Abstract
In vertebrates, the liver is the central metabolic organ of the body, which carries out an estimated 500 functions that range from general detoxification to protein synthesis, bile production, metabolism of fats, carbohydrates, proteins, bilirubin, vitamin and mineral storage and it even has an immune function. Hepatocytes are considered the professional liver cells, which carry out all of these functions. With such a variety of tasks to perform, it is not surprising that more than 400 rare monogenic disorders of hepatic origin have been described. For many of these, liver transplantation remains the only curative strategy, however, this is limited by organ availability and requires lifelong immune suppression. The fact that liver transplantation is curative led to the assumption that the restoration of the expression of the defective gene would result in the resolution of the disease. Indeed, liver-directed gene therapy trials for hemophilia A and B have demonstrated the potential of gene therapy to provide long-lasting clinical benefit in the treatment of monogenic liver disorders. Thus, liver-directed gene therapy and gene editing strategies have emerged as promising alternatives to transplantation in inherited monogenic liver disorders. Herein, we review the advances and limitations of gene therapy for such disorders, covering therapeutic strategies based on gene addition and gene editing and the exciting clinical results obtained with the use of ribonucleic acid as therapeutic molecules.
General Introduction
In vertebrates, the liver is the central metabolic organ of the body, which carries out an estimated 500 functions that range from general detoxification to protein synthesis, bile production, metabolism of fats, carbohydrates, proteins, bilirubin, vitamin and mineral storage, and it even has an immune function. Hepatocytes are considered the professional liver cells that carry out all of these functions. With such a variety of tasks to perform, it is not surprising that more than 400 rare monogenic disorders of hepatic origin have been described. For many of these, liver transplantation (LT) remains the only curative strategy; however, this is limited by organ availability and requires lifelong immune suppression.
The fact that LT is curative led to the assumption that the restoration of the expression of the defective gene would result in the resolution of the disease. Indeed, liver-directed gene therapy trials for hemophilia A and B have demonstrated the potential of gene therapy to provide long-lasting clinical benefit in the treatment of monogenic liver disorders. Thus, liver-directed gene therapy and gene editing strategies have emerged as promising alternatives to transplantation in inherited monogenic liver disorders.
Herein, we review the advances and limitations of gene therapy for such disorders, covering therapeutic strategies based on gene addition and gene editing and the exciting clinical results obtained with the use of ribonucleic acid (RNA) as therapeutic molecules.
Inherited Metabolic Liver Diseases
The liver is a critical organ for most metabolic pathways and thus is the target tissue for many inherited metabolic liver diseases (IMLDs). IMLDs are a diverse group of rare genetic disorders and, in many cases, the cause is a single gene mutation, occurring in ∼1 in 800 live births. 1 They are typically inherited as autosomal recessive, but can also occur by X-linked or codominant inheritance. 2 –4 A common feature of all these disorders is a deficiency in the synthesis or in the functionality of proteins involved in biochemical pathways essential for metabolism and have enzymatic, receptor, or transporter functions. 5 In general, the defect causes the accumulation of toxic substrates or failure to synthesize downstream essential nutrients. Conventional enzyme replacement therapy is generally not an option for these patients and the only curative treatment option is liver transplantation (LT), which remains high risk. 6,7 Long-term complications associated with chronic immunosuppression to prevent graft rejection are common.
Figure 1 shows hepatic IMLDs for which LT is the only curative solution. All of them are considered rare diseases (a disease is defined as rare in Europe when it affects <1 in 2000, and in the United States when it affects fewer than 200,000 Americans at any given time, however, prevalence can widely vary. Table 1 gives examples of some IMLDs, including their prevalence, characteristics, and presently available treatments. 8,11

Schematic representation of inherited metabolic liver diseases, for which liver transplantation remains the only curative strategy.
Inherited metabolic liver diseases: main characteristics and the treatments currently available
Source: Orphanet.
