Acta Histochemica xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Acta Histochemica
journal homepage: www.elsevier.com/locate/acthis

The MTT-formazan assay: Complementary technical approaches and in vivo

validation in Drosophila larvae
⁎
Raquel Pascua-Maestro , Miriam Corraliza-Gomez1, Sergio Diez-Hermano1,
Candido Perez-Segurado1, María D. Ganfornina, Diego Sanchez
Instituto de Biología y Genética Molecular, Departamento de Bioquímica y Biología Molecular y Fisiología, Universidad de Valladolid, CSIC, Valladolid, Spain

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

Keywords: The MTT assay was the first widely accepted method to assess cytotoxicity and cell viability. However, there is
MTT assay controversy on whether this indicator is a useful tool. In this work we intend to expand the interpretability of the
Light scattering MTT study by its combination with widely used cellular biology techniques. We propose complementary ap-
Tetrazolium salt proaches to the colorimetric assay, based on the use of measurements in three different settings: confocal mi-
Confocal microscopy
croscopy, multi-well plate assay and flow cytometry. Using confocal microscopy, we confirmed that MTT uptake
Multi-well plate
and reduction by cells is a time-dependent process, and that formazan accumulates in round-shaped organelles.
Flow cytometry
Fruitfly Quantitative measurements with a multi-well fluorimeter combined with nuclear staining result in a useful
method, yielding a ratio between formazan production and cell number that informs about the average cell
metabolic state. We also found that flow cytometry is a suitable technique to measure MTT reduction in large cell
populations. When assaying the effect of an oxidizing agent such as paraquat (PQ), this approach allows for the
distinction of subpopulations of cells with different reducing power. Finally, we prove that it is feasible to
monitor MTT reduction in an in vivo model, the Drosophila larvae, without affecting its survival rate. Formazan
accumulates exclusively in the larval fat body, confirming its lipid solubility. The methods explored in this work
expand the MTT potential as a useful tool to provide information of the physiological state of cells and or-
ganisms.

1. Introduction
The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bro-
mide (MTT) and other tetrazolium salts-based assays have become one
of the most popular techniques for the quantitative assessment of cell
proliferation, viability and cytotoxicity (Mosmann, 1983), due to being
a low-cost, fast and simple procedure. The typical MTT assay consists in
a colorimetric determination performed in microtiter plates, where
absorbance measurements are obtained at the end of the assay. Its
foremost assumption relies in tetrazolium as an indicator of the in-
tracellular reducing potential, which in turn informs of the overall cell
state and viability. Enzymatic reduction of tetrazolium by cytosolic
dehydrogenases and reducing agents results in the formation of violet-
blue, water insoluble formazan product that has been proposed to ac-
cumulate mainly in hydrophobic cytoplasmic domains (e.g. lipid dro-
plets) (Stockert et al., 2012; Diaz et al., 2007). Interference in formazan
formation by glycolysis inhibitors such as 3-bromopyruvate has been
recently shown (van Tonder et al., 2015), supporting the notion that the
main source of reducing power might come from the glycolytic NAD(P)
and not from mitochondria. However, both the subcellular localization
of MTT after its intake, and where the reduction to formazan actually
takes place, are still controversial (Surin et al., 2017).
The absorption of formazan in the visible region should apparently
correlate with the number of metabolic intact and alive cells in the
sample. In practice though, there seems to exist ambiguity in the in-
terpretation of routine multi-well plate MTT assays, as they may lack
specificity to reliably distinguish growth arrest from total loss of via-
bility (Mirzayans et al., 2017): changes in the average cell population
enzymatic activity (metabolic deceleration)might be confused with
changes in the overall cell number. Recent efforts have been oriented to
potentiate the MTT assay informative power, especially by combining it
with microfluidic technology and image cytometry (Lim et al., 2015;
Halter, 2012). It is also noteworthy the almost non-existent studies
regarding MTT assay applications in vivo (James and Davey, 2007),
even though the pharmacological effects of formazan compounds have
been largely known (Shawali and Samy, 2015).

