An Inherited Disorder of Lymphocyte Apoptosis: The Autoimmune Lymphoproliferative Syndrome FREE

Dr. Stephen Straus (Laboratory of Clinical Investigation, National Institute of Allergy and Infectious Diseases [NIAID], National Institutes of Health [NIH], Bethesda, Maryland): The autoimmune lymphoproliferative syndrome (ALPS) is a recently defined illness that arises in early childhood and can have fatal complications (1 - 8). It is associated with prominent nonmalignant lymphadenopathy, hepatosplenomegaly, and autoimmune manifestations. Underlying ALPS are heritable mutations in genes that regulate lymphocyte survival by triggering programmed death of lymphocytes, or apoptosis. More important than the mere description of ALPS, however, are the novel insights that this description affords into the mechanisms that regulate lymphocyte homeostasis and contribute to autoimmunity.

Selected features of ALPS have been recognized for decades, but the full-blown syndrome is rare and has only recently been appreciated. Several authors have described families with significant adenopathy and splenomegaly (9 - 13) and other families with hemolytic anemia, thrombocytopenia, or neutropenia in association with circulating autoantibodies (14 - 20).

Splenomegaly is a feature of autoimmune diseases (such as the Felty syndrome), and moderate lymphadenopathy is seen in up to 70% of patients with lupus (21). The constellation of lymphadenopathy, splenomegaly, and autoimmune cytopenia, however, was described by Canale and Smith in 1967 (22). Weisdorf and Krivit (23) and others (24 - 25) noted that similar patients had decreased proportions or function of lymphocyte subsets. Contemporary evaluation of some of these older cases, including a family followed at the NIH, has shown them to be cases of ALPS (4, 13; Straus SE. Unpublished data).

Clues to the nature of some of these familial diseases emerged in the early 1990s, when we realized that affected patients resembled mice with the MRL/lpr phenotype; these mice exhibit progressive lymphoproliferation and autoantibody-mediated renal and vascular disease (1,26 - 27). Moreover, the usually rare subset of T cells that show neither the CD4 nor the CD8 co-receptors (CD3⁺ CD4 ⁻CD8 ⁻), or “double-negative” cells, circulates in increased numbers both in these mice and in our patients.

In 1989, Trauth and coworkers (28) reported that a protein called Apo-1 triggered apoptosis of lymphocytes. In 1992, Watanabe-Fukunaga and colleagues (29) found that lpr mice failed to express that same antigen, which they called Fas and, later, CD95. In 1994, the APT1 gene encoding the human homologue of murine Fas was cloned (30). Shortly thereafter, Rieux-Laucat and associates in France (2) and Fisher and colleagues at the NIH (3) demonstrated defective apoptosis and specific Fas mutations in eight children with ALPS. In the ensuing years, more than 40 similar patients have been described (4 - 8).

The autoimmune lymphoproliferative syndrome (sometimes called the Canale-Smith syndrome (4)] represents a failure of apoptotic mechanisms that help maintain normal lymphocyte homeostasis, with a consequent accumulation of lymphoid mass and persistence of autoreactive cells. It is operationally defined as chronic, nonmalignant lymphoproliferation in patients with 1) an elevated percentage [>1%] of double-negative T cells and 2) defective lymphocyte apoptosis that produces a characteristic, if not pathognomic, pathologic picture on microscopic section of the lymph node or spleen (31). Autoimmunity is evident at some point in almost all affected patients. Most cases of ALPS are associated with specific Fas mutations, and yet-undefined mutations in other apoptosis genes are thought to underlie cases in patients with normal Fas. One remarkable case from our clinic shows many features of ALPS (Figure 1).

