*1.5. Lipid-Conjugated Oligonucleotides: Method of Delivery and Example of Conjugation* 1.5.1. Method of Enhanced Delivery and Lipid-Conjugated Structure

Finding the most optimal and efficient delivery method for therapeutic ON is still an ongoing campaign for the goal of achieving the maximal clinical outcome. Scientists usually implement one of the two following popular approaches: (1) external delivery capsules by utilizing nanoparticles and (2) covalent conjugation of endogenous biomolecules. Naturally occurring substances are preferable with some exceptions for artificial materials. Among these, hydrophobic or lipid moieties have gained much-wanted attention. It is abundant in biological systems and carries out essential functions such as executing signaling transportation. More importantly, it is a body of phospholipid bilayer that can help ON to mimic the hydrophobic properties.

The first exciting investigation of utilizing lipid nanoparticles as ON delivery was conducted by Felgner et al. He incorporated plasmid DNA with cationic lipid such as 1,2-O-octade-anyl-3-trimethylammonium propane (DOTMA) and dioleyl phosphatidyl ethanolamine (DOPE) to induce in vivo transfection into cells. This successful discovery leads to the use of LNP to be drug delivery carriers for small molecules as well. [28]. Significantly, there are eight LNP structures approved by the FDA. Patisiran is an example of ON carried by LNP approved in the market for the treatment of hereditary transthyretin amyloidosis. As by 2021, the most advanced LNP formulation was applied for delivery of two mRNA vaccines, BNT162b2/Comirnaty (Pfizer-BioNTech) and mRNA-1273 (Moderna) to counteract the global SARS-CoV-2 pandemic. For formulation of Comirnaty, ALC-0315 was the main component of this nanoformulation recipe. It was a synthetic lipid-like substance that has an ethanolamine headgroup along with two biodegradable branched ester tails. The LNP was formulated via mixture of ALC-0315/cholesterol/DSPC (Distearoylphosphatidylcholine)/PEG-lipid. In term of mRNA-1273, synthetic lipid SM-102 was selected as primary nanoparticles components. Its structure was similar to ALC-0315 containing an ethanolamine headgroup with difference of one mono and one branched degradable ester tails. mRNA-1273 was formulated strictly with SM-102/DSPC/cholesterol. Both synthetic lipids illustrated in Figure 5 were hypothesized to obtain cone-shaped structure from the branching tails, which boosted the strength of endosomal escape for mRNA molecules [29]. As advancing to clinical trials, Comirnaty was fully approved for individuals 16 years and older while mRNA-1273 is still in the third trial (approved for emergency use). They both encodes for stabilized full-length spike protein but their mRNA contents (100 µg and 30 µg for mRNA-1273 and BNT162b2, respectively). A review by Schoenmaker outlined the abridged and detailed information regarding of dosing and LNP components for both vaccines [30]. However, some ionizable lipids were feared to produce unwanted toxicity and in need of continuously monitoring due to the uncontrolled activation of cytokines after systemic administration [31].

**Figure 5.** Chemical structure of two most advanced LNP that assisted in delivery of 2 mRNA COVID vaccines. ALC-0315 and SM-102 contains similar structure with ethanolamine head (one is shorter than others). ALC-0315 has two branched degradable ester tail, while SM-102 only has one branched ester tail.

Alternatively, the second approach by lipid conjugation is plausible. Scientific evidence suggests a unity of lipid conjugated oligonucleotide (LON) can reduce the risk of immunogenesis while can maintain tolerance with a high dose in vivo [32]. Similar to LNP, LON's hydrophobicity is enhanced and be more accessible to membrane permeability and higher rate of internalization. In contrast, LONs are relatively smaller than LNP, which contributes to a higher leakage rate from endothelial lining and perfuse to various tissue types. However, the majority of lipid derivatives are highly accumulated in primary excretory organs, liver and kidney. Administration route is a crucial concept when mentioning LON delivery because it dictates the targeted tissue. In vivo experiments, either intravenous or subcutaneous injections can result in LON migration to clearance organ and some miscellaneous (spleen, skeletal muscle, etc.); while intrathecal or direct cranial injection can occupy parts of the brain [33,34]. Therefore, optimizing delivery location is theoretically a balancing act of hydrophobicity adjustment and understanding the chemical nature of bioconjugates.

