Professional Documents
Culture Documents
Bacterias Anaerobias0f
Bacterias Anaerobias0f
*Corresponding author. Department of Clinical Pharmacy, Institute of Pharmacy, Martin-Luther-Universitaet Halle-Wittenberg, Wolfgang-LangenbeckStr. 4, 06120 Halle, Germany. Tel: 49-345-5525190, Fax: 49-345-5527257; E-mail: charlotte.kloft@pharmazie.uni-halle.de
Keywords: in vitro models, antibiotics, dilution models, dialysis/diffusion models, static models
Introduction
The dosing regimens of antibiotics are often not optimal and the
doseresponse relationships not well known.1 One important
reason is that in the patient the pure antibiotic effect, i.e. the
pharmacodynamic (PD) characteristics,2 cannot clearly be separated from other factors determining the response to the antibacterial treatment. The effect also has to be regarded along with
pharmacokinetic (PK) properties,2 such as the ability of the
drug to reach its target. Thus, the PK, and the PD,2 are characteristics of an antibacterial agent and should be considered in the
development and prediction of the efficacy of the antibacterial
therapy. By linking the concentration time course (at the site
of action) to the drug effect (PK/PD), various dosing regimens
for different pathogens can be investigated in silico, enabling
the identification of potentially effective dosing regimens.
However, there is no standardized procedure for PK/PD evaluation for antibiotics, although the European Medicines Agency
(EMEA)1 and the FDA3 clearly recommend these investigations
for new compounds.
For characterizing the PD of an antibiotic, bacterial growth
and death under antibiotic exposure have to be investigated.
Since these are difficult to measure in human tissue, animal
and in vitro models have been developed. Animal models
provide similar growing conditions for bacteria, closely imitating
the characteristics of a human infection, and the endpoint of an
# The Author 2009. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: journals.permissions@oxfordjournals.org
186
In vitro pharmacodynamic (PD) models are used to obtain useful quantitative information on the effect of
either single drugs or drug combinations against bacteria. This review provides an overview of in vitro PD
models and their experimental implementation. Models are categorized on the basis of whether the drug
concentration remains constant or changes and whether there is a loss of bacteria from the system. Further
subdifferentiation is based on whether bacterial loss involves dilution of the medium or is associated with
dialysis or diffusion. For comprehension of the underlying principles, experimental settings are simplified and
schematically illustrated, including the simulations of various in vivo routes of administration. The different
model types are categorized and their (dis)advantages discussed. The application of in vitro models to
special organs, infections and pathogens is comprehensively presented. Finally, the relevance and perspectives
of in vitro investigations in drug discovery and clinical research are elucidated and discussed.
JAC
Review
Fundamentals
Characteristics of in vitro models
Classification
As a consequence of these definitions, all in vitro models for antibiotics can be classified according to the change of the drug concentration and whether or not there is bacterial loss (Table 1):
I models with a constant drug concentration and no bacterial
loss;
II models with changing drug concentration and bacterial
loss; and
Constant
Changing
II
no (closed systems)
I
static models
IIIa
IIIb
187
Review
188
Table 2. Special applications of in vitro models
Application and rationale
Models accounting for the immune system
realistic cell/fluid environment, examine the
penetration into immune cells,46 and
include the effect of immune cells
Special features
Ref.
static model
dilution model with stepwise substitution
dilution model with continuous dilution
dialysis model
46, 109
133
80, 81
93
static model
dialysis model
47, 135138
100, 132, 139
Intracellular models
PK and PD of intracellular bacteria different
from extracellular
static model
dialysis model
Site of infections/diseases
different PK profiles or bacterial growing
conditions41
bladder/bacterial cystitis
otitis media
endocarditis
chronic pneumonia
infected fibrin clots
dental infections
skin infections
tuberculosis
Multicompartmental models
simulate PK of drugs with numerous
theoretical compartments, no. of
compartments determines no. of (culture)
vessels
dialysis model
dialysis model
static model
dialysis model
static model
dialysis model
static model
dialysis model
dilution model with continuous dilution without and with
filters
140, 141
103108, 141
77, 142145
92, 146
131
140
131, 147, 148
149
149151
152
105, 141
Biofilm models
bacteria in biofilms: properties distinct from
single organisms, e.g. increased
antimicrobial resistance134
JAC
Production of biofilms by growing slime-producing bacteria, e.g. Staphylococcus aureus 47 and Pseudomonas aeruginosa 135 on various surfaces.
