A computational overview of integrase strand transfer inhibitors (INSTIs) against emerging and evolving drug-resistant HIV-1 integrase mutants

AIDS (Acquired immunodeficiency syndrome) is one of the chronic and potentially life-threatening epidemics across the world. Hitherto, the non-existence of definitive drugs that could completely cure the Human immunodeficiency virus (HIV) implies an urgent necessity for the discovery of novel anti-HIV agents. Since integration is the most crucial stage in retroviral replication, hindering it can inhibit overall viral transmission. The 5 FDA-approved integrase inhibitors were computationally investigated, especially owing to the rising multiple mutations against their susceptibility. This comparative study will open new possibilities to guide the rational design of novel lead compounds for antiretroviral therapies (ARTs), more specifically the structure-based design of novel Integrase strand transfer inhibitors (INSTIs) that may possess a better resistance profile than present drugs. Further, we have discussed potent anti-HIV natural compounds and their interactions as an alternative approach, recommending the urgent need to tap into the rich vein of indigenous knowledge for reverse pharmacology. Moreover, herein, we discuss existing evidence that might change in the near future. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00203-023-03461-8.


Introduction
HIV (human immunodeficiency virus) infection remains a global public health concern and is predominantly serious in low-and middle-income countries (Notario-Pérez et al. 2017). As stated in the 2021 report of the United Nations Programme on HIV/AIDS (UNAIDS), about 37.7 million people were currently living with HIV/AIDS worldwide in 2021, among which 28.2 million people were accessing antiretroviral therapy (ART). An estimated 1.5 million people were newly infected, the majority being above the age of 15. Despite the scientific advances in this field, nearly 680,000 people have died due to AIDS-associated diseases, with a large population being from Eastern and Southern Africa (Siwe-Noundou et al. 2019;UNAIDS 2021).
Due to the inability to completely cure HIV, averting its spread is of paramount priority. To achieve the goal to end AIDS by 2030, UNAIDS has formulated the 90-90-90 target (Duwal et al. 2018;Mikasi et al. 2021). According to the National AIDS Control Organization 2020 Annual Report, the majority of people living with HIV in India projected Mizoram with the most cases of adult HIV prevalence (2.37%), followed by Nagaland and Meghalaya. Although in the past decade there has been a 33.3% decline, the same organization reported the national adult HIV prevalence to be about 0.22% with 0.23% being males and 0.20% being females, accounting for about 23.19 lakh people living with HIV (majority originating from Maharashtra). By March 2020 about 15 lakh people were enrolled for the ART programme (NACO 2019). Due to this, the Indian government has pledged 22 million USD to the Global Fund's Sixth Replenishment for 2020-2022 to support the global response to HIV (Avert 2020). There is no published evidence as to what proportion of ART-eligible people living with HIV are lost before starting the therapy or the dropout rates of people exiting the programme.
HIV is an enveloped virus grouped in the genus Lentivirus within the family of Retroviridae and subfamily Orthoretrovirinae. It is 50 to 60 nm in radius and has a lipid envelope that is embedded with glycoproteins, beneath but the infected macrophages and microglia release lethal constituents, like inflammatory cytokines and chemokines along with viral proteins, which damage the neurons and cause HIV-1-associated neurocognitive disorders (Gong et al. 2020). The innate quiescent nature of the virus and its latency makes it difficult to identify compounds that could be employed as anti-HIV agents (Dahabieh et al. 2015).

Highly active antiretroviral therapy (HAART)
A personalized cocktail of diverse classes of drugs commonly referred to as the HAART exists to manage the disease. At present, about 26 antiretroviral drugs are FDA approved, which do not eradicate HIV infection but suppress its activity (Crisan and Bora 2021), as depicted in Fig. 2. Each HAART drug targets and suppresses different stages of HIV's replication: (i) viral entry (co-receptor CCR5/ CXCR4 antagonists; (ii) fusion of virion (fusion inhibitors); (iii) viral complementary DNA (vcDNA) synthesis (reverse transcriptase inhibitors); (iv) vcDNA-host DNA integration (INSTIs); and (v) virion release and maturation (protease inhibitors) (Notario-Pérez et al. 2017;Siwe-Noundou et al. 2019). HAART effectively reduces the viral load (viraemia) and assists the immune functions to recover in HIV/AIDS patients, prolonging death by more than 7 to 10 years or even longer when compared to single drug-treated patients (Park et al. 2021).
The World Health Organization and the National AIDS Control Organization (NACO) initially recommend a combination of ART for HIV patients, which includes reverse transcriptase inhibitors-2 nucleoside NRTIs and a non-nucleoside NNRTIs or INSTIs or a ritonavir-assisted protease (PR) inhibitor (Akanbi et al. 2012;NACO 2013;Huang et al. 2019;Raj et al. 2021; National Institutes of Health 2022). Despite the substantial success (decreased Fig. 1 Schematic representing the structure of HIV virion. This has been simplified for clarity. For instance, the gp120 and gp41 envelope glycoproteins are trimers in reality and not monomers as depicted here (Blut and Blood 2016;Rossi et al. 2021). This figure has been modified from Servier Medical Art (http:// smart. servi er. com/), licenced under a Creative Commons Attribution 3.0 Unported License mortality and morbidity of virus) achieved with the aid of HAART, the continued emergence of cross-resistant viral strains and drug-related adverse effects on patients persist significant challenges to a sustained battle against HIV. Hidden virus in cellular reservoirs is the foremost barrier to functional therapy. This implies a pressing need for the identification of novel anti-HIV agents with enhanced pharmacokinetic profiles, potency, and least side effects and cross-resistance, to defeat HIV on a global scale (Wallet et al. 2019).
