Comparative physicochemical and structural characterisation studies establish high Biosimilarity between BGL-ASP and Reference Insulin Aspart

doi:10.21203/rs.3.rs-3461609/v1

Biosimilar insulin analogues are increasing market access for diabetic patients globally. Scientific establishment of biosimilarity is cornerstone of this key change in the medical landscape. BGL-ASP is a biosimilar insulin aspart developed by BioGenomics Limited, India. BioGenomics has considered a stepwise approach in generating the totality of evidence required to establish similarity with reference product. Insulin aspart is a recombinant rapid-acting human insulin analogue utilised in the treatment of type-1 and type-2 diabetes mellitus. The single amino acid substitution at position B28 where proline is replaced with aspartic acid results in a decreased propensity to form hexamers, thus increasing the absorption rate on subcutaneous administration compared to native insulin. In order to establish the safety and efficacy of BGL-ASP, the critical quality attributes (CQAs) of BGL-ASP are identified based on the impact created on biological activity, pharmacokinetic/ pharmacodynamic (PK/PD), immunogenicity and safety. The CQAs of insulin aspart are related to product structure, purity and functionality and are characterised using a series of state-of-the-art orthogonal analytical tools. The primary protein sequence, the secondary, tertiary and quaternary structure are found to be highly similar for BGL-ASP and reference product. The product related impurities of insulin aspart and the assay content are determined using high performance liquid chromatography (HPLC) based analysis and is similar for BGL-ASP and reference insulin aspart sourced from United States of America (US), Europe Union (EU) and India. The safety, efficacy and immunogenicity of BGL-ASP is also found to be comparable with reference product and is confirmed through the clinical trials conducted as recommended by International Council for Harmonisation of Technical Requirements of Pharmaceuticals for Human Use (ICH) and European Medicines Agency (EMA) guidelines. The data encompassed in this study demonstrates that reference insulin aspart and BGL-ASP are highly similar in terms of structural, physicochemical, and biological properties, thus confirming its safety and efficacy for usage as potential alternative economical medicinal treatment for diabetes mellitus.

Biological sciences/Biotechnology

Biological sciences/Biological techniques

Biological sciences/Biochemistry/Hormones/Peptide hormones

BGL-ASP has been developed as a biosimilar to insulin aspart reference product

BGL-ASP is found to be highly similar to insulin aspart reference product in terms of structural

physicochemical and biological properties.

The discovery of insulin has completed 100 years, however, the accessibility and affordability of the molecule to patients is still a major concern, especially individuals in the low-income group worldwide. Based on the reported data it is observed that in individuals with type 2 diabetes, the access is restricted to 50% patients globally and 15% patients in the sub-Saharan Africa [1]. As the cases of type 2 diabetes are increasing annually, international guidelines prefer insulin analogues to human insulin because of lesser weight gain, better HbA1C control and effectivity [2]. However, the high cost of these products become a deterrent in usability, resulting in either terminating the treatment or resorting to lower usage than the prescribed dose.

The manufacturing of insulin and analogues are limited to two-three large companies creating monopoly with respect to the product pricing. In order to reduce the product cost and enhance the accessibility and affordability of insulins, it is imperative that newer manufacturers introduce safe and efficacious product manufactured in a GMP-compliant facility to increase the competition. Thus, to make affordable treatment on controlling blood glucose levels accessible to diabetes patients, BioGenomics has developed BGL-ASP, a biosimilar developed in reference to the marketed insulin aspart 100 units/ml (NovoRapid^® by Novo Nordisk A/S, Denmark).

Insulin aspart is a recombinant rapid-acting human insulin analogue utilized in the treatment of type-1 and type-2 diabetes mellitus. It is an analogue of human insulin made by replacing the proline residue at position B28 of the B-chain with aspartic acid [3, 4]. The switch in amino acids results in a decreased tendency of the insulin aspart to form hexamers, thus enhancing the rate of absorption from the subcutaneous injection sites. The properties of insulin aspart are suitable for “basal-bolus” regimen representing a more physiological plasma insulin profile, in which basal insulin concentrations are provided by long-acting insulin preparations and bolus or high levels of insulin are provided by rapid-acting insulin analogues [5, 6].

The production of insulin aspart by innovator is undertaken in Saccharomyces cerevisiae while BGL-ASP is manufactured in a non-pathogenic strain of Escherichia. Coli [4]. As the manufacturing processes of insulin aspart are different for BGL-ASP and reference product there are possibilities of differences arising in quality attributes and these comparisons are initiated in the early phase of development prior to clinical trials. The objective of the comparative data analysis is to demonstrate that the critical quality attributes of BGL-ASP and reference product are similar even though the cultivation systems and manufacturing processes are different [7].

BioGenomics has considered a stepwise approach in generating the totality of evidence required to establish similarity with reference product as shown in Fig. 1 [8]. In the first step the focus was on establishing comprehensive structural, physicochemical and biological similarity between reference product and BGL-ASP using state-of-the-art orthogonal analytical tools. These studies were highly sensitive, and the aim was to detect minor difference in quality attributes of clinical relevance between reference product and BGL-ASP. The characterization studies demonstrated a high similarity in the critical quality attributes of BGL-ASP and reference product.

The second step involved comparative, pre-clinical studies comprising of: four single dose, two repeat dose and one skin sensitization studies. The repeat dose studies were conducted in comparison with reference product and no abnormal clinical signs or adverse effects were observed post studies. In the third step bioequivalence of pharmacokinetic (PK) and pharmacodynamic (PD) parameters were demonstrated for BGL-ASP and reference product through phase I clinical studies. The EMA-compliant phase III clinical studies demonstrated that the immunogenicity profiles and drop in HbA1c values were identical for Type 2 (T2DM) diabetes mellitus patients treated with BGL-ASP and reference product, thus confirming the product safety, efficacy and immunogenicity [11].

In this report, the data related to structural and physicochemical characterization of insulin aspart is presented. Functional characterization studies such as cell-based glucose uptake, Insulin Receptor-B (IR-B) phosphorylation, lipogenesis, in vivo rabbit bioidentity tests and mitogenicity as an off-target assay for BGL-ASP have been completed in comparison with the reference product and the results were found to be highly similar, however, these studies are beyond the scope of this paper and are published separately [12].

1.1 Risk Ranking of Quality Attributes

Based on the analysis of multiple batches of reference insulin aspart, the quality target product profile (QTPP) was determined. The QTPP assisted in understanding the attributes related to product quality, safety and efficacy and formed the basis for identifying critical quality attributes (CQAs). The CQAs assisted in developing a control strategy for the manufacturing process essential in controlling the product quality throughout its lifecycle. The risk ranking of quality attributes for insulin aspart was assessed based on the impact created on safety and efficacy. The criticality score for a quality attribute was calculated as a function of impact and uncertainty [13, 14]. The impact score (such as very high, high, moderate or low) was calculated based on the impact created on biological activity, pharmacokinetics (PK), pharmacodynamics (PD), immunogenicity and safety [15]. The uncertainty score was based on the either the literature available for an impurity or the presence of the variant in material used for clinical studies. The Table 1 below demonstrates the varied quality attributes and the classification of criticality. The categorization of the attributes was performed as per the principles and concept mentioned in International Council for Harmonisation (ICH) Q8 and ICH Q9.

