Colocalization of TDP‐43 and stress granules at the early stage of TDP‐43 aggregation in amyotrophic lateral sclerosis

Abstract

TDP‐43 aggregates (skeins and round inclusions [RIs]) are frequent histopathological features of amyotrophic lateral sclerosis (ALS). We have shown that diffuse punctate cytoplasmic staining (DPCS) is the earliest pathologic manifestation of TDP‐43 in ALS, corresponding to nonfibrillar TDP‐43 located in the rough endoplasmic reticulum. Previous in vitro studies have suggested that TDP‐43 inclusions may be derived from stress granules (SGs). Therefore, we investigated the involvement of SGs in the formation of TDP‐43 inclusions. Formalin‐fixed spinal cords of six ALS patients with a disease duration of less than 1 year (short duration), eight patients with a disease duration of 2–5 years (standard duration), and five normal controls were subjected to histopathological examination using antibodies against an SG marker, HuR. In normal controls, the cytoplasm of anterior horn cells was diffusely HuR‐positive. In short‐duration and standard‐duration ALS, the number of HuR‐positive anterior horn cells was significantly decreased relative to the controls. DPCS and RIs were more frequent in short‐duration ALS than in standard‐duration ALS. The majority of DPCS areas and a small proportion of RIs, but not skeins, were positive for HuR. Immunoelectron microscopy showed that ribosome‐like granular structures in DPCS areas and RIs were labeled with anti‐HuR, whereas skeins were not. These findings suggest that colocalization of TDP‐43 and SGs occurs at the early stage of TDP‐43 aggregation.

Immunostaining of three serial sections of a spinal anterior horn cell containing a round inclusion (RI) (arrows). The RI is immunopositive for HuR (G), nTDP‐43 (H) and pTDP‐43 (I). In particular, the RI is encapsulated by HuR (G). Bar = 10 μm.

1. INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is characterized pathologically by the loss of upper and lower motor neurons with the frequent occurrence (>95%) of transactivation response DNA‐binding protein 43 (TDP‐43) inclusions. TDP‐43 is an RNA‐binding protein that plays multiple roles in RNA metabolism and is predominantly localized in the nucleus under normal conditions. This protein is also present in the cytoplasm and shuttles between the nucleoplasm and the cytoplasm. In patients with ALS, TDP‐43 becomes abnormally phosphorylated (pTDP‐43) in the cytoplasm, leading to conformational changes that eventually lead to the formation of TDP‐43 inclusions. A growing body of evidence suggests that nuclear‐to‐cytoplasmic mislocalization of physiological TDP‐43 or abnormal aggregation of pTDP‐43 induces toxicity through both loss‐ and gain‐of‐function mechanisms. Recently, Kon et al. [1] reported that DPCS is the earliest manifestation of TDP‐43 pathology in ALS and that DPCS corresponds to nonfibrillar TDP‐43 located at the ribosomes of the rough endoplasmic reticulum (ER).

Previous studies using cultured cells have suggested that TDP‐43 inclusions may be derived from degenerated stress granules (SGs) based on the presence of numerous SG markers in the inclusions [2]. SGs are dynamic, reversible biomolecular condensates, which assemble in the cytoplasm of eukaryotic cells under various stress conditions including oxidative, osmotic and heat stress, UV radiation, and viral infections [3, 4, 5]. They range in size from 200 to 1000 nm and are composed mainly of SG‐marker proteins and non‐translating mRNAs [6]. Formation of SGs typically occurs upon stress‐induced translational arrest and polysome disassembly [7]. SGs were first defined as cytoplasmic foci containing polyadenylated RNA, small ribosomal sub‐units, translation initiation factors (eIF3, eIF4E, and eIF4G), and RNA‐binding proteins such as TIA‐1, HuR, PABP, and TTP that form following eIF2α phosphorylation [8]. These proteins have been considered as SG markers.

Several investigators have demonstrated colocalization of TDP‐43 with SG proteins in the brains of patients with ALS, raising the question of whether TDP‐43 pathology truly arises from SGs [6, 9]. Recently, Diaz‐Garcia et al. reported that TDP‐43 aggregates were occasionally encapsulated by HuR in motor neurons of patients with ALS [10]. In the present study, therefore, we investigated the involvement of SGs in the formation of TDP‐43 inclusions in anterior horn cells (AHCs) in ALS using immunohistochemistry and immunoelectron microscopy with antibodies against an SG marker, HuR.

