Abstract
While severe immune dysregulation is an established risk factor for non-Hodgkin lymphoma (NHL), it is unclear whether subclinical immune system function influences lymphomagenesis. To address this question, we conducted a nested case-control study within the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial to investigate whether circulating levels of cytokines and other immune markers are associated with future risk of NHL. Selected cytokines [interleukin (IL)-4, IL-6, IL-10, tumor necrosis factor (TNF)-α] and other immune markers [soluble TNF receptor 1 (sTNF-R1), sTNF-R2, C-reactive protein (CRP), sCD27] were measured in prediagnostic serum specimens from 297 incident NHL cases and 297 individually matched controls. Odds ratios (OR) and 95% confidence intervals (CI) relating quartiles of analyte concentration to NHL risk were calculated using conditional logistic regression. Statistically significant associations with increased NHL risk were observed for elevated serum levels of sTNF-R1 (quartile 4 vs. quartile 1: OR 1.7, 95% CI 1.1–2.8; Ptrend=0.02) and sCD27 (OR 5.3, 95% CI 2.9–9.4; Ptrend<0.0001). These associations remained in analyses of cases diagnosed 6+ years following blood collection (sTNF-R1: OR 2.1, 95% CI 1.0–4.0, Ptrend=0.01; sCD27: OR 4.1, 95% CI 1.9–8.5, Ptrend=0.0001). Elevated levels of IL-10, TNF-α and sTNF-R2 were also significantly associated with increased risk of NHL overall; however, these associations weakened with increasing time from blood collection to case diagnosis, and were null for cases diagnosed 6+ years post-collection. Our findings for sTNF-R1 and sCD27, possible markers for inflammatory and B-cell stimulatory states respectively, support a role for subclinical inflammation and chronic B-cell stimulation in lymphomagenesis.
Keywords: non-Hodgkin lymphoma, cohort studies, serum, cytokines, soluble markers, immune system
INTRODUCTION
Epidemiologic studies of individuals with AIDS and organ transplantation recipients have established severe immune dysregulation as a risk factor for non-Hodgkin lymphoma (NHL) (1–4). The effects of AIDS and drugs for combating transplant rejection include immune suppression, viral reactivation and chronic antigenic stimulation, all of which cause increased B cell division, which increases the potential for unrepaired genetic errors that can lead to lymphomagenesis (5). It is unclear, however, whether subclinical variation in immune system function influences NHL risk in the general population. Insight into this question may come from the study of cytokines -- secreted proteins that modulate various activities of B and T lymphocytes including proliferation, antibody production, chemotaxis, interaction with immunoglobulins, and the production of other cytokines. Cytokines of diverse functions operate in a complex, tightly-regulated network controlling immune system function. Perturbations of this network may increase the potential for chronic B cell division and, ultimately, lymphomagenesis.
Limited evidence from genetic and molecular epidemiologic studies support a role for cytokines in lymphomagenesis. Associations between genetic variants in cytokine genes (e.g., TNF, IL4, IL10) and susceptibility to NHL have been observed (6–10). More recently, findings from prospective studies of circulating levels of cytokines and other secreted immune markers and future NHL risk have been reported. Investigations within a cohort study of AIDS patients have identified associations with increased NHL risk for elevated levels of several immune markers, including interleukin-6 (IL-6), IL-10, and soluble CD27 (sCD27), a proposed marker of immune activation (11–13). A case-control study (92 cases, 184 controls) within the New York University – Women’s Health Study (NYU-WHS) cohort identified associations with decreased NHL risk for elevated serum levels of IL-5 and IL-13, and with increased risk for levels of sIL2R, TNF-α and soluble TNF receptor type 2 (sTNF-R2), a possible marker of TNF-α-driven inflammation (14). Another study (86 NHL cases, 86 controls) nested within the EPIC-Italy cohort observed associations with reduced risk for higher levels of IL-2 and TNF-α, and an association with increased risk for intercellular adhesion molecule (ICAM) (15). As well, a small study of prediagnostic plasma C-reactive protein (CRP) and multiple cancers nested within the EPIC-Greece cohort reported a statistically significant positive association with leukemia/lymphoma (27 cases) (16).
The sample sizes of these studies of cytokines and other immune markers were small, limiting their statistical power to detect associations of moderate magnitude and their ability to explore, through analyses stratifying on follow-up time, the possibility that associations are the result of reverse causation. The latter point is particularly relevant given that abnormal levels of immune markers can be caused by lymphomas, both directly (through production of cytokines by cancer cells) and indirectly (through changes in the immune system in response to the presence of disease) (17, 18).
In this report we present findings from a comparatively large case-control study (297 cases, 297 controls) nested within the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) designed to investigate whether serum levels of selected cytokines (IL-4, IL-6, IL-10, TNF-α) and other immune markers (sTNF-R1, sTNF-R2, CRP, sCD27) are associated with future risk of NHL. We previously reported from this study an association between prediagnostic circulating soluble CD30 levels and NHL that persisted for the longest period of follow-up, from 6–10 years (19).
MATERIALS AND METHODS
Detailed descriptions of the PLCO Trial have been previously reported (20, 21). In brief, between 1993 and 2001, approximately 155,000 subjects in ten cities (Birmingham, AL; Denver, CO; Detroit, MI; Honolulu, HI; Marshfield, WI; Minneapolis, MN; Pittsburgh, PA; Salt Lake City, UT; St Louis, MO; and Washington, DC) aged 55–74 were recruited from the general population and randomized to the screening or non-screening arm of the study. All screening-arm subjects provided nonfasting baseline blood samples that were processed and frozen within two hours of collection, and stored at −70°C. Individuals were followed up for all cancer diagnoses by annual mailed questionnaire, in addition to the Trial disease outcomes by annual screening examinations during the first six years of follow-up. All cancer diagnoses were pathologically confirmed through medical record abstraction. The institutional review boards of the US National Cancer Institute and the ten study centers approved the trial, and all participants provided written informed consent.
After follow-up through January 31, 2006, 297 cases of NHL (ICD-O-2-M 9590–9595, 9670–9677, 9680–9688, 9690–9698, 9700–9717, 9823, 9827) were identified from 54,829 eligible screening-arm participants (eligibility criteria: signed informed consent, no history of cancer at enrollment, two or more unthawed serum vials available, did not develop cancer or exit the cohort within the first year of follow-up). Controls were individually matched to cases on a one:one ratio on the basis of age at baseline (five-year categories), sex, race, PLCO center, and date of baseline blood draw (three-month categories) from among subjects who had not been diagnosed with any type of malignancy except non-melanoma skin cancer at the time of the case diagnosis date.
