Therapeutic Benefit of Galectin-1: Beyond Membrane Repair, a Multifaceted Approach to LGMD2B

Two of the main pathologies characterizing dysferlinopathies are disrupted muscle membrane repair and chronic inflammation, which lead to symptoms of muscle weakness and wasting. Here, we used recombinant human Galectin-1 (rHsGal-1) as a therapeutic for LGMD2B mouse and human models. Various redox and multimerization states of Gal-1 show that rHsGal-1 is the most effective form in both increasing muscle repair and decreasing inflammation, due to its monomer-dimer equilibrium. Dose-response testing shows an effective 25-fold safety profile between 0.54 and 13.5 mg/kg rHsGal-1 in Bla/J mice. Mice treated weekly with rHsGal-1 showed downregulation of canonical NF-κB inflammation markers, decreased muscle fat deposition, upregulated anti-inflammatory cytokines, increased membrane repair, and increased functional movement compared to non-treated mice. Gal-1 treatment also resulted in a positive self-upregulation loop of increased endogenous Gal-1 expression independent of NF-κB activation. A similar reduction in disease pathologies in patient-derived human cells demonstrates the therapeutic potential of Gal-1 in LGMD2B patients.


Introduction
Dysferlin is a 230-kDa protein that is highly expressed in skeletal muscle and involved in membrane repair [1]. Membrane repair is essential for maintaining cell integrity and is crucial for stressed skeletal muscle fibers [2][3][4]. Mutated dysferlin leads to compromised membrane integrity and repair, resulting in the phenotypes associated with Miyoshi Myopathy (MM) and Limb-Girdle Muscular Dystrophy 2B (LGMD2B). In addition to diminished membrane repair, muscles in LGMD2B are characterized by wasting, fatty infiltrates (especially in the hip and lower leg), and chronic inflammation [5][6][7][8][9].
Chronic inflammation is responsible for many pathologies seen in LGMD2B [10][11][12]. In particular, the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling complex is highly upregulated in this disease [13]. The NF-κB signal can work through either the canonical or non-canonical pathway. Studies show that Galectin-1 activates the canonical NF-κB pathway in osteoarthritis chondrocytes [14]. The non-canonical NF-κB pathway is associated with the production of pro-inflammatory cytokines in cardiac tissues and the development of tumor necrosis [15,16]. In this study, we Recombinant Human Galectin-1 (rHsGal-1) was produced and purified as described by Vallecillo-Zúniga et al. [31]. In summary, LGALS1 gblock was cloned into the pET29b(+) vector and transformed into BL21(DE3) competent E. coli cells, grown, and induced with 0.1 mM IPTG. rHsGal-1purification was accomplished as described in Stowell et al. [32]. rHsGal-1 bacterial lysate was passed through a lactoyl-Sepharose column (Sigma-Aldrich, St. Louis, MO, USA). Galectin was eluted with 14 mM β-mercaptoethanol (β-ME) and 100 mM Lactose in 1X PBS. Protein fraction absorbance was read at 280 nm on BioPhotometer (Eppendorf, Hamburg, Germany), and fractions with an absorbance of greater than 0.5 were kept and stored at −80 • C. To activate the CRD, factions were passed over a PD-10 desalting column (GE Health Care, Chicago, IL, USA) and stored at 4 • C.
Monomeric Galectin-1 (mGal-1) was constructed, produced, and generously provided by Dr. S. Stowell at Emory University. This was achieved by capping the N-terminus with Lys, which severely inhibits Gal-1 dimerization [33]. Dimeric Galectin-1 (dGal-1) was constructed using a 2 Gly residue linking the N-terminus of one subunit to the C-of the other. The rHsGal-1GG construct, which contains two rHsGal-1 CRDs, was connected by a glycine-glycine linker that maintains the CRD orientation of wild-type galectin-1. The nucleotide sequences were constructed in SnapGene (GSL Biotech, Chicago, IL). Vector assembly was completed by Twist Bioscience (San Francisco, CA, USA). dGal-1 purification was accomplished as described in Stowell et al. [32]. mGal-1 and dGal-1 forms were either left in a native reduced form or oxidized using 1 µM H 2 O 2 [27].

