Making uHTS a workhorse in Drug Discovery: No gain without pain

Maulik P. Suthar,
Abbreviations: CD: Candidate Drug, NCE: New Chemical Entities, NDA: New Drug Application, IND: Investigational New Drug, HCS: High Content Screening, USD: US Dollar/s

Outline:
Introduction
Drug discovery & HTS
HTS instrumentation
Consumables, Reagents
Library & Data analysis
Assay Development
Conclusion
References

1. INTRODUCTION:

Role of HTS in pharmaceutical industry was developed & evolved dramatically since the early 1980s when Dr. Solomon Snyder at Johns Hopkins University of School of medicine began lobbying about the potential merits of HTS in modern drug discovery paradigm.

Definition of HTS: ‘Processes in which batches of compounds are tested for binding activity or biological activity against target molecules’. HTS tests the binding activity of thousands of compounds against a molecular disease target, usually a protein. This binding activity is a measure of the compound's ability to inhibit the activity of the protein, the first step in judging whether the compound is a potentially promising candidate drug.

Objective of HTS: ‘To interrogate large chemical collections in context of biological target to accurately identify active chemotypes.’

HTS is an evolving process that is today a discrete activity and may tomorrow become more highly integrated into a rapidly changing drug discovery paradigm. HTS is evolving not only as a discrete activity, but as a perspective that is expanding backward toward target identification and validation and forward to converting assay hits to qualified leads via information generated either within screens or through downstream, high-throughput ADME/tox (absorption, distribution, metabolism, and excretion) and toxicity testing.

Within a span of last four years uHTS(ultra/miniaturised HTS) has become the focal point for drug discovery programs in pharmaceutical & biotechnology industry. Typical screening facilities of pharmaceutical companies mount 20-35 campaigns by leading uHTS leaders which generates 2-18 millions primary screening data points within a year. However, some major players are planning to mount 50-100 screening campaigns a year with the compound collection of 0.5-1.5 million compounds.

Since the cost of biological assay can vary from a few cents to dollars per well, with the potential cost of a moderately priced screen exceeding USD 300,000, So compound-pooling strategy can save the significant cost. In the cutting edge & cutthroat completion in the field of drug discovery, it is essential for generic pharmaceutical companies to devise new blockbuster molecules as large numbers of patented molecules are expiring.

Reagent expense is the key bottleneck for HTS assay. Specialised reagents such as fluorescent probes, antibodies, conditioned media or reporter enzyme substrates can cost as much as 1USD/HTS compound tested when they are commercially available. So, assay miniaturization is inevitable by use of higher density plate format to reduce reagent expense, however instrumentation for compound libraries adoption and for higher density plate format can be significant. Hardware, software & reagents need to be reinvented to handle the increased numbers of targets and data points that are required in order to deal with the influx of genomics based entities.

Current Trends include: (1) the influence of genomics on increasing both the quantity and quality of new drug targets; (2) advances in combinatorial organic synthesis to increase both the quantity and quality of compound libraries; (3) the influence of high-throughput screening in providing an increased supply of new lead compounds; and (4) the enhanced use of bioinformatics for process integration.
[High-Throughput Screening in Review, Drug and Market Development (6 December 2001), http://container.pharmalicensing.com/features/disp/1005568086_3befc0562a952]

Following a difficult 2002, the result of the economic downturn, reductions in biotech/pharma R&D spending, and consolidations a slight increase in R&D spending was experienced in 2003. There are now positive signs of recovering pharma budgets and enhanced outlays. Overall, the outlook was one of "cautious optimism" for previoius year. Global drug discovery spending, which was $19.6 billion in 2002, is expected to reach $25.1 billion in 2006. [ http://www.iptonline.com/newproducts.asp ]

2. DRUG DISCOVERY PROCESS & HTS

After the second level of triage by HTS campaign, hits become lead compounds. Further synthesis to build up a focused library may then be required to provide a variety of compounds structurally related to the lead. These sublibraries can then be screened against targets in order to choose optimal structures. At this stage, some basic indicators of toxicity or bioavailability may be considered in an attempt to eliminate potential failures as early in the discovery process as possible.

90 % CD fail in clinical phase of development. Among of 60 % CD fails because of ADME/tox problems. Earlier the CD fail, more money can be saved. Utilisation of the secondary screening can be useful in the early development to extract the inadequate compounds. Many years of development time & hundreds of millions of dollars are being saved due to its expanding benefits of HTS.

In fact, we have very good system to show that compounds have specific activity, but we don’t have enough models to evaluate where that compound fails in earlier clinical studies. Pharmaceutical & Drug Discovery industries expend more money and resources on compounds that fail in development than succeed in market place; major reasons for failure are toxicity & poor kinetics.

Selecting a right target:
Targets are plentiful but concerning to drugable targets are less and once decision on target is made, inhibitors need to be found which can modify target function to obtain desired effect. There are complex targets that cannot be adequately configured or reconstituted in a biochemical assay e.g. complex interactions between receptors, co-activators, co-repressors, response elements and other cellular factors that cannot be adequately reproduced outside cell. Such targets are involved in signal regulating pathways, target themselves regulated by poorly understood partners along a signalling pathway, and targets those require assembly of a transcriptional regulatory complex.

Compounds are assayed in magnitude of hundreds of thousands, to identify QSAR. Random screening is used to identify screen from existing compound libraries. These leads can be used to design backbone for focused libraries. In next stages leads are randomly modified by adding molecular segments of other agents to improve the activity of initial backbone. Good IC50 value in the range of micro-nano litre is the indication of the success. But, the good leads should be novel in order to obtain patents application to have competitive advantage in drug market.

Once significant animal safety testing on various models is completed, Drug Discovery Company can submit results with an Investigational New Drug (IND) application to FDA. Once having reviews of animal and pre-clinical data FDA decide to approve IND to conduct clinical trials.

Only 20% of INDs survive procedure from clinical trail phase I-III and further views of FDA on results. To this milestone, bill comes nearly to 200-600USD & 10 years. [Thesis ref: Markus Wunder_Diplomabeit]. Compound requirements: logP, solubility, ionic state, compound reactivities, biological effects, ADME/tox etc.

