Chromosomal Aberrations in Peripheral Blood Lymphocytes and Risk for Non-Hodgkin Lymphoma

Abstract

Chromosomal aberrations (CAs) are thought to be integrative biomarkers that reflect exposure to chromosome-damaging carcinogens and host factors. To investigate whether CAs indicate non-Hodgkin lymphoma (NHL) risk, we evaluated 200 metaphase spreads each for 67 incident low-grade, untreated NHL cases and 57 controls matched on age, sex, and storage time of cryopreserved lymphocytes. Hyperdiploidy of 47 chromosomes was statistically significantly associated with increased NHL risk with odds ratios of 1.4 (97% confidence interval [CI] = 0.6–3.5) and 3.5 (95% CI = 1.1–10.9) for medium and high levels of hyperdiploidy, respectively, compared to the lowest level (P-trend = .04). Hypodiploidy of 43 and 44 chromosomes increased NHL risk 3.3-fold (95% CI = 1.2–8.7) and 2.2 (95% CI = 1.0–5.2), respectively, compared to those without the event; total hypodiploidy was only moderately associated with risk. Chromosome and chromatid breaks were not associated with NHL risk. Our data suggest for the first time that aneuploidy identified in cultured, peripheral lymphocytes may be potential indicators of NHL risk.

Genetic damage in a nontarget tissue such as peripheral blood lymphocytes may reflect similar events in target tissue. If targeted events can be identified for specific cancers, such events may serve as early indicators of DNA lesions and potentially be used to identify individuals at highest future risk for disease. Chromosomal aberrations (CAs) are thought to be integrative biomarkers that reflect exposure to chromosome-damaging carcinogens as well as host susceptibility factors, such as metabolism of environmental exposures, DNA repair capacity, and genomic maintenance. CAs can comprise aneuploidy (chromosomal loss or gain) and chromosomal translocations, both well-recognized hallmarks of clonal chromosomal abnormalities in non-Hodgkin lymphoma (NHL) tumor tissue (1).

There is a growing body of evidence from cohort studies that indicate CAs in peripheral lymphocytes may be predictive of subsequent risk of all cancers (2–6), with one study specifically demonstrating increased risk for hematopoietic malignancies (7). Initial studies in a Nordic cohort study of cancer incidence and an Italian cohort of cancer mortality reported doubling of cancer incidence and mortality, respectively, for all cancers with high levels of CAs, when compared with the lowest level (2,7,8). Increased risk with increasing frequency of CAs was also reported (6). Standardized mortality ratios have also been reported to be highest for lymphoid neoplasms (7). Associations were further found to be independent of some carcinogenic exposures, consistent with the notion that increased CA levels also reflect a “cancer-prone state” that can be attributed to host factors (8). Another nested case–control study that evaluated specific chromosome-type frequencies found much higher elevated risks for cancer among individuals with chromosome-type aberrations (5). However, these studies have relatively small numbers of total cancers and minimal numbers of subjects for any single cancer.

To specifically investigate whether CAs in peripheral blood lymphocytes are associated with NHL risk, we conducted a study of CAs within the population-based multicenter National Cancer Institute-Surveillance Epidemiology and End Results (SEER) NHL case–control study.

Methods

Study Population

The study population was derived from a subset of participants from a case–control study of non-Hodgkin lymphoma that has previously been described in detail (9). Briefly, the case–control study comprised 1321 newly diagnosed NHL cases identified in four SEER registries (Iowa; Detroit, MI; Los Angeles, CA; Seattle, WA) aged 20–74 years between July 1998 and June 2000 without evidence of HIV infection. Each registry provided NHL pathology and subtype information derived from abstracted reports by the local diagnosing pathologist. All cases were histologically confirmed and coded according to the International Classification of Diseases for Oncology, 2nd Edition (10). In all, 1057 population controls were identified by random digit dialing (<65 years) and from Medicare eligibility files (≥65 years). Overall participation rates were 76% in cases and 52% in controls; overall response rates were 59% and 44%, respectively. Written informed consent was obtained from each participant before interview, in accordance with US Department of Health and Human Services guidelines. This study was approved by the institutional review boards at the National Institutes of Health and at each participating SEER site (Iowa, Seattle, Los Angeles, and Detroit).

