JOURNAL OF TISSUE ENGINEERING AND REGENERATIVE MEDICINE RESEARCH J Tissue Eng Regen Med (2014) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/term.1891

ARTICLE

Improved physiological properties of gravityenforced reassembled rat and human pancreatic pseudo-islets R. A. Zuellig1†, G. Cavallari2†, P. Gerber1, O. Tschopp1, G. A. Spinas1, W. Moritz3 and R. Lehmann1* 1

Division of Endocrinology, Diabetes and Clinical Nutrition, University Hospital Zurich, Switzerland Nephrology, Dialysis and Transplantation Unit (Stefoni), S.Orsola-Malpighi Hospital, University of Bologna, Italy 3 InSphero AG, Schlieren, Switzerland 2

Abstract Previously we demonstrated the superiority of small islets vs large islets in terms of function and survival after transplantation, and we generated reaggregated rat islets (pseudo-islets) of standardized small dimensions by the hanging-drop culture method (HDCM). The aim of this study was to generate human pseudo-islets by HDCM and to evaluate and compare the physiological properties of rat and human pseudo-islets. Isolated rat and human islets were dissociated into single cells and incubated for 6–14 days by HDCM. Newly formed pseudo-islets were analysed for dimensions, morphology, glucose-stimulated insulin secretion (GSIS) and total insulin content. The morphology of reaggregated human islets was similar to that of native islets, while rat pseudo-islets had a reduced content of α and δ cells. GSIS of small rat and human pseudo-islets (250 cells) was increased up to 4.0-fold (p < 0.01) and 2.5-fold (p < 0.001), respectively, when compared to their native counterparts. Human pseudo-islets showed a more pronounced first-phase insulin secretion as compared to intact islets. GSIS was inversely correlated to islet size, and small islets (250 cells) contained up to six-fold more insulin/cell than large islets (1500 cells). Tissue loss with this new technology could be reduced to 49.2 ± 1.5% in rat islets, as compared to the starting amount. With HDCM, pseudo-islets of standardized size with similar cellular composition and improved biological function can be generated, which compensates for tissue loss during production. Transplantation of small pseudo-islets may represent an attractive strategy to improve graft survival and function, due to better oxygen and nutrient supply during the phase of revascularization. Copyright © 2014 John Wiley & Sons, Ltd. Received 19 July 2013; Revised 20 December 2013; Accepted 26 February 2014

Keywords

pseudo-islets; hanging drop; diabetes; reaggregation; islet cell transplantation; survival

1. Introduction Islet transplantation has been shown to successfully restore endogenous insulin production and glycaemic stability without severe hypoglycaemia and, therefore, represents a therapeutic alternative to standard insulin treatment or pancreas transplantation for selected patients with type 1 diabetes mellitus (Lehmann et al., 2008). In spite of the different improvements described

*Correspondence to: Roger Lehmann, Division of Endocrinology, Diabetes and Clinical Nutrition, University Hospital Zurich, Switzerland. E-mail: [email protected] † These authors contributed equally to this study. Copyright © 2014 John Wiley & Sons, Ltd.

in the ‘Edmonton protocol’ (Shapiro et al., 2000), the transplanted islet volume is still considered a crucial determinant for the prediction of graft function (Korsgren et al., 2005). The assessment of islet volume is based on mathematical conversion to standard islet equivalents (IEQ), with a diameter of 150 μm. According to the Edmonton experience, ~12 000 IEQ/kg recipient body weight are required to achieve insulin independence (Shapiro et al., 2000). Besides allograft rejection, several factors have been described that lead to poor engraftment and primary graft non-function and therefore contribute to the high demand of donor islets (Bennet et al., 1999; Carlsson et al., 2001; Heuser et al., 2000; Reffet and Thivolet, 2006). Among these, hypoxia is considered to play a crucial role. Pancreatic islets have a dense

R. A. Zuellig et al.

glomerular-like capillary network, which entails a high blood perfusion, resulting in a significantly higher oxygen tension (Carlsson et al., 2001). However, during the process of isolation and in vitro culture, islet vasculature degenerates. Immediately after transplantation pancreatic islets are supplied with oxygen and nutrients solely by diffusion, which penalizes islets with larger dimensions. Moreover, the oxygen tension in the portal blood stream is only 3–5 mmHg, while initial islet revascularization in the liver parenchyma requires 7–10 days (Jones et al., 2007; Menger et al., 1989). Inevitably, these critical near-anoxic conditions lead to apoptosis/necrosis in the core regions of larger islets, resulting in early graft loss (Giuliani et al., 2005). In an elegant study, MacGregor et al. (2006) demonstrated that if a marginal islet volume composed of large islets was transplanted into diabetic rats, all recipients failed to achieve normoglycaemia, while 80% of the transplantations using small islets were successful (MacGregor et al., 2006). In vitro, human islets of small dimensions (diameter < 150 μm) showed an improved glucose-stimulated insulin secretion in perifusion assays when compared to large islets (diameter 150 μm) (Lehmann et al., 2007). Additionally, large islets, after transplantation in animal models, exhibit a reduced capacity to produce angiogenic factors and stimulate regrowth of blood vessels (Kampf et al., 2006) and a higher tendency to cause microthrombosis and release of proinflammatory hypoxia-dependent cytokines (Su et al., 2010). Thus, small islets seem to be more favourable for clinical islet transplantation than large islets, due to a superior secretory function and diminished vulnerability to hypoxic conditions (Lehmann et al., 2007, 2008). It has long been known that isolated islets can be dissociated into single cells, which spontaneously reaggregate into islet-like clusters, so-called ‘pseudo-islets’. Such pseudo-islets have been described to be comparable to intact islets in terms of cellular composition, architecture and secretory function, while their immunogenicity in the transplant setting seems to be attenuated, but greater islet loss was observed during production (Halban et al., 1987; King et al., 2007; Matta et al., 1994; Shizuru et al., 1985; Wolf-Jochim et al., 1995). Formation of cellular aggregates can also be achieved by the ‘hanging-drop’ culture method, which was described previously for a variety of systems, e.g. for the study of embryo development (Potter and Morris, 1985), embryonic stem cell differentiation (Yoon et al., 2006) or tumour biology (Kelm et al., 2003). In a previous short meeting report we described the possibility of applying the hanging-drop technology to cultures of single rat islet cell suspensions, thereby generating pseudo-islets of standardized size and regular shape (Cavallari et al., 2007). The aims of the present study were: (a) to generate human pseudo-islets with defined dimensions by the hanging-drop culture method; (b) to compare the insulin secretion of human pseudo-islets of different sizes; (c) to test whether small pseudo-islets secrete more insulin; and (d) to evaluate and compare the cellular composition and in vitro functional capacity of human with rat pseudo-islets. Copyright © 2014 John Wiley & Sons, Ltd.

