Periodic Reporting for period 1 - T2D-EOMICS (High throughtput electrophysiological measurements coupled with transcriptomics to reveal cellular dysfunction in type 2 diabetes)
Reporting period: 2023-04-01 to 2025-03-31
Type 2 diabetes is a long-lasting condition in which the body loses its ability to regulate blood sugar levels. The pancreas is the organ responsible for this function. One of the main reasons for the progression of type 2 diabetes is that the pancreas becomes less capable of producing insulin. This occurs because the cells responsible for making insulin, beta cells, stop working properly. It is not yet fully understood what causes these cells to malfunction. Recent research examining individual cells has shown that not all beta cells are the same—they exhibit different patterns of gene activity. This discovery suggests that it is important to study how these differences relate to the loss of function observed in diabetes. However, a major challenge lies in figuring out how these different types of beta cells actually work and how they communicate with other cells in the pancreas: alpha and delta cells.
Cell functionality is largely determined by the genes expressed in each cell, which are eventually translated into proteins. Nowadays, the technique for quantifying how many genes or mRNA a cell has is single-cell RNA transcriptomics. However, transcriptomics alone does not reveal how well a cell performs its tasks. Therefore, the ultimate goal of this project is to couple function and transcriptome in human pancreatic samples. To achieve this, we are developing new methods to simultaneously measure transcriptomes and electrophysiology in islet cells with single-cell resolution. Traditionally, patch-clamp techniques have been used to perform electrophysiological measurements to study pancreatic cell secretion. A researcher can measure about 10 cells per day, and each cell can be measured only once. To scale these measurements, we use high-density microelectrode arrays (HD-MEAs) on small 3 × 3 mm chips. This approach enables the measurement of electrical activity from thousands of cells simultaneously, without damaging them, and facilitates repeated measurements over time, enhancing throughput and reproducibility. This provides us with the framework to understand the individual and collective roles of each cell type in diabetes progression.
Cell functionality is largely determined by the genes expressed in each cell, which are eventually translated into proteins. Nowadays, the technique for quantifying how many genes or mRNA a cell has is single-cell RNA transcriptomics. However, transcriptomics alone does not reveal how well a cell performs its tasks. Therefore, the ultimate goal of this project is to couple function and transcriptome in human pancreatic samples. To achieve this, we are developing new methods to simultaneously measure transcriptomes and electrophysiology in islet cells with single-cell resolution. Traditionally, patch-clamp techniques have been used to perform electrophysiological measurements to study pancreatic cell secretion. A researcher can measure about 10 cells per day, and each cell can be measured only once. To scale these measurements, we use high-density microelectrode arrays (HD-MEAs) on small 3 × 3 mm chips. This approach enables the measurement of electrical activity from thousands of cells simultaneously, without damaging them, and facilitates repeated measurements over time, enhancing throughput and reproducibility. This provides us with the framework to understand the individual and collective roles of each cell type in diabetes progression.
During this project, we established the technological foundation to perform high-throughput electrophysiology in pancreatic cells using HD-MEA. We optimized the culture and receiving conditions for multiple pancreatic cell types. In particular, we are able to measure rat beta cell line Ins1, primary cells from mouse and human, as well as human stem-cell-derived islets. Additionally, we set up the protocol to perform immunohistochemistry after electrical measurements to identify which cell types the electrical activity originates from. Finally, we are performing single-cell sequencing from individual chips after recording using the Low Input PARSE protocol.
To perform an initial validation of HD-MEA devices on endocrine cells, we used the rat cell line INS-1. We used this cell line because it is a well-established model for β-cells, and it can be easily cultured and expanded. This simplifies the testing of optimal conditions for measuring electrical activity. We investigated multiple coatings of the microelectrode array chips to facilitate the attachment and growth of these cell types since cells do not typically adhere well to metallic substrates. We found that using laminin helped in the attachment and survival of cells, and in establishing contact with the electrodes. Next, we developed a protocol to measure the electrophysiological responses of INS-1 cells in culture, finding that the HD-MEA array offers a good signal-to-noise ratio. We measured responses to different stimulatory agents such as glucose, diazoxide, and tolbutamide. Diazoxide blocks insulin secretion by opening KATP channels, causing membrane hyperpolarization and inhibiting action potentials. Tolbutamide, on the other hand, stimulates insulin secretion in β-cells by closing ATP-sensitive K+ (KATP) channels in the cell membrane of β-cells and triggering cell depolarization. We found that INS-1 cells respond under high glucose conditions (20 mM glucose) and to tolbutamide, and that they can also be silenced by reducing the glucose concentration (0 mM glucose) and by the addition of diazoxide.
