Tumor-induced immune suppression involves the accumulation of suppressive infiltrates in the tumor microenvironment such as regulatory T-cells (Tregs). Previous studies demonstrated that NK-dependant increases in CCL22 secretion selectively recruit Tregs toward murine lungs bearing Lewis Lung Carcinoma (LLC). To extend the in vitro studies, the present studies utilized in vivo depletion of NK cells to ascertain the contribution of NK-derived CCL22 toward total CCL22 and subsequent Treg levels in both normal and LLC-bearing lungs. However, NK depletion had the unexpected effect of increasing both CCL22 secretion and Treg levels in the lungs of NK-depleted LLC-bearing mice. This was concurrent with an increase in tumor burden. Flow cytometry and a series of both immunomagnetic and FACS isolations were used to identify the CCL22-producing cellular fractions in LLC-bearing lungs. A novel CD11b+CD11c+ cell population was identified that accumulates in large numbers in NK-depleted LLC-bearing lung tissue. These CD11b+CD11c+ cells secreted large amounts of CCL22 that may overcompensate for the loss of NK-derived CCL22 in the lungs of NK-depleted LLC-bearing mice. Taken together, these data suggest that NK cells play both a positive and negative role in the regulation of CCL22 secretion and, in turn, the recruitment of Tregs toward LLC-bearing lungs.
Tumor-induced immune suppression remains the largest obstacle toward effective immunotherapies against cancer.1, 2 Although solid tumors employ numerous mechanisms to evade immune responses, most involve the accumulation of suppressive infiltrates in the tumor microenvironment. One of the most potent and well studied of these suppressive infiltrates is the regulatory T-cell (Treg).3 High Treg numbers have been correlated with the expression of CCR4-associated chemokines within tumor tissue,4 and one of these chemokines, CCL22, has been shown to actively recruit Tregs toward human ovarian carcinoma and is indicative of a negative prognosis.5
Lewis Lung Carcinoma (LLC) is a murine squamous cell carcinoma derived from lung epithelial cells. Upon tumor formation, LLC elicits potent immune suppression,6 largely related to immune suppressive cytokines,7 suppressive myeloid infiltrates8–10 and Tregs.11 Particular attention should be paid to Tregs as these cells pre-exist at basal concentration throughout the body12 and thus, likely play a role in the induction of the suppressive microenvironment. Recent in vitro studies were carried out investigating the CCR4/chemokine axis in selective Treg recruitment toward LLC-bearing lung tissue. These studies investigated the CCR4-associated chemokines CCL22, and CCL17. Although low levels of CCL17 were secreted from both normal and LLC-bearing lung tissue, CCL22 secretion was highly upregulated in LLC-bearing lung tissue. These elevated levels of CCL22 were found to selectively recruit Tregs toward LLC-bearing lungs.13 Interestingly, CCL22 was not secreted by the LLC cells themselves. Isolation of cellular fractions and ex vivo depletion studies revealed that a large percentage of the increased CCL22 secretion from LLC-bearing lung tissue originates from NK cells, and that NK cells taken from the lungs of LLC-bearing mice secrete more CCL22 than those taken from normal controls.13 Despite the strong evidence suggesting that NK-derived CCL22 plays a large role in selective Treg recruitment toward LLC, in vivo depletion studies have yet to be done to confirm this mechanism.
The current study investigates the role of NK cells on CCL22 secretion and subsequent Treg levels by in vivo depletion of NK cells during LLC tumor growth. Because increased CCL22 secretion from LLC-bearing lungs is NK-dependant, it was intuitive to hypothesize that CCL22 secretion would be positively regulated by NK cells within the tumor-bearing tissue. However, the present study reveals that NK-mediated regulation of CCL22 secretion is more dynamic in nature and that compensatory mechanisms exist for CCL22 secretion and Treg recruitment in the absence of NK cells.
Material and Methods
A metastatic cell line of LLC (LN7) was used in all experiments. LLC and normal mouse lung epithelial cells (MLE-12) were kept in culture medium (RPMI-1640 media, Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 μg/ml streptomycin, 0.02 M Hepes buffer, and 5 × 10−5 M β-mercaptoethanol). Cells were passed as needed just prior to confluence. The LN7 variant of LLC cells are nonadherent in culture and, thus, do not require trypsinization prior to passage.
