M Joan Taylor, Krishan P Chauhan, Tarsem S Sahota
School of Pharmacy, De Montfort University, Leicester, UK
Address for correspondence: Joan Taylor
Professor of Pharmaceutics, School of Pharmacy, De Montfort University, Leicester LE1 9BH, UK
E-mail: mjt@dmu.ac.uk
https://doi.org/10.15277/bjd.2019.228
People with type 1 diabetes must use insulin and a large fraction of those with type 2 condition also do so. Many therefore struggle with the unpredictable balancing of insulin dose with calorie intake and utility. A healthy pancreas makes meticulous adjustment on a continuous basis that present therapeutic insulin administration cannot match. However, much progress has been made to make it simpler to inject both background and fast-acting boost insulins with a view to better mimicking normal pancreatic output. The present fast insulins are reviewed with accent on the primary amino acid structures of the biosynthetic types that diffuse more quickly than regular insulin that associates in hexamers. This makes boost doses kinetically and clinically more effective, allowing people to inject better estimated boost and corrective doses. Formulation advances are discussed for their present and potential contributions. The newer slow-acting insulins are also described and compared, their advantage also being kinetic with a lower likelihood of inducing overnight hypoglycaemia when used optimally. Finally, the appreciation of the advantages of alternative routes of administration such as oral and peritoneal are included in this review because of the possibility of altering the hepatic to peripheral ratio, the reasons for which are more effective but less obesogenic insulin activity. The logistics of oral insulin are summarised in terms of the risks to the insulin structure, the facilitation of paracellular uptake at the apical surface and the paradoxically advantageous hepatic first pass. Other non-invasive routes are also included in the review.
Br J Diabetes 2019;19:124-130
Key words: new insulins, insulin amino acid modifications, conjugations, formulatory changes, unfolding and aggregation protection, closed loop, delivery routes, hepatic to peripheral ratio
Many patients with type 2 diabetes (T2D), all patients with type 1 diabetes (T1D) and eventually almost all those with latent autoimmune diabetes of adulthood (LADA) need insulin, the latter − as with conventional T1D − to prevent ketoacidosis. For all insulin users, the aim should be to match insulin delivery to the pattern found physiologically in healthy counterparts, in order to keep the blood glucose within the target range for most of the time. For T1D, success is often only partial and so HbA1c values are commonly higher than target. This is true even for children for whom preventive measures for complications are so important.1–3 The reason is that the risk of hypoglycaemia is commonly perceived to outweigh the risk of long-term complications in the day-to-day management of many users, perhaps particularly for children. This may relate to concomitant glucagon and/or glycogenolysis failures or glucagon resistance.4 It is the case, therefore, that a significant fraction of the diabetic patient population needs improvements in the methods used to deliver insulin in order to improve safety and efficacy. Here, we chart progress in the design of approaches from the conventional to the innovative and futuristic. We will review various strategies including formulations to deliver, stabilise and protect the insulin molecule, improve comfort, convenience and compliance with the ultimate aim of optimising glycaemic control.5
Open loop and tight control
Since the demonstration that tight control of blood glucose (BG) is beneficial for the prevention of diabetes complications, multi-dose injection (MDI) with basal and bolus (prandial and corrective) insulin accompanied by frequent BG testing has become the standard for T1D and not uncommon for T2D. Structured education programmes such as DAFNE and DESMOND have enabled people to take advantage of these approaches. The tools for the job include the increasingly improved BG meters, flash and continuous glucose monitoring systems (CGMS), smart phones for capturing, logging and manipulating data, internet guidance, improved injection devices and new insulins.
