The median lymphocyte count at diagnosis is 20 to 30 × 10⁹/L, and in most patients, there is a continuous increase in the lymphocyte count over time.¹³, ²⁴³, ²⁴⁴ In half the patients, it takes more than 12 months for the lymphocyte count to double; cyclic fluctuations of up to 50 × 10⁹/L can occur in the lymphocyte counts of untreated patients, and in others, the count may remain stable for years.²⁴⁴ The CLL cells are small to medium-sized lymphocytes with clumped chromatin, inconspicuous nucleoli, and a small ring of cytoplasm. Cytoplasmic inclusions occasionally may be observed in CLL cells and may be crystalline, globular, tubular, or rod-shaped.²⁴⁸ Smudge cells (basket cells or shadow cells of Gumprecht) are commonly seen in the peripheral blood smear in CLL, but not in other lymphoid disorders, and are caused by a decrease in the cellular content of vimentin, a cytoskeletal protein required for maintaining cell structure.²⁴⁹, ²⁵⁰ The number of smudge cells can vary from 1% to 75% (median 28%) and appears to be remarkably consistent in individual patients.²⁵⁰ One study suggests that patients with ≥30% smudge cells are more likely to have mutated IgV_H and have a better prognosis than those with <30% smudge cells.²⁵⁰ There can be variations in cell morphology, with some cells being PL, whereas others are larger with abundant cytoplasm, and some are plasmoid or cleaved.²⁰⁴, ²⁵¹ The French/American/British classification system divides patients into 3 groups depending on the percentage of abnormal cells.²⁵¹ In classical CLL, >90% of cells are small, and when 11% to 54% of the cells are PL, it is termed CLL/PLL. When >15% of the lymphocytes are plasmoid or cleaved and <10% are PL, it is termed atypical CLL.²⁰³, ²⁰⁴, ²⁵¹ Approximately 80% of patients have classical CLL, and 20% have CLL/PLL or atypical CLL. If ≥55% of the cells are PL, the patient has PLL.

Marrow infiltration in CLL may be interstitial, nodular, mixed (nodular and interstitial), or diffuse, with mixed being the most common and nodular the least common.³, ²⁵² Diffuse involvement, in which there is effacement of the fat spaces by tumor, carries the worst prognosis.³, ²⁵² The marrow involvement is random and contrasts with follicular lymphomas, in which paratrabecular involvement is the rule. In contrast to marrow, involvement of the lymph node is diffuse. Proliferation centers with PL and paraimmunoblasts are typically seen in both marrow and lymph nodes.

CLL can usually be readily differentiated from other disorders by immunophenotyping¹, ², ³, ²⁵³, ²⁵⁴, ²⁵⁵ (Table 90.2). The leukemic cells have the B-cell markers CD19, CD20 (low), CD43, and CD79b (low) and must be CD5⁺. In addition, the cells show clonal light chain restriction, weak expression of sIgM and sIgD, and are CD23⁺ and CD10^–. The cells are also CD27⁺,⁴³ consistent with being memory B-cells.⁵ Alternatively, the presence of CD27, CD5, and CD23 could reflect the activated nature of the CLL cell.⁴² Matutes et al.²⁵⁴, ²⁵⁶ recommend 5 markers to differentiate CLL from other B-cell malignancies (Table 90.3). Typical CLL should be surface Ig (weak), CD5⁺, CD23⁺, CD79b or CD22 (weak), and FMC7^–.

