Novel Highly Sensitive IL-10–β-Lactamase Reporter Mouse Reveals Cells of the Innate Immune System as a Substantial Source of IL-10 In Vivo

The host’s immune responses are indispensable for protection against infection and cancer. However, uncontrolled, overexuberant immune responses are associated with adverse effects, leading, for example, to autoimmune diseases and allergy. Thus, a tight regulation of the immune responses is a necessity to prevent overreaction of the immune system and limit host damage. For this reason, the immune system has evolved regulation mechanisms that enable balanced, nonpathological immune responses and tolerate self- or harmless Ags.

IL-10 is a pleiotropic cytokine that has emerged as an important factor that mediates immune tolerance by regulating the functions of lymphoid and myeloid cells (1). However, before the discovery of the TLRs, the research of the adaptive immune system predominated throughout decades over that of innate immunity. Consequently, when the concept of immune regulation to avoid Paul Ehrlich’s predicted horror autotoxicus (autoimmunity) (2) was revisited, cells of the adaptive immune system, mainly T cells, were again in the focus of the new discipline concerning immune regulation and tolerance (3, 4). As IL-10 has been widely implicated as a functional marker for immune regulation and tolerance activities, great attention has been given mainly to CD4 T cell subpopulations producing IL-10 and their role in immune regulation and tolerance in homeostasis and under immune-activating conditions (5).

However, the focus on IL-10–producing T cells has also been partly due to the limitation of experimental tools, such as intracellular immunostaining, that enable detection of IL-10 production, mainly only in T cells. Furthermore, intracellular immunostaining usually requires ex vivo restimulation of the cells, which could lead to artifacts.

A promising tool to approach this problem can be provided by autofluorescent protein (AFP) reporter mice (6). However, AFP reporter mice have major limitations, including the long turnover rate of AFP, such as enhanced GFP (EGFP), and the high AFP amount required for fluorescence detection (7) and thus do not enable kinetic studies nor allow marking of cells weakly expressing the target gene.

In this study, we present a novel approach in engineering valuable, highly sensitive reporter mice that helps to efficiently overcome these limitations, using for the first time, to our knowledge, the reporter gene bla encoding the bacterial enzyme TEM-1 β-lactamase (Bla). The Bla reporter activity can be monitored in living cells using the fluorogenic cephalosporin derivative coumarin-cephalosporin-fluorescein (4)-acetoxymethyl (CCF4-AM) (8). CCF4-AM is a membrane permeant ester derivative that consists of a cephalosporin core bridging a 7-hydroxycoumarin to a fluorescein. The lipophilic, esterified form of CCF4-AM efficiently enters cells. In the cytosol, the ester groups are readily hydrolyzed by cytoplasmic esterases, releasing the negatively charged form, CCF4, which is thus trapped in the cytosol. CCF4 emits green light because of fluorescence resonance energy transfer from the coumarin donor to the fluorescein acceptor upon excitation of coumarin at 409 nm. Cleavage of the β-lactam ring of CCF4 by Bla splits off the fluorescein, disrupts energy transfer, and shifts the emission to blue. Bla production is reflected by its enzymatic activity and can be quantified in living cells by measuring the ratio of blue (product) to green (substrate) fluorescence. This ratiometric signal is minimally affected by variations in cell size, substrate uptake, excitation intensity, and emission sensitivity and thus provides more consistent results (8).

We report in this study on the generation of the IL-10–Bla reporter mouse and verify its efficiency and applicability in comparison with an IL-10–EGFP reporter mouse line. In contrast to the latter, the IL-10–Bla reporter (ITIB) mouse enables us to easily analyze and quantify IL-10 production at the single-cell level in all myeloid and lymphoid cell types. Interestingly, ITIB mice revealed that F4/80-positive macrophages represent the major IL-10 producer population in several organs in the steady state. Furthermore, using ITIB mice, we identified tumor-infiltrating macrophages (CD11b⁺F4/80⁺CD11c⁺) as the major source for IL-10 in B16-F10 melanoma in vivo. Finally, during systemic infection with Yersinia enterocolitica, we show that neutrophils represent the major cell type among IL-10–producing cells at the site of infection. Thus, we conclude that cells of the innate immune system are a substantial source of IL-10 in homeostasis during infection-triggered inflammation and in tumor microenvironments.

The gene-targeting strategies are depicted in Supplemental Figs. 1A and 2A. A 7.1-kb SacII-BamHI genomic fragment containing part of the promoter sequence and all five exons, but not the 3′ untranslated region (UTR) of Il10, was amplified from the vector pTZ18R10 (gift of Prof. Werner Müller, Faculty of Life Sciences, University of Manchester, Manchester, U.K.) and cloned into the SacII-BamHI sites of pBluescript-II-KS (Stratagene). An internal ribosomal entry site (IRES)-eGFP sequence was amplified from pIRES2-EGFP (BD Biosciences Clontech) and cloned into the SalI site, directly after the stop codon of Il10. With NotI and AflII sites introduced at the end of eGFP, a floxed neomycin resistance cassette was amplified from the pEasyFlox vector (M. Alimzhanov, Institute of Genetics, Cologne, Germany) and cloned into the NotI-AflII sites. A 2.2-kb genomic sequence containing the 3′ UTR of exon 5 of Il10 was amplified directly from genomic DNA (isolated from 129Sv mice) and cloned into the AflII/XhoI sites.

