Scholarly article on topic 'Effect of anti-laronidase antibodies on efficacy and safety of laronidase enzyme replacement therapy for MPS I: A comprehensive meta-analysis of pooled data from multiple studies'

Effect of anti-laronidase antibodies on efficacy and safety of laronidase enzyme replacement therapy for MPS I: A comprehensive meta-analysis of pooled data from multiple studies Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
Academic journal
Molecular Genetics and Metabolism
OECD Field of science
Keywords
{ADA / Antibody / "Enzyme replacement therapy" / Laronidase / "Mucopolysaccharidosis type I" / "Urinary glycosaminoglycan"}

Abstract of research paper on Clinical medicine, author of scientific article — Yong Xue, Susan M. Richards, Asif Mahmood, Gerald F. Cox

Abstract Enzyme replacement therapy (ERT) with laronidase has an important role in the treatment of patients with mucopolysaccharidosis type I (MPS I). Laronidase is safe and has demonstrated effectiveness in terms of stabilizing or improving conventional clinical and laboratory markers of the disease. However, like most ERTs, laronidase produces an anti-drug IgG antibody response in more than 90% of patients during the first few months of treatment. Preclinical data from the MPS I canine model suggest that anti-drug antibodies (ADA) impair enzyme uptake in target tissues. In patients, the effects on tissue glycosaminoglycan (GAG) clearance are difficult to assess directly but data from clinical studies have suggested an association between ADA and both a reduced pharmacodynamic response and hypersensitivity reactions. This comprehensive meta-analysis of pooled data from patients in three clinical studies of laronidase (including one study with an extension) was undertaken to provide a more robust assessment of the relationship between the ADA response to laronidase, clinical and laboratory markers of MPS I, and hypersensitivity reactions. The meta-analysis demonstrated an inverse relationship between the ADA response and the percent reduction in urinary GAG (uGAG) levels. However, no relationships between the ADA response and changes in percent predicted forced vital capacity and six-minute walk test were seen. The study also re-assayed stored serum samples from the original trials with a novel method to determine the inhibitory effect of ADA. Patients with higher ADA exposure over time were found to have higher inhibition of enzyme uptake into cells. High ADA exposure can result in a commensurate level of enzyme uptake inhibition that decreases the pharmacodynamic effect of the exogenously administered therapeutic enzyme, but with no clear effect on clinical efficacy.

Academic research paper on topic "Effect of anti-laronidase antibodies on efficacy and safety of laronidase enzyme replacement therapy for MPS I: A comprehensive meta-analysis of pooled data from multiple studies"

Effect of anti-laronidase antibodies on efficacy and safety of laronidase enzyme replacement therapy for MPS I: A comprehensive meta-analysis of pooled data from multiple studies

Yong Xue a, Susan M. Richards b, Asif Mahmoodc, Gerald F. Cox d'*

a Clinical Development, Rare Diseases Group, Sanofi Genzyme, Naarden, The Netherlands b Clinical Laboratory Sciences, Sanofi Genzyme, Framingham, MA, USA c Global Pharmacovigilance and Epidemiology, Sanofi Genzyme, Cambridge, MA, USA d Clinical Development, Rare Diseases Group, Sanofi Genzyme, 500 Kendall Street, Cambridge, MA 02142, USA

ARTICLE INFO ABSTRACT

Enzyme replacement therapy (ERT) with laronidase has an important role in the treatment of patients with mucopolysaccharidosis type I (MPS I). Laronidase is safe and has demonstrated effectiveness in terms of stabilizing or improving conventional clinical and laboratory markers of the disease. However, like most ERTs, laronidase produces an anti-drug IgG antibody response in more than 90% of patients during the first few months of treatment. Preclinical data from the MPS I canine model suggest that anti-drug antibodies (ADA) impair enzyme uptake in target tissues. In patients, the effects on tissue glycosaminoglycan (GAG) clearance are difficult to assess directly but data from clinical studies have suggested an association between ADA and both a reduced pharma-codynamic response and hypersensitivity reactions. This comprehensive meta-analysis of pooled data from patients in three clinical studies of laronidase (including one study with an extension) was undertaken to provide a more robust assessment of the relationship between the ADA response to laronidase, clinical and laboratory markers of MPS I, and hypersensitivity reactions. The meta-analysis demonstrated an inverse relationship between the ADA response and the percent reduction in urinary GAG (uGAG) levels. However, no relationships between the ADA response and changes in percent predicted forced vital capacity and six-minute walk test were seen. The study also re-assayed stored serum samples from the original trials with a novel method to determine the inhibitory effect of ADA. Patients with higher ADA exposure over time were found to have higher inhibition of enzyme uptake into cells. High ADA exposure can result in a commensurate level of enzyme uptake inhibition that decreases the pharmacodynamic effect of the exogenously administered therapeutic enzyme, but with no clear effect on clinical efficacy.

© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

CrossMark

Article history:

Received 19 December 2015

Received in revised form 18 February 2016

Accepted 18 February 2016

Available online 20 February 2016

Keywords:

Antibody

Enzyme replacement therapy Laronidase

Mucopolysaccharidosis type I Urinary glycosaminoglycan

1. Introduction

Mucopolysaccharidosis type I (MPS I) is one of several lysosomal storage diseases for which enzyme replacement therapy (ERT) is safe and effective [15]. MPS I is a clinical spectrum of disease that encompasses severe (Hurler syndrome) and attenuated (Hurler-Scheie and Scheie syndromes) forms without clear delineation. In clinical trials,

Abbreviations: % Predicted FVC, percent of predicted normal forced vital capacity; 6MWT, six-minute walk test; ADA, anti-drug antibody; AE, adverse events; AUC/time, area under the curve over time (reflects a weighted average of exposure over a period from baseline through a specific timepoint - see Section 2.4); GAG, glycosaminoglycan; IV, intravenous; RIP, radioimmunoprecipitation; uGAG, urinary glycosaminoglycan; ULN, upper limit of normal.

