The Use of a Humanized NSG-β2m-/- Model for Investigation of Immune and Anti-tumor Effects Mediated by the Bifunctional Immunotherapeutic Bintrafusp Alfa

Abstract

The lack of serial biopsies in patients with a range of carcinomas has been one obstacle in our understanding of the mechanism of action of immuno-oncology agents as well as the elucidation of mechanisms of resistance to these novel therapeutics. While much information can be obtained from studies conducted with syngeneic mouse models, these models have limitations, including that both tumor and immune cells being targeted are murine and that many of the immuno-oncology agents being evaluated are human proteins, and thus multiple administrations are hampered by host xenogeneic responses. Some of these limitations are being overcome by the use of humanized mouse models where human peripheral blood mononuclear cells (PBMC) are engrafted into immunosuppressed mouse strains. Bintrafusp alfa (M7824) is an innovative first-in-class bifunctional fusion protein composed of the extracellular domain of the TGF-βRII to function as a TGF-β "trap" fused to a human IgG1 antibody blocking PD-L1. A phase I clinical trial of bintrafusp alfa showed promising anti-tumor efficacy in heavily pretreated advanced solid tumors, and multiple clinical studies are currently ongoing. There is still much to learn regarding the mechanism of action of bintrafusp alfa, including its effects on both human immune cells in the periphery and in the tumor microenvironment (TME), and any temporal effects upon multiple administrations. By using the NSG-β2m^-/- mouse strain humanized with PBMC, we demonstrate here for the first time: (a) the effects of bintrafusp alfa administration on human immune cells in the periphery vs. the TME using three different human xenograft models; (b) temporal effects upon multiple administrations of bintrafusp alfa; (c) phenotypic changes induced in the TME, and (d) variations observed in the use of multiple different PBMC donors. Also discussed are the similarities and differences in the data thus far obtained employing murine syngeneic models, from clinical trials, and in the use of this humanized mouse model. The results described here may guide the future use of this agent or similar immunotherapy agents as monotherapies or in combination therapy studies.

Keywords: NSG-β2m−/−; TGF-β; bintrafusp alfa; humanized mouse; immunotherapy.

Figure 1

Characterization of the PBMC humanized NSG-β2m^−/− mouse model. 1 ×10⁷ human healthy donor PBMCs were injected i.p. into NSG-β2m^−/− mice at day 0. (A) Animals were bled weekly with circulating lymphocytes isolated and immunostained for human CD3. Graph depicts frequency of human CD3⁺ cells of total live circulating cells at weeks 2–7 post-PBMC engraftment (n = 19; week 7, n = 10; data reported as mean ± SEM). (B) Histograms from animals representing the frequency of circulating human T cells at weeks 2, 3, 4, and 5. (C) Percent change in humanized mouse body weight from the time of initial PBMC engraftment (n = 50); vertical blue line indicates the timepoint prior to the development of overt GvHD, red line denotes mean change in animal weight. (D) Parallel changes in body weight vs. relative proteinuria. (E) PBMC donor-dependent difference in the reduction of body weight post-PBMC engraftment.

Figure 2

Bintrafusp alfa treatment results in anti-tumor efficacy in the HTB-1 subcutaneous bladder model. (A) Timeline of tumor injection, human PBMC engraftment, and bintrafusp alfa treatments. (B) Titration of the number of human PBMCs and their effect on bintrafusp alfa-mediated anti-tumor efficacy on day 28 post-initiation of bintrafusp alfa treatment (n > 4; *p < 0.05, student's t-test, error bars represent mean ± SEM). (C) Average growth curves of HTB-1 s.c. tumors in animals treated with bintrafusp alfa in the absence of human PBMCs (red) or in the presence of human PBMCs without bintrafusp alfa (black) (n ≥ 7). (D) Average growth curves of HTB-1 s.c. tumors in control-Ig and bintrafusp alfa-treated mice (data combined from 3 independent experiments, n ≥ 26, 2-way ANOVA, ***p < 0.0001; red arrows indicate bintrafusp alfa administration). HTB-1 tumor T-cell infiltrates for both CD4⁺ (E) and CD8⁺ (F) T cells per milligram tumor weight at day 21 (left, n = 9), or after day 28 (right, n = 22) post-initiation of bintrafusp alfa treatment. Error bars represent mean ± SEM; *p < 0.05. (G) Representative immunofluorescent images at day 28 post-initiation of bintrafusp alfa treatment, comparing CD8⁺ expression (red) in HTB-1 tumors from IgG control vs. bintrafusp alfa-treated mice.

