Urinary N-acetyltyramine-O,β-glucuronide in Persons with Onchocerciasis-Associated Epilepsy

We investigated urinary N-acetyltyramine-O,β-glucuronide (NATOG) levels as a biomarker for active Onchocerca volvulus infection in an onchocerciasis-endemic area in the Democratic Republic of Congo with a high epilepsy prevalence. Urinary NATOG was measured in non-epileptic men with and without O. volvulus infection, and in O. volvulus-infected persons with epilepsy (PWE). Urinary NATOG concentration was positively associated with microfilarial density (p < 0.001). The median urinary NATOG concentration was higher in PWE (3.67 µM) compared to men without epilepsy (1.74 µM), p = 0.017; and was higher in persons with severe (7.62 µM) compared to mild epilepsy (2.16 µM); p = 0.008. Non-epileptic participants with and without O. volvulus infection had similar NATOG levels (2.23 µM and 0.71 µM, p = 0.426). In a receiver operating characteristic curve analysis to investigate the diagnostic value of urinary NATOG, the area under the curve was 0.721 (95% CI: 0.633–0.797). Using the previously proposed cut-off value of 13 µM to distinguish between an active O. volvulus infection and an uninfected state, the sensitivity was 15.9% and the specificity 95.9%. In conclusion, an O. volvulus infection is associated with an increased urinary NATOG concentration, which correlates with the individual parasitic load. However, the NATOG concentration has a low discriminating power to differentiate between infected and uninfected individuals.


Introduction
The filarial nematode Onchocerca volvulus is transmitted by blackflies (Simuliidae) and causes skin disease (itching and rash), eye disease (progressive loss of vision), and epilepsy (onchocerciasis-associated epilepsy (OAE)) [1]. A recent study showed that children with a high microfilarial (mf, larval stage) density (>200 mf/skin snip) were 28 times more likely to develop epilepsy as compared to children without mf [2]. OAE is an important public health problem, since it is estimated that there are approximately 100,000-400,000 persons with OAE in Africa, which could have been prevented through improved onchocerciasis control [3].
Currently, elimination programs are ongoing in onchocerciasis-endemic regions through the mass distribution of ivermectin (community-directed treatment with ivermectin (CDTI)) [4,5]. Ivermectin Pathogens 2020, 9,191 2 of 10 kills the microfilariae and temporarily represses the production of new microfilariae by the female worms, albeit without killing the adult worms [6]. Onchocerciasis is diagnosed through the detection of microfilariae in skin snips. However, obtaining skin snips requires an invasive, slightly painful procedure and is therefore not accepted everywhere, especially in low-endemic regions. Alternatively, exposure to O. volvulus can be estimated by means of a rapid diagnostic test (RDT) detecting IgG4 antibodies against the OV16 antigen of O. volvulus [7]. The OV16 RDT was reported to have a specificity of 98-100% and a sensitivity of 76.5%, 81% or 90%, depending on the population [7,8]. A disadvantage of the aforementioned test is that it does not differentiate between active infection and past exposure and that it does not provide information about infection load.
Previous research exploring the metabolomic profile of O. volvulus-infected individuals living in villages in Ghana and Cameroon, revealed that N-acetyltyramine-O,β-glucuronide (NATOG) was enriched in the urine of O. volvulus-infected individuals [9]. NATOG is the inactivated form of the O. volvulus neurotransmitter tyramine, which is excreted in the urine of infected individuals [9,10]. Since the discovery of NATOG, several studies have tried to determine a cut-off value for urinary NATOG concentration to diagnose active O. volvulus infection [9][10][11][12]. However, the results of these studies are conflicting. Therefore, the urinary NATOG threshold indicating an active O. volvulus infection remains unknown. Furthermore, the correlation between urinary NATOG concentration and infection load is unclear.
In this study, we aimed to further explore the potential of urinary NATOG as a biomarker for O. volvulus infection in ivermectin naive individuals with and without epilepsy in an onchocerciasis-endemic area of the Democratic Republic of Congo (DRC), where an epilepsy prevalence of 4.6% was documented in 2016 [13]. In addition, we investigated the association between urinary NATOG concentration and skin mf density in persons with epilepsy.

