[¹⁸F]EKZ-001 is a positron emission tomography (PET) tracer targeting histone deacetylase 6 (HDAC6), an enzyme responsible for intracellular transport and clearance of misfolded proteins. HDAC6 modulation is a promising treatment strategy in neurodegenerative disorders, including amyotrophic lateral sclerosis (ALS). Apart from motor symptoms, people with ALS (pwALS) can show a variable degree of cognitive impairment as part of the ALS-frontotemporal spectrum disorder (ALS-FTSD). This work assessed [¹⁸F]EKZ-001 binding in pwALS with variable involvement of FTSD.

Methods

Twenty-four pwALS (13M/11F, 61 ± 10 years) and 12 healthy controls (HC) (6M/6F, 58 ± 3 years) were included. Thirteen pwALS were cognitively normal (ALS-CN), and eleven pwALS presented with FTSD (ALS-FTSD) ranging from mild cognitive or behavioral impairment to FTD, according to their performance on the Edinburgh cognitive and behavioral ALS screen (ECAS). All subjects underwent dynamic PET-MR imaging with arterial sampling, and regional distribution volumes (V_T) were calculated using a Logan graphical analysis.

Results

[¹⁸F]EKZ-001 V_T was significantly lower in pwALS compared to HC. For ALS-CN, the largest reduction was found in the brainstem. For ALS-FTSD, reductions were more widespread in both gray and white matter. No differences in V_T were found between pwALS with and without a C9orf72 mutation. [¹⁸F]EKZ-001 V_T was not correlated with ECAS scores, age, or disease duration.

Interpretation

[¹⁸F]EKZ-001 binding is lower throughout the brain in pwALS compared to HC. This may be related to a compensatory mechanism to repair intracellular transport defects in ALS or to reduced HDAC6 enzyme availability for [¹⁸F]EKZ-001 binding due to sequestration of HDAC6 within protein aggregates.

1 Introduction

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive neurodegenerative disorder affecting both upper and lower motor neurons. ALS also affects cognition, with up to 50% of people with ALS (pwALS) developing mild–severe cognitive or behavioral impairment, and concomitant frontotemporal degeneration (FTD) occurring in about 10% of pwALS [1].

Histone deacetylase 6 (HDAC6) is a class IIb HDAC paralog predominantly present in the cell cytoplasm. HDAC6 deacetylates α-tubulin, impacting microtubule dynamics and cell motility, and multiple additional cytoplasmic deacetylation substrates, such as cortactin, tau, and HSP90 [2-5]. HDAC6 also facilitates clearance of misfolded proteins through its ubiquitin binding domain by promoting aggresomal formation and autolysosome maturation. Consequently, HDAC6 is integral in protective responses to the aggregation of cytotoxic proteins and has also been implicated in the pathogenesis of neurodegenerative disorders such as Alzheimer's and Parkinson's disease, where α-tubulin acetylation is decreased [6-8].

The role of HDAC6 in ALS is complex and discordant findings have been reported. HDAC6 mRNA is upregulated in fruit flies overexpressing human TAR DNA-binding protein 43 (TDP-43) mutations [9], and HDAC6 expression is reduced in fruit flies and in human cellular systems following loss of TDP-43 function [10]. Furthermore, in ALS induced pluripotent stem cell (iPSC)-derived motor neurons and skin fibroblasts, HDAC6 inhibition improved axonal transport defects, rescued defects in intracellular mitochondria distribution, reduced TDP-43 pathology, and enhanced neuromuscular junction formation [11-14]. In SOD1^G93A mouse studies, HDAC6 deletion improved compound muscle action potential and neuromuscular junction innervation [15], and enhanced accumulation of mutant SOD1 [16], with conflicting results on motor neuron count and disease progression. Another study in SOD1^G93A mice reported that decreased HDAC6 protein levels in the spinal cord were related to disease progression. When HDAC6 was overexpressed in SOD1^G93 mice that exhibited decreased HDAC6 expression, motor neuron loss was ameliorated concomitant with increased clearance of toxic SOD1 protein [17]. These results align with observations of increased TDP-43 levels and SOD1 aggregation in cultured cells following HDAC6 knockdown [18, 19]. Taken together, these studies emphasize the importance of HDAC6 for toxic protein clearance in ALS, and support selective HDAC6 pharmacologic inhibition to rescue axonal transport defects, but also highlight the importance of maintaining HDAC6 protein clearance activity.

Quantifying HDAC6 in the human brain is therefore of great interest to investigate disease pathogenesis, progression, and possible treatment response in ALS. [¹⁸F]EKZ-001 ([¹⁸F]Bavarostat) is a PET tracer with specific affinity for HDAC6. [¹⁸F]EKZ-001 has favorable kinetic properties for in vivo evaluation of HDAC6 binding [20, 21] and can be quantified by Logan graphical analysis (LGA) modeling with appropriate test–retest reproducibility in healthy volunteers [22]. The aim of this study was to assess [¹⁸F]EKZ-001 binding in the brain of pwALS with variable involvement of frontotemporal spectrum disorder (FTSD) compared to HC using PET-MR. Additionally, the relationship of [¹⁸F]EKZ-001 binding with C9orf72 mutation status, as it is the most common genetic mutation in ALS, was explored as well as the correlation with neuropsychological assessment, age, and disease duration.

