NK cells and CD8+ T cells cooperate to improve therapeutic responses in melanoma treated with interleukin-2 (IL-2) and CTLA-4 blockade

Combination immunotherapy with IL-2 and CTLA-4 blockade results in significantly delayed
tumor growth and prolonged survival

IL-2 therapy and CTLA-4 blockade have individually been shown to improve anti-tumor
responses and are approved as monotherapies for the treatment of metastatic melanoma
12],14],25]. Since each of these monotherapies may work through a unique mechanism, we explored
the effect of combination IL-2 and CTLA-4 blockade immunotherapy using the murine
B16-F10 (B16) melanoma model. Specifically, we challenged mice with B16 (1–1.2 x 10⁵ cells by intradermal injection) and treated with CTLA-4 blockade (?CTLA-4; 100 ?g
via intraperitoneal injection [i.p.] on days 3, 6 and 9 after tumor challenge) only,
IL-2 (100,000 units i.p every 12 hours on days 4–8 post tumor challenge) only, or
the combination IL-2 and CTLA-4 blockade. We designed the treatment regimen to correspond
with the manner in which these monotherapies are utilized in the clinical setting
(Figure 1A). We found that both monotherapies, IL-2 and CTLA-4 blockade individually delayed
tumor growth compared to the control (IgG?+?PBS) treatment (29 mm² and 14 mm² versus 76 mm²; p??0.01 and p??0.001, respectively, at day 14) (Figure 1B, C) and prolonged survival (30% and 50% versus 0%; p??0.05 and p??0.001, respectively,
at day 23) (Figure 1D). Combination IL-2 and CTLA-4 blockade resulted in significantly delayed tumor growth
compared to IL-2 only or CTLA-4 blockade only (2 mm² versus 29 mm² [p??0.01] and 14 mm² [p??0.01], respectively, at day 14) and significantly prolonged survival (100% versus
30% [p??0.01] and 50% [p??0.05], respectively, at day 23) (Figure 1B-D). No added toxicity was detected with combination IL-2 and CTLA-4 blockade compared
to either treatment alone (Table 1). These findings demonstrate an augmented anti-tumor effect with combination IL-2
and CTLA-4 blockade immunotherapy compared to either monotherapy alone against established,
poorly immunogenic melanoma.

Figure 1. Combination IL-2 and CTLA-4 blockade immunotherapy results in reduced tumor growth
and prolonged survival. (A) Schematic of the experimental design. (B) Cumulative graph of mean tumor size (mm²) per group from experiment described in (A). (C) Tumor size (mm²) of individual mice in each group from experiment described in (A). (D) Cumulative graph of mean percent (%) survival per group from experiment described
in (A). Seven to ten mice were included in each group. Graphs represent one experiment of
six conducted with similar results. *P??0.05, **P??0.001, ***P??0.001, ns?=?not
significant.

Table 1. Chemistry screen IL-2 and of CTLA-4 blockade immunotherapy

CTLA-4 blockade promotes immune cell infiltration within the tumor

Immune infiltration within the tumor is a positive prognostic indicator of response
to immunotherapy 33],34]. Thus, we assessed immune infiltration of B16 by determining the proportion of CD45+
cells within the tumor following treatment. To distinguish between tumor-resident
immune infiltration and immune cells restricted to the vasculature, we intravenously
injected anti-CD45 conjugated to a fluorophore three minutes prior to sacrificing
the mice. Following tumor dissection and tissue dissociation, we stained cells in vitro with anti-CD45 conjugated to a different fluorophore than the one used for intravenous
in vivo staining. This permitted the identification of tumor-resident leukocytes by flow
cytometry (Figure 2A). CTLA-4 blockade alone and in combination with IL-2 promoted CD45+ immune infiltration
(among live cells) compared to the control treatment (mean %CD45+ cells in the tumor:
48% and 28% versus 11%, p??0.001 and p??0.05, respectively) (Figure 2A,B). No significant difference in immune infiltration was observed between the combination
group and the CTLA-4 blockade treatment group (48% versus 28%, respectively, p??0.05)
or between IL-2 alone and the control group (18% versus 11%, respectively, p??0.05)
(Figure 2B). Further, to confirm that the combination immunotherapy enhanced immune infiltration,
we used epifluorescence microscopy. This demonstrated that IL-2 and CTLA-4 blockade
combination immunotherapy results in augmented CD45+ immune cell infiltration within
the tumor (1×10⁵ versus 4×10⁴ CD45+ cells/10⁶ nuclei, respectively, p??0.05) (Figure 2C). These data demonstrate that CTLA-4 blockade, individually and when combined with
IL-2, promotes immune cell infiltration within the tumor microenvironment.

