Impact of physicochemical properties on dose and hepatotoxicity of

Impact of physicochemical properties on dose and hepatotoxicity of oral drugs. Paul D Leeson. Chem. Res. Toxicol. , Just Accepted Manuscript. DOI: 10...
2 downloads 3 Views 1MB Size
Subscriber access provided by Kaohsiung Medical University

Article

Impact of physicochemical properties on dose and hepatotoxicity of oral drugs Paul D Leeson Chem. Res. Toxicol., Just Accepted Manuscript • DOI: 10.1021/acs.chemrestox.8b00044 • Publication Date (Web): 03 May 2018 Downloaded from http://pubs.acs.org on May 3, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

Chemical Research in Toxicology

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

    Impact of physicochemical properties on dose and hepatotoxicity of oral drugs   Paul D Leeson  Paul Leeson Consulting Ltd, The Malt House, Main Street, Congerstone, Nuneaton, Warks, CV13 6LZ, UK  [email protected]  ORCID: 0000‐0003‐0212‐3437   

 

ACS Paragon Plus Environment

Chemical Research in Toxicology

Page 2 of 27

          Table of contents graphic   

DILI model using dose, Fsp3, cLogP

Oral drugs

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

140 120 100 80 60 40 20 0

Most DILI (n=163)

No DILI (n=163)

Predicted No DILI Predicted Most DILI

ACS Paragon Plus Environment

 

 

Page 3 of 27

Chemical Research in Toxicology

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract.  A database containing maximum daily doses of 1841 marketed oral drugs was used to examine  the influence of physicochemical properties on dose and hepatotoxicity (drug induced liver injury, DILI).   Drugs in the highest ~20% dose range had significantly reduced mean lipophilicity and molecular weight,  increased fractional surface area, increased % of acids and decreased % of bases, versus drugs in the lower  ~60% dose range.  Drugs in the ~20‐40% dose range had intermediate mean properties, similar to the mean  values for the full drug set.  Drugs that are both large and highly lipophilic almost invariably do not have  doses in the upper ~20% range.  The results show that oral druglike physicochemical properties are  different according to these dose ranges, and this is consistent with maintenance of acceptable safety  profiles as efficacious exposure increases.  Verified DILI annotations from a compilation of >1000 approved  drugs (Chen et al, Drug Discov. Today, 2016, 21, 648) were used.  The drugs classified as No DILI (n=163)  had significantly lower dose and lipophilicity, and higher Fsp3 (fraction of carbon atoms that are sp3  hybridised) versus the Most DILI (n=163) drugs.  The % of acids was reduced, and bases increased, in the No  DILI versus the Most DILI groups.  Drugs classified as Less DILI or Ambiguous DILI had intermediate mean  values of dose, lipophilicity, Fsp3 and % acids and bases.  The impact of lipophilicity and Fsp3 on DILI  increases in the upper 20% versus the lower 80% dose range, and a simple decision tree model predicted  No versus Most DILI outcomes with 82% accuracy.  The model correctly classified 19 of 22 drugs (86%) that  failed in development due to human hepatotoxicity.  Because many oral drugs lacking DILI annotations are  predicted to be Most DILI, the model is best used preclinically in conjunction with experimental DILI  mitigation.   

 

ACS Paragon Plus Environment

Chemical Research in Toxicology Introduction.  The concept that increasing drug dose and exposure increases risk of toxicity, especially  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