Animal Care
Three-to nine-month-old male Bla/J mice (Dysf −/− (B6.129-Dysf tm1Kcam /J)) were provided by the Jain Foundation. Mice were housed in an approved facility at Brigham Young University. All procedures were performed in accordance with Brigham Young University IACUC-approved protocol. Purified rHsGal-1 was used as a treatment in in vivo studies. Protein concentration was analyzed using a Bradford assay protocol following the manufacturer's instructions. BLA/J mice were weighed and dosed with various doses of rHsGal-1. Treatment was delivered with 1 mL insulin syringes (BD, San Jose, CA, USA Cat #329420) intraperitoneally (IP).

Cell Culture
Immortalized primary H2K A/J −/− (A/J −/− ) myoblasts were cultured as described in Vallecillo-Zúniga et al. [31]. Mouse myoblasts were plated onto T75 flasks (SARSTED No. 83.3911.302, Newton, NC, USA) or standard 35 mm glass-bottom dishes (Mat Tek, No. 1.5 Coverslip, 14 mm Glass Diameter, Uncoated. Ashland, MA, USA), seeding at a density of 5555 cells/cm 2 and incubated at 33 • C in 10% CO 2 . At 80-90% confluency, myoblasts were differentiated and treated with or without 0.11 µM rHsGal-1 for 48-72 h at 37 • C and 5% CO 2 . Twitching of myotubes was quantified using ImageJ by selecting 10 different fields of view and calculating the number of visible twitching myotubes over the total number of myotubes present.

Muscle Fiber Isolation
Muscle fiber isolation was performed as described in Vallecillo-Zúniga et al. [31]. In summary, a 6-well plate (Cat. No. 665 180, Grenier Bio-One) was used as a digestion plate and prepared as described in Demonbreun et al. [35]. When the mice were sacrificed, the flexor digitorum brevis (FDB) was excised. Non-muscle tissues, including tendon, nerve, and overlying fascia, were carefully removed, and muscles were incubated in Collagenase II (2.5 U/mL, ThermoFisher, #17101015) in DMEM for 60-90 min at 37 • C. Next, by using a small-bore pipette, the fibers were transferred to 35 mm glass-bottom microwell dishes (Cat. No. P35GCol-1.0-14-C, MatTek, Ashland, MA, USA) and allowed to attach for at least 15 min.

RayBio Mouse Inflammation Array
Media from NT and rHsGal-1 treated A/J −/− myotubes were collected after 48 h. Cytokine expression was measured using the Mouse Inflammatory Array C1 (RayBiotech, Cat.No.AAM-INF-1-8, Peachtree, GA, USA) according to the manufacturer's instructions. Membranes were imaged using a FluorChem imaging system (Alpha Innotech, San Jose, CA, USA). The membranes were quantified using ImageJ software as described in Schindelin et al. [36].

Immunofluorescence
A/J −/− myoblasts were cultured onto 35 mm glass-bottom microwell dishes (Cat. No. P35GCol-1.0-14-C, MatTek, Ashland, MA, USA), fixed in 4% paraformaldehyde, permeabilized in 0.1% triton X-100 (in PBS), and blocked using PBST blocking solution for 1 h at room temperature. The myotubes were then incubated overnight at 4 • C with the appropriate primary antibody: NF-κB p65 ( No. 14-6657-80 Invitrogen, Rockford, IL, USA, 1:1000), were diluted in the wash buffer (10% EZBlock in PBST). The slides were left with the primary antibody overnight at 4 • C. The following day, the slides were washed three times with the wash buffer for 5 min, after which the slides were blocked with 3% hydrogen peroxide in methanol. The slides were then washed three times with PBS for 5 min each. The slides were then treated with the HRP conjugated secondary antibody (1:500, Invitrogen, Rockford, IL, USA, Cat. No. G-21234) and left for 1 h at room temperature. The slides were then rinsed once with PBST; after which they were washed with PBST twice PBST for 5 min each. The slides were then washed with PBS once for 5 min and then rinsed once with PBS. DAB (Thermo Scientific, 34002) was prepared according to the manufacturer's instructions. The slides were treated with DAB for 15 min and counterstained with hematoxylin for 4 min, washed with 100% ethanol, allowed to air dry, and then washed three times with Histo-Clear for 5 min each. Finally, the slides were mounted using Organo/Limonene mounting medium (Sigma-Aldrich, Burlington, MA, O8015) and cured overnight at 37 • C before imaging.

Histology
Bla/J psoas muscles were dissected, embedded in optimum cutting temperature compound (OCT, Cat. No.4583, Sakura, Torrance, CA, USA), and frozen isopentane cooled in a liquid nitrogen bath. Ten micrometers of frozen tissue were cut using a TNF50 Semi-Automatic Cryostat (Tanner Scientific). Sections were placed on 3P white extra slides (Cat. No. 3800200, Leica Microsystems, Buffalo Grove, IL, USA). Sections were processed for immunofluorescence. Digital images were captured with an Olympus microscope (Model BX51).

