Identifying CD1c as a potential biomarker by the comprehensive exploration of tumor mutational burden and immune infiltration in diffuse large B cell lymphoma

Introduction

Diffuse large B-cell lymphoma (DLBCL) is the most common type of aggressive non-Hodgkin lymphoma with a high morbidity and mortality rate (Liu & Barta, 2019). The standard treatment approach for patients with DLBCL is immunochemotherapy utilizing combinations of rituximab and anthracycline-based chemotherapy (Spinner & Advani, 2022). However, 30–40% of patients experience treatment failures or relapses, necessitating alternative treatment options (Poletto et al., 2022). In the past decade, targeted therapy and immunotherapy have emerged as promising alternatives for the management of DLBCL (Cheson, Nowakowski & Salles, 2021). Recent advances in understanding the molecular and immunological heterogeneity of DLBCL have led to the identification of novel therapeutic targets. Studies have reported frequent genetic alterations in various genes in DLBCL, including BCL2, BCL6, and MYC, which could be helpful targets for precision medicine approaches (Sermer et al., 2020).

Additionally, recognizing immune evasion mechanisms employed by tumor cells and developing immunotherapy agents offer new opportunities for durable and targeted treatment options for patients with DLBCL (Ye et al., 2022). Despite these advancements, DLBCL remains a challenging disease with a wide range of clinical subtypes, molecular alterations, and varied response rates to different treatments (Liu & Barta, 2019). Further research is needed to develop novel therapeutic approaches and better understand the complex interactions between the tumor microenvironment and DLBCL.

Recent advances in sequencing technologies have enabled the identification of genetic changes and mutations that play a crucial role in the pathogenesis and heterogeneity of DLBCL. Tumor mutational burden (TMB) is defined as the total number of mutations per megabase (Mb) of the genome. It has been shown to correlate with the efficacy of immunotherapy in various tumor types (Jardim et al., 2021). High TMB is associated with increased neoantigen production, which can promote a more robust immune response against the tumor. Therefore, TMB has garnered significant attention as a predictor of response rate to immune checkpoint inhibitors (ICIs) (Killock, 2020). However, the significance of TMB in DLBCL remains controversial. Recent studies have reported higher TMB in the ABC-DLBCL subtype than in the GCB-DLBCL subtype, indicating the potential of TMB as a prognostic marker and therapeutic target in ABC-DLBCL (Schmitz et al., 2018). A study tested 101 DLBCL patients and found that those with a higher TMB had poor OS and progression-free survival. Furthermore, specific genetic mutations such as MYD88, CD79B, and PIM1 were also associated with higher TMB levels, which suggests that TMB is a potential predictor of survival and prognosis in DLBCL patients (Chen et al., 2021).

Recent studies have shown the tumor microenvironment (TME) plays a crucial role in the pathogenesis and progression of DLBCL, and molecular targets related to the tumor microenvironment invasion may become the key to immunotherapy (Cioroianu et al., 2019). Based on advances in gene sequencing and expression profiling, studies have shown that the prognosis of DLBCL patients is related to TME (Ciavarella et al., 2019). For example, immune checkpoint inhibitors (ICI) have demonstrated significant efficacy in some refractory hematological malignancies (Thanarajasingam, Thanarajasingam & Ansell, 2016). These findings highlight the importance of studying the TME in DLBCL and suggest potential novel therapeutic strategies targeting the TME. Further research is needed to elucidate the complex interactions between diverse TME components and to identify biomarkers that can predict patient outcomes and guide personalized therapy. Tumor cells may evade immune recognition by the immune system by altering antigen presentation or modulating the TME, resulting in immune escape (Riaz et al., 2016). Therefore, exploring the relationship between TMB and immune regulation and identifying relevant genes or biological mechanisms will lead to a better understanding of the role of immunotherapy in cancer treatment.

In this study, we collected somatic mutation data, transcriptome data, and clinical information from the TCGA database, aiming to study the association between TMB and gene mutation, immune response, and prognosis of DLBCL combined with immune infiltration. We attempted to elucidate the relationships between TMB groups and clinicopathological factors, between TMB groups and different immune-infiltrating cells, and between TMB groups and prognosis. The results of these studies may provide novel biomarkers and potential treatment options for DLBCL.

