To enhance the interaction between ketolides and the ribosome, alpha halogenation on C2 by a fluorine atom has been undertaken in a stereospecific manner. The presence of this fluorine atom in an alpha configuration on C2, would help prevent the enolization of the C3 ketone found with Telithromycin or other ketolides. This chemical variation, coupled with the replacement of the imidazole-pyridine couple of Telithromycin by a 1,2,3-trizolyl-aniline, led to the development of Solithromycin (Figure 8). Solithromycin or CEM-101 actually presented better chemical stability, better affinity for the 50S subunit of the bacterial ribosome by the creation of a third interaction site [133] - [138]. It does not inhibit nicotinic acetylcholine receptors nAChRs like Telithromycin, so should not have the harmful side effects of Telithromycin [135].

Solithromycin also showed good activity in vivo and in vitro on most respiratory pathogens including strains resistant to other macrolides as well as atypical germs. It has been shown to be 8 to 16 times more effective than Azithromycin both on sensitive germs and on those resistant to Azithromycin [135] - [140]. Furthermore, the interaction of Solithromycin with the third ribosomal site makes it possible to limit the development of bacterial resistance [135] - [141]. Solithromycin also has a broad spectrum of activity against methicillin-resistant Staphylococcus aureus strains, Enterococci and the genus Plasmodium in animal models. It also has a relevant activity on bacteria responsible for urinary tract infections such as Neisseria gonorrhoeae including the multidrug resistant strains Chlamydia, Mycoplasma, Genitalium and Ureaplasma [141] - [146]. For some reasons in particular of insufficient data on hepatotoxicity, the FDA has

not approved the marketing authorization for this first fluoroketolide. However, the search for other 2-fluorinated derivatives of Solithromycin was continued. Thus, the variations undertaken were aimed at reducing hepatotoxicity and widening the performance of fluoroketolides. To do this, the first chemical variations used an atypical approach which called on the ribosomes involved in resistance mechanisms and mutations as a platform for developing new, more effective and more selective fluoroketolides [147]. Thus, by this method which also uses a structurally diverse library of alkynes, the aniline of the triazole of Solithromycin has been replaced by various aryls and heteroaryls. Among these new derivatives, those in which the triazole carries a thiophene heterocycle are particularly illustrated by their effectiveness. In fact, two positional isomers of thiophene (4-thiophene isomer and 5-thiophene isomer) (Figure 9) exhibited significant activities against strains of E. coli and against standard and resistant strains of S. pneumoniae [148]. Other modulations of Solithromycin have focused on the entire structural link of the C11-C12 cyclic carbamate. This is how the butyl-triazolyl-aniline chain of Solithromycin was replaced by a butenyl-quinoline link (compound 1) or by a propoxyl-quinoline (compound 2) (Figure 9). These compounds have shown in vitro antibacterial activities that can be superimposed or even superior to those of Clarithromycin and Telithromycin against sensitive and resistant macrolide germs. The activity of compound 1 with the butenyl-quinoline chain was also 8 times greater than that of telithromycin against the strain CSMS 12-1 of M. catarrhalis [149].

The butyl-triazolyl-aniline chain on the carbamate of Solithromycin has also been replaced this time by a rigid propargyl-like link carrying the triazole-pyridine couple. Among these new derivatives of Solithromycin, the compounds FSM-100573 and FSM-100563 (Figure 10) have shown antibacterial performances superior to those in particular of Azithromycin, Telithromycin and Solithromycin against some resistant strains extremely difficult such as S. pneumoniae and Pseudomonas aeruginosa [150]. Another chemical variation operated is to move the link of the carbamate cycle on the oxygen atom in C6 as in the case of Cethromycin. These new fluoroketolides carry at O6, a rigid propargyl-like link carrying a terminal heteroaryl-isoxazolyl group [151]. The leading compound in this series or compound 3 (Figure 10) is effective against

the strains of S. pneumoniae, S. pyogenes and H. influenzae with an antibacterial properties 10 times that of Clarithromycin against the sensitive strains of S. pneumoniae. The same compound is also 100 to 16,000 times more effective against strains of S. pneumoniae resistant to Erythromycin [151].

Always looking for new derivatives of Telithromycin that are more effective, in particular with regard to respiratory tract infections of community origin that are sensitive and multidrug-resistant and especially less hepatotoxic, the new chemical variations undertaken using Clarithromycin have made it possible to dot the series of azetidinyl carbamate ketolides [152] [153]. The chemical modulation undertaken essentially concerned the butyl imidazolyl-pyridine group of Telithromycin: the butyl chain was heterocyclized into azetidine, the nitrogen of which carries a methyl diazanaphthalene group of the 1,5- or 1,8-naphthyridine-like substituted (Figure 11). The first derivatives comprising the 1,8-naphthyridine heterocycle (compounds 4 and 5) were found to be favourable for the metabolism induced by human aldehyde oxidase (AO). On the other hand, the 1,5-naphthyridine derivative carrying a hydroxyl group in position 3 has been shown to be poorly metabolizable by AA even if the regioselectivity of the hydroxylation has not been established. Attenuation of the effect of AO appeared not with other polar functional groups but rather with the positional rearrangement of

the nitrogen atoms of naphthyridine. This is how the 1,5-naphthyridine derivative (compound 6 or PF-04287881) (Figure 11) has been shown to be more stable in the human hepatic cytosol. This molecule has a good ability to attenuate hepatotoxicity while minimizing hepatic renewal and inactivation as a function of time of the CYP3A isoforms in the liver [152]. In addition, this azetidinyl ketolide (PF-04287881) has an efficacy and a significant antibacterial activity against the germs responsible for infections of respiratory tracts of community origin that are sensitive or multi-resistant [152]. This is why it was selected for clinical development.

In order to optimize therapeutic activities against both sensitive and resistant respiratory pathogenic bacteria, another chemical variation consisted in replacing the methyl initially present on the C10 or C12 carbon of ketolides with non-polar groups [147] [154] [155]. The aim of such an option was to keep in ketolides series, the same overall hydrophobic character observed at the level of the underside of the neomacrolide macrocycle [156] [157] [158] [159]; which could lead to new ketolides with an antibacterial spectrum and more effective pharmacokinetic properties. The pharmacomodulation undertaken consisted in replacing methyl in C12 by a vinyl-like group [154] or to introduce in addition on C10, a hydroxyl [147] or also in replacing methyl in C10 by an alkylamine group [155].

