Ionizing radiation directly damages DNA leading to genomic instability and cell death through apoptosis or mitotic catastrophe (O’Neill and Wardman 2009). Investigators evaluating radiation effects on DNA have found that high linear energy transfer results in “clustered DNA lesions” which are inefficiently repaired and are thus thought to be biologically relevant to the acute and delayed effects associated with radiation exposure (O’Neill and Wardman 2009). The ionizing event also results in formation of reactive oxygen species (ROS), however the levels generated by therapeutic doses of radiation are negligible compared to those generated by aerobic metabolism (Ward 1994). Within hours of the initial ionizing exposure, there is a robust activation of secondary sources of oxidative stress leading to persistent production of free radical species that can participate in cell signaling. A number of transcription factors and pro-inflammatory/pro-fibrogenic molecules, such as NFκB, HIF-1α, and TGF-β, are sensitive to changes in the redox status of the cell. As the level of free radical production is thought to be beyond the capacity of the cell’s antioxidant defense mechanisms to eliminate them, the imbalance between free radicals and antioxidants leads to chronic upregulation of transcription factors, growth factors, and cytokines involved in radiation injury (Robbins and Zhao 2004; Benderitter et al. 2007; Zhao et al. 2007). The environment of chronic oxidative stress results in oxidative damage to DNA leading to chromatin remodeling and altered gene expression, chronic protein activation/inactivation, and oxidation/peroxidation of membrane lipids that can result in changes in cell signaling (Mikkelsen and Wardman 2003). The ability of potent superoxide dismutase (SOD)-like compounds and other antioxidants to mitigate radiation-induced injury is further evidence that redox-regulated signaling plays a substantial role in the development of radiation injury (Epperly et al. 1999a, b, 2002; Gauter-Fleckenstein et al. 2007, 2010; Rabbani et al. 2005, 2007; Vujaskovic et al. 2002; Rosenthal et al. 2011).

While the exact mechanism through which radiation-induced free radical production affects the activity of signal transduction pathways and transcription factors within the irradiated tissue is not clear, it is well known that many signaling mechanisms and gene expression pathways are sensitive to changes in intracellular oxidation/reduction status (Mikkelsen and Wardman 2003) leading to changes in signal transduction, transcription factor activation, and gene expression (Mikkelsen and Wardman 2003; Reboucas 2010; Zhang and Hogg 2005). It is theorized that the increase in ROS/RNS production in the early radiation response may result in oxidative modification of macromolecules participating in signal transduction pathways, resulting in signaling changes that perpetuate tissue injury over time (Reboucas 2010).

It is hypothesized that the initial ionizing event is followed by a ROS/RNS mediated signaling cascade through downstream activation of metabolic sources of prooxidant production. This includes mitochondria, nitric oxide synthases, and oxidoreductase enzymes, most notably the family of NADPH oxidases.

Non-phagocytic cells predominantly generate ROS/RNS through activity of the cell membrane-associated family of NADPH oxidases (von Lohneysen et al. 2010). A number of NADPH oxidase family members exist and have remarkably varied activation mechanisms and tissue distribution. For example, endothelial cells express Nox1, Nox2 (gp91^phox), Nox4, and Nox5 isoforms; whereas vascular smooth muscle cells express the Nox1, Nox4, and Nox5 isoforms (Lassegue and Griendling 2010). NADPH oxidase-derived superoxide anion (O₂^{. –)} is involved in vascular maintenance and regulation of vascular smooth muscle tone under physiological conditions (Bengtsson et al. 2003). However, pathological overproduction of O₂^{. −} by NADPH oxidases may participate in radiation injury by activating redox-sensitive signaling pathways involved in inflammation and fibrosis. Collins-Underwood et al. (2008) found that increased expression of NADPH oxidase subunits (Nox4, p22^phox, and p47^phox) following in vitro irradiation of rat brain microvascular endothelial cells was associated with a pro-inflammatory response. In these studies, elevated intracellular generation of ROS, activation of the pro-inflammatory transcription factor NFκB, and expression of vascular adhesion molecules ICAM-1 and PAI-1 could be prevented by inhibition of NADPH oxidase (Collins-Underwood et al. 2008).

