- Email: [email protected]
Culturing INS-1 cells on CDPGYIGSR-, RGD- and fibronectin surfaces improves insulin secretion and cell proliferation
Acta Biomaterialia 8 (2012) 619–626
Contents lists available at SciVerse ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Culturing INS-1 cells on CDPGYIGSR-, RGD- and fibronectin surfaces improves insulin secretion and cell proliferation Carina Kuehn, Evan A. Dubiel, Georges Sabra, Patrick Vermette ⇑ Laboratoire de bio-ingénierie et de biophysique de l’Université de Sherbrooke, Department of Chemical and Biotechnological Engineering, Université de Sherbrooke, 2500, boulevard de l’Université, Sherbrooke, Québec, Canada J1K 2R1
a r t i c l e
i n f o
Article history: Received 28 July 2011 Received in revised form 20 September 2011 Accepted 28 October 2011 Available online 4 November 2011 Keywords: INS-1 beta cells Insulin secretion and proliferation Integrin and E-cadherin expression Fibronectin and RGD CDPGYIGSR
a b s t r a c t Rat insulinoma cells (INS-1), an immortalized pancreatic beta cell line, were cultured on low-fouling carboxymethyl-dextran (CMD) layers bearing fibronectin, the tripeptide Arg–Gly–Asp (RGD) or CDPGYIGSR, a laminin nonapeptide. INS-1 cells were non-adherent on CMD and RGE but adhered to fibronectin- and peptide-coated CMD surfaces and to tissue culture polystyrene (TCPS). On CMD bearing fibronectin and the peptides, INS-1 cells showed higher glucose-stimulated insulin secretion compared to those on TCPS, bare CMD and RGE. INS-1 cells experienced a net cell growth, with the lowest found after 7 days on CMD and the highest on fibronectin. Similarly, cells on RGD and CDPGYIGSR showed lower net growth rates than those on fibronectin. Expression of E-cadherin and integrins avb3 and a5 were similar between the conditions, except for a5 expression on fibronectin, RGD and CDPGYIGSR. Larger numbers of Ki-67-positive cells were found on CDPGYIGSR, TCPS, fibronectin and RGD. Cells in all conditions expressed Pdx1. Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction One of the most promising treatments of insulin-dependent diabetes in humans is to graft isolated pancreatic islets to potentially restore the patient’s insulin production. The loss of extracellular matrix (ECM) during the isolation process is one of the many reasons for occurring problems after transplantation. Cell–ECM interactions are crucial for cell differentiation, migration, morphology, growth, function and survival [1]. The islet ECM is constituted of the interstitial matrix and the basement membrane. Laminin is the major component of the basement membrane, whereas fibronectin is found more abundantly beneath endothelial cells and epithelial ducts in the interstitial matrix [2,3]. Fibronectin has been shown to improve the proliferation of neonatal rat islet cells and furthermore prevent apoptosis [2,3]. It is secreted endogenously as an inactive dimeric complex. When bound to its corresponding cell surface receptors (i.e. integrins) the complex is activated, which results in unfolding and subsequent fibril assembly of fibronectin. Thus, formation of the cell supporting network is initiated [4,5]. Laminins are heterotrimers formed by a, b and c chains in multiple combinations [6]. It has been shown that laminins, either in purified form or as part of a cell-secreted matrix, improved pancreatic beta cell function and survival [7]. YIGSR, a synthetic laminin pentapeptide and the adhesion sequence of the b1 chain of laminin, encapsulated in a hydrogel exhibited a pro-survival ef⇑ Corresponding author. Tel.: +1 819 821 8000x62826; fax: +1 819 821 7955. E-mail address: [email protected] (P. Vermette).
fect on beta cells [7]. The tripeptide Arg–Gly–Asp (RGD) is an integrin recognition sequence found in multiple ECM proteins like fibronectin, laminins, collagen type I, vitronectin, fibrinogen and others [8]. Approximately one-third of the integrin receptor family recognizes RGD. The tripeptide itself improves the resistance of isolated islets against apoptosis [9,10]. Most recent efforts to immobilize ECM proteins or peptides have consisted of encapsulating these molecules in hydrogels using the cell-secreted 804G matrix, constituted of several ECM components [11]. Others have applied unspecific coating of surfaces using incubation of ECM molecules in solution [11]. In the present study, fibronectin, CDPGYIGSR and RGD were covalently grafted on carboxymethyl-dextran (CMD)-covered tissue culture polystyrene (TCPS) plates. CMD itself has shown excellent non-adherence properties towards several cell types and was therefore used as a negative control [12]. The cell type used for this study was INS-1, a rat insulinoma cell line that secretes insulin following glucose stimulation. The culture of INS-1 cells on fibronectin-, CDPGYIGSR- and RGD-covered CMD substrates was performed to compare the influence of the different molecules on the insulin secretion, cell proliferation and cell–ECM interactions. 2. Experimental section 2.1. Surface preparation The preparation of the surfaces were performed by functionalizing TCPS 12-well plates (3.84 cm2 surface area per well) with primary amines using plasma polymerization of n-heptylamine
1742-7061/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2011.10.036
620
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
(126802, Sigma–Aldrich, Oakville, ON, Canada) in a custom-built reactor as described elsewhere [13]. Briefly, n-heptylamine monomer pressure was stabilized at 0.040 torr and glow discharged for 70 s at an ignition power of 80 W and an excitation frequency of 50 kHz. The distance between the electrodes was 10 cm. CMD was then immediately grafted on the amine-functionalized plates. The synthesis of dextran into 25% CMD and the grafting on the plates have been described in detail elsewhere [12]. Briefly, CMD was dissolved in Milli-Q water to yield a 2 mg ml1 solution, which was then filtered through a 0.45 lm filter. The CMD was activated by the addition of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, E1769, Sigma–Aldrich) and N-hydroxysuccinimide (NHS, 130672, Sigma–Aldrich), to a concentration of 17 mM, for 10 min. The freshly n-heptylamine plasma polymer (HApp)-functionalized surfaces were covered with activated CMD solution and the reaction was allowed to proceed for 2 h at ambient temperature under agitation. The surfaces were then rinsed in 1 M NaCl solution (2 12 h), rinsed in Milli-Q water (2 12 h) and stored at 4 °C until use. CMD surfaces have been shown to exert excellent low-fouling behaviour towards several cell types and proteins [14,15]. Peptide grafting was performed as described previously [12]. Briefly, GRGDS and GRGES (44-0-23 and 44-0-51, respectively, American Peptide, Sunnyvale, CA, USA) solutions were prepared to a final concentration of 50 lg ml1 in 1 (0.1 M) phosphate-buffered saline (PBS; BP6651, Fisher Scientific, Ottawa, ON, Canada). CMD surfaces were activated in a 200 mM aqueous solution of EDC + NHS, rapidly rinsed in Milli-Q water and then covered in peptide solution. Samples were placed under agitation and the reaction was allowed to proceed for 2 h at ambient temperature followed by rinsing twice in 1 PBS. All peptide grafting was performed under sterile conditions. CDPGYIGSR (87-0-53, American Peptide Company) and fibronectin (human plasma fibronectin; FC010, Millipore, Temecula, CA, USA) were covalently immobilized on CMD-grafted 12-well TCPS plates using maleimide and carbodiimide chemistries, respectively. 2.2. X-ray photoelectron spectroscopy (XPS) Surfaces were analysed using an AXIS ULTRA DLD spectrometer (Kratos Analytical Ltd., Manchester, UK) with a 225 W Al Ka source to validate the presence of the immobilized peptides. Elemental composition was determined with a survey spectrum at a pass energy of 160 eV. CasaXPS (Casa Software Ltd.) was used to calculate the atomic concentrations of each element. Three separate areas of each surface were analysed. 2.3. Cell culture Rat insulinoma cells (INS-1, clone 832/13) [16,17] were cultured in RPMI-1640 with L-glutamine (31800-022, Invitrogen) containing 10% foetal bovine serum (FBS; 12483, Invitrogen), 10 mM HEPES (H3375, Sigma–Aldrich), 1 mM sodium pyruvate (S8636, Sigma–Aldrich) and 50 lM b-mercaptoethanol (M7522, Sigma–Aldrich). The culture medium was changed every 3 days. Upon confluence, cells were trypsinized and seeded onto the prepared surfaces at a density of 50,000 cells ml1. Cells were cultured for 7 days in all conditions. 2.4. Glucose-stimulated insulin secretion (GSIS) A GSIS test was performed with the 7 day cultures. For this purpose, the media were discarded and replaced with a low-glucose Krebs–Ringer buffer with HEPES (KRBH) solution supplemented with 2.8 mM D-glucose. Samples were then incubated for 1 h.
