Unit
1: The
Science of Animal Research, Part 1
II
Basic Experimental Design
 Recommended:
Experimental
Design and
Selection
of Animal Models
For more info on Banting & his friends:
Discovery of
Insulin
|
|
|
“The use of
animals as research substrates mandates that the situation be controlled
at all times. When all parties understand the importance of their
roles in the experimental situation, accidents are minimized, the animals
are less anxious, procedures are...less painful, and experimental success
is more likely."
Sheldon Scher
("Setting Up an Animal Experiment"
in Veterinary Technician 7(5): 217) |
-
Influences on Experimental Outcomes
The design of an experiment has a critical effect on its outcome and is therefore extremely important. Every component
of the experiment must be explicitly described and carefully planned in
advance. An improperly designed study may yield results
that can't be interpreted or used.
The welfare of the animals must be considered and
maximized during every phase of the design process.
During the experimental design process, it is important to minimize:
-
Nonexperimental factors that may
interfere with the test results (such as room temperature
variations)
-
Bias that predetermines the test
results (i.e. we see the result that we expected because we are looking
for it)
-
Discovery of the Etiology and Treatment of Diabetes
Mellitus
The ancient Romans knew that high levels of
sugar in blood and urine were associated with severe illness and
death. They named the disease "diabetes mellitus"--sweet
water. They, and physicians through the centuries that followed,
did not know the cause (etiology) of this disease nor had they any way
to treat it.
Research during the nineteenth century revealed that removal of the pancreas caused diabetes
mellitus. Additional studies showed that small groups of cells,
called the islets of Langerhans, produced a critical anti-diabetic
substance. In 1889, Oscar Minkowski wrote about an experiment in
which he ground up fresh pancreas and injected it into diabetic
dogs. No improvement occurred and research stalled.
In 1920,
Dr Fred Banting read Minkowski's paper while writing a lecture on pancreatic physiology at the
University of Western Ontario. He
hypothesized that digestive enzymes produced in the pancreas of
Minkowski's dogs caused the
destruction of the anti-diabetic substance. He believed that ligating
(tying off) the pancreatic duct would cause the enzymatic portion of the
pancreas to atrophy (waste away) without damaging the rest of the organ. Dr
J.J.R. MacLeod was the chairman of the
Physiology Department at the University of Toronto and an expert on glucose
metabolism. Banting presented his ideas to MacLeod and asked for
his assistance in performing an experiment to test his hypotheses.
Although he doubted that Banting was correct, MacLeod provided him with
laboratory space, along with ten experimental dogs and Charles Best, a
medical student.
Banting and Best surgically removed the pancreas in several of the dogs
(causing them to become diabetic). They ligated the pancreatic duct in the
other dogs; after six weeks, two of the dogs were anesthetized, but
surgery revealed that the duct had not been properly ligated and the
enymatic pancreas had not degenerated. Banting carefully re-ligated the
pancreatic duct on these dogs. After an additional six weeks, he
was able to harvest atrophied pancreatic tissue, which he ground up with
saline solution and injected into the vein of one of the pancreatized
diabetic dogs. This injection decreased the blood sugar level and
improved the health of the dog.
Dr MacLeod was impressed but not convinced by Banting's research:
he required the entire procedure to be repeated under his
supervision. The additional experiments confirmed the initial
results and Dr MacLeod announced the discovery of "insulin" at
a meeting in Toronto on November 14, 1921. J.P. Collip, a biochemist, joined the team to increase the purity of the
insulin extract. In January of 1922, insulin extract was injected
for the first time into a human, Leonard Johnson, a 14-year-old boy in a
diabetic coma. This treatment saved Leonard's life and established
an effective treatment for this previously fatal disease.
-
Steps in Designing an Experiment
Any experimental
design must be complete, detailed and reproducible:
-
As a building block to lead to further
research
(Banting built on the research of previous
experimenters. After reading Minkowski's research, Banting
decided to ligate the pancreatic duct, which kept the
insulin-producing cells alive but stopped the production of
pancreatic enzymes).
-
For confirmation--to ensure that the
results are valid
(Dr MacLeod required Banting to repeat the entire
experiment. Without detailed description of all of the
procedures,
such reproduction is difficult or even impossible).
The first step in
designing an experiment is to identify and define a problem, setting
achievable goals (Dr Banting knew about diabetes mellitus and the
causative role of the pancreas--he believed that he knew how to isolate
the anti-diabetic chemical that had eluded previous researchers).
After defining the problem, there are a number of additional steps
needed to develop a research plan:
-
Literature
search:
It is essential to find out everything that is known about the
problem and to ensure that the research is not unnecessarily
repeating an already proven study (Banting and Best
worked together to compile a complete picture of known
information).
-
Develop a
hypothesis:
A testable question must be clearly defined (Banting
hypothesized that ligation of the pancreatic duct would allow him to
collect a chemical that
would
decrease blood glucose in dogs with diabetes mellitus).
