Inhibition of brain NADH oxidationby chloramphenicol
in the freely moving rat
Stéphane Mottin1, Pierre Laporte1 and Raymond Cespuglio2
1
2
LTSI, CNRS UMR 5516, F-42023 St-Etienne Cedex 02, France
INSERM U480, F-69373 Lyon, France.
Proofs and reprint requests should be addressed to:
Stéphane Mottin, LTSI, CNRS UMR 5516, F-42023 St-Etienne, France
Email: [email protected]
Mottin, S. identifier: http://orcid.org/0000-0002-7088-4353
Abbreviations:
CAP, Chloramphenicol; 5HT, 5-hydroxytryptamine; Cx, Cortex; nRD; nucleus raphe
dorsalis; ps, Paradoxical Sleep; TRES, time-resolved emission spectroscopy
Reference BibTeX
@article{2003_mottin_21,
TITLE = {The inhibition of brain NADH oxidation by Chloramphenicol in the freely moving rat
measured by picosecond time-resolved emission spectroscopy.},
AUTHOR = { Mottin, Stéphane and Laporte, Pierre and Cespuglio , Raymond },
JOURNAL = {J. Neurochem.},
VOLUME = { 84},
NUMBER = {4},
PAGES = { 633-642},
YEAR = {2003},
DOI = {10.1046/j.1471-4159.2003.01508.x},
KEYWORDS = {Time-resolved fluorescence;Neurotoxicity;Antibiotic;NADH;Behaving
Rat;Sleep;Chloramphenicol;mitochondrial effects;brain;metabolism},
}
DOI DataCite: https://doi.org/10.5281/zenodo.439021
Abstract
Owing to the lack of methods capable to monitor the energetic processes taking place within small
brain regions (i.e. nucleus Raphe Dorsalis, nRD), the neurotoxicity of various categories of
substances, including antibiotics and psycho-active drugs, still remains difficult to evaluate. Using
an in vivo picosecond optical spectroscopy imaging method, we report that chloramphenicol (CAP),
besides its well-known ability to inhibit the mitochondria protein synthesis, also influences the
NADH/NAD+ redox processes of the respiratory chain. At a 200 mg/kg dose, CAP produces indeed
a marked increase in the fluorescent signal of the nRD which, according to clear evidence, is likely
to be related to the NADH concentration. This effect also implies an efficient inhibition of complex
I of the respiratory chain by CAP. It refers to the mechanism through which the adverse effects of
the antibiotic may take place. It could explain why paradoxical sleep, a state needing aerobic energy
to occur, is suppressed after CAP administration.
The present approach constitutes the first attempt to determine by fluorescence methods the effects
of substances on deep brain structures of the freely moving animal. It points out that in vivo ultrafast
optical methods are innovative and adequate tools for combined neurochemical and behavioural
approaches.
Keywords: Time-resolved fluorescence, Neurotoxicity, Antibiotic, NADH, Behaving Rat, Sleep.
License CC-BY-NC-ND
2
Introduction
To date, the "seasoned" chloramphenicol (CAP) antibiotic still remains widely used in
developing countries (Kumana et al.,1993) and the importance of its supply is so crucial that it often
leads to forgery (Payet,1997). This substance, among the most famous mitochondriotropic drugs
used in research (Kroon and Arendzen,1972; Yunis,1988; Bories and Cravendi,1994) is obviously
precious in the case of serious epidemic (Niel et al.,1997) and of multiresistant treatments
(Barie,1998). Owing to its excellent accessibility to the cerebro-spinal fluid and brain tissue, this
antibiotic is incomparably efficient against meningitis and typhoid fever (Meulemans et al.,1986).
Nevertheless, it must be mentioned that its administration to patients often induces adverse effects
including mental confusion, headache, appetite loss, ophtalmoplegia, selective inhibition of
paradoxical sleep (PS) and epileptogenic manifestations (Abou-Khalil et al.,1980; Yunis,1988; Holt
et al.,1993; Bories and Cravendi,1994).
Most of the antibiotic adverse effects on the brain have often been misinterpreted
(Snavely,1984; Thomas,1994; Norrby,1996; Kanemitsu et al.,1999). As yet, many CAP metabolites
have been studied in regard to mitochondria metabolism. But, approaches related to the in vivo study
of the brain mitochondria networks (Glinka et al.,1998; Yaffe,1999) still remain difficult. Owing to
the progress of neurophotonics, direct brain investigations have been started in the unanesthetized
animal (Mottin et al.,1997; Cassarino and Bennet,1999).
Besides, it is also known that mitochondria are elements sensitive to antibiotics (Ramilo et al.,1988;
Snavely and Hodges,1984, Degli Esposti,1998) and, in this respect, CAP has been widely studied.
After discovery of the CAP ability to inhibit mitochondria respiratory processes (Stoner,1964),
several other reports (Freeman and Haldar,1967; Freeman and Haldar,1968; Freeman and
Haldar,1970; Kroon and Arendzen,1972; Abou-Khalil et al.,1980) further pointed out that this
compound is an inhibitor of the mitochondria complex I (NADH-ubiquinone oxidoreductase, EC
1.6.5.3) on isolated mitochondria preparations. They allowed the conclusion that the concentration
of CAP necessary for the inhibition of complex I is far superior to the concentration necessary for
License CC-BY-NC-ND
3
the mitochondria protein synthesis inhibition. Since this period, CAP has been used, even in vivo,
as a "pure" mitochondria protein synthesis inhibitor. Its ability to inhibit the mitochondria complex
I has been neglected since considered as effective only in vitro at strong overdoses (Abou-Khalil et
al.,1980; Yunis,1988; Holt et al.,1993; Bories and Cravendi,1994). CAP presents, nevertheless, an
unusual excellent accessibility to the cerebro-spinal fluid and brain tissue where its accumulation
may reach a concentration efficient enough to inhibit mitochondria respiration (Meulemans et
al.,1986).
Throughout this report, we provide answers to the unsolved problems attached to the adverse effects
of CAP not related to the inhibition of mitochondria protein synthesis. For this purpose,
NADH/NAD+ redox processes taking place in the nucleus raphe dorsalis (nRD) of the freely moving
rat were first monitored with a picosecond time-resolved fluorescence method. The nRD target was
chosen because of its involvement in sleep triggering (Cespuglio et al.,1992) and CAP was
