In vitro studiesGLA transgene expression and characterization in cultured hepatocytes

To evaluate transduction efficiency of FLT190 vectors in vitro, Huh7 cells (Fig. 1A) and human primary hepatocytes (Fig. 1B) were transduced at different MOIs (1 × 104 vg/cell, 1 × 105 vg/cell, and 1 × 106 vg/cell) and the activity level of α-Gal A secreted in the media was assessed by an enzyme assay at 48 h post-transduction. For Huh7 cells, a 15-point dose-response curve (0 to 1 × 106 vg/cell) was constructed. The effective concentration that gives 50% of the maximal response (EC50) was determined as a measure of the potency of vectors expressing the GLA transgene. A dose-dependent increase in α-Gal A secretion was observed in both cell types. Huh7 cells showed a non-linear dose-response in kinetics of α-Gal A expression with EC50 value of 5 × 104 vg/cell and maximal activity level of 47 nmol/hr/mL. For human primary hepatocytes, three FLT190 MOIs were tested. A dose-dependent effect on the kinetics of α-Gal A expression was also observed, although expression was lower than observed in the Huh7 cell line.

Fig. 1: GLA transgene expression and characterization in cultured hepatocytes.

Kinetics of α-Gal A secretion following FLT190 vector transduction in A Huh7 hepatocyte cell line and B human primary hepatocytes. C α-Gal A secretion from Huh7 cells engineered to overexpress α-Gal A. D Glycosylation analysis of α-Gal A produced from Huh7 cells following PNGase F digestion – representative Western blot analyses are shown. Agalsidase alfa was used as a comparator. Data are mean ± SD (n = 3). α-Gal A = alpha galactosidase A; AAV = adeno-associated virus; Huh7 = human hepatocyte cell line; MOI = multiplicity of infection; PNGase F = Peptide:N-Glycosidase F; SD = standard deviation; vg = vector genomes.

A stable cell line, consisting of Huh7 cells engineered to overexpress α-Gal A from an integrated lentiviral vector (Lenti-CMV-eGFP-HPGK-GLAco), was also created as an additional α-Gal A producer cell line for secretion-uptake studies. The level of α-Gal A secretion was monitored from 0 to 77 h post-culturing (Fig. 1C), and a time-dependent increase in α-Gal A activity in the culture media was observed; activity levels at 24 h, 48 h, and 72 h were 77 nmol/hr/mL, 226 nmol/hr/mL, and 637 nmol/hr/mL, respectively.

The stable cell line has the same GLA codon-optimized transgene (GLAco) as the FLT190 expression cassette but driven by the CMV promotor. Regardless of the two delivery systems, the α-Gal A protein secreted from the stable cell line is expected to have the same glycosylation and maturation status as the AAV-transduced α-Gal A and will be a valuable tool for modelling in vitro uptake studies to characterize supraphysiological levels of exposure of the enzyme to the target cells and its efficiency on storage clearance.

The mature protein is comprised of two subunits of 398 amino acids (approximately 51 KD), each of which contains three N-linked glycosylation sites. To assess glycosylation status, supernatants of the culture media from the two expression systems were collected and treated with PNGase F before being analyzed by Western blot. As shown in Fig. 1D, the deglycosylation patterns and the shifting of bands were similar for both AAV- and lentiviral-vector produced α-Gal A and comparable to agalsidase alfa, which is currently in use as ERT in Fabry patients. These results confirmed that hepatic cell line-expressed α-Gal A was appropriately glycosylated similar to that produced by agalsidase alfa and can be used for uptake studies using Fabry-relevant cells.

α-Gal A secretion in apical and basal layers

FLT190 transduction in cultured Huh7 cells in the apical layer of the AAV transwell system (Fig. S1) was examined at two different MOIs (2.5 × 104 vg/cell, and 2 × 105 vg/cell) at Day 3 post-transduction. As shown in Fig. 2A, FLT190 transduction of Huh7 liver cells cultured apically led to a dose-dependent increase in recombinant α-Gal A protein secretion into culture media. Measurement of vg/cell in the apical transwell layer following FLT190 transduction showed a dose-dependent increase in transduced cells that was consistent with the dose-dependent increase α-Gal A enzyme secretion (Fig. 2B). Finally, Q-PCR analyses were conducted to evaluate whether any AAV viral particles leaked across the membrane into the basal transwell during co-culturing, which could yield false-positive results. Results of these analyses showed that only apical transduced Huh7 cells contained a high level of vector genome copies. These data demonstrate that co-cultured cells in basal transwells did not contain AAV viral particles after co-culturing.