Figure 2. Combination IL-2 and CTLA-4 blockade immunotherapy increases tumor immune infiltration.
(A) Flow cytometry plots of tumors dissected at day 14 and analyzed by flow cytometry
for CD45 expression from the experiment described in Figure 1A. Only tumor-infiltrating lymphocytes were analyzed (as determined by comparison
of intravenous CD45 staining compared to in vitro CD45 staining). (B) Cumulative graph showing mean percent CD45+ T cell infiltration (of live cells) in
the tumor from three independent experiments described in (A). (C) Representative immunofluorescence microscopy images of CD45 (red) and DAPI (4?,6-diamidino-2-phenylindole;
blue) staining of tumors from experiment in (A). Scale bars = 20 microns. Three to five mice were included in each group per experiment.
*P??0.05, ***P??0.001, ns?=?not significant.

CTLA-4 blockade results in increased CD8+ T cells in the tumor microenvironment

CD8+ T cells are vital mediators of anti-tumor responses. Therefore, we characterized
the tumor immune infiltrate to determine the effect of combination IL-2 and CTLA-4
blockade on CD8+ T cells. CTLA-4 blockade alone and in combination with IL-2 resulted
in an increased proportion of CD8+ T cells among the immune cell infiltrate (CD45+
cells) of the tumor compared to the control group (35% and 19% versus 7%, p??0.01
and p??0.05, respectively) (Figure 3A,B). No significant difference in immune infiltration was observed between the combination
group and the CTLA-4 blockade only treatment group (19% versus 34%, respectively,
p??0.05) or between IL-2 alone and the control group (6% versus 7%, respectively,
p??0.05) (Figure 3A, B). These data suggest that CTLA-4 blockade may be the primary driver of CD8+ T
cell tumor infiltration after CTLA-4 blockade with or without IL-2 immunotherapy.

Figure 3. Combination IL-2 and CTLA-4 blockade immunotherapy increases the proportion of tumor-infiltrating
CD8+ T cells. (A) Flow cytometry plots of tumors dissected at day 14 and analyzed for CD3?+?CD8+ T
cells (of CD45+ cells within the tumor) by flow cytometry from the experiment described
in Figure 1A. (B) Cumulative graph showing the mean percent of CD3?+?CD8+ T cells from experiment in
(A). (C) Cumulative graph showing the mean percent of CD3?+?CD8+ T cells in the tumor-draining
lymph nodes from experiment in (A). (D) Representative histograms showing expression of cell markers from CD3?+?CD8+ T cells
in the tumor and from a representative tumor-draining lymph node (LN). Numbers represent
mean fluorescence intensity (MFI). (E) Cumulative graph of mean percent of CD3?+?CD8+ T cells expressing PD-1 from experiment
in (A). Cumulative figures are from at least three independent experiments (with 3–5 mice
per group in each experiment). *P??0.05, ** P??0.01, ns?=?not significant.

To determine whether the increased proportion of CD8+ T cells in the tumor microenvironment
was the result of an overall systemic increase in CD8+ T cells, we analyzed the tumor-draining
(inguinal) lymph nodes and spleen for CD8+ T cell numbers. No increase was observed
in the proportion of CD8+ T cells within the tumor-draining lymph nodes as a result
of IL-2 alone, CTLA-4 blockade alone, or combination IL-2 and CTLA-4 blockade compared
to the control treatment (in all groups CD8+ T cells constituted approximately 27%
of CD45+ cells, p??0.05 for all comparisons) (Figure 3C). Similarly, no differences in CD8+ T cells amongst the groups were seen in the
spleen (data not shown). This demonstrates that the increased proportion of tumor-infiltrating
CD8+ T cells is not the result of a systemic increase in peripheral CD8+ T cells,
but rather changes specific to the tumor microenvironment.

