This article is a US government work and, as such, is in the public domain in the United States of America
Monoclonal B-cells can be detected in the peripheral blood of some adults without B-cell malignancies, a condition recently termed monoclonal B-cell lymphocytosis (MBL). The risk of individuals with MBL progressing to a B-cell malignancy is unknown. Polyclonal B-cell lymphocytosis (PCBL) has not been systematically studied in the general population.
We obtained lymphocyte subset counts on 1,926 residential adults aged 40–76 years in a series of environmental health studies between 1991 and 1994. We then conducted two follow-ups in 1997 and 2003 on consenting participants with B-cell lymphocytosis, which included nine participants with MBL. To ascertain the clinical implications of MBL, we reviewed medical records and death certificates.
The overall prevalence of MBL was 0.57% (11/1,926): nine cases at baseline and two additional cases identified at follow-up. Two (19%) MBL cases subsequently developed a B-cell malignancy; MBL persisted in the remaining nine cases (81%). All PCBL cases where no clone emerged regressed to normal B-cell counts over the follow-up period. MBL was significantly more frequent in residents near a hazardous waste site than in the control populations (age-adjusted OR 6.2; 95%CI 1.1–36.2).
MBL confers an elevated risk for developing a B-cell malignancy, although it occurs only in a minority of cases. PCBL is most often a transient state, but a monoclonal population can emerge and persist. Prospective studies are needed to distinguish stable from progressive forms of B-cell lymphocytosis and to clarify the etiologic role of environmental exposures. Published 2007 Wiley-Liss, Inc.
B-cell lymphocytosis may be due to either a monoclonal or a polyclonal expansion. The presence of circulating monoclonal B-cells is the hallmark of B-cell chronic lymphocytic leukemia (B-CLL), and such cells can sometimes be detected in the peripheral blood of patients with non-Hodgkin lymphoma. Previous studies have reported that monoclonal expansions of B-lymphocytes can be detected in the peripheral blood of some healthy adults (1–5). The prevalence of these monoclonal expansions, recently termed monoclonal B-cell lymphocytosis (MBL) (6), has been estimated as high as about 5% among older adults who had a normal blood cell count (4, 5). The natural history of MBL is currently unclear: no prospective studies have investigated its propensity toward transient, stable, or progressive forms.
Elevated B-cell counts without an apparent monoclonal population, usually termed polyclonal B-cell lymphocytosis (PCBL), have been found in a rare syndrome associated with female smokers (7, 8), a severe autoimmune syndrome (9), and Gaucher disease (10). Isolated case reports of PCBL have also been described in sarcoidosis (11), Hodgkin lymphoma (12), and a lymphoproliferative disorder resembling hairy cell leukemia (13). The natural history and prevalence of PCBL in the general population is unknown.
We conducted two medical follow-up examinations of participants selected from a series of seven baseline environmental health studies that compared the status of the immune system among residents near hazardous waste sites to residents in comparison communities (3). The follow-up cohort included 11 individuals with MBL and 51 with PCBL; the observation period from baseline to second follow-up spanned up to 12 years. The main purpose of this follow-up study was to determine whether MBL and PCBL were transient states, stable conditions, or progressive disorders that led to B-cell malignancies.
The base population consisted of 1,926 participants at least 40 years old from seven cross-sectional studies conducted in the United States from 1991 to 1994 (Fig. 1). Each of the original studies included a community near a hazardous waste site (target area) and a demographically similar community with no point source of hazardous waste exposure (comparison area) (3). An age- and sex-stratified random sampling scheme was used to select participants from each target and comparison areas, using door-to-door census data for sampling frames. The same standardized study protocol, questionnaire, and immune test panel were used in each study. The immune test panel included complete blood count (CBC), immunophenotyping for the three major lineages of lymphocytes (T-cells, B-cells, NK-cells) and T-cell subsets (CD4 and CD8), serum immunoglobulin levels, and C-reactive protein. CD5 + B-cells were determined on a subset of participants.
