Gene Therapy

Germline versus somatic gene therapy

Why germline therapy is attractive: it offers the chance to actually cure a genetic disease, especially a monogenic one (caused by a mutation in a single gene). You would simply provide a fully functional copy of the defective gene to the germ cells.

Problems with germline therapy:

• Extremely difficult to achieve.

• Potential for unforeseen negative effects on the individual and on future generations.

• Considered unethical: society does not like scientists altering the genome of germ cells.

Case mentioned: a Chinese scientist claimed to have edited the genes of twin girls. The lecturer was unsure whether the editing was successful, but the reaction showed that germline therapy is still considered extremely unethical. [Added] This was announced in 2018.

For somatic therapy, the number and type of cells that acquire the therapeutic gene depends on the delivery strategy.

3. Things to consider when designing a gene therapy

1. What is the genetic disease, which cells are affected, and which are the target cells?

2. How will the gene be delivered, and are the target cells accessible? (Strategies: ex vivo, in situ, in vivo.)

3. What proportion of target cells must be altered to see a therapeutic effect: the majority, or would 5–10% be enough?

4. What happens if the gene is overexpressed? Expression above the normal level for that cell type could cause side effects, organ failure, or even be fatal.

5. How long must the gene be expressed? For an inherited disease you usually want expression for the person’s lifetime. For a tumour you may need expression only while the tumour is present.

4. Delivery strategies

A therapeutic gene cannot simply be added as naked DNA or RNA: it will not get into cells and will not be expressed. It must first be packaged into a vector. The vector carrying the transgene (therapeutic gene) is then delivered in one of three ways.

Which strategy is used depends on the disease, the affected cell types, the tissues to target and whether they are accessible.

5. The three approaches to gene therapy

The three approaches are gene addition, gene correction/alteration (genome editing) and gene knockdown.

5.1 Gene addition

• The longest-established approach, used in many studies and relatively simple.

• The defective gene stays in place and you add an extra, fully functional (wild-type) copy, called the corrective gene.

• Two ways to do it:

5.2 Gene correction / alteration (genome editing or genome engineering)

• The least common approach because it is new and difficult.

• Ideal aim: replace the mutated DNA with a corrected, fully functional version of the gene.

• Mechanism: homologous recombination (HR). HR is a natural cellular process with several functions, including repair of double-strand breaks (which occur frequently during DNA replication). You introduce a donor DNA (the gene without the mutation). It has very high homology with the target gene (differing only at the mutation), so the cell’s HR machinery can swap the mutated version for the corrected one.

Advantages:

• You alter the genome very little: just one mutation in one gene.

• The corrected gene is under the endogenous promoter, so expression is regulated and at the correct levels in each cell type.

Disadvantages:

• Very difficult. The frequency of homologous recombination is very low, so the frequency of random integration of the donor DNA elsewhere in the genome is too high by comparison.

Solution: introduce a double-strand break specifically within the target gene.

• This brings the HR machinery together at the target gene, so the donor DNA is much more likely to be incorporated in the correct position and much less likely to integrate elsewhere.

• [Added] Cells can also repair double-strand breaks by non-homologous end joining, which is why a donor DNA is supplied for correction.

Ways of making a targeted double-strand break:

• Genome editing is not only for correction. The same technology can do gene addition, gene deletion, gene disruption and more.

5.3 Gene knockdown (RNA interference, RNAi)

• Used where a protein expressed at high levels causes the disease. You knock down expression of a specific gene by targeted gene silencing using RNA interference.

• RNAi is a natural cellular process that cells use to fine-tune protein levels. Gene therapy takes advantage of this existing, highly complex machinery.

• Two approaches:

• [Added] In cells, the enzyme Dicer processes dsRNA into siRNAs, which are loaded into the RISC complex that cuts the target mRNA. In practice, synthetic siRNAs of about 21–23 nucleotides are often delivered directly, because long dsRNA triggers an interferon response in mammalian cells; the lecture’s description of a ~200-nucleotide dsRNA being chopped up is given as presented.

