Introduction

Nodding syndrome (NS) is a mentally and physically disabling disease that mainly affects children between the ages of 5 and 15 years old leading to progressive cognitive dysfunction, neurological deterioration, stunted growth, and a characteristic nodding of the head [1]. Initially described by Jilek-Aall in Mahenge village of Southern Tanzania in 1970 [2], NS was considered a mysterious neurological disease occurring exclusively in onchocerciasis-endemic areas in Uganda, South Sudan and Tanzania. Today, NS has been recognized, together with Nakalanga syndrome, as part of onchocerciasis-associated epilepsy (OAE) and has been reported in several onchocerciasis-endemic regions with high prevalence of epilepsy and high ongoing or past Onchocerca volvulus transmission, such as in the Democratic Republic of the Congo, South Sudan, Cameroon and the Central African Republic [3–5].

The pathophysiological mechanism of OAE, including NS, remains to be elucidated. Thus far, the only identified risk factor for OAE is the degree of O. volvulus infection, specifically the O. volvulus microfilarial load [6]. In onchocerciasis-endemic regions, a high prevalence of OAE has been observed among households living near breeding sites of blackflies (the vector of onchocerciasis) [7]. Many children in these households may suffer from OAE, but not all individuals with a high microfilarial load develop OAE [6]. Therefore, co-factors, such as genetic factors, may play a role in the aetiology of OAE.

The Human Leukocyte Antigens (HLA) or Human Major Histocompatibility complex (MHC) are a group of proteins encoded by HLA genes on the short arm of the sixth chromosome [8,9]. The HLA complex is one of the most gene-dense and diverse regions of the human genome and has been linked to a series of genetic and infectious diseases [10–12]. Certain HLA types, especially HLA-class II variants, have been associated with either immunity or increased susceptibility to generalized or localized onchocerciasis [13]. More specifically, a correlation was found between allelic variants of HLA-DQA1 (protective: *04:01; risk: *01:02, *01:03 and *03:01) and changes in clinical manifestations in north-western Ecuador [14]. In a case-control study including 48 patients with NS and 51 healthy controls conducted in South Sudan’s Mundri Area by Benedek et al., an association was reported between NS and the HLA haplotype: HLA-B*42:01, C*17:01, DRB1*03:02, DQB1*04:02, DQA1*04:01, and the susceptible motif: Ala24, Glu63 and Phe67 in the HLA-B peptide-binding groove of HLA-B*35:01, *51:01 and *53:01 [15]. These amino acids create a hydrophobic and sterically closed peptide-binding HLA pocket, favoring the proline residue [15]. However, five cases in the South Sudanese study were from the same family, and cases and controls were from different villages and ethnicities. Moreover, cases and controls were not tested for onchocerciasis. Hence, it is possible that the observed HLA association was related to the families and ethnicities involved in the study or to susceptibility to onchocerciasis rather than NS.

While previous studies in Mahenge have focused primarily on infectious and environmental contributors, little is known about host genetic susceptibility [16–18]. To further investigate the potential association between OAE/NS and HLA haplotypes, a case-control study was conducted in 2021–2022 in Mahenge, an onchocerciasis-endemic area with a high-prevalence of OAE/NS in Tanzania [17,19].

Materials and methods

Study setting

A case-control study was conducted in Mahenge, Ulanga District, Tanzania, in 2021–2022. Participants were selected from individuals previously identified with epilepsy during door-to-door surveys conducted in 2017, 2018 and 2021 [17–19]. These participants came from nine rural villages in the Mahenge area (Mzelezi, Mdindo, Sali, Msogezi, Ebuyu, Euga, Isyaga, Isongo and Mgolo) and a suburban village (Vigoi) (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Map of Mahenge Area showing the ten surveyed villages (map prepared by A Hendy).

[16,20].

https://doi.org/10.1371/journal.pntd.0012971.g001

The fast-flowing rivers in the Mahenge Mountains are known to be suitable breeding sites for blackflies [21]. The surrounding area was infamous for having a high onchocerciasis endemicity before the implementation of annual community-directed treatment with ivermectin (CDTI) in 1997 [22–24], which was switched to biannual CDTI in 2019 [19]. The Wapogoro are the predominant ethnic group in the area. These people are sustaining themselves through agriculture, livestock farming (chickens, goats and pigs, with the latter mostly kept indoors and in suburban villages) and gem mining. Other filarial worms are not known to be endemic in the area [16].

