in vitro biocompatibility performance of physioneal

Kidney International, Vol. 64, Supplement 88 (2003), pp. S57–S74

In vitro biocompatibility performance of Physioneal

CATHERINE M. HOFF

Renal Division Research, Baxter Healthcare Corporation, McGaw Park, Illinois

In vitro biocompatibility performance of Physioneal. Peri-toneal dialysis (PD) has been a successful and effective form ofchronic renal replacement therapy since its introduction over20 years ago. Despite its overall success, there is a growingbody of evidence that suggests shortcomings in the preser-vation of membrane integrity. This has led to the develop-ment of several second-generation PD solutions that demon-strate improved biocompatibility. Physioneal, a neutral pH,bicarbonate/lactate-buffered solution, was one of the first ofthese new PD solutions to become commercially available.This review will focus on one of the first preclinical stages inthe development of Physioneal: studies on in vitro biocom-patibility testing. Studies in leukocyte, mesothelial cell, andfibroblast populations demonstrated significantly improvedbiocompatibility of neutral pH, bicarbonate/lactate-based so-lutions compared to conventional solutions. The solutionscontributed to improved leukocyte viability and response tobacterial infection (e.g., phagocytosis, superoxide radical gen-eration, and endotoxin-stimulated cytokine release). Studieson peritoneal mesothelial cells demonstrate improved cell via-bility, proliferation, and response to proinflammatory stimuli,and a reduced potential for angiogenesis and peritoneal fibro-sis, all suggesting a better preservation of membrane structureand function. The bicarbonate/lactate-based solutions demon-strated decreased cytotoxicity and preserved cell growth infibroblast cultures as well. In vitro biocompatibility testinghas clearly demonstrated that neutral pH, bicarbonate/lactate-buffered Physioneal solutions are superior to conventionalsolutions in preserving cell viability and function in cell popu-lations that contribute to peritoneal homeostasis. This positiveassessment now provides a foundation and rationale for mov-ing forward with the next stages in preclinical testing: in vivoanimal models and human ex vivo studies.

Peritoneal dialysis (PD) has been a successful andeffective form of chronic renal replacement therapy sinceit was introduced over 20 years ago. Despite its over-all success, technique survival can be compromised bycomplications of peritonitis and alterations in membranephysiology that occur in some patients with time on dial-ysis, leading to peritoneal fibrosis, increased membranepermeability, and eventual ultrafiltration failure. The con-tinued success of PD as a long-term treatment option

Key words: in vitro biocompatibility, bicarbonate/lactate-bufferedsolutions, peritoneal dialysis.

C© 2003 by the International Society of Nephrology

depends, therefore, on minimizing peritonitis and pre-serving the overall structure and function of themembrane.

There is a growing body of evidence that suggeststhat, while the conventional PD solutions perform wellin terms of metabolic waste removal and maintenance offluid and electrolyte balance, there are shortcomings interms of their ability to preserve membrane integrity andlongevity that appear to be associated with componentsof the solutions themselves. This realization has led to thedevelopment of several second-generation PD solutionsthat demonstrate improved biocompatibility. Physioneal,a neutral pH, bicarbonate/lactate-buffered solution, wasone of the first of these new PD solutions to become com-mercially available.

This review will focus on one of the first preclinicalstages in the development of Physioneal: studies on invitro biocompatibility. The role that these studies play informulation development, their intrinsic advantages anddisadvantages, an overview of the studies, the knowledgegained from them, and finally, whether these studies areultimately predictive of in vivo solution performance willbe discussed.

OVERVIEW OF BIOCOMPATIBILITY

The biocompatibility of a PD solution can be definedas the ability of the solution formulation to support long-term dialysis without causing clinically significant changesin the functional characteristics of the membrane. Morespecifically, the solution must support the dialytic func-tions of both fluid balance and waste removal while pre-serving membrane physiology. Qualities of PD solutionsthought to affect biocompatibility include solution pH,buffer components, osmolality, glucose concentration,and the glucose degradation product (GDP) profile.

The potential effects of biocompatibility factors onclinical outcomes are illustrated in Figure 1. Solutioncomponents can affect the normal physiology of thecell populations of the peritoneal cavity (e.g., leukocyte,mesothelial and endothelial cell, and fibroblast popula-tions), resulting in alterations in cytokine, chemokine,and growth factor networks, upregulation of proin-flammatory and profibrotic pathways, and induction of

S-57

S-58 Hoff: In vitro biocompatibility performance of Physioneal

PD solution biocompatibility factors

Inhibition of normal leukocyte, mesothelial, endothelial, and fibroblast functionAlteration of cytokine, chemokine, and growth factor networks

Disturbances in inflammatory and fibrotic mediatorsCross-linking of structural proteins

Carbonyl and oxidative stress

Impaired peritonealhost defense

Temporal changes in peritonealstructure and function

Altered peritoneal permeability;Altered ultrafiltration

Ultrafiltration failure

Fig. 1. Potential role of biocompatibility inclinical outcomes. Chronic exposure to bioin-compatible peritoneal dialysis (PD) solu-tions can result in perturbations in normalperitoneal membrane physiology. Over time,these changes in membrane structure andfunction can contribute to altered transportcharacteristics, peritoneal fibrosis, and scle-rosis, and ultimately to ultrafiltration failureand discontinuation of PD therapy. Reprintedwith permission from reference [122].

carbonyl and oxidative stress. These perturbations in nor-mal membrane physiology are thought to contribute toimpaired host defense, changes in peritoneal structureincluding peritoneal fibrosis and vasculopathy, andchanges in dialytic functions, including increased solutepermeability and ultrafiltration failure. The importanceof each of these biocompatibility factors and their hypoth-esized roles in cellular and membrane dysfunction hasbeen extensively reviewed [1–3]. This article will brieflydiscuss each of these factors and focus on their relativecontributions to the Physioneal solution formulation andbiocompatibility.

PHYSIONEAL PD SOLUTION

PD solutions were originally intended to be formu-lated with bicarbonate, as it is a physiologically rele-vant buffer. However, due to technological challenges inmanufacturing a stable bicarbonate-based PD solution,the first commercially available solutions were acetate-or lactate-based. The advent of solution sterilization inmultichambered bags facilitated the commercial devel-opment of bicarbonate-based solutions because they par-tition glucose at low pH in one chamber, which minimizesheat-catalyzed degradation, while keeping the bicarbon-ate component in a separate chamber at an alkaline pH.Mixing the two chambers just prior to use provides themeans to deliver bicarbonate-based, glucose-containing

Table 1. Composition of Physioneal and Dianeal formulations

Component Dianeal PD-4 Physioneal 40

Lactate mEq/L 40 15Bicarbonate mmol/L — 25pH 5.2 7.4Glucose g/L 13.6, 22.5, 38.6 13.6, 22.5, 38.6Osmolality mOsm/kg 358, 401, 511 358, 401, 511Sodium mEq/L 132 132Chloride mEq/L 95 95Magnesium mEq/L 0.5 0.5Calcium mEq/L 2.5 2.5

solutions at neutral pH. Physioneal PD solutions differfrom first generation low pH, lactate-buffered fluids suchas Dianeal, as they use 25 mmol/L bicarbonate combinedwith 10 to 15 mmol/L lactate as the buffer, and are for-mulated for delivery at a neutral pH (Table 1). WhilePhysioneal solutions are formulated with the same glu-cose concentrations as Dianeal, and therefore have thesame osmolarities, this new container system has resultedin significantly lower levels of potentially cytotoxic GDPs(Table 2). How do these changes—neutral pH, bicarbon-ate/lactate buffer and low GDPs—result in improved so-lution biocompatibility?

ASPECTS OF BIOCOMPATIBILITY

In order to fully evaluate how the Physioneal for-mulation results in improved biocompatibility, we first

Hoff: In vitro biocompatibility performance of Physioneal S-59

Table 2. Comparison of glucose degradation product concentrations in Dianeal and Physioneal

Compound Methylglyoxal Glyoxal 3-Deoxyglucosonelmol/L% Glucose Dianeal Physioneal Dianeal Physioneal Dianeal Physioneal

1.36% 2.05 ± 0.56 0.59 ± 0.28 2.17 ± 0.27 2.24 ± 0.46 97 ± 13.5 70 ± 6.62.27% 2.71 ± 0.19 0.8 ± 0.15 2.96 ± 0.51 2.49 ± 0.58 164 ± 6.1 111 ± 5.73.86% 4.08 ± 0.86 1.12 ± 0.03 3.13 ± 0.28 3.3 ± 0.61 298 ± 48 161 ± 14.1

The GDP profile was obtained from Dawnay et al: J Am Soc Nephrol 10:312A, 1999.

need to understand the effect of the individual so-lution components on normal cellular structure andfunction.

Low pH and elevated lactate

The first-generation PD solutions were based on a pH5.2, 40 mmol/L lactate formulation. It has been exten-sively documented that this combination of acidic pH,well below physiologic pH of 7.4 and superphysiologiclevels of lactate, is inhibitory to a number of cell func-tions in vitro, and is therefore a primary determinant ofPD solution bioincompatibility.

The cytotoxicity of low pH, lactate-based solutionsappears to be primarily due to the profound drop in intra-cellular pH (pHi) that occurs almost immediately uponexposure to fresh PD solution. Duwe et al [4] were thefirst to document a reduction in phagocytosis, bacteri-cidal capacity, and respiratory burst in neutrophils uponexposure to commercial PD solutions. Similarly, exposureof PMφ to conventional dialysate resulted in impairedchemiluminescence, a decreased capacity for oxygen rad-ical generation, and a decline in bacterial cell killing [5].While the decline in pHi is most severe in the presence ofboth low pH and high lactate, intracellular acidificationand impaired respiratory burst in PMφ were also notedat pH 6.5 and 7.0, indicating that the effect is at leastpartially independent of pH [abstract: Chaimovitz et al,J Am Soc Nephrol 3:407A, 1992]. A slightly different out-come was noted by Liberek et al [6], who found that whilePMφ pHi and chemiluminescence were reduced upon ex-posure to low pH, lactate-containing fluids, they were notsignificantly altered in lactate-containing neutral pH so-lutions. Increasing lactate concentration resulted in de-creased phagocytic activity in polymorphonuclear cells(PMN) at both acidic and neutral pH, although theinhibition was much more severe at pH 5.2, while cellviability and leukotriene B4 generation were not sig-nificantly changed from control under these conditions.Topley et al [7] demonstrated decreased chemilumines-cence in macrophages only in the presence of both factors,as lowering the pH of culture media to 5.2, or increasingthe lactate content at neutral pH, had no effect. Finally,improved neutrophilic superoxide production was notedin lactate-based solutions when the pH was neutralized[8].

The combination of acidic pH and lactate in con-ventional peritoneal dialysis fluids (PDF) also adverselyaffected cytokine release from endotoxin-stimulatedleukocytes. Jorres et al [9] demonstrated a decrease ininterleukin (IL)-6 and tumor necrosis factor a (TNFa)release from mononuclear leukocytes, and Douvdevaniet al [10, 11] found that inhibition of IL-1 and TNFa re-lease was maximal at low pH and high lactate, minimalwith lactate at neutral pH, and unaffected by pH alone.While the specific effects of low pH, lactate-based solu-tions appear to vary by cell type and individual parameterunder investigation, there is a consensus that exposure tothese solutions has a rapid and negative impact on cellularfunction.

Because dialysis solution equilibrates upon infusion,with pH reaching 6.8 within 30 minutes and near physio-logic pH within 120 minutes, the recovery of cell func-tion after initial exposure to PDF was examined in anumber of in vitro studies. Neutrophils reportedly failedto generate a respiratory burst after a one-hour recov-ery period in normal media, even though pHi had beenrestored [12]. Recovery of function in PMφ was delayedas well, with the failure of pHi, cytoskeletal alterations,proinflammatory cytokine release, and phagocytosis toreturn to normal levels [13–15]. It would appear that thelow pH, lactate combination has a significant impact notonly on cell function but on recovery of function as well.