* Corresponding author.

E-mail addresses: yong.xue@genzyme.com (Y. Xue), susan.richards@genzyme.com (S.M. Richards), asif.mahmood@genzyme.com (A. Mahmood), gerald.cox@genzyme.com (g.F. Cox).

intravenously administered recombinant human a-L-iduronidase (laronidase/Aldurazyme®, EC 3.2.1.76) demonstrated clinical benefit (improved pulmonary function and walking ability) while lowering markers of glycosaminoglycan (GAG) storage (liver volume and urinary GAG [uGAG] level) in patients with MPS 1 disease. As laronidase does not cross the blood-brain barrier, it is used to treat the non-neurological symptoms of MPS I, including patients with Hurler and Hurler-Scheie syndromes, and Scheie syndrome with moderate-to-severe symptoms. For Hurler patients younger than two years of age and developmental quotients above 70, hematopoietic stem cell transplantation (HSCT) is considered the standard of care [9]. Based on data from the MPS I Registry, laronidase has been used as an adjunctive therapy in the peri-transplant period in over 90% of cases since 2007, to improve the clinical status of patients prior to HSCT and to reduce GAG storage prior to en-graftment [4].

Most ERTs produce an anti-drug antibody (ADA) response, which in some cases can reduce efficacy or lead to hypersensitivity reactions

http://dx.doi.org/10.1016/j.ymgme.2016.02.006

1096-7192/© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

[1,19]. In clinical trials of MPS I, more than 90% of treated patients developed IgG antibodies to laronidase during the first few months of treatment [21]. (Hereafter, anti-drug IgG antibodies will be referred to as ADA and anti-drug IgE antibodies will be referred to as IgE ADA.) In these studies, higher ADA titers were associated with less uGAG reduction, despite overall evidence of clinical benefit [3,22].

Development of an immune response has the potential to impair the desired biological effects of the therapeutic enzyme through several means, including altered enzyme targeting, increased enzyme turnover, and/or inhibition of enzyme activity [1]. Data from a canine model of MPS I suggest that ADA may impair enzyme uptake in tissues, leading to less tissue GAG clearance and greater uGAG excretion [5]. In patients, the effects of ADA on tissue GAG clearance are difficult to assess since biopsies from clinically relevant organs and tissues cannot readily be performed. However, data from clinical studies of laronidase have suggested that the reduction in uGAG excretion is more robust in patients with low ADA titers, while patients with higher titers had more variable reductions in uGAG clearance [22]. This apparent correlation between the ADA response and uGAG clearance has been consistently observed in multiple clinical studies, despite differences in the patient population, dose and duration of treatment, and method for determination of ADA titer [7,21,22]. The impact of the ADA response on safety and efficacy of ERT with laronidase continues to be investigated.

This comprehensive, systematic meta-analysis of pooled data from patients in three clinical studies of laronidase (including one study with an extension) was undertaken to provide a more definitive assessment of the relationship between laronidase ERT, the ADA response, clinical outcome and laboratory measurements, and the potential for allergic reactions.

2. Methods

2.1. Studies

The designs of the three studies - ALID-003-99/ALID-006-01 (phase 3 study and long-term extension) [21], ALID-014-02 (under 5 study) [22] and ALID-017-03 (dose optimization study) [7] - are summarized in Table 1. ALID-003-99 was a multicenter, multinational, phase 3 trial of laronidase administered intravenously for 26 weeks in 45 patients, all of whom had attenuated MPS I. ALID-006-01 was a 182-week

extension study of all 45 patients who completed ALID-003-99. The two trials were treated as one study for this meta-analysis. ALID-014-02 was a 52-week, phase 2 trial to assess the safety, efficacy, and phar-macokinetics of laronidase in 20 patients with MPS I who were five years of age or younger at enrollment and who were not being considered for hematopoietic stem cell transplantation. ALID-017-03 was a phase 4, 26-week dose-optimization study, undertaken to determine whether alternative dosing regimens could further reduce lysosomal storage in 33 patients across the range of MPS I phenotypes. All patients enrolled and treated in all four studies had a confirmed deficiency of iduronidase activity and a clinical diagnosis of MPS I, and had not been previously treated with laronidase. Written informed consent, and assent if applicable, was obtained for all patients in all studies.

2.2. Patient population

This meta-analysis involved the pooling of data from 73 MPS I clinical trial patients who received at least one IV infusion of laronidase at the labeled dose (0.58 mg/kg of body weight per week [qw] [100 U/kg qw]). Where patients received multiple dose levels, only data from the time that they received the labeled dose were included. The number of patients from each study that was used in the meta-analysis is shown in Table 1, as are the demographics of the pooled population which spanned all three MPS I phenotypes. Results of genotyping for a-L-iduronidase mutations were available for 34 patients, either from the clinical study database or from the global MPS I Registry.

2.3. Seroconversion and anti-drug antibody titer

Periodic testing for ADA was performed for all patients in all four studies. ADA in patient serum samples was initially screened by enzyme-linked immunosorbent assay (ELISA) and reactive samples were confirmed by radioimmunoprecipitation (RIP) assay. Once a sample was confirmed to be positive, ADA levels were quantified either by absorbance at 450 nm (expressed as optical density [OD] units/|L) in study ALID-003-99/ALID-006-01, or by endpoint titration (expressed in titer) in studies ALID-014-02 and ALID-017-03. Titer represents the last serial dilution in which the signal is positive, whereas a single OD unit/|L can be associated with a limited range of titers. Therefore, stored serum samples from study ALID-003-99/ALID-006-01 were re-

Table 1

Summary of clinical trials included in meta-analysis.