Fig. 2: Uptake of α-Gal A by Fabry relevant target cells following transwell co-culture with FLT190 transduced Huh7 cells.

A α-Gal A activity was measured in the culture media of basal transwells following co-culturing of AAV transduced Huh7 cells. B vg copy numbers. C Colocalization of internalized α-Gal A and lysosomes. Representative confocal images of co-cultured knock-down cell lines co-stained with HA-tag and LAMP-1 antibodies. α-Gal A = alpha galactosidase A; AAV, adeno-associated virus; HA-tag = hemagglutinin tag; Huh7 = human hepatocyte cell line; LAMP-1 = lysosomal-associated membrane protein 1; vg = vector genomes.

To demonstrate α-Gal A enzyme secretion in the AAV and stable co-culture systems, the level of enzyme activity in the basal culture media was measured across multiple co-culturing experiments. A dose-dependent increase in α-Gal A activity was observed 48 h following AAV transduction in all experiments, with an average of a 2-to-3-fold increase from baseline in secreted α-Gal A activity levels (Fig. 1A). The level of secreted α-Gal A in the stable cell line co-culturing system was also evaluated, and an approximate 12.5-fold increase from baseline was observed 48 h after co-culturing (Fig. 1C). These data demonstrate that AAV transduction of apical Huh7 cells, and/or seeding of stable cell line, led to an increase in basal secretion of α-Gal A enzyme, with the latter providing higher levels of enzyme production.

Uptake of α-Gal A in key human cell linesConfocal microscopy

The internalization of α-Gal A into lysosomal compartments was evaluated by confocal microscopy 48 h after co-culturing in both wild-type and knockdown cell lines of kidney epithelial cells, kidney podocytes, cardiomyocytes, and Fabry disease patient fibroblasts. The subcellular localization of α-Gal A was determined by differential immunofluorescent staining of HA-tag [24] and lysosomal associated membrane protein-1 (LAMP-1) following transduction with FLT190 at 3 different MOIs (0 vg/cell [untreated], 2.5 × 104 vg/cell, and 2 × 105 vg/cell).

Confocal microscopy confirmed intracellular distribution of α-Gal A in all cocultured cell lines as detected by anti-HA tag antibody (red channel), and lysosomes were visualized in the perinuclear region by immunodetection of LAMP-1 (green channel) (Fig. 2D). The merger of HA tag and LAMP-1 signals showed co-localization of α-Gal A and lysosomes across all cell types. These data provide evidence that all the cell lines tested were capable of internalizing into the cytoplasm α-Gal A expressed from hepatocytes, and a proportion of internalized enzymes was directed to the lysosomes (Fig. S2).

Western blot and densitometric analyses

To demonstrate uptake of α-Gal A in co-cultured cells, we measured total cellular levels of α-Gal A protein in key cell lines by Western blot and densitometric analysis following co-culturing with the AAV system. At 48 h post-co-culturing, a dose-dependent increase in total cellular α-Gal A protein normalized to GAPDH in wild-type and knockdown kidney epithelial, kidney podocytes and cardiomyocyte cell lines, as well as Fabry fibroblasts, was observed compared with an un-transduced control (Fig. S3). A direct relationship was observed between secreted enzyme levels and protein uptake in each cell line.

Clearance of Gb3 storage from Fabry fibroblasts

Two different patient-derived Fabry fibroblast cell lines (81 and 107), patient-derived Fabry endothelial cells (IMFE1), and healthy (non-Fabry) fibroblasts were co-cultured with stable α-Gal A expressing Huh7 cells. Seventy-two hours post-co-culturing, densitometric analysis showed an increase in total cellular α-Gal A protein in these enzyme-deficient cell lines compared with untreated controls (Fig. S4A–C). The level of cellular α-Gal A protein in Fabry fibroblasts co-cultured with the stable cell line was found to be higher than those co-cultured with AAV-transduced Huh7 cells (20-fold increase versus 8-fold increase). This observation may be explained by the fact that Fabry fibroblasts were exposed to higher levels of α-Gal A via the stable cell line system (>150 nmol/hr/mL), as compared with AAV transduction (20 nmol/hr/mL), and further supports the finding that cellular exposure to higher levels of secreted enzymes leads to greater cellular uptake.