Witowski et al [16] reported that exposure of hu-man peritoneal mesothelial cells (HPMC) to standardPD solutions resulted in an inhibition of IL-6 andprostaglandin E2 (PGE2) release that appeared to berelated to low pH. Low pH/high-glucose media had aninhibitory effect on fibrinolytic parameters in HPMC,with the effect of low pH stronger than that due to highglucose/hyperosmolality [17]. Lactate has been shown tomodulate glucose-mediated effects on cellular function,potentially through enhancement of intracellular sorbitolaccumulation [18, 19]. A recent study in HPMC suggeststhat exposure to glucose in the presence of lactate leadsto an increase in TGF-b and monocyte chemoattractantprotein-1 (MCP-1) to levels above those induced by glu-cose or lactate alone, possibly via enhanced activation ofthe polyol pathway and sorbitol accumulation [20]. Whilethe degree of cellular impairment varies with cell typeand functional parameter under evaluation, there is aconsensus that the combination of low pH and lactate


compromises both host defense and physiologic functionsof the membrane and its cellular components.

Glucose

Glucose has been used successfully as the soleosmotic agent in PD solutions for over 20 years. Thereis growing concern, however, regarding the long-termeffect of chronic exposure of peritoneal tissues and cellsto superphysiologic levels of glucose and the impact onperitoneal physiology. Various aspects of glucotoxicity—mechanisms by which glucose can adversely affect mem-brane structural and functional parameters—have beenextensively reviewed [3, 21–23] and will be briefly sum-marized here.

In assessing mechanisms of glucotoxicity, it is impor-tant to consider not only the direct effects of glucose oncellular parameters, but also how glucose may indirectlyinfluence cell functions through hyperosmolality andglucose degradation production (GDP)- and advancedglycation end product (AGE)-mediated reactions.Glucose-based solutions are hyperosmolar relative tonormal serum (358 to 511 mOsm vs. 280 mOsm) and anumber of studies have shown negative effects of glucoseand/or hyperosmolality on cellular function. In HPMC,the generation of reactive oxygen species, upregulationof the polyol and p38 MAPK pathways, enhanced expres-sion of genes such as transforming growth factor-beta(TGF-b), fibronectin, laminin, and MCP-1, increasedapoptosis, and increased oxidative mitochondrial dam-age have all been noted after exposure to high glucose[3, 21, 22, 24–32]. In a recent study, Tamura et al [33]found that high glucose (140 mmol/L), and to a lesserextent, high osmolality, inhibited mesothelial cell migra-tion. This effect on migration, combined with glucose-mediated damage to intercellular junctions [34], suggestshow chronic exposure to high glucose PDF could promotemembrane damage by interfering with remesothelializa-tion of the membrane after injury. Increased expressionof PGE2 [35], tissue plasminogen activator (tPA) [36], andaquaporin-1 (AQP-1) [37, 38] in mesothelial cell popula-tions appeared to be largely due to hyperosmolality.

The effects of high glucose/hyperosmolality have alsobeen documented in leukocyte populations and includereductions in cell viability, phagocytosis, bactericidalactivity, respiratory burst, and superoxide and cytokineproduction [6, 39–42]. Finally, the rate of fibroblast pro-liferation was altered in the presence of glucose, withan initial increase in cell number followed by a decreaseat later timepoints [43]. The synthesis of procollagen IIIpeptide (PIIIP) was also increased in these cells, as wasthe production of hyaluronan through enhanced proteinkinase C activity [44, 45], indicating a potential glucoseeffect on expansion of the extracellular matrix and inter-stitial space. Finally, the synthesis of vascular endothe-

lial growth factor (VEGF) by rat peritoneal vascularendothelial cells, and its upregulation by high glucose,lactate-based PDF indicates the potential for dialysis-mediated changes in the vasculature and membranepermeability [46]. Increased expression of VEGF in en-dothelial cells could contribute to an increased effectiveperitoneal surface area, increased permeability to glucoseand small solutes, and potentially predispose the mem-brane to ultrafiltration failure [47].

In summary, glucose and hyperosmolarity have the po-tential to adversely affect normal leukocyte, mesothelial,and fibroblast physiology. These factors could enhanceprocesses of inflammation, angiogenesis, peritonealfibrosis, extracellular matrix (ECM) expansion, and cellu-lar damage, all contributing factors to compromised hostdefense, decreased cell viability, profibrotic changes inthe membrane. The end result may be altered peritonealtransport and ultrafiltration failure.

Glucose degradation products

Glucose degradation products have received increasedscrutiny regarding their contribution to PD solutionbioincompatibility and membrane dysfunction. Theselow molecular weight compounds are formed primarilyfrom the breakdown of glucose during heat sterilization,but can also form during storage of glucose-containingsolutions. GDPs include acetaldehyde, formaldehyde,2-furaldehyde, 5-hydroxymethyl-furfural, glyoxal,methylglyoxal (MG), 3-deoxyglucosone (3-DG), andthe recently identified 3,4-dideoxyglucosone-3-ene(3,4-DGE)[48], as well as compounds not yet chemicallyidentified [49–51]. GDPs exert their effects via twomechanisms: directly, by affecting cell viability and cellfunction (e.g., cell proliferation, gene expression), andindirectly through increased carbonyl and oxidativestress, and by contributing to the formation of AGEs andadvanced lipoxidation end products (ALEs).

In vitro cell culture studies have demonstrated GDPcytotoxicity in a variety of cell types. Conventional heat-sterilized PD solutions inhibited cell growth in L-929 cells,an established fibroblast cell line, to a significantly greaterdegree than did the comparable filter-sterilized solutions[49]. These results were confirmed in different cell pop-ulations, with an inhibition of cellular proliferation andTNFa release in an established murine macrophage line,and diminished capacity for superoxide radical gener-ation in freshly prepared human leukocytes [52]. Theacceleration of apoptosis in peripheral neutrophils wasproposed to be due to GDPs in conventional, glucose-based PD solutions [53], as the effect was independent oflactate and osmolarity, and reduced upon exposure to lowGDP-containing bicarbonate/lactate-based solutions.

In HPMC, long-term exposure (i.e., 36 days)of cells to a mixture of GDPs (acetaldehyde,


formaldehyde, 2-furaldehyde, glyoxal, MG, and5-hydroxymethylfuraldehyde) at the concentrationsfound in high glucose conventional PDF led to a sig-nificant impairment of cell viability and function [54].The authors noted a decreased proliferative capacity,reduced constitutive and interleukin-1 (IL-1)-stimulatedexpression of IL-6, decreased secretion of fibronectin,and impaired cellular attachment to substrates, the lattersuggesting an effect on wound-healing capacity. Thesesame authors had previously concluded that short-termexposure (e.g., 24 hours) to individual GDPs had noeffect, and only in combination was a subtoxic effectseen [55], pointing out the importance of exposure timein evaluating potential effects in vitro. In contrast, aneffect of individual GDPs on HPMC cellular parameterswas demonstrated in a recently published study. Weltonet al [56] showed an uptake of 3-DG and MG by HPMC,resulting in the formation of early AGEs. A shortexposure to individual GDPs resulted in an increasedproinflammatory response as evidenced by inductionof vascular cell adhesion molecule-1 (VCAM-1) andelevated production of IL-6 and IL-8. In addition, both3-DG and MG induced apoptosis and reduced cellproliferation [56]. In HPMC, the upregulation of heatshock protein-72 (HSP-72), an indicator of cellularstress, appeared to be due to acidic pH and GDPs ina standard PDF; HSP-72 expression was significantlyreduced in neutral pH, low GDP solutions [57]. Finally,a recent study in L-929 fibroblasts showed that of all theidentified GDPs, only 3,4-DGE was inhibitory to cellularproliferation at concentrations found in conventionalPDF, leading the authors to suggest that 3,4-DGE wasthe most biologically active of all GDPs [48].

The manufacturing of Physioneal in multichamber bagshas resulted in a reduction in GDPs compared to con-ventional solutions, implying a decreased potential forGDP-mediated cellular and membrane alterations. Thisreduction is significant, especially for patients requiringhigh glucose solutions to achieve adequate ultrafiltration,as GDP levels in 3.86% Physioneal are either below orequivalent to levels in 1.36% Dianeal (Table 2).

Advanced glycation end products

Glucose can also contribute indirectly to membrane al-terations and compromised membrane function throughthe formation of AGEs. These compounds are formedwhen reducing sugars such as glucose, or reactive car-bonyl compounds (RCOs) such as GDPs, react nonenzy-matically with amino groups of proteins to form a Schiffbase, which can then undergo an irreversible reaction toform AGEs. AGEs and their lipid counterparts, ALEs,have been implicated in a variety of activities that can ad-versely affect the peritoneal membrane, including proteinand ECM crosslinking, inflammation, angiogenesis [58],

vascular smooth muscle cell proliferation [59], and in-creased nitric oxide production [60]. Peritoneal biopsiesfrom long-term PD patients show pathologic alterationsthat are strikingly similar to those of patients with dia-betiform disease, with thickening of the basement mem-brane, vascular alterations, and AGE deposition [61–64],implicating exposure to glucose and/or its degradationproducts in membrane pathogenesis. A recent histochem-ical analysis of peritoneal membrane biopsies has demon-strated a colocalization of AGEs with TGF-b , VEGF, andM-CSF, suggesting that AGE-receptor binding activatessignaling pathways leading to the expression of growthfactors and development of peritoneal fibrosis [65]. Theseobservations, when taken together with patient clinicaldata, suggest a correlation between time on dialysis, peri-toneal tissue AGE content, and membrane dysfunction[64, 66].

Several in vitro studies support a link between AGEmodification and alterations in cellular function. A studyin human microvascular endothelial cells indicated thatAGEs stimulated VEGF production [67], an associa-tion that could have a significant impact on peritonealmembrane permeability. Rashid et al [68] have demon-strated that exposure of rat PMφ to AGE-modified serumalbumin resulted in increased levels of both constitu-tive and endotoxin-stimulated TNFa under conditionswhere glucose had no significant effect. In HPMC, AGEswere shown to stimulate VCAM-1 expression [69] andreduce viability in vitro in a dose-dependent fashion[70]. Finally, Amadori modified, or early glycated albu-min, increased plasminogen activator inhibitor-1 (PAI-1)expression in HPMC [71], indicating how glycation ofproteins may directly contribute to the fibrinolytic im-balance seen in PD patients. Taken together, thesestudies provide an indication of how AGEs can affecta variety of (patho)physiologic processes in peritonealcell populations, all potentially contributing to membranedysfunction.

A study by Millar et al [72] demonstrated that a neutralpH, bicarbonate/lactate-based solution has significantlyreduced potential for AGE formation in vitro (Fig. 2). Thelower rate and degree of AGE formation over time com-pared to lactate-based solutions of corresponding glucosecontent is probably due, at least in part, to a decreasedformation of GDPs during sterilization and not to glu-cose per se. This suggests that bicarbonate/lactate-basedsolutions have a decreased potential for AGE formationin vivo, which would minimize membrane damage andpreserve membrane structure and function.

OVERVIEW OF BIOCOMPATIBILITY TESTING

This section will focus on the role of preclinical biocom-patibility testing as an integral part of PD solution devel-opment. Biocompatibility testing should ideally follow a


1400

1200

1000

800

600

400

200

0

Flu

ores

cenc

e, U

/g H

SA

/L

0 5 10 15 20Time, days

Fig. 2. In vitro advanced glycation end product (AGE) formation withstandard and bicarbonate/lactate solutions. Increase in human serumalbumin–associated fluorescence following incubation in either stan-dard lactate-buffered (L) or bicarbonate/lactate-buffered (B/L) peri-toneal dialysis (PD) fluids containing either 1.36% or 3.86% glucose.Fluorescence is in arbitrary units per gram of hyaluronan synthase(HAS)/L. Symbols are: (�) 1.36% B/L, (�) 3.86% B/L, (�) 1.36% L,( ) 3.86% L. Reprinted from reference [72] with permission from AdvPerit Dial.

hierarchical approach as previously suggested by Holmes[73, 74] in this supplement. This review will focus on thefirst stage of this process, in vitro biocompatibility test-ing, while the other stages are reviewed elsewhere in thisKidney International supplement.