To generate the targeting vector for the ITIB mice, the TEM-1 bla gene was amplified from the plasmid pBR322 using the forward primer 5′-CGCCCACAACCATGGGACACCCAGAAACGCTGGTGAAAG-3′ containing a BstXI site and the backward primer 3′-CGCGCGGCCGCTTACCAATGCTTAATCAGTGAGGC-5′ containing a NotI site and cloned in BstXI-NotI sites of the targeting vector instead of eGFP (directly after the IRES2 sequence). Using these primers, a bla fragment was generated that lacks the coding region of the N-terminal signal sequence of bla (which was exchanged for the sequence ATGGGA coding for methionin und glycin) to avoid the secretion of Bla (9). A second IRES2-bla sequence was amplified from this modified targeting vector using the forward primer 5′-CGCGCGGCCGCCCGCGGGCCCGGGATCCG-3′ and the backward primer 3′-CGCCTTAAGTTACCAATGCTTAATCAGTGAGGC-5′ containing NotI and AflII sites, respectively, and cloned into NotI-AflII sites. Thus, the so generated targeting construct contains a tandem of IRES2-bla sequences. To enrich the selection for homologous recombinant embryonic stem (ES) cell clones, a Cre recombinase gene (Cre) was inserted into the XhoI site, outside the 3′ homology arm of the targeting vector, as a negative selection marker against ES cells, which have randomly integrated the targeting vector (10). An AscI site was introduced at the end of Cre that allows linearization of the targeting vector.

The integrity of the targeting vectors was confirmed by sequencing. The IL-10–1×(IRES-eGFP) (ITIG) and ITIB targeting vectors were linearized with XhoI and AscI, respectively, and electroporated into R1 ES cells. G418-resistant ES cell clones were screened by Southern blot, and two to four independently targeted homologous recombinant ES clones were injected into C57BL/6 blastocysts to produce chimeric mice. The targeted mice were confirmed by PCR genotyping (Supplemental Figs. 1B, 2B). The neomycin resistance gene was removed by crossing the chimeric mice to a general Cre-expressing deletor mouse strain (11) (kindly provided by the Institute for Genetics of the University of Cologne). The reporter mice were then backcrossed to C57BL/6 for at least four generations.

The ITIG and ITIB mice are viable, fertile, and show normal survival and no phenotypic defect through as many as seven generations and as late as 2 y of life.

ITIG and ITIB mice express Il10-eGFP and Il10-bla fusion mRNA, respectively, as could be shown by RT-PCR (Supplemental Figs. 1C, 2C). Furthermore, the basal level of Il10 mRNA in wild-type (WT) mice was comparable to that in ITIB/ITIB mice (Supplemental Fig. 2C), indicating that the integrity of Il10 mRNA expression is not affected by insertion of the two IRES-bla sequences upstream of the 3′ UTR of the Il10 gene.

The expression of Il10, Il10-eGFP, and Il10-bla hybrid mRNA was analyzed by RT-PCR using the following primer: for Il10, the forward primer 5′-GCACTGCTATGCTGCCTGCTCTTACTG-3′ (targeting exon1) and backward primer 3′-ATGGCCTTGTAGACACCTTGGTCTTGG-5′ (targeting exon5); for Il10-eGFP hybrid mRNA, the forward primer 5′-GAGCAAGGCAGTGGAGCAGGTG-3′ (targeting exon4 of Il10) and backward primer 3′-CTCCAGCTTGTGCCCCAGGA-5′ (targeting eGFP); and for Il10-bla hybrid mRNA, the forward primer 5′-GCACTGCTATGCTGCCTGCTCTTACTG-3′ (targeting exon1 of Il10) and backward primer 3′-CCCACTCGTGCACCCAACTGATCTTCAGC-5′ (targeting bla). As a loading control, the housekeeping gene hypoxanthine-guanine-phosphoribosyltransferase was amplified using the forward primer 5′-GTCCCAGCGTCGTGATTAGCGATG-3′ and backward primer 3′-GGCTGGCCTATAGGCTCATAGTGC-5′.

Mice were injected i.p. with 1.5 ml sterile filtrated 10% Proteose peptone in PBS. Three days postinjection, peritoneal cells were collected, washed with PBS, resuspended in RPMI 1640 medium with l-glutamine (0.3 g/l), 5% FBS, 10 mM HEPES, and penicillin (100 U/ml)/streptomycin (100 mg/ml) (Invitrogen), and seeded at 1 × 10⁶ cells/well on coverslips in a 24-well plate (Techno Plastic Products). After incubation for at least 5 h, the wells were washed with RPMI 1640 medium to remove nonadherent cells.

Bone marrow was obtained by flushing tibia and femur with ∼5 ml PBS. Cells were centrifuged at 300 × g for 5 min and resuspended in RPMI 1640 medium with l-glutamine (0.3 g/l), 10% FBS, 10 mM HEPES, penicillin (100 U/ml)/streptomycin (100 mg/ml) (Invitrogen), and 15 ng/ml murine recombinant GM-CSF (Natutec). A total of 2 × 10⁶ cells were seeded per plastic Petri dish and incubated at 37°C. On days 3 and 6, fresh medium containing GM-CSF was added to each Petri dish. Bone marrow-derived dendritic cells (BMDCs) were harvested on day 9.

A total of 1 × 10⁶ cells were seeded per well on coverslips in a 24-well plate and stimulated. After stimulation for the indicated time, the cells were washed once with PBS and covered with CCF4-AM staining solution supplemented with probenecid (efflux inhibitor), prepared according to the manufacturer’s instructions, and incubated for 90 min at room temperature (RT). CCF4-AM solution was then removed and the cells washed once with PBS and fixed with 3.7% formaldehyde. Cells were then washed two times with PBS and analyzed by microscopy using an Olympus BX-61 fluorescence microscope (Olympus) equipped with cell^P software (Olympus Soft Imaging Solutions). Images were converted to RGB, colored, and overlaid.

Analysis of cytokines from cell-culture supernatants was performed using a cytometric bead array (CBA) kit (BD Biosciences) and measured on an FACSCanto II flow cytometer (BD Biosciences). The assay was performed according to the manufacturer’s instructions. A total of 25–50 μl cell supernatant was subjected to the assay. FACS data were acquired on an FACSCanto II (BD Biosciences) and analyzed using the FCAP Array software (Soft Flow). Standard curves were determined for each cytokine from a range of 10–5000 pg/ml.