Multiplex assays for use on a Luminex multianalyte profiling (xMAP) system (Luminex Corp., Austin, TX) were used to measure several analytes. The xMAP system is a liquid array system that utilizes bead-bound analyte-specific reporter molecules to permit the simultaneous quantitative analysis of up to 100 different analytes in a single microtiter well. The specific multiplex array system used in these analyses was the Bio-Rad Bio-Plex 200 (Hercules, CA). Analytes measured using the Bio-Plex 200 system included the following: IL-4, IL-6, IL-10 and TNF-α, measured in duplicate (34 subjects per batch) using a 4-plex high-sensitivity kit (Millipore, Billerica, MA); sTNF-R1 and sTNF-R2, measured in duplicate (36 subjects per batch) using a 2-plex kit (Invitrogen/BioSource, Carlsbad, CA); and CRP, measured singly (72 subjects per batch) using a single-plex kit (Invitrogen/BioSource, Carlsbad, CA). Serum sCD27 was measured in duplicate (36 subjects per batch) by enzyme-linked immunosorbent assay (Bender Medsystems, Burlingame, CA). Values below assay limits of detection (LOD) were imputed using a parametric model-based estimation procedure (22). Using measurements among controls for a target analyte, we used maximum likelihood methods to estimate parameters for the log-normal distribution. For each measurement < LOD, we randomly sampled a value from the appropriate log-normal distribution as the imputed value. Samples from matched case-control sets were assayed consecutively within the same batch.
Masked replicate samples (N=72) were interspersed within and across assay batches to assess quality control. Intraclass correlation coefficients (intra-batch/inter-batch) were 0.98/0.93 for IL-4, 0.97/0.94 for IL-6, 0.96/0.79 for IL-10, 0.96/0.85 for TNF- α, 0.95/0.93 for sTNF-R1, 0.94/0.87 for sTNF-R2, and 0.96/0.93 for sCD27. Coefficients of variation (intra-batch/inter-batch) were 8.0%/26.4% for IL-4, 5.5%/27.5% for IL-6, 14.0%/38.3% for IL-10, 5.6%/31.0% for TNF-α, 3.5%/8.9% for sTNF-R1, 2.8%/11.8% for sTNF-R2, and 6.3%/19.8% for sCD27. For CRP, measurements from blinded quality control replicates suggested possible problems with one of the assay batches. We present results with this batch excluded [resulting sample size: 261 cases, 261 controls; coefficients of variation (intra-batch/inter-batch) 11.9%/40.3%; intraclass correlation coefficients 0.95/0.89], although our findings for the total data set of 297 cases and 297 controls were virtually identical (results not shown).
All statistical tests conducted in the analysis were two-sided. The Wilcoxon signed rank test was used to test for significant differences in levels of each analyte among the matched pairs of cases and controls. For each analyte, odds ratios (OR) and 95% confidence intervals (CI) relating concentration categories and NHL risk were then computed using conditional logistic regression modeling. For analyses of all NHL, analyte concentrations were categorized using control quartiles as cut-points. An exception is IL-4, measurements of which were below the LOD for 40% of all samples. For this analyte, samples below the LOD formed one category, while samples above the LOD were categorized using control tertiles. Tests for trend were performed by modeling the intra-category medians as a continuous parameter.
We evaluated the impact of follow-up time from blood collection to case diagnosis on our findings using two analytic methods. A “matched approach” involved fitting conditional logistic regression models using data from cases and their matched controls within strata defined by time from blood collection to case diagnosis (1–2, 3–5 and 6–10 years). An “unmatched approach” involved fitting a polytomous regression model comparing case subcategories defined by follow-up time to diagnosis (1–2, 3–5 and 6–10 years from blood collection) to all controls, with model adjustment for baseline age, sex, race, PLCO center, and enrollment year. The two methods have complementary strengths; the matched approach minimizes the impact of measurement error from between-batch assay variability on our findings, while the unmatched approach provides more stable risk estimates through the use of all controls as a single referent group. We tested for differences in analyte associations with NHL across strata of follow-up time in conditional models using the likelihood ratio test.
Analyses within strata defined by sex and age group were performed using conditional logistic regression modeling, with tests of interaction performed using the likelihood ratio test. Polytomous regression models were also fit to compute associations between dichotomized analyte levels (cut-point: control median) and individual NHL histologic subtypes [chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL); diffuse large B-cell lymphoma (DLBCL); follicular lymphoma (FL); other or not otherwise specified histologies (Other/NOS)].
RESULTS
Cases and controls had identical distributions of matching factors (Table 1). The majority of participants were male (63%) and self-identified as white (94%), with a median age at enrollment of 65 years. The most common NHL histologic subtypes were CLL/SLL (N=117, 39% of cases), DLBCL (62, 21%) and FL (44, 15%). The median length of follow-up from blood collection to case diagnosis was 5.0 years (range 1.0–10.8 years). In analyses among controls of immune marker levels, including sCD30 (19), several markers were correlated with one another (Supplemental Table 1), most notably IL-4 and IL-6 [Spearman rank correlation (r) = 0.80], sTNF-R1 and sCD30 (r = 0.39), sTNF-R1 and sTNF-R2 (r = 0.35), sTNF-R2 and TNF-α (r = 0.34), and sCD27 and sCD30 (r = 0.32). Levels of sTNF-R1, unlike sTNF-R2, were only weakly correlated with TNF-α (r = 0.15).
Table 1.
Selected characteristics of non-Hodgkin lymphoma (NHL) cases (N=297) and individually matched controls (N=297) selected from the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (1993–2006).
| Cases (N=297)
|
Controls (N=297)
|
|
|---|---|---|
| N(%) | N(%) | |
| Age at enrollment | ||
| 55–59 | 61 (20.5) | 61 (20.5) |
| 60–64 | 87 (29.3) | 87 (29.3) |
| 65–69 | 89 (30.0) | 89 (30.0) |
| 70–74 | 60 (22.2) | 60 (22.2) |
| Sex | ||
| Female | 111 (37.4) | 111 (37.4) |
| Male | 186 (62.6) | 186 (62.6) |
| Race | ||
| White | 280 (94.3) | 280 (94.3) |
| Black | 6 (2.0) | 6 (2.0) |
| Other | 11 (3.7) | 11 (3.7) |
| PLCO Center | ||
| Birmingham, AL | 9 (3.0) | 9 (3.0) |
| Denver, CO | 31 (10.4) | 31 (10.4) |
| Detroit, MI | 22 (7.4) | 22 (7.4) |
| Honolulu, HI | 7 (2.4) | 7 (2.4) |
| Marshfield, WI | 48 (16.2) | 48 (16.2) |
| Minneapolis, MN | 55 (18.5) | 55 (18.5) |
| Pittsburgh, PA | 46 (16.2) | 46 (16.2) |
| Salt Lake City, UT | 35 (11.8) | 35 (11.8) |
| St. Louis, MO | 24 (8.1) | 24 (8.1) |
| Washington, DC | 20 (6.7) | 20 (6.7) |
| Year of enrollment | ||
| 1993–1995 | 105 (35.4) | 105 (35.4) |
| 1996–1997 | 109 (36.7) | 109 (36.7) |
| 1998–2001 | 83 (28.0) | 83 (28.0) |
| NHL histologic subtype | ||
| CLL/SLL | 117 (39.4) | |
| DLBCL | 62 (20.9) | |
| FL | 44 (14.8) | |
| Other/NOS | 74 (24.9) | |
The distributions of analytes among cases and controls are summarized in Table 2. Cases had statistically significant elevated levels of IL-10, TNF-α, sTNF-R1, sTNF-R2 and sCD27 compared to their matched controls, with P-values ranging from 0.003 to <0.0001. Serum levels of IL-4, IL-6 and CRP did not differ between cases and controls (P ≥ 0.27). Similarly, in analyses of analyte categories (Table 3), statistically significant associations with increased NHL risk were observed for the highest categories of IL-10 (fourth quartile vs. first: OR 2.1, 95% CI 1.3–3.4, Ptrend=0.0001), TNF-α (OR 2.2, 95% CI 1.3–3.9, Ptrend=0.005), sTNF-R1 (OR 1.7, 95% CI 1.1–2.8, Ptrend=0.02) and, in particular, sCD27 (OR 5.3, 95% CI 2.9–9.4, Ptrend<0.0001). Categories of sTNF-R2 concentration were not associated with NHL at a level of statistical significance (OR 1.4, 95% CI 0.9–2.2, Ptrend=0.10).