2 Methods

2.1 Study Participants

pwALS were prospectively recruited at the Motor Neuron Disorder Clinic of the neurology department at the University Hospital Leuven between 2019 and 2024. Eligible patients were diagnosed with “probable, laboratory supported ALS,” “probable ALS,” “possible ALS,” and “suspected ALS” according to the El Escorial criteria [23]. Main exclusion criteria for pwALS consisted of other major internal pathology (e.g., cancer) and contra-indications for MRI or for PET (e.g., pregnancy). As part of the routine clinical work-up, all pwALS underwent genetic testing for C9orf72, TARDBP, GRN and FUS. The revised ALS functional rating scale (ALSFRS-R) and the Edinburgh cognitive and behavioral ALS screen (ECAS), evaluated by an experienced clinical neuropsychologist, were acquired in all pwALS. pwALS were classified as ALS-CN or ALS-FTSD [24] according to their performance on the ECAS and behavioral assessment [25]. Cerebrospinal fluid (CSF) neurofilament light chain (NfL) levels were determined for a subset of pwALS either as part of the routine clinical work-up or through participation in other research studies. No other biomarkers (e.g., Alzheimer's disease biomarkers) were determined. HC were recruited as previously reported [22].

2.2 Data Acquisition

[¹⁸F]EKZ-001 PET-MR acquisitions were performed on a 3T GE Signa PET-MR scanner (General Electric, Milwaukee, MI, USA). PET scans were performed in list mode and were reconstructed using ordered subset expectation maximization (28 subsets, 6 iterations) with isotropic Gaussian post-smoothing (4 mm full width at half maximum [FWHM]), including corrections for decay, scatter, deadtime, and randoms and time-of-flight detector response. Zero echo time MR-based attenuation correction was applied [26].

HC fasted for at least 4 h prior to the PET scan and underwent a full 120-min dynamic [¹⁸F]EKZ-001 acquisition with arterial blood activity and radiometabolite sampling. Coffee-break imaging protocols (CBP) were installed and validated to improve feasibility of [¹⁸F]EKZ-001 PET in pwALS [22]. PwALS did not fast prior to the PET scan. Ten pwALS underwent a coffee-break PET acquisition consisting of a 60-min dynamic acquisition followed by a 30-min scan from 90 to 120 min post injection (coffee-break protocol 1 [CBP1]). This was extended to a longer coffee-break protocol for some pwALS, with an initial 30-min dynamic scan followed by a second 30-min scan from 90 to 120 min (coffee-break protocol 2 [CBP2]). Quantitative validation of the different coffee-break protocols in healthy volunteers was performed (Data S1). The average bias introduced by CBP2 compared to the full dynamic acquisition ranged from 1.7% in the occipital cortex to 2.6% in the white matter (Data S1). Arterial sampling and radiometabolite analysis were performed according to the same protocol as for HC.

3D T1-weighted and fluid-attenuated inversion recovery (FLAIR) MRI scans were acquired using a vendor-supplied 8-channel brain-phased array head coil on the same PET-MR scanner. FLAIR images were used to determine Fazekas scores by a nuclear medicine specialist in training (C.C.).

2.3 Image Processing

Reconstructed [¹⁸F]EKZ-001 images were corrected for motion using a rigid frame-by-frame co-registration in PMOD v4.1 (PMOD technologies, Zurich, Switzerland). Next, PET images were co-registered to the structural T1-weighted MRI scans. MRI segmentation was performed by the CAT12 toolbox of SPM12 (Statistical Parametric Mapping, Welcome Trust Centre for Neuroimaging, University College, London, UK) [27]. Ten composite bilateral volumes-of-interest (VOIs) were defined from the Neuromorphometrics atlas: frontal cortex (excluding primary motor cortex), parietal cortex, occipital cortex, temporal cortex, primary motor cortex, striatum, thalamus, brainstem, cerebellum, and cerebral white matter. The LGA algorithm from the PKIN tool (PMOD v4.1) was used to calculate V_T for all VOIs, as was previously validated [22]. The time t* in an LGA can be described as the time after which the plotted samples can be fitted by a linear regression and no further significant increase in the slope is observed. To enable direct comparisons between all groups, regardless of the imaging protocol, t* was set to 26 min for all subjects since this was the latest available timepoint before the coffee break for subjects who were scanned with CBP2. For group comparisons, the PXMOD tool (PMOD v4.1) was used to construct parametric V_T images based on the same LGA. Parametric V_T maps were analyzed both with and without partial volume correction (PVC), using a region-based voxel-wise (RBV) algorithm, as described previously [28] with PET resolution modeling with a Gaussian kernel of 5 mm FWHM. Data were also processed using the two-tissue compartment model, as validated previously [22], to obtain VOI-based K₁ values.

Gray matter probability maps for voxel-based morphometry (VBM) and parametric V_T images (both with and without PVC) were normalized to MNI space using the CAT12 toolbox of SPM12 and were smoothed by an isotropic Gaussian kernel of 8 mm FWHM.

2.4 Statistical Analysis

General statistical analyses were performed in GraphPad Prism v10 (GraphPad Software, La Jolla, CA). Normality was assessed using Shapiro–Wilk tests. VBM and voxel-based group comparisons were performed using an unpaired t-test in SPM12 (cluster height p_FWE < 0.05, peak height p_uncorrected < 0.001). For VOI-based comparisons, when comparing more than two groups, ANOVA tests were performed with post hoc two-sample t-tests. For these post hoc two-sample t-tests, p-values lower than 0.005 were considered significant (Bonferroni correction for 10 VOIs). To assess the influence of age on group differences in [¹⁸F]EKZ-001 V_T, linear regression models were tested per VOI where [¹⁸F]EKZ-001 V_T was assessed as the dependent variable and group (HC/ALS-CN/ALS-FTSD) and age as independent variables. To assess the influence of white matter lesions on [¹⁸F]EKZ-001 V_T in the ALS group specifically, linear regression models were tested where [¹⁸F]EKZ-001 V_T was assessed as the dependent variable and Fazekas (0/1/2/3) as an independent variable.

Associations of [¹⁸F]EKZ-001 V_T PVC with scores of the neuropsychological assessment (ECAS), CSF NfL levels, age, and disease duration were assessed through Pearson or Spearman correlation analyses as appropriate. ECAS cognitive subdomains include an ALS-specific part (language, verbal fluency, and executive domains) and an ALS non-specific part (memory and visuospatial domains). For correlation analyses, residual scores for each cognitive domain were used. These residual scores were calculated using predicted values for each participant based on age, level of education, and sex [25]. pwALS for whom there was more than 3 months between ECAS assessment and the PET scan were excluded from the correlation analyses (17/24 included). Due to the exploratory character of the correlation analyses, no correction for multiple comparisons was performed.

