Immune defects in patients with chronic lymphocytic leukemia

Abstract

Over the past decade, the introduction of nucleoside analogs and monoclonal antibodies into the treatment of patients with chronic lymphocytic leukemia (CLL) has resulted in higher rates and longer duration of response. This is a significant step towards achieving the ultimate goal of disease-eradication and improved survival. A continuing problem, however, is the susceptibility of these patients to infections. Profound dysregulation of the host immune system in patients with CLL and its impact on the clinical course of the disease are well established. A number of investigators have sought to identify the mechanisms underlying this innate immune dysfunction, which is further exacerbated by the actions of the potent therapeutic agents. The early recognition of infections as well as prophylactic administration of appropriate antibiotics has been the mainstay of managing infections in patients with CLL. Hopefully, increasing understanding of the molecular events underlying the neoplastic change in CLL will lead to more targeted and less immunosuppressive therapeutic modalities. Furthermore, the understanding of the mechanisms of immune dysfunction in CLL is of pivotal importance in the novel immune-based therapeutic strategies currently under development.

Introduction

Infections are a major cause of morbidity and mortality in patients with CLL and it has been estimated that up to 50% of patients suffer from recurrent infections [1–3]. In addition, infections are the cause of death in 60% to 80% of patients [4–6]. Both the underlying disease and the sequelae of its treatment are responsible for this high incidence of infections. A number of disease-related factors including immunoglobulin deficiency, abnormal T-cell function, and neutropenia resulting from infiltration of the bone marrow, predispose patients with CLL to common as well as opportunistic infections (Table 1) [2]. Treatment with nucleoside analogs and monoclonal antibodies such as alemtuzumab may worsen cytopenias and exacerbate the immunological incompetence of the patient [7]. For example, some of the immunosuppressive effects of fludarabine may be mediated through its specific depletion of signal transducer and activator of transcription-1 (STAT1) protein (and mRNA), a transcription factor important in lymphocyte function [8].

Table 1.

Immune defects in patients with CLL

Immune defect
Humoral defects
Hypogammaglobulinemia
Suppression of helper T-cells
IgA deficiency
IgG subset deficiency
Deficiencies in secondary antibody response
Cellular defects
T-cell subset imbalances
Suppressed helper T-cell function
Diminished T-cell response to proliferative signals
Suppressed NK cell function
Suppressed lymphocyte activated killer (LAK) cell function
Functional and maturation defects of dendritic cells
Decreased CD4 T-cell number with fludarabine therapy
Lymphopenia with alemtuzumab therapy
Neutropenia: disease-related and drug-induced
Deficiencies of complement system
Suppressed classical pathway components
Suppressed alternate pathway components

Inherent immune defects in CLL

Hypogammaglobulinemia

A high incidence of hypogammaglobulinemia has been described in patients with CLL [4, 9–11]. Furthermore, a progressive decline in immunoglobulin levels occurs with increasing disease duration [4, 10, 12, 13]. Some patients who achieve a response when treated with fludarabine-based regimens may have an improvement in their immunoglobulin levels, but many patients, even those in complete remission, may not [14, 15].

The etiology of hypogammaglobulinemia in CLL is not clear and a number of factors such as functional abnormalities of T cells resulting in dysregulation of nonclonal CD5⁻ B cells, may be partly responsible [16, 17]. CLL cells are capable of inhibiting the interaction of activated T lymphocytes and normal B lymphocytes in vitro [18]. In particular, CLL cells can suppress the CD40 ligand-mediated helper signal necessary for normal lymphocyte differentiation and development [19]. Furthermore, CLL cells secrete TGF-β, which is a potent inhibitor of B-cell proliferation [20] and produce high levels of circulating IL-2 receptor, which can remove endogenous IL-2 and downregulate T-helper cell function [21]. Such perturbation of T-cell/B-cell interaction may be an important mechanism underlying the various immune deficiencies of patients with CLL, including humoral defects [18].

Several reports have evaluated the frequency of hypogammaglobulinemia, the type of immunoglobulin (Ig) class deficiency, and its impact on survival. Rozman et al. [10] reported that patients with an initial immunoglobulin level <700 mg/dl had a shorter survival compared to those patients with levels ≥700 mg/dl (P=0.03). This was particularly significant for reduced levels of IgG and IgA at diagnosis (P=0.027 and P=0.014, respectively), with the prognostic value of the IgA level being independent of the clinical stage on multivariate analysis. Low baseline IgM levels had no influence on survival [10]. The predominant deficiency of IgA in CLL patients may account for the increased frequency of sinopulmonary infections, as seen in patients with selective IgA deficiency. Montserrat-Costa et al. [22] reported that the most common immunoglobulin class deficiency was IgM, in contrast to Foa et al. [23] who identified IgA as the most common immunoglobulin class deficiency.

