Nodding syndrome is unlikely to be an autoimmune reaction to leiomodin-1 after infection by Onchocerca volvulus

Abstract

Nodding syndrome is a neurological disease of children in northern Uganda. Infection with the nematode parasite Onchocerca volvulus has been epidemiologically implicated as the cause of the disease. It has been proposed that an autoantibody directed against the human protein leiomodin-1 cross reacts with a tropomyosin-like nematode protein, thus suggesting that nodding syndrome is an autoimmune brain disease due to extra-cerebral parasitism. This hypothesis is dependent on constitutive neuronal expression of leiomodin-1. We tested this hypothesis by studying the distribution of leiomodin-1 in the normal human brain and other human tissues using immunohistochemistry. We found that immunostaining for leiomodin-1 follows a smooth muscle cell specific pattern. In the brain, it is confined to the smooth muscle cells of cerebral blood vessels and is not generally present in neurons or glia. However, immunoreactivity was identified in human Purkinje cell membrane and the body wall of C. elegans (as a proxy for Onchocerca volvulus) but only when immunostained with an antibody recognizing the N-terminal of leiomodin-1. Homology between leiomodin-1 and tropomodulin, specifically at the N-terminus, could explain why leiomodin-1 antibody cross reactivity between human Purkinje cells and C. elegans. However, we cannot provide proof confirming that the immunoreactivity in the membranes of Purkinje cells is specifically caused by the expression of tropomodulin. To overcome this limitation, further investigations using additional immunohistochemical and biochemical studies are required to corroborate our findings and provide more comprehensive evidence. Nevertheless, our findings do not support to the autoimmunity hypothesis involving Onchocerca volvulus and leiomodin-1. To gain a more comprehensive understanding of the cause and pathogenesis of NS, it is essential to explore alternative hypotheses.

Highlights

• •In the brain, Leiomodin-1 is mostly expressed in smooth muscles cells of blood vessels.
• •Leiomodin-1-like immunoreactivity is present in human Purkinje cell membranes.
• •Positive Leiomodin-1 immunostaining in the body wall of C. elegans.
• •Cerebellar and nematode immunopositivity may be due to tropomodulin cross reactivity.
• •Leiomodin-1 autoimmunity after nematode infection is unlikely cause of nodding syndrome.

1. Introduction

Nodding syndrome (NS) is a debilitating neurological disorder in rural communities in the Eastern sub-Saharan African countries of Tanzania [1], South Sudan [2] and more recently, Uganda [3]. The disease only affects children aged 5–15 years [4], at the onset. In Uganda, the outbreak started with daily bouts of atonic seizures manifesting as stereotypical head dropping movements; progressing to cognitive impairment, grand mal seizures, and neurological deterioration that can lead to death [4,5]. Initial neuropathologic findings indicate the presence of tau pathology in the neocortex, the midbrain, and pons, with peculiar absence of pathology in the hippocampus [6,7]. The cause of NS is unknown, but an epidemiological association with infection by the parasitic nematode that causes river blindness, Onchocerca volvulus, has been reported [5,8]. Indeed, case-control studies demonstrated a positive, but not invariant, relationship between NS case status and infection with Onchocerca volvulus [8]. Interestingly, despite this epidemiological relationship, there has been no evidence of Onchocerca volvulus in the cerebrospinal fluid (CSF) nor in brains of NS patients, leading to doubt in the validity of this theory.

A proteomics’ assay of sera from NS patients conducted by Johnson et al. [9] demonstrated putative autoantibodies to leiomodin-1 (LMOD1), a smooth muscle cell protein [9]. It was claimed that LMOD1 protein expression occurred in the hippocampus, Purkinje cells of the cerebellum and cortical neurons of the mouse brain [9]. Thus, these investigators proposed that LMOD1 in neurons is the target of an autoimmune attack, and this could explain how Onchocerciasis could cause NS. Accordingly, a subsequent study sought to explore the expression of LMOD1 in neurons developed in vitro from iPSC (induced pluripotent stem cells). Nauen et al. [10] detected the presence of LMOD1 in glia, along with expression in the membrane of newly formed neurons [10]; further supporting the hypothesis put forth by Johnson et al. [9].

