Abstract
Objectives:
In many pediatric acute liver failure (PALF) cases, a diagnosis is not identified, and the etiology is indeterminate (IND-PALF). Our pilot study found dense CD8 T-cell infiltrates and increased T-cell clonality in liver specimens from IND-PALF patients. We aimed to validate these findings in a multicenter cohort with investigators blinded to diagnosis.
Methods:
PALF Study Group registry subjects with IND-PALF (n = 37) and known diagnoses (DX-PALF) (n = 18), ages 1 to 17 years, with archived liver tissue were included. Liver tissue slides were stained for T cells (CD8 and CD4), B cells (CD20), macrophages (CD163), perforin, and tissue resident-memory T cells (Trm, CD103), and scored as minimal, moderate, or dense. Lymphocytes wereisolatedfromfrozenlivertissueforT-cellreceptorbeta(TCRβ)sequencing.
Results:
Dense hepatic CD8 staining was found in significantly more IND-PALF (n = 29, 78%) compared with DX-PALF subjects (n = 5, 28%) (P = 0.001). IND-PALF subjects were more likely to have dense or moderate perforin (88% vs 50%, P = 0.03) and CD103 (82% vs 40%, P = 0.02) staining compared with DX-PALF subjects. TCRβ sequencing of 15 IND-PALF cases demonstrated increased clonal overlap compared with 6 DX-PALF cases (P = 0.002).
Conclusions:
Dense infiltration of effector Trm CD8 T cells characterizes liver tissue from IND-PALF subjects. Increased clonality suggests the T-cell expansion is antigen(s)-driven as opposed to a nonspecific inflammatory response. These findings support CD8 staining as a new biomarker of the activated CD8 T-cell PALF phenotype. Future studies are needed to characterize potential antigens, host risk factors, and inflammatory pathways with the goal of developing targeted therapies.
Keywords: acute hepatitis, aplastic anemia, fulminant hepatic failure, immune dysregulation
Pediatric acute liver failure (PALF) is a rare and severe condition in which a child with no known prior history of liver disease rapidly develops liver synthetic dysfunction. A variety of conditions can lead to PALF; however, in over 30% of cases no known cause is identified, and the diagnosis is labeled indeterminate (IND-PALF). Recent studies show that many patients with IND-PALF have evidence of T-cell activation, including high levels of inflammatory mediators, such as serum soluble interleukin-2 receptor (sIL-2R), and the liver injury is thought to be immune-mediated (1–5). IND-PALF patients are more likely to develop aplastic anemia (AA), a condition similarly driven by dysregulated immune cells (6–8). Results from our pilot study identified dense hepatic CD8 T-cell infiltration as a characteristic feature of IND-PALF (9). Liver biopsy specimens from IND-PALF patients were significantly more likely to have dense CD8+ immunohistochemical (IHC) staining (82%) compared with known diagnosis (DX-PALF) cases (7%) (P < 0.0001). Furthermore, the IND-PALF CD8 T cells appeared armed for effector functions (perforin-positive), displayed a tissue resident memory phenotype (CD103-positive), and exhibited increased clonal overlap by T-cell receptor beta (TCRβ) sequencing. Flow cytometry of intrahepatic T cells from 2 IND-PALF patients indicated that most T cells were double-negative for the memory markers, chemokine (C-C motif) receptor 7 (CCR7) and CD45RA, consistent with an effector memory phenotype. We propose that CD8 T cells are a biomarker of the IND-PALF “activated CD8 T-cell hepatitis” phenotype and may be used to identify and define this group. Here we aimed to validate these findings and evaluate the T-cell phenotype with investigators blinded to diagnosis in a large multicenter cohort of subjects from the National Institutes of Health-funded prospective observational Pediatric Acute Liver Failure Study Group (PALFSG) registry.
METHODS
This was an ancillary study of the PALFSG registry and utilized archived liver tissue specimens. Specimen analysis was conducted by investigators at Ann & Robert H. Lurie Children’s Hospital of Chicago (LCH) and Children’s Hospital of Philadelphia (CHOP).
