Research ArticleBrain Imaging

Insights into neuroepigenetics through human histone deacetylase PET imaging

See allHide authors and affiliations

Science Translational Medicine  10 Aug 2016:
Vol. 8, Issue 351, pp. 351ra106
DOI: 10.1126/scitranslmed.aaf7551

Brain epigenetics revealed

Certain enzymes called histone deacetylases, or HDACs, are part of the epigenetic machinery that regulates gene transcription. In neurological disorders, HDACs change expression in regions throughout the brain, but their dynamic contribution to human disease development over time is unknown. Wey et al. therefore developed and applied an HDAC imaging probe, called Martinostat, to visualize HDAC expression in the living brain. Martinostat was previously tested in rodents and nonhuman primates, and here, it is used for the first time in humans. The authors saw surprisingly conserved regions of HDAC expression in the healthy brain, suggesting tightly regulated epigenetic processes. In human stem cell–derived neural progenitor cells, Martinostat engaged the subset HDACs that regulate downstream genes important for neuroplasticity, memory, and neurodegeneration, supporting its use in monitoring and understanding brain pathologies like Alzheimer’s disease.

Abstract

Epigenetic dysfunction is implicated in many neurological and psychiatric diseases, including Alzheimer’s disease and schizophrenia. Consequently, histone deacetylases (HDACs) are being aggressively pursued as therapeutic targets. However, a fundamental knowledge gap exists regarding the expression and distribution of HDACs in healthy individuals for comparison to disease states. Here, we report the first-in-human evaluation of neuroepigenetic regulation in vivo. Using positron emission tomography with [11C]Martinostat, an imaging probe selective for class I HDACs (isoforms 1, 2, and 3), we found that HDAC expression is higher in cortical gray matter than in white matter, with conserved regional distribution patterns within and between healthy individuals. Among gray matter regions, HDAC expression was lowest in the hippocampus and amygdala. Through biochemical profiling of postmortem human brain tissue, we confirmed that [11C]Martinostat selectively binds HDAC isoforms 1, 2, and 3, the HDAC subtypes most implicated in regulating neuroplasticity and cognitive function. In human stem cell–derived neural progenitor cells, pharmacologic-level doses of Martinostat induced changes in genes closely associated with synaptic plasticity, including BDNF (brain-derived neurotrophic factor) and SYP (synaptophysin), as well as genes implicated in neurodegeneration, including GRN (progranulin), at the transcript level, in concert with increased acetylation at both histone H3 lysine 9 and histone H4 lysine 12. This study quantifies HDAC expression in the living human brain and provides the foundation for gaining unprecedented in vivo epigenetic information in health and disease.

INTRODUCTION

Disorders of the central nervous system (CNS), including Alzheimer’s disease (AD), schizophrenia, depression, and addiction, are increasingly recognized to involve dysregulation of epigenetic machinery. Among all, histone deacetylases (HDACs)—a family of chromatin-modifying enzymes that dynamically regulates gene transcription—are the most frequently implicated (1, 2). A subset of HDACs has already been linked to neuronal development, synaptic plasticity, and cognition (3, 4). For example, postmortem human brain tissue analyses and in vivo rodent studies exposed HDAC1, HDAC2, and HDAC3 as antagonists of learning and memory and contributors to AD and mood disorders (3, 59). Genetic manipulations or pharmacologic inhibition of aberrant HDAC2 and HDAC3 activity rescued behavioral defects in rodent models of both AD and mood disorders (6, 7, 1014). HDAC inhibitors were also proposed as a targeted treatment of frontotemporal lobar degeneration, owing to mutations that cause haploinsufficiency of the progranulin-encoding gene GRN (14). Collectively, these studies implicate a direct relationship between the levels of class I HDACs (isoforms 1, 2, and 3) and neuronal function.

In addition to the overall level of HDAC expression within the brain, spatially localized variation of HDACs is also highly impactful in neuronal plasticity, memory, and behavior. For example, intrahippocampal injection of short hairpin RNA against Hdac2 selectively normalized HDAC2 levels and restored neuroplasticity-associated gene transcription, synaptic density, and cognitive behavior in a mouse model of AD (6). In contrast to the high level of hippocampal HDAC2 in animal models and postmortem human tissue from AD patients, deficient HDAC2 expression was observed in the frontal cortex of postmortem AD tissue, highlighting the importance of tightly regulated localized HDAC expression (15). Analogously, focal genetic deletion of Hdac3 in the hippocampus and the nucleus accumbens enhanced long-term memory and acquisition of cocaine-associated place preference in mice, respectively (5, 16). Although understanding of the full compendium of genes under HDAC-dependent regulation in defined regions of the brain is incomplete, HDAC2 chromatin immunoprecipitation studies in hippocampal tissue have identified several immediate-early genes (for example, BDNF and CDK5) involved in learning and memory, as well as multiple genes involved in synaptic plasticity (for example, SYP and SYT1) as downstream targets (3, 6, 17). Collectively, these studies provide support that localized HDAC expression levels drive pivotal epigenetic mechanisms that modulate neuronal function.

