Background

Toxocara canis and T. cati are parasitic roundworms that are causative agents of toxocariasis, a widespread neglected zoonotic disease [1,2,3] that afflicts millions of people worldwide [4]. The disease is transmitted to humans through ingestion of soil contaminated with T. canis eggs from faeces of dogs and T. cati from faeces of cats [5]. Determining the extent of environmental contamination with Toxocara spp. eggs, among which the relative importance of different definitive hosts as sources of ova, is considered one of the knowledge gaps in the epidemiology of Toxocara [2].

Despite the development of sensitive quantitative real-time polymerase chain reaction (qPCR) assays for the detection of T. canis and T. cati eggs [6], low numbers of Toxocara eggs in environmental samples still constitute a shortcoming in the control of toxocariasis [7]. This has resulted in low sensitivity of DNA extraction methods available for Toxocara eggs in soil samples. Although several protocols for extracting helminth DNA from environmental samples have been developed [8], they have not displayed high sensitivity for Toxocara eggs in soil samples [9]. An experimental method combining Toxocara egg enrichment by the flotation technique and subsequent DNA extraction from soil samples spiked with Toxocara eggs resulted in a Toxocara detection rate of 41.7% in 10 g of soil samples spiked with 10 eggs, and only 8.3% for samples spiked with one egg [9]. These results underscore the need for alternative methods which provide sensitive detection of Toxocara eggs in soil samples. The aim of this study was to optimize and validate DNA extraction of T. canis eggs from soil and sand samples. This entailed the assessment of (i) efficient Toxocara egg disruption, (ii) efficient DNA extraction and (iii) removal of PCR inhibitors which might be present in DNA solutions. The optimisation of these analytical steps was expected to improve PCR-based detection of Toxocara in environmental samples.

Methods

Rationale of the study

The rationale supporting this study was a stepwise selection of an optimised analytical protocol exhibiting the highest qPCR positivity rate for the presence of Toxocara eggs in sand and soil samples spiked with T. canis eggs. A series of methods and commercially available extraction and DNA purification kits were used, assessed and compared following a three-step procedure as detailed herein (Additional file 1: Figure S1). The ultimate goal was to use the best protocol on collected (n = 40) environmental soil samples where the potential detrimental effect of qPCR inhibitors on the sensitivity of Toxocara spp. egg detection is well known, and to compare it to the standard method, which is the conventional flotation-microscopic observation.

Soil sample collection

Egg-spiking experiments were carried out in presumably Toxocara-free clean sand and soil samples—commercial sand and soil from a backyard in Poznań, Poland (52° 24′ 52″ N, 16° 55′ 16″ E) without any history of dog, cat or fox presence and doubly confirmed negative for Toxocara spp. eggs with the flotation method described below. Environmental soil samples of 250 g (n = 40) were collected according to a systematic unaligned sampling method, from the upper soil layer (3 cm) [10] in and around Tuliszków, Poland, (52° 04′ 35″ N, 18° 17′ 37″ E), also covering nearby rural villages. The sites included playgrounds (n = 20) and backyards close to households (n = 20). After collection, the samples were dried for 24–48 h and sifted through a 2 mm sieve to remove stones and larger organic particles.

Egg stock solution and serial dilutions

Toxocara canis fertilized eggs were isolated from uteri of adult female worms and suspended in nuclease-free molecular biology-grade water (HyClone™ HyPure). Four aliquots of 5 µl of the unembryonated egg suspensions were observed under the light microscope (100-fold magnification). The number of eggs was calculated per field and then reported as the grand mean of the eight squares ± standard deviation (SD). Serial dilutions (i.e., 104, 103, 102, 10, 1) eggs were prepared in DNA/RNA-free water.

Eggshell disruption methods

Six methods for disruption of T. canis eggs (1, 10, 102 and 103 egg suspensions) were compared: (i) enzymatic lysis with proteinase K (PK) [incubation of egg solution with 0.2 unit of proteinase K in 40 µl solution containing 10% (w/v) of SDS at 56 °C under agitation at 800 rpm for 2 h using a thermomixer (Eppendorf, Hamburg, Germany)]; (ii) thermal disruption (TD) (5 freeze–thaw cycles: 3 min of freezing in liquid nitrogen, followed by 3 min of thawing in boiling water under agitation at 800 rpm in a thermomixer); (iii) mechanical disruption of eggs using FastPrep® tubes containing the lysing matrix A beads (FPA) (MP Biomedicals, Santa Ana, CA, USA) under shaking at 6 m/s for 40 s in a FastPrep-24 homogenizer (three cycles); (iv) the same protocol as the previous but using lysing matrix D beads (FPD) instead; (v) TD followed by FPD (TD-FPD); and (vi) TD-FPD followed by PK (TD–FPD-PK). Following the disruption step, DNA was extracted using the NucliSENS® MiniMag® Kit (bioMérieux, Boxtel, Netherlands) according to the manufacturer’s protocol. DNA solutions were stored at −20 °C until use.

