Peritoneal Transport Dynamics of Glucose and Icodextrin: the in Vitro

Advances in Peritoneal Dialysis, Vol. 21, 2005

Peritoneal Transport Dynamics of Glucose and Icodextrin: The In Vitro

Krystyna Czyzewska, Beata Szary, Teresa Grzelak Comparative Studies

We performed in vitro experiments with the isolated Key words rabbit parietal peritoneum to evaluate the importance Diffusive transport, glucose, icodextrin, unstirred fluid of fluid stirring intensification and of chemical modi- layers, sodium deoxycholate fication of mesothelium and interstitium to the peritoneal transport of glucose and icodextrin. We used a Introduction mathematical model of mass transport to calculate Glucose is usually added to dialysis solution for the the diffusive permeability coefficient, P, in centime- development of sufficient ultrafiltration during peri- ters per second. In control conditions (intact tissue; toneal dialysis (1,2). However, with increase in dwell stirring rate: 11 mL/min), the rate of glucose (2.0 g/dL) time, ultrafiltration decreases because of rapid glu- transfer remained constant, and no differences were cose absorption from the peritoneal cavity to the vas- observed for transport from the interstitial to the me- cular bed, with its resulting decrease in the osmotic sothelial (I→M) side of the membrane or in the oppo- pressure gradient (3,4). Recently, dialysis fluids with site direction (M→I). The value of P (± standard error glucose polymer (icodextrin) have been introduced of the mean) was 2.731 ± 0.472 × 10–4 cm/s. In con- to provide sustained ultrafiltration and to avoid the trast, the icodextrin (7.5 g/dL) I→M transport rate disadvantageous effects of glucose that are related to was higher than that for M→I (P: 0.319 ± 0.038 × long-term application of a hyperosmotic glucose-based 10–4 cm/s and 0.194 ± 0.035 × 10 –4 cm/s respectively). fluid during standard peritoneal dialysis (5,6). Icodex- Dynamics of the icodextrin M→I transfer were con- trin has been proposed as a carrier solution during stant, but I→M increased by 50% over time. The in- intraperitoneal chemotherapy for high doses of drugs tensification of the stirring rate increased the value in the peritoneal space and for reduction of its con- of P at varying rates: the increase was greater for centration in the systemic circulation, and as a pre- icodextrin than for glucose, and greater for the I→M vention measure for adhesion formation caused by transport direction than for the M→I direction for both intraperitoneal pharmacotherapy (7–9). solutes. Chemical modification (by 2.5 mmol/L sodium Diffusion and convection are the main mechanisms deoxycholate) increased glucose and icodextrin I→M of transperitoneal transport. Transport can occur in transfer a mean of 41% and 81% respectively, but in- two directions—from the peritoneal cavity to the creased M→I transfer by 70% and 224% respectively. blood, and from the vascular bed to the abdominal The dynamics of glucose and icodextrin perito- space—according to the concentration gradient and neal transfer in vitro are different: glucose diffusion the rate of dialysate and blood flow. Diffusion is the is constant, but I→M icodextrin transfer increases main mechanism of glucose (molecular weight: over time and is greater than M→I transfer. Fluid stir- 180 Da) transport, as it is for other small molecules ring intensification and chemical injury to the perito- during peritoneal dialysis. Peritoneal absorption of neum enhance diffusion of glucose and icodextrin. glucose increases during peritonitis as a consequence Glucose and icodextrin M→I transfer, but not I→M of peritoneal barrier modifications (10–12). Knowl- transfer, is restricted more by tissue barriers than by edge of the peritoneal transport of icodextrin, a high stagnant fluid layers. molecular weight compound (13,000 – 19,000 Da), is not extensive. Intraperitoneally introduced icodextrin From: Department of Chemistry and Clinical Biochemis- is probably absorbed at a low rate mainly by the lym- try, Poznan University of Medical Sciences, Poland. phatic pathway because of convective fluid movement 54 Czyzewska et al. out of the peritoneal cavity; the diffusive component (16). The principle of the assay involves exhaustive is limited, but not excluded (13,14). hydrolysis of glucose polymer using the enzyme During peritoneal dialysis, transperitoneal trans- amyloglucosidase (Sigma, St. Louis, MO, U.S.A.) port encounters several resistances: capillary endothe- from Aspergillus niger. lium with basement membrane, interstitium, Four experiments were carried out separately for mesothelium with basement membrane, and stagnant glucose and icodextrin: fluid layers on the surfaces of mesothelium and endo- thelium. Capillary wall is considered the major trans- • Control conditions (120 minutes; intact tissue; port barrier for transperitoneal exchange. Mesothelium fluid stirring rate: 11 mL/min) is usually expected to be highly permeable. Informa- • Before (15 – 60 minutes) and after (75 – 120 min- tion about the transport characteristics of the intersti- utes) change of stirring rate from 5.5 mL/min to tium, which can be a significant transport barrier, 11 mL/min especially for small molecules, is scarce. Stagnant fluid • Before (15 – 60 minutes) and after (75 – 120 min- layers are not often investigated (10,15). utes) change of stirring rate from 11 mL/min to In the present study, we compared the bidirectional 22 mL/min diffusion transfer of glucose and icodextrin across • Before (15 – 75 minutes) and after (90 – 150 min- peritoneal membrane in vitro and its change as a re- utes) chemical modification of the peritoneum sult of fluid stirring intensity and chemical modification of the peritoneal membrane. The fourth procedure used sodium deoxycholate (2.5 mmol/L, 104 mg/dL) added to the medium on the Material and methods mesothelial side of the peritoneum. After 3 minutes We studied transfer of glucose and icodextrin (initial of incubation with the detergent, the chamber was concentration gradients, 2.0 g/dL and 7.5 g/dL respec- thrice washed with fresh Hanks solution. tively) from the interstitial to the mesothelial side Assuming that the transmembrane transport in the (I→M) of rabbit peritoneal membrane and in the op- Ussing cell is purely diffusive and that mixing of fluid posite direction (M→I). The experiments were car- in each compartment of the cell is perfect, we used a ried out on fragments of parietal peritoneum taken mathematical model to estimate the diffusive perme- from the anterior abdominal wall of New Zealand rab- ability of the peritoneum (P in centimeters per sec- bits and mounted in a modified Ussing-type chamber. ond), expressed as a mean value ± standard error of The active surface of the membrane was 1.1 cm2. The the mean (SEM). The changes of P attributable to the chamber was filled with Hanks solution of the fol- experimental modifications were determined as a per- lowing composition (all mmol/L): NaCl, 136.88; KCl, centage of the control value before the change, indi-

