Differential Effector Function of Tissue-Specific Natural Killer Cells against Lung Tumors

Natural killer (NK) cells are a crucial component of the immune system, with diverse functions that remain the subject of ongoing research [1]. Although NK cells were first described in the 1970s [2‒5], researchers still debate the lineage and subsets of NK cells [6].

Currently, the knowledge of NK cell heterogeneity is limited outside the recognition hypothesis [7, 8] that delineates between licensed (containing self-specific MHC-I receptors) and unlicensed (lacking self-specific MHC-I receptors) NK cell subsets [9]. These subsets exhibit distinct functional characteristics during immune challenges [10], and these functions are influenced by the specific immune environment context [11]. Following the discovery of innate lymphoid cells (ILCs), NK cells were recategorized as group 1 ILCs mainly based on sharing an innate lymphoid precursor and secretion of TNF-α and IFN-γ [12]. This recent research into ILC biology and renewed insights about NK lineage have increased our understanding of the plasticity of NK cells in different contexts [13‒15].

However, NK cells represent immune cells with complex and multifaceted functions beyond producing cytotoxic effects [16]. In addition to their ability to directly eliminate infected, stressed, or malignant cells [17], they also play a crucial role in regulating and exhibiting immunoregulatory functions toward other immune cells [18, 19]. Moreover, recent studies have revealed phenotypic and functional variances in NK cells exposed to different physiological and pathological conditions, including cancer [20‒23]. Limited knowledge of tissue-specific phenotypic NK cell plasticity represents a challenge for developing new NK cell-based therapies. Therefore, it is crucial to comprehend the tissue-specific diversity of NK cells and its impact on the immune response against the solid tumor microenvironment (TME) for cancer immunotherapy.

In this study, C57BL/6 WT mice and the syngeneic LL/2 lung tumor models were employed in vitro and in vivo to examine the distinct expression profiles of NK cell surface receptors associated with an effector stage (CD200R, CD244, NKp46/CD335, NKG2D/CD314, ITGB2/CD18, IFNAR1, Notch1, Notch2, Notch3, Notch4, IL15Rα, and CD69) across distinct anatomical sites, including the lymph node, bone marrow, liver, spleen, peripheral blood, thymus, and lung tissues.

The findings indicate that in addition to the conditions fostered by the TME, the tissue-specific origin of NK cells and their specific expression patterns targeted toward the tumor site can significantly impact the efficacy of the immune response against the tumor. This highlights the critical importance of identifying these specific expression patterns for the development of more targeted and efficacious NK cell-based immunotherapies.

The experiments were performed on C57BL/6 WT mice aged between 8 and 16 weeks, including both males and females in equal proportions. The mice were kept under pathogen-free conditions at the Meharry Medical College (MMC) Animal Care Facility. They were cared for in accordance with the guidelines outlined in the National Institute of Health Guide for the Care and Use of Laboratory Animals and the Institutional Animal Care and Use Committee (IACUC). The MMC facility is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) and follows the Public Health Service policy on Human Care and Use of Laboratory Animals.

The murine metastatic lung epidermoid carcinoma LL/2 (LLC1) (ATCC^®, CRL-1642™) was cultured in Iscove’s Modified Dulbecco’s Medium (Gibco™, Thermo Fisher Scientific, Waltham, MA, USA). The medium was supplemented with 10% FBS, glucose, sodium bicarbonate, sodium pyruvate, l-glutamine, Antibiotic-Antimycotic, β-mercaptoethanol, HEPES, and MEM NEAA. The tumor cells were maintained at low passages (<10) for experimentation and were regularly verified to ensure fidelity.

The protocol for generating tumor spheroids was adapted from Foty et al. [24]. LL/2 tumor cells were cultured until they achieved 90% confluence, followed by two washes with PBS. Subsequently, 1.5 mL of 0.05% trypsin-1 mM EDTA (for a 75 cm flask) was administered, and the culture was incubated at 37°C for 1 min for cell detachment. The trypsinization process was then halted by adding 7.5 mL of complete media, and the cells were mixed until they were fully resuspended. The single-cell suspension was subsequently prepared at a viable concentration of 2.5 × 10⁶ cells/mL.

Tumor spheroids were formed by placing 24 drops of 25 μL LL/2 single-cell suspension on the sterile lid, inverting it, and placing it on top of the PBS-filled (12.5 mL) 100 × 20 mm Petri dishes. The dishes were incubated at 37°C with 5% CO₂ for 10–14 days until spheroids formed, followed by transferring them to 96-well U-bottom tissue culture plates, with one spheroid per well, for experimentation. Each 25 μL drop that forms a spheroid with a visible necrotic core begins with approximately 6.25 × 10⁴ cells and ends with approximately 712,500 (after 10 days) and 1M cells (after 14 days), respectively, corresponding to the LL/2 doubling time every 21 h.

