Dr T. E. Mollnes, Institute of Immunology, Rikshospitalet, Sognsvannsveien 20, N-0027 Oslo, Norway. E-mail: email@example.com
Microdialysis emerges as a useful tool to evaluate tissue inflammation in a number of clinical conditions, like sepsis and transplant rejection, but systematic methodological studies are missing. This study was undertaken to determine the recovery of relevant inflammatory mediators using the microdialysis system, comparing microdialysis membranes with two different molecular weight cut-offs at different flow rates. Twenty and 100 kDa pore sizes CMA microdialysis catheters were investigated using velocities of 0.3, 1.0 and 5.0 μl/min. Reference preparations for cytokines [tumour necrosis factor (TNF)-α, interleukin (IL)-1β, IL-6 and IL-10; m.w. 17–28 kDa] and chemokines (IL-8, MCP-1, IP-10 and MIG; m.w. 7–11 kDa) were prepared from plasma after incubating human whole blood with lipopolysaccharide. Reference preparation for complement anaphylatoxins (C3a, C4a, C5a; m.w. 9–11 kDa) was prepared by incubating human plasma with heat-aggregated immunoglobulin G. The reference preparations were quantified for the respective inflammatory molecules and used as medium for the microdialysis procedure. Through the 20 kDa filter only the four chemokines passed, but with low recovery (3–7%) and limited to the 1.0 μl/min velocity. The recovery with the 100 kDa filter was as follows: IL-1β = 75%, MCP-1 = 55%, MIG = 50%, IL-8 = 38%, C4a = 28%, IP-10 = 22%, C5a = 20%, C3a = 16%, IL-6 = 11, IL-10 = 8% and TNF-α = 4%. The highest recovery for all chemokines and anaphylatoxins were consistently at velocity 1.0 μl/min, whereas IL-1β and IL-10 recovered most efficiently at 0.3 μl/min. Thus, microdialysis using catheters with a cut-off of 100 kDa is a reliable method to detect inflammation as judged by a defined panel of inflammatory markers. These findings may have important implications for future clinical studies.
During the last two decades microdialysis has been established as a method to evaluate tissue metabolisms, in particular to detect tissue ischaemia. Small molecular weight metabolites (glucose, lactate, pyruvate, glutamate and glycerol) have been extensively investigated in many tissues using the traditional 20 kDa microdialysis filter. The microdialysis system consists of a pump with a syringe filled with perfusion fluid coupled to a double tubular microdialysis catheter (Fig. 1). The perfusion fluid is driven through the space between the inner tube and the outer dialysis membrane, then entering at the distal end the inner concentric tube. The dialysate then enters a microvial connected at the end of the tube. The amount of the substance appearing in the dialysate is dependent on the membrane pore size (cut-off), the length of the membrane, the size and properties of the substances that are to be studied and the flow rate (usually in the rage of 0.1–5 μl/min). Recovery is defined as the concentration of the substance in the dialysate in per cent of the actual concentration outside the membrane.
The introduction of 100-kDa membranes extends the application of microdialysis to include inflammatory mediators, enabling evaluation of the local inflammatory response as seen in, e.g. ischaemic injury and transplant rejection. However, many inflammatory mediators do not pass even the 100-kDa filter. Proteins of 10 kDa or larger has in general been found to have a low recovery in the range of 1–5% . A 3000-kDa filter, which is not available for routine use, enables diffusion with higher recovery as documented for human interleukin (IL)-1β (28%), IL-6 (45%) and neural growth factor (22%) in the brain  and IL-6 (42–50%) in connective and muscle tissues .
Despite the size-limited passage through the micropores in clinically used filters several studies have shown that cytokines can be evaluated using microdialysis. In particular IL-1β and IL-6 have been studied [4–8], but also MCP-1  and tumour necrosis factor (TNF)-α [10, 11], the latter limited to animal studies. Recovery is only occasionally examined in these studies and when reported it varies from below 1% to up to 50% depending upon the conditions used.
Studies to systematically evaluate the efficacy of microdialysis to reflect inflammation by measuring the recovery of a broad panel of cytokines and other inflammatory mediators are missing. Complement activation is an essential event in inflammation with formation of anaphylatoxins (C3a, C4a and C5a). It is unknown whether local complement activation can be evaluated using microdialysis. The aim of the present study was therefore to analyse, in vitro, the efficacy of microdialysis to reflect the concentration of the cytokines TNF-α, IL-1β, IL-6 and IL-10, the chemokines IL-8, MCP-1, IP-10 and MIG and the complement anaphylatoxins C3a, C4a and C5a, using reference preparations with known concentrations of the analytes. The effect of filter sizes (20 and 100 kDa) and flow rate (0.3–5 μl/min) was examined. The data indicate that the 100-kDa filter displays significant recovery of all the tested mediators. Thus, this catheter appears promising for future clinical studies to evaluate local inflammation.