There is no general agreement about the cause of NS; the LMOD1 autoimmunity hypothesis is controversial [11]. Indeed, Johnson et al. [9] did not determine the cellular distribution of LMOD1 in human brain tissue. In addition, neuropathologic studies of NS brains did not reveal direct infection by the nematode; Onchocerca volvulus [12]. Furthermore, Nanda et al. showed that LMOD1 expression in brain is low, which is atypical for a neuronal protein [13]. Additionally, LMOD1 immunoreactivity was not observed in neurons in the mouse brain [13], in contrast to findings of Johnson et al. [9]. Furthermore, a recent study [14] seeking to replicate findings of Johnson et al. [9] did not reproduce the results of the initial study. Based on this controversy, we further tested the LMOD1 autoimmunity hypothesis by studying the immunohistochemical distribution of LMOD1 in normal human tissues and the nematode C. elegans, as model for Onchocerca volvulus. Our results are the first to explore the histologic distribution of LMOD1 in human tissues. We found no evidence of LMOD1 generalized expression in human brain neurons along with a lack of overlap with tau pathology. In humans, LMOD1 is predominantly a smooth muscle protein with expression in the brain in blood vessels. On this basis, we cannot support the LMOD1 autoimmunity hypothesis. This is significant because this theory was the primary scientific support for onchocerciasis as the cause for NS. Therefore, it is critical that other hypotheses regarding the cause and pathogenesis of NS are now investigated.

2. Methods

2.1. Tissues

The project's main objective was to test the LMOD1 autoimmunity hypothesis by examining the histologic distribution of LMOD1 in the human brain, visceral tissue samples, and a model nematode. For this purpose, we utilized the brains of eight normal individuals ranging in age from 2 days to 15 years for immunostaining. We selected this age range, based on the young age of symptom onset for NS. In all cases, there was no pre-existing neurological disease and death was attributed either to trauma or drug intoxication. Immunostaining was also undertaken in five fatal cases of NS, as described previously [7,12]. We conducted postmortem examination and post-fixation examination of the brain based on consent from the affected families. Furthermore, we performed immunohistochemical analyses on tissue microarrays of 16 human tissues (Table 2) gathered from one normal individual [15]. Institutional ethical review was waived for use of normal brain and visceral tissue since the anonymized tissues had already been obtained for diagnostic purposes.

Table 2.

Tissues used for immunohistochemical analysis of tissue microarray slides using NBP1-89398 and PA5-44224 antibodies. Optimal conditions = 1:200 dilution with epitope retrieval solution 2 for pre-treatment. Level of staining = -, none; +/−, sparse; +, mild; ++, moderate; +++, marked.

Tissue Sample	Staining Pattern	NBP1 (Primary antibody to LMOD1)	PA5 (Primary antibody to N-terminal of LMOD1
Thalamus	Blood vessels	++	+
	Blush staining	–	+/−
Adrenal Cortex	Blood vessels	–	++
	Zona glomerulosa	+++	++
Spleen	Blood vessels	–	+
	Trabeculae	+/−	+
Cerebellum	Blood vessels	+	–
	Purkinje cells	–	+++
Liver	Blood vessels	–	+
	Hepatocytes	–	+/−
White matter	Blood vessels	++	–
Pons	Blood vessels	++	+
	Pontine nuclei	–	+/−
Cerebral cortex	Blood vessels	++	+
	Blush staining	–	+/−
Thyroid	Blood vessels	+	+
	Follicular cells	–	+/−
Pancreas	Blood vessels	+	+
Lymph node	Blood vessels	+	++
	Histiocytes	–	++
Lung	Blood vessels	+	+
	Wall of terminal bronchiole	+	–
Heart	Blood vessels	+	++
	Punctate staining of cardiomyocytes	–	++
Kidney	Blood vessels	+	++
Skeletal muscle	Blood vessels	+	++
Stomach	Blood vessels	+	++
	Muscle Propria	+	++

Leiomodin-1 antibody cross-reaction in C. elegans.