For the in vivo experiments, 200 µL of LL/2 cells at a concentration of 2 × 10⁶ cells/mL (0.4 × 10⁶ LL/2 cells) were injected subcutaneously (s.c.) near the fourth mammary fat pad. Following injection, the mice were monitored daily for welfare. After 14 days, tissues, including lymph nodes, bone marrow, liver, spleen, blood, thymus, and lung, were harvested for further analysis. For intravenous (i.v.) injections, 0.5 × 10⁶ LL/2 cells were resuspended in 50 µL and injected into the lateral tail veins. The lungs were harvested 25-day of postinjection for further experimentation.

After sedation with 2,2,2-Tribromoethanol (Millipore Sigma, St Louis, MO, USA) and cervical dislocation, bone marrow, liver, lung, lymph nodes, peripheral blood, spleen, and thymus tissues were collected from mice to prepare single-cell suspensions. For the isolation of NK cells from lymph nodes, we chose the inguinal, brachial, and axillary lymph nodes due to their larger size and abundant cell yield. Additionally, when feasible, superficial cervical and mediastinal lymph nodes were also used for this purpose. The tissues were homogenized using 40 µm cell strainers in complete IMD medium. These cell suspensions were transferred to 15 mL conical tubes, counted, and centrifuged at 163 G for 7 min, followed by resuspension at an optimum concentration of approximately 2.5 × 10⁶ cells/mL across all tissues. Prior to counting, spleen, liver, and peripheral blood samples were treated with 500 µL ACK Lysing Buffer (Gibco^™, Thermo Fisher Scientific, Waltham, MA, USA) to lyse erythrocytes. The lysing process was neutralized by resuspending the cells in complete IMD medium.

The coculture of different tissue cell suspensions with tumor spheroids was conducted at a 1:1 ratio using the hanging drop tumor spheroid protocol. Subsequently, NK cells were detected through flow cytometry analysis. For fluorescence intensity analysis, approximately 6.25 × 10⁴ purified tissue-specific NK cells were added to each well containing a tumor spheroid before incubation and imaging. For cytotoxicity assays, tumor spheroids were dissociated using pipetting, and the resulting cells were added to an assay plate at a 1:1 ratio of tumor cells to purified NK cells.

We utilized the NK Cell Isolation Kit II to isolate NK cells for in vitro applications (Miltenyi Biotec, Germany). The cells were counted, centrifuged at 244 G for 5 min, and the supernatant was removed. The cell pellets were then resuspended in 40 μL MACS buffer per 1 × 10⁷ total cells. Subsequently, 10 μL of NK Cell Biotin-Antibody cocktail was added, and the mixture was incubated for 5 min at 4°C. Following incubation, the cells were washed, and 80 μL of MACS buffer and 20 μL of Anti-Biotin MicroBeads were added. The mixture was then incubated for 10 min in the dark at 4°C. Finally, the magnetic separation was performed using the LS column.

The following anti-mouse antibodies used were anti-NK1.1 APC, PE, FITC (clone PK136), anti-CD314 PE, APC (clone 29A1.4/CX5), anti-CD49b PerCP/Cy5.5, PerCP, FITC, APC, PE (clone DX5), anti-CD16/32 FITC (clone 93), anti-CD3 FITC, PerCP, PerCP/Cy5.5 (clone 17A2), anti-CD45 PerCP, FITC (clone 30-F11), anti-CD335 PE (clone 29A1.4), anti-CD244.2 FITC (clone m2B4(B6)458.1), anti-CD200R FITC (clone OX-110), anti-IFNAR-1 APC (clone MAR1-5A3), anti-IL15Rα APC (clone 6B4C88), anti-CD18 PE (clone H155-78), anti-CD69 PerCP (clone H1.2F3), anti-Notch1 APC (clone HMN1-12), anti-Notch2 PE (clone HMN2-35), anti-Notch3 PE (clone HMN3-133), anti-Notch4 APC (clone HMN4-14). All monoclonal antibodies were purchased from BioLegend^® or eBioscience. Cells were resuspended and fixed in 200 µL of 4% paraformaldehyde.

Cell events were collected from a live gating region using the Guava easyCyte 8HT Base System (Luminex Corporation, Austin, TX, USA) to determine the expression of the NK cell receptors. The data were then analyzed using FlowJo version 10.8.1 software (BD Biosciences, Franklin Lakes, NJ, USA). The gating for the specific cell population was determined by single-color controls. Unstained and isotype controls were used to verify the absence of any nonspecific antibody binding.

Purified tissue-specific NK cells were labeled using the Vybrant^™ DiO cell-labeling solution (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) according to the “Labeling of Cells in Suspension” Experimental Protocol and were cocultured with LL/2 cells. After coculturing for 4 h, the cells were washed with PBS and then imaged using the BZ-X series All-in-One Fluorescence Microscope (Keyence, Osaka, Japan). The fluorescence intensity of the cells was measured using ImageJ software. Corrected total cell fluorescence (CTCF) measurements were obtained by taking the pixel count over a selected area and subtracting the area of the selected cell times the mean fluorescence of the image background.

