The triangle, in memory of Prof. Sung Wan Kim

Professor Sung Wan Kim, my dear friend, passed away on February 24, 2020, at the age of 79 in Salt Lake City, USA. I met Sung Wan for the first time in 1972, when I went with my family to Utah to work with Prof. Donald Lyman and Prof. Willem (Pim) Kolff in a time when the famous Utah Artificial Heart Project was active. Later, I returned and was appointed first as Visiting Associate Professor in the Dept. of Applied Pharmaceutical Science, College of Pharmacy, University of Utah and later as Adjunct Professor of Pharmaceutics working closely together with Sung Wan. In the 1970s, Prof. Sung Wan Kim, Prof. Jim Anderson (Case Western University, USA) and I had several intensive discussions and developed many new concepts for drug-polymer synthesis, controlled release systems, and experimental techniques that became the early foundation for the future Triangle Collaboration. Already in 1975, Wim Sederel went from The Netherlands to Case Western for a period of 1.5 years to work on (co)poly(α-amino acids) as potential biomaterials. The biennial series of International Symposia on Recent Advances in Drug Delivery Systems in Salt Lake City (the Utah Meetings) organized by Sung Wan and Jim Anderson provided a further impetus to the Triangle collaboration by catalyzing international scientific relationships. The proceedings of these meetings were published in the J. Control. Release, founded by Dr. Jorg Heller and myself. The Triangle collaboration later developed into an international research alliance between the University of Utah, University of Twente, Tokyo Women's Medical College, and University of Tokyo. More than 40 Ph.D. students and post-doctoral scientists, and sometimes research staff members, were exchanged between these participating laboratories, resulting in close to 50 joint research publications. In general, the PhD researchers would spend at least three months at one of the Triangle laboratories and could then use their special facilities and collaborative relationships to mutual benefit. Each researcher had to submit a work plan in advance for Triangle supervisor approval. In the late seventies little funding was available for international exchange of students. To cover any additional costs for the students, such as travel and accommodation, we used our own consultancy income, which was transferred into a non-profit foundation. Later, Prof. Yong-Hee Kim (Hanyang University, Korea) was also included in the network. Of the participating students, more than ten were later appointed as professors. The PhD students greatly appreciated our exchange program, which offered them the possibility to perform multidisciplinary research, to work within a different research team, and to become familiar with cultural aspects of the host country. Sung Wan and I have very good memories of these times, although there were also a few ‘unfortunate’ incidents with these exchanges. For example, one of our young researchers (Ivan Vulić) was arrested in Salt Lake City, because he joined other students in starting his own personal collection of city traffic signs. In court, he haggled with the judge over the fine. He surprised his American friends, when he succeeded to have his fine reduced after referring to his very limited financial resources. Such typical Dutch entrepreneurship was unheard of in Salt Lake City! Another Japanese exchange student (Naoki Negishi) was planning to come to Salt Lake City. Based on his experience, Sung Wan and I decided that it would be better to send him to my lab in Holland. After a last-minute change to his flight plan, he arrived safe and sound in Amsterdam and managed to find his way to our University. In Enschede, we provided him with an old but still working Russian Lada automobile, also used by former students. One day he called my wife Hilda telling that his Lada had a flat tire. After some confusion – his English was not particularly good – Hilda was able to explain to him that the Lada was equipped with a kind of bicycle pump which could be used to re-inflate the car tire. Subsequently, he managed to reach a local garage. You Han Bae and his family went to Bruges (Belgium) for a weekend and sought to stay in a hotel for one night. After numerous unsuccessful efforts to find a hotel in Bruges, they decided to drive back to Enschede. On their way back, in the middle of the night, their old car broke down completely and they were finally rescued by my PhD student, Harry Cremers. These are only a few stories, but I shall refrain from telling more, since the memories still make me somewhat nervous. Fortunately, many of our Triangle PhD exchange participants were able to publish one or more good papers. I have received numerous enthusiastic reactions from former Dutch students who worked with Sung Wan and his group about the great times they had enjoyed in Salt Lake City. I have to mention here the continuous support of Janny and Prof. Piet Dijkstra, who helped the exchange students to settle down and to feel at home in The Netherlands. In this paper, I will briefly address the research contributions of the participants in the Triangle program. I have attempted to track down all the participants, but I do


Introduction
Professor Sung Wan Kim, my dear friend, passed away on February 24, 2020, at the age of 79 in Salt Lake City, USA. I met Sung Wan for the first time in 1972, when I went with my family to Utah to work with Prof. Donald Lyman and Prof. Willem (Pim) Kolff in a time when the famous Utah Artificial Heart Project was active. Later, I returned and was appointed first as Visiting Associate Professor in the Dept. of Applied Pharmaceutical Science, College of Pharmacy, University of Utah and later as Adjunct Professor of Pharmaceutics working closely together with Sung Wan. In the 1970s, Prof. Sung Wan Kim, Prof. Jim Anderson (Case Western University, USA) and I had several intensive discussions and developed many new concepts for drug-polymer synthesis, controlled release systems, and experimental techniques that became the early foundation for the future Triangle Collaboration. Already in 1975, Wim Sederel went from The Netherlands to Case Western for a period of 1.5 years to work on (co)poly(α-amino acids) as potential biomaterials. The biennial series of International Symposia on Recent Advances in Drug Delivery Systems in Salt Lake City (the Utah Meetings) organized by Sung Wan and Jim Anderson provided a further impetus to the Triangle collaboration by catalyzing international scientific relationships. The proceedings of these meetings were published in the J. Control. Release, founded by Dr. Jorg Heller and myself. The Triangle collaboration later developed into an international research alliance between the University of Utah, University of Twente, Tokyo Women's Medical College, and University of Tokyo. More than 40 Ph.D. students and post-doctoral scientists, and sometimes research staff members, were exchanged between these participating laboratories, resulting in close to 50 joint research publications. In general, the PhD researchers would spend at least three months at one of the Triangle laboratories and could then use their special facilities and collaborative relationships to mutual benefit. Each researcher had to submit a work plan in advance for Triangle supervisor approval. In the late seventies little funding was available for international exchange of students. To cover any additional costs for the students, such as travel and accommodation, we used our own consultancy income, which was transferred into a non-profit foundation. Later, Prof. Yong-Hee Kim (Hanyang University, Korea) was also included in the network. Of the participating students, more than ten were later appointed as professors.
The PhD students greatly appreciated our exchange program, which offered them the possibility to perform multidisciplinary research, to work within a different research team, and to become familiar with cultural aspects of the host country. Sung Wan and I have very good memories of these times, although there were also a few 'unfortunate' incidents with these exchanges. For example, one of our young researchers (Ivan Vulić) was arrested in Salt Lake City, because he joined other students in starting his own personal collection of city traffic signs. In court, he haggled with the judge over the fine. He surprised his American friends, when he succeeded to have his fine reduced after referring to his very limited financial resources. Such typical Dutch entrepreneurship was unheard of in Salt Lake City! Another Japanese exchange student (Naoki Negishi) was planning to come to Salt Lake City. Based on his experience, Sung Wan and I decided that it would be better to send him to my lab in Holland. After a last-minute change to his flight plan, he arrived safe and sound in Amsterdam and managed to find his way to our University. In Enschede, we provided him with an old but still working Russian Lada automobile, also used by former students. One day he called my wife Hilda telling that his Lada had a flat tire. After some confusion -his English was not particularly good -Hilda was able to explain to him that the Lada was equipped with a kind of bicycle pump which could be used to re-inflate the car tire. Subsequently, he managed to reach a local garage. You Han Bae and his family went to Bruges (Belgium) for a weekend and sought to stay in a hotel for one night. After numerous unsuccessful efforts to find a hotel in Bruges, they decided to drive back to Enschede. On their way back, in the middle of the night, their old car broke down completely and they were finally rescued by my PhD student, Harry Cremers. These are only a few stories, but I shall refrain from telling more, since the memories still make me somewhat nervous.
Fortunately, many of our Triangle PhD exchange participants were able to publish one or more good papers. I have received numerous enthusiastic reactions from former Dutch students who worked with Sung Wan and his group about the great times they had enjoyed in Salt Lake City. I have to mention here the continuous support of Janny and Prof. Piet Dijkstra, who helped the exchange students to settle down and to feel at home in The Netherlands. In this paper, I will briefly address the research contributions of the participants in the Triangle program. I have attempted to track down all the participants, but I do apologize when I was not able to find the details of some Triangle exchange students. For instance, the first student who came to Holland from Salt Lake City was Eun-Soo Lee at the end of the 1970s. He worked on the role of protein adsorption in the blood-material interaction of polymer surfaces, but I do not have further details from his stay.

