Abstract
Peritoneal dialysis (PD) is a more continuous alternative to haemodialysis, for patients with chronic kidney disease, with considerable initial benefits for survival, patient independence and healthcare cost. However, longterm PD is associated with significant pathology, negating the positive effects over haemodialysis. Importantly, peritonitis and activation of macrophages is closely associated with disease progression and treatment failure. However, recent advances in macrophage biology suggest opposite functions for macrophages of different cellular origins. While monocyte-derived macrophages promote disease progression in some models of fibrosis, tissue resident macrophages have rather been associated with protective roles. Thus, we aimed to identify the relative contribution of tissue resident macrophages to PD induced inflammation in mice. Unexpectedly, we found an incremental loss of homeostatic characteristics, anti-inflammatory and efferocytic functionality in peritoneal resident macrophages, accompanied by enhanced inflammatory responses to external stimuli. Thus, alterations in tissue resident macrophages may render longterm PD patients sensitive to developing peritonitis and consequently fibrosis/sclerosis.
Introduction
An estimated 5-10 million people worldwide die every year due to chronic kidney disease (1). In Europe, an average of 850 people per million population (pmp) require renal replacement therapy (RRT) and 120 new patients pmp commence treatment annually (2). The average 5-year survival rate of patients receiving RRT is only 50.5%. This can be improved to over 90% if patients receive a kidney transplant, but rates of transplantation remain low (32 pmp), primarily due to organ availability and the majority of patients rely on dialysis as a therapy to substitute excretory kidney function (2).
Peritoneal dialysis (PD), utilizing the body’s own peritoneal membrane as a filter during dialysis, is a cost-effective alternative to haemodialysis (HD). Although PD has been associated with better initial survival rates (3), lower costs for the health system (4) and increased patient autonomy (5) as compared to HD, the incidence rate of PD over several decades has dropped in Europe (3). There are a variety of reasons for this reduction in PD uptake, but the significant risk of adverse effects and in some cases fatal outcomes, has limited the general adoption of PD in adult patients across Europe (6). Treatment failure is commonly associated with repeated episodes of peritonitis (i.e. inflammation) and a progressive thickening and vascularisation of the peritoneum, leading to impaired filtration and thus reduced efficacy of PD (6, 7). In rare cases, the fibrotic changes to the peritoneum become so extreme that they form a fibrous cocoon encapsulating the internal organs leading to persistent or recurring adhesive bowel obstruction (7-9). The recognition of this Encapsulating Peritoneal Sclerosis (EPS) is an indication to discontinue PD with the mortality approaching 50% one year after diagnosis (10). Management of EPS includes surgery but there is a relatively high frequency of symptom recurrence (10). In contrast, immunosuppressive treatments and the use of anti-fibrotic agents, like tamoxifen, have shown noticeable benefits to patient survival (10). Therefore, aberrant activation of the immune system appears to be linked to both alteration of peritoneal structure and PD treatment failure, as well as the progression of the fibrotic sequelae. Indeed, experimental rodent models of the disease, have suggested inappropriate or excessive activation of macrophages as a major cause of the pathology (11-14).
Recent advances in macrophage biology have highlighted the intrinsic heterogeneity of macrophage populations (15). Grossly simplified, macrophages can be split into tissue resident macrophages (MΦres) which are present in tissues during homeostasis, and monocyte-derived macrophages (MΦmono), which are recruited to the tissue during inflammatory conditions (16). Both macrophage populations can and do respond to external stimuli, like infection, but they possess distinct response profiles and adopt distinct functional properties (17, 18).
In health, peritoneal MΦres are essential for maintaining tissue homeostasis by silently removing apoptotic cells through efferocytosis (19, 20) and by providing a source of tissue-reparative cells that infiltrate surrounding tissues (e.g. the liver) during injury (21). Importantly, upon encountering inflammatory signals, MΦres undergo the Macrophage Disappearance Reaction (MDR) (22). Following injection of inflammatory agents (e.g. bacterial antigens) or infection the number of peritoneal MΦres detectable in peritoneal lavage will drop significantly within a few hours (23). The exact mechanism underlying MDR is still poorly understood, but it has been proposed that MΦres undergo activation-induced cell death (18) or adhere to the mesothelial lining of the peritoneal cavity reducing their recovery (21, 24, 25). Of note, the degree of detectable MΦres -loss directly correlates to the amount of inflammatory stimulus (e.g. cfu of bacteria) and the recruitment of inflammatory MΦmono (18). Following resolution of the inflammation, MΦres can re-populate the peritoneal cavity and return to homeostatic numbers through proliferative expansion of remaining MΦres (26).
