New Insights into Common Variable Immunodeficiency

Dr. Michael C. Sneller (Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases [NIAID], National Institutes of Health [NIH], Bethesda, Maryland): Common variable immunodeficiency (CVI) is a heterogenous syndrome characterized by hypogammaglobulinemia and recurrent bacterial infections. Patients with this syndrome also have an increased incidence of autoimmune disease and malignancy. This rare syndrome has an estimated prevalence ranging from 1:50 000 to 1:200 000 [1, 2]. Unlike X-linked agammaglobulinemia, CVI affects men and women equally. The disorder may occur at any age, but in most patients the onset is in the second or third decade of life [3-5]. This is in contrast to X-linked agammaglobulinemia, in which recurrent infections develop in the first 2 years of life [6].

Clinically, patients with CVI have recurrent bacterial infections of the respiratory tract such as sinusitis, otitis media, bronchitis, and pneumonia. The most common etiologic agents are encapsulated bacteria such as Streptococcus pneumoniae and Haemophilus influenzae. These infections are a direct result of the deficiency in antibody production (specifically immunoglobulin G) that is the hallmark of this syndrome. Immunoglobulin G is the major heat-stable opsonin of extracellular fluid, and the occurrence of repeated infections with encapsulated bacteria in IgG-deficient patients shows the importance of this molecule in host defense against these pathogens. If patients with CVI are not diagnosed (and hence not treated), recurrent pulmonary infections can lead to irreversible chronic lung disease with bronchiectasis. In such patients the spectrum of bacterial pathogens is broader and includes Staphylococcus aureus and Pseudomonas aeruginosa. Septicemia and recurrent infections of the skin, urinary tract, joints, or central nervous system also occur in patients with CVI but are less frequent [3-5].

In rare instances, patients with CVI can become infected with mycobacteria, Pneumocystis carinii, and various fungi. The occurrence of these opportunistic infections may be due to clinically important abnormalities of cell-mediated as well as humoral immunity. Patients with CVI tolerate viral infections normally but some exceptions exist. Herpes zoster (shingles) occurs in up to one fifth of patients with CVI [3]. This represents an abnormally high incidence of viral reactivation in such young patients (ages 20 to 30 years). Frequent and severe recurrences of herpes simplex infections have also been reported in a small number of patients with CVI [7], as have severe cytomegalovirus infections [8].

An unusual syndrome of severe enteroviral infection is seen in patients with primary antibody-deficiency syndromes [9]. This type of infection is most commonly associated with X-linked agammaglobulinemia, but several well-documented cases have occurred in patients with CVI. The most common pathogens are echoviruses, especially echovirus type 11. This syndrome usually occurs as a meningoencephalitis that has a chronic progressive and usually fatal course. Some patients may also have extraneurologic manifestations of disseminated enteroviral infection, including fever, a dermatomyositis-like syndrome, edema, rashes, and hepatitis. In a few cases these manifestations have preceded the development of neurologic symptoms. This syndrome is best diagnosed by examination of the cerebrospinal fluid. Patients almost always have elevated cerebrospinal fluid protein and a lymphocytic pleocytosis; frequently, their cerebrospinal fluid cultures are positive for enterovirus.

Patients with CVI have various infectious and noninfectious gastrointestinal disorders. The protozoan Giardia lamblia is a common cause of infectious diarrhea in these patients. In normal hosts, Giardia usually causes a self-limited, mild diarrheal syndrome lasting for days to weeks. In contrast, patients with CVI frequently develop chronic diarrhea with clinically significant malabsorption and weight loss that may last for months to years. Diagnosis is made by the presence of cysts or trophozoites in the stool, but because this method is frequently unreliable, diagnosis may require duodenal or jejunal aspirate or duodenal biopsy. Treatment with quinacrine or metronidazole is usually successful, but some patients with CVI require multiple courses of single or combination drug therapy to eradicate Giardia.

Patients with CVI are also at increased risk for infection caused by bacterial enteric pathogens such as Salmonella, Shigella, and Campylobacter species. In addition to such infections, we have seen several patients with CVI in whom diarrhea was apparently caused by a fastidious gram-negative bacterium (rod) known as dysgonic fermenter-3 [10]. These patients, who had either acute or chronic diarrhea, negative microbiologic work-up for other enteric pathogens, and dysgonic fermenter-3 isolated from their stool, responded to appropriate antibiotic treatment with resolution of their diarrhea and eradication of this organism from their stool. Thus, in selected patients, stool should be cultured for dysgonic fermenter-3 using cefoperazone-vancomycin agar plates incubated at 35 C [10].