The reduction in immunoglobulin levels is likely to be an important factor predisposing patients with CLL to infections [4, 10, 24] (Table 2). However, patients with a normal serum immunoglobulin level may suffer from recurrent infections, whereas those with marked reductions of their immunoglobulins may remain infection-free. Similarly, patients responding to treatment tend to have fewer and milder infections despite having significant hypogammaglobulinemia [23–25].

Table 2.

Association between hypogammaglobulinemia and infection rate in CLL

Reference	Total	IgG level (g/l)	Patients	Severe infection (%)	P value
24a	59	<6.0	32	15	<0.001
		>6.0	27	3
12	52	<6.0	20	11c	<0.001
		>6.0	32	1c
1b	120	≤6.5	37	51	<0.001
		>6.5	83	16

^a Low levels of pneumococcal antibodies were particularly associated with severe or multiple infections (P≤0.00001)

^b Other risk factors for severe infection were advanced clinical stage (P< 0.001) and diffuse bone marrow histology (P< 0.05)

^c Patients dying from severe infections, P value calculated for survival

The relationship between reductions in specific immunoglobulin classes and the infection risk has also been investigated. A specific association between low serum IgG levels and recurrent infections has been reported [4]. More recently, Morrison et al. [26] reported a higher frequency of infections and a higher incidence of severe infections in patients with low serum IgA level as compared with those with normal serum IgA level. They were unable to detect a similar association between serum IgG and IgM levels and the infection risk [26]. In another study decreased IgA level was an independent risk factor for infection (P=0.03) and, on multivariate analysis, only a low IgA level was an independent risk factor for infection [27]. Among the IgG subclasses, decreased concentrations of IgG2 and IgG4 were associated with increased susceptibility to infection, and these correlated with IgA levels [27]. Similarly, Copson et al. [28] reported significant reductions in serum IgG3 and IgG4 and only moderate reductions of IgG1 and IgG2 levels in patients with CLL compared to those in age- and sex-matched controls. A normal IgG3 level is known to protect against infection with herpes viruses and deficiency of IgG3 may contribute to the increased risk of these infections in patients with CLL.

Deficiencies in secondary antibody response to common antigens, such as low IgG titers to Escherichia coli and S. pneumoniae have also been reported in patients with CLL [24]. Low levels of pneumococcal antibodies were reported in 23 (39%) of 59 patients with CLL compared to 6 (11%) of 56 controls (P=0.005) [24]. Such low levels were more frequently seen in patients with recurrent, severe infections than otherwise (P<0.00001) [24]. Low, specific anti-pneumococcal antibodies were not directly associated with hypogammaglobulinemia, but the majority of patients with severe or multiple infections (13 of 18) had low levels of both total IgG and pneumococcal antibodies [24].

Levels of serum and mucosal (salivary) IgA, IgG, and IgM have been measured in patients with CLL and compared to control individuals [17, 29]. Salivary IgM levels were profoundly decreased. However, there were no differences in salivary IgG or IgA levels. Furthermore, an association between salivary immunoglobulin levels and infection rate was not apparent.

Defects of cellular immunity

Several quantitative and qualitative abnormalities of the accessory non-leukemic T-cells have been reported in patients with CLL [30–36]. An increase in T-cell numbers with a reduced CD4/CD8 ratio is common, particularly in the more advanced stages of the disease [11, 34, 35, 37, 38]. Both increased and decreased numbers of the suppressor-inducer phenotype (CD45Ra) have been reported [39–42]. Other subpopulations such as NK cells, suppressors T cells, and CTLs appeared to be increased [41, 43]. The precise mechanisms responsible for the imbalance in the T-cell subsets is unclear, but a differential sensitivity of CD4 and CD8 T-lymphocytes to the Fas-expressing CLL cells has been demonstrated [44].

Several investigators have evaluated the presence of qualitative T-cell dysfunction, and a suppressed helper T-cell function has been reported [45, 46]. An excessive, normal, or reduced CD8 cell function has also been reported [38, 45]. Prieto et al. [33] reported a diminished proliferative response of T cells to several mitogenic signals, including phytohemagglutinin (PHA), anti-CD3, and phorbol. In an another study, CLL T cells had a normal proliferative response to PHA but augmentation of response by interleukin-1 (IL-1) or IL-2 was diminished in comparison with that of normal T cells [47]. T-cells dysfunction in patients with CLL is thought to be mediated by soluble factors elaborated by the neoplastic CLL cells [48]. Purified T cells from patients with CLL have neither quantitative nor qualitative differences in cytokine production and proliferative capacity as compared with those from normal donors, when stimulated by different pathways, including CD3, CD2, and costimulation with CD28. Addition of autologous accessory cells (aAC), however, dramatically influenced the cytokine pattern of normal as compared to B-CLL-derived T cells. CLL B-cells as aAC caused a marked increase of IL-2, whereas IFN-γ was only slightly induced and IL-4 was not influenced [49]. In contrast, aAC from normal individuals (predominantly monocytes) led to a significant increase of IFN-γ and a reduction of IL-4 secretion [49]. The authors concluded that T-helper cells from patients with CLL were intrinsically normal but that the predominance of the neoplastic B cells as accessory cells alters the immune function of T cells [49]. Similarly, down-modulation of CD40-ligand (CD154) on activated T cells by the CLL B cells has been reported and may account in part for the immune defects of patients with CLL [50].