Open loop treatment of diabetes is the term used when the judgement of quantity and timing of insulin is left to the users, carers and/or prescribers.6 This may be as unsophisticated as to impose a relatively constant daily macronutrient regimen (content and timing) while prescribing a fixed, matched quantity of insulin. The adjustment is usually made by healthcare providers on behalf of the individual in an attempt to keep BG within reasonable limits. This is no longer regarded as a suitable approach for the majority of people with T1D, for whom a more intensive MDI or insulin pump regimen is recommended which is sufficiently flexible to achieve tight glucose control despite variable amounts and timing of food, physical activity etc, thus placing the decision with the user rather than the prescriber. Intensive control therefore requires the user to measure the BG fairly frequently and to calculate and administer bolus doses before food and as corrections. This method clearly relies on education and accurate BG information, often entering the latter on bolus advisor meters or phone apps that are available to help cope with the complexity and continual nature of the demands. The advent of flash and CGMS, that measure tissue fluid glucose rather than BG, gives information that is collected many times per hour, albeit with a potentially significant time lag in comparison with BG. Achieving tight control in order to reduce HbA1c to the T1D optimum 48 mmol/mol (6.5%) is challenging because, not only is the BG a moving target, but a distorted sympathetic neuronal control of the glucagon-based counter-regulatory system may complicate the effects of imposition of harsh antihyperglycaemic doses/ pump dose rates. Resulting hypos are common,7 estimated at CGMS-measured 2.1 events per 24 hours that often go unnoticed even in hypo-aware people. This mitigates against safe glucose control despite the likely lower HbA1c results and is especially true in people with hypoglycaemia unawareness. This has led to the concept of glucose targets being set for ‘time in range’ rather than HbA1c.
Development of injectable insulins for intensive control regimens
A variety of innovative fast- and slow-acting insulins are now available, including the ultra-long-acting degludec and fast-acting insulins such as glulisine and fast aspart. Their purpose is to make tight control achievable by providing the tools for imposing sustained background control and tailored fast-acting doses around mealtimes to prevent post-meal hyperglycaemia without delayed hypoglycaemia. Over the last decades, specific insulins have been marketed that claim to have advantages when used in MDI regimens.
The structure, design strategies and formulation are comprehensively reviewed by various authors4,8,9 but briefly, have mainly involved the biosynthetic alteration of the amino acid content, sequence and/or conjugation, resulting in changes in charge, hydrogen bonding and hydrophobicity. These modifications create the required pharmacokinetics either by solubility, competitive binding or changes in quaternary association, although modified insulins are generally inherently less resistant to physical destabilisation than the native molecule that normally groups in dimers and hexamers in concentrated solution (Figures 1 and 2).10
Formulatory changes in the solvent medium for insulins are additional common approaches,11,12 with polyethylene glycol (PEG), glycerol and ethanol having been tried for their effects on electrostatic interactions in the highly polar insulin molecule. Chelating agents, antioxidant sugars, surfactants and amino acids such as lysine, arginine and glycine have also featured in developmental work, often to create the monomeric form. Linjeta (Viaject) was formulated with EDTA and citric acid to chelate zinc, for example, but has not been pursued. Two agents have been incorporated with aspart (Novorapid®, Novo Nordisk), a monomeric insulin that is given as a preprandial bolus and is commonly used in insulin pumps13,14 because it may be cleared from the subcutaneous tissue quickly enough to transmit pump rate changes to appropriate modification of BG adjustment. Loss from skin to plasma is a more blunted process with soluble insulins that persist in hexamers, because the diffusion coefficient is much lower. Fiasp® (Novo Nordisk), like the normal version of aspart, contains glycerol, metacresol, phenol, zinc and pH adjustment by phosphate, hydroxide and hydrochloric acid. However, it achieves an even more prompt onset of action and physiological profile by the addition of niacinamide (nicotinamide, niacin, vitamin B3) as an absorption enhancer that works not only as a localised vasodilator, but by increasing the fast diffusing monomer fraction by about 35%.15 L-arginine is an additional agent working as a refolding protector and thus a stabiliser against aggregation. In this context, aggregation means an unwanted grouping in unspecified numbers of large molecules, usually peptides and proteins. It is different either from association into quaternary structuring such as hexamers or amorphous precipitation such as occurs with excess zinc or at the isoelectric zwitterionic pH point (pI), which can each preserve activity, as does crystallisation, even