Characterization of polymer surfaces for bacterial and endothelial adhesion studies
One of the earliest students who went to Salt Lake City was André Hogt (1983). With Sung Wan's group and Utah Bioengineering colleagues, Profs. Andrade and Gregonis, he worked on polymers and copolymers of different methacrylates, which were later used to perform model studies on bacterial adhesion and growth together with Jaap Dankert and studies on the adhesion of cultured endothelial cells carried out by Paulien van Wachem.
The surface properties of glass slides coated with films of methyl methacrylate (MMA) and hydroxyethyl methacrylate (HEMA) copolymers were varied both by MMA/HEMA content and the introduction of charged methacrylate monomers (anionic methacrylic acid or cationic trimethylaminoethyl methacrylate). Surfaces were characterized by water contact angle (CA) and ζ potential measurements. Most remarkable result of water contact angle measurements was that the advancing contact angles of HEMA copolymers increased with an increased level of charged units, probably because they reoriented at the water-airpolymer three-phase line, becoming buried inside the hydrogel when exposed to air. In addition, receding angles of a separate series of poly (alkyl methacrylates) were found to first increase and then decrease with increasing alkyl side-chain lengths, which was ascribed to a change in mobility of the longer alkyl methacrylate polymers related to their lower glass transition temperature. ζ potentials of the MMA and HEMA copolymers were more negative or less negative and positive in accordance with the level of negatively or positively charged monomers incorporated, respectively [1].

Blood-contacting polymer test devices
In 1980 Jan Olijslager went to Salt Lake City to cooperate with the Utah group in the design and application of new test devices to evaluate the blood compatibility of polymer surfaces under flow. Platelet adhesion and activation did not only depend on the type of material used but were also strongly influenced by the flow conditions applied [2]. These conditions were very relevant to the pioneering blood-contacting biomaterials efforts of the Kolff-led Utah Artificial Heart program receiving global attention at that time.

Heparin-releasing polymers
After the return of Jan Olijslager, Jim McRea came to The Netherlands in 1981 to work on the loading and release of heparin from polyurethanes. The aim was to develop a monolithic drug delivery formulation of heparin based on Biomer™, an early commercial polyurethane used in the Utah totally implantable artificial hearts and numerous other blood pumps. Drug release was measured utilizing Azure A dye as a quantitative assay. Further characterization of this system and its release behavior were established with TGA (Thermo-Gravimetric Analysis), DSC (Differential Scanning Calorimetry) PTT (Pro Thrombin Time, bioassay) and contact angle (CA) measurements [3]. To illustrate the early cultural results of the international exchange, Dr. McRea provided the following quote: "I was picked up after arrival in Amsterdam by Jan Olijslager

Plasma-induced surface modification
Utah's Ki Dong Park came to Holland in 1990 and worked on plasma-induced surface modification of polymers. Modified surfaces were characterized and protein adsorption on the modified surface was evaluated.

Albumin-heparin conjugates
In the 1980's Wim Hennink (1982) and Lou Dost went to Salt Lake City to study the synthesis and properties of human serum albuminheparin (Alb-Hep) conjugates. These conjugates were applied to improve the blood-material interaction of polymer surfaces either by preadsorption or by covalent coupling. In 1982 Charles Ebert came to Twente to participate in the same program. This resulted in a number of joint publications [4][5][6][7][8]. Alb-Hep conjugates were prepared by condensing Alb with Hep using 1-ethyl-3-(dimethylaminopropyl)-carbodiimide (EDC) and conjugates were subsequently purified by chromatography. Thrombin time (TT), inhibition of Factor Xa and activated partial thromboplastin time (APTT) assays were used to compare the activity of Hep in Alb-Hep conjugates with that of Hep used for the conjugate synthesis. Broad-molecular weight distributions were indicated by GPC. Alb-Hep conjugates were further fractionated using immobilized antithrombin III (AT III) to provide high and low affinity Alb-Hep (Alb-Hep-HA) and Alb-Hep (Alb-Hep-LA), respectively. Polystyrene (PS) surfaces were first pre-adsorbed with either Alb or Alb-Hep and the adsorption of AT III onto these surfaces was evaluated using a two-step enzyme immuno-assay. The highest AT III adsorption values (ATIII buffer solutions) were measured for Alb-Hep-HA pretreated surfaces. Using plasma or plasma diluted with buffer, AT III was only detected on surfaces pre-adsorbed with Alb-Hep-HA or non-fractionated Alb-Hep. It was concluded that the Hep moiety of the adsorbed conjugate is directly exposed to the solution phase whereas Alb interacts with the PS surface. Alb-Hep conjugate was also coated onto glass, PVC, Biomer™ and cellulose acetate. The adsorption and desorption behavior of Alb-Hep on glass was studied applying radiolabeled ( 3 H and 51 Cr) conjugates. Materials pre-coated with Alb-Hep showed a significant prolongation of the Lee-White clotting time as compared with noncoated materials. The prolonged clotting time for glass pre-adsorbed with Alb-Hep was due to surface-bound conjugate. Also, prolonged recalcification times of plasma exposed to glass, Biomer™, and PVC, preadsorbed with Alb-Hep were measured. Glass pre-coated with Alb-Hep-HA showed longer clotting times than glass pre-coated with Alb-Hep-LA. It is therefore concluded that the mechanism of action of Hep in pre-adsorbed conjugates is comparable to that of heparin in solution.