-
Select the appropriate model:
A model is one animal species used to represent another, usually
man (Banting used dogs as his model). Statistical
analysis should be used to determine the appropriate number of
animals to use to ensure that the results are valid.
-
Develop
a set-by-step protocol for the experiment:
The exact procedures must be described, including how to do
them, when to do them and by whom (Banting was responsible
for performing the pancreatic duct ligation and Best determined blood glucose
levels).
-
Describe
the method for data collection and evaluation:
The equipment for data collection and analysis, such as blood
chemistry machines and computer programs, must be evaluated
and selected.
-
Develop
a budget:
The anticipated costs must be calculated and funding must be
obtained. It is important to ensure not only that there
are animals; sufficient food and bedding,
plus salaries and equipment must be available (Banting, after the failure of
the initial ligatures, sold his truck to obtain enough money to
prolong the experiment for the additional weeks needed).
-
Specify
personnel qualifications:
According to the United States Animal Welfare Act regulations,
training, certifications and licenses must be identified and
described for all
personnel.
-
Animal
Care and Use Plan (ACUP):
The ACUP is a document required by the Animal Welfare Act and describes all activities involving animals,
including the test procedures, pain control methods and basic
husbandry. It must be submitted and approved by the
Institutional Animal Care and Use Committee of the facility
where the research is being performed.
Only after the
experimental plan is completed, submitted and approved can research
begin. Statistical
analysis of the experimental results is used to determine if the results
are significant. It may be necessary to revise
and revise again when unexpected results occur (Banting
essentially had to start over when his early surgeries did not cause the
effects that he had anticipated).
-
Characteristics of Experiments
Any factor or condition that can change is called a
"variable."
An experimental
variable is something that is selected to affect the result of a study: it
causes a reaction. Banting's variable was the procedure of
ligating the pancreatic duct that he believed would allow him to extract
the anti-diabetic substance.
A nonexperimental variable produces results that we do not want to
measure, and is usually the result of environmental or genetic
factors. Nonexperimental variables may invalidate research and should be minimized by careful
planning and adherence to the research protocols. Any
nonexperimental variable should be reported to your supervisor and
remedied if possible. There are several general types of
nonexperimental variables, including:
Chemical (ex. ammonia from
urine, pesticides & cleaning chemicals)
Physical (ex. change in temperature, loud
noises or overcrowding)
Pathogenic (ex. disease-causing viruses,
bacteria & parasites)
Genetic (ex. accidental matings of different strains of mice)
Biologic (ex. pain and
distress)
Animals in a research
protocol are categorized by subject group. Animals in the experimental group receive
the treatment, such as Banting's pancreatized dogs that were injected
with the insulin extract.
It is also important to have a
control group. These animals should be cared for in the same way as
the animals in the experimental group, except for the specific treatment.
This provides a standard for comparison with the test group. In
Banting's experiment, a control group of dogs could have been injected
with a placebo, such as saline solution, to ensure that the
injected substance and not the injection itself caused the change in
blood glucose. Assignment of an
animal to a subject group should be random and blind, to eliminate bias
(seeing the results we expected) and possible misinterpretation of results. Random selection is
achieved
when numbers are drawn out of a hat or a table of random numbers are
used to assign animals to the control and experimental groups.
In a blind experiment, the person observing the results does not know if
the animal is in the experimental or control group.
- The Influence of the Technical Animal Personnel
on Experiments
|
 |
Laboratory animal personnel have critical roles in maintaining the integrity of
experiments.
Animal husbandry is the responsibility of
the animal technicians. It is essential that the husbandry
procedures be standardized and meticulously adhered to. Different frequencies of cage cleaning
may affect the environment within a cage, with the cage cleaned less
frequently having higher levels of ammonia and pathogenic
bacteria. Favoritism toward certain individuals, such as a
preferred dog, may affect that animal's response to the experimental
procedure.
|
In addition, errors made during technical
procedures, such as incorrect weights, misidentification of animals or
careless recording of information, can compromise the validity of an
experiment.
-
Animal
Models
An animal model is the use of one living system, such as a dog, to study
another living system, usually man. The model is analogous
to man, not identical.
|
|
Mongolian gerbils, for example, are prone
to audiogenic seizures--they suffer from epileptic-like seizures as a
response to loud noises. We use them as a model to study epilepsy
in although, audiogenic seizures are not identical to human
epilepsy. They are similar enough, however, to apply data about the
effect of anti-epileptic drugs on gerbil seizures to humans. |
Recently, many models have been created
in animals through genetic engineering, creating transgenic
primates that contain human DNA and are susceptible to HIV or knock-out
mice that lack specific genes, rendering them more prone to
cancer.
There are many advantages to using
animals instead of people in experiments:
There are moral, ethical and legal restrictions
on human experimentation.
