Gene targeting as a therapeutic avenue in diseases mediated by the complement alternative pathway

Summary

The complement alternative pathway (AP) is implicated in numerous diseases affecting many organs, ranging from the rare hematological disease paroxysmal nocturnal hemoglobinuria (PNH), to the common blinding disease age‐related macular degeneration (AMD). Critically, the AP amplifies any activating trigger driving a downstream inflammatory response; thus, components of the pathway have become targets for drugs of varying modality. Recent validation from clinical trials using drug modalities such as inhibitory antibodies has paved the path for gene targeting of the AP or downstream effectors. Gene targeting in the complement field currently focuses on supplementation or suppression of complement regulators in AMD and PNH, largely because the eye and liver are highly amenable to drug delivery through local (eye) or systemic (liver) routes. Targeting the liver could facilitate treatment of numerous diseases as this organ generates most of the systemic complement pool. This review explains key concepts of RNA and DNA targeting and discusses assets in clinical development for the treatment of diseases driven by the alternative pathway, including the RNA‐targeting therapeutics ALN‐CC5, ARO‐C3, and IONIS‐FB‐LRX, and the gene therapies GT005 and HMR59. These therapies are but the spearhead of potential drug candidates that might revolutionize the field in coming years.

OUTLINE

This review gives the reader a broad overview of the gene‐targeting field including RNA targeting via RNA interference (RNAi) or antisense oligonucleotides (ASO), DNA targeting using clustered regularly interspaced short palindromic repeats (CRISPR) technology, and gene targeting via adeno‐associated virus vectors (AAV). A brief overview for each technology is given, and the rationales for using gene targeting in alternative pathway‐mediated diseases are highlighted. The review discusses current relevant clinical and preclinical applications and critically assesses challenges of gene targeting. To avoid overlap with other reviews elsewhere in this volume, only a brief introduction to complement or animal models in complement research is given and the reader will be referred to more substantial reviews.

1. INTRODUCTION TO GENE TARGETING

To alter the expression of genes, biology offers several mechanisms that are utilized by researchers not only to study the role of genes in health or disease, but also for therapeutic development. Levels of a specific gene can be increased (usually achieved by gene therapy), silenced or knocked down (RNAi or ASO), or can be permanently knocked out. In the latter case, this is mostly achieved using CRISPR‐Cas9, although there are other knock‐out systems including zinc finger nucleases, transcription activator‐like effector nucleases [TALEN]‐based and microRNAs. The complement system, in particular the alternative pathway (AP), has been identified as a target for these new therapeutics. Progress and clinical successes in the treatment of rare or monogenetic diseases will pave the way for further development of gene‐targeted therapies for the treatment of complement‐mediated diseases.

1.1. RNA targeting

The two main approaches to target RNA are RNAi and ASO, although there are also emerging technologies that utilize CRISPR‐Cas13 or a “dead” Cas9 for RNA‐targeting approaches. ¹ RNA‐targeting therapies are applied to inhibit the translation of disease‐associated genes transiently and selectively.

RNAi is a natural mechanism employed by eukaryotic cells to directly control gene expression and activity; in recent years, understanding of the mode of action has facilitated its use as a therapeutic avenue in drug development. RNAi requires a double‐stranded (ds) RNA molecule of 19–25 nucleotides which is sufficient for stable duplex formation and recognition by the RNA‐induced silencing complex (RISC) (described below) but short enough to avoid production of interferons in response to presence of dsRNA. ² These so‐called small interfering RNAs, or siRNAs, consist of an antisense (the “guide”) strand and a sense strand with 3′ dinucleotide overhangs. Alternatively, RNAi can also be delivered by plasmids and expressed as short hairpin RNAs (shRNAs) connected by a stem loop that is cleaved by the endoribonuclease Dicer to yield siRNA structure. ³ The siRNAs must be delivered to the cytoplasm of the target cell type where they are taken up by Argonaute (AGO) protein as part of RISC. The duplex strands are dissociated, with the antisense strand remaining bound to function as a guide to align the complex to target messenger RNA sequences and promote their subsequent degradation ⁴ (Figure 1a). The sense strand merely functions as a “drug delivery device”. Endogenous RNAi requires dsRNAs to be cleaved into 21–23 nucleotide long fragments by Dicer protease prior to binding to RISC, while RNAi therapies deliver the double‐stranded siRNAs directly into cells.

FIGURE 1.

RNA targeting via RNA interference (RNAi) or antisense oligonucleotides (ASO). (a) In RNAi, a double‐stranded small interfering (si) RNA is delivered to the target cell where it is taken up by the RNA‐induced silencing complex (RISC). The RNA strands are dissociated, and only the antisense strand remains bound to RISC and functions as a guide to align the complex to target sequences and promote subsequent degradation. (b) Gene knockdown via antisense oligonucleotide works in a similar manner to RNAi; however, a gene‐specific single‐stranded RNA is directly delivered to the target cells where it binds to its target sequence and promotes RNA degradation via RNaseH1. Figure was created with BioRender.com

ASOs are single‐stranded nucleotides of 19–25 base pairs in length that bind to and induce degradation of their target sequence. Apart from silencing, translation can also be blocked and/or mis‐splicing can be prevented by use of morpholinos, a subclass of ASOs. ⁵ As with RNAi, gene expression is knocked down; however, ASOs can also enhance target translation by manipulation of alternative splicing. ⁶ Once bound to the target RNA, the RNA/ASO duplexes are recognized by RNase H1, an endogenous endonuclease that cleaves the phosphodiester bonds of RNA in an RNA/DNA heterocomplex (Figure 1b). Subsequently, RNA fragments are further degraded by canonical RNA degradation pathways involving a 5′ → 3′ exoribonuclease, XRN1, also called PACMAN, and the exosome complex. ⁷ One key difference to RNAi is that ASOs must find their target sequence alone prior to binding of the enzyme, while siRNAs are bound by AGO and only then does the enzyme/oligonucleotide complex bind to the target RNA.

Contrary to genomic editing, RNA‐targeted therapies are transient in nature, with expression levels returning to baseline after several days or cell divisions, a phenomenon likely due to dilution rather than degradation. However, expression of siRNAs from a plasmid or chemical modification to increase stability or delivery to non‐dividing cells has been reported to result in a sustained silencing period over several months (Conference abstract: Development of IONIS‐FB‐LRx to Treat Geographic Atrophy Associated with AMD. Presented at the 2020 ARVO Annual Meeting). ⁸ , ⁹ , ¹⁰ Currently, there are three RNA therapies in clinical trials that target complement: Cemdisiran and ARO‐C3, two RNAi therapies targeting C5 and C3, respectively, and IONIS‐FB‐LRX, an ASO targeting Factor B (FB). All three therapies are described in detail in the sections below.

