Impact of histone deacetylase inhibition and arimoclomol on heat shock protein expression and disease biomarkers in primary culture models of familial ALS.

Protein misfolding and mislocalization are common themes in neurodegenerative disorders, including motor neuron disease, and amyotrophic lateral sclerosis (ALS). Maintaining proteostasis is a crosscutting therapeutic target, including the upregulation of heat shock proteins (HSP) to increase chaperoning capacity. Motor neurons have a high threshold for upregulating stress-inducible HSPA1A, but constitutively express high levels of HSPA8. This study compared the expression of these HSPs in cultured motor neurons expressing three variants linked to familial ALS: TAR DNA binding protein 43 kDa (TDP-43)G348C, fused in sarcoma (FUS)R521G, or superoxide dismutase I (SOD1)G93A. All variants were poor inducers of Hspa1a, and reduced levels of Hspa8 mRNA and protein, indicating multiple compromises in chaperoning capacity. To promote HSP expression, cultures were treated with the putative HSP coinducer, arimoclomol, and class I histone deacetylase inhibitors, to promote active chromatin for transcription, and with the combination. Treatments had variable, often different effects on the expression of Hspa1a and Hspa8, depending on the ALS variant expressed, mRNA distribution (somata and dendrites), and biomarker of toxicity measured (histone acetylation, maintaining nuclear TDP-43 and the neuronal Brm/Brg-associated factor chromatin remodeling complex component Brg1, mitochondrial transport, FUS aggregation). Overall, histone deacetylase inhibition alone was effective on more measures than arimoclomol. As in the FUS model, arimoclomol failed to induce HSPA1A or preserve Hspa8 mRNA in the TDP-43 model, despite preserving nuclear TDP-43 and Brg1, indicating neuroprotective properties other than HSP induction. The data speak to the complexity of drug mechanisms against multiple biomarkers of ALS pathogenesis, as well as to the importance of HSPA8 for neuronal proteostasis in both somata and dendrites.

Reversal of cognitive deficits in FUSR521G amyotrophic lateral sclerosis mice by arimoclomol and a class I histone deacetylase inhibitor independent of heat shock protein induction.

Protein misfolding and mislocalization are common to both familial and sporadic forms of amyotrophic lateral sclerosis (ALS). Maintaining proteostasis through induction of heat shock proteins (HSP) to increase chaperoning capacity is a rational therapeutic strategy in the treatment of ALS. However, the threshold for upregulating stress-inducible HSPs remains high in neurons, presenting a therapeutic obstacle. This study used mouse models expressing the ALS variants FUSR521G or SOD1G93A to follow up on previous work in cultured motor neurons showing varied effects of the HSP co-inducer, arimoclomol, and class I histone deacetylase (HDAC) inhibitors on HSP expression depending on the ALS variant being expressed. As in cultured neurons, neither expression of the transgene nor drug treatments induced expression of HSPs in cortex, spinal cord or muscle of FUSR521G mice, indicating suppression of the heat shock response. Nonetheless, arimoclomol, and RGFP963, restored performance on cognitive tests and improved cortical dendritic spine densities. In SOD1G93A mice, multiple HSPs were upregulated in hindlimb skeletal muscle, but not in lumbar spinal cord with the exception of HSPB1 associated with astrocytosis. Drug treatments improved contractile force but reduced the increase in HSPs in muscle rather than facilitating their expression. The data point to mechanisms other than amplification of the heat shock response underlying recovery of cognitive function in ALS-FUS mice by arimoclomol and class I HDAC inhibition and suggest potential benefits in counteracting cognitive impairment in ALS, frontotemporal dementia and related disorders.

Strength and muscle specificity of a compact promoter derived from the slow troponin I gene in the context of episomal (gutless adenovirus) and integrating (lentiviral) vectors.

