Murine hepatocyte cell lines promote expansion and differentiation of NK cells from stem cell precursors



While fetal liver is a major hematopoietic organ, normal adult liver provides a suitable microenvironment for a variety of immune cells and, in several pathological conditions, may become a site of extramedullary hematopoiesis. The direct influence of hepatocytes on hematopoietic cell differentiation is poorly understood. We have previously reported that the Met murine hepatocyte (MMH) untransformed hepatocytic lines retain several morphological and functional features of hepatocytes in vivo and are able to support the survival, self-renewal, and differentiation of hematopoietic precursors in a cell-cell contact system. Here we report the effects of soluble factors released by MMH lines on bone marrow–derived cells. Lymphohematopoietic cells were cultured in two different cell contact-free systems: transwell inserts on MMH feeder layers, and MMH conditioned medium (MMH-CM). Both culture systems were able to promote a substantial expansion of bone marrow-derived cells and their differentiation to natural killer (NK) cells that express the NK1.1 and U5A2-13 markers. Purified hematopoietic stem cells (Sca-1+Lin-), either plated as a bulk population or as single cells, were also able to differentiate into NK cells, when cultured in MMH-CM; thus, soluble factors secreted by MMH lines promote the expansion and differentiation of NK precursor cells. MMH-CM-derived NK cells are functionally active; stimulation by interleukin (IL)-12 together with IL-18 was required to induce interferon-gamma (IFNγ) expression and to enhance their cytotoxic activity. In conclusion, our findings may imply a direct role of hepatocytes in NK cell development, and the system we have used may provide a tool for studying the molecular mechanisms of NK cell differentiation. (HEPATOLOGY 2004;39:1508–1516.)

The liver is an organ with its own unique subset of resident immune cells. Most of hepatic lymphocytes are innate effector cells, such as natural killer (NK), NK-T and γδ T cells. These intrahepatic lymphocytes are generally considered to play important roles in antitumor, antiviral and pathogenic inflammatory responses.

The fetal liver is an active site of the production of lymphoid precursors1–3; whether the adult liver is a site of differentiation of lymphoid cells is a topic of considerable interest. Several lines of evidence support the existence of an extrathymic T cell differentiation pathway in the liver of the adult mouse.4–6 More importantly, adult mouse liver contains hematopoietic stem cells that give rise to leukocytes of all lineages.7–9 Moreover, the hepatic environment may also influence the distribution of T cell subpopulations present in the liver without necessarily promoting their differentiation. Cytokines, such as interleukin (IL)-2, IL-12, IL-15, and IL-18, can induce the selective expansion of particular T cell subpopulations and may cause phenotypic changes of infiltrating cells.10–14 During liver disease, such as viral hepatitis infection, the increased production of any of these cytokines may result in expansion of local populations of lymphocytes and/or increased recruitment of circulating lymphocytes.15–20

In vitro studies have emphasized the hematopoietic supportive role of stromal cells; they have focused on the nonhematopoietic adherent cell component obtained from long-term cultures.21, 22 Several stroma-dependent in vitro systems have been developed for the differentiation and expansion of specific hematopoietic cell lineages, such as hematopoietic stem cells,23 early B lymphocytes,24 and NK cells.25, 26 On the other hand, the identification of several cytokines that have pivotal effects on immature hematopoietic cells has led to the development of stroma-free, cytokine-driven specific lineage systems.27, 28

However, in vitro studies on the direct role of hepatocytes in determining the cellular fate of hematopoietic cell precursors have been hampered by difficulties in growing stable hepatocytic cell lines that retain differentiated hepatic functions. We have overcome this problem by taking advantage of hepatocyte cell lines generated from liver explants of transgenic mice that express an active truncated form of the human Met receptor (cyto-Met).29 Several immortalized untransformed cell lines, known as MMH (Met murine hepatocyte), were shown to retain in vitro several features characteristic of in vivo hepatocytes. Indeed, we have shown that MMH maintains the morphology, epithelial polarity, transcriptional factors, expression of products, and complex physiological functions characteristic of mature hepatocytes.29–32 In particular, MMH enables the role of hepatocytes in promoting and sustaining hematopoiesis to be studied. We have reported that hepatocytic MMH lines express a wide spectrum of hematopoietic growth factors and sustain, by cell-cell contact, long-term proliferation and differentiation of hematopoietic cells of fetal liver.33