1.2. Gene therapy

The goal of gene therapy is to introduce genetic material into specific cells of a patient, potentially providing a durable and one‐time treatment dependent on the type of cell that is transduced. A one‐time treatment requires transduction of a cell type that no longer or only slowly divides, such as hepatocytes, retinal cells, myocytes, adipocytes and neurons, or a vector that integrates the transgene within the host cell genome like lentiviral vectors. Hereditary diseases caused by gene mutations are particularly amenable to gene therapy as the treatment targets the underlying genetic cause. Key to effective treatment is efficient gene delivery to the target tissue/cells, and this can be achieved by using delivery vehicles such as non‐viral or viral vectors. ¹¹ , ¹² Furthermore, vector‐based gene therapy can be divided into two categories, ex vivo and in vivo gene therapies. Clinical approaches to ex vivo gene therapies, often associated with a subclass of cell therapy, require genetic modification of autologous cells isolated from the patient to elicit a therapeutic effect. In contrast, in vivo strategies involve either the direct infusion of gene therapies into the patient's blood stream or direct injection into target organs. ¹³ , ¹⁴

To date, viral vectors have been derived from engineered viruses including adenovirus (Ad), adeno‐associated virus (AAV), herpes simplex virus, and lentivirus, and have demonstrated their potential to efficiently deliver gene therapies in the clinic. ¹⁵ , ¹⁶ , ¹⁷ Recombinant AAV (rAAV) vectors are the dominant vector type for in vivo gene therapies and are under intensive investigation in numerous preclinical or clinical studies owing to their superior efficacy and safety profiles; several AAV‐based vectors recently have received a marketing authorization (LUXTURNA™ an AAV2 vector containing human RPE65 cDNA for the treatment of Leber congenital amaurosis, ZOLGENSMA™ an AAV9 vector containing the human SMN1 cDNA for the treatment of spinal muscular atrophy and, recently, ROCTAVIAN™ an AAV5 vector containing the human Factor VIII for the treatment of hemophilia A [press release: BioMarin Announces Record Revenues in Second Quarter 2022; Increases Full‐year 2022 Top and Bottom‐line Guidance. Published on August 24, 2022, on www.biomarin.com]). ¹⁸ , ¹⁹ Currently, development of complement gene therapies has been primarily based on the use of AAV vectors. Wildtype AAV (wtAAV) contain a single‐stranded DNA genome of approximately 4.7 kilobases (kb) in length, consisting of Rep and Cap genes, required for successful amplification and packaging of AAV virions, flanked by two inverted terminal repeats (ITRs) (Figure 2a). Vector engineering has demonstrated that a recombinant DNA sequence can be inserted between the ITRs in an AAV vector genome and encapsulated into pseudovirus by providing Rep and Cap genes, as well as adenovirus helper elements in trans. ²⁰ To confer transgene expression, the recombinant (r) AAV genome is typically designed to include four main components as follows: (1) a constitutive or tissue/cell‐specific promoter sequence that dictates the overall strength and cell‐specificity of expression; (2) protein‐coding transgenes for gene replacement or addition therapies, or siRNA expression cassettes for gene silencing and editing therapies; (3) polyadenylation signal sequence, critical for nuclear export and mRNA stability and (4) cis‐elements essential for expression, such as linker sequences, signal peptides (Figure 2a). ²¹ Moreover, selection of specific Cap genes derived from different AAV serotypes allows for generation of virions with desired tissue tropisms. ²² , ²³ Transduction of a target cell entails uptake of the rAAV vector by receptor‐mediated endocytosis and subsequent escape from the endosome to enter the nucleus where the DNA cargo is released by uncoating of the capsid. The delivered transgene is transcribed into mRNA which is secreted into the cytoplasm where protein is translated (Figure 2b). Like wtAAVs, rAAV are inherently non‐pathogenic; due to the absence of helper viral elements, the rAAV vector genome largely remains non‐integrative in host cells. More importantly, rAAV vectors can efficiently transduce both dividing and non‐dividing cells and, in the latter, confer prolonged expression by forming circular concatemers that persist as episomes in the nucleus of transduced cells. ²⁴

FIGURE 2.

Gene targeting via adeno‐associated virus (AAV) gene therapy. (a) The recombinant (r) AAV genome is packaged in an icosahedral capsid and comprises a promoter, the protein‐encoding sequence and a polyadenylation site (as well as other cis‐regulatory elements required for expression) with two inverted terminal repeats at either end required for efficient packaging of the genome into its capsid. The genome elements are modular and are often optimized specifically for each gene therapy, although almost all have the inverted terminal repeats (ITRs) of AAV2. The route of administration coupled with the capsid serotype determines tropism of the viral vector to specific cell types to reach the target cells. (b) On binding to its cell surface receptor, the AAV vector is taken up by receptor‐mediated endocytosis and resides in the endosomes until it escapes and enters the nucleus. The viral vector uncoats and releases it DNA cargo into the nucleus where it exists as extrachromosomal episomes. The single‐stranded AAV genome is converted to double‐stranded DNA via second strand synthesis. The rAAV promoter drives transgene expression which then follows default transcription and translation pathways, that is, export to the cytoplasm, translation into protein, post‐translational modifications and secreted in the presence of a signal peptide. Figure was created with BioRender.com

Collectively, these attributes have made the rAAV vector an ideal in vivo gene delivery tool for different disease applications. ²⁵ Distribution of AAV‐mediated gene therapy in clinical settings has indicated that over 70% of trials are based on gene replacement, and over 20% are gene addition. ²⁶ Only a handful of trials are based on AAV‐mediated gene silencing or gene editing, indicating their infancy in human application. Two complement gene therapies are currently in clinical development for the treatment of age‐related macular degeneration (AMD), discussed below.

1.3. DNA editing

Gene therapy, which can either be directly therapeutic or can inhibit a therapeutic target, has shown great potential for autosomal recessive disorders or diseases that require augmentation of gene products. However, it is less useful in autosomal dominant disorders where the mutated gene should either be knocked out, silenced, or replaced by a healthy copy. For treatment of the latter, gene and base editing can permanently introduce changes to the DNA; introduction of stop codons, frameshift mutations, or splice site mutations can prevent expression of the functional gene product, and activation or silencing of regulatory elements can alter gene transcription. Gene editing also offers the opportunity to replace genetic risk variants with protective variants by altering one or more amino acids in the protein sequence. CRISPR/Cas is a naturally occurring defence mechanism employed by bacteria to defend against bacteriophages that has been adopted by scientists and drug developers as a genome editing tool since 2012. ²⁷ Although still in early stage, a small number of CRISPR treatments are currently in clinical development, including treatment for sickle cell disease, beta‐thalassemia, and Leber congenital amaurosis (Press release: Editas Medicine Announces Positive Initial Clinical Data From Ongoing Phase 1/2 BRILLIANCE Clinical Trial Of EDIT‐101 For LCA10. Published September 29, 2021, on www.editasmedicine.com). ²⁸

In the traditional approach, there are two components to the system, Cas9 (a DNA endonuclease) and a guide RNA, which can be encoded in one plasmid or several (Figure 3a). The guide RNA further consists of two parts, the CRISPR RNA (crRNA), an 18–20 bp sequence that binds immediately upstream of so‐called protospacer adjacent motifs (PAM) to the target DNA, and the transactivating CRISPR RNA (tracrRNA), a long stretch of loops that enable Cas9 to bind to the specific locus. ²⁹ The crRNA and tracrRNA can be encoded within one RNA stretch, resulting in a so‐called single guide RNA (sgRNA), or are separated into two RNA molecules. If the DNA sequences of the target match the sequence of the guide RNA and the correct PAM sequence is present, the two nuclease domains of Cas9 cleave the DNA backbone resulting in a double‐strand break ²⁷ (Figure 3b). The cell repairs this break by non‐homologous end joining mechanisms (NHEJ) or by homology‐directed repair (HDR). ³⁰ , ³¹ The first is an error prone mechanism that results in random gene disruption by insertion or deletion at the target region. This approach can truncate the transcribed mRNA sequence which, without the subsequent polyA tail, results in nonsense‐mediated decay of the transcript, effectively causing a knockout. Usually, the guide RNA is targeted to the start of a gene to ensure the gene is knocked out in its entirety. By providing a correct template, CRISPR can also introduce corrections of point mutations by HDR, albeit by a highly inefficient process, especially in non‐dividing cells. ³² , ³³ To ensure successful delivery into the nucleus, CRISPR components are often encoded in an AAV backbone and target cells are transduced with the AAV vector. However, the components can also be delivered using single or multiple plasmids using traditional transfection methods or nanoparticles (Figure 3a). Cas9 “nickases” are mutated in the nuclease domain and only cut a single strand. If two nickases are paired together (one directed to each strand), a double‐strand break with long overhangs on each of the cleaved ends is produced instead of blunt ends. This results in a much greater control over gene integration and therefore reduced off‐target effects. ³⁴ In recent years, base editing has emerged as a new CRISPR technology to precisely alter single nucleotides without the introduction of double‐strand breaks into the DNA; this approach utilizes inactivated Cas9 that cannot cleave DNA to direct deaminases to specific genomic locations where they can chemically modify bases from purines to pyrimidines (and vice versa). ³⁵ , ³⁶

FIGURE 3.

