Hepatitis after gene therapy, what are the possible causes?

Hepatitis is a common adverse event following gene therapy for haemophilia, often associated with a loss of transgene expression. Investigating the potential causes and implications of this is crucial for the overall success of treatment. Gene therapy trials using adeno‐associated virus (AAV) vectors have demonstrated promising results marked by increases in factor FVIII and FIX levels and reductions in episodes of bleeding. However, hepatocellular injury characterised by elevations in alanine aminotransferases (ALT) has been noted. This liver injury is typically transient and asymptomatic, posing challenges in determining its clinical significance. Proposed causes encompass immune‐mediated responses, notably T cell cytotoxicity in response to the AAV vector, direct liver injury from the viral capsid or transcribed protein via the unfolded protein response and pre‐existing liver conditions. Liver biopsy data conducted years post‐gene therapy infusion has shown sinusoidal infiltration without significant inflammation. The overall safety profile of gene therapy remains favourable with no evidence drug‐induced liver injury (DILI) based on Hy's Law criteria. Essential pre‐therapy monitoring and identifying patients at high risk of liver injury should involve liver function tests and non‐invasive fibroscans, while novel blood‐based biomarkers are under exploration. Further research is required to comprehend the mechanisms underlying transaminitis, loss of transgene expression and long‐term effects on the liver, providing insights for optimising gene therapy for haemophilia.

concentrates and the isolation of the FVIII and FIX genes allowed for safer, efficacious treatment options that require regular administration.However, the hope of a single-dose treatment that could provide a curative solution in the form of gene therapy remained unaddressed by these therapeutic advances. 1eno-associated virus (AAV) is a single-stranded DNA virus that has been used as a vector for haemophilia gene therapy. 1 It has the benefit of low pathogenicity, a high degree of liver specificity through modification and a predominant lack of integration into host DNA once infused.The transgene containing genetic information for the production of FVIII and FIX is transferred to the hepatocytes; here it enters the cytoplasm and then the cell's nucleus. 2 The vector capsid uncoats and the genetic material contained forms circular episomal vector DNA in the hepatocyte nucleus. 3The genetic material then undergoes the complex process of transcription and translation to produce the factor protein required. 2

| LIVER COMPLI C ATI ON S OF G ENE THER APY IN HAEMOPHILIA
Almost all of the recent Phase 1 to 3 clinical trials using gene therapy to treat haemophilia A and B have successfully demonstrated significant, although variable, increases in FVIII and FIX levels, with notable reductions in bleeding episodes and the need for prophylactic treatment. 4However, the most notable adverse event (AE) related to gene therapy for haemophilia is liver inflammation-'hepatitis', defined as an increase in serum aminotransferases. 5The rise in serum alanine aminotransferase (ALT) is often accompanied by a rise in serum aspartate aminotransferase (AST) but markers of the synthetic function of the liver such as the total bilirubin level and coagulation remain within normal parameters and, to-date, no case of significant hepatotoxicity has been reported. 6However, the hepatitis associated with gene therapy is variably associated with a loss of transgene expression, 7 suggesting that damage to the hepatocytes compromises the function of the AAV delivered DNA.
Corticosteroids and other forms of immunosuppression have been used as treatment for, and prophylaxis against, ALT elevations to protect transgene expression with varying degrees of success and with the consequence of steroid related side effects. 8ses in ALT are generally transient-1 to 3 times the upper limit of normal (ULN) but levels have been seen up to 20 times the ULN or noted to persist over months or years. 2 As many as 86% of patients developed elevations in their liver enzymes, particularly ALT at 7-14 weeks post-drug administration in one of the pivotal clinical trials for haemophilia A. 9 The median time to the first ALT elevation was 8 weeks, with a median duration of 15 days.Steroid treatment was administered for a median duration of 230 days, and up to 71.8% of patients experienced steroid-related AEs.
The HOPE-B trial for haemophilia B similarly reported a median time of 5.1 weeks to first ALT elevation, with a median duration of 17 days. 10The steroid treatment duration required was a median of 74 days.The general practice during these trials has been to introduce steroids prophylactically at 2-4 weeks before the onset of transaminitis. 7It remains uncertain what the time to resolution of hepatitis is without steroids.
The instances of elevations in aminotransferases vary for FVIII and FIX studies, occurring more commonly in haemophilia A and with a severity that appears to be dose-dependent. 2Trials for haemophilia B use AAV vectors that express FIX with a gain of function Padua mutation, which codes for FIX protein that has increased activity, this allows a lower dose to be given to haemophilia B patients and may, perhaps, mitigate the ALT elevations. 11Transgene expression for FIX is more durable and stable over time which may reflect the degree of liver inflammation.

