BTBD9 attenuates manganese-induced oxidative stress and neurotoxicity by regulating insulin growth factor signaling pathway

Abstract

Manganese (Mn) is an essential mineral, but excess exposure can cause dopaminergic neurotoxicity. Restless legs syndrome (RLS) is a common neurological disorder, but the etiology and pathology remain largely unknown. The purpose of this study was to identify the role of Mn in the regulation of an RLS genetic risk factor BTBD9, characterize the function of BTBD9 in Mn-induced oxidative stress and dopaminergic neuronal dysfunction. We found that human subjects with high blood Mn levels were associated with decreased BTBD9 mRNA levels, when compared with subjects with low blood Mn levels. In A549 cells, Mn exposure decreased BTBD9 protein levels. In Caenorhabditis elegans, loss of hpo-9 (BTBD9 homolog) resulted in more susceptibility to Mn-induced oxidative stress and mitochondrial dysfunction, as well as decreased dopamine levels and alternations of dopaminergic neuronal morphology and behavior. Overexpression of hpo-9 in mutant animals restored these defects and the protection was eliminated by mutation of the forkhead box O (FOXO). In addition, expression of hpo-9 upregulated FOXO protein levels and decreased protein kinase B levels. These results suggest that elevated Mn exposure might be an environmental risk factor for RLS. Furthermore, BTBD9 functions to alleviate Mn-induced oxidative stress and neurotoxicity via regulation of insulin/insulin-like growth factor signaling pathway.

Introduction

Restless legs syndrome (RLS) is a common neurological disorder estimated to affect ~10% of the US population, with symptoms ranging from mild to severe (1). RLS is characterized by involuntary twitching leg movements when muscles are not actively used, especially when asleep (2). When at rest, the patients tend to have feelings such as itching, prickling, crawling or pulling in their legs and these uncomfortable sensations can be relieved by walking, stretching or shaking legs. The symptoms usually create an overwhelming urge to move the legs during sleep, thus disrupting sleep quality. RLS is associated with symptoms such as sleep deprivation, anxiety, depression and attention-deficit/hyperactivity disorder (ADHD) (1). Moreover, RLS may induce serious consequences, including hypertension, heart disease and stroke (3). The etiology of RLS remains largely unknown, while the pathobiology has been linked to deficits in dopaminergic (DAergic) function and iron (Fe) deficiency in the brain (4,5). The DAergic agonists pramipexole and ropinirole have been approved to treat moderate-to-severe RLS. The marked therapeutic effects of these agonists strongly suggest DAergic system involvement in RLS pathogenesis (2). Magnetic resonance imaging (MRI) studies have shown mild presynaptic and postsynaptic deficits in nigrostriatal DAergic pathways in RLS patients (4). Furthermore, cerebrospinal fluid, MRI and autopsy studies have consistently revealed Fe deficiency in brains of both familial and non-familial forms of RLS, particularly in basal ganglia circuits (5). Fe supplementation has proven to be clinically effective in reducing RLS symptoms (6,7). RLS exhibits both familial and non-familiar (idiopathic) forms, with ~60% of cases having a family history of the disease (8). BTBD9 has been identified as one of the most prevalent genetic risk factors for RLS. Studies in mouse and fly models have also demonstrated a functional interaction between BTBD9, Fe homeostasis and DAergic activity (9,10). Btbd9 knockout mice show motor restlessness, sleep disruption, altered sensory perception and serum iron levels, as well as enhanced fear memory (9). In Drosophila melanogaster, loss of BTBD9 is associated with increased motor activity, sleep fragmentation and decreased dopamine (DA) levels (10).

Manganese (Mn) is an essential metal but excess exposure represents an occupational health hazard, leading to neurotoxicity. Mn is preferentially accumulated in subcortical structures of the basal ganglia, including the substantia nigra, globus pallidus and the caudate/putamen of the striatum (11). These brain areas are highly sensitive to oxidative stress, containing metabolically active cell types, particularly tonically active motor neurons that require high levels of adenosine triphosphate (ATP) for optimal function (12). High Mn levels in the blood plasma and cerebrospinal fluid are associated with subclinical parkinsonian movements and postural instability, increased risk for ADHD (13,14), Parkinson’s disease (PD) (15,16), Alzheimer’s disease (AD) (17), Huntington’s disease (18), amyotrophic lateral sclerosis (ALS) (19) and upon exposure to high levels of Mn, one may develop manganism (20). The etiology of these diseases is not fully understood, which probably involves a gene– environment interaction. However, these diseases share two common pathologies, oxidative stress and mitochondrial dysfunction, which are also the fundamental molecular mechanism of Mn-induced neurotoxicity.

Mn and Fe have similar chemical and physical characteristics. Mn shares numerous homeostatic and transport pathways with Fe, including the divalent metal transporter (DMT), the transferrin receptor system (TfR) and the Fe exporter, ferroportin (21). It has been demonstrated that Fe deficiency, a cardinal feature of RLS, is associated with increased brain deposition of Mn and upregulation of both Mn importers DMT and TfR (22,23). These findings indicate a disruption in metal homeostasis in RLS, Fe and Mn in particular. We hypothesize that the cell phenotype and ultimately the symptoms of RLS patients are the results of Fe deficiency or elevated concentrations of another metal that opportunistically increases when Fe levels are low. Given the established link between Fe and Mn biology, we deemed it worthwhile to evaluate whether systemic and/or neuronal alterations in Mn may contribute to the etiology of RLS.

Here we showed that the levels of BTBD9 were regulated by Mn in human cohort and cell culture. BTBD9 functioned to protect against Mn-induced toxicity, including oxidative stress, mitochondrial dysfunction and DAergic neurodegeneration in Caenorhabditis elegans. This protection is achieved by regulating the insulin/insulin-like growth factor (IGF) signaling pathway, a pathway known to be regulated by Mn (24–26). Our results illustrate an etiological role of Mn in RLS and provide novel translational insights into the biological basis of RLS.

