Atypical lymphocytosis (AL) is characterized by increased numbers of large reactive lymphoid cells identified on a Wright-stained peripheral smear and is most often seen in acute viral infection, especially herpesviruses Epstein-Barr virus (EBV), cytomegalovirus (CMV), and varicella-zoster virus (VZV), but also pertussis, brucellosis, syphilis, toxoplasmosis, drug reactions, and lymphoid leukemia. The single most common clinical syndrome associated with AL is EBV-associated acute infectious mononucleosis (IM), in which patients typically present with fever, pharyngitis, and cervical lymphadenitis. In IM, EBV-infected B lymphocytes trigger an intense T cell-mediated immune response within oropharyngeal lymphoid tissues and increased atypical lymphocytes in the peripheral blood. Given the association of EBV with many forms of lymphoid malignancy, the AL in IM has been of enduring interest to pathologists, hematologists, virologists, and immunologists.
Previous phenotypic analyses of AL have been limited to EBV-associated IM. We present results of a comprehensive immunophenotypic analysis of peripheral blood lymphocyte subsets from 97 patients with AL, defined as an absolute lymphocyte count of >4000/μl with >10% atypical forms. Results from AL were compared with results obtained from 37 normal controls to identify AL-associated lymphocyte abnormalities. All cases of AL were examined so as to not only identify immunophenotypic abnormalities common to all cases of AL regardless of underlying etiology but also identify abnormalities specific for EBV-positive IM.
MATERIALS AND METHODS
Ninety-seven samples of freshly discarded EDTA-anticoagulated whole peripheral blood (less than 24 h old) were collected from patients with absolute AL after appropriate consent was obtained. Absolute AL was defined as absolute lymphocytosis as per complete blood count with >10% atypical lymphocytes on peripheral smear. Thirty-seven samples of freshly discarded blood from University of Texas Medical Branch (UTMB) patients with completely normal complete blood counts were collected as normal controls. Clinical diagnoses in the normal control population were (in descending order of prevalence) normal delivery, cardiac disease, renal disease, and infection, psychiatric disease, cancer, rheumatologic disease, diabetes, pancreatitis, and trauma.
Aliquots of EDTA-anticoagulated whole blood were incubated with combinations of fluorochrome-conjugated ((fluorescein isothiocyanate (FITC), phycoerythrin (PE), peridinin chlorophyll protein (PerCP)) monoclonal antibodies to human lymphocyte surface antigens (BD Bioscience, San Diego, CA) (Table 1). Erythrocytes were lysed with FACS Lysing Solution (BD Biosciences, San Diego, CA), cells rinsed with phosphate-buffered saline, and fixed briefly in 1% paraformaldehyde. Samples were analyzed by multiparametric flow cytometry using a FACSCalibur instrument equipped with CELLQuest software (BDIS, San Diego, CA). The nongranular lymphoid cell population was selected for analysis by standard forward-side scatter gating, and quadrant markers set with isotype control antibodies (BD Bioscience, San Diego, CA).
Table 1. Antibody Combinations for Flow Cytometry
The mean, standard deviation, and median of absolute counts, percentages, and ratios of various phenotypic subsets of T, B, and NK cells from all AL cases, the heterophile-positive IM AL subset, and the non-EBV AL subset were calculated and compared with normal controls. Statistical significance of the results was calculated using the Student's t-test with a P value of <0.05 considered statistically significant. Absolute numbers of T helper cells (CD3/CD4-positive), T suppressor/cytotoxic cells (CD3/CD8-positive), NK cells (CD3-negative CD16/56-positive), B cells (CD19-positive), etc., were calculated from the absolute lymphocyte count and the phenotype percentage.
The most common clinical presentation of patients presenting with AL was IM (55%) with heterophile antibody or EBV virus capsid antigen (VCA) IgM positivity confirmed in 33%, while the remaining cases consisted of other diseases, including viral syndromes of undetermined etiology, viral hepatitis, other herpesvirus infections (VZV, CMV), cancer, and drug reactions (Table 2). Patients with EBV-associated disease were significantly younger than patients with AL due to other causes (16 years versus 27 years, P = 0.0083).
