According to the Centers for Disease Control and Prevention (CDC), diabetes affected over 38.4 million individuals in the US in 2021, representing 11.6% of the population (1). Type 1 diabetes (T1D) is an autoimmune disease characterized by the destruction of pancreatic β-cells, resulting in an inability to produce insulin. For individuals diagnosed with T1D, insulin replacement therapy remains the only viable treatment (2). Type 2 diabetes (T2D), on the other hand, is a complex metabolic disorder associated with insulin resistance and impaired glucose regulation, often linked to obesity and aging. While T2D patients still produce insulin, many require insulin therapy to achieve desired glycemic control, due in part to the development of insulin resistance (3, 4).

In healthy individuals, insulin secretion from the pancreas is precisely regulated by blood glucose levels through a finely tuned feedback mechanism. The pancreatic β-cells within the islets of Langerhans act as glucose sensors, detecting fluctuations in circulating glucose concentrations and responding accordingly (5). After a meal, glucose levels rise, prompting β-cells to secrete insulin in a biphasic manner. The first phase is a rapid release of pre-stored insulin granules that occur within minutes, followed by a more prolonged second phase with sustained insulin secretion and de novo insulin synthesis (6). Conversely, during fasting or periods of low glucose availability, insulin secretion decreases to a basal rate sufficient for maintaining glucose homeostasis (7).

Endogenous insulin is stored in pancreatic β-cells as hexamers (insulin clusters of six molecules) stabilized by zinc ions within secretory granules (8). These hexamers serve as an inactive storage form, preventing premature insulin degradation (premature breakdown of insulin molecules that may affects its biological activity) and facilitating controlled release. Upon stimulation by elevated blood glucose, insulin-containing granules undergo exocytosis, releasing insulin into the bloodstream. Once in circulation, the stabilized hexamers dissociate into monomers (individual insulin molecules), converting insulin into its biologically active form (9). Monomeric insulin first exerts its effects on the liver, the primary site of insulin metabolism (10). Hepatic insulin action is crucial for maintaining glucose balance, as insulin suppresses glycogenolysis (breakdown of stored glycogen into glucose) and inhibits gluconeogenesis (production of new glucose); by reducing hepatic glucose output, insulin helps prevent excessive postprandial hyperglycemia (10). The liver also plays a key role in insulin clearance, degrading up to 80% of the insulin it receives before the hormone reaches the systemic circulation (11). The remaining insulin is distributed to peripheral tissues, where insulin facilitates glucose uptake, suppresses triglyceride breakdown in adipose tissue, enhances energy metabolism, and is ultimately cleared by the kidneys (10).

Structurally, insulin is a polypeptide hormone composed of 51 amino acids, arranged in two distinct chains: the A chain (21 amino acids) and the B chain (30 amino acids) (12). These chains are connected by two interchain disulfide bonds (CysA7–CysB7 and CysA20–CysB19), which are necessary for insulin’s stability and bioactivity. The two interchain disulfide bridges, an additional intrachain disulfide bond (CysA6–CysA11) within the A chain along with the A and B chain primary sequences maintain the higher order structure of insulin that ensures binding to specific receptors and concomitant insulin functions (13). Overall, the three-dimensional structure of the polypeptide is essential for insulin’s ability to regulate glucose metabolism effectively.

Since the discovery of insulin in the 1920s, our understanding of diabetes and insulin therapy has advanced significantly (4). Early commercial insulin was extracted from porcine or bovine pancreatic tissue, requiring extensive purification to remove contaminants that could trigger immune responses. Bovine insulin differs from human insulin by three amino acids, while porcine insulin differs by a single amino acid at position B30 (alanine instead of threonine), making the porcine sequence less immunogenic but still capable of eliciting antibody formation and allergic reactions [Table 1; (15, 16)]. Early purification methods, such as ethanol precipitation and acid-alcohol extraction, were crude, often leaving behind trace amounts of proinsulin, glucagon, and other pancreatic proteins, contributing to increased immunogenicity (17), and batch-to-batch content variability. By the mid-20th century, insulin purification techniques were optimized with the introduction of gel filtration, ion-exchange chromatography, and crystallization methods, which helped improve the purity of insulin preparations (18). These purification improvements resulted in a 60% reduction in injection site reactions and a 40% decrease in insulin antibody formation compared to earlier preparations, significantly improving patient tolerability and therapeutic outcomes (15, 16). In parallel, formulation science introduced hexamer-stabilizing agents (e.g., zinc, phenol, and protamine) to control insulin release kinetics (the rate at which insulin becomes available in the body). The full sequence of bovine insulin was published in several articles by Frederic Sanger between 1945 and 1955, identifying the peptide amino acid composition (19–26). In 1969, Dorothy Crowfoot-Hodgkin resolved a three-dimensional structure of porcine insulin using X-ray crystallography (27). These breakthroughs along with technological advances paved the way for the development of the first fully synthetic insulin in 1975 by Ciba-Geigy and the production of the first recombinant human insulin using E. coli in 1978 by David Goeddel and colleagues (4). The introduction of recombinant DNA technology in the 1980s revolutionized insulin manufacturing, providing a safer and more consistent alternative to animal-derived insulin while reducing cross-species immunogenicity concerns (4).