Neoplasms of the gastrointestinal tract (specifically, adenocarcinoma of the stomach and intestinal lymphomas) appear to occur with increased frequency in patients with CVI. In a report from Great Britain [11], 14 of 220 patients with CVI developed malignancies, of which 7 were gastric carcinomas. This represents an incidence of gastric carcinoma 50 times greater than that of the general population. In a U.S. study of 46 patients with CVI, 4 patients had gastric carcinoma [5]. This increased risk should be kept in mind when evaluating upper gastrointestinal symptoms in such patients. Patients with CVI are also at increased risk for developing lymphomas (see below), some of which may primarily involve the bowel.

Malabsorption, associated with weight loss and diarrhea, is another gastrointestinal manifestation of CVI. Laboratory findings of malabsorption are usually present and may include hypoalbuminemia, hypocalcemia (due to malabsorption of vitamin D), and decreased levels of vitamin A and carotene. Fecal fat determinations and measurements of D-xylose absorption are usually abnormal. Small-bowel biopsy shows flattening of the villi along with a lymphocytic infiltration in the lamina propria. Histologically, this condition resembles celiac sprue (gluten-sensitive enteropathy) except that plasma cells, a prominent feature of the lymphoid infiltrate in celiac sprue, are absent from the lymphoid infiltrate in patients with CVI. The cause of this syndrome is unknown. Despite the resemblance to celiac disease, gluten-free diets do not correct the malabsorption in most patients, nor has empiric treatment with broad-spectrum antibiotics, corticosteroids, or oral immunoglobulin proven effective. Treatment is supportive with vitamin and mineral replacement as indicated.

Inflammatory bowel disease (Crohn disease or ulcerative colitis) appears to occur with increased frequency in patients with CVI [4]; however, before a patient with CVI is diagnosed as having inflammatory bowel disease, a thorough search for an infectious process must be done. Treatment is the same as in nonimmunocompromised patients, except that therapy with immunosuppressive agents should be avoided if possible.

Approximately 20% of patients with CVI will develop one or more autoimmune diseases [3-5], indicating that CVI is a disease of abnormal immune regulation as well as immunodeficiency. Autoimmune (Coomb positive) hemolytic anemia and idiopathic thrombocytopenic purpura are the two most common autoimmune diseases seen [3-5]. Neutropenia is also seen in many patients with CVI, and in some cases antigranulocyte antibodies have been shown [3, 12]. Paradoxically, these patients are unable to mount an antibody response to infecting microorganisms but retain the ability to produce autoantibodies against erythrocytes, platelets, and granulocytes. Treatment of these autoimmune blood dyscrasias in patients with CVI is difficult because standard treatments such as corticosteroids, cytotoxic agents, and splenectomy have the potential to increase further the patient's immunodeficiency and susceptibility to infections. High-dose intravenous immunoglobulin has been used successfully in one patient with CVI and autoimmune hemolytic anemia [13] and may be an effective alternative to immunosuppressive therapy in these patients.

Pernicious anemia occurs in about 10% of patients with CVI. Although the average age of onset of pernicious anemia in nonimmunodeficient patients is in the sixth decade, patients with CVI usually develop pernicious anemia between the ages of 20 to 40 years [14]. Autoimmune thyroid disease occurs with an increased frequency in patients with CVI. Both Grave disease and hypothyroidism have been reported, with manifestations similar to those in nonimmunodeficient patients. Other autoimmune diseases can occur in association with CVI, including rheumatoid arthritis, systemic lupus erythematosus, and the Sjogren syndrome [3-5].

Patients with CVI frequently develop lymphoproliferative disorders, which can take several forms. Malignant lymphoma occurs with increased frequency in these patients, although the exact magnitude of this increase is unclear. In the study of 220 patients from Great Britain, 3 lymphomas were seen, which represents a 30-fold increase in the incidence of malignant lymphoma compared with the general population [11]. In a separate U.S. study, 7 lymphomas were seen in 98 patients with CVI [15]. Interestingly, all the lymphomas in this study occurred in female patients. Comparing these data with the age-adjusted figures for the expected incidence of lymphoma in the general population, this number represents a 400-fold increase in the incidence of lymphomas in female patients. This high incidence of lymphomas (and cancers in general) is another indication that some patients with CVI may have clinically important defects in cell-mediated immunity.

