Background
Infiltration of immune effector cells, such as T cells, natural killer (NK) cells, and macrophages, into tumors is known to suppress tumor cell growth [
1‐
3]. The number of tumor-infiltrating lymphocytes may serve as a valuable prognostic marker for immunotherapy in many cancers, including breast cancer, melanoma, non–small-cell lung cancer, and ovarian cancer [
4‐
7].
Chemotherapeutic agents used to treat cancer are generally believed to directly kill tumor cells [
8]; however, accumulating evidence suggests that chemotherapeutic agents facilitate infiltration of immune effector cells into tumors and thus sensitize tumor cells to immune cell attack [
9]. In fact, chemotherapy-induced immune response may serve as a predictor of therapeutic outcome in cancer patients [
10]. Supporting this view are several reports showing that some chemotherapeutic compounds exhibit a more promising antitumor effect in patients who have higher levels of tumor-infiltrating lymphocytes after treatment than in patients who have lower levels of these cells [
11]. In our previous study, we found that the combination of doxorubicin (Dox) and interleukin-12 (IL-12) had much greater antitumor efficacy than did either agent alone in 4T1 tumor–bearing mice [
12]. This increased antitumor efficacy was associated with a substantial increase in CD8
+T cell infiltration at tumor sites, but the mechanism for this CD8
+T cell accumulation at the tumor site is largely unknown.
Natural killer group 2, member D receptor (NKG2D)–dependent immune surveillance plays a key role in suppressing tumor progression [
13‐
15]. NKG2D is a lectin-like receptor protein that is expressed on NK, NKT, γδT, and some αβCD8
+ T cells [
16,
17]. NKG2D is detected on CD8
+T cells in both humans and mice, in mice on
active CD8
+T cells only [
18,
19]. As an activating receptor, NKG2D regulates innate and adaptive immune responses against infections and cancers [
20]. In melanoma patients, tumor-infiltrating NKG2D-positive T cells were shown to have promising antitumor efficacy [
21]. In the mouse tumor microenvironment, NKG2D-positive CD8
+T cells were critical in recognizing tumor cells for tumor immunosurveillance [
22]. We reasoned that a therapeutic strategy that increases the expression of NKG2D receptor on CD8
+T cells may contribute tumor infiltration. Treatment with IL-12 modestly enhanced NKG2D expression on NK cells
in vitro[
23], but the effects of IL-12 on NKG2D expression on NK cells and CD8
+T cells
in vivo are unknown.
Our purpose for this study was to determine whether Dox plus IL-12 induces NKG2D expression in T cells and whether accumulation of NKG2D-positive CD8
+T cells in tumors is dependent on NKG2D induction. Our central hypothesis was that Dox enhances IL-12–mediated NKG2D expression on CD8
+T cells and that this increased NKG2D expression facilitates the accumulation of CD8
+T cells in tumors and therefore enhances the antitumor efficacy of this combination [
12]. We have confirmed this hypothesis by using
in vivo and
in vitro approaches. This study for the first time reveals that Dox plus IL-12 increases expression of the NKG2D receptor in CD8
+T cells, thereby increasing accumulation of NKG2D-positive CD8
+T cells in tumors to promote antitumor immune surveillance.
Discussion
Our results show that treatment with Dox plus IL-12 increased the number of NKG2D-positive CD8
+T cells in tumor-bearing mice (Figures
1 and
2) and promoted the localization of NKG2D-positive CD8
+T cells in tumors (Figures
3,
4,
5). This combination treatment also increased NKG2D-dependent antitumor efficacy
in vivo (Figure
6).
Several known mechanisms account for IL-12–mediated antitumor efficacy, including induction of IFN-γ, promotion of type 1 helper T cell response, and stimulation of CD8
+T cell antitumor response [
27‐
29]. Likewise, Dox has been shown to act through several mechanisms, including disruption of DNA synthesis, initiation of DNA damage, and compromise of the cell membrane, to cause cell apoptosis [
30]. This study revealed that NKG2D induction on CD8
+T cells serves as an important mechanism for Dox-augmented IL-12–mediated tumor growth inhibition because the increased NKG2D expression on CD8
+T cells plays a role in tumor-specific localization (Figures
3 and
5).
