A METHOD FOR DETERMINING MYELOID NATURAL KILLER (NK)-CELLS AND USE THEREOF

Abstract

The present invention relates to ex-vivo methods for determining myeloid NK-cells, methods for diagnosis of a disease associated with and/or caused by myeloid NK-cells as well as depletion of myeloid NK-cells for use in treating. The present is also related to methods for determining whether a candidate agent reduces a myeloid NK-cell population.

Claims

1. An ex-vivo method for determining myeloid NK-cells comprising the steps of: a) providing a sample containing myeloid NK-cells; b) marking the myeloid NK-cells of the sample by means of at least one marking reagent; and c) detecting the myeloid NK-cells, wherein the myeloid NK-cells are characterized by the expression of cell surface marker phenotypical for NK-cells and by an expression of Il6ra.

2. The method of claim 1, further comprising a step b) after the step b) b) separating the marked myeloid NK-cells from the sample.

3. The method of claim 1, further comprising a step d) after the step c): d) quantifying the marked myeloid NK-cells.

4. The method of claim 1, wherein the myeloid NK-cells are mature myeloid NK-cells.

5. The method of claim 1, wherein the myeloid NK-cells are further characterized by an expression of cell Csf11r.

6. The method of claim 1, wherein the myeloid NK-cells are further characterized by an activation of Stat3.

7. The method of claim 1, wherein the myeloid NK-cells are characterized by upregulation Il6ra and of at least 50% of the genes selected from the following group consisting of Pla2g7, Fos, Csf1r, Cd93, Mpegl, Cybb, Ctss, Spi1, Cd74, Plbd1, Cd14, Clec10a, Il1rn, Sirpa, Pid1, Ptafr, Ly86, Grn, Tgfbi, Ctsh, C1qc, C1qb, Mrc1, Lrp1, Csf2ra, Ncf1, Cxcl9, Cd302, Cd300lb, Nfam1, Trem2, Emilin2, App, Sdc3, Ifi30, Csf2rb, Igsf6, Marcks, Ctsb, Cst3, Hp, Cfp, Lgals3, Cd300ld, Ifngr2, Rasgrp4, Scpep1, Fgd4, Basp1, Ctsz, Slc11a1, Clec12a, Gm2a, Adap2, Msrb1, Trib1, Msr1, Il1b, KIf4, Hck, Acer3, Plekho1, Mafb, Ciita, Axl, Adam15, Mef2c, Cebpb, Cebpa, Dusp1, Ext1, C1qa, Unc93b1, Naaa, Tmem86a, Lst1, Atf3, Ptpro, Nav1, Pld4, Tlr1, Pou2f2, Lacc1, Themis2, Ccdc109b, Ms4a7, and Rassf4.

8. The method of claim 1, wherein the myeloid NK-cells are characterized by upregulation Il6ra and Csf1r and of at least 50% of the genes selected from the following group consisting of Pla2g7, Fos, Cd93, Mpegl, Cybb, Ctss, Spi1, Cd74, Plbd1, Cd14, Clec10a, Il1rn, Sirpa, Pid1, Ptafr, Ly86, Grn, Tgfbi, Ctsh, C1qc, C1qb, Mrc1, Lrp1, Csf2ra, Ncf1, Cxcl9, Cd302, Cd300lb, Nfam1, Trem2, Emilin2, App, Sdc3, Ifi30, Csf2rb, Igsf6, Marcks, Ctsb, Cst3, Hp, Cfp, Lgals3, Cd300ld, Ifngr2, Rasgrp4, Scpep1, Fgd4, Basp1, Ctsz, Slc11a1, Clec12a, Gm2a, Adap2, Msrb1, Trib1, Msr1, Il1b, KIf4, Hck, Acer3, Plekho1, Mafb, Ciita, Axl, Adam15, Mef2c, Cebpb, Cebpa, Dusp1, Ext1, C1qa, Unc93b1, Naaa, Tmem86a, Lst1, Atf3, Ptpro, Nav1, Pld4, Tlr1, Pou2f2, Lacc1, Themis2, Ccdc109b, Ms4a7, and Rassf4.

9. The method of claim 1, wherein the myeloid NK-cells are characterized by upregulation Il6ra and Csf1r and of at least 50% of the genes selected from the following group consisting of Pla2g7, Fos, Cd93, Mpegl, Cybb, Ctss, Spi1, Cd74, Plbd1, Cd14, Clec10a, Il1rn, Sirpa, Pid1, Ptafr, Ly86, Grn, Tgfbi, Ctsh, C1qc, C1qb, Mrc1, Lrp1, Csf2ra, Ncf1, Cxcl9, Cd302, Cd300lb, Nfam1, Trem2, Emilin2, App, Sdc3, Ifi30, Csf2rb, Igsf6, Marcks, Ctsb, Cst3, Hp, Cfp, Lgals3, Cd300ld, Ifngr2, Rasgrp4, Scpep1, Fgd4, Basp1, Ctsz, Slc11a1, Clec12a, Gm2a, Adap2, Msrb1, Trib1, Msr1, Il1b, KIf4, Hck, Acer3, Plekho1, Mafb, Ciita, Axl, Adam15, Mef2c, Cebpb, Cebpa, Dusp1, Ext1, C1qa, Unc93b1, Naaa, Tmem86a, Lst1, Atf3, Ptpro, Nav1, Pld4, Tlr1, Pou2f2, Lacc1, Themis2, Ccdc109b, Ms4a7, and Rassf4, and by an activation of Stat3.

10. Depletion of myeloid NK-cells for use in medicine.

11. Depletion of myeloid NK-cells according to claim 10 for use in the treatment of obesity, insulin resistance, diabetes, autoimmune diseases, cancer, chronic infections or inflammation.

12. A method for diagnosis of a disease associated with and/or caused by myeloid NK-cells comprising the steps: a) providing a sample containing myeloid NK-cells; b) marking the myeloid NK-cells of the sample by means of a marking reagent; and c) detecting the marked myeloid NK-cells, wherein the myeloid NK-cells is characterized by the expression of cell surface marker phenotypical for NK-cells and by an expression of Il6ra.

13. The method of claim 12, further comprising steps d) and e) after the step c): d) quantifying the marked myeloid NK-cells; and e) comparing the quantified value of the marked myeloid differentiated NK-cells with a standard value.

14. The method of claim 12, wherein the disease associated with and/or caused by myeloid NK-cells is selected from a group consisting of obesity, insulin resistance, diabetes, autoimmune diseases, cancer, chronic infections and inflammation.

15. The method of claim 14, wherein the disease associated with and/or caused by myeloid NK-cells is is obesity-induced diabetes.

16. The method of claim 12, further comprising a step f): f) detecting and quantifying STAT3 and p-STAT3 of the sample.

17. A method for determining whether a candidate agent reduces or inhibits a myeloid NK-cell population comprising: a) providing a sample containing the myeloid NK-cell population; b) contacting the sample containing the myeloid NK-cell population with the candidate agent; c) marking the myeloid NK-cell population of the sample by means of a marking reagent; d) detecting the myeloid NK-cell population; and e) determining if the candidate agent reduces or inhibits the myeloid NK-cell population, wherein the myeloid NK-cells is characterized by the expression of cell surface marker phenotypical for NK-cells and by an expression of Il6ra.

18. The method of claim 17, wherein the candidate agent is a small organic non-peptidic molecule, a peptidic compound, a nucleic acid, or a metal complex.

19. The method of claim 17, wherein step e) reads: e) determining if the candidate agent inhibits activity of myeloid NK-cells.

20. The method of claim 17, wherein step e) includes determining if the candidate agent inhibits the activity of at least one protein of a group containing IL6Ra, Csf1r, gp130, JAK1/2, Spi1 and STAT3.

Description

DESCRIPTION OF THE FIGURES

[0315] FIG. 1: NK cells in obese C57/B16 wildtype mice upregulate IL6Ra expression Non-parenchymal cells were isolated from PGAT, blood and liver, and NK cells were defined as CD45.sup.+ CD3.sup. NK1.1.sup.+ Ncr1.sup.+ viable, single cells via flow cytometry.

[0316] A) Absolute numbers of NK cells per organ mass or blood volume are shown from PGAT, blood and liver of B16 wildtype mice after 16 weeks of HFD- or CD-feeding (n=18vs18).

[0317] B) Flow cytometric analysis of IL6Ra expression on NK cells derived from organs of obese versus lean mice after 16 weeks of HFD-feeding. Histograms show representative results. Quantification of IL6Ra.sup.+ NK cells (defined as viable, single CD45.sup.+CD3.sup.NK1.1.sup.+ cells, n=6vs8).

[0318] C) Gene set enrichment analysis of transcriptomes in PGAT derived NK cells of obese versus lean mice. The most significantly enriched gene ontologies are shown.

[0319] D) Quantification of CD25.sup.+, CD69.sup.+ and CCR2.sup.+ NK cells in liver, PGAT and blood (n=6vs8). (Statistics: unpaired 2-sided t-test corrected for multiple testing; n.s., not significant; *p0.05, **p0.01, ***p50.001)

[0320] FIG. 2: CD11b.sup.hi NK-cells are morphological distinct from mature NK-cells and express myeloid as well as activated NK-cell gene sets

[0321] CD11b.sup.hi and mature CD11b.sup.+ NK-cells were FACS-purified from PGAT and blood to perform quantitative gene expression analysis via mRNA deep-sequencing.

