Autoantibodies and Lymphoproliferative Diseases in Acquired C1-Inhibitor Deficiencies

Abstract

Angioedema due to acquired C1-inhibitor (C1-INH) deficiency (also referred to as “acquired angioedema”) is a rare, life-threatening disease with poorly defined etiology, therapy, and prognosis. To define the profile of acquired C1-INH deficiency and to facilitate the clinical approach to these patients, we report on 23 patients with acquired C1-INH deficiency followed for up to 24 years (median, 8 yr), and review the literature.

We measured C1-INH activity with chromogenic assay and detected autoantibodies to C1-INH by enzyme-linked immunosorbent assay (ELISA). Median age at onset of angioedema was 57 years (range, 39–75 yr). All patients had C1-INH function and C4 antigen below 50% of normal. C1q was reduced in 17 patients. Autoantibodies to C1-INH were present in 17 patients. Long-term prophylaxis of attacks with danazol was effective in 2 of 6 patients, and with tranexamic acid, in 12 of 13 patients. Therapy with C1-INH plasma concentrate was necessary in 12 patients: 9 had rapid positive response and 3 became progressively resistant. Associated diseases at the last follow-up were non-Hodgkin lymphomas (3 patients), chronic lymphocytic leukemia (1 patient), breast cancer (1 patient), monoclonal gammopathies of uncertain significance (13 patients). In 4 patients no pathologic condition could be demonstrated.

Compared with the general population, patients with acquired C1-INH deficiency present higher risk for B-cell malignancies, but not for progression of monoclonal gammopathies of uncertain significance to malignancy. Antifibrinolytic agents are more effective than attenuated androgens in long-term prophylaxis. Patients with acquired C1-INH deficiency may be resistant to replacement therapy with C1-INH plasma concentrate.

INTRODUCTION

Abbreviations used in this article: C1-INH, inhibitor of the first component of human complement, MGUS, monoclonal gammopathies of uncertain significance.

Angioedema due to acquired deficiency of the inhibitor of the first component of human complement (C1-INH) is a rare condition with little more than 100 cases reported in the literature (^{2,3,5–7,9,12,13,15–18,24,28,29,32,34,35,37,38,40,43,44,46,48,50–54,56–60,62,63,66–68,70,71,73–75,78,79,81,82,84–86,90}). It is frequently referred to as “acquired angioedema,” but we will not use this term since acquired angioedema can also occur with normal C1-INH. Characteristics of acquired C1-INH deficiency are the increased consumption of C1-INH and the hyperactivation of the classical pathway of human complement (⁶⁴). As a consequence, these patients have almost undetectable serum levels and/or activity of C1-INH, C1q, r, s, C4, and C2. Usually, these abnormalities are constantly present, but temporary normalization of 1 or more of these parameters has been reported (⁸²).

The clinical manifestations of the disease mimic those of the more common genetic defect of C1-INH, hereditary angioedema, and include subcutaneous, nonpruritic swelling without accompanying urticaria, involvement of the upper respiratory tract manifested as dysphagia, voice change or respiratory stridor, and partial obstruction of the gastrointestinal tract presenting as colicky abdominal pain (¹). Acquired C1-INH deficiency differs from hereditary angioedema by the absence of a family history of angioedema and the late onset of symptoms (fifth decade of life or later); response to treatment varies due to the C1-INH hypercatabolism characteristic of acquired C1-INH deficiency (²¹).

Acquired C1-INH deficiency is frequently reported in association with B-lymphocyte proliferation, either true malignancies or simple monoclonal gammopathies of undetermined significance (MGUS) (³⁷). In some cases neoplastic lymphatic tissues were shown to consume C1-INH (⁷⁸) and/or classical pathway complement components (⁴⁹), suggesting that they were directly involved in the pathogenesis of acquired C1-INH deficiency. Scattered reports describe acquired C1-INH deficiency associated with nonhematologic neoplasm, infections, or autoimmune diseases, while 14% of patients with acquired C1-INH deficiency have no other disease.

