You are here

  • Home
  • Online First
  • Systemic messenger RNA replacement therapy is effective in a novel clinically relevant model of acute intermittent porphyria developed in non-human primates

Article Text

Article menu
Hepatology
Original research
Systemic messenger RNA replacement therapy is effective in a novel clinically relevant model of acute intermittent porphyria developed in non-human primates
  1. Karol M Córdoba1,
  2. Daniel Jericó1,
  3. Lei Jiang2,
  4. María Collantes3,4,
  5. Manuel Alegre4,5,
  6. Leyre García-Ruiz4,6,
  7. Oscar Manzanilla4,5,
  8. Ana Sampedro1,
  9. Jose M Herranz7,8,
  10. Iñigo Insausti6,
  11. Antonio Martinez de la Cuesta6,
  12. Francesco Urigo1,
  13. Patricia Alcaide9,
  14. María Morán10,11,
  15. Miguel A Martín10,11,
  16. José Luis Lanciego12,13,
  17. Thibaud Lefebvre14,
  18. Laurent Gouya14,
  19. Gemma Quincoces3,4,
  20. Carmen Unzu15,
  21. Sandra Hervas-Stubbs4,16,17,
  22. Juan M Falcón-Pérez7,18,
  23. Estíbaliz Alegre4,19,
  24. Azucena Aldaz20,
  25. María A Fernández-Seara4,6,
  26. Iván Peñuelas3,4,
  27. Pedro Berraondo4,16,17,21,
  28. Paolo G V Martini2,
  29. Matias A Avila4,7,8,21,
  30. Antonio Fontanellas1,4,7,21
  1. 1Hepatology: Porphyrias & Carcinogenesis Lab. Solid Tumors Program, CIMA Universidad de Navarra, Pamplona, Spain
  2. 2Moderna Inc, Cambridge, Massachusetts, USA
  3. 3Translational Molecular Imaging Unit (UNIMTRA), and Nuclear Medicine-Department, Clínica Universidad de Navarra (CUN), University of Navarra, Pamplona, Spain
  4. 4Instituto de Investigación Sanitaria de Navarra (IdiSNA), Pamplona, Spain
  5. 5Department of Clinical Neurophysiology, Clinica Universitaria de Navarra, Pamplona, Spain
  6. 6Radiology Department, Clinica Universitaria de Navarra, Pamplona, Spain
  7. 7Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBEREHD), Instituto de Salud Carlos III, Madrid, Spain
  8. 8Hepatology Laboratory, Solid Tumors Program, CIMA-University of Navarra, Pamplona, Spain
  9. 9Centro de Diagnóstico de Enfermedades Moleculares, Universidad Autónoma de Madrid, Madrid, Spain
  10. 10Mitochondrial Diseases Laboratory, Instituto de Investigación Sanitaria Hospital 12 de Octubre (imas12), 12 de Octubre University Hospital, Madrid, Spain
  11. 11Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER), Instituto de Salud Carlos III, Madrid, Spain
  12. 12Neurosciences Department, CIMA Universidad de Navarra, Pamplona, Spain
  13. 13Centro de Investigación Biomédica en Red de Enfermedades Neurodegenerativas (CIBERNED), Instituto de Salud Carlos III, Madrid, Spain
  14. 14APHP. Nord-Université de Paris Cité, Centre Français des Porphyries, Hôpital Louis Mourier, Paris, France
  15. 15Gene Therapy and Regulation of Gene Expression Program, CIMA Universidad de Navarra, Pamplona, Spain
  16. 16Program of Immunology and Immunotherapy, CIMA-University of Navarra, Pamplona, Spain
  17. 17Centro de Investigación Biomédica en Red de Enfermedades Oncológicas (CIBERONC), Instituto de Salud Carlos III, Madrid, Spain
  18. 18Exosomes Lab. & Metabolomics Platform. Center for Cooperative Research in Biosciences (CIC bioGUNE), Bizkaia Technology Park, Derio, Spain
  19. 19Service of Biochemistry, Clinica Universitaria de Navarra, Pamplona, Spain
  20. 20Pharmacokinetics Division, Pharmacy Departement, Clínica Universidad de Navarra (CUN), Pamplona, Spain
  21. 21Cancer Center Clínica Universidad de Navarra (CCUN), Pamplona, Spain
  1. Correspondence to Professor Matias A Avila; maavila{at}unav.es; Dr Paolo G V Martini; Paolo.Martini{at}modernatx.com; Dr Antonio Fontanellas; afontanellas{at}unav.es

Abstract

Objective Acute intermittent porphyria (AIP) is a rare metabolic disorder caused by haploinsufficiency of hepatic porphobilinogen deaminase (PBGD), the third enzyme of the heme biosynthesis. Individuals with AIP experience neurovisceral attacks closely associated with hepatic overproduction of potentially neurotoxic heme precursors.

Design We replicated AIP in non-human primates (NHPs) through selective knockdown of the hepatic PBGD gene and evaluated the safety and therapeutic efficacy of human PBGD (hPBGD) mRNA rescue.

Results Intrahepatic administration of a recombinant adeno-associated viral vector containing short hairpin RNA against endogenous PBGD mRNA resulted in sustained PBGD activity inhibition in liver tissue for up to 7 months postinjection. The administration of porphyrinogenic drugs to NHPs induced hepatic heme synthesis, elevated urinary porphyrin precursors and reproduced acute attack symptoms in patients with AIP, including pain, motor disturbances and increased brain GABAergic activity. The model also recapitulated functional anomalies associated with AIP, such as reduced brain perfusion and cerebral glucose uptake, disturbances in hepatic TCA cycle, one-carbon metabolism, drug biotransformation, lipidomic profile and abnormal mitochondrial respiratory chain activity. Additionally, repeated systemic administrations of hPBGD mRNA in this AIP NHP model restored hepatic PBGD levels and activity, providing successful protection against acute attacks, metabolic changes in the liver and CNS disturbances. This approach demonstrated better efficacy than the current standards of care for AIP.

Conclusion This novel model significantly expands our understanding of AIP at the molecular, biochemical and clinical levels and confirms the safety and translatability of multiple systemic administration of hPBGD mRNA as a potential aetiological AIP treatment.

  • ENERGY METABOLISM
  • DRUG METABOLISM
  • GENE THERAPY
  • HEPATOBILIARY DISEASE

Data availability statement

Data are available on reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/gutjnl-2024-332619

Statistics from Altmetric.com

    Request Permissions

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

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • In rare metabolic diseases, the high phenotypic variability makes it challenging to gain a comprehensive understanding of the disease’s pathogenesis and natural history.

    • Developing clinically relevant large animal models is crucial for assessing the safety and translatability of new drugs just prior to clinical trials.

    • Given the limited availability of eligible patients for clinical trial recruitment, this information is highly relevant for rare disorders. It allows for better clinical trial design and endpoints assessment, based on homogeneous biomarkers indicating both disease severity and rescue status.

    WHAT THIS STUDY ADDS

    • Three advanced technologies were applied to develop a novel model of metabolic disease in non-human primates (NHPs). Specifically, a recombinant adeno-associated viral vector with an inducible promoter is used to express RNA interference, targeting a liver enzyme associated with acute intermittent porphyria (AIP).

    • The accuracy of this model allowed us to validate and expand our understanding of the transcriptomic, metabolomic, lipidomic and clinical features associated with the pathogenesis of AIP.

    • This insight is particularly relevant for rare disorders like AIP, where multiple metabolic pathways are affected, and the mechanisms of neurological involvement remain not fully understood.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • This study marks the first instance of a genetic metabolic disease model developed in NHPs by means of RNA interference technology.

    • Our findings validate the safety and translatability of multiple systemic administrations of human porphobilinogen deaminase mRNA as a potential aetiological treatment for AIP, restoring the normal phenotype and addressing metabolic disturbances in the liver and in the brain.

    Introduction

    Acute intermittent porphyria (AIP) is a rare autosomal dominant disorder caused by loss-of-function mutations leading to a decreased hepatic activity of porphobilinogen deaminase (PBGD, EC 2.5.1.61), the third enzyme of the heme synthesis pathway. AIP typically presents abdominal-psychoneurological acute attacks, characterised by abdominal pain often accompanied by nausea, vomiting and/or constipation, along with hypertension and/or tachycardia and peripheral neuropathy. Mental disturbances such as anxiety, depression or a confused state may also occur.1–4

    Both endogenous and exogenous factors triggering acute porphyria attack are associated with the upregulation of the first and rate-limiting enzyme of the hepatic heme biosynthesis, the aminolaevulinic acid (ALA) synthase (ALAS), leading to the accumulation of porphyrin precursors ALA and porphobilinogen (PBG).4 However, many aspects of AIP pathobiology remain poorly understood, such as interference with other metabolic pathways as TCA cycle and one-carbon metabolism (OCM), functional anomalies in mitochondrial respiratory chain complexes, reduced cerebral glucose uptake and brain perfusion, or even alterations in the neural GABAergic activity.

