Variability of fluorescence intensity distribution measured by flow cytometry is influenced by cell size and cell cycle progression

Abstract

The distribution of fluorescence signals measured with flow cytometry can be influenced by several factors, including qualitative and quantitative properties of the used fluorochromes, optical properties of the detection system, as well as the variability within the analyzed cell population itself. Most of the single cell samples prepared from in vitrocultures or clinical specimens contain a variable cell cycle component. Cell cycle, together with changes in the cell size, are two of the factors that alter the functional properties of analyzed cells and thus affect the interpretation of obtained results. Here, we describe the association between cell cycle status and cell size, and the variability in the distribution of fluorescence intensity as determined with flow cytometry, at population scale. We show that variability in the distribution of background and specific fluorescence signals is related to the cell cycle state of the selected population, with the 10% low fluorescence signal fraction enriched mainly in cells in their G0/G1 cell cycle phase, and the 10% high fraction containing cells mostly in the G2/M phase. Therefore we advise using caution and additional experimental validation when comparing populations defined by fractions at both ends of fluorescence signal distribution to avoid biases caused by the effect of cell cycle and cell size.

Full text

1
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports
Variability of uorescence intensity
distribution measured by ow
cytometry is inuenced by cell size
and cell cycle progression
Radek Fedr
1,2, Zuzana Kahounová
1, Ján Remšík
3, Michaela Reiterová
4, Tomáš Kalina
4 &
Karel Souček
1,2,5*
The distribution of uorescence signals measured with ow cytometry can be inuenced by several
factors, including qualitative and quantitative properties of the used uorochromes, optical properties
of the detection system, as well as the variability within the analyzed cell population itself. Most of
the single cell samples prepared from in vitrocultures or clinical specimens contain a variable cell cycle
component. Cell cycle, together with changes in the cell size, are two of the factors that alter the
functional properties of analyzed cells and thus aect the interpretation of obtained results. Here, we
describe the association between cell cycle status and cell size, and the variability in the distribution
of uorescence intensity as determined with ow cytometry, at population scale. We show that
variability in the distribution of background and specic uorescence signals is related to the cell
cycle state of the selected population, with the 10% low uorescence signal fraction enriched mainly
in cells in their G0/G1 cell cycle phase, and the 10% high fraction containing cells mostly in the G2/M
phase. Therefore we advise using caution and additional experimental validation when comparing
populations dened by fractions at both ends of uorescence signal distribution to avoid biases caused
by the eect of cell cycle and cell size.
Cell cycle is an essential biological process that signicantly contributes to the transcriptional heterogeneity in cell
dierentiation1, cell death2, and carcinogenesis3. Exploration of data obtained with single-cell RNA sequencing
(scRNA-seq) revealed that the cell cycle and cell volume can act as sources of bias, introducing within-cell-type
phenotypic and functional heterogeneity4–7. Unbiased cell clustering may therefore be obtained by correcting
for cell cycle eects7,8. Several strategies were developed to remove cell cycle eects from scRNA-seq (for review
see8) and mass cytometry data6. Besides scRNA-seq and mass cytometry, current state-of-the-art uorescence-
based ow cytometry allows measurement of 40+ colours simultaneously9,10, and represents a re-emerging
technology for large scale single-cell analysis11, with deeper understanding the cell cycle and cell volume eects
in polychromatic ow cytometry data now becoming more than necessary. Two major technical limitations of
ow cytometry are background uorescence, sometimes referred to as autouorescence, and spreading error,
which can contribute to the incorrectly identied heterogeneity within cell populations12. e background, native
uorescence is a normal characteristic of every particle, cell and tissue. Background uorescence is inuenced
by cellular phenotype13,14, metabolic state15,16, and proliferation rate17. A number of endogenous uorophores
have been described, including aromatic amino acids, cytokeratines, collagen and elastin, NAD(P)H, avins,
fatty acids, vitamin A derivatives, porphyrins and lipofuscin, and these can be exploited as intrinsic biomarkers18.
ese molecules are excited by and emit over a broad range of wavelengths and oen overlap the spectra of
commonly used uorescent probes19. is interesting phenomenon, together with the technological advance-
ments, opened a large eld of investigation and application of autouorescence in biological research14,20,21
OPEN
1Department of Cytokinetics, Institute of Biophysics of the Czech Academy of Sciences, Královopolská 135,
612 00 Brno, Czech Republic. 2International Clinical Research Center, St. Anne’s University Hospital Brno, Brno,
Czech Republic. 3Human Oncology and Pathogenesis Program, Memorial Sloan Kettering Cancer Center, New
York, NY 10065, USA. 4CLIP - Childhood Leukaemia Investigation Prague, Department of Pediatric Haematology
and Oncology, Second Faculty of Medicine, Charles University and University Hospital Motol, Prague,
Czech Republic. 5Department of Experimental Biology, Faculty of Science, Masaryk University, Brno, Czech
Republic. *email: ksoucek@ibp.cz
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
Vol:.(1234567890)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
and biomedical diagnosis22,23. On the other hand, it must be noted as an obstacle and a potential pitfall of
uorescence-based techniques12,24.
Here, we investigated the association between cell cycle status, cell size, and the variability of uorescence
intensity distribution as measured by ow cytometry. We demonstrated that the variability in the distribution
of both background and specic uorescence signal is related to the cell cycle state of the measured cell popula-
tion. Cells with low uorescence signal are enriched in smaller cells, mostly in G0/G1 phase, while the cells with
high uorescence signal are larger and in G2/M phase. We argue that the data interpretation from experiments
comparing the populations dened as “low” versus “high” in terms of symmetric selection of fractions at both
ends of uorescence signal distribution could be misleading. Investigators should take into account the eect of
cell cycle and cell size and corroborate such ndings with other techniques.
Results
Fluorescence background distribution is related to the cell cycle status in living and xed
cells. Dierences in the cell cycle stage of sorted cells can have profound eect on downstream analyses.
To systematically test whether the distribution of background signals, or autouorescence, of cells analyzed
with ow cytometry relates to their cell cycle status, we rst analyzed the cell cycle prole of lower and upper
10% of cells gated based on their background uorescence. We labelled two cell lines, HCT 116 (human colon
cancer) and cE2 (murine prostate cancer) with a series of commonly used DNA stains in both native (Hoechst
33342) and xed states (DAPI or propidium iodide). We recorded their uorescence at a single cell level using
ow cytometry across all detectors, including the empty, background channels. We then focused on these back-
ground channels and applied a back-gating strategy, separating the bottom 10% of the lower intensity population
and the top 10% of the higher intensity population in background uorescence channels (Fig.1). To control for
a possible uorescence spillover eect, the background uorescence was assessed on dierent optical line of the
BG FL 639//710/50
HighLow
101102103104105
Count
Count
Hoechst 33342
Count
Hoechst 33342
Low
High
Low
101102103104105
BG FL 488//586/42
Count
Count
DAPI
Count
DAPI
High
Propidium iodide Propidium iodide
BG FL 405//450/50
405 High
405 Low
101102103104105
Count
Count
Count
DNA content
Propidium iodide Propidium iodide
405 nm
488 nm
639nm
488 Low 488 High
639 Low 639 High
Count
Count
Count
Count
G0/G1=96
G2/M=0
G0/G1=75
G2/M=12
G0/G1=97
G2/M=0
G0/G1=29
G2/M=50
G0/G1=82
G2/M=4
G0/G1=50
G2/M=28
Figure1. Fluorescence background distribution is associated with cell cycle state in live and xed cells.
HCT 116 cells were stained using a series of DNA dyes, in xed (propidium iodide, DAPI) or native (Hoechst
33342) conditions. Background uorescence was analyzed using 405, 488, and 639nm lasers and an array of
detectors (425 up to 810nm) that were separated from the optical line for the particular DNA dye. Samples
were analyzed using ow cytometry (BD FACSAria II SORP). Dead cells were excluded based on LIVE/DEAD
staining, fractions of the cells with low (bottom 10%) and high (top 10%) background uorescence were gated
for DNA content (cell cycle) analysis. e values of G0/G1 and G2/M phase represent the proportion of cells.
Data are representative from at least three independent repetitions. For total DNA content distribution of the
entire population, see Supplementary Fig.8A.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
instrument (Sup. Table1) than the one used for DNA dye excitation/detection. Our results showed that in both
cell lines, in both native and xed detection conditions, and in all DNA dye conditions, the population of cells
with “low” background uorescence intensity was enriched in cells that were in G0/G1 phase of their cell cycle.
Inversely, cells selected based on the “high” background uorescence intensity were dominated by the cell popu-
lation in the G2/M phase (Fig.1, Sup. Anim. 1A and 1B). is phenomenon was observed on a conventional
ow cytometer using the three most commonly used lasers (wavelengths 405, 488, and 639nm) and detecting
background uorescence on three detectors with spectral bandpasses of 450/50, 525/50, and 780/60 (Fig.1 and
Sup. Fig.1). To extend these observations across the full detection spectrum, we performed similar analysis
using spectral ow cytometry (see “Material and methods” section for details). is approach allowed us to
subtract the signal from the DNA dye (FxCycle Far Red Stain) and observe the total uorescent background of
HCT 116 cells over the entire wavelength range. We dened the “lower” and “upper” fraction background cells,
similarly to conventional cytometry, in all 32 channels simultaneously. We then used two approaches to analyze
the data: First, we visualized the population of cells with “low” and “high” background in the cell cycle parameter
as we did for conventional ow cytometry (Fig.2A). Second, we compared the background uorescence of cell
populations at dierent phases of the gated cell cycle based on the amount of DNA labeled with the DNA dye
(Fig.2B). Both approaches conrmed our observations that the fraction of cells with “low” levels of background
uorescence represents cells predominantly in the G0/G1 phase of the cell cycle, while cells with “high” back-
ground uorescence reside predominantly in the G2/M phase (G0/G1: 89% vs. 13%, G2/M: 0% vs. 72%; G0/
G1 background MFI 152 vs. G2/M background MFI 370, Fig.2A,B). Moreover, we provide evidence that this
phenomenon is spectrally independent and can be observed in all uorescent channels used in conventional and
spectral ow cytometers. With these experiments we show that the background uorescence, as assessed with
Count
FxCycle Far Red BG FL
405/488/638//32CH
BG FL
405/488/638//32CH
BG FL
405/488/638//32CH
Count
Low High
DNA content
BG FL 405/488/638//32CH
Count
FxCycle Far Red
Count
FxCycle Far Red
Count
A)
B)
Count
Count
G0/G1 SDNA content G2/M
MFI=152 MFI=296 MFI=370
G1
Count
Count
G2/M
G0/G1
405 Low 405 High
G0/G1=89
G2/M=0
G0/G1=13
G2/M=72
Figure2. Spectral ow cytometry conrms the association of background uorescence and cell cycle state.
HCT 116 cells were stained using LIVE/DEAD Fixable Dead Cell Stain Kit, xed in 4% PFA and DNA was
labelled using FxCycle Far Red Stain. e background uorescence was measured in the range of 420–800nm
using 32 detectors with a spectral analyzer (SONY SP6800). (A) Representative image of background
uorescence and cell cycle prole of HCT 116 cells detected aer simultaneous 405, 488, and 638nm excitation.
Dead cells were excluded and fractions of cells with low (bottom 10%) and high (top 10%) background
uorescence were gated for DNA content (cell cycle) analysis. e values of G0/G1 and G2/M phase represent
the proportion of cells. (B) Examples of reversed gating strategy, when modelled cell phases (FlowJo) were gated
and analyzed for background uorescence. Median uorescence intensity (MFI) was then calculated for each
phase. Data are representative from two independent repetitions. For total DNA content distribution of the
whole population, see Supplementary Fig.8B.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