Requirements from industry:
The pharmaceutical industry currently has a pressing need for improvement in HTS technology because of insatiable appetite for new lead compounds; it is also under continuing pressure to reduce the costs of discovery and development. Pharmaceutical companies are recognizing that their future success is tied to viewing the entire drug discovery process as a single entity.

Manufacturers of HTS:
Manufacturers have risen to the occasion with a plethora of new offerings, including new assays, new formats, new technologies, and new instruments. Microfluidics technologies, for example, are being applied to the development of systems that dispense subnanoliter quantities of reagents. New alternative technologies and systems come at a cost, and manufacturers are continually faced with assessing their true benefit:cost ratios.


Business of HTS:
The field of HTS has generated several types of strategic alliances. The technology access agreement is a common mechanism whereby a large company gets to use a new technology and participate in its late-stage development. For that privilege, the pre–commercial stage partner tries to obtain up-front payments, research and development payments, milestone payments, and royalty payments for resulting products. Many collaborations of this sort have indeed provided all or most of these kinds of payments.

From primary screening to the preclinicalstudies estimated time requirement is about 6 years ( 40 % of overall time of drug discovery) , which cost about 200 million USD ( 33 % of overall cost of drug development). HTS helps to reduce the time for early market opportunities, that is why investor are interested for HTS business.

Hig throughput screening- where we are today? ,
http://www.images.technologynetworks.net/resources/evalrev1.asp


Growth of HTS:
The overall pattern for the industry serving pharmaceutical HTS is one of steady growth at rates exceeding pharmaceutical sales growth rates by a considerable margin. Growth rates for HTS exceed even rates of growth for pharmaceutical R & D. Even these high growth rates for screening lag behind the actual needs of the industry. Growth in HTS product and service revenues will be constrained by caps placed on pharmaceutical R & D expenditures in order to maintain acceptable profit margins. The revenue growth rates reflect a balance between burgeoning opportunity and the realities of spending constraints. The pharmaceutical industry is challenged to increase its output of innovative new drugs while maintaining profit levels that are acceptable to the investment community.

Eichma M and Goldstein I (1999) , The hig throughput screening mrket for drug discovery,

HTS approaches in industry & academic:
In industry: disease related pathways and more commercial screening.
In academic: variety of pathways in diverse organism (basic research), so require flexibility for hardware.

Fig1: general targets for drug discovery

3. HTS INSTRUMENTATION

Plate Format: (Need for miniaturised assay compared to traditional HTS)
96 well plate format is standard format for screening. A typical 96 well plate assay uses about 100microL test solution but reactions in 1536 well plate use only 0.5-10microL. See table No 1. It gives higher information density per unit volume of sample that is screened. Screening of lots of compounds at rapid rate allows larger amount of information to be gathered in a given unit of time. Compounds with no downstream value can be triaged earlier in drug discovery process.

If it assumed that 200 plates per day processed than, with 96 well total =19,200 individual analytes , with 384 well plate total = 76,800 individual analytes & with 1536 well total = 307,200 individual analytes can be generated.

Miniaturisation is the way to increase throughput & decrease cost per data point. However, road of miniaturisation is not easy. Implementation of miniaturisation poses many technical problems;
Tolerances in positioning of the plates, dispensing compounds & recognising a hit must be greatly improved.
Integrate the data of subsequent screens; because little is known about what mechanism & reaction is happening in well, even modification of the screen can yield the different results.
What is about the large molecule screening, they might not react in the small assay plate format?
Again lowering well size increase the evaporation rate, so assays have to perform very fast to minimise evaporation effect.

Interaction between engineer (robotics) & screenist (biology) is not an easy; we need biologist who can talk to the engineer because each have to learn what they can do & what cannot when designing & integrating for development and merging two advanced technologies.


Table 1: Comparison of the different plate formats

[Reference: Houstan & Banks, the chemical-biological interfaces: developments & miniaturised screening technology, pp 737,]

Plate format not only differs on the basis of material of composition, well shape, well depth & transparency but there are plates available with pre-coated various chemicals & reagents that are required for various assays. Standards of microplate are being led by Society for Biomolecular Screening (SBS).

Plate requirements:
Eliminates the performance-limiting well-to-well light cross talk

LHS:
Machines capable of deep well pipetting are usually more expensive than one capable of shallow well pipetting. Most high accuracy LHS instruments combine accurate low volume liquid handling with other functions such as library reformatting, cherry picking or pin-transfer.

Automation:
It must have stand-alone capabilities and can be integrated with each other for automation of sequential steps in screening protocols. This affects the choice of software and hardware components to be chosen. Integration of screening instruments can be done by contracting vendor or by employing an on staff robotic programmer. Prerequisite of such system is that it must be able to transfer micropaltes from one instrument to other instrument.

Undoubtedly, productivity and data quality can be significantly improved by automation, but there are cost considerations. In particular, staff had to be trained, technical experts employed, and it is often necessary to involve external contractors to maintain maximum system efficiency. In addition, when less stable reagents are used, regular, even hourly, manual intervention could be necessary. Liquid handlers must be able to pipette accurately & reproducibly into smaller wells, washers and dispenser must be able to handle wide variety of plate format & readers must be able to detection. Robot grippers can handle 384 well plates since external form factor is the same, but in many cases more accurate positioning of the plate is required on the target device because of the smaller wells and higher density of wells. Most of the issues pertaining to this are resolved for 384 well plate format & after that it’s a chance of 1536 now.

Advantage of the workstation based uHTS systems enable drug discovery companies to divide & conquer: perform screening activities in remote locations & share data over networks such that all authorised users can access same data set that is being generated by HTS campaign and analyse this data set in real time. This offers following advantages;
More leads (more screening = more hits = more leads), Increase screening prowess (increase capacity via miniaturisation & automation)
Careful assay development can improve lead quality
Higher quality leads are more likely to result in drugs that are successful in marketplace

The three factors, which can broke down reliability of components:
Functional quality of device
Ease of physical interaction between device & transport mechanism
Software routines used to control devices from different manufacturers

Ideally, a system should be capable to accommodate different devices from different manufacturers. This makes the system flexible and allows changes for different approaches. Maintenance of system is also a one important factor from changing bulbs in detector to avoiding crystallisation in a plate washer.

