Intraindividual Temporal miRNA Variability in Serum,Plasma,and White Blood Cell Subpopulations

Abstract

Blood microRNAs (miRNAs) are ideal biomarkers, and blood derivatives are often collected in the scope of miRNA research projects. However, knowledge of temporal variations of miRNAs in healthy individuals is lacking. In this study, miRNA variability was measured over a 1-year period in different blood derivatives, collected every 2–3 months from two healthy donors. There is a continuum of intraindividual temporal variability, with particularly stable (coefficient of variation [CV] <20%–30%) and particularly unstable (CV >100%–130%) miRNAs in serum, plasma, and specific white blood cell subpopulations. The temporal intraindividual variability of miRNAs should be taken into consideration in experimental design of biospecimen collections and validation of diagnostic biomarkers.

Introduction

Evaluation of blood microRNAs (miRNAs) involves a rapid and minimally invasive collection technique with analysis performed on increasingly accurate analytical platforms.¹ Extracellular miRNAs found in plasma and serum, thus, make ideal candidates for clinically useful diagnostic and prognostic biomarkers. These small noncoding RNAs found both intracellularly, as well as in the circulatory system, have a broad range of gene regulatory activity. In the blood, they are found in exosomes or microvesicles² with approximately two-thirds of microvesicle-encapsulated plasma miRNAs commonly expressed in peripheral blood mononuclear cells (PBMCs).³

More recently, it has been shown that miRNAs are also transported by high density lipoproteins,⁴ with up to 90% of the circulating miRNAs not in fact encapsulated in vesicles, but rather associated with protein complexes.⁵ Circulating miRNAs are considered to act in the manner of a hormone, exerting distant effects after uptake by cells in another region of the body.⁶ The mechanisms by which miRNAs are generated in biological fluids are not well understood.

Analysis of miRNA expression in a range of disease types with quantitative reverse transcription polymerase chain reaction (RT-qPCR) allows precise quantitative analyses.⁷ Optimal technical approaches, as well as pitfalls to avoid when analyzing miRNA in plasma and serum samples in this context, have been reported.^8,9 Potentially critical in vivo and in vitro preanalytical factors were identified, including diurnal variations, fasting, needle gauge, precentrifugation delays, white blood cell (WBC) count, and storage duration. Another study has shown the impact of the anticoagulant type (EDTA is the anticoagulant of choice in plasma) and RNA extraction method on the sensitivity of miRNA detection.¹⁰

The quality of extracted total-RNA samples can generally be controlled by RNA integrity number (RIN) measurements in cell subpopulations. However, the RIN algorithm is based on ribosomal RNAs, which are neither consistent nor sufficiently concentrated in serum and plasma samples and, thus, cannot be used for quality assessment of serum or plasma RNA, although that of the corresponding cell-derived RNA may be used as a surrogate quality indicator.

While miRNA quality is relevant in the scope of analytical validation of miRNA biomarkers, clinical validation entails knowledge of baseline reference ranges and of the reference change value (RCV), defined as the difference between two consecutive measurements of a given miRNA in an individual that represents a significant change in health status.¹¹ This validation is performed by analysis of intraindividual miRNA temporal variability occurring independent of disease occurrence.

We assessed an extensive miRNA panel for stability with a profile variability analysis performed from blood samples obtained every 2–3 months over 12 months from two healthy donors. Profiles were analyzed in serum and plasma, as well as in WBC subpopulations (granulocytes, monocytes, and T cells, the latter being representative of lymphocytes), using the WaferGen platform.

Methods

Blood collection and processing

Blood samples were collected from two healthy Caucasian donors (62-year-old male and 46-year-old female) who provided informed consent. Donors were nonsmokers with no recent illnesses or treatment for any chronic medical conditions. In total, 30 mL blood per sampling was collected between 9:00 and 11:00 am every 2–3 months over 12 months. Donors were asked to report any unusual food intake, exercise, and any infections/illnesses over the 12-month period.

