Insulin The first peptide medicine, the analogue family that followed, and the standard every later peptide is measured against

01Fifty-one residues in two chains

Insulin is not a short peptide in the modern marketing sense. It is a two-chain protein hormone: an A chain of 21 amino acids and a B chain of 30, joined by two interchain disulfide bridges (A7–B7 and A20–B19), with a third disulfide inside the A chain (A6–A11). The human gene is INS; the UniProtKB accession for the preproprotein is P01308. Those identity facts organise every later sentence in this file: when the prose says “insulin,” it means this substance — endogenous human insulin or pharmaceutical regular / human insulin — unless an analogue is named.

A 2003 mini-review of the human insulin genome map and the biochemical structure underlying recombinant DNA insulin production restates the same architecture in the language of industrial biotechnology (Chakraborty et al., 2003). The point for a research monograph is not novelty; it is boundary. Later peptide drugs will be smaller, longer-acting, or multi-agonist. Insulin remains the reference case of a disulfide protein that became a medicine.

021921–1923: extract, clinic, and the idea of a peptide hormone

The founding clinical story is older than the sequence. In a 1945 Banting Memorial Lecture, Best reviewed insulin and diabetes in retrospect and prospect — a primary-era voice looking back across the first quarter-century of the hormone as a clinical substance (Best, 1945). A 1954 historical note titled “Banting, insulin and diabetes” records the same founding names in the mid-century literature (Gill, 1954). A 2020 pathology review returns to the laboratory and autopsy contributions that made the discovery possible (Wright, 2020). None of these papers is a modern randomised trial; they are cited here as historical and secondary historical instruments, labelled as such.

What matters for the peptide series is the category they opened. A tissue extract was refined into a glucose-lowering medicine decades before solid-phase peptide synthesis or recombinant expression existed. Everything later in this document — hexamers, analogues, essential-medicine status — sits on that category change.

03Sequence, crystals, and what “knowing insulin” meant

Insulin was among the first proteins whose amino-acid sequence and three-dimensional structure became public scientific property. The crystallographic line associated with Hodgkin’s laboratory is represented in the harvest by mid-1960s zinc-insulin crystal studies: X-ray work on zinc insulin crystals (Adam et al., 1966), and the Journal of Molecular Biology crystal-structure papers that report packing and symmetry evidence for the rhombohedral form (Harding et al., 1966; Dodson et al., 1966). Those are structural papers, not clinical ones. They are why later formulation science could talk about hexamers without waving its hands.

Sanger’s sequencing of insulin is the other half of that “knowing.” The reading-set harvest is thin on the original Sanger titles themselves; this build therefore does not invent PMIDs for those papers. The recombinant-era reviews below treat the completed sequence as the prerequisite for biosynthetic human insulin (Chance et al., 1981; Chance et al., 1993; Francis et al., 2024). Where a plate needs the Sanger primary, the Apparatus will say the primary is still to be promoted into the local corpus.

04From animal glands to recombinant human insulin

For decades the clinical supply was animal insulin. Recombinant DNA technology changed the industrial source without changing the human sequence of regular insulin. Chance and colleagues described the chemical, physical and biologic properties of biosynthetic human insulin as the product entered the literature (Chance et al., 1981), and later reviewed research, development, production and safety of that product class (Chance et al., 1993). Contemporary reviews of biosynthetic human insulin and its analogues restate the same industrial arc (Rodbard et al., 2020; Francis et al., 2024).

The regulatory year this monograph locks for Humulin R is taken from the retained DailyMed SPL, which states Initial U.S. Approval: 1982 for insulin human. Broader slogans such as “first recombinant DNA medicine” remain historically important but are not printed here as a plate fact beyond what that label year and the biosynthetic literature support. Genentech–Lilly process history belongs in Apparatus notes until an approval-package PDF is retained locally.

What the recombinant gate changed for every later peptide file is simple to state and easy to forget. Before 1982, a peptide medicine meant extraction or chemical synthesis at punishing scale. After it, sequence and expression became industrial verbs. Chance and Frank’s later production-and-safety review of biosynthetic human insulin is therefore not only a product paper; it is a template paper for the series (Chance et al., 1993). Rodbard and Rodbard’s update of biosynthetic human insulin and analogues restates that template in modern therapeutic language without collapsing analogues into the parent (Rodbard et al., 2020). Francis and colleagues’ 2024 evolution review is the same story with another twenty years of analogue chemistry attached (Francis et al., 2024). Part Three will name those analogues. Part One only needs the gate.