Generation of LLC-bearing lung tissues
Single cell suspensions of 1.0 × 107 cells/ml washed in Hank's Buffered Saline Solution (HBSS, Invitrogen) were prepared from LLC cell cultures that were in log phase. Female C57BL/6 mice were injected i.v. into the tail vein with 0.2 ml of phosphate buffered saline or LLC cell suspension resulting in the delivery of 2.0 × 106 LLC cells per animal. In previous studies, mice were euthanized ∼16–18 days later when signs of labored breathing first appeared in the LLC-injected animals.13 However, with NK-depleted animals, signs of labored breathing generally appeared around Day 12. Hence, Day 12 post-tumor inoculation became the end point for all groups. Lung tissue was harvested and placed in HBSS prior to subsequent assays, or was frozen into blocks in Optimum Cutting Temperature (OCT; Sakura, Torrance, CA) compound using liquid N2.
In vivo depletion of NK cells
Mice were injected with 0.2 mg of anti-asialo GM1 antibody (Wako Pure Chemicals, Richmond, VA) i.v. into the tail vein concurrently with a PBS or LLC injection. Control animals received injections of 0.2 mg normal rabbit serum. Multiple injections resulted in anaphylaxis; hence, singular injections of anti-asialo GM1 were used. The effectiveness of NK depletion was tracked using flow cytometric analysis of NK1.1+CD49b+ cell levels in dissociates from the lungs of both normal and LLC-bearing mice.
Generation of conditioned media
For the generation of conditioned media from cell cultures, the cells were washed with HBSS 3 times before reconstitution in serum-free culture media. Cultures were left at 37°C overnight (16 hr). The following morning, the media was collected and filtered. For the generation of conditioned media from intact tissue, whole intact lungs were rinsed in PBS and placed overnight (16 hr) at 37°C in 12 ml of serum-free culture media. The following morning, the media was collected and filtered.
Enzyme-linked immunosorbent assays (ELISA)
CCL22 was measured in various conditioned media using Quantikine Immunoassay kits (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions.
Dissociation of solid lung tissues
Following harvest, entire sets of both normal and LLC-bearing lungs were placed in an enzyme digest (1 μg/ml collagenase IV, 0.6 μg/ml hyaluronidase, 0.1 μg/ml DNAse; Sigma, St. Louis, MO) dissolved in HBSS for 90 min at room temperature. Each specimen was then homogenized in a Stomacher™ 80 homogenizer (Seward, Bohemia, NJ) set on medium for 120 sec. The resulting homogenate was then rinsed with HBSS and filtered to remove debris, resulting in single cell suspensions.
Preparation of splenocytes and bone marrow
Whole spleens were harvested from normal C57BL/6 mice and homogenized via a Stomacher 80 homogenizer (Seward) set on medium for 60 sec. Red blood cells were lysed using ACK lysis buffer. Bone marrow aspirate was obtained from the femurs of mice and passed through a 40-μm cell strainer (BD Falcon, San Jose, CA). Red blood cells were lysed using ACK lysis buffer.
Lung weight analysis
Following harvest, each set of lungs was washed with PBS and gently shaken to remove excess PBS drops. Each lung was then weighed. The amount of time utilized for weight analysis was minimized to maintain cell viability in subsequent assays.
Nonspecific staining was blocked with FBS and 1 μl/1.0 × 106 cells of CD16/32 (BD PharmMingen, San Jose, CA) before staining with 5 μl of the following antibodies: PE-NK1.1, PerCP-CD4, (BD PharmMingen); PE-CD25, FITC-FoxP3, and FITC-CD49a (DX5), FITC-F4/80, PE-Ly-6G/Gr-1, FITC-CD31, APC-CD11b, and PE-CD11c (eBioscience, San Diego, CA). Intracellular staining for FoxP3 required overnight fixation with fixation/permeabilization solution (eBioscience). Asialo GM1 staining was achieved using anti-asialo GM1 rabbit antiserum (Wako Pure Chemicals), followed by a secondary incubation with goat FITC-F(ab)2 anti-rabbit IgG (Jackson Immunoresearch). The extent and frequency of labeled cells was visualized using flow cytometry (FacsCanto; BD Biosciences, San Jose, CA).