3 Results

3.1 Demographics

Demographic variables are summarized in Table 1. A total cohort of 24 pwALS (ALS-T) (13M/11F, mean 61 ± 10 y/o, range 44–80) was included. This ALS-T cohort consisted of two people with “probable ALS, laboratory supported,” six people with “probable ALS,” 14 people with “possible ALS” and two people with “suspected ALS” at time of diagnosis, which were later confirmed with laboratory supported ALS. The mean disease duration at the time of scanning was 2.2 ± 2.2 years (range 0.7–7.9). Thirteen pwALS were categorized as ALS-CN (7M/6F, mean 57 ± 9 y/o, range 44–80) and 11 as ALS-FTSD (6M/5F, mean 65 ± 9 y/o, range 51–80). Of the ALS-FTSD cohort, three met the diagnostic criteria for concomitant FTD [24], five had a cognitive impairment (CI), one presented with isolated behavioral impairment (BI), and two showed both cognitive and behavioral impairment (CBI). ALSFRS-R was acquired with an average of 58 ± 53 days (range 1–166) between ALSFRS-R and scan acquisition. ECAS was acquired in 21 out of 24 pwALS with a mean of 43 ± 49 days (range 0–176) between ECAS and scan acquisition. ALS-FTSD had significantly lower ECAS scores than ALS-CN (mean difference of 21 ± 6.0, p < 0.01). Compared to ALS-CN, ALS-FTSD had a longer disease duration (mean difference of 2.2 ± 0.8 years, p < 0.01) and were older (mean difference of 8.4 ± 3.7 years, p = 0.03). ALS-FTSD were also significantly older compared to HC (mean difference of 7.1 ± 2.7 years, p = 0.02).

TABLE 1. Demographic information.

	ALS-T	ALS-CN	ALS-FTSD	HC
N	24	13	11 (3 FTD, 5 CI, 1 BI, 2 CBI)	12
Age at scan (years)	61 ± 10 (44–80)	57 ± 9 (44–80)	65 ± 9 (51–80)	58 ± 4 (52–63)
Male/female (N)	13/11	7/6	6/5	6/6
Disease duration (years)	2.2 ± 2.2 (0.7–7.9) (n = 23)	1.2 ± 0.7 (0.7–3.3) (n = 13)	3.4 ± 2.8 (0.7–7.9) (n = 10)	NA
ALS-FRS-R (0–48)	38.0 ± 6 (19–46)	39 ± 7 (19–44)	37 ± 6 (28–46)	NA
Time between ALS-FRS-R and HDAC6 PET (days)	58 ± 53 (1–166)	50 ± 50 (2–166)	67 ± 57 (1–144)	NA
ECAS (0–136)	98 ± 17 (41–116) (n = 21)	109 ± 7 (97–116) (n = 10)	88 ± 17 (41–99) (n = 11)	NA
Time between ECAS and HDAC6 PET (days)	43 ± 49 (0–176) (n = 21)	47 ± 61 (0–176) (n = 10)	38 ± 33 (6–109) (n = 11)	NA
Bulbar/spinals (side) onset	11/13 (9L4R)	7/6 (3L3R)	4/7 (6L1R)	NA
Mutations	7 C9orf72 1 TARDBP 1 GRN	4 C9orf72 1 TARDBP	3 C9orf72 1 GRN	NA
Injected activity (MBq)	126 ± 14 (102–169)	127 ± 16 (104–169)	125 ± 13 (102–145)	138 ± 33 (116–201)

Note: Values as mean ± SD (range).
Abbreviations: ALS, amyotrophic lateral sclerosis; CI, cognitively impaired; CN, cognitively normal; HC, healthy controls; L, left; NA, not applicable; R, right; T, total.

The ALS group was exploratively divided into pwALS with known genetic mutations (“mutation carriers”) and those without (“no mutation carriers”), regardless of their cognitive status. In ALS “mutation carriers,” 4 out of 9 pwALS were categorized as cognitively impaired (3 CI and 1 CBI). In pwALS “no mutation carriers” this was 7 out of 15 (3 FTD, 2 CI, 1 BI and 1 CBI).

3.2 Group Comparison of Gray Matter Atrophy

VBM showed multiple clusters of significantly lower gray matter concentration in pwALS compared to HC, with clusters in the right precentral gyrus, left inferior temporal gyrus, left superior frontal gyrus, right temporal pole, and left posterior orbital gyrus (Figure 1). Therefore, group comparisons and correlation analyses were performed with PVC. Results without PVC can be found in the Supporting Information.

Details are in the caption following the image — **FIGURE 1**
Open in figure viewer PowerPoint

Voxel-based morphometry for ALS-T versus HC. Axial slices indicating clusters with significant reduction of gray matter probability in pwALS. Images shown in radiological convention. Color bar indicates SPM t-values. Significance thresholds were set at cluster height p_FWE < 0.05, peak height p_uncorrected < 0.001. ALS, amyotrophic lateral sclerosis; HC, healthy controls; SPM, statistical parametric mapping.

3.3 Group Comparison of [¹⁸F]EKZ-001 Binding

Group mean [¹⁸F]EKZ-001 V_T images as well as individual example [¹⁸F]EKZ-001 V_T images are shown in Figures S1 and S2. Voxel-based group analyses showed a widespread cluster, most prominent in the right parietotemporal and occipital regions, with significantly lower [¹⁸F]EKZ-001 V_T in ALS-T compared to HC. [¹⁸F]EKZ-001 V_T was lower in ALS-CN compared to HC in the brainstem, splenium of the corpus callosum, occipital white matter, and superior occipital gyrus. Even greater reductions in [¹⁸F]EKZ-001 V_T were evident in ALS-FTSD compared to HC throughout most gray and white matter regions, with peak voxels in the postcentral gyrus (Figure 2). The voxel-based analysis showed no significant differences in [¹⁸F]EKZ-001 V_T between ALS-FTSD and ALS-CN. Additional results, with and without PVC, can be found in Tables S1–S7 and Figure S3.

