Delivery of shRNA via lentivirus in human pseudoislets provides a model to test dynamic regulation of insulin secretion and gene function in human islets

Citations Over TimeTop 10% of 2018 papers

Abstract

Rodent islets are widely used to study the pathophysiology of beta cells and islet function, however, structural and functional differences exist between human and rodent islets, highlighting the need for human islet studies. Human islets are highly variable, deteriorate during culture, and are difficult to genetically modify, making mechanistic studies difficult to conduct and reproduce. To overcome these limitations, we tested whether pseudoislets, created by dissociation and reaggregation of islet cell suspensions, allow for assessment of dynamic islet function after genetic modulation. Characterization of pseudoislets cultured for 1 week revealed better preservation of first-phase glucose-stimulated insulin secretion (GSIS) compared with cultured-intact islets and insulin secretion profiles similar to fresh islets when challenged by glibenclamide and KCl. qPCR indicated that pseudoislets are similar to the original islets for the expression of markers for cell types, beta cell function, and cellular stress with the exception of reduced proinflammatory cytokine genes (IL1B, CCL2, CXCL8). The expression of extracellular matrix markers (ASPN, COL1A1, COL4A1) was also altered in pseudoislets compared with intact islets. Compared with intact islets transduced by adenovirus, pseudoislets transduced by lentivirus showed uniform transduction and better first-phase GSIS. Lastly, the lentiviral-mediated delivery of short hairpin RNA targeting glucokinase (GCK) achieved significant reduction of GCK expression in pseudoislets as well as marked reduction of both first and second phase GSIS without affecting the insulin secretion in response to KCl. Thus, pseudoislets are a tool that enables efficient genetic modulation of human islet cells while preserving insulin secretion. The loss of functional beta cell mass is the central pathology for both type 1 and type 2 diabetes (Kahn 2001; Atkinson et al. 2014; Chen et al. 2017). As the three-dimensional structure of pancreatic islets supports viability and function of beta cells through cell-cell and cell-matrix communications (Rutter and Hodson 2015; Arous and Wehrle-Haller 2017; Reissaus and Piston 2017; Briant et al. 2018), it is critical to address beta cell pathophysiology in pancreatic islets. Rodent islets are readily available, cost effective, can be easily genetically manipulated, and can be compared with syngeneic animals to connect in vitro observations to in vivo phenotype. However, human islets differ substantially from their rodent counterparts anatomically and functionally. Humans and mice show distinct islet innervation, cell distribution, and ratio of beta to alpha cells (Arrojo e Drigo et al. 2015). Glucose transporters, ion channels, the ratio of first/second phase of glucose-stimulated insulin secretion (GSIS), and amyloid deposition also differ between human and mouse islets (Arrojo e Drigo et al. 2015; Dai et al. 2016; Skelin Klemen et al. 2017). Thus, studies of human islets from organ donors are important for understanding the regulation of islet function and beta cell viability in humans. However, human islets are limited in availability, costly, difficult to maintain in culture, and challenging to genetically modify. Islet function, including GSIS, reduces over time in culture (Paraskevas et al. 2000; Arzouni et al. 2018). Genetic manipulation of intact islets by liposomal or viral-mediated vehicles has low efficiency and typically requires partial dispersion or enzyme digestion that compromises insulin secretion and removes cell-cell communication. The enormous heterogeneity of islet sizes also introduces high variability in assays. To overcome the variability of isolated human islets, islet spheroids or pseudoislets composed of dissociated reaggregated islet cells has been used. Reaggregated islet cells form uniformly sized pseudoislets that maintain similar spatial distribution of beta and alpha cells with better first phase GSIS compared with dispersed cells (Hopcroft et al. 1985; Halban et al. 1987) and original islets (Zuellig et al. 2017; Yu et al. 2018). Pseudoislets also lend themselves to more efficient gene modification (Caton et al. 2003; Arda et al. 2016; Peiris et al. 2018). Caton et al. reported that lentiviral-mediated overexpression of connexin cDNA does not interfere with pseudoislet formation and allows for the transduction of a large proportion of cells (Caton et al. 2003). Thus, pseudoislets appear to offer a unique and useful model to assess human islet function after the modulation of gene expression. Using readily available reagents and resources, we first characterized pseudoislets for insulin secretion in response to secretagogues by perifusion and analyzed their expression of markers for