2. Materials and methods 2.1. Enzymes and substances Collagenase NB8 was purchased from SERVA Electrophoresis (Heidelberg, Germany) and DNAse was obtained from Roche Molecular Biochemicals (Basel Switzerland). Cell culture media, HBSS, RPMI 1640, CMRL 1066, Glutamax, sodium pyruvate and antibiotic–antimycotic solution were all obtained from Life Technologies, Gibco BRL. Insulin was measured by radioimmunoassays (Insulin-CT, CIS Bio International, Schering AG, Baar, Switzerland). The intra- and interassay variations for insulin were 5.9% and 7.9%, respectively.

2.2. Isolation of rat islets All procedures and experimental protocols were performed in accordance with the Swiss animal protection laws. Male Lewis rats (weight 200–300 g; Harlan, The Netherlands) were anaesthetized with Isoflurane (Baxter, Switzerland) and subsequently sacrificed by neck dislocation. Islet isolation and purification was performed according to a modified procedure described earlier (Gotoh et al., 1985). In brief, after cannulation of the common bile duct, 10 ml enzyme solution (collagenase NB8 1.3 U/ml and DNAse 0.1 mg/ml in HBSS) was instilled by retrograde perfusion of the pancreas. After organ procurement, enzymatic digestion took place for 12–15 min at 37 °C in a water bath, with two intermitted 1 min periods of vigorous shaking. Following a wash step with 10% FCS/HBSS and filtration through cheesecloth, islets were obtained by centrifugation in a discontinuous density gradient of Histopaque (Sigma) and HBSS, with subsequent hand-picking. Islet purity was generally 95%. Isolated rat islets were cultured in RPMI 1640 supplemented with 1 mM sodium pyruvate, 10% FCS, 1% Glutamax (200 mM) and 1% antibiotic–antimycotic solution (10 000 U/ml penicillin, 10 mg/ml streptomycin and 250 μg/ml amphotericin B as fungizone in 0.85% saline) in a humidified incubator at 37 °C, 5% CO2/95% air, unless specified otherwise.

2.3. Human islets Human islets were obtained by the ‘islet distribution service’ from the Juvenile Diabetes Research Fund (JDRF) and the Islets for Research Distribution Programme of the European Consortium for Islet Transplantation (ECIT). The providing centres were the Cell Isolation and Transplantation Centre, Geneva University Hospitals; the Islet Processing Facility, S. Raffaele Scientific Institute Milan; and the Third Medical Department, University of Giessen. The islets are used in accordance with the ethical requirements. J Tissue Eng Regen Med (2014) DOI: 10.1002/term

Improved pseudo-islets

2.4. Single cell preparation Immediately after isolation or upon receipt, rat or human islets were washed three times with ice-cold Ca-free phosphate-buffered saline (PBS) and thereafter incubated for 30 min on ice. Islets were then dissociated into single cells with 0.017% trypsin (diluted into cell dissociation solution, Sigma) by pipetting up and down through a siliconized Pasteur pipette for 4–6 min. A trypsin concentration of 0.017% proved to be optimal in terms of dissociation efficiency without considerably affecting cell viability, which could be maintained at > 93%. The islet cell suspension was washed three times with hanging-drop medium (for rat islets, RPMI 1640, 11 mM glucose, supplemented with 20% FCS, 1 mM sodium pyruvate, 2 mM Glutamax, 100 μg/ml streptomycin and 100 U/ml penicillin; for human islets, CMRL 1066 medium, 5.5 mM glucose containing 20% FCS, 100 μg/ml streptomycin, 100 U/ml penicillin, 1 mM sodium pyruvate, 2 mM Glutamax and 2 mM HEPES). The cells were counted and viability was assessed by trypan blue exclusion test.

perifused with Krebs–Ringer–HEPES buffer (131 mM NaCl, 4.8 mM KCl, 1.2 mM CaCl2
2H2O, 25 mM HEPES, 1.2 mM KH2PO4 and 1.2 mM MgSO4
7H2O, pH 7.4) containing 0.1% bovine serum albumin (BSA; Sigma) and glucose at a flow rate of 100 μl/min, and samples of 300 μl were collected. During the first 114 min, the islets were perifused with 2.8 mM (human islets) or 3.3 mM (rat islets) glucose, from 115 to 174 min with 20 mM (human islets) or 16.7 mM glucose (rat islets), and from 175 to 210 min with 2.8 mM (human islets) or 3.3 mM (rat islets) glucose. Insulin secretion was assessed as the area under the curve (AUC) of the insulin secretion profile, with the first phase in the range 115–123 min and the second in the range 124–174 min. The AUC was computed using the GraphPad Prism program (GraphPad Software, San Diego, CA, USA). The amount of insulin secreted was normalized for IEQ in order to adjust for differences in total islet volume. The insulin secretion stimulation index was calculated by dividing the amount of stimulated insulin by the amount of insulin produced at basal conditions.

2.8. Immunohistochemistry 2.5. Hanging-drop culture Dissociated cells were taken up in hanging-drop medium, as described above, at densities of 250–1500 cells/30 μl medium, and drops of 30 μl were distributed onto the lids of large tissue-culture dishes. The lids were inverted and placed on dishes prefilled with 1% antibiotic–antimycotic solution (10 000 U/ml penicillin, 10 mg/ml streptomycin and 250 μg/ml amphotericin B as fungizone in 0.85% saline) diluted into PBS. Hanging-drop cultures were incubated in a humidified incubator at 37 °C, 5% CO2/95% air for up to 14 days.

2.6. Islet morphometry Intact or reaggregated islets were harvested into small cell-culture dishes with HBSS. Pictures were recorded with a Zeiss AxioCam MRc camera connected to a Zeiss Axiolab microscope (Carl Zeiss AG, Feldbach, Switzerland). Islet numbers and dimensions were assessed using NIH ImageJ software. Volumes were calculated based on average diameter, as determined by area measurements, assuming a perfect spherical shape. Islets volumes were expressed as multiples of an islet equivalent (IEQ), corresponding to a hypothetical sphere with a diameter of 150 μm.