A clear indication of impaired cellular immunity in CLL patients is the lack of response to recall antigens on skin testing. A prolonged skin graft survival in patients with CLL, suggestive of depressed cell-mediated immunity, has been reported [51]. Furthermore, a diminished response to skin testing for tuberculin, mumps, and Candida in patients with recurrent infections has also been shown [52]. Similarly, anergy to recall skin testing has been demonstrated [53].

Other immune effectors are dysfunctional in patients with CLL. NK-cell activity, the capacity of cytotoxic effecter cells to bind to their target, and lymphocyte-activated killer (LAK) cell functions are all suppressed [23, 54, 55, 56]. Similarly, significant maturation and functional defects of the dendritic cell (DC) compartment with an inability to stimulate an effective T-cell-mediated immune response has been demonstrated in patients with CLL [57]. Specifically, circulating DCs from patients with CLL showed defective expression of the maturation antigen CD83 and the costimulatory antigens CD80 and CD40, and had a reduced ability to release interleukin-12 (IL-12) and to drive a type-1 T-cell response [57, 58]. Similarly, they demonstrated a reduced capacity to stimulate the proliferation of allogeneic T cells. Furthermore, the addition of CLL peripheral blood mononuclear cells (PBMNCs) to DCs generated from normal monocytes induced similar markers of abnormal maturation and functional impairment and resulted in an inhibition in the expression of costimulatory molecules [57]. The addition of CLL cells was able to modify the cytokine production profile of normal DCs with a reduction in the number of IL-12-producing DCs, and a concomitant increase in the ability to produce IL-10. Furthermore, purified CD1a+ DCs, which matured in the presence of CLL PBMNCs, demonstrated a reduced allostimulatory capacity, as compared with the same DCs cultured in the presence of control medium [57]. The addition of monoclonal antibodies against IL-10, VEGF, and in particular IL-6 reverted the inhibitory effects of CLL cells in a proportion of patients, suggesting that the defects observed in the in vivo DC compartment are, at least in part, because of factors derived from the leukemic clone.

The inhibitory effects of the neoplastic CLL cells on the in vivo DC compartment is also suggested by some reports, suggesting the feasibility of generating apparently normal monocyte-derived DCs in vitro, even in patients with advanced CLL [59, 60]. However, these cultured CLL DCs were found to release an abnormal pattern of cytokines and expressed reduced amounts of CD40 and CD80 co-stimulatory molecules [58–60]. It is of significant interest, however, that CLL patients in remission showed no significant difference in the number of large DCs generated or in the expression of surface markers, including co-stimulatory molecules. The expression of both CD80 and CD40 was significantly higher in DCs-derived CLL patients in remission than in those derived from patients with active disease[58]. Therefore, monocyte-derived DCs from CLL patients with active disease showed phenotypic abnormalities that were no longer detectable after achieving remission following therapy [58]. The clinical influence of these immune defects has not been fully elucidated. However, it is likely that these abnormalities will have to be taken into consideration in the design of trials of immunotherapy and vaccination in patients with CLL.

Phagocytic cell defects

Although neutropenia is not a feature of early stages of CLL, disease progression and treatment with chemotherapy contribute to its development [5]. Factors leading to development include marrow infiltration by disease, chemotherapy-related myelosuppression, and suppression of neutrophil progenitors by disease-related cellular and humoral factors [5, 61, 62].

A deficiency of neutrophil function with associated enzyme deficiencies (lysozyme and myeloperoxidase) has been reported by some authors [63]. Such functional abnormalities of granulocytes have been thought to predict infection risk [63, 64]. Similarly, a reduction in monocyte number has been suggested to be a risk factor for infection in CLL [65].

Defects of the complement system

Abnormalities of the complement system have also been reported in patients with CLL [66–68]. Fust et al. [69] measured the components of the classic complement pathway in 46 patients with CLL and reported a reduction in mean C1 and C4 levels in more than 50% of the samples tested. No abnormalities of the alternative complement pathway were noted. In another study, depressed classic complement pathway activity in CLL was noted [70]. Low hemolytic activity of the classic complement pathway has been correlated with decreased survival in patients with CLL [71].

Schlesinger et al. [66] examined the complement system in patients with CLL and their relatives. They measured the levels of serum complement proteins over a 2-year period and found lower serum levels than in a group of sex- and age-matched healthy controls (P<0.0001). The decreased levels remained constant throughout the follow-up period and involved classic and alternate pathway components [66]. A reduced level of properdin was most frequently seen (11 of 18 patients). Complement defects correlated with disease stage, occurring in all patients with stages II–IV disease and only in 40% of patients with early stage disease (P<0.004). No correlation was found between incidence and type of infections and complement deficiency, probably because of the small sample size [66]. No relationship has consistently been established between complement abnormalities and the risk of infection in patients with CLL.