At age 18 months, NIH ALPS patient 2 was noted to have adenopathy, splenomegaly, and anemia (1,3,6). She underwent splenectomy at age 2 years for refractory anemia. Biopsies of splenic tissue, liver, lymph node, and bone marrow were nondiagnostic. Clinical photographs (Figure 1 A) and computed tomographic scans (Figures 1F and 1G) confirmed the presence of persistent, massive enlargement of all lymph node chains and organomegaly. Laboratory studies showed continuing anemia, polyclonal gammopathy, and positive results on a direct Coombs test. Autoimmune manifestations have included glomerulonephritis at age 2 years, idiopathic thrombocytopenic purpura (ITP) at age 9 years, and autoimmune biliary disease since age 10 years. Lymphocyte phenotyping showed absolute increases in B-cell and T-cell counts with a polyclonal expansion of T cells, 25% of which were double negative. The lymph nodes showed lymphoid hyperplasia and plasmacytosis (Figure 1 B), and more than 50% of T cells were double negative (Figures 1C, 1D, and 1E). In vitro studies of peripheral blood mononuclear cells showed defective Fas-mediated apoptosis, and genomic sequencing identified a point mutation in the APT1 gene encoding Fas.

Dr. Michael C. Sneller (Laboratory of Immunoregulation, NIAID, NIH): Between 1990 and 1997, the NIH evaluated 45 patients under approved protocols for unexplained chronic lymphadenopathy, splenomegaly, or both. Among these patients, ALPS was subsequently diagnosed in 26 (12 male and 14 female). The remaining patients had some features of ALPS or had entirely different conditions. The Table shows the clinical and salient laboratory features of these 26 patients, of whom 23 had documented Fas mutations. The clinical and laboratory features of ALPS did not differ between patients who had mutations and those who did not. Clinical data on 9 of these patients are reported elsewhere (1,6).

All 26 patients initially presented because of lymphadenopathy or splenomegaly at a median age of 11.5 months (range, 1 month to 9 years). Most had both splenomegaly and lymphadenopathy. Splenomegaly was frequently of massive proportions, and hepatomegaly was also common (Figure 1 G). Sixteen patients underwent splenectomy, most often for severe hypersplenism.

Lymphadenopathy was massive and distorted normal anatomic landmarks in some patients (Figure 1 A). Enlargement of abdominal and thoracic lymph nodes was frequently seen on computed tomography (Figure 1 F). Regardless of its extent, lymphadenopathy persisted for 2 or more years in almost all patients.

Histopathologic analyses of lymph nodes from patients with ALPS show architectural preservation, florid reactive follicular hyperplasia, and marked paracortical expansion with immunoblasts and plasma cells (1,6,31) (Figure 1 B). The paracortical expansion may be extensive enough to suggest a diagnosis of immunoblastic lymphoma, with many cells expressing the Ki-67 antigen indicative of active proliferation (32). However, the tissues show no chromosomal abnormalities or evidence of clonality (31). Increased numbers of double-negative T cells are also seen in the paracortical region of lymph node tissue (Figures 1C, 1D, and 1E). This combination of follicular hyperplasia and paracortical expansion by a mixed polyclonal infiltrate containing double-negative T cells differentiates ALPS from other benign and malignant lymphoproliferative lesions.

Circulating autoantibodies, overt autoimmune disease, or both were found in 23 of the 26 patients. Potentially pathogenic autoantibodies were detected in 22 patients (Table) but were not always associated with disease. For example, the direct Coombs test detected antibodies to erythrocytes in 19 patients, but 6 of these patients had no evidence of hemolysis. At least one autoimmune disease was documented in 17 patients (Table) and was evident in 4 of the 17 at the time of initial presentation with lymphoid hyperplasia. In the remaining 13 patients, autoimmune disease developed 6 months to 17 years later, suggesting that the proportion of patients with autoimmune disease increases over time.

The most common autoimmune diseases were hemolytic anemia and ITP (Table). Nine patients had at least one episode of hemolysis during which hemoglobin levels decreased to less than 4.4 mmol/L (7 mg/dL). In seven of eight patients with ITP, platelet counts decreased to less than 20 × 10⁹ cells/L. Five patients also had hemolytic anemia, either concomitantly with active ITP or as isolated episodes with normal platelet counts.

Neutropenia (absolute neutrophil count <1.0 × 10⁶ cells/L) in six patients seemed to result from autoimmune mechanisms because it developed after splenectomy and in the setting of normal myeloid cellularity on bone marrow examination.

Several nonhematologic autoimmune diseases also occurred in this group of patients. One patient developed the Guillain-Barré syndrome. Of two patients with glomerulonephritis, one later developed ITP and autoimmune biliary disease (Figure 1 A). Thus, in ALPS, multiple autoimmune diseases involving different organ systems may occur over time in a single patient.