Clinical and Laboratory Data
PALFSG is a multicenter, multinational study of children with acute liver failure enrolled December 1999 to 2014 (NIDDK; UO1-DK072146). The registry contains demographic, clinical, laboratory, and outcome data for 1144 participants ages 0 to 17 years. Criteria for enrollment included: no known evidence of chronic liver disease, biochemical evidence of severe liver injury occurring within 8 weeks of onset of illness, and hepatic-based coagulopathy (not corrected by vitamin K) defined as a prothrombin time (PT) at least 15 seconds or international normalized ratio (INR) ≥1.5 in the presence of hepatic encephalopathy (HE) or a PT ≥20 seconds or INR ≥2.0 with or without HE (10). Site-attending physician(s) directed the diagnostic evaluation, medical management, and assignment of final diagnoses consistent with standard of care. Demographic, clinical, and diagnostic data were recorded daily for up to 7 days after enrollment. The final diagnosis was registered as indeterminate in the absence of evidence for a specific diagnosis. The first outcome of death, liver transplantation (LT), or hospital discharge with native liver within 21 days following enrollment was recorded. Minimum and maximum laboratory values included values up to 7 days after enrollment or the first outcome event.
For this ancillary study, PALFSG subjects ages 1 to 17 years who had liver tissue specimens available in the biorepository were eligible for inclusion. Subjects with acute liver failure secondary to hemophagocytic lymphohistiocytosis (HLH), septic shock, or ischemia were excluded. Subjects with autoimmune hepatitis (AIH) or suspected autoimmune disease with positive auto-antibodies were excluded because of previous findings that liver tissue from these cases may exhibit a range of CD8-staining density (9). Subjects were included if they had at least 1 liver tissue paraffin block and/or at least 5 unstained formalin-fixed paraffin embedded (FFPE) liver tissue slides. Frozen liver tissue specimens were provided if available. De-identified samples were delivered to study investigators, who were blinded to final diagnosis. After all analyses were completed, the PALFSG clinical dataset was provided by the Data Coordinating Center for statistical analysis. Variables extracted included demographic data, laboratory values, degree of hepatic encephalopathy, worst values within the first 7 days after study enrollment, presence of multisystem organ failure (defined as need for ventilator, pressor, or dialysis support), treatment with corticosteroids before sample collection, development of AA, and 21-day outcomes. Subjects were divided for comparison based on final diagnosis: IND-PALF or DX-PALF. The study was approved by the LCH and CHOP Institutional Review Boards.
Immunohistochemistry and Liver Histopathology
IHC staining of liver tissue specimens was performed by the clinical pathology laboratories at LCH and CHOP as previously described (9). Cases with paraffin block tissue had slides stained with hematoxylin and eosin (H&E), Masson trichrome, and antibody for cytotoxic T cells (CD8), helper T cells (CD4), B cells (CD20), macrophages (CD163), perforin (a marker of lymphocyte cytolytic activity), and tissue resident-memory T cells (Trm, CD103). Cases with limited FFPE slides were stained with H&E and anti-CD8 antibody. Results were reviewed by investigators blinded to final diagnosis, including review by pediatric pathologists (H.M.A. and P.K.). CD8, CD4, CD20, CD163, perforin, and CD103-staining patterns were subjectively scored as dense, moderate, or minimal. Quantitative analysis of CD8 cells was performed using semi-automated microscopy to calculate mean CD8 cells per mm2. Images were captured by a 40× objective (400×) with Nikon 80i microscope with DS-Ri2 color camera and stitched with a Prior Proscan III (Rockland, MA) 8-slide state and digital encoder. Analysis was performed with NIS-Elements Advanced Imaging (version 5.1).
Immunofluorescence Staining
Indirect immunofluorescence assay (IFA) staining of liver tissue sections was performed using markers for CD8 and the T-cell memory markers CCR7 and CD45RA. Sections were deparaffinized, autofluorescence was reduced (MaxBlock, Dianova, Hamburg), and antigen retrieval was performed (Target Retrieval Solution, Dako-Agilent, Santa Clara, AC). Staining was performed at 4 °C with antibodies anti-CD8 at 1 : 50 dilution (Abcam, Cambridge, MA), anti-CCR7 at 1 : 100 dilution (Abcam, Cambridge, MA), and anti-CD45RA at 1:100 dilution (Novus Biologicals, Centennial, CO). Sections were prepared using a mounting reagent with 4′,6-diamidino-2-phenylindale (DAPI) for nuclear staining (Invitrogen, Carlsbad, CA). Imaging was performed on Zeiss LSM 880 Microscope (Carl Zeiss Microscopy, Thornwood, NY). Fluorescence staining for degree of CD8/CCR7 and CD8/CD45RA overlap was scored as dense, moderate, or minimal.