To achieve such a feat, an effective synthesis of LON is required. A fully functional LON consists of three distinct fragments as illustrated in Figure 6: (1) the designed ON (ssRNA, siRNA, aptamer, or any forms), (2) attaching linker, and (3) lipid derivatives. Synthetic LON was produced via either pre-synthetic or post-synthetic approach, which lipoid conjugate would be introduced in a different fashion. An articulate description of both LON's synthetic routes was reviewed by Raouane and Li et al. [2,35]. We would like to briefly compare two approaches and highlight vital conjugating procedures for lipoid species. When tackling LON with pre-synthetic approach, it provides more flexible lipid point of attachment options. Conjugation can occur either at 30 , 50 or even between consecutive nucleotides. The most popular and convenient technique is attaching the hydrophilic group at the 50 end as presynthesized phosphoramidite. However, 30 -lipoid attachment can be arduous because bioconjugation is required to be pre-tethered onto a solid support. For instance, Setsinger premade cholesteroyl solid support via oxidative phosphoramidation [36], while Nikan et al. built a solid support with a pre-formed amide bond with docosahexaenoic acid [37]. Some other example [38] using an alcoholic moiety attaching to a solid support via succinyl linker while bearing another hydroxyl group for cholesterol to attach. Ueno [39] selected glycerol to bridge lipophilic group and mRNA. Interchain bioconjugation was attempted through examples of Guzaev and Durand [40,41]. In contrast, the post-synthetic approach requires two completely independent entities of ON and lipoid conjugation with their complementary reactive group for coupling. Some available techniques could be the formation of triazole linkers resulting from click chemistry reaction between dibenzocycloctyne (DBCO) and azido-lipid conjugates [42]. Raoulane demonstrated the effectiveness of thiol-malamide bridge of RNA and squalene [43].

**Figure 6.** General molecular structure of conjugated oligonucleotides including: (1) synthetic oligonucleotide, (2) linkers, and (3) bioconjugates.

## 1.5.2. Example of Conjugations

#### A. Cholesterol Conjugates

Covalently attached cholesteryl moiety as non-vehicle delivery for oligonucleotides was conceptualized as early as the late 1990s [44]. Manoharan's group claimed 30 - cholesterolconjugated ON produced the best silencing effect and continued in unfolding the delivery mechanism. They observed a 2-fold uptake increase in comparison to naked ON for silencing murine ICAM-1 and proposed liver uptake mechanism associated with scavenger receptors [45]. Later in the 2000s, scientists from Alnylam synthesized and examined a plethora of cholesteryl derivatives at two terminals of either sense or antisense strands [46,47], which confirmed better in vivo efficacy of 30 -cholesterol ON (30 -Chol ON). Wolfrum et al. expanded Manoharan's mechanistic notion and elucidated receptor-mediated endocytosis as a key for ON uptake. The cholesterol-conjugated ON was highly recognized and attached to either LDL (low-density lipoprotein) or HDL (high-density lipoprotein); thus, the resulting complex docked to scavenger receptors (SR-BI) and proceed to internalization [48]. At the same time, cholesterol conjugates enhanced the hydrophobicity of oligonucleotide, which ameliorated the drug-like properties. For researching a more feasible pharmacokinetic characterization technique, Godinho et al. attested 30 -Chol ON with rapid distribution phase (*t*1/2 <sup>α</sup> = 18–33 min) and slow elimination phase (*t*1/2 <sup>β</sup> = 8–14 h) [49].

A complete pharmacokinetic parameter was summarized in Table 2 for intravenous administration. Nevertheless, it was noticeable that silencing capability can only reach near 50% efficacy even at a low dosage. It was believed that conjugated ON was sequestered during late endosome, which coiled the term endosomal entrapment [50]. Hence, chemical enhancers were used to damage vesicles along the endosomal system to promote escape [51,52]. Additionally, the delivery scope of cholesterol conjugated oligonucleotide was very limited to hepatic cells and to some extend pancreatic cells [53].


