Background
Mature gametocytes need to be present in human blood for transmission of
Plasmodium spp. to anopheline vectors. Reliable prevalence data of gametocyte carriage in the population are needed to know the infectious reservoir and battle the ongoing transmission of malaria. This is especially important in the light of renewed malaria elimination and eradication efforts [
1].
Gametocyte carriage alone does not cause clinical symptoms and gametocytes can remain in circulation for weeks after treatment. Microscopy can fail to detect gametocyte densities below 20 to 50 gametocytes/μl [
2].
In-vivo and
in-vitro feeding assays have shown, that even gametocyte densities below 1/μl can frequently result in mosquito infection. Although such low densities of circulating gametocytes reduce efficiency of malaria transmission [
3], missing submicroscopic gametocytaemia represents a significant gap in mapping the infectious reservoir.
Modern molecular detection techniques of
Plasmodium parasite nucleic acids like quantitative real-time PCR (qPCR) or QT-NASBA offer a limit of detection (LOD) down to 0.02 gametocytes per μl and thus can bridge this gap [
4‐
6]. These amplification methods use gametocyte specific mRNA for quantification. Therefore, transport of fresh blood samples to a suitable laboratory has to be in a cool chain and samples need to be frozen until they are analysed in order to prevent significant and unpredictable loss of mRNA. However, areas heavily affected by malaria often do not offer sufficient infrastructure.
In order to facilitate sample collection and processing, dried blood spots (DBS) on filter paper can be used. This method does not provide results directly in the field [
7], but increases the ability to store and transport large quantities of samples. Both, qPCR and QT-NASBA from DBS samples have been used for detection and quantification of
Plasmodium falciparum gametocytes [
8,
9]. However, no investigation has been performed so far as to what extent storage conditions affect detection. This study was designed to create settings which could realistically be encountered in resource-poor conditions. Assays were standardised with
in-vitro gametocyte cultures spiked in whole blood. QT-NASBA was performed on fresh whole blood (FWB) samples as well as DBS filter paper samples stored under different conditions. The results were compared in detail and recovery and decay rates for all storage conditions were calculated.
In-vivo samples consisted of DBS filter papers as well as thick blood films collected from malaria patients in Jimma, Ethiopia.
Methods
Gametocytes for the dilution series were obtained from a continuous
in-vitro parasite culture of
Plasmodium falciparum (strain NF 54, Amsterdam) synchronised with sorbitol [
10]. The parasite density of the cultures ranged from 2-10% infected erythrocytes. Culture was grown in sterile filtered medium (RPMI 1640 supplemented with 25 mM HEPES, 25 mM sodium bicarbonate, 25 mg/l gentamicin sulphate, 2% D-glucose, and 10% human type AB serum, pH 7.4) with additional erythrocytes to uphold a haematocrit of 5%. New medium to feed the culture was added twice daily. The cultures were kept under a gas phase of 3% O
2, 5% CO
2 and 92% N
2[
11,
12].
Cultures were harvested at a parasitaemia of approximately 10%, pooled until 100 ml of total volume was reached and centrifuged at 1,700 g for 10 minutes. The resulting pellet was resuspended in 10 ml prewarmed pure RPMI medium. To increase the density and to concentrate gametocytes, the diluted pellet was filtered through MACS® Separation Columns (Miltenyi Biotec, Bergisch Gladbach, Germany) as described in more detail by Ribaut and colleagues [
13]. This resulted in 250 μl of enriched solution which was analysed for gametocyte density. Gametocyte counts were independently performed by three trained microscopists using an oil immersion microscope (Axiolab, Carl Zeiss MicroImaging GmbH, Göttingen, Germany). The enumeration was based on three 1 μl Giemsa-stained thick films. The total surface area of each individual spot was measured and 20% thereof was counted. The mean gametocyte density was calculated based on these data.
To investigate the quantification accuracy and the LOD of real-time QT-NASBA, a 10-fold dilution series ranging from 2.5 × 10-3 to 2.5 × 103 gametocytes per μl was prepared in type AB fresh blood. The calculated mean gametocyte density from each dilution was used for further calculations. To microscopically control the dilution process, three 1 μl Giemsa-stained thick films were prepared for densities of 2.5 × 100 to 2.5 × 102 gametocytes per μl. All these thick films were entirely counted and the mean density was determined. Densities below 100 per μl were not considered evaluable by microscopy.
Dried blood spots on filter paper (Standard-Whatman Cellulose Chromatography paper 3MM, GE Healthcare, Fairfield, CT, USA) were prepared from 50 μl of the whole blood dilution series, dried for three hours at room temperature, labelled and individually sealed in plastic bags. Samples were then subjected to different storage conditions (as outlined below).