4
Vol:.(1234567890)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
ow cytometry, shows an association with cell cycle, with highly autouorescent cells being enriched in cells in
later stages of cell cycle.
Experimental modulation of cell cycle progression aects background autouorescence of
the cells. Whether the background uorescence reects the state of cultured cells remains, in the context
of cell cycling, largely unknown. To demonstrate that the cell cycle distribution does indeed aect the back-
ground uorescence intensity, we used several experimental strategies to perturb the cell cycle progression in the
HCT 116 cells invitro. We compared cells collected in the subconuent state of cell culture (control, asynchro-
nously proliferating) with the cells that are in fully conuent state (predominantly in the G0/G1 phase), and
cells that are arrested in the G2/M phase aer nocodazole treatment (commonly used synchronization tech-
nique)25,26. In both native and xed states, fully conuent cells showed lower background uorescence compared
to the subconuent cells (native background uorescence MFI 1349 vs. 2629). On the other hand, cells with
nocodazole-induced cell cycle arrest in the G2/M phase showed a signicant increase in their background uo-
rescence (native background uorescence MFI 2629 vs. 6592, Fig.3). Taken together, the cell cycling reects on
background uorescence of analyzed cell population and vice versa.
Characterization of low and high background autouorescence fractions of the cells. We next
wanted to empirically validate the association between cell cycle state and background uorescence. To achieve
this, we performed a more detailed characterization of cell populations sorted based on “low” and “high” back-
ground uorescence (Fig.4A). In parallel, we sorted cells based on their cell cycle state (Fig.4B) and performed
similar characterization. One of the functional dierences between cells in the cell cycle interphase and mitosis
is linked to the cell adhesion27. erefore, we used sorted cell fractions and performed cell adhesion assay using
a label-free, real-time, impedance-based system28,29. Our data showed signicant dierences in the cell adhesion
between fractions sorted based on “low” and “high” background uorescence (Fig.4C). Similarly, we observed
analogous pattern for cell fractions sorted based on their cell cycle state, i.e. G0/G1 versus G2/M (Fig.4D).
We further performed analysis of protein content in sorted fractions, focusing on the key components of cell
cycle regulation—cyclins30. We sorted cells based on their “low”, “medium”, and “high” background uorescence
Count
Vybrant Violet
Count
BG FL 488//525/50
101102103104105
050K 100K 150K 200K 250K
Live
Count
Vybrant Violet
Count
BG FL 488//525/50
Count
Vybrant Violet
Count
BG FL 488//525/50
101102103104105
050K 100K 150K 200K 250K
Nocodazol
e
Count
FxCycle Far Red
Count
BG FL
488//525/50
101102103104105
050K 100K 150K 200K 250K
Fixed
Count
FxCycle Far Red
Count
BG FL
488//525/50
101102103104105
050K 100K 150K 200K 250K
Count
FxCycle Far Red
Count
BG FL 488//525/50
101102103104105
050K 100K 150K 200K 250K
Background
DNA content
Background
DNA content
MFI=2629
MFI=1349
MFI=6592
MFI=1333
MFI=1762
MFI=5804
G0/G1=40
G2/M=23
G0/G1=82
G2/M=8
G0/G1=74
G2/M=10
G0/G1=55
G2/M=26
G0/G1=9
G2/M=90
G0/G1=21
G2/M=70
Figure3. Experimental modulation of cell cycle progression aects the background autouorescence.
HCT 116 cells were synchronized to G0/G1 phase by cultivation to the full conuency (red line plots) or
arrested in the G2/M phase with nocodazole treatment (blue line plots). Control HCT 116 cells were cultivated
in standard subconuent conditions (green line plots; see Methods for details). Dead cells were excluded from
analysis using LIVE/DEAD Fixable Dead Cell Stain Kit. e cell cycle was then analyzed in both native (Vybrant
DyeCycle Violet) and xed (FxCycle Far Red Stain) conditions. Together with background uorescence. e
numbers of G0/G1 and G2/M phases represent a percentage of cells. e values of G0/G1 and G2/M phase
represent the proportion of cells. Data are representative from two independent repetitions. MFI median
uorescence intensity.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

