Full Text
Lymphoma is a heterogeneous and complex group of lymphocytic neoplasms, which is accompanied by a broad range of immune regulatory disorder phenotypes. Although great progress has been obtained using chemotherapy and targeted approaches over recent decades, the overall treatment efficacy of lymphoma is still disappointing. Therefore, it is of vital importance to explore its detailed mechanism and find novel therapeutic targets for lymphoma.
Abnormal immune mechanism is considered critical in the progress of lymphoma.1 Immune dysfunction exists in lymphoma patients, and T cells, as the most important component of the immune system, are found to be numerically and functionally defective.2–4 Our previous results also showed that circulating Th17 immune response in lymphoma patients may be associated with the response to treatment and with different stages of diseases.5 Furthermore, targeting the immune molecules has been shown to be effective against lymphoma.4,6 Therefore, further researches on pathological mechanism in the function of the immune system are required.
It has been reported that there were various identified pathogenic mutations in immune-related genes, which were involved in BCR and TLR signaling, NF-κB pathway, cell apoptosis, and other functional pathways.7 Moreover, many other genetic polymorphisms of immune-related genes were reported to be closely related to the susceptibility and risk of lymphoma. Butterbach et al. found that polymorphisms of nine genes including IFNG, IL12A, and STAT family were significantly associated with the risk of lymphoma.8 A polymorphism in the interferon regulatory factor 4 gene, rs12211228, has a significant association with NHL risk.9 And the gene polymorphisms of IL-10 (rs1800890, rs1800896, and rs1800872) and IL-18 (rs1946518) were also associated with the risk or susceptibility of lymphoma.1,10 Patients with AA genotype in rs1585215 of NFKB1 in the NF-κB pathway were more likely to develop Stage IV and extranodal lesions when diagnosed. Besides, AG genotype of TNF-α was more likely to develop EBV-related chronic lymphocytic leukemia than homozygous GG.11 As one of the genes that control the activity of T lymphocytes, protein tyrosine phosphatase nonreceptor 22 (PTPN22) has also been reported to play an important role in the development of many autoimmune diseases in recent years. PTPN22 dephosphorylates NLRP3 inflammatory cells to limit the retention of NLRP3 in autophagy and to degrade it in autolysosomes,12 and its rs2476601 genotype can down-regulate the TCR signal system and reduce the affinity of Csk and T cells.13 In addition, Integrin Alpha M (ITGAM) and Poly ADP ribose polymerase-1 (PARP1) were also proved to be a new prognostic marker in diffuse large B-cell lymphoma (DLBCL)14 and mantle cell lymphoma15 Moreover, we have previously reported the positive associations between NLRP3 inflammasome associated polymorphisms and certain hematological diseases in the hematopoietic system, including lymphoma.1
To deeply discover new markers related to the etiology, progression and prognosis of lymphoma, we used the Agena Massarray technology platform to detect 25 candidate SNPs of 21 genes (TNFA, IFNG, IFNL3, IL2, IL4, IL6R, IL9, IL10, IL12A, IL22, TGFB1, TBX21/TBET, STAT1, STAT3, STAT5B, STAT6, GATA3, NFKBIA, ITGAM, PTPN22, PARP1) in 245 lymphoma patients and 213 healthy controls.8,10,16–57 All the 25 candidate SNPs were based on previous research studies about immune-related genetic polymorphisms on the susceptibility, prognosis, progression and stage of many cancers. And we confirmed them by SNP database, ClinVar or dbVar database that the genetic polymorphisms were associated with drug response, risk factors, and pathogenesis. Then we further analyzed the susceptibility of lymphoma and the relationship between SNPs and clinical characteristics. Our results will provide evidence for clinical diagnosis and management of lymphoma.
MATERIALS AND METHODS Characteristics of the subjectsA total of 245 patients with lymphoma (131 males and 114 females) were included in this study from March 2008 to October 2021 at Qilu Hospital of Shandong University with a median age of 53 years (14–90 years). The diagnosis and classification of lymphoma were confirmed strictly according to the WHO (2016) classification and Ann Arbor stage.58 Meanwhile, 213 healthy controls (125 males and 88 females) participated in the study, with a median age of 51 years (16–92 years). The basic characteristics of all participants, such as age, gender, the peripheral blood count, lactate dehydrogenase (LDH), and stratification of lymphoma patients were shown in Table 1. Informed consent was obtained from all patients before enrollment in the study in accordance with the Declaration of Helsinki. This study was conducted with the approval of the institutional ethics committee (KYLL-202204-059).