05β-cell production: preproinsulin, proinsulin, C-peptide

Endogenous insulin is not synthesised as the mature two-chain hormone. The β-cell builds a preproprotein, cleaves the signal peptide to yield proinsulin, then excises C-peptide to leave the disulfide-linked A and B chains. That biosynthetic outline is standard endocrinology; this monograph treats it as physiology context for the pharmaceutical substance, not as a dosing story. The structural and recombinant reviews already cited in Part One presuppose the same maturation path when they discuss human insulin as a gene product (Chakraborty et al., 2003; Francis et al., 2024).

C-peptide appears in clinical research as a marker of endogenous secretion. Papers in the harvest that use C-peptide for diabetes classification or hypoglycemia workup are measurement papers; they are not cited here as insulin pharmacology unless the sentence is about the hormone itself.

06The insulin receptor

Insulin acts through the insulin receptor (INSR), a receptor tyrosine kinase. A 2022 review marks fifty years since the receptor’s discovery and surveys what that anniversary means for the field (De Meyts, 2022). A structure-based survey of ligand binding in the human insulin receptor maps how ligands occupy the receptor’s binding surfaces (Kumar et al., 2022). Both are receptor biology: they are not human outcome trials, and they do not license claims about athletic performance.

Downstream of INSR, textbook pathways (IRS proteins, PI3K/Akt, MAPK) organise glucose uptake and metabolic gene expression. This monograph does not expand those pathways into a signalling textbook. Where later parts discuss analogues, the load-bearing claim is that the analogues still target the same receptor pharmacophore while changing absorption or depot behaviour (Vajo et al., 2001; Francis et al., 2024).

The receptor anniversary review is also a reminder about study type (De Meyts, 2022). Fifty years of receptor biology produced structural maps, binding assays and signalling models; it did not, by itself, produce a single outcome trial. The structure-based ligand survey likewise stays inside pharmacology (Kumar et al., 2022). Both are load-bearing for Part Two and deliberately silent for Part Four’s clinical harm questions.

That silence is not a defect in the receptor papers. It is a discipline the peptide series needs whenever a molecule has both a deep structural literature and a deep clinical one. Insulin is the extreme case: the hexamer and the essential-medicine list are about the same substance, and mixing their evidence tiers without labels is how slogans get written. The next two sections stay inside assembly chemistry for that reason.

Solution structures of the R6 human insulin hexamer and circular-dichroism work on hexamer dissociation are therefore placed here as biophysical instruments (Chang et al., 1997; Melberg et al., 1990), not as clinical proofs. They answer why a zinc-containing formulation can behave like a depot of hexamers that must come apart before receptor engagement. They do not answer whether any particular product is “better” for a patient. Part Three will need that distinction again when LysB28ProB29 assembly papers are cited beside lispro trial reviews.

07Hexamers, zinc, and why formulation is pharmacology

Pharmaceutical regular insulin is not delivered as a lonely monomer in a vacuum. In zinc-containing formulations the hormone assembles into hexamers; dissociation toward dimers and monomers is part of the absorption story. Solution and crystallographic work on the R6 human insulin hexamer (Chang et al., 1997) and circular-dichroism work showing secondary-structure change as hexamers dissociate (Melberg et al., 1990) are biophysical papers. They explain why chemists could redesign the B-chain C-terminus and expect a kinetic change without inventing a new receptor.

That redesign is the hinge into Part Three. Crystal and assembly studies of LysB28ProB29 insulin (insulin lispro) hexamers show how C-terminal B-chain residues alter assembly (Ciszak et al., 1995; Birnbaum et al., 1997). Those papers name the analogue. They are not evidence about unmarked “insulin.”

08What animal and in-vitro systems can and cannot say

Receptor structures, hexamer crystallography and transgenic immunogenicity models answer chemical and immunological questions. They do not answer human outcome questions by themselves. This document keeps that fence: a mouse immunogenicity assay is labelled as such (Ottesen et al., 1994, in Part Four); a meta-analysis of human trials is labelled as such (Melo et al., 2019, in Part Three). Crossing the fence without a label is the failure mode this series was built to refuse.

09Why “insulin” is not “any insulin product”

Recombinant DNA technology did more than replace animal glands. It made sequence editing a pharmaceutical strategy. An Endocrine Reviews survey of that strategy — insulin analogues produced by recombinant DNA methods — is the conceptual map for this Part (Vajo et al., 2001). Later reviews of biosynthetic human insulin and analogues update the industrial catalogue without collapsing products into one word (Rodbard et al., 2020; Francis et al., 2024).