2.7. Islet perifusion assay The technique of perifusion has been described previously (Spinas et al., 1998). In short, ca. 100 intact islets or reaggregated pseudo-islets were placed in a chamber of 300 μl, embedded in Sephadex G-10 (Pharmacia, Uppsala, Sweden), and mounted in a perifusion device, preheated to 37 °C (Superafusion 1000, Model SF-06; Brandel, Gaithersburg, MD, USA). The islets were Copyright © 2014 John Wiley & Sons, Ltd.

Intact or reaggregated islets were fixed in Bouin’s solution for 20 min, either directly after isolation or after 1 to several days in culture. The fixed islets were embedded in 2% agarose in PBS. Intact pancreatic tissue was fixed in formalin overnight. Tissue specimens were dehydrated and paraffin-embedded in a tissue processor (Leica TP1020, Leica Microsystems, Glattbrugg, Switzerland); 3 μm sections were cut on a microtome (Leica RM2255). Sections were dewaxed in xylol and rehydrated in a descending series of ethanols (99%, 95%, 90%, 80%, 70%, water) and stained with haematoxylin and eosin (H&E). For immunohistochemical staining, the sections were permeabilized for 5 min in methanol and blocked for 30 min at room temperature (RT) with 10% normal goat serum (NGS) in PBS. The sections were incubated with the first antibody, diluted from 1:100 to 1:1000 in (PBS/1% NGS), according to the antibody, for 1 h at 37 °C, and with the second antibody, diluted from 1:200 to 1:400 in (PBS, 1% NGS), for 1 h at RT. Antibodies used were guinea-pig anti-insulin (dilution 1:100; A0564, Dako Cytomation, Denmark); mouse anti-glucagon (dilution 1:1000; Sigma, G-2654); rabbit anti-somatostatin (dilution 1:100; A0566, Dako Cytomation, Denmark). Secondary antibodies used were donkey anti-guinea-pig FITC (dilution 1:200; 706-095-148); donkey anti-rabbit Cy5 (dilution 1:400; 711-175-152) and donkey anti-mouse Cy3 (dilution 1:400; 715-165-150; all Jackson ImmunoResearch Laboratories, West Grove, PA, USA). The slides were mounted in Hydromount (National Diagnostics, Atlanta, GA, USA) and studied with a Zeiss Axioplan 2 microscope equipped with a Zeiss AxioCam HRc camera and the imaging acquisition software AxioVision 3.1 (Carl Zeiss AG, Feldbach, Switzerland). Areas for positive insulin (FITC, green), glucagon (Cy3, red) and somatostatin (Cy5, blue) staining were assessed with NIH ImageJ software upon splitting into RGB J Tissue Eng Regen Med (2014) DOI: 10.1002/term

R. A. Zuellig et al.

colours, thresholding and black/white conversion. Areas of individual hormone staining were expressed as a percentage of total hormone staining, including insulin, glucagon and somatostatin. For each condition, at least 100 islets were analysed.

2.9. Statistical analysis Data are presented as mean ± standard deviation (SD). Statistical comparison of two groups was carried out by twosided Student’s t-test for independent samples (GraphPad Prism). Comparison of more than two groups was performed by one-way ANOVA followed by Bonferroni post hoc test. Values of p < 0.05 were considered to be significant.

3. Results 3.1. Islets size distribution and secretory capacity of isolated rat and human islets Size and volume distribution of native rat and human islets are shown in Figure 1A, B, respectively. Islet size is

more or less normally distributed around a diameter of 150 μm (Figure 1A, B). Small islets (< 150 μm) make up around two-thirds of the total number of islets, but conversion to IEQ reveals that they contribute only one-fifth to one-third of the total islet volume. The opposite relationship is true for islets > 150 μm in diameter, making up one-third of the total number but about two-thirds to four-fifths of the total islet volume (Figure 1C, D). To evaluate whether islet size is affecting glucosestimulated insulin secretion, we separated isolated rat islets, based on their size, into small (~0.15 IEQ; ~80 μm), medium (~0.5 IEQ, ~120 μm) and large (>1 IEQ, >150 μm) fractions. Islets were selected from the same isolation after either 1 or 6 days of cultivation under regular cell culture conditions. As depicted in Figure 1E, insulin secretion of size-fractionated islets, upon stimulation with 16.7 mM glucose, follows an exponential relationship dependent on the average islet size. Small islets (~0.15 IEQ, ~80 μm) secrete up to 2.5-fold more insulin than islets of mixed sizes (> 0.5 IEQ). Of note, insulin secretion capacities were comparable among size-matched islets, irrespective of the culture time up to 8 days.

Figure 1. Islet size distribution and secretory function. Morphometric analysis of islet diameters and volumes (IEQ) of representative samples from three different rat (A) and human (B) islet isolations are shown. One IEQ corresponds to an islet 150 μm in diameter. (C, D) Contribution of small islets (diameter < 150 μm, or IEQ < 1, white bar) and large islets (diameter ≥ 150 μm, or IEQ ≥ 1, grey bar) in percentage of total number or total volume are presented for rat (C) and human (D) islets. (E) Glucose (16.7 mM)-stimulated insulin secretion of sizefractionated rat islets after 1 or 6 days of culture. Insulin secretion is normalized for islet volume and size ranges of individual fractions are indicated as mean ± SD Copyright © 2014 John Wiley & Sons, Ltd.

