Expression of CD38 on Macrophages Predicts Improved Prognosis in Hepatocellular Carcinoma

Background: CD38 is involved in the adenosine pathway, which represents one of the immunosuppressive mechanisms in cancer. CD38 is broadly expressed across immune cell subsets, including human macrophages differentiated in vitro from monocytes, but expression by tissue-resident macrophages remains to be demonstrated. Methods: Tissue samples were obtained from 66 patients with hepatocellular carcinoma (HCC) from Singapore and analyzed using immunohistochemistry. Tumor-infiltrating leukocytes (TILs) were further examined using DEPArray™, and the phenotype of freshly isolated TILs was determined using flow cytometry. Results: CD38 was frequently co-expressed with the macrophage-specific marker CD68. CD38+CD68+ macrophage density was associated with improved prognosis after surgery, while total CD68+ macrophage density was associated with poor prognosis. DEPArray™ analysis revealed the presence of large (>10 μm), irregularly shaped CD45+CD14+ cells that resembled macrophages, with concurrent CD38+ expression. Flow cytometry also revealed that majority of CD14+HLA-DR+ cells expressed CD38. Conclusion: CD38 expression was clearly demonstrated on human macrophages in an in vivo setting. The positive association identified between CD38+ macrophage density and prognosis may have implications for routine diagnostic work.


INTRODUCTION
Human hepatocellular carcinoma (HCC) is the most common primary malignancy of the liver (1,2), and represents a severe, worldwide threat to human health and quality of life. Patient survival after surgery remains relatively low, with 5-year survival rates after resection for early-stage disease ranging between 17 and 53%, and recurrence rates being as high as 70% (2)(3)(4). Therefore, it is important to identify biomarkers that reliably distinguish patients at high risk of recurrence. HCC progression is known to be driven by chronic inflammation, which may arise from viral infection, hemochromatosis and alcoholic or nonalcoholic steatosis (5). The innate immune system also appears to serve a key role in the progression of HCC. For example, tumor-associated macrophages (TAMs) are widely considered to be correlated with poor prognosis in patients with HCC (6)(7)(8)(9).
CD38 is a multifunctional, membrane-associated ectozyme belonging to the ribosyl cyclase family, and is widely expressed across the immune system. This molecule was initially discovered on the surfaces of thymocytes and T cells. Subsequently, it was found to be expressed by other types of immune cell and certain non-lymphoid tissues, including brain, eye, gut and prostate tissues (10). CD38 cleaves NAD + and NADP + to generate the Ca 2+ -mobilizing compounds adenosine diphosphate ribose (ADPR), cyclic ADPR (cADPR), and nicotinic acid adenine dinucleotide phosphate (NAADP) (10). It also functions as a receptor, mediating the transduction of signals associated with activation and proliferation. These dual enzymatic and receptorial functions mean that CD38 is involved in a diverse range of cellular activities. In the context of the immune system, CD38 is known to induce pro-inflammatory and regulatory cytokine production in monocytes and dendritic cells, and to induce proliferation and inhibit apoptosis in lymphocytes (11). Downregulation of CD38 alters the migration of neutrophils and monocytes (12), and impairs the innate immune response against Listeria monocytogenes infection (13). CD38 is also known to be present in monocytes, where it acts as a co-accessory molecule for MHC Class II-induced T lymphocyte activation by superantigen (14).