Table 2.
Summary of serum levels of selected cytokines and other immune markers in non-Hodgkin lymphoma cases (N=297) and matched controls (N=297) selected from the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (1993–2006).
| Analyte | Units | Limit of Detection (LOD) | % of samples above LOD | Cases: Median (P5, P95) | Controls: Median (P5, P95) | P* |
|---|---|---|---|---|---|---|
| IL-4 | pg/ml | 0.13 | 60.4 | 8.3 (<0.1, 493.3) | 8.8 (<0.1, 467.6) | 0.94 |
| IL-6 | pg/ml | 0.10 | 98.8 | 5.5 (0.9, 68.2) | 5.9 (1.3, 79.8) | 0.84 |
| IL-10 | pg/ml | 0.15 | 95.3 | 5.5 (0.2, 59.7) | 4.0 (0.1–52.4) | 0.0002 |
| TNF-α | pg/ml | 0.05 | 99.8 | 6.5 (2.2, 13.8) | 5.8 (2.1–11.9) | 0.0003 |
| sTNF-R1 | pg/ml | 15 | 99.8 | 1170 (700, 2170) | 1095 (660, 1720) | 0.004 |
| sTNF-R2 | pg/ml | 15 | 99.8 | 910 (350, 1930) | 855 (320, 1410) | 0.003 |
| CRP | pg/ml | 2 | 100.0 | 620.9 (88.2, 3704.1) | 629.8 (83.5, 4806.7) | 0.27 |
| sCD27 | U/ml | 0.20 | 97.8 | 75.0 (24.2, 163.9) | 58.7 (27.3, 108.0) | <0.0001 |
Abbreviations: P5, 5th percentile; P95, 95% percentile.
P-value from Wilcoxon signed rank test of matched case-control pairs.
Table 3.
Associations between serum levels of selected cytokines and immune markers and non-Hodgkin lymphoma: the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (1993–2006).
| Analyte | Levels | NControls/NCases | OR* (95% CI) | PTrend |
|---|---|---|---|---|
| IL-4 (pg/ml) | ≤0.14 | 116/119 | 1.0 | 0.63 |
| 0.15–23.27 | 61/61 | 1.0 (0.6–1.5) | ||
| 23.28–102.38 | 61/52 | 0.8 (0.5–1.3) | ||
| >102.38 | 59/65 | 1.1 (0.7–1.8) | ||
| IL-6 (pg/ml) | ≤2.74 | 75/77 | 1.0 | 0.33 |
| 2.75–5.88 | 74/75 | 1.0 (0.6–1.6) | ||
| 5.89–16.11 | 74/59 | 0.8 (0.5–1.3) | ||
| >16.11 | 74/86 | 1.2 (0.7–1.9) | ||
| IL-10 (pg/ml) | ≤1.93 | 76/60 | 1.0 | 0.0001 |
| 1.94–3.96 | 74/49 | 0.8 (0.5–1.3) | ||
| 3.97–7.77 | 74/73 | 1.3 (0.8–2.0) | ||
| >7.77 | 73/115 | 2.1 (1.3–3.4) | ||
| TNF-α (pg/ml) | ≤3.96 | 76/60 | 1.0 | 0.005 |
| 3.97–5.78 | 74/68 | 1.3 (0.8–2.1) | ||
| 5.79–7.98 | 74/72 | 1.5 (0.9–2.7) | ||
| >7.98 | 73/97 | 2.2 (1.3–3.9) | ||
| sTNF-R1 (pg/ml) | ≤945 | 78/60 | 1.0 | 0.02 |
| 946–1095 | 73/59 | 1.1 (0.7–1.8) | ||
| 1096–1330 | 74/91 | 1.8 (1.1–2.9) | ||
| >1330 | 72/87 | 1.7 (1.1–2.8) | ||
| sTNF-R2 (pg/ml) | ≤670 | 81/84 | 1.0 | 0.10 |
| 671–855 | 69/49 | 0.7 (0.4–1.2) | ||
| 856–1050 | 76/65 | 0.9 (0.5–1.4) | ||
| >1050 | 71/99 | 1.4 (0.9–2.2) | ||
| CRP (pg/ml) | ≤242.98 | 66/57 | 1.0 | 0.42 |
| 242.99–629.75 | 65/74 | 1.4 (0.8–2.4) | ||
| 629.75–1671.60 | 65/76 | 1.5 (0.9–2.5) | ||
| >1671.60 | 65/54 | 1.0 (0.6–1.7) | ||
| sCD27 (U/ml) | ≤47.82 | 75/50 | 1.0 | <0.0001 |
| 47.83–58.71 | 74/41 | 0.9 (0.5–1.6) | ||
| 58.72–72.73 | 74/49 | 1.3 (0.8–2.3) | ||
| >72.73 | 74/157 | 5.3 (2.9–9.4) |
Statistically significant (P < 0.05) results shown in bold-face type.
Computed using conditional logistic regression modeling.
However, several of these associations were found to be affected by the length of time from blood collection to case diagnosis (Table 4). In both unmatched and matched analyses, prediagnostic levels of IL-10, TNF-α, sTNF-R2 were most strongly associated with cases diagnosed 1 or 2 years following blood collection. Associations with cases diagnosed 3–5 years after collection were weaker for IL-10 and TNF-α, and null for sTNF-R2. All three markers were not associated with NHL cases diagnosed 6 years or longer from collection. A test of heterogeneity in analytes associations across strata of follow-up time was statistically significant for sTNF-R2 (P=0.005) but not for TNF-α (P=0.22) or IL-10 (P=0.22). The association with NHL for sTNF-R1 was strongest for cases diagnosed 6 years or longer since blood collection (Ptrend = 0.07 and 0.01 for matched and unmatched analyses respectively). Serum levels of sCD27 were significantly associated with NHL at each time period. When we fit a polytomous model adjusting simultaneously for sCD27 and sTNF-R1, the association with sCD27 among cases diagnosed 6+ years post-collection remained statistically significant (Ptrend = 0.0004), while the sTNF-R1 association became of borderline significance (Ptrend = 0.07). When all other immune markers, including sCD30, were additionally adjusted for in the model, the findings did not materially change (sCD27, Ptrend = 0.002; sTNF-R1, Ptrend = 0.08). No evidence of statistical interaction with sex, age at baseline or other immune markers was observed.
Table 4.