J Tissue Eng Regen Med (2014) DOI: 10.1002/term

Improved pseudo-islets

3.2. Pseudo-islets formation by hanging-drop technology Reassociation of islet cells to spheroid clusters was accomplished after 4–6 days (Figure 2). The diameter of reaggregated rat pseudo-islets composed of 250 cells was 86.0 ± 6.1 μm (mean ± SD), 131.1 ± 10.6 μm for islets with 750 cells and 176.7 ± 16.9 μm for islets with 1500 cells, compared to 143.5 ± 42.8 μm of native intact islets (Figure 3A). The 25th–75th percentile of the diameter of pseudo-islets is in the range 81.8–89.8 μm for 250 cells, 125.6–138.2 μm for 750 cells and 169.0–187.6 μm for 1500 cells, compared to 111.6–167.7 μm of native intact islets. The cell number seeded/droplet correlated strongly with the volume (IEQ) of the resulting pseudo-islets (r2 = 0.994). The diameter of reaggregated human pseudo-islets composed of 250 cells was 95.4 ± 11.5 μm (mean ± SD), 137.2 ± 12.7 μm for islets with 750 cells and 182.0 ± 15.1 μm for islets with 1500 cells, compared to 138.2 ± 53.13 μm of native intact islets (Figure 3B). The 25th–75th percentile of the diameter of pseudo-islets is in the range 88.8–100.0 μm for 250 cells, 131.0–143.6 μm for 750 cells and 174.0–188.7 μm for 1500 cells, compared to 93.5–173.6 μm of native intact islets. The cell number

seeded/droplet correlated strongly with the volume (IEQ) of the resulting pseudo-islets (r2 = 0.994). Thus, the size and homogeneity of human pseudo-islets were almost identical to those of rat pseudo-islets.

3.3. Tissue loss in the production of pseudo-islets In order to assess tissue loss along the pseudo-islet production process, pictures taken from rat isolates prior to dissociation/reaggregation and after pseudo-islet formation during a 6 day culture period were analysed by computer-assisted morphometry. The recovery of islet volume was 50.8 ± 1.5% (n = 4). For human islets the variability was much larger, with a recovery of 18–67% (n = 4).

3.4. Cellular and morphological characteristics of non-dissociated intact and dissociated/ reaggregated pseudo-islets In H&E-stained sections, isolated intact rat or human islets occasionally reveal pycnotic nuclei or even central necrosis, predominantly in large islets, even after only

Figure 2. Reaggregation sequence of dissociated islets in hanging-drop cultures. Phase-contrast images of dissociated rat islet cells (upper row) and human islets cells (lower row) at various time points after seeding into hanging drops (250 cells/drop). Bar = 50 μm

Figure 3. Standardization of islet size. Size distribution of isolated rat (A) and human (B) intact islets and pseudo-islets composed of 250, 750 and 1500 cells after 6 days of hanging-drop culture. Boxes denote median with 25th and 75th percentiles and whiskers denote lowest and highest values Copyright © 2014 John Wiley & Sons, Ltd.

J Tissue Eng Regen Med (2014) DOI: 10.1002/term

R. A. Zuellig et al.

1 day of culture (Figure 4B, E). Pseudo-islets derived from hanging-drop cultures of dissociated rat or human islet cells are more regular in size and do not exhibit morphological signs of apoptosis/necrosis, despite a culture period of 6 days (Figure 4C, F). Immunostaining for insulin, glucagon and somatostatin revealed similar distribution patterns in pancreatic sections and pseudo-islets formed out of single cell suspensions (Figure 4G–J), with glucagon-secreting α cells and somatostatin-secreting δ cells located at the periphery of rat islets. Due to the more lobular structure of human islets, α and δ cells appear more

intermingled with centrally located β cells, a cellular organization which is also maintained in human pseudo-islets. The relative abundance of either cell type in pancreatic sections, freshly isolated islets and pseudoislets has been carefully assessed and is summarized in Table 1. Compared to native pancreatic rat islets, the relative areas of α and δ cells in rat pseudo-islets were largely reduced. As opposed to rat islets, human pseudoislets retained their cellular composition, with the exception of a slightly higher proportion of glucagonproducing α cells.

Figure 4. Islets architecture and cellular composition. H&E-stained paraffin sections of rat (A–C) or human (D–F) pancreatic islets in situ (A, D), islets 1 day after isolation (B, E) and after 6 days of reaggregation into pseudo-islets (C, F). Necrotic cores are observed predominantly in large islets (arrow), but also scattered cells with pycnotic nuclei can be identified (arrowheads). Immunofluorescent staining of insulin (green), glucagon (red) and somatostatin (blue) in rat (G, H) or human (I, J) pancreatic sections (G, I) and reaggregated pseudo-islets (H, J). Bar = 50 μm Copyright © 2014 John Wiley & Sons, Ltd.

J Tissue Eng Regen Med (2014) DOI: 10.1002/term

Improved pseudo-islets Table 1. Relative distribution of hormone-expressing cells Pancreas %

Isolated islets %

p

Pseudo-islets %

RAT Insulin Glucagon Somatostatin

82.1 ± 7.9 14.3 ± 7.2 3.6 ± 2.1

HUMAN Insulin Glucagon Somatostatin

77.2 ± 11.0 75.0 ± 9.2 ns 75.1 ± 7.2 16.0 ± 10.0 18.5 ± 8.5 ns 19.5 ± 5.9 6.8 ± 4.1 6.5 ± 2.8 ns 5.4 ± 2.7

p

87.4 ± 5.3 ns 95.9 ± 1.3 < 0.001 9.6 ± 4.7 ns 2.6 ± 0.9 < 0.001 3.0 ± 1.5 ns 1.5 ± 0.9 < 0.001 ns < 0.01 ns

Relative abundance of hormone-expressing cells indicated as percentage of total hormone-positive area in paraffin sections stained for insulin, glucagon and somatostatin. Data are mean ± SD from at least 100 islets originating from three different islet isolations. ns, not significant. All comparisons vs ‘pancreas’.

3.5. Assessment of insulin secretion by perifusion experiments Intact rat islets from three independent isolations (average IEQ 0.63 ± 0.42) and pseudo-islets from three independent reaggregation experiments, containing 250 cells/drop (average IEQ 0.33 ± 0.04) were placed in perifusion chambers and perifused with 3.3 mM glucose prior to stimulation with 16.7 mM glucose. Insulin secretion adjusted for islet volume was comparable at baseline, but strongly elevated in small pseudo-islets when compared to intact islets after