Table 2. Causes of EBV Negative Atypical Lymphocytosis (Number of Cases)
|Viral syndrome of undetermined cause (24)|
|Hepatitis (8) (three HCV, one HBV)|
|Upper respiratory infection, presumed viral (6)|
|Varicella infection (3)|
|Lymphoma (one Hodgkins, one non-Hodgkins)|
|Acute precursor B cell lymphoblastic leukemia|
|Lymphadenopathy of undetermined cause|
|Drug reaction (dapsone)|
|Acute HIV infection|
Absolute Cell Counts
The total lymphocyte count, percent atypical lymphocyte, CD8 T cell count, and NK cell count are all significantly increased in both AL subgroups, i.e., EBV-positive and EBV-negative groups. (Table 3) Nevertheless, in comparison with EBV-negative cases, EBV-positive cases are characterized by significantly higher CD8-positive T and NK cell counts (Tables 4 and 5). In the EBV-positive group, CD8-positive T cells (nearly all are CD45RO-positive) are increased 10-fold; NK cells increased nearly 5-fold, and total lymphocytes increased >3-fold compared with normal controls. In contrast, no increase in absolute CD4-positive T cell count and CD19-positive B cell count was seen in any AL group. (Tables 6 and 7) Gamma-delta (γδ)-positive T cells are markedly increased (nearly sevenfold) in the EBV-positive group, while no increase was seen in the EBV-negative group (Table 8).
Table 3. General Findings
| ||<0.0001||0.0019||<0.0001|| |
| ||ND||0.0014||ND|| |
| ||<0.0001||<0.0001||<0.0001|| |
Table 4. CD8 Positive T Cells
| ||<0.0001||0.0018||0.0002|| |
| ||0.0093||0.0053||0.0059|| |
|ABS CD8 RA||0.93||0.87||0.65||0.34|
| ||<0.0001||0.0026|| || |
|ABS CD8 RO||3.21||4.29||1.81||0.14|
| ||<0.0001||0.0016||0.0001|| |
|CD8 CD25+/−||1.57 (.02)||0.02||0.08||0.06|
| ||0.0008||0.0182|| || |
| ||<0.0001||<0.0001||0.0006|| |
|ABS CD8 CD28+||2.75||3.29||1.4||0.49|
| ||<0.0001||<0.0001||0.0007|| |
|ABS CD8 CD28−||1.21||1.74||1.00||0.2|
| ||<0.0001||0.0001||0.0146|| |
| ||0.0004||<0.0001|| || |
|CD8 CD58+/−||415 (21)||158 (19.6)||40.8 (17)||3.95 (1.98)|
| ||0.0497|| || || |
|CD8 CD95+/−||72.5 (22)||65.6 (22)||92.8 (31)||1.25 (1.04)|
| ||0.0001||0.0629|| || |
|CD8 CD43+/−||1419 (453)||468 (278)||245 (245)||1489 (874)|
| || ||0.0612||0.021|| |
| ||0.0132||0.0117||0.0016|| |
Table 5. NK Cells
| ||<0.0001||0.0144||0.0073|| |
Table 6. CD4 Positive T Cells
| ||0.0001||0.0012||0.0007|| |
| || ||0.0195|| || |
| ||<0.0001||<0.0001||0.0207|| |
|ABS CD4 CD62L−||0.21||0.18||0.18||0.08|
| ||<0.0001||0.0227||0.0369|| |
| ||0.0457||0.0004||0.0005|| |
| ||0.016||0.0022||0.0157|| |
| ||0.019||0.0228||0.0107|| |
| ||0.0235||0.0271|| || |
Table 7. B Cells
| || ||0.093||0.0206|| |
|ABS CD20 CD23−||0.27||0.35||0.19||0.13|
| ||0.1||0.0447|| || |
Table 8. T Cell Receptor Utilization
|CD3 αβ/γδ ratio||142 (18)||17.9 (16.6)||69.6 (37)||29.6|
| || ||0.0375|| || |
|ABS CD3 γδ||0.32||0.41||0.07||0.06|
| ||0.0317||0.0308|| || |
| ||0.002||0.0876|| || |
T Cell Activation Markers
The most dramatic change was seen in the CD8-positive T cell subset with a 6- to 10-fold change in the CD45RA/RO ratio and a 10- to 30-fold absolute increase in CD8/CD45RO-positive cells, with a significant increase in the EBV-positive subgroup compared with the EBV-subgroup (Table 4). In contrast, for CD4-positive T cells there was an insignificant change in the CD45RA/RO ratio, and a small but significant (twofold) absolute increase in CD4+/CD45RO+ cells in all AL groups (Table 6). Expression of the activation antigen HLA-DR was markedly increased on both CD4-positive and CD8-positive T cells in both groups, with expression on CD8-positive T cells (>30-fold) greatly exceeding that seen on CD4-positive T cells (greater than fivefold). In contrast to HLA-DR, CD38 expression, while increased greater than twofold on CD8-positive cells, was reduced greater than twofold on CD4-positive T cells in both subgroups. Thus, unlike HLA-DR, CD38 expression does not appear to serve well as a discriminating marker of lymphocyte activation since it is expressed on most CD4- and CD8-positive cells even in normal controls (CD4 CD38+/− ratio = 4.56; CD8 CD38+/− ratio = 5.56). In contrast to the dramatic changes in HLA-DR and CD38 expression, the only significant change in CD25 (interleukin (IL)-2 receptor) expression was moderately decreased expression (less than twofold) by CD4-positive T cells in the EBV-positive group only.