T-cell Receptor Beta Sequencing
Cases were selected for TCRβ sequencing based on having paraffin block tissue available, and CD8 staining density preliminary results with a goal to select a representative sample of dense, moderate, and minimal CD8-staining cases while remaining blinded to diagnosis. Genomic DNA was isolated from frozen liver sections banked in a −80 °C freezer. Sequences were generated from genomic DNA using primers situated near FR2 and Jb (BIOMED2 protocol). Samples were run in duplicate and libraries from each sample were pooled together. Libraries were created using the Illumina Nextera XT (2× 300 bp) paired end kit and run on an Illumina MiSeq instrument in the Human Immunology Core facility at the University of Pennsylvania. Data were analyzed using an in-house pipeline for assessing degree of clonality and diversity of repertoire.
Statistical Analysis
Continuous variables are described by median and inter-quartile range. Categorical variables are described by frequency and percentage. To test for differences between IND-PALF and DX-PALF, Fisher exact test or chi-square test was used for categorical variables and Wilcoxon-Mann-Whitney test was used for continuous variables. IHC-staining density scores were compared between groups using chi-square and Fisher exact tests. For quantitative analysis of CD8 cells/mm2, Mann-Whitney test was used to determine significant differences between groups. P values ≤0.05 were considered statistically significant. Statistical analyses were performed using SAS 9.4 software (SAS Institute, Cary, NC).
RESULTS
Subjects
Among 1144 subjects in the PALFSG registry, 716 met age and diagnosis inclusion criteria, 61 of whom had liver tissue specimens available. Six subjects were excluded as they had been subjects in our pilot study. The final cohort consisted of 55 subjects: 37 IND-PALF and 18 DX-PALF (Figure, Supplemental Digital Content 1, http://links.lww.com/MPG/B913). Most liver tissue samples were collected from the explant at the time of LT (n = 46, 84%). Known diagnoses included drug toxicity (n = 6; 2 Isoniazid exposure and 1 each Acetaminophen, Iron and Acetaminophen, Imatinib Mesylate, and l-asparaginase), viral hepatitis (n = 6; 1 each with Cytomegalovirus, Epstein-Barr virus, Herpes simplex virus, Influenza A virus, Influenza B virus, and Parvovirus infection), Wilson disease (n = 5), and veno-occlusive disease (n = 1). Subject characteristics, laboratory values, and outcomes are provided in Table 1. IND-PALF subjects were younger than DX-PALF subjects with median age 4.0 years (P = 0.0008). DX-PALF subjects were more likely to require ventilator (P = 0.02) or pressor (P = 0.004) support during the first 7 days. Eight IND-PALF and 2 DX-PALF subjects were treated with corticosteroid therapy for at least 24 hours before collection of the liver tissue sample. Median serum sIL-2R level was higher in IND-PALF (8481 IU/mL, n = 7) compared with DX-PALF (2432 IU/mL, n = 3) subjects (P = 0.03). One IND-PALF subject developed AA. Most IND-PALF (n = 36, 97%) and DX-PALF (n = 13, 72%) subjects had the 21-day outcome of LT.
TABLE 1.