Results

Mn exposure down-regulates BTBD9/hpo-9 levels

Given the link between Fe and Mn biology, we determined whether elevated Mn might regulate BTBD9 levels. Blood samples from a human cohort were collected and the levels of Mn and Fe were analyzed by inductively coupled plasma mass spectrometry (ICPMS). Twenty-four age- and sex-matched individuals were selected from the follow-up of Mn-exposed workers healthy cohort (MEWHC) in 2020 as previously described (27,28). These individuals were separated into two groups (12 each, including 8 women and 4 men) defined as Control-Low Mn (Blood Mn from 7.9–16.5 μg/l) and High Mn (Blood Mn from 28.8–55.2 μg/l) (Fig. 1A, Table 1). We found that individuals with higher whole blood Mn levels had significantly lower BTBD9 mRNA levels (~50%) (Fig. 1B). As the prevalence of RLS in women is much higher (almost double) than in men (29), we compared the sex-dependent effects on BTBD9 and Fe levels. In both men and women, high blood Mn led to lower BTBD9 levels (Fig. 1A and B). Notably, high Mn was associated with significantly lower Fe levels only in women, but not in men or combined groups (Fig. 1C), suggesting a sex-dependent pattern of Fe levels may exist. We also determined BTBD9 protein levels in adenocarcinomic human alveolar basal epithelial cells (A549 cells) exposed to MnCl₂ (24 h). Our results showed that BTBD9 protein levels were significantly decreased by Mn exposure in a dosage-dependent pattern, together with decreased viability (Fig. 1D–F). These results indicated that excessive Mn was able to downregulate BTBD9/hpo-9 expression at both transcriptional and translational levels.

Figure 1

Mn exposure downregulated BTBD9/hpo-9 levels. (A) Mn levels in the whole blood of the human cohort was quantified by ICPMS. (B) relative BTBD9 mRNA levels in the groups with low and high Mn levels. In each age- and sex-matched group, the mRNAs of the individual with high Mn were normalized to the individual with low Mn. (C) Fe levels in the whole blood of the human cohort. (D) BTBD9 protein levels in A549 cells by western blot after Mn exposure. (E) Quantitative BTBD9 protein levels normalized to GAPDH at 0 mm Mn. (F) cell survival rate measured by MTT assay. (G) hpo-9 mRNA levels in L1 larva and young adult nematodes by qPCR, without Mn exposure. hpo-9 levels in the tm3719 animals was normalized to the WT control worms. (B) hpo-9 mRNA levels in L1 larva of control and tm3719 animals exposed to Mn. All hpo-9 levels were normalized to the control animals at 0 mm of Mn. Data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test, except E and F by one way ANOVA (mean ± SD; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001; NS, non-significant. A–C, 8 females, 4 males, 12 in total; D–F, n = 3; G and H, n = 3).

Table 1

Mn and Fe levels in the blood of Age- and Sex- matched groups (F, female; M, male; Work Years, years working in the ferro-Mn alloy production plant)

Group	Sex	Age	Work years	Mn (μg/l)	Fe (mg/l)
1	F	49	31	10.1	462.1
	F	52	33	32.9	506.2
2	F	50	30	13.0	525.5
	F	50	34	35.4	372.2
3	F	49	24	9.5	418.7
	F	48	29	45.0	377.2
4	F	41	24	15.9	499.7
	F	44	25	34.4	381.2
5	F	45	25	15.9	565.9
	F	44	25	34.4	381.2
6	F	50	24	16.5	522.0
	F	46	26	35.2	318.5
7	F	48	29	12.8	483.0
	F	48	29	45.1	377.1
8	F	48	23	7.9	442.6
	F	46	26	35.2	318.5
9	M	52	34	12.5	610.2
	M	55	36	30.8	719.1
10	M	55	31	14.9	544.3
	M	57	36	28.8	648.9
11	M	47	23	14.7	645.7
	M	48	22	31.8	651.6
12	M	56	36	16.3	535.5
	M	56	36	55.2	283.7

Next, we validated the results in C. elegans

The nematode has a single BTBD9 mammalian homolog known as hpo-9 (Supplementary Material, Fig. S1A). A mutant allele tm3719 (hpo-9^{tm3719/tm3719}) results in an in-frame deletion (761 bp), which still produces a truncated hpo-9 mRNA (Supplementary Material, Fig. S1B). To detect the gene products of both wild-type (WT) and tm3719 allele (with deletion in Exon 2), a Taqman probe with Exon 4–5 boundary was selected for quantitative PCR (qPCR). Similar to the human study, Mn exposure resulted in a significant decrease (~40%) in hpo-9 mRNA in the WT animals, but not the mutants (Fig. 1G). Interestingly, we found a significant decrease (~50%) of hpo-9 mRNA levels in the mutant at both larva (L)1 and young adult stages (Fig. 1H), indicating the tm3719 allele is deficient in gene transcription. Thus, this mutant strain was selected for subsequent studies in the nematode.

Loss of hpo-9 renders worms more susceptible to Mn toxicity

Excessive Mn exposure can cause oxidative stress and finally lead to death in C. elegans. We found that tm3719 animals were more sensitive to MnCl₂ exposure (Fig. 2A), but not FeCl₂, CuCl₂ or ZnCl₂ when compared with the WT N2 control worms (Fig. 2B–D). However, we did not observe significant Mn accumulation in tm3719 worms (Fig. 2E), indicating HPO-9 either does not regulate total intracellular Mn content or it is expressed in a limited number of tissues and therefore does not affect overall internal Mn. To confirm the results from the mutant study, we knocked down hpo-9 using RNA interference (RNAi) feeding and found that these animals were also more susceptible to Mn exposure (Fig. 2F). qPCR confirmed that hpo-9 level was knocked down by ~40% (Fig. 2G).