TABLE 1 The demographic and clinical characteristics of lymphoma patients and healthy controls.
| Parameters | Cases, n (%) | Controls, n (%) |
| Age | ||
| Median(range) | 53 (14–90) | 51 (16–92) |
| >60 | 91 (37.14) | 65 (30.52) |
| ≤60 | 154 (62.86) | 148 (69.48) |
| Sex | ||
| Male | 131 (53.47) | 125 (58.69) |
| Female | 114 (46.53) | 88 (41.31) |
| WBC (×109/L) | ||
| Median(range) | 6.07 (0.88–69.82) | NA |
| >15 | 16 (6.53) | NA |
| ≤15 | 229 (93.47) | NA |
| PLT (×109/L) | ||
| Median(range) | 231 (9–687) | NA |
| >250 | 100 (40.82) | NA |
| ≤250 | 145 (59.18) | NA |
| HGB (g/L) | ||
| Median(range) | 125 (33–162) | NA |
| >110 | 174 (71.02) | NA |
| ≤110 | 71 (28.98) | NA |
| LDH (U/L) | ||
| Median(range) | 239 (117–2493) | NA |
| >Normal | 123 (52.34) | NA |
| ≤Normal | 112 (47.66) | NA |
| Subtype | ||
| MBN | 200 (81.63) | NA |
| Non-MBN | 45 (18.37) | NA |
| Ann Arbor stage | ||
| I | 21 (8.57) | NA |
| II | 26 (10.61) | NA |
| III | 88 (35.92) | NA |
| IV | 110 (44.90) | NA |
Abbreviations: HGB, hemoglobin; LDH, lactate dehydrogenase; MBN, Mature B-Cell Neoplasms; NA, not applicable; Non-MBN, Non-mature B-Cell Neoplasms; PLT, platelet count; WBC, white blood cell.
Leukocytes of peripheral blood or bone marrow were collected from the participants involved in our study. For the same patient, the DNA and RNA were extracted from the same type of biological sample.59–62 Genomic DNA was isolated by standard methods according to the manufacturer's instruction of TIANamp blood DNA kit (Tiangen Biotech). The quality and concentration of DNA were determined using a DeNovix DS-11 Spectrophotometer (DeNovix Inc.).
The detection of genotypingThe selected functional SNPs or hot SNPs were all summarized and aligned through the dbSNP and the UCSC databases. Then Agena's Assay Designer 4.0 software was used to evaluate the primers which were designed for multiple SNP sites. Finally, the PAGE primer purification method was used to synthesize three primers which corresponded with each SNP site, two PCR primers and one UEP primer. The genomic DNA was applied for genotyping assay of the selected SNPs using the Massarray platform (Agena Bioscience). Approximately 20 ng of genomic DNA from each sample was used for genotyping. After the DNA samples were amplified via multiplex PCR, allele detection was performed through MALDI-TOF mass spectrometry.
Reverse transcription and quantitative polymerase chain reaction (Total RNA from bone marrow or peripheral blood was extracted with TRIZOL™ reagent (REF: 15596018; Life Technologies) and detected by DeNovix DS-11 Spectrophotometer. Then the RNA was reversely transcribed into cDNA using Evo M-MLV RT Premix for qPCR (Code No. AG11706; Accurate Biotechnology Co., Ltd). Quantitative PCR (qPCR) was conducted by SYBR Green qPCR Mix (AH0104; SparkJade), and the relative gene expression was determined using a Roche LightCycler 480II system. All amplification primers and housekeeping gene β-actin primers were designed by Primer Premier 5.0, and the sequences were as follows: IL-12A forward: CTC CAG AAG GCC AGA CAA ACT C, IL-12A reverse: GCC AGG CAA CTC CCA TTA GTT A; IFNG forward: AAG TGA TGG CTG AAC TGT CG, IFNG reverse: TAC TGG GAT GCT CTT CGA CC; β-Actin forward: GAA GAG CTA CGA GCT GCC TGA; β-actin reverse: CAG ACA GCA CTG TGT TGG CG. And the mRNA expression was analyzed by the comparative cycle threshold (Ct) method.
Statistical analysisThe chi-square test was used to determine the difference in gender or age between the lymphoma patients and controls. Efficient implementations of exact tests were used to conduct the Hardy–Weinberg equilibrium of all 25 candidate SNPs. SNPs in HWE (p > 0.001) and minor allele frequency (MAF) >1% in the general population were included in the study. The relationships between SNPs phenotype or allele frequency and lymphoma susceptibility, blood routine indices, lymphoma stages, or survival analysis were determined in our study. Briefly, univariate binary logistic regression was used to analyze the differences between SNPs and susceptibility, peripheral blood count and LDH. Multivariate logistic regression analysis was used to analyze the relationship between Ann Arbor stage and lymphoma SNPs. All odds ratios (ORs) and corresponding 95% confidence intervals (95% CIs) were analyzed, and also adjusted for age and gender and corrected by Bonferroni to correct the analytic results. The purpose of Bonferroni correction in our study was to reject all the possibility of false positive results, and eliminate false positive results by the p-value adjusted.