One 10-mL tube of venous blood was collected from untreated cases and a subset of controls and shipped at room temperature from the study field site to the biorepository where the blood was processed and lymphocytes were separated and cryopreserved via a static liquid nitrogen vapor phase freezing method. All samples were handled and processed using the same protocol in the field, during cryopreservation, and during culturing and scoring.

For the current analysis, 100 untreated cases and 100 controls, matched on age, sex, and storage time of cryopreserved lymphocytes, were selected for study. To avoid the possibility that chemotherapy would affect CA rates in cases, any patient who was treated (eg, had received chemotherapy) was excluded from the analysis. In addition, we also selected cases with less severe disease (eg, stage 1) to reduce the possibility that the disease state itself could influence even nonclonal chromosomal events.

From 2003 to 2004, the cryopreserved lymphocytes were shipped overnight in a dry shipper saturated with liquid nitrogen to Cancer Genetics (Cambridge, MA) where they were subsequently cultured for 72 hours. The median length of storage in liquid nitrogen was 4.6 years (range: 3.4–5.5 years). The laboratory was blinded to case and control status. Of the 100 cases and 100 controls selected, 78 cases and 70 controls had viable cells where metaphase spreads could be adequately and subsequently prepared. Reasons for missing data included no growth from the lymphocytes (n = 23), no evidence of mitotic cells (n = 17), low mitotic index (n = 7), no cells (n = 3), bacterial contamination (n = 1), and missing data (n = 1). We further restricted our analyses to the 67 cases and 57 controls for whom at least 100 metaphase spreads were evaluable.

Metaphase Scoring

For viable samples, 200 metaphase spreads were prepared and scored for CAs in the unbanded form according to the International System for Human Cytogenetic Nomenclature. Numerical chromosomal aberrations (NCAs) and structural chromosomal aberrations (SCA) were measured. NCA included hypodiploidy (<46 chromosomes), specifically for <35, ≤42, 43, 44, and 45 chromosomes, and hyperdiploidy (>46 chromosomes), specifically for 47, 48, 49, and ≥50 chromosomes. SCA included both chromatid-type aberrations: chromatid gap, break, deletion, triradial formation, quadriradial formation, complex formation, and total chromatid-type exchanges; and chromosome-type aberrations: chromosome gap, break, double minute, dicentric chromosome, ring chromosome, acentric fragment, marker chromosome, premature centromere division, and total chromosome-type exchanges. Total NCA, a measure of aneuploidy, and SCA, a measure of total CA, were also evaluated. Quality control samples (triplicates for two samples) were also evaluated for all CAs.

Statistical Analysis

For descriptive characteristics (sex, race, study center, age, smoking), we first compared cases and controls with CA data to those without CA data to confirm that cell viability was independent of these characteristics. For all NCA and SCA, we evaluated the presence and median number of events per 100 cells for cases compared to controls.

For all NCA and SCA variables, we evaluated the distribution among controls and categorized dichotomously (by median), by tertiles and by quartiles. Based on these categories, we calculated the association between the variables and NHL with analysis by logistic regression to determine odds ratios and 95% confidence intervals. Based on these estimates, we further identified whether there was a trend or a threshold effect; if the latter was evident, we combined categories to increase our sample size and power to evaluate the threshold effect. All estimates were adjusted for age, sex, race, and study site. All statistical analyses were conducted using SAS version 9.1 (SAS Institute, Cary, NC).

Results

Demographic characteristics between cases and controls were equivalent (Table 1). Although the two main NHL subtypes in the parent case–control study comprised diffuse large B-cell lymphoma (32%) and follicular lymphoma (24%), our selection for untreated cases resulted in a case group defined largely by small lymphocytic lymphoma or chronic lymphocytic leukemia (28%) and follicular lymphoma (28%).

Table 1.