Hard rule for every sentence below: if the evidence is about insulin lispro, insulin aspart or insulin glargine, the sentence names that analogue. Regular / human insulin remains the parent identity from Parts One and Two.

10Rapid-acting exemplars: lispro and aspart

Insulin lispro. Inverting B28–B29 (Pro-Lys to Lys-Pro) weakens hexamer persistence relative to human insulin. Hexamer crystallography and assembly comparisons of LysB28ProB29 insulin are biophysical evidence for that claim (Ciszak et al., 1995; Birnbaum et al., 1997). A pediatric Type 1 diabetes review covering twenty years of insulin lispro evidence summarises the clinical literature for that analogue (Kaiserman et al., 2017). A basal-bolus review that discusses insulin glargine together with insulin lispro must be read as a joint regimen paper, not as unmarked insulin evidence (Candido et al., 2018). The Humalog DailyMed SPL retained for this build records Initial U.S. Approval: 1996.

Insulin aspart. B28 proline to aspartate is the parallel rapid-acting idea. A biophysical paper on a T3R3 hexamer of the B28Asp human insulin variant sits on the chemistry side (Palmieri et al., 2013). On the clinical side, a comparison of insulin aspart with regular human insulin in Type 1 and Type 2 diabetes (Wojciechowski et al., 2015) and an earlier efficacy and safety comparison of rapid-acting insulin aspart with regular human insulin (Rys et al., 2011) are human clinical evidence about aspart, not about regular insulin alone. The NovoLog DailyMed SPL retained here records Initial U.S. Approval: 2000.

A systematic review comparing short-acting insulin analogues with regular human insulin on postprandial glucose and hypoglycemia is useful precisely because it keeps the comparator named (Melo et al., 2019). It is a review of trial evidence, not a use instruction.

11Basal exemplars: glargine, with class notes

Insulin glargine. The chemical idea is an isoelectric shift: A21 asparagine to glycine plus arginine extension of the B chain, producing a depot after injection. A 2026 “silver jubilee” review frames insulin glargine as the opening of the long-acting analogue era (Bolli et al., 2026). A Type 1 diabetes review of clinical trials and real-world evidence for insulin glargine across two decades is analogue evidence labelled as such (Saboo et al., 2024). The Lantus DailyMed SPL retained here records Initial U.S. Approval: 2000.

Detemir and degludec. Acylation and multi-hexamer strategies appear in the biophysical literature on ligand-controlled hexamer and multihexamer assembly (Steensgaard et al., 2013). This monograph treats them as brief class notes unless a later packet expands them into sibling sections. They are not substitutes for glargine evidence.

12What transfers, and what does not

Head-to-head and analogue-versus-regular comparisons transfer one kind of knowledge: relative postprandial control and hypoglycemia signals under trial conditions (Rys et al., 2011; Wojciechowski et al., 2015; Melo et al., 2019). They do not transfer a license to speak of “insulin” as if every product were interchangeable, and they do not transfer performance or body-composition claims. Where a paper studies a biosimilar of an analogue, the sentence must name the biosimilar and the reference analogue — a discipline already required by the sibling-margin rules in pipeline/compound.py.

Two further boundary cases belong here before Part Four. First, basal-bolus reviews that discuss insulin glargine together with insulin lispro are joint regimen papers (Candido et al., 2018). They are useful for understanding how clinicians combine siblings; they are not evidence that “insulin” as a single substance did both jobs. Second, jubilee and decade reviews of insulin glargine are historical framing devices for one analogue’s class role (Bolli et al., 2026; Saboo et al., 2024). They may summarise many trials; they still do not license renaming glargine as regular insulin in a sentence about those trials.

The biophysical papers on B28Asp hexamers and on ligand-controlled multihexamer assembly sit on the same boundary from the other direction (Palmieri et al., 2013; Steensgaard et al., 2013). They explain why sequence and formulation edits change absorption kinetics. They are not human outcome evidence, and Part Four will not cite them as such.

13Essential-medicine status and the industrial template

Insulin is not only a research molecule. On the WHO essential-medicines portal HTML retained for this build, the Insulins section lists human short-acting and intermediate-acting insulin and rapid-acting and long-acting insulin analogues. That is a public-health classification instrument, not a prescribing guide. It is why later peptide monographs keep measuring themselves against insulin: the series’ founding approved peptide is also an essential-medicine class.