Study # ALID-003-99 ALID-006-01a ALID-014-02 ALID-017-03

Phase 3 3 2 4

Randomized X X

Double blind, placebo-controlled X

Open-label X X X

Treatment duration (wks) 26 182b 52 26

Age in years, mean (range) 15.8 (6.3-43.3) 3.0 (0.7-5.1) 8.6 (2.9-17.2)

Total number treatedc 45d 20 33

# analyzed in meta-analysis 45 20e 8f

Gender, M/F 22/23 12/8 3/5

Race, W/B/A/O/unknown 37/0/2/2/4 18/1/0/1/0 4/1/0/2/1

Phenotype (H/H-S/S) 0/38 (84%)/7 (16%) 16 (80%)/4 (20%)/0 2 (25%)/4 (50%)/2 (25%)

a ALID-003-99 and ALID-006-01 were conducted under separate study protocols but are considered to be one study for the purpose of the meta-analysis.

b For patients randomized to active treatment in ALID-003-99, the total period of laronidase treatment was up to 208 weeks. Patients randomized to placebo in ALID-003-99 received the first infusion of laronidase in ALID-006-01.

c Includes only those enrolled patients who received at least 1 infusion of laronidase (at any dose). For ALID-017-03, the number in this row include some patients whose data were not included in the meta-analysis.

d In ALID-003-99,22 patients were randomized to laronidase and 23 patients were randomized to placebo. All 45 patients subsequently enrolled in ALID-006-01 and received laronidase in that study.

e Four of the 20 patients had uGAG levels > 200 |og/mg creatinine at week 22 and received a subsequent dose increase to 1.2 mg/kg per week (qw) (200 U/kgqw) between week 26 and week 30. Data for these 4 patients were analyzed only through week 26.

f Patients in this study were randomized to receive IV infusions of laronidase at a dose of 0.58 mg/kg qw, 1.2 mg/kg qw, 1.2 mg/kg every other week (qow), or 1.8 mg/kg qow (300 U/kg qow). Eight of the 33 patients treated in the study received laronidase at the labeled dose of 0.58 mg/kg qw.

assayed using endpoint titration in available samples (91% of stored samples).

2.4. Cumulative anti-drug antibody exposure: area under the curve/time

In addition to peak ADA (IgG) titer, each patient's overall exposure to ADA over time was assessed to provide a more continuous distribution of the data. A derived parameter, the ADA area under the curve/time (AUC/time), was determined for 26 and 52 weeks of laronidase treatment. AUC/time reflects a weighted average of ADA exposure over a period from baseline through a specific time point, where the weights represent the proportion of time between successive titer assessments. Although this approach has been used to evaluate the therapeutic immune response to vaccines [14,20], the use of this derived exposure parameter to evaluate the immune response to a therapeutic enzyme is novel [2], and there are limitations to its general use for determining the immunogenicity of biologics.

2.5. Enzyme inhibition

To better understand the biological basis of any observed effects of ADA, stored serum samples from each seropositive patient in the pooled population were screened for the presence of inhibitory ADA. At every time point that was positive for ADA, quasi-quantitative assays [19] were used to determine both the ADA's inhibition of enzyme activity (measured as a percent of inhibition of 4-methylumbelliferone laronidase activity in vitro) and its competitive inhibition of cellular uptake of laronidase (measured as an ADA titer). The latter assay uses flow cytometry to measure the dose-dependent uptake of a fluorophore (AlexaFluor-488)-conjugated form of laronidase into cultured human fibroblast cells expressing the mannose-6-phosphate receptor (M6PR), a cell surface receptor known to mediate cellular uptake of laronidase. Two-fold serially diluted patient sera, starting with a minimal required dilution (MRD) of 1/10, were incubated with a constant amount of AF488-conjugated laronidase and then added to M6PR-expressing cultured fibroblast cells, and cellular uptake was measured as the median fluorescence intensity (MFI) by flow cytometry. The presence of inhibitory antibodies was detected as a reduction in MFI values at the chosen dose (the AF488-laronidase concentration that gives 50% of the maximal response), and the magnitude of the inhibitory antibody response was determined by endpoint titration of the patient samples. Samples having an enzyme uptake inhibition greater than the assay cutoff at two or more sera dilutions were considered positive, and endpoint titer was defined as the reciprocal of the highest serum dilution where the inhibition signal was greater than the assay cutoff established by human serum studies.

2.6. Urinary glycosaminoglycan level

Baseline uGAG data for 72 of the 73 patients in the pooled analysis population were available for the meta-analysis. Urinary GAG excretion, a commonly used pharmacodynamic biomarker of the treatment effect in MPS disorders, was measured in first morning void samples approximately every 4 weeks through week 26 in studies ALID-017-03 and ALID-003-99/ALID-006-01; every 3 months thereafter in study ALID-003-99/ALID-006-01; and at baseline and weeks 13, 22, 26, and 52 in study ALID-014-02. Urinary GAG levels were quantitated using a validated, automated dimethylmethylene blue dye-binding procedure in a central laboratory (BioMarin Pharmaceutical Inc., data on file) and expressed as ^g/mg creatinine. Patient uGAG values were defined as normal if they were less than or equal to the upper limit of normal (ULN) for age in healthy volunteers. In healthy volunteers, the ULN urinary GAG levels (expressed as |ag GAG/mg creatinine) decreased with age as follows: >6-12 months, 133; > 1-2 years, 110; >2-3 years, 93; >3-6 years, 72; >6-12 years, 52; > 12-20 years, 27; >20 years, 15) [7].

2.7. Statistical methods

For all patients, data were analyzed from baseline, defined as the measurement prior to the first infusion of laronidase, through completion of the study. Data were reported for the following pooled visit time points relative to the first infusion of laronidase, as appropriate for each study cohort: week 0 (baseline), weeks 1,4, 8,12,16, 20, 26, 38, 52, 60, 72, 84, 96,108, 120, 132,144, 156, 168, 182,194, and 208. For ADA, uGAG and clinical outcomes data, actual study visits for each study cohort were matched to pooled visits using windowing rules to capture the most data from related time points. Based on these windowing rules, actual study visits were within — 4 weeks to + 2 weeks of the matched pooled visit.

