Stem Cell Niche and the Radiation Response

Epithelial cells of the intestine undergo continuous turnover every four to five days. Renewal requires intestinal base columnar stem cells (CBC) located within the crypt basement membrane to cycle daily with an outcome of generating transit amplifying (TA) cells that migrate up along the villus and differentiate into enterocytes, goblet and enteroendocrine cells (37) (Supplementary Fig. S2;  http://dx.doi.org/10.1667/RR15059.1.S1). TA progenitor cells can also generate Paneth cells that migrate down into the base and may provide stem cell niche signals to self-renewing CBC cells (37). In the absence of Paneth cells, mesenchymal cells may provide compensatory niche signaling (37). Tian et al. (38) identified two stem cell pools in the small intestine. Proliferating CBC cells that express Lgr5, Ascl2 and Olfm4 provide constitutive intestinal renewal. The second pool consists of quiescent Bmi1+ cells located at positions +3 to +6. Diphtheria toxin-specific ablation of Lgr5+ CBCs revealed that Bmi1-expressing cells function as a reserve stem cell pool for regeneration of the crypt (38).

CBC cells undergo radiation-induced p53-dependent apoptosis in a dose-dependent manner. The intestinal villus-crypt can tolerate doses of up to 10 to 12 Gy, although radiation-induced apoptosis can be observed at doses as low as ∼0.02 Gy (39). For doses less than 10–12 Gy there is a dose-dependent/strain-dependent reduction of villus-crypt size that is subsequently followed by proliferation of remaining stem cells that regenerate the crypt (40). The identity of the cells that regenerate crypts after irradiation remains a critical question.

Barker et al. (41) found that crypt cells located at position +4 are significantly more radiosensitive than CBC cells: exposure to 1 Gy causes apoptosis in approximately 60% of +4 cells compared to only 20% of CBC cells. Yan et al. (42) examined the proliferation status, location and radiation response of Bmi1-expressing cells using Bmi1-CreER;Rosa26-YFP mice exposed to tamoxifen (42). Nonirradiated YFP-labeled cells were quiescent and located at position +4 (ranging from +1 to +6) (42). A 12 Gy whole-body dose induced a proliferative response in the YFP-labeled cells that subsequently resulted in expansion of the Bmi1+ lineage to repopulate crypts and villi. This work has been expanded on by Kuruvilla et al. (43), who found that Bim1-mediated stem cell expansion after irradiation requires KLF4 (43).

The response of Lrg5+ CBC cells to radiation was investigated using Lgr5-eGFP-IRES-CreERT2 mice (42). These mice harbor a knock-in allele within the first ATG codon of Lgr5 that is composed of an eGFP-IRES-CreERT2 cassette. The knocked-in cassette results in loss of Lgr5 gene function and expression of eGFP (42). In contrast to the radiation response of the +4 Bmi1 CreER;Rosa26 YFP cells, CBC Lgr5-eGFP-IRES-CreERT2 cell numbers were significantly reduced 48 h after 12 Gy irradiation, although sporadic regeneration was observed 7 days postirradiation. Yan et al. (44) then expanded on these studies using RNA-seq to characterize stem cell inter-relatedness. They found that the CBC signature genes (e.g., Lgr5, Ascl2 and Olfm4) were also expressed at low levels in Bmi1-GFP+ cells while Lgr5+ cells express 4+ markers Bmi1, mTert and Hopx, but at low levels. These results suggest that crypt position and individual cell markers may be insufficient for cell identification.

Metcalfe et al. (40) approached the question of stem cell regeneration after irradiation using Lgr5^DTR mice in which the diphtheria toxin receptor is knocked into the endogenous Lgr5 locus. Administration of diphtheria toxin results in Lrg5+-specific cell death (40). A regenerating phenotype was produced after exposure to 10 Gy of radiation. In contrast, diphtheria toxin-mediated ablation combined with 10 Gy irradiation of Lgr5+ cells resulted in attenuation of the regenerative response (40). Metcalfe et al. (40) reconciles his results with that of Yan et al. (42) by citing the work of Buczackie et al. (45), who identified a distinct, nondividing radiation-resistant population of Lgr5+ cells that are also positive for Bmi1+/4+ markers. These cells reside in positions 1–6 and have the potential to regenerate crypts after exposure to 6 Gy or less (40).