1) 12-vinyl ketolide derivatives

In the 12-vinyl ketolides series, two derivatives are particularly illustrated both by their antibacterial activities and by their pharmacokinetic properties [154]. These are Telithromycin derivatives which differ from one another by the presence on the pyridinyl nucleus of the N-butyl imidazolyl-pyridine sequence, either of a methyl (compound 7) or of a chlorine atom (compound 8) (Figure 12).

The antibacterial activities of compounds 7 and 8 were comparable in vitro to those of telithromycin. However, the presence of the vinyl group, doubled by that of methyl or chlorine on pyridine, reinforced the lipotropy of the hydrophobic surface of the macrocycle.

Thus, this contributing to a profound modification of the pharmacokinetic

and pharmacodynamic aspects of these compounds in vivo in rats. Indeed, vinyl ketolides had larger lung distribution volumes and longer half-life times; so that they had a much better pharmacodynamics by their better activity in vivo against Streptococcus pneumoniae and Haemophilus influenzae compared to Telithromycin [154].

2) C10-substituted ketolide derivatives

Following the performance of 12-vinyl ketolides [154], other regioselective modifications have been undertaken. They consisted in introducing either alongside the methyl at C10, a hydroxyl group [147], or by substituting this methyl with an aryl group or with an N-arylamine function [155]. Thus, three series were obtained: the series of 10-hydroxyl ketolides, that of 10-alkylaryl ketolides and that of 10-alkylamine ketolides and analogues. The objective was also to maintain the structural coherence of the macrolides with their bifacial characteristic. The results of the in vitro antibacterial screening of these new ketolides against a panel of respiratory pathogens and compared to the activities of Telithromycin and Clarithromycin, unfortunately showed that the presence of hydroxyl was not favourable to the expected results. The 10-hydroxyl ketolides (compounds 9, 10, 11) and their analogue 10,11-epoxide (compound 12) (Figure 13), were clearly less effective than Telithromycin and Clarithromycin [147]. Such a poor performance could be attributed to a conformational modification induced by the presence of hydroxyl in the prepared ketolides; which undoubtedly changed the tight 10-hydroxyl ketolide binding capabilities with their ribosomal target. In fact, the methyl at C10, initially in the indispensable beta configuration, is this time oriented in alpha with respect to the plane of the macrocycle (Figure 13).

In series of 10-alkylaryl ketolides and 10-alkylamine ketolides (Figure 14), the SAR studies undertaken noted that to induce good antibacterial activity, the nitrogen of ketolide 11,12-carmabate should not be substituted. In addition, only the C10-benzyl (compound 13) and C10-benzylamine (compound 14) derivatives are found to be twice as effective as Clarithromycin against sensitive and resistant strains of S. pneumoniae. Replacing the phenyl of benzyl with a basic heterocycle like pyridine does not improve the antibacterial activities [155]. Against certain resistant strains of S. aureus (BAA976 M mef and BAA977 iMLS

erm), the benzylamine derivative (compound 14) was found to be 32 times more active than Clarithromycin [155]. As for the analogue C10-spiro-pyrazole (compound 15) (Figure 14), this one presented against the 3914 MLS strain of S. pneumoniae, a moderate activity compared to Clarithromycin. On the other hand, like benzylamine, spiro-pyrazole was also found to be 32 times more effective than Clarithromycin against resistant strains of S. aureus [155].

In order to study the structural factors at the origin of resistance to macrolides, the amino sugar D-desosamine of Telithromycin undergoes several pharmacomodulations in particular a C2’-deoxygenation (compound 16), a 3’-desmethylation (compound 17) and an N-butylation of this desmethylated derivative (compound 18) (Figure 15). All the new derivatives showed moderate antibacterial activities against the sensitive strains ATCC29213 of S. aureus. However, they were completely inactive against the constitutive resistance of S. aureus. These results could indicate that the steric repulsion between the desmethylated cycle A2058 and the desosamine cycle cannot be considered as a main reason for resistance against the ermA strain ATCC3359 of S. aureus [160].

A series of acyclic carbamate derivatives which can be considered as originating from the opening of the oxazolidinone (11,12-carbamate) nucleus of Telithromycin and its rearrangement into 11-O carbamoyl derivatives have been recently proposed [161] [162]. In fact, these derivatives are obtained directly from 3-ceto clarithromycin by 11-O-carbamoylation (Figure 16). Such a chemical variation led to the production of a series of carbamoyl ketolides, two derivatives of which exhibited effective activities compared to that of Clarithromycin. Indeed, the compounds 17 and 18 carrying on the carbamate nitrogen, a propyl chain 1,2,3-triazole variously substituted (Figure 16) have shown in vitro a good activity not only against bacterial germs sensitive to Clarithromycin but also on resistant bacterial strains. These O-carbamoyl ketolides are in particular more effective in vitro on strains resistant to Erythromycin, in particular Streptococcus pneumoniae and Streptococcus pyogenes. In terms of sensitive germs, these triazole compounds have shown better activity against the strains of Staphylococcus aureus ATCC25923 and Bacillus Subtilis ATCC9372 [161] [162].

Carbonate ketolides were designed following the replacement of 1,3-oxazolidinone from Telithromycin by a cyclic carbonate or 1,3-dioxolanone at C11 and C12. In addition, these new ketolides are likely to undergo a second modification in their amino sugar (D-desosamine). Thus, two series of ketolide carbonates have been developed: those without modification of the amino sugar and those which have also undergone a modification at the level of D-desosamine. Like 11,12-carbamate ketolides or oxazolidinone ketolides, the creation of a cyclic carbonate or 1,3-dioxolan-2-one in position C11 and C12 would lead to an improvement in the antibacterial performance of the macrolides obtained. In addition, these compounds would have much better stability in an acid medium as well as better pharmacokinetic properties [109] [163] [164] [165] [166]. Among these carbonates, those from Erythromycin (compounds 19) or 3-ceto-Clarithromycin or compound 20 (Figure 17), exhibited antibacterial activities twice that of Erythromycin A [109] [163] [164]. This superior antibacterial performance of 11,12-carbonate derivatives is due to a better affinity of these derivatives for the 50S ribosomal subunit [164]. Thus, these derivatives presented MICs twice lower than those of Erythromycin and Clarithromycin against strains of Streptococcus pyogenes resistant to macrolides by a constitutive or inductive MLSB mechanism [165]. In contrast, compound 20 from 3-ceto-Clarithromycin (Figure 17) has induced excellent antistreptococcal activities [164] [166]. However, the 11,12-carbonate derivatives did not experience the same development as their 11,12-carbamate analogues probably due to their high hepatotoxic potential [167].