VOI-based analyses showed significantly lower [¹⁸F]EKZ-001 V_T values in the ALS-T group compared to HC in the brainstem, cerebellum, and thalamus as well as in the occipital, parietal, and temporal cortex (Figure S4). The most pronounced reduction was found in the brainstem (22%). Despite widespread trends of lower tracer binding across brain regions in ALS-CN compared to HC, [¹⁸F]EKZ-001 V_T values were only significantly lower in the brainstem VOI with a mean reduction of 20% (Figure 3). When comparing ALS-FTSD with HC, we found significantly lower tracer binding in all VOIs except for the frontal cortex and white matter (Figure 3). The largest reductions were found in the brainstem (24%) and in the occipital and temporal cortex VOIs (both 23%). Comparing the ALS-FTSD to the ALS-CN group showed widespread trends of lower tracer binding (Figure 3). Percent differences are summarized in Table 2 (with PVC) and Table S8 (without PVC). Group differences in [¹⁸F]EKZ-001 V_T remained significant after correction for age. [¹⁸F]EKZ-001 V_T levels were significantly associated with Fazekas scores in the cerebellum VOI (p = 0.02). However, this association did not remain significant after Bonferroni correction for multiple comparisons.

TABLE 2. Mean and standard deviation of [¹⁸F]EKZ-001 V_T values per VOI with relative group differences and uncorrected p-values compared to healthy controls.

VOI	HC	ALS-T	Diff	p-value	ALS-CN	Diff	p-value	ALS-FTSD	Diff	p-value
Brainstem	23.0 ± 3.1	18.0 ± 3.0	22%	< 0.001	18.3 ± 2.4	20%	< 0.001	17.5 ± 3.6	24%	< 0.001
Cerebellum	39.5 ± 5.8	33.0 ± 6.0	17%	0.004	34.9 ± 6.0	12%	0.064	30.6 ± 5.3	22%	0.001
Frontal cortex	38.8 ± 5.4	35.0 ± 5.3	10%	0.055	37.0 ± 5.6	5%	0.418	32.7 ± 4.1	16%	0.007
Motor cortex	33.7 ± 5.2	28.6 ± 5.1	15%	0.008	29.7 ± 5.9	12%	0.085	27.3 ± 3.8	19%	0.003
Occipital cortex	34.6 ± 4.3	28.3 ± 4.9	18%	< 0.001	29.9 ± 5.3	14%	0.023	26.6 ± 3.8	23%	< 0.001
Parietal cortex	37.0 ± 4.9	31.5 ± 5.1	15%	0.004	33.2 ± 5.3	10%	0.079	29.5 ± 4.2	20%	< 0.001
Striatum	41.5 ± 5.0	36.4 ± 7.0	12%	0.031	38.3 ± 7.7	8%	0.238	34.1 ± 5.7	18%	0.003
Temporal cortex	39.5 ± 5.5	32.9 ± 5.1	17%	0.001	35.2 ± 5.0	11%	0.053	30.2 ± 3.8	23%	< 0.001
Thalamus	29.8 ± 3.7	24.5 ± 4.4	18%	< 0.001	25.6 ± 4.3	14%	0.015	23.2 ± 4.2	22%	< 0.001
White matter	16.0 ± 1.8	13.8 ± 2.5	14%	0.011	14.3 ± 2.2	11%	0.047	13.2 ± 2.8	17%	0.010

Abbreviations: ALS-CN, ALS-cognitively normal; ALS-FTSD, ALS-cognitively impaired; ALS-T, total ALS group; Diff, relative difference; HC, healthy controls; VOI, volume-of-interest; V_T, distribution volume.

[¹⁸F]EKZ-001 V_T was not significantly different between “mutation carriers” and “no mutation carriers” (all p > 0.05) (Figure S5). Two-tissue compartment model K1 values were not significantly different between ALS-T and HC, nor between ALS-CN, ALS-FTSD and HC (Figure S6).

3.4 Correlation of [¹⁸F]EKZ-001 Binding With Neuropsychological Measures

[¹⁸F]EKZ-001 V_T was not significantly correlated with residual scores of the cognitive domains from the neuropsychological assessment, age, or disease duration (all p ≥ 0.07). Correlations between [¹⁸F]EKZ-001 V_T and neuropsychological assessment domain residual scores are summarized by a correlation matrix (Figure S7). For a subset of 14 pwALS, CSF NfL levels were determined, but these were not significantly associated with [¹⁸F]EKZ-001 V_T (p_uncorrected between 0.08 and 0.3 for all VOIs).

4 Discussion

The goal of this study was to investigate the binding of the HDAC6 PET tracer, [¹⁸F]EKZ-001, in pwALS compared to HC. Our primary finding is that [¹⁸F]EKZ-001 binding as measured by V_T was lower in pwALS compared to HC in most VOIs, including the brainstem, occipital, parietal, and temporal cortex, thalamus, and cerebellum. This suggests that HDAC6 expression may be lower in pwALS compared to HC, in line with reduced HDAC6 expression observed in SOD1^G93A transgenic mice [17]. However, reduced [¹⁸F]EKZ-001 binding may not be indicative of reduced enzymatic activity, as HDAC6 can be modulated by GSK3β-mediated phosphorylation to increase deacetylase activity [29]. We did not find evidence that known underlying genetic forms of ALS (specifically a C9orf72 expansion) influence [¹⁸F]EKZ-001 binding. The group comparison of K₁ values showed no significant differences, indicating that [¹⁸F]EKZ-001 influx is comparable between groups and cannot explain the differences in V_T between pwALS and HC. In the ALS-T group, cerebellar [¹⁸F]EKZ-001 V_T was significantly associated with Fazekas scores, but this association was not significant in any other VOI and would not have survived correction for multiple comparisons, indicating that differences in tracer binding were not likely to be driven by white matter lesions. There was a trend towards an association of [¹⁸F]EKZ-001 V_T with CSF NfL levels in the ALS-T group. However, these NfL levels were only available for a subset of 14 pwALS, resulting in limited statistical power for this analysis and warranting the need for replication in larger cohorts.

