Analytical Characterization, Safety and Clinical Bioequivalence of Recombinant Human Insulin from Transgenic Plants Boothe J.G., Nykiforuk C.L., Kuhlman P.A., Whelan H., Pollock W.B.R., Clark S., Yuan S., Kumar R., Murray E.W., Visser F., Martens K., Wu J., Pollock E., Given B., Szarka S., Zaplachinski S., Harry I., Keon R. and Moloney M.M. SemBioSys Genetics, 110, 2985 23rd Avenue NE, Calgary, Alberta, T1Y 7L3 kDa 250 150 100 75 50 37 25 20 15 Coomassie Blue Anti-Inslin Immunoblot ABSTRACT This work demonstrates the quality, safety and pharmacological equivalence of recombinant human insulin produced from transgenic safflower plants. A cGMP process was developed for the purification of insulin expressed in transgenic seed. Physical-chemical, structural, and biological analysis confirmed the identity and functionality of the resulting product. The plant-made insulin (SBS-1000) was also shown to meet all compendial specifications and to be free of host-specific impurities. Repeat dose (28-day) toxicity studies comparing the effects of SBS-1000 with Humulin ® R were performed in Fischer 344 rats and Cynomolgus monkeys. Parameters evaluated include clinical signs, injection site changes, tissue/organ pathology and toxicokinetics. Data from these studies indicate that SBS-1000 had a safety profile indistinguishable from Humulin ® R. A 3-way crossover study in healthy volunteers (n=23) using a euglycaemic clamp was conducted to compare the pharmacokinetics and pharmacodynamics of SBS-1000 with two reference insulins, Humulin ® R and Humulin ® S. Results from this study showed that the three insulins displayed similar adverse event profiles. Although an analytical error resulted in a reduced concentration in the formulated SBS-1000 (accounting for lower values for the point estimate ratios), the 90% CI of all primary endpoints fell within the 80-125% range required for bioequivalence to Humulin ® R. Similar results were obtained with Humulin ® S for all endpoints except GIRmax. This difference between SBS-1000 and Humulin ® S was also observed between Humulin ® R and Humulin ® S and was exacerbated by the greater variability observed for this parameter. Together these results demonstrate the feasibility of our transgenic plant manufacturing platform for the production of pharmaceutical-quality human insulin. II. NON-CLINICAL TOXICOLOGY III. CLINICAL PHARMACOLOGY PHARMACEUTICAL QUALITY INSULIN FROM PLANTS The growing incidence of diabetes worldwide has resulted in an escalating demand for human insulin. Plants offer a high-capacity, low-cost solution to meet this demand. SemBioSys has engineered safflower (upper left) to produce human insulin. The insulin is expressed as a fusion protein and accumulates specifically in the seed (upper right). The transgenic plants are grown and harvested according to a rigorous set of procedures that maintain strict containment of the crop and ensure the quality of the seed. SemBioSys has developed a cGMP process for the extraction and purification of insulin from seed. The resulting product meets all of the standards for pharmaceutical-quality insulin (lower table). These include all of the compendial specifications for identity, functional potency and purity with respect to product-related substances. The analysis additionally demonstrates the removal of all major types of host and process specific impurities. SAFETY PROFILE IS COMPARABLE TO REFERENCE LISTED DRUG The safety profile of the plant-produced insulin was established in side-by-side toxicology studies (see design in upper table) comparing the plant product with a reference listed insulin, Humulin®R. The studies were performed in two species, rats and monkeys, with animals receiving daily injections of insulin for 28 days at or near the maximum tolerated dose level for each species determined in previous range finding experiments. Results from the studies in monkey (lower table) revealed comparable safety profiles for both products with no unique toxicities associated with the plant-produced insulin. Toxicokinetic analysis indicated no significant differences in Cmax or AUC for either insulin or C-peptide. Reduction in blood glucose levels was also found to be similar for both insulins. Apart from some signs of hypoglycemia (* in table) attributed to the high dosing (resulting in the loss of one animal treated with plant-produced insulin), there were no other adverse clinical findings. Body weight, ophthalmoscopy, electrocardiography, hematology, clinical chemistry and urinalysis were all normal and no gross organ weight or histopathological changes were observed. Similar results were obtained in the rat studies. DEMONSTRATION OF CLINICAL BIOEQUIVALENCE AND SAFETY A Phase 1 human clinical trial was performed comparing the pharmacological behavior (pharmacokinetics and pharmacodynamic response) of plant-produced human insulin with that of two reference listed insulins, Humulin ® R and Humulin ® S (see design in upper table). The trial was conducted in healthy (non-diabetic) subjects and employed a euglycaemic clamp to measure glucose infusion rates (GIR) in response to a single subcutaneous injection of insulin. An ELISA assay