It has been suggested that conventional chemotherapy induces cancer cell elimination not only through direct cytostatic or cytotoxic mechanisms but also through modulation of host's immune cell function . DNA damage caused by therapies that disrupt cell cycle has direct immune stimulatory effects. Tumors do not regress in immune depleted animals treated with doxorubicin given in combination with IL-12 and several other cytotoxic agents can induce powerful immune responses against cancer . Although the effects of chemotherapeutic agents is currently believed to be due to the induction of cellular damage and subsequent activation of innate immune responses, no information is available about the transcriptional signatures within the tumor microenvironment that could support this hypothesis. A few experimental models suggest that chemotherapy acts by inducing the release of cytokines and chemokines such as CXCL-9/Monokine induced by IFN-γ (MIG), CXCL-10/IFN-γ -induced protein 10 kDa (IP-10), CCL-2/MCP-1, CCL-3/MIP-1α, CCL-5/RANTES, IL-6, IFN-γ and TNF [21, 22]. This study assessed the effect of chemotherapy at the global transcript level.
The study focused on the early effects of treatment which are more likely to provide mechanistic information about the up-stream events leading to the observed phenomenology as we have previously shown in preclinical  or clinical models [24, 25]. The results support the current view that chemotherapy acts, at least in part, by modulating immune responses (Figure 2c) resulting in enhancement of immune infiltrates (Figure 5c). Of note, CPT-11, modulated the high mobility group box-1 pathway, confirming the central role that it plays in the induction of anti-cancer activity by chemotherapeutics through activation of innate immunity [26, 27]. The effects of CPT-11 on immune responses at the dose used in this study were, however, ambivalent with a balanced proportion of genes related to inflammatory processes being up-regulated or down-regulated over their baseline activation within the tumor microenvironment (Figure 2). CPT-11 effects were centered on the regulation of apoptosis by NF-kB including activation of genes related to mitochondrial fatty acid oxidation, sodium pump regulation such as the ATPase, Na+/K+ transporting, alpha 1 polypeptide (ATP1A1), the tumor necrosis factor receptor super-family member 12A (Fn14/TNFRF12A/TWEAKR) inducer of apoptosis, the tumor necrosis factor AIP3 interacting protein 2 (TNIP2), the inositol 1,4,5-trisphosphate 3-kinase B (ITPKB). These genes have predominantly pro-apoptotic effects, while genes with clear anti-apoptotic effects and related to cell growth such as lactate dehydrogenase A (LDHA) and the transferrin receptor (TFRC) were down-regulated; thus, it appears that CPT-11 counteracts the anti-apoptotic activity of NF-kB inducing apoptosis and alterations of cellular metabolism. In contrast, pro-inflammatory pathways downstream of NF-kB regulation were predominantly down-regulated at this early time point perhaps because the tissue damage induced by chemotherapy had not reached a sufficient intensity to induce immune activation (Figure 4a and 5c). The complex activation of NF-kB-dependent transcripts leading to apoptosis and/or inflammation deserves discussion. The anti-apoptotic effects of NF-kB are balanced in cancer by the tumor suppressor effects of IRF-1 and/or IRF-5 which directly inhibit its cyto-protective activity [14, 28, 29]. In particular, IRF-5 can sensitize tumors treated with CPT-11 to DNA damage-induced apoptosis and cell death . In baseline conditions IRF-5 was found to be down-regulated in Colon 38 tumors compared to normal tissues while the expression of IRF-1 was neutral compared with other tissues (Figure 2a). We speculate that CPT-11 activates pro-apoptotic pathways by suppressing IRF-1 expression (Figure 5c) while leaving unaltered the constitutive down-regulation of IRF-5 in these tumors. The combined down-regulation of the two tumor suppressor genes may allow NF-kB to partially counteract apoptosis while the acute inflammatory process fostering cancer rejection is hampered by the down-regulation of the IRFs .
PHY906 was originally used to treat gastrointestinal symptoms including those related to the toxic effects of chemotherapy. Work from our group  demonstrates that PHY906 can revert intestinal damage caused by CPT-11 through the anti-inflammatory properties of the herb and the repopulation of intestinal progenitor cells. This study, did not address the complex effects that PHY906 has on the gut that will be addressed in the future, but focused on the early transcriptional changes observable at the tumor site. To address whether the effects of PHY906 were tissue specific, we compared the transcriptional changes at the tumor site with those in two normal organs characterized by high metabolic activity (liver) or immunologic functions (spleen). These organs were also exposed to systemic concentrations of orally administered PHY906 and/or CPT-11 (i.p.) different from those experienced by the front line exposure in the gut. Comparisons of the transcriptional profiles of the three tissues demonstrated clear baseline differences independent of any treatment which made cross comparisons of the various treatment effects among various tissues uninformative. Nevertheless, this approach clearly demonstrated that the effect of PHY906 and of chemotherapy with CPT-11 is exquisitely tissue-specific with very little overlap of pathways regulated by the various combinations in different tissues. This is not surprising since each tissue is enriched with different potential target cells that may react differently to CPT-11 or PHY906.
While tissues were not micro-dissected or single cell preparations were not obtained ex vivo, a general overview of the global changes induced by PHY906 could be obtained. In particular, we were interested in understanding how a herbal product with supposed anti-inflammatory properties could paradoxically increase the anti-tumor effects induced by chemotherapy; this concept is opposed to the current hypothesis of how the host immune system cooperates with the effects of chemotherapy . Besides its immune modulatory effects on inflammation, PHY906 could have a direct impact on cancer cell survival by modulating apoptosis, autophagy or necrosis in cells undergoing exposure to chemotherapy. Although we have previously observed that PHY906 could inhibit tumor cell growth in culture (Additional file 4, Data in S4), the in vitro data could be misleading since not all the chemicals in PHY906 are necessarily absorbed into circulation and some that are absorbed will be metabolized in vivo. However, it is possible that some of the in vitro effects mediated by PHY906 on cancer cells may be retained. In vivo, PHY906 induced a predominantly downward modulation of transcription and down-regulation of the pro-inflammatory activity naturally present within the tumor. This anti-inflammatory effect was corroborated by the decreased amount of tissue macrophages seen by immunohistochemistry (Figure 5b).This is in line with its postulated functions in the gut . Since no tumor regression was seen in response to PHY906 administered alone, it is unlikely that these effects have relevant bearing on tumor regression.
The surprising finding of this study emerged when we examined the effects of the combination of PHY906 plus CPT-11. Contrary to the anti-inflammatory properties observed when administered alone, PHY906 strongly counteracted the CPT-11 induced depression of inflammation through the IRF-mediated pathways. In particular, PHY906 counteracted the down-regulation of the master regulator of the pro-inflammatory switch and apoptosis IRF-1 [14, 31, 32]. Additionally, PHY906 enhanced the expression of IRF-5, another potent pro-inflammatory transcription factor associated with immune-mediated, tissue-specific rejection [33–36] as well as induction of apoptosis of CPT-11 treated cancers . Thus, we speculate that PHY906, in line with other experimental models , enhances the anti-tumor properties of chemotherapy with CPT-11 by imparting a pro-inflammatory state that is not observed in the same CPT-11 naïve cancerous tissue. The reasons for the contradictory behavior remain to be ascertained, although it is likely that they involve a modulation of the balance between the anti-apoptotic and pro-inflammatory functions of NF-kB while, at the same time, inducing the expression of interferon stimulated genes with chemo-attractant properties such as CCL-2/MCP-1 and CCL-5/RANTES .