LKT, liver-kidney transplantation; LT, liver transplantation.
Gene Delivery to the Liver
A growing toolbox has become available for liver-directed gene delivery. Although viral vectors have long been the preferred approach to target hepatocytes, an increasing number of nonviral vectors are emerging as highly efficient delivery vehicles, in particular for RNA molecules.
The most simple gene delivery method to the liver is hydrodynamic injection, a technique based on the injection of a relatively high volume of a solution containing DNA or RNA. Fast injection of a large volume induces an increase in hepatic pressure and facilitates the entrance of the genetic material into the hepatocytes. This technique is very frequently used in mice for proof-of-concept studies and has also been tested in large animal models such as pigs, sheep, or nonhuman primates (NHPs). 45 –49 Another technique for delivery of naked genetic material is ultrasound-mediated gene delivery in combination with microbubbles. 50 –52 It has proven to be an effective method for hepatic gene delivery and was recently tested in a canine hemophilia model. 53
The delivery of packaged genetic material includes nonviral and viral vectors. The use of nonviral vectors has several advantages over the use of viral vectors: easy production, unlimited size of the transgene, and a reliable safety profile. Limitations are the stability of these particles, their cellular uptake, and thus far a limited ability to achieve long-lasting transgene expression. 54,55 For hepatic gene delivery, the delivery particles must possess a diameter of 100 nm or less to be able to cross fenestrated liver endothelial cells and gain access to hepatocytes. The use of galactosylated polyethylenimine or chitosan, targeting the hepatic asialoglycoprotein receptor (ASGPR), proved to be very efficient in transducing human hepatic cell lines and the livers of mice and rats. 56 –58 Very recently, the use of liver-targeting lipid nanoparticles (LNP) was demonstrated to be a potential new generation of safer, although temporary, treatments to restore metabolic liver functions in patients. 54,55
Viral vectors that have been used for liver-directed gene therapy include adeno-associated virus (AAV), lentivirus (LV), and high-capacity adenoviral (HC-Ad) vectors.
Several features of AAV vectors make them ideal for in vivo liver-directed gene therapy as, for example, they can efficiently transduce quiescent cells and have a good safety profile. 59 The majority of AAV-delivered transgene expression cassettes generally remain episomal and do not integrate and thus carry a reduced risk for insertional mutagenesis. Compared with other viral vectors AAV vectors are considered weakly immunogenic as they do not elicit strong innate or T cell responses. The main limitations of AAV vectors are their relatively small cloning capacity (4.5 kb) and pre-existing humoral immunity. A large proportion of the population experienced natural infection during their lives and existing anti-AAV antibodies presently preclude their suitability for AAV-based therapies. 60,61 Furthermore, studies performed in mice with humanized livers and nonhuman primates revealed important differences in the transduction efficiency of the different AAV serotypes between human/NHP and murine hepatocytes. 62 The selection of new capsids with increased capacity to transduce the human liver is a field of intense research. 63
First-generation Ad vectors are highly efficient in transducing hepatocytes; however, they induce a strong immune response that leads to the rapid rejection of transduced cells. 64 Third-generation adenovirus vectors, or helper-dependent adenovirus (HD-adenovirus) in which all viral sequences were removed except for the terminal and packaging signal sequences, allow delivery of large genes (up to 36 kb) and provide long-term transgene expression in mice and NHPs. 65 –67 Furthermore, they are highly efficacious in transducing the liver. 68,69 Today, the lack of an appropriate clinical-grade production platform remains the main limitation of these vectors for clinical use. 64
Because both AAV and HC-Ad vectors do not actively integrate into the host cell genome, they are diluted upon cell division in the growing liver or during regeneration after liver injury. This makes their use in pediatric patients very challenging. LVs, on the contrary, integrate into the target cell chromatin and replicated as cells divide. In the clinic, LVs have been successfully used for the correction of hematological disorders, but their use for the treatment of liver disease is very limited. 70,71 More recently, systemic administration of an LV to adult mice and dogs resulted in selective targeting of the liver and spleen and stable factor IX expression in the liver. However, mild acute toxicity and low efficacy were observed in dogs as a result of LV uptake by phagocytes. 72 Strategies to avoid uptake by phagocytic cells are at present under investigation to improve hepatocyte transduction and to reduce systemic toxicity. The incorporation of human phagocytosis inhibitor CD47 into lentiviral particles resulted in a significant increase of transgene expression and a reduction of toxicity in NHP. 73 These studies provide further support by preclinical assessment and clinical evaluation of liver-directed LV-based gene therapy. In addition to in vivo correction, LVs were and are more frequently used for ex vivo modification of hepatocytes. 74
