Prenyltransferase Inhibitors: Treating Human Ailments from Cancer to Parasitic Infections

2.1 Biochemistry of Protein Prenylation

FTase and GGTase I use the same mechanism to prenylated their respective substrates (Figure 3). The prenylation reaction begins when the prenyltransferase enzyme binds the isoprenoid substrate, either the C₁₅ farnesyl or the C₂₀ geranylgeranyl, and forms a binary enzyme-isoprenoid diphosphate complex. Following binary complex formation, the CAAX protein substrate binds adjacent to the isoprenoid molecule in the enzyme active site. As the CAAX protein binds, the cysteine residue coordinates to a Zn²⁺ ion (as a thiolate) in the active site; the Zn²⁺ is necessary for catalysis.^23,24 This ternary complex is short lived, as the cysteine thiolate reacts at the C-1 position of the isoprenoid, forming a stable, covalent, thioether linkage to the isoprenoid. Upon binding of an additional prenyl diphosphate molecule, the isoprenoid portion of the newly formed prenylated protein translocates into an exit groove in the enzyme, and subsequently the prenylated protein product is released from the enzyme; this last step (product release) is the rate-limiting step for the protein prenylation reaction. In addition, FTase requires millimolar levels of Mg²⁺ for efficient catalysis to occur, as this ion stabilizes the diphosphate leaving group of FPP during the reaction. In GGTase I, however, a positively charged lysine residue near the diphosphate stabilizes it and removes the requirement of exogenous Mg²⁺.

In terms of the chemical mechanism for the enzymatic reaction, both FTase and GGTase I catalyze C-S bond formation with inversion of configuration at C-1 of the isoprenoid.^25,26 Fluorine substituted FPP analogues manifest an attenuated rate suggesting a transition state with some cationic character.^27,28 Kinetic isotope effect measurements are also consistent with partial C-O bond cleavage as well as partial C-S bond formation in the transition state.^29,30 QM/MM calculations indicate that no stable carbocationic intermediate occurs in the reaction pathway.³¹ Although not observed crystallographically, single turnover experiments and calculations suggest that a conformational change in the enzyme and/or substrate(s) occurs prior to C-S bond formation.^32,33

The RabGGTase reaction proceeds with some similarities and some key differences (Figure 3). The major difference is that RabGGTase only recognizes a Rab substrate protein when it is bound to a Rab escort protein (REP). Similar to FTase and GGTase I, Rab prenylation begins when RabGGTase is loaded with a molecule of GGPP. The next step involves a portion of the REP, the Rab-binding platform (RBP), recognizing a Rab protein substrate and binding to it. This binding is enhanced when the C-terminal binding region (CBR) of the REP interacts with the CBR interacting motif (CIM) on the Rab protein substrate, enhancing the affinity of the REP for the Rab protein substrate (further information on the CBR/CIM interaction can be found in section 2.2). The binary complex then binds with the RabGGTase and GGPP binary complex to form a quaternary complex of REP/Rab/RabGGTase/GGPP, in which the C-terminal cysteine of the Rab substrate binds to the active site Zn²⁺ of RabGGTase as a thiolate. Catalysis then occurs, with the thiolate of the Rab substrate reacting with C-1 of the GGPP isoprenoid to form a covalent thioether linkage. Next, an additional molecule of GGPP binds to the active site of RabGGTase by displacing the previous isoprenoid and the second round of prenylation occurs. Substrate release occurs analogously to FTase and GGTase I, in which the binding of an additional GGPP molecule forces displacement of the REP/di-prenylated Rab complex from RabGGTase (this is also the rate limiting step in Rab geranylgeranylation).¹² The REP/di-prenylated Rab complex is then trafficked to its cognate target membrane by a yet unknown mechanism. The REP protein is not only responsible for Rab protein binding and escort, but also for solubilization of the di-prenylated Rab product until its delivery to a membrane.³⁴ Since Rab proteins lack the C-terminal AAX motif, they are not subject to proteolysis and methylation the way FTase and GGTase I substrates are.