Materials and Methods

Results

The landscape of mutation profiles in DLBCL

The research strategy is presented in (Fig. 1). Somatic mutation profiles of 37 DLBCL samples were downloaded from the TCGA database. We used the “maftools” R package to visualize mutation data in VAF format. In general, missense mutation accounted for the most significant proportion of mutation types (Fig. 2A), and the occurrence frequency of single nucleotide polymorphism (SNP) was higher than insertion and deletion (Fig. 2B). The most common type of base substitution was C>T (Fig. 2C). The boxplot (Figs. 2D and 2E) showed different mutation types in DLBCL patients, and Fig. 2F showed the top 10 genes with mutation frequency, including PIM1 (22%), IGLV3-1 (38%), IGLL5 (27%), IGHG1 (22%), IGHV2-70 (27%), BTG (27%), IGHM (24%), KMT2D (32%), IGLC2 (24%), CARD11 (22%). The mutation landscape displayed the mutation information of each sample, in which the mutation frequency of IGLV3-1 and KMT2D accounted for 38% and 32%, respectively (Fig. 2G). Heatmap of gene correlations shows gene-to-gene relationships. For example, there is a synergistic effect between MUC16 and FAT4, while SOCS1 and KMT2D are mutually exclusive (Fig. 2H). Meanwhile, the Genecloud plot displayed the frequency of mutations in genes (Fig. S1), and the higher the mutation frequency, the larger the gene name.

TMB correlated with survival outcomes and clinical pathological characteristics

We calculated the mutation event per million bases as the TMB for DLBCL patients, worked out the optimal cutpoint using the surv_cutpoint function in the “survival” R package, and set the parameter cutpoint = 2.8 to divide patients into low- (18 cases) and high-(19 cases) TMB groups. TMB ranged from 0.14 to 6.92 with a median of 1.9 per MB (Fig. 3A). Kaplan-Meier survival analysis was carried out, and the result showed that the 5-year survival rate of the high-TMB group was lower than the low-TMB group (Fig. 3B). In addition, none of the clinical traits was significantly correlated with TMB level, which may be due to the small samples (Table S1).

Identifying differentially expressed genes based on TMB grouping and functional enrichment analysis of GO, KEGG, and GSEA

Differentially expressed genes (DEGs) were calculated by R software version 4.2.2 (R Core Team, 2023). A total of 62 DEGs were identified in the low- and high-TMB group using the “limma” R package by setting the P < 0.05 and | log₂ (FC) | > 1 (Table 1). The heatmap visualized DEGs between the low- and high-TMB groups (Fig. 4A). The volcano map showed 42 up-regulated genes and 20 down-regulated genes (Fig. 4B). Subsequently, we conducted GO enrichment analysis on DEGs and found that the differential genes were mainly involved in immune-related pathways, such as lymphocyte-mediated immunity, adaptive immune response based on somatic recombination of immune receptors built from immunoglobulin superfamily domains, immunoglobulin mediated immune response and complement activation, classical pathway (Fig. 4C and Table S2). The enrichment information of the GO pathways is illustrated in Fig. 4D. CD1c, CCL21, TP63, ORM1, ACTG2, IGHG3, IGHM, TRPM4, and so on are involved in all of the top GO pathways (including Molecular Function, Cellular Component, Biological Process) and were identified as hub genes. KEGG pathways illustrated that the differential genes were mainly enriched in vascular smooth muscle contraction, hematopoietic cell lineage, carbon metabolism, and neuroactive 210 ligand−receptor interaction pathway (Fig. S2 and Table S3). In addition, we further selected the GSEA results of the top TMB-related items in Figs. 4E–4G, including one carbon pool by folate, rig-i like receptor signaling pathway-creative diagnostics, and the tight junction, which were associated with the TMB level (P < 0.05).

Table 1:

Differentially expressed genes between low-TMB and high-TMB groups.