Access and affordability are part of that story. A 2020 review of the origins of the insulin affordability crisis in the United States is policy and health-services evidence (Luo et al., 2020). Biomaterials work on cold-chain resilience framed as a global-access problem is engineering evidence with an access claim (Maikawa et al., 2021). Neither paper is cited here as advice about which product a reader should obtain.

The industrial template matters to the rest of the peptide series for a second reason: once recombinant human insulin existed, analogue engineering became a repeatable playbook rather than a one-off miracle (Chance et al., 1993; Vajo et al., 2001; Francis et al., 2024). Essential-medicine listing and analogue proliferation are therefore not separate stories. They are the same molecule class under two institutions — public-health lists and pharmaceutical chemistry — and Part Three’s sibling margins exist so those institutions do not get conflated in prose.

14Hypoglycemia, weight, and immunogenicity

Hypoglycemia is the dominant mechanism-linked adverse class for insulin therapy. Comparative reviews of short-acting analogues versus regular human insulin report postprandial glucose and hypoglycemia outcomes under trial conditions (Melo et al., 2019). Analogue-specific reviews of insulin lispro and insulin glargine likewise discuss hypoglycemia as part of the clinical evidence package (Kaiserman et al., 2017; Saboo et al., 2024). Product labels retained under instruments/ carry the same adverse class in regulatory language. None of that is rewritten here as a titration schedule.

Immunogenicity was a louder problem in the animal-insulin era than in the recombinant human era, but it did not vanish as a research question. A 1994 Diabetologia study evaluated the potential immunogenicity of human insulin and insulin analogues in a transgenic mouse model (Ottesen et al., 1994). That is animal immunology. Modern biosimilar immunogenicity trials for specific analogues exist in the harvest; they are analogue-named clinical immunology and are not collapsed into unmarked insulin claims in this build.

15Contested long-term safety questions

Mitogenicity and cancer-signal debates around some insulin analogues are conflict literature, not a slogan. This monograph refuses two failure modes: (1) declaring analogues “safe because insulin-like,” and (2) declaring them carcinogenic from selective citation. Retracted items in the reading corpus — including a retracted letter responding to mitogenicity claims — are excluded from supportive citation (see manifest/02e_integrity.json). What remains is a contested literature that must be opened with study type and withdrawn when withdrawn.

16Non-medical misuse literature: evidence boundaries only

Pharmaceutical insulin appears in bodybuilding and performance discourse. Where peer-reviewed material discusses that adjacency, this series may report epidemiology or toxicity case reports with study-type labels. It may not provide doses, stacks, or “how to” language. The RUO constraint in the masthead is absolute for this section. No misuse paper is required for the controlling thesis of this monograph; the section exists to state the fence.

The same fence applies to reading labels for PK language. A DailyMed SPL may state onset, peak and duration figures for a named product. Those figures are regulatory description. Reprinting them here as a how-to would convert a research monograph into a formulary. This file therefore uses labels for identity, approval year and adverse-class language, and leaves labelled PK curves as pointers to the instruments under instruments/ rather than as instructions.

17Why later monographs keep measuring themselves against this file

Insulin is the template case for therapeutic peptides even when the market conversation has moved to incretins. It established (i) a sequenced protein hormone as a drug substance, (ii) recombinant manufacture at industrial scale, (iii) analogue engineering as pharmacokinetics by chemistry, and (iv) essential-medicine status as a public-health fact about a peptide. Reviews of biosynthetic human insulin and analogues are, among other things, histories of that template (Chance et al., 1993; Vajo et al., 2001; Rodbard et al., 2020; Francis et al., 2024).

When a later Radix monograph argues that a short peptide is or is not “like insulin,” the comparison only works if this file has kept regular insulin distinct from lispro, aspart and glargine, and has kept animal and crystallographic results distinct from human trials.

That is why the controlling thesis in the evidence dossier was not “insulin is old and important.” It was that insulin is the template case: sequence, structure, recombinant manufacture, analogue PK chemistry, and essential-medicine status, held together without collapsing siblings or smuggling use advice into research prose. Parts One through Four are that template written out. Part Five is only the reminder to use it.

18Open gaps

This build is honest about what it has not finished.