5.36; NaHCO3, 4.16; CaCl2, 1.26; KH2PO4, 0.44; vidually for each experiment (that is, separately for Na2HPO4 × 12 H2O, 0.34; MgCl2 × 6 H2O, 0.49; each piece of peritoneal membrane), and they are pre- MgSO4 × 7 H2O, 0.41. A total of 15 mL of solution sented as mean ± SEM for the series. In this way, each was circulated at the rate 5.5 mL/min, 11 mL/min, or piece of membrane served, in the initial part of the 22 mL/min. Adequate oxygen content in the medium experiment, as a control for the second part. and a constant pH of 7.4 were maintained by continu- For statistical analyses, we used the Student t-test ous bubbling with a gas mixture consisting of 5% CO2 for paired data (14). and 95% O2. Polyethylene tubes were used to con- nect the chamber to the fluid reservoir and a peristal- Results tic pump. The whole system was placed in a thermostat In control conditions, the rate of glucose transfer was box at 37°C (14). Sampling of the medium was car- stable for 120 minutes of the experiment, and no dif- ried out at regular 15-minute intervals. Glucose (Polfa, ferences were observed for transport directed from the Kraków, Poland) was measured using glucose oxidase interstitial to the mesothelial side of membrane (I→M) (Cormay, Kraków, Poland). Icodextrin (average or in the opposite direction (M→I). Values of P were weight: 14,556 Da), obtained from ML Laboratories 2.878 ± 0.390 × 10–4 cm/s and 2.584 ± 0.554 × plc (Liverpool, U.K.), was determined by the rapid 10–4 cm/s, respectively. In contrast, the values for ico- and simple assay developed by Wang and colleagues dextrin were different for transport in both directions: Peritoneal Transport Dynamics of Glucose and Icodextrin 55