[0322] A) Cytomorphology of Pappenheim-stained CD11b.sup.hi and mature NK-cells from blood directly sorted onto glass slides (630-fold magnification) and, lower part, flow-cytometric quantification of the cell size and granularity by median fluorescence intensity (MFI) of forward (FS) and sideward (SS) scatters (n=4vs4) (statistics: unpaired 2-sided t-test; ***p50.001).

[0323] B) Flow cytometric analysis of IL6Ra-expression in mature and CD11b.sup.hi NK-cells derived from PGAT or blood. Dot blots either gated on mature (left) or CD11b.sup.hi (right) NK-cells show representative results. Cumulative quantification shown as percent of IL6Ra.sup.+ NK-cells of the respective parental population (n=4) (statistics: unpaired 2-sided t-test with Bonferroni multiple comparison correction; ***p50.001).

[0324] C) Exemplary flow cytometry zebra blots of Csf1r expression gated either on mature, i.e. CD11b.sup.+, (left) or CD11b.sup.hi (right) NK-cells.

[0325] FIG. 3: Single cell sequencing and transcriptome analysis identifies formation of a distinct, obesity-associated NK-cell population in adipose tissue of obese mice.

[0326] Single, viable CD45.sup.+CD3.sup.NK1.1 CD11b.sup.+ NK-cells were isolated from PGAT of lean or obese wildtype C57/B16 mice and a total of 768 single cells were individually sorted.

[0327] A) Unsupervised cluster analysis (tSNE map) of individually cells based on their gene expression patterns results in three clusters (blue=cluster 1; green=cluster 2; purple=cluster 3).

[0328] B) Alignment of a published gene signature (Bezman et al., Nat. Immunol 2012) of late activated NK-cells (expression in log scale).

[0329] C) Alignment of a publically available signature (www.immgen.org) of genes expressed in macrophages and

[0330] D) alignment of genes up-regulated in bulk sequenced NK-cells from adipose tissue of obese mice (expression in log scale).

[0331] E) StemID analysis based on the same tSNE map as shown in A), which only shows 3 clusters, however the outlier cells in the clusters (the ones, which fit the least into the cluster) get their own cluster (cluster 5 and 4). StemID runs on the dataset including these outlier clusters. The black lines reflect a minimum spanning tree for the cluster medoids.

[0332] F) Lineage tree analysis shows the projections of all cells in t-SNE space. The color of the links indicates the log 10p-value of the link and the color of the vertices indicates the delta-entropy. The width of the connections in the right panel indicates the link score. Line width corresponds to link score multiplied by 5.

[0333] FIG. 4: Circulating IL6Ra.sup.+ NK-cells are detectable in obese humans and correlate with markers of metaflammation.

[0334] NK-cells defined as viable, single CD45.sup.+ lineage (lin) negative (CD3-CD14-CD19-) CD56.sup.+ cells were analyzed in peripheral blood mononuclear cells (PBMC) of lean (ctrl.) and obese humans by flow cytometry.

[0335] A) Representative flow cytometry dot plots of classical CD56.sup.bright and CD56.sup.dim NK-cell subpopulations in a ctrl. and obese individual. Quantification of CD56.sup.bright and CD56.sup.dim NK-cell subpopulations within the entire study cohort (ctrl. n=16; obese n=14).

[0336] B) Analysis of IL6Ra expression in CD3.sup.+ CD56.sup.+ NK-cells exemplarily shown in a control and an obese individual. Quantification of IL6Ra expression as median fluorescence intensity (MFI) in CD3.sup. CD56.sup.+ NK-cells.

[0337] C) Relative numbers of IL6Ra.sup.+ NK-cells (in % of CD3.sup. CD56.sup.+ cells).

[0338] D-F) Clinical data and blood samples from obese (n=14) and lean (n=16) subjects were prospectively collected and comprised

[0339] D) the body-mass-index,

[0340] E) the homeostatic model assessment of insulin resistance (HOMA-IR) as a measure of insulin sensitivity, and

[0341] F) plasma concentrations of high-sensitivity C-reactive protein (hsCRP) as a measure of systemic low-grade inflammation.

[0342] G) Linear regression analysis of hsCRP levels and the number of circulating IL6Ra.sup.+ NK-cells in each lean and obese individual.

[0343] FIG. 5: Depletion of CD11b.sup.hi NK-cells reduces obesity and improves metabolism in HFD fed mice

[0344] A) Breeding scheme to generate transgenic NK.sup.Csf1r_DTR mice, in which during diet-induced obesity (DIO) the human diphtheria toxin receptor (DTR) is expressed in myeloid-gene expressing NK (myNK) cells under the control of the Csf1r gene promotor.

[0345] B) Experimental layout of longitudinal depletion of myNK-cells via intraperitoneal diphtheria toxin (DT) injections at a dose of 5ng/g body weight (BW) every 5 days.

[0346] C) Body weight curves of NK.sup.Csf1r_DTR (n=7) and NK.sup.fOx littermate control (n=8) mice subjected to repetitive DT injections starting two weeks after the beginning of HFD-feeding (statistics: 2-way ANOVA with Sidak's multiple comparison test).

[0347] D) After two weeks of HFD-feeding but before DT injections NK.sup.Csf1r_DTR (n=7) and NK.sup.f OX littermate control (n=8) mice were analyzed by insulin tolerance tests (ITT) and glucose tolerance tests (GTT).

[0348] E) ITT and GTT analyses were repeated at end of the study in the same cohort of mice which then had received repetitive DT-injections (statistics: 2-way ANOVA with Sidak's multiple comparison test, *p0.05, **p0.01).

[0349] F) Body composition of lean and fat masses determined by computed tomography was analyzed at the end of the 14-week HFD-feeding period in DT-injected NK.sup.Csf1r_DTR (n=7) and NK.sup.flox littermate control (n=8) mice.

[0350] G) Flow cytometry analyses of NK-cells and CD11b expressing subpopulations from liver, PGAT and blood in DT-injected NK.sup.Csf1r_DTR and NK.sup.flox mice at the end of the DT-mediated depletion experiment (statistics: 2-way ANOVA with Sidak's multiple comparison test, *p50.05, **p50.01; n.s. not significant). H) Cytokine levels were determined in circulation from DT-injected NK.sup.Csf1r_DTR and NK.sup.fOx mice at the end of the experiment (statistics: unpaired, two-sided t-test, *p50.05).

[0351] H) Cytokine levels were determined in circulation from DT-injected NKCsf1r_DTR 632 and NKflox mice at the end of the experiment (statistics: unpaired, two-sided t-test, *p0.05).

[0352] FIG. 6: Abrogation of IL-6R-signaling from NK-cells of mice prevents diet-induced obesity

[0353] A) Quantification of CD11b.sup.hi NK-cells (gated on CD3.sup.NK1.1.sup.+NK-cells) in liver, PGAT and blood of IL6Ra.sup.fl/fl (n=9) and IL6Ra.sup.NK (n=11) mice after 16 weeks of HFD feeding (statistics: 2-way ANOVA with Sidak's multiple comparison test; *p50.05; n.s. not significant).

[0354] B) Glucose (GTT) and insulin (ITT) tolerance tests were performed in IL6Ra.sup.fl/fl (n=7) and IL6Ra.sup.NK (n=8) mice after 16 weeks of HFD-feeding (statistics: 2-way ANOVA with Sidak's multiple comparison test; *p0.05, **p0.01).

[0355] C) Hepatic glucose production at the basal and clamped state and organ specific glucose uptake rates were determined in the same cohort of mice.

[0356] FIG. 7: NK-cell-specific disruption of Stat3 protects from obesity and CD11b.sup.hi NK-cell formation

[0357] A) Body weights were assessed over a period of 12 weeks in HFD-fed mice lacking Stat3 expression in NK-cells (Stat3.sup.NK) and littermate controls (Stat3.sup.fl/fl) (n=9vs9) (statistics: 2-way ANOVA with Sidak's multiple comparison test).

[0358] B) Body composition after 12 weeks of HFD-feeding was analyzed in Stat3.sup.fl/fl (n=9) and Stat3.sup.NK (n=9) mice by CT scans (statistics: unpaired, two-sided t-test, *p0.05).

[0359] C+D) Glucose (GTT) and insulin (ITT) tolerance tests were performed in Stat3.sup.fl/fl and Stat3.sup.NK mice (n=7vs6) E) prior to HFD-feeding and F) after 12 weeks of HFD-feeding (statistics: 2-way ANOVA with Sidak's multiple comparison test; *p0.05, **p0.01).

[0360] E) Representative dot blots of flow cytometry CD11b/CD27 expression analysis in PGAT derived NK-cells after 12 weeks of HFD-feeding in Stat3.sup.fl/fl and Stat3.sup.NK mice. Quantification of CD11b.sup.hi NK-cells in liver, PGAT and blood of Stat3.sup.fl/fl and Stat3.sup.NK mice (n=3vs4) and of the classic NK-cell subpopulations based on CD11b/CD27 expression after 12 weeks of HFD-feeding (statistics: 2-way ANOVA with Sidak's multiple comparison test; *p50.05; n.s. not significant).