Host cells grown until a stable culture appears (continuous layer), bacterial suspension directly added, culture is incubated and investigations start when the infection is positive.
c
If antibiotics with different half-lives are simultaneously investigated in dynamic models, the flow of the medium, which decreases the drug, has to be adjusted to the shortest halflife41 or, more appropriately, the drug with the longer half-life has to be substituted to the reservoir.
b
dialysis model
static model
dilution model with continuous dilution
with filters
without filters
static model
dilution model with stepwise substitution (with filters)
dilution model with continuous dilution
Models for combination therapyc
synergistic effects
investigations on the relation/s of concentration- and timedependent bacterial growth. Specific organs and their special
conditions appear separately in Table 2 (Applications). The
new, extended classification (see Table 1) is based on both mentioned main characteristics of in vitro models, i.e. drug concentration and bacterial concentration, whereby all commonly
used models can be categorized.
189
without filters
with filter
161, 162
121
163
72
164
162, 165
48, 153
73
154
20, 155157
51, 117, 158160
Review
Review
Rationale
Implementation
waterbath, incubator
shaking or stirring
Model developments
Static models (No. I)
Static models consist of a closed culture vessel (Figure 2a). These
vessels are available in a variety of shapes, such as tubes,46,47
flasks,43,54 cell culture flasks44 or spinner flasks,55 and may be
made of glass43 or polystyrene.47 The first time kill investigations in static models were established by Garrett et al. 43
in 1966.
Models with continuous simple dilution reflect the in vivo conditions of a drug much more closely than a stepwise decline of
the drug. The decisive improvement in this field was made by
Grasso et al. 57 The Grasso model consists of a flask containing
the bacteria (culture vessel), a reservoir and a waste container
(Figure 2b). Fresh medium is continuously pumped from the
reservoir into the flask and used medium leaves the culture
vessel by the pressure of the incoming medium. Drug and bacterial samples can be taken from the vessel. A magnetic stirrer
ensures a homogeneous distribution of the drug and bacteria.
In the Grasso model, the bacteria are diluted by the incoming
medium and flow out with the outgoing medium, which
demands a mathematical correction for bacterial counts. In
theory, flow rates that are faster than the bacterial growth
rates would lead to a complete loss of bacteria; however, Haag
190
Review
Method
Principle
online
measurementa
direct
measurementb
detection of live
cells only
differentiation
between live
and dead cells
Comment
NA
Turbidimetry
43
126
NA
NA
Impedance
Bioluminescence
Microscope
Fluorescence
quantification
RNA profiling
Ref.
43, 56, 57, 97
76, 77, 84
128, 129
JAC
191
Downloaded from http://jac.oxfordjournals.org/ by guest on April 27, 2014
Review
Dynamic models
Working principle
Dialysis/
diffusion models
Yes
Simple
dilution
Adding and
removing
of medium
Substitution
Membrane material
Artificial
barrier
Natural
barrier
Artificial membranes
Continuous dilution
Continuous dilution models operating without outlet, resulting in
increasing volumes, were described by Sanfilippo and Morvillo,76
and OGrady and Pennington.77 They mostly reflect only selected
aspects of the in vivo situation. Pumps transport the medium
from a reservoir into the culture flask (Figure 2d). Since there is
no outlet, the volume of the second flask continuously increases,
the drug concentration changes and the bacteria will be
diluted.76 The increasing volume does not necessarily allow
exact exponential decline of the drug concentrations (see also
bladder/bacterial cystitis in Table 2).77
Continuous dilution without bacterial loss can also be achieved
with filters. The model by Greenwood and Tupper78 consists of a
vessel separated by a filter membrane in two chambers, with
the bacteria and drug being added to the upper one
(Figure 2e).78 However, this model has not been further used.