HDAC6 downregulation in SOD1^G93A mice is progressive and occurs in older mice, as they become symptomatic [17], suggesting that HDAC6 expression may be dynamic in ALS. In presymptomatic disease stages, there may be a transient increase in HDAC6 expression to aid in the clearance of toxic proteins, such as TDP-43 and SOD1. Consistently, HDAC6 is upregulated in acute models of Parkinson's disease [30, 31]. Reduced HDAC6 expression in symptomatic disease stages could be a compensatory response to the excessive deacetylating activity of the enzyme [32-34]. Future longitudinal [¹⁸F]EKZ-001 PET studies in pre-symptomatic carriers would be interesting to gain further insight into HDAC6 dynamics throughout the disease course of ALS. Alternatively, increased demand for HDAC6-dependent autophagic protein aggregate clearance may lead to sequestration of HDAC6 within protein aggregates [35, 36], rendering the HDAC6 catalytic domain 2 inaccessible to the binding of [¹⁸F]EKZ-001. Finally, the difference in [¹⁸F]EKZ-001 signal could be due to off-target binding of the tracer to metallo-beta-lactamase domain-containing protein 2 (MBLAC2), a palmitoyl-CoA hydrolase involved in endo- and exocytosis. Hydroxamate-based HDAC6 inhibitors, including EKZ-001 (unpublished results), inhibit MBLAC2 in vitro with nanomolar potency [37]. Although the relevance and role of MBLAC2 in ALS is unknown, differences in MBLAC2 expression between ALS and HC might account for the observed findings. To further characterize potential [¹⁸F]EKZ-001 off-target binding, quantitative binding assays and autoradiography in healthy and ALS brain samples are needed.

These findings may have implications for the development of disease-modifying therapies for ALS. Therapeutic development currently focuses on small-molecule HDAC6 inhibitors. Importantly, selective inhibition of HDAC6 catalytic domain 2, which has broad substrate specificity and is the primary target of HDAC6 inhibitors [38], does not impede the c-terminal zinc-finger ubiquitin binding domain activities of HDAC6 in protein aggregate clearance but instead stimulates autophagic flux within neurons [39]. Thus, despite the apparent lower HDAC6 expression in ALS compared to HC described here, inhibition of the catalytic activity of the existing pool of HDAC6 in pwALS has the potential to (1) facilitate intracellular transport by increasing α-tubulin acetylation and (2) stimulate autophagic clearance of toxic protein aggregates like TDP-43 in ALS. In support of this possibility, HDAC6 inhibition is efficacious in preclinical animal and cellular models of ALS and FTD [11, 13, 14, 40]. Furthermore, in line with preclinical evidence, HDAC6 overexpression combined with HDAC6 catalytic domain 2 inhibition could be beneficial through interactions with the ubiquitin-proteasome and autophagy-lysosome systems, which could mitigate neuronal toxicity caused by TDP-43 pathology [17-19].

To date, one previous study explored differences in HDAC binding in pwALS using [¹¹C]Martinostat PET, a pan-HDAC tracer that binds to paralogs 1, 2, and 3, as well as to HDAC6. The authors reported no significant differences in [¹¹C]Martinostat binding [41]. Since HDAC6 is significantly different in pwALS compared to HC, we hypothesize that HDAC6 may have a more specific role compared to the other HDAC paralogs in ALS. In contrast, in patients with Alzheimer's disease, reduced HDAC levels as measured by [¹¹C]Martinostat were associated with elevated amyloid-β and tau PET levels [42]. Moreover, [¹¹C]Martinostat binding was shown to mediate the deleterious effects of amyloid-β and tau on brain atrophy and cognitive impairment. To the best of our knowledge, [¹⁸F]EKZ-001 PET has not been studied in other neurodegenerative disorders and a head-to-head comparison of [¹¹C]Martinostat and [¹⁸F]EKZ-001 is still lacking.

The lack of correlation with neuropsychological assessment, age, and disease duration, in combination with the observed difference in group comparisons for ALS-FTSD and ALS-CN, could suggest differences in underlying disease pathophysiology between ALS-FTSD and ALS-CN. Future longitudinal evaluation should be conducted to elucidate the possible differential effect related to extra-motor involvement. It would also be interesting to compare cerebral [¹⁸F]EKZ-001 binding with plasma assessments of TDP-43 and NfL, as it was recently shown that both plasma TDP-43 and NfL are associated with disease severity in ALS, as measured by ECAS and ALSFRS-R [43, 44].

Limitations of this study include a relatively small sample size, particularly for the various genetic variants and FTSD, as well as the necessity of progressive coffee-break protocols for the acquisition of the [¹⁸F]EKZ-001 PET scans due to the inability of advanced pwALS to undergo long dynamic scanning. To ensure that group differences were not driven by differences in processing, all images were processed using the same t* values. In HC, the average bias introduced by the minimal coffee-break protocol (CBP2) compared to the full dynamic acquisition ranged from 1.7% in the occipital cortex to 2.6% in the white matter (Data S1), indicating minimal influence on the obtained results. The pwALS sample included in this study is not entirely representative of the ALS population since we included a relatively large subsample of C9orf72 mutation carriers. Moreover, the recruited C9orf72 mutation carriers were in an early disease stage, associated with a lower prevalence of cognitive impairment in this subcohort. Furthermore, due to the small sample size per group, we did not include age and sex as covariates in the analyses even though sex differences have been shown previously for [¹⁸F]EKZ-001 PET [22]. However, the number of males and females per group was matched to ensure sex differences would not influence the findings.

