Activation of B-1 Cells Promotes Tumor Cell Killing in the Peritoneal Cavity

The majority of patients who succumb to cancer do so not from primary tumors, but from metastatic disease (1). In particular, the spread of malignant cells to the peritoneal cavity carries a grave prognosis, especially when associated with ascites development (2). The peritoneal surface and cavity may be affected by malignant epithelial (carcinomatosis), mesenchymal (sarcomatosis) and more rarely, lymphoid (lymphomatosis) cells (3). Peritoneal carcinomatosis due to cancers derived from malignant ovarian, colon, appendiceal, as well as breast (infiltrating ductal carcinoma) tissue (2, 4), involves extensive spread and implantation of tumors and eventually, ascites development. The therapeutic options are limited and treatment plans are often palliative rather than curative, with cytoreductive surgery and hyperthermic intraperitoneal chemotherapy currently representing the most common treatments (5). Immunotherapeutic approaches to treat peritoneal malignancies have been limited, although results obtained from some mouse models offer hope for future treatments (5).

Understanding the distinctive environment of the peritoneal cavity is key to devising optimal immunotherapies for peritoneal metastasis. The peritoneal space represents a unique immune environment (6). Monocytes and macrophages comprise the majority of leukocytes in the cavity under normal conditions. Innate-like B-1 cells, composed of CD5⁺ B-1a cells and CD5⁻ B-1b cell populations, are the second most numerous (6, 7). These B cells have been studied most in mice, but have been identified in non-human primate peritoneal and omental tissue (7, 8) as well as in human blood (9). B-1 cells produce natural IgM and IgA as well as pathogen-specific antibodies, which are critical for host defense and clearance of apoptotic debris (10). Although they are known to have a critical role in protection against infectious diseases, their role in cancer is not well understood. Exchange of plasma components supplies the peritoneal fluid with many of the proteins found in the circulation, including B-1 cell–derived antibodies (6). However, additional soluble factors present in the peritoneal cavity, including IL10 produced by B-1 cells and macrophage-produced prostaglandins, indoleamine 2,3-dioxygenase, and nitric oxide drive immunosuppression (11–15). Under normal conditions, both peritoneal B-1 cells and macrophages inhibit T-cell activation and peritoneal macrophages additionally inhibit B-cell proliferation and antibody production (11–13, 15, 16). Ascites from patients with carcinomatosis contains high levels of IL10, TGFβ, regulatory T cells, and immunosuppressive macrophages (17), suggesting suppression within the peritoneal cavity is maintained, if not augmented, in peritoneal metastases.

The use of pathogen-associated molecular pattern molecules (PAMP) represents one strategy that is being investigated to overcome immune suppression and bolster antitumor responses (18). This is founded on the evidence of several successful therapeutic approaches in both pre-clinical models and patients using bacterial-derived products. Indeed, over a century ago, studies documented that injecting “Coley's toxin” (consisting of heat-killed Streptococcus pyogenes and Serratia marcescens) into tumors lead to tumor regression and sometimes even cure. Killed Bacille Calmette-Guérin mycobacterium, one of the most widely used antitumor adjuvant therapies in humans, induces regression of transitional bladder cell carcinomas by stimulating a robust local immune response (18). Toll-like receptor (TLR) agonists, such as Imiquimod, have also been developed to treat several different forms of cancer (18). C-type lectin receptor (CLR) ligands such as fungal-derived beta-glucans (19) and cord factor (trehalose-6′,6-dimycolate) derived from mycobacteria (20) also have antitumor effects. Although both TLR and CLR agonists show promise as therapies for promoting antitumor responses, their mechanisms of action are not completely understood.