In models of fibrotic disorders of the lung or kidney, influx of monocytes and monocyte-derived macrophages has been linked to disease progression and induction of pathology (27-29). Similarly, in rodent models of peritoneal fibrosis, preventing the influx of monocytes or depleting all macrophages limits the degree of peritoneal thickening and improves glomerular filtration (12, 30, 31). Moreover, injection of MΦmono, activated ex vivo using bacterial antigens (i.e. Lipopolysaccharide), often referred to as M1 macrophages, exacerbates disease progression (13). Together these data highlight the role of inflammation and infiltration of MΦmono in the progression of peritoneal fibrosis and seem to provide a cohesive picture explaining the enhanced risk of PD failure associated with repeated episodes of peritonitis (32). In this context, it is interesting to note that continuing peritoneal irrigation and thus continual removal of peritoneal cells, including any inflammatory infiltrate, in patients discontinuing PD, has been suggested to prevent subsequent EPS formation (33).
Other studies have found a role for anti-inflammatory, IL-4 activated macrophages (also called M2), characterised by the expression of CD206, Arg1 and Ym1, in promoting peritoneal dialysis fluid (PD fluid) induced fibrosis (12, 34, 35). Additionally, chronic fibrotic kidney disease is linked to a switch from inflammatory M1 to predominantly M2 activated macrophages (36). Indeed, both types of macrophage activation seem to have the capacity to promote kidney fibrosis, possibly acting during different phases of the pathology (37). Importantly, some markers used to define M2 are differentially expressed on MΦres and MΦmono. For example CD206 is expressed constitutively by MΦmono but not MΦres (17). Thus, the described role of M2 macrophages may merely reflect an enhanced influx of MΦmono rather than IL-4 mediated activation. In fact, transfer of exogenously activated M2 macrophages showed no impact on disease progression in a model of peritoneal fibrosis (13) and even reduced pathology in a model of renal fibrosis (38). Moreover, using different strategies to deplete macrophages in kidney fibrosis yielded opposing results with regard to disease progression indicating that depletion of different subsets of macrophage may lead to different outcomes (39). Indeed, renal MΦres as compared to MΦmono have been shown to be protective in a model of kidney pathology (29). This suggests that the cellular origin of macrophages may play a prominent part in determining their role during fibrotic disorders and influence the overall disease outcome.
Here we analysed the effects of repeated PD fluid injection on macrophage population dynamics and responses to activating signals. Our data indicate a significant change in MΦres phenotype over time during PD fluid administration. MΦres lost expression of anti-inflammatory and efferocytic markers correlating with the duration of PD fluid treatment. Moreover, MΦres from PD fluid treated animals showed enhanced inflammatory responses to external stimulation. Importantly, the enhanced inflammatory phenotype of MΦres persisted even when PD fluid administration was stopped. Thus, repeated exposure to PD fluid may render patients more susceptible to peritonitis, and by extension to peritoneal fibrosis, due to exaggerated inflammatory responses of MΦres.
Results
Dialysis fluid induces the disappearance of tissue resident macrophages
To determine the impact of peritoneal dialysis fluid (PD fluid) injection on peritoneal macrophage populations (see suppl. fig 1 for gating strategy) we first characterised the cellular dynamics induced after a single application of PD fluid. Six hours after PD fluid injection, tissue resident macrophages (MΦres) underwent a pronounced disappearance reaction with the number of F4/80 high MΦres reduced to approximately 30% of the levels found in naive control animals (Fig 1A). Simultaneously, a significant influx of neutrophils and Ly6C high monocytes were detected, indicative of an inflammatory response (Fig 1B&C). Of note, similar to previous reports (23) tissue dwelling, monocyte-derived MΦmono (F4/80 low MHC-II high) also underwent a disappearance reaction and were reduced in numbers by approximately 40 % (Fig 1D).
F4/80 high MΦres displayed limited signs of activation at this early time point post PD fluid injection. No increase in expression of MHC-II could be detected (Fig 1E). In contrast intracellular Ym1 and CD206 expression were significantly enhanced following PD fluid injection, but expression was restricted to less than 10 % of cells (Fig 1F&G). The disappearance of MΦres in this context seemed in part due to enhanced cell death as indicated by increased levels of Annexin V staining (Fig 1H). By 24 hours post PD fluid injection, the numbers of MΦres had returned to baseline levels and expression of CD206 was no longer significantly different, whereas Ym1 expression remained elevated compared to naïve controls (Fig S2).