Table 2: Cost analysis of the uHTS components

HTS platforms ranged in size from small, partly automated, desktop machines costing between 18,600 USD and 93,000 USD, to fully integrated, robotic systems. Costing around 372,000 USD, tended to be operated continuously in dedicated laboratories.

Detection technologies:
Not all fluorescence techniques are amenable to uHTS when the reagent and test compound savings are to be considered. By use of florescence techniques most detailed information on bimolecular processes can be obtained by detecting fluorescence from single molecules. As a result of that florescence based assay using Single Molecule Detection (SMD) techniques are evolving as an important tool.

Table 3: Amenability of detection technology to particular assay type

Cell based imaging –automated microscopes:

Cell based HTS formats could represent the fastest approach to screen poorly characterised targets. Cell based assay provide a means to discriminate agonist, allosteric modulation and antagonist activity at receptors that cannot be determined in binding assays.

Automated microscopes can be used to monitor changes at the level of individual cells within an assay well. These instruments perform iterative auto-focusing to acquire images of each well. It is important to level of throughput required. Since more information is retrieved from such instruments, obviously it takes more time compared to conventional plate readers. Reading times are longer approx 45-90 minutes for 384 well plate format by automated microscopy compared to 3 minutes by plate readers. Points must be considered are; whether multiple excitation/emission wavelengths can be used, incubation at higher than ambient temperature required or not.
Requirements of cell based assays:
1. Generation and characterisation of appropriate cell model
2. Production of sufficient cells for HTS
3. Plating cells for assay
4. Effects of compound exposure and capture of assay signal

Advantages of cell based functional HTS (HCS):

Ability to distinguish between agonists, inverse agonists & neutral antagonists
Rapid selection of modulators with desired pharmacological properties
No limitation in ligand binding assays with isolated targets
Potential to reveal indirect modulation of target response in the complex network of interwoven cellular signalling pathways
Compound can interrogate an entire signalling pathway rather than a single isolated target with multiple potential points of interaction
Reagents are general rather than being distinct for each different target

Table 4: Comparison HTS/HCS
Validation of HTS:

Validation of:
1. Kinetic and pharmacology of assay
1. DMSO tolerance
2. Reagent stability
3. Signal stability

Calibration and maintenance are important. Usually LHS arrive with the calibration certification from instrument company but conditions under which certificate given is not necessary that corresponds with actual laboratory. So, it is good practice to verify accuracy & precision of a machine through a wide range of parameters and to continually monitor findings at regular intervals (monthly or quarterly).
Levelling of a machine: bench top & floors may not be in level, surface that is only slightly off-level can produce significant inaccuracies in volume of liquid transferred due to small size of dispense needles.
Liquid Handling System Validation parameters:
Liquid class properties: soluble reagents, cells, beads, and compounds, radioactive/fluorescent?
Volumes needed
Accuracy/precision requirements
Carry over (contact/non-contact)
Time between dispense/read
Interface/architecture permit batch processing?
Reliability/ease of use/support
Detector Validation parameters:
Z heights/focal heights
Plate definitions-drift observations
Sensitivity
Variability
Read times
Flat field corrections (imagers)
Equipment selection criteria:
Low dollar spending in consumables per lead discovered
Cost per assay point & reliability of results
Ability for multiple integration with different components
Kind of after-sale services provided by vendor, service network.

Design facilities for workspace:
It can vary depending upon the kind of instruments, number & type assays to be run and number of personnel to be accommodated.
Followings are the main facilities required:
Tissue culture area:
Water-jacketed CO2 incubator for mammalian cell based assays,
Only standard 37 C incubator for yeast or bacterial assays
Six-foot, laminar flow tissue culture
Room for an automated plate filter: to dispense cells into assay plate
Clinical centrifuge: for compound stock plate as well as assay plate (mother plate & daughter plate)
Cold room
Refrigerator in tissue culture area: temporary storage of cell media and assay reagents
Additional free workspace: for individual researcher to prepare & carry out screen.
Standard lab benches: to carry out small number of freestanding machine.
Service facilities:
Vacuum and gas (e.g CO2 and air) services can be unreliable alternatively individual vacuum pumps & compressors
Computer network connections (data jacks): for assay detection & data capture phases of screens. As a large amount of data generated during HTS operation the final destination (e.g. Data-server and/or database) should be planned before screening begins.
Carts for laboratory equipments: should be flexible, these can be used separately or connected together to provide modular integrated workstations.
Close relationships: between local computing support group and data management group to ensure a smooth transfer of data from the screening facility to servers & databases

Ancillary Instrumentation:
Plate filters: to perform all of liquid handling steps of some screens, these are machines use a manifold (8- or 16- channel) to rapidly dispense cells or reagents into assay plates with an acceptable level of accuracy. Plate filters are straightforward in their operation and can be independently used by biologist.
Low volume automated pippetors: more accurate but expensive

Plate washing: semi automated or fully automated plate filter are available which add and wash reagents attached to a vacuum line for aspiration step, be used. Usually the washing operation (aspiration followed by buffer addition) can be carried out within 30-60 seconds for 384 well plate format. Time me vary depending upon assay type & plate format type.
Disadvantage: needles that perform these tasks become clogged easily, despite regular cleaning. So plate washers require more care and maintenance than other LHS machines.

4. CONSUMABLES & REAGENTS

Reagents:
Instrumentation & reagents make up the crucial elements of screening technology. Reagents need to tune in instruments so that end user (biologist/screenist!!!) receives complete solution out of box rather than a set of products that need to be configured in order. In a typical set-up, instrumentation cost for HTS is usually smaller that yearly costs of consumables used to perform screens. That is why the demand of highly flexible system is growing to integrate with both purification & detection instruments. Integrating products from various vendors have become an issue to standardisation. It is estimated that every dollar spent in equipment, eight are spent in consumables. This ratio suggests the need of equipment manufacturers to design sophisticated solution that will increase efficiency of consumables use.
Industry trends in high throughput screening ,
http://www.images.technologynetworks.net/resources/evalrev3a.asp


Broad range of reagents are utilised in drug discovery. Reagents need to be modified to be used in HTS quantities and adapted for specific detection technologies like fluorescence, scintillation (radioactive), and chromatography. However, there is no trend toward which reagents are going to be used in future. Web search of reagents is becoming the standard for information transfer, as BCC assume the segmented market of 60,000 commercial compounds. So with little orders of the same reagents, finding a stable market niche is like ‘ finding a needle in a haystack’.

