manuscript - helda.helsinki.fi
TRANSCRIPT
1
Glutamine metabolism controls stem cell fate reversibility and long-term maintenance in the hair
follicle
Christine S. Kim1,2#, Xiaolei Ding3#, Kira Allmeroth1,2§, Leah C. Biggs4,5,6§, Olivia I. Kolenc7,
Nina L’Hoest1,2, Carlos Andrés Chacón-Martínez1,2, Christian Edlich-Muth1, Patrick
Giavalisco1, Kyle P. Quinn7, Martin S. Denzel1,2, Sabine A. Eming2,3,8,9* and Sara A.
Wickström1,2,4,5,6*
1Max Planck Institute for Biology of Ageing, Cologne, Germany 2 Cluster of Excellence Cellular Stress Responses in Ageing-associated Diseases (CECAD), University of Cologne, Cologne, Germany 3Department of Dermatology, University of Cologne, Cologne, Germany 4Helsinki Institute of Life Science, Biomedicum Helsinki, University of Helsinki, Helsinki, Finland 5Wihuri Research Institute, Biomedicum Helsinki, University of Helsinki, Helsinki, Finland 6 Stem Cells and Metabolism Research Program, Faculty of Medicine, University of Helsinki, Helsinki, Finland 7Department of Biomedical Engineering, University of Arkansas, Fayetteville, AR, USA 8Center for Molecular Medicine Cologne (CMMC), University of Cologne, Cologne, Germany 9Institute of Zoology, Developmental Biology Unit, University of Cologne, Cologne, Germany
# These authors contributed equally § These authors contributed equally
*To whom correspondence should be addressed: [email protected]
Lead contact: [email protected]
Manuscript
2
Summary
Stem cells reside in specialized niches that are critical for their function. Upon activation hair
follicle stem cells (HFSCs) exit their niche to generate the outer root sheath (ORS), but a
subset of ORS progeny returns to the niche to resume a SC state. Mechanisms of this fate
reversibility are unclear. We show that the ability of ORS cells to return to the SC state
requires suppression of a metabolic switch from glycolysis to oxidative phosphorylation and
glutamine metabolism that occurs during early HFSC lineage progression. HFSC fate
reversibility and glutamine metabolism are regulated by the mammalian target of rapamycin
complex 2 (mTORC2)-Akt signaling axis within the niche. Deletion of mTORC2 results in a
failure to re-establish the HFSC niche, defective hair follicle regeneration, and compromised
long-term maintenance of HFSCs. These findings highlight the importance of spatiotemporal
control of SC metabolic states in organ homeostasis.
Introduction
A defining property of stem cells (SCs) is their capacity for self-renewal and multilineage
differentiation. Cell-intrinsic networks cooperate with signals from the microenvironment to
fine-tune these two key properties of SCs to maintain tissue homeostasis (Blanpain and
Fuchs, 2014; Chacón-Martínez et al., 2018). Hair follicles are one of the few tissue
compartments that undergo natural regeneration in mammals. Hair follicles regenerate
through cyclical bouts of rest (telogen), growth (anagen), and degeneration (catagen). At the
start of the hair cycle, quiescent HFSCs residing in the so-called bulge niche are activated
to give rise to progenitor cells (also known as transit amplifying cells) which subsequently
differentiate into outer root sheath (ORS) cells to supply the cells needed for hair follicle
down-growth. These ORS cells have exited the bulge niche and lost the expression of the
key bulge HFSC marker CD34, but still express SOX9 and continue to proliferate during
early- to mid-anagen (Hsu et al., 2011). After the completion of the growth cycle, a subset
of ORS cells returns to the bulge niche, resume quiescence and CD34 expression, and
remain in this state until the next hair cycle (Hsu et al., 2011; Lay et al., 2016). The
establishment of this new quiescent HFSC state is associated with the generation of a new
bulge niche adjacent to the old one (Gonzales and Fuchs, 2017; Muller-Rover et al., 2001).
Thus, HFSC activity is tightly controlled in time and space, making the hair follicle an ideal
system to investigate how SC quiescence, activation, and differentiation are regulated, and
what are the factors that determine whether a progenitor/transit-amplifying cell will return to
the quiescent SC state or fully commits to terminal differentiation.
3
Differentiation of embryonic and hematopoietic SCs associates with a change in the
carbohydrate metabolic program, characterized by a transition from glycolytic metabolism
to oxidative phosphorylation (OXPHOS) (Ito and Ito, 2016; Shyh-Chang et al., 2013).
Generation of energy by anaerobic glycolysis is inefficient, yielding only 2 ATP molecules
per glucose molecule in contrast to the 36 generated by OXPHOS, but superoxide anions
are produced in this process, generating mitochondrial reactive oxygen species (ROS).
Excessive production of ROS can abrogate various SC properties including quiescence and
self-renewal, which is a major reason why many SCs prefer glycolysis as their energy source
(Ito and Ito, 2016).
In addition to ATP production, SCs utilize glucose-derived carbons for lipid synthesis and
for production of hexosamines, ribose, glycerol, serine and glycine. Excess glucose-derived
carbons can be secreted as lactate. Differentiation, on the other hand, is generally
associated with proliferation and increased metabolic requirements, explaining the need for
OXPHOS at this stage to maximize ATP production (Kobayashi and Suda, 2012; Miyamoto
et al., 2007). To increase the flux of the cycle for biosynthesis, differentiating SCs and cancer
cells have also been shown to utilize additional carbon sources, such as glutamine, which
can be converted into both Acetyl-CoA and TCA cycle intermediates (Jin et al., 2015;
Oburoglu et al., 2014). In contrast to the role of metabolic changes in SC differentiation and
cancer, the role of metabolism in SC fate reversibility and reprogramming is largely unknown.
An important master regulator of metabolism is the mammalian target of rapamycin (mTOR)
pathway. mTOR is a serine/threonine protein kinase that integrates growth factor receptor
signaling with cell growth, proliferation and survival through two distinct multi-protein
complexes, referred to as mTOR complex 1 and 2 (mTORC1 and mTORC2). Both
complexes contain mTOR, mLST8 and Deptor. Binding of mTOR to Raptor defines the
mTORC1 complex, whereas Rictor binding to mTOR generates the mTORC2 complex.
mTORC1 regulates protein translation and anabolic metabolism in response to amino acid
and nutrient levels and is an important regulator of cell growth (Albert and Hall, 2015;
Eltschinger and Loewith, 2016; Laplante and Sabatini, 2012). mTORC2 acts downstream of
growth factor signaling to activate Akt, Pkcα, and Sgk1 (Sarbassov et al., 2005), but its
functions in SC regulation and metabolism are less well understood.
4
Here we address the molecular mechanisms that define the tipping point between SC fate
reversibility and commitment to terminal differentiation. We show that HFSC lineage
progression towards the ORS is controlled by a metabolic switch to increased OXPHOS and
glutaminolysis, and the ability of ORS progenitors to return to the SC state depends on their
ability to suppress this metabolic switch. The metabolic reprogramming is regulated by
TORC2-Akt signaling-mediated suppression of glutaminase expression and glutamine
metabolism. Consequently, deletion of mTORC2 in the epidermis results in loss of HFSC
fate reversibility, defective hair follicle regeneration, and impaired long-term maintenance of
HFSCs. Collectively, our work demonstrates the remarkable metabolic flexibility of HFSCs,
which is required to maintain a stable SC population throughout the lifespan of mice.
Results
HFSCs and early progenitor cells have distinct metabolic profiles
To gain insight into the molecular mechanisms that govern HFSC fate, we utilized an ex vivo
skin organoid culture system (3C culture) that specifically promotes growth of CD34-positive
and CD34-negative cells, which based on transcriptome and marker expression represent
a mixture of HFSCs, hair follicle ORS progenitors and inner bulge cells (Chacón-Martínez
et al., 2016) (Fig. 1A; Supplementary Fig. S1A). Both ORS progenitors and the inner layer
of bulge cells (CD34-negative, Keratin 6-positive) represent HFSC progeny and act as niche
cells for bulge HFSCs (Hsu et al., 2014). Specifically, most of these CD34-negative cells in
3C cultures expressed the HFSC and ORS progenitor marker and master transcriptional
regulator Sox9, and ORS identity gene Barx2, whereas interfollicular epidermis markers
(Joost et al., 2020) were not enriched (Supplementary Fig. S1A, B). A subset of these CD34-
negative cells was also positive for the inner bulge marker Keratin 6 and the hair follicle-
specific marker Keratin 16 (Supplementary Fig. S1A) (Hsu et al., 2011; Nowak et al., 2008).
As Keratin 6 and 16 can also be induced upon wound healing (Joost et al., 2018), we
confirmed that no enrichment of other wound healing marker genes was observed in the
CD34-negative population (Supplementary Fig. S1C), indicating that Keratin 16 and 6
represented HF identity markers of the CD34-negative population. As expected based on
the HFSC-like transcriptional profile, the cultured and subsequently fluorescence-activated
cell sorting (FACS)-purified CD34+/6 integrin+ cells displayed higher proliferative potential
in a colony-forming assay than the cultured CD34-/6 integrin+ cells (Supplementary Fig.
S1D).
5
We have previously shown that these cultured CD34+/6 integrin+ HSFCs and CD34-/6
integrin+ ORS progenitors (from here on referred to as 3C-HFSCs and 3C-ORS progenitors,
to distinguish them from in vivo HFSC and ORS cells) co-exist in a 50:50 population
equilibrium driven by bidirectional, dynamic interconversion of the two states, thus providing
a perfect model system to study SC fate reversibility (Chacón-Martínez et al., 2016). By
analyzing the transcriptomes of FACS-purified 3C-CD34+/6 integrin+ and 3C-CD34-/6
integrin+ cells (Chacón-Martínez et al., 2016), we identified genes involved in mitochondrial
ATP synthesis and OXPHOS to be most significantly overrepresented in the CD34-negative
3C-ORS progenitor population (Fig. 1A, B). To investigate if these differential gene
expression patterns would reflect differences in the levels of mitochondrial respiration in the
two populations, we quantified the amounts of mitochondria in CD34-positive HFSCs and
CD34-negative progenitors using transgenic mice expressing a fluorescent marker for
mitochondria (mito-YFP; (Sterky et al., 2011), as well as by analyzing expression levels of
genes involved in mitochondrial biogenesis (Scarpulla et al., 2012). FACS analyses of both
freshly isolated in vivo HFSCs and 3C-HFSCs showed significantly less fluorescently
labeled mitochondria as well as lower expression levels of mitochondrial biogenesis genes
in HFSCs compared to progenitors (Fig. 1 C; Supplementary Fig. S1E). To establish if the
higher amounts of mitochondria in CD34-negative progenitors would reflect their increased
OXPHOS activity, we performed Seahorse metabolic flux measurements of OXPHOS and
glycolysis. In agreement with their greater mitochondrial mass, the flux analyses revealed
that 3C-progenitors had significantly higher rates of basal respiration as well as higher
respiratory capacity compared to 3C-HFSCs (Fig. 1D, E). In contrast, no major differences
were observed in glycolytic activity (Fig. 1F).
