Model Tumor Systems

1
Model Tumor Systems

• Sara Rockwell, Ph.D.
• Departments of Therapeutic Radiology and
  Pharmacology
• Yale University School of Medicine
• New Haven, CT
• Residents course, February 5, 2009

2
References

• E. J. Hall and A.J. Giaccia, Radiobiology for the
  Radiologist, 6th edition, Chapter 20
• S. Rockwell and K. R. Rockwell, Mouse Models for
  Experimental Cancer Therapy, in Sourcebook of
  Models for Biomedical Research
• R.F. Kallman (ed.) Rodent Tumor Models in
  Experimental Cancer Therapy.

3
Cells in culture as models for tumors

• Uniform, well defined cell populations
• Good quantitative endpoints for cell survival
• Very useful model systems for some studies
• Examine effects of proliferation, cell cycle
  phase, etc.
• Examine effects of cellular environment
• Examine mechanisms of action
• Examine interactions between drugs and radiation
  and establish the mechanisms underlying such
  interactions
• Provide insights into effects of sequence, time
  and dose on effects of single agent and combined
  modality treatments

4
What can we learn from cell cultures ?

• Many things
• Study biology of cells, how they grow, how they
  interact
• How they metabolize drugs
• How they respond to treatment with drugs and/or
  radiation
• How therapeutic agents interact
• Example Breast cancer cells in culture, treated
  with Adriamycin, alone or in combination with a
  commonly used herbal medicine called black
  cohosh.
• Question what happens to tumors and normal
  tissues in vivo?

5
Tumor cell lines in culture differ from tumor
cells in vivo

• adapted to survive and grow in culture
• altered proliferation
• may have been cloned
• altered gene expression, enzyme activity
• altered shape and motility
• altered metabolism
• altered differentiation
• altered response to external signals

6
The environment of cells in ordinary cell
cultures differs from that of cells in vivo

• Limited or no contact with similar cells
• Limited or no contact with other cell types
• Oxygen levels are high
• Nutritional environment artificial and limited
• pH
• Cytokines other external signals
• Temperature, motion, growth surface, etc.

7
Some culture systems are better models for
tumors in vivo

• Primary cell explants
• Three dimensional cultures
• Perfused cultures
• Physiological growth surfaces
• Co-cultures containing multiple cell types
• Tissue and organ cultures
• None of these fully model tumors in vivo - they
  are all still models, with inherent limitations

8
Spheroids Sutherland et al.
9
(No Transcript)
10
Human cells are sometimes needed

• E.g. in studies of human cytokines, antibodies,
  or gene therapy
• But remember that cultures of human cells are
  also artificial model systems that do not fully
  model human tumors in vivo
• Malignant cell lines are heavily selected
• Normal cells are probably not normal if they
  grow well in culture, even if they have been
  cultured only for a limited period of time

11
Animal tumor models are closer, but still have
limitations as models for human patients

• Mice are not furry little people
• Neither are rats
• Or cats
• Or dogs
• Or non-human primates
• In vivo model systems must also be chosen with
  care to ensure that the model is appropriate for
  the specific question being addressed

12
Large animals with tumors are used only rarely in
experimental cancer therapy

• Some studies with veterinary patients
• Generally designed like human clinical trials
• Have all limitations of human clinical trials
• Potentially curative regimens
• Tolerable regimens
• Clinical grade drugs
• Oversight as strict as human studies (possibly
  stricter)
• Multiple layers of oversight IACUC, NIH,
  USDA, AAALAC, FDA

13
Other problems with large animal studies

• Vet patients more difficult to treat
• Anesthetized for irradiation
• Anesthetized for imaging
• Post procedure care and monitoring
• Patient variability is greater than in human
  clinical trials
• Characteristics of tumors may differ from those
  of human tumors of same origin
• Breast cancer in dogs
• Grey horse melanoma
• Dose-limiting toxicities may differ from those in
  people

14
Mice and rats are the most common models

• Why? Because inbred strains are available
• Genetically uniform - all identical twins
• Well defined phenotypes
• Increasingly well defined genotypes
• Some have strain-specific tumors at high
  frequencies (sometimes in essentially 100 of
  the animals)
• Tumors can be transplanted within an inbred
  strain
• Early passage tumors
• Serially transplanted tumors

15
Limitations of transplanted tumors must be
remembered in translating results to develop
treatment for cancer in people