Afterwards, low-glucose KRBH was discarded and samples were rinsed with PBS. KRBH containing 28 mM D-glucose was then added and samples incubated for 1 h. A rinsing step followed. Samples were then incubated for 1 h in KRBH supplemented with 28 mM D-glucose + 50 lM 3-isobutyl-1-methyl-xanthine (IBMX; I5879, Sigma–Aldrich) in KRBH. After a final rinsing step, samples were again incubated in low-glucose KRBH for 1 h. All media were collected and stored at 20 °C for further insulin content measurement. Insulin contents in the culture media were measured using an insulin (rat) high-range enzyme-linked immunosorbent assay kit (0.52 ng ml1 sensitivity; 80-INSRTH-E01, Alpco Diagnostics, Salem, NH, USA). The insulin content was normalized to the cell number for each condition. Absorbance was measured using a Synergy™ HT Multi-Detection Microplate Reader (Bio-Tek, Winooski, VT, USA).
2.5. Immunofluorescence Non-adherent INS-1 cultures (i.e. RGE and CMD) were collected in centrifuge tubes and centrifuged for 5 min at 1200 rpm. Adherent cells (i.e. RGD, CDPGYIGSR, fibronectin and TCPS) were trypsinized (1 min, ambient temperature) and collected in stopping solution (10% FBS in PBS), then centrifuged for 5 min at 1200 rpm. All cell pellets were rinsed with PBS, centrifuged and then fixed in 4% paraformaldehyde for 2 h at 4 °C, followed by 2% agarose and paraffin embedding according to standard protocols [1]. Paraffin sections were cut consecutively at a thickness of 4 lm. Primary antibodies to insulin (ab7842), E-cadherin (ab53033), aVb3 (ab7166), Ki-67 (ab16667), pancreatic and duodenal homeobox 1 (Pdx1, ab47267) were purchased from Abcam (Cambridge, MA, USA). The primary antibody for the a5 integrin (sc-10729) was purchased from Santa Cruz (Santa Cruz, CA, USA). Secondary antibodies were obtained from Invitrogen. All sections were blocked in 10% goat serum and counterstained with 40 ,6-diamidino-2-phenylindole, dihydrochloride (DAPI; D1306, Invitrogen). Manufacturer recommendations were used for pretreatment, dilutions, incubation time and temperature. Samples were imaged using an epifluorescence microscope (Nikon Eclipse).
2.6. Ki-67 and Pdx1 analysis Ki-67 and Pdx1 stained sections, processed as described in Section 2.5, were manually counted from fluorescent images taken at 60 magnification using the Simple PCI software (Hamamatsu Corporation). To estimate the percentage of Pdx1-positive cells, five independent fields of each condition were counted in triplicate. The average total cell number varied between 706 and 885 cells per repeat. Ki-67 sections were treated accordingly. Average cell numbers ranged between 327 and 838 cells per repeat. Data are presented as an average ± standard deviation (SD).
2.7. Assay for assessing cell number Cell numbers were quantified using a CyQUANTÒ assay (C35006, Invitrogen). Prior to carrying out this assay, non-adherent and adherent cells were collected as described in Section 2.5. After centrifugation, the pellets were rinsed and resuspended in Hanks’ balanced salt solution supplied with the assay and centrifuged a second time at 1200 rpm for 1 min. Supernatant was removed and the cell pellet was resuspended in dye solution for 1 h at 37 °C in a 96-well TCPS plate. Fluorescence was measured using a Synergy™ HT Multi-Detection Microplate Reader (Bio-Tek).
621
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
2.8. Statistical analysis Data are presented as mean ± SD for three independent experiments. Differences in results were assessed by a single-factor analysis of variance. A probability (p) value of < 0.05 was considered statistically significant. 3. Results and discussion In this study, low-fouling surfaces with well-defined surface properties were used to graft ECM components in an attempt to compare their effect on proliferation and insulin secretion of INS1 cells. The RGD and CDPGYIGSR peptides as well as fibronectin were covalently bound to low-fouling CMD surfaces, while TCPS served as a positive control. Bare CMD and RGE surfaces were used as negative controls. This approach allows the covalent linking of ECM-derived proteins and peptides, namely fibronectin, RGD and CDPGYIGSR, to low-fouling CMD layers using maleimide and carbodiimide chemistry, thus enabling the study of specific cell– ECM interactions while limiting non-specific interactions and the influence of FBS-derived proteins. The XPS results (Fig. 1 and Table 1) confirmed that RGD, fibronectin, CDPGYIGSR and RGE were covalently linked to the CMD surfaces. For each surface, there was a marked increase in the N/C atomic concentration ratio (Table 1) upon the addition of CDPGYIGSR, fibronectin, RGD and RGE to the CMD layers. This is indicative of the addition of a nitrogen-rich layer, typical of amino acids and proteins, to CMD surfaces. Furthermore, high-resolution C 1s spectra (Fig. 1) showed perturbations at binding energies corresponding to ether (286.5 eV, C–O–H and C–O–C), amide (288 eV, N–C@O) and ester/
Table 1 XPS atomic concentration (%) and atomic ratio of HApp, HApp + CMD, HApp + CMD + CDPGYIGSR, HApp + CMD + fibronectin, HApp + CMD + RGD and HApp + CMD + RGE. Atomic concentration (%)
HApp CMD CDPGYIGSR Fibronectin RGD RGE
Atomic ratio
O 1s
C 1s
N 1s
O/C
N/C
1.8 19.7 20.03 19.0 20.2 19.6
90.8 76.1 74.18 73.2 75.0 76.0
7.4 4.1 5.61 8.1 4.4 4.4
0.02 0.26 0.27 0.26 0.27 0.26
0.08 0.05 0.08 0.11 0.06 0.06
acid (289 eV, O–C@O) groups, also indicative of the addition of a peptide/protein to the CMD layers. Net cell growth had occurred for each culture condition, as shown in Fig. 2 (CyQUANT assay), Fig. 3 (percentage of Ki-67-positive cells) and supplementary Fig. A1 (Ki-67-positive cells). Attached INS-1 cells showed higher net cell growth rates in general when compared to the postulated 100 h doubling time [18]. It was not surprising to obtain higher cell numbers on bioactivated surfaces in comparison with RGE and bare CMD surfaces. Cell net growth on fibronectin measured with the CyQUANT assay was significantly higher than on all other surfaces (p < 0.03) except RGD (p > 0.05). Due to the cell–cell contacts [19] within the INS-1 clusters formed in suspension on CMD or loosely bound on RGE, proliferation still occurred despite the low-fouling nature of the CMD and RGE surfaces. The largest numbers of cells positively stained for Ki-67 were found on CDPGYIGSR and TCPS (p < 0.004). One reason for this
Fig. 1. High-resolution XPS C 1s spectra of CDPGYIGSR, fibronectin, RGD and RGE surfaces and interlayers.
622
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
*
INS-1 cell number
3,5E+05 3,0E+05 2,5E+05
*
2,0E+05 1,5E+05 1,0E+05 5,0E+04 0,0E+00 CMD
RGE
RGD
CDPGYIGSR Fibronectin
TCPS
Ki-67 positive cells (%)
Fig. 2. INS-1 cell number after 7 days of culture on the different surfaces. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
*
60 50 40 30 20 10 0 CMD
RGE
TCPS
Fibronectin
RGD
CDPGYIGSR
Fig. 3. Percentage of INS-1 cells stained positively for Ki-67 after 7 days of culture on the different surfaces. Total cell numbers were obtained from DAPI staining. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
could be that cells on fibronectin and RGD were grown close to confluence, and were therefore in a lag phase of growth. Despite that, more cells stained positively for Ki-67 on fibronectin and RGD than on CMD and RGE (p < 0.005 and p < 0.005, respectively). Additionally, Ki-67 is a proliferation marker found in all stages of the cell cycle except G0 [20]. Consequently, cells staining positively for Ki-67 are not necessarily undergoing mitosis. When INS-1 cells were exposed to RGD, CDPGYIGSR, fibronectin and TCPS surfaces, they adhered to the surface tightly and began forming well-defined surface-bound clusters on RGD, CDPGYIGSR and fibronectin. Conversely, INS-1 cells were non-adherent when seeded on bare CMD and adhered only loosely on RGE, forming ‘‘islet-like’’ aggregates in both conditions (Fig. 4). These findings indicate that INS-1 adherent cultures use their ECM receptors to bind to the specific surface ligands. Consequently, a5 and aVb3 integrin expression was assessed. The a5-subunit binds exclusively to b1integrin allowing an estimation of the a5b1-expression by staining for a5 only [21]. In previous studies, a5 expression has been observed during foetal rat pancreatic development and b1 expression, when suppressed in adult rat pancreatic beta cells, was found to be pivotal for survival and migration of rat beta cells on 804G matrix [10,11,22]. In contrast to those findings, a5 was expressed on all surfaces. This is surprising as cells on bare CMD and RGE surfaces have no access to ECM cues. Due to the low-fouling properties of both CMD and RGE, cells cannot attach and therefore stay in suspension. In the non-adherent CMD and RGE systems, the a5 and aVb3 integrins appeared to be expressed predominantly on those cells residing in the core of the aggregates, whereas on adherent surfaces they were expressed throughout, as shown in Fig. 5 (see also supplementary Fig. A2, 100 magnification). One possible explanation for this strong expression could be the presence of ECM molecules synthesized by the INS-1 cells into the interstitial space of the aggregates. FBS-associated proteins are not able to coat the CMD-covered surfaces as these surfaces are low fouling.