Seroconversion was a dichotomous variable indicating whether a patient had any ADA titer (seroconversion = Yes) or not (seroconversion = No). Time to seroconversion was calculated (in weeks) as the time from first infusion of laronidase to the first occurrence of an ADA titer.

For the analysis of ADA titer, baseline antibody titers were set equal to zero. Patients who were seronegative (i.e., negative in the RIP confirmatory assay) at any analysis timepoint had their ADA titer value for that time point set equal to zero (i.e., the RIP result was reported as "N/A") or set to missing (if the RIP result was reported as "NT" [Not Tested]). Titer values preceded by the symbol "<" were analyzed as absolute titer values (e.g., "< 100" was treated as "100"), whereas titer values reported as "< MIN QUANT" were set equal to zero.

Peak ADA titer was defined as the maximum ADA titer after the first infusion of laronidase, and time to peak ADA titer was calculated (in weeks) from the first infusion of laronidase to the first occurrence of the peak ADA value. Patients who did not seroconvert had their peak ADA titer value set to zero. For the meta-analysis of peak ADA titer data, patients were categorized into one of four peak ADA titer groups: Group 1 (no seroconversion): 3 patients (4.11%); Group 2 (100 — 3200): 16 patients (21.92%); Group 3 (6400-25,600): 29 patients (39.72%); Group 4 (51,200-204,800): 25 patients (34.25%).

Inhibition of enzyme uptake endpoint titer exposure over time (AUC/time) was computed over 26 weeks and 52 weeks for each patient. Only one patient had positive results for inhibition of enzyme activity.

Seroconversion, ADA titer, peak ADA titer, and inhibition of enzyme uptake endpoint titer were summarized by study cohort and over the pooled population. Mean ADA titers were plotted over time by study cohort. Plots of individual ADA titers over time were also generated. Plots for ADA titer AUC/time and inhibition of enzyme uptake titer AUC/time were generated over 26 and 52 weeks of treatment. Pearson correlation coefficients and p-values were calculated for selected scatter plots with individual patient data.

The relationships between uGAG and ADA exposure, ADA titer, and peak ADA titer were examined at selected timepoints, for each study cohort. Results for the six-minute walk test (6MWT) distance, percent of predicted normal forced vital capacity (% predicted FVC), and liver volume were descriptively summarized over the pooled population.

2.8. Clinical outcomes 2.8.1. Efficacy

The clinical efficacy of laronidase in the pivotal study (ALID-003-99) was evaluated by use of the co-primary endpoints 6MWT [6] and % predicted FVC [17]. Liver volume [10] was a secondary endpoint reflecting lysosomal storage. In study ALID-003-99/ALID-006-01, the 6MWT distance (in meters) and % predicted FVC were measured approximately every four weeks from baseline through week 26, and every 12 weeks thereafter. In study ALID-017-03, the 6MWT (but not % predicted FVC) was measured at baseline, week 12, and week 26. Liver volume, expressed as a percentage of body weight to account for changes due to patient growth, was measured by MRI at baseline and week 26 for

patients in studies ALID-003-99/ALID-006-01 and ALID-017-03. Liver data from study ALID-014-02 were not included in the pooled analysis because liver size measurements in that study of very young patients were estimated by manual palpation rather than by MRI due to the risk of anesthesia in pediatric patients with compromised airways.

2.8.2. Safety/hypersensitivity

Patients were observed for Adverse Events (AEs) during and for at least 30 min after completion of each laronidase infusion. Observed AEs assessed as related to laronidase administration were reported as infusion-associated reactions (IARs) and, for the purposes of this meta-analysis, were considered potential allergic reactions. In addition, any AE was identified for this meta-analysis as a potential allergic reaction if it either met the protocol definition for IARs, or occurred within one day of the infusion and was assessed as laronidase-related but not as an IAR. All such AEs were reviewed to determine if they were consistent with labeled IARs for Aldurazyme. For patients experiencing moderate or severe IARs, samples were to be drawn soon after the occurrence of the IAR for testing of IgE ADA, complement activation, and serum tryptase levels.

3. Results and comparative analyses

3.1. Anti-laronidase IgG antibody response

All 73 patients in the pooled population were seronegative at baseline, and nearly all (70, or 95.9%) seroconverted during treatment with no marked differences in seroconversion rates by disease syndrome or genotype. The median time to seroconversion was 4.1 weeks (range 2.1-37.7). Patients with Hurler syndrome tended to seroconvert earlier than patients with attenuated MPS I (median 3.1 vs. 6.7 weeks, respectively).

There was substantial variability in ADA titers observed among patients. Titers tended to be slightly higher on average for Hurler patients, particularly those with two nonsense mutations, than for patients with attenuated MPS I. Consequently, mean ADA titers in study ALID 014-02 which enrolled predominantly Hurler patients, were slightly higher than those in studies ALID-017-03 and ALID 003-99/ALID-006-01. The mean peak ADA titer for all patients in the pooled population was 31,972. (Peak titer was set to zero for the 3 patients who did not seroconvert; these patients were included in the analysis.)

Sixteen patients had consistently low ADA titers (<3200) and 25 patients had high peak titers (> 51,200). The median time from the first infusion to peak titer was 11.1 weeks (range: 2.9 to 120.9). Thereafter, ADA titers decreased gradually over time and ultimately fell to low or undetectable levels in most patients in ALID-003-99/ALID-006-01.