    In rare metabolic diseases like AIP, the high phenotypic variability makes it challenging to gain a comprehensive understanding of the pathogenesis of the disease. Experimental models are critical for advancing insights into disease mechanisms and assessing responses to emerging treatments.5 Developing clinically relevant large animal models is crucial for assessing new drug safety, gathering efficacy, identifying intervention windows and obtaining translatability data just before clinical trials. This information is highly pertinent for rare disorders such as AIP, where multiple metabolic pathways are affected, and neurological involvement remains unknown or not fully understood. Furthermore, it would be advantageous to directly link observed changes with the downregulation of a particular gene, eliminating its potential impact during embryogenesis or postnatal development.

    Advances in RNA interference (RNAi) technology now enable the partial and specific silencing of a particular gene, facilitating the attempt to replicate a disease in large animals. Thus, this work aimed to develop and characterise a new model of AIP in non-human primates (NHPs) through sustained specific inhibition of endogenous hepatic PBGD expression by intrahepatic administration of a recombinant adeno-associated viral (rAAV) vector containing a short hairpin RNA (shRNA) against PBGD mRNA of Macaca fascicularis. The efficacy and safety profile of recurrent administration of both conventional and an emerging therapy based on systemic messenger RNA therapy were also evaluated.

    Materials and methods

    Production of rAAV vector and lipid nanoparticles (LNPs) containing modified mRNAs encoding human PBGD

    The shRNA335 directed against the target mRNA was designed from conserved regions found in transcript variants of Macaca fascicularis hydroxymethylbilane synthase (HMBS) sequences to establish rAAV2/8-maPBGD shRNA. The shRNA is under the control of the liver-specific human α−1-antitrypsin promoter with regulatory regions from the human albumin enhancer (EalbAAT promoter)6 and two ALAS drug-responsive sequences (ADRES) upstream of the promoter.7 Production of LNPs containing modified mRNAs encoding human PBGD (hPBGD) was performed as previously described.8 Further sequence details are available in online supplemental materials and methods.

    Supplemental material

    Porphyria model in NHPs

    Female and male juvenile cynomolgus NHPs (Macaca fascicularis) received intrahepatic administration of an rAAV8-maPBGD shRNA (4.18×1013 gc/kg BW)(figure 1A). To replicate acute attacks, NHPs were administered with porphyrinogenic drugs, starting 4 weeks after the rAAV injection. NHPs were subjected to a 3-day regimen of increasing doses of phenobarbital (35, 45 and 50 mg/kg; i.m.). Only for CNS studies, this protocol involved a single dose of phenobarbital (45 mg/kg, i.m.) 2 days before and two i.m. doses of 400 mg, 8 hours apart, of Soltrim (Trimethoprim-Sulfamethoxazole; Almirall S.A., Barcelona, Spain) 1 day before studies. To evaluate the therapeutic efficacy of current and emerging therapies, AIP NHPs received multiple doses of givosiran (n=1, male), hemin (n=1, male) or hPBGD mRNA8 (n=3, 1 male) encapsulated in LNPs from month 3 to the end of the study (7 months after rAAV administration). Full methods are provided as online supplemental materials. All procedures were conducted in accordance with the European Council guidelines (protocol CEEA142-16). We used the ARRIVE reporting guidelines: Percie du Sert N The ARRIVE Guidelines 2.0: updated guidelines for reporting animal research.

    Figure 1
    Figure 1

    Schematic timeline representation of the study and biochemical status and gene expression of heme synthesis pathway. (A) Schematic timeline representation of the characterisation of the AIP model in NHPs. The rAAV8 vector was administered directly via the hepatic blood flow at a dosage of 4.18×1013 gc/kg BW. To replicate acute attacks, NHPs were administered with porphyrinogenic drugs, starting 4 weeks after the rAAV injection. Given that acute barbiturate challenge can transiently alter brain perfusion,44 affect the insulin signalling pathway in the brain,45 and enhance inhibitory GABAergic neurotransmission,46 a separate protocol was employed before positron-emission tomography (PET) and cerebral blood flow (CBF) studies. One male AIP NHP received a total of 5 doses of givosiran (2.5 mg/kg, administered s.c. every three 3 weeks; Alnylam Pharmaceuticals, Cambridge, Massachusetts, USA). Another male AIP NHP was treated with 19 doses of hemin (2 mg/kg intravenous Normosang, Orphan Europe, Recordati, Milan, Italy), over 6 cycles, with three doseases in each. The cycles were administered every 2-weeks, and an additional dose was given 1 week before sacrifice. Three AIP NHPs (one male) received a total of 7–8 doses of mRNA, with one dose of 0.5 mg/kg of hPBGD mRNA administered every 2 weeks. Lastly, three other AIP NHPs did not receive any treatment. A comprehensive description of the methods used is provided in online supplemental materials and methods. Mean peak urinary excretion of (B) ALA and (C) PBG after repeated acute attacks and multidose administration of current and emerging therapies. (D) Hepatic expression of ALAS1 in NHPs with AIP after treatment (three measurements per animal). (E) ALAS1 immunoblot in liver samples from WT NHPs and NHPs with AIP after recurrent acute attacks and repeated administrations of treatments. (F) Hepatic PBGD activity of NHPs with AIP after treatment (determined in three different lobes per animal). (G) Hepatic expression of HO1 in NHPs with AIP after treatment (three measurements per animal). Data were analysed using one-way ANOVA. Pairwise comparisons were made using Bonferroni’s multiple comparison tests. Data are expressed as mean±SD. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 vs WT, untreated or indicated groups. AIP, acute intermittent porphyria; ANOVA, analysis of variance; rAAV, recombinant adeno-associated viral; NHPs, non-human primates; PBGD, porphobilinogen deaminase.

    Results

    Assessment of the in vitro and in vivo efficacy and specificity of maPBGD shRNAs

    The maPBGD shRNA335 induced partial inhibition in the PBGD activity when transfected into primary fibroblasts obtained from corneal epithelial biopsy of Macaca fascicularis (online supplemental figure 1A). Furthermore, overexpression of shRNA335 either from a plasmid vector (online supplemental figure 1B) or using rAAV8 (online supplemental figure 1C) did not disrupt hepatic expression of human PBGD (hPBGD) in AIP mice after systemic administration of 0.5 mg/kg of hPBGD mRNA formulated in LNPs.8

    Supplemental material

    Three female NHPs were sacrificed at 1, 2 and 6 months post-rAAV administration. Body weight, liver function and haematological parameters were assessed weekly for the initial 3 weeks (online supplemental table 1). No signs of toxicity were observed. At sacrifice, specific detection of viral sequences in DNA samples revealed efficient hepatocyte transduction (online supplemental figure 1D). No differences in vector genome copies were observed 1 or 2 months after rAAV administration. However, vector genome levels were markedly diminished in the liver of the animal sacrificed 6 months postinjection. Transgene loss following rAAV vector administration in juvenile NHPs has been previously described.9 To counteract the progressive decline in shRNA expression, we incorporated ADRES sequences upstream of the vector promoter.7 These ADRES elements, described in the 5’-flanking regions of human ALAS1 gene, allow recurrent challenges with porphyrinogenic drugs to both induce hepatic ALAS1, triggering acute attacks and also to stimulate transgene expression in our model. Following recurrent challenges with porphyrinogenic drugs (figure 1A), ALAS1 expression increased in the NHP livers sacrificed at 2 and 6 months (online supplemental figure 1E). Additionally, PBGD activity remained inhibited at levels consistent with AIP throughout the study. Initially, the NHP livers exhibited PBGD levels around 45% of normal values, increasing to 65% after 6 months (online supplemental figure 1F). Gene expression measured through real-time qPCR (online supplemental figure 1G) and protein levels assessed by PBGD immunoblot (online supplemental figure 1H) were indicative of significant PBGD inhibition across all hepatic lobes.