More common than malignant lymphoma is the occurrence of benign lymphoproliferative disorders. Approximately 30% of patients with CVI will have splenomegaly, diffuse lymphadenopathy, or both. Histologically, this lymphoid hyperplasia can take several forms [16]. In reactive follicular lymphoid hyperplasia, the lymph node architecture is preserved; in atypical reactive lymphoid hyperplasia, however, the lymph node architecture is disrupted by a polymorphic lymphocytic infiltrate. This latter type of hyperplasia can be mistaken for malignant lymphoma. Atypical, reactive lymphoid hyperplasia can usually be differentiated from lymphoma by immunohistochemical studies that show a mixture of T and B cells. In addition, clonally rearranged T- or B-cell antigen receptor bands are not detected on Southern blot analysis of DNA. Lymphocytic infiltrates with these same histologic patterns can also occur extranodally, usually in the gastrointestinal tract. Nodular lymphoid hyperplasia is defined as multiple discrete nodules consisting of lymphoid aggregates confined to the lamina propria and submucosa of the intestinal tract. This condition, whose cause is unknown, can be seen in normal persons (especially children) but occurs frequently in patients with CVI. Radiographically, nodular lymphoid hyperplasia appears as multiple discrete nodules that are predominantly seen in the small intestine. Histologically, these nodules consist of local accumulations of B lymphocytes organized into follicles. It can persist unchanged for years and requires no specific therapy. Rare cases of intestinal lymphomas have been reported in association with nodular lymphoid hyperplasia, but their relation is unknown. Atypical lymphoid infiltrates can occur at other extranodal sites including the lung, nasopharynx, and bone marrow. These lesions usually do not require specific therapy and frequently undergo spontaneous regression and expansion. Because they are seen in a group of patients at risk for developing lymphomas, their main clinical significance is that atypical lymphoid infiltrates occurring extranodally may be confused with malignant lymphoma and result in unnecessary chemotherapy or irradiation.

Several types of arthritis, including rheumatoid and infectious arthritis, can occur in patients with CVI. The latter is not common, especially in patients receiving immunoglobulin replacement. The infecting organisms are usually encapsulated bacteria; however, reports exist of Mycoplasma species causing septic arthritis in patients with CVI [17]. More commonly, patients with CVI develop an asymmetrical, oligoarticular arthritis with a predilection for the larger joints (knees and ankles). The cause of this form of arthritis is unknown, but it rarely leads to joint destruction and often improves with immunoglobulin treatment.

A sarcoid-like syndrome, characterized by noncaseating granulomas in multiple organs, can occur in patients with CVI. The granulomas most commonly occur in the lung, lymph nodes, skin, bone marrow, and liver. Their cause is unclear, and they often undergo spontaneous expansion and regression without therapy. Two important points about this syndrome need to be emphasized. First, whenever these granulomas are identified, it is important to determine whether mycobacterial and fungal infection exist using appropriate cultures and special stains. Second, in most cases no treatment of these lesions is necessary. Occasionally this granulomatous process will become aggressive and result in tissue destruction and then corticosteroid therapy can be useful.

Common variable immunodeficiency should be suspected in any patient with abnormally recurrent or severe bacterial infections of the upper and lower respiratory tract. Patients need to be evaluated for other conditions that may predispose them to recurrent respiratory tract infections such as allergies, anatomic abnormalities, complement deficiencies, defects of ciliary function, and cystic fibrosis. Patients with either selective IgG subclass deficiency or impaired antibody responses to polysaccharides may develop recurrent bacterial respiratory infections [18, 19]. These patients have normal total IgG levels, which serves to distinguish them from patients with CVI. The laboratory workup for humoral immunodeficiency (including CVI) consists of measuring serum immunoglobulin levels, including subclasses of IgG. In addition to measuring immunoglobulin levels, qualitative aspects of the humoral immune response should be evaluated by measuring the IgG response to immunization with protein and polysaccharide antigens. This can be accomplished by immunizing patients with tetanus toxoid and pneumococcal polysaccharide vaccines and measuring pre- and postimmunization (4 to 6 weeks) titers. A fourfold increase in antibody titers is considered a normal response. The decision to treat a patient with immunoglobulin replacement should not be based on the absolute level of serum IgG (or IgG subclass) but rather on the frequency and severity of infections and on the inability to respond to immunization.

Dr. Warren Strober (Laboratory of Clinical Investigation, NIAID, NIH, Bethesda, Maryland): Although the basic immunologic defect(s) that cause CVI are unknown, a number of in vitro immunologic abnormalities have been identified in patients with this syndrome.