DNA extraction from sand and soil samples using commercial kits

Toxocara-free soil and sand samples (10 g) were spiked in triplicates with serial tenfold dilutions (104 to 1) of T. canis eggs. For the selection of the more efficient DNA extraction method in sand and soil samples, two kits representing the best disruption methods (i.e., DNeasy® PowerMax® Soil Kit, Qiagen, Hilden, Germany, and the FastDNA™ SPIN Kit for Soil, MP Biomedicals, Santa Ana, CA, USA) were used according to the manufacturer’s instructions. The kit producing the best result was then used for DNA extraction from soil samples (n = 40) collected in a rural area around Tuliszków in Poland for detection of T. canis/T. cati eggs using our optimised T. canis- and T. cati-specific duplex qPCR [6].

DNA purification (clean-up step)

The impact of a clean-up step for removal of PCR inhibitors was assessed by performing duplex qPCR for T. canis and T. cati on DNA before and after a clean-up step. A magnetic beads DNA method (Agencourt® AMPure®, Beckman Coulter, MA, USA) was used for DNA clean-up (i.e., for removal of PCR inhibitors) in all samples. Briefly, 1.8 volume of AMPure® beads was added to one volume of extracted DNA. The DNA-beads complex was placed on a magnetic stand and the solution was discarded. The DNA was then washed 2× with 70% ethanol. DNA was eluted in one volume of DNA/RNA-free water. All purified DNA solutions were kept at −20 °C until use. DNA quantity was measured by nanodrop but the concentration values were under the dynamic range of the instrument; consequently Toxocara-specific qPCR—as described below—was used to quantify Toxocara DNA from each DNA extracted sample.

Quantitative real-time PCR

Specific quantitative real-time polymerase chain reaction (qPCR) targeting T. canis (for samples spiked with eggs) or T. canis and T. cati (for environmental samples) was carried out in triplicate on extracted undiluted and diluted (1:10) DNA samples according to the procedure previously described by Durant et al. [6] on a CFX96 thermocycler (Bio-Rad, Hercules, CA, USA). An Ascaridoidea-generic qPCR was used as internal quality control. The qPCR results were expressed as quantification cycle (Cq). Cq values were flagged as “undetermined” by the thermocycler software when reaching 40 cycles, and this threshold was defined as a negative result.

Flotation method and microscopic observation

Environmental samples were processed using the flotation-centrifugation method [11] using the Sheather’s sugar solution specific gravity of 1.27. Two coverslips were used to recover eggs from each sample, examined under light microscopy for counting of T. canis eggs. Recovered eggs were characterized based on their size, the thickness of eggshells, transparency and visibility of semi-circular cavities on their surfaces.

Statistical analysis

For each of the egg disruption methods tested, extraction yield was calculated on the basis of the qPCR Cq values. When only one PCR reaction from the three replicates led to a missing value, this missing value was excluded from downstream analysis, as described and recommended in a recent survey on qPCR data analysis [12]. A calibration curve was built for each method and the limit of detection (LoD) [13] was calculated at the intersection between the limit of blank (LoB) and the prediction interval of the calibration curve. In the current study, the LoD is therefore the lowest number of eggs likely to be reliably distinguished from the LoB and at which detection is feasible. To determine the concordance between results obtained with tested extraction methods, Cohen’s kappa statistic was calculated in field samples. All statistical analyses were performed using R 3.4 and SPSS® Statistics software (IBM Corporation, Armonk, NY, USA).

Results

Dilution series of egg disruption methods

The qPCR Cq values obtained with each Toxocara egg disruption method are shown in Additional file 2: Figure S2. TD, FPD and TD–FPD-PK (LoD = 7 eggs) were the three disruption methods displaying the lowest LoD with the best yield. FPD was selected over TD–FPD-PK as it is handy and easy to use in the field. Accordingly, DNA extraction kits which include a mechanical egg disruption step were selected for assessment of the presence of T. canis in sand and soil samples spiked by serial dilutions of T. canis eggs.