Single-cell suspensions from mouse organs were prepared according to the gentleMACS protocols of Miltenyi Biotec. Livers and lungs were partially dissociated using the gentleMACS Dissociator (Miltenyi Biotec). Dissociated livers were incubated at 37°C for 30 min in Krebs–Ringer buffer (154 mM NaCl, 5.6 mM KCl, 5.5 mM glucose, 20.1 mM HEPES, and 25 mM NaHCO₃, adjusted to pH 7.4 with NaOH) containing 200 collagen digestion units/ml collagenase IV (Sigma-Aldrich) and 150 U/ml DNase I solution (AppliChem). Dissociated lungs were incubated at 37°C for 30 min in HEPES buffer (10 mM HEPES, adjusted to pH 7.4 with NaOH, 150 mM NaCl, 5 mM KCl, 1 mM MgCl₂, and 1.8 mM CaCl₂) containing 3 collagen digestion units/ml collagenase D (Roche Diagnostics) and 40 U/ml DNase I solution (AppliChem). The digested livers, lungs, whole spleens, and tumors were homogenized using the gentleMACS Dissociator according to the manufacturer’s instructions (Miltenyi Biotec), and cell suspensions were passed through a 100-μm nylon cell strainer (BD Biosciences). In the case of the livers, hepatocytes were pelleted by low-speed centrifugation (∼20 × g for 4 min at 4°C). RBCs were lysed by incubating the cell suspensions in erythrocyte lysis buffer (0.154 M NH₄Cl, 10 mM KHCO₃, and 0.1 mM EDTA [pH 7.4]) for 5 min at RT. Cells were centrifuged at 4°C, 300 × g, for 10 min. Cells from ITIB and control littermate mice were resuspended in CCF4-AM staining solution supplemented with probenecid, prepared according to the manufacturer’s instructions (Invitrogen), and incubated for 75 min, protected from light, at RT. Cells were then washed with FACS buffer (PBS, 0.5% BSA or 2% FBS, and 2 mM EDTA) and directly analyzed by flow cytometry or subjected to immunofluorescence staining. In this case, samples of 10⁶ cells were each resuspended in 100 μl FACS buffer containing 1 μg/10⁶ cells anti-FcγII/III (eBioscience) (to block Fc receptors) and incubated for 15 min at 4°C. Subsequently, cells were incubated for a further 30 min at 4°C with combinations of Abs against the following surface markers: CD45–PE-Cy7, CD11c–allophycocyanin-Alexa 750, F4/80–PE-Cy5, Ly6G-PE, CD49b-allophycocyanin, B220/CD45R-PE, CD8-PerCP, CD4-PE, CD4–allophycocyanin-H7, and CD25-allophycocyanin (purchased from BD Biosciences or eBioscience). Cells were then washed with FACS buffer, and 10⁵ cells were acquired on an FACSCantoII (BD Biosciences). Data were collected and analyzed with FlowJo 7.6.1. software (Tree Star).

CD4⁺ T cells were purified with MACS (Miltenyi Biotec), and 10⁶ CD4⁺ T cells were cultured in RPMI 1640 medium (supplemented with l-glutamine, 10% FBS, penicillin and streptomycin, and 0.1 mM 2-ME) with Dynabeads Mouse T-Activator CD3/CD28 (Invitrogen) at a bead to cell ratio of 1:1, in the presence or absence of 10⁻⁷ M dexamethasone (Sigma-Aldrich) and 4 × 10⁻⁸ M 1α,25-dihydroxyvitamin D3 (Sigma-Aldrich).

B220⁺ cells were purified with MACS (Miltenyi Biotec), and 10⁶ B220⁺ cells were cultured in RPMI 1640 medium (supplemented with l-glutamine, 5% FBS, penicillin and streptomycin, and 0.1 mM 2-ME). The stimulation of B220⁺ cells was performed by adding 5 μM oligodeoxynucleotides (ODNs; InvivoGen/Cayla SAS) or 10 μg/ml anti-IgD (eBioscience).

BMDCs and peritoneal exudate cells (PECs) were stimulated with the following TLR agonists (purchased from InvivoGen/Cayla SAS): LPS from Escherichia coli K12 (at the indicated concentration), 50 μg/ml zymosan (sometimes in combination with 40 ng/ml TGF-β, purchased from LABGEN/NatuTec), 5 μM CpG ODNs, and 5 μM CpG ODN (synthetic ODNs containing unmethylated CpG dinucleotides in particular sequence).

BMDCs were stimulated for 3 h with 50 μg/ml zymosan and 40 ng/ml TGF-β. After 3 h, medium was exchanged with another one containing 100 μg/ml cycloheximide (Sigma-Aldrich), and cells were further incubated at 37°C. Supernatants were collected at the indicated time points, and BMDCs were loaded with CCF4-AM and analyzed by FACS.

Mice were inoculated i.p. with 500 μl PBS containing 5 × 10⁴ CFU Y. enterocolitica, serotype O:8, strain WA-314 (12) or 1 × 10⁴ CFU Salmonella enterica serovar Typhimurium, strain SL1344 (SB300) (13). Mice were sacrificed by CO₂ asphyxiation at indicated days, spleens were dissected, and single-cell suspensions were prepared for further staining procedures as described above.