Figure 3

Bintrafusp alfa treatment is associated with increased T-cell infiltration/activation in the TME. Mice harboring HTB-1 subcutaneous tumors were treated with bintrafusp alfa or an IgG1 isotype control as outlined in Figure 2A. Number of IFNγ-producing CD4⁺ (A) and CD8⁺ (B) T cells per milligram tumor weight and frequency of (C) CD8⁺ NKG2D⁺ T cells and (D) IFNγ/GzmB⁺/NKG2D⁺ of total CD8⁺ tumor infiltrating cells at day 21 post-initiation of bintrafusp alfa treatment (*p < 0.05, student's t-test).

Figure 4

Robust peripheral and intratumoral changes associated with the administration of bintrafusp alfa in the humanized HTB-1 subcutaneous bladder model. (A) Timeline of PBMC engraftment, blood collection, and treatment for subcutaneous HTB-1 tumor-bearing PBMC humanized NSG-β2m^−/− mice. (B) Total serum TGF-β1 levels at 30 min and 6 days post first and second bintrafusp alfa injection (n = 6). (C) Active TGF-β1 levels as a percent of total protein in HTB-1 tumor lysates from control-IgG1 or bintrafusp alfa-treated mice at day 21 post-initiation of bintrafusp alfa treatment (n = 4; *p < 0.05). (D) In vivo localization of AF647-labeled bintrafusp alfa to HTB-1 tumors 2 weeks post i.p. injection.

Figure 5

Bintrafusp alfa confers a robust anti-tumor effect in the HPV⁺ SiHa cervical and MDA-MB-231 triple negative breast cancer models. (A) Average SiHa tumor volumes from control (black) or bintrafusp alfa-treated (red) PBMC humanized mice (n = 21, 2-tailed ANOVA, ***p < 0.0001, data combined from 2 independent experiments). (B,C) Tumor infiltrating CD4⁺ (B) and CD8⁺ (C) IFNγ-producing T cells per milligram tumor weight at day 21 post-initiation of bintrafusp alfa treatment (n = 9, *p < 0.05, student's t-test). (D) Representative dot plots from the tumor of control (left) and bintrafusp alfa-treated (right) mice bearing SiHa tumors. IFNγ is represented on the x-axis and granzyme B on the y-axis. (E) Waterfall plot of the percent change in SiHa tumor volume from the initiation of treatment in one independent experiment at day 28 post-initiation of bintrafusp alfa treatment. (F) Average MDA-MB-231 tumor volumes from control-IgG1 (black) or bintrafusp alfa-treated (red) PBMC humanized mice (n = 9, 2-tailed ANOVA, ***p < 0.0001, data combined from 2 independent experiments). (G) Immune activation score derived from gene expression profiles (as described in Figure 6) from HTB-1 and SiHa tumors from control and bintrafusp alfa-treated mice (n = 3, *p < 0.05, **p < 0.01, student's t-test).

Figure 6

Distinct transcriptional changes within the TME following bintrafusp alfa treatment. (A,B) Heatmap RNA expression profile from a NanoString pan-cancer immune profiling panel assayed on enriched CD45⁺ cells from tumors of control and bintrafusp alfa-treated mice bearing HTB-1 tumors at day 21 post-initiation of bintrafusp alfa treatment (A) or SiHa tumors at day 22 post-initiation of bintrafusp alfa treatment (B) (fold change >1.5, ANOVA p < 0.1). (C) Ingenuity pathway analysis of HTB-1 and SiHa RNA expression profiles describing those pathways most altered with bintrafusp alfa treatment; (*) denotes pathways altered in both HTB-1 and SiHa tumors; (X) denotes pathways altered in SiHa tumors only. (D) String analysis of those genes altered in the top 5 pathways identified by Ingenuity Pathway Analysis. Red nodes denote those genes/pathways identified post-bintrafusp alfa treatment in both HTB-1 and SiHa tumors; blue nodes denote those genes/pathways identified post-bintrafusp alfa treatment in SiHa tumors only.

Figure 7

Analyses of peripheral and intratumoral CD8⁺ T-cell PD1 expression following bintrafusp alfa administration. (A) Association between PD1 expression in splenic (A) or intratumoral (B) CD8⁺ T cells vs. tumor weight at day 21 post-initiation of bintrafusp alfa treatment in bintrafusp alfa-treated PBMC humanized mice. (C) Splenic or (D) tumor CD8⁺ PD1 expression in animals responding to bintrafusp alfa (tumor weight <263 mg) vs. those not responding (tumor weight >263 mg) in HTB-1 humanized mice (***p < 0.0001, student's t-test).