if some reversible unfolding happens. In aggregation, however, the normal tertiary and quaternary structures are lost because it involves irreversible unfolding. Correct folding is accomplished in the beta cell’s endoplasmic reticulum and is normally vital for activity whether the insulin is endogenous or biotechnically synthesised. Unfolding is therefore a serious degradative change that may facilitate further permanent transformation, often involving amyloid fibrous structures that are inactive, potentially antigenic and can accumulate at injection sites. An aggregation protector is therefore an important formulation success. In the past, soluble Hoe21PH insulin for pump use was stabilised against aggregation by the poloxamer micellisation agent Genapol, in line with surfactant strategies for proteins stabilised in general.16,17 In the case of Fiasp®, the combined changes to the previous aspart formulation halve the time of appearance in plasma with 74% greater insulin action within the first 30 min, the clear aim being better postprandial control.14,18
By contrast, detemir and degludec were synthesised as soluble, long-acting products both involving the covalent addition of polymeric lipophilic side chains.9,19 These alter the kinetics to produce low level basal dosage with a flattened plasma profile; this was an improvement on what could be achieved with suspension products like isophane (NPH) insulins and with insulin zinc suspension (IZS). The clinical value with these newer products in MDI regimens was the reduction in overnight hypoglycaemic events. Detemir has a myristic (C14) acid substitution at a terminal B29 lysine (no threonine), and its slow action is attributable to hydrophobic self-association in tissues and also to binding of the acyl chain to fatty acid binding sites on serum albumin.6,20 A further flattened profile is associated with degludec (t1/2 ~25 hours, duration >40 hours) and is enabled because, despite its superficial similarity to detemir, its des-30-structure (again no threonine) forms very long unique hexameric sequences stabilised by zinc, phenol and hydrophobic contact between covalently attached hexadecanoic (C16) diacid chains linked at B29 with glutamic acid.9,19 These lipophilic agents therefore differ from an alternative soluble long-acting insulin, glargine, which has a replacement of glycine for asparagine at position A21 and addition of two arginine molecules at positions B29–30. This alters the pI such that precipitation occurs at pH 7.4 instead of 5.4 as with native insulin (Figure 3), thus creating a subcutaneous depot.
A variety of other new kinetic approaches included a liver- specific analogue that was an interesting departure. Peglispro (LY2605541) was a developmental effort to slow the release but also to raise the ratio of hepatic to peripheral insulin, changing the ratio of hepatic glycogenolysis to peripheral glucose disposal and thus to suppress the weight gain associated with insulin medication. Other strategies such as oral insulin have a similar aim. The 5.8 kDa insulin lispro monomer had a 20 kDa linear PEG moiety covalently bound to the B28 lysine (not B29 as in native insulin), and promoting the binding of three water molecules thus giving a hydrodynamic radius equivalent to a globular molecule of about 78 kDa. This slowed the renal excretion as well as the absorption from the subcutaneous administration site. Peglispro binding to the insulin receptor is much reduced and thus the effective molar dose is greater. Madsbad21 reviewed the evidence for the hepatic specificity and Hirose22 pointed out the additional protective effect of the pegylation towards proteolytic enzymes. By 2015, however, this compound had failed its clinical assessment because of unacceptable lipid values and liver function tests.23
Closed loop: the automation of tight control
The closed loop system is the long-term aim of diabetes control using exogenous insulin because it takes decision-making by the user out of the system24,25 and, in the optimum embodiment, operates by several components including a frequent coupling with real-time sensing to correct BG in small steps, fast response for each corrective move and prompt catabolic mechanisms so that excess insulin is removed (Figure 4).
These points all emulate the physiological process and, to complete the best possible outcome, the insulin output should ideally be oscillatory with a frequency of minutes (ie, within the BG adjustment cycles).26 Oscillatory release is thought to minimise the downregulation of tissue receptors and oppose resistance developing which is a risk even in T1D.27 Closed loop systems working to these principles should avoid the danger of hypos yet be able to respond in real time to postprandial and counter-regulatory peaks, keeping BG within normal tolerances 100% of the time.
Currently there are three approaches to closed loop systems: biological (pancreas or islet transplants), electronic (artificial pancreas) and chemical (including smart insulin formulations), of which only the first two are in human development and use, and have been reviewed elsewhere.
The remainder of this article is the formulation of insulins that interact with tissue/BG such as to regulate insulin output.
