Intestinal T lymphocyte homing is associated with gastric emptying and epithelial barrier function in critically ill: a prospective observational study

Enteral nutrition via the gastrointestinal tract is accepted to improve health outcomes in critically ill patients [1]. Gastrointestinal motility disorders, in particular impaired gastric emptying frequently impede enteral nutrient delivery. The advantages of enteral nutrition include a decrease in infectious complications and mucosal barrier protection against bacterial translocation [2, 3]. The highest incidence of delayed gastric emptying in the critically ill has been reported in patients with sepsis [4].

The intestine participates in the inflammatory response locally with cytokine production and leukocyte recruitment. To perform their immunological functions, T lymphocytes must exit the blood and enter into different tissues in the body. Recruitment of leukocytes during an inflammatory response involves the expression and activation of adhesion molecules. CD4+ T lymphocytes responding to the antigen in intestinal lymph nodes start to express high levels of the integrin α4β7 and the chemokine receptor CCR9. The intestinal ligand of CCR9 is CCL25 and it is mainly expressed in the mucosal epithelium of the small intestine. Thus, co-expression of α4β7 and CCR9 delineates a subset of T lymphocytes preferentially migrating to the lamina propria of the small intestine [5].

While we have previously identified association between CD4 + α4β7 + CCR9+ T lymphocytes and gastric emptying [6], we hypothesized that CD4 + α4β7 + CCR9+ T lymphocytes are associated with enhanced pro-inflammatory cytokine secretion, intestinal barrier disruption and delayed gastric emptying in the critically ill. While assessment of gastrointestinal permeability by the frequently applied sugar absorption tests are not reliable in ICU patients [7] we choose to quantify plasma zonulin, ileal bile acid binding protein (ileal bile-acid binding protein) and intestinal fatty acid binding protein (intestinal fatty acid binding protein). Scintigraphy is considered the gold standard for evaluation of gastric emptying. Unfortunately, all currently available tests for the assessment of gastrointestinal motility, including scintigraphy, have limited value in the ICU setting. Major obstacles include lengthy procedures, the need to transfer patients, lack of accuracy in multi-organ dysfunction, costs and availability [8]. In the present prospective observational study, we have chosen the paracetamol absorption test for the assessment of gastric emptying as it is a feasible and reliable technique in the ICU setting [9, 10].

The inability of enteral feeding due to delayed gastric emptying frequently necessitates total parenteral nutrition. Total parenteral nutrition itself leads to mucosal inflammatory response resulting in a loss of epithelial barrier function [2], further contributing to intestinal motility disturbances. Increased intestinal permeability predicts the development of multiple organ failure and early enteral nutrition is a key factor for preservation of gut barrier function reducing the risk of organ failure [11]. Furthermore, enterocyte damage has been reported to be associated with ICU mortality [12].

While in most studies feed-intolerant patients or patients receiving parenteral nutrition have been examined, we investigated immune activation with epithelial barrier dysfunction and gastric emptying at admission prior to any enteral or parenteral nutritional support. Zonulin increases intestinal permeability in the epithelia of the small intestine and increased plasma zonulin has been reported in sepsis [13]. I-FABP is particularly expressed in the jejunum and if released into plasma indicates enterocyte destruction [14, 15], while I-BABP is exclusively present in the ileum [16]. We identified increased Zonulin, I-BABP and I-FABP levels in patients with delayed gastric emptying, indicating marked differences in upper gastrointestinal mucosal barrier integrity at the onset of critical illness.

In our study patients with delayed gastric emptying had significantly higher CD4 + α4β7 + CCR9+ T lymphocytes compared to patients with normal gastric emptying at ICU admission. Indeed CCR9 is required for effector CD4+ T lymphocyte entry into small intestinal lamina propria [17]. Under inflammatory conditions intra-epithelial lymphocytes alter barrier function, consequently enhancing intestinal permeability and further enhancing chemokine production [18]. In particular, IFABP and IBAP were inversely correlated with gastric emptying, indicating small bowel immune activation with intestinal barrier alterations associated with gastric emptying. There is further evidence linking intestinal immune activation with CD4 + α4β7 + CCR9+ T lymphocytes and gastrointestinal motility. The detection of CCR9-expressing T lymphocytes in the bloodstream of patients with post-operative ileus has been reported to potentially reflect dysmotility [19].