In vitro biocompatibility testing

The cell culture–based assay is the primary test sys-tem of in vitro biocompatibility testing, and can bedesigned to ask specific, physiologically relevant ques-tions concerning the biological effects of solutions and/orsolution components. One of the advantages of this typeof testing is that it can focus on specific molecular path-ways in distinct homogeneous cell populations. Inherentin this approach, however, is that it obviously does not,and cannot, duplicate the in vivo environment, and there-fore does not capture the multifactorial interactions andmolecular crosstalk that occurs in vivo. With this short-coming in mind, these studies should be as representa-tive of the in vivo environment as possible, and shouldinclude testing of all relevant cell populations and cellu-lar functions that can affect the dialytic properties of themembrane.

There are a number of considerations in the design ofexperiments for in vitro biocompatibility testing. Theseinclude the selection of cell populations, identification ofrelevant functional parameters, and experimental design(see [73] for a detailed review).

Cell populations

In choosing cell populations for in vitro biocompatibil-ity testing, it is important to consider each cell type that

plays a role in maintaining peritoneal cavity physiologyand that is exposed to dialysis solutions or solution com-ponents. The cell populations examined in these studiesshould therefore include peritoneal and peripheral leuko-cytes, peritoneal mesothelial cells, peritoneal fibroblasts,and the endothelial and vascular smooth muscle cells ofthe peritoneal vasculature. The suitability of using estab-lished or immortalized cell lines versus primary culturesin these studies has been questioned. The general consen-sus seems to favor the use of primary cultures; althoughestablished cell lines are more readily available and moreadapted to cell culture, primary cultures may be morephysiologically representative of the in vivo environmentand response.

Assay parameters

In vitro biocompatibility testing should focus not onlyon basic parameters such as cell viability, proliferation,and on functions associated with peritoneal homeosta-sis, but also on involvement of pathogenic processesthat may influence clinical outcomes and membrane dys-function. The peritoneal cell populations and poten-tial assay parameters are described below and listed inTable 3.

Peripheral and peritoneal leukocytes. Peritoneal andperipheral leukocytes have been used extensively inin vitro testing. Leukocytes have been obtained fromperipheral blood of healthy volunteers, peritonealmacrophages from nonuremic patients undergoinglaproscopy, and peritoneal macrophages and neutrophilsisolated from the effluents of continuous ambulatoryperitoneal dialysis (CAPD) patients. The use of effluent-derived cells that have been exposed to PD solutionsraises questions about preconditioning or activation, andhow that may influence subsequent testing. It should notpreclude the use of these cells, but their origin should befactored into any interpretation of the results.

Resident and circulating leukocytes perform a varietyof functions depending on their cell type. They act as thefirst line of host defense in the peritoneal cavity, and in-teract with the mesothelial cell and fibroblast populationsthrough the production of various signaling molecules. Alarge body of evidence has shown that conventional PDFmarkedly suppresses functional capabilities of periph-eral blood leukocytes and PMφ. Therefore, parameterssuch as phagocytosis, bacterial cell killing, and productionof reactive oxygen species should be routinely tested asindicators of biocompatibility in these cells.

Peritoneal mesothelial cells. Peritoneal mesothelial cellshave also been extensively used in in vitro biocompati-bility testing. Mesothelial cells are isolated from omentaltissue of consenting patients undergoing elective surgery,expanded in cell culture, and generally used within thefirst several passages. HPMC have also been isolated from


Table 3. Peritoneal cell populations and assay parameters

Cell type Assay parameter Function

Leukocytes Proliferation (N, M)LDH release, ATP content

(N, M)pHi (N, M)

Cell viability

Phagocytosis (N, M)Bactericidal activity (N, M)Respiratory burst/

chemiluminescenceSuperoxide production (N)Cytokine expression (M

Host defense

IL-1, TNFa, IL-8, TGF-b Inflammatory signalingMesothelial

cellsProliferationOxidative damageApoptosis

Cell viability

Migration Tissue repair, woundhealing

Cytokine and growth factorexpression:

IL-1, IL-6, MCP-1,RANTES, PGE2, PGI1,

Host defense

tPA, PAI-1, TF, MMPs,TIMP

Coagulation, fibrinolyticbalance

TGF-b , CTGF, fibronectin,laminin, PIIINP, PIICP,collagen

ECM deposition, tissuerepair

VEGF

Aquaporins

Angiogenesis, vascularpermeability, watertransport

HSP-72 Stress responseHA, GAGs Lubrication

Fibroblasts Proliferation Cell viabilityTransdifferentiation to

myofibroblastsTissue repair, response to

injury and inflammationLeukocyte trafficking Response to inflammationHAPICP [119],PIIINP [120]

Tissue repair, ECMexpansion

Endothelialcells

Proliferation Cell viability

VEGF, VEGFr AngiogenesisAquaporin expression Water transporteNOS, iNOS NO signaling

Abbreviations are: N, polymorphonuclear neutrophils; M, monocytes,macrophages; LDH, lactate dehydrogenase; pHi, intracellular pH; IL-1,interleukin-1; TNFa, tumor necrosis factor a; IL-6, interleukin-6; IL-8,interleukin-8; TGF-b , transforming growth factor-b ; MCP-1, monocyte chemoat-tractant protein-1; RANTES; regulated upon activation, normal T cell ex-pressed and secreted; PGE2, prostaglandin E2; tPA, tissue plasminogen activator;PAI-1, plasminogen activator inhibitor-1; MMP, metalloproteinase; TIMP, tissueinhibitor of metalloproteinases; TGF-b , transforming growth factor b ; CTGF,connective tissue growth factor; PIIINP, procollagen III peptide; PIICP, procol-lagen type II carboxyterminal propeptide; VEGF, vascular endothelial growthfactor; VEGFr, VEGF receptor; HSP-72, heat shock protein-72; HA, hyaluro-nan; GAG, glycosaminoglycan; ECM, extracellular matrix; PICP, procollagen IC-terminal peptide; eNOS, endothelial nitric oxide synthase; iNOS, inducible ni-tric oxide synthase; NO, nitric oxide.

peritoneal effluent, but these are probably best suited forimmediate phenotypic characterization.

Mesothelial cells play an active role in peritoneal mem-brane physiology. They maintain a fibrinolytic balance inthe peritoneal cavity with the production of fibrinolyticand coagulant molecules [75–78], participate in mem-brane repair and remodeling [79–85], are active in hostdefense and response to inflammation with the produc-tion of a variety of cytokines, chemokines, and growth

factors [29, 35, 86–91], and affect peritoneal transportthrough the production of molecules that have a directeffect on the vasculature [37, 92–96]. The mesothelial cellparameters assayed in biocompatibility studies shouldtherefore include measures of mesothelial cell health,fibrinolytic and coagulant activity, host defense, tissuerepair, and angiogenesis.

Peritoneal fibroblasts. Peritoneal fibroblasts can also beisolated from omental tissue and established in cell cul-ture. There has been considerably less biocompatibilitytesting in primary fibroblasts as compared to leukocytesand mesothelial cells to date, although solution biocom-patibility has been evaluated in murine L-929 cells, anestablished fibroblast cell line.

Peritoneal fibroblasts are involved in response to in-flammation and leukocyte trafficking, and contributeto membrane structure through the synthesis of ECMcomponents, and participate in tissue repair and woundhealing. Cell proliferation is increased in cultured fibrob-lasts in response to the proinflammatory cytokines IL-1and TNFa, as is the synthesis of prostaglandins PGE2

and PGI1 [97], chemokines and cytokines such as MCP-1, IL-6, and IL-8 [98, 99], and ECM components andglycosaminoglycans such as hyaluronan [45, 97, 100].Proliferation of fibroblasts and their conversion to a my-ofibroblast phenotype, a prelude to developing a profi-brotic state, has been observed upon repeated exposureto infected PD effluent [101], and may be enhanced bychronic exposure to glucose and/or GDPs and AGEs.Biocompatibility testing in fibroblasts should include in-dicators that are reflective of maintaining the normalfibroblast physiology, response to infection, and tissuerepair.

Endothelial cells. Peritoneal endothelial cells and vas-cular smooth muscle cells comprise and dictate the struc-tural and functional characteristics of the peritonealvasculature, and like mesothelial cells and fibroblasts, canbe isolated from omental tissue. Capillary endothelialcells perform a number of functions, including regula-tion of angiogenesis and vessel permeability through theexpression of VEGF and VEGF receptors, and aquapor-ins or water channels equivalent to the ultrasmall poresof the peritoneum. The production of nitric oxide byendothelial nitric oxide synthase (eNOS) and induciblenitric oxide synthase (iNOS) is thought to play a signif-icant role in the regulation of peritoneal permeability[102, 103] and in the induction of angiogenesis. Biopsystudies from long-term PD patients have revealed anincreased number of vessels per area, vascular wall thick-ening in small arteries, and capillary vasodilatation [104]and increased expression of eNOS in the vasculature [60],suggesting that changes in endothelial cell physiologycontribute to declining peritoneal transport and even-tual ultrafiltration failure. Evaluation of markers of vas-cular permeability and angiogenesis would therefore be


recommended for biocompatibility testing in endothelialcells.

The cellular parameter(s) selected for assay willdepend on the solution component being investigatedand the cell population being tested. For example, itis physiologically relevant to assay pHi in PMNs be-cause of acute exposure upon instillation of PD solutionin vivo, but not practical to assay in fibroblasts becausethey are not directly exposed to solution. The effect ofGDPs on mesothelial cell function should be assayed be-cause these cells are directly and chronically exposed tosolution. The effects should be evaluated in fibroblastsand endothelial cells as well. Even though these cells arenot in direct contact with solution, it is known that GDPsare transported from the peritoneal lumen across themembrane into the interstitium [105–107], thereby mak-ing fibroblasts and endothelial cells vulnerable to GDP-and AGE-mediated modifications.

Experimental design

Once both the cell population and functional pa-rameter to be investigated are chosen, studies must bedesigned to be as representative as possible of in vivoconditions. One aspect to consider in experimental de-sign is time of exposure to the solution under evaluation.A short exposure time, for example, up to 60 minutes, maybe suitable to measure a rapid response, such as changesin pHi and host defense characteristics. Jorres [108] hasrecommended exposure times of less than 30 minutes toundiluted solutions to determine the effects of pH andbuffer components, and times of less than 4 hours to ex-amine the effects of osmolarity. Evaluation of changes ingene expression may require longer periods of exposure(e.g., 24 to 72 hours) in order for changes in mRNA orprotein levels to become detectable.

Whether or not the test solution should be dilutedwith culture medium is also a consideration. For acute,short-term exposures of less than 30 minutes, the use ofundiluted solution may be acceptable. This is also ap-propriate for exposure/recovery studies in which cellsare transferred from a short incubation in test solutioninto normal culture medium. The functional parametermay be then assayed directly, or a specific stimulus (e.g.,endotoxin, cytokine) added and cellular response moni-tored. A short, acute exposure would be appropriate forevaluating the effect on intracellular pHi and responseto endotoxin challenge. It is well known that PD solu-tion undergoes a rapid equilibration upon instillation intothe peritoneal cavity, with neutralization of pH, dilutionwith residual solution, and absorption of glucose overtime, so it is unrealistic to conduct longer-term incuba-tions in undiluted solutions. On a practical note, cells thatnormally obtain their nutrients from the interstitial fluidneed an exogenous supply of nutrients to maintain basic

growth parameters over longer periods of time, and it isbecoming the standard practice in long-term experimentsto dilute dialysis solutions two-fold with serum-free orlow serum culture medium.

