Analytical Biochemistry 654 (2022) 114825

Contents lists available at ScienceDirect

Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio

Intermittent fasting-induced biomolecular modifications in rat tissues

detected by ATR-FTIR spectroscopy and machine learning algorithms
Taha Ceylani a, Hikmet Taner Teker b, Gizem Samgane c, Rafig Gurbanov d, e, *
a
Department of Food Quality Control and Analysis, Muş Alparslan University, Muş, Turkey
b
Department of Medical Biology, Ankara Medipol University, Ankara, Turkey
c
Department of Biotechnology, Bilecik Şeyh Edebali University, Bilecik, Turkey
d
Department of Bioengineering, Bilecik Şeyh Edebali University, 11230, Bilecik, Turkey
e
Central Research Laboratory, Bilecik Şeyh Edebali University, 11230, Bilecik, Turkey

A R T I C L E I N F O A B S T R A C T

Keywords: This study aimed to reveal the intermittent fasting-induced alterations in biomolecules of the liver, ileum, and
Intermittent fasting (IF) colon tissues of rats using Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA) algorithms
FTIR Spectroscopy developed on infrared spectrochemical data. LDA prediction accuracies were generally calculated in the range of
Machine learning
95–100%, while training and validation accuracies of SVM were in the range of 91–100% and 83–91%,
Rat
respectively. The quantitative measurements of spectral bands at the CH (lipids), Amide (proteins), and PO2
Metabolism
antisymmetric (nucleic acids) stretching regions were performed to monitor modulated metabolic processes. The
concentration of biomolecules and phosphorylation rate of proteins were found higher in studied tissues. The
altered conformations and low rates of carbonylation (oxidation) were also common in proteins. No significant
change was recorded for the length of fatty acid acyl chains (A2922/A2955 band area ratio) in the liver, whereas
the shortening of acyl chains was calculated as 23% and 27% in ileum and colon tissues, respectively. Enhanced
membrane dynamics (Bw2922/Bw2955 bandwidth ratio) were depicted in the liver (35% increase), while a decline
in dynamics was apparent in the ileum (36% decrease) and colon (31% decrease). The study revealed important
alterations in major biomolecules of studied tissues.

1. Introduction decrease in blood sugar levels. Accordingly, the blood insulin level di­
minishes, which may lead to the prevention of insulin resistance during
Intermittent fasting (IF) is one of the dietary approaches that in­ metabolic diseases [8–10].
volves the restriction of regular food intake. Preventive roles of IF Fatty liver disease is a very common condition in modern societies,
against metabolic illnesses, aging, as well as cardiovascular and neuro­ and it can be a very important risk factor for many diseases such as
logical diseases have recently been recognized by the scientific com­ diabetes, liver cirrhosis, cardiovascular diseases, and cancer [11]. It is
munity. IF diet is suggested as a prophylactic and therapeutic observed that the rate of fat in the liver begins to decline rapidly after
intervention for various chronic diseases [1]. Many studies have been metabolism enters the fat-burning mode during IF [12,13]. One of the
conducted focusing on fasting practices, revealing the beneficial effects most striking features of fasting may be its ability to suppress inflam­
of various fasting practices on both the metabolism and cognitive status mation, which is one of the most important risk factors for many dis­
of the person [2]. IF, lasting from 16 to 48 h, is seen as one of the most eases, from cancers to heart disease [14]. Recent shreds of evidence have
preferred fasting practices but there are also certain discussions on shown that fasting significantly reduces indicators of inflammation such
possible negative effects. However, with recent pieces of evidence IF as C-reactive protein and interleukin-6 in the circulatory system [15]. In
might be both safe and highly effective for health [3,4]. One of the most addition, IF provides significant protection against cardiovascular dis­
striking effects of IF might be healthy weight loss, but recent studies also eases by suppressing blood sugar, unhealthy cholesterol, and inflam­
find out many other beneficial effects at the cellular level [5–7]. Because mation [16,17]. Also, significant normalizations in blood pressure are
of a calorie restriction for a certain period, there is also a healthy observed in many patients who are included in the IF program due to

* Corresponding author. Department of Bioengineering, Bilecik Şeyh Edebali University, 11230, Bilecik, Turkey.
E-mail addresses: [email protected], [email protected] (R. Gurbanov).

https://doi.org/10.1016/j.ab.2022.114825
Received 19 February 2022; Received in revised form 17 July 2022; Accepted 18 July 2022
Available online 30 July 2022
0003-2697/© 2022 Elsevier Inc. All rights reserved.
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

Fig. 1. Intermittent fasting affects the whole biomolecular profile in the liver samples. a) LDA discrimination plot and b) baseline-corrected average spectra in full
(4000-650 cm− 1) spectral region. CLI (control rats), FLI (rats on intermittent fasting).