B cells originate from pluripotential stem cells in the bone marrow. One of the first B cells is the pre-B cell (Figure 1). Through a series of genetic recombination events, this cell has assembled a functional heavy-chain (but not light-chain) variable-region gene and is able to express cytoplasmic immunoglobulin heavy-chain protein. The pre-B cell eventually rearranges and expresses its light-chain gene and thus gains the ability to produce a functional, membrane-associated IgM molecule. At this point, the cell is an immature B cell that migrates from the marrow to peripheral lymphoid tissues, where it undergoes further maturation into a mature B cell, which expresses both membrane-associated IgM and IgD and is capable of interacting with antigen. When a mature B cell is stimulated by antigen, it becomes activated and undergoes clonal expansion. During this clonal expansion, progeny of the parent B cell undergo a series of immunoglobulin heavy-chain gene rearrangements that result in the juxtaposition of the variable-region DNA sequences with different heavy-chain constant-region genes. This process, known as isotype class switching, results in B cells capable of producing IgG, IgE, or IgA molecules with the same variable (antigen-binding) region as the parent IgM B cell. A final step of B-cell differentiation involves terminal differentiation of B cells into immunoglobulin-producing plasma cells. The process of normal B-cell differentiation is very dependent on T cells, which control both quantitative and qualitative aspects of B-cell differentiation through the secretion of lymphokines such as interleukin-2, interleukin-4, interleukin-5, and -interferon.

Unlike X-linked agammaglobulinemia, in which there is a virtual absence of mature B cells (because of a block in B-cell maturation at the pre-B-cell stage), most patients with CVI have normal numbers of mature B cells in the peripheral blood and lymphoid tissues. However, B cells from these patients are unable to normally differentiate into immunoglobulin-secreting cells. Although the exact immunologic lesion(s) responsible for this defect in B-cell differentiation is not known, abnormalities of either T- or B-cell function at any one of the steps along the B-cell differentiation pathway could result in CVI.

Research into the cellular basis of CVI began some 20 years ago when investigators at this and other institutions studied patients using newly developed techniques of in vitro cell culture. In these initial studies, peripheral blood lymphocytes from patients with CVI were stimulated in vitro with the lectin pokeweed mitogen, and culture supernatants were assayed for immunoglobulin production. A consistent finding, in this type of study, was the inability of peripheral blood lymphocytes from patients with CVI to secrete normal amounts of immunoglobulins [20, 21]. Because pokeweed mitogen-induced immunoglobulin secretion is strictly dependent on the presence of functional helper T cells, it was unclear whether this decreased immunoglobulin production by patient B cells was due to an abnormality intrinsic to the B cell or was secondary to a T-cell abnormality.

To address this question, partially purified T cells and B cells of patients were separately analyzed in pokeweed mitogen-driven cultures containing allogeneic normal T cells and B cells. This approach showed that most patients with CVI appeared to have a B-cell defect in that co-culture of their B cells with normal allogeneic T cells did not correct the defect in immunoglobulin production, and T cells from most patients were able to support immunoglobulin production by normal B cells [22, 23]. Occasionally patients were identified whose T cells were capable of suppressing normal B-cell immunoglobulin production, but in most patients shown to have excessive suppressor T-cell activity, co-culture of their B cells with normal T cells failed to rescue B-cell function; this result indicated that these patients also had a B-cell defect [22, 23]. To define further the B-cell abnormality, a number of studies were done in which partially purified B cells from patients with CVI were exposed to various T-cell and monocyte-independent stimuli, and their ability to produce IgM and IgG was examined. Studies using combinations of B-cell mitogens or soluble T-cell factors or both to stimulate CVI B cells showed that most patients could be classified into one of three subgroups based on their response. Thus, patients were found to produce either little or no immunoglobulin, IgM with little or no IgG, or normal levels of IgM and IgG [24-26]. The exact proportion of patients in each subgroup varied somewhat among studies, but the patterns of response were similar regardless of the type of stimulation used. These studies indicate that the defect in B-cell immunoglobulin production in CVI is not absolute because most of the patient's B cells can secrete some immunoglobulin if given the appropriate in vitro stimulus. These results also suggest that B cells from at least some patients with CVI may not be intrinsically abnormal and require only the appropriate stimuli (which are presumably lacking in vivo) to mature into immunoglobulin-secreting plasma cells.

In addition to these defects in B-cell function, many patients with CVI also have in vitro evidence of abnormal T-cell function. Kruger and colleagues [27] have shown that patients with CVI, as a group, have T cells that proliferate less and produce less interleukin-2 than normal T cells when exposed to various T-cell stimulants. Further studies from our laboratory defining the T-cell abnormalities in CVI will be discussed below.

One way of interpreting these studies of T- and B-cell function in CVI is based on the well-established fact that the process of normal B-cell differentiation and maturation is dependent on T-cell lymphokines at virtually every step. Thus, the failure of B cells from patients with CVI to produce immunoglobulin in various in vitro assays (even in the presence of normal allogeneic T cells or soluble T-cell factors) does not necessarily point to an intrinsic B-cell defect. It is possible that if a B cell does not receive the appropriate T-cell signal during a critical early stage of development (because of a defect in T-cell lymphokine production), it may subsequently be unable to differentiate into an immunoglobulin-secreting plasma cell with in vitro stimulation. This hypothesis is compelling and has certain therapeutic implications (see below), but the data supporting it are incomplete. Although defects in T-cell lymphokine production are seen, they are not present in every patient with CVI. Until it can be shown that the functional abnormalities of CVI B cells can be completely reversed by exposure to the appropriate T-cell lymphokine(s), we will assume that CVI is associated with some form of primary B-cell defect.