2. MATERIALS AND METHODS

2.1. Subjects

This study included 14 cases of sporadic ALS and five normal controls (Table 1). All the patients in the ALS cohort had been diagnosed, followed up, and died at Hirosaki University Hospital or Aomori Prefectural Central Hospital. Pathological diagnosis of ALS was confirmed by the loss of upper and lower motor neurons and the presence of pTDP‐43‐immunoreactive neuronal cytoplasmic inclusions (NCIs). Considering the natural history of ALS [11], we defined ALS cases with a disease duration of ≤1 year as short‐duration ALS (n = 6) and those with a disease duration of 2–5 years as standard‐duration ALS (n = 8). The five control subjects were matched with the patients for age and sex (Table 1). The clinical and neuropathological findings in one of the ALS cases (case 4) have been reported previously [12].

TABLE 1.

List of subjects.

No.	Age at death (years)	Sex	Diagnosis	Disease duration (months)	Nishihira type	No. of remaining AHCs	Area of remaining AHCs (μm2)
1	61	M	ALS‐D	5	2	87	664
2	65	M	ALS	6	2	77	841
3	42	M	ALS	7	1	75	661
4	68	M	ALS + CIDP	11	1	58	1048
5	76	F	ALS	12	2	76	531
6	77	M	ALS	12	1	35	638
Average	64.8			8.8		68.0	730.5
7	59	M	ALS‐D	24	2	66	1267
8	63	M	ALS	24	1	47	1187
9	54	M	ALS	34	1	50	677
10	46	M	ALS‐D	48	2	64	520
11	80	F	ALS	53	1	30	869
12	73	F	ALS	30	1	22	860
13	67	F	ALS	36	1	13	1034
14	69	F	ALS	46	1	30	864
Average	63.9			36.9		40.2	909.8
15	64	M	Control	NA	NA	90	1139
16	72	M	Control	NA	NA	76	1641
17	40	M	Control	NA	NA	88	1050
18	84	M	Control	NA	NA	77	1269
19	83	F	Control	NA	NA	63	1176
Average	68.6					78.8	1255.0

Abbreviations: AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; ALS‐D, ALS with dementia; CIDP, chronic inflammatory demyelinating polyneuropathy; NA, not applicable.

2.2. Morphological analysis

We divided TDP‐43‐immunoreactive NCIs into three types according to Kon et al. [1]: (1) DPCS, fine punctate granules scattered diffusely in the cytoplasm, (2) round inclusions (RIs), large spherical inclusions about 1–15 μm in diameter, and (3) skein‐like inclusions (SLIs), aggregates of thread‐like structures. When DPCS areas were accompanied by SLIs and RIs, we classified them as SLIs and RI, respectively.

2.3. Immunohistochemistry and morphometry

Three sections (4 μm thick) were serially cut from paraffin blocks of the fourth lumbar segment of the spinal cord in all cases and pretreated for heat‐induced epitope retrieval for 10 min in 10 mmol/L citrate buffer (pH 6.0) using an autoclave. The first section was immunostained with rabbit anti‐pTDP‐43 (pSer409/410; Cosmo Bio Co., Ltd.; 1:5000), the second section with mouse anti‐HuR (sc‐5261; Santa Cruz Biotechnology; 1:50), and the third section with rabbit anti‐native TDP‐43 (nTDP‐43) (10782‐1‐AP; Proteintech; 1:2000), using the avidin‐biotin‐peroxidase complex method with diaminobenzidine as the chromogen.

