The relationship between cytochrome c oxidase, CBF and CMRO₂ in mouse cortex: A NIRS-MRI study

Abstract

Quantifying relationships between cerebral blood flow (CBF), mitochondrial function (cytochrome c oxidase oxidation state), and metabolic rate of oxygen (CMRO₂) could provide useful insight into normal neurovascular coupling, as well as regulation of oxidative metabolism in neurological disorders. This paper uses a multimodal NIRS-MRI method to quantify these parameters in rodent brain and, in so doing, provides novel information on the regulation of oxygen metabolism by stimulating with hypercapnia or variations in oxygenation. Under hypercapnia, although oxygenation, oxidation state, and CBF increased, there was no increase in CMRO₂. Also, there was no correlation between CBF and CCO oxidation state. Conversely, changing oxygenation resulted in a strong correlation between the oxidation of CCO and CBF. This proves that the association between CBF and the redox state of CCO is not fixed and depends on the type of perturbation. Having a means to measure CBF and CCO oxidation state simultaneously will help understanding their contribution to intact neurovascular coupling and detecting abnormal cellular oxygen metabolism in many neurological disorders.

Introduction

Simultaneous measures of parameters associated with cerebral oxidative metabolism provide useful insight into the coupling of oxygen consumption (CMRO₂), with mitochondrial function and oxygen delivery. Cerebral blood flow (CBF) ensures a continuous supply of oxygen needed for energy generation. ¹ Changes in tissue oxygenation and oxygen flux will impact mitochondrial redox state—in particular the mitochondrial enzyme cytochrome c oxidase (CCO). Having the ability to monitor these parameters simultaneously, would provide insight into neurovascular coupling, the link between CBF and metabolism in the healthy brain, and how such coupling changes in brain disorders.

CCO is responsible for the majority of oxygen consumption in the cell as the final electron transfer in the electron transport chain (ETC) occurs between CCO and oxygen. CCO accepts and donates electrons in the ETC, and therefore it can be either oxidized (oxCCO) or reduced (reCCO). The oxidation state of CCO is affected by many factors which may include oxygen availability in the cell, electron flow through the ETC, ATP metabolism in the brain, and mitochondrial inhibitors. ² Therefore, CCO oxidation state might be impacted by changes in CBF, or CMRO₂.

In this paper, we use the interventions of hypercapnia and varying inspired oxygen, which are known to impact blood flow,^3,4 to perturb the system and show how it is regulated in healthy brain.

Investigating oxidative metabolism under a hypercapnia challenge is important as it is often used as a calibration procedure for MRI methods associated with quantifying CMRO₂.^5–7 It is assumed that hypercapnia, does not change CMRO₂. It is known that inducing hypercapnia increases CBF; ⁸ however, there have been conflicting reports about whether CMRO₂ is maintained at normal levels during such an increase of CBF and oxygen delivery.^9–11

Varying levels of inhaled oxygen affects oxygen content in the blood and so, for CMRO₂ to remain constant, there has to be a change in CBF and/or oxygen extraction fraction (OEF). Oxygen consumption on the mitochondrial level and the oxidation state of CCO might be affected under both hypercapnia and varying oxygenation.

We applied a multimodal technique that combines Near-Infrared Spectroscopy (NIRS) and MRI.^12,13 The multimodal NIRS-MRI technique was introduced in our previous studies as a non-invasive method to quantify absolute CBF and CMRO₂ in the rodent cortex. ¹² The ability of the technique to provide absolute data from the mouse and rat cortex was validated using a hypothermia challenge. ¹² This technique was used as well to investigate oxygen metabolism in the cuprizone mouse model of demyelination and remyelination, commonly used to study Multiple Sclerosis. ¹⁴ The development of a new NIRS algorithm that allows the quantification of absolute oxidation state and content of the mitochondrial enzyme CCO, ¹⁵ enables us to study the relationship between oxygen supply (CBF) and oxygen consumption (CCO and CMRO₂).

In this study, arterial Spin Labeling (ASL) MRI was applied to measure CBF. NIRS was used to quantify absolute measures of the different absorbers or chromophores. This includes, oxCCO, reCCO, oxyhemoglobin (HbO) and deoxyhemoglobin (dHb). The oxidation state of the enzyme was calculated as the difference between the concentration of oxCCO and reCCO. As a measure of tissue oxygenation, we calculated the microvascular saturation (HbO/(dHb + HbO)) in the cerebral tissue (S_tO₂). OEF and CMRO₂ were calculated using the modified Fick Principle.^16,17

Material and methods

Animals

C57BL/6J male mice were housed and maintained in the University of Calgary Animal Care facility with a 12 h light and dark cycle with access to water and food pellets ad libitum. Animal protocols were approved by the Health Sciences Animal Care Committee (HSACC) of the University of Calgary and conformed to the guidelines established by the Canadian Council of Animal Care (CCAC). The experiments conducted in this study have been reported in compliance with the Animal Research: Reporting of In Vivo Experiments (ARRIVAL) guidelines. ¹⁸

NIRS-MRI imaging

The experimental setup of the multimodal NIRS-MRI system was described in detail in our previous publications.^12,14,15 A schematic design of the system is shown in Figure S1 in the Supplemental Materials. The system consists of a 9.4 T horizontal bore MRI (Bruker Avance console, Bruker Biospin GmbH, Rheinstetten, Germany) with a 35 mm quadrature volume coil, and a continuous-wave broadband NIRS system, which were operated by two independent computers. MRI and NIRS data were manually synchronized with a time resolution of less than a second.