The etiopathogenetic puzzle of acquired C1-INH deficiency is completed by the description of autoantibodies inactivating C1-INH (⁵³). Initially identified in otherwise healthy patients, they were described later in patients with associated diseases, also. These autoantibodies are common in patients with MGUS, and frequently exhibit the same isotype of the M component (^19,37). It becomes clear from the above that the entire array of B-lymphocyte disorders (autoimmunity, MGUS, overt neoplasia) can be present in acquired C1-INH deficiency.

After angioedema due to acquired C1-INH deficiency has been diagnosed, it remains a challenge for the clinician to define etiology, establish a therapeutic program, and make a prognosis. Apart from a recent paper that collected complete information on 19 patients (³⁷), data from the literature are scattered and based mostly on single reports. Therefore it is difficult to predict the patient’s outcome. To help define the profile of acquired C1-INH deficiency and to facilitate the clinical approach to these patients, we report our experience with 23 patients with acquired C1-INH deficiency followed for up to 24 years, and review the literature.

PATIENTS AND METHODS

Twenty-three patients (median age, 72 yr; range, 47–83 yr) with the diagnosis of angioedema due to acquired C1-INH deficiency were followed for 1–24 years (median, 8 yr). The diagnosis was based on a history of recurrent angioedema without urticaria, which began during or after the fourth decade of life; absence of a family history of angioedema; and detection of C1-INH functional levels below 50% of normal.

All patients routinely underwent physical examination; blood testing including complete blood cell count, liver and renal function, protein electrophoresis, basic autoantibody screening (antinuclear, antimitochondria, antismooth muscle, rheumatoid factor); urinalysis; abdominal echography; and standard chest X-ray. Further specific analyses were performed upon detection of abnormalities. If no abnormalities were detected the above-mentioned tests were repeated every other year.

Complement parameters

Serum or plasma samples were stored at −80 °C until tested. C1-INH, C4, C3, and C1q antigens were measured with radial immunodiffusion plates (Nor Partigen, Low Partigen for C1q Behring, Marburg, Germany). C1-INH activity was measured with a commercially available chromogenic assay (Baxter). All values were expressed as percentages of the normal mean. Autoantibodies to C1-INH in serum were measured by enzyme-linked immunosorbent assay (ELISA) using the method described by Alsenz et al with slight modifications (^3,21). Immunoblotting for C1-INH was performed on plasma samples collected in EDTA-polybrene as previously described (¹⁹).

RESULTS

The age at onset of angioedema symptoms ranged from 39 to 75 years (median, 57 yr), the age at diagnosis ranged from 42 to 76 years (median, 60 yr) (Table 1). All patients except 1 had angioedema symptoms localized to the subcutaneous tissues. The face was the most common site, but extremities and external genitals were also sometimes affected. Thirteen patients (57%) suffered from recurrent abdominal pain, lasting 24–48 hours without other evident disease, that was thus considered to be due to angioedema of the intestinal wall. In 1 patient recurrent abdominal pain was the only manifestation of angioedema. The larynx was involved in 17 patients (74%), in 1 of whom the problem was so frequent and resistant to treatment that a permanent endotracheal cannula was placed.

TABLE 1: Characteristics of 23 patients with acquired C1-INH deficiency

Complement parameters

Median plasma levels of the complement proteins are reported in Table 2. All patients had C1-INH activity below 50% of normal values. C1-INH antigen was within the normal range in 3 patients and C1q, in 5. Normal C1-INH antigen was due to elevated levels of its cleaved form of 96 kD, as detected in immunoblotting experiments (not shown). Autoantibodies to C1-INH were present in 17 patients (74%). In these patients the levels of complement proteins were not significantly different than those of patients without autoantibodies. Autoantibodies belonged to all 3 major immunoglobulin classes: 7 were IgG, 5 IgM, and 4 IgA. In 1 patient both IgG and IgM were present.

TABLE 2: Plasma levels of complement proteins expressed as percentage of normal values in 23 patients with acquired C1-INH deficiency

Associated diseases

Most of our patients had an associated pathologic condition that was present at the onset of angioedema or developed thereafter. MGUS was by far the most frequent condition associated with acquired C1-INH deficiency, followed by B-lymphocyte malignancies. In 2 patients, non-Hodgkin lymphoma was already present at the onset of angioedema symptoms. In 1, the diagnosis of low-malignancy non-Hodgkin lymphoma had been made 6 years before and since that time the patient had been on different chemotherapy regimens. At the time the angioedema started, neoplastic cells were present in the bloodstream and the patient was on chlorambucil. In the other patient, the diagnosis of low-malignancy non-Hodgkin lymphoma was almost coincident with the appearance of the abdominal symptoms that led to the diagnosis of acquired C1-INH deficiency.