Gene targeting via CRISPR. (a) Components of the CRISPR system, Cas9 and guide RNAs, can either be delivered via a single plasmid or on separate plasmids. (b) The guide RNA binds upstream of so‐called protospacer adjacent motif (PAM) sequences which are recognized by Cas9. If the guide RNA matches the sequence, then Cas9 unwinds the DNA and introduces a double‐strand break into the DNA backbone. These breaks can either be repaired via non‐homologous end joining, resulting in a gene knockout if the break is at the start of the gene, or via homology‐directed repair if a donor DNA is provided to replace the original DNA. Abbreviations: sgRNA, single guide RNA; PAM, protospacer adjacent motifs. Figure was created with BioRender.com

Application of CRISPR technology to the complement field is still in early development, although reports have been published on a human CD46, CD55, and CD59 knockout HAP1 cell line, a CD46 knockout hTERT cell line, and a C3‐deficient pig generated using CRISPR/Cas9. ³⁷ , ³⁸ , ³⁹ In June 2021, Apellis Pharmaceuticals and Beam Therapeutics announced a partnership to develop up to six gene editing programs targeting C3 and other complement proteins for the treatment of diseases of the eye, liver, and brain (Press release: Apellis and Beam Therapeutics Enter Exclusive Research Collaboration to Apply Base Editing to Discover Novel Therapies for Complement‐Driven Diseases. Published June 30, 2021, on www.apellis.com). Application of CRISPR technology to complement therapeutics could dramatically change the landscape of complement‐mediated disorders in years to come.

2. RATIONALE FOR GENE TARGETING AS A THERAPEUTIC STRATEGY IN COMPLEMENT‐DRIVEN DISEASES

Complement is a key arm of innate immunity, providing a first line of defence against infection, opsonizing pathogens, apoptotic cells, and debris, enabling their removal and in the case of some target cells, causing cytolysis and death. ⁴⁰ Complement activation triggers production of inflammatory mediators, such as the activation fragment C5a which promotes cell infiltration, phagocytosis, and inflammatory reactions such as the oxidative burst. ⁴¹ A limited level of complement opsonization ensures that dying cells and debris are disposed of in a non‐inflammatory manner, and that immune complexes are transported to the reticuloendothelial system for clearance. Non‐canonical roles of complement are also described, including the long‐known augmentation of humoral immunity by crosstalk to the adaptive immune system, and the more recently described phenomenon of intracellular complement activation that serves multiple roles including interplay with T cell immunity and maintenance of metabolic homeostasis. ⁴²

The vascular complement system is a protein‐based system comprising of circulating activating proteins and a multitude of inhibiting or controlling proteins that exist either in a soluble form or attached to membranes where they directly modulate complement activation. These proteins work in concert to ensure that inadvertent deposition of activation fragments on self‐cells does not progress to cell damage or inflammation. Key activating components include those of the amplification loop that transform a small trigger into a large downstream effect. These proteins include the activated form of C3, C3b, that binds factor B (FB) to form a bimolecular complex, C3bB. FB complexed with C3b is in turn activated when it is cleaved by factor D (FD) into Ba and Bb, the latter remains bound to C3b within the C3 convertase (C3bBb). The proteolysis of C3 by C3bBb generates many more molecules of C3b and C3a, thus amplifying the complement system. Nascent C3b can bind covalently to surfaces, triggering formation of a C5‐cleaving enzyme and cleavage of C5 to C5a and C5b. This step marks the start of the terminal pathway and triggers downstream inflammation through the actions of the chemotactic and anaphylactic fragment, C5a, and the proinflammatory membrane attack complex (MAC): a lytic pore that is triggered by association of C5b67 with membranes. MAC can cause lysis of simple cells, such as erythrocytes, but rarely lyses nucleated cells where instead it causes cell activation and inflammation through multiple signaling pathways and via the NLRP3 inflammasome. ⁴³ , ⁴⁴

Proteins within the amplification loop responsible for complement amplification and thus C5 activation have become targets for therapeutic treatment of complement‐mediated disease. ⁴⁵ In normal circumstances, cell‐associated regulatory proteins, including CD35, CD55, CD46, and CD59, and soluble regulators including factor H (FH) and factor I (FI), control the amplification loop and prevent inappropriate activation. However, in many settings, complement control is insufficient and damaging inflammation ensues that in some circumstances may result in disease. Diseases associated with complement dysregulation and the mechanisms underlying pathology are numerous and are described in detail in other reviews. ⁴⁶ , ⁴⁷ In brief, mechanisms might include autoantibodies that drive activation on self‐tissues, for example in myasthenia gravis (MG) or neuromyelitis optica (NMO), or genetic mutations in complement proteins that interfere with the normal regulatory process. For example, mutations that result in haploinsufficiency of complement regulators or that reside in protein domains critical for regulatory function have been identified to cause loss‐of‐function in regulators or gain‐of‐function in activating proteins. ⁴⁸

The association of complement with many diseases has driven the development of anti‐complement drugs. Early anti‐complement therapeutics were biologics, including the well‐described molecule, eculizumab, a C5‐blocking antibody used in diseases where MAC or C5a caused damage and/or inflammation. ⁴⁹ Small molecules have more recently emerged onto the therapeutic landscape, including the C5a‐receptor inhibitor, avacopan, and the FB‐blocker, iptacopan. ⁵⁰ , ⁵¹ Many recent reviews describe the plethora of anti‐complement therapeutics in developement, ⁵² , ⁵³ , ⁵⁴ although development over many decades has resulted in a small number of drug registrations due to the difficulties associated with drugging the complement system. ⁵⁵

Some of the challenges associated with targeting the complement system include the high concentration and rapid turnover of many of the key players in the system. The component C5 is present in plasma at concentrations of approximately 50 μg/mL with a half‐life of around 60 h, leading to frequent and high‐dose administration of first‐generation drugs. The advent of recycling antibodies has enabled the transition from biweekly intravenous (i.v.) administration of 1.2 g of Alexion's anti‐C5 monoclonal antibody eculizumab/Soliris™ for adults with PNH to bimonthly i.v. infusion of 3.3 g of the recycling anti‐C5 antibody, ravulizumab/Ultomiris™. ⁵⁶ Roche are also exploring subcutaneous (s.c.) administration of the recyclable anti‐C5 antibody, crovalimab. ⁵⁷ Pegcetacoplan (EMPAVELI™), a compstatin derivative developed by Apellis and approved for PNH, comprises two 13 amino acid cyclic peptides that bind C3, a target with a concentration in plasma of between 1 and 2 mg/mL; pegcetacoplan comprises a PEG moiety to extend half‐life and is dosed at 1.08 mg twice weekly by s.c. infusion. ⁵⁸ Small molecules are dosed even more frequently; iptacopan, in clinical development for multiple indications, is dosed at 200 mg twice daily in order to block FB, which is present in plasma at approximately 150 μg/mL. Gene‐targeting techniques, such as ASO or RNAi, present an opportunity to knockdown targets that are primarily made in the liver, although of course liver‐targeting technologies cannot overcome local biosynthesis that can drive certain indications. Approaches in clinical development are described herein.