Another pharmacomodulation has been undertaken on ketolide carbonates. It consisted in introducing on 5-O-desosamine, a hydroxyl group in position C4’. The presence of this group has made it possible to increase the antibacterial activity by increasing the interactions between the molecule and the ribosomal subunit. This is due to the fact that the desosamine sugar binds to the V domain of the 23S rRNA by a hydrogen bond. Thus, the existence of a hydrophilic group of hydroxyl-like on the amino sugar could have created another hydrogen bond with the ribosomal target in particular on the amino acid residues of the binding pocket [168]. On the other hand, when the 4’-hydroxyl group of the desosamine is alkylated by hydrophobic groups of alkyl or aryl like, there is a non-improvement of the antibacterial activities [168]. These different chemical variations on desosamine have made it possible to access new ketolides whose antibacterial activities in vitro against several sensitive or resistant microbial agents have shown relevant activities. Thus the 4’-hydroxylated ketolide carbonate or compound 21 (Figure 17), unlike its 4’-O-alkylated derivative, has shown a potent activity against all pathogens sensitive or resistant to methicillin. Thus, compound 21 would be more active than Clarithromycin for the strains ATCC29213 (MSSA), 09L075 (MSAR), 09M124 (MRSE), 09H071 (ESSP) and 09U070 (ESSPy). Compared to Telithromycin, compound 21 has shown the same potency of activity against all pathogens sensitive or resistant to methicillin [168].

Another series of neoketolides, in particular those carrying at C11 and C12 a γ-lactone, have been prepared and evaluated for their activity against respiratory pathogenic bacteria. These are derivatives derived from the replacement of 1,3-oxazolidinone (carbamate) of Telithromycin by a cyclic ester or lactone. In fact, these new ketolides have two lactone groups: that belonging to the macrocyclic aglycone and that carried by the carbons C11 and C12. This γ-lactone very often carries an α function (thiol or amine) thus leading to the class of α-thio-γ-lactone ketolides and to that of α‑amino-γ-lactone ketolides.

1) γ-lactone ketolides derivatives

Among these compounds, it has been established that only N-alkylated 6-carbamate derivatives with a rigid allyl-aryl link, in this case allyl phenyl-pyrimidine, are capable of inducing excellent antibacterial activities [169]. Furthermore, the stereochemistry at the C10 methyl level was of paramount importance for antibacterial performance. Indeed compound 22, the most active with an activity superimposable on that of Telithromycin, was that endowed with a type R stereoisomerism at the level of methyl in C10 (Figure 18) [169].

2) α-thio-γ-lactone ketolides derivatives

Following a Michael intramolecular addition reaction of an aryl alkylthio acetic acid ester enolate to an α, β-unsaturated ketone, a series of γ-lactone ketolides were prepared. The new lactones fused in C11 and C12 carried in the alpha position, a thiol function variously substituted by alkyls, alkyl-aryls or by aryls (Figure 19) [170]. Some of these neoketolides have been the subject of an in vitro antibacterial evaluation against sensitive and resistant pathogenic respiratory strains and compared to Telithromycin. Thus, against sensitive germs such as S. aureus and S. pneumoniae and H. influenzae, the best compounds of the diastereoisomer type (compounds 23, 24 and 25) (Figure 19), have shown a similar antibacterial spectrum and an activity comparable to that of Telithromycin. However, such compounds were less active against resistant strains of S. pneumoniae and S. aureus.

The S-ethyl adenyl derivative (compound 24), was particularly illustrated by its antibacterial performance against certain Gram-negative germs such as E. coli [170]. Compared to Erythromycin and Clarithromycin, this chemical variation led to much better pharmacokinetic characteristic. Thus compound 23 (S-ethyl pyrazolyl-pyridinyl) and compounds 24 and 25 or S-alkyl adenyls, have shown in oral mice bioavailability (around 60%) and significantly higher plasma levels. Furthermore, their plasma half-lives were of the order of 1.1 to 1.2 hours in mice, thus suggesting a once-daily administration in humans [170]. Consequently, an in vivo screening in mice of compounds 23, 24 and 25 was carried out against a sensitive strain of Streptococcus pyogenes. It has been noted that compounds 23 and 25 were slightly less active, but nevertheless in the same performance range as Telithromycin. However, all of these compounds were significantly more effective than Erythromycin [170].

3) α-amino-γ-lactone derivatives ketolides and analogues: Nafithromycin

The α-amino-γ-lactone ketolides can be considered to be derived from the 11,12-carbazate ketolides or N-amino 11,12-carbamate ketolides [132] by removing nitrogen from the carbamate. Indeed, they were proposed following the antibacterial performances of 11,12-carbamate. The aim of such a chemical variation was to modulate, by various substituents, the α-amine function of γ-lactone. The α-amino-γ-lactone ketolides are therefore considered to be versatile intermediates for accessing neoketolides with optimized antibacterial activity and enhanced pharmacokinetic performance. Thus, from the 3-ceto clarithromycin carrying the α-amino-γ-lactone group in C11 and C12, two chemical series were prepared: the chemical series or the α-amine function was substituted by various side chains of the type aryl or heteroaryl following an amidation reaction (N-acylamine derivatives) [171], and that or the α-amine has been replaced by an amidoxime [172].

In the series of α‑N-acylamine γ-lactone ketolides, a derivative is particularly illustrated by its performance. This is compound 26 carrying the N-2-methylthio quinoxaline link (Figure 20). Indeed, this compound is particularly effective in vitro against Gram-positive pathogenic germs sensitive to Erythromycin. In addition, it was effective against strains of S. aureus resistant to MLSB, S. pneumoniae resistant to efflux and S. pneumoniae and S. pyogenes resistant to MLSB.