5
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
Cell index
Time (hrs)
Cell index
LowHigh
*
A)
C)
cyclin A
cyclin B1
cyclin D1
cyclin D3
cyclin E
Low
Med
High
-actin
BG FL
G)
E)
Pre sort
BG FL 405//450/50
Count
Sorted
Cell index
Time (hrs)
D)
DNA content
I)
*
Low High
405 Low 405 High
101102103104105
BG FL 488//525/50
Count
Low High
Medium
488 Low 488 High
488 Medium
B)
Vybrant DyeCycle Violet
Count
Pre sort
G0/G1 G2/M
Cell index
G1 G2M
*
Sorted
G2/M
G0/G1
G0/G1 G2/M
G0/G1
S
G2/M
cell cycle
cyclin A
cyclin B1
cyclin D1
cyclin D3
cyclin E
-actin
H)
F)
J)
G2M
G0G1
*
G0/G1 G2/M
050K 100K 150K 200K 250K
Vybrant DyeCycle Violet
Count
G0G1
S
G2M
G0/G1 G2/M
S
55
55
36
36
55
55
55
55
36
36
55
55
Figure4. Assessment of low and high background autouorescence cell fractions invitro. Representative gure showing fractions
selected for direct functional comparison of live HCT 116 cells sorted based on 10% low and 10% high background uorescence (A)
or G0/G1 and G2/M cell cycle phases aer staining with cell-penetrant, native DNA dye (Vybrant DyeCycle Violet), (B). Sorted cell
fractions were subjected to real-time cell adhesion monitoring with signal being recorded every 15min and cell index being a function
of cell adhesion. e adhesion pattern of cells sorted based on the extent of their background uorescence (C) resemble cells sorted
based on their corresponding cell cycle phase (D). Data are pooled from three technical replicates per condition and three independent
experiments are shown, for details see Methods (*~ P < 0.05 for cell index at 10h), see Sup. Fig.2 for post-sorting purity assessment.
Similarly, distribution of selected cyclins is similar between cells sorted based on their background uorescence (E,G) or cell cycle
phase (F,H). Representative blots are from two independent replicates, uncropped membrane scans are provided in Sup. Fig.7. Lastly,
cell volume as determined with CASY TT follows the same pattern for cells sorted based on their background uorescence (I) or cell
cycle phase (J). Data pooled from three independent experiments and plotted as mean ± S.D. (*~ P < 0.05).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

6
Vol:.(1234567890)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
(Fig.4E), or cell cycle phases aer native, cell-penetrant DNA dye staining (Vybrant DyeCycle Violet, Fig.4F;
post-sort purity shown in Sup. Fig.2A, B was assessed before the sample lysis). Analysis of cyclin expression con-
rmed the expected pattern in the samples sorted based on cell cycle phase (Fig.4H). Strikingly, we observed an
almost identical pattern of cyclin distribution in fractions sorted based on their background uorescence inten-
sity (Fig.4G). We next hypothesized that the increase in cell autouorescence in G2/M is related to a concomi-
tant increase in cell volume/size. To test this, we analyzed the volume of cells from sorted cell fractions using an
electronic cell counter and analyzer system, CASY TT. Our data showed the expected dierences between “low”
versus “high”, and G0/G1 versus G2/M sorted fractions. is analysis conrmed the size similarity between
“low” and G0/G1 sorted fractions, and between “high” and G2/M sorted fractions (Fig.4I,J respectively). Our
characterization of the cell fractions sorted based on their background uorescence showed intriguing similarity
to the cells sorted based on their cell cycle phase. ese results provide additional evidence for a direct relation-
ship between cell cycle state and background uorescence intensity.
Association of cell size and background uorescence is reproducible on dierent ow cytom-
eters. Since the generalization of our observation was unknown, we aimed to address the reproducibility
and robustness of the association between cell cycle/cell size and intensity of background autouorescence. We
performed additional measurements and analyses that involved several routinely used, state-of-the-art ow
cytometers and several cell lines with dierent cell sizes, growth conditions, and species of origin. We included
a human lymphoblast-like cell line, SU-DHL-4, that grows in suspension and hence does not require detach-
ment from the cell culture plastic. e median diameter of SU-DHL-4, HCT 116 and E2 cell lines measured
on the CASY TT system ranged from 12 to 19µm. We analyzed the background uorescence of these cell lines
using four dierent ow cytometers. Our systematic assessment showed that the background uorescence signal
increases together with cell size in all channels and aer dierent excitations, independently of the used cytom-
eter (Fig.5).
One of the outstanding questions that remained unanswered during these analyses was whether this phenom-
enon associates only with cellular objects, or whether it applies to particles in general. We analyzed polystyrene
particles with specic sizes, ranging from 2 to 14.7µm in diameter, on 5 dierent ow cytometers. First, we
compared populations of particles with dierent sizes on forward and side scatter (Sup. Fig.3A). Second, we
analyzed these populations on uorescence channels (Sup. Fig.3B). Based on the quantied green uorescence
channel medians we conrmed that increasing particle size is associated with the increase in signal in uores-
cence channels (Sup. Fig.3C). is observation was conrmed in all uorescence channels, and the pattern of
increasing signal with particle size was also evident for all used lasers (Sup. Fig.4A). Finally, we performed the
same measurements with a spectral ow cytometer in 32 uorescence channels, splitting the light spectra from
420 to 800nm into small fractions (Sup. Fig.4B). e connection between increasing uorescence background
and increasing particle size was present throughout the entire range of the 32 detectors. In summary, the rela-
tionship between cell size and background uorescence was reproducible across dierent ow cytometers and
can be generalized to a non-cellular particles, such as polystyrene beads.
Distribution of the uorescence signal within asynchronous cell population is associated with
cell cycle state. Flow cytometry is used to assess the presence or quantify the amount of expression of
selected antigens with uorochrome-tagged antibodies. e next logical step was therefore to assess the eects
of cell cycle on the distribution of specic immunouorescent stain. To gain a comprehensive understand-
ing of such relationship, we measured the expression of 332 cell surface markers and 10 isotype controls in
HCT 116 cells along with the DNA staining, allowing for simultaneous cell cycle analysis. Following ow cyto-
metric analysis, we used similar gating strategy as shown in Sup. Fig.1 and delineated the upper and lower 10%
of cells in terms of each surface marker expression. With such strategy, we examined the cell cycle distribution
prole of the “low” and “high” populations in the commonly observed scenarios: (1) negative expression—anti-
gen not present, with a signal of intensity similar to that of isotype control, (2) medium expression—weakly
expressed antigen that exhibits only a “shi” in the intensity, and (3) positive expression—highly expressed
antigen by the entire cell population. For each scenario, we selected a representative group of cell surface mark-
ers (Fig.6). Direct comparison between isotype controls and the three scenarios described above conrmed that
the cell cycle distribution was related to the uorescence intensity in extensive array of antigens. e fraction of
cells dened based on the lower 10% values of uorescence intensity is mainly enriched for cells in the G0/G1
phase of the cell cycle, whereas the fraction from the upper 10% values represents mainly cells in the G2/M phase
(see data on the proportion of cells in G0/G1 and G2/M in Fig.6). is phenomenon is obvious in all categories,
even in the positive population with strong specic uorescence signals. For independent conrmation, we per-
formed cell sorting in the native state based on the low, medium, and high uorescence intensities of two model
surface antigens, EpCAM and integrin β5 (Fig. 7A,B; see Sup. Fig.6 for post-sort purity assessment). ese
sorted fractions were then stained for DNA content with Vybrant DyeCycle Violet, and reanalyzed immediately
(Fig.7C,D). e cell cycle distribution recapitulated the previously observed pattern, with low-uorescence
sorted fraction being enriched in the G0/G1 cells, medium-uorescence population enriched in the G0/G1/S
cells, and the high-uorescence sorted population enriched in the G2/M cells (see data on the proportion of
cells in G0/G1 and G2/M in Fig.7). Overall, we provide strong evidence that the variability in the distribution of
background and specic uorescence signal is related to the cell cycle status. Orthogonally, the cell cycle distri-
bution aects the distribution of both background and specic uorescence signals.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

