The Role of IgM Antibodies in T Cell Lymphoma Protection in a Novel Model Resembling Anaplastic Large Cell Lymphoma

We previously demonstrated that anti-dsDNA IgM Abs are protective against lupus nephritis in MRL/lpr mice, a model of systemic lupus erythematosus (1, 2). In that study, we examined the specific role of IgM autoantibodies by comparing the phenotype of activation-induced deaminase (AID)–deficient MRL/lpr mice, secreting only IgM, to AID-deficient MRL/lpr mice lacking any secreted Abs. This was accomplished by generating MRL/lpr mice (mS.AID^–/–.MRL/lpr) with a defect in the secreted exon of IgM that are also deficient in AID, a molecule responsible for the generation of high-affinity, isotype-switched Abs, such as IgG (3). Lupus nephritis was abrogated in both groups. However, despite the lack of any autoimmunity-associated pathology, μS.AID^–/–MRL/lpr mice still experienced significantly higher mortality compared with the IgM-secreting mice. In this article, we report that the higher mortality observed in Ab-lacking, μS.AID^–/–.MRL/lpr mice is the result of an increased incidence and severity of T cell lymphoma. The development of lymphoma in MRL/lpr mice is likely associated with the defect in the FAS pathway underscoring the lpr mutation (4–7).

Fas apoptotic pathway defects result in massive expansion of TCR/CD3⁺CD4⁻CD8⁻B220⁺ “double-negative” (DN) T cells in older mice, with clonal expansion of B cell populations (8, 9). In addition, older lpr mice in the C3H background progressed to lymphoma, mostly of B cell origin, and in humans, FAS-mediated apoptosis defects yield both B and T cell lymphomas (4, 10–12). Thus, it is likely that the lpr mutation predisposed these mice to lymphoma that only became apparent when the lupus nephritis was abrogated because of AID deficiency. However, because the lpr defect was present in both strains, this does not explain the increased severity and incidence of lymphoma in the Ab-deficient μS.AID^–/–.MRL/lpr mice, relative to IgM only–secreting (AID^–/– MRL/lpr) mice. This therefore suggests that autoreactive IgM Abs also protect against T cell lymphoma.

Several studies have demonstrated a significant association between circulating autoantibodies and a variety of cancers (13–17). However, the role of autoantibodies in cancer is controversial, with studies suggesting functions ranging from protective to pathogenic and even involved in severe disease. In this study, we demonstrate that in the absence of secreted Abs, μS.AID^–/–.MRL/lpr mice experienced accelerated onset and increased severity of T cell lymphoma that can be mitigated with reconstitution of IgM-secreting B cells. We also show that treating lymphoma-prone p53 mice with anti-nuclear IgM Abs prevented development of high-grade T cell lymphoma and significantly improved survival. Finally, we demonstrate that the lymphoma in these mice originated from DN B220⁺ T cells (i.e., CD3⁺B220⁺CD4⁻CD8a⁻) that are often expanded in MRl/lpr mice and that the phenotype of the lymphoma cells resembled anaplastic large cell lymphoma (ALCL) cells in humans, both histologically and phenotypically with respect to expression of ALCL markers.

C57BL/6 and MRL/lpr strains were purchased from The Jackson Laboratory. The AID^–/–MRL/lpr strain was developed by crossing AID deficiency onto the MRL/lpr background for at least 15 generations, as described (1). B6;129S4-Igh-6^tm1Che/J (The Jackson Laboratory, stock no. 003751) mice were bred with AID wild-type (WT) or AID-deficient MRL/lpr strains to generate µS^–/–MRL/lpr or µS^–/–AID^–/–MRL/lpr strains, respectively, also as described. The strains with at least 99.8% of the MRL/lpr background were used in the experiments. The B6.129S2-Trp53^tm1Tyj/J strain (shortened in this article as P53^–/–B6) was provided by The Jackson Laboratory (stock no. 002101). All strains were maintained in specific pathogen-free conditions in the animal facility of the National Institute of Environmental Health Sciences, National Institutes of Health and the La Jolla Institute for Allergy and Immunology, La Jolla, California. Mouse blood was collected by cardiac puncture. Serum was prepared and stored at −20°C until use. Monoclonal autoantibodies were prepared as described previously (2). TACSAnnexin V Kit was purchased from Trevigen (Gaithersburg, MD). Polyclonal anti-CD3, anti–Pax-5, and anti-myeloperoxidase Abs as well as monoclonal anti-CD45 and anti-F4/80 Abs were purchased from Abcam (Cambridge, MA), Santa Cruz Biotechnology (Santa Cruz, CA), Dakocytomation (Carpinteria, CA), GeneTex (San Antonio, TX), and BioLegend (San Diego, CA), respectively. The OmniMap Anti-Rabbit DAB Detection and Vectastain Elite ABC Kits were provided by Ventana Medical Systems (Tucson, AZ) and Vector Laboratories (Burlingame, CA), respectively.

