Most drugs exert their effects via multi-target interactions, as hypothesized by polypharmacology. Here we introduce Drug Profile Matching (DPM) which is able to relate complex drug-protein interaction profiles with effect and target profiles. Structural data and registered effect profiles of all small-molecule drugs were collected and interactions to a series of non-target protein binding sites of each drug were calculated. Statistical analyses confirmed close relationships between the studied 177 effect and 77 target categories and the in silico generated interaction profiles of cca. 1,200 FDA-approved small-molecule drugs. Receiver Operating Characteristic analysis and 10-fold cross-validation was performed to assess the accuracy and robustness of the method. Based on the found relationships, the effect and target profiles of drugs can be revealed in their entirety, and hitherto uncovered effects and targets can be predicted in a systematic manner. In order to investigate the predictive power of DPM, four effect categories (PPAR agonist, angiotensin-converting enzyme inhibitor, cyclooxygenase inhibitor and dopamine agent) were selected and predictions in the set of the FDA-approved small-molecule drugs were verified by literature analysis and experimental tests. Moreover, a large set consisting of 600,000 druglike molecules was selected from a database of 50 million compounds and their interaction profiles were generated. Based on these profiles and chemical similarity considerations, predictions were calculated and tested experimentally to find new candidates that are chemically dissimilar to the reference drugs.

1. Drug Profile Matching
Drug discovery by polypharmacology-based
interaction profiling

2. Binding pattern to the
members of the proteome
Bioactivity:
effects / side effects
Target
Effect
Drug

3. ?
Similar binding patterns to
the proteome
Similar bioactivity:
effects / side effects
?
Scientific question and starting hypothesis

4. Interaction Profile (IP)
generation
Effect profile (EP)
generation
Interaction Profile Matrix
Against random (non-target) proteins.
Patterns instead of individual interactions.
Effect Profile Matrix
Contains binary effects
Drug
Simon et al, J Chem Inf Model 2012
Target profile
(TP)
generation
Target
Profile Matrix

5. Checking similarity, round one:
Vector distance vs. angle
Protein 2
Protein 1
Drug 2Drug 1

6. Predictions based on angle distance:
predicted and real profile of ziprasidone
QT prolongation: 9 out of the 13 nearest neighbors
Predicted property Described property
Effect
(category)
Adrenergic agent Antipsychotic
Anaesthetic Dopamine antagonist
Anti-anxiety agent Serotonine antagonist
Antiemetic agent
Antihypertensive agent
Antipsychotic
Dopamine antagonist
Metabolizing enzyme
Cytochrome P450 3A4 (CYP3A4) Cytochrome P450 3A4 (CYP3A4)
CytochromeP450 2D6 (CYP2D6) Cytochrome P450 2D6 (CYP2D6)
Mechanism of action
Adrenergic Adrenergic
Dopamine antagonism Dopamine antagonism
Histamine antagonism Histamine antagonism
Serotonine antagonism Serotonine antagonism

7. Identification of PPARγ ligands by
One-Dimensional Drug Profile Matching
Kovacs et al
Drug Des Devel Ther 2013

8. Effect Profile
Matrix
Interaction
Profile
Matrix
Negative
Positive
Simon et al, J Chem Inf Model 2012
Effect Probability Matrix
Target Profile Matrix → Target Probability Matrix

9. Effect Profile Matrix Effect Probability Matrix
„False” positives
True positives
Simon et al, J Chem Inf Model 2012

10. ROC (Receiver Operating Characteristic) curves
TPR
FPR
Sensitivity:
1-Specificity:

11. 0.839±0.081
Comparison of the distributions of
the effect and target profile-based AUC values
%
%
129 effect categories
Average AUC = 0.791±0.147
77 targets
Average AUC = 0.839±0.081
Penicillin-binding proteins 0.969
Beta-2 adrenergic receptor 0.934
Beta-1 adrenergic receptor 0.925
Mu-type opioid receptor 0.915
Gamma-aminobutyric acid receptor 0.890
Histamine H1 receptor 0.886
5-hydroxytryptamine 1D receptor 0.859
Alpha-1A adrenergic receptor 0.858
D(2) dopamine receptor 0.855
Muscarinic acetylcholine receptor M1 0.842
Monoamine oxidase A 0.841
5-hydroxytryptamine 2A receptor 0.838
Cyclooxygenase 1 0.835
D(1) dopamine receptor 0.827
5-hydroxytryptamine 1A receptor 0.826
5-hydroxytryptamine 1B receptor 0.823
Peroxisome proliferator-activated receptor 0.811
D(3) dopamine receptor 0.794
Reverse transcriptase 0.791
Angiotensin-converting enzyme 0.754

12. In vitro tests: COX inhibition
Telmisartan Proline
Novobiocin Dasatinib
Captopril Captopril
COX-1 COX-2
Alpha-linolenic acid Nitroxoline
COX-1 COX-2
Out of the 43
high-probability compounds:
- 4 confirmed by the literature
- 2 falsified by the literature
-10 reached Ki<50 µM
Végner et al, J Med Chem 2013

13. In vitro tests: ACE inhibition
Drug name Ki (µM)
Telmisartan 6
Proline 86
Novobiocin 167
Adenine 246*
Creatine 254*
Lamotrigine 287*
Tipranavir# 285*
Rosiglitazone 356*
Chloramphenicol 419*
Cimetidine 439*
Nelfinavir 542*
Dasatinib 715
Pentoxifylline 1540*
Sulpiride 2840*
Telmisartan
Végner et al, J Med Chem 2013

14. In vitro tests: dopamine agonism/antagonism
Name
Predicted
probability
Dopamine D1 receptor Dopamine D2 receptor
Ki (µM) in
Agonist mode
Ki (µM) in
Antagonist
mode
Ki (µM) in
Agonist mode
Ki (µM) in
Antagonist
mode
Celecoxib 0.995 <1 <1
Doxazosin 0.991 1
Cyclobenzaprine 0.977 2 2.6
Mitoxantrone 0.976 52 245 12
Flavoxate 0.971
Promethazine 0.966 19 2.6
Imipramine 0.952
Desipramine 0.951
Desogestrel 0.936 26.6 6.8
Epinastine 0.916 4.2
Clomipramine 0.907 1 669 4
Olopatadine 0.881 400 46
Thioguanine 0.878 188
Rimantadine 0.864 72
Mefloquine 0.854 22 122 6
Etodolac 0.796 335
Raloxifene 0.796 7
Fosfomycin 0.761 122
Végner et al, J Med Chem 2013
10 out of
18
active

15. %
Initial screening of 600,000 druglike compounds
JChem Base filtering for dissimilar compounds
Ki (µM)
10 20 30 40
8 compounds
between
40-70 µM

16. Drug Profile Matching is able
- to predict effects/targets of druglike molecules
DRUG DISCOVERY
and
- to reveal hidden effects/targets of FDA-approved drugs
DRUG REPOSITIONING
DRUG SAFETY

17. Acknowledgement
• Ágnes Peragovics
• László Végner
• Balázs Jelinek
• Barna Peitl
• Zoltán Szilvássy
• Péter Hári
• István Bitter
• Pál Czobor
• András Málnási-Csizmadia