Gene symbol	logFC	AveExpr	t	P.Value	adj.P.val	B
PRAME	2.922405	1.672344	5.237783	8.14E−06	0.184374	2.785421
SLC12A3	1.100708	0.871668	5.168989	1.00E−05	0.184374	2.580099
EIF5AP2	1.037704	0.7309014	4.540696	6.55E−05	0.446506	0.725776
TP63	1.498609	1.72559	4.29477	0.000135	0.446506	0.015043
C17orf99	1.701547	1.3456096	4.276447	0.000143	0.446506	−0.03746
CNTNAP4	1.047761	0.3234867	4.189251	0.000184	0.446506	−0.28634
LINC02562	1.440104	0.682375	3.902082	0.000421	0.597317	−1.09357
LY86	−1.36579	4.6202479	−3.65037	0.000859	0.628234	−1.78272
GTSF1	2.491803	3.3402724	3.647969	0.000865	0.628234	−1.78919
TRPM4	2.112655	2.5636372	3.589926	0.001018	0.685802	−1.94525
AL627309.6	1.130337	1.5450149	3.531998	0.001195	0.691416	−2.09986
MIR5195	−1.90974	4.7542273	−3.45277	0.001487	0.77525	−2.30939
TCTN1	1.056882	1.3681228	3.418234	0.001635	0.799445	−2.40003
ASB2	−1.38053	3.2815049	−3.30631	0.002216	0.84459	−2.69058
SLC9A7	1.001124	3.0139012	3.296243	0.002277	0.84459	−2.71648
AL627309.7	1.147779	1.3959171	3.215333	0.002829	0.902441	−2.92307
ORM1	2.689855	2.1889532	3.185368	0.003065	0.934341	−2.99888
IL4I1	1.275177	5.8378104	3.159182	0.003286	0.945784	−3.06482
PES1P2	1.059463	0.3808462	3.144121	0.003419	0.955589	−3.10261
ACTG2	2.614232	2.8996668	3.131267	0.003538	0.96048	−3.13478
PTGIR	1.373384	2.2671073	3.119858	0.003646	0.96048	−3.16327
IGKV1D-8	−1.26761	1.4899645	−3.11549	0.003688	0.96048	−3.17417
AP000593.3	1.96134	1.6328582	3.061174	0.004254	0.979536	−3.30891
SSTR2	1.027005	1.0412898	3.029081	0.004626	0.999998	−3.38788
AF127936.1	1.028164	0.7326819	2.986073	0.005172	0.999998	−3.49296
ATF5	1.412084	6.6579457	2.967457	0.005427	0.999998	−3.53817
IGHV5-78	−1.64874	4.4896813	−2.9616	0.00551	0.999998	−3.55236
ARLNC1	1.842147	2.1616004	2.883652	0.006729	0.999998	−3.7396
OTOF	1.054667	0.5352778	2.858965	0.007165	0.999998	−3.79827
AL137026.2	1.234373	0.6788708	2.818765	0.007932	0.999998	−3.89314
CTSLP2	1.010085	0.3901514	2.797078	0.008378	0.999998	−3.94397
NFIL3	1.002554	3.2248748	2.796697	0.008386	0.999998	−3.94486
AC005083.1	−1.00218	1.4038589	−2.70415	0.010562	0.999998	−4.15895
MIR4538	−2.1038	2.6705054	−2.65241	0.011997	0.999998	−4.2766
PSMA8	1.058776	0.8946082	2.611077	0.01327	0.999998	−4.3695
ORM2	1.39914	1.4257549	2.597839	0.013704	0.999998	−4.39905
DNAJC5B	1.282866	2.7580284	2.563015	0.014907	0.999998	−4.47631
AC024475.4	1.115043	0.6922118	2.532595	0.016037	0.999998	−4.54321
GPR160	−1.02655	2.3532824	−2.52493	0.016334	0.999998	−4.55998
GLDC	−1.14561	1.4819458	−2.50791	0.017011	0.999998	−4.5971
KRT17	1.215934	0.8071803	2.505569	0.017106	0.999998	−4.60218
FAM129A	1.01797	2.6609447	2.484965	0.017965	0.999998	−4.64685
LRRC32	−1.22325	3.030123	−2.48445	0.017987	0.999998	−4.64797
IGHG3	−2.28464	5.678394	−2.48286	0.018055	0.999998	−4.65141
ETV7	1.003919	2.1068744	2.45744	0.019174	0.999998	−4.70613
CR2	−1.8187	3.1337495	−2.44458	0.019763	0.999998	−4.73365
CD1C	−1.56179	2.9112972	−2.42331	0.020775	0.999998	−4.77898
TREML2	1.251118	2.1785378	2.409244	0.02147	0.999998	−4.80879
TREM1	1.013126	1.0170469	2.369286	0.02356	0.999998	−4.89281
USP2	1.059806	0.9148026	2.362492	0.023933	0.999998	−4.907
RPL10P6	1.261382	2.1649398	2.339933	0.025211	0.999998	−4.95389
RPL10P9	1.236246	3.0773338	2.337468	0.025354	0.999998	−4.95899
PPP1R14A	1.094098	1.1728365	2.324857	0.026099	0.999998	−4.98504
CCL21	−2.82779	5.2870774	−2.27481	0.029253	0.999998	−5.08742
FBP2	1.072932	0.6428013	2.256376	0.030498	0.999998	−5.1247
TRBJ2-2P	−1.42691	3.7776966	−2.24323	0.031415	0.999998	−5.15115
TRBJ2-1	−1.05611	2.6341454	−2.16798	0.037154	0.999998	−5.30037
IGKV1-5	−1.88826	3.856146	−2.16686	0.037245	0.999998	−5.30255
CPA6	1.006875	0.7805608	2.165899	0.037325	0.999998	−5.30444
IGHV1-69	−1.05659	1.4366253	−2.14157	0.039378	0.999998	−5.35181
IGHV3-23	−1.56581	3.7508495	−2.08555	0.044491	0.999998	−5.45934
IGHM	−2.13098	9.540321	−2.04243	0.048817	0.999998	−5.54061

DOI: 10.7717/peerj.16618/table-1

Differential abundance of immune cells in the low- and high-TMB groups using CIBERSORT

After DEG screening, to further compare the difference in the degree of immune cell infiltration between low- and high-TMB groups, we calculated the composition ratio of immune cells per sample by the “CIBERSORT” R package. The boxplot in Fig. 5A showed a specific portion of 22 immune cells in each DLBCL sample. We also calculated the proportion of immune cells in the whole DLBCL cohort, accounting for the most, including B cells naive, CD8+ T cells, M2 macrophages, and M0 macrophages (Fig. 5B). The heatmap showed the distribution of immune cells between low- and high-TMB groups, and the result displayed that the high-TMB group had a lower immune score (Fig. 5C). In addition, the Wilcoxon rank-sum test demonstrated that monocytes, dendritic cells activated, and dendritic cells resting were more deficient in the high-TMB group (P < 0.05) (Fig. 5D). According to the above analysis results, the high-TMB group inhibited the level of immune cells infiltration in DLBCL samples.