In the current study, we tested the effectiveness of monophosphoryl lipid A (MPL) and a synthetic cord factor analog, trehalose-6,6′-dicorynomycolate (TDCM) with 0.4% squalene in mouse models of peritoneal carcinomatosis and lymphomatosis. This treatment had potent and rapid effects in suppressing tumor growth and ascites development. Unexpectedly, the antitumor effects required B-1 cells, which harbor natural specificities for tumor-associated carbohydrate antigens (TACA). Treatment elicited B-1 cell production of tumor-reactive IgM, which when combined with complement, elicited potent tumor cell killing. Collectively, our results highlight an unexpected role for the elicitation of tumor reactive-IgM through PAMP stimulation of innate B-1 cells. Our findings have implications for the development of future treatment strategies for peritoneal metastases in human patients.

Wild-type (WT) C57BL/6, CAF₁, A/J, C3^−/−, C4^−/−, μMT (Ighm^tm1Cgn), MHC class II^−/−, IL10^−/−, and μS^−/− mice were from The Jackson Laboratory and Charles River Laboratories. CD19^−/− mice were as described previously (21). Studies were approved by Wake Forest Animal Care and Use Committee.

TA3-Ha cells were obtained from Dr. Richard Lo-Man (Pasteur Institute, Paris, France) in 2010. This stock was tested for rodent pathogens and Mycoplasma (IMPACT IV testing, IDEXX-RADIL). E0771 and EL4 cells (Wake Forest Cell and Viral Vector Core Laboratory) were also negative for mouse pathogens and Mycoplasma and were tested in 2016 (Impact III, IDEXX). For TA3-Ha challenges, aliquots of one ascites stock were used. TA3-Ha cells were expanded in culture in cRPMI+10% FCS for 3 days before injection. EL4 and E0771 were grown in the same medium and were similarly expanded in vitro from a frozen stock immediately before injection. Additional cell line authentication was not performed. TA3-Ha cells stably expressing GFP-luciferase were generated in house using a lentiviral vector expression system with blasticidin selection and flow sorting on GFP⁺ cells, followed by single-cell subcloning. Mice were given 1–2 × 10⁴ TA3-Ha or 2 × 10³-2 × 10⁴ TA3-Ha-GFP-luciferase cells as indicated, in 200 μL PBS intraperitoneally. For IL10 blocking, mice were given 100 μg of IL10–neutralizing monoclonal antibody (mAb; JES5-2A5) or control rat IgG mAb (BioXcell) intraperitoneally on days 0, 2, 5, and 8. Mice developing ascites with signs of distress (lethargy, dehydration, reduced/impaired movement, reduced grooming, labored breathing, etc.) were humanely euthanized. Mice were injected intraperitoneally with 10 μg LPS (Escherichia coli O111:B4, Sigma) or Sigma adjuvant system [containing 10 μg monophosphoryl lipid A (MPL) and 10 μg synthetic trehalose dicorynomycolate (TDCM; mixed in 0.4% squalene; Sigma] in a 200 μL volume.

Mice challenged with TA3-Ha-GFP-luciferase cells were injected intraperitoneally with 3 mg D-luciferin (Promega) 15 minutes before bioluminescent imaging using an IVIS Lumina Series III (PerkinElmer). All images compared within an experiment were captured using the same exposure time (either 1 or 2 seconds).

Unless otherwise indicated, peritoneal B-1a cells were purified using Thy 1.2-Dynal beads, and biotinylated F4/80, GR1, DX5, CD11c mAbs (Biolegend) in conjunction with magnetic bead negative depletion (Biotin-binder beads, Dynal) according to the manufacturer's instructions, followed by positive anti-Ly5.1 (CD5) bead selection according to the manufacturer's instructions (Miltenyi Biotec). CD43⁻ spleen B cells were purified using anti-CD43 magnetic beads (Dynal). B-cell purity was assessed by flow cytometry.