To determine whether the activation status of CD8+ T cells correlates with the anti-tumor
immune responses we observed (Figure 1B,C), we determined the expression of CD8+ T cell activation markers. Tumor-infiltrating
CD8+ T cells showed an activated phenotype (with increased CD44, CD25, CD69, and Tbet
expression and decreased CD62L expression) compared to CD8+ T cells in the tumor-draining
lymph nodes (Figure 3D). However, we observed no differences in these activation markers when comparing
the monotherapies (IL-2 or CTLA-4 blockade) or the combination IL-2 and CTLA-4 blockade
immunotherapy to the control group (IgG?+?PBS) (Figure 3D). Since PD-1 is known to be upregulated during T cell activation, we likewise determined
its expression on tumor-infiltrating CD8+ T cells. IL-2 and CTLA-4 blockade monotherapies
as well as the combination IL-2 and CTLA-4 blockade immunotherapy resulted in an increased
proportion of CD8+ T cells expressing PD-1 (among all CD8+ T cells) within the tumor
(46%, 43%, 53% versus 24%, p??0.01, p??0.05, and p??0.01, respectively) (Figure 3E). However, no differences in PD-1 expression between the combination immunotherapy
and the CTLA-4 blockade and IL-2 monotherapies (53% versus 46% and 43%, p??0.05 for
both, respectively) were detected. These data demonstrate the activation status of
CD8+ T cells does not correlate with the improved anti-tumor immune responses we observed
for combination IL-2 and CTLA-4 blockade immunotherapy.

Combination CTLA-4 blockade and IL-2 immunotherapy increases regulatory T cells within
the tumor immune infiltrate

Regulatory T cells (Tregs) promote tumor growth through the suppression of anti-tumor
immune responses. Tregs suppress CD8+ T cell and NK cell responses specifically through
CTLA-4 35] and by acting as IL-2 sinks through expression of high levels of the high affinity
IL-2 receptor, CD25 36]. Thus, we determined whether combination IL-2 and CTLA-4 blockade immunotherapy affects
Tregs in the tumor. While CTLA-4 blockade alone and IL-2 alone did not increase the
proportion of Tregs (among CD45?+?CD3+ cells) within the tumor compared to the control
treatment (1% and 2% versus 2%, p??0.05 for both, respectively), the combination
of these immunotherapies significantly increased Tregs among the immune infiltrate
compared to the control treatment (5% versus 2%, respectively, p??0.01) (Figure 4A, B). These findings demonstrate that the improved responses observed with combination
IL-2 and CTLA-4 blockade immunotherapy compared to either monotherapy are associated,
unexpectedly, with an increase in the proportion of Tregs in the tumor.

Figure 4. Combination IL-2 and CTLA-4 blockade immunotherapy increases the proportion of tumor-infiltrating
Tregs. (A) Flow cytometry plots of tumors dissected at day 14 and analyzed for regulatory CD4?+?Foxp3+
T cells (Tregs; of CD45?+?CD3+ cells within the tumor) by flow cytometry from the
experiment described in Figure 1A. (B) Cumulative graph showing mean percent of Tregs from experiment in (A). (C) Ratio of CD8+ T cells to Tregs in the tumor from experiment in (A). (D) Cumulative graph showing mean percent of Tregs from the tumor-draining lymph nodes
from experiment in (A). (E) Ratio of CD8+ T cells to Tregs in the tumor-draining lymph node from experiment in
(A). Cumulative figures are from at least three independent experiments (with 3–5 mice
per group in each experiment). *P??0.05, **P??0.01, ns?=?not significant.