[0361] FIG. 8. Analysis of immune cell infiltrates in obese C57BL/6 wildtype mice A) Body weight analysis weekly assessed in BI/6 wildtype mice during 16 weeks of HFD- or CD-feeding (n=18vs18) (statistics: 2-way ANOVA with Sidak's multiple comparison test).

[0362] B) Organ mass analyses of PGAT and livers in these animals (n=18vs18) after 16 weeks diet (statistics: unpaired, 2-sided t-test; **p50.01, ***p50.001).

[0363] C+D) Immune cell infiltration of C) CD45.sup.+ leucocytes and D) CD3.sup.+ T cells in PGAT and livers quantified per gram of tissue weight (statistics: unpaired, 2-sided t-test; *p50.05).

[0364] E) Distribution of NK-cell maturation states determined by CD11b/CD27 expression (CD11b.sup.CD27.sup. immature, CD11b.sup.CD27.sup.+ intermediate 1, CD11b.sup.+CD27.sup.+ intermediate 2 and CD11b.sup.+CD27.sup. mature) and

[0365] FIG. 9. Comparison of differentially regulated genes in CD11b.sup.hi NK-cell to gene sets of innate lymphoid cell populations

[0366] A total of 1137 genes have been identified to be specifically regulated in CD11b.sup.hi NK-cells compared to mature CD11b.sup.+ NK-cells. FIG. 9 shows a table containing most important, significantly regulated genes (cut-off p50.05; fold-change 2).

[0367] FIG. 10. Expression analyses of Ncr1-Cre recombinase in NK, T and B cells

[0368] To examine the immune cell subpopulations in which the Ncr1.sup.Cre-transgene is expressed, a fluorescent reporter mouse strain, tdTomato.sup.loxSTOPlox mice (stock no. 007905, Jackson Lab), was crossed with Ncr1Cre mice leading to tdTomato-fluorescence upon cre-mediated recombination.

[0369] A) Representative contour flow cytometry plots of tdTomato-expression in PBMC of Ncr1.sup.Cre+tdTomato.sup.loxsTOPlox mice further stained for NK-cells (NK1.1 and Ncr1), T cells (CD3) and B cells (CD19).

[0370] FIG. 11. Abrogation of IL6Ra signaling in NK-cells attenuates obesity in HFD-fed mice

[0371] A) Body weight development was assessed weekly over a period of 16 weeks in transgenic mice lacking IL6Ra expression in NK-cells (IL6Ra.sup.NK) and littermate control mice (IL6Ra.sup.fl/fl), which were either subjected to high-fat-diet (HFD) (IL6Ra.sup.fl/fl n=10; IL6Ra.sup.NK n=13) or control diet (CD) (IL6Ra.sup.fl/fl n=10; IL6Ra.sup.NK n=6) feeding (statistics: 2-way ANOVA with Sidak's multiple comparison test).

[0372] B) Organ masses after 16 weeks of HFD-feeding (IL6Ra.sup.fl/fl n=10; IL6Ra.sup.NK n=13).

[0373] C) Body fat content of IL6Ra.sup.fl/fl (n=10) and IL6Ra.sup.NK (n=8) after 16 weeks of HFD was determined by NMR analysis (statistics: unpaired, two-sided t-test, *p0.05).

[0374] FIG. 12. NK-cell specific knockout of the IL6Ra attenuates metaflammation in PGAT

[0375] To assess the level of metaflammation, the levels of cellular infiltration were analyzed by FACS of

[0376] A) CD45.sup.+ leucocytes, B) CD3.sup.+ T cells and C) CD3.sup.NK1.1.sup.+NK-cells in liver and PGAT of IL6Rafl/fl (n=9) and IL6Ra.sup.NK (n=11) mice after 16 weeks of HFD-feeding. Given numbers are cells per tissue mass. (Statistics: unpaired, two-sided t-test; *p0.05, **p0.01, ***p0.001)

[0377] D) NK-cell maturation states were analyzed in these mice by CD11b/CD27 staining and flow cytometry. (Statistics: 2-way ANOVA with Sidak's multiple comparison test; n.s. not significant, *p0.05, **p50.01, ***p0.001).

[0378] FIG. 13. Endometrium Carcinoma

[0379] Carcinoma in the endometrium associated with myeloid NK-cells is shown. A magnified view of the said carcinoma can be seen in the lower right section of the figure.

[0380] FIG. 14. Mamma Carcinoma

[0381] Mamma carcinoma associated with myeloid NK-cells is shown. A magnified view of the said carcinoma can be seen in the lower right section of the figure.

[0382] FIG. 15. Lymph Node Metastasis (breast cancer, obese male)

[0383] Lymph node metastasis in case of breast cancer of an obese male associated with myeloid NK-cells is shown. A magnified view of the said metastasis can be seen in the lower right section of the figure.

[0384] FIG. 16. Colon Carcinoma

[0385] Carcinoma in the colon associated with myeloid NK-cells is shown on the left side of the figure. A magnified view of the said carcinoma can be seen on the right side of the figure.

[0386] FIG. 17. Pancreatic Carcinoma

[0387] Carcinoma in the pancreas associated with myeloid NK-cells is shown. A magnified view of the said carcinoma can be seen on the right side of the figure.

[0388] FIG. 18. Gastric Carcinoma

[0389] Carcinoma in the gastric associated with myeloid NK-cells is shown. A magnified view of the said carcinoma can be seen in lower right section of the figure.

[0390] FIG. 19. Esophagus Carcinoma

[0391] Carcinoma in the pancreas associated with myeloid NK-cells is shown. A magnified view of the said carcinoma can be seen on the lower left section of the figure.

[0392] FIG. 20. Prostate Carcinoma

[0393] Carcinoma in the prostate associated with myeloid NK-cells is shown on the left side of the figure. A magnified view of the said carcinoma can be seen on the right side of the figure.

[0394] FIG. 21. Hepatocellular Carcinoma

[0395] Hepatocellular carcinoma associated with myeloid NK-cells is shown on the left side of the figure. A magnified view of the said carcinoma can be seen on the right side of the figure.

[0396] FIG. 22. Multiple Myeloma

[0397] Multiple Myeloma associated with myeloid NK-cells is shown on the left side of the figure. A magnified view of the said carcinoma can be seen on the right side of the figure. The arrows are directing to myeloma.

EXAMPLES

Abbreviations

[0398] CD-feeding Control diet feeding [0399] DAB 3, 3-diaminobenzidine, (chromogen substrate) [0400] DTR Diphtheria toxin receptor [0401] EtOH Ethanol [0402] FPKM Metric of RNA sequencing, meaning: fragments per kilobase of transcript per million mapped reads [0403] HFD-feeding High-fat diet feeding [0404] HOMA-IR Homeostatic Model Assessment for Insulin Resistance [0405] HRP Horseradish peroxidase [0406] hsCRP high-sensitivity C-reactive protein [0407] moAB Monoclonal antibody [0408] MPwater Micropure (i.e. distilled) water [0409] PGAT perigonadal adipose tissue [0410] RNAscope Patented technology of RNA in-situ detection (ACD Bio-techne) [0411] SCAT Subcutaneous adipose tissue

[0412] Material and Methods

[0413] Animal Care and Generation of Transgenic Mouse Strains

[0414] All mouse experiments were approved by the local authorities (Bezirksregierung Kln; Germany) and conducted in accordance with NIH guidelines. Mice were housed in groups of 3-5 animals at 22-24 C. and a 12-hour light/dark cycle. Animals had ad libitum access to food and water at all times, and food was only withdrawn if required for an experiment. In general, experiments started with 6-week old animals.

[0415] C57Bl/6JR wild-type mice were purchased from Janvier (Janvier Labs, France) at the age of 4 weeks and acclimatized to our facility over two weeks prior to the start of experiments. NK-cell specific transgenic mouse models, used in this study, were generated by crossing previously published mouse strains all of which have been backcrossed to C57BL/6 mice for at least ten generations. Briefly, mice with a conditional knockout of the IL6Ra gene in NK-cells (IL6Ra.sup.NK mice) were generated by crossing Ncr1.sup.Cre mice (kindly provided by Dr. Emilio Casanova, Ludwig Boltzman Institute for Cancer Research, Vienna, Austria) with IL6Ra-floxed mice and line breeding was maintained with Ncr1.sup.Cre+/ IL6Ra.sup.fl/fl crossed with Ncr1.sup.Cre/ IL6Ra.sup.fl/fl mice.

[0416] Mice in which NK-cells are specifically enabled to express DTR upon activation of the Csf1r-gene (NK.sup.Csf1r/DTR mice), a myeloid marker gene, were generated by crossing Ncr1.sup.Cre+/ mice with heterozygous Csf1r.sup.LoxStopLox-DTR mice (3) (kindly provided by Dr. Ana Domingos, Instituto Gulbenkian de Cincia, Oeiras, Portugal). An NK-cell specific knockout of the Stat3 gene was generated by crossing Ncr1.sup.Cre+/ mice with Stat3-floxed mice and line breeding was maintained with Ncr1.sup.Cre+/ Stat3.sup.fl/fl crossed with Ncr1.sup.Cre/ Stat3.sup.fl/fl mice.