⁎
Corresponding author at: Instituto de Biología y Genética Molecular, c/Sanz y Forés 3, Universidad de Valladolid-CSIC, 47003 Valladolid, Spain.
E-mail address: raquel.pascua@ibgm.uva.es (R. Pascua-Maestro).
1
MCG, SDH and CPS contributed equally to this work and are listed in alphabetical order.

https://doi.org/10.1016/j.acthis.2018.01.006
Received 2 January 2018; Received in revised form 16 January 2018; Accepted 16 January 2018
0065-1281/ © 2018 Elsevier GmbH. All rights reserved.

Please cite this article as: PASCUA-MAESTRO, R., Acta Histochemica (2018), https://doi.org/10.1016/j.acthis.2018.01.006
R. Pascua-Maestro et al.
Until a few years ago, there were no procedures to study the re-
duction of MTT without preventing cell disintegration, because for-
mazan precipitates, forming deposits that were claimed to be non-
fluorescent (Ladyman et al., 2016). However, the use of scattered light
(SL) has allowed the development of new detection methods of MTT-
formazan (Bernas and Dobrucki, 2004).
In the present work, we study the time-course of MTT-formazan
intracellular accumulation and localization using confocal microscopy.
We also combine quantitative formazan detection in a multi-well
fluorimeter with nuclear labeling to measure the reducing power per
cell. By detecting formazan formation in a flow cytometer, we are able
to resolve the metabolic state of different subpopulations of cells subject
to oxidative stress. Finally, we show the application of the MTT assay in
an in vivo model of Drosophila melanogaster larvae, where we describe
the differential distribution of the formazan product throughout the
larval internal organs and test whether organism viability is compro-
mised.
2. Material and methods
2.1. Cell culture and treatments
HeLa cells (ATCC) were grown at 37 °C in humidity-saturated at-
mosphere containing 5% CO2. Culture medium [Dulbecco-modified
Eagle’s medium (DMEM; Lonza), supplemented with heat-inactivated
5% fetal bovine serum (FBS), 1% L-glutamine (final concentration
2 mM), and 1% P-S-A stock (final concentration: 100 U/ml penicillin,
100 U/ml streptomycin, 0.25 μg/ml amphoterycin B)] was replaced
twice a week, and cells were subcultured at 50% confluence.
Cells treated with Paraquat (PQ; 500 μM; 1–24 h) were cultured in
phenol red-free DMEM supplemented with 1% L-glutamine, 1% P-S, and
0.2% charcoal stripped FBS.
Cells treated with MTT (62.5 μg/ml; 0–3 h) were cultured in phenol
red-free DMEM supplemented with 1% L-glutamine, 1% P-S, and
without FBS. For confocal images, cells were maintained in DMEM
medium with HEPES (25 mM).
2.2. Confocal image acquisition and analysis
Confocal images were obtained with a 63× oil immersion objective
(HCX PL Apo CS NA = 1.4; Leica) attached to a DMI 6000B confocal
microscope with a TCS SP5 confocal system (Leica) equipped with
AOBS and AOTF systems. Cells were excited with a white laser and a
405-line controlled by LAS AF software (Leica). Emissions were col-
lected with the AOBS system and three spectral detectors.
Excitation and emission spectra of formazan were analyzed to select
suitable wavelengths for illumination and detection of scattered light.
Excitation spectra were collected at 470–670 nm, every 10 nm with a
50 nm detection window. Emission spectra wavelengths were collected
at 660–800 nm after excitation at 640 nm, with a 40 nm detection
window. The final wavelengths used for the detection of formazan in
our experiments were: λexc = 640 nm; λem = 670–800 nm.
We ensured to obtain images with similar dynamic range, and ad-
justed gain and offset using LUTs. Negative control images (MTT un-
treated cells) showed very weak and homogeneous background. We
obtained confocal sections under constant conditions (temperature and
CO2 concentration) to minimize image acquisition variation. Images
were stored as 1024 × 1024 pixels and 8-bit TIFF files.
Z-stacking was performed using the cell membrane to calculate the
number of optical sections. The optimal value of the step size was
calculated for the wavelength used to fulfill the Nyquist theorem. The
optical section thickness was 0.772 μm. Some images were zoomed in
4× to reduce field size. Voxel size is 0.06 * 0.06 * 0.3777 μm3. Scanning
was performed with a 1.0 Airy unit pinhole size.
After confocal imaging, using the same cells, we performed the
conventional MTT colorimetric assay in order to validate the results
2
Acta Histochemica xxx (xxxx) xxx–xxx
obtained.
Time lapse (3 h) experiments were performed to study the formazan
production. Images were collected every minute from the same field,
and processed and analyzed with the FIJI program. Thirty cells were
analyzed for each condition.
2.3. Normalized multi-well plate FSL detection method
For multi-well assays, cells were seeded to a final density of 1500
cells/mm2 in a 96 well-plate with black walls (NUNC). After removal of
medium and washes with PBS, cells were exposed to MTT as described
above.
Formazan was measured by illuminating the sample with a
λexc = 612 nm and collecting with λem = 670 nm in a GENios Pro
Fluorimeter (MTX Lab Systems) equipped with the XFluor4 GENiosPro