Furthermore, compound 26 was approximately 4 times more effective against BML-resistant S. pyogenes and H. influenzae than Telithromycin [171]. The presence of the N-acylamine chain in α of the γ-lactone ketolide is therefore important for inducing serial antibacterial performances of the neoketolides.

In series of antibacterial α-amidoxime 11,12-γ-lactone ketolides, derived from amines by replacement of the latter by an amidoxime, only one representative is currently in clinical development. These are WCK 4873 or Nafithromycin developed by the Research Center in Wockhardt, India [172] (Figure 20).

Nafithromycin is structurally presented as a 3-keto Clarithromycin with a γ-lactonic heterocycle joined to C11 and C12 of the aglycone. The γ-lactone carries in α an amidoxime whose oxygen atom is alkylated in position 1 by an ethyl group itself fixed in position 2 of a 1,3,4-thiadiazole carrier in position 5 of a 2-pyridinyl. Nafithromycin with this structural characteristic based on the α-amidoxime group and a hydrophilic side alkyl chain, would have a much better interaction with the target domain II and the target domain V including modified, of rRNA. Such an affinity would be at the origin of potent antibacterial activity against strains collected worldwide [173] [174] [175] [176] [177] including against strains resistant to macrolides, in particular Telithromycin and Solithromycin [178] [179] [180]. In terms of pharmacokinetics and tolerance, Nafithromycin has been shown to be generally well tolerated at all doses. No serious or severe side effects have been observed [181] [182]. Unlike most macrolides and ketolides, Nafithromycin has also been shown to be non-inhibitor of all major human CYP isoforms, even at significantly higher concentrations. These results suggest a lower clinical drug interaction potential for Nafithromycin [183].

Thus, the excellent antimicrobial profile, pharmacodynamics, pharmacokinetics and the safety and human tolerability data of Nafithromycin, mean that it is currently in clinical development III in India. Nafithromycin is thus presented as the future broad-spectrum neomacrolide antibiotic for the treatment of bacterial pneumonia of community origin endowed with a notable activity against the typical and atypical respiratory pathogenic germs and the resistant strains [173] - [180].

As previously in the neomacrolide series, chemical modulations of the C9 ketone have also been undertaken in the ketolide series. The objective was the same, namely to prevent the formation of inactive 6,9-hemiketal for better stability of ketolides in gastric acid medium. Furthermore, it is also a question of optimizing the therapeutic and pharmacokinetic performances as well as the tolerability of the neoketolides obtained. In this perspective, the C9 ketone function of kétolides, like that of Roxithromycin, has been transformed into its oxime analogue and O-alkyl oxime derivatives.

1) Oxime ketolides

From the 11,12-carbamate ketolides or from the 11,12-carbonate ketolides a C9 oximation reaction was undertaken to give rise to the 9-oxime ketolides. The latter, in particular those of the 11,12-carbonate ketolides series (Figure 21) with compound 27 [184] exhibited better activity both on sensitive and resistant strains than other macrolides including Staphylococcus aureus resistant to Erythromycin by a constitutive MLSB mechanism [184] - [190]. However, the activity although being more stable in gastric acid medium, compound 27 has proved to be ineffective against S. pneumoniae resistant to Erythromycin [184]. To improve the activity of these oximes, in particular against H. influenzae and S. pneumoniae resistant to Erythromycin, two types of chemical modulations have been undertaken: either a chemical variation on the D-desosamine of the oxime ketolides, or either O-alkylation of these oximes (Ether-oxime ketolides). The first chemical modulation conserved the oxime in C9 but introduced at C4’ of D-desosamine of 11,12 carbonate 3-keto-Clarithromycin, a hydroxyl group or a carbamate group [191]. Such variations are thought to cause damage to the center of peptidyl transferase (CPT) in the exit tunnel and inhibit the formation of peptide bonds thanks to the presence of the 4’-O-carbamate group on desosamine. This activity would be comparable to that identified in 16-membered ring macrolides carrying mycaminosis-mycarose disaccharide in position C5 [191]. Thus, the derivative 11,12-carbonate C9 oxime (compound 28) and its derivative of 5-O-4’-carbamate (compound 29) (Figure 21), presented good antibacterial activities on respiratory pathogens comparable to those of Telithromycin and Clarithromycin [191].

2) Ether-oxime ketolides

The O-alkylation of C9 oxime, as in Roxithromycin, constitutes the second chemical variation operated in order to improve the antimicrobial activity of the ether oxime ketolides obtained. Chemical modulations have been undertaken on 11,12-carbamate ketolides and 11,12-carbonate ketolides. Several series of ether-oxime ketolides have been developed, four of which are currently under intense investigation. These are methyloxime derivatives, N-(aryl-alkyl) acetamidyl oximes, 9-O-arylpropenyl oximes and 9-O-arylpropargyl oximes.

Methyloxime ketolides were accessed from Telithromycin following the replacement of the C9 ketone by a methyloxime and the replacement of the butyl group by a propyl chain but still carrying the imidazo-pyridine chain. Furthermore, in order to improve the pharmacokinetic parameters as well as the antibacterial performance of these new derivatives compared to Telithromycin, various chemical variations have also been made, in particular by the introduction at C2 of modulators of hydroxyl, methyl, fluorine like, etc. [188] [189] [190]. Among these, the derivative 2-fluoro-9-oxime (CP-654743) analogue to Solithromycin, demonstrated the best activity both on strains sensitive and resistant to Erythromycin. CP-654743 is slightly more active than the non-fluorinated analogue (CP-605006) (Figure 22), but its activity on the main respiratory pathogens is similar to that of Telithromycin. In animal models, CP-654743 has shown efficacy similar to Telithromycin and superior to the non-fluorinated analogue CP-605006 against Streptococcus pneumoniae. The pharmacokinetic profile of CP-654743 was similar to that of CP-605006 and Telithromycin in middle ear infections [188] [189] [190].