These strict requirements are addressed by employing several computational techniques, like molecular docking (Tewtrakul et al. 2015;Ercan et al. 2019;Arslan 2019;Dogan and Durdagi 2021;Esmaeili et al. 2021), pharmacophore modelling (Martin 2014;Bhatt et al. 2014;Xue et al. 2014;Islam and Pillay 2016), molecular dynamics (MD) simulation (Chen et al. 2014;Islam and Pillay 2016;Chitongo et al. 2019;Samorlu et al. 2019), quantitative structure-activity relationships (QSAR) (Gupta et al. 2013;Reddy et al. 2013;Zhou et al. 2021;Xuan et al. 2021), and many more. The human cells lack structural or functional integrase homologues, making IN a desirable drug target for HAART (Siwe-Noundou et al. 2019); further IN depicts greater tolerability, high efficacy, and fewer drug-drug interactions as compared to other HAART classes (Brooks et al. 2019). Ross (2015) conducted a QikProp (developed by Schrödinger) analysis for the FDA-approved INSTIs to detect the compounds with unfavourable drug pharmacology and concluded that all of the IN inhibitors had a star score of 0, except for cabotegravir which had 1 depicting that one of its chemical properties (human serum albumin binding prediction) did not lie within the range of drug pharmacology. The more the star score, the more the compound is less drug-like owing to its predicted properties and outlying descriptors. Moreover, its potency is lower than the other two second-generation drugs and the development of Q148R/K mutation with its regimen can lead to a high-level cross-resistance to all INSTIs Oliveira et al. 2018). Therefore, in this study, we have focused particularly on integrase inhibitors (raltegravir, elvitegravir, dolutegravir, and bictegravir) owing to their gaining popularity and aforementioned reasons. C-C chemokine receptor 5, CXCR4 C-X-C chemokine receptor type 4 (Fusin), NRTIs Nucleoside reverse transcriptase inhibitors, NNRTIs Non-nucleoside reverse transcriptase inhibitors, INSTIs Integrase strand transfer inhibitors (Li and De Clercq 2016;Rossi et al. 2021)

Integrase inhibitors
IN is a crucial nucleotidyltransferase enzyme in the HIV's life cycle that decides the virion's fate, as its establishment leads to persistent HIV-1 infection (Brooks et al. 2019;Mbhele et al. 2021). It is present at the 3'-end of the pol gene and is produced during the cleavage of gag-pol polyprotein during early virus maturation (Mbhele et al. 2021). This enzyme catalyses the insertion of virally reverse-transcribed cDNA into the patient's genome via dual stages (S N 2 nucleophilic reaction) (Siwe-Noundou et al. 2019;Zhou et al. 2021). The first step is '3'-processing' (3'P), wherein a dinucleotide (GT) pair is cleaved from both the cDNA's 3'-ends, forming the pre-integration complex (PIC). This complex then integrates with the host's cellular genome with the aid of lens epithelial-derived growth factor (LEDGF or p75) in the second step called strand transfer (ST), infecting the individual (Varadarajan et al. 2013;Park et al. 2021). Both these reactions are energetically independent as well as set apart, both spatially and temporally as the 3'P occurs in the cytoplasm of the infected cells, whereas the ST activity takes place in the nuclei . Once integrated, it is now referred to as a provirus, which acts as a post-integrative template for viral infection, allowing its self-transcription to release more infectious copies of itself (Thierry et al. 2017).
HIV-1 IN has a molecular weight of 32 kDa and comprises 288 amino residues, which can be distributed into three particular functional domains: (i) Residues 1-49: N-terminal domain (NTD) The zinc finger motif HHCC (histidines 12 and 16, cysteines 40 and 43) present in NTD is responsible for the stability of the whole enzyme, favouring the multimerization of the protein.
(ii) Residues 50-212: Catalytic core domain (CCD) High enzymatic activity is maintained by proper chelation of the CCD catalytic triad DDE motif (i.e., Asp64-Asp116-Glu152) of the active site and the 2 divalent cofactor ions (Mn 2+ or Mg 2+ ). Mutation of either of the DDE residues inactivates the enzyme. In the apoprotein, one Mg 2+ is coordinated between D64 and D116. Once the vcDNA joins, the other cation is coordinated between D64 and E152. This domain also encompasses a flexible 10 residues loop (140-149) that is adjacent to the catalytic site, involved in vcDNA substrate recognition. (iii) Residues 213-288: C-terminal domain (CTD) This region serves as a non-specific binding site for the vcDNA and further contributes to stabilizing the IN-vcDNA complex Ross 2015;Thierry et al. 2017;Nusrath Unissa et al. 2017;Zhou et al. 2021).
IN inhibitors (INSTIs) are a potentially attractive class of next-generation antiretroviral (ARV) drugs that have gained an audience in the past 15 years (Zhou et al. 2021). These comprise a polycyclic core with heteroatom triads and a halogenated benzyl group that interacts with vcDNA attached to the CCD by a flexible linker (Smith et al. 2021a). These disrupt the strand transfer step during viral infection by binding to the CCD of IN by chelating the catalytically essential divalent metal ions and blocking the enzyme's active site, thereby displacing the reactive 3'-hydroxyl group of the terminal A17 away from the active site, obstructing the virion from hijacking the CD4 + cells (Chitongo et al. 2019;Park et al. 2021). The first INSTI to get approved was raltegravir (RAL) in 2007 as a therapeutic INSTI by the US Food and Drug Administration (FDA), followed by elvitegravir (EVG) in 2012, dolutegravir (DTG) in 2013 (Thierry et al. 2017), bictegravir (BIC) in 2018 and injectable formulation of cabotegravir (CAB) in 2021 (Phillips et al. 2021); their chemical structures have been depicted in Fig. 3. These are acclaimed for having a higher genetic barrier to resistance compared to other PR and RT inhibitors, especially DTG and BIC (Mbhele et al. 2021).