In 2 other patients, non-Hodgkin lymphoma developed after the onset of angioedema. Neither of these patients had MGUS at the onset of angioedema. In the first patient, T-cell rich B-cell lymphoma was diagnosed upon splenectomy 7 years after the first episode of angioedema. The lymphoma had a short remission on chemotherapy, but relapsed 4 months later. The second patient had Waldenström disease 2 years after the onset of angioedema that remitted on cyclophosphamide and relapsed 7 years later evolving into a rapidly progressive lymphoma despite chemotherapy and rituximab. This patient died 14 years after the onset of angioedema. In both patients complement parameters normalized at the time of the first remission and remained normal despite the disease relapse. At this writing, the first patient is still on chemotherapy.

Treatment of angioedema

Following the same policy used in hereditary angioedema, we suggested a long-term prophylactic treatment for those patients who suffered from more than 1 angioedema attack per month (that is, from 5 days of disability/month). Danazol was given to 6 patients and was successful in 2 (Table 3). The other 4 patients had no reduction in disease severity; indeed, 2 of them noticed an aggravation of symptoms. These 4 patients switched with good results to oral tranexamic acid, 1 g 3 times a day. A total of 13 patients were on long-term tranexamic acid: in 8 patients angioedema drastically decreased in frequency (less than 3 attacks per year); in 4, angioedema decreased to 3–6 attacks per year; in 1 patient there was no improvement and treatment was stopped. Despite its effectiveness, tranexamic acid did not always protect from laryngeal attacks: 7 such events were recorded in treated patients.

TABLE 3: Efficacy of treatment in 23 patients with acquired C1-INH deficiency

Replacement therapy with C1-INH plasma concentrate is our treatment of choice for life-threatening attacks. Such treatment was necessary in 12 patients for a total of 42 infusions. Eight patients were treated 3 times or fewer. In all of them 1,000–2,000 units (the average dose used for hereditary angioedema) reverted the attacks within 1 hour or less. The remaining 4 patients were treated from 5 to 10 times with C1-INH infusions. In 1 of them, 1,000–2,000 units continued to be effective; in the other 3 the dosage had to be progressively increased to revert the angioedema, in 1 patient up to 12,000 units for a single attack. Nevertheless, even with such a high dose and despite clinical effectiveness, C1-INH function and C4 failed to increase. As shown in Figure 1, C1-INH given with the concentrate was completely catabolized to its inactive form of 96 kD.

FIGURE 1.: Immunoblotting analysis of C1-inhibitor (C1-INH) in a normal subject (lane 1) and in a patient with acquired C1-INH deficiency during a laryngeal edema treated with C1-INH plasma concentrate (lanes 2–4). Lane 2, before infusion; lane 3, 10 minutes after 6,000 units; lane 4, 1 hour later 10 minutes after an additional 6,000 units. Native functional C1-INH of 105 kD, which represents the major band in the normal subject, was never detectable in the patient’s samples.