7
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
Discussion
Both cell cycle and cell volume are well-known sources of bias, introducing within-the-cell-type phenotypic
and functional heterogeneity in the scRNA-seq-generated results4–7. With recent technological advancements in
cytometry, such consideration of the cell cycle/cell volume eects in polychromatic ow cytometry data becomes
necessary. Here, we describe the association between cell cycle state/cell size and the distribution of uorescence
intensity, systematically dissected by ow cytometry. First, we demonstrated that the “low” fraction of back-
ground uorescence signal was enriched mostly in the cells in G0/G1 phase, while the “high” fraction contained
cells mostly in the G2/M phase. Employing dierent instrumental setups, we showed that this phenomenon is
spectrally independent and can be observed in all assessed uorescent channels used in conventional and spectral
ow cytometers. is relationship between cell cycle and cell size was conrmed by additional experiments in
which DNA was rst labeled natively, and its intensity analyzed by ow cytometry was used to determine the
intensity of background uorescence in dierent spectral ranks. Experimental manipulation of the cell cycle
prole and subsequent analysis of autouorescence also corroborated this relationship. For an ultimate valida-
tion, we chose to sort cells based on the intensity of background uorescence and analyze the expression levels
of key cell cycle-regulating cyclins, along with a functional approach that tested their ability to adhere to the cell
culture surface. ese analyses showed that the cells sorted using “low” and “high” approach diered in their
ability to adhere, and these results are consistent with previously published studies in which a higher ability to
adhere was demonstrated for cells in the G2/M phase of the cell cycle27. Moreover, the cyclin expression prole
Aria Attune
Calibu
rV
erse
12.1
SU-DHL-4
17.0
HCT 116
19.2
E2
12.1
SU-DHL-4
17.0
HCT 116
19.2
E2
12.1
SU-DHL-4
17.0
HCT 116
19.2
E2
12.1
SU-DHL-4
17.0
HCT 116
19.2
E2
Size (µm, lin)
Median (log)
Figure5. Average cell size correlates with the background uorescence intensity. e size of the three
dierent cell lines SU-DHL-4 (average diameter 12µm), HCT 116 (17µm), and E2 (19µm) was determined in
suspension using the CASY TT cell counter. e unlabeled cells were then analyzed using four dierent ow
cytometers and median background uorescence for all available lasers (BD FACSAria II SORP) or detectors
(TFS Attune, BD FACSCalibur, BD FACSVerse) was determined. Data are plotted as median ± S.D. from at least
three biological replicates.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

8
Vol:.(1234567890)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
κ
ISONegativeMediumPositive
Parental Low High
DNA content
G0/G1=90
G2/M=2
G0/G1=89
G2/M=3
G0/G1=91
G2/M=1
G0/G1=88
G2/M=3
G0/G1=89
G2/M=2
G0/G1=89
G2/M=3
G0/G1=78
G2/M=6
G0/G1=69
G2/M=8
G0/G1=85
G2/M=3
G0/G1=81
G2/M=16
G0/G1=63
G2/M=13
G0/G1=74
G2/M=4
G0/G1=28
G2/M=45
G0/G1=27
G2/M=46
G0/G1=35
G2/M=41
G0/G1=14
G2/M=58
G0/G1=14
G2/M=54
G0/G1=18
G2/M=55
G0/G1=37
G2/M=35
G0/G1=30
G2/M=46
G0/G1=14
G2/M=60
G0/G1=46
G2/M=30
G0/G1=42
G2/M=33
G0/G1=33
G2/M=41
Figure6. High-throughput cell surface marker screen conrms general association between uorescence
distribution and cell cycle. Histograms in the rst column represent characteristic examples of isotype (negative)
controls, markers with undetectable expression (< 1% positivity), medium expression (~ 50% positivity) and
markers with high, positive expression (> 99% positivity). e fractions of the cells with 10% low and 10% high
specic uorescence (PE channel) intensities were gated and analyzed for cell cycle distribution and are shown
in the second and third column. e values of G0/G1 and G2/M phase represent cell proportions. Values above
gating lines in the rst histogram column of the histograms represent median uorescence intensities of gated
cell fraction. Screen was performed with LEGENDScreen human PE kit, for details see Methods. For total DNA
content distribution of the whole population and gating strategy see Sup. Fig.5.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

9
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
corresponded with that of cells that were sorted based on the DNA amount. e simplest explanation was that
the rise in cellular autouorescence, linked to the cell cycle progression, is related to the change in the cell size31.
e relationship between autouorescence and cell size has been previously demonstrated in several studies that,
however, did not provide a direct link to changes in the cell cycle distribution32–34. We therefore conrmed that
cells sorted based on high background uorescence are typically larger in cell volume, corresponding to cells
sorted based on the amount of DNA in the G2/M phase of the cell cycle. e reproducibility and robustness of
these associations were addressed by measurements involving several ow cytometers and cell lines with dier-
ent cell sizes. We conrmed that the relationship between cell size and background uorescence is reproducible
across dierent ow cytometers and is not only related to dierences in cell size but is also observed for other
particles, such as polystyrene beads. Finally, we conrmed that the relationship between cell cycle and back-
ground uorescence distribution also remains valid in the case of specic uorescence. Analysis of 342 surface
molecules together with cell cycle conrmed that the variability of specic uorescence distribution (as a measure
of individual surface antigen expression) corresponded to the cell cycle distribution observed for background
uorescence. Overall, we showed that cell cycle status is related to both background and specic uorescence
signals of dierent abundance. Additionally, we conclude that cell cycle distribution aects the distribution of
both background and specic uorescence signals. We are aware of some of the limitations of our study, in par-
ticular, we are unable to simply distinguish between the consequences of intrinsic cell size changes and separate
A) C) Low Medium High
Count
Vybrant DyeCycle Violet
EpCAM
Count
EpCAM
Low
Medium High
B)
Integrin  - 5
Count
Integrin

