Improvements in Tritrichomonas foetus molecular testing

Abstract

Bovine trichomoniasis is a sexually transmitted disease that results in infertility, abortion, and calf age variability. To date, management strategies include testing for Tritrichomonas foetus and culling of infected males. Challenges associated with testing include cost of culture medium, time and labor burden of sample incubation and processing, and adverse effects of bacterial growth on detection sensitivity. To overcome these challenges, we developed a direct reverse-transcription quantitative real-time PCR (direct RT-qPCR) utilizing smegma, eliminating the use of culture medium. In an analysis of 166 field samples (56 positives and 110 negatives as determined using microscopic reading of cultures as the reference test), the direct RT-qPCR exhibited 100% diagnostic sensitivity and 100% specificity, whereas the currently employed qPCR (culture qPCR), which utilizes cultured samples, exhibited 95% diagnostic sensitivity and 100% specificity. Agreement between direct RT-qPCR and culture qPCR was 98%. Moreover, direct RT-qPCR identified 3 more positive samples and exhibited lower quantification cycle (Cq) values among positives by culture reading than did culture qPCR (direct RT-qPCR Cq range = 14.6–32.3 vs. culture qPCR Cq range = 18.7–37.4). The direct RT-qPCR enables simplified sample collection, elimination of culture medium, faster results, applicability in cows, and lower cost than culture qPCR.

Bovine trichomoniasis, a sexually transmitted disease caused by the protozoan Tritrichomonas foetus, is distributed worldwide, and predominates in areas that utilize natural mating such as Asia, Australia, South America, South Africa, Spain, and the United States.¹⁸ T. foetus infects the preputial mucosa of bulls and the reproductive tract of cows. Bulls of any age may become infected; however, those 3 years and older are more susceptible and become asymptomatic carriers that transmit parasites to cows during coitus.^15,18 Bull-to-cow transmission has been reported to be highly efficient; a single mating service with an infected bull resulted in 95% infections in susceptible nulliparous cows.¹³ There is evidence that cows can carry an infection through post-pregnancy and potentially serve as an infection reservoir for virgin or uninfected bulls.¹⁷

To date, detection is accomplished by 1) culture followed by microscopic reading, 2) conventional end-point PCR, or 3) quantitative PCR (qPCR) using direct preputial samples (smegma) or cultured smegma. PCR has been widely adopted because a single negative test result, as opposed to multiple culture test results, is required to confer negative T. foetus status (USAHA 2014, Resolution 10, General standards for trichomoniasis interstate movement requirements of bulls. Available at: https://goo.gl/j64QMr).

Although the use of PCR has provided significant improvements over culture, such as the ability to detect non-motile organisms (pseudocysts),¹⁴ fast turnaround time, and high specificity, challenges associated with these methods still exist. The requirement for culture medium increases time and labor cost and adversely affects test sensitivity (Se). Unfavorable shipping conditions and varied incubation processes of culture medium may lead to less than optimal growth conditions resulting in lower T. foetus detection Se^5,7 Additionally, the use of culture medium has been shown to promote the growth of smegma-derived bacteria that has been shown to adversely affect Se of T. foetus culture examination and qPCR.⁶ The mean quantification cycle (Cq) values for T. foetus–positive samples without spiked smegma-derived bacteria were 8.1–8.4 cycles lower than those with spiked smegma-derived bacteria.⁶ The authors hypothesized that bacteria in the culture medium can replicate more rapidly than T. foetus, deplete nutrients, and produce metabolic end-products that may result in acidic pH and conditions that adversely affect the growth of T. foetus.⁶ The acidic pH environment may also favor the enhancement of T. foetus DNase activity, which could result in DNA damage, and subsequently lead to lack of detection of T. foetus DNA.⁶