Screening TMB-related immune genes and verifying the prognosis of the screened genes using the GEO database

The immune-related genes were downloaded from the ImmPort database and intersected with the selected DEGs. A total of 13 TMB-related immune genes were obtained, including CD1c, ORM1, ORM2, CCL21, CR2, IGHG3, IGHM, IGHV1-69, IGHV3-23, IGKV1-5, IGKV1D-8, SSTR2, TRBJ2-1 (Fig. 6A). Subsequently, univariate regression analysis was performed on the above genes, and it was found that there was no significant correlation between these genes and prognosis (P > 0.05, Fig. S3), possibly due to the small samples in the TCGA-DLBC cohort. Therefore, we expanded the sample size and screened a total of 1,133 samples of DLBCL gene expression microarray datasets (GSE31312, GSE10846, GSE32918) from the GEO database, as well as the clinical information of the corresponding samples. After ID translation, data homogenization and standardization, and removal of batch effect, five genes including CD1c, CCL21, ORM1, CR2, and SSTR2 (the rest of the eight genes were not included in the expression profile data) were obtained after joint analysis of the three datasets. Kaplan-Meier survival analysis was performed, and the results showed that the level of CD1c expression was significantly correlated with prognosis (P < 0.05, Figs. 6B–6F). Kaplan Meier survival curves were plotted in six GEO datasets, including GSE31312, GSE10846, GSE32918, GSE53786, GSE87371, and GSE181063, respectively, and the results showed that the prognosis was poor in the high CD1c expression groups (Fig. S4).

CNV of CD1c, immune cells, and survival in DLBCL using TIMER database

In general, CNV refers to an increase or decrease in the copy number of large segments of the genome that are more than 1 kb in length. CNV was observed in prognostic TMB-related immune genes, and to verify the CNV of CD1c and the relationship between immune cell content and prognosis, we utilized the TIMER database to obtain CD1c expression between normal and tumor tissues in various cancers, and it shows that the expression level of CD1c is the highest in Thymoma (THYM), followed by that of DLBCL (Fig. 7A). Especially in DLBCL, CD1c expression was positively correlated with B cells, neutrophils, dendritic cells and negatively associated with CD8+ T cells, CD4+ T cells, and macrophages, among which, the correlation with B cells was the highest (cor = 0.693, P = 1.44E−03, Fig. 7B). In addition, high amplification of CD1c was significantly different compared to other CNVs (P < 0.01, Fig. 7C). As for the relationship between immune cells content and prognosis, high levels of CD8+ T cells and dendritic cells indicate a superiors survival result. In contrast, low expression of CD1c may promote better survival (Fig. 7D).

Relationship between CD1c and drug sensitivity and regulation of CD1c transcription factors

According to the correlation analysis of target genes and drug sensitivity in the CellMiner database, it was found that there was a significant correlation between the expression level of CD1c and clinical drug sensitivity, mainly with nelarabine, methylprednisolone, chelerythrine, ribavirin, fluphenazine was positive (Figs. 8A and 8B). Therefore, the higher the expression of CD1c, the more sensitive cells are to these drugs. The KonckTF database results showed that the regulation of CD1c may be related to the effects of transcription factors such as CREB1, AHR, and TOX, resulting in the corresponding biological effects (Table 2).

Table 2:

Transcription factors regulating CD1c in the Haematopoietic and lymphoid tissue by knockTF database.

Target gene	TF	Knock-method	Tissue type	Biosample name	Fold change	Log2FC
CD1c	CREB1	shRNA	Haematopoietic and lymphoid tissue	K562	0.43886	−1.18818
CD1c	AHR	siRNA	Haematopoietic and lymphoid tissue	THP-1	0.36694	−1.4464
CD1c	TOX	shRNA	Haematopoietic and lymphoid tissue	CCRF-CEM	0.32445	−1.62392

