Chemically modified hCFTR mRNAs recuperate lung function in a mouse model of cystic fibrosis

Gene therapy has always been a promising therapeutic approach for Cystic Fibrosis (CF). However, numerous trials using DNA or viral vectors encoding the correct protein resulted in a general low efficacy. In the last years, chemically modified messenger RNA (cmRNA) has been proven to be a highly potent, pulmonary drug. Consequently, we first explored the expression, function and immunogenicity of human (h)CFTR encoded by cmRNAhCFTR in vitro and ex vivo, quantified the expression by flow cytometry, determined its function using a YFP based assay and checked the immune response in human whole blood. Similarly, we examined the function of cmRNAhCFTR in vivo after intratracheal (i.t.) or intravenous (i.v.) injection of the assembled cmRNAhCFTR together with Chitosan-coated PLGA (poly-D, L-lactide-co-glycolide 75:25 (Resomer RG 752 H)) nanoparticles (NPs) by FlexiVent. The amount of expression of human hCFTR encoded by cmRNAhCFTR was quantified by hCFTR ELISA, and cmRNAhCFTR values were assessed by RT-qPCR. Thereby, we observed a significant improvement of lung function, especially in regards to FEV0.1, suggesting NP-cmRNAhCFTR as promising therapeutic option for CF patients independent of their CFTR genotype.

Cell culture and Transfection. CFBE41o− and 16HBE14o-cells (from Gruenert's lab) were incubated at 37 °C in a humidified atmosphere containing 5% CO 2 until they reached 80-90% confluency. Cell lines were washed with cold, sterile PBS and detached by Trypsin-EDTA. Trypsinization was stopped by adding minimum essential medium (MEM; www.thermofisher.com) containing 10% fetal calf serum. Cells were collected and spun down at 500 × g for 5 minutes before resuspension in fresh MEM. One day before transfection, 250,000 cells/well/1 ml were plated in 12-well plates and grown overnight in MEM without antibiotics. At confluence of 70-90%, cells were then transfected with 1000 ng (c)mRNA hCFTR or equivalent (in nmol) pDNA hCFTR using Lipofectamine 2000 (www.invitrogen.com) following the manufacturer's instructions and after changing the media to the reduced serum media, Opti-MEM (www.thermofisher.com). After 5 hours, the complexes were removed by replacement with fresh culture medium. Cells were kept for 24 h and 72 h before further analyses.
Flow cytometry analyses. All flow cytometry analyses were performed using a BD LSR Fortessa X-20 SORP (www.bdbioscience.com). For detection of hCFTR protein in 16HBE14o-and CFBE41o− cell lines, cells were transfected as described above and subsequently prepared for intracellular staining using a Fixation/ Permeabilization Solution Kit as directed in the manufacturer's instruction (www.bdbioscience.com). As primary antibody mouse anti-human hCFTR clone 596 (1:500, kindly provided by the cystic fibrosis foundation therapeutics Inc.) has been used. As secondary antibody served Alexa Fluor 488 goat anti-mouse IgG (1:1,000, www. lifetechnologies.com). At least 20,000 gated cells per tube were counted. Data were analyzed with FlowJo software, version 10.

Western blot analysis. Protein isolated from cell lines was separated on Bolt NuPAGE 4-12% Bis-Tris
Plus gels and a Bolt Mini Gel Tank (all from www.lifetechnologies.com). Immunoblotting for hCFTR was performed by standard procedures according to the manufacturer's instructions using the XCell II Mini-Cell and blot modules (www.lifetechnologies.com). After blocking for 1 h in 5% dry milk at room temperature, primary antibody against hCFTR clone 596 (1:500, kindly provided by the cystic fibrosis foundation therapeutics Inc.) or anti-GAPDH (1:1000) (www.scbt.com) was incubated overnight, horseradish peroxidase-conjugated secondary antibodies (anti-mouse from www.dianova.com) were incubated for 1 h at room temperature. All blots were processed by using ECL Prime Western Blot Detection Reagents for 30 min exposure time (www.gelifesciences. com). Semiquantitative analysis was performed using the ImageJ software and overexposure has been avoided as per as digital image and integrity policies.