Study Population and Sample Collection
The study was performed in the Logo health zone, Ituri province, DRC, in onchocerciasis-endemic villages where ivermectin was never distributed. Ethical approval was obtained by the ethical committee of the university hospital of Antwerp (May 24, 2017, B300201733011) and the ethical committee of the university of Kinshasa (February 28, 2018, ESP/CE/013/2018). Men older than 20 years, having no epileptic seizures and living in these villages for more than 10 years, were asked to participate in a rapid epidemiological mapping of onchocerciasis (REMO) to determine the degree of onchocerciasis-endemicity in the area. They were examined for the presence and number of palpable nodules and tested for exposure to O. volvulus using the OV16 IgG4 rapid diagnostic test (OV16 RDT, SD Bioline Onchocerciasis IgG4 rapid test, Abbott Standard Diagnostics, Inc., Yongin, Republic of Korea). Twenty men with evidence of O. volvulus infection (OV16 RDT positive with the presence of at least one nodule) and 19 men without evidence of O. volvulus infection (OV16 RDT negative and no nodules) were asked to provide a urine sample. Skin snip testing was not performed on these subjects.
In addition, persons with epilepsy (PWE) and with parasitological and/or serological evidence of O. volvulus infection were recruited during a clinical trial investigating the effect of ivermectin on seizure frequency [14]. PWE were randomized in two groups according to a 1:1 ratio, with one group receiving phenobarbital and an ivermectin dose of 150 µg/kg and the other group receiving only phenobarbital [15]. After informed consent was obtained, participants were interviewed to document their medical history and a physical examination was performed, including a neurological assessment. Participants were examined for the presence of nodules, but a nodule count was not recorded. Skin snips were obtained from the left and right iliac crests to determine the mf density; the OV16 RDT was used to detect O. volvulus antibodies and participants provided a urine sample. PWE having two or more seizures per month were classified as having a severe form of epilepsy whereas those with less than two seizures per month were considered to have a mild form of epilepsy. The PWE had not received antiepileptic drug treatment within the last two weeks before enrolment in this study. Four months after enrolment, a follow-up clinical examination was performed, skin snip testing was repeated, and a second urine sample was obtained.
Persons without epilepsy were considered to have an active O. volvulus infection if they were OV16 RDT positive, had never received ivermectin and had palpable nodules, in the absence of skin snip results. Ivermectin-naive PWE were considered to have active infection when mf was present in skin snips, or if they had a positive OV16 RDT, with no mf but with nodules. Ivermectin-naive PWE with a positive OV16 test with no mf in their skin snips and no nodules were not considered to have an active infection, but a pre-patent infection (whereby adult females are not yet producing mf).

Mass Spectrometry Analysis
NATOG concentrations were determined in urine samples by liquid chromatography tandem mass spectrometry (LC-MS/MS) at Janssen Global Public Health (Beerse, Belgium), as described earlier [12]. Quality control and calibration samples were prepared as described before [12] and a standard curve (1-100 µM) was constructed by spiking stock solutions of NATOG to 25-fold diluted urine. The LOQ was defined at 0.7 µM. Urine samples were diluted 25-fold in MilliQ-water and were first separated using a UPLC (Acquity UPLC; HSS T3 column; 1.8 µm, 2.1 × 50 mm, Waters Corporation) before analysis by electrospray ionisation (ESI) with triple quadrupole MS/MS (API 4000, AB Sciex). Data were collected using SRM in positive ESI mode with Q1 Mass 356.1 Da and Q3 Mass 180.1 Da. The calibration data were used to estimate a linear regression curve, that was subsequently considered to determine the NATOG concentrations of the collected urine samples.

Statistical Analysis
Medians and interquartile ranges (IQRs) were used to describe continuous variables, whereas categorical variables were characterized using absolute and relative frequencies. The correlation between urinary NATOG concentration and the number of nodules was examined using a Kendall's tau-b rank correlation coefficient. The association between mf density and urinary NATOG concentration was assessed using a Quasi-Poisson regression model while accounting for the dependence between pre-and post-treatment measurements on the same subjects using a generalized estimating equations (GEE) approach [16]. A logistic regression model with parameter estimation using a GEE approach was considered to assess the relationship between urinary NATOG concentration and active O. volvulus infection in ivermectin-naive PWE. Receiver operating characteristic (ROC) curves were constructed to illustrate the diagnostic ability of classifying persons with and without O. volvulus infection in two groups based on their observed urinary NATOG concentrations when the discrimination threshold is varied. Two-sided p-values <0.05 were considered statistically significant. Analyses were performed using SAS 9.4, SAS Institute Inc. and R, version 3.6.1.

Ethical Considerations and Informed Consent
The study was approved by the Ethics Committee of the School of Health of the University of Kinshasa and the University of Antwerp, Antwerp, Belgium. All eligible candidates provided written informed consent before enrolment into the study.

Urinary NATOG as a Biomarker for Active O. volvulus Infection
The discriminative power of urinary NATOG concentration to discriminate between active O. volvulus infected and non-infected individuals was found to be low, with an area under the curve (AUC) of 0.721 (95% CI: 0.633-0.797, Figure 2). Furthermore, using the previously proposed cut-off value of 13 µM, 17 individuals were correctly classified as having active infection and 47 individuals were correctly classified as not having active infection, leading to a sensitivity of 15.9% and a specificity of 95.9% (Table 5).