BACKGROUND: Gutless adenovirus (helper-dependent adenoviral vector; HDAd) and lentiviral vectors (LV) are attractive vectors for the gene therapy of muscle diseases. Because the organization of their DNA (episomal versus integrated) differs, we investigated whether the strength and specificity of ΔUSEx3, a novel muscle-specific promoter previously tested with plasmid, were maintained in the context of these vectors. METHODS: Two HDAds expressing ß-galactosidase regulated by ΔUSEx3 or CAG [cytomegalovirus (CMV) enhancer/ß-actin promoter], and three LV expressing green fluorescent protein regulated by ΔUSEx3, CMV or a modified skeletal α-actin promoter (SPcΔ5-12), were constructed. Gene expression was compared in cell culture and after intravenous (HDAd only) and intramuscular injection of mice. RESULTS: Irrespective of the vector used, ΔUSEx3 remained poorly active in nonmuscle cells and tissues. In myotubes, ΔUSEx3 was as strong as CMV and SPcΔ5-12, although it was ten-fold weaker than CAG, a proven powerful promoter in muscle. In cell culture, ΔUSEx3 activity in the context of LV was more stable than CMV, indicating it is less prone to silencing. In the context of HDAd, the behavior of ΔUSEx3 in skeletal muscle mirrored that of cell culture (10% of the CAG activity and half the number of transduced fibers). Surprisingly, in muscles treated with LV, ΔUSEx3 activity was five-fold lower than SPcΔ5-12. CONCLUSIONS: The data obtained in the present study confirm that ΔUSEx3 is a strong and robust muscle-specific promoter in the context of HDAd (cell culture and in vivo) and LV (cell culture). However, it was less efficient in vivo in the context of LV.

Genomic integration of adenoviral gene transfer vectors following transduction of fertilized mouse oocytes.

Adenoviral vectors (AdV) are popular tools to deliver foreign genes into a wide range of cells. They have also been used in clinical gene therapy trials. Studies on AdV-mediated gene transfer to mammalian oocytes and transmission through the germ line have been reported controversially. In the present study we investigated whether AdV sequences integrate into the mouse genome by microinjecting AdV into the perivitelline space of fertilized oocytes. We applied a newly developed PCR technique (HiLo-PCR) for identification of chromosomal junctions next to the integrated AdV. We demonstrate that mouse oocytes can be transduced by different recombinant adenoviral vectors (first generation and gutless). In one transgenic mouse line using the first generation adenoviral vector, the genome has integrated into a highly repetitive cluster located on the Y chromosome. While the transgene (GFP) was expressed in early embryos, no expression was detected in adult transgenic mice. The use of gutless AdV resulted in expression of the transgene, albeit the vector was not transmitted to progeny. These results indicate that under optimized conditions fertilized mouse oocytes are transduced by AdV and give rise to transgenic founder animals. Therefore, adequate precautions should be taken in gene therapy protocols of reproductive patients since transduction of oocytes or early embryos and subsequent chromosomal integration cannot be ruled out entirely.

Modulation of coxsackie and adenovirus receptor expression for gene transfer to normal and dystrophic skeletal muscle.

BACKGROUND: Efficient adenovirus (AdV)-mediated gene transfer is possible only in immature muscle or regenerating muscle, suggesting that a developmentally regulated event plays a major role in limiting AdV uptake in mature skeletal muscle. Previously, we showed that the expression of the primary coxsackie and adenovirus receptor (CAR) is severely down-regulated during muscle maturation and that, in muscle-specific CAR transgenic mice, there is significant enhancement of AdV-mediated gene transfer to mature skeletal muscle. METHODS: To evaluate whether increasing CAR expression can also augment gene transfer to dystrophic muscle that has many regenerating fibers, we crossed CAR transgenics with dystrophin-deficient mice (mdx/CAR). We also tested a two-step protocol in which CAR levels were increased in the target muscle, prior to administration of AdV, through the use of recombinant adeno-associated virus (AAV2) expressing CAR. Lastly, we assessed the effect of histone deacetylase inhibitors on CAR and AdV transduction efficiency in myoblasts and mdx muscle. RESULTS: Although somewhat higher rates of transduction can be achieved in adult mdx mice than in normal mice as a result of ongoing muscle regeneration in these animals, CAR expression in the mdx background (mdx/CAR transgenics) still markedly improved the susceptibility of mature muscle to AdV-mediated gene transfer of dystrophin. Prior administration of AAV2-CAR to normal muscle led to significantly increased transduction by subsequent injection of AdV. The histone deacetylase inhibitor valproate increased CAR transcript and protein levels in myoblasts and mdx muscle, and also increased AdV-mediated gene transfer. CONCLUSIONS: We have developed a method of increasing CAR levels in both normal and regenerating muscle.