employed on the basis of its ability to suppress PS (Petitjean et al.,1979). Afterwards, the effect of
CAP on the NADH/NAD+ redox balance was checked.
Methods
Experimental procedure.
In fifteen OFA male rats (IFFA CREDO, France) weighing 280-300 g and anaesthetised
with chloral hydrate (400 mg/kg, IP), a guide canula was implanted in the nRD according to a
procedure previously described (Mottin et al.,1997). After 10 days of recovery (12h-12h light-dark,
temperature at 24 ± 0.5 °C, food and water ad libitum) time-resolved fluorescence measurements
were carried out (daily sessions of 4-8 h). At the end of the experimental sessions, the animals were
sacrificed with a lethal dose of Nembutal and the position of the working sensor checked. CAP
hemisuccinate (Solnicol©, Synthelabo, France) and saline solution were administered
intraperitoneally (IP). For the 337 nm excitation wavelength experiments, two CAP doses were
used, i.e. 200 and 400 mg/kg. With the 355 nm excitation wavelength, experiments were conducted
with a 300 mg/kg dose.
License CC-BY-NC-ND
4
Time-resolved emission spectroscopy (TRES)
An application of ultrafast neurophotonics enabling both spectral and temporal analysis of
tissue fluorescence in behaving animals has been achieved in this study. The first generation of the
set-up used was described before (Mottin et al.,1997). Briefly, delivery and collection of the optical
signals (laser excitation and emission) were performed through a thin optical fibre allowing a good
anatomical resolution (Ø core=200 µm). A streak camera equipped with a special VUV window
(StreakScope from Hamamatsu, Japan) and a 270M spectrograph (Spex Jobin-Yvon, France) were
also used in the second generation of the optical design. Images obtained with the streak camera
were registered through a two-dimensional single-photon counting mode (Watanabe,1994).
TRES methodology in vivo
When the measurements are limited in intensity, without a real-time TRES analysis, the link
existing between the photo-electron counts and the fluorophore concentration variation cannot be
defined (Mottin et al,1993). The optical signal is proportional to the NADH concentration variation
only if the spectrum and the decay-time remain unchanged. Thus, if the NADH quantum efficiency
changes, if another emission overlaps the NADH fluorescence or if the inner filters absorb NADH
emission, then the conventional fluorimetric methods fail. TRES imagery avoids these
inconveniences and allows a more objective analysis of tissue optics. In order to add strength to our
methodology, we also introduced the control of the photon counting rate. For this purpose, the laser
intensity was set at a low level : 0.15 mW, 30 Hz, 5 J/pulse. The excitation wavelengths were (1)
337 nm, with a nitrogen laser at a repetition rate of 30 Hz and a FWHM (full width at half max) of
300 ps (LN 100, Laser Photonics, USA), (2) 355 nm with a tripled YAG laser at a repetition rate of
30 Hz and a FWHM of 3.5 ns (OPO901, BMIndustrie, France). Despite this long FWHM, the
355 nm wavelength was of a great interest with regard to the recent picosecond YAG microchip
laser developments. For the 337 and 355 nm wavelengths, we used a time window of 10 ns and
20 ns respectively, the integration time being set at one minute. In the 487-508 nm emission
wavelength window, the magnitude of the noise (measured in deionized water) was 2% and 8.5%
of the basal nRD fluorescent signal for 337 and 335 excitation wavelengths respectively.
License CC-BY-NC-ND
5
Significance of the increase in fluorescence observed after CAP administration can be
analysed in using different statistical tests. Since our in vivo results are time series data we used a
paired t-test well adapted to evaluate the significance of the changes observed.
In pharmacology, temporal distributions of such time-dependent variables are usually
studied by non-linear regression analysis. Thus, in order to quantify all aspects of the mean increase
in the autofluorescence induced by CAP, a mathematical pharmacokinetic model was defined, i.e.
y = a+b (1-exp(-(t-d)/c)), y being the value of the fluorescent signal expressed in single photoelectron count units (SPE) and t the temporal scale in minutes. Coefficients a, b, c and d represent
respectively the basic autofluorescence level (in SPE count), its increase (in SPE count), the time
lapse covering this variation (min) and the delay (min) existing between the injection procedure and
the beginning of the signal increase. To assess the validity of the model, the regression coefficient
(R), was always setted above 0.95.
Results
TRES imagery in vivo
A typical TRES image, derived from the nRD, is illustrated in Figure 1A. The
autofluorescence spectrum is measured in the 377-554 nm window (Fig. 1B). The temporal analysis
of the fluorescence shape gives a mean decay time of 900 ± 50 ps within the 487-508 nm window
(Fig. 1C).
Changes induced by CAP
In freely moving animals, saline administration did not induce behavioural changes nor
variations in the TRES signal. However, all the intraperitoneal (i.p.) injections of CAP succinate
(200-400 mg/kg) performed in the same conditions induced a highly significant increase in the nRD
blue fluorescence (Fig. 2). The means values of the basal counting (a) given ± the standard error are
respectively for 200 mg/Kg, 300 mg/Kg and 400 mg/Kg : 12721 ± 2310 SPE, 48794 ± 2474 SPE
and 16013 ± 1051 SPE. We further noticed that the basal counting rate was lower for the 337 nm
License CC-BY-NC-ND
6
wavelength excitation than for the 355 nm one. This was mainly due to the difference in the laser
beam quality and the coupling into the optical fibre.
The means values of (c) also given ± the standard error are respectively for 200 mg/Kg, 300
mg/Kg and 400 mg/Kg : 71 ± 33 min, 107 ± 24 min and 126 ± 35 min.
The delay (d) existing between the injection procedure and the beginning of the signal
increase is in a 2 -14 min time window with a mean peaking at 4 min.