All three prenyltransferase enzymes manifest nanomolar affinities for their respective isoprenoid substrates. Additionally, the rate-limiting step in the prenylation reaction is the release of product, which relies on the binding of an additional isoprenoid molecule. These two factors combine to give the prenyltransferases relatively slow turnover rates of approximately 3 min⁻¹, depending on the substrate.¹⁹ The rates of C-S bond formation catalyzed by the three enzymes (obtained from pre-steady state measurements) span an order of magnitude, with FTase reacting the fastest, followed by GGTase I and lastly RabGGTase (k_chem values are 12–17 s⁻¹, 0.5 s⁻¹, and 0.16 s⁻¹ for the three enzymes, respectively).³⁵ Furthermore, RabGGTase catalyzes its first geranylgeranylation four times faster than the second (k₁ = 0.16 s⁻¹ and k₂ = 0.04 s⁻¹).³⁵ This difference in rates is likely due to steric hindrance between the first prenylated cysteine residue and the next, as the two C-terminal cysteine residues are generally adjacent to one another.

2.2 Substrate Specificity

The specificity of the FTase and GGTase I catalyzed reactions arises mainly from the CAAX motif on the substrate protein, although analogs of FPP can alter the substrate specificity of the FTase enzyme.³⁶ Recently, methods for preparing C-terminal peptide libraries have been developed that have considerable potential for future experiments aimed at defining the global determinants of protein sequence specificity.³⁷ Additionally, structural differences in FTase and GGTase I account for the ability of the enzymes to discriminate between FPP or GGPP. In the isoprenoid binding pocket of FTase in the β-subunit, there are two residues that limit the length of isoprenoid that can bind: a Trp102 at the end of the binding cavity and a Tyr365 that prevents anything longer than a farnesyl group from rotating to fit (Figure 4A). In GGTase I, these same positions are occupied by less bulky residues, Thr49 and Phe324, respectively. These smaller residues, especially the Thr49, allow the longer GGPP substrate to fit into the active site. This simple ‘molecular ruler’ allows GGTase I to bind its longer, cognate isoprenoid. In contrast, FTase uses a different mechanism to control isoprenoid selectivity. While FTase binds GGPP with high affinity, the longer isoprenoid bulges out into the protein binding site due to the limited depth of the binding pocket (Figure 4B). The resulting binary complex cannot bind a protein substrate and is hence catalytically inactive.³⁸

Much less is known about the precise determinants of substrate specificity in the Rab GGTase reaction. In contrast to FTase and GGTase, RabGGTase does not act on simple peptide substrates. The first crystal structure of RabGGTase in complex with the REB wasn’t solved until 2003.³⁹ This has hampered detailed mechanistic studies of the RabGGTase reaction and how this enzyme recognizes its substrates. Recently, Guo and coworkers solved RabGGTase structures complexed with substrates and products which enabled a more complete picture to be revealed.³⁵ RabGGTase has the broadest lipid substrate specificity among all prenyltransferases (i.e. it is the most promiscuous of the prenyltransferases).⁴⁰ Even though the CBR of REP is a participating protein not involved directly in catalysis, it has been proposed that the interactions between the CIM of the Rab protein and the CBR of the REP are responsible for the substrate specificity of RabGGTase, analogous to the ‘AAX’ residues in FTase and GGTase I.³⁵ The REP controls the specificity of RabGGTase by selecting and binding to Rab proteins and facilitating their prenylation. The RabGGTase enzyme itself is highly promiscuous and operates stochastically, geranylgeranylating the protein substrate whenever a cysteine happens be in its active site.³⁵

2.3 Types of Prenyltransferase Inhibitors

Synthetic inhibitors of FTase and GGTase I fall into one of three categories: isoprenoid analogs, peptidomimetics (also known as CAAX competitive inhibitors), and bi-substrate inhibitors combining both the isoprenoid analog and peptidomimetic portions (Figure 5). Specific inhibitors of RabGGTase include those that are phosphonocarboxylates as well as peptidomimetics (Figure 5). Highly potent natural product-derived inhibitors of these enzymes have also been identified.⁴¹