software package (TECAN). Cells were washed after measurement with
PBS, and cell nuclei were labeled by incubation with Hoechst 33342
(0.2 μg/ml; ImmunoChemistry Technologies, LLC) for 30 min under
agitation. Cells were washed again with PBS and Hoechst fluorescence
was measured (λexc = 340 nm, λem = 535 nm). The collected for-
mazan signal was normalized to the total number of cells dividing by
the Hoechst signal.
Finally, using the same cells, we performed the conventional MTT
colorimetric assay in order to validate the results obtained. Fig. 2A
shows a schematic representation of the complete procedure.
2.4. FSL-flow cytometry assay
Cells were cultured in 25 cm2 flasks up to 90% confluence. Cell
suspensions were obtained by trypsinization of adherent cultures after
different treatments. Cells were analyzed in a Gallios flow cytometer
(Beckman Coulter) and data were processed with Kaluza Analysis
software v.1.3 (Beckman Coulter). At least 20000 cells of each condi-
tion were analyzed. A red laser line (638 nm) was used for illumination.
Three detectors were tested: 710 DC SP/660 BP 20, 725 BP 20 and 755
LP. The same cells analyzed by flow cytometry were recovered from the
cytometer tube and used to carry out the MTT colorimetric assay.
2.5. Standard MTT colorimetric assay
To validate the results obtained by using the proposed com-
plementary techniques we carried out, with the same samples, the
conventional MTT colorimetric assay. This procedure ensures that the
signal obtained is due to formazan and allows for comparisons of sen-
sitivity and dynamic range of each technique.
Briefly, we treated cells with an isopropanol-10% Triton X-100 so-
lution, collected the solubilized formazan, and transferred triplicate
aliquots to a 96-well standard plate (NUNC). We measured formazan
production by spectrophotometry using the SOFTmax Pro microplate
reader (Molecular Devices) equipped with the SOFTmax Pro software
package (v.4.7). Absorbance was collected at λ = 570 nm, subtracting
the background measured at λ = 690 nm.
2.6. Fly lines and maintenance
Drosophila melanogaster flies were grown in a temperature-con-
trolled incubator at 25 °C, 60% relative humidity, under a 12 h light–-
dark cycle. They were fed with standard medium (wet yeast 84 g/l,
NaCl 3.3 g/l, agar 10 g/l, wheat flour 42 g/l, apple juice 167 ml/l, and
propionic acid 5 ml/l). For all the experiments, the Drosophila line yw
(Bloomington Drosophila Stock Center, BL1495) was used.
2.7. Larvae collection and selection
Fruit fly larvae from stages L1–L2 (24–48 h post hatching) were
isolated using a nylon grid with 0.5 mm pores. Larvae collected from
R. Pascua-Maestro et al.
culture vials were placed in the grid. They were washed with distilled
water to remove food remains. Larvae were selected by length
(2–3 mm) to ensure no larvae from later (non-feeding) developmental
stages were used.
2.8. In vivo MTT assay
Properly staged larvae were collected and placed in 35 mm Nunc TC
Dishes (10 larvae per dish) containing 3 ml of 10% sucrose, 0.5% agar
and 50 μg/ml MTT. Control plates contained the same recipe without
MTT. They were incubated at 25 °C with controlled humidity for 24 h.
Larvae were observed in a dissecting scope (Nikon SMZ1000 ste-
reomicroscope equipped with a Plan Apo 1 x WD70 objective). A violet-
blue product is formed inside the larvae when MTT is reduced to
Formazan. Larvae were transferred to a microcentrifuge tube and
frozen. The same number of larvae without MTT treatment was used as
control to measure background signal. Larvae were homogenized with a
pestle and resuspended in 300 μl of isopropanol-10% Triton-X100. The
suspension was mixed by vortexing. A short spin was used to precipitate
larval tissues before collecting the supernatant. To assess the possible
interferences of larval components, a standard curve with ascorbic acid
was performed, mixing 0.5 μg/μl MTT with 2 μg/μl ascorbic acid. After
24 h of incubation, supernatant was removed and the precipitated for-
mazan was solubilized in isopropanol-10% Triton-X100.
Spectrophotometric studies were performed using a Shimadzu
UV–1800 spectrophotometer.
2.9. Survival assay
Survival of MTT-fed larvae was tested after 24 h of MTT treatment.
Larvae with formazan signal were placed in vials (5–8 larvae per vial)
containing our standard medium (see Fly lines and maintenance) at
25 °C with controlled humidity. Pupae were scored every 24 h. The
same process was performed with untreated larvae as control.
2.10. Larvae dissection
Larvae with formazan signal were transferred to a drop of PBS on a
dissection plate. A dissection pin was placed between the posterior
spiracles and another in the larval head. Longitudinal opening of the
body wall facilitated the dissection of all internal organs. The process
was visualized and photographed in a Nikon SMZ1000 stereomicro-
scope equipped with a Plan Apo 1× WD70 objective and a Nikon DS-
Fi3 digital camera.
2.11. Statistical analysis
Statistical analyses were performed with SigmaPlot v.11.0 (Systat)
software. A p value < 0.05 was used as a threshold for significant
changes. The tests used for each experiment are stated in figure legends.