Heparin-containing block copolymers
In another program, we aimed at synthesizing heparin (Hep)-containing block-copolymers, where the hydrophobic block interacts with a hydrophobic polymer support and the hydrophilic Hep block with hydrophilic spacer extends into the plasma or blood to limit thrombosis. Ivan Vulić, went to Salt Lake City in 1985, and Teruo Okano first visited Salt Lake City and also came to Twente later. Ivan worked closely together with Teruo on the design and synthesis of new block copolymers composed of polystyrene (PS), a spacer-block of poly(ethylene oxide) (PEO) and a bioactive block of Hep. Block copolymers were prepared by first reacting PS containing a terminal amino group with toluene 2,4diisocyanate (TDI) followed by a reaction with amino-telechelic PEO to produce AB type block copolymers terminated with an amino group. Coupling of PS-PEO-NH 2 with Hep was performed by first activating carboxylic groups of Hep with EDC (pH 5.1-5.2) followed by a reaction with PS-PEO-NH 2 at pH 7.5. Depending on the composition of the di-block copolymer used, 25-29% w/w Hep could be incorporated [9]. Subsequently, an alternative procedure for the preparation of the PS-PEO-Hep block copolymers was developed. Amino-semi-telechelic PS was synthesized by anionic polymerization of styrene. The terminal amino group was coupled with amino-telechelic PEO using TDI to produce amino-semitelechelic PS-PEO di-block copolymer (PS-PEO-NH 2 ). PS-PEO-Hep triblock copolymer was also prepared by a coupling reaction of PS-PEO-NH 2 with nitrous acid-degraded Hep containing aldehyde groups via reductive amination at pH 7. The PS-PEO-Hep copolymers contained 18-32% w/w Hep corresponding to about 1 Hep molecule per PS-PEO chain [10]. PS-PEO-Hep block copolymers, were coated on either aluminum, glass, polydimethylsiloxane (PDMS), PS or Biomer™ substrates. Coated films and surfaces were characterized by transmission electron microscopy (TEM), contact angle measurements (CA) and X-ray photoelectron spectroscopy (XPS). It was shown that thin coated films of PS-PEO and PS-PEO-Hep block copolymers had a heterogeneous micro-phase separated structure. Coatings of PS-PEO and PS-PEO-Hep block copolymers became more hydrophilic during immersion in water. PS-PEO and PS-PEO-Hep block copolymers films enriched in PEO in the top layers of the coatings (XPS), which was more pronounced for hydrated surfaces. Relatively small amounts of heparin were detected at the surface of PS-PEO-Hep block copolymer coatings [11]. PS-PEO-Hep were also coated either onto glass, PDMS, polyurethane or PS substrates. The blood-material interaction of coated surfaces was evaluated by determining surface-bound Hep activity, adsorption of ATIII, plasma re-calcification time assays, adhesion of platelets and by using an ex vivo rabbit A-A shunt model. Hep was shown to be present at the surfaces of all PS-PEO-Hep-coated materials and coated surfaces interacted with AT III and thrombin to prevent blood clot formation. The highest Hep surface activity found was 5.5 × 10 −3 U cm −2 . PS-PEO-Hep coated surfaces induced a significant prolongation of plasma re-calcification times as compared to controls. Platelets reacted only minimally with PS-PEO-Hep coatings and seemed to pacify the surfaces. In ex vivo A-A shunt experiments, which were carried out under low flow and low shear conditions, PS-PEO-Hep coatings exhibited prolonged occlusion times, indicating that PS-PEO-Hep coatings reduce thrombus formation [12]. David Grainger came to the Netherlands in 1986 and continued with the strategy to use heparin containing block copolymers to improve the blood-compatibility of various surfaces [13]. Amphiphilic block copolymers of polydimethylsiloxane, PEO and Hep (PDMS-PEO-Hep) were prepared via coupling reactions using functionalized pre-polymers, di-isocyanates, and amino-derivatized Hep. Commercial sodium Hep as well as Hep degraded by nitrous acid were converted into their benzyltrimethyl ammonium salts to enhance their organic solvent solubility so to covalently couple Hep in organic solution with the semi-telechelic copolymers. Glass beads and tubes were coated with PDMS-PEO-Hep block copolymers and subsequently tested for surface-immobilized Hep activity. Solvent-cast PDMS-PEO-Hep films have a heterogeneous microphase-separated structure (TEM) and surface restructuring occurs upon hydration (Wilhelmy plate contact angle measurements). PDMS-PEO-Hep cast surfaces revealed significant surface concentrations of Hep (toluidine blue assay). Re-calcification times, thrombin times, and Factor Xa assays showed that part of the surface-bound Hep maintains its bioactivity. In vitro platelet adhesion and activation assays using PDMS-PEO-Hep coated beads and rabbit platelet-rich plasma indicated that PDMS-PEO-Hep coatings induced low levels of platelet adhesion and serotonin release. These results were correlated with Hep surface concentrations to determine the efficacy of PDMS-PEO-Hep as a bloodcompatible material or coating [14].

Controlled release of prostaglandins
In order to reduce the contact activation of platelets on polyurethane surfaces, John Lin, Lou Dost and Teruo Okano in 1986 designed a system for the controlled release of prostaglandin E1 from polyurethane surfaces [15].

