MAFB is a biomarker for cancer severity, prognostic marker, and a potential target for cancer immunotherapy

MAFB is a transcription factor specically expressed in macrophages. Using in vitro and in vivo in mouse tumor models, our previous study suggested that MAFB could be a suitable marker for tumor-associated macrophages (TAMs), besides MAFB is expressed in anti-inammatory alternatively activated M2 macrophages in vitro. TAMs play a key role in the tumor microenvironment (TME) by inducing immunosuppression, angiogenesis, tumor invasion, and metastasis. However, nding a suitable specic biomarker and target for TAMs is challenging. Here, we demonstrated that MAFB could be used as a biomarker for TAMs and consequently, for severity in various human cancers, including lung, liver, colon, and pancreatic cancers, according to the immunohistochemical analysis of the expression of MAFB, CD68, and CD204. Moreover, In a cohort of lung adenocarcinomas patients (n = 120), increased MAFB expression was related to increased tendency towards metastasis and poor overall survival rate. Further, we showed that MAFB expression was positively correlated to the expression of CD204 and CD68 in both human hepatocarcinoma and colon cancers. Our ndings indicate that MAFB as a specic biomarker can be used as prognostic marker for Metastasis potential in Lung adenocarcinomas patients and also a biomarker for the severe Liver, Colon and pancreatic cancers. In addition, we showed that MAFB was expressed in Tumor associated macrophages expressing Programmed cell death protein-1 and/or Programmed cell death ligand 1 (TAM PD-1+ and TAM PD-L1+) cells in both human lung adenocarcinomas and Lewis lung carcinoma (LLC) mouse model. These ndings indicate that MAFB can be a potential target for drug development against TAM PD-1+ and TAM PD-L1+ cells. In summary, transcriptional factor MAFB can be used as a specic biomarker, prognostic marker and a potential target for cancer immunotherapy against TAMs.