The use of smegma without prior culturing would provide significant benefits to T. foetus detection testing by reducing testing cost, eliminating challenges associated with culture sample shipment and incubation, including the adverse effects of bacterial growth, and reducing turnaround time with minimal disruption of livestock movement. Although the use of smegma for qPCR has been reported to have similar Se with culture samples, the mean Cq value for the qPCR for smegma was 31.98 ± 0.46, significantly higher (p < 0.01) than the mean Cq value of 26.39 ± 0.45 for cultured samples,¹⁰ which suggests lower analytical sensitivity (ASe) for smegma samples. To improve the Se of T. foetus detection in smegma, we developed a detection workflow (hereafter denoted as direct RT-qPCR) that consists of a reverse-transcription quantitative real-time PCR (RT-qPCR) targeting RNA and an automated nucleic acid purification method. The workflow exhibits enhanced Se, overcomes the aforementioned challenges, and has the potential application for detection of T. foetus in cows. Primers and probe sequences for detection of T. foetus were selected based on bioinformatics sequence analysis of the T. foetus sequence (GenBank accession AF339736) containing the 5.8S ribosomal (r)RNA gene (Supplementary Fig. 1). The direct RT-qPCR targets and amplifies the 5.8S rRNA purified from smegma. Specificity of oligonucleotide sequences was evaluated empirically by RT-qPCR using nucleic acid from near-neighbor organisms Trichomonas vaginalis, Tetratrichomonas gallinarum, and Pentatrichomonas hominis, and in vitro transcribed RNA containing Simplicimonas sp. sequences (KY410341). No nonspecific amplification was detected using nucleic acid from the aforementioned organisms. Target-specific RT-qPCR using the nucleic acid from each of these near-neighbor organisms produced positive amplification, indicating the presence of the nucleic acid and providing support for the specificity of the oligonucleotides. Oligonucleotide sequence information is provided in Table 1.

Table 1.

Reverse-transcription quantitative real-time PCR (RT-qPCR) and qPCR primers and probes sequences, amplicon size, and oligonucleotide concentrations.

Target/ Oligonucleotide	Sequence (5’–3’) and reporter dye	Amplicon size (bp)	Final 25-µL reaction concentration (nM)	Gene	Reference
Direct RT-qPCR
Tritrichomonas foetus
TfoeqF2	GAACGTTGCATAATGCGATAAGC	109	500	5.8S rRNA	Current study
TfoeqR1	AACATATATGCGTGTTCTAGCAAGCT		500
Tfoeqpb1	FAM-ATCTTTGAATGCACATTGCGCGCC-BHQ1		125
Exogenous internal positive control
XIPC_F	TTCGGCGTGTTATGCTAACTTC	69	200	Synthetic construct	16
XIPC_R	CCACTGCGCCCAACCTT		200
XIPC_pb	CFO560-CTCCGCAGAAATCCAGGGTCATCG-BHQ1		125
Culture qPCR
Tritrichomonas foetus
Tfoe_F	GCGGCTGGATTAGCTTTCTTT	58	500	5.8S rRNA	12
Tfoe_R	GGCGCGCAATGTGCAT		500
Tfoe_pb	FAM-ACAAGTTCGATCTTTG-MGB		125
Exogenous internal positive control
XIPC_F	TTCGGCGTGTTATGCTAACTTC	69	250	Synthetic construct	16
XIPC_R	CCACTGCGCCCAACCTT		250
XIPC_pb	CFO560-CTCCGCAGAAATCCAGGGTCATCG-BHQ1		60

BHQ1 = black hole quencher 1; CFO560 = CAL Fluor Orange 560; FAM = 6-carboxyfluorescein; MGB = minor groove binder.