Description of cases and controls evaluated for chromosomal aberrations in the NCI-SEER NHL case–control study, restricted to those with 100 or more metaphases

Characteristics	Cases (n = 67)		Controls (n = 57)		P
	N, %		N, %
Age
Median (years)	61	(30–74)	60	(30–74)	.8
Sex
Male	32	47.8%	27	47.4%	1.0
Female	35	52.2%	30	52.6%
Race
Black	3	4.5%	9	15.8%	.1
White	59	88.1%	44	77.2%
Other or unknown	5	7.5%	4	7.0%
Study center
Detroit	9	13.4%	8	14.0%	.7
Iowa	22	32.8%	13	22.8%
Los Angeles	14	20.9%	14	24.6%
Seattle	22	32.8%	22	38.6%
NHL subtype
SLL or CLL	19	28.4%
Follicular	19	28.4%
Marginal zone	14	20.9%
DLBCL	5	7.5%
Mycosis fungoides	3	4.5%
Mantle cell	1	1.5%
Burkitt	1	1.5%
T/NK-other	1	1.5%
Not otherwise  specified	4	6.0%

SLL = small lymphocytic lymphoma, CLL = chronic lymphocytic leukemia, DLBCL = diffuse large B-cell lymphoma.

The presence of specific CAs varied widely among our controls (Table 2). In general, of the NCA evaluated, hypodiploidy occurred more frequently (100%) than hyperdiploidy (75%). We found no evidence of polyploidy. Of SCA evaluated, chromatid-type exchanges largely comprised chromatid breaks. Chromosome-type exchanges included acentric fragments, premature centromere division, and marker chromosomes. Chromosome-type exchanges occurred in virtually all controls (96%) and were slightly more frequent than total chromatid-type exchanges (88%). Of NCA events evaluated, presence of hyperdiploidy was generally higher in cases (84%) than controls (75%). Specifically, hyperdiploidy of 49 chromosomes was statistically significantly present more in cases than controls (P = .03). We found no statistically significant difference between cases and controls in the presence of any SCA events evaluated.

Table 2.

Presence and median number of NCA and SCA-specific events in cases and controls per 100 metaphases