Materials and methods
Microdialysis equipment and performance. The CMA Microdialysis System (Stockholm, Sweden) was used (Fig. 1). The microdialysis pump was CMA107, a portable battery-driven pump adjustable for different flow rates (0.1–5 μl/min). Two different microdialysis catheters were used. The 20-kDa CMA61 has a polyurethane shaft of 40 mm length and diameter of 0.9 mm and a polyamide membrane of 30 mm length and diameter of 0.6 mm. The 100-kDa CMA71 differs from the CMA 61 catheter only with respect to the membrane cut-off. Standard CMA perfusion fluid was used with the 20-kDa catheters, containing 147 mm sodium, 4 mm potassium, 2.3 mm calcium and 156 mm chloride. Plasmodex Pharmalink perfusion fluid was used with the 100-kDa catheters as recommended by CMA, containing 30 g/l Dextran 60, 130 mm sodium chloride, 4 mm potassium, 2 mm calcium, 1 mm magnesium, 32 mm acetate, 110 mm chloride; osmolality of 270 mosmol/kg and pH 6. The microdialysis syringes were filled with perfusion fluid, applied to the pumps and connected with the catheters. The catheters were perfused for 1 h before the start of the experiment to ensure proper fluid filling. The following three velocities were used in all experiments: 0.3, 1.0 and 5.0 μl/min. Dialysis time was 70 min in all experiments, which was the time required at low velocity to obtain adequate amounts of samples for analysis (>20 μl) at low velocity. The microdialysis catheters were placed in tubes with 2.5 ml of plasma containing various amounts of cytokines or serum containing various amounts of complement activation products. Sample volumes were as follows: 21, 60 and 300 μl, respectively, for the three velocities 0.3, 1.0 and 5.0 μl/min.
Reference preparation for cytokines and chemokines. Cytokine- and chemokine-containing plasma was produced from human whole blood obtained from 10 healthy voluntary donors after informed consent (approved by the local ethical committee). The blood was anticoagulated with 2.5 mg/ml Refludan (lepirudin; Hoechst Marion Roussel, Frankfurt am Main, Germany) as described previously . The blood samples were split into four tubes. One tube was immediately centrifuged at 900 g for 15 min at 4°C and the plasma (T-0) was stored at −70 °C until microdialysis. Lipopolysaccaride (LPS) (L-8274, Sigma Aldrich, Steinheim, Germany) was added to the other tubes to a final concentration of 100 ng/ml. The samples were incubated at 37 °C on a roller mixer (SRT 2, BIBBY Stuart Scientific, Staffordshire, UK) for 1, 4 and 24 h (in order to obtain optimal concentrations of the different cytokines), then centrifuged at 900 g for 15 min at 4 °C. The plasma samples (named T-1, T-4 and T-24 respectively) were stored at −70 °C. Cytokines and chemokines were quantified in the plasma samples as described below and the four samples (T-0 to T-24) were used as reference preparations for microdialysis.
Reference preparation for complement anaphylatoxins. C3a-, C4a- and C5a-containing samples was produced from lepuridin-anticoagulated plasma obtained from the volunteers by centrifuging the whole blood samples at 900 g for 15 min at 4 °C. Complement activation was obtained by incubating plasma with heat-aggregated immunoglobulin (HAIGG), 1 mg/ml, for 30 min at 37 °C. The plasma was centrifuged at 6000 g for 30 min at 4 °C and stored at −70°. Anaphylatoxins were quantified in the plasma as described below. Three reference preparations were used for microdialysis: (i) baseline concentration (normal plasma); (ii) modest concentration (10% activated plasma made by mixing one part of HAIGG-activated plasma with nine parts of normal plasma); and (iii) high concentration (100% HAIGG-activated plasma).