We conducted a rigorous data evaluation using the Shapiro-Wilk (W) test to ascertain its homogeneity, normality, and independence. Once the parametric conditions were observed, we analyzed the data using two-way ANOVA with Tukey’s multiple comparisons tests. Furthermore, we explored the main effects of the independent variables using linear regression models and calculated values for R, R squared, adjusted R², and p value. All tests were performed in GraphPad Prism v10.2.3. A p value of ≤0.05 was considered statistically significant.

To examine the impact of the tumor cell microenvironment on the expression of surface markers that regulate the activation and effector function of NK cells, we assessed the expression patterns of activation-associated receptors (CD314, CD244, CD335, and CD18) (online suppl. Fig. 1, online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000542078) using two distinct models: tumor spheroids (in vitro) and syngeneic subcutaneous tumor injection into C57BL/6 WT mice models (in vivo).

We observed multiple tissue-specific differences in the percentage of CD314⁺ NK cells in vitro. Without LL/2 exposure, we observed a decrease in the expression of CD314 in bone marrow NK cells compared with NK cells from lymph nodes (p = 0.0354), blood (p = 0.0001), and lung (p = 0.0044) (shown in Fig. 1a). Similarly, NK cells in the spleen and thymus showed lower CD314 expression compared to blood (p = 0.0003 and p = 0.0021, respectively) and lung (p = 0.0094 and p = 0.0480, respectively).

After exposure to LL/2, we observed that the expression of CD314 in bone marrow, spleen, and thymus NK cells was lower compared to lymph node NK cells (p = 0.0366, p = 0.0053, and p = 0.0103, respectively) (shown in Fig. 1a). Spleen and thymus CD314⁺ NK cells were also lower than blood NK cells (p = 0.0087 and p = 0.0165, respectively) after exposure to LL/2 tumor. Interestingly, we did not find significant intra-tissue differences in the percentage of CD314⁺ NK cells with or without LL/2 exposure. The in vitro analysis also revealed a significantly higher expression of CD244 in NK cells derived from liver and spleen tissues compared to NK cells from other tissue sites (p < 0.0001) both before and after exposure to LL/2 (shown in Fig. 1b). Additionally, we found significantly higher expression of CD335⁺ NK cells from the liver and blood tissues as compared to other NK cell-sourced tissue sites (p < 0.0001) (shown in Fig. 1c). Moreover, significant differences were observed in the percentage of CD18⁺ NK cells from thymus and lymph node compared to NK cells from other tissues (shown in Fig. 1d). However, without LL/2 exposure, thymus NK cells exhibited higher CD18 expression than liver and lung NK cells (p = 0.0008 and p = 0.0031, respectively). CD18 expression from lymph node NK cells was also notably higher than liver NK (p = 0.0441). Intra-tissue analysis also shows significantly less CD18 expression in blood NK cells after LL/2 exposure (p = 0.0010), in addition to that expression in blood NK cells being less than liver and spleen NK cells (p = 0.0266 and p = 0.0137, respectively) (shown in Fig. 1d).

The intra-tissue difference of CD314⁺ NK cells from the lung showed significantly higher in the presence of LL/2 tumors (p < 0.0001) (shown in Fig. 1e). However, consistent with the in vitro results, bone marrow, spleen, and thymus NK cells positively expressed CD314 at lower expression than other NK cell tissues. However, a higher CD244⁺ expression remained consistent in vivo with NK cells from the liver (shown in Fig. 1f). Also, we observed intra-tissue differences in the blood NK cells with diminished CD335 expression after exposure to the LL/2 tumor. In comparison, the higher CD335 expression from the liver remained consistent in vivo (shown in Fig. 1g). Moreover, the intra-tissue differences in blood remained consistent with in vitro results, with less CD18 expression observed post-LL/2 injection (p < 0.0001) (shown in Fig. 1h). Furthermore, in the absence of LL/2 tumor (saline injection condition), blood NK cells showed less CD18 than lymph node (p = 0.0050), bone marrow (p = 0.0108), spleen (p = 0.0025), and thymus (p < 0.0001) NK cells. Thymus NK cells also exhibited higher CD18 than liver (p = 0.0037) and lung (p = 0.0053) and remained higher than both liver (p = 0.0032) and lung (p = 0.0031) with LL/2 injection (shown in Fig. 1h). Finally, after the LL/2 exposure, the lymph node (p = 0.0328) and spleen (p = 0.0247) NK cells also had higher CD18 expression than the liver.

These results reveal consistent tissue-specific NK receptor expression patterns across the in vitro and in vivo experiments. Notably, lower expression of CD18 on NK cells from blood and lymph nodes was observed, whereas the liver NK cells exhibited higher expression of CD244 and CD335. Similarly, both CD18 and CD314 are expressed moderately in the thymus, while CD314 is prominent in the spleen and bone marrow. Furthermore, marked variations in receptor expression levels are evident among specific tissues, such as the lungs, blood, and lymph nodes, for CD314, CD244 in the liver, and CD18 in the thymus. These disparities suggest independent fluctuations in receptor expression within the same tissue, irrespective of NK cell exposure to the tumor.