Disease duration and age

CLL is a disease of the older adults with a median age at presentation of about 66 years [72]. The increased incidence of immune dysfunction associated with aging is well described [73, 74]. Although immune responses against recall antigens may still be conserved, the ability to mount primary immune responses against new antigens declines significantly with advancing age. This results in a high susceptibility to infectious diseases and may limit the efficacy of vaccination strategies in elderly people. As such, elderly patients with CLL may have an inherent susceptibility to infections which is accentuated by their disease.

The immune incompetence associated with CLL, which is secondary to the profound hypogammaglobulinemia and the impaired cell-mediated immunity to recall antigens, is directly linked to the disease duration and eventually develops in all patients [1, 5]. Molica et al. [1] examined the frequency of infections in 125 patients with CLL over a 10-year period. They reported a total of 199 infections, including 47 severe infections (9.8 per 100 person-years), and a 5-year risk for severe infections of 26% (95% confidence interval: 24.7–27.3%) Twenty-one of 71 (29.5%)) of deaths were directly attributable to infections. However, the increased risk of infection with time is confounded by the greater likelihood of therapeutic intervention and the immunosuppression caused by treatment.

Treatment-related defects of immune function

Purine analogs

The use of nucleoside analogs, in particular fludarabine, in the treatment of patients with CLL is now well established [75]. Treatment with fludarabine is associated with exacerbation of immune defects in patients with CLL and further predisposes them to infections (Table 3) [8]. The introduction of fludarabine has been associated with a different spectrum of infections than those classically encountered in this population, including opportunistic infections with Pneumocystis carinii, Listeria monocytogenes, Mycobacterium tuberculosis, and herpesviruses [76–80]. This is likely related to the selective T-cell abnormalities induced by this agent as well as its myelosuppressive properties [81].

Table 3.

Myelosuppression and infection risk using purine analog therapy in CLL

Agent	Reference	Patients (no.)l	% of patients developing grade 3/4 neutropeniam	% of patients developing grade 3/4 infectionm
Fludarabine or fludarabine-based combinations	130	77a	39	29
	131	35 [6]	5	27
	132	68 [33]	56c [53]	17 [9]
	85	100 [52]b	19	4
	87	188d	10	29
	87	141d,e	19	45
	88	128 [35]f	48c	50
Pentostatin or pentostatin-based combinations	133	26 [13]	3	34
	134	26	Median ANC day 21:1,200/μl	16
	94	29	17	31g
	92	23h	35	9
Cladribine	135	26	31	27
	136	90 [20]	NS	21 [30]i
	137	67 [46]j	12	37
	138	43 [33]k	10 [15]	20 [15]
	139	52	63	27

ANC Absolute neutrophil count; NS not significant

^a Eighteen patients with non-Hodgkin’s lymphoma included

^b Randomized to fludarabine versus cyclophosphamide/doxorubicin/prednisone

^c ANC <500/l

^d Randomized to fludarabine versus chlorambucil versus fludarabine and chlorambucil

^e Fludarabine in combination with chlorambucil

^f Fludarabine in combination with cyclophosphamide

^g Grade 2–4 infections reported

^h Pentostatin in combination with cyclophosphamide

ⁱ Three atypical infections in patients receiving steroid

^j All <55 years

^k All >70 years

^l No. of patients with no prior therapy given in []

^m Percentage of previously untreated patients given in []

Treatment with fludarabine based regimens leads to a profound decrease in CD3 cells with a fall in both CD4 and CD8 counts [15, 81]. Wijermans et al. [81] studied T-cell subpopulations during and after therapy with fludarabine by flow cytometry. A pronounced decrease in the various T-cell subpopulations was seen in all patients evaluated; the CD4 lymphopenia was still present one year from the end of the therapy. CD8 and natural-killer (NK) cell numbers normalized from elevated pre-therapy levels [81]. Bergmann et al. [82] observed a decrease in CD4 cells from a median of 2,935 cells/μl to 1,316 cells/μl in 10 patients with refractory CLL, with the median CD8 cell count decreasing from 5,281 cells/μl to 1,131 cells/μl. The CD4/CD8 ratio remained unchanged. O’Brien et al. [15] evaluated the CD4 count in 217 patients with CLL treated with fludarabine and prednisone. The median CD4 count before therapy was 1,015 cells/μl (range, 47–17,791 cells/μl); decreasing to 169 cells/μl (range, 0–4,669 cells/μl) after 3 months of therapy, and to 148 cells/μl (range, 6–1,908 cells/μl) after 6 months. Juliusson et al. reported that CD4 and CD8 levels in 68 patients with CLL were treated with cladribine with the median baseline CD4 level reported as 660 cells/μl (range, 409–1,169 cells/μl) falling to 399 cells/μl (range, 325–486 cells/μl) at 12 months. Corresponding levels for CD8 lymphocytes were 546 cells/μl (range, 300–875 cells/μl) and 429 cells/μl (range, 301–661 cells/μl) [83].