**Table 2.** Pharmacokinetic parameter of Chol-hsiRNA after intravenous injection adapted by Godinho et al.

Most studies have demonstrated that cholesteryl conjugated ON was delivered effectively, and specifically to liver tissue. Hence, hepatic-related disorder would be its ideal target. For an instant, hypercholesterolemia is the excessive circulation of cholesterol in blood, which caused by either habitual diet or genetic condition. Antisense technology has provided a therapeutic platform through silencing of PCSK9 [54–56] or hepatic ApoB [57–60]; however, the unconjugated ASO treatments suffered mild to serious toxicity while giving questionable efficacy. Henceforth, studies from Wada and Nakajima demonstrated coupling ASO with cholesterol would enhance liver uptake while improving the degradation efficacy of PCSK9 (−50% silencing and 2 µmol/kg) [61] and ApoB (−85% silencing and 0.5 mg/kg) [62], respectively. Both research groups also pinpointed the cleavage of phosphodiester linkage as quintessential for liberating ASO, which showed a 3- to 5-fold increase in vivo potency [63]. Furthermore, the application of cholesterol conjugated ON is extended into the realm of cancer treatment. Liu et al. cholesterol-conjugated let-7a miRNA mimics could downregulate both transcriptional and translational levels of RAS in vitro, and minimize murine xenograft tumor in vivo [64]. Chernolovskaya group delved into silencing MDR1 (multidrug resistance protein), which overexpressed in oncogenic cells to efflux impending drug and amplify resistance [65,66]. 21-mer MDR1 targeting siRNA both as monomer and trimer (63-mer) were compared in vivo demonstrating monomeric siRNA obtained superior silencing efficacy while trimeric derivative accumulated highly in tumors [67,68]. Additional examples of other applicable disease would be including Huntington's disease [37,69,70], diabetes nephropathy [71], and herpes simplex virus-2 protection [72].

Some cholesterol conjugated ONs were successfully introduced to clinical trials. For example, ARC-520-HBV was the first RNA interference therapeutic for treatment against hepatitis B virus (HBV). ARC-520 injection consisted of a pair of synthetic cholesterolconjugated siRNAs to augment its delivery to hepatocytes. It also contained polymer-based excipients such as dynamic polyconjugates that enable endosomal escape [73]. Mechanistically, it reduced all RNA transcripts from covalently closed circular DNA (cccDNA) leading to the diminishing of both viral antigen and DNA. ARC-520 was at phase II clinical studies with the promising pharmacokinetic profile in a single-dose study in healthy volunteers. The clinicians found out that IV injection with a dose of 3 mg/kg can increase the curative effect and reduce the viral antigen level by 81–96%. The dosage regimen was given 2 mg/kg once every 4 weeks for 3 total doses. As results, the degree of viral decline and duration of the effect was consistent with the previous animal experiments. Unfortunately, the inclusion of hepatocyte-targeted excipient ARC-EX1 melittin-derived peptide linked to *N*-acetylgalactosamine caused detrimental toxicity in nonhuman primates which rendered the trial to be terminated [74]. In another application, ARO-AAT (SEQUOIA) is currently in phase II/III of clinical trial for the treatment of Alpha-1 antitrypsin deficiency (ATTD)-associated liver disease. The subcutaneous dose of iRNA selectively degraded ATT mRNA caused by Pi\*Z mutation. The trial was aimed to determine the safety, tolerability, and pharmacodynamic effect of the drug by gauging the level of plasma and intrahepatic Z-AAT levels. iRNA were given in incremental multiple doses and up to 300 mg in a single shot. It was well-tolerated and resulted in more than 91% serum reduction which was

sustained for 6 weeks. To understand dosage window in practical term, this iRNA therapy could be administered four time a year or less to maintain desired silencing effect. The clinical trial is anticipated to be completed by July 2022 (ARO-AAT2001; NCT03945292).

Apart from this, the cholesterol conjugated ON is broadly used for connective tissue growth factor (CTGF) to battle against fibrotic disorders. For instance, Hwang et al. reported a novel application of this modified 20 -OMe phosphorothioate nucleic acid for antifibrotic skin therapeutics. The drug is composed of a cell-penetrating asymmetric interfering RNA (cp-asiRNA) known by OLX10010 as shown in Table 3 (cholesterol conjugate). This iRNA targets the expression of CTGF, and it is currently examined in an ongoing phase 2a clinical trial [75]. When compared to unconjugated siRNA, 1 mmol/L of cp-asiCTGF achieved more than 85% silencing knockdown of CTGF at mRNA level without the assistance of transfection media. The calculated IC<sup>50</sup> was 0.315 nM in cell lines (the best efficiency was observed in keloid fibroblast cell). Hwang et al. also discussed in vivo studies on rat skin to demonstrate a significant gene-specific silencing capability with 1 mg intradermal injection of lipid modified siRNA after 72 h. Furthermore, the conjugated siRNA exhibited 10-fold lower in dosage efficacy as compared to the commercially available siRNA [73]. A recent study by Choe et al. suggested to co-administrate L-type calcium channel blockers to further facilitate cellular internalization. As result, silencing of cp-asiRNA was potentiated without significant adverse effect [76]. Likewise, RXI-109, a cholesterol conjugated siRNA discovered by RXi Pharmaceuticals', exhibited a reduction of CTGF during the course of wound healing followed by keloidectomy. This therapy was applicable for patients suffering from age-related macular degeneration with high risk of subretinal fibrosis (www. rxipharma.com/technology/rxi-109, accessed on 15 November 2021). Therefore, targeting CTGF with conjugated siRNA is a good direction for fibrotic disorders such as hypertrophic scars and keloids. Moreover, these ONs are anticipated to treat excess collagen from injury or after surgery which was conventionally treated with less effective silicon sheets with the application of pressure.