idiosyncratic toxicity, is well documented.1,2  Daily dose can be viewed as the ultimate composite or  multiparameter property of an oral drug, since it relies on the effective therapeutic concentration, the  required target occupancy over time, absorption, clearance, volume of distribution, and dosing frequency.3   Absorption in turn is dependent on permeability and solubility.  A combination of dose, solubility and  lipophilicity criteria has been recommended to guide selection of high quality oral drug candidates.4   Physicochemical properties are known to influence each of the parameters affecting dose, and free drug  concentrations are inversely related to lipophilicity.5,6 This is important to consider if both dose and  physical properties are to be used together in structure‐toxicity studies.  An example is hepatotoxicity or  drug induced liver injury (DILI), the most common form of idiosyncratic toxicity in humans.  DILI is a  complex, rare, multifactorial event, occurring after weeks or months of treatment; drug exposure and  formation of reactive metabolites are generally accepted as key factors.7,8,9,10  The ‘rule of two,’ derived  from 164 oral drugs, proposed that high daily dose (>100 mg) combined with high lipophilicity (cLogP >3)  increases DILI risk.11  This could be useful guidance but might be misleading because the dataset is small  (100  mg, but not lipophilicity, was associated with various human hepatic adverse effects among a group of 975  oral drugs.13  In contrast, hepatoxicity models have been developed using the rule of two in combination  with reactive metabolite formation14 and hepatic metabolism.15  High DILI risk is associated with  Biopharmaceutics Drug Disposition Classification System (BDDCS) class 2 drugs,16 which are poorly soluble  but highly permeable and generally highly lipophilic.  Topological molecular properties17,18 and sub‐ structural chemical features19 could also be linked to hepatoxicity.  Quantitative structure‐hepatotoxicity  relationships,17,19,20,21,22 developed using differing datasets, provide better performance than the rule of  two, but are not easy to understand on a physicochemical basis.  Clearly a major problem hampering  analysis of DILI has been the use of datasets based on differing DILI assessments;23.24 the need for a large,  well‐curated and consistently accurate database was recently addressed by combining annotations from  FDA drug labelling with human causality information.25 The resulting ‘verified’ dataset for severity of DILI  annotated >1000 approved drugs as No DILI, Most DILI, Less DILI, and Ambiguous DILI.22,25  ACS Paragon Plus Environment

Page 4 of 27

Page 5 of 27

Chemical Research in Toxicology In this study, we have assembled an oral drug dose database containing >1800 marketed oral drugs.  We 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

show that 1) there are meaningful trends between dose and oral drug properties including lipophilicity,  size, and ionisation state; and 2) daily dose, in combination with the fraction of carbon atoms that are sp3‐ hybridised (Fsp3) and lipophilicity (cLogP) can distinguish the verified Most DILI from No DILI categories  with >80% accuracy.  Oral drug dose database.  An oral drugs database26 was updated to 2016 approvals and annotated with  maximum daily dose (MDD, mg), the highest reported total dose per day, obtained from literature1,9,20,27  and online28 sources.  A total of 1841 oral drugs were assigned an MDD value.  The defined daily dose  (DDD), ‘the assumed average maintenance dose per day for a drug used for its main indication in adults’29 is  useful single‐source comparative dataset13 and a total of 1261 oral drugs were assigned a DDD value.   Although drug doses by weight, in milligrams, are almost universally used, the application of molar doses,  instead of doses by weight, is best practice for quantitative structure‐activity studies.30  Daily dose values  were therefore converted to the logarithm of the molar doses, ‐p[MDD or DDD] = Log10 ((Dose (mg)/Mol  Wt)/1000), adjusting for salts where necessary.  Increasing values of ‐p[MDD or DDD] reflect increasing  dose size.  Standard physicochemical properties of the oral drugs were calculated.31  Relationships between physical properties and oral dose.  Mean and quantile DDD values are overall less  than two‐fold lower than MDD values (Table 1) and both sets showed similar dose versus property trends.   The MDD values are used in all analyses.  It is interesting to note that mean –p[MDD] values have reduced  over time but have been constant for the past 4 decades; in contrast cLogP and molecular weight have  increased significantly over the same period32 (supplementary Figure S1).  Hence oral drug molecular  inflation33 is not on average associated with changes to daily dose.   Dose  value  Median  Mean  10%,  90% 

MDD n = 1841  ‐p[MDD]  mg  ‐3.23  200  ‐3.68  853  ‐4.75,  7.5,  ‐2.11  2000 

DDD n = 1261  ‐p[DDD]  mg  ‐3.48  100  ‐3.59  542  ‐4.97,  4.1,  ‐2.30  1500 

 

Table 1.  Mean and quantile dose data for oral drug data sets.  MDD is maximum daily dose; DDD is  defined daily dose from the World Health Organisation.  ‐p[MDD or DDD] = Log10 ((Dose (mg)/Mol  Wt)/1000).    ACS Paragon Plus Environment

