1 Options for Automated Solubility Measurement Options for Automated Solubility Measurement In Discovery and Developments: In Discovery and Developments: Science Issues and Practical Solutions Science Issues and Practical Solutions Arnon Chait ANALIZA, Inc.
60
Embed
1 Options for Automated Solubility Measurement In Discovery and Developments: Science Issues and Practical Solutions Arnon Chait ANALIZA, Inc.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Options for Automated Solubility Measurement In Options for Automated Solubility Measurement In Discovery and Developments:Discovery and Developments:
Science Issues and Practical SolutionsScience Issues and Practical Solutions
Arnon ChaitANALIZA, Inc.
2
OUTLINEOUTLINE
Solubility: the basics Solubility assays Particles in solution: why should we care? Focus: elemental solubility assay
– How does it work?– Sample data for pH, excipient screening
Solid particles are an integral part of the solubility assay– Particles are always present– They must be present for turbidity to work– They are artifacts in absorbance/elemental
assays
The effects of particles on the data must be considered when examining the data– Subtle to substantial influence on quality of
DMSO effect can be very large, and is unpredictable
Starting material has significant influence on the data
DMSO effect can not be neglected, even at small amounts
Late discovery and development should avoid DMSO if at all possible
6
A Simple ConceptA Simple Concept
Solubility = Concentration of a dissolved compound in equilibrium with its solid
But:– Which solid?
Equilibrium (most stable form) vs. apparent (other forms)
– Which solvent? Buffers (intrinsic?) and co-solvents (kinetic)
– Which equilibrium? Time (kinetic) and temperature
7
But It’s All in the DetailsBut It’s All in the Details
Conceptually a very easy experiment to conduct
But:– We don’t always have the right form (discovery)– We don’t always have time to wait for
equilibrium– Everything matters:
Probably one of the most sensitive thermodynamic experiments to conduct
– Different users refers to same term with different meanings
8
Who Are The Users?Who Are The Users?
Discovery:– Obtain a measure of solubility of relevance to HTS and early
ADMET Can we determine if the compound will crash out of
solution?
– Obtain ADMET information of relevance for selecting amongst hits
Development:– Early lead characterization– Lead optimization– Formulation excipient selection
9
Who Are The Users? (2)Who Are The Users? (2)
Will the compound crash out? solubility
Everyone else needs good data:
– For promoting a compound: binning may be sufficient
– For lead optimization: accurate data
– For formulation development: accurate data
– For anything regulatory: accurate data
Arnon Chait
Binning does not mean that you can relax HOW you get the data - the method has to provide faithful solubility data, but could have a larger error bar for binning.
10
How Does a Compound Come Out of How Does a Compound Come Out of Solution?Solution?
Homogeneous vs. heterogeneous nucleation
Growth by dislocations (crystalline) vs. amorphous
Turbidity:No, solubility is inferred from dilution factor off a standard
13
What Controls the Size Distribution in Time?What Controls the Size Distribution in Time?
Ostwald coarsening: Particles nucleate at a minimum radius (critical radius)
Growth is an interplay between:oversaturation level in the liquidgain/loss from dissolving/growing neighbors
particles
The particle size distribution is a solution this interplay
Lifshitz-Slezov and Marqusee and Ross theories predict a continuous growth of larger particles at the expense of smaller ones: 1/3( , )f r t t
14
But Why Should We Care?But Why Should We Care?We Can Always Filter the ParticlesWe Can Always Filter the Particles
You can NOT filter all the particles– You must have solid at solubility
Filtering simply places a dynamic cutoff limit to the size distribution function – but it will again change in time!
Filtering is necessary to reduce the potentially deleterious effects of particles
Log(Size) Log(Size)
Time
15
Take Home Message #1Take Home Message #1
Particles are a fact of life:without them - we have no solubility
Particle size is changing dynamically
You can only interfere – for a while - with the natural size distribution by filtering
Each assay technique is sensitive to particles to a different degree
16
Optical Properties of Particles in SolutionOptical Properties of Particles in Solution
Van de Hulst, 1957
Gold particles
Scattering is very small in comparison with absorbance when the solid particles are small
MeasuredSignal
Arnon Chait
This is for spherical gold particles.Note that scattering is very low compared with absorbance until large particles appear. Also, these particular data is for gold, which has a specific complex refractive index. What's happens for screening, in which we don't know the particle refractive index, size, or shape?