The performance of the direct RT-qPCR was evaluated using T. foetus culture and field samples. A live T. foetus culture (strain 66, Biomed Diagnostics, White City, OR) was incubated (InPouch, Biomed Diagnostics) at 37°C; parasites were counted using a hemocytometer, and used for assessment of ASe. Field bulls (n = 166; 56 positives and 110 negatives as determined by microscopic reading of cultures) were sampled for evaluation of assay diagnostic sensitivity (DSe) and diagnostic specificity (DSp); these samples are hereafter denoted as reference samples. These reference samples were identified through diagnostic submissions to the Texas A&M Veterinary Medical Diagnostic Laboratory (TVMDL; College Station, TX) as follows. Owners and veterinarians were contacted after bulls tested positive or negative using the current TVMDL-employed T. foetus culture qPCR (targeting DNA), and bulls were resampled with the owner’s consent. The culture qPCR targets and amplifies T. foetus DNA purified from samples cultured in InPouch. Samples were collected by first washing the prepuce area with at least 10 mL of saline, followed by scraping the inside of the prepuce ~10 times with a sterile artificial insemination pipette. Smegma was collected into a 4-mL tube pre-filled with 1,500 µL of phosphate-buffered saline (PBS). Each sample was mixed to homogeneity using a Pasteur pipette; ~200 μL of each sample was reserved in the original tube and the rest of the sample was deposited into an InPouch. The smegma samples were stored with ice packs, and the pouches were stored without ice packs; all samples were transferred to the laboratory within 8 h of collection. Upon arrival at the laboratory, the smegma samples were refrigerated for use in direct RT-qPCR, and the pouches were incubated at 37°C (Precision incubator, Thermo Fisher Scientific, Waltham, MA) for up to 96 h, at which time the microscopic reading was completed. Four readings, each 24 h apart, were performed; the first reading occurred after 24 h incubation. If no motile T. foetus were observed, the pouch was identified as negative, and additional incubation and 3 readings 24 h apart were required. If motile T. foetus were observed during any of the 4 readings, the pouch was identified as positive and no further incubation and readings were required.

In addition to the 166 field reference samples, an additional 543 field smegma samples from low-risk T. foetus herds that tested negative by culture qPCR were collected to assess the direct RT-qPCR assay specificity. Microscopic reading of cultures was not performed for this sample set based on a low risk of T. foetus infection and cost and labor considerations.

A small set of cow samples was also collected to assess the application of the direct RT-qPCR for T. foetus infection detection in cows. Sixteen cervical mucus, cervical wash, or uterine wash samples were collected from aborting cows or from nonpregnant replacements. For the cow sample collection, the vulvar area was first washed with at least 10 mL of saline. Subsequently, a sterile mare inseminating pipette attached to a 12-mL syringe was used to apply suction (negative pressure) and aspirate cervical or uterine fluid; samples were placed into a 4-mL tube pre-filled with 1,500 µL of PBS.

Culture and field samples were processed (MagMAX 96 viral RNA isolation kit, Life Technologies, Carlsbad, CA; KingFisher-96 automated magnetic particle processor, Thermo Fisher Scientific), and an exogenous internal positive control (XIPC)¹⁶ was spiked into each purification. For smegma or cow cervical or uterine fluid, ~50 μL was transferred (using a 1,000-µL pipette tip given the high viscosity of the sample) to a 96 deep-well sample plate containing 20 μL of Bead Mix (Thermo Scientific, Grand Island, NY). The sample plate was placed on a titer plate shaker (Thermo Scientific) at setting 7 for at least 2–5 min, and 400 μL of lysis binding solution (200 μL lysis binding concentrate, 1 μL carrier RNA, 1 μL XIPC RNA [10,000 copies/μL], and 200 μL 100% isopropanol) was added. For the pouch sample, 270 μL was transferred to a 96 deep-well sample plate containing 20 μL Bead Mix, shaken as above at setting 7 for at least 2–5 min, and 600 μL of lysis binding solution (300 μL lysis binding concentrate, 1 μL carrier RNA, 1 μL XIPC RNA [10,000 copies/μL], and 300 μL 100% isopropanol) was added. Subsequently, the sample plate and reagent plates were loaded onto the magnetic particle processor and processed as described previously.¹⁶

The nucleic acid from the 166 reference pouch samples were used for qPCR. The qPCR consisted of 5 μL of nuclease-free water, 10 μL of Path-ID qPCR master mix buffer (Life Technologies), 1 μL of 20× primers–probe mix (Table 1), and 4 µL of nucleic acid. The 20× primers–probe mix consisted of T. foetus¹² and the XIPC¹⁶ primers and probe oligonucleotides. Thermocycler conditions were 95°C for 10 min (single cycle), and 95°C for 15 s and 55°C for 45 s (40 cycles; Applied Biosystems 7500 Fast real-time PCR instrument, Life Technologies). The Cq threshold settings were set at 5% of the maximum fluorescence value of the amplification signal of the T. foetus and XIPC positive control PCR containing each sequence target at 5,000 copies per reaction. Samples with a T. foetus Cq ≤ 35.0 and XIPC Cq < 40.0 were considered positive; a T. foetus Cq cutoff of ≤35.0 was selected based on previously published data.⁵