[0417] Diet Induced Obesity, High Fat Diet Feeding

[0418] For all feeding experiments the same conditions were used. Animals at the age of six weeks were either fed a high-fat diet (HFD) containing 60% calories from fat, 21% calories from carbohydrates and 19% calories from protein (ssniff D12492 (1) mod., ssniff Spezialditen GmbH, Germany) or a low-fat control diet containing 13% calories from fat, 67% calories from carbohydrates and 20% calories from protein (ssniff D12450B mod. LS, ssniff Spezialditen GmbH).

[0419] Diphtheria Toxin Mediated Cell Ablation

[0420] In order to deplete myeloid (CD11b.sup.hi) NK-cells, intra-peritoneal injections of diphtheria toxin (Sigma Aldrich) at a dose of 5ng per gram body weight every five days were given to NK.sup.Csf1r/DTR-mice and littermate controls.

[0421] Human Samples

[0422] After obtaining informed consent, anthropometric data were prospectively assessed and collected blood samples from obese and lean individuals and cryopreserved peripheral blood mononuclear cells (PBMCs) immediately after Ficoll density gradient centrifugation and washing steps.

[0423] Plasma samples were cryopreserved in liquid nitrogen immediately after centrifugation (15,000g, 10 minutes). The study (NK-ADIPO) was approved by the institutional review board (No. 15-042, Medical Faculty, University of Cologne, Cologne, Germany).

[0424] Immune Cell Isolation and Flow Cytometry

[0425] Leucocytes were isolated from adipose tissue, liver and peripheral blood according to modified protocols published elsewhere (Miltenyi Biotech). Briefly, organs were dissociated with a tissuelyser (GentleMACS, Miltenyi) and digested enzymatically for 20 min at 37 C. while continuous shaking: adipose tissue: type I collagenase 500 U/ml; DNAse1 150 U/ml. liver: type IV collagenase 500 U/ml; DNAse1 150 U/ml (all from Worthington, Lakewood, USA). Immune cells were separated from stromal cells by centrifugation: adipose tissue homogenates 400g for 5 minutes and liver homogenates via 20%-histodenz (Sigma Aldrich) density-gradientcentrifugation. Leucocytes from blood samples were generated by standard erythrocyte lysis in ammonium chloride solution (eBioscience) for 5-10 minutes. Finally, cell suspensions were resuspended in FACS buffer (MACS-Buffer, Miltenyi Biotech) and passed through a 40 m strainer (BD Biosciences) to remove large cellular debris.

[0426] For flow cytometry analysis, mouse (m) or human (h) leucocytes were stained after FC-blocking with either anti-mouse CD16/32 or human-IgG (Trustain, Biolegend) followed by fixable dead cell staining (LIVE/DEAD, Invitrogen). Directly fluorochrome-conjugated anti-mouse or anti-human antibodies or the respective isotype control were used for specific immunostainings (1:50-100 dilutions, all from Biolegend, San Diego, if not otherwise indicated):

[0427] mCD45-BV510 (30F11), mCD3-PacificBlue (17A2), mNK1.1-A700 (PK136), m/hCD27-PerCP-Cy5.5 or -PE-Cy7 (LG.3A10), m/hCD11b-PE-TexR (M1/70.15; Invitrogen) or m/hCD11b-APC or -APC-Cy7 (M1/70), mCD25-PE (7D4; Miltenyi Biotech), mCD69-PE-Cy7 (H1.2F3), mCCR2-FITC (FAB5538F, R&D Systems), mNKG2D-PE (CX5), mNKG2A-APC (16A11), mLy49D-AF647 (4E5), mLy49H-PE (3D10), mLy49C/I-Fitc (5E6; BD Pharmigen), mNKp46-Fitc or -PE or -BV421 (29A1.4), mKLRG1-PE-Cy7 (2F1/KLRG1), mCsf1r-PE or -Fitc (AFS98), mIL6Ra-PE or -APC (D7715A7), hCD3-PacificBlue (OKT3), hCD16-Fitc (3G8), hCD56-PercP-Cy5.5 (HCD56), hIL6Ra-APC (UV4).

[0428] Cells were analyzed using an 8-color flow cytometer (MACSquant-10, Miltenyi Biotec) or a 10-color flow cytometer (Gallios, Beckman Coulter) and respective data analysis was performed with FlowJo (Treestar) or Kaluza (Beckman Coulter) software packages in the recent versions. In all analyses, the first gate identified lymphocytes by forward and sideward scatter followed by exclusion of doublets. Analysis of NK-cells, defined as CD3 NK1.1.sup.+ and/or Ncr1.sup.+ (for murine cells) and CD3.sup.CD19.sup.CD14.sup. CD16.sup.+ and/or CD56.sup.+ (for human cells), was always based on viable (dead stain negative) CD45.sup.+ cells.

[0429] Cell Sorting

[0430] NK-cell subpopulations from murine or human samples were purified from single cell suspensions using FACSAria-11l or FACSAria-Fusion cell sorters (BD Bioscience) after immunostainings as described above. To sort mouse CD11b.sup.hi and CD11b.sup.+ mature NK-cells from organs and blood, gates were set on single, viable CD45.sup.+CD3.sup.NK1.1.sup.+ lymphocytes. To sort human CD11b.sup.hiIL6Ra.sup.+ versus CD11b.sup.+IL6Ra.sup. NK-cells from blood, gates were set on single, viable CD3-cells single or double positive for CD16 and CD56. Purified cells were directly sorted into RNA-protect cell reagent (Qiagen, Germany).

[0431] RNA Isolation from Purified Cells and mRNA Sequencing

[0432] Stabilized, purified NK-cell populations were pelleted by centrifugation and total RNA was extracted using the Arcturus RNA picopure Kit (KIT0204, ThermoFisher Scientific) following the manufacturer's instructions. RNA integrity was assessed with the Agilent 2100 Bioanalyzer.

[0433] RNA libraries were prepared from a minimum of 100ng total RNA using the TruSeq RNA sample preparation Kit v2 (Illumina). Complementary DNA (cDNA) was transcribed from poly-A selected RNA, which served for library generation. Libraries were sequenced in replicates for 30 million reads on an Illumina HiSeq 2000 sequencer with a paired-end (1017101 cycles) protocol.

[0434] Single Cell RNA-Seq Analysis

[0435] Read 2 of the read pair was first 3 trimmed for adapters, base quality and poly-A tails using cutadapt 1.9.1 (http://dx.doi.org/10.14806/ej.17.1.200). Remaining reads were mapped to the mouse genome GRCm38 (primary assembly) using STAR-2.5.2b (https://www.ncbi.nlm.nih.gov/pubmed/23104886). Gene models were used according to gencode version M9 (https://www.ncbi.nlm.nih.gov/pubmed/26187010). Gene summarization was done using feature Counts 1.5.0-p1 (Liao et al., 2014) collapsing exons to genes and excluding pseudogenes and transcripts with a biotype related to decay. Multimapping reads were discarded. Cell and gene demultiplexing was done using the cell-barcode and the unique molecular identifier (UMI) present in the first 12 nt of Read 1 of the read pair.

[0436] Data analysis was performed using RacelD2 and StemID algorithm (Grin et al., 2016). Downsampling to 800 transcripts was used for data normalization. K-medoids clustering was performed using log-pearson correlation as a distance metric. The minimum suitable cluster number (=3) characterizing the dataset was determined by computing Jaccard's similarity for each cluster by bootstrapping for k-medoids clustering with different cluster numbers. The minimum number yielding a Jaccard's similarity >0.6 for all clusters was selected. The t-distributed stochastic neighbor embedding (t-SNE) algorithm was used for dimensional reduction and cell cluster visualization (Maaten and Hinton, 2008). RacelD2 was executed with the probability threshold value for outlier identification set to <10-3. The StemID algorithm was used to infer a dedifferentiation trajectory. A p-value threshold of 0.05 was chosen to assign significance to the links.

[0437] Bioinformatic Analysis

[0438] Gene expression analysis was performed at the CECAD bioinformatic core facility (Dr. Peter Frommolt, University of Cologne, Germany) using the publically available QuickNGS software platform (http://athen.cecad.uni-koeln.de/quickngs/web) which integrates well-established algorithms (Tophat2, Cufflinks2, DESeq2 and DEXSeq). Heatmaps and unsupervised cluster analyses were generated from log 2-transformed FPKM values. After filtering the genes according to the variance of their FPKM values across the samples, only the 1000 genes with highest variance were displayed.

[0439] Gene set enrichment and gene ontology analyses of were performed with publically available programs (http://geneontology.org/) and the Ingenuity software package (Qiagen, Germany). Statistical analysis of enriched gene ontologies was corrected for multiple testing and only significant results (p50.05) were considered. Analyses were performed with The Ingenuity software package was also employed to align the CD11b.sup.hi NK-cell gene signature to those of different ILC populations and B220.sup.+ pre-mNK-cells derived from the Gene Expression Omnibus Repository (http://www.ncbi.nlm.nih.gov/geo/). The immunological genome project data base (https://www.immgen.org/) enabled us to align the CD11b.sup.hi NK-cell specific gene set to a large panel of immune cell specific gene sets.