H&E-stained tissue sections were prepared following standard protocols. T cell lymphomas were graded by pathologists based on their degree of differentiation and systemic dissemination using the following criteria. For grade 1, >75% of the cells are well differentiated and minimally abnormal; lymphoma mainly involves spleen, lymph nodes, and/or bone marrow; and mitotic index is low. For grade 2, ?50–75% of cells are mildly abnormal and moderately differentiated cells; lymphoma may spread beyond spleen, lymph nodes, and/or bone marrow; and mitotic index is low. For grade 3, ?50–75% of cells are poorly differentiated and moderately abnormal; lymphoma may spread beyond spleen, lymph nodes, and/or bone marrow; and mitotic index may be increased. And for grade 4, >75% of cells are undifferentiated and markedly abnormal; lymphoma has spread beyond spleen, lymph nodes, and/or bone marrow; and mitotic index is increased. All other lesions, including lymphoid dysplasia, were graded using a standard grading scheme: 1 (minimal), 2 (mild), 3 (moderate), and 4 (marked).

Slides were stained for CD3 on the Discovery XT. Standard CC1 retrieval was performed. The primary Ab (anti-CD3; Abcam, Cambridge, MA) was applied for 60 min at a dilution of 1:250 without heat. Normalized rabbit serum at the equivalent dilution was applied to the negative controls in lieu of primary Ab. The anti-rabbit HRP polymer label was applied for 16 min. The slides were treated with hematoxylin for 8 min, then incubated in bluing reagent for 4 min. For Pax-5 stain, slides were deparaffinized, hydrated, and then quenched in 3% H₂O₂ for 15 min to block endogenous peroxidase activity. Heat-induced epitope retrieval was done using Decloaker with 1× Citrate Buffer for 5 min at 120°C. Tissues underwent protein blocking with 10% normal horse serum followed by endogenous biotin blocking using the avidin/biotin blocking kit for 15 min. This was followed by a 60-min incubation with the primary Ab (anti–Pax-5; Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:500. Normalized negative normal goat serum was used for the negative control slides. A 1:500 horse anti-goat secondary incubation was performed for 30 min. The streptavidin peroxidase label was applied for 30 min. Liquid Dako DAB Chromagen was applied to the slides for 6 min under cover followed by a 20-s counterstain with hematoxylin. For F4/80, CD45, and myeloperoxidase stains, similar protocols were used with dilutions of primary Abs at 1:100, 1:250, and 1:1200, respectively (anti-myeloperoxidase, anti-CD45, and anti-F4/80 Abs were purchased from Dakocytomation, GeneTex, and BioLegend, respectively).