We next validated that the HD-MEA can also be used in primary cells by using mouse pancreatic islets. Here, we expected a mixture of multiple endocrine cell types (α, β, δ) that respond differently to glucose and other stimulatory agents. However, in mice, the predominant cell type in islets are β-cells, representing over 75% of the total. After optimizing culture conditions, we could also detect action potential firing using the same stimulatory protocol used for INS-1 cells. We found that the electrophysiological activity in most electrodes is consistent with that of a β-cell, with activity under high glucose conditions and in the presence of tolbutamide, and silencing under low glucose conditions and in the presence of diazoxide. When aggregating the activity across all electrodes, we also noticed that about 10-20% of electrodes show activity under low glucose conditions, which is consistent with the presence of α- and δ-cells in the cell mixture.
We have also successfully performed experiments using human primary islet cells. This was an important achievement because activity in single-cell dispersions of human islets is less pronounced and more difficult to detect than in mouse cells. For human cells, most activity was detected under high glucose conditions, which could be silenced by the addition of diazoxide. However, tolbutamide doesn’t seem to further increase the activity as observed in mouse primary cells.
Next, we optimized the protocol to perform immunostaining on the chips after electrophysiological recordings. This allows us to map electrophysiological activity patterns to specific cell types or clusters of cells. We have successfully achieved this for human cells by staining the chips using antibodies for insulin, glucagon, and somatostatin to map the location of β-, α-, and δ-cells, respectively. We image the chips using upright confocal microscopy to identify individual cells and their proximity to each electrode of the chip.
We have also started experiments with stem-cell-derived islets. The development of insulin-secreting β-cells and islets from stem cells is a promising approach for T1D therapeutics. However, to date, it has been challenging to fully recapitulate the bona fide physiology of mature human islets using stem cells. An in-depth functional characterization of their maturation process at the single-cell level is of great interest to improve their performance, make them more similar to human islets, and reduce safety concerns. Therefore, we collect data from three maturation points at weeks 1, 3, and 5 of differentiation.
We applied our sequential stimulatory protocol and found results consistent with a variable composition of cell types. When seeding the same cell density onto the chips, we observed that the number of active electrodes increased with maturation. Across all maturation stages, the largest responses were observed through the modulation of K-ATP channels activity using tolbutamide and diazoxide, while glucose modulation was less pronounced than in primary human and mouse islets. In fact, low glucose conditions show equal or slightly greater activity than high glucose conditions. In week 5 matured sc-islets, we observed a larger response to tolbutamide compared to weeks 1 and 3. Another difference from primary mouse and human cells was the approximately twofold increase in response. We are currently analysing single-cell datasets for each measured chip using the Low Input PARSE protocol to identify genetic signatures of cell maturation associated with cell functional activity.
We noticed that both primary cells and stem-cell-derived islets can migrate and reorganize on the chip. Under low-cell density conditions, cells tend to adapt to the grid pattern of the HD-MEA, which is optimal for measuring the electrophysiological activity of single cells. At high-density conditions, cells reorganize as pseudo-islets that can be used to investigate cell-to-cell interactions and the paracrine regulation of islet cell activity.
To perform an initial validation of HD-MEA devices on endocrine cells, we used the rat cell line INS-1. We used this cell line because it is a well-established model for β-cells, and it can be easily cultured and expanded. This simplifies the testing of optimal conditions for measuring electrical activity. We investigated multiple coatings of the microelectrode array chips to facilitate the attachment and growth of these cell types since cells do not typically adhere well to metallic substrates. We found that using laminin helped in the attachment and survival of cells, and in establishing contact with the electrodes. Next, we developed a protocol to measure the electrophysiological responses of INS-1 cells in culture, finding that the HD-MEA array offers a good signal-to-noise ratio. We measured responses to different stimulatory agents such as glucose, diazoxide, and tolbutamide. Diazoxide blocks insulin secretion by opening KATP channels, causing membrane hyperpolarization and inhibiting action potentials. Tolbutamide, on the other hand, stimulates insulin secretion in β-cells by closing ATP-sensitive K+ (KATP) channels in the cell membrane of β-cells and triggering cell depolarization. We found that INS-1 cells respond under high glucose conditions (20 mM glucose) and to tolbutamide, and that they can also be silenced by reducing the glucose concentration (0 mM glucose) and by the addition of diazoxide.