Although there is strong evidence for localized HDAC dysfunction in CNS disease, epigenetic models cannot recapitulate dynamic human-environment interactions and therefore may not accurately reflect in vivo human biology. Moreover, until now, there has been no method to visualize in vivo epigenetic mechanisms in humans. We developed the positron emission tomography (PET) epigenetic imaging agent, [11C]Martinostat, previously described in rodents and nonhuman primates (NHPs) (12, 18, 19). Our previous work in rodents demonstrated the specific and reversible binding properties of [11C]Martinostat and that the agent engaged recombinant class I HDACs (isoforms 1, 2, and 3) and class IIb HDAC (isoform 6) with low nanomolar affinities (18). Because [11C]Martinostat demonstrated excellent brain penetrance, it was used to determine whether clinically relevant HDAC inhibitors, such as suberoylanilide hydroxamic acid (SAHA) and CI-994, crossed the blood-brain barrier and exhibited target occupancy in rodents (12). Most recently, we performed studies in NHPs to characterize the kinetic properties of [11C]Martinostat and to estimate nondisplaceable binding of [11C]Martinostat with pharmacologic blockades in preparation for human studies (19). In addition to the brain, [11C]Martinostat showed high specific binding and fast binding kinetics appropriate for PET imaging in heart, pancreas, spleen, and kidneys (18, 19). Here, we translate [11C]Martinostat for clinical research use and quantify human epigenetic regulation.

RESULTS

In vivo human PET imaging reveals conserved regional HDAC expression patterns in the healthy brain

To visualize HDAC expression in the living human brain, we performed [11C]Martinostat PET imaging on eight healthy volunteers (four males and four females; mean age ± SD, 28.6 ± 7.6 years) (table S1). The uptake of [11C]Martinostat reached a maximum at ~30 min after injection and showed minimal decrease during the 90-min scan (fig. S1). The retention of radioactivity is a unique feature of [11C]Martinostat, which allows for a stable quantification of HDAC expression levels. Regional heterogeneity, such as different levels of [11C]Martinostat uptake between gray and white matter tissues, was observed at the individual subject level (Fig. 1). Quantitative analysis using compartmental modeling on individual subjects’ dynamic PET data allowed us to determine the distribution volume (VT), a measure of radiotracer binding that is normalized to the activity present in circulating blood, and rate constants describing the pharmacokinetics of [11C]Martinostat (fig. S1 and tables S2 and S3). VT values were stable beyond 50 min, with less than 10% variability when compared to the 90-min data (fig. S2).

Fig. 1. [11C]Martinostat images of all subjects show high cortical binding and distinct gray-white matter differences.

(A) [11C]Martinostat (injected dose, 4.7 mCi; specific activity, 1.1 mCi/nmol) images averaged from 60 to 90 min after radiotracer injection (SUV60-90 min; SUV = radioactivity per injected dose per body weight) from a representative subject overlaid on anatomical magnetic resonance (MR) image. (B) [11C]Martinostat SUVR60-90 min images of individual subjects. To facilitate intersubject comparison of regional HDAC distribution, we normalized regional SUV60-90 min to an individual subject’s white matter SUV60-90 min as SUV60-90 min ratios (SUVR60-90 min). The SUVR60-90 min images were also coregistered with an MNI152 standard human atlas brain.

Regional standardized uptake values from 60 to 90 min after radiotracer administration (SUV60-90 min), an image-based indicator of binding to HDACs (Fig. 2A), correlated positively with VT values (Fig. 2B). The image-based SUV60-90 min had less intersubject variability [coefficient of variation (CV) is 11.2 to 19.2% across brain regions] than the blood data–derived VT values (CV is 22.0 to 39.2% across brain regions) (Fig. 2B). SUV60-90 min may therefore be an appropriate surrogate outcome measurement for VT and can be used in future studies to eliminate arterial blood sampling and reduce sample size because of its smaller variation. As with all surrogate measures, validation relative to a full treatment of the data using arterial blood in each patient population will be required.

Fig. 2. Small intersubject variation of localized regional [11C]Martinostat binding in the human brain.

(A) Mean images (left) and standard deviation (inset, to the lower right of each composite image) of SUV60-90 min from healthy volunteers (n = 8). The images are overlaid onto the MNI152 standard brain, where x, y, and z indicate the coordinate of each image plane shown. (B) Correlation of regional VT values, derived from a two-tissue compartmental model using metabolite-corrected arterial plasma as an input function and SUV60-90 min. Data are means ± SD (n = 6 subjects), and each circle symbol represents a separate brain region (n = 14 brain regions). P value determined with Pearson correlation analysis. (C) Regional SUV60-90 min and SUV ratios (SUVR60-90 min) of cortical, subcortical, cerebellar, and white matter volumes of interest (VOIs). Individual pairs of brain regions that are significantly different from each other are listed in table S4. Each dashed line represents SUVR60-90 min from a single subject (n = 8).

Group-level analyses showed that the average gray matter SUV60-90 min was nearly double that of white matter (Fig. 2C, fig. S3, and table S4), and heterogeneous binding was observed among gray matter regions examined. Besides the white matter, the lowest [11C]Martinostat uptake was observed in the hippocampus and amygdala, and the highest was observed in the putamen and cerebellum (Fig. 2C, fig. S3, and table S4). To facilitate intersubject comparison of regional HDAC distribution, we normalized regional SUV60-90 min to individual subjects’ white matter SUV60-90 min as SUV60-90 min ratios (SUVR60-90 min). SUVR60-90 min showed that the regional distribution patterns of [11C]Martinostat binding were consistent in all subjects (Fig. 2C) and on consecutive scans in single subjects. In preliminary test/retest scans (3 hours apart) in the same individual, SUVR60-90 min showed less than 3% variability (fig. S4).