No patient developed opportunistic infections or other clinical evidence of immunodeficiency. However, five patients who had splenectomy developed Streptococcus pneumoniae septicemia, usually despite appropriate antibiotic and vaccine prophylaxis.

The most prominent abnormalities seen with routine immunologic testing of these patients were T-cell and B-cell lymphocytosis, increased numbers of circulating double-negative T lymphocytes, and polyclonal hypergammaglobulinemia. The magnitude of these abnormalities varied (Table). Patients with the most severe lymphoid hyperplasia also had the most pronounced lymphocytosis, the largest numbers of double-negative T lymphocytes, and the highest serum immunoglobulin levels.

Episodes of autoimmune hemolytic anemia and ITP usually required treatment with high doses of glucocorticosteroids (prednisone, ≥ 1 mg/kg of body weight per day). High-dose intravenous immunoglobulin was ineffective or produced only transient benefit with respect to ITP. Recurrent ITP was interrupted in several patients with the use of monthly pulses of dexamethasone, 4 to 6 mg/kg per day for 5 days. Recombinant granulocyte colony-stimulating factor produced sustained increases in neutrophil counts in two patients with ALPS who had neutropenia and recurrent infections.

On the basis of studies in lpr mice in which interleukin-2 and cyclosporine were effective, patient 2 received empirical trials of these agents (33 - 34). Moreover, this patient and many other patients received prednisone or intravenous immunoglobulin for the treatment of autoimmune disease. These therapies afforded, at best, a transient decrease in the degree of lymphadenopathy (1,6).

Because most of our patients have been followed for relatively short periods, we do not yet know the ultimate prognosis associated with ALPS. So far, none of our 26 patients has died. However, the major determinants of illness seem to be the severity of autoimmune disease and the occurrence of sepsis after splenectomy.

Some information on the long-term prognosis associated with ALPS can be gained by considering three of our patients who are now adults (6,8). Their medical records show manifestations of ALPS in early childhood. One patient has persistent lymphadenopathy but is otherwise well at age 26 years. The second patient had splenectomy at age 2 years (13). Her lymphadenopathy resolved during adolescence, but she developed ITP at age 18 years and autoimmune neutropenia at age 32 years. Thus, patients with ALPS have a lifelong risk for autoimmune disease. The third patient developed Hodgkin disease at age 26 years. Of note, this patient's brother also had a Fas mutation, had clinical features of ALPS, and died of lymphoma (6). These cases suggest that Fas mutations that impair antigen-induced lymphocyte apoptosis may be associated with an increased risk for lymphoma.

Dr. Michael Lenardo (Laboratory of Immunology, NIAID, NIH): The striking phenotype of lymphocyte expansion in patients with ALPS led to the hypothesis that its underlying cause is a defect in apoptosis, which is a crucial regulatory mechanism in the immune system (35). Apoptosis is a process by which the body rids itself of unneeded or harmful cells. It is essential during embryonic development and is thought to be conserved in all multicellular organisms for the homeostatic control of adult tissues. The term apoptosis has been used to describe a series of morphologic changes in a cell that is programmed to die; these changes include condensation and cleavage of the nuclear chromatin; blebbing of the cell membrane; and, finally, fragmentation of the cell into small, membrane-bound “apoptotic bodies” (36). The cellular remnants show phosphatidylserine and other ligands that interact with receptors on the surface of phagocytic cells, allowing rapid engulfment and removal of the dead cell. Hence, apoptosis eliminates cells with little of the inflammatory reaction that typically accompanies traumatic or necrotic cell death.

In the immune system, apoptosis has important functions (37). During lymphocyte differentiation in the thymus and bone marrow, it eliminates cells that are strongly autoreactive. In the peripheral immune system, it counteracts the potential accumulation of mature lymphocytes.