Mechanistic aspects of the anticoagulant activity of surfaceimmobilized heparin
In 1993, Youngro Byun came to Holland to study the effect of fibronectin (Fn) on the binding of AT III to surface immobilized Hep (IM-Hep). Hep was covalently coupled via a hydrophilic PEO spacer group onto a model polymer surface of randomly copolymerized styrene and p-amino-styrene (IM-Hep). One of the objectives was to study the influence of plasma proteins on the anticoagulant activity of IM-Hep. The competition and binding interaction between IM-Hep, ATIII and thrombin were studied. The strong anionic character of Hep induces specific and nonspecific binding with various plasma proteins. For instance, Fn with six active binding sites for Hep, may possibly interfere with the binding of Hep with AT III or thrombin. The binding interaction of IM-Hep with AT III was investigated (a) by first binding AT III to IM-Hep, followed by the introduction of Fn, (b) by binding of Fn to IM-Hep followed by incubation with ATIII, or (c) by simultaneous incubation of IM-Hep with AT III and Fn. The extent of AT III binding to IM-Hep was assayed using a chromogenic substrate for AT III. It was demonstrated that a reversible displacement of AT III from IM-Hep took place, which was proportional to the Fn concentration. Furthermore, it was concluded that the final concentration of AT III bound to IM-Hep was not dependent on the binding sequence (a-c) [16].

Albumin-heparin microspheres
In 1986, Glen Kwon came to Holland and worked closely with Harry Cremers and You Han Bae (1988) on the preparation and possible applications of Alb-Hep microspheres (MS).
Adriamycin (ADR)-loaded albumin microspheres (Alb-MS) have been extensively investigated for chemoembolization based on their biocompatible behavior and non-toxic degradation products. However, these MS have also some drawback's (i.e. drug loading is required during MS preparation, relatively low payloads and considerable burst effects). Some problems can possibly be overcome by the use of MS composed of Alb and Hep (Alb-Hep-MS). These MS were prepared as follows: (i) using a double crosslinking technique, in which soluble Alb and Hep were subsequently treated with 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and glutaraldehyde (Alb-Hep-MS) or (ii) by first preparing a soluble conjugate of Alb and Hep, which was then crosslinked with glutaraldehyde (Alb-Hep-Conj-MS). Alb-Hep-MS could be loaded with ADR with payloads of 15-30%. It was shown that Alb-Hep-Conj-MS could be used for sustained release of ADR. In vitro and in vivo, no harmful toxic effects were encountered when using Alb-Hep-MS [17].
The compositions of preformed soluble conjugates of Hep and Alb were assessed by amino acid analysis. The swelling properties of Alb-Hep-Conj-MS and Alb-MS were pH dependent and as expected, the swelling of Alb-Hep-Conj-MS was more susceptible to ionic strength than that of Alb-MS. Surface properties of Alb-Hep-Conj-MS and Alb-MS were determined by XPS, CA, electrophoresis, and scanning electron microscopy (SEM), respectively. XPS surface analysis confirmed the presence of Hep at the Alb-Hep-Conj-MS water interface [18]. Alb-Hep-Conj-MS and Alb-MS were also used for the controlled delivery of positively charged polypeptides and proteins by ion exchange. The adsorption isotherm of chicken egg lysozyme (CELys) on Alb-Hep-Conj-MS was linear until saturation. Adsorption data of human lysozyme (HLys) onto Alb-Hep-Conj-MS and Alb-MS fit the Freundlich equation, implying the presence of heterogeneous MS binding sites, consistent with multivalent electrostatic interactions between HLys and negatively charged MS. Scatchard plots for the adsorption of HLys on Alb-Hep-Conj-MS and Alb-MS, respectively, indicate negative cooperativity, while positive cooperativity for CELys adsorption on Alb-Hep-Conj-MS was observed. HLys loading of Alb-Hep-Conj-MS was 3 times higher than on Alb-MS and sustained release occurred via ion exchange. Apparent diffusion coefficients (ADCs) of 2.1 × 10 −12 and 3.9 × 10 −11 cm 2 /s, were obtained for the release of HLys from Alb-Hep-Conj-MS and Alb-MS, respectively. Lysozyme release rates were not dependent on diffusion, which leads to the conclusion that the rate determining steps in the release were likely adsorption/desorption processes. An ADC of 4.1 × 10 −12 cm 2 /s was determined for the release of CELys from Alb-Hep-Conj-MS. The release rate of lysozymes from Alb-Hep-Conj-MS in deionized water was low in line with the proposed ion exchange release mechanism. In general, Alb-Hep-Conj-MS demonstrated superior ion exchange characteristics over Alb-MS [19]. Alb-Hep-Conj-MS were also evaluated for the release of macromolecules. The loading of Alb-Hep-Conj-MS with FITC-dextran (Mw 17.200 Da) was carried out either concurrently or after MS preparation. Higher loading with FITC-dextran was achieved when loading was carried out concurrently with MS preparation. Sustained release of FITC-dextran from Alb-Hep-Conj-MS could be modulated by varying the crosslinking density of MS, their loading content and by the method of drug incorporation. For release of FITC-dextran from MS, assuming negligible interactions, a diffusion coefficient of 1.7 × 10 −9 cm 2 /s was determined [20]. Alb-Hep-MS were either prepared by a two-step process which involved the preparation of a soluble Alb-Hep conjugate, followed by formation of MS from this conjugate (Alb-Hep-Conj-MS) or by cross-linking involving both coupling of soluble Alb and Hep and microsphere stabilization in one step (Alb-Hep-MS, one step). The preparation of Alb-Hep-Conj-MS was superior as compared to the preparation of Alb-Hep-MS, one step, allowing better control over the composition and homogeneity of the MS. MS with diameters of 5-35 μm were prepared. The size of MS could be controlled by varying the emulsification conditions. As mentioned previously, the degree of swelling of the MS increased with increasing pH and decreasing ionic strength of the external medium [21]. Alb-Hep-Conj-MS were also investigated as carriers for ADR. Depending on the Hep content of the conjugate used, Alb-Hep-Conj-MS could be loaded with ADR reaching payloads up to 33% w/w. As expected, the release rate of ADR from Alb-Hep-Conj-MS was dependent on the ionic strength of the release medium. ADR release in saline was 90% within 45 min, whereas in distilled water only 30% was released in the same time. Modelling of drug release profiles could be performed by combining ion-exchange kinetics and diffusion controlled drug release models [22]. ADR-loaded (up to 33% w/w) Alb-Hep-Conj-MS were also investigated for sitespecific intraperitoneal (i.p.) delivery. One day after applying empty or ADR-loaded Alb-Hep-Conj-MS to male Balb/c mice, i.p. neutrophil levels were only moderately increased. After three days, neutrophil levels were comparable with controls, indicating an acute mild inflammatory response. The anti-tumor efficacy of these ADR-MS was evaluated using a L1210 tumor-bearing mouse model and in a CC531 tumor-bearing rat model. ADR-Alb-Hep-Conj-MS induced prolonged survival times of mice and delayed tumor growth in rats, as compared with free drug treated and untreated animals. In both animal models higher ADR doses were initially tolerated if the drug was formulated in MS. However, long-term ADR toxicity effects were observed in all treated groups [23]. In vitro degradation studies of Alb-Hep-Conj-MS and Alb-MS, respectively, show that these MS are degraded by trypsin, proteinase K and lysosomal enzymes and that degradation rates were inversely related to MS cross-link densities. After intrahepatic administration of Alb-Hep-Conj-MS (cross-linked with 0.5% glutaraldehyde) in male Wag/Rij rats by injection into a mesenteric vein (intraveno-portal: i.v.p.), MS became localized in terminal portal veins predominantly located at the periphery of the liver. Within 2 weeks after injection, Alb-Hep-Conj-MS were degraded by cellular enzymes (half-life of approximately 1 d). The tissue responses of ADR-loaded Alb-Hep-Conj-MS and Alb-Hep-Conj-MS control were evaluated by histological assessment of the mitotic activity of liver parenchyma and inflammatory response, and by determination of liver damage marker enzymes (LDMEs) during 4 weeks post-administration. LDMEs were not increased as compared to controls, nor were adverse effects observed upon histological examination. No differences in response between empty and ADR-loaded Alb-Hep-Conj-MS were observed [24]. ADR was also formulated in Alb-Hep-Conj-MS to improve site-specific delivery and to reduce the toxicity of the drug. ADR-loaded Alb-Hep-Conj-MS were intrahepatically administered to male Wag/Rij rats. After intravenous-portal (i.v.p.) administration of ADR-loaded Alb-Hep-Conj-MS, ADR peak plasma concentrations were reduced 10-fold and ADR tissue levels of non-target tissues were significantly reduced as compared to i.v.p. administration of ADR. At a dose of 7.5 mg/kg, free ADR showed signs of acute toxicity. At the corresponding dose of ADR-loaded Alb-Hep-Conj-MS, no signs of toxicity were found. Cardiac function parameters using an isolated working heart model did not change after of i.v.p. administration of free ADR or ADR-loaded Alb-Hep-Conj-MS at a dose of 7.5 mg/kg. However, heart weights of animals treated with ADR loaded Alb-Hep-Conj-MS or free ADR were significantly lower than controls. Dose tolerances of free ADR, empty Alb-Hep-Conj-MS or ADR loaded Alb-Hep-Conj-MS after intrahepatic-arterial (i.h.a.) administration in rats were assessed using animal survival times. Empty Alb-Hep-Conj-MS were tolerated up to a dose of 45 mg/kg, free ADR up to a dose of 4 mg/kg, whereas ADR loaded Alb-Hep-Conj-MS were tolerated up to 10 mg ADR/kg [25].