Cox proportional hazards model was used to analyze the relationship between the Overall survival (OS) of lymphoma patients and SNPs, and the hazard ratio (HR) and corresponding 95% CI were also adjusted and corrected by Bonferroni. The gene expression was calculated by the Mann–Whitney test. Adjusted two-tailed p-value < 0.05 was considered statistically significant. Efficient implementations of exact tests were performed by Plink software, and other statistical analyses were performed using IBM SPSS Statistics 22 software (SPSS Inc.) and R (version 4.1.2).
To further validate the accuracy of our results, the Least Absolute Shrinkage and Selection Operator (LASSO) regression was applied to feature selection and model construction. The “lasso” minimizes the residual sum of squares subject to the sum of the absolute value of the coefficients being less than a constant. Because of the nature of this constraint it tends to produce some coefficients that are exactly zero and hence gives interpretable models.63 In our study, the R-package “glmnet” was used to select the most effective prognostic markers in 25 candidate SNPs and its corresponding regression coefficient by the penalty parameter λ with the least error through the ten-fold cross-validation. The “caret” package was used to get the 70% training data from all the SNPs data sets randomly, and each genetic model was run 1000 times to select the most accurate model. The performance of prediction model was evaluated by receiver operating characteristic (ROC) curves. Besides, the hazard ratios of the selected SNPs obtained from the LASSO regression model were calculated by univariate Cox analysis, and the results were also adjusted by age and gender. In addition, based on the 30% testing data predicted by LASSO regression model, we performed Wilcoxon analysis for the two survival states.
RESULTS The characteristics of all participantsThe characteristics of lymphoma patients and controls are shown in Table 1. There was no significant difference in the distribution of age (median 53, range 14–90 versus median 51, range 16–92) and gender (male/female: 131/114 vs. 125/88) between lymphoma patients and controls. Among 245 lymphoma patients, 15 were Hodgkin's Lymphoma and 230 were non-Hodgkin's lymphoma, which including 200 mature B-cell neoplasms and 30 other types of non-Hodgkin's lymphoma. According to the criteria of Ann Arbor stage, the lymphoma group was further divided into four subgroups, including 21 cases at Stage I, 26 at Stage II, 88 at Stage III, and 110 at Stage IV. In this research, we calculated the Hardy–Weinberg equilibrium and minor allele frequency (MAF) and evaluated the applicability of all 25 candidate SNPs. Due to the low detection rate of rs2227491, the SNP was excluded to ensure the reliability of the result, and the remaining 24 SNP loci were used for the following analysis (Table 2).
TABLE 2 Selected genes and SNPs.
Abbreviations: HWE, Hardy–Weinberg equilibrium; MAF, minor allele frequency (>1%); SNP, single-nucleotide polymorphisms.
Relationship between immune-related genetic polymorphisms and lymphoma susceptibilityFive genetic models (dominant model, recessive model, overdominant model, codominant model and additive model) and allele model were used to analyze the relationship between immune-related SNPs and lymphoma. After preliminary screening by χ2 test or Fisher's exact test, the influence of age or gender on the susceptibility of lymphoma was excluded. The results of Univariate binary logistic regression analysis showed that rs12979860 of IFNL3 was significantly associated with lymphoma susceptibility under dominant, additive and allele models, and patients with the T genotype were 2.049 times more susceptible to lymphoma than patients with C genotype (all p < 0.05). The rs1800629 of TNFA was also significantly associated with the susceptibility to lymphoma in the dominant, overdominant, codominant, additive and allelic models, and the risk of patients with genotype A was lower than that with genotype G (all p < 0.05), indicating the protective role for genotype A. For IL4, the rs2070874 and rs2243250 were correlated with the risk of lymphoma in the dominant, overdominant and codominant models, and the patients with CT genotype were significantly lower than those with the TT genotype. The OR of these two SNPs were 0.575 (0.386–0.857) and 0.568 (0.379–0.849), respectively, suggesting that the C genotype may be a protective factor. In addition, the risk of the AA genotype in rs1805414 of PARP1 was significantly lower than that of the GG genotype (OR = 0.375, 95% CI = 0.148–0.950, p = 0.043) under the codominant model.