DOI: 10.7717/peerj.16618/table-2

Verified the expression of CD1c at mRNA and protein levels

CD1c mRNA levels were compared between normal tissues downloaded by GETx and tumor tissues of TCGA-DLBC. We found a significant difference in the expression level of CD1c between normal and tumor tissue (P < 0.05, Fig. S5). We performed RT-qPCR to detect the expression of CD1c mRNA in three cell lines, and the results showed that SU-DHL-4 (P = 0.0017) had the highest CD1c expression level, followed by SU-DHL-2 (P = 0.0008) and HMy2.CIR had the lowest expression level (Fig. 9A). We collected 76 DLBCL pathological sections and 34 sections of reactive lymphoid hyperplasia (Table 3). CD1c was significantly overexpressed in DLBCL tissues compared with normal lymphoid tissue hyperplasia (P < 0.0001, Fig. 9B). Western blot results demonstrated that SU-DHL-4 cell lines expressed high levels of CD1c protein. In contrast, the HMy.C1R cell line showed low expression of CD1c protein. It should be noted that SU-DHL-2 may be due to the cell line activity being too low, with no reference gene β-Tubulin expression (Fig. 9C). The immunohistochemical results showed that in 50 DLBCL tissue samples, 20 samples had high CD1c expression, and 30 samples had low CD1c expression. The positive rate of CD1c in DLBCL reached 90%. Among 19 cases of reactive lymphoid hyperplasia, 12 had low CD1c expression, and seven had high CD1c expression. The positive rate in the control group was 68.42% (Table S4). The protein expression of CD1c in DLBCL tissues was significantly higher than in reactive lymphoid hyperplasia tissues (Figs. 9D and 9E). Wet experiments further confirmed the reliability of the bioinformatics findings. To investigate the relationship between long-term prognosis and CD1c, we obtained OS data according to CD1c expression in DLBCL patients. The results showed that the prognosis of the CD1c high-expression group was poor. However, the difference was not statistically significant for the high and low CD1c protein expression groups, possibly due to the small sample size (Figs. 9F and 9G).

Table 3:

