1 INTRODUCTION

Fabry disease (OMIM 301500) is an X-linked lysosomal storage disorder (LSD) originally found to be due to impaired hydrolysis of the terminal galactose of the sphingolipid ceramide trihexoside (globotriaosylceramide, Gb₃) by the enzyme alpha-galactosidase A (α-gal A), encoded by the GLA gene.^{1, 2} The realisation that it was a lack of the enzyme activity resulting in accumulated sphingolipids with terminal galactose residues has been the mainstay of understanding this disease. Principal treatments have involved replacement of this enzyme activity using various products or enhancement of misfolded lysosomal activity using chaperones.^3-6 Therapeutic goals focus on reducing the clinical burden of the disorder from cardiac dysfunction and renal insufficiency, along with reducing the higher risk of stroke and lessening chronic pain.⁷ Since the start of enzyme replacement therapy (ET), this method of treatment has been the most commonly used worldwide. ET involves biweekly enzyme infusions, with products that have a half-life of less than 2 h.^{8, 9} A pegylated form of the enzyme has been approved in the United States that has demonstrated a half-life between 53 and 121 h,¹⁰ which still requires biweekly intravenous infusions. An oral chaperone (migalastat) has also been approved but only certain GLA mutations are amenable.⁴ Migalastat was not approved for clinical use at the time of initiation of this study. The use of ET means a biweekly intravenous infusion for life, requiring infusions at a hospital, infusion centre or at home, repeated vascular access, the potential use of central lines and the complications that accompany them. Reimbursement issues are problematic for intravenous or oral drugs. Patients may not be able to move to a geographical region where coverage is provided thereby limiting lifestyle choices and career goals. As an alternative to the burden and cost of lifelong use of medication, this prompted a need for a ‘one and done’ approach. There are two broad approaches to gene therapy in this context: lentiviral or adeno-associated virus (AAV) modalities. An ex vivo lentiviral approach, which transduces multi-potent haematopoietic stem cells, such as CD34⁺ lymphocytes, that populate the bone marrow, can provide systemic production of α-gal A. Direct AAV injection has a higher tropism for the liver, but this approach cannot be used in patients who have a cross-reactive antibody to the AAV vector serotype being used. Since the AAV transduction is primarily performed in non-dividing cells, the effective dose of the vector can also reduce over time. In Fabry disease, α-gal A can be taken up by uninfected bystander cells, similar to exogenous intravenous ET, to provide cross-correction in those cells.

The FACTs (Fabry Disease Clinical Research and Therapeutics) study involved recombinant lentivirus (LV)-mediated transfer of a codon-optimised cDNA for human α-gal A into haematopoietic stem/progenitor cells (HSPCs) from Fabry patients followed by autologous re-infusions of the cell products (clinicaltrials.gov, #NCT02800070). The study was a non-randomised, open-label, multi-centre clinical trial to evaluate the safety and efficacy of LV-transduced HSPCs for 5 years in males with classical Fabry disease who were already on ET. The primary endpoints (safety) were determined using criteria from the National Cancer Institute of Canada (NCIC) Common Terminology Criteria for Adverse Events (CTCAE). Safety monitoring assessments were performed for cardiac and renal function as well as for stroke, which are the three principal clinical outcome-related organ systems in Fabry disease. Secondary analyses included: (a) changes in α-gal A enzyme activity within the plasma, lymphocytes and bone marrow, (b) changes in Gb₃ levels in plasma and urine, (c) changes in lyso-Gb₃ and related analogue levels in plasma and urine, (d) persistence of LV-transduced cells as measured by quantitative polymerase chain reaction (qPCR) and (e) determination of anti-α-gal A antibody titres.

Interim results and safety data were published in 2021.¹¹ In addition to the interim publication on safety and clinical outcomes, a recent paper showed sustained reductions in Gb₃ and globotriaosylsphingosine (lyso-Gb₃) levels in our treated Fabry patients.¹² Longitudinal evaluations of engraftment in our Fabry disease patients receiving LV-mediated gene therapy were also performed; there was evidence for persistent polyclonal haematopoiesis and no evidence of clonal dominance in any patient.¹³ Examining clonal expansion is relevant because lentiviral constructs integrate into host DNA and have the potential to dysregulate proto-oncogenes, which can manifest with clonal expansion and haematological malignancies.¹⁴ Herein, we present the 5-year End-of-Study results from that FACTs trial, with ongoing safety monitoring outcomes, adverse clinical event assessments, and renal and cardiac function analyses.