- 5
Low
Medium
High
Low Medium High
Count
Vybrant DyeCycle Violet
D)
Pre sort Sorted
G0/G1=75
G2/M=6
G0/G1=64
G2/M=21
G0/G1=35
G2/M=46
G0/G1=83
G2/M=4
G0/G1=75
G2/M=19
G0/G1=34
G2/M=41
Figure7. Post-sorting analysis of cell cycle distribution in the fractions of cells with dierent levels of specic
stain uorescence. Viable HCT 116 cells were sorted in their native state based on the low, medium, and high
specic uorescence intensity aer cell surface staining for EpCAM (A) or integrin β5 (B). Sorted fractions
were subsequently stained for DNA content with cell permeable DNA dye (Vybrant DyeCycle Violet) and
immediately re-analyzed for cell cycle. Values for G0/G1 and G2/M phases in the sorted low, medium, and
high uorescence for EpCAM (C) and integrin β5 (D) cells represent cell proportions. e values of G0/G1 and
G2/M phase represent cell proportions. Data are representative from at least three independent repetitions. For
post-sort purity assessment, see Sup. Fig.6.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

10
Vol:.(1234567890)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
them from those associated with cell cycle phase changes, further recognizing that cell dierentiation/maturation
may be fundamentally involved in the spectrum of these changes. Nevertheless, based on the evidence presented
we argue that the interpretation of data obtained solely from comparisons of populations dened in terms of
symmetric uorescence signal distribution could be misleading. Without sucient validation, these results can
be confounded by cell cycle/size distribution, and we advise further conrmation using other, complementary
techniques. ese observations, specically the eect of cell cycle state and cell size, should be considered also
during visualization of polychromatic ow cytometry data using t-SNE and other popular algorithms.
Material and methods
Cells and cell culture. e mouse prostate cancer cell line cE2 and E235 (a kind gi from Dr. Pradip Roy-
Burman, University of Southern California, CA, USA) was maintained in Dulbecco’s modied Eagle’s medium
(DMEM) high glucose with GlutaMAX (32430, Gibco, ermo Fisher Scientic, USA, TFS) supplemented with
rhEGF (6ng/mL, Sigma-Aldrich, Merck, USA), insulin (5μg/mL, Sigma-Aldrich, Merck), bovine pituitary
extract (25μg/mL, Hammond Cell Tech, USA), penicillin (100 U/mL) and streptomycin (0.1mg/mL; PAA,
Austria), and 10% fetal bovine serum (PAA). Human colon adenocarcinoma cells HCT 116 (a kind gi from
Dr. Bert Vogelstein, Johns Hopkins University, MD, USA) were maintained in McCoy’s 5A (modied) medium,
GlutaMAX (36600, TFS) supplemented with penicillin (100 U/mL) and streptomycin (0.1mg/mL,TFS) and
10% heat-inactivated fetal bovine serum (TFS). Human B lymphoblasts SU-DHL-4 (a kind gi from Dr. Mar-
tin Trbušek, Masaryk University, Czech Republic) were cultured in Roswell Park Memorial Institute’s medium
(RPMI) 1640 with GlutaMAX (72400, TFS) and 10% fetal bovine serum (TFS), penicillin (100U/mL) and strep-
tomycin (0.1mg/mL; TFS) addition36. All cell lines were maintained in cell culture plastic from TPP (Switzer-
land) or BD Falcon (BD Biosciences, CA, USA) in a humidied incubator at 37°C in an atmosphere of 5%
CO2. e cells were harvested by incubation in 0.05% EDTA in PBS followed by trypsinization (0.25% w/v
trypsin/0.53mM EDTA in PBS) and counted with CASY TT automatic cell counter (Innovatis AG, Germany).
Cell suspensions were ltered through sterile 70- or 100-μm syringe lters (Filcons, Germany) before analysis
or sorting.
Instrumentation. Cell sortings and some of the experiments were performed on FACSAria II SORP system
(BD Biosciences) equipped with ve lasers (excitation wavelengths: 355, 405, 488, 561 and 639nm, respectively).
For all sortings, a 100-μm nozzle (20 psi) was used, and post sorting purity was analyzed immediately aer sort-
ing. We used four additional ow cytometers to conduct the experiments: FACSVerse (BD Biosciences), Attune
(1st generation, TFS), FACSCalibur (BD Biosciences), and SP6800 spectral analyzer (SONY). e advantage of
including spectral analyzer on the top of conventional ow cytometers is that it allows for more detailed spectral
detection. SP6800 contains similar laser excitation sources, and we used 32 channels with narrow bandpasses
starting at 420and ending at 800nm for detection (see Sup. Table1 for details). A specic feature of this system
is its ability to calculate a so-called virtual parameter that collects the signal from all 32 channels. Furthermore,
the analyzer allows to apply a spectral unmixing algorithm, detecting signal of the other uorescent markers in
the panel. Spectral unmixing is calculated with the signal previously collected from individually stained controls
over the entire 420–800nm spectrum range. Details about the instruments’ conguration are shown in Sup-
plementary Table1.
Flow cytometry staining and cell sorting. Samples of xed HCT116 cells stained for viability (LIVE/
DEAD Fixable Violet Dead Cell Stain Kit, TFS) and cell cycle (FxCycle Far Red Stain, TFS) were analyzed on
SP6800 spectral analyzer aer 405 and 488 together with 638nm excitation on 32 channels in narrow bands for
uorescence detection. Detectors covered the range from 420 to 800nm, using SONY’s soware we combined all
32 channels into one parameter “AF”. Dead cells were previously excluded. For purpose of real-time adherence
monitoring, we sorted HCT 116 cells (80K per group) in 2 repetitions based on autouorescence on 405nm
laser and cell cycle phase (staining with Vybrant DyeCycle Violet Stain, V35003, TFS) (Fig.3A). Dead cells were
excluded by LIVE/DEAD Fixable Far Red Dead Cell Stain Kit. e purity of sorted samples was controlled prior
to seeding. HCT 116 cells (800K cells per group and repetition) were also sorted for protein analysis with west-
ern blot (see below) based on autouorescence on 488nm laser (Fig.4A). Dead cells were excluded using pro-
pidium iodide. Alternatively, HCT 116 cells originated from the same ask (750K cells per group and repetition)
were sorted based on cell cycle distribution (staining with Vybrant DyeCycle Violet Stain) (Fig.4C). e purity
of sorted cells was reanalyzed on the sorter immediately aer sorting (Sup. Fig.2). HCT 116 cells were stained on
viability (LIVE/DEAD Fixable Far Red Dead Cell Stain Kit) together with biotin-conjugated CD326 (EpCAM)
Monoclonal Antibody (1B7) (1:200, eBioscience, TFS, cat. no. 13-9326-82) or unconjugated Puried anti-
human integrin β5 Antibody (1:100, BioLegend, cat. no. 345202). For unspecic binding and secondary staining
were used streptavidin FITC (1:2000, eBioscience, cat. no. 11-4317-87) or donkey anti-Mouse IgG (H + L) highly
cross-adsorbed secondary antibody Alexa Fluor 488 (1:500, eBioscience, TFS, cat. no. A21202) antibodies. Cells
were sorted into low, medium and high populations divided into thirds on both markers. Immediately aer
sorting, post sort purity was analyzed and sorted cells from each fraction were stained for DNA content using
Vybrant DyeCycle Violet Stain (1:1000, Invitrogen) as described below.
Data analysis. Cell doublets, aggregates and debris were excluded from the analysis based on a dual-param-
eter dot plot in which the pulse ratio (signal area/signal high; y-axis) versus signal area (x-axis) was displayed.
Dead cells were excluded from the analysis by staining with propidium iodide (Sigma-Aldrich, Merck) or LIVE/
DEAD Fixable Dead Cell Stain (dierent uorescence reactive dyes; Invitrogen, TFS). Cytometric data were
recorded using FACSDiva soware (Version 6.1.3; BD Biosciences), Attune Cytometric Soware (Version 2.1;
Content courtesy of Springer Nature, terms of use apply. Rights reserved

