Neurotherapeutics (2022) 19:1617–1633 https://doi.org/10.1007/s13311-022-01279-8 ORIGINAL ARTICLE Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune Encephalomyelitis Ángel Edo1,2 · Laura Calvo‑Barreiro3,4 · Herena Eixarch3,4 · Assumpció Bosch1,2,4 · Miguel Chillón1,2,4,5,6 · Carmen Espejo3,4 Accepted: 11 July 2022 / Published online: 28 July 2022 © The Author(s) 2022 Abstract The pathogenic role of the interleukin 21 (IL-21) in different autoimmune diseases, such as multiple sclerosis (MS), has been extensively studied. However, its pleiotropic nature makes it a cytokine that may exhibit different activity depending on the immunological stage of the disease. In this study, we developed a gene therapy strategy to block the interaction between IL-21 and its receptor (IL-21R) by using adeno-associated vectors (AAV) encoding a new soluble cytokine receptor (sIL21R) protein. We tested this strategy in a murine model of experimental autoimmune encephalomyelitis (EAE), obtaining differ- ent clinical effects depending on the time at which the treatment was applied. Although the administration of the treatment during the development of the immune response was counterproductive, the preventive administration of the therapeutic vectors showed a protective effect by reducing the number of animals that developed the disease, as well as an improve- ment at the histopathological level and a modification of the immunological profile of the animals treated with the AAV8. sIL21R. The beneficial effect of the treatment was also observed when inducing the expression of the therapeutic molecule once the first neurological signs were established in a therapeutic approach with a doxycyline (Dox)-inducible expression system. All these clinical results highlight the pleiotropicity of this cytokine in the different clinical stages and its key role in the EAE immunopathogenesis. Keywords IL-21 · Soluble receptor · EAE · Multiple sclerosis · AAV Introduction is understood as a set of bidirectional interactions between the components of the adaptive and innate immune system Multiple sclerosis (MS) is a chronic inflammatory and demy- in the periphery, which includes T cells, B cells, and myeloid elinating disease of the central nervous system (CNS) medi- cells, and resident cells of the CNS, such as microglia and ated by the immune system. MS is the most prevalent demy- astrocytes [2]. Effector CD4 T cells play an important role elinating disease in the world, with more than 2.8 million in MS pathogenesis and are located within inflammatory patients worldwide [1]. Nowadays, MS immunopathogenesis lesions of the CNS and in the cerebrospinal fluid of MS patients [3]. Initially, it was thought that the principal cause for experimental autoimmune encephalomyelitis (EAE) Carmen Espejo and Miguel Chillón contributed equally to this work. 3 * Miguel Chillón Servei de Neurologia‑Neuroimmunologia, Centre d’Esclerosi miguel.chillon@uab.cat Múltiple de Catalunya, Vall d’Hebron Institut de Recerca, Hospital Universitari Vall d’Hebron, Vall d’Hebron * Carmen Espejo Barcelona Hospital Campus, Pg. Vall d’Hebron 119‑129 carmen.espejo@vhir.org (Ed. Collserola, Lab. 149), 08035 Barcelona, Spain 4 1 Institut de Neurociències (INc), Department of Biochemistry Universitat Autònoma de Barcelona, Bellaterra, and Molecular Biology, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Spain (Campus UAB), Bellaterra, Cerdanyola del Vallès (Ed. H 5th 5 Institució Catalana de Recerca I Estudis Avançats (ICREA), level), 08193 Barcelona, Spain Barcelona, Spain 2 Vall d’Hebron Institut de Recerca, Hospital Universitari Vall 6 Vector Production Unit (UPV), Universitat Autònoma de d’Hebron, Barcelona, Spain Barcelona, Barcelona, Spain 13 Vol.:(0123456789) 1618 Á. Edo et al. and MS was the induction of Th1 response by IL-12 [4]; disease, as well as an improvement in the histopathological however, the cellular subtype Th17 (the main producer of signs of the CNS. On the contrary, the blockade of IL-21 IL-17) was later identified as a key player in autoimmun- during the development of the immune response (before the ity [5]. Although the encephalitogenic role has historically onset of clinical signs) was counterproductive, increasing been attributed to autoreactive CD4 T cells, CD8 T cells are EAE clinical severity. Additionally, treatment with induc- more abundant in human inflammatory lesions, and their ible AAV vectors that allowed sIL21R expression once the frequency strongly correlates with axonal damage [3, 6]. On first neurological symptoms were stablished, led to a clinical the other hand, there are components of the immune system improvement. All these results reveal the pleiotropic role of capable of modulating the inflammation by switching to Th2 IL-21 in the different clinical and immunological stages of immune responses or maintaining self-tolerance (i.e., regula- EAE and its interest as a therapeutic target for the treatment tory T (Treg) cells), thus generating a cycle of inflammation of MS. and its resolution [7]. There are different key cytokines involved in the patho- Methods genesis of MS; however, several evidences have pointed to interleukin 21 (IL-21) as one of the main cytokines involved Viral Vector Production and Purification in the initiation of the immune response. IL-21 regulates the activity of both innate (monocytes/macrophages, natural AAV8 and adenovirus 5 (Ad5) vectors were produced, puri- killer (NK) cells, and dendritic cells (DCs)) and the adaptive fied, and manipulated in the biosafety level 2 facilities at immune system (T and B cells). Although IL-21 can exert the Universitat Autònoma Barcelona. Briefly, AAV8 vec- certain suppressive functions on immunity, it is mainly known tors were generated using the triple transfection system in for its broad capacity to stimulate the proliferation and activity HEK293 cells. After 48 h, AAV vectors were harvested, of pro-inflammatory pathways involved in the immunopatho- treated with benzonase (71,206–3, Merck, Kenilworth, NJ, genesis of MS and other autoimmune diseases such as rheu- USA), purified in an iodixanol gradient, and titrated using matoid arthritis (RA) or systemic lupus erythematosus (SLE) the Picogreen system [20]. To produce Ad5 vectors, PacI- [8, 9]. The functions of IL-21 include the differentiation and linearized plasmids were transfected into HEK293 cells, proliferation of pathogenic Th17 cells [10, 11], the increase and the virus was recovered 8 to 10 days post-transfection. of follicular T helper cell proliferation [12], the stimulation of Then, viruses were sequentially amplified, purified by CsCl cytotoxicity and proliferation of CD8 T, NK, and NKT cells [8, gradients, and infective particles measured by an endpoint 13], as well as the inhibition of the differentiation and survival dilution assay that counted the number of hexon-producing of Treg cells [14]. Various genetic studies have established a cells in triplicate using the AdEasy Viral Titer Kit (972,500, relationship between the polymorphisms detected at the IL-21 Agilent Technologies, Santa Clara, CA, USA) [21]. All Ad5 receptor (IL-21R) locus and the susceptibility to develop MS vectors used in the in vitro experiments contained CMV pro- [15]. In addition, histopathological studies of CNS lesions of moter. While AAV8 vectors used in the in vivo experiments patients in the acute or chronic phases of MS have shown that with the animal model contained CMV, CAG or TetOn- both IL-21 and IL-21R are expressed on CD4 T cells, and CMV promoter depending on the therapeutic approach. the latter has also been detected on the cortical neurons of MS lesions [16]. These observations correlate with a higher Immunocytochemistry and Immunodetection expression of IL-21 mRNA during inflammatory relapses [17] of sIL21R and the detection of a high number of IL-21-producing CD4 T cells in the CSF of patients in the progressive stage [18]. HEK293 cells were infected with a multiplicity of infection Also, patients treated with Alemtuzumab (anti-CD52 treat- (MOI) of 2 or 5 for 24–48 h with Ad5.CMV.sIL21R and ment) have decreased levels of IL-21 in serum [19]. All these Ad5.CMV.Null or Ad5.CMV.eGFP as negative controls. highlights expose the involvement of IL-21 in the inflamma- For immunocytochemistry, anti-IL21R antibody (AF596, tory response that occurs in MS. R&D Systems, Minneapolis, MN, USA) and Alexa Fluor In the present work, we have studied the effect of IL-21 488 donkey anti-goat secondary antibody (A11055, Invit- sequestration in different clinical stages by administering