The IL-10GFP (VeRT-X) mouse strain is not suitable for the detection of IL-10 production by granulocytes during lung inflammation

The clear and unequivocal identification of immune effector functions is essential to understand immune responses. The cytokine IL-10 is a critical immune regulator and was shown, for example, to limit pathology during various lung diseases. However, the clear identification of IL-10-producing cells is challenging and, therefore, reporter mouse lines were developed to facilitate their detection. Several such reporter lines utilize GFP, including the IL-10GFP (VeRT-X) reporter strain studied here. In line with previous reports, we found that this IL-10GFP line faithfully reports on the IL-10 production of lymphoid cells. However, we show that the IL-10GFP reporter is not suitable to analyse IL-10 production of myeloid cells during inflammation. During inflammation, the autofluorescence of myeloid cells increased to an extent that entirely masked the IL-10-specific GFP-signal. Our data illustrate a general and important technical caveat using GFP-reporter lines for the analysis of myeloid cells and suggest that previous reports on effector functions of myeloid cells using such GFP-based reporters might require re-evaluation.


Introduction
Regulatory cytokines are important players during pulmonary inflammation to limit immunopathology without hampering effective pathogen clearance [1]. The regulatory cytokine IL-10 is involved in various lung diseases, like asthma, allergic airway disease (AAD), chronic obstructive pulmonary disease (COPD), and pulmonary infections [1]. IL-10 producing cells in the lung; FoxP3 + and FoxP3regulatory T cells [1], CD8 + T cells [2], B cells, alveolar macrophages (AM) [3], interstitial macrophages (IM) [4], airway and alveolar epithelial cells [1], and airway-associated dendritic cells (DCs) [1] were shown to limit disease pathology. However, due to the low percentage of IL-10 + cells within most cell populations and due to the low intensity of the flow cytometric staining of intracellular IL-10 the clear identification of IL-10 producing cells is challenging [5]. Therefore, various reporter mouse lines were developed to avoid the need for intracellular cytokine staining to identify IL-10 producing cells [6]. The IL- 10 GFP (VeRT-X) strain studied here, expresses an (IRES)-enhanced green fluorescent protein (eGFP) fusion protein downstream of the exon 5 of the il10 gene [7], and was reported to enable the identification of IL-10 + lymphoid but also myeloid cells [7]. Several myeloid cells possess an autofluorescence around 525 nm [8], which coincides with the emission maxima of GFP around 530 nm [9]. However, the impact of this myeloid autofluorescence on the detection of GFP-reporter signals has not been fully clarified. Here, using an LPS-induced lung inflammation model, we demonstrate that the autofluorescence of myeloid cells conceals the IL10 GFP -specific signal. This was mainly due to a large increase in the myeloid autofluorescence during the inflammation. These data demonstrate that not all GFP-reporter mouse strains are suitable to analyse effector functions of myeloid cells during inflammation.

PLOS ONE
The IL-10 GFP strain is not suitable for the IL-10 detection by granulocytes To clarify which cell types could produce IL-10, we directly compared the IL-10 signal derived from the GFP-signal or from intracellular cytokine staining (ICCS) for CD4 + and CD8 + T cells, neutrophils, macrophages, Siglec-F + cells (eosinophils, alveolar macrophages), monocytes, and ILCs. Cells from the lungs, spleens, and mLN of PBS (control) and LPS-challenged C57BL/6 and IL-10 GFP mice were analysed (S1 Fig). Although intracellular IL-10 staining is challenging, we obtained a clear IL-10 signal with the commercial antibody (S2 Fig). With this side-by-side comparison, we found that the IL-10-signal derived from GFP or ICCS correlated well for CD4 + (Fig 2A) and CD8 + T cells (Fig 2B). These data indicate that the IL-10GFP-signal faithfully reports on the IL-10 production of lymphoid cells. However, the GFP-signal from neutrophils (Fig 2C), macrophages (Fig 2D), and Siglec-F + (Fig 2E) cells was significantly higher than the ICCS-derived signal. Representative dot-plots for all cell types and organs are provided in S3 Fig. Furthermore, the additional staining with a secondary αGFP-AF488-antibody was not able to improve the specificity of the IL-10-signal (S4 Fig). Interestingly, the GFP-signal of monocytes (Fig 2F) and ILCs (Fig 2G) was substantially higher only in the lung tissues, indicating that the increase in the GFP-channel autofluorescence is organ-specific for some cell types.

