Synergistic defects of different molecules in the cytotoxic pathway lead to clinical familial hemophagocytic lymphohistiocytosis

Familial hemophagocytic lymphohistiocytosis (FHL) is an autosomal recessive immune disorder with defective lymphocyte granule–mediated cytotoxicity.^1,2  Several genes have been implicated in FHL. FHL2, -3, -4, and -5 and Griscelli syndrome involve PRF1, UNC13D (MUNC13-4), STX11, STXBP2, and RAB27A, respectively.^3-6  All of these genes, except for PRF1, play a role in the degranulation of cytotoxic lymphocytes, which affects the precise steps of docking, priming, and fusion of the cytotoxic lymphocytes to the target cell.^7,8  Perforin is then released from granules to assist in the penetration and delivery of granzymes into the cytosol of the target cell, where the granzymes cleave key substrates to initiate apoptotic cell death.^3,7  An uncompromised and coordinated function of all of these molecules is essential for normal lymphocyte cytotoxic activity. Although it is well known that FHL is inherited in an autosomal recessive manner, symptomatic heterozygous mutation carriers have been presented in several reports.^9,10  In these cases, additional synergistic defects in the molecules of the cytotoxic pathway may contribute to the final development of hemophagocytic lymphohistiocytosis, which suggests a digenic inheritance of FHL.

In digenic inheritance, the combination of concurrent partial defects in 2 genes within the same pathway gives rise to a clinical phenotype, whereas a heterozygous state in either of the genes alone results in a less severe phenotype or none at all. Examples of a digenic mode of inheritance have been described for an array of disorders, including retinitis pigmentosa, holoprosencephaly, deafness, epidermolysis bullosa, Hirschsprung disease, insulin resistance, and polycystic kidney disease.^11,12  To investigate potential digenic inheritance for FHL, we reviewed patients with clinical FHL and identified those with heterozygous variants in 2 FHL-associated genes; we then examined the age of onset, defective degranulation, perforin levels, and natural killer (NK) cell function activity compared with patients with one (heterozygous) or 2 sequence variants in a specific gene. We present data suggestive of synergistic heterozygosity in FHL between 2 genes involved in cytotoxic lymphocyte degranulation.

This was a retrospective chart review, approved by the Cincinnati Children’s Hospital Medical Center institutional review board, which included 2701 patients referred by their physicians for genetic testing because of clinically suspected FHL. This cohort included affected individuals with homozygous or compound heterozygous mutations in a single gene, single heterozygous individuals with a single mutation in one gene, and double heterozygous (digenic) individuals with 2 heterozygous mutations in 2 separate genes. We analyzed clinical information, clinical immunology testing results, and/or genetic testing results that were available for these patients.^13-15 

Interpretation of sequence variants was based on the algorithm described by the Human Genome Variation Society (http://www.hgvs.org/mutnomen/recs.html) and included the use of predictions with Alamut 2.2.2 (Interactive Biosoftware, Rouen, France) for SIFT (Sorting Intolerant From Tolerant),¹⁶  PolyPhen-2¹⁷  and the Grantham Scale.¹⁸  Alamut also integrates databases for variant frequencies (dbSNP, 1000 Genomes Project, and Exome Variant Server) as well as splice site prediction algorithms (Human Splicing Finder) and the Human Gene Mutation Database for reported mutations (data not shown).

In this study, a total of 2701 patients with clinically suspected FHL underwent genetic and immunologic testing. Among these patients, 28 patients (P1 to P28) were found to be heterozygous for either a known mutation or a likely pathogenic variant in 2 distinct cytotoxic pathway genes (PRF1, MUNC13-4, STXBP2, STX11, and RAB27A). Genetic variants and clinical information from these patients were used to examine the possibility of digenic inheritance in FHL (Table 1). Fifteen patients were found to have heterozygous variants in PRF1 and MUNC13-4, 6 in PRF1 and STXBP2, 4 in MUNC13-4 and STXBP2, 1 in MUNC13-4 and STX11, 1 in STXBP2 and STX11, and 1 in STXBP2 and RAB27A.