Ex vivo biochemistry of postmortem tissue confirms that Martinostat binds to HDAC1, HDAC2, and HDAC3 in the healthy brain

To assess regional differences in [11C]Martinostat binding in the human brain, we biochemically profiled postmortem brain tissue from gray matter regions [superior frontal gyrus (SFG), dorsolateral prefrontal cortex, hippocampus, and anterior cingulate] and a white matter region [corpus callosum (CC)] (table S5). Quantitative protein levels of HDAC1, HDAC2, HDAC3, and HDAC6 were determined by Western blotting (Fig. 3A). Significantly lower amounts of HDAC2 and HDAC3 were found in the CC relative to the SFG (Fig. 3B). No significant differences in HDAC expression were noted among the dorsolateral prefrontal cortex, hippocampus, or anterior cingulate—all gray matter regions (Fig. 3C). The average expression levels of HDAC2, HDAC3, and HDAC6 were similar in the SFG (0.12 to 0.16 pmol/mg total protein), with the notable exception of HDAC1 (1.7 pmol/mg total protein). Although high HDAC1 expression was observed across all brain regions tested (Fig. 3), we cannot exclude the possibility that these values are driven by postmortem neuronal death (20, 21).

Fig. 3. HDAC2 and HDAC3 expression levels are higher in cortical gray matter than in white matter.

Whole-cell lysates were prepared from postmortem human SFG and CC (n = 3 replicate donor pools with two donors per pool), as well as dorsolateral prefrontal cortex (DLPFC), hippocampus (Hipp), and anterior cingulate (Ant Cing) (n = 3 replicate pools with three donors per pool). (A) Equivalent amounts of total protein were compared to human recombinant HDAC standards through Western blotting. #The HDAC2 recombinant standard was tagged with glutathione S-transferase (GST), resulting in increased molecular weight. (B and C) Comparison of HDAC expression between white matter (CC) and gray matter (SFG) regions (B) and among different gray matter regions (C). HDAC immunoreactive band intensity values were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) intensity values. HDAC expression levels were calculated per milligram of total extracted protein. Solid lines represent mean expression values. Donor pools are denoted by black, gray, and open circles. P values were determined by unpaired t test (B) and ordinary one-way analysis of variance (ANOVA) (α = 0.05 with Tukey’s multiple comparisons correction) (C).

HDAC2 and HDAC3 expression level differences between the SFG and the CC could not be attributed to nuclear density, according to quantification of the number of nuclei per field of view in postmortem baboon brain tissue (fig. S5). We observed that the CC had an increased number of nuclei compared to the SFG, which suggested that lower HDAC expression in the CC was not due to a depletion of cells in this brain region (fig. S5). As nuclear size (area per nucleus) was smaller in the CC than in the SFG, the total nuclear area per field of view was equivalent between these regions, further refuting that HDAC expression levels are driven by nuclear density.

Thermal shift assays evaluate target engagement, such that inhibitor binding increases the thermal stability of a target protein, as compared to a vehicle control (22, 23). To determine the HDAC isoform selectivity of Martinostat, thermal shift assays were performed with clarified human brain homogenate and increasing concentrations of Martinostat. Martinostat stabilized HDAC1, HDAC2, and HDAC3 in both the SFG and the CC at nanomolar concentrations (Fig. 4A, with individual biological replicates in fig. S6). No significant stabilization of either HDAC6 or HDAC8 (negative control) was observed. The former suggests differences between the accessibility of endogenous HDAC6 complex and Martinostat binding, relative to recombinant protein (18). To assess heterogeneity in HDAC isoform selectivity across gray matter regions, we compared SFG binding to the dorsolateral prefrontal cortex, hippocampus, and anterior cingulate. On the basis of thermal stabilization data, Martinostat exhibited a relatively uniform binding profile in gray matter with target engagement observed at concentrations around and above 0.160 μM (Fig. 4B, with individual biological replicates in figs. S7 to S9).

Fig. 4. Martinostat engages HDAC1, HDAC2, and HDAC3 in the human brain.

(A) Whole-cell lysates were prepared from postmortem human SFG and CC (n = 3 replicate donor pools with two donors per pool). Thermal shift assays were performed with increasing concentrations of Martinostat (0, 0.0032, 0.016, 0.080, 0.40, 2.0, and 10 μM). Thermal stabilization of HDACs 1, 2, 3, 6, and 8 was compared through Western blotting with scaled immunoreactive band intensity values represented as an averaged heat map (n = 3). The imaging-derived dissociation constant (Kd) for [11C]Martinostat in NHP brain is indicated by the black arrow (19). See fig. S6 for original Western blotting data. (B) Whole-cell lysates were prepared from postmortem human SFG (n = 3 replicate donor pools with two donors per pool), as well as dorsolateral prefrontal cortex, hippocampus, and anterior cingulate (n = 3 replicate donor pools with three donors per pool). Thermal shift assays were performed with increasing concentrations of Martinostat (0, 0.16, 0.80, 4.0, 20, and 100 μM). Thermal stabilization of HDACs 1, 2, 3, 6, and 8 was compared through Western blotting with scaled immunoreactive band intensity values represented as an averaged heat map (n = 3). See figs. S7 to S9 for original Western blotting data. (C) Baboon brain (n = 1) was sectioned to include gray matter and white matter regions in the same slice. Tissue was coincubated with ~100 μCi of [11C]Martinostat and either 0 or 2 μM nonradiolabeled Martinostat. Grayscale autoradiographic images were colored using a standard lookup table (royal scale in Image J) to reflect [11C]Martinostat intensity (left). Region-specific baseline and blocking intensity values were quantitated from each slice (right). Data are means ± SD (n = 22 0-μM slices, n = 10 2-μM slices; one image per slice; one region of interest per brain region). P values were determined by ordinary two-way ANOVA (α = 0.05 with Sidak’s multiple comparisons correction).