Hematoxylin, eosin and erythrosine staining of frozen tissue
Lung tissue was frozen in OCT, cryosectioned into 10-μm sections, and placed onto glass slides. Sections were then fixed in 10% Neutral Buffered Formalin (Fischer Scientific, Pittsburgh, PA) for 5 sec, and rinsed 5 times with tap water. The sections were then stained with hematoxylin for ∼5 min, and rinsed again in tap water 5 times. The sections were then decolorized by dipping each slide 5 times in 0.25% acid alcohol (1.25 ml con. HCl in 500 ml 70% ethanol). The slides were then rinsed in tap water 6 times and dipped in 0.17 M lithium carbonate (Sigma) 10 times. Each slide was then rinsed in distilled water 5 times followed by 10 rinses in 95% ethanol (Harleco, Gibbstown, NJ). The slides were then placed in a mixture of eosin and erythrosine [0.1% Eosin Y, 0.2% Erythrosine B and 5% glacial acetic acid (Sigma) in 95% ethanol] for 30 sec. The slides were then dehydrated by rinsing 10 times in 95% ethanol followed by 10 rinses in 100% ethanol. Finally, slides were placed in xylene for 30 sec twice. Once dry, slides were mounted with Aquamount (Harleco) and visualized using light microscopy.
Immunofluorescent staining of frozen tissue sections
Lung tissue was frozen in OCT, cryosectioned into 10-μm sections and placed onto glass slides. Tissue sections were fixed in ice cold methanol for 10 min, and nonspecific staining was then blocked using 2% donkey serum in PBS for 1 hr at room temperature. After being rinsed in PBS, 1:100 dilutions in PBS of primary antibodies (CD11b, eBioscience; CD11c, BD Pharmingen) were added to the slides and allowed to incubate overnight at room temperature in a moist environment. The slides were then washed in PBS and donkey FITC-F(ab)2 anti-rabbit and donkey Rhodamine-F(ab)2 anti-rat secondary antibodies (Jackson labs, Bar Harbor, ME) were added at a dilution of 1:100 each in PBS and allowed to incubate for 45 min in room temperature in the dark. Following secondary incubations, slides were then rinsed in PBS, counterstained with Hoechst dye (Sigma), and then washed with distilled H2O and mounted using Vectashield fluorescent mount media (Vector Laboratories, Burlingame, CA). Fluorescent staining was visualized microscopically.
CD11b and CD11c fractions were isolated from lung tissue homogenates using isolation kits (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's instructions. CD3 and CD31 fractions were isolated from lung tissue homogenates by incubating with biotin-conjugated primary antibodies for 15 min on ice (biotin-CD3, biotin-CD31 eBioscience), followed by a secondary incubation with streptavadin microbeads (Miltenyi Biotec) for 15 min on ice. The labeled cells were then positively collected using LS magnetic columns (Miltenyi Biotec).
Isolation using FACS
CD11b+ subpopulations were first immunoflourescently labeled as described above for flow cytometry and then washed with PBS containing 1 μg/ml propidium iodide (Molecular Probes, Eugene, OR). Sorting was performed using a Dako MoFlo High-Speed Cell Sorter (Becton Dickinson, Franklin Lakes, NJ).
Data were reported using the mean as a measure of central tendency ± standard error of the mean. To compare one variable condition between groups, the 2-tailed Student's t-test was used. Significance was reported in the 95% confidence interval.
Injection with anti-asialo GM1 effectively depletes NK cells during LLC growth
CCL22 selectively recruits Tregs to the local tumor microenvironment.5 However, in the murine LLC model, CCL22 is not secreted by cancer cells or normal epithelial cells, but rather in part by resident NK cells in the tumor-bearing tissue as suggested by ex vivo studies.13 To investigate the role that NK cells have in LLC-associated CCL22 secretion, the present studies utilized anti-asialo GM1 antibody to deplete NK cells in vivo and then measured CCL22 and Treg levels.
Prior studies have utilized multiple injections of anti-asialo GM1 to deplete NK cells in various experimental models.14 However, we observed anaphylaxis in animals that received multiple anti-asialo GM1 injections as well as in animals that received multiple control rabbit serum injections. Therefore, only one injection in each mouse was used to avoid anaphylaxis and to ensure proper experimental conditions.