The fluorescence increment (b) induced by CAP injection is shown on figure 3. The paired
t-test comparisons performed indicate that the differences existing between the CAP doses are
significant. The differences existing between the CAP doses between 200 mg/Kg and 300 mg/Kg
are significant. Between 300 mg/Kg and 400 mg/Kg or between 200 mg/Kg and 400 mg/Kg the
differences are highly significant.
Regarding the 337 nm excitation, spectra obtained before CAP injection exhibited a high
variability in the UV-purple part. Below 450 nm, several patterns of the spectra and decay times
were also measured This variability might be due to the presence or the absence of a UV-purple
shoulder coming probably from different endogeneous fluorophores also combined with the Soret
Band of the hemoglobins (inner effect). Concerning again the above variability, the 450-480 nm
window was in an intermediate position while above 480 nm, the UV-purple shoulder was less
sensitive (Fig. 4).
In the case of the 355 nm excitation, basal spectra were more reproducible. Figure 5
illustrates the variations induced by a CAP injection on the whole spectral window. In the 450-550
nm window, the increase in fluorescence obtained was greater than in the 380-440nm window (4
positive effects / five trials).
Finally, we also checked that, for a 300 mg/kg dose of CAP, the overall CAP
pharmacokinetics (increase and decrease down to the basic fluorescence level) occurred within 67 h.
License CC-BY-NC-ND
7
Changes induced by the animal death
Concerning the ability of CAP (or metabolites) to inhibit the NADH/NAD+ redox processes
of the respiratory chain in the nRD, it is hard to give an absolute evaluation of the inhibition strength.
In order to overcome this difficulty, we compared the changes occurring in the signal during the
animal death (lethal dose of barbiturates: 120 mg/kg) with those obtained after a CAP injection. The
lethal dose was used when the basal level fluorescent signal before CAP injection was fully reached.
Results obtained indicate that during death the increase in the fluorescent signal is faster and higher
than after a CAP dose of 300 mg/kg (Fig.3). The magnitude of the NADH fluorescence increase
induced by the CAP is close to 40% of the death effect.
Discussion
Data obtained indicate that CAP induces a significant increase in the laser (335-337 nm
excitation wavelengths) induced NADH fluorescent signal of the nRD. For a CAP dose of
300 mg/Kg the effect obtained is close to 40% of the death triggered variation..
NADH dependence of the fluorescent signal
Since the pioneer work of Chance (Chance et al.,1962), several optical designs have been
published (Mottin et al.,1997). Many authors discussed the link existing between the UV-induced
brain fluorescence and the NADH intramitochondria concentration (Sick et al.,1999; Rex et
al.,1999, Hashimoto et al.,2000; Schuchmann et al., 2001). Again, the recent and important changes
reported in the mitochondria glucose-stimulated NADH fluorescence from intact pancreatic islets
(Eto et al.,1999; Patterson et al., 2000) confirm this aspect. It is thus very likely that the brain
autofluorescence measured in the 480-540 nm window might be attached to NADH (for 337 or
355 nm excitation wavelengths). However, since the TRES imagery, used in the present approach,
offers, over conventional spectrofluorimetric methods, the beneficial access to a complete analysis
of the tissue fluorescence, we again considered the dependence of the fluorescent signal measured
on the NADH concentration. In this respect, we further analysed whether the signal obtained could
License CC-BY-NC-ND
8
be produced by CAP itself. In the 480-540 nm window, the changes observed might indeed result,
not only from an increase in the NADH concentration, but also from: - the optical properties of
xenobiotic compounds; - the modifications occurring in the optical properties of the tissue; - the
increase in quantum efficiency of endogenous fluorophores.
Regarding the optical properties of xenobiotic compounds, it can be underlined that the first
spectral component observed in UV absorption, with CAP in water at pH 7, occurs at 278 nm.
Within the large 320-600 nm wavelength window, the CAP and metabolites absorption (Bories and
Cravendi,1994) are thus negligible. Therefore, the optical properties of CAP and metabolites cannot
interfere with the signal measured.
For the optical properties of the brain tissue, some modifications may occur when CAP is
strongly infused intravenously (Sangiah and Burrows,1989). In such conditions, hypotension is
triggered together with an increase in the cerebral blood volume. Both events could well be at the
basis of a haemodynamic artefact. Regarding this aspect, we emphasize that our experimental
protocol used only i.p. administrations of CAP and that our methodological set-up exhibited a
photon counting rate in the same range throughout the different experimental sessions. This
homogeneity underlines that our sensor was at a scale avoiding angioarchitectonic influences of the
nRD. This nucleus is, indeed, poorly vascularized (Descarries et al.,1982) since about 12-24
capillary lumens are present in the section of our sensor. In the 480-540 nm window, however, the
tissue absorption is lower than in UV and might increase the absorption and the scattering effects
produced by the nRD capillaries. Despite this assumption, we nevertheless observed that the 480540 nm spectral shapes do not change after CAP administration. Whatever the complexity of the
optical tissue properties might be, the time-course of the transient hypotension attached to CAP
administration is inferior to 10 minutes (Sangiah and Burrows,1989) and cannot in itself explain the
exponential increase observed in the fluorescent signal over two hours. Moreover, we underline that
the design of the monofibre sensor employed: - limits the number of scattering events; - allows a
License CC-BY-NC-ND
9
probing in small volumes as well as the largest collection of photons. This sensor design further
