Article Text

Download PDFPDF

Small-molecule signal-transduction inhibitors: targeted therapeutic agents for single-gene disorders
  1. D Mark Davies,
  2. Julian R Sampson
  1. Institute of Medical Genetics, Department of Genetics, Haematology and Pathology, School of Medicine, Cardiff University, Cardiff, UK
  1. Correspondence to Dr D Mark Davies, Institute of Medical Genetics, Cardiff University, Cardiff CF14 4XN, UK; daviesdm{at}cardiff.ac.uk

Abstract

Mutations affecting over 2000 of the 20 000 or so genes in the human genome have been linked so far to specific inherited diseases, most of which are rare and have been poorly understood. Many of the genes involved encode components of intracellular signalling pathways that regulate processes such as the growth, proliferation, differentiation and survival or programmed death of cells during development and the maintenance of tissues and organs. Mutations that change the function of genes encoding signalling proteins thereby cause disorders ranging from birth defects to cancer. For Mendelian disorders, the essentially causal relationship between mutation and disease may present direct opportunities to therapeutically manipulate intracellular signalling. Here, we review recent examples of the use of small-molecule drugs to target components of signalling networks in single-gene disorders. We also consider the limitations of these “molecularly targeted” approaches and the difficulties in their clinical development as therapies for rare genetic diseases.

  • Molecularly targeted therapy
  • inherited disease
  • getting research into practice
  • clinical genetics
View Full Text

Statistics from Altmetric.com

Targeting tumours

Mutations in oncogenes and tumour suppressor genes result in the deregulation of intracellular signalling pathways that control apoptosis, proliferation, angiogenesis, metabolism, stress responses and other cellular functions. In recent years, many agents that manipulate intracellular signalling have been developed as potential cancer treatments. The successes of signal transduction inhibitors such as imatinib, gefitinib and erlotinib show this approach can yield valuable clinical benefit, but experience of their use has highlighted the importance of patient selection.1 When signal-transduction inhibitors have been used in patient groups defined by tumour histology, the results have been generally disappointing, reflecting underlying heterogeneity in molecular pathology. In contrast, successes have occurred when these drugs have been used against tumours that share a specific molecular defect. Examples include gastrointestinal stromal tumours, most of which carry mutations in the c-kit oncogene, the target for imatinib, and non-small cell lung cancer in which specific mutations in epidermal growth factor receptor (EGFR) predict response to gefitinib or erlotinib.2 3 Such tumours exhibit a dependence upon a specific aberrant signalling pathway or gene product, and this phenomenon has been termed “oncogene addiction” or “tumour suppressor hypersensitivity”, depending on the nature of the responsible gene.4 5 Tumours arising in specific Mendelian tumour predisposition syndromes share at least one key molecular abnormality, for example, mutations in the tumour suppressors TSC1/2 in tuberous sclerosis, VHL in Von Hippel-Lindau disease or NF1 in neurofibromatosis type 1. This relative molecular homogeneity may facilitate selection of therapeutic agents to target the pathways fundamental to tumour growth and/or survival.

Tuberous sclerosis is an autosomal dominant disorder that has a prevalence of approximately 1 in 10 000 and very variable manifestations including seizures, intellectual disability, autism and the growth of benign tumours in multiple organs.6 Women with tuberous sclerosis are at risk of developing lymphangioleiomyomatosis (LAM), a disorder characterised by the progressive cystic destruction of lung tissue and proliferation of abnormal smooth muscle-like “LAM” cells. LAM also occurs as a very rare sporadic disease, generally in premenopausal women. Both tuberous sclerosis and sporadic LAM are associated with renal angiomyolipomas, benign tumours composed of adipose cells, smooth muscle cells and blood vessels. It has been suggested that pulmonary LAM cells may arise in renal angiomyolipomas (or another remote site) and spread to the lung by a process that has been termed “benign metastasis”.