Approximately half of the patients (52.1%) developed inhibitory ADA that interfered with enzyme uptake into cells, but ADA that inhibited enzyme activity were detected in the serum from only one patient in the pooled analysis population. Overall, inhibitory ADA titers remained low throughout treatment for nearly all patients. At week 26, all patients testing positive for inhibitory ADA had enzyme uptake inhibition titers of 20 or 40, with the exception of one patient with an enzyme uptake inhibition titer of 80. Since titer is not a quantitative assessment, an evaluation of the percent inhibition at the MRD was done. For patients in ALID-017-03, a titer of 20 reflected on average 31.7% (±9.3) inhibition of enzyme uptake, a titer of 40 averaged 43.1% (±7.4) inhibition, and a titer of 80 averaged 57.8% (±11.2) inhibition. Patients in ALID-014-02 had slightly higher percent inhibition at the MRD, on average, a 55.5% (±11.7) inhibition for a titer of 20, 59.1% (±7.9) inhibition for a titer of 40, and 68.9% (±8.9) inhibition for a titer of 80. The greater number of Hurler syndrome patients in ALID-014-02 likely accounts for the higher inhibition values relative to ALID-017-03. In general, the percent inhibition was not directly linear with dilution, further indicating that antibody assessments represent a composite readout of the nature of ADA and potency of binding

interactions incorporating relative concentration, antibody affinity, and patient matrix effects. The majority of tested patients receiving long-term treatment with laronidase (in ALID-003-99/ALID-006-01) tested negative for enzyme uptake inhibition at the completion of treatment.

3.2. IgG antibodies vs. Inhibitory antibody exposure

A positive relationship between inhibition of enzyme uptake titer and ADA titer was seen at week 26 (see Fig. 1), with similar findings for weeks 12 and 52. As the IgG antibody titer increased, the inhibition of enzyme uptake into cells also increased.

There were moderate to strong correlations between ADA titer and inhibition of enzyme uptake titer when both were measured by cumulative exposure over time (r = 0.476 and p = 0.0022 [r = 0.756 and p < 0.00001 if the two outliers are excluded] for baseline through week 26 - see Fig. 2; r = 0.823 and p < 0.0001 for baseline through week 52).

Overall, inhibition of enzyme uptake titers showed more variability over time than ADA titers.

3.3. Urinary glycosaminoglycan levels

All patients had elevated uGAG levels for age at baseline, and all showed reduced uGAG levels following laronidase treatment. Generally, there was an initial sharp reduction within the first few weeks followed by only small additional decreases. Most patients with uGAG data at early time points achieved a substantial reduction by week 8 (— 64.7 ± 12.26%, n = 27) with plateauing at week 26 and later time points (— 59.9 ± 16.16%, n = 71). In ALID-003-99/ALID-006-01, the mean percent reduction at week 182 was slightly greater ( — 75.0 ± 12.65%, n = 37).

The three patients from ALID-003-99/ALID-006-01 who did not se-roconvert (peak titer Group 1) showed uGAG reductions of 86.8%, 62.9%, and 74.0% over the course of 182 weeks of treatment. Twenty-four of the 73 patients (32.8%) in the pooled analysis population achieved normal uGAG levels at one or more time points.

Fig. 1. Inhibition of enzyme uptake titer versus IgG anti-drug antibody (ADA) titer. A positive relationship between inhibition of enzyme uptake titer and ADA titer was seen at week 26 for the pooled analysis population (n = 23). Data represent individual patients from all three trials, including the extension trial.

>> t go

100 60

<5 Q. 3

O o o O 0 o o

OO&JDO

0 20 40 60 80 100

IgG Antibody Titer (AUC/Time x103)

Fig. 2. Inhibition of enzyme uptake titer area under the curve (AUC)/time versus IgG antidrug antibody (ADA) titer AUC/time. As measured by cumulative exposure from baseline through week 26, ADA titer and inhibition of enzyme uptake titer for the pooled population showed a correlation (r = 0.476, p = 0.0022; r = 0.756, p < 0.00001 without outliers). Data represent individual patients from all three trials (including the extension trial).

Fig. 3. Relationship between percent reduction in urinary glycosaminoglycan (uGAG) level and antibodies. A. Percent reduction in uGAG level versus IgG anti-drug antibody (ADA) titer was examined graphically via scatter plots as the percent reduction from baseline in uGAG versus log10 ADA titer for the pooled population. In this graph, each data point represents an individual patient at week 26. A moderate inverse correlation (r = — 0.460, p = 0.0001) is seen. B. Percent uGAG level versus ADA titer AUC/time. At week 26, there was an inverse correlation between percent reduction in uGAG level and ADA titer AUC/Time for the pooled population. Data represent individual patients from all three trials, including the extension trial.

3.4. Urinary glycosaminoglycans and antibodies

At week 26, there was a moderate inverse correlation (r = — 0.460, p = 0.0001) between percent reduction in uGAG levels and ADA titer (see Fig. 3A). Patients with absent or low ADA titers had robust uGAG clearance, generally exceeding 50% reduction, while patients with high titers had more variable uGAG clearance, with similar numbers of patients above and below 50% reduction. Similar findings were observed at weeks 12 and 52, and when uGAG was compared to ADA titer AUC/ time at weeks 26 (see Fig. 3B) and 52.

The relationship between uGAG clearance and ADA titer was also analyzed by disease syndrome (data not shown). Similarly to the data presented in Fig. 3, the results observed for patients with attenuated MPS I at weeks 12, 26, and 52 showed an inverse correlation; the trend was less evident at weeks 96, 144, and 182 and not evident in patients with Hurler syndrome.

Percent reductions in uGAG level at the time of peak titer were variable for patients in each peak titer group. However, there was a trend towards a smaller reduction in uGAG level as the peak ADA titer increased, particularly in Titer Group 4. Patients in Titer Group 4 also were less likely to have achieved a normal uGAG level during treatment with laronidase. Of the 73 patients in the pooled analysis population, 24 patients achieved normal uGAG levels at one or more time points during the study; however, given the limited data, these results should be interpreted with caution. Overall, the percent uGAG reduction versus peak titer results corroborated the finding that higher ADA titers were associated with less robust uGAG reduction. Since titers decreased over time in Study ALID-006-01, eventually there ceased to be any apparent relationship between antibody titer and uGAG clearance based on that study's data.