76 DLBCL samples clinical information.

Sample	SampleID	PaththID	Gender	Age	Sites	Hans	Status	OStime
Sample01	T1	L2000519	Male	69	Nodal	ABC	0	1,393
Sample02	T2	L2000214	Female	64	Nodal	ABC	1	38
Sample03	T3	L2100150	Male	47	Nodal	ABC	0	709
Sample04	T4	L2100157	Female	56	Nodal	ABC	0	695
Sample05	T5	L2100202	Female	80	Nodal	NA	0	653
Sample06	T6	L2100204	Female	65	Nodal	GCB	0	674
Sample07	T7	L2100213	Male	59	Nodal	ABC	0	658
Sample08	T8	L2100246	Female	35	Nodal	ABC	0	624
Sample09	T9	L2100254	Female	66	Nodal	ABC	0	660
Sample10	T10	L2100271	Female	63	Nodal	ABC	0	1,089
Sample11	T11	L2100287	Female	43	Nodal	ABC	0	581
Sample12	T12	L2100408	Male	49	Nodal	ABC	1	544
Sample13	T13	L2100295	Female	56	Nodal	ABC	0	620
Sample14	T14	L2100417	Male	71	Nodal	ABC	1	4
Sample15	T15	L2100415	Male	49	Nodal	ABC	0	513
Sample16	T16	L2100454	Male	61	Nodal	ABC	0	498
Sample17	T17	Z2139511	Female	59	Nodal	ABC	0	631
Sample18	T18	Z2152204	Female	94	Nodal	ABC	1	98
Sample19	T19	Z2152985	Male	48	Nodal	ABC	0	507
Sample20	T20	Z2157414	Female	77	Nodal	ABC	0	504
Sample21	T21	L1900028	Male	30	Nodal	ABC	0	1,514
Sample22	T22	L1900160	Male	21	Nodal	ABC	0	1,431
Sample23	T23	L1900164	Male	59	Nodal	ABC	1	1,153
Sample24	T24	L1900191	Female	66	Nodal	ABC	0	1,428
Sample25	T25	Z1919871	Male	72	Gastrointestinal	ABC	1	1,198
Sample26	T26	Z1927492	Female	72	Gastrointestinal	ABC	0	1,380
Sample27	T27	Z1959932	Male	38	Gastrointestinal	ABC	0	1,232
Sample28	T28	Z1931327	Male	78	Gastrointestinal	ABC	1	421
Sample29	T29	Z1924788	Male	44	Gastrointestinal	ABC	0	1,366
Sample30	T30	Z1918661	Male	64	Nodal	GCB	0	1,400
Sample31	T31	Z1905302	Male	84	Nodal	ABC	1	376
Sample32	T32	Z1911133	Female	66	Gastrointestinal	ABC	1	1,258
Sample33	T33	Z1941053	Female	34	Nodal	ABC	1	42
Sample34	T34	Z1934519	Male	65	Gastrointestinal	GCB	0	1,348
Sample35	T35	Z1933930	Male	55	Gastrointestinal	GCB	1	26
Sample36	T36	Z1900756	Male	57	Gastrointestinal	ABC	1	1,409
Sample37	T37	Z2008503	Male	79	Gastrointestinal	ABC	0	1,078
Sample38	T38	Z2013238	Male	49	Gastrointestinal	GCB	0	1,049
Sample39	T39	Z2016528	Male	70	Nodal	GCB	1	1,209
Sample40	T40	Z2017865	Female	77	Gastrointestinal	ABC	0	1,025
Sample41	T41	Z2017001	Male	51	Gastrointestinal	ABC	1	487
Sample42	T42	Z2043475	Female	53	Gastrointestinal	ABC	1	290
Sample43	T43	Z2121652	Male	31	Gastrointestinal	GCB	0	651
Sample44	T44	Z2114970	Male	63	Nodal	NA	1	1
Sample45	T45	Z2122930	Female	57	Gastrointestinal	ABC	1	476
Sample46	T46	Z2129658	Male	64	Gastrointestinal	ABC	0	576
Sample47	T47	Z2130933	Female	54	Nodal	ABC	0	630
Sample48	T48	Z2131460	Female	66	Gastrointestinal	ABC	0	651
Sample49	T49	Z2131475	Female	52	Gastrointestinal	GCB	0	632
Sample50	T50	Z2131621	Female	55	Nodal	ABC	0	630
Sample51	T51	Z2137167	Male	71	Gastrointestinal	GCB	0	609
Sample52	T52	L1900193	Female	68	Nodal	ABC	0	1,424
Sample53	T53	L1900272	Male	53	Nodal	ABC	1	3,511
Sample54	T54	L2000538	Male	46	Nodal	ABC	0	821
Sample55	T55	L2000580	Female	50	Nodal	ABC	0	795
Sample56	T56	L2000381	Female	76	Nodal	ABC	0	2,044
Sample57	T57	L2000293	Female	75	Nodal	ABC	0	952
Sample58	T58	L2000076	Male	50	Nodal	ABC	0	1,074
Sample59	T59	L2000023	Female	34	Nodal	ABC	0	1,142
Sample60	T60	L1900498	Female	60	Nodal	ABC	0	1,229
Sample61	T61	L1900460	Male	69	Nodal	ABC	0	1,253
Sample62	T62	L1900202	Male	22	Nodal	GCB	0	1,395
Sample63	T63	L1900008	Female	84	Nodal	ABC	1	3,652
Sample64	T64	L1900135	Male	64	Nodal	ABC	1	729
Sample65	T65	L1900138	Female	62	Nodal	ABC	0	1,477
Sample66	T66	Z2108388	Female	53	Nodal	ABC	0	732
Sample67	T67	Z2058451	Male	72	Gastrointestinal	GCB	0	820
Sample68	T68	Z2038723	Male	56	Gastrointestinal	GCB	0	886
Sample69	T69	Z2054305	Female	54	Nodal	GCB	1	226
Sample70	T70	Z2017133	Male	39	Gastrointestinal	ABC	0	1,520
Sample71	T71	Z1953852	Male	30	Gastrointestinal	GCB	1	42
Sample72	T72	Z1900528	Female	70	Gastrointestinal	ABC	0	1,484
Sample73	T73	Z1941882	Female	55	Nodal	GCB	0	1,285
Sample74	T74	Z2165832	Male	77	Gastrointestinal	GCB	0	452
Sample75	T75	Z2161904	Female	35	Gastrointestinal	ABC	0	498
Sample76	T76	Q2046427	Male	71	Nodal	GCB	0	780

DOI: 10.7717/peerj.16618/table-3

Discussion

Based on TCGA DLBCL mutation data, TMB may not independently predict DLBCL prognosis. Previous studies explored the predictive value of a 69-gene panel-TMB, indicating its association with poor prognosis and potential as a predictive indicator when analyzed through a nomogram mode (Chen et al., 2021). Additionally, higher TMB in untreated cancer patients often leads to poor outcomes (Bevins et al., 2020), aligning with our research trend. Although not statistically significant, it is speculated that a small sample size or the need for combined use of TMB with other prognostic factors may yield better predictive effects. Unlike the results of the above two studies, our study screened the genes with differential mRNA expression in the high and low TMB groups through the mutation profile information of DLBC in the TCGA database and screened the predictive genes among them that were highly correlated with prognosis by combining the immune infiltration-related genes. 69-gene panel TMB results (Chen et al., 2021) were obtained by studying the mutation information of the genes, whereas our study mainly screened downstream differentially expressed mRNA genes by high and low mutation genomes, and then got the high and low mutation and prognostic factors affecting DLBCL patients at the transcriptional level.

