Rat CD4⁺CD8⁺ Macrophages Kill Tumor Cells through an NKG2D- and Granzyme/Perforin-Dependent Mechanism

Rats

Closed-colony Wistar rats were purchased from Charles River. The EGFP-transgenic rats ubiquitously expressing GFP (5) were obtained from the YS Institute (Utsunomiya, Japan) and Tohoku University (Sendai, Japan). All experiments using rats were done according to the Guidelines for the Care and Use of Laboratory Animals in the Hokkaido University Graduate School of Medicine.

Cell lines

The cloned malignant epithelial thymoma cell line MT was established from the transgenic rats carrying the human T cell leukemia virus type I pX gene in our laboratory (6). The rat mammary cancer cell line SST-2 was provided from Dr. J. Hamada (Hokkaido University, Sapporo, Japan). The COS-7 cell line, rat colon cancer cell line RCN-9, rat bladder cancer cell line NBT-II, and murine mastocytoma cell line P-815, which expresses FcγRs, were purchased from RIKEN Cell Bank (Tsukuba, Japan). MT, SST-2, and RCN-9 were derived from F344 rats and NBT-II from Wistar rats.

Antibodies

Murine mAbs used were anti-rat CD4 (OX-35; BD Pharmingen), anti-rat CD8 (OX-8; BD Pharmingen), anti-rat CD11c (8A2; Serotec), anti-rat CD25 (OX-39; BD Pharmingen), anti-rat CD68 (ED-1; Serotec), anti-rat CD80 (3H5; BD Pharmingen), anti-rat CD103 (OX-62; BD Pharmingen), anti-rat CD163 (ED-2; Serotec), anti-rat MHC class II (MRC OX-6; Serotec), anti-rat Fas ligand (MFL4; BD Pharmingen), and anti-rat NKR-P1A (10/78; BD Pharmingen). The hamster mAb for rat CD49b (Ha1/29; BD Pharmingen) was also used. Rabbit and goat polyclonal Abs used were anti-FLAG (Sigma-Aldrich) and anti-granzyme B (N-19; Santa Cruz Biotechnology), respectively. Mouse IgG1 and IgG2 (CBL600P and CBL601P, respectively; Chemicon International) and rabbit or goat IgG (Sigma-Aldrich) served as controls.

Immunization of rats

Killed tuberculosis germs were added to IFA (Sigma-Aldrich) to reach the concentration of 100 mg/ml, which were then emulsified with an equal volume of PBS. The resultant emulsion containing killed tuberculosis germs was inoculated into bilateral footpads (100 μl/site) of 8-wk-old rats.

Tumor inoculation into rats

Eight-week-old Wistar rats (n = 9) were divided into two groups. One group of rats was immunized with adjuvants containing killed tuberculosis germs (n = 5) and the other group was not treated (n = 4). Seven days later, syngenic NBT-II carcinoma cells were inoculated s.c. into the back of each rat (1 × 10⁷/0.5 ml per site). Then, the tumor size was measured three-dimensionally (length, width, and height) every other day and the volume was calculated as (length × width × height)/2.

Flow cytometry (FCM) and MACS

Peripheral blood cells were stained by the direct method without removal of serum. After reaction with Abs, erythrocytes were depleted by treatment with ammonium chloride. Splenic mononuclear cells isolated with Histopaque-1083 (Sigma-Aldrich) were resuspended in PBS supplemented with 3% FBS and then stained. Expression of cell surface molecules was analyzed using a FACSCalibur (BD Biosciences) with CellQuest software (BD Biosciences). MACS was done using a Magnetic Cell Separator (Miltenyi Biotec) as previously described (1).

Isolation of CD4⁺CD8⁺ macrophages and NK cells

Eight-week-old rats were immunized with the adjuvant containing killed tuberculosis germs. One week later, mononuclear cells were isolated from the spleen using Histopaque-1083 and incubated in plastic dishes for 20 min at 37°C. Resultant adherent cells were collected and then divided into CD8⁻ or CD8⁺ cells using the MACS system. We previously showed that the majority of CD8⁺ cells were CD4⁺CD8⁺ macrophages (1). NKR-P1A⁺ cells were collected from nonadherent mononuclear cells using the MACS system and used as splenic NK cells.