Table 1

Risk Ranking of Quality Attributes and testing methods
Test Classification	Quality Attributes	Methods	Criticality
Primary Structure	Amino Acid Sequence	Reduced peptide mass fingerprinting using Trypsin (LC-MS/MS)	Very High
	Molecular Weight	Intact protein molecular mass analysis (LC-MS)
	Disulphide Linkage	Reduced peptide mass fingerprinting using V8 enzyme (LC-MS/MS)
	Disulphide Linkage	Non-reduced peptide mass fingerprinting (LC-MS/MS)
Higher Order Structure	Secondary Structure	Far UV-CD Spectroscopy	High
	Secondary Structure	FTIR Spectroscopy
	Tertiary Structure	NMR Spectroscopy
		Near UV-CD Spectroscopy
		Fluorescence Spectroscopy
	Quaternary Structure	Differential Scanning Calorimetry (DSC)
Strength	Concentration	RP-HPLC-UV	High
Product Related Impurities	Higher Molecular Weight Protein (HMWP)	SE-HPLC-UV	High
Product Related Impurities	Impurities such as deamidated, oxidized and isomerized forms	RP-HPLC-UV	High
Formulation Components	Phenol Content	RP-HPLC based quantitation	Moderate
	M-cresol Content	RP-HPLC based quantitation	Moderate
	Zinc	Atomic Absorption Spectroscopy	Moderate
	pH	Potentiometry	Moderate
Additional Characterization Tests not related to Biosimilarity	Host Cell Protein (HCP)	ELISA	Considered critical to product safety and immunogenicity but not related to biosimilarity
	Host Cell DNA (HCDNA)
	Single Chain Precursor (SCP)
	Residual Enzyme Content
	Residual Solvent Content	GC-MS based analysis

LC-MS/MS: Liquid Chromatography Tandem Mass Spectrometry, UV-CD: Ultraviolet-Circular Dichroism, FTIR: Fourier Transform Infrared Spectroscopy, NMR: Nuclear Magnetic Resonance, RP-HPLC-UV: Reverse Phase-High Performance Liquid Chromatography-Ultraviolet, SE-HPLC-UV: Size Exclusion-High Performance Liquid Chromatography-UV, ELISA: Enzyme-Linked Immuno Sorbent Assay, GC-MS: Gas Chromatography-Mass Spectrometry

The section 351(i) of the Public Health Service (PHS) Act of United States Food and Drug Administration (USFDA) defines biosimilarity to mean “that the biological product is highly similar to the reference product notwithstanding minor differences in clinically inactive components” and that “there are no clinically meaningful differences between the biological product and the reference product in terms of the safety, purity, and potency of the product”[16]. The similarity assessment studies were performed in congruence with the regulatory guidelines using the developed scientific understanding of the molecule.

As mentioned in Table 1, the amino acid sequence and the native disulphide linkages are the basic elements governing the function and safety of insulin aspart, hence its criticality was considered as very high. The criticality was assigned as high to structural attributes related to pharmacological activity, protein concentration determined by RP-HPLC in relation to dosage accuracy and product related impurities. The deamidated, oxidized and isomerized forms are considered as triggers for Higher Molecular Weight Protein (HMWP) formation [17]. The HMWP are inactive forms of insulin and hence the criticality was considered as high. The preservatives, pH, and zinc content are important elements of insulin aspart formulation related to product efficacy, found to be effective over a wide range and hence their criticality was considered as moderate. The HCP, HCDNA, SCP, residual enzyme content and residual solvent content are important quality attributes related to product safety, however, these impurities are present in low quantities in the product and hence the criticality was considered as low. Additionally, these impurities are process-specific and hence not considered as part of biosimilarity [18].

1.2 Product Related Impurities of Insulin Aspart

The asparagine (Asn) at position A21 (A21Asp) and B3 (B3Asp), and aspartic acid (Asp) at B28 (B28Asp) position are susceptible to degradation resulting in the formation of product related impurities as shown in Fig. 2. The Asn at A21 is susceptible to deamidation, the Asn at B3 and the Asp at B28 position are susceptible to isomerization. The asparagine (Asn) deamidation and aspartate (Asp) isomerization reactions are non-enzymatic, intra-molecular rearrangement reactions occurring in peptides and proteins, and these non-native amino acid residues are a source of major stability concern in the formulation of these biomolecules [19].

Additionally, the other product related impurity observed with insulin aspart is the presence of higher molecular weight protein (HMWP) [20]. The HMWP formation typically occurs due to intermolecular disulphide bonds.

2.1 Manufacturing of BGL-ASP

The manufacturing of insulin aspart incorporates the utilization of the bacterial cultivation system (BCS). The BCS provides distinct advantages such as absence of post-translational modifications and fast growth kinetics. The disadvantage is the formation of the inactive inclusion bodies with non-native conformation. The developmental work involved rigorous optimisation of refolding/ renaturation conditions that generates native disulphide linkage and conformation [21–28].

Insulin aspart is expressed in its precursor form as inclusion bodies in a non-pathogenic laboratory strain of E.coli as the production organism. The laboratory strain is classified as non-hazardous by the 2012 Occupational Safety and Health Administration (OSHA) Hazard Communication Standard (29 CFR 1910.1200) and regulation (EC) 1272/2008 [CLP]. The seed culture is prepared from a characterized and qualified working cell bank. The seed flasks are transferred to the inoculums flasks on achieving the desired growth criteria. The inoculum is further added to the seed fermenter, which acts as feed source for the production fermenter. The production fermenter is operated as per the predefined conditions and the insulin aspart precursor (IAP) is overexpressed in the presence of suitable inducers at optimal concentrations and temperature. The overexpression of insulin aspart precursor results in the formation of inclusion bodies. The cell harvest is isolated using continuous centrifugation and subjected to cell-lysis using high pressure homogenization systems. The inclusion bodies are dense particles and are isolated from the cell debris through a series of centrifugation cycles. The isolated inclusion bodies of IAP are subjected to the process of renaturation resulting in the native conformation with expected disulphide linkages. The IAP is converted to insulin aspart in the presence of recombinant enzymes such as Trypsin and Carboxypeptidase B [27, 28]. The purification process involves low pressure and high pressure-based chromatography for separation of process and product related impurities and generation of a purified form of insulin aspart qualifying as per the predefined acceptance criteria [29]. The insulin aspart drug substance is manufactured in the freeze-dried form and is formulated in the presence of pharmacopeial grade excipients such as phenol, m-cresol, glycerol, zinc chloride, disodium phosphate dihydrate, sodium chloride, hydrochloric acid (for pH adjustment), sodium hydroxide (for pH adjustment) and water for injections, at a concentration of 3.5 mg/ml equivalent to 100 units/ml. The formulated insulin aspart drug substance is referred to as BGL-ASP in this study.

2.2 Physicochemical and Structural Characterization Studies

The physicochemical characterisation of insulin aspart involves tests related to purity, assay content, HMWP content, preservatives and zinc content. Comprehensive structural characterisation of insulin aspart was performed using techniques such as NMR, FTIR, CD, DSC and intrinsic fluorescence spectroscopy. The peptide mass fingerprinting under reduced and non-reduced conditions and mass analysis was performed using LC-MS technique. The process related impurities such as HCP, HCDNA, residual enzyme, SCP were also analyzed and the data for the same was recorded. The N-terminal and C-terminal sequencing using LC-MS/MS technique, sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS PAGE) [30] and western blot analysis [31] was also performed (data not shown).