0.194 ± 0.035 × 10 –4 cm/s (I→M) and 0.319 ± 0.038 × M→I transfer. The reasons for the described in vitro 10–4 cm/s (M→I). icodextrin transport phenomena are not known. They The dynamics (15 – 60 minutes vs. 75 – 120 min- may be connected with partial degradation of the glu- utes) of M→I icodextrin transport remained constant, cose polymer, an increase in fluid osmolality, and but I→M increased by 50% over time (p < 0.01; modifications in peritoneal transport function (18,19). Table I). The increase in the stirring rate to 11 mL/ Glucose polymer degradation in the rat is related min from 5.5 mL/min enhanced values of P for glu- to high amylase activity, but no data are available for cose by about 74% (I→M) and 58% (M→I) and for the rabbit. In contrast, in humans, intraperitoneal me- icodextrin by about 95% (I→M) and 61% (M→I) tabolism of icodextrin is small, and amylase activity [Figure 1(A)]. The change in stirring from 11 mL/min in peritoneal effluents from peritoneal dialysis patients to 22 mL/min increased glucose transport by about is undetectable or very low (13,17,18). Previous 42% in both directions, and increased icodextrin trans- in vitro results showed that glucose polymer modifies port by about 80% in the I→M direction and by about the diffusive permeability of the peritoneum and in- 25% in the M→I direction [Figure 1(B)]. Chemical duces asymmetry of uric acid and albumin transport injury to the peritoneum augmented glucose transfer (20). Frajewicki and co-authors detected a significant by 41% (I→M) and 70% (M→I), but augmented ico- increase in the peritoneal excretion of protein after a dextrin transfer by 81% (I→M) and 224% (M→I) single use of glucose polymer (7.5 g/dL) in rats and, [Figure 1(C)]. more importantly, after long-term administration. Moreover, the latest investigations underline the in- Discussion fluence of icodextrin on the peritoneal membrane: Earlier investigations showed that, in contrast to chemical peritonitis, reduction in mesothelial cell vol- glucose, intraperitoneally introduced icodextrin is ume, and membrane lipid peroxidation (21,22). transported slowly to blood. About 40% of this macro- Unstirred fluid layers are a barrier to the perito- molecule (7.5 g/dL) is absorbed during a 12-hour neal transport of molecules, but the significance of dwell (17). By comparison, glucose absorption is much the phenomenon is not clear. A variety of methods higher; about 85% is transferred during 2 hours of have been used in vivo and in vitro to reduce the stag- peritoneal dialysis (4). In in vitro conditions, the trans- nant fluid resistance on the peritoneal surface: for peritoneal transport of icodextrin (average weight: example, vibration of the abdominal wall, rhythmic 14,556 Da; initial concentration gradient: 7.5 g/dL) shaking of experimental animals, or changes in fluid was about ten times smaller than that of glucose (mo- flow rate (14,23,24). lecular weight: 180 Daltons; initial concentration gra- In the present in vitro study, we examined the im- dient: 2.0 g/dL). The peritoneal transport of glucose portance of stagnant fluid layers by changing the fluid was stable during the 120 minutes of the experiment, flow speed on both the mesothelial and interstitial sides and differences between the two transfer directions of the peritoneal membrane. An increase of fluid mix- were not observed. Icodextrin transport across the ing to 11 mL/min from 5.5 mL/min augmented glu- peritoneal membrane from the interstitial to the me- cose and icodextrin peritoneal transport bidirectionally sothelial side of the tissue increased over time and by 61% – 95% (higher percentages for glucose poly- showed asymmetry: I→M transfer was greater than mer than for glucose). Further increasing the fluid flow