As JAK–STAT and NF-κB pathways were involved in lymphoma, we first found that the rs744166 of STAT3 and rs6503691 of STAT5B contributed to lymphoma susceptibility under codominant, recessive, codominant, and allele models (all p < 0.05). Secondly, our results showed that the rs2233406 of NFKBIA in NF-κB pathway was significantly associated with the susceptibility of lymphoma in a variety of models (dominant, overdominant, codominant, additive, or allele models). And the mutation frequency of genotype A in lymphoma patients was significantly lower than that in controls, while genotype G was significantly higher in patients with lymphoma (p = 0.006), which indicated that genotype G is a risk factor for lymphoma susceptibility (Table 3). Therefore, our results reveal that some of our immune-related genetic polymorphisms are associated with the lymphoma susceptibility.
TABLE 3 Association between selected SNPs and lymphoma susceptibility.
Abbreviations: CI, confidence interval; OR, odds ratio; SNP, single nucleotide polymorphisms.
The immune-related genetic polymorphisms are associated with the Ann Arbor stage in lymphomaPatients with lymphoma have various degrees of lymph node or organ involvement, and the accurate and repeatable staging is crucial for the treatment and prognosis of lymphoma. In this study, according to the Ann Arbor Stage, this batch of lymphoma patients were divided into four Stages: I, II, III and IV.
For all lymphoma patients, the results of logistic regression analysis showed that patients with CA/CC genotype of IL6R rs2228145 had higher stage than patients with AA genotype under the dominant model (OR = 2.682, 95% CI = 1.084–7.033, p = 0.034), and patients with CA genotype had higher stage than patients with AA genotype under codominant model (OR = 2.741, 95% CI = 0.999–7.520, p = 0.048). In addition, the rs6503691 CT genotype of STAT5B was significantly different from that of the CC/TT genotype in different stages under the overdominant model (p = 0.030) (Table 5).
Due to the large scale of the MBN group, we further analyzed the relationship between the Ann Arbor stage and SNPs for the MBN lymphoma patients. Logistic regression analysis showed that the differences between the rs2228145 and the rs6503691 at different stages in MBN patients were similar to that in all lymphoma patients. Moreover, the rs6887695 genotype of IL12A was significantly different in four stages under codominant, recessive, additive and allele models (p < 0.05), and the risk of progression of lymphoma with C allele was much lower than that of the G allele (OR = 0.529, 95% CI = 0.280–1.001, p = 0.048), indicating that genotype G is a risk factor for lymphoma progression (Table 4).
TABLE 4 Association between Ann Arbor stage and SNPs in lymphoma.
Abbreviations: CI, confidence interval; OR, odds ratio; SNP, single nucleotide polymorphisms.
Relationship between immune-associated genetic polymorphisms and peripheral blood count andClinical variables, especially the peripheral blood count and LDH, were reported to be involved in the lymphoma.64 We analyzed the relationship between immune-related SNPs and peripheral blood count or LDH. In our study, a WBC higher than 15 × 109/L was defined as a high WBC, no higher than 15 × 109/L was defined as a low WBC,65 a PLT not higher than 250 × 109/L is defined as a low PLT, PLT higher than 250 × 109/L is a high PLT; HGB 110 g/L or lower is a low HGB, higher than 110 g/L is defined as a high HGB.66 In addition, according to the upper limit of the reference range of our hospital (≤ 230 U/L), we called abnormal increased LDH when it was >230 U/L. We performed binary logistic regression analysis in all lymphoma cases to determine the SNPs that may affect the clinical indicators, and to explore the roles of WBC, PLT, HGB or LDH in the prognosis of lymphoma.
The results of the binary logistic regression analysis showed that lymphoma patients with GA genotype in rs744166 of STAT3 were more likely to have low PLT than AA or AA/GG under the overdominant and codominant models (OR = 1.873, 95% CI = 1.060–3.307, p = 0.032; OR = 1.827, 95% CI = 1.071–3.117, p = 0.028) (Table 5). As for HGB, the genotype frequency of rs3824662 in GATA3 under dominant, overdominant or codominant models was significantly different between high and low HGB groups. Patients with CA or CA/AA genotype were more likely to develop anemia than those with CC genotype (OR = 2.019, 95% CI = 1.105–3.690, p = 0.021; OR = 1.867, 95% CI = 1.048–3.326, p = 0.031), indicating that A allele may be a negative factor. And the patients with rs907187 carrying CG genotype of PARP1 was 2.091 and 2.209 times higher than that in patients with CC or CC/GG (OR = 2.091, 95% CI = 1.092–4.005, p = 0.030; OR = 2.209, 95% CI = 1.239–3.940, p = 0.009) (Table 5). However, we cannot observe a significant relationship between SNPs and WBC count.