Preparation of poly(ethylene oxide)
In 1996 Young Kwon Choi came to the Netherlands and prepared poly(ethylene oxide)s, PEO's, using yttrium isopropoxide as an initiator. At 80°C in the bulk, functional PEO's terminated with one hydroxyl group and one isopropoxy group ( 1 H and 13 C NMR) with predetermined molecular weights and low polydispersity's (Mw/Mn~1.1) were synthesized. PEO's prepared by this living polymerization could also be further extended with for instance ɛ-caprolactone to prepare well-defined block copolymers, indicating that the initiator is suitable for macromolecular engineering [26].
We still remember that during his stay he went to the hospital for one night because he mentioned that he had smelled phosgene during the synthesis of NCA-amino acids.
Naltrexone was modified at the 3-and 14-hydroxyl positions and covalently coupled to a biodegradable poly(α-amino acid) backbone through a labile bond. Selective acetylation of naltrexone with acetic anhydride gave naltrexone-3-acetate, which was subsequently succinylated to naltrexone-3-acetate-14-hemisuccinate with succinic anhydride. The polymeric backbone chosen for initial coupling experiments was poly-N 5 -(3-hydroxypropyl)-L-glutamine (PHPG). The side-chain hydroxyl functionality permitted covalent bonding of the hemisuccinate through an ester linkage. Hydrolysis of covalently bound drug to give naltrexone or its derivatives should be much slower than diffusion of drug through the polymer matrix. While hydrolysis of naltrexone from the polymer side chain is first order, release of drug from the matrix can be zero order due to the geometry of the device and the physical and chemical interactions between naltrexone and the polymer matrix. In vitro studies of PHPG-naltrexone conjugate in disk form did not show constant release because of the hydrophilic nature of the polymer backbone and the changing local chemical environment upon hydrolysis of drug-polymer linkages. The conjugated system was made more hydrophobic by coupling drug to copolymers of hydroxypropyl-Lglutamine (HPG) and L-leucine. Conjugates of the hemisuccinate and co-poly(HPG-70/Leu-30) demonstrated a nearly constant, but slightly declining release rate of naltrexone and its derivatives for 28 days in vitro [27]. Somewhat later, Dave Bennett came to The Netherlands and also worked on the coupling of naltrexone, an opioid receptor antagonist as well as the antihypertensive, minoxidil, to biodegradable poly(αamino acids) based on L-glutamic acid to obtain polymeric prodrugs. The synthesis and release characteristics of these two systems were investigated. As mentioned above, the rate of hydrolytic cleavage of the polymer-drug linkage should be much slower than the rate of diffusion through the polymer matrix; thus, release rates approaching zero-order can be achieved. In vitro and in vivo release studies demonstrate the potential of these types of polymeric delivery systems [27][28][29].
In 1981 Mart Eenink went to Utah to work on glycosylated insulin derivatives, which possibly could be applied in self-regulating insulin delivery systems. Seo Young Jeong visited us in Enschede in 1983 and he also worked on the preparation and characterization of different glycosylated insulins, in which the sugar was coupled to insulin via succinate or glutarate spacers. These compounds could be used in selfregulating insulin delivery systems based on the competitive binding between glucose and glycosylated insulin with Concanavalin A (lectin)based polymer matrices [30].
In 1991 Masahiro Nukui came from Japan to the Netherlands and worked on the association behavior of conjugates of adriamycin (ADR) and poly(L-glutamic acid), (PGA). The association behavior in aqueous solution of two types of amphiphilic conjugates consisting of ADR bound to PGA, either directly or via a peptide spacer, was investigated using absorbance and fluorescence spectroscopy, GPC and low angle laser light scattering (LALLS) measurements. ADR was bound via the aminoribosyl moiety either directly with PGA side-chain groups or via oligopeptide (GlyGlyGlyLeu) spacer groups. UV/VIS, GPC and LALLS data showed that in buffer solution both ADR conjugates associate inter-molecularly. Directly bound ADR conjugates form multimers with different degrees of association, which are in equilibrium with single polymer chains. In contrast, peptide spacer-containing ADR conjugates associate to give stable uniform multimers. UV/VIS and fluorescence spectroscopy performed at very low conjugate concentration show that polymer-bound ADR residues associate intra-molecularly as well. The degree of intra-and inter-molecular association of the conjugate-bound ADR molecules depends on the type of conjugate (i.e., with or without spacer), ADR load and conjugate concentration as well as solvent ionic strength, the presence of organic co-solvents and temperature. ADRconjugates containing the peptide spacer are more stable compared with directly bound ADR-conjugates, probably due to enhanced flexibility and the presence of hydrophobic leucine residues in the spacer residues. ADR conjugates of PGA via GlyGlyGlyLeu residues preferentially fold intramolecular and are therefore preferred drug carriers over directly coupled ADR conjugates, which form less defined multimers [31].