2 RESULTS

2.1  α-Gal A activity and vector copy number

Our study design was described in an earlier interim report.¹¹ This open-label clinical trial was divided into four phases: Phase 1 (screening), Phase 2 (mobilisation), Phase 3 (treatment), and Phase 4 (post-treatment follow-up). Seven males with classic Fabry disease were screened for participation in the trial; five met the inclusion and exclusion criteria in order to proceed. All patients signed informed consent. Our earlier report characterised the initial results for all five patients up to 18 months post-infusion.¹¹ Patients treated with autologous CD34+ cells engineered to express α-gal A were shown to synthesise the missing enzyme in their leukocytes resulting in secretion of the deficient enzyme into their plasma. This resulted in continuous enzyme production and a decrease in glycosphingolipid accumulation, most notably plasma lyso-Gb₃. Antibody titres to α-gal A were shown to decrease after gene therapy. This End-of-Study report analyses subjects for 5 years post-treatment.

We have elected to focus on results from years 1 to 5 in this report, as earlier time points were previously described.¹¹ Plasma α-gal A activity was shown to be significantly increased (Supporting Information Figure 1) in all five patients from 1 to 5 years after stem cell transplantation (Figure 1A). Plasma levels were elevated at 1 year, then reached asymptotes throughout the course of the study. All patients were below the reference range for plasma α-gal A activity, with the exception of Patient 5. Leukocyte α-gal A-specific activity showed a similar response to plasma α-gal A activity (Figure 1B). Stable hydrolase specific activity was significantly increased post-therapy from years 1 to 5 of the study (Supporting Information Figure 2). In this case, all patients were within or slightly below the reference range, with the exception of Patient 5 whose leukocyte α-gal A-specific activity was supranormal.

Vector copy number (VCN) was also analysed in white blood cells obtained from all treated patients. The methods for VCN determination have been described previously.¹⁵ VCN was significantly increased in all patients (Supporting Information Figure 3). VCN decreased more rapidly at earlier time points as short-term HSPCs were extinguished, but the rate of decline was more moderate from years 1 to 5 of this study (Figure 1C). VCN mirrors leukocyte α-gal A activity as we previously reported.¹¹

2.3 Renal outcomes

Urinary protein levels, collected over 24 h, were determined for all patients (Figure 2A). Patient 2 had elevated initial measurements far exceeding the reference range (0–.15 g/24 h). However, these levels were observed to decrease throughout the trial to year 3 at which point stabilisation was observed, almost approaching the level at the time of transplantation. Patient 2 had Stage 3 chronic kidney disease at the time of enrolment. Urinary protein decreased for Patient 1 during years 2–4.5, but raised to pre-transplant levels by year 5. Patients 3–5 all had low levels of protein excretion near the upper reference range. There was a biphasic slope for urinary protein for Patient 2 in this study (Supporting Information Figure 6). A slope of −.37 was observed until year 3, at which point there was stabilisation and a slope of .27 was observed for the final 2 years of Phase 4. Because of the small sample size, there is no specific reason we could attribute to stabilisation of the proteinuria 3 years post-transplant in Patient 2. The slope calculations ranged from −.06 (Patient 1) to −.01 (Patient 3), .003 (Patient 4) and  .006 (Patient 5). All patients maintained use of other medications during this period. Patients 1 and 2 were prescribed losartan, but due to the voluntary recall of that drug during the FACTs study, they were switched to an equivalent dose of irbesartan. By comparison, the annualised mean change in estimated glomerular filtration rate (eGFR) in patients with Fabry disease treated with migalastat was −.7 mL/min/1.73 m²/year¹⁶ and −.97 mL/min/1.73 m²/year in healthy male controls who initially had normal kidney function.¹⁷ The therapeutic target is up to -3 mL/min/1.73 m²/year⁷ and can be slightly less than this in patients on ET.¹⁸