3. Results
3.1. Exploration of MTT assays with confocal microscopy, multi-well
fluorimetry, and flow cytometry
Analysis of the redox power of cells by the typical MTT assay pre-
sents the shortcoming of permanently altering the cell culture. It is
necessary to homogenize the sample and disrupt the cells to analyze
their content. An alternative method consists of measuring the for-
mazan production using confocal microscopy in live cells. With this
method, we are able to detect intracellular formazan deposits (Fig. 1A,
first panel) that correspond to those observed with DIC optics (Fig. 1A,
second and third panels). Direct observation avoids any accidental as-
sessment of deposits formed outside the cell. Higher magnification of
confocal images superimposed to DIC optics (Fig. 1B) reveals the
3
Acta Histochemica xxx (xxxx) xxx–xxx
granular appearance of formazan accumulations, although exact co-
localization with particular organelles cannot be assigned due to the
higher depth of focus of the non-confocal DIC image. The morphology
of granules helps to discard mitochondria as a candidate location of
subcellular reduction of MTT, since HeLa cells have long filamentous
mitochondria (Burman et al., 2017). To validate the method, we per-
formed the classical formazan absorbance assay in the same cell po-
pulations after confocal imaging (Fig. 1C).
Our experiments also reveal that formazan production is not im-
mediate in HeLa cells. Most cells need more than 50 min to start re-
ducing the internalized MTT to levels of formazan signal above the
detection threshold by confocal imaging (Fig. 1D, E). Our results show
that DIC optics has a lower threshold, and thus, signal is detected
sooner. However, the confocal approach can be particularly useful to
perform comparative analyses of individual cells, since it allows for
unambiguous measure of their reducing power with higher spatial re-
solution.
The multi-well fluorimetry assay is able to clearly differentiate cells
treated with MTT from the background signal of untreated cells. Our
approach, collecting formazan and Hoechst signals (Fig. 2A), allows for
normalization of the MTT-to-formazan conversion with the number of
cells (Fig. 2B), yielding a measure of the reducing power per cell. The
well-established colorimetric assay performed with the same samples
allowed us to ensure that the compound we are detecting was actually
formazan, resulting from the reduction of MTT by cells (Fig. 2C), as is
also the case for the confocal experiments reported above (Fig. 1C).
As a third methodology, we propose flow cytometry as a suitable
technique to detect cells producing formazan, as well as to perform
quantitative measurements of formazan signal in large cell populations.
First, we determined the ability of three different detectors to dis-
criminate between cells treated and untreated with MTT. The 755 LP
detector showed the best resolution in terms of separation of popula-
tions (Fig. 3A–F). Again, we confirmed by spectrophotometry that the
compound we were detecting was formazan (Fig. 3G). After estab-
lishing a gate that includes all cells in the sample (excluding debris), the
histograms representing the percentage of gated cells versus the for-
mazan signal intensity show two different peaks belonging to MTT-
positive and MTT-negative cells. Thus, we found that flow cytometry is
a suitable technique to discriminate cells according to its ability to re-
duce MTT into formazan.
3.2. Flow cytometry FSL assay detects subpopulations of cells with distinct
metabolic states upon oxidative insult
It is already known that oxidative stress-producing agents decrease
cell viability and, particularly, the ability of cells to reduce MTT (e.g.
Charão et al., 2015; Liu et al., 2018). The next issue we addressed was
whether our flow cytometry assay would help in the interpretation of
cell responses to oxidative insults. For this purpose, we established two
control groups (without PQ treatment), either not exposed (Fig. 4A) or
exposed to MTT for 3 h (Fig. 4B), that would be compared with two
experimental groups (Fig. 4C, D). HeLa cells were exposed to PQ for 1 h
after treatment with MTT (Fig. 4C) or for 24 h before MTT incubation
(Fig. 4D). Short PQ treatment after MTT treatment resulted in a slight
change in the shape of the formazan signal histogram and the density
plot, when compared to cells treated only with MTT (Fig. 4B). This shift
suggests that PQ, an oxidizing agent, is partially re-oxidizing the for-
mazan produced in the cells, rendering a sub-population with lower
formazan signal.
We exposed cells to PQ for 24 h before treatment with MTT, a
standard paradigm to test the effect on the physiological state after a
stress stimulus (Fig. 4D). Analysis by flow cytometry resulted in the
appearance of three different cell populations: one still able to reduce
MTT into formazan, an intermediate population, and a third one, si-
milar to the control cells untreated with MTT. This last population is
incapable of reducing the MTT, reflecting the oxidative stress induced
R. Pascua-Maestro et al. Acta Histochemica xxx (xxxx) xxx–xxx