Blood was processed within 2 hours following collection. Plasma was isolated from blood collected in EDTA Vacutainer tubes (Becton Dickinson) and centrifuged at 2000 g for 20 minutes. Aliquots were stored immediately at −80°C. Serum was isolated from blood clotted in a CAT Vacutainer tube becton dickinson for 30–45 minutes, then centrifuged at 2000 g for 10 minutes. Aliquots were stored immediately at −80°C. Complete blood counts in whole blood were performed with the ABX Micros CRP 200 (Horiba) according to the manufacturer's instructions.

WBC subpopulations were isolated by magnetic separation using an autoMACS^® Pro Separator (Miltenyi Biotec). MicroBeads (CD3 Miltenyi Biotec, CD14 Miltenyi Biotec, CD15 Miltenyi Biotec, and CD45 Miltenyi Biotec) were added to blood samples collected in EDTA Vacutainer tubes and incubated for 15 minutes at 2°C–8°C. For CD45 and CD15 extractions, 1 mL blood was used, and for the CD3 and CD14 extractions, 2 × 1 mL blood was used. Cold autoMACS running buffer was added, and tubes were centrifuged at 445 g for 10 minutes. Pellets were resuspended in autoMACS running buffer placed in a precooled Chill 5 rack and immediately sorted with the Posselwb program.

Subpopulation purity was verified by flow cytometry using monoclonal antibodies CD3-PerCP (Miltenyi Biotec), CD14-PE (becton dickinson), CD15-PECy7 (becton dickinson), and CD45-FITC (becton dickinson) incubated for 30 minutes at 2°C–8°C, centrifuged at 300 g for 5 minutes, resuspended in autoMACS running buffer, and analyzed on a BD FACSVerse. Purified WBC subpopulations were quantified on a CASY Model TTC Cell Counter and Analyzer (Roche Applied Science) according to the manufacturer's instructions.

Hemoglobin measurement

Hemoglobin was quantified using a DetectX Hemoglobin Detection Kit (Arbor Assays) according to the manufacturer's instructions, using the standard supplied. Absorbance was measured with a Synergy MX spectrophotometer with Gen5 software (Biotec).

RNA extraction

Plasma/serum samples (250 μL) were centrifuged at 3000 g for 5 minutes at 4°C. Supernatant (200 μL) was incubated with 750 μL QIAzol (QIAGEN) supplemented with 1 μg MS2 RNA (Roche Life Science) for 5 minutes at room temperature. Samples were mixed with 200 μL chloroform and incubated for 2 minutes then centrifuged at 12,000 g for 15 minutes at 4°C. The upper aqueous phase was collected, 1.5 volumes of ethanol were added, and the total sample was transferred on an RNeasy Mini spin column (QIAGEN) and centrifuged for 30 seconds at room temperature, 13,000 g. Samples were washed with RWT buffer (QIAGEN) and then washed thrice with RPE buffer (QIAGEN). RNA was eluted with 50 μl DNase/RNase-free H₂O by centrifugation for 1 minute at room temperature, 13,000 g, and then stored at −80°C.

For WBC subpopulation samples, duplicate fractions were pooled (monocytes, T cells) and centrifuged for 5 minutes at 300 g. Pellets were suspended in 200 μL phosphate-buffered saline and recentrifuged. The QIAzol Lysis Reagent (700 μL) was added to the pellet and this solution was thoroughly vortexed. Samples were stored at −80°C.

After thawing, Chloroform (140 μL) was added and thoroughly vortexed and 2 minutes incubation was followed by 15 minutes centrifugation at 12,000 g at 4°C. The upper aqueous phase was collected, 1.5 volumes of ethanol was added and transferred on to an RNeasy Mini spin column, and then centrifuged at 8000 g for 15 seconds at room temperature. Samples were washed with RWT buffer and then twice with RPE buffer. RNA was eluted with 30–50 μL DNase/RNase-free H₂O, centrifuged for 1 minute at room temperature, 8000 g, and stored at −80°C. Extracted RNA from plasma/serum and WBC subpopulation samples were dried in a Labconco CentriVap and resuspended in 8 μL H₂O.

RNA concentrations were determined by spectrophotometry (Synergy MX; BioTek Instruments). RNA integrity was analyzed in terms of RIN using RNA Nano Chip (Agilent) evaluated on an Agilent Bioanalyzer 2100 according to the manufacturer's instructions.