Depending on the stress, histone deacetylase inhibitors act as heat shock protein co-inducers in motor neurons and potentiate arimoclomol, exerting neuroprotection through multiple mechanisms in ALS models.

Upregulation of heat shock proteins (HSPs) is an approach to treatment of neurodegenerative disorders with impaired proteostasis. Many neurons, including motor neurons affected in amyotrophic lateral sclerosis (ALS), are relatively resistant to stress-induced upregulation of HSPs. This study demonstrated that histone deacetylase (HDAC) inhibitors enable the heat shock response in cultured spinal motor neurons, in a stress-dependent manner, and can improve the efficacy of HSP-inducing drugs in murine spinal cord cultures subjected to thermal or proteotoxic stress. The effect of particular HDAC inhibitors differed with the stress paradigm. The HDAC6 (class IIb) inhibitor, tubastatin A, acted as a co-inducer of Hsp70 (HSPA1A) expression with heat shock, but not with proteotoxic stress induced by expression of mutant SOD1 linked to familial ALS. Certain HDAC class I inhibitors (the pan inhibitor, SAHA, or the HDAC1/3 inhibitor, RGFP109) were HSP co-inducers comparable to the hydroxyamine arimoclomol in response to proteotoxic stress, but not thermal stress. Regardless, stress-induced Hsp70 expression could be enhanced by combining an HDAC inhibitor with either arimoclomol or with an HSP90 inhibitor that constitutively induced HSPs. HDAC inhibition failed to induce Hsp70 in motor neurons expressing ALS-linked mutant FUS, in which the heat shock response was suppressed; yet SAHA, RGFP109, and arimoclomol did reduce loss of nuclear FUS, a disease hallmark, and HDAC inhibition rescued the DNA repair response in iPSC-derived motor neurons carrying the FUSP525Lmutation, pointing to multiple mechanisms of neuroprotection by both HDAC inhibiting drugs and arimoclomol.

Downregulation of CD46 during muscle differentiation: implications for gene transfer to human skeletal muscle using group B adenoviruses.

Adenoviral vectors that use the coxsackievirus and adenovirus receptor do not transduce mature muscle efficiently. Group B adenoviruses use CD46 as their cell attachment receptor. To evaluate the utility of vectors based on group B adenoviruses for gene transfer to human skeletal muscle, we assessed the expression of CD46 in biopsied normal skeletal muscle samples and in muscles from patients with Duchenne muscular dystrophy. Transcript levels of CD46 were extremely low in mature muscle and CD46 immunoreactivity was detected only on blood vessels in the muscle sections. Although myoblasts cultured from biopsied samples had robust cell surface CD46 expression by flow cytometry, CD46 transcript levels were barely detectable after differentiation of the myoblasts into myotubes. The myotubes were also much less susceptible to infection with an adenoviral vector carrying the fiber of serotype 35 adenovirus (AdF35). These results suggest that for skeletal muscle, vectors derived from group B adenoviruses may not be a suitable alternative to the commonly used Ad5 vectors.

Strategies for muscle-specific targeting of adenoviral gene transfer vectors.