Reagent selection criteria:
Bulk manufacturing capabilities
Cost, quality and delivery time
Support to modify reagents for different HTS detection technologies
Breadth of product line

Consumables:
For each new assay run new assay formulations are required to be developed. As the trend is going on miniaturisation, plates & pipettes are widely used. HTS pipettes will become smaller and more precise. Users decide the plate format depending on detection technology. Again special plate material needed for specific applications. Decision is usually made on basis of price & requirements of particular assay.


5. COMPOUND LIBRARY & DATA ANALYSIS

Combinatorial chemistry is no longer viewed as a speculative technique for R&D in the Pharmaceutical and Drug Discovery industries but has become an established important technology. High throughput screening and combinatorial chemistry are altering the character of discovery research.

Requirements of Compound library:
Quality of compound library depends upon the chemical diversity of library, purity, and drug likeliness, synthetic feasibility of compounds. If compound libraries become more focused, this can be accelerated as HCS will have high enough throughput to screen entire libraries.

Screening of the entire library is governed by some factors:
Should provide comprehensive coverage of high value targets
Complete annotation for data mining purposes is necessary
Provides momentum to the screen
Certain assay cannot be adapted to uHTS
Time, money & energy plus infrastructure (robotics) required
Clogging of hit verification due to overload
Data overload (uHTS/HCS)
Physical depletion of compound library collection

Compound purchase:
Compound selection criteria: cost per library, quality/purity of compounds, re-supply cost, availability of individual compounds for follow up experiments. Vendor purchased compounds are expensive but vendors can provide more follow-up options like guaranteed re-supply/re-synthesis, discount to institution set-ups, none place of intellectual property restrictions. As time passes vendors to the library add new compounds so availability of older compounds generally falls. Therefore, it is important to ask vendor how long re-supply of their compounds is guaranteed & what options are available when compounds are out of stock.
Options of shipment of compound are also a critical decision including the type of plate and number of rows left empty per plate. It is wise to purchase compound already dissolved in DMSO because of considerable time require to dissolve thousands of dry compounds in DMSO. Deep well formatting instruments are required for compound re-formatting into 384 well screening stock plates as compounds are almost always shipped from the supplier in 96-deep well plates or in 96-tube racks.

Compound storage:
There is no common ideal solution for compound storage as characteristics of compounds within collections vary greatly. Usually compound stocks are dissolved in DMSO and stored frozen. Such compound stock plates are required to be stored in a desiccated environment because DMSO is hygroscopic and many compounds used in screening are not soluble at high concentration in water. To promote compound stability, number of freeze/thaw cycles experienced should be limited by fewer than 15-25 cycles.

Expensive plate storage devices and severs are available to organise and store compound stock plates under controlled atmosphere at controlled temperatures. These usually require bar-coding of stock plates and integration with other screening instruments via robotic arms.

Compound transfer:
Liquid transfers in microtitre range (transfer of library compound in into assay plates already containing cells etc) can be performed in nanolitlre range by use of carefully machined steel pin arrays, pin arrays can be rapidly washed, dried with compressed air or blotted with inexpensive paper, and reused with undetectable levels of carryover between stock plates. Pin array require considerable time for calibration but steel pin array can transfer 100nl of small molecules in DMSO into assay plates.
Advantage: steel pin transfer approach is cost effective, conserves reagents and compatible with types of assay used. It is not require purchasing and disposal of pipette tips. So consumables costs are for only methanol, which is used to wash pin arrays between transfer and blotting paper. Steel pin array may wear over time so need to send back to manufacturer approx once a year for refurbishment. Initial cost of pin array range from 5,000 – 9,000 USD, and refurbishment cost approx 500 USD.

Compound libraries are stored in DMSO. It is desirable that amount of DMSO transferred to an assay well is less than 5% of final well volume (usually, <1% preferred). As assay volumes usually range from 20-50 microL per well (for 384 well plate) approx 100nl of compound stock solution should be transferred to maintain the DMSO concentration in desirable range.

Building High Throughput Screening Facility in an academic setting , (2003) , Harvard medical school ,
http://iccb.harvard.edu/screening/


ASRS:
‘’Retrieval of the compound library for subsequent screening is termed as a Automated Storage and Retrieval System or ASRS’’. The libraries themselves need to manage in an automated way in order to effectively continue ongoing assay.

Storage of the compound library can be done in two ways
Storage in tube (vial) based system
Storage in ubiquitous micropalte format (termed as mother plate)

Later is a more common approach because tube based system require liquid handling automation at some point to transfer compounds from the tubes into microplates for screening. As compound library size vary from few thousand to several millions compound, it directly affects design and scale of ASRS system. There can be two approaches, among one if compounds are stored in vials then a 100,000 compound library require 100,000 individual vials to be stored and second If 384 well microplate then 300 such plates would be enough to store 100,000 compounds. The combination of storage & retrieval system with automation robotics outside the system serve to link compounds with downstream process. It includes liquid handling operations to transfer aliquots from motherplate to daughterplate for second approach or tube to plate transfer for first approach. Trend for ASRS system is going on to make it more modular, simple, fast & economical.

DRUG DISCOVERY AND BIOTECHNOLOGY TRENDS: LABORATORY AUTOMATION : BURSTING THROUGH THE BOTTLENECKS find the onlinelink


An effective uHTS operation cannot be performed without converting library (in vials) into a format manageable for automation. It is as essential as electronic data handling systems & transfer of libraries into microtitration plates. Barcoding by automatic system solve the problem for compound tracking. Sample preparation e.g. solubilising or storage can compromise integrity of samples. This risk must be balanced against the risk of sample transfer becoming a bottleneck.

