Professional Documents
Culture Documents
Anti-CTLA-4 Immunotherapy Does Not Deplete FOXP3þ Regulatory T Cells (Tregs) in Human Cancers
Anti-CTLA-4 Immunotherapy Does Not Deplete FOXP3þ Regulatory T Cells (Tregs) in Human Cancers
Cancer
Research
Anti-CTLA-4 Immunotherapy Does Not Deplete
FOXP3þ Regulatory T Cells (Tregs) in Human
Cancers
Anu Sharma1, Sumit K. Subudhi1, Jorge Blando2, Jorge Scutti2, Luis Vence2,
Jennifer Wargo3, James P. Allison2, Antoni Ribas4, and Padmanee Sharma1,2
www.aacrjournals.org 1233
Sharma et al.
IHC
Translational Relevance IHC was performed on FFPE tumor tissue sections. The tumor
In preclinical murine models, anti-CTLA-4 mAbs promote tissues were fixed in 10% formalin, embedded in paraffin, and
cancer regression by increasing the frequency of effector T cells transversely sectioned. Four-micron sections were used for the
within the tumor microenvironment and selectively depleting IHC analyses. The sections were stained with mouse anti-human
intratumoral regulatory T cells (Treg) via an Fc-dependent mAbs against CD4 (Novocastra, CD4-368-L-A), CD8 (Thermo
mechanism. However, it is unclear whether this occurs in Scientific, MS-457-S), FOXP3 (Abcam, ab96048-206D), and
patients with cancer treated with anti-CTLA-4 therapies. Here- CD68 (Dako, M0876). All sections were counterstained with
in, we show that treatment with IgG1 ipilimumab and IgG2 hematoxylin, dehydrated, mounted, and processed with
tremelimumab mAbs, which have distinct FcgR-mediated peroxidase-conjugated avidin/biotin and 30 -3-diaminobenzidine
mechanisms, increases the density of intratumoral CD4 and (DAB) substrate (Leica Microsystems). The IHC slides were
CD8 T cells, and does not deplete FOXP3þ Tregs across scanned and digitalized using the scanscope system from
different human tumors. These data suggest that there is an Scanscope XT, Aperio/Leica Technologies. The IHC analysis was
1234 Clin Cancer Res; 25(4) February 15, 2019 Clinical Cancer Research
Anti-CTLA-4 Therapy Does Not Deplete FOXP3þ Cells
A Melanoma
CD4 CD8
P < 0.0001 P = 0.0043
1,400 4,000
3,000
1,200
600 1,400
1,200
1,000
400
800
600
200 400
200
0 0
ipilimumab (Fig. 1A; Supplementary Fig. S1A) and tremelimu- cells within the tumor microenvironment. Of note, pretreatment
mab (Fig. 1B) significantly increased the intratumoral densities of biopsies from two patients with melanoma, which were negative
both CD4þ and CD8þ cells. Similar results have been reported in for FOXP3þ cells, showed FOXP3þ cellular infiltration post-
human melanoma upon treatment with tremelimumab (26). tremelimumab treatment (Fig. 2B).
Consistent with these findings, we observed increased densities Furthermore, in bladder and prostate tumors obtained from
of intratumoral CD4þ and CD8þ cells in bladder and prostate patients treated with ipilimumab, there was no reduction in
tumors obtained from patients treated with ipilimumab (Sup- intratumoral FOXP3þ cells (Fig. 3; Supplementary Fig. S3). Taken
plementary Fig. S2A and S2B). together, the data demonstrate that both ipilimumab and treme-
limumab, despite having two different IgG isotypes with poten-
Anti-CTLA-4 immunotherapy does not deplete intratumoral tially different mechanisms of action, similarly augment the
FOXP3þ cells in human cancers densities of intratumoral CD4þ and CD8þ cells, but do not
As it is postulated that human IgG1 mAbs promote ADCC to a deplete FOXP3þ cells.
greater extent than IgG2 (21), we evaluated the potential differ-
ential effects of ipilimumab (IgG1) and tremelimumab (IgG2) on Presence of tumor-associated macrophages in baseline and
the densities of FOXP3þ cells in tumor tissues from patients with ipilimumab-treated biopsies
advanced melanoma. Surprisingly, both ipilimumab (Fig. 2A; In humans, ADCC is primarily mediated by FcgRIIIA (CD16a;
Supplementary Fig. S3) and tremelimumab (Fig. 2B) did not ref. 27), which is expressed on macrophages and natural killer (NK)
deplete, but instead significantly increased the density of FOXP3þ cells. The lack of intratumoral Treg depletion following treatment
A B
P = 0.0019 P = 0.0029
Figure 2.