	Control (n = 57)	Case (n = 67)		Control (n = 57)	Case (n = 67)		Case:control median ratio
Chromosomal aberrations	N (%)	N (%)	P	median (range)	median (range)	P
NCA
Hypodiploidy
<35 Chromosomes	51 (89%)	54 (81%)	.2	2.5 (0–15.0)	1.7 (0–15.2)	.4	0.68
≤42 Chromosomes	56 (98%)	64 (96%)	.4	4.5 (0–20.0)	4.0 (0–23.5)	.5	0.89
43 Chromosomes	51 (89%)	62 (93%)	.5	1.6 (0–11.3)	2.0 (0–7.0)	.9	1.25
44 Chromosomes	55 (96%)	63 (94%)	.5	3.0 (0–14.0)	4.0 (0–15.0)	.2	1.33
45 Chromosomes	57 (100%)	67 (100%)	1	4.5 (0.5–23.0)	5.5 (0.6–25.5)	.2	1.22
Total hypodiploidy   (<46 chromosomes)	57 (100%)	67 (100%)	1	16.9 (2.5–54.5)	18.5 (3.0–59.5)	.6	1.09
Hyperdiploidy
47 chromosomes	35 (61%)	50 (75%)	.1	0.5 (0–6.2)	1.0 (0–24.5)	.02	2.00
48 chromosomes	17 (30%)	19 (28%)	.9	0 (0–1.0)	0 (0–3.9)	.4	—
49 chromosomes	6 (11%)	17 (25%)	.03	0 (0–1.0)	0 (0–1.6)	.04	—
≥50 chromosomes	14 (25%)	19 (28%)	.6	0 (0–1.5)	0 (0–2.9)	.4	—
Total hyperdiploidy   (>46 chromosomes)	43 (75%)	56 (84%)	.3	1.0 (0–7.1)	1.5 (0–25.5)	.02	1.5
Polyploidy
Triploidy	0 (0%)	0 (0%)	—	0 (0–0)	0 (0–0)	—	—
Tetraploidy	0 (0%)	0 (0%)	—	0 (0–0)	0 (0–0)	—	—
Total polyploidy	0 (0%)	0 (0%)	—	0 (0–0)	0 (0–0)	—	—
Total NCA	57 (100%)	67 (100%)	1	17.7 (2.5–55.5)	19.5 (3.0–59.5)	.2	1.10
SCA
Chromatid-type exchanges
Chromatid break	35 (61%)	33 (49%)	.2	0.5 (0–3.0)	0 (0–5.7)	.6	—
Deletion of one or   portion of chromatid	0 (0%)	0 (0%)	—	0 (0–0)	0 (0–0)	—	—
Triradial Formation of the   chromatids	4 (7%)	4 (6%)	.8	0 (0–1.0)	0 (0–0.9)	.8	—
Quadriradial formation of   the chromatids	1 (2%)	0 (0%)	.3	0 (0–0.5)	0 (0–0)	.3	—
Complex formation of   the chromatids	0 (0%)	0 (0%)	—	0 (0–0)	0 (0–0)	—	—
Chromosome-type   exchanges
Chromosome break	8 (14%)	7 (10%)	.5	0 (0–1.7)	0 (0–0.6)	.2	—
Double minute	17 (30%)	14 (21%)	.3	0 (0–2.5)	0 (0–3.7)	.6	—
Dicentric chromosome	5 (9%)	11 (16%)	.2	0 (0–0.5)	0 (0–1.0)	.2	—
Ring chromosome	1 (2%)	2 (3%)	.7	0 (0–0.5)	0 (0–0.6)	.6	—
Acentric fragment	30 (53%)	32 (48%)	.6	0.5 (0–3.5)	0 (0–4.3)	.8	—
Marker chromosome	21 (37%)	20 (30%)	.4	0 (0–1.6)	0 (0–4.8)	1.0	—
Premature centromere   division	24 (42%)	34 (51%)	.3	0 (0–11.4)	0.5 (0–7.5)	.9	—
Total SCA	56 (98%)	65 (97%)	.7	4.0 (0–22.4)	4.0 (0–21.7)	.4	1.0

NCA = Numerical chromosomal aberrations; SCA = Structural chromosomal aberrations.

Although nearly all individuals demonstrated at least one aberration among the 200 metaphases evaluated, the median number of events was low. Among controls, total hypodiploidy was more common (16.9 per 100 metaphases) than hyperdiploidy, which occurred on average only 1.0 per 100 metaphases (Table 2). The median number of total hypodiploidy and total hyperdiploidy per 100 metaphases were both higher in our cases compared to our controls. Of hyperdiploidy events, the median number of hyperdiploidy of 47 and 49 chromosomes and total hyperdiploidy were statistically significantly higher in cases than controls. We found no statistically significant difference between the median number of chromatid- and chromosome-type exchanges between cases and controls.

Based on evaluation of NCA and SCA by categories—dichotomously by tertiles and by quartiles we identified thresholds of risk for hypodiploidy of 43, 44, and 45 chromosomes. For hypodiploidy of 43 chromosomes, there was a 3.3-fold (95% CI = 1.2–8.7) increase in NHL risk among those with greater than 0.5 events per 100 metaphases (Table 3). Elevated risks among the highest quartiles were observed for both hypodiploidy of 44 and 45 chromosomes with risks of 2.2 (95% CI = 1.0–5.2) and 1.3 (95% CI = 0.6–3.0), respectively. The cumulative measurement of total hypodiploidy did not confer a more pronounced or statistically significant risk increase for NHL.

Table 3.