2.2.1 Reference Product Details

The reference product, NovoRapid^® by Novo Nordisk, in a 3 ml cartridge containing 100 units insulin aspart (equivalent to 3.5 mg), glycerol, phenol (1.5 mg/mL), meta-cresol (1.72 mg/mL), zinc chloride, disodium phosphate dihydrate, sodium chloride, hydrochloric acid and sodium hydroxide (to adjust pH) and water for injection was utilized for the physicochemical characterization and structural studies [32]. The reference product batches were sourced from India, Europe and United States of America (USA). The European Pharmacopeia Commission of Reference Substance (EPCRS) for insulin aspart manufactured by European Directorate for the Quality of Medicines and HealthCare (EDQM) was procured from Sigma-Aldrich (Cat No. Y0000349).

2.2.2 RP-HPLC based analysis of Insulin Aspart

The related impurities of insulin aspart were monitored using RP-HPLC based analysis. The column used for HPLC analysis was an octadecylsilyl silica gel column procured from Kromasil (250 mm x 4 mm; 5 µm) (Cat No: E118882). The mobile phase A was prepared by dissolving 142.0 g of anhydrous sulfate R (Cat. No. 59977, Sisco Research Laboratories) in water and further addition of 13.5 ml of phosphoric acid (Cat No. 1.93403.0521/ 79606, Sigma-Aldrich). The volume was adjusted to 5000 ml with water and pH was adjusted with sodium hydroxide (Cat. No. 1.93502.0521, Merck). Nine volumes of this solution was merged with one volume of acetonitrile (Cat No. 61803025001730, Merck). The mobile phase B was prepared by mixing equal volume of acetonitrile and water. The separation was performed at 35^oC, at a flow rate of 1 ml/min, with a total run time of 60 min. The separation methodology involved an isocratic step of 42% B, followed by a gradient of 42–80% B over 5 min. The monitoring wavelength was set at 214 nm. The RP-HPLC method can distinctly separate related impurities of insulin aspart such as B28isoAsp, B3Asp, A21Asp and B3isoAsp. The method was also utilized to calculate assay content against insulin aspart chemical reference standard from European Pharmacopeia (EP) (Order Code: Y0000349) [33].

2.2.3 Peptide Mass Fingerprinting - LC-MS based analysis of intact, reduced and V8 enzyme digested peptides of Insulin Aspart

In order to demonstrate that the disulphide linkage in BGL-ASP batches were highly similar to reference insulin aspart, a comparative peptide mapping analysis was performed wherein insulin aspart was digested with V8/ Endoproteinase GluC (Cat no: P2922, Sigma-Aldrich) enzyme and the mixture generated was analyzed under reducing and non-reducing conditions using Liquid Chromatography-Electron Spray Ionization-Mass Spectrometry/ Mass Spectrometry (LC-ESI-MS/MS). The reduction reaction was set up in the presence of Dithiothreitol (DTT) (Cat No: 10197777001, Sigma-Aldrich). The comparison between the theoretical and generated molecular weights for the peptides, assisted in demonstrating the presence of the native disulphide linkages. The untreated insulin aspart mass was also determined and compared with the innovator product. The column with the dimensions C18, 150 x 2.1 mm, 3 µm, (Cat No: TA12S05-15Q1PTH, YMC) was used for LC-MS based analysis. The detection wavelength was set at 215 nm, flow rate of 0.2 ml/min. The AB SCIEX Triple TOF 4600 was set to mass range of m/z 200–2000 Da, source temperature of 450^oC and collision energy of 35 V.

2.2.4 SE-HPLC based analysis of Higher Molecular Weight Proteins (HMWP) in Insulin Aspart

The HMWP content in insulin aspart was determined using SE-HPLC based analysis. The column used for SE-HPLC analysis was a hydrophilic silica gel column procured from Waters (7.8 x 300 mm; 5 µm) (Cat No: WAT201549). The mobile phase was prepared by mixing glacial acetic acid (Cat No: 85801(0129168), SRL), acetonitrile (Cat No: 61803025001730 / A2094, Merck), and 1.0 g/L solution of arginine R (Cat No: 014879, SRL) in a ratio of 15:20:65. Isocratic elution was performed at a flow rate of 0.5 ml/min at 25ºC, with a total run time of 35 min. The monitoring wavelength was set at 276 nm. The sample and reference solution were acidified and injected at a concentration of 100 units/ml [33].

2.2.5 Nuclear Magnetic Resonance (NMR) Spectroscopy

The NMR spectroscopy was performed using Bruker 800 MHz NMR spectrometer (Bruker Biospin, Switzerland, Avance AV 800) at 25°C. The spectrometer was equipped with gradient 5.0 mm probe. The data acquisitions and processing were carried out using Topspin 4.1.4. The NMR spectroscopy was performed using drug substance of insulin aspart extracted from formulation of BGL-ASP and NovoRapid^®. The drug substance of BGL-ASP and reference product were initially reconstituted in 10 mN HCl at a concentration of 10 mg/ml. The stock solution of drug substance was diluted with acetonitrile-d3 to achieve a final concentration of 7 mg/ml and utilized for NMR based analysis. The reconstituted sample was utilized to obtained proton 1D, COSY, TOCSY, NOESY and HSQC spectra. The comparison of the spectra was performed to establish similarity between BGL-ASP and reference product. The acquisition parameters for NMR experiments are listed in the Table 2.

Table 2

NMR Acquisition Parameters
Experimental Name	Encoded Nucleus		Complex Points		Spectral Width (Hz)		Carrier offset (ppm)		Number of Scans, Notes
	F1	F2	F1	F2	SW1	SW2	1H	13C
Proton 1D	¹H	-	16384	-	9615.385	-	4.392	-	64
2D ¹H-¹H COSY	¹H	¹H	2048	512	9615.385	9615.385	4.385	-	16
2D ¹H-¹H TOCSY	¹H	¹H	4096	512	9615.385	9615.385	4.385	-	32, mixing time of 80 ms
2D ¹H-¹H NOESY	¹H	¹H	4096	512	9615.385	9615.385	4.385	-	48, mixing time of 250 ms
2D ¹H-¹³C HSQC	¹H	¹³C	2048	256	9615.385	33195.160	4.382	75.009	64

2.2.6 Fourier-transformed infrared (FTIR) spectroscopy

The FTIR analysis was performed using FT-IR ALPHA II spectrometer from Bruker using attenuated total reflection (ATR) mode, equipped with software OPUS. The BGL-ASP and reference product were subjected to FTIR analysis at a concentration of 3.5 mg/ml, with the IR spectra recorded between 4000 cm^-1 and 600 cm^-1, with a 4-wave-number resolution. Each spectrum was averaged over 32 scans and five such spectra were recorded for each sample.

2.2.7 UV Circular Dichroism Spectroscopy

The UV circular dichroism (UV-CD) spectroscopy based analysis was used to study tertiary and secondary structural elements of insulin aspart. The UV-CD experiments were performed using Jasco J-815 spectrometer. The circular dichroism (CD) spectra in the near-UV region (250 nm – 350 nm) and the far-UV region (190 nm – 250 nm) were recorded at a concentration of 0.2 mg/ml. The spectra were recorded with path length of 0.1 cm and banwidth of 0.1 nm in a quartz cell, at 25^°C with a scanning speed of 200 nm/min. For recording the spectra, 6 accumulations were performed for each sample [34].

2.2.8 Intrinsic Fluorescence Spectroscopy

Intrinsic fluorescence spectroscopy-based analysis was performed using Synergy H1 hybrid Multi-mode reader (Biotek instruments). Intrinsic fluorescence analysis was performed by excitation at 278 nm and emission was scanned from 300 to 400 nm [34].