A final consideration about studies of the pathogenesis of CVI relates to whether a genetic predisposition to this syndrome exists. Evidence for a genetic influence was noted early in the study of CVI, when family studies showed that many patients had first-degree relatives with other immunologic diseases such as isolated IgA deficiency and various autoimmune diseases [28]. Recent studies by Schaffer and colleagues [29] provide further evidence for a genetic influence in the pathogenesis of CVI. These investigators examined the class III major histocompatibility region in patients with CVI or isolated IgA deficiency. This region is located on chromosome 6, between the HLA-B and HLA-DQ regions, and it contains genes encoding complement proteins as well as the 21-hydroxylase A and tumor necrosis factor genes. In this study, 63% of patients with CVI and 56% of those with IgA deficiency had rare C2 alleles or C4A and 21-hydroxylase A deletions or both, whereas these features were found in only 15% of normal patients [29]. These results suggest that CVI and IgA deficiency are related disorders, and susceptibility is determined by a gene or genes near the major histocompatibility class III region.

Dr. Eli Eisenstein (Laboratory of Clinical Investigation, NIAID, NIH, Bethesda, Maryland): As alluded to above, CVI is a primary humoral immunodeficiency syndrome that has both expected and unexpected clinical manifestations. Recurrent bacterial infections are explained by the defect in antibody production that is the hallmark of this syndrome. In contrast, a number of infectious, autoimmune, and neoplastic diseases occur in these patients that are not well explained by a simple antibody deficiency, suggesting that many patients have clinically important abnormalities of T-cell function. Such abnormalities could contribute to the hypogammaglobulinemia of CVI by failing to provide the appropriate signals necessary for normal B-cell maturation.

We examined the expression of lymphokine and activation genes in mitogen-stimulated T cells from a group of patients with CVI [30]. For this study, we chose patients with a normal peripheral blood T-cell phenotype. Specifically, T cells from these patients did not differ from normal persons in total T-cell number or CD4/CD8 ratio. Thus, any observed abnormalities in lymphokine production would not be due to quantitative differences in the absolute numbers of CD4+ or CD8+ T cells. In initial studies, peripheral blood mononuclear cells from CVI patients and normal persons (controls) were stimulated with phytohemagglutinin (a T-cell mitogen), and mRNA was then isolated from these cells at specified periods. The mRNA was blotted onto nitrocellulose filters, hybridized with cDNA probes, and exposed to x-ray film. The resultant autoradiographs were analyzed by scanning densitometry to quantitate mRNA levels. In addition, supernatants from these cultures were analyzed for secreted interleukin-2 and -interferon. Using this approach, we found that phytohemagglutinin-activated patient T cells produced less interleukin-2 mRNA and protein than did control T cells (P < 0.05; Figure 2 and 3).

When -interferon expression was examined, it was found that, although patient T cells had normal levels of -interferon mRNA at 6 hours (suggesting that the initial activation of the -interferon gene was normal), mRNA levels at 24 hours were decreased (P < 0.05), as were levels of -interferon protein in the culture supernatants (see Figure 2). In normal T cells, the initial [1 to 6 hour] increase in -interferon mRNA expression seen after mitogen activation of human T cells is independent of the presence of interleukin-2 protein. However, in the later stages of activation, -interferon mRNA and protein production are augmented by interleukin-2 [31]. This finding suggests that the deficient production of -interferon by these CVI T cells does not represent a primary abnormality in -interferon gene activation but rather is secondary to the deficient interleukin-2 production. This argument is strengthened by our finding that -interferon production by patient T cells could be normalized by the addition of recombinant interleukin-2 to these cultures [30]. In contrast to these abnormal findings, phytohemagglutinin-activated T cells from these patients were found to express normal amounts of mRNA for the chain or the interleukin-2 receptor (CD25) and the proto-oncogene known as c-myc (see Figure 2). Thus, rather than having a global abnormality of T-cell activation, these patients seem to have a selective defect in the ability to activate their interleukin-2 gene normally.