Isoprenoid analog inhibitors of all three prenyltransferases have been the least developed and studied, because of three important reasons:⁴² 1) The affinity of the isoprenoids for the prenyltransferases makes it difficult to design inhibitors that can be competitive with the native substrate (for example, the K_D of FPP for FTase is 2 nM⁴²). 2) The diphosphate moiety is negatively charged and this may affect the uptake properties of the molecule versus a prodrug or uncharged molecule.⁴³ While biosynthetic intermediates in the isoprenoid pathway (present as diphosphates) and anionic inhibitors clearly enter the cell, the precise mechanism by which this occurs remains unclear.⁴⁴ 3) Several additional enzymes (e.g. squalene synthase, undecaprenyl pyrophosphate synthase, heme O synthase, etc.) utilize isoprenoids as substrates and an isoprenoid analog inhibitor may have significant off-target effects. Bisubstrate inhibitors that contain both a CAAX competitive portion and an isoprenoid analog portion have also not been well developed, as complications in synthesis and water solubility have hampered their progress. The next section will focus on those inhibitors that are peptidomimetic as they have been the most studied and refined of all the classes of prenyltransferase inhibitors.

2.4 Evolution of FTIs

Peptidomimetic FTIs have been developed by three strategies: “reengineering” of substrate peptides,⁴⁸ use of rigid nonpeptidic scaffolds that mimic peptide secondary structural elements,^49,50 and large scale screening of compound libraries.⁵¹ The initial development of peptidomimetic FTIs began with the native K-Ras CAAX box CVIM peptide by mutating the A₂ position to a phenylalanine (CVFM), giving an inhibitor with a 25 nM IC₅₀ (Figure 6).⁵² Placing bulky substituents in the A₂ position displaces the peptide in the binding pocket resulting in a catalytically inactive conformation and enzyme inhibition. Other inhibitors of this type have been able to achieve low nanomolar inhibition of FTase (see L-739,750, Figure 6),⁵³ but the peptide based scaffold made them less than ideal for in vivo use due to the possibility of proteolytic degradation. The next generation of FTIs generally contained fewer amide bonds in the backbone to circumvent this problem. For example, FTI-276 (Figure 6) was prepared from a related scaffold and binds in a similar manner to the FTase active site as L-739,750, achieiving a 0.5 nM IC₅₀ (with respect to FTase and 50 nM with respect to GGTase I) without potential degradation issues.⁵⁴ Interestingly, changing the C-terminal methionine of FTI-276 to a leucine changes the specificity of inhibition (25 nM for FTase and 5 nM for GGTase I, Figure 6).⁵⁵

It was presumed in the early development of FTIs that a thiol moiety was necessary in the inhibitor structure to preserve the zinc coordination. Thiols can be oxidized to sulfoxides and sulfones, and the search for thiol replacements for zinc coordination led to the development of FTIs containing phenols, carboxylic acids, and nitrogen heterocycle derivatives. The most potent of these are the pyridine⁵⁶ and imidazole⁵⁷ based inhibitors (see Figure 7) with low nanomolar IC₅₀ values. Placing the zinc-coordinating imidazole in a cyclic structure led to one of the most potent FTIs developed with an IC₅₀ of 20 pM (Figure 7).⁵⁸

The last evolutionary step for FTIs occurred by removing the zinc-coordination on the peptidomimetic inhibitor. A class of inhibitors from Schering-Plough, originally based on an antihistamine scaffold, exemplifies the effects of removing the zinc-coordination (Figure 8). SCH 66366 (lonafarnib) lacks any zinc-coordinating moiety but still has an impressive IC₅₀ of 1.9 nM.⁵⁹ When a similar scaffold was modified slightly to contain an imidazole zinc-coordinating moiety, the IC₅₀ improved nearly 50-fold to 0.04 nM (SCH 211618).⁶⁰

Tuning the FTI to contain various structural elements has resulted in the creation of hundreds of FTIs with low nanomolar IC₅₀ values. Several of the best FTIs that combine potency, oral-bioavailability, and efficacy in animal models have entered clinical trials for cancer treatment (see Section 3). Additionally, an important goal in the development of FTIs has been to provide specificity for FTase over GGTase I, as they share considerable similarity, especially in the active site. Section 2.5 will discuss inhibitors that have been developed specifically to target GGTase I while having minimal impact on FTase.

