Seventh Addendum to Research and License Agreement by and between Technion Research and Development Foundation Ltd. and Eloxx Pharmaceuticals Ltd., dated March 6, 2016

Prof. Timor Baasov

of our most recently developed leads (Eloxx funded second year’s research) NB156-NB158, along with the proposed new structures NB159-NB165. These new structures are actually the derivatives of NB158. NB158 serves as the scaffold to which three already established pharmacophores including 6’-(R)-Me, 5”-(S)-Me and N1-AHB groups are sequentially introduced: the only one pharmacophore (compounds NB159, NB160 and NB161), two pharmacophores (compounds NB162, NB163 and NB164) and all three pharmacophores (NB165). The rationale in selecting NB159-NB165 as potential new leads is based on the following our past observations. First, NB158 exhibits similar to better activity to that of NB30, while its eukaryotic and prokaryotic translation inhibition are significantly lower to that of NB30. These observations suggest that NB158 may serve better scaffold than NB30 for further development. This expectation is supported by: (i) while the eukaryotic inhibition of translation of NB158 is only about two-fold poorer than that of NB30 (IC₅₀^euk values of 70 and 31 µM, respectively) this gap in the prokaryotic inhibition of translation is increased up two orders of magnitude (IC₅₀^prok values of 36.45 and 0.45 µM, respectively), suggesting significantly increased specificity and selectivity of NB158 towards eukaryotic versus prokaryotic ribosome than that of NB30 (the IC₅₀^prok/IC₅₀^euk values of 1.9 and 69 for NB158 and NB30; see Table 1 of the 2^nd year’s research report). (ii) Since our previous data indicated that the reduction in prokaryotic ribosome specificity can be correlated with the reduction of mitochondrial ribosome specificity, the data suggest that the NB158 and its follow-up derivatives NB159-NB165 probably will enjoy with reduced mitochondrial inhibition and subsequently with reduced toxic side-effects. (iii) Once the synthesis of NB158 has already established in our lab, the generation and subsequent biological tests of the suggested new derivatives NB159-NB165 should be a relatively easy task.

The synthesis and evaluation of the suggested new structures NB159-NB165 will be done according to our second year’s research report for the synthesis of NB158 and our previous reports for the introduction of the 6’-(R)-Me, 5”-(S)-Me and N1-AHB groups on the previously developed NBs.

Note: one can argue that since NB156 showed better activity than its parent NB74, why we will not pursue this structure for further modification (e.g. attachment of N1-AHB) and development as the potential drug. Three arguments: (1) NB156 along with NB157 and NB158 are currently under the cochleotoxicity tests in Schachts’ lab; the observed data will help us to see the potential therapeutic index of these compounds versus their parent NBs. If the data will found to be satisfactory, we will then attempt to continue their further development as well. (2) Because the lack of important data on the rescue of the functional protein in various disease models, we are largely hindered to make proposal in this direction (problems to obtain MTA from other collaborative groups). (3) Finally, since the establishment of the synthesis scheme towards the introduction of chiral 6’,7’-diol in these series of compounds took us much efforts, and the current scheme includes rather several complicated synthetic steps, along with the highly increased cytotoxicity of the NB157 versus NB124, I decided to put those further perspectives in hold and promote those projects and compounds such as NB159-NB165 (Fig. 1) that I believe and fill that may open us new opportunities and discoveries.

Aim #2a – Modification of the developed NBs at N1-position for the production of non-ototoxic lead compounds: The major limiting side effects of aminoglycoside antibiotics is permanent hearing impairment, which is detected in approximately 20% of treated patients using conventional audiometry. Even though, our developed lead NB-compounds were proved in exhibiting significantly reduced ototoxicity potential, this toxicity issue is still somewhat problematic and introduces “red light mark” for all aminoglycosides for clinical development as a drug. This issue is especially more relevant when the aminoglycoside-based drug is used for the treatment of genetic diseases; because the compound must be administered to the patients for the lifelong. Therefore, the continuous attempts towards new designer structures towards eliminating their ototoxic effects, while preserving their potent readthrough activity, should be of our top priority.

Towards these ends I suggest here to apply those most recent developments in this field that provided important and interested results. One of such most recent approaches considers the modification of aminoglycosides so to reduce the ability of the drug to entry into hair cells via mechanotransducer (MET) channels (Fig 2A) and as such reducing its ototoxic effects.¹ In this work, nine derivatives of the aminoglycoside antibiotic sisomicin have been synthesized by modification of the parent drug at N1 position (ring II), at N3’’ position (ring III) and at both N1 and N3”. The modifications included acylation of the amine moiety(s) either by different simple acyl groups (e.g. acetate or benzoate) or by alkyl- and aryl sulfonyl groups (e.g. methylsulfonyl, arylsulfonyl). It was hypothesized that since the resulted N-acyl derivatives of sisomicin