To characterize further the abnormality in interleukin-2 production in this group of patients, we did studies using purified CD4+ T cells rather than whole mononuclear cells. Patient CD4+ T cells were purified by negative selection using antibody-coated immunomagnetic beads and were then stimulated with phytohemagglutinin in the presence of normal allogenic macrophages [32]. As shown in Figure 4(left), patient CD4+ T cells cultured in this manner had deficient interleukin-2 production. Thus, the abnormality in interleukin-2 production is in the CD4+ T cell and is not due to abnormal macrophage function.

Although this T-cell abnormality in CVI may explain certain clinical features of the syndrome (such as the increased incidence of malignancy and the occasional occurrence of certain types of opportunistic infections), it does not clearly account for the hypogammaglobulinemia. Thus, the abnormal B-cell function cannot usually be corrected by the in vitro addition of normal T cells or their secreted lymphokines (including interleukin-2). We believe that CVI is characterized by a defect common to both T cells and B cells, which leads to a functional abnormality of lymphokine and immunoglobulin production, respectively.

In speculating about the defect, we know that B cells and T cells use many of the same signal-transduction mechanisms and second messengers. For example, tyrosine kinases are important in transducing signals generated by the T-cell receptor/CD3 complex in T cells and by the immunoglobulin receptor in B cells [33, 34]. In addition, in both T and B cells, nuclear binding proteins are known to be important in regulating the expression of genes for secreted cell products (lymphokines in T cells and immunoglobulin in B cells). Thus, CVI could occur as a result of a biochemical defect in activation pathways common to both T and B cells, which results in the reduced ability of T and B cells to undergo normal differentiation and produce secreted products.

Dr. Jonathan S. Jaffe (Laboratory of Clinical Investigation, NIAID, NIH, Bethesda, Maryland): Within the CVI syndrome, there is a subgroup of patients with a distinct T-cell phenotype. Peripheral blood T cells from these patients have an abnormally low CD4/CD8 ratio that is due to an increase in the absolute number of CD8+ T cells (CD8^hi patients). Clinically, CD8^hi patients have a greater incidence of splenomegaly and depressed in vivo T-cell responses to recall antigens than do CVI patients with normal CD4/CD8 ratios [35]. The expanded population of CD8+ T cells from these patients has an abnormal surface phenotype characterized by increased expression of the surface proteins CD57 and HLA-DR [35]. CD8 (+) T lymphocytes that coexpress CD57 have been shown to mediate both cytotoxic and suppressor cell activity, and HLA-DR is a surface marker normally associated with T-cell activation [36, 37]. In vivo expansion of CD8+ T cells expressing the CD57 and HLA-DR surface markers has been observed in a number of clinical conditions including chronic cytomegalovirus infections, acute Epstein-Barr virus infections, and human immunodeficiency virus (HIV) infections [38-42]. Thus, the expanded population of CD8+ T cells from these CD8^hi patients has a surface phenotype (CD8+ CD57+ HLA-DR+) characteristic of virus-induced, activated cytotoxic T cells.

To define further the immunologic abnormalities in patients with the CD8^hi phenotype, we did detailed functional studies on purified T-cell subpopulations from four patients in this subgroup [43]. T cells were purified into CD4+ and CD8+ populations by a negative selection process using antibody-coated immunomagnetic beads.

We first tried to determine if CD4+ T cells from these CD8 (^hi) patients had the same abnormality of interleukin-2 production noted in CVI patients with normal CD4/CD8 ratios. When purified CD4+ T cells from CD8^hi patients were stimulated with phytohemagglutinin, they produced normal amounts of interleukin-2 (Figure 4). Thus, the abnormality of interleukin-2 production in CD4+ T cells from patients with CVI, who have normal CD4/CD8 T-cell ratios, is not present in CD4+ T cells from the CD8^hi subgroup of patients. We next examined the functional properties of these patient's CD8+ T cells. We found that, compared with control CD8+ T cells, CD8+ cells from CD8^hi patients had increased cytotoxic activity and produced markedly increased levels of -interferon [43]. Further studies were done to determine if CD8+ T cells from the CD8^hi patients were more potent suppressors of B-cell immunoglobulin production than were control CD8+ T cells. In these studies, purified CD8+ T cells (from patients and controls) were added at various T-cell/B-cell ratios to normal tonsillar B cells stimulated with the combination of the B-cell mitogen Staphylococcus-A Cowan's and interleukin-2. As shown in Figure 5, CD8+ T cells from the CD8^hi patients consistently suppressed IgG secretion in a dose-dependent manner, whereas CD8+ T cells from normal controls did not suppress IgG production at any of the T-cell/B-cell ratios [43].