DISCUSSION

The etiopathogenesis of acquired C1-INH deficiency has been a matter of controversy. Because of the association with malignancy, specifically B-cell malignancies, it has been suggested that the associated disease could be the cause for accelerated C1-INH consumption. Experimental evidence has shown that such consumption could result from either neoplastic tissue itself (^49,78) or complement fixation by peculiar immune complexes as the idiotypeantiidiotypes (³⁹). Based on the finding that the first patients with autoantibodies to C1-INH looked otherwise healthy (^3,53), it was proposed that 2 separate forms of acquired C1-INH deficiency existed: type I, paraneoplastic, mainly associated with lymphatic malignancies; and type II, autoimmune, caused by autoantibodies to C1-INH. The latter form appeared to be further characterized by elevated serum levels of cleaved C1-INH (⁴). This division has been questioned, however (^30,88). We previously showed that cleaved C1-INH was not invariably present in the serum of patients with so-called autoimmune acquired C1-INH deficiency (¹⁹). With regard to the absence of malignancy in autoimmune acquired C1-INH deficiency, the literature shows that of 42 acquired C1-INH-deficient patients reported to have malignancies, only 7 had been tested for anti-C1-INH autoantibodies, and 2 came out positive (^{2,7,12,17,23,56,85,86}). In addition, a recent survey of 19 patients with acquired C1-INH deficiency type 2 lists, at the end of the follow-up period, 3 lymphatic malignancies, 3 solid malignancies of other origin, and 9 cases of MGUS (³⁷). Our series, where all 23 patients were tested for autoantibodies, demonstrates that half of the patients with malignancies (5 B-cell proliferations and 1 breast cancer) also had autoantibodies to C1-INH, either at the time of onset of angioedema or later in the course of the disease. Likewise, only 4 of 18 patients with autoantibodies were otherwise healthy; the others had echinococcus infection (1 patient), MGUS (10 patients), or, as already mentioned, malignancies (3 patients). These data all together indicate that autoimmune acquired C1-INH deficiency is not distinct from the acquired C1-INH deficiency that occurs in the setting of malignancies or other diseases. Detection of autoantibodies to C1-INH in a patient with acquired C1-INH deficiency does not decrease the importance of taking into account the possibility of an associated pathologic condition: a patient with acquired C1-INH deficiency may or may not develop malignancy irrespective to the fact that autoantibodies to C1-INH are detectable in serum. Thus, dividing acquired C1-INH deficiency into 2 categories, based on the presence of neoplasia and the presence of antibodies, is artificial and inappropriate since in many patients the 2 conditions coexist.

The actual risk of malignancy, particularly of B lymphocytes, remains once diagnosis of acquired C1-INH deficiency has been made. Reports in the literature list 105 acquired C1-INH deficiency patients with 49 malignancies, 41 of them lymphatic. In our series of 23 patients, there were 5 malignancies, 4 of them lymphatic (Table 4). While the ratio of lymphatic to nonlymphatic malignancies is similar in the present study and in the literature, the overall incidence of malignancy is strikingly higher in the literature than in our study (45% versus 22%, respectively). The bias that uncomplicated acquired C1-INH deficiency may not be published as a case report likely accounts for this discrepancy. However, the incidence of 4 lymphatic malignancies in 23 patients over a median period of 8 years is unquestionably high compared to the general population. The incidence of new cases of non-Hodgkin lymphoma in 2 parts of Italy between 1985 and 1992 was 14.7 and 7.4 per year per 100,000 people. Even taking into account that throughout Europe the mean rate of increase of these diseases is 4.2% per year, the incidence in our series remains much higher (¹⁴). Hence, we can conclude that the risk of B-cell malignancy is elevated in patients with acquired C1-INH deficiency, while the occurrence of a single malignancy other than lymphatic in our series does not confirm or deny the possibility that a patient with acquired C1-INH deficiency is also at risk for such malignancies.

TABLE 4: Diseases detected in 128 patients with acquired C1-INH deficiency^*

With regard to the presence of MGUS, it is known that 25% of all patients with MGUS show disease progression to myeloma after a median of 10 years (⁷²). The median follow-up of our 13 patients with MGUS was 8 years. The fact that none of them has evolved to myeloma suggests that their risk of malignancy is not higher than other patients with MGUS.