To conclude, we observed that HDAC6 tracer binding is lower throughout the brain in pwALS compared to HC, as measured by [¹⁸F]EKZ-001 PET. The reduction was more pronounced in ALS-FTSD and was not influenced by C9orf72 mutation status or related to performance on cognitive domains of the ECAS, age, or disease duration.

Author Contributions

K.V.L., P.V.D., D.V.W., T.M.G., and J.E.K. contributed to the study concept and design. G.V., C.C., J.D.V., F.O., P.M., C.D.W., and P.V.D. contributed to participant recruitment. G.V., C.C., and C.D.W. contributed to data acquisition. G.V., C.C., M.K., R.E.J., F.A.S., J.M.H., T.M.G., and K.V.L. contributed to data analysis and interpretation. G.V., C.C., and K.V.L. drafted the manuscript. Manuscript revision was performed by all authors.

Acknowledgments

We thank the participants for their willingness to participate in this study. We thank the PET-MR technologists for their contribution to data acquisition. We thank the PET radiopharmacy team and nuclear medicine medical physics team for their skilled contributions. Financial support was provided through the Target ALS Foundation and AFTD, and a research grant to KU Leuven by Eikonizo Therapeutics.

Conflicts of Interest

The authors declare no conflicts of interest.

Open Research

Data Availability Statement

Anonymized data will be deposited in an access-controlled file server used by the academic research PET imaging group, which can be shared upon reasonable request from any qualified investigator on approval by the Ethics Committee of the local university hospital.