Associations between immune marker levels and non-Hodgkin lymphoma (NHL), by time from blood collection to case diagnosis: the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (1993–2006).
| Analyte | Level | NControls | NHL cases, categorized by time from blood collection to case diagnosis (years)
|
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 – 2
|
3 – 5
|
6 – 10
|
|||||||||
| NCases | OR1* (95% CI) | OR2† (95% CI) | NCases | OR1* (95% CI) | OR2† (95% CI) | NCases | OR1* (95% CI) | OR2† (95% CI) | |||
| IL-10 (pg/ml) | ≤1.93 | 65 | 14 | 1.0 | 1.0 | 21 | 1.0 | 1.0 | 25 | 1.0 | 1.0 |
| 1.94–3.96 | 74 | 15 | 0.6 (0.2–1.5) | 1.1 (0.5–2.4) | 12 | 0.7 (0.3–1.7) | 0.6 (0.3–1.4) | 22 | 1.1 (0.5–2.5) | 0.9 (0.5–1.9) | |
| 3.97–7.77 | 74 | 15 | 1.2 (0.4–3.3) | 1.2 (0.5–2.7) | 27 | 1.3 (0.6–2.7) | 1.4 (0.7–2.7) | 31 | 1.3 (0.6–2.8) | 1.3 (0.7–2.4) | |
| >7.77 | 73 | 35 | 3.1 (1.2–8.5) | 2.8 (1.3–5.9) | 43 | 2.4 (1.1–5.3) | 2.2 (1.2–4.2) | 37 | 1.6 (0.8–3.2) | 1.5 (0.8–2.9) | |
| PTrend | P = 0.004 | P = 0.0006 | P = 0.009 | P = 0.001 | P = 0.20 | P = 0.13 | |||||
| TNF-α (pg/ml) | ≤3.96 | 76 | 15 | 1.0 | 1.0 | 23 | 1.0 | 1.0 | 22 | 1.0 | 1.0 |
| 3.97–5.78 | 74 | 18 | 1.2 (0.5–3.0) | 1.3 (0.6–2.9) | 20 | 1.1 (0.5–2.6) | 0.9 (0.5–1.9) | 30 | 1.6 (0.7–3.6) | 1.4 (0.7–2.7) | |
| 5.79–7.98 | 74 | 17 | 1.1 (0.4–3.6) | 1.7 (0.7–4.1) | 26 | 1.8 (0.7–4.8) | 1.3 (0.6–2.6) | 29 | 1.5 (0.7–3.7) | 1.3 (0.6–2.8) | |
| >7.98 | 73 | 29 | 3.6 (1.0–13.4) | 3.2 (1.4–7.5) | 34 | 3.3 (1.2–8.9) | 1.9 (0.9–3.8) | 34 | 1.5 (0.6–3.7) | 1.4 (0.7–2.9) | |
| PTrend | P = 0.07 | P = 0.004 | P = 0.01 | P = 0.05 | P = 0.55 | P = 0.46 | |||||
| sTNF-R1 (pg/ml) | ≤945 | 78 | 21 | 1.0 | 1.0 | 18 | 1.0 | 1.0 | 21 | 1.0 | 1.0 |
| 946–1095 | 73 | 20 | 1.2 (0.5–2.8) | 1.3 (0.6–2.6) | 21 | 1.3 (0.6–2.9) | 1.2 (0.6–2.5) | 18 | 0.8 (0.3–2.1) | 1.0 (0.5–2.0) | |
| 1096–1330 | 74 | 16 | 1.1 (0.5–2.8) | 0.9 (0.4–1.9) | 37 | 2.0 (0.8–4.6) | 2.3 (1.1–4.5) | 38 | 2.3 (1.0–5.2) | 2.1 (1.1–4.2) | |
| >1330 | 72 | 22 | 1.8 (0.7–4.9) | 1.4 (0.7–2.9) | 27 | 1.6 (0.7–3.6) | 1.7 (0.8–3.5) | 38 | 1.8 (0.8–4.2) | 2.1 (1.0–4.0) | |
| PTrend | P = 0.23 | P = 0.52 | P = 0.31 | P =0.11 | P = 0.07 | P = 0.01 | |||||
| sTNF-R2 (pg/ml) | ≤670 | 81 | 20 | 1.0 | 1.0 | 30 | 1.0 | 1.0 | 34 | 1.0 | 1.0 |
| 671–855 | 69 | 7 | 0.5 (0.2–1.7) | 0.4 (0.2–1.2) | 16 | 0.6 (0.3–1.4) | 0.6 (0.3–1.2) | 26 | 0.9 (0.4–1.9) | 0.9 (0.4–1.6) | |
| 856–1050 | 76 | 18 | 2.5 (0.9–7.0) | 1.0 (0.5–2.2) | 26 | 1.1 (0.5–2.4) | 0.9 (0.5–1.6) | 21 | 0.4 (0.2–0.9) | 0.6 (0.3–1.2) | |
| >1050 | 71 | 34 | 5.6 (1.8–2.2) | 2.3 (1.2–4.8) | 31 | 1.1 (0.5–2.4) | 1.1 (0.6–2.1) | 34 | 0.8 (0.4–1.8) | 1.1 (0.6–2.1) | |
| PTrend | P = 0.0007 | P = 0.004 | P = 0.53 | P = 0.60 | P = 0.41 | P = 0.88 | |||||
| sCD27 (U/ml) | ≤47.82 | 75 | 16 | 1.0 | 1.0 | 21 | 1.0 | 1.0 | 13 | 1.0 | 1.0 |
| 47.83–58.71 | 74 | 11 | 0.7 (0.3–1.9) | 0.8 (0.3–1.9) | 8 | 0.6 (0.2–1.6) | 0.4 (0.2–0.9) | 22 | 1.6 (0.7–4.1) | 1.7 (0.8–3.9) | |
| 58.72–72.73 | 74 | 12 | 2.0 (0.6–6.3) | 1.1 (0.4–2.5) | 13 | 0.8 (0.3–2.1) | 0.6 (0.3–1.4) | 24 | 1.9 (0.8–4.5) | 1.9 (0.8–4.1) | |
| >72.73 | 74 | 40 | 11.3 (2.7–48.0) | 3.6 (1.8–7.6) | 61 | 4.8 (1.9–12.4) | 3.3 (1.7–6.2) | 56 | 4.3 (1.7–11.1) | 4.1 (1.9–8.5) | |
| PTrend | P = 0.0005 | P < 0.0001 | P <0.0001 | P < 0.0001 | P = 0.002 | P = 0.0001 | |||||
Statistically significant (P < 0.05) results shown in bold-face type.
Computed using conditional logistic regression modeling within strata defined by time from blood collection to case diagnosis (1–2, 3–5 and 6–10 years).
Computed using polytomous regression modeling, comparing case subcategories defined by follow-up time to diagnosis (1–2, 3–5 and 6–10 years from blood collection) to all controls, with model adjustment for baseline age, sex, race, PLCO center, and enrollment year.
