Peritoneal Dialysis Research in the UK: The Cardiff Contribution

Membrane Biology: Setting the Scene

While others were investigating membrane function, the focus in Cardiff was to improve our understanding of the biology of the dialyzed peritoneal membrane. Crucial to this was the development of new laboratory methods, most notably the successful isolation and culture of mesothelial cell monolayers (4). These methods were later developed into in vitro systems that enabled simulation of the membrane in vivo (5) and events related to infection (6). Key early observations were the demonstration that mesothelial cells were capable of producing many inflammatory cytokines [e.g., interleukin (IL)-6 and IL-8] and their synergistic induction by other co-stimulatory molecules (e.g., IL-1β and tumor necrosis factor-alpha) (7,8). From these early studies it was possible to demonstrate that all these cell types were capable of active participation in a cytokine network that has since been shown to be involved in determining membrane function and regulating the response to infection and, potentially, mechanisms of membrane injury, such as progressive fibrosis (9).

Biocompatibility: Description and Definition

In parallel with these studies, the Cardiff group was able to demonstrate the adverse effects of dialysis fluid on the normal biological functioning of constituents of the membrane, including its host defenses. They found that the combination of low pH with lactate was especially toxic to the function of neutrophils and macrophages by further reducing intracellular pH (10) and that exposure to dialysate collected throughout the PD dwell cycle impaired cytokine production by phagocytes in vitro (11,12). As a consequence of these studies, they developed ways of assessing this “bioincompatibility,” both in vitro (i.e., within the test tube) and ex vivo, taking the crucial components from the dialyzed membrane (e.g., macrophages), where they had been exposed in vivo to various dialysis solutions, and studying them in vitro. What emerged was a multidimensional picture of toxicity, with low pH, lactate buffer, high osmolality, and high glucose concentration affecting intracellular metabolism via the polyol pathway (13) and later the damaging effects of glucose degradation products on, for example, wound healing (14), all causing overlapping adverse effects. These observations led to the conceptualization of biocompatibility as a local intraperitoneal process distinct from the systemic effects of PD (e.g., lipid abnormalities, obesity) and became the gold standard approach to assessing the biocompatibility of newly developed PD solutions (15-18).

Membrane Morphology: The Peritoneal Biopsy Registry

The logical next step was to understand in what way these adverse biological processes might be affecting peritoneal membrane morphology; hence the inception of the biopsy registry. Again, new methodology was called for, including the careful harvesting of material to prevent tissue damage and the terminology that would be needed to describe and quantify abnormal findings. In this respect, the “other” Williams in the team, Geraint, a gastrointestinal histopathologist, played a crucial role. This methodology included consideration of how numbers of vessels per area or length of examined tissue should be quantified and the categorization of membrane into mesothelium, submesothelial compact zone, and underlying fatty tissue. The important findings of this study, which include the unique diabetiform occlusive vasculopathy that involves both arterioles and venules combined with time-related membrane thickening, resulted in Cardiff's most cited publication (19). They also noted that the membrane in uremic controls is already abnormally thickened and that the vascular and membrane thickening abnormalities were both associated with each other and clearly related to time on therapy, despite the cross-sectional nature of this study (20). Undoubtedly, these observations lend significant weight to the whole biocompatibility argument and the etiological role of glucose (and glucose degradation products), in particular to long-term membrane injury.

Biocompatibility: Solutions to a Problem?

Having followed the bench-to-bedside argument through thus far, the inevitable consequence was to develop dialysis solutions with improved biocompatibility. Again, Cardiff played a major role in in vitro, ex vivo, and subsequently in vivo evaluation of these solutions (21-23). There can be little doubt that these solutions have a favorable biocompatibility profile with improved cellular function, as indicated by improved phagocytosis and mesothelial cell response to wound injury (24, 25). It has also been a universal finding of all available biocompatible solutions that randomized controlled trials have shown evidence of preserved mesothelial cell mass, as judged from increased dialysate effluent concentrations of cancer antigen 125 (22,23). The effects on other potential mediators of membrane inflammation and healing have been less clear, in part because we still do not really know how to interpret them. Does an increase in a particular cytokine (e.g., IL-6, given that mesothelial cells are a potentially potent source and conventional dialysate suppresses in vitro production) mean that there is more inflammation in the membrane or there is a healthier mesothelial cell population? Demonstrating a clinically relevant benefit (lower peritonitis rate, reduced membrane failure, less encapsulating peritoneal sclerosis) has proved much more difficult, however, not least because of the required size and the length of follow-up of a clinical trial powered for these end points. The exception to this has been the undoubted benefit of normal pH solutions on in-flow pain.

