Clinical improvement of sepsis by extracorporeal centrifugal leukocyte apheresis in a porcine model

Although there has been substantial progress in the management of sepsis recently, it is still a great challenge for the health care system in high-income countries, with a reported annual incidence of 31.5 million affected cases and 5.3 million deaths, which is a much higher mortality rate than that of other diseases [1, 2]. This fact facilitates perpetual efforts in exploring novel methods for the treatment of sepsis. According to the current knowledge, a dysregulated host immune response to infection is considered the key underlying pathophysiological cause for life-threatening organ dysfunction presented in sepsis. [3, 4] Therefore, methods aimed at modulating immune function have become a research approach to seeking novel treatment options for sepsis. There have been numerous investigations of the role of extracorporeal blood purification therapies in the treatment of sepsis through the removal of proinflammatory substances such as cytokines and endotoxins from blood (reviewed in [5]). Such therapies may be proposed as adjuncts to standard treatments (such as antibiotics, management of organ dysfunction, and surgical treatment as required). As reported, these methods include high-volume hemofiltration (HVHF) [6], cascade hemofiltration [7], hemoperfusion or plasma perfusion using endotoxins/cytokine adsorption devices such as Toraymyxin^® (polymyxin B-immobilized fiber column), [8] CytoSorb^{® [9]}, the AN69ST membrane [10], and oXiris^®[11]. However, the effectiveness of these methods remains controversial, with most trials failing to prove their superiority over standard treatments [5]. The failure of methods targeting the removal of plasma toxic solutes should turn research interests to targeted intervention of immune cells, which are the major sources of cytokine production. Sepsis induces immune alterations in innate and adaptive cells [3], and marked elevation of white blood cells (WBCs) in circulation is a physical response to infection in the early stage of sepsis. Based on these considerations, it is worth investigating the effect of eliminating dysfunctional WBCs from the circulation during the early stage of sepsis using centrifugal cell separation techniques such as leukocytapheresis (LCAP). To address this issue, we performed a randomized controlled experimental study to explore the effect of LCAP on immune modulation and improvement of clinical conditions in a peritonitis-induced septic porcine model.

All animal experiments were performed in the Department of Comparative Medicine, Jinling Hospital, with a protocol previously approved by the Institutional Animal Care and Use Committee of Jinling Hospital, Nanjing University (grant number 2017NZGKJ-019). The treatments of animals were fully in compliance with the National Research Council’s 1996 Guide for the Care and Use of Laboratory Animals [12].

Two pigs died before the induction of sepsis, one because of a catheterization-associated accident and one because of an intubation-associated accident. Finally, 12 pigs were randomized equally into the control and treatment groups. All of them reached the criteria of sepsis [2] 12 h after induction, with blood LPS positivity, and survived to the end of the study, except one pig in the treatment group, which experienced resistant hypotension during LCAP and died 8 h after the end of treatment, with data included in the final analysis.

There were no significant differences in any measured data between the control group and treatment group at baseline (Table 1).

This preliminary study demonstrated that extracorporeal cellular apheresis therapy yields an immunomodulatory effect accompanied by clinical improvement of systemic hemodynamics and oxygenation status in a peritonitis-induced septic porcine model, which to our knowledge is the first study to use the extracorporeal centrifugation technique in sepsis therapy.

In this study, we established a sepsis model in domestic pigs by intraperitoneal injection of a suspended solution containing previously incubated autologous feces (1.0 g/kg). This model has an identical size, endotoxin sensitivity, tissue antigenicity and immune response as that in human beings and is suitable for investigating the effect of extracorporeal blood therapy in animal models [19]. Unlike other models using a single infusion of specific live bacteria [20, 21] or endotoxin [22] to induce a transient reaction other than a sustained process, we used a peritonitis-induced sepsis model to mimic polymicrobial and progressive characteristics of clinical situations, most importantly, with an appropriate sepsis severity to avoid rapid death within the first 6 to 12 h due to an overly serious condition [23, 24] and to guarantee a sufficient survival time for the completion of the study. As reported in some experiments, an observed time that is too short may not be enough to obtain a full view of sepsis at the end of the investigation [25, 26]. Notably, we administered the intervention 12 h after the induction of sepsis, which more closely conforms to the clinical real world, rather than before or immediately after the induction of sepsis, as reported in some studies [21, 25]. Since time is needed for the treatment to produce the therapeutic effect, we also spared a 24 h time window to observe the effect of the treatment.

Organ dysfunction is the clinical feature of sepsis, involving hypotension, acute respiratory distress syndrome, disseminated intravascular coagulation and so on [27], and the cumulative effect of organ dysfunction is the strongest predictor of mortality [28]. We observed that after 12 h of sepsis induction, animals showed a decrease in PaO2 and PaO2/FiO2 and an increase in lactate levels, and the trend was further obvious if not treated, which indicated lung injury. After 18 h of sepsis induction, blood pressure showed a downward trend, which is an indication of circulatory disorders. At the 12 h time point, there was a significant difference in the pH value between the control and treatment groups; however, in view of clinical practice, the pH difference of 0.1 units had no clinical significance, and the model of the two groups was consistent. Organ dysfunction is a complicated pathophysiologic process and is a result of the joint action of multiple responses to inflammation, one of which is immune dysregulation [29].