Effect of CTLA-4 blockade on FOXP3þ cells in
Density of FOXP3+ cells/mm2
with ipilimumab could be attributable to a paucity of macrophages established, its effect on human FOXP3þ Tregs within the tumor
or NK cells and/or FcgRIIIA expression within the tumor micro- microenvironment remains unclear. While some studies report a
environment. Therefore, we sought to evaluate the effect of ipili- decrease in the frequency of circulating or intratumoral Tregs (5, 7,
mumab on intratumoral CD68þ macrophages in patients with 9, 29, 30), other studies did not result in Treg depletion following
advanced melanoma. Compared with intratumoral CD4þ and anti-CTLA-4 treatment (11, 25, 26, 31, 32). This study has been
CD8þ cell densities, there was considerable infiltration of CD68þ undertaken to gain a better understanding of the effect of anti-
macrophage in the untreated tumor samples, which did not CTLA-4 therapy on intratumoral FOXP3þ Tregs in human cancers.
significantly change following treatment with ipilimumab (Fig. 4). Preclinical studies have reported that CTLA-4 blockade results
in an increase in the Teff to Treg ratio (5–7). Multiple studies have
been conducted to evaluate Tregs in human tumors following
Discussion anti-CTLA-4 therapy (24–26, 28, 32). However, most of these
Anti–CTLA-4 immunotherapy has demonstrated significant studies relied on the evaluation of intratumoral Tregs by flow
antitumor activity in preclinical and clinical studies (3, 6, 18, cytometry with the calculation of FOXP3þ Tregs as a percentage of
19, 28), and it has emerged as a paradigm shift in cancer therapy. total cells. In previous studies, FOXP3þ Tregs were evaluated in
Although the effect of CTLA-4 blockade on Teff cells has been well bladder tumor tissues pre- and post-ipilimumab treatment. We
A B
P = 0.0267
Density of FOXP3+ cells/mm2
Figure 3.
Effect of CTLA-4 blockade on FOXP3þ cells in
bladder and prostate tumors. Stage-matched
untreated and ipilimumab (Ipi)-treated bladder
tumors (n ¼ 9; A), and prostate tumors (n ¼ 16; B)
were analyzed by IHC for the presence of FOXP3þ
cells. Each plot shows mean with SD, and each
symbol represents an individual patient. Statistical
significance is defined as P < 0.05.