The therapy of acquired C1-INH deficiency should have 2 aims: treatment of the associated diseases and prevention/reversal of angioedema symptoms. The possibility of reversing biochemical and/or clinical abnormalities of acquired C1-INH deficiency by curing the associated disease was first reported by Cohen and colleagues and subsequently confirmed (^{5,12,24,32,40,49,66,77,81,86}). However, the response can be temporary, even without evidence of relapse of the associated disease (^24,40). In our patients modifications of angioedema symptoms and biochemical parameters show an uneven relation to the associated disease. In 2 patients, symptoms and complement parameters consistently reverted to normal with remission of the associated lymphatic malignancy. It is noteworthy that, in both patients, when the disease became aggressive at relapse of malignancy, C1-INH deficiency remained normal and angioedema symptoms did not recur. Complete clinical remission was achieved in 2 other patients. One began to have angioedema shortly after an echinococcal cyst was removed from the liver (²²). Seven years later, when we began seeing this patient, the echinococcal cyst had relapsed and was removed. After surgery the patient experienced a transient normalization of C1-INH lasting a few weeks; echinococcus did not recur, and she remained asymptomatic for angioedema until she died 20 years later from causes unrelated to acquired C1-INH deficiency. In the second patient, angioedema symptoms ended abruptly, independent of any treatment or other apparent reason. The patient remained deficient in C1-INH until he died 18 years later from complications of hepatitis C, without further angioedema. In summary, a cause and effect relationship between acquired C1-INH deficiency and the associated disease clearly can be found in some patients. Nevertheless, such a relationship may be broken or absent, and symptoms may not be present regardless of persisting biochemical abnormalities (^{13,15,32,47,65,69}). It is clear that the biologic mechanisms connecting associated disease, C1-INH deficiency, and angioedema symptoms remain superficially elucidated.

Prevention and treatment of angioedema attacks in patients with acquired C1-INH deficiency is usually shaped on the policy used in patients with hereditary angioedema, that is, attenuated androgens as prophylaxis and C1-INH concentrate for acute attacks (^41,87). Acquired C1-INH deficiency patients may be resistant to these treatments, however (^{4,12,21,31,69}). We found previously that acquired C1-INH deficiency patients resistant to attenuated androgens can benefit from antifibrinolytic agents (^21,25), whose effectiveness in hereditary angioedema was demonstrated long ago (^36,80). In the present report, two-thirds of our patients needed prophylactic treatment because angioedema recurred with high frequency. In our experience, attenuated androgens have worked in only a few patients with acquired C1-INH deficiency. Antifibrinolytic agents appear much more effective and now represent our first choice for long-term prophylaxis in patients with acquired C1-INH deficiency.

Although studies have not shown a definite increased risk for thrombosis (^8,26) in patients treated with antifibrinolytic agents, there have been a few case reports of thromboembolic complications associated with this treatment (^{27,33,76,83,89}), and we reported 1 patient who developed myocardial infarction while chronically on tranexamic acid (²⁰). Thus there may be some concerns about starting this treatment in patients already presenting with increased risk of thromboembolism. Hence in 4 such patients (2 had coronary artery disease and 2 had recurrent venous thrombosis), tranexamic acid was given in association with oral anticoagulants. At this writing, the 4 patients have been on this regimen for 4 years or more without recurrence of thromboembolic events.

The treatment of life-threatening laryngeal edema is a key problem in patients with C1-INH deficiency. Replacement therapy with C1-INH concentrate has been available in different European countries for more than 20 years. Several studies demonstrated its effectiveness in patients with hereditary angioedema (^10,55,87), while acquired C1-INH deficiency patients partially resistant to this treatment have been reported (^4,11,21). Laryngeal attacks in our patients were always treated with C1-INH concentrate, and no need for invasive emergency treatment or fatalities were registered. Nevertheless, 3 patients appeared to be slow responders due to the rapid catabolism of C1-INH, and in 1 of them whose attacks were particularly severe and response to treatment slow, a permanent endotracheal cannula was applied preventively. Slow responsiveness is a high-risk condition because it seems to increase with subsequent treatment. In 1 of our patients the amount of C1-INH required to revert an attack increased 12 times, and even at such a high dosage functional C1-INH was degraded within minutes (see Figure 1). A candidate mediator of symptoms in C1-INH deficiency is the vasoactive peptide bradykinin, which is released from high-molecular weight kininogen upon kallikrein activation. In the past we used aprotinin, a bovine kallikrein inhibitor, to treat angioedema attacks as an alternative to C1-INH (⁶¹). This treatment was abandoned because of the risk of anaphylactic reactions. Dx88, a synthetic kallikrein inhibitor based on the same functional domain (Kunitz domain) of aprotinin, has been shown to revert increased vascular permeability in C1-INH knockout mice (⁴⁵). This peptide is under study for treating angioedema in patients with hereditary angioedema (⁴²). Bypassing the problem of accelerated C1-INH catabolism, this peptide represents a promising alternative for patients with acquired C1-INH deficiency who are resistant to C1-INH concentrate.