Immunobiology of Peritonitis: A Developing Story

Another research thread that has run alongside the biocompatibility story in Cardiff is the peritoneal membrane's immunobiological response to infection. Having established that the dialyzed membrane is, in effect, an immune-compromised site driven by the bio-incompatibility observations described above, the initial focus became documentation of the early events associated with an infectious episode (26). The Cardiff group was able to show that bacterial infection is associated with a massive increase in cytokine production (e.g., IL-6) by mesothelial cells, a process that was significantly enhanced or “superinduced” by the presence of dialysis effluent and IL-1β (27). They also demonstrated the mechanisms driving neutrophil recruitment into the membrane, showing its dependence on the chemokine gradient and expression of intercellular adhesion molecule-1, again indicating a central role for the mesothelial cell in this process (6). Lately, the focus of attention has been more on the signaling mechanisms, in particular the role of IL-6 (28,29), for control and resolution of infectious episodes, pursuing the hypothesis that repeated episodes of infection might result in alteration in the regulation of this process, thus exposing the membrane to the risk of progressive fibrotic injury. Their recent demonstration that microbe-responsive γδ T cells within the dialyzed peritoneum, which are able to induce monocyte differentiation, appear to play an important role in this context, being specifically activated by infecting organisms associated with a worse outcome in patients (HMB-PP positive; e.g., most gram-negative infections) (30). Most recently, the Cardiff group has identified a population of functional effector memory T cells that are likely to modulate the host's response to infection and, again, they hypothesize that an alteration in the membrane's resident T cell populations might determine whether progressive fibrosis might subsequently ensue (31,32). The pieces of this jigsaw puzzle are by no means complete but there is plenty more to come.

Biomarkers and Biometrics: The Global Fluid Study

Given the difficulties of conducting randomized trials to evaluate the effects of the newer biocompatible solutions, Nick Topley proposed about a decade ago that there should be a longitudinal cohort study linking solution use to dialysate effluent biomarkers, clinical measures (including membrane function), and important clinical end points such as peritonitis, membrane injury, and patient and technique survival. The result was the GLOBAL Fluid Study, which has now recruited well over 1000 patients from around the world (e.g., UK, Canada, Korea) with substantial follow up and paired plasma, dialysate, and clinical observations. Since its inception, two things have changed: (1) increasing awareness and concern related to the problem of encapsulating peritoneal sclerosis, and thus the need for biomarkers/metrics to reassure the clinician that continuing PD is safe; and (2) dramatic developments in the science of biomarkers, with emphasis on patterns of change rather than single cytokines to give us the answers. For these reasons we (I am clinical lead on this project) have deliberately held back from premature analysis, ensuring that the most up-to-date approaches to sample and data analyses are used. Preliminary work does indeed suggest that the peritoneal cavity varies in its inflammatory status and that this is largely independent of systemic inflammation, observations strengthened by the multicenter nature of the study in which center effects are also apparent. There is much work to be done but it is already clear that the GLOBAL Fluid Study represents a very powerful resource for the future in helping to answer these important questions.

In conclusion, the Cardiff group has made a considerable contribution to the science and development of PD as a therapy. It would not be fair to finish this summary without mentioning a more general but nevertheless important contribution they have made: they raised the bar of high quality research in PD, especially when it comes to the contribution of basic science. The Cardiff group includes founding members of EuroPD, a meeting that remains a very successful now biannual conference (come to Birmingham, UK, in October 2011!); they also instigated regular meetings to bring scientific rigor to the use of animal models in PD research (ANIMOD). In this, as with their own research, they led the way.