CD4-Positive T Cell Subsets
There was no change in absolute CD4-positive T cell count in AL (Table 5). However, significant proportional increases in CD7-negative (threefold), CCR5-positive (threefold), CD43-negative (greater than fivefold), and CD48-negative (greater than fivefold) subsets, as well as an absolute increase in CD62L-negative (suppressor-inducer) CD4-positive T cells, were identified.
CD8-Positive T Cell Subsets
There were dramatic changes in CD8-positive T cell subsets in AL, including proportional increases in CD57-negative (4-fold), CD58-positive (20-fold), CD48-negative (4-fold), and CD95-positive (>20-fold) subsets (Table 4). A significant proportional increase in the CD43-negative subset was limited to the EBV-negative subgroup, while an increase in the CD57-negative subset was limited to the EBV-positive group. Given the increased absolute CD8 count, the number of both CD28-negative and CD28-positive T cells is increased in both groups when compared to normal controls (CD28-negative cells increased more than fivefold; CD28-positive cells increased more than threefold). While no significant difference between EBV-positive and EBV-negative subgroups in the absolute number of CD28-negative T cells was seen, there were significantly more CD28-positive T cells in the EBV-positive subgroup. By gating on the larger atypical cells, an increased proportion of CD8/CD28-positive cells was noted, suggesting that many of the larger atypical T cells, of which there were significantly more in the EBV-positive group, are effector cytotoxic T cells.
No change in absolute CD19-positive B cell count was seen in AL (Table 7). However, a significant increase (two- to threefold) in CD23-negative CD20-positive B cells was noted in total AL and EBV-positive subgroups. There was a trend toward an increased proportion of CD23-negative B cells in all groups, which reached statistical significance only in the EBV-negative subgroup (greater than twofold).
Specific Abnormalities in EBV-Positive IM
When the EBV-associated (heterophile and/or EBV VCA IgM-positive) IM subgroup was segregated and compared with the unselected (total) AL group, many of the lymphocyte abnormalities became more significant. In addition, there were eight findings in the EBV-positive subgroup that were significantly different from the EBV-negative subgroup (marked * in tables, summarized in Table 9). First, EBV-positive patients were significantly younger than EBV-patients (16 years versus 27 years, P = 0.0083). The degree of AL was more pronounced in EBV-positive cases, with a higher absolute lymphocyte count (7.9 versus 4.8 × 106/ml, P = 0.0019) and higher atypical lymphocyte percentage (24% versus 14%, P = 0.0014). Although there was no difference in the number of CD4-positive T cells between groups, there was a significant difference in the CD8-positive T cell count with a much higher count in the EBV-positive group (5.2 versus 2.46 × 106/ml, P = 0.0018). In addition, there were two subsets of CD8-positive T cells (CD45RO-positive, CD28-positive) that were very significantly increased in the EBV-positive group compared with the EBV-negative group (P = 0.0016 and <0.0001, respectively). A very striking difference between the two subgroups was the marked increase in γδ-positive T cells seen in the EBV-positive group (0.41 versus 0.07 × 106/ml, P = 0.03) — a finding completely absent in the EBV-negative group. EBV-positive AL was also characterized by a significantly increased number of NK cells when compared with EBV-negative cases (0.89 versus 0.48 × 106/ml, P = 0.0144).