Tandem mass spectrometry was used to determine if uptake of transgene α-Gal A expressed from hepatocytes can functionally correct Gb3 lipid levels in cultured fibroblasts from Fabry patients. Gb3 lipid levels were measured in cell lysates of cultured fibroblasts exposed to α-Gal A following co-culturing as in Fig. S4A. As shown in Fig. S4D, the level of Gb3 in treated fibroblasts was normalized to a level similar to that observed in healthy controls. These data demonstrate that uptake of transgene-expressed α-Gal A can functionally correct excessive Gb3 accumulation in cultured Fabry fibroblasts.

In vivo studiesPharmacokinetics of AAV8-FLT190 in WT mice

AAVS3 was developed for increased human hepatocyte transduction, and is less efficient at transducing mouse hepatocytes; therefore, FLT190 genome (FRE1-GLAco) pseudotyped with an rAAV8 vector (AAV8-FLT190) was used for mouse studies. All AAV8-FLT190-treated animals exhibited a rapid dose-dependent increase in plasma α-Gal A activity by Week 2 and this high level of expression was maintained for the duration of the study (Fig. 3A, Table 1). The lowest dose of AAV8-FLT190 (6 × 1011 vg/kg) yielded an 830-fold increase in α-Gal A compared with the control group. Increasing the AAV8-FLT190 vector dose to 6 × 1012 vg/kg resulted in a marked elevation in plasma α-Gal A enzyme activity to 2500-fold greater than controls (Fig. 3A and Table 1). An analysis of area under the curve (AUC) in the AAV8-FLT190 dose groups showed a linear pharmacokinetic profile (R2 = 0.9901).

Fig. 3: α-Gal A activity in plasma following administration of AAV8-FLT190 or ERT in wild-type (WT) mice.

A Time course of α-Gal A enzyme activity levels in the plasma of WT mice that received AAV8-FLT190 6 x 1011, 2 × 1012 or 6 x 1012 vg/kg. Data are mean ± SD, n = 4–8 animals per time point. Plasma was collected by tail vein bleeding at the indicated times and the enzyme activity was measured using the fluorogenic substrate 4-methylumbelliferyl-α-D-galactopyranoside. B Plasma clearance curves after intravenous injection of agalsidase alfa at 0.2 mg/kg, in male WT mice. ERT pharmacokinetics using one-phase decay model data are mean ± SD, n = 5 animals per time point. Dotted line indicates plasma α-Gal A at steady state levels following AAV8-FLT190 in wild-type mice. α-Gal A = alpha galactosidase A; AAV8 = adeno-associated virus serotype 8; ERT = enzyme replacement therapy; SD = standard deviation; vg = vector genomes; WT = wild type.

Table 1 Plasma α-Gal A enzyme activity in wild-type mice following treatment with AAV8-FLT190 or enzyme replacement therapy.

An evaluation of the pharmacokinetic profiles of current ERTs (agalsidase alfa and agalsidase beta) was performed in parallel to AAV8-FLT190. Agalsidase alfa (0.2 mg/kg) or agalsidase beta (1 mg/kg) was administrated via tail vein injection to C57BL6 wild-type mice at 2 months of age (n = 15 per group). Mice were sacrificed at 10 min (n = 5), 24 h (n = 5) or 1 week (n = 5) post-injection for determination of plasma α-Gal A levels. Following administration of ERT, α-Gal A was cleared rapidly from the plasma. Assuming exponential decay, a modelled half-life of approximately 40 min was calculated with Cmax assumed to be reached within 10 min following injection. The observed Cmax and AUC of α-Gal A activity for agalsidase beta were greater than those of agalsidase alfa, which may be due to differences in dose (1 mg/kg and 0.2 mg/kg, respectively) (Table 1).

A comparison of the pharmacokinetic profiles of α-Gal A produced by AAV8-FLT190 with those produced by ERTs showed substantially greater α-Gal A enzyme exposure with AAV8-FLT190 than with both agalsidase beta and agalsidase alfa (Table 1). The α-Gal A expressed with the lowest dose of AAV8-FLT190 (6 × 1011 vg/kg) was comparable to the Cmax of agalsidase alfa 0.2 mg/kg. At this level, greater AUC was observed for the AAV8-FLT190 dose group due to the steady state and stable α-Gal A levels in plasma (Fig. 3B).