Near-Infrared spectroscopy (NIRS)

The NIRS system included a broadband white light source (77501, Oriel Instruments Inc., USA), two 5 m long Hard-Clad Silica Core Multimode optical fibres (FT1000EMT, Thorlabs Inc, USA) with a 1000 µm core diameter and 0.39 NA, a spectrograph (Shamrock 303i, Andor Technology Inc, Northern Ireland), and a CCD camera (iDus 420, Andor Technology Inc, Northern Ireland). A GRIN lens with a 1.80 mm diameter and 0.55 NA (#64-525, Edmund Optics, USA) and a 90° prism with a leg length of 2 mm (#45-524, TECHSPEC NSF11, Edmund Optics, USA) were glued to the end of the fibers to collimate the light and direct it from the source fibre into the tissue and from the tissue into the detector fibre. Attenuation spectra were collected from the mouse cortex over the range 705–960 nm.

Hair was removed from the animal head, and the fibers, with a thin layer of glycerol ¹⁹ on the prisms, were secured on top of the head, near bregma, with the posterior edge of the prisms placed on the line between the two external auditory meatus (within 1 mm of the interaural line). Fibers were spaced with a black rubber separator (4 mm width) and secured by covering the area with masking tape connected to the outside of the MRI coil cradle.

Magnetic resonance imaging (MRI)

Arterial Spin Labeling (ASL) was applied to measure perfusion. A single axial slice was acquired around the bregma, where the optic fibers were located, a CASL-HASTE sequence with the following parameters: TR = 3000 ms, TE = 13.5 ms, FOV = 25.6 × 25.6 mm, matrix size = 128 × 128 pixels, slice thickness = 1.5 mm, 16 averages. Four perfusion images were collected per measurement: 2 control images and 2 tagged images, to correct for magnetization transfer. Following these images, a T1 map was obtained in the same location using a RARE-VTR sequence where effective TE = 20 ms, TR = 100, 500, 1000, 3000, and 7500 ms. Together, the four perfusion images and the T1 map were collected over a period of 14 min.

Study design

Anesthesia was induced with 5% isoflurane added to a gas mixture of 70% N₂ and 30% O₂. During data acquisition, anesthesia was maintained at 1.7%–2% isoflurane. The isoflurane level was reduced if respiration rate declined. The isoflurane was added after the O₂ and N₂ was mixed to the specified percentages. To ensure a stable physiology, heart rate, breathing rate and S_aO₂ were monitored by the MouseOx MRI-compatible pulse oximeter (Starr Life Sciences, USA) placed on the shaved left thigh of the animal. Core temperature was monitored by a rectal thermometer and maintained at 36.5 ± 0.1°C. The animal was placed on a heating pad in the MRI coil, together with the NIRS fibers and the pulse oximeter, and positioned for optimal image quality. Two sets of experiments were conducted to investigate the association of CCO with CBF and CMRO₂ and are detailed below.

Hypercapnia study

9 mice (24–26 g, 17 weeks old from The Jackson Laboratory, US) were used in this study. The mice were physiologically stabilized prior to data acquisition. NIR spectra were collected over a time course of 65 minutes, where the gas mixture was intermittently varied. The experiment started with a baseline measurement where the mice were supplied by a 70% N₂, 30% O₂, 0% CO₂ mixture for 15 minutes. Then, MRI-ASL acquisition was initiated for perfusion measurement, and required 14 minutes. Afterwards, the gas mixture was changed to 65% N₂, 30% O₂, 5% CO₂. After stabilization for 10 minutes, MRI-ASL measurement was repeated. The gas mixture was restored to 70% N₂, 30% O₂, 0% CO₂. An anoxia pulse (100% N₂, 0% O₂, 0% CO₂) was given after stabilization of 5 minutes, for a short period of 50 seconds. Afterwards, the oxygen level was restored to 30% for recovery. NIRS data were averaged over the time of the ASL acquisition period for the extraction of chromophores concentrations.

Varying oxygenation study

5 male mice (19–24 g, 8 weeks old from Charles River, Québec, Canada) were used in this study. The mice were physiologically stabilized prior to data acquisition. NIR spectra were collected over a time course of 80 minutes, under 4 different levels of inhaled oxygen (100%, 30%, 20%, and 10%) balanced with N₂. At each oxygen level, the mice were monitored for 5 minutes to ensure stable physiology, and then MRI-ASL was acquired for 14 min. At the end, a short anoxia pulse (50 seconds) was given to the mice, where the air mixture ratio was altered to 100% N₂/0% O₂. After the anoxia pulse, the oxygen level was restored to 30% for recovery. In this study, normoxia is defined as the inhalation of 30% O₂ under anestheisa. The inhalation of higher O₂ levels (100%) was defined as hyperoxia, while the inhalation of lower levels of O₂ was defined as very mild (20%) and severe (10%) hypoxia. NIRS data were averaged over the time of the ASL acquisition period at each oxygenation level for the extraction of CCO concentration.