    Supplemental material

    Monitoring well-being, safety of multidose administration of treatments and heme biosynthesis pathway during the study

    A total of four female and four male NHPs were included in this study (figure 1A). Among them, AIP NHPs received multiple doses of hPBGD mRNA (n=3), givosiran (n=1) or hemin (n=1) from month three to the end of the study. BW and serum liver function tests remained within the normal range throughout the duration of the study (online supplemental table 2). Coagulation parameters, particularly serum fibrinogen (FIB) levels, temporarily increased just after rAAV administration but were unaffected by repeated challenges with porphyrinogenic drugs or treatments (online supplemental table 2). Red blood cells, haematocrit and haemoglobin showed a gradual decline following AIP induction but cannot be directly attributed to the surgical procedure, as three animals maintained stable haematological parameters in the initial 16 days but began to reduce haematocrit and haemoglobin levels weeks after surgery. These parameters tended to stabilise during recurrent acute attacks and subsequent treatments, except in the animal treated with givosiran (online supplemental table 2). No changes were noted in total white cell count or their major subcategories, ruling out infectious processes or allergic reactions during the study (online supplemental table 2). No infusion reactions at the injection site were observed after multidose (MD) administration of treatments and histological analyses showed no alterations in liver tissue (data not shown). Finally, no significant differences were observed in serum sodium levels (147.5±5.7 in controls vs 147.0±4.6 mEq/L in AIP NHPs; p=0.753).

    Supplemental material

    Sustained urinary heme precursor overexcretion was observed throughout the study, particularly associated with acute attacks (figure 1B,C, untreated group). Enhanced production and significant serum accumulation of ALA precursor were detected 48 hours after the last dose of the porphyrinogenic drugs (online supplemental figure 1I). Consistent with these data and the recurrence of acute attacks, ALAS1 expression (fold change ~3; figure 1D) and protein levels (fold change ~12; figure 1E) in untreated animals were significantly induced at sacrifice, 1 week after the final drug challenge. Following three cycles of 3-week drug challenges, MD administration of therapies commenced in the third month and continued until sacrifice 7 months post-rAAV injection (figure 1A). In untreated AIP NHPs or those treated with current treatment regimens (hemin and givosiran), hepatic PBGD exhibited 71.2% of normal activity compared with WT animals. MD administration of hPBGD mRNA led to increased PBGD protein levels, as demonstrated by enzymatic activity (figure 1F) and Western blot analysis of liver tissues (online supplemental figure 2A). This activity surge was 2.1-fold higher than that observed in AIP NHPs and 1.4-fold higher compared with WT animals. Importantly, this augmentation persisted even after the administration of up to 8 doses of mRNA, suggesting an absence of an immune response against hPBGD or the vehicle. The lack of a cellular immune response targeting hPBGD or LNPs was further validated through IFN-gamma ELISpot assays conducted on peripheral blood lymphocytes (online supplemental figure 2B).

    Supplemental material

    Treatment with hPBGD mRNA or current therapies effectively protected against ALAS1 mRNA and protein overexpression (figure 1D,E). Consequently, urinary heme precursor accumulation significantly decreased in treated animals (figure 1B,C). These results confirm treatments’ efficacy in reversing drug-induced upregulation of heme synthesis. Hepatic heme-oxygenase-1 mRNA showed no overexpression in AIP NHPs, except in the hemin-treated animal (figure 1G).

    At sacrifice, serum and cerebrospinal fluid (CSF) samples were collected from untreated (n=3) and hPBGD mRNA treated (n=3) animals. One week after the last drug dose, serum precursor levels showed preattack values, but there was a clear trend towards PBG accumulation in the CSF of untreated animals (online supplemental figure 1J). These data suggest that hepatic PBG can cross the blood–brain barrier (BBB), with elevated levels persisting longer in CSF than in serum.

    Functional changes in the central and peripheral nervous systems of AIP NHPs

    At baseline, positron emission tomography (PET)/CT imaging revealed a prominent distribution of [11C]flumazenil, a gamma-aminobutyric acid (GABA) antagonist radiotracer, in the brain cortex and hippocampus (figure 2A,B and online supplemental figures 3A and 4A). A significant reduction of brain [11C]flumazenil uptake was observed in AIP NHPs, starting 1 month after AIP induction. Reduced uptake remained consistent following recurrent attacks in untreated animals, while MD administration of hPBGD mRNA uniformly restored baseline uptake across all regions in one animal (figure 2A,B and online supplemental figure 3B). Unfortunately, technical issues with the synthesis equipment hindered the completion of the [11C]flumazenil study in other treated animals.

    Supplemental material

    Supplemental material

    Figure 2
    Figure 2

    Therapeutic efficacy of multidose intravenous administration of hPBGD mRNA against CNS alterations in AIP NHPs. (A) Representative PET/CT images of [11C]flumazenil brain distribution from an untreated NHP (left) with AIP and an NHP injected with hPBGD mRNA (right) before (baseline) and after (porphyria) PBGD inhibition, and at the end of the study after recurrent acute attack and treatment. (B) Quantification of [11C]flumazenil brain uptake in NHPs at different time points described in A. (C) Representative PET/CT images of [18F]FDG brain uptake and (D) quantification from untreated NHPs with AIP and animals treated with hPBGD mRNA, hemin and givosiran. (E) Quantification of CBF and (F) representative MRI CBF maps overlaid over the anatomical images of NHPs at baseline, with induced porphyria and after treatment with hPBGD mRNA, hemin and givosiran. Significant heterogeneity was observed among animals in baseline CBF values; however, no gender difference was noted in the detection of hypoperfusion in the model. Data were analysed using one-way ANOVA. Pairwise comparisons were made using Bonferroni’s multiple comparison tests. Data are expressed as mean±SD. *p<0.05; **p<0.01 vs WT, untreated or indicated groups. AIP, acute intermittent porphyria; ANOVA, analysis of variance; CBF, cerebral blood flow; CNS, central nervous system; hPBGD, human porphobilinogen deaminase; NHPs, non-human primates; PET, positron emission tomography.

    Region-specific analysis of brain glucose uptake, assessed through PET/CT imaging with the glucose analogue [18F]FDG, revealed the highest radiotracer uptake in the prefrontal motor cortex and subcortical ganglia at the brain’s base, associated with movement initiation and planning (online supplemental figures 4B and 5). Porphyria induction in NHPs led to a notable reduction in the radiotracer signal, declining further with repeated attacks (figure 2C,D). Of note, treatment with hPBGD mRNA significantly improved brain [18F]FDG uptake and restored baseline levels (figure 2C,D and online supplemental figures 4B and 5).

    Supplemental material

    MRIs and quantification indicated a reduction in cerebral blood flow (CBF) in AIP NHPs starting 1 month post-rAAV injection, indicative of the porphyria state (figure 2E,F). Remarkably, hypoperfusion persisted during acute attacks, with hPBGD mRNA treatment being the sole intervention capable of restoring baseline CBF levels (figure 2E).

    After recurrent attacks, untreated AIP NHPs exhibited slowed movements, impaired coordination and difficulty in swallowing (data are not shown). These behavioural changes were not observed after current and emerging treatments. Electrophysiological studies in AIP NHPs revealed a decreased amplitude of sensory action potentials in the median, ulnar and plantar nerves (figure 3A). Treatment with hPBGD mRNA prevented amplitude reduction in the ulnar and plantar nerves (figure 3A). In contrast, nerve conduction velocity remained comparable to baseline values (figure 3B). Likely, the AIP NHPs model, with moderate porphyrin precursor accumulation, may necessitate more than 7 months to reveal significant reductions in conduction velocity.

    Figure 3
    Figure 3

    Therapeutic efficacy of multidose intravenous administration of hPBGD mRNA against peripheral nervous system in AIP NHPs. (A) Amplitude of median, ulnar and plantar sensitive nerve, and (B) velocity of nerve conduction in electrophysiological studies performed after repeated acute attacks and the administration of repeated doses of hPBGD mRNA, hemin and givosiran. Comparisons between two groups were analysed by paired t-tests. Data are expressed as the percentage respect to baseline values (mean±SD). *p<0.05; **p<0.01 vs baseline data. AIP, acute intermittent porphyria; hPBGD, human porphobilinogen deaminase; NHPs, non-human primates.