Peritoneal cells were harvested using 10 mL of PBS to lavage the peritoneal cavity. Cells were resuspended in wash buffer (PBS containing 2% newborn calf serum; 2 × 10⁶/mL). Peritoneal cells were pre-incubated with 0.5 μg/mL Fc Block and stained with 2.5 ug/mL biotin-conjugated Helix Pomatia agglutinin (HPA; Sigma) and mAbs conjugated to fluorochromes (from Biolegend, BD Biosciences, and eBioscience): CD4(RM4-5), CD8(Ly-2), NK1.1(PK136), CD49b(DX5), CD25(PC61.5), MHC class I(34-2-12), CD5(53-7.3), Ly6G(Gr-1), CD86(GL-1), CD138(281-2), CD19(1D3), CD11b(M1/70), F4/80(BM8), and CD11c(N418). Cells were washed and stained with streptavidin-conjugated fluorochrome for 20 minutes, washed and fixed with 1.5% buffered formaldehyde. TA3-Ha cells were identified as HPA⁺CD11b^negCD138⁺MHC-I⁺CD19⁻ (22). To detect Ig and C3 bound to tumor cells harvested from peritoneal cavities, cells were incubated with Fc Block for 15 minutes before addition of biotinylated HPA (2.5 μg/mL) and mAbs against CD138, MHC class I, CD19, CD11b, and goat anti-mouse Ig (H+L; Invitrogen) and anti-mouse C3 IgG–FITC (Cappel). Cells were washed in wash buffer and streptavidin-BV510 was added to detect HPA. Isotype controls were used to define positive and negative populations. Serum IgM and IgG binding (1:10 dilution) to tumor cells was assessed as previously described (22). Bovine submaxillary mucin (BSM; MP Biologicals) was included as a blocking agent (100 μg/mL) in some experiments. Cells were fixed in 1.5% buffered formaldehyde and analyzed using a FACSCanto II or Fortessa-X20 cytometer (Becton Dickinson).

Data are shown as means ± SEM with differences assessed using the unpaired Student t test. Differences in Kaplan–Meier survival curves were assessed using the log-rank test.

We tested the effectiveness of MPL and a synthetic cord factor analog, TDCM, in 0.4% squalene (MPL/TDCM) against the aggressive mouse mammary adenocarcinoma, TA3-Ha, which models rapid tumor growth and ascites development in the peritoneal cavity as observed in peritoneal carcinomatosis (23–26). CAF₁ (BALB/c x A/J F1) mice are susceptible to syngeneic (A/J) TA3-Ha cells, with 100% of mice succumbing to the tumor within 4 weeks following intraperitoneal challenge with 2 × 10³ TA3-Ha-GFP-luciferase cells, and 100% of mice succumbing to challenge with 2 × 10⁴ cells within 2 weeks (Fig. 1A and B). However, CAF₁ mice treated with MPL/TDCM had significantly prolonged survival (Fig. 1A and B). IVIS imaging revealed a notable reduction in TA3-Ha-GFP-luciferase as early as 6 days post challenge, with minimal detection of tumor cells at 10 days post challenge in MPL/TDCM-treated mice, whereas untreated mice showed high levels of bioluminescence with extensive ascites development by day 10 (Fig. 1C; 2 × 10⁴). Combined MPL/TDCM and cyclophosphamide treatment modestly improved survival (Supplementary Fig. S1), although remaining survivors did not survive a rechallenge on day 40. Similar results were observed in A/J mice, which are highly susceptible to challenge, with MPL/TDCM-treatment limiting tumor growth at 10 days post treatment (Fig. 1D) and significantly prolonging survival (Fig. 1E). This result is consistent with previous work demonstrating protective effects of natural Mycobacterium tuberculosis cord factor and lipopolysaccharide in the TA3-Ha model (27).