We report here our evaluation of the ability of the soluble factors released by MMH to sustain the expansion and differentiation of specific subsets of lymphohematopoietic cells. Experiments using a cell contact-free system, without the addition of exogenous cytokines, indicate that the soluble factors released by MMH promote the expansion of hematopoietic stem cells and their differentiation into NK cells.


NK, natural killer; IL, interleukin; MMH, Met murine hepatocyte; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; PBS, phosphate-buffered saline; MMH-CM, MMH-conditioned medium; IMDM, Iscove's modified Dulbecco's medium; BM, bone marrow; SP, side population; FACS, fluorescence-activated cell sorter; IFN, interferon; RT-PCR, reverse-transcriptase polymerase chain reaction.

Materials and Methods


Five- to 10-week-old C57BL/6 mice (Charles River, Lecco Italy) received human care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals,” prepared by the National Academy of Sciences and published by the National Institutes of Health.

Culture Conditions for the Production of Conditioned Media.

MMH (E14, D3, and D6 lines) cells were grown in RPMI-1640 as previously described.32 HuH7 cells were grown at 37°C in 5% CO2 and Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mmol/L L-glutamine, 100 u/mL penicillin, and 100 μg/mL streptomycin (Gibco, Carlsbad, CA). To obtain MMH-conditioned medium (MMH-CM) and HuH7-conditioned medium (HuH7-CM), semiconfluent (70%) cultures were washed once with phosphate-buffered saline (PBS) and the respective media were replaced by Iscove's modified Dulbecco's medium (IMDM), supplemented with 10% FBS and antibiotics. After 48 hours, medium, MMH-CM or HuH7-CM, was filtered (0.2 μm) and used.

Isolation and Culture Conditions of Splenic and Bone Marrow Cells.

Spleens were washed once with PBS containing 5% FBS, manually homogenized, and filtered through a mesh (70 μm diameter). Recovered splenic cells were washed twice with PBS and mononuclear cells were obtained by Lympholyte density gradient centrifugation (Cedarlane, Hornby, Canada).

Bone marrow (BM) cells were isolated from murine femurs by flushing once with PBS supplemented with 5% FBS; they were then passed through a nylon mesh (70 μm), centrifuged once, and plated in complete IMDM. Both splenic and BM cells were incubated for 1 hour on culture dishes; nonadherent cells were used.

BM cells or splenocytes were cultured at 37°C in 5% CO2 and MMH-CM, HuH7-CM, or IMDM at a starting density that ranged from 20 to 40 × 103 cells/cm2 in a T-25 flask (Falcon-BD, Franklin Lakes, NJ). Starting at day 3, half of the medium was replaced with fresh MMH-CM, HuH7-CM, or IMDM every 2 days. To retain floating cells, the replaced medium was centrifuged, and cells were resuspended in fresh medium, and then added to the original culture.

To generate the transwell culture system (0.4-μm filter; Falcon), 6-well plates were used. The lower chamber contained the semiconfluent MMH or HuH7 feeder layer. Fresh IMDM was replaced and after 48 hours 7–12 × 103/ cm2 BM cells were added to the upper chamber. Starting at day 3 of coculture, the medium was replaced with fresh IMDM every 2 days.

Isolation and Culture Conditions of Sca-1+Lin- and Side Population (SP) Stem Cells.