In addition, CD38 is involved in the adenosinergic pathway via its NADase function. This pathway is of significant interest to the field of cancer immunotherapy, as it represents a major alternative immunosuppressive mechanism to the PD-1/PD-L1 pathway. Here, the CD38-CD203a (also known as ENPP1 or PC1)-CD73 salvage pathway generates adenosine through the degradation of pyridine metabolites, such as NAD + . Specifically, CD38 hydrolyses NAD + to ADPR, and CD203a further degrades ADPR to adenosine monophosphate (AMP). Following this, CD73 dephosphorylates AMP to adenosine (15)(16)(17). Tumor proliferation, survival, adhesion and migration may be regulated through the adenosine pathway. For example, in immune cells, adenosine molecules hamper vital effector cell functions and may be involved in mediating T cell exhaustion (18). A recent study using a murine lung cancer model reported that CD38 expression on cancer cells was upregulated in response to PD-1/PD-L1 blockade, resulting in the inhibition of CD8 + T-cell function via adenosine receptor signaling (19).
Our group recently established the relevance of CD38 to HCC by identifying a correlation between CD38 + tumor-infiltrating leukocyte (TIL) density and improved patient survival (20). We studied the expression of CD38 on lymphocytes, Natural Killer (NK) cells, NKT cells and monocytes, but not on macrophages. Indeed, CD38 expression has previously been reported on all major leukocyte populations, including B cells, T cells, NK cells, monocytes and dendritic cells (10). However, with respect to macrophages, data concerning CD38 expression is limited. Upregulation of CD38 has been observed on murine and human macrophages following in vitro stimulation with IFN-γ ± lipopolysaccharide (LPS), suggesting an association between CD38 and the pro-inflammatory M1 state (21,22). However, in vivo expression of CD38 by tissue-resident macrophages remains to be demonstrated, and in vitro cultures may not represent in vivo conditions accurately. Considering the known correlation between TAMs and poor prognosis in HCC (6)(7)(8)(9), this requires further investigation.
In the present study, we sought to ascertain the in vivo expression of CD38 on macrophages from tumor tissues obtained from patients with HCC. Using immunohistochemistry (IHC) and Multiplex immunofluorescence (mIF), we confirmed the co-expression of CD38 with the macrophage-specific marker CD68. Notably, through Kaplan-Meier survival analysis and multivariate Cox regression, we established that the presence of CD38 + CD68 + macrophages predicted improved prognosis after surgery, while total CD68 + macrophage density was associated with poor prognosis. In vitro functional studies using THP-1 derived macrophages revealed CD38 expression to be associated with the M1 state, characterized by CD80 expression and secretion of TNFα and IL-6, all of which contribute to the antitumor immune response. Using the DEPArray TM , an automated platform that is able to identify and recover single cells with high resolution and purity, we visualized the expression of CD38 on macrophage-like cells isolated from HCC tumors. Finally, using flow cytometry, the co-expression of CD38 with CD14 + HLA-DR + cells was further confirmed.