Supporting Information

References

1P. Masrori and P. van Damme, “Amyotrophic Lateral Sclerosis: A Clinical Review,” European Journal of Neurology 27, no. 10 (2020): 1918–1929.
10.1111/ene.14393
CAS PubMed Web of Science® Google Scholar
2C. Hubbert, A. Guardiola, R. Shao, et al., “HDAC6 Is a Microtubule-Associated Deacetylase,” Nature 417, no. 6887 (2002): 455–458.
10.1038/417455a
CAS PubMed Web of Science® Google Scholar
3X. Zhang, Z. Yuan, Y. Zhang, et al., “HDAC6 Modulates Cell Motility by Altering the Acetylation Level of Cortactin,” Molecular Cell 27, no. 2 (2007): 197–213.
10.1016/j.molcel.2007.05.033
CAS PubMed Web of Science® Google Scholar
4H. Trzeciakiewicz, D. Ajit, J.-H. Tseng, et al., “An HDAC6-Dependent Surveillance Mechanism Suppresses Tau-Mediated Neurodegeneration and Cognitive Decline,” Nature Communications 11, no. 1 (2020): 5522.
10.1038/s41467-020-19317-4
CAS PubMed Web of Science® Google Scholar
5J. J. Kovacs, P. J. M. Murphy, S. Gaillard, et al., “HDAC6 Regulates Hsp90 Acetylation and Chaperone-Dependent Activation of Glucocorticoid Receptor,” Molecular Cell 18, no. 5 (2005): 601–607.
10.1016/j.molcel.2005.04.021
CAS PubMed Web of Science® Google Scholar
6H. Choi, H. J. Kim, J. Yang, et al., “Acetylation Changes Tau Interactome to Degrade Tau in Alzheimer's Disease Animal and Organoid Models,” Aging Cell 19, no. 1 (2020): e13081.
10.1111/acel.13081
CAS PubMed Web of Science® Google Scholar
7B. Hempen and J. P. Brion, “Reduction of Acetylated Alpha-Tubulin Immunoreactivity in Neurofibrillary Tangle-Bearing Neurons in Alzheimer's Disease,” Journal of Neuropathology and Experimental Neurology 55, no. 9 (1996): 964–972.
10.1097/00005072-199609000-00003
CAS PubMed Web of Science® Google Scholar
8A. R. Esteves and S. M. Cardoso, “Differential Protein Expression in Diverse Brain Areas of Parkinson's and Alzheimer's Disease Patients,” Scientific Reports 10 (2020): 13149.
10.1038/s41598-020-70174-z
CAS PubMed Web of Science® Google Scholar
9K. Miskiewicz, L. E. Jose, W. M. Yeshaw, et al., “HDAC6 Is a Bruchpilot Deacetylase That Facilitates Neurotransmitter Release,” Cell Reports 8, no. 1 (2014): 94–102.
10.1016/j.celrep.2014.05.051
CAS PubMed Web of Science® Google Scholar
10F. C. Fiesel, A. Voigt, S. S. Weber, et al., “Knockdown of Transactive Response DNA-Binding Protein (TDP-43) Downregulates Histone Deacetylase 6,” EMBO Journal 29, no. 1 (2010): 209–221.
10.1038/emboj.2009.324
CAS PubMed Web of Science® Google Scholar
11W. Guo, M. Naujock, L. Fumagalli, et al., “HDAC6 Inhibition Reverses Axonal Transport Defects in Motor Neurons Derived From FUS-ALS Patients,” Nature Communications 8, no. 1 (2017): 861.
10.1038/s41467-017-00911-y
PubMed Web of Science® Google Scholar
12K. Heo, S. M. Lim, M. Nahm, et al., “A De Novo RAPGEF2 Variant Identified in a Sporadic Amyotrophic Lateral Sclerosis Patient Impairs Microtubule Stability and Axonal Mitochondria Distribution,” Experimental Neurobiology 27, no. 6 (2018): 550–563.
10.5607/en.2018.27.6.550
PubMed Web of Science® Google Scholar
13R. Fazal, S. Boeynaems, A. Swijsen, et al., “HDAC6 Inhibition Restores TDP-43 Pathology and Axonal Transport Defects in Human Motor Neurons With TARDBP Mutations,” EMBO Journal 40, no. 7 (2021): e106177.
10.15252/embj.2020106177
CAS PubMed Web of Science® Google Scholar
14K. Stoklund Dittlau, E. N. Krasnow, L. Fumagalli, et al., “Human Motor Units in Microfluidic Devices Are Impaired by FUS Mutations and Improved by HDAC6 Inhibition,” Stem Cell Reports 16, no. 9 (2021): 2213–2227.
10.1016/j.stemcr.2021.03.029
CAS PubMed Google Scholar
15I. Taes, M. Timmers, N. Hersmus, et al., “Hdac6 Deletion Delays Disease Progression in the SOD1G93A Mouse Model of ALS,” Human Molecular Genetics 22, no. 9 (2013): 1783–1790.
10.1093/hmg/ddt028
CAS PubMed Web of Science® Google Scholar
16J.-Y. Lee, Y. Kawaguchi, M. Li, et al., “Uncoupling of Protein Aggregation and Neurodegeneration in a Mouse Amyotrophic Lateral Sclerosis Model,” Neurodegenerative Diseases 15, no. 6 (2015): 339–349.
10.1159/000437208
CAS PubMed Web of Science® Google Scholar
17S. Chen, X.-J. Zhang, L.-X. Li, Y. Wang, R. J. Zhong, and W. le, “Histone Deacetylase 6 Delays Motor Neuron Degeneration by Ameliorating the Autophagic Flux Defect in a Transgenic Mouse Model of Amyotrophic Lateral Sclerosis,” Neuroscience Bulletin 31, no. 4 (2015): 459–468.
10.1007/s12264-015-1539-3
CAS PubMed Web of Science® Google Scholar
18S. Lee, Y. Kwon, S. Kim, et al., “The Role of HDAC6 in TDP-43-Induced Neurotoxicity and UPS Impairment,” Frontiers in Cell and Development Biology 8 (2020): 581942.
10.3389/fcell.2020.581942
PubMed Web of Science® Google Scholar
19J. Gal, J. Chen, K. R. Barnett, L. Yang, E. Brumley, and H. Zhu, “HDAC6 Regulates Mutant SOD1 Aggregation Through Two SMIR Motifs and Tubulin Acetylation,” Journal of Biological Chemistry 288, no. 21 (2013): 15035–15045.
10.1074/jbc.M112.431957
CAS PubMed Web of Science® Google Scholar
20M. G. Strebl, A. J. Campbell, W.-N. Zhao, et al., “HDAC6 Brain Mapping With [18F]Bavarostat Enabled by a Ru-Mediated Deoxyfluorination,” ACS Central Science 3, no. 9 (2017): 1006–1014.
10.1021/acscentsci.7b00274
CAS PubMed Web of Science® Google Scholar
21S. Celen, J. Rokka, T. M. Gilbert, et al., “Translation of HDAC6 PET Imaging Using [18F]EKZ-001—cGMP Production and Measurement of HDAC6 Target Occupancy in NHPs,” ACS Chemical Neuroscience 11, no. 7 (2020): 1093–1101.
10.1021/acschemneuro.0c00074
CAS PubMed Web of Science® Google Scholar
22M. Koole, D. Van Weehaeghe, K. Serdons, et al., “Clinical Validation of the Novel HDAC6 Radiotracer [18F]EKZ-001 in the Human Brain,” European Journal of Nuclear Medicine and Molecular Imaging 48, no. 2 (2021): 596–611.
10.1007/s00259-020-04891-y
CAS PubMed Web of Science® Google Scholar
23B. R. Brooks, R. G. Miller, M. Swash, T. L. Munsat, and World Federation of Neurology Research Group on Motor Neuron Diseases, “El Escorial Revisited: Revised Criteria for the Diagnosis of Amyotrophic Lateral Sclerosis,” Amyotrophic Lateral Sclerosis and Other Motor Neuron Disorders 1, no. 5 (2000): 293–299.