Inherited mutations in either TSC1 on chromosome 9 or TSC2 on chromosome 16 cause tuberous sclerosis. Consistent with Knudson's two hit tumour suppressor model, acquired loss of the second allele of TSC1 or TSC2 has been documented in most types of tuberous sclerosis-related tumours.7 Sporadic LAM is caused by acquired biallelic mutations, generally in TSC2.8 The protein products of TSC1 and TSC2 form a complex within cells that inhibits the mammalian target of rapamycin complex 1 (mTORC1), a mediator of many cellular processes including growth, autophagy and cell cycle progression.9 Loss of TSC1 or TSC2 activates mTORC1 signalling and leads to tumour growth (figure 1). Sirolimus (also known as rapamycin) is an inhibitor of mTORC1 and is used clinically as an immunosuppressant in transplant recipients. When research showed that sirolimus normalised mTORC1 signalling in TSC1/2-deficient cells and that tumour growth was inhibited in mice engineered to carry TSC1 or TSC2 mutations,10 11 phase II clinical trials were initiated to look at the efficacy of sirolimus in patients with tuberous sclerosis or sporadic LAM in the USA and Europe.12 13 These have recently reported significant reduction in the size of angiomyolipomas in both groups of patients. On the basis of these findings, randomised control trials have been initiated or to assess sirolimus or related derivatives as treatments for a variety of TSC- or LAM-related clinical problems.

Figure 1

The mammalian target of rapamycin complex 1 (mTORC1) and mitogen-activated protein kinase (MAPK) signalling network. This simplified diagram shows how multiple inputs from a number of stimuli converge at the TSC1/TSC2 complex. Signalling molecules encoded by genes mutated in Mendelian disorders are coloured light blue. These include Noonan syndrome (SOS1, SHP2, RAF1, KRAS, BRAF), neurofibromatosis type 1(NF1), Costello syndrome (HRAS), LEOPARD syndrome (RAF1, BRAF), cardiofaciocutaneous syndrome (KRAS, BRAF, MEK1, MEK2), tuberous sclerosis (TSC1, TSC2), PTEN hamartoma syndromes (PTEN), Peutz-Jegher syndrome (LKB1) and familial cardiac hypertophy (AMPK).

Targeting cysts

Sirolimus is being investigated as a potential treatment for autosomal dominant polycystic kidney disease (ADPKD), the most common hereditary cause of end-stage renal disease. ADPKD is caused by mutation in PKD1 on chromosome 16, encoding polycystin-1(PC1), and PKD2 on chromosome 4, encoding polycystin-2 (PC2).14 PC1 and PC2 interact with multiple signalling pathways. However, inappropriate activation of mTORC1 has been demonstrated in the epithelium lining a proportion of renal cysts from human patients with APKD and in multiple rodent models, suggesting this is a common, convergent event in renal cystogenesis.15 Sirolimus reduced cyst size in two mouse models of polycystic kidney disease and slowed disease progression in a rat model.15 16 In a retrospective analysis of patients with APKD who had received a renal transplant, use of sirolimus as an immunosuppressant was associated with a reduction in size of the native polycystic kidney.15 On the basis of these results, a number of trials of sirolimus in patients with ADPKD have been initiated.

In renal epithelial cells both PC1 and PC2 localise to the primary cilia and have been suggested to act as mechanosensors, responding to luminal flow, and to influence cell proliferation and planar cell polarity.17 Mutations in genes involved in cilia function cause diverse disease phenotypes and work is ongoing to establish if mTORC1 activation is a common event in these conditions. Inhibition of mTORC1 may have therapeutic potential not only for polycystic kidney disease but also for a wide range of ciliopathies.18

Targeting the brain

Many Mendelian disorders are associated with cognitive deficits. Recently, interest has become focused on whether some of these deficits might be amenable to the therapeutic targeting of intracellular signalling pathways.