TABLE 5 Association between peripheral blood counts and SNPs in lymphoma.
Abbreviations: CI, confidence interval; OR, odds ratio; SNP, single nucleotide polymorphisms.
As for the importance of LDH in lymphoma,67 we further analyzed the relationship between the genetic polymorphism and LDH. Our results showed that the rs2069762 of IL2 had a significant difference between the high and low groups of LDH used by overdominant, codominant and recessive models, and the risk of abnormal increased LDH in patients with CA genotype was 2.021 or 2.289 times higher than that in patients with AA or AA/CC genotype, respectively (p < 0.05). Under the codominant and recessive model, the risk of LDH abnormality in patients with rs1800871 GG genotype of IL10 was 3.603 times higher than that in patients with AA genotype. In the codominant, additive or allele models, there was a significant difference of rs3824662 genotype of GATA3 between the two LDH groups, and the patients with AA genotype had a lower risk of increased LDH than those with CC genotype (OR = 0.392, 95% CI = 0.160–0.965, p = 0.040) (Table 6).
TABLE 6 Association between LDH count and SNPs in lymphoma.
Abbreviations: CI, confidence interval; OR, odds ratio; SNP, single nucleotide polymorphisms.
The immune-related genetic polymorphisms are associated with the survival of lymphoma patientsFirst, we investigated the relationship of OS with clinical characteristics. We found that there existed significant differences between the age ≥ 60 group and the age < 60 group (p = 0.044). And there was no significant difference in OS between males and females (p > 0.05). In addition, OS among the four stages of Ann Arbor stage exhibited remarkable differences (p = 0.0037), among which the OS was negatively related to the clinical stages (Figure 1A).
Enlarge this image.FIGURE 1. The OS of Ann Arbor stages and two genetic polymorphisms in lymphoma patients. (A). The OS of Ann Arbor Stage I, II, III, IV. (B).The OS of lymphoma patients with AA, AG, GG genotypes in rs2069718 under codominant and additive models. (C). The OS of lymphoma patients with AA, AG/GG genotypes in rs2069718 under dominant model. (D). The OS of lymphoma patients with CC, GC, GG genotypes in rs6887695 under codominant model. (E). The OS of lymphoma patients with GG, GC/CC genotypes in rs6887695 under dominant model (F). The OS of lymphoma patients with GG/CC, GC genotypes in rs6887695 under overdominant model.
We further analyzed the relationship between SNPs and the OS of lymphoma patients under five genetic models and allele models. Univariate Cox proportional hazards regression model showed that the frequency of rs2069718 genotype of IFNG was significantly associated with the survival of lymphoma patients under dominant, codominant, additive and allele models (all p < 0.05) (Figure 1B,C). The risk about survival of lymphoma patients with the AG genotype was 1.760 times higher than that of the AA genotype, and the G allele genotype was 1.726 times higher than that of the A allele, suggesting that the G genotype was a poor prognostic factor. Consistently, the result of Log-rank (Mantel-Cox) analysis showed that the survival time of AG/GG genotype was significantly shorter than that of AA genotype (p = 0.032). In addition, the rs6887695 of IL12A was also significantly associated with the survival of lymphoma under the dominant, overdominant or codominant models. In codominant or overdominant models, the risk about survival of patients with GC genotype was 2.3 times higher than that of GG or GG/CC genotype (2.301 (1.254–4.225) or 2.380 (1.391–4.072)), and the survival time was shorter than that of GG or GG/CC (Figure 1D,F). After Bonferroni multiple corrections, the rs6887695 was also significantly associated with OS under the overdominant model. Besides, in the dominant model, patients with GC/CC genotype had an increased risk of OS compared with that genotype GG (HR = 1.825, 95% CI = 1.011–3.294) in rs6887695 of IL12A (Figure 1E). Therefore, genotype GC may be an adverse prognostic factor (Table 7).
TABLE 7 Association between OS and SNPs in lymphoma.
Abbreviations: CI, confidence interval; HR, Hazard ratio; SNP, single nucleotide polymorphisms.