Polydepsipeptides
In 1984 Shuji Sato came to assist Janneke Helder with her polydepsipeptide project to synthetize copolymers of glycine and D,L-lactic acid by ring-opening of 2,5-morpholine-dione derivatives [32], and in 1985 Seung Jin Lee also came to The Netherlands and worked together with Janneke Helder on the synthesis of copolymers of glycine and D,Llactic acid with different compositions of hydroxy-and amino acids by ring-opening polymerization [33]. In 1991 Zhengrong Shen came from China and contributed to the program with additional work on glycine/ glycolic acid based biodegradable copolymers, prepared by ringopening homo-polymerization of morpholine-2,5-dione, and ringopening copolymerization of glycolide and morpholine-2,5-dione. Optimal polymerization conditions were found for the homo-polymerization of morpholine-2,5-dione using stannous octoate as an initiator (3 min at 200°C and 17 h at 130°C), providing a yield of 60% and polymers with intrinsic viscosities [η] = 0.50 dL/g (DMSO, 25°C) using a molar ratio of monomer and initiator of 1000. The homo-polymer with alternating glycine and glycolic acid residues had a Tg of 67°C and a Tm of 199°C. Random copolymers of glycine and glycolic acid were synthesized similarly by copolymerizing glycolide and morpholine-2,5dione. These copolymers were either semi-crystalline or amorphous, depending on the mole fraction of glycolic acid residues [34].

Poly(lactide) release matrices
Phil Triolo came to The Netherlands in 1984 and later in the period between 1985 and 1987 and worked on the release of macromolecules from poly(lactide) polymer matrices using albumin as a model compound.
Phil met his later wife, Willy, in The Netherlands during this stay, perhaps producing the ultimate in Triangle cultural exchange experiences.

Stereo-complex formation of poly(lactides)
In 1988, Nobuhiko Yui came to Holland to synthesize sequential diblock copolymers composed of L-and D-lactic acid residues, respectively, via a living ring-opening polymerization of L-and D -lactide using aluminum tris(2-propanolate) as an initiator. Block copolymers with different compositions were synthesized by varying the reaction conditions and monomer over initiator ratio. The polymers were characterized by 1 H NMR, molecular weight and optical rotation measurements. Molecular weights ranged from 1.3-2.0 · 10 4 with M̄w/ M̄n varying between 1,2-1,4. Stereo-complex formation in these blockcopolymers was proven by their melting temperatures of about 205°C (DSC) [35]. Chaulmin Pai came to Holland in 1993 and worked on stereo-complexed poly-lactide systems for controlled drug delivery.

Thermo-sensitive polymers
Herman Feil went to Utah from 1989 to 1993 to perform a joint PhD study with Sung Wan Kim, You Han Bae (member of his PhD committee) and myself on thermo-sensitive (co)polymers based on N-isopropylacrylamide (NIPAAm). Different ionizable, thermosensitive polymers were prepared with the general formula of poly (N-isopropylacrylamide-co-butylacrylamide-co-X), in which X is an ionizable co-monomer, like acrylic acid or N,N-(diethylamino)ethyl methacrylate. It was shown that the Lower Critical Solution Temperatures (LCST's) of these polymers are affected by polyelectrolyte complex formation when polyelectrolytes of opposite charge of the LCST polymer are present [36]. Subsequently, poly(N-isopropylacrylamideco-butyl methacrylate-co-X), with X being a hydrophilic, hydrophobic, cationic, or anionic co-monomer were prepared. DSC measurements of aqueous solutions of these polymers were used to elucidate the mechanism of temperature-induced phase separation as a function of comonomer content, hydrophilicity, and charge. The endothermic heat of phase separation decreased linearly with the LCST, suggesting that hydrophobic interactions between polymer side groups increase with increasing temperature. This is caused by reduced water structuring around hydrophobic groups with increasing temperatures. An important conclusion from this study is that changes in LCST caused by the incorporation of co-monomers are due to changes in overall hydrophilicity of the polymer. Co-monomer hydrophilicity or charge does not have a direct effect on the structuring of water around hydrophobic residues in the co-polymers [37]. Poly(NIPAAm-co-BMA-co-DEAEMA) hydrogels, where BMA is butyl methacrylate and DEAEMA is (diethylamino)ethyl methacrylate were obtained using ethylene glycol di-methacrylate (EGDMA) as a crosslinking agent. Collapse of the gels (pH = 7.4) occurs at high temperatures even when more than 20 mol% of DEAEMA was incorporated. This can be explained by the increased neutralization of DEAEMA with increasing temperature. The fraction of ionizable DEAEMA monomer in the network determines the temperature range in which the temperature-induced swelling transition occurs. The pH-induced swelling transitions in these networks were controlled by temperature, which can be related to the influence of NIPAAm on the degree of ionization of DEAEMA. These results assist the design of polymer networks with specific temperature-or pH-induced swelling transitions [38]. Temperature-sensitive hydrogel membranes based on crosslinked poly(N-isopropylacrylamide-co-butyl methacrylate 95:5 mol%) were applied for the separation of molecules of different size. By applying different swelling levels of the hydrogel membrane via a change in temperature at the appropriate time, two dextrans with molecular weights of 150,000 and 4400 g/mol, respectively, and uranine with a molecular weight of 376 g/mol could be effectively separated in a continuous way. The inverse membrane hydration was linearly related to the rate of diffusion of uranine as well as dextran (molecular weight 4400 g/mol) in line with the free-volume theory [39]. In 2001 Young Min Kwon came to Holland and worked on the synthesis and aqueous phase behavior of thermo-responsive biodegradable triblock copolymers based on D,L-3-methylglycolide (MG) and poly(ethylene glycol) (PEG). Ring-opening polymerization of MG was initiated with PEG and Ca[N(SiMe 3 ) 2 ] 2 (THF) 2 in THF and triblock copolymers with alternating lactyl/glycolyl sequences of controlled molecular weight and low PDI were obtained. A 27 wt.-% aqueous solution of PMG-PEG-PMG (1400-1450-1400) shows gelation near 30°C and syneresis above 41°C. The system is somewhat more hydrophilic than the previously reported lactide/glycolide (molar ratio 3:1) based system PLGA-PEG-PLGA. Upon increasing the temperature of aqueous solutions of PMG-PEG-PMG, sol-gel transitions or an increase in viscosity without gel formation were detected by rheology and dynamic differential scanning calorimetry (DDSC). Transitions were dependent on molar mass ratios of PMG and PEG blocks [40].

