We previously reported preventive and therapeutic effects of Smad7, a multifunctional protein, on radiotherapy (RT)-induced mucositis in mice without promoting human oral cancer cell survival or migration in vitro. The current study aims to determine whether a Smad7-based biologic can treat existing oral mucositis during radiotherapy for oral cancer and whether this treatment compromises RT-induced cancer cell killing in neighboring oral cancer.
Experimental Design: We transplanted human oral cancer cells into the tongues of mice and applied craniofacial irradiation to simultaneously kill tumor cells and induce oral mucositis, thus modeling RT and mucositis in oral cancer patients. We topically applied a recombinant human Smad7 protein fused with the cell-penetrating Tat tag (Tat-Smad7) to the oral mucosa of tumor-bearing mice post RT when oral mucositis began to develop.
Topically applied Tat-Smad7 penetrated cells in both the oral mucosa and oral cancer, attenuating TGFβ and NF-κB signaling as well as inflammation at both sites. Tat-Smad7 treatment alleviated oral mucositis with reductions in DNA damage and apoptosis in keratinocytes, but increased keratinocyte proliferation compared with vehicle-treated mucositis lesions. In contrast, adjacent oral cancer exposed to Tat-Smad7 did not show alterations in proliferation or direct DNA damage, but showed increased oxidative stress damage and apoptosis compared with tumors treated with vehicle.
Our results suggest that short-course Tat-Smad7 application to oral mucositis promotes its healing but does not compromise the cytotoxic effect of RT on oral cancer and has context-specific effects on oral mucosa versus oral cancer.
Oral mucositis, painful oral ulcerations, is one of the most common toxic effects of radiotherapy (RT) and chemotherapy in cancer patients and can lead to therapy withdrawal or dose reduction. In head and neck cancer (HNC) patients, a major challenge in treating oral mucositis is to repair ulcerated mucosa without promoting neighboring cancer growth. To date, there is no FDA-approved drug to treat oral mucositis in HNC patients. We studied a novel biological agent for treating oral mucositis in a mouse model mimicking RT-induced oral mucositis adjacent to oral cancer. Our preclinical data demonstrate the feasibility of a novel therapeutic approach for treating existing oral mucositis without compromising radiotherapy in neighboring cancer.
Oral mucositis, painful ulceration of the oral mucosa, is a common toxic effect of radiotherapy (RT) for bone marrow transplant, craniofacial RT for head and neck cancer (HNC), and chemoradiotherapy for all cancer types (1). Approximately 70% of HNC patients develop oral mucositis during treatment, which can be severe enough to cause reduction in oral intake or premature withdrawal from cancer treatment (1, 2). New RT technologies such as intensity-modulated RT and stereotactic body RT (SBRT) more precisely target cancer lesions and spare more normal tissue. However, these treatments do not reduce acute toxicity in the oral mucosa of HNC patients (3), because neighboring mucosa is still exposed to high intensity RT, albeit at a lower dose than the cancer. Further, certain patients are at high risk for oral mucositis regardless of treatment regimen (4). Palifermin, recombinant human keratinocyte growth factor (KGF), is the only FDA-approved targeted therapy for preventing oral mucositis in bone marrow transplant patients (4% of the at-risk population), but it has no effect on existing mucositis (5). Palifermin clinical trials in oral cancer patients showed modest prevention of severe oral mucositis (6, 7). A major challenge in treating oral mucositis is to repair ulcerated mucosa without promoting cancer, as growth factors (e.g., KGF) and their receptors are often overexpressed in cancer cells. To date, there is no FDA-approved drug to treat oral mucositis in cancer patients. Clinical trials for GC4419, a superoxide dismutase mimetic, to treat oral mucositis in HNC patients revealed reductions in severe oral mucositis cases (8). GC4419 requires i.v. infusions 1 hour prior to each of 14 doses of RT. Because it is difficult to predict which patients will develop severe oral mucositis, we sought to develop a therapeutic intervention that is topically applied to existing oral mucositis and targets multiple pathogenic processes of oral mucositis.
