Most insulin used by people with diabetes today is made by genetically engineered bacteria or yeast in large manufacturing facilities. These microorganisms are programmed with the human gene for insulin, so they produce a molecule identical to what a healthy pancreas would make. This is a dramatic shift from insulin’s early decades, when every vial came from the ground-up pancreases of cows and pigs.
The Original Source: Animal Pancreases
From the 1920s through the early 1980s, all commercial insulin was extracted from the pancreatic tissue of cattle and pigs, sourced from slaughterhouses. Pork insulin was the closer match to human insulin, differing by just one amino acid. Beef insulin differed by three amino acids. Both worked well enough to keep people alive, but the slight molecular differences meant some patients developed immune reactions, including antibodies that made the insulin less effective over time. Early animal-derived insulin was also only about 70% pure, which contributed to injection-site reactions and other side effects.
How Bacteria and Yeast Make Human Insulin
The shift away from animal sources began in 1978 with recombinant DNA technology. Scientists designed synthetic genes matching the two protein chains that make up human insulin (called the A chain and B chain), using genetic sequences optimized for the common gut bacterium E. coli. These synthetic genes were inserted into circular pieces of bacterial DNA called plasmids, which act like instruction manuals the bacteria follow as they grow and divide.
Two separate strains of E. coli are each given one of the two chain genes. As the bacteria multiply in fermentation tanks, they churn out large quantities of each chain as part of a larger fused protein. Technicians then chemically snip the insulin chain free, purify the A and B chains separately, and combine them through a chemical reaction that links them together, just as they would be in the human body.
Yeast cells (specifically baker’s yeast, Saccharomyces cerevisiae) are also widely used as an alternative to bacteria. The yeast-based process has a practical advantage: yeast cells can fold the insulin molecule into its correct three-dimensional shape and secrete it directly into the surrounding liquid. The bacterial method, by contrast, requires extra steps to unfold and refold the protein into its functional form. Both systems produce insulin that is chemically identical to what the human pancreas makes.
From “Human Insulin” to Insulin Analogs
Once scientists could manufacture exact copies of human insulin, the next step was improving on it. The pancreas releases insulin in precise bursts timed to meals and also maintains a low, steady background level throughout the day. A single type of injected insulin can’t replicate both patterns well, so researchers began tweaking the amino acid sequence to change how the molecule behaves after injection.
Rapid-acting insulins are designed to break apart quickly at the injection site so they absorb into the bloodstream fast, mimicking the burst your pancreas would release when you eat. Lispro (sold as Humalog) swaps the position of two amino acids near the end of the B chain. Aspart (NovoLog) replaces one amino acid at the same spot with a different one. These small changes prevent insulin molecules from clumping together, which means individual molecules reach your blood within minutes rather than the 30 to 60 minutes regular human insulin takes.
Long-acting insulins take the opposite approach. Glargine (Lantus, Toujeo, Basaglar) adds two extra amino acids to the B chain and swaps one on the A chain. These changes cause the insulin to form tiny crystals under the skin that dissolve slowly over roughly 24 hours, providing a steady baseline. Degludec (Tresiba) attaches a fatty acid chain to the insulin molecule, which causes it to bind to a protein in the blood and release gradually over an even longer period. All of these analogs start as the same recombinant manufacturing process in bacteria or yeast, with the genetic instructions modified to include the desired amino acid changes.
Purity Standards for Modern Insulin
Today’s insulin products must meet strict purity requirements set by the U.S. Pharmacopeia. Unwanted protein contaminants, including larger protein fragments and breakdown products, are each capped at no more than 1% of the total. Proinsulin, a precursor molecule the body normally processes before insulin is active, must be present at no more than 10 nanograms per milligram. Bacterial toxins are limited to no more than 10 endotoxin units per milligram, and total bacterial counts cannot exceed 300 colony-forming units per gram. For insulin derived from a single species (relevant to the small remaining supply of animal-derived products), cross-contamination from another species must stay below 1%. These thresholds are orders of magnitude stricter than what was achievable in the early decades of insulin therapy.
Biosimilar Insulins and Pharmacy Substitution
Because insulin is a biological product rather than a simple chemical compound, generic versions are called “biosimilars.” A biosimilar insulin must demonstrate it is as safe and effective as the original brand-name product it references. Some biosimilars earn an additional designation: “interchangeable.” An interchangeable biosimilar can be substituted at the pharmacy counter without needing the prescribing doctor’s approval, similar to how a generic pill replaces a brand-name one. This distinction matters practically because it can mean lower out-of-pocket costs without an extra doctor visit. Both biosimilars and interchangeable biosimilars undergo rigorous testing to confirm they perform the same way in the body.
Experimental Sources on the Horizon
Researchers are exploring whether plants could serve as a cheaper insulin factory. In laboratory studies, tobacco and lettuce plants genetically modified to produce proinsulin accumulated the protein at remarkably high levels: up to 47% of total leaf protein in tobacco and 53% in lettuce. When this plant-produced proinsulin was given to diabetic mice, either by injection or orally (with the plant cell walls protecting it through digestion), it lowered blood glucose levels comparably to commercial insulin. Plant-based production could theoretically reduce manufacturing costs dramatically, since growing plants requires far less infrastructure than maintaining sterile bacterial fermentation facilities. This approach remains in the research stage and is not yet available for human use.