To isolate the Sca-1+lineage- (Sca-1+Lin-) population, BM cells were incubated with phycoerythrin (PE)-conjugated mAbs to Mac-1, CD45R/B220, CD11c, Gr-1, Ter-119, CD4, CD8, and U5A2-13, and fluorescein (FITC)-conjugated mAb to Sca-1, at 4°C for 30 minutes. Cells were washed twice with washing medium (PBS, 1 % bovine serum albumin, 0.1 % sodium-azide) and the Sca-1+Lin- population was electronically gated and sorted using a fluorescence-activated cell sorter (FACS Vantage, Becton Dickinson, San Jose, CA). The purity of this population was analyzed using a FACS Calibur.

For the isolation of the SP, Hoechst staining was undertaken as described by Goodell.34 Cells were analyzed and sorted using a MoFlo cell sorter (DakoCytomation, Inc., Milan Italy). Purity was consistently >90%.

Collected cells (Sca-1+Lin- and SP) were plated at a density that ranged from 3 to 18 × 103/cm2. Single Sca-1+Lin- cells were sorted into the individual wells of a 96-well, round-bottom tissue culture plate using the Clon-Cyt system of FACS Vantage. All cultures were incubated for 14 days in MMH-CM; the medium was replaced every 3 days.

Monoclonal Antibodies and Cytokines.

The monoclonal antibodies PE-conjugated anti-CD45R/B220(RA3-6B2), anti-Mac-1(M1/70), anti-CD11c(HL3), anti-CD4(H129.19), anti-CD8(53-6.7), anti-Ter119(TER119), anti-Ly6G(RB6-8C5), NK/NK-T cell antigen(U5A2-13), FITC-conjugated anti-NK1.1(PK136), anti-CD3(17A2), anti-Sca-1 (E13-161.7), APC-conjugated PanNK(DX5), PE-Cy5-conjugated anti-CD45(30-F11), and biotinylated anti-MHC-II(AF6) were all purchased from Pharmingen (Erembodegem, Belgium). FITC-conjugated anti-F4/80(MCA497F) was obtained from Serotec (Oxford, UK). Murine rIL-12 and rIL-18 were purchased from Peprotech (London, UK).

Flow Cytometry Analysis for Surface Antigens.

Staining was undertaken in 96-well U-bottom plates. Cells were washed in washing medium and treated with FcBlock (Pharmingen) for 5 minutes at 4°C to block Fc receptors prior to incubation for 15 minutes at 4°C with conjugated mAbs. Cells were washed twice in washing medium and fixed in 1% paraformaldehyde. Flow cytometry was undertaken using a FACS Calibur (Becton Dickinson). For each sample 104 events were acquired and analyzed using the CellQuest software (Becton Dickinson). Quadrants were determined using appropriate isotype controls.

Cytotoxicity Assay.

The murine NK-susceptible YAC-1 target cells were labeled with 51Cr and incubated for 4 hours in U-shaped microtiter wells (1 × 104 cells/well) with MMH-CM-derived BM effector cells at different effector:target cell ratios (E:T). 51Cr release was determined in the supernatants of triplicate cultures; specific lysis was calculated as the ratio: (experimental 51Cr-release − spontaneous 51Cr-release) / (maximum 51Cr-release [targets in 1% Triton] − spontaneous 51Cr-release) × 100.

Reverse-Transcriptase Polymerase Chain Reaction (RT-PCR).

Total RNA was extracted from either MMH lines (E14, D3, and D6) or BM cells cultured in MMH-CM for 7 or 14 days, using Trizol reagent (Invitrogen, UK), according to the manufacturer's protocol. Oligo-dT-primed complementary DNA was synthesized from 1 μg of RNA. Complementary DNA was amplified by PCR using specific primers for: murine β-actin (forward-5′-ATGGATGACGATATCGCTGCG-3′,reverse-5′-ATCTTCATGAGGTAGTCTGTCAGG-3′);IL-15 (forward-5′-TCCCTAAAACAGAGGCCAAC-3′, reverse-5′-TTTCTCCTCCAGCTCCTCAC-3′); perforin (forward-5′-TGCTACACTGCCACTCGGTCA-3′,reverse-3′-TTGGCTACCTTGGAGTGGGAG-5′); IL-7 (forward-5′-AGAGTGTACTGATGATCAGC-3′, reverse-5′-GCAGTTCACCAGTGTTTGTG-3′); and interferon (IFN) γ (forward-5′-GACAATCAGGCCATCAGCAAC-3′, reverse-5′-CGCAATCACAGTCTTGGCTAA-3′). The reaction mixtures were first heated to 95°C for 15 minutes, and then subjected to 30 amplification rounds, each lasting of 30 seconds at 94°C, followed by 1 minute at 57°C for IL-15 and perforin, 55°C for IL-7, and IFNγ, and 60°C for β-actin, and then 1 minute at 72°C.