The use of anti-asialo GM1 as a means to deplete NK cells in the lungs of normal and LLC-bearing mice was validated by demonstrating that asialo GM1 expression in lung dissociates is restricted to NK cells. Both normal and LLC-bearing lung dissociates were stained with anti-asialo GM1 antibody followed by a FITC-fluorescent secondary antibody. Flow cytometric analysis revealed that in normal and LLC-bearing lung tissue, asialo GM1 is largely restricted to NK1.1+ cells (Fig. 1a). Importantly, treatment with anti-asialo GM1 may also deplete NKT cells that can also express NK1.1, or CD49b. However, in normal or LLC-bearing lung tissues, no significant population of cells exist that expresses both NK and T-cell markers,13 suggesting that NKT cells do not play a significant role within this tumor model system.
LLC-bearing animals injected with anti-asialo-GM1 displayed signs of labored breathing as soon as Day 11, and averaging 12 days after tumor injection. This was considered the endpoint for all test groups and lung tissue was then harvested. Flow cytometric analysis of fluorescently stained dissociates demonstrates that NK cell numbers remain depleted after Day 12 post-LLC and anti-asialo GM1 injection (Fig. 1b).
Depletion of NK cells results in increased LLC tumor burden
Upon resection of the various lung tissues, it became apparent that LLC tumor growth was greatly affected by NK depletion, while no visible effect occurred in normal lung tissues after NK cell depletion. LLC-bearing lung tissue typically forms numerous, but distinct, nodules concentrating around the outside periphery of the lung tissue, where the vascular bed is thickest. However, in NK-depleted animals, nodules were absent despite obviously heavy tumor burden (Fig. 2a). The NK-depleted LLC-bearing lung tissue was also firm to the touch and felt as though it were a solid mass.
The apparent increase in tumor burden in NK-depleted animals was confirmed by lung weight. Each excised set of lungs was weighed following a brief rinse in PBS. NK depletion had no effect on the weight of normal rabbit serum-injected control tissue. LLC-bearing lungs injected with rabbit serum were slightly heavier than normal and NK-depleted normal lungs (p = 0.002). However, NK-depleted LLC-bearing lungs were by far the heaviest, weighing over twice as much as rabbit serum-injected LLC-bearing lungs (p < 0.001) (Fig. 2b).
Histological analysis also revealed large increases in tumor burden associated with NK cell depletion. Rabbit serum-injected control LLC-bearing lung tissue displayed signs of invasion and moderately heavy tumor burden, yet rudimentary boundaries could be seen with each respective nodule. As such, many nonepithelial structures remained intact such as blood vessels and bronchioles. In NK-depleted LLC-bearing lung tissues, no distinct nodules could be discerned. Instead, heavy tumor burden permeated every structure suggesting heavy invasiveness. PBS and rabbit serum-injected normal control lungs displayed normal histology. NK depletion had no visible effect on lung histology in PBS-injected animals (Fig. 2c).
NK depletion paradoxically raises CCL22 levels in LLC-bearing mice
Because NK cells were identified as a major CCL22 secreting component of LLC-bearing lung tissue,13 it was hypothesized that NK depletion would result in a reduction of CCL22 secretion from LLC-bearing lung tissue. Interestingly, media conditioned by NK-depleted LLC-bearing lung tissue contained CCL22 levels that were twice as high as the levels in media conditioned with control LLC-bearing lung tissue (2,208 ± 286 ng/ml vs. 1,414 ± 229 ng/ml, p = 0.049) (Fig. 3). Media conditioned with control LLC-bearing lung tissue still contained higher CCL22 levels than media conditioned with the lungs of normal PBS and rabbit serum-injected control animals (1,414 ± 229 ng/ml vs. 715 ± 59 ng/ml, p = 0.01), consistent with previously observed increases in CCL22 secretion by LLC-bearing lungs.13 Additionally, NK depletion had no significant effect on CCL22 secretion from normal PBS-injected lung tissue.