Pediatric acute liver failure subject demographic, laboratory, and clinical characteristics
| Variable | Total Cases n = 55 |
IND-PALF n = 37 |
DX-PALF n = 18 |
P value |
|---|---|---|---|---|
| Median age, years (Q1–Q3) | 6.0 (3.0–13.0) | 4.0 (2.0–8.0) | 13.0 (7.0–16.0) | 0.0008 |
| Male gender, n (%) | 32 (58) | 24 (65) | 8 (44) | 0.2436 |
| Developed aplastic anemia, n (%) | 1 (2) | 1 (3) | 0 (0) | 0.4815 |
| Explant sample used for study, n (%) | 46 (84) | 35 (95) | 11 (61) | 0.0036 |
| Treated with corticosteroids before sample collection, n (%) | 10 (18) | 8 (22) | 2 (11) | |
| Median days (Q1–Q3) | 3 (1.0–5.8) | 1.5 (1.0–4.3) | 7.5 (6.8–8.3) | 0.1726 |
| At study enrolment | ||||
| HE level 0 to 1, n (%) | 43 (78) | 29 (78) | 14 (78) | 0.9596 |
| ALT (IU/L), n | 53 | 36 | 17 | |
| Median (Q1–Q3) | 1468 (606–3040) | 1786 (1043–3189) | 719 (223–2527) | 0.0610 |
| Total bilirubin (mg/dL), n | 52 | 35 | 17 | |
| Median (Q1–Q3) | 16.9 (13.6–21.1) | 17.2 (14.1–21.3) | 15.7 (8.2–19.1) | 0.1605 |
| INR, n | 54 | 36 | 18 | |
| Median (Q1–Q3) | 2.7 (2.3–4.0) | 2.6 (2.4–3.9) | 3.0 (2.2–4.8) | 0.7490 |
| Total WBC count (1000/μL), n | 55 | 37 | 18 | |
| Median (Q1–Q3) | 6.6 (5.2–10.3) | 6.2 (5.2–10.2) | 7.4 (5.2–10.3) | 0.6301 |
| Hemoglobin (g/dL), n | 55 | 37 | 18 | |
| Median (Q1–Q3) | 11.2 (10.1–12.6) | 11.2 (10.1–12.2) | 11.4 (10.2–13.0) | 0.6555 |
| Platelet count (1000/μL), n | 55 | 37 | 18 | |
| Median (Q1–Q3) | 198 (128–257) | 217 (156–352) | 126 (84–193) | 0.0008 |
| Worst values within first 7 days | ||||
| HE level 2 to 4, n (%) | 32 (58) | 22 (59) | 10 (56) | 0.7830 |
| Highest ALT (IU/L), n | 54 | 36 | 18 | |
| Median (Q1–Q3) | 1409 (604–3210) | 1802 (1043–3189) | 662 (105–3210) | 0.0824 |
| Highest total bilirubin (mg/dL), n | 51 | 33 | 18 | |
| Median (Q1–Q3) | 19.6 (16.0–25.4) | 19.6 (16.0–24.6) | 20.1 (14.5–30.5) | 0.7316 |
| Highest INR, n | 54 | 36 | 18 | |
| Median (Q1–Q3) | 4.3 (3.3–6.2) | 4.4 (3.3–6.7) | 3.9 (3.0–5.5) | 0.5114 |
| Lowest total WBC count (1000/μL), n | 55 | 37 | 18 | |
| Median (Q1–Q3) | 5.4 (4.1–7.7) | 5.2 (4.0–7.7) | 5.8 (4.3–7.0) | 1.000 |
| Lowest hemoglobin (g/dL), n | 55 | 37 | 18 | |
| Median (Q1–Q3) | 9.6 (8.4–11.0) | 9.9 (8.6–11.3) | 9.4 (7.8–10.8) | 0.2708 |
| Lowest platelet count (1000/μL), n | 55 | 37 | 18 | |
| Median (Q1–Q3) | 143 (99–217) | 158 (122–262) | 86 (66–154) | 0.0002 |
| Anytime during the first 7 days | ||||
| Ventilator, n (%) | 19 (35) | 9 (24) | 10 (56) | 0.0223 |
| Pressors, n (%) | 9 (16) | 2 (5) | 7 (39) | 0.0036 |
| Dialysis, n (%) | 14 (25) | 7 (19) | 7 (39) | 0.1855 |
| Soluble IL-2 receptor (IU/mL), n | 10 | 7 | 3 | |
| Median (Q1–Q3) | 6773 (3545–9434) | 8481 (5753–10814) | 2432 (943–2977) | 0.0291 |
| 21 Day outcomes, n (%) | ||||
| Liver transplantation | 49 (89) | 36 (97) | 13 (72) | |
| Alive with native liver | 4 (7) | 1 (3) | 3 (17) | |
| Death | 2 (4) | 0 | 2 (11) | 0.0165 |
ALT = alanine aminotransferase; DX-PALF = known diagnosis pediatric acute liver failure; HE = hepatic encephalopathy; IND-PALF = indeterminate pediatric acute liver failure; INR = international normalized ratio; WBC = white blood cell.