Individuals with tuberous sclerosis manifest a range of neuropsychological and behavioural phenotypes,19 which may reflect an interplay between the effects of seizures, developmental structural abnormalities of the brain, mass lesions and abnormal neuronal function.20 21 It has been suggested that inhibiting mTORC1 signalling in the brain may be sufficient to improve cognitive function in tuberous sclerosis. Ehninger et al22 investigated the effects of sirolimus on cognitive function in a heterozygous Tsc2 +/− mouse model, which does not manifest macroscopic or microscopic brain lesions or seizures but none the less shows learning and memory deficits. Treatment of adult Tsc2 +/− mice with sirolimus for 5 days ameliorated these cognitive deficits. In another mouse model, in which Tsc1 is conditionally knocked out primarily in glial cells, sirolimus treatment normalised mTORC1 signalling and neuronal disorganisation and suppressed seizures.23 These findings suggest that mTORC1 inhibitors may have a therapeutic role in improving cognitive function and seizures in tuberous sclerosis.

Inhibition of mTORC1 has also been proposed as a therapeutic strategy for neurodegenerative conditions such as Huntington disease where the formation of abnormal protein aggregates may contribute to pathogenesis.24 Inhibition of mTORC1 with sirolimus in mouse models of Huntington disease enhanced the clearance of these proteins, possibly by upregulating autophagy, and ameliorated behavioural abnormalities.25

The difficulties of extrapolating from mouse models to humans, particularly in the area of cognitive function, has been illustrated in a recent trial of targeted therapy in neurofibromatosis type 1 (NF1).26 This autosomal dominant condition is characterised by a number of neurocutaneous manifestations, including tumour growth and the frequent occurrence of intellectual disability and specific neuropsychological deficits. NF1 is caused by mutations in the NF1 gene, which encodes a GTPase-activating protein whose normal role is to regulate RAS signalling.27 Increased RAS signalling has been implicated in the neuronal plasticity defects and spatial learning and attention problems seen in NF1 mouse models.28–30 RAS signalling can be inhibited by farnesyl transferase inhibitors and 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors, which block the isoprenylation required for RAS activity. Treatment of NF1 mouse models with these agents for just a few days was found to lead to a reversal of cognitive function deficits.28–30 On the basis of these findings, Krab et al26 conducted a clinical trial to assess the effect on cognitive function of the HMG-CoA reductase inhibitor simvastatin in children with NF1. Sixty-two children were randomised to receive either simvastatin or placebo for 12 weeks. However, no significant differences were observed in the primary outcome measures of cognitive function.

Targeting tissue integrity

The maintenance of cell and tissue integrity requires a balance between cell death and division, control of cell differentiation and regulated interplay between cells and the extracellular matrix. Marfan syndrome is characterised by abnormalities of the skeletal, cardiovascular and ocular systems and affects approximately 1 in 20 000 individuals.31 It is caused by mutations in the FBN1 gene, which encodes fibrillin-1. This microfibril protein contributes to the mechanical integrity of organs and the vasculature. Aortic root dissection is a major cause of mortality in patients with Marfan syndrome. Recent studies have shown that aortic aneurysm in Marfan syndrome may arise not simply as a result of inherent weakness of connective tissues due to deficiency of fibrillin-1 but also because of defective interaction between fibrillin-1 and transforming growth factor beta (TGFβ), leading to impaired tissue maintenance.32 The TGFβ family of growth factors regulate many cellular processes such as proliferation, differentiation and survival. TGFβ is normally sequestered by fibrillin-1 and is released in a tightly controlled manner. Deficiencies of fibrillin-1 may thereby lead to excessive TGFβ activity and promote processes such as matrix degradation, which may contribute to vascular pathology.33 Analysis of Fbn1 knockout mouse models that develop fatal aortic aneurysms showed upregulation of TGFβ signalling in affected tissue, and treatment of these mice with TGFβ neutralising antibodies reduced aortic root dilatation.34 TGFβ-neutralising antibodies are not currently available for clinical use in humans. However, another way of reducing TGFβ signalling is to target the angiotensin pathway with a small-molecule angiotensin II type 1 receptor blocker such as losartan, a widely used antihypertensive agent. AT1-receptor blockade decreases TGFβ-mediated intracellular signalling by incompletely understood mechanisms that may involve a direct effect on TGFβ synthesis and modulation of cross-talk between signalling pathways.35 36 Habashi et al37 treated fibrillin-1 knockout mice with losartan, the β-blocker propranolol or placebo. The doses of losartan and propranolol were titrated to obtain comparable changes in blood pressure. Propranolol reduced the aortic root dilatation compared with placebo, but losartan completely inhibited aortic root dilatation, and losartan-treated animals were indistinguishable from wild-type controls. In a small-cohort study of 18 patients with Marfan syndrome, treatment with an angiotensin II type 1 receptor blocker was associated with a dramatic slowing in the rate of aortic root dilatation.38 In light of these promising results, a number of clinical trials of losartan for Marfan syndrome have begun.