Data analysis

CBF quantification

Perfusion maps were generated out of the collected perfusion images and the T₁ map.^12,20 In the perfusion map, CBF was calculated on a voxel-by-voxel basis using the following equation:^21,22

$$f=\frac{1}{T_1^{\mathrm{\hspace{0.17em}\prime }}}\frac{\lambda }{2\alpha }\left(\frac{M_c-M_L}{M_c}\right)$$ (1)

where $f$ is the tissue perfusion in (ml ‧ 100 g⁻¹ ‧ min⁻¹), $\lambda =0.9$ is the blood-brain partition coefficient,^23,24 $M_c$ is the average signal of the two control images, $M_L$ is the average signal of the two tagged images, $\alpha =0.675$ is the spin-labelling efficiency, ²⁰ and $T_1^{\mathrm{\hspace{0.17em}\prime }}\hspace{0.17em}$ is the measured $T_1$ value in each voxel. Individual perfusion maps were calculated for each animal. An ROI from the cortex was created using a tagged image (better contrast), and which encompassed a region that is estimated to be similar to the region of sensitivity for NIRS. This ROI was then used on the perfusion maps to obtain a mean ± SD (ml ‧ 100 g⁻¹ ‧ min⁻¹) for perfusion at each time point.

A primary and a secondary analysis individuals drew the ROIs in a blinded fashion and the results for each were compared. If the difference was over 5% the analysis was repeated. If not, the data from the primary analysis person were used.

Chromophores quantification

The absolute concentrations of the main chromophores in the tissue: deoxyhemoglobin (dHb), oxidized CCO (oxCCO) and reduced CCO (reCCO), were quantified from the measured attenuation light using the algorithm described in our previous studies.^12,15,25 The well-defined spectral absorption feature of water at 800–850 nm and the cerebral water content, which is assumed to be 80% in adult rodents, ²⁶ were used to estimate the pathlength that light travels through the tissue. The algorithm applied multilinear regression to fit the measured attenuation spectra to the specific absorption coefficients of each chromophore, downloaded from the University College of London medical physics website, ²⁷ over a specific range of wavelengths.

The total concentration of CCO (totCCO) was calculated as the sum of oxCCO and reCCO. We are defining the difference between the concentration of oxCCO and reCCO as the “oxidation state” of the enzyme. In much of the biochemistry literature, oxidation state is defined as the ratio between oxCCO and reCCO. In order to be able to compare our data to other NIRS studies monitoring the difference spectra of CCO and reporting the change of oxidation state (ΔµM), we chose to show the oxidation state of the enzyme as the difference between oxCCO and reCCO.

In normocapnia (inhalation of 70% N₂ and 30% O₂), the concentration of tHb in the cerebral tissue was determined using the anoxia pulse method,^25,28 in which it is assumed that during the anoxia pulse the total hemoglobin is approximated by the total dHb concentration ([tHb] = [dHb]). The anoxia pulse was obtained by switching the inspired gas to 0% O₂ and 100% N₂ for 50 seconds. This period was sufficient to reach steady state of dHb while minimizing the effect on other physiological parameters.^25,28

This concentration of tHb does not represent the accurate concentration of tHb during the inhalation of 5% CO₂, since CO₂ affects cerebral blood volume (CBV). ²⁹ Ideally, another anoxia pulse should be given right after the CO₂ intervention period. Nevertheless, to avoid unnecessary stress for the animals, only one anoxia pulse was applied. To estimate tHb concentration during the inhalation of 5% CO₂, the Grubb’s exponent^30,31 was used to correlate between the change in CBV and CBF:

$$\mathit{log}\frac{\mathit{CBV}}{{\mathit{CBV}}_0}=0.40\times \mathit{log}\frac{\mathit{CBF}}{{\mathit{CBF}}_0}$$ (2)

tHb concentration changes proportionally to CBV, ³² i.e.:

$$\mathit{CBV}=k·\mathit{tHb}, k=\frac{1}{Hb·R·D}$$ (3)

where Hb is the hemoglobin content in the large vessels blood, R is the cerebral to large vessel hematocrit ratio, and D is the brain density. The Hb content in the blood along with its spatial distribution are assumed to be constant before and after the inhalation of 5% CO₂. Therefore, substituting equation (3) in equation (2), would provide the correlation between the change in tHb and CBF:

$$\mathit{log}\frac{\left[\mathit{tHb}\right]}{{\left[\mathit{tHb}\right]}_0}=0.40\times \mathit{log}\frac{\mathit{CBF}}{{\mathit{CBF}}_0}$$ (4)

where [tHb]₀ and CBF₀ are tHb and CBF measured when no CO₂ was inhaled and [tHb] and CBF are the concentration of tHb and CBF during the 5% CO₂ inhalation.

S_tO₂ and OEF quantification

The oxygen saturation in tissue (S_tO₂) was determined using dHb and tHb concentrations and the following equation:

$$S_t{O}_2=\frac{\left[\mathit{tHb}\right]-\left[\mathit{dHb}\right]}{\left[\mathit{tHb}\right]}$$ (5)

OEF is related to the arteriovenous oxygen saturation difference:^33,34 arterial (S_aO₂) and (S_vO₂):

$$\mathit{OEF}=\frac{{S_a{O}_2-S}_v{O}_2}{S_a{O}_2}$$ (6)

S_aO₂ is the arterial oxygen saturation, measured with the MouseOx MRI-compatible pulse oximeter from the thigh of the animal. S_vO₂ is thevenous oxygen saturation, estimated from S_aO₂ and S_tO₂ values: ¹²

$$S_v{O}_2=\frac{4}{3}{S}_t{O}_2-\frac{1}{3}{S}_a{O}_2$$ (7)

as tissue [Hb] consists of 25% arterial [Hb] and 75% venous [Hb].^35,36 Therefore, OEF can be rewritten as the following:

$$\mathit{OEF}=\frac{4}{3}\frac{\left(S_a{O}_2-S_t{O}_2\right)}{S_a{O}_2}$$ (8)