Statistical Analyses
Data analyses were completed by using Tukey's multiple comparison test 1-way and 2-way ANOVA and Student's t-test through the GraphPad Prism Software, version 9.0. For membrane repair analysis, the data conferred are the averaged values for all the myotubes used in the analysis and the treatment at individual time points. p-values are indicated in the figure when statistical significance is determined for all groups as described in the figure legends.

Reduced Dimeric Galectin-1 Is Responsible for Optimal Membrane Repair
We designed several synthetic forms of Gal-1 to determine the effect of dimerization and redox states on membrane repair in models of dysferlinopathy ( Figure S1). These forms include wild type Gal-1 (WTGal-1), transiently reduced recombinant Human Gal-1 (rHsGal-1), fixed monomeric Gal-1 (mGal-1), and fixed dimeric Gal-1 (dGal-1). WTGal-1 and rHsGal-1 are structurally identical except for the addition of a 6X His-tag on the C-terminus of rHsGal-1 and are comprised of the human LGALS1 gene. For mGal-1, an induced V5D mutation prevents dimerization, as shown by Cho et al. [33]. We formed dGal-1 by fusing two human Gal-1 constructs with a flexible Gly-Gly linker, as described in Earl et al. [37].
Previous work from our lab showed that rHsGal-1 with a 6X his-tag diluted in PBS at a pH of 7.4 beneficially increased membrane repair in dysferlin-deficient myotubes and explants. However, we did not evaluate the effects such as the 6X His tag, alkylation, dimerization, and redox state [31]. Stowell et al. showed that alkylation of Gal-1 provides permanent protein stability through the prevention of cysteine residue oxidation, thus stabilizing the CRD of Gal-1 [28,31]. β-ME is a well-known temporary reducing agent that helps stabilize the protein of interest by cleaving the disulfide bonds between cysteine residues [38]. As explained in the methods section, we purified transiently reduced rHsGal-1 (under a β-ME environment) to ensure protein stability. Prior to the experimental use of rHsGal-1, β-ME was removed by passing the protein through a PD-10 column. To test the activity of rHsGal-1 and WT Gal-1 in membrane repair, A/J −/− myotubes were treated for 48 h with these two forms of Gal-1. After injury, no significant differences were observed in fluorescent intensity between rHsGal-1 and WT Gal-1 treated myotubes ( Figure 1A).
We next sought to compare the effects of permanently reduced rHsGal-1 (alkylated rHsGal-1) and rHsGal-1 on membrane repair. After 48 h treatment, myotubes were subjected to a laser injury. The change in fluorescent intensity was nearly identical between the alkylated rHsGal-1 and rHsGal-1 ( Figure 1B). Additionally, we found that the final fluorescence intensity between rHsGal-1 and WTGal-1 was 54% and 40% lower when compared to non-treated myotubes, but not statistically significant between the two forms. These results indicate no deleterious effects from the His-tag or alkylation.
The known interactions of Gal-1 are irrevocably connected to the oxidation state. Thus, identifying the relationship between membrane repair and oxidation state is vital to therapeutic development. Previously, our lab showed that a 10 min treatment with rHsGal-1 markedly improved membrane repair [31]. To test the relationship between oxidative state of mGal-1 and repair, we treated A/J −/− myotubes for 10 min with oxidized or reduced mGal-1, using rHsGal-1 as a positive control ( Figure 1C). We saw no improvement in myotubes treated with reduced mGal-1, but we observed an increased dye entry of 40% in myotubes treated with oxidized mGal-1 compared to NT myotubes, indicative of decreased repair even when compared to NT A/J −/− myotubes. We then treated A/J −/− myotubes for 48 h with either reduced mGal-1, oxidized mGal-1, or rHsGal-1 ( Figure 1D). After 48 h, we saw that both forms of mGal-1 decreased dye entry by 41% compared to NT myotubes, while rHsGal-1 treated myotubes decreased by 72%. The dichotomy of results from the 10 min versus the 48 h treatments with mGal-1 suggests that a longer treatment is necessary for membrane repair when using this form of Gal-1. This implies that signaling pathways must be activated for either redox state of mGal-1 to have a therapeutic effect. This is consistent with reports that monomeric Gal-1 is primarily involved in intracellular signaling [37,39].