Mice were sacrificed via CO₂ inhalation, and their lymph nodes, spleens, thymii, and livers were harvested and fixed in zinc formalin for 24 h at room temperature (RT). Femurs and tibias were cut near the joints, and bones were spun down in a centrifuge tube to release bone marrow cells. Samples were kept in 70% isopropanol until processing on an automated tissue processor (TissueTek VIP150; Sakura) using a graded series of isopropanol and Pro-Par clearant (no. 510; Anatech). Samples embedded in paraffin blocks were sectioned at 4 µm, adhered to positively charged slides, and dried overnight. Sections were dewaxed with Pro-Par, rehydrated, and subjected to Ag retrieval in sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6.0 in ultrapure water). The slides were placed into room-temperature buffer and heated in an electric pressure cooker (no. CPC-600; Cuisinart) at a high setting for 20 min then removed and cooled at RT for 30 min. The slides were then washed twice in PBS-T wash buffer (0.1% Tween 20, no. 10113103; Thermo Fisher Scientific) in PBS (no. P3813; MilliporeSigma). Blocking buffer (5% normal donkey serum and 0.3% Triton X-100 in PBS) was applied to each slide and incubated for 1 h at RT. The primary Abs were applied, including CD3 (no. 17617-1-AP, 1:200 in PBS-T; Proteintech), B220 (clone RA3-6B2, no. 14-0452-82, 1:500 in PBS-T; Thermo Fisher Scientific), CD4 (clone 4SM95, no. 14-9766-82, 1:100 in PBS-T; Thermo Fisher Scientific), and CD8α (clone 4SM15, no. 14-0808-82, 1:200 in PBS-T; Thermo Fisher Scientific) and incubated overnight at 4°C. The primary Abs were detected with either anti-Rabbit Alexa Fluor Plus 555 (no. A32732, 1:500 in PBS-T; Thermo Fisher Scientific) or anti-Rat Alexa Fluor 568 (no. A11077, 1:500 in PBS-T; Thermo Fisher Scientific). The slides were then washed four times with 1 ml of PBS followed by counterstaining with Hoechst (no. H3570, 1:1000 in PBS; Thermo Fisher Scientific) for 5 min. The Hoechst was washed off with 1 ml of PBS three times, and the slides were removed from the Freequenza Rack and treated with TrueBlack Lipofuscin Autofluorescence Quencher (no. 23007, 1:20 in 70% reagent alcohol; Biotium) for 30 s before being washed three times in PBS in a Coplin jar. Finally, no. 1.5 coverslips were applied to slides using ProLong Gold Antifade Mountant (no. P36930; Thermo Fisher Scientific), and slides were scanned with a ZEISS AxioScan Z1 slide scanner using a 20×/0.8NA objective.

Spleens and lymph nodes were passed through 70-µm filters (no. DGN258368; BioPioneer) into staining buffer (PBS with 2% FBS, 2 mM EDTA, and 0.1% sodium azide). Cells were centrifuged for 3 min at 500 × g at 4°C, resuspended in staining buffer, and counted using Vi-CELL counter (Beckman Coulter). Cells (4 × 10⁶) were added to a 96-well plate, resuspended in Fc block to reduce nonspecific binding, and incubated with Zombie NIR (no. 77184; BioLegend) viability dye in PBS for 10 min at 4°C. Cells were stained with combinations of fluorescently labeled anti-mouse Abs against CD4 (GK1.5), CD3ε (145-2C11), TCRβ (H57-597), CD45R (RA3-6B2), CD8α (53-6.7), CD45 (30-F11), and CD30 (mCD30.1) (Thermo Fisher Scientific and BioLegend). Cells were preincubated for 15 min at 37°C with CD30, after which the remaining Abs were added. Cells were stained for 30 min at 4°C, washed twice in staining buffer, and fixed in 4% paraformaldehyde. Flow cytometry data were acquired on a BD FACSCelesta flow cytometer (BD Biosciences) and were analyzed with FlowJo software (Tree Star, Ashland, OR) after eliminating dead and doublet cells. Frequencies of normal T cells and lymphoma cells within the parent gate and live cells were calculated in FlowJo.

Cell apoptosis was examined with TACS Annexin V kit by following the manufacturer’s instructions. Briefly, cells were harvested and washed once with PBS. Cells were stained in the dark in binding buffer containing propidium iodide and annexin V-FITC for 15 min at RT. After adding 0.4 ml of binding buffer, the cells were examined with an LSR II flow cytometer (BD Biosciences), and data were analyzed using FlowJo software (Tree Star).

TK-1 cells were plated into 48-well plates at 2 × 10⁵ cells/500 µl per well and pretreated with DNase I (Sigma-Aldrich) at 5, 50, or 100 µg/ml at 37°C for 1 h. MRL/lpr sera and anti-nuclear IgM monoclonal Abs were pretreated with DNase I at 5, 50, or 100 µg/ml at 37°C for 1 h. After another 20-h incubation step at 37°C, cells were harvested for staining apoptosis, as described above.