Injectable and implantable
All people with T1D and about 25% of those with T2D are prescribed recipients of formulations intended mainly for subcutaneous injection. While this route is not the fastest route for absorption, it has the advantage of being an easily-learned technique for self-administration. Its exploitation therapeutically has been the norm, especially as it was safe for the longer acting suspension formulations that have been used for many years. However, some people with diabetes are needle phobic, develop injection site lipohypertrophy or localised allergy or bruise because of concomitant anticoagulant therapy. As a general consequence, efforts have been made to exploit alternative parenteral delivery paradigms such as microneedles, pumps, patch pumps and patches, as well as other routes of administration that are less invasive. Other considerations are important, so that the peritoneal route circumvents the delays due to the dense, fatty subcutaneous skin layer and can be built into viscerally administered insulin from devices, possibly with a similarly placed sensor.25,28–32 This may solve the potential iatrogenic problem of the non-physiological effects of distorting the peripheral and hepatic concentration ratio of released insulin, as discussed also below. In a different context, Rhea et al33 report access to the CNS mainly by intrathecal, cerebral, ventricular but also nasal and ocular routes, for idiopathic CNS insulin resistance that is inadequately treated by peripheral administration.
Oral
Alternatives to parenteral delivery feature high in diabetes pharmaceutical research.34 The oral route has been a long-term goal for insulin delivery, not only for convenience and compliance but because the mesenteric to portal drainage implies a potentially normal ratio of hepatic to peripheral insulin, fostering normal liver regulation of hyperglycaemia and differing markedly from parenteral delivery.35,36 However, the harsh environment of proteolytic digestive enzymes, food interaction and the obstacles to the absorption of large molecules has meant that development has been slow, waiting for tactics to overcome the problems. Protective strategies include chitosan, an aminopolysaccharide gel derived from crustacean shell.37,38
Access to gut vasculature via the apical aspect of the gut epithelium is an equal challenge. The transport of large polar molecules depends largely on the paracellular route, implying the traversing of a complex mixed environment gap between cells that is also size limited. Information about the paracellular gap size is variable. Anderson comments that the biology and pharmaceutics approaches differ, and typically drug delivery studies use a series of labelled tracers that may or may not be charged, while transcytosis may be ignored in the assessment.39 Chitosan and its quaternised derivatives such as trimethylchitosan have the ability to influence permeation by accessing claudin and actin structures.37,38,40–43 However, the balance of alkyl chain length and charge on the water solubility permeation effectiveness and tight junction molecular recovery is critical, as discussed by Benediktsdóttir et al.44 Maher et al categorise both paracellular and transcellular mechanisms, highlighting sucrose laurate, a surfactant-like material, as well as the phosphatase inhibitor PIP 640 and some ionic liquids such as choline geronate (CAGE).45 These may influence tight junctions to increase gut uptake of insulin, as may the formulation as nanoparticles.45
Biologically, it seems that two paracellular routes may co-exist, namely a high capacity one with steep size dependence, a cut-off of 4 Å and a larger gap that may occur in intact tight junctions under the influence of inflammatory cytokines. A gap size of 2–5 Å is thus often quoted but, depending on tissue, where normal transport can admit molecules of 200 Da such as mannitol (3.5 Å) but not usually the fructal inulin (molecular weight variable but possibly 3–5 kDa and about 11 Å, thus approaching insulin monomer size). However, Taverner holds that ileal tight junctions can exist in open and shut form, as a function of the phosphorylation of light chain myosin and its regulatory kinase system that maintains a circumferential tension. Several efforts have been made to set the default to open the junctions for insulin transport through the apical surface using kinase system blockers that have been shown to permit transport of labelled dextran of 70 kDa.46 In another approach, chitosans have been used not only for their protective effect but because they can open tight junctions in a transient and reversible manner by a disruption of claudin-4, a tight junction protein.43
Other non-invasive routes