Tissue engineering
In 2008, Zheng Zhang went to Japan and participated in a program to investigate the applicability of human periodontal ligament-derived cells (hPDL cells) as a reliable source for cyto-therapeutic use. Periodontal ligament (PDL) is an attractive source for cells to be used for periodontal regeneration. A protocol for the extraction, expansion, and characterization of human PDL (hPDL) cells for clinical trials was developed. Human PDL tissues were obtained from 41 surgically extracted teeth and digested with collagenase/dispase (29/30 (96.7%)) enzymes to obtain hPDL cells. These cells exhibited osteogenic potential both in vitro and in vivo, proliferated rapidly at low cell densities and frequently differentiated into a cementoblastic/osteoblastic lineage (~60%). NCAM1, S100A4, and periostin genes were more expressed in hDPL cells as compared to controls (bone marrow-derived mesenchymal stem cells (hBMMSCs), and gingival fibroblasts (hGFs)). S100A4 and periostin genes were also expressed in hPDL tissue as detected by immunohistochemical analysis. These results led to a protocol for the successful application of hPDL cells for clinical trials [41].

Non-viral gene delivery and delivery of nucleic acid based drugs
Zhiyuan Zhong went to Utah in 2004 and worked on novel ABA triblock copolymers consisting of low molecular weight linear polyethylenimine (L-PEI) as the A block and poly(ethylene glycol) (PEG) as the B block. Linear PEI-PEG-PEI triblock copolymers with controlled compositions were prepared by polymerizing 2-methyl-2-oxazoline (MeOZO) using PEG-bis(tosylate) as a macro-initiator and subsequent acid hydrolysis. Two copolymers, PEI-PEG-PEI (2100-3400-2100) and PEI-PEG-PEI (4000-3400-4000), were prepared. Both copolymers were able to condense plasmid DNA (pDNA) effectively to provide polymer/ pDNA complexes (polyplexes) with sizes lower than 100 nm and moderate positive zeta-potentials of about +10 mV at polymer/plasmid weight ratios ≥1.5/1. These polyplexes were used for successful transfection of COS-7 cells and primary bovine endothelial cells (BAECs), respectively. The transfection efficiencies of polyplexes based on PEI-PEG-PEI (4000-3400-4000) and polyplexes of branched PEI (25,000) were similar. Transfection of BEACs with luciferase reporter gene was threefold higher using polyplexes of PEI-PEG-PEI (4000-3400-4000) as compared to L-PEI (25,000)/DNA polyplexes. The transfection activity of PEI-PEG-PEI polyplexes was not inhibited by the presence of serum in the medium. The cytotoxicity of PEI-PEG-PEI triblock copolymer systems was very low under conditions to achieve high transgene expression, whereas systems using high molecular weight PEIs were rather toxic. Therefor linear low molecular weight PEI-PEG-PEI triblock copolymers are promising systems for nonviral gene delivery [42]. Lane Christensen came to the Netherlands in 2005 and worked closely together with Zhiyuan Zhong and Johan Engbersen in a joint program with Sung Wan Kim on triggered intracellular gene delivery. Various poly(amido ethylenimine)s with multiple disulfide bonds (PAEIs-MDS), were prepared for effective and triggered delivery of pDNA into cells. Three PAEIs-MDS were synthesized by Michael addition reactions of cystamine bisacrylamide (CBA) with various ethylene amine monomers (ethylenediamine (EDA), diethylenetriamine (DETA), or triethylenetetramine (TETA)). As shown by 1 H NMR, these addition reactions all went to completion. PAEIs-MDS were characterized by GPC, acid-base titration, and liquid chromatography-mass spectroscopy (LC-MS). Polyplexes with pDNA were characterized by gel electrophoresis and their particle size and zeta-potential were determined. Polyplexes of PAEIs-MDS and pDNA had diameters less than 200 nm and positive surface charges of about 32 mV. Mouse embryonic fibroblast (NIH3T3), primary bovine aortic endothelial (BAEC), and rat aortic smooth muscle (A7R5) cell lines were used for transfer of a reporter gene (pCMV-Luc reporter plasmid) using the corresponding polyplexes of the PAEIs-MDS. High levels of reporter gene expression (about 20× higher transfection efficiency than polyplexes based on linear polyethylenimine (L-PEI, 25 k)) were obtained. Serum in the transfection medium (10%) did not affect the transfection efficiency. Confocal microscopy using labeled pDNA indicated that polyplexes with PAEIs-MDS were most probably degraded by intracellular reductive agents (e.g. glutathione) leading to an improved intracellular pDNA distribution as compared to polyplexes based on PEI. Summarizing, PAEIs-MDS are promising transfection agents combining high gene expression with low toxicity levels [43]. A reducible poly(amido ethylenimine) (PAEI-MDS) synthesized by addition copolymerization of TETA and cystamine bisacrylamide (poly(TETA/CBA)) was developed as a carrier for small interference RNA (siRNA). Poly(TETA/CBA) efficiently condenses siRNA to form stable complexes under physiological conditions. siRNA was completely released from the complexes in a reductive environment (glutathione). After incorporation in human prostate cancer cells (PC-3), VEGF-directed siRNA-poly(TETA/CBA) complexes suppressed VEGF more efficiently than corresponding complexes based on linearpolyethylenimine (L-PEI, 25 k). Using the poly(TETA/CBA) formulation, five hours after transfection, an improved intracellular distribution of dissociated siRNA was observed as compared to the L-PEI based complexes. Reductive degradation of poly(TETA/CBA) in the cytoplasm (e.g. glutathione) triggers release of siRNA in the cytoplasma, which also may affect the RNAi activity. These results suggest that non-viral carriers should not only be stable at physiological conditions, but should also be designed for effective intracellular dissociation and trafficking of the nucleic acid drug [44]. Transfection of cells with hypoxia-inducible vascular endothelial growth factor (RTP-VEGF) plasmid using a novel reducible disulfide poly(amido ethylenediamine) (PAEI-MDS) carrier was studied. In primary rat cardiomyoblasts (H9C2), transfection with PAEI-MDS based luciferase reporter gene complexes at a weight ratio of polymer/DNA (12:1) leads to 16-fold higher expression of luciferase compared to corresponding complexes based on an optimized branched PEI control. FACS analysis revealed up to 57 (+/− 2%) GFP positive H9C2s. The efficiency of plasmid delivery to H9C2 using the PAEI-MDS was depended on intracellular glutathione (GSH) levels. PAEI-MDS mediated delivery of RTP-VEGF plasmid in both H9C2 and rat aortic smooth muscle cells (A7R5) induced significantly higher levels of VEGF expression (up to 76-fold) under hypoxic compared to normal oxygen conditions. In a rabbit myocardial infarct model, delivery of 100 micrograms of RTP-VEGF by PAEI-MDS resulted in up to 4-fold increase in VEGF protein expression in the region of the infarct compared to injections of PAEI-MDS /RTP-Luc. It is concluded that this PAEI-MDS can be applied to improve ischemia-inducible VEGF gene therapy both in vitro and in vivo, thereby promoting neovascular formation and improving tissue function in ischemic myocardium [45]. A variety of cancer and non-cancer cell lines was transfected with pDNA using poly(amido-butanol) (pABOL), which is an effective bioreducible poly(amido amine) gene carrier. A mouse renal carcinoma (RENCA) cell line was used to transfect survivin-inducible plasmid DNA, which was constructed to express the soluble VEGF receptor, sFlt-1, an antagonist of VEGF, downstream of the survivin promoter (pSUR-sFlt-1). A dual bio-responsive gene delivery system was developed based on p(ABOL) and the survivin-inducible gene expression system (pSUR-sFlt-1 or pSUR-Luc reporter gene) which proved that increased gene expression can be effected in cancer cells as compared to normal cells [46]. In 2013 Young-Wook Won came to The Netherlands and worked on poly(amido amine)s with agmatine and butanol side chains as efficient gene carriers. Bioreducible poly(amido amine) copolymers were prepared by Michael addition polymerization of cystamine bisacrylamide (CBA) with mixtures of different ratios of 4-aminobutylguanidine (agmatine, AGM) and 4-aminobutanol (ABOL). It was shown previously that both the presence of positively charged guanidinium groups of AGM and hydroxybutyl groups of ABOL in the side chains of poly(amido amines) improve their overall transfection efficiency. We therefore synthesized poly(CBA-ABOL/AGM) polymers with various ratios of both components as gene carriers. Luciferase gene transfection efficiencies of these polymers in various cell lines were determined to select the optimal ratio of AGM and ABOL. Poly(CBA-ABOL/AGM) containing a molar ratio of AGM and ABOL of 4 showed the best transfection efficiency and the lowest cytotoxicity, indicating that this polymer is very promising as a potent and nontoxic gene carrier [47].