2.5 Inhibitors for GGTase I

Providing selective GGTase I inhibitors was initially as simple as modifying the peptide backbone residues to contain larger, more hydrophobic elements to better fit into the GGTase I active site and exclude them from FTase (see Figure 6). This strategy only provided approximately 5-fold selectivity for GGTase I over FTase. Structure-activity design efforts to introduce non-oxidizable zinc-coordinating groups, such as an imidazole, into the scaffold from GGTI-287 has led to the development of an inhibitor with greater than 250-fold selectivity for GGTase I over FTase (GGTI-2154, Figure 9).⁶¹ Additional tuning and optimization of this C-terminal leucine scaffold led to a dramatic selectivity increase of over 5,000-fold for GGTase I over FTase (GGTI-2418, Figure 9).⁶² Clearly, the use of structure-activity relationships has been successful for the development of selective inhibitors of GGTase I; furthermore, the use of library screening has proven immensely beneficial. The inhibitor P5-H6 was developed after screening a 171 compound library for inhibitors of GGTase I (Figure 9).⁶³ This compound exhibits an approximate 100-fold selectivity for GGTase I over FTase. Many potent and selective inhibitors for GGTase I have been developed, however, only one inhibitor, GGTI-2418 (Figure 9), has advanced to clinical trials (see Section 3).

2.6 RabGGTase inhibitor development

The design of inhibitors of RabGGTase has progressed significantly more slowly, as the main effort has been in pursuing FTIs for cancer therapy based on the importance of Ras proteins in malignancy. However, recently, RabGGTase has been shown to be a valid cancer target as many Rab proteins and RabGGTase itself, are overexpressed in several cancer types, such as ovarian,⁶⁴ breast,⁶⁴ hepatocellular carcinoma,⁶⁵ and thyroid adenomas.⁶⁶ The difficulty in the development of RabGGTase inhibitors has been in creating molecules that specifically inhibit only this enzyme, as there is considerable similarity with the other two prenyltransferases (Figure 2). Further, RabGGTase has the largest amount of lipid promiscuity of all three prenyltransferases and doesn’t recognize a single, direct substrate, but rather a dimer of Rab and REB, complicating inhibitor development. Nonetheless, progress has been made to identify specific RabGGTase inhibitors with high potency, with the current generation achieving low nanomolar inhibition.

The first generation of RabGGTase inhibitors was based on the scaffold of risedronate, a bisphosphonate drug that is used to strengthen bone and treat osteoporosis through osteoclast inhibition, leading to increased bone mass.⁶⁷ Interestingly, the molecular target of risedronate is farnesyl diphosphate synthase (FPPS), which is responsible for the synthesis of farnesyl diphosphate in the cell.⁶⁸ Inhibition of FPPS directly decreases the available pool of FPP, in addition to GGPP, which prevents geranylgeranylation of several proteins and leads to osteoclast inhibition.⁶⁹ Inhibition of geranylgeranylated proteins in osteoclasts is crucial for the treatment of osteoporosis, and this drug specifically inhibits the isoprenoid pathway to prevent protein geranylgeranylation, while also inhibiting the RabGGTase enzyme, thus functioning via a dual mode of action.^70,71

Structure-activity relationship studies on risedronate led to the synthesis and evaluation of phosphonocarboxylate derivatives of the bisphosphonates with low micromolar inhibition of RabGGTase (Figure 10).⁷² The first such compound was 3-PEHPC (also known as NE10790), which was originally reported to have an in vitro IC₅₀ value of approximately 600 μM;⁴⁶ however, in later assays with a modified buffer system and optimized concentrations of RabGGTase, isoprenoid and protein substrates, and Rab escort protein, the newly reported in vitro IC₅₀ value of 3-PEHPC for RabGGTase was 24.1 μM.⁷² It was believed that the α-hydroxyl in 3-PEHPC was important for bone mineral binding, so a series of halogen substitutions on the α-carbon was synthesized with the goal of improving binding and enhancing RabGGTase inhibition; this, however, only led to minor improvements (Figure 10).⁷² Substitution of the pyridine ring structure in 3-PEHPC with a imidazo[1,2-a]pyridine cyclic moiety led to the most potent inhibitor of RabGGTase in this compound class, with an IC₅₀ value of 1.3 μM, nearly 25-fold better than the initial inhibitor structure.⁷³ It was recently discovered that this class of inhibitors inhibits only the second geranylgeranylation step of RabGGTase, not the first, and it was suggested that the phosphonocarboxylate inhibitors bind in a site adjacent to the active site, preventing the mono-geranylgeranylated protein from translocating to provide room for the second prenylation step to occur.⁷³ Some Rab proteins are only mono-geranylgeranylated and as a consequence are not inhibited by the phosphonocarboxylates; the development of specific inhibitors of all Rab protein geranylgeranylation has become a high priority.