Chemical Research in Toxicology The intention here is not to establish structure‐activity relationships for oral dose, a challenging  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

proposition,27 but to examine how druglike physicochemical properties vary, if at all, by dose.  In this sense,  the properties, not the dose, are the dependent variables.  Scatter plots of physical properties (y‐axis)  versus ‐p[MDD] (x‐axis) were therefore assembled and trends in the complex data visualised and simplified  by using moving average plots and 5% or larger dose bins, as shown in Figure 1 for lipophilicity (cLogP).   There is no impact of dose on cLogP when the ‐p[dose] is  ~‐2.5 (the DDD set behaves similarly).  Mean  cLogP in the 5% dose bins follows the moving average (Figure 1b; the binned cLogP distributions are shown  in Figure 1c).  The downwards trajectory of lipophilicity with increasing dose affects significant numbers of  drugs: ‐p[MDD] values above ‐3 and ‐2.5 comprise 40% and 20% of the total respectively (Figure 1d).  The  general pattern seen with cLogP versus –p[MDD] is recapitulated with all other measures of lipophilicity  examined (supplementary Table S1), including cLogDpH7.4 and the property forecast index34 (PFI =  chromatographic cLogDph7.4 plus number of aromatic rings).  [Figure 1 here]  In Figure 2a, the 5% dose bin mean properties for cLogP (Figure 1c) are reproduced and Figures 2b‐i show  the same 5% dose bin analysis for several other physicochemical properties, each referenced to the mean  property value for the whole set.  Molecular weight (Figure 2b) is highest in the lowest 5 percentile dose  group and decreases below the average in the highest 15 percent dose range.  Total polar surface area  shows a complex pattern (Figure 2c) but when normalised for size using heavy atom count (Figure 2f) the  pattern is consistent with the lipophilicity trend, showing increasing polar fraction in the highest 20% dose  range.  Aromatic ring count is reduced at both the lowest 5% and highest 20% of doses (Figure 1d), while  Fsp335 shows little change until the lowest 5% dose range, where it is increased (Figure 1e).  It is notable  that in the upper 20% dose range the proportion of drugs that are acids increases above the average, while  bases are reduced (Figures 2g and 2h); there is a complex pattern in the distribution of neutral drugs by  dose (Figure 2i).  [Figure 2 here]  ACS Paragon Plus Environment

Page 6 of 27

Page 7 of 27

Chemical Research in Toxicology In Figure 2, each of the property distributions appears near to their average at the upper 20 percentile dose 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

value.  Using relative property values in each 5% bin, as % changes from the average value, shows that oral  drugs in the ~20‐40 percentile dose range have mean property values close to the overall averages, unlike  the upper 20% and lower 60% ranges where there is also greater variability (Figure 3).   The data in Figure 3  show that the ~20‐40% dose range represents a cross‐over region for those properties that change as dose  increases.  These percentile dose ranges are not hard cut‐offs but approximate values based on visual  inspection of the dose‐property trends in Figures 1‐3; it should be noted there is considerable property  variability in each of these dose ranges (Figures 1c and 1d).  The mean physical properties of the upper 20%  versus the 20‐40% and lower 60% dose ranges is listed in Supplementary Table S1.  Only mean O+N count  and chiral atom count show no differences between the dose ranges, but because molecular weight is  reduced in the upper 20% dose range, these (and other) properties show dose‐dependent differences when  normalised for size (supplementary Table S1).  [Figure 3 here]  The compression of lipophilicity and molecular weight in the upper 20% dose range has practical  consequences relevant to selection of drug candidates.  Of the 368 oral drugs with upper 20% dose values,  only 9 (2.4%) have both cLogP >4 and molecular weight >400, where ADMET risk is increased  (supplementary Figure S2).36  Of the 9 drugs, three are iodinated radiocontrast agents, three are prodrugs  and only two, nelfinavir and vemurafenib, were discovered in the last 40 years.  The cLogP >4 and  molecular weight >400 region is a popular chemical space for current drug discovery, since it is the most  populated of the four possible cLogP/molecular weight categories in the recent patent literature37 (45% of  compounds, supplementary Figure S2).  These observations suggest that candidates having these  properties together with ‐p[MDD] values >‐2.5 (the upper 20 percentile value) are highly unlikely to  succeed.  Hence avoiding doses in the upper 20% range is critical for success in oral candidates possessing  more extreme physical properties.  Hepatotoxicity.  Of the 1036 drugs given a ‘verified’ DILI annotation in humans,25 731 were assigned oral  MDD values (Table 2).  The remainder are biologicals, injectables, topicals and seven oral clinical candidates  that were not marketed due to human hepatotoxicity.  These seven compounds were combined with other  candidates that failed because of hepatotoxicity to provide a set of 22 compounds useful for testing  ACS Paragon Plus Environment