[0440] Analysis of Plasma Samples in Mice and Humans

[0441] Mouse Samples:

[0442] Insulin plasma concentrations were determined by ELISA with mouse standards according to the manufacturer's guidelines (mouse high-sensitivity Insulin ELISA, Crystal Chem, USA). Blood glucose levels were determined from tail vein blood using an automatic glucometer (Bayer Contour, Bayer, Germany). Cytokines (TNF-alpha, IL-1beta, IL-12p70 and GM-CSF) were detected in replicates of undiluted 50 l plasma samples using a multiplex magnetic bead immunoassay (Life Technologies) and quantification was performed on a Bio-Plex 200 reader (BioRad) according to the manufacturer's instructions.

[0443] Human Samples:

[0444] Analyses of insulin, glucose and hsCRP plasma concentrations were performed in the central clinical laboratory of the University Hospital Cologne (Cologne, Germany). HOMA-IR was calculated as previously described [glucose (mg/dl)insulin (mU/I)/405].

[0445] Glucose and Insulin Tolerance Tests

[0446] Glucose tolerance tests (GTT) were performed with 6 h-fasted mice at the indicated age and time of diet-feeding given for each experiment. Blood glucose concentrations (mg/dl) were measured following fasting, prior to the test, and 15, 30, 60 and 120 minutes after intraperitoneal injection of glucose 20% (1.5 mg/g BW) (DeltaSelect). Blood glucose levels were determined from tail vein blood using an automatic glucometer (Bayer Contour, Bayer, Germany).

[0447] Insulin tolerance tests (ITT) were performed with 2 h-fasted mice at corresponding time points to GTTs. Blood glucose concentrations (mg/dl) were measured following fasting, prior to the test, and 15, 30, 60 and 120 minutes after intraperitoneal injection of insulin (0.75 mU/g BW, Insuman rapid, Sanofi Aventis).

[0448] Hyperinsulinemic-Euglycemic Clamp Studies in Awake Mice

[0449] Surgical implantation of catheters into the jugular vein was performed as described. After 5-6 days of recovery, only mice that had lost less than 10% of their preoperative weight were included. Each animal was deprived of food for 4 h in the morning of the experiment. All infusions used in the experiment were prepared with a 3% plasma solution obtained from fasted donor mice. A primed-continuous infusion of tracer d-[3-3H]-glucose was initiated 50 min before the clamping (5 Ci priming at a rate of 0.05 Ci/min; PerkinElmer). After a 50-minute basal period, a blood sample was collected from the tail tip for determination of basal parameters. The clamping began with a primed-continuous insulin (INSUMAN rapid; Sanofi-Aventis) infusion (40 U prime per gram body weight, followed by a continuous rate of 4 U per g body weight per min) and glucose concentrations in blood were measured every 10 min (Bglucose analyzer, Hemocue). Euglycemic serum levels (120-140 mg/dl) were maintained by adjustment of a 20% glucose infusion (DeltaSelect). Approximately 120 min before the end of the experiment, 2-[1-14C]-deoxy-d-glucose (10 Ci; American Radiolabeled Chemicals) was infused, and blood samples were collected until steady state was reached. The steady state was considered, when a fixed glucose-infusion rate maintained the glucose concentration in blood constant for 30 min. During the steady state, blood samples were collected for the measurement of steady-state parameters.

[0450] At the end of the experiment, mice were killed by cervical dislocation, and brain, liver, PGAT, SCAT and skeletal muscle were dissected, snap-frozen in liquid nitrogen and stored at 80 C. The [3-3H] glucose radioactivity of plasma in basal conditions and at steady state was measured as described. The radioactivity of 2-[1-14C]-deoxy-d-glucose in plasma was measured directly in a liquid scintillation counter. Lysates of adipose tissue and skeletal muscle were processed through ion-exchange chromatography columns (AGR1-X8 formate resin, 200-400 mesh dry; Poly-Prep Prefilled Chromatography Columns; Bio Rad Laboratories) for the separation of 2-[1-14C]-deoxy-d-glucose from 2-[1-14C]-deoxy-d-glucose-6-phosphate, 2-[1-14C] (2DG6P).

[0451] The glucose-turnover rate (mgkg.sup.1min.sup.1) was calculated as described (Konner et al., 2007). The uptake of glucose in brain, PGAT, SCAT and skeletal muscle in vivo (nmolg.sup.1min.sup.1) was calculated on the basis of the accumulation of 2-[1-14C]-deoxy-d-glucose-6-phosphate, 2-[1-14C] in the respective tissue and the disappearance rate of 2-[1-14C]-deoxy-dglucose from plasma, as described.

[0452] Body Composition Analysis

[0453] Body weights were assessed weekly. Lean and fat mass were determined using the NMR Analyzer Minispeq mq7.5 (Bruker Optik). Alternatively, body composition was analyzed by computed tomography (CT) in isoflurane-anesthetized mice (Drager and Piramal Healthcare).

[0454] For data acquisition on an IVIS Spectrum CT scanner (Caliper LifeScience, USA) IVIS LivingImage Software V4.3.1 was used. Quantification of lean and fat mass contents were determined with a modification of the previously described Vinci software package 4.61.0 developed at our institution.

[0455] Indirect Calorimetry

[0456] Indirect calorimetry was performed using an open-circuit, indirect calorimetry system (PhenoMaster, TSE systems). Mice were trained for three days before data acquisition to adapt to the food/drink dispenser of the PhenoMaster system. Afterwards mice were placed in regular type II cages with sealed lids at room temperature (22 C.) and allowed to adapt to the chambers for at least 24 hours. Food and water were provided ad libitum. All parameters were measured continuously and simultaneously.

[0457] Immunoblots

[0458] Protein extraction from cryopreserved tissues and SDS-page immunoblot procedures were done as previously described. Membranes were blocked with 5% WB-blocking reagent (Roche, Switzerland) in Tris-buffered saline containing 0.2% Tween-20 (TBS-T), and incubated with primary antibodies at 4 C. overnight. The following primary antibodies and dilutions were used: anti-phosphoAKT.sup.Ser473 (1:1000, #4060, Cell Signaling), anti-panAKT1-3 (1:1000, #9272, Cell Signaling) and, as a loading control, anti-calnexin (1:2000, Cat 208880, Merck Milipore). Quantification of chemiluminescent signals was performed with the ImageJ (Fiji)-software package.

[0459] JNK Kinase Assay

[0460] Analysis of c-jun N-terminal kinase (JNK) activity in white adipose tissue and liver protein lysates was performed with the non-radioactive JNK kinase assays kit (#8794, Cell Signaling) according to the manufacturer's instructions. Briefly, phosphorylated JNK (p-JNK) was immunoprecipitated from tissue lysates with rabbit anti-mouse-p-JNK.sup.Thr183,Tyr185 (#8794; Cell Signaling) coupled to Sepharose beads. In vitro phosphorylation of recombinant c-Jun protein as a JNK-substrate was performed in ATP containing kinase buffer. Finally, the amount of phosphorylated c-Jun.sup.Ser63, as a measure of JNK-activity, was determined by immunoblot analysis. Probes of the initial lysates before immunoprecipitation were used as loading controls.

[0461] Histological Analyses

[0462] Liver and adipose tissues were harvested following mouse scarification, fixed in paraformaldehyde 4% and embedded in paraffin. Hematoxylin and eosin (HE) staining of tissue sections (4-5 m) were performed according to standard procedures. F4/80 immunohistochemical staining was performed with 1:100 diluted primary anti-F4/80 antibody (MCA497G, Serotec) as previously described. F4/80 positive crown-like structures were quantified with a LeicaDM1000 LED microscope (Leica, Germany). Quantitative image analyses were performed with the ImageJ (Fiji) software package (NIH) including the Adiposoft plugin (http://imagej.net/Adiposoft).

[0463] Statistical Analyses

[0464] All data, unless otherwise indicated, are shown as mean valuesstandard error of the mean (SEM). In box-and-whisker plots the upper and lower whiskers indicate the minimum and maximum values of the data, centerlines indicate the median, and the mean is represented by a plus sign. Comparison of two independent groups was performed with unpaired two-tailed Student's t-test. Data sets with more than two groups were analyzed using one-way analysis of variance (ANOVA) followed by Tukey's posthoc test. For statistical analyses of longitudinal data, i.e. body weight curves, GTTs, ITTs and clamp studies (glucose infusion rate), two-way ANOVA was performed corrected by Sidak's multiple comparison test. All figures and statistical analyses were generated using the GraphPad Prism 6 software. P-values 0.05 were considered statistical significant.

Example 1: Obesity Promotes Formation of Myeloid NK-Cells in Mice

[0465] In light of the recently described role of NK-cells in the pathophysiology of obesity and obesity-associated insulin resistance, the inventors analyzed NK-cell development and tissue distribution during the course of obesity development in mice. Mice, which had been fed either a normal chow (CD) or a high-fat diet (HFD) from the age of 6 weeks on, were analyzed. As expected, mice exposed to HFD exhibited a significant, 140% increase in body weight, an almost threefold increase in perigonadal adipose tissue (PGAT) mass and significantly increased liver weight compared to CD-fed animals (FIG. 7A, B). In parallel to progressive obesity during 16 weeks of HFD-feeding, tissue mass related total numbers of CD45.sup.+ immune cells increased in PGAT, but not in liver (FIG. 7C). This was paralleled by increased T-cell numbers in PGAT, a fact previously reported (FIG. 7D). As observed for T cells, also the numbers of NK-cells significantly increased in PGAT of mice upon HFD-feeding (FIG. 1A).