As proteinuria is associated with kidney decline, we calculated eGFR for all patients throughout the study (Figure 2B). Treatment phase data were omitted for clarity as melphalan treatment resulted in a transient increase in eGFR activity (data not shown). Patient 2 displayed Fabry-associated kidney disease and had an eGFR level that decreased throughout the study until year 3 when stabilisation was observed. Patient 1 had a more intermediate eGFR reading, consistent with Stage 2 kidney disease in this subject. Finally, Patients 3–5 had stable eGFR measurements near 90 mL/min/1.73 m². eGFR slope is a documented marker of kidney function over time.¹⁹ Patient 2 demonstrated a biphasic response with a slope of −7.48 until year 3 and then a slope = .11 from years 3 to 5 of this study, similar to the 24 h urinary protein results illustrated in the previous figure (Figure 3C). Patients 3–5 had eGFR slopes that were decreasing (Patient 3 = −.59, Patient 4 = −1.80 and Patient 5 = −.38). In contrast, the eGFR slope for Patient 1 was closer to 0 (.09; Figure 3B).

2.5 Globotriaosylceramide (Gb₃)

Plasma Gb₃ levels decreased throughout the course of this study; all five patients had plasma Gb₃ levels within the reference range at year 5 (Figure 4A). Note that Patient 1 stopped ET on day 548, Patient 4 stopped ET on day 214 and Patient 3 did not resume ET after transplantation of transduced cells. Interestingly, there were no significant differences in plasma Gb₃ in any patient observed before and after the gene therapy treatment (Supporting Information Figure 8).

Plasma lyso-Gb₃ is a recognised metabolite of interest in Fabry disease. Elevated serum lyso-Gb₃ correlates with disease outcome and it is a more soluble metabolite than serum Gb₃.^{21, 22} On the other hand, it is present in much lower concentrations. Plasma lyso-Gb₃ levels were stable from years 1 to 5 in all patients (Figure 4B). Analysis of plasma lyso-Gb₃ levels before and after gene therapy revealed statistically significant decreases of plasma lyso-Gb₃ in Patient 1 (gene therapy alone), Patient 2 (ET + gene therapy), Patient 3 (gene therapy alone) and Patient 5 (ET + gene therapy; Figure 5). Interestingly, Patient 4 showed a small, but significant increase in plasma lyso-Gb₃ (pre-therapy = 29.09 nM; post-therapy = 36.31 nM).

Urine Gb₃ levels varied depending on whether patients selected post-treatment ET therapy or not (Figure 4C). Urine Gb₃ increased when each patient stopped their biweekly ET administration (Figure 4C). Patient 1 showed a biphasic urine Gb₃ curve: a sharp increase until year 2.1, followed by a more gradual increase until year 5. Urine Gb₃ for Patient 3 remained elevated but fairly constant throughout the follow-up period. Patient 4 showed elevation of urine Gb₃ until 3.5 years post-transplantation, followed by a decrease until year 5. In contrast, Patients 2 and 5 remained on ET and their urine Gb₃ levels remained close to baseline levels (Supporting Information Figure 9). Differences in urine Gb₃ were observed to be statistically significant in both cohorts: both for the group that was on gene therapy alone and the two patients on ET and gene therapy (Supporting Information Figure 10).

Urine lyso-Gb₃ levels fluctuated throughout the year 1–5 period, but remained relatively low (Figure 4D). Interestingly, there was a significant decrease in urine lyso-Gb₃ in Patients 2 and 5 pre- and post-therapy who were on both ET and gene therapy (Supporting Information Figure 11). The evaluation of the complete biomarker profiles (with all the isoforms and analogues) for these five Fabry patients has recently been published by our group.¹²

2.6 Antibody titre

Antibodies that recognise recombinant α-gal A protein are a significant impediment to ET treatment. Patients with elevated immune-associated reactions were excluded from this FACTs study. As we reported earlier,¹¹ there was an initial transient rise in anti-α-gal A antibody titre in Patients 1, 3 and 4. However, this reactivity decreased in the early stages of the trial.¹¹ We now illustrate that anti-α-gal A antibody titres remain at or near baseline beyond 18 months in this study (Figure 6).