rHsGal-1 Lowers Expression of Proteins in the NF-κB Pathway
Chronic inflammation is a common disease pathology of LGMD2B and a desirable therapeutic target. We hypothesized that Gal-1 would reduce certain NF-κB markers that are upregulated in LGMD2B. We first examined the ability of our various recombinant forms of Gal-1 to modulate the activation of the NF-κB pathway. A/J −/− myotubes were treated for 48 h with oxidized and reduced forms of Gal-1, and lysates were probed for p65. We found that only rHsGal-1, oxidized mGal-1, and oxidized dGal-1 reduced p65 levels (Figure 2A,B). These results were confirmed via immunofluorescent imaging of p65 in A/J −/− myotubes ( Figure 2C). Based on the fact that only rHsGal-1 produced optimal effects in both the injury repair assay and NF-κB pathway activation, we used this form in the rest of our experiments.
There are two activation pathways of NF-κB inflammation: canonical activation via TAK1 and non-canonical activation via NIK. It is unknown how Gal-1 regulates the noncanonical NF-κB pathway. As such, we wanted to determine which inflammatory pathway Gal-1 regulates in LGMD2B models. We achieved this by probing for the respective proteins of each pathway. The rHsGal-1 treatment lowered levels of TAK1 by 84%, but did not induce changes in NIK, indicating that the reduction in inflammation is due to modulation of the canonical NF-κB pathway ( Figure 2D,E). We further examined multiple proteins downstream of TAK1, including IKBα, p50, and phospho-p65 (P-p65) ( Figure 2F-I). These results show a dramatic decrease in inflammatory transcription factors p50 and P-p65, by 56% and 55%, respectively, as well as an increase in the inhibitory protein IKBα by 40%. To verify the presence of P-p65, we visualized its location using immunofluorescent microscopy. We saw a decreased expression of P-p65 in A/J −/− myotubes when treated with rHsGal-1 (Supplementary Figure S2). This overall reduction in inflammatory markers indicates that the monomeric rHsGal-1 is largely responsible for the likely decrease of NF-κB inflammation in cellular models of LGMD2B. 3.3. The Therapeutic Window for In Vivo rHsGal-1 Is from 1.35 to 8.1 mg/kg A broad therapeutic window is a desired trait for drug therapies. Based on our in vitro dosage and the dosages used in the mdx DMD model, we predicted the therapeutic dosage for our in vivo model to be 2.7 mg/kg [30,31]. Our previous work showed that an ex vivo treatment of myofibers taken from SJL/J (Dysf −/− ) and Bla/J mice 2 h prior to injury resulted in an increased membrane repair capacity [31]. Here, we used ex vivo membrane repair assays to determine the optimal rHsGal-1 dose and to define the in vivo therapeutic window in LGMD2B murine models ( Figures 3A,B, S3 and S4, and Table 1). Intraperitoneal injections of Gal-1 have been successfully utilized in other dystrophic mouse models [30].
Wuebbles et al. showed that intravenous (IV) rHsGal-1 doses greater than 2.5 mg/kg were lethal due to the ability of Gal-1 to induce hemostasis and thrombosis [40]. Thus, all doses were given intraperitoneally (IP). Day 7 (D7) doses were given 2 h prior to sacrificing the mouse, based on previous work in the Bla/J mouse model [31]. The change in fluorescent intensity for each rHsGal-1 treatment was normalized to PBS-treated groups to allow comparisons between experiments ( Figures 3B, S3 and S4). We found that most doses of rHsGal-1 either positively improved membrane repair or had no detrimental effect. Only 27 mg/kg rHsGal-1 given IP three times a week proved detrimental to membrane repair, although we observed no other signs of animal distress or toxicity. Three doses, 1.35 mg/kg, 2.7 mg/kg, and 5.4 mg/kg, given at Days 0 (D0) and 7, proved to have "clinically significant" impacts on membrane repair (greater than 2-fold change from PBStreated mice) ( Figure 3B, inset). Since 9 out of the 12 dosing schemes included a Day 7 treatment, we evaluated whether the rHsGal-1 dose provided 2 h prior to sacrifice was necessary to gain maximum improvement in membrane repair. Mice were treated on either Day 0 only, Day 7 only, or Days 0 and 7. Laser injury assay showed that the combined Days 0 and 7 treatments improved membrane repair the most (71% decrease in final fluorescence intensity). The individual Day 0 and Day 7 treatments also showed significant improvements to membrane repair (final fluorescence intensity decreased 25% and 47%, respectively). This suggests that rHsGal-1 provides immediate and cumulative benefits to membrane repair ( Figure 3).
To assess the relationship of treatment schedule to the amount of rHsGal-1 in tissues, we used a western blot analysis to probe for the His-tag present in our rHsGal-1. We found that tissues from mice treated on Days 0 and 7 had 87% more rHsGal-1 than mice treated on Day 0 and 220% more rHsGal-1 than mice treated on Day 7 ( Figure 3D,E). Oxidized 3,3 -Diaminobenzidine (H-DAB) staining showed that only tissues from Bla/J mice treated with rHsGal-1 on Days 0 and 7 had significantly more rHsGal-1 in the tissues than nontreated mice ( Figure 3F,G). These results show that rHsGal-1 has benefits at both 2 h and week-long treatments and has additive effects on membrane repair with both treatments.