Cytotoxic assay

Effector cells were added to the culture of tumor cells. After incubation for 6 or 18 h, cytotoxicity was measured using a CytoTox 96 test kit (Promega).

Adhesion assay

Tumor cells were incubated in 8-well chamber slides (Nalge Nunc) to form a monolayer at 37°C. CD4⁺CD8⁺ macrophages isolated from EGFP-transgenic rats were added to the culture. After incubation for 2 h, the chamber slides were gently washed with PBS. Then, GFP⁺ cells that adhered to the tumor cell sheet were observed using a fluorescence microscope (ECLIPSE E600; Nikon).

RT-PCR

Total RNAs were extracted from cells using a RNeasy Mini Kit (Qiagen) and then reverse-transcribed with M-MLV reverse transcriptase (Invitrogen Life Technologies). PCR was done using the cDNAs, 2 mM dNTP mix (GeneAmp dNTPMix; Applied Biosystems), TaqDNA polymerase (AmpliTaq Gold; Applied Biosystems), and primer sets for 30 cycles of 95°C for 40 s, 56°C for 40 s, and 72°C for 40 s. The primer sequences for PCR are listed in Table I⇓.

Construction of the rat RAET1 expression plasmid

The entire coding region of rat RAET1, excluding the leader sequence, was obtained by PCR using cDNA derived from the bladder cancer cell line NBT-II. The primer sequences were 5′-CCAAGCTTGTACACTCTCTTAGTTGCAA-3′ (with a HindIII site at its 5′ end) and 5′-GTGGATCCTCAGCTCCCAGTACTTGAAAACA-3′ (with a BamHI site at its 5′ end). After digestion with HindIII and BamHI, the PCR products were ligated to the HindIII/BamHI-digested pFLAG-CMV-3 expression vector (Sigma-Aldrich). Accuracy of the construct was verified by DNA sequencing. This construct, designated as RAET1-pFLAG-CMV-3, enabled the expression of rat RAET1 with an N-terminal FLAG tag when transfected into mammalian cells. DNA for transfection was isolated with a plasmid purification kit purchased from Qiagen.

Transfection

To obtain transient transfectants expressing rat RAET1, COS-7 cells were transfected with the RAET1-pFLAG-CMV-3 plasmid using the FuGENE 6 Transfection Reagent (Promega). The efficiency of transfection was monitored by FCM using the anti-FLAG Ab.

Production of rat NKR-P2 and human Ig (hIg) fusion proteins

The entire extracellular region of rat NKR-P2 was obtained by PCR using cDNA derived from rat CD8⁺ lymphocytes as templates. The coding sequence for human IgG1-Fc was obtained by PCR using the pFUSE-hIgG-Fc2 expression vector (Invivogen Life Technologies) as templates. The primer sequences for NKR-P2 were 5′-GTGGTACCATCGAGGGCCGCTTGTTATGCAACAAGGAA-3′ (with a KpnI site at its 5′ end) and 5′-CTAGCTAGCTTACACTGCCCTTTTCATAC-3′ (with an NheI site at its 5′ end) and those for IgG1-Fc were 5′-GAAGATCTGTGGAGTGCCCACCTTGCCC-3′ (with a BglII site at its 5′ end) and 5′-GTGGTACCTTTACCCGGAGACAGGGAGA-3′ (with a KpnI site at its 5′ end). After digestion with KpnI and NheI, NKR-P2 PCR products were ligated to the BglII/KpnI-digested IgG-Fc products. This chimeric gene was amplified by PCR using the primers 5′-GAAGATCTGTGGAGTGCCCACCTTGCCC-3′ (sense) and 5′-CTAGCTAGCTTACACTGCCCTTTTCATAC-3′ (antisense). After digestion with BglII and NheI, the PCR products were ligated to the pFUSE-hIgG-Fc2 expression vector devoid of the IgG1-Fc sequence. Accuracy of the construct was verified by DNA sequencing. This construct, designated as NKR-P2-hIg-pFUSE, enabled the expression of rat NKR-P2 dimers, each with an N-terminal human IgG1-Fc tag. The plasmid DNA was transfected into COS-7 cells using the FuGENE 6 Transfection Reagent. Soluble NKR-P2-hIg was purified from the culture supernatant of the transfectants using the HiTrap Protein G HP column (Amersham Biosciences).