For immunostaining of a model nematode, a collection of 10000 Caenorhabditis elegans worms growing on two 10-cm Petri plates were washed with M9 buffer and centrifuged at 2000 rpm for 1 min, yielding a pellet of 50 μL in volume. The pellet was fixed in 4–6 mL of 10% neutral buffered formalin. For processing, the pellet was solidified in molten agar [16]. The resulting pellet was fixed in formalin and processed for embedding in paraffin.

2.2. Immunohistochemistry

We used 4-μm thick sections made from formalin-fixed paraffin embedded blocks from: (a) various brain regions, including the cerebral cortex, hippocampus, brainstem and cerebellum; (b) the tissue microarrays; and (c) C. elegans for immunohistochemistry. Immunostaining was conducted with two different antibodies to LMOD1: (a) a polyclonal antibody (NBP1-89398, 1:100 dilution) to the full-length recombinant protein; and (b) a polyclonal antibody (PA5-44224, 1:200 dilution) to a synthetic peptide directed towards the N-terminal of human LMOD1. Tissue sections were pre-treated with BOND epitope retrieval solution 2 (Leica Biosystems) for heat-induced epitope retrieval. In all cases, the immunoperoxidase method was utilized with diaminobenzidine as the chromogen. The omission of the primary antibody conducted negative controls. Stained slides were examined by light microscopy for the cellular distribution of immunoreactivity.

2.3. Sequence homology search

Human LMOD1 and C. elegans amino acid sequences were retrieved from UniProt. Sequence homology search was performed by PDBe (EMBL-EBI), and pairwise alignment was completed using Blast (NCBI Blast).

3. Results

3.1. Expression of leiomodin-1 in cerebral development

We examined LMOD1 immunoreactivity in a developmental spectrum of normal human brains ranging from the neonatal period to adolescence (Table 1). The age range for this study was determined by the age of symptom onset for NS, typically from 5 to 15 years of age. However, since the putative cerebral insult that causes NS could occur prior to symptom onset, younger brains were also examined by LMOD1 immunostaining. Using an antibody against human LMOD1 (NBP1-89398), we found that LMOD1 immunoreactivity was limited to the smooth muscle cells of the media in cerebral blood vessels (Fig. 1A–E). No immunostaining was generally present in any neurons, glia in the cerebral cortex (Fig. 1A and B), or any other brain region. All neuronal cell populations within the hippocampus were consistently unstained (Fig. 1D). This finding included the neuronal populations during development, such as migrating neurons and the external granular cell layer of the cerebellum (Fig. 1E). The LMOD1 immunostaining pattern in the cerebral cortex of five cases of NS was identical to that observed in the normal brains. There was no LMOD1 immunostaining of neurofibrillary tangles (Fig. 1F). In addition, epithelial cells of choroid plexus and ependymal cells were also unstained.

Table 1.

Tissues used to study the immunohistochemical distribution of LMOD1 in normal human brain.

Age/Sex	Diagnosis	Brain regions
2 day/female	Sudden death in an unsafe sleep environment	Cerebral cortex, hippocampus, pons, and cerebellum
45 day/female	Sudden death in an unsafe sleep environment	Cerebral cortex, hippocampus, pons, and cerebellum
7 weeks/male	Sudden death in an unsafe sleep environment	Cerebral cortex, hippocampus, pons, and cerebellum
4 years/male	Head injury	Cerebral cortex, pons, and cerebellum
9 years/male	Hanging	Cerebral cortex, thalamus, midbrain, and cerebellum
11 years/male	Hanging	Cerebral cortex, hippocampus, striatum, Pons, cerebellum, and cervical spinal cord
15 year/female	Oxycodone toxicity	Cerebral cortex, amygdala, midbrain, and cerebellum
15 year old/male	Hanging	Cerebral cortex, hippocampus, pons, medulla, and cerebellum

Fig. 1.