To identify IL-10–producing cells in vivo, we first generated an IL-10–EGFP reporter mouse model in which an internal ribosomal entry site (IRES)-eGFP cassette was inserted by homologous recombination immediately before the 3′ UTR of the Il10 gene (Fig. 1A, Supplemental Fig. 1). In this configuration, eGFP is transcribed along with the Il10 gene under the control of the endogenous Il10 promoter. Il10 translation remains cap-dependent, whereas the translation of eGFP is driven by IRES. Furthermore, the posttranscriptional regulation through the endogenous 3′ UTR of Il10 is not affected. Mice carrying the allele IL-10–1×(IRES-eGFP) were designated ITIG (T stands for 10). In the remaining text, the ITIG mice denote homozygous reporter mice.

To examine whether the ITIG mice would be able to track IL-10–producing cells, BMDCs and splenic B and T cells from the ITIG mice were prepared and stimulated in vitro. BMDCs were stimulated for 24 h with 500 ng/ml LPS or 50 μg/ml zymosan. Both agonists are known as potent inducers of IL-10 production by DCs (14, 15). CD4 T cells were stimulated for 3 d with a combination of the IL-10 inducers, vitamin D3, and dexamethasone (16) in the presence of anti-CD3 and anti-CD28. B cells were stimulated for 2 d with 5 μM synthetic unmethylated CpG dinucleotides (CpG ODNs) that are known to trigger IL-10 production by B cells (17).

Unfortunately, although a significant amount of IL-10 could be measured in the supernatants of the stimulated ITIG cells (Fig. 1C), we failed to detect significant EGFP fluorescence in the ITIG BMDCs and B cells (Fig. 1B). However, a considerable EGFP-positive population could be detected in stimulated ITIG T cells.

Interestingly, ITIG T cells produced more IL-10 than BMDCs and B cells (Fig. 1C). Thus, we conclude that the ITIG reporter mouse seems not to be sensitive enough to track low IL-10 amounts. This is illustrated in more detail with the following example of the zymosan-stimulated ITIG BMDCs.

Considering that ∼1.5 μg IL-10 were produced by 10⁶ BMDCs (Fig. 1C), the estimated average number of IL-10 molecules produced per cell should be ∼5 × 10⁴. Keeping in mind that on one hand GFP’s limit of detection is ∼10⁵ molecules per typical cell of 1 to 2 pL volume (this estimate assumes perfect GFP maturation) (7), and on the other hand, IRES-dependent gene expression is ∼20–50% of that of the cap-dependent gene (Ref. 18 and H. Bouabe, unpublished observations), the expected number of GFP molecules per zymosan-stimulated ITIG DC should be 50–80% lower than that of IL-10 (≤2.5 × 10⁴) and thus significantly below the GFP’s limit of detection. Moreover, the cellular size and autofluorescence of cells, such as B cells and hematopoietic cells of myeloid lineages, as well as incomplete maturation of GFP molecules would raise the threshold copy number of GFP even further (7). In contrast, T cells, which show much higher IL-10 protein expression and have very small cytosols and low autofluorescence, might require fewer numbers of GFP molecules to be readily detected.

In conclusion, our IL-10–EGFP reporter mouse model is, in the best case, suitable to track IL-10 production only in those cells that are relatively small, show low autofluorescence, and produce a very high amount of IL-10, as is the case for T cells.

In an attempt to overcome the experimental limitations of the ITIG mouse in monitoring weak IL-10 expression in all cell types, we decided to design and develop an alternative IL-10 reporter mouse model using for the first time, to our knowledge, the enzyme TEM-1 bla as reporter gene in mice. We chose bla for different reasons: 1) a membrane-permeable, fluorogenic substrate for Bla (CCF4-AM) is available that allows analyzing of Bla expression in intact, live cells by flow cytometry and fluorescence microscopy (8); and 2) as few as 50 molecules of Bla are sufficient to lead a detectable signal (8).

The IL-10–bla reporter mice were generated by inserting two copies of a cassette containing IRES and bla (IRES-bla-IRES-bla) at the end of the last exon and before the polyadenylation signal of the Il10 gene (Fig. 2A, Supplemental Fig. 2). The insertion of two IRES-bla cassettes is expected to double the expression of bla (19). Mice bearing the allele IL-10–2×(IRES-bla), were designated ITIB (T stands for 10). In the remaining text, the ITIB mice denote homozygous reporter mice (ITIB/ITIB).

Taking into account the drawbacks of the IL-10–EGFP reporter mouse shown above, we wondered if the ITIB mouse would be able to mark more efficiently IL-10–producing cells, especially B cells and myeloid cells that were difficult to track with EGFP. Thus, we first analyzed Bla reporter activity in proteose peptone-elicited PECs, which comprise mainly monocytes/macrophages. PECs were stimulated with 5 μM CpG ODNs or 500 ng/ml LPS for ∼13 h. PECs were then analyzed by fluorescence microscopy. Interestingly, Bla reporter activity could be monitored for both LPS- and CpG ODN-stimulated ITIB PECs (CCF4 conversion to blue) (Fig. 2B). No blue fluorescent cells could be detected in nonstimulated cells. Furthermore, Bla reporter activity correlated with the release of IL-10 into the cell-culture supernatants (Fig. 2C), confirming the fidelity and specificity of Bla reporter activity. The fact that fluorescence microscopy is much less sensitive than flow cytometry speaks in favor of the prominent sensitivity of Bla.

We next tested the ability to detect Bla reporter activity in different cell types by FACS. Purified splenic CD4 T cells from ITIB mice were stimulated in vitro for 3 d with a combination of the IL-10 inducers vitamin D3 and dexamethasone. In contrast to ITIG T cells that were stimulated under the same conditions and showed EGFP positivity only in ∼13% of the cells (Fig. 1B), ∼50% of the activated ITIB T cells exhibited Bla reporter activity (Fig. 2D). Moreover, in contrast to BMDCs and B cells from ITIG mice, which did not respond with detectable EGFP fluorescence, although they produced IL-10 (Fig. 1B, 1C), Bla reporter activity could be easily detected in significant subsets of purified splenic B220⁺ B cells stimulated with CpG ODNs and of ITIB BMDCs stimulated with zymosan or LPS (Fig. 2D). However, no IL-10 production as well as no Bla reporter activity were observed when BMDCs were stimulated with LPS in the presence of polymyxin B, an agent known to neutralize the activity of LPS (20), thus validating the fidelity and specificity of Bla reporter activity (Supplemental Fig. 3A, 3B).