We explored further the associations with sTNF-R1 and sCD27 through analyses of common NHL subtypes (Table 5). The strongest subtype-specific association with elevated sTNF-R1 was observed for CLL/SLL, although associations with cases diagnosed 6 years or longer from blood collection were weaker and not statistically significant. Conversely, associations with DLBCL, FL and other/unclassified NHL cases were strongest for cases diagnosed 6–10 years post-collection. For sCD27, associations with CLL/SLL and FL were strongest for cases diagnosed 1–2 years from blood collection, and FL was not associated with elevated sCD27 for cases diagnosed 6–10 years later. DLBCL was significantly associated with elevated sCD27 only for cases diagnosed 6–10 years post-collection. Analyses of the other immune markers did not suggest clear associations with specific NHL subtypes, although elevated levels of IL-10, TNF-α and sTNF-R2 were significantly associated with increased risk of CLL/SLL cases diagnosed 1–2 years following blood collection (Supplemental Table 2).
Table 5.
Associations between immune marker levels and non-Hodgkin lymphoma (NHL) subtypes, by time from blood collection to case diagnosis: the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (1993–2006).
| Analyte | N Controls | Cases
|
Cases, categorized by time from blood collection to case diagnosis (years)
|
||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 – 2
|
3 – 5
|
6 – 10
|
|||||||
| NHL Subtype Level | N | OR* (95% CI) | N | OR* (95% CI) | N | OR† (95% CI) | N | OR† (95% CI) | |
| sTNF-R1 (pg/ml) | |||||||||
| Chronic lymphocytic leukemia/small lymphocytic lymphoma | |||||||||
| <1096 | 151 | 39 | 1.0 | 11 | 1.0 | 11 | 1.0 | 17 | 1.0 |
| ≥1096 | 146 | 78 | 2.1 (1.3–3.4) | 20 | 2.5 (1.0–5.8) | 32 | 2.9 (1.3–6.3) | 26 | 1.5 (0.7–3.1) |
| Diffuse large B-cell lymphoma | |||||||||
| <1096 | 151 | 30 | 1.0 | 11 | 1.0 | 8 | 1.0 | 11 | 1.0 |
| ≥1096 | 146 | 32 | 1.1 (0.6–2.0) | 3 | 0.2 (0.1–0.9) | 9 | 1.6 (0.6–4.4) | 20 | 2.3 (1.0–5.6) |
| Follicular lymphoma | |||||||||
| <1096 | 151 | 21 | 1.0 | 9 | 1.0 | 6 | 1.0 | 6 | 1.0 |
| ≥1096 | 146 | 23 | 1.2 (0.6–2.4) | 7 | 1.0 (0.3–3.0) | 8 | 1.3 (0.4–4.1) | 8 | 1.6 (0.5–5.7) |
| Other or not otherwise classifiable NHL | |||||||||
| <1096 | 151 | 29 | 1.0 | 10 | 1.0 | 14 | 1.0 | 5 | 1.0 |
| ≥1096 | 146 | 45 | 1.9 (1.1–3.3) | 8 | 0.6 (0.2–1.8) | 15 | 1.5 (0.6–3.7) | 22 | 5.7 (1.9–17.0) |
| sCD27 (U/ml) | |||||||||
| Chronic lymphocytic leukemia/small lymphocytic lymphoma | |||||||||
| <58.72 | 149 | 30 | 1.0 | 8 | 1.0 | 11 | 1.0 | 11 | 1.0 |
| ≥58.72 | 148 | 87 | 3.2 (2.0–5.3) | 23 | 4.2 (1.6–10.6) | 32 | 2.9 (1.4–6.3) | 32 | 2.8 (1.2–6.3) |
| Diffuse large B-cell lymphoma | |||||||||
| <58.72 | 149 | 22 | 1.0 | 8 | 1.0 | 7 | 1.0 | 7 | 1.0 |
| ≥58.72 | 148 | 40 | 1.7 (0.9–3.0) | 6 | 0.8 (0.2–2.4) | 10 | 1.9 (0.7–5.3) | 24 | 2.9 (1.1–7.5) |
| Follicular lymphoma | |||||||||
| <58.72 | 149 | 15 | 1.0 | 5 | 1.0 | 3 | 1.0 | 7 | 1.0 |
| ≥58.72 | 148 | 29 | 2.4 (1.2–5.0) | 11 | 5.6 (1.5–20.6) | 11 | 4.8 (1.1–20.2) | 7 | 0.8 (0.2–2.7) |
| Other or not otherwise classifiable NHL | |||||||||
| <58.72 | 149 | 24 | 1.0 | 6 | 1.0 | 8 | 1.0 | 10 | 1.0 |
| ≥58.72 | 148 | 50 | 2.3 (1.3–4.1) | 12 | 2.5 (0.8–7.6) | 21 | 3.3 (1.3–8.1) | 17 | 1.6 (0.7–3.9) |
Statistically significant (P < 0.05) results shown in bold-face type.
Computed using polytomous regression modeling, comparing NHL histologic subtypes to all controls, with model adjustment for baseline age, sex, race, PLCO center, and enrollment year.
Computed using polytomous regression modeling, comparing, for each NHL subtype, subcategories defined by follow-up time to diagnosis (1–2, 3–5 and 6–10 years from blood collection) to all controls, with model adjustment for baseline age, sex, race, PLCO center, and enrollment year.
DISCUSSION
In this investigation of serum levels of immune markers and subsequent risk of NHL, elevated pre-diagnostic levels of sTNF-R1 and sCD27 were significantly associated with increased risk, both overall and for cases diagnosed 6 years or longer from the date of blood collection. Elevated levels of IL-10, TNF-α and sTNF-R2 were also significantly associated with NHL overall; however, these associations weakened with increasing time from blood collection to case diagnosis, and were null for cases diagnosed 6 years or longer post-collection. We did not observe associations with risk for levels of IL-4, IL-6 or CRP.
sTNF-R1 and sTNF-R2 are the respective soluble fragments of tumor necrosis factor receptors I and II, the cell surface receptors through which TNF-α effects are mediated (23). These soluble receptors, shed from the cell surface in response to TNF-α (24) and other inflammatory stimuli (25–27), can bind with TNF-α and may play a regulatory role by limiting its circulating levels in the blood (28). sTNF-R1 and sTNF-R2 have been measured in several studies as possible markers of TNF-mediated inflammation, given the stability of their levels over time (29) and the high sensitivities of assays for these markers (28). In our study, elevated levels of both sTNF-R1 and sTNF-R2 were associated with future NHL risk. However, whereas the association with sTNF-R1 was consistently observed with increasing follow-up time, the association with sTNF-R2 was present only for cases diagnosed within two years of blood collection – CLL/SLL cases in particular. Given previous evidence that CLL cells release sTNF-R2 (30), our findings for sTNF-R2 may reflect reverse causation. As a possible marker for inflammation, our association with sTNF-R1 supports the hypothesis that pro-inflammatory effects may increase risk of NHL (31). The exact mechanism responsible for elevated sTNF-R1, however, is unclear. We note that among our controls TNF-α was more strongly correlated with sTNF-R2 than sTNF-R1; this is consistent with evidence suggesting that sTNF-R2 secretion is triggered by TNF-α, while sTNF-R1 may be regulated through different inflammatory triggers (32–34). An association with sTNF-R2, but not sTNF-R1, was observed in a smaller study conducted within the NYU-WHS cohort (14), and was found to remain after excluding from the analysis cases diagnosed <2 years from the blood collection date, although the impact of excluding cases with specimens collected longer than 2 years from diagnosis was not explored. Additional evidence from prospective studies is needed to clarify the relationship with NHL risk for these soluble markers.