The Role of Nrf2

Yang et al. (46) addressed the question of whether Nrf2 protected intestinal stem cells from radiation injury. Their approach was to administer 13 Gy to the abdomen of C57BL/6J mice rather than to the whole body. This targeting spares the majority of the bone marrow. All (100%) Nrf2-null mice survived for 30 days postirradiation, whereas only 80% of the wild-type survived. Abdominal irradiation resulted in apoptosis and a significant loss of crypt-villus anatomy in wild-type mice. Abdominal irradiation-induced death day 14 and is a consequence of loss of intestinal stem cells. Strikingly, the crypt-villus structure was spared in the Nrf2-null irradiated mouse. At the stem cell level, the number of Lgr5+ cells was greater in the Nrf2-null crypt compared to crypts in wild-type mice. These investigators found that loss of Nrf2 resulted in greater radiation-induced activation of NFκB compared to wild-type intestine. A control cohort of C57BL/6J mice received 7.5 Gy whole-body irradiation. The LD50₃₀ for a C57BL/6 mouse is in the range of 8 to 9 Gy (47) and lethality is a consequence of bone marrow stem cell ablation. Of the wild-type mice that received 7.5 Gy whole-body irradiation, 43% survived to day 30. In contrast, less than 10% of Nrf2-null mice survived a 7.5 Gy whole-body dose. This result is consistent with the results reported by McDonald et al. (47). The observation that the Nrf2-null mouse was to resistant to radiation after abdominal exposure is unexpected but intriguing and bears further investigation.

Triterpenoids are electrophiles (48) that undergo Michael reaction and inactivate Keap1 (19), thus allowing Nrf2 to avoid ubiquitination and proteasome degradation, and thereby producing elevated Nrf2 signaling. CDDO (1-(2-cyano-3,12-dioxooleana-1,9-dien-28-oyl)) is a triterpenoid that results in robust Nrf2 activation. Kim et al. (49) fed mice a diet supplemented with 400 mg/kg CDDO for 3 days prior to whole-body irradiation. CDDO induced Nrf2 and the Nrf2 DNA repair target gene 53BP1. CDDO reduced radiation-induced apoptosis, and attenuated radiation-induced loss of crypt length and cell density in the colon and small intestine. Mechanistically the results appeared to be due to enhanced signaling of the DNA damage response pathway (49).

Stem Cell Niche and the Radiation Response

Long-term hematopoietic stem cells [LT-HSCs: Lin^–, Sca1⁺, c-Kit⁺, CD150⁺, CD48^–, CD41^– (50)] in bone marrow can be found adjacent to sinusoid vessel niches (50) that are supported by endothelial and nestin-expressing mesenchymal cells that produce SCF and CXCL12, as well as other factors (50, 51). Approximately 20% of LT-HSCs are also found adjacent to osteoblasts in bone endosteum niches (50, 52). Endosteal trabecular niches gain importance after ablative doses of ionizing radiation. Forty-eight hours after exposing mice to 1.125 Gy irradiation, Dominici et al. (53) observed proliferation and expansion of osteoblasts located proximal to the endosteal surface of trabecular bone while noting extensive loss of endothelial niches. These investigators then identified osteoblast niches as the sites of engraftment after transplantation of GFP-expressing donor HSCs into the lethally irradiated mice (53). The Sp7/Osx transcription factor is a Runx2 target gene required for osteoblast differentiation (54). Recently, Himburg et al. (51) demonstrated that Sp7-expressing cells secreting dickkopf-1 (Dkk1) promoted hematopoietic stem cell regeneration after ablative irradiation.

The Role of Nrf2

Rana et al. (55) found that Nrf2 was critical to osteoblast survival after ablative irradiation. Significantly fewer osteoblasts survived irradiation in Nrf2-null compared to wild-type mice, suggesting that the niche's ability to recover from radiation injury is dependent, in part, on Nrf2. Within the niche, LT-HSCs and hematopoietic stem progenitor cells (HSPCs) exhibit a hypersensitivity to low-dose radiation (0.02 Gy), a consequence of elevated reactive oxygen species (ROS) that induces autophagy and mitophagy (56). Radiation-induced ROS stabilizes Nrf2, resulting in nuclear translocation and induction of Nrf2-mediated antioxidant gene expression. Failure to induce Nrf2 in Nrf2^–/– mice increased HSC radiohypersensitivity, whereas inducing Nrf2 prior to irradiation in Mx-Cre-Keap1^flox/flox mice treated with poly (I:C) abolished hypersensitivity (56). Additionally, Nrf2^–/– LSK were sorted, irradiated at 0.02 Gy and engrafted in recipient mice. Analysis performed 6 months after engraftment showed a dramatic decrease in LT-HSC (56). Nrf2 expression also influences the effectiveness of myoablative doses of radiation: loss of Nrf2 lowers the LD50₃₀ (47, 57). Conversely, pharmacological induction of Nrf2 increases the LD50₃₀ (49, 57).