Immunofluorescence. CFBE41o− and 16HBE14o-were plated on a cell culture insert (0.75 × 10 6 cells per insert) containing a PET membrane (0.4 μm pore size) (www.corning.com) to provide an air-liquid interface. Cells were transfected 12 h after plating with 5000 ng cmRNA hCFTR or equivalent (in nmol) pDNA hCFTR using Whole blood assay. Ethical approval for using whole blood from healthy donor was obtained from Ethics Commission University Clinic of Tuebingen, Germany (349/2013BO2) and experiments were conducted in accordance with relevant guidelines and regulations. Informed consent form (following WHO guideline) was signed by each volunteer (healthy donor) and kept safely by principal investigator for privacy requirement. Blood samples from three healthy donors were collected in EDTA collection tubes (www.sarstedt.com). For each treatment group, 2 ml of EDTA-blood was transferred into 12-well plates and treated accordingly. R848 (Resiquimod, www.sigmaaldrich.com) was added at a concentration of 1 mg/ml to the respective blood positive controls. cmR-NA hCFTR and pDNA hCFTR (7 picomol each) were complexed to NPs at a ratio of 1:10. The samples were kept at 37 °C incubator maintaining 5% CO 2 . At 6 h and 24 h, 1 ml of whole blood was transferred into microtubes containing serum gel (www.sarstedt.com) and spun down at 10,000 × g for 5 min to obtain serum. Sera were stored at −20 °C for further cytokine measurement analyses. Serum was used to conduct IFN-α, TNF-α and IL-8 ELISA at manufacturer's instruction (www.thermofisher.com).
Animal experiments. All   Pulmonary mechanics. Lung function for each group was evaluated using a FlexiVent ® equipped with FX1 module and NPFE extension and was operated by the flexiWare v7.2 software (www.scireq.com). Prior to tracheostomy, mice were anesthetized intraperitoneally as described above. After anesthesia, a 0.5 cm incision was performed in rostral to caudal direction. A flap of skin was retracted, the connective tissue was dissected, and the trachea was exposed. The trachea was then cannulated between the second and third cartilage ring with a blunt-end stub adapter. The mouse was connected to the FlexiVent ® system and quasi-sinusoidally ventilated 27 with a tidal volume of 10 ml/kg. A breathing frequency of 150 breaths per min was maintained with an inspiratory to expiratory ratio of 2:3. Airway resistance (Rn), which is dominated by the resistance of the large conducting airways was considered in this study when the coefficient of determination of the model fit was ≥0.9. Compliance (Cst) was calculated straight from deflating arm of the pressure volume (PV) loops and ramp style pressure-driven maneuver (PVr-P). For obtaining FEV 0.1 data a NPFE maneuver was performed which results in FV loops and FE-related parameters. The mice lung was inflated by a pressure of +30 cmH 2 O over 1.2 s and rapidly deflated to a negative pressure of −55 cmH 2 O to generate an imposed negative expiratory pressure gradient.
Salivary assay. Prior to tracheostomy, anesthetized mice were injected with 50 µl of 1 mM acetylcholine (ACh) in the cheek to stimulate the production of saliva. The fluid was collected via glass capillaries and a chloride assay was performed using the Chloride (Cl − ) Assay Kit according to the manufacturer's protocol (www.sigmaaldrich.com). Briefly, saliva was diluted at a ratio of 1:100 with water in a total volume of 50 µl and subsequently 150 µl chloride reagent was added. After 15 min incubation at room temperature in the dark, absorbance was measured at 620 nm using an Ensight Multimode plate reader (www.perkinelmer.com).
injection of differently modified cmRNA hCFTR in Cftr −/− mice (CFTR tm1Unc ), the lungs were isolated at day 6 (experimental endpoint), homogenized and lysed in 600 µl RIPA-buffer and 5 µl protease inhibitor cocktail with tubes of the Precellys Ceramic Kit 1.4/2.8 mm at 6,500 rpm for 10 s for a total of three cycles, each interrupted by a 15 s break in a Precellys Evolution Homogenizer for protein isolation (all from www.peqlab.com). Subsequently, supernatants were kept on ice and additionally homogenized 10 times with a 20G needle and incubated for 20 min (www.bdbioscience.com). Lysates were spun down for 20 min at 13,000 × g and 4 °C. The supernatant was collected and stored at −20 °C for further use. Prior to hCFTR ELISA detection, protein concentration was measured using the Pierce BCA protein assay kit (www.thermofisher.com). For each sample, an equal amount of 15 µg whole protein lysate was used. A human CFTR ELISA kit (www.elabscience.com) was used for hCFTR detection according to manufacturer's instructions.