As with sTNF-R1, levels of sCD27 were associated with future NHL risk, both overall and for cases diagnosed 6–10 years from blood collection. CD27 is a member of the TNF receptor superfamily, and is expressed exclusively on peripheral blood T-cells and a subset of B-cells (35). CD27 appears to be involved in activation of both T-cells and B-cells; the receptor, upon activation by its ligand CD70, stimulates T-cell proliferation (36–38) and enhances immunoglobulin production in B-cells (39, 40). The extracellular portion of CD27 is proteolytically cleaved from CD27-positive cells to produce a soluble form of the receptor (sCD27) detectable in serum. Elevated levels of sCD27 have been observed with many infectious and autoimmune diseases, and has been proposed as a marker of immune activation (35), similar to sCD30 (19). The B-cell stimulatory role of CD27 and the suspected importance of chronic B-cell activation in lymphomagenesis led Widney et al. to investigate the association between sCD27 and subsequent risk of AIDS-NHL in a small study (36 cases, 133 controls) nested within a cohort of HIV+ patients (13). The investigators found elevated sCD27 measured up to 18 months prior to diagnosis to be significantly associated with an increased risk of AIDS-NHL. Our study has now replicated that finding within a general population cohort with follow-up of up to 10 years post-collection. The associations with sCD27 for CLL/SLL and FL weakened with increased time from blood collection to case diagnosis. It is notable that these malignancies are commonly indolent in nature and can remain undetected for a long time. Elevated serum sCD27 has been previously reported in patients with lymphoid malignancies and associated with tumor load, and CD27 is commonly expressed in lymphoma tissue, supporting the idea that the elevated sCD27 levels among some lymphoma patients may be tumor-induced (41). However, we found that the association between sCD27 and DLBCL, an aggressive lymphoma type, was associated with elevated sCD27 only among cases diagnosed 6–10 years from blood collection, consistent with an etiologic effect. Further research is needed to clarify the basis of the association between prediagnostic sCD27 and future risk of NHL.
Serum levels of TNF-α and IL-10 were associated with future NHL risk in our study. However, these associations weakened with longer follow-up time, and were not statistically significant for cases diagnosed 6–10 years after blood collection. Tests of interaction across strata of follow-up time for these analytes were not statistically significant, although our study lacks power to detect such differences. In two smaller prospective studies of circulating immune markers and NHL risk in general population cohorts, inconsistent findings were reported for TNF-α; an association with increased NHL risk was observed in the NYU-WHS study (14), and with decreased risk in the EPIC-Italy investigation (15). In both studies, prediagnostic IL-10 levels were not associated with NHL risk. Our finding of statistically significant associations with levels of these cytokines only for cases diagnosed within 5 years of blood collection is suggestive of tumor-induced effects, although these results are also compatible with etiologic effects in the late stages of NHL development. Studies of TNF and IL10 knock-out mice (42–44) and of genetic polymorphisms in humans (6–10) strongly implicate both cytokines in B-cell lymphomagenesis. TNF-α is thought to influence lymphoma development through up-regulation of pro-inflammatory and anti-apoptotic signals, possibly via the nuclear transcription factor (NF)-κB pathway (6), while IL-10 may influence NHL risk through its B-cell stimulatory effects (12). Voorzander et al. reported that elevated IL-10 levels observed among NHL patients appeared to be caused by increased expression from lymphoma cells and stimulated lymphoma proliferation, suggesting that IL-10 is an autocrine lymphoma growth factor (45). In contrast, elevated levels of TNF-α among NHL patients appear to be produced by normal immune cells (45), although it is possible that lymphomas may play an indirect role in stimulating such production from local cells.
An important strength of our study is its sample size; the case series (N=297) has more than three times the number of cases included in the three previously published prospective studies of circulating immune markers and NHL risk in community-based cohorts. This comparatively large sample size provided us with good study power to detect associations of moderate magnitude, and to explore associations with individual NHL subtypes. Moreover, it enabled us to more aggressively investigate the effects of time from blood collection to case diagnosis on our findings beyond just the exclusion of cases diagnosed within one or two years from diagnosis, as was done in previous studies. As a result, we were able to discover that the apparent associations with IL-10, TNF-α and sTNF-R2 were not observed for cases diagnosed beyond five years of follow-up, suggestive of reverse causation. Another strength of this study is the standardized blood collection and processing protocol employed in the PLCO Trial, which specified that blood specimens be centrifuged, processed and frozen within two hours of blood collection, thus minimizing blood collection and processing procedures as sources of variation in serum analyte levels. This is particularly important for the measurement of immune markers, given concerns regarding the potential of ex vivo expression from leukocytes and enzymatic digestion of cytokines with prolonged time to processing (46, 47). By comparison, serum collected in the NYU-WHS cohort was frozen within approximately 3 hours after collection (48), while the median time from blood collection to storage of samples used in the EPIC-Italy cohort investigation was 4 hours (15).
This study also has limitations. The median follow-up time from blood collection to case diagnosis (5 years) in our study was relatively short for an investigation of NHL etiology, although our analyses restricted to cases diagnosed 6–10 years post-collection were less likely to be affected by reverse causation. As with other recent investigations that have used Luminex-based multiplex assays (14, 15, 49), we observed high between-batch assay variability for some of our analytes. However, our ICCs were high, indicating that the components of analyte variability due to within- and between-batch effects are relatively small compared to the variability between individuals. Measurement error from between-batch variability may have introduced bias, likely towards the null, in our unmatched analyses. Our matched analyses, however, would not have been affected by such bias, given that members of each case-control matched pair were assayed in the same batch.
In conclusion, in this prospective study we observed associations between elevated serum levels of sTNF-R1 and sCD27, possible markers for inflammatory and B-cell stimulatory states respectively, and increased future risk of NHL. These findings support a role for subclinical inflammation and chronic B-cell stimulation in lymphomagenesis, although more research is needed to better understand the biologic basis underlying these relationships
Supplementary Material
Acknowledgments
Financial support: This research was supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics and by contracts from the Division of Cancer Prevention, National Cancer Institute, NIH, DHHS.
This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. HHSN261200800001E. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. The authors thank Drs. Christine Berg and Philip Prorok, Division of Cancer Prevention, National Cancer Institute, the Screening Center investigators and staff of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, Mr. Tom Riley and staff, Information Management Services, Inc., Ms. Barbara O’Brien and staff, Westat, Inc., Ms. Jackie King and staff, BioReliance, Inc, and Ms. Georgina Mbisa, the Viral Technology Laboratory at SAIC-Frederick. Most importantly, we acknowledge the study participants for their contributions to making this study possible.