In addition to the phosphonocarboxylate derivatives, several attempts to make specific peptidomimetic inhibitors for RabGGTase have been attempted, with recent generations performing exceptionally well. Bon and coworkers developed the most selective RabGGTase inhibitors to date by modifying the scaffold of the inhibitor BMS3 (Figure 11), which, at the time of their study, was the most potent RabGGTase inhibitor, but that was not selective as it also inhibited FTase.⁷⁴ Through structure-activity relationship guided design, their best molecule inhibited RabGGTase with an in vitro IC₅₀ of 42 nM (Figure 11) and displayed high selectivity, as it showed no inhibition of FTase or GGTase I at concentrations up to 9,700 nM and 99,500 nM, respectively; the compound also inhibited the proliferation of several cancer cell lines, demonstrating potency in cell-based assays.⁷⁴ To achieve this remarkable selectivity, the authors capitalized on a structural element present only in RabGGTase, called the TAG tunnel. This tunnel provides extra space in the active site for molecular diversity; placing a benzylcarbamate substituent in position A fills this TAG tunnel in RabGGTase, while excluding the ability of the molecule to bind FTase or GGTase I (Figure 11). Additional modification of the cyano group (B) with a furonitrile derivative provides hydrophobic interactions with the lipid substrate pocket further enhancing RabGGTase selectivity. Continued development of RabGGTase inhibitors in this manner should allow potent inhibitors to be developed with the potential to be used in a clinical setting.

3.1 FTIs as cancer therapeutics

Initial interest in developing farnesyltransferase inhibitors (FTIs) was spurred by the finding that the oncogenic Ras protein is farnesylated²⁰ and that the farnesyl group is necessary for membrane association and proper function of Ras.⁷⁵ Activating mutations in Ras, and specifically K-Ras, are involved in more than 30% of all human cancers and have been implicated in several tissue specific cancers with alarmingly high frequency; for example, greater than 90% of pancreatic cancers and 50% of colorectal cancers harbor K-Ras mutations.⁷⁶

The first report of an inhibitor of FTase was published in 1990⁷⁷ and since this time hundreds of different FTIs have been developed, with four progressing to clinical trials: tipifarnib, lonafarnib, BMS-214662 and L-778,123 (Figure 12).⁶ In the 75 different clinical trials using these four compounds alone or in combination with other anti-cancer drugs, a mere 2.3% of all patients treated had an objective response (either a partial or complete response) with monotherapy, while only 11.4% had a similar response with combination therapy.⁶ The results of these studies have clearly been a disappointment. The most probable explanation for these unanticipated results is likely due to the fact that the majority of clinical trials enrolled patients harboring mutations in the K-Ras oncogene. In the presence of FTIs, both the K-and N-Ras proteins can be alternatively prenylated by GGTase I to yield geranylgeranylated forms that retain their biological function.^78–81 Patients that had been given an FTI to treat their tumors bearing K-Ras mutations likely suffered from this unfortunate biochemical event. Additionally, a majority of the patients enrolled in these clinical trials had advanced or metastatic disease, that may have been too advanced for effective treatment with an FTI. As an example, mice bearing H-Ras mammary tumors showed tumor regression upon daily FTI administration, but when the FTI was withdrawn the tumor reappeared; this may indicate that treating H-Ras bearing tumors early, prior to metastasis, may prove effective.⁸² It is likely that clinical trials utilizing more selective patient enrollments, especially with those patients harboring H-Ras mutations may have been more effective.