In a preliminary study, we attempted immunostaining in our postmortem samples (ALS and control cases) using antibodies against SG marker proteins [9] including anti‐EIF2C1 (LS‐C99286, LifeSpan Biosciences), anti‐EIF3eta (sc‐16,377, Santa Cruz Biotechnology), anti‐EIF4G1 (HPA028487, SIGMA), anti‐G3BP1 (HPA004052, Sigma‐Aldrich), anti‐HuR (sc‐5261, Santa Cruz Biotechnology; #11910‐1‐AP, Proteintech), anti‐PABP1 (#4992, Cell Signaling Technology), anti‐TIA‐1 (sc‐1751, Santa Cruz Biotechnology), anti‐TIAR (610352, BD Biosciences). In ALS and control cases, cytoplasm and/or Nissl bodies were immunostained with most antibodies. However, ALS‐specific TDP‐43 inclusions (DPCS and round) were immunostained with anti‐HuR antibodies (sc‐5261, Santa Cruz Biotechnology; #11910‐1‐AP, Proteintech) moderately and anti‐EIF4G1 (HPA028487, SIGMA) slightly. Since Diaz‐Garcia et al. reported that TDP‐43 aggregates were occasionally encapsulated by HuR in motor neurons of ALS patients [10], we adopted HuR immunoreactivity as an SG marker.

Digital images of the anterior horn on both sides were captured by a digital camera (Provis AX‐70, Olympus). AHCs were defined as cells with a somal diameter of >37 μm containing Nissl bodies and nucleoli in Rexed laminae VIII and IX [13]. AHCs were numbered on enlarged prints (×140). When the same neuronal cell body was recognized on three contiguous sections, the same number was marked on the neuron to determine the total number of AHCs and the area of the neuronal cell bodies. HuR immunoreactivity of the neuronal cytoplasm and NCIs was also evaluated in four groups: AHCs with DPCS, RIs or SLIs, or AHCs without inclusions. To confirm the HuR immunoreactivity in DPCS areas and RIs, selected sections were also immunolabeled with another HuR‐antibody (#11910‐1‐AP, Proteintech) that had been used in a recent study by Diaz‐Garcia et al. [10]. ImageJ/Fiji 1.46 software was used to calculate the area and number of AHCs in digital images. All specimens were coded and morphometric analyses were performed by two authors (Fumiaki Mori and Hina Yasui) without knowledge of the cases.

2.4. Double‐labeling immunofluorescence

Paraffin sections from the lumbar cord from short‐duration ALS and standard‐duration ALS were processed for double‐label immunofluorescence. Deparaffinized sections were blocked with donkey serum and then incubated overnight at 4°C with a mixture of mouse anti‐HuR (sc‐5261; Santa Cruz Biotechnology; 1:20) and polyclonal anti‐nTDP‐43 (10782‐1‐AP; Protein Tech Group, Inc.; 1:200) or rabbit anti‐pTDP‐43 (pSer409/410; Cosmo Bio Co., Ltd.; 1:500) antibodies. The sections were then rinsed and incubated for 1 h at 38°C with a mixture of anti‐mouse IgG tagged with Alexa Fluor 488 (Invitrogen; 1:200) and anti‐rabbit IgG tagged with Alexa Fluor 594 (Invitrogen; 1:200). The sections were then mounted with Vectashield (Vector) and examined with an Olympus Provis AX‐70 fluorescence microscope (Olympus). The proportion of HuR‐positive inclusions (DPCS, RIs, or SLIs) relative to the total number of inclusions positive for pTDP‐43 was calculated in each case.

2.5. Immunoelectron microscopy

About 50‐μm‐thick vibratome sections of the lumbar cord from two cases of short‐duration ALS (cases 1 and 5) were incubated with a mouse anti‐HuR antibody (sc‐5261; Santa Cruz Biotechnology; 1:20) and a rabbit polyclonal anti‐pTDP‐43 antibody (pSer409/410; Cosmo Bio Co., Ltd.; 1:500) for 2 days at 4°C, followed by incubation with a 1.4‐nm gold‐coupled Fab' fragment of goat anti‐mouse IgG (Nanoprobes; 1:50) or a 1.4‐nm gold‐coupled Fab' fragment of goat anti‐rabbit IgG (Nanoprobes; 1:50). The sections were visualized using a silver‐enhancing kit (BB International), then treated with osmium tetroxide (1% in 0.1 M phosphate buffer), block‐stained with uranyl acetate, dehydrated in graded ethanol series and flat‐embedded on glass slides in Poly/Bed 812 Resin (Polysciences Inc.). Regions containing target structures (DPCS, RIs, or SLIs) were cut away, glued on a flat‐surfaced plastic block, and then cut into serial 0.5‐μm‐thick semithin sections. The sections were stained with toluidine blue to identify the immunolabeled target structures. After the targets had been identified, subsequent blocks were cut into 90‐nm‐thick ultrathin sections, which were then stained with uranyl acetate and lead citrate and viewed with a JEOL1230 transmission electron microscope (JEOL Ltd.).

