Lentiviral haematopoietic stem-cell gene therapy for early-onset metachromatic leukodystrophy: long-term results from a non-randomised, open-label, phase 1/2 trial and expanded access

Abstract

Background: Effective treatment for metachromatic leukodystrophy (MLD) remains a substantial unmet medical need. In this study we investigated the safety and efficacy of atidarsagene autotemcel (arsa-cel) in patients with MLD.

Methods: This study is an integrated analysis of results from a prospective, non-randomised, phase 1/2 clinical study and expanded-access frameworks. 29 paediatric patients with pre-symptomatic or early-symptomatic early-onset MLD with biochemical and molecular confirmation of diagnosis were treated with arsa-cel, a gene therapy containing an autologous haematopoietic stem and progenitor cell (HSPC) population transduced ex vivo with a lentiviral vector encoding human arylsulfatase A (ARSA) cDNA, and compared with an untreated natural history (NHx) cohort of 31 patients with early-onset MLD, matched by age and disease subtype. Patients were treated and followed up at Ospedale San Raffaele, Milan, Italy. The coprimary efficacy endpoints were an improvement of more than 10% in total gross motor function measure score at 2 years after treatment in treated patients compared with controls, and change from baseline of total peripheral blood mononuclear cell (PBMC) ARSA activity at 2 years after treatment compared with values before treatment. This phase 1/2 study is registered with ClinicalTrials.gov, NCT01560182.

Findings: At the time of analyses, 26 patients treated with arsa-cel were alive with median follow-up of 3·16 years (range 0·64-7·51). Two patients died due to disease progression and one due to a sudden event deemed unlikely to be related to treatment. After busulfan conditioning, all arsa-cel treated patients showed sustained multilineage engraftment of genetically modified HSPCs. ARSA activity in PBMCs was significantly increased above baseline 2 years after treatment by a mean 18·7-fold (95% CI 8·3-42·2; p<0·0001) in patients with the late-infantile variant and 5·7-fold (2·6-12·4; p<0·0001) in patients with the early-juvenile variant. Mean differences in total scores for gross motor function measure between treated patients and age-matched and disease subtype-matched NHx patients 2 years after treatment were significant for both patients with late-infantile MLD (66% [95% CI 48·9-82·3]) and early-juvenile MLD (42% [12·3-71·8]). Most treated patients progressively acquired motor skills within the predicted range of healthy children or had stabilised motor performance (maintaining the ability to walk). Further, most displayed normal cognitive development and prevention or delay of central and peripheral demyelination and brain atrophy throughout follow-up; treatment benefits were particularly apparent in patients treated before symptom onset. The infusion was well tolerated and there was no evidence of abnormal clonal proliferation or replication-competent lentivirus. All patients had at least one grade 3 or higher adverse event; most were related to conditioning or to background disease. The only adverse event related to arsa-cel was the transient development of anti-ARSA antibodies in four patients, which did not affect clinical outcomes.

Interpretation: Treatment with arsa-cel resulted in sustained, clinically relevant benefits in children with early-onset MLD by preserving cognitive function and motor development in most patients, and slowing demyelination and brain atrophy.

Funding: Orchard Therapeutics, Fondazione Telethon, and GlaxoSmithKline.

Figure 1

High-level engraftment of gene-corrected HSPCs (A) Assessed over time by disease subtype by percentage of lentiviral vector-transduced cells in bone marrow clonogenic progenitors. (B) Mean VCN in PBMCs. LLQ is 0·0037 VCN per cell. Zero values are plotted as 0·001. (C) Mean VCN in bone marrow-derived CD34⁺ cells. LLQ is 0·0037 vector copy number/cell. Zero values are plotted as 0·001. (D) Mean ARSA activity in PBMCs. LLQ is 25·79 nmol/mg/h. ARSA activity measured in PBMC in the intention-to-treat set after treatment at years 2 (coprimary endpoint) and 3 was compared with pre-treatment values using a mixed-model repeated measures model. (E) Mean ARSA activity in cerebrospinal fluid (CSF). LLQ is 0·0032 nmol/mg/h. Geometric means and 95% CIs are presented where there are at least three patients with non-missing data. ARSA=arylsulfatase A. GM=geometric mean. LLQ=lower limit of quantification. PBMCs=peripheral blood mononuclear cells. VCN=vector copy number. *95% CI for 60-month timepoint: 24·35–128·73. †95% CI for 72-month timepoint: 6·69–303·48. ‡From adult reference donors. §From paediatric reference donors. In all panels, values less than the LLQ were imputed as the LLQ as this represents a conservative approach to treatment evaluation in those cases.