17
Can We Predict Optical Effects of Particles Can We Predict Optical Effects of Particles In Practice?In Practice?
Can You Tell The Saturated Solution?Can You Tell The Saturated Solution?
20
So Let’s Filter the StuffSo Let’s Filter the Stuff
Dynamic Light Scattering:
Bendroflumethiazide, saturated, filtered
Filter,m
Particle size,nm
Polydispersity
0.2 481 – 4,000 0.226
0.45 270 0.096
1 543 - 2,000 0.327
21
Particle Effects in Optical Assays (1)Particle Effects in Optical Assays (1)
Turbidity:– Relies on particles to produce scattering– No filtering is desired– Particle evolution is unpredictable
Absorbance:– Scattering is an undesired artifact
– Filtering is desired– Scattering contribution to data is unpredictable
0 scattered transmittedI I I
22
Particle Effects in Optical Assays (2)Particle Effects in Optical Assays (2)
I0It
No particles:
No scattering
Beer’s law works
Is
Particles in solution:
Scattering
Beer’s law does not work
~A C Determine from dissolved samples (undersaturated)
Use in saturated solutions
23
Caffeine concentration, mg/ml
0 20 40 60 80 100
Ana
lytic
al S
igna
l
0
10
20
30
40
50
60
70
The Ideal ExperimentThe Ideal Experiment
All of these data have particle effects!
24
Saturation is a Difficult Place To AssaySaturation is a Difficult Place To Assay
nm solids artifacts:
- Assay specific- Significant for optical techniques
25
Take Home Message #2 Take Home Message #2
Particles will affect solubility data:
Turbidity:Increase, perhaps
significantly
Absorbance:Increase, unpredictably
Elemental:Increase, more predictably
26
What Do We Want to Study?
Time
Temperature
DMSO Crystalform
Salt
Co-Solvent/Excipient
pH
Solubility
Time
Temperature
DMSO Crystalform
Salt
Co-Solvent/Excipient
pH
Solubility
27
How We Chose to Do the Job:How We Chose to Do the Job:Automated Solubility WorkstationAutomated Solubility Workstation
Philosophy of Design:– Accuracy is everything– Assay flexibility is a must– Medium throughput is acceptable (up to 250/day)– System must be compatible with dry and DMSO-
dissolved– Minimize number of assumptions– Native assay relies on elemental assays: no
standards required– Eliminate the need for experienced chemist by using
intelligent software– Simplify and modernize everything significant
28
Solubility: Where Do You Spend Solubility: Where Do You Spend YourYour Time? Time?
What is the Effect of DMSO?What is the Effect of DMSO?
DMSO is a fact of life during screening
DMSO solubility is a current issue in HTSe.g. K. Balakin, “DMSO Solubility and Bioscreening”, Current Drug Discovery, (August 03)
Key Issue: How did you start?– DMSO-dissolved sample: NO solid
Precipitating solid at saturation crystalline form
Difficult experiment requiring high concentrations
– DMSO added to solid form
Arnon Chait
DMSO has unpredictable effects in different assays. Bioassays may be less sensitive - but not alwaysSolubility is more difficult to modify than bioactivity, so we have to be much more careful Only rational way for working with DMSO is to consult multiple independent parameters and interrogate the data vs. structure for common sense
46
DMSO – Time ExperimentsDMSO – Time Experiments
30 mM DMSO stock solutions, 20-fold dilution Powder with buffer or with 5% DMSO 50 mM phosphate buffer at pH 11
Source % DMSOIncubationTime, hrs.