Analysis of cytokines, chemokines and anaphylatoxins. A total of 84 samples were analysed: four plasma samples containing different amounts of cytokines and three serum samples containing different amounts of anaphylatoxins underwent microdialysis using three different velocities and two different micodialysis catheters. Two independent experimental series were run on separate days. All mediators were analysed using cytometric bead array (CBA; BD Pharmingen, San Diego, CA, USA). The cytokines IL-1β, TNF-α, IL-6 and IL-10 were quantified using the Human Inflammation CBA-kit (Cat. no: BD 55181). The chemokines IL-8, MCP-1, IP-10 and MIG were quantified using the CBA-kit for Human Chemokines (Cat. no: BD 552990). The anaphylatoxins C3a, C4a and C5a were quantified using the CBA-kit for Human Anaphylatoxins (Cat. no: 552363). The assay principle is briefly as follows: mixed particles dyed with different fluorescence intensities, separately detected by flow cytometry, are coated with specific antibodies directed against the different analytes. Thus, the different capture beads will bind their specific analytes, which is then detected by specific antibodies conjugated with phycoerythrin (PE) present in the buffer solution. Thereby, several analytes can be quantified in the same sample with small amounts of sample volume. Data for all the kits were analysed using the CBA software (BD Pharmingen). Standard curves were generated for each analyte using the mixed analyte standard provided with the kit. Final sample volumes were 20 μl. Cytokine and chemokine samples were diluted 1:2 and the anaphylatoxin samples diluted 1:10 (low concentration) to 1:40,000 (high concentration). The assays were performed according to the description from the manufacturer.
Ethics. The study was approved by the Norwegian Regional Committee for Medical Research Ethics.
The inflammatory mediators studied are listed in Table 1. In the presentation below they are divided into three groups termed cytokines, chemokines and anaphylatoxins. The cytokines comprises IL-1β, TNF-α, IL-6 and IL-10 (m.w. 17–26 kDa). The chemokines, a subgroup of cytokines with structurally related molecules, comprises IL-8, MCP-1, IP-10 and MIG (m.w. 8–11 kDa). The anaphylatoxins are complement activation products with the same size as the chemokines and comprises C3a, C4a and C5a (m.w. 9–11 kDa).
Table 1. Characteristics of the inflammatory mediators included in the study.
Molecular weight (kDa)
Some biological properties
CD121a and b
A number of pro-inflammatory effects; T-helper cell stimulation
TNF ligand superfamily (TNFSF2)
CD120a and b
A number of pro-inflammatory effects; cytotoxic to tumour cells
A number of pro-inflammatory effects; a major acute phase inducer
Inhibits synthesis of a number of cytokines (‘anti-inflammatory’)
Chemotactic for all migratory immune cells; activates only neutrophils
Chemotactic for monocytes but not neutrophils, cytokine expression
Role in hypersensitivity reactions of the delayed type
Modulating growth, motility and activation of inflammatory cells
Modestly potent anaphylatoxin; inflammatory effects in the airways
No known biological effects
Potent anaphylatoxin; highly chemotactic; induces a broad range of inflammatory effects
Preparation of reference samples
Cytokines (Fig. 2, upper panel). All cytokines were generated in large amounts reaching ng/ml amounts. IL-1β and IL-6 started to increase after 1 h and reached maximum values after 24 h. TNF-α started to increase at 1 h, reached maximum after 4 h and decreased after 24 h. IL-10 increased later and reached maximum level after 24 h.
Chemokines (Fig. 2, lower panel). All cytokines were generated in large amounts reaching ng/ml amounts. They uniformly started to increase after 1 h and reached maximum levels at 24 h.
Anaphylotoxins (Fig. 3). All anaphylatoxins were generated in large amounts reaching μg/ml amounts as indicated in right columns in Fig. 3 (100%). A preparation of modest concentration (10%) was made by mixing normal plasma with activated plasma 9:1 (middle columns in Fig. 3).
Microdialysis and recovery
Cytokines. IL-1β, TNF-α, IL-6 and IL-10 did not pass the 20-kDa filter (0% recovery) at any of the velocities. The reference preparations with the highest concentrations yielded measurable recoveries using the 100-kDa filter, but the variation between the cytokines was considerable, i.e. IL-1β passed efficiently compared with the others. Thus, the 100 kDa recovery for IL-1β was 75% at 0.3 μl/min, decreasing to 35% at 1.0 μl/min and to 0% at 5.0 μl/min (Fig. 4, upper left panel). The corresponding recoveries for IL-10 were 8%, 2% and 0% (Fig. 4, upper middle panel). The maximal recoveries for TNF-α and IL-6 were 4% and 11% respectively (data not shown).
Chemokines. IL-8, MCP-1, IP-10 and MIG showed a homogenous passage pattern. They all consistently passed the 20-kDa filter with low recovery (4%, 7%, 3% and 12% respectively) at 1.0 μl/min, but did not pass (0% recovery) at 0.3 and 5.0 μl/min (data not shown). The 100 kDa recovery was also constantly highest at 1.0 μl/min. Using the reference preparation with highest concentration (24-h sample) the recoveries at this velocity were: IL-8 = 38%, MCP-1 = 55%, IP-10 = 22% and MIG = 50% (Fig. 4, upper right and lower three panels). The corresponding recoveries using the lower concentration reference (4 h sample) and 1.0 μl/min were 25%, 42%, 10% and 37%. The chemokine recovery was markedly dependent on velocity, as they could hardly be detected at 0.3 μl/min and amounted only 1–12% at 5.0 μl/min (Fig. 4).