Quantitative polymerase chain reaction

0.5 μg of extracted RNA was tailed with poly As using Poly(A) Polymerase Tailing Kit (Epicentre) with RT buffer from the High capacity complementary DNA (cDNA) RT Kit and RNase inhibitor (both Applied Biosystems) for one cycle of 30 minutes at 37°C and then 5 minutes at 70°C, in a Thermal Cycler (C1000; Bio-Rad Laboratories). Samples were stored at 4°C followed by reverse transcription using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) with SmartChip miRNA RT primers (WaferGen), with one cycle of 60 minutes at 40°C and then 5 minutes at 85°C. cDNA samples were stored at −80°C.

For qPCR, the total cDNA sample was used, mixed with 500 μL LightCycler 480 DNA SYBR Green 1 Mix (Roche Applied Science) with 1 mg glycogen (Roche Life Sciences) and 1 μg M13mp18-ssDNA (New England Biolabs) spiked in yeast controls (2.0 μL from Saccharomyces cerevisiae S288c; WaferGen) and made up to 1 mL with PCR grade H₂O (Ambion). Samples were loaded onto a SmartChip Human miRNA Panel v3.0 (WaferGen) with a SmartChip Nanodispenser and read with a SmartChip Cycler using WaferGen Human miRNA v3 Protocol Rev B, or loaded chips were frozen at −20°C until analysis. Quadruplicate analysis was performed for each miRNA, time point, and specimen type.

Data analysis

miRNA qPCR profiles were analyzed with the WaferGen qPCR v2.5 software. See Supplementary Table S1 (Supplementary Data are available online at www.liebertpub.com/bio) for detailed analysis settings. Values were exported to qBase v 2.6.1 software (Biogazelle) and analyzed by specimen type and gender. Global normalization and target scaling to average were used for qBase parameter calculation settings. Calibrated normalized relative quantity (CNRQ) data were exported to Excel (Microsoft) from samples with the following QC criteria: global normalization factor between 0.3 and 3.33, reference target stability with an average GNorm M < 1.0, and coefficient of variation (CV) <0.6. Negative results were replaced by minimal CNRQ value of 0.050 to avoid artificially decreasing CVs. miRNA variability was expressed in terms of CV of the CNRQ, and miRNAs were ordered accordingly.

For each miRNA, the CVs of the CNRQ values for each gender and specimen type were ranked, together with the mean of the corresponding replicate non-normalized Ct values, of all the time points passing the acceptance criteria, and together with the number of these time points. The miRNAs with an accepted replicate number of Ct values of <2 time points per gender and specimen type were excluded. The miRNAs with at least two accepted replicate time points were sorted on CNRQ CV to generate a list of miRNA reliable ranking from the ones with the most stable levels to the ones with the most unstable levels per gender and specimen type.

Results

Characterization of plasma, serum, and WBC subpopulation samples

The two donors were healthy throughout the study duration, with no infectious episodes reported. Both followed a consistent diet and exercise schedule. There was no significant hemolysis in any of the serum or plasma samples: hemoglobin concentration was lower than 1.4 mg/mL in all samples (data not shown). The WBC subpopulation distribution was defined as being normal by HORIBA Medical Normal Ranges (Supplementary Table S2). Total yields of sorted cells were reproducible along the duration of the study: 2.7 × 10⁶ (±13%) CD45⁺, 1.5 × 10⁶ (±21%) CD3⁺, 7.3 × 10⁵ (±13%) CD14⁺, and 2.1 × 10⁶ (±16%) CD15⁺ for the female donor and 3.1 × 10⁶ (±29%) CD45⁺, 1.0 × 10⁶ (±40%) CD3⁺, 7.7 × 10⁵ (±25%) CD14⁺, and 2.5 × 10⁶ (±22%) CD15⁺ for the male donor. The purity of each sorted WBC subpopulation was >90% (data not shown).