Another type of study that is appropriate for experi-mental testing involves the use of preconditioned media.In this system, one cell population is exposed to the testsolution, the supernatant collected, and then transferredto a second culture of the same or different cell type,and response monitored. This experimental design allowsevaluation not only of the effects of solution but also theeffects of the factors produced in response by the firstcell population. For example, leukocyte or mesothelialcells could be exposed to the test solution, their super-natants collected and transferred to fibroblast cultures,and the effect(s), such as changes in gene expression,assayed. In this scenario, the fibroblasts are exposed notonly to the solution components, but to factors producedby the leukocytes and mesothelial cells. This type of studyrecognizes the importance of secondary cell signaling andattempts to duplicate the cross-talk between cells and cellpopulations. It is probably also more indicative of the invivo situation where a cell type may respond only mini-mally to a component of the PD solution, but exhibit amuch more significant response to factors produced byother cell populations.

Cell culture systems have been described that are morerepresentative of in vivo conditions than the traditionalgrowth in two-dimensional cultures. Beavis et al [101]have described a unique cell culture system in whichfibroblasts are grown in a three-dimensional collagen ma-trix, a system that better reflects the in vivo environmentthan the traditional two-dimensional cell culture. Cellculture media and supernatants can be added directlyto these cultures and responses such as proliferation,cytokine and growth factor secretion, and migrationmonitored.

IN VITRO BIOCOMPATIBILITY STUDIESON BICARBONATE/LACTATE SOLUTIONS

The studies examining the biocompatibility ofbicarbonate/lactate-based solutions are listed in Table 4and reviewed below.

Studies with leukocyte populations

A study by Rogachev et al [109] was the first to exam-ine the biocompatibility of bicarbonate/lactate- and purebicarbonate–buffered solutions in PMφ. Cells were iso-lated from the overnight effluent of CAPD patients whohad been peritonitis free for the previous six months,and the effects of the solutions on pHi and cytokineproduction measured. Upon exposure to 1.5% Dianeal(1.5-D), the pHi dropped sharply to below 6.2; however,


Table 4. In vitro biocompatibility studies outcomes

BioincompatibleCell type Lactate/Dianeal Bicarbonate/Lactate factor Reference

Peritoneal Decreased pHi < 6.2 Maintained pH 6.6–6.8 pH [109]macrophage

Reduced endotoxin-stimulated Improved TNFa release pH [109]TNFa release

PMN Decreased cell viability No change pH [110]Decreased chemiluminescence No change pH [110]Decreased phagocytosis to Decreased phagocytosis to pH [110]

<20% of control 40–55% of controlDecreased superoxide generation Improved superoxide generation GDPs [111]Decreased phagocytosis Improved phagocytosis GDPs [111]

PBMC Decreased TNFa secretion Improved TNFa secretion GDPs [111]HPMCs Decreased cell viability Improved cell viability pH [110]

Decreased cell viability Improved cell viability pH; secondary effect [113]of GDPs

Increased DNA damage No effect pH, lactate [113]Decreased proliferation Improved proliferation Osmolarity [113]Decreased proliferation Improved proliferation [121]Decreased IL-6 No effect pH [110]Increased VEGF Increased VEGF GDPs [114]Increased TGF-b No change GDPs [114]Increased PIIINP Increased PIIINP GDPs [114]

Fibroblasts Decreased proliferation Improved proliferation GDPs [115]

pHi, intracellular pH; TNFa, tumor necrosis factor a; GDP, glucose degradation product; IL-6, interleukin-6, VEGF, vascular endothelial growth factor; TGF-b ,transforming growth factor-b ; PIIINP, procollagen III N-terminal peptide.

adjustment of Dianeal to pH 7.3 prior to testing main-tained the pHi between 6.6 and 6.8. Exposure to a 1.5%25 mmol/L bicarbonate/15 mmol/L lactate solution, pH7.3 (1.5-B/L) reduced pHi relative to the control medium,but it remained above pH 6.5 for the duration of the as-say. Having demonstrated that the neutral pH, bicarbon-ate/lactate formulation prevented a drastic decline in pHi,the authors examined whether other previously noted ad-verse effects of standard solutions were alleviated. Secre-tion of TNFa in response to an endotoxin challenge wasselected as a representative cellular parameter and in-dicative of PMφ participation in host defense.

PMφ were exposed to test solutions for 15 to 30 min-utes, then transferred to growth medium, challengedwith lipopolysaccharide (LPS), and TNFa secretion mea-sured. 1.5-D reduced TNFa production by 59% and by>95% at 15 and 30 minutes, respectively, compared tocontrol medium (Fig. 3). 1.5-B/L improved PMφ responseat both time points with a significant change from controlonly at 30 minutes. TNFa secretion in bicarbonate/lactatesolutions containing 4.25% dextrose decreased comparedto 1.5% at both times, but the levels were still improvedover those of 1.5-D at 15 minutes. This study suggests thatbicarbonate/lactate-buffered solutions are more biocom-patible with regards to PMφ function, maintaining pHi

closer to physiologic levels and preserving cytokine se-cretion in response to endotoxin, although the levels werestill somewhat below control. The improved PMφ func-tion in 4.25-B/L over that in 1.5-D suggests that pH, andnot high glucose, is primarily responsible for altered PMφ

physiology.

45004000

35003000

25002000

15001000500

0

TN

Fα,

pg/

106

cells

HBSS 1.5−B/L 4.25−B/L 1.5−D

15 minute exposure

30 minute exposure

Fig. 3. Tumor necrosis factor a (TNFa ) secretion from macrophage(PMφ) following exposure to lactate- and bicarbonate/lactate-bufferedperitoneal dialysis (PD) solutions. PMφ were incubated in PD so-lutions for 15 or 30 minutes and then stimulated in medium withlipopolysaccharide (10 lg/mL). Shown is the median supernatant lev-els of TNFa after a 5-hour stimulation. HBSS, Hank’s Balanced SaltSolution; 1.5-B/L, bicarbonate/lactate-buffered solution, 1.5% dex-trose; 4.25-B/L, bicarbonate/lactate-buffered solution, 4.25% dextrose;1.5-D, Dianeal, 1.5% dextrose. ∗Significantly lower than HBSS (P <

0.05); †Significantly lower than 1.5-B/L (P < 0.05). Adapted fromreference [109].

A study on PMN functional parameters conductedby Topley et al [110] also suggested that neutral pH,bicarbonate/lactate-buffered solutions show improvedbiocompatibility over conventional solutions. PMNs iso-lated from healthy donors were exposed to test solutionsfor 30 minutes and the effect on cell viability, phago-cytosis, and chemiluminescence analyzed. Test solutions


included Dianeal, and 25 mmol/L bicarbonate/15 mmol/Llactate-based solutions at pH 6.8 and 7.4, all contain-ing 1.36% glucose. Only Dianeal caused a significantdecrease in cell viability, as evidenced by a decline incellular ATP content to 48% of control levels. Thebicarbonate/lactate solutions at pH 6.8 and 7.4 producedresults equivalent to the control medium, as did Dianealadjusted to pH 7.4 prior to assay. Similarly, LDH release,another measure of cytotoxicity, was elevated above con-trol levels only with Dianeal, pH 5.2.

Chemiluminescence was severely limited by the con-ventional solution, while other tested solutions includingDianeal pH 7.3 were not significantly different from con-trol. Phagocytic capacity was decreased to 40% to 55%of control by all tested solutions except Dianeal, whichdecreased levels to <20% of control. Taken together,these results illustrate a common finding that not all cellfunctions are affected to the same degree by a partic-ular solution (i.e., the bicarbonate/lactate solution hadno significant effects on cell viability and chemilumines-cence, but caused a 50% decrease in phagocytic capac-ity). Nonetheless, they do show that bicarbonate-based,neutral pH solutions are more biocompatible and inducea lower degree of cellular impairment than the conven-tional solution and that impairment of leukocyte functionis mediated by pH.

Studies by Topley [110] and Rogachev [109] com-pared the biocompatibility of pure bicarbonate- andbicarbonate/lactate-buffered formulations in leukocytepopulations. While there were some minor differencesbetween these two buffer systems with regards to a spe-cific cell function, there was no evidence to establish theoverall superiority of either buffer system.

The improved biocompatibility of bicarbonate/lactate-buffered solutions on leukocyte function was confirmedin another series of in vitro assays [111]. PBMCs fromhealthy donors were incubated in two-fold diluted testsolutions for four hours, challenged with endotoxin, andTNFa release assayed. Tested solutions included 1.5-D,4.25-D, 1.5-B/L, 4.25-B/L, and 4.25% filter sterilizedDianeal (4.25-DF), all neutralized prior to testing to elim-inate the effect of pH. TNFa secretion decreased afterexposure to all solutions, with the greatest inhibitionseen with 4.25-D, suggesting an effect of glucose concen-tration or hyperosmolality (Fig. 4A). However, becausethere was no significant difference between 1.5-B/L and4.25-B/L, it is more likely that the inhibitory effect inthis test system is due to the presence of GDPs in theDianeal solutions. This theory was strengthened by thesignificant improvement in secretion with 4.25-B/L and4.25-DF.

Superoxide production, phagocytic capacity, anddegranulation, three indicators of PMN function, werefound to be better preserved with the 25 mmol/L bicar-bonate/15 mmol/L lactate solutions [111]. As shown in

Figure 4B, superoxide production was not significantlychanged from control levels with either 1.5-D or 1.5-B/L,while the corresponding 4.25% solutions resulted in a sig-nificant decrease. The finding that superoxide productionwas least affected by 4.25-DF suggests that it is influencedin part by factors other than glucose and osmolality, suchas GDPs.

Normal phagocytic function was also maintained toa greater extent with bicarbonate/lactate solutions overconventional solutions of the same glucose concentra-tion and osmolality (Fig. 4C). Increasing the glucose con-centration had an adverse effect on phagocytic activity,regardless of the buffer system. 4.25-DF preservedphagocytosis better compared to 4.25-D and 4.25-B/L,again emphasizing the potential negative effect of GDPson cellular function.

These authors also examined PMN degranulationcapacity, as measured by endotoxin-stimulated BPI pro-tein release. PMN exposed to bicarbonate/lactate solu-tions showed better preserved degranulation capacitycompared to those exposed to Dianeal solutions of thesame glucose concentrations. At the higher glucose con-centration of 4.25%, degranulation capacity was signifi-cantly reduced only in Dianeal. There was no significantdifference in BPI release between 4.25-B/L and 4.25-DF,and these levels approximated those found with 1.5-Dand 1.5-B/L. Sundaram concluded that the improvedbiocompatibility of bicarbonate/lactate solutions may beattributable, at least in part, to their lower GDP profile.Levels of acetaldehyde and 5-HMF in the tested solu-tions were highest in the heat-sterilized fluids and lowestin the filter-sterilized and bicarbonate/lactate solutions[111], supporting this hypothesis.

In summary, in vitro testing in leukocytes con-firmed the improved biocompatibility of neutral pH,bicarbonate/lactate-buffered PDF. The solution pre-served pHi on acute exposure, improved cell viability,and response to bacterial infection (e.g., phagocytosis,superoxide radical generation, and endotoxin-stimulatedcytokine release). While not assayed in these studies, itcan be inferred that the increased apoptosis and necrosisobserved upon exposure to conventional fluids, thoughtto be due to GDPs [53, 112], would be ameliorated withthe new solution. These observations suggest that Phys-ioneal would preserve leukocyte function in vivo, result-ing in a more normal response to, and decreased damagefrom, infectious peritonitis.

BIOCOMPATIBILITY STUDIES WITHPERITONEAL MESOTHELIAL CELLS

A number of studies have examined the biocompat-ibility profile of neutral pH, bicarbonate/lactate-basedsolutions, such as Physioneal, in mesothelial cells.An early study by Topley et al [110] demonstrated


16

14

12

10

8

6

4

2

0

RPMI

1.5−D

1.5−B

/L

4.25

−D

4.25

−B/L

4.25

−DF

HBSS1.

5−D

1.5−B

/L

4.25

−D

4.25

−B/L

4.25

−DF

TN

Fα,

pg/

2.5

×

nm

ol/L

of O

2, 5

×

50

40

30

20

10

0

P < 0.001

P < 0.001

P < 0.001

P = 0.002

P = 0.01

P = 0.006 NS

NS NS

A B

106

PB

MC

106

PM

Ns/

10 m

in

HBSS1.