Figure 2

hpo-9 mutants were more sensitive to MnCl₂-induced toxicity. (A–D) synchronized L1 animals were exposed to MnCl₂, FeCl₂, CuCl₂ and ZnCl₂, respectively. Survival rate was scored two days post exposure. (E) Mn levels were determined by ICPMS after exposure. (F) survival rates in young adult animals after RNAi feeding and Mn exposure. (G) hpo-9 mRNA levels were determined by qPCR. −, control bacteria carrying the L4440 empty vector; +, hpo-9 RNAi bacteria. (H) ROS levels by measuring DCFDA fluorescence intensity after Mn exposure. Fluorescence intensity was normalized to 0 mm Mn for each strain. (I) mitochondrial morphology in Mn-exposed young adult nematodes: top, intact; middle, intermediate and bottom, fragmented. (J) quantitative analysis of mitochondrial morphology by scoring worms with intact mitochondria. (A–F, H & J) Data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test, while data in (G) were analyzed by student’s t test (P < 0.0001 and P = 0.0007, respectively) (mean ± SD; A–E and H, n = 6; F, G and J, n = 3. ^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001).

Elevated reactive oxygen species and mitochondrial dysfunction in hpo-9 mutant

Mn exposure promotes generation of reactive oxygen species (ROS), which results in elevated oxidative stress and mitochondrial dysfunction (30–33). A fluorescent probe 2′,7′- DCFDA was used to determine ROS levels in the nematodes. We found that Mn exposure resulted in elevated ROS levels in a dose-dependent pattern in both control and the mutant animals; ROS levels were consistently higher in tm3719 mutants than the control animals (Fig. 2H).

As Mn is known to accumulate within mitochondria (34,35) and cause mitochondrial dysfunction associated cytotoxicity (30,32,36), next, we tested whether hpo-9 protects mitochondria from Mn exposure. The strain SD1347 with green fluorescent protein (GFP) labeled nucleus and mitochondria in the body wall muscle cells was crossed with the tm3719 mutant. We examined mitochondrial morphology in the nematode and found the mutant animals unexpectedly showed a slight but significant decrease (~15%) in intact mitochondria even without Mn exposure (Fig. 2I and J), indicating BTBD9 regulates mitochondrial homeostasis. After Mn exposure, the defect was further amplified (Fig. 2I and J). These results suggest that HPO-9 not only maintains mitochondrial morphology at normal physiological conditions but also protects against Mn-induced oxidative stress.

Expression pattern of hpo-9 in C. elegans

As RLS patients are associated with DAergic neurological dysfunction (10), next, we determined whether hpo-9 is expressed in DAergic neurons of the nematodes and functions to maintain normal DAergic activities. A transcriptional GFP reporter strain MAB336 (hpo-9p::GFP) was generated. We observed a ubiquitous GFP expression from L1 to adult stage, primarily in the pharynx and head neurons but weaker in the posterior intestine, body wall muscle cells and hypodermal cells (Fig. 3A and B). C. elegans hermaphrodites have 8 DAergic neurons, including 4 cephalic neurons (CEPs) and 2 anterior deirid neurons (ADEs) in the head (Fig. 3C and H) and 2 postdeirid neurons (PDEs) in the posterior. We crossed MAB336 with OH7193, which has mCherry labeled DAergic neurons and found that the GFP signal was co-localized with mCherry labeled DAergic neurons (Fig. 3D), indicating hpo-9 was expressed in DAergic neurons. Moreover, a translational reporter strain MAB335 was created and found that HPO-9 was localized in the cytoplasm (Fig. 3E–G), consistent with the previously findings in the human and Drosophila (10).

Figure 3

HPO-9 expression pattern and Mn-induced neuropathology in the nematode. (A–D) Expression pattern of hpo-9 using hpo-9 promoter-driven GFP reporter (hpo-9p::GFP). (A) hpo-9 expression in a Larva 3 animal. Scale bar, 100 μm. (B) The transcriptional reporter hpo-9p::GFP expression in the head. (C) mCherry labeled DAergic neurons. (D) Merged GFP and mCherry. Scale bar, 15 μm. (E–G) HPO-9 intracellular expression pattern using a translational GFP reporter (unc-54::HPO-9::GFP). (E) Intracellular expression of HPO-9::GFP in the body wall muscle cells. (F) DsRed protein in the cytoplasm. (G) Merged GFP and DsRed. Scale bar, 15 μm. (H–K) DAergic neurons in MAB300 worms exposed to 0 mm (H) or 10 mm MnCl₂ for 1 h (I–K). (H) Intact DAergic neurons in the head, including 4 CEPs and 2 ADEs. (I–K) Impaired DAergic neurons, arrowheads indicate damaged neuronal processes, asterisk indicates alternations in cell bodies. Scale bar, 10 μm. (L and M) DA staining by FIF in worms exposed to 0 mm (L) and 25 mm (M) MnCl₂. The dash square indicates the region with individual DAergic neurons for measuring fluorescence intensity. Scale bar, 10 μm. (N)Qquantitative analysis of worms with intact DAergic neurons after MnCl₂ exposure. Worms were recovered for 2 h before analysis. (N) Quantitative analysis of DA levels by measuring FIF. The fluorescence intensity was normalized to WT animals at 0 mm Mn. (O) Basal slowing response quantified by scoring body bends in 20 s. After Mn exposure, worms were grown to L4 stage for the behavioral assay. (N–P) Data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test (mean ± SD; ^**P < 0.01, ^***P < 0.001 and ^****P < 0.0001. L, n = 9; O, n = 3; P, n = 6).

Loss of hpo-9 resulted in susceptibility to Mn-induced neurotoxicity

Exposure of L1 worms to Mn induced morphological changes characteristic of neurodegeneration in the 6 head DAergic neurons (37) (Fig. 3I–K). Nematodes expressing dat-1p::GFP (MAB300) were crossed with tm3719 mutant. In the absence of Mn, we found no difference between the control and mutant worms in the morphology of DAergic neurons. However, tm3719 worms had a significantly lower percentage of morphologically normal DAergic neurons upon Mn exposure (Fig. 3N), indicating loss of HPO-9 leads to increased susceptibility to Mn-induced neurotoxicity. We also measured DA levels in these animals using formaldehyde-induced fluorescence (FIF) technique (Fig. 3L and M) (38,39). After Mn exposure, L1 animals were stained by FIF and the fluorescent images were taken by confocal microscopy. We found that Mn exposure decreased DA levels in both WT and tm3719 animals, while tm3719 worms showed significantly lower DA levels than WT worms (Fig. 3O). To determine whether the morphological changes were associated with DAergic dysfunction, we performed a behavioral assay (basal slowing response) which was dependent on DAergic neurons. We found Mn exposure caused a significant decrease in basal slowing response in tm3719 animals, compared with the control (Fig. 3P). These results showed that (1) under normal physiological conditions, loss of hpo-9 did not result in DAergic neuronal defect; and (2) the presence of Mn rendered tm3719 worms more susceptible to DAergic dysfunction than the control.