Complete blockade of the complement system to treat disease has been approached with caution due to safety reservations, largely centered on infection risk. The anti‐C5 approach was explored early in the history of anti‐complement therapeutics. Blocking C5 has the potential to abrogate the terminal pathway and prevent C5a and MAC formation; however, this approach enabled C3b opsonization which was considered an avenue to fight infection. C3‐targeting approaches, such as compstatin‐based molecules, entered clinical development at a much later date (2015); pegcetacoplan has potential to completely block the complement system beyond C4b opsonization. However, as risk has been mitigated by appropriate vaccination and drugs with blocking activity such as eculizumab and pegcetacoplan have been approved, concerns have been assuaged, increasing interest in the exploration of gene therapy which could target the complement system in different ways. This approach has potential to overcome the challenge of frequent dosing (such as drug burden/noncompliance), to achieve complete coverage of the target, and to reach effective concentration at the site of disease (such as in the eye). In some cases, local administration of a gene therapy can place the “drug factory” close to the site of pathology, enabling localized dosing. Historically, gene therapies have been developed for monogenic disease to replace a non‐expressing or mutant protein with a normal gene and is considered an achievable and potentially safe approach to treatment. With regard to the complement system, no gene therapy blocking agents are currently in clinical development, possibly due to the potentially irreversible nature of gene therapy. The approaches that are emerging in the clinic are primarily those that are focussed on supplementation of additional regulators, such as FI or CD59. Local expression in the eye has emerged as an avenue to explore gene therapy, largely driven by the unmet need for treatment of the common blinding disease AMD and the accessibility of this isolated organ within the body. These approaches are detailed in the following section.

3. COMPLEMENT‐TARGETING GENE AND RNA MODIFYING THERAPIES IN CLINICAL DEVELOPMENT

3.1. RNAi and ASO therapies currently in clinical trials

Cemdisiran: Alnylam Pharmaceuticals is developing cemdisiran, ALN‐CC5, an RNAi therapeutic that targets liver‐expressed C5 and is evaluated in the treatment of AP‐driven diseases. Although C5 is not a component of the AP, in some diseases including PNH that are driven purely by dysregulation or overactivation of the amplification loop, C5 is the downstream driver of pathogenesis. Blockade of either the AP or C5 could have therapeutic benefit in such diseases. Cemdisiran is a chemically modified double‐stranded siRNA whose 3′ end of the positive strand is conjugated to a GalNAc ligand for targeted delivery to the liver via binding to ASGPR on hepatocytes. ¹⁰ , ⁵⁹ , ⁶⁰ , ⁶¹ Cemdisiran prevents C5 biosynthesis, decreasing the amount of circulating C5, thus decreasing inflammation and subsequent complement‐mediated hemolysis that can ensue when C5 is cleaved and activated. RNAi therapies targeting C5 have promise for diseases currently treated with other C5‐targeting modalities, such as eculizumab, by reducing treatment frequency. Patients carrying the c.2654G → A mutation in C5 may also benefit from C5 gene therapy since the mutation disrupts the eculizumab‐binding site, making them refractory to treatment. ⁶²

Cemdisiran has been tested in several clinical trials including in both healthy individuals and patients (PNH, IgA nephropathy, generalized myasthenia gravis, and atypical hemolytic uremic syndrome). Administration to healthy individuals showed that cemdisiran was safe and well tolerated. Levels of C5 were decreased by up to 98%–99%, and serum hemolytic activity was reduced by up to 61% (Press release: Alnylam Reports Positive Initial Clinical Results for ALN‐CC5, an Investigational RNAi Therapeutic Targeting Complement Component C5 for the Treatment of Complement‐Mediated Diseases. Published June 12, 2015, on www.alnylam.com). The same press release also reports that within 48 h after s.c. administration, plasma levels of cemdisiran or its main active metabolite, AS(N‐2)3′‐cemdisiran, were below the lower limit of quantification due to efficient uptake into hepatocytes while minimal renal excretion was observed. Maximum reduction of circulating C5 protein was achieved between 3 and 4 weeks after dosing, and this level of reduction was maintained for 10–12 months following either a single administration or biweekly doses of 600 mg cemdisiran. The long treatment duration is likely possible due to both the metabolic stability of siRNA that enables the drug to have a long half‐life within hepatocytes and the low turnover rate of hepatocytes. Six patients with PNH (treatment naive N = 3; patients receiving eculizumab N = 3) were also treated with cemdisiran. However, although a maximum C5 knockdown of ~98% and a mean maximum classical pathway inhibition of ~94% were reached, the goal of lactate dehydrogenase levels <1.5x of upper limit of normal (to monitor intravascular hemolysis) was not achieved (Poster presentation: A Subcutaneously Administered Investigational RNAi Therapeutic (ALN‐CC5) Targeting Complement C5 for Treatment of PNH and Complement‐Mediated Diseases: Preliminary Phase 1/2 Study Results in Patients with PNH. Presented December 3–6, 2016, at the 58th ASH Annual Meeting and Exposition). Moreover, residual hemolysis was observed, highlighting a therapeutic need to further eliminate remaining C5 biosynthesis. Thus, in PNH patients who require complete inhibition of C5, cemdisiran only effectively inhibited residual C5 when co‐treated with an anti‐C5 antibody (eculizumab or its successor ravulizumab); of note, the C5 antibody could then be lowered in dose and/or frequency than the prescribed regimen to control C5. The requirement for additional C5 blockade could be due extra‐hepatic C5 produced locally, the need of complete (>99%) C5 inhibition to achieve a therapeutic effect or insufficient cemdisiran treatment or dosing. These observations have given rise to exploration of multi‐modal combination approaches (eg, RNAi and mAb), where C5 is targeted at both transcriptional and translational axes. Cemdisiran is currently being evaluated in multiple complement‐mediated indications including IgA nephropathy, PNH, and generalized myasthenia gravis as a monotherapy or in combination with other C5 inhibitors (NCT04940364, NCT04888507, NCT04811716, NCT04601844, NCT05070858, NCT05133531, NCT05131204). Recently, a positive topline result was presented from a Phase 2 study testing cemdisiran monotherapy in IgA nephropathy where a clinically meaningful reduction in proteinuria 32 wk after treatment was demonstrated (Press release: Alnylam Reports Positive Topline Results from Phase 2 Study of Investigational Cemdisiran for the Treatment of IgA Nephropathy. Published June 9, 2022, on www.alnylam.com).

ARO‐C3: Arrowhead Pharmaceuticals is developing ARO‐C3, an RNAi therapeutic designed to silence liver‐produced C3 using Arrowhead's proprietary TRiM™ platform. Encouraged by positive results of superior activity of pegcetacoplan over eculizumab in the PEGAGUS trial (NCT03500549), ARO‐C3 is currently being tested in a randomized, double‐blinded, placebo‐controlled phase 1 clinical trial in heathy volunteers (N = 24), in PNH, and in complement‐mediated renal disease (C3 glomerulopathy or IgA nephropathy) (N = 38) (NCT05083364) (Press release: Arrowhead Files for Regulatory Clearance to Begin Phase 1/2a Study of ARO‐C3 for Treatment of Complement Mediated Diseases. Published October 25, 2021, on www.arrowheadpharma.com). Part 1 of the study assesses safety, tolerability, pharmacokinetics, and pharmacodynamics of multiple ascending doses of ARO‐C3 (50, 100, 200, and 400 mg) in healthy volunteers while part 2 consists of two doses (low and high dose to be determined in part 1) of ARO‐C3 injected into adult subjects with either PNH, C3 glomerulopathy or IgA nephropathy (Webinar: ARO‐C3 KOL Webinar. Held October 26, 2021, and accessible on www.arrowheadpharma.com). No results have been reported to date.