Gene Addition Strategies for Inherited Liver Diseases
Recombinant AAV vectors
Because of their many advantages, AAV vectors have emerged as the leading candidates for gene therapy in IMLD. After a number of studies demonstrated the feasibility of delivering genes to the liver in animal models of hemophilia, 75 the first hepatic IMLD treated with an AAV vector was the familial hypercholesterolemia (FH). 76 Infusion of an AAV vector carrying the very-low-density lipoprotein (VLDL) receptor (VLDLR) into the portal circulation of FH mice resulted in a 40% reduction in serum cholesterol and triglyceride. Importantly, the reduction was stable for the duration of the study. Since then, the number of studies reporting therapeutic efficacy in IMLD animal models has increased exponentially, and proof-of-concept studies include acute intermittent porphyria (AIP), 77 –80 primary hyperoxaluria type 1 (PH1), 81 Wilson disease (WD), 82 –84 methylmalonic academia (MMA), 85 propionic academia (PA), 86 –88 phenylketonuria (PKU), 89 –92 tyrosinemia type 1 (HT1), 93,94 progressive familial intrahepatic cholestasis 3 (PFIC3), 95 Crigler–Najjar syndrome (CN), 96 –103 glycogen storage disease 1a (GSD1a), 104 –108 and urea cycle disorders. 109 –117 Although different AAV serotypes, promoters, and delivery routes were used, in general, all studies showed long-term therapeutic efficacy in the absence of safety issues.
Recombinant HD-adenoviral vectors
The use of HD-adenovirus expressing the LDL-receptor represents a promising strategy for the treatment of homozygous FH (HoFH). The injection of a single dose of vector improves lipid profile and reduces aortic atherosclerosis in rodents and nonhuman primates. 66,69 The therapeutic efficacy of this vector has been also clearly demonstrated in animal models of PH1, 118 CN, 119 –121 AIP, 122 GSD1a, 123,124 and ornithine transcarbamylase (OTC). 125,126 However, clinical data are still pending for this highly liver tropic vector mainly due to manufacturability difficulties.
Recombinant lentiviral vectors
Pioneering studies based on systemic administration of LV were performed for the treatment of CN and hypercholesterolemia. 127 –130 However, after systemic injection, genotoxicity was observed after fetal injections of mice and nonhuman primates and the initial expectation declined. 130,131 The risk of insertional mutagenesis associated with LV has meanwhile been diminished with the development of self-inactivating (SIN) vectors, in which long terminal repeat enhancer/promoter elements have been deleted. The capacity of LV to cure was shown in HT1 in utero gene therapy in a large animal model 132 and in mouse models of MMA and GSD1a. 133,134
Furthermore, LV-mediated liver-directed ex vivo gene therapy has been successfully reported in a pig model of HT1, and in CN rats. 135 –138 In HT1, the modified hepatocytes using their selective advantage repopulated the diseased organ. The corrected cells generated enough liver mass to not only clinically ameliorate the disease but to fully normalize tyrosine levels and liver function tests. 138
Treatment of IMLD based on the administration of RNA molecules
Nucleoside-modified messenger RNA (mRNA) encapsulated in liver-targeting LNP can be successfully delivered to the liver by intravenous (i.v.) administration. In the last 2–3 years, the therapeutic efficacy of this approach has been demonstrated in preclinical mouse models of a number of IMLDs, such as MMA, 139 PH1, 140 GSD1a, 141 citrine deficiency, 142 and AIP. 143 Recently, at the annual meeting of the ASGCT (2019), the efficacy of mRNA-LNP was reported for additional indications such as maple syrup urine disease, 144 arginase deficiency, 145 OTC, 146 and PFIC3. 147 In general, the duration of action of a single i.v. dose of mRNA-LNP lasted only between 2 and 3 weeks and required the repeat administration of multiple doses to prolong a curative effect. The expression of the functional proteins attenuated, prevented, or normalized the characteristic metabolic abnormalities typical for these disorders. These encouraging results support further studies to investigate the potential of mRNA therapeutics as possible treatments for IMLDs.
