Murine regulatory T cells utilize granzyme B to promote tumor metastasis

Abstract

Regulatory T cells (Tregs) possess a wide range of mechanisms for immune suppression. Among them, Granzyme B (GzmB) and perforin expressed by Tregs were shown to inhibit tumor clearance in previous reports, which contradicted the canonical roles of these cytotoxic molecules expressed by cytotoxic T cells and NK cells in antitumor immune responses. Given the ability of the tumor to manipulate the microenvironment, Treg-derived GzmB function may represent an important approach to aid in tumor growth as well as facilitating tumor metastasis. In this study, we utilized Treg-specific GzmB knockout (Foxp3^creGzmB^fl/fl) mice to test whether Treg-derived GzmB can aid in tumor progression and metastasis. Using an IL-2 complex to activate GzmB expression in the non-immunogenic B16-F10 tumor model, we provide evidence to show that GzmB produced by Tregs is important for spontaneous metastasis to the lungs. In addition, we depleted CD8 + T cells to selectively measure the impact of Treg-derived GzmB in an experimental lung metastasis model by intravenous injection of B16-F10 tumor cells; our results demonstrate that Treg-derived GzmB plays an important role in increasing the metastatic burden to the lungs.

Introduction

Regulatory T lymphocytes (Tregs) are a specific subset of CD4⁺ T cells that play an important role in inducing and maintaining peripheral tolerance. These cells were initially discovered in the early 1970s as tolerance inducing population of cells that were thymus derived [1, 2]. It was not until 1995, when Sakaguchi and colleagues showed that CD25, the high affinity IL-2 receptor α chain, could be a phenotypic marker for suppressive CD4⁺ T cells [3]. In 2003, studies identified the importance of the transcription factor forkhead box P3 (FoxP3) in the development and function of CD4⁺CD25⁺ regulatory T cells [4]. Mice deficient in FoxP3 do not develop Tregs and developed a fatal lymphoproliferative disease [5]. In fact, this is also observed in humans with FoxP3 mutations that lead to an autoimmune condition called immune dysregulation, polyendocrinopathy, enteropathy, and X-linked (IPEX) syndrome [5]. Overexpressing FoxP3 in mice leads to a greater number of suppressive cell subsets even in CD4⁺CD25⁻ and CD8⁺ cells [6–8].

The generation and function of Tregs are extremely important during early development for maintaining tolerance later in life. Depletion of thymic-derived Tregs after birth results in severe multi-organ autoimmunity [9]. It is well documented now that Treg deficiency or dysfunction can lead to severe autoimmunity [5]. Although Tregs are important for the prevention of autoimmune diseases such as type 1 diabetes and the suppression of chronic inflammatory diseases such as asthma, this suppressive strength is now known to work as a double-edged sword [10, 11]. Tregs have been shown to greatly suppress antitumor immunity and in some cases to prevent the killing of certain pathogens [12–14]. In addition, in patients with malignant pathologies, increased level of peripheral Tregs before therapy correlates with shorter progression free survival, whereas elevated Tregs in circulation and tumor tissues of patients with breast cancer, non-small cell lung carcinoma (NSCLC), ovarian cancer, and hepatocellular carcinoma are associated with poorer prognosis and higher risk of recurrence [15–19].

Tregs express a wide range of suppressive mechanisms [8]. Among them, Granzyme B and perforin expressed by Tregs were shown to inhibit tumor clearance in previous reports [19–21], which contradicted the canonical roles of these cytotoxic molecules expressed by cytotoxic T cells and NK cells in antitumor immune responses [22, 23]. Given the ability of the tumor to manipulate the microenvironment, Treg-derived GzmB function may represent an important approach to aid in tumor growth as well as facilitating tumor metastasis. In this study, we utilized Treg-specific GzmB knockout (Foxp3^creGzmB^fl/fl) mice to test whether Treg-derived GzmB can aid in tumor progression and metastasis. Using an IL-2 complex to activate GzmB expression in the non-immunogenic B16-F10 tumor model, we provide evidence to show that GzmB produced by Tregs is important for spontaneous metastasis to the lungs. In addition, we depleted CD8⁺ T cells to selectively measure the impact of Treg-derived GzmB in an experimental lung metastasis model by intravenous injection of B16-F10 tumor cells, our results demonstrate that Treg-derived GzmB plays an important role in increasing the metastatic burden to the lungs.

