Aim 3 To develop therapies that inhibit NANOG and NANOGP8 expression in vivo Lentivirus is commonly used for the introduction of shRNA into cells in vitro. Our preliminary data suggested that lentivirus (LV) might be a useful vector to deliver shRNA to inhibit gene expression in established tumors. However, we found that the use of LV as an in vivo gene therapy appears not to be practical for this nonreplicating virus. Intralesional treatment of 3 - 6 mm subcutaneous nodules of CX-1 in NOD/SCID mice with a 5 multiplicity of Infection (MOI) of LV decreased tumor growth by 50% for up to 8 days after injection. This is likely to be the maximum clinical dose. After 8 days there was rapid rebound of tumor growth. When mice were sacrificed 3 days later, the tumors in the shRNA groups were similar in volume and weight to the controls. NANOG total transcript levels in these tumors paralleled the growth response in that NANOG levels were low 2 days after tumor injection compared to tumors injected with control shRNA but levels were the same as controls at the time of sacrifice. Thus, LV shRNA to NANOGP8 or NANOG (since both are present in both tumors and inhibiting one seems to inhibit the other because both promoters contain NANOG binding sites) controls tumor growth in transduced cells but non-transduced cells eventually grow, leading to progressive tumor growth after single injection. As a result, we created replicating oncolytic adenoviruses that express shRNA. Adenovirus type 5 (Ad5) was the backbone of ONYX 15 that was a conditionally replicating adenovirus (CRAd) that replicated in tumor cells that lost function of p53. Early trials demonstrated that the virus replicated in patients with head and neck squamous carcinoma but that host immunity limited the utility of systemically administered CRAd. Also Ad5 is hepatotoxic. We have created adenoviruses with a chimeric fiber that combines a type 3 knob with a type 5 fiber to create Ad5/3. Such chimeras have reduced hepatotropism (and hepatotoxicity). This converts the receptor for the virus from the coxsackie adenovirus receptor (CAR gene) to desmoglein 2 (DSG2), a protein in hemidesmosomes expressed in epithelia and endothelial cells. Our viruses also contain the H1-driven shRNAs of the control shNEG and the allele-specific shNG-1 and shNp8-1 along with a green fluorescent protein (GFP) reporter. These viruses will replicate in and lyse human but not mouse cells and have the Wild Type E1a viral promoter and are designated Ad5/3-E1aWT-shNEG, Ad5/3- E1aWT -shNp8 and Ad5/3- E1aWT (this last without shRNA or GFP reporter). In addition, we have created a conditionally replicating adenovirus with a 5/3 fiber (CRAd5/3) where the first 1 Kb of the NANOGP8 promoter drives a part of E1a. This CRAD 5/3-E1aNp8-shNp8 also contains shNp8-1 as well as the GFP reporter and replicates in cells that support NANOGP8 expression but not in NANOGP8 negative human cells. The Ad5/3's with wild type promoters were first tested in mid-January, 2014 with Ad5/3-E1aWT and Ad5/3-E1aWT-shNp8 in Clone A and CX-1 at various dilutions of viral particles per CRC cell. As expected, increasing the viral MOI increases the infection rate but even at MOIs of only 3 5 -10% of CRC are infected and by day 7 80% of cells express GFP. In our LV experiments we achieved a maximal decrease in CRC cell survival of 50% inhibition at 3 days. In contrast, exposure of CRC cells in 3-D spheroids and aggregates to Ad5/3-E1aWT or Ad5/3-E1aWT-shNp8 decreased cell growth in 10 days to 15% of that of the untreated controls. The CRAd Ad5/3-E1aNp8-shNp8 is a little less active but achieves a 50% inhibition of growth in 10 days of suspension culture. Data indicate that both the wild type and CRAd viruses spread from cell to cell in spheroids as they lyse the human CRC cells, possibly by traversing through the desmosomes that connect one cell to another. In addition, co-cultures with murine 3T3 cells that do not support replicationof Ad5/3 inhibited the lytic activity of the adenoviruses but only when stromal:CRC cell ratio was 1:1 which is higher than occurs in most human CRC metastases. In addition, data demonstrate that the Ad5/3 with shRNA to NANOGP8 was more active than the wild type virus without shNp8. Intratumoral treatment studies are just underway to assess the effect of Ad5/3's and the CRAD in vivo. Thus, we have developed an in vitro 3-D tumor growth model that may be a beneficial model for predicting treatment effects in vivo as well as several new viral agents whose activity in vivo will be assessed shortly.
Aim 4. Assess whether inhibition of NANOG or NANOGP8 increases responses to cancer therapies. We have submitted a manuscript that demonstrates that inhibition of the NANOG/NANOGP8 gene expression inhibits MCL-1 expression as well as the expression of AKT and its activation. This effect potentiates the efficacy of BH3 peptides both in vitro and in vivo in a systemic treatment model in NOD/SCID mice as well as inhibiting the ability of CSC to support clonogenic regrowth. This is currently under review at Clinical Cancer Research (Mattoo AR, Jingyu Zhang J, Luis A. Espinoza LA, Jessup JM. Inhibition of NANOG/NANOGP8 results in MCL-1 down regulation in colorectal cancer cells to enhance the therapeutic efficacy of BH3 mimetics.). This has led to a follow-up study in which the CSC inhibitor selinomycin is used in place of shRNA to the NANOGs. The drug also potentiates the effect of the BH3 inhibitors in 3 of 4 human CRC cell lines. The 1 cell line that selinomycin does not affect is the one in which it does not inhibit MCL-1. Treatment with shNp8 did inhibit MCL-1 in all 4 lines and increased the activity of ABT-737 and ABT-199. We are now using Nanostring gene expression arrays to assess how shRNA alters gene expression. Last year we reported that the LVshRNAs to the NANOGs potentiate the activity of Topotecan on CRC. This has not been published because the molecular mechanism is still unclear and we focused on the interaction between the NANOGs and BH3 death proteins. Finally, we have found that inhibition of NANOGP8 by LV shNp8-1 activates Caspase 9 and the intrinsic pathway of apoptosis during apoptotic stress caused during culture in suspension in serum-free medium in ULLA plates. This pathway is the subject of a manuscript under review (Mattoo AR, Zhang J, Espinoza LA, Korokhov N, Jessup JM. Inhibition of NANOGP8 or NANOG Activates the Intrinsic Pathway of Apoptosis. Under review) and may also help increase the activity of drugs that may benefit from caspase-dependent cell death. Preliminary results with the Nanostring gene expression system suggests that inhibition of NANOG or NANOGP8 leads to the alteration of 5 - 10 important transcription factors that then lead to secondary modifications in gene expression. Preliminary work suggests that in some CRC lines shNp8-1 and to a lesser extent shNG-1 activate autophagy and that this also leads to apoptosis when the level of autophagy is quite high. In summary, our data support the thesis that NANOGP8 and NANOG are master regulators of stem cell function in CRC and that inhibition of their expression may cause a pronounced decrease in stemness and possibly an increase response to agents of precision medicine.