CMRO₂ quantification

CMRO₂ was quantified using the modified Fick principle: ³⁶

$${\mathit{CMRO}}_2=k\times \left[Hb\right]\times S_a{O}_2\times \mathit{CBF}\times \mathit{OEF}$$ (9)

where k = 1.39 (ml(O₂)/g(Hb)) is a factor describing the amount of oxygen bound to Hb when completely saturated, ³⁷ and [Hb] is the concentration of hemoglobin (g/l). By substituting equation (8) in equation (9), and combining the average CBF value (ml ‧ 100g⁻¹ ‧ min⁻¹) from the ROI, with NIRS and pulse oximeter data, CMRO₂ (ml_(O2) ‧ 100 g⁻¹ ‧ min⁻¹) can be quantified using the following equation:

$${\mathit{CMRO}}_2=k\times \mathit{CBF}\times \frac{4}{3}\left(S_a{O}_2-S_t{O}_2\right)\times \frac{\left[\mathit{tHb}\right]}{\mathit{CBV}\times \rho }$$ (10)

S_aO₂, S_tO₂ and tHb values were averaged over the period of perfusion acquisition. $\mathit{CBV}\times \rho$ is the cerebral blood volume to tissue volume ratio, where the CBV in the cortical GM is assumed to be 0.045 ml/g, ³⁸ and the cerebral tissue density $\rho =1.05\hspace{0.17em}\text{g}/\text{m}\text{l}$ . ³⁹ This factor was used to normalize the NIRS measurement of tHb to determine the concentration of Hb in the blood compartment of tissue of interest.

Statistical analysis

Sample size (n) was determined using a power analysis. One of the parameters with a large variance is the perfusion data. Based on the expected mean and variance at baseline in control mice, a difference of 20 ml ‧ 100g⁻¹ ‧ min⁻¹ (α = 0.05, power level: 95%) could be detected with n = 4. To avoid type I and II errors, sample size was increased to be n = 9 in the hypercapnia study and n = 5 in the varying oxygenation study.

A paired t-test was conducted to determine whether there was a significant change in oxidative metabolism correlates when measured before and during 5% CO₂ intervention. A one-way repeated measured ANOVA, followed by an LSD post-hoc test, was conducted to determine whether there was a significant change between the varying oxygen levels. All data were expressed as mean ± SD, and p < 0.05 was considered statistically significant. All statistical analyses were performed in IBM SPSS Statistics v24.

Results

Hypercapnia study

Multimodal NIRS-MRI was used to quantify the oxidation state of CCO and CBF, in addition to tHb, S_tO₂, OEF, and CMRO₂ from the cortex of mice under anesthesia, before and during the inhalation of 5% CO₂.

Since biochemistry literature reports the oxidation state of CCO as the ratio between the oxidized and the reduced forms of the enzyme, and we report it as the difference between these forms, we found it useful to show the relationship between these calculations. Figure S2 in Supplemental Materials, shows a linear relationship (r = 0.94, p < 0.0001) between these two metrics, where: $\frac{\left[\mathit{oxCCO}\right]}{\left[\mathit{reCCO}\right]}=0.84·\left(\left[\mathit{oxCCO}\right]-\left[\mathit{reCCO}\right]\right)+0.9$

The tHb during normocapnia was quantified at the end of the protocol, with the anoxia pulse. Changes in tHb during hypercapnia were quantified using the Grubb equation.

Figure 1 shows the absolute values of these metabolic parameters under both conditions. There was a significant increase in oxCCO (+2.9%, p = 0.003) and a significant decrease in reCCO (−3.1%, p = 0.008) while the total amount of the enzyme remained constant (p = 0.1) during hypercapnia. An example of the absorption spectra collected from the mouse cortex under each condition, and their second derivative is shown in Figure S3 in the Supplemental Materials. Mean values of CBF were quantified from the corresponding ROI in the perfusion maps. A representative example of the perfusion maps is shown in Figure S4 in the Supplemental Materials. CBF increased significantly (+10.4%, p = 0.004) during hypercapnia while OEF values decreased significantly (−5.3%, p = 0.006) to maintain a fixed CMRO₂ (p = 0.1). There was a significant increase in tHb concentration (+4.1%, p = 0.009) and S_tO₂ (+0.7%, p = 0.003) during hypercapnia, whereas S_aO₂ values didn’t change (p = 0.4). Table 1 summarizes the values (mean ± SD) of these parameters during the inhalation of 0% and 5% CO₂.

Figure 1.

Quantification of oxCCO (a), reCCO (b), totCCO (c), CBF (d), OEF (e), CMRO₂ (f), tHb (g), S_tO₂ (h), and S_aO₂ (i), in each of the 9 mice before (0% CO₂) and during hypercapnia (5% CO₂) Each symbol represents a different animal. Statistical analysis was performed comparing normocapnia and hypercapnia (** – p ≤ 0.01) using paired t-test.

Table 1.

The mean ± SD of the measured metabolic correlates from (n = 9) mice during the inhalation of 0% and 5% CO₂.