The dependence of antitumor efficacy on NKG2D was not surprising because others have discovered that silencing NKG2D, or its associated molecules DAP10 or DAP12, reduces the cytotoxicity of activated CD8
+T cells and NK cells
in vitro[
31]. The surprising observation from this study is the CD8
+T cell–specific induction of NKG2D by treatment with Dox plus IL-12 (Figure
1A). This is surprising because NKG2D is constitutively expressed on NK cells, and others found that IL-12 recombinant protein induces modest NKG2D expression on NK cells
in vitro[
23]. In contrast, our
in vivo results showed that Dox plus IL-12 treatment expanded the NKG2D-positive CD8
+T cell population (Figure
1A) but failed to change the NKG2D-positive NK cell population (Figure
1A). This observation was validated in NK cell–depletion and T cell–depletion studies (Figure
2).
In vitro results from others found that IL-12 induces modest NKG2D expression in NK cells [
23], but our
in vivo results could not confirm this observation. Instead, we found that NKG2D is induced in CD8
+T cells. This observation may not be easily accepted but has been carefully validated in our study. The discrepancy between our observations and those of other authors is most likely due to the differences in model systems. Other investigators used an
in vitro system to observe the modest NKG2D increases in NK cells, whereas we used an
in vivo system; others used IL-12 alone, whereas we used Dox plus IL-12; finally, others used recombinant IL-12 protein, whereas we used an
IL-12 gene therapy approach, which yields only a very low level of IL-12 expression. In future study, we will explore the mechanism by which IL-12 plus Dox induces NKG2D
in vivo. Finding this mechanism will not be easy, based on our preliminary study, in which direct treatment of splenocytes with IL-12 plus Dox did not induce NKG2D.
It is still a puzzle how NKG2D-positive CD8
+T cells accumulate in tumors after IL-12 plus Dox treatment. Markiewicz
et al.[
32] indicated that NKG2D ligand Rae-1 expression recruits NKG2D-positive cytotoxic T cells independent of antigen recognition. NKG2D ligands are primarily present in tumor cells or infected cells but not in normal tissues [
17,
33,
34]. Recently, NKG2D ligand Rae-1 expression was also observed in tumor vasculature [
35]. These discoveries shed light on the possibility that NKG2D ligand expression in tumor cells or tumor vasculature attracts NKG2D-positive CD8
+T cell accumulation. Moreover, it is known that chemokines affect the migration of lymphocytes to tumors. An earlier study stated that the expression of CCR5 and CXCR3 has a positive correlation with the accumulation of CD8
+ and CD4
+ T lymphocytes in invasive colorectal tumors [
36]. Another study demonstrated that chemokine CCL2 and its receptor CCR2 are needed for human Vδ1T cell infiltration into various tumors including lung, prostate, liver, or breast cancers [
37]. Moreover, CXCL10 was found to enhance tumoral lymphocytic infiltrate in multiple cancers. Mulligan
et al.[
38] indicated that in breast cancers, CXCL10 expression showed significant association with tumor-infiltrating CD4
+ and CD8
+ lymphocytes. Also, induction of CXCL10 and CCL5 in colorectal tumors by hyperactivated NF-κB selectively promoted the accumulation of T effector lymphocytes but reduced the T regulatory lymphocytes [
39]. Therefore, IL-12 plus Dox treatment possibly reprograms chemokine expression in the tumor microenvironment, which boosts the NKG2D induction-associated recruitment of NKG2D-positive CD8+ T lymphocytes.
Other investigators found that a modest dose of Dox had the potential to boost immune response and potentiate the IL-2 effect against tumor cells [
40]. In fact, one report demonstrated that the Dox-mediated therapeutic effect against cancer requires CD8
+T cells and IFN-γ [
41]. Although the mechanism was unknown in both cases, we speculate that the immune response may be boosted by upregulating NKG2D through a combination of Dox plus IL-2 or Dox plus IFN-γ.
Conclusions
In summary, we have presented
in vivo evidence that Dox plus IL-12 induces CD8
+T cell–specific NKG2D induction, which facilitates the accumulation of NKG2D-expressing CD8
+T cells in tumor sites. Others have found that induction of NKG2D ligands boosts NKG2D-mediated tumor cell death [
17,
24,
26,
42,
43]. We expect that developing a strategy to simultaneously boost induction of the NKG2D ligand in tumors and NKG2D expression in immune cells, which will be the focus of our future effort, will greatly enhance the antitumor immune response and the treatment’s antitumor efficacy.