Use of corticosteroids, either concomitantly or prior to or after fludarabine therapy, is associated with an increased risk for these infections [77, 80]. Anaissie et al. [77] reviewed their experience with 402 patients with CLL, who received fludarabine with or without prednisone. The incidence of infections was higher in previously treated (58%) than untreated (34%) patients (P<0.001). Infections with L. monocytogenes or P. carinii occurred in 12 (7%) of 170 previously treated patients who received fludarabine plus prednisone, in none of 78 previously treated patients who received fludarabine alone, and in 2 (1%) of 154 previously untreated patients who received the combination (P=0.003). Other factors in addition to the low CD4 count associated with therapy with fludarabine predispose to these infections. Rai stages III and IV, prior therapy, and elevated serum creatinine concentration were identified as significant independent risk factors for major infection. A baseline granulocyte count greater than 1.0×10⁹/l was protective (odds ratio, 0.54; 95% CI: 0.29 to 0.99). Five (26%) of 19 patients with CD4 counts <50 cells/μl, compared with 9 (6%) of 139 patients with a CD4 count >50 cells/μl, experienced cutaneous zoster (P=0.01). These data suggested a correlation between very low CD4 counts and reactivation of herpes [77]. A number of strategies to reduce the incidence of these infections have been proposed [77].

Several randomized trials have compared fludarabine with alkylating-agent based regimens used to treat patients with CLL [84–86]. The impact of therapy with chlorambucil, fludarabine, or the combination of the two drugs on the incidence and spectrum of infections was evaluated among a series of previously untreated patients with CLL [87]. Five hundred and fifty-four patients with intermediate/high risk disease were randomized into the above three treatment arms. Concomitant therapy with corticosteroids was prohibited but prior use of steroids for autoimmune phenomena was allowed. A total of 1,107 infections including 241 major infections occurred in 518 patients over a period commencing at study entry and lasting until reinstitution of the initial therapy, initiation of therapy with a second agent or occurrence of death [87]. Patient receiving the combination of chlorambucil and fludarabine had more infections than those receiving either drug alone (P<0.0001). The mean number of infections per month of infection follow-up was also significantly higher for patients treated with the combination (P=0.008). When the single agent fludarabine and chlorambucil arms were compared, the incidence of infections, including major infections, was higher in the fludarabine-treated patients (P=0.055) [87]. The spectrum of infections in fludarabine-treated patients included common bacterial as well as opportunistic infections. The majority of viral infections were herpes virus infections and these were more likely to occur in the combination arm (P<0.0001). Herpes virus infections including varicella zoster were more frequent in fludarabine-treated patients compared to chlorambucil-treated patients (P=0.0006). Three patients developed mycobacterial infections and only one case of legionella infection was reported in a patient treated with the combination. Similarly, fungal infections were uncommon with the majority of infections being caused by candida species in patients treated with fludarabine, either alone or in combination [87]. The increased risk of opportunistic and viral infections in fludarabine-treated patients is clearly suggestive of a further decline in T-cell-mediated immunity following therapy with this agent.

Combinations of fludarabine and alkylating agents such as cyclophosphamide with or without monoclonal antibodies have been recently investigated in an effort to increase the response rates [88, 89]. Such combination regimens have the propensity for an increased incidence and severity of myelosuppression and as such may be expected to exacerbate the immune dysfunction associated with the disease and therapy. In the study by O’Brien et al. [88] combining fludarabine and cyclophosphamide, neutropenia to less than 500×10⁹/l was noted in 48% of patients who received cyclophosphamide 300 mg/m² . Pneumonia or sepsis occurred in 25% of patients, and fever of unknown origin occurred in another 25%. This was likely related to the significant myelosuppression induced by the regimen. Six atypical infections were noted, including one P. carinii pneumonia, cryptococcal bronchitis, cryptococcal meningitis, Vibrio sepsis, strongyloides, and cytomegalovirus. Two of those patients were also receiving corticosteroids. Seven patients (5%) developed herpes zoster infections, and 10 patients (8%) had reactivation of herpes simplex. Infections were significantly more frequent in patients who were refractory to fludarabine at the start of combination chemotherapy [88].

A similar incidence of myelosuppression and infections has been reported with other nucleoside analogs, including cladribine and pentostatin, alone or in combination with alkylating agents (Table 3) [90–92]. Van Den Neste et al. [93] compared the incidence of infections during 6 months preceding treatment with cladribine with the 6 months following therapy in 95 patients, and showed a doubling of the infection rate after treatment with cladribine. Robak et al. [90] conducted a randomized, prospective trial comparing the efficacy and toxicity of cladribine and prednisone with chlorambucil and prednisone. Drug-induced neutropenia was more frequently observed in patients treated with cladribine and prednisone than those treated with chlorambucil and prednisone (P=0.02) [90]. Furthermore, infections occurred more frequently in the cladribine-treated group (P=0.02) [90]. A similarly increased incidence of infections has been noted with the use of pentostatin, despite the perceived diminished myelosuppression with the single agent [92, 94]. Weiss et al. [92] used a combination of pentostatin and cyclophosphamide to treat 23 patients with relapsed CLL. Despite the routine use of filgrastim, grade 3 or 4 neutropenia was reported in 35% of patients. Fifteen patients developed infections but only two had a grade 3 or 4 infection [92].

