CERE-120 Prevents Irradiation-Induced Hypofunction and Restores Immune Homeostasis in Porcine Salivary Glands

Abstract

Salivary gland hypofunction causes significant morbidity and loss of quality of life for head and neck cancer patients treated with radiotherapy. Preventing hypofunction is an unmet therapeutic need. We used an adeno-associated virus serotype 2 (AAV2) vector expressing the human neurotrophic factor neurturin (CERE-120) to treat murine submandibular glands either pre- or post-irradiation (IR). Treatment with CERE-120 pre-IR, not post-IR, prevented hypofunction. RNA sequencing (RNA-seq) analysis showed reduced gene expression associated with fibrosis and the innate and humoral immune responses. We then used a minipig model with CERE-120 treatment pre-IR and also compared outcomes of the contralateral non-IR gland. Analysis of gene expression, morphology, and immunostaining showed reduced IR-related immune responses and improved secretory mechanisms. CERE-120 prevented IR-induced hypofunction and restored immune homeostasis, and there was a coordinated contralateral gland response to either damage or treatment. CERE-120 gene therapy is a potential treatment for head and neck cancer patients to influence communication among neuronal, immune, and epithelial cells to prevent IR-induced salivary hypofunction and restore immune homeostasis.

Keywords: AAV-neurturin; CERE-120; fibrosis; gene therapy; head and neck cancer; humoral immunity; innate immunity; irradiation damage; salivary gland; salivary hypofunction.

Graphical abstract

Figure 1

CERE-120 Treatment Pre-IR, as Compared to Post-IR, Prevents Long-Term Hyposalivation in Murine SMGs. (A) Structure of the CERE-120 vector and the AAV2-GFP vector. (B) Schematic illustration showing the two arms of the mouse study design. (C–E) Pilocarpine-stimulated whole saliva was measured at 90 (C), 120 (D), or 300 (E) days post-IR in AAV2-GFP (10¹⁰ vp/g) or CERE-120 (10⁶, 10⁸, and 10¹⁰ vp/g)-treated animals and compared to non-IR mice (baseline). Saliva flow normalized to non-IR and shown as %. Mean ± SEM. N = 5–10 mice. Dots represent saliva measurements of individual mice. ANOVA with post hoc Dunnett’s test, ∗∗∗p < 0.001; ∗∗p < 0.01, ∗p < 0.05, not significant (ns) compared to non-IR.

Figure 2

Gland Anatomy and Morphology Improves after CERE-120 Pre-IR Treatment. (A–C) Analysis of the body weight (g) (A), submandibular gland weight (mg), (B) and normalized ratio of gland weight to body weight (mg/g) (C) at 300 days of non-IR mice, and mice treated with AAV2-GFP (10¹⁰ vp/g) or CERE-120 (10⁶, 10⁸, and 10¹⁰ vp/g) pre-IR. Dots represent measurement of individual mice. Mean ± SEM. N = 3–10 mice. ANOVA with post hoc Dunnett’s test, ∗∗∗p < 0.001, ∗p < 0.05, not significant (ns) as compared to non-IR. (D) H&E and Masson’s trichrome (MT) staining of SMGs of non-IR mice, and mice treated with AAV2-GFP (10¹⁰ vp/g) or CERE-120 (10¹⁰ vp/g) pre-IR. Images are representative of results from N ≥ 3 mice. Scale bar, 50 μm.

Figure 3

CERE-120 Treatment Reduces Levels of IR-Induced Fibrosis, Immune Response, and Innervation and Increases Proacinar Marker SMGc. (A) Fold changes in gene expression of stromal/ECM remodeling, immune, and epithelial-related genes in AAV2-GFP (10¹⁰ vp/g) and CERE-120 (10⁸ vp/g) pre-IR treated glands. Data were normalized to Gapdh and non-IR glands (dotted line). Mean ± SEM, N > 3. ANOVA with post hoc Dunnett’s test. ∗∗p < 0.01, ∗p < 0.05, as compared to non-IR. (B–G) Confocal imaging of SMGs of non-IR, AAV2-GFP 10¹⁰ vp/g, and CERE-120 10⁸ vp/g immunostained for (B) SMGc (magenta), E-cadherin (ECAD) (green), and nuclei (blue). Scale bar, 10 μm. (C) Quantification of SMGc and ECAD staining. (D) Immunostained for MMP2 (red), ECAD (green), and nuclei (blue). Scale bar, 20 μm. (E) Quantification of MMP2 and ECAD staining. (F) Immunostained for TUBB3-expressing neurons (green) and ECAD (red). (G) Quantification of TUBB3 and ECAD. All staining was normalized to the non-IR SMGs and nuclei. (C–G) Graphs show mean ± SEM. N > 3 glands with at least 5 images collected per gland. ANOVA with post hoc Dunnett’s test. ∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05 compared to non-IR.

