ABSTRACT
Background: Pediatric acute liver failure (PALF) cause is indeterminate in 30-50% of cases (iPALF). Children with acute severe hepatitis of unknown etiology (SH-u) may evolve to iPALF. Hemophagocytic Lymphohistiocytosis (HLH), a hyper-inflammatory state with marked T-cell activation (ACT) often involves the liver. SH-u evolving to iPALF could have hyperinflammatory immune signatures before fulfilling PALF criteria, and may overlap with HLH. We compared immune dysregulation signatures of SH-u, HLH, known PALF, and healthy controls (HC). Method(s): From 2019-2021, 15 patients (pts) hospitalized with SH-u and 7 pts with known PALF were prospectively enrolled. Age dependent standard diagnostic studies were performed. SH-u was defined as ALT>500, INR <2, and no hepatic encephalopathy. HLH enrollees met 2004 criteria. High dimension T-cell immunophenotyping, cytokine and chemokine profiling was done for SH-u, HLH (n=5), and HC (n=16) blood samples. T cell ACT was identified by co-expression of surface ACT markers HLA-DR and CD38. CD8 effector memory (EM) ACT of >9% identified pts with significant T cell ACT vs HC. Normally distributed data were compared by a two-tailed t-test or an ordinary One-Way Anova test with Turkey's multiple comparison test. Non-normally distributed data were compared by the Mann-Whitney test or KruskalWallis test with Dunn's multiple comparisons test. P values <0.05 were significant. Result(s): Pts ranged from 4days to 19years old. No age/sex differences existed between groups. One SH-u pt had a prior COVID infection. No pts met MIS-c criteria. Two SH-u pts ultimately met PALF criteria (INR>2). All pts with SH-u had higher CD8 EM T-cell ACT (mean +/- SEM = 43.7+/-6.3%;range 9.2 to 81.3%) than HC (2.9+/-0.5%) or PALF (4.0+/-0.9%);p<0.0001. T-cell ACT in SH-u was < HLH (90.3+/-2.7%;p<0.0001;Figure 1). T cell ACT was noted in the CD4 compartment, but CD8 compartment ACT was much > CD4. SH-u CD4/CD8 ratio was reduced compared to PALF pts. Though T cell ACT was greater in SH-u vs PALF, ferritin was lower in SH-u compared to PALF (1212+/-568 vs. 39517+/-32149;p<0.01) and very significantly < HLH (32415 +/-14845;p =0.002). Ferritin was <500 mg/L in 47% of SH-u pts. Despite overlapping T cell ACT with HLH, only 2 pts SH-u pts had cytopenia. IFN-gamma driven chemokines, CXCL9 and CXCL10, were elevated in SH-u, compared to HCs but similar to HLH pts. As a proof of concept, 2 patients with SH-u were treated with Emapalumab (an IFN-gamma blocking antibody) along with other immune modulators resulting in complete liver and immune function recovery. Conclusion(s): SH-u is associated with T-cell ACT. HLH and SH-u with T cell ACT have similarly increased IFN-gamma activity. Ferritin, however, may not be a reliable screening for SH-u patients with significant T cell ACT. If validated in larger studies of SH-u, the findings suggest a role for IFN-gamma blocking agents in a subgroup of SH-u prior to the development of PALF or possible bone marrow failure.
ABSTRACT
Background: 30-50% of pediatric acute liver failure (PALF) is of unknown cause, or indeterminate PALF (iPALF), which frequently results in transplantation. A subset of iPALF is characterized by T-cell activation. Some children with acute severe hepatitis of unknown etiology (SH-u) can evolve to iPALF. Hemophagocytic Lymphohistiocytosis (HLH) is a well-defined hyper-inflammatory condition characterized by marked T-cell activation and frequent severe liver involvement. We postulated SH-u evolving to iPALF has hyper-inflammatory immune signatures that are identifiable before fulfilling PALF criteria, and might overlap with those seen in HLH. We compared the immune dysregulation signatures of children with HLH to children with SH-u, PALF cases with known etiologies, and healthy pediatric controls (HC). Method(s): Between 2019-2021, we prospectively enrolled 14 patients hospitalized with SH-u and 7 patients with PALF of known etiologies. Age dependent standard of care diagnostic studies were performed. SH-u was defined as ALT> 500, INR < 2, and no hepatic encephalopathy. HLH enrollees fulfilled the 2004 diagnostic criteria. High dimension T-cell immunophenotyping, cytokine and chemokine profiling (71-plex) was done for SH-u, HLH (n=5), and HC (n= 16) peripheral blood samples. T cell activation was prospectively identified by co-expression of surface activation markers HLA-DR and CD38. Based on immune studies in HC, CD8 effector memory (EM) activation of >9% distinguished patients with significant T cell activation from HC. This cutoff of >9% was therefore used to identify SH-u patients with T cell activation. Normally distributed data were compared by either a two-tailed t-test or an ordinary One-Way Anova test with Turkey's multiple comparison test. Non-normally distributed data were compared by either the Mann-Whitney test or Kruskal-Wallis test with Dunn's multiple comparisons test. P Values < 0.05 were deemed significant. Result(s): Subjects ranged in age from 4 days to 19 years old. There were no age or sex differences between the groups. One SH-u patient had prior COVID infection, but no subject met MIS-c criteria. Two SH-u patients ultimately evolved to PALF criteria with INR> 2. All patients with SH-u had higher CD8 EM T-cell activation (mean +/- SEM = 43.7+/-6.3%;range 9.2 to 81.3;p<0.0001), which was significantly higher than HC (2.9+/-0.5%) and PALF of known etiology (4.0+/-0.9%) . However, the amplitude of T-cell activation was lower in the SH-u group relative to the HLH group (90.3+/-2.7%;p<0.0001), as shown in Figure 1. A similar trend in T cell activation was noted in the CD4 compartment. Overall, the activation in the CD8 compartment was much greater than in CD4. SH-u patients had a decreased CD4/CD8 ratio compared to the PALF group. Despite higher T cell activation in patients with SH-u compared to PALF, ferritin, often used to screen for hyper-inflammation, was lower in the SH-u group when compared to PALF group (1240+/-609 vs. 39517+/-32149;p<0.05) and very significantly lower than HLH (32415 +/- 14845;p =0.002). 50% of patients with SH-u etiology had ferritin < 500 mg/L. Cytopenia (hemoglobin < 9 g/dL, ANC < 1000/mL, platelets < 100,000/mL) is characteristic of patients with HLH. Despite overlapping T cell activation with HLH, the SH-u cohort had only 2 patients with this feature: one with thrombocytopenia and one with neutropenia. Supportive of this higher T cell activation, we noted chemokines driven by IFN-gamma, CXCL9 and CXCL10, to be elevated in SH-u compared to HCs and comparable to HLH patients. As a proof of concept, 1 patient with SH-u and thrombocytopenia underwent treatment with Emapalumab (an IFN-gamma blocking antibody) along with other immune modulators both with complete liver, immune, and platelet count recovery. Conclusion(s): Our cohort of SH-u was associated with significant T-cell activation. In addition, our patients with HLH and SH-u with T cell activation had similar increased IFN-gamma activity. Despite this T cell activation, ferritin values were significantly lower in SH-u compared to PALF without T cell activation. Ferritin may not be a reliable screening test to identify SH-u patients with significant T cell activation. If validated in a larger well-defined population of SH-u, the results may suggest a role for IFN-gamma blocking agents in a subgroup of SH-u prior to PALF or before bone marrow failure development.