Nanomedicine in diabetes management: where we are now and where next

Keywords::

• diabetes
• nanomedicine
• nanotechnology

‘Nanos’ is the Greek for dwarf and, in its broadest sense, nanotechnology involves making or measuring things on a very small scale. When this technology is applied to the problems of medicine it is called ‘nanomedicine’. As with many buzz words in bioscience (e.g., biotechnology, molecular medicine), nanotechnology and nanomedicine are somewhat loosely and arbitrarily defined and use aspects of several sciences that are not totally new [1]. How small must a structure be to be included in nanomedicine, for example? Although a nanometer (nm) is 10^-9 meters, by convention nanotechnology has so far mostly dealt with structures from 1 to 100 nm in size – larger than an atom but smaller than a cell.

Although we are interested in nanoscale substances in medicine because small size allows improved access across biological barriers and into tissues, there is another, perhaps more important, feature of the nanoscale: many substances have quite markedly altered properties at this level, for example, changing strength, color and chemical reactivity [2]. A more satisfactory and illuminating description of nanomedicine then is the manufacture, measurement and clinical application of very small (nano)-scale structures, molecular systems and devices with fundamentally altered properties. Among the structures that are commonly researched in nanomedicine are nanoparticles, nanovesicles, nanofilms and nanoporous membranes, nanotubes, nanocrystals and engineered proteins. The enhanced or changed properties that may be associated with these nanostructures include altered metabolite sensing, biocompatibility, controlled porosity, immuno-isolation, targeted imaging, targeted drug delivery, bioactivity (e.g., kinetics, substrate or ligand specificity) and coupled detection and treatment (theranostics).

Applications of nanomedicine in diabetes are in their infancy and none have reached routine clinical use. However, some outstanding problems in diabetes care, such as the need for improved glucose sensing, oral insulin formulations and transplantation of islets with enhanced survival, are likely to have nanomedicine solutions and research is already very active in these areas [3]. It is valuable then to consider the potential of nanomedicine from the perspective of some well-described nanostructures and to indicate how applications of these in diabetes management are being researched.

Nanofilms for making glucose sensors

The layer-by-layer (LBL) technique involves the electrostatic deposition of alternating layers of positively and negatively charged polymers, thereby building up a very thin film with tunable permeability and controlled biocompatibility [4–10]. A typical thickness is approximately 10 nm for six bilayers. An early use of such nanofilms has been in the construction of microvesicles for glucose sensing [5–10]. These might eventually be implanted in the dermis or subcutaneous tissue as a ‘smart tattoo’, with the semi-permeable capsules allowing entry of glucose from the interstitial fluid but containing and protecting the sensor materials [3,7]. If the sensing mechanism involves a near infrared (NIR)-based fluorescence assay, the glucose sensors can be excited and the fluorescence emission interrogated from outside the body, since NIR light passes through several centimeters of tissue. Therefore, this might be the basis for a noninvasive glucose-sensing technology.

Among the strategies for making microvesicles by the LBL technique are sequential absorption of polyelectrolytes around crystals of a glucose-sensing enzyme such as glucose oxidase [5], or absorption of a glucose-recognition molecule such as glucose oxidase, apo-glucose oxidase (the enzyme without the prosthetic group), concanavalin A or a glucose-binding protein (GBP) onto a sacrificial template (e.g., alcium carbonate), followed by stepwise addition of the charged polypeptides (e.g., poly-l-lysine and poly- l-glutamic acid). The template is then dissolved with ethylenediaminetetraacetic acid (EDTA), leaving a hollow microcapsule containing the sensor.

A number of glucose sensors with nanoengineered microcapsules have been described based on competitive binding, for example, a reduction in fluorescence-resonance energy transfer (FRET) when fluorophore-labeled dextran is displaced by glucose from fluorophore-labeled apo-glucose oxidase or concanavalin A [6,7]. As an alternative, we have constructed glucose-sensing microcapsules with a noncompetitive binding mechanism: capsules contain an engineered mutant of bacterial GBP labeled near the binding site with the environmentally (polarity) sensitive fluorophore, Badan [10]. Glucose binding causes a marked change in the conformation of the GBP, leading to a decrease in the polarity around the binding site and the fluorophore label, and thereby an increase in fluorescence intensity and lifetime.