The activity of the TCA cycle is not only required for electron transport chain (ETC) function,
but it can also be used to provide carbon for biosynthetic reactions. As OXPHOS but not
aerobic glycolysis was increased in progenitors, we sought to analyze the status of the TCA
cycle by quantifying key metabolic intermediates using liquid chromatography mass
spectrometry (LC-MS). The analyses showed that metabolites from TCA cycle were
increased in in vivo CD34-/6+ progenitors (Fig. 1G). In contrast, metabolic intermediates
of glycolysis were not altered (Supplementary Fig 1F). Collectively these data indicate that
the progenitor state is associated with increased levels of OXPHOS and TCA cycle activity.
Anagen entry is associated with increased oxidation in upper hair follicle cells in vivo
6
Given that cell dissociation and FACS may cause metabolic changes in a metabolite-,
compartment-, or cell type-specific manner (Binek et al., 2019; Llufrio et al., 2018), we aimed
to directly quantify metabolic changes in HFSCs and their progeny during their activation
and subsequent lineage progression. For this, we utilized label-free multiphoton microscopy
that can be used to quantify autofluorescence from NADH and FAD and second harmonic
generation (SHG) from fibrillar collagen (Jones et al., 2018; Kolenc and Quinn, 2019). The
quantum yield of NADH and FAD autofluorescence is an order of magnitude lower than
typical exogenous fluorophores used to stain tissues. However, the relatively weak natural
fluorescence of NADH and FAD can be efficiently measured through multiphoton
microscopy to assess relative changes in metabolic activity (Kolenc and Quinn, 2019). An
optical redox ratio of [FAD/(NADH+FAD)] was computed from the NADH and FAD
fluorescence intensities within the HFSC bulge and ORS area in C57BL/6J mice at the
transition between postnatal day (P) 21 (telogen resting stage) and P23 (early anagen
growth phase). Decreases in the redox ratio of cells or tissues have been attributed to
increased glycolysis, whereas increases are associated with increased glutaminolysis and
OXPHOS (Georgakoudi and Quinn, 2012; Kolenc and Quinn, 2019; Varone et al., 2014). A
higher optical redox ratio was detected within the upper hair follicle bulge/ORS area in the
early anagen stage (P23) follicles relative to telogen stage (P21) hair follicles (Fig. 2A, B).
This significant increase in optical redox ratio over two days is indicative of increased
oxidation. The in vivo optical measurements during the phase of HFSC
activation/differentiation were consistent with the metabolic flux profiles of FACS-purified
progenitors compared to HFSCs (Fig. 1G) and were further corroborated by a measured
increase in the LC/MS-derived redox ratios of FAD/(NADH+FAD), ATP/ADP, ATP/AMP and
SAM/SAH in freshly isolated cells (Supplementary Fig. S2).
Low oxygen tension promotes the HFSC state
To understand if the metabolic switch towards increased OXPHOS upon anagen induction
would control HFSC fate, we probed if limiting pO2 would influence HFSC activation (i.e. the
switch from a quiescent to an actively cycling state) or lineage progression. To this end, we
performed 3C cultures in ambient air (20% O2) or in reduced oxygen (2% O2). FACS
analyses revealed that 14 d of 3C culture in 2% O2 significantly increased the proportion of
3C-CD34+/6+ cells within the cultures (Fig. 3A). qRT-PCR analysis of key HFSC fate
master regulators and identity genes showed that expression of these genes was
upregulated in the total cell population under reduced oxygen, whereas the expected
enrichment of these transcripts in FACS purified 3C-CD34-positive cells compared to 3C-
7
CD34-negative cells remained unchanged, confirming that the low pO2 -triggered increase
reflected an increase in 3C-HFSCs within the organoids (Fig. 3B, Supplementary Fig. S3A).
To test if the increased proportion of HFSCs in the cultures was caused by a pO2-dependent
activation of 3C-HFSCs, or alternatively a selective growth advantage of 3C-HFSCs
compared to progenitors in low pO2, we analyzed the growth and apoptosis rates of both
populations in 20% and 2% O2 but observed no differences (Supplementary Fig. S3B).
Absence of differential proliferation/survival suggested that 2% O2 did not control HFSC
activation, but rather could either prevent HFSC lineage progression to ORS or to promote
progenitor reversal to the HFSC state. To test this hypothesis, we performed lineage tracing
experiments where we FACS-purified CD34+/6+ 3C-HFSCs expressing GFP (LifeAct-GFP
(Riedl et al., 2010) with CD34-/6+ progenitors lacking GFP expression, mixed them 1:1 to
ensure comparable population composition of the cultures, and traced the conversion of 3C-
HFSCs to progenitors and vice versa (Fig. 3C). FACS analyses after 14 d of tracing revealed
that 2% O2 both enhanced return of progenitors to the HFSC state as well as reduced HFSC
lineage progression (Fig. 3D, Supplementary Fig. S3C).
HFSC lineage progression requires mitochondrial metabolism and glutaminolysis
To identify the molecular mechanisms by which low pO2 prevented HFSC lineage
progression and promoted progenitor return to the HFSC state, we examined the activation
of Hypoxia-inducible factor (Hif) pathway (Majmundar et al., 2010). To this end we subjected
3C organoid cultures to acute low pO2 (12 h) and analyzed Hif1 stabilization by western blot
and Hif1/2 target gene expression using qRT-PCR. As expected, Hif1 protein was
stabilized in 2% O2 (Supplementary Fig. S4A). Hif1/2 target gene analyses further
revealed that a subset of Hif targets (Vegf, Pgf, Glut1, Pai1) were mildly upregulated both in
3C-CD34-positive and 3C-CD34-negative cells, indicating that both populations respond to
low pO2 (Supplementary Fig. S4B). Intriguingly, however, the Hif1 target and regulator of
glycolysis Pdk2 was not induced, but it was expressed at higher levels in 3C-CD34-positive
cells compared to in 3C-CD34-negative cells, and the difference became slightly more
pronounced in low pO2. In addition, low pO2 suppressed expression of the Hif1/2 target
glutaminase (Gls), the amidohydrolase that generates glutamate from glutamine
(DeBerardinis and Cheng, 2010), specifically in 3C-CD34-positive cells (Supplementary Fig.
S4B), suggesting that the gene expression changes could not be attributed solely on Hif1
induction. Consistent with this notion, the stabilization of Hif1/2 using CoCl2 was not
8
sufficient to alter the population ratio of 3C-CD34-positive and 3C-CD34-negative cells
(Supplementary Fig. S4C).
We thus hypothesized that low pO2 could regulate HFSCs mainly through more complex,
possibly Hif-independent, effects on metabolism, and to test this we manipulated cellular
respiration pathways (Fig. 4A). To address if HFSC lineage progression required OXPHOS
and therefore the ETC, we blocked the ETC with two inhibitors, antimycin A and rotenone
(Boveris and Chance, 1973). Both inhibitors reduced the relative amount of CD34-negative
progenitors in the 3C cultures (Fig. 4B) without affecting the total numbers of live cells
(Supplementary Fig. S4D). To test if the HFSC state would depend on glucose oxidation we
interfered with pyruvate entry into the TCA cycle using dichloroacetate (DCA) and UK5099
(Hildyard et al., 2005; Stacpoole, 1989). Surprisingly, boosting glucose oxidation by DCA or
blocking it with UK5099 did not alter the 3C-HFSC-progenitor equilibrium (Fig. 4C,
Supplementary Fig. S4E). We further inhibited glycolysis using the non-metabolizable
glucose analogue 2-deoxyglucose (Wick et al., 1957) and by replacing glucose with
galactose that provides no net ATP generation through oxidation, forcing cells to rely
increasingly on OXPHOS for ATP production (Supplementary Fig. S4F, G). Strikingly, none
of these manipulations had a significant effect on the 3C-HFSC-progenitor equilibrium (Fig.
4C; Supplementary Fig. 4E-G). The lack of effect of blocking glycolysis on HFSC lineage
progression was unexpected, as previous studies had indicated that HFSC activation,
requires anaerobic glycolysis and lactate production (Flores et al., 2017). To assess the role
of anaerobic glycolysis and lactate on 3C-HFSC proliferation in our system, we inhibited
lactase dehydrogenase using FX11 (Flores et al., 2017), and observed that inhibiting lactate
production attenuated proliferation of both HFSCs and ORS progenitors (Supplementary Fig.
S4H), but did not affect the proportions of the two populations (Fig. 4D). This indicated that,
consistent with Flores et al., anaerobic glycolysis is required for HFSC activation, but it does
not influence HFSC to ORS lineage progression.
Glutamine has been shown to provide an alternative carbon source to glucose in cancer
cells and certain SCs (Jin et al., 2015; Oburoglu et al., 2014), and its increased consumption
is associated with increased optical redox ratio (Varone et al., 2014). As blocking glucose
entry to the TCA cycle had no major effect on HFSC fate, we asked if progenitor lineage
progression could instead require glutaminolysis. Supporting this idea, LC-MS analyses
showed higher levels of glutamine and glutamate in in vivo FACS-purified progenitors
9
compared to HFSCs (Supplementary Fig. S4I). Intriguingly, blocking glutaminase using
CB839 (Gross et al., 2014) resulted in an increase in the relative amounts of 3C-HFSCs in
organoid cultures, indicative of reduced lineage progression (Fig. 4E). Further, consistent
with the requirement for glutamate to fuel the TCA cycle to promote 3C-HFSC differentiation,
LC-MS analyses of 13C-labeled glutamine confirmed decreased entry of glutamine into the
TCA cycle in 3C-HFSCs under 2% O2 conditions where lineage progression was attenuated.
In contrast, no difference in entry of glycolytic metabolites was observed with 13C-glucose
tracing (Fig. 4F). The analyses further confirmed the pronounced differences in the
metabolic profiles of CD34-positive 3C-HFSCs and CD34-negative progenitors, with
progenitors displaying increased glutamine metabolism and absent metabolic response to
low pO2 (Fig. 4F). Importantly, the reduced lineage progression induced by the 2% O2 culture
conditions could be fully rescued by supplementing the TCA cycle with exogenous,
membrane permeable dimethyl--ketoglutarate, the direct downstream metabolite of
glutamate (Fig. 4G). Collectively, these data indicate that the HFSC state is metabolically
flexible and does not depend on glycolysis or TCA cycle. In contrast, HFSC lineage
progression to progenitors requires OXPHOS and glutamate entry into the TCA cycle.
mTORC2 controls HFSC metabolism and ORS cell fate reversibility through Akt
To identify the mechanisms that control HFSC metabolism and fate during lineage
progression, we returned to the RNAseq analyses to identify upstream regulators that could
explain the differential expression of metabolic genes in 3C-HFSCs and progenitors.