2.6. Statistical analysis

One‐way analysis of variance with post hoc analysis by Scheffe's F test was used to examine differences in the number and area of AHCs in patients with ALS and control subjects and differences in the proportion of HuR‐immunoreactive inclusions (DPCS, RIs, and SLIs). Calculations were performed using Statcel software (OMS Publishing). Statistical significance was set at a p value of <0.05.

2.7. Ethical approval

This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee of Hirosaki University Graduate School of Medicine, Hirosaki, Japan. Written informed consent was obtained from all of the patients' families.

3. RESULTS

3.1. Clinical and histopathological findings

Clinical and histopathological findings are shown in Table 1. Short‐duration ALS, standard‐duration ALS, and the controls had a similar age at death. Disease duration was significantly shorter in short‐duration ALS (8.8 months) than in standard‐duration ALS (36.9 months) (p < 0.001). In comparison with controls, the number and area of AHCs were decreased in short‐duration and standard‐duration ALS (Table 1). DPCS and RIs were more frequent in short‐duration ALS than in standard‐duration ALS (Table 2). There was no significant difference in the number of SLIs between standard‐duration and short‐duration ALS.

TABLE 2.

Incidence of TDP‐43 inclusions in short‐duration and standard‐duration ALS cases.

	DPCS	RI	SLI	No inclusion	Total AHCs
Short‐duration ALS	24.5*	9*	2.5	32.0	68.0
	(36.0%)	(13.2%)*	(3.7%)	(47.1%)	(100%)
Standard‐duration ALS	7.6	2.6	3.4	26.6	40.2
	(18.9%)	(6.5%)	(8.4%)	(66.1%)	(100%)

Abbreviations: AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; DPCS, diffuse punctate cytoplasmic staining; RI, round inclusion; SLI, skein‐like inclusion.

^* p < 0.05.

3.2. HuR immunoreactivity of AHCs

In controls, the majority of AHCs showed moderate to weak HuR immunoreactivity in the cytoplasm and nucleus (Figure 1A). In short‐duration ALS, AHCs were also immunoreactive for HuR, and a small proportion of AHCs showed no HuR immunoreactivity (Figure 1B, arrows). In standard‐duration ALS, immunoreactivity for HuR was severely decreased in AHCs (Figure 1C, arrows).

FIGURE 1.

Immunoreactivity for HuR in the anterior horn of the lumbar spinal cord from normal controls (A) and patients with short‐duration ALS (B) and standard‐duration ALS (C). (A) In controls, the majority of anterior horn cells (AHCs) show moderate immunoreactivity for HuR in the cytoplasm and nucleus. (B) In short‐duration ALS, AHCs occasionally show no HuR immunoreactivity (arrows). (C) In standard‐duration ALS, the immunoreactivity is severely decreased in AHCs (arrows). Bar = 100 μm. The number of HuR‐positive AHCs (D), the somatic area of HuR‐positive AHCs (E), and the proportions of HuR‐positive AHCs in relation to the total number of AHCs (F) in control cases (n = 5), short‐duration ALS (n = 6) and standard‐duration ALS (n = 8). ALS, amyotrophic lateral sclerosis. *p < 0.05.

Relative to controls (70.0), the number of HuR‐immunoreactive AHCs was significantly decreased in standard‐duration ALS (19.5; p < 0.05), but not in short‐duration ALS (44.8) (Figure 1D). The area of HuR‐immunoreactive AHCs was significantly decreased in short‐duration ALS (783.1 μm²; p < 0.05) and standard‐duration ALS (858.5 μm²; p < 0.05) relative to controls (1339.8 μm²) (Figure 1E). The proportion of HuR‐immunoreactive AHCs relative to the total number of AHCs was decreased in standard‐duration ALS (45.1%; p < 0.05) and short‐duration ALS (61.9%, not significant) relative to controls (89.1%) (Figure 1F).