Nox4 is especially interesting as recent studies have shown TGF-β1 induced cell death to be dependent upon Nox4-derived H₂O₂ (Carnesecchi et al. 2011). Nox4 is a unique isoform as it produces H₂O₂ rather than O₂^{. −} due to an extra 23-amino acid sequence in its E-loop (Takac et al. 2011). In recent studies by Zhang et al. (2012), Nox4 expression was colocalized with increased PTEN expression, which correlated to a progressive and persistent increase in cell death between 24 h and 6 months after whole thorax irradiation. The increase in PTEN expression coincided with downregulation of Akt signaling.

Within the first week following exposure to radiation, there is a decrease in inflammatory cells followed by an increase in neutrophils and lymphocyte accumulation that quickly returns to normal. Over time, fluctuations in inflammatory cell subsets infiltrating the irradiated tissue can be observed. The “perpetual cascade of cytokines” in the lungs after thoracic irradiation was first proposed by Rubin et al. 1995 which demonstrated an absence of a “latent period” at the molecular and cellular level in late responding tissues.

Our group has helped to pioneer the use of increasingly sophisticated molecular biologic methods to assess and analyze late effects in normal tissue (LENT). In so doing, we have extended the paradigm o the clinical pathologic course of events following irradiation or chemotherapy from target cells per se to include and emphasize the intercellular communication. In our current paradigm, during the course of LENT induction, the so-called cytokine cascade can be considered to involve four basic cell components in tissues or organs—the parenchymal cell, the inflammatory cell, the endothelial cell, and the interstitial cell. These are illustrated in Fig. 2a, are a few of the proinflammatory and profibrotic cytokines that are thought to be expressed concurrently (Rubin and Williams 2001).

An intercellular conversation is initiated at the moment of irradiation, when injury to cell components occurs (e.g., membrane, cytoplasmic body, or DNA), which leads to altered gene expression. This reaction is often in the form of an immediate release of cytokine mRNA; in time, this reaction may provoke a series of downstream events through cell signaling, which is illustrated in Fig. 2b. Through signal transduction, the receptor cells are activated; such activation may result in as little as additional or sequential cytokine expression or it can lead to proliferation or production of extracellular matrix proteins, depending on the species of receptor cell. In the specific case of the receptor cell being a fibroblast, activation of the collagen gene, which has been seen within 24 h of injury, can persist for days, weeks, or even months. This extended persistence may continue until the pathologic or clinical expression of injury is manifested (Rubin and Williams 2001).

Using a mixture of molecular biologic techniques and in vivo/in vitro assays, a number of in-field effects can be appreciated. The release of cytokines occurs shortly after irradiation and persists until the pathologic and clinical expression of late effects, and there is an arbitrary temporal division of cytokine expression:

This scenario can be seen to correlate and give further credence to the postulated shape of the clinical pathologic course of events. The potential to use cytokines to alter the therapeutic ratio favorably is of great interest to clinicians in the protocol design of new clinical trials.

The most predominant cytokines involved in inflammation appear to be IL-1α/β, IL-6, TNF-α, and TGF-β (Rubin et al. 1995; Chiang et al. 1997; Herskind et al. 1998; Rube et al. 2004; Brush et al. 2007), however the pattern and timing of cytokine expression appears to vary among studies (Hill 2005), as well as among animal models. The variation of cytokine expression in response to radiation between mouse strains may present one possible mechanism for the observed difference in strain sensitivity and response to radiation. Likewise, baseline differences in the number and types of resident inflammatory cell types among strains as well as genetic differences in leukocyte recruitment may contribute to the variation in radiation response. For example, fibrosing strains have a higher ratio of active to latent TGF-β than non-fibrosing strains (Franko et al. 1997). Epperly et al. (1999) found a time dependent progression in pathological fibrosis coincided with increased IL-1 mRNA levels and a biphasic wave of TGF-β expression during the development of fibrotic disease in C57BL/6 J mice (Epperly et al. 1999). However, targeting cytokine production has not been completely effective in improving survival and respiratory function, nor has any single circulating cytokine been identified as a potential biomarker for susceptibility to development of radiation-induced injury.