AAV8-FLT190 in Gla-deficient (Fabry) miceHuman α-Gal A expression

Fabry mice received AAV8-FLT190 via tail vein injection at 4 months of age; 2 × 109 vg/kg (n = 10, 5 male and 5 female), 2 × 1010 vg/kg (n = 10, 3 male and 7 female), 2 × 1011 vg/kg (n = 10, 5 male and 5 female), and 2 × 1012 vg/kg (n = 11, 5 male and 6 female). Consistent with the results of studies in wild-type mice, Fabry mice treated with AAV8-FLT190 exhibited a rapid elevation in plasma α-Gal A activity levels, reaching peak levels by Week 4 post-treatment (Fig. 4A). This level of expression was maintained for the duration of the study (14 weeks) in all treated animals, with mean α-Gal A activity ranging from 4357 to 10,514 nmol/hr/mL. α-Gal A levels observed in male mice were significantly higher than in female mice (p < 0.002). Compared with untreated wild-type controls, the relative fold increases in α-Gal A activity levels in AAV8-FLT190 treated male and female Fabry mice were 1061 and 183, respectively, 14 weeks post-treatment. A strong positive correlation between α-Gal A enzyme activity in the liver and α-Gal A secretion to the plasma following AAV8-FLT190 administration was also demonstrated (Fig. 4B and Fig. 4C).

Fig. 4: AAV8-FLT190 in Gla-deficient (Fabry) mice.

A α-Gal A enzyme activity levels in plasma of Gla-deficient (Fabry) mice that received AAV8-FLT190 2 × 1012 vg/kg. Plasma was collected by tail vein bleeding at the times indicated and enzymatic activity was measured using the fluorogenic substrate 4-methylumbelliferyl-α-D-galactopyranoside. Data are expressed as mean ± SD, n = 5 to 6 animals per time point, males (n = 5) and females (n = 6). Wild-type α-Gal A level is shown as x1 normal, based on the measured activity from 4 animals. Data were analyzed using one-way ANOVA with Bonferroni correction for multiple comparisons. α-Gal A activity was significantly higher in AAV8-FLT190-treated mice when treatment group means were compared with untreated control means (p < 0.0001). B α-Gal A expression and secretion relationship for individual animals showing a positive correlation for liver homogenate to plasma α-Gal A enzyme activity levels. C Levels of vector genome present in murine liver lysate following administration of AAV8-FLT190 in Fabry mice. Amount of vector genome present in liver lysate was determined by QPCR using sets of primers targeting the FRE1 promoter. GAPDH was quantified using primers targeting mouse GAPDH. α-Gal A = alpha galactosidase A; AAV8 = adeno-associated virus serotype 8; FRE1 = Freeline-derived promoter; GAPDH = glyceraldehyde-3-phosphate dehydrogenase; QPCR = quantitative polymerase chain reaction; SD = standard deviation; vg = vector genomes; WT = wild type.

α-Gal A uptake and biodistribution

α-Gal A enzyme activity was determined in tissue homogenates of liver, kidney, heart, spleen, and skin of Fabry mice at 14 weeks following administration of AAV8-FLT190. Tissue α-Gal A level was significantly elevated in the liver of AAV8-FLT190-treated mice, reflective of local production of the enzyme (Fig. 5A). Mean α-Gal A enzyme activity in the liver was 36.98 ± 7.31 nmol/hr/mg protein in WT mice and 1.25 nmol/hr/mg protein in Fabry mice. AAV8-FLT190 2 × 1012 vg/kg produced α-Gal A activity of 1527 nmol/hr/mg protein in female and 32,673 nmol/hr/mg protein in male Fabry mice, which were 41 times and 883 times greater, respectively, than observed in WT mice. These results demonstrated that the AAV8-FLT190 vector genome was efficiently expressed in liver.

Fig. 5: α-Gal A enzyme activity levels in the liver, kidney and heart of Fabry mice administered AAV8-FLT190 2 × 1012 vg/kg.