The nucleic acid from smegma and cervical or uterine samples were used for RT-qPCR. The RT-qPCR consisted of 6.2 μL of nuclease-free water, 12.5 μL of Path-ID qPCR master mix buffer (Life Technologies), 0.05 μL of ArrayScript reverse transcriptase (200 units/µL, Life Technologies), 1.3 μL of 20× primers–probe mix (Table 1), and 5 µL of nucleic acid. Thermocycler conditions were 48°C for 10 min (single cycle), 95°C for 10 min (single cycle), and 95°C for 15 s and 55°C for 45 s (40 cycles; Applied Biosystems 7500 Fast real-time PCR instrument, standard mode). Samples with a T. foetus Cq ≤ 35.0 and XIPC Cq < 40.0 were considered positive, for alignment with the culture qPCR, which is an implemented routine laboratory test.

Statistical analysis for linear regression, qualitative test performance (DSe and DSp), kappa, McNemar test, and paired t-test were completed (Analyse-It v.2.3, Analyse-It, Leeds, UK). Results were determined to be significant at p < 0.05. Probit analysis was performed (NCSS v.7.0, NCSS, Kaysville, UT) to assess the limit of detection (LOD) of the direct RT-qPCR. Repeatability of the direct RT-qPCR assay was assessed by calculation of the intra- and inter-assay coefficient of variation (CV) of Cq for T. foetus input dilution samples. Intra-assay variability was analyzed in triplicate RT-qPCR for each sample, and inter-assay variability was analyzed in 4 independent experiments (2 operators, 2 runs per operator). CV for each sample = [(Cq SD)/Cq average) × 100%]. Intra-assay CV = [average of CV for all samples per run]. For inter-assay CV calculation, total CV was first calculated using each run’s average Cq for each sample using the formulae: total CV= [(SD of averages of 4 runs/average of averages of 4 runs) × 100%]; inter-assay CV = [average of total CV for each sample].

The dynamic range, efficiency, and relative ASe of the direct RT-qPCR were assessed using spiked serially diluted T. foetus organisms (estimated by cell counting) into smegma and linear regression analysis. Direct RT-qPCR exhibited 97% efficiency (95% CI: 94.2–100.3%) and R² of 0.999, for a 10-fold dilution series that spanned a 5 log dynamic range with 12 replicates for each T. foetus input dilution (Fig. 1). The T. foetus intra-assay CVs were 0.6%, 0.7%, 1.3%, and 0.6% for operator 1 run 1, operator 1 run 2, operator 2 run 1, and operator 2 run 2, respectively. The T. foetus inter-assay CV was 1.0%. The inter-assay average XIPC Cq value for all direct RT-qPCRs performed by both operators was 29.4 ± 0.2, and the CV was 1.1%. The low variations within and between runs for the different amounts of T. foetus organism inputs demonstrate robust performance. Relative ASe was assessed by Probit analysis. Relative ASe in this context refers to the minimum amount of T. foetus nucleic acid that can be detected based on a serial dilution input and not on organisms per reaction given that T. foetus organism counting is inherently inaccurate. A total of 8 replicates for each 2-fold organism input dilution were used to determine the direct RT-qPCR 95% detection rate (LOD); Cq values <40 were considered a positive response or amplification for this analysis. The LOD was determined to be at a dilution of log₁₀ – 1.6489 at 34.2 Cq. The RNA copy number equivalent for the LOD was estimated using a separate set of RT-qPCR targeting serial dilutions of an in vitro transcribed T. foetus control RNA and linear regression analysis. The direct RT-qPCR LOD was ~130 RNA copies per reaction. The Probit plot is shown in Supplementary Figure 2 (probability vs. log₁₀ [dose] T. foetus input dilution). The XIPC internal control target was successfully amplified in all reactions used for Probit analysis; the mean Cq was 30.1 ± 0.2.

Figure 1.

