CXCR6+ NK Cells in Human Fetal Liver and Spleen Possess Unique Phenotypic and Functional Capabilities

Abstract

Tissue-resident Natural Killer (NK) cells vary in phenotype according to tissue origin, but are typically CD56^bright, CXCR6⁺, and CD69⁺. NK cells appear very early in fetal development, but little is known about when markers of tissue residency appear during gestation and whether the expression of these markers, most notably the chemokine receptor CXCR6, are associated with differences in functional capability. Using multi-parametric flow cytometry, we interrogated fetal liver and spleen NK cells for the expression of a multitude of extracellular markers associated with NK cell maturation, differentiation, and migration. We analyzed total NK cells from fetal liver and spleen and compared them to their adult liver and spleen counterparts, and peripheral blood (PB) NK. We found that fetal NK cells resemble each other and their adult counterparts more than PB NK. Maturity markers including CD16, CD57, and KIR are lower in fetal NK cells than PB, and markers associated with an immature phenotype are higher in fetal liver and spleen NK cells (NKG2A, CD94, and CD27). However, T-bet/EOMES transcription factor profiles are similar amongst fetal and adult liver and spleen NK cells (T-bet^-/EOMES⁺) but differ from PB NK cells (T-bet⁺EOMES^-). Further, donor-matched fetal liver and spleen NK cells share similar patterns of expression for most markers as a function of gestational age. We also performed functional studies including degranulation, cytotoxicity, and antibody-dependent cellular cytotoxicity (ADCC) assays. Fetal liver and spleen NK cells displayed limited cytotoxic effector function in chromium release assays but produced copious amounts of TNFα and IFNγ, and degranulated efficiently in response to stimulation with PMA/ionomycin. Further, CXCR6⁺ NK cells in fetal liver and spleen produce more cytokines and degranulate more robustly than their CXCR6^- counterparts, even though CXCR6⁺ NK cells in fetal liver and spleen possess an immature phenotype. Major differences between CXCR6^- and + NK cell subsets appear to occur later in development, as a distinct CXCR6⁺ NK cell phenotype is much more clearly defined in PB. In conclusion, fetal liver and spleen NK cells share similar phenotypes, resemble their adult counterparts, and already possess a distinct CXCR6⁺ NK cell population with discrete functional capabilities.

Keywords: CXCR6; Eomes; T-bet; innate immunity; liver; natural killer cell; spleen; tissue residency.

Figure 1

Hepatic and splenic fetal NK cells share a similar immature tissue-resident phenotype. Phenotypic profiling of fetal liver and spleen NK cell subsets. Gating strategy based on FMOs in all tissues is shown in Figures S1–S8. (A) Proportion of CD56^bright and CD56^dim NK cells (left) and CD56 mean fluorescence intensity (MFI) (right) in fetal liver (blue circles), spleen (green circles), and adult PB NK cells (gray squares). (B) Comparison of extracellular markers important in NK cell differentiation, maturation, and migration in fetal liver (blue circles), fetal spleen (green circles), and adult PB NK cells (black squares). (C) Adult liver and spleen NK cells are phenotypically similar and are more mature than fetal liver and spleen NK cells. Comparison of extracellular markers on fetal liver and spleen NK cells versus adult liver and spleen NK. Individual data points are shown. Bars represent mean and standard deviation for each marker. Bars represent mean and standard deviation for each marker. A one-way ANOVA with either Holm-Sidak's Multiple Comparison Test (A) or Tukey's Multiple Comparisons Test (B,C) was used to analyze unpaired samples (shown). Paired Student's t-tests were used to compare matched fetal tissue pairs where possible (see text). ^*p < 0.05; ^**p < 0.01; ^***p < 0.001; ^****p < 0.0001. n = 4–8.

Figure 2

Phenotypic differences in CXCR6⁻ and CXCR6⁺ NK cell subsets in fetal liver and spleen. (A) Representative dot plots for gating CXCR6⁻ and + NK cells are shown. Comparison of NK cell markers on CXCR6⁻ (black) and CXCR6⁺ (red) NK cells in fetal liver, spleen, and adult PB NK cells (n = 8–26). (B) Representative flow plots showing expression of CD49e in fetal liver and spleen, adult liver and spleen, and adult PB NK cells in total, CXCR6⁻, and CXCR6⁺ NK cell subsets. (C) CXCR6⁺ NK cells express lower levels of CD49e than CXCR6⁻ NK cells in fetal liver and spleen. Individual data points showing CD49e expression on total, CXCR6⁻, and CXCR6⁺ NK cell subsets in fetal liver (blue circles) and spleen (green circles), adult liver (blue squares) and spleen (green squares), and adult PB NK cells (gray squares) (n = 4–11). Bars represent mean and standard deviation for each marker. A one-way ANOVA with Tukey's Multiple Comparisons Test was performed for unpaired samples and a Student's t-test for paired CXCR6⁻ and + NK cells within the same donor. P-values are as described in the legend to Figure 1.

