Both sepsis and HLH often present with overlapping clinical symptoms and hyperinflammation. Hence, early distinction between the two syndromes and timely intervention remains a challenge, more so when the underlying cause of HLH is infection. Since, both the conditions show an obvious dysregulation of the immune system, in this study we compared the immune profiles; i.e., soluble and cellular markers of activation of immune system and cytokines/chemokines reflecting the state of inflammation in children with sepsis and those which develop sepsis-associated HLH. We report differences in the T cell subsets, their activation states and the cytokines associated with T cell activity in sepsis children with and without HLH.
In this study, SHLH children presented with similar T cell frequencies but lower CD4/CD8 ratios in comparison to the sepsis children without HLH. Similar findings were also reported in another study in children with EBV-associated HLH when compared with sepsis children14. Increased apoptosis of T cells, preferentially in activated CD4 is a common phenomenon noticed in sepsis patients15–18. Whether a marked reduction in the CD4/CD8 ratio in SHLH patients is either governed by preferential apoptosis of CD4 T cells or excessive proliferation of activated CD8 T cells which is often reported in HLH patients; or both, needs further investigation. A number of studies have clearly demonstrated expansion of activated T cell population as a characteristic feature in HLH patients; more evidently in CD8 T cells10,12,14 while in sepsis, suppressive T cell phenotypes or function was noticed17,19–21. More recently, the diagnostic utility of CD38hi HLADR + activated T cells has also been reported in distinguishing active HLH from early sepsis in children10 (Chaturvedi et al). In keeping with these studies, we also report higher frequencies of activated (CD38 + HLADR+) CD4 T cells in children with SHLH compared to those with sepsis without HLH. In contrast to these studies12,14, however, the levels of activated CD8 T cells (though elevated in SHLH), did not reach statistical significance in our study. Besides HLADR and CD38, PD1 expression also upregulates upon activation of T cells, and at the same time its expression is often linked with exhaustion in T cells22,23. Our data showed that similar to the activated T cells, the PD1 expressing CD4 T cells were also remarkably elevated in SHLH children than those with sepsis alone which is in line with previous studies24; however, we find that these differences were not statistically significant for CD8 PD1 + cells. Whether these cells exhibit suppressive immune function may not be concluded from our study, however a recent study have shown partial impairment of cytotoxic function in PD-1 + T cells in HLH patients24. Also, the lower CD4:CD8 ratio observed in SHLH may indicate reduced CD4 help to CD8 cells, which in turn can result in impaired CD8 T cell proliferation and effector functions25–27. Although previous studies on HLH patients have emphasized more on activated CD8 T cells, our data in contrast suggest CD38 + HLADR + or PD1 expressing CD4 T cells to be more useful in differentiating SHLH from sepsis cases at an early stage or a may be few weeks later as well; as seen in a couple of follow up samples in this report. These differences in results may be explained by the differences in our study design or patient characteristics, differences in sample size & statistical power, heterogeneity in etiological factors, differences in genetic composition, as well as technical variability in laboratory methodologies across study cohorts. Also, activation of CD8 T cells is common in viral infections and other intracellular pathogens which can further amplifies with the severity of disease28,29. As a significant proportion of sepsis children in this study were severe with septic shock/multiorgan dysfunction, it is likely that the background activation of CD8 T cells is already very high in some patients and hence making it less suitable for early diagnosis of SHLH.
Multiple studies have shown cytokines/chemokines related to both T cell (IFNg, CXCL9, CXCL10, IL-10 etc.) and monocyte/macrophage activation (IL-18, IL-6, IL-1b, IL-1RA, TNFa etc) to be elevated in HLH in comparison to those with sepsis or other inflammatory conditions11,30–33. However, in this study where we compared the cytokine profile between sepsis and sepsis-associated HLH, only a set of cytokines which are largely involved in T cell function/regulation were found to be different. Thus, in corroboration with our T cell data, we also report higher levels of IFNg (secreted by effector Th1 cells), IFNg induced chemokines MIG/CXCL9 (monokine induced by interferon-gamma, a chemoattractant for CXCR3 expressing effector T cells), IL-2Ra (which is considered as a surrogate for T cell activation) and IL-10 (an anti-inflammatory cytokine produced by regulatory T cells, Th-2 cells or activated macrophages; regulating T cell response to viral infections) (ref Ejrnaes et al). These findings are also supported by a recent study31 where the both gene expression and plasma cytokine profile indicated upregulation of IFNg inducible genes in pediatric HLH patients in comparison to those with severe sepsis/SIRS. Though IFN-g levels were not seen to be elevated in all the sepsis-HLH samples in this study, yet higher MIG levels reflects an active and increased IFNg signaling predominantly in SHLH children. Thus, as reported previously, a shorter half-life may explain poor detection of IFNg in some of the SHLH patients34. When the same set of cytokines were evaluated for their diagnostic potential to distinguish between the two syndromes, IL-10, MIG and IL2Ra also demonstrated a good diagnostic capacity with > 80% specificity. A few reports have also suggested altered IL6/IL-10 or IL-6/CXCL9 ratio in HLH patients with lower IL-6 and higher IL-10 or CXCL9 in differential diagnosis of HLH from sepsis. We observed a similar trend in primary HLH children, however between SHLH and sepsis IL-6 levels were similar. In keeping with this a recent study also indicated higher levels of IL-6 in infection associated HLH (other than EBV) when compared to other forms of HLH11.
Previous studies have reported a positive association of IL10 levels with the poor outcomes in both pediatric and adult HLH patients33–36. We also observed, higher levels of IL10 in SHLH children who died than those who survived. Interestingly, none of the IFN-g regulated cytokines were found to be further elevated in children who died. Instead, those who died showed a higher concentration of IL6, IL8, ferritin, IL10 in comparison to those who survived. These immune signatures were previously shown to be associated with poor survival and high severity/SOFA score in sepsis children37. Additionally, higher absolute total leukocyte and neutrophil counts indicate that the hyperinflammation in those who died, likely to be contributed by enhanced monocyte/macrophage and neutrophils activation. Whereas, a poor percentage of activated T cells among the non-survivors may be a result of either enhanced apoptosis of activated T cells or their enhanced recruitment to the sites of infection or inflammation38. Similar trend of lower proportion of activated T cells in a small number of poor-outcome HLH patients was also noticed in another report by Chaturvedi et al, hence this needs further validation and investigation on a larger sample size.
Though the study’s limitation is its smaller sample size as it was exploratory, yet the findings on increased T cell activation along with elevated IFN-g and it’s inducible proteins are similar to the other two similar studies published recently10,31. In addition, this study is unique, as the pediatric HLH patients in our study are a clean infection associated HLH children and those with primary HLH were analyzed separately. Another, important factor is the study design where we assessed the immune profile at an early timepoint within a couple of days of diagnosing sepsis, thus the results indicate early markers to differentiate SHLH from sepsis. Thus, together our data show, reduced CD4 to CD8 ratio, increased T cell activity with significantly higher proportions of activated CD4 T cells and IFNg, MIG and IL-10 levels in SHLH than those with sepsis. Out of these, CD4 to CD8 ratios, percentage of HLADR + CD38+/PD1 + CD4 cells and IL-10 plasma concentrations; together may provide an advantage in early diagnosis of SHLH from those with sepsis without HLH.