islet cell types, beta cell function, cell stress, and extracellular matrix (ECM). We compared transduction efficiency between adenovirus and lentivirus side by side and determine the impact of transduction on GSIS by perifusion. Following characterization of the pseudoislet platform, we tested short hairpin RNA (shRNA) delivered with a lentiviral vector targeting glucokinase (GCK) as a model target to test dynamism of insulin secretion by perifusion. Our data demonstrate that lentiviral-mediated gene downregulation combined with a simple protocol to form human pseudoislets is a useful tool that enables assessment of the impact of gene function on islet GSIS. Human islets from nondiabetic donors from Integrated Islet Distribution Program or PRODO laboratories (Table 1) with reported viability and purity above 80% were cultured in CMRL1066 containing 1% human serum albumin (HSA), 1% Pen-Strep, and 1% l-Glutamate (1% HSA CMRL) overnight at 37°C and 5% CO2 upon arrival for recovery from shipping. Then, islets were divided into fresh, cultured-intact, or pseudoislets. While fresh islets were harvested on the next day, cultured-intact islets were maintained in CMRL1066 containing 10% heat inactivated FBS, 1% Pen-Strep, and 1% l-Glutamate (10% HI-FBS CMRL) for 1 week at 37°C and 5% CO2 before harvesting. For pseudoislets preparation (Fig. 1A), single cell suspension was prepared first as follows. Human islets were washed once with PBS, digested with Accutase (A6964, MilliporeSigma, St Louis, MO) at 37°C for 5 min, pipetted through 1 mL tip for 15 times, digested for additional 4 min at 37°C, and passed through 40 μm strainer using a plunger of 1 mL syringe (Butcher et al. 2014). Filtered single cell suspension was counted, washed with PBS once, resuspended in 10% HI-FBS CMRL at 102 cells/μL, and seeded in a 96-well spheroid microwell plate (Corning, Corning, NY) at 3000 cells/well. The microwell plate was centrifuged at 270g at room temperature for 7 min and cultured after addition of 100 μL/well of 10% HI-FBS CMRL at 37°C and 5% CO2 until analyses. The study was reviewed by IRB at University of Iowa and approved as nonhuman study. Islets in 10% HI-FBS CMRL were incubated with 10 μg/mL Hoechst 33342 for 30 min at 37°C and 5% CO2. Then, Z-stack images captured by Leica DMi8 Microscope (Leica Microsystem, Buffalo Grove, IL) were analyzed with the ImageJ macro “Measure Spheroid Shape.ijm.”(https://www.researchgate.net/publication/3268440 33_ImageJ_Macro_to_Quantify_Spheroid_Volume_and_Size). The maximal intensity Z-stack was converted into a single plane and a binary threshold was applied to create a mask surrounding the islet cell mass. A series of erosions and dilations were performed to remove debris and the ImageJ plugin “Measure Particles” was used to find the dimensions of the islet fit using an ellipse approximation as well as directly measure the “Area” of the max projected islet cross section. The major and minor diameters of the calculated ellipse were then used to estimate the ellipsoidal volume of the islet that was defined as V = (4/3)*π*a*b2, where (a) is the major ellipse axis and (b) is the minor ellipse axis. Diameter was calculated as 0.5*square root of (area/π). The macro was run with the restrictions of circularity between 0.5 and 1 and islet area <18% of the total image area. Images which did not pass these quality control metrics were flagged, reviewed, and the image threshold was set manually. BioRep Perifusion System (Biorep Technologies, Miami Lakes, FL) was used to perifuse human islets at 120 μL/min and perfusates were collected every minute between 49 to 58 min and every 2 min for the rest of the run. After 52 min in 2.8 mmol/L glucose in Krebs-Ringer bicarbonate (KRB) buffer, islets were perifused for 16 min with 16.7 mmol/L glucose, 10 μmol/L glibenclamide in 5.6 mmol/L glucose, or 30 mmol/L KCl in 2.8 mmol/L glucose followed by 2.8 mmol/L glucose in vivo ose alone in KRB unless specified otherwise. Total insulin contents were obtained from islets incubated overnight at 4°C in acidified ethanol. Insulin was measured using STELLUX Chemiluminescent Human Insulin ELISA (ALPCO, Salem, NH). Insulin secretion was expressed by taking total insulin contents as 100% (% total) when the comparison is made among islets from the same donor or taking the average of insulin secretion during perifusion at 2.8 mmol/L (basal) as 1 when data from multiple donors were combined due to large variation of % total among donors and the lack of correlation between % total and efficiency of GSIS (Butcher et al. 2014). Stimulation index (SI) for the first phase was determined as the average insulin secretion between 53 and 56 min and SI for the second phase as the average of insulin secretion between 57 and 70 min, both divided by average basal insulin secretion. RNA was isolated from islets using TRIzol reagent (ThermoFisher Scientific, Waltham, MA) according to manufacturer's protocol