Fig. 1. Confocal analyses of HeLa cells treated with MTT.

A: Formazan signal (red LUT) and DIC views of HeLa cells treated for 3 h with MTT. B: Z-series and z-projection of cells treated with MTT (4× zoom). C: MTT colorimetric assay
performed by absorbance collection at 570 nm and background subtraction (690 nm) in the same sample of cells after confocal microscopy. D: Integrated intensity of formazan-positive
cells obtained by time-lapse confocal microscopy (n = 30 cells) during the 3 h of MTT exposure (first 12 min were not recorded). E: Confocal time-lapse analysis of formazan production in
cells treated with MTT. Selected images were taken each 30 min. Calibration bars in B: 10 μm, and A, E: 40 μm.

4
R. Pascua-Maestro et al. Acta Histochemica xxx (xxxx) xxx–xxx

Fig. 2. In vitro MTT reduction to formazan

monitored by measuring scattered light in a
multi-well plate assay.
HeLa cells were treated with MTT for 3 h,
and the formazan was measured in a multi-
well fluorimeter (n = 18 wells/condition).
A: Schematic representation of the experi-
ment. B: Formazan production was esti-
mated by measuring light scattered when
cells are illuminated at λ = 612 nm and
collected at λ = 670 nm (FSL), and normal-
izing by cell number (Hoechst signal;
λexc = 340 nm, λem = 535 nm). C: MTT
colorimetric assay collecting absorbance at
570 nm and subtracting background (at
690 nm).