avoids the geometrical blindness of a multi-optic fibre configuration in scattering media.
Concerning the changes occurring in the quantum efficiency of the fluorophores, the
fluorescence decay time analysis performed is well suited. Indeed, in vitro, the quantum efficiencies
of either free or protein-bound forms of NADH exhibit decay time variations in a range running
from 0.3 ns to 4 ns (Rosset et al.,1982). In vivo, however, we have not observed significant
variations in the temporal shape of the fluorescence in the 480-540 nm window. Thus, the increase
in the nRD fluorescence obtained after CAP injection is not dependent on the changes occurring in
the quantum efficiency. It might thus be directly linked with NADH concentration changes and
complex I inhibition.
Death versus CAP effect
It is clear that the increase in fluorescence induced by death cannot be directly used as a
perfect anoxic test of reference since its amplitude depends on the nature of the anaesthetic (Holt et
al.,1993; Bories and Cravendi,1994) and the concomitant transient modifications occurring in the
optical properties of the tissue (Delpy et al.,1988). Whatever these inconveniences might be, when
death occurs, mitochondria redox processes are totally suppressed and NADH remains fully
reduced. If the strength of this inhibition is referenced at 100%, then the in vivo effect obtained with
CAP reaches 40 % of the death-related changes. To fulfil this aspect from an experimental point of
view, the use of different inhibitors of the complex I, for example rotenone (Degli Esposti,1998),
would be useful. Finally, if we assume that the maximal level of the NADH fluorescence occurs
after death, the complex I inhibition could be estimated around 40% for a 300 mg/kg CAP injection.
As discussed above, if the optical properties of CAP and metabolites cannot not interfere directly
with the signal measured, the inhibition of the complex I could come from CAP itself or some of its
metabolites.
Is the complex I inhibition induced by CAP or by some of its p-NO2 metabolites?
CAP offers a unique example in terms of metabolic pathway diversity (Glazko,1987; Bories
and Cravendi,1994). It is questioned here whether some of the CAP metabolites could lead to the
License CC-BY-NC-ND
10
rise observed in the nRD NADH fluorescence after CAP administration. In this respect, CAP, poorly
soluble in water, is often formulated as an biologically inactive ester. CAP succinate, however, is
hydrolysed into the active form of CAP in the liver, lungs and kidneys. The rate at which this
hydrolysis occurs in the liver appears to be highly variable among individuals (Kroon and de
Jong,1979) and this is confirmed and quantified by our data. The increase in CAP concentration in
brain tissue is a composite function of its hydrolysis rate, the excretion of CAP succinate, the
glucuronidization into CAP glucuronide and the blood-brain barrier transfer. Since the
concentration of CAP into the cerebro-spinal fluid of the rat injected with a 165 mg/kg dose is 23 ±
5 mg/l during the first hour post-injection (Meulemans et al.,1986), in our experiments a 200 mg/kg
dose might lead to a CAP concentration around 26-28 mg/l (80-110 µM). Further, in in vitro
mitochondria preparations, a 50% inhibition of the complex I is achieved by CAP in the 4001000 µM range (Freeman and Haldar,1968; Freeman and Haldar,1970; Kroon and Arendzen,1972;
Abou-Khalil et al.,1980), while for oxidative phosphorylation a 7-17% inhibition is obtained at
100 µM (Kroon and Arendzen,1972; Abou-Khalil et al.,1980). In our experimental conditions, the
nRD complex I inhibition might thus be in the above range. It is not excluded, however, that the
large NADH rise obtained after CAP administration could come from one of its metabolites. The
consistent investigations conducted as yet on CAP metabolism (Abou-Khalil et al.,1980;
Glazko,1987; Yunis,1988; Holt et al.,1993; Bories and Cravendi,1994) point out nitroso-CAP (NOCAP) as a putative candidate. Although not identified in clinical samples (Holt et al.,1993; Bories
and Cravendi,1994), this compound might contribute to the CAP effect reported here. Thus, our in
vivo results raise once more the question related to the CAP toxicity. In this respect, several studies
(Freeman and Haldar,1968; Freeman and Haldar,1970; Abou-Khalil et al.,1980; Glazko,1987;
Yunis,1988; Holt et al.,1993; Bories and Cravendi,1994) have already suggested that the p-NO2
group may be related to the complex I inhibition. This is also supported by the fact that
thiamphenicol (TAP), differing from CAP by a methylsulfonyl moiety replacing the p-nitro group,
is inactive on complex I (Freeman and Haldar,1968; Freeman and Haldar,1970; Abou-Khalil et
al.,1980). TAP remains, however, capable to induce an inhibition of the protein synthesis like CAP.
License CC-BY-NC-ND
11
In this sense, our preliminary results (not shown) indicate that in vivo TAP does not increase the
brain autofluorescence.
Does the complex I inhibition achieved by CAP occur in preferential neuronal sets?
The complex I appears to be concentrated in brain regions containing a high density of
excitatory synapses (Higgins and Greenamyre,1996). A preference for the dendrites (60%) has also
been reported (Wong-Riley,1989; Higgins and Greenamyre,1996). The nRD area probed (200 µm3,
about 200-300 nerve cell bodies) in our experiments exhibits numerous dendrites (Descarries et
al,1982). The nRD comprises also the largest collection of 5HT cell bodies (about 50% of the whole
nerve cells) in rats (Descarries et al,1982), in cats (Chazal and Ralston,1987) and in humans (DorphPetersen,1999). Finally, the area occupied by mitochondria (10%) was estimated nearly identical in
5HT and non-5HT neurons (Descarries et al,1982). Thus, the complex I inhibition does not occur
exclusively in 5HT neurons.
Is the complex I inhibition tissue-specific in vivo?
In in vitro preparations, the inhibition induced by CAP has been observed at doses 5 to 10fold higher than those used in our experiments (Stoner,1964; Freeman and Haldar,1967; Freeman