In addition to its high sensitivity enabling the detection of IL-10 expression in all hematopoietic cell types, Bla has an additional advantage against EGFP: a short t_1/2 time that is ∼3 h (8) and thus comparable to that of IL-10, for which t_1/2 time is in the range of 60–180 min (21, 22). Furthermore, cycloheximide assay confirmed that the stability of Bla is parallel to that of IL-10. Bla reporter activity and IL-10 production level already decreased 6 h poststimulation to a very low level (Supplemental Fig. 3C, 3D). Thus, Bla appeared to be superior when reporting transient gene expression. For this reason, we investigated the ability of the ITIB mouse to perform kinetic studies. BMDCs from ITIB mice were stimulated with LPS, and Bla reporter activity was analyzed at different time points by FACS. As shown in Fig. 3A, the percentage of Bla⁺ cells increased over time and reached its maximum at ∼12 h. After 48 h, only a minor population of Bla⁺ cells could be detected (Fig. 3A). The increase and decrease of Bla⁺ cells correlated with the level of IL-10 secreted in the respective cell culture supernatants (Fig. 3B). Furthermore, the up- and downregulation of IL-10 expression could be reflected by the level of Bla activity at the single-cell level that was quantified by measuring the ratio of blue to green fluorescence of the Bla⁺ population (the ratio of blue to green is an indirect parameter of the amount of Bla within a cell) (Fig. 3C). Thus, ITIB cells enable to enumerate in real time the frequency of IL-10 producer cells.

Whole single-cell suspensions from spleen, liver, and lung of naive ITIB mice were loaded with CCF4-AM, stained with Abs against immune cell-surface markers, and analyzed by FACS. A significant basal Bla reporter activity could be detected in ∼2 to 3% of splenic and hepatic leukocytes (Fig. 4A) as well as lung leukocytes (Fig. 4E). In comparison, no EGFP-positive splenic leukocytes (Fig. 4B), hepatic, and lung leukocytes (data not shown) could be detected in naive ITIG mice.

The IL-10 reporter expression in the peripheral lymphoid organ spleen and in the nonlymphoid organs liver and lung of naive ITIB mice is restricted to hematopoietic cells (Figs. 3E, 4A). Interestingly, each of the examined myeloid and lymphoid cell types, including, for example, B220⁺(CD11c⁻) B cells, CD4⁺ T cells, CD8⁺ T cells, CD49b⁺ NK cells, CD11c⁺ DCs, Ly6G⁺ neutrophils, and F4/80⁺ macrophages, showed significant subsets of Bla⁺ cells (Supplemental Fig. 4A and data not shown). However, when analyzing the percentage of cell types represented in the splenic, hepatic, and lung Bla⁺ cell populations, the majority (≥50%) of the Bla⁺ leukocytes in each of these organs were F4/80⁺ macrophages (Fig. 4C–E). Interestingly, although only a minority of Bla⁺F4/80⁺ splenic macrophages express the CD11b myeloid marker, the majority of Bla⁺F4/80⁺ hepatic and lung macrophages are CD11b⁺ (Fig. 4C, 4D). Moreover, significant subsets of splenic and lung Bla⁺F4/80⁺ macrophages were also positive for CD11c (Fig. 4C, 4E).

The second large cellular population that produces IL-10 was CD4⁺ T cells that constitute ∼15% and ∼30% of Bla⁺ leukocytes in spleen and liver, respectively (Fig. 4C, 4D). In addition, one third of splenic Bla⁺CD4⁺ leukocytes were CD25⁺ (Fig. 4C).

Furthermore, B220⁺ cells (including B200⁺CD11c⁻ and B220⁺CD11c⁺ cells) constitute ∼25% of the Bla⁺ leukocytes in the spleen (Fig. 4C).

IL-10 has emerged as a key regulator of immune responses during infection, inhibiting excessive immune activation and ameliorating immunopathology (23). Previously, it has been demonstrated that Y. enterocolitica induces an IL-10 response through the virulence Ag LcrV, which is recognized by TLR2 (24). To identify the cell types contributing to IL-10 production during infection with Yersinia, we infected ITIB mice i.p. with Y. enterocolitica and 5 d later analyzed splenic and hepatic leukocytes for Bla expression.

In comparison with naive mice (Fig. 5A), the fraction of Bla-positive splenic leukocytes increased upon infection at least two to three times (Fig. 5A). In contrast, no or only a marginal GFP⁺ population (≤1%) of splenic leukocytes could be detected in Yersinia-infected ITIG mice (Fig. 5B). However, the ratio of Bla-positive leukocytes was dependent on the bacterial burden in the spleen. The higher the bacterial burden, the larger the Bla⁺ population (Supplemental Fig. 3E, 3F and data not shown). Moreover, the level of IL-10 production by splenocytes from infected mice paralleled Bla activity, and both (Bla activity and IL-10 level) increased with infection time and with the level of bacterial burden in the examined organ (Supplemental Fig. 3E, 3F), thus confirming the fidelity and specificity of Bla reporter activity and demonstrating the suitability of ITIB mice for kinetic studies in vivo.

Interestingly, there was significant increase of the IL-10⁺ cell subsets in each myeloid and lymphoid cell type that was present in the Yersinia-infected spleen (Supplemental Fig. 4B). However, the analyses of the percentage of cell types represented in the splenic Bla⁺ cell population revealed that the majority were CD11b⁺Ly6G⁺ neutrophils, followed by CD11b⁺Ly6G⁻ (F4/80⁻CD11c⁻) monocytes (Fig. 5C). Both Bla⁺ populations constituted ∼50% of splenic Bla⁺ leukocytes. Coincidentally, CD11b⁺ cells represent the major cell population that was freshly recruited to the spleen during Yersinia infection (Supplemental Fig. 4C).