To investigate the effect of a tumor cell environment on the expression of NK immunoregulatory surface markers across tissues, we compared the expression of CD200R in tissue-specific NK cells (CD3⁻NK1.1⁺) with and without exposure to LL/2 using CD49b to discriminate between circulating (CD3⁻/NK1.1⁺/CD49b⁺) and tissue-resident (CD3⁻/NK1.1⁺/CD49b⁻) NK cells [25, 26] (online suppl. Fig. 2; online suppl. Table 1). Through tissue-specific representative histograms of CD200R expression in circulating and tissue-resident NK cells (shown in Fig. 2a), we observed that the expression of CD200R was higher in circulating NK cells in vitro and in vivo across tissues compared with tissue-resident NK cells. Particularly, in the in vitro settings, in the absence of LL/2 exposure, bone marrow, and liver NK cells exhibited significantly higher CD200R expression (p < 0.0001) compared with other tissues (shown in Fig. 2b). This expression pattern remained increased after exposure with the LL/2 tumor (p < 0.0001). Also, liver NK cells were the sole tissue source among tissue-resident cells with significant intra-tissue expression differences upon LL/2 exposure (p = 0.0054). When examining circulating phenotypic NK cells in vitro, only blood NK cells in the LL/2 exposed group presented a significantly lower expression of CD200R compared to the liver (p = 0.0255) and spleen (p = 0.0235) (shown in Fig. 2c). Furthermore, the in vitro differential expression patterns of CD200R were validated using our in vivo model. The percentage of blood and lung NK cells (CD3⁻NK1.1⁺CD49b⁺CD200R⁺) was notably lower compared to other tissues (p < 0.001). This corroborates with the in vitro results (shown in Fig. 2d). Regarding the circulating NK cells, blood and lung NK cells consistently express CD200R at a lower percentage compared to other tissues, both with and without LL/2 exposure (shown in Fig. 2e). This aligns with our results in the in vitro setting. These variations in the expression patterns of CD200R across different tissues imply that tissue-specific NK cells possess distinct capabilities in recognizing tumors through the CD200R-CD200 axis. This also suggests that, besides the differential patterns of activating receptors, the specific expression patterns of inhibitory receptors may selectively influence the antitumor response of these tissue-specific NK cells.

To comprehensively characterize the interactions between tissue-specific NK cells and LL/2 tumors and to validate the implications of NK-specific receptor expression patterns on the NK effector function, an analysis of the infiltrating capacities and the cytotoxicity effects was conducted. This procedure involved utilizing DiO-labeled tissue-specific NK cells and LL/2 tumor cells in a spheroid, with a 1:1 effector-to-target ratio (shown in Fig. 3a). The calculation of the CTCF in the spheroid revealed that the coculture of blood or thymus NK cells with the tumor exhibited significantly less CTCF compared to NK cells from lymph nodes, bone marrow, and liver tissues (shown in Fig. 3b).

We also noticed a significant increase in tissue-specific cytotoxicity, as measured by LDH release. Specifically, liver and bone marrow NK cells exhibited significantly higher cytotoxicity when compared to other tissues such as the spleen (p < 0.037), thymus (p < 0.0393), lung (p < 0.0133), and blood (p < 0.01) (shown in Fig. 3c).

We also examined changes in IL15Rα and IFNAR1 across tissues-specific NK cells with and without exposure to LL/2. IL15Rα and IFNAR1 are two cytokine receptors crucial for the development, activation, proliferation, survival, and homeostasis of NK cells [27‒30]. Additionally, we analyzed CD69, historically known as a stimulatory receptor for NK cell function [31, 32] (online suppl. Fig. 3; online suppl. Table 2). Interestingly, prior to exposure to LL/2, blood NK cells exhibited higher IL15Rα expression levels compared to bone marrow (p = 0.0043), spleen (p = 0.0025), and thymus NK cells (p = 0.0022). Conversely, after exposure to the tumor, the proportion of IL15Rα⁺ NK cells within each tissue significantly increased compared to unexposed conditions except in blood, spleen, and thymus NK cells. Notably, the highest IL15Rα expression levels were shown in lymph node and liver NK cells (shown in Fig. 3d). In addition, a significant difference in intra-tissues was observed for IFNAR1 expression in both lymph nodes and blood between no LL/2 vs. LL/2 condition (shown in Fig. 3e). However, no significant differences for CD69 expression were observed irrespective of LL/2 tumor exposure. However, NK cells in the liver and bone marrow exhibited a significant increase in CD69 expression compared to other tissue-specific NK cells (shown in Fig. 3f), which could be related to a higher activation level in these tissue-specific NK cells. These findings indicate that tissue-specific NK cells within a single organism exhibit differential effector capacities to target the same tumor. These differences may be due to the specific expression patterns of activating receptors, such as those in the liver and bone marrow. Potentially, NK cells from these two tissues could lead to more effective responses against LL/2 tumors.