Immune dysfunction of both the native and adaptive immune systems is thought to be the essence of the pathophysiology of sepsis [3]. In the traditional view, sepsis is a process that shifts from initial proinflammatory reactions to subsequent anti-inflammatory reactions, manifested as an initial significant increase in peripheral blood leukocytes, especially neutrophils and monocytes, immediately following lymphopenia [30]. However, in this study, we observed a significant increase in neutrophils and monocytes and a decrease in lymphocytes (including B cells, Th cells and Tc cells) in the initial 12 h of sepsis, although without a significant difference. As in the profile of serum cytokines, a simultaneous proinflammatory and anti-inflammatory response was present. This conforms to the characteristics of the immune response in sepsis recently described by Xiao WZ et al. [31], in which they revealed a simultaneous rapid and sustained upregulation of genes involved in the innate immune and pro/anti-inflammatory response, as well as a downregulation of genes involved in adaptive immunity in patients suffering from severe trauma or burn injury.

Leukocytapheresis is a technique using extracorporeal centrifugation to separate blood constituents according to their density, and we used the MNC program in this study to remove blood MNCs from the circulation, with the expectation of attenuating the negative reactions of sepsis. As revealed by the results, various kinds of blood-formed elements were removed by LCAP, including B cells, CD4 + T cells, CD8 + T cells, DC cells, monocytes, platelets, and neutrophils, all of which are involved in sepsis-induced innate and adaptive immune derangements relevant to sepsis recovery and survival [3]. The cell selectivity (different cells were cleared in different amounts and ratios) may have been related to the cell density because LCAP stratified the blood components by centrifugal force according to the specific gravity of different cells. The effects of LCAP on sepsis as revealed by the study should be a wide-spectrum and bidirectional effect on sepsis-activated immune cells, manifested as a flattening of the change trends of various immune cells induced by sepsis, accompanied by improvement of two vital clinical conditions of sepsis, that is, hemodynamics and oxygenation.

An interesting phenomenon shown in our experiment is that a considerable quantity of neutrophils was removed by LCAP via the MCN program because the density difference between granulocytes and mononuclear cells will generally make them appear in different layers under density gradient centrifugation. This may be explained by the fact that abnormally dense neutrophils, also called low-density neutrophils (LDNs), may be present in the blood of patients with infections, cancer, pregnancy and autoimmune diseases [32]. However, thus far, it is difficult to distinguish LDNs from normal-density neutrophils (NDNs) by common markers such as CD11b, CD15, CD66b and CD33, although a difference in the expression level may exist [33]. We used CD11b as a marker [34] to show activated neutrophils accounting for 59.6% of the collection solution by LCAP and a lower level of them in circulation after LCAP, but with a significantly higher level of MPO, an indicator of the antimicrobial activity of neutrophils [35]. This result may imply that elimination of overactivated neutrophils from the circulation does not impair but rather enhances the antimicrobial capacity of the body. Another interesting phenomenon is that all the cells rebounded after 24 h of treatment, although only the rebounding of DC cells and monocytes was statistically significant. The apoptosis of DC cells engages in the immune suppression of sepsis [36]. Multiple reports have shown that prevention of sepsis-induced DC apoptosis or augmentation of DC function enhances sepsis survival [37, 38]. The long-term effect (at least 24 h) of LCAP increases the number of DC cells, and cell function needs to be explored in the future.

Other methods have also been reported for the extracorporeal removal of activated immune cells. One of them is a selective cytopheretic device (SCD) with regional citrate anticoagulation, which was reported to absorb activated leukocytes (mainly neutrophils) and improve the 60-day mortality when the postfilter ionized calcium (iCa) level was  ≤ 0.4 mmol/L [39]. Cellsorba is a device with polyethylenephtarate fibers that unselectively capture monocytes, granulocytes, and lymphocytes and is used for the treatment of inflammatory bowel disease in Japan and Europe, but its effects remain unclear [40]. These are all complicated extracorporeal circulation techniques, requiring a high dose of anticoagulants, central venous catheterization to afford a high blood flow rate, and a specialized team to handle the complex circuits, which obviously hinder their clinical applications regardless of the clinical effects. In contrast, extracorporeal centrifugal cellular apheresis, as a mature technique with low blood flow rates of 40–60 ml/min, a simple filter-free circuit and low clotting risk, may be worth further exploring for its application in the treatment of sepsis.

The obvious limitations of our study should be stated. First, the sepsis model of the study was not very severe, with modest pathological lesions in organs and a survival time beyond 48 h in order to guarantee the accomplishment of the study. Second, no antibiotics, as recommended by consensus [2], were used for the treatment of sepsis; third, only one treatment at a specific time point of sepsis was administered in the study, considering that sepsis is a sustained inflammatory response process, and one shot is obviously not enough. Finally, no hard endpoints were considered in the study, and consequently, further studies are needed to investigate the effect of LCAP on sepsis mortality.

In conclusion, this preliminary study using LCAP to treat septic pigs revealed an immunomodulatory effect with improvement of vital clinical conditions. However, further studies are still needed to verify its role in improving the mortality of sepsis.