5−D

1.5−B

/L

4.25

−D

4.25

−B/L

4.25

−DF

60C

50

40

30

20

10

0

Pha

gocy

tosi

s in

dex

P < 0.001

P < 0.001 P < 0.001

P < 0.001P = 0.02

that a neutral pH, bicarbonate/lactate solution betterpreserved cell viability and cytokine production. Cellviability, measured as intracellular ATP content, wasseverely depressed after a 30 minute exposure of growtharrested cells to 1.36% Dianeal. Exposure to a 25 mmol/Lbicarbonate/15 mmol/L lactate-based formulation at pH7.4, and to Dianeal adjusted to pH 7.4 improved cellviability, although it remained depressed relative to con-trol. When the cells were transferred to M199 culturemedium post-exposure, ATP levels returned to normalwithin a 60-minute recovery period, with no differencesbetween test and control cells. However, the ATP con-tent of cells exposed to pH 5.2 Dianeal declined there-after, implying a delayed effect of exposure to low pHsolution.

The effect on cell viability was also investigated byHa et al [113]. In a slightly different experimental for-mat from that of Topley, growth-arrested synchronizedHPMC were exposed to undiluted solutions for up to

Fig. 4. Effects of bicarbonate/lactate- and lactate-based peritonealdialysis (PD) solutions on leukocyte functions. The effects of 1.5% Di-aneal (1.5-D), 4.25% Dianeal (4.25-D), filter-sterilized 4.25% Dianeal(4.25-DF), 1.5% bicarbonate/lactate solution (1.5-B/L), and 4.25% bi-carbonate/lactate solution (4.25-B/L) were assayed in peripheral bloodmononuclear cells (PBMC) and polymorphonuclear cells (PMN) pop-ulations. RPMI culture medium and Hank’s Balanced Salt Solution(HBSS) served as a control media. (A) PBMC were exposed to test so-lution for 4 hours, resuspended in RPMI containing 10 ng/mL endotoxinfor an additional 20 hours, and TNFa production assayed. (B) PMNwere exposed to test solution for 30 minutes, resuspended in HBSS,and superoxide production assayed. (C) PMN were exposed to test so-lution for 30 minutes, resuspended in HBSS, and uptake of heat-killed14C-labeled Staphylococcus aureus over a 30 minute period measured.Phagocytosis is expressed as a ratio of cell-associated counts/total countsadded. Adapted from reference [111].

96 hours, and cell death measured by LDH release.Exposure of cells to 1.5-D, 4.25-D, and 4.25% Physioneal40 (4.25-P) resulted in a 60% increase in LDH releaseat 24 hours, with levels remaining elevated at 96 hours.1.5-P had no significant effect. When the exposure timeto undiluted solutions was shortened to 60 minutes withsubsequent recovery in M199, LDH was increased onlyin 1.5-D (4.25-D was not tested), with no effect seen with1.5-P and 4.25-P. These results imply that cell viability isinfluenced primarily by pH. However, when test solutionswere diluted two-fold in M199 prior to exposure (effec-tively neutralizing solution pH), there was no increasein LDH over control at 24 hours, while at 96 hours only4.25-D increased LDH values (65% vs. 20% for control).The finding that a two-fold dilution of 4.25-D, but not4.25-P, significantly increased LDH release suggests thathigh glucose may not be a major determinant of HPMCcytotoxicity, and that another component of 4.25-D, suchas GDP content, is more influential.


DNA damage, another marker of cell viability, wasmeasured after a 60-minute exposure to undiluted testsolutions. Exposure to both 1.5-D and 4.25-D, but not toeither 1.5-P or 4.25-P, resulted in measurable damage, theextent of which could be attenuated by increasing the pHof Dianeal to 7.4 prior to assay.

Ha et al [113] evaluated the effect of Physioneal ona third parameter of HPMC function—proliferation.Quiescent HPMC were incubated for up to 96 hours witheither control solution (M199) or test solutions dilutedtwo-fold with M199. Cell proliferation was unchanged incells exposed to 1.5-P, but was decreased in 1.5-D, 4.25-D,and 4.25-P at 24 hours, and in 4.25-D and 4.25-P at96 hours, suggesting an effect of glucose or hyperosmolal-ity. When the cells were subjected to an acute 5-minuteexposure to undiluted solution followed by recovery inM199, cell proliferation was significantly inhibited with1.5-D and 4.25-P, but not by 1.5-P or, interestingly, 4.25-D.The authors’ explanation for this apparently inconsis-tent finding is that cell proliferation is adversely affectedby hyperosmolality, which accounts for the inhibition by4.25-D and 4.25-P when cells are exposed over time to thetwo-fold diluted solutions, and for the inhibition of 1.5-Dand 4.25-P upon acute exposure. However, after acuteexposure to undiluted 4.25-D, cellular mechanisms acti-vated in response to high osmolality, low pH, and highGDPs may have overcome the inhibitory effects of thesolution components, giving the impression that 4.25-Ddoes not negatively affect HPMC proliferation. Thiswould imply that these “response” mechanisms were notactivated in the presence 4.25-P.

A study recently published in abstract form com-pared commercially available 25 mmol/L bicarbonate/15 mmol/L lactate- and 40 mmol/L lactate-based so-lutions with filter-sterilized 40 mmol/L lactate-bufferedsolutions formulated with 1.36%, 2.27%, and 3.86%glucose [abstract; Boulanger et al, Perit Dial Int 23(Suppl1):S8, 2003]. HPMC proliferation was inhibited to a lesserdegree in bicarbonate/lactate versus lactate solutions atall glucose concentrations tested. Even when the buffersystems were equalized with the addition of lactate orbicarbonate, the inhibitory effects were greater in thelactate solutions, suggesting the importance of GDPs. Inaddition, cell death and apoptosis were higher in 3.86%lactate-buffered PDF than in the filtered and bicarbon-ate/lactate solutions, implying that the high levels ofGDPs in the conventional lactate solutions reduce cellviability independently of buffer and pH.

Solution biocompatibility in HPMC was also evaluatedin terms of cytokine production. IL-1–mediated IL-6 pro-duction was significantly impaired relative to control afterexposure to undiluted 1.36-D, but not to a 1.36-B/L, pH7.4 solution, or Dianeal that had been adjusted to pH 7.4prior to exposure, suggesting pH as the causative agent[110].

The effect of PD solutions on the secretion of growthfactors and extracellular matrix proteins in HPMC wasalso evaluated. Cells were incubated for 48 hours in testsolutions diluted two-fold with M199 culture medium,and the secretion of VEGF, TGF-b , procollagen IC-terminal peptide (PICP), PIIINP, and fibronectin as-sayed [114]. VEGF secretion was significantly enhancedrelative to control upon exposure to 1.5-D, 4.25-D, and4.25-P, but not 1.5-P (Fig. 5A). For each solution tested,VEGF secretion was higher upon exposure to 4.25% ver-sus 1.5% solutions, suggesting a role for glucose. How-ever, the extreme difference between 4.25-D and 4.25-P,and the fact that VEGF levels with 1.5-D, 1.5-P, and 4.25-Pwere all significantly lower than 4.25-D (22, 18, 28 vs.65 ng VEGF/mg protein, respectively) suggests amore substantial role of GDPs in VEGF produc-tion. Supporting evidence comes from a study byMandl-Weber et al [93], who found that nonen-zymatic glycation products upregulated VEGF pro-duction in HPMC, while glucose had only a slightinhibitory effect. HPMC production of PIIINP dis-played the same pattern of secretion as VEGF, with in-creased synthesis in all solutions, and levels in 4.25-Dand 4.25-P higher than 1.5-D and 1.5-P, respectively, and4.25-P below 1.5-D (Fig. 5B). In addition, TGF-b produc-tion was increased only in response to 4.25-D (Fig. 5C),again implicating GDPs in regulation of protein secre-tion. All solutions upregulated PIIIN to various degrees,with both Dianeal solutions producing higher levels thanthe Physioneal solutions, and the 4.25% solutions alwaysthe most inhibitory of the two glucose concentrations. Theauthors suggest that both high glucose and GDPs mayinfluence the production of VEGF, TGF-b , and PIIINPto varying degrees.

The results from Ha illustrate how solution compo-nents affect HPMC production of angiogenic factors andECM components, both of which contribute to peritonealfibrosis and alterations in membrane ultrafiltration, andhow these changes are greatly reduced in the presenceof Physioneal. That there was no significant increase insecretion of either PICP or fibronectin (Fig. 5D) withany of the solutions tested should be interpreted withcaution. These investigators have shown in an earlierstudy that high glucose upregulates both TGF-b and fi-bronectin through activation of protein kinase C [83],and suggest that conditions in the present study may notbe conducive to fibronectin expression or that the solu-tion may also contain negative regulators of fibronectin[114].

In summary, the studies on mesothelial cellsindicate an improved biocompatibility of neutralpH, bicarbonate/lactate-based PDF, as evidenced byimproved cell viability and proliferation, and responseto proinflammatory stimuli. The changes in expressionof various growth factors and ECM components imply a


Fig. 5. Effects of peritoneal dialysis (PD) so-lutions on the secretion of growth factor andextracellular matrix (ECM) proteins by hu-man peritoneal mesothelial cells (HPMC).HPMC were cultured in various PD fluidsdiluted two-fold with serum-free M199 for48 hours, and the secretion of variousgrowth factors and ECM proteins mea-sured by enzyme-linked immunosorbentassay (ELISA) [vascular endothelial growthfactor (VEGF), transforming growth factor-b(TGF-b), and fibronectin) or radioimmunoas-say (RIA) [procollagen III peptide (PIIINP)].Test solutions used were: 1.5% Dianeal(1.5-D), 4.25% Dianeal (4.25-D), 1.5% Phys-ioneal (1.5-P), and 4.25% Physioneal (4.25-P).Results are given as ng protein/mg total cel-lular protein. ∗P < 0.05 compared with M199;†P < 0.05 compared to 1.5-D). Adapted fromreference [114].

reduced potential for changes in angiogenesis, peritonealtransport, and peritoneal fibrosis, all suggesting the possi-bility for better preservation of membrane structure andfunction.

PERITONEAL FIBROBLASTS

The improved biocompatibility of a bicarbon-ate/lactate solution was demonstrated in L-929 fibroblastsupon exposure to two-fold dilutions of 3.86-D, filter-sterilized 3.86%D (3.86-DF), and a 25 mmol/L bicar-bonate/15 mmol/L lactate solution also containing 3.86%glucose. The proliferative capacity of cells exposed tothe bicarbonate/lactate solution showed a minor, non-significant decrease relative to control and the filter-sterilized Dianeal, while exposure to 3.86-D resulted ina greater than 60% inhibition of cell growth [115]. Thisinhibition was attributed to the presence of GDPs, asboth the bicarbonate/lactate and the filtered solutionsshowed decreased or undetectable levels of acetalde-hyde. These results support an earlier study in primaryperitoneal fibroblasts that suggested that bicarbonate-buffered solutions were less cytotoxic than conventionalsolutions as measured by preservation of IL-1 mediatedIL-6 release, and that peritoneal fibroblasts appeared tobe more resistant to PD solution–mediated cytotoxicitythan primary mesothelial cells [108].

OVERVIEW OF IN VITROBIOCOMPATIBILITY TESTING

In vitro cell culture studies, the first level of preclini-cal biocompatibility testing, provide a platform for eval-uating the effect of specific solution components on cellviability and function. These studies show quite clearlythat in leukocyte, mesothelial, and fibroblast cell pop-ulations Physioneal is more biocompatible and betterpreserves cellular function in vitro than conventionalacidic lactate-based PDF. The improved biocompatibil-ity appears largely attributable to the change to neutralpH, bicarbonate/lactate buffer and the reduction in GDPlevels (see Table 4).