Currently, adenoviral transfer of therapeutic genes such as dystrophin is hampered by low transduction efficiency of adult skeletal muscle. This is largely due to the lack of appropriate virus attachment receptors on the myofiber surface. Recent studies in transgenic mice revealed that upregulation of Coxsackie- and adenovirus receptor improves gene transfer efficiency by approximately ten-fold. Conversely, the vector load that needed to be administered to achieve sufficient gene transfer could be lowered significantly. Reduced viral vector loads may help to control virally mediated toxicity and immunogenicity. To date, there are no drugs or methods known to increase Coxsackie- and adenovirus receptor expression in skeletal muscle that would be easily applicable in humans. However, alternative strategies such as vector retargeting are currently being investigated that may allow for an increase in binding of adenoviral vectors to skeletal muscle. Recent experiments have shown that directed mutagenesis of the adenoviral fiber knob allows for a significant reduction in Coxsackie- and adenovirus receptor binding and for introduction of a new binding domain. Therefore, vector retargeting towards efficient and specific infection of skeletal muscle may be achieved by directed genetic alteration of adenoviral capsid proteins.

Inhibition of the protein tyrosine phosphatase SHP-1 increases glucose uptake in skeletal muscle cells by augmenting insulin receptor signaling and GLUT4 expression.

The protein tyrosine phosphatase (PTPase) Src-homology 2-domain-containing phosphatase (SHP)-1 was recently reported to be a novel regulator of insulin's metabolic action. In order to examine the role of this PTPase in skeletal muscle, we used adenovirus (AdV)-mediated gene transfer to express an interfering mutant of SHP-1 [dominant negative (DN)SHP-1; mutation C453S] in L6 myocytes. Expression of DNSHP-1 increased insulin-induced Akt serine-threonine kinase phosphorylation and augmented glucose uptake and glycogen synthesis. Pharmacological inhibition of glucose transporter type 4 (GLUT4) activity using indinavir and GLUT4 translocation assays revealed an important role for this transporter in the increased insulin-induced glucose uptake in DNSHP-1-expressing myocytes. Both GLUT4 mRNA and protein expression were also found to be increased by DNSHP-1 expression. Furthermore, AdV-mediated delivery of DNSHP-1 in skeletal muscle of transgenic mice overexpressing Coxsackie and AdV receptor also enhanced GLUT4 protein expression. Together, these findings confirm that SHP-1 regulates muscle insulin action in a cell-autonomous manner and further suggest that the PTPase negatively modulates insulin action through down-regulation of both insulin signaling to Akt and GLUT4 translocation, as well as GLUT4 expression.

Efficient and fast functional screening of microdystrophin constructs in vivo and in vitro for therapy of duchenne muscular dystrophy.

Duchenne muscular dystrophy (DMD) is an X-linked, lethal genetic disorder affecting the skeletal muscle compartment, and is caused by mutation(s) in the dystrophin gene. Gene delivery of microdystrophin constructs using adeno-associated virus (AAV) and antisense-mediated exon skipping restoring the genetic reading frame are two of the most promising therapeutic strategies for DMD. Both approaches use microdystrophin proteins either directly as a desired construct for gene delivery, using the capacity-limited AAV vectors, or as the therapeutic outcome of gene splicing. Although functionality of the resulting artificial dystrophin proteins can be predicted in silico, experimental evidence usually obtained in transgenic mice is required before human trials. However, the enormous number of potential constructs makes screening assays for dystrophin protein function in vitro and in vivo highly desirable. Here we present data showing that functionality of microdystrophins can be assessed using relatively simple and fast techniques.

Successful compensation for dystrophin deficiency by a helper-dependent adenovirus expressing full-length utrophin.