One of the typical features of the CLL cell is the overexpression of membrane CD23, which is related to deregulation of NOTCH2 signaling and may play a role in the decreasing apoptosis.²⁵⁷, ²⁵⁸ The presence of CD23 is useful to differentiate CLL from mantle cell lymphoma, which is also CD5⁺.²⁵⁷ Cell surface CD23 undergoes spontaneous proteolysis, producing elevated serum levels of CD23, the level of which is a marker of disease stage and progression.²⁵⁷

The BCR complex is required for the proliferation of B-cells after immune stimulation and is a complex formed by sIg and Igα/Igβ (CD79a/CD79b). CLL cells lack CD79b, which is related to overexpression of an alternatively spliced form of the gene.²⁵⁹ The FMC7 antibody identifies an epitope of CD20 and usually strongly stains hairy cell leukemia and PLL; however, only 16% of CLL cases stain positively, presumably because CD20 is only weakly expressed in typical stable CLL.²⁵³, ²⁶⁰ In general, those patients who are FMC7⁺ have high levels of surface IgM, low expression of CD23, and poor prognosis.²⁵³ Although the myelomonocytic antigens (CD11b, CD13) may be expressed in multiple myeloma, acute lymphoblastic leukemia, and in the CD5- chronic lymphoid leukemias, they are not expressed in CLL.²⁶¹

The CD5 antigen is most commonly associated with mature T-cells and is expressed weakly on thymocytes.², ²⁶² However, normal B-cells carrying the CD5 marker are located in the mantle zone of the lymph node, and small numbers of these cells are also present in the peripheral blood.²⁶² The CD5 molecule has been cloned and appears to be involved in the activation of T lymphocytes. The function of CD5 on the B-cell remains unknown, but its induction has been shown to be inhibited by the T-cell-derived cytokine IL-4. CD5⁺ B-cells can be stimulated to secrete anti-DNA antibodies and rheumatoid factors in vitro, and increased numbers are found in the peripheral blood of patients with rheumatoid arthritis, systemic lupus erythematosus, and Sjögren syndrome.²⁶² It has been suggested that the normal counterpart of the CLL cell is the CD5⁺ B-cell,¹⁹ although detailed immunophenotyping demonstrated significant differences between the two groups.²⁶³ Thus, the CD5⁺ B-cell appears to be a resting antigen-naive cell, whereas the CLL cell is activated and antigen-experienced. Moreover, the specific activation marker expression pattern varies between cells that have IgV_H gene mutations and those that do not, with the unmutated group resembling B-cells at an earlier state of activation.⁴¹ Older studies suggested that about 5% of cases of CLL could be CD5^– and were more likely to be FMC7⁺, CD23^–, CD11b⁺, and CD13⁺ and have a poor prognosis.²⁶¹ However, these CD5^– cases should be reclassified, as CD5⁺ is essential for the diagnosis of CLL.¹, ³

Both CD38 and ZAP-70 have been evaluated as surrogate markers for IgV_H mutational status, as these can be measured readily by flow cytometry. To define the stage of maturation of the CLL cell, Damle et al.⁴² measured CD38 levels in CLL. Approximately 50% of patients had >30% CD38⁺ cells, and these patients had unmutated IgV_H and had a worse prognosis than those with <30% CD38⁺ cells. However, not all have found such a correlation.² CD38 acts as a receptor and enzyme and can produce cell replication and survival with a variety of signals.²⁶⁴ Activation of CD38⁺ cells, but not CD38^– cells, through sIgM induces apoptosis and through IgD prolongs cell survival and induces differentiation.²⁶⁴ In addition, recent evidence suggests that the replicating cells in CLL are the CD38⁺ cells and the number of CD38⁺ cells is higher
in lymph nodes, likely reflecting the higher rate of cell proliferation at that site.¹¹⁵, ²⁶⁵ Although CD38 can be measured easily by flow cytometry there is disagreement about the number of cells that are required to define positivity, with values ranging from 5% to 30%.²⁶⁶, ²⁶⁷, ²⁶⁸ Moreover, the number of CD38 positive cells can vary over time.²⁶⁶, ²⁶⁷, ²⁶⁸