One-Month rHsGal-1 Treatment Improves Translational and Biochemical Measures of LGMD2B
After determining that a 2x/week treatment of 2.7 mg/kg rHsGal-1 improved membrane repair, we tested its efficacy during a one-month study. Nine-month-old Bla/J mice were treated weekly with 2.7 mg/kg/wk rHsGal-1 for four weeks. At the end of the four-week study, we assessed membrane repair capacity, functional activity, histopathology, and inflammation. Using activity monitoring cages, we saw significant increases in rearing events (Z counts) and horizontal movement (X counts) after the one-month treatment with rHsGal-1 (1.22 and 1.54-fold difference, respectively; Figure 4B,C). Additionally, we found that membrane repair was improved with a final fluorescence decrease of 51% compared to PBS treated Bla/J mice ( Figure 4A). Studies in mdx mice and our previous study of A/J −/− myotubes showed that rHsGal-1 treatment resulted in a positive feedback loop resulting in upregulating endogenous Gal-1 [31]. To determine if this apparent positive feedback loop existed in vivo, we used RT-qPCR to determine Gal-1 transcription levels in the psoas, the most affected muscle in the Bla/J mouse model [41]. We found a 7.5 and 18-fold increase in LGALS1 after 1-week (D0, D7 regiment) and 1-month treatments, respectively ( Figure 4D). This data provided evidence of a positive feedback loop, where exogenous Gal-1 induces expression of endogenous Gal-1. An ELISA assay was used to measure the concentration of Gal-1 in serum. Results showed that there was a baseline concentration of Gal-1 of 6.9 ng/mL. A one-time IP treatment with 2.7 mg/kg revealed that the serum Gal-1 levels reached peak concentration at approximately 3 h and by 12 h had returned to baseline concentration. The half-life of rHsGal-1 treatment in the blood of Bla/J mice is 2.9 h ( Figure 4I).
In order to determine if the reduction in inflammatory markers seen in vitro was recapitulated in vivo, we probed for p65 in the psoas of mice treated with rHsGal-1 and PBS for one month. Immunofluorescence imaging revealed a reduction in the normalized area of p65 ( Figure 4G,H). This was confirmed via western blot of the gastrocnemius muscle ( Figure 4E,F). Additionally, the expression of other NF-κB proteins, p50 and phospho-p65, was significantly reduced.
Patient and animal model histopathology shows upregulation of fibroadipogenic progenitors in slow twitch myofibers, which correlates with disease pathophysiology [42][43][44]. To investigate the effect of rHsGal-1 on this aspect of the disease, we probed psoas tissue sections for perilipin, a marker of lipid infiltrate. We saw that the one-month treatment reduces perilipin area by 50% ( Figure 4J,K). Additionally, we observed that after 15 days, A/J −/− myotubes treated with rHsGal-1 showed significantly more twitching ability compared to NT myotubes ( Figure S5 and Supplementary Video S1). Together, these results show that treatment with rHsGal-1 improves muscle function (activity), muscle membrane repair, pathology, and inflammation.