Chemical Research in Toxicology physiochemical models (see later).  The 731 drugs having a DILI annotation are about evenly distributed  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

between the four categories, leaving 1110 drugs as a reference set lacking DILI annotation (Table 2).  The  most commonly prescribed 200 drugs (Top 200)38 contain 155 oral drugs, of which 17 (11%) are classified as  No DILI and 16 (10%) most DILI (Table 2).  It is interesting to note that 59% of the most prescribed oral  drugs are classified as Less DILI, suggesting that this designation has not significantly restricted therapeutic  application.  The No DILI group contain older drugs than the other groups (Table 2), perhaps because it  takes longer to identify unambiguously ‘clean’ molecules.  Supplementary Spreadsheet S1 contains the DILI  annotated oral drugs and their doses. 

Oral drug  category  No DILI  Most DILI  Less DILI  Ambiguous DILI  No data 

n,  MDD  163  163  232  173  1110 

n, % of Top  200 (155 orals)  17, 11%  16, 10%  91, 59%  22, 14%  9, 6% 

Median year of  publication  1955  1969  1971  1969  1966 

  Table 2.  Distribution of verified DILI assignments and oral drugs in the Top 200 prescribed  medicines, together with median year of first drug publication for each group.  Note the No DILI  drugs are appreciably older than the other groups.   The physical property distributions of the five groups in Table 2 were examined in detail and the key results  are shown in Figure 4.  Four properties significantly distinguished the No DILI versus Most DILI groups: dose  (‐p[MDD]), lipophilicity (cLogP), Fsp3, and ionisation class.  The No DILI group had the lowest mean ‐p[MDD]  and cLogP values, and the highest mean Fsp3 value.  Although the mid‐50% quantile property ranges of all  groups overlap, it is notable that the Less DILI, Ambiguous DILI and No data sets showed no differences in  these properties, yet their mean values are intermediate between the No and Most DILI groups.  There was  varying statistical significance in the differences between the Less DILI and Ambiguous DILI groups versus  the No and Most DILI groups (Figure 4).  The ion class observations are striking: acids comprise 9% and  bases 54% of the No DILI group compared with 27% and 25% respectively in the Most DILI group (Figure 4).   Other properties that were significantly different between the No DILI and Most DILI groups were:  alternative measures of lipophilicity (LogD7.5, LogD6.5, LogP (all Chemaxon) and PFI), aromatic ring count,  sp2 atom count and sp2‐sp3 atom count (Most DILI greater in all cases); and sp3 and chiral atom counts  ACS Paragon Plus Environment

Page 8 of 27

Page 9 of 27

Chemical Research in Toxicology (Most DILI lower).  Properties that were not different across all DILI groups were molecular weight, heavy 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

atom count, rotatable bond count, hydrogen bond donors and acceptors, and total polar surface area  (supplementary Table S2).  [Figure 4 here]  Since the upper 20% daily dose range of drugs show changed properties versus lower doses (Figure 3), the  impact of cLogP and Fsp3 in the upper 20% versus lower 80% doses was examined for the No and Most DILI  classes (Figure 5).  As expected from Figure 2, amongst the combined set of No and Most DILI drugs, cLogP  was lower in the upper 20% dose range versus the lower 80%, but the cLogP difference between No and  Most DILI was increased by 2 units in the higher dose group (Figure 5a).  Again, in agreement with Figure 2,  the combined set of No and Most DILI drugs did not differ in mean Fsp3, but the difference between No  and most DILI is significantly increased by 0.23 in the higher dose group (Figure 5b).  The results show that  cLogP and Fsp3 become increasingly important in influencing hepatotoxicity as dose is increased.  [Figure 5 here]  Mean  properties  (std. dev.) ‐p[MDD] cLogP Fsp3  

Acids

Bases

Neutrals

No DILI n=15

Most DILI n=44

No DILI n=88

Most DILI n=41

No DILI n=48

Most DILI n=72

‐3.30 (1.87) ‐0.4 (2.58) 0.54 (0.28)

‐2.70a (0.78) 3.7b (2.06) 0.2b (0.20)