[0466] To further investigate NK-cell maturation in obesity, multicolor flow cytometry analyzing CD11b/CD27 immunoreactivity in (CD45.sup.+NK1.1.sup.+CD3.sup.) NK-cells as well established maturation markers were employed. Here, no consistent major alterations in the proportion of immature, intermediate 1, intermediate 2 and mature NK-cells in circulation or the investigated tissues were detected, neither early nor late during obesity development (FIG. 1B and FIG. 7E). However, in PGAT as well as in the circulation, HFD-feeding provoked the appearance of a previously unidentified NK-cell subpopulation characterized by higher levels of CD11b expression compared to CD11b mature NK-cells (FIG. 1B, C). In addition, increased NK-cell numbers expressing the activation markers CD25 and CD69 as well as the chemokine receptor CCR2, most pronounced in PGAT of HFD-fed mice (FIG. 1D) were detected. In contrast, expression of the activatory and inhibitory NK-cell receptors NKGD2, NKG2A, Ly49D, Ly49H, Ly49C/I, NKp46 (Ncr1) and KLRG1 remained largely unaltered upon obesity development. This allows to conclude that during obesity development, NK-cells are an important immune cell population infiltrating PGAT. Most strikingly, obesity promoted the expansion of a distinct NK-cell population in circulation and PGAT characterized by high level expression of CD11b, referred to as CD11b.sup.hi NK-cells.

Example 2: Further Analysis of Murine CD11b.SUP.hi .NK Subpopulation, Morphological Analysis and Expression Profile

[0467] To further characterize this obesity-associated CD11b.sup.hi NK-cell population, the inventors analyzed morphologic features of flow-sorted CD11b.sup.hi and CD11b (mature) NK-cells (CD45.sup.+NK1.1.sup.+CD3.sup.). CD11b.sup.hi NK-cells were significantly larger and more granulated compared to mature NK-cells, irrespective whether they were isolated from PGAT or circulation (FIG. 2A). Furthermore, cytomorphological analysis of both NK-cell types confirmed the distinct large cell appearance of CD11b.sup.hi NK-cells compared to their mature CD11b.sup.+ counterpart (FIG. 2A).

[0468] To further define CD11b.sup.hi NK-cells at a molecular level, gene expression profiles of flow-sorted CD11b.sup.hi and mature NK-cells isolated from circulation or PGAT were compared by total mRNA deep-sequencing. Unsupervised hierarchical cluster analyses revealed a trunk of similar gene expressions in all probes and a comparable gene expression pattern of mature NK-cells isolated from circulation or PGAT (FIG. 2B). Also, gene expression profiles of CD11b.sup.hi NK-cells isolated from both compartments revealed a high degree of similarity (FIG. 2B). However, when gene expression between CD11b.sup.hi NK and mature NK-cells was analyzed pronounced differences in gene expression signatures were uncovered, with 1675 and 2499 genes increased and 171 and 510 genes decreased in CD11b.sup.hi and mature NK-cells respectively (cut-off: fold change 2; p-value 0.01). A shared signature of 1137 genes differentially expressed in CD11b.sup.hi NK-cells from circulation and PGAT was found to be up-regulated compared to mature NK-cells (cut-off: fold change 2; p-value 0.01). This gene set was subjected to gene ontology analysis, which revealed a significant enrichment of gene sets in CD11b.sup.hi NK-cells affecting cytokine-associated signaling, myeloid differentiation, NK-cell activation and chemotaxis (FIG. 2C).

[0469] Among the most differentially regulated genes were the IL-6 receptor alpha (IL6Ra) and the CSF-1 receptor (Csf1r) with a 10-fold upregulation for IL6Ra and 15-fold upregulation for Csf1r (FIG. 2C). Of note, both genes were previously reported to be either absent or expressed only at low levels in classical NK-cells. Their expression was thus validated in CD11b.sup.hi NK-cells by flow cytometry. While only 9% of mature NK-cells express the IL6Ra, the proportion of IL6Ra.sup.+ cells was increased to 90% in CD11b.sup.hi NK-cells. Similarly, the proportion of Csf1r-expressing NK-cells was significantly increased (3-fold) in the CD11b.sup.hi-population.

[0470] The inventors next related the 1137 genes differentially expressed in CD11.sup.hi NK-cells to other immune cells and aligned this gene signature to publically available gene expression profiles of B cells, T cells, dendritic cells (DCs), monocytes, macrophages, granulocytes and NK-cells. The highest degree of overlap was found with myeloid immune cells (i.e. DCs, monocytes, macrophages, B cells and granulocytes). The overlap with other cell types was less prominent. Of note, the CD11b.sup.hi NK-cell gene signature shared only very limited overlap with gene expression signatures of defined innate lymphoid cell (ILC)-populations or pre-mNK-cells (formerly described as B220.sup.+ NK-cells or interferon producing killer dendritic cells, (IKDC)) (FIG. 8A). However, comparison of upstream regulatory pathways, which resulted from gene set enrichment analyses in CD11b.sup.hi NK-cells and pre-mNK-cells, uncovered potential similarities of both cell types although pre-mNK-cells share only 4.3% of the differentially expressed genes found in CD11b.sup.hi NK-cells (FIG. 8B).

[0471] Collectively these analyses allow to conclude that CD11b.sup.hi NK-cells, which drastically increase in obesity, share highest morphological and molecular similarity with cells of the myeloid lineage, members of which have well-established functions in obesity associated insulin resistance. Therefore, the cells of the newly identified CD11b.sup.hi NK-cell subpopulation is herein called myeloid NK-cells (myeloid NK-cells). The increased IL6Ra expression appears to be a discriminative marker for myeloid NK-cells compared to mature NK-cells.

Example 3: Identification and Investigation of Myeloid NK in Human Subjects

[0472] Having identified increased myeloid NK-cells in PGAT and circulation of obese, insulin resistant mice, the inventors next investigated whether a similar NK-cell population is detectable in obese humans. To address this point, cohorts of lean and obese human subjects were recruited and the number and marker gene expression of NK-cells in circulation was analyzed. Analysis of NK-cell subsets based on CD56 and CD16 expression revealed no differences in circulation between lean and obese human subjects (FIG. 3A). However, when NK-cells were stained for CD11b and CD27 to further define NK-cell maturation states, a significant increase in the proportion of CD11b.sup.hi NK-cells in the circulation of obese versus lean human subjects was observed (FIG. 3B). In line with the findings in mice, NK-cells from obese individuals displayed increased IL6Ra expression and the numbers of CD11b/IL6Ra double-positive NK-cells were significantly increased in the circulation of obese human subjects (FIG. 3C, D).

[0473] To further investigate a possible dynamic regulation of myeloid NK-cells during weight gain and loss in humans, blood samples from human subjects undergoing a controlled weight reduction program (Optifast) were prospectively collected. This program comprised of a 3-month phase of massive weight reduction due to caloric restriction and a weight maintenance phase of another 3 months. Blood samples were taken before (t0) and at the end of the caloric restriction phase (t1) as well as when the individuals reached the steady state phase at the end of the program (t2), which was usually six to eight months after the starting point (t0). The analysis of that samples revealed a significant reduction in circulating myeloid NK-cells upon weight reduction and maintenance in parallel to improved insulin sensitivity measured by HOMA-IR and reduced systemic inflammation as assessed via circulating concentrations of the high-sensitivity C-reactive protein (hsCRP) (FIG. 3E-H). In line with the observations in mice, obese human subjects display increased levels of CD11b.sup.hiIL6Ra.sup.+NK-cells which decrease upon weight loss alongside with improved systemic inflammation and insulin sensitivity.

[0474] Next, the gene expression signatures of murine and human CD11b.sup.hi NK-cells were compared. Therefore, CD11b.sup.hiIL6Ra.sup.+NK-cells and CD11b.sup.+IL6Ra.sup. NK-cells were purified by flow cytometry cell sorting (FACS) (gated on viable CD3.sup. cells excluding CD16.sup.CD56.sup. cells) from the circulation of lean and obese human subjects and mRNA deep-sequencing analyses was performed. Unsupervised hierarchical cluster analysis of differentially expressed genes allowed for separation of gene expression signatures between cell types irrespective of body weight. When the lists of genes upregulated in CD11b.sup.hi NK-cells in humans and mice were compared, 553 genes were found regulated in both species representing an overlap of 48.6% of the genes found in murine CD11b.sup.hi NK-cells. Gene ontology analyses of genes similarly regulated in both species revealed over-representation of myeloid marker genes in human CD11b.sup.hiIL6Ra.sup.+NK-cells versus CD11b.sup.+L6R.sup. NK-cells. Taken together, our analysis confirmed the increased occurrence of a previously undescribed myeloid NK-cell subpopulation in murine and human obesity.