2.2.9 Differential Scanning Calorimetry (DSC)

Differential scanning calorimetry based analysis was performed using NanoDSC from TA Instruments, New Castle, DE. The instrument was equipped with DSCrun and NanoAnalyzer software. The thermal stability measurements were performed in the range of 20–100^°C at ramp rate of 1.5^°C /min. The BGL-ASP and reference product were subjected to analysis without any dilution. The placebo was used as control. The generated thermograms were compared by Pearson’s correlation testing.

The characterization study focuses on identifying differences between BGL-ASP and reference insulin aspart using state-of-the-art analytical techniques related to structural and physicochemical analysis. The batches used for clinical trials of BGL-ASP were utilized in the characterization studies and hence considered as representative batches of the manufacturing process.

3.1 Higher Order Structural analysis

A native and stable higher-order structure of a biotherapeutic is essential for in-vivo efficiency and safety of the molecule. Biotherapeutics are large molecules with complex three-dimensional structure. The correct three-dimensional structure of the molecule is a critical requirement for the functionality and stability of the molecule. The disulphide linkage and hydrogen bonds predominantly dictate the native conformation of the molecule. The conformation of the molecule might be affected by the production and purification process and hence establishing conformational similarity is a critical requirement. In this study, the primary structure analysis for BGL-ASP was determined using LC-MS and LC-MS/MS methodology. The far-UV CD study (190 nm – 250 nm region) was used to predict the percentage of secondary structural element such as α-helix, β-sheet and random coil. The near-UV CD and intrinsic fluorescence data provides an insight into the tertiary conformation of insulin aspart. The secondary structure for BGL-ASP was determined using FTIR based analysis and the tertiary/ quaternary structure was determined using DSC based analysis. Higher order structural analysis was also aided by NMR based analysis of BGL-ASP.

The excipients present in the formulation tend to interfere with the spectroscopic analysis and hence for CD, Fluorescence and NMR spectroscopy, the drug substance was extracted from the drug product. For techniques such as FTIR and DSC spectroscopy, the drug product was used to identify the structural conformation.

3.1.1 Mass Analysis of Insulin Aspart

The LC-MS based analysis demonstrates that the molecular weight of intact (as shown in Table 3) and reduced insulin aspart (as shown in Table 4) was highly similar for BGL-ASP and reference product. The reduction of insulin aspart results in the formation of A-chain and B-chain and the theoretical molecular weights for the peptides were compared with data obtained for BGL-ASP and reference product.

Table 3

Intact mass data of insulin aspart
Product Details	Theoretical Mass of insulin aspart (Da)	Generated Masses (Da)
Reference insulin aspart	5825.64	5825.51(± 1 Da)
BGL-ASP	5825.64	5825.52 (± 1 Da)

The intact molecular mass of BGL-ASP was determined to be 5825.5 (± 1 Da) and was found to be highly similar to reference insulin aspart.

Table 4

Reduced mass data of insulin aspart
Product Details	A-Chain		B-Chain
	Theoretical mass (Da)	Generated mass (Da)	Theoretical mass (Da)	Generated Mass (Da)
Reference insulin aspart	2382.71	2382.86 (± 1 Da)	3447.93	3447.30 (± 1 Da)
BGL-ASP	2382.71	2382.86 (± 1 Da)	3447.93	3447.31 (± 1 Da)

The molecular mass for A-chain and B-chain of BGL-ASP was determined to be 2382.8 (± 1 Da) and 3447.3 (± 1 Da), respectively and was found to be highly similar to reference insulin aspart.

3.1.2 Peptide Mass Fingerprinting under reducing and non-reducing conditions

The Staphylococcus strain V8 protease specifically cleaves peptide bonds on the carboxyl side of aspartic and glutamic acid residues and was utilized in peptide mapping of insulin aspart to demonstrate the presence of native disulphide bonds. The Fig. 3 demonstrates the structure of insulin aspart with the potential cut sites and the fragments generated pre and post reduction reaction with dithiothreitol (DTT). Comparison of the RP-HPLC profile of the peptide mixture for BGL-ASP and reference product and sequencing of the peptides formed assisted in demonstrating the presence of native disulphide bonds and confirmation of the amino acid sequence.

The Fig. 3 demonstrates the structure of insulin aspart held intact by the presence of two interchain and an intrachain disulphide bond. The enzymatic digestion with V8 protease resulted in the formation of fragments/peptides such as F1, F2, F3 and F4. The fragments F1 and F2 are held together by the presence of the interchain disulphide bonds. The reduction of the disulphide bonds resulted in the formation of additional fragments F1A, F1B, F2A and F2B. Based on the generation of a particular set of peptides under non-reducing and reducing conditions, the presence of native disulphide linkages was confirmed.

RP-HPLC profile of V8 enzyme digested Insulin Aspart

The Fig. 4 demonstrates a comparative RP-HPLC profile of the peptide mixture generated post V8 enzymatic digestion of BGL-ASP and reference product. The profile shows high similarity in peak intensities and retention times. The trace peaks were also comparable and no new peaks were apparent in BGL-ASP chromatograms, on comparison with the reference product.

The Fig. 4A and 4B, demonstrates a comparative RP-HPLC profile for peptides formed post V8 enzymatic digestion for BGL-ASP and reference insulin aspart, under non-reducing and reducing conditions, respectively. The chromatograms were found to be highly similar and the generated mass of the fragments determined using LC-MS analysis (shown in Table 5 and 6) were found to be identical to the theoretical mass, thus confirming the presence of the native disulphide linkage and similarity to reference product. The RP-HPLC retention times for the generated fragments under non-reducing and reducing conditions were also found to be similar.

Table 5

Amino acid sequence and retention time profile of peptide fragments formed under non-reducing conditions.
Peptide Details	Amino Acid Composition	Theoretical Mass (Da)	Generated Mass (Da)	Retention Time of the peptides (min)
Fragment 4	GIVE	416.22	416.22 (± 1 Da)	2.0 ± 0.1
Fragment 1	QCCTSICSLYQLEFVNQHLC GSHLVE	2967.30	2968.23 (± 1 Da)	31.0 ± 0.1
Fragment 2	NYCNALYLVC GE	1376.57	1376.57 (± 1 Da)	14.0 ± 0.1
Fragment 3	RGFFYTDKT	1134.20	1133.55 (± 1 Da)	8.3 ± 0.1

The Table 6 demonstrates the RP-HPLC retention times for the fragments generated under reducing conditions.

Table 6

Amino acid sequence and retention time profile of peptide fragments formed under reducing conditions.
Peptide Details	Amino Acid Composition	Theoretical Mass (Da)	Generated Mass (Da)	Retention Time of the peptides (min)
Fragment 4	GIVE	416.22	416.22 (± 1 Da)	2.3 ± 0.1
Fragment 1B	QCCTSICSLYQLE	1490.70	1488.84 (± 1 Da)	12.6 ± 0.1
Fragment 2A	NYCN	494.50	498.17 (± 1 Da)	20.1 ± 0.1
Fragment 3	RGFFYTDKT	1134.20	1133.58 (± 1 Da)	10.6 ± 0.1
Fragment 2B	ALYLVCGE	867.00	865.43 (± 1 Da)	4.0 ± 0.1
Fragment 1A	FVNQHLCGSHLVE	1481.60	1480.79 (± 1 Da)	31.3 ± 0.1

The fragments shown in Table 6 were subjected to LC-MS/MS analysis and it was observed that the amino acid sequence of BGL-ASP and reference product are highly similar, thus confirming the primary structure of the molecule.