Treatment of liver diseases based on gene silencing
As described, some of the inherited hepatic metabolic disorders lead to the accumulation of toxic metabolites. Patients affected by these disorders can benefit from therapeutic strategies based on the reduction of the precursor molecules, also called substrate reduction therapies (SRT). The principle of SRT is to decrease the synthesis of any toxic intermediates that accumulate because of primary defects in their metabolic pathway, often due to loss-of-function mutations in genes encoding a key metabolic enzyme. 148 Inhibition of the pathway upstream of the lacking/defective enzyme results in the reduction of the toxic substrate. The inhibition of the upstream enzymatic activity can be achieved by using chemical compounds, and also by the reduction of enzyme expression by means of RNA interference (RNAi). 149 RNAi is a naturally occurring process that makes use of double-stranded RNA molecules (siRNAs) to inhibit the expression of specific mRNAs. 150 siRNAs can be modified by the addition of ligands specific for hepatocytes allowing high-efficiency liver targeting after systemic administration. 151 This strategy has been used for the treatment of PH1 and AIP in preclinical animal models and in clinical trials. In PH1, the administration of liver-targeted siRNA molecules targeting LDH and hydroxyacid oxidase 1 (HAO1), which encode proteins that act upstream of AGXT, significantly reduced the accumulation of oxalate. 152 –154 In AIP, inhibition of the enzyme that regulates the heme synthesis pathway resulted in significant reduction of a toxic porphyrin precursor. 155,156
Curing Genetic Disorders by Gene Editing of Hepatocytes
Genome or gene editing is a genetic engineering technique used to modify the DNA of a cell in a specific and precise way. Gene editing tools have greatly evolved in the past decades, opening a landscape for the treatment of many genetic disorders. Three major platforms are used to target specific DNA sequences: zinc finger nucleases (ZFN), transcription activator-like effector nucleases (TALEN), and clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (CRISPR/Cas9). 157
Because it is very easy to target the liver efficiently, it makes an ideal organ for the experimental application of gene editing-based therapies in animal models of disease. The discovery of the potential of CRISPR/Cas9 systems to edit eukaryotic cells 158 has given a huge push for liver gene editing, because it is a simple, highly efficient, and easy-to-manipulate system. CRISPR/Cas9 tools generate double-strand breaks (DSBs) in a specific region of the genome, which are subsequently repaired by either of the 2 cellular mechanisms: the error-prone nonhomologous end-joining (NHEJ) that introduces insertions or deletions (indels) in the repaired region, or homology-directed repair (HDR), which uses a template for exact repair. NHEJ is favored in adult cells, while HDR is more frequent during development and growth. The therapeutic strategies developed to date took advantage of both repair mechanisms to edit genes involved in hepatic genetic disorders. 157
Because of the low efficiency of HDR, several NHEJ-based gene knockout (KO) strategies have been designed to treat cell autonomous metabolic disorders of hepatic origin. These therapies are based on permanently knocking out a gene directly involved in the disease or indirectly in the generation of symptoms. One of the examples is the inhibition of PCSK9 gene, which encodes a hepatic protease involved in cholesterol regulation, as a treatment for FH. 159 –161 The gene editing systems were delivered using 2 viral vectors: adenovirus for Streptococcus pyogenes Cas9 (SpCas9) 159 and AAV for Staphylococcus aureus Cas9 (SaCas9) 160 in mice and AAV for meganucleases in macaques. 161 Gene KO strategies have also been used as SRT in metabolic disorders involving the accumulation of a toxic metabolite, such as HT1 162 and PH1. 163 In HT1, a specific CRISPR/Cas9 system was delivered by hydrodynamic tail-vein injection, and in PH1, by administration of AAV vector. In both cases, enzymes acting upstream in the pathway were inhibited successfully, which resulted in the reduction of the toxic metabolite and amelioration of disease phenotype.