Storage procedures
The DBS filter papers of each dilution series were analysed after 24 hours, 28 days and 92 days. The processing involved nucleic acid extraction, amplification and quantification with QT-NASBA. To investigate the effect of different storage procedures, the filter papers were subjected to the following conditions:
(I):
Transport simulation (+5-30°C, alternating relative humidity (RH))
(II):
Room temperature (+25°C, 38% RH)
(III):
Warm storage (+37°C, 25% RH)
Right before analysis, the DBS containing 50 μl of blood were cut out of the filter paper with a distance of approximately 3 mm to the blood spot using heat sterilised scissors. The blood spots were immediately soaked in 2 ml NucliSens® easyMAG® (bioMérieux, Lyon, France) lysis buffer containing (guanidiumisothiocyanate, GuSCN) and rocked at 150 rpm for 30 minutes at room temperature. The solution was subsequently centrifuged at 1,500 g for 5 minutes and the filter paper was removed. Fresh whole blood was treated according to the manufacturer’s instructions. After addition of 50 μl of silica particle solution, the original RNA extraction method described by Boom et al. [
14] was performed. RNA was extracted with the NucliSens® - miniMAG™ unit (bioMérieux, Lyon, France) according to the manufacturer´s instructions. Nucleic acids were eluted in 30 μl elution buffer as supplied by the manufacturer.
After the extraction, the specimen was either immediately amplified using NASBA technology or stored at −80°C for up to 24 hours for further analysis. Filter papers spotted with 50 μl of plasmodium-negative full blood were used as negative controls for all steps of analysis.
Real-time quantitative nucleic acid sequence-based amplification (QT-NASBA)
The
Pfs 25-mRNA was chosen as target sequence as it is expressed only in
P. falciparum stage V gametocytes [
15]. Primers
Pfs 25.F (5‘-GACTGTAAATAAACCATGTGGAGA-3) and
Pfs 25.R (5‘-T7-CATTTACCGTTACCACAAGTTA-3‘) and the molecular beacon probe (5‘-TexasRed-CGATCGCCCGTTTCATACGCTTGTAA-CGATCG-DABSYL-3‘) were used at a final concentration of 290 nM and 145 nM, respectively, as described elsewhere [
15‐
17]. The amplification was performed using the NucliSENS EasyQ® Basic Kit (bioMérieux, Lyon, France) according to the manufacturer’s instructions at a KCl concentration of 80 mM. In brief, 5 μl of RNA eluate was incubated at 65°C for 2 min and subsequently at 41°C for 2 min together with 10 μl reaction mixture including primers and beacons. Prior to isothermal amplification at 41°C, an enzyme mixture containing AMV-RT, RNAse H and T7 RNA polymerase was added to a total reaction volume of 20 μl. Real-time amplification was allowed to run for 90 minutes.
Samples were considered positive when the time-point of amplification at which the fluorescence detecting target amplicons (time to positivity = TTP) exceeded the mean fluorescence of three negative controls + 20 SD as described by Schneider
et al[
6].
Data analysis
The raw data from the QT-NASBA reactions were analysed with Mathematica 8.0 (Wolfram Research, Champaign, IL, USA) and Sigma Stat (systat Software GmbH, San José, CA, USA) software. Student’s t-test was used for significance testing of normally distributed data sets. Otherwise, Whitney-Man rank sum test was performed. P-values of <0.05 were considered statistically significant. All P-values are indicated within the figures or the text as appropriate.
Discussion
This report evaluates the stability of gametocyte-specific
Pfs 25-mRNA in dried blood spots on filter paper subjected to different storage conditions. Two methods are widely used to quantify gametocytes in different materials: QT-NASBA and qPCR. There are several advantages of QT-NASBA amplification over qPCR. Primarily, the reverse transcription step to produce complementary DNA (cDNA) is dispensable resulting in faster sample processing and less opportunities to lose RNA or to contaminate samples. As isothermic amplification does not require cycles, amplification time can be reduced [
6,
15,
18]. If QT-NASBA could be performed without extraction, this would further speed up sample preparation [
7]. The primer attachment sites for qPCR and QT-NASBA can be chosen to be at similar positions within the coding sequence. However, as many genes in
Plasmodium do not contain introns, potential cDNA is identical to genomic sequences. Hence, qPCR is totally dependent on complete removal of genomic DNA in contrast to QT-NASBA [
15]. In summary, real-time QT-NASBA seems to be more convenient and faster for gametocyte quantification.