**Table 3.** A selective example of ON conjugated with lipoid moieties in corresponding with each bioconjugates section.


Notation: red—2'MOE modification, green—cET modification, underline—PS backbone modification, and **X**—bioconjugates.

#### B. Fatty Acid Conjugates

Like cholesterol, fatty acid is an attractive entity for bioconjugation since it offered hydrophobicity customization and mimicked the uncanny composition of the phospholipid membrane. Currently, unbranched fatty acids were heavily delved such as the study conducted by Prakash et al. An array of fatty acid tethered to 16-mer-ASO through phosphodiester–linked hexaylamino spacer was synthesized. Two structure-activity relationship (SAR) studies were conducted and examined two revolving concepts: carbon length and degree of unsaturation. The first SAR involved with eight different fatty acids' lengths (C10 to C22) conjugated to ASO revealed two findings: (1) protein binding property

with chain length shorter than 16-C was lower than their counterparts showed in Tables 3 and 4) Malat-1 expression was more significantly reduced by ASO with fatty.


**Table 4.** Adapted protein binding data from Prakash et al. displayed the trends depending on carbon lengths.

Fatty acid chain longer than 12 in quadriceps while all remained similar in the heart [77]. Furthermore, the second SAR delved with 12 different unsaturated fatty acids. Protein binding to albumin, LDL, and HDL were slightly improved while there was no significant effect attributed to the double bond position. The activity of the representative unsaturated ASO displayed significant improvement compared to unconjugated ASO; however, there was no remarkable difference from their saturated counterparts. An interesting observation was none of the unsaturated moiety could outmatch palmitoyl's silencing activity. Hence, palmitoyl conjugated ASO was selected as to be the most efficacious and subjected for elucidating muscle uptake mechanism in rodent model. Chappel et al. examined the consequential efficacy of palmitoyl-ASO after injection to endocytosis receptor knockdown mice (CAV1-/-, FcRN-/- and Alb-/-) [78]. In CAV knockdown mice, ED<sup>50</sup> of palmitoyl ASO in quadriceps decreased by four-fold compared to wild type (9.7 µmol/kg versus 2.4 µmol/kg). In FCRN -/- mice, attenuation of palmitoyl ASO's activity was observed compared to controlled groups with similar outcome (ED<sup>50</sup> of 5.5 µmol/kg to 16 µmol/kg). A contrast was observed in Alb -/- mice with unchanged activity in quadriceps (0.73 µmol/kg in Alb -/- and 0.71 µmol/kg in controlled BL6). Thus, muscle uptake of palmitoyl-ASO was facilitated by caveolin-receptor-mediated endocytosis into endothelium cells once bound to albumin. Simultaneously, silencing FcRN could weaken the recycling of albumin into circulation thus impairing the albumin binding of ASO. However, Alb -/- mice contradicted the hypothesis, which would question if other proteins would be upregulated in compensation of drastic albumin reduction, and some would exist sufficient affinity for palmitoyl ASO binding.

Khvorova group compared the pharmacokinetic distribution property of diverse lipid moiety with emphasis on four fatty acids: myristic (Myr), docosahexaenoic (DHA), docosanoic (DCA), and eicosapentaenoic (EPA) acid. Length and degree of unsaturation constituted the hydrophobicity, which resulted in various in vivo distribution outcome. This study concluded with two premises: (1) more hydrophobic conjugates offered higher retention and (2) hydrophobicity instituted tissue accumulations [53]. Furthermore, shorter and less hydrophobic fatty acid such as myristic was synthesized and PK was analyzed as mono-, di-, or trimer. As discussed, the impact of hydrophobicity was profoundly shown in different behavior: (i) mono-lipid conjugates was quickly released with high kidney accumulation, (ii) di-lipid conjugates functioned as in-between showing preferential liver accumulation while flexibly distributed to other tissue (lung, heart, and fat), and (iii) tri-lipid conjugates resides at the injection site with no significant systemic exposure [79].