Immunohistochemical Staining Results
IND-PALF cases had a common pattern of liver injury characterized by diffuse portal and lobular inflammation, significant necrosis and collapse, periportal ductular reaction, and none or mild fibrosis. IHC scoring results for CD8, perforin, CD103, CD4, CD20, and CD163 are reported in Table 2. IND-PALF subjects were more likely to have dense CD8 staining (n = 33, 77%) compared with DX-PALF subjects (n = 5, 28%) (P = 0.001). Representative images of CD8-staining patterns in IND-PALF and DX-PALF cases are shown in Figure 1A. The CD8+ cells in IND-PALF cases with dense or moderate scores were distributed diffusely throughout the lobule and in the portal areas. Five of the 8 IND-PALF patients who were treated with corticosteroids before liver sample collection had dense CD8 staining, 2 had moderate staining, and 1 had minimal staining. Semi-automated quantification of CD8 cells agreed with the subjective scoring results: median CD8 cells/mm2 were higher for IND-PALF cases (2681, IQR 1931–3726) compared with DX-PALF cases (861, IQR 505–2009) (P = 0.002) (Figure, Supplemental Digital Content 2, http://links.lww.com/MPG/B914). Most of the 5 DX-PALF cases with dense CD8 staining had liver failure secondary to viral infection (Cytomegalovirus, Epstein-Barr virus, Influenza A virus, and Parvovirus), and one had drug-induced liver injury (Isoniazid). All 5 subjects with Wilson disease had minimal CD8 staining. See Table, Supplemental Digital Content 3, http://links.lww.com/MPG/B915, which lists subject diagnoses and corresponding IHC-staining scores. IND-PALF subjects were more likely to have dense or moderate perforin (P = 0.03) and CD103 (P = 0.02) staining compared with DX-PALF subjects. There were no significant differences in CD4, CD20, or CD163-staining patterns between groups. In all cases, CD4+ and CD20+ cells were concentrated in portal areas with minimal lobular staining, and most cases had moderate or minimal CD4 (93%) and CD20 (91%) staining. Dense CD163 staining was observed in 84% of cases, distributed in portal areas and throughout the parenchyma, and concentrated in areas of liver injury. See Figure, Supplemental Digital Content 4, http://links.lww.com/MPG/B916, which includes additional staining results from representative IND-PALF and DX-PALF cases.
TABLE 2.
Immunohistochemical staining patterns by pediatric acute liver failure diagnosis group
| Immunohistochemical stain | Total cases n = 55 |
IND-PALF n = 37 |
DX-PALF n = 18 |
P value |
|---|---|---|---|---|
| CD8-staining pattern, n (%) | 55 | 37 | 18 | 0.0011 |
| Dense | 34 (62) | 29 (78) | 5 (28) | |
| Moderate | 5 (9) | 2 (5) | 3 (17) | |
| Minimal | 16 (29) | 6 (16) | 10 (56) | |
| Perforin-staining pattern, n (%) | 44 | 34 | 10 | 0.0353 |
| Dense | 14 (32) | 13 (38) | 1 (10) | |
| Moderate | 21 (48) | 17 (50) | 4 (40) | |
| Minimal | 9 (20) | 4 (12) | 5 (50) | |
| CD103-staining pattern, n (%) | 44 | 34 | 10 | 0.0212 |
| Dense | 17 (39) | 16 (47) | 1 (10) | |
| Moderate | 15 (34) | 12 (35) | 3 (30) | |
| Minimal | 12 (27) | 6 (18) | 6 (60) | |
| CD4-staining pattern, n (%) | 44 | 34 | 10 | 0.3544 |
| Dense | 3 (7) | 3 (9) | 0 | |
| Moderate | 26 (59) | 18 (53) | 8 (80) | |
| Minimal | 15 (34) | 13 (38) | 2 (20) | |
| CD20-staining pattern, n (%) | 44 | 34 | 10 | 0.6653 |
| Dense | 4 (9) | 3 (9) | 1 (10) | |
| Moderate | 21 (48) | 15 (44) | 6 (60) | |
| Minimal | 19 (43) | 16 (47) | 3 (30) | |
| CD163-staining pattern, n (%) | 43 | 34 | 9 | 0.0948 |
| Dense | 37 (86) | 31 (91) | 6 (67) | |
| Moderate | 5 (12) | 2 (6) | 3 (33) | |
| Minimal | 1 (2) | 1 (3) | 0 |
DX-PALF = known diagnosis pediatric acute liver failure, IND-PALF = indeterminate pediatric acute liver failure.
FIGURE 1.
(A) Images at 20× magnification of representative CD8, perforin, and CD103 immunohistochemical staining patterns from an IND-PALF case with dense CD8 staining and a DX-PALF case with minimal CD8 staining. (B) Images at 40× magnification of representative immunofluorescence assay double-staining showing rare overlap between (1) CD8+ (red) and CD45RA+ (green) T-cells and (2) CD8+ (red) and CCR7+ (green) T-cells for an IND-PALF case with dense CD8 staining. Nuclear staining with DAPI is shown in blue. DAPI = 4′,6-diamidino-2-phenylindole; DX-PALF = known diagnosis pediatric acute liver failure; IND-PALF = indeterminate pediatric acute liver failure.