While there was an inverse relationship between percent reduction in uGAG level and inhibition of enzyme uptake titers at weeks 12, 26 and 52, the interpretation of these data is limited by the narrow range of titers across patients. However, there was a moderate inverse correlation between percent reduction in uGAG level and cumulative exposure over time to ADA that inhibit enzyme uptake through weeks 26 (r = —0.479, p = 0.0023) and 52 (r = — 0.485, p = 0.0119). The correlation through week 26 was statistically significant only after excluding data from an outlier who had a substantial reduction in uGAG level at week 26 despite a high (115) inhibition of enzyme uptake titer AUC/ time.

3.5. Clinical outcomes

3.5.1. Efficacy

After 26 weeks of laronidase treatment, the mean 6MWT distance increased by 20 m and the mean % predicted FVC decreased by 0.9% (absolute change) compared to baseline. Values for both the 6MWT and % predicted FVC varied substantially between patients, at baseline and over the course of treatment. For this reason, the meta-analysis also examined the potential impact of ADA on liver volume. After 26 weeks of treatment, mean liver volume decreased by 22.3%.

3.5.2. Safety/hypersensitivity

Thirty-eight (52%) of the patients in the pooled analysis population experienced one or more AEs that were identified as potential allergic reactions. A 17-year-old male patient with Hurler-Scheie syndrome from ALID-003-99/ALID-006-01 experienced an AE of anaphylaxis; this patient required emergency tracheotomy for an associated airway obstruction at week 62, after 44 laronidase infusions [21]. The patient recovered but continued to require ventilatory support via a tracheotomy at the time of study discontinuation. This patient had experienced severe IARs of dyspnea at week 34 and hypoxia at week 50. Among the 10 patients tested for IgE ADA following moderate or severe IARs, this was the only patient who was IgE-positive (at weeks 38, 52 and 60).

This patient seroconverted for ADA 8 weeks after receiving his first laronidase infusion, with a titer of 25,600 [3].

Nine patients were tested for complement activation, and four were found to be positive. Serum tryptase samples were obtained in five patients, and all results were normal.

3.6. Efficacy and antibodies

Absolute changes from baseline in the 6MWT and % predicted FVC at week 26 were compared with peak ADA titer by time point, with patients classified into the above-described titer groups. No apparent relationship was seen for either clinical measure, at weeks 26 (see Fig. 4A for 6MWT and Fig. 5A for % Predicted FVC) and 52.

In addition, no consistent relationship was observed between ADA cumulative exposure through week 26 and the change from baseline in 6MWT distance (see Fig. 4B), % predicted FVC (see Fig. 5B), and liver volume. Similar findings were obtained for cumulative ADA exposure through week 52 versus the change from baseline in 6MWT distance, % predicted FVC, and liver volume.

Fig. 5. Relationship between percent of predicted normal forced vital capacity (% predicted FVC) and antibody titer. Change from baseline in % predicted FVC versus peak IgG antidrug antibody (ADA) titer (A) and ADA titer area under the curve (AUC)/time (B). No apparent relationship was seen between the change from baseline in % predicted FVC and peak ADA titer or AUC/time at week 26 for the pooled population. Data represent individual patients for study ALID-003-99/ALID-006-01 only.

Most enzyme uptake inhibition titers were either 20 or 40 and, as expected due to the narrow range of values, no relationships with clinical outcome were found. Subsequent correlative analyses were performed using AUC/time exposure which provided a more dynamic range. No consistent relationship was observed between exposure to ADA that inhibit enzyme uptake and the change in % predicted FVC or liver volume at weeks 26 and 52. However, a marginally significant inverse correlation was observed for the 6MWT at week 26 (r = — 0.417, p = 0.0476) although not at week 52 (r = — 0.072, p = 0.7674).

3.7. Potential allergic reactions and antibodies

A small number of patients experienced potential allergic reactions at each pooled visit time point. There were no consistent trends in ADA titer between patients who had potential allergic reactions and those who did not.

4. Discussion

Fig. 4. Relationship between six-minute walk test (6MWT) distance walked and antibody titer. Change from baseline in 6MWT versus peak IgG anti-drug antibody (ADA) titer (A) and ADA titer area under the curve (AUC)/time (B). There was no apparent relationship between the change from baseline in 6MWT distance walked and peak ADA titer or AUC/time at week 26 for the pooled population. Data represent individual patients from trials ALID-003-99/ALID-006-01 and ALID-017-03.

This meta-analysis utilized the largest data set to date, from three studies involving 73 patients of various ages and MPS I phenotypes, to analyze the relationships among laronidase treatment, ADA, clinical outcome, and potential allergic reactions. The data demonstrated a clear relationship between exposure to ADA and a pharmacodynamic

biomarker, uGAG, but not with clinical outcomes, including 6MWT, % predicted FVC, and liver volume. Significant inverse correlations were observed between percent reduction in uGAG level and ADA titer as well as AUC/time for ADA titer and enzyme uptake inhibition titer, but not enzyme activity. More robust reductions in uGAG levels were seen in patients with lower ADA titers. This inverse relationship was clear in patients with attenuated Hurler-Scheie and Scheie forms of MPS I. In contrast, patients with higher titers - including most of the Hurler patients - showed more variable and generally less uGAG reduction. This meta-analysis was limited to data from the time that patients received the labeled dose of laronidase, 0.58 mg/kg (100 U/kg). However, it is noteworthy that, in the under 5 study (ALID-014-02), the four Hurler patients who received twice the labeled dose weekly (i.e. 200 U/kg, administered after the week 24 uGAG response was considered suboptimal), although excluded from the meta-analysis, had among the highest percent uGAG reductions at week 52, suggesting that the higher dose may have overcome the effect of the ADA.