When a resting T cell is stimulated, it proliferates in response to interleukin-2 or other cytokines. This renders it sensitive to apoptosis. The fate of the proliferating T cell depends on the amount of antigen encountered. A large amount of antigen induces the expression of death cytokines, including tumor necrosis factor (TNF), and members of the family of similar proteins and their receptors (38) (Figure 2). The chief member of this family involved in lymphocyte apoptosis is the Fas receptor and its ligand (FasL; CD95L) (39). The induction of apoptosis in dividing T cells prevents the overexpansion of activated cells in the face of persistent or recurrent antigen exposure. A different form of apoptosis occurs after the antigen has been eliminated and the immune response is waning. Production of proliferative cytokines, such as interleukin-2, decreases. Dividing cells that are no longer needed for a protective response undergo lymphokine-withdrawal apoptosis. Somehow, a small number of the activated cells escape apoptosis and persist as “memory” T cells.

The Fas and TNF death receptors and homologous proteins are the principal cell surface receptors that induce apoptosis in mammals, thereby regulating the deletion of autoreactive thymocytes and principal mature peripheral T cells (29,39 - 40). The TNF receptors (TNFRs) and Fas share significant amino acid sequences and structural homology. The extracellular portions of these proteins are important for ligand binding. Their cytoplasmic portions contain “death domains” that bind cytoplasmic signaling proteins essential for inducing apoptosis (41) (Figure 2). After their respective ligands bind to them, three molecules of Fas or TNFRs assemble into complexes. The cytoplasmic portions of the Fas trimer attract a cytoplasmic adapter protein known as FADD (Fas-associated death domain) or MORT1 (mediator of receptor-induced toxicity), and the cytoplasmic tails of a TNFR bind adapter molecules known as TRADD (TNFR-associated death domain) and RIP (receptor-interacting protein) (41 - 42). These latter proteins also harbor death domains, which enable the further recruitment of the apoptosis-inducing protease caspase 8 (42 - 48). Caspases, including caspase 8, are cysteine-containing proteinases that cleave themselves at aspartate residues to generate highly active, mature enzymes that proteolytically process further proteases in a cascade. The last members of this cascade damage mitochondria and degrade chromosomal DNA.

To investigate the possibility that ALPS could be the result of altered lymphocyte apoptosis, it was necessary to test the efficiency with which apoptosis could be induced in our patients' lymphocytes. Patient cells were stimulated with interleukin-2 and then treated with an antibody directed at Fas. The percentage of cells killed in our patients was significantly lower than the percentage of cells killed in healthy controls (Figure 3).

These data strongly suggested that defects in apoptosis were associated with ALPS and that the failure of apoptosis might result from abnormal function of the Fas receptor. It remained, then, to determine whether the gene encoding Fas is mutated in patients with ALPS.

Dr. Jennifer M. Puck (Immunologic Genetics Branch, National Human Genome Research Institute, NIH): Figure 4 shows the organization of the APT1 gene. It contains nine exons and has a genomic span of about 25 kb on human chromosome 10q23 (30). Exons 2 through 5 encode the extracellular domain of Fas. A transmembrane region is encoded in exon 6. Exon 9 encodes the death domain referred to in the preceding section. The Fas protein is expressed in the heart and liver as well as in B and T lymphocytes. Particularly high amounts are found in activated T cells.

To date, more than 40 patients with ALPS have been identified. APT1 gene mutations have been found in more than 21 families, some of which have more than one affected member (2 - 8). The localization and type of mutations in patients with ALPS are depicted in Figure 4. These mutations are not polymorphisms because they have not been detected in the screening of hundreds of unrelated healthy persons. All of the mutations to date are single-nucleotide changes except for the homozygous deletion found in a severely affected child of related French parents (2,5) (Figure 4).

By far, the most common form of ALPS is that associated with heterozygous Fas mutations; ALPS is inherited in an autosomal dominant fashion. The region of the APT1 gene most often mutated is the death domain. These mutations are predicted to result in early termination (frameshift insertions and deletions; amino acid changes to stop codons) or in single amino-acid substitutions (missense mutations) that disrupt the three-dimensional structure of the death domain (44). In vitro transfection experiments proved that Fas proteins bearing death-domain mutations inhibit the function of normal Fas proteins. This explained why heterozygous Fas mutations behaved in an autosomal dominant manner (3). Because Fas molecules form trimeric complexes to signal apoptosis, we predicted that a single abnormal Fas molecule in these complexes would impair apoptosis. When we expressed both mutated and normal Fas proteins in cells, we found that mutant proteins strongly inhibited the transmission of a death signal by the normal, wild-type protein.