Patients and Tumor Samples
A total of 66 archival formalin-fixed, paraffin-embedded (FFPE) specimens from patients diagnosed with HCC between January 2001 and December 2011 at the Department of Anatomical Pathology, Division of Pathology, Singapore General Hospital, were analyzed. All samples were obtained prior to chemo-or radiotherapy. Clinicopathological parameters, including patient age, tumor size, histologic growth pattern, grade, subtype, lymphovascular invasion and axillary lymph node status, were reviewed (Supplementary Table 1). Tumors were staged and graded according to the AJCC staging system (23) and the Edmonson-Steiner grading system (24). Ishak fibrosis scoring (25,26)

Tissue Microarray Construction
Tumor regions for tissue microarray (TMA) construction were selected based on pathological assessment, with >50% of the sample representing tumor area. For each sample, two or three representative tumor cores of 1 mm diameter were transferred from donor FFPE tissue blocks to recipient TMA blocks using a MTA-1 Manual Tissue Arrayer (Beecher Instruments, Inc., Sun Prairie, WI, USA).

ELISA
ELISA was performed to detect IL-6, TNFα, IL-10, EBI3 (IL-35 subunit), and IL-12(p70). IL-6, TNFα and IL-10 kits were purchased from eBioscience (San Diego, CA, USA); EBI3 kit was purchased from R&D Systems, Inc. (Minneapolis, MN, USA); IL-12(p70) kit was purchased from RayBiotech (Norcross, GA, USA). Wells of microtiter plates were coated (18 h, room temperature) with respective capture antibody in 100 µl of coating buffer and were then blocked with 1% BSA in PBS (reagent diluent) for 1 h at room temperature. Samples or standards (100 µl) were loaded in duplicates and incubated for 2 h at room temperature, followed by the addition of 100 µl detection antibody for additional 2 h at room temperature. HRP-conjugated with Strepavidin (1:20,000) in reagent diluent was added (20 min, room temperature) and the reaction was visualized by the addition of 50 µl chromogenic substrate (TMB) for 10 min. The reaction was stopped with 100 µl H 2 SO 4 and absorbance at 450 nm was measured using ELISA microplate reader (Perkin Elmer's Enspire 2300). Plates were washed three times with washing buffer (PBS, pH 7.4, containing 0.05% (v/v) Tween 20) after each step. As a reference for quantification, individual standard curves were established by serial dilution of respective recombinant cytokines.

Tissue Dissociation and Isolation of Leukocytes
Liver tissue was cut into fine pieces and digested with 0.5 mg/ml Collagenase IV (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 0.05 mg/ml DNAse I (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) in complete RPMI (Gibco; Thermo Fisher Scientific, Inc.) for 30 min at 37 • C. Digested tissue was filtered using a 70 µm cell strainer. Cells were pelleted and treated with red blood cell lysis buffer (G-Biosciences, St Louis, MO, USA) for 5 min at room temperature. FIGURE 2 | High density of CD38 + , but not total, CD68 + TAMs was associated with improved overall survival rates in patients with HCC. (A,B) Kaplan-Meier curves showing overall and disease free survival rates, respectively, for patients grouped according to high or low density of CD38 + CD68 + macrophages in the tumor microenvironment. (C,D) Kaplan-Meier curves showing overall and disease free survival rates, respectively, for patients grouped according to high or low density of CD68 + total macrophages in the tumor microenvironment. Cut-offs for high/low density of CD38 + CD68 + macrophages (0.5690) and CD68 + total macrophages (0.1918) were based on optimal values calculated using statistical software.

Flow Cytometry
Cells were incubated with Fixable Viability Dye eFluor TM 455UV (eBioscience; Thermo Fisher Scientific, Inc.) for 30 min at 4 • C for live/dead discrimination. Fc receptors were blocked using Human TruStain FcX TM (BioLegend; San Diego, CA, USA) for 5 min at room temperature. Cell surfaces were labeled with antibodies against markers of interest (refer to Supplementary

Validation, follow-Up, and Statistical Analysis
Follow-up data were obtained from medical records. Diseasefree survival (DFS) and overall survival (OS) were defined as the time from diagnosis to recurrence or death/date of last follow-up, respectively. Statistical analysis was performed using RStudio 1.

Expression of CD38 by Tumor-Associated Macrophages (TAMs) in HCC Is Visualized by Immunohistochemistry (IHC)
IHC staining of FFPE tumor sections revealed the expression of CD38 by cells residing in the liver sinusoids. As shown in Figure 1A, cells that were stained had an irregular shape with many cytoplasmic extensions. We therefore hypothesized that these CD38-expressing cells may be macrophages, based on both their morphology and anatomical location (42). Moreover, previous studies have demonstrated that CD38 is expressed on macrophages in animal models (13,22,43,44) and that CD38 expression on macrophages is correlated with inflammation in humans (21). Thus, we performed Opal-Vectra multiplex IHC on the specimens, using protocols previously optimized and reported by our group and other laboratories (20,27,(30)(31)(32)(33)(34)(35), to identify the co-expression of CD68 and CD38 in liver immune cells. CD38 and CD68 were frequently co-expressed by the same cells (Figures 1B-D), with 62.9 ± 19.2% of CD38 + cells bearing the macrophage marker CD68 and 32.3 ± 24.1% of CD68 + cells being CD38 + (Figures 1E,F). This suggested that a large proportion of CD38 + cells were macrophages.
High CD38 + Macrophage Density Is Associated With Improved Survival in HCC Next, we investigated whether CD38 + macrophage density in tumors affected survival outcome in patients with HCC. Patients with high density of CD38 + CD68 + macrophages experienced significantly improved overall survival (OS) compared with those with low density of CD38 + CD68 + macrophages (P = 0.0367; Figure 2A); in contrast, disease free survival (DFS) did not differ significantly between groups (P = 0.2400; Figure 2B). Multivariate analysis further supported the association between a high density of intratumoral CD38 + macrophages and significantly improved OS (HR = 0.41; 95% CI 0.18-0.93; P = 0.0322; Table 1).