All WBC subpopulation fractions had RIN values of 7.0 or higher (Supplementary Table S3). RIN score was not determined in serum and plasma samples.

miRNA temporal variability

miRNA variability was determined on the basis of CVs calculated for six evaluations corresponding to samples collected at six time points (every 2–3 months) over the 1-year period. From two time points (No. 1 and No. 3), the results failed the acceptance criteria and corresponding data were excluded from the analyses. We found that temporal variability of miRNAs in any of the blood derivatives that were examined (serum, plasma, lymphocytes, granulocytes, monocytes, or total leucocytes) corresponds to a continuum, with few miRNAs having extremely stable levels (0.7%–4.6% of the detected miRNAs have CV <10%), few miRNAs having very unstable levels (3.9%–7.8% of the detected miRNAs have 110% < CV <140%), and a substantial proportion of the detected miRNAs having extremely unstable levels (9.0%–19.6% of the miRNAs have CV >140%) (Fig. 1).

FIG. 1.

Number of miRNAs with different temporal CV% in blood derivatives. CV F and CV M correspond to the results obtained from the female and male donor, respectively. neg, absent miRNAs; CV, coefficient of variation; F, female; M, male; miRNAs, microRNAs.

Table 1 summarizes the least variable miRNAs in serum/plasma (CV <30%) and WBC subpopulations (CV <20%) as found in both female and male donors. The miRNAs with the most stable levels found in both serum and plasma were miR-106B*, miR-3652, miR-503, and miR-99A and in WBC were miR-4289, miR-4297, miR-3158, and LET-7A-2*. Supplementary Table S4 includes the miRNAs, which were consistently found to be absent (CV = 0%) in the blood derivatives (serum, plasma, and WBC subpopulations) of both the male and the female donor. The most variable miRNA for both female and male donor in plasma was miR-1260, while the most variable miRNAs in serum were miR-101*, miR-146B-3P, and miR-218-1*.

Table 1.

miRNAs with Stable Levels (CV < 20% for Cells and < 30% for Plasma and Serum) in Both Male and Female Donors, over a 1-Year Period

Plasma	Serum	CD45 ⁺ (leucocytes)		CD3 ⁺ (T cells)	CD14 ⁺ (monocytes)	CD15 ⁺ (granulocytes)
Let-7e^*	Let-7d^*	Let-7a^*	Mir-4267	Let-7a-2^*	Let-7a-2^*	Mir-106a
miR-100	miR-106b^*	let-7b^*	miR-4288	miR-103	let-7a^*	miR-1225-5p
miR-105^*	miR-1231	let-7g^*	miR-4289	miR-106a	let-7b^*	miR-1237
miR-106b^*	miR-1237	miR-1224-3p	miR-4297	miR-1208	miR-1238	miR-126
miR-1228^*	miR-1273c	miR-1225-3p	miR-4300	miR-1279	miR-146a	miR-1307
miR-1288	miR-1285	miR-1225-5p	miR-4304	miR-17	miR-152	miR-154
miR-1469	miR-1294	miR-1238	miR-4319	miR-200c	miR-18b^*	miR-24-2^*
miR-150^*	miR-1306	miR-1279	miR-4325	miR-202	miR-24	miR-3137
miR-1538	miR-142-5p	miR-1307	miR-433	miR-218	miR-28-5p	miR-3158
miR-183	miR-155	miR-148a^*	miR-483-3p	miR-26a	miR-29b-1^*	miR-4289
miR-19b-1^*	miR-222	miR-152	miR-485-3p	miR-26a-2^*	miR-3142	miR-4318
miR-3193	miR-29b	miR-15b^*	miR-490-3p	miR-3158	miR-339-5p	miR-564
miR-320c	miR-302d^*	miR-16-2^*	miR-498	miR-324-3p	miR-345	miR-574-3p
miR-342-3p	miR-31	miR-17	miR-500b	miR-3616-3p	miR-3665	miR-924
miR-342-5p	miR-3180-3p	miR-187	miR-517c	miR-3676	miR-4289	miR-92a
miR-3652	miR-3192	miR-1915^*	miR-518b	miR-425^*	miR-4297	miR-98
miR-3918	miR-3652	miR-195	miR-532-3p	miR-4297	miR-4318
miR-3937	miR-370	miR-196b	miR-542-5p	miR-512-3p	miR-4325
miR-4325	miR-371-5p	miR-24-2^*	miR-548aa	miR-525-3p	miR-486-5p
miR-503	miR-423-3p	miR-26a	miR-550a	miR-574-3p	miR-490-3p
miR-664	miR-4252	miR-28-5p	miR-572	miR-625^*	miR-500a
miR-92b^*	miR-4278	miR-30e	miR-582-3p		miR-548q
miR-939	miR-4286	miR-31	miR-625^*		miR-551b
miR-99a	miR-4297	miR-3158	miR-628-5p		miR-598
	miR-503	miR-331-3p	miR-634		miR-676^*
	miR-543	miR-339-3p	miR-708^*		miR-875-3p
	miR-611	miR-33b^*	miR-7-1^*		miR-99a
	miR-675	miR-342-5p	miR-766
	miR-873	miR-3656	miR-767-3p
	miR-99a	miR-3665	miR-769-5p
		miR-3676	miR-874
		miR-378	miR-924
		miR-425^*	miR-92a