rHsGal-1 Treatment Upregulates Anti-Inflammatory Cytokine Secretion In Vitro and In Vivo
To further investigate the effect of rHsGal-1 treatment on the NF-κB pathway, we quantified changes in secreted cytokines in dysferlin models with treatment. We tested cell culture media of A/J −/− myotubes NT or treated with 0.11 µM rHsGal-1 using a mouse cytokine profiler ( Figure 5A,B). This assay revealed that IL-4, CXCL-1, MCP-1, and TIMP-2 cytokines were upregulated during 48 h rHsGal-1 treatment by 15.5, 1.4, 1.5, and 1.5-fold, respectively ( Figure 5C). Although there are several functions for each cytokine, the commonality between them is that each has either anti-inflammatory properties or promotes tissue remodeling and regeneration ( Figure 5D). For example, IL-4 is a multifunctional cytokine critically involved in inflammation by promoting alternative macrophage activation [13,[45][46][47]. Our results show that IL-4 is the foremost upregulated cytokine in myotubes after 48 h rHsGal-1 treatment. To confirm the cytokine secretome results, we tested cytokine expression in tissue lysates from Bla/J mice treated with PBS or rHsGal-1 for one month. We found a significant increase in IL-4, MCP-1, and TIMP-2 in mice treated with rHsGal-1 compared to the PBS by 38.5, 1.9, and 1.4-fold, respectively ( Figure 5E-H). These results may explain the beneficial effect of rHsGal-1 treatment in inflammation and membrane repair.

rHsGal-1 Improves Membrane Repair in Dysferlin Deficient, Patient-Derived Myotubes
With the goal to bring Gal-1 treatment to patients, we used dysferlin-deficient patientderived myotubes to verify that the therapeutic effects measured in mouse models would translate to a human model. Patient myotubes treated with 0.11 µM rHsGal-1 experienced a 79% reduction in final fluorescent intensity in laser injury assay, indicating increased repair ( Figure 6A,B). Changes in inflammatory protein markers were investigated using immunofluorescence and western blots. Quantification of p65 confocal immunofluorescent images revealed a 47% reduction in p65 expression with the same trend using western blot analysis. (Figure 6C-F). There is a direct relationship between mouse muscle health with cage exploration and rearing [41]. Both of these indices were significantly improved in mice treated with rHsGal-1. We suspect this increase in movement is due to the ability of rHsGal-1 to reduce inflammation and promote muscle membrane repair. Mechanically, we believe the rHsGal-1 is working to facilitate the formation of the membrane patch, an integral component in membrane repair. Together, our results from the one-week and one-month experiments demonstrate that rHsGal-1 affects muscle membrane mechanically through Gal-1 localization to the membrane and biologically through inflammatory signaling [31]. More investigation is needed to show the proteins that rHsGal-1 interacts with as it mediates membrane repair. Regardless of the mechanism, the treatment with rHsGal-1 in BLA/J mice suggests an improvement in muscle health as evidenced in the laser injury, CLAMS cages, biochemical, and histological assays.
Furthermore, rHsGal-1 showed the same therapeutic potential when administered to patient-derived dysferlin-deficient cells ( Figure 6). rHsGal-1 shows promise at diminishing the symptoms of LGMD2B in two areas of pathology: inflammation and muscle membrane repair. This two-pronged mechanism would be extremely useful as a therapeutic and may stem from the ability of rHsGal-1 to function as either a monomer or a dimer. The dimer form of rHsGal-1 is clearly more beneficial in assisting in the membrane repair process, while the monomeric version helps to reduce the markers of inflammation. Although more testing is required, these two parallel processes position rHsGal-1 as a highly effective therapeutic against LGMD2B.

