Biocompatibility of peritoneal dialysis solutions determined by genomics of human leucocytes: a cross-over study

Sir,

Peritoneal dialysis (PD) is based on passive movement of water and soluble molecules across the peritoneum. In continuous ambulatory peritoneal dialysis (CAPD), the patient's abdomen is filled with a dialysate fluid introducing an osmotic gradient driven by electrolytes and glucose, or macromolecules such as icodextrin. Biocompatibility of PD fluids is the most important criterion to enable long-term dialysis without introducing clinically significant changes in the functional characteristics of the peritoneum and systemic inflammatory effects [1]. The effects of biocompatibility on clinical outcome include changes in the physiology of cell populations constituting the peritoneal cavity (leucocyte, mesothelial and endothelial cells, and fibroblasts) and the gene expression of peripheral blood mononuclear cells (PBMCs) triggering alterations in cytokine, chemokine and growth factor networks, upregulation of proinflammatory and profibrotic pathways, and induction of carbonyl and oxidative stress [2–4].

Our study objective was to compare the genome-wide gene expression signature of PBMCs of PD patients using glucose-based (GBF) and icodextrin-based peritoneal fluids (IBF) to allow a direct comparison of biocompatibility relevant intracellular processes with respect to the PD fluid used. This pilot study should give us first insights into the alterations in gene expression of leucocytes triggered by different PD fluids and should provide an informative basis for future research.

Therefore, we conducted a random cross-over study in five stable ESRD patients being treated with CAPD between 4 and 18 months (demographic data are provided on our laboratory homepage in Table 1 (http://www.meduniwien.ac.at/nephrogene/data/pd/)). Blood samples (10 ml) were collected immediately after a 4- to 6-h dwell of GBF (Physioneal® 40, Glucose 2.27% w/v, 395 mOsmol/l) and an overnight dwell of IBF (Extraneal®, icodextrin 7.5%, 284 mOsmol/l) [study approved by the local Institutional review board (Ethical Committee # EK-318/06, see http://ohrp.cit.nih.gov/search/asearch.asp)]. Oligoarrays were obtained from the Stanford University Functional Genomics core facility. All microarray experiment protocols can be found on the Stanford University webpage at http://cmgm.stanford.edu/pbrown/protocols/index.html. Stratagene Universal human reference RNA was used as a reference. Raw data files as well as the MIAME checklist are available at our laboratory webpage.

Table 1.

Biological processes separating IBF- and GBF-treated patient groups as derived on the level of PBMC differential gene expression