Nrf2 Modulates Hematopoiesis

Short-term HSCs (ST-HSCs; Lin^–, Sca-1⁺, c-Kit⁺, CD34⁺ Flt3^–), multipotent progenitor cells (MPP; Lin^–, Sca-1⁺, c-Kit⁺, CD34⁺ Flt3⁺), committed myeloid and lymphoid lineages, but not LT HSCs are increased in Nrf2-null mice compared to wild-type mice (58). While LT-HSCs cell numbers were independent of Nrf2, fewer of these cells were quiescent compared to wild-type controls, as measured by expression of proliferation and differentiation markers, a consequence of elevated cyclin D1 activity and loss of CXCR4-mediated signaling (58). These defects in Nrf2-deficient HSCs resulted in accelerated hematopoietic exhaustion, as assessed in serial competitive transplantation experiments using lethally irradiated mice (58, 59). Thus, an Nrf2 deficiency has the potential to reduce engraftment after myoablative irradiation.

Kim et al. (57) used Mx-Cre;Keap1^flox/flox mice to study the effects of Nrf2 overexpression on HSC function. In LSK cells (Lin^– Sca1⁺ c-Kit⁺; enriched for LT and ST HSCs, as well as MPP cells) loss of Keap1 resulted in elevated Nrf2 target gene expression (57). Under steady-state conditions there was an increase in LSK cell numbers in Mx-Cre;Keap1^flox/flox mice compared to control. Use of colony-forming unit assays revealed an increase in committed progenitors (GM-CFU and G-CFU). Competitive transplantation of LSK cells into lethally irradiated mice revealed increased engraftment of Mx-Cre;Keap^flox/flox poly (I:C) treated donor cells compared to donor cells from Keap1^flox/flox mice (57). These results were interpreted to indicate that prolonged Nrf2 signaling enhances hematopoietic reconstitution after myelosuppressive irradiation due to activation of Notch signaling. Use of tamoxifen-treated CMV-Cre:Keap1^flox/flox mice and pharmacological activation of Nrf2 confirmed the results obtained with of Mx-Cre;Keap^flox/flox poly (I:C) treated mice (57).

Vav1-Cre:Keap1^flox/flox mice were used to test the hypothesis that Nrf2 promotes HSC proliferation (60). Disruption of Keap1 exons 4 through 6 in LSK cells and LT HSCs elevated expression of Nrf2 and Nrf2 target genes. Similar to LSK cells in Mx-Cre;Keap1^flox/flox mice, there was a 1.4-fold enrichment of LSK cells in Vav1-Cre:Keap1^flox/flox mice (60). LT HSC numbers in Vav1-Cre:Keap1^flox/flox did not increase; rather, the cells moved from a G₀ quiescent state into G₁. This exit from quiescence was not observed in Vav1-Cre:Keap1^flox/flox::Nrf2^–/– mice, demonstrating that the loss of quiescence was Nrf2 driven (60). Mechanistically, Nrf2-mediated loss of quiescence was a consequence of JAK-STAT signaling. These results suggested that the inability to increase cell number coupled with loss of quiescence would impair self-renewal, thereby limiting the ability of LT HSCs to reconstitute lethally irradiated bone marrow. This hypothesis was tested using a competitive transplantation assay. Indeed, LT HSCs obtained from Vav1-Cre:Keap1^flox/flox mice exhibited a deficiency in reconstitution capacity compared to control (60). These results were confirmed using Mx-Cre;Keap1^flox/flox mice treated with poly (I:C) or administered the Nrf2 activator 1-(2-cyano-3,12-dioxooleana-1,9-dien-28-oyl) imidazole (CDDO-Im). In these model systems it appears that prolonged activation of Nrf2 reduces LT HSC self-renewal and quiescence while driving differentiation, thereby diminishing the ability of LT HSCs to reconstitute irradiated bone marrow (60). The differences between the study of Kim et al. (57) and that of Murakami et al. (60) are not currently understood and require further investigation.