Soluble Factors Released by MMH Promote the Expansion of Bone Marrow–Derived Cells and Their Differentiation Into NK Cells

To explore the effects of soluble factors released by MMH on the proliferation and differentiation of specific subpopulations of hematopoietic cells, we developed two cell-contact free culture conditions: (1) murine bone marrow cells cocultured in Transwell inserts on semiconfluent MMH feeder layers (MMH-Transwell), and (2) BM cells cultured in MMH-CM. The HuH7 cell line (HuH7-Transwell and HuH7-CM) or IMDM alone were used as controls. No exogenous cytokines were added to any culture.

BM-derived cells underwent an expansion of 5 ± 1- and 6 ± 2-fold when cultured for 2 weeks in MMH-Transwell or in MMH-CM, respectively (Fig. 1). In contrast, BM cells did not increase in number and did not survive more than 3–6 days when cultured in IMDM alone, HuH7-CM, or HuH7-Transwell.

Figure 1.

In vitro expansion of bone marrow-derived cells. Bone marrow cells, cultured as indicated in the inset, were counted at the indicated times. Results are shown as the mean SD of the ratio (number of cells counted after culture/number of seeded cells) (n = 4). MMH, Met murine hepatocyte; MMH-CM, MMH conditioned medium; HuH7-CM, HuH7 conditioned medium; IMDM, Iscove's modified Dulbecco's medium.

To characterize the subpopulations of BM cells undergoing expansion, we analyzed the phenotype of the resulting cells after 14 days of culture in MMH-CM. FACS analysis was undertaken using mAbs directed against surface markers preferentially expressed by T lymphocytes, B lymphocytes, NK cells, dendritic cells, myeloid cells, macrophages, and erythrocytes. The immunological activation of BM-derived cells was evaluated using a mAb anti-MHC-II. The common leukocyte antigen CD45 was used to verify that all cells had a hematopoietic origin.

Table 1 compares the phenotype of cells either uncultured, soon after the isolation, or after culture for 14 days in MMH-CM. Interestingly, after culture in MMH-CM the relative percentage of BM cell subpopulations had changed appreciably. The majority of cells became positive for markers associated with the NK phenotype (75± 12% was NK1.1+); 13 ± 5% was positive for CD11c, a dendritic cell marker shared by other cell populations, including NK cells.35, 36 Almost all cells expressed high levels of MHC-II, suggesting that they had a functionally mature/activated immunological phenotype. Other cell subsets, such as T cells, B cells, macrophages, and erythrocytes, were not supported by MMH-CM.

Table 1. Phenotypic Characterization of BM-Derived Cells
Cell Population (phenotype)t = 0 (%)t = 14 (%)
  1. Abbreviation: ND, not determined.

  2. NOTE. BM cells were stained for FACS analysis either after isolation (time [t] = 0) or after 2 weeks of culture in MMH-CM (t = 14). Results are shown as the mean number of positive cells ± SD (n = 4).