Indicates the opposite arm of the predominant miRNA on the predicted precursor.

CV, coefficient of variation; miRNA, microRNA.

Table 2 summarizes the most variable (CV >130%) miRNAs in serum and plasma, and Table 3 summarizes the most variable (CV >100%) miRNAs in WBC subpopulations as found in both female and male donors. Figure 2 shows representative examples of the CNRQ temporal variation for different miRNAs. Complete lists of the CV% of CNRQ of all miRNAs are available in Supplementary Table S5. The annual variation range of the raw (non-normalized) Ct values of the miRNAs with the most stable levels was generally lower than 2Ct, while the annual variation range of the raw Ct values of the miRNAs with the most unstable levels was often as high as 6Ct (Fig. 3).

FIG. 2.

Relative quantities of four miRNAs in blood derivatives isolated from a female and male donor, at six time points over 1 year. (A) miR-3652, stable in serum (SER) and plasma (PLA) (average CV 17.1%); (B) miR-2160, unstable in serum and plasma (average CV 92.5%); (C) miR-152, stable in WBCs (average CV 18%); (D) miR-3179, unstable in WBCs (average CV 78%). CD3, lymphocytes; CD14, monocytes; CD15, granulocytes; CD45, total leucocytes; WBCs, white blood cells.

FIG. 3.

Examples of raw Ct value variations of stable miRNAs in PLA/SER (A) and WBCs (B) and of unstable miRNAs in PLA/SER (C) and WBCs (D).

Table 2.

Highly Variable (CV >130%) miRNAs in the Female and Male Plasma and Serum Samples, over a 1-Year Period

Plasma (F)	Serum (F)	Plasma (M)	Serum (M)
miR-1260	miR-101	miR-1225-3p	miR-101^*
miR-1274B	miR-101^*	miR-1260	miR-135b
miR-1283	miR-1179	miR-196b	miR-146b-3p
miR-129-5p	miR-1208	miR-19a^*	miR-154
miR-138-1^*	miR-129-5p	miR-302d	miR-200a
miR-192^*	miR-1298	miR-3124	miR-218-1^*
miR-217	miR-1305	miR-3148	miR-30c-2^*
miR-24-2^*	miR-138	miR-3157	miR-3681
miR-302c^*	miR-144	miR-33a	miR-3919
miR-3074	miR-146b-3p	miR-34c-3p	miR-421
miR-30c-2^*	miR-15a	miR-3658	miR-4311
miR-3158	miR-15b^*	miR-3667-5p	miR-4330
miR-3622b-3p	miR-16	miR-3919	miR-525-5p
miR-3662	miR-1913	miR-493^*	miR-548l
miR-3665	miR-192	miR-520f	miR-556-3p
miR-3917	miR-195	miR-526a
miR-4264	miR-218-1^*	miR-548d-5p
miR-4276	miR-25	miR-548f
miR-4330	miR-302c^*	miR-548i
miR-502-3p	miR-31^*	miR-556-3p
miR-548a-5p	miR-3137	miR-890
miR-618	miR-3156
miR-888	miR-338-5p
	miR-33a
	miR-3620
	miR-3622a-5p
	miR-363
	miR-369-5p
	miR-411^*
	miR-449b
	miR-451
	miR-486-3p
	miR-486-5p
	miR-487a
	miR-496
	miR-515-5p
	miR-548a-5p
	miR-548n
	miR-553
	miR-558
	miR-595
	miR-640
	miR-642b
	miR-643
	miR-671-3p
	miR-875-3p
	miR-888

The miRNAs listed were at least in two of the time points analyzed, positive per sex and sample type. In bold are the miRNAs, which are highly variable for both female and male donors.