The early clearance of TA3-Ha cells in CAF₁ and A/J mice following MPL/TDCM treatment suggested a rapid and yet effective early innate-like response was playing a role in protection. Of note, the TA3-Ha cell line is highly invasive and mucinous and is so effective at evading the immune system that it is lethal in even non-syngeneic strains (23). To dissect the role of immune cells in MPL/TDCM-induced protection, we took advantage of the amenability of the TA3-Ha peritoneal carcinomatosis tumor model to growth in mice on a C57BL/6 background. 10⁴ TA3-Ha cells injected intraperitoneally into naïve WT C57BL/6 mice induces visible ascites between days 8 to 10 with mice rapidly succumbing to tumor growth and ascites development (22), similar to CAF₁ and A/J mice (Fig. 2A). As observed in CAF1 and A/J mice, C57BL/6 mice were protected by MPL/TDCM treatment (Fig. 2A). Notably, E. coli-derived LPS elicited partial, yet significant protection (∼50% survival; Fig. 2B), whereas MPL or TDB with squalene had only a partial effect (Supplementary Fig. S2).

Five days post treatment, CD11b⁺GR1⁺ myeloid cells and CD4⁺ and CD4⁺CD25⁺ T-cell frequencies were decreased in the peritoneal cavity, whereas B-cell frequencies were increased (Fig. 2C). This corresponded with a significant increase in the overall number of B cells, including B-1a, B-1b, and B-2 cells (Fig. 2D). Moreover, the frequency of Ab-secreting plasmablasts (a rare population in the peritoneal cavity) was significantly increased (Fig. 2E). Finally, consistent with IVIS imaging results, the frequency and number of TA3-Ha cells recovered from peritoneal cavities on day 5 were significantly reduced (≥6-fold) in MPL/TDCM-treated versus untreated mice (Fig. 2F).

We next assessed the importance of B cells in MPL/TDCM-induced antitumor protection. As shown in Fig. 3A, MPL/TDCM treatment did not protect B-cell–deficient (μMT) mice against TA3-Ha challenge. In addition, mice lacking CD19, a B-cell signaling molecule critical for supporting B-cell function and development of B-1a and marginal zone B cells, were not protected (Fig. 3B). Because CD19^−/− mice are severely deficient in B-1a cells (21), we assessed whether transfer of bead-selected WT B-1a cells into CD19^−/− and B-cell–deficient mice could rescue the protective effects of MPL/TDCM treatment (Supplementary Fig. S3). As shown in Fig. 3C, transfer of WT B-1a cells into CD19^−/− mice restored the protective effects of MPL/TDCM treatment, with 80% of B-1a cell recipient CD19^−/− mice surviving challenge relative to 20% of nonreconstituted CD19^−/− mice treated with MPL/TDCM. Similar effects were observed in μMT mice, with 60% survival achieved with B-1a cell reconstitution (Fig. 3D). Importantly, transfer of CD43⁻ spleen (largely B-2 cells) B cells did not restore MPL/TDCM-induced protection in μMT mice (Fig. 3D), even with transfers of up to 10⁷ B cells. FACS-sorted peritoneal B-1a, but not B-1b or B-2 cells, similarly protected μMT mice against tumor challenge (Supplementary Fig. S4). Nonetheless, transfer of peritoneal B-1a cells from MPL/TDCM-treated WT mice that had survived TA3-Ha challenge into μMT mice did not provide protection (Supplementary Fig. S5), indicating residual peritoneal “memory” B-1a cells (perhaps low-Ab secreting) are not sufficient for protection in the absence of MPL/TDCM stimulation. Collectively, these results demonstrate that B-1a cells are critical for the protective antitumor response elicited by MPL/TDCM.

B-1a cells produce a large amount of IL10 (14, 15), which in some cases, actually support antitumor immune responses (28). To determine whether IL10 was playing a role in treatment effects, WT mice were given either anti–IL10 (neutralizing) or control Ab during tumor challenge. No significant difference in survival was observed between mice that received control Ab versus mice that received IL10 blocking Ab (75% vs. 100%, Fig. 4A) or in IL10^−/− mice (Supplementary Fig. S6). Thus, IL10 is not required for MPL/TDCM-induced antitumor effects.