Competition autoradiography was performed in postmortem baboon brain tissue to compare the specific binding of [11C]Martinostat in gray and white matter. [11C]Martinostat binding in white matter was more biased by nonspecific uptake than in gray matter (Fig. 4C). Together, our in vivo imaging and ex vivo biochemistry data indicate that [11C]Martinostat binds to a subset of class I HDACs (isoforms 1, 2, and 3) across the human and baboon brains.

In vitro biochemistry of human neural progenitor cells reveals downstream targets of Martinostat-bound HDACs

To link [11C]Martinostat uptake with downstream HDAC substrate signaling and gene expression, we treated human stem cell–derived neural progenitor cells with increasing concentrations of Martinostat. Acetylation levels of established class I HDAC substrates, histone H3 lysine 9 (H3K9) and histone H4 lysine 12 (H4K12), were determined using Western blotting (3, 24). Treatment with 2.5 and 5.0 μM Martinostat increased H3K9 and H4K12 acetylation levels as compared to vehicle control (Fig. 5A). Treatment with 5.0 μM Martinostat elevated acetylation to a level equivalent to or greater than 10 μM SAHA (Fig. 5A). Messenger RNA (mRNA) transcript levels of memory-related (3, 6, 24), neuroplasticity-related (3), and neurological disease–related genes (17) were measured through quantitative polymerase chain reaction (qPCR). Treatment with 2.5 and/or 5.0 μM Martinostat increased brain-derived neurotrophic factor (BDNF), early growth response protein 1 (EGR1), cyclin-dependent kinase 5 (CDK5), synaptotagmin (SYT1), synaptophysin (SYP), and progranulin (GRN) expression compared to vehicle control, but not frataxin (FXN) (Fig. 5B). Treatment with 2.5 μM Martinostat elevated BDNF and SYP (about 20- and 10-fold, respectively) to a level equivalent to or greater than 10 μM SAHA (Fig. 5B). Together, these results indicate that Martinostat engages the subset HDACs that deacetylate targets including H3K9 and H4K12, to regulate downstream genes important for neuroplasticity (BDNF, EGR1, CDK5, SYT1, SYP, and GRN).

Fig. 5. Martinostat increases histone acetylation and gene expression levels in human neural progenitor cells.

Human neural progenitor cells were treated with DMSO (Veh), Martinostat (MSTAT; 0.5, 2.5, or 5.0 μM), and SAHA (10 μM) for 24 hours. (A) Whole-cell lysates were prepared (n = 3). #Because treatment with 5.0 μM Martinostat was toxic to cells, whole-cell lysates from three replicates were combined into one pool to obtain sufficient protein for this dose. Equivalent amounts of total protein were compared through Western blotting. Histone acetylation immunoreactive band intensity values were normalized to GAPDH intensity values. Data are means ± SD (n = 3). P values compare drug treatments to Veh, determined by repeated-measures two-way ANOVA (α = 0.05 with Dunnett’s multiple comparisons correction). (B) RNA was extracted (n = 3) and converted into complementary DNA (cDNA). mRNA transcript levels of memory/neuronal plasticity–related (BDNF, EGR1, CDK5, SYT1, and SYP) and monogenic neurological disorder–related (GRN and FXN) genes were compared through qPCR and normalized to GAPDH mRNA levels. Data are means ± SEM (n = 3 cDNA per condition with three technical qPCR replicates per cDNA). P values compare drug treatments to Veh, determined by repeated-measures two-way ANOVA (α = 0.05 with Dunnett’s multiple comparisons correction).

DISCUSSION

This first-in-human epigenetic imaging study with [11C]Martinostat establishes that HDACs are highly expressed throughout the healthy brain with region-specific distribution, including distinct differences between gray and white matter and differences between cortical and subcortical gray matter regions. On the basis of our previous in vitro profiling with recombinant HDACs (18) and our ex vivo profiling with postmortem human and baboon brain tissues, the [11C]Martinostat signal in the brain originated from binding class I HDACs (isoforms 1, 2, and 3), which are relevant to cognition, memory, and mood regulation (3, 13, 16, 25). Notably, Martinostat stabilized these isoforms at a concentration of ~0.1 μM, which is consistent with the imaging-derived dissociation constant (Kd) for [11C]Martinostat in the NHP brain (18). In contrast with previous in vitro recombinant inhibition data (18), Martinostat did not appear to stabilize HDAC6 in the brain regions that we assessed, although it is worth noting that the recombinant assay provided a more than fivefold lower median inhibitory concentration for HDAC6 when compared to isoforms 1, 2, and 3. At high concentrations, α-tubulin acetylation may be increased by Martinostat, thus implicating potential HDAC6 binding at therapeutic-relevant concentrations.