11
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
TFS) and FACSuite (Version 1.0.5.3841 and 1.0.6; BD Biosciences). Data analysis was performed using FlowJo
soware (Version 7.6.5 and 10.0.7, BD Biosciences). List mode data are uploaded into the Flow Repository data-
base of ow cytometry experiments (https:// owr eposi tory. org/ id/ FR- FCM- ZYFP).
Cell cycle analysis. Trypsinized and PBS-washed cE2 or HCT 116 cells were stained for cell cycle imme-
diately in their native state, or aer xation (70% ethanol or 4% paraformaldehyde) and permeabilization (0.1%
Triton-X100). Live cells were stained in complete media with Hoechst 33342 (Sigma-Aldrich, Merck) or Vybrant
DyeCycle Violet Stain (Invitrogen) for 45min at 37°C. Fixed and permeabilized cells were stained with Vin-
delov’s solution37, DAPI, or FxCycle Far Red Stain (Invitrogen, TFS). Staining was performed for 30min at 37°C
for Vindelov, at room temperature for DAPI, and at 4°C for FxCycle Far Red Stain.
Cell cycle synchronization. HCT 116 cells were synchronized in G1 and G2/M phases prior to the cell
cycle staining. Cells were maintained for 8days in the same culture dish, with media change every 2–3days, to
reach 100% conuence and synchronize in G0/G1. Subconuent (70–80%) cells were used as a control sample.
For G2/M arrest, cells were treated for 24h with nocodazole (nal concentration 100ng/mL, Sigma-Aldrich,
Merck), and only oating cells were collected for further processing.
Cell surface markers screening. HCT 116 cells were expanded, harvested and 3 × 108 cells was stained for
cell cycle (Vybrant DyeCycle Violet Stain) and viability (LIVE/DEAD Fixable Far Red Dead Cell Stain Kit) and
dispensed into LEGENDScreen Human Cell Screening (PE) Kit plates for surface staining with 332 cell surface
markers and 10 isotype controls (cat. no. 700001, BioLegend, CA, USA). Further processing of cells was done
according to the manufacturer’s recommendation. Cell from each plate well were recorded on BD FACSVerse for
2min per well on medium speed. Only viable (LIVE/DEAD negative), single cells (FSC-A vs. FSC-H followed
by single-cell selection on Vybrant-A vs. Vybrant-W plot) without debris (FSC-A vs. SSC-A) were selected for
further analysis.
Real‑time cell adherence analysis. Cell adherence of sorted cell populations was monitored in real-time
using the xCELLigence real-time cell analysis (RTCA) DP system in combination with E-plate View inserts,
equipped with the RTCA Soware v1.2 (Acea Biosciences, USA). Adherence was inferred by the measure-
ment of electrical impedance across microelectrodes that integrated into the apical surface of the well bottom
of E-plates38. Every cell attached to microelectrodes acts as electrical insulator in conductive cell culture media
and is measured as an increase in total impedance. First, a standard background measurement was recorded
using 200µL of complete culture medium every minute for 5min. Next, 20,000 of sorted HCT 116 cells were
seeded manually per each well (in multiplicate of 3 wells for each subpopulation of G0/G1, G2/M phase, 10% low
and 10% highly background uorescent cells). We then used cell index, which represents normalized electrical
impedance, to reect the cell adherence. Impedance signal was recorded continually every 15min for up to 10h.
Cell volume measurement. Cells from dierent populations (G0/G1, G2/M phase cells, 10% “low” and
10% “high” background uorescent cells—channel 405//450/50) were sorted as described above and analyzed
on CASY TT cell counter for cell volume and viability. At least 800 cells per replicate were analyzed. For sorted
HCT 116 fractions in Fig.4 at least 250 cells were analyzed.
SDS‑PAGE and western blot analysis. Sorted cells were briey spun, and cell pellets were ash frozen
on dry ice, stored at − 80°C, and then thawed on ice and lysed in radioimmunoprecipitation assay buer (RIPA)
with the addition of Protease inhibitor mix G (3910102, Serva) and Phosphatase inhibitor mix II (39055.02,
Serva, Germany). RIPA was prepared fresh in-house and consisted of 150mM NaCl; 50mM Tris/HCl, pH 7.4;
1% Igepal CA-630 (I8896, Sigma-Aldrich, Merck); and 0.25% sodium deoxycholate (D6750, Sigma-Aldrich,
Merck). Lysates were briey sonicated, cleared, and the concentration of proteins was assessed using DC Protein
Assay Kit (BioRad, CA, USA). Lysate concentrations were adjusted so they were all equal by dilution with RIPA
and mixed with 5 × Laemmli loading dye (nal: 2% SDS; 50mM Tris, pH 6.8; 0.02 bromophenol blue; 100mM
DTT; 1% glycerol). Samples were boiled for 10min at 90°C and 10µg of proteins were loaded. Proteins were
separated by SDS-PAGE using Hoefer miniVE vertical electrophoresis unit), blotted onto PVDF Immobilon P
Transfer Membrane (IPVH00010, Millipore, Merck) and blocked in 5% nonfat dry milk, pH 7.2 in TBS (20mM
Tris–HCl pH 7.2; 140mM NaCl containing 0.05% Tween-20) for 1h at room temperature. Membranes were
incubated with following primary antibodies at 4°C overnight: cyclin A (1:500 in 5% milk, sc-751, Santa Cruz
Biotechnology, CA, USA, SCBT); cyclin B1 (1:300 in 5% BSA, sc-245, SCBT); cyclin D1 (1:500 in 5% milk,
sc-20044, SCBT); cyclin D3 (1:500 in 5% milk, sc-182, SCBT); cyclin E (1:500 in 5% milk, sc-481, SCBT). Follow-
ing secondary antibodies were used: ECL anti-mouse HRP linked whole antibody (1:3000 in 5% milk, NA931,
GE Healthcare Biosciences) and ECL anti-rabbit HRP linked whole antibody (1:3000 in 5% milk, NA934, GE
Healthcare Biosciences). Chemiluminescent signals were detected using Immobilon Western HRP Substrate
(WBKLS05000, Millipore, Merck) and visualized on X-ray lms (Agfa, Germany). Detection of ß-actin (1:4000
in 5% milk, A5441, Sigma-Aldrich, Merck) served as a control of equal loading. Blotting membranes were cut
prior to hybridization with the antibodies, scans of stained membranes with visible protein ladders and edges, in
their entirety, are presented in Supplementary Fig.7.
Particle size analysis. e mixture of PBS and polystyrene particles of all sizes from (Sphero Particle Size
Standard Kit, Spherotech) was prepared by dispensing 2 drops of each particle size into 1mL PBS. Background
Content courtesy of Springer Nature, terms of use apply. Rights reserved