Powder 0 20
DMSO stock 5 20
DMSO stock 5 1
Powder 0 1
Powder 5 20
47
Test Compounds – Solubility (mg/ml)Test Compounds – Solubility (mg/ml)
1 bendroflumethiazide 0.3507
2 nifedipine 0.0013
3 nitrofurazone 0.1615
4 perphenazine 0.0006
5 trimethoprim 0.3440
6 nafoxidineHCl 0.0040
7 nadolol 0.0038
8 imipramineHCl 0.0007
9 verapamilHCl 0.0064
10 clofazimine 0.0006
11 bifonazole 0.0009
12 clomipramineHCl 0.0004
13 labetalolHCl 0.5946
14 chlorpromazineHCl 0.0023
15 triflupromazineHCl 0.0013
16 thioridazineHCl 0.0007
48
Group I: Fully DissolvedGroup I: Fully Dissolved
#1 #5 #13So
lub
ility
Re
lativ
e t
o T
ota
l Co
mp
ou
nd
Con
cen
tra
tion
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Powder in Buffer alone, 20 hr incubation30 mM DMSO stock, diluted 20-fold in bufferPowder in buffer with 5% DMSOPlot 1 Upper Control Line
49
Group II: Similar ResultsGroup II: Similar Results
#3 #7 #11 #14 #15
So
lub
ility
Rel
ativ
e t
o P
ow
der
in B
uff
er
Alo
ne
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Powder in Buffer alone, 20 hr incubation30 mM DMSO stock, diluted 20-fold in bufferPowder in buffer with 5% DMSOPlot 1 Upper Control Line
50
Group III: Significant DMSO Stock Effect Group III: Significant DMSO Stock Effect
#4 #2 #6 #10 #16
Sol
ubili
ty R
elat
ive
to P
owde
r in
Buf
fer
Alo
ne
0
10
20
30
40
Powder in Buffer alone, 20 hr incubation30 mM DMSO stock, diluted 20-fold in bufferPowder in buffer with 5% DMSOPlot 1 Lower Control Line
51
Group IV: Significant DMSO Effect Group IV: Significant DMSO Effect
#4 #8 #9 #12
So
lubi
lity
Rel
ativ
e t
o P
owde
r in
Buf
fer
Alo
ne
0
5
10
15
20
25
Powder in Buffer alone, 20 hr incubation30 mM DMSO stock, diluted 20-fold in bufferPowder in buffer with 5% DMSOPlot 1 Upper Control Line
52
Powder Plus DMSO: Reasonable at Low %Powder Plus DMSO: Reasonable at Low %
Allopurinol
Bendroflumethiazide
Butamben
Clofazimine
Nitrofurazone
Theophylline
Nifedipine
Perphenazine
Phenacetin
Sulfanilamide
Trimethoprim
pH 7.4
53
Powder Plus DMSO: Reasonable at Low %Powder Plus DMSO: Reasonable at Low %
54
DMSO Stock Effect is Still Significant at 1%DMSO Stock Effect is Still Significant at 1%
DMSO effect is most pronounced when starting material is DMSO-dissolved
DMSO effect is still significant and unpredictable in practice, even at low DMSO %, when starting from DMSO-dissolved samples
DMSO-dissolved samples are fine for early discovery questions, but can not be relied on for later discovery and development
56
How Do We Deal With Impure Samples?How Do We Deal With Impure Samples?
Impurities are a fact of life:– Combi libraries are 90-95% pure– Lead synthesis is similarly pure
Turbidity:– No need, usually, but you don’t know what
came out of solution
Absorbance:– Use spectral techniques – not trivially
Elemental:– Clean sample: inline column + fast gradient
57
Final ThoughtsFinal Thoughts
If you need good numbers, you have to watch your steps:– Particle effects in optical methods– Filter effects– DMSO effects when starting from dissolved
samples Elemental method provides for automated,
accurate determinations with full accessibility to later discovery/development questions: pH, temperature, salts, etc…
Hardware is expensive Software is very expensive Gold-standard assays are always that Only big pharma can own “one of each” In silico predictive analysis works much
better on same series, if properly trained Chemists will use your data if it is GOOD,
Automated Discovery Workstation: Multi-assay on the same hardware platform:
– Solubility (dry/DMSO, pH, salts,…)– LogD (pH)– BBB (near term GIT) permeability– Coming: Tox (cardiac), protein binding
Nothing proprietary, no magic, only gold-standard assays
Never have to weigh or prepare standards Never have to believe your library concentration Optional on-line sample cleanup Future integration with predictive software