Instead, for the already mentioned decisive Grasso model
(dilution, but bacterial loss),57 different modifications have been
suggested to improve the accuracy. Filters are inserted between
the culture vessel and the waste, and the outlet has moved
from the side panel to the bottom of the flask.79 81 A practical solution to prevent the bacterial loss in the Grasso model was found
by Lowdin et al. 55 (Figure 2e). The base of a spinner flask is modified, including an outlet and a perforated metal support, on which
a filter membrane and a pre-filter are adjusted. Above the membrane, a magnetic stirrer is placed to prevent membrane blockage.
Fresh medium is pumped in via one side arm in the spinner flask.
The other arm is prepared with a silicon membrane for repeated
sampling.55,82,83
A multicompartment model based on the Grasso model
with retention of bacteria was presented by Navashin et al. 84
192
Direct adding or
direct removing
of medium
Yes
No
JAC
Review
Natural membranes
The principle of dialysis models with natural membranes is
almost the same as for those with artificial membranes. Bacteria
(a)
are captured behind a barrier and the drug has to pass the
barrier to reach the bacteria. Haller102 described a tissue
culture model, where tissue cells are grown on a dialysis ultrafilter until a continuous layer is formed. The membrane, consisting
of the filter and the tissue cell layer, is placed on a cylinder
(central compartment). Another cylinder located above serves
as the peripheral compartment with bacteria. The antibiotic is
administered in the lower part of the chamber by syringes and
diffuses through the cells to the upper part (Figure 2g). The
model was suggested to investigate the penetration of the
drug through intercellular spaces and was later used by numerous groups to investigate drug bacteria effects.103 107 An intracellular model implementing tissue cultures in PD in vitro models
is presented by Hulten et al. 103 Tissue cells are grown in inserts
in a glass chamber, similar to a closed Petri dish (central
(b)
(c)
W
B
(d)
(e)
1.
or
2.
3.
or
4.
(f)
193
Review
B Peripheral
Central
Central
Peripheral
Peripheral
Central
B
Peripheral
W
Central
Caption:
B
R
W
Figure 2. Continued
Natural membranes
ref.
85
86
52, 94, 97
96
89
95
material
ref.
agarose gel
130
cells
102
cell membranes 102105
fibrin
131
slime
47, 132
194
JAC
Review
(a) In dilution models with the example of a continuous simple dilution model
Bolus administration
Extravascular
administration
No absorption
0-order absorption 1st-order absorption
Infusion
A
R
or
Infusion
Extravascular
administration
Bolus administration
No absorption
D
Central
A
R
D
R
A
B
W
or
B
Peripheral
drug
reservoir
additional vessel, mimicking absorption (optional)
culture vessel with bacteria
waste
Figure 3. Schematic depiction of in vitro implementation of the different in vivo routes of administration.
Applications
A substantial number of in vitro PD models have been developed
to simulate specific conditions. Even if not all of these models
can imitate the designated PK profiles, they are useful tools for
specific conditions. In Table 2, the models are grouped by their
main aspects and may appear in different categories.
195
Review
196
static models. They do not offer a continuous dilution and, therewith, not the same exposition profile for bacteria as in vivo.
Nevertheless, it is possible to achieve more realistic results
than with static models.
Dialysis models are extensively used, as well. Their main
advantage is the closed system, whereby bacteria cannot
escape and no further filters need to be installed. Dialysis
models enable simultaneous investigations of different bacterial
strains in separated vessels, but in one model.97 However, bacteria accumulate at the membrane, which might become
blocked (as in the case of dilution models).52,85 Furthermore,
the changing drug concentration in the bacteria compartment
does not necessarily follow the designated PK profile,14 since
the drug has to pass a barrier. Diffusion of a drug is a first-order
process. The extent of drug transfer across the barrier depends
on the site of membrane permeation and varies with time.
