Molecular Imaging and Targeting of Hypoxic Microenvironment
Balaji Krishnamachary1, Louis Dore-Savard1, Santosh Kumar Bharti1, Flonne Wildes1, Yelena Mironchik1, and Zaver M Bhujwalla1

1Radiology, Johns Hopkins University, Baltimore, MD, United States

Synopsis

Hypoxic environments frequently exist in solid tumors and result in resistance to therapy and the evolution of a more lethal phenotype. Here, we have genetically engineered human prostate cancer PC-3 cells to report on hypoxia and also express yeast cytosine deaminase under the control of hypoxia response elements to convert the prodrug 5-fluorocytosine to 5-Fluorouracil. We also show that selective killing of hypoxic cells significantly reduces tumor growth.

Introduction

Cancer cells display an adaptive response to hypoxia through the activation of several genes mediated by the binding of hypoxia inducible factors (HIFs) to hypoxia response elements (HRE) in the promoter region of targeted gene that results in their increased transcription [1]. HIFs have been reported to promote key steps in tumorigenesis, including angiogenesis, metabolism, proliferation, metastasis, and differentiation [1]. HRE-driven imaging reporter systems report on HIF expression, since the binding of HIF to the HRE drives the expression of an imaging reporter such as luciferase [2]. The bacterial and yeast enzyme cytosine deaminase (CD) converts the nontoxic prodrug 5-FC to the anti-cancer drug 5-FU that is widely used in the treatment of a range of cancers [3]. Controlling the expression of CD by HRE, in cells that also report on HIF-1a expression with luciferase expression provides the ability to detect hypoxia and generate 5-FU from CD directly within cells only when HIF-1a is present. Here we have established a human prostate cancer cell line, PC-3, that reports on hypoxia with bioluminescence imaging (BLI) and expresses yCD under hypoxia to further understand the effects of targeting hypoxia on aggressive tumor populations and their elimination.

Methods

A plasmid expressing the luciferase gene under the control of five tandem repeats of hypoxia response elements (5X-HRE) and yCD were sub cloned into a lentiviral vector (HRE-Luc, HRE-yCD). Virion generation was accomplished as previously published [3]. Virions were used to transduce PC-3 cells obtained from ATTCC. PC-3 cells expressed HRE-Luc (PC-3 Luc) or HRE-yCD (PC-3-yCD) or both (PC-3-yCD+Luc). Transduction efficiency and reporter activity in response to hypoxia was evaluated by performing luciferase assays, and BLI of cells in vitro or in vivo using a Xenogen IVIS Spectrum system. Cell viability in vitro in response to 48h of hypoxia followed by 24h of 5-FC treatment in genetically engineered PC-3 cells was assessed by modified MTS assay. In vivo studies were performed by inoculating 2x106 engineered PC-3 cells on either side of the flank of 5-week-old male severely combined immune deficient (SCID) mice. BLI was performed once tumors reached ~200mm3 followed by 5-FC injection through the tail vein (200mg/kg) and intraperitoneally (250mg/kg). BLI was performed 3 days after the first 5-FC injection and continued through the treatment protocol. At the end of the treatment protocol, tumors were excised, and a part of the tumor was processed for immunohistochemistry and the other for high-resolution 19F MR spectroscopy.

Results and Discussion

As shown in Fig. 1B, bioluminescence was detected in both PC3-HRE-Luc and PC-3-HRE-yCD+Luc cells in response to the hypoxia mimetic cobalt chloride but not under normoxia, confirming the regulation of luciferase by hypoxia. Both cell lines also showed a robust increase in the enzyme activity in response to 1% O2. To validate yCD gene activity and its ability to convert the prodrug 5-FC to 5-FU, we performed a cell viability assay. As shown in Fig. 2, a statistically significant increase of cell kill was observed in both PC-3-HRE-yCD and PC-3-HRE-yCD+Luc cells that were exposed to hypoxia followed by treatment with 1mM of 5-FC. There was no cell kill in these cells when maintained under normoxia following treatment with 5-FC. These engineered cells were further tested in vivo to evaluate the ability of HRE-driven luciferase to report on hypoxia and HRE-driven-yCD to selectively kill hypoxic cells. Representative BLI images following treatment shown in Fig. 3A demonstrate reduction of hypoxic regions and tumor volume in the tumor with HRE-driven yCD. A significant reduction of tumor volume was observed by day 11 as shown in Fig. 3B. Also evident is extensive necrosis in the tumor with HRE-driven-yCD (Fig. 3C). The feasibility of detecting the activity of yCD from a conversion of 5-FC to 5-FU with 19F MRS is shown in Fig. 3D [4]. We are currently evaluating the effects of eliminating hypoxic cancer cells on distant metastasis as well as on aggressive subpopulations such as cancer stem cells in the primary tumor. We are also evaluating the formation of 5-FU metabolites in the tumor using high-resolution 19F MRS of extracts.

Acknowledgements

This work was supported by NIH R01CA136576 and P50 CA103175. We thank Mr. Gary Cromwell for valuable technical assistance.

References

[1] Philip, B., et al., Carcinogenesis, 2013. 34(8): p. 1699-707. [2] Ma, Y., et al., PLoS One, 2011. 6(12): p. e29170, [3] Krishnamachary B. et al., PLoS One, 2012; [4] Chen Z, et al., ACS Nano. 2012 6(9):7752-62.

Figures

A) Line drawing showing elements that constitute lentiviral plasmid expressing luciferase (a) and yCD gene (b) under the control of HRE. B) Representative images showing bioluminescence in PC-3 cells expressing HRE driven constructs in response to CoCl2 and 5-FC drug treatment. C) Luciferase assay measured by quantifying the activity of the enzyme in various PC-3 cell lines in response to 1% O2 for 48 hours. Values represents mean ±SEM of three independent experiments. ***p<0.0005.


Cell viability following treatment with 1mM of 5-fluorocytosine (5-FC) for 24h in genetically engineered PC-3 cells that were maintained under normoxia or subject to 48h of 1% O2.Values represent mean ±SEM of four independent experiments. *p<0.05, ***p<0.0005.

A) Representative BLI images showing larger hypoxic area in control (left) and reduced hypoxic area together with a reduction of tumor volume in a yCD tumor (right). B) Average tumor volume of Luc and yCD+Luc tumors following 3 injections of 5-FC administered at 3-4 day intervals. C) Representative H&E staining of control and yCD tumors with viable and non-viable areas. D) Representative 19F magnetic resonance spectra shows the feasibility of detecting the conversion of 5-FC to 5-FU by CD [4].



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2495