There are moral, ethical and legal restrictions
on human experimentation
There are adequate numbers of
genetically similar animals available
It is less expensive to obtain and
maintain animals
Better standardization and control of
the environment and genetics
The disease state can be manipulated
The shorter lifespan allows simulation
of "long-term" effects in less time
There is unlimited access to
specimens, such as blood and tissues
There
are also many
disadvantages to using animals, however:
Extrapolation of animal data to humans
may be unreliable
There are species variations in
anatomy, physiology and metabolism between
humans and animals
The results and their interpretation
are limited to the standardized conditions
of the study
Concern about animal welfare and
animal rights
To
minimize some of these possible disadvantages, a plurispecies approach is
recommended. This means that more than one species of animal
should be used in assess the effectiveness of an experimental
procedure. Penicillin was first administered to a
guinea pig--and it died. Fortunately, penicillin was also tested
on other species of animals and proven to be a potent antibacterial
drug. The initial selection of the guinea pig as a model was not
appropriate for the test.
After a model has been tentatively
selected, it should be validated. Validation is a formal
testing procedure that proves the model is significantly similar to the
human system it represents--it helps prevent inappropriate "guinea
pigs".
|
-
Examples of Biomedical Research:
There are many types of research and many different procedures have been
developed to effectively utilize animals. Common types of animal
research protocols are toxicology studies, experiments that assess
immune function, cancer research and new drug development.
Toxicology
Toxicology is the study of poisons.
Man has identified or invented hundreds of thousands of different
chemicals. It is important to determine the deleterious effects of
the chemicals, the safeguards necessary to prevent harm, and the levels
of these substances that are safe or dangerous.
Acute toxicity tests determine the effect
of a single dose of a substance. The Draize
test is an example of an acute toxicity test. In this test,
a chemical is applied to the skin or eye of an animal and the level of
irritation is determined.
The LD50 is another
acute toxicity test, used to determine the dose that kills 50% of
animals in a trial. . An
LD50 is the weight of drug in milligrams per kilogram of
animal. Some sample LD50 include:
-
Sugar:
27,700.0 mg/kg
-
Salt:
3,000.0 mg/kg
-
Caffeine:
192.0 mg/kg
-
Cyanide: 6.4 mg/kg
(Later lectures will address alternatives to these
experiments)
|
Development of New Drugs
|
|
The research process is complicated,
time-consuming, and costly and the end result is never
guaranteed. Literally hundreds and sometimes thousands
of chemical compounds must be made and tested in an effort to
find one that can achieve a desirable result.
-F.D.A. CDER
Handbook |
The US Food and Drug Administration estimates that it
takes over eight years and hundreds of millions of dollars to market a new drug.
The federal government closely regulates each of the many steps
involved in this process.
There are many different methods by which drugs are
"discovered." Researchers study the pathophysiology
of diseases, determining biochemical reactions that contribute the
disease development. Further research seeks to discover drugs
that counteract these chemicals.
Other researchers may discover that a currently
known drug works for a totally unrelated problem. For
example, thalidomide, which was prescribed for morning sickness in
the 1950s (causing horrific birth
defects), is now being evaluated for use in treating
multiple
myelomas and signs associated with
HIV.
Most drug research begins in the chemistry or
computer lab. Computers are used to predict the effects of a
potential drug on a chemical reaction. In the chemistry lab,
chemicals are mixed and the reactions are studied. Chemicals
are assayed to determine their composition, purity and effect.
Many of these chemicals are added to cell or tissue cultures. After
research has indicated that a chemical might be effective, it
is tested in animals. This testing has several purposes:
To determine if the chemical may have toxic side
effects (is it harmful?) To determine the efficacy of the chemical on the
disease process (does it work?) To determine how much of the chemical is
needed (what is the dose?) To determine the effect of the body on the
chemical (ex. how much is absorbed?)
Animal testing usually lasts several months to years
and more than one species of laboratory animal is often used, because
different animals may react differently to a drug. Only
if a chemical is proven safe and effective in animals do
human clinical trials begin. Clinical trials consist of four
phases:
Phase 1- safety evaluation (tested
in 20 -100 people for several months) Phase 2- efficacy evaluation (up to
several hundred people for months to 2 years) Phase 3- evaluate safety, effectiveness
and dosage (several hundred to several thousand people for 1
- 4 years) Phase 4- postmarketing evaluation of
safety, efficacy and dosage (on-going monitoring of drug after it
is available for general use)
Each species of animal is unique, and has biochemical
and physiological processes that may be different from those of
humans. Nonhuman animals are used in place of people during
early testing so that a chemical can be fully evaluated before
exposing people to it. It is important to understand and
recognize that humans and other animals are not identical: a
positive test result on animals does not mean that identical results
will occur during the clinical trials. A negative result from
animal testing, however, can prevent subsequent trials on people (such
as your mom or your kid), that may have harmful effects.
Return to the top of the page
|