‐3.82 (0.89) 2.2 (0.28) 0.51 (0.22)

‐3.09b (0.53) 3.5b (0.42) 0.44a (0.24)

‐3.56 (1.52) 2.2 (3.25) 0.50 (0.27)

‐2.94b (0.63) 2.5a (2.47) 0.30b (0.25)

Table 3.  Mean properties and standard deviations of No and Most DILI drugs according to ion class.   a

 No difference between No DILI and Most DILI.  b No DILI and Most DILI are significantly different (p 

3, dose >100mg; No DILI:  cLogP3 quartiles from the  median.  d) Box and whisker plots of the distributions of cLogP in the lower 60% range of doses, the  20‐40% range and the upper 20%, these cutoffs coming from inspection of the trends shown in a‐c.  In  c and d, categories connected by the same letter are not different (p >0.05, Tukey HSD).      ACS Paragon Plus Environment

Page 21 of 27

Chemical Research in Toxicology

 

a

450

100

425

95

3

400

90

375

85

2

2.48

1.5 1 0.5 0

Mean TPSA

4

Mean Mol Wt

Mean cLogP

c

3.5 2.5

350 325

349

300 275

75 70 65 55

20%

200 ‐7

‐6

‐5

‐4

‐3

‐2

‐1

d

‐6

‐5

‐4

‐3

‐2

‐1

5 percentile mean ‐p[MDD]

e

2.1

‐7

1.3

0.475 0.45 0.425 0.4

1.1

20% 0.9 ‐7

‐6

‐5

‐4

‐3

‐2

g

0.423

4 3.5

‐7

‐6

‐5

‐4

‐3

‐2

i

70

60

35

60

55

10

16%

40 30 20

20% 0 ‐7

‐6

‐5

‐4

‐3

‐2

5 percentile mean ‐p[MDD]

‐1

‐3

‐2

‐1

45

37.3%

40 35

25

20%

0

‐4

30

10

5

‐5

50

41%

% Neutral

% Base

15

‐6

5 percentile mean ‐p[MDD]

40

20

20% ‐7

‐1

5 percentile mean ‐p[MDD]

3.28

3 2

50

‐1

5

2.5

h

25

‐2

4.5

0.35

30

‐3

5.5

0.375 ‐1

5 percentile mean ‐p[MDD]

‐4

6

Mean TPSA/HA

Mean FSp3

1.5 1.61

‐5

6.5

20%

0.5 1.7

‐6

5 percentile mean ‐p[MDD]

f

0.55 0.525

1.9

20%

50 ‐7

5 percentile mean ‐p[MDD]

75.4

60

225

20%

‐1

Mean Ar ring

80

250

‐0.5

% Acid

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

b

20%

20 ‐7

‐6

‐5

‐4

‐3

‐2

5 percentile mean ‐p[MDD]

‐1

‐7

‐6

‐5

‐4

‐3

‐2

5 percentile mean ‐p[MDD]

  Figure 2.  Property means of 5 percentile dose bins versus mean –p[MDD]. a) cLogP (data taken from  Figure 1c).  b) Molecular weight. c) Total polar surface area. d) Aromatic ring count.  e) Fsp3 (fraction  of carbon atoms that are sp3 hybridised). f) Total polar surface area divided by heavy atom count.  g) %  Acids (negatively charged at physiological pH). h) % Bases (positively charged at physiological pH). i) %  Neutrals (uncharged at physiological pH).  Mean ‐p[MDD] values for acids, bases, neutrals and  zwitterions are respectively ‐2.91, ‐3.59, ‐3.36, ‐3.09.  Acids have significantly higher doses than neutral  or bases (p >0.05, Tukey HSD).  The mean values for each property are shown in red and referenced to  the 20 percentile upper dose range in blue.           ACS Paragon Plus Environment

‐1

 

Chemical Research in Toxicology

Page 22 of 27

 

40

Fsp3

20

cLogP Mol Wt

% Deviation from mean 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

0

Ar ring ‐20

TPSA

‐40

Acid

‐60

Base Neutral

‐80

0% deviation ‐100

20% ‐p{MDD]

20%

40%

40% ‐p[MDD]