3.3. HuR immunoreactivity of AHCs with and without TDP‐43 inclusions

In ALS, AHCs were divided into four groups: AHCs without inclusions (Figure 2A–C), AHCs with DPCS (Figure 2D–F), RIs (Figure 2G–I), or SLIs (Figure 2J–L). The cytoplasm in the majority of AHCs without inclusions was immunopositive for HuR (Figure 2A), but not for nTDP‐43 (Figure 2B) or pTDP‐43 (Figure 2C). The nucleus was also immunopositive for HuR (Figure 2A) and nTDP‐43 (Figure 2B), but not for pTDP‐43 (Figure 2C). HuR immunoreactivity in the cytoplasm and nucleus of DPCS‐containing and RI‐containing AHCs was relatively preserved (Figure 2D–I). The cytoplasm and nucleus of SLI‐bearing AHCs were not immunolabeled with anti‐HuR (Figure 2J–L). Some RIs (Figure 2G–I), but not SLIs (Figure 2J–L), were immunopositive for HuR.

FIGURE 2.

Immunoreactivity for HuR (A, D, G, J), nTDP‐43 (B, E, H, K), and pTDP‐43 (C, F, I, L) in serial sections of AHCs in ALS. (A–C) A normal‐looking AHC showing immunopositivity for HuR (A), but not for nTDP‐43 (B) or pTDP‐43 (C), in the cytoplasm. The nucleus is diffusely immunopositive for HuR (A) and nTDP‐43 (B). (D–F) A slightly atrophic AHC containing diffuse punctate cytoplasmic staining. The cytoplasm and nucleus are immunopositive for HuR (D), but not for nTDP‐43 (E) or pTDP‐43 (F). (G–I) A slightly atrophic AHC containing a round inclusion (RI) (arrows). The cytoplasm and nucleus are immunopositive for HuR (G), but not for nTDP‐43 (H) or pTDP‐43 (I). The RI is also immunopositive for HuR (G). (J–L) A severely atrophic AHC containing skein‐like inclusions (SLIs) (arrowheads). The cytoplasm and nucleus are immunonegative for HuR (J), nTDP‐43 (K), and pTDP‐43 (L). The SLIs are also immunonegative for HuR (J). AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis. Bars = 10 μm.

In short‐duration ALS, the proportion of HuR‐positive AHCs without inclusions relative to the total number of AHCs was 73.0%. The proportion of HuR‐positive AHCs with DPCS or RIs relative to the total number of AHCs with each inclusion was 54.6% or 64.2%, respectively. All of the SLI‐containing AHCs were immunonegative for HuR (Figure 3A). In standard‐duration ALS, the proportion of HuR‐positive AHCs without inclusions relative to the total number of AHCs was 45.2%. The proportion of HuR‐positive AHCs with DPCS or RIs relative to the total number of AHCs with each inclusion was 59.8% or 43.4%, respectively. All of the SLI‐containing AHCs were immunonegative for HuR (Figure 3B).

FIGURE 3.

Cytoplasmic HuR‐immunoreactivity of AHCs without inclusions and with DPCS, RI, and SLI in short‐duration (A) and standard‐duration ALS (B). (A) In short‐duration ALS, the proportion of HuR‐positive AHCs without inclusions relative to the total number of AHCs was 73.0%. The proportion of HuR‐positive AHCs with DPCS or RIs relative to the total number of AHCs with each inclusion was 54.6% or 64.2%, respectively. All of the SLI‐containing AHCs are immunonegative for HuR. (B) In standard‐duration ALS, the proportion of HuR‐positive AHCs without inclusions relative to the total number of AHCs was 45.2%. The proportion of HuR‐positive AHCs with DPCS or RIs relative to the total number of AHCs with each inclusion was 59.8% or 43.3%, respectively. All of the SLI‐containing AHCs were immunonegative for HuR. AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; DPCS, diffuse punctate cytoplasmic staining; RI, round inclusion; SLI, skein‐like inclusion. *p < 0.05; **p < 0.01.

3.4. HuR immunoreactivity of TDP‐43 inclusions

Double‐labeling immunofluorescence revealed that HuR‐immunoreactive AHCs in ALS had slight to moderate immunoreactivity for nTDP‐43 in the nucleus (Figure 4A–C). pTDP‐43‐immunoreactive DPCS (Figure 4D–F) and RIs (Figure 4G–I), but not SLIs (Figure 4J–L), were immunopositive for HuR.