Of the original 1,926 participants, we offered medical follow-up to 74 (3.8%) who met one or more of the following criteria: (1) an absolute number of B-lymphocytes ≥830 cells/μL; (2) an absolute number of CD5 + B-lymphocytes ≥350 cells/μL; (3) a percentage of CD5 + B-lymphocytes among all B lymphocytes ≥60%. The three criteria reflected the upper 2.5 percentile of their respective distributions in the base population; 51 (69%) met the first criterion and 7 (9%) met all three criteria. Three individuals who had already been diagnosed with CLL or lymphoma at baseline were excluded.
The 74 eligible participants were first contacted in 1997; 59 (79.7%) consented to participate, 11 (14.9%) declined, one (1.4%) was deceased, and three (4.1%) were lost to follow-up. In 2003, we offered a second follow-up for the 59 individuals who participated in the first follow-up; 49 (83.1%) consented to participate, two (3.4%) declined, five (8.5%) were deceased, and three (5.1%) were lost to follow-up. We searched the National Death Index and Social Security Death Index databases through December 31, 2002 for 17 of the 74 eligible individuals, whose vital status was not known during the second follow-up. No additional deaths were identified. Therefore, six deaths (8%) occurred among the 74 individuals, and 49 (72%) of the 68 living individuals participated in both follow-ups.
Our institutional review boards approved the original studies and both follow-up studies. Informed consent was obtained from all participants. We obtained medical records for the participants who consented to release of their records. For all decedents identified, we obtained copies of their death certificates and medical records from the certifying physicians or from the hospitals where they died.
Blood samples were obtained from study participants at local hospital laboratories that performed the CBC and shipped specimens to the US Centers for Disease Control and Prevention. Baseline immunophenotyping was performed by two-color flow cytometry at CDC (14). MBL was confirmed at baseline by immunophenotyping at the US Food and Drug Administration (15). For both follow-up studies, a single clinical hematopathology laboratory performed diagnostic immunophenotyping by flow cytometry. In addition to CD45 gating, B-cells were identified as CD19+ or CD20+, CD3− events on two-color (baseline), three-color (first follow-up) (16), or four-color (second follow-up) (17) panels.
MBL was defined as the presence of a discernable monoclonal B-cell population and/or by κ-λ ratio, in the absence of a lymphoproliferative disorder (6). Three participants whose baseline B-cell immunophenotypes clearly showed CLL-like clusters, identified by CD20, CD19 and CD5 staining patterns, were also classified as MBL, without follow-up light chain immunophenotyping. The remainder of the follow-up cohort was classified as “non-MBL.” The non-MBL individuals in the follow-up cohort, whose absolute B-cell count was ≥830 cells/μL, were further classified as PCBL.
We first examined the relationship between the absolute counts of lymphocyte subsets (B-cell, T-cell, and NK-cell) and age in the base population by calculating the Spearman correlation coefficients (rs). We also employed the Jonkheere–Terpstra trend test to examine age and absolute cell counts as categorical variables using quartiles. To determine whether changes in absolute B-cell counts over the follow-up period could be explained solely by aging, we used generalized estimating equation (GEE) modeling (18) with an exchangeable correlation-structure. This method accounts for the inherent correlation between baseline and follow-up measurements within subjects.
Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated using logistic regression analyses to examine association between demographic characteristics (i.e., sex, age, and area of residence) and absolute lymphocyte counts (i.e., total lymphocytes, B-cells, T-cells, and NK-cells) at baseline, comparing MBL cases with PCBL cases. To account for the confounding effect of age, adjusted ORs were calculated including age (45–54 years vs. 55 and older) in all models. The baseline cell counts were dichotomized using the median values of the PCBL cases as cut points. The baseline demographic and clinical characteristics of MBL cases were also compared with those of controls randomly selected from the base population, stratified by the year of entry into the original studies. The 74 individuals selected for medical follow-up and the three B-cell malignancy cases identified at baseline were excluded from the sampling frame for controls. All data were analyzed using SAS software.