NK-depletion results in highly elevated levels of Tregs in LLC-bearing lung tissues
The elevated CCL22 secretion seen from NK-depleted LLC-bearing lung tissue was unexpected. However, CCL22 was previously implicated in selective Treg recruitment toward LLC-bearing lung tissue13 and CCR4 expression and CCR4-associated chemokine levels have been repeatedly correlated with high levels of Tregs.4, 5, 15–18 Hence, the levels of Tregs in NK-depleted LLC-bearing lungs were measured. To measure Treg levels, NK-depleted and control rabbit serum-injected lung tissues from both LLC-bearing and PBS-injected control animals were dissociated and immunoflourescently stained for CD4, CD25 and FoxP3. Treg levels were quantified as a percentage of the total CD4 compartment (Fig. 4a). Treg levels in LLC-bearing lung tissues were significantly higher than in both NK-depleted and rabbit serum-injected normal PBS control tissues (13% ± 1% vs. 5.4% ± 0.6% and 4.9% ± 0.3% respectively, p < 0.001) (Fig. 4b). Treg levels in NK-depleted LLC-bearing lung tissues were significantly higher than in normal rabbit serum-inject LLC-bearing lung tissues (20% ± 2% vs. 13% ± 1%, p = 0.005) (Fig. 4b). For both rabbit serum-injected and NK-depleted LLC-bearing lungs, total CD4+ levels were similar or higher than that for normal rabbit serum-injected or NK-depleted normal PBS control lungs.
Increased CCL22 secretion remains in the myeloid fraction of NK-depleted LLC-bearing lungs
The paradoxical rise in CCL22 secretion from NK-depleted LLC-bearing lungs leads to the hypothesis that another cell population may compensate for CCL22 secretion in the absence of NK cells within the LLC microenvironment. In an attempt to narrow down possible CCL22-secreting cells, immunomagnetic isolations of CD11b+, CD11c+, CD3+ and CD31+ cells were performed on lung tissue dissociates from animals injected with PBS or LLC along with rabbit serum or anti-asialo GM1. Following the isolations, each fraction was plated at equal cell numbers and allowed to incubate overnight to generated conditioned media. CCL22 was not secreted in any significant amounts from CD31+ fractions isolated from the lungs of any of the groups of mice (Fig. 5). CD3+ cells isolated from PBS and rabbit serum-injected normal control lungs secreted moderate amounts of CCL22 (466 ± 43 ng/ml). Lower CCL22 secretion from isolated CD3+ cells taken out of both rabbit serum-injected and NK-depleted LLC-bearing mouse lungs, as well as PBS-injected NK-depleted mouse lungs indicated that endogenous T-cells do not compensate for the loss of NK-derived CCL22 in NK-depleted LLC-bearing lung tissue. CCL22 secretion from isolated CD11c+ fractions was moderately high from both rabbit serum-injected and NK-depleted PBS-injected control lungs (735 ± 38 ng/ml, and 776 ± 21 ng/ml, respectively). CCL22 secretion from isolated CD11c+ fractions was less moderate from both rabbit serum-injected and NK-depleted LLC-bearing lungs (429 ± 66 ng/ml and 465 ± 136 ng/ml, respectively).
Isolated CD11b+ fractions displayed the most interesting results (Fig. 5). CD11b+ cells taken from both rabbit serum-injected and NK-depleted PBS-injected control lungs secreted very low amounts of CCL22 (155 ± 51 ng/ml and 94 ± 30 ng/ml, respectively). However, CD11b+ cells taken from both rabbit serum-injected and NK-depleted LLC-bearing lungs secreted significantly higher levels of CCL22 (1,079 ± 137 ng/ml and 867 ± 61 ng/ml, respectively) than did CD11b+ cells from either NK-depleted or rabbit serum-injected control lungs (p < 0.001 across all 4 comparisons). Increased CCL22 secretion from CD11b+ cells taken from LLC-bearing lungs issue is consistent with previous observations.13 However, the fact that CD11b+ cells taken from NK-depleted LLC-bearing lung tissue also display an increase in CCL22 secretion suggests that any cells compensating for lost NK-derived CCL22 still reside in the myeloid compartment. Because the isolated fractions were all plated at equal cell concentrations, this experiment does not take into account the relative frequencies of each fraction found in the various lung tissue types toward total CCL22 secretion.