Although the mentioned INSTIs share various properties, including mechanism of action, antiviral potency, good tolerability, and safety profile, they have other significant traits that differ from each other, like pharmacokinetics, interactions, resistance profile, and pharmacodynamics (Blanco et al. 2015;Roulet et al. 2018), refer to Table 1. However, the advent of INSTI-resistant strains is an impending task (Xuan et al. 2021). Although numerous structurally diverse INSTIs have been reported, like coumarin derivatives, quinolone acids, diketo acids, salicylhydrazide derivatives, aryl cyclic compounds, sulphur nitrogen thiazepines, diazonaphthalene derivatives, pyrimidine ketones, benzenesulfonamides, and so on, the need to develop novel molecules still persists (Bhatt et al. 2014;Zhou et al. 2021).
Some of the IN inhibitor-based HAARTs have revealed antiviral potency against Middle East respiratory syndrome (MERS), severe acute respiratory syndrome (SARS), and viruses associated with COVID-19 (Rameshrad et al. 2020;Makoti and Fielding 2020). However, no clinical improvements beyond standard care were noticed, when administered to severe COVID-19 patients (Crisan and Bora 2021).

Computational studies
Computational approaches have created substantial interest by decreasing the time essential for preclinical assessments, clinical trials, as well as in terms of expenses and resources. These aid in predicting the significance of mutations involved in making the strain resistant, elucidating the molecular mechanisms of resistance associated to IN-ligand complex, and lastly, disclosing precise insights vital to designing new potent INSTIs, refer to Fig. 4 for major INSTI resistance-associated mutations (RAMs) (Islam and Pillay 2016;Han et al. 2016;Zhou et al. 2021;Crisan and Bora 2021). Consequently, the approved anti-HIV agents can be employed either as reference molecules to guide the design of novel ones or can be scrutinized for their re-application to other viruses or non-virus diseases using computational as well as experimental methods (like in the case of COVID-19) (Rameshrad et al. 2020;Indu et al. 2020;Alavian et al. 2021;Crisan and Bora 2021).

Raltegravir (isentress, isentress HD)
The era of the therapeutic INSTIs for HIV treatment began with the approval of RAL in 2007 (Brooks et al. 2019); it predominantly suppresses the ST activity while inhibiting the 3'P activity only up to a certain limit (100-fold less specific than ST). The study of RAL's binding to the HIV-1 IN and its conformational inclinations represent an essential task for identifying the molecular factors that contribute to its pharmacological action. This pyrimidinone analogue acts as a multipotent drug owing to the presence of 2 polydentate pharmacophores (1,3,4-oxadiazole-2-carboxamide and carbonylamino-1-N-alkyl-5-hydroxypyrimidinone) which are capable of hitting more than one target in HIV-1, the vcDNA, the unbound IN, or IN-vcDNA complex (Arora and Tchertanov 2013). The inhibition process initiates by recognition of RAL by the processed vcDNA bound to a transient intermediate IN, providing a potentially favourable route to the design of INSTIs with enhanced affinity and selectivity . Since no crystal structures of full-length HIV-1 IN are available, Tn5 transposase and prototype foamy virus (PFV) IN are considered surrogate models or scaffolds for studying the mechanism of action of INSTIs (Arora 2013;DeAnda et al. 2013) (Fig. 5).
The most impactful INSTIs inhibit both wild-type (WT) IN as well as a variety of familiar mutants. However, because there are mutations that reduce the potency of all available INSTIs, novel and enhanced compounds are   (2019) and Kolakowska et al. (2019) needed (Nusrath Unissa et al. 2017;Smith et al. 2021a). With recurrent use, resistance patterns are likely to impact future INSTI-based regimen selection specifically concerning the suitability for use of long-acting injectables. The combinations of these mutations along with accessory mutations can lead to a greater reduction in drug susceptibility representing a threat to the ARV drug's long-term efficacy (Brooks et al. 2019;Cook et al. 2020;Smith et al. 2021a). RAMs impend the enduring triumph of combination ART (cART) consequences for HIV-1 treatment (Chitongo et al. 2019). Of the available INSTIs, the first-generation RAL and EVG share an overlapping resistance profile (N155H, Q148K/R/H, and Y143R/C/H) and a low genetic barrier, resulting in a greater propensity for loss of virologic activity secondary to resistance. Due to this, patients failing one of the either containing regimens should not be transitioned to another (Nusrath Unissa et al. 2017;Brooks et al. 2019;Chitongo et al. 2019;Smith et al. 2021b, a). On the other hand, DTG and BIC have a comparatively lower possibility of developing resistance due to higher genetic barriers owing to their extended tricyclic scaffolds (Wainberg and Han 2015b;Smith et al. 2021a). RAL further depicts E/Z configurational isomerism and high structural flexibility due to the presence of non-amide bonds joining the aromatic ring of the 8 aliphatic single bonds, characterizing the relative locus of the vicinal 1-4 and 1-5 oxygen atoms. This also determines the chelating properties of deprotonated or protonated drugs (Arora and Tchertanov 2013;. These oxygen atoms of carbonylamino-1-N-alkyl-5-hydroxypyrimidinone and β-ketoenolate can chelate the pair of cations simultaneously by competing with the metal-binding sites (Arora 2013). Subsequently, when RAL is docked to the unbound IN, it adopts different configurations within a large binding cavity formed by the catalytic active site and the adjacent loop , further contributing to the recognition and binding of different substituents in topologically distinctive regions of IN by relating its molecular and conformational characteristics. It acts as a divalent metal cation chelating ligand and can bind to them by both pharmacophores in varying isomerisation states (Arora and Tchertanov 2013). All observed RAL poses show the absence of specific interactions with the targets lacking vcDNA intermediate . Moreover, the structural flexibility of RAL allows itself to fit into the relatively large binding pocket of IN-vcDNA PIC with low docking scores, thus yielding various RAL docked conformations/poses and indicating it to be a biologically relevant target of the INSTI. This also imposes that the DNA substrate is required to bind for the precise configuration of residues that are required for catalysis. (Arora and Tchertanov 2013; The MD simulation studies have verified the significance of the catalytic site loop flexibility for catalysis. A major conformational change, concerning the E152 residue of the DDE motif, was observed upon vcDNA binding. This occurs due to a change in the position of the Mg 2+ cation in the active site that induces a switch of E152 side-chain conformation towards the active site, without any changes in the configuration of the overall catalytic loop. This is crucial to decrease the flexibility induced, for instance, the G140A/ G149A mutations result in lower levels of activity despite minimal effects on vcDNA binding. MD simulations of the influence of RAL-selected mutations on IN structure and, chiefly, on the structure of the catalytic loop, indicated strict conservation of IN CCD structure, including the Ω-shaped hairpin in the active site loop (Arora 2013).