Immunocytochemistry

Mononuclear cells separated from rat spleen were cultured in chamber slides (Nalge Nunc International) for 1 h. Resultant adherent cells were fixed using cold acetone for 5 min or 4% paraformaldehyde for 15 min at room temperature and then stained. After washing with PBS, the slides were mounted in Fluorescent Mounting Medium (DakoCytomation). Immunofluorescence was detected using the fluorescence microscope.

Immunohistochemistry

After fixation by 10% phosphate-buffered formaldehyde, tissues were embedded in paraffin blocks and cut into 4-μm sections. For immunohistochemistry, the sections were reacted with the anti-rat CD68 Ab (ED-1) at a 1/250 dilution at 4°C overnight. After removing the unbound Ab, the sections were exposed to the alkaline phosphatase-labeled anti-mouse IgG Ab (Histofine AP; Nichirei) at room temperature for 30 min. New Fuchcin (Envision AP; DakoCytomation) was used for detection of CD68 representing alkaline phosphatase activity. After washing, the sections were reacted with the biotinylated anti-rat CD8 Ab (OX-8) at a 1/250 dilution at 4°C overnight. To detect the expression of CD8, the avidin-biotin immunoperoxidase kit (LSAB2 kit; DakoCytomation) was used.

Immunofluorescent staining

For immunofluorescent staining, formaldehyde-fixed paraffin-embedded sections were reacted with the anti-rat CD68 Ab (ED-1, mouse IgG1) and goat anti-serum for granzyme B (N-19) at 4°C overnight. After removing the unbound Abs, the sections were reacted with Alexa 488-labeled anti-mouse IgG and Alexa 568-labeled anti-goat IgG Abs (Invitrogen Life Technologies) at room temperature for 60 min. Fluorescence was detected using Eclipse E600 (Nikon).

Image processing

Microscopic photographs were taken using the objective lens (×40/0.75 numeric aperture) in the DP70 system (Olympus). DP Controller software (Olympus) was used for image processing.

CD4⁺CD8⁺ monocytes/macrophages are a cell population distinct from DCs with cytotoxic phenotypes

CD4⁺CD8⁺ monocytes/macrophages, rat splenic DCs, and IKDCs all exhibit killer activities against tumor cells (2, 3, 4). To confirm that CD4⁺CD8⁺ monocytes/macrophages are a cell population distinct from DCs with cytotoxic phenotypes, we examined the expression of surface markers characteristic of DCs on CD4⁺CD8⁺ monocytes/macrophages. Rat monocyte/macrophage lineage cells were identified in the CD4^medium population (R2 gate) in mononuclear cells (R1 gate) (Fig. 1⇓A) (1). CD4⁺CD8⁺ monocytes/macrophages, which were CD8⁺ and CD4^medium, were abundantly seen in the peripheral blood (52.9% of monocytes) and spleen (26.7% of macrophages) in adjuvant-immunized rats. Histograms were obtained by setting a gate for the CD4^mediumCD8⁺ cells (Fig. 1⇓B). CD4⁺CD8⁺ monocytes in the peripheral blood and CD4⁺CD8⁺ macrophages in the spleen expressed CD11c, CD80, MHC class II, CD68, and CD163. However, they were negative for CD25, CD103 (both are characteristically expressed on rat splenic DCs (2, 7)) or the IKDC marker CD49b (3, 4). Thus, CD4⁺CD8⁺ monocytes/macrophages differ from rat splenic DCs and IKDCs in expression of surface markers.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Phenotypic markers of CD4⁺CD8⁺ monocytes/macrophages. Peripheral blood cells and splenocytes obtained from Wistar rats that had been immunized with adjuvants containing killed tuberculosis germs 1 wk before were analyzed on FCM. In each experiment, at least three rats were used. Representative data are shown. A, Peripheral blood cells and splenocytes were stained with FITC-conjugated anti-CD4 (OX-35) and PE-conjugated anti-CD8 (OX-8) Abs, followed by depletion of erythrocytes. At first, peripheral blood cells were divided according to the forward scatter (FSC) and side scatter (SSC) patterns. Then, mononuclear cells in region 1 (R1) were further divided according to the FL1 (FITC) and FL2 (PE) patterns. The CD4^medium population (R2) contains rat monocyte/macrophage lineage cells. In the right two panels, the expression of CD8 on the CD4^medium cells is shown. The numbers in the panels represent the percentage of CD8⁺ cells in monocytes or macrophages. B, Peripheral blood cells and splenocytes were stained with FITC- or PE-conjugated anti-CD4 (OX-35), PerCP-conjugated anti-CD8 (OX-8), and FITC- or PE-conjugated anti-CD11c (8A2), CD80 (3H5), MHC class II (MRC OX-6), CD68 (ED-1), CD163 (ED-2), CD25 (OX-39), CD49b (Ha1/29), or CD103 (OX-62) Abs, followed by depletion of erythrocytes. Intracellular CD68 molecules were stained after permeabilization of the cells. Unfilled histograms represent the expression profiles of respective markers on CD4⁺CD8⁺ monocytes or macrophages. Gray-filled histograms represent negative controls using isotype-matched Abs.