Table 5 summarized the pharmacokinetic parameter of three myristic variants. DCAconjugated ON shared similar PK property as dimeric Myr and was able to silence the expression of myostatin (Mstn) in skeletal muscle after subcutaneous injection at 20 mg/kg dosage. Mstn, a growth factor expressed in skeletal muscles, negatively modulates muscle mass; hence, its

inhibition was a potential therapeutic treatment against muscle wasting [80–82] or Duchenne muscular dystrophy (DMD) [83–85]. Interestingly, the toxicity profile of fatty acid conjugates was safer compared to cholesterol conjugates showing low activation of cytokine at high dose (100 mg/kg) [86].

**Table 5.** Pharmacokinetic parameters of three myristic variants after 7 days period injection adapted from Biscan et al.


Fatty acid would serve as an ideal conjugate to deliver therapeutic ON to muscle tissue. Currently, two ASO-splicing modulated therapy are approved by the FDA for muscle-related index such as eteplersen for Duchenne muscular dystrophy (DMD) [87] and nusinersen for spinal muscular atrophy [88]. Although eteplersen received such speedy approval with promising application, overall clinical efficacy [87,89] and renal toxicity from high dose [90] remained controversial. Thus, fatty acids can aspire to be a delivery platform to ameliorate both therapies for patients in need. Aside from muscularrelated disorder, GRN163L (Imetelstat sodium) currently resides at phase III of clinical trials as a treatment for myelofibrosis as shown in Table 3 (fatty acid conjugate). It is a 13-mer phosphorothioate ASO with covalently attached palmitic acid at the 50 terminal that exuded telomerase inhibitory activity. Observation of telomerase shortening was detected across multiple cancerous cells derived from glioblastoma [91], multiple myeloma [92], Barrett's esophageal adenocarcinoma [93], breast [94], lung [95], and liver [96]. From in vivo delivery perspective, IC<sup>50</sup> values were seven-fold higher [97], and efficacy increased up to 56% compared to naked counterpart after 24 h followed by intravenous injection (50 mg/kg) [98]. In follow-up studies, a group of researchers managed to explore the effects of long-term GRN163L exposure on the maintenance of telomeres and lifespan of 10 pancreatic cancer cells. They summarized the study with IC<sup>50</sup> value was ranged from 50 nM to 200 nM, and suggested continuous exposure of GRN163L eventually led a complete loss of viability after several doubling times. Conversely, telomerase reactivation and elongation were observed in the absence of GRN163L. This observation reinforced that GRN163L could target the RNA template region of telomerase and proven to produced outstanding inhibitory effect. Overall, these outcomes demonstrated that the lifespan of pancreatic tumor cells can be shortened by continuous exposure and can be used in patients in the future [99]. Additionally, co-administration of GRN163L with trastuzumab revealed to produce synergistic effect, which GRN163L reversed the resistance of HER 2 + metastatic breast cancer against trastuzumab.

The clinical application of fatty acid conjugate is extended to ameliorate antibacterial resistance and antibiotic treatment as well. The attachment of ketal bis C15 and cyanine to 25-mer oligonucleotide at 50 or 30 terminal proved the efficient strategies in cell delivery. It decreased the minimum inhibitory concentration (MIC) of laboratory and clinical resistant strains to cephalosporin drug (i.e., ceftriaxone) by 25-fold than the naked equivalence. The decrease of beta-lactamase activity was dose-dependent and 5µM was found to be efficacious. Furthermore, 30 lipid modification was less efficient than 50 . The 3'-attachment could propel the destabilization of heteroduplex structure of mRNA-LON, which enhanced steric hinderance to prevent RNase cleavage rather than uptaking into the bacteria [100].

