Ecology, evolution, and antibiotic resistance

1
Ecology, evolution, and antibiotic resistance
Carl T. Bergstrom Department of
Biology University of Washington
University of MichiganDecember 8th, 2005
2
Humankind has conquered infectious disease.
3
(No Transcript)
4
The SARS virus
The SARS virus
5
H5N1 Avian Influenza
6

• The New York Times June 13, 2000
• Antibiotic Misuse Turns Treatable to Incurable

7
Vancomycin-resistant Enterococcus in US hospital
intensive care National Nosocomial Infections
Surveillance System Report, 2003
8
How evolution works

• Variation different individuals have
• different traits.
• Heritability offspring tend to be somewhat
  like their parents.
• Selection individuals with certain traits
• survive better or reproduce more.
• Time successful variations accumulate
• over many generations.

9
From Battling bacterial evolution The work of
Carl Bergstrom Understanding Evolution,
University of California.
10
(No Transcript)
11
(No Transcript)
12
1
Antibiotic-sensitive
Antibiotic-resistant
Dead
13
Transformational Process
14
1
1. Where does the variation come from? 2. What
does the selecting?3. What are the
consequences? 4. How can we intervene?
15
1

1. 1. Where does the variation come from?2. What
   does the selecting?3. What are the consequences?
2. 4. How can we intervene?

16
Mutation

• Macrolide antibiotics block protein synthesis by
  binding to bacterial ribosomes.

From Hanson et al (2002) Molecular Cell
17
Mutation

• A single point mutation in the green binding
  region can prevent macrolide binding and confer
  resistance.

Modified from Hanson et al. (2002) Molecular Cell
18
Mutation

• Genome size 5 x 106 base pairs
• Mutation rate 2 x 10-3 per genome
• Population size 1010 to 1011 per g fecal
  matter
• A single gram of fecal matter is likely to
  contain a novel point mutation conferring
  macrolide-resistance!

19
Natural ecology of antibiotics
Soil microbes produce antibiotics to kill
competitors.
20
Lateral Gene Transfer
Electron micrograph Dennis Kunkel.
http//www.denniskunkel.com
21
Lateral gene transfer
Vancomycin Resistant Enterococcus
22
2

1. Where does the variation come from? 2. What
does the selecting?3. What are the
consequences? 4. How can we intervene?
23
Most resistant strains are commensals

24
Extremely high rate of drug use

25
Hospital staff act as disease vectors

26
High rate of patient turnover

27
Agricultural use
25 million pounds per year into animal
feed! Union of Concerned Scientists, 2001 Much
of this being erithromycin, one of the macrolides
discussed earlier.
28
Agricultural use
400,000 excess days of diarrhea a year due to
floroquinilone resistance (mostly?) in
Camphylobacter from chickens.
29
3

1. 1. Where does the variation come from?2. What
   does the selecting?3. What are the consequences?
2. 4. How can we intervene?

Doesn't take a rocket scientist, let alone an
evolutionary biologist. 1 million resistant
infections acquired each year in US
hospitals. Imposing a financial cost 4-5 billion
dollar cost and considerable extended stay times
and mortality.
30
Resistance in the Intensive Care UnitNational
Nosocomial Infections Surveillance System Report,
2003
Pseudomonas aeruginosa
Klebsiella pneumoniae
23
10
52
28
Staphylococcus aureus
Enterococcus sp.
31
In the Community Macrolide resistance
Streptococcus pneumoniae
Helicobacter pylori
32
20-90
Up to 70
Ineffective
Streptococcus pyrogenes
Haemophilus influenzae
32
Methicillin against

. macrolide resistance
Vancomycin used .
against MRSA
MRSA
33
Methicillin against

. macrolide resistance
Vancomycin used .
against MRSA
Linezolid against VRE
MRSA
VRE
34
1
1. Where does the variation come from? 2. What
does the selecting?3. What are the
consequences? 4. How can we intervene?
35
Antimicrobial cycling

• One-time shift of drugs clears up resistance
  outbreaks.
• Antimicrobial cycling takes the same idea
  further
• Try repeated, scheduled rotations among different
  drugs.
• Gentamicin, Piperacillin/Tazobactam and
  ceftazidime for gram-negatives in a neonatal ICU
  (Toltzis et al., Pediatrics 2002)
• Imipenem/cilastatin, pip / tazo, and ceftazidime
  clindamycin / cefepime in a pediatric ICU
  (Moss et al., Critical Care Medicine 2002)
• Carbapenems and ciprofloxacnin clindamycin,
  followed by cefepime metronidazole and pip /
  tazo in postoperative patients (Raymond et al.
  Critical Care medicine 2001)

36
Antibiotic cycling

• "The crop rotation' theory of antibiotic use
• suggests that if we routinely vary our go to'
• antibiotic in the ICU, we can minimize the
• emergence of resistance because the
• selective pressure for bacteria to develop
• resistance to a specific antibiotic would be
• reduced as organisms become exposed to
• continually varying antimicrobials." - M.
  Niederman (1997) Am. J. Respir. Crit.
  Care Med.

37
In our black boxBegin with a traditional SI
model
38
Community
Hospital
39
Translate the gearbox into equations

• S patients colonized with sensitive bacteria
  
• R patients colonized with resistant
  bacteria
• X uncolonized patients

40
We can solve explicitly for equilibrium behavior

• For example, resistance will be endemic when
• Left side is R0 for the resistant strain.
• Right side measures the availability of
  colonizable hosts

41
We can study the dynamics using numerical
solution

• E.g., things change fast.
• Non-specific control
• does appreciably
• reduce resistance.
• Formulary changes
• can rapidly eradicate
• resistant bacteria.When resistance is rare
  in the community

42
Extend our model to multiple resistant strains
Community
Hospital
43
An ODE model

• Two resistant strains, one sensitive strain.
• No dual resistance yet.

44
Dynamics of cycling90 day cycles
45
How do we judge whether cycling works?

• Total resistant infections R1 R2
• Probability of dual resistance arising by
  lateral gene transfer R1 R2
• Baseline for comparison In each case, compare
  the outcomes under cycling to an approximation of
  the status quo Mixing of the two drugs, in which
  at any given time half of the patients receive
  drug 1, the other half drug 2.

46
Total resistant infections
Cycling
Mixing
47
Total resistant infections by cycle length
Cycling
Mixing
48
Average total resistanceincreases with cycle
period
Cycling
Mixing
49
Rate of emergence of dual resistance
50
Rate of dual resistance evolution is greater
with cycling.
51
Why doesn'tcycling work?
Time
52
Why doesn'tcycling work?
Time
53
Mixing creates more heterogeneous environment
than does cycling!
Time
54
US infectious disease mortalitythroughout the
20th century
55
Acknowledgements

• Diane Genereux
• Department of Biology
• University of Washington