Leukocytes (CD45+)≥8595 ± 3
T Cells (CD3+)5 ± 1<1
B cells (B220+)19 ± 8<1
Macrophages (F4/80+)ND<1
Erythrocytes (Ter-119+)12 ± 1<1
Myeloid cells (Mac-1+)53 ± 695 ± 3
Dendritic cells (CD11c+)1 ± 0.513 ± 5
Natural killer (NK1.1+)4 ± 275 ± 12
(MHC-II+)1 ± 295 ± 3

Cytometric analysis of BM-derived cells cultured in MMH-Transwell revealed no significant differences to results obtained using MMH-CM (data not shown). Since MMH-CM permits easier cell manipulation, all subsequent experiments were undertaken using MMH-CM.

We analyzed the kinetics of the expansion of NK1.1+ cells in MMH-CM in time course experiments. A gradual increase in NK1.1+ cells started at day 3 and peaked at day 12, when 82 % of the cells expressed the NK1.1 marker (Fig. 2).

Figure 2.

Time course analysis of NK1.1+ cells cultured in MMH-CM. BM cells were cultured in MMH-CM. Cells were harvested at the indicated times and stained with anti-NK1.1 mAb. Numbers are expressed as the percentage ± SD of positive cells. NK, natural killer.

These data indicate that MMH-secreted soluble factors are able to promote the expansion of specific BM cell subpopulations and particularly NK1.1+ cells.

Characterization of the Phenotype of BM-Derived NK1.1+ Cells Generated by MMH-CM

To further characterize the NK1.1+ cells generated by MMH-CM, we analyzed the expression of other NK cell-associated and stage-specific markers, as well as the expression of the perforin gene.

The NK/NK-T cell antigen, recognized by the U5A2-13 mAb, is known to be expressed by 85% of NK1.1+CD3+ cells and by 55% of NK cells in mouse liver.37, 38 After 14 days in MMH-CM, almost all NK1.1+ cells were U5A2-13+ (Fig. 3A). All U5A2-13+ cells were CD3−, indicating that soluble factors released by hepatocytes support NK but not NK-T cell expansion. Moreover, all NK1.1+ cells coexpress integrin Mac-1 (Fig. 3A), a marker associated with a mature NK phenotype.39 However, these cells did not express integrin α2β1 (DX5), another NK-associated marker specific for cell maturation (Fig. 3A).

Figure 3.

Natural killer (NK) phenotypic analysis and perforin-expression of BM-derived cells cultured in MMH-CM. (A) Two-color flow cytometric analysis of the coexpression of NK1.1 and U5A2-13, Mac-1 or DX5 markers, or the coexpression of CD3 and U5A2-13 markers was undertaken. (B) RT-PCR analysis was undertaken using RNA extracted from BM-derived cells that had been cultured for 7 or 14 days in MMH-CM. Primers specific for perforin and β-actin were used.

Finally, RT-PCR analysis was undertaken using RNA extracted from BM cells that had been cultured in MMH-CM. The analysis indicated that perforin mRNA was already detectable by day 7 of culture and was clearly evident at day 14 (Fig. 3B), thereby confirming a progressive process of NK cell expansion and/or differentiation.

MMH-CM Promotes the NK Differentiation of Hematopoietic Stem Cell Precursors

We next addressed the question whether MMH-CM promotes either the proliferation of preexisting mature NK cells or the expansion and differentiation of NK cell precursors.

NK progenitor cells are more abundant in the BM than in the spleen; the spleen is not considered to be an organ essential for the generation of NK cells.40 We cultivated mature splenocytes in MMH-CM. We observed a 430-fold expansion of the U5A2-13+CD3− population of BM-derived cells, and only a 3-fold expansion of spleen-derived cells (Table 2). Thus, MMH soluble factors appear to promote the expansion and differentiation of progenitor cells rather than the proliferation of preexisting mature NK cells.

Table 2. MMH-CM Induces NK Cells Expansion From Different Lymphohematopoietic Sources
Source of Hematopoietic Progenitor CellsU5A2-13+CD3- (% t = 0)U5A2-13+CD3- (% t = 14)Fold Increase of U5A2-13+CD3- Cell Numbers
  1. NOTE. BM and splenic cells were stained for FACS analysis either after isolation (time [t] = 0) or after 2 weeks of culture in MMH-CM (t = 14). Results are shown as the mean number of positive cells ± SD (n = 4). Fold increase of U5A2-13+CD3- cells is reported as the mean ± SD of the ratio (number of cells counted after MMH-CM culture)/(number of seeded cells).