Indicates the opposite arm of the predominant miRNA on the predicted precursor.

F, female; M, male.

Table 3.

Highly Variable (CV >100%) miRNAs in the Female and Male Total WBCs and Specific WBC Subpopulation Samples, over a 1-Year Period

CD45 ⁺ (F, M)	CD3 ⁺ (F, M)	CD14 ⁺ (F, M)	CD15 ⁺ (F, M)
miR-1206	miR-1182	miR-1179	miR-921
miR-1224-5p	miR-129^*	miR-1206	miR-100^*
miR-1247	miR-19b-2^*	miR-1273c	miR-1179
miR-1258	miR-3074	miR-1278	miR-1256
miR-1295	miR-3160	miR-1298	miR-1262
miR-1301	miR-380	miR-133b	miR-1267
miR-1305	miR-449b^*	miR-139-5p	miR-1278
miR-144^*	miR-515-5p	miR-148b^*	miR-1295
miR-193b	miR-524-3p	miR-153	miR-1297
miR-200b	miR-548p	miR-193b	miR-133b
miR-202^*	miR-548^*	miR-194	miR-146b-3p
miR-299-5p	miR-563	miR-200b	miR-18b
miR-302b^*	miR-614	miR-210	miR-19a^*
miR-3121	miR-656	miR-218-1^*	miR-202^*
miR-3128		miR-218-2^*	miR-216b
miR-3152		miR-24-1^*	miR-224
miR-320d		miR-302a	miR-2277-5p
miR-34a^*		miR-3123	miR-2355-5p
miR-363^*		miR-3125	miR-23b^*
miR-3691		miR-3147	miR-299-5p
miR-384		miR-3157	miR-302b
miR-4285		miR-3193	miR-302b^*
miR-4302		miR-3194	miR-3138
miR-493		miR-326	miR-3144-3p
miR-511		miR-335^*	miR-3151
miR-515-3p		miR-361-3p	miR-3153
miR-516b		miR-367	miR-3166
miR-518c^*		miR-3682	miR-3189
miR-518d-3p		miR-3689b	miR-3198
miR-548c-3p		miR-376b	miR-323-3p
miR-548k		miR-379	miR-3606
miR-550b		miR-3923	miR-3662
miR-578		miR-4290	miR-3692
miR-612		miR-452	miR-3923
miR-623		miR-487b	miR-423-3p
miR-641		miR-509-3p	miR-4265
miR-645		miR-514b-3p	miR-4266
miR-770-5p		miR-515-3p	miR-4307
miR-885-5p		miR-518e^*	miR-4312
miR-9		miR-519d	miR-448
miR-921		miR-520a-5p	miR-486-3p
		miR-520f	miR-513a-3p
		miR-548e	miR-517a
		miR-548k	miR-517b
		miR-550b	miR-520c-3p
		miR-551a	miR-520g
		miR-553	miR-524-3p
		miR-604	miR-562
		miR-616^*	miR-616^*
		miR-620
		miR-624
		miR-655
		miR-671-5p
		miR-711
		miR-885-5p
		miR-9

Indicates the opposite arm of the predominant miRNA on the predicted precursor.

WBC, white blood cell.

Discussion

Although differences between the miRNA profiles of serum and plasma have been shown,¹² many of the miRNAs with the most stable levels had stable levels in both serum and plasma (Table 1), for example, miR-3652, miR-503, and miR99A. Zhao et al. found the miR-346 and miR-134 to be stably detected in plasma extracted from PAXgene tubes, over various in vitro preanalytical parameters, including processing delay, storage conditions, and storage time.¹³ These miRNAs showed high intraindividual temporal variability in our study. Therefore, it appears that specific miRNA targets can be robust to in vitro preanalytical conditions, when specimen collection occurs at one time point and has high in vivo intraindividual variability over time.