Figure 2

GMFM scores and age at severe motor impairment or death (A) GMFM scores for patients with late-infantile and early-juvenile MLD compared with age-matched and disease subtype-matched untreated natural history controls at 2 and 3 years after treatment. Adjusted least-squares means, treatment difference (atidarsagene autotemcel [arsa-cel] minus natural history), and the associated 95% CI and p value were reported overall and by disease subtype (late infantile, early juvenile) for the null hypothesis of 10% or less difference in total GMFM scores between treated and natural history patients at years 2 and 3. (B) Kaplan-Meier plot showing age at severe motor impairment or death in patients with late-infantile MLD versus untreated natural history late-infantile MLD controls. (C) Kaplan-Meier plot showing age at severe motor impairment or death in patients with pre-symptomatic and early-symptomatic early-juvenile MLD versus untreated natural history early-juvenile MLD controls. Severe motor impairment-free survival is defined as the interval from birth to the earlier loss of locomotion and sitting without support (GMFC level 5 or 6) or death from any cause; otherwise severe motor impairment-free survival is censored at the last GMFC assessment date. Symptomatic status refers to arsa-cel treated patients at the time of treatment. GMFM=gross motor function measure. MLD=metachromatic leukodystrophy. *p values calculated using an unstratified log-rank test.

Figure 3

Cognitive performance and verbal age-equivalent profiles (A) Age-equivalent cognitive performance in late-infantile patients. (B) Age-equivalent cognitive performance in early-juvenile patients. (C) Verbal-age equivalent in late-infantile patients. (D) Verbal-age equivalent in early-juvenile patients. Age-equivalent corresponds to the chronological age at which, on average, typically developing children reach a given raw score. Cognitive age-equivalent for late-infantile (A) and early-juvenile (B) at each visit has been derived as follows. For Wechsler Preschool & Primary Scale of Intelligence (WPPSI) and Wechsler Intelligence Scale for Children (WISC): (development quotient performance x chronological age)/100 for which development quotient is derived by dividing the age-equivalent by the chronological age and then multiplying by 100. For Bayley III: cognitive raw scores have been compared with the tabulated values in the Bayley III manual to calculate cognitive age-equivalent. For Bayley II and in cases for which a neuropsychological assessment has been done but a questionnaire could not be completed because of severe clinical condition, cognitive age-equivalent is based on mental development age as reported on the case report form (CRF). Verbal age-equivalent for late-infantile (C) and early-juvenile (D) at each visit has been derived as follows. For WPPSI and WISC: (development quotient language x chronological age)/100. For Bayley III: expressive communication and receptive communication raw scores have each been compared with the tabulated values in the Bayley III manual and based on the average of the mental development ages for the two scores. For Bayley II and in cases for which a neuropsychological assessment has been done but a questionnaire could not be completed because of severe clinical condition, verbal age-equivalent is based on mental development age as reported on the CRF. Arsa-cel=atidarsagene autotemcel. MLD=metachromatic leukodystrophy.

Figure 4

Longitudinal evaluation of brain MRI of patients treated with arsa-cel vs natural history (A) Axial T2-weighted MRI obtained from a patient with late-infantile MLD (patient 3) over time showing, at baseline, a physiological T2 signal corresponding to myelin maturation, typical of the first year of life, followed by appearance of subtle posterior periventricular T2 hyperintensities that stabilised over time, with normal maturation of the remaining brain regions. (B) The comparison with the corresponding images obtained from his untreated sibling at the same age. (C) The comparison with an unrelated untreated patient with late-infantile MLD with similar age of onset, followed along disease course demonstrating white matter alterations (T2 hyperintensities) involving, in the earliest phases, periventricular areas and corpus callosum, spreading to subcortical regions, cerebellum and corticospinal tracts, and associated with progressive brain atrophy (enlargement of ventricular system and subarachnoid spaces). (D and E) Comparison of brain MRI total scores of patients with late-infantile (D) and early-juvenile by symptomatic status (E) treated with atidarsagene autotemcel [arsa-cel] versus natural history through an non-linear mixed-effects model based on a published methodology and strategy for model selection. MRI severity scoring system is a modified Loes’ score using methodology as previously described. The maximum total score is 31·5. MLD=metachromatic leukodystrophy.