Evaluation of anti-PD-1-based therapy against triple-negative breast cancer patient-derived xenograft tumors engrafted in humanized mouse models

Abstract

Background: Breast cancer has been considered not highly immunogenic, and few patients benefit from current immunotherapies. However, new strategies are aimed at changing this paradigm. In the present study, we examined the in vivo activity of a humanized anti-programmed cell death protein 1 (anti-PD-1) antibody against triple-negative breast cancer (TNBC) patient-derived xenograft (PDX) tumor models.

Methods: To circumvent some of the limitations posed by the lack of appropriate animal models in preclinical studies of immunotherapies, partially human leukocyte antigen-matched TNBC PDX tumor lines from our collection, as well as human melanoma cell lines, were engrafted in humanized nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (hNSG) mice obtained by intravenous injection of CD34⁺ hematopoietic stem cells into nonlethally irradiated 3-4-week-old mice. After both PDXs and melanoma cell xenografts reached ~ 150-200 mm³, animals were treated with humanized anti-PD-1 antibody or anti-CTLA-4 and evaluated for tumor growth, survival, and potential mechanism of action.

Results: Human CD45⁺, CD20⁺, CD3⁺, CD8⁺, CD56⁺, CD68⁺, and CD33⁺ cells were readily identified in blood, spleen, and bone marrow collected from hNSG, as well as human cytokines in blood and engrafted tumors. Engraftment of TNBC PDXs in hNSG was high (~ 85%), although they grew at a slightly slower pace and conserved their ability to generate lung metastasis. Human CD45⁺ cells were detectable in hNSG-harbored PDXs, and consistent with clinical observations, anti-PD-1 antibody therapy resulted in both a significant reduction in tumor growth and increased survival in some of the hNSG PDX tumor lines, whereas no such effects were observed in the corresponding non-hNSG models.

Conclusions: This study provides evidence associated with anti-PD-1 immunotherapy against TNBC tumors supporting the use of TNBC PDXs in humanized mice as a model to overcome some of the technical difficulties associated with the preclinical investigation of immune-based therapies.

Keywords: Anti-PD-1; Humanized mouse model; Immunotherapy; PD-L1; TNBC; Triple-negative breast cancer.

Fig. 1

Analysis of human immune cell engraftment. a Evolution of the percentage of human CD45⁺ cells after intravenous (i.v.) injection of hCD34⁺ hematopoietic stem cells. Cells were identified by flow cytometry in circulating blood collected from humanized mice at the indicated time intervals (n = 8). b Analysis of hCD45⁺ and corresponding subpopulations, including hCD20⁺ (B cells), hCD3⁺ (T cells), hCD33⁺ (myeloid lineage), hCD56⁺ (natural killer [NK] cells), and hCD68⁺ (macrophages) cells, was determined by flow cytometry in blood, bone marrow, and spleen samples collected from humanized nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (hNSG) mice after 22 weeks of i.v. injection of human hematopoietic stem cells (n = 8). c Representative IHC analysis of human CD45⁺, CD20⁺, CD68⁺, and CD56⁺ cells performed in preparations of spleen from humanized (upper row) and nonhumanized (lower row) NSG mice. Counterstain, hematoxylin; magnifications, 20× and 4× (inset)

Fig. 2

In vivo effects of humanization of nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (NSG) mice in the growth and engraftment of triple-negative breast cancer (TNBC) patient-derived xenograft (PDX) tumor line MC1 (a) and human melanoma A375 cell line (b). Both humanized and nonhumanized female NSG mice (n = 10 in each group) were transplanted orthotopically with pieces of either the PDX tumor line MC1 (into the cleared mammary fat pad) or A375 cells (into the skin) and allowed to grow. Tumor volume was determined twice weekly. NS Nonsignificant; *P < 0.05, *** P < 0.001. c Flow cytometric analysis of human CD45⁺ cells and hCD20⁺ (B cells), hCD3⁺ (T cells), hCD33⁺ (myeloid lineage), hCD56⁺ (natural killer [NK] cells) and hCD68⁺ (macrophages) cell subpopulations determined in blood, spleen, bone marrow, and MC1 PDX tumors of the corresponding samples shown in (a) (n = 10)

Fig. 3

IHC analysis of human CD45⁺, CD20⁺, CD68⁺, CD56⁺, CD4⁺, and CD8⁺ cells and cells present in MC1 tumor xenografts. Representative images (from a total of 8–10 processed samples in each group) of IHC performed in preparations of MC1 tumor samples grown in either humanized or nonhumanized nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (NSG) mice corresponding to samples shown in Fig. 2a or c, respectively. 4× (inset) and 20× magnifications are shown; counterstain, hematoxylin.