#### C. Vitamin E (*α*-tocopherol)

Vitamin E is a group of fat-soluble compounds consisting of either tocopherol or tocotrienols. Naturally occurring α-tocopherol is an essential dietary supplement so it would be a safe and interesting selection for chemical bioconjugation. Additionally, its structure is composed of hydroxyl chromane and a hydrophobic saturated side chain that potentially enhances ON membrane permeability. Nishina et al. synthesized a 17-mer gapmer targeting murine hepatic ApoB. Structurally, it consisted of a parent 13-mer gapmer flanked by two wings of LNA with additional 4-mer modified RNA as the second wing directly linked to α-tocopherol via a phosphodiester bond. In vivo efficacy examination, tocopherol 17-mer ON showed −70% ApoB mRNA silencing capability after murine injection at 0.75 mg/kg [101]. The mRNA silencing potency was heavily dependent on dosage level (drastic reduction of ApoB expression as dose increased to 1.5 and 3 mg/kg) and prolonged duration of exposure (maximum response occurs from day 3 to 14). A followed-up pharmacokinetic study using Alexa Fluor-647 tagged tocopherol 17-mer ON at 50 -end revealed more than 3.5-fold higher of accumulation in the liver compared to non-conjugated parent, while tocopherol 17-mer ON also possessed higher serum content (10,000 µg/L) than naked parent ON (>1000 µg/L) at 5 min after 5 mg/kg dose of injection. The pharmacokinetic parameter of tocopherol 17-mer ON was summarized in Table 6. Interestingly, western blot analysis suggested cleavage of full-length tocopherol 17-mer ON into naked 13-mer unit once arrived at the liver which hypothesized the second wing tagged tocopherol acting as a delivery enhancer and release the main frame of 13-mer to initiate RNase-H cleavage mechanism toward targeted ApoB mRNA.

**Table 6.** Pharmacokinetic parameter of Toc-17-mer ASO.


AUC—area under the serum concentration time curve, CLtot—total body clearance, MRT—mean residence rate constant, Kα—initial elimination rate constant, Kβ—terminal elimination rate constant, and Vdss—steady-state volume of distribution.

Another study conducted by Østergaard et al. comparing three different bioconjugates (cholesterol, tocopherol, and palmitate) ASO duplex targeting dystrophia myotonic protein kinase (DMPK), which caused myotic dystrophy (DM1) as the product of toxic repetition of nucleotide in the 30 - untranslated regions [102]. Structure of tocopherol-conjugated ON was illustrated in Table 3 (tocopherol conjugate). In vivo rat models, palmitate conjugated ASO responded with more improved silencing potency in skeletal muscle and heart compared to cholesterol and tocopherol after 10 mg/kg injection dose. However, in the monkey model, tocopherol conjugated-ASO came as more advantageous than the other two displaying a lower ED<sup>50</sup> value of 7 mg/kg across three different DMPK expressed tissues (heart, quadriceps, and tibialis) [103]. Additionally, tocopherol moiety displayed tolerable high dose while cholesterol struggled with toxicity issues (in mice and unable to advance for primal testing). In plasma pharmacokinetics, tocopherol conjugates tended to co-elute with HDL and LDL, which displayed from size exclusion chromatography suggesting the essential of plasma protein binding was essential for receptor-mediated endocytosis. Benizri et al. disclosed additional pharmacokinetic data showing elevated liver accumulation after 6-h injection at a dose of 3 mg/kg (9–14 µg/g for tocopherol-ON versus 2–5 µg for naked-ON) [98]. However, tocopherol is remained understudied and required further preclinical investigation; thus, limited cases of human studies are often acquired.

