Lifetime exposure to a soluble TGF-beta antagonist protects mice against metastasis without adverse side effects

Abstract

TGF-betas play diverse and complex roles in many biological processes. In tumorigenesis, they can function either as tumor suppressors or as pro-oncogenic factors, depending on the stage of the disease. We have developed transgenic mice expressing a TGF-beta antagonist of the soluble type II TGF-beta receptor:Fc fusion protein class, under the regulation of the mammary-selective MMTV-LTR promoter/enhancer. Biologically significant levels of antagonist were detectable in the serum and most tissues of this mouse line. The mice were resistant to the development of metastases at multiple organ sites when compared with wild-type controls, both in a tail vein metastasis assay using isogenic melanoma cells and in crosses with the MMTV-neu transgenic mouse model of metastatic breast cancer. Importantly, metastasis from endogenous mammary tumors was suppressed without any enhancement of primary tumorigenesis. Furthermore, aged transgenic mice did not exhibit the severe pathology characteristic of TGF-beta null mice, despite lifetime exposure to the antagonist. The data suggest that in vivo the antagonist may selectively neutralize the undesirable TGF-beta associated with metastasis, while sparing the regulatory roles of TGF-betas in normal tissues. Thus this soluble TGF-beta antagonist has potential for long-term clinical use in the prevention of metastasis.

Figure 1

Antagonist design and in vitro validation. (a) Soluble TGF-β antagonist SR2F. This antagonist consists of the extracellular domain of the human type II TGF-β receptor fused to the Fc domain of human IgG₁. It can bind TGF-β1 and TGF-β3, but not TGF-β2. (b) Transgene construct. Transgene expression is driven by the mammary-selective MMTV-LTR promoter/enhancer. Small introns are present in the Fc domain of SR2F and the SV40 3′ UTR to enhance transgene expression in vivo. (c) Reversal of the growth-inhibiting effects of TGF-β1 by transfected SR2F. MDA MB435 cells stably transfected with empty vector (pcDNA3), membrane-bound dominant-negative type II TGF-β receptor (DNR), or SR2F were assayed for their resistance to the growth-inhibiting effects of added TGF-β1 (5 ng/ml). Cell proliferation is normalized to the “no TGF-β” condition for each construct. Results are the mean ± SD of three determinations. (d) Determination of the molar ratio of purified SR2F to TGF-β1 required for neutralization of TGF-β activity. The ability of increasing amounts of purified SR2F to reverse the growth-inhibiting activity of 2 pM TGF-β1 on Mv1Lu cells was determined. Results are the mean ± SD for three determinations.

Figure 2

TGF-β antagonist expression and function in the MMTV-SR2F mouse. (a) Expression of SR2F mRNA in different tissues. RNA was prepared from an adult virgin female transgenic mouse and probed for SR2F. (b) Cell type–specific expression of SR2F in the mammary gland. In situ hybridization showed specific expression of SR2F mRNA in the mammary ductal epithelial cells (ECs) but not in the mammary fat pad (FP) or blood vessels (BVs) of a virgin transgenic mouse. (c) SR2F protein in the transgenic mammary gland. Western blots of mammary gland extracts from wild-type or transgenic (TG) mice were probed with anti–human Fc antibody (anti-huFc), with or without an excess of human immunoglobulin G1 (huIgG) to demonstrate specificity. Purified SR2F was the positive control. (d) SR2F protein levels in serum and tissues from transgenic mice. Sera and tissue extracts from 2.5-month-old virgin transgenic mice were assayed for SR2F protein. Results are the mean ± SD for three mice of each sex. (e) In vivo stability of SR2F. Serum levels of SR2F in the wild-type offspring of hemizygous transgenic dams were determined various times after birth. The serum half-life (T_1/2) of SR2F was calculated from the decay curve. (f) Ability of SR2F in transgenic serum to bind TGF-β1. Serum from wild-type or transgenic mice was probed for TGF-β binding proteins by ligand affinity crosslinking with ¹²⁵I–TGF-β1. Purified SR2F spiked into saline or wild-type serum were the positive controls. Mam gl., mammary gland; Sal. Gl., salivary gland; Mes. LN, mesenteric lymph node ; Sal. gl, salivary gland; Sm. int., small intestine; SV, Seminal vesicle; MWM, molecular weight markers; α₂M, α₂-macroglobulin.

Figure 3

Effect of SR2F on metastatic efficiency in a tail-vein injection assay. (a) Macroscopically detectable liver metastases at different times after inoculation. Wild-type or transgenic (SR2F) mice were injected in the tail vein with 10⁶ isogenic 37-32 melanoma cells. After 21, 28, and 35 days, mice were necropsied and the number of macroscopically visible metastases on the liver surface was quantified. The 21-day timepoint had five mice per genotype group, while the remaining timepoints had ten mice per genotype group. (b) Effects on metastatic efficiency in different internal organs. The number of histologically evident metastases was quantified in all tissues that showed evidence of gross metastases in mice necropsied at 35 days after inoculation. Two representative cross sections of each lobe were assessed for the liver and a single representative cross section was analyzed for other organs. Ten mice were used per genotype group. (c) Dose dependency of SR2F effect on metastatic efficiency. In a replicate experiment of the one described in a, the number of histologically detectable metastases in representative cross sections of the liver were quantitated and plotted against the circulating SR2F levels, as measured by sandwich ELISA on serum prepared at necropsy. The data for the wild-type cohort (0 ng/ml SR2F) are presented as mean ± SD (n = 4). R represents the correlation coefficient for the linear curve fit.

Figure 4

Effects of SR2F on primary tumorigenesis and metastasis in the MMTV-neu transgenic model of metastatic mammary cancer. MMTV-neu mice were crossed with FVB/NCr control mice or MMTV-SR2F mice to generate cohorts that either expressed neu alone (neu) or were bigenic for the neu and SR2F transgenes (neu/SR2F). Mice were cycled through one round of pregnancy and then monitored for development of tumors. (a) Tumor latency. Latency was determined as the time to first appearance of a palpable mammary tumor. The mean latency was 34 weeks for the neu group and 32.5 weeks for the bigenic neu/SR2F group (not statistically different). (b) Metastasis. Lungs from mice in both genotype groups were examined histologically for the presence of metastases.

Figure 5

Effects of prolonged exposure to SR2F on memory T cell phenotype. The percentage of CD4⁺ T cells with a memory T cell phenotype (CD4⁺CD44^highCD62L^low) was determined by FACS analysis of spleens from wild-type and MMTV-SR2F transgenic mice at different ages. TGF-β1 null (TGF-β1^–/–) and age- and strain-matched wild-type (TGF-β1^+/+) mice are shown for comparison. The TGF-β1 null mice do not survive beyond about 3 weeks of age.