Expression of hpo-9 in the tm3719 animals restores the mutant defects

To further confirm the protection afforded by HPO-9 and to exclude the possibility that defects in tm3719 were caused by other mutations, we expressed WT hpo-9 gene in the tm3719 animals to study whether the defects could be rescued. The strain MAB415 (eft-3p::hpo-9::FLAG; hpo-9(tm3719)) was generated and expression of HPO-9 was confirmed by western blot (Supplementary Material, Fig. S2A). We found that transgenic over-expression of hpo-9 (OE) not only recovered the lethality defect in the mutant, but also led to greater resistance to Mn than in controls (Fig. 4A). Similarly, the ROS levels and mitochondrial morphology defects in m3719 animals were also restored by hpo-9 OE (Fig 4B and C). To validate HPO-9 function in the DAergic neurons, a strain MAB414 was generated with HPO-9 exclusively expressed in DAergic neurons (Supplementary Material, Fig. S2C). We found that hpo-9 OE animals had significantly higher percentage of intact DAergic neurons (Fig. 4D) and higher DA levels (Fig. 4E), compared with tm3719 mutants.

Figure 4

Expression of hpo-9/BTBD9 in the mutant animals restores the defects. (A and F) Survival rate of synchronized L1s after Mn exposure. (B) ROS levels measured by DCFDA fluorescent intensity after Mn exposure. Fluorescence intensity was normalized to 0 mm Mn for each strain. (C) Quantitative analysis of mitochondrial morphology by scoring worms with intact mitochondria. (D and G) Quantitative analysis of worms with intact DAergic neurons after Mn exposure. (E) DA levels measured by FIF. The fluorescence intensity was normalized to WT animals at 0 mm Mn. H, the expression pattern of GFP tagged hpo-9 and BTBD9 in Neuro2A cells. Green, GFP tagged proteins in the cytosol; blue, nucleus by DAPI staining. Scale bar, 15 μm. I, cell viability analyzed by MTT assay. All Data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test when main effects were observed (mean ± SD; n = 3. ^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001. A and F, n = 9; B, C, D, G and I, n = 6; E, n = 3).

Next, we determined whether the human homolog BTBD9 shared the same function as HPO-9. Similarly, we created MAB425 and MAB405 with BTBD9 expressed in the whole animal and DAergic neurons, respectively (Supplementary Material, Fig. S2B and D). We found that expression of BTBD9 restored m3719 defects in survival and DAergic neurons upon Mn exposure (Fig. 4F and G). Our results suggested HPO-9 and BTBD9 share the same function in protecting against Mn-induced lethality and neurotoxicity.

BTBD9/hpo-9 protects against Mn-induced toxicity in Neuro2A cells

Given the protective role of BTBD9/hpo-9 in C. elegans, we assumed that it has a similar function in mammalian systems. We expressed GFP tagged hpo-9 and BTBD9 in mouse neuroblastoma-Neuro2A cells. Both proteins were present in the cytoplasm (Fig. 4H), and expression of these significantly increased cell viability (Fig. 4I), consistent with the findings in C. elegans.

HPO-9 overexpression upregulates Fe levels

As hpo-9 re-expression attenuated Mn-induced toxicity, we first assumed that it might downregulate Mn levels in the nematode. Previously, we failed to find differences between the control and tm3719 animals, which might be due to the limited expression in worms (Fig. 2A). Now, with whole tissue expression of hpo-9 in MAB415 worms, we expected to see a significant decrease of internal Mn levels. However, this was not the case as the hpo-9 OE strain had similar internal Mn levels as the control and tm3719 animals (Fig. 5A). This result suggests that HPO-9 alleviates Mn-induced toxicity not by regulating Mn transport. Interestingly, we found that Fe levels were significantly elevated in hpo-9 OE worms, compared with both the control and tm3719 animals, regardless of Mn exposure (Fig. 5B). Next, we tested whether pre-incubation of Fe would protect worms upon Mn exposure. L1 animals were pre-incubated with FeCl₂, followed by Mn exposure. We found that low levels of FeCl₂ (0.0001 and 0.001 mm) significantly increased C. elegans survival rate upon later Mn exposure (Fig. 5C) in tm3719 animals. In contrast, Fe pre-incubation had no effect in WT control animals (Fig. 5D). These results were consistent with the facts that Fe levels are lower in the serum and brains of RLS patients, and corroborates the clinical study of Fe supplementation in reducing RLS symptoms (6,7).

Figure 5

Mn and Fe levels in treated animals and the protective effect of Fe supplementation in in mutant animals. (A and B) Mn and Fe concentrations were measured by ICP-MS, respectively. (C and D) survival rates of tm3719 (C) and WT (D) animals with Fe supplementation. Synchronized L1 animals were pre-incubated with FeCl₂ (0–0.01 mm) for 30 min. Fe was then washed and worms were exposed to Mn as described in Fig. 1A. (A and B) Data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test when main effects were observed; C and D, data were analyzed by one-way ANOVA (mean ± SD; ^*, P < 0.05; ^**, P < 0.01; A and B, n = 3; C and D, n = 9).