obesity and diabetes [18–20]. statistics, signal processing, experimental design, modeling, calibration,
Recent studies revealed that the most crucial effect of IF may be its optimization, pattern recognition, classification, artificial intelligence
role in the initiation and progression of the autophagy cascade. With the approaches, image processing, and information and system theory.
autophagy process, aged proteins, organelles, and also damaged DNA is Chemometric approaches (multivariate pattern recognition methods or
digested and degraded to be used as an energy source and replaced with machine learning approaches) are used for the mining and classification
newly generated ones [21]. The ability of IF to balance sugar and of data gathered by a variety of analytical procedures in fields such as
cholesterol metabolism and suppress inflammation is also reflected in analytical chemistry, biology, archaeology, as well as clinical and
the brain as well as in all other organs. Brain-derived neurotrophic factor forensic medicine [33].
(BDNF) begins to be secreted in the brain after fasting for a certain Studies on fasting practices show that the hunger period is as
period. As it is known, BDNF has very crucial functions in the regener­ important as the satiety condition for human metabolism. It seems that
ation of brain cells [22–24]. Another important impact of IF is its human metabolism can show the ability to turn the hunger period in its
possible role in cancer protection [25]. It has been shown that there is an favor. Many studies have mentioned the numerous effects of fasting on
increase in the level of certain so-called “anticancer proteins” in the regeneration. The regeneration that starts at the cellular level as a result
blood during fasting exceeding 4 weeks. These substances slow down the of autophagy is then reflected in all tissues and organs [21]. However,
reproduction of cancer cells and also increase the sensitivity of cancer the initiation of autophagy requires a certain period of starvation.
cells to chemotherapy [26]. Therefore, the time interval of daily food restriction in the fasting pro­
Fourier Transform Infrared (FTIR) Spectroscopy, which detects gram must be long enough to initiate the autophagy process. Although
molecule vibrations and generates molecular spectral bands in the mid- fasting for 16 h does not normally result in ketosis, it is sufficient to
infrared region, is a technique used to gather broad-spectrum data in a encourage many pathways associated with prolonged fasting such as
rapid, easy, and non-invasive manner in biological analyses [27–29]. autophagy [34]. Because of the crucial role of the liver and intestine in
The Attenuated Total Reflection (ATR) mode of FTIR spectroscopy is a the regulation of metabolism, digestion, also host-microbiome interac­
useful tool for studying biological samples [30–32]. Chemometrics is a tion; this study aimed to investigate the impact of 18-h IF administered
branch of chemistry that covers the computer-assisted analysis of for 5 weeks on the liver, ileum, and colon tissues of rats. For this pur­
chemical data, as well as statistics and mathematics. The appeal of these pose, two different supervised machine learning techniques (Support
chemometric approaches is due to their ability to provide flexible and Vector Machine/SVM and Linear Discriminant Analysis/LDA) were
adaptable solutions for obtaining rapid, accurate, precise, and trust­ applied to big data obtained from FTIR spectral measurements of tissues.
worthy findings in the analysis of complicated materials. Therefore, it The quantitative measurements of FTIR spectral bands specific to lipids,
encompasses a range of disciplines, including descriptive and inferential proteins, and nucleic acids and their ratio indices were also calculated

2
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

Table 1 2. Materials and methods

SVM classification for liver samples in full (4000-650 cm− 1) spectral region. CLI
(control rats), FLI (rats on intermittent fasting). SVM type: Classification (nu- 2.1. Animal studies
SVC). Method: Radial Basis Function. * Highlights the wrong classification.
Accuracy (%) The male Wistar rats (12-month-old) were used as a model organism
Training 91.66 Validation 87.50 in the study. IF was applied to the rats (n = 6) in the IF experimental
group for 35 days. While the rats in the IF group were always able to
Classification
drink water, their access to food was restricted for 18 h, and they were
Samples Class only allowed to feed for 6 h. The food access interval of the animals in
CLI1 1 1 CLI the IF group was determined to be between 09:00 a.m. and 03:00 p.m.
CLI2 2 2 CLI The control group (n = 6) was allowed access to water and food for 24 h.
CLI3 3 3 CLI
The animals were fed a standard rodent diet on an ad libitum basis [35].
CLI4 4 4 CLI
CLI5 5 5 CLI
The body weight of the animals, the feed, and the consumed water were
CLI6 6 6 CLI followed for 35 days. The blood glucose levels were also measured at the
CLI7 7 7 CLI end of the application. One day after the end of the application, the
CLI8 8 8 CLI animals in the control and IF groups were slightly stunned by treatment
CLI9 9 9 CLI
with ether and sacrificed. The extracted tissues were shocked on dry ice
CLI10 10 10 CLI
CLI11 11 11 CLI and left in the − 80 ◦ C deep freezer until the time to be studied. All
CLI12 12 12 CLI animals were housed under standard animal care conditions. The study
FLI1 13 13 FLI was carried out with the approval of the Ethics Committee (approval
FLI2 14 14 FLI number: 2021/05) from the Saki Yenilli Experimental Animal Produc­
FLI3 15 15 *CLI
FLI4 16 16 FLI
tion and Practice Laboratory.
FLI5 17 17 FLI
FLI6 18 18 FLI
2.2. Analysis of samples by Attenuated total reflectance Fourier
FLI7 19 19 FLI
FLI8 20 20 FLI Transform Infrared (ATR-FTIR) spectroscopy
FLI9 21 21 FLI
FLI10 22 22 FLI Liver, ileum, and colon tissues of all animals were compressed on the
FLI11 23 23 FLI Zn/Se crystal of the ATR unit (PerkinElmer) without any pretreatment
FLI12 24 24 FLI
and each tissue was examined twice with an ATR-FTIR spectrometer
(PerkinElmer) at a resolution of 4 cm− 1 and a scan number of 32. The
for the monitoring of modulated metabolic processes at the biomole­ spectra were obtained with the Spectrum One (PerkinElmer) software in
cular level. Both quantitative data analyses and machine learning ap­ the wavelength range of 4000–650 cm− 1 [27].
proaches effectively enabled rapid and accurate detection of IF-induced
biomolecular changes. 2.3. Prediction studies with different machine learning approach based on
big spectral data

Linear Discriminant Analysis (LDA), a machine learning approach,

was applied to differentiate the experimental groups from each other.

Fig. 2. The changes in the FTIR spectral band areas for liver samples. The area values of a) CH3 antisymmetric (2955 cm− 1), b) CH2 antisymmetric (2922 cm− 1), c)
Amide I (1653 cm− 1), d) Amide II (1545 cm− 1), e) PO2 antisymmetric (1239 cm− 1) bands. The indices for f) acyl chain length of fatty acids (A2922/A2955), g) protein
phosphorylation (A1239/A2955+A2922), and h) protein conformation (A1653/A1545). CLI (control rats), FLI (rats on intermittent fasting), A (Absorbance).

3
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

prediction and confusion matrices [39,40].

Support Vector Machine (SVM) is another very popular machine
learning approach. SVM classification method was accomplished via
The Unscrambler® X 10.3 (CAMO Software AS, Norway) multivariate
analysis (MVA) software. All spectra were pre-processed as explained
above and subsequently different sample categories were used to
generate a training set. Classification (nu-SVC) was chosen as SVM type
using a linear method as Kernel type. Nu value was set to 0.5, weights as
all 1.00. The 9 segments of cross-validation were used in the calculation
of training and cross-validation accuracies. Finally, the generated
training dataset was applied to all sample datasets to obtain an SVM
classification model.