Partial screening approach:
It enables to synthesise focused libraries
It enables rapid identification of starting points for library mining
Refinement or enrichment of particular library is possible
Target drugability (amenability to modulate its activity by small molecules)
Candidate Drug expected to bind can be tested

Focused screening approach:
It can progress a smaller number of compounds, pre-filtered by the medicinal chemists, and so that the screen only select compounds that would make a good starting point. This can be significantly quicker overall than a comparable HTS. In a current environment of aggressive patenting strategies, particularly around families of tractable targets like kinases, the speed factor is potentially a major advantage of focused screening.
Data Management in screening:
As millions of data points are generated, reliable & effective data handling software is required. Features of databases should include compound inventory and management, compound logistics, automated data processing, and identification of potential hits and integration of communication interfaces for exchanging data.

Software tools: are required to support HTS fall into two categories
1. To facilitate data collection and analysis
2. To organise and search compound structures

Software packages are require much thought and work to configure correctly so that their features are utilised appropriately.
Advantages: capable to store large amounts of data from many assays and allow comparison with other screens or access to compound structure.
Drawbacks: not feasible due to cost, personnel, or simply because the number of assays to be run is small. It is possible to work with assay data points by using much simpler and cheaper tools like Excel.

For a biologist it is common to analyse massive data and compare them with other screen data or access to compound structures. Very simple databases can be constructed to maintain analysed data provided it should be easily accessible to all users.

Plate readers: this generate data file in the text file format. So it comes in small size (usually 3KB for single plate) but enough to fill up hard-disk drive immediately.
Automated microscopes; this generate data files in image file format. So it comes in bigger size (usually 650 KB for single image). So for 20,000 image files it requires approx 51 GB of hard-disk space.Therefore, it is very important to have a server for data storage.

Average screen numbers, one sample per well can be screened, one well per sample at one concentration of compound and about 400,000 samples can be processed per screen. With the average rate of 16 compounds per 384 well plate with 200 screens result 3,200(16x200) different concentration for IC50. And total 200x384= 768,000 data points generated per day. Obviously, there is need for algorithms, which can extract best.

Most of the assay run for HTS operation have high amount of inherent variability and error. So it is wise to run all assays in duplicate whenever feasible by simply duplicating entire assay in a new set of assay plates. This sounds reliable rather than re-analysing or re-reading same assay plates twice. Advantage: dual data points from an assay allow the biologist to concentrate only on positive results detected in both assay and can result in reduction of false positive rates by upto one half.

Data Analysis:
Control readings:
For screens without appreciable time based or plate based signal to background ratio should forego the fold-induction calculation and simply be normalised on a plate by plate basis by calculating the z-score, or number of standard deviations from mean fro each readout value. Z-score can then be used as an indication of probability that a screening positive is not due to background noise.

Most of the assay readouts are time dependent so, having background and signal levels that will vary over time by plate. Any screen liable to change signal to background ratio from plate to plate should first be scaled using fold induction by dividing the observed value in each well by plate median or the palte control well medians, depending upon experimental design. Usually, plate median is more reliable to use for re-scaling or normalisation than plate mean as it is less affected by outlier values

Controls can be classified into main two classes,
1. Plate based controls: ‘ controls that are placed on each individual assay plate’. These are required to identify plate-by-plate variability and for detection of assay background level. Assay that are prone to plate-plate variability are luciferase readouts (decay over time) should make use primarily of plate based controls and normalisation. Usually stock compound plates are formatted with empty wells for the purposes of controls and it is good practice to use all available wells.
2. Assay wide controls: ‘are separate plates containing only control wells and no screening compounds’. These are required to determine background levels of an assay and should be used to determine whether an assay has sufficient signal to be reliably detected.

Considering a data point that run in duplicate having z-score of 0.5 & 2.0. The data point with a z –score of 0.5 represents an event that is about 61.7% probable to have occurred randomly, whereas the data point with a z-score of 2.0 represents an event with a 4.5 % chance of having randomly occurred. Average of these two probabilities gives 33.1% chance of random occurrence. As a comparison a data point which receives a z-score of 1.5 in both duplicates has a 13.4% chance of random occurrence in both cases making it a better screening positive to follow up.


6. ASSAY DEVELOPMENT

Assay development process for uHTS:

HTS system relies strongly on complexity of assay method. The factors are assay protocol, optimisation, adaptability to automation and availability of high quality reagents simple assay procedure is preferable.

Variables for assay:
Time, temperature & order of addition
Buffer composition, pH
Reagent concentration
Readout conditions

Assay format: 96/384/1536 well plate
Assay type: radioactive, fluorescence, absorbance
DMSO sensitivity
Reference compound: 100% effect
Homogeneous (automation-able) or heterogeneous (semi-automation)
Number of steps involved (additions, washing etc)
CV (co-efficient of variation)
Cost: reagents & consumables


Hit to miss ratio vary significantly from 0.2-16% while primary screening. The factors are assay methodology, biology developed, target selected, well plate format. Computational algorithm can help a lot in selection & prioritisation of Candidate Drug. Development of the assay in initial step should be limited to 96 well format once a successful development of signal: background ratio it can be further adopted to higher density well plate format in subsequent stages of assay optimisation & validation from HTS to uHTS. Most of the assay development can be done with time span of 3-4 months with current facilities & expertise.

Several steps are required to transfer a new assay to small-scale lab test system on a 96 well plate format of HTS. A robust assay result gives the same result in hands of any researcher. Target protein must be well characterised (validated) with respect to purity, concentration, specific activity & kinetics. First & foremost step is the determination of optimum assay conditions. Detection & automation selection is crucial but it is dictated by assay itself. Cell based assays are more complicated than enzyme assay, such assay also require separation. Final step in assay development is validation. Stability of assay is assured by investigation of intra-day variation by comparing variation between different assay runs. The purpose of screening is to avoid false negative compounds (compound active against target but not scoring in the rest) & false positive compounds (compound scoring in assay but not active against the target). Table No 5 indicates the approximate time required for assay development & validation. False negative compounds are critical when not enough hits are generated from primary screening. To rule out false positive both cost & time is required. So assay should be designed as straight forward and suitable controls must be added (0-100%). Designs of follow-up/secondary assay are needed to confirm hits retrieved from primary assay.