To investigate the alternative mechanisms of activation and migration of tissue-specific NK cells, we conducted an expression analysis of NOTCH family proteins with or without exposure to the LL/2 tumor (online suppl. Fig. 4; online suppl. Table 2). Our findings revealed that following exposure to the LL/2 tumor, the expression of NOTCH1 (shown in Fig. 4a) and NOTCH4 (shown in Fig. 4d) decreased in all tissues but slightly increased in blood and lung NK cells (p < 0.0001). Conversely, the expression of NOTCH3 (shown in Fig. 4b) and NOTCH2 (shown in Fig. 4c) increased in all tissues prior to exposure to the LL/2 tumor. However, after exposure to LL/2 tumor, NK cells from lymph nodes, bone marrow, and liver displayed significant increases in NOTCH3 expression compared with other tissues (p = 0.0092, p = 0.0262, and p < 0.0001, respectively). These results suggest that, in addition to the canonical activation mechanisms of NK cells, there are also less studied alternative mechanisms that exhibit differential changes in NK cells from various tissues.

To comprehend and forecast the behavior of NK cell intrinsic systems in response to tumors, we conducted a mathematical analysis of the factors that linked to the effector function using the differential expression data indicated in the above result section. Initially, we performed a correlation analysis to examine the variables directly associated with cytotoxic ability across all tissues in scenarios with or without exposure to LL/2 tumor (shown in Fig. 5a, b). Interestingly, some variables, such as CD244 and CD69 receptors, showed a strong positive correlation with cytotoxic capacity before tumor exposure and are maintained after tumor exposure. In contrast, other variables like CD335 and IL15Rα showed an increase in the positive correlation value post-exposure with the tumor (shown in Table 1). This analysis suggests that some of the receptors within the intrinsic NK cell system might be crucial in developing an effective cytotoxic response, as they are expressed positively in certain tissue-specific NK cells even before exposure to the tumor.

Linear regression models have been widely applied to study tumor-infiltrating immune cells before using RNA-seq data from blood and bulk tissues [33, 34]. Based on this, we employed linear regression models to examine specific correlations of analyzed variables and independent effector functions of NK cells from various tissues based on our correlation matrices. Initially, we evaluated the correlation between the expression of CD69 as a dependent variable of the activation capacity of the cells and the independent variables (receptors) influencing activation. Our findings indicate that NK cells in the liver, lymph nodes, and bone marrow exhibit a more significant number of positively correlated variables that influence the expression of CD69 (shown in Table 2). Additionally, after NK cell exposure to the LL/2 tumor, the number of variables increases, with proteins including CD244, CD335, and NOTCH3 being the most influential variables affecting the expression of CD69 (shown in Table 3).

We also assessed the implications of increased CD69 expression on the dependent variable (cytotoxicity) through another regression model. Our prediction model revealed that the positive correlation of CD69 in liver NK cells exerts a significant influence on the cytotoxic responses compared to other tissues (shown in Table 4). This correlation contrasted with the multiparametric analysis using three variables where we observed the positive influence of these variables at both pre- and post-LL/2 tumor exposure (shown in Fig. 5c, d). In summary, our findings showed that NK-sourced tissues expressing similar receptors exhibit significant coefficient parameter estimates and correlations between different receptors that could potentially affect the activity of NK cells against LL/2 lung cancer models.

Following the regression analysis, we examined the activation receptors on tumor-infiltrating NK cells from mice injected with LL/2 via s.c. and i.v. injections (shown in Fig. 6a), identifying positive expression of specific markers such as CD335, CD244, NOTCH3, and CD69 (shown in Fig. 6b). Additionally, a correlation matrix was used to assess the similarity between the immunophenotypes of these tumor-infiltrating NK cells and other tissue-specific NK cells, revealing a stronger correlation with the phenotypes observed in the liver (s.c. injection) and thymus (s.c. and i.v. injection) (shown in Fig. 6c). However, when focusing specifically on CD69 receptor expression, we observed a weaker correlation between NK cells at the tumor site and those from other highly cytotoxic tissues, such as the liver. In contrast, a strong correlation was noted with blood NK cells, which exhibit the lowest cytotoxicity (shown in Fig. 6d).

These observations indicate that while tumor-infiltrating NK cells share immunophenotypic similarities with other tissue-specific NK cells with highly cytotoxic potential, they have substantially reduced CD69 expression. Based on our regression analysis, CD69 expression is closely associated with cytotoxic capacity. This reduction may highlight the role of the TME in suppressing the cytotoxic response of NK cells, underscoring its pivotal influence in modulating the immune response mediated by these cells. We also confirmed that mathematical correlations can be instrumental in predicting how NK cells from various tissues expressing the same set of receptors may respond differently to immune challenges, resulting in varying antitumor responses. Observed differential expression patterns could shed light on tissue-specific phenotypic variations in NK cells exposed to LL/2 or other tumors.