Real-time RT-PCR. After i.t. or i.v. injection of cmRNA hCFTR the lungs were isolated at day 6 (experimental endpoint), homogenized and lysed with tubes of the Precellys Ceramic Kit 1.4/2.8 mm at 5,000 rpm for 20 s in a Precellys Evolution Homogenizer for subsequent RNA-isolation (all from www.peqlab.com). Reverse transcription of 200 ng RNA was carried out using an iScript cDNA synthesis kit (www.bio-rad.com) and 1:20 dilution of the cDNA product had been used for further experiment. Detection of mRNA hCFTR was performed by SYBR-Green based quantitative Real-time PCR in 15 μl reactions on a ViiA7 (www.lifetechnologies.com). In all involved procedures, we strictly followed the MIQE protocols for RealTime experiments 28 . Pre-and post-reaction rooms were strictly separated. Reactions were incubated for 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 2 min at 50 °C (annealing and extension), followed by standard melting curve analysis. The following primer pairs were used: hCFTR fwd 5′-GAGATGCTCCTGTCTCCTGG-3′, rev 5′-CCTCTCCCTGCTCAGAATCT-3′; 18S rRNA fwd 5′-GGGAGCCTGAGAAACGGC-3′, rev 5′-GACTTGCCCTCCAATGGATCC-3′. Differences in mRNA expression between groups were analyzed by pair-wise fixed reallocation randomization tests with REST 2009 software after collection of the data from Viia7.

Immune response in vivo.
To assess immune responses to (c)mRNA hCFTR and pDNA hCFTR , C57BL/6 (Jackson Laboratory (www.jax.org)) mice (n = 4 per group) were treated as described for Cftr −/− mice. As positive controls a group of mice received two administrations of E. coli mRNA-NPs (20 µg) i.v. or i.t. C57BL/6 mice received two injections of 20 µg cmRNA hCFTR complexed to NPs i.v. or i.t. After 6 h, 24 h, and 72 h of second injection mice were sacrificed and blood was collected. For cytokine measurement, blood from mice was used to obtain serum using a serum separator (www.sarstedt.com) and tested for IFN-α and TNF-α production as directed in the manufacturer's instructions (www.thermofisher.com).
Statistics. All analyses were performed using the Kruskal-Wallis test with GraphPad Prism Version 6 (www. graphpad.com). Most of the data are represented as mean ± SD; box plot data are represented as a mean ± minimum to maximum values. P ≤ 0.05 was considered statistically significant.

(c)mRNA hCFTR and hCFTR protein quantification in vitro.
To evaluate the influence of chemical nucleoside modification, we first conducted a set of in vitro analyses to characterize the expression and functionality of hCFTR protein. First, we compared the expression profile of plasmid-encoded hCFTR (pDNA hCFTR ), unmodified hCFTR mRNA (mRNA hCFTR ) and two well-defined nucleoside modifications (cmRNA CFTR To confirm and substantiate those findings, we performed Western blot analyses of protein lysates taken from transfected CFBE41o-cells at 24 h and 72 h post treatment (Fig. 1C). As a positive control served protein lysate from untransfected 16HBE14o-cells, and GAPDH was used to normalize band intensities. At Fig. 1C). This drastic increase of hCFTR expression after pDNA transfection goes well in line with the observations in flow cytometry. As well as the quick All bar graph data are depicted as means ± SDs while box plots data are depicted as the means ± minimum to maximum values. *P ≤ 0.05 versus unmodified mRNA hCFTR ; § P ≤ 0.05 and § § P ≤ 0.01 vs. pDNA hCFTR . 1 0 resulting in the most robust hCFTR expression among all (c)mRNA transfections (All the blots are separately provided in Supplement Fig. S4).