5

Prof. Timor Baasov

exhibit one less positive charge, the compounds can still bind to their ribosomal target with sufficient affinity to exert the required antibacterial activity, while in parallel their entry to the cochlear cells will be significantly reduced due to inhibition of MET channels. Interestingly, all 9 derivatives were displayed no or reduced ototoxicity and the lead compound N1MS was found 17 times less ototoxic and with significantly reduced penetration of hair cell MET channels in rat cochlear cultures (Fig 2). Moreover, systemic sisomicin treatment of mice resulted in 75% to 85% hair cell loss and profound hearing loss, whereas N1MS treatment preserved both hair cells and hearing. These are indeed excellent observations! Note that they did not test whether the observed reduced ototoxicity of N1MS is primarily due its reduced entry into the mitochondria, or may because the reduced inhibition of mitochondrial translation. Nevertheless the observed data with N1MS is very important and definitely provides support to the concept that the targeted modification of aminoglycoside so to inhibit its entry into the mitochondria through the mechanotransducer channels provides important new tool for the generation of nonototoxic aminoglycosides.

Based on these observations, the first question that I asked was: why we will not do the similar modifications on our NBs as well? After all, this tool can definitely be supplementary way to further decrease the ototoxicity potential of our NBs, which should be our most important priority. Of course, the question is how much damage such modification can cause to the readthrough potency of NBs? Note that 3 out 9 compounds synthesized and tested by Hurt et al. exhibited comparable anti E. coli activities to the parent compound sisomicin.¹ This is an excellent observation and encouraged us to pursue this strategy towards synthesis of new derivatives of NBs with diminished ototoxicity.

Initially we propose to synthesize the compounds of set1 and set2 structures (Fig. 3). The only difference between these two sets of compounds is the presence of chiral 5’’-Me group in set2 compounds while this is lack in set1 compounds. The rationale in choosing these particular structures is that all these 10 different compounds can be easily accessed from the known intermediate compound A as illustrated in Scheme 1. Selective N-acylation of the compound A with different acylating groups followed by selective acetylation will afford a series of fife different acceptors B-F. Glycosylation of each of these acceptors separately by either the known trichloroacetimidate (TCAI) donor G or H followed by sequential two deprotecting steps will afford the target library of 10 different compounds of set1 and set2. The final products will then be subjected for the comparative readthrough and toxicity tests as we performed in our earlier studies. Compounds of set1 are the N1-acyl derivatives of NB74 and the compounds of set2 are similar derivatives of NB124. Therefore, NB74 and NB124 will used as the parent compounds of the comparative studies of these sets of compounds. The compounds with good readthrough activity and low cytotoxicity will then be subjected for the initial cochleotoxicity tests in the laboratory of Prof. Jochen Schacht at the University of Michigan (?).

Prof. Timor Baasov

As the preliminary data in this direction (Master student is working on the project already!) we have prepared multigram quantities of compound A (Scheme 1) and successfully assembled selected examples of the acceptors B-F in small quantities for the tuning of the these crucial steps. Further coupling and deprotection steps are under current investigation.

Finally, since the lead compound N1MS (Fig. 2)¹ is the derivative of sisomicn that has very unique ring I (unsaturated ring), one can argue that structurally this ring may has some critical influence; the combination of ring-I structure with N1-acylation is critical for the efficient inhibition of MET channels by N1MS. This dilemma is supported by the fact that until now except sisomacin no other aminoglycosides of this class (4,6-disubstituted) or of the 4’,5’-disubstituted class have been reported that their N1-modification could lead (or not) to the compounds with reduced ototoxicity. To test this hypothesis, we propose to assemble additional set of 3 compounds of set3 (Fig 3). All these compounds exhibit similar unsaturated ring I of sisomicin (and of N1MS) and contain the most favorable N1-methylsulfonyl group as in the reported lead N1MS. Thus, the likelihood of set3 compounds in exhibiting the desirable reduced ototoxicity is very high. Compound 11 can be easily accessed from the intermediate compound from the synthesis of our developed NB158 (Fig. 1). Similarly, the compounds 12 and 13 will be accessed from the intermediate pseudo-disaccharide compounds of the synthesis of NB159 and NB162, respectively.