• Artificial model system
• Inoculation of single focus of malignant cells,
  often at site selected for convenience or to
  minimize stress or injury to host
• Cells often selected for rapid growth, high
  clonogenicity
• Sometimes selected for other features
• Metastatic rate/pattern
• Response to specific agents
• Ability to grow in vitro
• Presence of a specific gene or marker

16
Some considerations in preclinical cancer therapy
studies using animals

• Activation/metabolism/clearance of drugs and drug
  carriers vary with species and substrain
• Different mouse strains can be very different
• Drug metabolism may also vary with husbandry
  (e.g. bedding in rodent cages) and microbiology
• Pharmacokinetics, biodistribution, clearance may
  be very different in rodents and people
• Area under curve generally longer in people
• Maximal attainable peak tumor levels may be lower
• Can lead to inaccurate predictions of efficacy in
  humans
• Severity and patterns of toxicity may differ in
  mice and people

17
More considerations in preclinical cancer therapy
studies using animals

• Good Microbiology is critical to good science
• Radiation and anticancer drugs are
  immunosupressive - subclinical infections can
  become clinical or lethal after treatment
• Marrow
• Gut
• Lung
• Subclinical infections can change tumor growth
  and response of tumors to treatment
• Subclinical infections can change proliferation
  patterns in bone marrow and gut and therefore
  change response to treatment and toxicity of
  treatment
• Tumors and cell lines can carry and transmit
  pathogens this poses hazards to the mice and to
  people
• Use of specific pathogen free (SPF) mice is
  critical in experimental cancer therapy.

18
Still more considerations in preclinical cancer
therapy studies using animals

• Match tumor burden to that in patients
• Common problem treating unrealistically large
  tumors in mice
• Growing mice are not good models for adult people
• Mice are often sold very young (soon after
  weaning)
• Juvenile mice respond differently to drugs and
  radiation than adult mice proliferation
  patterns in growing tissues hormones
• Young mice can be unstable models a week of
  difference in age from batch to batch can make a
  huge difference in effect.

19
Still more considerations in preclinical cancer
therapy studies using animals

• Modeling of clinical treatment regimens requires
  different treatment times in mice
• Cell proliferation rates (tumors and normal
  tissues)
• Daily treatment in people doesnt equal daily
  treatment in mice
• Modeling of tumor and normal tissue responses to
  therapy requires different follow up times in
  mice
• Mice have a maximum life span of 3 years
• Mouse tumors have doubling times of days
• 3 months of follow up in mice equals years of
  follow up in people

20
Human tumor xenografts in immune deficient mice
widely used model for human cancers

• Only the malignant cells are human
• Tumor cells have adapted for rapid growth in mice
• The stroma and vascular bed are mouse
• The other normal tissues are mouse
• Pharmacokinetics, biodistribution clearance are
  mouse
• Activation and metabolism of drug may reflect
  metabolism by mouse cells
• Hosts have other defects which affect results
• Nudes thermoregulation
• SCIDs - DNA repair defects make the vascular bed
  and stroma sensitive to injury and alter the
  balance of direct and indirect tumor cell killing
  from that found in a syngeneic system

21
Special traps with the use of human tumor
xenografts in immune deficient mice to study
human-specific agents (e.g. MAb, siRNA)

• Only the malignant cells are human
• Tumor stroma is mouse
• Vasculature is mouse
• Normal tissues are mouse
• Pharmacokinetics, biodistribution, targeting of
  tumor-specific agents will not resemble those in
  a person, because only the tumor will have the
  target
• High tumor levels
• Little or no accumulation in normal tissues
• May have unrealistically large effects on
  xenografts
• May have little toxicity to normal (mouse)
  tissues
• Toxicities in the mouse tissues probably will not
  predict toxicities in human patients
• No way to predict therapeutic ratio in patients

22
Another trap - transplanted tumors in genetically
engineered mice (GEM)

• 50 generations of inbreeding are needed to
  produce inbred mouse lines
• Many GEM lines have been inbred only 5-10
  generations from founders with outbred or mixed
  genetic backgrounds
• GEM may be uniform at the locus of interest,
  because this is verified, but they are not
  uniform at many other critical loci
• Not homozygous not syngeneic
• Transplants of tumors arising in GEM are
  problematic
• Studies using tumor lines transplanted from a
  parental mouse strain are also very problematic

23
Assays of tumor response

• Tumor cell survival assays
• TD50 assay
• Lung Colony Assay
• Spleen colony Assay
• Tumor control assay (TCD50 assay)
• Tumor growth assay