This means there is a specific concentration gradient between
the drug concentration in the central compartment and peripheral compartment at each timepoint. Unfortunately, only a few
groups have determined the drug concentration in the bacteria
compartment,52,53,85,86,89,92,94,97 99 where the concentration
gradient was confirmed. The gradient can be improved by
higher circulation (addition of pumps) and contraflow of the
medium in the central compartment and peripheral compartment.52,53,97 The early dialysis models with artificial membranes
as well as most with natural membranes suffered from a small
surface area of the membrane to volume ratio. This led to diminished diffusion or membrane blockage.85,87,93 Later, the exact
ratio between the membrane surface area and the volume of
the peripheral compartment was estimated and changed, e.g.
by Vance-Bryan et al.,92,122 for the Shah model.93 In intracellular
models, determination of the drug concentration is even more
problematic as the site of action is inside the cells. Here, the
flow is not directed for an optimal exchange between extraand intracellular fluid, which may lead to different PK profiles
that the bacteria will be exposed to. Only with special equipment
and procedures such as fluorescence microscopy is it possible to
measure the drug concentration inside the cells. So, the effect is
often related to the drug decline in the central compartment,
because of easier determination.
In summary, in spite of their simplicity, static models will still
play an important role for antibacterial PK/PD studies in the
future, but should be regarded as a starting point. For more
complex PK designs, dynamic models will be more important
and their use will hopefully increase. The dilution models,
such as the Grasso model,57 have existed since the 1970s
and have been intensively diversified. Almost at the same
time, dialysis models have been introduced and, later,
improved. Currently, the ratio of using dialysis or dilution
models is balanced. Both types of dynamic models have been
further developed in the past and are presented in the
current literature. New developments combine the ideas of a
one-compartment dilution model with filters and a twocompartment dialysis model, resulting in a computer-controlled
semi-automated in vitro model for industrial purposes.111 In
future, this trend of combining models for different purposes,
as well as automation, might lead to more frequent use and,
eventually, they might become an inherent part of drug discovery and development. Comprehensive understanding of the PD
of antibiotics should facilitate the development of rational
a stronger competition for nutrients can lead to a higher production of antimicrobial drug targets, resulting in a higher susceptibility in vitro.43,114 In spite of this, in vitro models allow good
prediction of bacterial growth in vivo and comparison of different
dosing regimens of one drug as well as comparison between different drugs. Finally, they contribute to dose optimization.18,26,100,112
The choice of a specific in vitro model is determined by the
objectives of a PK/PD study as well as the advantages and disadvantages of in vitro models. Static in vitro models are extensively
used as they are easy to handle and well investigated. They
provide basic information on the interaction between the antibiotic and bacteria. In contrast to these favourable economic
aspects is the unrealistic nature of the unchanged drug and
medium in static models,41 which fail to mirror two important
aspects of in vivo conditions, namely exchange of nutrients
and dilution of the drug. They are not useful for prolonged treatment studies, because nutrient depletion, space limitations and
toxic metabolites lead to growth restrictions.38 In our opinion,
static models should be used as the starting point for PD
studies of the effect over time. The quick-and-easy findings
from static models are useful preliminary knowledge for
dynamic investigations.
Dynamic in vitro models represent the in vivo conditions with
respect to the changing drug and medium much more
closely.14,41 Beside this, dynamic models also enable prolonged
treatment studies (up to 5,116,117 10118,119 and also 15 days120)
with multiple dosing.52,71 Associated potential problems include
the Haag factor58 and membrane blockage. On the other
hand, they require large volumes of growth medium for changing
the drug concentration according to the half-lives of the drug.
Hence, antibiotics with a long half-life have very low flow rates,
a low volume of medium replacement and thus nutrient
depletion, and an increase of toxic metabolites. The two main
principles for changing drug concentrations have been developed
and are both in use: dilution models and dialysis models.