    Characterisation of the hepatic transcriptome in AIP NHPs

    Out of the 10 760 different mRNAs detected, 4805 (45%) exhibited statistically significant differences (adjusted p<0.05) when comparing WT (n=15, 7 males) and AIP NHPs experiencing recurrent porphyria attacks. A primary discovery from our transcriptomic analyses in AIP NHPs was the remarkable alteration in the expression of key hepatic genes involved in metabolic homoeostasis. This included transcription factors NR1H4 (FXR) and HNF4α, the translation regulator BZW1, and the secreted proteins FST and IGFBP1, all showing impaired expression in AIP NHPs Notably, treatment with hPBGD mRNA restored the expression of these and other relevant genes (figure 4A). Metabolic processes involving lipid homoeostasis and fatty acids oxidation (log FDR adjusted p value~3.5), xenobiotic response (~1.7), as well as carbohydrate metabolism (~2), and energy production (~1.8) were identified as the functions showing the most pronounced differential expression patterns. The number of genes and whether they are deregulated in those specific metabolic pathways is shown in online supplemental figure 6. The expression of about 76% of these genes was restored following treatment with both emerging, hPBGD mRNA and current therapies for AIP.

    Supplemental material

    Figure 4
    Figure 4

    Transcriptomic analysis of liver-specific genes and therapeutic efficacy of multidose intravenous administration of hPBGD mRNA against metabolomic alterations on the TCA cycle. (A) HeatMap illustrating the expression levels of key hepatic genes critical for metabolic homoeostasis. Transcriptomic analyses were conducted on liver samples of 15 age-matched WT NHPs (including 7 males), and from 3 lobes of three AIP NHPs. Additionally, three liver lobes from each of the three AIP NHPs receiving MD of hPBGD mRNA, three liver lobes from one male NHP who underwent MD hemin administration and another male injected with givosiran were also included. RNA-Seq Data Analysis was performed with a Benjamini-Hochberg correction to get log-adjusted p values, without the loss of precision from undoing and redoing the log-transformations. (B–G) Metabolomic analysis reveals notable differences in TCA metabolites between AIP NHP livers and the impact of treatments. Data were analysed using one-way ANOVA. Pairwise comparisons were made using Bonferroni’s multiple comparison tests. Data correspond to three hepatic lobes from each AIP NHP and are expressed as mean±SD *p<0.05; **p<0.01; ***p<0.001 vs WT animals or indicated groups. AIP, acute intermittent porphyria; ANOVA, analysis of variance; hPBGD, human porphobilinogen deaminase; MD, multidose; NHPs, non-human primates.

    Figure 5
    Figure 5

    Metabolomic study focusing on the OCM and TCA cycle. (A) HeatMap illustrating levels of key hepatic metabolites for OCM and TCA cycles. Analyses were conducted on liver samples of 7 age-matched WT NHPs (including 4 males) and from 3 lobes of three AIP NHPs. Additionally, three liver lobes from each of the three AIP NHPs receiving MD of hPBGD mRNA, three liver lobes from one male NHP who underwent MD hemin administration and another male injected with givosiran were also included. (B-J) Metabolomic analysis reveals notable differences in OCM metabolites between AIP NHP livers and the impact of treatments. Data were analysed using one-way ANOVA. Pairwise comparisons were made using Bonferroni’s multiple comparison tests. Data correspond to three hepatic lobes from each AIP NHP and are expressed as mean±SD *p<0.05; **p<0.01; ***p<0.001 **** p<0.0001 vs WT animals or indicated groups. AIP, acute intermittent porphyria; GPC, glycerophosphocholine; hPBGD, human porphobilinogen deaminase; MD, multidose; NHPs, non-human primates; OCM, one-carbon metabolism; SAMe, S-adenosylmethionine; SAH, S-adenosylhomocysteine; dcSAMe, decarboxylated S-adenosylmethionine; MTA, 5′-deoxy-5′-methylthioadenosine.

    Altered expression on genes encoding molecules regulating energy metabolism (PPARGC1A), transmembrane glucose transport (SLC2A3) or cellular processes under insulin control (IRS1 and IRS2) exhibited altered expression (online supplemental figure 7A). AIP NHPs also showed upregulation on lipid uptake, beta-oxidation, triglycerides (TGs) synthesis and lipid droplet lipolysis (online supplemental figure 7B). Moreover, the genes encoding the TCA cycle also exhibited altered expression in AIP NHPs (online supplemental figure 8A). Considering that the TCA cycle oxidises acetyl-CoA derived from carbohydrates or fatty acids to produce NADH and FADH2, these data suggest disrupted energy metabolism in AIP NHPs. Importantly, treatment with hPBGD mRNA restored the expression of these and other relevant genes, a result not consistently observed in animals treated with current therapies.

    Supplemental material

    Supplemental material

    Heatmaps summarising expression patterns on OCM, mainly folate, polyamine metabolism and glutathione synthesis pathways (online supplemental figure 8B), also showed genes differentially expressed between WT and AIP NHPs. Remarkably, hPBGD mRNA therapy showed the highest rate of restoration in the expression of altered genes. Finally, the expression levels of genes coding for essential hemoproteins (such as CAT, HBB, IDO, TDO2, NOS1 or NOX1) and the essential cofactor of CYP450 enzymes were significantly altered in AIP NHPs before treatment, indicating changes in xenobiotic and detoxification processes (online supplemental figure 9).

    Supplemental material

    Liver metabolism in the AIP NHP model

    To further evaluate the functional impact of these transcriptional alterations, we performed metabolomic and lipidomic studies using liver and serum samples from the AIP NHP model. Within the TCA cycle (figures 4B-G and 5A), a reduction in succinate (figure 4B) and malate (figure 4C) levels indicated a decrease in TCA cycle intermediates downstream of succinyl-CoA, but not in citrate levels (figure 4D), which initiate the cycle. Furthermore, a decline in hepatic glucose (figure 4E) and glutamine (figure 4F) concentrations was observed alongside elevated levels of glutamate (figure 4G), which contribute metabolites to the TCA cycle via citrate and α-ketoglutarate (α-KG), respectively. Remarkably, MD administration of hPBGD mRNA restored hepatic succinate and malate levels, a result not consistently observed in the animal treated with hemin. Of note, givosiran treatment resulted in hepatic concentrations of glucose, citrate, fumarate and malate well above levels found in control NHPs, suggesting an overactivation of intermediary metabolism (figure 4C,D,E). Quantification of serum metabolites involved in the TCA cycle revealed no significant changes over time (online supplemental figure 10B–F).

    Supplemental material

    Concerning metabolites involved in OCM (figure 5A), we observed a marked depletion in the hepatic levels of s-adenosyl methionine (SAMe) (figure 5B) and methylthioadenosine (MTA) (figure 5C), a product of SAMe decarboxylation generated in the polyamine biosynthetic pathway (online supplemental figure 11).10 Putrescine, the first metabolite in the polyamine pathway, markedly accumulated in the liver of AIP NHPs (figure 5D). However, the higher polyamines spermidine and spermine, which need decarboxylated SAMe for their synthesis, remained unchanged or were reduced (figure 5E,F and online supplemental figure 11). Interestingly, while all three treatments reduced putrescine accumulation (figure 5D), only hPBGD mRNA treatment significantly recovered SAMe and MTA levels (figure 5B,C). Regarding the transsulfuration pathway, the hepatic enzymatic activity of cystathionine beta-synthase (CBS) was not modified in AIP NHPs when compared with WT livers (figure 5G). However, porphyric animals exhibited a notable decrease in vitamin B6 (PLP), which is a crucial cofactor for enzyme activity (figure 5H). Interestingly, metabolomic studies showed no changes in cystathionine levels (product of CBS) (figure 5I) but low levels of glutathione (GSH) (figure 5J), a downstream product of SAMe metabolism in the transsulfuration pathway (online supplemental figure 12). Remarkably, the levels of GSH and total glutathione (GSH +GSSG) were restored in hPBGD mRNA-treated animals (figure 5J and online supplemental figure 12).