PLOS ONE
The IL-10 GFP strain is not suitable for the IL-10 detection by granulocytes In conclusion, our data indicate that the IL-10 GFP (VeRT-X) reporter strain is not suitable to analyse IL-10 production of myeloid cells during inflammation, due to the strong increase in the autofluorescence, which masks the IL-10-specific GFP-signal. Although we only analysed lung, spleen, and mLN of the VeRT-X IL-10 GFP reporter strain, it appears likely that this problem will also be relevant to other organs. Importantly, the strong increase of the granulocytic autofluorescence in the GFP-channels was independent of the strain analysed. Therefore, it is expected that other fluorescent reporter lines that utilize the GFP-channel would face similar problems at distinguishing the reporter-specific signal from the autofluorescence signal when analysing granulocytes. Our data indicate that granulocytes are the source of the background GFP-signal, in line with previous publications reporting high levels of autofluorescence [11], although other sources, like collagen deposition [12,13], cannot be excluded. This large increase in false-positive GFP-signals for myeloid cells during inflammation could have been missed previously due to a focus on lymphoid cells or due to a lack of the WT-treated controls. Our data illustrate a general and important technical caveat using GFP-reporter lines

PLOS ONE
The IL-10 GFP strain is not suitable for the IL-10 detection by granulocytes for the analysis of myeloid cells and suggest that previous reports on effector functions of myeloid cells using such GFP-based reporters might require re-evaluation.

Mice
C57BL/6J mice and IL-10 GFP (also known as Vert-X) B6(Cg)-Il10 tm1.1Karp /J mice were originally obtained from Jackson Laboratory (Bar Harbor, USA). All mice were housed in the vivarium of the Izmir Biomedicine and Genome Center (IBG, Izmir, Turkey) in accordance with the respective institutional animal care committee guidelines. All mouse experiments were performed with prior approval by the institutional ethic committee ('Izmir Biomedicine and Genome Center's Ethical Committee on Animal Experimentation'), in accordance with national laws and policies. All the methods were carried out in accordance with the approved guidelines and regulations.

Reagents and monoclonal antibodies
Monoclonal antibodies against the following mouse antigens were used in this study: CD3ε Details on the antibody used in this study are given in the S1 Table. Anti-mouse CD16/32 antibody (2.4G2) and Zombie UV Dead Cell Staining Kit were obtained from Tonbo Biosciences (San Diego, USA) and from BioLegend, respectively. Unconjugated mouse and rat IgG antibodies were purchased from Jackson ImmunoResearch (West Grove, USA).

Cell preparation
Lungs were removed and minced into smaller pieces in a 6-well plate (Greiner, Germany). The digestion mixture, composed of 1 mg/mL collagenase D and 0.1 mg/mL DNase I (both from Roche, Switzerland) in complete RPMI medium (Gibco, USA), was added to the samples and incubated for 45 min at 37˚C on a lateral shaker. The lung samples were filtered through 100 μm mesh with PBS, washed twice, and the red blood cells were eliminated by ACK lysis buffer (Lonza, USA). Spleens and mediastinal lymph nodes were homogenized by filtering through a 76 μm mesh with ice-cold PBS (Lonza), washed twice, and red blood cells were eliminated by ACK lysis buffer (Lonza).

Flow cytometry
Flow cytometry was performed as described [14]. In brief, for staining of cell surface molecules, cells were suspended in staining buffer (PBS, 1% BSA, 0.01% NaN 3 ) and stained with fluorochrome-conjugated antibodies (0.1-1 μg/10 6 cells, or according to the manufacturer's recommendations) for 15 min in a total volume of 50 μl at 4˚C for 30 min. FcεR-blocking antibody αCD16/32 (2.4G2) and unconjugated rat and mouse IgG (Jackson ImmunoResearch) were added to prevent non-specific binding. If biotin-conjugated antibodies were used, cellbound antibodies were detected with streptavidin conjugates (1:200, or according to the manufacturer's recommendations) in a second incubation step. Dead cells were labelled with a commercially available Zombie UV Dead Cell Staining Kit (BioLegend). For the analysis of intracellular cytokines, cells were fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences) for 10 min at 37˚C. Cells were washed twice and incubated overnight at 4˚C with the fluorochrome-conjugated antibodies and unconjugated rat and mouse IgG in Perm/Wash solution (BD Biosciences), which was followed by additional 5 min incubation in Perm/Wash solution without antibodies. For the analysis of the GFP signal, the cells were fixed with freshly prepared 2% formaldehyde solution for 40 minutes on ice as described [15]. This fixation allows the detection of the GFP-signal together with other intracellular proteins [15]. Then, the cells were permeabilized using the BD Perm/Wash solution. Cells were washed twice and incubated overnight at 4˚C with the fluorochrome-conjugated antibodies and unconjugated rat and mouse IgG in Perm/Wash solution, which was followed by additional 5 min incubation in Perm/Wash solution without antibodies. Cells were analysed with FACSCanto or LSR-Fortessa (BD Biosciences), and data were processed with CellQuest Pro (BD Biosciences) or Flow Jo (Tree Star) software.

Statistical analysis
Data are presented as mean ± standard error of the mean (SEM). The statistical analysis was performed with GraphPad Prism 7.0 software (GraphPad Software, San Diego, USA). Oneway ANOVA followed by Holm-Sidak posthoc test are used to compare p-values regarded as � p�0.05, �� p�0.01, and ��� p�0.001.