Table 9. Statistically Significant Immunophenotypic Differences Between EBV-Positive and EBV-Negative Atypical Lymphocytosis (EBV-Pos > EBV-Neg)
|Percent atypical lymphs|
|Absolute lymph count|
|Absolute CD8-positive T cell count|
|Absolute CD8/CD45RO-positive T cell count|
|Absolute CD8/CD28-positive T cell count|
|Absolute γδ T cell count|
|Absolute NK cell count|
Several studies of peripheral blood lymphocytes in IM have demonstrated increased numbers of activated CD8-positive T lymphocytes as indicated by increased expression of the activation markers CD29, CD38, HLA-DR, and CD45RO (1–5). In the present study, a fivefold increase in CD3/CD8-positive suppressor/cytotoxic T cells was noted with a marked increase in HLA-DR and CD45RO positivity (only a slight increase in CD38), indicative of activated primed T cells (4). There was a marked increase (>20-fold) in CD95 expression by T cells in AL — another finding consistent with T cell activation. CD95-positive cytotoxic T cells induce Fas-mediated apoptosis when bound to CD95L-positive target cells (6).
CD28-positive T cells are activated by interaction with B7-expressing antigen-presenting cells. Although most reports indicate that CD28 is expressed by cytotoxic T cells (CTL) and not by suppressor T cells (7, 8), some reports indicate that CD8-positive CD28-negative T cells may be cytolytic effectors (9). Increased numbers of CD8/CD28-positive cytotoxic T cells have been reported in IM (10). CD8/CD28-positive T cells were increased in the present study, especially in the EBV-positive subgroup (EBV-positive subgroup, greater than sixfold; EBV-negative subgroup, less than threefold). However, CD28-negative CD8+ T cells were increased to an even greater extent (EBV-positive group, greater than eightfold; EBV-negative group, fivefold), a finding consistent with downregulation of CD28 expression on CD8-positive T cells reported in acute viral infection (11). Interestingly, when the larger lymphocytes are gated, there is a proportional increase in CD8/CD28-positive cytotoxic T cells, thus suggesting that the larger “atypical” T cells, unlike the smaller T cells, are cytotoxic T cells rather than suppressor T cells.
Expression of CD25, the p55 subunit of the IL-2 receptor, on CD8-positive cells has been shown to be decreased in IM (1, 5, 10, 12). In the present study, decreased CD25 expression was noted on both CD4 and CD8-positive T cells, but only in the EBV-positive subgroup. Despite the lack of CD25 expression by CD8-positive T cells in IM, these cells have been shown to be capable of responding to exogenous IL-2 in vitro through normal expression of the p70 subunit of the IL-2 receptor (12).
Although CD4-positive CD62L-negative suppressor-inducer T cells are reportedly decreased in IM (10), in the present study cells of this phenotype were significantly increased (two- to threefold). Decreased CD62L expression is also characteristic of interferon-γ-producing T effector memory (Tem) cells targeted to peripheral tissues (13). In chronic persistent EBV infection, CD62L-negative Tem and CD62L-positive Tcm (central memory) EBV-specific memory T cells are present (14). Perhaps the increase in CD62L-negative T cells in the peripheral blood reflects lymphocyte homing directed to the infected pharyngeal mucosa in early acute viral infection.
Some reports indicate that in addition to increased numbers of activated CD8-positive T cells, IM is characterized by increased numbers of activated CD4-positive T helper-inducer cells (1, 2, 4). Although in the present study CD4-positive T cells were not increased, there was a very significant increase in HLA-DR positivity, a finding consistent with CD4-positive T cell activation. Curiously, no similar increase in expression of activation markers CD25 and CD38 was detected on CD4-positive T cells.
Th2 cells promote B cell-mediated humoral immune responses rather than T cell-mediated cytotoxic T cell responses, and thus may impair T cell-mediated antiviral immunity. Although immunophenotypic differentiation of Th1 and Th2 cells is admittedly imprecise, several markers, including CD7, CD43, and CCR5, have been used in previous studies. CD7 negativity is characteristic of a Th0/Th2 T helper-inducer subset found in normal blood and increased in HIV infection, renal transplantation, and bone marrow transplantation (15, 16). In the present study, CD7-negative CD4-positive T cells were increased 2.7-fold. CD7 negativity has also been previously reported in a dimCD4-positive T cell population in IM (17). The proportional increase in CD4-positive CD43 (sialophorin)-negative T cells seen in AL is also consistent with an increase in Th2-type T cells (18). Perhaps Th2 cells in IM not only inhibit the T cell-mediated cytotoxic response to EBV, but also promote growth of EBV-infected B cells. On the other hand, in this study AL is also characterized by a proportional increase in CCR5-positive CD4-positive T cells. Since the chemokine receptor CCR5 is selectively expressed on human Th1 cells (19), the present data are consistent with an increase in Th1 cells — a result at odds with the previous results.
