Every islet matters: improving the impact of human islet research

Abstract

Detailed characterization of human pancreatic islets is key to elucidating the pathophysiology of all forms of diabetes, especially type 2 diabetes. However, access to human pancreatic islets is limited. Pancreatic tissue for islet retrieval can be obtained from brain-dead organ donors or from individuals undergoing pancreatectomy, often referred to as 'living donors'. Different protocols for human islet procurement can substantially impact islet function. This variability, coupled with heterogeneity between individuals and islets, results in analytical challenges to separate genuine disease pathology or differences between human donors from experimental noise. There are currently no international guidelines for human donor phenotyping, islet procurement and functional characterization. This lack of standardization means that substantial investments from multiple international efforts towards improved understanding of diabetes pathology cannot be fully leveraged. In this Perspective, we overview the status of the field of human islet research, highlight the challenges and propose actions that could accelerate research progress and increase understanding of type 2 diabetes to slow its pandemic spreading.

Fig. 1.. Comparative analysis of living and organ donors as the source of pancreatic tissue for islet studies.

The schematic representation illustrates some key opportunities and limitations specifically associated with either organ donors (left) or living donors (right). These range from clinical and laboratory profiling of the donor, amount and properties of the retrieved pancreatic tissue, possibilities for analysis of other organs and follow-up investigations of the same donor. The lower central box highlights challenges in the field, such as the need to increase the number of donors and recruiting centers by increasing public awareness and funding to support the infrastructures required for these activities. Best practice requires the standardization of the clinical phenotyping of the donors. This figure was created with BioRender.com.

Fig. 2.. Models for recruitment of donors and pancreatic tissue for islet isolation and characterization.

Model 1 (top) relies on a single isolating centre obtaining whole pancreas organs or pancreatic fragments from donors for tissue processing. Enzymatic isolated islets, pancreatic tissue slices and sections or islet fragments obtained by laser capture microdissection are then distributed to individual laboratories, which independently perform assays in parallel, some of which may be duplicated. Data analysis is restricted to the assays performed by a single laboratory, leading to the generation and publication of complementary, or potentially overlapping datasets, which are then deposited in public databases post-publication. This approach, which prevailed until recently, offers the opportunity to integrate different datasets from different groups and verify the reproducibility of data, but also entails a greater risk for redundancy and it is not well suited in cases where a finite amount of pancreatic tissue is available, as in the case of surgical samples from living donors. Model 2 (bottom) follows a more centralized workflow, with multiple sites each being responsible for the recruitment of living and/or organ donors, their phenotyping, and the retrieval of the islets/islet fragments by enzymatic isolation, laser capture microdissection or the generation of pancreatic tissue slices/sections are carried out according to shared standardised operating procedures. Biological samples are further distributed to different laboratories, each well-qualified for the performance of a specific predefined assay. Collected clinical, laboratory and meta-data arising from the analysis are then jointly integrated and stored in a database accessible to all members of the network. Following their joint publication and deposition in public databases, data become available to the whole community. This figure was created with BioRender.com.

Fig 3.. Models for islet isolation and characterization.

Whole pancreas from organ donors can be used for enzymatic isolation of islets, to obtain tissue slices and sections as well as islet fragments upon laser capture microdissection. In the case of surgical samples from living donors, the limited tissue availability precludes the enzymatic isolation of islets in numbers adequate in quality and quantity. As indicated, each of these different sources of islet material is best suited for a given set of analysis and comes with specific advantages (+) and limitations (−) relative to the others. Challenges and best practices common to all approaches are indicated. Xenotransplantation of human pancreatic tissue slices in immunodeficient mice is a potential future opportunity. This figure was created with BioRender.com.

Fig 4.. Schematic representation of data lifecycle management needs and challenges.

(left) Islet preparations are assigned a unique sample and donor identifier prior to data acquisition to ensure that each sample is unique and can be traced back to its origin (right). Data should be acquired using standardized protocols and be processed using common workflows for each technology to minimize batch and other confounding factors, ensuring data of the same type can be compared. Clinical information, sample metadata and molecular, functional, and imaging results need to be described according to international semantic standards to ensure interoperability between datasets. This enables them to be either integrated into a single database or be components of a federated network of interoperable datasets (as shown). Finally, standardized acquisition, processing and description of terminologies and pipelines enables clinical and laboratory data to be aligned and analyzed across donor cohorts, increasing statistical power to link findings to disease outcomes or molecular pathways. This can be done either by accessing a single database containing the combined data, or (as shown) via federated networks enabling interactive analysis of distributed databases (Federated analysis) or decentralised learning strategies such as Swarm learning where AI models can be learnt through exchange of model parameters between distributed databases. Current challenges and best practice are indicated. This figure was created with BioRender.com.