TGFβ inhibition may have a therapeutic role in a wider range of inherited conditions. Increased TGFβ activity has been implicated in the pathogenesis of the X-linked Duchenne and Becker muscular dystrophies, caused by mutations in the DMD gene. These conditions are characterised by progressive skeletal muscle weakness and wasting. The more severe Duchenne form is usually fatal by early adult life. Early in the disease both muscle cell death and regeneration are seen, but regeneration slowly fails and fibrogenesis occurs.39 This process has been linked to increased TGFβ activity.40 41 Treatment of DMD knockout mice with losartan resulted in improvement of muscle architecture and function by maintaining muscle regeneration.41 These findings have led to proposals for clinical trials of losartan for the human disease.

Treatment in utero using molecularly targeted agents

Many Mendelian disorders are characterised by congenital anomalies, which reflect abnormal development in utero. Recent studies in mouse models have suggested that molecularly targeted therapies might be used in utero to treat malformation syndromes. Apert syndrome is characterised by craniosynostosis and syndactyly of the hand and feet. It is caused by activating mutations in fibroblast growth factor receptor 2 (FGFR2), most commonly S252W.42 43 Such mutations upregulate a number of downstream signalling pathways including the mitogen-activated protein kinase (MAPK) pathway.44 Shukla et al45 developed a mouse model of Apert syndrome in which embryonic expression of a mutant FGFR2 allele led to craniosynostosis. Introduction of a short hairpin RNA (shRNA) complementary to the mutant allele prevented its expression, normalised signalling in the MAPK pathway, and corrected the phenotype. The therapeutic use of RNA interference to knock down the expression of target genes is an exciting prospect, but issues of delivery and safety will have to be overcome.46 However, Shukla et al45 also treated pregnant transgenic mice with U0126, a small-molecule MEK1/2 inhibitor, and this led to complete repression of craniosynostosis in affected embryos. Cranial abnormalities emerged in the postnatal period, but were ameliorated by reinitiating treatment with U0126.

MAPK pathway inhibitors may have a role in other inherited disorders that exhibit increased MAPK signalling such as the phenotypically related conditions cardiofaciocutaneous, Costello and Noonan syndromes, all caused by mutations in genes encoding components of the MAPK pathway47 (figure 1). However, significant difficulties exist for the translation of in utero therapy to the clinical setting, including the need for sufficiently early prenatal diagnosis and the risk of drug-related teratogenicity.

Challenges to molecularly targeted therapy

Simply inhibiting one component of a perturbed signalling pathway may fall far short of restoring normal signalling, with its nuanced responses to multiple stimuli, and in practice it is difficult to titrate dose against molecular response. Unsurprisingly therefore, drugs that target important intracellular signalling networks inevitably have side effects. For example, sirolimus is an immunosuppressant, impairs wound healing, and causes a variety of metabolic abnormalities, particularly hyperlipidaemia. These effects will potentially restrict patient selection, dose and the duration of therapy. Furthermore, many signal transduction inhibitors are “promiscuous” drugs, inhibiting multiple signalling molecules, and these off-target effects may also contribute to a wide side-effect profile.

Feedback mechanisms and cross-talk characterise many signalling networks and may lead to unintentional effects when a pathway is targeted. For example, multiple feedback loops exist in the TSC/mTORC1 pathway, and normalisation of mTORC1 signalling could promote cell survival via feedback effects on AKT in TSC1/2-deficient cells.48 Functional redundancy between pathways may also limit drug efficacy. Often, drugs will not be available to target directly the protein encoded by a mutated gene but rather a component downstream of it or even in a related pathway. Sirolimus does not directly target the TSC1/TSC2 complex but rather downstream mTORC1, and the drug is not expected to correct any mTORC1-independent functions that TSC1/TSC2 may have. Combination therapy or the use of “promiscuous” drugs to target multiple signalling pathways or multiple components of the same pathway may overcome some of these difficulties.