by long term PQ treatment. An overlay of histograms belonging to the
four experimental conditions (Fig. 4E) allows for comparisons of the
reducing power of cell populations in each condition. The formazan
signal values of the intermediate population in the long-term PQ ex-
periment are similar to those obtained with the PQ short-term treat-
ment post-MTT. These three distinct cell responses would be indis-
tinguishable when using the standard MTT colorimetric assay.
This observation is confirmed by the MTT conventional assay
(Fig. 4F) performed in the same samples, showing a statistically sig-
nificant decrease in the amount of formazan in the group of cells treated
with PQ for 24 h, compared to the untreated ones.
3.3. In vivo application of MTT assay in Drosophila larvae
Finally, we checked whether it was feasible to adapt the MTT col-
orimetric protocol to an in vivo experimental setup. Drosophila was
chosen as it ranks amongst the most versatile animal models in terms of
its wide applicability to genetic screening. Fruitfly larvae were
incubated in an agar matrix containing MTT. After 24 h, the presence of
formazan is easily observed in the larvae (Fig. 5A, B). After homo-
genization of larvae in isopropanol −10% Triton-X100, the spectro-
photometric analysis of solubilized formazan revealed an absorption
peak at 570 nm, as expected by comparison with the spectrum of MTT
after in vitro reduction by ascorbic acid (Fig. 5C). Other larval compo-
nents present in the homogenate do not affect the measurement.
The MTT-treated violet-colored larvae were followed up to test
whether formazan deposits might affect their survival (Fig. 5D). The
surviving larvae molt successfully into pupae, and healthy adults
emerge. Apparently, larvae eliminate all formazan deposits by its
normal digestive process, because they lose labeling over time. The
pupation ratio (over 80%) of the MTT treated larvae does not differ
from control ones (Fig. 5D). No significant difference in pupation
probability between control and MTT-treated larvae was found (sur-
vival analysis, p > 0.05, data not shown), indicating that formazan
deposits do not interfere with the larval life cycle.
An analysis of formazan tissue distribution revealed that deposits

Fig. 3. Evaluation of the ability of three different fluorescence detectors to discriminate between formazan-producer cells and negative control cells by flow cytometry.
Each panel shows two different cell populations: (+)MTT = cells treated with MTT for 3 h; (−)MTT = control cells not treated with MTT. The same populations were analyzed
simultaneously with different detectors. A–C: Density plots of “complexity” (SS: side scattering) vs. formazan light signal collected by each detector. D–F: Frequency histograms re-
presenting the percentage of cells vs. the formazan signal intensity. A gate selecting detected events with SS and FS (FS: forward scattering, size signal) characteristic of cells (discarding
debris) is used for all analyses. Detector 710 DC SP/660 BP 20 (A, D); Detector 725 BP 20 (B, E), Detector 755 LP (C, F). G: Conventional MTT colorimetric assay, collecting absorbance at
570 nm and subtracting background (at 690 nm).

5
R. Pascua-Maestro et al. Acta Histochemica xxx (xxxx) xxx–xxx

Fig. 4. Identification of different physiological states in PQ treated cells using FSL flow cytometry assay.
A–D: Formazan signal intensity histograms and density plots of cells untreated with MTT (A), treated with MTT for 3 h (positive control) (B), cells treated with MTT for 3 h and then
incubated in PQ for 1 h (C), and cells pre-incubated with PQ for 24 h and then exposed to MTT for 3 h (D). E: Overlap of the frequency histograms corresponding to each of the
experimental conditions. Plots were normalized by using the parameter “%gated”. F: Conventional MTT colorimetric assay, collecting absorbance at 570 nm and subtracting background
(at 690 nm).

are almost exclusively localized in the larval fat body (Fig. 5F, K). No
violet deposits are observed after the fat body is removed (Fig. 5G).
Neither the brain nor the digestive track showed signs of formazan
accumulations (Fig. 5H–J).
4. Discussion
Application of the classical MTT assay as an indicator of cell via-
bility has always been controversial (Funk et al., 2007). In the present
work, we aimed to expand the interpretability of the MTT study by
combination with other widely used techniques in cellular biology.
Some studies support that MTT measures the reducing potential of
mitochondria (Surin et al., 2017). Our data and previous studies
(Stockert et al., 2012) show that the staining of formazan over time
locates in vesicles different from the filamentous mitochondria of HeLa
cells, making mitochondria an unlikely candidate for MTT localization.
Although it is known that in stressed cells the mitochondrial network
undergoes fission and mitochondria round up, this factor could not
explain the observed structures in our confocal analysis, which was
performed in standard, non-stressed conditions. Fat body localization in
Drosophila larvae also support a non-mitochondrial location. In addi-
tion, our results support the hypothesis that MTT does not alter orga-
nelle morphology or distribution once it has been reduced to formazan.
In this work, we quantitate formazan production in live cells using
three different technical approaches: confocal microscopy, multi-well
fluorimetry and flow cytometry. The formazan optical signals obtained
with these techniques are assumed to be generated by light scattering,
according to Bernas and Dobrucki (2004), though the contribution of a
fluorescent signal might be considered, as recently proposed by
Stockert and Blázquez-Castro (2017). However, it is worth noticing
that, in the time-lapse analysis, the blue-colored formazan structures
start to increase, in both number and size, well before red scattering
reach the threshold for detection in our confocal-microscopy experi-
ments.
Formazan signal acquisition in a confocal microscope or with multi-
well fluorimeter allows us to relate formazan production with cell
number at the end of the assay, by either counting cells in the micro-
scopy field or using Hoechst labeling of nuclei. This feature is important