Figure 4

CERE-120 Treatment Pre-IR Does Not Increase the Weight of the IR PG, but the Contralateral Gland Is Similar to Non-IR Control. (A) Schematic illustration of the study design with 52 animals. AAV2-GFP or CERE-120 gene therapy was delivered to one PG in each animal pre-IR. One week later, animals in the IR group received 15 Gy of radiation to the treated gland. Saliva was collected bi-weekly until 16 weeks, when tissues were harvested for analysis. (B) Representative images of the parotid salivary gland at the 16-week time point from various groups (non-IR-GFP, IR-GFP, and IR-CERE [10⁹ and 10¹¹]). Scale bar, 5 cm. (C–G) Graphs representing animal weight (C), gland weight of the treated side (D), normalized treated gland weight (data from C/data from D) (E), gland weight of the contralateral non-treated side (F), and normalized contralateral non-treated gland weight (data from F/data from D) (G). Mean ± SEM. Dots represent data from individual animals. N > 4. ANOVA with post hoc Dunnett’s test, ∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05, not significant (ns), as compared with non-IR-GFP.

Figure 5

IR-Induced Alterations in Acinar Morphology and KRT19+ Ducts in Porcine PGs Are Prevented by CERE-120 Treatment before IR. (A and B) Representative pictures of H&E and Masson trichrome (MT) staining on glands from various groups (non-IR-GFP, IR-GFP, and IR-CERE 10⁹ groups). Scale bar, 50 μm. (C) Single 2 μm confocal sections of glands stained for Keratin 19 (red). Scale bar, 100 μm. (D) Pictures of MT staining showing enlarged acinar-tubular structures with IR-GFP. Scale bar, 40 μm. (E) Confocal imaging of AQP5 (red), ECAD (green), and nuclei (blue) highlight the large luminal membranes of the enlarged acini structures. Scale bar, 20 μm. (F) Maximum intensity projections of confocal sections of glands stained for AQP5 (red), SMA (green), and nuclei (DAPI). Scale bar, 10 μm. Arrows indicate the diameter of an individual acinus. (G) Graph depicting the area of acinus diameter in glands from different treatment groups. Mean ± SEM. N > 3. ANOVA with post hoc Dunnett’s test, ∗∗∗p < 0.001, as compared to non-IR-GFP. (H) Gene expression of acinar and ductal-related markers. Mean ± SEM. N > 4. ANOVA with post hoc Dunnett’s test compared to non-IR-GFP.

Figure 6

IR-Induced Changes in Neuronal, ECM Remodeling, and Immune Cells Are Restored by CERE-120 Treatment before IR in Minipigs. (A) Representative confocal images of PGs from non-IR-GFP, IR-GFP, and IR-CERE (10¹¹, 10¹⁰, and 10⁹ vp/g) stained for TUBB3 (yellow), KRT19 (red), Peanut Agglutinin (PNA) (gray), and nuclei (blue). The TUBB3 channel is shown in the lower panels for clarity. Scale bar, 100 μm. (B) Quantification of immunostaining showing fold increase in staining compared to non-IR-GFP. Graphs show mean ± SEM. N > 3 pigs with a minimum of 3 images collected per sample. ANOVA with post hoc Dunnett’s test. ∗∗∗p < 0.001, ∗p < 0.05 compared to non-IR-GFP. (C) Fold changes in gene expression of neuronal genes, TUBB3, GFRA2, and TH; genes involved in ECM remodeling and fibrosis, SERPING1 and MMP2; and immune-related genes, C3 and LYZ. Data were normalized to RPS29 and non-IR glands (dotted line). Graphs show mean ± SEM. N > 3. ANOVA with post hoc Dunnett’s test. ∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05 compared to non-IR. (D) Confocal images of PGs from non-IR-GFP, IR-GFP, and IR-CERE 10¹⁰ vp/g, stained for the leukocyte marker CD45 (green), ECAD (magenta), and nuclei (blue). Scale bar, 10 μm. (E) Quantification of CD45 normalized to the non-IR-GFP PGs. Graphs show mean ± SEM. N > 3. ANOVA with post hoc Dunnett’s test. ∗∗p < 0.01 compared to non-IR.

Figure 7

Low Doses of CERE-120 Gene Therapy Pre-IR Prevents IR-Induced Hyposalivation in Porcine PGs. (A and B) Pilocarpine-stimulated saliva collected from the treated (A) and contralateral (B) parotid salivary gland was collected every 2 weeks. Saliva collected from 2–16 weeks was averaged and normalized to the baseline (dotted line, 100%). Mean ± SEM. N = 4–8 animals per group. ANOVA with post hoc Dunnett’s test. ∗p < 0.05, ∗∗∗p < 0.001, not significant (ns) when compared with non-IR-GFP group. (C) Time-course of saliva flow for various treatment groups. Black line, AAV2-treated gland. Gray line, non-IR non-treated contralateral gland. The gray area with dotted line represents ≤50% reduction in flow from baseline (100%). Mean ± SEM. N = 4–8 animals per group.