IgM levels in blood were determined with a commercial ELISA kit (Bethyl Laboratories, Montgomery, TX), using manufacturer instructions. Briefly, Costar 96-Well EIA/RIA Plates (Corning Inc, Tewksbury, MA) were coated with goat anti-mouse IgM in PBS at 100 µl/well at 4°C overnight, followed by incubation at RT for 1 h in blocking buffer (PBS, pH7.4, with 1% BSA) at 200 µl/well. Mouse sera were diluted with sample diluent (blocking buffer plus 0.05% Tween 20) at 1:20 and added to wells at 100 µl/well. Goat anti-mouse IgM-HRP conjugator (Bethyl) was used at 1:100,000 and incubated for 1 h. TMB enzyme substrate (KPL, Gaithersburg, MD) was added at 100 µl/well and incubated for 30 min at RT. The reaction was stopped by adding 100 µl of 1 M H₂SO₄.The absorbance at 450 nm was measured in a Multiskan Ascent microplate reader (Thermo Electron, Finland).

Bone marrow was collected from AID^–/–MRL/lpr and µS^–/–AID^–/–MRL/lpr mice by flushing femora and tibiae with PBS. A single-cell suspension was prepared by repeated pipetting up and down. Cells were washed and resuspended in PBS. µS^–/–AID^–/–MRL/lpr mice that had been irradiated at 8 Gy received 2 × 10⁶ bone marrow cells from AID^–/–MRL/lpr mice and µS^–/–AID^–/–MRL/lpr mice (a cell mix at 1:1 ratio) or 1 × 10⁶ bone marrow cells only from µS^–/–AID^–/–MRL/lpr mice. Ten weeks later, blood was collected for quantifying circulating IgM by ELISA, and mice were euthanized for necropsy.

Anti-nuclear IgM monoclonal Abs were prepared as previously described. A mixture of several IgM Abs (four anti-dsDNA and one anti-Smith) was quantified, and its concentration was adjusted to 1 mg/ml. The Abs were aliquoted and stored at −20°C until ready for use. p53^–/–B6 mice, all male, 11–12 wk old, were divided into two groups. One group (n = 10) was injected via the i.p. route with the IgM Abs at 100 µg/mouse, twice weekly, for 3 mo while another group (n = 9) served as a control. During the experimental period, the mice were closely monitored. Any dead mice were recorded and their tissues fixed in 10% neutral-buffered formalin. At the end of the experiment, all remaining mice were euthanized for histological examination. Lymphomas were evaluated using the grading criteria, as described above.

The Mann–Whitney rank sum test was used for most statistical analyses. The Student t test was used in which normal distributions and equal variances were found. Log-rank (Mantel–Cox) test was used to compare the death curves in the survival experiment, and p values <0.05 were considered significant.

In a previous study, we found that MRL/lpr mice lacking Abs from a combined defect in AID and IgM secretion (μS.AID^–/–.MRL/lpr mice) had much higher mortality than MRL/lpr mice with AID deficiency but normal IgM secretion, despite both groups experiencing a complete abrogation of lupus nephritis (Supplemental Fig. 1). Necropsies of mice euthanized at 6 and 12 mo of age unequivocally demonstrated that the cause of death in both groups was lymphoma (Fig. 1). Blood cell counts revealed an increase in WBCs, and immunohistochemical analysis suggested the lymphoma was of lymphocytic origin (data not shown and Supplemental Fig. 2A. Lymphomas of much earlier onset and of significantly higher grade were seen in μS.AID^–/– MRL/lpr mice (without secreted Abs), compared with MRL/lpr mice capable of secreting all Ig class Abs (WT) and those secreting only IgM (AID^–/–) (Fig. 1). Indeed, at 6 mo of age, only 24% of AID-deficient MRL/lpr mice with secreted IgM developed lymphoma, compared with 100% of μS.AID^–/–.MRL/lpr mice. There was also a difference in severity at 6 mo of age, wherein over half of the mice lacking secreted Abs (but no Ab-secreting mice) had T cell lymphoma of at least grade 3 (Fig. 1A). By 12 mo of age, the incidence in Ab-secreting mice increased to 46%, most with grade 2 or 3 lymphoma, whereas all mice in the Ab-deficient group had advanced lymphomas of at least grades 3 and 4, with significant involvement of nonlymphoid organs (Fig. 1B). Cellular morphology of tumors suggested a lymphocytic or myeloid origin (Supplemental Fig. 2A), prompting us to more definitively establish the cellular origin of the tumors. As an initial screen, we performed immunohistochemistry and immunofluorescence, using Abs recognizing Ags on the surface of B cells, T cells, myeloid cells, or macrophages. Without exception, all tumors were CD3⁺, indicating a lymphoma of T cell origin (Table I) and suggesting the culprit cell population may be the enigmatic CD3⁺CD4⁻CD8⁻ (but B220⁺) expanded population of T cells observed in MRL/lpr mice (i.e., DN T cells) (8). Immunohistochemical and flow cytometric analysis of the lymphoma-derived cells confirmed a B220⁺CD3⁺CD4^low/negCD8^neg phenotype, consistent with this population (Supplemental Figs. 2B, 3) and, further, were CD19⁻, ruling out a B cell origin (Supplemental Fig. 3). In addition, these DN lymphoma T cells expressed slightly higher levels of CD30 than normal T cells (Supplemental Fig. 3). We could not detect significant anaplastic lymphoma kinase (ALK) expression in the lymphoma cells (data not shown).