Analysis concerning CD4⁺ T cells revealed that although there was an increase of the IL-10⁺ cell subsets in CD4⁺ T cell population (Supplemental Fig. 4B), the percentage of CD4⁺ T cells (including CD4⁺CD25⁺ and CD4⁺CD25⁻) in the Bla⁺ splenic leukocyte populations of infected mice was comparable with that of naive mice (Fig. 4C, 5C). Similarly, the frequency of B220⁺ cells (including B200⁺CD11c⁻ and B220⁺CD11c⁺ cells) in the Bla⁺ splenic population also remained constant postinfection (Fig. 4C, 5C).

As neutrophils are the first immune cells that rapidly migrate to the site of inflammation, we wondered whether they also produce IL-10 at an early time during Yersinia infection. Thus, we infected ITIB mice with Y. enterocolitica and analyzed Ly6G⁺ neutrophils for Bla expression 1 d postinfection. As shown in Fig. 5D, along with strong neutrophil infiltration of the spleen, the ratio of Bla-positive neutrophils increased significantly among the Bla⁺ splenic leukocyte population (Fig. 5D).

Because rapid recruitment of neutrophils is not an exception for infection with the extracellularly multiplying Y. enterocolitica, we wondered whether infection with an intracellular pathogen, such as Salmonella, would also result in induction of IL-10 production by neutrophils. We therefore infected ITIB mice with S. Typhimurium, an enteropathogenic bacterium that is known to induce recruitment of neutrophils to the infected organs (25). Five days postinfection, mice were sacrificed, and the presence of Bla⁺ neutrophils in the spleen was investigated. Similarly to Yersinia infection, the fraction of Bla⁺ splenic leukocytes increased upon Salmonella infection at least two to three times (Fig. 5E). Interestingly, the ratio of the Bla⁺Ly6G⁺ neutrophil population was almost half that of the whole Bla⁺ splenic leukocyte population (Fig. 5E).

There is increasing evidence that subsets of tumor-infiltrating leukocytes (TILs) suppress immune responses against tumors and promote their escape and probably dissemination. Among the suppressor cells are regulatory T cells and tumor-associated macrophages. One possible immune suppressor mechanism governed by these suppressor cells has been suggested to be the expression of IL-10 (26–28). Thus, the identification of IL-10–producing cells in tumor microenvironments in vivo is a prerequisite to investigate their implication in tumorigenesis.

In this study, using the example of an implantable murine skin B16-F10 melanoma model, a tumor for which growth is modulated by IL-10 (27, 29), we demonstrate that the ITIB mouse provides a valuable tool for the identification of IL-10–producing cells in tumor microenvironments in vivo.

ITIB mice were s.c. injected with 2 × 10⁵ B16-F10 melanoma cells into the lower dorsal quadrant. Two weeks later, tumors were removed, from which single-cell suspensions were prepared, loaded with CCF4-AM, stained with Abs against immune cell-surface markers, and analyzed by flow cytometer. Interestingly, ∼80% of the Bla⁺ population displayed the markers F4/80, CD11b, and CD11c (Fig. 6). However, the appearance of all three cell-surface markers on the same cell was dependent on the tumor growth stage. Usually, they seem to appear in the following order: CD11b, F4/80, and then CD11c (data not shown). Thus, our ex vivo data (without ex vivo restimulation of TILs) confirm the in vitro and indirect in vivo (e.g., using IL-10 knockout mice) data showing that tumor-associated macrophages have a tendency to release IL-10 (29–31).

The development of cell type-specific IL-10–modulating agents is a promising therapeutical strategy in cancer and inflammatory and autoimmune diseases. We therefore investigated the prospective ability of ITIB mice to perform in vivo screening of cell type-specific IL-10–inducing agents. We used the example of the mouse CD28 superagonist, a specific mAb against the T lymphocyte costimulatory receptor CD28, which has been shown to activate T cells without the need for TCR ligation and preferentially induce the expansion of IL-10–producing CD4⁺CD25⁺ regulatory T cells with therapeutic potential for the treatment of autoimmune diseases (32).

ITIB mice were injected i.p. with 400 μg CD28 superagonist, and 4 d after the injection, splenocytes were analyzed for Bla production. Although there was no major change in the number of CD4⁺ T cells in the spleen after a single treatment with CD28 superagonist, the frequencies of CD4⁺CD25⁺ as well as Bla⁺CD4⁺CD25⁺ T cell populations each increased to up to ∼2.5 times compared with nontreated mice (Fig. 7). Thus, the ITIB reporter mouse is a valuable tool for in vivo screening of cell type-specific IL-10–inducing agents.

IL-10 is one of the key cytokines involved in immune regulation and tolerance, and thus is a promising therapeutic agent, for example, for autoimmune diseases and cancer. However, systemic administration of recombinant IL-10 to treat, for example, immune-mediated inflammatory diseases showed no or only modest therapeutic benefits (33). Thus, the localization and timing of IL-10 production/activities may determine its effectiveness. Furthermore, cell type-specific inactivation of the Il10 gene revealed a critical nonredundant role of different IL-10–producing cell types in immune homeostasis and in control of particular immune responses (34–37). Thus, a successful therapeutic use of IL-10 has also to take into account the IL-10–producing cell types under given physiological and immunological conditions; for example, by developing drugs that specifically control, temporally and locally, IL-10 production by a given cell type.