The significant differences we identified in the miRNA profiles of lymphocytes, monocytes, and granulocytes in both healthy donors support previous recommendations to use selected WBC subpopulations for miRNA studies,¹⁴ and at a minimum to accurately document blood cell counts in the original blood specimens if using RNA from whole PBMCs. Many miRNAs were consistently absent from all blood cell subpopulations (Supplementary Table S4, miRNAs shown in bold under CD45⁺). Examples of miRNAs whose CNRQs had the lowest CV% are miR-4289 and miR-490-3P in total leucocytes, miR-106A and miR-3676 in lymphocytes, miR-3665 and miR-598 in monocytes, and miR-3137 and miR-924 in granulocytes (Supplementary Table S5).

Although miRNA circadian oscillations have been found in Drosophila¹⁵ and mice,¹⁶ circannual variability in humans has not previously been studied. Our analysis demonstrated high temporal in vivo variability for many miRNA normalized relative quantities in two unrelated donors. Furthermore, this variability was seen in different biospecimen types—serum, EDTA plasma, lymphocytes, monocytes, granulocytes, and total leucocytes. A range of causes may contribute to miRNA temporal instability. Seasonal variations have been reported to affect blood clinical chemistry parameters,^17,18 and may also affect some miRNAs. However, a multiannual study will be necessary to assess seasonal variations.

Among other in vivo and in vitro parameters with a potential impact on miRNAs, age and gender have previously been reported as not influencing miRNA concentrations in either PBMC or plasma microvesicles.³ Some miRNAs have been reported to increase in the event of hemolysis, including miR-15b, miR-16, miR-24,¹⁹ and miR-451.²⁰ Among these, the miR-16 and miR-451 were among the most variable miRNAs in female serum.

The choice of the WaferGen platform for detecting differences between miRNAs was based on published data demonstrating very high sensitivity and reproducibility.²¹ The data analysis pipeline used here combines the power of both the WaferGen and the qBase analysis programs. This allowed us to determine a list of miRNAs sorted according to temporal stability at 2 to 3-month intervals over a period of 1 year, for two donors and different sample types.

As with any validated clinical diagnostic biomarker, knowledge of the analytical imprecision (CV_a), biological intraindividual variation (CV_w), and interindividual variation (CV_b) is required. A diagnostic laboratory must establish the following parameters to assess clinically significant differences from a patient's homeostatic baseline¹: desirable assay imprecision (usually corresponding to half the intraindividual variability; 0.5 CV_w),² index of individuality (ratio of the intraindividual to interindividual variation),³ and RCV.²² The simplest method to calculate RCV, assuming that consecutive measurements are not correlated, is RCV = 2.77√(CV_w² + CV_a²).¹¹ In the case of miRNA RT-qPCR, with very low CV_a, this is RCV = 2.77 CV_w. The higher the CV_w the more difficult it is to prove the clinical relevance of a blood miRNA-based signature and to use it in clinical practice.

Our study showed that roughly 20% of blood miRNAs exhibited random intraindividual variations with CV_w > 100% over time. This highlights the need to establish platform-specific reference ranges and databases of biological variability for all miRNAs, as has been done with other clinical chemistry parameters.²³ While this study was based on 2 donors, establishment of reference ranges in healthy population would require 120 donors, according to international recommendations.²⁴ When collecting blood specimens in the scope of miRNA biomarker discovery, we suggest collecting and storing specific blood derivatives (WBC subpopulations, serum, or plasma) and at multiple time points, to avoid bias from intraindividual temporal variations.

Footnotes

Acknowledgments

The authors are grateful to Sarah MacKenzie, PhD, for scientific writing assistance, to Stefaan Derveaux for advice on statistical analysis of the data, and to the two blood donors.

Author Disclosure Statement

No competing financial interests exist.

References

Cortez

, Calin

. MicroRNA identification in plasma and serum: A new tool to diagnose and monitor diseases. Expert Opin Biol Ther, 2009; 9:703–711.

Valadi

, Ekström

, Bossios

, Sjöstrand

, Lee

, Lötvall

. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol, 2007; 9:654–659.

Hunter

, Ismail

, Zhang

, et al. Detection of microRNA expression in human peripheral blood microvesicles. PLoS One, 2008; 3:e3694.