Bone marrow2 ± 188 ± 16430 ± 262
Spleen6 ± 132 ± 103 ± 1

To prove this hypothesis, we analyzed the effect of the MMH-CM on purified hematopoietic stem cells. Either the SP34 or Sca-1+Lin- cells were isolated and cultured as described. After two weeks MMH-CM promoted a significant expansion of SP- and Sca-1+Lin-derived cells (>100-fold and >40-fold, respectively) (Table 3). Interestingly, a large proportion of these cells were U5A2-13+ (74% and 43% for SP- and Sca-1+Lin-derived cells, respectively; Table 3), indicating that hematopoietic progenitor cells differentiated into the NK cell phenotype when cultured in MMH-CM.

Table 3. MMH-CM Induces NK Cells Differentiation From Hematopoietic Stem Cells
Source of Hematopoietic Stem CellsU5A2-13+ (% t = 14)Fold Increase of Total Cell Numbers
  1. NOTE. Hematopoietic progenitor cells were stained for FACS analysis after 2 weeks of culture in MMH-CM (time [t] = 14). Results are shown as the mean number of positive cells ± SD (n = 3). Fold increase of MMH-CM-derived cells is reported as the mean ± SD of the ratio (number of cells counted after culture)/(number of seeded cells).

Side population74 ± 5103 ± 28
Sca-1 + Lin-43 ± 2147 ± 37

Although the population of sorted Sca-1+Lin- cells used in our experiments was 95% pure (Fig. 4A) and Lin+ cells were almost undetectable in this population (0.5%), we could not rule out the possibility that NK cells arose from the selective proliferation of contaminating NK cells. To address this issue, single sorted Sca-1+Lin- cells were isolated, deposited by FACS into individual wells, and cultured in MMH-CM. After 14 days, 27 out of the 120 single cells plated yielded a clone; among them 6 were analyzed singularly and shown to be U5A2-13+ (74 ± 10%) (Fig. 4B). Similar results were obtained using the pool of remaining clones (data not shown). Because of the purity of Sca-1+Lin- cells (95%), the 22.5% efficiency of cloning obtained cannot be ascribed to the proliferation of contaminating U5A2-13+ cells.

Figure 4.

MMH-CM generates U5A2-13+ NK cells from a single Sca-1+Lin- cell. (A) Purification of BM-derived Sca-1+Lin- cells. Cells were sorted from the gate R3 (left panel) and their purity determined by FACS reanalysis (right panel); (B) single Sca-1+Lin- sorted cells were cultured in MMH-CM for 2 weeks. Six independent clones (average of 5,000 cells each) were analyzed by FACS for the presence of the U5A2-13 marker. One representative plot is shown. Percentages of each quadrant are indicated. SSC, side scatter.

Taken together, these data demonstrate that the soluble factors released by MMH cells drive hematopoietic stem cell differentiation into the NK cell lineage.

Functional Characterization of BM-Derived NK Cells Generated in MMH-CM Cultures

NK cells mediate MHC-unrestricted cytotoxicity and produce cytokines, such as IFNγ and tumor necrosis factor-α, that further enhance their cyototoxic effects and contribute to modulation of both innate and acquired immune responses. We determined whether MMH-CM-derived NK cells produce IFNγ and mediate lytic activity. In view of the striking synergy between IL-12 and IL-18 in inducing IFNγ,41 BM cells were cultured for 14 days in MMH-CM and then stimulated for 24 hours with these 2 cytokines.