Nanofilms for islet cell encapsulation

The problem of preventing the immune rejection of transplanted islet cells might also make use of LBL nanofilms. Permselective islet encapsulation with macro- or micro-coverings, with the intention of excluding immune proteins and cells while allowing glucose and nutrient access and insulin secretion, has been investigated for a number of years, but it has not been completely successful [11]. LBL nanofilms promise more-complete covering of islets (with better immune isolation), faster responses times and better nutrient access because of the thinness of the films and, with the use of modified polymers in the outer bilayer, the possibility of enhanced biocompatibility.

In an experimental model, we used ‘pseudoislets’ derived from the insulinoma cell line MIN6, which self-assemble into β-cell spheroids. Using LBL nanocoating with chitosan/alginate multilayers and a final protein-repelling nanolayer of phosphorylcholine-modified chondroitin sulfate, we demonstrated preserved glucose-dependent insulin secretion, long-term viability in cell culture and exclusion of antibody to MHC class I antigens [12]. These encapsulated islets thus have the characteristics required for efficient transplantation and in vivo transplantation studies with isolated rodent islets encapsulated in the LBL nanofilms are now taking place.

Nanoporous membranes for immunoisolation of islet cells

An alternative approach to protecting transplanted islets from immune rejection is to contain the islets in a micromachined nanoporous capsule based on silicon with etched pores down to 7 nm in diameter, 10–100-nm in length and a typical membrane surface area of 7 mm²[13]. It is speculated that xenografts of, for example, pig islets can be used for human studies using such nanoporous capsules, but this has not yet been rigorously tested.

Protein engineering for glucose sensing

There are likely to be many applications in the coming years of engineered protein nanostructures in nanomedicine, and as an example it is worth noting the use of GBP in glucose sensing (see earlier). GBP is a periplasmic-binding protein from Escherichia coli and other bacteria and it has excited much research in recent years as a glucose-recognition molecule because, as noted previously, glucose binding causes a marked change in the conformation of the protein. Monitoring this change in the tertiary structure of GBP, usually by change in fluorescence, can be a useful glucose-sensing technology. However, the binding constant (Kd) is in the micromolar range and unmodified GBP is thus unsuitable for clinical use as a sensor where blood glucose concentrations up to 30 mmol/l must be regularly measured [14].

In an attempt to increase the Kd of GBP, we synthesized by site-directed mutagenesis a number of molecular variants of the protein with amino acid substitutions suggested by the crystallographic structure as likely to affect glucose binding. The triple mutant H152C/A213R/L238S, which was labeled with the fluorophore, Badan, at position 152 near the binding site had a Kd of 11 mM and an operating range of 1–100 mmol/l, and is thus useful for development as a clinical sensor [15]. We have successfully incorporated this mutant of GBP in a fluorescence lifetime-based fiber-optic glucose sensor intended for implantation in the subcutaneous tissue of diabetic patients.

Nanoparticles for oral insulin delivery

Much research has been performed over the years in trying to develop a formulation of insulin that is orally active, thus, protecting insulin from denaturation at acidic pH in the stomach, preventing the proteolytic degradation of insulin by gut enzymes and enhancing the low permeability of the intestine to all proteins [16]. Several polymer nanoparticles incorporating insulin have been shown to lower blood glucose levels in animals when administered orally, including those consisting of biodegradable cyanoacrylate, polycaprolactone/polyacrylic polymers, casein and the polycationic polysaccharide, chitosan [16–18]. The latter polymer is mucoadhesive and prolongs residence time in the gut by binding the nanoparticle to the mucosa, but nanoparticles composed of chitosan and insulin have also been found to open the tight junctions between gut mucosal cells and improve paracellular absorption of insulin [18].

Ceramic nanoparticles such as those made of calcium phosphate, silica or alumina have the advantages of high biocompatibility and small size (<50 nm) and have also been investigated in animal studies as oral insulin formulations [16]. Other insulin nanoparticles that have been tested include those based on gold/chitosan [19] and dextran–vitamin B12 [20]. In most cases, nanoparticle formulations for oral insulin delivery have not progressed beyond animal testing.