As not all types of signalling molecules are currently “drugable”, new classes of drugs that target protein–protein interactions or protein–DNA binding are likely to be required to expand the range of genetic diseases amenable to a molecularly targeted approach.

Clinical development of molecularly targeted therapy for single-gene disorders

Conducting clinical research in rare genetic conditions is challenging as the patient population is usually small and geographically dispersed and may include vulnerable groups such as children or people with intellectual disability. The Committee for Medicinal Products for Human Use (CHMP) has produced guidelines for clinical trials in small populations which recognise the difficulties of performing adequately powered trials in these settings.49 For extremely rare disorders, the combined evaluation of single-case studies may be the only source of evidence. CHMP recommends that such studies adhere to guidelines for good clinical practice in clinical trials and ideally standardise treatment conditions and data collection to facilitate systematic review. Innovative and efficient trial design and data analysis methods may allow randomised control trials to be carried out in rare disorders. Surrogate markers of outcome may have to be used as alternatives to clinical endpoints to allow adequately powered studies to be carried out. For conditions where potential participant numbers will not support randomised control trials a higher degree of uncertainty in the evidence base will have to be accepted by regulatory authorities and for clinical decision making. Patient registers will have an important role to play as therapeutic interventions become available for increasing numbers of rare inherited disorders, providing historical controls, information on the natural course of a disease and facilitating recruitment to trials. Regional clinical genetic services have already developed registers for some disorders and recently the families of patients affected by Duchenne muscular dystrophy were instrumental in setting up a national register of patients and their genotypes specifically to facilitate clinical trials.50 Clinical trials will often require international cooperation and coordination, and this would benefit from further simplification and standardisation of regulatory and ethical approval processes.

Cost is a major challenge in the development of novel agents to treat rare genetic diseases. Pharmaceutical companies may select to explore drug development in some rare but biologically informative conditions to open niche markets or to gain insights into disease processes relevant to more common conditions. Worldwide a number of programmes have been introduced to promote the development of “orphan” medicinal products for rare diseases. The EU defines an “orphan” medicinal product as a product intended for the diagnosis, prevention or treatment of life-threatening or chronically debilitating conditions that affect no more than 5 in 10 000 people in the European Union, or one which, for economic reasons, would be unlikely to be developed without incentives. The EU ‘orphan’ medicinal products programme provides community wide economic incentives to pharmaceutical companies such as market exclusivity provision, regulatory fees waivers and protocol assistance.

Most drugs currently in trials for genetic disorders have already been used to treat other sporadic conditions. It is likely that repositioning of such agents will remain the most realistic strategy for future treatment of many of the disorders that constitute the workload of clinical genetics services. This repositioning may be viewed by pharmaceutical companies as a relatively low-risk strategy to optimise product pipelines, but conflict may arise between the interests of these companies and those of healthcare purchasers for older, well-established drugs nearing the end of their patent period compared with newer, more expensive derivatives.

Despite programmes to support orphan medicinal product development, economic considerations mean that, for many single-gene disorders, translational research will have to be driven from outside the pharmaceutical industry. Clinical and laboratory genetics services have been quick to translate knowledge of the human genome into diagnostic advances. We suggest that they should now accept the challenge of delivering the therapeutic advances for inherited disorders that were anticipated by the Human Genome Project.

References

View Abstract

Footnotes

  • Funding The Trial of Efficacy and Safety of Sirolimus for Tuberous sclerosis and Lymphangioleiomyomatosis (TESSTAL) is funded by the Tuberous Sclerosis Association, the James Tudor Foundation and the Wales Gene Park. Sirolimus is provided free of charge by Wyeth. DMD has also received a contribution to laboratory research costs from Wyeth.

  • Competing interests Both authors participate in research into genetic conditions and could benefit from an increase of funding in this area.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.