Our imaging data revealed that in vivo HDAC expression is higher in cortical gray matter than in white matter, which was confirmed for HDAC2 and HDAC3 by postmortem human tissue analyses. We postulate that HDAC complexes in each brain tissue type may affect the selectivity of Martinostat and other HDAC inhibitors, including those currently used as U.S. Food and Drug Administration–approved drugs. HDAC complex–directed selectivity of HDAC inhibitors has been shown previously through chemoproteomic approaches (26, 27), and additional work will be required to elucidate the HDAC complexes most represented by the [11C]Martinostat signal.

Beyond regional differences in HDAC distribution, the most striking observation was the consistency of [11C]Martinostat binding patterns between individual subjects. Because epigenetic machinery, and thus HDAC expression, is a highly dynamic process, we did not fully expect a spatially conserved pattern of HDAC expression between individuals. This result not only suggests that HDAC expression is tightly regulated and may represent a state function, but also reiterates the importance of localized levels of HDACs as they directly relate to gene transcription (1). We anticipate that regional [11C]Martinostat uptake differences between healthy and diseased individuals will be detectable given the conversed baseline expression that we have measured. The use of [11C]Martinostat imaging may eventually enable precision medicine approaches for disease stratification and treatment based on epigenetic aberrations in the human brain. As hippocampal HDAC2 overexpression has been found in postmortem brain tissue from AD patients (6), [11C]Martinostat PET imaging holds great potential for detecting aberrant hippocampal HDAC expression and assessing novel HDAC therapeutics in AD patients.

Because we envision and will apply [11C]Martinostat to measure HDAC expression in patient populations, it is critical that the outcome measurements are reliable, reproducible, and noninvasive. By comparing the standard deviation of the mean of VT and SUV60-90 min across brain regions, we found that intersubject variability was smaller using SUV60-90 min analysis than VT. These results support the use of SUV60-90 min in future studies to eliminate arterial blood sampling when patient enrollment would be limited by the invasiveness and risk of this procedure. Perhaps as important, PET studies with [11C]Martinostat may be sufficiently powered with a smaller sample size when SUV60-90 min is chosen as the outcome measurement instead of VT. However, validation studies will be required to evaluate whether SUVs are appropriate surrogates for VT values in different patient populations.

To begin to connect HDAC imaging with [11C]Martinostat to gene regulation in the human brain, we compared mRNA transcript level changes elicited by pharmacologically relevant doses of Martinostat in human stem cell–derived neural progenitor cells. The concentrations of Martinostat used to treat neural progenitor cells were ~1000-fold higher than tracer-level doses used for in vivo [11C]Martinostat imaging. Tracer-level doses are intended to achieve low occupancy and thus should not perturb HDAC enzyme activity and downstream gene expression. However, by using pharmacologically relevant doses for neural progenitor cell studies, the downstream targets of Martinostat-bound HDACs were revealed and provide insight into imaging signal interpretations. For example, these data suggest that in regions where [11C]Martinostat binding in the human brain is lowest, such as the hippocampus, the levels of HDAC-regulated genes, such as BDNF, are elevated. The hippocampus was previously shown to be consistently enriched in BDNF (17, 2831).

Besides genes implicated in memory and neuroplasticity, Martinostat enhanced the mRNA expression of GRN encoding the glycoprotein progranulin. GRN mutations are a major cause of autosomal dominant frontotemporal lobar degeneration (14). The demonstration here that Martinostat treatment increases GRN mRNA levels supports the value of HDAC-targeted therapies as a disease-modifying treatment for this type of dementia. Moreover, because HDAC inhibitors are the subject of current clinical investigation for frontotemporal lobar degeneration, measuring HDAC expression in the human brain with [11C]Martinostat imaging may provide a critically needed tool for determining optimal doses of therapeutics and for patient stratification should levels of HDACs change in the disease state.

We recognize several limitations in our current study. First, the imaging data presented here are from a cohort of eight healthy subjects and thus we cannot characterize changes in “normal” HDAC expression (for example, as a function of age). Future studies will expand our imaging cohort to include more healthy subjects as well as multiple HDAC dysfunction-associated patient populations, including AD and schizophrenia, to investigate the in vivo relevance of HDAC expression in neurological and psychiatric diseases. Another limitation is that quantitative HDAC levels in postmortem brain tissue are relative to recombinant HDAC standards and do not reflect the absolute values of HDAC expression in the living brain, as postmortem HDAC levels may be affected by artifacts such as postmortem interval. Additionally, owing to the low throughput of thermal shift assays with Western blot–based detection and limited tissue availability, we found it necessary to pool multiple postmortem brain samples into three biological replicates, rather than analyze individual thermal shift assays for each donor, which may have revealed a higher variability of Martinostat selectivity. Last, neural progenitor cell studies uncovered only a subset of downstream Martinostat-bound HDAC substrates and gene targets. Future studies using acetyl proteomic profiling, RNA sequencing, and chemoproteomics are needed to fully understand the biological pathways detected by [11C]Martinostat.