Figure 3

Tbet and EOMES transcription factor profiles show distinct patterns in CXCR6⁻ and CXCR6⁺ NK cell subsets. (A) Representative flow plots showing gating strategy for Tbet, EOMES, and Tbet/EOMES subsets in total NK cells (see Figure S9 for FMOs). (B) Expression of Tbet/EOMES subsets in total, CXCR6⁻ and CXCR6⁺ fetal liver (n = 8), fetal spleen (n = 8), PB (n = 18), adult liver (n = 3), and adult spleen (n = 7) NK cells. Total, CXCR6⁻, or CXCR6⁺ NK cells were gated according to FMOs and percentages of four subsets of Tbet/EOMES-expressing NK cells were quantified as parts of a whole: Tbet⁻EOMES⁻ (purple); Tbet⁻EOMES⁺ (blue); Tbet⁺EOMES⁺ (red); and Tbet⁺EOMES⁻ (green). Donor-matched fetal liver and spleen were compared using Student's t-test, as were CXCR6⁻ and CXCR6⁺ NK cells within the same donor. Unmatched adult liver and spleen NK cell subsets were compared to fetal liver and spleen NK cells and peripheral blood NK cells using a one-way ANOVA with Tukey's Multiple Comparisons Test (see Table). (C) Representative flow plots for Tbet/EOMES expression in fetal liver and spleen NK cells analyzed by gestational age. (D) Summary of the expression of Tbet/EOMES-expressing NK cell subsets in fetal liver and spleen as a function of gestational age. The majority of time points are represented by one sample (n = 1), with the exception of weeks 17 and 19 (n = 2). Tbet⁻EOMES⁻ (purple line); Tbet⁻EOMES⁺ (blue line); Tbet⁺EOMES⁺ (red line); Tbet⁺EOMES⁻ (green line).

Figure 4

CXCR6⁺ NK cells exhibit a less mature phenotype in terms of CD27, CD11b. (A) Representative flow plots showing expression of four subsets of CD27 and CD11b-expressing NK cells within total NK cells of fetal liver, spleen, and PB (n = 3). FMOs are in Figure S10. (B) CD27, CD11b-expressing NK cells were quantified as parts of a whole as follows: CD27⁻CD11b⁻ (green); CD27⁺CD11b⁻ (blue); CD27⁺CD11b⁺ (purple); CD27⁻CD11b⁺ (orange). Statistics were as described in the table. (C) Expression of CD27, CD11b subsets in total fetal liver and spleen NK cells as a function of gestational age. Colored lines correspond to subpopulations as described in (B) (n = 1 for each time point).

Figure 5

Fetal liver and spleen NK cells have low killing capacity but produce cytokines and degranulate in response to stimulation. (A) Bulk fetal liver (n = 4) and spleen (n = 2) lymphocytes were plated at various effector to target cell ratios (E:T) with ⁵¹Cr-labeled target cells. K562 were used as targets for cytotoxicity assays in the presence or absence of IL-2. Raji B lymphoma cells were used as targets for ADCC, plus or minus rituximab. PBMCs from healthy donors were used as controls. (B) Intracellular flow cytometry (ICFC) was used to assess fetal liver and spleen NK cell function. Total NK cells were gated for markers of cytotoxic function including production of IFNγ, TNFα, perforin, Granzyme B, and CD107a (degranulation). Total NK cells were stimulated with CFSE-labeled K562 target cells at a 1:1 ratio or PMA/ionomycin (P/I). Unstimulated cells (unstim) were incubated in media alone. Incubation was for 4 h. Fetal liver (blue circles), fetal spleen (green circles), and PB NK cells (black squares) (n = 4). (C–E) CXCR6⁺ fetal liver and spleen NK cells possess a unique functional profile with respect to mediators of NK cell effector function. Differences in the contribution of CXCR6⁻ NK cells (black circles) and CXCR6⁺ NK cells (red circles) in the production of IFNγ, and TNFα, (C) perforin, and CD107a (D) and Granzyme B, (E) in fetal liver and spleen NK cells following stimulation. PB NK cells were used as controls. Matched pairs were analyzed for statistical significance using the Student's t-test. P-values are as shown or as detailed in the legend to Figure 1.