6
R. Pascua-Maestro et al. Acta Histochemica xxx (xxxx) xxx–xxx

Fig. 5. Formazan effect and distribution in Drosophila larvae fed with MTT.
A: Healthy larva fed with standard diet without MTT. B: Larva after 24 h treatment with MTT. MTT is reduced inside the larva forming formazan violet deposits with a heterogeneous
distribution. C: Absorption spectra of formazan extracted by homogenization of larvae in isopropanol-Triton X-100 after 24 h treatment (solid line), and MTT reduced in vitro by ascorbic
acid (dashed line). D: Formazan accumulation in MTT-fed Drosophila larvae does not compromise pupation. E: Schematic representation of Drosophila larval internal organs. F–K: Third-
instar Drosophila larvae after MTT treatment for 48 h reveals a distinct distribution of formazan deposits. A dissected larva fed with MTT is shown in F. The same larva is shown in G after
gut and fat body are removed, showing no formazan deposits in the body wall. Head parts (H), brain (I), and gut (arrow in J) show no signs of formazan deposits, while fat body
(arrowhead in J, and K) accumulate formazan deposits in the live larva. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this
article.)

to overcome one of the main concerns about conventional MTT assays:
whether the differences observed in formazan production between cell
samples are due to the experimental condition tested (Stepanenko and
Dmitrenko, 2015). Different cell number (increased by cell division or
decreased by cell death) or the same number of cells in different me-
tabolic states (with changing reducing power) could not be dis-
tinguished.
Here we show that formazan signal measurement by flow cytometry
is particularly useful to detect subpopulations of cells in different me-
tabolic states. Our experiments suggest that upon oxidative stress, the
ability of cells to reduce MTT is compromised, yielding a lower pro-
duction of formazan. These observations fit the previously reported
mild mitochondrial redox imbalance after MTT exposure (Stockert
et al., 2012), that might be exacerbated by PQ-driven oxidation.
However, it is worth to point out that the effects of oxidative stress on
cells are diverse, and that after a pro-oxidant stimulus there might be
different subpopulations with different redox states in the same cell
sample, and consequently different abilities to produce formazan. In
this paradigm, flow cytometry could be combined with cell sorting,
which would enable to separate cells according to its ability to reduce
MTT, and even to measure other additional parameters and specific cell
markers.
We have also described a Drosophila larvae in vivo model that is
suitable to the classical MTT assay protocol. The presence of MTT-for-
mazan deposits exclusively in the larvae fat body agrees with the pro-
posed affinity of the compound for lipid-enriched cellular environments
(Stockert et al., 2012; Diaz et al., 2007). This accumulation seems not to
interfere with the proper metabolic and energy-supply function of the
organ, as seen in the survival assay. Since no signs of formazan toxicity
affecting survival exist, the methods opens the possibility of assaying
effects of genetic modifications altering relevant biological processes.
This will undoubtedly allow us to expand the uses of the MTT technique
and gain further insight in its potential physiological applications.
Funding statement
This work was supported by grants to MDG and DS (Ministerio de
Ciencia e Innovación (MICINN) grant BFU2015-68149-R, co-financed
by European Regional Development Fund). MCG was supported by a
University of Valladolid fellowship to young researchers (call#2016).
RPM was supported by a Junta de Castilla y León (JCyL) fellowship to
young researchers (call#EDU/1883/2013), financed by the European
Social Fund, Operational Programme for Castilla y León and managed
by Consejería de Educación (JCyL). The funders had no role in study
design, data collection and analysis, decision to publish, or preparation
of the manuscript.
Acknowledgments
We thank A. Martín Muñoz and C. Sánchez-Vicente at the IBGM
Flow Cytometry and Confocal Microscopy Service for technical assis-
tance. J. Rojo-Ruiz provided technical advice on Drosophila larva
management and dissection. We thank Alfonso Blázquez-Castro and
Juan C. Stockert, for helpful discussions and inspiring ideas.
References
Bernas, T., Dobrucki, J.W., 2004. Backscattered light confocal imaging of intracellular
MTT-formazan crystals. Microsc. Res. Tech. 64 (2), 126–134.
Burman, J.L., et al., 2017. Mitochondrial fission facilitates the selective mitophagy of
protein aggregates. J. Cell Biol. 216 (10), 3231–3247.
Charão, M.F., et al., 2015. Protective effects of melatonin-loaded lipid-core nanocapsules
on paraquat-induced cytotoxicity and genotoxicity in a pulmonary cell line. Mutat.
Res./Genet. Toxicol. Environ. Mutagen. 784–785 (Suppl. C), 1–9.
Diaz, G., et al., 2007. Localization of MTT formazan in lipid droplets. An alternative
hypothesis about the nature of formazan granules and aggregates. Eur. J. Histochem.
51 (3), 213–218.
Funk, D., Schrenk, H.-H., Frei, E., 2007. Serum albumin leads to false-positive results in
the XTT and the MTT assay. Biotechniques 43 (2), 178–186.
Halter, M., 2012. Modernizing the MTT assay with microfluidic technology and image