In conclusion, this first-in-human epigenetic imaging study reveals that HDACs are highly expressed throughout the healthy brain with a conserved regional distribution between individuals. Our study uncovers region-specific variations in HDAC inhibitor binding, which we postulate is due to differences between the HDAC complex identities in those regions. Together, our neuroimaging and biochemical experiments provide a critical foundation for how to quantify epigenetic activity in the living brain and in turn accomplish HDAC inhibition in the CNS as a therapy for human brain disorders.

MATERIALS AND METHODS

Study design

Our main research objective was to quantify in vivo regional HDAC expression in the healthy human brain, using [11C]Martinostat PET. As a first-in-human PET imaging study, a cohort of eight individuals was included to evaluate intrasubject and intersubject variability of [11C]Martinostat uptake. These are critical information for appropriate power calculations when designing future studies. No data from the eight subjects were excluded as outliers for [11C]Martinostat uptake values. Regional VT and SUVs were the image-based endpoints assessed. We furthered our imaging findings through ex vivo biochemistry to ascribe the HDAC subtype selectivity of Martinostat using human and NHP brain tissues. Thermal shift and HDAC expression level assays included three biological replicates, with lysates pooled from two to three human donors per replicate. Nuclear density and autoradiographic assays included a minimum of four NHP brain slices per region from one baboon (Papio anubis). For these assays, we excised a contiguous section of baboon brain spanning a gray and white matter boundary to remove external variables from our analyses. The availability of an intact baboon brain is very rare; thus, we were only able to access one biological replicate through multiple slices. We also furthered our imaging findings through in vitro analyses of Martinostat-dependent substrate acetylation and gene expression levels. Acetylation and mRNA profiling assays included three biological replicates of human neural progenitor cells. Imaging and biochemical studies were not blinded.

Participants

Eight participants (four females and four males; mean age ± SD, 28.6 ± 7.6 years) were included in this study (eIND #123154). Participants were healthy volunteers with no history of hepatic, renal, neurological, or psychiatric disease and were not taking any prescription medication, as evaluated by medical examinations. Participants had not smoked tobacco products within the past 5 years and were not using any illicit drugs, as assessed by a urine drug test (Discover 12 Panel Test Card, American Screening Corp). Additionally, a serum pregnancy test (Sure-Vue serum hCG-STAT, Fisher HealthCare) was performed for female participants to ensure no pregnancy at the time of the scan. Participants provided written informed consent to take part in the study, which was approved by the Institutional Review Board and the Radioactive Drug Research Committee at Massachusetts General Hospital. Volunteers were compensated for their participation in the study.

Radiosynthesis of [11C]Martinostat

[11C]Martinostat was synthesized as described in Supplementary Methods.

MR/PET imaging

Participants had no magnetic resonance imaging (MRI) or PET contraindications to safely undergo brain imaging. An arterial line (A-line) was placed in the radial artery of one arm, and an intravenous catheter was placed in the antecubital vein of the other arm. A licensed nuclear medicine technologist administered [11C]Martinostat into the intravenous catheter as a manual bolus, and an experienced nurse practitioner drew blood samples from the A-line during the scan to determine plasma radioactivity and radioactive metabolites. Participants were instructed to remain still for the total duration of each scan. PET and MRI images were acquired on a 3T Siemens TIM Trio with a BrainPET insert (Siemens). A PET-compatible circularly polarized transmit coil and an eight-channel receive array coil were used for MRI data acquisition. A high-resolution anatomical scan using a multi-echo MPRAGE (magnetization-prepared rapid acquisition gradient echo) sequence [repetition time (TR), 2530 ms; echo time 1 (TE1), 1.64 ms; TE2, 3.49 ms; TE3, 5.35 ms; TE4, 7.21 ms; inversion time (TI), 1200 ms; flip angle, 7°; and isotropic resolution, 1 mm] was acquired.

Dynamic PET image acquisition was initiated concomitant with the start of intravenous bolus injection of ~5 mCi (4.8 ± 0.4 mCi for the eight scans) [11C]Martinostat to the subject. PET data were acquired for 90 min, stored in list mode format, and binned into 28 frames of progressively longer duration (10 s × 8, 20 s × 3, 30 s × 2, 60 s × 1, 120 s × 1, 180 s × 1, 300 s × 8, and 600 s × 4). The corresponding images were reconstructed using the three-dimensional ordinary Poisson ordered-subset expectation maximization (3D OP-OSEM) algorithm with detector efficiency, decay, dead time, attenuation, and scatter corrections applied. The attenuation correction map was derived using a Statistical Parametric Mapping (SPM)–based, pseudo–computed tomography method (32), which combines segmentation and atlas-based approaches. Simultaneously collected MR sequences consisting of an echo-planar imaging readout were used to measure subject motion during the scan and an MR-based motion correction was applied to the PET data (33). The final PET images were reconstructed into 153 slices with 256 × 256 pixels and a 1.25-mm isotropic voxel size, in the units of radioactivity concentrations (becquerels per milliliter) and SUVs (mean radioactivity per injected dose per weight). Three subjects completed a second PET scan, which was accomplished 3 hours after the first scan on the same day, using identical imaging methods.