Overall, these data are consistent with the induction of low grade inflammation caused by PD fluid injection accompanied by a significant change in the prevalence of various myeloid cell populations within the peritoneal cavity.
Repeated PD fluid treatment induces a gradual change in MΦres phenotype
We next sought to determine the long term effect of PD fluid administration on peritoneal macrophage populations. For this animals were injected once per day, 5 times a week with PD fluid i.p. and the peritoneal exudate cells (PEC) analysed 24 hours after 1, 4, 9 and 14 injections, respectively (Fig 2A). Over the course of the experiment the number of MΦres remained consistently lower in Physioneal treated animals as compared to naive mice (Fig 2B). Simultaneously, enhanced influx of Ly6C monocytes and successive accumulation of F4/80 low MΦmono could be detected in the peritoneal cavity following multiple rounds of PD fluid injection (Fig 2C& D).
Despite the overall reduced numbers of MΦres, repeated injection of PD fluid led to marked induction of Ki67 expression in MΦres, indicative of proliferative expansion (Fig. S3). This is in line with a previous report demonstrating repopulation of the peritoneal cavity by MΦres following an inflammatory insult through proliferation (40). Thus, it is likely that PD fluid injection leads to a repeated cycle of MΦres disappearance followed by re-population. Similarly, it is likely that the gradual influx of monocytes and MΦmono underlies cyclic fluctuations after each round of PD fluid injection. Thus, this data indicated that macrophage populations within the peritoneal cavity undergo dynamic but limited changes following PD fluid instillation, with a slight accumulation of MΦmono over time.
Closer investigation of the MΦres phenotype, however, revealed a progressive loss of characteristics that define the tissue resident phenotype. In particular markers associated with the efferocytic and anti-inflammatory function of MΦres were expressed at increasingly lower levels following repeated injection of PD fluid (Fig 2E - G). T cell immunoglobulin and mucin domain containing 4 (Tim4), a molecule associated with the efficient removal of apoptotic cells (41, 42) as well as V-set and immunoglobulin domain– containing 4 (Vsig4), associated with limiting inflammatory responses (43, 44)], are specifically expressed in MΦres during steady state conditions (45). MΦres gradually lost expression of these markers with increasing number of PD fluid injections (Fig 2 E & G). Similarly, CD73, an anti-inflammatory effector molecule specifically expressed by MΦres (46, 47) was found to be significantly reduced after 14 injections of PD fluid (Fig 2 F). In contrast, expression of CD102 (Icam2) was not altered by PD fluid injection (Fig 2 H). A similar phenotype (loss of MΦres marker expression with sustained CD102 expression) has previously been described in mice lacking the transcription factor Gata6, indicating that repeated PD fluid injection may exert its effects via affecting Gata6 expression (48). However, no significant difference in Gata6 expression by MΦres was found after 9 injections of PD fluid measured in subsequent experiments (Fig 2 I). Notably, Gata6 expression levels varied following PD fluid injection with diminished expression detected in approximately 50 % of animals (Fig 2 I). Thus, any effect of PD fluid injection on Gata6 expression may be transient or reflect the impact of other factors, like the degree of inflammation.
Taken together, following repeated PD fluid injection, MΦres, while retaining their tissue identity (F4/80 high CD102+ Gata6+), lose some of their functional characteristics and in particular anti-inflammatory and efferocytosis associated functions.
Repeated PBS injection but not chronic bacterial infection leads to a loss of MΦres characteristics
As shown in Figure 1 and previously reported (13), repeated injection of PD fluid is associated with the induction of low grade inflammation as evidenced by influx of neutrophils and Ly6C+ monocytes. Thus, to assess whether the loss of MΦres functional markers (i.e. Tim4, CD73, Vsig4) was due to the elicited inflammatory response, we analysed peritoneal exudate cells from animals subjected to prolonged bacterial infection. Animals were infected orally with attenuated Salmonella enterica ser. Typhimurium (SL3261, ΔaroA) and peritoneal cells analysed during the chronic phase of infection, 34 and 55 days post infection. In line with previous data (18) and confirming the inflammatory environment, MΦres from SL3261 infected animals showed clear upregulation of MHC-II as well as enhanced expression of Sca-1 on day 55 (Fig 3A & B). Moreover, persistent influx of neutrophils was detected in the peritoneal cavity of SL3261 infected animals (Fig 3C) as well as live, cfu-forming bacteria (Fig 3D) confirming an ongoing pro-inflammatory activation.