Results

A total of 98 persons with epilepsy and 112 controls were included in the HLA study (Table 1). There were no statisticaly significant differences in socio-demographic characteristics of persons with epilepsy and controls for sex (p-value (p)=0.86), age (p = 0.93) and ethnicity (p = 0.10). Most persons with epilepsy (cases) met the criteria of OAE (93, 94.9%) and about one-third (37, 37.8%) met the criteria of probably NS (Fig 2). Only one person with NS (male; 45 years old) was identified with nodding syndrome with seizure onset at age five did not meet the strict OAE criteria because it was not clear whether there were other causes for epilepsy during childhood. Cases had a significantly higher prevalence of onchocerciasis-related skin disease than controls (4.1% versus 0.0%, p = 0.05), and a non-significantly higher prevalence of onchocercal nodules (19.4% versus 10.7%, p = 0.08). Anti-Ov16 seroprevalence was significantly higher among cases (76.5%) than controls (58.9%, p = 0.01). Most cases and controls (>90%) had taken ivermectin in the past.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic and clinical characteristics of the study population.

https://doi.org/10.1371/journal.pntd.0012971.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Venn diagram of population distribution based on epilepsy, onchocerciasis-associated epilepsy, nodding syndrome and anti-Ov16 seropositivity.

https://doi.org/10.1371/journal.pntd.0012971.g002

After applying the Bonferroni correction for multiple comparisons, no significant differences in HLA patterns were observed between persons with epilepsy and controls (p = 0.36). Similarly, no specific HLA patterns were associated with OAE (p = 0.26), probable NS (p = 0.29) or anti-Ov16 seropositivity (p = 0.46). However, several HLA patterns showed potential relevance before multiple test adjustment (Table 2): Considering non-adjusted p-values, HLA-C*07:01 was associated with epilepsy, HLA-DQB1*06:02 was associated with OAE, HLA-B*35:01 was associated with probable NS, and HLA-C*08:02 and HLA-DRB1*03:01 were associated with anti-Ov16 seropositivity. All frequency tables based on the different case-control datasets (epilepsy/OAE/NS/anti-Ov16 seropositivity) are available in S1 File.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Association statistics of identified alleles of interests and post-hoc power analysis.

https://doi.org/10.1371/journal.pntd.0012971.t002

Based on the post-hoc power analysis the minimal sample size to provide enough power would be 219 based on the highest frequency allele, whereas the needed sample size for the lowest perceived frequency would be 16,539. With the sample size used, we reached a minimal power of 0.05 and maximally 0.70, depending on the allele frequency and effect size. To provide further context, Table 2 presents the achieved power and the estimated sample size required for each of the alleles identified.

Population statistics showed that all loci, except for DRB3, presented in HWE. However, looking at individual genotypes, HLA class I (A, B, and C) show at least 11 genotypes deviating from HWE. As for class II, incl. DRB1, DQB1, DQA1, DPB1, DRB3, DRB4, DRB5 and DPA1, showed maximally 6 genotypes deviating from HWE. From the previously identified alleles, C*07:01 was shown to deviate from HWE in co-occurrence with C*17:01 (p = 0.02), as did DQB1*06:02 + DQB1*04:02 (p = 0.02). While testing for neutrality in the population, loci B and DRB1 showed significant deviation (p = 0.01, p = 0.04 respectively) from neutrality.

Discussion

In this case-control study conducted in Mahenge, Tanzania, no specific HLA patterns could be statistically significantly associated with epilepsy, OAE, probable NS nor to anti-Ov16 seropositivity. A case control study conducted in 2020 by Benedek et al. in Mundri, South Sudan, suggested that NS was linked to specific HLA patterns [15]. However, the HLA patterns described by Benedek et al were not associated with epilepsy, including OAE and NS, in our study population in Mahenge [15]. Both studies had limited sample sizes, so comparisons based on non-adjusted p-values may still provide useful insights. Benedek et al. also reported additional HLA alleles linked to NS with significant p-values before adjustment [15]. Similar to Benedek et al. our study also identified HLA-B*35:01 to be significantly associated with NS before adjustment. Given that both studies independently identified this specific HLA haplotype to be a potential risk factor for developing NS, and Benedek et al. also found a specific HLA-B binding groove susceptible motif (Ala24, Glu63, and Phe67) to be significantly present in this haplotype, HLA-B*35:01 might be an interesting focus for further investigation.

Considering non-adjusted p-values, HLA-C*07:01 was perceived to be a risk factor for epilepsy. This allele has been linked to febrile infection-related epilepsy syndrome (FIRES), a rare neuroinflammatory disease characterised by a febrile illness prior to seizures onset with high morbidity in previously healthy children and young adults [30,31]. FIRES’s pathogenesis remains unknown, with current hypotheses pointing to a cytokine-mediated autoinflammatory mechanism rooted in microglial inflammation [31].

Similarly, considering non-adjusted significant p-values, HLA-DQB1*06:02 was associated with OAE. This allele has been linked to several auto-immune diseases, such as multiple sclerosis, systemic lupus erythematosus, narcolepsy and sarcoidosis [32]. Also based on non-adjusted p-values, HLA-C*08:02 and HLA-DRB1*03:01 were associated with the presence of Ov16 antibodies. However, none of these HLA types have previously been reported in association with O. volvulus antibody seropositivity, which has mostly been linked to HLA-DQ alleles [13,14]. However, HLA-DRB1*03:01 has been reported as a major risk factor for systemic lupus erythematosus [33] and acts as a disease modifier for autoimmune hepatitis type 1 [34].