TABLE I Transperitoneal glucose and icodextrin transport in vitro expressed as diffusive permeability (P) in control conditions (15 – 60 minutes and 75 – 120 minutes)

Initial Experiments Transport Diffusive permeability (× 10–4 cm/s) Compound concentration (n) direction 15–60 Min p Value 75–120 Min

Glucose 2 g/dL 10 I→M 2.797±0.388 NS 2.959±0.393 10 M→I 2.733±0.679 NS 2.436±0.429 Icodextrin 7.5 g/dL 11 I→M 0.280±0.038 0.01 0.394±0.046 13 M→I 0.195±0.039 NS 0.207±0.035

I = interstitial side of the peritoneal membrane; M = mesothelial side of the peritoneal membrane; NS = statistically nonsignificant. 56 Czyzewska et al.

rate (to 22 mL/min from 11 mL/min) evoked less-evi- A dent changes (25% – 80%). When animals were shaken, in vivo diffusion of glucose was more than four times enhanced, but transfer of macromolecules (protein) increased only by about 50% (24). During continuous flow-through peritoneal dialysis in dogs, glucose clearance was 11 ± 5 mL/min for a flow rate of 3.6 L/h and 16.5 ± 6 mL/ min for a flow rate of 6 L/h (25,26). These in vitro and in vivo findings are probably all connected with an increase of fluid contact with the peritoneal membrane and the resulting reduction in stagnant fluid layers. In contrast, in transport experiments with a plastic chamber affixed to rat parietal peritoneum, fluid mix- B ing did not affect the transfer of mannitol, just as in a previous in vitro investigation, an increase in fluid flow did not change the transport of urea, but enhanced transfer of albumin (14,27). The reasons for the described diversity are unknown, but complexity of the transperitoneal transport mechanisms, adaptation of the peritoneal membrane, and differences in the experimental models cannot be excluded. In our in vitro study, chemically modified peritoneum was used as a surrogate for change in the peritoneal membrane during long-term dialysis with recurrent peritonitis. Within the peritoneum, peritoneal inflammation induces structural and functional C disturbances that influence peritoneal transfer dynamics (13,28,29). To injure the experimental peritoneal membrane, we used 2.5 mmol/L (104 mg/dL) sodium deoxycholate, which is known to denude the peritoneal and pleural mesothelium and the epithelium of renal tubules, but not to damage the associated basement membranes (28,30). Other data from our study showed that the detergent induces apoptosis and ne- crosis of mesothelial cells, lipid peroxidation, and generation of free radicals. Morphologic and mor- phometric examinations showed that sodium deoxycholate removed mesothelial cells and induced loosening of adjacent connective tissue. The detergent increased the thickness of the rabbit parietal peritoneum by about 40% (p < 0.001, unpublished data). FIGURE 1 The changes of diffusive permeability (P) for glucose and icodextrin in rabbit peritoneum expressed as a percentage of This chemical injury to the peritoneum increased dif- the control value ± the standard error of the mean (SEM). (A) Before fusion transport of glucose and icodextrin in both and after the change of mixing rate from 5.5 mL/min to 11 mL/ transport directions, but to a higher degree for glu- min; (B) before and after the change of mixing rate from 11 mL/ cose polymer then for glucose. Similar effects were min to 22 mL/min; (C) before and after sodium deoxycholate influence on the mesothelial side of the membrane. I = interstitial obtained for albumin and urea after sodium deoxy- side of the peritoneal membrane; M = mesothelial side of the peri- cholate was administered in peritoneal dialysis in toneal membrane; n = number of experiments. animals (28). Peritoneal Transport Dynamics of Glucose and Icodextrin 57