Monoclonal antibodies

Monoclonal antibodies are increasingly used to treat hematological malignancies and solid tumors. Their attractiveness stems from their ability to target specific antigens expressed preferentially on the malignant cells. As such, their therapeutic efficacy is minimally hampered by non-specific toxicity. A number of monoclonal antibodies have been evaluated in treating patients with CLL [95].

Rituximab is a chimeric human-mouse monoclonal antibody directed against the CD20 antigen present on the surface of B cells. Treatment with standard dose rituximab is associated with a significant reduction in the peripheral B-lymphocyte count within 3 days (by approximately 90%) followed by a slow recovery over 9–12 months [96–98]. Interestingly, there is a mild decrease in serum immunoglobulin levels in up to 20% of patients, and there is no effect on serum complement. Isolated but reversible neutropenia of unclear etiology may develop after therapy with rituximab [99]. A direct toxic effect of rituximab on granulocytes or myeloid precursors is unlikely as they do not express the CD20 antigen. IgG-type antibodies bound to the surface of neutrophils have been described in some patients and an autoimmune mechanism related to the altered immune repertoire of the patients has been proposed [99]. Single agent rituximab has been used in treating patients with CLL and a dose-response correlation has been demonstrated [100, 101]. Myelosuppression and infection are generally uncommon in patients treated with rituximab alone despite the higher dose employed in some studies [100, 101]. In fact, the incidence and severity of infections do not appear to be different from those in patients with CLL not receiving rituximab. Rituximab has also been administered in combination with a number of chemotherapy agents in an effort to capitalize on potential synergistic activity [89, 102–104]. The addition of rituximab does not appear to significantly affect the incidence or severity of infections, although an increased incidence of grade 3 or 4 neutropenia has been reported [102].

Alemtuzumab is a humanized monoclonal antibody targeting the CD52 antigen, which is expressed on the surface of all lymphocytes at various stages of differentiation [105]. The administration of alemtuzumab results in severe lymphopenia with a reduction in normal blood B and T cells in all patients. The CD16+ NK cells and CD14+ monocytes decrease marginally [106]. Osterborg et al. [107] treated 29 patients with CLL, who had relapsed after an initial response or were refractory to chemotherapy with alemtuzumab administered as a 30-mg 2-h intravenous infusion thrice weekly for a maximum period of 12 weeks. Lymphopenia (<0.5×10⁹/l) occurred in all patients. The ensuing immunosuppression predisposed patients to a number of infections with the most common being viral, including varicella-zoster virus (VZV), herpes simplex virus (HSV) and CMV. Other infections include those caused by common bacterial pathogens, L. monocytogenes, invasive fungi, and atypical mycobacteria [108]. In the pivotal multicenter US study, 93 fludarabine-refractory patients with CLL were treated with alemtuzumab. Overall, 51 patients (55%) developed at least one infection during the course of the study; grade 3 and 4 infections occurred in 26.9% of patients [109]. Eleven patients, all with advanced disease, developed opportunistic infections, including P. carinii pneumonia, invasive fungal infections, viral infections and listeria, with some experiencing multiple episodes. A further of seven opportunistic infections occurred in the follow-up period. The most commonly reported opportunistic infection was CMV reactivation, which occurred in 7 patients during treatment. No cases of CMV were reported in the post-study period. Not all opportunistic infections were associated with concomitant use of systemic corticosteroids, but seven of the above cases occurred in patients who were treated with steroids for longer than 5 days. Furthermore, patients who developed opportunistic infections were, in general, heavily pretreated [109].

Rai et al. [110] reported their experience in 24 previously treated patients with CLL, who received intravenous alemtuzumab. Lymphocyte subset analysis by flow cytometry was performed before the start of alemtuzumab in 19 patients. These analyses were repeated in varying numbers (13 to 15) of patients at 4 weeks, at the end of alemtuzumab therapy, and again 28 days after this therapy. A profound decrease in B cells, T cells, and natural killer cells was noted at week 4 and at the end of treatment, but there was evidence of recovery of T cells in some patients at the end of treatment, and this recovery seemed to continue when patients were tested again 28 days later. At that time, there was also evidence of a re-emergence of B cells and of CD52⁺ lymphocytes in some patients. Ten patients (eight nonresponders and two responders) experienced major infections on-study [110].