Footnotes
Potential conflicts of interest: none to declare
References
- 1.Penn I. Cancers complicating organ transplantation. N Engl J Med. 1990;323:1767–9. doi: 10.1056/NEJM199012203232510. [DOI] [PubMed] [Google Scholar]
- 2.Opelz G, Henderson R. Incidence of non-Hodgkin lymphoma in kidney and heart transplant recipients. Lancet. 1993;342:1514–6. doi: 10.1016/s0140-6736(05)80084-4. [DOI] [PubMed] [Google Scholar]
- 3.Beral V, Peterman T, Berkelman R, Jaffe H. AIDS-associated non-Hodgkin lymphoma. Lancet. 1991;337:805–9. doi: 10.1016/0140-6736(91)92513-2. [DOI] [PubMed] [Google Scholar]
- 4.Mbulaiteye SM, Biggar RJ, Goedert JJ, Engels EA. Immune deficiency and risk for malignancy among persons with AIDS. J Acquir Immune Defic Syndr. 2003;32:527–33. doi: 10.1097/00126334-200304150-00010. [DOI] [PubMed] [Google Scholar]
- 5.Fisher SG, Fisher RI. The epidemiology of non-Hodgkin’s lymphoma. Oncogene. 2004;23:6524–34. doi: 10.1038/sj.onc.1207843. [DOI] [PubMed] [Google Scholar]
- 6.Rothman N, Skibola CF, Wang SS, Morgan G, Lan Q, Smith MT, et al. Genetic variation in TNF and IL10 and risk of non-Hodgkin lymphoma: a report from the InterLymph Consortium. Lancet Oncol. 2006;7:27–38. doi: 10.1016/S1470-2045(05)70434-4. [DOI] [PubMed] [Google Scholar]
- 7.Lan Q, Zheng T, Rothman N, Zhang Y, Wang SS, Shen M, et al. Cytokine polymorphisms in the Th1/Th2 pathway and susceptibility to non-Hodgkin lymphoma. Blood. 2006;107:4101–8. doi: 10.1182/blood-2005-10-4160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Purdue MP, Lan Q, Kricker A, Grulich AE, Vajdic CM, Turner J, et al. Polymorphisms in immune function genes and risk of non-Hodgkin lymphoma: findings from the New South Wales non-Hodgkin Lymphoma Study. Carcinogenesis. 2007;28:704–12. doi: 10.1093/carcin/bgl200. [DOI] [PubMed] [Google Scholar]
- 9.Skibola CF, Curry JD, Nieters A. Genetic susceptibility to lymphoma. Haematologica. 2007;92:960–9. doi: 10.3324/haematol.11011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Skibola CF, Bracci PM, Nieters A, Brooks-Wilson A, De SS, Hughes AM, et al. Tumor necrosis factor (TNF) and lymphotoxin-alpha (LTA) polymorphisms and risk of non-Hodgkin lymphoma in the InterLymph Consortium. Am J Epidemiol. 2010;171:267–76. doi: 10.1093/aje/kwp383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Breen EC, van der Meijden M, Cumberland W, Kishimoto T, Detels R, Martinez-Maza O. The development of AIDS-associated Burkitt’s/small noncleaved cell lymphoma is preceded by elevated serum levels of interleukin 6. Clin Immunol. 1999;92:293–9. doi: 10.1006/clim.1999.4760. [DOI] [PubMed] [Google Scholar]
- 12.Breen EC, Boscardin WJ, Detels R, Jacobson LP, Smith MW, O’Brien SJ, et al. Non-Hodgkin’s B cell lymphoma in persons with acquired immunodeficiency syndrome is associated with increased serum levels of IL10, or the IL10 promoter -592 C/C genotype. Clin Immunol. 2003;109:119–29. doi: 10.1016/s1521-6616(03)00214-6. [DOI] [PubMed] [Google Scholar]
- 13.Widney D, Gundapp G, Said JW, van der Meijden M, Bonavida B, Demidem A, et al. Aberrant expression of CD27 and soluble CD27 (sCD27) in HIV infection and in AIDS-associated lymphoma. Clin Immunol. 1999;93:114–23. doi: 10.1006/clim.1999.4782. [DOI] [PubMed] [Google Scholar]
- 14.Gu Y, Shore RE, Arslan AA, Koenig KL, Liu M, Ibrahim S, et al. Circulating cytokines and risk of B-cell non-Hodgkin lymphoma: a prospective study. Cancer Causes Control. 2010;21:1323–33. doi: 10.1007/s10552-010-9560-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Saberi HF, Krop EJ, Scoccianti C, Krogh V, Palli D, Panico S, et al. Plasma cytokines and future risk of non-Hodgkin lymphoma (NHL): a case-control study nested in the Italian European Prospective Investigation into Cancer and Nutrition. Cancer Epidemiol Biomarkers Prev. 2010;19:1577–84. doi: 10.1158/1055-9965.EPI-09-1237. [DOI] [PubMed] [Google Scholar]
- 16.Trichopoulos D, Psaltopoulou T, Orfanos P, Trichopoulou A, Boffetta P. Plasma C-reactive protein and risk of cancer: a prospective study from Greece. Cancer Epidemiol Biomarkers Prev. 2006;15:381–4. doi: 10.1158/1055-9965.EPI-05-0626. [DOI] [PubMed] [Google Scholar]
- 17.Hsu SM, Hsu PL. Autocrine and paracrine functions of cytokines in malignant lymphomas. Biomed Pharmacother. 1994;48:433–44. doi: 10.1016/0753-3322(94)90004-3. [DOI] [PubMed] [Google Scholar]
- 18.Whiteside TL. Immune responses to malignancies. J Allergy Clin Immunol. 2010;125:S272–S283. doi: 10.1016/j.jaci.2009.09.045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Purdue MP, Lan Q, Martinez-Maza O, Oken MM, Hocking W, Huang WY, et al. A prospective study of serum soluble CD30 concentration and risk of non-Hodgkin lymphoma. Blood. 2009;114:2730–2. doi: 10.1182/blood-2009-04-217521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gohagan JK, Prorok PC, Hayes RB, Kramer BS. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: history, organization, and status. Control Clin Trials. 2000;21:251S–72S. doi: 10.1016/s0197-2456(00)00097-0. [DOI] [PubMed] [Google Scholar]
- 21.Hayes RB, Reding D, Kopp W, Subar AF, Bhat N, Rothman N, et al. Etiologic and early marker studies in the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial. Control Clin Trials. 2000;21:349S–55S. doi: 10.1016/s0197-2456(00)00101-x. [DOI] [PubMed] [Google Scholar]
- 22.Lubin JH, Colt JS, Camann D, Davis S, Cerhan JR, Severson RK, et al. Epidemiologic evaluation of measurement data in the presence of detection limits. Environ Health Perspect. 2004;112:1691–6. doi: 10.1289/ehp.7199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Brockhaus M, Schoenfeld HJ, Schlaeger EJ, Hunziker W, Lesslauer W, Loetscher H. Identification of two types of tumor necrosis factor receptors on human cell lines by monoclonal antibodies. Proc Natl Acad Sci U S A. 1990;87:3127–31. doi: 10.1073/pnas.87.8.3127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lantz M, Malik S, Slevin ML, Olsson I. Infusion of tumor necrosis factor (TNF) causes an increase in circulating TNF-binding protein in humans. Cytokine. 1990;2:402–6. doi: 10.1016/1043-4666(90)90048-x. [DOI] [PubMed] [Google Scholar]