Materials and methods

Cell lines and mice

B16-F10 cells were obtained from ATCC and cultured as previously described [24]. Wild-type mice were obtained from Jackson Laboratories (C57BL/6 J). GzmB knockout (KO) and Perforin KO mice were generated and maintained as previously described [25–33]. FoxP3^cre mice were obtained from Jackson Laboratories (B6.129(Cg)-Foxp3^{tm4(YFP/icre)Ayr}/J and bred with GzmB^fl mice to create homozygous Foxp3^creGzmB^fl/fl mice as previously described [24]. All mice were maintained in specific pathogen-free housing, and all experiments were conducted in accordance with the animal care guidelines from the Office of Animal Welfare Assurance at the University of Maryland School of Medicine Veterinary Resources using protocols approved by the Institutional Animal Care and Use Committee.

Flow cytometry

Briefly, cells were washed using flow buffer (PBS with 2% FBS), and Fc receptors were blocked with the addition of unlabeled anti-CD16/CD32 (BD Biosciences, 553,142) for 20 min in 4 °C. Extracellular markers and fixable LIVE/DEAD Fixable Aqua (Invitrogen, L34966A) were stained together in PBS for 40 min in 4 °C and washed two times with flow buffer. Intracellular stains were performed using the eBioscience FoxP3/Transcription Factor Staining Buffer Set (Invitrogen, 00,552,100). Cells were fixed overnight in 4 °C using the Intracellular Fixation Buffer. Cells were resuspended in 1 × permeabilization buffer and incubated for 5 min. Cells were washed using 1 × permeabilization buffer, and potential intracellular Fc receptors were blocked with the addition of unlabeled anti-CD16/CD32 for 10 min in 4 °C. Samples were split, and test Abs or isotype control Abs were added and incubated for 1 h in 4 °C and washed twice in 1 × permeabilization buffer before resuspending in flow buffer. Samples were run on the Aurora spectral flow cytometer (Cytek Biosciences) in the Center for Innovative Biomedical Resources at the University of Maryland School of Medicine. Unmixed samples were analyzed using FlowJo software (FlowJo). Background staining was assessed with unstained controls, stained isotype controls, and experimental negative controls when possible.

IL-2 complex treatment for Treg activation in vivo

Murine IL-2 cytokine (PeproTech) was incubated with IL-2mAb (JES6-1A12) (BioXCel) at a 1 µg:5 µg ratio for 30 min at 37 °C. We then brought up the solution to 200 µl in 1xPBS and injected peritoneally for 3 consecutive days. For Treg expansion in the subcutaneous tumor model, we injected every other day for four total injections.

ELISPOT

We activated Tregs in vivo by injecting IL-2c for three consecutive days in WT, GzmB^ko and FoxP3^creGzmB^fl/fl mice. Five days following the first injection, we harvested the spleen from the mice and isolated CD4⁺CD25⁺ T cells, which were then plated on enzyme-linked immunospot PVDF membrane coated with anti-GzmB antibody at various densities and cultured for various of times as indicated on figure legend, in 37 °C and 7% CO₂. The production of spots using the GzmB ELISPOT (R&D systems) according to manufacturer’s instructions. The number of blue-black colored spots were counted with ImmunoSpot CTL. Pictures were taken using Zeiss Stemi 508 Stereomicroscope.

Tumor growth and metastasis in vivo

2 × 10⁵ B16-F10 tumor cells were injected subcutaneously in the right flank. Tumor growth was measured using an automated caliper. Tumors were allowed to grow to 4200mm³ or 20 mm in any dimension for survival studies. Upon endpoint, mice were euthanized, and lungs were harvested to measure metastasis. For experimental tumor metastasis, 3 × 10⁵ B16-F10 tumor cells were injected intravenously. Mice were then euthanized at variously time point after injection to examine metastasis in the lungs and liver.