and cDNA was synthesized using Superscript IV VILO Master Mix (ThermoFisher Scientific). Gene expressions were assessed using ABI TaqMan commercial primers (Applied Biosystems, Foster City, CA) and results were expressed taking human PPIB as an internal standard. Lentiviral vector expressing GFP under CMV promoter (LV-CMV-GFP) was obtained from University of Iowa Viral Vector Core (Iowa city, IA). Scramble and ShRNA sequence targeting human GCK obtained from Genetic Perturbation Platform (https://portals.broadinstitute.org/gpp/public) were cloned into PLKO.1 vector (10878, Addgene, Cambridge, MA) under human U6 promoter. Seventy percent confluent HEK293T cells in a 10 cm plate were transfected with 15 μg PLKO1-shRNA, 9 μg pMDLg/pRRE (12251, Addgene), 5.5 μg pRSV-Rev (12253, Addgene), and 5 μg pMD2.G (12259, Addgene) mixed with 103.5 μg PEI (MilliporeSigma) for 6 h and grown for 60 h in 10% FBS DMEM with 1% Pen-Strep, and 1% l-Glutamate. Thereafter, virus-containing media were centrifuged at 250g and filtered through a 0.45 μm filter. Viruses were pelleted by ultracentrifugation at 125,000g for 2 h, resuspended in 100 μL PBS, and stored in −80°C. For pseudoislets transduction, islet single cell suspension in 10% HI-FBS CMRL prepared as above was mixed with lentiviruses at approximately 0.8 x 106 TU/3000 islet cells and seeded in 96-well spheroid microwell plates at 30 μL/well as above except that 100 μL 10% HI-FBS CMRL was added after the overnight culture at 37°C and 5% CO2. For cultured-intact islet transduction, intact islets after overnight culture were resuspended in 0.1 mmol/L EGTA in serum free CRML and incubated with adenovirus expressing GFP under CMV promoter (Ad-CMV-GFP from Vector Biolabs, Malvern, PA) at 10,000 pfu/IEQ for 1 h at room temperature with mixing every 15 min before transferring to 10% HI-FBS CRML for culture. Data are presented as mean ± SEM or SD as specified. Differences of between were assessed with test was used when are between was applied when between were by test using 7 A was As of the of pseudoislets is their uniform compared to islets et al. 2018), we first assessed the variability of pseudoislet compared with islets for human area of pseudoislets and intact islets was measured and the ellipsoidal volume was from Z-stack images of (Fig. islets showed variability in area and volume donor as well as between donors (Fig. and As pseudoislet was highly both and between donors compared to intact islets. The of variation for area and volume donors was reduced from ± SD = ± and ± SD = ± for cultured-intact islets to ± SD = ± and ± SD = ± for pseudoislets. of cultured islets from donors from ± 52 μm to ± while that of pseudoislets were from ± to ± 9 μm mean ± among donors (Fig. Using a perifusion we compared GSIS of fresh islets with cultured-intact and pseudoislets that were cultured for 1 week (Fig. GSIS expressed as area under the was not among the (Fig. However, the first-phase insulin secretion in fresh human islets was in cultured-intact islets (Fig. perifusion profiles of pseudoislets from donors showed reduction in first phase compared with fresh islets, the first phase to be more that of cultured-intact islets (Fig. Thus, we calculated the ratio of SI for first and second phase GSIS to the of first phase GSIS as we that the ratio is a useful index when first-phase insulin secretion was compared between human islets from nondiabetic and type 2 islets (Butcher et al. 2014). The ratio was reduced to in cultured-intact islets ± SD = ± compared with fresh islets ± SD = ± pseudoislets showed a of reduction in first/second phase SI ± SD = ± compared with fresh islets = the ratio was compared with cultured-intact islets To demonstrate the ratio of first/second phase SI with the of first phase perifusion profiles of cultured-intact and pseudoislets from donor are in as donor first/second SI of for cultured-intact and for pseudoislets, of mean from we tested whether pseudoislets maintain response to islets insulin secretion to of in response to 10 μmol/L glibenclamide with in second phase secretion and did not insulin secretion after glibenclamide was (Fig. 30 mmol/L KCl insulin secretion from fresh islets and to of with first glucose with to after of KCl (Fig. were followed by pseudoislets (Fig. except was a of first-phase insulin secretion after glibenclamide in pseudoislets compared with fresh islets (Fig. = Thus, the in first/second phase SI between glibenclamide and KCl was significant in a fresh islet (Fig. The of expressed as did not differ between fresh islets and pseudoislets for both glibenclamide (Fig. and KCl (Fig. that pseudoislets and pancreatic cells (Zuellig et al. 2017; Yu et al. and indicated that the ratio of cells in the original intact islets and pseudoislets is similar (Zuellig et al. 2017). However, it is whether minor of cells in islets are into pseudoislets. Thus, we analyzed RNA expression of cell markers to a of cell including