was used to determine insulin levels in the blood. All three insulins were tested in each subject. All products showed similar profiles with respect to insulin levels (middle left) and GIR (middle right) and were well-tolerated with no atypical adverse events. Although an analytical error resulted in a reduced concentration in the formulated plant insulin (giving rise to lower point estimates for the endpoints), the trial demonstrated that the plant product was bioequivalent to Humulin ® R for all primary endpoints (Insulin AUC0-8hrs. Insulin Cmax, GIR AUC0-8hrs., GIRmax). Plant-produced insulin was also bioequivalent to Humulin ® S for all endpoints except GIRmax. However a similar discrepancy was observed for Humulin ® R vs. Humulin ® S and these results are likely due in part to the greater variation observed in the trial for this parameter. CONCLUSION Our results demonstrate that human insulin produced in transgenic plants can be manufactured to achieve the same standards of quality obtained using conventional microbial production systems. The resulting product has a safety profile that is indistinguishable from that of pharmaceutical-grade insulin. In a human clinical trial the plant-produced insulin was shown to be bioequivalent to listed insulin drug products. Together these findings support the potential for using plant-based production to meet the rising global demand for human insulin. I. PRODUCTION AND ANALYSIS OF INSULIN FROM PLANTS API Batch Analysis Tabulated Results Summary (Monkey) Comparison of Blood Insulin Levels Between SBS-1000 and Reference Insulins Comparison of Glucose Infusion Rates Between SBS-1000 and Reference Insulins 0 50 100 150 200 250 0 10.00 8.00 6.00 4.00 2.00 0.0 0 500 400 300 200 100 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 200.0 Hum S SBS-1000 Hum R Hum S SBS-1000 Hum R Category Attribute Result General Quality Characteristics Appearance White, dry, powder Residual Moisture 7.6% Identity USP Identification A RT corresponds to USP Insulin USP Identification B Fragments correspond to USP Insulin Zinc Content 0.1% Potency USP Insulin Assay 30.27 units/mg Insulin Receptor Binding Ratio IC 50 USP Insulin: 1.20 IGF-1 Receptor Binding Ratio IC 50 USP Insulin: 1.06 Glucose Uptake Ratio USP Insulin: 1.04 USP 121 Bioassay (quantitative) 26.2 units/mg Product Purity USP Human Insulin Assay 99.53% High Molecular Weight Compounds 0.06% Insulin Related Compounds 0.8% A21 desamido 0.47% other compounds Product Related Impurities Residual scFvD9 < 2 ppm Residual C-peptide < 1 ppm Process Related Impurities Host Cell Protein < 10 ppm Host Cell DNA < 1 ppm Pesticides < LOD (LOD = 50 ppb) Residual solvents Conforms with ICH Q3C Residual Process Enzymes < LOD (LOD = 0.03%) Microbial Limits < LOD (LOD = 33 cfu / g) Bacterial Endotoxins 8 EU/mg Study Design 28-day repeat dosing (daily) comparing plant-derived insulin (SBS-1000) to Reference Listed Drug (Humulin ® R). Species Dosing Fischer 344 Rats (12 animals/group), Cynomolgus monkeys (4 animals/group) Rats: 2.2-2.4 U/kg, SQ Monkeys: 2.0-1.8 U/kg, SQ Parameters Toxicokinetics, body weight, food consumption, ophthalmoscopy, glucose levels, clinical chemistry, hematology, urinalysis, local site reaction, histopathology Study Design Double-blind, randomized, active-controlled, 3-way crossover, euglycaemic glucose clamp study in healthy adults comparing SBS-1000 with two reference listed insulins, Humulin ® R and Humulin ® S. Patient Population Healthy (non-diabetic) adult subjects, n=23 Dosing 0.2 U/kg, SQ Primary Endpoints Pharmacokinetics: Insulin AUC0-8hrs and Cmax. Pharmacodynamic: GIR AUC0-8hrs GIRmax Secondary Endpoints Pharmacokinetic: Insulin AUC0-predose baseline, tmax and t1/2. Pharmacodynamic: GIR AUC0-3hrs, AUC0-6hrs and AUC0-predose baseline. Safety and tolerability. SBS-1000:Humulin ® R SBS-1000:Humulin ® S Humulin ® R: Humulin ® S Parameter Ratio (%) 90% CI Ratio (%) 90% CI Ratio (%) 90% CI Insulin AUC0-8 hrs. 91 87-95 90 85-95 99 95-103 GIR AUC0-8 hrs. 92 84-100 92 85-100 101 92-110 Insulin Cmax 94 87-102 86 80-93 92 84-100 GIRmax 93 81-108 81 68-97 87 74-103 Parameter Control Humulin ® R SBS-1000 M F M F M F Insulin: AUC 0-6 (mIU•hr/mL) Day 1 70.7 22.6 2070 2150 2270 1890 Day 28 12.9 18.2 1620 1830 2010 1640 C-Peptide: AUC 0-6 (pmol/L) Day 1 1180 1530 418 404 153 183 Day 28 1080 1350 223 356 92.5 141 Noteworthy Findings Died or Sacrificed Moribund 0 0 0 0 0 1 Body Weight (% diff.) na na –2.3 –0.3 –1.9 +5.1 Clinical Observations – – – – +(*) +(*) Dermal Irritation – +(mild) – – – – Ophthalmoscopy – – – – – – Electrocardiography – – – – – – Glucose (mg/dL) Day 1 Prior to dosing 80.5 83.5 90.0 70.5 76.0 76.0 30 minutes 68.3 73.8 30.8 27.5 34.0 33.5 120 minutes 63.5 60.8 31.8 10.8 19.0 19.5 360 minutes 66.0 73.8 63.3 72.3 75.3 75.3 Day 28 Prior to dosing 60.0 65.5 67.0 62.5 70.8 70.8 30 minutes 53.5 66.8 18.5 20.8 14.0 19.8 120 minutes 75.3 72.8 37.5 28.3 33.5 42.8 360 minutes 82.8 81.5 88.0 78.0 76.8 90.5 Hematology – – – – – – Clinical Chemistry – – – – – – Urinalysis – – – – – – Gross Pathology – – – – – – Organ Weights – – – – – – Histopathology – – – – – – Injection Site Changes 0 3 1 2 3 2 pmole/L*min hr min mL/h