It is notable that none of the reviewed studies identifiedglucose as a primary causative agent of bioincompatibil-ity, despite numerous in vitro studies that have showndefinite adverse effects when added directly to culturedcells. While high glucose (4.25%) lactate PDF provedmore detrimental to cell viability and cell function thanthe 1.5% lactate solutions, 4.25% bicarbonate/lactatesolution preserved normal cellular physiology to a greaterdegree, often having an effect similar to 1.5% lactate, oreven control media. In comparisons of filter- and heat-sterilized solutions, the biocompatibility of high glucosefilter-sterilized media was decidedly superior, and of-ten comparable to that of low glucose or control media.


These two observations suggest that a factor in PDF otherthan, or in addition to, glucose has a strong influence onbiocompatibility, and the other factor is most likely thepresence of GDPs. This should not be interpreted todownplay the contributions of glucose to bioincompat-ibility or to negate the numerous studies showing aglucotoxic effect, but suggests that under the conditionsassayed in the reviewed studies, GDPs may have a higherdegree of cytotoxicity.

The majority of the studies reviewed here suggestGDPs are causative agents of bioincompatibility, as theyare implicated in impaired leukocyte, mesothelial cell,and fibroblast function (Table 3). This is an interest-ing observation, especially in light of cell culture studiesusing pure forms of individual GDPs. These studies showvariable outcomes, ranging from effects only after long-term [54] versus short-term [56] exposure, from no effectsto effects only with selected GDPs [48, 56].

Physioneal has achieved increased biocompatibilityover conventional solutions through a change to abicarbonate/lactate-based, neutral pH solution, with asignificantly lower GDP content despite its continued useof glucose as the osmotic agent. In light of the changed for-mulation, how significant will glucose-mediated effects beon membrane function and longevity? It is possible that asoverall solution biocompatibility improves, as appears tobe the case with Physioneal, normal membrane structuraland functional characteristics will be better maintainedover the course of the therapy, resulting in preservationof ultrafiltration capacity and transport status. Improvedmembrane viability may reduce or moderate the overallcytotoxic effects of glucose in two ways. First, preserveddialysis capacity may reduce the need to use high glucosesolutions to achieve adequate ultrafiltration, and resultover time in lower solution glucose exposure. Second, ahealthier membrane may be more resistant to, and betterable to recover from, glucose-mediated damage.

An important message derived from these studies isthat solution components can affect cellular functions indifferent ways. The observations that cell viability is af-fected mainly by GDPs and low pH, DNA damage isrelated to pH, cell proliferation is affected by hyperosmo-lality, and protein expression in leukocytes and mesothe-lial cells is strongly regulated by GDPs, indicate thatbiocompatibility is multifactorial. In addition, specific celltypes have different sensitivities to, and are affected indifferent ways, by PD solutions. Therefore, when extrap-olating observations on in vitro biocompatibility to whatmight be expected in vivo, it is essential to take intoaccount the effects on all cell populations. The physiologyof the peritoneal membrane is determined by the interac-tions of resident and transient cell populations. Therefore,when evaluating biocompatibility, one must look not onlyat the direct effects of exposure on all relevant cell pop-ulations, but at secondary effects as well—how the initial

exposure can set in motion signaling cascades from onecell population to another. As such, it will be the collec-tive and blended responses of all affected cell types thatultimately determine overall membrane health.

ADEQUACY OF STUDIES IN DEFININGBIOCOMPATIBILITY

The biocompatibility studies reviewed here provideinformation on solution cytotoxicity and insight intothe effects on inflammatory response, angiogenesis andcapillary permeability, fibrosis, and ECM deposition,cellular and molecular changes relevant to membraneperformance in vivo. These studies provide a fairly com-prehensive evaluation of solution biocompatibility inleukocyte and mesothelial cell populations; in contrast,there was a single study in L-929 fibroblasts and no stud-ies to date using cells of the peritoneal vasculature. Asthese cells are significant contributors to both normalmembrane physiology, as well as processes of fibrosis,neoangiogenesis, and vasculopathy, hallmarks of mem-brane deterioration and ultrafiltration failure, inclusionof these cell types in future biocompatibility studies isstrongly recommended.

Also recommended are studies based on the use ofcell culture supernatants or preconditioned media. Thesestudies would be of considerable value in assessingsolution biocompatibility in endothelial and fibroblastpopulations. As residents of the interstitial space,endothelial cells, vascular smooth muscle cells, and fi-broblasts are not directly exposed to solutions in vivo,although they are exposed to glucose and GDPs trans-ported from the peritoneal lumen. These cells dorespond to molecular signals from peritoneal leukocytesand mesothelial cells, however, and therefore would beaffected in large part through these signaling pathways.Therefore, culture of these cells with supernatants fromleukocytes and mesothelial cells exposed to PDF wouldbe a valid means to assess solution biocompatibility.These studies would also provide needed information onthe cell-cell interactions and offer sights into the indirect(but no less significant) effects of PD solutions.

CONTRIBUTIONS TO THE NEXTLEVELS OF STUDIES

In vitro biocompatibility studies should provide a ba-sis for moving forward with the next level(s) of testing.Improvements in cellular functional parameters demon-strated in cell culture studies (Table 3) can now beinvestigated in animal models of acute and chronic ex-posure to PDF. A major objective of the animal studies isto determine whether the conclusions drawn from the invitro biocompatibility studies can be confirmed in vivo.One of the limitations of in vitro studies is the difficulty in


evaluating processes that occur over long periods of time,such as AGE formation and its subsequent effects. Thereduced levels of GDPs in Physioneal, and the decreasedrate and level of AGE formation in vitro [72], imply a re-duction in AGE formation in vivo. Confirmation of thisand subsequent changes in membrane functional param-eters must now come from animal studies and clinicalbiopsies.

ARE IN VITRO STUDIES PREDICTIVE OF INVIVO BIOCOMPATIBILITY?

One of the main questions regarding in vitro studiesis whether they are predictive of, or accurately reflectin vivo biocompatibility. One would like to assume thatbiocompatibility could be predicted with some degree ofaccuracy on the basis of preclinical biocompatibility test-ing. In fact, preliminary results in PD patients indicatethat Physioneal is more biocompatible based on effluentmarkers indicating improved membrane health and pre-served mesothelial cell function [116, 117]. A definitiveanswer will come from prospective clinical studies on pa-tients using Physioneal for a long enough period so thatany solution effects are reflected not only in structuralcharacteristics of the membrane (e.g., AGE deposition,vasculopathy) but functional parameters as well.

When evaluating solution biocompatibility, it must beremembered that solution components are not the onlyfactors influencing membrane function and longevity asa dialyzing organ. Membrane structure and function inrenal disease patients are also affected by the uremic envi-ronment. It has been proposed that the peritoneal cavityof PD patients is in a state of severe overload of carbonylstress compounds derived not only from glucose-baseddialysis solutions, but from the uremic circulation as well[106, 118]. Components of the submesothelial space andthe interstitium—ECM components, fibroblasts, residentmacrophages, and cells of the peritoneal vasculature—are exposed not only to glucose and GDPs moving fromthe peritoneal cavity into the submesothelial space andinterstitium, but to RCOs and glucose moving from theuremic circulation into the peritoneal lumen. This bidi-rectional transport places these cellular components atincreased risk for AGE modification. AGEs and ALEshave been localized to the peritoneal membrane of nor-moglycemic uremic patients prior to the initiation of PD,an observation that emphasizes the contributions of theuremic circulation and oxidative stress, per se, to AGEformation and membrane damage [105], and indicatesthat this process is at least partially independent of PD.

The health of the membrane and its longevity as adialyzing tissue is influenced both by the uremic envi-ronment and chronic exposure to PDF. This necessitatesa dual approach to improving PD as a viable, long-termrenal replacement therapy. We must not only continue

to investigate innovative strategies to improve the bio-compatibility of dialysis solutions, but at the same timetreat the underlying kidney disease and uremic condi-tion, as well as comorbid conditions such as diabetes andhypertension.

CONCLUSION

In vitro biocompatibility testing has clearly demon-strated that neutral pH, bicarbonate/lactate-bufferedPhysioneal solutions are superior to conventional solu-tions in preserving cell viability and function. They haveproven valuable in confirming the basic safety of thesolution, identified predictive markers of cell function,and provided a basis for moving forward with the nextstages in preclinical testing. The goal for all in vitro bio-compatibility studies should be the use of an experimentaldesign that involves all relevant cell populations, mim-ics cell-to-cell interactions and molecular signaling path-ways, and recreates the in vivo environment as closely aspossible, so that in vivo solution biocompatibility can bepredicted with the highest degree of confidence. Whilethese studies cannot take the place of long-term clinicalstudies, and cannot fully address the comorbid conditionsthat influence technique survival, they do play an integralrole in PD solution development.

ACKNOWLEDGMENT

The author wishes to thank Jessica Svatek for her crit-ical reading of the manuscript.

Reprint requests to Catherine M. Hoff, Ph.D., Renal Division, BaxterHealthcare Corporation, 1620 Waukegan Road, MPGR-A2N, McGawPark, IL 60085-6730.E-mail: Cathy [email protected]

REFERENCES

1. CHUNG SH, STENVINKEL P, BERGSTROM J, LINDHOLM B: Biocompat-ibility of new peritoneal dialysis solutions: What can we hope toachieve? Perit Dial Int 20(Suppl 5):S57–S67, 2000

2. GOKAL R: Newer peritoneal dialysis solutions. Adv Ren ReplaceTher 7:302–309, 2000

3. HOLMES CJ, SHOCKLEY TR: Strategies to reduce glucose exposurein peritoneal dialysis patients. Perit Dial Int 20 (Suppl 2):S37–S41,2000

4. DUWE AK, VAS SI, WEATHERHEAD JW: Effects of the compositionof peritoneal dialysis solution on chemiluminescence, phagocyto-sis, and bactericidal activity in vitro. Infect Immunol 33:130–135,1981

5. FISCHER F-P, SCHENK U, KIEFER T, et al: In vitro effects ofbicarbonate- versus lactate-buffered continuous ambulatory peri-toneal dialysis fluids on peritoneal macrophage function. Am JKidney Dis 26:924–933, 1995

6. LIBEREK T, TOPLEY N, JORRES A, et al: Peritoneal dialysis fluid inhi-bition of polymorphonuclear leukocyte respiratory burst activityis related to the lowering of intracellular pH. Nephron 65:260–265,1993

7. TOPLEY N, ALOBAIDI HM, DAVIES M, et al: The effect of dialysate onperitoneal phagocyte oxidative metabolism. Kidney Int 34(3):404–11, 1988


8. YU AW, ZHOU XJ, ZHOU FQ, et al: Neutrophilic intracellular acido-sis induced by conventional, lactate-containing peritoneal dialysissolutions. Int J Artif Organs 15:661–665, 1992

9. JORRES A, JORRES D, GAHL GM, et al: Leukotriene B4 and tu-mor necrosis factor from leukocytes: Effect of peritoneal dialysate.Nephron 58:276–282, 1991

10. DOUVDEVANI A, RAPOPORT J, KONFORTI A, et al: The effect ofperitoneal dialysis fluid on the release of IL-1b and TNFa bymacrophages/monocytes. Perit Dial Int 13:112–117, 1993

11. DOUVDEVANI A, RAPOPORT J, KONFORTY A, et al: Intracellular acid-ification mediates the inhibitory effect of peritoneal dialysate onperitoneal macrophages. J Am Soc Nephrol 6:207–213, 1995

12. ING TS, YU AW, ZHOU FQ, et al: Failure of neutrophils to recovertheir ability to produce superoxide after stunning by a conven-tional, acidic, lactate-based peritoneal dialysis solution. Int J ArtifOrgans 17:191–194, 1994

13. ALOBAIDI HM, COLES GA, DAVIES M, et al: Host defense in con-tinuous ambulatory peritoneal dialysis: The effect of the dialysateon phagocyte function. Nephrol Dial Transplant 1:16–21, 1986

14. DOBOS GJ, BOHLER J, ZHOU XJ, et al: Persistent inhibition ofneutrophil function by glucose based dialysis solutions. ASAIOJ 40:M435–M439, 1994

15. DOUVDEVANI A, ABRAMSON O, TAMIR A, et al: Commercialdialysate inhibits TNFa mRNA expression and NF-jB DNA-binding activity in LPS-stimulated macrophages. Kidney Int47:1537–1545, 1995