Protective effects of HPO-9 are dependent on FOXO/DAF-16

As HPO-9 does not regulate internal Mn concentrations, it is likely that HPO-9 functions to attenuate Mn-induced oxidative stress and mitochondrial dysfunction. It is known that Mn can act as a co-factor for the insulin receptor (IR) and the IGF receptor (IGFR) (24,25), and that the IGF signaling pathway is Mn-regulated (26). Therefore, we tested whether HPO-9 acts via DAF-16/FOXO signaling. To test this hypothesis, we first analyzed survival rate of the control (WT N2), mu86 mutant [CF1038, daf-16(mu86)] and daf-16 OE [TJ356, zIs356 [daf-16p::daf-16a/b::GFP + rol-6(su1006)]], in the presence of Mn exposure. We found that the mutant mu86 was more sensitive to Mn exposure; in contrast, daf-16 OE worms were more resistant, compared with the control (Fig. 6A). Next, we crossed mu86 mutant with MAB415 (hpo-9 OE) worms. We found that these animals became more sensitive to Mn-induced lethality, with similar survival rate to tm3719 mutants but significantly lower than hpo-9 OE animals (Fig. 6B), suggesting the protection of hpo-9 OE was completely eliminated by the mu86 mutant. Moreover, ROS levels in MAB456 animals were elevated analogous to tm3719 animals, compared with hpo-9 OE worms (Fig. 6C). Furthermore, in DAergic neurons, loss of daf-16 also diminished the protection of hpo-9 OE and resulted in more severe morphological damages (Fig. 6D). Taken together, our results indicated that DAF-16/FOXO is required for HPO-9 protection against Mn-induced toxicity in both the whole tissue and DAergic neurons.

Figure 6

HPO-9 function is dependent on FOXO/DAF-16. daf-16(mu86) is a loss of function mutant; daf-16 OE is a transgenic strain overexpressing daf-16. (A and B) Survival rate of L1 animals exposed to MnCl₂. (C) ROS levels by measuring DCFDA fluorescence intensity after Mn exposure. Fluorescence intensity was normalized to 0 mm Mn for each strain. (D) Quantitative analysis of worms with intact DAergic neurons. All Data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test when main effects were observed. (Mean ± SD; ^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001. A and D, n = 3; B and C, n = 9.).

HPO-9 down-regulates AKT level and elevates DAF-16 and downstream signaling

As HPO-9 is dependent on DAF-16, we assumed that it acts upstream of DAF-16. To study whether HPO-9 regulates DAF-16 directly, tm3719 and hpo-9 OE animals were crossed with TJ356 strain with a translational GFP reporter (DAF-16::GFP, Fig. 7A). We found that tm3719 animals (MAB447) had a slight but significant decrease in DAF-16::GFP level; surprisingly, DAF-16 level was increased to ~4 fold in hpo-9 OE animals (MAB444), when compared with the control (Fig. 7A and B). After Mn exposure, DAF-16 was significantly decreased in all three strains; however, the decrease in hpo-9 OE animals was subtle (Fig. 7A and B). Interestingly, there was no significant difference in DAF-16 levels between the control and tm3719 animals after Mn exposure. To further study whether DAF-16 is still able to activate downstream signaling upon Mn exposure, CF1553 strain (muls84[sod-3p::GFP + rol-6(su1006)]), with a transcriptional GFP reporter of sod-3, was crossed with hpo-9 OE animals (MAB445) (Fig. 7C). No significant difference was seen in sod-3p::GFP levels between WT and hpo-9 OE animals without Mn exposure; in the presence of Mn, sod-3 level was slightly decreased in WT animals; surprisingly, sod-3 was significantly increased in hpo-9 OE animals (Fig. 7C and D), indicating DAF-16 was still functional and was able to respond to oxidative stress caused by Mn. In contrast, no difference was seen in gst-4p::GFP, a downstream target of the nuclear factor erythroid 2–related factor 2 (Nrf2/SKN-1) signaling pathway (Supplementary Material, Fig. S3A and B).

Figure 7

HPO-9 regulates IGF signaling. (A) The translational reporter DAF-16::GFP in the WT, tm3719 or hpo-9 OE animals. (C) The transcriptional reporter sod-3p::GFP in the WT and hpo-9 OE animals. (A and C) Animals were synchronized at L1 stage and exposed to MnCl₂ (0, 25 and 50 mm) for 2 h, and recovered for 2 h before analysis. Scale bar, 50 μm. (B and D) quantitative analysis of GFP intensity. GFP intensity was normalized to the WT animals at 0 mm Mn. (E) Top, the translational GFP reporter AKT-1::GFP; (G) bottom, the translational GFP reporter PDK-1::GFP. Animals at larva stage 3 (L3) were analyzed. Scale bar, 50 μm. (F and G) Quantitative analysis of AKT-1::GFP (F) and PDK-1::GFP (G), respectively. GFP intensity was normalized to the WT animals at 0 mm Mn. H and I, quantitative analysis of DAF-16::GFP and AKT-1::GFP at Adult Day 3, after knocking down of hpo-9 by RNAi feeding. Bacteria carrying L4440 empty vector as the control. For each experiment, ~15 animals was analyzed. All data were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test when main effects were observed, except (F and G) data were analyzed by one-way ANOVA (mean ± SD; ^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001; ^****, P < 0.0001. B, D, F and G, n = 3; H and I, n = 9).

To further identify the HPO-9 substrate in IGF signaling, proteins acting upstream of DAF-16/FOXO were investigated. Translational reporter strains expressing GFP tagged AKT-1 [GR1672 (AKT-1::GFP)] and PDK-1 [GR1674 (PDK-1::GFP)] (Fig. 7E) were crossed with tm3719 and hpo-9 OE animals. We found that AKT-1 level was significantly decreased by ~50% in hpo-9 OE animals, but not in the tm3719 worms (Fig. 7E top and F). In contrast, PDK-1 level remained unchanged in both tm3719 and hpo-9 OE animals (Fig. 7E bottom and G). AKT is known to phosphorylate DAF-16/FOXO and facilitate its export and degradation in the cytosol by 14-3-3 proteins (40,41), consistent with our results that hpo-9 expression increases DAF-16 levels and its downstream target SOD-3. Therefore, AKT is likely a downstream target of hpo-9 in the IGF signaling pathway. To further validate these changes were due to hpo-9 OE, we knocked down hpo-9 by RNAi feeding. When hpo-9 was knocked down, DAF-16 levels were significantly decreased (Fig. 7H), and AKT levels were significantly increased (Fig. 7I) in both WT and hpo-9 OE animals, while no changes seen in tm3719 animals. Taken together, our results suggested that HPO-9 might specifically regulate FOXO/DAF-16 signaling, but not NRF-2/SKN-1 signaling. These data are consistent with a mechanism by which hpo-9 down-regulates AKT protein levels, in turn increasing FOXO levels. The activation of FOXO/DAF-16 downstream target genes offers protection against Mn-induced cellular stress and neurotoxicity.