The incidence of infections with alemtuzumab appears to be significantly lower in previously untreated patients. Lundin et al. [111] evaluated the efficacy and safety of alemtuzumab delivered subcutaneously as first-line therapy, over a prolonged treatment period of 18 weeks in 41 patients with symptomatic, previously untreated, B-cell CLL. Transient grade IV neutropenia developed in 21% of the patients. Infections were rare, but 10% of the patients developed cytomegalovirus (CMV) reactivation. These patients rapidly responded to intravenous ganciclovir. One patient, allergic to co-trimoxazole prophylaxis, developed P. carinii pneumonia [111]. In a further analysis, Lundin et al. [43] presented the results of their long-term follow-up data for peripheral blood lymphocyte subsets analyzed by flow cytometry in this group of patients. All lymphoid subsets were significantly (P<0.001) and profoundly reduced; the median end-of-treatment counts for CD4+, CD8+, CD3–56+ (natural killer/NK), CD3+56+ (NK-T) and CD19+5- (normal B) cells were 43, 20, 4, 1 and 8 cells/μL, respectively. Granulocytes and monocytes were also significantly (P<0.01) depleted from a median of 4.1×10⁹/l (range 0.6–12.6×10⁹/l) at baseline to 2.2×10⁹/l (range 0.7–5.7×10⁹/l) at the end of therapy, and from 0.6×10⁹/l (range 0–7.3×10⁹/l) to 0.1×10⁹/l (range 0–0.7×10⁹/l), respectively. Granulocytes returned to baseline levels early during the follow-up.

The recovery of immune cells was slow and the median cell count of all lymphoid subpopulations remained at less than 25% of the baseline values for greater than 9 months post-treatment. Thereafter, cell numbers increased more rapidly except for normal B cells, which remained at a low level at 18 months. At 4 months after the completion of treatment, CD4+ and CD8+ levels in the blood had reached >100 cells/μL in more than half of the patients. One patient had CMV reactivation and one patient developed P. carinii pneumonia during therapy [43]. No opportunistic or other major infections were recorded during unmaintained, long-term follow-up. The cumulative dose of alemtuzumab did not correlate with the severity or the duration of immunosuppression. T-cell subsets lacking CD52 expression were found during the treatment and comprised more than 80% of all CD4+ and CD8+ cells in the blood at the end of therapy. These subpopulations declined gradually during unmaintained follow-up [43].

Several recent studies have evaluated the role of alemtuzumab in consolidation of the response to fludarabine-based regimens [112–114]. In the study by O’Brien et al. [113], alemtuzumab after chemotherapy improved the response rates in approximately 50% of the treated patients. Reactivation of CMV was also reported in 22% of patients with no correlation between the dose of alemtuzumab and risk of reactivation. One patient died of CMV-related complications [113]. The German CLL study group terminated their randomized study as a result of a significant increase in the incidence of opportunistic infections, in particular CMV reactivation, despite a significant improvement in responses in the alemtuzumab arm [112]. Overall the incidence of severe infections appeared to be higher in the studies that administered alemtuzumab shortly after the completion of fludarabine therapy [112, 114]. On the contrary, when there was a longer time interval elapsing before the initiation of alemtuzumab therapy, serious infections including CMV were less frequent [113]. Therefore, treatment with alemtuzumab is associated with a dose-independent and long-lasting depletion of lymphocyte subsets that are likely to exacerbate the pre-existing immune deficiency. Sequential therapy after fludarabine may be associated with a more profound immunosuppression and should be carefully monitored.

Stem cell transplantation

Both autologous and allogeneic transplantation have been utilized in the treatment of CLL, often in the setting of advanced disease [115–118]. However, published reports have not described an increase in opportunistic infections beyond that typically noted in the transplantation setting [116, 119]. A detailed description of the immune defects associated with transplantation in patients with CLL is provided elsewhere in this symposium.

Splenectomy and splenic irradiation

Indications for splenectomy in the setting of CLL include autoimmune hemolytic anemia, immune thrombocytopenia, splenomegaly, and cytopenias caused by hypersplenism. Asplenia is associated with an increased risk of infection by encapsulated bacteria (e.g., S. pneumonia, Haemophilus influenzae), Neisseria species, and Capnocytophaga canimorsus (formerly DF-2 bacillus), as well as by intraerythrocytic parasites (e.g., Babesia microti) in certain geographic areas. Other organisms causing infections in asplenic patients include group B streptococci, S. aureus, Enterobacteriaceae, and even L. monocytogenes. However, S. pneumoniae is by far the most frequent and serious pathogen in splenectomized patients, accounting for 50% to 90% of infections and associated with a mortality of up to 60% [120, 121]. Therefore, loss of splenic function through splenectomy or splenic irradiation further contributes to the infection risk in patients with CLL.