3.1.3 UV Circular Dichroism Spectroscopy

The comparative far and near-UV CD spectra for BGL-ASP and reference product is shown in Fig. 5. The far-UV CD spectra of BGL-ASP and reference marketed insulin aspart were found to be highly similar with characteristics spectra of an alpha helix with positive maxima at 193 nm and negative maxima at 208 nm and 222 nm [35]. The near-UV CD spectra of insulin aspart were generated to demonstrate the equivalence of tertiary structures.

Table 7 represents the generated data set from secondary structural characterization of insulin aspart samples by means of far-UV CD spectroscopy.

Table 7

Analytical data of secondary structural elements for insulin aspart
Product Analyzed	Secondary Structural Elements (%)
Product Analyzed	αhelix	βsheets	β turn	Random coil	Total
Reference Product	26.3	51.8	0	22	100
BGL-ASP	26	53	0	21	100

The generated far UV CD data confirmed the presence of highly similar α-helix, β-sheets and random coil content for reference product and BGL-ASP.

3.1.4 NMR Spectroscopy

NMR Spectroscopy is one of the most powerful techniques to decipher the structural information of biomolecules in solution at atomic resolution. Magnetically active nuclei like ¹H, ¹⁵N, and ¹³C behave like a small spinning magnets, and when kept under the influence of an external magnetic field, will precess about the applied magnetic field. The frequency of precession depends upon chemical environment of the atom and gives rise to a particular resonance frequency, known as the Larmor frequency (ω). This frequency also depends on the applied magnetic field. Thus, in order to have uniform scale across different magnetic fields under which NMR data can be measured, the precession frequency is expressed as chemical shift (δ). In case of a protein and peptide, each amino acid is not only influenced by neighbouring amino acids in the polypeptide chain but also by amino acids that come close due to the three-dimensional arrangement in the tertiary structure. This leads to a different local environment for each atom in the protein and hence a different chemical shift.

The chemical shift of the nucleus is extremely sensitive to chemical environment and any subtle changes caused by changes in conformation, aggregation, and/or changes in buffer conditions will manifest as changes in the chemical shifts observed in the NMR spectra. Hence, NMR signals in the signature regions of 1D and 2D NMR spectra act as a fingerprint of the molecule in a particular condition. These signature regions can be used to compare the conformation of peptide molecule in different formulations and assist in establishing the structural similarity. Proton 1D spectrum and a range of 2D NMR spectra can be used as independent methods for assessing structural similarity. The 2D-COSY (homonuclear correlation spectroscopy), 2D-TOCSY (Total Correlation Spectroscopy) and 2D-¹H,¹³C-HSQC (heteronuclear single quantum coherence) experiments provide through-bond correlations. NOESY (Nuclear Overhauser Effect Spectroscopy) experiments provide information of through-space correlation among atoms. The NOESY patterns are used to identify secondary structural features such as α-helices and β-strands for certain regions of the sequence. Similarly, secondary ¹³C chemical shifts from ¹H,¹³C-HSQC provide secondary structure information as a function of the amino acid sequence. The NMR chemical shifts are highly sensitive to changes in the chemical environment of the nucleus and identical spectra of samples unequivocally confirm the identical structure.

Proton 1D spectra of BGL-ASP and Reference Product

The overlay of 1D NMR spectrum for both the samples demonstrated the presence of the hydrogen peak with same integral value, with same line width at definite ppm value. Down fielded signals above 9 ppm and up-fielded signals below 0 ppm, which are indicative of a well folded polypeptide conformation were not observed in the 1D spectrum. Additionally, the spectral dispersion observed in the 1D ¹H spectrum of both the samples indicated that the peptide is likely to be present either in random coil or a helical conformation. The overlay spectra (Fig. 6) revealed that both the spectra were identical and hence indicated the structural similarity of BGL-ASP and reference product.

2D-¹H,¹H-COSY spectra of BGL-ASP and Reference Product

The 2D COSY spectrum provides spectral correlations for protons that are coupled via 3 or less than 3 covalent bonds. In the 2D-¹H,¹H COSY, spectra the peak observed diagonally corresponds to the signals observed in 1D NMR spectra, whilst positions of cross peaks observed in the COSY corresponds to pairs of protons or equivalent groups of protons which are coupled by 3 or less than 3 covalent bonds. The H^N-H^α signature region consists of correlations between backbone amide proton H^N and alpha proton H^α and hence has one signal per amino acid except for proline which lacks the backbone amide proton. The number of cross peaks observed in H^N-H^α signature region of COSY spectrum for both the samples was 42. The expected number of correlations in this region was about 52. The missing peaks could be overlapped or could be broadened due to conformational exchange. The overlay 2D-¹H,¹H COSY spectra (Fig. 7) showed that, the position of diagonal peaks and position of cross peaks for both samples are exactly overlapped with each other and that the number of peaks observed in H^N-H^α signature region of the COSY spectrum were identical. These observations strongly supported the structural similarity of BGL-ASP and reference product.

2D-¹H,¹H-TOCSY spectra of BGL-ASP and Reference Product

In the 2D-¹H,¹H TOCSY spectrum, the peaks observed diagonally correspond to the signals observed in 1D 1H NMR spectrum, whilst positions of cross peaks observed correspond to the side chain protons of the respective aromatic and aliphatic amino acids which are in a network of coupled spins. Thus, the TOCSY spectrum provides correlations between all protons which are part of a coupled network at every proton and enables easy identification of types of amino acids present in the molecules. The H^N-H^α signature region of the TOCSY spectrum is similar to that of COSY spectrum except that the correlations between H^N and H^β of Serine and Threonine residues also show up in this region. The overlay of ¹H,¹H TOCSY spectra (Fig. 8) for both the samples showed that the position of diagonal peaks and cross peaks match exactly with each other. Thus, the TOCSY spectra strongly suggested the structural similarity of BGL-ASP and reference product.

2D-¹H,¹H-NOESY spectra of BGL-ASP and Reference Product

In the NOESY spectrum, positions of cross peaks observed are due to spin coupling resulting from interactions of spins via space. Hence, a NOESY correlation between 2 protons indicates that the pair of protons are close to each other in space, usually < 6 Ǻ. Folding of protein generates secondary and tertiary structure and, in the process, brings different groups of protons close to each other, resulting in different NOESY patterns. The overlay of NOESY spectra (Fig. 9) for the both the samples showed the same peak position in the entire spectra. Additionally, similarity in the NOESY correlations in the H^N-H^α and H^N-H^N regions indicated the similarity of the secondary structure and tertiary structure exhibited by the BGL-ASP and reference product. Similarity in the NOESY spectra confirmed the presence of identical disulphide linkages for BGL-ASP and reference product.

2D-¹H,¹³C-HSQC spectra of BGL-ASP and Reference Product

The 2D ¹H,¹³C-HSQC spectrum provides correlations between protons and the directly attached carbons. Carbon chemical shifts are sensitive to the conformation of the molecule and hence similarity of signals in the ¹H¹³C-HSQC spectrum strongly suggests similarity of the conformation of the molecule. The overlay of ¹H,¹³C-HSQC spectra (Fig. 10) revealed that the carbon chemical shift values for aromatic carbon, alpha carbon and aliphatic carbons of BGL-ASP and reference product were identical which confirmed the conformational similarity of both the molecules.

The different NMR methodologies confirmed that atom to atom spatial positioning was identical for BGL-ASP and the reference product.

3.1.5 FTIR Spectroscopy

FTIR spectroscopy is utilized to study the secondary structure of peptides and proteins as it probes the amide (peptide) bonds responsible for the distinct infrared (IR) signals for the differently folded peptides and proteins. The Amide I, II and III are most commonly used IR spectral regions used for protein structure-function analysis. The FTIR spectrum between 1700 cm^-1 and 1500 cm^-1 consists of the amide I and II bands resulting from the absorption of the carbonyl amide present in the peptide bond. The shape of these bands are analysed for the determination of the secondary structure content [36–38]. The comparative FTIR spectra for BGL-ASP and reference product is shown in Fig. 11.