Results

IL-2 complex activates GzmB and expands Tregs but also activates GzmB in CD8 + T cells in vivo

We investigated methods to test how GzmB impacts tumor growth and metastasis in vivo. We first tested the ability of tumor cells to grow in WT versus GzmB knockout (KO) and Perforin (Prf1) KO mice. We injected 2 × 10⁵ B16-F10 cells in the flank of WT, GzmB^ko, or GzmB-Prf1^dko mice. We found that there was no difference in the tumor growth or the survival of these mice. Notably, flow cytometry analyses revealed that Granzyme B protein expression was not detected in either CD8 + T cells or Tregs in the tumor-bearing mice, probably due to the low immunogenicity of the B16-F10 tumor cells.

To study the impact of Granzyme B in tumor immunity in vivo, we investigated the ability of IL-2 complex to activate Tregs [34, 35]. We incubated 1 µg of murine IL-2 cytokine and 5 µg of IL-2 monoclonal antibody (JES6-1A12) for 30 min at 37 °C to allow the antibody to form a complex with the cytokine (IL-2c). We then injected the complex, or PBS and isotype control (rat IgG2a,k), intraperitoneally into mice once a day for 3 consecutive days. Five days following the first injection, we collected the spleen to measure the phenotypic changes. We found that the IL-2c increases the frequency of FoxP3⁺ cells as well as increasing GzmB expression in the FoxP3 + cells (Fig. 1). Interestingly, we found that the IL-2c biased for Treg activation also activates CD8 + T cells, inducing GzmB expression.

Fig. 1.

IL-2 complex induces Treg expansion and GzmB expression in both Tregs and CD8 + T cells in vivo. IL-2 complex was injected for 3 consecutive days. Five days following the first injection, the spleens of mice (n = 6 in each group combined from 2 independent experiments) were harvested to analyze cellular frequency and phenotypes. Representative flow plots: A Frequency of Tregs (gated on live cells). B Frequency of CD8 + T cells (gated on live cells). C Intracellular GzmB expression in Tregs (gated on CD4 + cells). D Intracellular GzmB expression in CD8 + T cells (gated on TCRβ + Cells)

Generating Treg-specific deletion of GzmB in vivo

To test the contribution of Treg-expressed GzmB, we generated a novel mouse model to specifically delete GzmB from FoxP3 expressing Tregs (FoxP3^creGzmB^fl/fl). Because IL-2c can activate GzmB expression in FoxP3 + cells and CD8 + T cells, we used this IL-2c method to test the efficiency and specificity of our FoxP3^creGzmB^fl/fl mouse model. We found that GzmB has specifically been deleted from FoxP3 + cells, while not affecting the expression of GzmB in CD8 + T cells (Fig. 2). This result is pivotal for our study of the impact of Treg-derived GzmB in the tumor model.

Fig. 2.

Treg-specific deletion of GzmB expression in vivo. Flow plots show the efficiency and specificity of the Cre-loxP model. WT, GzmB^ko, and FoxP3^creGzmB^fl/fl mice (n = 6 in each group combined from 2 independent experiments) receiving daily intraperitoneal injections of Treg-biased IL-2 complex for three days shows activation of Tregs and CD8 + T cells. Five days after first injection, splenocytes were stained for T cell markers and intracellular FoxP3 and GzmB. A Representative plots of live T cells showing the activation of CD4 + and CD4- T cells. This confirms the specificity of FoxP3^creGzmB^fl/fl. B Representative plots of live cells showing CD4 + FoxP3 + cells. Gated the double positive population is a representative plot of CD25 and GzmB

Due to the rare nature of Tregs and GzmB expression within the Tregs, we conducted an additional experiment to confirm the specificity and efficiency of our Cre-loxP model. On day 5 following the first IL-2c injection, we isolated CD4 + and CD8 + T cells from the spleens and plated 2 × 10⁶ cells on αCD3-αCD28 coated plates for 48 h. Following the culture, we conducted flow cytometric analysis and found that GzmB expression to be upregulated in both CD4⁺ and CD4⁻ T cells (Fig. 3). GzmB expression is also found in FoxP3⁺ cells in the WT mice but was not found in both GzmB-Perf^dko and FoxP3^creGzmB^fl/fl mice.

Fig. 3.