cells and pancreatic cells cells and (Fig. islets showed gene expression similar to fresh islets except for a in (Fig. Pseudoislets showed a of in compared with fresh islets = and reduction in compared with cultured-intact islets (Fig. = However, the cell type markers in fresh islets were at similar in pseudoislets that dissociation and reaggregation not for cell for in the pseudoislets. we measured expression of beta cell markers (Fig. and genes that GSIS (Fig. beta cell markers and and 2015; and were in pseudoislets compared with fresh islets (Fig. and were also in cultured-intact islets compared with fresh islets. was between cultured-intact and pseudoislets in expression of these genes (Fig. The expression of GCK and was also in cultured-intact and pseudoislets compared with fresh islets, while was in cultured-intact islets compared with fresh and pseudoislets (Fig. expression of these genes in pseudoislets is similar to cultured-intact and does not in of first phase GSIS among of islets. the loss of during human islet is to to of insulin secretion and viability in cultured islets, genes expressed in islets were analyzed et al. 2014; et al. 2018). to beta cell markers and GSIS genes that showed between cultured-intact and pseudoislets, was expression of and type 1 between cultured-intact and pseudoislets in in pseudoislets (Fig. Pseudoislets compared with fresh islets and cultured-intact islets and reduced expression of and type 4 compared with fresh islets (Fig. We did not in type 6 or that the loss of human islets during culture is with including et al. and et al. 2018), we tested genes to be in beta cells under (IL1B, CCL2, and stress and was a significant reduction in expression of proinflammatory markers CCL2, and for pseudoislets compared with fresh islets (Fig. However, pseudoislets did not show in and when compared with fresh or cultured-intact islets. significant in of and stress markers were for cultured-intact islets compared with fresh islets (Fig. The modulation of gene expression in intact human islets typically low efficiency when using adenovirus, which the into islets among Thus, we compared the efficiency of transduction of human cultured-intact islets using the with pseudoislets during the dissociation phase using a lentiviral As in the pseudoislets the as was in The by the in the cultured-intact islets as the culture was to at the of islet (Fig. GSIS tested by perifusion did not differ between transduced and islets, pseudoislets showed better preservation of first phase GSIS cultured-intact in with and (Fig. preservation of first phase was by the in SI ratio of first phase in compared with (Fig. To demonstrate the of the pseudoislet as a tool to study the of gene downregulation on dynamism of GSIS, we performed using targeting GCK GCK was genetic in et al. et al. and and of GCK are to to a in GSIS, making it an with in significant downregulation of the GCK gene in comparison with transduced pseudoislets when measured by qPCR (Fig. both the first and second phase of GSIS was in pseudoislets while the pseudoislets first and second phase GSIS (Fig. and similar to pseudoislets without (Fig. and the loss of glucose 30 mmol/L KCl insulin secretion in pseudoislets transduced by or (Fig. and is in with in of of in beta cells and the transduction glucose without from the study demonstrate that human pseudoislets can as a tool to study the dynamic regulation of insulin secretion using gene downregulation by side comparison of intact islets and showed that a better first phase of GSIS compared with We gene that is to GSIS et al. et al. to efficiency of gene downregulation in human pseudoislets by lentivirus both at gene expression and by GSIS. reduced GCK expression and in first and second phase GSIS while preserving KCl response in human pseudoislets a of for assessment of both of insulin secretion in genetically human pseudoislets. and of the pseudoislets a high in gene expression of cell type markers and response to secretagogues between human pseudoislets and the original fresh islets. Thus, human pseudoislets created by a simple protocol as a useful tool to assess the genetic control of the dynamic regulation of insulin secretion three-dimensional cell of genes as and showed of human pseudoislets in assessment of GSIS (Caton et al. 2003; Arda et al. of downregulation of et al. and study targeting GCK that pseudoislets are in the assessment of functional gene which requires high data that enables studies to the of genes of in a model with better first and second phase GSIS. GSIS in is to a critical in the of and is to be in diabetes at the of the et al. 2000; et al. et al. of first-phase insulin secretion in perifusion was when human islets isolated from donors were compared with from nondiabetic donors that the of first phase by first/second phase ratio is an important of human islet function (Butcher et al. 2014). of first phase impact control in humans. the