stimulation with high glucose (Figure 5A). However, neither the intact islets nor the pseudo-islets revealed the classical secretion profile, since the first-phase peak of insulin secretion was basically absent. The AUC at 16.7 mM glucose was increased by 2.9-fold (p = 0.003). The resulting stimulation factor for small pseudo-islets was 8.76 ± 2.68 vs 2.83 ± 0.33 (p = 0.02) for intact control islets. The reduced secretory capacity of intact islets was independent of culture time, with 1 day-old islets performing equally to 6 day-old islets. To rule out that improved secretory function of pseudoislets resulted from the differences in average islet sizes, we used hanging-drop cultures with either 250 or 600 cells/drop, which generated islet populations of two different sizes (IEQ 0.33 ± 0.11 vs 0.77 ± 0.13), the latter representing an average islet size which compares to the isolation index of a regular rat islet isolate, which is usually IEQ 0.6–1.0 and in this particular experiment was IEQ 0.68 ± 0.67. In perifusion experiments, baseline secretion was similar in all three groups. Insulin secretion in intact fresh islets upon stimulation with 16.7 mM glucose was enhanced by 2.76 ± 0.63-fold, while the stimulation factors for small (< 0.5 IEQ, < 120 μm) and large (> 0.5 IEQ, > 120 μm) pseudo-islets were 10.58 ± 4.22 and 7.63 ± 3.75, respectively (Figure 5B). Glucose-stimulated insulin secretion was inversely correlated to the size of rat pseudo-islets (Figure 5C), with small pseudo-islets releasing about twice the amount of insulin per IEQ than large pseudo-islets (p = 0.0002).

Figure 5. Glucose-stimulated insulin secretion of intact rat islets and reaggregated pseudo-islets. (A) Insulin secretion profiles from perifusion assays of intact rat islets (○) and reaggregated 250-cell rat pseudo-islets (●) at low (3.3 mM) and high (16.7 mM) glucose. Data are normalized to total islet volume. Values are mean ± SD from three independent isolations. (B) Data from perifusion experiments, indicating the stimulation index of basal vs stimulated insulin secretion from fresh intact, and small (< 0.5 IEQ) and large (> 0.5 IEQ) rat pseudo-islets after 6 days of culture. Data are mean ± SD from 10 independent isolations. (C) Relationship of pseudo-islet size and glucose-stimulated insulin secretion. Pseudo-islet preparations of different sizes, comprising 200–750 cells, were subjected to perifusion assays and glucose (16.7 mM)-stimulated insulin secretion is plotted against the average islet size; *p < 0.001, **p < 0.01 Copyright © 2014 John Wiley & Sons, Ltd.

J Tissue Eng Regen Med (2014) DOI: 10.1002/term

R. A. Zuellig et al.

Unlike rat islets, freshly isolated human islets or reaggregated human pseudo-islets revealed a more physiological secretion pattern, with a distinct first-phase insulin release. However, while extended culture of intact islets did not affect the secretion capacity, the secretory properties of human pseudo-islets improved markedly over time (Figure 6). Particularly small pseudo-islets consisting of 250 cells (IEQ 0.23), cultured for 14 days, were superior to medium (750 cells, IEQ 0.63) and large (1500 cells, IEQ 1.40) pseudo-islets, as well as to a mixture of intact islets with respect to a pronounced first-phase insulin secretion and overall secretory capacity (Figure 6B). Analysis of the area under the curve (AUC) for firstand second-phase insulin secretion demonstrates the quantitative differences observed in reaggregated pseudo-islets compared to intact islets (Figure 7). Despite the lack of a pronounced first-phase insulin secretion, rat pseudo-islets secreted 4.1-fold (p < 0.01) more insulin compared to freshly isolated islets, which did not further improve beyond 6 days of culture. Small human pseudo-islets (250 cells/islet) showed an increase in first- and second-phase insulin release after a culture period of 14 days, resulting in a significant overall enhancement of insulin secretion by a factor of 2.5 (p < 0.001) as compared to freshly isolated islets. This improved secretion capacity was statistically significant only for pseudo-islets cultured for the period of 14 days. While reaggregated pseudo-islets performed best after 14 days of culture, the insulin secretion of isolated intact islets was largely unaffected by time.

3.6. Relationship of islet size and cellular insulin content To test whether differences in insulin secretion could be related to the cellular insulin content, values of total insulin were plotted against the average IEQ of different populations of non-dissociated intact human islets and dissociated/reaggregated human pseudo-islets of various sizes (Figure 8A). Interestingly, both native and pseudoislets followed a curvilinear relationship, predicting the cellular insulin content in relation to islet size with a probability of 85% and 72%, respectively. Measurements of total insulin revealed that intact small rat and human islets contained up to six-fold more insulin/cell than intact large islets (Figure 8B, C). Likewise, the cellular insulin content of small rat and human pseudo-islets (250 cells) was increased three- to five-fold (4.8-fold rat, 2.8-fold human) when compared to their large counterparts (1500 cells). Hence, insulin secretion from intact islets or reaggregated pseudo-islets roughly matched cellular insulin content, which was inversely correlated to islet size.

4. Discussion The present data demonstrate that it is possible to generate reaggregated islets (so-called pseudo-islets) with improved physiological properties by the hanging-drop culture method.

Figure 6. Glucose-stimulated insulin secretion profile of intact human islets and reaggregated human pseudo-islets. Insulin secretion profiles from perifusion assays of intact human islets (A) and reaggregated human pseudo-islets obtained from 250 cells (B), 750 cells (C) and 1500 cells (D). Islets were measured either fresh (♦) or after 7 (●) and 14 days (○) of culture. Glucose concentrations applied for basal and stimulated insulin secretion are indicated, representative for all panels. Data are normalized to total islet volume; mean values are presented from four independent isolations Copyright © 2014 John Wiley & Sons, Ltd.