RAL is preferred over NRTIs and is employed clinically in first-line and second-line ART regimens because of its efficacy and better tolerance (Huang et al. 2019). Like other antiretroviral inhibitors, even RAL develops/induces resistance, which is linked with amino acid substitutions, specifically either N155H or Q148R/H/K or Y143R (in rare cases) primary mutations (Quashie et al. 2012;Arora and Tchertanov 2013;Smith et al. 2021b). RAL forms links with the catalytic triad and interacts with the five residues, T66, E92, Y143, Q148, and N155, involved in primary resistance (Arora 2013). RAL's oxadiazole ring interacts via van der Waals contact with Y143 and P145 and makes a π-stacking connection with them (Ross 2015). G140S, a secondary mutation, further increases the RAL resistance by rescuing the replication defects from the Q148R/H/K mutation. This chiefly occurs by disturbing the position of the active site Mg 2+ ions and not by affecting the affinity of the metals (Arora and Tchertanov 2013). RAL is independent of HIV-1 co-receptor tropism (CCR5 and CXCR4) owing to its mechanism of action and is also active against other classes of ARV (Arora 2013).
A study conducted by Nusrath Unissa et al. (2017), showed RAL forming 4 conventional H-bonds with Asp116 (2 bonds), Ile151 and Glu152 residues. Alike EVG, the halogen atom of RAL was also involved in interacting with Pro142 residue. The primary mutations Q148H/R/K decrease IN susceptibility to RAL by about 5 to 50 folds, whereas the double mutants E138K/A with Q148H or Q148H/R/K and G140S/A or Y143C/R with T97A (naturally occurring due to immune escape) reduces susceptibility by more than 100-fold (Ross 2015;Acosta et al. 2019). The role of these primary mutation pathways has been elaborated by Anstett et al. (2017), in dominating RAL selections in vivo and in vitro.
MK-2048 is a second-generation INSTI which is formulated as an intravaginal ring as pre-exposure prophylaxis for HIV prevention (Hoesley et al. 2019;Brooks et al. 2019), that has a duration of action fourfold longer than RAL and a dissociation half-life of 32 h on WT-IN (Arora 2013). It can specifically select a G118R resistance mutation, as demonstrated in its selection studies (Quashie et al. 2012). This ring is clinically proven safe and has depicted good tolerance along with detectable drug levels in the cervical tissue, cervicovaginal fluid, and plasma when continuously used for more than 28 days (Brooks et al. 2019). Another novel drug that advanced into phase 1 clinical trials is a noncatalytic site IN inhibitor, BI-224436 (BI2), which inhibits the virion replication via binding to a conserved allosteric pocket of IN, disrupting the interactions between IN and LEDGF protein, located in the CCD dimer interphase at residues A128, E170, A124, T125, and H171. Despite BI2's promising results, it was put on hold (Ross 2015).

Elvitegravir (Vitekta, Stribild, Genvoya)
EVG is the next-generation INSTI therapy to garner FDA approval in 2012 for clinical use as monotherapy or in combination with other anti-HIV drugs (Brooks et al. 2019). It is a cobicistat-boosted quinolone-3-carboxylic INSTI that is administered in a quad pill formulation with the RT inhibitors as Stribild (Ross 2015). EVG inhibits the process of vcDNA strand transfer with a promising 50% maximal inhibitory concentration (IC 50 ) of less than 0.06 μmol L −1 in vitro, thereby averting virion from being integrated into the patient's chromosome and obstructing provirus formation and infection proliferation (Abdelhameed 2015). The emergence of certain mutational patterns in IN affects inhibitor binding, thereby modifying the degree of IN mutant susceptibility to EVG, and limiting their therapeutic efficiency (Masso et al. 2014). Due to multiple primary mutations at E92Q, T97A, R263K, S147G, E138K/A/T, T66I/A/K, G140A/S/C, Q148K/H/R, G118R, and N155H, it leads to resistance against EVG and hence virologic failure both in vitro and in vivo (Ross 2015;Unger et al. 2016;Stanford HIV-1 Drug Resistance Database 2019;Smith et al. 2021b). EVG substantially inhibits the RAL-resistant mutant Y143R owing to the absence of chemical moiety that stacks with the phenol group of Y143 (Smith et al. 2021b).
A study conducted by Khoury et al. (2019) showed that EVG binding to intasome has a slight preference over DTG and RAL, indicating that additional interactions of the diketo acid groups of the drugs with DNA could be essential to further enable preferential binding of DTG. EVG affects not only the residues in direct contact but also the outer-shell residues indirectly bound, particularly the double mutations, like Q148R/H and G140S/A, which has a synergetic effect on its efficacy. Its halobenzyl group targets the highly conserved 5′CpA 3′/5′TpG 3′ on the vcDNA ends with its centralized pharmacophore coordinating the pair of catalytic divalent cations. The X-ray structural analysis demonstrated that in the IN-vcDNA complex, the halobenzyl ring stacks over C16 (upstream), while the C-X bond points towards the centre of the G4 (downstream) of the second strand. Therefore, a surge in the drug-vcDNA complex stability draws a parallel with an increase in drug activity and a decline in viral resistance, indicating the vitality of vcDNA end recognition for the binding affinity of INSTIs to the intasome (Khoury et al. 2019).