1236 Clin Cancer Res; 25(4) February 15, 2019 Clinical Cancer Research
Anti-CTLA-4 Therapy Does Not Deplete FOXP3þ Cells
Melanoma cells; likewise, staining for FOXP3 alone can represent Teff and
CD68 Treg cells. To address this, we collaborated our IHC results with
multiparametric CyTOF analysis of tumor-infiltrating cells from
paired pre- and post-ipilimumab-treated melanoma tissues. We
5,000 found that greater than 95% of the CD8þ population were T cells
4,500
(CD3þCD8þ); whereas less than 5% of these were NK cells
(CD3CD8þCD56þ). More importantly, ipilimumab significant-
Density of CD68+ cells/mm2
observed an increase in the number of total T cells, but a decrease Disclosure of Potential Conflicts of Interest
in frequency of FOXP3þ Tregs, which was measured as a percent of J. Wargo reports receiving speakers bureau honoraria from Dava, and is
total cells (24, 25). Because the total number of T cells increased in a consultant/advisory board member for Bristol-Myers Squibb, AstraZeneca,
posttreatment tumor tissues, our calculation of FOXP3þ Tregs as a Novartis, Illumina, and Merck. J.P. Allison is an inventor and recipient of
royalty from intellectual property related to activation of ICOS signaling
percent of the total T cells led to data indicating a decrease in for cancer immunotherapy issued to The University of Texas System and
frequency of FOXP3þ Tregs, even though the absolute number of Memorial Sloan Kettering Cancer Center (MSKCC) and licensed to Jounce
FOXP3þ cells may not have changed. Therapeutics; is an inventor and recipient of royalty from intellectual
In this study, we evaluated immune subsets including CD4þ, property related to the use of Newcastle Disease Virus as a biotherapeutic
CD8þ, and FOXP3þ cells in tumor tissues obtained from patients issued to Icahn School of Medicine at Mount Sinai and MSKCC and licensed
to Merck; and is an inventor and recipient of royalty on a now expired
treated with anti-CTLA-4 mAbs by quantitative IHC. We measured
patent for use of antagonism of the CTLA-4 signaling pathway for cancer
the density of T-cell subsets in stage-matched human tumor immunotherapy issued to the University of California and licensed to
tissues and found that anti-CTLA-4 therapy increases the density Bristol-Myers Squibb. P. Sharma is listed as a co-inventor of a patent
of intratumoral CD4þ and CD8þ cells, but without significant regarding the targeting of ICOS for immunotherapy that is owned by Jounce
change or depletion of FOXP3þ cells. In addition, evaluation of Therapeutics, and is a consultant/advisory member for Jounce, Neon
paired pretreatment and post-ipilimumab treatment melanoma Therapeutics, Constellation, Oncolytics Biotech, BioAtla, Fourty-Seven,
tissues revealed higher densities of intratumoral CD8þ cells (Sup- Apricity, Polaris, Marker Therapeutics, Codiak, Pieris Pharmaceuticals, Kite
Pharma, Merck, and BioMx. No potential conflicts of interest were disclosed
plementary Fig. S1B) and FOXP3þ cells (Supplementary Fig. S1C). by the other authors.
Importantly, the timing of biopsy collection post-anti-CTLA-4
therapy (ipilimumab and tremelimumab) did not impact the
infiltration of immune subsets (Supplementary Fig. S4). Consis- Authors' Contributions
Conception and design: J.P. Allison, A. Ribas, P. Sharma
tent with this, an IHC study evaluating melanoma tumor samples
Development of methodology: A. Sharma, S.K. Subudhi, J. Scutti, P. Sharma
(n ¼ 19) demonstrated higher infiltrates of FOXP3þ cells in the Acquisition of data (provided animals, acquired and managed patients,
responding lesions following treatment with tremelimumab (26). provided facilities, etc.): J. Scutti, L. Vence, A. Ribas
Taken together, these results indicate that IHC may be a more Analysis and interpretation of data (e.g., statistical analysis, biostatistics,
reliable tool to examine treatment-related changes in immune computational analysis): A. Sharma, S.K. Subudhi, J. Blando, J. Scutti, L. Vence,
subsets within the tumor, as it provides a more direct measure of J. Wargo, J.P. Allison, A. Ribas, P. Sharma
Writing, review, and/or revision of the manuscript: A. Sharma, S.K. Subudhi,
the absolute numbers of cells in a given surface area.
J. Blando, J. Wargo, J.P. Allison, A. Ribas, P. Sharma
The limitation of single-stain IHC is that an individual marker Administrative, technical, or material support (i.e., reporting or organizing
may be expressed on more than one cellular population. For data, constructing databases): A. Sharma
example, single staining for CD8 can represent both T cells and NK Study supervision: P. Sharma
References
1. Walunas TL, Lenschow DJ, Bakker CY, Linsley PS, Freeman GJ, Green JM, 18. Hodi FS, O'Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, et al.
et al. CTLA-4 can function as a negative regulator of T cell activation. Improved survival with ipilimumab in patients with metastatic melanoma.
Immunity 1994;1:405–13. N Engl J Med 2010;363:711–23.
1238 Clin Cancer Res; 25(4) February 15, 2019 Clinical Cancer Research