first-phase insulin secretion is more in human islets in mouse islets, in human islets is to the first-phase insulin secretion (Arrojo e Drigo et al. 2015). However, dispersion of human islets to or or culture of intact islets first-phase insulin secretion. Thus, preservation of first-phase GSIS in pseudoislets an over both dispersed and cultured-intact islets. While study on dynamism of GSIS, Yu et al. that human pseudoislets made by of dispersed islet an by as insulin secretion by cell highlighting the of pseudoislets over dispersed islet cells for the efficiency of insulin secretion et al. 2018). of human pseudoislets to KCl and glibenclamide was also similar to that of fresh islets the of human pseudoislets to the regulation of insulin secretion in human islets. To gene expression and insulin secretion data by using pseudoislets with reduced compared with fresh and cultured-intact islets, we created pseudoislets with 3000 islet cells and obtained pseudoislets of in first-phase GSIS of pseudoislets be by of pseudoislets. et al. used cells pseudoislets and correlation between GSIS and of pseudoislets. to form pseudoislets, as (Zuellig et al. a low well et al. and microwell et al. function of pseudoislets. perifusion was not a microwell that available was reported to allow the formation of human pseudoislets with GSIS using a similar simple as et al. and first-phase GSIS compared with a low well we used in the study. cell types, and markers were expressed in pseudoislets, at similar as fresh islets. with studies that in cell and distribution (Zuellig et al. 2017; Yu et al. 2018), human pseudoislets the original intact islets well for both and mouse and cells were alpha and cells were to form beta cells are critical for the formation of pseudoislets and Piston 2017). it to be determined whether beta cell formation of pseudoislets in data that the formation of pseudoislets in or of cell to be for both and While expression of beta cell markers and beta cell function markers and in pseudoislets compared with fresh islets of beta cell in pseudoislets, we did not in these markers between cultured-intact and pseudoislets. Thus, the first-phase GSIS in pseudoislets compared with cultured-intact islets be by better preservation of in pseudoislets. we that expression was while expression was reduced in pseudoislets compared with cultured-intact islets of unique to pseudoislets. The loss of during islet is to viability and function of human islets and Wehrle-Haller 2017). The of as or cells functional mass of cultured human islets the of in islet (Arrojo e Drigo et al. 2015; Arzouni et al. 2018). While the of in expression of these genes on the of in pseudoislets to be pseudoislets to maintain insulin secretion over culture. a expressed highly in and in pancreatic et al. 2017). While on pancreatic islets is was of expressed genes in of mouse islets after high et al. et al. 2014). with the reduced expression of proinflammatory (IL1B, CCL2, and in pseudoislets compared with fresh islets, and reduced to beta cell function in pseudoislets. As for cultured intact islets, expression was both fresh and pseudoislets. the in expression is with the of secretion in cultured-intact islets, the of GSIS by to reduced first phase GSIS in intact islets. expression of beta cell markers and genes with beta cell function in cultured-intact islets was that a study reported reduction in and with when human islets cultured for were compared with fresh islets captured by et al. be due to that the islet is with a of cells that culture (Paraskevas et al. 2000; Arzouni et al. 2017). functional heterogeneity among beta more and and 2017; and Hodson 2018). be to determine whether more beta cells are during culture of human islets. the fresh islets we used were cultured for to RNA which markers compared with islets obtained by used in the study et al. Our study has a of qPCR the of of cell markers including low or single cell is to of and pseudoislets. of and proinflammatory cytokine gene expression by qPCR need at before their for the preservation of function of pseudoislets is As we did not gene expression be additional genes in pseudoislets compared with fresh and cultured-intact islets that to be we that human pseudoislets created using a simple protocol maintain functional and of the original islets, first-phase GSIS after culture, and as an efficient gene transduction to test gene function in human islets in culture. of the of to

Related Papers

• → Construction of shRNA lentiviral vector(2010)19 cited
• Research on the Effect of SOCS-3 RNA Interference Mediated by Lentiviral Vector in Adipocytes with Leptin Treatment(2010)
• Effect of lentivirus expressing TLR4 shRNA on function of rat alveolar macrophage(2010)
• Construction and package of recombinant lentivirus expressing shRNA targeting TIPE2 gene(2014)
• → Supplementary Figure 5 from Inhibition of BET Bromodomain Targets Genetically Diverse Glioblastoma(2023)