Use of human islets to understand islet biology and diabetes: progress, challenges and suggestions

Improved access to human islets for research

Initially, efforts to isolate human islets focused primarily on isolation of a sufficient number and quality for transplantation into individuals with type 1 diabetes. With the support of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) in the USA, the JDRF, the Special Diabetes Program and others, clinical human islet isolation centres in the USA, Canada, Europe and Australia began to isolate and distribute human islets for research-specific purposes through programmes like the Islet Cell Resource Center Consortium [1, 2]. These research-supporting islet isolation and distribution programmes have continued to evolve worldwide and now include the Integrated Islet Distribution Program (IIDP), the European Consortium for Islet Transplantation (ECIT), the Nordic Network for Clinical Islet Transplantation, the Australian Islet Transplant Consortium, and the IsletCore Program and the Alberta Islet Distribution Program (both at the University of Alberta in Canada). Over the past 15 years, these programmes and commercial sources have supplied millions of human islets to hundreds of scientists, leading to a dramatic increase in the number of scientific publications describing research with human islets.

Improved understanding of human islet biology

Greater availability of human islets and expanded research efforts have led to an improved understanding of numerous facets of human islet biology. Regrettably, space limitations prevent more than a partial listing of selected advances and references (see Text box 1 and Table 1, Human column).

An important result of greater human islet accessibility has been a greater appreciation of the similarities and differences between human and rodent islets. Improved understanding of the characteristics of both types of islets has led to the establishment of research initiatives, such as the Human Islet Research Network (HIRN; https://hirnetwork.org), to better understand human islet biology. Table 1 summarises selected features of human and rodent islets. We do not question the use of rodent islets, but instead emphasise that to fully understand the role of the islet in human diabetes and to translate findings from commonly used models, the field needs integrated approaches incorporating multiple sources of islets and experimental designs that recognise the shared and disparate characteristics of human islets, non-human islets, and insulin-producing cells derived from pluripotent stem cells.

Factors affecting functional properties of human islet preparations

Several studies have identified donor, pancreas and pancreas-processing characteristics that influence islet isolation yield and transplant outcomes [4–11]. In contrast, few studies have examined factors that affect the functional quality of human islets distributed for research or clinical islet transplantation. Surprisingly, the characteristics and assays that define human islet quality are not universally accepted or standardised. Insulin secretion, assessed either in static assay or in a dynamic perifusion system, is the most commonly reported measurement of islet quality. Secretion of other islet hormones, such as glucagon, and other islet metrics (e.g. oxygen consumption) have not been assessed in a large set of human islet preparations. Thus, for this review, we considered insulin secretion as the primary in vitro measure of islet quality. This discussion about the criteria for islet quality only applies to human islets used for research; clinical islet transplantation programmes have developed quality metrics for human islets used for clinical transplantation [12].

Three recent reports have assessed the in vitro islet insulin secretory function of a large number (>50) of human islet preparations by perifusion or static culture [9, 13, 14]. Two reports described human islet preparations generated at one or two islet isolation sites (the University of Alberta or the University of Louvain and the University of Lille) [9, 13], and the third report described preparations from multiple IIDP-sponsored isolation centres [14]. These studies used different methodologies to examine the determinants of human islet functional variability. Nonetheless, they represent a strong foundation to address sources of human islet functional inconsistency that may arise from variation in donor characteristics, pancreas processing and/or isolation centres [9, 13, 14].

The study that examined IIDP-derived islet preparations found that most human islet preparations secreted insulin in response to increased extracellular glucose and increased glucose with cAMP stimulation during islet perifusion. However, nearly a quarter of human islet preparations were dysfunctional (i.e. unstable basal insulin secretion and/or impaired stimulated insulin secretion) [14]. This, and other studies [9, 13, 15], suggest that the following characteristics affected in vitro islet insulin secretion (i.e. basal and/or stimulated insulin secretion): isolation centre [14], cause of death (COD) [14], estimated islet purity [13, 14], BMI [13, 14], islet size [13], pancreas cold ischaemia time (CIT) [9], age [9, 15] and HbA_1c [9]. The Alberta-based study noted several similar factors impacted human islet function but concluded that these made a modest contribution to the overall human islet functional heterogeneity observed [9].

Overall, these studies emphasise important considerations for research using human islets, including (1) considerable functional heterogeneity exists between human islet preparations; (2) a significant percentage of human islet preparations are dysfunctional as defined by insulin secretion; (3) explanations for this heterogeneity or dysfunction are not clear, and multiple biological, donor, transport, isolation and as-yet unknown characteristics likely influence islet quality; and (4) the functional state of an islet preparation cannot be predicted or assumed a priori. In light of these challenges, we believe that investigators, journals and funding agencies must develop more holistic, nuanced and sophisticated standards and criteria for assessing and reporting critical characteristics of human islets used for research.