Anaphylatoxins. C3a, C4a and C5a showed a homogenous passage pattern with a recovery similar to the chemokines with respect to velocity, i.e. the highest recovery was consistently found at 1.0 μl/min. They did not, however, pass the 20-kDa filter. The 100-kDa recoveries using the 100% reference preparation were 16%, 28% and 20% for C3a, C4a and C5a, respectively, at 1.0 μl/min (Fig. 5). The corresponding values at 0.3 μl/min were 5%, 8% and 6% and at 5.0 μl/min they were 3%, 7% and 5% (Fig. 4). Using the low concentration reference preparation (10% of fully activated plasma) the maximum recoveries were also seen at 1.0 μl/min, amounting 3%, 10% and 5%, respectively, for C3a, C4a and C5a (data not shown).
Summary of recovery data
The maximum recoveries obtained from high-concentration reference preparations for the different inflammatory mediators and the corresponding optimal velocities are indicated in Table 2. Of the 11 inflammatory molecules tested, nine showed recovery above 10%. The molecular relatively homogenous chemokines and anaphylatoxins recovered efficiently, whereas the recovery of the heterogenous cytokines differed substantially.
Table 2. Recovery of inflammatory mediators after microdialysis using a 100-kDa filter.
†Maximum recovery was obtained using the indicated velocity. The velocities included in the experiments were 0.3, 1.0 and 5.0 μl/min.
Microdialysis has mainly been used to study biochemistry and metabolism using conventional 20 kDa molecular weight cut-off membranes, enabling measurement of small molecules such as glucose, lactate, pyruvate, glycerol and glutamate. This has enabled the clinician to diagnose ischaemia in patients early enough to intervene at a time when the organ is salvageable. Introduction of 100-kDa filters have extended the application of microdialysis to evaluate tissue inflammation by measuring larger molecules like cytokines. This applies to patients with reperfusion injury, rejection or other inflammatory reactions, like for instance sepsis. The 20- and 100-kDa filters seem to be equal with respect to recovery of small molecules . No systematic study has been performed comparing the two filter sizes for recovery of a broader spectrum of inflammatory molecules including cyto/chemokines and complement anaphylatoxins. In the present study, we investigated three groups of totally 11 inflammatory molecules. The first group comprised four conventional cytokines (m.w. 17–16 kDa), the second four chemokines (m.w. 7–11 kDa) and the third three complement anaphylatoxins (9–11 kDa). Complement activation, an important inflammatory event, has not previously been investigated with microdialysis. Our data clearly indicate that the 100-kDa filter in contrast to the 20 kDa can be used to detect all these molecules, but the molecular weight alone does not predict the degree of recovery. Thus, the recovery of IL-1β was remarkably high (75% at highest concentration) compared to a lower recovery for TNFα, IL-6 and IL-10 (4–11%), although all three are in the same molecular weight range. The chemokines and anaphylatoxins have similar molecular weights, but still there was a difference in the recovery, ranging 38–55% for the chemokines and 16–28% for the anaphylatoxins.
Thus, other factors than the molecular weight contribute to the degree of recovery. These are in particular the diffusion coefficient, including the physical and chemical properties of the actual molecule, and the velocity of the perfusion fluid. The molecular quaternary structure and the degree of hydrophilicity are important factors determining the diffusion across membranes. While small hydrophilic molecules pass readily with high recovery, hydrophobic proteins stick to surfaces including the filter membrane and recovery is reduced. This may explain why the anaphylatoxins in general showed a lower recovery than the chemokines although they are of the same molecular weight. Still the recovery for the anaphylatoxins was substantial and these molecules are definitely candidates together with the chemokines when microdialysis is used for evaluation of tissue inflammation.
The perfusion fluid velocity influences the recovery substantially. In general, the recovery and velocity correlates inversely , consistent with our findings for the cytokines IL-1β and IL-10, belonging to the group with highest molecular weight. In contrast, all chemokines and anaphylatoxins consistently showed markedly higher recovery at 1.0 μl/min than at 0.3 and 5.0 μl/min. Mixup of samples and technical errors explaining this finding can be excluded for the following reasons: firstly, the whole experimental setup was performed two times with virtually identical results; secondly, the cytokines and chemokines were analysed in the same sample (muliplex assay) with consistent results in each run; thirdly, the low- and high-concentration preparations were run separately and showed consistent results. As the chemokines belongs to a homogenous group of proteins with similar chemical properties, as do the anaphylatoxins, in contrast to the higher molecular weight cytokines, the optimum reached at 1.0 μl/min for the former may be explained by more extensive hydrophobic binding of the proteins to the inner side of the membrane at low velocity (0.3 μl/min). By increasing the velocity the recovery increases due to less time for binding keeping the substance dissolved in the dialysate, until the recovery again decreases due to reduced passage through the pores (5.0 μl/min).