All in vitro results are also underlined by the conducted immunofluorescence imaging. All tested samples show a higher amount of hCFTR positive cells compared to the negative control (CFBE41o-cells; Fig. 2A). Additionally, transfection with unmodified mRNA hCFTR produced a lower amount of hCFTR positive cells compared to both pDNA hCFTR and cmRNA hCFTR with the highest amount of hCFTR positive cells in the samples transfected with cmRNA CFTR . (Fig. 2A). Looking at the fluorescence image itself transfection of pDNA hCFTR shows a quite dispersed appearance of hCFTR within the cells compared to cmRNA hCFTR transfection seeming to have a higher abundance of hCFTR towards the cell membrane ( Fig. 2A, left panel). In general, the CFBE41o-cells compare to untransfected CFBE41o-and 16HBE14o-cells. Image J has been used for calculating means ± SDs of hCFTR positive cells; (B) Quenching efficacy of pDNA hCFTR or (c)mRNA hCFTR transfected CFBE41o-and CFTR null A549 cells relative to un-transfected controls was measured at 24 h, 48 h and 72 h post-transfection. *P ≤ 0.05 versus un-transfected controls; (C) 2 ml whole blood, each from three different healthy human donors, were incubated with either R848 (1 mg/ml) or 7 pmol pDNA hCFTR or 7 pmol (c) mRNA hCFTR (providing the same total number of nucleic acid molecules) and NPs at a 1:10 ratio; after 6 h and 24 h the immune response was determined by ELISA in the sera; The blue area represents the variance of the negative controls which are biological replicates. n.d., not detectable and red dotted lines mark the detection limit as specified in the respective ELISA kit. All bar graph data are depicted as means ± SDs while box plots data are depicted as the means ± minimum to maximum values. *and § P ≤ 0.05 ( § § P ≤ 0.01) versus control at 6 h and 24 h, respectively. Immunofluorescence imaging confirms that transfection with pDNA hCFTR as well as (c)mRNA hCFTR leads to increased levels of hCFTR protein within the transfected cells.

hCFTR (c)mRNA functionality test in vitro.
For functional analysis of the (c)mRNA hCFTR -encoded CFTR channel, we performed a YFP-based functional assay using CFTR null A549 cells or ΔF508 CFBE41o-cells which stably express halide-sensitive YFP-H148Q/I152L 30 . Quenching of the YFP signal induced by hCFTR channel-mediated I − influx is reciprocally proportional to hCFTR channel function 25,32 . Figure 2B shows the quenching efficacy after transfection of 250 ng (c)mRNA hCFTR , for three different time points, normalized to mock-transfected cells. In pDNA hCFTR transfected cells, the quenching efficacy was significantly higher after 48 h and stayed high even after 72 h (P ≤ 0.05), while mRNA hCFTR as well as modified cmRNA hCFTR transfected cells revealed a single peak quenching at 48 h (P ≤ 0.05), which was undetectable at 72 h in A549 cells. In CFBE41ocells mRNA hCFTR could not provide any detectable quenching but . showed very significant quenching at 48 h (P ≤ 0.001), which is in line with expression patterns seen in Fig. 1A,B. (c)mRNA hCFTR immunogenicity ex vivo by an adapted human whole blood assay. Due to lack of a reliable method to detect immune responses that therapeutic mRNAs may trigger in a living organism, we focused on an innovative approach to using whole blood from humans. Blood was collected from three healthy donors and used fresh to conduct whole blood assays. Interestingly, the negative control groups (blood only and NP only) did not raise IFN-α values above the detection limit (Fig. 2C, red dotted lines), while TNF-α and IL-8 were already measurable in human blood untreated or treated only with NPs. That is the reason why we adapted the graphical presentation, using a blue colored area that represents the variance of the negative controls, which are biological replicates. The positive control (R848) lead to a strong and significant production of IFN-α (6 h and 24 h, respectively; P ≤ 0.05), IL-8 (6 h and 24 h, respectively; P ≤ 0.01) and TNF-α (6 h and 24 h, respectively; P ≤ 0.05) (Fig. 2C). All cmRNA hCFTR showed a very similar result in cytokine expression as observed for negative controls: the IFN-α levels did not reach the detection limit of the ELISA; IL-8 and TNF-α responses were not statistically significant at 6 h and 24 h, respectively (Fig. 2C). Unmodified mRNA hCFTR resulted in a significant increase of IFN-α at 6 h and 24 (P ≤ 