α-Gal A enzyme activity was measured using the fluorogenic substrate 4-methylumbelliferyl-α-D-galactopyranoside. Data are mean ± SD, n = 4 to 6 animals per time point. One-way ANOVA with Bonferroni correction was used to compare mean of all treatment groups (AAV8-FLT190 treated vs. untreated (vehicle) controls, p < 0.0001). Fold increase was calculated based on WT untreated controls (n = 4). α-Gal A = alpha galactosidase A; AAV8 = adeno-associated virus serotype 8; SD = standard deviation; vg = vector genomes; WT = wild type.

Other tissues exhibited modest increases in α-Gal A activity compared with untreated mice, indicative of metabolic cross-correction and α-Gal A uptake (Fig. 5B, Fig. 5C, Fig. S5). With AAV8-FLT190, α-Gal A enzyme activity levels were 6- to 8-times greater in the kidneys and 14- to 37-times greater in the heart than the levels observed in WT mice. There was a strong positive correlation between plasma α-Gal A levels and tissue uptake, which varied across different tissues and organs.

Exposure and clearance of substrate

Sustained expression of α-Gal A resulted in significant reductions in levels of Gb3 and Lyso-Gb3 in plasma and organs relevant to Fabry disease (Fig. 6).

Fig. 6: Levels of Gb3 and Lyso-Gb3 in plasma, urine and organs of Fabry mice administered AAV8-FLT190 2 × 1012 vg/kg.

Data are mean ± SD, n = 4 animals per time point, analyzed 14 weeks after treatment at 7.5 months of age. n = 4 for untreated age-matched control group, n = 2 for WT mice. The relative reduction of Lyso-Gb3 is shown as % remaining storage. Unpaired t-test was used to compare means of all treatment groups. AAV8-FLT190 treated vs. untreated controls: plasma Gb3 (p = 0.0084); plasma Lyso-Gb3 (p < 0.0001); urine Gb3 (p = 0.02); kidney Gb3 (p = 0.045); kidney Lyso-Gb3 (p = 0.022); heart Gb3 (p < 0.0001); heart Lyso-Gb3 (p ≤ 0.0001); liver Gb3 (p = 0.0075). AAV8 = adeno-associated virus serotype 8; lyso-Gb3 = globotriaosylsphingosine; Gb3 = globotriaosylceramide; SD = standard deviation; vg = vector genomes; WT = wild type.

In liver (Fig. 6), Gb3 levels were similar to those observed in WT mice, whereas untreated control Fabry mice continued to accumulate Gb3 and Lyso-Gb3. In heart, the increase in α-Gal A activity following uptake completely normalized Gb3 and Lyso-Gb3 substrate levels to those observed in WT controls 14 weeks after treatment. In addition, plasma Gb3 and Lyso-Gb3 were practically eliminated 14 weeks after AAV8-FLT190 injection compared with the untreated age-matched controls. Gb3 levels in urine were also effectively normalized after injection of AAV8-FLT190.

Results are also expressed as residual Gb3 or Lyso-Gb3 content in the treated animals relative to age-matched untreated controls. Fourteen weeks after treatment, Gb3 levels were reduced by 91% in the plasma (p = 0.008), 64% in the kidney (p = 0.045), 98% in the heart (p < 0.0001), 97% in the spleen (p < 0.0001), and 99% in the liver (p = 0.0075) compared with untreated controls. Lyso-Gb3 levels were reduced by 98% in the plasma (p < 0.0001), 94% in the kidney (p = 0.022), and 98% in the heart (p < 0.0001) compared with untreated controls (Fig. 6F).

Minimal therapeutic α-Gal A prediction

α-Gal A-mediated Gb3 clearance was dose and time dependent. The range of plasma α-Gal A levels achieved in the AAV8-FLT190-treated mice varied from normal, to above normal (3 to 10-fold), to supra-physiological range (>1000 fold normal). Similarly, there was a wide range in the reductions in Gb3 storage in plasma and tissues. A positive correlation was observed between increase in plasma α-Gal A levels and a reduction in Gb3 storage in plasma, kidney, and heart (Fig. 7A‒C). The observed relationship that Gb3 reduction correlated with plasma α-Gal A activity enabled a prediction of the α-Gal A activity required to deliver a given level of efficacy, as defined by reduction in Gb3. α-Gal A vs. Gb3 analysis revealed a positive correlation between an increase in plasma α-Gal A levels and a reduction in Gb3 storage (R2 > 0.72). This approach allowed modelling of plasma α-Gal A required for specific levels of plasma/kidney/heart Gb3 clearance. Predicted α-Gal A for minimal efficacy (defined as 50% reduction in Gb3 storage in plasma/tissues) was within the normal physiological range, indicating the importance of continuous exposure to α-Gal A.

