Reducing levels of toxic RNA with small molecules

Abstract

Myotonic dystrophy (DM) is one of the most common forms of muscular dystrophy. DM is an autosomal dominant disease caused by a toxic gain of function RNA. The toxic RNA is produced from expanded noncoding CTG/CCTG repeats, and these CUG/CCUG repeats sequester the Muscleblind-like (MBNL) family of RNA binding proteins. The MBNL proteins are regulators of alternative splicing, and their sequestration has been linked with mis-splicing events in DM. A previously reported screen for small molecules found that pentamidine was able to improve splicing defects associated with DM. Biochemical experiments and cell and mouse model studies of the disease indicate that pentamidine and related compounds may work through binding the CTG*CAG repeat DNA to inhibit transcription. Analysis of a series of methylene linker analogues of pentamidine revealed that heptamidine reverses splicing defects and rescues myotonia in a DM1 mouse model.

Figure 1

Structure of pentamidine linker analogs. Methylene carbon linker highlighted by gray parenthesis (n = 5 for pentamidine).

Figure 2

CUG repeat RNA levels after treatment with pentamidine or pentamidine analog. a-c) Northern blot analysis and quantification of HeLa cells expressing 960 CUG repeats treated with a) propamidine, b) pentamidine, and c) heptamidine. Propamidine and pentamidine were observed to significantly decrease CUG levels while heptamidine did not. The CUG RNA produced from the plasmid is observed to be two different sized bands. These different bands could be the product of alternative splicing or polyadenylation of the transcript, or possibly a contraction of the repeats, as CTG repeat DNA is notoriously unstable.

Figure 3

Inhibition of in vitro transcription by pentamidine. a) Representative gel images of transcription reactions. Each template was treated with increasing concentrations of pentamidine. b) IC₅₀ curves for each template, plotted on a log scale. c) AT content of each template plotted against the IC₅₀ for that template. Data for non-repeat templates (APP3, APP5 and pTRIEX) could be fit by an exponential decay curve. Pentamidine inhibited the transcription of CTG/CAG repeat templates much better than would be predicted by AT content alone. Error bars are standard deviation of three observations.

Figure 4

Binding of fluorescent pentamidine analogs to CTG*CAG repeat hairpin DNA. a) Structure of fluorescent pentamidine analogs prop-BAPPA (n=3), pent-BAPPA (n=5), and hept-BAPHA (n=7). b-d) Representative CTG*CAG titrations into 0.5 μM b) prop-BAPPA, c) pent-BAPPA, and d) hept-BAPHA. e) Maximum emission of each DNA concentration curve. Data were zeroed and fit with rectangular hyperbola in order to obtain an apparent K_d. Error bars are standard deviation of three observations.

Figure 5

Pentamidine analogs rescue mis-splicing of minigene reporters in a HeLa cell DM1 model. a) Jitter plot representation of TNNT2 splicing. Each point is one experiment and the line represents the average of all experiments for that condition (at least three for each concentration). Butamidine, pentamidine, hexamidine, and heptamidine rescued TNNT2 mis-splicing in the presence of 960 CUG repeats. Octamidine had a slight effect. b) Plots of INSR splicing. All pentamidine linker analogs are able to fully or partially rescue INSR mis-splicing in the presence of 960 CUG repeats. Gray area denotes range between typical splicing and DM mis-splicing.

Figure 6

Plot of EC₅₀ values versus methylene linker length. When INSR splicing is monitored, the inverse relationship between methylene linker length and EC₅₀ holds. The trend is less clear with TNNT2. Error bars are standard deviation of at least three separate measurements.

Figure 7

Heptamidine rescue of endogenous mis-splicing events and myotonia in HSA^LR mice. a) Clcn1 showed a complete rescue by 20 mg kg⁻¹ per day for 7 d. Each symbol represents the splicing outcome for vastus muscle of a single mouse. After treatment, withdrawal mice (WD) were maintained for 10 d with no additional heptamidine injections. These mice showed a complete return to the basal level of splicing impairment. Atp2a1 showed partial rescue of mis-splicing. At 30 mg kg⁻¹ heptamidine, exon 22 inclusion is approximately 50% rescued. WD mice again reverted to the pre-treatment splicing levels. Errors are standard deviation. Gray area denotes range between typical splicing and DM mis-splicing. b) Myotonia rescue in HSA^LR mice treated with 0, 20, or 30 mg kg⁻¹ heptamidine for 7 d. Mice not treated with heptamidine showed myotonic discharge with nearly all electrode insertions (grade 3), whereas those treated with 20 or 30 mg kg⁻¹ heptamidine had occasional myotonic discharges with less than 50% of insertions (grade 1) or no myotonia (grade 0). c) qRT-PCR analysis of HSA levels in HSA^LR mice treated with a 5% glucose solution or 15 mg kg⁻¹ heptamidine for 7 d. Mice treated with heptamidine showed significant (p-value < 0.02) reduction in HSA levels as compared to glucose. Error bars are standard error. d) qRT-PCR analysis of HSA levels in HSA^SR mice treated with a 5% glucose solution or 15 mg kg⁻¹ heptamidine for 7 d. No significant difference was observed between the glucose and heptamidine treated mice with only 5 CUG repeats.