Interestingly, pathway analysis predicted Rictor, the defining component of mTORC2, as a
potential upstream regulator of the observed gene expression changes (Fig. 5A). To assess
the role of mTORC2 in HFSC lineage progression and fate reversibility, we isolated
epidermal cells from epidermis-specific Rictor knockout mice (RicEKO) (Ding et al., 2016) and
subjected them to 3C organoid culture. Intriguingly, RicEKO cells showed reduced 3C-HFSC
content in organoid cultures compared to control cells, and failed to respond to low pO2 with
an increase in 3C-HFSCs (Fig. 5B). The defect in HFSC fate dynamics was confirmed by
qRT-PCR analysis of key HFSC identity genes which were low in RicEKO organoids and were
not further upregulated in response to low pO2 (Supplementary Fig. 5A). Importantly, the
overall cell composition within the 3C cultures was comparable between RicEKO and controls,
both cultures consisting of Sox9-positive/CD34-positive 3C-HFSCs and Sox9-
positive/CD34-negative ORS progenitors (Supplementary Fig. 5B).
10
To assess if mTORC2 signaling was required to prevent excessive HFSC lineage
progression or to promote progenitor return to HFSC state, we FACS-purified HFSCs and
progenitor populations from control and RicEKO mice and placed them into 3C cultures
separately. Notably, RicEKO organoids showed comparable rates of 3C-HFSC lineage
progression to control cells, but the return of RicEKO progenitors to the CD34+/6+ HFSC
state was severely compromised (Fig. 5C; Supplementary Fig. S5C), indicating that
mTORC2 controls HFSC fate reversibility.
To address the molecular mechanism by which mTORC2 impacts HFSC fate reversibility
and the progenitor cell state, we analyzed the activation status of the key downstream target
of mTORC2, Akt-pS473 (Sarbassov et al., 2005), and observed its phosphorylation to be
significantly attenuated in RicEKO cells (Fig. 5D). The mTORC1 downstream target pS6 was
not altered, confirming the specific effect of Rictor deletion on mTORC2 activity (Fig. 5D).
Consistent with the reduced levels of Akt-pS473 in RicEKO cells, global inhibition of Akt in 3C
cultures phenocopied RicEKO cells in their inability to maintain a 50-50 population equilibrium
of 3C-HFSCs and progenitors, and to respond to low pO2 by promoting the HFSC state (Fig.
5E). Inhibition of Akt in Rictor-deficient cells did not induce an additive effect (Fig. 5E), further
confirming that Akt acts downstream of mTORC2. Interestingly, inhibition of mTORC1 using
rapamycin (Laplante and Sabatini, 2012; Loewith and Hall, 2011) did not affect 3C-HFSC
proportions (Supplementary Fig. S5D), indicating that the role of mTOR on HFSC to ORS
early fate reversibility is specific to mTORC2.
Finally, consistent with requirement for Akt downstream of hypoxia to promote the HFSC
state, we observed that 2% pO2 promoted Akt phosphorylation, particularly in the 3C-CD34-
negative progenitors (Fig. 5F). Hypoxia-dependent Akt activation was Rictor-dependent, as
shown by absence of pAkt induction in RicEKO cells in 2% pO2 (Fig. 5G). Further, activation
of Akt using SC79 (Jo et al., 2012) was sufficient to increase levels of 3C-HFSCs in 20% O2
conditions (Supplementary Fig. S5E). Collectively, the data indicate that the mTORC2/Akt
signaling axis is required downstream of low pO2 for the ability of progenitors to return to the
HFSC state.
mTORC2-Akt signaling axis regulates glutaminolysis and progenitor fate reversibility
As mTORC2-Akt signaling was required for the response of progenitors to low pO2, we next
sought to analyze the metabolic state of mTORC2-deficient cells. LC-MS quantification of
11
metabolites from cells in organoid cultures revealed that, as expected, lowering pO2 led to
accumulation of Acetyl-CoA and reduction of downstream TCA cycle metabolites (Fig. 6A).
This metabolic reprogramming depended on Akt activity, as Akt inhibition in low pO2 induced
an opposite effect of increased concentration of TCA metabolites (Fig. 6A). Furthermore,
whereas low pO2 led to reduced levels of glutamate, Akt inhibition increased the glutamate
concentration (Fig. 6B). Interestingly, inhibition of Akt had no substantial effects on
metabolites in high pO2 conditions (Fig. 6A, B).
To understand how low pO2 and Akt signaling could control the TCA cycle and in particular
glutamine metabolism, we analyzed the expression of Gls. Interestingly, whereas low pO2
suppressed Gls expression, inhibition of Akt counteracted this effect (Fig. 6C). The same
effect was observed in RicEKO cells (Fig. 6D).
Collectively these data suggested that low pO2 suppresses Gls expression in an mTORC2-
Akt-dependent manner to reduce glutamine hydroxylation to prevent the flux of glutamate
into the TCA cycle, and that this metabolic reprogramming is required to facilitate progenitor
return to the HFSC state. To challenge this prediction, we assessed if blocking glutamine-
to-glutamate hydroxylation would rescue the inability of RicEKO progenitors to return to the
HFSC state. Indeed, treating Rictor-deficient cells with the Gls inhibitor restored the ability
of RicEKO progenitor cells to convert into 3C-HFSCs (Fig. 6E). This effect on 3C-HFSC fate
was independent of cell proliferation or survival, as the Gls inhibitor had no major effects on
RicEKO progenitor proliferation or death (Supplementary Fig. S6A, B). Gls inhibition seemed
to mainly target 3C-ORS progenitors as 3C cultures from purified CD34-positive cells
showed a less pronounced increase in HFCS content upon Gls inhibition (Supplementary
Fig. S6C). Taken together these data indicate that 3C-HFSC lineage progression into the
progenitor state requires flux of glutamate into the TCA cycle to maintain the ETC, and that
this metabolic pathway needs to be suppressed by mTORC2-Akt-dependent
downregulation of Gls expression for progenitors to return to the HFSC state.
mTORC2 is required for HFSC fate reversibility and long-term maintenance in vivo
As the in vitro data so far assigned a critical role for mTORC2 signaling in metabolic
regulation and HFSC fate reversibility, we proceeded to assess the consequences of these
processes in vivo in the hair follicle (Fig. 7A). Newborn and early adult RicEKO mice are
characterized by a mild barrier defect and epidermal atrophy (Ding et al., 2016; Ding et al.,
2019), whereas hair follicle morphogenesis did not display obvious impairment in mutants
12
(Fig. 7B, C). In addition, careful inspection revealed overall normal hair follicle architecture
(Supplementary Fig. S7A). However, entry into the first postnatal telogen at P21 was
delayed in RicEKO mice (Fig. 7B, C). This was followed by later entry into anagen and again
a delayed entry into telogen in the following cycles (Fig. 7B, C), indicative of impaired HFSC
regulation.
More detailed analyses of the telogen hair follicles in RicEKO mice revealed striking absence
of the secondary bulge that is formed by the ORS progenitors when they return to
quiescence (Fig. 7D, E, Supplementary Fig. S7B) (Hsu et al., 2011). To assess whether the
absence of the second bulge was due to a club hair retention defect that can result in loss
of the first bulge (Higgins et al., 2009; Lay et al., 2016), we traced the fate of club hair through
a full hair cycle by staining the hair coat with a UV-visible dye. RicEKO showed comparable,
or even increased hair dye retention compared to controls, excluding a club hair retention
defect (Supplementary Fig. S7C). To assess if RicEKO were instead unable to generate a
new bulge, we performed lineage tracing analyses, where we labeled ORS cells by
administering BrdU in late anagen (P34 for control, P48 for RicEKO), when bulge HFSCs
have ceased proliferating (Hsu et al., 2011), and traced the fate of the label-retaining BrdU-
positive cells until the next telogen. Presence of BrdU-positive cells in the bulge niche
indicates that they are derived from ORS cells from previous anagen (Hsu et al., 2011; Lay
et al., 2016). As expected, no difference in initial BrdU label incorporation was observed
directly post BrdU administration in early anagen (Supplementary Fig. S7D), indicating
normal progenitor proliferation. Also, no defects in ORS identity, as marked by Barx2 and
Sox9, or abnormal apoptosis in catagen were observed (Supplementary Fig. S7E-G). In
contrast, when BrdU-labeled cells were traced to the following telogen phase, very few
labeled cells were found in RicEKO mouse bulges, whereas control mice showed significant
amounts of labeled cells in the new bulge (Fig. 7F). Importantly, and as expected, the old
bulge was negative for BrdU (Fig. 7F). This indicated that RicEKO ORS cells are unable to
form a new HFSC bulge niche.
To exclude that the hair follicle cycling defects and the lack of second bulge formation could
be indirect consequences of potential defects outside the hair follicle, we generated a HFSC-
specific deletion of Rictor using the bulge-specific Keratin19-Cre-ERT (Means et al., 2008).
As reported by others (Vagnozzi et al., 2015) Keratin 19-Cre-ERT-mediated Rictor deletion
was not complete (Supplementary Fig. S7H), but it was sufficient to trigger delayed entry
13
into telogen and subsequent compromised formation of the second bulge similar to the
Keratin 14-Cre deletion (Supplementary Fig. S7I-L), confirming that the phenotype of Rictor-
deficient hair follicles is HFSC-autonomous.
To test if the inability to form a second bulge was indeed due to the inability of ORS cells to
switch off glutaminolysis and the TCA cycle, we hypothesized that inhibition of Gls in RicEKO
mice should be sufficient to rescue this phenotype. We injected RicEKO mice intradermally
with the Gls inhibitor BPTES (Robinson et al., 2007) in mid-anagen and address timing of
telogen entry and bulge number. Intriguingly, inhibition of Gls enhanced formation of a
second bulge in RicEKO mice (Fig. 7G).
Consistent with the role of mTORC2-Akt in maintaining HFSC fate reversibility during HFSC
activation, Akt-pS473 levels were very low in telogen hair follicles, whereas a substantial
increase was observed in anagen specifically at the upper ORS region right below the bulge
HFSC niche (Fig. 7H). Further, and as predicted from the ability of low pO2 to trigger Akt-
pS473 in organoid CD34-/Sox9+ ORS progenitors, this increase in Akt phosphorylation in
anagen ORS coincided with pronounced hypoxia at the anagen-stage upper ORS
(Supplementary Fig. S7M). Akt activity remained high in catagen hair follicles, but it was
present only below and within the new and not the old bulge. As expected, Akt activity was
decreased in all hair follicle stages in RicEKO mice (Fig. 7H).