    Supplemental material

    Supplemental material

    Considering the significant dysregulation of lipid homoeostasis observed in the transcriptomic studies (online supplemental figure 7B), we characterised the hepatic lipidome in our AIP NHP model. Hepatic lipidomic profiles of AIP NHPs revealed increased levels of TGs (primary form of storage and transport of fatty acids), non-esterified fatty acids and acylcarnitine when compared with control WT animals (online supplemental figure 13). Taken together, these findings demonstrate a marked impairment in the hepatic handling of lipids in AIP NHPs, which may be reflected in the alterations in serum lipids concentration (figure 6). Changes in the expression of genes involved in key metabolic genes may underlie the phenotype of AIP NHPs. For instance, a marked upregulation of diacylglycerol O-acyltransferase 1 (DGAT1) expression, an essential gene in TG synthesis, was observed (online supplemental figure 7B). Concomitantly, a strong induction of carnitine palmitoyltransferase 1 (CPT1) expression was evidenced in AIP NHPs liver tissues. CPT1 is the key enzyme in the mitochondrial oxidation of long-chain fatty acids (online supplemental figure 7B). This aligns with the accumulation in long-chain acylcarnitines, suggesting a compensatory response to the observed buildup of fatty acids in the liver in this model. Recurrent administration of hPBGD mRNA efficiently reverted the accumulation of TGs, free fatty acids and acylcarnitines while recovering the levels of ChoES. Remarkably, among the different treatments, only hPBGD mRNA administration restored TG levels to control values (figure 6E,F), in accordance with its unique ability to downregulate diacylglycerol O-acyltransferase 1 (DGAT1) and carnitine palmitoyltransferase 1 (CPT1) gene expressions, which are essential genes in TG synthesis and the mitochondrial oxidation of long-chain fatty acids, respectively (online supplemental figure 7B).

    Supplemental material

    Figure 6
    Figure 6

    Serum lipidomic analysis of AIP NHPs and therapeutic efficacy of hPBGD mRNA. Serum levels of (A) phosphatidylethanolamines (PE) (36:4) and (B) PE (P-16:0/18:2); (C) Phosphatidylcholines (PC) (38:2) and (D) PC (O-18:0/18:2); (E) Triglycerides (TG) (51:0) and (F) TG(53:1); (G) Cholesterol esters (ChoE)(17:1) and (H) Sphingolipids (SM (43:1)) and (I) SM (d18:1/25:0). Data were analysed using one-way ANOVA. Pairwise comparisons were made using Bonferroni’s multiple comparison tests. Data correspond to different serum samples taken over two last months of the study from each AIP NHP and are expressed as mean±SD. *p<0.05; **p<0.01; ***p<0.001 vs WT animals or indicated groups. AIP, acute intermittent porphyria; ANOVA, analysis of variance; hPBGD, human porphobilinogen deaminase; NHPs, non-human primates.

    To comprehensively characterise hepatic metabolic alterations associated with the disease, we evaluated the pharmacokinetics of caffeine and acetaminophen, substrates of CYP1A2 and CYP3A4 cytochromes, respectively, in our model. The area under the curve, reflecting the serum concentration of caffeine (figure 7A,B left) and acetaminophen (figure 7A,B right) over time, increased in the animals following porphyria induction. Indeed, the elevated Cmax, Tmax, and t1/2 values suggested a diminished metabolising capacity in AIP NHPs (online supplemental table 3A,B). While all treatments improved drug metabolism, the group treated with hPBGD mRNA showed a metabolisation curve that closely resembled that of control baseline (figure 7A,B, online supplemental table 3A).

    Supplemental material

    Figure 7
    Figure 7

    Therapeutic efficacy of multidose intravenous administration of hPBGD mRNA against some haemoproteins function. We investigated potential alterations in drug metabolism following intragastric administration of a bolus of caffeine and acetaminophen. (A) Pharmacokinetic and (B) area under the curve quantification of caffeine (left) and acetaminophen (right) in the serum of untreated and treated animals. Data correspond to different serum samples taken over two last months of the study from each AIP NHP and are expressed as mean±SD. (C) Cumulative activity of the mitochondrial respiratory complexes I–IV. Given that Complex II-driven respiration is activated by succinyl-CoA, enzymatic activities were measured at the maximum rate with saturated substrates, ensuring independence from the contribution of the TCA cycle. (D) Immunodetection of mitochondrial complex as showed by Western-Blot from liver samples. Data correspond to three hepatic lobes from each AIP NHP and are expressed as mean±SD. Data were analysed using one-way ANOVA. Pairwise comparisons were made using Bonferroni’s multiple comparison tests. *p<0.05; **p<0.01; vs WT animals or indicated groups. AIP, acute intermittent porphyria; ANOVA, analysis of variance; hPBGD, human porphobilinogen deaminase; NHP, non-human primate.

    Finally, the activity of mitochondrial oxidative phosphorylation complexes was analysed in isolated liver mitochondria from the NHP model (figure 7C). The data indicate a significant decrease in mitochondrial complex I activity in AIP NHPs. Complex II and IV also showed a tendency towards reduced activity, with this reduction reaching significance in the givosiran-treated animal (figure 7C). Notably, MD administration of hPBGD mRNA successfully restored their activities to baseline levels (figure 7C). Immunoblot analysis confirmed reduced protein levels of complex I, except on those animals treated receiving MD administration of hPBGD mRNA (figure 7D).

    Discussion

    We described a clinically relevant model of AIP developed in NHPs using RNAi technology to downregulate PBGD gene expression in a liver-targeted manner. Continuous delivery of maPBGD-shRNA was achieved through intrahepatic administration of a high dose of an rAAV vector expressing the interfering sequence under the control of a robust hepatic promoter inducible with porphyrinogenic drugs. Consequently, the activity of the PBGD enzyme was consistently reduced for up to 7 months, reaching levels comparable to those observed in individuals with AIP.

    In this AIP NHP model, hepatic PBGD inhibition led to a modest increase in urinary ALA and PBG excretion 1 week after rAAV administration. Challenge with porphyrinogenic drugs markedly induced hepatic ALAS1, leading to increased accumulation of porphyrin precursors. Although its excretion was not as pronounced as observed in the AIP mouse,8 the NHP model replicates most typical clinical features, including pain and motor impairment. Online supplemental table 4 summarises altered functions described in patients with AIP and reproduced in various experimental models. Remarkably, this model provides new insights into the activation of the GABAA receptor in the cerebral cortex, the most prominent receptor for inhibitory neurotransmitters in the CNS. GABA binding triggers changes in the receptor’s conformation, allowing chloride ions to flow through, stabilising the neuronal resting potential and blocking excitatory signals.11 Since the molecular structure of ALA resembles GABA, it has been hypothesised that overproduced ALA might activate GABAA receptor, potentially causing an imbalance of excitation-inhibition within the CNS of patients.12 13

    Supplemental material

    The GABA agonist [11C]flumazenil gives a surrogate index for assessing the density and status of GABAA receptor in the animal’s brain. PET/CT imaging revealed a prominent distribution of the GABA agonist in the brain cortex of NHPs, including the hippocampus, because of its cortical nature. Reduced radiotracer brain uptake was observed during the induction of porphyria and the onset of recurrent acute attacks. Importantly, [11C]flumazenil uptake returned to baseline after MD administration of hPBGD mRNA, indicating restored GABAergic activity. As our model specifically modifies PBGD expression only in the liver, our data provide the first confirmation that hepatic porphyrin precursors may act as GABA agonists in the CNS, even at low levels due to limited access of hepatic-origin precursors to the brain.14–16

    Quantification of MRIs revealed reduced brain perfusion in AIP NHPs, sometimes developing even before the onset of recurrent attacks, and in the absence of hyponatraemia. Reduced CBF has been observed in AIP mice that never received phenobarbital challenges to induce acute attacks, and in two patients during acute attacks.17 None of these two patients presented posterior reversible encephalopathy syndrome or brain oedema, ruling out a breach in the BBB. Remarkably, increased PBGD expression in the liver through rAAV-gene therapy in AIP mice17 or MD administration of hPBGD mRNA in AIP NHPs restored brain perfusion (figure 2E).

    Reduced brain uptake of a radiolabeled glucose analogue was confirmed after the porphyria induction in NHPs, and repeated challenges resulted in further reductions. Experimental models of acute hepatic porphyrias have also demonstrated delayed glucose tolerance test (GTT), along with altered carbohydrate metabolism in the liver and reduced glucose uptake in the brain, an essential glucose-dependent organ.15 18 Abnormal results in GTT have been previously reported in patients with AIP.19 20 In our AIP NHP model, transcriptome analysis revealed that carbohydrate homoeostasis was among the metabolic processes exhibiting the highest disturbance scores in the GO functional classification compared with control WT livers (online supplemental figures 6 and 7A). Additionally, the upregulation of critical genes for regulating gluconeogenesis and glycogen synthesis, suggests enhanced hepatic glucose synthesis to maintain glucose homoeostasis (figure 4A). This expression pattern correlates with metabolomics data, indicating a depletion of hepatic glucose levels. Remarkably, the treatments tested in the model restored hepatic glucose concentration and normalised the expression of upregulated genes, with the esception of that of SLC2A3 which codes for the glucose transported GLUT3. Considering that all animals were sacrificed on the same time schedule, these data suggest a heightened sensitivity to fasting in the AIP animals, wherein gluconeogenesis may play a more prominent role in restoring glucose homoeostasis.