• The PubMed reading-set harvest kept 4,068 records against a MeSH-scale surface of hundreds of thousands; identity screening, not recall, chose the citations.
• PMC body-text matches exceed 90,000; the fetch target is the front-screened subset (7,287). Most of that surface is counted and not read.
• Landmark Sanger sequence primaries are not yet promoted into the local citation list from harvest titles; recombinant reviews stand in until they are.
• The Humulin R label locks 1982 as Initial U.S. Approval; a full FDA approval-package PDF for the “first rDNA drug” industrial narrative is still preferred.
• WHO evidence here is the essentialmeds.org portal HTML; a Model List PDF should replace it when retained.

Those gaps belong in the Apparatus. They are not licenses to invent identifiers.

19Standing constraint

Section 25References

Every entry below was resolved against the National Library of Medicine's records during this build and read back against its author, journal, year and title line. None was written from recall. The build refuses to run if any identifier fails to resolve, and a separate gate parses every author–year citation out of the prose above and asserts it against these records, so a citation that is correctly formatted and attached to the wrong paper is caught rather than printed.

Of the 30 records, 7 carry a PubMed Central identifier and 23 do not. This first draft resolved citations from NCBI metadata and retained regulatory instruments; stage 03 fulltext fetch was not yet run, so a PMCID here is not a claim that the OA full text was ingested into the local corpus. Abstract-only records are handled under Section 27.