The use of other routes involving transport across mucous membranes is reviewed by Easa et al.35 Briefly, inhaled insulin has undergone a revival as Afrezza since the abortive attempt with the unwieldy Exubera, because the theory is well-developed, as reviewed by Lin et al,47 and the practicalities have been improved. Transdermal delivery can also include microneedles and, for completeness, it should be mentioned here that iontophoresis systems have been widely studied and sometimes combined with microneedles. Buccal and nasal routes are conceivable, so that micelle-associated insulin in Generex’s Ora-lyn has been reported as beginning trials in 2019, but Aquestive’s bioadhesive film associated with glycan and insulin aspart-bearing gold nanoparticles (1–2 nm) has failed because, although the particles were renally eliminated and despite apparent absorption through buccal tight junctions, the product demonstrated impractically low bioavailability.35 Nevertheless, buccal formulations – as well as providing access paracellularly – could be developed to optimise as least basal dose.48–50 Rectal and vaginal routes seem impractical and neither finds many recent citations except in reviews51,52 or as a rectal instillation (for anti-inflammatory treatment of colitis rather than for its antihyperglycaemic properties).53 The ophthalmic route may be effective, as shown in accidental systemic toxicity after an insulin eye mishap,54 but predictability is an issue due to nasolacrimal loss and other dose impracticalities. The reality is also that administration of insulin by these alternative routes – or, indeed, by parenteral routes other than peritoneal – is unable to emulate the physiological hepatic to peripheral insulin ratio, except for the potential of oral delivery (see above). Ironically, in drug delivery generally, many alternative routes are developed to avoid hepatic first pass effects. However, for insulin, first pass would enable the advantageous liver effects that are suppressed with non-oral delivery. The major effect of hepatic targeting is to curb glycogenolysis and also gluconeogenesis compared with normal, and so glucose peaks are harder to control unless the liver is targeted. Hepatocentric activity would be an asset because a further factor is that glucagon is also often poorly regulated in diabetes so that hyperglycaemic peaks postprandially are potentially exaggerated for that reason also. Finally, the peripheral effects of excess insulin on adipose tissue are obesogenic, which affects treatment compliance.52
The quest for much more physiological insulin delivery has been ongoing for many decades, ever since the realisation that crudely dissolved insulin was life-saving but not as a long-term sustainable treatment. It is true that great strides have been made such that diabetic people can now be in possession of the information about their glucose status on a minute-by-minute basis. They have the tools to plan background dosage with bolus doses to cover meals and aim for low HbA1c values yet avoid repeated hypoglycaemic crises. Eating and exercising variably are possible, thanks to education programmes to take advantage of these strategies. The advent of smart phone apps has made this easier, yet still BG and HbA1c levels are too high. The problem is that the target moves, but the delivery of insulin cannot keep pace with the variable need. On the whole, people avoid hypoglycaemia at the expense of hyperglycaemia. A closed system is needed. The perception of a slow pace of commercial development has meant that some users themselves have become skilled enough to be able to combine pump delivery and sensor to create fairly adventurous and sophisticated artificial pancreases, as described on the Nightscout webpage.55 The physiological delays that limit the kinetics of skin sensors and delivery systems remain a barrier to development of this type, however, whatever the source. Since algorithm development cannot detect what has not yet happened as a meal is consumed, subcutaneous systems must undergo a change or replacement in the quest. Other systems such as oral, transdermal and pulmonary may ultimately prove to be superior to subcutaneous basal insulins.
This is an accompanying article of a talk given at the Royal Society of Medicine in London on 13 June 2019 on “New technologies in diabetes”.
Conflict of interest: None
Funding: Salaried by De Montfort. Work referred to (Taylor et al) funded by NIHR and Edith Murphy Foundation between 2009 and 2017.
1. Campbell MS, Schatz DA, Chen V, and T1D Exchange Clinic Network. A contrast between children and adolescents with excellent and poor control: the T1D Exchange Clinic Registry experience. Pediatr Diabetes 2014;15(2):110–17. https://doi.org/10.1111/pedi.12067
2. Christie D, Thompson R, Sawtell, M, et al. Effectiveness of a structured educational intervention using psychological delivery methods in children and adolescents with poorly controlled type 1 diabetes: a cluster- randomized controlled trial of the CASCADE intervention. BMJ Open Diabetes Res Care 2016;4(1):e000165. https://doi.org/10.1136/bmjdrc-2015-000165
3. Pinhas-Hamiel O, Hamiel U, Boyko V, Graph-Barel C, Reichman B, Lerner-Geva L. Trajectories of HbA1c levels in children and youth with type 1 diabetes. PloS One 2014;9(10):e109109. https://doi.org/10.1371/journal.pone.0109109
4. McCall AL, Farhy LS. Treating type 1 diabetes: from strategies for insulin delivery to dual hormonal control. Minerva Endocrinol 2013;38(2):145–63.