	0% CO2	5% CO2	% Change
CBF (ml ‧ 100 g−1 ‧ min−1)	223.6 ± 6.2	246.8 ± 18.0**	+10.4%
OEF	0.19 ± 0.02	0.18 ± 0.03**	−5.3%
CMRO2 (ml ‧ 100 g−1 ‧ min−1)	4.1 ± 1.3	4.5 ± 1.5	+9.7%
SaO2 (%)	97.6 ± 1.3	97.4 ± 1.3	−0.2%
StO2 (%)	83.8 ± 2.4	84.4 ± 2.6**	+0.7%
tHb (µM)	53.0 ± 16.4	55.2 ± 17.4**	+4.1%
oxCCO (µM)	2.77 ± 0.36	2.86 ± 0.37**	+2.9%
reCCO (µM)	1.51 ± 0.39	1.46 ± 0.39**	−3.1%
totCCO (µM)	4.28 ± 0.5	4.32 ± 0.50	+0.8%
Oxidation state (µM) (oxCCO–reCCO)	1.26 ± 0.53	1.39 ± 0.54**	+10.3%

CBF: cerebral blood flow; OEF: oxygen extraction fraction. % Change is the percent of change obtained by comparing the initial (0%) and the final (5%) values. ** – p ≤ 0.01 using paired t-test.

To understand the link between the different physiological parameters related to oxidative metabolism, we examined whether there is an association between these parameters following the change of the inspired CO₂ level. Figure 2(a) and (b) shows a significant positive linear correlation between S_tO₂ and CBF (r = 0.53, p = 0.02), and a strong negative correlation between OEF and CBF (r = −0.53, p = 0.02).

Figure 2.

Correlation of S_tO₂ (a, c) and OEF (b, d) with CBF and CCO oxidation state. S_tO₂ has a positive linear correlation (r = 0.53, p = 0.02) and OEF has a negative linear correlation (r = −0.53, p = 0.02) with CBF. S_tO₂ has a positive linear correlation (r = 0.55, p = 0.02) and OEF has a negative linear correlation (r = −0.64, p = 0.004) with the oxidation state of CCO. Each symbol represents a different animal. Best fit lines (solid lines) and 95% confidence intervals (blue dotted lines) are plotted. Correlation coefficients (r) and p-values (p) are displayed.

Figure 2(c) and (d) shows the correlation of the oxidation state of CCO, which is calculated as the difference between the concentration of oxCCO and reCCO (oxCCO–reCCO), with S_tO₂ and OEF. The oxidation state of CCO had a significant positive linear correlation with S_tO₂ (r = 0.55, p = 0.02), and a significant negative linear correlation with OEF (r = −0.64, p = 0.004). These correlations show that there is an association between the oxygen availability in the tissue, its extraction from the microvasculature, and the oxidation state of CCO. Figure 3 shows that there was no linear correlation between the oxidation state of CCO and CBF (r = 0.2, p = 0.4), indicating that under hypercapnia, there was no direct association between CBF and CCO oxidation state.

Figure 3.

Correlation of CCO oxidation state with CBF. The oxidation state of CCO is calculated as [oxCCO] minus [reCCO]. CCO oxidation state has no significant correlation with CBF (r = 0.2, p = 0.4). Each symbol represents a different animal. Best fit line (solid line) and 95% confidence intervals (blue dotted lines) are plotted. Correlation coefficients (r) and p-values (p) are displayed.

Varying oxygenation study

This study was set to define the response of CBF and the oxidation state of CCO in anesthetized mice, when an external oxygenation challenge is imposed. S_tO₂, OEF and CMRO₂ were not quantified in this study since tHb was not measured at each oxygenation level, to prevent additional stress on the animals induced by repeated anoxia pulses.

An example of the absorption spectra collected from the mouse cortex under each oxygenation level, and their second derivative is shown in Figure S5 in the Supplemental Materials. Figure 4 shows the absolute values of the measured parameters under the different conditions: Normoxia (O₂ = 30%), hyperoxia (O₂ = 100%), very mild hypoxia (O₂ = 20%), and severe hypoxia (O₂ = 10%). S_aO₂ varied between animals but declined in proportion to the inhaled oxygen level. The significant drop was between 30% and 20% O₂ (−1.2%, p = 0.04). Although the decline in S_aO₂ between 30% and 10% O₂ was −17.7%, it was not a statistically significant difference (p = 0.1). At 10% O₂, some mice had an S_aO₂ of more than 90% while others had a notable decrease (<70%). The mean S_aO₂ observed here at 10% O₂ (78.16 ± 17.74%) is consistent with S_aO₂ obtained in isoflurane-anesthetized rats (75 ± 3%) spontaneously breathing 9% O₂ premixed with balance N₂, ⁴⁰ but higher than S_aO₂ values (40.7 ± 9.9%) obtained in isoflurane-anesthetized piglets mechanically ventilated and breathing 10% O₂. ⁴¹ The high inter-subject variability might cause the non-significant change in S_aO₂ despite the large difference between the average arterial oxygenation under the 4 different conditions. This inter-subject variability could impede a clear interpretation of the change in other correlations.

Figure 4.

Quantification of SaO₂ (a) and CBF (b) in addition to the oxidation state (c) and the total concentration (totCCO) of CCO (d) in each of the 5 mice during the inhalation of different O₂ levels. Each symbol represents a different animal. A one-way repeated measured ANOVA, followed by an LSD post-hoc test, was conducted to determine whether there is a significant change between the different O₂ levels (* - p < 0.05, ** - p ≤ 0.01, *** - p ≤ 0.001).