Finally, although the young RicEKO mice did not show a reduction in hair follicle density or
HFSCs (Supplementary Fig. S7N, O), aging of RicEKO mice was accompanied by hair loss
(Fig. 7I), reduced hair follicle density (Fig. 7I) and reduction in HFSCs (Fig. 7J). Collectively,
these data indicate that mTORC2-Akt signaling is required for timely ORS progenitor return
to the HFSC state within the hypoxic upper ORS to establish the new bulge SC niche. The
failure to do so results in age-induced reduction of HFSCs, leading to reduction in hair
follicles and thinning of the hair coat.
Discussion
Changes in metabolic pathways actively influence SC differentiation in multiple tissues (Ito
and Ito, 2016; Shyh-Chang et al., 2013). Yet, little is known about the tenability of this
metabolic reprogramming and its role in SC fate reversibility. Our current data demonstrate
that the transition of HFSCs to the ORS progenitor state involves activation of OXPHOS and
14
flux of glutamine into the TCA cycle. This metabolic switch to OXPHOS/glutamine
metabolism needs to be suppressed to allow progenitors return to the bulge SC state at the
end of the anagen growth phase, identifying a previously unknown role for metabolic
flexibility in SC fate decisions. We further unravel a role for localized TORC2-Akt signaling,
through repressing Gls expression and thus glutamine metabolism within the bulge niche
region, for metabolic reprogramming required for ORS progenitor-mediated generation of
the new bulge niche. Consequently, the failure to locally activate mTORC2-Akt signaling
results in loss of SC fate reversibility and the ability of HFSCs to regenerate their bulge niche
after the hair cycle, leading to HFSC exhaustion and hair loss.
Various SCs reside in hypoxic niches and hypoxia has consequently been shown to be
important for the maintenance of hematopoietic, mesenchymal and neural SCs (Ito and Ito,
2016). Also the hair follicle niche is mildly hypoxic; whereas the epidermis has a partial O2
pressure (pO2) between 0.2-8% and the lower dermis reaches 10%, the pO2 of the hair
follicle is only 0.1-0.8 % (Bedogni et al., 2005; Evans and Naylor, 1967; Peyssonnaux et al.,
2008). Our data further identifies the anagen upper ORS as particularly hypoxic, which could
be a result from the expansion of the hypoxic hair follicle area during anagen. It was recently
shown that the anaerobic glycolysis, and in particular lactate production through lactate
dehydrogenase is critical for HFSC activation to proliferate at the onset of the anagen growth
phase (Flores et al., 2017, Miranda et al., 2018), further consistent with the hypoxic state of
the HFSC niche (Rezvani et al., 2011). In agreement with this, we observe that low pO2
conditions promote maintenance of the HFSC state, whereas HFSC differentiation is
associated with increased OXPHOS. Consistent with the Flores et al (2017) study, we
observe that anaerobic glycolysis and lactate are required for HFSC and ORS progenitor
proliferation to initiate the hair follicle anagen growth phase. However, glycolysis does not
regulate the lineage progression of HFSCs to ORS progenitors or the return of ORS cells to
the HFSC state. In fact, these cell states show similar levels of glycolysis, whereas Flores
et al (2017) showed that HFSCs show increased levels of glycolytic metabolism when
compared to the total epidermis (including terminally differentiated cells). This indicates that
both HFSCs and the early ORS progenitor state cells likely have relatively high levels of
glycolysis compared to the rest of the epidermis, and that this is critical for their proliferation,
whereas the role of OXPHOS in lineage progression is independent from proliferation and
primarily regulates the progenitor-to-HFSC conversion.
15
Interestingly, due to the lack of change in glycolysis, the increase in TCA cycle flux is
facilitated by increased glutamine metabolism. Glutamine can contribute to essentially every
core metabolic task of proliferating cells. It participates in bioenergetics, supports cell
defences against oxidative stress and can also replace glucose in the production of
macromolecules (DeBerardinis and Cheng, 2010). We observe that the increase in
glutaminolysis is regulated by the mTORC2-Akt signalling axis which, like hypoxia,
suppressed the expression of Gls. Our observation is consistent with previous cell culture
studies showing increased glutamine consumption in Rictor-deficient keratinocytes
(Tassone et al., 2017). Notably, it has previously been shown that cancer cell proliferation
under hypoxic conditions relies on mitochondrial glutamine metabolism (Sun and Denko,
2014). Thus, it is feasible to hypothesize that within the hypoxic niche, whereas lactate is
required for the early activation and proliferation of HFSCs, glutamine could provide the
necessary carbon source to facilitate efficient differentiation of HFSC immediate progeny,
similarly to what has very recently been reported in chondrocytes (Stegen et al., 2019).
Consequently, at the end of the anagen growth phase returning to the hypoxic niche, ORS
progenitors would need to reprogram their metabolism by locally activating mTORC2-Akt
signalling to suppress Gls expression facilitating the switch to back to the quiescent HFSC
state. This is consistent with the observed pattern of Akt activation that occurs specifically
within the upper ORS directly adjacent to the bulge niche and is temporally restricted to the
anagen-catagen phase. Interestingly, the HFSC-ORS lineage progression was rapamycin
insensitive. This is consistent with previous studies showing that mTORC1 activity is
restricted to the inner root sheath of anagen hair follicles (Castilho et al., 2009). Thus,
whereas mTORC2 regulates early HFSC fate determination, mTORC1 activity is critical for
hair follicle differentiation (Castilho et al., 2009).
In summary, the present study demonstrates the functional importance of precise
spatiotemporal control of metabolic flexibility to facilitate SC fate reversibility that is required
for the maintenance of a stable SC population for the lifetime of the organism.
Limitations of Study
One limitation of our study is the lack of methodology to trace and quantify metabolic fluxes
in vivo in the cycling hair follicle. The organoid model for hair follicle stem cells used here
recapitulates key early features but not all the steps of hair follicle differentiation. In addition,
16
a genetic model directly targeting glutamine metabolism would further clarify the role of this
pathway in hair follicle stem cell fate.
Author contributions
SAW conceived and supervised the study. CSK designed and performed most of the
experiments and analyzed data. XD and SAE developed the mouse models and
conceptualized, performed and analyzed most of the in vivo experiments. KA, NLH, LCB,
and CAC-M designed and performed cell culture experiments. OK and KPQ designed,
performed and analyzed the multiphoton redox ratio imaging experiments. CE and PG
designed and performed the metabolite analyses. MD conceived experiments and analyzed
data. SAW wrote the paper. All authors commented and edited the manuscript.
Acknowledgements
We thank Yekaterina A. Miroshnikova for discussions and help with experiments, Anu M.
Luoto for technical assistance, Anu Suomalainen and Carien Niessen for critical reading of
the manuscript, Nils-Göran Larsson (Karolinska Institute, Sweden) for Mito-YFP mice,
Roland Wedlich-Söldner (University of Münster, Germany) for LifeAct mice, Guoqiang Gu
(Vanderbilt University, USA) for Keratin 19-Cre ERT mice, Michael Hall and Markus Rüegg
(Biozentrum, University of Basel, Switzerland) for Rictor flox mice, and Catherin Niemann
and Anna Geueke (University of Cologne, Germany) for support with the Ric-K19-CreERT
mouse model. We thank the FACS & Imaging and Comparative Biology core facilities at MPI
for Biology of Ageing for support. This work was supported by the Max Planck Society (to
SAW), the Deutsche Forschungsgemeinschaft (DFG): (project number 73111208 - SFB 829
to MD, SAE, and SAW), the Cluster of Excellence CECAD (to MD, SAE, and SAW); the
Research Unit FOR2599, Research Unit FOR2240, Center for Molecular Medicine and the
DEBRA International Foundation (all to SAE); Helsinki Institute of Life Science, Wihuri
Research Institute, Jane and Aatos Erkko Foundation, European Research Council (ERC)
under the European Union's Horizon 2020 research and innovation programme (grant
agreement 770877) (all to SAW).
Declaration of interests
S.A.W. and C.A.C.M disclose that they are inventors of a patent application
PCT/EP2016/073675 (WO 2017/060240 (A1)) for the 3C Culture method of epidermal stem
cells. All other authors have no competing interests.
17
Main Figure Titles and Legends
Fig. 1. HFSCs and early progenitor cells have distinct metabolic profiles
(A) Schematic of the pilosebaceous unit and cell surface markers used to identify specific
sub-populations: interfollicular epidermis (IFE; Sca1+/CD34-/6 integrin+; in blue), outer
bulge hair follicle stem cells (HFSC; Sca1-/CD34+/6 integrin+; in orange) and outer root
sheath (ORS; Sca1-/CD34-/6 integrin+; in magenta). Epidermal cells were isolated and
placed in 3C organoid culture, after which 3C-CD34+/6 integrin+ and 3C-CD34-/6+ cells
were purified and analyzed.
(B) GO-term analyses from RNAseq data from FACS-purified 3C-CD34+/- cells.
(C) Normalized mean intensity of mito-YFP mice from freshly isolated (left) and 3C cultured
HFSCs and progenitors (n=4-6 mice; *P < 0.05, **P< 0.01, Mann-Whitney).
(D) Representative quantifications of real-time cellular oxygen consumption rates (ORC) (left)
and extracellular acidification rates (ECAR) (right) of FACS-purified 3C-CD34+/- cells.
(E) Quantification of basal respiration (left) and spare respiratory capacity (right) in FACS-
purified 3C-CD34+/- cells (n=3 mice; *P< 0.05, Ratio paired t-test).
(F) Quantification of glycolysis level, glycolysis capacity, and glycolysis reserve in FACS-
purified 3C-CD34+/- cells (n=3 mice; ns=not significant).
(G) Relative amount of key TCA cycle metabolites quantified by LC-MS from freshly isolated
CD34+/-; 6 integrin+ cells, normalized to CD34+ cells (mean SD; n=3-4 mice; *P< 0.05,
Mann-Whitney).
Fig. 2. Anagen entry is associated with increased oxidation in upper hair follicle cells
in vivo
(A) Schematic illustration of the pilosebaceous unit containing the interfollicular epidermis
(IFE), sebaceous gland (SG) and the hair follicle with the outer bulge HFSCs (in orange),
inner bulge niche cells (in yellow) and hair germ/outer root sheath (ORS) located below the
SG and above the dermal papilla (DP). HS=hair shaft; ROI=region of interest used for
measurements.
(B) Quantification of redox ratios from HFSC bulge and ORS areas (dashed lines in A)
(minimum to maximum box and whiskers plot; ***P=0.0135 Student’s t-test ;n=9 follicles
from 2 mice (anagen), n=20 follicles from 2 mice (telogen).