The Base Population
The median age among the base population (n =1,926) was 53 years, and the majority (94%) were white (Table 1). Among the base population, absolute B-cell counts were negatively correlated with age (rs = −0.17, P < 0.0001) (Fig. 2A), a trend that was more pronounced when known MBL and B-cell malignancy cases were excluded from the base population. Total lymphocytes (Fig. 2B) and T-cell counts (Fig. 2C) were also negatively correlated with age (rs = −0.14, P < 0.0001; rs = −0.19, P < 0.0001, respectively), whereas NK-cell counts were positively correlated with age (rs = 0.16, P < 0.0001) (Fig. 2D). These trends confirm previous reports in smaller populations (19, 20).
Table 1. Baseline Characteristics of 74 Individuals Eligible for Medical Follow-Up Study
Baseline data were obtained at the time of the original cross-sectional studies that were conducted between 1991 and1994 in the USA. The base population included the participants who were 40 years or older and had a B-lymphocyte count available at the time of the original studies.
Median values (10th, 90th percentiles).
Values in square brackets indicate percentages.
Target areas were near hazardous waste sites.
Based on n = 771 for base population and n = 41 for eligible participants.
Based on n = 1698 for base population and n = 69 for eligible participants.
The demographic characteristics of the 74 individuals eligible for follow-up were similar to those of the base population (Table 1). Reflecting the selection criteria, the median values for total B-cell and CD5 B-cell counts were notably higher among those eligible for follow-up. Their median white blood cell counts, total lymphocyte counts, T-cell counts, and NK cell counts were also somewhat higher than those of the base population.
Follow-up results of the 74 eligible individuals are summarized in Figure 3. A total of six deaths occurred in the cohort during the follow-up period: 4/11 (36%) were MBL and 2/63 (3%) were non-MBL. In addition to the nine MBL cases present at baseline, two MBL cases were detected during the first follow-up in 1997. None were detected in the second follow-up, and so a total of 11 MBL cases were identified during this study. Of these 11 MBL cases, four (36%) died between the first and second follow-up, and four (36%) maintained their MBL status through the second follow-up. The remaining three (27%) did not participate in either follow-up.
Review of the medical records and laboratory results of MBL cases are summarized for the four living cases (Fig. 4) and the four deceased cases (Table 2). The clinical course among the MBL cases varied widely, ranging from one case with apparent regression of a λ clone (Case no. 1016 in Fig. 4) to one case with rapid progression to CLL of a κ clone (Case no. 1021 in Table 2). MBL persisted over 9–12 years in three of four living cases without progressing to a B-cell malignancy; the fourth (Case no. 1026 in Fig. 4) was diagnosed with Waldenstrom's Macroglobulinemia. Among the deceased MBL cases, only one (Case no. 1021 in Table 2) had developed CLL, which was listed as a contributory cause on the death certificate. A κ clone was found in 7 (87.5%) of the eight MBL cases, whose κ/λ ratios were measured.
Table 2. Case Summary for Deceased MBL Cases
Male/died in 2002 at age 78 years
CD5 bright, CD20 dim, kappa-restricted MBL detected in 1997, preceded 3 years by a CD5+ B-cell lymphocytosis. Severe seronegative rheumatoid arthritis was present. CLL diagnosed with a striking increase in BALC prior to death. Thrombocytopenia was noted at death, but splenomegaly and lymphadenopathy were not seen.
Male/died in 2001 at age79 years
CD5 bright, CD20 dim, kappa-restricted MBL detected in 1994. Had a history of end stage, steroid dependent, COPD. Lymphadenopathy, splenomegly, or lymphocytosis was never noted.
Male/died in 2002 at age 82 years
CD5 bright, CD20 dim, kappa-restricted MBL detected in 1993. Remote history of a clinical EBV infection. Atypical lymphocytes and smudge cells noted in early 2002. This patient died from a stroke.