16. WITOWSKI J, TOPLEY N, JORRES A, et al: Effect of lactate-bufferedperitoneal dialysis fluids on human peritoneal mesothelial cellinterleukin-6 and prostaglandin synthesis. Kidney Int 46:282–293,1994

17. SHAO JC, YORIOKA N, NISHIDA Y, YAMAKIDO M: Effect of pH andglucose on cultured human mesothelial cells. Scand J Urol Nephrol33:248–256, 1999

18. BRUNKHORST R, MAHIOUT A: Pyruvate neutralizes peritonealdialysate cytotoxicity: maintained integrity and proliferation ofcultured human mesothelial cells. Kidney Int 48:177–181, 1995

19. MAHIOUT A, BRUNKHORST R: Pyruvate anions neutralize peritonealdialysate toxicity. Nephrol Dial Transplant 10:391–394, 1995

20. WONG TYH, PHILLIPS AO, WITOWSKI J, TOPLEY N: Glucose-mediated induction of TGF-b1 and MCP-1 in mesothelial cellsin vitro is osmolality and polyol pathway dependent. Kidney Int63:1404–1416, 2003

21. HA H, LEE HB: Effect of high glucose on peritoneal mesothelialcell biology. Perit Dial Int 20(Suppl 2):S15–S18, 2000

22. KANG DH, HONG YS, LIM HJ, et al: High glucose solution and spentdialysate stimulate the synthesis of transforming growth factor-b1of human peritoneal mesothelial cells: Effect of cytokine stimula-tion. Perit Dial Int 19:221–230, 1999

23. SHOSTAK A, WAJSBROT V, GOTLOIB L: High glucose accelerates thelife cycle of the in vivo exposed mesothelium. Kidney Int 58:2044–2052, 2000

24. KUMANO K, SHIMODA M, HYODO T, SAKAI T: The role of TGF-bin growth inhibition of peritoneal mesothelial cells in high-glucosedialysate. Perit Dial Int 15 (Suppl 7):S93–S95, 1995

25. KUMANO K, SCHILLER B, HJELLE JT, MORAN J: Effects of osmotic so-lutes on fibronectin mRNA expression in rat peritoneal mesothe-lial cells. Blood Purif 14:165–169, 1996

26. MEDCALF JF, WALLS J, PAWLUCZYK IZA, HARRIS KPG: Effects ofglucose dialysate on extracellular matrix production by humanperitoneal mesothelial cells (HPMC): The role of TGF-b . NephrolDial Transplant 16:1885–1892, 2001

27. WANG T, CHEN YG, YE RG, et al: Effect of glucose on TGF-bexpression in peritoneal mesothelial cells. Adv Perit Dial 11:7–10,1995

28. OGATA S, YORIOKA N: Effect of prednisolone on increased expres-sion of laminin by human peritoneal mesothelial cells cultured withhigh glucose. Adv Perit Dial 17:2–4, 2001

29. LEE SK, KIM BS, YANG WS, et al: High glucose induces MCP-1expression partly via tyrosine kinase-AP-1 pathway in peritonealmesothelial cells. Kidney Int 60:55–64, 2001

30. ISHIBASHI Y, SUGIMOTO T, ICHIKAWA Y, et al: Glucose dialysate in-duces mitochondrial DNA damage in peritoneal mesothelial cells.Perit Dial Int 22:11–21, 2002

31. YANG AH, CHEN JY, LIN YP, et al: Peritoneal dialysis solutioninduces apoptosis of mesothelial cells. Kidney Int 51:1280–1288,1997

32. XU Z-G, KIM KS, PARK HC, et al: High glucose activates the p38MAPK pathway in cultures human peritoneal mesothelial cells.Kidney Int 63:958–968, 2003

33. TAMURA M, OSAJIMA A, NAKAYAMADA S, et al: High glucose lev-els inhibit focal adhesion kinase-mediated wound healing of ratperitoneal mesothelial cells. Kidney Int 63:722–731, 2003

34. ITO T, YORIOKA N, YAMAMOTO M, et al: Effect of glucose on inter-cellular junctions of cultured human peritoneal mesothelial cells.J Am Soc Nephrol 11:1969–1979, 2000

35. SITTER T, HASLINGER B, MANDL S, et al: High glucose increaseprostaglandin E2 synthesis in human peritoneal mesothelial cells:Role of hyperosmolarity. J Am Soc Nephrol 9:2005–2012, 1998

36. SITTER T, MANDL-WEBER S, WORNLE M, et al: D-glucose increasesthe synthesis of tissue-type plasminogen activator (t-PA) in humanperitoneal mesothelial cells. Thromb Haemost 82:1171–1176, 1999

37. LAI KN, LI FK, LAN HY, et al: Expression of aquaporin-1 in humanperitoneal mesothelial cells and its upregulation by glucose in vitro.J Am Soc Nephrol 12:1036–1045, 2001

38. KUBOSHIMA S, OGIMOTO G, SAKURADA T, et al: Hyperosmotic stim-uli induces recruitment of aquaporin-1 to plasma membrane incultured rat mesothelial cells. Adv Perit Dial 17:47–52, 2001

39. JORRES A, GAHL GM, TOPLEY N, et al: In-vitro biocompatibility ofalternative CAPD fluids: Comparison of bicarbonate-buffered andglucose-polymer-based solutions. Nephrol Dial Transplant 9:785–790, 1994

40. DE FIJTER CWH, VERBRUGH HA, OE LP, et al: Biocompatibility of aglucose polymer-containing peritoneal dialysis fluid. Am J KidneyDis 4:411–418, 1993

41. THOMAS E, AL-BAKER E, DROPCOVA S, et al: Different kinetics ofsenescence in human fibroblasts and peritoneal mesothelial cells.Exp Cell Res 236:355–358, 1997

42. CENDOROGLO M, SUNDARAM S, JABER BL, PEREIRA BJG: Effect ofglucose concentration, osmolality, and sterilization process of peri-toneal dialysis fluids on cytokine production by peripheral bloodmononuclear cells and polymorphonuclear cell functions in vitro.Am J Kidney Dis 31:273–282, 1998

43. HIGUCHI C, SANAKA T, SATO T, et al: The effect of glucose on theproliferation of peritoneal fibroblasts. Adv Perit Dial 13:253–256,1997

44. HIGUCHI C, NIHEI H: The role of protein kinase C activity in the pro-liferation of peritoneal fibroblasts. Perit Dial Int 19(Suppl 2):S353–S357, 1999

45. BREBOROWICZ A, WISNIEWSKA J, POLUBINSKA A, et al: Role of peri-toneal mesothelial cells and fibroblasts in the synthesis of hyaluro-nan during peritoneal dilaysis. Perit Dial Int 18:382–386, 1998

46. SEO MJ, OH SJ, KIM SI, et al: High glucose dialysis solutions in-crease synthesis of vascular endothelial growth factors by peri-toneal vascular endothelial cells. Perit Dial Int 21(Suppl 3):S35–S40, 2001

47. MIYATA T, DEVUYST O, KUROKAWA K, VAN YPERSELE DE STRIHOU C:Toward better dialysis biocompatibility: Advances in the biochem-istry and pathophysiology of the peritoneal membranes. Kidney Int61:375–386, 2002

48. LINDEN T, COHEN A, DEPPISCH R, et al: 3,4-dideoxyglucosone-3-ene(3,4-DGE): A cytotoxic glucose degradation product in fluids forperitoneal dialysis. Kidney Int 62:697–703, 2002

49. WIESLANDER AP, NORDIN MK, KJELLSTRAND PTT, BOBERG UC:Toxicity of peritoneal dialysis fluids on cultured fibroblasts, L-929.Kidney Int 40:77–79, 1991

50. KJELLSTRAND P, MARTINSON E, WIESLANDER A, HOLMQUIST B: De-velopment of toxic degradation products during heat sterilizationof glucose-containing fluids for peritoneal dialysis: Influence oftime and temperature. Perit Dial Int 15:26–32, 1995

51. WIESLANDER AP, ANDREN AH, NILSSON-THORELL C, et al: Are alde-hydes in heat-sterilized peritoneal dialysis fluids toxic in vitro? PeritDial Int 15:348–352, 1995

52. WIESLANDER AP, NORDIN MK, MARTINSON E, et al: Heat ster-ilised PD-fluids impair growth and inflammatory responses of cul-tured cell lines and human leukocytes. Clin Nephrol 39:343–348,1993


53. CATALAN MP, REYERO A, EGIDO J, ORTIZ A: Acceleration of neu-trophil apoptosis by glucose-containing peritoneal dialysis solu-tions: Role of caspases. J Am Soc Nephrol 12:2442–2449, 2001

54. WITOWSKI J, WISNIEWSKA J, KORYBALSKA K, et al: Prolonged expo-sure to glucose degradation products impairs viability and functionof human peritoneal mesothelial cells. J Am Soc Nephrol 12:2434–2441, 2001

55. WITOWSKI J, KORYBALSKA K, WISNIEWSKA J, et al: Effect of glucosedegradation products on human peritoneal mesothelial cell func-tion. J Am Soc Nephrol 11:729–739, 2000

56. WELTEN AGA, SCHALKWIJK CG, TER WEE PM, BEELEN RJH: Sin-gle exposure of mesothelial cells to glucose degradation products(GDPs) yields early advanced glycation end-products (AGEs) anda proinflammatory response. Perit Dial Int 23:213–221, 2003

57. ARBEITER K, BIDMON B, ENDEMANN M, et al: Peritoneal dialysis fluidcomposition determines heat shock protein expression patterns inhuman mesothelial cells. Kidney Int 60:1930–1937, 2001

58. LU M, KUROKI M, AMANO S, et al: Advanced glycation end prod-ucts increase retinal vascular endothelial growth factor expression.J Clin Invest 101:1219–1224, 1998

59. SATOH H, TOGO M, HARA M, et al: Advanced glycation end prod-ucts stimulate mitogen-activated protein kinase and proliferationin rabbit vascular smooth muscle cells. Biochem Biophys Res Com-mun 239:111–115, 1997

60. COMBET S, MIYATA T, MOULIN P, et al: Vascular proliferation andenhanced expression of endothelial nitric oxide synthase in humanperitoneum exposed to long-term peritoneal dialysis. J Am SocNephrol 11:717–728, 2000

61. DI PAOLO N, SACCHI G: Peritoneal vascular changes in continuousambulatory peritoneal dialysis (CAPD): An in vivo model for thestudy of diabetic microangiopathy. Perit Dial Int 9(1):41–5, 1989

62. DOBBIE JW, LLOYD JK, GALL CA: Categorization of ultrastructuralchanges on peritoneal mesothelium, stroma and blood vessels inuremia and CAPD patients. Adv Perit Dial 6:3–12, 1990

63. HONDA K, NITTA K, HORITA S, et al: Morphological changes in theperitoneal vasculature of patients on CAPD with ultrafiltrationfailure. Nephron 72:171–176, 1996

64. NAKAYAMA M, KAWAGUCHI Y, YAMADA K, et al: Immunohisto-chemical detection of advanced glycation end-products in the peri-toneum and its possible pathophysiological role in CAPD. KidneyInt 51:182–186, 1997

65. NAKAMURA S, TACHIKAWA T, TOBITA K, et al: Role of advanced gly-cation end products and growth factors in peritoneal dysfunctionin CAPD patients. Am J Kidney Dis 41(Suppl 1):S61–S67, 2003

66. HONDA K, NITTA K, HORITA S, et al: Accumulation of advancedglycation end products in the peritoneal vasculature of continuousambulatory peritoneal dialysis patients with low ultra-filtration.Nephrol Dial Transplant 14:1541–1549, 1999

67. YAMAGISHI S, YONEKURA H, YAMAMOTO Y, et al: Advanced gly-cation end products-driven angiogenesis in vitro. J Biol Chem272:8723–8730, 1997

68. RASHID G, LUZON A-A, KORZETS Z, et al: The effect of advancedglycation end-products and aminoguanidine on TNFa productionby rat peritoneal macrophages. Perit Dial Int 21:122–129, 2001