Table 5: typical time span for assay development & validation
Sr. No
Assay type
Time (months)
1
Cell free assay in same condition
3.7
2
Cell based assay
2.7
3
Homogeneous assay
1.9

Ideal requirements of Assay

Wedin R , (1999), One step fluorescene HTS assays are getting faster, cheaper , smaller , and more sensitive, Modern Drug Discovery ,
http://pubs.acs.org/hotartcl/mdd/99/wedin.html

Preferred approach for primary screening is to make assay well characterised, functional, non-radioisotopic, cell free assay that can differentiate agonist from antagonist and give a well defined response.

Fast: Faster assays make good financial sense. Speeding a new blockbuster drug to market can increase revenues by $1,000,000 per extra day of market exclusivity.
Less expensive: Screening materials are a significant fraction of the total screening costs. In many cases, the costs of the chemical and biological reagents can add up to three-quarters of the total screening costs.
Miniaturized: Miniaturization is an attractive way to hold down material costs. Miniaturization also can save time because potentially thousands of miniaturized assays can be conducted on one standard-sized microplate, with the liquid-dispensing nozzles and the analytical sensors moving less distance between each sample. Also, miniaturized tests can help reduce the need for laboratory space.
More automated: Automation can help with speed and reproducibility of results. Companies are using microplates with 384 or 1536 wells; they've passed the physical limit at which humans can prepare the plates. If an assay can't be automated, then it won't be feasible in the HTS environment.
More sensitive: The sensitivity of some detection methods is directly related to the volume of the sample, so smaller samples lead to diminished sensitivity. If scientists must increase the detection time to compensate for smaller sample volumes, then the method is not a good candidate for use in HTS systems.
Nonradioactive: Although trends are going toward fluorescent assays, radioactive substances require special handling of waste, which can be time-consuming, space-consuming, and expensive. In addition, some of these radiometric tests require long detection times or large sample volumes.
One-step: Test methods that require filtration, separation, washing, quenching, or the use of solid phases add extra time, money, and complexity to the process. The ideal assay is performed in one step, in solution.
Method transfer: In the ideal assay, methods developed by the assay development group are easily transferred to the HTS environment.
Flexibility: The selection of best assay mode should, ideally, be based on the biology of the target, not on the limitations of the assay technologies. Therefore, drug companies prefer technologies and instruments that can be adapted to many methods.

Current trends in assay technologies:
ELISA is a good example of heterogeneous assay where separation by washing is required. Other drawbacks of heterogeneous assay are prolonged incubation & extensive washing which make the automation more demanding. Current drive is to drive toward easily formatted-user-friendly-assays. The deluge of targets and the need to triage early in the screening process is necessary in order to ensure that the companies do not chaseover false leads for tool long with associated cost. Concerning to assay technology, drug discovery companies are shifting from heterogeneous to homogenous assay format (mix & read) more information on homogenous assay can be found on http://www.htrf-assays.com/index.php4. Homogeneous assay formats do not require any washing steps so easier to automate. Vendors in HTS space are looking at converting all their assay offerings reagents & reagent kits to be deployed into homogeneous format assays and to ‘tune’ these assays to instrumentation platforms. So consumables (plate formats & reagents) can be configured across various operating systems-unlike IT application softwares configured around a given operating system. Demand for reagents are robust and configurable into generalised assay format for biological analytes, either in vitro assays or in vivo (cell based) systems, will grow. Vendors in the space are seeking to make available reagents are focusing on those sectors such as receptor assays (specifically GPCRs), enzyme assays, cell based assays & other kind of assays.
For cell free assay plenty of targets from genomics are available but difficult to develop assay & validate as well. So cell based assays will gain more importance in secondary screening process. In next future we expect significant consolidation with hardware & reagent companies coming together with which complement eachother’s platforms and product offerings. It is also expected that understanding of the variables, which affect the screening, will increase in future.

Personnel:
Wide varieties of skills are necessary. Capabilities to tackle mechanical problems and troubleshooting of computers. One staff member should be proficient in computer language Visual Basic 6.0: to integrate operation of individual screening robots with each other.
Compound collection, maintenance, investigation and operation of screening robots and training biologist for independent operation of some machines. Significant training is required to program, run and maintain high-end LHS instruments. So such instruments are operated exclusively by screening facility staff.

Functions of Assay Development Scientist:
Assay development scientist is usually in a good position to influence HTS users.
To users: they guide about technologies & suppliers to be selected.
On a Technical side: they guide clients how to use the technology & provide them with basic information about which consumables will work the best with their technology.
On a Sales side: create a link with the key decision makers & have superior access to customer information.


Example of strategic planning for screening:
If hit to miss ratio is assumed to be 0.001 then total 100 potential hit can be retrieved from 100,000 compounds. If assay cost per compound screened is estimated to be 1.5 USD then for 100,000 compounds total cost is 150,000USD. Only the strategic planning can save the cost. It can be summarised as follows.

Primary Screening
For the goal to get 100 hits, acquisite 5,000 compounds (from in house/vendor). Screen all compounds with 0.001 hit/miss ratio. Create ChemTree model to increase tenfold hit/miss ratio 0.001 to 0.01 & generate 5 hits.(screen #1)

Secondary screening (with focused library)
Synthesise or purchase 9,500 compounds totally focused for target with hit/miss ratio 0.01. Expect 95 hits(screen #2)

Cost saving using ChemTree approach
Primary screening cost= No. of compound screened x cost per assay
= 5,000 x 1USD = 7,500 USD
Secondary screening cost = No. of compound screened x cost per assay
= 9,500 x 3 USD = 28,000 USD

Total cost of screening to get 100 hits = 36,000 USD

Total saving cost per assay = 114, 000 USD
Total annual saving for 10 assay = 1.14M USD

7. CONCLUSION
HTS appears at first glance to be a relatively straightforward activity, a sort of numbers game. Current efforts are going to minimise the amount of material consumed in HTS & easiest approach is to decrease the assay volume. Screening more compounds against more targets per unit time should generate more hits, which should generate more leads, which should generate more products.