Other therapeutic strategies involve the HDR mechanism for the correction of a mutated gene or for insertion of a functional gene copy in a safe locus. This strategy has proven to be efficient in diseases such as HT1, in which the corrected hepatocytes have growth advantage over the faulty ones and repopulate the liver. 164 Therapeutic efficacy and complete repopulation of corrected hepatocytes were observed after treating a mouse model of HT1, with the Cas9 protein, sgRNA, and donor template (Fah), which were delivered by hydrodynamic injection of plasmids or a combination of Cas9 mRNA nanoparticles and AAV vectors expressing the sgRNA and the donor. 165,166 In the absence of a selective advantage, the low efficiency of HDR may result in inefficient gene correction and thus fail to reduce disease symptoms. However, for diseases caused by the deficiency of a secreted protein, such as hemophilia, the insertion of a promoterless construct containing the cDNA of the therapeutic gene in the highly expressed albumin locus resulted in the production of therapeutic levels of the secreted enzyme. 167,168 Interestingly, insertion under the control of the endogenous promoter turned to be less potent. 169 Finally, since HDR is favored during development and growth, Yang et al. treated neonate and adult mice with OTC with 2 AAV vectors expressing the CRISPR/Cas9 compounds necessary to correct the mutation. 170 The percentage of corrected hepatocytes in the absence of a growth advantage was higher in mice treated as neonates than in mice treated as adults and resulted in a greater therapeutic effect.
Patients affected by alpha-1-antitrypsin deficiency could potentially benefit from NHEJ and HDR events in the mutated SERPINA1 allele. Delivery of a CRISPR/Cas9 (targeting the SERPINA1 allele) and the respective donor template to correct it using a combination of 2 AAVs could correct the disease in a mouse model. 171,172
The engineering of gene editing nucleases and delivery methods has greatly increased the diversity of gene editing strategies for in vivo treatment of hepatic genetic disorders. Another example are base editors that comprised a “dead” Cas9 (without nuclease activity) fused to a cytidine deaminase or an adenosine deaminase, which is able to generate mismatched intermediates by interconverting C-G or T-A, respectively. 173 This class of gene editors has already been tested in vivo to correct disease causing mutation, such as in PKU. Editing frequencies were similar to the ones previously observed for NHEJ. 174
Cas9 has also been fused to transcriptional activators or repressors to modify gene expression without permanently editing the genomic target region. A Cas9 fused to a Krüppel-associated box (KRAB) transcriptional repressor was used for treatment of FH in vivo by inhibiting PCSK9 expression. 175 The reduction of circulating PCSK9 was comparable with that obtained with NHEJ, although a transient, treatment-associated transaminase elevation was observed.