Discussion

The pathology of RLS is linked to Fe deficiency and dysregulation of the DAergic system (4,42). However, the etiology of RLS remains unknown. It has been reported that the level of symptoms and age of onset in RLS patients vary remarkably despite high penetrance (43), suggesting that environmental risk factors could be the triggering cause. Our results suggest that Mn is a potential environmental risk factor for RLS. High levels of Mn negatively regulate BTBD9 levels in human whole blood and A549 cells. Loss of BTBD9 increased sensitivity to Mn-induced oxidative stress and DAergic neurological dysfunction. Interestingly, an RNAseq study in human neuroblastoma cells found a significant decrease in BTBD3 (a BTBD9 paralog) after Mn exposure (44). As the major risk genotypes of BTBD9 are located in the noncoding regions (45–47), it is plausible that those single nucleotide polymorphisms regulate transcription and alter its mRNA levels. Our study indicates that the impact of Mn on BTBD9 gene transcription may mimic that of the risk alleles. Meanwhile, high levels of blood Mn were associated with decreased Fe levels in females (but not in males), which might explain why RLS is more common in women than in men. However, direct evidence in people with excessive Mn levels and RLS symptoms is needed to evaluate whether Mn is a causative risk factor for RLS. Due to the limited sample size in our study cohort and other RLS cohorts (48), high blood Mn levels have not been reported yet in those diagnosed with RLS. Investigation of Mn levels in the blood plasma, cerebrospinal fluid and basal ganglia in a large sample size of RLS cohort will be important to determine the causative role of Mn in RLS. Meanwhile, diagnosis of RLS in Mn exposed human cohorts will also be needed to further reveal the impact of Mn in the progression of RLS.

IGF signaling is remarkably conserved through evolution. There is strong in vivo and in vitro evidence supporting a role for Mn-dependent regulation of the IGF signaling (24–26). For example, in vitro biochemical evidence indicates that IR and IGFR exhibit increased kinase activity in the presence of Mn versus Mn as the kinase co-factor (24,25). In mouse neuronal and human stem cell-based models, Mn-dependent insulin/IGF signaling has been shown to be dependent on bioavailable Mn acting directly at the level of the IR/IGFR (26).

Our present studies provide evidence of additional putative mechanisms by which Mn may regulate IGF signaling: Mn exposure (a) down-regulates BTBD9 levels; (b) resulting in lower FOXO levels; (c) which inhibits activation of downstream stress response genes. It is known that BTBD9 can function as an adaptor for the cullin-3 (Cul-3) class of E3 ubiquitin ligases (10,49). Our results showed that expression of BTBD9 decreases AKT-1 level by ~50% and dramatically increases FOXO level by ~3 fold. Thus, it is reasonable to assume that BTBD9 may promote protein degradation and/or dephosphorylation of AKT, which eventually upregulates FOXO level. We assume BTBD9 may have similar function as phosphatase and tensin homolog (PTEN) and the serine/threonine phosphatase 2 (PP2A) to counteract IGF signaling. In addition to FOXO, AKT also negatively regulates transcription factor EB (TFEB), a master regulator of autophagy and lysosome (50). It is also noteworthy that FOXO interacts with TFEB and regulates mitochondrial uncoupling proteins (51). Previously, we have shown that Mn exposure down-regulates nuclear localization of TFEB and suppresses autophagic-lysosomal degradation of unhealthy mitochondria (52). Here, loss of BTBD9 resulted in accumulation of abnormal mitochondria and it was worsened by Mn exposure, revealing the role of BTBD9 in energy production. DAergic neurons lacking BTBD9 are less capable to produce ATP and defend oxidative stress, and thus more susceptible to Mn-induced DAergic insults, which might contribute to the etiology of RLS. Interestingly, a relationship between RLS and PD has been suggested and RLS is often perceived as one subgroup of PD symptoms. For example, MRI studies have shown mild presynaptic and postsynaptic deficits in nigrostriatal DAergic pathways in RLS patients without PD (4,42). Increased prevalence of PD is noted in RLS, and vice versa (53). Given the function of BTBD9 on ATP production and oxidative stress defense in DAergic neurons, it is possible that loss of BTBD9 might contribute to DAergic neurodegeneration and worsen the progression of PD. Further research is needed to determine the role of BTBD9 in PD. Together, our study sheds new light on the links between Mn, BTBD9, FOXO and mitochondrial homeostasis. Last, but not least, as BTBD9 is able to regulate AKT, we expect that this protein might have a boarder function in addition to mitigating Mn-induced oxidative stress. However, further studies are needed to investigate the specific interaction between BTBD9 and Akt/Akt-regulated signaling pathways and confirm the role of BTBD9 in more general oxidative stress.

Conclusions

BTBD9, as one of the most prevalent genetic risk factors for RLS, is widely expressed in the central nervous system (54,55). However, its molecular function, especially its role in the etiology of RLS, remains largely unknown. Our results establish the links between BTBD9, Mn neurotoxicity, RLS and IGF signaling pathway for first time. Elevated Mn exposure down-regulates BTBD9 mRNA and protein expression, which further diminishes cells’ ability to defend Mn induced oxidative stress and mitochondrial dysfunction. Loss of BTBD9 in animals increases their susceptibility to Mn secondary to DAergic dysfunction, suggesting that excessive Mn levels may play a role in the etiology of RLS. Our results also show that BTBD9 might be a novel component of the IGF signaling pathway, negatively regulating AKT levels. Given its function in cellular stress and mitochondria, BTBD9 might have protective roles in other neurological disorders, such as PD, AD and ALS to name a few. Further research is directed to verify the causative role of Mn in RLS and confirm BTBD9 function in other diseases.