However, unlike following repeated PD fluid injection, MΦres isolated from animals harbouring S. Typhimurium did not show any noticeable loss of Tim4, CD73 or Vsig4 expression in the chronic phase of the infection (Fig 3 E-G). Rather to the contrary MΦres from SL3261 infected animals expressed elevated levels of Tim4 (Fig 3 E). Thus, chronic inflammatory conditions alone did not cause the progressive loss of MΦres phenotype as observed following repeated Physioneal injection.
Importantly, injection of sterile PBS instead of PD fluid induced similar, albeit less pronounced changes in peritoneal macrophage phenotype (Fig 4A). This would indicate that repeated disturbance of the peritoneal immune system alone, rather than specific constituents of the dialysis fluid, was sufficient to drive the observed alterations in peritoneal MΦres. However, PD fluid enhanced these effects (Fig 4A).
Furthermore, it has been suggested that male peritoneal dialysis patients have a significantly reduced survival rate on PD than female patients (49). Although the reason for this discrepancy is unknown and likely due to multiple factors, differences in the inflammatory response may contribute to the observed effects. Moreover, male and female mice differ considerably in the maintenance and cellular dynamics of peritoneal macrophages (50). Thus, we assessed the effect of repeated PD fluid injection on myeloid cell populations in male and female mice. MΦres from naive male or female mice showed comparable expression levels of CD73 and Vsig4. In contrast naive male mice showed considerably lower expression of Tim4 (Figure 4B-D). However, independent of these differences in the steady state, Physioneal induced the loss of Tim4, CD73 and Vsig4 in both sexes to a similar degree.
Taken together this data shows that repeated disturbance of the peritoneal environment triggers low grade inflammation which alters the MΦres phenotype.
Prolonged PD fluid injection alters the response of MΦres to external stimuli
Our previous data had highlighted a significant loss of Tim4 as well as Vsig4 and CD73, markers which have been associated with MΦres core functions (i.e. efferocytosis and anti-inflammatory activity) (41, 44, 51). Thus, we next aimed to assess whether repeated injection of dialysis fluid altered macrophage functional responses.
To test the capacity of MΦres to take up and ingest apoptotic cells, whole PEC from animals injected for various times with Physioneal or from naive controls were incubated in the presence of apoptotic thymocytes labelled with pHrodo. pHrodo labelled cells emit a very low, nearly undetectable fluorescent signal after staining, but will become brightly fluorescent and clearly detectable by flow cytometry after encountering an acidic environment, as found inside a phagolysosome (52). Thus, use of pHrodo allows the reliable detection of ingested apoptotic cells as compared to labeled cells bound to the surface of a phagocyte.
Repeated injection of Physioneal gradually reduced the proportion of myeloid cells (total CD11b+lin-) capable of ingesting apoptotic cells (Fig 5A). This was in part due to the increased proportion of Ly6C high monocytes and F4/80 low, monocyte-derived macrophages within the myeloid cell pool (Fig 2C&D), cells which possess reduced efferocytic activity (53). However, when the analysis was restricted to MΦres a similar progressively reduced capacity to ingest apoptotic cells was detected (Fig 5B). Thus, MΦres from Physioneal injected animals seemed to lose the functionality to carry out efferocytosis efficiently. Of note, this loss of efferocytic capacity was not restricted to the use of dialysis fluid, as repeated injection of sterile PBS induced a similar reduction in the capacity of MΦres to ingest apoptotic cells (Fig S4A).
To further analyse whether MΦres from Physioneal injected animals were in general less responsive to external stimuli, we subjected whole PEC to in vitro stimulation with LPS and rIFNγ. MΦres from PD fluid treated animals showed a gradually increasing inflammatory response towards bacterial stimulation (Fig 5C&D). The proportion of cells expressing NOS2 or Sca-1, markers of pro-inflammatory, M1 activation (54, 55), was consistently higher in cells from PD fluid treated animals as compared to cells from naive animals (Fig 5C&D). Indeed, Sca-1 was not found to be upregulated on naive MΦres after 6h stimulation with IFNγ/LPS in vitro, indicating repeated Physioneal injections resulted in a stronger and more rapid response of MΦres to pro-inflammatory stimuli.
Importantly, an enhanced activation profile was not only observed in response to pro-inflammatory stimuli, but also in response to rIL-4, a driver of M2, anti-inflammatory macrophage activation (56). MΦres stimulated with rIL-4 for 24h showed increased expression of Ym1 and at a later timepoint also Relmα (Fig 5E&F).