IONIS‐FB‐LRX: Ionis Pharmaceuticals and Roche are developing IONIS‐FB‐LRX to target FB. This is a 2'‐O‐methoxyethyl (2'MOE) second‐generation ASO (20 bp) conjugated to an N‐Acetylgalactosamine (GalNAc) ligand. Upon s.c. administration, the liver‐targeted GalNAc ligand moiety specifically binds to and is taken up by the asialoglycoprotein receptor 1 (ASGPR1) expressed on hepatocytes for enhanced liver targeting. In a placebo‐controlled, masked phase 1 clinical trial in healthy volunteers (N = 54), IONIS‐FB‐LRX reduced plasma FB levels by approximately 56% and 72% after 36 days of multiple, s.c. drug administrations of 10 and 20 mg, respectively (Conference abstract: Development of IONIS‐FB‐LRx to Treat Geographic Atrophy Associated with AMD. Presented at the 2020 ARVO Annual Meeting). The higher dose of 20 mg achieved on‐target attenuation of AP activity with a 62% reduction in AH50 (measured by Wieslab AP assay) and a reduction of plasma FB and Bb levels of 72% and 73%, respectively. No meaningful classical pathway change (12% reduction in CH50, measured by hemolytic assay) was detected and no safety signals were observed, including no clinically relevant changes in blood chemistry or hematology.

Ionis Pharmaceuticals is currently in phase 2 clinical trials in geographic atrophy (GA), an archetypal AP‐driven disease with genetic associations in complement components (C3, FB, C9) and regulators (FH, FI, FH‐related proteins) and abundant evidence of complement activation in the outer retina of affected eyes. ⁶³ , ⁶⁴ Although FB protein is primarily synthesized in the liver, circulating FB passes through the ocular blood vessels within the retina and choroid. Due to the fenestrated nature of the blood vessels beneath the Bruch's membrane, it is likely that systemic FB bathes the choroidal environment and Bruch's membrane. ⁶⁵ , ⁶⁶ , ⁶⁷ A non‐human primate study showed that despite significant reduction of ocular FB protein and systemic FB mRNA following administration of FB ASO, ocular FB mRNA remained relatively stable at <1% of liver mRNA levels. ⁶⁸ These data suggest that FB is made locally in the eye beyond the blood–retinal barrier. Indeed, FB can be measured in ocular fluids of healthy individuals, ⁶⁹ raising the interesting question of whether local or systemic complement, or both, contribute to disease pathogenesis.

The outcome of the phase 2 GOLDEN trial in GA (NCT03815825) will provide critical mechanistic information regarding GA pathogenesis and may support the utility of targeting systemic FB in AP‐driven diseases, despite that an abundance of literature shows the important role for locally derived complement in disease pathogenesis, including in GA and in other organ‐centric conditions such as kidney transplant. ⁷⁰ IONIS‐FB‐LRX is also evaluated in a phase 2A trial in primary IgA nephropathy (NCT04014335). Over the coming years, the outcome of clinical trials with either local or systemically delivered gene‐targeting therapies will provide evidence to guide future drug development strategies for GA and other diseases where pathology is localized.

3.2. Gene therapies currently in clinical trial

There are two gene therapies targeting the complement pathway that are currently in clinical trials. GT005 is an AAV2 vector expressing full length, native FI, a critical controller of the amplification loop. FI acts by proteolytically cleaving the activated form of C3, C3b, to its inactive form, iC3b. ⁷¹ This degradation fragment cannot bind FB and drive the amplification loop. FI also inactivates the central component of the classical and lectin pathways, C4b, supporting its inactivation by formation of the fragments C4d and C4c. The second gene therapy in clinical development is HMR59, or AAVCAGsCD59; this is an AAV2 vector resulting in the production of a soluble (engineered) form of CD59 (sCD59), an inhibitor of MAC that marks the endpoint of the complement cascade. While CD59 controls complement downstream of the amplification loop, it has entered clinical development as a treatment for the AP‐driven ocular disease, AMD.

3.2.1. GT005

Gyroscope Therapeutics Limited started its first clinical trial with its lead candidate, GT005, in 2019. Clinical trials are evaluating the delivery of this AAV2 vector via subretinal transvitreal delivery or via Gyroscope Therapeutics' proprietary Orbit™ Subretinal Delivery System (Orbit SDS™) device, which accesses the subretinal space via suprachoroidal cannulation and does not require a vitrectomy. Once in the subretinal space, GT005 transduces the retinal pigment epithelium (RPE) and photoreceptors leading to secretion of FI at levels that can be measured in vitreous and aqueous humor (Press release: Gyroscope Therapeutics Announces Presentation of Positive Interim Phase I/II Data for Investigational Gene Therapy GT005 at Retina Society Annual Scientific Meeting. Published September 30, 2021, on www.gyroscopetx.com). GT005 is currently being tested in FOCUS (NCT03846193), an open‐label phase 1/2 clinical trial evaluating the safety and dose response of three doses of GT005 given as a single subretinal injection to patients with GA secondary to AMD. Interim results were first reported at the Angiogenesis, Exudation, and Degeneration virtual meeting 2021, and an updated data set was presented at the Retina Society's 54th Annual Scientific Meeting: intraocular overexpression of FI following subretinal GT005 administration was shown to be well tolerated and FI was elevated by 122% compared to baseline (N = 13, P = 0.002) Press release: Gyroscope Therapeutics Announces Positive Interim Data from Phase I/II FOCUS Trial of Investigational Gene Therapy GT005. Published February 12, 2021, on www.gyroscopetx.com). Furthermore, target engagement and regulation of the amplification loop were demonstrated by a 46% reduction in Ba levels in vitreous humor compared to baseline (N = 11; P = 0.001) and a 46% reduction in C3 breakdown proteins (C3b, iC3b, and C3c) compared to baseline (N = 13; P = 0.001) (see above press release, September 30, 2021). Continuous FI expression and downregulation of complement were demonstrated to have long‐lasting effects with transgene expression and target engagement evident for almost 2 years after GT005 dosing. EXPLORE (NCT04437368) and HORIZON (NCT04566445) are phase 2 trials that evaluate GT005 in a population of GA patients with CFI rare genetic variants associated with low plasma FI levels (EXPLORE) and in a broader GA population, that is, AMD associated variants in other complement genes and without rare FI variants (HORIZON).