| A SS E SS ING THE S E VERIT Y OF HEPATITIS
Elevations in liver enzymes are a good indicator of hepatocellular injury and, in other contexts, have been associated with increased morbidity and mortality. 5,12However, in this setting, they are ineffective at ascertaining the severity of injury. 12Where liver injury has occurred as the result of a drug such as gene therapy, this is referred to as drug-induced liver injury (DILI) 13 and Hy's Law is a criterion used to characterise severe DILI.A positive Hy's Law is defined as an elevation in ALT/AST >3× ULN and total bilirubin >2× ULN in the absence of other causes of liver injury.Hy's Law is a good predictor of severe and fatal liver injury as severe DILI carries a 10%-50% risk of death or requirement for liver transplantation. 3,14e liver injury observed following gene therapy for haemophilia is documented as an AE, and in the pivotal FVIII trial, 15 the severity was graded according to the Common Terminology Criteria for Adverse Events (CTCAE).CTCAE grading ranges from mild (Grade 1), moderate (Grade 2), severe (Grade 3), life-threatening (Grade 4) to death (Grade 5) and is based on required interventions for an AE and the effects of the AE on activities of daily living (ADLs) (Table 1). 16st ALT elevations seen in gene therapy trials are Grade 1 to 2 in severity. 17 the aforementioned trial for haemophilia A, 86% of 115 participants who received gene therapy had AEs in the form of ALT elevations. 9Only 8.2% of them were deemed to have Grade 3 transaminitis based on the CTCAE but, critically, none of them met Hy's Law for drug-induced liver injury.Most trial data in haemophilia follows a similar pattern concerning the severity of liver injury, 3 this implies that the safety profile for gene therapy in haemophilia does not include the morbidity and mortality associated with significant DILI.
In haemophilia trials, the liver injury appears to be transient and asymptomatic, and the significance to long-term liver health remains unclear. 18Given that no clinically meaningful consequences of the hepatitis have been reported to-date it is probable that there will be no adverse health sequelae.However, patients studied in clinical trials underwent a vigorous pre-treatment assessment for liver health and individuals with known liver disease, active and chronic liver infections with hepatitis B and C and patients with HIV on hepatotoxic medications were not recruited. 2This criterion meant that a large proportion of haemophilia patients were excluded as those treated with plasma-derived factor products in 1980s may have iatrogenically acquired these conditions from donated blood products. 3nce, although the current safety data suggests that the AAV associated hepatitis in patients with haemophilia is clinically unimportant, care will be required as the treatments are increasingly used in patients with pre-existing liver damage.It is the author's view that the degree of liver fibrosis should be assessed before initiating gene therapy by a non-invasive technique such as transient elastography.
Patients with significant pre-existing liver fibrosis should be considered carefully for their suitability for gene therapy and assessed on a case by case basis.
The issue of hepatitis and its association with a decline in transgene expression 19 is much more relevant for clinical care as it leads to a reduction in factor level and undermines the success of gene therapy.It is important to note that some individuals develop hepatitis without loss of transgene expression and vice versa. 17There is also a gradual decline of factor expression over time that seems to be more pronounced in patients with haemophilia A. The mechanism for liver injury is incompletely understood and precludes a more evidence based approach to therapy but there are currently a number of possible causes.