In humans using 1.5% glucose dialysis solution, Nephrol 1987; 27:51–5. peritoneal transport of glucose from the peritoneal 5 Wolfson M, Ogrinc F, Mujais S. Review of clinical cavity to the blood increased during peritonitis by trial experience with icodextrin. Kidney Int 2002; 74%. In in vivo investigations with animal models, 62(Suppl 81):S46–52. peritonitis increased peritoneal absorption of glucose 6 Cooker LA, Holmes CJ, Hoff CM. Biocompatibility of icodextrin. Kidney Int 2002; 62(Suppl 81): by 155% during a 2-hour dwell with 4.25% glucose S34–45. solution (14,29). No data are available about the ef- 7 Conroy SE, Baines L, Rodgers K, Deviren F, Verco fect of peritonitis on the absorption of icodextrin dur- SJ. Prevention of chemotherapy-induced intraperito- ing peritoneal dialysis in humans. An increase of neal adhesion formation in rats by icodextrin at a glucose polymer degradation and dialysate osmolal- range of concentrations. Gynecol Oncol 2003; 88: ity were observed in rats during peritoneal inflamma- 304–8. tion (13). 8 Hosie KB, Kerr DJ, Gilbert JA, et al. A pilot study of adjuvant intraperitoneal 5-fluorouracil using 4% ico- Conclusions dextrin as a novel carrier solution. Eur J Surg Oncol Our study results indicate that the in vitro peritoneal 2003; 29:254–60. transfer dynamics of glucose and icodextrin are dif- 9 Kerr DJ, Young AM, Neoptolemos JP, et al. Pro- longed intraperitoneal infusion of 5-fluorouracil ferent: glucose diffusion is constant, but I→M ico- using a novel carrier solution. Br J Cancer 1996; 74: dextrin transfer increases over time and is greater than 2032–5. → M I transfer. Intensification of dialysate stirring 10 Venturoli D, Rippe B. Transport asymmetry in perito- (a decrease in stagnant fluid resistance) and chemi- neal dialysis: application of a serial heteroporous cal injury to the peritoneum can both induce an in- peritoneal membrane model. Am J Physiol Renal crease in peritoneal diffusive transport of glucose and Physiol 2001; 280:F599–606. icodextrin. Glucose and icodextrin M→I transfer, but 11 Park MS. Peritoneal transport with different dialysis not I→M transfer, is restricted more by tissue barri- solutions: clinical and experimental studies [Thesis]. ers than by stagnant fluid layers. Changes in diffu- Stockholm: Huddinge University Hospital, Karolin- sive transport may be unfavorable for peritoneal ska Institute; 1995. dialysis efficiency. 12 Szary B, Czyzewska K. Diffusive permeability of rabbit peritoneum for small solutes in vitro: effect of gentamicin and insulin. Int J Artif Organs 2002; 25: Acknowledgments 290–6. This work was supported in part by State Committee 13 Wang T, Cheng HH, Heimbürger O, Waniewski J, for Scientific Research Grant No. 3 P05A 122 22, Bergström J, Lindholm B. Effect of peritonitis on 2002 – 2004 (pathophysiology of macromolecules peritoneal transport characteristics: glucose solution transperitoneal transport). versus polyglucose solution. Kidney Int 2000; 57: 1704–12. References 14 Czyzewska K, Szary B, Waniewski J. Transperito- 1 Feriani M, La Greca G, Kriger FL, et al. CAPD sys- neal transport of glucose in vitro. Artif Organs 2000; tem and solutions. In: Gokal R, Nolph KD, eds. The 24:857–63. textbook of peritoneal dialysis. Dordrecht: Kluwer 15 Flessner MF. Peritoneal transport physiology: in- Academic Publishers; 1994: 233–70. sights from basic research. J Am Soc Nephrol 1991; 2 Waniewski J, Werynski A, Heimbürger O, Park MS, 2:122–35. Lindholm B. Effect of alternative osmotic agents on 16 Wang R, Moberly JB, Martis L, et al. A rapid assay peritoneal transport. Blood Purif 1993; 11:248–64. for icodextrin determination in plasma and dialysate. 3 Henderson LW, Leypoldt JK. Ultrafiltration with Adv Perit Dial 2002; 18:91–5. peritoneal dialysis. In: Nolph KD, ed. Peritoneal di- 17 Moberly JB, Mujais S, Gehr T, et al. Pharmacokinet- alysis. Dordrecht: Kluwer Academic Publishers; ics of icodextrin in peritoneal dialysis patients. Kid- 1989: 117–32. ney Int 2002; 62(Suppl 81):S23–33. 4 Krediet RT, Boeschoten EW, Zuyderhoudt FM, Arisz 18 de Waart DR, Zweers MM, Struijk DG, Krediet RT. L. The relationship between peritoneal glucose ab- Icodextrin degradation products in spent dialysate of sorption and body fluid loss by ultrafiltration during CAPD patients and the rat, and its relation with di- continuous ambulatory peritoneal dialysis. Clin alysate osmolality. Perit Dial Int 2001; 21:269–74. 58 Czyzewska et al.