CD11b+CD11c+ cells represent a novel component to the myeloid compartment in NK-depleted LLC-bearing lungs
In previous studies, 2 and 3 major CD11b+ populations were identified in the CD11b+ compartments of normal and LLC-bearing lungs respectively.13 These populations were CD11b+NK1.1+ and CD11b+Gr-1+ cells in both normal and LLC-bearing lungs, along with the addition of CD11b+F4/80+ cells in LLC-bearing lungs. Within the myeloid fraction, these cellular markers were mutually exclusive from each other and accounted for the majority of the total CD11b+ cells in both normal and LLC-bearing lungs.13 However, these markers account for less than half of the total CD11b+ fraction in NK-depleted LLC-bearing lungs suggesting that additional subpopulations are present.
A very large subpopulation was discovered within the myeloid compartment of NK-depleted LLC-bearing lung tissue that expresses both CD11b and CD11c (Fig. 6). Amazingly, this population accounted for 45% of all CD11b+ cells in NK-depleted LLC-bearing lung tissues. Additionally, this population is unique to NK-depleted LLC-bearing lung tissue as no CD11b+CD11c+ cells were detected in the lungs of either rabbit serum-injected or NK-depleted PBS-injected control animals. Minimal numbers were detected in rabbit serum-injected LLC-bearing lungs.
The uniqueness of CD11b+CD11c+ cells to NK-depleted LLC-bearing lung tissue was confirmed microscopically. Frozen lung sections from animals injected with PBS or LLC along with rabbit serum or anti-asialo GM1 were prepared and immunoflourescently stained for CD11b and CD11c. CD11c+ cells were rare in lungs from PBS and rabbit serum-injected control animals, and were distinctly different from CD11b+ cells (Fig. 7a). The same was true for CD11c+ cells in the lungs of anti-asialo GM1 and PBS-injected animals (Fig. 7b), suggesting that cells that express both CD11b and CD11c are absent from these tissues. Conversely, CD11b+CD11c+ cells could be found in rabbit serum-injected LLC-bearing lung tissue, although at miniscule frequencies (Fig. 8a). There was also a large increase in CD11b+ cells that stained negatively for CD11c in the lungs of rabbit serum-injected LLC-bearing animals, reiterating previous observations of increased myeloid infiltration in LLC-bearing lung dissociates.13 Although CD11b+CD11c+ cells were rare in the lungs of rabbit serum-injected LLC-bearing animals, the lungs of anti-asialo GM1-injected LLC-bearing animals contained highly elevated levels of CD11b+CD11c+ cells (Fig. 8b). This suggests that high numbers of CD11b+CD11c+ cells may be unique to the lungs of anti-asialo GM1-injected LLC-bearing animals.
CD11b+CD11c+ cells secrete CCL22 in NK-depleted LLC-bearing lungs
We hypothesized that the paradoxical increase in CCL22 secretion seen in NK-depleted LLC-bearing lungs can be attributed to compensation by a novel cell population. Immunomagnetic isolations suggested that the majority of increased CCL22 secretion remained within the myeloid fraction in NK-depleted LLC-bearing lungs (Fig. 5). However, a unique subpopulation of CD11b+ cells that also express CD11c was identified in the lungs of NK-depleted LLC-bearing animals.
FACS was used to isolate each previously described CD11b+ subpopulation along with this novel CD11b+CD11c+ fraction from lung dissociates from animals injected with PBS or LLC along with rabbit serum or anti-asialo GM1. Each isolated population of cells was plated at equal cell concentrations and allowed to incubate overnight to create conditioned media from which CCL22 secretion could be measured (Fig. 9a). Similar to previous observations,13 isolated CD11b+NK1.1+ cells from rabbit serum-injected LLC-bearing lungs secreted elevated levels of CCL22 compared to rabbit serum-injected PBS-injected control lungs (141 ± 4 ng/ml vs. 53 ± 4 ng/ml respectively, p = 0.003). C11b+NK1.1+ cells were not present in anti-asialo GM1-injected PBS or LLC-bearing lungs. CD11b+Gr-1+ cells did not secrete detectable CCL22, nor did CD11b+F4/80+ cells isolated from rabbit serum-injected LLC-bearing lungs. Interestingly, no CD11b+F4/80+ cells were detected in anti-asialo GM1-injected LLC-bearing lungs, despite the fact that F4/80+ cells do not stain positively for asialo GM1. No CD11b+F4/80+ cells were found in rabbit serum or anti-asialo GM1-injected PBS-injected control lungs.