#### D. Squalene

Squalene is a naturally occurring triterpene molecule that is frequently harvested from shark's liver and some variety of vegetable oil. It is an important precursor for human cholesterol synthesis. As hydrophobic moiety squalene is also a candidate for ON conjugation that can couple either at 30 or 50 terminal of the sense strand. Thiol-maleimide or DBCO via click chemistry are usually generated, and squalene-ON can spontaneously form nanoparticles. Due to the amphiphilic nature of squalene, these nanoparticles could assemble in different shapes. Raouane et al. synthesized a 50 squalene attached mRNA duplex employing a thiol-maleimide linker. This spherical nanoparticle was characterized by a drastic increase in lipophilicity while maintaining exceptional stability in serum media as a negative suspension (zeta potential= −26 mV). Cytotoxicity MTT assay in BHP-10-3 cell lines demonstrated > 95% cell viability at 50 nM maximal concentration of squalene-ON nanoparticle, while qRT-PCR depicted −80% RET/PTC1 silencing capability in vitro. Mice implanted with tumor were intravenously administered with a dose of 2.5 mg/Kg in vivo also demonstrated approximately 80% silencing of RET/PTC1 through qRT-PCR. Tumor biopsy showed significant shrinkage compared to controlled naked mRNA duplex after 15 days of the injection [43]. In another oncogenic targeting study, Masaad et al. investigated the silencing outcome of 50 squalene attached ON against TMPRSS2-ERG fusion oncogene. This group employed Cu-free click chemistry to functionalize reactive DBCO group tethering to the spacer of siRNA duplex to azido squalene. The structure was shown in Table 3 (squalene conjugate). Nanoformulation of 50 end was characterized to be temperature sensitive and degradable at 37 ◦C, while its structure was constricted to be spherical and quite anionic (zeta = −37 mV). This formulation was subjected to in vitro inhibitory efficacy test with VCap cells. Wherein, 50 nM of 50 -squalene nanoparticle showed a similar silencing effect (−50%) as naked siRNA transfected by lipofectamine after 3 different time points. Additionally, xenografted mice with VCap tumors showed significant size growth inhibition by −60%. siRNA treated mice were sacrificed and collected with excretory organs to analyze biodistribution by detecting radioactive <sup>32</sup>P label. The majority of siRNA nanoparticles resided in either liver or kidney; however, it was interesting to see a significant accumulation directly at prostate tumors [42]. Hence, squalene conjugation was an exciting concept for ON's design. Nevertheless, squalene harvesting can be controversial due to the revolving of endangering the shark population.

#### **2. Conclusions and Future Outcome**

Hydrophobic modifications such as cholesterol, fatty acid, α-tocopherol, and squalene still have room to mature compared to medicinal nanoformulation such as micelle, lipoplex, or, even, LNP. Of course, the primary goal of bioconjugation is elevating hydrophobic profile of ON-based therapy but a deeper quantitative understanding of structure related to delivery efficacy is still underappreciated. As mentioned in Biscan et al., the hydrophobicity profile of three distinctive lipid conjugates (dimeric Myr, cholesterol, and tocopherol succinate) appear to be similar as quantified via HPLC (measured in retention time); however, the biodistribution pathways are concluded to be diverse. Countless observation of multiple lipoid conjugates is accumulated at large in the secretory organ (liver) but detection at other tissues includes the spleen, kidney, and, even, skeletal muscle tissue will pave the way to develop novel delivery techniques to extrahepatic tissues [53]. Some fatty acids, such as docosanoic acid, had the ability to penetrate to skeletal muscle and, even, in the brain, which required direct spinal injection (unfavorable for human application) [37]. Aimed with current understanding, additional structural explorations would lead to better optimization for highly stable and selective modified ON; thus, current drawbacks in pharmacokinetic and biodistribution can be properly addressed. Moreover, there were still potential and unexplored lipophilic moiety both naturally occurred and artificial that can be examined to potentiate the delivery of ON-based therapy. More so, the profound pharmacokinetic, efficacy, and toxicity data from the previous conjugation can be utilized to develop a learning-based artificial intelligence

to predict of other lipid species or even fabricate novel artificial structures in the quest of advancing ON-based therapy in the new height.

In the future, the hope of better delivery of ON therapy can reduce the need of a large dose to patients which can significantly cut down the cost of treatment. Currently, the patient-affordable cost for ON therapy is astronomical for individuals in need. Eteplirsen, the current treatment for DMD, was marketed in 2017 with the price of \$300,000/patient a year; while nusinersen charges patients up to \$750,000 for the first year following with \$350,000 for consecutive years [104]. Such skyscraper cost of therapy can associate to denial of coverage from insurance companies. Even with approval, the insurance coverage may increase annually, which will devastate other members within the same insurance network. Therefore, the work of uncovering the most optimized delivery is not only limited to certain method but it is a combination effort of both bioconjugation and nanoparticle formulation.

**Author Contributions:** Conceptualization, H.-y.L., P.T. and T.W.; writing—original draft preparation, P.T. and T.W.; writing—review and editing, P.T., Z.L. and H.-y.L.; visualization, P.T. and T.W.; supervision, H.-y.L.; project administration, H.-y.L.; funding acquisition, H.-y.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by Helen Adams & Arkansas Research Alliance Endowed Chair Fund.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest. This manuscript reflects the views of the authors and does not necessarily reflect those of the Food and Drug Administration. Any mention of commercial products is for clarification only and is not intended as approval, endorsement, or recommendation.

#### **References**