Discussion
Mutations of the dysferlin gene lead to impaired sarcolemma repair with limited treatment options [19,20]. In addition to defective membrane repair, muscles that lack functional dysferlin exhibit chronic muscle inflammation and abnormal lipid metabolism [41,[48][49][50]. Steroid, stem cell, or gene replacement therapies to restore the functionality of dysferlin are under investigation [22,51,52]. However, these treatments are still far from clinical application. Our previous study showed that in vitro and ex vivo rHsGal-1 treatment improves myogenesis and membrane repair in dysferlin-deficient models [31]. However, the mechanism behind this improvement is not clear. The multiplicative roles of Gal-1 align with our current results. Here, we provide evidence that rHsGal-1 acts via two discrete mechanisms: restoration of membrane integrity and decrease of inflammatory response in LGMD2B models.
We present data uncovering the biological activity of different types of Gal-1 in membrane repair and inflammation by testing multiple forms of Gal-1 in various oxidation states. An effective therapeutic for LGMD2B patients should address both muscle repair and chronic inflammation in order to reverse pathophysiology. Both rHsGal-1 and reduced dGal-1 effectively improved membrane repair in vitro and ex vivo ( Figure 1E-G). Because monomeric and oxidized dGal-1 treatments were ineffective, the reduced dimeric form of Gal-1 is likely responsible for the bulk enhancements in membrane repair in the rHsGal-1 treatment. This aligns with previous studies, which state that the CRD domain of Gal-1 is inactivated via oxidation and that the CRD is necessary for membrane repair [27,31]. The fact that rHsGal-1 and oxidized forms of mGal-1 and dGal-1 both reduced levels of p65 illustrates different niches for the various forms of Gal-1, with oxidized Gal-1 playing a larger role in inflammatory signaling. The dynamic nature of rHsGal-1 allows it to excel at both signaling and membrane repair.
We posit that micro-cellular environmental changes of the various monomer/dimer and redox states of rHsGal-1 are responsible for decreased LGMD2B manifestations. Because of the monomer/dimer modulation, we infer that Gal-1 assists in membrane repair processes as a dimer and simultaneously decreases inflammation as a monomer in our A/J −/− , Bla/J, and patient dysferlin-deficient models. Further studies on rHsGal-1 cellular surface and intracellular interactome are needed to deduce additional information on the pathways and mechanisms by which Gal-1 decreases inflammation.
Abnormal expression in the NF-κB pathway causes detrimental effects that accompany inflammation [53]. In accordance with previous studies, we gathered indirect evidence that Gal-1 is associated with downregulated NF-κB activity [31]. For example, phosphorylated p65, which is required for NF-κB relocation to the nucleus, was significantly downregulated in vitro and in vivo [51]. While encouraging, reductions in P-p65 and other elements of the NF-κB pathway do not always correlate with changes in nuclear NF-κB activity or changes in genes being transcribed. However, based on the data we have, we conclude that NF-κB pathway activation is likely also downregulated, although more research, including promotor studies, is needed to confirm this directly. The NF-κB pathway is activated by two different signaling cascades, the canonical and the non-canonical pathways, each with unique signaling and biological functions [54,55]. This study demonstrates that rHsGal-1 is affecting the NF-κB response through the canonical pathway, TAK-1, or the receptor for TAK-1. Since dysregulation of the non-canonical pathway is associated with lymphoid malignancies and autoimmune diseases, it is beneficial that rHsGal-1 is not eliciting this response.
NF-κB activation triggers gene expression for a broad range of pro-inflammatory cytokines, chemokines, and adhesion molecules; therefore, it is unsurprising that Gal-1 treatment affects all of these. The cytokines significantly upregulated with treatment (IL-4, CXCL1, MCP1, and TIMP-2) play a unique role in the cell and may further explain the therapeutic effect of exogenous Gal-1 treatment. IL-4 is a well-studied anti-inflammatory cytokine involved in myogenesis [13,46,47] and tissue repair. In a recently published study, direct treatment with IL-4 improved muscle differentiation [13,56]. This is interesting since Gal-1 treatment in mouse and human models of both Duchenne and LGMD2B both increase muscle differentiation. It should be noted that many cytokines and chemokines have alternative promotor elements that may have consequences unrelated to the NF-κB pathway, which require further investigation. Thus, it is reasonable to conclude that Gal-1 upregulation of Il-4 might drive these changes in differentiation.
Additionally, IL-4 is involved in changing the polarization of macrophages from M1 to M2 [57][58][59]. M1 macrophages are upregulated in dysferlin deficient muscle, which has been shown to contribute to muscle damage [60]. It is possible that rHsGal-1 might polarize macrophages in Bla/J mice through the upregulation of IL-4. This signaling may be the avenue that ultimately leads to greater muscle health. Therefore, upregulation of IL-4 in response to rHsGal-1 treatment may reduce the negative effects of chronic inflammation in LGMD2B and lead to greater muscle health ( Figure 5). The CXCL1, MCP-1, and TIMP-2 cytokines are involved in tissue regeneration, wound healing, and ECM regulation, respectively, all of which can contribute to overall muscle health in vivo [61][62][63]. We suspect that these cytokines contribute toward the therapeutic action of Gal-1, but more investigation is necessary.