B-1a cells have also been shown to present MHC class II-restricted Ag to CD4⁺ T cells, which could promote antitumor effects (29). However, MPL/TDCM treatment significantly increased overall survival (∼40%) and prolonged time to death in MHC class II^−/− mice during TA3-Ha challenge (Fig. 4B). Nonetheless, the effect of MPL/TDCM treatment in MHC class II^−/− mice was not as striking as in WT mice. We therefore assessed whether B-1a cells lacking MHC class II could provide protection in the context of MPL/TDCM treatment. Reconstitution of CD19^−/− mice with MHC class II^−/− B-1a cells provided a similar level of protection to reconstitution with WT B-1a cells (Fig. 4C). Thus, B-1a cells elicit antitumor protection independently of MHC class II presentation.

B-1a cells play a central role in producing natural Abs (nAb) important for clearance of cellular debris and surveillance against pathogens and tumor cells (10). These Abs are spontaneously produced in the absence of exogenous Ag stimulation and are weakly reactive with self-carbohydrates, lipids, and proteins. PAMPs such as LPS and MPL potently stimulate increases in nAb production (30). We therefore assessed Ab levels bound to tumor cells harvested from the peritoneal cavities of WT and CD19^−/− mice 5 days post TA3-Ha cell challenge. WT mice that received MPL/TDCM treatment had significantly higher levels (∼8-fold) of Ig bound to recovered TA3-Ha cells relative to untreated mice (Fig. 4D). CD19^−/− mice had very little Ig deposited on recovered TA3-Ha cells, regardless of MPL/TDCM treatment. Follow-up experiments demonstrated that increased IgM was bound to tumor cells recovered from treated WT mice (Supplementary Fig. S7).

Serum harvested from WT C57BL/6 mice 5 days post treatment similarly showed significantly increased IgM reactivity with tumor cells relative to untreated (10-fold; Fig. 4E), whereas tumor-reactive IgM was not induced in CD19^−/− mice (Fig. 4E). In contrast, serum IgM reactivity against another mouse mammary adenocarcinoma, E0771, was only modestly increased (∼2-fold) in treated WT mice (Fig. 4F). Increases in TA3-Ha–reactive serum IgG were not observed in either WT or CD19^−/− mice following treatment (Fig. 4E and F). A/J and CAF₁ mice treated with MPL/TDCM also exhibited significant (>12-fold) increases in TA3-Ha-reactive IgM (Fig. 4G). Thus, MPL/TDCM treatment significantly increases IgM levels bound to TA3-Ha cells in vivo and serum IgM binding to TA3-Ha cells in vitro. Notably, TA3-Ha–reactive serum IgM levels were also significantly higher in mice treated with MPL/TDCM in the absence of tumor challenge (Supplementary Fig. S8), and cultured peritoneal B cells activated with either MPL/TDCM or LPS produced significantly increased levels of IgM, but not IgG, reactive with TA3-Ha cells (Supplementary Fig. S9), indicating that TA3-Ha–reactive IgM rapidly arose from non-specific activation and differentiation of B cells.

B-1 cells produce carbohydrate-reactive IgM, including IgM reactive against (22). We therefore assessed whether increased IgM reactivity included increased specificity for TACAs using BSM that shares tumor-associated antigens, such as TF, Tn, and sialyl Tn, with the epiglycanin mucin expressed at high levels by TA3-Ha cells (31, 32). As shown in Fig. 4H and I, inclusion of BSM during staining significantly reduced serum IgM binding in WT mice treated with MPL/TDCM, but had no effect on serum IgM binding from untreated WT mice or treated CD19^−/− mice. Thus, increased IgM reactivity induced by MPL/TDCM is in part, due to mucin-associated antigen reactivity.

Serum IgM reactivity against EL4 cells, a mouse thymoma, was also significantly increased (∼6-fold) in MPL/TDCM-treated WT mice, but not CD19^−/− mice (Fig. 5A). In contrast with the reduced binding observed with TA3-Ha cells, BSM had no effect on serum IgM binding to EL4 cells (Fig. 5B). This is not unexpected as this tumor is not described to express TACAs found on mucins, but expresses other TACAs, including the glycosphingolipid, GD2 (33), for which increased IgM reactivity was found in WT, but not CD19^−/−-treated mice (Supplementary Fig. S10). Thus, MPL/TDCM rapidly and significantly increases the level of IgM reactive against some tumor cells known to express TACAs, including TA3-Ha and EL4 cells.