10.1080/146608200300079536
CAS PubMed Web of Science® Google Scholar
24M. J. Strong, S. Abrahams, L. H. Goldstein, et al., “Amyotrophic Lateral Sclerosis—Frontotemporal Spectrum Disorder (ALS-FTSD): Revised Diagnostic Criteria,” Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration 18, no. 3–4 (2017): 153–174.
10.1080/21678421.2016.1267768
PubMed Web of Science® Google Scholar
25L. A. Bakker, C. D. Schröder, L. A. Spreij, et al., “Derivation of Norms for the Dutch Version of the Edinburgh Cognitive and Behavioral ALS Screen,” Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration 20, no. 1–2 (2019): 19–27.
10.1080/21678421.2018.1522352
CAS PubMed Web of Science® Google Scholar
26G. Schramm, M. Koole, S. M. A. Willekens, et al., “Regional Accuracy of ZTE-Based Attenuation Correction in Static [18F]FDG and Dynamic [18F]PE2I Brain PET/MR [Internet],” Frontiers in Physics 7 (2019): 211, https://doi.org/10.3389/fphy.2019.00211.
10.3389/fphy.2019.00211
Web of Science® Google Scholar
27C. Gaser, R. Dahnke, P. M. Thompson, F. Kurth, E. Luders, and the Alzheimer's Disease Neuroimaging Initiative, “CAT: A Computational Anatomy Toolbox for the Analysis of Structural MRI Data,” GigaScience 13 (2024): giae049.
10.1093/gigascience/giae049
PubMed Web of Science® Google Scholar
28N. Mertens, L. Michiels, G. Vanderlinden, et al., “Impact of Meningeal Uptake and Partial Volume Correction Techniques on [18F]MK-6240 Binding in aMCI Patients and Healthy Controls,” Journal of Cerebral Blood Flow and Metabolism 42, no. 7 (2022): 1236–1246.
10.1177/0271678X221076023
CAS PubMed Web of Science® Google Scholar
29S. Chen, G. C. Owens, H. Makarenkova, and D. B. Edelman, “HDAC6 Regulates Mitochondrial Transport in Hippocampal Neurons,” PLoS One 5, no. 5 (2010): e10848.
10.1371/journal.pone.0010848
CAS PubMed Web of Science® Google Scholar
30W. Jian, X. Wei, L. Chen, et al., “Inhibition of HDAC6 Increases Acetylation of peroxiredoxin1/2 and Ameliorates 6-OHDA Induced Dopaminergic Injury,” Neuroscience Letters 658 (2017): 114–120.
10.1016/j.neulet.2017.08.029
CAS PubMed Web of Science® Google Scholar
31M. Su, J.-J. Shi, Y.-P. Yang, et al., “HDAC6 Regulates Aggresome-Autophagy Degradation Pathway of α-Synuclein in Response to MPP+-Induced Stress,” Journal of Neurochemistry 117, no. 1 (2011): 112–120.
10.1111/j.1471-4159.2011.07180.x
CAS PubMed Web of Science® Google Scholar
32U. B. Pandey, Z. Nie, Y. Batlevi, et al., “HDAC6 Rescues Neurodegeneration and Provides an Essential Link Between Autophagy and the UPS,” Nature 447, no. 7146 (2007): 859–863.
10.1038/nature05853
CAS PubMed Web of Science® Google Scholar
33C. Boyault, Y. Zhang, S. Fritah, et al., “HDAC6 Controls Major Cell Response Pathways to Cytotoxic Accumulation of Protein Aggregates,” Genes & Development 21, no. 17 (2007): 2172–2181.
10.1101/gad.436407
CAS PubMed Web of Science® Google Scholar
34J.-Y. Lee, H. Koga, Y. Kawaguchi, et al., “HDAC6 Controls Autophagosome Maturation Essential for Ubiquitin-Selective Quality-Control Autophagy,” EMBO Journal 29, no. 5 (2010): 969–980.
10.1038/emboj.2009.405
CAS PubMed Web of Science® Google Scholar
35Y. Kawaguchi, J. J. Kovacs, A. McLaurin, J. M. Vance, A. Ito, and T. P. Yao, “The Deacetylase HDAC6 Regulates Aggresome Formation and Cell Viability in Response to Misfolded Protein Stress,” Cell 115, no. 6 (2003): 727–738.
10.1016/S0092-8674(03)00939-5
CAS PubMed Web of Science® Google Scholar
36H. Ouyang, Y. O. Ali, M. Ravichandran, et al., “Protein Aggregates Are Recruited to Aggresome by Histone Deacetylase 6 via Unanchored Ubiquitin C Termini,” Journal of Biological Chemistry 287, no. 4 (2012): 2317–2327.
10.1074/jbc.M111.273730
CAS PubMed Web of Science® Google Scholar
37S. Lechner, M. I. P. Malgapo, C. Grätz, et al., “Target Deconvolution of HDAC Pharmacopoeia Reveals MBLAC2 as Common Off-Target,” Nature Chemical Biology 18, no. 8 (2022): 812–820.
10.1038/s41589-022-01015-5
CAS PubMed Web of Science® Google Scholar
38S. Jha, J. H. Kim, M. Kim, et al., “Design, Synthesis, and Biological Evaluation of HDAC6 Inhibitors Targeting L1 Loop and Serine 531 Residue,” European Journal of Medicinal Chemistry 265 (2024): 116057.
10.1016/j.ejmech.2023.116057
CAS PubMed Web of Science® Google Scholar
39P. Guedes-Dias, J. de Proença, T. R. Soares, et al., “HDAC6 Inhibition Induces Mitochondrial Fusion, Autophagic Flux and Reduces Diffuse Mutant Huntingtin in Striatal Neurons,” Biochimica et Biophysica Acta 1852, no. 11 (2015): 2484–2493.
10.1016/j.bbadis.2015.08.012
CAS PubMed Web of Science® Google Scholar
40A. J. Phipps, S. Dwyer, J. M. Collins, et al., “HDAC6 Inhibition as a Mechanism to Prevent Neurodegeneration in the mSOD1G93A Mouse Model of ALS,” Heliyon 10, no. 14 (2024): e34587.
10.1016/j.heliyon.2024.e34587
CAS PubMed Web of Science® Google Scholar
41A. M. Dios, S. Babu, E. J. Granucci, et al., “Class I and II Histone Deacetylase Expression Is Not Altered in Human Amyotrophic Lateral Sclerosis: Neuropathological and Positron Emission Tomography Molecular Neuroimaging Evidence,” Muscle & Nerve 60, no. 4 (2019): 443–452.
10.1002/mus.26620
CAS PubMed Web of Science® Google Scholar
42T. A. Pascoal, M. Chamoun, E. Lax, et al., “[11C]Martinostat PET Analysis Reveals Reduced HDAC I Availability in Alzheimer's Disease,” Nature Communications 13, no. 1 (2022): 4171.
10.1038/s41467-022-30653-5
CAS PubMed Web of Science® Google Scholar
43M. Chatterjee, S. Özdemir, C. Fritz, et al., “Plasma Extracellular Vesicle Tau and TDP-43 as Diagnostic Biomarkers in FTD and ALS,” Nature Medicine 30, no. 6 (2024): 1771–1783.
10.1038/s41591-024-02937-4
CAS PubMed Web of Science® Google Scholar
44T. J. Simkins, S. Kupfer, F. I. Malik, et al., “Plasma Neurofilament Analysis in VITALITY-ALS,” Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration 26, no. 112 (2025): 103.
10.1080/21678421.2024.2423707
CAS PubMed Google Scholar

Volume12, Issue7

July 2025

Pages 1350-1359

Histone Deacetylase 6 Brain PET in Amyotrophic Lateral Sclerosis-Frontotemporal Spectrum Disorder