The amide I band is due to carbonyl stretching vibrations of the peptide bonds and are influenced by the secondary structures such as α-helix, β-sheet, etc. The amide II band is attributed to NH₂ deformation in primary amides and form a mixed vibration of N-H bending and C-N stretching in secondary amides. The position and shape of the amide I and amide II bands was recorded between 1700 cm^-1 and 1500 cm^-1. The position of amide I and amide II bands were identified at 1658.5 cm^-1 and 1547.7 cm^-1, respectively for BGL-ASP and NovoRapid^® batch. The Pearson’s correlation coefficient was used to compare the FTIR spectra and a value closer to + 1.0 indicated a better correlation. The Pearson’s correlation coefficient for these two spectra was found to be 0.9674, thus confirming the high similarity between BGL-ASP and reference product.

3.1.6 Differential Scanning Calorimetry (DSC)

DSC is a useful tool for monitoring equilibrium thermodynamic stability of biomolecules. The DSC technique is routinely utilized to monitor the thermal stability and folding mechanism of biopharmaceuticals. The formulation process and the excipients such as phenol, m-cresol and zinc are critical formulation components related to the conformation of insulin aspart present in solution. The melting temperature (T_m), identified from the DSC thermogram can be effectively compared for reference product and BGL-ASP to demonstrate similarity in hexamer stability in formulation. The Table 8 shows the melting temperature for each of the three phases of insulin aspart.

Table 8

Summary of nanoDSC results.
Phases	Reference Product	BGL-ASP
Phases	T_m (˚C)	T_m (˚C)
Phase I	66.86	67.10
Phase II	76.15	77.36
Phase III	80.26	83.07

The T_m measured by DSC can be utilized as a predictive tool for identifying the impact of aggregation and chemical modifications on higher order structural attributes [39–42]. The DSC data demonstrated the presence of three different melting phases and similar T_m for BGL-ASP and reference product, thus confirming the presence of highly similar quaternary structures. The Pearson’s correlation coefficient for the two thermograms was found to be 0.9893, thus confirming high similarity between BGL-ASP and reference product.

3.1.7 Intrinsic Fluorescence Spectroscopy

The intrinsic fluorescence spectroscopy is utilized to study the tertiary structure of insulin aspart. The fluorescence spectra acts as an orthogonal tool to near-UV CD data and further confirms the native conformation of the molecule. The comparative intrinsic fluorescence spectroscopy data for BGL-ASP and reference product is shown in Fig. 12.

The intrinsic fluorescence spectra was found to be highly similar for reference product and BGL-ASP, thus confirming the similarity in tertiary structure.

3.2 Physicochemical Characterization

The physicochemical characterization of BGL-ASP involves the use of RP-HPLC based analysis to determine the assay, impurity and excipient content. The higher molecular weight protein (HMWP) content is monitored using SE-HPLC based analysis.

3.2.1 Biosimilarity Limits

A quality range approach was used to establish the biosimilarity between BGL-ASP and reference insulin aspart. The quality range limits were determined by analyzing 12 reference product batches and expressing limits as mean ± 3 standard deviation. The values for BGL-ASP batches were expected to fall within the established limits to confirm the similarity [43].

3.2.2 RP-HPLC based Related Impurity Analysis of Insulin Aspart

The related impurity analysis of insulin aspart was performed using the RP-HPLC technique. The typical impurities formed for insulin aspart are the deamidated (A21Asp) and isomerized forms (B28isoAsp, B3Asp, B3isoAsp) and the RP-HPLC analysis was able to resolve these impurities, thus assisting in quantitative analysis and establishing biosimilarity limits. The RP-HPLC based analysis was also utilized in determining the assay content and the concentration of the preservatives used in insulin aspart formulation such as phenol and m-cresol. The comparative RP-HPLC chromatogram for BGL-ASP and reference product is shown in Fig. 13.

The related impurities are marked in the chromatogram and were found to be well resolved. The percentage purity and assay content of reference insulin aspart batches was determined and the biosimilarity limits were calculated using a multiplier of x = 3 (mean ± 3*SD) as a generally appropriate multiplier for the similarity assessment. The biosimilarity limits and the data obtained for percentage purity and assay content for BGL-ASP and reference product is shown in Fig. 14.

The data demonstrates that the percentage purity and assay content of BGL-ASP batches are within the established biosimilarity limits, identified by analyzing multiple batches of reference product.

3.2.3 SE-HPLC based HMWP Analysis in Insulin Aspart

The SE-HPLC based analysis is utilized to determine the HMWP content in insulin aspart formulation. The dimeric and multimeric forms of insulin aspart are resolved using SE-HPLC. The comparative SE-HPLC chromatogram for BGL-ASP and reference product is shown in Fig. 15.

The HMWP are marked in the chromatogram and were found to be well resolved. The HMWP content in reference product and BGL-ASP batches and the biosimilarity limits are demonstrated in Fig. 16.

The HMWP content in BGL-ASP batches was found to be within the established biosimilarity limits, identified by analysing multiple batches of reference product.

3.2.4 Preservative Content, pH and Zinc content in Insulin Aspart

RP-HPLC based methodology has been utilized to determine the preservative content such as phenol and m-cresol in insulin aspart formulations. The Table 9 demonstrates the content of the excipient in reference product and BGL-ASP formulation.

Table 9

Excipient content in reference product and BGL-ASP batches.
Excipient		Identified Range	Minimum-maximum observed range (number of lots tested)
Excipient		Identified Range	Reference Product	BGL-ASP
Phenol	1.41–1.61 mg/ml		1.45–1.55 (12)	1.43–1.61 (9)
m-Cresol	1.61–1.85 mg/ml		1.66–1.79 (12)	1.57–1.79 (9)
pH	7.10–7.50		7.27–7.42 (12)	7.20–7.42 (9)

The phenol, m-cresol content and pH were found to be within the identified range for BGL-ASP. The zinc content in insulin aspart was determined using atomic absorption spectroscopy. The zinc content for BGL-ASP and reference product was found to be within the range of 15–25 ug/100 IU.

3.2.5 Process Related Impurities of Insulin Aspart

The design of the manufacturing process ensures that the process related impurities are effectively removed at varied process stages. A significant reduction in process related impurities was observed at the capture chromatography stage. The process related impurities were mapped throughout the entire manufacturing process, thus providing a detailed understanding of the process capabilities. BGL-ASP was tested for a series of process related impurities tabulated in Table 10.

Table 10

Process related impurities of BGL-ASP
Process related impurities of insulin aspart	Impact	Complies as per established safety limits
Bacterial Endotoxin	Safety and Immunogenicity	✓
Residual Solvent Content		✓
Residual Single Chain Precursor		✓
Residual Host Cell Protein		✓
Residual Host Cell DNA		✓
Residual Enzyme Content		✓

The process related impurities were found to be within the established safety limits in BGL-ASP, thus confirming the product safety.

Physicians play a crucial role in prescribing biosimilar insulin to patients, and it is essential that the health care professionals are convinced about the quality, safety and efficacy of the drugs. Biologics such as insulins are less complex than monoclonal antibodies, however, the complexity is much higher than generic drugs which are typically smaller molecules with simpler structures. Thus, due to the complex biopharmaceutical manufacturing process which is unique for every protein molecule, the concerns of physician and healthcare would generally revolve across physicochemical biosimilarity, batch to batch consistency and similarity in impurity profiles. The onus to resolve these concerns lies with the manufacturer.