1. Adam MG, Coller L, Hodgkin DC, Dodson GG. X-ray crystallographic studies on zinc insulin crystals. Am J Med. 1966;40(5):667-71.
   PMID 5941965 · doi:10.1016/0002-9343(66)90146-x
2. Best CH. Insulin and Diabetes-In Retrospect and in Prospect: The Banting Memorial Lecture, 1945. Can Med Assoc J. 1945;53(3):204-12.
   PMID 20323538 · PMC1582550
3. Birnbaum DT, Kilcomons MA, DeFelippis MR, Beals JM. Assembly and dissociation of human insulin and LysB28ProB29-insulin hexamers: a comparison study. Pharm Res. 1997;14(1):25-36.
   PMID 9034217 · doi:10.1023/a:1012095115151
4. Bolli GB, Home PD, Lepore M, Riddle MC, Porcellati F, Fanelli CG, et al.. The Silver Jubilee (2025) of Insulin Glargine: Introducing the Era of Long-Acting Insulin Analogues for Diabetes Mellitus. Diabetes Obes Metab. 2026;28(7):5527-5541.
   PMID 42046184 · doi:10.1111/dom.70751 · PMC13243952
5. Candido R, Wyne K, Romoli E. A Review of Basal-Bolus Therapy Using Insulin Glargine and Insulin Lispro in the Management of Diabetes Mellitus. Diabetes Ther. 2018;9(3):927-949.
   PMID 29654514 · doi:10.1007/s13300-018-0422-4 · PMC5984925
6. Chakraborty C, Mungantiwar AA. Human insulin genome sequence map, biochemical structure of insulin for recombinant DNA insulin. Mini Rev Med Chem. 2003;3(5):375-85.
   PMID 12769691 · doi:10.2174/1389557033488024
7. Chance RE, Kroeff EP, Hoffmann JA, Frank BH. Chemical, physical, and biologic properties of biosynthetic human insulin. Diabetes Care. 1981;4(2):147-54.
   PMID 7011716 · doi:10.2337/diacare.4.2.147
8. Chance RE, Frank BH. Research, development, production, and safety of biosynthetic human insulin. Diabetes Care. 1993;16 Suppl 3:133-42.
   PMID 8299470 · doi:10.2337/diacare.16.3.133
9. Chang X, Jorgensen AM, Bardrum P, Led JJ. Solution structures of the R6 human insulin hexamer,. Biochemistry. 1997;36(31):9409-22.
   PMID 9235985 · doi:10.1021/bi9631069
10. Ciszak E, Beals JM, Frank BH, Baker JC, Carter ND, Smith GD. Role of C-terminal B-chain residues in insulin assembly: the structure of hexameric LysB28ProB29-human insulin. Structure. 1995;3(6):615-22.
    PMID 8590022 · doi:10.1016/s0969-2126(01)00195-2
11. De Meyts P. [The insulin receptor discovery is 50 years old - A review of achieved progress]. Biol Aujourdhui. 2022;216(1-2):7-28.
    PMID 35876517 · doi:10.1051/jbio/2022007
12. Dodson E, Harding MM, Hodgkin DC, Rossmann MG. The crystal structure of insulin. 3. Evidence for a 2-fold axis in rhombohedral zinc insulin. J Mol Biol. 1966;16(1):227-41.
    PMID 5917732 · doi:10.1016/s0022-2836(66)80275-9
13. Francis D, Chacko AM, Anoop A, Nadimuthu S, Venugopal V. Evolution of biosynthetic human insulin and its analogues for diabetes management. Adv Protein Chem Struct Biol. 2024;142:191-256.
    PMID 39059986 · doi:10.1016/bs.apcsb.2024.06.004
14. GILL DG. Banting, insulin and diabetes. J Med Assoc State Ala. 1954;24(3):75-7.
    PMID 13192438
15. Harding MM, Hodgkin DC, Kennedy AF, O'Conor A, Weitzmann PD. The crystal structure of insulin. II. An investigation of rhombohedral zinc insulin crystals and a report of other crystalline forms. J Mol Biol. 1966;16(1):212-26.
    PMID 5917731 · doi:10.1016/s0022-2836(66)80274-7
16. Kaiserman K, Jung H, Benabbad I, Karges B, Polak M, Rosilio M. 20 Years of insulin lispro in pediatric type 1 diabetes: a review of available evidence. Pediatr Diabetes. 2017;18(2):81-94.
    PMID 27390032 · doi:10.1111/pedi.12401
17. Kumar L, Vizgaudis W, Klein-Seetharaman J. Structure-based survey of ligand binding in the human insulin receptor. Br J Pharmacol. 2022;179(14):3512-3528.
    PMID 34907529 · doi:10.1111/bph.15777
18. Luo J, Gellad WF. Origins of the Crisis in Insulin Affordability and Practical Advice for Clinicians on Using Human Insulin. Curr Diab Rep. 2020;20(1):2.
    PMID 31997036 · doi:10.1007/s11892-020-1286-3 · PMC9036473
19. Maikawa CL, Mann JL, Kannan A, Meis CM, Grosskopf AK, Ou BS, et al.. Engineering Insulin Cold Chain Resilience to Improve Global Access. Biomacromolecules. 2021;22(8):3386-3395.
    PMID 34213889 · doi:10.1021/acs.biomac.1c00474 · PMC8627795
20. Melberg SG, Johnson WC. Changes in secondary structure follow the dissociation of human insulin hexamers: a circular dichroism study. Proteins. 1990;8(3):280-6.
    PMID 2281088 · doi:10.1002/prot.340080309
21. Melo KFS, Bahia LR, Pasinato B, Porfirio GJM, Martimbianco AL, Riera R, et al.. Short-acting insulin analogues versus regular human insulin on postprandial glucose and hypoglycemia in type 1 diabetes mellitus: a systematic review and meta-analysis. Diabetol Metab Syndr. 2019;11:2.
    PMID 30622653 · doi:10.1186/s13098-018-0397-3 · PMC6317184
22. Ottesen JL, Nilsson P, Jami J, Weilguny D, Dührkop M, Bucchini D, et al.. The potential immunogenicity of human insulin and insulin analogues evaluated in a transgenic mouse model. Diabetologia. 1994;37(12):1178-85.
    PMID 7895946 · doi:10.1007/BF00399790
23. Palmieri LC, Fávero-Retto MP, Lourenço D, Lima LM. A T3R3 hexamer of the human insulin variant B28Asp. Biophys Chem. 2013;173-174:1-7.
    PMID 23428413 · doi:10.1016/j.bpc.2013.01.003
24. Rodbard HW, Rodbard D. Biosynthetic Human Insulin and Insulin Analogs. Am J Ther. 2020;27(1):e42-e51.
    PMID 31876563 · doi:10.1097/MJT.0000000000001089
25. Rys P, Pankiewicz O, Łach K, Kwaskowski A, Skrzekowska-Baran I, Malecki MT. Efficacy and safety comparison of rapid-acting insulin aspart and regular human insulin in the treatment of type 1 and type 2 diabetes mellitus: a systematic review. Diabetes Metab. 2011;37(3):190-200.
    PMID 21333580 · doi:10.1016/j.diabet.2010.12.003
26. Saboo B, Chandalia H, Ghosh S, Kesavadev J, Kochar IPS, Prasannakumar KM, et al.. Insulin Glargine in Type 1 Diabetes Mellitus: A Review of Clinical Trials and Real-world Evidence Across Two Decades. Curr Diabetes Rev. 2024;20(1):e100323214554.
    PMID 36896906 · doi:10.2174/1573399819666230310150905 · PMC10909813
27. Steensgaard DB, Schluckebier G, Strauss HM, Norrman M, Thomsen JK, Friderichsen AV, et al.. Ligand-controlled assembly of hexamers, dihexamers, and linear multihexamer structures by the engineered acylated insulin degludec. Biochemistry. 2013;52(2):295-309.
    PMID 23256685 · doi:10.1021/bi3008609
28. Vajo Z, Fawcett J, Duckworth WC. Recombinant DNA technology in the treatment of diabetes: insulin analogs. Endocr Rev. 2001;22(5):706-17.
    PMID 11588149 · doi:10.1210/edrv.22.5.0442
29. Wojciechowski P, Niemczyk-Szechowska P, Olewińska E, Jaros P, Mierzejewska B, Skarżyńska-Duk J, et al.. Clinical efficacy and safety of insulin aspart compared with regular human insulin in patients with type 1 and type 2 diabetes: a systematic review and meta-analysis. Pol Arch Med Wewn. 2015;125(3):141-51.
    PMID 25644227 · doi:10.20452/pamw.2705
30. Wright JR. Essential Contributions of Pathologists and Laboratory Physicians Leading to the Discovery of Insulin. Arch Pathol Lab Med. 2020;144(7):894-904.
    PMID 31825669 · doi:10.5858/arpa.2019-0400-HP

