FUNCTIONAL PANCREATIC ISLET NETWORKS IN NUTRIENT HOMEOSTASIS IN MEN AND MICE – FUN-NET

We used genetically modified mouse models that allow for the inducible ablation of a specific cell or hormone type in the islets of Langerhans: alpha-cell-deficient islets (GluDTR), beta-only islets, and alpha/beta-deficient islets. The main advantage of these models is that untreated mice provide an excellent control group, and the ablation occurs in adulthood, thus avoiding developmental problems.

We also adapted the parameters to more physiological values: whereas previously only glucose was used, we simulated a meal by adding a physiological mixture of amino acids.

The nutritional status of the mice was first assessed in vivo using standard tests, and then their electrical activity was continuously monitored in vivo using microelectrode arrays. We also used high-density gratings to localize electrical activity to specific areas of the islets. Finally, microfluidics enabled the gradual administration of nutrients, mimicking their arrival in the islets in vivo after a meal. We also determined hormone secretion using standard insulin and/or glucagon assays.

Finally, electrical recordings at precisely known nutrient concentrations allowed us to transpose islet activity into a widely recognized in silico model of human metabolism. This enables us to monitor blood glucose levels over several days in this human in silico model and determine whether a given cell type is sufficiently capable of regulating homeostasis.

T2.1. GluDTR animals, devoid of alpha-cells, mice showed reduced tolerance to glucose and amino-acids (mimicking meals). In-vitro analysis demonstrated reduced insulin secretion in response to glucose (G) in the presence of a physiological amino-acid mix (AAM). This was recovered by addition of the alpha-cell hormone glucagon. Collectively the data underline the relevance of alpha-cell input on beta-cell function under physiological meal conditions.

Electrophysiological analysis showed similar results. We developed novel algorithms to analyse functional β-cell networks using our data from high density MEA recordings. The characteristics of leading regions were preserved in GluDTR islets, but synchrony, cluster size and signal propagation speed were largely reduced. Thus, α-cells are required for nutrient homeostasis by regulating the dynamics of β-cell networks.

T1.1. and T2.2. (beta-only, islets containing only beta-cells).

A new microfluidic set-up was developed in the CBMN group to combine electrophysiological recordings in conjunction with hormone measurement. The in-vitro characterisation confirmed the improvement of glucose homeostasis when challenged with glucose only in the “beta-only” mouse model. In stark contrast, mimicking meals using G and AAM a considerable diminution in glucose tolerance was apparent in beta-only animals. In-vitro we noted a strong reduction in insulin-secretion in beta-only islets stimulated by G and AAM, in line with our in-vivo data. Using our novel microfluidic set-up we have subsequently performed stepwise increases in G and AAM and measured and analysed the electrophysiological responses. A stark reduction in activity was apparent upon addition of AAM to G, mainly in the amplitudes (which are a good indicator of physiological coupling among islet beta-cells) as given in Fig. 1.

T1.2. and T2.2. alpha/beta-only (islets containing only alpha and beta-cells, no gamma or delta).

We have investigated the alpha/beta model. We observed initially an astonishing phenotype, which, as shown by transcriptomic and immunohistological examination, is due to considerable loss of alpha cells as the DTR is here under the PPY promoter, the reason is most likely due to low expression of the (typical gamma hormone) PPY-gene also in alpha cells. We also resorted to a homozygous model, that still contains the delta- and gamma cells but no longer expresses their corresponding hormones. Its characterisation is currently still in progress.

T3.1. In-Silico Model Integration and T3.2. In-Silico Evaluation

Electyrophysiological data were used for simulations on an FDA-approved in-silico model of nutrient homeostasis in man (Fig. 2). They clearly demonstrate that alpha-cells are required to obtain physiological nutrient homeostasis (ie in presence of G and AAM) (Fig. 2). Collectively these observations show, that optimal nutrient homeostasis needs both, beta- and alpha cells to face a normal meal in man.

First, we have developed and used novel approaches that will be important to study islet physiology and pathophysiology: microfluidic electrophysiology. This expertise is also of importance for ongoing developments of microfluidic chips that contain more than just one organ. These chips may be extremely useful for future research in physiology but also pathophysiology and could largely replace animal experiments.