Next, we examined the morphological features of CD4⁺CD8⁺ macrophages and the expression profiles of TLRs in these cells. CD4⁺CD8⁺ macrophages were morphologically similar to CD8⁻ conventional macrophages, but different from rat splenic DCs displaying characteristic dendrites (Fig. 2⇓A). Recent work has shown that the expression profiles of TLRs differ between DCs and monocyte/macrophage lineage cells (8). For example, TLR3 and TRL10 are expressed at high levels in DCs. When we examined TLR expression profiles, TLR3 and TLR10 were abundantly expressed in rat splenic DCs, but only at low levels in CD4⁺CD8⁺ and CD8⁻ macrophages (Fig. 2⇓B). These observations support our conclusion that the CD4⁺CD8⁺ cells characterized in this study are macrophage/monocyte lineage cells rather than DC lineage cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Morphology and TLR expression profiles of CD4⁺CD8⁺ monocytes/macrophages. Eight-week-old Wistar rats were immunized with adjuvants containing killed tuberculosis germs. One week later, mononuclear cells were separated from the spleen and incubated in plastic dishes for 20 min at 37°C. Resultant adherent cells were collected and divided into CD8⁻ and CD8⁺ cells using the MACS system. They were used as CD8⁻ macrophages and CD4⁺CD8⁺ (double positive (DP)) macrophages. DP cells were further purified using a cell sorter (FACS Vantage; BD Biosciences). Splenic DCs were purified from mononuclear cells using the rat DC separation kit. A, Samples were cultured in chamber slides for 2 h at 37°C. Adherent cells were stained with H&E after removing nonadherent cells and then photographs were taken. Representative results are shown. Total magnification, ×1000. B, The expression profiles of TLRs in DP macrophages were compared with those in splenic DCs and CD8⁻ macrophages. RT-PCR was conducted in triplicate. Representative data are shown.

CD4⁺CD8⁺ macrophages kill tumor cell lines selectively

In our previous study, we demonstrated that CD4⁺CD8⁺ macrophages were able to kill the malignant thymoma cell line MT in vitro (1). Killing was dose dependent and reached the peak when the E:T ratio was 30. Because the effector/target combination was allogenic (CD4⁺CD8⁺ macrophages were derived from Wistar rats and MT from F344 rats), killing was induced by a MHC-unrestricted mechanism. After coculture with CD4⁺CD8⁺ macrophages for 18 h, almost all MT cells were destroyed, whereas control CD8⁻ macrophages barely killed the target cells (Fig. 3⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Tumor cell lines display differential sensitivities to CD4⁺CD8⁺ macrophages. A, Splenic CD4⁺CD8⁺ (DP) macrophages and CD8⁻ macrophages were added to the culture of MT cells with an E:T ratio of 30 (1.2 × 10⁶ effector cells and 4 × 10⁴ target cells/well in 24-well plates). Experiments were done in triplicate. Supernatant was removed after incubation for 18 h and then photographs were taken. Representative results are shown. Spindle-shaped tumor cells were alive in the upper panel. Arrows in the lower panel indicate apoptotic cells. Total magnification, ×200. B, NKR-P1A⁺ cells were purified from the splenic nonadherent mononuclear cells using the MACS system and served as NK cells. Splenic NK cells, CD8⁻ macrophages, and DP macrophages were added to the culture of the MT, SST-2, RCN-9, or NBT-II cell line with an E:T ratio of 30 (2.4 × 10⁵ effector cells and 8 × 10³ target cells/well in 96-well plates). After incubation for 18 h, cytotoxicity was measured using a CytoTox 96 test kit. Data are represented as mean ± SD of experiments done in triplicate. ∗, p < 0.05.