As regards the N-(aryl alkyl) acetamidyl oxime ketolides derivatives, these were obtained from the 11,12-carbonate ketolides as a result of the attachment of an N-(aryl alkyl) acetamide group to the hydroxyl oxime [185] [186]. Optimization of activity was dependent on the length of the alkyl chain and the nature of the aryl. In fact, it was optimal when it was a propyl chain carrying a heteroaryl of the quinoline or quinoxaline nature (Figure 22). The importance of these heterocycles, in particular quinoline, has also been reported to be an excellent inducer of antibacterial activity in macrolide series [149] like that of the C6 side chain of Cethromycin. In addition, an extensive SAR study has shown that the amide function with its free amine group is of paramount importance for the induction of good activity. Thus, N-(arylalkyl) acetamidyl oxime ketolide derivatives have shown potent antibacterial activities against Haemophilus influenzae and Streptococcus pneumoniae including Erythromycin-resistant strains [185] [186].

Despite their excellent in vitro activity, these ketolide acetamide derivatives have unfortunately proved to be ineffective in vivo in animal models, probably due to their poor pharmacokinetic profiles [185] [186]. In order to improve the pharmacokinetic parameters of ether-oximes while optimizing their antibacterial

activities against resistant respiratory pathogenic strains, a new modulation of oxime was operated. This is the introduction of a degree of rigidity in the alkyl linker linking oxime and the aryl. Thus, propenyl and propargyl groups carrying in this case a heterocycle of the quinoline, isoquinoline, pyridine, pyrimidine, thiophene etc. type have been introduced on the oxime of the ketolides to give respectively the series of 9-O-arylpropenyl oxime ketolides and that of 9-O-arylpropargyl oxime ketolides.

In series of the 9-O-arylpropenyloxime ketolides carrying diols at C11 and C12, this variation enables to improve the antibacterial activity of these molecules on Gram-positive germs. The activity was linked to the nature of the aryl. It was higher for compounds in which the aryl is quinoline or thiophene like (Figure 23). These compounds had potent activities against sensitive and resistant strains including Staphylococcus aureus, Staphylococcus epidermidis, Streptococcus pneumoniae and Streptococcus pyogenes [192]. In this same series of 9-O-arylpropenyloxime ketolides, this time with a cyclic carbonate at C11-C12, the introduction of a fluorine atom at C2 has the consequence of obtaining analogues of Solithromycin [193] [194]. Among these, compound 30, the oxime of which is alkylated by the propenyl-quinoline chain, has been shown to be particularly effective against the strain ATCC49619 of S. pneumoniae sensitive to Erythromycin (Figure 23). With an IC₅₀ > 5 mM, compound 30 also had a weak inhibition of CYP3A4 [194]. The fluorine atom was not involved in the interactions of the H bond with the ribosomal tunnel. Its presence only resulted in an improvement in the lipophilicity of the less hydrophilic side of the aglycone [193] [194].

Still in the 11,12-carbonate ketolide series, the other chemical variation undertaken consisted in replacing the allyl linker by a propargyl also carrying most of the time a nitrogenous heteroaryl. This variation led to the 9-O-arylpropargyloxime ketolides derivatives characterized by the formation of a more rigid bond; which turned out to be essential for a suitable conformation for binding with RNA 23S. Molecular docking studies have elucidated the disparity of SAR in this series of propargyls, and have suggested the possibility of this new binding mode different from Telithromycin and Cethromycin [192]. In addition, when propargyl is

carried by a nitrogen heterocycle, in this case quinoline, optimal antibacterial activity is obtained. The position of the nitrogen atom in the cycle is essential for the optimal binding of the molecule and its ribosomal target, for example in a 4-isoquinoline heterocycle (Figure 23) [192]. Furthermore, these 4- or 5-isoquinoline derivatives have an interesting antibacterial activity against inducible pathogenic germs resistant to MLSB and resistant to efflux, whatever the phenotypes sensitive to methicillin or resistant to methicillin [192].

Faced with the performance of azalides, in particular those of Azithromycin which has a reassuring safety profile, good stability and exceptional pharmacokinetic, pharmacodynamic and therapeutic properties, it was normal for the chemical modulations that led to the development of the Azithromycin are also taken up in series of Ketolides. The “ketoazalides or azaketolides” obtained should be at least as effective against infectious germs as Azithromycin, or even provide better molecular stability in gastric acid medium.

1) 3-keto azithromycin derivatives and their pharmacomodulation sites

The first ketoazalides were prepared from a 3-keto-9-oxime-erythromycin A. Thus, from this compound, the Beckmann rearrangement reaction was carried out successively, the reduction of the intermediate iminoether obtained, then N-methylation of the nitrogen atom introduced into the macrocycle. The first ketoazalides with 15-membered ring obtained, were all derivatives of 3-keto azithromycin or derivatives 9a-aza homoerythromycin A ketolides (Figure 24). Despite the addition of an additional nitrogen atom which has led azalides in series to an enhancement of antimicrobial performance, at the level of ketoazalides this modification alone is not enough to induce good activities. This poor performance is related to possible instability reactions which would lead to the formation of inactive 3,6-hemiketals, if the C6 hydroxyl group was not previously protected, for example by a methyl [195]. However, even the 6-O-methylated derivative (Figure 24) has been found to be largely ineffective against respiratory pathogens compared to Azithromycin and Telithromycin [116] [196]. To induce the expected antibacterial activity, 6-O-methyl has been replaced by the side chain of Cethromycin. Such a structural variation has led to the production of 6-O-(3-quinolyl) allyl-3-keto-azithromycin (Figure 24) which has effectively

become active against S. aureus, S. epidermidis, Escherichia coli and Nocardia asteroides [197].

In the end, in the ketoazalides series, the replacement of L-cladinose in C3 by a ketone function doubled or not by 6-O-methylation, are not sufficient in themselves to induce good antibacterial activities. Also, to develop ketoazalides with effective antibacterial activities, use was made of the structural study of Azithromycin adapted to that of 3-keto azithromycin. It has identified functional sites likely to undergo chemical variations beneficial for antibacterial activities [198]. This is particularly to (Figure 25):

· the alkylation by various groups other than methyl, of the C6 hydroxyl function,

· the introduction of C11 or C11-C12, respectively of an acyclic carbamate and of an cyclic carbamate,

· the halogenation of carbon C2,

· the substitution in C4’of D-desosamine or of the O-alkylation of hydroxyl in position C2’, or even of the introduction of ester or amide functions on methyl in C5’,

· the desmethylation of the nitrogen atom 9a or of its displacement in position 8a or 11a,

· narrowing or widening of the aglycone at 15-membered ring to new aglycones respectively at 14- or 16-membered ring,

· the introduction of a lactam function in which the nitrogen atom is either in position 9a or in 8a.

