIK3CA exon 20 mutations in serum cfDNA were detected in 2 out of 7 patients with PIK3CA exon 20 mutations in plasma cfDNA (kappa 0.18; CI $-$ 0.22–057). In one of 2 patients with PIK3CA exon 20 mutations in serum cfDNA, we could not detect PIK3CA exon 20 mutations in paired plasma cfDNA.

Figure 1.

A. Concentrations of cfDNA (pg/ $\mu$ L) in paired plasma (open circle) and serum (cross) in each of the 33 MBC cases. There were no statistically significant differences between the concentrations of cfDNA in plasma (median: 117 pg/ $\mu$ L, range: 14–1018) and serum (median: 73 pg/ $\mu$ L, range: 14–1375). B. Peak fragment sizes (bp) of cfDNA in paired plasma (open circle) and serum (cross) in each of the 33 MBC cases. The peak fragment size of serum cfDNA was consistently larger than that of paired plasma cfDNA (median difference: 41 bp, range of difference: $-$ 21–121 bp). C. The genomic status of cfDNA ESR1 and PIK3CA in plasma (open circle) and serum (cross) from the 33 MBC patients; contingency table for each mutation detected in plasma versus serum (left) and for each mutation fraction in plasma and serum in the 33 MBC cases (right). ESR1, PIK3CA exon 9, and PIK3CA exon 20 from the upper to the bottom are shown. Abbreviations: cfDNA, cell-free DNA; MBC, metastatic breast cancer; bp, base pair; MT, mutation; WT, wild type.

4. Discussion

We investigated the presence of ESR1 and PIK3CA mutations in cfDNA extracted from plasma and serum samples of 33 MBC patients since the different frequency of ESR1 and PIK3CA mutations between plasma and serum have not been fully investigated. Although the concentrations of cfDNA between the plasma and serum were similar, we found differences in the presence of ESR1 and PIK3CA mutations. ESR1 mutations in serum cfDNA were detected only in one out of 7 patients with ESR1 mutations in plasma cfDNA; PIK3CA exon 9 mutations in serum cfDNA were detected in 2 out of 3 patients with PIK3CA exon 9 mutations in plasma cfDNA; and PIK3CA exon 20 mutations in serum cfDNA were detected in 2 out of 7 patients with PIK3CA exon 20 mutations in plasma cfDNA.

This difference may be explained by that serum samples are easily affected by DNase activity and contamination from lysis of white blood cells (WBCs). Barra et al. reported that serum cfDNA was more affected by degradation than EDTA-treated plasma cfDNA, since DNase activity was inhibited by EDTA; a 14.9-fold higher DNase activity was detected in the serum compared with EDTA-treated plasma [14]. Several groups reported that serum cfDNA is contaminated by genomic DNA from white blood cells (WBCs). Lee et al. showed that the extra DNA in serum is generated during the process of clotting due to WBC lysis [15]. Gautschi et al. indicated that serum cfDNA concentration strongly correlated with WBC counts [16]. In addition, some groups suggested that longer DNA fragments originate from WBCs in serum samples than in plasma samples [17, 18]. Indeed, in our study the peak of the fragment size of serum cfDNA was consistently larger than that of paired plasma cfDNA. Thus, serum cfDNA can lose potentially important diagnostic information, such as tumor-derived genetic alterations. However, the present study has some limitations; it is a retrospective, single-institute study with a relatively small patient cohort.

However, it is unlikely that the following two results could be explained only by such reasons. First, according to the Fig. 1C showing the actual MAF of cfDNA, almost half of such cases represented very large MAF ( $>$ 20%) in plasma while not in serum at all. For example, in case #7, the MAF of PIK3CA exon 20 ctDNA was greater than 50% in plasma and negative in serum. Similarly, #8, 17, 22, 27 showed MAFs of ESR1 was greater than 20% in plasma but 0% in serum. In our results, since the detection amount of total ctDNA was small, MAF tended to be high when mutant DNA was present. Second, in patient #12, PIK3CA exon 20 mutation was negative in plasma and positive in serum. This was not consistent with the conclusion. In order to ascertain its clinical significance, it is necessary to analyze how this mutation changes due to treatment and disease progression. Therefore, further studies are required to confirm our results.

5. Conclusions

Our results showing a higher frequency of ESR1 and PIK3CA mutations in the plasma than in the serum in 33 MBC patients suggest that the use of serum as a source of cfDNA may be discouraged.