Sources with no PubMed record. Guideline, database and structural-archive material is listed separately so that the numbered list above stays wholly machine-verified.

1. U.S. National Library of Medicine (DailyMed). HUMULIN R (insulin human) injection, solution — SPL labeling. Initial U.S. Approval: 1982 and the identity of the drug substance as insulin human were read from the DailyMed SPL HTML retained under instruments/label_humulin_r_*. Label PK and warning language is descriptive regulatory text, not use advice..
   https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=b519bd83-038c-4ec5-a2
2. U.S. National Library of Medicine (DailyMed). HUMALOG (insulin lispro); NOVOLOG (insulin aspart); LANTUS (insulin glargine) — SPL labeling. Initial U.S. Approval years 1996 (lispro), 2000 (aspart), 2000 (glargine) and sequence/formulation identity statements were read from retained SPL HTML under instruments/label_humalog_*, label_novolog_*, label_lantus_*..
   https://dailymed.nlm.nih.gov/
3. World Health Organization. WHO Model List of Essential Medicines — Insulins section (portal extract). Human short- and intermediate-acting insulin and rapid- and long-acting insulin analogues appear under Insulins on the essentialmeds.org portal HTML retained as instruments/who_eml_24th_html.html. Prefer the Model List PDF if obtained later; this build cites the retained portal instrument..
   https://list.essentialmeds.org/
4. UniProt Consortium. P01308 (INS_HUMAN) — insulin preproprotein. Gene symbol INS and accession P01308 are taken from UniProtKB. Residue counts and disulfide connectivity in the prose are aligned to this record and to the crystallographic literature cited from the harvest..
   https://www.uniprot.org/uniprotkb/P01308/entry

Section 26How this document was assembled

The corpus for this first Insulin build was assembled against project 05 identity rules and against PubMed and PubMed Central. The interesting arithmetic is the volume problem: bare insulin is one of the largest MeSH-scale surfaces in biomedicine, so the harvest used a reading-set union rather than the whole Title/Abstract surface, and PMC was front-screened before any fetch budget was spent.

Identity

“Insulin” is an ordinary English scientific word. Admission therefore requires peptide/hormone/diabetes/receptor context, not the bare token in unrelated biochemistry. Lispro, aspart, glargine, detemir and degludec are siblings with margin: analogue-only papers do not silently become findings about regular / human insulin. IGF-1 and IGF-2 are context for mechanism, not rivals that refuse the subject. The matcher configuration lives in pipeline/compound.py (HOUSE_STYLE A9).

Local store

This first draft did not complete the local fulltext sweep or PMC fetch stages. Local files opened: 0. Local admissions: 0. Peer-reviewed local full texts: 0. Commercial/derived: 0. Those zeros are real absences, not hidden vendor pages.

External harvest

The reading-set PubMed query returned 4,633 records, of which 4,068 passed record-level relevance. The wider Title/Abstract surface counted beside that harvest is 459,685 and was not treated as the reading corpus.