Example 4: Contribution of Myeloid NK-Cells in Obesity and Insulin Resistance in Mice

[0475] To understand the contribution of the CD11b.sup.hi NK-cell population in the development of obesity and/or insulin resistance, the inventors decided to selectively deplete these cells in mice. To this end, they capitalized on the fact, that classical NK-cellsin contrast to myeloid NK-cellslack expression of the Csf1r gene. By crossing Csf1r.sup.loxSTOPlox-DTR-mice to Ncr1.sup.Cre-mice a mouse line was generated, in which NK-cells inducible express the human diphtheria toxin receptor (DTR) only upon Csf1r-gene activation, which enable to specifically deplete Csf1r.sup.+ NK-cells (i.e. CD11b.sup.hi NK-cells) by diphtheria toxin (DT) injections in vivo (FIG. 4A). NK-cell specific expression of the Ncr1.sup.Cre-transgene was confirmed by flow cytometry in PBMC derived from Ncr1.sup.Cre/tdTomato-fluorescent-protein reporter mice (FIG. 9A).

[0476] To investigate the role of Csf1r.sup.+Ncr1.sup.+ (i.e. CD11b.sup.hi NK-cells) in obesity, the inventors subjected NK.sup.Csf1r_DTR mice and the respective control littermates to HFD-feeding from the age of 6 weeks on (FIG. 4B). At the age of 8 weeks, following two weeks of HFD-feeding and prior to DT-application, neither body weight nor insulin and glucose tolerance differed between NK.sup.Csf1r_DTR-mice and floxed controls (FIG. 4C, D). From that time point on, the animals received repetitive intraperitoneal injections of DT (5ng/g body weight) every 5 days (FIG. 4B). Monitoring of body weights during the course of DT-applications revealed that 6 weeks after the beginning of DT-treatment, NK.sup.Csf1r/DTR-mice exhibited reduced body weights compared to DT-treated control mice (FIG. 4C). After 9 weeks, the reduction of body weight reached statistical significance and, at the end of the experiment, DT-treated NK.sup.Csf1r/DTR-mice weighed 15% less than their DT-treated controls (FIG. 4C). Strikingly, at this point DT-treated NK.sup.Csf1r/DTR-mice presented with significantly improved insulin and glucose tolerance (FIG. 4E) and significantly reduced adiposity as assessed by computed tomography (CT) scans (FIG. 4F). Flow cytometry analyses confirmed the successful reduction of CD11b.sup.hi NK-cells in the circulation of DT-treated NK.sup.Csf1r/DTR-mice albeit the overall number of total or mature NK-cells remained unaltered (FIG. 4G). The improvements in obesity and insulin resistance were accompanied by decreased concentrations of circulating pro-inflammatory cytokines in DT treated NK.sup.Csf1r/DTR-mice (FIG. 4H). Collectively, specific, DT-mediated ablation of Csf1r.sup.+Ncr1.sup.+DTR, i.e. myeloid NK-cells, reduced obesity, insulin resistance and the systemic pro-inflammatory tone associated with HFD-feeding.

Example 5: Investigation of the Effect of Abrogation of IL6Ra-Signaling from NK-Cells of Mice

[0477] IL6Ra-expression discriminates between mature and myeloid (CD11b.sup.hi) NK-cells and IL-6 is consistently increased in circulation of obese mouse models and humans. Thus, the contribution of IL6Ra-signaling in NK-cells during the development of obesity and obesity-associated insulin resistance in vivo was investigated. Therefore, mice carrying a loxP-flanked IL6Ra-gene were crossed to Ncr1-Cre mice to specifically delete IL6Ra expression in Ncr1.sup.+ NK-cells. Such IL6Ra.sup.NK-mice and the respective control littermates were then subjected to either CD- or HFD-feeding from the age of 6 weeks on. While body weight remained unaltered between CD-fed IL6Ra.sup.NK- and control mice (FIG. 10A), HFD-fed IL6Ra.sup.NK-mice exhibited significantly reduced body weights (FIG. 10A), PGAT and liver masses as well as reduced total fat contents (FIG. 10B, C) compared to HFD-fed control mice. Histological examination of PGAT revealed a significant reduction in adipocyte size in IL6Ra.sup.NK-mice compared to controls. Reduced body weights in IL6Ra.sup.NK-mice upon HFD-feeding resulted from increased energy expenditure as assessed by indirect calorimetry, while food intake of these mice remained unaltered during the night phase and was even slightly increased during the light phase in IL6Ra.sup.NK-mice. Consistent with protection from obesity, IL6Ra.sup.NK-mice also exhibit reduced hepatic steatosis. Analysis of NK-cells isolated from HFD-fed IL6Ra.sup.NK- and control mice confirmed successful and specific disruption of IL6Ra expression in NK-cells. These data led us to conclude that, abrogation of IL6Ra-signaling in NK-cells protects from obesity and hepatic steatosis upon HFD-feeding.

[0478] To investigate the consequence of NK-cell specific IL6Ra-deletion for the manifestation of obesity-associated inflammation, the inventors determined total (CD45.sup.+) immune cell infiltration in liver and PGAT of 22-week old IL6Ra.sup.NK- and littermate control mice after 16 weeks of HFD-feeding. This analysis revealed a significant reduction of CD45.sup.+ immune cells in PGAT of HFD-fed IL6Ra.sup.NK-mice compared to controls (FIG. 11A). Similarly, the number of PGAT-infiltrating T cells and NK-cells were reduced in these mice (FIG. 11B, C). In line with the observations during obesity development in wildtype mice, no major differences in NK-cell maturation between IL6Ra.sup.NK- and control mice (FIG. 11D) were detected. In contrast, relative numbers of myeloid NK-cells were significantly reduced both in PGAT and circulation of IL6Ra.sup.NK-mice (FIG. 5A). In addition, IL6Ra.sup.NK-mice exhibited a clear reduction of F4/80 immunoreactive crown-like structures in PGAT compared to controls. Collectively, IL6Ra-inactivation in NK-cells protected from obesity paralleled by a reduction of myeloid NK-cells and obesity-associated macrophage infiltration in PGAT.

[0479] To further define the consequences of reduced CD11b.sup.hi NK-cell-formation in obese IL6Ra.sup.NK-mice, gene expression profiles were analyzed by mRNA deep-sequencing in PGAT and livers of HFD-fed IL6Ra.sup.NK- and control mice. The resulting gene expression profiles were subjected to gene set enrichment analyses, which revealed that genes found to be down regulated in livers and PGAT of IL6Ra.sup.NK-mice related to immune response, leukocyte migration and activation compared to control mice. Conversely, over represented gene ontologies in genes overexpressed in IL6Ra.sup.NK-mice compared to controls clustered in pathways associated with glucose and lipid homeostasis. Taken together, the alterations of immune responses in IL6Ra.sup.NK-mice translated into reduced inflammatory gene expression signatures and improved metabolic control in liver and PGAT of these animals.

Example 6: Determination of Insulin Resistance in IL6Ra.SUP.NK.-Mice During FD-Feeding

[0480] To study the consequences of reduced myeloid NK-cell-formation and attenuated obesity-associated inflammation on glucose homeostasis in IL6Ra.sup.NK-mice, the inventors performed glucose and insulin tolerance tests in HFD-fed IL6Ra.sup.NK-mice and littermate controls. This analysis revealed improved glucose tolerance and insulin sensitivity in IL6Ra.sup.NK-mice compared to control animals (FIG. 5B).

[0481] To further dissect tissue-specific effects of attenuated IL6Ra-signaling in NK-cells on insulin sensitivity and glucose homeostasis, hyperinsulinemic-euglycemic clamp studies were performed in HFD-fed IL6Ra.sup.NK-mice and littermate controls. Here, IL6Ra.sup.NK-mice required a significantly higher glucose infusion rate compared to control mice to maintain a similar degree of glycaemia, further underscoring the clearly improved overall insulin sensitivity in these animals (FIG. 5C). Similarly, IL6Ra.sup.NK-mice exhibited drastic improvement in insulin's ability to suppress hepatic glucose production and to promote glucose uptake into skeletal muscle and PGAT (FIG. 5C). On a molecular level this was reflected by an increased ability of insulin to stimulate AKT-phosphorylation in skeletal muscle and PGAT of HFD-fed IL6Ra.sup.NK-mice. Consistent with an improvement in inflammatory gene expression and insulin action, also obesity-induced JNK-activation was attenuated in liver and PGAT of IL6Ra.sup.NK-mice.

Example 7: Investigation of the Effect of Stat3 Deletion in NK-Cells

[0482] Stat3 is a central signaling component downstream of IL-6. Given the importance of IL6Ra-signaling for myeloid NK-cell formation, the inventors therefore conditionally inactivated Stat3 in NK-cells by crossing Ncr1.sup.Cre-mice to Stat3.sup.flox-mice. Six-week old Stat3.sup.NK-mice and their control littermates were subjected to HFD-feeding, and body weights were weekly assessed over a period of 12 weeks. Prior to HFD-feeding, body weights were similar in both 13 genotypes. In contrast, beginning with four weeks of HFD-feeding, Stat3.sup.NK-mice gained significantly less weight than the controls, and at the end of the experiment (12 weeks of HFD-feeding) Stat3.sup.NK-mice weighed 15% less than the controls (FIG. 6A). Reduced weight gain was reflected in a 44% reduction of fat mass assessed by computed tomography (CT) scans at this time point (FIG. 6B). Histological analyses of organs revealed less PGAT associated crown-like structures and reduced hepatic steatosis in Stat3.sup.NK-mice. Strikingly, Stat3.sup.NK-mice presented with drastically improved insulin sensitivity and glucose tolerance after 12 weeks of HFD-feeding. Of note, a slightly improved insulin action could be detected prior to HFD (FIG. 6C, D). These metabolic improvements correlated with a significant reduction of myeloid NK-cells in PGAT, whereaslike our previous observations in IL6Ra.sup.NK-miceother NK-cell subpopulations remained unaltered (FIG. 6E).