5. Bailey CJ. Glucose-lowering therapies in type 2 diabetes: opportunities and challenges for peptides. Peptides 2018;100:9–17. https://doi.org/10.1016/j.peptides.2017.11.012
6. Jayakrishnapillai P, Nair SV, Kamalasanan K. Current trend in drug delivery considerations for subcutaneous insulin depots to treat diabetes. Colloids and Surfaces B: Biointerfaces 2017;53:123–31. https://doi.org/10.1016/j.colsurfb.2017.02.017
7. Reno CM, Litvin M, Clark AL, Fisher SJ. Defective counterregulation and hypoglycemia unawareness in diabetes: mechanisms and emerging treatments. Endocrinol Metab Clin North Am 2013;42(1):15–38. https://doi.org/10.1016/j.ecl.2012.11.005
8. Weiss MA. Design of ultra-stable insulin analogues for the developing world. J Health Spec 2013;1(2):59–70. https://doi.org/10.4103/1658-600X.114683
9. Woo VC. New insulins and new aspects in insulin delivery. Can J Diabetes 2015;39(4):335–43. https://doi.org/10.1016/j.jcjd.2015.04.006
10. Weiss MA, Lawrence MC. A thing of beauty: structure and function of insulin's "aromatic triplet". Diabetes Obes Metab 2018;20(Suppl 2):51–63. https://doi.org/10.1111/dom.13402
11. Gast K, Schüler A, Wolff M, et al. Rapid-acting and human insulins: hexamer dissociation kinetics upon dilution of the pharmaceutical formulation. Pharmaceut Res 2017;34(11):2270–86. https://doi.org/10.1007/s11095-017-2233-0
12. Mitchell DE, Fayter AER, Deller RC, Hasan M, Gutierrez-Marcos J, Gibson MI. Ice-recrystallization inhibiting polymers protect proteins against freeze-stress and enable glycerol-free cryostorage. Materials Horizons 2019;6(2):364–8. https://doi.org/10.1039/C8MH00727F
13. Mathieu C, Bode BW, Franek E, et al. Efficacy and safety of fast-acting insulin aspart in comparison with insulin aspart in type 1 diabetes (onset 1): a 52-week, randomized, treat-to-target, phase III trial. Diabetes Obes Metab 2018;20(5):1148–55. https://doi.org/10.1111/dom.13205
14. Heise T, Zijlstra E, Nosek L, Rikte T, Haahr H. Pharmacological properties of faster-acting insulin aspart vs insulin aspart in patients with type 1 diabetes receiving continuous subcutaneous insulin infusion: a randomized, double-blind, crossover trial. Diabetes Obes Metab 2017;19(2): 208–15. https://doi.org/10.1111/dom.12803
15. Kildegaard J, Buckley ST, Nielsen RH, et al. Elucidating the mechanism of absorption of fast-acting insulin aspart: the role of niacinamide. Pharm Res 2019;36(3):49. https://doi.org/10.1007/s11095-019-2578-7
16. Walter HM, Timmler R, Mehnert H. Stabilized human insulin prevents catheter occlusion during continuous subcutaneous insulin infusion. Diabetes Res (Edinburgh, Scotland) 1990;13(2):75–7.