2.4. Quantification studies of FTIR spectral bands

Spectral data analysis was performed using OPUS 5.5 (Bruker) soft­
ware. The average spectra of 2 replicates from each sample were base­
line corrected using the Rubberband correction method with 128
baseline points prior to the band quantification analyses. In detailed
band analyses, the bands with the highest absorbance values in different
spectral regions of the spectra were selected and the beginning and
ending frequencies of the bands were determined with precision. The
areas of bands specific to various biomolecules were analyzed by taking
the integral areas of the determined frequency ranges with the OPUS 5.5
(Bruker) software. In addition, a virtual line was drawn vertically from
the midpoint of the band baseline to the peak of the band, and the length
of the line was measured with the help of a virtual ruler. Then, by
marking the point where 0.75 times the length of the line coincides with
the line, a horizontal line was drawn along the band from this point and
bandwidth values were obtained.

2.5. Statistics

Fig. 3. The changes in the FTIR spectral bandwidths for liver samples. The Statistical evaluations and graph plots of the results were made using
bandwidth values of a) CH3 antisymmetric (2955 cm− 1), b) CH2 antisymmetric GraphPad Prism 6.01 (GraphPad, USA). The data were analyzed using
(2922 cm− 1), c) Bw2922/Bw2955 ratio index, and d) Amide I (1653 cm− 1). CLI an unpaired t-test, and the significance levels were stated as P ≤ 0.05 *,
(control rats), FLI (rats on intermittent fasting), Bw (Bandwidth).
P ≤ 0.01 **, P ≤ 0.001 ***, and P ≤ 0.0001 ****. Results are presented
as mean ± SEM (standard error of the mean).

Spectral data were used in pattern recognition analysis. To make the 3. Results
analyzes as independent as possible from the FTIR spectrometers, each
sample spectrum was preprocessed on The Unscrambler® X 10.3 (CAMO 3.1. Intermittent fasting affects the water and feed consumption of
Software AS, Norway) software with a baseline offset transformation in animals
the 4000-650 cm− 1 region. Spectra processed in this way were first
subjected to unsupervised Principal Component Analysis (PCA) The body weights of the rats in the IF group did not change signifi­
[36–38]. Spectra were passed from standard deviation normalization cantly (P = 0.795). However, there was a significant increase in the body
(mean centering normalization) and full-cross random validation. Sub­ weights of the rats in the control group (P ≤ 0.001). The difference
sequently, the spectra were examined in lipid (3000-2700 cm− 1), pro­ between the control and IF groups was also significant (P ≤ 0.0001)
tein (1700-1500 cm− 1), nucleic acid (1200-650 cm− 1), and full (Fig. S1a). There was also a significant difference in food (P ≤ 0.0001)
(4000-650 cm− 1) regions by Singular Value Decomposition (SVD) and water consumption (P ≤ 0.0001). In the days after the application
algorithm. started, the rats in the IF group tended to eat and drink more as time
LDA is a supervised classifier in which n-dimensional feature samples went on (Figs. S1b–c). At the end of the application, the blood glucose
are linearly transformed into an m-dimensional space. PCA only uses levels of rats were measured. While the glucose values recorded in the
sample spectra to determine the transformation, while LDA also uses control group were seen as more stable, there was no significant change
class information in training samples leading to better classification. between the two groups (P = 0.25) (Fig. S1d).
PCA data were used as LDA model inputs with The Unscrambler® X 10.3
(CAMO Software AS, Norway) multivariate analysis (MVA) software.
3.2. Intermittent fasting affects lipid, nucleic acid, and protein profiles in
The category variable column was included in a data matrix and then all
the liver
spectra of different sample categories were used to generate a training
set. The linear method using the projections of the 9 PCA components
When the spectra of samples taken from the liver tissue at the end of
was used for the prediction. Prior probabilities were calculated from the
the application were examined by the LDA method, there was a signif­
training set. The results are presented as a discrimination plot, as well as
icant differentiation in terms of the whole biomolecular profile between

4
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

Fig. 4. Intermittent fasting affects the whole biomolecular profile in the ileum samples. a) LDA discrimination plot and b) baseline-corrected average spectra in full
(4000-650 cm− 1) spectral region. CIL (control rats), FIL (rats on intermittent fasting).

the control and IF groups, with a high accuracy rate (95.83% for whole functional group of the relevant molecule is directly proportional to the
biomolecular profile) (Fig. 1a, Tables S1–S2). As seen in the LDA concentration of that molecule. Therefore, we used areas under the
discrimination plot, the data for the control and fasting groups are bands to obtain relative concentration information about biomolecules
clustered in completely different places (Fig. 1a). This indicates the [41]. To monitor the possible origin of IF-associated differentiation in
absolute effect of IF on liver tissue. We observed the same effect when the liver, the changes in FTIR spectral bands related to different func­
we examined the lipid profiles of the liver (100% for lipids) (Fig. S2, tional groups in biomolecules were calculated. Accordingly, the spectral
Tables S3–S4). As shown in Fig. S2, the data obtained from the control parameters such as band areas, band area ratios, and bandwidths were
and IF groups in terms of lipids are located in completely different analyzed for the elucidation of molecular modifications in lipids, pro­
places. This IF-induced differentiation was similarly realized in nucleic teins, and nucleic acids. The analyses mainly covered the bands at 2955
acid and protein profiles of the liver with the highest accuracies (100% cm− 1 (CH3 antisymmetric stretching: lipids and proteins), 2922 cm− 1
for nucleic acids, 100% for proteins) (Figs. S3–S4, Tables S5–S8). When (CH2 antisymmetric stretching: lipids), 1653 cm− 1 (Amide I: α-helical
Fig. S3 and Fig. S4 are examined, it is seen that the data obtained for structure of proteins), 1545 cm− 1 (Amide II: β-sheet structure of pro­
both nucleic acid and protein profiles are located in completely different teins), and 1239 cm− 1 (PO2 antisymmetric stretching: nucleic acids)
clusters, and IF has a significant effect on these parameters. A compa­ positions. The band areas were significantly increased (76% in 2955
rable classification was obtained with the SVM method with 91.66% cm− 1, 88% in 2922 cm− 1, 14% in 1653 cm− 1, 69% in 1545 cm− 1, and
training and 87.50% cross-validation accuracies, for the whole bio­ 93% in 1239 cm− 1) for liver tissues of IF group (Fig. 2a–e).
molecular profile of liver tissues (Table 1). The acyl chain length of fatty acids can be calculated by band area
Average spectra (full infrared region/4000-650 cm− 1) for liver ratio A2922/A2955. However, no significant changes were recorded in the
samples of control and IF groups demonstrate visible changes in many liver of rats subjected to IF (Fig. 2f). In the literature, the band area ratio
spectrochemical bands, each associated with specific functional groups A1239/A2955+A2922 is usually referred to as the protein phosphorylation
of biomolecules (Fig. 1b). According to Beer-Lambert law, the intensity index. This index was found to be increased by 50% in the liver of rats
and/or the area of the infrared absorption bands rising from a particular subjected to IF, indicating high concentrations of phosphorylated