ZAP-70 is a member of the Syk-ZAP-70 protein kinase family and is expressed in T and natural killer (NK) cells and is important in T-cell signaling.²⁶⁹, ²⁷⁰ However, recent evidence suggests that normal B-cells may also express ZAP-70, particularly when activated.²⁶⁹, ²⁷⁰ Gene expression studies in CLL have also demonstrated that cells with unmutated IgV_H have an increased expression of ZAP-70 whereas those with mutations have low levels.⁷ Subsequent studies have shown a 70% to 90% correlation between ZAP-70 expression and IgV_H mutational status, regardless of whether ZAP-70 was measured by flow cytometry (≥20% cells positive), Western blot analysis, or immunohistochemistry.²⁷¹, ²⁷², ²⁷³, ²⁷⁴ Kröber et al.²⁷⁵ have shown that discordant cases may have poor prognostic features including deletions of 17p13 or 11q22-q23, or IgV_H3-21 expression. ZAP-70 positivity also correlates moderately with CD38 positivity and the presence of poor-risk cytogenetics, that is, deletion 11q22-q23, deletion 17p13, and trisomy 12.²⁷⁵, ²⁷⁶ The poor prognosis associated with ZAP-70 positivity has been related to the magnified signaling by the BCR in ZAP-70-positive CLL cells.²⁷⁷ A major difficulty presently in the routine use of ZAP-70 is the variability in the assay as a result of the sensitivity of different antibodies and the use of different internal controls.²⁶⁶, ²⁶⁷, ²⁶⁸ Moreover, although initial studies suggested that ZAP-70 differed from CD38 in remaining stable during the disease course, more recent evidence indicates that ZAP-70 may also change over time.²⁶⁷, ²⁶⁸

Apart from having either κ or λ light chains on the cell-surface membrane, clonality is confirmed by the presence of unique idiotypic specificities of the Igs produced by CLL cells²⁷⁸ and by IG gene rearrangements.², ¹⁸ CLL cells have low to undetectable amounts of monoclonal polyreactive IgM autoantibodies, frequently of the rheumatoid factor type, on their surface.²⁶², ²⁶³ Thus, the same monoclonal autoantibody produced by a leukemia cell may react with a variety of antigens (e.g., IgG, cardiolipin, histones, or single- or double-stranded DNA).²⁶² Surface and cytoplasmic Igs contain either κ or λ light chains but never both. Most cells display a single heavy chain class, usually µ, although some display µ and δ. Less commonly, γ, α, or no heavy chain determinant is found. Although the CLL cell is believed to be frozen at a particular stage in maturation, up to 50% of CLL patients have IgM⁺ leukemia cells that are able to undergo isotype switching to IgA (usually) or to IgG.²⁷⁹, ²⁸⁰ Rearrangements of the IGH and IGL genes are seen in CLL and these remain constant over time.² Interestingly, the T-cell receptor β gene can be aberrantly re-arranged in 6% of CLL cases and this is associated with chromosome 6q deletions.²⁸¹, ²⁸²

CLL cells may secrete idiotypic IgM and in most cases can be induced by mitogens to secrete IgM that can react to a variety of autoantigens.²⁸³, ²⁸⁴ Using a sensitive immunoblotting technique, monoclonal proteins can be detected in the serum of virtually all patients although the light chain of the protein is only the same as that on the CLL in half the cases.²⁸⁵ These findings indicate that the monoclonal protein in many cases is not derived from the tumor cells. Serum electrophoresis detects a monoclonal protein in approximately 5% of CLL cases and an abnormal serum-free light chain ratio is observed in one third of patients.²⁸⁶, ²⁸⁷ The presence of an abnormal ratio correlates with advanced disease, markers of aggressive disease, and poor prognosis.²⁸⁷

Patients with CLL are immunosuppressed, as a result of hypogammaglobulinemia, alterations in T-cell function and to abnormalties in complement activation and neutrophil/monocyte function.²⁸⁸, ²⁸⁹ Moreover, the immunosuppression is potentiated by chemotherapy or immunotherapy. As a result, about 50% of patients have recurrent infections, with sepsis being an important cause of death. Apart from typical bacterial infections, CLL patients are also susceptible to opportunistic infections, particularly if they have received nucleoside analogues, steroids, or monoclonal antibodies. The nucleoside analogues are highly toxic to T lymphocytes, whereas monoclonal antibodies may be cytotoxic to B-cells (e.g., rituximab) or to both B- and T-cells (alemtuzumab).