3.2.2. HMR59

Hemera Biosciences (Janssen Pharmaceuticals Inc. acquired the rights to HMR59 in 2020) was the first company to start a phase 1 clinical trial using a complement regulator in 2017 (NCT03144999). HMR59 is delivered intravitreally (into the vitreous humor) to the eye. It results in overexpression of sCD59 and is explored for the potential treatment of AMD. Native CD59 is a glycosylphosphatidylinositol (GPI)‐anchored membrane protein that regulates and inhibits the action of the MAC on cell membranes by preventing incorporation of C9 into the C5b‐8 complex. Although early versions of recombinant soluble (eg, non‐membrane targeted) forms of sCD59 were able to inhibit MAC in vitro, they were not successful in vivo probably because of rapid clearance by the kidney and loss of activity in serum. ⁷³ These challenges could be overcome by AAV gene therapy that resulted in higher and constitutive expression of sCD59 in targeted tissues. ⁷⁴ , ⁷⁵ HMR‐1001, a phase 1 study, evaluated intravitreal injection of HMR59 in patients with GA (NCT03144999). The treatment was found to be safe, well tolerated, and no dose‐limiting toxicity was observed. While not powered to show efficacy and uncontrolled against a true placebo arm, HMR59 treatment demonstrated an overall 23% slower rate of GA progression versus a historical control arm, with 9 out of 11 (81%) eyes illustrating a slower rate of progression. No patients converted to neovascular AMD at 18‐month follow‐up; however, 3 out of 11 (27%) patients experienced mild uveitis and vitritis, requiring 6–8 weeks of anti‐inflammatory therapy and subsequent protocol institution of systemic steroids (CLINICAL TRIAL DOWNLOAD: Data on a Gene Therapy for Dry and Wet AMD A phase 1 clinical trial program is targeting both disease states. Published April 17, 2020, on www.retinalphysician.com. SUBSPECIALTY NEWS: Angiogenesis meeting highlights, elevated IOP after intravitreal injection, and more. Published on March 16, 2019, on www.retinalphysician.com). HMR‐1002 is a second phase 1 study that tested intravitreal HMR59 administration in 2 dose cohorts in wet AMD patients that received a single injection of anti‐VEGF 7 days prior to HMR59 administration (NCT03585556). Of 22 patients who had 6 months of minimum follow‐up, 18% did not require retreatment, although retreatment criteria have not been fully disclosed. Despite an initial short course of oral steroids, 13% of patients experienced inflammation and required either topical or further oral steroids. Overall, these results highlight the therapeutic promise of complement gene therapy for AMD yet indicate potential safety challenges with intravitreal delivery. It remains to be seen whether Janssen will continue to develop HMR59 for AMD; the program is no longer listed in the company's pharmaceutical pipeline and Hemera's phase 2 trials are administratively withdrawn due to licensing of the program to Janssen.

4. GENE THERAPIES AND RNAi TARGETING THE COMPLEMENT SYSTEM IN PRECLINICAL ANIMAL MODELS

Recent advances using genetic manipulation for therapy necessitate in vitro and in vivo animal models for early‐stage research and preclinical development. Preclinical models of ocular, renal, and systemic complement‐driven diseases are established as described elsewhere in this volume. ⁷⁶ The technological advances described above and current drives to optimize capsid design and delivery have resulted in a plethora of gene‐targeting approaches to treat systemic and local complement‐mediated disease. The dominant approaches are supplementation of AP regulators in order to provide more effective control, or attenuation of activating components, thereby decreasing the complement activating potential.

4.1. Gene targeting in ocular disease models

The eye provides an ideal environment to test gene therapies due to its encapsulated nature, smaller required vector volumes, limited exposure of vector to the systemic system and by being a relatively immune‐privileged site in the human body. Vectors can be applied locally, different routes of administration and capsids can be explored, and the impact of the microenvironment can be readily interrogated. Many of the cells in the ocular environment, such as RPE, ganglion cells, and photoreceptor cells, are terminally differentiated, making them ideal targets for sustained gene therapy. Several gene‐targeting approaches have been tested in preclinical models of ocular disease, particularly in models that translate to common human diseases (AMD or glaucoma) or monogenic diseases, such as Stargardt's disease, which might be particularly amenable to genetic manipulation for therapeutic benefit.

Preclinical models of AMD are challenging due to the increase in complexity of the human eye compared to the eyes of smaller mammals such as rodents. While the mouse eye does not have a typical macular region, it has a similarly photoreceptor‐dense central retina that models the human peripheral macula well. ⁷⁷ Laser‐induced choroidal neovascularization (CNV) has evolved as a benchmark for testing of complement therapeutics. This model is more representative of the wet (neovascular) form of AMD than GA, but never‐the‐less the laser‐induced damage is propagated by complement and the model can be used to evaluate anti‐complement activity in the ocular environment.

When treating disease driven by complement dysregulation, approaches that upregulate or supplement complement control proteins are a rational strategy to restore control. Alternatively, knockdown of critical activating proteins, such as FB or FD, might diminish the activating drive. The former approach was adopted by Kumar‐Singh and colleagues, who used AAV vectors to supplement the natural membrane‐associated regulators CD55 (DAF), CD46 (MCP) or a soluble version of the MAC inhibitor, CD59 (sCD59), in the ocular environment in laser‐induced CNV rodent models. ⁷⁴ , ⁷⁸ , ⁷⁹ CD55, CD46, and sCD59 were delivered via subretinal Ad or AAV vectors. Ad encoding for CD46, AdCAGCD46, resulted in a modest 24% reduction of MAC deposition in the RPE. ⁷⁸ CD55 gene therapy was explored in a similar fashion in an Ad5 construct (AdCAGCD55). Subretinal delivery of AdCAGCD55 conferred significant reduction in MAC deposition on mouse RPE compared to vector control, highlighting a potential therapeutic approach via decay acceleration. ⁷⁹ Similarly, to overexpress FI protein in the eye as a potential treatment for AMD, Dreismann et al packaged CFI cDNA expressing wildtype FI protein into an AAV2 vector and administered the viral vector subretinally into mice, resulting in transduction of RPE and photoreceptor cells and elevated levels of FI in vitreous humor. ⁸⁰ These experiments contributed to the development of GT005, a FI gene therapy currently in phase 2 clinical development by Gyroscope Therapeutics (see above).

While expression of a secreted protein leads to its widespread distribution and therapeutic potential in tissues beyond the site of expression, membrane‐associated molecules limit complement control to those cells that are transduced. Recognition of this potential limitation led to testing of engineered, secreted regulators of complement in the laser‐induced CNV model. Vectorized PRELP (Prolargin, or Proline and Arginine Rich End Leucine Rich Repeat Protein) was administered via subretinal injection via an AAV8 vector where it significantly reduced CNV area by ~60% (P < 0.0001) and MAC staining by ~20%. ⁷² , ⁸¹ PRELP is a leucine‐rich protein found in the extracellular matrix of connective tissue that has various binding partners, one of which is the terminal component, C9. ⁸¹ As described above, the classical MAC inhibitor, CD59, was also administered in a soluble form, sCD59, where it limited CNV lesion growth, even though the delivery was not localized to the site of laser injury and was delivered intravitreally packaged in an AAV5 vector. ⁷⁴ These studies paved the way for development of the clinical drug, HMR59, described above.

To increase the potential of soluble regulators to ameliorate complement activation at lower doses, chimeric molecules containing the C3d‐binding domain of complement receptor 2 (CR2) have been developed. These molecules localize to sites of complement activation through the iC3b/C3d‐binding domain in CR2, and provide therapeutic complement control through the second moiety, for example, the regulatory domains of FH. Vectorized TT30, a fusion molecule of CR2 and FH, also known as CR2‐FH, attenuated lesion growth in the mouse laser‐induced CNV model when packaged in an AAV5 vector and delivered via the subretinal route. ⁸² In a smoke‐induced ocular pathology model, targeting of CR2‐FH to RPE cells (using AAV5) rather than ganglion cells (using AAV2YF) was more effective at reducing complement activation, and moreover, subretinal delivery of CR2‐FH resulted in a better outcome than intravitreal injection. ⁸³ These data highlight the importance of correct serotype and route of administration for a therapeutic effect. ⁸³ TT30 was originally developed by Taligen Therapeutics (acquired by Alexion Pharmaceuticals in 2011). The soluble biologic, TT30 (ALXN1102 and ALXN1103), entered phase 1 testing for PNH, a disease characterized by overt C3b attack and MAC‐mediated lysis of erythrocytes (NCT01335165), and was considered safe, with no apparent dose‐related safety risks. While TT30 resulted in pharmacologically relevant inhibition of the alternative pathway, resulting in reduced lactate dehydrogenase levels, the therapeutic inhibition was short‐lived. ⁸⁴ This highlights the potential advantageous characteristics for gene therapy to deliver sustained levels of molecules that do not normally have the necessary bioavailability when administered as soluble biologics in vivo. Indeed, in 2022, Kriya Therapeutics announced the licensing of AAV CR2‐FH for treatment of GA, indicating the potential to translate biologics to a gene therapy platform (Press release: Kriya Licenses Next Generation Complement‐Targeted Gene Therapies for the Treatment of Geographic Atrophy and Other Ocular Diseases, published January 26, 2022, on https://kriyatherapeutics.com/).