5
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

Table 2 proteins (Fig. 2g). The reduction (67%) in area ratio A1653/A1545, an
SVM classification for ileum samples in full (4000-650 cm− 1) spectral region. CIL indicator of protein conformation, was also seen in the liver of the IF
(control rats), FIL (rats on intermittent fasting). SVM type: Classification (nu- group, pointing out conformational alterations in liver proteins
SVC). Method: Linear. * Highlights the wrong classification. (Fig. 2h).
Accuracy (%) The bandwidth of CH2 antisymmetric stretching bands gives infor­
Training 100 Validation 83.33 mation about membrane dynamics since it is related to the motion rates
of the lipid molecule [42]. The increase in the bandwidths of these bands
Classification
implies an increase in membrane dynamics. It was also reported that the
Samples Class disruption of membrane organization is directly associated with the
CIL1 1 1 CIL decrease in membrane fluidity [43]. Therefore, the changes in mem­
CIL2 2 2 CIL brane dynamics were assessed through the calculation of bandwidths of
CIL3 3 3 CIL
2955 cm− 1 and 2922 cm− 1 bands, and the most reliable Bw2922/Bw2955
CIL4 4 4 CIL
CIL5 5 5 CIL
ratio index. It was estimated that IF elevates the membrane dynamics
CIL6 6 6 CIL (35% in Bw2922/Bw2955 bandwidth ratio) in liver tissues (Fig. 3a–c).
CIL7 7 7 CIL Broadening of Amide I band i.e., an increase in bandwidth of band at
CIL8 8 8 CIL 1653 cm− 1 is an indicator of protein carbonylation. The formation of
CIL9 9 9 CIL
additional carbonyls on some amino acid residues results from protein
CIL10 10 10 *FIL
CIL11 11 11 CIL oxidation [42]. IF lowered the protein carbonylation (oxidation) index
CIL12 12 12 CIL in liver tissues by 9% (Fig. 3d).
FIL1 13 13 FIL
FIL2 14 14 FIL
FIL3 15 15 FIL 3.3. Intermittent fasting affects lipid, nucleic acid, and protein profiles in
FIL4 16 16 FIL
the ileum
FIL5 17 17 FIL
FIL6 18 18 FIL
FIL7 19 19 FIL The effect of IF was seen when the spectra of samples obtained from
FIL8 20 20 FIL the intestinal tissue’s ileum region were examined using the LDA
FIL9 21 21 FIL method (Fig. 4a, Tables S9–S10). The differentiation of groups in terms
FIL10 22 22 FIL
FIL11 23 23 FIL
of the whole biomolecular profile was calculated with a higher degree of
FIL12 24 24 FIL accuracy than that of liver tissue, and the data of control and IF groups
are clustered in completely different places (100% for the whole bio­
molecular profile). This difference in accuracy rate was also reflected in
the distribution, and the effect of IF on the ileum occurred with greater
clarity. When the data distribution was examined in terms of lipid,

Fig. 5. The changes in the FTIR spectral band areas for ileum samples. The area values of a) CH3 antisymmetric (2955 cm− 1), b) Amide II (1545 cm− 1), c) PO2
antisymmetric (1239 cm− 1) bands. The indices for d) acyl chain length of fatty acids (A2922/A2955), e) protein phosphorylation (A1239/A2955+A2922), and f) protein
conformation (A1653/A1545), CIL (control rats), FIL (rats on intermittent fasting), A (Absorbance).

6
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

ileum (Fig. 6a–c). As in the liver, the protein carbonylation (oxidation)

index which is the bandwidth of 1653 cm− 1 was found to be decreased
by 14% in the ileum (Fig. 6d). Although many biomolecular changes in
the ileum and liver of rats are similar, it is obvious that the membrane
fluidity parameters (membrane dynamics values) change in the opposite
direction.