Linear dynamic range and efficiency of the direct reverse-transcription quantitative real-time PCR (RT-qPCR). Tritrichomonas foetus organism input serial dilutions, spanning a 5 log dynamic range, were used for nucleic acid purification and detection; 10,000 copies of XIPC RNA were spiked into each nucleic acid purification. Each input dilution was tested 12 times; thus, each data point is the average of 12 nucleic acid purifications and RT-qPCR. Cq = quantification cycle.

DSe and DSp of the direct RT-PCR and culture qPCR were calculated using microscopic reading of cultures as the reference test for 166 field reference samples. Direct RT-qPCR exhibited 100% DSe and DSp; the agreement between the direct RT-qPCR and culture readings was 100% (kappa = 1.0). The positive samples average XIPC Cq was 32.6 ± 1.1, and the negative samples average XIPC Cq was 31.8 ± 1.3, indicating proper functionality of nucleic acid purification and detection methods. Culture qPCR exhibited 95% DSe (3 of 56 positives by culture reading were negative by culture qPCR) and 100% DSp; the agreement between the culture qPCR and culture readings was 98% (kappa = 0.96; McNemar test p = 0.25), indicating no statistically significant difference between the 2 tests. The positive samples average XIPC Cq was 33.4 ± 2.0, and the negative samples average XIPC Cq was 32.8 ± 1.4, indicating proper functionality of workflow. The agreement between culture qPCR and direct RT-qPCR was 98% (kappa = 0.96; p = 0.25), indicating no statistically significant difference between the 2 tests (Table 2). However, the direct RT-qPCR identified 3 more positive animals and exhibited lower Cq values among the positives by culture reading than the culture qPCR: 14.6–32.3 for direct RT-qPCR versus 18.7–37.4 for culture qPCR (Fig. 2). The average Cq value for the direct RT-qPCR was 23.0 ± 4.3, whereas the average Cq value for the culture qPCR was 26.5 ± 4.3, and a paired t-test indicated a statistically significant difference (t(55) = −7.67, p < 0.0001). The mean difference between Cq values (direct RT-qPCR minus culture qPCR) was −3.5 Cq (95% CI: –4.42 to −2.59). The lower Cq range of the direct RT-qPCR results enabled better data interpretation because all Cq values were outside the typical inconclusive or suspect range (35–40 Cq).

Table 2.

Performance comparison between the direct reverse-transcription quantitative real-time PCR (RT-qPCR) and culture qPCR. A total of 166 field samples (56 positives and 110 negatives) were analyzed.

Culture qPCR	Direct RT-qPCR			PPA	NPA	Agreement	Kappa	95% CI	McNemar test (p)
	Positive	Negative	Total
Positive	53	0	53	0.97%	0.99%	0.98%	0.96	0.91–1.00	0.2482
Negative	3	110	113	(0.94–1.00)	(0.97–1.00)	(0.94–1.00)
Total	57	110	166

NPA = negative percent agreement; PPA = positive percent agreement. Numbers in parentheses are 95% confidence intervals.

Figure 2.

Direct reverse-transcription quantitative real-time PCR (RT-qPCR) and culture qPCR quantification cycle (Cq) values plot for 56 field sample positives by culture readings.

Specificity of the direct RT-qPCR was also assessed using 543 field smegma samples collected from bulls that were negative by initial culture qPCR testing. These re-collected samples were found to be negative by direct RT-qPCR and the second culture qPCR testing that was performed concurrently with the direct RT-qPCR.

Direct RT-qPCR performance was also evaluated using a small set of cervical mucus and uterine fluid samples collected from aborting cows and nonpregnant purchased replacements. Five of the 16 samples were T. foetus positive (Cq range = 9.9.–29.4; Table 3), 1 was suspect positive (Cq = 35.6), and 10 were negative. Two positive, and the suspect positive, samples were from cows that had aborted in the final trimester of gestation, and 3 positive samples were from nonpregnant replacements. Three negative samples were from aborting cows; one of these cows was infected with bovine viral diarrhea virus. The remaining 7 negative samples were from nonpregnant purchased replacements.

Table 3.