The most common delivery method of gene editing nucleases has been the use of viral vectors, in particular AAV vectors. However, unlike gene therapy strategies using the delivery of an entire functional gene, most of the therapeutic strategies described above only require a transient expression of the editing system. In fact, to date, no studies on the long-term safety of continuously expressed nucleases are available. Another possible impediment to the use of especially AAV vectors is the increase in size of engineered nucleases, which can compromise the packaging of the constructs expressing the nuclease. Therefore, efforts are also being directed at the development of a nonviral delivery method to increase the safety and efficiency of therapy. 176 –179
Nuclease-free strategies
AAV has been shown to naturally induce a certain rate of homologous recombination (HR), and Barzel et al. optimized a promoterless nuclease-free approach (AAV-HR) for genome editing by exploiting the induction of HR upon AAV transduction. 180 They deliver the hFIX gene flanked by murine albumin homology arms to allow integration of the hFIX gene into the albumin locus in a nondisruptive way. Thus, the chimeric mRNA produced translated to both the endogenous albumin protein and the introduced hFIX. This technology has since been successfully tested for the treatment of metabolic liver diseases such as CN and ZZ A1AT deficiency (the most severe form of A1AT deficiency). 181,182 Nevertheless, the percentage of edited cells upon AAV-HR-mediated gene transfer to the liver was low and with only 0.4–1% of total hepatocytes corrected it is insufficient for the majority of IMLDs.
On the contrary, a group of Clade F AAV isolated from human CD34+ hematopoietic stem cells (AAVHSCs) have been recently described to very efficiently mediate HR in the absence of nuclease expression. 183 AAVHSC-mediated gene insertion to correct a disease phenotype was successfully assessed in Pahenu2 mice, the standard PKU murine model. 184
Gene Therapy Clinical Trials for IMLDs
The frontrunner for the application of liver-directed gene therapy is without doubt hemophilia. It is an X-linked bleeding disorder that results from deficiency of factor VIII (FVIII) or FIX due to mutations in the F8 or F9 genes, respectively. 185,186 Because residual factor levels directly correlate with the severity of the phenotype and patients with 6–30% present with mild disease only (bleeding as a consequence of more severe trauma or invasive procedures), relatively low levels of transduced hepatocytes are sufficient for disease amelioration and correction. 185,186 Decades of effort have accumulated to 13 clinical trials that have been completed or are still ongoing and that have shown great success. However, the number of clinical trials targeting liver disorders is very limited (Table 2). The first human experience in liver metabolic disease was the treatment of a patient with FH who received hepatocytes transfected with a retroviral vector expressing LDL-R, resulting in a moderate reduction of 17% of the initial cholesterol values. 187
Gene therapy clinical trials for IMLDs
AIP, acute intermittent porphyria; CN, Crigler–Najjar syndrome; HoFH, homozygous familial hypercholesterolemia; MMA, methylmalonic acidemia; OTC, ornithine transcarbamylase; PH1, primary hyperoxaluria type 1.
The first clinical trial for a metabolic inherited hepatic disease was performed in 1999 to treat OTC deficiency. A first-generation adenovirus was used with devastating consequences for the field due to an uncontrolled inflammatory reaction to the vector in 1 patient. 188 Only years later, another trial was performed, the first using an AAV vector for an IMLD. In 2014, D'Avola et al. used an AAV5 expressing human porphobilinogen deaminase (PBGD) to treat patients with very severe AIP. 189 This therapy proved to be safe and produced partial symptomatic relief, but failed to reduce porphyrin precursor levels. This latter result was likely due to insufficient liver transduction with the vector doses used in that study. However, it is noteworthy that a significant improvement on was observed in all the patients and 2 of the 8 treated patients have stopped standard-of-care treatment for more than 3 years now (personal communication). The use of more efficient vectors for the transduction of the human liver deserves consideration for the treatment of this disease.
To date, several AAV-based clinical trials for the treatment of IMLDs are running. Two phase 1/2, open-label, dose-escalation studies are ongoing in patients with severe CN syndrome who require phototherapy. In these studies, the safety and efficacy of an i.v injection of an AAV8 vector carrying the UGT1A1 gene are evaluated. The primary endpoints are the evaluation of safety and the reduction of bilirubin levels after weaning off phototherapy. While one study is run in Europe sponsored by Genethon, the other is performed in the United States by Audentes.