Another study by Kang Birken et al. (2019) discussed 2 cases involving HIV rebound due to a possible drug-drug interaction between EVG, cobicistat, emtricitabine, tenofovir alafenamide, and calcium-containing products, which were the first reports of the rapid emergence of mutation. Further, a study involving ligand-EVG complexed with WT-IN protein demonstrated a conventional H-bond between EVG and the residue Asp116 along with two carbon H-bond formations with the same residue and an unfavourable negative-negative interaction between EVG and Asp60. Fascinatingly, two halogens (F and Cl) bonds were formed between EVG and with residues Pro142 and Asn144 of WT, respectively (Nusrath Unissa et al. 2017). Cross-resistance with RAL is most prominent in over 90% of cases as the RAL-resistant virus also retains resistance to EVG; however, despite viral mutations conferring resistance to RAL and EVG, 64% of the mentioned cases were susceptible to second-generation DTG (Unger et al. 2016).

Dolutegravir (Tivicay, Triumeq)
DTG was the earliest second-generation and first unboosted INSTI to get approval in 2013, owing to its favourable tolerability profile against RAL and EVG. This prominently distinctive resistance profile with a higher resistance barrier has decreased the pill burden and has further maintained activity against several INSTI resistance mutations as witnessed with first-line INSTIbased regimens. This high barrier is usually attributed to its long binding half-life to IN (Wainberg and Han 2015b;Malet et al. 2017;Brooks et al. 2019).
Metal ion chelation is a unifying feature mutual to all approved INSTIs. Structural and computational analyses reveal that the INSTI-resistant double mutation G140S/ Q148H can affect all INSTI binding by disturbing the secondary coordination shells of the divalent ions in the IN active site (Nusrath Unissa et al. 2017;Smith et al. 2021a). However, variations to the INSTI scaffold can recompense for this loss of binding affinity, like in the case of DTG (presence of oxazine ring) and BIC (presence of oxazepine ring), which allow these compounds to have stabilizing engagements with backbone atoms of G118 and N117 in the β4-α2 loop of the IN, contributing to the enhanced activity of the second-generation INSTIs against the mutants (Cook et al. 2020). DTG chelates the Mg2+ ions in the active site, and the halo benzyl group forms a hydrophobic π-π stack with the base of the penultimate cytosine, interacting with the terminal adenine of the vcDNA (see Fig. 6), and the longer the linker, the better it allows a bound DTG to flexibly adjust to the alterations in the active site of IN mutants (Smith et al. 2021b).
Multiple reports on binding modes of all INSTIs to a surrogate for the HIV-1 intasome, the PFV intasomes, have revealed that an extended linker of DTG allows its halogen groups to enter beyond the pocket within the IN active site than other INSTIs. Further, DTG can regulate its structure and conformation in retort to structural changes within the active sites of ARV drug-resistant INs (Wainberg and Han 2015b, a). Using residue interaction network (RIN) analysis and MD simulation, Xue et al. (2013) disclosed that P145 in the 140S loop (Gly140-Gly149) of the intasome depicts strong hydrophobic interactions with INSTIs and is involved in a conformational rearrangement at the active site of the HIV-1 intasome which leads to a decrease in the chelating ability in INSTIs to Mg 2+ in the active site of the resistant IN further causing cross-resistance.
As per the Stanford HIV-1 Drug Resistance Database-HIVDB (https:// hivdb. stanf ord. edu/), residues T66K, E92Q, G118R, E138K/A/T, G140A/S/C, Q148H/R/K, N155H, and R263K are liable for IN resistance against DTG (Ross 2015; Stanford HIV-1 Drug Resistance Database 2019). Several non-polymorphic mutations (S153, R263, and G118) confer low-level in vitro DTG phenotype changes (Malet et al. 2017). A study conducted by Nusrath Unissa et al. (2017) demonstrated that DTG simultaneously formed two H-bonds between Gln62 and Asp116 residue of WT-IN and 3 conventional H-bonds amid DTG and Asp116, Gly149, and Glu152 residues. Remarkably, a halogen bond between the fluorine atom of the DTG's pharmacophore and residue Asp116 was also formed. In addition, π-anion bonds were formed between DTG and residues Asp116 and Glu152, respectively, highlighting the role of residue Asp116 in multiple interactions. Further, alkyl bonds were also formed between DTG and residues Ile141 and Ile151, respectively. In the WT complex, the DTG formed a pair of stacking interactions, the first amid its 2, 4-difluorobenzyl ring and cytosine C16 of vcDNA and the next amid the metalchelating pharmacophore and the viral 3′-deoxyadenosine (Malet et al. 2018).
This monocyclic carbamoyl pyridine derivative (Smith et al. 2021b) extends to the full width of the active site but intasome with BIC as ligand. The coiling orange structure represents the viral complementary DNA (vcDNA) fragment, wherein the ligand can be seen interacting via the π-stacking bonding with the terminal 3′-deoxyadenosine 17 (DA17) of the vcDNA. Image created with the help of PyMol (Schrödinger and DeLano 2020) only spans half of its height (DeAnda et al. 2013;Ross 2015). Few mutations among other accessory mutations involved in IN resistance to DTG both in vitro and in vivo include E138K, F121, G118R, S153Y, and R263K (Wainberg and Han 2015b;Seatla et al. 2021). The most typical mutation that arises in subtype B and recombinant A/G viruses when DTG is employed in vitro is R263K. These mutations not only confer low-level cross-resistance to DTG but either alone or in combination with other secondary mutations (like M50I, Q95K, T97A, V206I, G149A, H51Y/H, D67N, E157Q, A128T, and D232N) also diminish the viral fitness, and are rarely testified in patients (Wainberg and Han 2015b;Nusrath Unissa et al. 2017;Brooks et al. 2019;Seatla et al. 2021;Mikasi et al. 2021). Molecular modelling conducted by Quashie et al. (2012) proposes that R263K can initiate catalytic and structural changes within IN via partially impairing the IN-vcDNA binding, elucidating its selection by DTG.