Apart from increased CD4 + α4β7 + CCR9+ T lymphocytes, we also identified increased serum cytokine levels and pro-inflammatory cytokines produced in vitro by PBMC. Intestinal dysmotility is known to be accompanied by increased expression of pro-inflammatory cytokines such as TNF-α and IL-1β, followed by infiltration of lymphocytes [20]. Based on animal models, TNF-α and IL-1β are known to directly inhibit gastric emptying [21] and decrease small bowel muscle contractility [22]. The increased release of pro-inflammatory cytokines is followed by infiltration of T lymphocytes and macrophages [23]. Macrophages activated by CD4+ T lymphocytes in the intestine initiate the release of nitric oxide, directly inhibiting smooth muscle cell function [24] consequently contributing to intestinal dysmotility. In a previous study, both cytokines were enhanced in patients with idiopathic gastroparesis even in the absence of an infectious stimulus [6]. Thus, our data clearly suggest an association between increased pro-inflammatory cytokines with delayed gastric emptying in the critically ill. Furthermore, TNF-α is a critical factor for epithelial barrier dysfunction [25]. A significant increase in TNF-α in bowel segments is linked to an impaired epithelial barrier [3] within these segments. We therefore believe that our findings of enhanced epithelial barrier function reflected by increased zonulin, I-BABP and I-FABP in subjects with delayed gastric emptying depict a putative causal link between initial immune activation and gastric dysmotility.

Utilizing the paracetamol absorption test, we found 70% of patients, who were admitted to the ICU and needed immediate invasive mechanical ventilation, had impaired gastric emptying. This is in line with published data showing delayed gastric emptying in 42–81% of ICU patients [26, 27]. While the diagnosis on admission exerts a modest impact on gastric emptying in the critically ill, it appears to be most prominent in sepsis. Indeed, the majority of patients included in the present study were admitted to ICU with an infectious complication leading to sepsis or acute respiratory failure. Against the background of published data demonstrating an association between gastric emptying and severity of illness [4] it could be argued that the observed differences in immune activation and epithelial barrier function parameters between patients with normal and delayed gastric emptying are simply a function of baseline severity of illness. However, we did not observe differences in the APACHE II, MODS or SOFA in our study population.

Other potentially confounding factors include sedation and glycemic control. It has been demonstrated that patients receiving morphine and midazolam are more likely to have slow gastric emptying than those receiving propofol [28]. The majority of the patients included in our study received midazolame and sufentanil, without significant differences between the groups. However, an effect of sedation on the overall rate of delayed gastric emptying cannot be excluded. There is abundant literature on glycemic control in the critically ill and blood glucose concentration could be one of a number of factors affecting gastric emptying [29]. Our patients had comparable blood glucose levels and equal insulin doses to achieve glycemic control. Therefore, a profound effect on gastric emptying cannot be assumed.

It also needs to be acknowledged that the majority of patients in our study had an underlying malignancy. There are some data indicating T lymphocyte dysfunction in patients with hematologic malignancies, potentially differing from the general population of critically ill patients. Different degrees of T lymphocyte dysfunctionality, including impaired proliferation, altered cytokine secretion and altered expression of surface molecules, have been described in various hematologic malignancies. In particular T cell lymphoma, chronic myeloid leukemia and chronic lymphoid leukemia are known to be associated with T lymphocyte dysfunctionality. The invasive spreading ability of leukemic cells may depend on the expression of surface molecules including α4β7-Integrin.

There are limited data on neoplastic cells in T cell lymphoma, not achieving remission, expressing α4β7-Integrin when infiltrating bone marrow [30]. Increased expression of α4β7-Integrin is also observed in hematopoietic stem cell recipients with acute gastrointestinal graft versus host disease [31]. Therefore, it needs to be emphasized that we did not include patients with relapse, disease progression or acute graft versus host disease in this study. In addition, the majority of patients with an underlying hematological malignancy had acute myeloid leukemia, which is not associated with a T lymphocyte proliferation defect or impairment of cytokine secretion [32]. Furthermore patients in remission or patients who have undergone intensive chemotherapy may have reduced CD4+ T lymphocytes but without impairment of proliferation [33]. Only when T lymphocyte counts are very low are the remaining T lymphocytes functionally impaired. Thus, it is important to emphasize that patients in our study had normal white blood cell counts and T lymphocyte populations and therefore do not represent severely neutropenic (<1.500 neutrophils/μl) patients with cancer, whose inclusion could potentially confound the immune activation results. However, the observed ICU mortality of 50% reflects current mortality rates of critically ill patients with cancer [34] and is therefore certainly due to the high percentage of patients in our study who had malignancies.

To our knowledge this study is the first to identify a link between gastric dysmotility and increased small bowel-directed immune activation and mucosal barrier impairment in ICU patients at the onset of critical illness, in the absence of any nutritional support. Our result suggests that initial immune activation leads to increased release of pro-inflammatory cytokines followed by infiltration of lymphocytes. While cytokines are able to directly impair gastrointestinal motility, lymphocytes may contribute to delayed gastric emptying by activating macrophages releasing nitric oxide as a motility inhibitor. Cytokines also impair intestinal permeability, leading to amplification of the inflammatory response. Thus, this cascading inflammation may drive organ failure including impaired intestinal barrier and motility as an underlying cause of further intestinal dysfunction, regardless of the overall severity of the initial illness.