On the other hand, no clear or consistent relationship was seen between the clinical efficacy of laronidase ERT and ADA response. Based on the drug's mechanism of action, such physiologic assessments that are far-removed from the underlying metabolic defect may prove more difficult to influence than pharmacodynamic parameters that directly reflect the underlying metabolic defect. It can be challenging to determine precise associations with functional parameters (such as % predicted FVC and 6MWT) that can be influenced by multiple factors, including interactions of different organ systems, developmental changes, and secondary pathophysiologic effects, such as fibrosis. These factors, undoubtedly, contribute to disease heterogeneity as evidence by the broad phenotypic expression and differing rates of progression.

There are several other possible explanations for the apparent inconsistency between the ADA impact on pharmacodynamics and clinical efficacy. The pharmacodynamic response measured by uGAG excretion, which is thought to be derived from tubular epithelial cells from kidney, may not reflect GAG clearance in other patient tissues. Other published reports have demonstrated a pharmacodynamic impact of antibodies on uGAG excretion but without a discernable clinical effect [1,12]. Kidney may be more sensitive to the effects of ADA than other tissues. Of note, the liver is a major source of GAG storage and its response to treatment did not appear to be impacted by ADA in this meta-analysis. While uGAG excretion did correlate with tissue GAG clearance in a preclinical canine model, the ADA levels were much higher than those observed in patients, which may have led to a greater observed effect [5]. Urine is a readily available source of GAG, but it might not be the most clinically relevant biomarker because MPS I patients do not manifest renal disease. Measurement of GAG in other fluid compartments (e.g. blood), might be more informative but historically has been difficult to measure owing to its low concentration in the circulation; newer techniques employing tandem mass-spectrometry or measurement of pathologic GAG through their non-reducing ends could help to address this possibility.

It is also possible that the clinical endpoints examined in this meta-analysis were inherently too variable to permit the observation of a relationship. However, the findings of this meta-analysis are supported by a recent evaluation of cross sectional data from 24 patients who received laronidase ERT over 12 months in two prospective open-label studies [12]. A sustained ADA response was found to be possibly related to an impaired biomarker response but no correlation between uGAG levels and clinical outcome was found. In that publication, the clinical outcome scores included cardiac and pulmonary function, 6MWT distance, surgical interventions, growth, infections, cognitive development, obstructive sleep apnea, and hearing problems, together with patients' overall clinical condition evaluated each year as stable, improved, or worsened as compared to the previous year. Unlike bio-markers, these long-term clinical outcomes might not adequately reflect acute changes. On the other hand, an association between ADA

formation and reduction in sleep apnea was recently proposed based on a comparison of responses between patients receiving laronidase versus hematopoietic stem cell transplantation [1]. The ADA and its impact on clinical outcome was investigated by Pal and coworkers [16]. Obstructive sleep apnea measured by ODI4% (oxygen desaturation index) was more advanced in patients with higher levels of inhibitory antibodies. Given the small sample size and retrospective nature of data collection, further investigation is warranted using sleep oximetry to evaluate immunogenicity and its impact on clinical outcome.

The results of this meta-analysis do not change the risk-benefit profile of laronidase, which has shown clinical benefit despite the presence of ADA. In patients with attenuated disease, the ADA generally decline or are no longer detectable over time, indicating clinical tolerization.

The meta-analysis aggregated data from separate studies while retaining each patient's study number as a means of categorization. A post hoc re-evaluation classifying patients by clinical diagnosis of disease phenotype, i.e. Hurler, Hurler-Scheie, or Scheie also failed to demonstrate a clear relationship between ADA titers and clinical outcome or the potential for allergic reactions. As neither 6MWT nor % predicted FVC testing was performed in the original ALID-014-02 trial due to age limitations, no data for patients with Hurler syndrome are available to correlate ADA titers with these clinical parameters.

It should be noted that findings regarding the nature of the relationship between ADA and ERT efficacy differ among lysosomal storage diseases. In Gaucher disease type 1, relatively few patients develop an immune response to treatment, and no correlation has been shown between anti-imiglucerase ADA and treatment efficacy [18]. In Pompe disease, on the other hand, evidence suggests that high titers of anti-alglucosidase alfa ADA are associated with a poorer clinical response to treatment and that reducing the ADA response using immunosup-pressant agents improves the efficacy of ERT [11]. In Fabry disease, patients who developed anti-agalsidase alpha or beta ADA during ERT showed significantly less reduction in urinary globotriaosylceramide than those who were ADA-negative [13].

In a phase 3 randomized placebo-controlled study in patients (n = 176) with Morquio A syndrome (MPS IVA), the ADA and its impact on efficacy and safety of ERT with elosulfase alfa was studied. Despite the high incidence of ADA, decreases in uGAG (keratin sulfate) excretion and improvements in 6MWT were observed in treated subjects. No correlations were detected between higher total antibody titers or neutralizing antibody (antibodies that neutralize the effects of antigens or infectious bodies) positivity and reduced 6MWT results or increased incidence of hypersensitivity AEs [8].

ADA formation frequently occurs in ERT administered for lysosomal storage diseases. Patient disease enzyme mutations influence the residual enzyme produced, both in terms of protein structure and concentration, and such factors influence the magnitude of the ADA response to ERT. However, the impact of ADAs can vary among diseases and among individual patients. More research is needed to evaluate the impact of ADA in proper clinical settings, with standardized assays and pharmacodynamic biomarkers that can be correlated with clinical outcomes across a spectrum of patients and clinical centers.

4.1. Conclusion

This meta-analysis provides a comprehensive, systematic review of immunogenicity data across four clinical studies of laronidase, including those with long-term follow-up, and involves patients with all three MPS I phenotypes. The results presented herein show a statistically significant inverse correlation between percent reduction in uGAG level and ADA titer. However, the presence of ADA had no significant impact on clinical outcomes or the occurrence of hypersensitivity reactions. Overall, these results do not alter the interpretation of previous analyses, and pose no change to the risk-benefit profile of laronidase.