Biological process	Gene symbols	P-value
DEGs up-regulated by IBF treatment
Immunity and defence	CIITA, UNQ3033, SCGB1C1, CLEC1B, CTSW, CLEC4E, TNFRSF7, CLEC10A	<0.001
Natural killer cell-mediated immunity	UNQ3033, CLEC1B, CTSW	<0.001
T-cell-mediated immunity	CIITA, CTSW, TNFRSF7	0.001
Cell communication	UNQ3033, SCGB1C1, CLEC1B, STAT4, CLEC10A	0.008
Other neuronal activity	SP110, RASGRP2	0.009
Macrophage-mediated immunity	CLEC4E, CLEC10A	0.010
Ligand-mediated signalling	STAT4, UNQ3033, SCGB1C1	0.010
Other immune and defence	SCGB1C1, CLEC4E	0.012
Glucose haemeostasis	STAT4	0.021
Signal transduction	LST1, STAT4, UNQ3033, SCGB1C1, RASGRP2, CLEC1B, TNFRSF7, CLEC10A	0.022
MHC I-mediated immunity	CTSW	0.023
Cytokine- and chemokine- mediated signalling pathways	STAT4, TNFRSF7	0.029
MHC II-mediated immunity	CIITA	0.036
Glycolysis	HK3	0.048
DEGs up-regulated by GBF treatment
Ectoderm development	CELSR2, FOXA2, HLF, KRT80, TNFRSF21, COBLL1, NTN4, CRABP1, NLGN2, FGFR3, THSD3	<0.001
Signal transduction	FRAS1, DOC1, CELSR2, MGP, RND3,CGA, GNG4, RAB23, FOXA2, AXL, CAP2, CDH13, INPP5F, TACSTD2, TNFRST21, MFAP4, DIRAS1, CRABP1, NLGN2, SFRP2, THSD3, GPR161, FGFR3, NTN4	<0.001
Neurogenesis	CELSR2, FOXA2, HLF, TNFRSF21, COBLL1, NTN4, NLGN2, FGFR3, THSD3	<0.001
Cell communication	FRAS1, CELSR2, MGP, CGA, FOXA2, CAP2, CDH13, MFAP4, NTN4, CRABP1, NLGN2, SFRP2, THSD3	<0.001
Oncogenesis	DOC1, AXL, CDH13, MAGEA12, NTN4, MLF1, FGFR3, THSD3	<0.001
Developmental processes	DOC1, CELSR2, MGP, FOXA2, HLF, KRT80, TTK, MAGEA12, EFHD1, TNFRSF21, COBLL1, NTN4, CRABP1, NLGN2, FGFR3, THSD3	0.001
Other oncogenesis	MAGEA12, FGFR3, THSD3	0.002
Cell proliferation and differentiation	DOC1, FOXA2, AXL, TACSTD2, C9orf58, UHRF1, NTN4, MLF1, GINS2, FGFR3	0.002
Cell structure	DLG5, CELSR2, COL7A1, FOXA2, KRT80, PHLDB1, TJP1	0.006
Cell structure and motility	DLG5, CELSR2, COL7A1, FOXA2, KRT80, PHLDB1, TJP1, RND3, CAP2	0.011
DNA replication	DOC1, CDC2, GINS2	0.014
Homeostasis	CGA, HEPH, FSTL1	0.025
Stress response	MOCOS, C9orf58, GPX3	0.026
Other cell cycle process	UHRF1	0.028
DNA metabolism	DOC1, CDC2, DNTT, GINS2	0.028
Other receptor-mediated signalling pathway	FOXA2, TACSTD2, TNFRSF21	0.030
Proteolysis	DOC1, DGC, C1R, MMP15, CAP2, SERPINA5, TIMP3	0.033
Cell surface receptor-mediated signal transduction	CELSR2, RND3, GNG4, FOXA2, AXL, TACSTD2, TNFRSF21, FGFR3, THSD3, GPR161	0.035
Other steroid metabolism	SC5DL	0.041
Cell cycle	DOC1, CDC2, FOXA2, TTK, C9orf58, UHRF1, GINS2	0.042
Sex determination	TTK	0.044
Cell cycle control	DOC1, CDC2, FOXA2, C9orf58	0.045
Neurotransmitter release	STXBP1, EHD2	0.046
Cell adhesion	CELSR2, COL7A1, CDH13, MFAP4, NLGN2	0.049

Categories are ranked by the P-value (comparison of expected number of genes and observed number of genes in each biological process) indicating the relevance of a particular process.

A paired t-test (P < 0.05) of log-transformed expression values was used to evaluate differences between IBF and GBF treatment. Differentially expressed genes (DEGs) were hierarchically clustered and graphically represented using the MultiExperiment Viewer (MeV) (Pearson correlation, complete linkage) [5]. DEGs were furthermore analysed with respect to their molecular functions, biological processes and interaction partner using gene ontology terms (GO-Terms), PANTHER (Protein ANalysis THrough Evolutionary Relationships) ontologies and Online Predicted Human Interaction Database (OPHID).

A total of 124 genes (fold change over two, 34 up-regulated and 90 down-regulated in the IBF group) were identified as being significantly differentially expressed in PBMCs comparing patients under IBF and GBF usage (Figure 1 online on our homepage).

A total of 27 up-regulated genes assigned to IBF treatment and 81 up-regulated genes associated with GBF treatment could be classified according to PANTHER ontologies (Table 1). A number of the genes up-regulated in the course of IBF usage were found to be involved in immune response and inflammatory processes. Genes up-regulated by GBF usage in contrast are found to be assigned to development and signal transduction processes.

Our study provides full genome differential gene expression profiles of PBMCs after peritoneal dialysis on a genome-wide scale comparing GBF and IBF peritoneal dialysis fluids confirming the differential involvement of inflammation. A limitation of our study is the small sample size of five CAPD patients. Therefore, we used a random cross-over design and computed a paired t-test.

These pilot data suggest reduced inflammation and consequently an improved biocompatibility of GBF peritoneal fluids compared with IBF fluids. Certainly, further evaluation in larger studies is needed.