The concentration of the analytes outside the filter membrane influences the degree of recovery, i.e. the higher the concentration of the substance, the better the recovery. The local concentration of inflammatory mediators in tissues is low under physiological conditions, but can increase to very high levels during an inflammation. It is difficult to anticipate the local concentration of mediators as the plasma level does not reliably reflect the local process. In the present study, we therefore examined the recovery in reference preparations containing high and moderate concentrations of mediators. The recovery was clearly higher using high-concentration preparations, but for the chemokines and anaphylatoxins the recovery was substantial even when the low-concentration preparation was used, indicating that these two groups can be studied using microdialysis in tissues with variable concentrations of the molecules. Indeed, the general recovery in our study was substantial and even the low recoveries observed for the cytokines may be of significance for clinical studies. Techniques to improve the recovery during microdialysis has been developed, based on immunoaffinity electrophoresis to study low concentrations of cytokines in culture supernatants  and the use of antibody-coated microspheres in the dialysis fluid to increase the diffusion coefficient . By these approaches, the recovery was markedly improved and reached 50–95% for several of the cytokines. Still it is not the amount of recovery per se that is important for a correct clinical diagnosis, but the ability of the method to detect a correct trend. Thus, a 10-fold increase, as might be found in tissue inflammation, is as well detected for substances with low or moderate recovery as with high recovery.
Clinical studies using microdialysis has particularly been performed during neurosurgery for monitoring brain metabolism . There is, however, a need for local monitoring of tissue homeostasis in a number of clinical conditions like ischaemia in flaps in reconstructive surgery or a transplanted organ, during reperfusion injury or rejection. For instance, the diagnosis of rejection might be aided by microdialysis. Thus, evaluation of liver metabolism using small molecule microdialysis have been used in liver transplantation [18–20]. Blood measurements of cytokines in liver transplant patients indicated sTNF-RII, sIL-2R and IL-10 levels were significantly elevated 3 days prior to or at the onset of acute steroid-resistant rejection, IL-8 and sTNF-RII was indicative of severe infection 3 days prior to onset and the pattern of cytokines during the course could discriminate between rejection and infection . It is reasonable to suggest that the diagnostic sensitivity and specificity will increase by measuring inflammation parameters directly in the liver tissue rather than in the systemic blood. In fact, preliminary data from our laboratory using liver microdialysis, according to the principles presented in this study, to evaluate post-transplant inflammation indicate that this is the case.
A number of studies indicate that activation of complement is an important pathogenetic factor in rejection of kidney and heart transplants [22, 23]. Systemic complement activation has been observed in patients undergoing liver transplantation [24–26] and portal capillary complement deposition has been demonstrated during liver rejection , indicating that complement may be of importance in liver rejection as well. Thus, complement anaphylatoxins should be included in the panel of inflammatory mediators measured locally in liver tissue after transplantation.
In conclusion, we have showed that the 100-kDa microdialysis catheter, in contrast to the 20-kDa filter, indeed can be used to measure a panel of clinically relevant cytokines and anaphylatoxins. Chemokines and anaphylatoxins (m.w. 7–11 kDa) showed very good recovery at 1.0 μl/min whereas cytokines with higher molecular weight (17–26 kDa) showed a lower recovery, except for IL-1β, which recovered remarkably good. It is not the amount of recovered substance per se that is most important, but the increase indicating inflammation. Thus, a 10- to 100-fold increase of cytokines, chemokines and complement factors is likely in an inflamed tissue. Therefore, even substances with a low recovery might be used as a clinical diagnostic. As microdialysis catheters is used in an increasing number of clinical conditions, our findings is likely to be applicable clinically for early diagnosis of tissue inflammation.
The study was financially supported by The Research Council of Rikshospitalet, The Norwegian Council on Cardiovascular Disease, The Norwegian Foundation for Health and Rehabilitation, Sigvald Bergesen d.y. og hustru Nanki's almennyttige stiftelse, Odd Fellow Foundation, Anders Jahre's Fund for the Promotion of Science, The Sonneborn Charitable Trust, The Family Blix Foundation and the NIH grant no. EB003968.