Image analyses

Dynamic PET data were motion-corrected to a late time point image (frame 20; 39 to 44 min after radiotracer injection) of the time series using rigid body linear registration (6 df) implemented in FSL [FMRIB (Oxford Centre for Functional MRI of the Brain) Software Library] (MCFLIRT) (34). A PET mean image from the motion-corrected time series was calculated for each subject and registered and resampled to the subject’s T1-weighted structural scan (MPRAGE) using spmregister from FreeSurfer (http://surfer.nmr.mgh.harvard.edu) (35). The PET mean image was further registered to the Montreal Neurological Institute (MNI) space using a linear [FLIRT (FMRIB’s linear image registration tool)] and a nonlinear [FNIRT (FMRIB’s nonlinear image registration tool)] algorithm implemented in FSL (http://fsl.fmrib.ox.ac.uk/fsl) (36). Finally, dynamic PET images (in both radioactivity concentration and SUV units) were normalized to the MNI space, using a combined transformation matrix derived from the PET mean image, for further analyses.

Kinetic modeling was performed using PMOD 3.4 (PMOD Technologies Ltd). Twenty-eight VOIs were defined according to the Automated Anatomical Labeling human brain atlas distributed with PMOD (37). A two-tissue compartmental model was applied to the regional time-activity curves (TACs) extracted from the VOIs and using the metabolite-corrected arterial plasma as input function to derive VT and microparameters describing the pharmacokinetics of the radiotracer (table S3). The following equation was used for compartmental model fitting:Embedded ImageEmbedded Imagewhere CP is the arterial input function, CND represents the nondisplaceable compartment, CS represents the specific binding compartment, and CND + CS is the radioactivity that we measured with PET.

The minimum scan duration required for stable VT value estimation was also evaluated (fig. S2). An averaged SUV image (SUV60-90 min) was calculated from 60 to 90 min after radiotracer injection. Regional cortical VT and SUV60-90 min values were combined for cortical lobes using a weighted average to reduce the total number of VOIs (resulting in a total of 14 VOIs). Voxel-wise, group mean, and standard deviation maps of the SUV60-90 min were calculated and overlaid on an MNI152 template brain after spatial smoothing with a 6-mm full width at half maximum Gaussian filter (Fig. 2A). In addition, SUV60-90 min values were normalized to individual subjects’ white matter SUV60-90 min (SUVR60-90 min) to evaluate intersubject variability for different VOIs (Fig. 2B).

Human tissue samples

Postmortem frozen human brain tissue was obtained from the National Institutes of Health (NIH) NeuroBioBank; specifically, tissue was obtained from the Harvard Brain Tissue Resource Center, University of Miami Brain Endowment Bank, Human Brain and Spinal Fluid Resource Center, and Brain Tissue Donation Program at the University of Pittsburgh Medical Center. For all donors, informed consent was obtained from next of kin. Donor brains had a neuropathology diagnosis of normal (table S5). Tissue lysates were prepared as described in Supplementary Methods.

HDAC expression levels

Known concentrations of recombinant HDAC enzymes (Reaction Biology Corp KDA-21-365, KDA-21-277, and KDA-21-213; Abcam ab82071) were diluted in twofold increments and compared to 12 μg of total protein from human brain lysates (n = 3 replicate pools per region) through Western blotting. Notably, recombinant HDAC2 contained a GST tag, which increased its detected size. Immunoreactive band intensity was quantified with ImageJ (Image Processing and Analysis in Java, NIH) (38). Standard curves were calculated for each recombinant HDAC isoform with GraphPad Prism software, and the concentrations of HDACs per lane of lysate were determined.

Histone acetylation changes in human neural progenitor cells with Martinostat and SAHA

Human induced pluripotent stem cell–derived neural progenitor cells from a healthy control subject fibroblast cell line GM08330 (Coriell Institute for Medical Research) were generated as described in (39) and cultured as described in (40) and Supplementary Methods. Cell pellets from human neural progenitor cells (n = 3 per condition) were lysed in radioimmunoprecipitation assay (RIPA) buffer (Boston BioProducts #BP-115) with EDTA-free protease inhibitors (Sigma #4693159001) and rocked at 4°C for 30 min. The lysate was centrifuged at 14,000 rpm at 4°C for 25 min, and the supernatant was collected. Protein quantification was determined by a bicinchoninic acid assay (Thermo Scientific #23227). Lysates were diluted to 800 ng/μl in RIPA buffer and stored at −80°C until ready for use. H3K9 and H4K12 acetylation levels were measured by Western blotting. Mean immunoreactive band intensities from each replicate were quantified with ImageJ. Gene expression changes were determined as described in Supplementary Methods.

Statistical analyses

Statistical tests were performed using GraphPad Prism (Prism6, GraphPad Software Inc.). For PET imaging analyses, a nonparametric Friedman test (α = 0.05 with Dunn’s multiple comparisons correction) was carried out to compare SUV60-90 min between brain regions (Fig. 2 and table S4). A Pearson correlation analysis was performed between VT and SUV60-90 min values for the 14 VOIs (Fig. 2B) to evaluate whether an image-based outcome measurement (SUV60-90 min) is an appropriate surrogate to that estimated with the full kinetic modeling data (VT). Differences in postmortem HDAC expression levels as well as differences in nuclear density, size, and total area between the SFG and the CC were evaluated with an unpaired t test (Fig. 3B and fig. S5). Differences in postmortem HDAC expression levels between the dorsolateral prefrontal cortex, hippocampus, and anterior cingulate were evaluated with an ordinary one-way ANOVA (α = 0.05 with Tukey’s multiple comparisons correction) (Fig. 3C). Differences in histone acetylation and gene expression levels as compared to vehicle were evaluated with a repeated-measures two-way ANOVA (α = 0.05 with Dunnett’s multiple comparisons correction) (Fig. 5). In autoradiographic assays, differences between [11C]Martinostat baseline and blocking intensity values, in gray matter and white matter, were evaluated with an ordinary two-way ANOVA (α = 0.05 with Sidak’s multiple comparisons correction) (Fig. 4C).