Aim #2b – Modification of the developed NBs at C4’-position for the production of non-ototoxic lead compounds:

The second approach that recently been proved as a valid strategy towards developing new aminoglycoside antibiotics with reduced ototoxic potential is based on the works of Bottger and coworkers.^2,3 In this these works the authors demonstrate that simple alkylation of paromomycin at 4’-OH results new derivatives with similar or little reduced antibacterial activity but with significant reduction in their action on the mammalian cytoplasmic and mitochondrial ribosomes and subsequently with significant reduction in their ototoxicity potential. These works (and the follow-up publications) especially highlighted two new derivatives of paromomycin Par¹ and Par² (Fig. 5)³ with excellent selectivity at the ribosomal target, promising antibacterial activity, and little, if any, ototoxicity upon chronic administration (note that in vivo ototoxicity of these lead compounds were tested by our collaborator J. Schacht who is also on the paper!). I would like to further note that even though they also resolved x-ray crystal structures of several new derivatives in complex with the bacterial ribosome, still they could not provide important structural information why these compounds exhibit such a large reduction in their ototoxicity potential. And the main reasons mentioned and tested were the selectivity increase of these compounds towards prokaryotic versus mammalian cytoplasmic and mitochondrial ribosomes. They determined the ribosomal damage index (Fig. 5) and found this damage are especially lower towards mitochondrial and cytoplasmic ribosomes for Par1 and Par2 in comparison to the parent paromomycin. Comparative in vivo ototoxicity test in guinea pigs showed that at 400 mg/kg dose compound Par1 has only little and Par2 no outer hair cell loss.

7

Prof. Timor Baasov

Based on these observations, we suggest testing similar modifications on our NBs with the aim to gain new derivatives with significantly reduced ototoxicity. Towards these ends we suggest to synthesize the new set4 structures (Fig. 6). Initially, we will use the NB74 as the pseudo-trisaccharide scaffold and by introducing 4’-O-Ethyl, 4’-O-Propyl and 4’-O-Benzyl groups we will generate the target compounds 14-16 as potential leads with substantial read-through activity and reduced ototoxicity. The acceptors used for the preparation of these compounds will also serve us for the coupling with the second donor exhibiting the chiral methyl group to generate the desired second set of compounds, 17-19 (Fig. 6). The proposed general synthetic scheme towards the assembly of set4 structures is illustrated in Scheme 2.

We will use G418 as starting material to assemble the intermediate disaccharide A (Scheme 2). The rational is to introduce here the 4’,6’-benzylidene protection which can selectively opened by reductive opening to the 4’-OH intermediate. This intermediate then can be easily alkylated by attaching the desired 4’-ethyl and 4’-propyl alkyl groups selectively at 4’-position to afford the compounds B and C. Note that the intermediate A cannot be used for the synthesis of other, desired 4’-benzyl derivatives (compounds 16 and 19, Fig. 6); since the other hydroxyls in A have already benzyl groups as protection, its selective removal from the 4’ position cannot be done. Thus, for the preparation of 16 and 19 we should choose other groups that can be easily removed in the presence of benzyl ether! The acceptors B and C will then be separately subjected to the glycosylation step by two different donors D and E, followed by the de-protection steps to afford the desired structures 14-15 and 17-18.

The final products will then be subjected for the comparative readthrough and toxicity tests as we performed in our earlier studies. Compounds of 14-15 are 4’-alkyl derivatives of NB74 and the compounds 17-18 are similar derivatives of NB124. Therefore, NB74 and NB124 will be used as the parent compounds of the comparative studies of these sets of compounds. The compounds with good readthrough activity and low cytotoxicity will then be subjected for the initial cochleotoxicity tests in the laboratory of Prof. Jochen Schacht at the University of Michigan (?).

Finally, it is of note that while our previous data suggests that the ex-vivo tests of cochleotoxicity well correlated to the in vivo ototoxicity data, still this correlation should be considered as a preliminary observation and for the proper ototoxicity evaluation we will need in vivo studies to establish this issue appropriately. For this reason, the compounds that will show significant reduction in their initial cochleotoxicity tests, will further subjected for the in vivo tests.

REFERENCES

8

Prof. Timor Baasov

III. Proposed Budget (in US Dollars) and Time Frame

NOTE that the budget can only allow selected structures synthesis and evaluation and also does not include ex vivo studies (functional protein synthesis in MEFs for example) and in vivo studies required for the project completion and only can be done with additional funding of collaborators labs!

IV. Detailed Budget (in US Dollars)

Personnel:

Supplies:

9

Prof. Timor Baasov

Budget Justification:

The requests for laboratory assistant and postdoctorant for the duration of the grant period are in recognition of the amount of work required in this project. Considerable effort will be expended in the syntheses of various lead compounds discussed in the proposal, their structure determination and analysis, assays for their activity. The request for materials, supplies, and chemicals/biochemicals is an important part for a successful development and completion of the project.

V. Investigators’ Curriculum Vitae

Surname:                Baasov                 First name:    Timor

Birthdate:        January 3, 1954

(a) Education Background

Major research interest: Carbohydrate chemistry, Bioorganic and medicinal chemistry, Drug design and development, Rational design of substrate and inhibitors, Mechanistic enzymology.

(b) Employment

VI. T. Baasov - List of Publications last three years (2012-2015)

10

Prof. Timor Baasov

Chapters in Books and other Publications

11