Cellular protein expression of LMOD1 in developmental spectrum of human brains. (A) LMOD1 immunoreactivity in medial smooth muscle cells of blood vessels in the subarachnoid space, Scale bar: 100 μm. (B) LMOD1 immunoreactivity of medial smooth muscle cells of blood vessels in the thalamus with no staining of neurons or glia, Scale bar: 250 μm. (C) LMOD1 immunoreactivity of medial smooth muscle cells of blood vessels in the thalamus with no staining of cortical neurons, including pyramidal cells, Scale bar: 50 μm. (D) No LMOD1 immunoreactivity in dentate granular cells of the hippocampus with adjacent vascular immunoreactivity, Scale bar: 50 μm. (E) No LMOD1 immunoreactivity in the developing cerebellum, including the external granular cell layer, Scale bar: 50 μm. (F) Cerebral cortex in nodding syndrome with neurofibrillary tangles (arrowheads) showing LMOD1 immunoreactivity only in a cortical arteriole, Scale bar: 25 μm.

3.2. Cellular localization of leiomodin-1

Immunostaining of a diverse set of human tissues with both NBP1 and PA5 showed a smooth muscle cell specific pattern of LMOD1 expression. Indeed, LMOD1 immunoreactivity was observed in the media of the blood vessels of various tissues such as the brain (pons, thalamus, white matter, cerebral cortex), heart, kidney, skeletal muscle, thyroid, and lung (Table 2, Fig. 2A–D & F–I). Additionally, we observed staining of the gastrointestinal tract's walls and the zona glomerulosa in the adrenal cortex (Table 2, Fig. 2C,D,H,I). Interestingly, we discovered further expression of LMOD1 in the membrane of Purkinje cells of cerebellar samples stained with PA5 (Fig. 2J) but not NBP1 (Fig. 2E). The membranous immunostaining pattern by PA5, but not NBP1 could be explained by a homologous protein to LMOD1 expressed in Purkinje cell membranes. We also confirmed the pattern of Purkinje cell membrane staining in the developmental spectrum of normal human brains (Fig. 3).

Fig. 2.

Cellular protein expression of LMOD1 in various human tissues with NBP1-89398 (antibody to LMOD1) and PA5-44224 (antibody directed towards N-terminal of LMOD1). (A–E) Immunostaining with antibody to LMOD1. (F–J) Immunostaining with to N-terminal of LMOD1. (A&F) LMOD1 immunoreactivity of medial smooth muscles cells of blood vessels in the heart (black arrows), Scale bar: 100 μm. (B&G) LMOD1 immunoreactivity of medial smooth muscles cells of cortical blood vessels (black arrows) with no staining of cortical neurons, Scale bar: 100 μm. (C&H) LMOD1 immunoreactivity of the zona glomerulosa in the adrenal cortex (arrowheads), Scale bar: 100 μm. (D&I) LMOD1 immunoreactivity of the muscularis propria in the stomach, Scale bar: 100 μm. (E) No LMOD1 immunoreactivity in the cerebellum including external granular cell layer, Purkinje cells and dendrites, Scale bar: 25 μm. (J) LMOD1 immunoreactivity of membrane of Purkinje cells and dendrites in the cerebellum, Scale bar: 25 μm.

Fig. 3.

Localization of LMOD1 in cerebellar samples using PA5-44224 (antibody directed towards N-terminal of leiomodin-1). (A–E) Representative cerebellar sections from developmental spectrum of brains, Scale bar: 25 μm. Arrows indicate membranous staining of Purkinje cells. Arrowheads indicate staining of dendrites.

We investigated if PA5 and NBP1 antibodies immunostained the nematode C. elegans, as a model of Onchocerca volvulus. We found that only the PA5 antibody immunostained C. elegans (Fig. 4D). The staining pattern was limited to the contractile elements of the body wall [17].