Materials and Methods

Plasmid constructs

hpo-9 coding sequence (~2.1 kb, with C-terminal FLAG tag) and hpo-9 promoter sequence (~1.5 kb upstream of the start codon) was PCR amplified using from genomic DNA isolated from WT N2 worms. The human BTBD9 complementary DNAs (cDNAs; with C-terminal FLAG tag) were amplified from cDNAs provided by M. Di. Neely (Vanderbilt University, Nashville, TN). The GFP tag was fused to hpo-9 and BTBD9 C-terminal by fusion PCR. The transcriptional GFP reporter hpo-9p::GFP::unc-54 3′untranslated region (UTR) was created by fusing hpo-9 promoter sequence, GFP coding sequence and unc-54 3′UTR together. Using Gateway recombinational cloning (Invitrogen), the above PCR products were recombined with the pDONR221 vector to create pENTRY clones. For somatic expression in worms, hpo-9 and BTBD9 pENTRY constructs were then recombined into pDEST-eft-3 vector, under the promoter from eukaryotic translation elongation factor (eft-3) gene. For expression in DAergic neurons, the pENTRY constructs were recombined into pDEST-dat-1 vector, under the promoter from the DA transporter (dat-1) gene. For body wall muscle expression, HPO-9::GFP pENTRY constructs were then recombined into pDEST-unc-54 vector, under the promoter from a muscle myosin class II heavy chain (unc-54) gene. These plasmids were then used to create transgenic worms. For expression in Nuro2A cells, hpo-9 and BTBD9 (without stop codon) pENTRY constructs were recombined into pcDNA-DEST47 vector for later transformation.

C. elegans strains and culture

Nematodes were grown and maintained using standard procedures (56). The strain MAB200 carrying a hpo-9 mutant (allele tm3719 with 761 bp deletion) was ordered from National BioResource Project: C. elegans in Japan. Strains CF1038, CF1553, CL2166, GR1672, GR1674, OH7193, SD1347 and TJ356 were ordered from C. elegans Genetic Center (CGC). Transgenic worms were created using microinjection as previously described (57). Plasmids were injected at a concentration of 50 mg/ml. For whole worm expression, eft-3::hpo-9 or BTBD9 alone was co-injected into tm3719 strain with pG2M36 (myo-3::dsRed) and pBCN27-R4R3 (rpl-28::PuroR, Addgene), which were used as the selective markers for transformation. For expression in DAergic neurons, dat-1p::hpo-9 or BTBD9 alone were co-injected into MAB300 [dat-1::GFP(vtIs1) V, smf-2(gk133) X] strain with elt-2::mCherry and pBCN27-R4R3. The hpo-9 transcriptional GFP reporter was generated by injecting hpo-9p::GFP::unc-54 3′UTR, pRF4 and pBCN27-R4R3 into OH7193 (otIs181 [dat-1p::mCherry; ttx-3p::mCherry], him-8(e1489) IV) worms. For the translational reporter, unc-54p::HPO-9::GFP was co-injected with pG2M36 (myo-3::dsRed) into MAB200 worms. For each injection mixture, at least three stable lines were generated and analyzed. Representative lines were selectively integrated by using UV irradiation using a Spectroline UV crosslinker with an energy setting of 500 mJ/cm2. All strains and genotypes are listed in Supplementary Material, Table S1.

Metal exposure

Worms were synchronized at L1 or L4 stages as per experimental needs. MnCl₂, FeCl₂, CuCl₂ and ZnCl₂ exposures were performed as previously described (37,58).

Lethality analysis

This assay was performed as previously described (37). Experiments were performed in at least three independent replicates.

RNA interference in C. elegans

RNAi by feeding was performed as previously described (59). Worms were first synchronized at L1 stage and then grown on RNAi plates with corresponding RNAi bacteria until they reached desired stages for later analysis. The bacterial clone producing dsRNA for hpo-9 was obtained from the Ahringer RNAi library (Source BioScience). The control bacteria contained the empty RNAi expression vector pL4440.

Measurement of intracellular ROS

After Mn exposure, L1 nematodes were washed and resuspended in M9 buffer. ~500 L1s (in 50 μl M9) were then transferred to 96-well plates, with 50 μl of 50 μm fluorescent probe 2′,7′- DCFDA (at a final concentration of 25 μm). Worms were gently shaken for 2 h, then the fluorescence intensity was quantified by a plate reader (FLUOstar OPTIMA, BMG LABTECH) at an excitation wavelength of 490 nm and an emission wavelength of 510–570 nm.

Mitochondrial morphology

Nematodes with mitochondria-localized GFP in body wall muscle were used to assess alterations in mitochondrial morphology. Synchronized L4 stage larve were treated with MnCl₂ for 2 h, followed by 2-h recovery. Mitochondrial morphology was assessed in at least 20 animals for each condition and blindly scored. The morphological categories of mitochondria were defined according to Momma, Homma, Isaka, Sudevan and Higashitani (60) with slight modification: (1) tubular: a majority of mitochondria were interconnected and elongated like tube shape; (2) intermediate: a combination of interconnected and fragmented mitochondria; (3) fragmented: a majority of round or short mitochondria in the image taken.

DAergic neurodegeneration assay

This assay was performed as previously described (37). For each experiment, at least 20 animals were quantitated in three independent replicates. The scoring was performed by the same investigator blind to the nematode genotypes and treatments.

Confocal microscopy

This assay was performed as previously described (61) using a Nikon CSU-W1 Spinning Disk confocal microscope. Worms expressing DAF-16::GFP and sod-3p::GFP were analyzed at L1 stage; worms expressing AKT-1::GFP and PDK-1::GFP were analyzed at L3 stage. RNAi animals were analyzed at Day 3 adult stage.

Dopamine staining

DA)levels in the worms were measured using FIF technique, as previously described (38,39). After Mn exposure and recovery, five L1 stage worms were placed on a slide in a drop of 5 μl PFA solution (4% paraformaldehyde in 0.1 M Na₂HPO₄/KH₂PO₄ buffer, pH 7.2) and incubated at room temperature for 5 min. Excess liquid was wicked off and the slide was heated for 10 min on an aluminum block at 100°C. After cooled to room temperature, a drop of glycerol and a coverslip were placed on top of the treated worms. Fluorescence was observed with a Nikon CSU-W1 Spinning Disk confocal microscope filter set (excitation, 395–440 nm; emission, 470 nm long pass).