To examine the killing specificities of CD4⁺CD8⁺ macrophages, four tumor cell lines were used as targets (Fig. 3⇑B). CD4⁺CD8⁺ macrophages killed MT and NBT-II cells efficiently (percent specific lysis was 73.8 ± 6.8 and 77.4 ± 20.4, respectively) and SST-2 cells to some extent (percent specific lysis was 11.8 ± 1.8). However, RCN-9 cells were resistant to killing (percent specific lysis was −1.8 ± 0.5). These findings indicate that not all tumor cell lines are sensitive to lysis by CD4⁺CD8⁺ macrophages and that killing is critically dependent on putative factors expressed by target cells.

By contrast, the four cell lines were susceptible to killing by purified NK cells although the extent of killing showed considerable variation (percent specific lysis was 31.7 ± 4.5 for MT, 39.9 ± 14.3 for SST-2, 16.1 ± 2.7 for RCN-9, and 68.1 ± 4.3 for NBT-II). Interestingly, MT cells were more susceptible to killing by CD4⁺CD8⁺ macrophages than by NK cells, whereas the opposite was the case for SST-2 cells. These observations ruled out the possibility that the cytotoxic activities of CD4⁺CD8⁺ macrophages were mediated by contaminating NK cells.

Cell-cell contact is required for killing by CD4⁺CD8⁺ macrophages

To examine whether cell-cell contact is required for killing by CD4⁺CD8⁺ macrophages, adhesion of the effector cells to target cells was blocked using a cell culture insert. Cytotoxicity against NBT-II was completely inhibited by blocking cell-cell contact (percent specific lysis was decreased from 62.5 ± 8.6 to −2.4 ± 15.2; Fig. 4⇓A). This result indicates that adhesion to the target cells is essential for the killing process by CD4⁺CD8⁺ macrophages.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Cell-cell contact is required for killing by CD4⁺CD8⁺ macrophages. A, Splenic CD4⁺CD8⁺ macrophages were added to NBT-II cells with an E:T ratio of 30 (1.2 × 10⁶ effector cells and 4 × 10⁴ target cells/well in 24-well plates). For blocking the cell-cell contact, a cell culture insert (pore size, 0.4 μm; BD Biosciences was used. After incubation for 18 h, cytotoxicity was measured. Data are represented as mean ± SD of experiments done in triplicate. B, Tumor cells were incubated in 8-well chamber slides to form a monolayer at 37°C. Then, splenic CD4⁺CD8⁺ macrophages isolated from EGFP-transgenic rats were added to the culture. After incubation for 2 h, the chamber slides were gently washed with PBS and then GFP⁺ cells that adhered to the tumor cell sheet were observed using a fluorescence microscope. Total magnification, ×200. C, GFP⁺ cells that adhered to the tumor cell sheet in three high-power fields (×400) were counted. Data are represented as mean ± SD. ∗, p < 0.05.