Female/died in 2002 at age 70 years
CD5 bright, CD20 bright, kappa-restricted MBL detected in 1994. This patient died from abdominal carcinomatosis.
The baseline demographic characteristics of MBL cases were compared with those of non-MBL cases, as well as with those of controls (Table 3). Compared with non-MBL cases, MBL cases were more likely to be above 55 years (OR 7.2; 95% CI 1.7–30.4). Compared with controls, MBL cases were more likely to reside near a hazardous waste (age-adjusted OR 6.2; 95% CI 1.1–36.2). Male gender, which is a known risk factor for B-CLL (21, 22), was not a significant risk factor for MBL in our study.
Table 3. Baseline Characteristics Associated with the Risk of MBL
MBL, monoclonal B-cell lymphocytosis (total n = 11).
Non-MBL, Medical follow-up participants without a discernable B-cell monoclonal population (total n = 63).
Control: A random sample drawn from the base population, excluding three cases with lymphoproliferative disease at baseline and the 74 follow-up eligible individuals, stratified by the study entry year (total n = 44).
Age-adjusted for sex, location of residence, and cell counts.
Values in parentheses indicate percentages.
Values in square brackets indicate 95% confidence intervals.
Areas near a hazardous waste site.
Dichotomized cell counts, by using median values of the non-MBL group (for MBL vs. non-MBL) and the control group (for MBL vs. control).
When baseline lymphocyte counts for MBL cases were compared to the median values for controls at baseline, MBL cases were more likely to be above the median for absolute lymphocyte counts (OR 9.7; 95% CI 1.1–86.0) and absolute B-cell counts (OR 13.5; 95% CI 1.5–125.9). T-cell counts (OR 3.8; 95% CI 0.8–18.4) and NK-cell counts OR (3.5; 95% CI 0.6–19.3) were also more likely to be above the median, but these elevations were not statistically significant (Table 3). On the other hand, when lymphocyte counts for MBL cases were compared with the median values for non-MBL follow-up cases, MBL cases were more likely to be below the median for both T-cell counts (OR 0.2; 95% CI 0.05–1.1) and NK cells (OR 0.5; 95% CI 0.10–1.9), but above the median for absolute counts of lymphocytes and B-lymphocytes.
Of the 51 individuals with PCBL at baseline, 42 participated in first follow-up and 32 participated in both follow-ups. One participant with PCBL at baseline developed MBL by the first follow-up. The distribution of B-cell counts in the remainder of this group regressed substantially by the first follow-up and was indistinguishable from the baseline distribution by the second follow-up (Fig. 5). None of the 32 participants who participated in both follow-ups had PCBL by the second follow-up. The results of GEE analysis showed that the magnitude of this change was considerably greater than the age-related decline in B-cell counts observed in the base population and could not be explained by normal aging during the follow-up period.
This report describes the first long-term prospective study of MBL and PCBL in a cohort selected from well-defined population. Although MBL was described soon after immunophenotyping become widely used (1–5), the nomenclature and diagnostic criteria have only recently been established (6). These criteria exclude anyone with a diagnosable B-cell malignancy. In our original analyses on the base population, we found 11 individuals with detectable monoclonal B-cells in peripheral blood at baseline, and two more were found in the first follow-up. Of these 13, only two had been diagnosed with a B-cell malignancy at the time the monoclonal B-cells were first detected in our baseline study; therefore, among those with monoclonal B-cells first detected in peripheral blood, the prevalence of MBL (11/1,926) was five times higher than the prevalence of B-cell malignancies (2/1,926). The mortality rate among MBL over the entire follow-up period was 36%, much higher than the 3% mortality among non-MBL. However, since MBL cases were significantly older than non-MBL, and the number of observations is too small for a formal analysis, we do not know whether MBL is an independent predictor of mortality.