Association between numerical chromosomal aberrations by final categories, restricted to subjects with 100 or more metaphases

		Control (n = 57)	Case (n = 67)
Numerical chromosomal abberations	Number of events per 100 metaphases	N (%)	N (%)	Unadjusted OR (95% CI)	Adjusteda OR (95% CI)
Hypodiploidy
43 Chromosomes	≤0.5	17 (30)	9 (13)	1.0 (ref)	1.0 (ref)
	>0.5	40 (70)	58 (87)	2.7 (1.1–6.8)	3.3 (1.2–8.7)
44 Chromosomes	≤4.5	44 (77)	40 (60)	1.0 (ref)	1.0 (ref)
	>4.5	13 (23)	27 (40)	2.3 (1.0–5.0)	2.2 (1.0–5.2)
45 Chromosomes	≤6.5	39 (68)	39 (58)	1.0 (ref)	1.0 (ref)
	>6.5	18 (32)	28 (42)	1.6 (0.7–3.3)	1.3 (0.6–3.0)
Total hypodiploidy	≤16.94	29 (51)	27 (40)	1.0 (ref)	1.0 (ref)
	>16.94	28 (49)	40 (60)	1.5 (0.8–3.1)	1.4 (0.6–3.1)
Hyperdiploidy
47 Chromosomes	0	22 (39)	17 (25)	1.0 (ref)	1.0 (ref)
	>0 to ≤1.5	27 (47)	27 (40)	1.3 (0.6–3.0)	1.4 (0.6–3.5)
	>1.5	8 (14)	23 (34)	3.7 (1.3–10.4)	3.5 (1.1–10.9)
				P-trend = .01	P-trend = .04
48 Chromosomes	≤0.5	54 (95)	59 (88)	1.0 (ref)	1.0 (ref)
	>0.5	3 (5)	8 (12)	2.4 (0.6–9.7)	2.0 (0.5–8.8)
49 Chromosomes	0	51 (89)	50 (75)	1.0 (ref)	1.0 (ref)
	>0	6 (11)	17 (25)	2.9 (1.1–7.9)	2.5 (0.9–7.1)
Total hyperdiploidy	≤0.5	25 (44)	21 (31)	1.0 (ref)	1.0 (ref)
	>0.5 to ≤2	21 (37)	22 (33)	1.2 (0.5–2.9)	1.2 (0.5–3.1)
	>2	11 (19)	24 (36)	2.6 (1.0–6.5)	2.2 (0.8–6.2)
				P-trend = .047	P-trend = .14
Total NCA	≤12.5	20 (35)	18 (27)	1.0 (ref)	1.0 (ref)
	>12.5	37 (65)	49 (73)	1.5 (0.7–3.2)	1.4 (0.6–3.4)

OR = odds ratio.

^aAdjusted for age, race, sex, and study site.

The highest categorical levels for hyperdiploidy of 47, 48, and 49 chromosomes were associated with increased NHL risk with risk estimates of 3.5 (95% CI = 1.1–10.9), 2.0 (95% CI = 0.5–8.8), and 2.5 (95% CI = 0.9–7.1), respectively (Table 3). We further observed a trend for increasing risk among increasing levels for hyperdiploidy of 47 chromosomes when compared to the lowest quartile (P-trend = .04). Finally, increasing numbers of total hyperdiploidy (>46 chromosomes) indicated increased risk for NHL. In analysis that included all data (Supplemental Table available online), the highest categorical level for total hyperdiploidy was associated with a 2.6-fold (95% CI = 1.1–6.1) increased NHL risk with increasing numbers of total hyperdiploidy statistically significantly increasing NHL risk (P-trend = .03). In analysis restricted to those with 100 or more metaphases, the increased risk was also evident but the P-trend was statistically significant only in unadjusted analysis (Table 3). Although high levels of total NCA indicated increased risk for NHL, the estimate was not statistically significant.

Evaluation of all SCA events dichotomously, by tertiles and by quartiles, yielded no pattern of risk increase for any event. We thus observed no associations between chromatid- or chromosome-type exchanges with NHL risk.