Another ongoing clinical trial is evaluating the safety and preliminary efficacy of a single i.v. infusion of an AAV8 vector expressing OTC in adults with symptomatic late-onset OTC deficiency. So far, no serious adverse events have been reported. Two patients had mild, asymptomatic ALT increases that were resolved with corticosteroids. Importantly, in 2 patients (1 in cohort 1 and the other in cohort 2), the ureagenesis rate normalized, which allowed them to discontinue all ammonia scavenger medications and reduce dietary protein restrictions. 190
A phase 1/2 trial assessing AAV8-mediated liver-directed gene therapy in adults with GSDIa is also under way to evaluate the safety, tolerability, and efficacy of the vector. Twelve weeks postvector administration, all patients demonstrated a positive biologic response, reflected by an increased time to a hypoglycemic event during a controlled fasting challenge. In 2 patients, this improvement was clinically meaningful and all patients have been able to decrease their baseline total daily cornstarch use. 191
Very exciting results are coming from clinical trials assessing the therapeutic efficacy of liver-targeting siRNA molecules. Sardh et al. reported the results of a phase 1 trial in AIP patients treated with subcutaneous injections of givosiran, an anti-ALAS1 siRNA linked to GalNAc for hepatocyte targeting via the ASGPR. Givosiran was able to induce a potent reduction in urinary aminolevulinic acid and PBG, and a significant reduction in the attack rate. In addition, the number of hemin doses required by the patients was reduced by 64%. However, some adverse events were abdominal pain, nausea, and diarrhea. Phase 3 is ongoing.
RNAi targeting liver HAO1 mRNA (lumasiran) reduces the expression of glycolate oxidase, which results in a reduction of hepatic oxalate. Multiple doses of lumasiran demonstrated an acceptable safety profile and a mean maximal reduction in urinary oxalate of 75% (range: 43–87%) relative to baseline. Phase 3 is also ongoing.
A First-in-Human phase 1/2 study has been initiated to evaluate the therapeutic efficacy of LNP-mRNA in patients between 1 and 18 years with MMA. The study is designed to determine the safety, pharmacokinetics, and pharmacodynamics of different doses of LNP-mRNA in patients as part of the dose escalation phase.
Current Limitations and Possible Future Directions
The majority of metabolic diseases are cell autonomous, meaning that each hepatocyte is a diseased cell and needs to be corrected. Thus, the successful application of gene therapy for metabolic liver disorders requires the transduction of a high percentage of hepatocytes. An exception are only those diseases in which therapeutic transgene expression confers a positive selective advantage and the correction of a small percentage of cells is enough to repopulate the entire liver, such as in HT1 or MMA. However, integrative vector or gene editing strategies are required to achieve stable expression. 132,133 For other IMLDs such as CN or Wilson, it has been shown that the correction of a relatively low percentage results in a significant clinical benefit. This implies that the uncorrected cells can benefit from the activity of the corrected cells. 82,97 However, in metabolic disorders such as hyperoxaluria or AIP, more than 50% of the cells need to be corrected. A high vector dose can be used to reach a high enough number of cells, but clinical studies have indicated that administration of high doses of AAV vectors may induce a higher risk of treatment-related adverse effects, including cytotoxic T cell responses against transduced hepatocytes. 192
The liver is a growing organ and possesses the capacity of regeneration after an insult—a characteristic that affects the duration of the therapeutic effect of episomal vectors such as AAV vectors. This is of particular importance for the treatment of metabolic disease in infancy or childhood. The use of integrative vectors or gene editing strategies can solve this problem; however, careful assessment of efficacy and safety issues requires full attention. An alternative to permanent genetic modification offers the readministration of the therapeutic vector. Nonviral vectors can be repeatedly administered, but as their effect lasts only 1–2 months, lifelong repeat treatments are a necessity. On the contrary, the readministration of viral vectors poses another problem that is at present excluding many patients from therapy altogether: the presence of neutralizing antibodies (NAbs) against the vector. Therefore, the development of strategies that prevent the development of a humoral immune response after each administration or the elimination of pre-existing NAbs is urgently required.