Helper-dependent adenovirus vector (AdV)-mediated full-length dystrophin expression leads to significant mitigation of the dystrophic phenotype of the mdx mouse. However, dystrophin, as a neoantigen, elicits antibody formation. As an alternative approach, we evaluated gene transfer of full-length murine utrophin, a functional homologue of dystrophin that is normally present only at the neuromuscular junction. A single injection in the tibialis anterior (TA) muscle of the helper-dependent adenovirus vector encoding utrophin provided very good transduction, with 58% of fibers demonstrating sarcolemmal utrophin expression in the neonates, and 35% utrophin-positive (Utr(+)) fibers in adults. The presence of utrophin prevented extensive necrosis in the neonates, halted further necrosis in the adults, and led to restoration of sarcolemmal expression of dystrophin-associated proteins up to 1 year after injection. Marked physiological improvement was observed in both neonates and adults. Neither increased humoral responses nor cellular immune responses were evident. However, there was a time-related decline of the initial high utrophin expression. Although viral DNA persisted in animals that were injected in the neonatal stage, viral DNA levels decreased in muscles of adult mice. These results demonstrate that although utrophin gene transfer leads to amelioration of the dystrophic phenotype, the effects are not sustained upon loss of utrophin expression.

Simultaneous dystrophin and dysferlin deficiencies associated with high-level expression of the coxsackie and adenovirus receptor in transgenic mice.

The Coxsackie and adenovirus receptor (CAR), a cell adhesion molecule of the immunoglobulin superfamily, is usually confined to the sarcolemma at the neuromuscular junction in mature skeletal muscle fibers. Previously, we reported that adenovirus-mediated gene transfer is greatly facilitated in hemizygous transgenic mice with extrasynaptic CAR expression driven by a muscle-specific promoter. However, in the present study, when these mice were bred to homozygosity, they developed a severe myopathic phenotype and died prematurely. Large numbers of necrotic and regenerating fibers were present in the skeletal muscle of the homozygous CAR transgenics. The myopathy was further characterized by increased levels of caveolin-3 and beta-dystroglycan and decreased levels of dystrophin, dysferlin, and neuronal nitric-oxide synthase. Even the hemizygotes manifested a subtle phenotype, displaying deficits in isometric force generation and perturbed mitogen-activated protein kinase (MAPK-erk1/2) activation during contraction. There are few naturally occurring or engineered mouse lines showing as severe a skeletal myopathy as observed with ectopic expression of CAR in the homozygotes. Taken together, these findings suggest that substantial overexpression of CAR may lead to physiological dysfunction by disturbing sarcolemmal integrity (through dystrophin deficiency), impairing sarcolemmal repair (through dysferlin deficiency), and interfering with normal signaling (through alterations in caveolin-3 and neuronal nitric-oxide synthase levels).

Impact of the coxsackie and adenovirus receptor (CAR) on glioma cell growth and invasion: requirement for the C-terminal domain.

Expression of the coxsackie and adenovirus receptor (CAR) is downregulated in malignant glioma cell lines and is barely detectable in high-grade primary astrocytoma (glioblastoma multiforme). We determined the effect of forced CAR expression on the invasion and growth of the human glioma cell line U87-MG, which does not express any CAR. Although retrovirally mediated expression of full-length CAR in U87-MG cells did not affect monolayer growth in vitro, it did reduce glioma cell invasion in a 3-dimensional spheroid model. Furthermore, in xenograft experiments, intracerebral implantation of glioma cells expressing full-length CAR resulted in tumors with a significantly reduced volume compared to tumors generated by control vector-transduced U87-MG cells. In contrast, U87-MG cells expressing transmembrane CAR with a deletion of the entire cytoplasmic domain (except for the first 2 intracellular juxtamembrane cysteine amino acids) had rates of invasion and tumor growth that were similar to those of the control cells. This difference in behavior between the 2 forms of CAR was not due to improper cell surface localization of the cytoplasmically deleted CAR as determined by comparable immunostaining of unpermeabilized cells, equivalent adenoviral transduction of the cells and similar extent of fractionation into lipid-rich domains. Taken together, these results suggest that the decrease or loss of CAR expression in malignant glioma may confer a selective advantage in growth and invasion to these tumors.

The short MCK1350 promoter/enhancer allows for sufficient dystrophin expression in skeletal muscles of mdx mice.