CD38 Up-regulation Was Associated With M1 Macrophages, Characterized by Expression of Co-stimulatory CD80 and a Pro-inflammatory Cytokine Profile
To ascertain the function of CD38 + macrophages, we took advantage of the THP-1 cell line to investigate surface expression and cytokine secretion of CD38 + macrophages under in vitro culture conditions. THP-1 cells were polarized into M1/M2 macrophages that showed differential expression of CD80 and DC-SIGN (Figures 3A,B), consistent with the results of a previous study (45). Notably, robust expression of CD80 was observed on M1 macrophages but not M0 or M2. Additionally, we found up-regulation of CD38 on M1 macrophages (Figures 3C,D), concomitant with pro-inflammatory IL-6 and TNFα secretion ( Figure 3E). By contrast, M0 or M2 macrophages did not secrete IL-6 or TNF-α ( Figure 3E). As compared to IL-6 (786.7 ± 117.1 pg/ml) and TNFα (18.2 ± 1.9 ng/ml), the level of IL-35 secretion was minimal (82.0 ± 4.2 pg/ml) and there was no detectable IL-10 and IL-12(p70) secretion from M1 macrophages ( Figure 3F). Altogether, our results suggest that CD38 + macrophages are associated with an M1 phenotype that is characterized by higher CD80 expression and a pro-inflammatory cytokine profile.

Confirmation of CD38 Expression on Macrophages by DEPArray TM and Flow Cytometry
To further confirm the expression of CD38 on TAMs from patients with HCC, tumor-infiltrating leukocytes (TILs) were examined using the DEPArray TM , which is an automated platform capable of visualizing and isolating single cells with high resolution and purity. Images captured by the DEPArray TM revealed the presence of large (>10 µm), irregularly shaped CD45 + CD14 + cells that resembled macrophages (Figure 4). CD38 expression was clearly detected on the surfaces of these cells. Next, flow cytometry was performed to analyse TILs in greater detail. Monocytes/macrophages were identified in tumor tissues from four HCC patients using the gating strategy shown in Figure 5A. The vast majority of CD14 + HLA-DR + cells expressed CD38 (84.9 ± 7.4%; Figure 5B), indicating enrichment of CD38 + monocytes/macrophages in the tumors of these four patients. In terms of total CD38-expressing leuckocytes, monocytes/macrophages constituted only 34.6 ± 26.6% (Figure 5B), indicating the presence of other CD38expressing immune lineages.