Fig. 7: Correlating Plasma α-Gal A to Gb3.

Correlating plasma α-Gal A to Gb3 using a nonlinear fit regression model Marquardt method in SAS System. Percentage (%) of remaining storage relative to untreated controls, Sigmoid (4PL) Fit: X = Log α-Gal A, Y = Logged Gb3 analysis is shown. The R2 values are the following: plasma = 0.89 (A), for kidney = 0.72 (B), and for heart = 0.82 (C). The main purpose of the curve fitting was to predict plasma α-Gal A values for various levels of Gb3 reduction (e.g. 50%), exploring log(plasma Gb3), log(kidney Gb3) and log(heart Gb3) values as the y responses. α-Gal A prediction at 10%, 20%, 50% and 70% remaining storage are shown in tables. Statistical packages SAS and GraphPad Prism were used for the analysis. α-Gal A = alpha galactosidase A; Gb3 = globotriaosylceramide.

Pathological correction and durability of expression

Storage of substrate in kidney and heart and endothelial abnormalities similar to those observed in Fabry patients previously have been described in the Gla-deficient mouse model [22]. We noted significant reductions in storage inclusion bodies in renal cell types and heart in AAV8-FLT190-treated animals (Fig. 8) with stable serum α-Gal A activity. Sustained, durable levels of plasma α-Gal A were demonstrated in Fabry mice up to 14 months of age after a single dose of an early development self-complementary (sc) codon optimized wild-type GLA proof-of-concept construct scAAV2/8-LP1-GLAco-SV40p 2 × 1012 to 2 x 1013 vg/kg (Fig. S6).

Fig. 8: Representative electron microscopy images of kidney and heart tissues from Fabry mice untreated (left) and following administration of AAV8-FLT190 (right).

Tissues from kidney, renal cortex and apex of the heart were processed for electron microscopy. The arrows indicate pathological changes resulting from Fabry disease-like features (storage inclusion bodies). Clearance of substrate was observed in AAV8-FLT190-treated mice compared with vehicle control Fabry mice. Representative pictures shown. The scale bar represents 10 µm and 2 µm for kidney, and 5 µm for heart. AAV8 = adeno-associated virus serotype 8; WT = wild type.

NHP studies

The pharmacokinetics of transgene expression following administration of FLT190 (pseudotyped with the AAVS3 vector) were investigated in rhesus macaques in the following two studies: a 13-week GLP toxicity study of a single dose of 3 x 1013 vg/kg that included 3 males and 3 females, and a 26-week investigational study of a single dose of 6 x 1012 vg/kg vector dose (3 males) and vehicle control (3 males and 3 females). Data on plasma α-Gal A activity and hGLA mRNA copy number from the two studies are combined and shown in Fig. 9A, B, respectively.

Fig. 9: FLT190 evaluation in NHPs.

A α-Gal A enzyme activity levels in the plasma of NHPs administered FLT190 3 × 1013 vg/kg or 6 × 1012 vg/kg dose. B hGLA mRNA levels in liver biopsies. Plasma was collected at the times indicated and the enzyme activity was measured using the fluorogenic substrate 4-methylumbelliferyl-α-D-galactopyranoside (nmol/hr/mL). Data are geometric mean ± SD, n = 6 animals per time point (vehicle control and 3 × 1013 vg/kg) or n = 3 animals per time point (6 × 1012 vg/kg). Two-way ANOVA with repeated measures was used to compare treatment groups. FLT190-treated vs. untreated (vehicle) control, p = 0.0003. α-Gal A = alpha galactosidase A; hGLA = human α-galactosidase A gene; mRNA = messenger ribonucleic acid; NHP = nonhuman primate; RNA = ribonucleic acid; SD = standard deviation; vg = vector genomes.