3.2 FTIs currently in clinical trials

Only two clinical trials remain active, both of which are using tipifarnib. The first is a Phase 2 trial to treat patients with relapsed or refractory lymphoma, while the second is a Phase 3 observational trial to determine whether tipifarnib or observation alone is better at preventing cancer recurrence in patients with acute myeloid leukemia (www.clinicaltrials.gov).⁸⁷ With the remaining two trials nearing completion, it is likely that clinical trials with FTIs will no longer be a priority (see section 3.4).

3.3 GGTase I inhibitor development for cancer

In early 2009, GGTI-2418 (Figure 9) entered Phase I clinical trials investigating tolerability, dosing, and toxicity; patients that had metastatic tumors for which the current therapies were unsuccessful or unavailable were selected for participation. Early results demonstrated that this compound was well-tolerated and had minimal toxicity.⁸⁸ However, the Phase I trial of GGTI-2418 has been stopped due to its lack of efficacy in patients (http://kiraxcorp.com/ggti2418.html, personal communication with Kirax Corp. representative). GGTI-2418 was the first GGTase I inhibitor to enter clinical development. Given the ability of FTase to circumvent GGTase inhibition and vice versa, monotherapy with these inhibitors remains a distinct challenge.

3.4 Future of prenyltransferase inhibitors in cancer

As discussed previously, inhibitors of both FTase and GGTase I have been ineffective in clinical trials for cancer treatment and have had disappointing results. To date, no inhibitors of RabGGTase have entered clinical trials, as this enzyme was only recently shown to be a potential cancer target (see section 2.6). One of the biggest problems in these clinical trials has been how patients have been selected (see section 3.1–3.2); however, an even more pressing issue is finding the exact molecular targets of prenyltransferase inhibitors. It has been shown that in most systems tested, from cells in culture to human patients, the ability of FTIs to inhibit tumor growth does not correlate with mutations in Ras.^83,89,90 This information indicates that even though the Ras protein is oncogenically mutated, it is the inhibition of other prenylated proteins that is responsible for the anti-tumor activity of prenyltransferase inhibitors. Methods to detect the specific proteins that are affected by prenyltransferase inhibition will be required to develop more potent and specific inhibitors as well as to design clinical trials with patients that are likely to have the most success with prenyltransferase inhibitor treatment. Some strategies for addressing this goal have been developed, including various proteomic approaches and expression analysis (see references 40,91–97 for examples). More work needs to be conducted to identify specific proteins affected by prenyltransferase inhibition for more effective cancer therapeutics to be developed that will impact cancer therapy significantly.

4.1 Hutchinson-Gilford Progeria

Hutchinson-Gilford progeria syndrome (HGPS, progeria) is a rare, autosomal dominant, fatal disorder characterized by early onset aging.⁹⁸ Approximately 1 in 4,000,000 children is stricken with this disease and display other phenotypes as a result, such as osteoporosis and alopecia. Those suffering from progeria have a mean survival time of 12.6 years.⁹⁹ This disease is the result of a single base pair mutation (C to T) at nucleotide 1824 in the gene (LMNA) coding for the nuclear lamin A protein.^100,101 The lamin A protein is initially synthesized as a precursor protein, called prelamin A, that undergoes a series of modifications (Figure 13). First, prelamin A is farnesylated, followed by proteolysis of the ‘AAX’ residues and methylation of the C-terminus. Lastly, an internal cleavage by the enzyme ZMPSTE24 gives rise to non-farnesylated, mature lamin A.¹⁰² In HGPS, the single base pair mutation results in a cryptic mRNA splice site that results in an internal 50 amino acid deletion in prelamin A.¹⁰³ Because this form of prelamin A still contains the CAAX box, it is still farnesylated, proteolyzed (by –AAX removal), and methylated; however, the internal 50 amino acid deletion causes the loss of the ZMPSTE24 internal cleavage site resulting in a permanently farnesylated form of lamin A called progerin (Figure 13).¹⁰⁴ Progerin is targeted to the nuclear rim where it interferes with the nuclear lamina, causing mis-shaped nuclei and chromosomal disorganization.¹⁰⁵