DISCUSSION
Through multiplex IF and conventional IHC, our group was able to identify and enumerate CD68 + TAMs present in the tumor samples from HCC patients. We established that CD68 + (B) % of monocytes/macrophages that expressed CD38 (black column) and % of CD38 + leukocytes that were monocytes/macrophages (gray column). The data represents the means ± s.d.
TAM density was correlated with poor prognosis in HCC, in line with the finding of a previous report (6). The tumor microenvironment is known to be dominated by cytokines and growth factors that establish a Th2-type, anti-inflammatory immune environment that favors the survival of tumor cells. Under such environmental influences, TAMs are driven toward a M2 phenotype. This transition is accompanied by increased pro-tumoral activities, including the suppression of the adaptive immune response and the promotion of cancer proliferation, angiogenesis and extracellular matrix remodeling (46). To the best of our knowledge, the present study represents the first instance that CD38 expression has been demonstrated on human macrophages in an in vivo setting. Previously, CD38 expression was reported on macrophages isolated from mice (13,47), on a murine macrophage cell line (48) and on human macrophages under ex vivo conditions (21). We initially observed the co-localization of CD38 and CD68 signals in the sinusoids of liver sections, and subsequently from freshly isolated TILs. CD38 was visually confirmed on macrophagelike CD14 + cells using DEPArray TM , and also on CD14 + HLA-DR + monocytes/macrophages using flow cytometry. Notably, statistical analysis of the IHC data revealed that CD38 + CD68 + macrophage density was positively associated with improved prognosis, in direct contrast to total CD68 + macrophage density. This suggested the existence of functionally-distinct macrophage subsets, with CD38 marking a subpopulation of proinflammatory macrophages (21) that may play a role in tumor suppression. Supporting this hypothesis was data from our in vitro functional study of THP-1 cells polarized to M1 or M2 macrophages. Specifically, M1 macrophages were distinguished by higher expression of CD80-a costimulatory molecule that signals through CD28 to amplify T cell activation and contributes to anti-tumor immunity (49,50). We also found that CD38 was up-regulated on M1 macrophages, concomitant with a proinflammatory cytokine profile characterized by robust IL-6 and TNFα secretion, with minimal IL-35 secretion and no secretion of IL-10 and IL-12 cytokines. By contrast, M0 or M2 macrophages did not produce detectable levels of all cytokines examined. A previous study reported that deletion of IL6 or TNFα in a mouse model of HCC accelerated tumor development (51). Altogether, our results suggest that CD38 + macrophages are associated with an M1 phenotype that is characterized by higher CD80 expression and a pro-inflammatory cytokine profile, all of which contribute to the anti-tumor immune response. This effect may account for the improved prognosis of HCC patients with higher proportion of tumor-infiltrating CD38 + macrophages. To further appreciate the complexities of TAMs in liver cancer, a future single cell study could be conducted. In view of this, the DEPArray TM platform may be useful for isolating single cells with high purity, as required for sensitive downstream genomic and expression analyses (52)(53)(54).
It is important to note that CD38 is involved in the production of adenosine from NAD + . Adenosine is known to be an inhibitor of effector T cells and is produced by various regulatory cells expressing CD38, including myeloid-derived suppressor cells, mesenchymal stem cells and NK cells (55). In view of this, the expression of CD38 on tumor-infiltrating macrophages in the present study raises the possibility of participation in this mode of immunosuppression. This suggests the presence of a complex interplay between the inflammatory response and immune suppression via adenosine production.
When flow cytometry was used to investigate TILs from four patients with HCC, high levels of CD38 expression were found among the monocytes/macrophages. However, these monocytes/macrophages represented only 34.6% of total CD38 + leukocytes, which indicated the presence of other CD38 + immune lineages. In our previous study concerning TILs in HCC, we reported expression of CD38 on lymphocytes, NK cells, NKT cells and monocytes (20). Such broad expression was in line with the reported ubiquity of CD38 across the immune system (10).
Of interest to us in future studies is CD157, a paralogue of CD38 derived from gene duplication. CD157 also possesses dual receptorial and NADase functions and is widespread across lymphoid tissues, including macrophages (56). Given the high level of sequence similarity between CD38 and CD157, it is important to establish whether CD157 + macrophages also predict improved prognosis in patients with HCC.
In conclusion, this study confirmed the expression of CD38 by human macrophages in vivo, and established that CD38 + macrophages are correlated with improved prognosis after surgery. The function of CD38 in myeloid cells warrants further study, and this marker may be utilized in routine diagnostic work in the era of cancer immunotherapy.

ETHICS STATEMENT
The