A similar approach has been used to reduce complement attack on retinal ganglion cells (RGC) to ameliorate glaucoma‐associated neuroretinal injury in DBA/2 J mice. ⁸⁵ Rodents have a unique cell‐associated complement regulator that is not present in humans, termed complement receptor 1‐related protein Y (Crry). ⁸⁶ Crry possesses the activities of human CD46 (cofactor activity) and CD55 (decay accelerating activity) and is similarly comprised of the same building blocks, short consensus repeats (SCR) that make it amenable to molecular engineering. Crry has been expressed as a chimeric molecule with the C3d‐binding domain of CR2, as described above for CR2‐FH. DBA/2 J mice usually develop anterior segment anomalies and elevated IOP at 8–9 months of age, ⁸⁷ however, intravitreal administration of AAV2.CR2‐Crry resulted in reduction of C3d deposits and neuroprotection of RGCs at 10 months and optic nerve at 15 months. Crry expressed in its native, membrane‐anchored form has also been tested in Abca4 knockout mice, a model of Stargardt macular degeneration. ⁸⁸ AAV‐Crry inhibited complement at the RPE level, slowed photoreceptor degeneration, and reduced accumulation of inflammatory products (bisretinoids) resulting in rescue of the Stargardt‐like phenotype. ⁸⁸ Such results pave the way for constitutive, AP‐targeted gene therapy in a range of ocular diseases involving C3 deposition.

4.2. Gene targeting in systemic disease models

Gene/RNA therapy has also been targeted to the liver as it is the main “factory” of plasma complement; blocking biosynthesis of complement components in the liver might enable treatment of systemic disease or disease affecting multiple organs/tissues. Bora and colleagues used systemic administration of FB siRNA in mice and demonstrated a 95%–98% reduction of FB transcripts in the eye, liver, and spleen. ⁸⁹ Inhibition of liver‐derived FB expression resulted in decreased levels of MAC and angiogenic factors, vascular endothelial growth factor, and tumor growth factor β2, in the laser‐induced CNV model. ⁸⁹ Taking the alternative approach, Logan and colleagues demonstrated that sustained increases in circulating FI could be achieved using AAV8 to target the liver; this induced a fourfold to fivefold increase in circulating FI levels and decrease in activity of the AP. ⁹⁰ This FI gene therapy reduced the classic hallmarks of renal disease, C3 and IgG deposition, in a mouse model of systemic lupus erythematosus (NZBWF1), indicating an impact on the pathogenic process. ⁹⁰

Engineered complement regulatory molecules have also been adapted for systemic (liver targeted) gene therapy. In FH knockout mouse models of C3 glomerulopathy (C3G), constructs that harness the functional (cofactor and cell‐surface binding) domains of FH (SCR 1–5_18–20) and FHR dimerization domains (in FHR1 and FHR2) have demonstrated efficacy in reducing C3b deposition in the kidney glomerulus and normalizing C3 levels in serum. ⁹¹ , ⁹² Packaging and delivery of the latter construct within AAV resulted in long‐lasting and potent amelioration of C3G phenotype in the FH knockout mouse, overcoming the need for frequent dosing of the soluble biologic used in previous studies. ⁹² , ⁹³ , ⁹⁴ The same approach also ameliorated renal complement attack in an FHR5 gain‐of‐function mouse model of C3G. ⁹⁵ A study performed by Kamala et al. also revealed some of the risks of treatment with AAV: It was reported that one FH knockout mouse suffered thrombotic microangiopathy shortly after C3 levels were returned to normal, a parallel observation has been reported in human trials after AAV9 exposure; in the latter case, treatment with eculizumab ameliorated complement activation and TMA. ⁹³ , ⁹⁶

Building upon their experience using AAV to express CD55, CD46, and sCD59 (see above), Kumar‐Singh and colleagues have used fusion approaches combining functional domains of all three regulators and packaging into AAV as a single chimeric molecule. ⁹⁷ Minigenes encoding the complement regulatory domains from human CD46, CD55 and CD59 (SACT) or CD55 and CD59 (DTAC) were cloned into an AAV8 vector and injected into the peritoneal cavity of mice. After 3 weeks, a mouse model of liver disease was induced by injection of an antibody against murine platelet/cell adhesion molecule, CD31/PECAM‐1, followed by perfusion with human serum (as a source of complement). ⁹⁷ Both SACT and DTAC significantly attenuated MAC deposition in the liver vasculature in this acute model. Although the use of chimeric fusion proteins may have technical challenges, such as potential immunogenicity, these approaches offer a creative means to simultaneously address multiple facets of complement‐driven disease not otherwise possible with single molecule approaches.

High levels of C3b deposition on surfaces result in the formation of C5 convertases, triggering the terminal complement cascade and driving tissue injury and inflammation via MAC formation. Thus, C5 and the MAC can mediate pathogenic effects in diseases of AP dysregulation and silencing of C5 can bring therapeutic benefit. The C5 RNAi in clinical development by Alnylam Pharmaceuticals, ALN‐CC5, is described above, and an expanded literature indicates further promise in preclinical models of disease. For example, systemic administration of C5 RNAi reduced immune cell infiltration into the choroid plexus of ApoE knockout mice with an Alzheimer's disease‐like phenotype. ⁹⁸ An alternative approach to ameliorate the C5a‐driven aspects of complement‐mediated disease is to target its receptor, C5aR1. In a model of renal ischemia, mice injected with 50 μg of C5aR siRNA 2 days before induction of ischemia were protected from renal injury and had lower levels of serum creatinine and blood urea nitrogen compared to the control group. ⁹⁹ Interestingly, proinflammatory cytokines and chemokines such as tumor necrosis factor‐α, MIP‐2, and KC were also reduced in the C5aR siRNA‐treated group which reduced cell necrosis and reduced neutrophil influx into the kidney. ⁹⁹

5. CHALLENGES

Challenges associated with RNA‐targeting approaches include off‐target effects, immune responses to the siRNA or single‐stranded RNAs, degradation by exonucleases and delivery into the target cells. ¹⁰⁰ Silencing of non‐target sequences has been reported due to binding of short regions of the mRNA 3′‐UTR containing imperfect matches to the small RNA ¹⁰¹ , ¹⁰² ; one siRNA therefore has the potential to affect expression of hundreds of transcripts. ¹⁰² These off‐target effects appear to be dose‐dependent and can be addressed by careful selection or optimization of the RNA sequence. ¹⁰³ , ¹⁰⁴ Systemic clearance can hamper delivery, and uptake into the target cell requires penetration of barriers, such as endothelial cells. Simple diffusion through membranes is prevented by the large and hydrophilic nature of siRNAs. Various delivery strategies have been optimized to overcome these challenges, including sugar modifications, rational design, or chemical modification of the siRNA such as replacement of the highly charged and unstable phosphodiester backbone with a phosphorothioate backbone. Approaches that facilitate delivery include formulation of siRNAs as lipid nanoparticles which promotes RNA packing, increases stability, and allows for passage through the lipid bilayer, or conjugation to a ligand that binds a specific high‐capacity receptor on the target cell. ⁹⁹ Suitable ligands include GalNAc sugars that bind to the abundantly expressed hepatocyte receptor ASGPR. Tri‐antennery GalNAc displays the highest affinity toward ASGPR and because of its rapid recycling and turnover, a single administration of GalNAc siRNA conjugate yields very high siRNA uptake. ¹⁰⁵ , ¹⁰⁶ Delivery of ASOs in a mouse study reported increased hepatocyte uptake by up to 10‐fold and a phase 1 study using a second‐generation antisense drug (using triantennery GalNAc) designed to reduce the synthesis of apolipoprotein in the liver found improvements of up to 20–30‐fold. ¹⁰⁵ , ¹⁰⁷ , ¹⁰⁸