The difference between the two MRL/lpr strains in this study is whether they have or lack the ability to secrete Abs. µS^–/–AID^–/–MRL/lpr have B cells but cannot secrete any Abs, whereas AID^–/–MRL/lpr secrete high levels of IgM (but no other isotypes). In fact, the former strain was derived from the latter strain, suggesting it is the absence of IgM Abs (or any kind of Abs) that leads to increased mortality and severe lymphoma. To more formally test the hypothesis that Abs from AID^–/–MRL/lpr mice B cells protected from development of severe T cell lymphoma, we reconstituted irradiated μS.AID^–/–.MRL/lpr mice with B cells from AID-deficient MRL/lpr mice and, as a control, with B cells from littermate μS.AID^–/–.MRL/lpr mice, which lack secreted Abs (Fig. 2A). Close to 50% of μS.AID^–/–.MRL/lpr mice attained high levels of circulating IgM following reconstitution with IgM-secreting B cells (Fig. 2B). Consistent with our hypothesis, none of the mice with high levels of circulating IgM developed T cell lymphoma, whereas the majority of those with little or no circulating IgM developed significant lymphoid dysplasia or lymphoma at rates similar to μS.AID^–/– MRL/lpr mice reconstituted with nonsecreting B cells (Fig. 2C). These results indicate IgM-secreting B cells protect AID^–/–MRL/lpr mice from developing severe T cell lymphoma.

To probe whether sera from these mice can directly or indirectly trigger apoptosis of T cell lymphoma cells, we incubated sera from AID-deficient or AID WT MRL/lpr mice with TK-1 cells, a T cell lymphoma murine cell line (18). As additional controls, serum was also incubated with normal splenic T cells or with cells from a murine leukemic B cell line, L1210 (19). Sera from MRL/lpr mice induced significant apoptosis of TK-1 cells, regardless of AID status, whereas sera from non-autoimmune mice or μS.AID^–/–MRL/lpr mice (without Abs) failed to induce apoptosis significantly (Fig. 3A, 3B, data not shown). In addition, none of the MRL/lpr sera treatments with normal murine splenic T cells significantly induced apoptosis above background (Supplemental Fig. 4A). Also of note, when normal thymocytes were induced to undergo apoptosis with dexamethasone, the protective purified IgM bound the apoptotic cells at high levels (Supplemental Fig. 4B). Interestingly, only serum from AID WT MRL/lpr mice induced high-level apoptosis in the L1210 B cell leukemic cell line (L1210, Supplemental Fig. 4C). Indeed, sera from AID-deficient MRL/lpr mice efficiently killed the T cell but not the B cell lymphoma cells. Collectively, these results suggest that isotype-switched autoantibodies were required to kill B cell lymphoma cells, whereas IgM autoantibodies were equally effective against the T cell lymphoma cell line, both while sparing normal nonapoptotic lymphocytes.