7
R. Pascua-Maestro et al.
cytometry. Cytometry A 81 (8), 643–645.
James, C.E., Davey, M.W., 2007. A rapid colorimetric assay for the quantitation of the
viability of free-living larvae of nematodes in vitro. Parasitol. Res. 101 (4), 975–980.
Ladyman, M.K., et al., 2016. Fluorescent formazans and tetrazolium salts–towards
fluorescent cytotoxicity assays. Comb. Chem. High Throughput Screen. 19 (5),
384–391.
Lim, S.-W., et al., 2015. Reduction of MTT to purple formazan by vitamin E isomers in the
absence of cells. Trop. Life Sci. Res. 26 (1), 10.
Liu, Z., et al., 2018. Silymarin attenuated paraquat-induced cytotoxicity in macrophage
by regulating Trx/TXNIP complex, inhibiting NLRP3 inflammasome activation and
apoptosis. Toxicol. In Vitro 46 (Suppl. C), 265–272.
Mirzayans, R., Andrais, B., Murray, D., 2017. Do multiwell plate high throughput assays
measure loss of cell viability following exposure to genotoxic agents? Int. J. Mol. Sci.
18 (8), 1679.
Mosmann, T., 1983. Rapid colorimetric assay for cellular growth and survival: application
Acta Histochemica xxx (xxxx) xxx–xxx
to proliferation and cytotoxicity assays. J. Immunol. Methods 65 (1–2), 8.
Shawali, A.S., Samy, N.A., 2015. Functionalized formazans: a review on recent progress in
their pharmacological activities. J. Adv. Res. 6 (3), 241–254.
Stepanenko, A., Dmitrenko, V., 2015. Pitfalls of the MTT assay: direct and off-target ef-
fects of inhibitors can result in over/understimation of cell viability. Gene 574
(2), 10.
Stockert, J.C., Blázquez-Castro, A., 2017. Fluorescence Microscopy in Life Sciences. E-
book, Bentham Science Publishers, Sharjah. U.A.E.
Stockert, J.C., et al., 2012. MTT assay for cell viability: intracellular localization of the
formazan product is in lipid droplets. Acta Histochem. 114 (8), 785–796.
Surin, A.M., et al., 2017. Disruption of functional activity of mitochondria during MTT
assay of viability of cultured neurons. Biochemistry (Mosc.) 82 (6), 737–749.
van Tonder, A., Joubert, A.M., Cromarty, A.D., 2015. Limitations of the 3-(4,5-di-
methylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assay when com-
pared to three commonly used cell enumeration assays. BMC Res. Notes 8, 47.