5.1. Technological risks

Although a promising approach, CRISPR gene editing also has several limitations and risks which challenge its use as a therapeutic tool to correct disease‐causing mutations in the human genome. These challenges include safe and effective delivery into target cells, off‐target effects, immunogenicity against Cas9 protein, ethics, and efficient excision of the target region or insertion of template DNA to correct a mutation. ¹⁰⁹ , ¹¹⁰ , ¹¹¹ , ¹¹²

Despite numerous rAAV‐mediated gene therapy trials being underway, rAAV vectors have not yet reached full maturity in clinical applications; successes in preclinical studies often do not translate to positive clinical outcomes. The emerging wealth of preclinical and clinical data has helped identify four fundamental challenges, restricting rAAV‐mediated gene therapy from realizing its full potential: immunogenicity toward the vector, managing treatment doses, controlling transgene expression, and vector manufacturing.

The host's immune system remains the most challenging barrier for rAAV‐mediated gene therapy approaches. Most humans harbor anti‐AAV neutralizing antibodies (NAbs) due to subclinical infections by wildtype virus during infancy which can greatly limit the efficiency of gene transfer. ¹¹³ , ¹¹⁴ Seroprevalence of pre‐existing NAbs in the human population worldwide varies for different AAV variants, with the highest being observed for AAV2, and because of sequence conservation in AAV capsids, NAbs can cross‐react with multiple AAV serotypes. ¹¹³ , ¹¹⁵ , ¹¹⁶ , ¹¹⁷ Interestingly, complement is also activated following systemic administration of high doses of rAAV, and Apellis Pharmaceuticals announced a program in which APL‐9, a C3 inhibitor, would be tested to control host immune responses to AAV vector administration in gene therapies (Press release: Apellis Pharmaceuticals Will Commence APL‐9 Program to Control the Complement System in Host Responses to AAV Vector Administration for Gene Therapies. Published July 18, 2019, on www.apellis.com). No results have been published. Innate immune responses against rAAV vectors and AAV cis‐regulatory sequences have further been implicated as a cause for toxicities in gene therapy trials, and NAbs in particular can activate the complement system via the alternative pathway. ¹¹⁸ rAAV‐associated pathogen‐associated molecular patterns (PAMPs) have been described, including ligands on rAAV capsids that bind to toll‐like receptor 2 (TLR2) on cell surfaces, ¹¹⁹ the binding of TLR9 to unmethylated CpG dinucleotides in the vector genome and double‐stranded RNA derived from the bidirectional promoter activity within AAV‐ITRs has also been described. ¹²⁰ , ¹²¹ , ¹²²

Over time, a cellular and humoral adaptive immune response may be mounted against the capsid or the foreign transgene protein, mediating the induction of either CD8+ T cell or CD4+ T cell responses via antigen presentation on Class I MHC or Class II MHC molecules, respectively, leading to destruction of transduced cells by cytotoxic T cells and generation of transgene product‐specific antibodies blocking re‐adminstration. ¹²³ , ¹²⁴ , ¹²⁵ , ¹²⁶ , ¹²⁷ , ¹²⁸ , ¹²⁹ In particular, ocular toxicity has been demonstrated to be promoter dependent, with ubiquitous promoters (CMV, CAG, Ubiquitin C) eliciting greater toxicity than cell‐specific promoters in the retina. ¹³⁰ , ¹³¹ It is hypothesized that ubiquitous promoters can drive transgene expression in multiple cell types including immune cells, resulting in higher off target transgene protein concentration and consequently higher immune response. In the eye, the route of administration strongly dictates the retinal transduction pattern, the biodistribution profile of the vector, shedding and the resulting immune responses. ¹³² , ¹³³ Preclinical studies show that innate and adaptive immune responses are clearly dose dependent; ocular immune responses were induced when doses of greater than 1 × 10¹¹ viral genomes per eye were injected, and these responses were absent when the administered dose was below this threshold. ¹³⁴ , ¹³⁵ Process and product‐related impurities are known to contribute to unwanted immune responses and heterogeneity and the ratio of empty to full AAV capsids creates the need for higher doses which can stimulate stronger immune responses, including activation of the complement classical pathway following high‐dose systemic administration. ¹¹⁸ , ¹²⁹ , ¹³⁴ , ¹³⁶ , ¹³⁷

Efforts to diversify AAV capsids from WT serotypes can help circumvent the neutralizing effect of pre‐existing humoral immunity and have the potential to reduce vector dose. AAV capsid engineering can be achieved through methods including rational design, directed evolution, isolation of natural variants, reconstruction of viral evolutionary lineage and machine learning, to generate diversified chimeric capsids with enhanced attributes such as immune evasion, improved transduction efficiency, and selected cell tropism. ¹⁸ , ¹³⁸ , ¹³⁹ , ¹⁴⁰ , ¹⁴¹ , ¹⁴² Modifications of the recombinant vector genome may include the use of cell/tissue‐specific or synthetic enhancers/promoters to reduce toxicity associated with constitutive strong promoters, codon optimization (removal of native CpG dinucleotides, CpG methylation, reduction of G/C content), and incorporation of TLR9 inhibitory nucleotides sequences. ¹²¹ , ¹³⁰ , ¹⁴³ , ¹⁴⁴ , ¹⁴⁵ , ¹⁴⁶ , ¹⁴⁷ Encapsulation of vector content remains a black box, and further investigation is required to understand the role of impurities and contaminants in inflammation. Rituximab and rapamycin inhibiting T and B cell activation, IgG cleaving endopeptidase, and epitope masking via exosome‐encapsulation have been shown to block anti‐capsid immunity. ¹⁴⁸ , ¹⁴⁹ , ¹⁵⁰ , ¹⁵¹ , ¹⁵²

6. CONCLUDING REMARKS

Overall, there are a multitude of genetic and RNAi therapies targeting the AP and/or its downstream effectors in preclinical development and a few in clinical development with huge potential for cutting‐edge treatment of disease (Figure 4). The field will need to address some of the challenges associated with gene targeting, such as route of administration, choice of AAV capsid, or avoidance of off‐target effects, in order to provide long‐lasting, inflammatory‐free solutions to complementopathies. The gene‐targeting field is still very much in its infancy, with very few therapies currently available to patients and still lessons being learned along the way, but they herald the dawn of new and exciting innovative genetic medicines that may soon become mainstream treatments for chronic human diseases in the future.

FIGURE 4.

Alternative pathway gene‐targeting therapies currently in clinical development or preclinical characterization. The diagram illustrates RNA therapies, CRISPR and AAV gene therapies discussed in this review and demonstrates that most of the development is in the treatment of ocular models using AAV gene therapy. Abbreviations: AAV, adeno‐associated virus; ASO, antisense oligonucleotides; CRISPR, clustered regularly interspaced short palindromic repeats; RNAi, RNA interference. Figure was created with BioRender.com

CONFLICT OF INTEREST

All authors are or were employed by Gyroscope Therapeutics Limited when the manuscript was written.

Dreismann AK, Hallam TM, Tam LC, et al. Gene targeting as a therapeutic avenue in diseases mediated by the complement alternative pathway. Immunol Rev. 2023;313:402‐419. doi: 10.1111/imr.13149

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this study.