Oral mucositis develops as a consequence of complex molecular and cellular pathobiology processes, leading to apoptosis of basal epithelia cells, loss of epithelial renewal, atrophic damage, excessive inflammation, and ulceration (9). Our previous study discovered that in addition to NF-κB activation, activated TGFβ signaling contributes greatly to RT-induced oral mucositis (10). TGFβ is a potent growth inhibitor and apoptosis inducer for epithelial cells and a proinflammatory cytokine in oral mucosa (11). To dampen both TGFβ and NF-κB pathways to treat oral mucositis, we developed a recombinant Smad7 protein that contains human Smad7 fused to the HIV-1 Tat protein transduction domain. Tat-Smad7 protein rapidly penetrates cells upon contact (10). Local delivery of Tat-Smad7 to mouse oral mucosa shows prophylactic and therapeutic effects on RT-induced oral mucositis (10). A remaining question is how to effectively utilize this therapeutic intervention without compromising RT-directed cancer cell killing. Smad7 can be either tumor suppressive or tumor promotive in different cancer types (12). For cancers outside the oral cavity, the concern for potential systemic tumor-promoting effects of Tat-Smad7 is relatively minor because orally administered Tat-Smad7 protein will be degraded as it passes through the digestive tract. However, when RT-induced oral mucositis in oral cancer patients is treated with Tat-Smad7, the oral cancer will also be exposed to Tat-Smad7. Although increasing epithelial cell proliferation and reducing apoptosis in oral mucositis promote wound healing, these effects could compromise RT-directed cancer cell death. We have previously shown that Tat-Smad7 increases survival of human oral keratinocytes but not head and neck squamous cell carcinoma (HNSCC) cells after RT. Similarly, Tat-Smad7 increases migration of normal keratinocytes but not HNSCC cells (10). These in vitro data indicate the necessity of in vivo testing of Tat-Smad7 using a model system mimicking RT in human oral cancer with an RT dose sufficient to induce oral mucositis. Although our previous study shows both preventive and therapeutic effect of Tat-Smad7 on oral mucositis (10), the current study focused on assessing the therapeutic effect of Tat-Smad7 on promoting healing of existing oral mucositis in a model of oral cancer. We chose two HNSCC cell lines: FaDu, with a SMAD4 deletion representing altered TGFβ signaling found in tobacco-associated oral cancer (13), and UM-SCC-1, with wild-type SMAD4, representing intact TGFβ signaling in tumor cells. We provide evidence that Tat-Smad7 application to RT-induced oral mucositis with neighboring oral cancer effectively promotes oral mucositis healing while paradoxically increasing tumor cell death in RT-treated oral cancer.
Materials and Methods
Two human oral cancer cell lines were used. FaDu was purchased from the ATCC, and UM-SCC-1 (14) was provided by MTA through University of Michigan. Cells were authenticated by short tandem repeat profiling (University of Colorado Cancer Center Protein Production, Monoclonal Antibody, Tissue Culture Shared Resource) prior to our experiments. Cells were cultured in DMEM containing 10% FBS.
Generation and application of Tat-Smad7
As previously described, we produced recombinant Tat-tagged, human Smad7 protein as a GST fusion protein from E. coli (10). Tat-Smad7 was cleaved from GST with PreScission protease followed by size exclusion chromatography to purify Tat-Smad7 protein as a monomer (Supplementary Fig. S1). For in vivo treatment, 1 μg Tat-Smad7 (in 30 μL PBS containing 30% glycerol) was applied to mouse oral cavity; food/water was withdrawn for 1 hour after treatment to minimize treatment disruption.
Orthotopic human oral cancer xenotransplantation
Animal experiments were performed in accordance to an Institutional Animal Care and Use Committee–approved protocol. We used female 8- to 10-week-old athymic nude mice (Charles River Laboratories) as xenotransplantation recipients. Mice were anesthetized with 80 mg/kg ketamine and 12 mg/kg xylazine (i.p. injection). Note that 105 oral cancer cells (FaDu or UM-SCC-1) were suspended in 20 μL of 50% PBS/50% Matrigel and injected directly into the anterior/middle tongue using a syringe with 30-gauge needle.
Craniofacial irradiation to oral tumor–bearing mice
Mice with tongue tumors (7 days after injection of cancer cells) were cranially irradiated to induce oral mucositis using a RS2000 biological irradiator (Rad Source Technologies) as previously reported (10). Based on our previous observation that a single 18 Gy dose has kinetics and severity of oral mucositis similar to 8 Gy x3 fractionated irradiation (10), we chose 18 Gy irradiation to reduce the death rate due to repeated anesthesia for fractionated irradiation to oral tumor-bearing mice that had deteriorating health. Each mouse was anesthetized with 80 mg/kg ketamine and 12 mg/kg xylazine and placed under a lead shield exposing only their head. The day of irradiation was designated day 1. All animals were provided soft food in addition to standard diet. On day 6, when mice began to lose weight due to oral ulcer-associated reduced food intake, mice were divided into two groups (of equal weight and tumor size) and treated daily with Tat-Smad7 or vehicle. BrdU (125 mg/kg) was administered i.p. 2 hours before euthanasia. Mice were sacrificed and tongue samples collected on day 10 for pathologic evaluation and immunostaining.
Pathologic evaluation, immunostaining, and terminal deoxynucleotidyl transferase dUTP nick end labeling assay
Tongue tissues were fixed in 10% formalin, embedded in paraffin, and cut in 5 μm sections. Tongue epithelium and tumor histology were evaluated using hematoxylin and eosin (H&E)–stained slides. Open ulcer size, defined as complete loss of epithelium, was measured using NIS (Nikon Intensilight) Elements software by two independent investigators, and the results averaged to determine ulcer size (mm). We performed immunohistochemical and immunofluorescent staining as previously described (10). The primary antibodies used were guinea pig anti-Keratin14 (1:200, Fitzgerald, 20R-CP200), chicken anti-Keratin5 (1:200, Biolegend, #905901), FITC-labeled antibody to BrdU+ (BD Bioscience, 347583), rat anti-mouse CD45 (1:50, BD Bioscience, 550539), rabbit anti-mouse F4/80 (1:400, Cell Signaling Technology, 70076), rat anti-mouse Ly6G (1A8, 1:3,200, Biolegend #127602), rabbit anti–NF-κB subunit p50 (1:100, Santa Cruz Biotechnology, SC-7178), rabbit anti-pSmad2 (1:200, Cell Signaling Technology, 3101), rabbit anti-pSmad3 (1:400, Abcam, ab52903), rabbit anti-pH2AX (1:100, Cell Signaling Technology, 9718), and mouse anti–8-OHdG (1:100, Alpha Diagnostic, 8OHG11-M). Smad7 antibody was produced using human Smad7 recombinant protein to immunize rabbits. Specificity of Smad7 antiserum was confirmed by Western blot (Supplementary Fig. S2). Secondary antibodies conjugated to Alexa Flour 594 (red) or 488 (green) were used (1:200 for all, Invitrogen). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was performed with a TUNEL kit (Promega) according to the manufacturer's directions to detect apoptotic cells. Slides were mounted with coverslips using Fluoromount-G or DAPI Fluoromount-G (SouthernBiotech).
Quantification of immunostaining
In oral mucosa, we quantified BrdU+, pH2AX-, or 8-hydroxy-2′-deoxyguanosine (8-OHdG)–positive cells as cells per mm basement membrane length (including all epithelial cells), TUNEL-positive cells as cells per mm basement membrane length (including all epithelial layers and stroma above the muscle layer), CD45-positve cells as DAPI-positive cells per mm2 epithelial and stroma area above muscle layer, nuclear pSmad2- or NF-κB p50–positive cells as the percentage of positive cells per total epithelial cell count (excluding sloughed epithelial cells induced by irradiation). Sequential 20x images along the basement membrane were quantified and averaged per sample. In oral cancer regions, K14- or K5-positive cells were defined as tumor cells. All tumor cells were measured together as the area of tumor (mm2), excluding stromal cells. We quantified percentage of nuclear pSmad2- or NF-κB p50–positive cells/total tumor (epithelial and stromal) cells. We quantified BrdU+, pH2AX+, 8-OHdG+, CD45+, and TUNEL+ cells as cells per mm2 tumor area. Ly6G+ or F4/80+ cells were quantified as cells per mm2 tumor area based upon staining intensity and morphology to distinguish them from nonspecific staining of other cell types. For pH2AX staining, cells with more than three nuclear foci were defined as pH2AX-positive cells. Five random 20x tumor images were quantified and averaged per sample.
Statistical analyses were performed using GraphPad Prism software. Normality was determined by D'Agostino and Pearson normality test (or Shapiro–Wilk normality test for smaller samples size, n < 8). Every dataset passed these normality tests; differences between two treatment groups were determined using the Student t test, and differences between more than two groups were determined by one-way ANOVA with Tukey multiple comparison test. Data are presented as the mean ± SEM.
Tat-Smad7 treatment reduced oral mucositis severity without protecting adjacent oral cancer
To mimic RT-induced oral mucositis in oral cancer patients, we used a human oral cancer orthotopic tongue xenograft model in athymic nude mice. Bone marrow transplant patients are immune suppressed and experience severe RT-induced oral mucositis (15), suggesting that T lymphocytes are not major contributors to mucositis. We transplanted two human oral cancer lines, FaDu or UM-SCC-1, into the tongues of athymic nude mice. The health of tumor-bearing mice rapidly deteriorated due to cancer burden and compromised oral intake. Therefore, after tumors were established, we irradiated mice with 18 Gy cranial RT instead of fractionated RT to minimize death due to repeated doses of anesthesia required for fractionated RT. This irradiation dose induces oral mucositis with the kinetics and severity similar to 3 × 8 Gy fractionated RT in mice with obvious mucosal damage starting on day 5 after RT (10). Six days after RT, when mice began to lose weight (10), we treated them with Tat-Smad7 produced as we previously described (ref. 10; Supplementary Fig. S1). We chose the dose and regimen of topical Tat-Smad7 application that effectively treated oral mucositis (1 μg/30 μL oral dose, daily). Vehicle treatment (30% glycerol in PBS) was used as control. Mice were sacrificed and tongues harvested 10 days after RT. We assessed if Tat-Smad7 protein was delivered to both oral mucosa and oral cancer with immunofluorescent staining using a human Smad7 antibody that also cross-reacts with mouse Smad7 (Supplementary Fig. S2). Similar to our previous report (10), in the oral epithelium of Tat-Smad7–treated mice, Smad7 was present in both the cytoplasm and the nucleus, whereas endogenous Smad7 levels in vehicle-treated mice were low (Fig. 1A). In tumors, Tat-Smad7 was detected in the cytoplasm of both tumor epithelia and stroma (Fig. 1A), suggesting that TGFβ in these cells drives Smad7 cytoplasmic translocation to block TGFβ signaling (16). We measured tumor area microscopically and found that both FaDu and UM-SCC-1 tumor sizes were reduced in RT-treated mice and Tat-Smad7 treatment did not significantly affect tumor size compared with vehicle in this short-course treatment (Fig. 1B and C). Tongue mucosa adjacent to tumor showed RT damage and mucositis formation primarily at the posterior dorsal surface which has fewer cornified layers compared with tongue papillae at the tip of the tongue (Fig. 1F and G). Tat-Smad7–treated mice had significantly smaller RT-induced tongue ulcers compared with vehicle-treated mice (Fig. 1D and E).
Histopathologic analysis revealed that irradiated epithelium exhibited atrophy, flattened tongue papillae, thinning cuticle, and ulceration compared with untreated epithelium (Fig. 1F and G). Irradiated mucosa treated with Tat-Smad7 showed less epithelial atrophy, smaller ulcerated area, and fewer immune cells compared with vehicle-treated mucosa (Fig. 1F and G; ref. 10). Compared with nonirradiated tumors, irradiated tumors near mucositis showed obvious cell damage with vacuolar degeneration and pyknotic nuclei or dysplastic cell shape with enlarged, smudged chromatin. Apoptotic cells, inflammation, and interstitial fibrosis were obvious in irradiated tumors. FaDu tumors were more sensitive to irradiation as demonstrated by more damaged tumor cells than UM-SCC-1 tumors (Fig. 1F and G).
Tat-Smad7 reduced NF-κB and TGFβ signaling in both oral mucositis and oral cancer
To examine on-target effects of Tat-Smad7 on oral mucosa and cancer, we performed immunostaining for pSmad2 and pSmad3, markers of TGFβ pathway activation, and nuclear NF-κB p50, a marker of NF-κB pathway activation. Oral mucosa after RT had more nuclear pSmad3+ cells compared with nonirradiated mucosa (Supplementary Fig. S3), similar to changes in pSmad2+ cells as previously reported (10). Fewer nuclear pSmad2+ and pSmad3+ cells were observed in irradiated oral mucosa and adjacent oral cancer in Tat-Smad7–treated mice compared with vehicle treatment (Fig. 2A, B, E, F, I, and J; Supplementary Fig. S3). In nonirradiated tongue tumors, both UM-SCC-1 and FaDu had more pSmad3+ cells than oral mucosa (Supplementary Fig. S3), indicating TGFβ-dependent Smad3 activation (even in Smad4-deficient FaDu tumors) as previously reported (13). In addition, fewer nuclear NF-κB p50+ cells were observed in both irradiated oral mucosa and adjacent oral cancer in Tat-Smad7–treated mice compared with vehicle treatment (Fig. 2C, D, G, H, K, and L). These data demonstrate the on-target efficacy of Tat-Smad7 against known Smad7 targets in tongue mucosa with oral tumors.
Tat-Smad7 treatment alleviated inflammation and DNA damage–associated cell death in oral mucositis lesions of oral cancer–bearing mice
We quantified leukocytes with CD45 staining and found Tat-Smad7–treated mucositis contained fewer CD45+ leukocytes (Fig. 3A and B). We next performed F4/80 and Ly6G staining to determine relative levels of F4/80+ macrophages and Ly6G+ cells [a marker of polymorphonuclear (PMN) neutrophils or PMN myeloid-derived suppressor cells (MDSC)]. RT-induced oral mucositis harbors primarily neutrophils and macrophages (9) that were obvious in oral mucositis lesions of tumor-bearing mice (Fig. 3A). Tat-Smad7–treated mucositis had fewer PMNs and macrophages than vehicle-treated lesions (Fig. 3A), consistent with smaller ulcers and fewer inflammatory cells observed by H&E and CD45 staining. To examine epithelial slough due to DNA damage–induced cell death and cessation of proliferation, we first performed pH2AX staining as a marker of DNA-damaged cells. Irradiated oral mucosa had numerous pH2AX+ cells, and Tat-Smad7–treated oral mucosa had fewer pH2AX+ cells compared with vehicle-treated mucosa (Fig. 3A and C). Next, staining of nuclear 8-OHdG was performed as a marker of oxidative DNA damage (17). There were fewer 8-OHdG+ cells in oral mucosa of Tat-Smad7–treated mice compared with vehicle-treated mice (Fig. 3A and D). We performed TUNEL assays to quantify apoptosis and noted numerous apoptotic cells in the oral mucosa of irradiated mice. Tat-Smad7–treated mucosa contained fewer apoptotic cells than vehicle control (Fig. 3A and E). Together, these data demonstrate that Tat-Smad7 treatment reduces inflammation, DNA damage, and apoptosis associated with RT-induced oral mucositis in tumor-bearing mice.
Tat-Smad7 reduced inflammation in irradiated oral cancer
Unlike nonirradiated oral mucosa that had no obvious inflammation (Fig. 1), leukocyte infiltration was apparent in nonirradiated tumors that contained numerous F4/80+ macrophages and Ly6G+ cells (Supplementary Fig. S4). Very few infiltrated B cells were detected (data not shown). In irradiated FaDu tumors, F4/80+ cells were still prominent, but Ly6G+ cells were largely diminished (Fig. 4A and C vs. Supplementary Fig. S4) even though Ly6G+ cells were numerous in adjacent mucositis (Fig. 3). Both F4/80+ cells and Ly6G+ cells were still pronounced in UM-SCC-1–irradiated tumors (Fig. 4D, F, and G). Intriguingly, Tat-Smad7 reduced the number of CD45+ leukocytes and F4/80+ macrophages in both types of irradiated oral tumors and reduced Ly6G+ cells in irradiated UM-SCC-1 tumors (Fig. 4).
Effects of Tat-Smad7 on irradiated oral cancer diverge from its effects on oral mucositis
Similar to oral mucositis, irradiated tumors had numerous pH2AX+ cells that were significantly more abundant than in nonirradiated tumors (Fig. 5A; Supplementary Fig. S5). FaDu tumors harbored more pH2AX+ cells than UM-SCC-1 tumors (Fig. 5D and E; Supplementary Fig. S5). However, unlike oral mucositis, there was no difference in the number of pH2AX+ cells in oral cancers treated with Tat-Smad7 versus vehicle in both tumor types (Fig. 5D and E). 8-OHdG+ cells were hardly detectable in nonirradiated tumors (Supplementary Fig. S5) but were induced significantly by RT (Fig. 5B). In contrast to Tat-Smad7 effects on oral mucositis, 8-OHdG+ cells in oral cancer were significantly increased in Tat-Smad7–treated mice compared with vehicle (Fig. 5F and G). Basal levels of TUNEL+ apoptotic cells were present in nonirradiated tumors (Supplementary Fig. S5) and RT induced numerous apoptotic cells in both tumor types (Fig. 5C). Apoptotic cells in oral cancer were increased in Tat-Smad7–treated mice compared with vehicle-treated (Fig. 5H and I), opposite of what was observed in Tat-Smad7–treated mucositis. We performed BrdU staining to quantitate proliferating cells. Adjacent to ulceration, Tat-Smad7–treated oral mucosa showed more proliferative BrdU+ cells along the epithelial basement membrane than vehicle-treated oral mucosa (Fig. 6A–D). In contrast, Tat-Smad7–treated oral cancer had no difference in proliferation compared with vehicle-treated oral cancer (Fig. 6A, B, E, and F). In summary, Tat-Smad7–treated mucositis lesions had reduced RT-induced DNA damage, inflammation, and apoptosis and increased keratinocyte proliferation, whereas adjacent irradiated tumors treated with Tat-Smad7 demonstrated no resolution to DNA damage, decreased inflammation, and unaltered proliferation.
One of the major challenges for treating oral mucositis in cancer patients is that potential therapies reviving normal mucosal regeneration could also pose risk by protecting cancer cells. We show here that Tat-Smad7 topical application reduced activation of NF-κB and TGFβ signaling and reduced inflammation in both oral mucosa and cancer, but differentially affected normal mucosa and oral cancer with respect to DNA damage and associated cell death, and cell proliferation.
Our current study confirms that in oral mucositis-induced acute inflammation associated with NF-κB and TGFβ activation (10, 18), macrophages and neutrophils are predominant leukocyte types (9). Fewer macrophages and neutrophils in Tat-Smad7–treated mucositis could reflect less damage (smaller ulcers and fewer dead cells) or a quicker resolution of inflammation. Oral cancer without RT showed chronic inflammation in the tumor microenvironment that harbored numerous F4/80+ putative M2 macrophages and Ly6G+ putative tumor-associated neutrophils or PMN-MDSCs; both are known to be tumor-promoting leukocytes attracted by TGFβ1 and also a major source of TGFβ1 (19). With RT, F4/80+ and Ly6G+ cells remained plentiful in UM-SCC-1 tumors, and Tat-Smad7 reduced numbers of these cells similar to TGFβ inhibitors and NF-κB inhibitors used in clinical trials of metastatic cancer, which has an inflammatory tumor microenvironment (20, 21). The subtypes of these leukocytes in irradiated tumors require further study, e.g., if they have shifted from PMN-MDSCs to neutrophils as part of acute inflammation induced by RT. Interestingly, Ly6G+ leukocytes were largely diminished in irradiated FaDu tumors despite their presence in the adjacent mucosa. Because FaDu cells lack Smad4 but retain nuclear pSmad2+ and pSmad3+, it is unclear if this altered TGFβ signaling affected RT response–associated changes in Ly6G+ cell infiltration or depletion in tumors. In contrast, Tat-Smad7 still reduced F4/80+ macrophages in FaDu tumors, possibly by blocking RT-induced TGFβ signaling through Smad2/Smad3 and by direct blocking of non–Smad-mediated NF-κB activation by TGFβ (22). These data highlight the complex biology and heterogeneity of tumors in response to RT. The anti-TGFβ/NF-κB effects of Smad7 could have a cascade effect on cytokine production directly or indirectly due to reduced leukocyte numbers. Therefore, future in-depth studies are needed to dissect subtypes of immune cells influenced by Smad7 as well as the production of cytokines by these cells.
Because reducing inflammation is insufficient to alleviate oral mucositis, other pathogenic processes must also be targeted for effective treatment. Among these processes, RT-induced DNA damage kills tumor cells but also initiates oral mucositis (1, 23). Therefore, reducing DNA damage or accelerating repair is key for oral mucositis healing, but such an effect is undesirable for eradicating tumor cells. The selective reduction of RT-induced DNA damage by Smad7 in oral mucosa (Fig. 3) might be explained by its ability to facilitate DNA damage repair through protein–protein interactions with the DNA repair protein complex containing ATM and Mre11-Rad50-Nbs1 proteins (24). However, Tat-Smad7 did not reduce RT-induced DNA damage in tumor cells (pH2AX+ cells). This may be part of the reason that intrinsic DNA repair defects in tumor cells (evident in nonirradiated tumors) cannot be reversed by Tat-Smad7. In addition, TGFβ-mediated DNA repair is an important mechanism of radioresistance in cancer treatment (25–29). In this context, blocking TGFβ signaling by Tat-Smad7 would enhance DNA damage in cancer cells. Further, RT-induced oxidative stress is required to amplify DNA damage, causing collateral damage leading to cancer cell death (30, 31) and oral mucositis (1, 23). Reduced oxidative stress marker 8-OHdG with Tat-Smad7 treatment in oral mucositis is consistent with previous reports that both TGFβ and NF-κB pathways can induce oxidative stress (32–35). Paradoxically, TGFβ is reported to reduce oxidative stress in a skin SCC model (36). This could explain why Tat-Smad7 treatment increased 8-OHdG+ cells in oral cancer. Given these effects, it is not surprising that DNA damage–associated apoptosis was reduced by Tat-Smad7 in oral mucositis but increased in adjacent oral cancer. The effect of Tat-Smad7 on apoptosis is also consistent with previous reports showing Smad7 reduced apoptosis in normal epithelia (10, 37) but enhanced apoptosis in cancer (38).
In a subset of HNSCCs, endogenous Smad7 is activated by TGFβ and NF-κB, which transcriptionally suppress Smad target genes that could either be tumor suppressive or promotive (22). Therefore, pharmacologic dosage of a Smad7-based biologic needs to be carefully assessed. Our study revealed that HNSCCs derived from SMAD4 mutant (FaDu) and SMAD4 wild-type (UM-SCC-1) cells responded to Tat-Smad7 similarly. This can be explained by the fact that the majority of human cancers escape Smad-dependent TGFβ growth inhibition, which is also supported by Tat-Smad7 not affecting cancer cell proliferation in our model. In addition, we have shown SMAD4-mutant epithelial cells rely primarily on Smad3-dependent signaling to mediate TGFβ1-induced inflammation (13). Therefore, the use of Tat-Smad7 in SMAD4-mutant cancer could follow the same principle as TGFβ inhibitor in treating SMAD4-mutant metastatic cancer including pancreatic cancer that has a high rate of SMAD4 mutation (39).
In summary, we provide evidence that local short-term Tat-Smad7 protein delivery alleviated RT-induced oral mucositis without compromising RT-induced killing of neighboring oral cancer. Potential mechanisms for the context-specific effects of Smad7 are as follows: In oral mucositis, Tat-Smad7 attenuates TGFβ-mediated growth arrest and apoptosis as well as TGFβ/NF-κB–mediated inflammation. Smad7 could also directly reduce DNA damage or accelerate repair (Fig. 6G). In neighboring oral cancer, Smad7 still attenuates TGFβ/NF-κB–mediated inflammation, but is insufficient to reduce DNA damage (Fig. 6H). In addition, Smad7 could block TGFβ-mediated suppression of oxidative damage and consequently induce apoptosis in tumor cells (Fig. 6H). Because high intensity/hypofractionated SBRT is increasingly used to treat oral cancers resistant to traditional low-dose fractionated RT, future studies that more closely mimic SBRT will be needed to assess therapeutic efficacy of Tat-Smad7 in oral mucositis treatment. In addition, HNSCC mouse models with an intact immune system (13, 40) will further determine if effects of Tat-Smad7 are influenced by T-cell–mediated tumor immunity. Future studies should carefully assess Tat-Smad7 dose-dependent efficacy to effectively treat oral mucositis but avoid potential oncogenic effects on neighboring oral cancer, and assess the effect of long-term Tat-Smad7 use on oral cancer progression with an intact immune microenvironment.
Disclosure of Potential Conflicts of Interest
D. Raben is a consultant/advisory board member for EMD Serono. X.-J. Wang is an employee of Veterans Affairs, reports receiving commercial research grants from NIH to Allander, and holds ownership interest (including patents) in Allander Biotechnologies. No potential conflicts of interest were disclosed by the other authors.
Conception and design: J. Luo, D. Raben, X.-J. Wang
Development of methodology: J. Luo, L. Bian, C. Liang, D. Du, H. Zhou, C.D. Young, X.-J. Wang
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Luo, L. Bian, M.A. Blevins, D. Wang, C. Liang, D. Du, F. Wu, B. Holwerda, R. Zhao
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Luo, L. Bian, C. Liang, D. Du, F. Wu, H. Zhou, C.D. Young, X.-J. Wang
Writing, review, and/or revision of the manuscript: J. Luo, L. Bian, M.A. Blevins, D. Raben, H. Zhou, C.D. Young, X.-J. Wang
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): F. Wu
Study supervision: R. Zhao, H. Zhou, X.-J. Wang
We thank Drs. Yosef Refaeli, Brian Turner, Steve Sonis, and Thomas Carey for advice and an anonymous donation for support of this work. D. Raben is supported by the Marsico Endowment fund for research.
This work was supported by NIH grant DE024659 to C.D. Young and X.-J. Wang, and DE015953 to X.-J. Wang. X.-J. Wang is also a Research Biologist in Department of Veterans Affairs. J. Luo and F. Wu were supported by the National Nature Science Foundation of China No. 81772898, and the State Scholarship Fund of China Scholarship Council No. 201506240122 and No. 201606240201. M.A. Blevins was supported by T32 CA174648.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.