19 Gokal R, Moberly J, Lindholm B, Mujais S. Meta- dialysis: flow-through and dialysate regeneration. bolic and laboratory effects of icodextrin. Kidney Int ASAIO J 1999; 45:372–8. 2002; 81(Suppl):S62–71. 27 Flessner M, Henegar J, Bigler S, Genous L. Is the 20 Grzelak T, Szary B, Czyzewska K. Effect of icodex- peritoneum a significant transport barrier in perito- trin on transperitoneal uric acid and albumin trans- neal dialysis? Perit Dial Int 2003; 23:542–9. port in vitro. Adv Perit Dial 2004; 20:47–50. 28 Gotloib L, Wajsbrot V, Cuperman Y, Shostak A. 21 Frajewicki V, Kushnir D, Wajsbrot V, Kohan R, Acute oxidative stress induces peritoneal Shostak A, Gotloib L. Peritoneal transport after long- hyperpermeability, mesothelial loss, and fibrosis. term exposure to icodextrin in rats. Nephron 2002; J Lab Clin Med 2004; 143:31–40. 92:174–82. 29 Verger C, Luger A, Moore HL, Nolph KD. Acute 22 Gokal R. Icodextrin-associated sterile peritonitis. changes in peritoneal morphology and transport Perit Dial Int 2002; 22:445–8. properties with infectious peritonitis and mechanical 23 Zakaria ER, Carlsson O, Rippe B. Limitation of injury. Kidney Int 1983; 23:823–31. small-solute exchange across the visceral perito- 30 Simon M. Peritoneal mesothelium in vitro: an elec- neum: effects of vibration. Perit Dial Int 1997; 17: trophysiologic study. Perit Dial Int 1996; 16:393–7. 72–9. 24 Levitt MD, Kneip JM, Overdahl MC. Influence of Corresponding author: shaking on peritoneal transfer in rats. Kidney Int Krystyna Czyzewska, MD, Department of Chemistry 1989; 35:1145–50. and Clinical Biochemistry, Poznan University of 25 Ash SR, Janle EM. Continuous flow-through peritoneal dialysis (CFPD): comparison of efficiency to Medical Sciences, Swiecickiego 6 Str., Poznan 60-781 IPD, TPD, and CAPD in animal model. Perit Dial Int Poland. 1997; 17:365–72. E-mail: 26 Roberts M, Ash SR, Lee DB. Innovative peritoneal [email protected]