Population statistics indicated that DRB3 was not in HWE, suggesting other driving factors of HLA diversity at this locus. These factors might be due to biological reasons such as selection pressure, population structure, or non-random mating [35]. As the ethnic groups are generally well matched, the Wahlund effect on this locus should be minor. Moreover, familial ties were kept to a minimum during case-control selection (8%), thereby minimal impact of non-random mating due to consanguinity is to be expected. Selection pressure could underlie this disequilibrium but no DRB3 alleles were linked to epilepsy in this study. However, due to the nature of the environment, multiple other infectious diseases are common in the population and could therefore drive this selection. This is in line with the well-documented tendency of deviation from HWE in HLA that are not uncommon as the immune-profile of populations are primarily driven by the pathogens circulating in the environment [36]. Nevertheless, the small sample size may also contribute to the observed equilibrium, which also extents to the deviation found in the individual alleles. Notably, two of our previously identified interesting alleles, C*07:01 + C*17:01 (p = 0.002) and DQB1*06:02 + DQB1*04:02 (p = 0.017), were found to be out of HWE during multi-locus analysis. Their respective links to epilepsy and OAE warrant further investigation into the co-occurrence with these alleles. However, due to their proximity, linkage is to be expected.

Similar evolutionary forces are likely to explain the absence of neutrality in HLA-B and DRB1, where long-term pathogen-mediated balancing selection influences allele frequencies over time, complementing the short-term genotype frequency deviations seen in HWE tests.

A slightly significantly higher proportion of cases presented with skin manifestations and nodules suggesting O. volvulus infection. There was also a significant difference in anti-Ov16 seropositivity between cases and controls. This is in line with the previously reported larger matched case-control study in which anti-Ov16 seroprevalence was positively associated with probable NS (OR: 5.05, 95%CI: 1.79-14.27) and overall epilepsy (OR: 2.03, 95%CI: 1.07-3.86) [25].

The pathogenesis of OAE, including NS, remains unresolved [37]. Epidemiological data indicate that higher microfilarial loads in childhood and the typical onset age (≈8–12 years) are important correlates of risk [37–39]. Cross-reactive autoimmunity (molecular mimicry with host neuronal proteins) has been proposed but remains unconfirmed and contested [40–42]. Evidence of chronic neuroinflammation, including gliosis and activated microglia on neuropathology [42], and immune/complement activation, supports an inflammatory contribution [43]. Although, cerebrospinal fluid cytokine signals have been inconsistent across small studies [44]. Macrophage migration inhibitory factor (MIF) may play a dual role: high-expression MIF genotypes appear protective, yet plasma MIF is elevated in NS and could amplify neuroinflammation [45]. Finally, metagenomics work has identified a rhabdovirus (OVRV1) in the O. volvulus virome across life-cycle stages and inducing an immunological response in people living in onchocerciasis endemic areas [46]. Studies are currently underway to test whether this virus could contribute to OAE pathogenesis.

Our case–control study confirms the now overwhelming epidemiological evidence for the association between onchocerciasis and epilepsy. As with many infectious disease-related conditions, genetic factors likely play a role; however, our study did not identify a clear association with any specific HLA haplotype. NS is certainly not a genetic disease. The primary risk factor for NS remains exposure to O. volvulus. It is crucial that both affected communities and public health decision-makers understand this, as NS is a preventable condition. Strengthening onchocerciasis elimination programs—such as switching from annual to bi-annual CDTI as implemented in Mahenge—dramatically reduced the incidence of NS [47]. Elimination of onchocerciasis transmission was shown in Uganda to also eliminated NS [48].

One major limitation of this study is the limited sample size. HLA genes are predominantly linked to ethnicity and are high in polymorphisms, making the ideal sample size for accurately representing a population and its disease associations exceptionally large. In addition, we were unable to compare our findings with HLA data from other African populations, which is particularly important given the extensive variation in HLA genes across Africa. However, due to the high costs of HLA typing, we were only able to conduct a case-control study with a limited number of participants. Based on the post-hoc analysis, expanding HLA typing of samples collected in future case-control studies and combining datasets from across sub-Saharan Africa will be needed to reach a sufficient sample size to identify potential genetic risk factors for OAE, including NS.

Conclusion

In our case-control study in Mahenge, Tanzania, no HLA patterns were found to be statistically significantly associated with epilepsy, OAE, probable NS or anti-Ov16 seropositivity. However, several alleles—particularly HLA-B*35:01, also reported in a previous South Sudanese study [15]—emerged as potential candidates for further investigation. A meta-analysis combining our data with that from the study conducted in South Sudan should be considered. Ultimately, larger case-control studies across sub-Saharan Africa will be essential to validate these potential associations and better understand the potential genetic factors contributing to OAE and NS.

Supporting information

Acknowledgments

We would like to thank the study participants, community members and leaders for accepting to participate in this study. Health authorities in Morogoro Region and Ulanga District for their cooperation and support to the study teams. Further, the data collectors for their commitment and the Belgian Red Cross for facilitating DBS sequencing.