2) Carbamate-ketoazalide derivatives with 15-membered ring

In order to boost the serial antimicrobial activities of ketoazalides, one of the chemical variations undertaken consisted in introducing an acyclic carbamate function on the C11 hydroxyl. The studies of SAR undertaken, have noted that the open and rigid nature of carbamoyl-ethylamide-like carbamate, is not conducive to the induction of good antimicrobial activities. Indeed, compounds 31 and 32 (Figure 26) were respectively 8 times and 4 times less active than Clarithromycin and Azithromycin against various bacterial strains, including clinical strains of sensitive S. pyogenes [161] [199]. To remedy this poor performance, the carbamate group was introduced on carbon C11 and C12 in an oxazolidinone-like heterocycle as in the Telithromycin structure. Other chemical modulations have been carried out on these new 11,12-carbamate ketoazalide derivatives. These are in particular the substitution of the nitrogen of the cyclic carbamate by the chain 1-butyl-1,2,3-trizolyl-aniline of Solithromycin and the removal of the methyl previously fixed on the nitrogen in 9a of ketoazalide, 6-O-methylation of hydroxyl or even C2 fluorination like Solithromycin [150] [197] [198].

It appears from these various pharmacomodulations that the presence on ketoazalides of a cyclic carbamate C11 and C12 carrying on its nitrogen atom of the same side chain as Solithromycin, leads to an improvement of the antibacterial activities on strains resistant to Erythromycin A by an inducible and efflux MLSB mechanism [150] [197] [198]. Among these compounds, the derivative not methylated in 9a but 6-O-methylated (compound 33) (Figure 26) was found in vitro on Gram (+) germs in particular on sensitive and resistant S. aureus, more effective than Azithromycin. This efficacy was however less good compared to that of Telithromycin and Solithromycin. The same trend was also observed with resistant S. pneumoniae and with sensitive E faecalis. Compound 33 had an activity comparable to Azithromycin and Telithromycin but superior to Solithromycin against Gram (−) germs in particular on H. influenzae. Furthermore, even if its efficacy remained equivalent to that of Azithromycin on E coli, compound 33 was significantly more effective compared to Telithromycin and Solithromycin [198]. Furthermore, the introduction of a C2 fluorine atom (compound 34) (Figure 26) improves the activity on both Gram (+) and Gram (−) germs. We are indeed witnessing a broadening of the antibacterial spectrum towards the germs of the genus S. aureus, S. pneumoniae, S. pyogenes, E. faecalis, E. coli, P. aeruginosa and H. influenzae, including the multiresistant strains [150] [197] [198].

Still in series with 11,12-carbamate Ketoazalide derivatives, another chemical variation consisted in mimicking the structure of Cethromycin, characterized by the presence of an allyl-quinoline link on C6 hydroxyl [150]. Thus, the analogue ketoazalide obtained (compound 35) (Figure 26) has been shown to perform

better in vitro than Azithromycin on sensitive and resistant strains of S. aureus. However, its efficacy was less good in the same species compared to Telithromycin and Solithromycin. The activity against sensitive and resistant S. pneumoniae was also more effective than Azithromycin. However, compared to Telithromycin, compound 35 was less effective on the resistant strain. Against Gram (−) germs, especially H. influenzae, the activity of compound 35 was low compared to that of Azithromycin and Telithromycin, but it remained comparable to that of Solithromycin. This ketoazalide analogue of Cethromycin has been shown to be superimposable on the activity of Azithromycin against E. coli and above all superior to Telithromycin and Solithromycin on E. coli [150].

Another site has been modified in the structure of 11,12-carbamate Ketoazalides. These include D-desosamine in position C5. The latter was substituted either by an ester function or by an amide group (Figure 27). The non-6-O-methylated derivative, carrying the methylene ester, in particular methylene benzoate and whose carbamate has the same link as Telithromycin (compound 36), has shown a relevant antibacterial activity comparable to that of Azithromycin and Solithromycin against most strains of S. pneumoniae and the strain ATCC 19615 of S. pyogenes [200] [201]. On the other hand, the derivative carrying an amide group, not methylated in 9α but 6-O-methylated (compound 37), has proved to be less effective both on Gram (+) and Gram (−) germ compared to Azithromycin, Telithromycin and Solithromycin [150].

3) 11α-aza ketolide derivatives with 15-membered ring

A new ketoazalide 15-membered ring has been developed following the displacement of the nitrogen atom to an atypical position, particularly by the introduction of a methylamine at 11a. Furthermore, the general structure of these ketoazalides undergoes a profound modification characterized by the presence of a C9 hydroxyl and the suppression of diols at C11 and C12 diols. The ethyl group protecting the lactone function of the macrolide was also deleted (Figure 28). Pharmacochemically, the introduction of nitrogen into this unusual position in the aglycone and the profound structural changes in the active base motif were not beneficial for the induction of effective antimicrobial activities. Indeed, compound 38 (Figure 28) is found to be ineffective against Gram (+) germs such as sensitive and resistant S. pneumoniae with MICs > 128 μg/ml [202].

4) 14- and 16-membered ring ketoazalides

By narrowing the aglycone from 15-membered ring, new 14-membered ring ketoazalides have been fully developed by total synthesis [150]. In summary, this involves replacing the C9 carbonyl group of Solithromycin with an amine function. Among its new ketoazalides, the non-fluorinated compound 39 (Figure 29) proved to be less effective compared to Solithromycin against sensitive and resistant Gram (+) pathogens such as S. aureus, S pneumoniae, S pyogenes and E. faecalis [203]. On the other hand, the introduction of a C2 fluorine atom like Solithromycin, resulted in an effectiveness and a broadening of the antibacterial spectrum of compound 40 (Figure 29). Indeed, the latter was more effective than Azithromycin both on sensitive and resistant Gram (+) germs, in particular S. aureus, S pneumoniae, S pyogenes and E. faecalis. On the other hand, on Gram (−) germs the activities of the fluorinated compound are less effective like that of Solithromycin [150].