No contemplation of high-throughput screening can be considered complete without recognition of two key factors:
(1) Drug discovery is in the midst of revolutionary and very rapid changes; and
(2) High-throughput screening cannot be considered in isolation from other aspects of drug discovery.
Findings:
Exists for great diversity in HTS technologies and systems.
Continual evolution, is driven by economic realities and pulled along by ever-increasing requirements for increased information content
The modern biologist must be capable of operating from multiple perspectives. Biologist must also be fluent in molecular biology, computer science, robotics, and instrumentation and overall process engineering.

Limitations of uHTS: It cannot directly identify a drug. Role of HTS is lead identification compounds & information for lead optimisation. Many of the properties like bioavalibility, pharmacokinetic, toxicity or specificity of CD cannot be assessed.
Some other factors are;
1. Value received for miniaturization versus resource inputs required;
2. The extent to which new technology provides value that extends beyond the primary screening process;
3. The information content provided by new technology; and
4. The technology’s “homogeneity index.”

HTS & Rational Drug Design: marriage or divorce? At first instance companies involved in rational drug design appear to be smaller in size than companies that use HTS. Big companies use the HTS not as a self-platform but as a complement to rational drug design. Combination of rational drug design with HTS can bring combination of scientific knowledge that may increase the profitability of drug discovery. There cannot be a market for HTS if target are not discovered. So, HTS is a tool being increasingly used for rational drug design. The rate with which screening technology is improving; there is a bright future it holds. It cannot be said for rational drug design.

HTS systems needed to be rapid, reliable and relatively inexpensive: technology needed to be robust, and easy to automate and minimise. HTS assays had to be sensitive enough to detect active compounds at very low concentrations. Limitation of HTS system is the capability of system to automate vast variety of targets. If HTS technologies are able to expand the base of targets they can screen, more will be the investment. Lot of technological innovations will bring improved cost structure for screening.

Future of HTS: Level of funding committed to that activity by pharmaceutical companies: determined by a complex equation for optimisation of the entire drug discovery process. Factors (cultures, strategies, and the influence borne by one or another faction).

In summary uHTS technologies has find a way so far in drug discovery industry and has indispensable element in lead discovery operations. At first instance, investing a large sum of money, energy & time can look as a big hurdle & but it should not be forgotten that it offers the unexpected rewards as well.


Glossary:
Cherry picking: ‘Secondary assay of compounds for further evaluation ‘
Cell-based assays: type of assay, which provides information on multiple parameters for a given target or multiple target proteins simultaneously in a biologically relevant context.
Cell based screening assays: use of cell based assay for primary high throughput screening
Compound validation: A process to rapidly identify drug candidates with optimal pathway selectivity as well as pharmacology and toxicity profiles.
Diversity screening: Screening with improved hit identification process, and are based on simple model of testing everything
Focussed screening: it is a successful hit generation strategy by use of an assay that is more appropriate, rather than one that works well at a large scale
High content analysis, high content assays:
Observation of multiple intracellular events in individual cells By using multiple fluorescent reporter systems, combined with high- resolution imaging and high- throughput image analysis.
High-content screening HCS:
Hit: Library component whose activity exceeds a predefined, statistically relevant threshold. [IUPAC Combinatorial Chemistry].
Or
A molecule with robust dose­ response activity in a primary screen and known, confirmed structure.
Hit to lead: Improvement of success rate by selecting & developing most promising hits generated by HTS with use of in depth assessment of chemical integrity, synthetic accessibility and functional behaviour as well as ADME features in parallel.
Homogeneous assay: assays, which require no separation steps. Pipette, incubate, and measure are the only steps required & reactions occur completely in solution generally without beads or solid phase attachments to interfere with low affinity interactions.
Lead discovery: The process of identifying active new chemical entities, which by subsequent modification may be transformed into a clinically useful drug. [IUPAC Medicinal Chemistry]
Lead identification: use of chemical genetics approaches for target identification and validation and use of virtual screening, affinity methods, and HTS analysis to identify hit. Predictive methods can highlight potential DMPK issues, and computational approaches that relate structure to biology have potential application in targeted design and compound acquisition.
Lead optimisation: The synthetic modification of a biologically active compound, to fulfil all stereoelectronic, physicochemical, pharmacokinetic and toxicologic required for clinical usefulness. [IUPAC Medicinal Chemistry]
Primary screening: use of High-content screening (HCS) technologies to approach throughput, automation, robustness, and assay development capabilities necessary to handle primary screening in a cellular environment, which is more predictive of how the drug may behave in humans. It also allows to study targets that are intractable using conventional in vitro assays. HCS in primary screening promises to increase the confidence in hits and reduce the need for secondary screens, leading to lower cost and time of drug development.
Screening: Pharmacological or toxicological screening consists of a specified set of procedures to which a series of compounds is subjected to characterize pharmacological and toxicological properties and to establish dose- effect and dose- response relationships. [IUPAC Tox]
Or
The use of in vitro biochemical assays, or tests, to detect compounds, which modulate the activity of a target (i.e., enzyme inhibitors, receptor agonists or antagonists). [Oxford Molecular]
Secondary screening: Use of the high- content cellular information on lead specificity, bioavailability, and ADME/ Tox to prioritise leads with more confidence and impact the bottom line by reducing late- stage attrition. As the primary screening technologies are becoming increasingly automated and higher-throughput, bottleneck is shifting downstream towards secondary screening and lead optimisation.
Small molecule screening: identification of chemically 'interesting' starting points for elaboration towards a drug by placing less emphasis on the number of data points that can be produced, and to focus instead on the quality of the data obtained.
Throughput: Output or production, rate at which something can be processed.