FIGURE 4.

Double‐labeling immunofluorescence for HuR (A, D, G, J), nTDP‐43 (B), or pTDP‐43 (E, H, K) in AHCs in ALS. HuR appears green, nTDP‐43 appears red, and overlap of HuR and nTDP‐43 or pTDP‐43 (C, F, I, L) appears yellow. (A–C) The cytoplasm and nuclei of AHCs without inclusions are diffusely immunopositive for HuR and nTDP‐43. (D–E) DPCS is immunopositive for HuR and pTDP‐43. (G–I) An RI (arrows) is immunopositive for HuR and pTDP‐43. (J–L) SLIs (arrowheads) are immunonegative for HuR. AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; DPCS, diffuse punctate cytoplasmic staining; Lf, lipofuscin granules; Nu, nucleus; RI, round inclusion. Bars = 10 μm.

In short‐duration ALS, the proportion of AHCs with HuR‐positive DPCS or RIs relative to the total number of AHCs with each inclusion was 54.6% or 21.1%, respectively (Figure 5A). All of the SLIs were immunonegative for HuR. In standard‐duration ALS, the proportion of AHCs with HuR‐positive DPCS or RIs relative to the total number of AHCs with each inclusion was 59.8% or 10.0%, respectively (Figure 5B). All of the SLIs were immunonegative for HuR.

FIGURE 5.

HuR immunoreactivity of DPCS, RI, and SLI in short‐duration (A) and standard‐duration ALS (B). (A) In short‐duration ALS, the proportion of AHCs with HuR‐positive DPCS relative to the total number of AHCs with DPCS (54.6%) was significantly higher than that of AHCs with HuR‐positive RIs (21.1%) and SLIs (0%). (B) In standard‐duration ALS, the proportion of AHCs with HuR‐positive DPCS or RIs relative to the total number of AHCs with each inclusion was 59.8% or 10.0%, respectively. All of the SLIs were immunonegative for HuR. AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; DPCS, diffuse punctate cytoplasmic staining; RI, round inclusion; SLI, skein‐like inclusion. *p < 0.05; **p < 0.01.

3.5. Immunoelectron microscopy

RIs were often located between accumulated lipofuscin granules and chromatolytic areas containing dispersed rough ER (Figure 6A, Figures S1A and S2A). The inclusions consisted of accumulated ribosome‐like structures (Figure 6B, Figures S1B and S2B) containing autolysosomes and autophagosomes (Figure 6B) and some early stage Bunina bodies (Figure 6C, Figure S1B). Gold labeling of HuR was evident in the ribosome‐like structures (Figure 6D) attaching to and detaching from the rough ER in and around the inclusions. Gold labeling of pTDP‐43 was also evident in the ribosome‐like structures (Figure S1C) attaching to and detaching from the rough ER in and around the inclusions. Gold labeling of pTDP‐43 was also found in ribosome‐like structures arrayed in a line (Figures S1D and S2C). Bundled filamentous structures corresponding to SLIs were evident near the RIs (Figure S2C,D) and among the DPCS areas (Figure 7A,B). In DPCS‐containing AHCs, HuR immunoproducts were localized in the ribosomes attached to and detached from the rough ER or fragmented Nissl bodies (white arrows, Figure 7C). Notably, HuR immunoproducts were not evident in SLIs (black arrow, Figure 7D).

FIGURE 6.

Immunoelectron microscopy using anti‐HuR antibody in ALS. (A) An AHC containing an RI (arrow). (B) A higher‐magnification view of the area indicated by the arrow in (A) showing gold particles in an accumulation of ribosome‐like structures containing autolysosomes (white arrowheads). (C) A higher‐magnification view of the area indicated by the asterisk in (B) showing accumulated ribosome‐like structures containing early stage Bunina bodies (white arrows). (D) A higher‐magnification view of the area indicated by the star in (C) showing ribosome‐like granular structures immunolabeled with gold (black arrowheads). Filamentous structures are not evident in this RI. AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; RI, round inclusion. Bars: A = 5 μm; B, C = 1 μm; D = 0.2 μm.

FIGURE 7.