0.05), only significant increase in IL-8 at 24 hours (P ≤ 0.05) and the TNF-α levels were in line with the negative control. While pDNA hCFTR triggered high TNF-α responses at 6 h (P ≤ 0.05), significant and detectable IFN-α and IL-8 responses after 6 h and 24 h (P ≤ 0.05). Due to both, significantly lower expression of mRNA hCFTR in vitro (Fig. 1) and unwanted higher immune responses of mRNA hCFTR , we focused on cmRNA hCFTR and pDNA hCFTR in the following therapeutic studies. significantly increased the compliance from 0.02 ± 0.01 ml/cmH 2 O (Cftr −/− mice) to 0.03 ± 0.01 ml/cmH 2 O (P ≤ 0.05), reaching equivalent values to those measured in Cftr +/+ mice (Fig. 3B). In contrast, the i.v. application of 40 µg  0.01, Fig. 6C). More importantly, we wanted to analyze if there is a significant increase in hCFTR protein levels in the lungs of treated mice by hCFTR ELISA (Fig. 6B,E). These analyses confirmed that mice treated with 40 µg  pDNA hCFTR i.v. or i.t. did not lead to detectable responses of key cytokines IFN-α or TNF-α (detected by ELISA) at all three-time points (Fig. 7A) 34,35 . Nanoparticles alone (used in all in vivo experiments) showed no immune response over the detection limit. However, as expected the positive control (E. coli extract total RNA) i.v. and i.t. resulted in a significant increase of IFN-α and TNF-α at 6 h and a trend increase of IFN-α at 24 h, while an effect at 72 h was not detectable (Fig. 7A). No immune response had been observed apart from positive control in groups treated intratracheally (i.t) (Fig. 7B).

Discussion
Although much progress has been achieved since the discovery of the CFTR gene 25 years ago, there is still a substantial need to restore robust CFTR function in patients suffering from cystic fibrosis 8 . With the recent approvals of the small molecule agents ivacaftor and lumacaftor, science has paved a possible way to overcome the hurdles caused by the disease-conferring gene. Those treatments can be more or less effectively applied to patients bearing CFTR mutations delF508 (Lumacaftor-ivacaftor/Orkambi) and G551D (ivacaftor) [36][37][38][39] . However, lung function, as one of the main outcome parameters probably having the most significant influence on life quality of CF patients, is rarely tested in preclinical models. In fact, actual effects of (modern) existing drugs on lung function, with forced expiratory volume in one second (FEV 1 ) as a key parameter, are quite low 40 . Here, by using cmR-NA hCFTR , we are presenting a proof of concept for a viable and potent therapeutic alternative. We have vigorously tested mRNA therapy with focus on in vivo lung function normalization while avoiding any possible, unwanted immune responses for a possibility of repeated dosing. The unique formulation utilized can be used both topically (intratracheally) and systemically (via i.v. injection), having in both cases a profound effect on normalizing the lung function parameters, including compliance, resistance and FEV 0.1 of treated Cftr −/− mice to values obtained from Cftr +/+ mice.
In vitro, using cmRNA hCFTR , CFTR protein expression in CFBE41o− cells was increased up to 5.5-fold compared to mRNA hCFTR , which is consistent with previous studies obtained by us and others 18,31,41 . Incorporation of naturally occurring chemically modified nucleosides has been shown to suppress inhibitory effects on translation by avoiding detection by pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs) TLR3, TLR7, and TLR8 34,35 . Those receptors play a crucial role in the detection, processing, and degradation of mRNA. Interestingly, depending on the mRNA modification, kinetics of hCFTR expression varies upon the different nucleosides used. In fact, after 72 h we only observe an increased quenching of Yellow fluorescent protein (YFP) in YFP assay in CFTR null A549 and CFBE41o− cells by pDNA hCFTR which would corroborate our findings from our flow cytometry and western blot analyses in CFBE41o− cells. In contrast there is a significant increase in I − influx by functional hCFTR channels and quenching of YFP at 48 h post transfection by cmRNA hCFTR . Consequently, we assume that upon different cell lines, kinetics by which the hCFTR protein is expressed varies. Earlier studies support our notion that differently modified mRNAs can have an impact on the translational effect between distinct cell lines 31,35 .