For this reason, there is a need to know the type and implication of cells that produce IL-10 in the steady state and under immune activating conditions in vivo. However, because of the low expression level of IL-10 and its rapid secretion, it is difficult to directly identify IL-10–producing cells in vivo or ex vivo.

Alternative, widely used, experimental tools to identify the in vivo cellular sources of a cytokine of interest offer reporter mice using autofluorescent proteins, such as EGFP (6).

Correspondingly, in an attempt to identify IL-10–producing cells in vivo, we first generated an IL-10–EGFP reporter (ITIG) mouse model. However, the ITIG mice allowed tracking IL-10 expression only in activated T cells. Furthermore, the ITIG mice did not enable to clear whether and which cell type produces IL-10 in vivo. In contrast to our results, in a recently published IL-10 reporter tiger mouse that is similar to our ITIG mouse model, Kamanaka et al. (38) could also detect GFP expression in BMDCs after stimulation with zymosan. The cause of this discrepancy is probably related to the stimulation protocol and the type of BMDCs used by us and by Kamanaka et al. (38). Whereas we generated BMDCs by incubating bone marrow cells 9 d with a defined concentration of recombinant GM-CSF, Kamanaka et al. (38) obtained BMDCs by culturing bone marrow cells 5 d in 3% supernatant of a GM-CSF–producing L929 cell line. However, fresh DCs from the spleen and Peyer’s patch of tiger mice that were stimulated in vitro with zymosan showed only a marginal GFP expression (38). Furthermore, in accordance with our results using the present ITIG mice, no significant GFP signals could be identified in macrophages, B cells, or DCs after in vivo stimulation of tiger mice (38).

Thus, IL-10–EGFP reporter mice (such as the present ITIG mice) are not suitable to track IL-10 production in cells with high autofluorescence (such as macrophages, DCs, and B cells) and/or low amount of IL-10.

The low expression of IL-10 is not due to its weak promoter activity alone, rather IL-10 expression, like many other cytokines, is strongly controlled by posttranscriptional regulation. Sequences in the 3′ UTR that contain mRNA destabilizing motifs confer a short t_1/2 to the Il10 mRNA (39–42). Powell et al. (41) showed that the t_1/2 of RNA containing the 3′ UTR of IL-10 was short in both nonstimulated (t_1/2 of 1 h) and PMA-stimulated (t_1/2 of 3 h) cells, and significant IL-10 protein production was shown only after stimulation. In contrast, when the 3′ UTR of IL-10 was exchanged for SV40 polyadenylation signals that ensure mRNA stability, the t_1/2 of Il10 mRNA was longer than 12 h, and the IL-10 protein level was much higher (41). Previously, we also confirmed with two different reporter genes, luciferase and EGFP, that the 3′ UTR of IL-10 negatively regulates the protein production (19). The rapid degradation of Il10 mRNA is mediated, for example, by binding of AU-rich element binding factor 1, tristetraprolin, or microRNA (hsamiR-106a) to the 3′ UTR (40, 43, 44). Taken together, this intrinsic instability of Il10 mRNA results in low and transient expression of IL-10 protein and consequently restricts the sensitivity of a reporter gene like EGFP that is known to have a high limit of detection (∼10⁵ molecules/cell) (7). Recently, IL-10 reporter mouse models, 10BiT, Vert-X, and Il10^Venus mice, have been reported in which the authors tried to overcome this limitation by exchanging the mRNA-destabilizing Il10-3′ UTR for the SV40 early intron-polyA (SV40 polyA) that ensures mRNA stability of the corresponding reporter genes (37, 45, 46). Although these reporter mouse models indeed showed improved reporter activity, they lack a decisive posttranscriptional regulating element, the Il10-3′ UTR. Thus, these reporter mouse strains report only the transcription of the Il10 gene but not its expression at protein level.

In this study, we presented a more suitable and powerful alternative to efficiently track IL-10 protein expression using for the first time, to our knowledge, the reporter enzyme Bla and the fluorescence resonance energy transfer substrate CCF4. The advantage of Bla, compared with other reporter enzymes such as luciferase, is the ability to combine Bla analysis with multicolor immunostaining, which enabled us to perform a detailed analysis of IL-10–producing cell populations.

In contrast to the ITIG mouse line, the ITIB mouse model enables tracking of the expression of IL-10 in all myeloid and lymphoid cell types.

Strikingly, ITIB mice revealed that macrophages represent the major IL-10 producer population in the peripheral lymphoid organ spleen and in the nonlymphoid organs liver and lung of naive mice examined in this study. This phenomenon is especially important in the case of the spleen, where macrophages constitute a relatively small percentage (5–10%) of splenic leukocytes. In contrast, in the previously described Vert-X mice, B cells were the dominant population of IL-10–expressing cells in the spleen of naive mice (37). How can this controversy be explained? The expression of the reporter genes, bla and eGFP, in the ITIB and Vert-X mice, respectively, are differently regulated at the posttranscriptional level; bla is under the control of the endogenous 3′ UTR of Il10, whereas eGFP under the control of the exogenous mRNA-stabilizing SV40 polyA. Thus, a likely explanation for the inconsistent cellular expression profiles of IL-10 in the spleen of naive ITIB and Vert-X mice may be that the expression of Il10 is cell type-specific differentially regulated at posttranscriptional level. It seems that, at least in the steady state, the Il10-3′ UTR mediates Il10 mRNA instability more profoundly in B cells than in macrophages and thus confers stronger suppression of IL-10 protein production in the former cell type. Recently, posttranscriptional regulation of Il10 gene expression has been shown to control specifically IL-10 secretion by NK cells (47). Furthermore, it has been shown that, compared with normal melanocytes, the melanoma cell line MNT1 exhibits reduced cytoplasmic levels of AU-rich element binding factor 1, an AU-rich element binding protein that mediate mRNA instability, resulting in IL-10 overexpression and tumor escape (40). Thus, a possible mechanism that may underlie a cell type-specific posttranscriptional expression regulation of Il10 could be that proteins and microRNAs that bind to the 3′ UTR and regulate the stability of Il10 mRNA display cell type-dependent differential expression/activation level.