| Immune responses in AAV associated hepatitis
Immune-mediated responses to the AAV vector in transduced hepatocytes have been hypothesized as a potential cause for the transaminitis found in haemophilia patients treated with gene therapy. 9These responses are thought to lead to transduced hepatocyte destruction and hence a loss of transgene expression.The response could be directed against the AAV vector capsid proteins, the transgene itself and potentially the expressed protein. 11The AAV vector used in gene therapy has no discernible viral genome products to cause an immune response however the vector capsid is antigenic and has the potential to induce a reaction. 20psid-specific CD8 + T cells can be triggered by hepatocytes which act as antigen-presenting cells by expressing capsid proteins on MHC class 1 receptors on the cell surface. 21The recruitment of CD8 + T cells may occur through innate and adaptive immunity as there is supposed widespread exposure, in the human population to wild-type AAV vector serotypes, leading to individuals with preexisting AAV-neutralising antibodies.Although patients with preexisting AAV antibodies at high titres were excluded from the clinical trials it is possible that antibody levels that were undetected by the assay may have been associated with a memory response to the viral vector.While liver biopsies were not routinely performed in gene therapy trials, the limited number performed supports this proposed mechanism as they show inflammatory infiltrates in the liver consisting mainly of CD8 + T cells. 8Interestingly, some studies show elevations in ALT without a significant cytotoxic T-cell response (detected by EliSpot assays). 4This suggests that either the assay does not detect all of the possible responses or that there are alternative mechanisms for the hepatitis and that there may be specific vector and recipient characteristics that influence the immune response. 4,21ere is also evidence of immune responses to the transgene as Toll-like receptor 9 (TLR9) has been implicated in triggering immunotoxicity and subsequent loss of transgene expression. 2,22TLR9 detects unmethylated cytidine-phosphate-guanosine (CpG) sequences in DNA on the transgene which are recognised as pathogenic leading to an immune response.Some clinical trials on FIX have noted that the presence of a higher number of these CpG motifs led to a loss of factor expression and ALT elevations, in which loss of FIX expression could not be recovered with steroids. 19A potential solution for this is CpG motif-depleted DNA but this is difficult to accomplish without compromising protein expression.A final potential immune mechanism for the observed hepatitis is an immune response to the expressed protein leading to hepatotoxicity, but the evidence for this remains limited. 23ere are additional putative mechanisms for immune mediated hepatitis in patients undergoing gene therapy for haemophilia.These b Self caring ADLs refers to bathing, un/dressing, feeding self, toileting but not bedridden.

TA B L E 1
The National Cancer Institute's Common Terminology Criteria for Adverse Events version 4.0.
include vector purity, and any non-human codon sequences in the transgene and the manufacturing platform. 7There is, however, little evidence to support these mechanisms.Host-dependant factors for immunogenicity may, of course, play a role in the incidence and severity of the hepatitis and these include host immune genotype and phenotype, pre-existing immunity, age and the prior presence of inflammation.Immune-related hepatitis is usually observed early post-gene therapy, however in haemophilia A, ALT elevations can persist over time 24 and the reasons for this are not well understood but seem unlikely to be linked to an immune response to the capsid as expression of this protein is dependent upon the initial dose which is time-limited.

| Direct liver injury
Proposed non-immune causes of AAV associated hepatitis include direct injury to the hepatocyte by the viral capsid, endoplasmic reticulum (ER) stress from misfolded proteins and potentially pre-existing liver disease. 8,25Beta domain deleted FVIII is integrated into viral vectors with a small carrying capacity leading to constructs which are 4.7 kb in size. 2 Once transduced these truncated cassettes undergo gene repair in the hepatocyte nucleus and are organised into monomeric or concatemeric circular episomes.In this form, persistent transgene expression is possible.The high viral capsid dose per hepatocyte post-transduction may be overwhelming due to the level of processing required, including viral capsid degradation and this could lead to cell death and liver injury. 26patocyte expressed FVIII is ectopic as this factor is normally synthesised in the liver sinusoidal endothelial cells. 2 The transgene encoded by the AAV vector is a modified protein (as noted above it lacks the beta domain) and is a large and complex protein whose, synthesis in hepatocytes may induce intracellular stress and a possible lack of the receptors normally found in sinusoidal cells may hinder its secretory pathway. 27If an overload of factor protein synthesis occurs in the ER, the protein may misfold, aggregate and degrade.
Unfolded or misfolded proteins can accumulate which may activate the unfolded protein response (UPR). 23The UPR regulates the production and secretion of proteins from the cell by producing more ER to process the accumulated protein, if the cellular stress is pronounced it can lead to cell death which may manifest as elevations in ALT and loss of transgene expression.No liver biopsy studies during the acute phase of the AAV-associated hepatitis have been performed to examine this hypothesis but biopsy studies many months after the hepatitis (see later) do not show any evidence of ER mediated stress at this time. 28Whether this indicates that ER mediated stress is not involved in AAV associated hepatitis or whether it is only present at very early time points post transduction is unknown.
Another possible cause for liver elevations in gene therapy for haemophilia is pre-existing liver conditions. 8Where gene therapy was used in children with spinal muscular atrophy (SMA), at a high dose of 1.1 × 10 14 vg/kg, asymptomatic transaminitis was observed that resolved with the administration of steroids. 3,8,22,29However, two of the participants developed Grade 3 acute liver injury which fulfilled Hy's law criteria for severe DILI. 29Liver biopsy samples were taken, which suggested inflammatory infiltration by CD8 + T cells but also indicated that pre-existing liver disease may have contributed to the development of severe liver injury.Autopsy data from a trial of gene therapy in X-linked myotubular myopathy (XLMTM) also found pre-existing liver disease as a main cause of liver failure that resulted in death but without an inflammatory infiltrate. 22In clinical trials of AAV gene therapy for haemophilia much lower doses of viral vectors were used and great care was taken to exclude patients with preexisting liver disease.In clinical practice, patients with pre-existing liver disease may require additional monitoring to identify and manage any clinically significant hepatic reactions and vigilance will be required in the management of such patients.