Finally, it is still not clear why NKG2D-positive NK cells fail to accumulate in tumors whereas NKG2D-positive CD8+T cells do accumulate. We speculate that independent engagement of another ligand and receptor between a tumor and CD8+T cells is required, a theory that we are currently investigating.
Materials and methods
Ethics statement
The mice used in this study were maintained under National Institutes of Health guidelines and according to procedures approved by the Institutional Animal Care and Use Committee of The University of Texas MD Anderson Cancer Center.
IL-12 gene construct and Dox
The mice were treated with a combination regimen of Dox plus an
IL-12 gene construct. The
IL-12 gene construct was obtained from Valentis, Inc. (Vilnius, Lithuania); the backbone of this construct was described in a previous publication [
44]. The control plasmid DNA consisted of the same construct with the
IL-12 gene deleted. Plasmid DNA was prepared by using the endotoxin-free Mega preparation kit from Qiagen, Inc. (Valencia, CA) according to the manufacturer’s instructions. Doxorubicin (Bedford Laboratories, Bedford, OH) was purchased from the pharmacy at the Louisiana State University School of Veterinary Medicine.
Tumor models and DNA delivery via intramuscular electroporation
Six- to eight-week-old female BALB/C mice weighing 18-20 g were obtained from the National Cancer Institute or Jackson Laboratory (Bar Harbor, ME) for this study. Murine invasive breast carcinoma 4T1 cells were maintained in Dulbecco modified essential medium containing 10% fetal bovine serum (Life Technologies, Grand Island, NY). Tumors were generated by subcutaneously inoculating BALB/C mice with 4T1 tumor cells (2×10
5) suspended in a 30-μL volume of phosphate-buffered saline solution (PBS). Tumor dimensions were measured with calipers every 3 days, and tumor volume was calculated with use of the following formula: V = (π/8)
a ×
b2, where V = tumor volume in cubic centimeters,
a = maximum tumor diameter, and
b = diameter at 90° to
a[
45].
Using protocols described previously, we injected
IL-12–encoding or control plasmid DNA into the two hindlimb tibialis muscles of each mouse via electroporation [
46]. This procedure yielded a blood IL-12 level greater than 100 pg/mL [
47]. The electroporation parameters for intramuscular injection, which had previously been identified as optimal, were set at 350 V/cm and 20 ms pulse duration for 2 pulses [
46].
Mice received one of four standard treatments: control plasmid DNA alone (control DNA), Dox plus control plasmid DNA (Dox plus control DNA), IL-12 plasmid DNA alone (IL-12), or Dox plus IL-12 plasmid DNA (Dox plus IL-12). The treatments were administered twice on days 7 and 17 after tumor cell inoculation. For each round of treatment, each mouse received 5 μg of DNA for each muscle, for a total of 10 μg of DNA. The dose of each Dox treatment was 5 mg/kg, and the Dox was administered intraperitoneally. Dox was administered at the same time as the plasmid DNA. Mice were euthanized via CO2 inhalation 4 days after the second treatment (day 21) and their tissues subjected to the analyses described in subsequent sections.
CD8
+
T/NK cell depletion in vivo
For immune cell–depletion experiments, CD8+T cell–depletion antibody (clone 2.43) or NK cell–depletion antibody (anti-Asialo GM1) was administered to deplete CD8+T cells or NK cells, respectively. Tumor-bearing BALB/C mice were inoculated intraperitoneally with one of the antibodies (50 μg of antibody in 50 μL PBS) on day 7 along with the first treatment. Injection of the cell-depleting antibody was repeated twice a week.
Flow cytometry analysis for detecting NKG2D-positive immune cells
Spleens from treated mice were homogenized gently in a 40-μm nylon strainer, and red blood cells were subjected to lysis with Puregene red blood cell lysis solution (Gentra Systems, Minneapolis, MN). Spleen cells (50,000 cells/sample) were stained with various antibodies to identify immune cell types: PE-Cyanine7 (PE-Cy7)-conjugated anti-mouse NKG2D (clone CX5) or isotype control antibody (eBioscience, San Diego, CA); fluorescein isothiocyanate (FITC)–conjugated anti-mouse CD4 (clone GK1.5) or CD8a antibody (clone 53-6.7) or the isotype control antibody (Pharmingen, San Diego, CA); PE-conjugated anti-mouse CD3ϵ (clone 145-2C11) and FITC-conjugated anti-mouse NKp46 antibody (clone 29A1.4) or its isotype control antibody (eBioscience, San Diego, CA). NKp46 was recently identified as a NK cell marker [
48]. Anti-NKG2D C7 antibody was generously provided by Dr. Wayne Yokoyama (Washington University School of Medicine). The stained cells were analyzed on an Attune acoustic focusing cytometer (Applied Biosystems, Inc., Carlsbad, CA). Data were analyzed by using Attune software (Applied Biosystems, Inc.) or FlowJo software (Ashland, OR).