However, treatment-emergent resistance mutations are of concern in patients, predominantly those with prior exposure to RAL or EVG (Cook et al. 2020). The second-generation INSTIs are effective against multiple RAMs that confer resistance to RAL and EVG; further, DTG considerably retains potency against the RAL-resistant mutants N155H, Y143R, and G140S/Q148H and the EVG-resistant mutants E92Q and T66I (Smith et al. 2021b). However, few standard mutations, mainly the ones involved in Q148 and N155 pathways, have been linked with the failure of DTG treatment, along with the selection of additional mutations, such as H51Y, L74M, T66A/I, E92Q, G118R, E138K/E/A, G140S/A, and S153Y/F, resulting in reduced susceptibility to DTG (Wainberg and Han 2015a;Brooks et al. 2019;Chitongo et al. 2019;Cook et al. 2020;Smith et al. 2021b, a). Furthermore, the effect of G140S mutation on subtype C IN has provided acumen into reduced drug affinity because of DTG resistance occurring due to reduced IN stability and greater flexibility around the 140 loop region in the mutant system (Chitongo et al. 2020).
A case study by Hardy et al. (2015) identified novel DTG mutational pathways in a RAL-resistant patient harbouring N155H when substituted with a DTG regimen for around 10 months. The mutations at T97A and E138K in IN led to a 37 FC resistance to DTG. After employing DTG for 10 further months, the sequential acquisition of mutations at L68FL, A49P, and L234V led to further resistance to DTG (~ 63 FC). Of these, except for L68FL, the other two are novel mutational pathways leading to emergent DTG resistance while on salvage therapy with the N155H mutation. In a similar case, an HIV-infected patient with high-level resistance to all PR and RT inhibitors and N155H, S119R, and E157Q mutations in IN after RAL failure was treated with DTG and RAL was subsequently removed from the regimen. A month later, the viral load was undetectable, and after about 8 months, a pair of novel mutations T97A and S147G in IN were detected (Wainberg and Han 2015b, a). These outcomes propose that DTG should be used with caution in INSTI salvage therapy for RAL failures or in patients with multiple RAMs (Alaoui et al. 2019).
Another study (Fulcher et al. 2018) showed that no INSTI is impervious to resistance, no matter how high its resistance barrier is. They showed virologic failure in treatment-naive individuals owing to the rapid evolution of IN mutations (T66I and M184V) in the presence of DTG, along with a simultaneous spike in plasma viraemia. In another study, the interesting association of polymorphic secondary mutation K156N (including other minor mutations leading to low susceptibility-K211R and E212T) with N155H was depicted, which was enough to confer impactful resistance to DTG; about fivefold change (FC, which is the decrease in potency of IN mutant against WT) was observed in IC 50 as paralleled with N155H alone (Malet et al. 2018). DTG failure has been detected predominantly via the acquisition of manifold mutations at a given time.

Bictegravir (Biktarvy, GS-9883)
BIC was the next second-generation INSTI to be approved in 2018. It was designed and developed using structure-activity relationship data of DTG by modifying its oxazine ring with a series of oxaza-bridging and a 2,4,6-trifluorophenyl group, which led to an enhanced potency against G140S + Q148R double mutant. Being potent it has several advantages over first-generation agents, including a low potential for drug-drug interactions, a high genetic barrier to resistance like DTG, and administration without regard to meals or a boosting agent (Brooks et al. 2019). Although BIC has a high resistance barrier, the in vitro IN resistance selection tests over a long time frame verify the emergence of HIV-1 resistance to BIC comprising S153Y/F or R263K along with or without M50I (FC ~ 7) mutations. These substitutions have low-level reduced susceptibility to BIC and may occur in vivo after its wider use (Acosta et al. 2019;Smith et al. 2021b).
Similar to other INSTIs, even BIC coordinates the pair of chelating motifs, Mg 2+ ions and stacks the halobenzyl moiety with the base of the penultimate cytosine. The favourable π-stacking interaction amid the halobenzyl tail with the 3' cytosine of the vcDNA aids in its potency and resistance profile. The displaced terminal adenosine adapts to multiple rotameric conformations. Its oxazepine ring makes several interactions (van der Waals) with the β4α2 loop near the IN active site, resulting in a tighter binding. The methylene bridge in its ring imparts rigidness to its structure, allowing BIC to retain the ability to bind despite the changes in the geometry of the mutant IN active site (Smith et al. 2021b), refer to Fig. 7.
As per the HIVDB, residues T66K, E92Q, G118R, E138K/A/T, G140A/S/C, Q148H/R/K, N155H, and R263K are liable for IN resistance against BIC (Stanford HIV-1 Drug Resistance Database 2019). A study conducted by Anstett et al. (2017) demonstrated that the second-generation INSTIs faced the highest FCs in resistance due to the Q148 pathway when at least a pair of secondary mutations were present. However, these FCs are almost always below those that have been observed with first-generation INSTIs.
In several phase III clinical studies, BIC has shown noninferiority to DTG in treatment-naïve patients. However, it has led to weight gain in some patients (Smith et al. 2021b).
The in vitro susceptibility studies and animal models depict a 14 FC and 6 FC spike in the EC50 (i.e. 50% effective concentration) due to the G118R and R263K mutations, respectively, whereas a simultaneous 4.8 FC decrease in susceptibility to BIC is observed due to G140S + Q148H combination of mutations. If E138K joins this latter combination, it further reduces the BIC susceptibility 10 FC. To avoid these susceptible mutations and those observed during virologic failure of DTG monotherapy, BIC is co-formulated with tenofovir alafenamide or emtricitabine (Acosta et al. 2019;Wensing et al. 2019 An experiment conducted by Ghasabi et al. (2022) showed that the mutational combination of a naturally occurring polymorphism, M50I with a primary mutation, R263K and a minor accessory mutation L74I, reduces the susceptibility to BIC potency. Another study (Smith et al. 2020) showed that the IN with triple or quadruple mutations having RAL-resistant G140S + Q148H substitutions along with an accessory mutation at G140S/Q148H/V75A, G140S/ Q148H/T122N, and G140S/Q148H/L74M/V75A reduced the BIC potency by more than 2.5 FC. The IN mutants developed by the team (C56S, G149A, and G140S + G149A) were all susceptible to BIC. Moreover, only BIC exhibited the potency to inhibit the RAL-resistant double mutant G140S/Q148H and the DTG-resistant IN mutants R263K, G118R, and H51Y/R263K with EC50 values less than 5 nM . Therefore, they concluded that RAMs with single substitutions that are located about 5 Å from the divalent ions bound at the IN active site generally do not cause an extensive reduction in susceptibility to BIC potency (Smith et al. 2020).