AID-deficient MRL/lpr mice have high levels of anti-nuclear IgM Abs (1). Given that killing was only observed with sera from autoimmune mice, we speculated that autoantibodies commonly found in MRL/lpr mice may contribute to tumor cell killing. Therefore, we examined whether anti-nuclear IgM Abs can trigger apoptosis of the TK-1 line and treated tumor cells with a mixture of five IgM monoclonal Abs: four with anti-dsDNA specificity and one with an anti-Smith IgM specificity. Significantly higher killing of T cell lymphoma cells was obtained with this anti-nuclear autoantibody mixture, when compared with a non-autoreactive IgM control (Fig. 4). Combined, these results strongly implicate autoreactive IgM Abs as major contributors to delayed onset and decreased severity of T cell lymphoma seen in AID-MRL/lpr mice, relative to μS.AID^–/–MRL/lpr mice. Given that non-autoimmune serum lacked such protective activity, it also suggests the possibility that B cells capable of targeting and killing certain cancer cells may be perceived as autoreactive and thus routinely depleted from the repertoire.

Given that only sera from autoimmune mice induced tumor cell killing and, in particular, incubation with anti-dsDNA IgM Abs also caused significant apoptosis of TK-1 cells, we tested the possibility that DNA bound to cell surfaces or in complex with Abs were required for this activity. TK-1 cells were pretreated with DNAse1 at various concentrations, then incubated with either sera from MRL/lpr mice or anti-dsDNA IgM Abs that had also been pretreated with DNase I. Importantly, the concentration of DNase I selected is within the range previously shown to disrupt DNA/Ig complexes (20). We found that regardless of concentration, DNase I pretreatment of tumor cells, monoclonal anti-dsDNA IgM Abs, or sera from MRL/lpr mice did not impact tumor cell apoptosis. These data suggest that DNA, either bound to the tumor cell surface or in complex with anti-dsDNA Abs, is not required for the antitumor activity of autoreactive Abs (Fig. 5).

Most mice homozygous for a defect in the tumor suppressor protein p53 develop a severe form of T cell lymphoma within the first 6 mo of life (21). If, in fact, autoreactive IgM Abs derived from AID-deficient MRL/lpr mice can also kill T cell lymphoma cells in vivo, one would expect that passive transfer of autoreactive IgM Abs into young asymptomatic p53-deficient mice would improve survival. Indeed, p53 mice receiving the autoreactive IgM mixture experienced a dramatic increase in survival compared with p53 mice receiving only PBS as control (Fig. 6A). Among mice still alive at 6 mo of age, most controls had developed stage 2 or 3 lymphoma, whereas most receiving the IgM Abs from AID-deficient MRL/lpr mice developed grade 1 lymphoma (Fig. 6B). The lymphomas were likely T cell lymphomas composed of a uniform lymphoblast population characteristic of thymic (T cell) lymphomas. A thymic origin for lymphomas in mice from this study would be consistent with that previously reported in a tumor spectrum analysis of p53-mutant mice (21). It is remarkable that these IgM monoclonal Abs isolated from MRL/lpr mice protected p53-deficient Bl6 mice.

Pathological examination of the lymphomas in IgM-treated and control mice revealed that p53 knockout mice treated with protective IgM autoantibodies not only had a significant delay of lymphoma onset and exhibited improved survival and the lymphomas tended to be lower grade (Fig. 5) but also had notable lymphoid depletion associated with tingible-body macrophages typical of phagocytosis of apoptotic bodies (Fig. 6C, Table II) in the thymus, the splenic white pulp, and/or mesenteric lymph nodes. Among mice in which this information was obtained, pathologists scored the presence and degree of lymphoid depletion (1–4, reflecting increasing depletion). Among those receiving the IgM, 7 out of 10 mice had evidence of depletion (Table II) and displayed high-grade (grade 3 or 4) depletion and tingible-body macrophages in at least two lymphoid organs (Fig. 6C, Table II). Among those that did not receive the protective IgM, two out of seven mice had notable lymphoid apoptosis but only in one tissue (mesenteric lymph nodes) and the depletion was of lower grade. These results suggest that ongoing killing of tumor cells helps control the lymphoma, and these Abs may protect by inducing apoptosis of lymphoma cells.

MRL/lpr mice typically succumb to glomerulonephritis or other complications of their lupus-like syndrome within the first year of life (22). In examining the role of Abs in lupus nephritis, we developed MRL/lpr mice that were also deficient in AID, a molecule critical for the generation of high-affinity, isotype-switched Abs. These mice secrete only IgM Abs, with particularly high levels of anti-nuclear IgM Abs, yet experience a dramatic improvement in survival. When compared with Ab-deficient MRL/lpr (μS.AID^–/–.MRL/lpr) mice (with an additional defect in the gene encoding the secretory exon of IgM), we were surprised to find that despite both groups experiencing substantial improvements in survival and lupus nephritis, Ab-deficient mice still experienced higher mortality when compared with mice secreting only IgM (2). In this study, we report that increased mortality in the Ab-deficient group was due to a severe form of T cell lymphoma.