Use of growth factors, IVIG, and prophylactic antibiotics in CLL

Hematopoietic growth factors

Several small studies have suggested that growth factor therapy is effective in reducing the length and severity of neutropenia associated with disease progression and its treatment, and as such may be beneficial in patients with CLL. In a study by O’Brien et al. [122], 25 previously treated patients with advanced-stage CLL were treated with monthly courses of fludarabine supported with granulocyte colony-stimulating factor (GCSF). There was a significant reduction in the incidence of neutropenia (absolute neutrophil count <1×10⁹/l) in the growth factor–treated group compared to matched historical controls (45% vs. 79%; P=0.002). This translated into a reduction in the incidence of delay in therapy (P=0.005) and the incidence of pneumonia (8% in the treated group versus 37% in the controls). The incidence of other infections (sepsis, fever of unknown origin, and minor infections) was similar in the two groups [122]. Other investigators have used GCSF routinely to reduce the severity and duration of neutropenia associated with the combinations of nucleoside analogs and alkylating agents [92, 123]. Although these reports suggest a possible beneficial effect from the prophylactic administration of GCSF, in the absence of data from randomized studies, routine use of growth factors has not been established.

Use of IVIG in CLL

Several early reports of administration of intravenous immunoglobulin (IVIG) in patients with CLL suggested a beneficial effect in correcting the immune deficiency and reducing the risk of infections [124, 125]. In a randomized, controlled, double-blind, multicenter trial comparing placebo with IVIG 400 mg/kg body weight every 3 weeks for 1 year in 81 patients with B-CLL and either hypogammaglobulinemia or history of prior infection [126]. Administration of IVIG was associated with a significantly reduced risk of bacterial infections (23 compared to 42; P=0.01). However, there was no significant difference in the incidence of viral and fungal infections and the survival duration between the two groups [126].

Most studies confirm the effectiveness of IVIG (in various doses) in reducing the incidence of bacterial infections in patients with CLL. Nevertheless, the expense of such therapy has generally limited the widespread use of this strategy. Furthermore, it can be argued that rapid institution of effective therapy with antibiotics would be a more cost-effective method of managing the immune deficit in patients with CLL. A widely utilized approach is to use IVIG in patients with hypogammaglobulinemia and repeated sino/pulmonary infections that are poorly controlled with antibiotic therapy or recur after it.

The scarcity of reliable data on prophylactic administration of antibiotics to patients with CLL has limited its universal application. However, a pre-emptive strategy of supplying anti-bacterial and anti-viral antibiotics for rapid initiation of therapy may reduce the morbidity of bacterial and viral infections.

Conclusions and perspectives

Progress in understanding the biology of CLL, as well as the development of more effective therapeutic strategies have led to significant increases in the rate and duration of responses. Although the modern regimens used in treating patients with CLL are capable of depleting the malignant lymphocyte population, they lack specificity and further contribute to the significant immune dysfunction that predisposes these patients to a serious infection risk. As such, the role of supportive care, already a pivotal factor in treating patients with CLL has been further emphasized. A number of strategies, in particular the use of effective prophylactic antibiotics, close monitoring for identification of infections such as CMV and their rapid treatment have become a common practice.

Passive immunotherapy has recently been established as a viable treatment modality for patients with CLL, particularly through the introduction of a number of monoclonal antibodies directed at antigens expressed on the surface of CLL lymphocytes. Combinations of these antibodies with chemotherapy have been employed successfully to further improve on the response rates and to eradicate minimal residual disease. Active immunotherapy using innovative approaches to direct the host immune system to target and eliminate the malignant B-cell clone is also under investigation, and a number of strategies to exploit the distinctive pattern of surface molecules and leukemic cell “altered self-antigens” as targets for an autologous or allogeneic immune response are being explored [127]. CLL B-cells have the inherent capability to evade autologous immune responses. However, leukemic cells can be effectively used as antigen-presenting cells to mediate proliferative T-cell responses and expansion of autologous T cells for immunotherapy is under investigation [128, 129].

The successful application of such strategies would require navigation of the inherent immune defects associated with the disease and its progression. Numerical and functional deficiencies of T cells and antigen-presenting cells in patients with CLL, down-modulation of CD40-ligand (CD154) expressed on activated T cells by CLL cells, as well as lack of expression of important costimulatory molecules such as CD80 by resting CLL B-cells are all factors that are likely to provide obstacles for the generation of meaningful vaccine-based therapies. Furthermore, most current treatment regimens deplete the components of immune system in CLL patients. Therefore, application of autologous T cells in immunotherapy of CLL patients would necessitate selection and expansion of these cells prior to initiation of therapy.

CLL B-cells have a negative impact on the generation of functional DCs, both in vivo and in vitro. In contrast, monocytic DCs generated from CLL patients in remission did not show differences compared with those obtained from normal donors [58]. These findings suggest that immunotherapeutic strategies that are dependent on the generation of autologous responses are likely to be more effective after the achievement of a low-leukemic burden. Furthermore, T-cell therapy can be used to reconstitute the defective immune profile of patients following therapy. Altogether, debulking of leukemia burden by combination chemotherapy, use of effective infection prophylaxis, and application of immune-based therapeutic strategies to eradicate minimal residual disease will likely lead to the achievement of the goal of “cure” in CLL.