Thus, MΦres altered their functional repertoire following repeated exposure to dialysis fluid with a reduced efferocytic capacity and enhanced responses to both M1 and M2 polarising stimuli.
Altered PD fluid-induced responsiveness of MΦres is maintained after treatment is discontinued
Next we examined whether the phenotypical and functional changes observed in MΦres following repeated PD fluid injection were temporary or persisted even after treatment ceased. For this mice were injected 5 times a week with Physioneal or PBS for a total of 9 injections, a timepoint when the altered MΦres phenotype is evident (Fig 2), and then rested for seven days. Subsequently, the PEC were collected and analysed for the expression of activation markers as well as re-stimulated in vitro to analyse their response to external stimuli as described above. MΦres from discontinued Physioneal or PBS injected animals maintained a reduced expression of CD73, Tim4 and Vsig4 compared to naive controls (Fig 6A). Similarly, MΦres from Physioneal and PBS injected animals showed an enhanced response to in vitro stimulation with LPS/IFNγ (NOS2, Sca1; stimulated for 6 h) (Fig 6B). These data highlight that repeated PD fluid injection sensitise peritoneal macrophages and, thus, potentially exacerbates the severity and the detrimental sequelae of peritonitis events.
In order to verify that repeated exposure to PD fluid alters the response of peritoneal macrophages to inflammatory stimuli in vivo we utilised a model of sterile inflammation.Animals were subjected to repeated PD fluid injection (3 times per week for a total of 9 injections) and an inflammatory response was induced with a single dose of sterile 4 % Brewer’s modified thioglycollate medium at day 6, mimicking an intermittent event of peritonitis (see Fig 6C). Twenty-four hours after the last PD fluid injection MΦres were collected and analysed for the expression of Tim4, CD206 and Ym1. As expected, repeated injection of dialysis fluid led to a significant loss of Tim4 expression on MΦres with limited induction of CD206 or Ym1 expression (Fig 6D-F). Additionally, in line with our in vitro findings (Fig 5), intermittent induction of inflammation did not further alter MΦres phenotype as determined by Tim4 expression, but led to enhanced activation of MΦres adopting a pro-fibrosis associated phenotype (Fig 6D-F). Importantly, expression of CD206, which has been closely linked to induction of fibrosis (12, 34, 35), was thought to not be expressed by MΦres (17).
Overall these findings indicate that prolonged and repeated exposure to PD fluid renders peritoneal macrophages more responsive to stimulation potentially exacerbating detrimental inflammatory responses.
Discussion
Macrophages are versatile cells implicated in many pathologies (57). They play both essential protective as well as detrimental / pathological roles, often within the same disease setting (58-60). Thus, the regulatory mechanisms governing macrophage responses have become the focus of current research aiming to dissect these contradictory behaviours. Moreover, scientists have proposed macrophages as an excellent target for therapeutic interventions, as altering their phenotype may not only improve disease outcome, but actively revert associated pathologies.
Recent discoveries have highlighted the inherent diversity of the macrophage pool, indicating the presence of different types of macrophages with different ontogeny (61), response profiles (17) and functional roles (18). In particular, MΦres and acutely recruited MΦmono have been identified as distinct mediators of pathology. During lung fibrosis recruited MΦmono have been shown to be essential drivers of the fibrotic pathology and disease progression (27). Similarly, in murine models of peritoneal dialysis associated fibrosis, infiltration of MΦmono has been shown to be detrimental to disease outcome (12, 13, 30, 62). Furthermore, significant changes in the prevalence of specific macrophage/monocyte populations have been described in dialysis patients, dependent on the history of peritonitis episodes (11). Thus, the pathology inducing effect of peritonitis associated inflammation in peritoneal dialysis seem to be due to the recruitment and accumulation of fibrosis promoting cells like MΦmono. Due to legal restrictions on the maximum number of i.p. injections allowed, we did not observe any clinical signs of fibrosis in our experimental setup. However, similar to the findings above, we observed low-grade inflammation and infiltration of MΦmono in our system. In addition, we demonstrated a drastic change in phenotype of MΦres. Importantly this change in phenotype translated into an enhanced response profile boosting both pro-inflammatory as well as M2 associated effector molecule production. Thus, the simple dichotomy of good and bad macrophages cannot be upheld in these settings. The increased risk of developing pathological sequelae following repeated episodes of peritonitis in PD patients is likely due to the damaging effect of the inflammatory response (63). Our data implies, that patients on longterm PD treatment are at an increased risk to develop clinical symptoms of peritonitis, not only because of an increased risk of exposure to bacterial antigens, but also due to an enhanced inflammatory response of peritoneal resident cells. Whereas in healthy people small numbers of infiltrating pathogens are silently removed by MΦres without triggering an inflammatory response, PD fluid conditioned MΦres may have lost some of this regulatory capacity leading to exacerbated pathology-promoting inflammation. Thus, targeting MΦres to restore some of their homeostatic, anti-inflammatory capabilities may provide a feasible approach to extend the therapeutic window of PD. Indeed, adding soluble factors (e.g. soluble TLR2) to the dialysis fluid targeting the inflammatory response has previously been suggested (64) and could be adapted to specifically target MΦres.