We next assessed the efficacy of MPL/TDCM treatment on other tumors growing in the peritoneal cavity. EL4 is a C57BL/6-derived mouse thymoma line that also forms ascites in the peritoneal cavity (referred to as peritoneal lymphomatosis). WT male mice challenged with 1 × 10⁶ EL4 cells intraperitoneally became moribund with ascites around 20 days post challenge. However, WT mice treated with MPL/TDCM were completely protected from EL4 tumor growth and ascites development (Fig. 5C and D). As observed with TA3-Ha cell challenge, MPL/TDCM-induced protection against EL4 growth was dependent on B cells and functional CD19 expression, as MPL/TDCM had no effect in CD19^−/− or μMT mice (Fig. 5C and D). Moreover, surviving mice that were given MPL/TDCM treatment following primary EL4 challenge were protected against EL4 rechallenge (Supplementary Fig. S11). In contrast, MPL/TDCM treatment, which had had modest effects in eliciting E0771-reactive IgM, had no protective efficacy against intraperitoneal E0771 challenge (Fig. 5E). Thus, MPL/TDCM induces significant IgM reactivity against EL4 cells, and induces B-cell–dependent protection against EL4 tumor challenge.

The significantly increased binding of IgM to TA3-Ha cells following MPL/TDCM treatment led us to investigate the possibility that IgM was eliciting classical complement activation to aid in tumor cell killing. WT mice treated with MPL/TDCM had significantly increased levels of C3 bound to tumor cells harvested from the peritoneal cavity compared with mice that received TA3-Ha cells only (Fig. 6A). Moreover, MPL/TDCM treatment offered little protection to C3^−/− mice against TA3-Ha challenge (Fig. 6B). Similarly, MPL/TDCM treatment did not provide protection in C4^−/− mice, which are unable to active classical and lectin-initiated complement pathways (Fig. 6C). Thus, C3 and C4 are required for MPL/TDCM-induced antitumor protection.

We hypothesized production of tumor-reactive IgM was the key mechanism by which B-1 cells elicited antitumor protection. Results of a passive serum transfer experiment using day 5 sera from treated WT and CD19^−/− mice demonstrated a mild protective effect with WT sera (33% survival) in contrast with CD19^−/− sera (0% survival), although this did not reach significance (Supplementary Fig. S12). This result may have been due to the relatively short half-life of IgM. We therefore obtained μS^−/− mice that lack secreted IgM. As shown in Fig. 7A, WT littermates treated with MPL/TDCM had high levels of Ig deposited on TA3-Ha cells 5 days post challenge relative to nontreated mice. In contrast, MPL/TDCM treatment did not increase Ig deposition on tumor cells recovered from μS^−/− mice. Along with this, the frequency of CD138⁺ (Ab-producing) B cells were significantly increased (4-fold) in WT(μS^+/+) but not in μS^−/− littermates, following MPL/TDCM treatment (Fig. 7B). Tumor cells recovered from WT littermates treated with MPL/TDCM had significantly increased C3 deposition relative to nontreated mice (Fig. 7C). However, MPL/TDCM treatment in μS^−/− mice did not induce C3 deposition on TA3-Ha cells. Thus, secreted IgM is required for increased C3 deposition on tumor cells in response to MPL/TDCM treatment.

Finally, we assessed the extent to which secreted IgM was involved in MPL/TDCM-mediated protection against TA3-Ha tumor challenge. Five days post treatment, WT mice had a significant reduction (>20-fold) in the frequency of TA3-Ha cells present in the peritoneal cavity relative to untreated mice (Fig. 7D). Relative to these results, MPL/TDCM treatment had a modest effect in reducing the tumor burden in μS^−/− mice. Consistent with these findings, MPL/TDCM treatment protected WT mice during TA3-Ha challenge, but did not induce significant protection in μS^−/− mice (Fig. 7E). Thus, MPL/TDCM-mediated protection against TA3-Ha cell peritoneal carcinomatosis is mediated by secretion of tumor-reactive IgM by B-1 cells.