The binding of INS-1 cells to RGE, but not CMD (Fig. 4), is most probably due to unspecific INS-1 surface receptor–RGE interactions. If the TCPS plate on which the CMD was plated and RGE was immobilized was only slightly agitated, INS-1 cell–RGE aggregates detached immediately. Also, cell aggregates on RGE were morphologically closer to floating INS-1 aggregates on CMD than to the surface-bound cell clusters on RGD, fibronectin, CDPGYIGSR and TCPS. To ensure that differences between the CMD and RGE surfaces were studied correctly, the detachment of cell aggregates on RGE had been avoided. On CDPGYIGSR, RGD and fibronectin the expression of a5 seemed to be elevated as the intensity of the staining was more pronounced when compared to the other conditions. Both a5 and aVb3 integrins have been identified as RGD-binding integrins in pancreatic tissues [1,10]. Fibronectin contains multiple RGD motifs [3]. Therefore, INS-1 cells express integrins that bind to both RGD and fibronectin. In the case of CDPGYIGSR, the positive integrin staining is most probably due also to cell-derived ECM after a 7 day culture. Additionally, the YIGSR motif interacts specifically with the 67 kDa laminin receptor, which has not been assessed in this study [23]. Cells on RGD and fibronectin as well as those on TCPS all expressed the aVb3 integrin. It has previously been described that integrin aVb3 plays a role in cell adhesion during pancreatic development, though it was recently shown to be less significant than other aV integrins in foetal human islet preparations. Interestingly, the adult islet expression levels of aVb3 were the same as those of foetal islets [24]. Our findings partially support those results as INS-1 cell expression of aVb3 was overall weaker when compared to the a5 integrin (Fig. 5). In contrast, our results suggest that aVb3 plays a role in INS-1 adhesion as it was expressed in all conditions. E-cadherin was shown to be essential for the islet formation and recognition of the different cell types within the islet as it is expressed in b-cells only. The cadherin family are Ca2+-dependent cell anchors which bind to cadherin homologues of neighbouring cells [25]. Here, b-cell E-cadherin expression is consistent with previous studies. E-cadherin is apparent and convincingly expressed throughout all conditions. Similar expression was detected on clustered cells grown on CDPGYIGSR, RGD and fibronectin, and also in the CMD and RGE aggregates, which are both shown in Fig. 6. CMD and RGE were selected to show the positive E-cadherin staining in general and to highlight the finding of E-cadherin expression in the non-adherent cultures. The cells on bare CMD surfaces, which stayed in suspension, and the loosely bound cells on RGE most probably established cell–cell contacts to avoid detachmentinduced apoptosis, known as anoikis [26]. As long-term cultures of insulin-producing cells are of great interest for further studies and transplantation purposes, 7 day cultures were chosen. To examine the possibility of cell-derived ECM deposition and alteration of the findings of this study, 3 day cultures were also stained for a5 and aVb3 integrins and E-cadherin. No differences in expression levels between the 3 and 7 day cultures could be found, as shown in supplementary Figs. A3 (a5 and aVb3 integrin expression) and A4 (E-cadherin expression). INS-1 cells were stained for insulin and glucagon to confirm the purity of the cell line and to ensure the viability and function of the insulin-secreting cells. INS-1 cells expressed insulin but not glucagon after 7 days of culture throughout all conditions (supplementary Fig. A5). Previous studies using insulinoma cells found pseudo-islet formation in suspension. The insulin secretion of the pseudo-islets was negatively influenced by blocking E-cadherin [27]. In another study, pancreatic precursor cells, not detached from their ECM, formed so-called ‘‘islet-like’’ cell aggregates [28]. Both these findings might help to explain the insulin expression and secretion seen on bare CMD and RGE surfaces supported by the E-cadherin-mediated cell–cell adhesion.
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
623
Fig. 4. Phase-contrast microscopy pictures of INS-1 cells grown on the different surfaces after 72 h. 20 magnification. Bars correspond to 100 lm.
While all surface conditions resulted in glucose-stimulated insulin secretion (Fig. 7), INS-1 cells on fibronectin and RGD secreted remarkably higher levels of insulin in comparison with the other culture conditions tested after stimulation with high glucose (p < 0.04) and 50 lM IBMX (p < 0.005). When stimulated with high glucose only, cells cultured on CDPGYIGSR did not secrete more insulin than cells cultured on TCPS. Inversely, when incubated with high glucose plus IBMX, cells grown on CDPGYIGSR secreted significantly more insulin than cells cultured on TCPS (p < 0.006). It has been shown that isolated human islets cultured in the presence of fibronectin, laminins, and collagens I and IV are able to maintain their abilities to adhere, survive and function in an improved manner in vitro [29]. Hence, the higher insulin secretion on fibronectin and RGD is consistent with previous studies [1,26]. The higher secretion of insulin on the ECM surfaces is most probably due to the resulting cell–ECM interactions [1]. Surprisingly, cells on RGE secreted approximately the same amount of insulin into the medium as cells on TCPS when normalized to the cell number. INS-1 aggregates found on RGE did not secrete significantly more insulin than those on CMD. INS-1 cells cultured on CMD and RGE both formed ‘‘islet-like’’ clusters
(Fig. 4). Therefore it is possible to compare them to MIN-6 pseudo-islets in both function and proliferation behaviour as they maintain their ability to secrete insulin in response to glucose [27]. Recent studies have shown that cancerous or transformed beta cell lines with high proliferation, e.g. INS-1 or MIN-6 cells, produce and secrete less insulin than primary beta cells [30]. The relatively low insulin secretion of cells grown on CDPGYIGSR when stimulated with 28 mM glucose could be explained by the observed high proliferation rates on CDPGYIGSR, as revealed from cells staining positively for Ki-67. This is an agreement with a previous study showing that insulin secretion from an immortalized human beta cell line can be affected by the cell proliferation state [31]. INS-1 cells on CMD and RGE with the lowest levels of positive Ki-67 staining secreted relatively less insulin than INS cells in all other conditions. It is possible that cell aggregates on CMD and RGE were negatively influenced by the lack of cell–ECM support, thus showing less integrin expression compared to the other conditions. Additionally cells in all conditions expressed Pdx1 (supplementary Fig. A6). The relative numbers of cells grown on fibronectin, RGD, CDPGYIGSR and TCPS and expressing Pdx1 were larger than those on CMD and RGE (p < 0.002 and p < 0.02, respectively;
624
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
Fig. 5. Expression of integrins a5 (red) and aVb3 (green) of INS-1 cells cultured for 7 days on the different surfaces. Nuclei are stained with DAPI (blue). 60 magnification. Bars correspond to 50 lm.
Fig. 6. Immunofluorescence staining for E-cadherin of INS-1 cells cultured for 7 days on CMD and RGE. Nuclei are DAPI stained in blue. 60 magnification. Bars correspond to 50 lm.
Fig. 8). The transcription factor Pdx1 is essential for pancreatic development and maintenance of beta cell function and survival [32]. The interaction between insulin and Pdx1 seems to prevent islets from undergoing apoptosis [33]. Our results suggest that
cell–ECM interactions promoted by the surface-immobilized fibronectin, RGD and CDPGYIGSR enhance insulin secretion and the number of cells expressing Pdx1 via integrin-receptor activation. This is supported by a study conducted by Krishnamurthy et al.
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
Insulin ng/106 cells
250
625
Appendix A. Supplementary data
*
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.actbio.2011.10.036.
200
150
Appendix B. Figures with essential colour discrimination
100
Certain figures in this article, particularly Figures 5 and 6, is difficult to interpret in black and white. The full colour images can be found in the on-line version, at doi:10.1016/j.actbio.2011.10.036.