We first noticed monoclonal B-cell populations in the baseline studies because of their immunophenotypic staining patterns for CD20, CD45, and/or CD5, which in most cases was accompanied by an altered light scatter pattern, suggesting a distinct subpopulation of smaller lymphocytes (14). All but one of these 11 cases were in the upper fifth percentile of the B-cell count distribution, and they accounted for about half (9/19, 47%) of the individuals in the uppermost percentile. Since the biomarker immunophenotype panel in our baseline studies was not targeted toward detection of MBL, it included only markers for the major lymphocyte subsets and not for κ/λ expression. The medical follow-up in 1997 was conducted to confirm the monoclonality of those cases we had putatively identified and to determine whether we had missed monoclonality in other individuals with high B-cell or CD5 B-cell counts. The criteria used to select the follow-up subjects were designed to identify persons with absolute increases in B-lymphocytes, or absolute or relative increases in CD5+ B-lymphocytes. All MBL cases identified at baseline met these criteria, suggesting that these criteria may identify individuals at risk for MBL. Of course, these criteria may not identify everyone at risk. Future studies should determine the optimal screening criteria for identifying individuals with an elevated risk of having or developing MBL.
By looking at individuals with the highest B-cell counts, we were able to observe the natural history of PCBL as well as MBL. Those with the highest CD5 B-cell counts who did not meet the total B-cell count criterion were included, because the CD5 B-cell is considered the primary target for the malignant transformation that leads to the most common phenotype of CLL (23). In fact, the MBL uncovered in the first follow-up that developed into aggressive CD5-B-CLL was eligible for follow-up only because of the CD5 criteria, as this participant did not have PCBL at baseline. This emphasizes the importance of including CD5 in any screening panel for B-cell lymphocytosis.
The immunophenotyping used for both medical follow-ups was performed at a clinical hematopathology laboratory using current methods for the differential diagnosis of B-cell malignancies (16, 17). The first follow-up included six of the original nine MBL cases, and monoclonality was confirmed in all six cases. Therefore, our interpretation of baseline immunophenotype staining patterns as monoclonal populations appears to have been highly specific. Two additional individuals with monoclonal populations were detected in the 1997 follow-up, and we cannot eliminate the possibility that those monoclonal populations were present at baseline and might have been detected by κ/λ analysis. However, the monoclonal populations were obvious phenotypically by 1997, suggesting that if clones had been present at baseline in those two individuals, they had expanded considerably by the first follow-up. As a rule, the κ/λ ratio is dependent upon whether or not heterologous antisera or monoclonal antibodies are used. The normal ratio is in the range of 2:1 for heterologous antisera and 1:1 for monoclonal antibodies. If either ratio is greater than four or less than 0.5, we make a greater effort during gating to detect the presence of a clone or an isolated population that is absent in one or the other distribution. We would point out, however, that the detection of a clone can occur in the presence of a normal κ/λ ratio. Conversely, the B cell clone light chain restriction is sometimes indeterminant.
The overall prevalence of MBL in our study population, 0.57%, was higher than the 0.14% (7/5,138) reported in a study of mid-Western US blood donors aged 39–80 years (24), but lower than the 3.5% (32/910) reported in a study of clinic outpatients over 40 years old (4). The prevalence among the adults aged 60 and older may reach up to 5% (4, 5). The highest reported prevalence of MBL to date is in unaffected family members of familial CLL cohorts: 18% (6/33) (25) and 13.5% (8/59) (26). It is now clear that a combination of gating on subpopulations of B-cells with distinct staining patterns for lineage markers and evaluating κ/λ ratios of the gated populations is the most sensitive approach to detecting monoclonality (4, 5). Thus, the observed prevalence of MBL can be highly dependent on the methods used to detect it, but differences in the study cohort compositions no doubt contribute to the wide range of reported prevalence rates. Since our base population was residential, our prevalence figure probably represents that of the general population more closely than previous reports.
Even from the limited number of cases we followed, our results suggest that MBL has certain distinct hematological features. The large majority (7/8, 87%) were kappa light chain restricted, which agrees with the previous findings by Ghia et al. (16/19, 84%) (5). The predominance of kappa-restriction would seem to functionally mimic the murine CD5 B1 population (27). MBL also differed from PCBL in having lower T-cell counts, as noted in an early observation of this phenomenon (1). Taken together, our findings are consistent with MBL arising from a persistent, innate, restricted B-cell immune response.