We performed the LASSO regression to further validate our analysis results. LASSO regression and univariate Cox regression indicated that the rs2069718 of IFNG and rs6887695 of IL12A were correlated with the survival of the lymphoma patients under dominant and codominant models (all p < 0.05), and the AUC (Area Under Curve) were all greater than 0.62. The rs2069718 of IFNG was associated with the patients survival under additive model, and the rs6887695 of IL12A was associated with the patients' survival under the overdominant model, and the AUC of two SNPs were 0.69 (HR = 2.876, 95% CI = 1.488–5.559, p = 0.002) and 0.76 (HR = 2.246, 95% CI = 1.230–4.101, p = 0.008) separately. However, we cannot obtain some meaningful markers under the recessive model. The LASSO regression of the overdominant model was shown in Figure 2 and other genetic models can be seen in (Figure S1: codominant model; Figure S2: additive model; Figure S3: dominant model). All the results of five genotype models we got by LASSO regression models were generally consistent with the results obtained by the Cox proportional hazards models. The LASSO does not focus on subsets but rather defines a continuous shrinking operation that can produce coefficients that are exactly zero. Therefore, we identified that the SNPs associated with lymphoma patients' survival were the rs2069718 of IFNG and the rs6887695 of IL12A.
Enlarge this image.FIGURE 2. The selection of the variables by LASSO regression under overdominant model. (A). The variation characteristics of the coefficient of variables. (B). The selection process of the optimum value of the parameter λ in the LASSO regression model by ten-fold cross-validation method. (C). ROC curve of overdominant model by LASSO regression. (D). The wilcoxon analyis between two survival states in test subsets under overdominant model.
Finally, combined with the age, gender, Ann Arbor stage and whole blood count, the Cox proportional hazards model and LASSO regression model with SNPs were used for multivariate analysis. After age and gender adjusted, it was found that Ann Arbor stage (OR = 1.764, 95% CI = 1.227–2.535) and HGB (OR = 2.613, 95% CI = 1.527–4.473) and LDH (OR = 2.087, 95% CI = 1.194–3.647) have independent effects on OS. And HGB still had a significant effect on survival after stepwise regression (OR = 2.214, 95% CI = 1.216–4.030, p = 0.009). By analyzing the two SNPs associated with HGB, we found that the CA/AA genotype of rs3824662 had poorer OS than CC genotype, and the CA genotype had shorter OS than CC/AA or CC genotype. And the CG genotype in rs907187 had shorter OS than CC/GG or CC genotype, which corresponded to the high risk of anemia of the CG genotype, indicating that the G genotype was a poor prognostic factor.
Association of the genetic polymorphisms andThe results of statistical analysis in our research found that rs2069718 of IFNG and rs6887695 of IL12A were significantly associated with OS in lymphoma patients. Therefore, in order to further determine the role of the two genetic polymorphisms on the progression and prognosis of lymphoma, we explored the association of the two genetic polymorphisms and the mRNA expression levels of IFNG and IL12A in some available lymphoma samples.
For IFNG, we analyzed the expressions in 26 lymphoma patients, which included the rs2069718 polymorphisms of 13 AA, 12 AG, and 1 GG. Under the codominant model, patients with AG genotype had a higher mRNA expression of IFNG than AA genotype (p = 0.0383, Figure 3A). And in the dominant model, the expression level of IFNG in lymphoma patients with AG/GG genotype was also higher than those with AA genotype (p = 0.0284, Figure 3B). These results were consistent with the results of rs2069718 about OS. Among the 40 patients participated in the analysis of mRNA expression of IL12A, 15 patients with the GG genotype, 15 with GC genotype and 10 with CC genotype. Under the codominant and dominant model, GC genotype and GC/CC genotype had much high mRNA expression level of IL12A than GG genotype (all p < 0.05) (Figure 3C,D). However, there was no significant difference of IL12A mRNA expression between GG/CC and GC genotypes under overdominant model (Figure 3E).
![View Image - FIGURE 3. The mRNA expression of IFNG and IL12A. (A). Expression of IFNG mRNA in lymphoma patients with the AA and AG genotypes under codominant model. (B). Expression of IFNG mRNA in lymphoma patients with the AA and AG/GG genotypes under dominant model. (C). Expression of IL12A mRNA in lymphoma patients with the GG and GC genotypes under codominant model. (D). Expression of IL12A mRNA in lymphoma patients with the GG and GC/CC genotypes under dominant model. (E). Expression of IL12A mRNA in lymphoma patients with the GG/CC and GC genotypes under overdominant model. *, p [less than] 0.05.](https://media.proquest.com/media/hms/THUM/6/psv1U?_a=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&_s=ZAfH%2FDIMfKSWZoh%2B99ewN2g79x0%3D "View Image - FIGURE 3. The mRNA expression of IFNG and IL12A. (A). Expression of IFNG mRNA in lymphoma patients with the AA and AG genotypes under codominant model. (B). Expression of IFNG mRNA in lymphoma patients with the AA and AG/GG genotypes under dominant model. (C). Expression of IL12A mRNA in lymphoma patients with the GG and GC genotypes under codominant model. (D). Expression of IL12A mRNA in lymphoma patients with the GG and GC/CC genotypes under dominant model. (E). Expression of IL12A mRNA in lymphoma patients with the GG/CC and GC genotypes under overdominant model. *, p [less than] 0.05.")