50
0 CMD
RGE low glucose
TCPS highglucose
RGD
CDPGYIGSR
highglucose+ IBMX
Fibronectin
low glucose
Pdx1 positive cells (%)
Fig. 7. GSIS of INS-1 cells cultured on different surfaces for 7 days. Insulin secretion is normalized per 106 cells. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
*
100 90 80 70 60 50 40 30 20 10 0
CMD
RGE
TCPS
Fibronectin
RGD
CDPGYIGSR
Fig. 8. Percentage of INS-1 cells stained positively for Pdx1 after 7 days of culture on the different surfaces. Total cell numbers were obtained from DAPI staining. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
[17] showing that perturbation of integrin function results in decreased human islet and INS-1 cell proliferation, Pdx1 expression and decreased insulin expression and secretion. 4. Conclusions Globally, we have demonstrated that INS-1 cells cultured on substrates bearing low-fouling CMD layers, on which fibronectin, RGD and CDPGYIGSR were covalently linked, exhibited higher insulin secretion and cell proliferation compared to the CMD, RGE and TCPS controls. These well-defined surfaces could be used for further characterization of ECM–beta cell interactions, thus leading to a better understanding of beta cell function and subsequently diabetes and pancreatic development. Additionally, the surfaces could be tested with isolated human islets to investigate their effect over islet survival and insulin secretion in response to glucose. Acknowledgements This research project was supported by the Université de Sherbrooke and the National Science and Engineering Research Council of Canada (NSERC) through a Discovery Grant. We also thank Dr. Rennian Wang from the Child Health Research Institute and Department of Physiology & Pharmacology (The University of Western Ontario, London, ON, Canada) for her insight on the culture of INS-1 cells and immunofluorescence protocols. The technical assistance of Pascale Roy is recognized. We appreciate the contribution of Jamie Sharp who provided language support in proof reading the article.
References [1] Wang RN, Rosenberg L. Maintenance of beta-cell function and survival following islet isolation requires re-establishment of the islet-matrix relationship. J Endocrinol 1999;163:181–90. [2] Hulinsky I, Cooney S, Harrington J, Silink M. In vitro growth of neonatal rat islet cells is stimulated by adhesion to matrix. Hormon Metab Res 1995;27:209–15. [3] Zhang Z, Vuori K, Reed JC, Ruoslahti E. The alpha 5 beta 1 integrin supports survival of cells on fibronectin and up-regulates Bcl-2 expression. Proc Natl Acad Sci USA 1995;92:6161–5. [4] Johnson KJ, Sage H, Briscoe G, Erickson HP. The compact conformation of fibronectin is determined by intramolecular ionic interactions. J Biol Chem 1999;274:15473–9. [5] Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 1992;69:11–25. [6] Burgeson RE, Chiquet M, Deutzmann R, Ekblom P, Engel J, Kleinman H, et al. A new nomenclature for the laminins. Matrix Biol 1994;14:209–11. [7] Weber LM, Hayda KN, Haskins K, Anseth KS. The effects of cell–matrix interactions on encapsulated beta-cell function within hydrogels functionalized with matrix-derived adhesive peptides. Biomaterials 2007;28:3004–11. [8] Ruoslahti E. RGD and other recognition sequences for integrins. Annu Rev Cell Dev Biol 1996;12:697–715. [9] Pinske GGM, Bouwman WP, Jiawan-Lalai R, Terpstra OT, Bruijn JA, de Heer E. Integrin signaling via RGD peptides and anti-b1 antibodies confers resistance to apoptosis in islets of Langerhans. Diabetes 2006;55:312–7. [10] Wang R, Li J, Lyte K, Yashpal NK, Fellows F, Goodyer CG. Role for beta1 integrin and its associated alpha3, alpha5, and alpha6 subunits in development of the human fetal pancreas. Diabetes 2005;54:2080–9. [11] Hammar E, Parnaud G, Bosco D, Perriraz N, Maedler K, Donath M, et al. Extracellular matrix protects pancreatic beta-cells against apoptosis: role of short- and long-term signaling pathways. Diabetes 2004;53:2034–41. [12] Monchaux E, Vermette P. Bioactive microarrays immobilized on low-fouling surfaces to study specific endothelial cell adhesion. Biomacromolecules 2007;8:3668–73. [13] Martin Y, Boutin D, Vermette P. Study of the effect of process parameters for nheptylamine plasma polymerization on final layer properties. Thin Solid Films 2007;515:6844–52. [14] Dubois J, Gaudreault C, Vermette P. Biofouling of dextran-derivative layers investigated by quartz crystal microbalance. Colloids Surf B 2009;71:293–9. [15] Monchaux E, Vermette P. Cell adhesion resistance mechanisms using arrays of dextran-derivative layers. J Biomed Mater Res Part A 2008;85A:1052–63. [16] Krishnamurthy M, Li J, Al-Masri M, Wang R. Expression and function of ab1 integrins in pancreatic beta (INS-1) cells. J Cell Commun Signal 2008;2:67–79. [17] Krishnamurthy M, Li J, Fellows GF, Rosenberg L, Goodyer CG, Wang R. Integrin a3, but not b1, regulates islet cell survival and function via PI3K/Akt signaling pathways. Endocrinology 2011;152:424–35. [18] Asfari M, Janjic D, Meda P, Li G, Halban PA, Wollheim CB. Establishment of 2mercaptoethanol-dependent differentiated insulin-secreting cell lines. Endocrinology 1992;130:167–78. [19] Carvell MJ, Marsh PJ, Persaud SJ, Jones PM. E-cadherin interactions regulate bcell proliferation in islet-like structures. Cell Physiol Biochem 2007;20:617–26. [20] Köhler CU, Kreuter A, Rozynkowski MC, Rahmel T, Uhl W, Tannapfel A, et al. Validation of different replication markers for the detection of beta-cell proliferation in human pancreatic tissue. Regul Pept 2010;162:115–21. [21] Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell 2002;110:673–87. [22] Yashpal NK, Li J, Wheeler MB, Wang R. Expression of beta1 integrin receptors during rat pancreas development-sites and dynamics. Endocrinology 2005;146:1798–807. [23] Graf J, Ogle RC, Robey FA, Sasaki M, Martin GR, Yamada Y, et al. A pentapeptide from the laminin B1 chain mediates cell adhesion and binds the 67,000 laminin receptor. Biochemistry 1987;26:6896–900. [24] Kaido T, Perez B, Yebra M, Hill J, Cirulli V, Hayek A, et al. Alphav-integrin utilization in human beta-cell adhesion, spreading, and motility. J Biol Chem 2004;279:17731–7. [25] Jain R, Lammert E. Cell–cell interactions in the endocrine pancreas. Diabetes Obes Metab 2009;11:159–67.
626
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
[26] Weber LM, Hayda KN, Anseth KS. Cell–matrix interactions improve beta-cell survival and insulin secretion in three-dimensional culture. Tissue Eng Part A 2008;14:1959–68. [27] Rogers GJ, Hodgkin MN, Squires PE. E-cadherin and cell adhesion: a role in architecture and function in the pancreatic islet. Cell Physiol Biochem 2007;20:987–94. [28] Hardikar AA, Marcus-Samuels B, Geras-Raaka E, Raaka BM, Gershengorn MC. Human pancreatic precursor cells secrete FGF2 to stimulate clustering into hormone-expressing islet-like cell aggregates. Proc Natl Acad Sci USA 2003;100:7117–22. [29] Daoud J, Petropavlovskaia M, Rosenberg L, Tabrizian M. The effect of extracellular matrix components on the preservation of human islet function in vitro. Biomaterials 2010;31:1676–82.
[30] Reers C, Hauge-Evans AC, Morgan NG, Wilcox A, Persaud SJ, Jones PM. Downregulation of proliferation does not affect the secretory function of transformed b-cell lines regardless of their anatomical configuration. Islets 2011;3:80–8. [31] Narushima M, Kobayashi N, Okitsu T, Tanaka Y, Li SA, Chen Y, et al. A human bcell line for transplantation therapy to control type 1 diabetes. Nat Biotechnol 2005;23:1274–82. [32] Bernardo AS, Hay CW, Docherty K. Pancreatic transcription factors and their role in the birth, life and survival of the pancreatic beta cell. Mol Cell Endocrinol 2008;294:1–9. [33] Johnson JD, Bernal-Mizrachi E, Alejandro EU, Han Z, Kalynyak TB, Li H, et al. Insulin protects islets from apoptosis via Pdx1 and specific changes in the human islet proteome. Proc Natl Acad Sci USA 2006;103:19575–80.