• RNAi therapy has been around a little longer than genome editing.

6. Clinical trial phases

• Phases run from phase 0 to phase 4 (see clinicaltrials.gov).

• Early phases are exploratory and then concerned with efficacy and safety; higher phases (3 and 4) are more concerned with things like cost-effectiveness and side effects.

• [Added] Standard summary: phase 0 exploratory first-in-human; phase 1 safety and dose in a small group; phase 2 efficacy and side effects; phase 3 larger trials confirming efficacy against standard care; phase 4 post-approval monitoring.

• Where each approach sits:

  • Genome editing (newest): mostly phase 0 and phase 1.

  • RNAi: mainly phase 1, 2 and 3 (the lecturer also said “mainly phase 1 and 2” at one point).

  • Gene addition (oldest): mostly phase 2 to 4.

7. Vectors

7.1 General principles

• A vector is needed to get the therapeutic gene into cells, into the nucleus, expressed at high levels and for a good period of time.

• Viral vectors are the most common. Viruses have evolved over thousands of years to infect cells, reach the nucleus and express their genes, and gene therapy takes advantage of this.

• To use a virus, you must mutate key viral genes to make the virus replication-incompetent (unable to replicate in us). A wild-type virus would replicate and make the person sick. The therapeutic gene is then packaged into the defective viral genome.

• Non-viral vectors are also possible but not popular because at present their disadvantages outweigh their advantages (section 7.5).

Desirable properties of a vector: safe; cost-effective; possibly targeted/selective; able to give expression for a good length of time; available in a highly concentrated and active form; immunologically neutral (a strong immune response would wipe out the vector and therefore the therapeutic gene).

Four barriers to successful gene therapy: a 2011 review article (recommended, to be posted on Moodle) sets out four barriers to getting the therapeutic gene into cells, into the nucleus, expressed at high levels and for a good period of time. The lecture did not go through the four barriers in detail; read the review.

7.2 Retroviruses

• RNA genome. On infection, reverse transcriptase converts the RNA genome into DNA, which randomly integrates into the host genome.

• Advantages: stable gene expression (integration); high transduction efficiency.

• Disadvantages: insertional mutagenesis from random integration; they only infect dividing cells, so you must consider how proliferative the target tissue is.

• Tumours as a target: the need for dividing cells can be an advantage, because tumours are more proliferative than surrounding normal tissue, so a retrovirus preferentially infects tumour cells.

Case study: X-linked severe combined immune deficiency (X-SCID)

• The disease: boys lack immune cells (the lecturer said B and T cells; see section 11). It is caused by a mutation in the gene for interleukin-2 receptor gamma (IL2RG).

• The trial: the wild-type IL2RG was packaged in a retrovirus, using an ex vivo strategy: bone marrow stem cells were removed, cultured, infected with the retrovirus, and returned to the patient.

• Success: many patients developed a functional immune system.

• Problem: 5 out of 20 successfully treated patients went on to develop leukaemia.

• Cause: the retrovirus activated an oncogene, LMO2 (LIM domain only 2). The same activation of LMO2 and leukaemia was seen in a trial for a different immune deficiency, Wiskott–Aldrich syndrome.

• Current hypothesis: LMO2 activation alone is not enough to cause leukaemia; you also need additional growth-promoting activity. The IL2RG transgene has its own independent growth-promoting activity, and the combination of the two leads to leukaemia. So the leukaemia is transgene-specific.

• Supporting evidence: in a trial using the same method for ADA deficiency (mutation in adenosine deaminase), no patients developed leukaemia, and adenosine deaminase has no growth-promoting activity.

• Take-home message: retroviruses can be used successfully and safely, but you must carefully consider whether the therapeutic gene has growth-promoting activity.