Fig. 4.

Sequence homology between human LMOD1 and tropomodulin in C. elegans. (A) Sequence homology between human LMOD1(Residues 13–77) and C. elegans tropomodulin (Residues 41–103) by pairwise alignment (NCBI Blast, blastp suite-2sequences). (B) Sequence homology between human LMOD1 (Residues 322–479) and C. elegans tropomodulin (Residues 228–386) by pairwise alignment (NCBI Blast, blastp suite-2sequences). (C) No LMOD1 immunoreactivity in C. Elegans with antibody to LMOD1, Scale bar: 25 μm. (D) Restricted LMOD1 immunoreactivity in body wall of C. Elegans with antibody recognizing N-terminal of LMOD1, Scale bar: 25 μm.

Since Purkinje cell membrane immunoreactivity was exclusive to the PA5 antibody, which is specific to the N-terminus of LMOD1, this suggested that the PA5 antibody likely cross-reacted with another protein that was homologous to LMOD1. On this basis, we conducted a homology search between human LMOD1 and well-studied proteins in C. elegans. C. elegans tropomodulin was the only protein identified. Human LMOD1 has 28.3% sequence identity to C. elegans tropomodulin. Even though the general homology was relatively low, regional homology analyses demonstrated that amino acids 13 to 77 of human LMOD1 showed 44% sequence identity and 63% sequence similarity to amino acids 41 to 103 of C. elegans tropomodulin (Fig. 4A). In addition, amino acids 322 to 479 of human LMOD1 showed 28% sequence identity and 55% sequence similarity to amino acids 228 to 386 of C. elegans tropomodulin (Fig. 4B). Overall, this homology could explain the shared PA5 immunoreactivity of human Purkinje cell membrane and C. elegans.

4. Discussion

We studied the immunohistochemical distribution of LMOD1 in human tissues and the nematode C. elegans. Our study was motivated by previous inconsistent results about the role of LMOD1 autoimmunity as a cause of NS. Specifically, it has been claimed that LMOD1 is an autoantigen in NS secondary to molecular mimicry initiated by infection with the nematode Onchocerca volvulus [9]. Despite this claim, LMOD1 has never been shown to exist in human neurons [13], but it has been observed in neurons created in vitro from iPSC [10]. To seek further clarity on this issue, we have now broadly studied the immunolocalization of LMOD1 in human tissues and the paradigm nematode, C. elegans.

The key observations in our study are: (a) lack of immunoreactivity of human neurons and C. elegans with an antibody to full-length LMOD1; (b) restricted immunoreactivity of human Purkinje cell membranes and C. elegans with an antibody to the N-terminus of LMOD1; (c) consistent immunoreactivity of smooth muscle cells (revealed by both antibodies) in a variety of human tissues, including blood vessels in the brain. Overall, our results indicate that LMOD1 is not generally expressed in neurons in the human brain, consistent with the work of Nanda et al. [13] in mice. However, LMOD1 was expressed widely in the brain within smooth muscle cells in blood vessels. This finding explains the detection of LMOD1 in immunoblots of brain homogenates [9].

Based on our findings and other reported results, LMOD1 is an unlikely target of autoimmune attack in NS. The reasoning for this view is twofold. First, neurons are not reservoirs of the protein. Second, if a pathogenic autoantibody to LMOD1 was implicated in NS, then the expected target would be vascular smooth muscle cells. This finding would be predicted to cause diffuse cerebral vasculitis, but cerebral vasculitis is not present NS. Therefore, the role of LMOD1 autoimmunity due to onchocerciasis is unsupported by both the lack of LMOD1 in neurons and its presence in blood vessels.