The manufacturing process of BGL-ASP is validated and is scientifically understood. The process parameters are evaluated and based on the impact on the critical quality attributes, the critical process parameters are identified. The design space for the critical process parameters is established and continuous monitoring of the process using in-process quality control checks ensures consistent manufacturing of BGL-ASP with established quality characteristics, thus maintaining batch to batch consistency. The similarity assessment studies were focused on identifying differences between the reference product and BGL-ASP and the data demonstrates high level of similarity for the critical quality attributes, thus providing the health care professional with an extensive data set related to the product’s purity, safety and potency.

A series of comprehensive structural, physicochemical and biological comparability (11) studies have been completed for BGL-ASP and reference product sourced from India, Switzerland, Denmark, Netherlands and USA. The data demonstrated high similarity for the structural and physicochemical quality attributes of BGL-ASP and reference product. The LC-MS data for the intact and reduced mass of insulin aspart and the peptide mass fingerprinting demonstrated that the primary structure was highly similar for BGL-ASP and reference product. A high level of similarity at secondary, tertiary and quaternary structural levels was established for BGL-ASP and reference product using state-of-the-art orthogonal spectroscopic tools such as NMR, FTIR, UV-CD, Intrinsic fluorescence and DSC. The physicochemical characterization focusses on impurity profiling, assay and excipient content determination and high similarity was observed for reference product and BGL-ASP. Stability studies such as real time, accelerated, in-use and stress stability studies have been completed for BGL-ASP in comparison to reference product and the rate change profiles were found to be similar. The successful clinical study performed as per EMA/ICH guidelines generates adequate evidence to confirm that the product safety, efficacy and immunogenicity of BGL-ASP was highly similar to the reference insulin aspart product.

The reference product is manufactured in genetically modified strain of Saccharomyces cerevisiae and BGL-ASP was manufactured in E.coli. The renaturation step was identified as a critical step in the manufacturing process of BGL-ASP and the structural characterization data confirmed a high similarity with reference product. The absence of glycosylated impurities in BGL-ASP proved advantageous with respect to purification strategy and the focus of the chromatographic separation was more towards the removal of deamidated and oxidized impurities.

Based on the totality of evidence and the detailed physicochemical and structural characterization, it can be concluded that BGL-ASP is highly similar to reference insulin aspart (NovoRapid^® from NovoNordisk). BGL-ASP thus can provide a safe, efficacious, and cost-effective choice and access to clinicians and patients for controlling blood sugar levels.

This study provides a template for performing biosimilarity studies which involves identifying the QTPP by analysing multiple batches of reference product. The CQAs are identified based on the impact created on product quality, safety and efficacy. The biosimilarity limits and the criteria for establishing similarity are identified. The similarity assessment is completed through a series of comparative physicochemical, structural and biological characterisation studies.

Funding: This research was funded by BioGenomics Limited, India

Conflicts of Interest: None declared.

Availability of data and material: All data generated or analyzed during this study are included in this published article.

Ethics approval: Not applicable.

Consent to participate: Not applicable.

Consent for publication: Not applicable.

Code availability: Not applicable.

Author contributions: Dr. Nikhil Ghade contributed towards the experimental setup, analysis and interpretation of the data. Dr. Nikhil Ghade, Dr. Damodar K. Thappa, Jeseena Lona, Dr. Archana R. Krishnan and Dr. Sanjay M. Sonar contributed to conception, design and development of the molecule.

Acknowledgement: We thank Abhishek Deshpande for continuous support during the development of the molecule.