3 DISCUSSION

We have shown that LV-mediated gene therapy employing autologous, peripheral blood-mobilised CD34+ HSPCs with non-myeloablative conditioning is a safe treatment modality in a Phase 1 cohort of five male Fabry disease patients. Importantly, our high-titre vector preparation and reasonable multiplicity-of-infection allowed us to treat all five patients with the same good manufacturing practice-grade batch of recombinant LV. This helped to considerably reduce costs and logistical complexity. After reduced-intensity melphalan treatment, all patients engrafted their autologous modified α-gal A expressing cells. Importantly, all patients synthesised and secreted α-gal A throughout the course of the study. Expression of α-gal A resulted in a decrease in plasma lyso-Gb₃ in four of five patients and stabilisation of kidney symptoms in all patients. This gene therapy approach is safe as only two SAEs were described throughout the course of this 5-year study.¹¹ Throughout the long-term follow-up interval of 5 years, there were no AEs of any grade attributable to the cellular gene therapy intervention or host conditioning. We demonstrate that this therapeutic approach has merit, is durable and should be explored in a larger clinical trial. While this study treated only five male Fabry patients, the fact that all five treated patients responded positively to the gene therapy bodes well for outcomes in a larger study.

All five patients synthesised and secreted α-gal A throughout this study. Plasma α-gal A was below the reference range in four of five patients. Similarly, three of five patients had leukocyte α-gal A-specific activity below the reference range. Patient 2 had leukocyte α-gal A-specific activity within the reference range. Plasma and leukocyte α-gal A enzyme activity remained above the reference range in Patient 5 throughout this clinical trial. Previous model system studies showed no adverse effects of supraphysiological α-gal A expression in transgenic mice.^{23, 24} The elevated expression of α-gal A in Patient 5 will continue to be monitored in our subsequent long-term follow-up study, but is not now a point of clinical concern. The low, but durable level of α-gal A synthesis and secretion observed in Patient 1 is still effective in the reduction of plasma lyso-Gb₃. Interestingly, this patient continued to have reduced urinary protein throughout the trial, resulting in a positive eGFR slope of .09. All five patients qualified to cease their ET therapy and three of five elected to do so.

This was a safety study and not powered to evaluate the correlation between gene therapy and any specific clinical parameters. We showed stable and prolonged expression of LV-transduced cells with no evidence of formation of anti-α-gal A antibodies, beyond 18 months post-transplant. In addition, all patients are clinically healthy. Not all patients showed the same responses in catabolites but work equating specific biomarkers and clinical outcomes in Fabry disease is still in its early stages. This is also evident from the lack of LGE advancement in any patient despite some increases in troponin levels and stabilisation of proteinuria despite a rise in urine Gb₃ in Patient 1. While Patient 1 developed intermittent atrial fibrillation with an atrial thrombus, paroxysmal atrial fibrillation can be seen in 13.3% of males with Fabry disease with a mean age of presentation of 67 ± 11 years suggesting that prevalence increases with age.^{25, 26}

Success of haematopoietic cell-based gene therapy requires conditioning with a chemotherapeutic agent to allow efficient engraftment of engineered autologous cells.²⁷ Since we were treating a chronic LSD, for which ET is the current standard-of-care, we elected to use a reduced dosage of melphalan, opting for partial ablation. We were also cognizant that metabolic cooperativity or cross-correction in LSDs like Fabry disease may decrease the threshold for α-gal A activity needed.^{15, 28} We reasoned that this reduced-intensity conditioning approach would lead to fewer SAEs, which would be better tolerated by patients and allow the gene therapy to be administered as an out-patient procedure. Polyclonal reconstitution was observed in all patients up to 18 months post-infusion.¹³ This therapy was delivered in an out-patient manner for four of five patients; one patient was kept overnight by the hospital for observation.

Busulfan is the current chemotherapeutic agent of choice for haematopoietic cell-based gene therapy treatment of paediatric autoimmune and malignant diseases with fatal outcome.^{27, 29-34} Delivery of busulfan must be closely monitored and use of this agent for full ablation requires intensive care unit (ICU) admittance. Busulfan is known to cause seizures in recipients³⁵ and its use can lead to long-term gonadal toxicities³⁶ along with an increased risk of secondary malignancies.^37-39 While a stronger myeloablation regimen has been used in malignant haematopoietic diseases, we hypothesised that for a metabolic disease, stronger myeloablation with an autologous transplant was unnecessary since even partial improvements in enzyme levels may be sufficient to displace the requirement of ET. Our results support the hypothesis that high-dose chemotherapy conditioning regimens may not always be necessary in autologous lentiviral-based gene therapy.