17. Lee HJ, McAuley A, Schilke KF, McGuire J. Molecular origins of surfactant-mediated stabilization of protein drugs. Adv Drug Deliv Rev 2011; 63(13):1160–71. https://doi.org/10.1016/j.addr.2011.06.015
18. Leelarathna L, Ashley D, Fidler C, Parekh W. The value of fast-acting insulin aspart compared with insulin aspart for patients with diabetes mellitus treated with bolus insulin from a UK health care system perspective. Ther Adv Endocrinol Metab 2018;9(7):187–97. https://doi.org/10.1177/2042018818766816
19. Tambascia MA, Eliaschewitz FG. Degludec: the new ultra-long insulin analogue. Diabetol Metab Syndr 2015;7:57. https://doi.org/10.1186/s13098-015-0037-0
20. Ma Z, Christiansen JS, Laursen T, Lauritzen T, Frystyk J. Short-term effects of NPH insulin, insulin detemir, and insulin glargine on the GH-IGF1-IGFBP axis in patients with type 1 diabetes. Eur J Endocrinol 2014; 171(4):471–9. https://doi.org/10.1530/EJE-14-0258
21. Madsbad S. LY2605541: a preferential hepato-specific insulin analogue. Diabetes 2014;63(2):390–2. https://doi.org/10.2337/db13-1646
22. Hirose T. Development of new basal insulin peglispro (LY2605541) ends in a disappointing result. Diabetol Int 2016;7(1):16–17. https://doi.org/10.1007/s13340-016-0255-1
23. Munoz-Garach A, Molina-Vega M, Tinahones FJ. How can a good idea fail? Basal insulin Peglispro [LY2605541] for the treatment of type 2 diabetes. Diabetes Ther 2017;8(1):9–22. https://doi.org/10.1007/s13300-016-0214-7
24. Battelino T, Omladič JŠ, Phillip M. Closed loop insulin delivery in diabetes. Best Pract Res Clin Endocrinol Metab 2015;29(3):315–25. https://doi.org/10.1016/j.beem.2015.03.001
25. Uduku C, Oliver N. Pharmacological aspects of closed loop insulin delivery for type 1 diabetes. Curr Opin Pharmacol 2017;36:29–33. https://doi.org/10.1016/j.coph.2017.07.006
26. Satin LS, Butler PC, Ha J, Sherman AS. Pulsatile insulin secretion, impaired glucose tolerance and type 2 diabetes. Mol Aspects Med 2015; 42:61–77. https://doi.org/10.1016/j.mam.2015.01.003
27. Priya G, Kalra S. A review of insulin resistance in type 1 diabetes: is there a place for adjunctive metformin? Diabetes Ther 2018;9(1):349–61. https://doi.org/10.1007/s13300-017-0333-9
28. Taylor MJ, Gregory R, Tomlins P, Jacob D, Hubble J, Sahota TS. Closed-loop glycaemic control using an implantable artificial pancreas in diabetic domestic pig (Sus scrofa domesticus). Int J Pharm 2016;500(1–2):371–8. https://doi.org/10.1016/j.ijpharm.2015.12.024
29. Dassau E, Renard E, Place J, et al. Intraperitoneal insulin delivery provides superior glycaemic regulation to subcutaneous insulin delivery in model predictive control-based fully-automated artificial pancreas in patients with type 1 diabetes: a pilot study. Diabetes Obes Metab 2017; 19(12):1698–705. https://doi.org/10.1111/dom.12999
30. Huyett LM, Dassau E, Zisser HC, Doyle FJ 3rd. Design and evaluation of a robust PID controller for a fully implantable artificial pancreas. Ind Eng Chem Res 2015;54(42):10311–21. https://doi.org/10.1021/acs.iecr.5b01237
31. Renard E. New modes of insulin delivery and new modes of monitoring of type 1 diabetes mellitus. Rev Prat 2018;68(6):620–7.
32. Zisser H. Clinical hurdles and possible solutions in the implementation of closed-loop control in type 1 diabetes mellitus. J Diabetes Sci Technol 2011;5(5):1283–6. https://doi.org/10.1177/193229681100500537
33. Rhea EM, Salameh TS, Banks WA. Routes for the delivery of insulin to the central nervous system: a comparative review. Exp Neurol 2019; 313:10–15. https://doi.org/10.1016/j.expneurol.2018.11.007
34. Frid AH, Kreugel G, Grassi G, et al. New insulin delivery recommendations. Mayo Clinic Proc 2016;91(9):1231–55. https://doi.org/10.1016/j.mayocp.2016.06.010
35. Easa N, Alany RG, Carew M, Vangala A. A review of non-invasive insulin delivery systems for diabetes therapy in clinical trials over the past decade. Drug Discovery Today 2019;24(2):440–51. https://doi.org/10.1016/j.drudis.2018.11.010