| Liver biopsy data
To accurately assess the causes of ALT elevations, liver biopsies have been proposed as the most sensitive and specific way to determine hepatocyte injury and death. 26They can also be used to determine FVIII and FIX distribution and expression, AAV capsid presence, T cell cytotoxicity, inflammation and evidence of the immune response.A study looking at interindividual differences in factor expression reviewed liver histology 2.1-4.1 years post infusion with valoctocogene roxaparvovec (AAV5-hFVIII-SQ), a gene therapy for haemophilia A. 28 The study took biopsy samples from five participants who had received this drug as part of a Phase 1/2 clinical trial.
Results of local and independent analysis of these samples by specialist liver pathologists found no clinically significant inflammation, but noted sinusoidal infiltrates, which may be found in nonpathological liver samples from individuals without liver disease and whose significance is uncertain.The findings noted normal liver architecture without evidence of fibrosis, necrosis or dysplasia and no evidence of ER stress.Four out of five patients had mild steatosis.
This finding was also seen in an earlier liver biopsy study that involved two participants who had received gene therapy for haemophilia A. 30 It is important to note that both studies had small sample sizes and biopsies were taken long after drug infusion and past the period in which ALT elevations are typically reported.It is possible that hepatocyte damage occurred, and that ER stress was induced but at an earlier point, given these limitations, it is difficult to draw firm conclusions. 28 addition, the latter study found evidence of episomal forms of AAV5-hFVIII-SQ vector DNA which were dose-dependent in the biopsy samples, suggestive of their long-term persistence of the transgene many months post-administration.Some of these forms were incomplete implying a limited ability to produce functional FVIII RNA.Potential mechanisms for the decline in FVIII expression over time were proposed and included the loss of episomal genomes due to hepatocyte turnover or additional cellular metabolism and epigenetic genome silencing. 28e research for haemophilia B, suggests no prolonged liver toxicity in a long-term follow-up study of 15 years but liver biopsies post-gene therapy infusion were not conducted on participants due to ethical considerations. 31Liver biopsy samples closer to the time of drug infusion and during the period of ALT elevation may provide clearer answers.However, the practicalities and risks involved in frequent liver biopsy sampling in patients with known bleeding disorders may raise concerns.Percutaneous liver biopsies are associated with a higher risk of bleeding complications. 32A trans-jugular approach may be a safer option in comparison although it is not entirely without risks.Perhaps, it may be justified during the early period of transaminitis that occurs following gene therapy infusion.
The decision to perform liver biopsies should be weighed carefully, considering the potential benefits and risks, as well as the individual patient's circumstances and health status.