DNA extraction from sand and soil samples using commercial kits

qPCR Cq values from sand and soil samples spiked with serial tenfold dilutions of T. canis eggs (from and 104 to 1) and processed with the DNeasy® PowerMax® Soil Kit and with the FastDNA™ SPIN Kit for Soil are shown in Table 1. The LoD was 4 and 46 T. canis eggs in 10 g of sand and soil samples, respectively, with a probability of 95% for both LoDs (Fig. 1). Typical amplification curves are shown in Additional file 3: Figure S3.

Table 1 DNA extraction efficiency on tenfold serial dilutions of Toxocara canis eggs (104 to 1) spiked in 10-g sand and 10 g-soil samples, by comparing two extraction kits, with or without an additional DNA purification and/or DNA dilution (1:10) step
Full size table
Fig. 1
figure 1

DNA extraction efficiency on serial dilutions of Toxocara canis eggs spiked in 10-g sand and 10-g soil samples. This figure shows standard curves expressing Cq values from sand and soil samples spiked with tenfold serial dilutions (104 to 1) of T. canis eggs. DNA extraction was performed using the DNeasy® Power Max® Soil Kit. DNA from sand samples was used unpurified and undiluted, whereas DNA from soil samples was purified and diluted 1:10

Full size image

Results without clean-up procedure

Results obtained with both DNA extraction kits for soil samples are in line with a higher level of qPCR inhibition in soil than in sand samples, and with higher efficacy of the DNeasy® PowerMax® Soil Kit for removing qPCR inhibitors. However, this removal effect was only visible with sand samples, as no positive signal was obtained with any soil sample, irrespective of the egg concentration. Diluting DNA weakened the positive signals from sand samples while failing to improve results obtained with soil samples, in line with the persistence of inhibitors in the latter samples.

Results comparison after clean-up procedure

The clean-up procedure confirmed its usefulness for removing soil inhibitors in soil samples. The DNeasy® PowerMax® Soil Kit indeed generated a positive signal on soil samples but only at the highest egg concentration. On the contrary, this step had a detrimental effect on strength of positive signals obtained with sand samples. Diluting DNA further decreased the positive signals from sand samples. Combining a clean-up and DNA dilution step on soil samples generated positive signals with both DNA extraction kits, whereas signals obtained with the DNeasy® PowerMax® Soil Kit were higher at low (102 and 103) than at high (104) egg concentration.

Environmental sample analyses

A total of 40 environmental samples collected from backyards and playgrounds were assessed for the presence of Toxocara eggs through the processing of 10 g of soil (from a total of 250 g) using the most effective protocol as identified from the above comparison of methods (i.e., DNeasy® PowerMax® Soil Kit, AMPure® clean-up step and DNA dilution 1:10). In parallel, 40-g soil samples were processed through conventional microscopic examination after an enrichment (flotation) step. qPCR and microscopic observation results are summarized in Additional file 4: Table S1.

Altogether, the duplex-specific qPCR was positive for T. canis in 8/40 samples while no signal relating to the presence of T. cati was generated and the Ascaridoidea-generic qPCR was consistently positive in all samples outlining the presence of Ascaridoidea DNA in all environmental samples examined.

Overall, there was moderate agreement between light microscopic observation of enriched soil samples after flotation, with Cohen's κ = 0.423 (P < 0.005). However, the molecular assay (i.e., DNA extraction using the DNeasy® PowerMax® Soil Kit plus AMPure® clean-up and duplex qPCR) enabled the detection of more positive samples when compared to light microscopy observation on flotation-enriched samples. Additional file 5: Figure S4 gives a global overview of the successive steps and main outcomes.

Discussion

Soil is considered the primary source of Toxocara spp. transmission to humans, especially children [14]. Soil contamination by eggs of Toxocara spp. occurs with faeces from infected dogs and cats. Furthermore, the survival of Toxocara eggs from dogs and cats in the environment for months or years is an additional risk factor for contamination [15]. However, detection of Toxocara eggs in soil samples in the frame of environmental surveillance programs is hampered by the lack of reliable and sensitive analytical methods. While enrichment by flotation and subsequent light microscopic examination remains the reference method, this method is difficult to perform in the field, is time-consuming, and requires experience for light microscopy and parasitological diagnosis.