T cell lymphoma can be associated with defects in the FAS apoptosis pathway, which is crucial for regulation of immune cells of all types, including lymphocytes (4, 5, 23, 24). MRL/lpr mice likely have a defect in the FAS pathway (the lpr mutation) that accelerates development of the lupus-like syndrome. In this study, most mice, regardless of Ab status, eventually progressed to lymphoma with earlier onset, with greater severity in the strain lacking IgM Abs. These results suggest that MRL/lpr mice would normally develop T cell lymphoma, if able to survive the nephritis associated with the lupus-like syndrome (as in the case for AID-deficient MRL/lpr mice). However, the absence of Abs accelerated this process dramatically, suggesting that IgM Abs, such as those secreted by AID-deficient MRL/lpr mice, ameliorated T cell lymphoma. Adoptive transfer of B cells capable of secreting IgM from AID-deficient MRL/lpr into Ab-deficient MRL/lpr mice resulted in a significant reduction in severity of T cell lymphoma among reconstituted mice with high levels of circulating IgM. These results isolated the protective factor in AID-deficient MRL/lpr mice to B cells and their product, Abs. In addition, sera from AID-deficient MRL/lpr mice or anti-dsDNA IgM Abs from these mice were associated with a highly significant increase in apoptosis of a murine T cell lymphoma cell line. The killing was specific to tumor cells, as apoptosis of normal splenic T cells was not enhanced by incubation with these reagents. Neither non-autoreactive IgM nor sera from non-autoimmune C57BL/6 mice were associated with increased apoptosis of tumor cells. Strikingly, passive transfer of autoreactive IgM Abs (mostly anti-dsDNA specific) into p53-deficient mice significantly improved survival and mitigated T cell lymphoma severity. Combined, these results directly implicate autoreactive Abs, particularly anti-dsDNA IgM Abs, as being protective against T cell lymphoma. It will be important to establish the mechanism of Ab-mediated T cell lymphoma killing, but it will likely involve binding of a cell surface Ag uniquely expressed or exposed in tumor cells. It also possible that binding to lymphoma T cells may happen through the IgM Fc receptors. IgM has three Fc receptors, including Fcα/μ receptor (Fcα/μR), polymeric Ig receptor (pIgR), and Fcμ receptor (FcμR); thus, carefully controlled blocking of all these receptors will be needed to determine if this is the case. However, because the main difference between non-autoimmune IgM and autoimmune IgM lies in the V region, we think the protective effect is primarily driven by recognition of a receptor through the Ag-binding pocket generated by the IgM pentamer, and current studies are addressing these possibilities. Finally, it is interesting that anti-dsDNA IgM specificity is associated with protection in this study, suggesting the possibility that TLR may be indirectly involved, given their role in the response to nucleic acid binding, inflammation, and cancer (25–27).

DNase 1 pretreatment of either tumor cells or anti-dsDNA IgM Abs did not disrupt Ab-mediated tumor cell killing, suggesting DNA on cell surfaces or as part of immune complexes is not required. This suggests that the anti-dsDNA specificity correlates to another specificity for these Abs that could, for example, target an apoptotic pathway in tumor cells. In fact, p53-deficient mice treated with protective IgM had strong evidence of lymphoid depletion through apoptosis in multiple lymphoid organs, including the thymus. In addition, although IgM-mediated killing of T cell lymphoma cells was not seen with normal thymocytes, normal T cells induced to apoptose were readily bound by the protective IgM Abs. Combined, these data point to an apoptosis-related receptor on tumor T cells as the target of the protective Abs, and ongoing studies are examining this possibility. Regardless of mechanism, it is apparent that the Ab-mediated protection eventually fails as even MRL/lpr mice, regardless of AID status, progressed to lymphoma. This might reflect the possibility that, as the lymphoma cells become less differentiated, they lose expression of a putative tumor Ag(s) these Abs recognize, essentially becoming invisible to the humoral immune system.