Despite the low numbers of CD11b+CD11c+ cells found in rabbit serum-injected LLC-bearing lung tissue, enough cells were isolated to allow for CCL22 measurement. The CD11b+CD11c+ cells isolated from both rabbit serum and anti-asialo GM1-injected LLC-bearing lungs secreted moderate amounts of CCL22 (54 ± 11 ng/ml and 67 ± 8 ng/ml, respectively). Each fraction was plated at equal cell concentrations in the experiment depicted in “part a” of Figure 9 and does not take the frequency of each population into account.
Before FACS isolation, a small aliquot was used to quantify the frequency of each subpopulation using flow cytometry. These frequencies were then used to adjust the measured amounts of CCL22 secretion in an attempt to better represent the CCL22 contribution of each fraction. The product of CCL22 and population frequency is depicted (Fig. 9b). Despite the fact that NK1.1+ cells isolated from rabbit serum-injected LLC-bearing lungs secrete more CCL22 than CD11b+CD11c+ cells on a per cell basis, the relative contribution of CD11b+CD11c+ cells in NK-depleted LLC-bearing lung tissue is much greater due to highly elevated CD11b+CD11c+ cell numbers. Taken together, these data suggest that elevated CD11b+CD11c+ cell numbers in NK-depleted LLC-bearing lungs compensate for CCL22 secretion in the absence of NK cells.
CD11b+CD11c+ cells do not accumulate in bone marrow or secondary lymph tissue in LLC-bearing NK-depleted mice
Although both CD11b and CD11c can be expressed on multiple cell types, the CD11b+CD11c+ cells accumulating in NK-depleted LLC-bearing lungs are likely myeloid dendritic cells (DC) distinct from lymphoid CD11b−CD11c+ DCs native to normal lung tissue. Unlike lymphoid DCs, myeloid DCs are of hematopoietic lineage originating from precursors in the bone marrow.19 This raises the question if the observed CD11b+CD11c+ cell accumulation seen in NK-depleted LLC-bearing lungs is unique to tumor-bearing tissue within the tumor-bearing animals. In an effort to ascertain possible systemic accumulation, bone marrow and splenocytes were harvested from animals injected with PBS or LLC along with rabbit serum or anti-asialo GM1. The cells were then immunofluorescently stained for CD11b and CD11c and analyzed using flow cytometry. So significant changes were detected for CD11b+CD11c+ cell numbers across any group in either bone marrow (Fig. 10a) or spleen (Fig. 10b). This suggests that NK-depletion does not induce the accumulation of myeloid DCs in hematopoietic or secondary lymphoid tissues.
Because NK cells secrete elevated levels of CCL22 and, consequently, recruit Tregs in the LLC microenvironment,13 we hypothesized that in vivo NK depletion would result in reduced levels of CCL22 secretion and Treg levels. Paradoxically, NK depletion resulted in a dramatic increase in CCL22 secretion from LLC-bearing lung tissue. This increase was not seen when NK cells were depleted from normal rabbit serum-injected controls, suggesting that the phenomenon is unique to tumor-bearing individuals. Although the increase in CCL22 secretion was unexpected, it became intuitive that increased levels of CCL22 would lead to increased numbers of Tregs. Indeed, NK-depleted LLC-bearing lung tissues contained highly elevated levels of Tregs that were approximately twice the levels in control rabbit serum-injected LLC-bearing lung tissue.
NK cells have a well established role in immune surveillance20–23 and are generally accepted as a beneficial cell population for anti-tumor immunity. NK depletion at the time of tumor inoculation resulted in drastic increases in LLC tumor burden, suggesting that the positive anti-tumor role that NK cells play outweighs any detrimental role in supporting tumor-induced immune suppression. The increases in tumor burden seen here reiterate the importance of NK cells to the process of immune surveillance, and are not surprising despite the implications of NK-mediated Treg recruitment.