Introduction
V-maf musculoaponeurotic brosarcoma oncogene homolog B (MAFB) is a member of the large Maf transcription factor family 2 which is expressed in several tissues and has been associated with the differentiation of various cell types, including macrophages (Mφs), 3 tumor-associated macrophages (TAMs), 1 kidney podocytes, 4 keratinocytes, 5,6 and pancreatic α-cells and β-cells. 7,8 In the hematopoietic cell lineage, MafB induces monocytes to differentiate into macrophages rather than dendritic cells, suggesting that MafB can shape the macrophage phenotype. 9,10 Moreover, MAFB expressed in macrophages is related to the restoration of homeostasis, which can be achieved by resolution of in ammation, removal of dead cells, 11 and the uptake of oxidized cholesterol. 12 Mutations in human MAFB have been reported to cause diseases, such as multicentric carpotarsal osteolysis and Duane retraction syndrome with focal segmental glomerulosclerosis. 13,14 Furthermore, MAFB is expressed in many speci c disease-related macrophages, including in ltrating macrophages at the site of bacterial infections, 15 monocyte-derived MHC-II-positive macrophages in myocardial infarction models, 16 atherosclerotic regions, 12,17,18 macrophages in rat hind-limb ischemia models, 19 and kidney-resident macrophages in renal artery stenosis models. 20 In addition, MAFB expression has been reported in pro brotic macrophages in lung brosis induced by bleomycin, 21 microglial cells following peripheral nerve injury of the spinal cord, 10,22,23 and peritoneal macrophages in human patients with peritoneal dialysis. 24 Moreover, MAFB expression in TAMs has been reported in mouse mammary tumors 25 and human skin, colon, and melanoma tumors. 26 Based on our previous work 1 we showed that the transcription factor MAFB could be a biomarker for TAMs in both lung adenocarcinomas, and in lewis lung carcinoma (LLC) mouse model as well as MAFB was expressed in anti-in ammatory alternatively activated M2 macrophages in vitro.
The ecosystem of the tumor microenvironment (TME) is characterized by diverse heterogeneity and is rapidly evolving. Homeostasis of the TME is regulated by an intimate crosstalk among its cellular components, involving malignant, endothelial, stromal, and immune cells. 27 While the TME comprises both adaptive and innate immune cells, the later can facilitate cancer cell growth and metastasis, induce tumor angiogenesis, suppress antitumor immune response, and modulate response to anticancer therapies. These innate immune cells include Mφs, dendritic cells, neutrophils, myeloid-derived suppressor cells, natural killer cells, and innate lymphoid cells, 28 among which monocyte-derived Mφs account for 50% of the tumor mass. 29 Mφs can be classi ed as in ammatory M1, which are classically activated and exert antitumor effect, or immune-suppressive M2, which are alternatively activated and has protumor effect. 30 However, Mφs are dynamic cells with functional plasticity. 31 In fact, M2 Mφs can be broadly classi ed into M2a, M2b, M2c, and M2d subsets 28 based on Mφs polarization inducers as follows: IFNγ and LPS for M1; IL4, IL10, and IL13 for M2a; TLR agonists for M2b; IL10, TNFα, and glucocorticoids for M2c; and TLR and adenosine A2A receptor for M2d. 32 Hence, the dynamicity of Mφs accounts for the high degree of heterogeneity of TAMs within the TME and re ects the ability of TAMs to acquire various phenotypic, metabolic, and functional pro les in response to ecosystem uctuations of the TME 33 among patients suffering from speci c cancer types or even within different lesion sites in the same patient. 27 Generally, de nite TAM subsets (M2) support oncogenesis, angiogenesis, immunosuppression, and therapy resistance, resulting in poor prognosis, whereas other TAM populations (M1) exhibit anti-tumor activity (tumoricidal) and increase the e cacy of various anticancer immunotherapies. 34 TAMs are suitable indicators of tumor malignancy 35 hereafter, multi-marker methods are available to distinguish TAMs. [36][37][38][39][40][41] There are many diagnostic and prognostic markers for TAMs within the TME. One of the most common is the CD68 (pan-macrophage antibody). 42 The limitation of depending on CD68 is that it is not singularly expressed by TAMs but also was expressed by many other TME cells, such as malignant epithelial and stromal cells. 43 Notably, M1 and M2 subsets cannot be recognized via the single labeling of Mφs with CD68 marker antibodies. Hence, multiplex markers for multiplex Mφs (usually two or three different Mφs markers) have been recently performed to identify M1 and M2 subsets but this approach is costly and time consuming. 44 Also, CD163 (hemoglobin-scavenger receptor), and CD204 (Mφ-scavenger receptor-1) markers may be used to identify Mφs. CD163 and CD204 have been proposed to affect tumor progression by stimulating metastasis, neovascularization, and immune evasion. They exhibit dual action by inhibiting the antitumor activity of both M1 pro-in ammatory TAM and T-helper (Th1) cells and activating regulatory T cells and Th2 cells. 42,45,46 Although both CD163 and CD204 are expressed speci cally on TAM, they have different expression levels, functional signi cance, and prognostic values in various cancer types. This suggests that CD163 and CD204 are not expressed in identical TAM populations. [36][37][38][39][40] Accordingly, both levels of CD163 and CD204 are monitored to evaluate the polarization of TAM subsets. 41 As abovementioned, TAMs support cancer cells invasion and migration, that can be achieved through regulating genes related to metastasis. Hence, also many TAM's identi cation markers can be used as metastasis markers in various tumors models and experimental systems. 47 For example, invasion and migration of cancer cells can be studied by checking the CD11b + /Gr1 mid/low TAMs in MC38 and LLC tumor models in mouse in vitro, 48 F4/80 + and/or CD11b + Gr1-F4/80 + TAMs in MMTV-PyMT in Mouse in vivo, 49,50 CD68 + CD163 + in THP-1 and patients samples in human in vitro and in human colorectal cancer samples, respectively, 51 . Moreover, CSF-1R + TAMs in Hepatocellular Carcinoma were related to Increased intrahepatic metastasis, tumor recurrence and reduced patients survival. 52 Previous studies have reported that TAMs express Programmed cell death protein 1 (PD-1) 53 and Programmed cell death ligand 1 (PD-L1) 54,55 within the TME.