The 1980s and 1990s marked a transformative period in insulin therapy with the advent of genetically modified insulin analogs, designed to optimize absorption, metabolism, and clearance. With utilization of recombinant DNA technology, insulin analogs were engineered to enhance pharmacokinetics, resulting in more predictable glucose control and improved patient outcomes. Utilizing recombinant DNA technology, Eli Lilly introduced the first recombinant human insulin, Humulin R^®, in 1982 and the first recombinant human insulin NPH, Humulin N, also in 1982 [Figure 1, (28, 29)]. The introduction of recombinant human insulin eliminated immunogenicity concerns associated with animal-derived products and provided consistent, reliable insulin therapy (30). This was followed by NovoNordisk in 1991 with a recombinant human insulin, Novolin R^®, and a recombinant human insulin NPH, Novolin N [Figure 1, (31, 32)], expanding patient access to human insulin products.

The first rapid-acting insulin, insulin lispro (Humalog^®) was released in 1996 by Eli Lilly, followed by Novo Nordisk’s rapid-acting insulin aspart (Novolog^®) in 2000 (33, 34). In that same year, Sanofi revolutionized basal insulin therapy with the introduction of insulin glargine (Lantus), the first long-acting insulin analog.

In subsequent years, additional analogs emerged, including insulin glulisine (Apidra^®), another rapid-acting insulin, as well as long-acting formulations such as insulin detemir (Levemir^®) and insulin degludec (Tresiba^®) (Figure 1). Multiple variants of these insulins have also been marketed (Figures 1, 2), including several “follow-on” products such as Admelog and Basalgar, which are insulin products that have been approved as biosimilars outside of the USA (35, 36). The first interchangeable biosimilar for insulin glargine, insulin glargine-yfgn U100 (Semglee^®) from Biocon was approved in 2020, providing alternatives while maintaining therapeutic equivalence (37). A non-exclusive timeline of prevalent US-FDA approved insulin products is shown in Figure 1 and highlights key milestones in the evolution of insulin therapy. A diagram presenting a categorized overview of prevalent commercial insulin products, grouped by insulin type is shown in Figure 2.

Like many proteins, insulin in aqueous solutions can degrade through deamidation (loss of amino group) or aggregation (when insulin molecules clump together to form larger structures). At low pH, deamination has been shown to primarily affect asparagine at A21, whereas under neutral or alkaline conditions, asparagine at B3 has been observed to be more susceptible to deamination (38). To maintain stability and potency, insulin production requires purification steps to eliminate impurities and tightly controlled formulations to maintain hormone stability and potency. Overall, advances in purification methodologies, including chromatography and filtration techniques, have been instrumental in producing more stable and potent insulin products with improved shelf-life and reduced immunogenic potential (4). Modern commercial insulin and insulin analog formulations typically include stabilizing agents such as zinc, preservatives, and buffers (Table 2). Upon subcutaneous injection, these additives facilitate the controlled dissociation of hexameric insulin (storage form) into its monomeric state (active form), enhancing insulin’s absorption into the bloodstream and uptake by the liver (39).