Next, to examine whether differential sensitivities to killing by CD4⁺CD8⁺ macrophages are dependent on the adhesive efficiency to the target cells, effector cells derived from EGFP-transgenic rats were cultured on the monolayer of tumor cells. After incubation for 2 h, a large number of GFP⁺ cells (219 ± 24.9/field) adhered to the monolayer of MT cells (Fig. 4⇑, B and C). In contrast, adhesion to NBT-II was significantly less extensive (83.3 ± 9.0/field) than that to MT cells. Because MT and NBT-II cells were killed by CD4⁺CD8⁺ macrophages equivalently (see Fig. 3⇑B), differences in adhesive efficiency alone cannot account for differential sensitivities of target cells to CD4⁺CD8⁺ macrophages. Interestingly, CD4⁺CD8⁺ macrophages also adhered to RCN-9 (56.3 ± 7.0/field) with efficiency comparable to that of NBT-II. However, RCN-9 was resistant to killing by CD4⁺CD8⁺ macrophages (see Fig. 3⇑B). Thus, adhesion to the targets is essential, but not sufficient for killing by CD4⁺CD8⁺ macrophages.

Interaction of NKR-P2 and RAET1 is critically involved in the killing mechanism of CD4⁺CD8⁺ macrophages

To identify the molecules involved in the killing by CD4⁺CD8⁺ macrophages, the expression profiles of genes for Fas, RAET1, CD48, CD112, and CD155, known to be involved in antitumor immunosurveillance (4, 9, 10, 11), were compared among the four tumor cell lines (Fig. 5⇓A). RCN-9 cells, which were resistant to lysis by CD4⁺CD8⁺ macrophages, expressed Fas and RAET1 at significantly lower levels than the other cells. Expression of CD48 was not detected in any of the samples. CD112 and CD155 were expressed at similar levels in the four cell lines. Based on these observations, we assumed that Fas and/or RAET1 could be involved in the killing mechanism of CD4⁺CD8⁺ macrophages. To examine the involvement of the Fas/Fas ligand system, we pretreated effector cells with the anti-Fas ligand Ab. This pretreatment did not alter cytotoxic activities of CD4⁺CD8⁺ macrophages, arguing against the involvement of the Fas-Fas ligand interaction (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Interaction of NKR-P2 and RAET1 is critically involved in the killing mechanism of CD4⁺CD8⁺ macrophages. A, The expression profiles of genes for Fas, RAET1, CD48, CD112, and CD155, known to be involved in antitumor immunosurveillance, were compared among the four tumor cell lines by RT-PCR. B, The RAET1 expression plasmid was transfected into COS-7 cells, the transfectants were stained with rabbit anti-FLAG Ab followed by FITC-conjugated goat anti-rabbit IgG Ab, and then analyzed on FCM. Gray and black unfilled histograms represent the expression of FLAG-tagged RAET1 when transfection was performed with 1 and 2 μg of DNA, respectively. The gray-filled histogram represents the negative control transfected with the empty vector DNA. The numbers in the panel represent the percentage of FLAG⁺ cells. C, Splenic CD4⁺CD8⁺ macrophages were added to the culture of COS-7 cells transfected with the empty vector DNA or 1 or 2 μg of RAET1 cDNA with an E:T ratio of 30 (2.4 × 10⁵ effector cells and 8 × 10³ target cells/well in 96-well plates). After incubation for 6 h, cytotoxicity was measured. Data are represented as mean ± SD of experiments done in triplicate. ∗, p < 0.05. D, COS-7 cells transfected with 2 μg of DNA of the rat RAET1 expression plasmid (unfilled histogram) or with the empty vector (gray-filled histogram) were incubated with soluble rat NKR-P2-hIg fusion protein followed by staining with FITC-conjugated anti-human IgG Ab and then analyzed on FCM. The numbers in the panel represent the percentage of hIg⁺ cells. E, To block the interaction of NKR-P2 and RAET1, soluble NKR-P2-hIg (sNKR-P2) was added to the coculture of the effector and target cells. For control experiments, the same volume of PBS was added. After incubation for 6 h, cytotoxicity was measured using a CytoTox 96 test kit. Data are represented as mean ± SD of experiments done in triplicate. ∗, p < 0.05.