12
Vol:.(1234567890)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
uorescence of suspended particles was recorded on following ow cytometers: BD FACSAria II SORP, TFS
Attune (1st gen.), BD FACSCalibur, BD FACSVerse, and SONY SP6800 spectral analyzer in dierent uorescent
channels at low speed (at least 50,000 events were recorded). Pellets from E2, HCT 116, and SU-DHL-4 cell lines
were prepared as described above. e mean cell diameter for each cell line was quantied with CASY TT cell
counter. Measurement of background uorescence for each cell line was then performed on all 4 cytometers.
Standardized suspension of each cell line was used for this analysis (2 million cells per 1mL of PBS).
Data reproducibility and statistical analysis. For the high-throughput antibody-based screen,
HCT 116 cell line was analyzed one well per antibody. e initial screen was performed once. All further cell
line-based experiments were performed independently at least three times. e percentage of G0/G1 and G2/M
phases were calculated using Dean-Jett-Fox modelling in FlowJo v10.7.2 (BD Biosciences). Statistical analyses
were performed in GraphPad Prism v9.2 (GraphPad Soware, USA). Plotting and analysis was performed in
SigmaPlot for Windows (Version 10.0, Systat Soware). P values were calculated with paired t-test and ratio
paired t-test (two-tailed), if not stated otherwise.
Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary
Information le). List mode data were deposited to the Flow Repository database of ow cytometry experiments
(https:// owr eposi tory. org/ id/ FR- FCM- ZYFP). Additional raw data les are available from the corresponding
authors upon reasonable request.
Received: 7 June 2022; Accepted: 21 March 2023
References
1. Pauklin, S. & Vallier, L. e cell-cycle state of stem cells determines cell fate propensity. Cell 155, 135–147. https:// doi. org/ 10.
1016/j. cell. 2013. 08. 031 (2013).
2. Xia, X., Owen, M. S., Lee, R. E. C. & Gaudet, S. Cell-to-cell variability in cell death: Can systems biology help us make sense of it
all?. Cell Death Dis. 5, e1261–e1261. https:// doi. org/ 10. 1038/ cddis. 2014. 199 (2014).
3. Pernicova, Z. et al. e role of high cell density in the promotion of neuroendocrine transdierentiation of prostate cancer cells.
Mol. Cancer 13, 113. https:// doi. org/ 10. 1186/ 1476- 4598- 13- 113 (2014).
4. Buettner, F. et al. Computational analysis of cell-to-cell heterogeneity in single-cell RNA-sequencing data reveals hidden subpopu-
lations of cells. Nat. Biotechnol. 33, 155–160. https:// doi. org/ 10. 1038/ nbt. 3102 (2015).
5. Padovan-Merhar, O. et al. Single mammalian cells compensate for dierences in cellular volume and DNA copy number through
independent global transcriptional mechanisms. Mol. Cell 58, 339–352. https:// doi. org/ 10. 1016/j. molcel. 2015. 03. 005 (2015).
6. Rapsomaniki, M. A. et al. Cell CycleTRACER accounts for cell cycle and volume in mass cytometry data. Nat. Commun. 9, 632.
https:// doi. org/ 10. 1038/ s41467- 018- 03005-5 (2018).
7. Barron, M. & Li, J. Identifying and removing the cell-cycle eect from single-cell RNA-Sequencing data. Sci. Rep. 6, 33892. https://
doi. org/ 10. 1038/ srep3 3892 (2016).
8. Liu, J., Fan, Z., Zhao, W. & Zhou, X. Machine intelligence in single-cell data analysis: Advances and new challenges. Front. Genet.
https:// doi. org/ 10. 3389/ fgene. 2021. 655536 (2021).
9. Sahir, F., Mateo, J. M., Steinho, M. & Siveen, K. S. Development of a 43 color panel for the characterization of conventional and
unconventional T-cell subsets, B cells, NK cells, monocytes, dendritic cells, and innate lymphoid cells using spectral ow cytometry.
Cytom. Part https:// doi. org/ 10. 1002/ cyto.a. 24288 (2020).
10. Park, L. M., Lannigan, J. & Jaimes, M. C. OMIP-069: Forty-color full spectrum ow cytometry panel for deep immunophenotyping
of major cell subsets in human peripheral blood. Cytom. Part A 97, 1044–1051. https:// doi. org/ 10. 1002/ cyto.a. 24213 (2020).
11. Brummelman, J. et al. Development, application and computational analysis of high-dimensional uorescent antibody panels for
single-cell ow cytometry. Nat. Protoc. 14, 1946–1969. https:// doi. org/ 10. 1038/ s41596- 019- 0166-2 (2019).
12. Mazza, E. M. C. et al. Background uorescence and spreading error are major contributors of variability in high-dimensional ow
cytometry data visualization by t-distributed stochastic neighboring embedding. Cytom. Part A 93, 785–792. https:// doi. org/ 10.
1002/ cyto.a. 23566 (2018).
13. Miranda-Lorenzo, I. et al. Intracellular autouorescence: A biomarker for epithelial cancer stem cells. Nat. Methods 11, 1161–1169.
https:// doi. org/ 10. 1038/ nmeth. 3112 (2014).
14. Larcher, V. et al. An autouorescence-based method for the isolation of highly puried ventricular cardiomyocytes. Cardiovasc.
Res. 114, 409–416. https:// doi. org/ 10. 1093/ cvr/ cvx239 (2018).
15. Shah, A. T., Cannon, T. M., Higginbotham, J. N., Coey, R. J. & Skala, M. C. Autouorescence ow sorting of breast cancer cell
metabolism. J. Biophoton. 10, 1026–1033. https:// doi. org/ 10. 1002/ jbio. 20160 0128 (2017).
16. Chacko, J. V. & Eliceiri, K. W. Autouorescence lifetime imaging of cellular metabolism: Sensitivity toward cell density, pH, intra-
cellular, and intercellular heterogeneity. Cytom. A 95, 56–69. https:// doi. org/ 10. 1002/ cyto.a. 23603 (2019).
17. Bagri-Manjrekar, K. et al. In vivo autouorescence of oral squamous cell carcinoma correlated to cell proliferation rate. J. Cancer
Res. er. 14, 553–558. https:// doi. org/ 10. 4103/ 0973- 1482. 172710 (2018).
18. Croce, A. C. & Bottiroli, G. Autouorescence spectroscopy and imaging: A tool for biomedical research and diagnosis. Eur. J.
Histochem. EJH 58, 2461. https:// doi. org/ 10. 4081/ ejh. 2014. 2461 (2014).
19. Mosiman, V. L., Patterson, B. K., Canterero, L. & Goolsby, C. L. Reducing cellular autouorescence in ow cytometry: An insitu
method. Cytometry 30, 151–156. https:// doi. org/ 10. 1002/ (SICI) 1097- 0320(19970 615) 30:3% 3c151:: AID- CYTO6% 3e3.0. CO;2-O
(1997).
20. Kolenc, O. I. & Quinn, K. P. Evaluating cell metabolism through autouorescence imaging of NAD(P)H and FAD. Antioxid. Redox
Signal. https:// doi. org/ 10. 1089/ ars. 2017. 7451 (2018).
21. You, S. et al. Intravital imaging by simultaneous label-free autouorescence-multiharmonic microscopy. Nat. Commun. 9, 2125.
https:// doi. org/ 10. 1038/ s41467- 018- 04470-8 (2018).
22. Tu, H. et al. Stain-free histopathology by programmable supercontinuum pulses. Nat. Photon. 10, 534–540. https:// doi. org/ 10.
1038/ nphot on. 2016. 94 (2016).
23. Kanchwala, N., Kumar, N., Gupta, S. & Lokhandwala, H. Fluorescence spectroscopic study on malignant and premalignant oral
mucosa of patients undergoing treatment: An observational prospective study. Int. J. Surg. 55, 87–91. https:// doi. org/ 10. 1016/j.
ijsu. 2018. 05. 029 (2018).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