Contents lists available at SciVerse ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Culturing INS-1 cells on CDPGYIGSR-, RGD- and fibronectin surfaces improves insulin secretion and cell proliferation Carina Kuehn, Evan A. Dubiel, Georges Sabra, Patrick Vermette ⇑ Laboratoire de bio-ingénierie et de biophysique de l’Université de Sherbrooke, Department of Chemical and Biotechnological Engineering, Université de Sherbrooke, 2500, boulevard de l’Université, Sherbrooke, Québec, Canada J1K 2R1
a r t i c l e
i n f o
Article history: Received 28 July 2011 Received in revised form 20 September 2011 Accepted 28 October 2011 Available online 4 November 2011 Keywords: INS-1 beta cells Insulin secretion and proliferation Integrin and E-cadherin expression Fibronectin and RGD CDPGYIGSR
a b s t r a c t Rat insulinoma cells (INS-1), an immortalized pancreatic beta cell line, were cultured on low-fouling carboxymethyl-dextran (CMD) layers bearing fibronectin, the tripeptide Arg–Gly–Asp (RGD) or CDPGYIGSR, a laminin nonapeptide. INS-1 cells were non-adherent on CMD and RGE but adhered to fibronectin- and peptide-coated CMD surfaces and to tissue culture polystyrene (TCPS). On CMD bearing fibronectin and the peptides, INS-1 cells showed higher glucose-stimulated insulin secretion compared to those on TCPS, bare CMD and RGE. INS-1 cells experienced a net cell growth, with the lowest found after 7 days on CMD and the highest on fibronectin. Similarly, cells on RGD and CDPGYIGSR showed lower net growth rates than those on fibronectin. Expression of E-cadherin and integrins avb3 and a5 were similar between the conditions, except for a5 expression on fibronectin, RGD and CDPGYIGSR. Larger numbers of Ki-67-positive cells were found on CDPGYIGSR, TCPS, fibronectin and RGD. Cells in all conditions expressed Pdx1. Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction One of the most promising treatments of insulin-dependent diabetes in humans is to graft isolated pancreatic islets to potentially restore the patient’s insulin production. The loss of extracellular matrix (ECM) during the isolation process is one of the many reasons for occurring problems after transplantation. Cell–ECM interactions are crucial for cell differentiation, migration, morphology, growth, function and survival [1]. The islet ECM is constituted of the interstitial matrix and the basement membrane. Laminin is the major component of the basement membrane, whereas fibronectin is found more abundantly beneath endothelial cells and epithelial ducts in the interstitial matrix [2,3]. Fibronectin has been shown to improve the proliferation of neonatal rat islet cells and furthermore prevent apoptosis [2,3]. It is secreted endogenously as an inactive dimeric complex. When bound to its corresponding cell surface receptors (i.e. integrins) the complex is activated, which results in unfolding and subsequent fibril assembly of fibronectin. Thus, formation of the cell supporting network is initiated [4,5]. Laminins are heterotrimers formed by a, b and c chains in multiple combinations [6]. It has been shown that laminins, either in purified form or as part of a cell-secreted matrix, improved pancreatic beta cell function and survival [7]. YIGSR, a synthetic laminin pentapeptide and the adhesion sequence of the b1 chain of laminin, encapsulated in a hydrogel exhibited a pro-survival ef⇑ Corresponding author. Tel.: +1 819 821 8000x62826; fax: +1 819 821 7955. E-mail address: [email protected] (P. Vermette).
fect on beta cells [7]. The tripeptide Arg–Gly–Asp (RGD) is an integrin recognition sequence found in multiple ECM proteins like fibronectin, laminins, collagen type I, vitronectin, fibrinogen and others [8]. Approximately one-third of the integrin receptor family recognizes RGD. The tripeptide itself improves the resistance of isolated islets against apoptosis [9,10]. Most recent efforts to immobilize ECM proteins or peptides have consisted of encapsulating these molecules in hydrogels using the cell-secreted 804G matrix, constituted of several ECM components [11]. Others have applied unspecific coating of surfaces using incubation of ECM molecules in solution [11]. In the present study, fibronectin, CDPGYIGSR and RGD were covalently grafted on carboxymethyl-dextran (CMD)-covered tissue culture polystyrene (TCPS) plates. CMD itself has shown excellent non-adherence properties towards several cell types and was therefore used as a negative control [12]. The cell type used for this study was INS-1, a rat insulinoma cell line that secretes insulin following glucose stimulation. The culture of INS-1 cells on fibronectin-, CDPGYIGSR- and RGD-covered CMD substrates was performed to compare the influence of the different molecules on the insulin secretion, cell proliferation and cell–ECM interactions. 2. Experimental section 2.1. Surface preparation The preparation of the surfaces were performed by functionalizing TCPS 12-well plates (3.84 cm2 surface area per well) with primary amines using plasma polymerization of n-heptylamine
1742-7061/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2011.10.036
620
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
(126802, Sigma–Aldrich, Oakville, ON, Canada) in a custom-built reactor as described elsewhere [13]. Briefly, n-heptylamine monomer pressure was stabilized at 0.040 torr and glow discharged for 70 s at an ignition power of 80 W and an excitation frequency of 50 kHz. The distance between the electrodes was 10 cm. CMD was then immediately grafted on the amine-functionalized plates. The synthesis of dextran into 25% CMD and the grafting on the plates have been described in detail elsewhere [12]. Briefly, CMD was dissolved in Milli-Q water to yield a 2 mg ml1 solution, which was then filtered through a 0.45 lm filter. The CMD was activated by the addition of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, E1769, Sigma–Aldrich) and N-hydroxysuccinimide (NHS, 130672, Sigma–Aldrich), to a concentration of 17 mM, for 10 min. The freshly n-heptylamine plasma polymer (HApp)-functionalized surfaces were covered with activated CMD solution and the reaction was allowed to proceed for 2 h at ambient temperature under agitation. The surfaces were then rinsed in 1 M NaCl solution (2 12 h), rinsed in Milli-Q water (2 12 h) and stored at 4 °C until use. CMD surfaces have been shown to exert excellent low-fouling behaviour towards several cell types and proteins [14,15]. Peptide grafting was performed as described previously [12]. Briefly, GRGDS and GRGES (44-0-23 and 44-0-51, respectively, American Peptide, Sunnyvale, CA, USA) solutions were prepared to a final concentration of 50 lg ml1 in 1 (0.1 M) phosphate-buffered saline (PBS; BP6651, Fisher Scientific, Ottawa, ON, Canada). CMD surfaces were activated in a 200 mM aqueous solution of EDC + NHS, rapidly rinsed in Milli-Q water and then covered in peptide solution. Samples were placed under agitation and the reaction was allowed to proceed for 2 h at ambient temperature followed by rinsing twice in 1 PBS. All peptide grafting was performed under sterile conditions. CDPGYIGSR (87-0-53, American Peptide Company) and fibronectin (human plasma fibronectin; FC010, Millipore, Temecula, CA, USA) were covalently immobilized on CMD-grafted 12-well TCPS plates using maleimide and carbodiimide chemistries, respectively. 2.2. X-ray photoelectron spectroscopy (XPS) Surfaces were analysed using an AXIS ULTRA DLD spectrometer (Kratos Analytical Ltd., Manchester, UK) with a 225 W Al Ka source to validate the presence of the immobilized peptides. Elemental composition was determined with a survey spectrum at a pass energy of 160 eV. CasaXPS (Casa Software Ltd.) was used to calculate the atomic concentrations of each element. Three separate areas of each surface were analysed. 2.3. Cell culture Rat insulinoma cells (INS-1, clone 832/13) [16,17] were cultured in RPMI-1640 with L-glutamine (31800-022, Invitrogen) containing 10% foetal bovine serum (FBS; 12483, Invitrogen), 10 mM HEPES (H3375, Sigma–Aldrich), 1 mM sodium pyruvate (S8636, Sigma–Aldrich) and 50 lM b-mercaptoethanol (M7522, Sigma–Aldrich). The culture medium was changed every 3 days. Upon confluence, cells were trypsinized and seeded onto the prepared surfaces at a density of 50,000 cells ml1. Cells were cultured for 7 days in all conditions. 2.4. Glucose-stimulated insulin secretion (GSIS) A GSIS test was performed with the 7 day cultures. For this purpose, the media were discarded and replaced with a low-glucose Krebs–Ringer buffer with HEPES (KRBH) solution supplemented with 2.8 mM D-glucose. Samples were then incubated for 1 h.