    Besides glucose, fatty acid oxidation provides acetyl-CoA to initiate the TCA cycle when glucose cannot provide pyruvate to generate acetyl-CoA. This way, hepatic transcriptome (online supplemental figure 7B) and lipidomic profiles (online supplemental figure 13) suggest increased fatty acid beta-oxidation to generate ATP and to provide acetyl-CoA to initiate the TCA cycle. Compared with control WT samples, serum lipidomic profiles of AIP displayed reductions in PE, phosphatidylcholines and sphingomyelin components (figure 6) of the high-density lipoprotein (HDL) particles.21 Previous reports indicate a trend towards elevated HDL cholesterol levels in asymptomatic patients with acute porphyrias.22 23 However, there are only isolated case reports documenting increased lipoproteins in patients experiencing recurrent attacks.24–26 Remarkably, MD administration of hPBGD mRNA displayed lipidomic profiles closest to those of the control group among the treatments evaluated. Systemic messenger RNA partially reversed the significant effects of AIP induction in NHPs, indicating a potential beneficial therapeutic effect.

    Liver energy metabolism is tightly controlled. Thus, upregulation of genes involved in lipid metabolism (LPIN2, and HMGCS2 (figure 4A)) and the protein encoded by PPARGC1A gene, a transcriptional coactivator that regulates genes involved in energy metabolism, suggest mitochondrial energy dysfunction. Remarkably, those genes recovered normal expression in animals receiving MD administration of hPBGD mRNA.

    Considering the strong interconnection of lipid and carbohydrate metabolism with the TCA cycle and energy metabolism via the electron transport chain (ETC), along with its role in providing the succinyl-CoA substrate for ALAS1, a metabolomic analysis was conducted on the livers of AIP NHPs. While the entry of metabolites into the TCA cycle through citrate and α-ketoglutarate, facilitated by glutamate, remained constant, metabolites downstream of succinyl-CoA were depleted. Probably, the sharp increase in ALAS1 activity during AIP crises consumes most of the available succinyl-CoA in the mitochondria, hindering the TCA cycle’s ability to generate essential redox factors, NADH/H+and FADH2, that are transferred to ETC to produce energy.27 Collectively, these observations support the previously noted cataplerosis of the TCA cycle observed in the mouse AIP model during phenobarbital-induced AIP crises.27 28 Notably, hepatic succinate and malate levels, substrates for FADH2 and NADH/H+, respectively, were recovered to normal levels following MD administration of hPBGD mRNA.

    The resolution of the porphyria crises with systemic hPBGD mRNA treatment was also accompanied by the restoration of activities in mitochondrial respiratory chain complexes. Complex I activity was significantly inhibited in the AIP NHPs, along with complex II activity in hemin-treated animals. Similar inhibition occurred in the liver of variegate porphyria rabbits,15 suggesting that oxidative stress during the porphyria crisis may affect iron-sulfur clusters containing these complexes,29 30 critical targets for oxygen-free radicals.31 32

    The heme ring serves as the prosthetic group of several proteins, including P450 enzymes, which are mono-oxygenase enzymes involved in phase I metabolism of endogenous substances, drugs and other xenobiotics.33 34 Our findings suggested a diminished hepatic capacity to metabolise caffeine and acetaminophen due to perturbations in the heme synthesis pathway, which were restored following therapy administration. However, variations in heme affinity among different CYP isoforms have been reported,35 potentially explaining the diverse upregulation or downregulation in the hepatic cytochrome expression in AIP NHPs.

    Heme and PLP are cofactors of CBS, a haemoprotein that catalyses the conversion of homocysteine (Hcy) to cystathionine, the first step in the transsulfuration pathway, partially responsible for Hcy clearance.36–39 However, the hepatic activity of CBS was not modified in AIP NHPs (figure 5G; p=0.14). Interestingly, metabolomic studies showed unchanged cystathionine levels (a product of CBS) but reduced levels of GSH and GSSG, which are the ultimate metabolites in the transsulfuration pathway. Since ALA has been described as a prooxidant agent and heme depletion also results in oxidative stress and inflammation,39 reduced GSH and GSSG could be due to their excessive consumption.

    Besides glutathione synthesis, OCM also involves folate, polyamine metabolism, sulfur compound biosynthetic process, and the provision of methyl groups for all transmethylation reactions in nucleic acids, amino acids, creatine and phospholipids metabolism. Previously data indicate that elevated plasma Hcy in patients with AIP,36–41 may be partially reversible by supplementation of vitamin B6 (PLP), folic acid or vitamin B12.36 37 40 Although vitamin supplements were provided to the NHPs, we found a reduction in SAMe and MTA, intermediate metabolites of the methionine cycle. At the same time, the hepatic concentration of polyamines was increased after the induction of AIP and acute attacks, particularly those of putrescine. Given the critical role of SAMe in the polyamine synthesis pathway, its reduced levels could be associated with the alteration in polyamine levels. These data suggest the inactivation of methionine adenosyltransferase (MATI/III), the enzyme responsible for the hepatic conversion of methionine into SAMe. Transcriptomic studies showed no changes in the expression of MAT1A, coding for MAT (MAT I/III), the enzyme responsible for SAMe synthesis in the liver. However, MAT I/III is readily inactivated under oxidative stress conditions,39 42 which likely occur in the liver of AIP NHPs as indicated by the low GSH levels and the increased expression of SOD2 (superoxide dismutase 2 (online supplemental figure 8B)). Interestingly, a significant induction of ornithine decarboxylase 1 (ODC1), together with decreased SAMe availability, may explain the accumulation of putrescine in AIP NHPs. Remarkably, hepatic levels of SAMe and GSH, as well as partial normalisation of the OCM pathway were observed following MD treatment with hPBGD mRNA.

    Recurrent systemic administration of mRNA formulated in LNPs was safe in AIP NHPs. This technology is relatively recent, and currently clinical trials are underway for the treatment of metabolic diseases using this approach. Recently, the first interim results from a phase 1/2 mRNA trial for propionic acidaemia examined the serum antibody response.43 Only one participant tested positive for anti-PEG antibodies (a component of the LNPs) at baseline. Antibody titres increased by fourfold on day 52 after the first three doses (0.30 mg/kg every 3 weeks), and then gradually decreased, returning to baseline levels at month 12. This participant experienced grade 1 or 2 infusion-related reactions (IRRs) during 11 out of 43 infusions received. The participant continued receiving the treatment and did not experience IRRs for over a year. No other patients showed antibody responses against LNP components. Regarding the recombinant protein expressed, one participant receiving 0.60 mg/kg every 2 weeks, tested positive for anti-propionyl-coenzyme-A carboxylase antibodies at baseline but tested negative at all subsequent assessments. These data support the progression towards clinical application of mRNA therapies for AIP and other metabolic liver diseases.

    Conclusions

    A clinically relevant model of AIP was successfully developed in juvenile NHPs utilising RNAi technology. Our findings represent the first confirmation of the feasibility of employing RNAi techniques to achieve tissue-specific gene silencing in large animal models. This breakthrough offers new insights into the underlying mechanisms of the disease, sheds light on CNS processes associated with hepatic porphyrias and elucidates regulatory mechanisms in metabolic pathways with a significant impact on drug discovery. Furthermore, our results support the safety and translatability of multiple systemic administrations of hPBGD mRNA as a promising aetiological treatment for patients with AIP.

    Supplemental material

    Data availability statement

    Data are available on reasonable request.

    Ethics statements

    Patient consent for publication

    Acknowledgments

    We thank L Vandenberghe, director of the Grousbeck Gene Transfer Vector Core and R Xiao (Schepens Eye Research Institute, Boston, USA) for producing the rAAV2/8-maPBGD shRNA vector. We thank S Arcelus, U Latasa, M Ecay, FP Marchese, M Burrell, M Arechederra, A Martínez de la Cruz and L Guembe for technical assistance. A Rico, G Abizanda, the Experimental Operating Room Staff and I Alkain are acknowledged for their invaluable technical assistance and management of NHPs. We thank Drs Sebastiaan Martjin van Liempd and Diana Cabrera from the CIC bioGUNE Metabolomics platform for their dedication and technical assistance in the metabolomics analyses.