Our confirmation of the observation of LMOD1 in smooth muscle cells related to the presence of actin. Since LMOD1 is an actin binding and modulating protein [18], the presence of LMOD1 in cells with actin-myosin contractile elements is understandable. Another interesting observation is the immunostaining of Purkinje cell membranes with an antibody recognizing the N-terminus of LMOD1. This finding is like the membrane immunostaining reported in young iPSC-derived neurons in vitro [10]. However, it is unclear if the immunostaining in human cerebellum represents recognition of LMOD1 or cross-reactivity with a neuronal actin-binding protein. We trust in the latter, considering evidence demonstrating sequence similarities, specifically in the N-terminus, between LMOD2, a paralog of LMOD1, and related tropomodulin isoforms such as Ubiquitous-Tmods (TMOD3) and Neural-Tmods (TMOD2) [19]. Furthermore, a sequence homology search of human LMOD1 against proteins in C. elegans, as a proxy for Onchocerca volvulus, showed sequence overlap between LMOD1 and C. elegans tropomodulin in the N-terminus. Based on this finding, we conducted immunohistochemical analyses of C. elegans. In this, we performed immunostaining exclusively with the antibody recognizing the N-terminus of LMOD1. This revealed restricted immunoreactivity in the body wall of C. elegans, which contains actin-binding contractile proteins. Overall, these findings offer some evidence of cross-reactivity between LMOD1 and tropomodulin, which tends to reduce the likelihood of LMOD1's role as an autoantigen in NS. Although we acknowledge that cross reaction with tropomodulin cannot be confirmed with absolute certainty, based on the weight of the evidence presented, we are inclined to believe that tropomodulin cross reacts antibodies to LMOD1, thus explaining the immunohistochemical findings. However, it is also conceivable that LMOD1 has restricted expression in human Purkinje cells. This can be further evaluated in the future with additional analysis assays (e.g., immunoblotting and immunolocalization of tropomodulins).

An interesting set of observations in our study was the preferential immunolocalization of LMOD1 in some non-smooth muscle cells. The non-smooth muscle cells included the secretory cells of the zona glomerulosa and adrenal medulla. This likely reflects the role of the actin-dependent process in hormonal exocytosis/secretion. For example, annexin-1 is critical for the exocytosis of vesicles and likely mediates this effect via interaction with actin. Therefore, it is not surprising that we found that the aldosterone-secreting cells of the zona glomerulosa expressed LMOD1. Similarly, our observations of the presence of LMOD1 in chromatin cells of the adrenal medulla correlates with previously reported actin-dependent release of catecholamine from chromaffin granules [20].

In summary, we show that LMOD1 is predominantly present in the brain within smooth muscle cells in blood vessels, not neurons or glia. In addition, we suggest that the potential presence of LMOD1 antibody cross-reactivity with neuronal actin-binding protein(s) such as tropomodulin, could explain why immunostaining of human Purkinje cells and newly formed neurons in vitro [10], could have been misattributed to LMOD1. This data, paired with the inability to replicate the Johnson et al. results [9], do not corroborate that onchocerciasis causes NS by molecular mimicry involving LMOD1. Moreover, although our conclusions are based solely on our immunolocalization observations, the lack of widespread neuronal or glia immunostaining, with only restricted cerebellar immunostaining does not corroborate the LMOD1 autoimmunity hypothesis. Thus, if onchocerciasis is the cause of NS, it is not by direct infection of the brain or indirectly through an autoimmune reaction to LMOD1. On this basis, other hypotheses on the cause and pathogenesis of NS must now be investigated.

CRediT authorship contribution statement

Kenneth G. Kodja: Conceptualization, Methodology, Formal Analysis, Investigation, Writing-Original Draft, Writing-Review & Editing, Visualization. Sylvester Onzivua: Writing – Review & Editing. David L. Kitara: Writing – Review & Editing. Amanda Fong: Investigation. Patrick Kim: Visualization. Michael S. Pollanen: Conceptualization, Methodology, Formal Analysis, Investigation, Writing-Original Draft, Writing-Review & Editing, Supervision.

Funding

The Raymond Chang Foundation supported this work financially.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.