rogen, Carlsbad, CA, USA) were used at a dilution of 1:100 adeno-associated vectors (AAV) encoding a soluble IL- and 1:200, respectively. For immunodetection, cell pellet 21R (sIL21R) capable of selectively blocking the function was homogenized in lysis buffer (RIPA) and total protein of IL-21 in an EAE mouse model of MS. Similar strategies was determined from lysed cells and supernatants using the have been tested in other inflammatory diseases such as SLE Pierce BCA Protein Assay (232,227, Thermo Fisher Scien- or RA, obtaining promising results [9]. Our results show tific, Waltham, MA, USA). Protein extracts and superna- that the preventive treatment with sIL21R produced a clini- tants were loaded onto denaturing acrylamide gels and then cal improvement due to a reduction in the incidence of the electrotransferred to PDVF membranes (10,600,023, GE 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1619 Healthcare, Chicago, IL, USA). Anti-IL21R primary anti- postimmunization (p.i), mice were intravenously (i.v.) body (AF596, R&D Systems) was used at 1:1000 dilution injected with 250 ng of pertussis toxin (P7208, Sigma- and goat HRP-anti-IgG secondary antibody (P0160, Dako, Aldrich). The clinical symptoms of EAE were monitored Agilent Technologies, Santa Clara, CA, USA) at 1:10,000 daily using the following criteria: 0, no clinical signs; 0.5, were used in the presence of 5% (w/v) of BSA (A9418, loss of caudal tail tone; 1, paralysis of whole tail; 2, mild Sigma-Aldrich, St. Louis, MO, USA). paraparesis of one or both hindlimbs; 2.5, severe paraparesis or paraplegia; 3, mild tetraparesis; 3.5, moderate tetrapare- sis; 4, tetraparesis (paralyzed hindlimbs); 4.5, quadriplegia In Vitro Blocking of Murine IL‑21 (paralyzed forelimbs); 5, moribund; and 6, death. Correc- tive measures and endpoint criteria to ensure EAE-incident sIL21R and eGFP conditioned media were produced in animals’ welfare were applied as described in [22]. All data HEK293 cells by the infection with Ad5.sIL21R, and Ad5. presented are in accordance with the ARRIVE guidelines for eGFP as a control, using a MOI of 10. Cells were harvested animal research and with the guidelines suggested for EAE 30 h post-infection, lysed by freeze–thaw cycles, centrifuged publications [23]. 10 min at 3000 rpm to obtain the cleared supernatants, and stored at −80 °C until use. Fresh splenocytes were cultured In Vivo AAV and Doxycycline Administration in 6-well plates at 6·106 cells/well, in DMEM supple- mented with 10% FBS (1,237,980, Thermo Fisher Scien- To determine the prophylactic effect of sIL21R expression tific). Recombinant murine IL-21 (12,340,212, Immunotools in EAE mice, a single dose of 5·1011 vg/mouse of AAV8 GmbH, Frieshoythe, Germany) at 10 ng/mL was incubated vectors (Null or sIL21R) was administered 21 days before for 1 h at 37 °C in sIL21R or eGFP conditioned culture EAE induction. This dose was determined on the basis of a media, and latter, added to the splenocytes for 15 min. previous study done by our group [25]. In the early thera- Then, splenocytes were harvested, medium discarded by peutic experiment, a dose of ­1012 viral genomes (vg)/mouse centrifugation, and cells resuspended in lysis buffer (RIPA). of AAV8 vectors was administered at days −1 or +1 p.i. The Results were analyzed by Western blot using anti-pY705- null control is a vector that has the same serotype as the Stat3 (9145S), anti-Stat3 (4904S), and anti-Gapdh (2118S; therapeutic vector but does not encode any transgene as it Cell Signaling Technology, Danvers, MA, USA) as primary contains a non-transcribed DNA. Finally, in the therapeutic antibodies, all at 1:1000 dilution. Rabbit HRP-anti-IgG sec- approach with doxycycline (Dox)-inducible vectors, a dose ondary antibody (P0399; Dako) at 1:10,000 dilution was of ­1012 vg/mouse was administered 15 days before EAE used in the presence of 5% (w/v) BSA. Band intensities were induction and the expression of the inducible system was quantified using Image Lab (Bio-Rad Laboratories, Hercu- daily activated with an intraperitoneal (i.p.) administration les, CA, USA). of 25 mg/kg Dox-Hyclate (D9891, Sigma-Aldrich) diluted in 100 μL of NaCl 0.9% from the day on which the mice had Animals and Ethical Permissions a clinical score of 0.5. In all experiments, the viral dose was diluted in 200 μL of NaCl 0.9% and was administered by i.v. Female (6- to 9-week-old) C57BL/6JOlaHsd mice were injection through the lateral tail vein. purchased from Envigo (Venray, The Netherlands) and housed under standard laboratory conditions. All proce- Splenocytes Proliferative Assay dures were performed in accordance with European Union (Directive 2010/63/EU) and governmental regulations (Real Splenocyte suspensions were obtained by griding murine Decreto 53/2013; Generalitat de Catalunya Decret 214/97). spleens through a 70 μm nylon cell strainer at 27 days p.i. The Ethics Committee on Animal Experimentation of the In all proliferative assays, 2·10 5 splenocytes/well were Vall d´Hebron Research Institute approved all procedures seeded in 96-well plates in 200 μL of proliferation medium described in this study (protocol number: 9533 CEEA). (X-VIVO™ 15 medium (BE02-060F, Lonza, Basel, Switzer- land) supplemented with 1% (v/v) L-glutamine (X0550, Bio- Induction of EAE and Clinical Follow‑Up west, Nuaillé, France), 0.4% (v/v) penicillin–streptomycin (L0022, Biowest), 0.1 M HEPES (H0887, Sigma-Aldrich), Anesthetized mice were immunized by subcutaneous and 6 μM 2-β-mercaptoethanol (M3148, Sigma-Aldrich)). injections of 100 μg ­MOG40-55 (Proteomics Section, Uni- Cells were stimulated with 5 μg/mL of M ­ OG40-55 or 5 μg/mL versitat Pompeu Fabra, Barcelona, Spain) emulsified in of phytohemagglutinin (PHA-L) (L2769, Sigma-Aldrich) 100 μL of CFA (IFA (F5506, Sigma-Aldrich) containing and cells without any stimulus were used as baseline con- 4 mg/mL Mycobacterium tuberculosis H37RA (231,141, trols. After 48 h, 75 μL of supernatant were harvested and BD Biosciences, San Jose, CA, USA)). At 0 and 2 days stored at −80 °C to further assess cytokine secretion. At 13 1620 Á. Edo et al. the same time, 1 μCi/well of [­ 3H]-thymidine (NET027Z, into 4-μm-thick transversal sections. Samples were immu- PerkinElmer, Waltham, MA, USA) was added and incubated nostained to evaluate several histopathological parameters: for 18 h. The levels of radioactivity were measured using a demyelination with anti-myelin basic protein (MBP) antibody β-scintillation counter (Wallac, Turku, Finland). Five repli- (AB980, Merck), axonal damage with anti-neurofilament H, cates per mouse and condition (control, ­MOG40-55, PHA-L) non-phosphorylated antibody (SMI-32) (801,701, Biole- were performed and results were expressed as the stimula- gend, San Diego, CA, USA), C ­ D3+ infiltrating cells with tion index. Stimulation index was calculated by dividing the anti-CD3 antibody (A0452, Dako) and reactive astroglia mean counts per minute of M ­ OG40-55 or PHA-L condition with anti-glial fibrillary acidic protein (GFAP) antibody by the mean of the control condition. (Z0334, Dako). For reactive microglia identification, lec- tin from Lycopersicon esculentum (LEA) (L0651, Sigma- Cytokine Detection Aldrich) was used. Briefly, samples were deparaffinized and rehydrated, and antigen retrieval (SMI-32 and CD3 Cytokine secretion was determined in the supernatants of stainings) and incubation with blocking solution for 1 h at ­MOG40-55 stimulated splenocytes using the Th1/Th2/Th9/ room temperature were performed. Then, primary antibod- Th17/Th22/Treg Cytokine 17-Plex Mouse ProcartaPlex ies diluted in blocking solution were incubated overnight Panel (EPX170-26,087–901, Invitrogen) and TGF beta 1 at 4 °C. After several washes, the samples were incubated Mouse Procartaplex Simplex Kit (EPX01A-20608–901, 1 h at room temperature with their respective secondary Invitrogen), according to the manufacturer’s instructions. antibodies diluted in blocking solution. Finally, cell nuclei Data were analyzed with a Magpix instrument (Luminex were stained with 4’, DAPI (D9542, Sigma-Aldrich). Images Corporation, Austin, TX, USA) and ProcartaPlex Analyst were taken using a Nikon Eclipse 90i fluorescence micro- software (Thermo Fisher Scientific). scope (Nikon, Minato, Japan) and quantification performed with the ImageJ analysis software (Wayne Rasband National Analysis of mRNA Levels by qRT‑PCR Institute of Health, Bethesda, MD, USA). Regarding anti- CD3 labelling, positive cells were counted manually and Total RNA was extracted from the isolated splenocytes using the results expressed as ­CD3+ cells/mm2 of the entire white the RNeasy Mini Kit (74,104, Qiagen), treated with DNase I matter area. Quantifications of MBP, LEA, SMI-32, and Amplification Grade (AMPD1, Sigma-Aldrich) and reverse GFAP labelling are expressed as the percentage of the posi- transcribed with iScript cDNA Synthesis Kit (1,708,891, tive area respect to the entire white matter area. In all cases, Bio-Rad Laboratories). Primers for Tgfb1 (Mm01178820, two transversal sections of caudal spinal cord per animal Applied Biosystems, Foster City, CA, USA), Rorc2 (cus- were analyzed. tom assay as described in [24]), Tbx21 (Mm00450960, Applied Biosystems), Foxp3 (Mm00475162, Applied Bio- Flow Cytometry systems), Gata3 (Mm00484683, Applied Biosystems), Il4 (Mm00445259, Applied Biosystems), Il10 (Mm01288386, To determine the immunological profile, spleen cell sus- Applied Biosystems), Ifng (Mm01168134, Applied Bio- pensions were prepared as described previously. First, systems), Il17a (Mm00439618, Applied Biosystems), and dead cells were labelled with Fixable Viability Stain (BD Gapdh (Mm99999915, Applied Biosystems) as well as Pharmigen, BD Biosciences) and cells were blocked with TaqMan Gene Expression Master Mix (4,369,016, Applied anti-mouse CD16/32 (553,142, BD Pharmigen). To analyse Biosystems) were used to perform qPCR according to the Treg cell population, antibodies against CD3 𝜀 (561,827, manufacturer’s instructions. All experiments were per- BD Pharmigen), CD4 (12–0251, eBiosciences), CD8a formed on CFX384 Touch Real-Time PCR Detection System (563,152, BD Horizon), CD25 (12–0251, eBiosciences), (Bio-Rad Laboratories) and the analysis of qPCR output data and Foxp3 (17–5773, eBiosciences) were used. For intra- followed the manufacturer-suggested 2­ −ΔΔCt method [26]. cellular labelling of Foxp3, the Foxp3 Fixation/Permea- The expression of the housekeeping gene (Gapdh) was used bilization Kit (00–5523-00, eBiosciences) was used fol- to normalize the expression of the different genes of interest lowing the manufacturer’s instructions. For B cell subsets, while the Null group was used as a calibrator. antibodies against B220 (552,772, BD Pharmigen), CD1d (553,846, BD Pharmigen), CD5 (550,035, BD Pharmi- Histological Studies gen), CD19 (560,143, BD Pharmigen), CD138 (562,610, BD Horizon), and MHCII (562,363, BD Pharmigen) were At the endpoint of the experiment (day 27 p.i), mice were used. For DC subpopulations and activation status, anti- euthanized, and spinal cords collected and fixed in 4% para- bodies against B220 (553,088, BD Pharmigen), CD11b formaldehyde at 4 °C for 24 h, embedded in paraffin, and cut (25–0112, eBiosciences), CD11c (47–0114, eBiosciences), 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1621 CD8a (563,152, BD Horizon), CD80 (553,769, BD Phar- Results migen), and MHCII (562,363, BD Pharmigen) were used. For analysis of macrophage, antibodies against CD11b Design and Characterization of sIL21R (562,605, BD Horizon), CD206 (141,712, BioLegend), F4/80 (565,410, BD Pharmigen), Ly6C (560,593, BD Phar- sIL21R was designed based on the transmembrane receptor migen), and Ly6G (551,460, BD Pharmigen) were used. (NCBI GenBank NM_021887.2) (Fig. 1a), since no natu- For NK cells and its activation status, antibodies against ral form of the soluble receptor has been identified, neither CD3ε (561,827, BD Pharmigen), CD69 (560,689, BD human nor murine. However, only the extracellular region Pharmigen) and NK1.1 (563,220, BD Horizon) were used. of the IL-21R was used as a therapeutic molecule, which For the analysis of cytokine production, splenocytes were contains both the secretion signal and the canonical binding stimulated ex vivo with 50 ng/mL PMA (P1585, Sigma- site for IL-21 (Fig. 1b). Aldrich) and 1 μg/mL ionomycin (I0634, Sigma-Aldrich) First, a complete characterization of the sIL21R was in the presence of GolgiPlug (555,029) and GolgiStop carried out to demonstrate that the recombinant protein (554,724, BD Bioscences) for 6 h. Extracellular stain- was soluble, secretable, and capable of selectively block- ing with anti-CD3ε (561,827, BD Pharmigen), anti-CD4 ing murine IL-21. HEK293 cells were infected with Ad5. (46–0042, eBiosciences), and anti-CD8a (563,152, BD sIL21R or Ad5.Null (control) with a low MOI (MOI = 2) Horizon) was performed. Intracellular staining of cytokines to avoid saturation. Immunocytochemistry of the infected was performed to detect IL-17A (561,020, BD Pharmigen), cells revealed a clear dispersed cytoplasmic localization of IFNγ (563,376, BD Horizon), IL-4 (554,436, BD Phar- sIL21R (Fig. 1c), which confirmed the solubility of sIL21R. migen), and IL-10 (563,276, BD Horizon). Anti-CD69 To assess its secretion capacity, HEK293 cells were infected (560,689, BD Pharmigen) was used as control of proper with Ad5.sIL21R or Ad5.eGFP, as a negative control of ex vivo stimulation. For intracellular staining, BD Cyto- secretion, using a higher MOI (MOI = 5) to favor its detec- fix/Cytoperm Fixation/Permeabilization Kit (554,714, BD tion. After 24 h, sIL21R was detected in the cell culture Biosciences) was used according to manufacturer’s instruc- medium, suggesting that the therapeutic molecule is highly tions. Fluorescence minus one controls were included in secretable (Fig. 1d). the analysis to properly interpret the data. Samples were When natural IL-21 binds to its membrane receptor, acquired in a CytoFLEX flow cytometer and data was ana- it signals through the phosphorylation of Stat3. In a pre- lyzed with the CytExpert 2.3 software (Beckman Coulter, liminary experiment with mouse primary splenocytes, we Brea, CA, USA). determined that an incubation time of 15 min with IL-21 was sufficient to detect the phosphorylated form of Stat3 Statistical Analysis (p-Stat3). In order to demonstrate that sIL21R blocks IL-21 signalling pathway, IL-21 and sIL21R were co-incubated for The data on the text are expressed as the mean of the values 1 h and then added to splenocytes. As a result, we observed obtained ± the standard deviation of the mean (SD), while a 90–95% reduction of the p-Stat3 signal (Fig. 1e). the graphs are represented as the mean of the values ± the standard error of the mean (SEM). When the data of the AAV8.CMV.sIL21R Treatment Prevents groups to be compared followed a normal distribution, the Development of EAE parametric analyses were applied. When comparing two independent groups, Student t analysis was used, while for Once the biochemical and functional characteristics of the comparison of three or more groups, ANOVA test with sIL21R were assessed, we studied its therapeutic potential a Tukey post hoc test was used. Nonparametric analyses in an experimental model of MS, starting with a preven- were applied when the data did not have a normal distribu- tive approach. AAV8.CMV.sIL21R and AAV8.CMV.Null tion. Mann–Whitney analysis was used for the comparison vectors were administered intravenously 21 days before the of two independent groups, while the Kruskal–Wallis test induction of EAE with a viral dose of 5·1011 vg per ani- with a Dunn’s post hoc test was used for the comparison mal. As seen in Fig. 2a, sIL21R-treated group presented a of three or more groups. Statistical analysis of the data was significantly lower area under curve (AUC) of the clinical performed with SPSS 26.0 (IBM, New York, NY, USA) and score (Null: 51.67 ± 19.27, n = 18; sIL21R: 30.49 ± 30.08, GraphPad Prism 8.0 (GraphPad Holdings, San Diego, CA, n = 19; p = 0.015) (Fig. 2b), and a lower maximum clinical USA), which was also used to generate the graphs. Bilateral score (Null: 3.97 ± 1.33, n = 18; sIL21R: 2.37 ± 2.17, n = 19; hypothesis analysis was applied in all cases and differences p = 0.02) due to a significant reduction in the number of ani- were considered statistically significant when the p value mals which developed EAE [Null: 17/18 (94.44%); sIL21R: was < 0.05. 