Plasma Gb₃ levels decreased in all five patients on ET, from years 1 to 5. However, these differences were not statistically significant. While elevated plasma Gb₃ is useful for diagnosis of Fabry disease, it does not seem to be informative as a biomarker with ET treatment or chaperone therapy.^{40, 41} Higher serum lyso-Gb₃ levels were shown to be independently associated with serum creatinine and cardiomyopathy in a cohort of 69 Fabry disease patients, including 28 males.²² In a follow-up report, plasma lyso-Gb₃ was also demonstrated to be a factor associated with adverse clinical outcomes in a long-term study of 26 male and 40 female Fabry disease patients.²¹ Plasma lyso-Gb₃ was significantly lower in Patients 1–3 and 5 in our FACTs trial. Post-treatment plasma lyso-Gb₃ from years 1 to 5 was elevated compared to pre-treatment values in Patient 4. Further characterisation of plasma lyso-Gb₃ in Patient 4 revealed that this metabolite increased 23% compared to baseline during years 1–5 of follow-up post-therapy, however, urinary protein in Patient 4 remained in the reference range and eGFR was indicative of Type I chronic kidney disease. This subject has daily urinary protein secretion near the upper limit of the reference range (.15 g/day; slope = .003) and an eGFR slope of −1.80. Given the proposed clinical importance of elevated plasma lyso-Gb₃, this patient is a candidate to resume ET in the near future.

Interestingly, we observed that urine Gb₃ levels rose in all three patients that paused ET (Supporting Information Figure 9). The reason(s) underscoring this observation are unclear. Increases in urine Gb₃ have been observed either as a consequence of decreased ET dosage or the emergence of anti-α-gal A antibodies for patients on ET therapy.⁴² No increases in anti-agalsidase antibody activity were observed in any patient (Figure 6), suggesting that the increased urine Gb₃ was associated with the low, but detectable α-gal A enzyme produced by these three patients. Despite the rise in urine Gb₃, kidney function remained normal in all three patients that paused ET with eGFR exceeding 75 in each and decreases observed in urinary protein secretion in Patients 1 and 3. In contrast, statistically significant decreases in pre- and post-treatment urine Gb₃ and urine lyso-Gb₃ were observed in Patients 2 and 5. Both of these patients remained on ET after the stem cell transplantation. Collectively, these data suggest that biweekly ET infusion is useful for decreasing these catabolites, however, the absolute lowest levels are achieved by the combination therapy; with the gene therapy continuously supplying α-gal A to each patient.

We examined whether there were correlations in enzyme activity, VCN, Gb₃ metabolites, stem cell dosage at infusion and age at transplant, with a selected cut-off of R² > .80. Plasma α-gal A, leukocyte α-gal A and VCN were highly correlated. Plasma α-gal A and VCN correlate to age at transplant. There were no correlations of enzyme or VCN with Gb₃ or lyso-Gb₃ in plasma or urine, underscoring the challenges of using these measures as bone fide Fabry disease biomarkers.

This was the first gene therapy clinical trial to be completed for Fabry disease.^{11, 13} In addition to existing gene therapy protocols, there is the potential to examine other means of introducing α-gal A into circulation and then into target cells. Our group has recently shown that Fabry patient T cells can be appropriate targets for implementing therapy, utilising our LV-based protocol.⁴³ These autologous cells can be collected easily, expanded, transduced and infused multiple times as needed in out-patient procedures. The transduced T cells distribute the enzyme systemically throughout the body. Transduced T cells can also be long-lived in the body and take up residence at key sites. We have shown that analogous vector-transduced T cells may also be effective for other LSDs, including Gaucher, Farber and Pompe diseases.⁴³ This concept of an alternate autologous cell source is especially timely given how adept transplant groups have become at collecting, transducing and expanding T cells for chimeric antigen receptor (CAR-T)-based cancer trials. Furthermore, implementation of an autologous T cell-based gene therapy protocol to treat eligible Fabry patients may cost even less than the HSPC-based protocol we used here.

Preclinical testing of AAV vector backbones to deliver GLA cDNA has been reported.^{44, 45} These studies have led to the development of clinical trials (NCT04046224 and NCT040400409), although the latter trial has been terminated. Two additional groups are also utilising analogous AAV vectors to introduce the GLA cDNA into Fabry patients (NCT04519749 and NCT06270316).