Enlarge this image.FIGURE 3. The mRNA expression of IFNG and IL12A. (A). Expression of IFNG mRNA in lymphoma patients with the AA and AG genotypes under codominant model. (B). Expression of IFNG mRNA in lymphoma patients with the AA and AG/GG genotypes under dominant model. (C). Expression of IL12A mRNA in lymphoma patients with the GG and GC genotypes under codominant model. (D). Expression of IL12A mRNA in lymphoma patients with the GG and GC/CC genotypes under dominant model. (E). Expression of IL12A mRNA in lymphoma patients with the GG/CC and GC genotypes under overdominant model. *, p [less than] 0.05.
In this study, we performed genetic polymorphisms analysis of 24 immune-related genes based on 245 lymphoma patients and 213 healthy controls to identify immune-related SNPs that may be related to the susceptibility, progression, and prognosis of lymphoma. After strictly statistical analysis, we found that 15 immune-related SNPs were associated with lymphoma susceptibility and clinical characteristics, including Ann Arbor stage, peripheral blood counts, LDH and OS. Especially, rs6887695 was closely related to the survival of lymphoma patients after bonferroni multiple correction.
Previous studies have demonstrated that most genetic polymorphisms were significantly associated with lymphoma susceptibility. For example, IL-10 polymorphisms were reported to be associated with the risk of lymphoma,10 and rs1800871-CC may play a pathogenic role in the occurrence of non-Hodgkin's lymphoma.31 Besides, the TG/GG of rs2069762 in IL-2 increased the risk of non-Hodgkin's lymphoma by more than three times, and was more likely to develop adverse outcomes.20 In our results, we found that a total of eight SNPs contributed to lymphoma susceptibility under at least one genetic model, which were significantly different between lymphoma patients and healthy controls. These SNPs include cytokines (IFNL3, TNFA, IL4), genes of JAK–STAT and NF-κB pathways (STAT3, STAT5B, NFKBIA), and PARP1 which plays an important role in DNA repair and apoptosis. A study on myeloproliferative neoplasms (MPN) revealed that the CC genotype of IFNL3 rs12979860 was significantly associated with hematologic response to interferon-α treatment, with a significantly higher complete response rate compared with non-CC.68 Besides, Annibali et al. found that the T allele of rs12979860 in IFNL3 significantly increased the risk of cytomegalovirus infection in patients with hematological malignancies undergoing autologous stem cell transplantation.19 And we also found that the T allele in rs12979860 increased the risk of lymphoma susceptibility compared with the C allele, indicating the adverse effect of T in lymphoma.
In addition, in the JAK–STAT pathway, we found the GG genotype of STAT3 rs744166 compared with AA and the T allele of STAT5B rs6503691 compared with C were significantly associated with lymphoma susceptibility. Consistent with our results, other study reported that the GG genotype of rs744166 correlated with immunotherapy failure in aplastic anemia patients46 and rs6503691 TT/CT genotypes were closely associated with recurrent genetic abnormalities in acute myeloid leukemia (AML) patients.36 For genes in NF-κB pathway, we found that lymphoma susceptibility of the A allele was significantly decreased in rs2233406 of NFKBIA. And Zhang et al. found that another genetic polymorphism in NFKBIA, rs696, also exhibited significantly decreased susceptibility in Hodgkin lymphoma,50 suggesting the important role of NFKBIA genetic polymorphism in lymphoma.
The immune-related genetic polymorphisms have been reported to be related to many clinical characteristics that are involved in poor prognosis, such as lower HGB levels, higher WBC, lower PLT counts, higher LDH levels, higher bone marrow involvement, and advanced diseases.20 We integrated and investigated the clinical characteristics of lymphoma patients, including peripheral blood counts, LDH counts, Ann Arbor stages, and OS. Rigorous statistical analysis found that a total of nine SNPs was significantly associated with clinical indices. The rs1800871 of IL10 showed significant difference in LDH level, and the CC genotype suggested a higher risk of abnormally increased LDH than the TT genotype. Consistently, Liu et al. also found that patients with TT genotype in rs1800871 had higher complete remission (CR) or overall response rate, and longer progression-free survival (PFS).32 Kim et al. also showed that rs1800871 had a significant correlation with the prognosis of NHL.30 Moreover, rs2228145, which is closely associated with Ann Arbor stage, has been reported to affect the balance of the IL-6 activities between pro-inflammatory and anti-inflammatory. In patients homozygous for the major alleles of rs2228145, serum CRP levels increased before transplantation, which suggested the major allele of rs2228145 was an adverse factor.27 In our study, patients with the major allele AA had much lower OS than CC genotype, while in the analysis of Ann Arbor stage, AA genotype has lower stage than CA/CC genotype, which needs to be studied in the future.