Immunofluorescence Assay Results
IFA staining of liver tissue slides was performed for 15 IND-PALF and 5 DX-PALF subjects. Overlap between CD8+ and CD45RA+ cells was rarely observed in IND-PALF cases, as expected for memory CD8 T cells (Fig. 1B). There was occasional overlap between CD8+ and CCR7+ cells, suggesting a small number of CD8 T cells acquired surface markers of central memory T cells (Fig. 1B). The majority of CD8 T cells were double-negative for CCR7 and CD45RA, categorizing them as effector memory T cells. There was no difference in degree of staining overlap between IND-PALF and DX-PALF groups.
T-cell Receptor Beta-sequencing Results
We performed next-generation sequencing of PCR products using common Vβ primers from 22 frozen liver specimens (15 IND-PALF and 7 DX-PALF) to assess clonality of T-cell receptor beta usage of intrahepatic lymphocytes. DX-PALF etiologies included Epstein-Barr virus infection, Influenza A virus infection, Acetaminophen toxicity, Iron and Acetaminophen toxicity, Isoniazid toxicity, and Wilson disease (n = 2). Both DX-PALF cases of viral infection had dense CD8 T-cell IHC staining, while the remaining 5 cases had moderate (n = 1) or minimal (n = 4) staining. Most IND-PALF cases (n = 11, 73%) had dense CD8 staining, with 2 moderate and 2 minimal. Samples were run in duplicate and showed no difference in quality between runs. We assessed clonality by measuring percentage of reads that composed the top 20 clones and measuring percent overlapping clones between duplicates of each sample, which is expected to be higher if specific clones dominate the specimen. The DX-PALF subject with Epstein-Barr virus had a highly clonal intrahepatic T-cell repertoire, as expected for a virus associated with direct hepatic injury, and was excluded from the analysis. By clonal overlap score, IND-PALF cases showed greater clonality compared with the 6 DX-PALF cases (Fig. 2A) (P = 0.003). There was a trend toward increased clonality by top 20 clones (Fig. 2B) (P = 0.08). No significant association was found between T-cell clonality and CD8 density.
FIGURE 2.
(A) Clonality as measured by overlap of T-cell receptor beta sequences between replicates for each sample: P = 0.003 IND-PALF versus DX-PALF, unpaired t test. (B) Clonality as measured by percentage of reads that compose the top 20 clones for all samples: P = 0.08 IND-PALF versus DX-PALF, unpaired t test. DX-PALF = known diagnosis pediatric acute liver failure; IND-PALF = indeterminate pediatric acute liver failure.
DISCUSSION
In a large multicenter cohort with investigators blinded to diagnosis, we confirm that children with IND-PALF are more likely to have dense hepatic CD8+CD103+Perforin+ T-cell staining compared with those with DX-PALF. This supports the use of CD8 T-cell infiltration as a new biomarker of IND-PALF (3,9,11). We further propose that PALF patients with dense CD8 staining on liver biopsy and no alternative identified etiology be classified as “activated CD8 T-cell hepatitis” with or without acute liver failure. We recommend considering liver biopsy when initial workup does not reveal a diagnosis and performing IHC staining for CD8 T cells in all cases in which liver biopsy is performed. Transjugular liver biopsy in PALF is well tolerated, not associated with an increased risk of major complications, and may help to guide diagnosis and management (12). Biopsy should be discussed at PALF presentation as disease trajectory and decision to list for LT are determined early in the disease course (13,14). Patients with dense CD8 staining are likely to have liver injury that is immune-mediated, whereas patients with minimal CD8 staining may have damage secondary to a metabolic or toxin-driven process. As previously discussed, it is important to note that activated CD8 T-cell hepatitis is not the only etiology associated with dense or moderate hepatic CD8 staining. Hepatic CD8 T-cell infiltration may be seen in cases of drug-induced hypersensitivity reactions, viral hepatitis, HLH and MAS (9,15–17). In addition, liver tissue from patients with autoimmune hepatitis can exhibit a range of minimal, moderate, or dense CD8 staining (9). Therefore, dense CD8 staining is not pathognomonic of “activated CD8 T-cell hepatitis”, but rather an additional diagnostic tool to be used in context with other available clinical information.