RT-PCR analysis revealed that BM-derived NK cells expressed IFNγ only when stimulated by IL-12 together with IL-18 (Fig. 5A). Basal cytotoxic activity of BM-derived cells against NK-sensitive YAC-1 target cells was low after 7 or 14 days of culture in MMH-CM (7% and 15% specific lysis at an E:T ratio of 100:1, respectively), but stimulation by IL-12 together with IL-18 strongly enhanced their lytic activity (17% and 42% after 7 or 14 days of culture, respectively) (Fig. 5B).

Figure 5.

IL-12 together with IL-18 stimulation induces interferon-γ expression and enhances cytotoxic activity of BM-derived NK cells cultured in MMH-CM. (A) RT-PCR analysis was undertaken using RNA extracted from BM cells that had been cultured for 14 days in MMH-CM, either untreated or stimulated with IL-12 (50 ng/mL), together with IL-18 (100 ng/mL) for 24 hours. Primers specific for interferon-γ and β-actin were used. (B) BM-derived cells cultured for 7 or 14 days in MMH-CM were used as effectors in a classical 51Cr-release assay against natural killer—sensitive YAC-1 target cells. BM-derived cells cultured in MMH-CM for 7 days (open triangle) or 14 days (open circle) mediated low cytotoxic activity; 24 hours of stimulation by IL-12 together with IL-18 greatly enhanced their lytic activity (day 7, closed triangle; day 14, closed circle). Experiments were undertaken in triplicate. Results are expressed as means of percentage of specific lysis at each effector:target cell ratio. IL, interleukin; IFN, interferon.

Taken together, our data show that the two major NK cell activities, cytotoxicity and IFNγ production, are both mediated by NK cells generated in MMH-CM.

MMH Cell Lines Express IL-15 and IL-7

We have previously shown that MMH cells express cytokine genes involved in the survival and self-renewal of early progenitor cells (stem cell factor and Flt3 ligand), as well as those that act at different stages of progenitor differentiation (IL-3, leukemia inhibitory factor, IL-6, granulocyte-macrophage colony-stimulating factor, granulocyte colony-stimulating factor, thrombopoietin).33 Since IL-15 and IL-7 have been shown to be important for NK cell maturation,42–44 we determined their expression in different MMH lines. RT-PCR was undertaken using RNA extracts from E14, D3, and D6 lines that had been cultured for 14 days in IMDM. All three MMH lines expressed IL-15 and IL-7 (Fig. 6). Therefore, MMH lines express cytokines, such as Flt3 ligand, stem cell factor, granulocyte-macrophage colony-stimulating factor, IL-7, and IL-15, that are known to be involved in the induction of proliferation and differentiation of NK cells. These findings provide an explanation for the ability of MMH to induce the differentiation of NK cells from stem cell precursors.

Figure 6.

MMH lines express IL-15 and IL-7. RT-PCR was undertaken using total RNA extracted from E14, D3, and D6 MMH lines that had been cultured for 14 days in IMDM without the addition of exogenous cytokines. Primers specific for IL-15, IL-7, and β-actin were used. IL, interleukin.


The hepatocytic cell lines MMH were shown to retain several phenotypic and functional features that characterize hepatocytes in vivo.29–33 They were shown to sustain survival, self-renewal, and differentiation of hematopoietic precursors in a cell-cell contact system.33 Therefore, MMH cells constitute an in vitro system for evaluating the influence of the hepatic environment on the proliferation and differentiation of specific subpopulations of lymphohematopoietic cells. These considerations provided a basis for our analysis of the effects of soluble factors released by MMH on hematopoietic cells derived from different sources.

The major finding in this study is the demonstration that soluble factors released by MMH are able to promote the expansion of hematopoietic precursor cells and their differentiation into NK cells. This process occurs in a cell contact-free system and without the addition of exogenous cytokines.