Author disclosures

All authors are employees of Sanofi Genzyme.

Acknowledgements

Katherine Kacena, and Jonathan Fallet participated in the preparation of this manuscript. In addition, the authors would like to acknowledge the invaluable contributions of the patients, investigators, and laboratory and staff members involved in the studies included in this meta-analysis: AL1D-003-99/AL1D-006-01 [21], AL1D-014-02 [22] and AL1D-017-03 [7].

References

[1 ] B.W. Bigger, M. Saif, G.E. Linthorst, The role of antibodies in enzyme treatments and therapeutic strategies, Best Pract. Res. Clin. Endocrinol. Metab. 29 (2015) 183-194.

[2] P.L. Bonate, C. Sung, K. Welch, S. Richards, Conditional modeling of antibody titers using a zero-inflated poisson random effects model: application to Fabrazyme®, J. Pharmacokinet. Pharmacodyn. 36 (2009) 443-459.

[3] L. Clarke, J.E. Wraith, M. Beck, et al., Long-term efficacy and safety of laronidase in the treatment of mucopolysaccharidosis 1, Pediatrics 123 (2009) 229-240.

[4] M.H. de Ru, J.J. Boelens, A.M. Das, et al., Enzyme replacement therapy and/or hematopoietic stem cell transplantation at diagnosis in patients with mucopolysaccharidosis type 1: results of a European consensus procedure, Orphanet. J. Rare Dis. 6 (2011) 55-63.

[5] P. Dickson, M. Peinovich, M. McEntee, et al., 1mmune tolerance improves the efficacy of enzyme replacement therapy in canine mucopolysaccharidosis 1, J.Clin. 1nvest. 118 (2008) 2868-2876.

[6] P.L. Enright, The six-minute walk test, Respir. Care 48 (2003) 783-785.

[7] R Giugliani, V.M. Rojas, A.M. Martins, et al., A dose-optimization trial of laronidase (Aldurazyme) in patients with mucopolysaccharidosis 1, Mol. Genet. Metab. 96 (2009) 13-19.

[8] C.J. Hendriksz, B. Burton, T.R. Fleming, et al., Efficacy and safety of enzyme replacement therapy with BMN 110 (elosulfase alfa) for Morquio A syndrome (mucopolysaccharidosis 1VA): a phase 3 randomised placebo-controlled study, J. 1nherit. Metab. Dis. 36 (2014) 979-990.

[9] E. Jameson, S. Jones, J.E. Wraith, Enzyme replacement therapy with laronidase (Aldurazyme®) for treating mucopolysaccharidosis type I, Cochrane Database Syst. Rev. 11 (2013), CD009354.

10] E.D. Kakkis, J. Muenzer, G.E. Tiller, et al., Enzyme-replacement therapy in mucopolysaccharidosis I, N. Engl. J. Med. 344 (2001) 182-188.

11] P.S. Kishnani, P.C. Goldenberg, DeAmey Sl, et al., Cross-reactive immunologic material status affects treatment outcomes in Pompe disease infants, Mol. Genet. Metab. 99 (2010) 26-33.

12] E.J. Langereis, N. van Vlies, H.J. Church, et al., Biomarker responses correlate with antibody status in mucopolysaccharidosis type I patients on long-term enzyme replacement therapy, Mol. Genet. Metab. 114 (2015) 129-137.

13] G.E. Linthorst, C.E. Hollak, W.E. Donker-Koopman, A. Strijland, J.M. Aerts, Enzyme therapy for Fabry disease: neutralizing antibodies toward agalsidase alpha and beta, Kidney Int. 66 (2004) 1589-1595.

14] B.A. Martell, E. Mitchell, J. Polinig, K. Gonsai, T.R. Kosten, Vaccine pharmacotherapy for the treatment of cocaine dependence, Biol. Psychiatry 58 (2005) 158-164.

15] Neufeld EF. Enzyme replacement therapy—a brief history. In: Mehta A, Beck M, Sunder-Plassmann G, editors., eds. Fabry Disease: Perspectives from 5 Years of FOS. Oxford, England: Oxford PharmaGenesis; 2006.

16] A.R. Pal, E.J. Langereis, M.A. Saif, et al., Sleep disordered breathing in mucopolysaccharidosis I: a multivariate analysis of patient, therapeutic and metabolic correlators modifying long term clinical outcome, Orphanet. J. Rare Dis. 10 (2015) 42-54.

17] G. Polgar, P. Varuni, Pulmonary Function Testing in Children: Techniques and Standards, first ed. Saunders, Philadelphia: PA, 1971 100.

18] K. Starzyk, S. Richards, J. Yee, S.E. Smith, W. Kingma, The long-term international safety experience of imiglucerase therapy for Gaucher disease, Mol. Gnet. Metab. 90 (2007) 157-163.

19] J. Wang, J. Lozier, G. Johnson, et al., Neutralizing antibodies to therapeutic enzymes: considerations for testing, prevention and treatment, Nat. Biotechnol. 26 (2008) 901-908.

20] D. Williamson, H.C. Flick-Smith, C. Lebutt, et al., Human immune response to a plague vaccine comprising recombinant F1 and V antigens, Infect. Immun. 73 (2005) 3598-3608.

21] J.E. Wraith, L.A. Clarke, M. Beck, et al., Enzyme replacement therapy for mucopolysaccharidosis I: a randomized, double-blinded, placebo-controlled, multinational study of recombinant human a-L-iduronidase (laronidase), J. Pediatr. 144 (2004) 581-588.

22] J.E. Wraith, M. Beck, R. Lane, et al., Enzyme replacement therapy in patients who have mucopolysacchardiosis I and are younger than 5 years: results of a multinational study of recombinant human a-L-iduronidase (laronidase), Pediatrics 120 (2007) e37-e46.