J Tissue Eng Regen Med (2014) DOI: 10.1002/term

Improved pseudo-islets

Figure 7. Secretory capacity of glucose-stimulated insulin secretion in intact and reaggregated rat and human pseudo-islets. The area under the curve (AUC) from insulin secretion profiles are indicated each for first phase (A), second phase (B) and the total (C) released insulin from either intact islets or pseudo-islets composed of 250 cells. Data are mean ± SD from three independent isolations of rat islets (grey bars) and four independent isolations of human islets (white bars); *p < 0.05, **p < 0.01, ***p < 0.001

Figure 8. Cellular insulin content of intact and reaggregated rat and human pseudo-islets. (A) Relationship of size and insulin content in human intact (●, —) and reaggregated (○, ) islets is described by an inverse relationship. Total insulin content expressed per cell from intact and reaggregated rat (B) and human (C) islets. Data are mean ± SD from four independent isolations; *p < 0.001 vs # small, p < 0.001 vs 250 cells

The size of these reaggregated clusters can be predetermined by the number of dispersed islet cells seeded per drop, which allows reaggregation into pseudo-islets yielding an optimal size for transplant purposes. Pseudo-islets obtained by hanging-drop technology show a cellular composition and architecture that is comparable to intact islets, but they have a more robust insulin-secretory response upon stimulation with glucose. The success of islet transplantation as an alternative therapeutic strategy to whole-organ transplantation for the treatment of type 1 diabetes mellitus is hampered by a dramatic islet loss during the implantation and revascularization period (Davalli et al., 1995), with a functional capacity of only ~ 30% (Ryan et al., 2001) and a gradual Copyright © 2014 John Wiley & Sons, Ltd.

loss of graft function, with up to ~80% of patients resuming insulin treatment 3 years after islet transplantation (Shapiro et al., 2006). The low functional capacity could be due to initial islet loss or a reduced function, due to a 10-fold lower oxygen supply in revascularized islets in the portal field of the liver as compared to the native islets in the pancreas (Carlsson et al., 2001). It is conceivable that graft survival and function could be enhanced by transplanting predominantly small islets, which have an advantage during the time of revascularization, when nutrient and oxygen supply depend on diffusion only, until the time when revascularization is re-established. Indeed, we were able to demonstrate that C-peptide concentrations and clinical outcome in patients receiving an J Tissue Eng Regen Med (2014) DOI: 10.1002/term

R. A. Zuellig et al.

islet transplant were dependent on the size and number of transplanted islets rather than the total graft volume as expressed by IEQ (Cavallari et al., 2007; Lehmann et al., 2008). This suggests that, with a given total islet mass, small islets are superior to large islets with respect to the restoration of normoglycaemia. The ability of single islet cells to re-associate into isletlike clusters with similar cellular composition and biological function to native or freshly isolated pancreatic islets has been described previously (Halban et al., 1987; Matta et al., 1994). In order to accelerate and standardize the reaggregation process, we used the hanging-drop culture method, which was initially described for the study of early embryonic development. In hanging-drop cultures, gravity facilitates the sedimentation and formation of well-defined spherical clusters. Located at the interface of the medium and the surrounding gas atmosphere, forming pseudo-islets are optimally supplied with oxygen without any interference from plastic surfaces. By using this technology as opposed to static cultures or cultures in rotating devices, we could show that the size of islets generated in hanging drops is predetermined by the number of cells seeded/drop, and therefore much more uniform, with a SD of < 10% of the average diameter. Tanaka et al. (2013), for example, obtained in a simulated microgravity culture system pseudo-islets of 180–270 μm diameter (smallest group). Moreover, hanging-drop cultures prevent fusion of reaggregated pseudo-islets, as occurs in conventional culture devices. Loss of islet tissue during the dissociation/reaggregation procedure is of particular concern. In our hands, a combination of mechanical shear forces and mild enzymatic digestion proved to be the most efficient and least harmful way to obtain a single-cell suspension with good viability. We experienced a higher yield with rat as compared to human pseudo-islets. Three different users obtained very similar, reproducible yields with rat pseudo-islets, but very variable yields with human pseudo-islets. A possible explanation for this is that rat pancreata originate from healthy animals of the same strain and age. In addition, the isolated islets are kept under optimal culture conditions and used within 1 day. The human islets, however, are from different centres, donor quality usually is lower (since good islets are transplanted) and the islets are shipped for 1–2 days under suboptimal conditions. Therefore, it is not surprising that human pseudo-islets have a much higher variability. In spite of this, the overall tissue recovery from the initial islet isolate to the final pseudo-islet volume was ca. 50%, which is several-fold higher than the 10% described by Callewaert et al. (2007). At first sight, a 50% cell loss may seem rather high; however, this number also includes the loss of damaged islet tissue that usually occurs simply as a consequence of the isolation procedure, leading to a disruption of the normal cell–matrix relationship (Thomas et al., 2001). It has been estimated that 20–40% of initial islet mass is lost post-isolation during in vitro culture, depending on medium supplements and incubation temperature (Cavallari et al., 2007; Fraga et al., 1998; Kim et al., 2005). Therefore, the observed two- to four-fold Copyright © 2014 John Wiley & Sons, Ltd.

improved insulin secretion performance of pseudo-islets will fully compensate for the additional tissue loss of 10–30% which might be caused exclusively by the dissociation/reaggregation procedure. Future efforts should focus on limiting tissue loss and also investigate on the capability to further stimulate islet cell proliferation, potentially by triggering regenerative processes in hanging-drop cultures. In this context, it has been reported that the aggregation yield of human dissociated islet cells was significantly improved by treatment with medium supplemented with amino acids, vitamins and inorganic salts (Tsang et al., 2007), all-trans-retinoic acid (Lee et al., 2008) or, maybe, GLP-1 receptor agonists by reducing apoptosis and maintaining islet integrity (Farilla et al., 2003). Treatment of reaggregating pseudo-islets with β-trophin could be another possibility to increase β-cell mass. Eight days after β-trophin injection in mice, the total pancreatic β-cell mass is three-fold higher than in control mice (Yi et al., 2013). In terms of biological function, both rat and human pseudo-islets displayed an improved glucose-stimulated insulin secretion as compared to their native counterparts. In rat islets a bi-phasic insulin release was barely noticeable, which reflects the current literature, where first-phase insulin secretion from isolated rat islets has generally been described as rather weak (Liang et al., 1990). Human pseudo-islets, on the other hand, showed an accentuation of first-phase insulin release, which was particularly observed after long-term culture of 14 days. A pronounced first-phase insulin release ensures a rapid response to a meal, conferring a profound and long-term inhibitory effect on hepatic glucose production and a sustained prandial glucose tolerance (Caumo and Luzi, 2004). The secretory capacity of pseudo-islets normalized to islet equivalents was inversely correlated to the average islet size, particularly in rat islets. The insulin content of small islets of rat and human origin was up to six-fold higher than large islets, a fact that could account for an improved glucosestimulated insulin secretion. However, it does not provide an explanation for the increased first-phase insulin release, since this was only observed in human pseudo-islets. Similar results were obtained by Ramachandran et al. (2013). They engineered pseudo-islets, called Kanslets, on a glass mould containing microrecesses. These Kanslets had a diameter of 41 μm and a morphology and function comparable to native islets. At this point we can only speculate as to the underlying reasons that lead to a more physiological response of reaggregated islets. Dissociation and subsequent reassociation of isolated pancreatic islets may lead to a removal of inflammatory or dead cells, and the rearrangement of islets cells may promote an adaptation to the non-perfused state of avascular islets. Data from Keymeulen et al. (1996) suggested that long-term graft function is hampered when transplanting small islets obtained by gradient centrifugation. However, small islets selected from an isolate are barely comparable to reaggregated pseudo-islets in terms of integrity, cellular composition and insulin content, and therefore not a reference for the performance of J Tissue Eng Regen Med (2014) DOI: 10.1002/term