While BIC has not yet been employed in patients with prior INSTI failure, in vitro data propose that BIC and DTG have an extensively overlapping resistance profile (Scutari et al. 2020). Although BIC retains potency (FC ~ 5) against the DTG mutants (T66I + L74M + E138K + S147G + Q148 R + S230N = FC ~ 33 towards DTG) by inhibiting the replication of a broad spectrum of IN mutants (Smith et al. 2020), it was susceptible to several other combinations of mutations, such as triple mutants G140S + Q148H + G149A (FC ~ 13), G140S + Q148H + N155H (FC ~ 30), E 1 3 8 K + G 1 4 0 A + Q 1 4 8 K ( F C ~ 1 1 7 ) , a n d V72I + E138K + Q148K (FC ~ 36) and the quadruple mutant L74M + G140S + S147G + Q148K (FC ~ 147). Moreover, these INSTI-resistant triple mutants when tested against Fig. 7 (a) Binding orientation and polar bonds (in black) of BIC*3OYA (PDB ID) interactions. The green spheres represent magnesium ions, the yellow sticks represent BIC, and the pink sticks are the amino acid residues involved in the ligand*receptor interaction. (b) Structural models of the PFV 3OYA intasome with BIC as ligand. The coiling orange structure represents the viral complementary DNA (vcDNA) fragment, wherein the ligand can be seen interacting via the π-stacking bonding with the terminal DA17 of the vcDNA. Image created with the help of PyMol (Schrödinger and DeLano 2020) CAB depicted a substantial decrease in their susceptibility (Smith et al. 2021b, a). These remarks emphasize the earlier mentioned tenacious need to develop novel INSTIs that can maintain potency against evolving IN mutations, particularly those that preserve potency against the variants capable of causing multiple mutations.
The mutation scores (MtS) refer to the summation of individual RAMs penalty score for a drug and determines its estimated resistance level.

Natural anti-HIV molecules
Recently, to treat HIV/AIDS there has been a surge in isolating the natural substances as active leads from plants and other natural sources. The chief obstacle is the affordability of the treatment that does not involve any risk of drug resistance and serious side effects, thereby creating a strong demand to assess drugs derived from plants as well as their derivatives (Kaur et al. 2020). Some plant compounds are known for modulating multiple cellular factors involved in HIV replication, which makes them potential anti-HIV candidates. Therefore, it is vital to search for new antiretroviral agents which can be combined with or can substitute with the present antiretroviral drugs. Although several studies on multiple natural compounds have depicted enormous potential to become drug leads owing to their anti-HIV activity that may yield effective and affordable therapeutic agents, it has not been well documented or its literature is vague and scanty (Kurapati et al. 2015).
Multiple bioactive compounds can inhibit HIV, these are broadly categorized as phenolics, flavonoids, triterpenes, alkaloids, polyphenolics, lectins, coumarins, saponins, lignans, tannins, lactones, iridoids, O-caffeoyl derivatives, xanthones, sulphated polysaccharides, polyketides, quinones, and phospholipids (Salehi et al. 2018). There are also some ribosome-inactivating plant proteins, like MAP30 and MRK29, from Momordica charantia fruits, Trichosanthin from Trichosanthes kirilowii roots, and GAP31 from Gelonium multiflorum seeds that inhibit the HIV-1 infection in T-cells and monocytes via IN inhibition, except for MRK29 which inhibits RT. Moreover, these have an N-glycosidase activity on 28S ribosomal RNA and a topological activity on long terminal repeats of HIV (Mandal et al. 2020;Kaur et al. 2020). More such plants comprising proteins employed in HIV/AIDS have been discussed by Kaur and their team (2020). Further, there are proofs where plant extracts depicted anti-HIV activity considerably, but need supplementary studies to isolate the major active compounds responsible for the same. Several such prominent compounds and plants have been summarized in Online Resource Table S1. Mandal et al. (2020) and Kurapati et al. (2015) have summarized the clinical trials of HIV/AIDS employing natural compounds. As reported in 2018, out of 717 species, HIV-RT, HIV-PR, and HIV IN inhibitory activities have been tallied to about 206, 254, and 43 species, respectively. Apart from these studies, researchers have also estimated nearly 390 species for other enzyme inhibition studies which are categorized under the umbrella of anti-HIV activities (Salehi et al. 2018). Since the main focus of this review paper are INSTIs, we will be elaborating on the docking interactions for a few bioactive compounds/plants exhibiting IN inhibition.
Online Resource Table S1 indicates that the potential IN enzyme inhibition is due to the presence of the catechol, galloyl, and sugar moieties. In the computational docking of the bioactive compounds composing the crude extract of Dioscorea bulbifera L, Chaniad et al. (2016) found myricetin to be a potent INSTI as it comprised the galloyl moiety which binds strongly with the IN amino acids. It interacts with several amino acid residues via hydrogen bonding-Thr66, His67, Asp116, Glu152, Asn155, and Lys159. Another compound, quercetin-3-O-β-d-glucopyranoside comprised both the catechol and the sugar moieties; the former was involved with the IN inhibition, whereas the latter enhanced this activity by increasing its solubility. It forms hydrogen bonds with the enzyme's Thr66, Glu92, Asp116, Gln148, and Lys159 residues, exhibiting an IC 50 of 21.80 μM (Kaur et al. 2020).