(C) Representative in vivo images of hair follicles. Top row: Two photon excited
autofluorescence from NADH (green) and FAD (blue) and collagen SHG (red) to identify
18
and segment the HFSC bulge area. Bottom row: Optical redox ratio maps of
FAD/(NADH+FAD) autofluorescence. Dashed line shows HFSC bulge. Scale bar 50 µm.
Fig. 3. Low oxygen tension promotes the 3C-HFSCs state
(A) FACS analysis of 3C-CD34+ cells (%) after 14 d of culture in 20% or 2% O2 (n=4
mice/condition; *P< 0.05, Mann Whitney).
(B) qRT-PCR analysis of key HFSC signature genes from 3C cultures in 20% or 2% O2
(mean SD, n=3 mice/condition; *P< 0.05, Mann-Whitney).
(C) Workflow of the lineage tracing experiments. GFP+ CD34+/6+ cells and GFP- CD34-
/6+ cells were FACS-purified, mixed 1:1 and 3C cultured under 20% or 2% O2 for 14 days.
(D) FACS analyses of GFP – cells (left panel) and GFP + cells (right panel) shows increased
conversion of GFP- cells into CD34+ and decreased conversion of GFP+ cells into CD34-
(mean SD n=3 mice/condition; *P< 0.05, Mann Whitney).
Fig. 4. HFSC differentiation requires mitochondrial metabolism and glutaminolysis
(A) Schematics summarizing flux of metabolites into the TCA cycle and the targets of the
inhibitors used.
(B) FACS analyses of HFSC content from 10-day 3C cultured cells treated with 1 nM
Antimycin A (AA) or 10 nM Rotenone (Rot) for 2 days (n= 5 mice (AA), 3 mice (Rot); *P<
0.05, **P< 0.01, Mann Whitney).
(C) FACS analyses of HFSC content from 3C cultured cells treated with 1 µM
dichloroacetate (DCA) (n= 3 mice/condition).
(D) FACS analyses of HFSC content from 3C cultured cells treated with 50 µM FX11 for 2
days (n= 3 mice).
(E) FACS analyses of HFSC content from 10-day 3C cultured cells treated with 20 nM
CB839 glutaminase inhibitor for 2 days (n=4 mice; *P< 0.05, Mann Whitney).
(F) LC/MS quantification of unlabeled (M0) and 13C-labeled (M4/M5) metabolite ratios from
12-day 20% or 2% O2 cultured cells treated with 13C-labeled glucose (pyruvate) or 13C-
glutamine (all other metabolites) for 1 h prior to sample collection and subsequent FACS-
purification of CD34+/- cells. (meanSD; n = 4 mice/condition; *P<0.05, **P<0.001,
***P<0.0007, ****P< 0.0001; ANOVA/Fisher’s LSD).
19
(G) FACS analyses of HFSC content from 10-day 3C cultured cells supplemented with 200
µM dimethyl-α-ketoglutarate for 2 days (n=6 mice (20% O2)/4 mice (2% O2); *P=0.049,
Kruskal-Wallis/Dunn’s; ns=not significant).
Fig. 5. mTORC2 controls HFSC metabolism and fate reversibility through Akt.
(A) Ingenuity pathway analysis from RNAseq data predicts Rictor as potential upstream
regulator of gene expression changes in CD34-/6+ cells.
(B) FACS analyses of 3C-HFSC content from cells isolated from Rictor-deficient mice
(RicEKO) and their control littermates (CTR) cultured in 3C under 20% or 2% O2 for 14 days
(n=6 mice/genotype; ***P= 0.0002, ****P< 0.0001, ANOVA/Dunnett’s).
(C) FACS analyses of HFSC content from FACS-purified CD34+/- cells from RicEKO and
CTR mice cultured for 3 days under 20% O2 (n=4 mice (CD34+)/3 mice (CD34-); *P< 0.05,
Mann-Whitney test).
(D) Representative western blots of pAkt and pS6K from RicEKO and CTR keratinocytes.
Quantification represents pAkt/total Akt ratio (n=2 mice/genotype from 4/4 analyzed). (E)
FACS analyses of HFSC content from 12-day 3C cultured RicEKO and CTR cells treated with
10 µM Akt inhibitor for 2 days (n=6 mice/genotype; ***P< 0.0007, ****P< 0.0001, all statistical
comparisons are to CTR 20% O2 and only significant differences are indicated; Sidak’s
multiple comparisons).
(F) Representative images and quantification of pAkt in 3C organoids grown in 20% or 2%
O2. Note increased pAkt-S473 in response low p02, particularly in 3C-CD34- cells (95%
confidence interval box-and-whiskers; n>17 organoids/condition pooled across 4 mice;
**P=0.0098, ***P=0.0005, ANOVA/Fisher’s; scale bars 70 µm).
(G) Representative western blots of pAkt from RicEKO and CTR keratinocytes exposed to 5h
of 20% or 2% O2. Quantification represents pAkt/total Akt ratio (n=2 mice/genotype from 4/4
analyzed).
Fig. 6. mTORC2-Akt signaling axis regulates glutaminolysis and progenitor fate
reversibility
(A, B) Cells were cultured in 3C under 20% or 2% O2 for 12 d after which 10µM Akt inhibitor
was added. 2 days after application of inhibitor key TCA metabolites (A) and glutamate (B)
were quantified by LC-MS (n = 3 mice/condition; * P< 0.05, Student’s t-test).
20
(C) qRT-PCR analysis of Gls expression from 14-day 3C cultured cells under 20% or 2% O2
with or without Akt inhibitor (n=3 mice/condition; *** P< 0.001, ****P< 0.0001, ANOVA/Holm-
Sidak).
(D) qRT-PCR analysis of Gls expression from 14-day 3C cultured RicEKO and control (CTR)
cells under 20% or 2% O2 (n=3 mice/condition; *P< 0.05, ***P< 0.001, ANOVA/Holm-Sidak).
(E) FACS-purified CD34- cells from RicEKO and CTR mice cultured in 3C for 3 days with or
without 20 nM CB839 glutaminase inhibitor after which CD34+ cells (%) were analyzed via
FACS (n=3 mice/condition; *P< 0.05, **P< 0.01, Ratio paired t-test).
Fig. 7. mTORC2 is required for HFSC fate reversibility and long-term maintenance in
vivo
(A) Schematic illustration of hair follicle (HF) cycling in mice. HFs undergo cycles of telogen
(rest), anagen (regeneration), and catagen (degeneration) throughout the lifetime of the
animal. Each HF generates a new bulge and new club hair with every anagen.
(B) H/E stainings of control (CTR) and RicEKO skin at various hair cycle stages. Scale bars
200 µm.
(C) Quantification of hair follicle (HF) cycling (mean SEM; n=3-4 mice/genotype/timepoint).
(D, E) Representative H&E staining images of CTR and RicEKO mice in 2nd telogen (D) and
quantification (E) of hair follicles with more than two bulges (red arrows) (n = 5
mice/genotype; **P< 0.01, Mann Whitney, scale bars 100 µm).
(F) Experimental timeline, representative immunofluorescence images and quantification of
lineage tracing experiments where proliferative ORS cells were labeled with a BrdU pulse in
late anagen after which BrdU labeled cells were traced during the following telogen phase.
(n = 4 mice/genotype; *P< 0.05, Mann Whitney; scale bars 100 µm).
(G) Representative H&E staining images and quantification of telogen bulge number of
RicEKO mice treated with BPTES to inhibit glutaminolysis during anagen (n = 4 mice/condition;
*P< 0.05, Mann Whitney; scale bars 100 µm).
(H) Representative immunofluorescence images and quantification of telogen, anagen and
catagen hair follicles shows no pAkt (magenta) in telogen hair follicle bulges (CD34; cyan),
whereas anagen whereas anagen and catagen upper ORS shows high activity in CTR. pAkt
is decreased in telogen, anagen, and catagen skin of RicEKO. Unspecific immunoreactivity
of the hair shaft is marked with an asterisk (n = 4 mice/genotype/time point; *P< 0.05, Mann
Whitney; scale bars 100 µm).
21
(I) CTR and RicEKO mice at 24 months age. Red arrow points to areas of sparse hair (right).
Quantifications of hair follicle (HF) density of 20 months-old mice (left) (n= 4 mice/genotype,
*P= 0.0286, Mann Whitney).
(J) Representative hair follicle bulges and quantification of CD34+ cells from 20 months-old
mice (n= 5 mice/genotype, *P=0.027 Student’s t-test; scale bars 100 µm).
STAR METHODS
RESOURCE AVAILABILITY
Lead contact
Further information and requests for resources and reagents should be directed to and will
be fulfilled by the Lead Contact, Sara A. Wickström ([email protected]).
Materials Availability
This study did not generate new unique reagents.
Data and Code Availability
Datasets and source data are available from the corresponding author upon request.
EXPERIMENTAL MODEL AND SUBJECT DETAILS
Primary cells and cell lines
Epidermal cells were isolated from telogen-stage (P65-69) C56BL/6 mice, LifeAct-GFP
(Riedl et al., 2010), or Rictor-deficient mice (Ding et al., 2016) and cultured in 3C conditions
(Chacón-Martínez et al., 2016). For this, epidermal single cell suspensions were generated
by incubating skin pieces in 0.8% trypsin for 50 min. Cells were embedded in a 1:1 mixture
of growth factor-reduced Matrigel and 3C medium (MEM Spinner’s modification (Sigma), 5
µg/mL insulin (Sigma), 10 ng/mL EGF (Sigma), 10 µg/mL transferrin (Sigma), 10 µM
phosphoethanolamine (Sigma), 10 µM ethanolamine (Sigma), 0.36 µg/mL hydrocortisone
(Calbiochem), 2mM glutamine (Gibco), 100 U/mL penicillin and 100 µg/mL streptomycin
(Gibco), and 10% chelated fetal calf serum (Gibco), 5 µM Y27632, 20 ng/mL mouse
recombinant VEGF, 20 ng/mL human recombinant FGF-2 (all from Miltenyi Biotec)). Cells
were cultured in 5% CO2 at 37 C unless otherwise indicated. For hypoxia experiments cells
were cultured in a hypoxia chamber (Hypoxic Glove Box, Cat# 8375250, CoyLabs) in 2%
O2 at 37 C. No cell lines were used in this study.