Photodynamic therapy
In 1998 and 1999, Henk Stapert went from the Netherlands to Tokyo University to work with the group of Kazunori Kataoka on photodynamic therapy as an EU-Fellow Science & Technology. This 2year fellowship was awarded by the EU-Japan Science and Technology fellowship program financed by the European Commission-Directorate General XII, Science and Technology. Core-shell type micelles with a hydrodynamic diameter of 52 nm composed of 38 molecules of zinc dendrimer porphyrin with 32 negatively charged carboxylate groups on the periphery and 39 positively charged poly(ethylene glycol)-poly(Llysine) (PEG-PLys) block copolymer molecules were prepared at a stoichiometric mixing ratio. At salt concentrations higher than 200 mM NaCl, these micelles were stabilized by the formation of hydrogen bonds. In contrast, micelles formed by a zinc dendrimer porphyrin with 32 positively charged trimethylammonium groups and negatively charged poly(ethylene glycol)-poly(aspartic acid) (PEG-PAsp) block copolymer were disrupted at NaCl concentrations exceeding 200 mM. However, both polyion dendrimer micelle systems showed a high stability upon dilution at physiological conditions and are potential delivery systems for light-harvesting ionic zinc dendrimer porphyrin sensitizers [48]. Third-generation aryl ether dendrimer porphyrins (DPs) with either 32 positively charged quaternary ammonium groups (32 DPZn) or 32 negatively charged carboxylic groups (32 DPZn) were assessed as promising, supramolecular photosensitizers for PDT. DPs showed different cell-interactions depending on the positive or negative charge on the periphery, and both DPs were eventually internalized via endocytosis, followed by localization in membrane-limited organelles such as lysosomes. In contrast, protoporphyrin IX (PIX), which is a hydrophobic and relatively low molecular weight photosensitizer used as a control in this study, became localized in the cytoplasm, but not in the nucleus. Confocal fluorescent imaging using organelle-specific dyes indicated that PIX induced severe photo-damage to disrupt membranes and intracellular organelles, including the plasma membrane, mitochondrion, and lysosome. In contrast, cells treated with DPs maintained the characteristic fluorescent pattern of such organelles even after photo-irradiation. Positively charged (32 DPZn) gave rise to substantially higher 1 O 2 induced cytotoxicity against Lewis Lung Carcinoma (LLC) cells than PIX. Additionally, both DPs had far lower dark toxicity as compared with PIX, demonstrating their highly selective photosensitizing effect combined with a reduced systemic toxicity [49].

Gelation behavior of polymers under electric current
In 1991 Ick Chan Kwon went from Utah to Japan (ICBS) and studied the gelation behavior of polymers under the influence of electric currents.

Conclusions
I feel that we have come to the end of an era, in which Prof. Sung Wan Kim played such an essential and influential role. Sung Wan was not only an excellent scientist, but also a person who cherished personal contacts and understood the value of global scientific teams. The numerous Triangle participants in the exchange program have been enriched by these productive international experiences and remain very grateful for the research opportunities offered by Sung Wan in facilitating and maintaining the unique Triangle exchange program. Some will never forget their time spent in Salt Lake City, with the scientific stimulation, with the different culture and the personal attention provided by Sung Wan. Others have developed academic or industrial research groups of their own in their advanced scientific careers, and continue with the international spirit to support a global research community inspired by their Triangle experiences.

Contributions of Early Exchange Students and Post-doctoral
Associates' between the Laboratories of Sung Wan Kim (University of Utah), Teruo Okano and Yasuhisha Sakurai (Tokyo Women's Medical College), Kazunori Kataoka (University of Tokyo) and Jan Feijen (University of Twente).

Declaration of Competing Interest
None.