The pathophysiology of activated CD8 T-cell hepatitis remains unknown but we postulate that effector CD8 T cells are directly involved in propagating the liver injury. These patients tend to be younger than those with established diagnoses, which may reflect a unique window during which the developing immune system is susceptible to this type of injury (18). Evidence for T-cell activation in IND-PALF patients includes prominent staining for perforin, a cytotoxic granule released from T cells armed to carry out effector functions, and serum sIL-2R level ≥5000 IU/mL in all subjects tested, which has been associated with increased risk of death and LT (1). The majority of IND-PALF hepatic CD8 T cells were CCR7- and CD45RA-negative, consistent with an effector memory phenotype. The increase in clonal overlap in IND-PALF subjects suggests the T-cell expansion may be antigen(s)-driven as opposed to a nonspecific inflammatory response. This could represent an initial antigenic trigger that leads to a self-propagating, dysregulated immune response. In addition, IND-PALF patients are at increased risk of developing AA and often have early evidence of bone marrow dysfunction (19–21). These disorders may share pathophysiologic features as AA is similarly an immune-mediated disease involving activated CD8 T cells and high levels of inflammatory cytokines (6,8).
Our findings raise the question of whether outcomes for activated CD8 T-cell hepatitis patients could be improved with immunosuppressive therapy. There is a growing trend at pediatric liver transplant centers to empirically treat IND-PALF patients with a trial of corticosteroids. Our review of corticosteroid therapy for 28 IND-PALF patients at 4 large pediatric liver transplant centers, however, found that outcomes may not be different from historical controls (22). The benefit of immunosuppressive therapy for this disorder remains unknown and future research to define the immune pathways involved is needed to develop targeted treatments.
Limitations of this study include that subjects were selected based on availability of archived liver tissue, and biased towards those who underwent LT. Our cohort included more subjects with final diagnosis of IND-PALF (67%) compared with the overall PALFSG registry (44%) (13). The demographic characteristics and presentation lab values of IND-PALF subjects in our study appear similar to those reported from the larger PALFSG cohort (13), thus we feel that our findings are generalizable. In addition, the sample size for known diagnoses in our study was small, constituted of a heterogenous group of diagnoses, and did not include the entire spectrum of disease processes. Analyses were limited by the depth and breadth of information in the PALFSG data set. Final diagnoses were accepted as reported by site investigators. Individual subject diagnostic workup, treatment, and decision to list for LT was at the discretion of treating physicians. Some subjects were treated with corticosteroids before tissue specimens were obtained, which may have influenced the degree of lymphocyte infiltration.
CONCLUSIONS
These findings in a large multicenter cohort validate our prior results and support dense hepatic CD8 T-cell staining as a biomarker of activated CD8 T-cell hepatitis. Liver biopsy with CD8 staining should be adopted as part of the standard PALF diagnostic evaluation when a known cause is not readily identified. Additional studies are required to better characterize the unique phenotype of these CD8 T cells, which appear ready-to-perform effector functions (perforin+) and bear the surface markers of Trm T cells (CD103+). Future work will explore the clonal nature of the T-cell response, the role of possible antigen triggers, and the cytokines and inflammatory pathways involved.
Supplementary Material
Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.jpgn.org).
What Is Known
Many children with indeterminate pediatric acute liver failure have clinical, laboratory, and liver biopsy features of immune dysregulation.
Pilot data suggest that patients with indeterminate pediatric acute liver failure are more likely to have hepatic CD8 T-cell inflammation compared with those with other diagnoses.
What Is New
In a multicenter cohort, liver specimens from indeterminate pediatric acute liver failure patients are characterized by dense infiltration of effector memory CD8 T cells that are activated for cytolytic functions and exhibit increased clonality.
We propose the new classification “Activated CD8 T-cell Hepatitis” to describe this phenotype.
Acknowledgments:
The authors are grateful for the support from the National Institutes of Health (Edward Doo, MD, Director Liver Disease Research Program) and for assistance from members of the Data Coordinating Center at the University of Pittsburgh (directed by Steven H Belle, PhD, MScHyg).
Funding for this study was provided by the National Institutes for Health (NIH-NIDDK U01-DK072146).
Footnotes
The authors report no conflicts of interest.