Immunoelectron microscopy using anti‐HuR antibody in ALS. (A) An AHC containing diffuse punctate cytoplasmic staining (DPCS) and skein‐like inclusions (SLIs, black arrows). (B) A higher‐magnification view of the area indicated by the asterisk in (A) showing DPCS (fragmented Nissl bodies) and SLIs (black arrows). (C) A higher‐magnification view of the area indicated by the black star in (B) showing fragmented Nissl bodies labeled by HuR immunoproducts (white arrows). (D) A higher‐magnification view of the area indicated by the white star in (C) showing ribosome‐like granular structures labeled with HuR immunoproducts (black arrowheads). HuR immunoproducts are not evident in SLIs (black arrow). AHC, anterior horn cell; ALS, amyotrophic lateral sclerosis; RI, round inclusion. Bars: A = 5 μm; B, C = 1 μm; D = 0.2 μm.

4. DISCUSSION

4.1. Significance of decreased HuR immunoreactivity in AHCs

In the present study, the proportion of HuR‐positive AHCs relative to the total number of AHCs was decreased in both short‐duration and standard‐duration ALS relative to controls. Many reports have shown that HuR plays an important role in mRNA stabilization, regulation of TDP‐43 and FUS proteins, and protection against neurotoxicity [14, 15, 16]. In HuR‐knockout mice, neuronal HuR deficiency resulted in redistribution of TDP‐43 to cytosolic granules, which is associated with motor neuron degeneration [17]. Although we cannot exclude the possibility that the decreased HuR immunoreactivity of AHCs in ALS is a secondary phenomenon caused by motor neuron degeneration, this reduction of HuR in the cytoplasm may accelerate the aggregation of TDP‐43 in AHCs.

4.2. TDP‐43 inclusions are derived from TDP‐43 bound to SGs

DPCS is the earliest manifestation of TDP‐43 pathology in ALS [1]. In the present study, DPCS and RIs were more frequent in short‐duration ALS than in standard‐duration ALS. The majority of DPCS areas and a small proportion of RIs, but not SLIs, were immunopositive for the SG marker, HuR. These findings suggest that TDP‐43 temporarily colocalizes with HuR, especially in the early phase of TDP‐43 inclusion formation. In addition, immunoelectron microscopy showed that the ribosome‐like structures comprising DPCS and RIs were immunolabeled with anti‐HuR antibody, suggesting that the ribosome‐like structures appear to be, at least in part, SGs. SGs are known to be dynamic entities that can enlarge and coalesce under prolonged stress [18]. Our histopathological findings are in line with in vitro studies showing that after overnight treatment of HeLa cells with paraquat, TDP‐43 colocalized to SGs together with HuR [19].

4.3. Implications of nuclear depletion of TDP‐43 during DPCS formation

In the present study, nTDP‐43 was decreased or depleted from the nucleus of AHCs containing DPCS, which is considered to represent the early phase of TDP‐43 inclusions [1]. The colocalization of TDP‐43 inclusions and HuR suggests that TDP‐43 may be trapped in SGs from the early phase of TDP‐43 accumulation. Koyama et al. have reported that in ALS motor neurons, especially neurons with mislocalized TDP‐43, the amount of TARDBP mRNA is increased in the cytoplasm, suggesting that the absence of nuclear TDP‐43 induces abnormal autoregulation and increases the amount of TARDBP mRNA. This vicious circle may accelerate the progression of ALS [20]. During inclusion body formation, sequestration of essential proteins such as TDP‐43 in SGs can directly impair their function, including the suppression of cryptic exon skipping, and cause neuronal degeneration [21, 22, 23].

4.4. Neuroprotection by HuR‐positive SGs?

HuR plays important roles in stabilizing mRNA, regulating TDP‐43 and FUS proteins, and providing protection against neurotoxicity [15, 16]. Lu et al. suggested that there is a balance between the positive effects of HuR and the negative autoregulatory effects of TDP‐43 to control the expression of TDP‐43 [15]. This balance could be perturbed with HuR overexpression, knockdown, or inhibition, resulting in upregulation or downregulation of TDP‐43. Considering that the majority of DPCS areas and a small proportion of RIs are immunopositive for HuR, the expression of HuR in these inclusions may imply a neuroprotective mechanism. Further studies will be necessary to clarify the neuroprotective roles of HuR in ALS.