Vickers

, Palmisano

, Shoucri

, Shamburek

, Remaley

. MicroRNAs are transported in plasma and delivered to recipient cells by high-density lipoproteins. Nat Cell Biol, 2011; 13:423–433.

Arroyo

, Chevillet

, Kroh

, et al. Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma. Proc Natl Acad Sci U S A, 2011; 108:5003–5008.

Cortez

, Bueso-Ramos

, Ferdin

, Lopez-Berestein

, Sood

, Calin

. MicroRNAs in body fluids—The mix of hormones and biomarkers. Nat Rev Clin Oncol, 2011; 8:467–477.

Keller

, Leidinger

, Bauer

, et al. Toward the blood-borne miRNome of human diseases. Nat Methods, 2011; 8:841–843.

Kroh

, Parkin

, Mitchell

, Tewari

. Analysis of circulating microRNA biomarkers in plasma and serum using quantitative reverse transcription-PCR (qRT-PCR). Methods, 2010; 50:298–301.

Moldovan

, Batte

, Trgovcich

, Wisler

, Marsh

, Piper

. Methodological challenges in utilizing miRNAs as circulating biomarkers. J Cell Mol Med, 2014; 18:371–390.

10.

Kim

, Linnstaedt

, Palma

, et al. Plasma components affect accuracy of circulating cancer-related microRNA quantitation. J Mol Diagn, 2012; 14:71–80.

11.

Harris

, Yasaka

. On the calculation of a “reference change” for comparing two consecutive measurements. Clin Chem, 1983; 29:25–30.

12.

Wang

, Yuan

, Cho

J-H

, McClarty

, Baxter

, Galas

. Comparing the microRNA spectrum between serum and plasma. PLoS One, 2012; 7:e41561.

13.

Zhao

, Shen

, Hu

, et al. Effects of preanalytic variables on circulating microRNAs in whole blood. Cancer Epidemiol Biomarkers Prev, 2014; 23:2643–2648.

14.

Petriv

, Hansen

, Humphries

, Kuchenbauer

. Probing the complexity of miRNA expression across hematopoiesis. Cell Cycle, 2011; 10:2–3.

15.

Yang

, Lee

, Padgett

, Edery

. Circadian regulation of a limited set of conserved microRNAs in Drosophila. BMC Genomics, 2008; 9:83.

16.

Cheng

, Papp

, Varlamova

, et al. MicroRNA modulation of circadian-clock period and entrainment. Neuron, 2007; 54:813–829.

17.

Brisson

, Bérubé

, Diorio

, Sinotte

, Pollak

, Mâsse

. Synchronized seasonal variations of mammographic breast density and plasma 25-hydroxyvitamin D. Cancer Epidemiol Biomarkers Prev, 2007; 16:929–933.

18.

Radke

, Izzo

. Seasonal variation in haemodynamics and blood pressure-regulating hormones. J Hum Hypertens, 2010; 24:410–416.

19.

McDonald

, Milosevic

, Reddi

, Grebe

, Algeciras-Schimnich

. Analysis of circulating microRNA: Preanalytical and analytical challenges. Clin Chem, 2011; 57:833–840.

20.

Blondal

, Jensby Nielsen

, Baker

, et al. Assessing sample and miRNA profile quality in serum and plasma or other biofluids. Methods, 2013; 59:S1–S6.

21.

Mestdagh

, Hartmann

, Baeriswyl

, et al. Evaluation of quantitative miRNA expression platforms in the microRNA quality control (miRQC) study. Nat Methods, 2014; 11:809–815.

22.

Fraser

, Harris

. Generation and application of data on biological variation in clinical chemistry. Crit Rev Clin Lab Sci, 1989; 27:409–437.

23.

Sebastián-Gámbaro

, Lirón-Hernández

, Fuentes-Arderiu

. Intra- and inter-individual biological variability data bank. Eur J Clin Chem Clin Biochem, 1997; 35:845–852.

24.

Clinical Laboratory and Standards Institute. How to Define and Determine Reference Intervals in the Clinical Laboratory; Approved Guideline. Second Edition. CLSI document C28-A2. Wayne, PA: CLSI; 2000: 13–22.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.03 MB

0.02 MB

0.00 MB

0.39 MB

0.02 MB