Time-resolved FRET screening identifies small molecular modifiers of mutant Huntingtin conformational inflexibility in patient-derived cells

Huntington’s disease (HD) is the most common monogenic neurodegenerative disease and is fatal. CAG repeat expansions in mutant Huntingtin (mHTT) exon 1 encode for polyglutamine (polyQ) stretches and influence age of onset and disease severity, depending on their length. mHTT is more structured compared to wild-type (wt) HTT, resulting in a decreased N-terminal conformational flexibility. mHTT inflexibility may contribute to both gain of function toxicity, due to increased mHTT aggregation propensity, but also to loss of function phenotypes, due to decreased interactions with binding partners. High-throughput-screening techniques to identify mHTT flexibility states and potential flexibility modifying small molecules are currently lacking. Here, we propose a novel approach for identifying small molecules that restore mHTT’s conformational flexibility in human patient fibroblasts. We applied an antibody-based time-resolved Förster resonance energy transfer (TR-FRET) immunoassay, measuring endogenous HTT flexibility using two validated HTT-specific antibodies. The ratio of TR-FRET signal at 4°C and 20°C differs between wtHTT and mHTT and allowed to perform a high-throughput screening using HTT flexibility as a read-out. We identified several small molecules that can partially rescue mHTT inflexibility, presumably by altering HTT post-translational modifications. This novel screening approach has the potential to identify previously unknown HD drugs and drug targets.

resulting in a decreased N-terminal conformational flexibility. mHTT inflexibility may 23 contribute to both gain of function toxicity, due to increased mHTT aggregation 24 propensity, but also to loss of function phenotypes, due to decreased interactions with 25 binding partners. High-throughput-screening techniques to identify mHTT flexibility 26 states and potential flexibility modifying small molecules are currently lacking. Here, we 27 propose a novel approach for identifying small molecules that restore mHTT's 28 conformational flexibility in human patient fibroblasts. We applied an antibody-based 29 time-resolved Förster resonance energy transfer (TR-FRET) immunoassay, measuring 30 endogenous HTT flexibility using two validated HTT-specific antibodies. The ratio of TR- 31 FRET signal at 4°C and 20°C differs between wtHTT and mHTT and allowed to perform a 32 of effector proteins. As a consequence, loss of normal HTT function can alter the activity 48 and cellular localization of different partners thus having a negative impact on cellular 49 physiology. 50 Despite advances to lower the cellular concentration of mHTT by antisense 51 oligonucleotides, certain technical challenges and medical risks remain (4,5). wtHTT's 52 physiological function is still not fully understood and reducing both the mutant and wild-53 type (wt) HTT alleles might come at the risk of neurodevelopmental defects (4,6-10). 54 Asymptomatic mHTT gene carriers, who in principle could benefit the most from 55 reduction in mHTT levels, might also be at a higher risk for detrimental neurological side-56 effects. Furthermore, the intrathecal administration of ASOs is a burdensome procedure  The enzymes responsible for N17 phosphorylation are unknown, but a few 79 molecules that possess N17-modifying capabilities have been identified and were shown 80 to rescue aggregation behavior and toxicity (21-23). However, the currently known small 81 molecule modifiers of HTT PTMs have not been clearly linked to a mHTT flexibility rescue.

82
In addition, these compounds do not all display drug-like properties. 83 Therefore, in order to identify novel modulators of mHTT PTMs, we developed a 84 high-throughput screening approach in HD patient fibroblasts expressing endogenous 85 levels of HTT. We utilize a previously described time-resolved Förster Resonance Energy Qs) (24-28) and is amendable to high-throughput testing. As the PTM status of mHTT can 91 play a central role in HD pathogenesis by influencing mHTT conformational flexibility, 92 aggregation and toxicity, we hypothesize that compounds that modulate mHTT PTMs 93 could provide a promising avenue for novel HD therapeutics.

95
Here, we use a previously described, scalable, and sensitive TR-FRET-based 96 immunoassay assay, which can faithfully detect conformational flexibility differences 97 between wtHTT and mHTT in response to variations in temperature (14-16). HTT 98 conformational changes can be assessed by the relative positions of two antibodies 99 labeled with acceptor and donor fluorophores, which recognize specific epitopes of the 100 target protein: the N-terminal 2B7 antibody, which recognizes both wtHTT and mHTT, 101 and the mHTT (polyQ-specific) MW1 antibody ( Figure 1A). This assay provides a useful 102 screening tool to identify small molecule modulators of mHTT conformational flexibility.

TR-FRET can distinguish HTT flexibility states 104
The presented assay is able to detect conformational constraints imposed by 105 polyQ expansion on HTT in lysates obtained from immortalized human HD fibroblasts, 106 where HTT is expressed at endogenous levels from the relevant genomic locus (14).

107
Screening for potentially PTM modifying small molecules must be performed in living 108 cells in order to guarantee the integrity of all signaling pathways that can alter PTMs. both 4°C and 20°C HTT. We hypothesized that the 4°C/20°C TR-FRET signal ratio 119 generated by the wtHTT protein would be higher than that of mHTT due to wtHTT's 120 increased flexibility at 20°C (Fig 1 A). In addition, we expected that the absolute TR-FRET 121 signal should be higher for mHTT at both 4°C and 20°C due to the presence of higher 122 number of polyQ epitopes, which increase the avidity of the MW1-D2 antibody for the 123 mutant protein. As predicted, we observed that for increasing plasmid concentrations, the 124 TR-FRET signal generated by the expression of mHTT was consistently higher than that 125 generated by wtHTT (Fig 1 B and C). In addition, we also observed a decrease in total TR-126 FRET signal intensity for both constructs when cell lysates where shifted to 20°C.