For the high and low TMB groups, DEGs were primarily enriched in immune cell-mediated immune responses, such as lymphocyte-mediated immunity, immunoglobulin-mediated immune response, B cell-mediated immunity, B cell activation, and immune response activating cell surface receptor signaling pathway. The role of immune cells in DLBCL is complex and multifaceted, involving clinical and histological prognostic factors, immune environment interactions, immune escape strategies, and tumor progression. Previous studies have used computational methods to analyze gene expression profiling (GEP) datasets of 175 DLBCL cases, identifying prognostic genes related to the immune microenvironment, and showed that patients with higher proportions of dendritic cells (DCs) had more prolonged overall survival (OS) (Ciavarella et al., 2019). Our GSEA analysis showed that high TMB groups were mainly enriched in the one-carbon pool by folate (OCPF) and rig I-like receptor (RLR) signaling pathways. Previous study has shown that defects or mutations in genes involved in the OCPF pathway are closely related to DNA methylation and play a crucial role in tumor development (Pan et al., 2021). In recent years, research has explored the correlation between the RLR signaling pathway and tumor mutation burden. Mutations or abnormal expression levels of specific genes in the RLR signaling pathway, such as MAVS, TBK1, and IRF3, can lead to changes in the immune signaling pathway, resulting in a decreased ability of immune cells to recognize and clear malignant tumors (Pecori et al., 2023). The low TMB group is mainly enriched in tight junctions (TJs). Some recent studies have found a specific association between TJs and low tumor mutation burden (Hashimoto & Oshima, 2022).

To elucidate the intrinsic connection between TMB and immune infiltrating cells in the tumor microenvironment, our further studies indicate that the high TMB group has a lower immune score, significant infiltration of M0 and M2 macrophages, activated dendritic cells, and activated immune responses. The above conclusion is consistent with previous literature reporting. Tumor-associated macrophages may provide a growth-promoting microenvironment for malignant B cells and are associated with poor prognosis (McCord et al., 2019). Activated dendritic cells are closely related to higher patient survival rates and prolonged progression-free survival (Ciavarella et al., 2018), which may be because activated dendritic cells can promote T cell activation and enhance tumor-specific cytotoxicity (Garris & Luke, 2020), thereby improving the treatment efficacy for DLBCL patients. However, studies also report a correlation between activated dendritic cells and poorer prognosis (Merdan et al., 2021), which may be because the number of activated dendritic cells is often replaced by other cells in patients with DLBCL, which is also a manifestation of immune system dysregulation in these patients. The low TMB group has an abundance of B cells, follicular helper T cells, regulatory T cells, γδT cells, and various resting immune cells. The study indicated that the impact of inhibiting follicular helper T cells and B cells on tumor immune response is more profound than that of inhibiting CD8+ T cells, indicating that B cells and follicular helper T cells play a critical role in tumor immune response (Hollern et al., 2019). Furthermore, higher levels of γδT cells have been associated with a better prognosis. One study identified two clusters (EC1 and EC2), with EC1 showing an abundance of TP53, MYD88, HIST1H1D, HIST1H1C, KMT2D, and EZH2 mutations, and poor prognosis. EC2 is often accompanied by B2M, CD70, and MEF2B mutations, which are related to DNA damage repair, cytokine-mediated and B cell activation immune signal transduction, elevated CD8+ T cell, γδT cell, and T helper cell levels, increased immune score and immunogenic cell death (ICD) regulatory factors, with good patient prognosis (Wang et al., 2022).

Generally, differences in immunogenicity may lead to differences in immune mechanism activation, and the activation of different types of immune cells in the high TMB group may indicate the inhibition of immune response. It is worth noting that in our study, PIM3 and KMT2D have prominent performances among mutated genes. Previous research has shown that high expression of PIM3 is a poor prognostic factor for DLBCL patients and is associated with common mutated genes in DLBCL, such as MYD88, MYC, and BTK (Wang et al., 2023). KMT2D is a methyltransferase that can reduce the transcriptional activity of specific genes. In a study of DLBCL, somatic mutations in KMT2D were most commonly observed (at 19.5%) and were associated with poor prognosis (Liu et al., 2021).