Cytokines expressed in irradiated tissue, such as interleukins, monocyte chemoattractant protein-1, and keratinocyte chemoattractant (KC), are consistent with what one would expect based on the inflammatory cell infiltrates found in the tissue and in circulating plasma. Inflammatory cells accumulate in irradiated tissue due to persistent upregulation of vascular adhesion markers (Epperly et al. 2002; Son et al. 2006; Muller et al. 2006; Jaal and Dorr 2005; Ikeda et al. 2000; Olschowka et al. 1997; Gaugler et al. 1997). Targeting inflammatory cell recruitment through inhibition of vascular adhesion molecules, specifically ICAM-1, reduces the number of infiltrating inflammatory cells and blocks the development of fibrosis in irradiated lungs of C57BL/6 mice (Hallahan et al. 2002; Hallahan and Virudachalam 1997). Although the impact of systemically blocking ICAM-1 on acute and chronic inflammation is significant, there is no improvement in respiratory function or survival among non-treated and treated animals exposed to radiation (Hallahan et al. 2002). This suggests an underlying cause for radiation pneumonitis/fibrosis, which impacts animal survival, distinct from inflammation alone.

It has long been acknowledged that the vascular endothelium may be an important target of ionizing radiation. Indeed, several studies have shown both gastrointestinal toxicity and xerostomia to be influenced by endothelial cell damage. Ionizing radiation can disrupt the structural integrity of the vascular architecture resulting in endothelial cell damage, barrier breakdown, leakage, and edema (Anscher et al. 2005; Martin et al. 2000; Stone et al. 2004). As a consequence of the disruption in vascular function, hypoxic regions can develop in the under-perfused tissue. Indeed, one of the key pathophysiological consequences of vascular injury after radiation is the development of tissue hypoxia. Tissue hypoxia can recruit inflammatory cells, where they are then activated to undergo the respiratory burst. Not only does this increase oxidative stress, but also further contributes to tissue hypoxia as oxygen is required for the respiratory burst in phagocytic cells. One of the primary cell types recruited by hypoxia is the macrophage. Jackson et al., demonstrated that macrophages incubated under hypoxic conditions exhibit a redox-dependent increase in TGF-β production and the hypoxia-inducible factor-1 alpha (HIF-1α) product VEGF (Jackson et al. 2007).

HIF-1α has observed in lung tissue as early as one day after radiation. The ability of free radical scavenging to inhibit stabilization of HIF-1α has implicated ROS/RNS as a critical activator of HIF transcriptional activity (Pouyssegur and Mechta-Grigoriou 2006; Li et al. 2007). The consequences of HIF stabilization in irradiated healthy tissue likely include greater endothelial dysfunction, enhanced vascular hyperpermeability, formation of temporary fibrin matrices, and angiogenesis (van Hinsbergh et al. 2001; van Hinsbergh 2001a, b; Semenza 2000). Vascular leakage and fibrin deposition facilitate the formation of new vessels (van Hinsbergh et al. 2001) through establishment of a provisional fibrin matrix which provides a stable scaffold for HIF-1α mediated inflammatory and endothelial cell migration and vessel sprouting (Haroon et al. 1999). Under physiological conditions, HIF has a relatively short half-life of less than 5 min (Wang et al. 1995; Huang et al. 1998). However, under hypoxia, or in the event of oxidative/nitroxidative modification of protein residues in the oxygen dependent domain, the alpha subunit of HIF is stabilized and translocated to the nucleus where it forms a heterodimer with its constitutively expressed HIF-1β subunit. The HIF heterodimer binds to the hypoxia response element (HRE) of a wide variety of genes regulating cell proliferation, migration, pH, apoptosis, energy metabolism, and most importantly, angiogenesis (Brahimi-Horn et al. 2005).

Vujaskovic et al. showed an increase in expression of the HIF product vascular endothelial growth factor (VEGF) between the time of radiation exposure and development of symptomatic disease in a rat model of radiation-induced lung injury (Fleckenstein et al. 2007