Improved pseudo-islets

pseudo-islet grafts. Small islets might also reflect overdigestion with many fragmented islets, which is not the same as native small islets. Rodent models are not ideal to demonstrate the superiority of small pseudo-islets, due to the large necrosis zone induced by intraportal transplantation of islets (Korsgren et al., 2008). Therefore, our data have to be confirmed in a large in vivo model, demonstrating an improved graft function of transplanted small pseudo-islets. Several strategies for the use of pseudo-islets in future clinical islet transplantation could be pursued: either patients could be given small pseudo-islets produced from islets of an entire donor pancreas approximately 14 days after isolation, with the option to apply immunosuppressive induction therapy or novel immunomodulatory strategies prior to transplantation to enhance graft survival; or, alternatively, islets could be transplanted in a two-step procedure, first the fraction consisting of small native islets (< 100 μm) would be sorted by complex object parametric analyser and sorter large-particle flow cytometry (COPAS) (Fernandez et al., 2005) and transplanted immediately, and in a second step pseudo-islets derived from dissociation/reaggregation of the large islet fraction (> 100 μm) would be transplanted by intraportal islet infusion 14 days later. Furthermore, extension of the culture period permitted by pseudo-islets could provide several advantages in the clinical setting of islet transplantation, by allowing additional time for the selection and treatment of the recipient with suitable pretransplant strategies and the performance of the transplantation under ideal immunosuppressive conditions (Shapiro, 2011). In addition, the use of pseudo-islets may facilitate islet shipment to remote transplant centres, increasing the flexibility of the transplant networks and allowing islet transplantation even

in centres at long distance (Kempf et al., 2005; Porrett et al., 2007). Finally, the use of pseudo-islets could benefit the field of research in pretransplant islet manipulation. For instance, it was reported that small aggregates of rat islet cells exhibit better results after transplantation in a barium alginate capsule (O’Sullivan et al., 2010), and that pseudo-islets facilitate lentiviral transduction when compared to intact islets (Callewaert et al., 2007). Another possibility is to precondition or co-cultivate the islets with mesenchymal stem cells, as this treatment improves graft revascularization (Cavallari et al., 2012). In conclusion, the hanging-drop culture method allows the generation of islets that meet the size criteria required for maximal viability and biological function for the purpose of therapeutic islet transplantation, in order to reduce the initial massive loss of islet mass or function observed after islet transplantation that can not be improved by anti-inflammatory treatment. Our findings warrant further exploration into the potential of hanging-drop technology to form small pseudo-islets with much better oxygen diffusion and functional capacity, as compared to normal islets in a large animal model.

Conflict of interest The authors declare no conflicts of interest.

Acknowledgements Funding of this project by the Olga Mayenfisch, Hartmann–Müller and Hans Paul Waelchli Foundations is gratefully acknowledged. We thank Heidi Seiler for her excellent technical support.

References Bennet W, Sundberg B, Groth CG et al. 1999; Incompatibility between human blood and isolated islets of Langerhans: a finding with implications for clinical intraportal islet transplantation? Diabetes 48: 1907–1914. Callewaert H, Gysemans C, Cardozo AK et al. 2007; Cell loss during pseudoislet formation hampers profound improvements in islet lentiviral transduction efficacy for transplantation purposes. Cell Transpl 16: 527–537. Carlsson PO, Palm F, Andersson A et al. 2001; Markedly decreased oxygen tension in transplanted rat pancreatic islets irrespective of the implantation site. Diabetes 50: 489–495. Caumo A, Luzi L. 2004; First-phase insulin secretion: does it exist in real life? Considerations on shape and function. Am J Physiol Endocrinol Metab 287: E371–385. Cavallari G, Olivi E, Bianchi F et al. 2012; Mesenchymal stem cells and islet cotransplantation in diabetic rats: improved islet graft revascularization and function by human adipose tissue-derived stem cells preconditioned with natural molecules. Cell Transpl 21: 2771–2781.

Copyright © 2014 John Wiley & Sons, Ltd.

Cavallari G, Zuellig RA, Lehmann R et al. 2007; Rat pancreatic islet size standardization by the ‘hanging drop’ technique. Transpl Proc 39: 2018–2020. Davalli AM, Ogawa Y, Ricordi C et al. 1995; A selective decrease in the β-cell mass of human islets transplanted into diabetic nude mice. Transplantation 59: 817–820. Farilla L, Bulotta A, Hirshberg B et al. 2003; Glucagon-like peptide 1 inhibits cell apoptosis and improves glucose responsiveness of freshly isolated human islets. Endocrinology 144: 5149–5158. Fernandez LA, Hatch EW, Armann B et al. 2005; Validation of large particle flow cytometry for the analysis and sorting of intact pancreatic islets. Transplantation 80: 729–737. Fraga DW, Sabek O, Hathaway DK et al. 1998; A comparison of media supplement methods for the extended culture of human islet tissue. Transplantation 65: 1060–1066. Giuliani M, Moritz W, Bodmer E et al. 2005; Central necrosis in isolated hypoxic human pancreatic islets: evidence for postisolation ischemia. Cell Transpl 14: 67–76.