A nanoparticle vaccine for antigen-specific immunotherapy of Type 1 diabetes

In a recent paper, Tsai et al. used iron oxide nanoparticles coated with disease-relevant MHC complexes to expand autoregulatory T cells in vivo in NOD mice, preventing disease in prediabetic mice and restoring normoglycemia in diabetic animals [21]. The nanovaccine is clearly an attractive candidate for clinical testing.

Carbon nanotubes for glucose sensing

Single-walled carbon nanotubes (SWCNTs) fluoresce in the NIR spectral region and, since they also suffer no photobleaching, SWCNTs are thus particularly suitable as fluorophore probes in glucose sensors designed for eventual in vivo use [22]. SWCNTs have been employed in a fluorescence-based competitive glucose-sensing strategy where dextran is bound to the carbon nanotubes, and binding of concanavalin A or apo-glucose oxidase to the dextran–SWCNT attenuates the fluorescence, which is reversed by the addition of glucose [22]. Apo-glucose oxidase has also been covalently attached to polyvinyl alcohol to make a glucose-responsive hydrogel that can be monitored by the fluorescence of SWCNT embedded in the hydrogel [23].

Quantum dots for glucose sensing

Quantum dots (QDs) are nanosized (2–10 nm) semiconductor crystals, such as cadmium selenide, coated with a shell, such as zinc sulfide [24]. They are especially useful as a fluorescence source because they display high-intensity fluorescence that is excitable over a broad range of wavelengths, but have an emission wavelength that is dependent on the particle size. QDs have been used as a fluorescent probe in several biosensor applications, often using FRET. Tang et al., for example, based a glucose sensor on FRET between QDs as a fluorescence donor and gold nanoparticles as an acceptor [25]. In this assay, glucose displaces concanavalin A-labeled QDs from gold-labeled cyclodextrin, thereby, reducing FRET and increasing fluorescence.

Imaging pancreatic islets in Type 1 diabetes using magnetofluorescent nanoparticles

Denis et al. have used superparamagnetic iron oxide nanoparticles (25 nm) coated with cross-linked dextran and linked to an Alexa fluorophore to create a highly fluorescent and MRI-detectable vascular probe, which can noninvasively image the islet cell inflammation associated with mouse models of Type 1 diabetes [26]. Such nanoparticles are nontoxic, nonimmunogenic and have a long circulation time. Hypothesizing that islet inflammation (insulitis) leads to microvascular leakiness, extravasation of nanoparticles and uptake into macrophages, these authors injected nanoparticles intravenously into animals and visualized the pancreatic area by MRI 24 h later. In general, the degree and progression of islet inflammation correlated with the MRI appearance. Such nanoparticles have been used in humans for cancer imaging, indicating that noninvasive islet imaging eventually may be possible in human diabetes.

The future

What can we expect from nanomedicine applied to diabetes management in the future? There is likely to be targeted imaging of diabetes complications, probably using antibodies or other molecules with an affinity for specific tissues and linked to a NIR-emitting fluorophore-labeled or MRI-visible nanoparticle. QDs with their strong fluorescence may be suitable fluorescent probes for imaging, probably passivated with silica to improve biocompatibility. Similarly, targeted drug delivery might use an antibody linked to a nanovesicle containing a specific drug for delivery directed at, for example, the liver.

Nanosized machines (‘nanorobots’) that circulate in the body, identifying disease and repairing and treating it have long been the subject of conjecture in nanotechnology and popular literature, and these devices continue be the subject of speculation in diabetes research [27]. Medical nanorobots, in a sophisticated sense of a machine with moving parts, seem to be far from clinical reality, but some automatic coupling of metabolite or disease-related tissue sensing with treatment in an integrated nanodevice or structure, which has been called ‘theranostics’, is quite possible. For example, there are several ongoing attempts to make a glucose-controlled insulin delivery system. One approach might be to use vesicles made of oxidation-sensitive polymers and containing glucose oxidase [28] and insulin. Hydrogen peroxide produced by the oxidation of glucose solubilizes the membrane of the vesicle, and if the vesicles were to contain insulin, it might be released in proportion to the glucose concentration.

To conclude, there are many exciting opportunities for research and development in nanomedicine applied to diabetes, and translation to routine clinical practice in at least some areas is not far away.

Financial & competing interests disclosure

The authors are grateful to the Engineering and Physical Sciences Research Council and the Diabetes Foundation for grant support. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.