The finding that patient CD8+ T cells are able to suppress IgG secretion by normal B cells suggests that patient B cells might function more normally in the absence of autologous CD8+ T cells. To test this hypothesis, we obtained highly purified B cells from four CD8^hi patients and assessed their ability to produce immunoglobulin in response to activation with Staphylococcus-A Cowan's in the presence of interleukin-2 (that is, in the absence of autologous T cells that might cause suppression of IgG secretion). We found that B cells from each of the four patients tested were able to produce normal amounts of IgM, but only one patient produced normal amounts of IgG [43]. These studies indicate that, in most cases, purification of CD8 Boron-hi cells does not result in restoration of normal immunoglobulin production, suggesting that there is no simple relationship between increased suppressor T-cell function and reduced B-cell function. Thus, either the CD8+ T cells are not pathogenetically related to the defect in antibody production (and the latter is due to an independent B-cell defect), or they are affecting B-cell differentiation at an early stage, so as to lead to the development of B cells that cannot respond to in vitro stimuli even when studied in the absence of suppressor T cells.

These in vitro findings have important implications for the pathogenesis of CVI in this CD8^hi subgroup of patients. As alluded to earlier, several acute and chronic viral infections induce the expansion of CD8+ T cells with a surface phenotype identical to that of the expanded population of CD8+ T cells seen in CD8^hi patients. Furthermore, CD8+ T cells from CD8^hi patients have the functional characteristics of activated cytotoxic/suppressor cells. One possibility is that the expanded population of CD8+ T cells in CD8^hi CVI are footprints of a chronic viral infection. Possibly, in a genetically predisposed person, a chronic viral infection could result in immunodeficiency either by directly impairing lymphocyte function, or indirectly by inducing abnormal cytotoxic or suppressive CD8+ T cells. Precedent for such a mechanism of immunodeficiency exists in the X-linked lymphoproliferative syndrome [44]. In this syndrome, genetically susceptible individuals are clinically and immunologically normal before infection with Epstein-Barr virus [45] but after infection surviving patients develop an immunodeficiency syndrome characterized by low CD4/CD8 ratios, hypogammaglobulinemia, and lymphoproliferative disease [44, 46]. Also, immunologic abnormalities seen in CD8^hi patients may either be unrelated to viral infection or may be the result of a chronic viral infection that is not a primary cause of the immunodeficiency. In the latter case, the observed T-cell abnormalities would be secondary to a primary B-cell immunodeficiency, which predisposes the person to viral infection.

Dr. Charlotte Cunningham-Rundles (Departments of Medicine and Pediatrics, Mt. Sinai Medical Center, New York, New York): The first mode of therapy in CVI is aggressive treatment of ongoing infections with antibiotics. When empiric antibiotic therapy is necessary, an agent with activity against Haemophilus influenzae and Streptococcus pneumoniae should be used. The course of treatment in patients with CVI should be longer than for persons with normal immunity, and intravenous administration may be required in some cases.

Aside from treating acute and chronic infections in CVI, it is essential to increase the antibody content of the blood by the administration of immunoglobulin concentrates. The prophylactic administration of immunoglobulin in hypogammaglobulinemia was initiated in 1952 [47]. The first patient was an 8-year-old boy who received 3.2 g of an immunoglobulin concentrate at monthly intervals. This dose increased the amount of immunoglobulin observed on serum protein electrophoresis. The choice of the time interval between doses was based on the observation that a gradual decline in the level of -migrating proteins occurred during a 3-to-4 week period. The dose was the same as that subsequently recommended by the World Health Organization for hypogammaglobulinemic patients (100 mg/kg per month given by intramuscular injection, weekly or biweekly). This dose increases the serum IgG level to 1.0 or 1.5 g/L, the minimum level necessary for the prevention of serious infections [48].

Although the use of the World Health Organization-recommended dose diminished rates of morbidity and mortality in hypogammaglobulinemic patients, it soon became clear that this was not the optimum dose. Twenty years ago, the Medical Research Council of Great Britain did a trial in which patients treated with the standard dose (100 mg/kg per month) were compared with patients receiving twice this amount. Patients receiving the larger dose developed fewer infections [49]. However, the volume of this larger dose was difficult to administer intramuscularly; for example, a patient weighing 60 kg needed the intramuscular administration of 20 mL of gammaglobulin concentrate each week. Thus, the dose eventually recommended for hypogammaglobulinemic patients was based more on expediency than on available clinical data.

Thus, various research groups set out to devise immunoglobulin preparations suitable for intravenous administration. These efforts resulted in a number of products, including the eight preparations currently licensed for use in the United States (Table 1). The first report establishing the efficacy of intravenous immunoglobulin treatment was a crossover trial of 34 patients that compared the standard intramuscular dose (100 mg/kg per month) with the same dose of a reduced and alkylated intravenous preparation [50]. This trial was followed by subsequent studies in which the standard intramuscular treatment was compared with progressively larger doses (150 to 600 mg/kg per month) of intravenous preparations [51-53]. These studies established that intravenous therapy with doses of 300 to 600 mg/kg given every 3 to 4 weeks was superior to intramuscular therapy in preventing infections. In addition, larger doses may have a particularly beneficial effect on rates of infection in patients with bronchiectasis [53].