In summary, we have shown that utilising LV-mediated gene therapy targeting HSPCs is a safe and efficacious means of treating Fabry disease. Five males with four distinct classical Fabry disease mutations were selected to participate in this clinical trial. Gene therapy was well tolerated with only two SAEs documented to date. All five patients were eligible to pause ET and three patients elected to do so. Compellingly, we estimate that pausing ET in these three patients alone has led to a savings of $4.8 million dollars to the Canadian health care system at the time of the last patient visit of Patient 5. This cost saving continues to increase daily as all five patients are now in long-term follow-up for 15 years. These data also demonstrate that our LV-mediated approach can be successfully performed in an academic setting. Our multi-site model illustrates that cells can be collected at one site, isolated and enriched at another, transduced at a third site and shipped back to the origin for infusions into patients. We have demonstrated that assays can also be performed at multiple locations. Canada has the ability to incorporate such true experimental, ‘bench-to-bedside’ approaches directly into the national health care system. There is already an example of such cooperative studies in Canada for individualised CAR-T cells engineered against CD19+ cancers.⁴⁶ Another approach would be to form an alliance of University hospitals to manufacture gene therapies, as has been done in Switzerland.⁴⁷ Collectively, can the lessons learned from CAR-T cell therapies for treating cancer, for example, be applied to rare diseases including LSDs such as Fabry disease? Future clinical trials in this area would aim to expand the patient cohort examined in this study as well as include female Fabry disease patients and younger male subjects.

4 MATERIALS AND METHODS

AUTHOR CONTRIBUTIONS

α-Gal A enzymatic assays were performed by the London Health Sciences Centre Clinical Biochemical Genetics Laboratory (Director C. Anthony Rupar), London, ON. Gb₃ and lyso-Gb₃ analyses were performed by the Auray-Blais laboratories at the Université de Sherbrooke, Sherbrooke, QC. Vector copy number assays and ELISAs to measure anti-α-gal IgG antibodies were performed by the Medin laboratories at the Medical College of Wisconsin, Milwaukee, WI. The FACTs team had confidential access to the data through the trial management company. Aneal Khan, Dwayne L. Barber and Jeffrey A. Medin wrote and edited the manuscript. All the authors edited and approved the final manuscript before publication.

CONFLICT OF INTEREST STATEMENT

Aneal Khan holds research grants from Actelion, Amicus, AVROBIO, Canadian Institutes of Health Research, Canadian Fabry Disease Initiative, Genome Canada/Genome Prairie, Mallinckrodt, Moderna, New Frontiers in Research Fund (Canada), Resverlogix, Sangamo, Sanofi/Genzyme, Shire HGT and Takeda. He has received consulting fees, speaker fees and travel grants from Actelion, Amicus, AVROBIO, Sanofi/Genzyme and Shire HGT. Dr. Khan is the President and Medical Director of M.A.G.I.C. Clinic (Metabolics and Genetics in Canada), Chief Executive Officer of Discovery DNA and Medical Director of Maternal Genomics. He is the President of rDNA (Rare Disease Network of Alberta). Dr. Khan is part of a revenue sharing agreement with University Health Network for Fabry gene therapy.

Dwayne L. Barber and Alexandra Berger were partially paid from a Sponsored Research Agreement by AVROBIO, Inc.

C. Anthony Rupar has the following financial relationships to disclose: the Biochemical Genetics clinical diagnostic laboratory at his home institution was contracted by AVROBIO, Inc. to assay enzymes on a fee for service basis. He is the laboratory director but received no personal compensation.

Christiane Auray-Blais has received a service contract and honoraria for biomarker analysis with AVROBIO, Inc. and a grant from CIHR.

Armand Keating has received consultancy fees from AVROBIO, Inc. unrelated to this study.

Michael L. West has received research grants, consulting fees or speaker fees with Amicus Therapeutics, Chiesi, Protalix, Sanofi and Takeda, revenue distribution agreement with University Health Network regarding gene therapy using technology from this work, shares IP in Fabry cardiac biomarker panel with University of Alberta.

Jeffrey A. Medin has the following financial relationships to disclose: Honoraria—Sanofi Genzyme and Shire; Scientific co-Founder—AVROBIO, Inc.; Shareholder—AVROBIO, Inc.; and Grants from Canadian Institutes of Health Research and Kidney Foundation of Canada and AVROBIO, Inc.

William M. McKillop, Graeme Fraser, Daniel H. Fowler and Ronan Foley have no financial relationships to disclose in relation to this trial.

ETHICS STATEMENT

Research Ethics Board (REB) approval was obtained from University Health Network, Alberta Health Services, Capital Health Services, Hamilton Integrated Research Ethics and the Medical College of Wisconsin Institutional Review Board for this study.