Characteristics of human islet preparations reported in recent scientific publications

Since human islet quality is a critical consideration for research and may be influenced by various characteristics, we assessed which characteristics of human islet preparations have been reported in recent publications. We conducted a PubMed literature search of four leading journals that publish human islet research (Cell Metabolism, Diabetes, Diabetologia and Journal of Clinical Investigation) from 2013 to mid-2017, and combined these search results with the IIDP publication list (Fig. 1a). We identified 241 publications that contained studies of isolated human islets. We then assessed whether each publication reported a selected set of 19 islet characteristics related to islet origin, islet function, donor demographics and pancreatic/islet handling that have been reported to, or would be predicted to, influence islet quality (Fig. 1b, c). Review of articles published in other journals would be expected to produce similar findings.

In these 241 publications, the information about the human islet preparations was often difficult to find, quite variable and in many cases did not adequately report the nature of the human islets used. Several studies reported none of the 19 selected islet characteristics examined (Fig. 1b). The most detailed publication reported 14 of the 19 characteristics (Fig. 1b). Almost 50% of human islet studies described four or fewer of the selected characteristics (Fig. 1b). The islet source and experimental culture time were the only two characteristics reported in more than 50% of the reviewed publications (Fig. 1c). We noted that ~15% of publications did not report where the investigators obtained the human islets from and ~70% did not report the islet isolation centre. Furthermore, the majority of publications did not report details of human islet donor demographic information. Age, sex and BMI were reported most commonly (reported in 45%, 41%, 39%, respectively), then COD, ethnicity and HbA_1c (15%,10%, and 9%, respectively). In addition, the vast majority of studies did not mention warm ischaemia time (1%) or CIT (7%) or the total time of islet culture (5%).

More than 50% of publications did not report an assessment of islet function. Thus, the quality and function of the human islets used in those publications were not relayed to the readership and were potentially not known. The most common functional islet assessment method was static islet culture (32%), followed by perifusion (13%) and electrophysiology (12%), with some studies reporting more than one measure of islet function. Estimated islet viability and purity and efforts to purify islets after arrival in the user’s laboratory (e.g. handpicked or sorted) were reported in 26%, 25% and 29% of studies, respectively. Since additional islet purification was not reported in most studies, and human islet preparation purity and determination of purity can vary considerably, the possibility of inclusion of non-islet cells (acinar, ductal, etc.) in RNA and protein studies is quite likely and could be problematic.

We also observed that in publications that studied numerous islet preparations, it was often difficult to discern which islet preparations were used for a particular experiment, highlighting the need for a clear description of human islets used in specific experiments within publications. Finally, we observed that human islet experiments were often a minor aspect of the manuscript. For example, many manuscripts had a single figure or figure panel with human islets that contained limited experiments. We appreciate that human islet scarcity and cost are substantial challenges and likely to be responsible for these experimental limitations (addressed in [3]). Nevertheless, if human islets are used in research studies, key islet characteristics should be reported and the experimental design should be transparent and rigorous to optimise data interpretation.

One can readily see how different interpretations or conclusions related to human islet biology could arise from different laboratories performing human islet experiments at different times with different islet preparations. Thus, standardised and more detailed assessment and reporting of biological and non-biological variables such as pancreas processing, donor and islet characteristics are needed. There is also certainly biological variability between human islet preparations and future efforts are needed to define these sources of this variability.

Need for a unique identifier for each islet preparation

We propose that each islet preparation be assigned a unique identifier, something equivalent to a Globally Unique Identifier (GUID) or Universally Unique Identifier (UUID) that is used to identify information in computer systems (https://en.wikipedia.org/wiki/Universally_unique_identifier). The IIDP has recently adopted an approach to assign each human islet preparation a Research Resource Identifier (RRID) based on the US National Institutes of Health (NIH) Resource Identification Portal created as part of their initiative to authenticate key biological resources (https://scicrunch.org/resources). The creation, use, and reporting of an RRID for each human islet preparation would allow investigators to identify which islet preparations were used in particular publications and if other investigators used the same islet preparation. A unique research identifier for each human islet preparation would also create an organisational framework, allowing establishment of centralised databases in which multidisciplinary human islet data from multiple sources and investigators could be stored, referenced, accessed and harmonised.

Infrastructure support

Funding agencies and academic medical centres directly or indirectly support several components of the human islet research ecosystem. Funding agencies should continue to provide critical financial and organisational support for human islet production and distribution. Funding agencies are also vital in ensuring that different components of the human islet research ecosystem work cooperatively and meet their responsibilities.