Afterwards, low-glucose KRBH was discarded and samples were rinsed with PBS. KRBH containing 28 mM D-glucose was then added and samples incubated for 1 h. A rinsing step followed. Samples were then incubated for 1 h in KRBH supplemented with 28 mM D-glucose + 50 lM 3-isobutyl-1-methyl-xanthine (IBMX; I5879, Sigma–Aldrich) in KRBH. After a final rinsing step, samples were again incubated in low-glucose KRBH for 1 h. All media were collected and stored at 20 °C for further insulin content measurement. Insulin contents in the culture media were measured using an insulin (rat) high-range enzyme-linked immunosorbent assay kit (0.52 ng ml1 sensitivity; 80-INSRTH-E01, Alpco Diagnostics, Salem, NH, USA). The insulin content was normalized to the cell number for each condition. Absorbance was measured using a Synergy™ HT Multi-Detection Microplate Reader (Bio-Tek, Winooski, VT, USA).
2.5. Immunofluorescence Non-adherent INS-1 cultures (i.e. RGE and CMD) were collected in centrifuge tubes and centrifuged for 5 min at 1200 rpm. Adherent cells (i.e. RGD, CDPGYIGSR, fibronectin and TCPS) were trypsinized (1 min, ambient temperature) and collected in stopping solution (10% FBS in PBS), then centrifuged for 5 min at 1200 rpm. All cell pellets were rinsed with PBS, centrifuged and then fixed in 4% paraformaldehyde for 2 h at 4 °C, followed by 2% agarose and paraffin embedding according to standard protocols [1]. Paraffin sections were cut consecutively at a thickness of 4 lm. Primary antibodies to insulin (ab7842), E-cadherin (ab53033), aVb3 (ab7166), Ki-67 (ab16667), pancreatic and duodenal homeobox 1 (Pdx1, ab47267) were purchased from Abcam (Cambridge, MA, USA). The primary antibody for the a5 integrin (sc-10729) was purchased from Santa Cruz (Santa Cruz, CA, USA). Secondary antibodies were obtained from Invitrogen. All sections were blocked in 10% goat serum and counterstained with 40 ,6-diamidino-2-phenylindole, dihydrochloride (DAPI; D1306, Invitrogen). Manufacturer recommendations were used for pretreatment, dilutions, incubation time and temperature. Samples were imaged using an epifluorescence microscope (Nikon Eclipse).
2.6. Ki-67 and Pdx1 analysis Ki-67 and Pdx1 stained sections, processed as described in Section 2.5, were manually counted from fluorescent images taken at 60 magnification using the Simple PCI software (Hamamatsu Corporation). To estimate the percentage of Pdx1-positive cells, five independent fields of each condition were counted in triplicate. The average total cell number varied between 706 and 885 cells per repeat. Ki-67 sections were treated accordingly. Average cell numbers ranged between 327 and 838 cells per repeat. Data are presented as an average ± standard deviation (SD).
2.7. Assay for assessing cell number Cell numbers were quantified using a CyQUANTÒ assay (C35006, Invitrogen). Prior to carrying out this assay, non-adherent and adherent cells were collected as described in Section 2.5. After centrifugation, the pellets were rinsed and resuspended in Hanks’ balanced salt solution supplied with the assay and centrifuged a second time at 1200 rpm for 1 min. Supernatant was removed and the cell pellet was resuspended in dye solution for 1 h at 37 °C in a 96-well TCPS plate. Fluorescence was measured using a Synergy™ HT Multi-Detection Microplate Reader (Bio-Tek).
621
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
2.8. Statistical analysis Data are presented as mean ± SD for three independent experiments. Differences in results were assessed by a single-factor analysis of variance. A probability (p) value of < 0.05 was considered statistically significant. 3. Results and discussion In this study, low-fouling surfaces with well-defined surface properties were used to graft ECM components in an attempt to compare their effect on proliferation and insulin secretion of INS1 cells. The RGD and CDPGYIGSR peptides as well as fibronectin were covalently bound to low-fouling CMD surfaces, while TCPS served as a positive control. Bare CMD and RGE surfaces were used as negative controls. This approach allows the covalent linking of ECM-derived proteins and peptides, namely fibronectin, RGD and CDPGYIGSR, to low-fouling CMD layers using maleimide and carbodiimide chemistry, thus enabling the study of specific cell– ECM interactions while limiting non-specific interactions and the influence of FBS-derived proteins. The XPS results (Fig. 1 and Table 1) confirmed that RGD, fibronectin, CDPGYIGSR and RGE were covalently linked to the CMD surfaces. For each surface, there was a marked increase in the N/C atomic concentration ratio (Table 1) upon the addition of CDPGYIGSR, fibronectin, RGD and RGE to the CMD layers. This is indicative of the addition of a nitrogen-rich layer, typical of amino acids and proteins, to CMD surfaces. Furthermore, high-resolution C 1s spectra (Fig. 1) showed perturbations at binding energies corresponding to ether (286.5 eV, C–O–H and C–O–C), amide (288 eV, N–C@O) and ester/
Table 1 XPS atomic concentration (%) and atomic ratio of HApp, HApp + CMD, HApp + CMD + CDPGYIGSR, HApp + CMD + fibronectin, HApp + CMD + RGD and HApp + CMD + RGE. Atomic concentration (%)
HApp CMD CDPGYIGSR Fibronectin RGD RGE
Atomic ratio
O 1s
C 1s
N 1s
O/C
N/C
1.8 19.7 20.03 19.0 20.2 19.6
90.8 76.1 74.18 73.2 75.0 76.0
7.4 4.1 5.61 8.1 4.4 4.4
0.02 0.26 0.27 0.26 0.27 0.26
0.08 0.05 0.08 0.11 0.06 0.06
acid (289 eV, O–C@O) groups, also indicative of the addition of a peptide/protein to the CMD layers. Net cell growth had occurred for each culture condition, as shown in Fig. 2 (CyQUANT assay), Fig. 3 (percentage of Ki-67-positive cells) and supplementary Fig. A1 (Ki-67-positive cells). Attached INS-1 cells showed higher net cell growth rates in general when compared to the postulated 100 h doubling time [18]. It was not surprising to obtain higher cell numbers on bioactivated surfaces in comparison with RGE and bare CMD surfaces. Cell net growth on fibronectin measured with the CyQUANT assay was significantly higher than on all other surfaces (p < 0.03) except RGD (p > 0.05). Due to the cell–cell contacts [19] within the INS-1 clusters formed in suspension on CMD or loosely bound on RGE, proliferation still occurred despite the low-fouling nature of the CMD and RGE surfaces. The largest numbers of cells positively stained for Ki-67 were found on CDPGYIGSR and TCPS (p < 0.004). One reason for this
Fig. 1. High-resolution XPS C 1s spectra of CDPGYIGSR, fibronectin, RGD and RGE surfaces and interlayers.
622
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
*
INS-1 cell number
3,5E+05 3,0E+05 2,5E+05
*
2,0E+05 1,5E+05 1,0E+05 5,0E+04 0,0E+00 CMD
RGE
RGD
CDPGYIGSR Fibronectin
TCPS
Ki-67 positive cells (%)
Fig. 2. INS-1 cell number after 7 days of culture on the different surfaces. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
*
60 50 40 30 20 10 0 CMD
RGE
TCPS
Fibronectin
RGD
CDPGYIGSR
Fig. 3. Percentage of INS-1 cells stained positively for Ki-67 after 7 days of culture on the different surfaces. Total cell numbers were obtained from DAPI staining. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
could be that cells on fibronectin and RGD were grown close to confluence, and were therefore in a lag phase of growth. Despite that, more cells stained positively for Ki-67 on fibronectin and RGD than on CMD and RGE (p < 0.005 and p < 0.005, respectively). Additionally, Ki-67 is a proliferation marker found in all stages of the cell cycle except G0 [20]. Consequently, cells staining positively for Ki-67 are not necessarily undergoing mitosis. When INS-1 cells were exposed to RGD, CDPGYIGSR, fibronectin and TCPS surfaces, they adhered to the surface tightly and began forming well-defined surface-bound clusters on RGD, CDPGYIGSR and fibronectin. Conversely, INS-1 cells were non-adherent when seeded on bare CMD and adhered only loosely on RGE, forming ‘‘islet-like’’ aggregates in both conditions (Fig. 4). These findings indicate that INS-1 adherent cultures use their ECM receptors to bind to the specific surface ligands. Consequently, a5 and aVb3 integrin expression was assessed. The a5-subunit binds exclusively to b1integrin allowing an estimation of the a5b1-expression by staining for a5 only [21]. In previous studies, a5 expression has been observed during foetal rat pancreatic development and b1 expression, when suppressed in adult rat pancreatic beta cells, was found to be pivotal for survival and migration of rat beta cells on 804G matrix [10,11,22]. In contrast to those findings, a5 was expressed on all surfaces. This is surprising as cells on bare CMD and RGE surfaces have no access to ECM cues. Due to the low-fouling properties of both CMD and RGE, cells cannot attach and therefore stay in suspension. In the non-adherent CMD and RGE systems, the a5 and aVb3 integrins appeared to be expressed predominantly on those cells residing in the core of the aggregates, whereas on adherent surfaces they were expressed throughout, as shown in Fig. 5 (see also supplementary Fig. A2, 100 magnification). One possible explanation for this strong expression could be the presence of ECM molecules synthesized by the INS-1 cells into the interstitial space of the aggregates. FBS-associated proteins are not able to coat the CMD-covered surfaces as these surfaces are low fouling.