11/19 (57.89%); p = 0.019]. The motor dysfunction of EAE 13 1622 Á. Edo et al. Fig. 1  Biochemical and functional characterization of the sIL21R. lysate (CL). Ad5.eGFP-infected HEK293 cells were used as a nega- a Schematic representation of the transmembrane IL-21R and the tive secretion control. Representative Western blot of three independ- designed sIL21R (in gray). b Sequence of the therapeutic molecule, ent experiments with primary anti-IL-21R and anti-eGFP antibody. e that contains the secretion signal and the extracellular region con- sIL21R blocked IL-21-Stat3 phoshorylation signalling by inhibiting taining the described IL-21 binding site. c Immunocytochemistry its binding to IL-21R. Mouse primary splenocytes were incubated with anti-IL-21R antibody (green) reveals cellular distribution of the for 15 min with 10 ng/mL of IL-21 that had previously been co- soluble receptor in HEK293 cells infected with Ad5.sIL21R. Ad5. incubated with or without sIL21R. Control: unstimulated splenocyte Null-infected cells were used as a negative control. Cell nuclei were cell lysate. Representative Western blot and quantification of the stained with Hoechst (blue). d Secretion of sIL21R into the extra- pStat3/Stat3 ratio from two independent experiments. Data are repre- cellular medium was detected in the supernatants (SN) of HEK293 sented as mean ± SEM. ****p < 0.0001 cells infected with Ad5.sIL21R. sIL21R was also detected in the cell is associated with weight loss, which is more pronounced less weight than the control group, however, the differences in the acute phase of the disease, reaching a partial recov- did not reach statistical significance (AUC of weight change ery that is maintained in the chronic phase. As observed for Null: −56.37 ± 197.00, n = 18; sIL21R: 48.23 ± 199.60, in Fig. 2c, sIL21R-treated mice showed a tendency to lose n = 19; p = 0.118). At the end of the experiment, strong 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1623 Fig. 2  The preventive treatment with AAV8.CMV.sIL21R pro- to initial weight on the day of EAE induction. The graphs represent tects against the development of EAE. Mice were immunized with the combined results of two independent experiments (Null, n = 18; ­MOG40-55 and treated with 5·1011 vg/mouse of AAV8.CMV.sIL21R sIL21R, n = 19). Treatment did not alter the (d) antigen-specific or AAV8.CMV.Null, 21 days prior to the induction of EAE. a Mean response ­(MOG40-55) or the (e) polyclonal response (PHA-L) in sple- daily clinical score for each group. b sIL21R-treated group showed nocytes. The graphs represent the combined results of two independ- a statistically significant milder clinical course of the disease com- ent experiments (Null: n = 16; sIL21R: n = 18). Data are represented pared with the Null group. c Average daily weight changes relative as mean ± SEM. *p < 0.05 sIL21R expression in liver was confirmed in all animals clinical symptoms developed a proper antigen-specific T cell (data not shown). However, when analyzing the clinical response due to the immunization. variables of the incident mice only, the therapeutic group did not show less AUC of the clinical score than the control sIL21R‑Preventive Treated Mice Present group (Null: 54.71 ± 14.76, n = 17; sIL21R: 52.66 ± 18.56, an Improvement in the CNS Pathology n = 11; p = 0.826). To find out whether the treatment had any effect on T-cell The mice that were preventively treated with AAV8.sIL21R reactivity to the MOG, proliferative assays of splenocytes presented less demyelination (Null: 72.83% ± 6.27%, n = 8; were performed with a specific response against M ­ OG40-55 sIL21R: 87.05% ± 8.49%, n = 8; p = 0.002) (Fig. 3b), micro- and polyclonal immune response with PHA-L, as positive glia activation (Null: 16.19% ± 12.90%, n = 8; sIL21R: control. No differences were observed in the proliferative 4.58% ± 3.14%, n = 8; p = 0.026) (Fig. 3c), axonal dam- capacity of splenocytes from sIL21R and Null group nei- age (Null: 2.30% ± 1.66%, n = 8; sIL21R: 0.81% ± 0.92%, ther when stimulated with M ­ OG40-55 (Null: 7.94 ± 3.46, n = 8; p = 0.043) (Fig. 3d) and T-cell infiltration (Null: n = 16; sIL21R: 7.63 ± 1.84, n = 18; p = 0.740), or with 390.20 cells/mm2 ± 501.4 cells/mm2, n = 8; sIL21R: 47.21 PHA-L (Null: 21.31 ± 9.23, n = 16; sIL21R: 20.12 ± 9.61, cells/mm2 ± 49.65 cells/mm2, n = 8; p = 0.015) (Fig. 3e). n = 18; p = 0.716) (Fig. 2d, e). However, all splenocytes Of note, no differences were seen in activated astroglia from sIL21R-treated mice presented a MOG-specific pro- (Null: 7.57% ± 2.77%, n = 8; sIL21R: 5.79% ± 3.19%, n = 8; liferation, demonstrating that mice that did not develop EAE p = 0.253) (Fig. 3f). 13 1624 Á. Edo et al. Fig. 3  Mice treated with AAV8.CMV.sIL21R have a better CNS his- treated with AAV8.sIL21R, while there were no differences in f the topathological status. a Selected images from a representative ani- GFAP marking. The CD3.+ T-cell infiltrate has been delimited with mal of each group. b Demyelination, c microglia activation, d axonal a dotted line. (Null: n = 8; sIL21R: n = 8). Data are represented as damage, and e T-cell infiltration was significantly reduced in mice mean ± SEM. *p < 0.05, **p < 0.01 Preventive Blockade of IL‑21 Modifies the Immune cytometry as well as the cytokine expression profile of sple- Profile nocytes. sIL21R-treated mice showed a lower percentage of IL-10 producing CD4 T cells (IL-10+ in ­CD3+CD4+) (Null: Once the protective effect of the treatment was observed, 2.07% ± 0.54%, n = 8; sIL21R: 1.46% ± 0.49%, n = 9; p = 0.028) we decided to study the peripheral immune response in the (Fig. 4a), but no significant differences were found in the rest chronic phase by analyzing various cell populations by flow of analyzed immune populations (Supplementary Table 1). In 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1625 addition, sIL21R preventive treatment increased the concentra- Early Therapeutic Approach with AAVs tion of TNFα (Null: 23.99 pg/mL ± 8.44 pg/mL, n = 8; sIL21R: of sIL21R‑Constitutive Expression Produces 36.36 pg/mL ± 12.47 pg/mL, n = 9; p = 0.032), IL-18 (Null: a Clinical Worsening of EAE 719.20 pg/mL ± 315.10 pg/mL, n = 8; sIL21R: 1095.00 pg/ mL ± 387.90 pg/mL, n = 9; p = 0.046) and tended to increase The next step consisted in testing the effect of blocking IFNγ (Null: 196.70 pg/mL ± 136.20 pg/mL, n = 8; sIL21R: IL-21 with a constitutive expression of AAVs in a therapeu- 418.00 pg/mL ± 287.10 pg/mL, n = 9; p = 0.060) in the super- tic approach. In previous studies, we showed a decrease in natants of the splenocytes stimulated with M ­ OG40-55 (Fig. 4b). the vector expression if it was administered after the induc- Subsequently, we decided to further investigate the tion of EAE (days 5 and 9 p.i) [25]. peripheral immunological response in sIL21R-treated mice. Based on this study, we decided to clone the therapeutic For this purpose, new experiments were carried out in which gene under the control of the CAG promoter and to double the mice were euthanized in the acute phase of the disease the viral dose administered in the preventive approach to (day 15 p.i), when the immune response is already stab- ensure that the transgene was expressed at high levels when lished. First, the expression of characteristic transcription the disease was established. factors and cytokines involved in EAE pathogenesis were Furthermore, considering that an AAV requires between analyzed in splenocytes by qPCR. sIL21R-treated mice 1 and 2 weeks to reach maximum expression [27, 28], we showed a higher mRNA expression of Tgfb1 (6.44 ± 4.33, decided to perform an early therapeutic approach in which n = 7; p = 0.024), Tbx21 (Th1 cells) (6.11 ± 4.61, n = 7; ­1012 vg per animal of AAV8.CAG.sIL21R or AAV8.CAG. p = 0.028) and Gata3 (Th2 cells) (3.37 ± 1.65, n = 7; Null were administrated the day before (day −1) or after p = 0.043) and a trend towards increased expression of Foxp3 (day +1) the immunization day. Surprisingly, when the (Treg cells) (3.66 ± 1.84, n = 7; p = 0.092) when results clinical variables of the EAE-incident mice were analyzed, are normalized with respect to Null group. No expression sIL21R-treated group developed a clinical worsening of the change was observed in Rorc2 (Th17 cells) (1.13 ± 0.25, disease. Although differences reached statistical significance n = 7; p = 0.467) (Fig. 4c). However, the expression of sev- in the group administered on day −1 (mean clinical score eral genes could not be detected (Il17a, Il4, Il10, and Ifng). AUC (Null: 28.63 ± 15.95, n = 7; sIL21R: 56.54 ± 20.16, Interestingly, this effect was no longer observed when ana- n = 7; p = 0.011), weight change AUC (Null: 90.58 ± 123.20, lyzing the expression in splenocytes from animals eutha- n = 7; sIL21R: -83.73 ± 161.50, n = 7; p = 0.042)) (Fig. 5a, b), nized in the chronic phase (day 27 p.i), when the immune the experimental group administered on day +1 also showed response is not so activated (data not shown). Thus, in a trend towards clinical worsening (mean clinical score AUC this experimental model, the effect of the sIL21R treat- (Null: 21.33 ± 15.70, n = 6; sIL21R: 35.70 ± 13.03, n = 7; ment mainly happens during the establishment of immune p = 0.138), weight change AUC (Null: 142.10 ± 79.28, n = 6; response at the periphery. sIL21R: 