7.3 Adenoviruses

• Very common viruses that infect humans. DNA genome, which exists episomally (does not integrate).

• Advantages: no insertional mutagenesis risk; infect both dividing and non-dividing cells; high transduction efficiency; very high levels of gene expression; can package a large amount of non-viral DNA. Because of these advantages they are the most widely used vector in clinical trials.

• Disadvantages:

  1. Transient expression: only for a matter of weeks.

  2. Highly immunogenic.

• Why the two disadvantages together are a headache: you usually need expression for longer than a few weeks, so you need repeat rounds of therapy. But after the first dose the body has antibodies against that adenovirus, which would wipe out the virus (and the therapeutic gene) almost immediately. The workaround is to package the gene in a series of different adenovirus serotypes, using a different one for each round. This is laborious and expensive.

7.4 Adeno-associated viruses (AAV)

• Very limited immune response, because when the therapeutic gene is packaged, most of the viral DNA is removed.

• Major disadvantage: limited packaging capacity, up to about 4 kb of non-viral DNA. [Added] The commonly quoted limit is about 4.7 kb.

• Mostly episomal. About 1% integrate, and when they do it is specifically at chromosome 19 (unlike the random integration of retroviruses). It is not known whether integration at chromosome 19 can be pro-oncogenic.

• Successful clinical example: Leber congenital amaurosis (LCA). An incurable blindness disorder caused by mutation in RPE65. Wild-type RPE65 was packaged in AAV and delivered by retinal infusion. Many patients showed marked improvement in measurable vision parameters.

• Animal example: in squirrel monkeys, an opsin gene packaged in AAV corrected colour blindness.

7.5 Non-viral vectors

• Advantages: no immune response; can package as much DNA as you want.

• Disadvantages: very difficult to get the vector into cells; even if it gets in, very difficult to get the gene expressed at high levels for a good period of time.

• At present the disadvantages outweigh the advantages, but if these problems were overcome they could become the vector of choice.

7.6 Vector comparison

8. Gene therapy in clinical trials

ClinicalTrials.gov lists gene therapy trials for a wide variety of diseases, from ADA-SCID and Wiskott–Aldrich syndrome to many cancers and muscular dystrophy. The lecture focused on two diseases: cystic fibrosis and cancer.

9. Case 1: cystic fibrosis (CF)

Why CF should be an ideal target

• It is a monogenic disease, caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene.

• Although CF affects many organs, the lung is the major organ whose failure usually causes death, and the lung is accessible: a vector can be squirted up the nose to reach the lung epithelium.

Normal CFTR function

• CFTR is an ion channel in the apical (luminal) membrane of airway epithelial cells. It acts as a chloride channel (promoting chloride secretion) and can also act as a bicarbonate channel.

• When active it influences the epithelial sodium channel (ENaC), which promotes sodium absorption. By regulating chloride secretion and sodium absorption, CFTR regulates fluid movement across the epithelium.

• Mucociliary clearance: inhaled bacteria, viruses, particles and toxins get stuck in a thin layer of mucus on the luminal surface. Cilia beat and push the mucus (with the trapped material) up and out of the lungs. This prevents lung infections.

What goes wrong in CF

1. CFTR is absent, does not reach the luminal membrane, or does not work.

2. Loss of proper chloride secretion and increased sodium absorption.

3. Hyperabsorption of airway surface fluid.

4. Mucus becomes thick, sticky and dehydrated instead of thin.

5. Cilia cannot beat, so mucociliary clearance fails.

6. An ideal environment for bacterial infection and growth.

7. Chronic, strong inflammation.

8. Tissue damage and progressive loss of lung function.

9. Organ failure in the lung, which usually causes death.

The gene therapy idea

Package a functional CFTR in a viral vector and squirt it up the nose. The virus infects lung epithelial cells, which express functional CFTR in the luminal membrane, restoring chloride secretion, sodium absorption and mucociliary clearance.