PD-1 is an immune checkpoint receptor upregulated in activated T cells for the induction of immune
tolerance. Tumor cells frequently overexpress PD-L1, thus facilitating immune escape. 56 Mouse and human TAMs expressing PD-1 (herein, TAM PD-1) are negatively correlated with phagocytic potency against tumor cells, and the blockade of PD-1/PD-L1 receptors in vivo increases macrophage phagocytosis, reduces tumor growth, and prolongs survival of cancer mouse models in a macrophagedependent manner. 53 Several recent studies have suggested that one of the immunosuppressive functions of TAM is performed by regulating PD-L1, which is expressed on TAM in many cancer types, and stimulating the apoptosis of T cells via interaction with PD-1 on CD8+T cells. 54,57 Therefore, we aimed to identify MAFB as a suitable biomarker for the TAM PD-1 + and/or TAM PD-L1 + , thereafter it can be used as a target for cancer immunotherapy development directing against these cell populations.

Animals
Mice with C57BL/6J background raised in our center (lab animal resource center, university of tsukuba, japan) were used. To generate MafB conditional knockout (cKO) mice, the Mafb gene was anked by a loxP constituent were used. 58 Subsequently, these mice were crossed with Lysm-Cre knock-in mice to obtain (Mafb ox/ ox ::Lysm-Cre) and the control WT (wild type) and (Mafb ox/ ox ) animals. Mafb GFP knock-in mice (Mafb heterozygous (Mafb gfp/+ ) were created by inserting GFP at the Mafb locus as previously described. 3 Genotypes were veri ed by analyzing the genomic DNA on mouse tail tips. The primer sequences used were as previously reported. 12 Mice were maintained under speci c pathogen-free conditions at the Laboratory Animal Resource Center, University of Tsukuba, Japan. All experiments were performed in compliance with Japanese institutional laws and guidelines and were approved by the University of Tsukuba Animal Ethics Committee (approval number 19-141).

Statistical analysis
Data are expressed as the mean ± S.E.M and analyzed using Welch's t-test. The Kaplan-Meier method was used to estimate overall survival time, the difference in survival was compared by log-rank test, the two paired groups by Wilcoxon test, and the different survival distributions by Tarone-Ware test. In survival analysis, we adjusted the p-value by Dunn-Sidak correction. Based on Patients data that we received from the hospital of university of Tsukuba, TNM staging system was used to classify the severity and the extent of spread among all human cancer Patients participated in this study. Differences were considered statistically signi cant at p < 0.05.

MAFB was expressed in various human cancers
To con rm that MAFB is indeed expressed in different human cancer types; 130 patients suffered from lung adenocarcinomas, whereas 6 were diagnosed with colon cancer, 6 with liver cancer, and 2 with pancreatic cancer. Immunohistochemical analysis using human anti-MAFB antibodies con rmed that MAFB was expressed in all cases (Fig. 1).