The CLL cell has no useful immune function and shows poor stimulatory activity in mixed lymphocyte culture²⁹⁰ and in response to B-cell mitogens, such as pokeweed mitogen, lipopolysaccharide, and the Epstein-Barr virus.²⁹¹, ²⁹² However, phorbol esters, Staphylococcus aureus protein A from Cowan I, anti-µ antibodies, anti-CD40, and loxoribine have been shown to be potent mitogens.²⁹³, ²⁹⁴ Moreover, these cells are poor antigen-presenting cells and, as discussed below, can interfere with normal B- and T-cell function.²⁹⁵

The risk of infection is closely related to the extent of hypogammaglobulinemia in CLL, and the severity increases with the duration and stage of disease.²⁸⁸, ²⁸⁹, ²⁹⁶ The Ig levels are all decreased, and within the IgG class, reduced levels of IgG2 and IgG4 correlate best with the risk of infection; however, the decline in IgA levels is the most important predictor of infection.²⁹⁶, ²⁹⁷ Interestingly, the correlation between Ig levels and infection rate is not absolute, and some patients with normal Ig will have repeated infections whereas others with hypogammaglobulinemia will remain infection-free.²⁸⁸ Patients with CLL have reduced primary and secondary responses to immunization.²⁸⁹ Although patients with higher levels of gammaglobulin usually show better responses than those with low levels, the responses of both groups are abnormal. The pathogenesis of the hypogammaglobulinemia is poorly understood. However, impaired B-cell function and regulatory abnormalities of T-cells (including the reversal of normal helper/suppressor cell ratios) probably play a role. In addition, CLL-derived NK-cells have been shown to suppress Ig secretion by normal B-cells in vitro.²⁹⁸

The absolute number of T-cells may be increased in untreated CLL, and there are marked abnormalities in the surface markers, including inversion of the T-helper to -suppressor cell ratio, suggesting perturbations in T-cell function.²⁹⁰, ²⁹⁹ In addition, the increase in the number of T-suppressor cells may correlate with the degree of hypogammaglobulinemia.³⁰⁰ The T-cells usually respond normally to mitogens, such as phytohemagglutinin, in vitro and produce IL-2 and γ-interferon.³⁰¹ However, there is decreased T-helper function with reduced reactivity to allogeneic and autologous B-cells.³⁰², ³⁰³ Spontaneous and antibody-dependent cytotoxicities are reduced, suggesting an abnormality in the large granular lymphocyte population, including NK-cells.³⁰⁴

The cause of these abnormalities is unclear, but the increase in T-cells has been ascribed to stimulation by CLL cells or chronic infection, such as cytomegalovirus (CMV).²⁸⁹, ³⁰⁵, ³⁰⁶ Alternatively, the B-, T-, and NK-cell functions may be suppressed by means of immunosuppressive factors produced by CLL B-cells.³⁰⁷, ³⁰⁸ The gene expression profile of CD4⁺ and CD8⁺ cells are different in CLL patients as compared to normal individuals.³⁰⁷ However, the abnormal expression could be induced in CD4⁺ and CD8⁺ cells from normal individuals by co-culturing them with CLL cells.³⁰⁷ Similar studies showed that CLL cells could affect the ability of normal T-cells to form normal immunologic synapses.³⁰⁸ One potential candidate that could be causing these changes is TGF-β, which is secreted by CLL cells and marrow stromal cells in CLL, and has been shown to be a potent inhibitor of normal B- and
T-cells.¹³³, ¹³⁴, ¹³⁵, ¹³⁶, ¹³⁷ In addition, CLL cells express both CD40 and the CD40L (CD154), and these cells decrease CD154 expression by normal T-cells with resultant effects on normal B-cell differentiation and isotype switching.³⁰⁹