127
However, while the TR-FRET signal reduction for wtHTT was especially pronounced at 128 this temperature and approached the detection limit (Fig 1 C), the signal for mHTT was 129 just slightly lower that that obtained at 4°C. As a result, the 4°C/20°C TR-FRET ratio for 130 wtHTT was consistently higher even at low plasmid concentrations than the temperature 131 ratio of mHTT (Fig 1 D). Together these results indicate that the TR-FRET-based 132 immunoassay is sensitive enough to detect specific HTT-type dependent conformational 133 changes in complex cell lysates overexpressing wt and mHTT proteins.

144
Next, we tested whether our TR-FRET assay could detect endogenously expressed 145 HTT proteins and whether the dual-temperature read-out was sensitive enough to detect 146 flexibility differences between wtHTT and mHTT in cell lysates. We tested different 147 primary HD patient fibroblast cell lines with varying polyQ lengths (S1 Fig A). The 148 GM04691 fibroblast line displayed the highest signal-to-background (S/B) ratio and was 149 selected for immortalization using hTERT + E6/E7 transduction (S1 Fig A). Most studies In order to determine the optimal cell density to be used for obtaining the highest mHTT were reduced compared to the signal at 4°C (Fig 2 B). This effect was most likely

187
Previous work from several laboratories identified, among other residues, the N-188 terminal HTT serine 13 (S13) to be sufficient for increasing mHTT flexibility when

195
We tested if the phosphorylation of endogenous HTT on S13 by previously Tb/MW1-D2 antibody mix was added. After an overnight incubation, plates were read at 223 20°C followed by an incubation at 4°C for two hours and a second read-out. The assay was 224 highly reproducible and robust between replicates: At 4°C (Fig 3 B) and 20°C (Fig 3 C)

237
FRET ratio approximately two-fold. Interestingly, these hits had no effect on mHTT 238 flexibility at 4°C and clustered together with the DMSO control (Fig 3 B), while at 20°C 239 most hits were among the lowest scoring compounds (Fig 3 C). A reduction of the signal

258
Identified hits represent multiple modes of action 259 We selected the ten strongest hits and characterized them further. For reasons of 260 simplicity the compounds are referred to as C1-C10. The majority of the ten compounds 12 261 have a previously determined biological mode-of-action and defined target protein (Table   262 1). Compound targets were previously identified by Sanofi in a variety of cell types using 263 a range of biochemical or cellular assays. As a result, some of the identified compounds 264 have multiple targets and additional targets are likely to exist.
265 The ten hits were selected based on a 4°C/20°C FRET ratio that was reproducibly higher than 267 three times the standard deviation of the DMSO control around its median value (see also Fig 3 268 D). All tested 3,378 compounds originate from the Sanofi mode-of-action library.

269
Next, we confirmed the hit compounds by assessing their effect on cell viability, 270 establishing a dose-response relationship, and testing their effect on the phosphorylation 271 of mHTT Ser13 (Fig 4). An ATP concentration-based cell viability assay showed that none 272 of the identified compounds led to a high degree of cell death in the tested concentration 273 range of 0.014 to 30 µM when incubated for 16 hours (Fig 4 A) ratio 20-30% compared to the DMSO control (Fig 4 B) which was in the same range as 279 observed during initial screening (Fig 3 D), indicating a mild increase in mHTT flexibility.

280
Especially compounds C1, C2, C4, C8, and C9 showed a dose-response relationship 281 towards an increased 4°C/20°C TR-FRET ratio. Effects were generally the highest in the 282 medium tested concentration range between 0.12 and 3.33 µM, which overlapped with 283 the 3 µM used during screening (Fig 3 D & 4 B). Since HTT N-terminal PTMs such as S13 284 phosphorylation have been described to influence mHTT's conformational flexibility, we 285 tested whether the identified compounds modified this residue's phosphorylation 286 relative to the level of HTT. Compound C1 and C10 led to a relative increase of S13 287 phosphorylation (Fig 4 C). Although, changes in S13 phosphorylation are generally small, and can be used reproducibly in screening-mode with compound libraries (Fig 5). We 324 detected several small molecular compounds which significantly increase mHTT 325 conformational flexibility, although their exact mechanism of action remains unidentified.

334
It is important to note that the effect of most described PTMs on mHTT conformational 335 flexibility is unknown and needs to be investigated further. The screening approach 336 presented here utilizes endogenous full-length mHTT and therefore has the potential to 337 identify other PTMs that influence the conformational flexibility between the N17 and the 338 polyQ region. Due to the lack of suitable antibodies, in the future detailed mass-339 spectrometric analysis of HTT fragments will be required to obtain more information 340 about mHTT regulative sites.

341
As we identified a small number of compounds that seem to influence mHTT 342 flexibility, but also the level of the protein (Fig 3 D & 4 C), our study raises the possibility 343 that dual-action compounds might exist that can be used to lower mHTT levels and in 344 addition partially rescue flexibility. In line with this hypothesis, phosphorylation of 345 certain HTT residues has been found to influence HTT stability. For example, the 16 346 phosphorylation of mHTT at S434 or S536 decreases proteolysis by caspase 3 and calpain, 347 respectively (39,40). Conceptually similar, phosphorylation at S13 and S16 enhances 348 degradation of both wtHTT and mHTT (29). Identifying new small molecular modulators 349 of mHTT PTMs using conformational flexibility as a read-out might therefore enable the 350 discovery of novel and potentially multifaceted drug targets to treat HD.

352
Cell lines and cell culture 353 All used cell lines were obtained from the Venezuelan Huntington Disease