Discussion

We observed hypodiploidy and, in particular, hyperdiploidy measured from peripheral blood lymphocytes more often in NHL patients than in controls, a pattern consistent with the known occurrence of chromosomal losses and gains in NHL (1). Although it is possible that hypodiploidy can be artificially increased, this association, if real, may reflect an underlying tendency toward genomic instability, which could be associated with increased risk of NHL. At the same time, such aneuploidy in peripheral lymphocytes could reflect exposure to carcinogens that alter chromosome copy, although it is uncertain about the extent to which the events we measured reflect long-term past exposure. As specific exposures and genotoxic agents that increase NHL risk remain largely unknown, detection of NCA as an intermediate marker of risk may shed further mechanistic light on potential environmental agents and genetic susceptibility that cause NHL. Further investigations to identify the specific chromosomal losses and gains that comprise the NCA events detected in peripheral blood lymphocytes are warranted.

We did not observe increased risk of NHL with SCA events (chromosome- or chromatid-type exchanges). Some of the previous studies that found elevated risks were conducted among populations exposed to specific occupational or environmental agents [eg, benzene (4), arsenic (5)], although SCAs were found to be associated with risk of all cancer regardless of previous carcinogenic exposure (8). In addition, chromosome-type aberrations were somewhat less common in our population than has been previously reported, so we may have had inadequate power. Alternatively, SCA in peripheral lymphocytes may not be associated with risk of NHL.

To date, there have been only a few studies that have evaluated the association of nonclonal CAs in cultured peripheral lymphocytes with risk of cancer. Although these studies were prospective cohort studies, very few NHL cases were identified, and for most studies, an association between CAs specific to NHL could not delineated. Measurement of CAs for almost all of these studies was also carried out in many different laboratories and aneuploidy was not evaluated in these studies.

Study strengths include our analyses on the largest number of NHL cases to date and the most comprehensive evaluation of nonbanded CAs, which included both NCAs and SCAs. Our investigation of lymphocytes in individuals with mostly low-grade, untreated NHL also allowed the measurement of early biologic effect markers in a tissue that is less likely to be altered due to disease. Despite this restriction to our cases, however, we cannot exclude the possibility that our results were driven by potential disease effects. Finally, our evaluation also addressed limitations of previous studies that lacked specific CA-type data, including aneuploidy, as well as had possible misclassification of CA measurements due to nonuniform laboratory analyses. Specifically, all samples in the current study were handled and processed using the same protocol in the field, during cryopreservation, and during culturing and scoring, and laboratories were blinded as to case and control status.

Study limitations include the modest risk estimates we observed; although the largest study of NHL to date to assess CAs, we may have lacked sufficient power for evaluation for many of these relatively uncommon events. However, identifying the specific causes of NCA events may result in clarification of risk factors that have a more substantial impact of risk on NHL and also inform our mechanistic understanding of lymphomagenesis. Another potential limitation of our study was the use of cryopreserved, rather than fresh, samples. Chromatid aberrations have been reported as consequences of cryopreservation (11). Although we matched on the duration of cryopreservation, the process may have introduced artifacts in our CA measurements; such artifacts, if present, would have occurred similarly among our cases and controls and biased our results to the null. Finally, as with other studies of CAs, our estimate of levels in peripheral blood lymphocytes is likely an underestimate due to the use of unbanded metaphase preparation (8). Although a recent evaluation of CA with cancers that included NHL did not identify an association, no data on aneuploidy were reported (12).

Our data therefore suggest for the first time that aneuploidy, and especially hyperdiploidy, may be genetic damage reflective of NHL risk that can be identified in cultured, peripheral lymphocytes. We hypothesize that some of these events reflect those that occur in the target tissue, and further evaluation of specific chromosomal losses, gains, and other events such as translocations by fluorescent in situ hybridization and polymerase chain reaction (PCR)–based methods are thus warranted within both the nontarget and target tissues. It is unknown whether the NCA events we detected similarly reflect a “cancer-prone state” of genetic predisposition and/or reflect exposure to relevant unidentified carcinogens that increase NHL risk. Evaluation of NCA events in the context of genetic predisposition and environmental exposures are therefore also warranted.

Funding

Public Health Service (PHS) contracts N01-PC-65064, N01-PC-67008, N01-PC-67009, N01-PC-67010, N02-PC-71105.