- 25.Bargetzi MJ, Lantz M, Smith CG, Torti FM, Olsson I, Eisenberg SP, et al. Interleukin-1 beta induces interleukin-1 receptor antagonist and tumor necrosis factor binding protein in humans. Cancer Res. 1993;53:4010–3. [PubMed] [Google Scholar]
- 26.Landmann R, Keilholz U, Scheibenbogen C, Brockhaus M, Gallati H, Denz H, et al. Relationship between soluble tumor necrosis factor (TNF) receptors and TNF alpha during immunotherapy with interleukin-2 and/or interferon alpha. Cancer Immunol Immunother. 1994;38:113–8. doi: 10.1007/BF01526206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Tilg H, Trehu E, Atkins MB, Dinarello CA, Mier JW. Interleukin-6 (IL-6) as an anti-inflammatory cytokine: induction of circulating IL-1 receptor antagonist and soluble tumor necrosis factor receptor p55. Blood. 1994;83:113–8. [PubMed] [Google Scholar]
- 28.Brockhaus M. Soluble TNF receptor: what is the significance? Intensive Care Med. 1997;23:808–9. doi: 10.1007/s001340050416. [DOI] [PubMed] [Google Scholar]
- 29.Aderka D, Engelmann H, Shemer-Avni Y, Hornik V, Galil A, Sarov B, et al. Variation in serum levels of the soluble TNF receptors among healthy individuals. Lymphokine Cytokine Res. 1992;11:157–9. [PubMed] [Google Scholar]
- 30.Trentin L, Pizzolo G, Zambello R, Agostini C, Morosato L, Sancetta R, et al. Leukemic cells in hairy cell leukemia and B cell chronic lymphocytic leukemia release soluble TNF receptors. Leukemia. 1995;9:1051–5. [PubMed] [Google Scholar]
- 31.Hjelmstrom P. Lymphoid neogenesis: de novo formation of lymphoid tissue in chronic inflammation through expression of homing chemokines. J Leukoc Biol. 2001;69:331–9. [PubMed] [Google Scholar]
- 32.Spinas GA, Keller U, Brockhaus M. Release of soluble receptors for tumor necrosis factor (TNF) in relation to circulating TNF during experimental endotoxinemia. J Clin Invest. 1992;90:533–6. doi: 10.1172/JCI115891. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Aukrust P, Froland SS, Liabakk NB, Muller F, Nordoy I, Haug C, et al. Release of cytokines, soluble cytokine receptors, and interleukin-1 receptor antagonist after intravenous immunoglobulin administration in vivo. Blood. 1994;84:2136–43. [PubMed] [Google Scholar]
- 34.Bemelmans MH, Gouma DJ, Buurman WA. LPS-induced sTNF-receptor release in vivo in a murine model. Investigation of the role of tumor necrosis factor, IL-1, leukemia inhibiting factor, and IFN-gamma. J Immunol. 1993;151:5554–62. [PubMed] [Google Scholar]
- 35.Lens SM, Tesselaar K, van Oers MH, van Lier RA. Control of lymphocyte function through CD27-CD70 interactions. Semin Immunol. 1998;10:491–9. doi: 10.1006/smim.1998.0154. [DOI] [PubMed] [Google Scholar]
- 36.Goodwin RG, Alderson MR, Smith CA, Armitage RJ, VandenBos T, Jerzy R, et al. Molecular and biological characterization of a ligand for CD27 defines a new family of cytokines with homology to tumor necrosis factor. Cell. 1993;73:447–56. doi: 10.1016/0092-8674(93)90133-b. [DOI] [PubMed] [Google Scholar]
- 37.Hintzen RQ, Lens SM, Lammers K, Kuiper H, Beckmann MP, van Lier RA. Engagement of CD27 with its ligand CD70 provides a second signal for T cell activation. J Immunol. 1995;154:2612–23. [PubMed] [Google Scholar]
- 38.Kobata T, Agematsu K, Kameoka J, Schlossman SF, Morimoto C. CD27 is a signal-transducing molecule involved in CD45RA+ naive T cell costimulation. J Immunol. 1994;153:5422–32. [PubMed] [Google Scholar]
- 39.Maurer D, Fischer GF, Fae I, Majdic O, Stuhlmeier K, Von JN, et al. IgM and IgG but not cytokine secretion is restricted to the CD27+ B lymphocyte subset. J Immunol. 1992;148:3700–5. [PubMed] [Google Scholar]
- 40.Agematsu K, Kobata T, Yang FC, Nakazawa T, Fukushima K, Kitahara M, et al. CD27/CD70 interaction directly drives B cell IgG and IgM synthesis. Eur J Immunol. 1995;25:2825–9. doi: 10.1002/eji.1830251017. [DOI] [PubMed] [Google Scholar]
- 41.van Oers MH, Pals ST, Evers LM, van der Schoot CE, Koopman G, Bonfrer JM, et al. Expression and release of CD27 in human B-cell malignancies. Blood. 1993;82:3430–6. [PubMed] [Google Scholar]
- 42.Korner H, Cretney E, Wilhelm P, Kelly JM, Rollinghoff M, Sedgwick JD, et al. Tumor necrosis factor sustains the generalized lymphoproliferative disorder (gld) phenotype. J Exp Med. 2000;191:89–96. doi: 10.1084/jem.191.1.89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Batten M, Fletcher C, Ng LG, Groom J, Wheway J, Laabi Y, et al. TNF deficiency fails to protect BAFF transgenic mice against autoimmunity and reveals a predisposition to B cell lymphoma. J Immunol. 2004;172:812–22. doi: 10.4049/jimmunol.172.2.812. [DOI] [PubMed] [Google Scholar]
- 44.Czarneski J, Lin YC, Chong S, McCarthy B, Fernandes H, Parker G, et al. Studies in NZB IL-10 knockout mice of the requirement of IL-10 for progression of B-cell lymphoma. Leukemia. 2004;18:597–606. doi: 10.1038/sj.leu.2403244. [DOI] [PubMed] [Google Scholar]
- 45.Voorzanger N, Touitou R, Garcia E, Delecluse HJ, Rousset F, Joab I, et al. Interleukin (IL)-10 and IL-6 are produced in vivo by non-Hodgkin’s lymphoma cells and act as cooperative growth factors. Cancer Res. 1996;56:5499–505. [PubMed] [Google Scholar]
- 46.House RV. Cytokine measurement techniques for assessing hypersensitivity. Toxicology. 2001;158:51–8. doi: 10.1016/s0300-483x(00)00401-7. [DOI] [PubMed] [Google Scholar]
- 47.Thavasu PW, Longhurst S, Joel SP, Slevin ML, Balkwill FR. Measuring cytokine levels in blood. Importance of anticoagulants, processing, and storage conditions. J Immunol Methods. 1992;153:115–24. doi: 10.1016/0022-1759(92)90313-i. [DOI] [PubMed] [Google Scholar]
- 48.Gu Y, Zeleniuch-Jacquotte A, Linkov F, Koenig KL, Liu M, Velikokhatnaya L, et al. Reproducibility of serum cytokines and growth factors. Cytokine. 2009;45:44–9. doi: 10.1016/j.cyto.2008.10.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Wong HL, Pfeiffer RM, Fears TR, Vermeulen R, Ji S, Rabkin CS. Reproducibility and correlations of multiplex cytokine levels in asymptomatic persons. Cancer Epidemiol Biomarkers Prev. 2008;17:3450–6. doi: 10.1158/1055-9965.EPI-08-0311. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.