The use of electrophysiology to study networks of electrical active cells, such as in the islets, will benefit from material sciences and we are currently trying to develop chips with closer spaced electrodes or transistors to increase resolution to the single cell level.

Our work has considerable impact in our understanding of islet function and of therapeutic approaches. Indeed, current development of therapy of Type 1 diabetes in the frame of iPSC-derived organ transplantation is largely centered on beta-cells only. Our work demonstrates that alpha cells are equally important and suggests strongly that future work on iPSC-derived surrogate islets has also to develop on alpha cells.

Finally, current therapy of T1D by “artificial pancreas” depends largely on algorithms to drive an insulin pump. These algorithms are strongly beta-cell oriented. As our data indicate, alpha-cells are also of prime importance here and their function should be considered when constructing the algorithms.

Pancreatic islets are central to nutrient homeostasis and diabetes. They contain 4 cell types (alpha/beta/gamma/delta), which alter their electric activity that tightly regulates hormone secretion in a distinct biphasic and pulsatile manner for up to 3 hours after a meal to restore homeostasis. The different cell types of the islet micro-organ form regulatory networks. Previous work addressed mainly glucose effects but not physiological mixed nutrients and lacked sufficient temporal resolution to provide unbiased insight into dynamic supracellular organisation. Moreover, multiple combinations of cell-type ablations have not been feasible. Models of islet function in-silico have been proposed but mostly for B-cells during steady-state (2nd phase) and never been transposed into human organism status. Consequently the precise physiological relevance of cell types is still only partially known in nutrient homeostasis and its pathological derangement in nutrient stress or diabetes.

FUN-NET is an interdisciplinary project of 3 groups with distinct expertise: (i) The HERRERA group (U Genève, co-coordinator) is specialized in endocrine pancreas plasticity and metabolic outcome in-vivo. Using CRISPR technology the group has generated new and highly versatile genetic mouse models for differential or combined specific cell ablation or hormone inactivation. They have also developed human pseudo-islets of defined cell-type composition. (ii) The LANG group (U Bordeaux CNRS, co-coordinator) has developed a microfluidic islet-on-chip for extracellular electrophysiology using micro-electrode arrays (MEAs, µMEAs) coupled to secretion analysis. This allows read-outs over hours/days with high temporal resolution on single cells or whole islets and unbiased direct network analysis; high-density MEAs provide considerable spatial information. (iii) The RENAUD group (Bordeaux INP), specialized in conceiving electronic systems for real time interaction with biology, has designed real time analysis and automated net-work analysis. Using an FDA approved human in-silico simulator of metabolism (T1DMS), they developed an electrical islet signature based model for glycaemia regulation during several meals over days.

We will address the following QUESTIONS:
(i) What is the specific contribution of B-cells alone on islet activity, phases, networks and homeostasis of different nutrients?
(ii) How does each of the non-beta cells impact B-cell function and gene expression?
(iii) How do they impact on islet activity under disease-mimicking challenges?
(iv) How do these cell-type specific contributions translate into human whole body homeostasis?

THE DETAILED OBJECTIVES are:
OBJECTIVE 1: In-vivo exploration of cell-type influence on mixed nutrient handling including metabolic challenges/high fat diet (Leader HERRERA)
OBJECTIVE 2: In-vitro exploration of cell-type influence on dynamic nutrient handling, functional intra-islet organization and gene expression (Leader LANG)
OBJECTIVE 3: Human in-silico exploration (T1DMS) of cell-type influence on B-cells in the human organism (Leader RENAUD)

FUN-NET will impact understanding of nutrient homeostasis and lead to a new concept of the islet micro-organ network closer to physiology, its pathological derangements and will provide substantial information for surrogate islet generation for diabetes therapy. It will also constitute the first translation of islet function into a human whole-body simulator. Moreover, harnessing functional islet principles, optimized in 0.5 billion years of evolution, may contribute to optimize the artificial pancreas via bio-inspired algorithms as current control algorithms are underperforming. FUN-NET will impact in terms of technology as the combination of genetic models, µMEA and human in-silico simulator, may provide a powerful, modular and cost-effective platform to address a number of fundamental and applied research questions as well as clinical issues not feasible in man.