and Haldar,1968; Freeman and Haldar,1970; Kroon and Arendzen,1972; Abou-Khalil et al.,1980;
Yunis,1988). Moreover, it was also shown that the CAP inhibition site fits in many aspects with that
of rotenone (Freeman and Haldar,1970). CAP belongs indeed to a class of polycyclic hydrophobic
inhibitors (rotenone-like) related to quinone. In vivo, the existence of brain complex I tissue-specific
isoenzymes have been suggested as well as the fact that rotenone impairs more strongly the brain
than skeletal muscles, the heart and kidneys (Higgins and Greenamyre,1996). In this respect, a
threshold effect has been proposed as an additional mechanism contributing to the tissue specificity
suggested (Davey et al,1998; Rossignol et al.,1999; Rossignol et al., 2000). The threshold value
quantifies how far the enzymatic activity can be reduced before the occurrence of significant
impairments of the oxidative phosphorylation. Data reported indicate a strong tissue difference for
the complex I, i.e. about 40-50% inhibition leads to brain energy impairments (Rossignol et al.,1999;
Rossignol et al., 2000). More detailed studies (Davey and Clark,1996; Davey et al.,1997; Davey et
License CC-BY-NC-ND
12
al,1998) further specify that the low threshold values differed among various brain regions
(hippocampus, cortex) when considering non-synaptic (60%) or synaptic mitochondria (25%).
When the 25% threshold is exceeded, mitochondria respiration is severely impaired, resulting in a
reduced synthesis of ATP. Below this threshold, the complex I activity changes and the protonelectron fluxes remain nearly unchanged. But, this is not exactly the case for NADH since its
variations are devoted to the maintenance of the fluxes at the same level. Moreover, in in vivo
situation, control of critical ratio (oxidative phosphorylation fluxes/free radical production) and
thresholds can also be influenced (Barrientos and Moraes,1999), for example by glutathione which
reduces complex I threshold (Davey et al,1998). Thus, the fact that very low doses of CAP are
capable to trigger the in vivo complex I inhibition might be related to: - complex I specific steric
factors towards rotenone binding sites; - very low threshold value of brain complex I; - in vivo redox
situation.
In vivo CAP neurotoxicity
Concerning this aspect, it is likely that brain mitochondria injury induced by CAP could
result from inhibition of complex I and protein synthesis. Our results do not imply that the inhibition
of the protein synthesis results secondarily from the complex I inhibition. They only suggest that
these two processes run in parallel when the CAP dose administered is sufficient enough for
triggering both of them. Most of the CAP side effects reported in neurological practice might be the
consequence of the brain complex I inhibition. CAP was used widely in paediatric practice until the
identification of the so-called "grey syndrome" in the late 1950s as a result of the antibiotic
treatment. Despite intensive research, the mechanism of the CAP-induced aplasia remain
unexplained (Glazko,1987; Holt et al. 1993). A tentative explanation could reside in the fact that in
vivo, CAP inhibits the complex I. Its persistent use worldwide (Kumana et al.,1993; Norris et al.
1995; Payet,1997; Niel et al.,1997; Barie,1998; Kanemitsu et al.,1999) would justify a renewed
interest in the toxicology related to this cheap and useful antibiotic. Although the precise
mechanisms of CAP neurotoxicity remain uncertain at the cellular level, the toxical manifestations
as yet reported might likely be a consequence of a reduced production of energy in the areas where
License CC-BY-NC-ND
13
the complex I threshold is exceeded. Finally, in practice, a cheap mitochondria neuroprotective
substance, given in association with CAP, would protect patients against adverse effects of the
antibiotic in several underdeveloped countries.
CAP and paradoxical sleep
The implications of our results extend beyond the field related to the antibiotic neurotoxicity.
In this respect, we recall that in 1974, in our laboratory, CAP was given orally, as an antibiotic, to
cats equipped only with the polygraphic electrodes allowing the sleep-wake states scoring. A
marked inhibition of ps occurrence was then noticed and later confirmed in mice and rats (Petitjean
et al.,1979; Fride et al.,1989; Prospero-Garcia et al.,1993). Up to now, the mechanisms related to
this effect have remained unexplained. They have been, nevertheless, at the basis of a fruitful
research on the nature of the link existing between protein synthesis and ps occurrence. These
investigations were based on the fact that PS inhibition related to CAP does not result from a specific
inhibition of the protein synthesis since TAP, a structural analogue of CAP achieving the same
protein synthesis inhibition, does not prevent PS occurrence (Petitjean et al.,1979; Fride et al.,1989,
Prospero-Garcia et al.,1993). Afterwards, the possibility of a PS dependence on energetic
metabolism emerged (Jouvet,1994; Mottin et al.,1997). Data reported here fulfil the hypothesis that
PS might indeed be energy-gated (Jouvet,1994). They underline that respiratory chain inhibition at
the complex I level is a determinant event in the CAP-related PS suppression.
Conclusion
During the past 25 years, mitochondria complex I inhibition by CAP has been considered to
be sensitive only at strong overdoses. Our report shows that, in vivo, this inhibition is effective at
clinical dosages of the substance. The CAP neurotoxicity is, at least in part, a consequence of the
complex I inhibition. This adverse property, limiting the oxidative production of ATP, might explain
why PS, an energy-gated state, is suppressed after the antibiotic administration. Moreover, the TRES
optical methods appear to be well-suited for probing brain mitochondria functions in relation with
License CC-BY-NC-ND
14
behaviour. They can be also of paramount importance for studies related to the brain toxicology of
substances.
References:

Abou-Khalil S., Abou-Khalil W.H. and Yunis A.A. (1980) Differential effects of
chloramphenicol and its nitrosoanalogue on protein synthesis and oxidative phosphorylation in rat
liver mitochondria. Biochem. Pharmacol. 29, 2605-2609.

Barie P.S. (1998) Antibiotic-resistant Gram-Positive Cocci: Implications for surgical
practice. World J. Surg 22, 118-126.

Barrientos A. and Moraes T. (1999) Titrating the Effects of Mitochondrial Complex I
Impairment in the Cell Physiology. J. Biol. Chem. 274, 16188-16197.

Bories G.F. and Cravendi J.-P. (1994) Metabolism of chloramphenicol. A story of nearly 50
years. Drug Metabol. Rev. 26, 767-783.

Cassarino D.S. and Bennet J.P. (1999) An evaluation of the role of mitochondria in
neurodegenerative diseases: mitochondrial mutations and oxidative pathology, protective nuclear
responses, and cell death in neurodegeneration. Brain Res. Rev. 29, 1-25.

Cespuglio R., Houdouin F., Oulerich M., El Mansari M. and Jouvet M. (1992) Axonal and
somato-dendritic modalities of serotonin release: their involvment in sleep preparation, triggering
and maintenance. J. Sleep Res. 1, 150-156.

Chance B., Cohen P., Jobsis F. and Schoener B. (1962) Intracellular oxidation-reduction
states in vivo. Science 137, 499-508.

Chazal G. and Ralston H.J. (1987) Serotonin-containing structures in the nucleus raphe
dorsalis of the cat: an ultrastructural analysis of dendrites, presynaptic dentrites and axon terminals.
J. Comp. Neurol. 259, 317.

Davey G.P. and Clark J.B. (1996) Threshold effects and control of oxidative phosphorylation
in nonsynaptic rat brain mitochondria. J. Neurochem. 66, 1617-1624.

Davey G.P., Canevari L. and Clark J.B. (1997) Threshold effects in synaptosomal and
nonsynaptic mitochondria from hippocampal CA1 and paramedian neocortex brain regions. J.
Neurochem. 69, 2564-2570.

Davey G.P., Peuchen S. and Clark J.B. (1998) Energy Thresholds in Brain Mitochondria.
Potential involvement in neurodegeneration. J. Biol. Chem. 273, 12753-12757.

Degli Esposti M. (1998) Inhibitors of NADH-ubiquinone reductase: an overview. Biochim.
Biophys. Acta 1364, 222-235.
License CC-BY-NC-ND
15

Delpy D.T., Cope M., van der Zee P., Arridge S., Wray S. and Wray J. (1988) Estimation of
optical pathlength through tissue from direct time of flight measurement. Phys. Med. Biol. 33, 14331442.

Descarries L., Watkins K.C., Garcia S. and Beaudet A.J. (1982) The serotonin neurons in
nucleus raphe dorsalis of adult rat: a light and electron microscope radioautographic study.
Comp.Neurol. 207, 239-254.

Dorph-Petersen, K.-A. (1999) Stereological estimation using vertical sections in a complex
tissue. J. Microsc. 195, 79-86.

Esposito L.A., Melov S., Panov A., Cottrell B.A. and Wallace D.C. (1999) Mitochondrial
disease in mouse results in increased oxidative stress. Proc. Natl. Acad. Sci. USA 96, 4820-4825.

Eto K., Tsubamoto Y., Teraushi Y., Sugiyama T., Kishimoto T., Takahashi N., Yamauchi
N., Kubota N., Murayama S., Aizawa T., Akanuma Y., Aizawa S., Kasai H., Yazaki Y. and
Kadowaki T. (1999) Role of NADH shuttle system in glucose-induced activation of mitochondrial
metabolism and insulin secretion. Science 283, 981-985.

Freeman K.B. and Haldar D. (1967) The inhibition of NADH oxidation in mammalian
mitochondria by chloramphenicol. Biochem. Biophys. Res. Comm. 28, 8-12.

Freeman K.B. and Haldar D. (1968) The inhibition of mammalian mitochondrial NADH
oxidation by chloramphenicol and its isomers and analogues. Can. J. Biochem. 46, 1003-1008.

Freeman K.B. and Haldar D. (1970) Effects of chloramphenicol and its isomers and
analogues on the mitochondrial respiratory chain. Can. J. Biochem. 48, 469-485.

Fride E., Ben-Or S. and Allweis C. (1989) Mitochondrial protein synthesis may be involved
in long-term memory formation. Pharmacol Biochem. Behav. 32, 873-878.

Glazko A.J. (1987) Early adventures in drug metabolism – Chloramphenicol. Ther. Drug
Monit. 9, 320-330.

Glinka Y., Tipton K.F. and Youdim M.B.H. (1998) Mechanism of inhibition of
mitochondrial respiratory complex I by 6-hydroxydopamine and its prevention by desferrioxamine.
Eur. J. Pharmacol. 351, 121-129.

Hashimoto M., Takeda Y., Sato T., Kawahara H., Nagano O. and Hirakawa M. (2000)
Dynamic changes of NADH fluorescence images and NADH content during spreading depression
in the cerebral cortex of gerbils, Brain Res. 872, 294-300.