4.5. Putative process of TDP‐43 inclusion formation

We have previously considered the maturation processes of TDP‐43‐immunoreactive SLIs and RIs in the spinal anterior horn cells of patients with ALS [24]. The RIs appeared to arise from dot‐like inclusions and small RIs, which were occasionally located in the perinuclear region (Nissl bodies were relatively well preserved). Recently, we demonstrated that autophagy‐related proteins and structures were localized in and around TDP‐43‐positive SLIs and RIs [25]. Moreover, we showed that pTDP‐43‐immunoreactive DPCS is the initial manifestation of TDP‐43 pathology in early stage ALS, and that pTDP‐43 accumulation may initiate from the rough ER in a nonfibrillar state [1]. In the present study, DPCS and RIs were more frequent in short‐duration ALS than in standard‐duration ALS. Nissl bodies were preserved in AHCs with DPCS and/or RIs, and the majority of DPCS areas and a small proportion of RIs were immunoreactive for HuR. However, SLI‐bearing AHCs showed no HuR immunoreactivity. Immunoelectron microscopy revealed that the bundled filamentous structures corresponding to SLIs were located in the vicinity of the RIs and DPCS. On the basis of the presence of SG markers within TDP‐43‐positive inclusions, Aulas et al. hypothesized that pathological TDP‐43/FUS‐containing inclusions originate from SGs that have failed to disassemble [2]. These findings indicate that TDP‐43 may initially accumulate in the ribosome‐like structures (SGs?) attached to or detached from the ER as DPCS, and then form small to large RIs. Both DPCS and RIs might change to SLIs via the autophagy‐lysosome pathway.

In conclusion, we have demonstrated that TDP‐43 aggregation starts with HuR‐positive SGs in association with nuclear depletion of nTDP‐43 in ALS. Therapeutic targeting of this initiation point might hold promise for prevention of ALS.

AUTHOR CONTRIBUTIONS

Fumiaki Mori: Study concept and design, drafting of the manuscript, acquisition and analysis of data (neuropathological examination). Hina Yasui: Acquisition and analysis of data (neuropathological examination). Yasuo Miki, Tomoya Kon, Akira Arai and Masahiko Tomiyama: Acquisition and analysis of clinical data. Hidekachi Kurotaki and Koichi Wakabayashi: Study concept and design, and study supervision. All authors read and approved the final version of the manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

FIGURE S1. Immunoelectron microscopy using anti‐pTDP‐43 antibody in ALS. (A) An AHC containing an RI (arrow). (B) A higher‐magnification view of the area indicated by the arrow in (A) showing gold particles in an accumulation of ribosome‐like structures containing an early stage Bunina body (white arrow). (C) A higher‐magnification view of the area indicated by the star in (B) showing rough endoplasmic reticulum and detached ribosome‐like structures labeled with gold (black arrowheads). (D) A higher‐magnification view of the area indicated by the asterisk in (B) demonstrating that ribosome‐like structures labeled with gold are arrayed in a line (white arrowheads), but filamentous structures are not evident. Bars: A = 5 μm; B = 1 μm; C, D = 0.2 μm.

FIGURE S2. Immunoelectron microscopy using anti‐pTDP‐43 antibody in ALS. (A) An AHC containing an RI (arrow). (B) A higher‐magnification view of the area indicated by the arrow in (A) showing accumulation of ribosome‐like structures and granulofilamentous structures labeled with gold. (C) A higher‐magnification view of the area indicated by the asterisk in (B) showing a bundle of filaments (star), ribosome‐like structures arrayed in a line (white arrowheads), and autophagosomes (black arrowheads). (D) A higher‐magnification view of the area indicated by the star in (C) showing a bundle of filaments corresponding to SLIs (white arrows). Bars: A = 5 μm; B, C = 1 μm; D = 0.2 μm.

Mori F, Yasui H, Miki Y, Kon T, Arai A, Kurotaki H, et al. Colocalization of TDP‐43 and stress granules at the early stage of TDP‐43 aggregation in amyotrophic lateral sclerosis. Brain Pathology. 2024;34(2):e13215. 10.1111/bpa.13215

DATA AVAILABILITY STATEMENT

The data sets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.