The use of oncolytic VACV is a promising approach for the treatment of various types of cancer
[7–10]. However, the therapeutic efficacy relies extensively on the efficient oncolysis of tumor cells
, indicating that an efficient viral spreading inside tumor mass is essential. In this context, proteins of the ECM have been identified as an obstacle for the intra-tumoral spreading of virus particles
Two recently published studies reported that adenovirus-mediated intra-tumoral expression of relaxin resulted in an increased tumor regression
In the present study, we showed for the first time that degradation of the tumoral ECM also enhanced therapeutic efficacy of oncolytic rVACV. To achieve tumoral ECM degradation, we used GLV-1h255 encoding mmp-9, for the treatment of PC-3 tumor-bearing mice.
PC-3 tumors of GLV-1h68 and PBS injected mice showed a high basal MMP-9 expression. This might be due to the presence of non-malignant stromal cells, e.g. inflammatory cells such as neutrophils, macrophages or lymphocytes, most of which express MMP-9
. Since PC-3 xenograft tumors have a high metastatic potential
[30–32] and the degradation of ECM proteins by MMP-9 may facilitate metastasis
, we analyzed the size of lumbar and renal lymph nodes in PC-3 tumor bearing mice, as these have been shown to be colonized by tumor cells, and increased in size compared to healthy ones
. However, the GLV-1h255 mediated over-expression of MMP-9 did not further enlarge these lymph nodes, as they had a similar size compared to GLV-1h68-treated animals and were significantly smaller than those in PBS treated animals. In accordance, Lavilla-Alonso et al. reported no increased tumor invasiveness by macrophage metalloelastase, another ECM degrading enzyme, when administered in combination with an oncolytic virus
As MMP-9 can also induce angiogenesis by increasing the bioavailability of VEGF
 which may stimulate primary tumor growth, we quantified the BVD in sections of tumors from infected and uninfected mice. The infected areas of both GLV-1h68- and GLV-1h255 treated tumors revealed similar BVD, which was significantly reduced compared to PBS-treated tumors.
Taken together, neither an increase in size of renal and lumbar lymph nodes, nor an induction of angiogenesis was observed in PC-3 tumor-bearing mice injected with GLV-1h255. Therefore, the final biological function of MMP-9 may strongly depend on the context and the local microenvironment of MMP-9 over-expression.
In this study, we propose that the enhanced oncolytic effect of GLV-1h255 could be due to the MMP-9-mediated collagen IV degradation in PC-3 tumors.
Interestingly, the collagen IV content was significantly increased in GLV-1h68 infected tumors compared to PBS-treated tumors, which might be a result of an inflammatory reaction caused by viral infection. It has already been reported that tumor-targeted microbes, such as bacteria and viruses
[27, 36], induce an overwhelming inflammatory reaction in the tumor microenvironment, which is accompanied by an uncontrolled collagen deposition. In normal wounded tissue, inflammation is tightly controlled and ultimately leads to scar formation and healing. Tumors however, which also have been described by Dvorak as “wounds that do not heal”
, do not provide an orderly environment for the resolution of an infection focus
. This may explain the excessive deposition of collagen IV in the context of infection. The significantly decreased collagen IV content in GLV-1h255 infected areas may be the reason for the significantly higher virus-titer in GLV-1h255 infected tumors, compared to those infected with GLV-1h68, due to an accelerated extracellular cell to cell spreading. This would be in line with the findings of Ganesh et al. and Kim et al., who reported an increased tumor infection by adenoviruses after the degradation of ECM proteins
[16, 17]. This may furthermore explain the accelerated regression of GLV-1h255 infected tumors, as higher titers could result in an increased oncolysis, which has been shown to be an important factor for the therapeutic efficacy of VACV
Collectively, our study revealed that the degradation of ECM within the tumor microenvironment can boost the oncolytic effect of rVACV. Moreover, we showed here that high levels of local, intra-tumoral ECM-degrading enzymes can be produced by virus-infected tumor cells themselves, certainly limiting systemic side effects.