25.89 ± 139.00, n = 7; p = 0.099)) (Fig. 5c, d). sIL21R-treated mice exhibited a significantly lower percent- When the clinical data of this experiment were analyzed, it age of activated CD4 T cells ­(CD25+Foxp3− in ­CD3+CD4+) was observed that the severity of disease, both in the sIL21R (Null: 1.80% ± 0.42%, n = 10; sIL21R: 1.44% ± 0.22%, and the Null-treated groups, seemed to become milder when n = 10; p = 0.029) (Fig. 4d) and a reduction in the percent- the AAVs were administered at day +1 compared to the group age of activated NK cells ­(CD69+ in NK1.1+CD3−) (Null: administered on day −1, so we hypothesized that the admin- 17.28% ± 3.40%, n = 10; sIL21R: 14.07% ± 3.08%, n = 10; istration of AAVs on days close to immunization produced p = 0.040) (Fig. 4e). On the other hand, in this acute clini- an alteration in the clinical course. To confirm this hypoth- cal phase, the sIL21R treatment induced a greater expres- esis, Null groups included in the early therapeutic experiment sion of the activation marker MHCII in myeloid DCs were compared with a Null group administered in preventive (mDCs: ­CD11b+B220−CD11chighCD8−) (Null: 13,816.00 (day −21 p.i) of an experiment carried out in parallel. When median fluorescence intensity (MFI) ± 2265.00 MFI, n = 10; analyzing the clinical variables of EAE-incident mice, we sIL21R: 16,603.00 MFI ± 2467.00 MFI, n = 10; p = 0.017) observed that the clinical severity was significantly lower (Fig. 4f). The sIL21R-treated group also exhibited a signifi- when vectors were administered close to the day of immu- cantly higher percentage of ­MHCII+ lymphoid DCs (lDCs: nization (Null day −21: 49.41 ± 13.48, n = 7; Null day −1: ­CD11b+B220−CD11chighCD8+) (Null: 87.67% ± 5.51%, 28.63 ± 15.95, n = 7; Null day +1: 21.33 ± 15.70, n = 6; n = 10; sIL21R: 92.19% ± 2.57%, n = 10; p = 0.031) (Fig. 4g). p = 0.012) (Fig. 6a, b). However, no statistically significant dif- No significant differences were found in the other immune ferences were found when weight change was analyzed (Null populations analyzed (Supplementary Table 2) or the cytokine day −21: 64.05 ± 119.60, n = 7; Null day −1: 90.58 ± 123.20, secretion profile (Fig. S1). n = 7; Null day +1: 142.10 ± 79.28, n = 6; p = 0.457). 13 1626 Á. Edo et al. Fig. 4  Preventive treatment with AAV8.CMV.sIL21R modifies IL-21 also produced a peripheral reduction in the frequency of the immunological profile of mice with EAE. In the chronic phase, activated CD4 T-cells (­CD25+Foxp3− in ­ CD3+CD4+) and e acti- sIL21R-treatment (a) induced a peripheral reduction in IL-10 pro- vated NK cells (­CD69+ in NK1.1+CD3−) during the acute phase of ducing CD4 T cells (­CD3+CD4+IL10+) and (b) an increase in EAE. However, f sIL21R promoted an increase in the activation of the levels of TNFα, IL-18 and a trend towards increased IFNγ in DCs which could be observed in a higher MFI of MHCII in mDCs the supernatant of splenocytes stimulated with M ­ OG40-55. (Null, ­(CD11b+B220−CD11chighCD8−) and g in a higher frequency of n = 8; sIL21R, n = 9). In the acute phase, sIL21R-treatment (c) ­MHCII+ lDCs (­ CD11b+B220−CD11chighCD8+MHCII+). (Null, increased the mRNA expression of TGFb1, Tbx21 and Gata3, rela- n = 10; sIL21R, n = 10). Data are represented as mean ± SEM (a, b, d, tive to Null group (Null, n = 8; sIL21R, n = 7). d The blockade of e, f, g) or as mean ± SD (c). #p < 0.06, *p < 0.05 Treatment with AAV8.TetOn.sIL21R Improves as those that reported acceptable liver expression in animals Established Symptoms of EAE (data not shown). The use of a Dox-inducible system forced us to consider To test the therapeutic potential of sIL21R in the absence the influence that the inductor could have on the clinical of viral capsid effect, Dox-inducible AAVs were developed. course and its toxicity in the animal model. Thus, compar- This vectors can be administered prior to the induction of ing both the mean clinical score (p = 0.559) and the per- EAE and induce the expression of the therapeutic protein centage of variation in weight (p = 0.611) between EAE with the appearance of the first neurological signs of the dis- animals treated with Dox and the vehicle (saline solution), ease. In a previous pilot study, the experimental conditions of no differences were observed (Fig. 7a, b). In addition, EAE ­1012 vg per animal and a Dox dose of 25 mg/kg were selected mice presented hepatic eGFP expression levels similar 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1627 Fig. 5  The early therapeutic treatment with AAV8.CAG.sIL21R induction. c Mean daily clinical score of incident animals treated at produces a clinical worsening of EAE. Mice were immunized with day +1. d Average daily weight change of incident animals treated at ­MOG40-55 and treated with 10.12 vg/mouse of AAV8.CAG.sIL21R day +1 relative to initial weights on the day of EAE induction. Data or AAV8.CAG.Null the day before (day −1) or the day after (day +1) were obtained from a single experiment (Null (day −1): n = 7; sIL21R the induction of EAE. a Mean daily clinical score of EAE-incident (day −1): n = 7; Null (day +1): n = 6; sIL21R (day +1): n = 7). Data mice treated at day −1. b Average daily weight change of incident are represented as mean ± SEM. *p < 0.05 animals treated at day −1 relative to initial weights on the day of EAE to those presented by healthy control animals both in the Discussion percentage of e­ GFP+ cells (Vehicle: 0.30% ± 0.04%, n = 3; Healthy: 19.07% ± 6.23%, n = 8; EAE: 16.15% ± 4.50%, The pleiotropy and the wide range of action of IL-21 on n = 8) (Fig. 7d) and the level of eGFP expression (Vehicle: the components of the immune system have placed this 3297 MFI ± 135.50 MFI, n = 3; Healthy: 8459 MFI ± 2890 cytokine in the central axis of various pathologies such as MFI, n = 8; EAE: 7100 MFI ± 1392 MFI, n = 8) (Fig. 7e). cancer and immune-mediated diseases like MS, RA, SLE, These results validated the use of these inducible vectors for or type 1 diabetes [29]. In the present study, we have devel- therapeutic approaches in this animal model of MS. oped a gene therapy strategy for blocking IL-21 signalling Finally, mice administered with AAV8.TetOn.sIL21R using viral vectors encoding a sIL21R and have tested it in a showed less severe symptoms once the expression of sIL21R mouse model of EAE. In in vitro studies, sIL21R exhibited was induced compared to the Null-control group (Null: the desired characteristics of solubility and secretability, as 31.64 ± 2.82, n = 7; sIL21R: 25.75 ± 7.46, n = 7; p = 0.043) well as decreased IL-21 signalling when incubated in the (Fig. 7f). No statistically significant differences were found presence of sIL21R in murine splenocytes. in the AUC of weight change (Null: −5 1.38 ± 41.47, n = 7; First, we tested the prophylactic potential of sIL21R by sIL21R: −31.77 ± 39.83, n = 7; p = 0.384) (Fig. 7g). administering constitutively expressed AAV8 vectors in 13 1628 Á. Edo et al. EAE animals. Since AAV vectors require at least 2 weeks Fig. 7  Dox administration does not alter the clinical course and the ◂ to reach the maximum level of transgene expression [27, 28], treatment with AAV8.TetOn.sIL21R improved established symptoms of EAE. Mice were immunized with ­MOG40-55, treated with 1­ 012 vg/ the preventive in vivo experiments involved the administra- mouse of AAV8.TetOn.Null, AAV8.TetOn.eGFP or AAV8.TetOn. tion of the viral vectors 3 weeks prior to mice immunization sIL21R 14 days before the induction of EAE, and transgene expres- (day −21 p.i). Vectors were systemically administered by sion was induced by daily i.p. administration of 25 mg/kg doxycy- i.v. injection since the origin of the autoimmune response in cline (Dox) from the day of onset of neurological symptoms. a Mean daily clinical score of EAE mice treated with vehicle or Dox. b Aver- the EAE model is in the periphery [30, 31]. After carrying age daily weight change of EAE animals treated with vehicle or Dox out two independent experiments, a clinical improvement relative to weight prior to onset of neurological symptoms. Day 0 is was observed in the sIL21R-treated group compared to the considered the first day of Dox administration. Data were obtained control Null-treated group. This improvement was caused by from a single experiment (Vehicle: n = 6; Dox: n = 7). c Representa- tive flow cytometry histogram of ­eGFP+ hepatocytes of the different the 40% reduction in the EAE incidence of sIL21R-treated experimental groups. d Frequency and e MFI of eGFP.