The computational docking of crude extract of Betula alnoides showed potential interactions of its compounds with IN active sites. The results directed that betulin interacted    105  120  45  45  ----G140S + Q148N  40  45  10  10  ----G140S + Q148N + G163R  55  60  10  10  ----with Thr66, His67, and Lys159 residues involved in the ST reaction and with Asp64, a catalytic triad residue involved in 3′P of the integration process. The binding energy of betulinic acid was lower than that of botulin as it interacted with the same residues except for His67. Its oleanolic and ursolic acids showed weak interactions with Thr66 and Gln148, and Gln148 and Lys159, respectively (Chaniad et al. 2019). Similarly, catechin interacts with Thr66, Gly148, and Glu152 in the CCD, whereas protocatechuic acid was reported to interact with Thr66, His67, Glu152, Asn155, and Lys159. Both these compounds interact with Glu152 which is a catalytic triad residue; further catechin possesses higher activity and lower binding energy than protocatechuic acid, making it a better INSTI (Panthong et al. 2015;Kaur et al. 2020). Another docking study of chicoric acid shows that it binds to several amino acid residues depending on its orientation, at times with hydrophobic interactions between its two phenyl arms and E152, I141, and Q148 and at times with E152, K156, C65, and E92 while pointing towards H67 and T66. Its phenyl arms are entirely anchored via hydrophobic interactions, one arm coordinated by the catalytic triad residue D116 and Q148 and the other arm by C65 (Nobela et al. 2018). Another study conducted by Siwe-Noundou et al. (2019) docked IN with L-chicoric acid (LCA) and Alchornea cordifolia's methylgallate (MG). This protein-ligand interaction analysis indicated that one of the vicinal OH-group in MG formed an H-bond with the thiol group, Cys65. Unique to LCA was its ability to interact with Gln148 (which is found in the 140 s loop) facing the active site, for which its aromatic rings have to be perpendicular for optimum interaction.
Apart from anti-HIV natural substances extracted from plants, INSTIs can be of other origins too, like microbial (fungal, algal, cyanobacterial), marine vertebrates and invertebrates, etc. Despite marine invertebrates producing a variety of antivirals, limited products have reached clinical trials or have been approved for use. One such example is tris-phenethyl urea, molleurea A, and cyclopeptides, mollamides E (IC 50 = 39 μM) and F, extracted from tunicate Didemnum molle's methanol extract, which demonstrated the ability to inhibit HIV-1 replication as well as behaved like an INSTI. There have also been studies on Lamellarin alkaloids produced by the marine ascidian (mollusc), Lamellaria sp. (Riccio et al. 2020).
If one carefully observes Online Resource Table S1, it is evident that some bioactive molecules act as dual inhibitors (DIs); these comprise a pharmacophore that enables it to bind to two varying target enzymes to produce synergistic or additive effects. These are usually either IN and LEDGF-dependent IN or IN and RT-associated ribonuclease H inhibitors. The latter is the most common type of DIs, sharing a common catalytic site residue and geometry, like quinolinonyl β-diketo acid derivatives, 2-hydroxyisoquinoline-1,3(2H,4H)-diones, madurahydroxylactone, and pyrrolyl derivatives. Some examples of the former category include CHI-1043, CHIBA-3002, CHIBA-3003, CHIBA-3053, GSK1264, Tertbutoxy(4-phenyl-quinolin-3-yl)-acetic acid, and CX04328 (Di Santo 2014; Choi et al. 2018;Ercan et al. 2019). A universal drawback of ribonuclease H is that it has a human homolog; this obstructs the designing of DI that specifically inhibits the viral enzyme and keeps the human homolog intact. Consequently, it is better to find biological targets other than IN that can be inhibited simultaneously with IN when formulated as a DI, tackling the resistance problems faced earlier (Mahboubi-Rabbani et al. 2021).

Conclusion
• Hitherto, INSTIs arguably are the most effective and well-tolerated ARV drugs to be discovered and developed. The introduction of INSTIs to the HIV therapeutic landscape provides safe, effective, and viable HIV treatment alternatives. • Selecting which regimen to use in initial HIV treatment acts as a crucial decision. It is predicted that the older first-generation INSTIs, RAL and EVG, will soon be phased out in near future owing to the potential for novel formulations to revolutionize patient care and advantages offered by second-generation INSTIs in terms of drug interactions, tolerability, resistance barrier, safety, cross-resistance, dosing strategy, and efficacy in both treatment-experienced and treatmentnaïve patients. • Although INSTIs are highly potent drugs, as the frequency of use of INSTIs continues to rise, inevitable mutations will develop that confer resistance against all currently available ARVs. Therefore, more computational studies (in silico) and clinical experiments (in vitro and in vivo) are required for the characterization of the resistance profile of all FDA-approved INSTIs. • Promising natural or synthetic compounds that retain potency against the initial set of mutants can be tested against more complex IN mutants and for their ability to select additional mutants. • More sensitive next-generation techniques, assays, and equipment are vital for identifying low-level viraemia and minority low-level resistance variants. • To infer based on the available data, BIC is the most suitable and promising option. This and its structurally related analogues can be further studied and employed accordingly in a novel INSTI structure-based design having a better resistance profile than present drugs. • In addition to that, INSTIs designed to interact more strongly with the 3'-adenosine nucleotide could potentially adapt to the modified geometry of the IN active site that is responsible for some of the most problematic RAMs. With this, there is a high possibility that the HIV-1 intasome structures addressing numerous popular IN mutants will be solved soon. • Bioactive compounds alone or as DI or in combination with HAART can be prescribed for accepted toxicity and fewer RAMs. This optimistic outcome is bound to inspire researchers for a positive breakthrough in the near future.
Author contributions All the authors have contributed to the study's conception and design. Material preparation, data collection, and data analysis were performed by SKS and TS. The first draft of the manuscript was written by MQ and SKS and all the other authors are involved in critical revision, project administration, and supervision. All the authors have read and agreed to the final version of the manuscript.
Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Data Availability All data generated or analyzed during this study are included in this published article [and its supplementary information files.