Compared to normoxia, CBF values decreased significantly under hyperoxia (−10.1%, p = 0.03), and severe hypoxia (−20.7%, p = 0.03), while there was no significant change under very mild hypoxia (p = 0.4). A representative example of the perfusion maps is shown in Figure S6 in the Supplemental Materials.

Oxidation of CCO increased significantly (+31.3%, p < 0.001) under hyperoxia and decreased significantly both under very mild (−25.7%, p < 0.001) and severe hypoxia (−205%, p < 0.001). The total amount of the enzyme did not change during the inhalation of different oxygen levels (p > 0.05). Table 2 summarizes the values (mean ± SD) of these parameters during the inhalation of the different oxygen levels, and the percentage change of these values compared to normoxic values.

Table 2.

The measured metabolic correlates (Mean ± SD, n = 5) from the mouse cortex during the inhalation of different levels of O₂: 30% O₂ (Normoxia), 100% O₂ (Hyperoxia), 20% O₂ (Mild Hypoxia), and 10% O₂ (Severe Hypoxia).

	Normoxia	Hyperoxia	Mild Hypoxia	Severe Hypoxia
SaO2 (%)	94.98 ± 3.60	97.28 ± 1.77	93.84 ± 3.39*	78.16 ± 17.74
% Change		+2.4%	−1.2%	−17.7%
CBF(ml ‧ 100g−1 ‧ min−1)	219.3 ± 31.8	197.1 ± 35.7*	222.4 ± 33.4	173.8 ± 21.6*
%Change		−10.1%	+1.4%	−20.7%
oxCCO (µM)	3.18 ± 0.15	3.38 ± 0.10*	2.94 ± 0.12**	1.28 ± 0.50**
%Change		+6.3%	−7.5%	−59.7%
reCCO (µM)	1.20 ± 0.42	0.78 ± 0.31**	1.47 ± 0.47**	3.37 ± 0.81**
%Change		−35.0%	+22.5%	+180.83%
totCCO (µM)	4.37 ± 0.53	4.16 ± 0.37	4.41 ± 0.53	4.65 ± 0.68
%Change		−4.8%	+0.91%	+6.41%
Oxidation State (µM) (oxCCO–reCCO)	1.98 ± 0.33	2.60 ± 0.28***	1.47 ± 0.44***	−2.09 ± 1.15***
%Change		+31.3%	−25.7%	−205%

% Change is the percent of change obtained by comparing the values measured at each condition and the values measured under normoxia. A one-way repeated measured ANOVA, followed by an LSD post-hoc test, was conducted to determine whether there is a significant change between inhaled O₂ levels (* – p < 0.05, ** – p ≤ 0.01, *** – p ≤ 0.001).

To examine the relationship between CBF and oxygen saturation, CBF was plotted versus S_aO₂ in Figure 5(a). There was an overall significant positive correlation between CBF and S_aO₂ (r = 0.51, p = 0.02). The relationship between the oxidation state of CCO and oxygen saturation was displayed in Figure 5(b). The oxidation state of CCO showed an overall significant positive correlation with S_aO₂ (r = 0.73, p = 0.0003). As both oxidation state and CBF changed significantly with the oxygenation level, the relationship of oxidation state and CBF was tested and plotted in Figure 5(c). There was a significant positive correlation between oxidation state and CBF (r = 0.56, p = 0.01) over the examined oxygenation range.

Figure 5.

Correlation between the oxidation state of CCO, CBF and S_aO₂. CBF (a) and CCO oxidation state (b) have a significant positive correlation with S_aO₂. CCO oxidation state (c) has a significant correlation with CBF (r = 0.56, p = 0.01) over the examined oxygenation range. Each symbol represents a different animal. Best fit lines (solid lines) and 95% confidence intervals (blue dotted lines) are plotted. Correlation coefficients (r) and p-values (p) are displayed.

Discussion

Non-invasive imaging of cerebral oxygen delivery and consumption is crucial to understand neurovascular coupling and oxidative metabolism. We combine NIRS and MRI to determine the correlation between mitochondrial status and perfusion in the cortex of mice, when exposed to hypercapnia or varying oxygen levels.

These global perturbations were chosen due to a spatial limitation imposed by the geometry of the optic fibres. The optic fibres used in this NIRS-MRI system allow us to look only at NIRS data within the cortex along the top of the brain. However, since this NIRS-MRI technique is applicable to other NIRS fibre geometries, localised data could be investigated in future, for instance, by implanting an optic fibre in the brain.^42,43

Hypercapnia stimulates CBF and tHb

Increasing the level of inhaled CO₂ caused a significant increase in CBF and tHb. Assuming that Hb content in the blood and its spatial distribution are constant, tHb can be used as an indicator for changes in CBV.^29,32,44 Thus, the increase in tHb is consistent with increased CBV reported in previous studies. ⁴⁵ Both human ⁴⁶ and animal studies^29,47 have showed that CO₂ induces a higher increase in the blood flow than the vascular diameter, therefore, the response of CBF to CO₂ is higher than that of CBV. ³² This relationship is characterized by the Grubb’s exponent.^30,31

A strong relationship between CO₂ levels in blood and CBF has been reported, both in animal^48,49 and human ⁹ studies. The increase in CBF (10.4%) and tHb (4.1%) (Table 1), when exposed to a relatively long period (10 minutes) of 5% CO₂ inhalation under isoflurane, was comparable to the increase in CBF (9.6%) and CBV (5.6%) obtained in awake mice ²⁹ during only 50 seconds of 5% CO₂ challenge. Moreover, an increase of 51% in CBF was reported in awake rats, while only 25% increase was observed in anesthetized rats under hypercapnia (5% CO₂). ⁵⁰ This shows that under isoflurane anesthesia, hypercapnia-induced CBF changes may be smaller than the change observed in awake animals.