We next validated that the HD-MEA can also be used in primary cells by using mouse pancreatic islets. Here, we expected a mixture of multiple endocrine cell types (α, β, δ) that respond differently to glucose and other stimulatory agents. However, in mice, the predominant cell type in islets are β-cells, representing over 75% of the total. After optimizing culture conditions, we could also detect action potential firing using the same stimulatory protocol used for INS-1 cells. We found that the electrophysiological activity in most electrodes is consistent with that of a β-cell, with activity under high glucose conditions and in the presence of tolbutamide, and silencing under low glucose conditions and in the presence of diazoxide. When aggregating the activity across all electrodes, we also noticed that about 10-20% of electrodes show activity under low glucose conditions, which is consistent with the presence of α- and δ-cells in the cell mixture.
We have also successfully performed experiments using human primary islet cells. This was an important achievement because activity in single-cell dispersions of human islets is less pronounced and more difficult to detect than in mouse cells. For human cells, most activity was detected under high glucose conditions, which could be silenced by the addition of diazoxide. However, tolbutamide doesn’t seem to further increase the activity as observed in mouse primary cells.
Next, we optimized the protocol to perform immunostaining on the chips after electrophysiological recordings. This allows us to map electrophysiological activity patterns to specific cell types or clusters of cells. We have successfully achieved this for human cells by staining the chips using antibodies for insulin, glucagon, and somatostatin to map the location of β-, α-, and δ-cells, respectively. We image the chips using upright confocal microscopy to identify individual cells and their proximity to each electrode of the chip.
We have also started experiments with stem-cell-derived islets. The development of insulin-secreting β-cells and islets from stem cells is a promising approach for T1D therapeutics. However, to date, it has been challenging to fully recapitulate the bona fide physiology of mature human islets using stem cells. An in-depth functional characterization of their maturation process at the single-cell level is of great interest to improve their performance, make them more similar to human islets, and reduce safety concerns. Therefore, we collect data from three maturation points at weeks 1, 3, and 5 of differentiation.
We applied our sequential stimulatory protocol and found results consistent with a variable composition of cell types. When seeding the same cell density onto the chips, we observed that the number of active electrodes increased with maturation. Across all maturation stages, the largest responses were observed through the modulation of K-ATP channels activity using tolbutamide and diazoxide, while glucose modulation was less pronounced than in primary human and mouse islets. In fact, low glucose conditions show equal or slightly greater activity than high glucose conditions. In week 5 matured sc-islets, we observed a larger response to tolbutamide compared to weeks 1 and 3. Another difference from primary mouse and human cells was the approximately twofold increase in response. We are currently analysing single-cell datasets for each measured chip using the Low Input PARSE protocol to identify genetic signatures of cell maturation associated with cell functional activity.
We noticed that both primary cells and stem-cell-derived islets can migrate and reorganize on the chip. Under low-cell density conditions, cells tend to adapt to the grid pattern of the HD-MEA, which is optimal for measuring the electrophysiological activity of single cells. At high-density conditions, cells reorganize as pseudo-islets that can be used to investigate cell-to-cell interactions and the paracrine regulation of islet cell activity.
Overall, these results show that HD-MEA devices can be used to measure electrophysiological activity in (1) pancreatic cell line INS-1, (2) primary mouse cells, (3) primary human cells, and (4) human stem-cell derived islets in multiple maturation time points, which had not been shown before. This is important, as these devices are more commonly used to detect electrophysiological activity of larger cells (e.g. neurons and cardiomyocytes) which are easier to detect using extracellular recordings with microelectrode arrays.
Coupling the electrophysiological measurements with immunohistochemistry allows us to establish a solid platform for combined functional and molecular phenotyping of pancreatic cell types across systems (2-4). We are currently conducting in-depth analysis of thousands of electrodes from multiple chips that have different cell type abundances.
Coupling the electrophysiological measurements with immunohistochemistry allows us to establish a solid platform for combined functional and molecular phenotyping of pancreatic cell types across systems (2-4). We are currently conducting in-depth analysis of thousands of electrodes from multiple chips that have different cell type abundances.