In contrast to the cycle restriction, the enlargement of the ketoazalides from 15-membered ring to a new aglycone with 16-membered ring consisted in the introduction of a methylene amine group (-CH₂-NH-) in place of the carbonyl C9

of Solithromycin. Among its 16-membered ring ketoazalides, non-fluorinated compound 41 (Figure 29) exhibited antibacterial activities on sensitive and resistant Gram (+) germs (S. aureus, S. pneumoniae, S. pyogenes and E. faecalis) superior to those of Azithromycin. However, the performances of this same compound were less good compared to Telithromycin and Solithromycin against these same germs. Furthermore, the activates of compound 41 were not very effective against Gram (−) germs compared to Azithromycin [150]. The introduction of a fluorine atom in C2 in series of the ketoazalides with 16-membered ring (compound 42) (Figure 29), proved not very favourable to the induction of a good antibacterial activity on Gram (+) germs [203]. In the end, unlike the series of ketoazalides with 14-membered ring or C2-fluoridation improve antimicrobial performance, in series of ketoazalides with 16-membered ring, the presence of this fluorine atom does not induce antibacterial efficacy on Gram (+) germs expected.

5) Ketolactams or derivatives of Aza-homoerythromycin A ketolide

This new series of ketoazalides is the result of the pharmacomodulation of 3-keto erythromycin A. These ketoazalides were obtained by introducing a nitrogen atom either between the C8 and the carbonyl at C9 or between the C10 and the carbonyl at C9 of 3-keto-erythromycin A, to obtain respectively 9a-lactams and 8a-lactams ketolides (Figure 30). The chemical synthesis of these compounds took place according to key strategic steps including an EZ isomerization followed by an aglycone enlargement reaction of the resulting 9(Z)-oxime via the Beckmann rearrangement [204] [205]. These are therefore ketoazalides with a lactam function also called Ketolactams, the most prominent of which are those of the 8a-aza-8a-homoerythromycin A ketolides series and of the 9a-aza-9a-homoerythromycin A ketolides series (Figure 30). The general structure of Ketolactams is therefore characterized by the presence of an aglycone with 15-membered ring, the creation of a lactam chain from C8 or C9, the presence of the diols C12 and C13 or their cyclization into carbonate, the maintenance of D-desosamine at C5 and by the presence of an ether chain functionalized at the level of the hydroxyl in C6. This O6 chain is most often carrying nitrogen heterocycle including quinolines or quinolones to give hybrid compounds (Figure 30).

In terms of the impact of these chemical variations on the induction of antibacterial activities, three modulations are to be retained. These are the lactam (CO-NH = 9a-aza) or retrolactam (NH-CO = 8a-aza) of the 15-membered ring aglycone, cyclization to carbonate or not of the C12 and C13 diols and of the nature of the ethereal sequence at the C6 level (Figure 30). Whatever the nature of the C6 ether chain, retrolactam-ketolides (8a-aza) performed better than their lactam-ketolide counterparts (9a-aza). Furthermore, the cyclization of the diols in C12 and C13 into carbonate, does not necessarily induce a higher antibacterial activity unless the presence of the cyclic carbonate is associated with that of an ethereal chain in C6 as effective on the ribosomal target as the chain C6 of Cethromycin [198] [204] [205] [206]. In the end, in series of ketolactams or Aza-homoerythromycin A ketolides, the whole problem of improving their antibacterial activities finally focused on retrolactams (8a-aza) with the rarely cycled C12 and C13 diols.

The only pharmacomodulation site for improved antimicrobial activity was the nature of the C6 ether chain. Therefore, we will approach this part of the SAR studies, classifying the retrolactams (8a-aza) according to the nature of the ether chain in C6. Thus, three groups are to be distinguished: the group of 6-O-alkyl retrolactams, that of 6-O-allyl carrying or not heteroaryl and that of 6-O-isopropanolyl-aminoheteroaryl. When the C6 hydroxyl is tethered by groups of the methyl, ethyl or allyl like, it is found that the antimicrobial properties, both on sensitive strains and on resistances, decreases from 6-O-methyl (compound 42) to 6-O-allyl (compound 44) through 6-O-ethyl (compound 43) derivatives (Figure 31) [204] [205]. The 6-O-methylated retrolactam-ketolide (compound 42) has shown better activity against inducible resistant S. aureus. It had a significant antibacterial activity not only against strains sensitive to Azithromycin, but also against strains resistant to Gram positive [198] [204] [205]. To improve the induction of antimicrobial activities with ether-allyls, these have been replaced by a quinoline-like heterocycle having a primordial role in the ligand-receptor binding, in particular at domain II of 23S rRNA [125] [128]

[149]. The 6-O-allyl-quinoline chain thus obtained is supposed to be an excellent inducer of antibacterial activity in series of ketolides like that of the C6 side chain of Cethromycin. In fact, the retrolactam-ketolide (8a-aza) obtained (compound 45) has shown, as the unsubstituted allyl derivative (compound 44), good activity against pathogens sensitive to Erythromycin (Figure 31). On the other hand, it was much more active than unsubstituted allyl (compound 44) against very resistant strains of S. pneumoniae, S. pyogenes and S. aureus. In particular, the 3-quinolyl compound (compound 45) was more effective against the strains of S. pneumoniae and of S. pyogenes resistant to MLSB compared to its positional isomer 4-quinolyl (compound 46) (Figure 31). However, this compound was still inactive against the B0330 strain of S. aureus, resistant to MLSB [206]. The cyclization of the diols C11 and 12 of compound 45 surprisingly strengthened the activities of the derivative 11,12 cyclic carbonate obtained (compound 47). Compared to its acyclic diol analogue (compound 45), it has a broad antibacterial spectrum extended against BML resistant S. aureus, S. pyogenes and H. influenzae. With the exception of its moderate activity against strains of S. pneumoniae resistant to BML, the overall performance of compound 47 was comparable on that of Telithromycin [206] (Figure 31).

