1. Dealing with PAINs in a drug discovery CRO
D Swift, N Bland, R Lane, D Laughton, R Satchell, L Birch, S Pollack, J Unitt
Sygnature Discovery Ltd, BioCity, Pennyfoot Street, Nottingham NG1 1GF, United Kingdom
Abstract ID:                       
Overview:
Pan-Assay INterference compounds (PAINs) is a term used to describe a broad range of compounds that interfere with biological screening assays by acting through a range of mechanisms (see box “what are PAINS”). As such, PAINs are frequent hitters in screens, producing often unknowingly false positives that could waste much effort in drug discovery research, although almost 4% of FDA approved drugs are known promiscuous aggregators, indicating that PAIN activity does not preclude a compound as a potential therapeutic.
At Sygnature Discovery, rather than removing known PAINs from screening libraries, which could miss legitimate hits, we have implemented a target-specific systematic strategy to avoid, identify and characterize false positives during the hit identification and validation stages of a drug discovery campaign. We illustrate these approaches with two case studies which highlight two common PAIN mechanisms: detection technology interference and redox activity due to compound impurities.
What are PAINs?
Frequent hitters, False positives, Non-specifics
Many mechanisms
Impurities e.g. heavy metals, reactive contaminants
Multiple binding sites on enzyme
Protein destabilisers
Form molecular aggregates
Technological interference (e.g. fluorescence quench)
Redox: Generate reactive species
Reactive
Widespread in biochemical enzyme activity assays
PAINs are often not recognized as non-specific  Confusion & wasted time
Rhodanines
A family of compounds that are frequently found to be non-specific hits in high throughput screens
Identifying PAINs:
IC50 > 3µM
Steep concentration-inhibition curve (super stoichiometric)
Non-competitive
Irreversible
Flat SAR
Time-dependent
Sensitive to detergent (aggregators)
Active in unrelated screens-metal ion chelators
No comparable activity in orthogonal assay
Assay with different physical basis: e.g. binding v. activity
A known or predicted PAIN structure
Case Study 2:
Dealing with PAINs in a high throughput screen (HTS)
244,000 diverse compounds were screened at 10 µM against a metabolic enzyme (with an active site thiol) target as part of a drug discovery program. Conditions in the HTS were tailored to minimize the potential for false positives.
Counter-screening for PAIN identification:
In addition, a set of counter-screens was introduced to identify PAINS. A significant source of false positives were compounds that showed redox activity to due contamination. Redox activity can be detected by the reduction of resazurin in the presence of a reductant such as the cysteine in our assay buffer.
Results of HTS and follow-up:
Of the 244,000 compounds 150 of these had IC50 values of ≤ 60 µM against the main target. These 150 actives were then analysed in a series of counter-screens, including the redox assay described above, with the following outcomes:
89 compounds with PAIN structural flag; 122 with no PAIN counter-screen liabilities for follow-up.
Conclusions:
A clear counter-screening cascade following the HTS allowed for an efficient removal of PAINs and triaging of actives to take forward as validated hits.
Eflornithine
PAIN-like characteristics (irreversible with IC50>20µM)
Used to treat African trypanosomiasis
Assay condition/screen
Purpose
L-cysteine in buffer
Quench thiol-reactive compounds
[Tween] 0.01%
Minimise effects of aggregators
Case Study 1:
Lysine Specific Demethylase 1 (LSD1)
Epigenetic target
Primary screen
H2O2 peroxidase-coupled to resorufin fluorescence
2 hit series identified from HTS
Some SAR
Hit to Lead:
Orthogonal screen: HCHO Production
HCHO inhibition 
Counter-screen: HRP+H2O2
No HCHO inhibition 
An Amplex Red technology counter-screen excluded series 1 but not series 2
An additional orthogonal assay based on HCHO product identified series 2 as a PAIN.
Counter-screen
Purpose
Outcome
Resazurin redox assay
Identify redox-active PAINS
15% with redox activity ≤ 50 µM
ADP-GloTM assay w/o enzyme
Identify detection step interferers
3% with ADP-GloTM activity ≤ 50 µM
Unrelated Cys enzyme assay
Flag for nonspecific thiol-reactive compounds
27% with IC50 ≤ 10 µM
6 fold increase [Tween]
Identify aggregators
1% with potency shift of ≥ 10-fold
Dealing with PAINs as a CRO:
The case studies described above demonstrate the importance of recognising PAINs early on during a drug discovery campaign.
Case study 1- stopped the project early
Saving client time and importantly costs associated with pursuing nuisance compounds.
Case study 2 – PAIN management
Multiple classes of PAINS identified, efficiently triaging compounds for hit to lead optimisation
Orthogonal screens
Employ FTSA and SPR as complementary techniques, not only to measure binding affinity, but to identify small molecules with undesirable mechanisms.
Profiled common types of PAIN in both technologies and these data are a valuable reference when screening novel compounds.
Summary:
Aim to identify and flag PAINS using:
Publications on new hits
Use prospective tools to identify PAIN risk
But keep PAINS in screening sets
Develop a robust PAIN assay cascade specific to target
Primary screen (to avoid PAINs)
Orthogonal screen (Enzyme binding and activity)
SAR
Confirm structure and activity
Early chemistry at risk to increase SAR confidence
Mechanism of Inhibition
Once hit series structure/activity confirmed
Share the PAIN
Publicize PAIN management best practice
Primary screen
Orthogonal screens:
biophysics
Chemical filtering
Confirm structure and activity
Clear SAR
MOI Studies
Hit Series
Minimise effects:
Detergent, chelator
Resynthesis
repurification
Rescreen
Flag potential PAINS
Not super stoichiometric
No technological interference
Inhibition = binding
Identify non-specifics
Chemical clustering
Reversible/irreversible
Competitive/non-competitive
Detergent sensitive
Medicinal chemist input
Compound binding Kinetics by SPR
Compound binding by FTSA