Commercial insulin products differ in structure, formulation and/or presentation (Table 2). These differences can be observed by analytical procedures sensitive to alteration in higher order structure (40–43). In addition, these differences can be observed using pharmacometrics data, such as pharmacokinetics (PK) and pharmacodynamics (PD), which are used to define the time-action profile of insulin products [ (44), Tables 3–5]: i) onset of action (the time it takes for the blood glucose to lower after injection); ii) time to peak (the time taken to reach maximum effect); and iii) duration of action (how long the effects last after an injection) (45). The ability of insulin to lower blood glucose depends on its ability to transition from hexameric storage forms to active monomers. Since only insulin monomers can pass through the capillary walls, the time to onset depends on the strength of the interactions that bind the insulin hexamers in the subcutaneous environment. When these interactions are strong, the time to onset and the duration of action can take longer as the hexamers need to be broken down into monomers before entering the bloodstream (45). Conversely, weaker interactions facilitate faster monomer release, resulting in a more rapid onset and shorter duration of action [refer to these comprehensive reviews for detailed visual comparisons of onset, peak, and duration among different insulin types (44, 45)]. To optimize insulin PK/PD properties, structural modifications and formulation advancements have been introduced. For example, rapid-acting insulins incorporate amino acid substitutions that reduce self-association, accelerating monomer release, while long-acting insulins are engineered with fatty acid modifications or amino acid substitutions that enhance hexamer stability and prolong absorption [refer to Table 2 and (12)]. Additionally, formulation excipients such as zinc, phenol, and protamine are used to modulate insulin aggregation and release kinetics [refer to Table 2 and (46)]. Based on their time-action profiles, insulin products can be categorized into four major groups: rapid-acting, short-acting, intermediate-acting, and long-acting insulins (44). These classifications guide clinical decision-making, allowing for the selection of insulin therapies that align with patient-specific glucose management needs.

Diabetes affects more than 39 million people in the United States and imposes an estimated annual cost of $412.9 billion (as of 2022), representing a significant economic burden through increased medical expenses, lost productivity, and reduced quality of life (80). Given insulin’s critical role in diabetes management, extensive efforts have been made to improve purification and develop formulations that better align with patients’ physiological needs. Advances in insulin therapy have led to a diverse range of products, from short-acting and rapid-acting insulins designed to mimic endogenous postprandial insulin secretion, to intermediate- and long-acting formulations that provide stable basal insulin levels (Tables 2-4). Structural modifications, such as amino acid substitutions and acylation, influence self-association and receptor binding, while formulation components like zinc, phenol, and protamine regulate aggregation and release kinetics (Table 1). These advancements have not only enhanced glycemic control but have also improved patient compliance by reducing injection frequency and minimizing the risk of hypoglycemia. However, insulin therapy still faces significant limitations. Patient variability in absorption, metabolism, and insulin sensitivity creates challenges in achieving optimal glycemic control for all individuals (62, 81, 82). Adherence challenges persist, particularly with complex regimens requiring multiple daily injections, timing considerations, and glucose monitoring. Additionally, the risk of hypoglycemia, weight gain, injection site reactions, and lack of physiologic insulin delivery patterns continue to impact patient quality of life and treatment satisfaction. The future of insulin therapy promises exciting innovations that address current limitations. Weekly insulin formulations currently in Phase III clinical trials aim to reduce injection frequency to once per week, potentially enhancing patient compliance and quality of life while addressing adherence challenges and reducing injection site complications (83). Smart delivery systems in various development stages from preclinical to early clinical trials, including closed-loop insulin pumps and glucose-responsive insulin formulations, are being developed to provide automated, physiologic insulin delivery that minimizes hypoglycemia risk and reduces the need of additional glucose monitoring (84, 85). Oral insulin formulations currently in Phase II/III clinical trials, despite decades of challenges, continue to show promise with novel delivery technologies and absorption enhancers that could eliminate injection-related complications (86, 87). Additionally, personalized insulin therapy approaches using pharmacogenomics and artificial intelligence may optimize dosing and timing for individual patients, minimizing adverse effects while maximizing therapeutic benefits (88, 89). With numerous recombinant insulin innovators and biosimilar products currently marketed in the USA, ongoing manufacturing and product quality research continue to refine insulin therapies, addressing the clinical and economic burden of diabetes while improving the quality of insulin products and, by extension, the quality of life for millions of patients. The continued evolution of insulin therapy represents one of medicine’s greatest success stories, with future innovations promising even greater improvements in diabetes care and patient outcomes.