The binding of INS-1 cells to RGE, but not CMD (Fig. 4), is most probably due to unspecific INS-1 surface receptor–RGE interactions. If the TCPS plate on which the CMD was plated and RGE was immobilized was only slightly agitated, INS-1 cell–RGE aggregates detached immediately. Also, cell aggregates on RGE were morphologically closer to floating INS-1 aggregates on CMD than to the surface-bound cell clusters on RGD, fibronectin, CDPGYIGSR and TCPS. To ensure that differences between the CMD and RGE surfaces were studied correctly, the detachment of cell aggregates on RGE had been avoided. On CDPGYIGSR, RGD and fibronectin the expression of a5 seemed to be elevated as the intensity of the staining was more pronounced when compared to the other conditions. Both a5 and aVb3 integrins have been identified as RGD-binding integrins in pancreatic tissues [1,10]. Fibronectin contains multiple RGD motifs [3]. Therefore, INS-1 cells express integrins that bind to both RGD and fibronectin. In the case of CDPGYIGSR, the positive integrin staining is most probably due also to cell-derived ECM after a 7 day culture. Additionally, the YIGSR motif interacts specifically with the 67 kDa laminin receptor, which has not been assessed in this study [23]. Cells on RGD and fibronectin as well as those on TCPS all expressed the aVb3 integrin. It has previously been described that integrin aVb3 plays a role in cell adhesion during pancreatic development, though it was recently shown to be less significant than other aV integrins in foetal human islet preparations. Interestingly, the adult islet expression levels of aVb3 were the same as those of foetal islets [24]. Our findings partially support those results as INS-1 cell expression of aVb3 was overall weaker when compared to the a5 integrin (Fig. 5). In contrast, our results suggest that aVb3 plays a role in INS-1 adhesion as it was expressed in all conditions. E-cadherin was shown to be essential for the islet formation and recognition of the different cell types within the islet as it is expressed in b-cells only. The cadherin family are Ca2+-dependent cell anchors which bind to cadherin homologues of neighbouring cells [25]. Here, b-cell E-cadherin expression is consistent with previous studies. E-cadherin is apparent and convincingly expressed throughout all conditions. Similar expression was detected on clustered cells grown on CDPGYIGSR, RGD and fibronectin, and also in the CMD and RGE aggregates, which are both shown in Fig. 6. CMD and RGE were selected to show the positive E-cadherin staining in general and to highlight the finding of E-cadherin expression in the non-adherent cultures. The cells on bare CMD surfaces, which stayed in suspension, and the loosely bound cells on RGE most probably established cell–cell contacts to avoid detachmentinduced apoptosis, known as anoikis [26]. As long-term cultures of insulin-producing cells are of great interest for further studies and transplantation purposes, 7 day cultures were chosen. To examine the possibility of cell-derived ECM deposition and alteration of the findings of this study, 3 day cultures were also stained for a5 and aVb3 integrins and E-cadherin. No differences in expression levels between the 3 and 7 day cultures could be found, as shown in supplementary Figs. A3 (a5 and aVb3 integrin expression) and A4 (E-cadherin expression). INS-1 cells were stained for insulin and glucagon to confirm the purity of the cell line and to ensure the viability and function of the insulin-secreting cells. INS-1 cells expressed insulin but not glucagon after 7 days of culture throughout all conditions (supplementary Fig. A5). Previous studies using insulinoma cells found pseudo-islet formation in suspension. The insulin secretion of the pseudo-islets was negatively influenced by blocking E-cadherin [27]. In another study, pancreatic precursor cells, not detached from their ECM, formed so-called ‘‘islet-like’’ cell aggregates [28]. Both these findings might help to explain the insulin expression and secretion seen on bare CMD and RGE surfaces supported by the E-cadherin-mediated cell–cell adhesion.
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
623
Fig. 4. Phase-contrast microscopy pictures of INS-1 cells grown on the different surfaces after 72 h. 20 magnification. Bars correspond to 100 lm.
While all surface conditions resulted in glucose-stimulated insulin secretion (Fig. 7), INS-1 cells on fibronectin and RGD secreted remarkably higher levels of insulin in comparison with the other culture conditions tested after stimulation with high glucose (p < 0.04) and 50 lM IBMX (p < 0.005). When stimulated with high glucose only, cells cultured on CDPGYIGSR did not secrete more insulin than cells cultured on TCPS. Inversely, when incubated with high glucose plus IBMX, cells grown on CDPGYIGSR secreted significantly more insulin than cells cultured on TCPS (p < 0.006). It has been shown that isolated human islets cultured in the presence of fibronectin, laminins, and collagens I and IV are able to maintain their abilities to adhere, survive and function in an improved manner in vitro [29]. Hence, the higher insulin secretion on fibronectin and RGD is consistent with previous studies [1,26]. The higher secretion of insulin on the ECM surfaces is most probably due to the resulting cell–ECM interactions [1]. Surprisingly, cells on RGE secreted approximately the same amount of insulin into the medium as cells on TCPS when normalized to the cell number. INS-1 aggregates found on RGE did not secrete significantly more insulin than those on CMD. INS-1 cells cultured on CMD and RGE both formed ‘‘islet-like’’ clusters
(Fig. 4). Therefore it is possible to compare them to MIN-6 pseudo-islets in both function and proliferation behaviour as they maintain their ability to secrete insulin in response to glucose [27]. Recent studies have shown that cancerous or transformed beta cell lines with high proliferation, e.g. INS-1 or MIN-6 cells, produce and secrete less insulin than primary beta cells [30]. The relatively low insulin secretion of cells grown on CDPGYIGSR when stimulated with 28 mM glucose could be explained by the observed high proliferation rates on CDPGYIGSR, as revealed from cells staining positively for Ki-67. This is an agreement with a previous study showing that insulin secretion from an immortalized human beta cell line can be affected by the cell proliferation state [31]. INS-1 cells on CMD and RGE with the lowest levels of positive Ki-67 staining secreted relatively less insulin than INS cells in all other conditions. It is possible that cell aggregates on CMD and RGE were negatively influenced by the lack of cell–ECM support, thus showing less integrin expression compared to the other conditions. Additionally cells in all conditions expressed Pdx1 (supplementary Fig. A6). The relative numbers of cells grown on fibronectin, RGD, CDPGYIGSR and TCPS and expressing Pdx1 were larger than those on CMD and RGE (p < 0.002 and p < 0.02, respectively;
624
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
Fig. 5. Expression of integrins a5 (red) and aVb3 (green) of INS-1 cells cultured for 7 days on the different surfaces. Nuclei are stained with DAPI (blue). 60 magnification. Bars correspond to 50 lm.
Fig. 6. Immunofluorescence staining for E-cadherin of INS-1 cells cultured for 7 days on CMD and RGE. Nuclei are DAPI stained in blue. 60 magnification. Bars correspond to 50 lm.
Fig. 8). The transcription factor Pdx1 is essential for pancreatic development and maintenance of beta cell function and survival [32]. The interaction between insulin and Pdx1 seems to prevent islets from undergoing apoptosis [33]. Our results suggest that
cell–ECM interactions promoted by the surface-immobilized fibronectin, RGD and CDPGYIGSR enhance insulin secretion and the number of cells expressing Pdx1 via integrin-receptor activation. This is supported by a study conducted by Krishnamurthy et al.
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
Insulin ng/106 cells
250
625
Appendix A. Supplementary data
*
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.actbio.2011.10.036.
200
150
Appendix B. Figures with essential colour discrimination
100
Certain figures in this article, particularly Figures 5 and 6, is difficult to interpret in black and white. The full colour images can be found in the on-line version, at doi:10.1016/j.actbio.2011.10.036.