Urinary NATOG as a Biomarker for Active O. volvulus Infection
The discriminative power of urinary NATOG concentration to discriminate between active O. volvulus infected and non-infected individuals was found to be low, with an area under the curve (AUC) of 0.721 (95% CI: 0.633-0.797, Figure 2). Furthermore, using the previously proposed cut-off value of 13 μM, 17 individuals were correctly classified as having active infection and 47 individuals were correctly classified as not having active infection, leading to a sensitivity of 15.9% and a specificity of 95.9% (Table 5). Using the previously proposed NATOG cut-off value of 13 μM for the diagnosis of active O. volvulus infection and an uninfected state [10], the sensitivity of urinary NATOG was 15.9% and specificity was 95.9% (Table 5). Using the previously proposed NATOG cut-off value of 13 µM for the diagnosis of active O. volvulus infection and an uninfected state [10], the sensitivity of urinary NATOG was 15.9% and specificity was 95.9% (Table 5).

Discussion
This is the first study investigating the association between urinary NATOG concentration and O. volvulus mf density and the effect of ivermectin treatment on urinary NATOG concentrations in individuals with OAE. The urinary NATOG concentration decreased after ivermectin treatment, with a corresponding drop in mf density. Furthermore, we observed higher mf densities and urinary NATOG concentrations in PWE compared to persons without epilepsy, and an increasing trend in NATOG concentration from uninfected individuals (lowest concentration) to infected individuals, persons with mild epilepsy, and persons with severe epilepsy (highest concentration). However, there was no difference in urinary NATOG between infected and uninfected people without epilepsy.
Both the urinary NATOG concentration and skin mf density decreased in the ivermectin-treated and untreated PWE. However, the decrease in mf density (100%) in the treated group compared to the untreated group (83%) was significant, while the decrease in urinary NATOG concentration between both groups was not significant. The explanation for the decrease in mf density and urinary NATOG concentration in PWE that were not treated with ivermectin is unclear. After performing skin snips, all PWE were started on phenobarbital. Currently, the effect of phenobarbital on O. volvulus is unknown, but this anti-epileptic drug acts as a gamma-aminobutyric acid (GABA) A receptor subunit agonist [17]. GABA-gated chloride channels are required for O. volvulus locomotion and are a drug target for ivermectin [18]. Therefore, one explanation could be that phenobarbital may have reduced mf densities or motility, leading to reduced emergence from the skin snip before counting, potentially by acting on the parasite GABA-receptor subunits. However, it cannot be excluded that pre-analytic differences in urine collection or differences in skin snip collection influenced the results.
The average urinary NATOG concentration in individuals with active infection in our study was 8.9 µM, which is close to the 8.4 µM observed in Guatemalan samples in a study by Globisch et al. [9] ( Table 6). This concentration is below the 13 µM threshold required to identify active O. volvulus infection in Africa, proposed by Globisch et al. This high threshold was proposed by Globisch et al. because, in a study in Ghana, high urinary NATOG concentrations were observed in O. volvulus uninfected endemic African controls (Table 6) [9,10]. However, the diagnosis of uninfected controls was mainly based on the absence of nodules. In only a limited number of participants were skin snips obtained, and mf densities were not taken into account. Another explanation for the elevated urinary NATOG concentrations in the participants from Ghana might be co-infections with other nematodes, such as Mansonella perstans and/or Loa Loa. Indeed, urinary NATOG concentrations in O. volvulus non-infected individuals with other nematodes was, on average, 9.29 µM [10]. Eight individuals with an O. volvulus, Loa. loa and M. perstans co-infection had very high NATOG concentrations 100.5 ± 33.5 µM [10]. None of the PWE in Ituri were Loa loa-infected, but co-infection with M. perstans was not assessed. Globisch et al suggested that the lower urinary NATOG in the Guatemala samples compared to their samples obtained in Ghana and Cameroon could be explained by the genetic diversity of O. volvulus. A more likely explanation is that the mf densities in Guatemala were lower than those in Ghana and Cameroon because of previous ivermectin exposure. In fact, very low urinary NATOG concentrations (1.06 µM) were also detected in another population in Ghana, investigated by Lagatie et al. In the latter study, low mf densities (0-10 mf/mg skin) most likely resulting from previous ivermectin use, may explain the low NATOG concentrations observed [12]. Indeed, the urinary NATOG concentrations and skin mf densities in the study by Lagatie et al. were similar to those observed in our study participants after ivermectin treatment (Table 6). Recently, a lateral flow immunoassay (LFIA) test was developed to detect urinary NATOG [19]. In a small study, this test accurately diagnosed 23 (85%) of the 27 African samples tested, with a cut-off concentration set at 25 µM [19]. However, the authors of the above-mentioned study did not provide detailed information about mf densities.