The Radiation Response

There are several excellent reviews with in-depth discussions on the subject of radiation-induced pneumonitis and pulmonary fibrotic injury (61–63). In brief, radiation pneumopathy is composed of two pathophysiologies: pneumonitis and fibrosis. Pneumonitis is a mostly resolvable inflammatory reaction that develops within weeks postirradiation, whereas fibrosis takes months to years to develop and is progressive (61). Fibrosis is considered a disease induced by recurrent and/or nonresolving injury to lung epithelium that results in cell death and/or phenotypic alterations to surviving cells, which leads to generation of myofibroblasts/fibrocytes and deposition of extracellular matrix (64–71). Radiation-induced pulmonary injury can be characterized by endothelium and pneumocyte apoptosis and senescence (71, 72) that are followed by recruitment of NADPH oxidase-expressing inflammatory cells and myofibroblasts to sites of injury, which in turn contribute to oxidative and cytokine stress. In published studies, it has been shown that suppression of TGF-β/Smad signaling, or overexpression of superoxide dismutase and/or overexpression of catalase abrogates development of radiation-induced pulmonary fibrosis (73–76). Such research supports the hypothesis that TGF-β1 and oxidative stress are two drivers of fibrosis. Interestingly, as a side bar, Favaudon et al. (77) found that short pulses (less than 500 ms) of radiation administered at dose rates in excess of 40 Gy/s did not induce pulmonary fibrosis if the total dose delivered did not exceed 20 Gy.

The Role of Nrf2

The severity of radiation-induced pulmonary fibrosis is enhanced in the Nrf2-null mouse. Loss of Nrf2 increases radiation-mediated pulmonary tissue remodeling and collagen deposition (78). Loss of Nrf2 in irradiated lung impairs ΔNp63 stem/progenitor cell mobilization and proliferation, regeneration of alveolar type 2 cells (78), with an outcome of elevated FSP1 expression (79) and enhanced Masson's trichrome staining of collagen deposition (78), two markers of fibrosis. In injured lung Δp63+/Krt5+ stem cells can be mobilized to proliferate significantly and differentiate into alveolar type 1 and 2 epithelial cells in vitro and in vivo. Diphtheria toxin targeting of these stem cells impairs regeneration of injured lung and pulmonary oxygenation while promoting fibrosis, as measured by FSP-1, Masson's trichrome, and α-smooth muscle staining (80, 81). Thus, it has been hypothesized that ΔNp63+ stem/progenitor cell mobilization contributes to alveolar type 2 cell regeneration after radiation injury (78). However, it is important to note that if there are hypoxic regions in injured lung then HIFα/Notch signaling reprograms Δp63+/Krt5+ cells to form basal-like metaplasia (82). Additionally, an Nrf2 deficiency promotes radiation-induced alveolar type 2 cell epithelial-mesenchymal transitions into myofibroblasts (78). At a molecular level it has been shown that Nrf2 can form a nuclear complex with nuclear pSmad3 at CAGA sites located in the proximal promoters of TGF-β target genes (79), suppressing gene expression (79, 83). Thus, loss of Nrf2 promotes TGF-β/RSmad signaling, a key factor for promoting epithelial-mesenchymal transitions (84). Conversely, TGF-β/RSmad signaling has been shown to suppress Nrf2-dependent gene expression. TGF-β/RSmad signaling induces ATF3, which in turn complexes with Nrf2. The ATF3/Nrf2 complex binds to ARE sites in the promoters of Nrf2 target genes, suppressing recruitment of CBP to AREs, with an outcome of suppressing Nrf2 target gene expression (37, 38, 85). Loss of Nrf2-dependent gene expression has been shown to promote myofibroblast development and increased oxidative stress (35, 86).

The molecular identity of radiation-induced pulmonary oxidant injury has been an enigma until recently. Mont et al. (87) provided direct biochemical evidence that isolevuglandin (IsoLG) adduction of proteins represents a radiation-induced, oxidant-mediated injury in lung. Genetic evidence linking oxidant challenge to IsoLG adduction has been provided in studies using genetically engineered mice, where it was shown that oxidative stress, generated by either loss of Nrf2 or NADPH oxidase activity, promoted IsoLG adduction (87). Isolevuglandins (IsoLGs), also known as isoketals, are formed by free radical oxidation of arachidonic acid (88). IsoLGs rapidly adduct the ϵ-amino group of protein lysines. Mont et al. (87) identified over 100 proteins susceptible to IsoLG adduction, including lysines on histones H3 and H4. Adduction of these histones would be expected to negatively affect epigenetic regulation and may represent a mechanism that contributes to fibrotic reprogramming (64). Extracellular collagen present in fibrotic lesions has been identified as another target for adduction (87). IsoLG adduction of collagen impairs the ability of MMP1 to degrade it, thus contributing to collagen accumulation. Chronic oxidative stress results in accumulation of IsoLG-adducted protein that can evolve into proteotoxic/apoptotic events (78), i.e., recurrent injury.