The efficacy of immunotherapies has been previously attributed to their ability to
increase the proportion of CD8+ T cells to Tregs 31]; therefore, we determined this ratio within the tumor. We observed a trend towards
an increased ratio of CD8+ T cells (of CD45+ cells) to Tregs (of CD45+ cells) in the
CTLA-4 blockade monotherapy group compared to the combination IL-2 and CTLA-4 blockade
immunotherapy group (47 versus 8, respectively, p?=?0.06) and compared to the control
group (47 versus 10, respectively, p?=?0.07) (Figure 4C). However, we found no statistically significant differences in the ratio of CD8+
T cells to Tregs when comparing the combination IL-2 and CTLA-4 blockade immunotherapy
group to any other group (p??0.05 for all comparisons) (Figure 4C and data not shown). These data suggest that the CD8+ T cell to Treg ratio does
not correlate with the improved anti-tumor responses observed with combination IL-2
and CTLA-4 blockade immunotherapy compared to either monotherapy in this model.

To determine whether the observed differences and trends in the proportion of Tregs
and the ratio of CD8+ T cells to Tregs was confined to the tumor microenvironment,
we determined these measures within the tumor-draining lymph nodes and spleen. Here,
we observed no differences or trends in either the proportion of Tregs or the ratio
of CD8+ T cells to Tregs with the monotherapies or the combination IL-2 and CTLA-4
blockade in the tumor-draining lymph nodes (in all groups Tregs constituted 3 to 5%
of CD45+ cells, p??0.05 for all comparisons; and in all groups CD8:Treg ratios were
6 to 9, p??0.05 for all comparisons) (Figure 4D, E) or in the spleen (data not shown). These data suggest that the observed differences
and trends in the proportion of Tregs and the ratio of CD8+ T cells to Tregs in the
tumor-draining lymph nodes and spleen do not parallel those found in the tumor microenvironment.

CTLA-4 blockade increases NK cells within the tumor infiltrate and IL-2 modulates
NK cell differentiation status

Since the therapeutic activity of the combination treatment did not depend on the
CD8+ T cell:Treg ratio, we sought to determine if NK cells were involved in the antitumor
activity. CTLA-4 blockade alone and in combination with IL-2 resulted in an increase
in the proportion of NK cells among the non-T cell (CD45?+?CD3-) tumor immune infiltrate
compared to the control (32% and 34% versus 15%, p??0.01 and p??0.05, respectively)
(Figure 5A,B). There were no differences in NK cell infiltration between the CTLA-4 blockade
only and the combination IL-2 and CTLA-4 blockade treatment groups (32% versus 34%,
respectively, p??0.05) (Figure 5A,B). As with our observation for CD45+ cells, CD8+ T cells, and Tregs, the increase
in the proportion of NK cells was not observed in the tumor-draining lymph nodes (in
all groups NK cells constituted approximately 2% of CD45+ cells, p??0.5 for all comparisons)
(Figure 5C) or spleen (data not shown).

Figure 5. Combination IL-2 and CTLA-4 blockade immunotherapy increases the proportion and changes
the differentiation of tumor-infiltrating NK cells. (A) Flow cytometry plots of tumors dissected at day 14 and analyzed for the proportion
of NK cells (of CD45?+?CD3- cells within the tumor) by flow cytometry from the experiment
described in Figure 1A. (B) Cumulative graph showing mean percent of NK cells from experiment in (A). (C) Cumulative graph showing mean percent of NK cells from the tumor-draining lymph nodes
from experiment in (A). (D) Representative histograms showing expression of cell markers by NK cells in the tumor
and from a representative tumor-draining lymph node (LN). Numbers represent mean fluorescence
intensity (MFI). (E) Flow cytometry plots showing expression of CD27 and CD11b on tumor-infiltrating NK
cells from experiment in (A). Cumulative figures are from at least three independent experiments (with 3–5 mice
per group in each experiment). *P??0.05, **P??0.01, ns?=?not significant.