Basu, S., Yudkin, J. S., Kehlenbrink, S., Davies, J. I., Wild, S. H., Lipska, K. J., Beran, D. Estimation of global insulin use for type 2 diabetes, 2018–30: a microsimulation analysis. The Lancet Diabetes & Endocrinology. (2018); 7(1):25–33 doi:10.1016/s2213-8587(18)30303-6
Hartman, I. Insulin Analogs: Impact on Treatment Success, Satisfaction, Quality of Life, and Adherence. Clinical Medicine & Research. 2008; 6(2): 54–67. doi:10.3121/cmr.2008.793
Furman, B. L. Aspart Insulin☆. Reference Module in Biomedical Sciences. 2017. doi:10.1016/b978-0-12-801238-3.97984-4
US Food and Drug Administration. Center for Drug Evaluation and Research. Application Number: 20–986: Approval letter for NovoLog (Insulin Aspart [rDNA origin] Injection). 2000. In: US Food and Drug Administration [Internet]. Silver Spring, MD: US Food and Drug Administration. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2000/20-986_NovoLog_Approv.pdf
Zoltan Vajo and William C. Duckworth. Genetically Engineered Insulin Analogs: Diabetes in the New Millenium. Pharmacological Reviews. 2000; 52(1): 1–10.
Zoltan Vajo, Janet Fawcett and William C. Duckworth. Recombinant DNA Technology in the Treatment of Diabetes: Insulin Analogs. Endocrine Reviews. 2001; 22(5): 706–717. doi: 10.1210/er.22.5.706
Reflection paper on statistical methodology for the comparative assessment of quality attributes in drug development, EMA/ CHMP/138502/2017. https://www.ema.europa.eu/en/documents/scientific-guideline/reflection-paper-statistical-methodology-comparative-assessment-quality-attributes-drug-development_en.pdf Accessed 26 July 2021.
Sujoy Ghosh, Saptarshi Bose, Sandeep Gowda and Pradip Mukhopadhyay. Biosimilar Insulins – What a Clinician Needs to Know? Indian J Endocrinol Metab. 2019; 23(4): 400–406.
McCamish M, Woollett G. Worldwide experience with biosimilar development. MAbs. 2011;3:209–17.
Alten R, Cronstein BN. Clinical trial development for biosimilars. Semin Arthritis Rheum 2015;44(6 Suppl):S2-8.
International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonized Guideline: Integrated addendum to ICH E6(R1): Guideline for Good Clinical practice E6(R2). 2016 Nov. In: ICH Database [Internet]. Geneva, Switzerland: International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Available from: https://database.ich.org/sites/default/files/E6_R2_Addendum.pdf
Akshay G. Mishra, Rutuja B. Deshmane, Damodar K. Thappa, Jeseena Lona, Nikhil S. Ghade, Sanjay M. Sonar and Archana R. Krishnan. In Vitro Biological Characterization of Recombinant Insulin Aspart from Biogenomics and Originator Insulin Aspart. BioDrugs. 2023; 37: 709–719
International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonised Tripartite Guideline: Pharmaceutical Development Q8(R2). 2009 Aug [cited 30 Oct 2020]. In: ICH Database [Internet]. Geneva, Switzerland: International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Available from: https://database.ich.org/sites/default/files/Q8_R2_Guideline.pdf
International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonised Tripartite Guideline: Quality Risk Management Q9. 2005 Nov 9 [cited 30 Oct 2020]. In: ICH Database [Internet]. Geneva, Switzerland: International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Available from: https://database.ich.org/sites/default/files/Q9%20Guideline.pdf.
Sarah Demmon, Swapnil Bhargava, Doreen Ciolek, Jennifer Halley, Nomalie Jayab, Marisa K. Jouberte, Edward Koepf, Phillip Smith, Melody Trexler-Schmidt, Philip Tsai. A cross-industry forum on benchmarking critical quality attribute identification and linkage to process characterization studies. Biologicals. 2020; 67: 9–20. doi:10.1016/j.biologicals.2020.06.008
Guidance for Industry. Reference Product Exclusivity for Biological Products Filed Under Section 351(a) of the PHS Act. U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research (CDER) Center for Biologics Evaluation and Research (CBER). August 2014. Available from: https://www.fda.gov/media/89049/download
Hjorth, C. F., Hubálek, F., Andersson, J., Poulsen, C., Otzen, D., & Naver, H. Purification and Identification of High Molecular Weight Products Formed During Storage of Neutral Formulation of Human Insulin. Pharmaceutical Research. 2015; 32(6): 2072–2085. doi:10.1007/s11095-014-1600-3
Guideline on similar biological medicinal products containing biotechnology-derived proteins as active substance: quality issues (revision 1). Committee for Medicinal Products for Human Use (CHMP) 22 May 2014. EMA/CHMP/BWP/247713/2012. Available from: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-similar-biological-medicinal-products-containing-biotechnology-derived-proteins-active_en-0.pdf
Mette Uve Jars, Aage Hvass and Dorte Waaben. Insulin Aspart (AspB28 Human Insulin) Derivatives Formed in Pharmaceutical Solutions. Pharmaceutical Research. 2002; 19 (5).
Chen Zhou, Wei Qi, E. Neil Lewis, John F. Carpenter. Characterization of Sizes of Aggregates of Insulin Analogs and the Conformations of the Constituent Protein Molecules: A Concomitant Dynamic Light Scattering and Raman Spectroscopy Study. Journal of Pharmaceutical Sciences. 2016; 105: 551–558.
Zhi-Song Qiao, Zhan-Yun Guo, and You-Min Feng. Putative Disulfide-Forming Pathway of Porcine Insulin Precursor during Its Refolding in Vitro. Biochemistry. 2001; 40: 2662–2668
Jeannette Winter, Hauke Lilie, and Rainer Rudolph. Renaturation of human proinsulin—a study on refolding and conversion to insulin. Analytical Biochemistry. 2002; 310 (2): 148–155. doi:10.1016/s0003-2697(02)00287-7
Castellanos-Serra, L.R., Hardy, E., Ubieta, R., Vispo, N.S.,Fernandez, C., et al. Expression and folding of an interleukin-2-proinsulin fusion protein and its conversion into insulin by a single step enzymatic removal of the C-peptide and the N-terminal fused sequence. FEBS Lett. 1996; 378: 171–176.
Cowley, D.J. and R.B. Mackin. Expression, purification and characterization of recombinant human proinsulin. FEBS Lett. 1997; 402(2–3): 124–130
Robert B. Mackin. Streamlined procedure for the production of normal and altered version of recombinant human proinsulin. Protein Expression and Purification. 1998; 15: 308–313.
Zhi-Song Qiao, Cheng-Yin Min, Qing-Xin Hua§, Michael A. Weiss§, and You-Min Feng. In Vitro Refolding of Human Proinsulin-Kinetic Intermediates, Putative Disulfide-forming pathway, Folding initiation site, and potential role of C-peptide in folding process. 2003; 278: 17800–17809.
Singh, S. M., & Panda, A. K. Solubilization and refolding of bacterial inclusion body proteins. Journal of Bioscience and Bioengineering. 2005; 99(4): 303–310. doi:10.1263/jbb.99.303
Luis Felipe Vallejo and Ursula Rinas. Strategies for the recovery of active proteins through refolding of bacterial inclusion body proteins. Microbial Cell Factories. 2004; 3(1): 11. doi:10.1186/1475-2859-3-11
Marcus Degerman, Niklas Jakobsson, Bernt Nilsson. Modeling and optimization of preparative reversed-phase liquid chromatography for insulin purification. Journal of Chromatography A. 2007; 1162(1): 41–49. doi:10.1016/j.chroma.2007.02.062.
LAEMMLI, U. K. Cleavage of Structural Proteins during the Assembly of the Head of Bacteriophage T4. Nature. 1970; 227(5259): 680–685. doi:10.1038/227680a0 10.1038/227680a0.
Okita, N., Higami, Y., Fukai, F., Kobayashi, M., Mitarai, M., Sekiya, T., & Sasaki, T. Modified Western blotting for insulin and other diabetes-associated peptide hormones. Scientific Reports 2017; 7(1). doi:10.1038/s41598-017-04456-4.
US Food and Drug Administration. Center for Drug Evaluation and Research. Application Number: 20–986: Final Printed Labeling for NovoLog (Insulin Aspart [rDNA origin] Injection). 2000. In: US Food and Drug Administration [Internet]. Silver Spring, MD: US Food and Drug Administration. Available from: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2000/20-986_NovoLog_prntlbl.pdf
European Pharmacopeia. 11.0. France: EDQM Council of Europe; 2019. Insulin Aspart; 3086–3087.
Goyal P, Pai HV, Kodali P, Vats B, Vajpai N, Annegowda S, et al. Physicochemical and functional characterization of MYL-1501D, a proposed biosimilar to insulin glargine. PLoS ONE. 2021; 16(6): e0253168. https://doi.org/10.1371/journal. pone.0253168
Greenfield, N. J. Using circular dichroism spectra to estimate protein secondary structure. Nature Protocols. 2007;1(6): 2876–2890. doi:10.1038/nprot.2006.202
Groß, P. C., & Zeppezauer, M. Infrared spectroscopy for biopharmaceutical protein analysis. Journal of Pharmaceutical and Biomedical Analysis. 2010; 53(1): 29–36.doi:10.1016/j.jpba.2010.03.009
Derenne, A., Derfoufi, K.-M., Cowper, B., Delporte, C., & Goormaghtigh, E. FTIR spectroscopy as an analytical tool to compare glycosylation in therapeutic monoclonal antibodies. Analytica Chimica Acta.2020; 1112:62–71. doi:10.1016/j.aca.2020.03.038.
Basic Aspects of the Technique and Applications of Infrared Spectroscopy of Peptides and Proteins. ACS Symposium Series. 2000; 750:2–37.
Arthur, K. K., Dinh, N., & Gabrielson, J. P. Technical Decision Making with Higher Order Structure Data: Utilization of Differential Scanning Calorimetry to Elucidate Critical Protein Structural Changes Resulting from Oxidation. Journal of Pharmaceutical Sciences.2015;104(4):1548–1554. doi:10.1002/jps.24313
Dzwolak, W., Ravindra, R., Lendermann, J., & Winter, R. Aggregation of Bovine Insulin Probed by DSC/PPC Calorimetry and FTIR Spectroscopy. Biochemistry (2003); 42(38): 11347–11355. doi:10.1021/bi034879h
Johnson, C. M. Differential scanning calorimetry as a tool for protein folding and stability. Archives of Biochemistry and Biophysics. (2013); 531(1–2):100–109. doi:10.1016/j.abb.2012.09.008
Lisi, G. P., Png, C. Y. M., & Wilcox, D. E. Thermodynamic Contributions to the Stability of the Insulin Hexamer. Biochemistry. (2014); 53(22): 3576–3584. doi:10.1021/bi401678n
Tsong, Y., Dong, X., & Shen, M. Development of statistical methods for analytical similarity assessment. Journal of Biopharmaceutical Statistics. 2016; 27(2); 197–205. doi:10.1080/10543406.2016.1272

No competing interests reported.

Comparative physicochemical and structural characterisation studies establish high Biosimilarity between BGL-ASP and Reference Insulin Aspart

Status:

Journal Publication

Version 1

Abstract

Figures

1 Introduction