4.2 Possible therapies for progeria using FTIs

The finding that the permanently farnesylated progerin was responsible for the phenotype of progeria led to considerable interest in the use of FTIs for treating this disease;¹⁰⁶ if farnesylation of prelamin A could be prevented, progerin would not form. Initial studies testing this hypothesis in cell culture were successful; the nuclear blebbing phenotype of HGPS was lost when fibroblasts from patients were treated with an FTI.¹⁰⁷ Following these in vitro studies, a mouse model of HGPS was developed, and it was shown that treatment of these mice with an FTI improved progeria-like disease phenotypes, such as rib fractures and reduced bone density).^106,108,109 However, the amelioration of these symptoms was not complete, and eventually all of the mice developed severe disease and died. Importantly, the amelioration of the progeria-like symptoms in these mouse models was directly due to blocking the farnesylation of progerin with an FTI.¹¹⁰

The prospect of using FTIs as a treatment for progeria in the clinic has been questioned recently due to the finding that progerin can be alternatively geranylgeranylated in the presence of an FTI.¹¹¹ Varela and coworkers used a combination of a statin, to deplete endogenous isoprenoids, and a bisphosphonate, to inhibit the farnesyl diphosphate synthase enzyme. This combination served to deplete farnesylated proteins, including lamin A.¹¹¹ These investigators showed that this combination improved progeria-like disease phenotypes in mice. Based on the current evidence, it appears that any therapy targeting protein farnesylation may improve the symptoms of progeria, but it is unclear whether it will completely ameliorate the disease in humans.

4.3 Clinical trials for the treatment of progeria with FTIs

Characterization of all HGPS phenotypes, especially those resulting in vascular problems, has been performed^112,113 and clinical trials are underway to test the treatment of progeria with FTIs. The first was a phase 2 trial, started in 2007, that enrolled 26 children to receive the FTI lonafarnib.¹¹⁴ The trial ended in early 2010 and the results have recently been published.¹¹⁵ In this study, children with progeria were administered lonafarnib every 12 hours for up to two years. The therapy was well tolerated and no children were removed from the study because of toxicity.¹¹⁵ Exciting successful outcomes of the study included decreases in arterial pulse wave velocity and carotid artery echodensity and increases in skeletal rigidity and sensorineural hearing, one or more of which occurred in all children in the study.¹¹⁵ For instance, the carotid-femoral pulse wave velocity, PWV_cf, measures the arterial stiffness in an individual, which normally increases with age. Those suffering from HGPS typically have PWV_cf values equivalent to individuals aged 60–69 (3.5 times greater than normal pediatric values); in this study the PWV_cf decreased in 18 patients by a median of 35%.115 Furthermore, areal bone mineral density (aBMD) demonstrated a clinically significant ≥ 3% increase from pretherapy to end of therapy at one or more sites in 76% of children.¹¹⁵ Although therapy with lonafarnib is not a cure, these results suggest that it may be a suitable therapy to improve vascular stiffness and bone structure in patients with progeria. Finally, it should also be noted that while there is good evidence to support the notion that inhibition of lamin A farnesylation is physiologically relevant for the treatment of this disease, the authors of the above study noted that there was “a lack of correlation between the pharmacodynamics marker used (HDJ-2) and clinical response”. This observation highlights the need for improved methods for examining the effects of FTIs on the prenylation status of multiple proteins simultaneously as noted in section 3.4.

A second phase 2 clinical trial began in 2009; 45 patients with HGPS received a triple combination of zoledronic acid (a bisphosphonate), pravastatin (a statin), and lonafarnib (an FTI).¹¹⁶ The trial lasted two years and was scheduled for completion in February 2012, with the results of the trial now pending. Because lonafarnib showed earlier success as a single agent in clinical therapy,¹¹⁵ using combination therapy to ensure complete farnesylation inhibition is an optimistic approach the may prove beneficial. The results from this study as well as the previous study should provide valuable information to better develop treatments for Hutchinson-Gilford progeria syndrome.