    References

    1. ↵
      1. Thapar M,
      2. Singh A,
      3. Robinson KM, et al
      . Obstacles to Early Diagnosis of Acute Hepatic Porphyria: Current Perspectives on Improving Early Diagnosis and Clinical Management. Clin Exp Gastroenterol 2024;17:1–8. doi:10.2147/CEG.S348507
      OpenUrlCrossRefPubMed
    2. ↵
      1. Puy H,
      2. Gouya L,
      3. Deybach J-C
      . Porphyrias. Lancet 2010;375:924–37. doi:10.1016/S0140-6736(09)61925-5
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Fontanellas A,
      2. Avila MA
      . Hydroxymethylbilane synthase (aka porphobilinogen deaminase): A novel metabolic tumor suppressor gene in hepatocellular carcinoma. J Hepatol 2022;77:912–4. doi:10.1016/j.jhep.2022.06.023
      OpenUrlCrossRefPubMed
    4. ↵
      1. Dickey AK,
      2. Leaf RK,
      3. Balwani M
      . Update on the Porphyrias. Annu Rev Med 2024;75:321–35. doi:10.1146/annurev-med-042921-123602
      OpenUrlCrossRefPubMed
    5. ↵
      1. Fontanellas A,
      2. Ávila MA,
      3. Anderson KE, et al
      . Current and innovative emerging therapies for porphyrias with hepatic involvement. J Hepatol 2019;71:422–33. doi:10.1016/j.jhep.2019.05.003
      OpenUrlCrossRefPubMed
    6. ↵
      1. Unzu C,
      2. Sampedro A,
      3. Mauleón I, et al
      . Sustained enzymatic correction by rAAV-mediated liver gene therapy protects against induced motor neuropathy in acute porphyria mice. Mol Ther 2011;19:243–50. doi:10.1038/mt.2010.210
      OpenUrlCrossRefPubMed
    7. ↵
      1. Serrano-Mendioroz I,
      2. Sampedro A,
      3. Alegre M, et al
      . An Inducible Promoter Responsive to Different Porphyrinogenic Stimuli Improves Gene Therapy Vectors for Acute Intermittent Porphyria. Hum Gene Ther 2018;29:480–91. doi:10.1089/hum.2017.056
      OpenUrlCrossRefPubMed
    8. ↵
      1. Jiang L,
      2. Berraondo P,
      3. Jericó D, et al
      . Systemic messenger RNA as an etiological treatment for acute intermittent porphyria. Nat Med 2018;24:1899–909. doi:10.1038/s41591-018-0199-z
      OpenUrlCrossRefPubMed
    9. ↵
      1. Unzu C,
      2. Hervás-Stubbs S,
      3. Sampedro A, et al
      . Transient and intensive pharmacological immunosuppression fails to improve AAV-based liver gene transfer in non-human primates. J Transl Med 2012;10:122. doi:10.1186/1479-5876-10-122
    10. ↵
      1. Avila MA,
      2. Garcı́a-Trevijano ER,
      3. Lu SC, et al
      . Methylthioadenosine. Int J Biochem Cell Biol 2004;36:2125–30. doi:10.1016/j.biocel.2003.11.016
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Ghit A,
      2. Assal D,
      3. Al-Shami AS, et al
      . GABAA receptors: structure, function, pharmacology, and related disorders. J Genet Eng Biotechnol 2021;19:123. doi:10.1186/s43141-021-00224-0
      OpenUrlCrossRefPubMed
    12. ↵
      1. Becker D,
      2. Viljoen D,
      3. Kramer S
      . The inhibition of red cell and brain ATPase by delta-aminolaevulinic acid. Biochim Biophys Acta 1971;225:26–34. doi:10.1016/0005-2736(71)90280-x
      OpenUrlPubMed
    13. ↵
      1. Meyer UA,
      2. Schuurmans MM,
      3. Lindberg RLP
      . Acute porphyrias: pathogenesis of neurological manifestations. Semin Liver Dis 1998;18:43–52. doi:10.1055/s-2007-1007139
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Yasuda M,
      2. Gan L,
      3. Chen B, et al
      . Homozygous hydroxymethylbilane synthase knock-in mice provide pathogenic insights into the severe neurological impairments present in human homozygous dominant acute intermittent porphyria. Hum Mol Genet 2019;28:1755–67. doi:10.1093/hmg/ddz003
      OpenUrlCrossRefPubMed
    15. ↵
      1. Jericó D,
      2. Córdoba KM,
      3. Jiang L, et al
      . mRNA-based therapy in a rabbit model of variegate porphyria offers new insights into the pathogenesis of acute attacks. Mol Ther Nucleic Acids 2021;25:207–19. doi:10.1016/j.omtn.2021.05.010
      OpenUrlCrossRefPubMed
    16. ↵
      1. Bonkowsky HL,
      2. Tschudy DP,
      3. Collins A, et al
      . Repression of the overproduction of porphyrin precursors in acute intermittent porphyria by intravenous infusions of hematin. Proc Natl Acad Sci U S A 1971;68:2725–9. doi:10.1073/pnas.68.11.2725
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Jericó D,
      2. Luis EO,
      3. Cussó L, et al
      . Brain ventricular enlargement in human and murine acute intermittent porphyria. Hum Mol Genet 2020;29:3211–23. doi:10.1093/hmg/ddaa204
      OpenUrlPubMed
    18. ↵
      1. Collantes M,
      2. Serrano-Mendioroz I,
      3. Benito M, et al
      . Glucose metabolism during fasting is altered in experimental porphobilinogen deaminase deficiency. Hum Mol Genet 2016;25:1318–27. doi:10.1093/hmg/ddw013
      OpenUrlCrossRefPubMed
    19. ↵
      1. Tschudy DP,
      2. Valsamis M,
      3. Magnussen CR
      . Acute intermittent porphyria: clinical and selected research aspects. Ann Intern Med 1975;83:851–64. doi:10.7326/0003-4819-83-6-851
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Solares I,
      2. Izquierdo-Sánchez L,
      3. Morales-Conejo M, et al
      . High Prevalence of Insulin Resistance in Asymptomatic Patients with Acute Intermittent Porphyria and Liver-Targeted Insulin as a Novel Therapeutic Approach. Biomedicines 2021;9:255. doi:10.3390/biomedicines9030255
    21. ↵
      1. Yetukuri L,
      2. Söderlund S,
      3. Koivuniemi A, et al
      . Composition and lipid spatial distribution of HDL particles in subjects with low and high HDL-cholesterol. J Lipid Res 2010;51:2341–51. doi:10.1194/jlr.M006494
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Fernández-Miranda C,
      2. De La Calle M,
      3. Larumbe S, et al
      . Lipoprotein abnormalities in patients with asymptomatic acute porphyria. Clin Chim Acta 2000;294:37–43. doi:10.1016/s0009-8981(99)00257-0
      OpenUrlCrossRefPubMed
    23. ↵
      1. Mustajoki P,
      2. Nikkilä EA
      . Serum lipoproteins in asymptomatic acute porphyria: no evidence for hyperbetalipoproteinemia. Metab Clin Exp 1984;33:266–9. doi:10.1016/0026-0495(84)90049-0
      OpenUrlCrossRefPubMed
    24. ↵
      1. Whitelaw AG
      . Acute intermittent porphyria, hypercholesterolaemia, and renal impairment. Arch Dis Child 1974;49:406–7. doi:10.1136/adc.49.5.406
      OpenUrlFREE Full Text
    25. ↵
      1. Shiue JW,
      2. Lee FY,
      3. Hsiao KJ, et al
      . Abnormal thyroid function and hypercholesterolemia in a case of acute intermittent porphyria. Taiwan Yi Xue Hui Za Zhi 1989;88:729–31.
      OpenUrlPubMed
    26. ↵
      1. Indika NLR,
      2. Kesavan T,
      3. Dilanthi HW, et al
      . Many pitfalls in diagnosis of acute intermittent porphyria: a case report. BMC Res Notes 2018;11:552. doi:10.1186/s13104-018-3615-z
    27. ↵
      1. Homedan C,
      2. Laafi J,
      3. Schmitt C, et al
      . Acute intermittent porphyria causes hepatic mitochondrial energetic failure in a mouse model. Int J Biochem Cell Biol 2014;51:93–101. doi:10.1016/j.biocel.2014.03.032
      OpenUrlCrossRefPubMed
    28. ↵
      1. Homedan C,
      2. Schmitt C,
      3. Laafi J, et al
      . Mitochondrial energetic defects in muscle and brain of a Hmbs-/- mouse model of acute intermittent porphyria. Hum Mol Genet 2015;24:5015–23. doi:10.1093/hmg/ddv222
      OpenUrlCrossRefPubMed
    29. ↵
      1. Barupala DP,
      2. Dzul SP,
      3. Riggs-Gelasco PJ, et al
      . Synthesis, delivery and regulation of eukaryotic heme and Fe-S cluster cofactors. Arch Biochem Biophys 2016;592:60–75. doi:10.1016/j.abb.2016.01.010
      OpenUrlCrossRef
    30. ↵
      1. Fukuda Y,
      2. Wang Y,
      3. Lian S, et al
      . Upregulated heme biosynthesis, an exploitable vulnerability in MYCN-driven leukemogenesis. JCI Insight 2017;2:e92409. doi:10.1172/jci.insight.92409
    31. ↵
      1. Rötig A,
      2. Sidi D,
      3. Munnich A, et al
      . Molecular insights into Friedreich’s ataxia and antioxidant-based therapies. Trends Mol Med 2002;8:221–4. doi:10.1016/s1471-4914(02)02330-4
      OpenUrlCrossRefPubMed
    32. ↵
      1. Hermes-Lima M,
      2. Castilho RF,
      3. Valle VG, et al
      . Calcium-dependent mitochondrial oxidative damage promoted by 5-aminolevulinic acid. Biochim Biophys Acta 1992;1180:201–6. doi:10.1016/0925-4439(92)90069-y
      OpenUrlPubMed
    33. ↵
      1. Correia MA,
      2. Sinclair PR,
      3. De Matteis F
      . Cytochrome P450 regulation: the interplay between its heme and apoprotein moieties in synthesis, assembly, repair, and disposal. Drug Metab Rev 2011;43:1–26. doi:10.3109/03602532.2010.515222
      OpenUrlCrossRefPubMed
    34. ↵
      1. Tsiftsoglou AS,
      2. Tsamadou AI,
      3. Papadopoulou LC
      . Heme as key regulator of major mammalian cellular functions: Molecular, cellular, and pharmacological aspects. Pharmacol Ther 2006;111:327–45. doi:10.1016/j.pharmthera.2005.10.017
      OpenUrlCrossRefPubMedWeb of Science
    35. ↵
      1. Mustajoki P,
      2. Mustajoki S,
      3. Rautio A, et al
      . Effects of heme arginate on cytochrome P450-mediated metabolism of drugs in patients with variegate porphyria and in healthy men. Clin Pharmacol Ther 1994;56:9–13. doi:10.1038/clpt.1994.94
      OpenUrlCrossRefPubMed
    36. ↵
      1. To-Figueras J,
      2. Lopez RM,
      3. Deulofeu R, et al
      . Preliminary report: hyperhomocysteinemia in patients with acute intermittent porphyria. Metab Clin Exp 2010;59:1809–10. doi:10.1016/j.metabol.2010.05.016
      OpenUrlCrossRefPubMed
    37. ↵
      1. Ventura P,
      2. Sardh E,
      3. Longo N, et al
      . Hyperhomocysteinemia in acute hepatic porphyria (AHP) and implications for treatment with givosiran. Expert Rev Gastroenterol Hepatol 2022;16:879–94. doi:10.1080/17474124.2022.2110469
      OpenUrlCrossRefPubMed
    38. ↵
      1. McCully KS
      . Homocysteine, vitamins, and vascular disease prevention. Am J Clin Nutr 2007;86:1563S–8S. doi:10.1093/ajcn/86.5.1563S
      OpenUrlAbstract/FREE Full Text
    39. ↵
      1. Fontanellas A,
      2. Ávila MA,
      3. Arranz E, et al
      . Acute intermittent porphyria, givosiran, and homocysteine. J Inherit Metab Dis 2021;44:790–1. doi:10.1002/jimd.12411
      OpenUrlCrossRefPubMed
    40. ↵
      1. To-Figueras J,
      2. Wijngaard R,
      3. García-Villoria J, et al
      . Dysregulation of homocysteine homeostasis in acute intermittent porphyria patients receiving heme arginate or givosiran. J Inherit Metab Dis 2021;44:961–71. doi:10.1002/jimd.12391
      OpenUrlCrossRefPubMed
    41. ↵
      1. Lu SC
      . Glutathione synthesis. Biochim Biophys Acta 2013;1830:3143–53. doi:10.1016/j.bbagen.2012.09.008
      OpenUrlCrossRefPubMed
    42. ↵
      1. Ramani K,
      2. Lu SC
      . Methionine adenosyltransferases in liver health and diseases. Liver Res 2017;1:103–11. doi:10.1016/j.livres.2017.07.002
      OpenUrlCrossRefPubMed
    43. ↵
      1. Koeberl D,
      2. Schulze A,
      3. Sondheimer N, et al
      . Interim analyses of a first-in-human phase 1/2 mRNA trial for propionic acidaemia. Nature New Biol 2024;628:872–7. doi:10.1038/s41586-024-07266-7
      OpenUrlCrossRef
    44. ↵
      1. Kamijo Y,
      2. Soma K,
      3. Kondo R, et al
      . Transient diffuse cerebral hypoperfusion in Tc-99m HMPAO SPECT of the brain during withdrawal syndrome following acute barbiturate poisoning. Vet Hum Toxicol 2002;44:348–50.
      OpenUrlPubMedWeb of Science
    45. ↵
      1. Wang M,
      2. Pu X,
      3. Feng B, et al
      . Alterations of Glucose Uptake and Protein Expression Related to the Insulin Signaling Pathway in the Brain of Phenobarbital-Addictive Rats by 18F-FDG PET/CT and Proteomic Analysis. J Proteome Res 2021;20:950–9. doi:10.1021/acs.jproteome.0c00703
      OpenUrlCrossRefPubMed
    46. ↵
      1. Söderpalm B
      . Anticonvulsants: aspects of their mechanisms of action. Eur J Pain 2002;6 Suppl A:3–9. doi:10.1053/eujp.2001.0315
      OpenUrlPubMed