MAFB was signi cantly increased in Lung adenocarcinomas showing metastasis potential.
By performing immunohistochemical analysis for 2 groups of human lung adenocarcinomas samples, group one was showing no tendency for metastasis (Metastasis -) (n = 60) ( Fig. 2A) and group 2 showed propensity for metastasis (Metastasis +) (n = 60) (Fig. 2B), later by morphological quantifying of the MAFB signals expression in relative to the total tissue area, we con rmed that MAFB expression was signi cantly increased in Lung adenocarcinomas showing metastasis potential ( Supplementary Fig.3 and Fig.2 C).
Moreover, we con rmed that increasing MAFB expression in group 2 was related to poor survival rate compared to group one (Fig.2 D). These ndings indicates that increased MAFB expression in human lung adenocarcinomas was related to increasing tendency for metastasis besides it was a prognostic indictor for increased risk and poor survival rate.
MAFB was sigini cantly increased in severe colon cancers simulatenously with CD204 and CD68.
We compared the expression of MAFB and other common malignancy markers for TAMs. Hence, we performed immunohistochemical analysis on 2 groups of human Colon cancers, group one representing the mild cases (n = 3) and group 2 representing the severe cases (n = 3) using human anti-MAFB, human Anti-CD204, and human anti-CD68. We found that MAFB was expressed at nearly the same expression sites of CD204 and CD68 markers in both mild cases (Fig.3 A) and severe cases (Fig.3 B). Furthermore, MAFB was speci cally expressed in the nuclei of macrophages, while CD204 and CD68 were expressed on the outer cell surfaces of many cells within the TME in the mild cases (Fig.3 C,E) and in the severe cases ( Fig.3 D,F), respectively. However, by using CD204 and CD68 only, it was di cult to mark the macrophages separately from other cells.
Then we performed morphological quantifying of the MAFB signals, CD204 and CD68 expression in relative to the total tissue area, and we con rmed that MAFB expression was signi cantly increased in simultaneously with increasing the expression of CD204 and CD68 in the group 2 representing the severe colon cancers cases (Supplementary Fig.1).
MAFB was signi cantly increased in severe Liver cancers simultaneously with CD204 and CD68.
In the same way to the Colon cancer samples, we performed immunohistochemical analysis on two groups of human Liver cancers, group one was representing the mild cases (n = 3), and group 2 representing the severe cases (n = 3) also using human anti-MAFB, human Anti-CD204, and human anti-CD68. As the previous result, we con rmed that MAFB was expressed at nearly the same expression sites of CD204 and CD68 markers in both mild cases (Fig.4 A) and severe cases (Fig.4 B). Similarly, MAFB was speci cally expressed in the nuclei of macrophages, while CD204 and CD68 were expressed on the outer cell surfaces of many cells within the tumor tissues in the mild cases (Fig.4 C, E) and in the severe cases ( Fig.4 D, F), respectively.
We also performed morphological quantifying of the MAFB signals, CD204 and CD68 expression in relative to the total tissue area, similarly to the abovementioned, the MAFB expression was signi cantly increased in simultaneously with increasing the expression of CD204 and CD68 in group 2 representing the severe liver cancers cases ( Supplementary Fig.2).

MAFB was expressed in severe pancreatic cancers
Finally, we performed immunohistochemical analysis on a group of severe human pancreatic cancers patients (n = 2) using human anti-MAFB (Fig. 5A), human Anti-CD204 (Fig. 5B) and human anti-CD68 (Fig.5 C). In the same way as the previous results for Colon and Liver cancers, we found that MAFB was speci cally expressed in the nuclei of macrophages, while CD204 and CD68 were expressed on the outer cell surfaces of many other cells within the TME.
GFP-MAFB signals were expressed within the PD-1 or/and PDL-1 positive population in Lewis lung carcinoma (LLC) mouse model.
We hypothesized that MAFB could be a suitable target for drug development targeting TAM PD-1 + and/or TAM PDL-1 + cells. Hence, we performed FACS analysis of 3-weeks old LLC tumors collected from Wild type mice (n = 5) and Mafb gfp/+ mice (n = 5). Then we checked the expression of GFP, F4/80, PD-1, and PD-L1 after gating the viable cells. Our results showed that GFP signals, indicating MAFB expression, were expressed within the PD-1 + or PDL-1 + populations separately (Fig.6 A) and also in about 35.3% of the double-positive TAM PD-1 + /PD-L1 + population (Fig.6 B).
MAFB was expressed in TAM PD-1 + or/and TAM PDL-1 + cells but did not regulate their expression in the LLC mouse model To con rm whether MAFB expressed in TAM can regulate the PD-1 or PD-L1 expression, we injected LLC cells in group of control mice (Mafb ox/ ox ) (n = 5) and in a group of mice were Mafb ox/ ox ::Lysm-Cre (n = 5). Later 3-weeks old LLC tumors collected from the 2 groups, and we checked the expression of F4/80, CD204, PD-1, and PD-L1 after gating the viable cells. We found no signi cant difference in the expression of PD-1 and PD-L1 between the WT group (Fig.7 A) and the MAFB-Conditional Knocked-out group (Fig.7 B). These ndings may suggest that MAFB is expressed in TAM PD-1 + or/and TAM PD-L1 + cells but does not regulate their expression.
To con rm the MAFB expression within PD-1 and PD-L1 populations in human cancers, we performed immunohistochemical analysis on cancer tissue serial sections collected from a group of human lung adenocarcinomas patients (n = 5) using human anti-MAFB, human Anti-PD-1, and human Anti-PD-L1. We con rmed that MAFB (Fig8. B) was nearly co-expressed in the same tissue area with PD-1 + (Fig.8 C) and/or PD-L1 + (Fig.8 D) expression. These results indicate that MAFB can be a suitable marker for TAM PD-1 + or/and PD-L1 + cells within lung adenocarcinomas cancers, and consequently, it could be an appropriate target for cancer immunotherapy.