Previous research provides evidence that activation of NF-κB can increase Gal-1 transcription [64]. Thus, seeing an increase in Gal-1 transcript would be expected when NF-κB is activated. Gal-1 is upregulated in diseases with chronic injuries, such as other types of muscular dystrophy [12]. Specifically, in LGMD2B, Gal-1 transcript levels are 3.8 times higher in patient tissue over non-diseased tissue [12]. Although the level of endogenous Gal-1 in Bla/J mice is unknown, WT and diseased cell models showed similar transcript levels of Gal-1, as shown in our previous study. However, the mRNA levels were upregulated 2-fold with the Gal-1 treatment [31]. Even though the impact of adding exogenous Gal-1 in WT mice has not been defined, studies have shown that IP injection of Gal-1 to treat inflammation in other diseased tissues leads to upregulated endogenous Gal-1 over a period of 24 h [65]. Our results show that, along with negative NF-κB modulation, there is an increase in endogenous Gal-1 transcription due to our exogenous rHsGal-1 treatment compared to PBS treated Bla/J mice. These data suggest a novel selfupregulation loop where Gal-1 positively modulates its own expression. This positive feedback loop could be related to the role of Gal-1 as an alternative pre-mRNA splicing factor, other unknown nuclear interactions, or a yet-to-be-defined signaling pathway [24]. This phenomenon deserves further consideration and experiments to clarify the mechanism of how Gal-1 is able to self-upregulate.
The results obtained during our one-week dose-response experiment show that there is a 6-fold therapeutic range at which rHsGal-1 was effective. The effective concentration seems to peak at 2.7 mg/kg rHsGal-1 every seven days. Although the following doses did not provide benefit: 0.27 mg/kg D0, D7; 0.54 mg/kg D0, D7; and 13.5 mg/kg D0, D7, they also did not result in worsening membrane repair. This shows a safety profile of doses between 0.54 and 13.5 mg/kg in Bla/J mice. This 25-fold safe dosing range, along with pharmacokinetic studies, shows that the treatment of rHsGal-1 takes approximately 12 h to return to pre-dosing levels of Gal-1, indicating that this biologic may be a safe option for human patients.
Functional, histological, and biochemical experiments in our one-month study provide additional evidence of therapeutic benefits with rHsGal-1 treatment. Decreased inflammation may be a fundamental reason for observed increased muscle integrity in treated animals, as inflammation has been shown to play a large role in the pathological symptoms of LGMD2B [12,66]. A primary histological marker for LGMD2B is lipid deposition in affected muscles. This fatty infiltration has been linked to the presence of inflammatory markers [66]. Decreased perilipin with rHsGal-1 treatment demonstrates a decrease in fat deposition within Bla/J myofibers. Biologically, rHsGal-1 decreases inflammation through the NF-κB pathway and upregulates cytokines with anti-inflammatory and regenerative effects. We hypothesize that the decrease in fat deposition is due to decreased activation of the canonical NF-κB pathway as a result of treatment.
There is a direct relationship between mouse muscle health with cage exploration and rearing [41]. Both of these indices were significantly improved in mice treated with rHsGal-1. We suspect that this increase in movement is due to the ability of rHsGal-1 to reduce inflammation and promote muscle membrane repair. We believe the rHsGal-1 works mechanically to facilitate the formation of the membrane patch, an integral component in membrane repair. Together, our results from the one-week and one-month experiments demonstrate that rHsGal-1 affects the muscle membrane mechanically by Gal-1 localization to the membrane and biologically through inflammatory signaling [31]. More investigation is needed to show the proteins that rHsGal-1 interacts with as it mediates membrane repair. Regardless of the mechanism, the treatment with rHsGal-1 in Bla/J mice suggests an improvement in muscle health as evidenced by the CLAMS cages, the laser injury, biochemical, and histological assays.
Furthermore, rHsGal-1 showed the same therapeutic potential when administered to patient-derived dysferlin-deficient cells ( Figure 6). rHsGal-1 shows promise at diminishing the symptoms of LGMD2B in two areas of pathology: inflammation and muscle membrane repair. This two-pronged mechanism would be extremely useful as a therapeutic and may stem from the ability of rHsGal-1 to function as either a monomer or a dimer. The dimer form of rHsGal-1 is clearly more beneficial in assisting in the membrane repair process, while the monomeric version helps to reduce the markers of inflammation. Although more testing is required, these two parallel processes position rHsGal-1 as a highly effective therapeutic against LGMD2B.

Patents
The University of Nevada-Reno has been issued a patent in the U.S. (# US20130065242 A1) and Australia (# 45557BOA/VPB) for "Methods for diagnosing, prognosing and treating muscular dystrophy". PMVR is an inventor on these patents. Strykagen currently holds the license for this technology.
Brigham Young University has a patent for "Galectin-1 immunomodulation and myogenic improvements in muscle diseases and autoimmune disorders." (#U.S. Pat. No. 62/161,027. PCT/US2021/026232). PMVR and MLVZ are the inventors of this patent. This does not alter our adherence to MDPI-Cells policies on sharing data and materials.