To better mimic in vivo human conditions, we performed an ex vivo whole blood assay (WBA) which offers a more complex environment to test for immune responses. This assay has already been used in a number of preclinical settings, and Coch and colleagues could demonstrate that it has the potential to reflect broad aspects of the in vivo cytokine release caused by oligonucleotides 42 . Indeed, we could show that the small molecule Resiquimod (serving as a positive control by activating TLR7 and TLR8) lead to a substantial release of IFN-α, TNF-α and IL-8. pDNA hCFTR , as well as unmodified mRNA hCFTR , also showed elevated cytokine levels probably due to the activation of innate immune receptors 34,35 . In contrast, incorporation of modified nucleosides into hCFTR mRNA (cmRNA hCFTR ) abolished such responses, with no detectable amounts of IFN-α. This is in concert with previously published data, demonstrating cmRNA's limited immune responses, mainly by evading detection from receptors such as TLRs, RIG-1, MDA-5 or PKR 34,41 . Interestingly, even though TNF-α or IL-8 could be detected, it rather shows donor-dependency than effects deriving from NPs and/or cmRNA hCFTR with cytokine levels being all within the variance of negative controls. Although it mirrors only the blood compartment and does not reflect the more complex in vivo situation, the WBA can give a prediction of how cytokines are released in the human system in response to systemically applied (c)mRNA prior to clinical testing.
To determine the clinical potential of CFTR-encoded cmRNA we compared not only different modifications in vivo but also two different routes of administration. I.t. application has been chosen for this study on the base of our previous findings of applying cmRNA i.t. in a surfactant protein-B deficient mouse model leading to significantly prolonged survival 26 . Given the fact that in patients suffering from CF one of the key barriers is the airway mucus layer in which inhaled particles are more likely to get trapped and removed, we sought to apply cmRNA h-CFTR /pDNA hCFTR complexed to NPs by i.v. injection as an alternative administration route. Systemic delivery via lipid-modified polymeric nanoparticles have been already shown to target the lungs efficiently 43 .
To support our notion of improved CFTR activity, we performed extensive lung function measurements using state-of-the-art technology to provide detailed in vivo information on different lung function parameters. There are doubts about Cftr −/− mice as a proper model for cystic fibrosis as it does not reflect the typical lung phenotype seen in CF patients 44 . However, the reason behind that seems to be in how deeply lungs or other affected organs had been investigated. A layer of material can be observed with characteristics of an acid mucopolysaccharide on the bronchiolar surface and is also evident in alveoli by using scanning electron microscopy in Cftr −/− mice, which is not evident in Cftr +/+ mice 45 . It has also been reported Cftr −/− mice shows similar effect of CF patients like, age-dependent pulmonary inflammation, death of respiratory epithelial cells and high vulnerability to severe Pseudomonas aeruginosa infection 46 . Recent studies could demonstrate reduced airway compliance and increased resistance in comparison to wild-type mice 47,48 . Indeed, we observed significantly higher and lower levels regarding resistance and compliance, respectively, in Cftr −/− controls and mock-treated Cftr −/− mice compared to homozygous wild-type mice (Cftr +/+ ) mice and demonstrated that treatment with cmRNA hCFTR -NPs improved compliance and resistance significantly equal to those seen in healthy Cftr +/+ mice. FEV 1 percentage (for mouse or small animal FEV 0.1 ) is related to survival in CF and a most important physiological parameter for CF patients. A previous study demonstrated that patients with a %FEV 1 of <30 compared to healthy individuals had a 50% chance of mortality within 2 years and hence are regularly examined in clinical setup 49 . A strong variance amid Cftr −/− controls and mock-treated Cftr −/− mice compared to homozygous wild-type mice (Cftr +/+ ) mice has been perceived in the case of FEV 0.1 . Our study provides a significant improvement of FEV 0.1 due to treatment with NP-cmRNA hCFTR . Interestingly, NP-pDNA hCFTR when administered via i.t. route improved parameters of lung function measurements including FEV 0.1 , but not as significant as cmRNA hCFTR . We also observed i.v. or i.t. Overall, we could demonstrate that certain protocols, applying cmRNA hCFTR either i.v. or i.t. efficiently restored lung function values equal to those of wild-type. Suggesting a more even distribution through arteries and the bronchial circulation by i.v. injection, this route and formulation could lead to a very potent therapy especially for newborns and young infants. By providing functional CFTR early in life, the lungs could be protected from irreversible damage. Nevertheless, when applied intratracheally, which mimics deep inhalation of a spray or powder formulation (primary application route in adults), an adjustment in dose and/or formulation (e.g. cmRNA CFTR s2U /m5C h 0 25 0 25 . . increased to 80 µg) might easily abrogate any negative effect of the Cftr −/− genetic background on lung function.