Unraveling the web of NK cell diversity is an essential step toward developing successful NK cell-based cancer immunotherapies. Numerous studies have investigated the tissue-specific NK cell functions in lymph nodes, spleen, liver, thymus, lung, and peripheral blood [35‒41]. The uniqueness of this study is in characterizing, within solid lung tumor experimental frameworks in vitro and in vivo, tissue-specific NK cells from seven different tissue sites: lymph nodes, bone marrow, liver, spleen, peripheral blood, thymus, and lung, focusing on translational NK cell surface receptor proteins that are common in both humans and mice.

NKG2D, or CD314, is a fundamental activating receptor associated with the first signal during NK cell activation through directly identifying compromised cells via disrupted human leukocyte antigen ligands such as major histocompatibility complex class I chain-related antigens A (MICA) or UL16-binding proteins. Furthermore, NKG2D-deficient mice have shown deficiencies in tumor immunosurveillance [42]. Previous studies have also reported that the expression of the MICA protein is crucial for NK cells to identify and eliminate LL/2 cells [43], indicating that the tumor’s identification through active interaction of the CD314 with their ligands at the membrane level varies between the distinct tissue-specific NK cell subsets having direct implications on their cytotoxicity. We observed that although some NK cells from specific tissues (e.g., liver, lymph nodes, blood, and lung) express CD314 significantly, these cells vary in their cytotoxic response to tumors. In this context, some tumors can shed NKG2D ligands that bind to CD314 receptors. Once the ligands bind to the corresponding receptors, the NK cell becomes desensitized [44]. Also, the loss of NKG2D ligands could alter the correct identification of these tumors through this interaction, which could be a suppression mechanism in LL/2 tumors, desensitizing the effector capacity of NK cells through this axis. However, during the development of NK cells, the activating receptor NKG2D plays a crucial role in regulating the NK cell reactivity by setting the activation threshold for the receptor NKp46 (also known as CD335) [45]. The CD335 receptor is responsible for recognizing and preventing tumor metastasis and promoting tumor eradication [46].

In our analysis, we discovered that the expression percentages of CD335 in NK cells from the liver and blood are upregulated compared to other tissues before exposure to the LL/2 tumor. Additionally, after LL/2 exposure, we observed a specific decrease in CD335 expression exclusively in the blood NK cells, which could be closely related to the decreased cytotoxicity. In addition to upregulated CD335, the liver NK also expressed upregulated CD314, as mentioned earlier, especially in the in vivo experimental setup. High expression of both CD314 and CD335 rationalizes the higher cytotoxicity seen in liver NK cells. Indeed, other studies suggest that the combined expression of CD314 and CD335 can aid in the proper clearance of lung tumors [47]. Our finding on liver NK cells, where the expression of both receptors is high after exposure to LL/2 cells, supports this study. Our results show that NK cells from bone marrow and spleen may still effectively eliminate target cells despite low expression of CD314 and CD335 as compared to other tissues. This could indicate that while the presence of CD314 and CD335 on liver and blood NK cells may be crucial for tumor elimination, other alternative activation receptors might regulate the cytotoxic capacity in other tissues or even contribute to the heightened cytotoxicity of liver NK cells.

Interestingly, an increased expression of CD244 was observed in NK cells from the liver compared to other tissues. CD244 receptor plays a dual role, serving as both an activating and inhibitory receptor, which is determined by its receptor expression, ligation, adapter molecules, and immunoreceptor tyrosine-based switch motifs [48, 49]. In humans, CD244 predominantly induces activation signals in NK cells [50, 51]. Upon activation, this receptor significantly enhances NK cell functions, including cytokine production, tumor invasion, and cytotoxicity [52]. Conversely, in certain pathological conditions such as head-and-neck carcinoma, studies have demonstrated that the activation of 2B4/CD244 results in increased programmed death 1 (PD1) expression, promoting an immunosuppressive TME [53]. This alternative pathway may regulate the activity of these tissue-specific NK cells against LL/2, which, in turn, may be controlled through a constant expression of PD-L1 on lung tumor cells [54]. However, in our data, an increased CD244 expression on the liver NK cells is consistent and positively related to activation and increased cytotoxicity against LL/2. This could indicate that within the NK cells of different tissues, the differential patterns expression of the same activation receptors has different repercussions on the effector capacity when facing the tumor even under the same microenvironment conditions.