Clinical trial outcomes

• Since CFTR was identified, there have been well over 20 clinical trials with various methods, but none achieved any major therapeutic effect.

• Only minor effects were seen, such as transgene mRNA expression or partial correction of chloride secretion. This suggests some elements of the technique work.

Why has it not worked?

• The lung is a barrier: one of its main jobs is to stop viruses infecting epithelial cells, and in CF the thick, sticky mucus makes it worse.

• Airway inflammation (a hallmark of CF) means you reach unacceptable toxicity quickly, so you cannot flood the lung with virus.

• Anatomical complexity: a branching series of tubes that get smaller and thinner, so intranasal delivery is unlikely to reach the lower lung.

• Complex pattern of CFTR expression in the lung, which cannot be recapitulated by squirting virus up the nose.

• It is not even known what the correct target cell is, so the method may be targeting the wrong cell type.

Outlook: not great in the near future until these problems are solved, but gene therapy could dramatically improve the lifespan and quality of life of CF patients if it can be made to work.

[Added] The lecture does not mention that, separately from gene therapy, small-molecule CFTR modulator drugs have since become available and have greatly changed CF treatment.

10. Case 2: cancer and conditionally replicating adenoviruses (CRAds)

Gene therapy approaches in cancer

A wide variety are in trials: immuno-gene therapy, suicide gene therapy, overexpression of tumour suppressor genes, and knockdown of oncogenes. The lecture focused on CRAds.

The concept

• Normally viral vectors are made replication-defective so they cannot replicate in our cells. But tumours proliferate very quickly, so a non-replicating virus has little effect.

• CRAds can replicate in cancer cells but not in normal cells.

• Oncolysis cycle: the virus infects a cancer cell, replicates (makes many copies), and the cell ruptures (oncolysis), releasing virus particles into the local environment. These infect further cancer cells, replicating and causing more oncolysis, so the tumour mass is depleted over time. The particles also infect surrounding normal cells but cannot replicate there.

Two ways to make a virus replicate only in cancer cells

1. Transcriptional regulation

• Put essential viral genes (needed for replication) downstream of a promoter that is highly active in tumour cells but not in normal cells.

• Example: the telomerase promoter. About 90% of cancers express telomerase at very high levels, whereas most normal cells do not.

• In cancer cells the viral genes are expressed and the virus replicates; in normal cells they are not, so the virus cannot replicate.

2. Deletion mutants: Onyx-015

• Onyx-015 carries a mutation (deletion) in the viral gene E1B-55kDa.

• Normal viral infection: a normal infected cell activates p53, p53 levels rise, and p53 promotes expression of genes that cause apoptosis, so the cell dies and the virus cannot replicate. Wild-type virus avoids this because E1B-55kDa binds p53 and prevents it from working.

• Onyx-015 does not make E1B-55kDa, so in a normal cell p53 triggers apoptosis and the virus cannot replicate.

• But in about 50% of cancers p53 is mutated and non-functional, so apoptosis cannot be triggered and Onyx-015 can replicate. Onyx-015 is therefore useful for the roughly 50% of cancers with mutated p53.

• Onyx-015 was the first CRAd to enter clinical trials, used for a variety of cancers. On its own its therapeutic effect was limited, but combined with other treatments such as chemotherapy the effects were more promising.

• [Added] The extent to which Onyx-015’s selectivity depends on p53 status has been debated in the literature; the lecture presents the p53 explanation.

CRAds as gene therapy vehicles

Using a virus to kill cancer cells is not itself gene therapy, but CRAds can carry a therapeutic gene, giving a double effect (oncolysis plus the therapeutic gene):

• A CRAd expressing siRNA against Ras. Ras is an important oncogene often upregulated in human cancer, so the knockdown of Ras adds to the oncolytic effect.

• A CRAd expressing shRNA against VEGF. VEGF promotes angiogenesis and tumour progression, so knocking it down adds to the oncolytic effect.