We developed a novel cancer screening strategy whereby tumors are forced to secrete a biomarker. We developed a conditionally replicative HSV, rQ-M38G, to selectively force tumor cells to secrete Gaussia luciferase as a demonstration of this screening strategy. Tumor-bearing animals given intravenous rQ-M38G expressed higher levels of biomarker compared with tumor-free animals, which showed only low, background expression. The utility of rQM-38G for screening appeared to be a function of the capability of a given tumor to support HSV replication. Regardless of tumor type or location, tumor screening by rQ-M38G was highly sensitive to tumor presence.
A critical feature for cancer screening is the ability to detect small tumors, ultimately in people. One a priori
concern was that small tumors may be associated with less vascular surface area available for HSV entry. This issue did not appear to be a limitation in mice, because in some of the models we were able to detect tumors as small as 4–5 mm in diameter (50 mm3
) and in another model we detected microscopic tumor burden. Scalability to humans (with larger blood volumes) is difficult to predict, but some estimates are possible. In vitro
evaluation of increasing numbers of SKNEP-Luc tumor cells indicates that infecting as few as 1,000 cells in a volume of cell culture media equal to a mouse blood volume yields detectable GLuc secretion. Therefore, simultaneous expression of GLuc from 1,000 cells at any time in a mouse is theoretically detectable. Under these conditions infecting 10% of the cells in a tumor of diameter 400 µm (104
cells) would be detectable in a mouse. When scaled from a mouse to a human the theoretical limit of detection is increased by a ratio of 1
2600 (25 mg mouse
65 kg human) and diluting the biomarker in a larger average human blood volume of 4.7 L. Using these numbers, the theoretical limit of detection when infecting 10% of the cells in a tumor is changed from a 394 um diameter mass in a mouse to a 1–4 mm tumor diameter in a human. If only 1% of tumor cells are transduced by virus, tumors as small as 8.5 mm in diameter might still be detectable using these calculations. This sensitivity suggests strong potential for identifying minimal tumor burdens even when scaled to human proportions.
The ability of this screening strategy to reveal tumors is dependent upon viral factors, tumor microenvironment, and biomarker detection limitations each of which can be enhanced for real world practicality. Here we employed a doubly attenuated HSV to effect a tumor specific transduction and gene expression. Such vectors have already been documented to be safe for human use 
. Alternative viruses with single or less-attenuating mutations demonstrate greater oncolytic capacity are also already in clinical trials (e.g., HSV1716, see www.clinicaltrials.gov
, NCT00931931). Coupling this screening strategy to less attenuated viruses will likely diversify its utility among the various solid malignancies as well as enhance the therapeutic component. Agents that are simultaneously both diagnostic and therapeutic have been dubbed “theragnostic.” It is possible our strategy could be further refined by the use of a more robust cancer-dependent promoter to drive biomarker expression. Improved vector delivery to the tumor may also be achieved using tumor targeting small molecule and nanoparticles uptake-enhancement strategies as recently described with the use of internalizing RGD peptides 
. Eventual commercialization of our proposed screening strategy would benefit from using biomarkers that are already in widespread use such as βHCG (pregnancy test), PSA (prostate cancer) and αFP (liver and germ cell malignancies). Biomarker assay sensitivity in the context of this screening strategy can still improve exponentially as is being realized with microfluidic technologies 
A significant concern for the prototypical cancer screening method developed in this work is the development of an immune response to the gene delivery agent or the transduced biomarker that would preclude repeated use of the screening tool. Previous work in the field has shown that pre-immunized animals still experience therapeutic benefit from HSV with sustained efficacy 
. Because 80% of the general population has been exposed to HSV, implementing this detection strategy with HSV will depend upon the efficacy of engineered HSVs administered to immunocompetent subjects who have previously been exposed to wild-type HSV strains. In vitro
screening for mouse tumor cell lines revealed that all mouse lines we tested exhibited comparably low susceptibility to infection by our attenuated mutant virus, on par with normal quiescent human cells. In vivo models of HGF116 demonstrated no virus sensitivity following i.t. or i.v. injection. Thus, we are unable to assess our cancer screening strategy in an immunocompetent model until the identification of mouse models that are more susceptible to human HSV. Implementation of this strategy may evolve with non-viral vectors 
or viral vectors masked in liposomes 
The challenge of population-wide cancer screening is the development of clinical assays that are fiscally practical, universal across a spectrum of cancer types, and easy to implement. Physical examination, while affordable, often fails to identify malignancies that are deep or asymptomatic. Imaging benefits from high sensitivity, but may not always differentiate nonspecific or benign masses from malignant disease (eg, lung nodules may be granuloma or cancer), is financially challenging and difficult to broadly implement. With the identification of appropriate biomarkers, cancer screening could potentially be economically viable with automated point of care testing (POCT) technologies (www.i-stat.com
, www.siloambio.com 
This work demonstrates the principle of inducing expression of a secretable transgene in cancer using a systemically administered gene delivery agent as a biomarker to screen for tumor presence. The novel screening principle proposed and demonstrated in this work could hold immediate implications for patients with known cancer risks such as genetic predispositions (BRCA-1, Nf1 mutations) or patients with a history of cancer who are at high risk for recurrence. Ultimately, exogenous administration of a cancer targeting gene delivery agent (infectious or not) could force any malignancy to secrete a biomarker and could be used as a universal first step for population-wide cancer screening. The impact of this approach would revolutionize the technology for cancer detection in industrialized nations and the developing world where imaging and biopsy-based diagnostics may not be as readily available.