FLT190 α-Gal A expression

In the 13-week 3 x 1013 vg/kg NHP study, plasma α-Gal A activity increased rapidly from baseline within 1 week following FLT190 infusion and remained stable from Weeks 5 to 13 (Fig. 9A). Of the 36 plasma samples from FLT190-treated animals analyzed, mean α-Gal A specific activity at terminal stage was 115 nmol/hr/mL, with values ranging from 23 nmol/hr/mL to 294 nmol/hr/mL; baseline activity was between 14 and 21 nmol/hr/mL. On average, the relative increase in plasma α-Gal A activity was 5.5 times higher than normal in females and 7.6 times higher than normal in males 13 weeks post-treatment. A two-way ANOVA revealed significantly higher α-Gal A expression in the FLT190 group compared with the control group (p = 0.0003). In treated animals there was ongoing uptake and clearance of α-Gal A by the tissues, and an equilibrium state was maintained by continuous secretion of α-Gal A from transgene expression in the liver. Tissue α-Gal A activity was elevated in the liver of FLT190-treated NHPs, reflective of AAV-mediated expression of the enzyme. NHPs injected with FLT190 3 x 1013 vg/kg showed a mean α-Gal A level of 167 nmol/hr/mg protein as compared with 77 nmol/hr/mg protein in control NHPs, reflecting a 2.2-fold change between groups (data not shown). Consistent with plasma α-Gal A activity, high levels of hGLA mRNA were detected in the liver using RT-qPCR in the FLT190-treated animals compared with background levels detected in reference samples (Fig. 9B).

Two high-responder NHPs (#2003 and #2503) had α-Gal A levels in the liver that were 3.7- and 3.2-fold higher than normal levels (p < 0.0008), respectively. Other tissues in these two high-responder NHPs exhibited modest physiological elevations of α-Gal A levels compared with control NHPs, indicative of α-Gal A uptake into the target organs. For example, the α-Gal A levels in the kidney cortex reached 1.8 and 1.9 times above normal (p < 0.0001) in the two high-responder NHPs, whereas no significant uptake was observed in spleen and heart. In the FLT190 treated group, mean α-Gal A levels of 387 nmol/hr/mg protein were detected in the kidney cortex, 1.5 above that of the control group at 13 weeks post-dose. As for the high-responding animals, no significant uptake was observed in the spleen or heart.

AAV-mediated dorsal root ganglion degeneration and dorsal axonopathy of the spinal cord have been observed at the cervical, thoracic, and lumbar segments following the administration of AAV vectors in NHP [25]. The neuronal degeneration was predominantly noted when vectors were administered directly to the cerebrospinal fluid and when the dose was 1 x 1013 vg/kg or greater. Hence, we performed a histological examination of cervical, thoracic and lumbar dorsal root ganglion, and spinal cord tissues from this study. No neuronal degeneration was noted in any FLT190-treated NHPs.

In the 26-week NHP study, the pharmacokinetic profile of α-Gal A in plasma was evaluated in animals dosed with FLT190 6 x 1012 vg/kg (3 males) at 7 time points: once in acclimation (baseline), and at Weeks 1, 3, 5, 13, 22 and 26 post FLT190 infusion. The plasma α-Gal A activity data showed a rapid increase in α-Gal A levels from baseline to 2 weeks post AAV infusion, and that plasma levels remained stable from Week 5 to 26 (Fig. 9A). Of the 21 plasma samples analyzed from FLT190-treated animals, the average specific activity of α-Gal A at the end of the study (Day 183) was 19.8 nmol/hr/mL, with values ranging from 11 nmol/hr/mL to 24 nmol/hr/mL, whereas the baseline activity measured at pre-dose time point was between 3 and 17 nmol/hr/mL. After FLT190 treatment, physiological levels of α-Gal A reaching up to 57.2 nmol/hr/mL were observed in animal #1003. However, on average, the relative increase of plasma α-Gal A activity was 2 times normal at most time points except at Day 15 where 4 times normal peaked levels were observed.

Anti-α-Gal A antibodies in plasma

The levels of anti-α-Gal A antibodies in sera of NHPs dosed with FLT190 were determined by an ELISA using a rabbit anti-human α-Gal A monoclonal antibody as a reference standard positive control. Of the six 3 x 1013 vg/kg FLT190-treated NHPs, one male (#2003) exhibited a positive signal for anti-α-Gal A antibodies at Day 57 and Day 92. The other five NHPs remained negative for anti-α-Gal A antibodies; no specific antibodies against α-Gal A were detected at Day 29, 57, and 92 post-dose as compared with pre-dose samples. The anti-α-Gal A positive NHP #2003 had a titer of 40 and 320 at Days 57 and 92, respectively, and there was no evidence of neutralizing effects or drop in α-Gal A expression at the corresponding time points.