Studies of family members of patients with ALPS identified additional persons carrying the same APT1 mutations. Some were completely free of the symptoms and signs of ALPS, whereas others met all of the diagnostic criteria for ALPS (2 - 4,7 - 8). Still other mutation-bearing relatives had some but not all of the features of ALPS, such as episodes of significant adenopathy or enlarged spleen but no autoimmune disease. For example, a kindred of 14 persons in four generations was followed for up to 25 years (8). Some members had histories of splenomegaly or other features of ALPS, but 11 of the 14 were shown to have a heterozygous mutation in the death domain of Fas (8) (Figure 5). All of the mutation-bearing members of this kindred had impaired lymphocyte apoptosis. Among those with clinical features of ALPS, lymphoproliferation tended to abate by adulthood. The only premature death in the family was due to sepsis after splenectomy in one boy; however, the occurrence of non-Hodgkin lymphoma in one person at age 50 years, as well as this person's development of ITP at age 23 years and hemolytic anemia at age 54 years, show the risk for malignancy and the unpredictable appearance of autoimmune complications.

Evaluation of families such as this one indicate that additional genetic or environmental factors must interact with defects in apoptosis to engender overt lymphoproliferation and autoimmunity. Such modifying factors might include other proteins in the Fas apoptosis pathway or related pathways (Figure 2).

In contrast to patients with ALPS and Fas mutations, whom we designate as having ALPS type Ia, both children and adults have been found to have autoimmune disease, chronic lymphoproliferation, and defective apoptosis but no Fas mutation. Some of these persons have expanded numbers of double-negative T cells, as well, and therefore have been designated as having ALPS type Ib if they have a Fas ligand mutation or ALPS type II if no genetic defect has been uncovered (6,49 - 50). The early studies of ALPS suggest that defects in apoptosis pathway proteins other than Fas might contribute not only to some cases of ALPS but also to a wider variety of syndromes with autoimmunity, lymphoproliferation, and possibly malignancy.

In all, many gene products regulating lymphocyte signaling networks have been implicated in lymphoproliferation and autoimmunity; many of these could be exacerbated by an apoptosis impairment and therefore may also be candidates for second mutations in ALPS. Other factors, such as excessive synthesis of the cytokine interleukin-10, may also play roles in the genesis of autoimmunity.

Dr. Warren Strober (Mucosal Immunity Section, Laboratory of Clinical Investigation, NIAID, NIH): Patients with ALPS have a B- and T-cell lymphocytosis and increased levels of double-negative T cells (1 - 8) (Table 1). Many of the T cells bear the MHC class II (DR antigens) surface marker, reflecting previous activation and indicating abnormal persistence. Moreover, cells in tissues often express an acute activation marker, CD69; this suggests that abnormal lymphocyte proliferation also contributes to the peripheral and tissue lymphocyte expansion seen in ALPS, not just to the prolonged survival of resting T cells.

To help understand what might be driving the lymphocyte proliferation, we measured plasma levels of cytokines in patients with ALPS (51). Circulating levels of the regulatory cytokines interleukin-4, interferon-γ, and interleukin-2 and levels of the inflammatory cytokines TNF-α, interleukin-1b, and interleukin-6 were generally normal. In contrast, plasma interleukin-10 values were strikingly elevated in almost all patients with ALPS: The median value was 183 pg/mL in 24 patients with ALPS but was 0 pg/mL in healthy controls and in 121 patients with lymphoma, vasculitis, or other autoimmune or lymphoproliferative disorders. Minimal increases in interleukin-10 levels were described in another study of Hodgkin disease (52). Of interest, the median level of interleukin-10 in family members with Fas mutations was only 7 pg/mL. Given that they all manifested defective apoptosis, it is evident that a high interleukin-10 level is related more to the presence of ALPS than to abnormal apoptosis.