+ hepatocytes mice. These results were confirmed by histopathological obtained from healthy and EAE mice treated with AAV8.TetOn. analysis, in which sIL21R-treated mice showed significantly eGFP. Baseline values were provided by control animals without vec- less CD3 T cell infiltration, demyelination, microglial acti- tor and without Dox (n = 2). Data were obtained from a single experi- ment (Vehicle: n = 3; Dox (Healthy: n = 8; EAE: n = 8)). f Mean daily vation, and axonal damage in the CNS. Our results were in clinical score of mice administered with AAV8.TetOn.Null or AAV8. line with a study in the EAE model where daily exogenous TetOn.sIL21R vectors and treated with Dox. g Average daily weight administration of IL-21 during 7 days prior to immuniza- change of EAE-incident mice administered with AAV8.TetOn.Null tion triggered a more severe clinical outcome [32]. Simi- or AAV8.TetOn.sIL21R vectors and treated with Dox relative to weight prior to onset of neurological symptoms. Day 0 was consid- lar results were found in a study in which knockout (KO) ered the first day of Dox administration. Data were obtained from a mice for IL-21 (IL-21−/−) developed milder EAE [33]. In single experiment (Null: n = 7; sIL21R: n = 7). Data are represented addition, a more recent study in the spontaneous EAE 2D2 as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001 IL-21R−/− mice determined that IL-21 signalling is criti- cal for the development of the pathology [34]. However, some studies in IL-21−/− and IL-21R−/− mice have showed applied. In addition, it is worth mentioning that KO ani- contradictory results [35, 36]. Thus, there are studies show- mals for elements of the immune system eventually develop ing that the absence of the IL-21 signalling induces the compensatory immune mechanisms capable of replacing the development of more severe clinical signs. Due to it, these functions of the missing protein [37, 38]. This highlights the contradictory results may be explained by the different importance of achieving homogeneous experimental designs genetic backgrounds of the transgenic animals, differences for comparative purposes, especially for strategies involving in immunization protocols and experimental approaches pleiotropic cytokines such as IL-21. Fig. 6  The administration of AAV vectors close to the induction of incident mice of all groups. b Null-treated mice at day −1 and day +1 EAE produces a reduction in clinical severity. Mice were immunized developed a milder disease compared with Null group preventively with ­MOG40-55 and treated with 1­ 012 vg/mouse of AAV8.CAG.Null treated at day −21. Data were obtained from a single experiment 21 days before (day −21), the day before (day −1) or the day after (Null (day −21): n = 7; Null (day −1): n = 7; Null (day +1): n = 6). (day +1) the induction of EAE. a Mean daily clinical score of EAE- Data are represented as mean ± SEM 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1629 Subsequently, immunological studies were performed cells in the periphery, contrary to expected. However, this on splenocytes in the chronic phase (day 27 p.i) and fur- effect could not be seen during the acute phase. In preven- ther extended in the acute phase (day 15 p.i) of the experi- tive approaches, the reduction of anti-inflammatory path- mental disease. In the chronic phase, the sIL21R-treated ways, such as those mediated by Treg cells, is a common group showed a reduction in the number of IL-10+ CD4 T occurrence. In a recent study, it was observed that mice 13 1630 Á. Edo et al. that had been protected from EAE showed a reduction in an increase either in the percentage of IL-4+ T cells or in the frequency of Treg cell in the periphery [39] whereas in the levels of IL-4 in the supernatant of ­MOG40-55 stimu- another study this phenomenon is justified by the migration lated splenocytes. This could be due to a delay between the of these suppressor cells into the CNS [40]. In our case, we increase in mRNA expression and the increase of activa- hypothesized that the sIL21R-treated mice experienced a tion and synthesis of cytokines characteristic of this cel- lower initial inflammatory response, which required a lower lular pathway. We also observed a reduction in the number compensatory response mediated by IL-10. In this clinical of activated NK cells in the periphery of sIL21R-treated stage, we also observed an increase in TNFα and IL-18, mice in the acute phase, which would evidence that IL-21 and a trend towards increased IFNγ levels in the superna- activity was reduced by the treatment since this cytokine tant of ex vivo MOG-stimulated splenocytes. Interestingly, has a stimulatory effect on NK cell activation, cytotoxicity increased levels of TNFα and IFNγ cannot be considered as and IFNγ production [55]. a univocally negative effect of the treatment as these mol- Mice treated with sIL21R also showed increased periph- ecules have been associated to controversial roles in EAE eral activation of DCs in the acute phase, which is in line and MS. Anti-TNFα treatments were found to be effective with the capacity of IL-21 to inhibit the maturation and acti- in the passive transfer-induced EAE model, contrary to the vation of DCs by reducing surface expression of MHCII, obtained results by active induction [41, 42]. Even exog- CD80 and CD86 and TNFα secretion [56, 57]. This DC enous administration of TNFα to TNFα KO mice rescued activation status could be related with the increased expres- them from further severe EAE, while clinical trials in MS sion of Tbx21, the main Th1 cell transcription factor, in this patients treated with a soluble TNFα receptor (Lenercept) clinical stage. resulted in clinical worsening and had to be discontinued The next step was to study the effect of constitutive [43, 44]. On the other hand, although passive transfer of sIL21R expression in a therapeutic approach using AAV8 IFNγ-producing autoreactive Th1 cells can induce EAE vectors. Previous results from our group showed that [45], anti-IFNγ treatments cause an exacerbation of clini- transgene expression was significantly lower when viral vec- cal symptoms [46–48] and exogenous administration of tors were administered after immunization [25]. Therefore, IFNγ generates a protective effect in EAE [47, 48]. In con- we increased hepatic expression of the viral vectors by using trast, the administration of IFNγ in MS patients results in the CAG promoter as well as administered a double viral an exacerbation of disease [49]. Furthermore, IFNγ has dose to the animals. We administered the vectors close to been shown to inhibit the development of Th17 cells [50], the induction of EAE (days −1 and +1 p.i) to synchronize although we did not observed such effect. the time of maximum transgene expression with the onset of In the acute phase, sIL21R-treated mice showed clinical symptomatology. In contrast to the results obtained increased gene expression of TGFβ-1 in the periphery, a in the prophylactic approach, the sIL21R-treated incident potent tolerance inducer whose main function is to reduce mice developed the disease earlier than control-treated mice CD4 T cell activation in this clinical stage [51, 52]. This and the clinical course was more severe. This result was in would explain the reduction in the percentage of activated line with another study in a relapsing–remitting model of CD4 T cells observed in the periphery. It is also widely EAE in which IL21R-Fc was administered 2 days before accepted that TGFβ is a key molecule in the balance of the immunization and also revealed a worsening of the clinical Th17/Treg response, dependent on the presence of IL-6, outcome [58]. Bioavailability of a specific cytokine and the but also IL-21 [53]. Although the treatment produced a relative abundance of its receptor in target cells is frequently trend towards increased FoxP3 expression in the periphery, associated with the variation of the cytokine functionality. we could not observe an alteration in the frequency of Treg This is the case of IL-2, an essential cytokine for Treg cell or Th17 cells by flow cytometry as well nor in the gene functionality and survival [59], since its partial blockade expression of RORγ t. Although IL-21 is able to induce increases the expansion and suppressive function of this cell differentiation towards Th17 in the presence of TGFβ, this type [60]. Taking this into account and since AAV vectors function can be effectively compensated by the presence can initiate their expression within few days after admin- of IL-6 [35]. This could explain why IL-21 blockade in istration [61], we believed that AAV administration close monotherapy does not