Dilution models can imitate virtually all in vivo PK profiles. The
drug and bacteria are in one compartment, so the bacteria are
directly exposed to the designated drug concentration. Hence,
these models apply the designated PK drug profile to the bacteria, but they should be monitored. The early dilution models
were designed without bacterial retention as open models. The
bacterial loss always has to be corrected.49,50 In the Grasso
model,57 the bacteria leave the culture vessel with the outgoing
medium and toxic waste is diluted, but not considered. Additionally, bacterial aggregation and adherence to the vessel wall was
found, where the bacterial populations should not be detectable.58 In the model by Murakawa et al. 40 the bacteria are distributed into the second compartment and also eliminated into
the waste. This complicates corrections for accurate bacterial
concentrations and, therefore, reliable predictions of the antibacterial effect. The inserted bacteria filter causes new problems:
the filter often becomes blocked with bacteria the longer the
experiment lasts.85,121 The implementation of pre-filters or stirrers has provided potential solutions,55,82 but also open models
have been reused. This has meant a step back regarding the
loss of bacteria.121 In the later developed closed dilution
models, the bacterial backgrowth into the reservoir presented
another problem. Dilution models with stepwise substitution
(and filters), such as the syringe model by Nolting et al.,45 are
easily practicable, but need even more laborious effort than
Review
Funding
This work was partially supported by a grant from the Dr. August und
Dr. Anni Lesmueller-Stiftung, Germany.
Transparency declarations
The authors do not have any financial, commercial or proprietary interest
in any drug, device or equipment mentioned in this paper.
15 Firsov AA, Zinner SH, Lubenko IY. In vitro dynamic models as tools to
predict antibiotic pharmacodynamics. In: Nightingale CH, Ambrose PG,
Drusano GL et al., eds. Antimicrobial Pharmacodynamics in Theory and
Clinical Practice, 2nd edn. New York, NY: Informa Healthcare, 2007;
45 78.
16 Rybak MJ, Allen GP, Hershberger E. In vitro antibiotic
pharmacodynamic models. In: Nightingale CH, Murakawa T, Ambrose
PG, eds. Antimicrobial Pharmacodynamics in Theory and Clinical Practice.
New York, NY: Marcel Dekker, Inc., 2002; 41 66.
17 Li RC, Zhu ZY. In vitro models for prediction of antimicrobial activity: a
pharmacokinetic and pharmacodynamic perspective. J Chemother 1997;
9 Suppl 1: 55 63.
18 MacGowan A, Bowker K. Developments in PK/PD: optimising efficacy
and prevention of resistance. A critical review of PK/PD in in vitro
models. Int J Antimicrob Agents 2002; 19: 2918.
19 MacGowan A, Rogers C, Bowker K. In vitro models, in vivo models, and
pharmacokinetics: what can we learn from in vitro models? Clin Infect Dis
2001; 33 Suppl 3: 21420.
20 Blaser J. In-vitro model for simultaneous simulation of the serum
kinetics of two drugs with different half-lives. J Antimicrob Chemother
1985; 15 Suppl A: 125 30.
21 Rowe EL, Morozowich W. A simple dilution analog computer for
simulation of drug distribution processes. J Pharm Sci 1969; 58: 1375 8.
22 Derendorf H, Meibohm B. Modeling of pharmacokinetic/
pharmacodynamic (PK/PD) relationships: concepts and perspectives.
Pharm Res 1999; 16: 17685.
23 Czock
D,
Keller
F.
Mechanism-based
pharmacokinetic
pharmacodynamic
modeling
of
antimicrobial
drug
effects.
J Pharmacokinet Pharmacodyn 2007; 34: 72751.
24 Schuck EL, Derendorf H. Pharmacokinetic/pharmacodynamic
evaluation of anti-infective agents. Expert Rev Anti Infect Ther 2005; 3:
36173.
25 Meibohm B, Derendorf H. Basic concepts of pharmacokinetic/
pharmacodynamic (PK/PD) modelling. Int J Clin Pharmacol Ther 1997;
35: 40113.