3.4. Intermittent fasting affects lipid, nucleic acid, and protein profiles in
the colon

The effect of IF was also significant on colon tissues, as demonstrated

by LDA. The effect of IF on the whole biomolecular profile of the colon
was prominent with high accuracy (95.83%) and adequate separation
rates (Fig. 7a, Tables S17–S18). When the impact of IF was evaluated for
lipid and nucleic acid profiles of the colon; the differentiation occurred
with the highest accuracy rates (100%) were similar to the lipid and
nucleic acid profiles of the liver and ileum (Figs. S8–9, Tables S19–S22).
In terms of the protein profile, the differentiation was detected with a
lower accuracy rate (75%) (Fig. S10, Tables S23–S24). When all the
tissues are evaluated; the effect of IF was most evident on lipid profiles,
while its effect on protein profiles followed a more interesting and
unique course. The SVM method revealed 95.83% training and 91.66%
cross-validation accuracies for the whole biomolecular profile of the
colon tissues (Table 3).
The full region infrared spectra and spectral band parameters for
colon biomolecules are given in Figs. 7b and Figs. 8–9, respectively.
Similar to other tissues, the increased concentrations of lipids (29% in
CH3 antisymmetric at 2955 cm− 1), proteins (13% in Amide II at 1545
cm− 1), and nucleic acids (19% in PO2 antisymmetric at 1239 cm− 1) were
calculated in colon tissues (Fig. 8a–c). The shorter acyl chain length of
fatty acids was also measured for the colon (27% diminish in A2922/
A2955 band area ratio index), as in the ileum (Fig. 8d). The changes in
Fig. 6. The changes in the FTIR spectral bandwidths for ileum samples. The other important indices such as protein phosphorylation (34% increase
bandwidth values of a) CH3 antisymmetric (2955 cm− 1), b) CH2 antisymmetric in A1239/A2955+A2922 ratio) and protein conformation (8% decrease in
(2922 cm− 1), c) Bw2922/Bw2955 ratio index, and d) Amide I (1653 cm− 1). CIL A1653/A1545 ratio) were also depicted (Fig. 8e–f). While the membrane
(control rats), FIL (rats on intermittent fasting), Bw (Bandwidth).
dynamics parameter (Bw2922/Bw2955 bandwidth ratio index) of the
colon decreases by 31% similar to the ileum (36%), it shows an opposite
change with the liver (35%) (Fig. 9a–c). The 14% diminish in protein
nucleic acid, and protein profiles with the same highest accuracies carbonylation (oxidation) index in colon samples was comparable with
(100% for all these biomolecules), the deepest differentiation occurred other tissues (Fig. 9d).
for lipids, then for nucleic acids and proteins, respectively (Figs. S5–S7,
Tables S11–S16). Similar to liver samples, there was a sharper differ­ 4. Discussion
entiation for lipids, while a similar differentiation for nucleic acids and
proteins was depicted in ileum samples. In the case of the SVM method Biological stress like IF improves metabolic health and extends life­
conducted for the whole biomolecular profile of ileum tissues, the ac­ span by activating stress resistance pathways, triggering weight loss, and
curacies for training and cross-validation were revealed as 100% and reducing blood pressure, LDL, cholesterol, and triglyceride levels. It also
83.33%, respectively (Table 2). systematically reduces fasting insulin levels, insulin resistance, and
Infrared spectra of ileum samples show the main differences between inflammation, further decreasing oxidative stress at cellular levels [44].
control and IF groups (Fig. 4b). However, the given full infrared region Not only animal models, but also dozens of human trials are demon­
(4000-650 cm− 1) does not visualize all the subtle changes. According to strated reductions in body weight and metabolic disease risk parame­
quantitative band analyses; IF led to significant increments in marker ters. The beneficial effect is observed at a cellular level initiating from
bands specific to lipids (22% in CH3 antisymmetric at 2955 cm− 1), gene transcription to protein translation and post-transcriptional mod­
proteins (10% in Amide II at 1545 cm− 1), and nucleic acids (8% in PO2 ifications [44–46]. The liver is a key player in providing metabolic ho­
antisymmetric at 1239 cm− 1) (Fig. 5a–c). In addition, IF caused a 23% meostasis for all other organs. IF induced serious increments in lipid,
decline in A2922/A2955 band area ratio index demonstrating shorter acyl protein, and nucleic acid biomolecules of the liver, indicating crucial
chains in fatty acids (Fig. 5d). Significant modifications were also changes in hepatocellular metabolism. As it is known, autophagy begins
measured in the band area indices for protein phosphorylation (10% at the cellular level when the subjects do not consume calories for at
increase in A1239/A2955+A2922) and protein conformation (7% decrease least 16 h. Autophagy enables the lysosomes to digest the old proteins,
in A1653/A1545) (Fig. 5e–f). organelles, and damaged DNA in certain periods, and make new mole­
In contrast to the increase in liver tissue, the Bw2922/Bw2955 band­ cules while using digested ones. Prolonged fasting elevated the endog­
width ratio index for membrane dynamics was diminished (36%) in the enous synthesis of molecules with antioxidant properties that function to

7
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

Fig. 7. Intermittent fasting affects the whole biomolecular profile in the colon samples. a) LDA discrimination plot and b) baseline-corrected average spectra in full
(4000-650 cm− 1) spectral region. CC (control rats), FC (rats on intermittent fasting).

reduce free radicals and thus improve metabolism [47]. Several human is recognized that fasting has a crucial role in intestinal health and
studies also revealed positive effects of IF on concentrations of various disease through endogenous metabolites as well as hormones and
lipid molecules circulating in blood [48]. A recent meta-analysis and growth factors [53].
systematic review study illustrated the potential of the IF diet in According to the findings of this study; the IF diet initiates serious
reducing lipid oxidative stress parameters and increasing total antioxi­ and mostly beneficial molecular modifications in the liver, ileum, and
dant capacities in humans [49]. Numerous animal and human studies colon tissues of rats. Moreover, some unique biological consequences of
have indicated an oxidative stress-reducing role of IF interventions in IF were revealed depending on tissue type. Both supervised machine
various metabolic disorders [50]. An elaborate metabolic response to learning methods (LDA and SVM) were effective in the prediction/
fasting is orchestrated by the liver and is largely dependent on tran­ classification of datasets with high accuracies. The prediction accuracy
scriptional regulation. Many transcription factors are activated in range of LDA was generally calculated to be 95–100%, whereas this
response to glucagon and glucocorticoid hormones and regulate various accuracy was 75% for colon samples in the protein (1700-1500 cm− 1)
genes involved in gluconeogenesis, fatty acid oxidation, ketogenesis, spectral region. The training and validation accuracies of SVM were
and amino acid metabolism [51]. Targeted metabolomics analysis of found in the range of 91–100% and 83–91%, respectively. The effect of
liver metabolites in mice subjected to IF revealed a >1.5-fold increase in IF was broad on specific biomolecular parameters. Particularly, the
12 molecules (PEP, citrate, succinate, NAD, NADH, NADPH, ATP, etc.) concentrations of lipids, proteins, and nucleic acids, as well as the
involved in glycolysis, tricarboxylic acid cycle, and oxidative phos­ phosphorylation rate of proteins were found higher in studied tissues of
phorylation which are central biochemical pathways in mammalian rats subjected to IF. The altered conformations and low rates of
energy metabolism [52]. The intestinal epithelium is a rapidly renewing carbonylation (oxidation) were also common in proteins. No significant
barrier of the body and the first-line sensor for dietary nutrients. The change was recorded for the length of fatty acid acyl chains (A2922/A2955
constant intestinal epithelial adaptation is required in response to band area ratio index) in liver tissues, whereas the shortening of acyl
stresses such as fasting stimuli alongside the changing nutrient stimuli. It chains was calculated as 23% and 27% in ileum and colon tissues,

8
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

Table 3 respectively. Another alteration was calculated for membrane fluidity in

SVM classification for colon samples in full (4000-650 cm− 1) spectral region. CC other words membrane dynamics considering the Bw2922/Bw2955
(control rats), FC (rats on intermittent fasting). SVM type: Classification (nu- bandwidth ratio index. Enhanced membrane dynamics were depicted in
SVC). Method: Linear. the liver (35% increase), while a decline in dynamics was apparent in
Accuracy (%) the ileum (36% decrease) and colon (31% decrease) tissues. The study
Training 95.83 Validation 91.66 demonstrated biomolecular findings on IF-induced modulations in he­
patic and intestinal tissues of rats, which may also be helpful for fasting-
Classification
related human studies and clinical trials.
Samples Class