The levels of different complement components are decreased in CLL, particularly in patients with advanced disease.³¹⁰ As well, multiple defects in neutrophil and monocyte function have been described in CLL, and these are associated with an increased risk of infection.²⁸⁸, ³¹¹

Despite being immune deficient, CLL patients have an increased incidence of autoimmune cytopenias secondary to autoantibody formation.³¹², ³¹³, ³¹⁴, ³¹⁵, ³¹⁶, ³¹⁷, ³¹⁸ The immune cytopenias may be present before or at the time of diagnosis in one third and during the course of disease in two-thirds.³¹⁵ CLL is the most common cause of autoimmune hemolytic anemia (AIHA), causing 14% of cases, followed by systemic lupus erythematosus.³¹² Moreover, patients with AIHA have an increased incidence of MBL and subsequently developing CLL, but this is not true for patients with immune thrombocytopenia (ITP).³¹⁴, ³¹⁶ Overall, 4% to 10% of CLL patients develop AIHA, and this is usually associated with a warm-type antibody against the Rhesus system and a positive direct antiglobulin test (DAT).³¹², ³¹³, ³¹⁴, ³¹⁵ Another 7% to 14% of cases may have a positive DAT without evidence of hemolysis. The incidence of this disorder is increased with male sex, older age, high lymphocyte count, advanced disease, and the presence of biologic markers of poor prognosis, such as high β2-microglobulin, high ZAP-70, and unmutated IgV_H.³¹⁴, ³¹⁵, ³¹⁷ The diagnosis of AIHA may be difficult in CLL, as multiple factors may produce anemia, a positive DAT can exist without hemolysis, and other diagnostic parameters may be influenced by the disease.³¹⁴, ³¹⁵, ³¹⁷ Zent et al.³¹⁷ have thus developed diagnostic criteria for the diagnosis of AIHA in CLL and include (a) hemoglobin <100 g/L; (b) ≥1 marker of hemolysis, i.e., reticulocytosis, increased indirect bilirubin without liver disease, increased lactate dehydrogenase (LDH) without another cause or increased marrow erythropoiesis; (c) positive DAT/cold agglutinins or ≥2 markers of hemolysis, without evidence of hypersplenism or bleeding. When fludarabine was introduced, AIHA was seen relatively frequently leading investigators to believe that this was a unique complication of the drug. However, in the LRF CLL4 trial the incidence of AIHA was similar in previously untreated patients who received chlorambucil (12%) as compared to fludarabine (11%), and the incidence was lowest in patients treated with a combination of fludarabine plus cyclophosphamide (5%) indicating that cyclophosphamide has a protective effect.³¹⁸ The likelihood of developing AIHA was greater if the patient had a positive direct DAT before treatment, had advanced disease, or had a high β2-microglobulin test.³¹⁸ The incidence of a positive DAT before treatment was 14% and approximately one third of patients receiving chlorambucil or fludarabine with a positive DAT developed AIHA. Conversely, the chance that a patient who was DAT-negative would develop AIHA was 7%. Thus, the high incidence of AIHA observed when fludarabine was first used was because patients receiving this drug usually had advanced and previously treated disease. The protective effect of cyclophosphamide has been confirmed in a subsequent study where the incidence of AIHA was 1%.³¹⁹ Although the addition of rituximab to chemotherapy would be expected to reduce the incidence of AIHA further, it is reported to occur in <1% to 6.5% of patients receiving FCR.²¹⁰, ³¹⁹ The impact of the development of immune cytopenias on survival has been recently evaluated in 2 studies, which have demonstrated that the prognosis of these patients is no worse than those without cytopenias.³¹⁵, ³¹⁷