Basal slowing response

This behavioral assay was performed as previously described (62), except that the worms were exposed to Mn at L1 stage and assayed at L4/young adult stage.

Quantitative real time PCR

qPCR was applied to determine the expression levels of target genes using TaqMan Gene Expression Assays (Applied Biosystems). The gene ama-1 (Ce02462726_m1), encoding an RNA polymerase II large subunit was used as the control. The hpo-9 (Ce02477190_g1) probe with Exon 4–5 boundary was able to detect the mRNAs from both the WT N2 and the mutant tm3719 (deletion in Exon 2) animals.

Semi-quantitative RT-PCR

Semi-quantitative RT-PCR was performed on worms to validate the expression of hpo-9(tm3719) allele and transgenes. RNA isolation was performed as described (59). To detect the truncated tm3719 allele, primers 5′-ATGAGCGATAACCATGCTTTTGG-3′ and 5′-TTATTTTATGGCAATTGGAACGTTTG-3′ are used. To detect FLAG tagged transgenes (hpo-9 and BTBD9) in DAergic neurons, the following primer pairs were used: for hpo-9::FLAG, 5′-ATGAGCGATAACCATGCTTTTGG-3′ and 5′-TCACTTGTCATCGTCGTCCTTGTAGTC-3′; for BTBD9::FLAG, 5′-ATGAGTAACAGCCACCCTCTTC-3′ and 5′-TCACTTGTCATCGTCGTCCTTGTAGTC-3′. For the ama-1 (RNA polymerase II large subunit) loading control, the forward primer was 5′-GCTACTCTGGCAAGACGTG-3′ and the reverse primer was 5′-CGAGCGCATCGATGACCC-3′.

Human blood studies

Individuals were randomly selected from the follow-up of MEWHC in 2020, who were recruited from one large ferro-Mn alloy production plant in Guangxi, China, as previously described (27,28). The present study was approved by the Medical Ethics Committee of Guangxi Medical University and a proof/certificate of approval is available upon request. Whole blood samples were collected in glass tubes containing ethylenediamine tetra acetic acid anticoagulant (Junnuo, China) from the participants after overnight fasting, and stored at −80°C until RNA extraction and metal analysis. Concentrations of Mn and Fe were quantified by ICPMS (Thermo Scientific) as previous described (63), except for the following details. Briefly, diluent containing 0.1% nitric acid (Fisher Scientific, Canada), 0.5% 1-Butanol (Acros, US) and 0.1% Triton ™ X-100 (Sigma, GB) was used to dilute whole blood samples. The Trace Elements Whole Blood (Sero, L-1, L-2 and L-3, Norway) and SRM1640a (Trace Elements in Natural Water, Nist, US) were used as reference materials, and the recovery rate was 79.68–88.20%. In the present study, all of the samples were above the limit of detection or Mn (2.37 ng/L) and for Fe (0.36 μg/L). Total RNA was extracted from 250 μl of whole blood using TRIzol LS Reagent (Invitrogen, USA). A total of 300 ng of RNA was used to obtain cDNA by reverse transcription, as described in our previous report (64). Real-time qPCR used 2 × RealStar Green Fast Mixture (GenStar, China) with a Real-time PCR Detection System (LightCycler96, Roche, Switzerland). Briefly, the amplification of primers was as follows: denaturation at 95°C for 2 min, followed by 40 cycles at 95°C for 15 s, 60°C for 30 s and 72°C for 30 s. The following primers were used: BTBD9: forward 5′-AGCCTGCCTCCTTCATCCGTATC-3′ and reverse 5′-CTCTGCTGCTCTGGACACTCAAAG-3′; Actin: forward 5′-CATGTACGTTGCTATCCAGGC-3′ and reverse 5′-CTCCTTAATGTCACGCACGAT-3′.

Cell culture

A549 and Neuro2a cells were maintained in an incubator with the supplementation of 5% CO₂ at 37°C. The normal culture medium is Dulbecco’s modified Eagle medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS, Gibco, USA), 10 U/ml penicillin and 100 μg/ml streptomycin. For transformation, neuro2a cells were placed in a 6-well plate and incubated in DMEM with 10% FBS until they reached to ≥80% of confluency. Cells were then transfected with pGFP, pGFP-HPO-9 and pGFP-BTBD9 encoding plasmid using TransIT^®-LT1 according to the manufacturer’s protocol, respectively. The cells were incubated with transfection complex for 24 h at incubator with the supplementation of 5% CO₂ at 37°C. After transfection, the cells were placed in 96-well plates and treated with 0, 0.125, 0.25 and 0.5 mm/l MnCl₂ for 24 h, respectively. Cellular viability was measured using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) tetrazolium reduction assay. The medium was removed and washes twice with PBS. 10 μl 5 mg/ml MTT solution was added to DMEM (100 μl per well), and incubated at 37°C for 4 h. Later, the medium was removed and 150 μl DMSO was added to each well. The absorbance was measured at 490 nm with a plate reader (FLUOstar OPTIMA, BMG LABTECH).

Statistical analysis

Statistical analyses were conducted in GraphPad Prism version 8 (GraphPad Software Inc.), and results were expressed as mean ± standard deviation (SD). One-way analysis of variance (ANOVA) were performed to analyze data with one factor. Data with two factors were analyzed by two-way ANOVA followed by Sidak’s multiple comparisons test when main effects were observed. Differences were considered statistically significant if P-value was <0.05.

Acknowledgements

We thank the patients and families for their participation in this study. We thank the Caenorhabditis Genetics Center (CGC) and Shohei Mitani for providing the C. elegans strains. We also thank Dr M. Diana Neely for the human cDNAs.

Conflict of Interest statement. We declare no conflicts of interest.

Funding

This work was supported by National Institute of Environmental Health Sciences (R21ES031315, R01ES10563 and R01ES07331); the German Research Foundation (DFG) for the financial support of the DFG Research Unit TraceAge (FOR 2558); National Natural Science Foundation of China (81860573 and 82073504) and Guangxi Natural Science Foundation for Innovation Research Team (2017GXNSFGA198003).