CC1 1 1 CC Declaration of interest statement

CC2 2 2 CC
CC3 3 3 CC
No potential conflict of interest was reported by the authors.
CC4 4 4 CC
CC5 5 5 CC
CC6 6 6 CC CRediT authorship contribution statement
CC7 7 7 CC
CC8 8 8 CC Taha Ceylani: Conceptualization, Data curation, Formal analysis,
CC9 9 9 CC
Funding acquisition, Investigation, Methodology, Resources, Software,
CC10 10 10 CC
CC11 11 11 CC Supervision, Validation, Visualization, Writing – original draft, Writing
CC12 12 12 CC – review & editing. Hikmet Taner Teker: Conceptualization, Data
FC1 13 13 FC curation, Formal analysis, Funding acquisition, Investigation, Method­
FC2 14 14 FC ology, Resources, Software, Supervision, Validation, Visualization,
FC3 15 15 FC
FC4 16 16 FC
Writing – original draft, Writing – review & editing. Gizem Samgane:
FC5 17 17 FC Investigation, Methodology, Resources, Software, Supervision, Valida­
FC6 18 18 FC tion, Visualization, Roles/Writing. Rafig Gurbanov: Investigation,
FC7 19 19 FC Methodology, Resources, Software, Supervision, Validation, Writing –
FC8 20 20 FC
review & editing, Visualization, Roles/Writing.
FC9 21 21 FC
FC10 22 22 FC
FC11 23 23 FC Data availability
FC12 24 24 FC
Data will be made available on request.

Fig. 8. The changes in the FTIR spectral band areas for colon samples. The area values of a) CH3 antisymmetric (2955 cm− 1), b) Amide II (1545 cm− 1), c) PO2
antisymmetric (1239 cm− 1) bands. The indices for d) acyl chain length of fatty acids (A2922/A2955), e) protein phosphorylation (A1239/A2955+A2922), and f) protein
conformation (A1653/A1545), CC (control rats), FC (rats on intermittent fasting). A (Absorbance).