RNA isolation and Northern blot analysis of gene expression
RNA was isolated from tumors with TRIzol reagent (Invitrogen, Carlsbad, CA) as described previously [
49]. The details of Northern blot analysis of gene expression were presented in a previous publication [
50]. Briefly, total RNA (4 μg) was subjected to denaturing by agarose gel electrophoresis, and ribosomal RNA was stained with ethidium bromide to ensure equal loading of all samples. The Northern blot was scanned with use of a Molecular Imager (Bio-Rad, Hercules, CA). The signal intensity was normalized to the level of the total ribosomal RNA.
Immunofluorescence staining analysis
Frozen tumor sections were fixed with cold acetone, acetone plus chloroform (1:1), and acetone. Tissue sections were blocked with blocking buffer (5% normal horse serum and 1% normal goat serum in PBS) and incubated with rat anti-mouse CD8α (clone YTS105.18, AbD Serotec, Raleigh, NC), rat anti-mouse CD4 (clone RM4-5, BD Pharmingen, San Jose, CA), or rat anti-mouse NKp46 antibody (clone 29A1.4, Biolegend, San Diego, CA) overnight at 4°C. The next day, tissue sections were blocked and incubated with goat anti-rat Alexa fluor 488 secondary antibody (Life Technologies, Grand Island, NY) for 1 hour at room temperature. Tissues were then blocked and incubated with second primary antibody NKG2D-biotin antibody (1:50; R&D Systems, Minneapolis, MN) overnight at 4°C and second secondary antibody streptavidin-conjugated Alexa fluor 594 (Life Technologies, Grand Island, NY) for 1 hour at room temperature. Rat IgG was used as the negative control. Nuclei were counterstained with Hoechst 33258 (1:10,000) (Life Technologies, Grand Island, NY). Tumor sections were mounted in antifade fluorescence mounting medium (Life Technologies, Grand Island, NY). Slides were visualized under the Nikon eclipse Ti fluorescence microscope (Nikon, Melville, NY) with use of appropriate filters (original magnification, 100×).
Statistical analysis
Tumor volume, lung metastases, and flow cytometry analyses were the primary outcomes measured. We used the 2-sided Student t-test to compare results between two treatment groups or one-way ANOVA to compare results from more than two treatment groups. GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA) was used to determine the P values, and P values < 0.05 were considered statistically significant.
Shulin Li, professor at the University of Texas Graduate School of Biomedical Science (UTGSBS) in Houston, Department of Pediatrics Research, Endowed Chair; Chair of Cellular Immune Response Committee for American Society of Gene and Cell Therapy. Dr. Li has co-authored articles published in the journals Science, Immunity, Journal of Experimental Medicine, Molecular Cell, Journal of the National Cancer Institute, and Nature Reviews.
Eugenie Kleinerman, professor at UTGSBS, Department of Pediatrics Research; head of Department of Pediatrics Research, MD Anderson Cancer Center; Charter of NIH Study Section. Dr. Kleinerman has published more than 100 cancer-related articles.
Liangfang Zhang, associated professor at the University of California San Diego, has created a physiological nanoparticle vehicle for tumor-targeted delivery of doxorubicin and other chemical agents. His recent publications have appeared in the Proceedings of the National Academy of Sciences of the United States of America, Nanobiotechnology, and Nanomedicine and Nanoscale.
Competing interest
The authors have declared that no conflict of interest exists.
Author contributions
JH generated most of the data and figures; XX performed Northern blotting and PCR analysis; SZ was a contributor to Figure
6; EK, LZ, and SL were primary contributors to the experimental design, MS integration, and editing. All authors’ read and approved the final manuscript.