Given the paradoxical rise in CCL22 levels in NK-depleted animals, we hypothesized that NK depletion may alter the profile or behavior of infiltrating populations, resulting in one or more novel sources of CCL22 in LLC-bearing lungs. Here, we identified a large accumulation of CD11b+CD11c+ cells in NK-depleted LLC-bearing lungs that are absent in the lungs from both PBS-injected control groups and nearly absent from rabbit serum-injected LLC-bearing lungs. This novel population of cells secreted moderate amounts of CCL22 which, when taken in the context of the high numbers of CD11b+CD11c+ cells, largely overcompensates for the lost NK-derived CCL22. There is, however, a disjunction between the observed CCL22 secretion from FACS-isolated CD11b+CD11c+ cells and CD11c+ cells isolated via immunomagnetic separation. CD11c+ cells isolated from both rabbit serum and anti-asialo GM1-injected LLC-bearing lungs secreted less CCL22 than did CD11c+ cells isolated from either rabbit serum or anti-asialo GM1-injected PBS control lungs. This is likely due to a lesser level of purity of isolated CD11c+ cells from the immunomagnetic column. Of all the broad spectrum makers isolated via immunomagnetic separation, CD11c cells were the least enriched.13 In addition, these cells would contain both CD11c+CD11b+ and CD11c+CD11b− cells. Finally, following immunomagnetic isolation, each cell fraction was plated at an equal concentration, meaning that the data collected in Figure 5 and “part a” of Figure 9 did not take cellular frequencies into account. Given the small prevalence of CD11c+ cells in both rabbit serum and anti-asialo GM1-injected PBS control lungs, and in rabbit serum-injected LLC-bearing mice, it is not likely that CD11c+-derived CCL22 is biologically significant. This is not true in NK-depleted LLC-bearing lung tissue, given the large cell numbers of CD11b+CD11c+ cell therein.
These results, combined with previous studies, suggest that NK-regulated CCL22 secretion is more complex than previously thought. Under normal circumstances, NK cells secrete minimal amounts of CCL22 in murine lungs. In the tumor microenvironment, this secretion is upregulated by unknown factors, possibly IL-4 or IL-13, which can induce CCL22 secretion from NK cells in vitro.24 At the same time, NK cells still play a positive role by inhibiting the accumulation of CD11b+CD11c+ cells which, at highly elevated cell numbers, contribute a larger pool of CCL22 than tumor-associated NK cells. Regardless, elevated levels of CCL22 lead to increased selective Treg recruitment and subsequent tumor-induced immune suppression (Fig. 11).
The nature by which NK cells inhibit the accumulation of CD11b+CD11c+ is unknown. Given the ability of NK cells to produce both inflammatory cytokines and chemokines, it is easy to speculate that their absence alters both recruitment and induction mechanisms. It is also possible that NK cells may not directly inhibit the observed accumulation of CD11b+CD11c+ cells, but rather that anti-asialo GM1 may indirectly inhibit their accumulation by affecting other populations systemically, most notably NKT cells. However, detectable levels of NKT cells were not previously observed in either normal or LLC-bearing lungs.13
A pressing question raised by these studies is the identity of the CD11b+CD11c+ cells accumulating in LLC-bearing lungs in the absence of NK cells. CD11b+CD11c+ cells have been described as an immature myeloid population in the normal continuum of myelopoietic differentiation.25 The idea that the CD11b+CD11c+ cells found in NK-depleted LLC-bearing lungs are immature myeloid cells is consistent with observed detrimental effects of tumor influence on myelopoietic maturation,26 and would explain the lack of more mature myeloid populations such as CD11b+F4/80+ cells. Yet, if NK depletion adversely affects hematopoiesis, it is reasonable to expect CD11b+CD11c+ cell accumulation in the bone marrow as well. Since this is not the case, the most likely explanation is that this accumulation affects more mature myeloid populations. One common mature cell type that expresses both CD11b and CD11c is the myeloid DC.19 These cells are distinct from lymphoid DCs in both origin and differentiation. As such they carry myeloid linage markers that are lacking in lymphoid DCs.19 Interestingly, lymphoid DCs and myeloid DCs retain unique chemokine receptors and thus respond to entirely different recruitment signals based on distinct precursor cells.27 By depleting NK cells, it is possible that the entire chemokine/cytokine milieu is shifted to one that favors the accumulation of myeloid DCs over resident lymphoid DCs. Further studies need to be conducted to verify the identity of these cells and their relationship to NK cells.
The authors thank Drs. Amanda LaRue and HaiQun Zeng for their technical expertise and assistance with cell sorting. They are also grateful to Dr. Jennifer Konopa-Mulligan, and Mr. Jarrett Walsh for their support and assistance.