‐120 ‐6.5

‐6

‐5.5

‐5

‐4.5

‐4

‐3.5

‐3

‐2.5

‐2

‐1.5

‐1

5 percentile mean ‐p[MDD]     Figure 3.  The percent deviation from the mean for each property in the 5 percentile dose bins,  versus –p[MDD], for eight molecular properties shown in Figure 2.  By visual inspection,  properties are close to the overall means in the ~20‐40 percentile dose range, which represents a  cross‐over region for those properties which are changing as dose increases.                              ACS Paragon Plus Environment

Page 23 of 27

Chemical Research in Toxicology

   

a

 

b

    c

d Acid

Base

Neutral

Zwitterion

60

% Oral drugs

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

54

50

44

40

35

30

29

20

0

27

34

38

25

18

19

13

10

9 7

36

4

7

43 38

14 5

No DILI Most DILI Less DILI Amb DILI No data

  

  Figure 4.  Distributions of a) –pMDD, b) Fsp3, c) cLogP and d) ion class in the annotated DILI groups and No  data oral drug groups (Table 2).  For box and whisker plot definitions, see Figure 1. Categories connected by  the same letter are not different (p >0.05, Tukey HSD).                     ACS Paragon Plus Environment

Chemical Research in Toxicology

Page 24 of 27

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

        a

Δ = ‐0.16

Δ = ‐0.39

b

Δ = 1.0

Δ = 3.0

          Figure 5.  Distributions of a) cLogP and b) Fsp3 in No and Most DILI groups, divided by lower 80% and upper  20% daily doses.  For box and whisker plot definitions, see Figure 1. Categories connected by the same  letter are not different (p >0.05, Tukey HSD). 

ACS Paragon Plus Environment

Page 25 of 27

Chemical Research in Toxicology

a

Bases No DILI

10

10

8

8

6

6

4

4

2 0

Most DILI

No DILI

2 0

‐2

‐2

‐4

‐4

‐6

‐6

‐8

Neutrals

cLogP

Most DILI

cLogP

cLogP

No DILI

‐8 ‐8

‐7

‐6

‐5

‐4

‐3

‐2

‐1

0

‐7

‐6

‐5

‐p[MDD]

‐4

‐3

‐2

Most DILI

14 12 10 8 6 4 2 0 ‐2 ‐4 ‐6

‐1

‐9

‐8

‐7

‐6

‐5

‐4

‐3

‐2

‐1

‐p[MDD]

‐p[MDD]

 

 

Acids Most DILI

No DILI

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Fsp3

Fsp3

No DILI

Bases

‐8

‐7

‐6

‐5

‐4

‐3

‐2

‐1

0

‐7

‐6

‐5

Most DILI

No DILI

‐3

‐2

‐1

‐9

2

4

6

8 10

cLogP

Most DILI

0

‐7

‐6

‐5

‐4

‐3

‐2

‐1

Neutrals No DILI

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 ‐8 ‐6 ‐4 ‐2

‐8

‐p[MDD]

Fsp3

Fsp3 0

‐4

Most DILI

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Bases

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 ‐8 ‐6 ‐4 ‐2

No DILI

‐p[MDD]

Acids No DILI

Most DILI

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

‐p[MDD]

c

Neutrals

Fsp3

b

Fsp3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Acids

2

4

6

8 10

cLogP

Most DILI

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 ‐6 ‐4 ‐2 0

2

4

6

8 10 12 14

cLogP

  Figure 6.  Properties of No DILI and Most DILI drugs by ion class. a) cLogP vs –p[MDD]; b) Fsp3 vs –p[MDD]  and c) Fsp3 vs cLogP.     

ACS Paragon Plus Environment

Chemical Research in Toxicology

a

No DILI

Most DILI

1

1st cut

‐4.2

2nd cut

b

No DILI

Most DILI

Mean 0.8

14

3rd cut

Mean  3.5

12

0.9

10

0.8

8 0.7

Fsp3

0.5

6 4

cLogP

‐p[MDD] ≥ ‐4.2 &  FSp3 ≥ 0.28

0.6

2.4

2 0

0.4

‐2

0.3

0.28

‐4

0.2

‐6

0.1

‐8 ‐10

0 ‐9

‐8

‐7

‐6

‐5

‐4

‐3

‐2

‐1

0

‐p[MDD] ≥ ‐4.2 & Fsp3 ≥0.28

‐p[MDD]

81

Fsp3 ≥0.28

 

‐p[MDD]