To examine whether RAET1 is involved in the killing process, we isolated RAET1 cDNA from NBT-II cells. This cDNA had a sequence completely identical to the predicted RAET1 sequence deposited in the GenBank database under the accession number XM_001054610. When 1 μg of the RAET1 expression plasmid was used for transfection, RAET1 was expressed on the surface of 30.2% of COS-7 cells, as assessed by staining with the anti-FLAG Ab (Fig. 5⇑B). COS-7 cells expressing REAT1 became susceptible to lysis by CD4⁺CD8⁺ macrophages (percent specific lysis was 19.0 ± 6.9 as compared with 3.0 ± 2.7 when COS-7 cells were transfected with an empty vector DNA; Fig. 5⇑C). When 2 μg of the RAET1 expression plasmid was used for transfection, the proportion of RAET1-positive COS-7 cells increased to 41.4%. Correspondingly, COS-7 cells became more susceptible to lysis by CD4⁺CD8⁺ macrophages (percent specific lysis was 24.2 ± 9.4; Fig. 5⇑C). These results indicate that RAET1 is involved in the killing mechanism of CD4⁺CD8⁺ macrophages.

We confirmed that the rat RAET1 molecule we identified has the ability to bind to soluble NKR-P2-hIg fusion proteins and thus functions as a ligand for NKG2D (Fig. 5⇑D). We then examined whether soluble NKR-P2-hIg can inhibit the cytotoxicity mediated by CD4⁺CD8⁺ macrophages. Addition of soluble NKR-P2-hIg significantly inhibited the cytotoxicity of CD4⁺CD8⁺ macrophages against the RAET1 transfectants (percent specific lysis was decreased from 17.8 ± 4.4 to 2.7 ± 6.5; Fig. 5⇑E). Because CD4⁺CD8⁺ macrophages express NKR-P2 (rat ortholog of human NKG2D) (1), collective evidence indicates that CD4⁺CD8⁺ monocytes/macrophages kill target cells via NKG2D ligand recognition.

NKR-P1A is involved in the killing mechanism of CD4⁺CD8⁺ macrophages

CD4⁺CD8⁺ monocytes in rat peripheral blood express the activating NK receptor NKR-P1A (1). We examined whether rat splenic CD4⁺CD8⁺ macrophages also express NKR-P1A (Fig. 6⇓, A and B). By RT-PCR and FCM, CD4⁺CD8⁺ macrophages were found to express NKR-P1A at higher levels than CD8⁻ macrophages. Furthermore, cross-linking of NKR-P1A with a specific mAb significantly augmented directed killing of P-815 cells by CD4⁺CD8⁺ macrophages (percent specific lysis were 26.7 ± 4.4 for Ab specific for NKR-P1A and 20.4 ± 3.8 for isotype-matched Ab; Fig. 6⇓C). These results indicate that NKR-P1A is partially involved in the killing mechanism of CD4⁺CD8⁺ macrophages.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

NKR-P1A is involved in the killing mechanism of CD4⁺CD8⁺ macrophages. A, Expression of NKR-P1A mRNA was examined by RT-PCR in splenic CD4⁺CD8⁺ (DP) macrophages and CD8⁻ macrophages. B, Expression of NKR-P1A molecules was analyzed by FCM. Unfilled and gray-filled histograms represent the expression of NKR-P1A on the DP and CD8⁻ macrophages, respectively. Experiments were repeated three times. Representative data are shown. C, For cross-linking experiments, total splenic macrophages were preincubated with the anti-NKR-P1A Ab followed by cocultivation with P-815 cells. For control experiments, macrophages were preincubated with the isotype-matched Ab. After washing with PBS, CD4⁺CD8⁺ macrophages were added to the culture of P-815 cells with an E:T ratio of 30 (2.4 × 10⁵ effector cells and 8 × 10³ target cells/well in 96-well plates). After incubation for 18 h, cytotoxicity was measured using a CytoTox 96 test kit. Data are represented as mean ± SD of experiments done in triplicate. ∗, p < 0.05.