We observed that culturing BM cells in MMH-CM led, within 12 to 14 days, to the selective expansion of a specific subpopulation of cells, most of which were positive for Mac-1, NK1.1 and negative for CD3, B220, Ter-119, and F4/80 markers. These cells were further characterized using other NK cell-associated antigens. The NK1.1+ cells were stained using U5A2-13 mAb. This mAb recognizes an antigen, which has not yet been characterized and which is expressed by most NK-T cells and by 55% of NK cells in the liver of various strains of mouse.37, 38 MMH-CM promoted a 430-fold increase in the subpopulation of U5A2-13+ cells; these cells constitute about 88% of surviving cells after 14 days of culture. These cells were shown to be CD3-, and, hence, did not represent an expansion of NK-T cells. Moreover, MMH-CM-expanded NK1.1+ cells were found to be positive for integrin Mac-1, but negative for integrin DX5, the well-known pan-NK cell marker. In a recent study, Kim and colleagues39 described, in vivo, the sequential acquisition and concomitant down-regulation of several markers during maturation of mouse NK cells. Both Mac-1 and DX5 integrins were described as markers of mature NK cells. In contrast, Arase45 reported that when splenic NK1.1+ cells were cultured with IL-2, they progressively lose reactivity with integrin DX5 antibody as a consequence of cellular proliferation. Thus, it seems reasonable to suppose that the experimental conditions we used may be responsible for the lack of detectable α2β1 integrin.

We found that MMH-CM induces the generation of NK cells from the stem cell compartment when cultured either as a bulk population or as a single cell; thus, these cells arise as a result of differentiation of progenitor cells rather than from the proliferation of preexisting NK cells. The generation of NK cells is induced solely by the addition of MMH-CM. Thus, MMH lines secrete all the essential factors needed for this process. Several conditions necessary for the differentiation of adult stem cells into NK cells have been described. In vitro systems require either the direct contact of progenitor cells with stroma cell lines (e.g., AFT024)46 or the addition of several cytokines, such as Flt3 ligand and IL-15.28, 47 Although our analysis indicated that MMH lines express an appreciable repertoire of cytokines known to support NK cell ontogeny, such as IL-7, IL-15, Flt-3 ligand, stem cell factor and granulocyte-macrophage colony-stimulating factor, further studies are required to determine the relevance of individual cytokines in MMH-mediated development of NK cells.

We also showed that MMH-CM-derived NK cells express the critical cytolytic mediator perforin gene, exhibit low cytotoxic activity against YAC-1 target cells, and respond to a combination of IL-12 and IL-18 by increasing IFNg gene expression and enhancing their cytotoxicity. Interestingly, these features are reminiscent of normal liver NK cells described by Trobonjaca and colleagues48; such cells exhibit low cytolytic activity against tumor cells and produce IFNγ only in response to stimulation with IL-12 and IL-18.

The number of liver-associated NK cells increases in response to either pathological conditions or biological agents49; currently, it is unknown whether NK cells are recruited as mature cells or as NK precursors followed by differentiation in situ. Indeed, the presence of CD34+CD56+ NK progenitor cells suggests that normal adult liver may be the site of local NK and NK-T cell maturation.50 It has been shown that in response to viral infections or immunological modulation, an increase in the number of NK cells is due, at least in part, to proliferation of a liver-associated NK cell population.16 However, the ablation of BM cellularity inhibits such an increase in NK cells.49 Our data demonstrate that the soluble factors released by MMH cell lines are able to sustain the differentiation of hematopoietic stem cells into NK cells. In vivo hepatocytes may provide a suitable environment for the differentiation of NK cells; an increased number of liver-associated NK cells may be due partially to the differentiation of NK cell precursors in situ. Further in vivo studies are needed to verify this hypothesis. Finally, hepatocytic cell lines MMH constitute a suitable model for in vitro studies of NK cell development and differentiation.


This work was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC), Ministero della Sanità (Ricerca Finalizzata), and MIUR Ministero dell'Università e Ricerca Scientifica. We thank Professor Angela Santoni for helpful discussion and critical reading of the manuscript, and Sarah Yeomans for editing it. We also thank Tittania Musella for her kind assistance, Dr. Federico Martini for assistance with cell sorting, and Gabriele De Luca for his care of experimental animals.