The current study reveals an unexpected role for B-1 cells in the protective antitumor response stimulated by PAMPs. Intraperitoneal injection of the TLR4 agonist, MPL, combined with the Mincle/MCL agonist, TDCM, significantly reduced tumor growth, ascites development, and morbidity and mortality associated with peritoneal carcinomatosis and lymphomatosis. Treatment resulted in a significant increase in peritoneal B cells, including antibody-secreting cells, as well as a significant increase (10–20-fold) in tumor-reactive IgM. Importantly, B-1a cells, but not splenic or peritoneal B-2 cells, were able to reconstitute protection in B-cell–deficient mice. B-1 cells rapidly secrete high levels of natural IgM in response to PAMPs (34) and it is known that some of these antibodies are reactive with carbohydrate antigens expressed by tumor cells (22). Consistent with this, MPL/TDCM treatment significantly increased IgM reactivity with tumor cells known to express TACAs. The elicited IgM was critical for C3 complement deposition on tumor cells, and both secreted IgM and the classical complement pathway were essential for MPL/TDCM-induced antitumor protection. Collectively, our results highlight a major role for B-1a cell-produced IgM in combating peritoneal metastases.

The role of B cells in cancer is only beginning to be understood (35, 36). Early studies using mouse models initially suggested B cells inhibit T-cell antitumor responses and thus, promote tumor growth (36). However, some studies using mouse models support a protective role for B cells in cancer (35, 36). Along with these data are divergent reports regarding the protumorigenic and antitumorigenic functions of B-cell–produced antibodies in cancer (35). Nonetheless, in human patients, the presence of tumor infiltrating B cells often correlates with improved survival (35, 36). B-1 cells also appear to play dual roles in cancer. Through the production of tumor-reactive natural IgM, B-1 cells likely play a significant role in immune surveillance. Increased production of these antibodies, through nonantigen-specific activation signals, may promote direct tumor cell killing as well as promote the induction of adaptive antitumor responses and memory, as supported by our EL4 rechallenge results. Through the production of antigen-induced TACA-reactive antibodies, B-1b cells can also participate in the protective adaptive humoral response to tumors (22). Nonetheless, due to their production of IL10 and other tumor-promoting soluble factors (37), B-1a (and potentially B-1b) cells may also exert a strong protumorigenic role in cancer. Our current study illustrates that the role of B-1 cells in cancer may be in fact, dynamic and context dependent. Indeed, our findings raise the possibility that PAMPs may skew protumorigenic B-1 cells toward potent drivers of antitumor immune responses.

Many human patients with cancer generate tumor-reactive antibody responses. These antibodies are often germ-line encoded and largely belong to the IgM subclass (38). Included within these specificities are TACAs found on glycoproteins and glycolipids, such as Tn, sialyl Tn, TF, LeY, GM2, GD2, and GD3 (38–40). Natural antibodies reactive with TACA have demonstrated oncolytic effects (38, 41). Additional studies indicate the production of TACA-specific antibodies are associated with increased survival rates in patients (42–45) and in mice (22). Although the breadth and fine specificity of PAMP-elicited IgM is not yet elucidated, our results indicate mucin-related TACAs expressed by BSM (sTn, Tn, TF, sTF; refs. 46, 47), and shared by the epiglycanin mucin that contributes to the aggressive nature of TA3-Ha cells (31, 32), represent at least a portion of the TA3-Ha-, but not EL4-, reactive IgM elicited by treatment. EL4 cells, in contrast, express high concentrations of the tumor-associated glycolipid, ganglioside D2 (GD2; ref. 33). Similar to the results described herein, an IgM mAb specific for GD2 works in concert with complement to protect mice against EL4 tumor growth and metastases (33). More detailed analyses of the tumor-protective IgM reactivities elicited by PAMPs are certainly warranted. Research into the efficacy of tumor-reactive IgM and complement in penetrating and eliciting efficacy against solid tumor masses is also needed.