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/8/351/351ra106/DC1

Methods

Fig. S1. TACs and compartmental model fitting (two-tissue compartmental model) results for superior frontal cortex and white matter.

Fig. S2. Stability of outcome measurement (VT) as a function of scan duration.

Fig. S3. Regional SUV60-90 min from all brain regions analyzed.

Fig. S4. Same day test-retest reproducibility of [11C]Martinostat SUVR60-90 min.

Fig. S5. Nuclear density, size, and total area in postmortem baboon brain tissue.

Fig. S6. Martinostat thermal shift assay in human SFG and CC biological replicates 1, 2, and 3.

Fig. S7. Martinostat thermal shift assay across human gray matter biological replicate 1.

Fig. S8. Martinostat thermal shift assay across human gray matter biological replicate 2.

Fig. S9. Martinostat thermal shift assay across human gray matter biological replicate 3.

Table S1. Biometric information for PET imaging participants.

Table S2. Goodness of fit for one- and two-tissue compartmental models to regional PET data.

Table S3. Kinetic rate constants and regional VT for [11C]Martinostat.

Table S4. Statistical comparison of [11C]Martinostat between different brain regions.

Table S5. Sample information for postmortem human brain tissue.

REFERENCES AND NOTES

  1. Acknowledgments: We are grateful to U. Mahmood, S. Stufflebeam, and O. Johnson-Akeju for consenting participants, and to E. Pierce and O. Johnson-Akeju for placing the arterial line in participants. We thank J. Sore, G. Gautam, K. Phan, and S. To for technical assistance in radiotracer synthesis and G. Arabasz, S. Hsu, M. Wentworth, and R. Butterfield for assistance with MR/PET imaging. We also thank G. Van de Bittner and M. Riley for assistance with autoradiographic experiments and L. Rogers for technical assistance with HDAC density experiments. Postmortem tissue was obtained from the NIH NeuroBioBank (requests #100 and #250). Funding: This research received funding from the National Institute on Drug Abuse of the NIH under grant numbers R01DA030321 (to J.M.H.) and K99DA037928 (to H-Y.W.). This research was also supported by the Harvard/MGH Nuclear Medicine Training Program from the Department of Energy under grants DE-SC0008430 (to H-Y.W., T.M.G., and C.W.) and HHSN-271-2013-00030C (to the Harvard Brain Tissue Resource Center). This research was carried out at the Athinoula A. Martinos Center for Biomedical Imaging at Massachusetts General Hospital, using resources provided by the Center for Functional Neuroimaging Technologies, P41EB015896, a P41 Biotechnology Resource Grant supported by the National Institute of Biomedical Imaging and Bioengineering, NIH. This work was conducted with support from Harvard Catalyst and the Harvard Clinical and Translational Science Center (National Center for Research Resources and the National Center for Advancing Translational Sciences, NIH Award UL1 TR001102) and financial contributions from Harvard University and its affiliated academic healthcare centers. Additional support was provided by the Bluefield Project to Cure Frontotemporal Dementia. This work also involved the use of instrumentation supported by the NIH Shared Instrumentation Grant Program, specifically grants S10RR017208, S10RR026666, S10RR022976, S10RR019933, and S10RR023401. Author contributions: H-Y.W., T.M.G., S.J.H., C.W., and J.M.H. designed the study. H-Y.W., N.R.Z., and A.B. collected in vivo human imaging data. H-Y.W. and N.R.Z. analyzed in vivo human imaging data. T.M.G., B.D.T., and F.A.S. collected ex vivo human biochemical data. T.M.G. collected ex vivo NHP biochemical data. T.M.G. and J.M.H. analyzed ex vivo human and NHP biochemical data. A.S. collected in vitro human NPC data. A.S., S.J.H., and T.M.G. analyzed in vitro human NPC data. H-Y.W., T.M.G., N.R.Z., and A.S. performed statistical analyses. H-Y.W., T.M.G., N.R.Z., A.S., A.B., F.A.S., C.W., S.J.H., and J.M.H. wrote and edited the manuscript. Competing interests: The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University, and its affiliated academic healthcare centers, or the NIH. Intellectual property (IP) has been filed around [11C]Martinostat by J.M.H., C.W., and F.A.S. A portion of this IP has been licensed. S.J.H. has financial interests in Rodin Therapeutics and is an inventor on HDAC inhibitor-related IP licensed to this entity that is unrelated to the present study. Data and materials availability: Tissues were provided by the Harvard Brain Tissue Resource Center, University of Miami Brain Endowment Bank, Human Brain and Spinal Fluid Resource Center, and Brain Tissue Donation Program at the University of Pittsburgh Medical Center.
View Abstract

Navigate This Article