[0483] Thus, the HFD-elicited, IL6Ra-dependent formation of myeloid NK-cells predominantly depends on Stat3-mediated signaling in NK-cells.

Example 8: Detection of Myeloid NK-Cells in the Microenvironment of Human Obesity-Associated Cancers

[0484] Formalin-fixed, paraffin embedded (FFPE) human tissue samples of obesity-associated cancer entities (Bhaskaran et al., The Lancet 2014, 384, 755-765) were selected to analyze whether obesity-associated myeloid NK-cells, as described by before in obese murine and human non-cancer individuals, could also be detected in cancer. Therefore a non-exclusive number of cancer entities was chosen from clinically documented overweight or obese patients with a minimum body-mass-index of 25 kg/m.sup.2.

[0485] The experiment was carried out using an RNAscope, immunohistochemistry (IHC) combi-stain (IL6Ra, CD56), and human FFPE tissue sections.

[0486] Materials

[0487] RNAscope: [0488] Hs-IL6R Probe (ACD bio-techne, USA; catalog no. 557201) [0489] 2.5 HD Reagent Kit-Red (ACD bio-techne, USA; catalog no. 322350)

[0490] IHC: [0491] mouse anti-human CD56 moAB (clone 123C3) (Invitrogen, cat. 07-5603) [0492] donkey anti-mouse 2.sup.nd antibody (Jackson ImmunoResearch, (Ref: 715-065-150) [0493] VECTASTAINABC-HRP Kit (Vector Labs, cat no. PK-6100) [0494] DAB (DAKO, Agilent Technologies, USA, cat no. K3468) [0495] Hematoxylin: (Vector Laboratories; QS H-3404)

[0496] Preparations (Cutting and Baking, Max. 1 Week Before Experiment):

[0497] FFPE samples were cut into 4 m sections (on 38-40 C. heating plate until finished). The slides were baked for 1 h for 60 C. (vertical position), and proceeded immediately or store at room temperature afterwards (max. 1 week).

[0498] Day of Experiment:

[0499] Prepare Buffers Such as Follows: [0500] Target Retrieval Buffer (prepare 700 ml) [0501] (do not boil longer than 15 min before use!) [0502] 630 mL distilled water (MPwater) [0503] +70 mL 10 Target Retrieval Reagent (REF 322000) [0504] Wash Buffer (prepare 3 L) [0505] 2.94 L distilled water (MPwater) [0506] +60 mL 50 Wash Buffer (REF 310091)

[0507] De-Paraffinize Sections: [0508] 5 Min in Xylol, agitate gently [0509] 5 Min in fresh Xylol, agitate gently [0510] 3 Min in 100% EtOH, agitate gently [0511] 3 Min in fresh 100% EtOH, agitate gently [0512] Air dry slides on absorbant paper with samples face-up for 5 minutes

[0513] RNA Scope

[0514] RNAscope Tissue Pretreatment:

[0515] HybEZ oven was warmed up to 40 C. The humidifying paper (wet with MPwater) was put in humidity control tray. The covered tray was warmed up to 40 C. for at least 30 min. (keep it there, when not in use). Target Retrieval Buffer was heated up to 98-102 C. which was checked with a thermometer. Afterwards, an incubation in RNAscope using H.sub.2O.sub.2 (3%; 5-8 drops to cover whole tissue) for 10 min @ room temperature (blocks endogenous peroxidase activity) followed. H.sub.2O.sub.2 was removed by tapping on an absorbent paper, then immediately slides was put in MPwater, and the tray was moved in the water up and down 3-5 times. The slides were washed again in fresh MPwater. The slides were boiled in the target retrieval buffer for 15 min (98-102 C.). The hot slides were immediately put into fresh MPwater at room temperature, and washed 3-5 times by up-down movements (use a big beaker in case of many samples to keep RT conditions). They were transferred into fresh 100% EtOH, then let them be dried in the air dry completely at room temperature. A barrier (ImmEdge Pen) is created around the tissue, let it dry. An incubation of the tissue in ProteasePlus for 25 min at 40 C. using approximately 4-5 drops/slide followed. (The Detection Kit materials was prepared during this incubation time). The remaining ProteasePlus was removed (tap onto absorbent paper) and was immediately put slides into fresh MPwater, and washed (by 3-5 up-dwn movements)

[0516] RNAscope Amplification

[0517] The reagents were equilibrate reagents to the required temperatures before. The following steps were carried out at 40 C. The remaining liquids were removed from slides by tapping on adsorbent paper. Afterwards, the target probe was hybridize using 4-5 drops per tissue in order to completely cover it, and incubated for 2 hours at 40 C. Then the excess liquids were removed, and the probe was immediately put into one wash buffer, and washed for 2 min. at room temperature. The washing step was repeated with a fresh wash buffer, and washed for 2 min. at room temperature. AMP 1 was hybrized for 30 min. at 40 C. using 4-5 drops per slide. The excess liquid was removed and washed in a wash buffer two time for 2 min. at room temperature. AMP 2 is hybrized for 15 min. at 40 C. using 4-5 drops per slide. Then the excess liquid were removed, and washed two times in one wash buffer for 2 min. at roomtemperature. AMP 3 is hybrized for 15 min. at 40 C. using 4-5 drops per slide. Then the excess liquid were removed, and washed two times in one wash buffer for 2 min. at roomtemperature. AMP 4 is hybrized for 15 min. at 40 C. using 4-5 drops per slide. Then the excess liquid were removed, and washed two times in one wash buffer for 2 min. at roomtemperature.

[0518] AMP 5 is hybrized for 15 min. at room temperature using 4-5 drops per slide. Then the excess liquid were removed, and washed two times in one wash buffer for 2 min. at roomtemperature. AMP 6 is hybrized for 15 min. at room temperature using 4-5 drops per slide. Then the excess liquid were removed, and washed two times in one wash buffer for 2 min. at roomtemperature.

[0519] It was immediately proceed with step 1 of the following procedure.

[0520] RNAscope Signal Detection

[0521] RED working solution by using a 1:60 ratio of RedB to RedA (e.g. 2.5 L RedB+150 L RedA) was prepared. It has to be used within 5 min (protect from UV or direct sun light), whereas 70-80 L/slide should be calculated.

[0522] The excess wash buffer was removed from the slides (tap onto adsorbent paper. Hybridization in RED working solution was performed for 10 min. at room temperature. Afterwards, the excess staining solution was removed from the slides (tap onto absorbent paper). The slides were washed in MPwater (two times fresh water each). Then it is directly proceed with the immunohistochemistry.

[0523] Immunohistochemistry (DAB):

[0524] Excess of water was removed and incubated in H.sub.2O.sub.2 (3%) for 10 min at room temperature in order to block the peroxidase activity from the RNAscope amplification. The tissue sections were blocked for 2 h at room temperature in donkey blocking solution (1PBS, Ac, 0.25% TritonX, and +3% donkey serum). Then an incubation with primary antibody at 4 C., overnight was carried out. The day after, the slided was washed (310 min) in PBS and 0.1% TritonX. A secondary antibody (donkey-anti-mouse Biotin) (1:500 in new blocking solution (=PBS and 0.1% TritonX) was added (2.sup.nd AB: Jackson ImmunoResearch, Ref: 715-065-150). Incubation with the secondary antibody was performed for 1 h at room temperature. Then slides was washed three time for 10 min. in PBS and 0.1% TritonX. The signal was enhanced with an ABC-Kit (VECTASTAINABC-HR; Kit; PK-6100, 2.5 mL PBS and 0.1% TritonX and 10 L A+10 L B (depending on no. of slides)) for 30 min. at room temperature. The slides were washed in PBS and 0.1% TritonX. The signal was detected with DAB (DAKO, Ref: K3468) at pH 7.5. For this 1 mL DAB and 1 drop of DAB substrate (calculate 100 L for number of slides) is used. Incubation in DAB was carried out for 10 min in the dark. The reaction was stopped and rinsed gently with MPwater. The staining (Hematotoxylin) was counted with QS H-3404; Vector Laboratories. Hematoxylin solution was dropped directly onto the slide until it is covered. Then, incubation for 30 sec (this is normally enough) followed. The reaction was stopped in MPwater and rinsed until the slide is clear.

[0525] Mounting and Sealing

[0526] The slides were dried at 60 C. for 15 min. (until completely dry). It was briefly dip into fresh, pure Xylol, and immediately mounted with EcoMount solution (before Xylol dries). Gently cover slip was put on top and air dry at room temperature for 10 min (or overnight). A long-term storage is at 4 C.

[0527] Result

[0528] The myeloid NK-cells characterized by IL6ra and CD56 co-expression could be found in the following cancers: endometrium carcinoma, lymph node metastasis (breast cancer, obese male), colon carcinoma, pancreatic cancer, gastric carcinoma, esophagus carcinoma, prostate cancer, hepatocellular carcinoma, and multiple myeloma as depicted in FIG. 13 to 22.