CD1c has been finally confirmed as a TMB-associated immune gene relevant to the prognosis of DLBCL, and its function has been further explored. CD1c encodes a transmembrane glycoprotein, a member of the CD1 family, and is associated with β2-microglobulin (Moody & Suliman, 2017). Analysis of CD1 expression in B-cell chronic lymphocytic leukemia shows that CD1 mediates immune deficiency, cytokine response polarization, adhesion changes, increased intracellular protein transfer, and leukemia cell processing (Zheng et al., 2002). CD1c+ restricted T cells exhibit potent anti-leukemia activity in mouse models, indicating that this lipid antigen may represent a new target for immunotherapy of hematological malignancies (Lepore et al., 2015). In non-small cell lung cancer (NSCLC), the CD1c+ DCs subset may play an important role in anti-tumor immunity (Lu et al., 2019). A study on early lung adenocarcinoma with EGFR mutation found that the infiltrating T cell types were mainly exhausted and regulatory T cells, which were associated with an increase in dendritic cells expressing the CD1c gene precisely (He et al., 2021). We found one article by searching for CD1b as a potential prognostic biomarker associated with tumor mutation burden and promotes antitumor immunity in lung adenocarcinoma. In this study, the authors showed that CD1b expression is associated with a better prognosis in Lung Adenocarcinoma (LUAD) and promotes anti-tumor immunity by constructing a TMB prognostic model that effectively predicts the prognosis of patients with LUAD, which can be used as a potential prognostic biomarker and immune-related therapeutic target for LUAD (Li et al., 2022). While our study focused on the prognostic value of CD1c in DLBCL, the function of the target gene was verified by various database applications, such as TIMER, CellMiner, konckTF and GTEx. Wet assays confirmed the expression of the target gene at the RNA and protein levels in DLBCL tissue and cell samples. The count of CD1c+ DCs in the blood of gastric cancer patients increases (Liu et al., 2018). In renal cell carcinoma, CD1c+ DCs predict progression-free survival (van Cruijsen et al., 2008). In hepatocellular carcinoma (HCC), the elevation of the ILT4+CD1c+ subset in tumor tissue may play an essential role in immune suppression (Wang et al., 2019). In breast cancer (BRCA), the expression of CD1c correlated with BRCA prognosis and clinical features (Chen et al., 2023). We have identified an increase in CD1c expression in DLBCL tissue, which is associated with poor prognosis according to survival analysis. Although CD1c is a marker of dendritic cell maturation, previous studies have reported CD1c antigen expression in both normal and neoplastic B cells. While CD1c has not been studied as a viable biomarker for DLBCL immunotherapy yet, in our study, we explored the relationship between CD1c expression and prognosis in DLBCL using bioinformatics analysis, revealing that low CD1c expression is indicative of better prognosis. Flow cytometry and immunohistochemistry results demonstrate CD1c antigen expression in a variety of B-cell subgroups, including mature and immature B-cell lines, as well as chronic lymphocytic leukemia (CLL) and hairy cell leukemia (HCL) (Smith, Thomas & Bodmer, 1988). HCL patients with high CD1c expression generally have a poor prognosis, possibly due to HCL cells being over-activated in cases of high CD1c expression, resulting in a more invasive phenotype (Bourguin-Plonquet et al., 2002). Additionally, high CD1c expression in dendritic cells may be closely associated with their immunological function in the immune system and play an essential role in tumor treatment. Therefore, CD1c expression likely has different mechanisms of action regarding prognosis in different cell types, and further research is needed to determine its specific impact.

After conducting further in-depth analysis of CD1c, we found that high amplification of CD1c in B cells and dendritic cells suggests that CD1c mutations inhibit the effective mediation and maintenance of normal immune response by antigen-presenting cells. The poor prognosis of patients with high CD1c expression supports this observation. Targeted gene sensitivity analysis indicates CD1c is associated with clinical sensitivity to various drugs. Molecular research suggests that transcription factors CREB1, AHR, and TOX drive tumor growth and metastasis and are associated with poor prognosis of DLBCL (Huang et al., 2022). The expression of CD1c mRNA and protein levels was verified by wet experiments at cell and tissue levels, indicating that CD1c was more expressed in tumor tissues than in reactive hyperplasia tissues, and the high expression group had a poor prognosis. Studies on the prognostic value of CD1c have shown that high CD1c expression at the mRNA level has a poor prognosis and low CD1c expression has a good prognosis. Prognostic analysis of CD1c in the TIMER database also showed that high mRNA expression within 10 years was associated with poor prognosis. Prognostic analysis of DLBCL clinical samples showed that patients with high CD1c expression at the mRNA level had a poor prognosis, and the same was true for patients with high CD1c expression at the protein level, but the difference was not statistically significant. Based on the experimental results and previous related literature, the possible mechanism and biological function of CD1c in tumor prognosis were explored. Up-regulation of CD1c expression inhibits the effective mediation and maintenance of normal immune responses by antigen-presenting cells, which in turn leads to poor prognosis of patients.

In summary, based on the joint analysis of TMB and immune infiltration, our study has identified immune genes associated with prognosis in DLBCL mutations and explored the inherent correlation between TMB and immune infiltration. CD1c has been identified as a potential biomarker for DLBCL, which may provide new insights for combination therapy.

Conclusions

In conclusion, based on the co-analysis of TMB and immune infiltration, our study identified the immune genes associated with prognosis in DLBCL mutations. It explored the internal correlation between TMB and immune cells infiltrated in the immune microenvironment. CD1c was recognized as a potential marker of DLBCL, which may provide new insights into the immunotherapy of DLBCL. CD1c is also a gene associated with TMB of DLBCL, which predicts poor survival. Bioinformatic analysis shows CD1c is involved in tumor-related signaling pathways and immune and metabolic processes. Thus, the study offers a novel target to investigate the underlying mechanism for diffuse large B cell lymphoma.

Supplemental Information