Treg-specific deletion of GzmB expression ex vivo. WT, GzmB^koPrf1^ko, and FoxP3^creGzmB^fl/fl mice (n = 6 in each group combined from 2 independent experiments) receiving daily intraperitoneal injections of Treg-biased IL-2 complex for three days shows activation of Tregs and CD8 + T cells. Five days after first injection, splenocytes were harvested and isolated Pan T cells were cultured on αCD3-αCD28 coated plates for 48 h. Following the culture, we stained for T cell markers and intracellular FoxP3 and GzmB. A Representative plots of live T cells showing the activation of CD4 + and CD4- T cells. This confirms the specificity of FoxP3^creGzmB^fl/fl B Plots of live cells showing CD4 + FoxP3 + cells. Gated the double positive population is a representative plot of GzmB. C Summary of GzmB + percentages in TCRβ + T cells. D Summary of GzmB + percentages in CD4 + FoxP3 + Tregs

To further examine GzmB production by Tregs in vivo, we activated Tregs in vivo by injecting IL-2c for three consecutive days in WT, GzmB^ko and FoxP3^creGzmB^fl/fl mice. Five days following the first injection, we harvested the spleen from the mice and isolated CD4⁺CD25⁺ T cells. Due to the small percentage of GzmB expression within the rare Treg population, we chose to plate 10,000 cells on an ELISPOT plate overnight. Our selection quality for CD4 + T cells was successful (65% CD4⁺ to 93% CD4⁺); however, the quality of CD25⁺ was less effective (10% CD25⁺ to 66% CD25⁺) (Fig. 4). We found a high number of spots following the overnight culture. The impurity of the CD25 selection could explain the number of spots for the FoxP3^creGzmB^fl/fl group; however, there is still a large decrease in the GzmB spots compared to WT. We surmise that this decrease is due to GzmB being deleted when FoxP3-Cre is expressed in Tregs while the remaining GzmB found is due to activation of CD4⁺ CD25⁻ T cells, which are likely to be FoxP3⁻ non-Tregs.

Fig. 4.

ELISPOT shows Treg-specific deletion of GzmB expression in vivo without additional stimulation. WT, GzmB^ko, and FoxP3^creGzmB^fl/fl mice (n = 6 in each group combined from 2 independent experiments) receiving daily intraperitoneal injections of Treg-biased IL-2 complex for three days. Five days after first injection, CD4⁺CD25⁺ T cells were selected from splenocytes and then plated on enzyme-linked immunospot PVDF membrane coated with anti-GzmB antibody. A A diagram showing the experimental process. B Flow plots showing percentages of CD4⁺CD25⁺ T cells pre and post selection, gated on TCRβ + and CD4 + cells, respectively. C ELISPOT images following overnight culture with 10,000 cells plated. D Summary graph showing number of spots

Depleting CD8 + T cells while activating Tregs promotes spontaneous metastasis

We reasoned that global deletion of GzmB would mask the complex and even opposite contribution of specific cell types especially CD8 + T cells versus Tregs. Based on the effects of IL-2c on CD8 + T cells and Tregs, we evaluated if the simultaneous activation of CD8 + T cells and Tregs would counteract any pro-tumoral versus anti-tumoral effect. We subcutaneously injected 2 × 10⁵ B16-F10 cells in the flank of C57Bl/6 WT mice. Two days prior to the injection, mice were either injected intraperitoneally with 200 µg of CD8α depleting antibody or the isotype control, rat IgG2a,k. Five days following the tumor inoculation, mice were intraperitoneally injected with IL-2c or PBS and isotype control, rat IgG2a,k, once every other day for 4 total injections. These combinations give us a total of four groups: Control, CD8-depleted, Treg-expanded, CD8-depleted-Treg-expanded. However, we found that there was no significant difference in tumor growth between different groups (Fig. 5A-D). Yet we did see a trend of prolonged survival in the Treg-expanded group, which may be due to the activation of CD8 + T cells slowing down tumor growth or metastasis. Indeed, the most interesting result we found was that the CD8-depleted-Treg-expanded group showed the greatest percentage of spontaneous metastasis to the lungs, four mice out of the five had metastatic nests in the lungs. These results suggest that GzmB produced by the increasing frequency of Tregs can increase the metastatic burden of B16-F10.

Fig. 5.