Intriguingly in PD patients receiving oral supplementation with vitamin D, a factor involved in the expansion of peritoneal MΦres in mice (65), antibacterial responses were enhanced (66), potentially through improved phagocytosis of invading bacteria (67). MΦres in virtually all tissues of the body are seeded during embryonic development and maintained locally through proliferative expansion. However, in many tissues, including the peritoneal cavity, these embryonic derived macrophages are, over time, replaced by monocyte-derived cells adopting features of tissue resident cells (50). Thus, it may be feasible to develop therapeutic approaches fostering the non-inflammatory removal of bacteria by promoting MΦres differentiation. However, the conversion of MΦmono in the peritoneal cavity under steady state seems inefficient and leads to a gradual accumulation of Gata6+, F4/80 high MΦres with low expression of Tim4 (50). Our data from Salmonella enterica ser. Typhimurium infected mice indicated that the late stages of infection were linked to an increase in Tim4 expressing MΦres as compared to naive animals. Similarly, previously published results from helminth infected animals indicate enhanced expression of Tim4 on MΦres during the chronic phase of these infection (68). Hence, factors associated with the resolution of inflammation or chronic Th2 immune responses may allow to ‘rejuvenate’ monocyte-derived MΦres or promote the conversion of MΦmono. Further research is required to determine the specific factors driving the MΦres phenotype as well as their potential use as therapeutic adjunct during PD.
In this context it is important to consider that peritonitis or PD induced inflammation will drive the disappearance of MΦres and, thus, may undo any previous accumulation / conversion attempts. It has previously been suggested that the loss of MΦres is at least in part due to enhanced cell death of resident cells during inflammation (18). Importantly, the degree of loss of MΦres is directly correlated to the degree of inflammation as well as the number of MΦres present within the peritoneal cavity (18). Thus, targeted expansion of MΦres prior to or during PD fluid treatment may provide an option to retain enhanced numbers of MΦres during inflammatory events. In contrast, recent work suggests that the apparent reduction in the number of MΦres detectable in the peritoneal wash is due to the adherence of these cells to the peritoneum rather than an actual loss in numbers (25). This is especially noteworthy in the context of peritoneal dialysis as it would bring these cells in close contact to the main site of pathology. Moreover, targeted disruption of the coagulation cascade reduced the apparent loss of MΦres and prevented their adherence (25) potentially opening a window for therapeutic intervention. Currently neither approach (increasing MΦres numbers or preventing the MDR) is likely suitable for use in human patients due to systemic and off-target effects of the applied methodology. However, both approaches indicate intriguing opportunities to develop novel techniques specifically targeting peritoneal MΦres.
Independent of the therapeutic potential discussed above, analysing the immune cells contained in the effluent of PD patients may yield a useful biomarker strategy. Changes in the composition of myeloid cells as described by Liao et al. (11) as well as the assessment of cellular activation markers as described in mice here, may allow risk-based patient stratification. Patients likely to develop pathological sequelae of PD treatment could then be prioritised for kidney transplant or transferred to haemodialysis prior to PD failure or before overt pathological changes occur.
Lastly, our data comparing PBS injected animals to Physioneal 40 injection indicated a limited impact of the composition of the PD fluid on the observed physiological changes. This seems in contrast to the commonly held believe that the constituents of PD fluid, in particular lactate, glucose-degradation-products (GDP) and low pH, are instrumental in driving PD related pathology (69, 70). However, our data is in line with a recent report investigating the use of so-called biocompatible PD fluid in patients. Despite considerably lower levels of GDP and lactate in these solutions, the morphological changes observed were very similar to those seen in patients utilising standard PD fluid and even an early increase in microvascular density was detected (71). Of note, it has been suggested that this adverse effect of biocompatible PD fluid may be due to an altered inflammatory response (72). Thus, together with our data this suggests it may be more effective to prevent pathological changes by targeting the elicited immune response to PD fluid instillation, even in the absence of peritonitis.