Dried blood on filter paper has already been used effectively for gametocyte detection by amplifying specific mRNA [
9]. In order to implement this approach in large-scale studies, the estimation of the original gametocyte density should preferably be feasible any time after sample collection - at least up to 3 months afterwards.
No significant differences in detectable mRNA levels for different storage procedures within the first 24 hours were observed. Apparently, temperatures in the range of −20 to 37°C seem to be of little importance for short-term storage as long as the samples have been handled and dried appropriately. Surprisingly, in all DBS subjected to different storage conditions for only 24 hours, significantly more Pfs 25-mRNA was detected than in freshly extracted whole blood samples (p = 0.002).
To exclude interference with single-stranded DNA (ssDNA), DNAse- and RNAse- treated samples were amplified. Results indicated no influence of ssDNA. Moreover, it was hypothesised that different makes of filter paper might interfere with the process. So far, the outcome seemed not to be affected by use of four different kinds of filter papers stored for 24 hours at room temperature.
To investigate if fresh whole blood interferes with kit reagents by i.e. inhibiting the extraction or amplification reaction, the whole study was also repeated with gametocyte-spiked whole medium producing similar results. Perhaps the drying process on filter paper might facilitate the lysis of parasites by damaging their walls, resulting in a more efficient lysis than in fresh blood.
By contrast, investigations performed with HIV-1-RNA on DBS demonstrated a significant loss of RNA during the drying process and within the first hours after preparation of filter paper samples compared to fresh whole blood [
19,
20]. Perhaps HIV-1-RNA is more accessible in the beginning or degrades faster during the drying process.
In conclusion, the correction factor for signal increase was found to be stable in the specific study settings and was not influenced by the tested storage procedures. It is however recommended to determine the individual correction factor depending on the workflow used for each study.
After 28 days, significant differences between the storage conditions could be observed. As expected, the freezer (−20°C) prevented most efficiently RNA decay (average recovery rate 98.7%) and 37°C performed worst (average recovery rate 66.7%) (see also Figure
1). Similar results could be obtained after 92 days of storage. At day 28, storage in the fridge showed a slight, but insignificant, trend towards higher decay rates. This trend was only found once in the experiment selected for presentation and not in the other repetitions. It can be assumed that this trend was due to statistical variation. The experiment was neverthless selected as it was the most extensive performed with storage times of up to 92 days (transport simulation, room temperature and freezer).
Exponential decay rates for each storage condition were calculated, allowing estimates of signal loss experienced for each individual sample as long as storage conditions and duration were known. Naturally, all calculations rely on standardised procedures and might vary depending on the individual setting.
The theoretical limit of detection (LOD) was calculated by linear regression with over 42 duplicates of 24 h-old DBS. A density of less than 10 gametocytes per ml could be detected (0.0096/μl, 95% CI 0.0025 to 0.038). Nevertheless, due to methodical non-linearity of the QT-NASBA reaction for target concentrations with absolute extremes, the linearity of regression is only an approximation. LODs close to one gametocyte per DBS are subject to huge dilution effects and stochastic noise. In addition, the LOD will decrease with prolonged storage due to target decay. However, an individual reasonable LOD can be estimated with the data presented above.
QT-NASBA amplification of DBS was performed from Ethiopian patients, who were tested negative for gametocytes by microscopy in 2008 and 2009. Interestingly, 22 (69%) of 32 cases were detected positive for gametocytes nevertheless with QT-NASBA (Additional file
4). These results were astonishing, as the samples had been stored for 2–12 months at room temperature and subsequently 18–30 months at −20°C. Significant decay had to be expected. However, successful RNA extraction has also been described after long-term storage if the initial load was high enough [
21].
Although it is unlikely that a mosquito blood meal of 2–3 μl at gametocyte densities below 0.20-0.30 gametocytes per μl contains at least one male and one female gametocyte required for fertilisation [
2], Schneider and colleagues could prove that some patients tested negative for
Pfs 25-mRNA with QT-NASBA could still infect mosquitoes [
16]. It is yet to be investigated to what extent these very low-level gametocyte densities contribute to the infectious reservoir and at what frequency they result in mosquito infection. Sub-microscopic gametocyte carriage over longer time scales is known to play an important role in sustaining transmission [
22].
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
VS, MP, NBR, MH, SS and AW designed the experiments. MP, VS, DP and NBR performed real-time QT-NASBA and cultured parasites. MP, TE and NBR performed microscopy. TE, NBR and LT provided clinical samples. JS, MP, VS, AW and NBR performed data analysis. MP, VS, DP, NBR, JS and AW wrote the manuscript. All authors read and approved the final manuscript.