Depletion of CD8 + T cells while expanding Tregs increases metastatic burden via Treg-derived GzmB-dependent fashion. (A-D) As shown in the diagram above, WT mice were depleted of CD8 + T cells and injected with B16-F10 cells subcutaneously. Following tumor injection, mice were injected with IL-2c intraperitoneally every other day for four total injections. Tumor size was measured for the duration of the experiment, n = 4–5 in each group. A Tumor volume for the duration of experiment. B Survival curve. C Representative images of lungs harvested from mice that succumbed to tumor. D Percentage of mice that developed metastasis in the lungs at endpoint (tumor size or upon death) between day 35–46. E–G WT, FoxP3^cre, GzmB^fl/fl, FoxP3^creGzmB^fl/fl mice (n = 5 in each group) were depleted of CD8 + T cells and injected with B16-F10 cells subcutaneously. Following tumor injection, mice were injected with IL-2c intraperitoneally every other day for four total injections. E Tumor volume for the duration of experiment. F Survival curve. G Percentage of mice that developed metastasis in the lungs upon endpoint

Treg-derived GzmB promotes spontaneous and experimental tumor metastasis

We then focused on the CD8-depleted-Treg-expanded condition to examine tumor growth and spontaneous metastasis in WT, GzmB^fl/fl, FoxP3^cre, and FoxP3^creGzmB^fl/fl mice. Unexpectedly, we found there was no difference in tumor growth and survival between the different groups (Fig. 5E-G). However, we did find that the FoxP3^creGzmB^fl/fl group had a decrease in the percentage of mice with spontaneous metastatic burden in the lungs (20%), compared to the WT and control groups (60, 80, 80%).

These were very interesting results; however, the timing of spontaneous metastasis to the lungs and subsequent lung harvest were variable. To more accurately measure the impact of Treg-derived GzmB on tumor metastasis, we adopted an experimental metastasis model by intravenous injection of tumor cells under the CD8-depleted-Treg-expanded condition. Specifically, we depleted CD8 + T cells, injected 3 × 10⁵ B16-F10 cells intravenously, and injected IL-2c intraperitoneally for three consecutive days. Then we measured experimental metastasis into the lungs at various time points. We found a trend of decrease in metastatic nests in the FoxP3^creGzmB^fl/fl mice at day 8 and day 16 following intravenous injection. Remarkably, we found at day 21 a significant difference in metastatic nests, weight of the lung, as well as the area of the lung infiltrated with tumor (Fig. 6A-D). Furthermore, to examine whether the different metastasis rate can impact host survival, we decreased the dose of tumor cells by injecting 1 × 10⁵ B16-F10 cells intravenously. We found that the FoxP3^creGzmB^fl/fl mice survived significantly longer than the WT mice (Fig. 6E). Based on these studies, we conclude that GzmB produced by Tregs contributes to an increase in metastatic burden and a reduced overall survival.

Fig. 6.

Treg-derived GzmB significantly increases metastatic burden at late timepoint and precipitates earlier host death. (A-D) As shown in the diagram above, WT and FoxP3^creGzmB^fl/fl mice were depleted of CD8 + T cells and injected with 3 × 10⁵ B16-F10 cells intravenously. Tregs were activated to produce GzmB using IL-2 complex (rIL-2/IL-2mAb) for 3 consecutive days. Lungs were harvested on day 21. A Representative images and H&E stain of lungs with metastatic burden. B Number of metastatic nests counted for each mouse, n = 6 in each group combined from 2 independent experiments. A student’s t-test was used to calculate statistics, p < 0.0001. C Weight of the lungs after harvesting, p = 0.042. D Percentage of the lungs being tumors, p = 0.0475. E) Survival curve of mice injected with 1 × 10⁵ B16-F10 cells intravenously, n = 10–11 in each group combined from 2 independent experiments. Log-rank (Mantel–Cox) test was used to analyze statistical significance, p = 0.0236

Discussion

The classical understanding of GzmB is that it functions as a cytotoxic effector molecule in lymphocyte-mediated immune response against transformed malignant cells or cells infected by intracellular pathogens [23]. However, our previous studies found that GzmB⁺ Tregs are important in aiding in tumor growth and inhibiting tumor clearance [19, 20]. Treg frequency as well as the GzmB expression increases as the tumors grow [20]. In this study, aided by the more advanced Treg-specific GzmB KO mice, we present evidence demonstrating that Treg-derived GzmB can also promote tumor metastasis. We found that depleting CD8 + T cells while expanding and activating Tregs, led to an abnormal increase in spontaneous metastasis to the lungs from a primary tumor site. We then showed that GzmB specifically produced by Tregs can increase the metastatic burden to the lungs.