Moreover, we observed relatively stable expression percentages of CD18 with slight decreases in the LL/2 condition across most tissues except for the liver and lung. We also noted a significant decrease in blood NK cell CD18 expression percentages after LL/2 exposure. CD18 or integrin subunit beta 2 (ITGB2) is a crucial integrin during the development and maturation of NK cells. Its expression enhances NK cell cytotoxicity against tumors and virus-infected cells [55]. Conversely, NK cells that lack CD18 expression have a reduced ability to recognize target cells and are less responsive to activating receptor stimulation [56]. Studies have found that patients with acute myeloid leukemia with lower ITGB2 expression have a higher infiltration of resting NK cells [57]. Recent research has shown that CD18 is essential for the final maturation of NK cells, which can be regulated by the microenvironment under specific conditions, leading to transcriptionally and phenotypically diverse NK cells [58].

Our study also revealed high CD200R expression among circulating NK cells (CD3⁻/NK1.1⁺/CD49b⁺) across tissues. While blood and lung NK expressed CD200R at lower expression among circulating both in vitro and in vivo, bone marrow and liver NK cells expressed CD200R higher among NK cells considered tissue-resident (CD3⁻/NK1.1⁺/CD49b⁻) in both models. Previously, CD49b has been used to classify conventionally circulating NK cells vs tissue-resident NK cells [59, 60]. This may rationalize our results with more tissue expression variability in the CD49b⁻ NK cells. Some studies have reported the heterogeneous levels of CD200R in NK cells from acute myeloid leukemia patients [61]. In mice with CD200⁺ metastatic melanoma tumors, knocking out CD200R significantly reduced the percentage of NK cells in the tumor, suggesting that CD200R plays a vital role in tumor identification and subsequently regulating tumor growth [62]. Also, cell carcinomas that release CD200 are associated with a low presence of NK cells in the TME, primarily due to the binding of CD200 to CD200R in situ, blocking NK cell-killing functions and IFN release [63]. This suggests that the effector activity of NK cells is mainly regulated through the CD200R axis and may become ineffective at destroying the tumor when facing targets with a CD200 low or null expression. In this sense, a recent study reported the absence of the CD200R ligand (CD200) in LL/2 tumor cells [64]. The lack of interaction between the CD200R and its ligand may explain why we observed heightened cytotoxicity in certain tissue-specific NK cells, particularly those from the liver and bone marrow, compared to NK cells from other tissues despite their high expression of CD200R. These differences could also be attributed to the differentially expressed levels of activating receptors, including CD69, which are not negatively regulated through the CD200R-CD200 axis.

Notch signaling is integral to NK cell development and function, regulating signaling pathways in the ILC populations within most tissues [65]. Thus, we analyzed the expression of the Notch family receptor proteins, Notch1–4, in tissue-specific NK cells. Prior to LL/2 exposure, Notch1 and Notch4 were expressed at low percentages across tissues, with blood and lung NK cells expressing both receptors at higher percentages post-LL/2 exposure. It has been found that stage 4A NK cell developmental intermediates are regulated by Notch1 [15]. In addition, Notch2, co-expressed with Notch1 at earlier NK cell development stages, enhances NK cell function, specifically NK cytotoxic activity [66]. As we were not analyzing NK cell development intermediates, this could yield insight into the observed low percentages of Notch1 we saw across tissues. Interestingly, in our data, Notch2 and Notch3 were expressed at high percentages prior to LL/2 exposure. Once exposed to LL/2, intra-tissue expression percentage differences were only observed in lymph nodes, bone marrow, and liver tissues. Notably, Notch3 expression percentages increased with LL/2 exposure in lymph nodes, bone marrow, and liver NK cells. Notch3 has previously been shown to be critical in the pro-inflammatory responses of activated macrophages [67], and those pro-inflammatory responses include releasing activating cytokines like TNF-α, which help promote antitumor cytotoxicity in NK cells [68].

Moreover, NK cells actively crosstalk with dendritic cells that preferentially express Notch ligands like Jagged2 under different conditions [66, 69]. In LL/2 lung cancer, an overexpressed glycan-binding protein called galectin-1 increases the expression of Jagged2 and promotes lung tumor metastasis partly by mediating dendritic cell anergy [70]. Thus, high expression percentages of Notch3 with and without LL/2 exposure can be rationalized through NK cell responses to dendritic cells and macrophages and in response to LL/2 Jagged2 expression. In addition, Notch2 regulates dendritic cells, leading to NK cell activation while promoting cytotoxic T lymphocyte differentiation [66, 71]. Notch2 expression is prominent in human NK cells from various tissues, with blood and decidual NK cells having different affinity levels for binding specific Notch ligands [72].

In addition, we found that some cytokine receptors play an important role as mediators of the cytotoxicity in these tissue-specific NK cells. Specifically, we observed that after the exposure to LL/2, lymph nodes, bone marrow, liver, and lung NK cells exhibited significantly increased IL15Rα expression. IL15Rα is a high-affinity receptor that promotes the trans-presentation of IL15, promoting the proliferation and differentiation of NK cells during development [29, 73]. Research suggests that IL15 and IL15Rα released from tumors can engage with NK cells and promote IL15/IL15Rα NK-mediated tumor immunosurveillance [74]. Additionally, significant increases in the percentage of IFNAR1⁺ NK cells were found in every tissue with LL/2 exposure except for the spleen and thymus. Type 1 IFN signaling can affect NK cell development, education, and tumor immunosurveillance in multiple ways [75]. Knocking out IFNAR1 in mice has been shown to contribute directly and indirectly to defects in NK maturation and cytolytic activity through other immune cells, such as dendritic cells [76, 77].