Duplication of the same allyl-quinoline ether chain at positions O6 and O13, on the lactam-ketolide analogue (9a-aza) of compound 45, considerably improved the antibacterial activity of compound 48 obtained (Figure 31). This performance was reflected in the activity against S. pneumoniae and S. pyogenes, which are resistant to MLSB as well as against Gram-positive bacteria resistant to MLSB. The lactam-ketolide (9a-aza) or compound 48 was approximately 8 to 16 times more effective against the strains of S. pneumoniae and S. pyogenes resistant to BML than its monosubstituted analogue or compound 45 [206].

1) Acetylimine 6,11-bicyclonides: Modithromycin and derivatives.

Another approach for obtaining neoketolides capable of avoiding intra-acetalization reactions, consisted in bridging in particular the hydroxyls in C6 and C11 of 3-keto erythromycin A, by means of an acetone oxime chain whose oxygen is alkylated by a methyl-heteroaryl group: [-CH₂-C(NO-CH₂-heteroaryl)-CH₂-]. This chemical modification led to the formation of a new O6-O11 cycle to give rise to a new chemical series, that of bicyclic ketolides or 6,11-bicyclonides. The best representative of this group is Modithromycin (Figure 32). This bicyclonide is also characterized structurally by the presence of an acetylimine function by replacing the C9 ketone. The heteroaryl of the methyl-heteroaryl chain of the bridged oxime is of the 2-pyrazolyl-3-pydinyl like [207] [208]. Modithromycin has good activity on most respiratory pathogens including strains resistant to Erythromycin by an inducible MLSB mechanism, on atypical germs as well as on methicillin-resistant Staphylococcus with a post-antibiotic effect comparable to that of Telithromycin [209] [210]. Furthermore, its activity on Mycobacterium avium is remarkable [208]. After a single oral dose to healthy volunteers, plasma concentrations were 0.2 mg/L and 1mg/L at 100 mg and 1200 mg doses respectively after an average of 3 hours. The apparent elimination half-life was approximately 15 hours. Modithromycin is currently recommended for the treatment of respiratory tract infections [211] [212]. Other 6,11-bicyclonides analogues of Modithromycin (Figure 32) have also been developed [213]. The modifications made consisted in replacing the C9 acetylimine group with the initial carbonyl (compound 49) or else in replacing acetylimine with another of a propionylimine nature (compound 50). At the level of the heteroaryl pyridinyl-pyrazole of Modithromycin, this has also been replaced by other heteroaryl fractions of pyridinyl-aminothiazole nature (compound 49) still thiazolyl-aminopyridine (compound 50) or pyridinyl-isoxazole (compound 51). Furthermore, by analogy with the interesting antibacterial effects stimulated by the presence of fluorine, we are also witnessing such an introduction of fluorine into C2 like Solithromycin (compound 50). All of these new bicyclonides have shown potent activities against H. influenzae strains. Compound 50 with a faction of thiazolyl-aminopyridine-like heterocycles, a C9 propionylimine and a C2 fluorine has risen like Modithromycin, active against the methicillin-resistant S. aureus [213].

2) Tricyclic ketolides or Tricyclonides

In pursuit of the quest for more effective macrolides against pathogenic and resistant respiratory germs, a new series of ketolides characterized by a 6-O-methyltricyclic skeleton bridged with C9-C11-C12 has been developed (Figure 33). These ketolides or Tricyclonides were prepared from Clarithromycin intermediates following a series of reactions introducing ethylenediamine, the formation of C11 and C12 carbamate followed by an intramolecular formation of imine and the construction of a 9,11-diazaheptene cycle [214] [215] [216]. The first derivatives of this series differed in the presence or absence of methyl substituents in alpha (C17) of the imine of the 9,11-diazaheptene ring. Thus, four derivatives, namely the compounds TE-802, TE-935, TE-943 and TE-806 were evaluated for their antibacterial activities (Figure 33). These note that these new ketolides are effective in vitro on strains sensitive to Erythromycin, on S. aureus resistant to Erythromycin, on certain Gram (−) bacteria, anaerobic bacteria and Mycoplasma pneumoniae. In particular, the compound not methylated on diazaheptene ring or compound TE-802, has proved to be the most effective and has exhibited activities comparable to or even superior to those of Clarithromycin and Azithromycin against the Gram (+). TE-802 was also effective against certain strains of S. pneumoniae resistant to erythromycin. In terms of pharmacokinetics, TE-802 presented in vivo in rats, good stability in gastric acid medium, a long half-life and better tissue penetration. However, all these 6-O-methyltricyclic ketolides did not show any effective activity against S. pneumoniae constitutionally resistant to MLSB [214] [215] [216].

Faced with the low activity of TE-802 and its methylated derivatives against Streptococcus pneumoniae resistant to Erythromycin by an MLSB mechanism and on Haemophilus influenzae, two types of chemical variation were operated in order to reinforce the activity antibacterial of new TE-802 derivatives on these bacteria. The first modulation consisted in introducing on the C2 of TE-802 compound, a fluorine atom as in the case of Solithromycin [217]. The analogue 2-fluoro-TE-802 or compound 52 obtained (Figure 33) effectively presented an improved antibacterial and pharmacokinetic profile. Indeed, the presence of fluorine widened the spectrum of action in vivo in the rat of this analogue of TE-802 to H. influenzae. Furthermore, the other analogues C2-substituted by hydroxyl or alkyl groups or even chlorine or bromine atoms, showed no improvement in the expected activities [217]. The second modification consisted to introducing alkoxy-aryl substituents in alpha position (C17) or in beta position (C16) of the imine of 9,11-diazaheptene ring of TE-802 compound (Figure 33). In general, the derivatives resulting from this modification of TE-802 have shown improved antibacterial activities, in particular on Streptococci resistant by an MLSB mechanism and on Haemophilus influenzae [218]. In addition, SAR studies show that the spatial and isomeric arrangements of alkoxy-aryls on the diazaheptene ring influence activity. Derivatives with C16 substitution carrying a quinoline have better in vitro activity against MLSB resistant Streptococci and H. influenzae. Conversely, the compounds resulting from the substitution for C17 would have better in vivo efficacy. These are indeed 2 to 3 times more effective

than Clarithromycin even if they have a performance comparable to that of TE-802. Among these substituted C17s, the bromobenzyloxy-methyl analogue (compound 53) was found to be the most effective (Figure 33). Substituted C17 derivatives or alpha derivatives of imine appear to preserve the antibacterial properties of TE-802 [218] [219].