REFERENCES
- 1.Bucuvalas J, Filipovich L, Yazigi N, et al. Immunophenotype predicts outcome in pediatric acute liver failure. J Pediatr Gastroenterol Nutr 2013;56:311–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.DiPaola F, Grimley M, Bucuvalas J. Pediatric acute liver failure and immune dysregulation. J Pediatr 2014;164:407–9. [DOI] [PubMed] [Google Scholar]
- 3.McKenzie RB, Berquist WE, Nadeau KC, et al. Novel protocol including liver biopsy to identify and treat CD8+ T-cell predominant acute hepatitis and liver failure. Pediatr Transplant 2014;18:503–9. [DOI] [PubMed] [Google Scholar]
- 4.Alonso EM, Horslen SP, Behrens EM, et al. Pediatric acute liver failure of undetermined cause: a research workshop. Hepatology 2017;65:1026–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zamora R, Vodovotz Y, Mi Q, et al. Data-driven modeling for precision medicine in pediatric acute liver failure. Mol Med 2017;22:821–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ren J, Hou XY, Ma SH, et al. Elevated expression of CX3C chemokine receptor 1 mediates recruitment of T cells into bone marrow of patients with acquired aplastic anaemia. J Intern Med 2014;276:512–24. [DOI] [PubMed] [Google Scholar]
- 7.Xing L, Liu C, Fu R, et al. CD8+HLA-DR+ T cells are increased in patients with severe aplastic anemia. Mol Med Rep 2014;10:1252–8. [DOI] [PubMed] [Google Scholar]
- 8.Zeng Y, Katsanis E. The complex pathophysiology of acquired aplastic anaemia. Clin Exp Immunol 2015;180:361–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chapin CA, Burn T, Meijome T, et al. Indeterminate pediatric acute liver failure is uniquely characterized by a CD103(+) CD8(+) T-cell infiltrate. Hepatology 2018;68:1087–100. [DOI] [PubMed] [Google Scholar]
- 10.Squires RH Jr, Shneider BL, Bucuvalas J, et al. Acute liver failure in children: the first 348 patients in the pediatric acute liver failure study group. J Pediatr 2006;148:652–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Patel KR, Bertuch A, Sasa GS, et al. Features of hepatitis in hepatitis-associated aplastic anemia: clinical and histopathologic study. J Pediatr Gastroenterol Nutr 2017;64:e7–12. [DOI] [PubMed] [Google Scholar]
- 12.Chapin CA, Mohammad S, Bass LM, et al. Liver biopsy can be safely performed in pediatric acute liver failure to aid in diagnosis and management. J Pediatr Gastroenterol Nutr 2018;67:441–5. [DOI] [PubMed] [Google Scholar]
- 13.Li R, Belle SH, Horslen S, et al. Clinical course among cases of acute liver failure of indeterminate diagnosis. J Pediatr 2016;171:163–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Squires JE, Rudnick DA, Hardison RM, et al. Liver transplant listing in pediatric acute liver failure: practices and participant characteristics. Hepatology 2018;68:2338–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Billiau AD, Roskams T, Van Damme-Lombaerts R, et al. Macrophage activation syndrome: characteristic findings on liver biopsy illustrating the key role of activated, IFN-gamma-producing lymphocytes and IL-6- and TNF-alpha-producing macrophages. Blood 2005;105:1648–51. [DOI] [PubMed] [Google Scholar]
- 16.Nakazawa A, Nakano N, Fukuda A, et al. Use of serial assessment of disease severity and liver biopsy for indication for liver transplantation in pediatric Epstein-Barr virus-induced fulminant hepatic failure. Liver Transpl 2015;21:362–8. [DOI] [PubMed] [Google Scholar]
- 17.Thimme R, Wieland S, Steiger C, et al. CD8(+) T cells mediate viral clearance and disease pathogenesis during acute hepatitis B virus infection. J Virol 2003;77:68–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Narkewicz MR, Horslen S, Hardison RM, et al. , Pediatric Acute Liver Failure Study Group. A learning collaborative approach increases specificity of diagnosis of acute liver failure in pediatric patients. Clin Gastroenterol Hepatol 2018;16:1801.e3–10.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Molina RA, Katzir L, Rhee C, et al. Early evidence of bone marrow dysfunction in children with indeterminate fulminant hepatic failure who ultimately develop aplastic anemia. Am J Transplant 2004;4:1656–61. [DOI] [PubMed] [Google Scholar]
- 20.Gonzalez-Casas R, Garcia-Buey L, Jones EA, et al. Systematic review: hepatitis-associated aplastic anaemia–a syndrome associated with abnormal immunological function. Aliment Pharmacol Ther 2009;30:436–43. [DOI] [PubMed] [Google Scholar]
- 21.Hadzic N, Height S, Ball S, et al. Evolution in the management of acute liver failure-associated aplastic anaemia in children: a single centre experience. J Hepatol 2008;48:68–73. [DOI] [PubMed] [Google Scholar]
- 22.Chapin CA, Horslen SP, Squires JE, et al. Corticosteroid therapy for indeterminate pediatric acute liver failure and aplastic anemia with acute hepatitis. J Pediatr 2019;208:23–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.