197
References
JAC
Review
198
52 Blaser J, Stone BB, Zinner SH. Two compartment kinetic model with
multiple artificial capillary units. J Antimicrob Chemother 1985; 15
Suppl A: 1317.
Review
199
JAC
Review
123 Eng RH, Smith SM, Cherubin CE et al. Evaluation of two methods for
overcoming the antibiotic carry-over effect. Eur J Clin Microbiol Infect Dis
1991; 10: 34 8.
124 den Hollander JG, Mouton JW, Bakker-Woudenberg IA et al. Enzymatic
method for inactivation of aminoglycosides during measurement of
postantibiotic effect. Antimicrob Agents Chemother 1996; 40: 48890.
125 den Hollander JG, Mouton JW, van Goor MP et al. Alteration of
postantibiotic effect during one dosing interval of tobramycin,
simulated in an in vitro pharmacokinetic model. Antimicrob Agents
Chemother 1996; 40: 784 6.
126 Hanberger H, Svensson E, Nilsson M et al. Effects of imipenem on
Escherichia coli studied using bioluminescence, viable counting and
microscopy. J Antimicrob Chemother 1993; 31: 24560.
127 Fang W. Quantification of Staphylococcus aureus and Escherichia coli
in the liquid medium by fluorimetry and its use in phagocytosis assay.
J Appl Bacteriol 1996; 80: 57782.
128 Fey A, Eichler S, Flavier S et al. Establishment of a real-time
PCR-based approach for accurate quantification of bacterial RNA
targets in water, using Salmonella as a model organism. Appl Environ
Microbiol 2004; 70: 361823.
129 Smith CJ, Osborn AM. Advantages and limitations of quantitative
PCR (Q-PCR)-based approaches in microbial ecology. FEMS Microbiol Ecol
2009; 67: 6 20.
130 Tomita T, Ohara-Nemoto Y, Moriyama H et al. A novel in vitro
pharmacokinetic/pharmacodynamic model based on two-compartment
open model used to simulate serum drug concentration time profiles.
Microbiol Immunol 2007; 51: 56775.
131 McGrath BJ, Kang SL, Kaatz GW et al. Bactericidal activities of
teicoplanin, vancomycin, and gentamicin alone and in combination
against Staphylococcus aureus in an in vitro pharmacodynamic
model of endocarditis. Antimicrob Agents Chemother 1994; 38: 2034 40.
132 Vergeres P, Blaser J. Amikacin, ceftazidime, and flucloxacillin against
suspended and adherent Pseudomonas aeruginosa and Staphylococcus
epidermidis in an in vitro model of infection. J Infect Dis 1992; 165:
2819.
133 Randolph JA, Buck RE, Price KE et al. Comparative bactericidal effect
of ceforanide (BL-S 786) and five other cephalosporins in an in vitro
pharmacokinetic model. J Antibiot (Tokyo) 1979; 32: 727 33.
134 Mah TF, OToole GA. Mechanisms of biofilm resistance to
antimicrobial agents. Trends Microbiol 2001; 9: 34 9.
135 Nickel JC, Wright JB, Ruseska I et al. Antibiotic resistance of
Pseudomonas aeruginosa colonizing a urinary catheter in vitro. Eur J
Clin Microbiol 1985; 4: 213 8.
136 Prosser BL, Taylor D, Dix BA et al. Method of evaluating effects of
antibiotics on bacterial biofilm. Antimicrob Agents Chemother 1987; 31:
1502 6.
137 Nickel JC, Ruseska I, Wright JB et al. Tobramycin resistance of
Pseudomonas aeruginosa cells growing as a biofilm on urinary catheter
material. Antimicrob Agents Chemother 1985; 27: 61924.
138 Ellen RP, Lepine G, Nghiem PM. In vitro models that support
adhesion specificity in biofilms of oral bacteria. Adv Dent Res 1997; 11:
33 42.
200
Review
201
JAC