There is growing evidence for the importance of Tregs in tumor progression [36, 37]. Tregs are extremely important in the induction and maintenance of peripheral tolerance to self-antigens [6, 9, 11]. Given that tumor cells are primarily expressing self-antigens that T cells should have been tolerized to, Tregs are playing a role to inhibit any ‘unintended’ damage to the host [8, 36, 37]. Several studies have focused on investigating the role of GzmB in Treg-mediated immune suppression. It has been found that Tregs can suppress B cell proliferation via a GzmB- and Perforin-dependent fashion [38]. Another study found that Tregs can suppress effector CD4 + T cells via a GzmB-dependent and Perforin-independent process [39]. Our laboratory previously discovered that the frequency of GzmB-expressing Tregs increases for the duration of tumor growth and that GzmB + Tregs inhibit tumor clearance [19, 20]. Furthermore, a study on oral squamous cell carcinoma showed higher prevalence of GzmB-expressing Tregs [40].

We recently performed and published pan-cancer expression analysis of GzmB and Foxp3 in a number of cancer types [41]. We have found higher GzmB mRNA expression in many cancers such as cholangiocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, kidney renal clear cell carcinoma, stomach adenocarcinoma compared to normal tissues [41]. However, cancers such as lung adenocarcinoma and lung squamous cell carcinoma have low GzmB expression [42]. What was interesting was the range of GzmB expression within the same cancer type. We analyzed Pan-cancer expression and found that some patients with cancers such as uveal melanoma and pancreatic adenocarcinoma have a better probability of survival when there is minor GzmB transcription compared to a large amount [43]. This seems to correlate with FoxP3 expression greatly in some patients [41]. The majority of cancers showed a greater probability of survival with greater expression of GzmB. What was interesting was finding high FoxP3 expression increased probability of survival of patients with cancers such as head and neck squamous cell carcinoma and stomach adenocarcinoma. One interesting finding was that some cancers had a large amount of FoxP3 expression with minor GzmB expression such as lower grade glioma (LGG), invasive breast carcinoma (BRCA), ovarian serous cystadenocarcinoma (OV) [41]. Understanding the correlation with FoxP3 and GzmB in different tumor conditions may prove to be important in the generation of personalized tumor targeting therapies.

Lastly, it would be important to investigate the mechanisms by which GzmB aids in tumor migration and metastasis. A previous study demonstrated that Foxp3 + Tregs induced the death of dendritic cells (DC) in tumor-draining lymph nodes, but not in the absence of tumor, via a tumor-antigen and perforin dependent mechanism [21], leading to a conclusion that Treg-dependent Perforin-mediated DC death in tumor-draining lymph nodes limits the onset of CD8( +) T cell responses. Additional studies are required to define other potential mechanisms by which GzmB promotes tumor metastasis, such as Perforin-independent extracellular and intracellular activities. One possible mechanism is through the activation of extracellular proteins such as matrix metalloproteases. Another possible mechanism is how Perforin-independent uptake of GzmB by tumor cells changes their migratory capacity or response to inflammatory danger signals [44]. This notion is in conjunction with a recent study demonstrating that GzmB can elicit non-lethal DNA damage through a Perforin-independent mechanism [45].

In conclusion, our results in this study expand the role of GzmB in Treg-mediated pro-tumor activity. Additional studies utilizing different inflammatory and tumor models may reveal that selective inhibition of GzmB has potential to improve cell-based immunotherapies in the treatment of malignancies that manipulate this pathway to evade the immune system.

Author contributions

E.T. and X.C. designed the project, performed the experiments and wrote the manuscript. R.R.K.K., D.J. and L.W. performed in vivo tumor experiments. All authors reviewed the manuscript.

Data availability

The data sets used and/or analyzed during current study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest.