Proton FLASH Radiotherapy Ameliorates Radiation-induced Salivary Gland Dysfunction and Oral Mucositis and Increases Survival in a Mouse Model of Head and Neck Cancer

Radiotherapy, chemotherapy, and surgical intervention are primary treatments for patients with head and neck cancer (HNC). Nearly 800,000 new cases of HNC are annually diagnosed worldwide (1, 2). With ongoing technical advances such as intensity-modulated radiation therapy (IMRT) and proton therapy, radiotherapy can be delivered in a more conformal manner, allowing targeted delivery of higher doses of radiation to cancer cells. Still, incidental damage to the underlying tissue remains a common and significant limitation of radiotherapy (3). One of the major complications following head and neck radiotherapy is xerostomia, a condition that manifests as a significant reduction in saliva flow and damages the secretory acinar cells (4). Patients undergoing radiotherapy for HNCs can also develop acute oral mucositis. This condition is characterized by severe damage of the mucosal epithelium and ulcerative lesions that lead to difficulty in swallowing and excessive pain, resulting in less oral food and liquid intake, dehydration, and weight loss (5). Other comorbidities of HNC radiotherapy include dysgeusia and dental complications (the most severe being osteoradionecrosis), any of which can negatively affect long-term patient quality of life (2, 6, 7).

Recent studies have reported that ultra-high dose rate “FLASH” radiotherapy decreases normal tissue toxicity while maintaining tumor control response compared with conventional dose rates used for patient treatments (8, 9, 10, 11). Unlike conventional or standard radiotherapy, which has a dose rate of 1–2 Gy/minute, FLASH radiotherapy is typically administered at a dose rate of >40 Gy/second (12). Most studies on FLASH radiotherapy have utilized electrons (8, 9). Compared with electrons, FLASH proton radiotherapy (F-PRT) provides deeper tissue penetration and a reduced penumbra and exit dose due to the Bragg peak, which does not exist in X-rays (11, 13, 14). Sparing of normal tissues by FLASH electron radiotherapy has been well established in the brain (15, 16), the gastrointestinal tract, and the skin (9, 11, 17). Although the mechanisms of the FLASH effect are not well understood, it is evident that these ultra-high dose rates can expand the therapeutic window of radiotherapy (12). Some of the mechanisms implicated include dose rate differences in hypoxia response, redox chemistry, tissue microenvironment (15, 18, 19, 20, 21), effect in stem cell proliferation (10, 21), and changes in inflammatory signaling cascade (8, 17, 22, 23).

The effect of F-PRT in providing equivalent tumor control compared with standard proton radiotherapy (S-PRT) has been documented in pancreatic adenocarcinomas, sarcomas, and head and neck squamous cell carcinomas (HNSCC; refs. 10, 11, 17). However, no studies so far have tested the normal tissue-sparing ability or tumor control capabilities of FLASH in orthotopic HNCs. Thus, in this study, we aimed to investigate the sparing effect of FLASH in radiation-induced salivary gland dysfunction and oral mucositis and in controlling orthotopic tumor growth in mice. The results from this study are the first to report that F-PRT ameliorates radiation-induced salivary gland dysfunction and oral mucositis while maintaining an equal tumor control response to that of S-PRT in an orthotopic HNC murine model.

Mouse oral squamous cell carcinoma cell line, MOC2 was a generous gift from Dr. Amit Maity's lab (Hunstman Cancer Institute, University of Utah, Salt Lake City, UT). MOC2 cells were cultured in Iscove's modified Dulbecco's medium MOC line media. Lentivirus-mediated transfection of MOC2 cells was done using LV-EF1α-fLuc-IRES-Puro (SignaGen Laboratories) and the transfected cells were selected using the antibiotic, puromycin (1 μg/mL). The Mycoplasma testing was done at Cell Culture Services core, University of Pennsylvania (Philadelphia, PA) after receiving MOC2 from Dr. Maity's lab and cell line authentication was done by IDEXX Impact II PCR Profile test.

Female and male C57BL/6J mice (Jackson Laboratory) ages 8–10 weeks old were used in the study. The mice were maintained in Association for Assessment and Accreditation of Laboratory Animal Care International–accredited facilities and all the procedures were approved by the Institutional Animal Care and Use Committee at the University of Pennsylvania (Philadelphia, PA). Mice were monitored daily and euthanized whether they demonstrated any overt signs of distress or were moribund or whether their weight decreased by 20% or more of the initial body weight.

Orthotopic head and neck tumors were generated by submucosal injection of 5 × 10⁴ MOC2-luc cells (suspended in a volume of 30 μL of PBS) directly into the lateral tongue using a 100 μL tuberculin syringe (Hamilton Co.). Seven days after tumor implantation, mice were randomized into different irradiated and non-irradiated groups. The assignment of mice in groups was based on the tumor bioluminescent signal that was acquired on an IVIS Spectrum Imager (Perkin Elmer). Briefly, mice were injected intraperitoneally with 150 mg/kg of d-Luciferin (IVISbrite D-Luciferin, RediJect, Perkin Elmer). At 10 minutes after injection (when tumor intake of luciferin maximizes), mice were anesthetized by 2% isoflurane and placed at supine position. Images were acquired selecting the autoexposure of the IVIS Acquisition Control Panel which automatically sets the exposure time, f/stop and binning to keep the signal within an optimal range for quantification. Equal-sized regions of interest were drawn to include the entire tumor bioluminescent signal and the total flux (photon/second) of the tumors was quantified using Living Image Software 4.7.3 (Perkin Elmer).

The head and neck regions of mice were irradiated either with a single dose of 14, 16, or 18 Gy of PRT or with a fractionated dose of 3 × 8 Gy delivered once every alternate day. The irradiated region contains both the salivary glands and the lower oral cavity. The mouse was positioned as shown in Fig. 1A. The irradiation was performed using a 230 MeV (range ∼32.0 g/cm²) proton beam generated by an IBA Proteus Plus Cyclotron. The fixed research beam line was used to deliver proton beams using a shoot-through technique (10). Alignment was performed with image guidance using the cone beam CT (CBCT) from the small animal radiation research platform (SARRP, Xstrahl Inc.) that is on rails and placed in front of the proton beam (24). The alignment of the proton beam isocenter and the CBCT isocenter was verified before each experiment.

A 12 × 20 mm² rectangular collimator was used to deliver the proton beam. EBT3 Gafchromic film (Ashland Advanced Materials) was used to verify field uniformity daily. The dose was measured before each experiment using a NIST-traceable Advanced Markus Chamber (PTW). An ion chamber placed in the beam path (Bragg peak chamber Type 34070, PTW) was cross-calibrated so that each mouse irradiated could have a recorded delivered charge corresponding to the dose.

During the image-guided alignment of the mouse (Fig. 1B), the isocenter was set near the bottom of the mouse so that the top of the field border was below the brain and sinus cavity. The left beam edge (from the beam's eye view) was set to avoid the esophagus. The irradiated region of the mouse was roughly 6 mm by 20 mm (Supplementary Fig. S1).

Following irradiations, the mice were given a dietary supplement (clear water recovery gel) for a period of consecutive 7 days and then returned to their normal diet. Mice were weighed twice per week (Supplementary Fig. S2).

Following irradiation, mice were anesthetized with 2% isoflurane and their saliva secretion was stimulated with an intraperitoneal injection of pilocarpine hydrochloride (0.5 mg/kg). Total saliva was collected from the mouths of anesthetized mice 2 minutes after pilocarpine injection using a pipette for 10 minutes in a preweighed tube. Following collection, the weight of the saliva-containing tubes was measured at each timepoint of the study.

To determine radiation-induced oral mucositis, the tongues of the mice of irradiated and non-irradiated groups were stained in a solution of 1% toluidine blue (TB) for 3 minutes. Following staining repeated wiping with an acetic acid–soaked swab was done until there was no further recovery of the dye. Deep blue staining in regions of epithelia indicated a positive result, whereas light or no dye uptake indicated a negative result (25).

The histology of harvested tongues, salivary glands, and tongues with orthotopic tumors of mice was done at the Comparative Pathology Core at the University of Pennsylvania School of Veterinary Medicine following the standard protocol (11). All the histopathology analyses were evaluated blindly.

Salivary gland tissues of 10-μm-thick sections were used to assess the expression of Aquaporin 5 (AQP5; ref. 26) by immunofluorescence (IF) staining following the standard protocol (11). Apoptosis of the salivary gland tissues was studied using terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay (In Situ cell death detection kit, TMR Red, Roche) following the manufacturer's protocol. Quantification of the IF images was conducted using Fiji image J. The quantified data are presented as the total stained area normalized to the total nuclei area.

Total RNA was isolated from the harvested submandibular glands using TRIzol Reagent (Invitrogen; Life Technologies) following standard protocol. The cDNA was synthesized using a high-capacity RNA-to-cDNA kit according to the manufacturer's instructions (Thermo Fisher Scientific). qRT-PCR was performed with TaqMan master mix. For data analysis, the QuantStudio 6 Flex Real-Time PCR System (Applied Biosystems) was used. The relative mRNA expression levels were defined using the delta-delta Ct method, following normalization to 18S rRNA. The TaqMan primer list is provided in Supplementary Table S1.

The expression of AQP5 was also studied by the Western blot technique, following standard protocol. The analysis of the bands obtained is done using Fiji image J software (27). The quantified data were normalized to that of the internal control protein Vinculin. The list of antibodies used is given in Supplementary Table S2.

Mouse heads were scanned by a μCT 45 micro CT scanner (SCANCO Medical AG) at 10.4 μm isotropic voxel size. The scanned images were three-dimensional reconstructed and analyzed by Dragonfly software (Object Research Systems Inc). Alveolar bone loss (ABL) was measured by adding the distance from the cementoenamel junction to the alveolar bone crest at mesial and distal sites of the second molar in sagittal slices as described previously (28).

Submandibular gland tissue homogenates were analyzed for cytokine levels using multiplex ELISA at the RIA Biomarkers Core of the Penn Diabetes Center, at the University of Pennsylvania (Philadelphia, PA). A total of eight different cytokines were studied.

Statistical analysis was carried out using R and GraphPad Prism V software. A t test was used to compare means, a Mann–Whitney test was used to compare the distribution and a log-rank test was used to compare survival. Mixed effects models are used to analyze repeated measures within an experiment. Data are presented as means ± SEM. P < 0.05 was considered significant.

Raw data generated in this study are available upon request from the corresponding author.

To test the potential sparing of F-PRT on normal tissues, mice were irradiated either with a single dose of 14 or 16 Gy or with a fractionated dose of 8 Gy × 3 fractions of F-PRT or S-PRT. Following irradiation, survival was studied for 60 or more days after radiotherapy for a single dose of PRT and 90 days for a hypofractionated dose of PRT. Female mice irradiated with a single dose of 14 Gy of PRT showed no sign of mortality and all the irradiated mice were alive at 90 days after radiotherapy when the experiment was terminated. However, female or male C57BL/6 mice treated with a single dose of 16 Gy of S-PRT experienced significantly greater mortality (P < 0.0001) compared with non-irradiated or F-PRT–treated mice (Fig 1C and E, respectively). By 20 days (about 3 weeks) after irradiation, 54.5% of the female mice died in the S-PRT–treated group compared with 9.1% in the F-PRT–treated group (Fig. 1C). Similar results were also seen in mice irradiated with a hypofractionated dose of PRT. Mice irradiated with a dose of 8 Gy × 3 of F-PRT showed significantly improved survival (P = 0.035) compared with ones treated with S-PRT (Fig. 1D). Histopathologic analysis of the deceased mice showed an increase in cell death and mitotic activity in the salivary glands, atrophy of the major and lingual salivary glands and an increase in severity of inflammatory infiltrates. These findings demonstrate that when delivered in FLASH mode, PRT displayed a significant advantage in morbidity and mortality compared with standard PRT.

One of the key toxicities in patients receiving head and neck irradiation is xerostomia. To investigate the impact of F-PRT compared with S-PRT on murine salivary glands after irradiation, pilocarpine-induced saliva production was measured as described previously (29). Saliva production was estimated at 5, 10, 14, 28, and 90 days after irradiation with a single dose of 14 Gy of S-PRT/F-PRT and at 5, 10, 14, and 28 days after irradiation with a single dose of 16 Gy of S-PRT/F-PRT, respectively. Following irradiation with 14 or 16 Gy, the salivary flow was initially reduced by both treatments. However, the F-PRT–treated mice showed improvement in salivary production post-14, 28, and 90 days after irradiation with a single dose of 14 Gy (Fig. 2A) and 10, 14, and 28 days after irradiation with a single dose of 16 Gy (Fig. 2B). Salivary production was also estimated following irradiation with a hypofractionated dose of 8 Gy × 3 fractions of PRT. No significant difference was observed at 14 and 30 days after irradiation. However, at 90 days after irradiation, F-PRT–treated mice showed that mean saliva production was significantly higher compared with S-PRT–treated mice, using the same hypofractionated regime (Fig. 2C).

To evaluate sex as a biological variable we studied pilocarpine-induced saliva production in C57BL/6 male mice after irradiation with a single dose of 16 Gy. Following irradiation, F-PRT–treated mice showed significant improvement in salivary production at 10, 14, and 30 days after irradiation (Supplementary Fig. S3).

Following PRT, submandibular glands were harvested and studied for histopathologic changes after irradiation with a single dose of 16 Gy. Compared with F-PRT, mice treated with S-PRT showed an increase in cell death (Fig. 2D and F) and increased mitoses (Fig. 2E and G) at 14 and 28 days (Supplementary Fig. S4) posttreatment. Inflammation was mostly mononuclear (lymphocytes and plasma cells) and neutrophilic when present in the affected salivary glands. At 60 days after irradiation, S-PRT–treated mice displayed an increase in the lingual gland atrophy score (30) in the salivary glands compared with those treated with F-PRT (Fig. 2H and I).

As hyposalivation is associated with radiotherapy-induced xerostomia, we sought to investigate the expression of AQP5 in the salivary glands after irradiation with S-PRT and F-PRT. Both IF and immunoblotting for AQP5 were conducted in mouse submandibular gland tissue at 2, 5, 10, and 14 days after irradiation. IF staining showed a significant decrease in AQP5 expression at 5 days (Fig. 3A and B) and 14 days (Fig. 3C and D) after irradiation with PRT. However, in comparison with S-PRT, the F-PRT–treated mice exhibited a significant preservation of the expression of AQP5 at both 5 (P = 0.003) and 14 days (P = 0.001) after irradiation. Consistent with the IF data, immunoblotting at 2, 5, 10, and 14 days after radiotherapy also showed significant downregulation of AQP5 expression in S-PRTs compared with F-PRT–treated mice (Fig. 3G). Interestingly, no significant difference in relative mRNA expression of AQP5 was observed at 2 days and 14 days after irradiation with S-PRT or F-PRT (Fig. 3E and F). These findings suggest that this water channel protein, which plays a critical role in saliva production, is somehow underexpressed or degraded at the protein level but not at the mRNA level in S-PRT–treated mice whereas F-PRT ameliorates this effect. The mechanism by which this occurs and whether it directly contributes to the observed hyposalivation phenotype is currently under investigation.

Fibrosis is associated with physiologic and pathologic dysfunction of both murine and human salivary glands. TGFβ1 has been associated with the profibrotic pathogenesis of after radiation salivary gland dysfunction, and Sjögren syndrome (31). The expression level of inflammatory cytokines was studied in murine submandibular glands irradiated with PRT using multiplex ELISA. The expression level of eight different cytokines (TGFβ1, IL4, IFNγ, IL10, IL1β, VEGF, IL2, and TNFα) was studied at 2, 10, and 28 days after irradiation (Fig. 4A), with only TGFβ1 being significantly different between the standard and FLASH irradiated groups at 2 days after irradiation. On day 28 after irradiation, the expression of most of the cytokines under study (with the exception of IL1β, and VEGF) was found to be significantly higher in S-PRT–treated compared with the F-PRT groups. However, on day 10, there were no significant differences observed in the expression of most of the cytokines under study between the two irradiated groups. Correspondingly, histopathologic analysis of submandibular gland tissues also showed an increase in inflammatory infiltrates on S-PRT–treated mice after 28 days of irradiation (Supplementary Fig. S4).

Because TGFβ1 is a key mediator of radiation-induced inflammatory reactions (8, 11), the expression of TGFβ1 was also studied using IF staining in tissue sections of submandibular glands 28 days after irradiation with PRT and in the control group (Fig. 4B and C). Mice treated with S-PRT exhibited a significantly higher expression of TGFβ1 (P = 0.001) compared with those being treated with F-PRT.

To investigate the potential mechanism behind the protective effect of F-PRT against mucosal ulceration, the tongues of the control and the irradiated groups of mice were harvested and stained with 1% TB at 5, 10, 14, and 28 days after irradiation. As shown in Supplementary Fig. S5, the deep blue staining with TB indicated tongue ulcers starting to appear at day 14, and showing more severity at day 28 after irradiation. S-PRT–treated mice demonstrated substantially stronger staining compared with F-PRT–irradiated mice. To further assess the extent of ulceration, blinded histopathology studies were conducted using the tongue sections. In contrast to F-PRT, the S-PRT–treated mice showed a significant increase in lingual gland atrophy scores (P = 0.047) at 28 (Fig. 5A and B) as well as 60 days after irradiation (Supplementary Fig. S6). Submucosal fibrosis, assessed on a Masson's trichrome stain of the tongue, was significantly increased at 28 (Fig. 5C and D) and 60 days after irradiation (Fig. 5E and F) in the S-PRT mice, compared with the F-PRT mice (P < 0.0001).

In patients with HNSCC undergoing radiation, radiographic bone loss is an early indicator of subsequent periodontal disease and tooth loss. The most common assay of measuring jawbone loss is by measuring the ABL level (32), which is defined as the distance between the cervical enamel and the alveolar bone crest of the mandible. Mice irradiated with a single dose of 16 Gy S-PRT showed a significant increase (P = 0.041) in the ABL level compared with the ones irradiated with F-PRT (Fig. 5G and I). Mice irradiated with a hypofractionated dose of 8 Gy × 3 also showed a significant (P = 0.053) increase in the ABL level (Fig. 5H and J). These data suggest that F-PRT not only reduces the toxicity leading to oral mucositis and saliva production but also mitigates the jawbone loss that is associated with the long-term effects of radiotherapy.

To evaluate the comparative effectiveness of F-PRT with S-PRT in a more relevant potential clinical scenario, orthotopic tongue tumors were generated using the MOC2 HNSCC cell line transfected with firefly luciferase and injected into the tongue of syngeneic C57Bl/6 mice. Following implantation and propagation of the tumor, a dose-escalation study with 14, 16, and 18 Gy of PRT was conducted (Fig. 6A) and survival was studied over a period of 90 days. The Kaplan–Meier plot (Fig. 6B) shows that the overall survival of the mice was increased significantly following irradiation with F-PRT at each dose level 14 Gy (P = 0.031), 16 Gy (P = 0.009), and 18 Gy (P = 0.032) compared with S-PRT.

In a separate cohort of mice with orthotopic tongue tumors, irradiated with 16 Gy, all the mice of the NR, S-PRT, and F-PRT-treated groups were euthanized on day 15 after implantation. Unfortunately, the luciferase signal from these tumors exhibited wide variations and did not track with observed tumor volumes. Therefore, we harvested the tongues with the embedded tumors and submitted them for IHC staining and histopathologic/morphometric analysis of tumor growth. Blinded evaluation of the hematoxylin and eosin (H&E) staining of the longitudinal tongue sections revealed that compared with the non-radiated (NR) group, both the S-PRT– and F-PRT–treated groups exhibited a significantly reduced tumor area. However, no significant difference was observed between the S-PRT– and F-PRT–treated groups (Fig. 6C, D, and E), suggesting that both S-PRT and F-PRT achieved similar tumor control in orthotopic head and neck murine models.

Salivary gland dysfunction and oral mucositis are two of the most common and severe side effects observed in patients receiving conventional radiotherapy for HNCs (2, 25). This study was undertaken to determine the toxicity-sparing effect of F-PRT on both radiation-induced side effects of salivary glands and on oral mucosa compared with that of S-PRT. Moreover, to determine the efficacy of F-PRT for clinical applications, we developed an orthotopic head and neck tumor model by injecting mouse oral squamous cell carcinoma (MOC2) into the dorsal tongue of the mice. Using the orthotopic tumor model, when compared with S-PRT, we found that F-PRT has similar antitumor efficacy, but with significantly improved overall survival. The antitumor efficacy of F-PRT has already been shown by us (10, 11) and others (8, 9, 33). Our results are consistent with those of Cunningham and colleagues (17), who conducted a study by growing flank tumors using both MOC1 and MOC2 cell lines. Along with orthotopic tumor controlling potency, our study is the first to report that F-PRT spares the normal tissue toxicity of the organs at risk after irradiation in the head and neck region in a mouse model. Thus, with these findings, our study offers a strong preclinical rationale to evaluate F-PRT in a clinical setting. Because the tumor-controlling potency of F-PRT and S-PRT were found to be similar, the improved survival of the mice with orthotopic tongue tumors treated with F-PRT is likely due to improved salivary gland function and reduced mucositis and preservation of mouse body weight. We believe our study is the first to demonstrate a survival benefit in mice implanted orthotopically with a head and neck tumor treated with FLASH radiation.

Importantly, we also observed improved survival in tumor-free mice with both a single-fraction and fractionated F-PRT. It should be noted that in another study by Yang and colleagues, photon (X-ray) irradiation, at a similar single dose (16.5 Gy), resulted in a much higher level of lethality compared with our S-PRT–treated group (25). This is likely due to two factors: First, in our study, we irradiate only the lower portion of the oral cavity and the salivary glands (Fig. 1), whereas Yang and colleagues irradiates the whole head area. Second, following PRT, we provide veterinary care to all the mice in the form of hydrating gel, to avoid dehydration and allow sufficient time for other toxicities (xerostomia, epithelial mucositis, etc.) to develop. Both single-dose and hypofractionated dose of F-PRT have been observed to preserve saliva production in both male and female C57BL/6 mice. In contrast, mice irradiated with a single dose of S-PRT were found to continue with a declining rate of saliva production even 3 months after irradiation with 14 Gy or with a hypofractionated dose of 8 Gy × 3 fractions. Because radiation-induced hyposalivation hampers the quality of life of patients with HNC, the significant improvement in saliva production after treatment with F-PRT suggests that this aspect of after radiotherapy management might be a key consideration for a clinical application. Although the cells of the salivary glands have slower turnover rates (33), the changes in saliva composition and quantity after radiotherapy give an indication of the acutely responding characteristics of the salivary gland tissue (34–38). However, the exact mechanism behind the reduced saliva secretion is unknown and it is still unclear whether the direct effects of radiation on the secretory acinar and ductal cells cause salivary gland dysfunction versus other factors such as radiation-induced inflammatory infiltration, increased capillary permeability, injury to the fine vascular structures, or interstitial edema (33). A few reports with photon radiation have shown that the expression of a water channel protein AQP5 is critical for saliva secretion (2, 39, 40). However, no studies have reported changes in response to proton irradiation. In the current study, we have shown that in mice treated with standard proton therapy, the expression of AQP5 was reduced at 2, 5, 10, and 14 days after irradiation while proton FLASH preserves AQP5 expression. co-IF study using antibodies for AQP5 and TUNEL showed no colocalization of AQP5 with the apoptotic marker (Supplementary Fig. S7), strongly suggesting that the salivary gland secretory cells are not lost due to PRT, but rather, there must be a reduction of AQP5 levels within live cells. The lack of significant changes in AQP5 mRNA levels indicates a likely posttranscriptional mechanism. It is possible that S-PRT may induce a more pronounced inhibition of AQP5 mRNA translation compared with F-PRT. Alternatively, the loss of AQP5 may involve an autophagic or mitophagic mechanism. Previous studies employing a salivary-specific conditional knockout model of autophagy reported a significant decrease in stimulated salivary flow rates following a single dose of targeted head and neck radiation (41, 42). Intriguingly, FLASH electron radiation has been shown to induce reduced membrane lipid oxidation compared with standard radiation (43). This would translate to reduced autophagic activation and therefore better preservation of salivary gland secretory activity, which might be behind the sparing effect of F-PRT on these glands. These studies pave the way for future investigations into the possible role of autophagy following irradiation with S-PRT/F-PRT in salivary gland function.

Moreover, in contrast to F-PRT, histopathology data suggested an increase in lingual gland atrophy scores in submandibular gland tissues of S-PRT–treated mice at 60 days after irradiation. Both cytokine analysis and histology staining of the submandibular glands also demonstrated the induction of inflammatory infiltrates in S-PRT–treated mice while the F-PRT–treated ones showed less induction of inflammatory infiltrates at 2 and 28 days after PRT. TGFβ1 is a key mediator of both acute and late toxicities after irradiation. Consistent with the previous reports from our group and others in cells and different tissues (8, 11, 23), we also observed that F-PRT induces less secretion of TGFβ1 at both acute and late phases after irradiation in the salivary gland tissues.

In addition to salivary gland dysfunction, we also evaluated the effect of proton FLASH on oral mucositis. Both humans and mice show similar pathogenesis of radiation-induced oral mucositis, composed of four phases: an initial inflammatory phase, an epithelial phase, an ulcerative phase, and a healing phase (5, 33). The ulceration induced by irradiation on the dorsal tongue has been observed in several studies (25, 44, 45). In our study, after irradiation with a single dose of 16 Gy of S-PRT, mice exhibited tongue ulceration starting at 14 days after radiotherapy as observed by deep blue staining with TB and by histopathology compared with those being irradiated with F-PRT. Histopathology study demonstrated that F-PRT-irradiated mice have significantly lower lingual gland atrophy scores at day 28 after radiotherapy compared with those treated with S-PRT. As a result of lingual gland atrophy, the papillae on the tongues are worn out affecting the function of the taste buds (46). Moreover, F-PRT reduced late fibrosis of the tongue epithelium as measured by average muscle layer thickness after staining with Mason Trichrome. Another novel finding of our study was the determination of the ability of F-PRT to ameliorate ABL in the jawbone after irradiation compared with S-PRT. Because osteoradionecrosis of the jaw is one of the most common and dreaded complications of head and neck radiotherapy in humans (47), this preclinical finding sets the stage for F-PRT to be considered in a clinical setting.

The current study has certain limitations that need to be addressed in the future. We have studied survival and saliva production in irradiated mice for a maximum of 3 months, which corresponds to early and intermediate toxicities. To fully determine the efficacy of F-PRT for controlling long-term toxicities, longer follow-up is required. Moreover, in the current study, we used a hypofractionated dose scheme of dose 8 Gy x 3 fractions. Additional hypofractionated dose schema, particularly those used for standard clinical patient care, need to be investigated in the future to fully assess the benefit of normal tissue sparing by FLASH radiation. In this study, we have used only one tumor model using MOC2 cells for HNC. However, our group and many other groups have already reported the effect of F-PRT in providing equivalent tumor control compared with S-PRT in pancreatic adenocarcinomas (10), sarcomas (11) and HNSCCs (17).

In conclusion, we have demonstrated that F-PRT ameliorates multiple pathophysiologic toxicities associated with irradiation in the head and neck region of mice, particularly salivary gland dysfunction and oral mucositis. F-PRT ameliorates induction of TGFβ1 thus reducing both acute toxicity and late fibrosis. Moreover, by employing an orthotopic head and neck tumor model we showed that F-PRT increases overall survival in a dose-escalation study while having similar tumor-controlling efficacy as that of S-PRT. We believe that these results support a future clinical trial of this promising modality for the management of head and neck malignancies.

I.I. Verginadis reports consulting fees from Mevion Medical Systems, honoraria from the University of Arkansas, and travel expenses support for attending FRPT. M.M. Kim reports grants from NIH during the conduct of the study. E.S. Diffenderfer reports grants from IBA and NIH during the conduct of the study. M. Putt reports grants from NIH-NCI during the conduct of the study. L. Qin reports grants from NIH/NIA during the conduct of the study. S. Girdhani reports other support from IBA during the conduct of the study; in addition, S. Girdhani has a patent for METHODS OF USE OF ULTRA-HIGH DOSE RATE RADIATION AND THERAPEUTIC AGENT issued to Varian. F. Vander Stappen reports personal fees from Ion Beam Applications during the conduct of the study; personal fees from Ion Beam Applications outside the submitted work. K.A. Cengel reports grants from NIH during the conduct of the study; other support from IBA and Simphotek outside the submitted work; in addition, K.A. Cengel has a patent for LET dependent radiosensitization issued. T.M. Busch reports other support from IBA during the conduct of the study; other support from Simphotek, Inc and personal fees from IBA outside the submitted work. J.M. Metz reports personal fees from IBA and Varian during the conduct of the study. L. Dong reports grants from NIH during the conduct of the study. A. Lin reports personal fees from Galera Therapeutics and Janssen Pharmaceuticals outside the submitted work. C. Koumenis reports grants and personal fees from IBA during the conduct of the study; and scientific founder of Veltion Therapeutics. This start-up company does not engage in any activities relevant to the work described here. No disclosures were reported by the other authors.

P. Chowdhury: Conceptualization, data curation, formal analysis, validation, methodology, writing–original draft. A. Velalopoulou: Data curation, formal analysis, validation, methodology, writing–original draft. I.I. Verginadis: Data curation, formal analysis, validation, methodology, writing–original draft. G. Morcos: Data curation, formal analysis, methodology, writing–original draft. P.E. Loo: Data curation, formal analysis, methodology. M.M. Kim: Conceptualization, data curation, visualization, methodology, writing–original draft. S.A.O. Motlagh: Data curation, methodology. K. Shoniyozov: Data curation, methodology. E.S. Diffenderfer: Visualization, methodology. E.A. Ocampo: Data curation, formal analysis, writing–original draft. M. Putt: Software, formal analysis, validation, visualization, methodology, writing–original draft. C.-A. Assenmacher: Data curation, formal analysis, validation, methodology, writing–original draft. E. Radaelli: Software, formal analysis, validation, visualization, methodology, writing–original draft. J. Lu: Data curation, formal analysis, methodology, writing–original draft. L. Qin: Data curation, formal analysis, validation, methodology, writing–original draft. H. Liu: Data curation, methodology. N.M. Leli: Data curation, methodology. S. Girdhani: Funding acquisition, visualization, methodology. N. Denef: Funding acquisition, visualization, methodology. F. Vander Stappen: Funding acquisition, visualization, methodology. K.A. Cengel: Conceptualization, supervision, funding acquisition, investigation, visualization. T.M. Busch: Conceptualization, supervision, funding acquisition, investigation, visualization. J.M. Metz: Conceptualization, supervision, investigation, visualization. L. Dong: Conceptualization, supervision, funding acquisition, investigation, visualization. A. Lin: Conceptualization, supervision, investigation, visualization, writing–original draft. C. Koumenis: Conceptualization, data curation, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft.

We acknowledge the service provided through the Small Animal Imaging Facility and the Cell and Animal Irradiation (CARC) Core at the University of Pennsylvania. The CARC core RRID is SCR_022377. We would also like to recognize the service provided through Penn Vet Comparative Pathology Core, Department of Pathobiology, School of Veterinary Medicine, University of Pennsylvania.

This work was supported by a sponsored Research Agreement from IBA to perform preclinical research on FLASH Proton Radiotherapy and partially supported by NIH 5P01CA257904-02. P.E. Loo and E.A. Ocampo thank SUPERS@PENN hosted by the Department of Radiation Oncology, University of Pennsylvania. The program is funded by the NIH 5R25-CA140116-13.