Eventually, we determined the impact of cmRNA hCFTR and pDNA hCFTR on another relevant physiological outcome such as the saliva chloride concentration to evaluate therapeutic effect and complement the lung function results. Sweat chloride concentration has become an accepted method as a diagnostic readout to assess treatment effects of CF patients 50 . As an analog, chloride concentration of β-adrenergic stimulated salivary glands of Cftr −/− mice can be investigated as it complies with findings in CF patients 33 . In this study, we could show a substantial difference in salivary Cl − content of cmRNA hCFTR and pDNA hCFTR treated mice -both, i.v. and i.t. -compared to their untreated counterpart. With end point-analysis, a significant decrease in Cl − to nearly 50% was observed, indicating a restoration of CFTR in the duct compartment of salivary glands and thus leading to an improved Cl − absorption. Previous studies estimated that a restoration of CFTR activity to 50% could lead to sweat chloride levels to near normal levels in CF patients. Given that, it is possible that cmRNA hCFTR treatment has the potential to improve CFTR activity to levels that are at least similar to those in patients with a mild CF phenotype 51 .
In this study, by applying cmRNA hCFTR consecutively, both modifications were successfully delivered to the lungs with the i.v. route being more efficient at doses of 40 µg (2 mg/kg body weight) per treatment. Intriguingly, in contrast to the results obtained in vitro, . .

cmRNA CFTR
s2U /m5C h 0 25 0 25 showed a significantly higher CFTR protein expression with higher accumulation of hCFTR mRNA in lung cells. Assuming differences of cmRNA-encoded transgene expression between distinct cell lines, it is plausible to consider such differences between in vitro versus in vivo applications, which is by far more complex. In this respect, the higher amount of cmRNA CFTR s2U /m5C h 0 25 0 25 . .
found in lung cells after i.v. injection, might be due to the fact that its nucleoside composition is more favorable to evade PRRs, thus being less degraded. However, regardless of cmRNA kinetics we also observed differences in the delivery route of cmRNA hCFTR /pDNA hCFTR -NPs. Our data suggest i.v. injection to be more efficient in delivering such complexes to the lung than topical administration. Tests of cmRNA hCFTR -NP's capacity of mucus penetration are in planning phase including detection of cmRNA hCFTR and CFTR protein (glycosylated) in a Cftr-deficient mouse model especially at the apical side of the bronchial epithelium. The upper airways are lined with mucus and mucociliary movements clear foreign particles immediately. In addition, the main barriers in the deeper areas are the alveolar lining, scavenger transporters and alveolar macrophages 52,53 . We, therefore, concluded that the original dosing by which cmRNA-NPs were delivered i.t. was not as efficient as using the i.v. route. Indeed, increasing the amount by doubling the dose (to 80 µg) for each treatment showed a hCFTR expression close to levels seen using the i.v. route.
To exclude immune reactions caused by either NPs or the cmRNA hCFTR itself, we conducted extensive immune assay tests in vivo. Except for the positive control (E. coli total mRNA), we could not detect any immunostimulatory effect in vivo that could arise from NPs or the cmRNA hCFTR . These results confirm our previous studies in which we showed that NPs, as well as modified mRNA, could be administered safely to the lungs without any substantial increase in cytokines, or inflammatory-related cells such as macrophages or neutrophils 26 . Systemic delivery has also been reported to have no impact on proinflammatory cytokine secretion 29 .
Taken together, this study is the first proof of concept of efficient application of NP-cmRNA hCFTR in vivo to restore lung function in a Cftr-deficient mouse model. Importantly, we could neither detect immune responses in vivo nor in a more defined setting ex vivo. Applying cmRNA hCFTR to Cftr −/− mice could efficiently restore lung function close to levels of healthy control mice. In addition, our study compared -apart from two well-known mRNA modifications and pDNA hCFTR -also two different delivery routes, demonstrating that systemic administration of cmRNA targets lung cells more efficiently at lower dosages. This study provides a proof of concept for alternative treatment of patients suffering from CF. cmRNA hCFTR transcript supplementation may be broadly applicable for most CFTR mutations, not only in adults but already in the postnatal state, thereby protecting the lungs from exacerbations from the very beginning of life.