9
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

[7] R. Freire, Scientific evidence of diets for weight loss: different macronutrient
composition, intermittent fasting, and popular diets, Nutrition 69 (2020), 110549,
https://doi.org/10.1016/j.nut.2019.07.001.
[8] R. Antoni, K.L. Johnston, A.L. Collins, M.D. Robertson, Effects of intermittent
fasting on glucose and lipid metabolism, Proc. Nutr. Soc. 76 (2017) 361–368,
https://doi.org/10.1017/S0029665116002986.
[9] L. Olansky, Strategies for management of intermittent fasting in patients with
diabetes, Cleve. Clin. J. Med. 84 (2017) 357–358, https://doi.org/10.3949/
ccjm.84a.16118.
[10] S. Furmli, R. Elmasry, M. Ramos, J. Fung, Therapeutic use of intermittent fasting
for people with type 2 diabetes as an alternative to insulin, BMJ Case Rep. (2018),
https://doi.org/10.1136/bcr-2017-221854, 2018.
[11] Y. Qian, J.-G. Fan, Obesity, fatty liver and liver cancer, Hepatobiliary Pancreat. Dis.
Int. 4 (2005) 173–177.
[12] H. Cai, Y.-L. Qin, Z.-Y. Shi, J.-H. Chen, M.-J. Zeng, W. Zhou, R.-Q. Chen, Z.-Y. Chen,
Effects of alternate-day fasting on body weight and dyslipidaemia in patients with
non-alcoholic fatty liver disease: a randomised controlled trial, BMC Gastroenterol.
19 (2019) 219, https://doi.org/10.1186/s12876-019-1132-8.
[13] S. Drinda, F. Grundler, T. Neumann, T. Lehmann, N. Steckhan, A. Michalsen,
F. Wilhelmi de Toledo, Effects of periodic fasting on fatty liver index-A prospective
observational study, Nutrients 11 (2019), https://doi.org/10.3390/nu11112601.
[14] D. Margină, A. Ungurianu, C. Purdel, D. Tsoukalas, E. Sarandi, M. Thanasoula,
F. Tekos, R. Mesnage, D. Kouretas, A. Tsatsakis, Chronic inflammation in the
context of everyday life: dietary changes as mitigating factors, Int. J. Environ. Res.
Publ. Health 17 (2020), https://doi.org/10.3390/ijerph17114135.
[15] F.B. Aksungar, A.E. Topkaya, M. Akyildiz, Interleukin-6, C-reactive protein and
biochemical parameters during prolonged intermittent fasting, Ann. Nutr. Metab.
51 (2007) 88–95, https://doi.org/10.1159/000100954.
[16] M. Allaf, H. Elghazaly, O.G. Mohamed, M.F.K. Fareen, S. Zaman, A.-M. Salmasi,
K. Tsilidis, A. Dehghan, Intermittent fasting for the prevention of cardiovascular
disease, Cochrane Database Syst. Rev. 1 (2021) CD013496, https://doi.org/
10.1002/14651858.CD013496.pub2.
[17] T.A. Dong, P.B. Sandesara, D.S. Dhindsa, A. Mehta, L.C. Arneson, A.L. Dollar, P.
R. Taub, L.S. Sperling, Intermittent fasting: a heart healthy dietary pattern? Am. J.
Med. 133 (2020) 901–907, https://doi.org/10.1016/j.amjmed.2020.03.030.
[18] V. Deretic, D.J. Klionsky, Autophagy and inflammation: a special review issue,
Autophagy 14 (2018) 179–180, https://doi.org/10.1080/
15548627.2017.1412229.
[19] F. Grundler, R. Mesnage, A. Michalsen, F. Wilhelmi de Toledo, Blood pressure
changes in 1610 subjects with and without antihypertensive medication during
long-term fasting, J. Am. Heart Assoc. 9 (2020), e018649, https://doi.org/
10.1161/JAHA.120.018649.
[20] N. Di Daniele, G. Marrone, M. Di Lauro, F. Di Daniele, D. Palazzetti, C. Guerriero,
Fig. 9. The changes in the FTIR spectral bandwidths for colon samples. The A. Noce, Effects of caloric restriction diet on arterial hypertension and endothelial
bandwidth values of a) CH3 antisymmetric (2955 cm− 1), b) CH2 antisymmetric dysfunction, Nutrients 13 (2021), https://doi.org/10.3390/nu13010274.
(2922 cm− 1), c) Bw2922/Bw2955 ratio index, and d) Amide I (1653 cm− 1). CC [21] M. Bagherniya, A.E. Butler, G.E. Barreto, A. Sahebkar, The effect of fasting or
calorie restriction on autophagy induction: a review of the literature, Ageing Res.
(control rats), FC (rats on intermittent fasting), Bw (Bandwidth).
Rev. 47 (2018) 183–197, https://doi.org/10.1016/j.arr.2018.08.004.
[22] B. Martin, M.P. Mattson, S. Maudsley, Caloric restriction and intermittent fasting:
Acknowledgments two potential diets for successful brain aging, Ageing Res. Rev. 5 (2006) 332–353,
https://doi.org/10.1016/j.arr.2006.04.002.
[23] D.A. Loeffler, Influence of normal aging on brain autophagy: a complex scenario,
The authors are grateful to the Department of Chemistry at Bilecik Front. Aging Neurosci. 11 (2019) 49, https://doi.org/10.3389/fnagi.2019.00049.
Şeyh Edebali University for providing an FTIR spectrometer. [24] S.-H. Baik, V. Rajeev, D.Y.-W. Fann, D.-G. Jo, T. V Arumugam, Intermittent fasting
increases adult hippocampal neurogenesis, Brain Behav 10 (2020), e01444,
https://doi.org/10.1002/brb3.1444.
Appendix A. Supplementary data [25] A. Nencioni, I. Caffa, S. Cortellino, V.D. Longo, Fasting and cancer: molecular
mechanisms and clinical application, Nat. Rev. Cancer 18 (2018) 707–719, https://
doi.org/10.1038/s41568-018-0061-0.
Supplementary data to this article can be found online at https://doi. [26] N.S. Rocha, L.F. Barbisan, M.L.C. de Oliveira, J.L.V. de Camargo, Effects of fasting
org/10.1016/j.ab.2022.114825. and intermittent fasting on rat hepatocarcinogenesis induced by
diethylnitrosamine, Teratog. Carcinog. Mutagen. 22 (2002) 129–138, https://doi.
org/10.1002/tcm.10005.
References
[27] R. Gurbanov, F. Yıldız, Molecular profile of oral probiotic bacteria to Be used with
functional foods, Journal of Food and Health Science 12 (2017) 117–131, https://
[1] J.H. Lee, N. Verma, N. Thakkar, C. Yeung, H.K. Sung, Intermittent fasting: doi.org/10.3153/jfhs17015.
physiological implications on outcomes in mice and men, Physiology 35 (2020) [28] B. Velmurugan, L. Devaraj Stephen, S. Karthikeyan, S. Binu Kumari, Biomolecular
185–195, https://doi.org/10.1152/physiol.00030.2019. changes in gills of Gambusia affinis studied using two dimensional correlation
[2] M.P. Mattson, K. Moehl, N. Ghena, M. Schmaedick, A. Cheng, Intermittent infrared spectroscopy coupled with chemometric analysis, J. Mol. Struct. 1262
metabolic switching, neuroplasticity and brain health, Nat. Rev. Neurosci. 19 (2022), 132965, https://doi.org/10.1016/j.molstruc.2022.132965.
(2018) 63–80, https://doi.org/10.1038/nrn.2017.156. [29] M.M. Kumar, S.B. Kumari, E. Kavitha, B. Velmurugan, S. Karthikeyan, Spectral
[3] A. Michalsen, C. Li, Fasting therapy for treating and preventing disease - current profile index changes as biomarker of toxicity in Catla catla (Hamilton, 1822)
state of evidence, Forschende Komplementärmed. 20 (2006) 444–453, https://doi. edible fish studied using FTIR and principle component analysis, SN Appl. Sci. 2
org/10.1159/000357765, 2013. (2020) 1233, https://doi.org/10.1007/s42452-020-3001-z.
[4] F. Wilhelmi de Toledo, F. Grundler, A. Bergouignan, S. Drinda, A. Michalsen, [30] S.G. Kazarian, K.L.A. Chan, Applications of ATR-FTIR spectroscopic imaging to
Safety, health improvement and well-being during a 4 to 21-day fasting period in biomedical samples, Biochim. Biophys. Acta Biomembr. 1758 (2006) 858–867,
an observational study including 1422 subjects, PLoS One 14 (2019), e0209353, https://doi.org/10.1016/j.bbamem.2006.02.011.
https://doi.org/10.1371/journal.pone.0209353. [31] C. Aksoy, F. Severcan, Infrared spectroscopy and imaging in stem cells and aging
[5] M.-C. Stockman, D. Thomas, J. Burke, C.M. Apovian, Intermittent fasting: is the research, Methods Mol. Biol. 2045 (2019) 201–215, https://doi.org/10.1007/
wait worth the weight? Curr Obes Rep 7 (2018) 172–185, https://doi.org/ 7651_2018_119.
10.1007/s13679-018-0308-9. [32] E. Kavitha, L. Devaraj Stephen, F.H. Brishti, S. Karthikeyan, Two-trace two-
[6] I. Templeman, J.T. Gonzalez, D. Thompson, J.A. Betts, The role of intermittent dimensional (2T2D) correlation infrared spectral analysis of Spirulina platensis and
fasting and meal timing in weight management and metabolic health, Proc. Nutr. its commercial food products coupled with chemometric analysis, J. Mol. Struct.
Soc. 79 (2020) 76–87, https://doi.org/10.1017/S0029665119000636. 1244 (2021), 130964, https://doi.org/10.1016/j.molstruc.2021.130964.
[33] E. Dinç, Kemometri çok değişkenli Kalibrasyon Yöntemleri, Hacettepe Üniversitesi
Eczacılık Fakültesi Dergisi. 27 (2007) 61–92.