| Identifying high-risk patients in gene therapy implementation
Identifying patients at high risk of liver injury in the context of gene therapy implementation will involve a multifaceted approach.As with most haemophilia trials, monitoring essential liver parameters (ALT, AST, GGT, ALP, bilirubin, albumin and prothrombin time) is useful as a screening tool for liver disease. 7 K18 with some success. 33By adhering to these methods and parameters, healthcare providers can effectively pinpoint high-risk patients, ensuring safer gene therapy practices and reducing potential complications during and after treatment.
Screening for pre-existing AAV neutralising antibodies (NAbs) has been common practice in trials.And while they have been implicated in the liver injury commonly seen, the screening process during these trials is to ensure that transfer of gene therapy can occur without an adaptive immune response. 21There is no role for screening patients for NAbs in clinical practice as gene therapy for haemophilia is not yet readily available.However, for trial recruitment exclusion is based on antibody titre threshold.In a trial for haemophilia B patients, NAbs at low titres did not prevent the successful transduction of gene therapy with Etranacogene Dezaparvovec. 10There seemed to be no predictive power to NAb titres and elevations in ALT for this trial, suggesting that NAbs titres at the level seen in the general population perhaps should not be a reason to exclude individuals from receiving gene therapy.

| Implications for current therapy
At present, research in AAV gene therapy for haemophilia has found no severe DILI based on Hy's Law and the available biopsy data, despite its limitations, suggests no significant long-term liver disease. 28,31The current knowledge regarding ALT elevations does however highlight a significant need for protection of the transgene, especially as expression of FVIII and FIX is the key measure of success in these trials. 2,7Immunosuppression with corticosteroids is commonly administered to mitigate the loss of the transgene during transaminitis, but these drugs have a wide side effect profile which should perhaps encourage cautionary rather than uniform usage. 25The use of immunosuppression for ALT elevations remains controversial and is discussed by [REFERENCE TO TUDDENHAM

CHAPTER TO BE PROVIDED HERE AT PROOF STAGE] in this
supplement.
There is a current gap in our understanding of the mechanisms of transaminitis and loss of gene expression in AAV mediated gene therapy for haemophilia.There is limited liver biopsy data within the relevant period to provide information on the architecture and composition of the liver tissue and the inflammatory process.The specific patient or gene therapy factors that determine individuals who have liver toxicity and those who do not needs to be elucidated, alongside characteristics that determine loss of transgene expression, with or without liver toxicity.More intriguingly, transaminitis occurs more often with haemophilia A than B and there is a loss of transgene expression over time with occasional re-occurrences of transaminitis. 23,24The reasons for this require further exploration and may have significant implications for the success of gene therapy in haemophilia A.

| Implications for the future & other considerations
The research landscape for gene therapy in haemophilia is exploring ways to reduce immunogenicity, hepato-toxicity and loss of transgene expression.The use of AAV vectors with CpG constructs removed from non-coding parts of the expression cassette to prevent immune responses to the transgene continues to be explored. 2,34The development of less immunogenic AAV vectors has some promising results 35 and the administration of lower doses of gene therapy has also proven to be less toxic. 36at remains to be seen relates to known unknowns.Can we be confident, based on current liver biopsy data, that there are no long-term effects on the liver?The data currently available is not robust but information to-date provides a high level of reassurance.
Can the stringent inclusion criteria be revisited to include haemophilia patients with liver diseases and HIV patients on potentially hepatotoxic medications?The cases of severe DILI seen in SMA and XLMTM patients imply underlying liver disease may contribute to the development of hepatic injury therefore, the risks of inclusion for gene therapy need to be carefully considered on a case by case basis. 8,37However, given the low doses of vector used in people with haemophilia and the current safety profile it seems probable that, with appropriate monitoring, such patients could be considered for gene therapy, albeit appropriate counselling and monitoring will be required.Despite the unknowns in gene therapy for people with haemophilia, it is clear that these therapeutics provide significant benefits to patients and the data to-date indicates that the relatively mild, transient liver dysfunction that is often seen does not lead to long-term sequelae and can be managed satisfactorily by current immunosuppressive approaches.

CO N FLI C T O F I NTE R E S T S TATE M E NT
Professor Foster has received consultancy fees from Abbvie, Biomarin, CSL Behring, Gilead, GSK and Pfizer.

4 Life threatening consequences; urgent intervention required 5
or mild symptoms, clinical and diagnostic symptoms only, no intervention required 2 Moderate; minimal, local or non-invasive intervention indicated, limiting instrumental ADLs a 3 Severe or medically significant but not life-threatening, hospitalisation or prolonged hospitalisation indicated; disabling and limiting self-care ADLs b Death related to adverse event a Instrumental ADLs refer to meal preparation, shopping, groceries and laundry.