50
0 CMD
RGE low glucose
TCPS highglucose
RGD
CDPGYIGSR
highglucose+ IBMX
Fibronectin
low glucose
Pdx1 positive cells (%)
Fig. 7. GSIS of INS-1 cells cultured on different surfaces for 7 days. Insulin secretion is normalized per 106 cells. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
*
100 90 80 70 60 50 40 30 20 10 0
CMD
RGE
TCPS
Fibronectin
RGD
CDPGYIGSR
Fig. 8. Percentage of INS-1 cells stained positively for Pdx1 after 7 days of culture on the different surfaces. Total cell numbers were obtained from DAPI staining. Results represent the mean ± SD of three independent experiments (⁄p < 0.05).
[17] showing that perturbation of integrin function results in decreased human islet and INS-1 cell proliferation, Pdx1 expression and decreased insulin expression and secretion. 4. Conclusions Globally, we have demonstrated that INS-1 cells cultured on substrates bearing low-fouling CMD layers, on which fibronectin, RGD and CDPGYIGSR were covalently linked, exhibited higher insulin secretion and cell proliferation compared to the CMD, RGE and TCPS controls. These well-defined surfaces could be used for further characterization of ECM–beta cell interactions, thus leading to a better understanding of beta cell function and subsequently diabetes and pancreatic development. Additionally, the surfaces could be tested with isolated human islets to investigate their effect over islet survival and insulin secretion in response to glucose. Acknowledgements This research project was supported by the Université de Sherbrooke and the National Science and Engineering Research Council of Canada (NSERC) through a Discovery Grant. We also thank Dr. Rennian Wang from the Child Health Research Institute and Department of Physiology & Pharmacology (The University of Western Ontario, London, ON, Canada) for her insight on the culture of INS-1 cells and immunofluorescence protocols. The technical assistance of Pascale Roy is recognized. We appreciate the contribution of Jamie Sharp who provided language support in proof reading the article.
References [1] Wang RN, Rosenberg L. Maintenance of beta-cell function and survival following islet isolation requires re-establishment of the islet-matrix relationship. J Endocrinol 1999;163:181–90. [2] Hulinsky I, Cooney S, Harrington J, Silink M. In vitro growth of neonatal rat islet cells is stimulated by adhesion to matrix. Hormon Metab Res 1995;27:209–15. [3] Zhang Z, Vuori K, Reed JC, Ruoslahti E. The alpha 5 beta 1 integrin supports survival of cells on fibronectin and up-regulates Bcl-2 expression. Proc Natl Acad Sci USA 1995;92:6161–5. [4] Johnson KJ, Sage H, Briscoe G, Erickson HP. The compact conformation of fibronectin is determined by intramolecular ionic interactions. J Biol Chem 1999;274:15473–9. [5] Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion. Cell 1992;69:11–25. [6] Burgeson RE, Chiquet M, Deutzmann R, Ekblom P, Engel J, Kleinman H, et al. A new nomenclature for the laminins. Matrix Biol 1994;14:209–11. [7] Weber LM, Hayda KN, Haskins K, Anseth KS. The effects of cell–matrix interactions on encapsulated beta-cell function within hydrogels functionalized with matrix-derived adhesive peptides. Biomaterials 2007;28:3004–11. [8] Ruoslahti E. RGD and other recognition sequences for integrins. Annu Rev Cell Dev Biol 1996;12:697–715. [9] Pinske GGM, Bouwman WP, Jiawan-Lalai R, Terpstra OT, Bruijn JA, de Heer E. Integrin signaling via RGD peptides and anti-b1 antibodies confers resistance to apoptosis in islets of Langerhans. Diabetes 2006;55:312–7. [10] Wang R, Li J, Lyte K, Yashpal NK, Fellows F, Goodyer CG. Role for beta1 integrin and its associated alpha3, alpha5, and alpha6 subunits in development of the human fetal pancreas. Diabetes 2005;54:2080–9. [11] Hammar E, Parnaud G, Bosco D, Perriraz N, Maedler K, Donath M, et al. Extracellular matrix protects pancreatic beta-cells against apoptosis: role of short- and long-term signaling pathways. Diabetes 2004;53:2034–41. [12] Monchaux E, Vermette P. Bioactive microarrays immobilized on low-fouling surfaces to study specific endothelial cell adhesion. Biomacromolecules 2007;8:3668–73. [13] Martin Y, Boutin D, Vermette P. Study of the effect of process parameters for nheptylamine plasma polymerization on final layer properties. Thin Solid Films 2007;515:6844–52. [14] Dubois J, Gaudreault C, Vermette P. Biofouling of dextran-derivative layers investigated by quartz crystal microbalance. Colloids Surf B 2009;71:293–9. [15] Monchaux E, Vermette P. Cell adhesion resistance mechanisms using arrays of dextran-derivative layers. J Biomed Mater Res Part A 2008;85A:1052–63. [16] Krishnamurthy M, Li J, Al-Masri M, Wang R. Expression and function of ab1 integrins in pancreatic beta (INS-1) cells. J Cell Commun Signal 2008;2:67–79. [17] Krishnamurthy M, Li J, Fellows GF, Rosenberg L, Goodyer CG, Wang R. Integrin a3, but not b1, regulates islet cell survival and function via PI3K/Akt signaling pathways. Endocrinology 2011;152:424–35. [18] Asfari M, Janjic D, Meda P, Li G, Halban PA, Wollheim CB. Establishment of 2mercaptoethanol-dependent differentiated insulin-secreting cell lines. Endocrinology 1992;130:167–78. [19] Carvell MJ, Marsh PJ, Persaud SJ, Jones PM. E-cadherin interactions regulate bcell proliferation in islet-like structures. Cell Physiol Biochem 2007;20:617–26. [20] Köhler CU, Kreuter A, Rozynkowski MC, Rahmel T, Uhl W, Tannapfel A, et al. Validation of different replication markers for the detection of beta-cell proliferation in human pancreatic tissue. Regul Pept 2010;162:115–21. [21] Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell 2002;110:673–87. [22] Yashpal NK, Li J, Wheeler MB, Wang R. Expression of beta1 integrin receptors during rat pancreas development-sites and dynamics. Endocrinology 2005;146:1798–807. [23] Graf J, Ogle RC, Robey FA, Sasaki M, Martin GR, Yamada Y, et al. A pentapeptide from the laminin B1 chain mediates cell adhesion and binds the 67,000 laminin receptor. Biochemistry 1987;26:6896–900. [24] Kaido T, Perez B, Yebra M, Hill J, Cirulli V, Hayek A, et al. Alphav-integrin utilization in human beta-cell adhesion, spreading, and motility. J Biol Chem 2004;279:17731–7. [25] Jain R, Lammert E. Cell–cell interactions in the endocrine pancreas. Diabetes Obes Metab 2009;11:159–67.
626
C. Kuehn et al. / Acta Biomaterialia 8 (2012) 619–626
[26] Weber LM, Hayda KN, Anseth KS. Cell–matrix interactions improve beta-cell survival and insulin secretion in three-dimensional culture. Tissue Eng Part A 2008;14:1959–68. [27] Rogers GJ, Hodgkin MN, Squires PE. E-cadherin and cell adhesion: a role in architecture and function in the pancreatic islet. Cell Physiol Biochem 2007;20:987–94. [28] Hardikar AA, Marcus-Samuels B, Geras-Raaka E, Raaka BM, Gershengorn MC. Human pancreatic precursor cells secrete FGF2 to stimulate clustering into hormone-expressing islet-like cell aggregates. Proc Natl Acad Sci USA 2003;100:7117–22. [29] Daoud J, Petropavlovskaia M, Rosenberg L, Tabrizian M. The effect of extracellular matrix components on the preservation of human islet function in vitro. Biomaterials 2010;31:1676–82.
[30] Reers C, Hauge-Evans AC, Morgan NG, Wilcox A, Persaud SJ, Jones PM. Downregulation of proliferation does not affect the secretory function of transformed b-cell lines regardless of their anatomical configuration. Islets 2011;3:80–8. [31] Narushima M, Kobayashi N, Okitsu T, Tanaka Y, Li SA, Chen Y, et al. A human bcell line for transplantation therapy to control type 1 diabetes. Nat Biotechnol 2005;23:1274–82. [32] Bernardo AS, Hay CW, Docherty K. Pancreatic transcription factors and their role in the birth, life and survival of the pancreatic beta cell. Mol Cell Endocrinol 2008;294:1–9. [33] Johnson JD, Bernal-Mizrachi E, Alejandro EU, Han Z, Kalynyak TB, Li H, et al. Insulin protects islets from apoptosis via Pdx1 and specific changes in the human islet proteome. Proc Natl Acad Sci USA 2006;103:19575–80.