Mouse models
22
Wild type male C57Bl6/J mice, as well as Mito-YFP (Sterky et al., 2011) and LifeAct (Riedl
et al., 2010) mice, have been described earlier and were used for cell isolation and for 3C-
HFSC organoids. Rictor was deleted in the epidermis as previously described (Ding et al.,
2016). Briefly, mice carrying a loxP-flanked Rictor allele were crossed with a transgenic
mouse line expressing Cre recombinase under the control of the human Keratin 14 promoter
(Bentzinger et al., 2008; Hafner et al., 2004), leading to epidermis-specific deletion of rictor
(RicEKO) in the entire epidermis. Rictor LoxP/wt-Krt14Cre+, Rictor LoxP/LoxP and Rictor
LoxP/wt mice were phenotypically indistinguishable from each other and were all used as
controls. To generate a deletion specifically in bulge HFSCs, mice carrying a loxP-flanked
Rictor allele were crossed with a mouse carrying a Cre recombinase fused to a mouse
estrogen receptor ligand binding domain that is inserted upstream of the Keratin 19 (Krt19)
gene (K19-CreERT) (Means et al., 2008). To activate K19-CreERT, mice were fed with
tamoxifen diet (TAM400/CreER, TD.55125, Envigo) for 3 weeks starting from P26, the
beginning of anagen. Tissue samples for histology were collected at various time points
indicated. Gender-matched littermates were used in all experiments. For 3C organoid
cultures, tissues were harvested when both RicEKO and CTR mice were in telogen stage
(P65-69).
Animals were housed and maintained according to FELASA guidelines in the animal facility
of the Max Planck Institute for Biology of Ageing, facility of the Department of Pharmacology,
University of Cologne, Germany and University of Helsinki. All experiments were approved
by local authorities (permit numbers 2017.A082, 2014.A067 and ESAVI-6357-2020).
METHOD DETAILS
3C organoid culture
Epidermal cells were isolated from male telogen-stage C56BL/6 mice, LifeAct-GFP (Riedl
et al., 2010), or Rictor-deficient mice (Ding et al., 2016) and cultured in 3C conditions as
described earlier (Chacón-Martínez et al., 2016). Briefly, epidermal single cell suspensions
were generated by incubating skin pieces in 0.8% trypsin for 50 min. Cells were embedded
in a 1:1 mixture of growth factor-reduced Matrigel and 3C medium (MEM Spinner’s
modification (Sigma), 5 µg/mL insulin (Sigma), 10 ng/mL EGF (Sigma), 10 µg/mL transferrin
(Sigma), 10 µM phosphoethanolamine (Sigma), 10 µM ethanolamine (Sigma), 0.36 µg/mL
hydrocortisone (Calbiochem), 2 mM glutamine (Gibco), 100 U/mL penicillin and 100 µg/mL
streptomycin (Gibco), and 10% chelated fetal calf serum (Gibco), 5 µM Y27632, 20 ng/mL
mouse recombinant VEGF, 20 ng/mL human recombinant FGF-2 (all from Miltenyi Biotec)).
23
Where indicated Akt inhibitor (1L6-Hydromethyl-chiro-inositol-2(R)-2-O-methyl-3-O-
octadecyl-sn-glycerocarbonate) (10µM, Merck), CB839 (20 nM, Cayman Chemical), CoCl2
(100 µM, Sigma), Sodium dichloroacetate (DCA, 1µM, Sigma), UK5099 (Cayman Chemical),
Antimycin A (1 nM, Sigma), Rotenone (10nM, Sigma), SC 79 (5 µg/ml, Tocris) and dimethyl-
alpha-ketoglutaric acid (200 µM, Sigma) were added. 2-Deoxy-D-glucose (2-DG, Agilent)
was used at various concentrations noted in the figure legend.
Cells were cultured in 5% CO2 at 37 C unless otherwise indicated. For hypoxia experiments
cells were culture in a hypoxia chamber (Hypoxic Glove Box, Cat# 8375250, CoyLabs) in 2%
O2 at 37 C.
Flow cytometry
For freshly isolated epidermal cells, single cell suspensions were generated as above. For
3C cultured cells, matrigel droplets were degraded for 8 min at 37°C in 0.5% Trypsin, 0.5
mM EDTA in PBS. Trypsin was neutralized using FACS buffer (2mM EDTA, 2% FBS in
PBS).
Cells were centrifuged for 5 min at 2000 rpm at 4°C, resuspended in FACS buffer and
stained for cell surface markers, CD34-eFluor 660 (Thermo Fisher Scientific, Cat # 50-0341-
82, 1:100) and ITGA6-PE/Cy7 (Biolegend, Cat #313622, 1:1000) on ice. 15 min before the
analysis, 7AAD (Thermo) was added to each sample. Cells were analyzed using
FACSCANTO II (BD Biosciences) or sorted using FACSAria Illu Svea (BD Biosciences)
and/or FACSAria Fusion (BD Biosciences). Sorted cells were collected into 15-mL conical
tubes with cell culture medium at 4ºC, centrifuged, re-suspended in medium and cultured in
3C.
Pathway analyses
Gene ontology (GO) term analyses of the published RNAseq dataset were carried out using
Panther. The pathway analysis was carried out using Ingenuity Pathway Analysis (IPA;
QIAGEN)
In vivo multiphoton autofluorescence imaging
In vivo multiphoton imaging experiments were approved and performed according to the
University of Arkansas IACUC (Protocol #18084). C57BL/6J male mouse aged 21 days and
23 days were anesthetized using isoflurane (2-5% induction, 1-3% maintenance. The dorsal
24
side of the ears were depilated from the base of the ear to the tip. The ear was then gently
mounted to a custom 3D printed platform using double sided tape (Li et al., 2012). Individual
hair follicles were located and imaged using a Bruker Ultima Investigator multiphoton
microscope (Middleton, Wisconsin) equipped with a Titanium:Sapphire laser (Spectra-
Physics; Santa Clara, California). Images were acquired at a resolution of 1024x1024 with
12-bit depth via a 20x, 1.0 NA water-immersion objective (Olympus; Tokyo, Japan) at 2.5x
digital zoom. Autofluorescence emission was collected through a 680 nm low pass filter
(Chroma, ET680sp-2p) and series of dichroic mirrors that separated emission into four
GaAsP photomultiplier tubes (PMTs) (Hamamatsu; H10770PB-40). NADH fluorescence
isolated at 755nm excitation using a PMT equipped with a 460(±20) nm filter, and FAD
fluorescence was isolated at 855nm excitation using a PMT equipped with a 525(±25) nm
filter. Collagen second harmonic generation (SHG) signal was collected at both excitation
wavelengths using a PMT equipped with a 430nm short pass filter. Hair follicles were easily
recognized due to the intensely autofluorescent hair shafts surrounded by dermal collagen
SHG signal. An overlay of the autofluorescence and SHG images was generated for each
image field and adjusted for contrast and brightness (Fig. 2C, top row) to identify the HFSC
bulge area. The HFSC bulge area was identified by the cellular morphology surrounding
the hair shaft below the strongly autofluorescent sebaceous glands. This bulge area,
excluding pixels containing the hair shaft, was segmented (Fig. 2C, bottom row, dashed
lines) using a manual tracing function written in MATLAB.
Image fluorescence intensities were normalized as previously described to account for
differences in laser power and PMT voltage (Jones et al., 2018; Quinn et al., 2013). The
pixels within the traced HFSC bulge area were then used to compute the average redox
ratio of [FAD/(NADH+FAD)] from the normalized NADH and FAD intensities in 29 different
hair follicles (n=9 anagen, n=20 telogen). Visualization of the redox ratio was achieved by
mapping values corresponding to the redox ratio of each pixel to a jet color map in MATLAB.
Differences in the average optical redox ratio between anagen and telogen bulge regions
were determined using an unpaired t-test.
EdU and Annexin V incorporation
For analysis of cell proliferation, cells were cultured for 12-13 days before EdU (10µM)
incubation for 24 h. Single cell suspensions were incubated with Fixable Viability Dye eFluor
506 (Thermo) for 25 min on ice, washed once with FACS buffer, incubated with antibodies
against cell surface markers as described above and subsequently fixed with 4% PFA. Cells
25
were washed once, permeabilized with Click-iT saponin-based permeabilization buffer for
15 min at room temperature in the dark and washed. The EdU reaction cocktail was
prepared following the manufacturer’s protocol for Click-iT Plus EdU Alexa Fluor 488 Flow
Cytometry Assay Kit (Thermo), incubated for 15 min at room temperature and subjected to
flow cytometry analysis.
For analysis of apoptosis using Annexin V staining, single cell suspensions were incubated
with PE-tagged anti-Annexin V (1:25; Biolegend) for 15min at room temperature. Cells were
washed and 7AAD was incubated for 5min at room temperature prior to flow cytometry
analysis.
Seahorse metabolic flux measurements
Cells were FACS sorted into CD34+ITGA6+ and CD34-ITGA6+ cells and seeded in 3C
conditions on Seahorse XF96 V3 PS cell Culture Microplates (Agilent) and allowed to
recover for 4d. Seahorse measurements were performed following the manufacturer’s
protocol from Agilent Seahorse XF Cell Mito Stress Test Kit and XF Glycolysis Stress Test
Kit. In short, a day before the experiment, calibrant solution was added to XFe96 sensor
cartridges and incubated overnight under 37 ºC without CO2. The day of the experiment,
cultured cells were washed twice and incubated for 45 min at 37 ºC without CO2 with
Seahorse assay media adjusted to pH 7.4 containing 10 mM Glucose and 2 mM L-Glutamine
for mito stress protocol and 2 mM L-Glutamine only for glycolysis stress protocol. For
OXPHOS measurements, 20 µM oligomycin, 20 µM FCCP, and 0.5 µM rotenone/antimycin
A were injected into port A, B, and C respectively. For glycolysis analysis, 100 mM glucose,
10 µM oligomycin, 500 µM 2-DG were injected into port A, B, and C respectively. After the
experiment, cells were lysed in ice-cold RIPA buffer, boiled at 65ºC for 10min, after which
dsDNA content was measured with a Qubit fluorometer (Thermo) according to the
manufacturer’s instructions. DNA concentrations were used to normalize Seahorse
measurements to cell numbers.
qRT-PCR
RNA was isolated using RNeasy Plus Mini Kit or RNeasy Plus Micro Kit (Qiagen). 500 ng of
RNA was subjected to reverse transcription using SuperScript IV VILO Master Mix (Thermo)
following the manufacturer’s protocol. PCR was performed with CFX384 Touch Real Time
PCR Detection System (Bio-Rad) using the DyNAmo Color Flash SYBR Green Mix
(Thermo). Gene expression was quantified using the Ct method using normalization to
26
S18 or Actb. For comparisons of freshly isolated and cultured HFSCs, normalization was
performed using the ERCC RNA Spike-in Mix (Invitrogen). The primer sequences are
provided in Table S1.