Isoflurane is known to increase CBF, and this increase is dose dependent. It has been shown that under 2% isoflurane, CBF in the cerebral cortex of anesthetized rats increased by ∼23%, compared to awake rats. ⁵⁰ A similar increase was found in an older study using quantitative autoradiographic technique in rats breathing 1.5–2% isoflurane. ⁵¹ Their reported mean values showed a trend of similar increase in CBF, but not statistically significant as the variance was high. A study measuring brain pO₂ in rats using electron paramagnetic resonance showed no increase in oxygenation when awake animals were anesthetized with 2% isoflurane, implying that there were no major changes in CBF. ⁵² Previous studies have consistently reported that the autoregulatory control is preserved under 0.7 to 2.8% isoflurane anesthesia.^53,54 In addition, it has been shown that neurovascular coupling (measured with local field potentials), following single-pulse stimulation and CO₂ stimulation, was unchanged in the rat somatosensory cortex, across the range of isoflurane from 1.1% to 2.1%, although the CBF response was dose-dependent.^55,56

In the current study we investigated the cerebral cortex using an isoflurane dose between 1.7% and 2%. Therefore, we believe that isoflurane had mildly increased CBF in our region of interest but did not impair cerebrovascular reactivity (CVR). It is true that the magnitude of change following hypercapnia induction is lower than the change expected in awake animals, but we still see the same trend of change, as expected. CVR is dependent on baseline CBF regardless of the anesthesia protocol used. ⁵⁷ Thus we suggest that although isoflurane may have impacted our results, we suggest that the impact is minimal, and the anesthesia is not affecting our overall conclusion.

Hypercapnia increases the oxidation state of CCO

The oxidation state of the mitochondrial enzyme CCO increased during hypercapnia. Such an increase has been reported in human studies under the inhalation of 6% CO₂.^58,59 The oxidation state of the enzyme is affected by several factors, such as oxygen availability in the tissue, pH, and electron flux through the enzyme. ² The tissue oxygen content was also higher, as indicated by an increase in S_tO₂. The correlation of CCO oxidation with S_tO₂ demonstrates that high oxygen concentration can increase oxidation state. The lack of correlation between CCO oxidation state and CBF indicates that there are other factors, in addition to oxygen supply, that may have changed due to hypercapnia and have caused to an increase both in CBF and oxidation of CCO. CCO oxidation state also correlated with oxygenation but not CBF in a study using hypercapnia in humans. ⁵⁸

During hypercapnia, the excess of CO₂ in the brain results in a decrease in the pH, i.e., an increase in the concentration of [H⁺] in the tissue. ⁶⁰ The accumulation of [H⁺] in the intermembrane of mitochondria slows down the process of electrons transport through the ETC enzymes, which is likely to affect the oxidation state of CCO and consequently decreases the consumption of oxygen. In addition, increased oxidation of CCO during hypercapnia can be caused by changes in the supply of the ETC substrates, and alteration of ATP metabolism in the brain. ⁶¹ It has been suggested, that increased oxidation of cerebral CCO in living piglets exposed to mild hypercapnia, was caused by alterations of carbohydrate metabolism. ⁶² In vitro studies showed that when isolated mitochondria were incubated with 5% and 15% CO₂, there was a significant decline in oxygen uptake and in the oxidation of the ETC substrates, especially succinate. ⁶¹

Moreover, hypercapnia increases nitric oxide (NO) production. ⁶³ NO stimulates blood flow and is a competitive inhibitor of CCO. ⁶⁴ NO may reduce CCO’s affinity to oxygen. ⁶⁵ Thus, the impact of NO on CCO could contribute to the maintenance of stable CMRO₂ despite the increase in CBF and oxygen availability (S_tO₂). ⁶⁶ These data indicate that hypercapnia may reduce the change in CMRO₂ for a given change in CCO oxidation state.

Hypercapnia does not change CMRO₂

There have been conflicting reports about whether hypercapnia can increase CMRO₂.^{9–11,67–69} In the current study, CBF and CCO oxidation state increased, but CMRO₂ did not change. This is consistent with studies which found little or no change in cerebral metabolism under similar hypercapnia conditions, or during moderate changes in CBF. ⁷⁰ For instance, in humans, an inhalation of 5% CO₂ increased CBF by ∼50% while no change was observed in the cerebral oxygen consumption. ⁹ This demonstrates that in case of hypercapnia, increases in CBF are caused primary by the effect that CO₂ has on blood vessel diameter and not due to cerebral tissue metabolism. PET studies have shown that a large increase in CBF (30–50%) was needed to support a relatively small change (∼5%) in CMRO₂. ⁷¹ The nonlinear coupling between CBF and CMRO₂ observed here, can be explained by the reduction in OEF values (−5.6%), and the strong negative linear correlation between OEF and CBF (Figure 2(b)). These data clearly indicate that when blood flow increases, less oxygen is extracted from the blood to maintain stable CMRO₂.^70,72

The increased oxidation of CCO, accompanied with decreased OEF indicate that higher oxCCO does not mean necessarily that oxygen consumption would be higher. We suggest that the regulation of CCO has changed due to hypercapnia, resulting in a tight coupling between the metabolic requirements of the brain, which are not expected to change during hypercapnia, and oxidative metabolism or CMRO₂.^62,66