69. BOULANGER E, WAUTIER M-P, WAUTIER J-L, et al: AGEs bind tomesothelial cells via RAGE and stimulate VCAM-1 expression.Kidney Int 61:148–156, 2002

70. NISHIDA Y, SHAO JC, KIRIBAYASHI K, et al: Advanced glycationend-products reduce the viability of human peritoneal mesothelialcells. Nephron 80:477–478, 1998

71. MANDL-WEBER S, HASLINGER B, SCHALKWIJK CG, SITTER T: Earlyglycated albumin, but not advanced glycated albumin, methylgly-oxal, or 3-deoxyglucosone increases the expression of PAI-1 inhuman peritoneal mesothelial cells. Perit Dial Int 21:487–494, 2001

72. MILLAR DJ, HOLMES C, FAICT D, DAWNAY A: Comparison of invitro AGE formation between standard PD fluid and a novel bi-carbonate/lactate formulation. Perit Dial Int 14:191–194, 1998

73. HOLMES CJ: Pre-clinical biocompatibility testing of peritoneal dial-ysis solutions. Perit Dial Int 20(Suppl 5):S5–S9, 2000

74. HOLMES CJ: Biocompatibility of peritoneal dialysis solutions. PeritDial Int 13:88–94, 1993

75. SITTER T, TOET K, QUAX P, KOOISTRA T: Fibrinolytic activity ofhuman mesothelial cells is counteracted by rapid uptake of tissue-

type plasminogen activator. Kidney Int 55:120–129, 199976. HOLMDAHL L, FALKENBERG M, IVARSSON M, RISBERG B: Plasmino-

gen activators and inhibitors in peritoneal tissue. APMIS 105:25–30, 1997

77. SITTER T, TOET K, FRICKE H, et al: Modulation of procoagulant andfibrinolytic system components of mesothelial cells by inflamma-tory mediators. Am J Physiol 271:R1256–R1263, 1996

78. VAN HINSBERGH VWM, KOOISTRA T, SCHEFFER MA, et al: Char-acterization and fibrinolytic properties of human omental tis-sue mesothelial cells. Comparison with endothelial cells. Blood75:1490–1497, 1990

79. MARTIN J, YUNG S, ROBSON RL, et al: Production and regulationof matrix metalloproteinases and their inhibitors by human peri-toneal mesothelial cells. Perit Dial Int 20:526–533, 2000

80. STOENOIU MS, DEVRIESE AS, CNOPS Y, et al: Diabetes induces per-meability changes and differential nitric oxide synthase (NOS)isoforms expression in the rat peritoneum. J Am Soc Nephrol11:653A, 2000

81. OFFNER FA, FEICHTINGER H, STADLMANN S, et al: Transforminggrowth factor-b synthesis by human peritoneal mesothelial cells.Am J Pathol 148:1679–1688, 1996

82. YANG WS, KIM BS, LEE SK, et al: Interleukin-1b stimulates theproduction of extracellular matrix in cultured human peritonealmesothelial cells. Perit Dial Int 19:211–220, 1999

83. HA H, YU MR, PARK MS, LEE HB: High glucose-induced TGF-b and fibronectin synthesis by human peritoneal mesothelial cell:Involvement of PKC. Perit Dial Int 18:111, 1998

84. YUNG S, THOMAS GJ, STYLIANOU E, et al: Source of peritoneal pro-teoglycans. Human peritoneal mesothelial cells synthesize and se-crete mainly small dermatan sulfate proteoglycans. Am J Pathol146:520–529, 1995

85. YUNG S, COLES GA, DAVIES M: IL-1b , a major stimulator ofhyaluronan synthesis in vitro of human peritoneal mesothelialcells: relevance to peritonitis in CAPD. Kidney Int 50:1337–1343,1996

86. BASOK A, SHNAIDER A, MAN L, et al: CD40 is expressed on hu-man peritoneal mesothelial cells and upregulates the productionof interleukin-15. J Am Soc Nephrol 12:695–702, 2001

87. DOUVDEVANI A, RAPOPORT J, KONFORTY A, et al: Human peritonealmesothelial cells synthesize IL-1a and b . Kidney Int 46:993–1001,1994

88. TOPLEY N, JORRES A, LUTTMANN W, et al: Human peritonealmesothelial cells synthesize interleukin-6: Induction by IL-1b andTNF-a. Kidney Int 43:226–233, 1993

89. TOPLEY N, BROWN Z, JORRES A, et al: Human peritoneal mesothelialcells synthesize interleukin-8. Am J Pathol 142:1876–1886, 1993

90. LIBEREK T, TOPLEY N, LUTTMANN W, WILLIAMS JD: Adherence ofneutrophils to human peritoneal mesothelial cells: Role of inter-cellular adhesion molecule-1. J Am Soc Nephrol 7:208–217, 1995

91. CHEN J-Y, YANG A-H, LIN Y-P, et al: Absence of modulating ef-fects of cytokines on antioxidant enzymes in peritoneal mesothelialcells. Perit Dial Int 17:455–466, 1997

92. SCHROPPEL B, FISCHEREDER M, WIESE P, et al: Expression of glu-cose transporters in human peritoneal mesothelial cells. KidneyInt 53:1278–1287, 1998

93. MANDL-WEBER S, COHEN CD, HASLINGER B, et al: Vascular en-dothelial growth factor production and regulation in human peri-toneal mesothelial cells. Kidney Int 61:570–578, 2002

94. MARGETTS PJ, KOLB M, GALT T, et al: Gene transfer of transform-ing growth factor-b1 to the rat peritoneum: Effects on membranefunction. J Am Soc Nephrol 12:2029–2039, 2001

95. SELGAS R, DEL PESO G, BAJO M-A, et al: Spontaneous VEGF pro-duction by cultured peritoneal mesothelial cells from patients onperitoneal dialysis. Perit Dial Int 20:798–801, 2000

96. WITOWSKI J, BENDER TO, FINN A, et al: TGF-b stimulates VEGFproduction in human peritoneal mesothelial cells: Superinductionby proinflammatory cytokines. Perit Dial Int 21(Suppl 1):S21, 2001

97. JORRES A, LUDAT K, SANDER K, et al: The peritoneal fibroblast andthe control of peritoneal inflammation. Kidney Int 50:S-22–S-27,1996

98. WITOWSKI J, THIEL A, DECHEND R, et al: Synthesis of C-X-C andC-C chemokines by human peritoneal fibroblasts: induction bymacrophage-derived cytokines. Am J Pathol 158:1441–1450, 2001


99. JORRES A, LUDAT K, LANG J, et al: Establishment and functionalcharacterization of human peritoneal fibroblasts in culture: Regu-lation of interleukin-6 production by proinflammatory cytokines.J Am Soc Nephrol 7:2192–2201, 1996

100. KLEIN ES, ASCULAI SS, BEN-ARI GY: Effects of hyaluronic acidon fibroblast behavior in peritoneal injury. J Surg Res 61:473–476,1996

101. BEAVIS MJ, WILLIAMS JD, HOPPE J, TOPLEY N: Human peritonealfibroblast proliferation in 3-dimensional culture: Modulation bycytokines, growth factors and peritoneal dilaysis effluent. KidneyInt 51:205–215, 1997

102. FERRIER M-L, COMBET S, VAN LANDSCHOOT M, et al: Inhibition ofnitric oxide synthase reverses changes in peritoneal permeabilityin a rat model of acute peritonitis. Kidney Int 60:2343–2350, 2001

103. DEVUYST O, COMBET S, CNOPS Y, STOENOIU MS: Regulation of NOsynthase isoforms in the peritoneum: Implications for ultrafiltra-tion failure in peritoneal dialysis. Nephrol Dial Transplant 16:675–678, 2001

104. WILLIAMS JD, CRAIG KJ, TOPLEY N, et al: Morphologic changes inthe peritoneal membrane of patients with renal disease. J Am SocNephrol 13:470–479, 2002

105. MIYATA T, HORIE K, UEDA Y, et al: Advanced glycation and lipidox-idation of the peritoneal membrane: Respective roles of serum andperitoneal fluid reactive carbonyl compounds. Kidney Int 58:425–435, 2000

106. UEDA Y, MIYATA T, GOFFIN E, et al: Effect of dwell time on carbonylstress using icodextrin and amino acid peritoneal dialysis solutions.Kidney Int 58:2518–2524, 2000

107. ZEIER M, SCHWENGER V, DEPPISCH R, et al: Glucose degradationproducts in PD fluids: Do they disappear from the peritoneal cavityand enter the systemic circulation? Kidney Int 63:298–305, 2003

108. JORRES A, BENDER TO, FINN A, et al: Biocompatibility and buffers:Effect of bicarbonate-buffered peritoneal dialysis solutions onperitoneal cell function. Kidney Int 54:2184–2193, 1998

109. ROGACHEV B, HAUSMANN MJ, YULZARI R, et al: Effect ofbicarbonate-based dialysis solutions on intracellular pH (pHi) andTNFa production of peritoneal macrophages. Perit Dial Int 17:546–553, 1997

110. TOPLEY N, KAUR D, PETERSEN MM, et al: In vitro effects of bicar-bonate and bicarbonate-lactate buffered peritoneal dialysis solu-

tions on mesothelial and neutrophil function. J Am Soc Nephrol7:218–224, 1996

111. SUNDARAM S, CENDOROGLO M, COOKER LA, et al: Effect of two-chambered bicarbonate lactate-buffered peritoneal dialysis fluidson peripheral blood mononuclear cell and polymorphonuclear cellfunction in vitro. Am J Kidney Dis 30:680–689, 1997

112. CENDOROGLO M, SUNDARAM S, GROVES C, et al: Necrosis and apop-tosis of polymorphonuclear cells exposed to peritoneal dialysisfluids in vitro. Kidney Int 52:1626–1634, 1997

113. HA H, YU MR, CHOI HN, et al: Effects of conventional and newperitoneal dialysis solutions on human peritoneal mesothelial cellviability and proliferation. Perit Dial Int 20(Suppl 5):S10–S19, 2000

114. HA H, CHA MK, CHOI HN, LEE HB: Effects of peritoneal dialysissolutions on the secretion of growth factors and extracellular ma-trix proteins by human peritoneal mesothelial cells. Perit Dial Int22:171–177, 2002

115. COOKER LA, LUNEBURG P, FAICT D, et al: Reduced glucose degra-dation products in bicarbonate/lactate-buffered peritoneal dialysissolutions produced in two-chambered bags. Perit Dial Int 17:373–378, 1997

116. RIPPE B, SIMONSEN O, HEIMBURGER O, et al: Long-term clinicaleffects of a peritoneal dialysis fluid with less glucose degradationproducts. Kidney Int 59:348–357, 2001

117. JONES S, HOLMES CJ, KREDIET RT, et al: Bicarbonate/lactate-basedperitoneal dialysis solution increases cancer antigen 125 and de-creases hyaluronic acid levels. Kidney Int 59:1529–1538, 2001

118. MIYATA T, IZUHARA Y, SAKAI H, KUROKAWA K: Carbonyl stress:increased carbonyl modification of tissue and cellular proteins inuremia. Perit Dial Int 19(Suppl 2):S58–S61, 1999

119. HJELLE JT, MILLER-HJELLE MA, DOBBIE JW: The biology of themesothelium during peritoneal dialysis. Perit Dial Int 15:S13–S23,1995

120. GRAFF J, JOFFE P, FUGLEBERG S, JENSEN LT: Collagen markers inperitoneal dialysis patients. Adv Perit Dial 11:24–27, 1995

121. BOULANGER E, WAUTIER MP, GANE P, et al: Bicarbonate/lactatebuffered peritoneal dialysis solutions increase proliferation andreduce apoptosis and necrosis of human peritoneal mesothelialcells. Perit Dial Int 23(Suppl 1):S8, 2003

122. COOKER LA, HOLMES CJ, HOFF CM: Biocompatibility of icodextrin.Kidney Int 62(Suppl 81), 2002

in vitro biocompatibility performance of physioneal

Documents