To determine whether the activation status of NK cells correlates with the anti-tumor
immune responses we observed (Figure 1B,C), we determined the expression of well-described NK cell markers. There were no
significant differences in tumor-infiltrating NK cell activation markers PD-1, Tbet,
NKG2D, Eomes, and CD127 in the IL-2 or CTLA-4 blockade monotherapy or combination
immunotherapy treatment groups versus the control group (Figure 5D). Because IL-2 has been reported to increase the proportion of immature NK cells
37] and regulate NK cell maturation, as determined by expression of CD27 and CD11b 38], we determined the effects of combination IL-2 and CTLA-4 blockade immunotherapy
on the maturation status of NK cells within the tumor. Treatment with IL-2 alone and
in combination with CTLA-4 blockade resulted in reduced expression of CD27 and CD11b
compared to the control (3% and 2% versus 12%, p??0.05 for both, respectively) on
NK cells suggesting a less differentiated phenotype (Figure 5E). CTLA-4 blockade had no effect on NK cell maturation compared to the control group
(16% versus 12%, respectively, p??0.05) (Figure 5E). These results demonstrate that CTLA-4 blockade and IL-2 work in combination to
increase the proportion of NK cells in the tumor infiltrate (via the effects of CTLA-4
blockade) and to promote a less differentiated NK cell population (via the effects
of IL-2) within the tumor microenvironment.

Since one possible mechanism through which NK cells might mediate anti-tumor activity
is to delete MHC class I negative tumor cells, and because melanoma has been reported
to evade immune detection through loss of MHC class I expression, we evaluated the
level of MHC class I on the B16 tumor cells after in vivo challenge. We observed that over 92% Melan-A+ cells in the tumor demonstrated surface
expression of MHC class I molecule H-2K^b (Additional file 1: Figure S1). This finding suggests that the importance on NK cells in the context
of combination IL-2 and CTLA-4 blockade immunotherapy is not based on loss of MHC
class I expression on tumor cells.

CD8+ T cells and NK cells are necessary for the efficacy of combination IL-2 and CTLA-4
blockade immunotherapy

NK cells can contribute to anti-tumor immune responses against B16 melanoma, as reported
for various other types of combination immunotherapy approaches 39]-41]. Thus, to determine if NK cells were involved in the anti-tumor activity of IL-2
and CTLA-4 blockade treatment, we repeated the tumor studies in Figure 1 in the absence of NK cell and/or or CD8+ T cells (Figure 6A). Individual depletion of CD8+ T cells or NK cells partially reduced the overall
efficacy of the combination immunotherapy compared to treatment in immune competent
mice (36 mm² and 22 mm² versus 11 mm², p??0.05 and p??0.001, respectively, at day 12) (Figure 6B, left panel). However, depletion of either subset individually was not sufficient
to completely ablate therapeutic efficacy compared to the control (36 mm² and 22 mm² versus 69 mm², p??0.05 for both, respectively, at day 12) (Figure 6B, left panel). Depletion of neither CD4 nor B cells had any effect on the efficacy
of the combination IL-2 and CTLA-4 blockade immunotherapy compared to no depletion
(9 mm² and 16 mm² versus 11 mm², p??0.05 for both, respectively, at day 12) (Figure 6B, right panel). When both CD8+ T cells and NK cells were depleted concurrently during
the course of combination IL-2 and CTLA-4 blockade immunotherapy, therapeutic responses
were abrogated compared to no depletion (82 mm² versus 14 mm², respectively, p??0.05, at day 16) and tumor growth resembled that of the control
group when both CD8+ T cells and NK cells were depleted concurrently (82 mm² versus 55 mm², respectively, p??0.05, at day 16) (Figure 6C,D). Further depletion of both CD8 T cells and NK cells concurrently reduced the
therapeutic response significantly compared to either CD8 or NK cell depletion alone
(p??0.01 for both comparisons). This demonstrates that both CD8+ T cells and NK cells
are necessary for the therapeutic effectiveness of combination IL-2 and CTLA-4 blockade
immunotherapy.

Figure 6. NK and CD8+ T cells are necessary for combination IL-2 and CTLA-4 blockade immunotherapy-mediated
effects. (A) Schematic of the experimental design. (B) Cumulative graph of the mean tumor size (mm²) per group after CD8 or NK cell depletion (left panel) and B cell or CD4+ cell depletion
(right panel) from the experiment described in (A). (C) Cumulative graph of the mean tumor size (mm²) per group after combination CD8 and NK cell depletion from the experiment described
in (A). (D) Tumor size (mm²) of individual mice in each group from experiment described in (C). Four to ten mice were included in each group. Graphs represent one experiment of
three conducted with similar results. *P??0.05, **P??0.001, ***P??0.001, ns?=?not
significant.