For maximising the disruption of the thick wall of Toxocara eggs, different DNA extraction methods were compared, as a first step. A bead-beating step was shown to be preferable to enzymatic and thermal lysis. The next step was then to use DNA extraction kits including beads for processing sand and soil samples spiked with serial dilutions of eggs. Results were assessed with and without a clean-up step for removal of PCR inhibitors. DNeasy® PowerMax® Soil Kit gave the best results in terms of Cq values with sand samples, but no soil sample was positive without a clean-up step. The AMPure® beads purification was compared with the PowerClean Pro Cleanup Kit (Qiagen, Hilden, Germany), and displayed similar efficiency (data not shown). The AMPure® clean-up was favoured, as it is compatible with the prospect of automation of the whole analytical processing. This additional step did not improve the qPCR signals obtained with sand samples but appeared to be necessary to generate a positive qPCR signal with soil samples, albeit only at the highest egg concentration. For soil samples, an additional dilution step was necessary. Using the best analytical workflow enabled us to achieve sensitivity of 4 and 46 eggs in 10 g of sand and soil, respectively. To the best of our knowledge, this is an improvement compared to previously published methods [6, 9]. A method combining flotation and qPCR achieved a detection threshold of 10 eggs for the flatworm Echinococcus multilocularis per 10 g of soil but failed to reach a similar figure for Toxocara eggs [9]. Although the causes behind this low sensitivity for detection of Toxocara DNA are not known, they might be associated with the use of methods which do not emphasise the mandatory disruption of eggs prior to DNA extraction. By improving this step, we have been able to achieve sensitivity comparable to that observed with other roundworms. However, current data show that total removal of PCR inhibitors from soil samples cannot be guaranteed, as pointed out by other authors [16]. This contradicts manufacturers’ claims that their extraction kits are efficient in removing PCR inhibitors from soil samples.

A third step was to then analyse a series of environmental soil samples collected from backyards or playgrounds, using the most efficient analytical workflow as previously defined on spiked soil samples. Results were compared with the conventional flotation-microscopic evaluation. Interestingly enough, the latter gave a lower positivity rate. When considering the potential hurdle of non-homogeneous distribution of Toxocara eggs, it is worth noting that samples 7 and 14, which came from the same playground, did not display similar qPCR results. A solution might reside in sampling more areas. As already commented above, PCR inhibitors as the cause of a PCR negative result cannot totally be ruled out. The same observation applies to samples 31, 32 and 40, which came from the same backyard. Of note, the alleged improved sensitivity on detection of helminths such as Echinococcus multilocularis was achieved by combining an enrichment (flotation) step with qPCR [9]. This renders the method cumbersome and time-consuming, two features which preclude the use of the method in the field for screening of hundreds of environmental samples. Another lingering question, not assessed in the current study, is whether positive qPCR results can be taken at value face as proof of soil contamination, given that dead Toxocara eggs can be associated with positive qPCR signals without also being able to cause toxocariasis when ingested by humans. There is no definitive answer to this question, but the long-term survival of Toxocara eggs in the environment [15] and the probability that viable eggs can be mistaken for dead eggs upon light microscopic observation should prompt us to consider qPCR positive as a reliable surrogate marker of soil contamination with Toxocara eggs.

Finally, we assessed the cost of the optimized workflow to process a batch of 10 samples with the optimal DNA extraction kit (~ 64€/sample, details in Additional file 6: Table S2). The optimized DNA-based analysis remains more expensive than the handmade classical method. However, it can already be anticipated that the automated processing of large numbers of samples will make DNA analysis less time-consuming, labour-intensive and expensive than manual sample processing, while also significantly increasing the egg detection positivity rate.

Conclusions

In conclusion, the sample processing developed here, which combines a mechanical disruption of Toxocara eggs, DNA extraction using the DNeasy® PowerMax® Soil Kit and a subsequent DNA clean-up for removal of PCR inhibitors, improves substantially the qPCR detection of Toxocara eggs in soil samples. While the analytical cost per batch of samples is higher than the conventional microscopic examination, current automated analytical procedures have the potential to substantially decrease the cost per sample, and to simplify and speed up the analysis, allowing timely results to be delivered within a short period. Combined with a significantly higher positivity rate, these advantages are worth considering when a large number of analyses need to be performed rapidly in field conditions.