Islet isolation, phenotyping, and distribution

As part of their regular activities, islet isolation centres and islet distribution programmes should continue to gather information on donor demographics, document pancreatic handling before and during the isolation procedures, assign a unique identifier to islet preparations, and make this information available to recipient investigators. Because the islet isolation centres are dispersed and function largely autonomously, phenotyping of human islets used for research has been performed at each isolation site and/or has been left to individual laboratories. As evident from our literature review, most publications have not reported the function and health of islet preparations and there is a lack of agreement on a standardised way of assessing human islet health and function. Investigators must establish and report the function and health of islets used in their studies. We support efforts to establish centralised islet phenotyping centres and/or standardised approaches to assess the health, function, number and purity of all research islets. To this end, the IIDP has recently developed the Human Islet Phenotyping Program (HIPP), which will provide phenotypic data on all IIDP-associated islet shipments. In this programme, a small aliquot of each IIDP-distributed islet preparation is evaluated in a centralised phenotyping centre using standardised protocols, allowing for enhanced comparison of experimental outcomes with different islet preparations. The HIPP is assessing human islet insulin and glucagon secretion, estimated viability, estimated purity and morphological characteristics. These data will be made available to investigators using IIDP islet preparations. The University of Alberta IsletCore also characterises each of their islet preparations and makes this information available upon request. By informing investigators in a timely manner about the function and health of islet preparations, centralised phenotyping centres will ensure that investigators know the quality and function of the islets used in an experiment. Additionally, centralised phenotyping centres could also reduce net experimental cost by eliminating the need for extensive islet phenotyping by individual laboratories. Furthermore, such efforts may foster collaboration between investigators using the same human islet preparation and lay the foundation for human islet data repositories, such as is being done in the Nordic Network for Clinical Islet Transplantation (https://nordicislets.medscinet.com/en/islet-isolation/nics-database.aspx). Since human islets are being distributed for research in the USA, Canada, Europe and Australia, we suggest, as a first step, that the human islet ecosystem work collaboratively to standardise islet assessment through workshops and other efforts, as has been done for standardisation of other assays (e.g. HbA_1c and islet cell autoantibodies). In this way, centralised phenotyping centres could also determine which assays are optimal for assessing human islet health and function. For example, whether additional assays beyond insulin secretion, such as glucagon secretion and oxygen consumption, should be considered for assessing human islet health. Unfortunately, it is not known whether the same assessment of islet quality is needed for islets used in very different experimental approaches (e.g. genetics, protein expression, RNA isolation).

Investigators using human islets for research

We propose that investigators using human islets should be responsible for ensuring that the appropriate donor and pancreas characteristics are compiled and recorded (Table 2). In addition, investigators should report how and when human islets were used, including the number of islets used in particular experiments and whether the islets were further purified in any manner. We suggest that investigators studying human islets (1) report stimulated hormone secretion; (2) determine at the receiving institution the quantity of human islets used in experiments (i.e. in-house calculation of islet equivalents/amount); (3) if conducting studies of islet proteins or nucleic acids, use purified islets or islet cells or report the relative contributions of various cell populations; and (4) record and report key islet characteristics and experimental use.

Scientific journals

Scientific journals publishing reports on human islet research should ensure that key human islet characteristics are reported and presented in a clear and accessible manner. To facilitate the reporting of key information about human islet preparations, we propose a checklist detailing human islet characteristics that would be completed by investigators and submitted with the publication, as is now required by several journals for methods and reagents (ESM Table 1). We hope the proposed checklist will stimulate discussion and will be modified and improved by investigators and journals moving forward.

Other considerations for handling and reporting human islet samples

In addition to human islets from non-diabetic individuals, improved pancreatic procurementisolation-delivery chains have increased the availability of pancreases and islets from individuals with diabetes (type 1, type 2, pancreatogenic, cystic fibrosis-related, monogenic) for use in research [16–26]. These valuable islets present their own unique set of challenges, including the cellular composition of the islet, which may differ from normal islets and affect their function. We suggest that investigators using isolated islets from individuals with diabetes collect and report as much clinical and research information as possible without compromising the anonymity of the islet donor. This is especially true for islets from individuals with type 2 diabetes, where clinical information (stage of disease, treatment, etc.) is often unavailable or not reported. Furthermore, recent studies highlight that clear and correct clinical diagnosis of type 1 and 2 diabetes, monogenic diabetes or pancreatogenic diabetes is more difficult than previously appreciated [27–32], emphasising the need for holistic investigation and integration of donor information and research with pancreatic tissue and islets.