Fig. 4

Analysis of breast cancer lung metastasis in humanized nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (hNSG) patient-derived xenograft (PDX). IHC analysis of human Ki-67, cytokeratin 19, and CD45⁺ expression in primary (breast) and metastatic (lung) triple-negative breast cancer PDX tumor lines BCM-2147, MC1, and BCM-4913 engrafted in hNSG mice. Amplifications, 4× and 20×; counterstain, hematoxylin

Fig. 5

Analysis of programmed death ligand 1 (PD-L1) protein expression in patient-derived xenograft (PDX) tumor samples engrafted in both nonhumanized and humanized nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (hNSG) mice performed by Western blotting (a, MC1) or IHC (b, upper panels, MC1; lower panels, BCM-4913). In Western blotting experiments, samples were blotted with an anti-β-actin antibody as a loading control. The blots were processed in parallel, and they were all sourced from the same experiment. c Comparative analysis of PD-L1 levels was performed using four different PDX tumor lines (MC1, BCM-4913, BCM-4664, BCM-5471) engrafted in hNSG mice. Three independent tumors (animals) of each PDX line were evaluated by Western blot analysis. Samples were blotted with an anti-β-actin antibody as a loading control. d PD-L1 analysis performed by IHC of BCM-4664 and BCM-5471 PDXs engrafted in hNSG mice. 4× magnifications are shown; counterstain, hematoxylin

Fig. 6

Response of triple-negative breast cancer (TNBC) patient-derived xenografts (PDXs) to the anti-programmed cell death protein 1 (anti-PD-1) therapy. a In vivo treatment with anti-PD-1 antibody (10 mg/kg intravenous [i.v.] once weekly) of either TNBC MC1 PDX-engrafted nonhumanized (left graph, n = 5) or humanized (right graph, n = 5) nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (hNSG) mice. Tumor volume was measured twice weekly. b Kaplan-Meier analysis of median survival of mice treated with vehicle (n = 6) vs. anti-PD-1 antibody (n = 6). c hNSG mice engrafted with an additional TNBC BCM-4913 PDX tumor line were treated with either vehicle control or anti-PD-1 antibody (10 mg/kg i.v. once weekly). Tumor volumes were measured twice weekly. d In vivo treatment with anti-PD-1 antibody (10 mg/kg i.v. once weekly) of TNBC BCM-4664 (n = 5) and HM-3818 (n = 5) PDXs engrafted in hNSG mice. Tumor volume was measured twice weekly. e Analysis of tumor-infiltrating lymphocyte (TIL) cytotoxic activity. TILs isolated by Ficoll gradient from vehicle- or anti-PD-1 antibody-treated MC1 PDX tumors engrafted in hNSG mice were cocultured with disaggregated MC1 tumor cells obtained from the corresponding PDX grown in nonhumanized NSG mice. Cytotoxic activity was measured using the CytoTox 96® Non-Radioactive Cytotoxicity Assay as per the manufacturer’s instructions. f Levels of granzyme B tumor were measured by incubating tumor protein lysates with antibody-immobilized magnetic beads and evaluated using a Luminex LX200 Multiplexing Assay System. **P < 0.01, ***P < 0.001. NS Nonsignificant

Fig. 7

Analysis of A375 melanoma cell line xenograft growth. Human melanoma cells (A375; 5 × 10⁵) were injected orthotopically into the skin of both nonhumanized nonobese diabetic/severe combined immunodeficiency IL2Rγ^null (NSG) and humanized NSG (hNSG) mice, after which (initial tumor volume 150–200 mm³) they were randomly sorted into treatment groups. Non-hNSG mice (a) or hNSG mice (b and c) were treated weekly with vehicle (control), anti-CTL4 (2.5/5 mg/kg (b), or anti-PD-1 (10 mg/kg) (c) antibodies. Tumor growth was evaluated twice weekly. If tumor volume reached 1500–2000 mm³, mice were killed as per humane animal welfare regulations. *P < 0.05, **P < 0.01, *** P < 0.001. NS Nonsignificant