A small-molecule ICMT inhibitor delays senescence of Hutchinson-Gilford progeria syndrome cells

A farnesylated and methylated form of prelamin A called progerin causes Hutchinson-Gilford progeria syndrome (HGPS). Inhibiting progerin methylation by inactivating the isoprenylcysteine carboxylmethyltransferase (ICMT) gene stimulates proliferation of HGPS cells and improves survival of Zmpste24-deficient mice. However, we don't know whether Icmt inactivation improves phenotypes in an authentic HGPS mouse model. Moreover, it is unknown whether pharmacologic targeting of ICMT would be tolerated by cells and produce similar cellular effects as genetic inactivation. Here, we show that knockout of Icmt improves survival of HGPS mice and restores vascular smooth muscle cell numbers in the aorta. We also synthesized a potent ICMT inhibitor called C75 and found that it delays senescence and stimulates proliferation of late-passage HGPS cells and Zmpste24-deficient mouse fibroblasts. Importantly, C75 did not influence proliferation of wild-type human cells or Zmpste24-deficient mouse cells lacking Icmt, indicating drug specificity. These results raise hopes that ICMT inhibitors could be useful for treating children with HGPS.


Introduction
Hutchinson-Gilford progeria syndrome (HGPS) is caused by the accumulation of progerin, a mutant form of prelamin A that is farnesylated and methylated within the nuclear envelope (De Sandre- Giovannoli et al., 2003;Eriksson et al., 2003). Farnesyltransferase inhibitors (FTIs) prevent progerin farnesylation and improve some clinical phenotypes of HGPS patients, including survival, but the effect is modest (Gordon et al., 2018;Young et al., 2005). Also, a potential limitation of this approach is that FTIs are anti-proliferative (Lee et al., 2010), and children with progeria would benefit from a therapy that supports cell proliferation. We found earlier that inhibiting the methylation of progerin by inactivating the isoprenylcysteine carboxylmethyltransferase (ICMT) gene overcomes senescence and increases proliferation of HGPS cells (Ibrahim et al., 2013). Also, a knockout of Icmt substantially improves clinical phenotypes and survival of Zmpste24-deficient mice, a model of progeria (Ibrahim et al., 2013). This result raises the possibility that inhibiting ICMT activity could be a useful therapeutic strategy. An important step in the preclinical validation of this strategy would be to determine whether knockout of Icmt improves phenotypes and survival in an authentic progerinexpressing HGPS mouse model. Another step would be to determine whether pharmacologic targeting of ICMT produces similar cellular effects as genetic inactivation. To address these issues, we defined the consequences of knocking out Icmt in progerin-knock-in mice; and we synthesized a potent cell-permeable ICMT inhibitor, compound 75 (C75) (Judd et al., 2011), and examined its effects on HGPS cells.