First-generation adenovirus vectors (AdV) have been used successfully to transfer a human dystrophin minigene to skeletal muscle of mdx mice. In most studies, strong viral promoters such as the cytomegalovirus promoter/enhancer (CMV) were used to drive dystrophin expression. More recently, a short version of the muscle creatine kinase promoter (MCK1350) has been shown to provide muscle-specific reporter gene expression after AdV-mediated gene delivery. Therefore, we generated a recombinant AdV where dystrophin expression is controlled by MCK1350 (AdVMCKdys). AdVMCKdys was injected by the intramuscular route into anterior tibialis muscle of mdx mice shortly after birth. Dystrophin expression was assessed at 20, 30, and 60 days after AdV-injection. At 20 days, muscles of AdVMCKdys-injected mdx mice showed a high number of dystrophin-positive fibers (mean: 365). At 60 days, the number of dystrophin-positive fibers was not only maintained, but increased significantly (mean: 600). In conclusion, MCK1350 allows for sustained dystrophin expression after AdV-mediated gene transfer to skeletal muscle of newborn mdx mice. In contrast to previous studies, where strong viral promoters were used, dystrophin expression driven by MCK1350 peaks at later time points. This may have implications for the future use of muscle-specific promoters for gene therapy of Duchenne muscular dystrophy.

Expression of dystrophin driven by the 1.35-kb MCK promoter ameliorates muscular dystrophy in fast, but not in slow muscles of transgenic mdx mice.

Successful gene therapy of Duchenne muscular dystrophy may require the lifelong expression of a therapeutic gene in all affected muscles. The most promising gene delivery vehicles, viral vectors, suffer from several limitations, including immunogenicity, loss of therapeutic gene expression, and a limited packaging capacity. Therefore, various efforts were previously undertaken to use small therapeutic genes and to place them under the control of a strong and muscle-specific promoter. Here we report the effects of a minidystrophin (6.3 kb) under the control of a short muscle-specific promoter (MCK 1.35 kb) over most of the lifetime (4-20 months) of a transgenic mouse model. Dystrophin expression remained stable and muscle-specific at all ages. The dystrophic phenotype was greatly ameliorated and, most importantly, muscle function in limb muscles was significantly improved not only in young adult but also in aged mice compared to nontransgenic littermates. Dystrophin expression was strong in fast-twitch skeletal muscles such as tibialis anterior and extensor digitorum longus, but weak or absent in heart, diaphragm, and slow-twitch muscles. Additionally, expression was strong in glycolytic but weak in oxidative fibers of fast-twitch muscles. This study may have important implications for the design of future gene therapy trials for muscular dystrophy.

Factors influencing the efficacy, longevity, and safety of electroporation-assisted plasmid-based gene transfer into mouse muscles.

Intramuscular injection of plasmid is a potential alternative to viral vectors for the transfer of therapeutic genes into skeletal muscle fibers. The low efficiency of plasmid-based gene transfer can be enhanced by electroporation (EP) coupled with the intramuscular application of hyaluronidase. We have investigated several factors that can influence the efficiency of plasmid-based gene transfer. These factors include electrical parameters of EP, optimal use of hyaluronidase, age and strain of the host, and plasmid size. Muscles of very young and mature normal, mdx, and immunodeficient mice were injected with plasmids expressing beta-galactosidase, microdystrophin, full-length dystrophin, or full-length utrophin. Transfection efficiency, muscle fiber damage, and duration of transgene expression were analyzed. The best transfection level with the least collateral damage was attained at 175-200 V/cm. Pretreatment with hyaluronidase markedly increased transfection, which was also influenced by the plasmid size and the strain and the age of the mice. Even in immunodeficient mice, there was a significant late decline in transgene expression and plasmid DNA copies, although both still remained relatively high after 1 year. Thus, properly optimized EP-assisted plasmid-based gene transfer is a feasible, efficient, and safe method of gene replacement therapy for dystrophin deficiency of muscle but readministration may be necessary.