10
T. Ceylani et al. Analytical Biochemistry 654 (2022) 114825

[34] A. Paoli, G. Tinsley, A. Bianco, T. Moro, The influence of meal frequency and [44] J. Martel, D.M. Ojcius, Y.F. Ko, P.Y. Ke, C.Y. Wu, H.H. Peng, J.D. Young, Hormetic
timing on health in humans: the role of fasting, Nutrients 11 (2019), https://doi. effects of phytochemicals on health and longevity, Trends Endocrinol. Metabol. 30
org/10.3390/nu11040719. (2019) 335–346, https://doi.org/10.1016/J.TEM.2019.04.001.
[35] M.P. Mattson, V.D. Longo, M. Harvie, Impact of intermittent fasting on health and [45] H.H. Pak, S.A. Haws, C.L. Green, M. Koller, M.T. Lavarias, N.E. Richardson, S.
disease processes, Ageing Res. Rev. 39 (2017), https://doi.org/10.1016/j. E. Yang, S.N. Dumas, M. Sonsalla, L. Bray, M. Johnson, S. Barnes, V. Darley-Usmar,
arr.2016.10.005. J. Zhang, C.-L.E. Yen, J.M. Denu, D.W. Lamming, Fasting drives the metabolic,
[36] R. Gurbanov, S. Tunçer, S. Mingu, F. Severcan, A.G. Gozen, Methylation,Sugar molecular and geroprotective effects of a calorie-restricted diet in mice, Nature
puckering and Z-form status of DNA from a heavy metal-acclimated freshwater Metabolism 3 (2021) 1327–1341, https://doi.org/10.1038/s42255-021-00466-9.
Gordonia sp, J. Photochem. Photobiol. B Biol. 198 (2019), 111580, https://doi. [46] K.A. Varady, S. Cienfuegos, M. Ezpeleta, K. Gabel, Clinical application of
org/10.1016/J.JPHOTOBIOL.2019.111580. intermittent fasting for weight loss: progress and future directions, Nat. Rev.
[37] R. Gurbanov, N.S. Ozek, S. Tunçer, F. Severcan, A.G. Gozen, Aspects of silver Endocrinol. 18 (2022) 309–321, https://doi.org/10.1038/s41574-022-00638-x.
tolerance in bacteria: i?nfrared spectral changes and epigenetic clues, J. Biophot. [47] F. Grundler, R. Mesnage, N. Goutzourelas, F. Tekos, S. Makri, M. Brack,
11 (2018), e201700252, https://doi.org/10.1002/jbio.201700252. D. Kouretas, F. Wilhelmi de Toledo, Interplay between oxidative damage, the redox
[38] R. Gurbanov, H. Karadağ, S. Karaçam, G. Samgane, Tapioca starch modulates status, and metabolic biomarkers during long-term fasting, Food Chem. Toxicol.
cellular events in oral probiotic Streptococcus salivarius strains, Probiotics 145 (2020), https://doi.org/10.1016/j.fct.2020.111701.
Antimicrob Proteins 13 (2021) 195–207, https://doi.org/10.1007/s12602-020- [48] H. Meng, L. Zhu, H. Kord-Varkaneh, H. O Santos, G.M. Tinsley, P. Fu, Effects of
09678-z. intermittent fasting and energy-restricted diets on lipid profile: a systematic review
[39] S. Tunçer, R. Gurbanov, A novel approach for the discrimination of culture medium and meta-analysis, Nutrition 77 (2020), https://doi.org/10.1016/j.
from Vascular Endothelial Growth Factor (VEGF) overexpressing colorectal cancer nut.2020.110801.
cells, Turk. J. Biochem. 45 (2020) 715–724, https://doi.org/10.1515/tjb-2020- [49] S.I. Sharsher, A.I. Ahmed, M. Metwally, A.H. Arisha, K.E.D. Ahmed, Intermittent
0058. fasting decreases oxidative stress parameters and increases total antioxidant
[40] R. Gurbanov, S. Tunçer, S. Mingu, F. Severcan, A.G. Gozen, Methylation,Sugar capacity, Biointerface Research in Applied Chemistry 12 (2022) 6763–6775,
puckering and Z-form status of DNA from a heavy metal-acclimated freshwater https://doi.org/10.33263/BRIAC125.67636775.
Gordonia sp, J. Photochem. Photobiol. B Biol. 198 (2019), 111580, https://doi. [50] F.R. de Azevedo, D. Ikeoka, B. Caramelli, Effects of intermittent fasting on
org/10.1016/J.JPHOTOBIOL.2019.111580. metabolism in men, Rev. Assoc. Méd. Bras. 59 (2013) 167–173, https://doi.org/
[41] A. Dogan, R. Gurbanov, M. Severcan, F. Severcan, CoronaVac (Sinovac) COVID-19 10.1016/j.ramb.2012.09.003.
vaccine-induced molecular changes in healthy human serum by infrared [51] I. Goldstein, G.L. Hager, Transcriptional and chromatin regulation during fasting –
spectroscopy coupled with chemometrics, Turkish Journal of Biology = Turk the Genomic era, Trends Endocrinol. Metabol. 26 (2015) 699–710, https://doi.
Biyoloji Dergisi. 45 (2021) 549–558, https://doi.org/10.3906/biy-2105-65. org/10.1016/J.TEM.2015.09.005.
[42] D. Yonar, L. Ocek, B.I. Tiftikcioglu, Y. Zorlu, F. Severcan, Relapsing-Remitting [52] J. Ma, Y. Cheng, Q. Su, W. Ai, L. Gong, Y. Wang, L. Li, Z. Ma, Q. Pan, Z. Qiao,
Multiple Sclerosis diagnosis from cerebrospinal fluids via Fourier transform K. Chen, Effects of intermittent fasting on liver physiology and metabolism in mice,
infrared spectroscopy coupled with multivariate analysis, Sci. Rep. 8 (2018) 1025, Exp. Ther. Med. 22 (2021), https://doi.org/10.3892/etm.2021.10382.
https://doi.org/10.1038/s41598-018-19303-3. [53] G. Calibasi-Kocal, O. Mashinchian, Y. Basbinar, E. Ellidokuz, C.-W. Cheng, Ö.
[43] R. Gurbanov, M. Bilgin, F. Severcan, Restoring effect of selenium on the molecular H. Yilmaz, Nutritional control of intestinal stem cells in homeostasis and
content, structure and fluidity of diabetic rat kidney brush border cell membrane, tumorigenesis, Trends Endocrinol. Metabol. 32 (2021) 20–35, https://doi.org/
Biochim. Biophys. Acta 1858 (2016) 845–854, https://doi.org/10.1016/j. 10.1016/j.tem.2020.11.003.
bbamem.2016.02.001.

11