Results
We bred mice with a hypomorphic Icmt allele (Icmt hm ; with~85% reduced ICMT activity [Ibrahim et al., 2013]) with progerin-expressing lamin A knock-in mice (Lmna G609G [Osorio et al., 2011]). As expected from previous studies, Lmna G609G/G609G Icmt +/+ mice developed alopecia, stunted growth, and weight loss, and all mice had died by 129 days of age; at that time, numbers of vascular smooth muscle cell (VSMC) nuclei in aortic arch sections were reduced by~75% and muscle fiber size in the quadriceps muscle were 50% smaller compared with wild-type mice (Figure 1a-f). In contrast, at 129 days of age, the Lmna G609G/G609G Icmt hm/hm mice were still alive with substantially higher body weights, and when they were sacrificed, VSMC numbers were found to be normalized and skeletal muscle fiber size increased (Figure 1a-  they should be interpreted with caution as the mice were difficult to breed and we only obtained three double homozygotes. Consequently, we also analyzed Lmna G609G/+ Icmt +/+ mice, which had a maximal life span of 290 days ( Figure 1g) (and no aorta and skeletal muscle phenotypes); importantly, all Lmna G609G/+ Icmt hm/hm mice were still alive at 290 days and their overall survival were increased ( Figure 1g). These results are important because progerin rather than prelamin A causes progeria in Lmna G609G mice, and because homozygous Lmna G609G/G609G mice exhibit a vascular phenotype, which are prominent in children with HGPS, but absent in Zmpste24-deficient mice used in earlier studies (Ibrahim et al., 2013)-and Icmt inactivation markedly improved this phenotype. We next synthesized the ICMT inhibitor C75 as described (Judd et al., 2011), and found that its IC50 was 0.5 mM (Figure 2a-figure 2 supplement S1). Prolonged C75 incubation was well tolerated by two different human HGPS cell lines and caused prelamin A accumulation and mislocalization of the RAS oncogene -markers of reduced ICMT activity-but did not affect the nuclear shape abnormalities (Figure 2b-d-figure 2 supplement S2a). C75 did not influence the electrophoretic mobility of HDJ2 which could have been indicative of effects on FTase-mediated farnesylation (figure 2 supplement S2b). Importantly, C75 increased proliferation of late-passage HGPS cell lines as judged by 45-to 70-day population-doubling assays (Figure 2e-f). The drug had increased cell viability already at 8 days, that is, before the increase in cell proliferation was evident (figure 2 supplement S2c-d). C75 also increased proliferation of Zmpste24-deficient mouse fibroblasts with normal Icmt expression but not in cells lacking Icmt, indicating drug specificity (Figure 2g-h). The drug did not affect proliferation of wild-type human fibroblasts ( Figure 2i). In contrast, the FTI lonafarnib rapidly reduced proliferation of HGPS cells and abolished the effect of C75 on cell growth (Figure 2jk). The latter finding makes sense because protein methylation cannot occur without protein farnesylation.
Consistent with increased proliferation, C75 increased the fraction of HGPS cells in the G1 and S/ G2/M phases of the cell cycle (figure 2 supplement S2e); reduced senescence-associated b-galactosidase activity ( Figure 2I-m); and normalized the expression of the senescence markers IL6 and CDKN2A (Figure 2o-p). The drug also normalized oxygen consumption rates and ATP production in HGPS cells, as judged by Seahorse analyses; and reduced the levels of oxidative stress (figure 2 supplement S2f-g). In contrast, C75 did not influence expression of endoplasmic reticulum (ER) stress markers in HGPS cells; and DNA damage signaling markers, including g-H 2 AX, remained unchanged (figure 2 supplement S2h-i). C75 did however reduce the fraction of cells with nuclei harboring multiple g-H 2 AX foci, examined by immunofluorescence, but not the total fraction of g-H 2 AX-positive nuclei (figure 2 supplement S2j). A potential interpretation of the latter finding is that C75 increases proliferation of cells with low levels of DNA damage which outcompete cells with high levels.
The signaling molecule AKT binds progerin and farnesyl-prelamin A (in HGPS and Zmpste24-deficient cells, respectively) and exhibits only low levels of phosphorylation (Ibrahim et al., 2013). C75 reduced the progerin-AKT interactions and increased AKT phosphorylation (Figure 3a-d), but C75 did not influence phospho-AKT levels in wild-type cells ( Figure 3b). Although these data don't reveal whether AKT is functionally involved in the improved phenotypes upon C75 administration, we showed previously that pharmacologic AKT inhibition prevents the increased proliferation following Icmt knockout in mouse cells (Ibrahim et al., 2013). Despite improvements in multiple cellular phenotypes, C75 increased the absolute levels of progerin in HGPS cells, a consequence of reduced progerin turnover (Figure 3e-g). Moreover, C75 mislocalized some progerin and farnesyl-prelamin A away from the nuclear membrane into the nucleoplasm (Figure 3h-i).