8. REFERENCES
1. Putt S. K. & Hergenrother P.J. et al (2004), An enzymatic, An enzymatic assay for poly (ADP-ribose) polymerase-1 (PARP-1) via the chemical quantitation of NAD+: application to the high-throughput screening of small molecules as potential inhibitors, Analytical Biochemistry, 326, 1, pp 78-86
2. Muckenschnabel I., Falchetto R., Mayr L. M. and Filipuzzi I. , (2004), SpeedScreen: label-free liquid chromatography–mass spectrometry-based high-throughput screening for the discovery of orphan protein ligands, Analytical Biochemistry, 324, 2, pp241-249
3. Parniak M. A., Min K. L., Budihas S. R., S. F. J. L. Grice and Beutler J. A. (2003) A fluorescence-based high-throughput screening assay for inhibitors of human immunodeficiency virus-1 reverse transcriptase-associated ribonuclease H activity, Analytical Biochemistry, 322, 1, pp 33-39
4. Wegener D., Hildmann C., Riester D. and Schwienhorst A., (2003), Improved fluorogenic histone deacetylase assay for high-throughput-screening applications, Analytical Biochemistry, 321, 2, 2003, pp 202-208
5. Barrie S. E., Amooquaye E. E., Hardcastle A., Platt G., Richards J., Bedford D., Workman P., Aherne W., Mittnacht S. And Garrett M. D., (2003), High-throughput screening for the identification of small-molecule inhibitors of retinoblastoma protein phosphorylation in cells, Analytical Biochemistry, 320, 1, pp 66-74
6. Zolli-Juran M., Cechetto J. D., Hartlen R., Daigle D. M. and Brown E. D., (2003) High throughput screening identifies novel inhibitors of Escherichia coli dihydrofolate reductase that are competitive with dihydrofolate, Bioorganic & Medicinal Chemistry Letters, 13, 15, pp 2493-2496
7. Kunapuli P., Ransom R., Murphy K. L., Pettibone D., Kerby J., Grimwood S., Zuck P., Hodder P., Lacson R., Hoffman I., (2003), Development of an intact cell reporter gene -lactamase assay for G protein-coupled receptors for high-throughput screening, Analytical Biochemistry, 314, 1, pp 16-29
8. Lin S. , Fischl A. S., Bi X. and Parce W. , (2003), Separation of phospholipids in microfluidic chip device: application to high-throughput screening assays for lipid-modifying enzymes, Analytical Biochemistry, 314, 1, pp 97-107
9. Hodder P., Cassaday J., Peltier R., Berry K., Inglese J., Feuston B., Culberson C., Bleicher L., Cosford N. D. P., Bayly C., (2003), Identification of metabotropic glutamate receptor antagonists using an automated high-throughput screening system, Analytical Biochemistry, 313, 2, pp246-254
10. Berdichevsky Y., Zemel R., Bachmatov L., Abramovich A., Koren R., PeramachiSathiyamoorthy, Golan-Goldhirsh A., Tur-Kaspa R.and Benhar I., (2003), A novel high throughput screening assay for HCV NS3 serine protease inhibitors, Journal of Virological Methods, 107, 2, pp 245-255
11. Li Y., Cummings R. T., Cunningham B. R., Chen Y. and Zhou G., (2003), Homogeneous assays for adenosine 5'-monophosphate-activated protein kinase, Analytical Biochemistry, 321, 2, pp 151-156
12. Zawadzke L. E., Wu P., Cook L., Fan L., Casperson M., Kishnani M., Calambur D., Hofstead S. J. and Padmanabha R. (2003), Targeting the MraY and MurG bacterial enzymes for antimicrobial therapeutic intervention, Analytical Biochemistry, 314, 2, pp243-252
13. N. Yamamoto, H. Hasegawa, H. Seki, K. Ziegelbauer and T. Yasuda, (2003), Development of a high-throughput fluoroimmunoassay for Syk kinase and Syk kinase inhibitors, Analytical Biochemistry, 315, 2, pp 256-261
14. Wu, Ge; Irvine, Jennifer; Luft, Chris; Pressley, David; Hodge, C Nicholas; Janzen, Bill, (2003), Assay development and high-throughput screening of caspases in microfluidic format, Combinatorial Chemistry & High Throughput Screening, 6, 4, pp 303-312
15. Chambers, Chris; Smith, Fiona; Williams, Christine; Marcos, Sandra; Liu, Zhen Han; Hayter, Paul; Ciaramella, Giuseppe; Keighley, Wilma; Gribbon, Phil; Sewing et al. (2003), Measuring intracellular calcium fluxes in high throughput mode, Combinatorial Chemistry & High Throughput Screening, 6, 4, pp 355-362
16. Eglen, Richard M; Singh, Rajendra , (2003) , Beta galactosidase enzyme fragment complementation as a novel technology for high throughput screening, Combinatorial Chemistry & High Throughput Screening, 6, 4, pp 381-387
17. Wee1 and Myt1 kinases
18. Zaman G.J.R.; Garritsen A.; Boer T.de; van Boeckel C.A.A. (2003), Fluorescence Assays for High-Throughput Screening of Protein Kinases , Combinatorial Chemistry & High Throughput Screening, 6, 4, pp. 313-320(8)
19. Drees B.E.; Weipert A.; Hudson H.; Ferguson C.G.; Chakravarty L.; Prestwich G.D. (2003), Competitive Fluorescence Polarization Assays for the Detection of Phosphoinositide Kinase and Phosphatase Activity, Combinatorial Chemistry & High Throughput Screening, 6, 4, pp. 321-330(10)
20. Pastula C.; Johnson I.; Beechem J.M.; Patton W.F. (2003), Development of Fluorescence-Based Selective Assays for Serine / Threonine and Tyrosine Phosphatases, Combinatorial Chemistry & High Throughput Screening, 6, 4, pp. 341-346(6)
21. Heyduk E.; Fei Y.; Heyduk T., (2003), Homogeneous Fluorescence Assay for Cyclic AMPCombinatorial Chemistry & High Throughput Screening, 6, 4, pp. 347-354(8)
22. Gee K.R.; Rukavishnikov A.; Rothe A.,(2003), New Ca2+ Fluoroionophores Based on the BODIPY Fluorophore ,Combinatorial Chemistry & High Throughput Screening, 6, 4, pp. 363-366(5)
23. Graeff R.M.; Lee H.Cheung, (2003), High Throughput Fluorescence-Based Assays for Cyclic ADP-Ribose, NAADP, and Their Metabolic Enzymes, Combinatorial Chemistry & High Throughput Screening, 6, 4, pp. 367-379(13)
24. Waller A.; Simons P.; Prossnitz E.R.; Edwards B.S.; Sklar L.A. , (2003) , High Throughput Screening of G-Protein Coupled Receptors via Flow Cytometry , Combinatorial Chemistry & High Throughput Screening, 6, 4, pp. 389-397(9)