PubMed Central returned 91,688 body-text matches. The front screen reduced that surface to 7,289 title/abstract-indexed targets; 84,900 documents were counted as surface and never read. Stage 03 fetch was not run for this first draft, so the reading corpus of scientific full texts is reported as 0 (honest zero until fetch completes). Citations in Sections 01–19 were resolved from verified PubMed metadata plus retained DailyMed and WHO instruments.

Route arithmetic pending a full 02d merge: PubMed-with-PMCID 1,159; PMC screened route 7,287; naive sum 8,446; keyed union recorded as 8,446; double-counts avoided 0 (zero until intersection is measured).

Retractions

Integrity screening checked 4,068 PMIDs tied to the harvest. In the reading corpus, 3 are retracted and 2 carry an expression of concern. Across the wider integrity query, 687 retracted and 2 concern-flagged records were seen. None of the 30 cited records is retracted or flagged, and no claim rests on a withdrawn paper.

The gap this document cannot close

23 of the 30 cited records have no PubMed Central identifier and were read at abstract/metadata level with NCBI verification. Landmark Sanger sequence primaries are not yet in the citation list from harvest titles. The Humulin R DailyMed label locks Initial U.S. Approval 1982; a full FDA approval-package PDF is still preferred for industrial history beyond that year. WHO evidence is the retained essentialmeds.org portal HTML. Section 18 states these absences as findings.

Section 27Evidence handling

Study type is named in the sentence that reports the finding. Crystallography is crystallography; a meta-analysis of randomised trials is a meta-analysis; a DailyMed SPL is a regulatory instrument. Animal immunology is not a human outcome trial.

Regular insulin and analogues are siblings with margin. A sentence that reports lispro, aspart or glargine evidence names that analogue. Head-to-head trials are kept; unmarked “insulin” never inherits analogue results (Parts Three and Four).

Labels and essential-medicine lists are not use advice. Initial U.S. Approval years and WHO Insulins listings are identity and classification facts. No dose, route or schedule in this document is offered for use by any person (Section 19).

Conflict is presented as conflict. Mitogenicity debates and retracted items are handled as contested or withdrawn literature, not as slogans (Section 15). Integrity-flagged PMIDs are excluded from supportive citation.

Absence is reported as a finding. Section 18 lists the open gaps for this build: unfinished fulltext fetch, Sanger primaries not yet promoted, FDA package PDF still preferred, WHO Model List PDF still preferred. Each is an absence in what was actually held, not a rhetorical gesture.

First-draft corpus honesty. The masthead corpus line for this build reports zero scientific full texts ingested by stage 03, because that stage was not run. The 30 PubMed references are still NCBI-resolved, and the DailyMed / WHO / UniProt instruments are retained on disk. A later rebuild that completes PMC fetch must raise the corpus figure by keyed union, not by restating the PubMed harvest count as if it were full text.

Misuse literature stays behind the fence. Section 16 states the boundary without instructional content. Performance and bodybuilding adjacency may be described as epidemiology or toxicity case material when peer-reviewed sources exist; this build did not require such sources for the controlling thesis and therefore does not invent them.

What a later Insulin rebuild should add. Promote Sanger sequence primaries once harvest titles or lawful archives supply them; retain an FDA approval-package PDF for the industrial “first rDNA medicine” narrative beyond the 1982 label year; replace the WHO portal HTML with the Model List PDF; run 02d merge with measured intersection; fetch and screen a budgeted slice of the 7,287 PMC front-screen targets. Until then, Section 18 remains the honest inventory of absences.

Sibling margin is an evidence rule, not a branding preference. Insulin lispro, insulin aspart and insulin glargine each have their own approval years on retained DailyMed SPLs (1996, 2000, 2000). A sentence that reports a lispro pediatric review (Kaiserman and colleagues) or an aspart meta-comparison (Rys; Wojciechowski; Melo) or a glargine jubilee review (Bolli; Saboo) must keep that name in the sentence. The Apparatus restates the rule because Part Three is where readers most often try to collapse the family.

Hypoglycemia is reported as the dominant adverse class, not as a titration guide. Comparative reviews and labels agree that hypoglycemia is the mechanism-linked harm class for insulin therapy. This document reports that agreement as a finding about the literature and the instruments. It does not translate those findings into dose, route or schedule language for any person.

Closing note on series use. Later Radix monographs that say a compound is or is not “like insulin” should mean the template facts in Parts One through Four of this file: two-chain disulfide protein, recombinant manufacture, analogue engineering as pharmacokinetics by chemistry, and essential-medicine status — with study type labelled and siblings named. Anything else is slogan, and slogan is not what this series is for.