24
Tumor cell survival assays

• Cell culture assay for tumor cell survival
• Analogous to assays of cell survival for cells
  grown and treated in culture
• Technique
• Prepare treated and control tumors
• Suspend cells from tumors
• Count tumor cells
• Plate cells at low density in cell culture
• Allow viable clonogenic cells to grow into
  colonies
• Count colonies, calculate plating efficiency ,
  calculate Surviving Fractions

25
Example Effect of Hypoxia on Radiation Response
of Cells in Solid Tumors
1.0
Hypoxic cells in vitro Hypoxic cells in tumors ?
0.1
Surviving Fraction
0.01
Normally aerated tumors in situ
0.001
Aerobic cells in vitro
0
10
20
30
Dose (Gy)
26
Strengths of cell survival assay

• Highly quantitative
• Good precision, reproducibility
• Not influenced by late host toxicities of
  treatment
• Measures cell survival directly
• Can be used to examine subpopulations of tumor
  cells
• Can be used to probe mechanisms
• Considers only clonogenic cells (not influenced
  by response of differentiated or non-clonogenic
  cells)

27
Problems with cell survival assay

• Not applicable to all tumors many tumor cells
  wont clone in vitro
• Requires preparing a good single cell suspension,
  with high viability can be a technical
  challenge
• Requires removing cells from tumor
• Environment of cells changes from that of cells
  remaining in vivo
• Some cells clonogenic in vivo may not be
  clonogenic in vitro (or vice versa)
• Limited to relatively low radiation doses
• Problems in measuring effects of fractionated or
  protracted treatments
• Cell loss (normal loss from differentiation or
  environment)
• Apoptosis and rapid cell death after treatment

28
Lung Colony Assay and Spleen Colony Assay

• Colony formation assays Analogous to cell
  culture assay discussed above and to normal
  tissue assays discussed last session
• Treat tumors
• Prepare single cell suspensions
• Inject cells into tail vein of syngeneic mouse
• Cells lodge in lung (certain carcinomas and
  sarcomas) or spleen (certain lymphomas) and form
  colonies
• Allow colonies to grow
• Kill animal, remove and fix tissue, count
  colonies arising from viable tumor cells

29
Strengths of and problems with these assays

• Similar to those for cell survival assays in
  culture
• Additional problems
• Requires good syngeneic system an immune
  response to the tumor can invalidate the assay
• Requires good microbiology (Specific Pathogen
  Free SPF host animals)
• Labor intensive
• Expensive

30
TD50 Assay

• Hewitt and Wilson first tumor cell survival
  assay
• Was used with wide variety of lymphomas and solid
  tumors
• Less widely used now (mostly because of expense,
  availability of cell culture assays)

31
Basic technique

• Prepare treated, control tumors
• Suspend cells from tumor
• Count cells
• Inject serial dilutions of cells into one or more
  isolated sites on syngeneic recipient mice
• Watch for development of tumor at each
  inoculation site
• Calculate (by probit or logit analysis) number of
  tumor cells required to form tumors in 50 of the
  inoculation sites TD50
• Use TD50 for treated and control tumors to
  calculate the Surviving fraction for tumor cells
  in treated animals
• SF TD50 treated cells / TD50 control cells

32
(No Transcript)
33
Some things to note about these curves

• Sigmoidal shape
• Refects probabilistic endpoint of tumor formation
• Heterogeneity in tumor inocula, host mice or
  recipient mice decreases the slope of the take vs
  cell number curve
• Need range of tumor cell inocula
• Need several mice per point
• Need long follow up

34
Stengths of TD50 assay

• Quantitative
• Not influenced by late toxicities of treatment to
  original host mouse
• Measures cell survival
• Can be used to examine subpopulations
• Can be used to probe mechanisms
• Considers only clonogenic cells (not influenced
  by response of differentiated or non-clonogenic
  cells)

35
Problems with TD50 Assay

• Very expensive and labor intensive
• Requires removing cells from tumor
• Environment of cells changes
• Some cells clonogenic in tumor may not be
  clonogenic in site of inoculation (or vice versa)
• Less precise than other cell survival assays
• Problems in measuring effects of fractionated or
  protracted treatments
• Cell loss
• Apoptosis and rapid cell death

36
Tumor growth delay assay

• Inoculate tumors into a large number of animals
• Identical hosts
• Quantitative injection of tumor cells
• Site readily accessible for repeated
  measurements
• 5-20 animals per treatment group
• Identify each animal
• Measure tumor volumes beginning when tumors first
  become palpable
• Daily to weekly
• Frequency depends on growth rate