13
Vol.:(0123456789)
Scientic Reports | (2023) 13:4889 | https://doi.org/10.1038/s41598-023-31990-1
www.nature.com/scientificreports/
24. Wizenty, J. et al. Autouorescence: A potential pitfall in immunouorescence-based inammation grading. J. Immunol. Methods
456, 28–37. https:// doi. org/ 10. 1016/j. jim. 2018. 02. 007 (2018).
25. Harper, J. V. In Cell Cycle Control: Mechanisms and Protocols (eds Humphrey, T. & Brooks, G.) 157–166 (Humana Press, 2005).
26. Langan, T. J., Rodgers, K. R. & Chou, R. C. In Cell Cycle Synchronization: Methods and Protocols (ed. Banfalvi, G.) 97–105 (Springer,
2017).
27. Jones, M. C., Zha, J. & Humphries, M. J. Connections between the cell cycle, cell adhesion and the cytoskeleton. Philos. Trans. R.
Soc. B Biol. Sci. 374, 20180227. https:// doi. org/ 10. 1098/ rstb. 2018. 0227 (2019).
28. Vistejnova, L. et al. e comparison of impedance-based method of cell proliferation monitoring with commonly used metabolic-
based techniques. Neuroendocrinol. Lett. 30, 121–127 (2009).
29. Slabakova, E. et al. Opposite regulation of MDM2 and MDMX expression in acquisition of mesenchymal phenotype in benign
and cancer cells. Oncotarget 6, 36156–36171. https:// doi. org/ 10. 18632/ oncot arget. 5392 (2015).
30. Lim, S. & Kaldis, P. Cdks, cyclins and CKIs: Roles beyond cell cycle regulation. Development 140, 3079–3093. https:// doi. org/ 10.
1242/ dev. 091744 (2013).
31. Amodeo, A. A. & Skotheim, J. M. Cell-size control. Cold Spring Harbor Perspect. Biol. 8, a019083–a019083. https:// doi. org/ 10.
1101/ cshpe rspect. a0190 83 (2016).
32. Tzur, A., Moore, J. K., Jorgensen, P., Shapiro, H. M. & Kirschner, M. W. Optimizing optical ow cytometry for cell volume-based
sorting and analysis. PLoS ONE 6, e16053. https:// doi. org/ 10. 1371/ journ al. pone. 00160 53 (2011).
33. Bertolo, A., Baur, M., Guerrero, J., Pötzel, T. & Stoyanov, J. Autouorescence is a reliable invitro marker of cellular senescence in
human mesenchymal stromal cells. Sci. Rep. 9, 2074. https:// doi. org/ 10. 1038/ s41598- 019- 38546-2 (2019).
34. Schaue, D., Ratikan, J. A. & Iwamoto, K. S. Cellular autouorescence following ionizing radiation. PLoS ONE 7, e32062. https://
doi. org/ 10. 1371/ journ al. pone. 00320 62 (2012).
35. Liao, C.-P. et al. Mouse models of prostate adenocarcinoma with the capacity to monitor spontaneous carcinogenesis by biolumi-
nescence or uorescence. Cancer Res. 67, 7525–7533. https:// doi. org/ 10. 1158/ 0008- 5472. can- 07- 0668 (2007).
36. Tao, K., Fang, M., Alroy, J. & Sahagian, G. G. Imagable 4T1 model for the study of late stage breast cancer. BMC Cancer 8, 228
(2008).
37. Vindelov, L. L. Flow microuorometric analysis of nuclear DNA in cells from solid tumors and cell suspensions. A new method
for rapid isolation and straining of nuclei. Virchows Arch. B Cell Pathol. 24, 227–242 (1977).
38. Staršíchová, A. et al. Dynamic monitoring of cellular remodeling induced by the transforming growth factor-β1. Biol. Proced.
Online 11, 316–324. https:// doi. org/ 10. 1007/ s12575- 009- 9017-9 (2009).
Acknowledgements
is work was supported by the Czech Science Foundation, grant nr. 20-22984S (KS) and 21-11585S (KS); the
Ministry of Health of the Czech Republic, grant nr. 18-08-00245 (KS, TK), NU20J-07-00028 (MR); the European
Structural and Investment Funds, Operational Program Research, Development and Education, Preclinical Pro-
gression of New Organic Compounds with Targeted Biological Activity” (Preclinprogress)—CZ.02.1.01/0.0/0.
0/16_025/0007381; e project National Institute for Cancer Research (Programme EXCELES, ID Project No.
LX22NPO5102)—Funded by the European Union—Next Generation EU (KS, TK, MR). JR is supported by the
Terri Brodeur Breast Cancer Foundation and MSKCC Support Grant P30 CA008748. e authors would like to
thank Dr. Pradip Roy-Burman, Dr. Bert Vogelstein, and Dr. Martin Trbušek for providing cell lines used in this
study, Šárka Šimečková for help with western blots, Iva Lišková, Martina Urbánková and Kateřina Svobodová for
technical assistance. We completed this work in memory of our colleague and friend Vlastimil Mašek.
Author contributions
R.F. performed the experiments and analyzed the data, interpreted the data and wrote and reviewed the manu-
script. Z.K. assisted with invitro experiments and reviewed the manuscript, J.R. assisted with LegendScreen and
reviewed the manuscript, T.K. and M.R. helped with the spectral measurements, data analysis, and reviewed the
manuscript. K.S. supervised, conceptualized and designed the study, interpreted the data, and wrote and reviewed
the manuscript. All authors approved the nal version of this manuscript.
Competing interests
e authors declare no competing interests.
Additional information
Supplementary Information e online version contains supplementary material available at https:// doi. org/
10. 1038/ s41598- 023- 31990-1.
Correspondence and requests for materials should be addressed to K.S.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access is article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
© e Author(s) 2023
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com