The effect of hyperoxia on CBF and CCO oxidation state

Hyperoxia induces vasoconstriction^9,48,73 leading to a decline in CBF. The decrease in CBF obtained here (−10.1%) is in a good agreement with the decline reported in the gray matter of isoflurane-anesthetized rhesus monkeys (−9.4 ± 2%) measured with ASL-MRI ⁷³ and in awake humans (−13%) measured by PET. ⁹

In hyperoxia, increased inhaled oxygen resulted in increased S_tO₂ and oxCCO. This shows that increased oxygen tension directly increased the oxidation of CCO. Increased oxidation of CCO also indicates that under normoxia CCO was not fully oxidized. This is consistent with previous studies reporting that CCO can be further oxidized by hyperoxia.^15,58,59 Although not quantified in this study, we expect OEF to increase as a response to the reduction in CBF, in order to maintain stable oxygen consumption.

The effect of hypoxia on CBF and CCO oxidation state

Hypoxia yielded a different response depending on the severity and duration of the hypoxic challenge. For example, at 20% O₂, although S_aO₂ was significantly low relative to 30% O₂, CBF did not change. In severe hypoxia (10% O₂) CBF decreased (−20.7%). Autonomic responses occur during hypoxia to compensate for the insufficient oxygen availability in tissue. These responses include vasodilation, increases in respiration rate, heart rate, CBV, and CBF. ⁷⁴ Hypoxia-induced increases in CBF have been reported in awake rats ⁷⁵ while decreases were reported in other studies where awake rats were exposed to mild hypoxia with non-controlled levels of CO₂.^74,76 No change in CBF in response to mild hypoxia has also been reported in humans and animals.^4,77 In our case, since CO₂ levels were not controlled, hypoxia may have induced hyperventilation, which reduced P_aCO₂. We suggest that the hypoxia-induced hypocapnia led to a reduction in CBF instead of a hypoxia-driven increase in CBF.

Under severe hypoxia, the compensatory mechanism of CBF failed and a significant decrease in CBF was observed. This also accords with hypoxia-induced CBF decreases found in isoflurane-anesthetized and spontaneously breathing ⁷⁴ or mechanically ventilated ⁷⁷ rats exposed to severe hypoxia (9% O₂). One contributing factor could be vascular failure during extreme hypoxia and a subsequent decline in blood pressure, ⁷⁴ which was not monitored or controlled for during the experiment. Some of the decrease in CBF is likely associated with hypocapnia induced by hypoxia and the suppression of the autonomic responses by isoflurane. Moreover, it was reported that CBF response to hypoxia is brain area-specific. ⁷⁸ During hypoxia (5% O₂), subcortical brain regions showed a high protection mechanism through cerebral autoregulation, but in the cerebral cortex the protection mechanism failed and CBF decreased with hypoxia. ⁷⁸

Mild and severe hypoxia induced by reducing S_aO₂ resulted in a reduction in cellular oxygen availability. This resulted in increase of reCCO and decrease of oxCCO. In other words, when the inhaled oxygen was lower than 30%, CCO was mostly in its reduced state. This is consistent with previous studies measuring the difference in the oxidation state of CCO under hypoxia in healthy adults.^58,59

The association of CCO oxidation state with CBF and oxygenation

Both CBF and the oxidation state of CCO state were affected by the change in oxygen saturation. Overall, there was a significant positive correlation between CBF and S_aO₂, and between the oxidation state of CCO and S_aO₂. When oxygenation was the varying factor, a significant association was observed between CBF and the oxidation state of CCO. This positive relationship is an indication of normal coupling between oxygen delivery and use. This contrasts with the case of hypercapnia where the oxidation state changed independently from CBF, indicating that the relationship between oxidation and CBF can be impacted by the local environment. The gas challenges imposed here showed that under varying oxygenation the oxidation state of CCO was directly affected by the oxygen levels, while during mild hypercapnia it was more likely to be affected by secondary metabolic effects. ⁵⁸

Conclusions

The multimodal NIRS-MRI technique was applied to healthy mice, to determine the correlation between CBF, oxygen metabolism, and the oxidation state of CCO in the cortex, when exposed to mild hypercapnia or varying oxygenation.

We induced changes in CBF through two mechanisms – hypercapnia and changes in inspired oxygen. During mild hypercapnia, there can be increased tissue oxygenation (S_tO₂) and oxidized CCO without an increase in CMRO₂. The hypercapnia associated increase in CCO oxidation was accompanied with decreased OEF, resulting in constant CMRO₂. Increased oxidation of CCO does not always result in increased CMRO₂. Oxygen availability in the tissue was not the only factor driving the change in oxidation state of CCO and its association with CBF.

When varying oxygen availability by changing inspired oxygen, there is a more tightly coupled relationship between S_tO₂, CCO redox state, and CBF. Thus, the association between CBF and the oxidation state of CCO was not fixed, as it depended on the type of perturbation.

This work shows how powerful the method of combining MRI with NIRS can be to measure metabolic and physiological changes in the brain. Monitoring CCO in vivo simultaneously with CBF and CMRO₂ lays the groundwork for investigations of the mechanisms underlying neurovascular coupling and will provide new insight into vascular regulation and mitochondrial damage in neurological diseases.

Supplemental material

Supplemental material for this article is available online.