CD4⁺CD8⁺ macrophages kill tumor cells via the granzyme/perforin-dependent pathway

CD4⁺CD8⁺ monocytes/macrophages characteristically express cytotoxic molecules such as granzyme B, perforin, and Fas ligand, which are normally seen in NK cells and CTLs (1). Synthesized granzyme B is stored within cytolytic granules along with perforin and is released to the targets upon the influx of Ca²⁺ into the effector cells (12, 13). We observed that granzyme B was granularly distributed in the cytoplasm of CD4⁺CD8⁺ macrophages (Fig. 7⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

CD4⁺CD8⁺ macrophages kill tumor cells via the granzyme/perforin-dependent pathway. A, Splenic CD4⁺CD8⁺ macrophages were stained for CD68 (ED-1, green) and granzyme B (N-19, red) (total magnification, ×4000). B, Splenic CD4⁺CD8⁺ macrophages were added to the culture of NBT-II cells with an E:T ratio of 30 (2.4 × 10⁵ effector cells and 8 × 10³ target cells/well in 96-well plates). After incubation for 18 h in the presence of 50 μM DCI or 5 mM EGTA, cytotoxicity was measured. Data are represented as mean ± SD of experiments done in triplicate. ∗, p < 0.05.

To examine whether granzyme B and perforin are involved in the killing mechanism of CD4⁺CD8⁺ macrophages, we inhibited granzyme activities and degranulation of the cytolytic granules using 3,4-dichloroisocoumarin (DCI; a broad serine protease inhibitor) and EGTA (a chelator of Ca²⁺), respectively (14). Both DCI and EGTA significantly inhibited cytotoxic activities of CD4⁺CD8⁺ macrophages against the NBT-II cell line (percent specific lysis was decreased from 89.8 ± 9.8 to 25.6 ± 8.8 with DCI treatment and to −1.1 ± 12.9 with EGTA treatment; Fig. 7⇑B). These results indicate that CD4⁺CD8⁺ macrophages kill tumor cells via the granzyme/perforin-dependent pathway.

CD4⁺CD8⁺ monocytes/macrophages are involved in antitumor immunity in vivo

To investigate how CD4⁺CD8⁺ monocytes/macrophages change in the number, phenotype, and function in tumor-bearing hosts, we inoculated NBT-II cells into the subcutis of rats with or without preimmunization with adjuvants containing killed tuberculosis germs. On day 4 after inoculation, the proportion of CD4⁺CD8⁺ monocytes in PBMC was significantly decreased in tumor-bearing rats that had received no immunization (Fig. 8⇓A). On day 6, CD4⁺CD8⁺ macrophages (detected as CD8⁺CD68⁺ cells) were identified in degenerative cystic spaces in the tumor tissues (Fig. 8⇓B). These findings suggest that CD4⁺CD8⁺ macrophages in the tumor tissues were recruited from CD4⁺CD8⁺ monocytes in the blood.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

CD4⁺CD8⁺ monocytes/macrophages are involved in antitumor immunity in vivo. NBT-II tumor cells were inoculated into the subcutis of syngenic rats with or without preimmunization with adjuvants containing killed tuberculosis germs. A, Proportion of CD4⁺CD8⁺ monocytes in PBMC obtained from nonimmunized rats. FCM was done using blood samples obtained on day 0 (preinoculation) and day 4 postinoculation. ∗∗, p < 0.01. B and C, Rats were killed on day 6 after tumor inoculation for histological examination. Representative photographs are shown (original magnification, ×400). D, Chronological changes in tumor size. ▴ and ▪, Preimmunized and nontreated rats, respectively. ∗∗, p < 0.01.

Next, we examined the expression of granzyme B in macrophages infiltrating into degenerative cystic spaces in the tumor tissues. Granzyme B was hardly detectable in CD68⁺ cells in nonimmunized rats, whereas CD68⁺ cells clearly expressed granzyme B in preimmunized rats (Fig. 8⇑C). These findings are consistent with our observation that granzyme B mRNA was undetectable in CD4⁺CD8⁺ monocytes obtained from nonimmunized rats (data not shown). Furthermore, the growth of inoculated tumor cells was significantly inhibited in preimmunized rats compared with rats without preimmunization (Fig. 8⇑D). Because preimmunization with adjuvants significantly increased the number of CD4⁺CD8⁺ monocytes/macrophages (1) and elevated the expression of granzyme B in these cells (Fig. 8⇑C), combined evidence suggests that CD4⁺CD8⁺ monocytes/macrophages are involved in antitumor immunity in vivo.