Although we have determined PAMP-elicited tumor protection requires rapidly secreted IgM and classical complement activation, it is not yet clear what precise mechanism(s) are involved in tumor cell killing. It seems plausible that complement mediated cytotoxicity (CDC) plays a role in tumor cell death given the evidence of C3 deposition as well as the finding that treated A/J mice (naturally impaired in CDC due to C5 deficiency) succumb to tumor challenge much more rapidly than CAF₁ mice. Nonetheless, it has previously been suggested that mucin-bearing tumors are resistant to this mechanism of killing (48). Importantly, expression of cell surface complement regulators may further limit CDC, yet still allow complement mediated killing by IgM (33). Under these circumstances, complement dependent cellular cytotoxicity (CDCC) may play a role in eliciting killing via recognition of complement fragments by effector cells, such as NK cells and macrophages (49). Although complement has been shown to have tumor-promoting potential (49), its role in PAMP-elicited protection against peritoneal carcinomatosis supports that, similar to the divergent roles described for B cells in cancer, complement can be tumor-promoting or protective. Further studies are required to determine the mechanisms by which complement and IgM eradicate tumor cells in the unique environment of the peritoneal cavity.

Using PAMPs to activate B-1 cells to secrete tumor-reactive natural IgM may offer yet another therapeutic strategy that when combined with other immune-activating agents, improve patient outcomes. Bacterial products as well as other PAMPs have previously been demonstrated to elicit antitumor effects in mouse tumor models as well as in patients with cancer (18). However, in most cases, the mechanisms involved have not been clearly elucidated. Thus, our novel findings specifically demonstrating that the TLR and CLR agonist combination of MPL and a cord factor analog elicit innate-like B-cell–dependent antitumor protection in the unique environment of the peritoneal cavity may stimulate additional research aimed at developing new therapeutic strategies for prevention and treatment of peritoneal metastases, and perhaps other TACA-expressing cancers in distinct anatomical locations. In order for such opportunities to be realized, it is critical that the mechanisms by which PAMPs elicit antitumor killing are elucidated. Thus, it remains to be established whether MPL and/or TDCM directly stimulate B-1 cells to produce tumor reactive IgM, and whether these agonists yield synergistic signals through TLR4 and Mincle/MCL, as has been described for other TLR/CLR agonist pairings (50). Moreover, whether CDC or CDCC is a primary mechanism of tumor cell killing, and whether other effector or accessory cells are involved remains to be determined. It is our hope that our study on mouse B-1a cells will offer insights into elicitation of protective TACA-reactive IgM in humans. Nonetheless, as there is much to learn about the newly characterized human B-1 population, additional work on human TACA-reactive B cells is necessary before therapeutic approaches specifically aimed at activating natural IgM production and potentially other tumor-promoting effects can be designed for human patients.

No potential conflicts of interest were disclosed.

Conception and design: M.A. Haro, K.M. Haas

Development of methodology: M.A. Haro, K.M. Haas

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.M. Dyevoich, J.P. Phipps

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.A. Haro, A.M. Dyevoich, J.P. Phipps, K.M. Haas

Writing, review, and/or revision of the manuscript: M.A. Haro, A.M. Dyevoich, K.M. Haas

Study supervision: K.M. Haas

This work was supported by DOD grant W81XWH-15-1-0585 and an American Cancer Society research scholar grant (RSG-12-170-01-LIB) awarded to K.M. Haas. Shared resources support was provided by NCI-CCSG grant P30CA012197 awarded to B.C. Pasche. M.A. Haro was supported by NIH/NCI F31CA183567 and A.M. Dyevoich was supported in part by National Institutes of Health Training Grant AI007401 awarded to M.A. Alexander-Miller.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.