Phenotypic characterization by immunohistochemistry, immunofluorescence, and flow cytometry of the lymphoma cells in the μS.AID^–/–.MRL/lpr revealed their origin to be DN (CD3⁺CD4^low/negCD8⁻) T cells that are expanded in MRl/lpr mice and naturally found in humans (8, 28–30). Because many of the cells had the “hallmark” eccentric nuclei of a horseshoe or kidney-shape appearance (Supplemental Fig. 2) often seen in ALCL, expression of CD45 (B220), CD30, and ALK was examined. The lymphoma cells expressed high levels of CD45 and moderately elevated levels of CD30 but were negative for ALK expression. These results suggest that the novel μS.AID^–/–.MRL/lpr strain may be a model of the ALK-negative subtype of ALCL (31). These results have two implications: ALCL may originate from the DN T cell population, and autoantibodies may be protective in the development of this lymphoma and may represent a novel therapeutic approach, perhaps involving an enhanced IgM product for Ig i.v. therapy (IvIg).

The relationship between autoantibodies and cancer is well documented but extremely complex. Defects associated with immune deregulation, such as defects in apoptotic pathways, are often directly associated with both disease processes (28, 29). In contrast, B cell immunodeficiency, acquired or inborn, is often independently associated with both cancer susceptibility and autoimmunity (32–35). In cancer, this correlation with immunodeficiency is likely a result of decreased immune surveillance, whereas in autoimmunity, it may result from expansion of residual immune cell populations (some of which may be autoreactive) to combat persistent infections. In addition, many autoimmune disorders are associated with increased susceptibility to later development of neoplasia. This may be directly correlated to organ damage and inflammation associated with an autoimmune process (36–40). For these reasons, it is not unusual to find significant correlations between circulating autoantibodies and malignancy and suggests that the correlation is one of common association rather than cause and effect. In this study, we show direct evidence implicating autoreactive Abs in protection against a hematological malignancy, T cell lymphoma. These results introduce the possibility that the autoantibody correlation with cancer seen in hematological malignancies, such as non-Hodgkin lymphoma and others, may reflect an ongoing immune response to malignant cells by autoreactive Abs (13, 37, 40–42). This may explain why cancer patients often develop autoimmune manifestations that are associated with the presence of autoantibodies, particularly to nuclear Ags such as dsDNA, which in some cases is associated with antitumor effects (43, 44). Interestingly, serum from AID WT MRL/lpr mice was as good or better than serum from AID-deficient MRL/lpr mice at killing tumor cells. Because the main difference between the two strains is the presence of switched Abs, particularly IgG, these results suggest that autoreactive IgG is also capable of protection from lymphoma. However, the mouse experiments described in this article, such as the adoptive transfer of Ab-secreting B cells into μS.AID^–/– MRL/lpr mice and passive transfer of anti-dsDNA Abs into p53^–/– mice, all used IgM and showed highly significant protection. From a therapeutic standpoint, autoreactive IgM would be preferable over IgG, as autoreactive IgG, especially anti-dsDNA IgG, is pathogenic, whereas anti-dsDNA IgM is actually protective (2). Also, given that the IgM Abs studied in this article fit the characteristics of natural Abs (44, 45) (that is, unmutated, due to AID deficiency) of IgM isotype and polyreactive, it may be that these anti-dsDNA IgM Abs are natural protective barriers to neoplasia, and harvesting them for treatment may be an effective novel weapon in the treatment of T cell lymphoma. It is interesting that the lymphoma cells in these mice clearly originate from the expanded DN CD3⁺B220⁺ T cell population seen in MRL/lpr mice. This population of T cells has been described in the appendix and other locations of normal humans and mice and appears to play a regulatory role in autoreactive T cells (46–49). That the lymphoma cells in the Ab-deficient mice morphologically resembled those seen in human ALCL suggests autoreactive IgM as a potential novel treatment for ALCL and μS.AID^–/– MRL/lpr mice as a potential novel mouse model of this lymphoma.

We are grateful to Greg Travlos, Deborah King, Natasha Clayton, Matthew Meyer, and Connie Cummings for assistance with blood chemistry, immunofluorescence, multicolor flow cytometry, and electron microscopy.

The authors have no financial conflicts of interest.