Our results also show that MBL can progress to B-CLL (9% or 1/11). Although this appears to be an uncommon pathway, it does suggest a highly elevated relative risk for developing B-CLL. The age-adjusted incidence of CLL in 2000 for the USA was 3.7 per 100,000 (28). The incidence of CLL increases dramatically with age: persons over 65 years of age had an incidence of 21.0 per 100,000, while those under 65 years of age had an incidence of 1.2 per 100,000. This risk may be influenced by sex: about half of our MBL cases were female, while B-CLL is at least twice as common in men than in women (21, 22, 28). Interestingly, the MBL case that developed CLL and the case that developed Waldenstroms Macroglobulinemia were both male.
Since MBL is easily detected, it provides a convenient biomarker to study the early stages of neoplastic transformation. Evolution of B-cell neoplasia is related to the mutational status of the immunoglobulin variable region genes VHDJH and VLJL (29). The restricted VH gene usage seen in CLL is also found in CLL-like MBL, and clonal homogeneity in the immunoglobulin heavy-chain gene is detectable in most MBL cases (30). Single cell PCR on purified B-cells would be informative in identifying mutated subpopulations within monoclonal expansions. Neoplastic transformation of B-cells is also clearly related to 13q14 interstitial deletions (31), which have been detected in ∼50% of CLL-like MBL (30). This deletion effects the functional and mutational status of the microRNA cluster coding for mir15 and mir16, critical regulators of bcl-2 mediated apoptosis (32). Determining the expression levels of cell-surface receptor CD38 (33, 34) and of cytoplasmic ZAP-70 (35, 36) in MBL would also shed light on the natural history of B-cell neoplastic transformation.
Our findings on PCBL show a striking regression to normal B-cell counts over the time course of the two follow-up studies (Fig. 5). This transient PCBL is clearly quite different than the PCBL associated with malignant or autoimmune diseases (9–13), or than the rare syndrome of persistent PCBL that is predominant in female smokers with the DR7 HLA haplotype (7, 8, 37–39). The regression to normal suggests that the PCBL we observed was largely a homeostatic response, i.e., a well-regulated component of humoral host defense. However, since 1 of the 51 (2%) individuals with baseline PCBL developed MBL, we cannot eliminate the possibility that PCBL is a risk factor for MBL. The relationship between PCBL, MBL, and B-cell malignancies can only be answered by longitudinal studies in a cohort that includes the entire spectrum of B-cell counts.
The apparent association between MBL and residence near a hazardous waste site also deserves further study. Other reports have suggested that certain environmental chemicals, such as pesticides and PCBs, increase the risk for B-cell malignancies (40–43). B-CLL has been considered a nonradiogenic form of cancer, but the basis for this conclusion has recently been questioned (44). This question is currently under investigation in an occupational cohort (45). Bauer has proposed that mutations that accumulate in B-cells can confer selective advantage when mutated B-cells undergo a proliferation in response to environmental stimuli (46). Thus, through their longevity, intrinsic genomic instability, and chronic responses to environmental stimuli, B-cells might serve as sentinel markers for oncogenic changes. A comprehensive analysis of the B-cell compartment in a longitudinal study using the most incisive laboratory techniques available would allow a closer look at this hypothesis in cohorts exposed to biological, chemical, and physical agents.
The authors thank Drs. Jeannine Holden and Robert Bray (Emory University Hospital) for performing flow cytometry analyses in the follow-up studies. We also thank Dr. Vikas Kapil for clinical support, Dr. Mohammed Uddin for developing a medical history questionnaire, Ms. Carolyn Harris for administrative support, and Dr. Jeff Lybarger for his support in the original studies. We are grateful to Dr. Neil Caporaso for many helpful discussions.