Higgins D.S and Greenamyre J.T. (1996) [3H]Dihydrorotenone binding to NADH:
ubiquinone reductase (complex I) of the electron transport chain: an autoradiographic study. J.
Neurosci. 16, 3807-3816.

Holt D., Harvey D. and Hurley R. (1993) Chloramphenicol toxicity. Adv. Drug. React.
Rev.Toxicol. Rev. 12, 83-95.

Jouvet M. (1994) Paradoxical sleep mechanisms. Sleep 17: S77-S83.

Kanemitsu K, Shimada J. (1999) Neurotoxicity of antibacterial agents (tetracyclines,
chloramphenicol, aminoglycosides, beta-lactams and others), Ryoikibetsu Shokogun Shirizu 27 (Pt
2), 545-551.
License CC-BY-NC-ND
16

Kroon A.M. and Arendzen A.J. (1972) Inhibition of mitochondrial biogenesis. FEBS
Meeting, 28, 71-83.

Kroon A.M. and de Jong L. (1979) The use of antibiotics to study mitochondrial protein
synthesis. In: Methods in Enzymology, (Fleischer, S., Packer L., ed.). Academic Press, New York.

Kumana C.R., Li K.Y. and Kou M. (1993) Do chloramphenicol blood dyscrasias occur in
Hong Kong. Adv. Drug. React. Rev. Toxicol. Rev. 12, 97-106.

Meulemans A., Vicard P., Mohler J., Vulpillat M. and Pocidalo J.J. (1986) Continuous
sampling for determination of pharmacokinetics in rat cerebrospinal fluid. Antimicrob. Agents
Chemother. 30, 888-891.

Mottin S., Tran-Minh C., Laporte P., Cespuglio, R. and Jouvet M. (1993) Fiber optic timeresolved fluorescence sensor for in vitro serotonin determination, Applied Spectroscopy, 47, 590597.

Mottin S., Laporte P., Cespuglio, R. and Jouvet, M. (1997) Determination of NADH in the
rat brain during sleep wake states with an optic fibre sensor and time resolved fluorescence
procedures. Neuroscience 79, 683-693.

Norrby S.R. (1996) Neurotoxicity of carbapenem antibacterials. Drug Saf. 15, 87-90.

Norris A.H., Reilly J.P., Edelstein P.H., Brennan P.J., Schuster M.G. (1995)
Chloramphenicol for the treatment of vancomycin-resistant enterococcal infections. Clin. Infect.
Dis. 20, 1137-1142.

Niel L., Lamarque D., Coué J.C., Soarès J.L., Milleliri J.M., Boutin J.P., Mérouze F. and
Rey J.L. (1997) Chronique d'une épidémie annoncée de méningite à méningocoque (Goma, Zaïre).
Bull. Soc. Path. Ex. 90, 299-302.

Patterson G.H., Knobel S.M., Arkhammar P., Thastrup O. and Piston D.W. (2000)
Separation of the glucose-stimulated cytoplasmic and mitochondrial NAD(P)H responses in
pancreatic islet beta cells. Proc. Natl. Acad. Sci USA 97, 5203-5207.

Payet M. (1997) Trafics sur ordonnance. Croissance 401, 14-18.

Petitjean F., Buda C., Janin M., David M. and Jouvet M. (1979) Effets du chloramphenicol
sur le sommeil du chat. psychopharmacology 66, 147-153.

Prospero-Garcia O, Jimenez-Anguiano A. and Drucker-Colin R. (1993) Chloramphenicol
prevents carbachol-induced REM sleep in cats. Neurosci. Lett. 154, 168-170.

Ramilo O., Kinane B.T. and McCracken G.H. (1988) Chloramphenicol neurotoxicity. Ped.
Infect. Dis. J. 7, 358-359.

Rex A., Pfeifer L., Fink F. and Fink H. (1999) Cortical NADH during pharmacological
manipulations of the respiratory chain and spreading depression in vivo. J. Neurosci. Res. 57, 359370.

Ross J.B.A., Subramanian S. and Brand L. (1982) Spectroscopic studies of the pyridine
nucleotide coenzymes and their complexes with dehydrogenases. In: The Pyridine Nucleotide
Coenzymes (Everse, J., Anderson, B., You, K-S., ed. ). Academic Press, New York.
License CC-BY-NC-ND
17

Rossignol R., Malgat M., Mazat J.-P., and Letellier T. (1999) Threshold Effect and Tissue
Specificity. Implication for mitochondrial cytopathies. J. Biol. Chem., 274, 33426-33432.

Rossignol R., Letellier T., Malgat M., Rocher C. and Mazat J.-P. (2000) Tissue variation in
the control of oxidative phosphorylation: implication for mitochondrial diseases. Biochem. J. 347,
45–53.

Sangiah S. and Burrows G.E. (1989) Hypotension produced by rapid intravenous
administration of chloramphenicol in anaesthetised dogs. Res. Vet. Sci. 46, 62-67.

Schuchmann S., Kovacsb R., Kanna O., Heinemanna U. and Buchheim K. (2001)
Monitoring NAD(P)H autofluorescence to assess mitochondrial metabolic functions in rat
hippocampal-entorhinal cortex slices. Brain Res. Protoc. 7, 267-276.

Sick T.J. and Perez-Pinzon M.A. (1999) Optical methods for probing mitochondrial function
in brain slices. Methods 18, 104-108.

Snavely S.R. and Hodges G.R. (1984) The neurotoxicity of antibacterial agents. Ann. Intern.
Med. 101, 92-104.

Stoner C.D., Hodges T.K. and Hanson J.B. (1964) Chloramphenicol as an inhibitor of
energy-linked processes in maize mitochondria. Nature 203, 258-261.