Following the application of mathematical correlation and linear regression models in our research, we discovered distinct relationships between the receptors associated with the activation function of NK cells with or without LL/2 exposure in each tissue. Notably, we observed positive predictive correlations in the liver among closely related factors such as CD69, CD335, CD244, and CD314. These correlations persisted even after NK cell exposure to the LL/2 tumor, suggesting potentially heightened cytotoxicity of hepatic NK cells compared to those from other tissues. Our controlled system supported these findings, and the adjusted R² values, accompanied by a low variation compared to the R values, enable us to forecast these correlation estimates in population systems, such as the population of hepatic NK cells. Additionally, we noted alterations in the correlations of receptors relative to other tissues during exposure to the LL/2 tumor, which could directly impact the effector capacity of these cells. These findings provide insight into the heterogeneity of tissue-specific NK cells and could contribute to the expanding research on tissue-specific NK cell diversity [78‒82].

Altogether, these data integrate into the topic of tissue-specific NK cell heterogeneity by adding more to the answer “What is the surface receptor expression of NK cells found in different tissues, and how does that affect cytotoxicity against tumors?”. Our research reveals complex, differential patterns in the expression of receptors associated with immunoregulatory and activation functions in NK cells from different tissues, especially in the context of interactions with the Lewis lung carcinoma metastatic cells. These receptors and their expression patterns showed a significant role in the proper functioning of tissue-specific NK cells within a living system, especially in the context of the implications of previous and after tumor interaction. Still, we also understand that the immune challenges of the microenvironment are highly complex and can impact the expression of activation and inhibitory associated surface receptors, influencing the NK cells’ ability to identify and kill tumor cells.

The future direction of this research includes building up the tumor spheroid model to mimic an accurate, complex TME by piecewise adding additional cells critical to both TME structure and function such as fibroblasts, cytokines, and other immune cells like T cells, macrophages, and dendritic cells. We would also like to include using kinetic time points to quantify the length of sustained cytotoxicity of these tissue-specific NK cells in the context of LL/2. Finally, adding additional immune cell types that are known to exhibit NK crosstalk, CD8+ T cells, dendritic cells, and macrophages [83] to study in concert with tissue-specific NK would be necessary for future research.

The implications of this research provide valuable insights into multiple critical areas: the potential to engineer a readily accessible, tissue-specific NK cell, likely derived from blood, to emulate the effector phenotype of another tissue-specific NK cell; the resolution of challenges related to NK cell persistence in vivo from a tissue environment perspective; the modulation of NK cell cytokine production; and the identification of tissue-specific antigen exposure. In conclusion, a comprehensive understanding of the receptor expression patterns and their impact on NK cell function could significantly enhance our ability to predict functional outcomes and optimize the clinical use of NK cells in cancer immunotherapy, thereby improving patient prognosis.

The manuscript presented here comprises part of the studies presented by ZD in his thesis under the supervision of AS for the award of the degree of Doctor of Philosophy at Meharry Medical College. The authors thank Girish Rachakonda, J. Shawn Goodwin, Harshana Rajakaruna, and Siddharth Pratap for their feedback, which helped refine the manuscript. The graphic abstract was produced under agreement number NI268WLYHU on Biorender.com.

The study protocol was thoroughly reviewed and received approval from the Institutional Animal Care and Use Committee (IACUC) at Meharry Medical College, under Approval No. #16-07-582, dated August 25, 2020.

The authors declare that the research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. The funding bodies had no role in the design or writing of the manuscript.

This work was supported by funds to A.S. from the following National Institutes of Health (NIH) grants: U54 CA163069, U54 MD007593, SC1 CA182843, SC1 CA182843-07s1, and R01 CA175370. Z.T.D. is supported by the NIH RISE training grant 5R25GM059994-22. The Meharry cores are supported by NIH grants G12 MD007586, R24 DA036420, and S10 RR0254970.

Conceptualization: Z.T.D. and A.S.; methodology: Z.T.D., S.G.O., and A.S.; formal analysis: Z.T.D. and S.G.O.; investigation and data curation: Z.T.D., S.G.O., and T.K.; resources and funding acquisition: A.S; writing – original draft: Z.T.D.; writing – review and editing: Z.T.D., S.G.O., T.K., A.I., and A.S.; and supervision and project administration: A.I. and A.S. All authors contributed to the article and approved the submitted version.

All data generated or analyzed during this study are included in this article and its online supplementary material. Further inquiries can be directed to the corresponding author.