produce an effective regulation of to EAE induction lead to a partial blockade of IL-21 that this immune population. This idea is supported by a pre- resulted in a potentiation effect of the pro-inflammatory vious study that combined a treatment of IL-6 and IL-21 function of the cytokine. blockade in the collagen-induced RA model triggering a Furthermore, the reduction in clinical severity observed synergistic effect and significantly decreasing the num- in the Null-treated groups administered close to the immu- ber of Th17 cells [54]. Our gene expression studies also nization day compared to the Null-treated group adminis- showed an increase in Gata3, a characteristic factor of the tered 21 days before EAE immunization, led us to hypoth- Th2 cell population, but this result was not supported by esize that an immune response induced by the viral capsid 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1631 likely reduces the response against the encephalitogenic encephalomyelitis; GFAP: Glial fibrillary acidic protein; i.p.: Intra- peptide. An additional hypothesis that could explain this peritoneal; i.v.: Intravenous; IL-21: Interleukin 21; IL-21R: Interleukin 21 receptor; KO: Knockout; LEA: lycopersicon esculentum Lectin; result is the increase in cortisol levels in the mice gen- MBP: Myelin basic protein; MFI: Median fluorescence intensity; erated by the administration of the viral vectors itself, MOI: Multiplicity of infection; MS: Multiple sclerosis; NK: Natural as such stress could also lead to a reduction in clinical killer; p.i: Postimmunization; RA: Rheumatoid arthritis; SLE: Sys- severity. To address this issue, we designed a therapeutic temic lupus erythematosus; SD: Standard deviation of the mean; SEM: Standard error of the mean; sIL21R: Soluble interleukin 21 approach where we administered a Dox-inducible system, receptor; Treg: T regulatory; vg: Viral genomes AAV8.TetOn.sIL21R, 14 days before EAE induction and we induced Dox-mediated transgene expression at the Supplementary Information The online version contains supplemen- onset of the neurological signs. tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 13311-0​ 22-0​ 1279-8. To study the effect of Dox in the onset and progres- sion of EAE, we analysed the clinical course and weight Required Author Forms Disclosure forms provided by the authors are available with the online version of this article. loss of the animals and found no differences. Furthermore, by assessing the eGFP hepatic expression, we concluded Funding Open Access Funding provided by Universitat Autonoma that the inducible system worked efficiently in the EAE de Barcelona. This project was supported by the Ministerio Ciencia model, validating its usefulness for future AAV-mediated Innovación, retos Sociedad (PI15/0271). A.E was a recipient of a VHIR therapeutic approaches in this model. Interestingly, mice fellowship. administered with AAV8.TetOn.sIL21R developed less Open Access This article is licensed under a Creative Commons Attri- severe clinical signs than the control group. However, bution 4.0 International License, which permits use, sharing, adapta- it seems that the contribution of IL-21 once the clinical tion, distribution and reproduction in any medium or format, as long symptoms are established is not as relevant as its con- as you give appropriate credit to the original author(s) and the source, tribution in the initiation of the immune response. Thus, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are although the highest levels of IL-21 are produced during included in the article's Creative Commons licence, unless indicated the onset of the immune response and in the acute phase otherwise in a credit line to the material. If material is not included in [36], the exogenous administration of IL-21 from day the article's Creative Commons licence and your intended use is not 9 p.i does not result in an appreciable worsening of the permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a clinical course [32]. On the other hand, IL-21 blockade by copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. therapeutic administration of IL21R-Fc has been effective in other animal models of RA [62] or SLE [63], produc- ing a reduction in clinical severity and an increase in the survival rate. Immunological studies should be carried References out to explain the clinical improvement observed in this approach. In addition, it would also be interesting to study 1. Walton C, King R, Rechtman L, et al. Rising prevalence of mul- the effect of the sIL21R-inducible vectors in the different tiple sclerosis worldwide: Insights from the Atlas of MS, third edition. Mult Scler. 2020;26:1816–21. inflammatory phases in a relapsing–remitting EAE model. 2. Filippi M, Bar-Or A, Piehl F, et al. Multiple sclerosis. Nat Rev Dis To sum up, although further clinical and immunological Primers. 2018;4:1–27. studies are needed to define the suitable therapeutic time 3. Traugott U, Reinherz EL, Raine CS. Multiple sclerosis. Distri- window for sIL21R treatment, this gene therapy strategy bution of T cells, T cell subsets and Ia-positive macrophages in lesions of different ages. J Neuroimmunol. 1983;4:201–21. has shed light on the pathogenic role of IL-21 in the dif- 4. Tada T, Takemori T, Okumura K, Nonaka M, Tokuhisa T. Two ferent clinical and immunological stages of EAE. It may distinct types of helper T cells involved in the secondary antibody represent a breakthrough in the current MS therapeutic land- response: independent and synergistic effects of Ia- and Ia+ helper scape where most treatments involve the use of antibodies T cells. J Exp Med. 1978;147:446–58. 5. Bettelli E, Carrier Y, Gao W, et al. Reciprocal developmental and recombinant proteins. Since current therapies involve pathways for the generation of pathogenic effector TH17 and regu- long-term routine administration, which explains the large latory T cells. Nature. 2006;441:235–8. financial outlay, and the ability to respond to these treat- 6. Frischer JM, Bramow S, Dal-Bianco A, et al. The relation between ments declines as the disease progresses [64], making the inflammation and neurodegeneration in multiple sclerosis brains. Brain. 2009;132:1175–89. effectiveness-safety ratio less favorable over time. AAV gene 7. Rosenblum MD, Remedios KA, Abbas AK. Mechanisms of therapy treatments might be a great promise for the long- human autoimmunity. J Clin Invest. 2015;125:2228–33. term treatment of MS. 8. Spolski R, Leonard WJ. Interleukin-21: a double-edged sword with therapeutic potential. Nat Rev Drug Discov. 2014;13:379–95. Abbreviations AAV: Adeno-associated vector; Ad5: Adenovirus 5; 9. Long D, Chen Y, Wu H, Zhao M, Lu Q. Clinical significance AUC​: Area under curve; CNS: Central nervous system; DCs: Den- and immunobiology of IL-21 in autoimmunity. J Autoimmun. dritic cells; Dox: Doxycycline; EAE: Experimental autoinmune 2019;99:1–14. 13 1632 Á. Edo et al. 10. Ivanov II, Zhou L, Littman DR. Transcriptional regulation of utility for understanding disease pathophysiology and treatment. Th17 cell differentiation. Semin Immunol. 2007;19:409–17. Handb Clin Neurol. 2014;122:173–89. 11. Wei L, Laurence A, Elias KM, O’Shea JJ. IL-21 is produced by 32. Vollmer TL, Liu R, Price M, Rhodes S, La Cava A, Shi FD. Th17 cells and drives IL-17 production in a STAT3-dependent Differential effects of IL-21 during initiation and progres- manner. J Biol Chem. 2007;282:34605–10. sion of autoimmunity against neuroantigen. J Immunol. 12. Zotos D, Coquet JM, Zhang Y, et al. IL-21 regulates germinal 2005;174:2696–701. center B cell differentiation and proliferation through a B cell- 33. Nurieva R, Yang XO, Martinez G, et al. Essential autocrine regu- intrinsic mechanism. J Exp Med. 2010;207:365–78. lation by IL-21 in the generation of inflammatory T cells. Nature. 13. Zeng R, Spolski R, Finkelstein SE, et al. Synergy of IL-21 and 2007;448:480–3. IL-15 in regulating CD8+ T cell expansion and function. J Exp 34. Lee Y, Mitsdoerffer M, Xiao S, Gu G, Sobel RA, Kuchroo VK. IL- Med. 2005;201:139–48. 21R signaling is critical for induction of spontaneous experimental 14. Attridge K, Wang CJ, Wardzinski L, et al. IL-21 inhibits T cell IL-2 autoimmune encephalomyelitis. J Clin Invest. 2015;125:4011–20. production and impairs Treg homeostasis. Blood. 2012;119:4656–64. 35. Coquet JM, Chakravarti S, Smyth MJ, Godfrey DI. Cutting edge: 15. Nohra R, Beyeen AD, Guo JP, et al. RGMA and IL21R show asso- IL-21 is not essential for Th17 differentiation or experimental ciation with experimental inflammation and multiple sclerosis. autoimmune encephalomyelitis. J Immunol. 