Discussion
Previously, we have shown that MAFB could be a suitable marker for TAMs in human lung adenocarcinomas and in LLC mouse models. 1 In this study, we proved that Transcription factor MAFB could be a biomarker for various human cancers, including Hepatic cancers, Colon cancers, and pancreatic cancers side by side with our previous ndings, human lung adenocarcinoma, and LLC mouse models.
Moreover, here in this study we rst showed that MAFB high expression was correlated to the increased metastatic potential of cancer cells in human lung adenocarcinomas patients. Besides, high MAFB expression was drew a parallel to increased risk and poor survival rate among these patients.
Additionally, our results harmonize with the previous studies stated that MAFB could have regulating roles in human hepatocellular carcinoma, 59 human colorectal cancers, 60 and pancreatic ductal adenocarcinoma, 61 but we rst showed that high MAFB expression in these cancers was intensely related with the increasing severity and poor prognosis.
What is more is that many markers can be used to mark M2 type TAMs within the TME in both human cancers samples or any other experimental systems, these markers including CD68, CD204, CD136, CD11b, F4/80, and CSF-1R. 42,[45][46][47][48][49][50][51][52] However, in clinical practice these markers have diverse expression levels that indicating distinct TAMs phenotypes and are correlated to different prognostic values, besides some others cells within TME or M1 TAM can express some of these markers as well. [36][37][38]41,43,44 While we showed that transcriptional factor MAFB was expressed speci cally in M2 TAMs in various human cancers and murine tumor models and brightly expressed within the TAMs nuclei making their morphological quanti cation easier. Furthermore, high MAFB expression within the studied cancers was correlated to overall poor prognostic values.
Finally, we showed that MAFB was expressed in TAM PD + and/or TAM PDL-1 + cell populations in both human lung adenocarcinomas and LLC mouse models which meets with the previous studies. [53][54][55] This nding indicates that MAFB can be a suitable marker for TAM PD-1+ or/and PD-L1+ population within lung adenocarcinomas TME and accordingly it could be a suitable target for cancer immunotherapies Targeting these cells populations.
In conclusion, we assumed that transcription factor MAFB could be used as a biomarker for many human cancers and murine tumor models. A marker for cancer severity in many human cancers including human lung adenocarcinomas, hepatic cancers and colon cancers. In our more detailed studies on human lung adenocarcinomas, MAFB can be a marker for many prognostic values including tumors severity, cancer cells migration, invasion, metastasis, and overall poor survival rates.
Finally, MAFB can be a potential target for cancer immunotherapies. However, more studies should be conducted to determine the molecular roles of MAFB within TAMs and to detect the exact roles of MAFB in TAM PD + and/or TAM PDL-1 + cells in different human cancer types. Figure 1 MAFB was expressed in various human cancers. Immunohistochemical staining of (A) human lung adenocarcinomas (n = 10), (B) Liver cancers (n = 6), (C) Colon cancers (n = 6), and (D) Pancreatic cancers (n = 2), with antihuman-MAFB staining Tumor associated macrophages (TAMs).