Discussion
We conclude that genetic Icmt inactivation improves survival and unique phenotypes of an authentic HGPS mouse model. We further conclude that pharmacologic inhibition of ICMT delays senescence, restores respiration rates and ATP production, and stimulates proliferation of HGPS and Zmpste24deficient cells, most of which is consistent with findings in Icmt-deficient cells (Ibrahim et al., 2013). Blocking methylation partially mislocalizes progerin to the nucleoplasm, disrupts its interaction with AKT, and increases AKT signaling. These positive phenotypes of blocking methylation with C75 outweighed any potential adverse effects from the modest amount of progerin accumulation. In wildtype cells, prelamin A is fully processed to mature lamin A and the cells grow and proliferate normally. In HGPS and Zmpste24-deficient cells, farnesylated and methylated progerin/prelamin A accumulates and causes senescence. Our data suggest that progerin/prelamin A methylation contributes to the toxicity of these proteins and their ability to induce senescence, and we propose that blocking progerin/prelamin A methylation mislocalizes the proteins into the nucleoplasm and thereby reduces their ability to induce DNA damage, metabolic alterations, and senescence. A limitation of C75 is that despite good apparent permeability (Figure 2 -supplement 1d) it is predicted to have poor bioavailability (i.e., very hydrophobic and high first-passage metabolism in in silico ADME analyses). Thus, new compounds will be required for in vivo studies in mice. Nonetheless, our study takes two important steps in the preclinical validation of ICMT as a potential drug target, and thereby raises hopes that ICMT inhibition could be an effective strategy for treating children with HGPS and progeroid disorders resulting from ZMPSTE24 deficiency (Michaelis and Hrycyna, 2013).
Drug synthesis, ICMT activity assay, and apparent permeability assay C75 was synthesized by Recipharm AB as described (Judd et al., 2011). ICMT activity was carried out as described (Choy and Philips, 2000;Zhou et al., 2016). Apparent permeability assay was carried out by analyzing the apical-to-basolateral (and vice versa) transport of C75 using Caco-2 cell monolayers as described (Hubatsch et al., 2007).

Senescence-associated b-galactosidase assay
Senescence-associated b-galactosidase (SA-b Gal) staining on primary MEFs and human HGPS cell lines was performed using the Senescence Detection kit (9860, Cell Signaling). Cells were incubated with SA-b Gal solution for 24 hr (mouse) and 4 hr (human), separately, at 37˚C. Results are reported as percent of blue cells.

Isolation of nuclear membrane and nucleoplasm fractions
Nuclear membrane and nucleoplasm separation was performed on MEFs and human fibroblasts using Minute Nuclear Envelop Protein Extraction Kit (NE-013, Invent Biotechnology), and Minute Detergent-Free Nucleoplasm Isolation Kit (NI-024, Invent Biotechnology).

Mitochondrial function assay
Mitochondrial function parameters were measured with the Cell Mito Stress Test kit using the Seahorse XFe96 Analyzer (Agilent). Cells were seeded in microplates (15,000 cells/well) (101085-004, Agilent) and cultured overnight at 37˚C in a CO 2 incubator. Freshly prepared DMEM-base medium supplemented with glucose, pyruvate, and glutamine, and adjusted to pH 7.4 were added to the cells and they were incubated for 45 min at 37˚C in a non-CO2 incubator and then analyzed at 37˚C in the XFe96 Analyzer. Basal and maximal respiration and ATP production data were normalized to viable cell numbers obtained from identically treated additional wells using the Presto Blue Cell Viability assay (A13262, ThermoFisher).

Flow cytometry analysis
HGPS cells were incubated with 250 ml fixation/permeabilization solution (554714, BD) for 30 min on ice, in the dark; washed twice; incubated with antibodies to Ki67 (5 ml/sample; 561277, BD) at room temperature for 1 hr and 45 min; washed with PBS + 10% FCS; stained with 7AAD (5 mg/sample; A9400-1MG, Sigma) at room temperature for 20 min; resuspended and filtered into flow tubes; and analyzed using a BD LSRFortessa X-20.

Statistics
Data are presented as mean ± SEM. For statistical analyses, we used Graphpad Prism software v.7; the log-rank test was used for survival, two-way ANOVA for cell-growth curves, one-way ANOVA with Bonferroni's post-hoc test when comparing three or more groups, and Student's t test when comparing two groups only. Experiments were repeated 2-4 times unless stated otherwise; n indicates biological replicates.
the Royal Institute of Technology. The study was supported by grants from the Progeria Research Foundation, Center for Innovative Medicine (CIMED), and the Swedish Research Council (to MOB). The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Data availability
Data generated in this study are presented in the manuscript and supporting files. A source file for exact P values is also included.