In the current study, we used a syngeneic minimal islet mass transplantation model in STZ-induced diabetic mice to demonstrate that maintaining normal islet size and morphology at the implantation site is beneficial for transplantation outcome. We used two different experimental approaches to maintain normal islet morphology by preventing islet fusion and thus limiting the formation of large amorphous endocrine aggregates. The first approach was to transplant islets at the renal subcapsular site by manually dispersing individual islets beneath the kidney capsule, as opposed to the standard procedure of transplanting them as a single pellet or cluster. In the second approach, islets were mixed with and transplanted in matrigel to ensure that the transplanted islets were physically separated beneath the kidney capsule by the solid matrigel support.
Our morphological measurements showed that manually dispersing islets beneath the renal capsule reduced the size of individual endocrine aggregates by approximately 75 percent compared to grafts of pelleted islets, consistent with maintaining normal endogenous pancreatic islet size at the subcapsular implantation site. Providing the transplanted islets with the physical support of a matrigel matrix was equally effective in maintaining individual islet morphology under the kidney capsule, producing similar effects on islet distribution and anatomy at the graft site. Immunostaining of islet α-cells for glucagon expression showed that islet architecture was maintained in both the dispersed and matrigel islet grafts, with the majority of graft sections showing a peripheral rim of α-cells surrounding the β-cell core, in contrast to the disorganised core-mantle cellular architecture in pelleted islet grafts 
. Both methods of maintaining islet structure were also associated with significantly enhanced revascularisation of the graft endocrine tissue, as demonstrated by increased vascular density when compared to conventional pelleted islet grafts.
Immunohistochemical analysis of the STZ pancreata at one month post transplantation revealed very low numbers of insulin-positive cells and all cured mice that were nephrectomised reverted to severe hyperglycaemia (blood glucose ≥33.7 mmol/l). This is consistent with our previous observations where we demonstrated that spontaneous pancreatic β-cell regeneration is unlikely to account for improved glycaemia in high dose STZ-diabetic mice over a 1 month monitoring period. Instead, the maintenance of islet anatomy in grafts consisting of both dispersed islets and matrigel-implanted islets is associated with improved transplantation outcome in the current study. Thus, when compared to conventional islet pellets, both methods to maintain islet morphology and size enhanced the rate and frequency of reversion to normoglycaemia in STZ-induced diabetic mice and showed significant improvements in overall glycemic control. Notably, we observed an initial decrease in blood glucose in all islet graft recipients, which we believe is not physiologically relevant. Instead, this is likely to be due to extensive islet cell death 
and subsequent insulin leakage from dying cells during the immediate post transplantation period. The real differences in glycaemia are present at 2–4 weeks post transplantation when the anatomical remodelling and revascularisation process are known to be completed 
Matrigel is a solubilised basement membrane preparation extracted from an Engelbreth-Holm-Swarm mouse sarcoma 
, in which the main components are ECM proteins such as laminin, collagen IV, fibronectin and perlecan 
. These basement membrane proteins are involved in interactions between intraislet ECs and endocrine cells 
and a number of studies have suggested that loss of integrin signalling between islets and the surrounding ECM proteins is detrimental to islet function 
. Conversely, entrapment of islets within ECM scaffolds is reported to enhance islet function 
and survival 
. In the present study we did not detect any additional in vivo
benefit of suspending the islets in matrigel over and above the improved function associated with the maintenance of islet morphology by physical dispersion below the renal capsule. This does not imply that islet-ECM interactions are unimportant, but suggests that interactions with the specific matrix components present in matrigel are neither beneficial nor detrimental for islet survival and function in vivo
when transplanted to the renal subcapsular site. Thus, the beneficial effects of matrigel in our experimental model can be attributed to its role as a physical support to maintain islet anatomy.
There are a number of mechanisms through which maintained islet architecture may have beneficial effects on graft function and transplantation outcome in our studies. Hypoxia-related dysfunction 
and cell death 
is an important confounding factor in the survival of avascular islets during the immediate post-transplantation period. Oxygen tension gradients across fused islet tissue have been demonstrated previously 
, with higher partial pressures of oxygen at the periphery of the islet graft compared with centrally located parts of the graft. Diffusion of oxygen and nutrients will be more effective in smaller endocrine aggregates enhancing their survival and function until the re-establishment of the islet vasculature. Small islet aggregates have previously been shown to be superior to large intact islets as graft material in diabetic mice, with improved transplantation outcomes being associated with reduced hypoxia-related necrosis in the small islet aggregates 
. Importantly, this benefit of small islet aggregates over large intact islets was demonstrated using encapsulated islets which do not revascularise in vivo
, so the improved islet function was independent of any influence on islet revascularisation.
Graft revascularisation is obviously important for subsequent function and inadequate revascularisation of transplanted islets at a number of implantation sites is associated with deleterious outcomes 
, whereas improvements in graft revascularisation are associated with improved islet function and long-term survival 
. Our results demonstrate that maintaining individual islets at the graft site resulted in a significant enhancement of revascularisation, consistent with a previous report of superior revascularisation of small, compared to larger islets 
. Similarly, in our previous study where we co-transplanted islets with MSCs, the resultant smaller endocrine aggregates had an enhanced vascular density compared to that of the large endocrine masses formed in mice implanted with islets alone 
Intra-islet interactions are known to be important for normal islet function 
and disruption of islet architecture is associated with impaired secretory responses to a range of physiological stimuli. Maintaining anatomically correct islet architecture may therefore further enhance graft function by facilitating the numerous interactions between islet cells 
that are required for normal insulin secretion 
Our observations using the renal subcapsular graft site are in accordance with recent studies of intramuscular islet transplantation, in which islets grafted as clusters developed central fibrosis 
, whereas transplanting the islets in a ‘pearls-on-a-string’ configuration, such that they are engrafted essentially as single islets, was associated with improved transplantation outcomes 
. This suggests that the beneficial impact of maintaining islet anatomy during transplantation is not graft site-specific.
In conclusion, there is mounting evidence that the current intraportal route for clinical islet transplantation places the grafts into a hostile microenvironment and confers multiple and perhaps avoidable stresses upon the transplanted islets 
, so efforts are being made to identify alternative optimal implantation sites for islets. The current study suggests that preventing the fusion of islets at extrahepatic sites represents an important strategy for promoting islet engraftment, which may contribute to achieving routine single donor islet transplantation 
, thereby increasing the availability of donor islet tissue and enabling the more widespread application of islet transplantation as a therapy for T1D.