37

• At specific volume, randomize into treatment and
  control groups
• Treat
• Continue monitoring growth until each tumor
  reaches a defined size
• Calculate time for each tumor to grow from the
  treatment volume to this size
• Calculate growth delays induced by treatments

38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
Strengths of assay

• Clinically relevant endpoint but it measures
  the success of failed treatments because all
  tumors must regrow for the assay to be valid
• Need little prior knowledge of treatment efficacy
  to plan experiments
• Relatively precise endpoint (if done well large
  numbers of similar tumors frequent, precise
  measurements, etc. )
• Can be used with fractionated and protracted
  treatment regimens

42
Problems with assay

• Uses non-curative treatments - all treatments
  fail
• Requires syngeneic tumor/host system or
  immunosuppressive effects of treatment will
  influence the results
• Cannot be used with tumors that metastasize early
• Labor intensive and expensive
• Toxicities to host can limit assay
• Lethal toxicities of some drugs preclude using
  doses that perturb tumor growth
• Drug toxicities may alter tumor growth
  indirectly, complicating interpretation of data
• An amazing amount of really bad, misinterpreted
  tumor growth data can be found in the literature

43
Some problematic variations of tumor growth
studies (discussed only as a warning, not as a
recommendation for taking these shortcuts)

• Kill all mice at a predetermined time, measure
  tumors (Classic chemotherapy approach gives
  volumes as T/C)
• Kill all mice when the control tumors get as big
  as the animal care committee will let them grow
• For relatively successful treatments, treated
  tumors may still be small or even undetectable
• Cannot distinguish control from slow recurrence
  or slow growth
• Measure time until tumors kill mice
• Gives growth delay as T/C
• For solid tumors, causes of death vary in
  different mice
• Inhumane

44
Tumor control assay

• Called by various names
• Tumor control dose 50 (TCD50 assay)
• Tumor cure dose 50 assay
• Effective dose 50 (ED50)
• Basic approach to determine the dose of
  radiation (alone or in combination with other
  agents) that controls 50 of a population of
  identical tumors

45
TCD50 Protocol

• Inoculate tumors into a large number of animals
• Quantitative injection
• Site readily accessible for repeated measurements
• 5-20 animals per treatment group
• Identify each animal
• Measure tumor volumes beginning when tumors first
  become palpable
• Measurement frequency depends on growth rate
• Daily to week
• At specific time or volume, randomize into
  treatment and control groups

46
TCD50 protocol, continued

• Treat with range of doses ranging from doses
  where 100 recurrence is expected to doses where
  100 control is expected
• Continue monitoring the volume of each tumor
  until
• It grows to a predetermined volume and is deemed
  not cured
• It regresses completely and has been gone for a
  long enough time that it is statistically
  unlikely to recur (time depends on tumor usually
  gt 3 months). It is then considered controlled
• Develop dose-response curves proportion of
  tumors cured as a function of dose
• Calculate TCD50s for using probit or logit
  analyses
• Compare TCD50s for different regimens

47
(No Transcript)
48
Comments on TCD50s

• Last slide showed data on probit scale
• On a linear scale, the curve has a sigmoidal
  shape
• Refects the probabilistic endpoint
• Heterogeneity in treatments, tumors or mice
  decreases the slope of the tumor control vs dose
  curves
• Uniformity of mice, tumors, and treatment
  regimens is essential for good data

49
Strengths of assay

• Clinically relevant endpoint
• Viewed as best assay by some regulatory
  agencies and clinical trial groups
• Applicable to radiation dose ranges and treatment
  intensities of clinical interest
• Can be used with fractionated and protracted
  treatments
• Can be relatively precise if experiment is done
  right (large numbers of animals, well matched
  tumors, etc)

50
Problems with assay

• Requires syngeneic tumor/host system
• Requires non-metastatic tumors
• Requires good microbiology in animal colony
• Exceedingly labor intensive and expensive
• Each TCD50 requires 50 animals, most of which
  will be followed for months
• Requires prior information on treatment agents
  used to chose curative doses for assay
• Requires cures to obtain any information
• Therefore if combining radiation with drug that
  does not cure tumors alone, cannot measure effect
  of drug alone
• Therefore Usually cannot be used to determine
  whether effects are additive, synergistic, etc

51
Conclusions

• In vitro and in vivo models can provide valuable
  insights into the therapeutic responses of solid
  tumors
• All model systems and all assays have limitations
  none are perfect models for human cancer
  patients
• Watch for poorly designed studies, misinterpreted
  studies, and erroneous conclusions as you read
  the literature

52
(No Transcript)