36. Chaturvedi K, Ganguly K, Nadagouda MN, Aminabhavi TM. Polymeric hydrogels for oral insulin delivery. J Control Release 2013;165(2):129–38. https://doi.org/10.1016/j.jconrel.2012.11.005
37. Grigoras AG. Polymer-lipid hybrid systems used as carriers for insulin delivery. Nanomed Nanotechnol Biol Med 2017;13(8):2425–37. https://doi.org/10.1016/j.nano.2017.08.005
38. Fonte P, Araújo F, Silva C, et al. Polymer-based nanoparticles for oral insulin delivery: revisited approaches. Biotechnol Adv 2015;33(6, Part 3):1342–54. https://doi.org/10.1016/j.biotechadv.2015.02.010
39. Anderson JM, Van Itallie CM. Physiology and function of the tight junction. Cold Spring Harb Perspect Biol 2009;1(2):a002584. https://doi.org/10.1101/cshperspect.a002584
40. Mukhopadhyay P, Mishra R, Rana D, Kundu PP. Strategies for effective oral insulin delivery with modified chitosan nanoparticles: a review. Prog Polymer Sci 2012;37(11):1457–75. https://doi.org/10.1016/j.progpolymsci.2012.04.004
41. Karavasili C, Fatouros DG. Smart materials: in situ gel-forming systems for nasal delivery. Drug Discovery Today 2016;21(1):157–66. https://doi.org/10.1016/j.drudis.2015.10.016
42. Tscheik C, Blasig IE, Winkler L. Trends in drug delivery through tissue barriers containing tight junctions. Tissue Barriers 2013;1(2):e24565. https://doi.org/10.4161/tisb.24565
43. Yeh T, Hsu L, Tseng MT, et al. Mechanism and consequence of chitosan-mediated reversible epithelial tight junction opening. Biomaterials 2011;32(26):6164–73. https://doi.org/10.1016/j.biomaterials.2011.03.056
44. Benediktsdóttir BE, Gudjónsson T, Baldursson Ó, Másson M. N-alkylation of highly quaternized chitosan derivatives affects the paracellular permeation enhancement in bronchial epithelia in vitro. Eur J Pharm Biopharm 2014;86(1):55–63. https://doi.org/10.1016/j.ejpb.2013.04.002
45. Maher S, Brayden DJ, Casettari L, Illum L. Application of permeation enhancers in oral delivery of macromolecules: an update. Pharmaceutics 2019;11(1):41. https://doi.org/10.3390/pharmaceutics11010041
46. Taverner A, Dondi R, Almansour K, et al. Enhanced paracellular transport of insulin can be achieved via transient induction of myosin light chain phosphorylation. J Control Release 2015;210:189–97. https://doi.org/10.1016/j.jconrel.2015.05.270
47. Lin Y, Mi F, Lin P, et al. Strategies for improving diabetic therapy via alternative administration routes that involve stimuli-responsive insulin-delivering systems. Adv Drug Deliv Rev 2019;139:71–82. https://doi.org/10.1016/j.addr.2018.12.001
48. Iyire A, Alaayedi M, Mohammed AR. Pre-formulation and systematic evaluation of amino acid assisted permeability of insulin across in vitro buccal cell layers. Sci Rep 2016;6:32498. https://doi.org/10.1038/srep32498
49. Lancina MG, Shankar RK, Yang H. Chitosan nanofibers for transbuccal insulin delivery. J Biomed Mater Res A 2017;105(5):1252–9. https://doi.org/10.1002/jbm.a.35984
50. Xu Y, Zhang X, Zhang Y, Ye J, Wang HL, Xia X, et al. Mechanisms of deformable nanovesicles based on insulin-phospholipid complex for enhancing buccal delivery of insulin. Int J Nanomedicine 2018;13:7319–31. https://doi.org/10.2147/IJN.S175425
51. Matteucci E, Giampietro O, Covolan V, Giustarini D, Fanti P, Rossi R. Insulin administration: present strategies and future directions for a noninvasive (possibly more physiological) delivery. Drug Des Devel Ther 2015;9:3109–18. https://doi.org/10.2147/DDDT.S79322
52. Vernon G. Do insulin injections make you fat? Br J Gen Pract 2018; 68(669):188. https://doi.org/10.3399/bjgp18X695537
53. Yassin M, Sadowska Z, Tritsaris K, et al. Rectal insulin instillation inhibits inflammation and tumor development in chemically induced colitis. J Crohn's Colitis 2018;12(12):1459–74. https://doi.org/10.1093/ecco-jcc/jjy112
54. Nakadate Y, Sato T, Sato H, Koeva V, Schricker T. Hypoglycaemia after accidental ocular insulin injection. Br J Anaesth 2017;118(4):640–1. https://doi.org/10.1093/bja/aex065
55. Nightscout Foundation. We are not waiting. Nightscout support CGM in the cloud. 2019. Available: http://www.nightscout.info/.