It is noteworthy that the ITIB mouse developed in this study combined with the previously described IL-10 transcriptional reporter mice, such as Vert-X and Il10^Venus mice (37, 46), may form an interesting basis of further works investigating the cell type-specific expression regulation of IL-10 at the transcriptional and posttranscriptional level and thus may help elucidating the various mechanisms that determine the amount of IL-10 production by a given cell type.

Resident macrophages are present in several organs where they perform several functions in steady-state tissue homeostasis including, among others, tissue repair, vascularization, and removal of apoptotic cells, aged erythrocytes, and cellular debris (48, 49). Macrophages display remarkable heterogeneity and adapt their physiology in response to environmental signals in the tissues where they reside (49, 50). Depending, for example, on the type of cytokines they produce, macrophages can be divided into two different functional subpopulations: classically activated (M1) macrophages (inflammatory and microbicidal activities) and regulatory macrophages (referred also as alternatively activated [M2] macrophages) (anti-inflammatory, wound healing, and tissue repair activities) (49, 51, 52). As IL-10 is considered as an important functional marker for regulatory macrophages, we suggest that the identified splenic, hepatic, and lung F4/80⁺Bla⁺ macrophages may perform homeostatic immune regulatory functions. M2 macrophages have been shown, among others, to induce immunological tolerance in the lung by protecting it from unwanted environmentally induced inflammation (53). Furthermore, a recent study has shown that IL-10–producing intestinal macrophages play an important role in the maintenance of gut homeostasis (54). Possible mechanisms underlying the induction of IL-10 production by tissue-resident macrophages may be associated, for example, with their function to phagocyte apoptotic cells and to uptake immune complexes. These processes have been shown to induce IL-10 production by macrophages (49, 55–57). Taken together, we suggest ubiquitous interdependence between homeostatic functions of resident macrophages and IL-10 production.

Interestingly, freshly recruited F4/80⁺ macrophages represented also the major IL-10–producing population in the microenvironment of an s.c. B16-F10 tumor. This raises the question about the relationship between processes inducing IL-10 production by macrophages in different organs and tumors. Healing wounds and developing tumors have long been described to share many similar features with repairing tissues (58–60). Furthermore, organs, like spleen, liver, and lung, are usually in continuous renewing, repairing, and/or remodeling processes of their tissues and cells. Consequently, as M2 macrophages play an essential role in wound repair and tissue remodeling, tumors, and organs share processes that provide a common reason for accumulation of IL-10–producing (regulatory) macrophages.

Neutrophils and monocytes are the first leukocytes to infiltrate infected tissue, where they contribute an essential role in defending many bacterial and fungal infections (61, 62). For instance, mice depleted of neutrophils/monocytes were more susceptible to systemic infection with Yersinia (63, 64), and Nod2-deficient mice showed increased production of the neutrophil chemoattractant, keratinocyte chemoattractant, accompanied with stronger neutrophil infiltration and improved killing of Yersinia (65). Strikingly, in systemic infection of ITIB mice with Y. enterocolitica, infiltrating CD11b⁺Ly6G⁺ neutrophils followed by CD11b⁺Ly6G⁻ (F4/80⁻CD11c⁻) monocytes represented major IL-10–producing cell populations in the spleen (5 d after inoculation). Y. enterocolitica is known to disseminate, among other locations, to the spleen, where they replicate predominately extracellularly and form microcolonies consisting primarily of extracellular residing Yersinia-bacteria and neutrophils (66). In contrast, S. Typhimuruim is an intracellular living pathogen that has been recently shown to specifically target splenic neutrophils and to survive and replicate inside them (67). However, despite the different life styles of Yersinia and Salmonella, Ly6G⁺ neutrophils represented also a major IL-10–producing population in the spleen of Salmonella-infected ITIB mice. Moreover, an increase of the proportion of Bla⁺ neutrophils has been also shown at an early time point (24 h) postinfection. Thus, IL-10 production seems to belong to the response program of activated neutrophils/monocytes, as early as they are recruited in the site of infection. Further work will be required to elucidate in more detail the contributions of IL-10 producing neutrophils and monocytes to systemic bacterial infections.

Since the discovery of the so-called innate immune receptors (e.g., TLRs and nucleotide-binding oligomerization domain-like receptors), the missing link between innate and adaptive immunity was established (68), and the cells of the innate immune system have moved from being an evolutionary remnant to a powerful and decisive player in all immune-response events. In accordance with this revised view, the present ITIB mice revealed that cells of the innate immune system that are involved in immune homeostasis or in immune responses are substantial source of the important immune regulatory factor IL-10.

For this reason, to gain accurate and complete insightful information about cytokine-mediated immune regulation, the efforts that are made to track and investigate cytokine-producing cells in vivo should not be predominantly concentrated on adaptive immune cells, like T cells. However, the current focus on T cells is partly due to the limited sensitivity of intracellular immunostaining and reporter mice (particularly those using autofluorescent proteins as reporter) that do not enable the detection of weakly expressed cytokines. In this study, we showed that the use of Bla as a reporter gene for mice helps to overcome these limitations and open up unique possibilities to study the expression of weakly expressed genes in all immune cell types that, all together, compose our immune system.

Finally, it is noteworthy that by developing new substrates for Bla, other yet-unrecognized uses of the ITIB mouse can be established.

We thank Thomas Hünig for the gift of CD28 superagonist, Werner Müller for the gift of the vector pTZ18R10, and Reinhard Fässler for facilities for ES cell culture. We also thank Joe Dramiga and Lukas Schneider for the English proofreading of the manuscript.

The authors have no financial conflicts of interest.