2008;180:7097–101. Genes Immun. 2010;11:279–93. 36. Liu R, Bai Y, Vollmer TL, et al. IL-21 receptor expression deter- 16. Tzartos JS, Craner MJ, Friese MA, et al. IL-21 and IL-21 receptor mines the temporal phases of experimental autoimmune encepha- expression in lymphocytes and neurons in multiple sclerosis brain. lomyelitis. Exp Neurol. 2008;211:14–24. Am J Pathol. 2011;178:794–802. 37. Kreiner G. Compensatory mechanisms in genetic models of neu- 17. Tegla CA, Cudrici CD, Azimzadeh P, et al. Dual role of Response rodegeneration: are the mice better than humans? Front Cell Neu- gene to complement-32 in multiple sclerosis. Exp Mol Pathol. rosci. 2015;9:1–6. 2013;94:17–28. 38. El-Brolosy MA, Stainier DYR. Genetic compensation: a phenomenon 18. Romme Christensen J, Börnsen L, Ratzer R, et al. Systemic in search of mechanisms. PLoS Genet. 2017;13:1–17. inflammation in progressive multiple sclerosis involves follicular 39. Zhu J, Liu JQ, Liu Z, et al. Interleukin-27 gene therapy prevents T-helper, Th17- and activated B-cells and correlates with progres- the development of autoimmune encephalomyelitis but fails sion. PLoS ONE. 2013;8:1–11. to attenuate established inflammation due to the expansion of 19. Zhang X, Tao Y, Chopra M, et al. Differential reconstitution of ­CD11b+Gr-1+ myeloid cells. Front Immunol. 2018;9:1–12. T cell subsets following immunodepleting treatment with alemtu- 40. Seifert HA, Gerstner G, Kent G, Vandenbark AA, Offner H. zumab (anti-CD52 monoclonal antibody) in patients with relapsing- Estrogen-induced compensatory mechanisms protect IL-10-deficient remitting multiple sclerosis. J Immunol. 2013;191:5867–74. mice from developing EAE. J Neuroinflammation. 2019;16:195–209. 20. Piedra J, Ontiveros M, Miravet S, Penalva C, Monfar M, Chillon 41. Teuscher C, Hickey WF, Korngold R. An analysis of the role of M. Development of a rapid, robust, and universal picogreen-based tumor necrosis factor in the phenotypic expression of actively method to titer adeno-associated vectors. Hum Gene Ther Meth- induced experimental allergic orchitis and experimental allergic ods. 2015;26:35–42. encephalomyelitis. Clin Immunol Immunopathol. 1990;54:442–53. 21. Chillon M, Alemany R. Methods to construct recombinant adeno- 42. Selmaj K, Raine CS, Cross AH. Anti-tumor necrosis factor virus vectors. Methods Mol Biol. 2011;737:117–38. therapy abrogates autoimmune demyelination. Ann Neurol. 22. Calvo-Barreiro L, Eixarch H, Ponce-Alonso M, et al. A commercial 1991;30:694–700. probiotic induces tolerogenic and reduces pathogenic responses in 43. van Oosten BW, Barkhof F, Truyen L, et al. Increased MRI activ- experimental autoimmune encephalomyelitis. Cells. 2020;9:1–22. ity and immune activation in two multiple sclerosis patients 23. Baker D, Amor S. Publication guidelines for refereeing and report- treated with the monoclonal anti-tumor necrosis factor antibody ing on animal use in experimental autoimmune encephalomyelitis. cA2. Neurology. 1996;47:1531–4. J Neuroimmunol. 2012;242:78–83. 44. Arnason BGW, Jacobs G, Hanlon M, et al. TNF neutralization 24. Ivanov II, McKenzie BS, Zhou L, et al. The orphan nuclear recep- in MS: results of a randomized, placebo-controlled multicenter tor RORgammat directs the differentiation program of proinflam- study. The Lenercept Multiple Sclerosis Study Group and The matory IL-17+ T helper cells. Cell. 2006;126:1121–33. University of British Columbia MS/MRI Analysis Group. Neurol- 25. Miralles M, Eixarch H, Tejero M, et al. Clinical and histopatho- ogy. 1999;53:457–65. logical amelioration of experimental autoimmune encephalomy- 45. Domingues HS, Mues M, Lassmann H, Wekerle H, Krishnamoorthy G. elitis by AAV vectors expressing a soluble interleukin-23 receptor. Functional and pathogenic differences of Th1 and Th17 cells in experi- Neurotherapeutics. 2017;14:1095–106. mental autoimmune encephalomyelitis. PLoS ONE. 2010;5:1–13. 26. Livak KJ, Schmittgen TD. Analysis of relative gene expression 46. Lublin FD, Knobler RL, Kalman B, et al. Monoclonal anti-gamma data using real-time quantitative PCR and the 2(-Delta Delta interferon antibodies enhance experimental allergic encephalomy- C(T)) Method. Methods. 2001;25:402–8. elitis. Autoimmunity. 1993;16:267–74. 27. Miao CH, Snyder RO, Schowalter DB, et al. The kinetics of rAAV 47. Heremans H, Dillen C, Groenen M, Martens E, Billiau A. Chronic integration in the liver. Nat Genet. 1998;19:13–5. relapsing experimental autoimmune encephalomyelitis (CREAE) 28. Xiao X, Li J, Samulski RJ. Production of high-titer recombinant in mice: enhancement by monoclonal antibodies against interferon- adeno-associated virus vectors in the absence of helper adenovi- gamma. Eur J Immunol. 1996;26:2393–8. rus. J Virol. 1998;72:2224–32. 48. Billiau A, Heremans H, Vandekerckhove F, et al. Enhancement 29. Spolski R, Leonard WJ. Interleukin-21: basic biology and of experimental allergic encephalomyelitis in mice by antibodies implications for cancer and autoimmunity. Annu Rev Immunol. against IFN-gamma. J Immunol. 1988;140:1506–10. 2008;26:57–79. 49. Panitch HS, Hirsch RL, Haley AS, Johnson KP. Exacerbations 30. Constantinescu CS, Farooqi N, O’Brien K, Gran B. Experi- of multiple sclerosis in patients treated with gamma interferon. mental autoimmune encephalomyelitis (EAE) as a model for Lancet. 1987;1:893–5. multiple sclerosis (MS). Br J Pharmacol. 2011;164:1079–106. 50. Yeh WI, McWilliams IL, Harrington LE. IFNγ inhibits Th17 dif- 31. Robinson AP, Harp CT, Noronha A, Miller SD. The experi- ferentiation and function via Tbet-dependent and Tbet-independent mental autoimmune encephalomyelitis (EAE) model of MS: mechanisms. J Neuroimmunol. 2014;267:20–7. 13 Therapeutic Effect of IL‑21 Blockage by Gene Therapy in Experimental Autoimmune… 1633 51. Gilbert KM, Thoman M, Bauche K, Pham T, Weigle WO. Trans- 58. Piao WH, Jee YH, Liu RL, et al. IL-21 modulates CD4+ CD25+ forming growth factor-beta 1 induces antigen-specific unrespon- regulatory T-cell homeostasis in experimental autoimmune siveness in naive T cells. Immunol Invest. 1997;26:459–72. encephalomyelitis. Scand J Immunol. 2008;67:37–46. 52. Chen CH, Seguin-Devaux C, Burke NA, et al. Transforming 59. Fan MY, Low JS, Tanimine N, et al. Differential roles of IL-2 signaling growth factor beta blocks Tec kinase phosphorylation, Ca2+ in developing versus mature tregs. Cell Rep. 2018;25:1204–13. influx, and NFATc translocation causing inhibition of T cell dif- 60. Wilkinson DS, Ghosh D, Nickle RA, Moorman CD, Mannie MD. ferentiation. J Exp Med. 2003;197:1689–99. Partial CD25 antagonism enables dominance of antigen-inducible 53. Gutcher I, Donkor MK, Ma Q, Rudensky AY, Flavell RA, Li MO. ­CD25high ­FOXP3+ regulatory T cells as a basis for a regulatory T Autocrine transforming growth factor-β1 promotes in vivo Th17 cell-based adoptive immunotherapy. Front Immunol. 2017;8:1–21. cell differentiation. Immunity. 2011;34:396–408. 61. Zincarelli C, Soltys S, Rengo G, Rabinowitz JE. Analysis of AAV 54. Roeleveld DM, Marijnissen RJ, Walgreen B, et al. Higher efficacy serotypes 1–9 mediated gene expression and tropism in mice after of anti-IL-6/IL-21 combination therapy compared to monother- systemic injection. Mol Ther. 2008;16:1073–80. apy in the induction phase of Th17-driven experimental arthritis. 62. Young DA, Hegen M, Ma HL, et al. Blockade of the interleukin-21/ PLoS ONE. 2017;12:1–16. interleukin-21 receptor pathway ameliorates disease in animal models 55. Brady J, Hayakawa Y, Smyth MJ, Nutt SL. IL-21 induces of rheumatoid arthritis. Arthritis Rheum. 2007;56:1152–63. the functional maturation of murine NK cells. J Immunol. 63. Bubier JA, Bennett SM, Sproule TJ, et al. Treatment of BXSB-Yaa 2004;172:2048–58. mice with IL-21R-Fc fusion protein minimally attenuates systemic 56. Brandt K, Bulfone-Paus S, Jenckel A, Foster DC, Paus R, Rückert lupus erythematosus. Ann N Y Acad Sci. 2007;1110:590–601. R. Interleukin-21 inhibits dendritic cell-mediated T cell activation 64. Rieckmann P, Traboulsee A, Devonshire V, Oger J. Escalating and induction of contact hypersensitivity in vivo. J Invest Derma- immunotherapy of multiple sclerosis. Ther Adv Neurol Disord. tol. 2003;121:1379–82. 2008;1:181–92. 57. Strengell M, Lehtonen A, Matikainen S, Julkunen I. IL-21 enhances SOCS gene expression and inhibits LPS-induced Publisher's Note Springer Nature remains neutral with regard to cytokine production in human monocyte-derived dendritic cells. jurisdictional claims in published maps and institutional affiliations. J Leukoc Biol. 2006;79:1279–85. 13