Metabolite extraction for LC-MS
Cells were either freshly FACS purified from skin or 3C cultured for 12-14 days after which
they were FACS purified or directly pelleted by centrifugation and snap frozen in liquid
nitrogen. For 13C glutamine/glucose flux analyses, on the day of sample collection 3C-
cultured cells were washed once with PBS, incubated with DMEM with 10% chelated FBS,
and without glucose and glutamine and phenol red (Gibco) supplemented with U-13C6-D-
Glucose (Cambridge Isotope Laboratories) and L-Glutamine (Sigma) or 13C5-L-Glutamine
(Cambridge Isotope Laboratories) and D-Glucose (Agilent) for 3 h under culturing conditions.
Cells were then washed with PBS, trypsinized, neutralized with ice-cold DMEM with 10%
chelated FBS, FACS sorted, and snap frozen in liquid nitrogen.
Metabolite extraction from each cell pellet was performed using 1 mL of a mixture of
40:40:20 [v:v:v] of pre-chilled (-20°C) acetonitrile:methanol:water (OptimaTM LC/MS grade,
Thermo Fisher Scientific). The samples were subsequently vortexed until the cell pellets
were fully suspended, before incubating them on an orbital mixer at 4°C for 30 min at 1500
rpm. For further disintegration, samples were sonicated for 10 min in an ice cooled bath
sonicator (VWR, Germany) before centrifuging them for 10 min at 21100x g and 4°C. The
metabolite-containing supernatant was collected in fresh tubes and concentrated to dryness
in a Speed Vac concentrator (Eppendorf). The protein-containing pellets were also collected
and used for protein quantification (BCA Protein Assay Kit, Thermo Fisher Scientific).
Anion-Exchange Chromatography Mass Spectrometry
Extracted metabolites were re-suspended in 100 µl of water (OptimaTM LC/MS grade,
Thermo Fisher Scientific,). 80 µl were used for the analysis of anionic metabolites from the
TCA cycle and glycolysis, while 20 µL were used for the analysis of amino acids (see below,
LC-MS analysis of amino acids).
Anions were analysed using a Dionex ionchromatography system (ICS 5000, Thermo Fisher
Scientific) essentially as described previously (Schwaiger et al., 2017). In brief, 10 µL of
polar metabolite extract were injected in full loop mode using an overfill factor of 3, onto a
Dionex IonPac AS11-HC column (2 mm × 250 mm, 4 μm particle size, Thermo Scientific)
equipped with a Dionex IonPac AG11-HC guard column (2 mm × 50 mm, 4 μm, Thermo
27
Scientific). The column temperature was held at 30°C, while the auto sampler was set to
6°C. A potassium hydroxide gradient was generated by the eluent generator using a
potassium hydroxide cartridge that was supplied with deionized water. The metabolite
separation was carried at a flow rate of 380 µL/min, applying the following gradient. 0-3 min,
10 mM KOH; 3-12 min, 10−50 mM KOH; 12-19 min, 50-100 mM KOH, 19-21 min, 100 mM
KOH, 21-21.1 min, 100-10 mM KOH. The column was re-equilibrated at 10 mM for 8 min.
For the targeted analysis of pool sizes the eluting metabolites were detected in negative ion
mode using ESI MRM (multi reaction monitoring) on a Xevo TQ (Waters) triple quadrupole
mass spectrometer (MS). For the analysis of isotope ratios the samples were measured on
a Q-Exactive HF high resolution MS (Thermo Fisher Scientific). The settings on the Xevo
TQ were the following: capillary voltage 1.5 kV, desolvation temperature 550°C, desolvation
gas flow 800 L/h, collision cell gas flow 0.15 mL/min. Two MRM transitions (one quantitative
and one confirmative MRM transition). The quantitative MRM transitions were: Citrate m/z
precursor mass (M-H+) 191, fragment mass (M-H+) m/z 111 : cone voltage 18V, collision
energy 10V. Isocitrate m/z precursor mass (M-H+) 191 fragment mass (M-H+) m/z 72: cone
voltage 18V, collision energy 12V. α-ketoglutarate m/z precursor mass (M-H+) 145 fragment
mass (M-H+) m/z 56: cone voltage 18V, collision energy 12V. Succinate m/z precursor mass
(M-H+) 117 fragment mass (M-H+) m/z 55: cone voltage 20V, collision energy 14V. Fumarate
m/z precursor mass (M-H+) 115 fragment mass (M-H+) m/z 27: cone voltage 22V, collision
energy of 8V. Malate m/z precursor mass (M-H+) 132 fragment mass (M-H+) m/z 71: cone
voltage 22V, collision energy of 16V.
Data analysis and peak integration for this datatype was performed using the TargetLynx
Software (Waters).
The high resolution Q-Exactive HF MS was operating in negative ionization mode monitoring
the mass range m/z 50-750. The heated ESI source settings of the mass spectrometer were:
Spray voltage 3.2 kV, capillary temperature was set to 275°C, sheath gas flow 60 AU and
aux gas flow 20 AU at a temperature of 300°C. The S-lens was set to a value of 60. Targeted
data analysis was performed as described in the section for the tracing analysis of amino
acids (see below).
LC-MS analysis of cellular pool sizes and 13C fluxes of amino acids
For amino acid analysis, the benzoylchlorid derivatization method (Wong et al., 2016) was
used. In brief: 20 µl of the aqueous sample, were mixed with 10 µl of 100 mM sodium
carbonate (Sigma) followed by the addition of 10 µl 2% benzoylchloride (Sigma) in
28
acetonitrile (VWR). Samples were then analyzed using an Acquity iClass UPLC (Waters)
connected to a Q-Exactive HF (Thermo Fisher Scientific) operating.
For each analysis 2 µl of the derivatized sample were injected onto a 100 x 1.0 mm HSS T3
column, which is packed with 1.8 µm particles (Waters). The flow rate was 100 µL/min and
the buffer system consisted of buffer A (10 mM ammonium formate, 0.15% formic acid in
water) and buffer B (acetonitrile). The gradient was: 0% B at 0 min; 0-15% B 0-0.1 min; 15-
17% B 0.1-0.5 min; 17-55% B 0.5-14 min, 55-70% B 14-14.5 min; 70-100% B 14.5-18 min;
100% B 18-19 min; 100-0% B 19-19.1 min, 19.1-28 min 0% B. The mass spectrometer was
operating in positive ionization mode monitoring the mass range m/z 50-750. The heated
ESI source settings of the mass spectrometer were: Spray voltage 3.5kV, capillary
temperature 275°C, sheath gas flow 40 AU and aux gas flow 20 AU at a temperature of
300°C. The S-lens was set to 60.
Data analysis of isotope ratios was performed using the TraceFinder software (Version 4.2,
Thermo Fisher Scientific). Identity of each compound was validated by authentic reference
compounds, which were analyzed independently. For the isotope enrichment analysis the
area of the extracted ion chromatogram (XIC) of each isotope [M + H]+ were determined
with a mass accuracy (<5 ppm) before calculating the proportions of each detected isotope
towards the sum of all isotopes of the corresponding compound. These proportions are
displayed as percent values for each isotope.
LC-MS analysis of cellular pools of acetyl and succinyl CoA
For the analysis of acetyl and succinyl CoA the extracted metabolites were re-suspended in
100 µl of UPLC grade acetonitrile:water (80:20 [v:v], OptimaTM LC-MS-grade, Thermo Fisher
Scientific). The samples were analyzed on an Acquity iClass UPLC (Waters), using a
SeQuant ZIC-pHILIC 5 µm polymer 150 x 2.1 mm column (Merck) connected to a Xevo TQ-
S (Waters) triple quadrupole mass spectrometer.
3 µl of re-suspended metabolite extract was injected onto the column and separated using
a flow rate of 350 µl/min of buffer A (20mM ammonium carbonate, 0.1% ammonium
hydroxide ) and buffer B (acetonitrile) using the following gradient: 0-1 min 20-70% A; 1-3
min 70-80% A; 3-4 min 80-98%A; 4-4.5 min 98% A; 4.5-5 min 98-20% A. The column is re-
equilibrated at 20% A for additional 5 min.
The eluting metabolites were detected in negative ion mode using ESI MRM (multi reaction
monitoring) applying the following settings: capillary voltage 1.5 kV, desolvation temperature
550°C, desolvation gas flow rate 800 L/h, collision cell gas flow 0.15 mL/min. The following
29
MRM transitions were used for relative compound quantification of acetyl CoA m/z precursor
mass (M+H+) 810, fragment mass (M+H+) m/z 303: cone voltage of 98V, collision energy
28V, while succinyl CoA was analysed using a precursor mass (M+H+) m/z 868, fragment
mass (M+H+) m/z 361: cone voltage of 2V, collision energy 34V. Data analysis and peak
integration were performed using the TargetLynx Software (Waters).
Gas chromatography mass spectrometry
Similar to the above described 13C flux analysis of glycolysis and TCA cycle intermediates
using ion chromatography coupled to high resolution MS, high resolution GC-MS (Q-
Exactive GC, Thermo Fisher Scientific) was employed.
For this purpose, dried metabolite pellets required a two-step derivatization using
methoxyamine (methoxyamine hydrochloride, Sigma) and N-Methyl-N-trimethylsilyl-
trifluoracetamide (MSTFA, Macherey-Nagel) before performing the GC-MS analysis.
In brief: Dried samples were re-suspended in 5 µL of a freshly prepared (20 mg/mL) solution
of methoxyamine in pyridine (Sigma). The re-suspended pellets were incubated for 90 min
at 40°C on an orbital shaker (VWR) at 1500 rpm. After adding additional 45 µL of MSTFA,
the samples were incubated for additional 30 min at 40°C and agitation at 1500 rpm. At the
end of the derivatisation the samples were centrifuged for 10 min at 21100x g and 40 µL of
the clear supernatant were transferred to fresh autosampler vials with conical glass inserts.
For the GC-MS analysis 1 µL of each sample was injected using a PAL autosample system
(Thermo Fisher Scientifc) using a Split/Splitless (SSL) injector at 300 C in splitless mode.
The carrier gas flow (helium) was set to 2 ml/min using a 30m DB-35MS capillary column
(0.250 mm diameter and 0.25 µm film thickness, Agilent). The GC temperature program was:
2 min at 85°C, followed by a 15°C per min ramp to 330°C. At the end of the gradient the
temperature was held for additional 6 min at 330°C. The transfer line and source
temperature are both set to 280°C. The filament, which was operating at 70 V, was switched
on 2 min after the sample was injected. During the whole gradient period the MS was
operated in full scan mode covering a m/z range between 70 and 800 at a scan speed of 20
Hertz.
Similar to the analysis of the isotope enrichment analysis in the amino acids, isotope
enrichment analysis in glycolysis and TCA cycle metabolites were determined using GC-MS
(Q-Exactive GC-Orbitrap, Thermo Fisher Scientific). For this purpose, metabolites were
derivatized using a two-step procedure starting with an methoxyamination (methoxyamine
30
hydrochloride, Sigma) followed by a trimethyl-silylation using N-Methyl-N-trimethylsilyl-
trifluoracetamid (MSTFA, Macherey-Nagel).