CD57 has been identified as a marker of CD8-positive effector CTL (11, 20–22). On the other hand, in one report CD8-positive CD57-positive cells were shown to inhibit generation of cytotoxic T cells to EBV-positive B lymphoblastoid cells (23). Within lymph nodes, CD57-positive T cells are present in B cell-rich reactive germinal centers and are increased in nodular lymphocyte-predominant Hodgkin lymphoma (24). The significantly decreased CD57 expression seen in the EBV-positive subgroup only is suggestive of an increased number of suppressor T cells in EBV-positive IM, complemented by the increase in CD4-positive CD62L-negative suppressor-inducer T cells.
A dual CD4/CD8-positive phenotype has been reported in a subset of cytotoxic T cell clones grown from patients with IM (2). However, significant numbers of dual positive T cells were not identified in the present study. The dual CD4/CD8-negative phenotype is characteristic of T cells that express the γδ T cell receptor, a population reportedly increased in IM (25, 26). In the present study, although they represent a minor population, a marked increase (greater than sevenfold) in γδ T cells was detected in the EBV-positive subgroup only. γδ-positive T cells, which preferentially localize to the alimentary tract, may be important in mucosal immunity (27, 28) and have been shown to respond to human EBV-infected B cells (29, 30). Perhaps circulating γδ-positive T cells, specifically increased in EBV-associated IM, are targeted to the EBV-infected pharyngeal mucosa and lymphoid tissues.
NK cells have been shown to be important in the early innate phase of the immune response to viral infection prior to initiation of an antigen-specific T cell-mediated response. Thus, it is not surprising to find that AL is characterized by a significant increase (threefold) in CD16/56-positive NK cells. An interesting new finding is that the NK lymphocytosis in EBV-positive IM is significantly higher than that seen in the EBV-negative subgroup of AL.
Two smaller previous studies identified a decreased number of B cells in IM (5, 31). In contrast, in the present study no significant change in absolute B cell count was noted. It is likely that the B cell count in early IM varies with time after infection and that lab results will be dependent upon sampling time. Although increased numbers of CD5-positive B cells have been reported in IM (32), we were unable to demonstrate an increase in this B cell subpopulation (data not shown). EBV transformation of B cells in vitro is accompanied by increased expression of the low-affinity IgE receptor CD23 (33), and increased CD23-positive B cells have been reported in IM (34). However, in the present study, a decrease in CD23-positive B cells was identified. Even if one assumes that circulating EBV-infected B cells do express CD23, these cells are very infrequent, with estimates ranging from 0.07–0.8% of the B cells (35, 36). It seems very unlikely that the low frequency of EBV-infected B cells in blood could lead to a detectable increase in CD23 expression by flow cytometry. EBV-infected B cells that express cytoplasmic immunoglobulin are CD23-negative (37). Since plasmacytoid cells are often seen in peripheral smears of IM, perhaps these cells represent CD23-negative EBV-infected B cells. Another factor contributing to the decreased CD23 expression in IM may be the rapid immune elimination of highly antigenic CD23-positive lymphoblastoid B cells that express the full complement of EBV latent genes with persistence of poorly immunogenic CD23-negative B cells expressing EBV LMP-2a only (38).
The present study was designed to further define the immunophenotype of AL and to identify immunophenotypic features that differentiate classic EBV-associated heterophile-positive IM from AL of other causes. Given the large number of cases examined, we have more carefully detailed previously described lymphocyte abnormalities in AL and have expanded the range of lymphocyte antigens examined. In addition, for the first time, we have identified eight specific immunophenotypic features that differ between EBV-positive and EBV-negative cases (Table 9). This information may prove useful in further understanding of the profound immunologic effects induced by primary EBV infection. Although it is unlikely that this information will lead to practical methods for diagnostic discrimination between EBV-positive and EBV-negative AL, it may allow for better discrimination between benign and malignant lymphocytosis.