    Supplementary materials

    Footnotes

    • KMC and DJ are joint first authors.

    • KMC and DJ contributed equally.

    • Correction notice This article has been corrected since it published Online First. The author name, Gemma Quincoces, has been corrected.

    • Contributors KMC, DJ, AS, FU, MAA and AF designed in vitro and animal experiments. KMC, DJ, PB, AS, FU, JLL and AF performed the experiments and processed animal samples and tissues. II and AMdlC performed intrahepatic administration of rAAV. CU and AF designed the rAAVs. KMC, DJ, PB, AS and AF performed behaviour assays and blood extractions. MC, GQ and IP performed PET studies. KMC, DJ performed IHC analysis. LG and TL performed the analyses of precursors in serum and CSF. MA and OM performed electrophysiological studies. PA and MM, MAM, analysed CBS and OXPHOS activities, respectively. LG-R, JLL and MAF-S performed MRI studies. SH-S and PB performed immunological analysis. JF, EA and AA performed metabolomic and serum sample studies, respectively. LJ and PGVM designed and produced mRNA formulations. LJ, MAA, AF and PGVM supervised mRNA production and supported administrative, technical and logistic tasks for sending and receiving samples. KMC, DJ, FU, JMH, PB and AF performed all statistical analysis. KMC, DJ, LJ, AS, FU, J-MH, PB, MAA, PB, PGVM and AF analysed the data. KMC, DJ, LJ, FU, PB, MAA, PGVM and AF wrote the manuscript, assisted by KMC, FU and DJ for figures and tables. All authors performed a critical revision of the manuscript for important intellectual content and final approval of the manuscript. AF is the guarantor.

    • Funding This research was supported in part by Moderna Inc. and grants from Spanish Institute of Health Carlos III (FIS) cofunded by European Union (PI21/00546 and PI24/00489, PI24/00489) and the Spanish Fundación Mutua Madrileña de Investigación Médica (2023). The financial sponsors had no role in the analysis or the development of conclusions. FU has received funding from the Fundación de Investigación Médica Aplicada (FIMA). KMC holds a Predoctoral Research Training Contract in Health (PFIS) from the Carlos III Health Institute.

    • Competing interests LJ and PGVM are employees of Moderna Inc. focusing on the development of therapeutic approaches for rare diseases. MAA is Editor of Gut Journal. The rest of the authors have no conflict of interest to declare.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

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

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.