Taken together we have shown here that repeated exposure of the peritoneal microenvironment to PD fluid instillation in mice led to a gradual change in phenotype as well as activation response of peritoneal resident cells. These changes were relatively long lasting and resulted in a more vigorous response to inflammatory triggers. Thus, targeting MΦres and preventing excessive inflammatory responses in PD patients may pose an exciting novel approach to limit PD-related pathology.
Methods and Methods
Ethics Statement
All animal experiments were performed in accordance with the UK Animals (Scientific Procedures) Act of 1986 under a Project License (70/8548) granted by the UK Home Office and approved by the University of Manchester Ethical Review Committee.
Mice and in vivo treatments
Eight to 13 week old male and female C57BL/6 mice were obtained from a commercial vendor (Envigo, Hillcrest, UK) and maintained in groups of 4-6 animals in specific pathogen–free facilities at the University of Manchester. Experimental mice were randomly allocated to treatment groups using a computer-based randomization technique and age and sex matched. Euthanasia was performed by asphyxiation in carbon dioxide in a rising concentration. Animal numbers were based on initial power calculations of key determinants derived from preliminary experiments.
Animals were injected i.p. with 500 μL peritoneal dialysis fluid (Physioneal 40, 3.86% glucose, Baxter HealthCare Ltd., Compton, UK) for the indicated number of injections. Injections were carried out daily or every other day for three or five days per week for up to 4 weeks (maximum 14 injections per animal). Control animals were left untreated or received an equal volume of PBS i.p. as indicated.
For the induction of inflammatory responses PD fluid treated mice received intra-peritoneal injections of 400 μL 4% Brewer modified thioglycollate medium (BD Biosciences, San Jose, CA) or sterile saline as control. The attenuated Salmonella enterica serovar Typhimurium strain SL3261 (ΔaroA) (73) was cultured overnight at 370C in a shaking incubator from frozen stock in Luria-Bertani broth with 50 μg/mL streptomycin. The following morning, culture was diluted in fresh Luria-Bertani broth with 50 μg/mL streptomycin and incubated at 370C in a shaking incubator to ensure the bacteria were in the growth phase. CFU/mL was estimated by the OD600 reading. Animals were pre-treated with 20 mg streptomycin 1 day prior to oral infection with ∼1×10^8 CFU Salmonella Typhimurium diluted in PBS. Infectious doses and peritoneal bacterial burdens were enumerated by plating inocula or peritoneal exudate cells in 10-fold serial dilutions in PBS on LB-Agar plates.
Cell-isolation
Peritoneal cavity exudate cells (PEC) were obtained by washing the cavity with 10 mL lavage media comprised of RPMI 1640 (Sigma-Aldrich, Dorset, UK) containing 2 mM EDTA and 1% L-Glutamine (Thermo Fisher Scientific, Waltham, MA). Erythrocytes were removed by incubating with red blood cell lysis buffer (Sigma-Aldrich). Cellular content was assessed by cell counting using Viastain AO/PI solution on a Cellometer ® Auto 2000 Cell Counter (Nexcelom Bioscience, Manchester, UK) in combination with multicolor flow cytometry.
Flow cytometry
Equal numbers of cells were stained with Zombie UV viability assay (Biolegend, London, UK). All samples were then blocked with 5 μg/mL anti CD16/32 (93; Biolegend) and heat-inactivated normal mouse serum (1:10, Sigma-Aldrich) in flow cytometry buffer (0.5% BSA and 2 mM EDTA in Dulbecco’s PBS) before surface staining on ice with antibodies to F4/80 (BM8), SiglecF (E502440, BD Biosciences), Ly6C (HK1.4), Ly-6G (1A8), TCRβ (H57-597), CD11b (M1/70), CD11c (N418), I-A/I-E (M5/114.15.2), CD19 (6D5), CD115 (AFS98), CD73 (TY/11.8), CD102 (3C4 (MIC2/4)), Tim4 (RMT4-54), CD206 (MR6F3), Vsig4 (NLA14; Thermo Fisher Scientific), CD226 (10E5), Sca1/Ly6A (D7). All antibodies were purchased from Biolegend unless stated otherwise.