For rs3824662 of GATA3 associated with HGB, we found that the CA genotype was more likely to develop a lower HGB than CC genotype. Consistently, it was reported that the presence of A allele of rs3824662 correlated with an increased risk of relapse and adverse outcomes in acute lymphoblastic leukemia patients.49 The mortality rate in AA genotype carriers was significantly higher than that in CA + CC carriers.48 That may be due to the rs3824662 located in the GATA3 enhancer region, which affected the gene expression and functions of GATA3. Similarly, we found that patients who carried the CG genotype contributed to a lower HGB than CC genotype in the PARP1 rs907187 which was located in the promoter region. A recent study showed that the SNPs of PARP1 significantly altered the response of cancer cells to treatment, due to affecting the expression of PARP1.55 However, Avitabile M showed that the G allele of rs907187 was associated with a better response to induction therapy in neuroblastoma patients, supporting the potential role of PARP1 as a candidate gene for neuroblastoma treatment failure.56 The opposite result may be due to the heterogeneity of different tumors.
IFN is considered to have antiviral and immunomodulatory effects in vivo, which triggers JAK–STAT pathway, and ultimately up-regulates the expression of interferon stimulating genes.69,70 Our results showed that IFNG rs2069718 was associated with the survival of lymphoma patients under dominant model, and the G allele of rs2069718 had a lower OS. Similarly, other study reported that G allele of rs2069718 contributed to the susceptibility and severity of mixed connective tissue disease (MCTD).71 IL-12 plays a central role in the cellular and humoral pathways, which is mainly produced by antigen-presenting cells and acts on T cells and natural killer (NK) by inducing IFN-γ production.72,73 Recently, the IPI-based immune prognostic model based on eight genes including IL12A had independent prognostic significance in DLBCL patients, which provided an immunological perspective for elucidating the tumor progression mechanism of DLBCL, and were also useful for DLBCL immunotherapy.74 We found that the GC genotype had the lowest OS, and the GG and CC genotypes had similar OS levels in lymphoma patients.
In summary, we found some novel molecular markers related to lymphoma susceptibility and disease progression. Especially, IFNG rs2069718 and IL-12A rs6887695 were significantly associated with OS, and the mRNA expressions of the two genes were also associated with OS of lymphoma patients. We believe that it can be used as a preliminary screen for gene polymorphisms studies and has important exploration value for the prognosis of lymphoma. However, there are also some limitations in our research. First, our patients were recruited all from one hospital which may be regionally biased due to the specific populations, so we may need a multicenter meta-analysis in the future. Second, our SNPs analysis did not cover all genes related to the immune regulatory pathways, which may result in insufficient or uncomprehensive results. Third, our findings are based on the normal peripheral blood and LDH counts of all healthy controls, which may be affected by the lack of the clinical parameters of healthy controls. Furthermore, due to the limited scale and the heterogeneity of the lymphoma patients, our results may need to be further confirmed in the future.
AUTHOR CONTRIBUTIONSChaoyang Gu: Formal analysis (equal); writing – original draft (equal). Can Can: Funding acquisition (equal). Jinting Liu: Conceptualization (equal). Yihong Wei: Data curation (equal). Xinyu Yang: Investigation (equal). Xiaodong Guo: Conceptualization (equal); data curation (equal). Ruiqing Wang: Methodology (equal). Wenbo Jia: Data curation (equal). Wancheng Liu: Data curation (equal). Daoxin Ma: Resources (equal); writing – review and editing (equal).
ACKNOWLEDGMENTSThis work was supported by grants from the National Natural Science Foundation of China (No. 32241005, 81873439).
CONFLICT OF INTEREST STATEMENTThe authors declare no conflict of interest.
DATA AVAILABILITY STATEMENTThe data that support the findings of this study are available from the corresponding author upon reasonable request.
ETHICS STATEMENTThe studies involving humans were conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee on Scientific Research of Shandong University Qilu hospital (KYLL-202204-059).
© 2023. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
1 Department of Hematology, Qilu Hospital of Shandong University, Jinan, China