Chem 195 Drug Discovery

1
Chem 195 - Drug Discovery

• Lecture 6
• Compound Collections, Assay Development, High
  Throughput Screening, and Informatics

2
Housekeeping

• Meet outside North end of Stanley hall for Chiron
  field trip at 130
• Next week no office hours
• Next Friday, 3/15, Dr. David Myles, from Kosan
  Biosciences will talk about medicinal chemistry
• 3/20 Guest lecture from Dr. Dirksen Bussiere on
  Structure-Based Drug Discovery 5-7 pm
• Mid-term is 4/5, Project outlines due 4/19

3
References on Web Site

• PfizerHTS.pdf - Pfizer perspective on what
  targets are more tractable
• Hts2.pdf - HTS and microfluidics
• Hts3.pdf - assay technologies
• Moldivrev.pdf - discussion of molecular diversity
  and combinatorial chemistry
• Compscreen.pdf - overview on computational (in
  silico) screening approaches

4
You have a target - What to Screen?

• Company compound collection
• Broad screening
• Med chem vs. combichem vs. natural products
• Single compounds or mixtures
• Focused screening (specific compound classes) or
  a diverse subset
• Collaborate with or purchase compounds from
  another group with a specific collection

5
Compound Collections

• Sources for Compound Collections
• Med chem derived - projects generate lots of
  compounds useful for subsequent screening
• Compounds tend to be related (analogs for
  structureactivity relationships
• Well characterized and often drug like
• History teaches that often leads for new targets
  can be derived from other target compounds
  (sulfonamides as an example)
• Big pharma compound collections have largely been
  built this way

6
Compound Collections

• Combichem derived - can generate very large
  numbers but are they useful as drug leads?
• Advantages - can design collection, ease of
  synthesis but quality issues
• Analysis based on activity
• Mixture screening vs. multi-parallel synthesis
• Can make 100,000 compounds (and more!) in a year
  with a few chemists and automation

7
Compound Collections

• Molecular Diversity - Is compound collection size
  everything?
• Concept - how many organic molecules are there in
  chemical diversity space?
• How to measure - what are the properties of
  molecules?
• Molecular descriptors - size, shape,
  lipophilicity
• A How many dimensional space?
• Orthogonality of descriptors
• Utility - why measure molecular diversity?
• Where does one want to be in that space?

8
Compound Collections

• Molecular Diversity
• Pharmacologically active space vs. ADME active
  space
• Does one only screen compounds with good ADME
  properties?
• Lipinskis rule of 5 - to improve odds on getting
  compounds with oral bioavailability
• MW lt 500
• ClogP lt 5
• H bond donors lt 5
• H bond acceptors lt 10

9
Compound Collections

• Natural products - fermentation, extracts
• Dereplication
• Activity and concentration
• Complexity of mixtures
• Reproducibility - obtaining samples, regrowing
  organisms
• Out of favor in the industry but there are very
  important drugs, especially in cancer and
  infectious diseases, which come from natural
  products
• Engineering of producing organisms to make new
  natural products - polyketide biosynthesis -
  combinatorial biology

10
Assay Development

• Screening and Selection
• Using genetics in drug discovery
• Growth and death - if theres an inhibitor your
  organism lives! (Positive selection)
• Example - protease inhibitors
• Chemical biology
• Genomics has led to a potential target explosion
• Target validation vs. drug lead identification
• Is chemistry good at this?

11
High Throughput Screening

• Screening
• In the simplest case need an assay to distinguish
  bound vs. free and a functional response
• Collection size (compounds x targets) has driven
  technology to miniaturization and homogeneous
  assay formats (mix and read)
• Assays need to be highly reproducible but a low
  level of false positives is common
• Hit rates are often 0.1 or less
• Robotics is required
• The biggest issue may be information handling

12
HTS

• Secondary Assays and Confirmation
• Given the logistics of handling many thousands of
  samples, confirmation of initial results is key
• New samples from archive
• Reproducible results
• Secondary assays often use a different format to
  reduce false positives, i.e., fluorescence
  quenching by colored compounds

13
HTS Assay Methods

• Assay Detection Methods
• Automation and Format
• Small is better
• 1, 24, 96, 384, 1536 well plates
• Radioactivity, Fluorescence, HPLC, FACS, reporter
  genes, CCD cameras

14
Detection Methods in HTS

• Radioactive Methods
• Highly sensitive - work for miniaturization
• Separation and/or washing makes laborious
• Preparation of labeled tracers can be difficult
• Scintillation Proximity or Flash Plate Assays as
  a second generation method
• Health and Waste Concerns

15
SPA
16
Flash Plates - SPA without Beads
Radiolabeled Tracer and Sample added
Only bound Tracer is Measured
Bind Receptor Or Target to Plate
17
Detection Methods in HTS

• Fluorescence Methods
• Highly sensitive - work for miniaturization
• Fluorescence anisotropy (FP)
• Time resolved fluorescence (TRF - Europium
  Chelates)
• Fluorescence Resonance Energy Transfer - donor
  quencher pairs - protease assays

18
Principle of FP
Low MW molecules rapidly rotate relative to F
relaxation, thus have low FP, whereas bound
molecules slowly rotate hence have Higher FP.
19
FP
Theoretical FP Behavior of a Series of
Fluorescent Dyes As a Function of MW
20
Time Resolved Fluorescence Assay
21
Fluorescence Resonance Energy Transfer
Two conditions must pertain for an effective FRET
assay 1) The donor fluorescence must be
significantly quenched 2) The starting material
and product need to differ significantly in
extent of quenching (Forster radius ca. 50 A)
22
FRET

• Example of use of FRET in a protease assay

Matayoshi et al., Science 247 954 (1990)
23
Informatics

• Relational databases are key for retrieval of
  information
• Chemical, screening, biological, genomic,
  pharmacological information
• Information explosion - big companies are
  generating terabytes of data per year
• What information is worth having/knowing?
• How to write intelligent queries?
• Balancing power with ease of use
• Often custom front-ends with Oracle database
  back-ends

24
Informatics

• Linking assay data, compound identity/location,
  and compound structure
• Representation of data to take advantage of
  things humans are good at (pattern recognition)
• Storing large numbers of chemical structures has
  required the development of new languages for
  their representation (Smiles, Grins, etc.)
• http//www.daylight.com/dayhtml/smiles/

25
Summary

• Compound Collections and Molecular Diversity
• High Throughput Screening
• Assay Methods
• SPA
• FP, TRF, FRET
• Informatics