Detection of intracellular activation markers was performed directly ex vivo. Cells were stained for surface markers then fixed and permeabilized for at least 16 h using the eBioscience™ Foxp3 / Transcription Factor Staining Buffer Set (Thermo Fisher Scientific). Cells were then stained with directly labeled Abs to NOS2 (CXNFT; Thermo Fisher Scientific), Arg1 (polyclonal, R&D Systems, Abingdon, UK), Ki67 (B56, BD Biosciences), Annexin V (Biolegend), Gata-6 (D61E4; Cell Signaling Technology Europe, Leiden, Netherlands) or purified polyclonal rabbit antiRelmα (PeproTech EC Ltd., London, UK) and biotinylated anti-Ym1/2 (polyclonal, R&D Systems) followed by Zenon anti–rabbit reagent (Thermo Fisher Scientific) or streptavidin BUV 737 (BD Biosciences), respectively. Samples were acquired on a BD LSR II or BD FACSymphony using BD FACSDiva software (BD Biosciences) and post-acquisition analysis performed using FlowJo v10 software (BD Biosciences).
Cell-culture experiments
For in vitro stimulation of PD fluid-conditioned cells, whole PEC were counted as described above and seeded to 96-well U bottom plates at 3×10^5 cells per well in RPMI 1640 containing 5 % foetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin and stimulated with lipopolysaccharide (LPS, 100 ng/mL; Salmonella enterica ser. Typhimurium; Sigma-Aldrich) and recombinant murine Interferon γ (IFNγ, 20 ng/mL PeproTech EC Ltd.) or medium alone for 6 h or with murine recombinant Interleukin–4 (rIL-4, 20 ng/mL, PeproTech EC Ltd.) for 24 h and analysed for MΦ activation markers by flow cytometry.
Apoptotic cell uptake assay
Uptake of apoptotic cells by peritoneal MΦres was assessed as previously described (52). Briefly, thymocytes were collected from naive animals by mincing thymi through 2 μm gauze until completely homogenized. Erythrocytes were removed by incubating with red blood cell lysis buffer (Sigma-Aldrich). Thymocytes were resuspended at 1×10^7 cells/mL in complete DMEM and incubated in the presence of 0.1 µM dexamethasone (Sigma-Aldrich) at 37 °C for 18 h. This produced >90 % apoptosis, as assessed by Viastain AO/PI staining measured on a Cellometer ® Auto 2000 Cell Counter (Nexcelom Bioscience). Subsequently, apoptotic thymocytes were washed twice with PBS and resuspended in PBS at 10^6 cells/mL containing 40 ng/mL pHrodo-SE (Thermo Fisher Scientific) and incubated at RT for 30 min. Thereafter the cells were washed twice with PBS and resuspended in RPMI containing 5 % foetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin. Unstained apoptotic cells served as staining control.
Statistical analysis
Statistical analysis was performed using Prism 8 for Mac OS X (v8.2.1, GraphPad Software, San Diego, CA). Differences between groups were determined by t-test or ANOVA followed by Tukey’s or Dunn’s multiple comparison-test. In some cases data was log-transformed to achieve normal distribution as determined by optical examination of residuals. Where this was not possible a Mann-Whitney or Kruskal-Wallis test was used. Percentages were subjected to arcsine transformation prior to analysis. Differences were assumed statistically significant for P values of less than 0.05.
Author contributions
- Tara E Sutherland (Resources, Writing – Review & Editing, Funding Acquisition)
- Tovah N Shaw (Investigation, Writing – Review & Editing)
- Rachel Lennon (Resources, Writing – Review and Editing, Clinical advice)
- Sarah E Herrick (Conceptualization, Writing – Review and Editing)
- Dominik Ruckerl (Conceptualization, Formal Analysis, Validation, Investigation, Writing - Original Draft Preparation, Writing – Review and Editing, Visualization, Project Administration, Funding Acquisition)
Competing interests
The authors declare that they have no competing interests.
Acknowledgements
This work was supported by the Medical Research Council UK (MR/P02615X/1, DR) and the Medical Research Foundation / Asthma UK (MRFAUK-2015-302, TES). TNS was supported by a BBSRC Discovery Research Fellowship (BB/S01103X/1, TNS), RL was supported by a Wellcome Trust Senior Fellowship award (202860/Z/16/Z, RL) and SEH and DR were supported by an MRC Research Grant (MRC MR/S02560X/1, SEH). We thank Dr. Gareth Howell from the Flow Cytometry Core Facility at the University of Manchester for assistance with the flow cytometry analyses. We’d also like to thank Sister Trish Smith from the Royal Manchester Children’s Hospital for the provision of Physioneal 40.
Footnotes
References have been updated to match JI submission guidelines.