With intravenous immunoglobulin therapy, postinfusion serum IgG levels increase by approximately 2.5 g/L for each 100 mg/kg infused. Individual patients have differences in the rate at which they catabolize IgG, and some investigators have tried to individualize doses to establish a minimum trough IgG level [54]. Most investigators consider that the dose of intravenous immunoglobulin should be sufficient to result in a trough level of no less than 5.0 g/L [55]. In most patients this level can be reached by the administration of 200 to 400 mg/kg of intravenous immunoglobulin every 3 to 4 weeks.

Most patients with CVI tolerate intravenous immunoglobulin therapy well. The most frequent adverse reactions are nonanaphylactic and characterized by back or abdominal pain, nausea, vomiting, chills, fever, and myalgias. These symptoms usually begin within the first 30 minutes of the infusion and are not associated with dyspnea or hypotension. Headaches and fatigue may occur at the end of an infusion and may persist for several hours. This type of reaction is most often seen in newly treated patients or those with an active infection and may be related to the binding of infused antibodies to microbial antigens. These reactions are best treated by interrupting the infusion until the symptoms subside and then restarting the infusion at a slower rate. Occasionally, patients with severe reactions of this type may require pretreatment with corticosteroids [56].

In patients with CVI, true anaphylactic reactions to intravenous immunoglobulin infusions are rare. Signs and symptoms of anaphylaxis begin seconds to hours after starting the infusion and consist of flushing, facial swelling, dyspnea, and hypotension. This type of reaction may be mediated by anti-IgA antibodies (of the IgE isotype) in the patient reacting to the IgA in the infused gammaglobulin [54]. One of the immunoglobulin products intended for intravenous infusion has been sufficiently depleted of IgA during manufacturing to permit its use in patients who have anti-IgA antibodies [54].

By arrangements with physicians or various home care services, patients can learn to administer their own infusions or have a family member or home care nurse do so. Depending on the arrangements, the cost of this form of treatment can be less than that in a hospital or clinic. Thus, an increasing number of hypogammaglobulinemic patients receive their treatments at home.

Patients with CVI who, as a result of repeated infections, have structural damage to the sinuses or lungs are a difficult problem. Despite treatment with antibiotics and adequate doses of immunoglobulin, these patients may continue to have recurrent sinus and lung infections. Unfortunately, sinus surgery offers little chance of long-term success. Patients with CVI who have developed bronchiectasis often require postural drainage or other physical therapies to enhance adequate clearance of pooled secretions. Some patients with CVI develop chronic bronchitis with bronchospasm and may require treatment with bronchodilators or aerosolized corticosteroid preparations or both. The use of systemic corticosteroids in patients with CVI should be avoided whenever possible.

Patients with CVI frequently develop various gastrointestinal disorders. Among the gastrointestinal infections frequently encountered are those due to Giardia lamblia, Campylobacter species, and other enteric pathogens. These infections improve after usually appropriate antimicrobial therapy. Occasionally, gluten-sensitive enteropathy or lactose intolerance is diagnosed, and these conditions improve if the patient avoids gluten or lactose. Unfortunately, most of the other gastrointestinal disorders occurring in patients with CVI (including inflammatory bowel disease and idiopathic diarrhea with malabsorption) have no clear cause or definitive treatment [57].

Although the introduction of intravenous immunoglobulin, home care services, and various new antibiotics have certainly improved the treatment of CVI, patients continue to have recurrent infections and to develop other manifestations of this syndrome, such as autoimmune disease and cancer. Clearly, an ideal treatment would be to reconstitute the patient's immune system. Because the underlying gene defect(s) in CVI are entirely unknown and likely to be diverse, gene therapy is a distant and perhaps impractical goal.

Another approach to the treatment of CVI is the development of some form of cytokine or other pharmacologic therapy that could normalize the immunologic defects in these patients. As discussed earlier, various in vitro studies suggest that B cells from certain patients with CVI can be induced to secrete normal amounts of IgG. In addition, normal immunoglobulin production has been restored in vivo in one CVI patient who became infected with HIV and restored in a few patients treated with cimetidine [58, 59]. These and other data [60] suggest that the immunologic defects in some patients with CVI could be corrected by immunomodulatory therapy, but further research on the nature of these defects is needed to make such therapy a reality.

CVI: common variable immunodeficiency