Gotoh M, Maki T, Kiyoizumi T et al. 1985; An improved method for isolation of mouse pancreatic islets. Transplantation 40: 437–438. Halban PA, Powers SL, George KL et al. 1987; Spontaneous reassociation of dispersed adult rat pancreatic islet cells into aggregates with three-dimensional architecture typical of native islets. Diabetes 36: 783–790. Heuser M, Wolf B, Vollmar B et al. 2000; Exocrine contamination of isolated islets of Langerhans deteriorates the process of revascularization after free transplantation. Transplantation 69: 756–761. Jones GL, Juszczak MT, Hughes SJ et al. 2007; Time course and quantification of pancreatic islet revascularization following intraportal transplantation. Cell Transpl 16: 505–516. Kampf C, Mattsson G, Carlsson PO. 2006; Size-dependent revascularization of transplanted pancreatic islets. Cell Transpl 15: 205–209. Kelm JM, Timmins NE, Brown CJ et al. 2003; Method for generation of homogeneous

J Tissue Eng Regen Med (2014) DOI: 10.1002/term

R. A. Zuellig et al. multicellular tumor spheroids applicable to a wide variety of cell types. Biotechnol Bioeng 83: 173–180. Kempf MC, Andres A, Morel P et al. 2005; Logistics and transplant coordination activity in the GRAGIL Swiss–French multicenter network of islet transplantation. Transplantation 79: 1200–1205. Keymeulen B, Korbutt G, De Paepe M et al. 1996; Long-term metabolic control by rat islet grafts depends on the composition of the implant. Diabetes 45: 1814–1821. Kim SC, Han DJ, Kim IH et al. 2005; Comparative study on biologic and immunologic characteristics of the pancreas islet cell between 24 °C and 37 °C culture in the rat. Transpl Proc 37: 3472–3475. King AJ, Fernandes JR, Hollister-Lock J et al. 2007; Normal relationship of βand non-β-cells not needed for successful islet transplantation. Diabetes 56: 2312–2318. Korsgren O, Lundgren T, Felldin M et al. 2008; Optimising islet engraftment is critical for successful clinical islet transplantation. Diabetologia 51: 227–232. Korsgren O, Nilsson B, Berne C et al. 2005; Current status of clinical islet transplantation. Transplantation 79: 1289–1293. Lee DY, Park SJ, Nam JH et al. 2008; Optimal aggregation of dissociated islet cells for functional islet-like cluster. J Biomater Sci Polym Edn 19: 441–452. Lehmann R, Spinas GA, Moritz W et al. 2008; Has time come for new goals in human islet transplantation? Am J Transpl 8: 1096–1100. Lehmann R, Zuellig RA, Kugelmeier P et al. 2007; Superiority of small islets in human islet transplantation. Diabetes 56: 594–603. Liang Y, Najafi H, Matschinsky FM. 1990; Glucose regulates glucokinase activity in cultured islets from rat pancreas. J Biol Chem 265: 16863–16866. MacGregor RR, Williams SJ, Tong PY et al. 2006; Small rat islets are superior to large

Copyright © 2014 John Wiley & Sons, Ltd.

islets in in vitro function and in transplantation outcomes. Am J Physiol Endocrinol Metab 290: E771–779. Matta SG, Wobken JD, Williams FG et al. 1994; Pancreatic islet cell reaggregation systems: efficiency of cell reassociation and endocrine cell topography of rat isletlike aggregates. Pancreas 9: 439–449. Menger MD, Jaeger S, Walter P et al. 1989; Angiogenesis and hemodynamics of microvasculature of transplanted islets of Langerhans. Diabetes 38(suppl 1): 199–201. O’Sullivan ES, Johnson AS, Omer A et al. 2010; Rat islet cell aggregates are superior to islets for transplantation in microcapsules. Diabetologia 53: 937–945. Porrett PM, Yeh H, Frank A et al. 2007; Availability of suitable islet donors in the United States. Transplantation 84: 280–282. Potter SW, Morris JE. 1985; Development of mouse embryos in hanging drop culture. Anat Rec 211: 48–56. Ramachandran K, Williams SJ, Huang HH et al. 2013; Engineering islets for improved performance by optimized reaggregation in a micromold. Tissue Eng A 19: 604–612. Reffet S, Thivolet C. 2006; Immunology of pancreatic islet transplantation. Diabetes Metab 32: 523–526. Ryan EA, Lakey JR, Rajotte RV et al. 2001; Clinical outcomes and insulin secretion after islet transplantation with the Edmonton protocol. Diabetes 50: 710–719. Shapiro AM. 2011; State of the art of clinical islet transplantation and novel protocols of immunosuppression. Curr Diabetes Rep 11: 345–354. Shapiro AM, Lakey JR, Ryan EA et al. 2000; Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 343: 230–238. Shapiro AM, Ricordi C, Hering BJ et al. 2006; International trial of the Edmonton

protocol for islet transplantation. N Engl J Med 355: 1318–1330. Shizuru J, Trager D, Merrell RC. 1985; Structure, function, and immune properties of reassociated islet cells. Diabetes 34: 898–903. Spinas GA, Laffranchi R, Francoys I et al. 1998; The early phase of glucosestimulated insulin secretion requires nitric oxide. Diabetologia 41: 292–299. Su Z, Xia J, Shao W et al. 2010; Small islets are essential for successful intraportal transplantation in a diabetes mouse model. Scand J Immunol 72: 504–510. Tanaka H, Tanaka S, Sekine K et al. 2013; The generation of pancreatic β-cell spheroids in a simulated microgravity culture system. Biomaterials 34: 5785–5791. Thomas F, Wu J, Contreras JL et al. 2001; A tripartite anoikis-like mechanism causes early isolated islet apoptosis. Surgery 130: 333–338. Tsang WG, Zheng T, Wang Y et al. 2007; Generation of functional islet-like clusters after monolayer culture and intracapsular aggregation of adult human pancreatic islet tissue. Transplantation 83: 685–693. Wolf-Jochim M, Wohrle M, Federlin K et al. 1995; Comparison of the survival of fresh or cultured pancreatic islets, pseudoislets and single cells following allotransplantation beneath the kidney capsule in nonimmunosuppressed diabetic rats. Exp Clin Endocrinol Diabetes 103(suppl 2): 118–122. Yi P, Park JS, Melton DA. 2013; Betatrophin: a hormone that controls pancreatic β cell proliferation. Cell 153: 747–758. Yoon BS, Yoo SJ, Lee JE et al. 2006; Enhanced differentiation of human embryonic stem cells into cardiomyocytes by combining hanging drop culture and 5-azacytidine treatment. Differ Res Biol Diversity 74: 149–159.

J Tissue Eng Regen Med (2014) DOI: 10.1002/term