Patients were separated into 2 groups: heterozygotes with variants in 2 genes involved in degranulation (MUNC13-4, STXBP2, STX11, and RAB27A [denoted as Deg/Deg]; 7 total) and heterozygotes with variants in PRF1 and one degranulation gene (denoted as PRF1/Deg; 21 total). Controls included a group of affected patients who were homozygous or compound heterozygous with PRF1 (120 patients identified), MUNC13-4 (72), or STXBP2 (33) mutations and a group of patients who were single heterozygous for single mutations within PRF1 (163), MUNC13-4 (119), or STXBP2 (29).

Deg/Deg patients had an earlier age of onset (≤24 months) for 71.4% (5 of 7) of patients (Figure 1A). In contrast, PRF1/Deg patients were observed to have an early age of onset (≤24 months) for only 14.3% (3 of 21) of patients (Figure 1A). The age of onset for Deg/Deg patients (≤24 months) was similar to that of patients with biallelic mutations in any one of these homozygous or compound heterozygous genes: 83.3% (100 of 120) of patients with PRF1, 83.3% (60 of 72) patients with MUNC13-4, and 69.7% (23 of 33) patients with STXBP2 (Figure 1A). In contrast, the single heterozygous patients were observed to have a later age of onset (Figure 1A). Thus, early age of onset in patients with 2 genetic defects within the degranulation pathway supports a potential digenic mode of inheritance for FHL.

Upon degranulation, NK cells express CD107a on their surface. Thus, decreased measure of CD107a expression is indicative of defective degranulation.^8,19  To determine which individuals displayed defective degranulation, CD107a expression data from immunologic testing was analyzed. CD107a expression, perforin expression, and NK function results were available for only a small portion of these patients. From the limited number of patients, we found 75% (3 of 4) of patients with Deg/Deg double heterozygotes had defective CD107a degranulation. This finding is similar to the defective degranulation seen in individuals with biallelic MUNC13-4 (91.7%; 11 of 12) and STXBP2 (92.3%; 12 of 13) mutations. In contrast, PRF1/Deg double heterozygotes had normal degranulation (0%; 0 of 8) (Figure 1C). As expected, defective NK function was observed in all patient groups (data not shown).

Patients with defects in PRF1 had decreased perforin levels. Numerically, 100% (41 of 41) of PRF1-affected, 63.6% (42 of 66) of PRF1 heterozygotes, and 72.7% (8 of 11) of PRF1/Deg patients had decreased perforin expression (Figure 1C). Interestingly, only 20% (1 of 5) of Deg/Deg double heterozygotes, 16% (4 of 25) of MUNC13-4–affected, 30.8% (4 of 13) of STXBP2-affected, 33.3% (14 of 42) of MUNC13-4 heterozygotes, and 40% (6 of 15) of STXBP2 heterozygotes had low perforin in any of these categories (Figure 1C). Patients without PRF1 mutations had increased or normal levels of perforin (data not shown).

Synergistic function was hypothesized for the genes in the degranulation pathway because of the known highly precise and coordinated functions that lead to the secretion of granzyme B to target cells. Findings from this retrospective study suggest that heterozygous defects in 2 degranulation pathway genes have a synergistic deleterious effect and thus imply that a digenic mode of inheritance can lead to FHL. Patients with mutations in 2 degranulation genes in this study had early childhood presentation of FHL. Those patients also displayed defective degranulation of lymphocytes as expected, unlike PRF1/Deg patients. This is the first retrospective study that suggests a potential digenic inheritance in FHL. However, it is worth noting several limitations of this study. Detailed clinical data were not available for several patients with suspected FLH. Data for analysis of NK function, CD107a degranulation, and perforin expression were also limited. In addition, parental and/or sibling samples were not available in most families for segregation studies. Ideally, future prospective studies should include a thorough clinical summary, outcome results, and data on genetic and functional studies performed on the proband and family members.

The authors thank the staff of the Cincinnati Children’s Hospital Medical Center (a Federation of Clinical Immunology Societies Center of Excellence) and the diagnostic immunology and molecular genetics laboratories that performed the clinical tests for this study.

Contribution: K.Z. and A.H.F. designed the research, analyzed and interpreted data, and wrote the manuscript; S.C., C.A.V., and H.C. analyzed and interpreted data and wrote the manuscript; and D.K., J.A.J., and A.H. performed the research and collected and analyzed the data.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Kejian Zhang, Division of Human Genetics, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, ML 4006, Cincinnati, OH 45229; e-mail: kejian.zhang@cchmc.org.