Quantification cycle (Cq) values of 5 Tritrichomonas foetus–positive and 1 suspect-positive cow cervical mucus samples. Cows 1–3 had late-term abortions; cows 4–6 were nonpregnant replacements.

Cow	Cq
	T. foetus	XIPC
1	9.9	32.7
2	17.3	29.0
3	35.6	31.8
4	29.4	30.0
5	25.9	29.5
6	25.3	29.8

XIPC = exogenous internal positive control.

Our results demonstrate improvements enabled by the direct RT-qPCR for detection of T. foetus infection. The RT-qPCR component of the workflow enables enhanced T. foetus detection Se and specificity by targeting the highly repetitive 5.8S rRNA gene. The T. foetus 5.8S rRNA gene has been shown to be present at 12 copies per T. foetus genome³; consequently, transcription of these gene copies produces high number copies of the RNA. Thus, targeting RNA provides enhanced detection Se that has eliminated the requirement of commercial pouch culture or equivalent medium. An additional benefit of the selected primers and probe is the nucleic acid target specificity. A previous study reported false-positive PCR results as the result of Simplicimonas moskowitzi observed in samples obtained from bulls and tested using the McMillen and Lew PCR assay,¹² in spite of 5 mismatches in the forward primer and 3 mismatches in the reverse primer sequences (Schommer et al, Cross-reaction of Simplicimonas spp. Trichomonads in Tritrichomonas foetus assays. Proc Am Assoc Vet Lab Diagn; Sept 2011; Buffalo, NY). In a 2017 published study, “Simplicimonas-like” sequences were amplified from vaginal swabs from cows exhibiting vaginitis.⁹ Melting curve and sequencing analysis indicated that the amplified sequences exhibited 91% homology to Simplicimonas spp., were subsequently described as “Simplicimonas-like,” and deposited in GenBank as accession KY410341. These Simplicimonas-like DNA were present in 44% of the collected vaginal swabs that were positive in qPCR (Cq 36–40) that utilized routine T. foetus primers.⁹ Unlike the aforementioned primers, the direct RT-qPCR reverse primer used in our study does not align to the “Simplicimonas-like” sequence (KY410341) or S. moskowitzi (GenBank accession GQ254636.1; Supplementary Fig. 1), did not amplify in vitro transcribed RNA containing “Simplicimonas-like” sequences (KY410341), and thus would likely enable optimal target specificity.

The nucleic acid purification component of the workflow utilizes a commercial automated nucleic acid purification method⁸ and enables the use of smegma, eliminating challenges associated with culture sample shipment and incubation, and adverse effects of bacterial growth. The automated nucleic acid purification method also reduces human labor and enables high-throughput sample processing. The direct RT-qPCR enables detection of T. foetus in cow cervical mucus and uterine fluids from aborting cows and nonpregnant replacement cows. The number of cow samples analyzed in our study was small given that the purpose of the analysis was to assess feasibility. The preliminary results indicate that direct RT-qPCR may have the potential to be utilized for T. foetus detection in cows from infected high-risk herds (e.g., cows experiencing known recent reproductive loss). However, caution should be exercised when negative results are obtained because the T. foetus organism often does not colonize the vagina but moves into the uterus making retrieval of the antigen impossible. The direct RT-qPCR can only detect T. foetus organisms that are accessible during the sample collection process and should not be used as a regulatory test to confirm negative females.

Maintenance of infection in cows is likely an overlooked factor in the epidemiology of T. foetus because cows are rarely tested and hence their infection status is unknown. If cows are infected, they may subsequently infect clean bulls. A prior study provided this evidence when 8 uninfected bulls mated to infected heifers resulted in 7 of 8 bulls becoming infected.⁴ Once infected, bulls can transmit to cows with high efficiency through a single service.^2,13 Cows usually eliminate the T. foetus organism and resume normal estrus, but a small proportion of females can maintain infection through gestation, deliver a normal calf, and maintain postpartum infection, serving as a reservoir of infection.^1,11,17 Thus, the management of T. foetus in cows is essential in the control of the disease. Direct RT-qPCR with enhanced detection Se may be valuable in cow testing. Additional performance evaluation using more cow samples would be beneficial in the management of T. foetus.