We next examined the capacity of lymphocytes in patients with ALPS to release various cytokines after stimulation in vitro (51). Previously activated DR⁺ CD4⁺ T cells from six patients produced 20-fold more interleukin-4 and 10-fold more interleukin-5 than did DR⁺ T cells from controls. In contrast, the same cells produced 2-fold to 4-fold less interferon-γ and interleukin-2 than did DR⁺ T cells from controls. The double-negative T cells from our patients proved to be relatively unresponsive to in vitro stimulation; this parallels the results of studies of double-negative cells in MLR/lpr mice (53). The origin of these cells in patients with ALPS is uncertain, although evidence in the mice indicates that these cells derive from CD8⁺ T cells that have lost the ability to express CD8 (54). These studies suggest that the double-negative T cells are not likely to be participating in immune responses in the development of the autoimmune problems in our patients.

Finally, we examined the ability of monocytes and macrophages from patients with ALPS to release various cytokines after stimulation in vitro. These cells from patients with ALPS produced, on average, 5-fold more interleukin-10 and substantially less interleukin-12 than control cells did. Because the monocyte-macrophage population is vastly expanded in ALPS as a result of the general lymphoid expansion, these studies strongly suggest that this population is a source of high interleukin-10 levels in patients with ALPS.

In asking how these combined data might shed light on the development of autoimmunity in ALPS, we should recall two basic principles. First, as Figure 6 shows, T-cell immune response patterns are classified as either T helper 1 (Th1) (driven by interleukin-12 and leading to interferon-γ and TNF-α release) or T helper 2 (Th2) (driven by interleukin-4 and leading to release of interleukin-4, interleukin-5, and interleukin-6). Second, many immunologic disorders are associated with imbalanced Th1 or Th2 responses. Thus, in patients with multiple sclerosis or Crohn disease and mice with similar diseases, greatly upregulated Th1 activity leads to inflammation (55 - 58). Conversely, in patients with systemic lupus erythematosus or allergic asthma, unbridled Th2 responses underlie inflammation and autoimmunity (59 - 63). Thus, the finding that the T cells of patients with ALPS show a pronounced Th2 profile begins to explain the humoral autoimmunity seen in ALPS; the Th2 cytokines provide the T-cell drive to B cells needed to produce autoantibodies.

The occurrence of a Th2 bias in ALPS is probably due to both elevated interleukin-10 production and decreased interleukin-12 production. Interleukin-10 inhibits interleukin-12 production and thus drives T-cell differentiation in the direction of Th2 responses (57,64). It has been claimed that increased interleukin-10 production accounts for the reduced interleukin-12 levels seen in patients with systemic lupus erythematosus (60,65). In ALPS, defective Fas-mediated apoptosis cannot be the sole basis of the interleukin-10 elevation because healthy relatives with Fas mutations do not have it. It is more likely that elevated interleukin-10 production is an independent abnormality.

Increased interleukin-10 production may have additional effects on B-cell and T-cell survival. Interleukin-10 induces the antiapoptotic protein Bcl-2 in both B cells and T cells and thus retards their death (66 - 68). Therefore, overexpression of interleukin-10 could lead to the persistence of autoreactive cell clones and even malignant cell clones. Together, these contribute to the clinical features and complications of ALPS.

The autoimmune lymphoproliferative syndrome is a newly recognized disorder caused by inherited defects in the mechanisms that induce lymphocytes to die. It permits the accumulation of lymphocytes in and the expansion of lymphoid organs. Some of the surviving lymphocytes react with a patient's own antigens, precipitating autoimmunity. This syndrome is the first known human disorder of these cell death pathways and is a novel genetic cause of autoimmune disease.

Apoptosis: A tightly regulated program of molecular and biochemical processes that leads to cell death.

Exon: A coding section of a gene retained in the gene's messenger RNA that is translated into a segment of protein.

Fas: A lymphocyte surface receptor protein that initiates apoptosis.

Intron: An intervening RNA segment that is spliced out of a gene's messenger RNA.

T helper 1: CD4⁺ cells that produce interleukin-2 and interferon-γ, which promote cellular immunity.

T helper 2: CD4⁺ cells that produce interleukin-4, which promotes humoral immunity.

Transcription: The making of an RNA molecule by using the information encoded in DNA.