Thomas R.J. (1994) Neurotoxicity of antibacterial therapy. South Med. J. 87, 869-874.

Watanabe M., Koishi M., Fujiwara M., Takeshita T. and Cieslik W. (1994) Development of
a new fluorescence decay measurement system using two-dimensional single-photon counting. J.
Photochem. Photobiol. A:Chem 80, 429-432.

Wong-Riley M.T.T. (1989) Cytochrome oxidase: an endogenous metabolic marker for
neuronal activity, Trends Neurosci. 12, 94-101.

Yaffe M.P. (1999) The machinery of mitochondrial inheritance and behavior. Science 283,
1493-1497.

Yunis A.A. (1988) Chloramphenicol: relation of structure to activity and toxicity. Ann. Rev.
Pharmacol. Toxicol.28, 83-100.
License CC-BY-NC-ND
18
Figure 1:
Typical time-resolved spectroscopy image derived from the nucleus Raphe Dorsalis (nRD) by using
a two-dimensional single photo-electron (SPE) counting. The spectrum of the autofluorescence is
shown in part B. Figure C shows the temporal shape corresponding to the spectral window 487508 nm. The colours of the Z-axis give the number of SPE counts for each pixel. The black part
corresponds to zero SPE, the white part to 1 SPE. Grey becomes darker throughout the SPE
counting. A rapid variation occurs from 1 to 10 SPE counting. Fifteen minutes were used to acquire
this image.
0
Time (ns)
2
4
6
8
(C)
1000
(A)
10
100
10
1
400
450
500
550
400
450
500
550
Counts
50
Counts
2500
60
40
(B)
2000
1500
30
20
10
1000
500
0
0
Wavelength (nm)
License CC-BY-NC-ND
19
Figure 2:
counts (SPE)
Time-resolved spectroscopic measurements achieved in the nRD. Part A, B and C are respectively
devoted to the 200 mg/kg dose, 400 mg/kg and 300 mg/kg. Each point represent the sum of single
photo-electron counts performed in time-resolved emission spectroscopy within the nRD (windows:
488-507 nm). The pharmacokinetic exponential fitting ( y = a+b(1-exp(-(t-d)/c)) is shown. Some
data are missing due to data processing and back-up procedure. Symbols are used for clarity.
25000
A
20000
15000
•
•
10000
-60
Time (min)
0
60
120
180
counts (SPE)
40000
20000
•
B
30000
•
Time (min)
10000
counts (SPE)
-60
0
60
120
180
C
60000
50000
Time (min)
40000
-60
License CC-BY-NC-ND
0
60
120
20
180
Figure 3:
The mean NADH fluorescence increment induced by CAP injections is quantified by (b). A paired t-test
comparison indicates that significant differences exist between the CAP doses (between 200 & 300 mg/Kg,;
0.0048
p-value=0.0007
20000
15000
10000
0.0946
Increment of fluorescence
(SPE)
between 300 & 400 mg/Kg and between 200 & 400 mg/Kg,).
5000
0
200
300
CAP (mg/kg)
License CC-BY-NC-ND
21
400
Figure 4:
In vivo fluorescence spectra derived from the nRD and induced by a Nitrogen laser excitation. Each
spectrum is the mean of seven spectra measured about 30 minutes before injection.
SPE
300
• •
B( )
250
200
150
100
50
•
A(
)
0
400
440
480
Wavelength (nm)
License CC-BY-NC-ND
22
520
Figure 5:
In vivo fluorescence spectra derived from the nRD and induced by a 355nm excitation. Each
spectrum is the mean of seven spectra measured about 30 minutes “before” or 115 to 125 minutes
“after” injection of CAP. The curves are marked by symbols which correspond to the symbols of
the Figure 2 C. The ratio (the spectrum “after” divided by the spectrum “before”) is indicated for
each curve.
Before
600
500
400
300
400
200
0
400
440
480
520
After/Before
After/Before
200
100
0
1.6
1.4
Wavelength (nm)
1.2
1
0.8
400
440
480
520 nm
800
700
600
500
400
300
200
100
C
800
600
400
200
0
0
1.6
1.4
400
440
480
520
Wavelength (nm)
1.2
1
0.8
After
400
440
480
520 nm
Before
After
800
800
700
600
500
400
300
200
100
0
500
400
400
300
200
200
100
0
440
480
520
After/Before
400
Wavelength (nm)
1.2
1
0.8
400
440
480
520 nm
Before
700
600
After
C
800
500
SPE
SPE
C
600
600
200
100
0
1.6
1.4
600
400
400
300
200
200
100
0
0
After/Before
700
1.6
1.4
1.2
1
0.8
400
440
480
520
Wavelength (nm)
400
440
License CC-BY-NC-ND
480
520 nm
23
0
1.6
1.4
1.2
1
0.8
400
440
480
520
Wavelength (nm)
400
440
480
520 nm
SPE
C
800
700
600
500
400
300
SPE
800
SPE
SPE
Before
After/Before
SPE
SPE
800
After
SPE
C
Before
SPE
800
700
600
After
Figure 6:
Events related to a CAP injection (A) or to the animal death (B). In the case illustrated (B) the animal
was sacrificed (lethal dose of barbiturates, i.e. 120 mg/2 ml for the entire animal) 8 hours after a
300 mg/kg i.p. injection of CAP succinate (the c coefficient is 3.8 ±0.5 min). The basal level of the
signal is fully reached 7 hours after the 300 mg/kg i.p. injection. The nRD fluorescence was
counts
measured at 355 nm excitation wavelength (emission wavelength: 484-508 nm).
B
50000
A
45000
40000
-60 -30
0
30
60
90
120 150
Time (min)
License CC-BY-NC-ND
24