Strategies Allowing Readministration of the Therapeutic Vector
The development of vector-specific antibodies is undoubtedly one of the biggest challenges that need to be overcome when utilizing virus-based vectors for gene therapy. The prevalence of pre-existing NAbs in humans due to natural infection is a major concern for systemic delivery, especially of AAV. 192 Furthermore, even if a patient does not present with pre-existing NAbs, readministration of the same vector serotype is not possible as NAbs are readily generated after the initial dose. Developing strategies to overcome this issue is the focus of many research groups and has been the subject of many excellent reviews 192 –195 and is briefly summarized in the following.
From a vector perspective, a very simple strategy is the switching of existing AAV serotypes. 98,196 However, as this requires the codevelopment and production of more than 1 therapeutic vector, this approach is not feasible from an economic point of view. The development of new serotypes that transduce their target tissue more efficiently while at the same time are less immunogenic is a very active field. Different strategies, including capsid engineering, directed mutagenesis, or shuffling of capsid libraries, have been used, 63,197 –199 and while many new serotypes have been produced, true advance in this area is often hampered by (partial) loss of tropism.
Chemical modification of the AAV surface, such as PEG-coating, 200 can be used to avoid detection by the immune system; however, excessive coating may also result in obscuring interaction with receptors required for infection.
Administration of therapeutic vector together with an excess of empty vector causes a stealth effect, which allows higher transduction rates. 201 On the downside, the administration of high amounts of vectors can also be counterproductive as it bears an increased risk of inducing T cell immunity and can also lead to even higher NAb titers.
From a patient perspective, several strategies have been investigated but although most proved successful, they all come with certain downsides for the patient. A recurrent and straightforward option already successfully applied in both preclinical and clinical trials is immunosuppressive treatment. Immunosuppression can involve selective depletion of immune populations involved in antibody production (rituximab for CD20+ or anti-CD4 for T helper cells) 202 or a more general approach: by blocking proteasome activity, bortezomib not only prevents generation of antibodies but also antigen presentation. 203,204 The most comprehensive approach so far has been rapamycin encapsulated in nanoparticles, which inhibited the formation of NAbs against an AAV vector in mice and NHP. 205 Despite being promising and attractive alternatives for cases in need of repeated treatment, such as pediatric patients, all carry the associated risk of leaving the patient immunocompromised for a certain amount of time and thus exposed to infection. Depending on the pathology of the disease, this is a downside to be carefully evaluated.
Other more invasive treatments aimed at increasing the transduction efficacy by evading an antibody response are plasmapheresis or immunoadsorption (physical removal of antibodies) and choosing a route of administration that delivers the therapeutic vector directly to the target tissue. 206,207 Both have been successfully applied in animal models and plasmapheresis/immunoadsorption is a validated treatment for autoimmune diseases caused by high autoantibody levels (e.g., Guillain–Barré or myasthenia gravis.
Conclusion
Great advances have been made in the field of gene therapy over the last decade, from our understanding of viral vectors to the development of new vectors and gene-editing tools. The efforts have culminated in the approval of several therapies and promising ongoing clinical trials for many more. The success of the liver-directed AAV-based therapies for hemophilia has opened the doors for treatment of IMLD even wider. Although many challenges remain, continuous effort and progress will hopefully allow the inclusion of further, presently untreatable genetic disorders in the near future.
Footnotes
Author Disclosure
G.G.-A. is founder and CSO of Vivet Therapeutics SAS.
Funding Information
G.G.-A. received funding from Secretaria de Estado de Investigación, Desarrollo e Innovación, Ministerio de Economia y Competitividad and Ministerio de Ciencia y tecnología (SAF2015-70028-R and RTI2018-101936-B-I00).