In brief: Dried samples were re-suspended in 5 µL of a freshly prepared (20 mg/mL) solution
of methoxyamine in pyridine (Sigma) to perform the methoxyamination. These samples were
then incubated for 90 min at 40°C on an orbital shaker (VWR) at 1500 rpm. In the second
step additional 45 µL of MSTFA were added and the samples were incubated for additional
30 min at 40°C and 1500 rpm. At the end of the derivatisation the samples were centrifuged
for 10 min at 21100x g and 40 µL of the clear supernatant was transferred to fresh auto
sampler vials with conical glass inserts (Chromatographie Zubehoer Trott). For the GC-MS
analysis 1 µL of each sample was injected using a PAL autosample system (Thermo Fisher
Scientifc) using a Split/Splitless (SSL) injector at 300 °C in splitless mode. The carrier gas
flow (helium) was set to 2 ml/min using a 30m DB-35MS capillary column (0.250 mm
diameter and 0.25 µm film thickness, Agilent). The GC temperature program was: 2 min at
85 °C, followed by a 15 °C per min ramp to 330 °C. At the end of the gradient the temperature
is held for additional 6 min at 330 °C. The transfer line and source temperature are both set
to 280 °C. The filament, which was operating at 70 V, was switched on 2 min after the sample
was injected. During the whole gradient period the MS was operated in full scan mode
covering a mass range m/z 70 and 800 with a scan speed of 20 Hertz.
For the data analysis peak areas of extracted ion chromatograms of each isotope of
compound-specific fragments [M - e-]+ were determined using the TraceFinder software
(Version 4.2, Thermo Fisher Scientific) with a mass accuracy (<5 ppm). Subsequently
proportions of each detected isotope towards the sum of all isotopes of the corresponding
compound-specific fragment were determined. These proportions are given as percent
values for each isotope.
Details on the compound-specific fragments of the analysed compounds: 3-phosphoglyceric
acid was analysed from a three carbon-containing fragment with the chemical formula
C14H36O7PSi4 and the m/z 459.12702. Phosphenolpyruvic acid was analysed from a three
carbon-containing fragment (C11H26O6PSi3) with a m/z of 369.07693. Pyruvic acid was
analysed from a three carbon-containing fragment (C6H12NO3Si) and the m/z of 174.05809.
Lactic acid was analysed from a three carbon-containing fragment (C8H19O3Si2) and a m/z
of 219.08672. Citric acid was analysed from a five carbon-containing fragment
(C11H21O4Si2) and a m/z of 273.09729. Alpha-ketoglutaric acid was analysed from a five
carbon-containing fragment (C8H12NO3Si) and a m/z of 198.0580963. Succinic acid was
analysed from a four carbon-containing fragment (C9H19O4Si2) and a m/z of 247.08164.
Fumaric acid was analysed from a four carbon-containing fragment (C9H17O4Si2) and a
31
m/z of 247.08164. Malic acid was analysed from a four carbon-containing fragment
(C9H17O4Si2) and a m/z of 247.08164. The retention time and therefore identity of each
compound was validated by authentic reference compounds which were analysed
independently.
H&E staining
The upper back skin of mice was dissected fixed in ice-cold 4% PFA for 1h on ice,
dehydrated with 50% ethanol overnight, and processed using ExcelsiorTM AS Tissue
Processor (Thermo), embedded in paraffin and sectioned. Subsequently, paraffin was
removed and H&E staining was performed using automated Gemini slide stainer (Thermo).
Stained tissues were mounted on No. 1 rectangular coverslips with Cytoseal XYL mounting
medium (Thermo). Images of H&E stained tissue samples were obtained using Nikon
Eclipse Ci (Nikon) using Nikon NIS Elements D software and a 20x dry lens.
Western blots
Cells were lysed in radio immunoprecipitation assay (RIPA) buffer (150 mM NaCl, 50 mM
Tris-HCl, 5mM EDTA, 0.1% SDS, 1% sodium deoxycholate and 1% Triton X-100,
supplemented with proteinase inhibitor protease and phosphatase inhibitor tablets (Roche).
Protein concentration was determined by Micro BCA Protein Assay Kit (Thermo Scientific)
after which 20 μg protein was subjected to SDS–PAGE. Separated polypeptides were
transferred onto PVDF membranes. After blocking (5% non-fat milk in TBS-Tween buffer),
membranes were incubated with primary antibodies. Rictor (1:1000, cat #2114), Akt (1:1000,
cat #9272), Akt-pS473 (1:1000, cat #9271), S6-pS240/244 (1:1000, cat #5364) were all from
Cell Signaling and used at 1:1000 dilution. β-actin (1:2000) was from Sigma (A2066) and
Hif1 from Novus Biologicals (1:1000, cat #NB100-479). After incubation with horseradish
peroxidase-conjugated secondary antibodies (DAKO), the blot was developed with ECL
substrate (Pierce) and exposed on X-ray film (Amersham Biosciences) or BioRad Chemidoc
Imaging System.
Immunofluorescence
Tissue biopsies were fixed (4% PFA), embedded in paraffin and sectioned. Sections were
de-paraffinized using a graded alcohol series, blocked in 10% normal goat serum, and
incubated with primary antibodies diluted in Dako Antibody Diluent over night at 4°C. Bound
primary antibody was detected by incubation with Alexa-Fluor 488- or Alexa Fluor 568-
32
conjugated antibodies (Invitrogen). Nuclei were counterstained with 4′, 6-diamidino-2-
phenylindole (DAPI, Invitrogen). After washing slides were mounted in Elvanol. The
following antibodies were used: Akt-pS473 (1:100; Cell Signaling cat #9271), CD34 (1:1000,
gift from Manuel Koch, University of Cologne), CD34 (1:100, Invitrogen cat #14-0341-82),
K16 (1:150, Abcam cat #Ab76414), K6 (1:500, Biolegend cat #905701), Sox9 (1:100, Santa
Cruz cat #sc-166505), and periostin (1:150, Santa Cruz cat #sc-398631). TUNEL staining
was performed with In Situ Cell Death Kit (Roche) according to manufacturer’s instructions.
All fluorescence images were collected by laser scanning confocal microscopy (SP8X; Leica)
with Leica Application Suite software (LAS X version 2.0.0.14332), using 40x, 63x or 100x
immersion objectives. Images where acquired at room temperature using sequential
scanning of frames of 1 µm thick confocal planes (pinhole 1) after which planes were
projected as a maximum intensity confocal stack. Images were collected with the same
settings for all samples within an experiment.
Epidermal whole mounts from back skin
Shaved back skin was dissected, after which adipose and connective tissue were removed
by scraping with a scalpel. Skin pieces were floated on thermolysine/PBS (Sigma-Aldrich)
at 37 °C (dermal side down) for 30-40 min. Epidermis was separated from dermis by peeling
and fixed in 4% PFA for 1.5 h at room temperature. After washing with PBS, epidermal tissue
was processed for staining as described above for immunofluorescence.
Image analyses
All images were analyzed using Fiji (Schindelin et al., 2012). To quantify intensity of hair
follicle immunofluorescence, a region of interest (ROI) was manually drawn around each
hair follicle of maximum intensity projected confocal stacks, after which mean grey value
within each ROI was quantified. BrdU- and Tunel-positive cells and hair follicle bulges were
counted manually. 3C images were analyzed by generating maximum intensity projections
of the acquired confocal stack. ROIs were manually drawn around individual organoids
based on DAPI staining. ROIs of interest were then scored for CD34 staining presence or
absence, and ultimately, mean gray value of the Akt-pS473 staining was determined for
each ROI.
BrdU injections and lineage trace
To identify the hair cycle stage, mouse back skin was shaved during predicted first telogen.
Skin color and morphology were monitored daily. BrdU (BD Pharmingen) was dissolved in
33
PBS and injected intraperitoneally at 50µg/g body weight in 150uL of volume twice at 6h
intervals to littermate mice during mid-anagen; control mice (~P34) and the hair follicle
cycling stage matched RicEKO mice (~P43) as described in Hsu et al. 2011. Back skin tissue
was collected at time points indicated and cryo embedded in OCT (Tissue-Tek). Tissue
sections were cut, fixed with ice-cold methanol, washed with PBS and permeabilized with
0.1% Triton-X-100/PBS for 5 min. DNA was denatured by soaking the samples in 2M HCl
for 30 min. Samples were then subjected to immunofluorescence analyses as described
above with anti-CD34 and anti-BrdU (1:100, BD Biosciences) antibodies.
Hair dye experiments
Mouse back skin was washed and dyed with glow-in-the-dark hair dye (Manic Panic, Hot
Pink) for 30 min during their first telogen hair cycle (~P21) as described in Lay et al 2016.
Hair samples were taken on P26, P39, P48, P53 and P65 and images were taken under an
epifluorescence microscope (Leica Sp8) using UV light.
BPTES injections
Anesthetized mice were shaved on their back skin during the mid-anagen phase (Control:
P34-36; RicEKO: P42-P45 ) and injected intradermally with BPTES (Cayman Chemical) at
12.5 mg/kg body weight in 10% DMSO in the total of 200 µL of PBS as described in (Xiang
et al., 2015) three times a week for 2 weeks or until the second telogen phase had been
reached. The back skin tissue was taken for histological analyses and processed as
described above for H/E stainings.
Hypoxyprobe injection and staining
Solid pimonidazole HCl (Hypoxyprobe-1; Hypoxyprobe Kit, Hypoxyprobe Inc) was
reconstituted in sterile saline and injected intraperitoneally at 60 mg/kg body weight at late
anagen (P38) or telogen (P75). Animals were sacrificed 1 hour post injection, and back skin
collected and fixed (4% PFA), embedded in paraffin and sectioned. Samples were
rehydrated through graded ethanol series and subjected to antigen retrieval in pH6 DAKO
reagent (Agilent Technologies, S2031). Tissue sections were blocked in 5% normal goat
serum/3% BSA/PBS for 1 hour. Sections were incubated with anti-pimonidazole mouse IgG1
monoclonal antibody (Mab1, 1:50, HypoxyprobeKit) in DAKO diluent at 4°C overnight.
Sections were washed with 1xPBS and then incubated with biotinylated secondary for 30
minutes at room temperature. Subsequently, sections were incubated in Vectastain Elite
34
ABC reagent for 30 minutes at room temperature and then washed with 1xPBS. Peroxidase
activity was visualized with Sigmafast 3,3’-Diaminobenzidine (D4168, Sigma). Slides were
then processed as for H&E staining.
QUANTIFICATION AND STATISTICAL ANALYSIS
Statistical analyses were performed using GraphPad Prism software (GraphPad, version 8).
Statistical significance was determined by the specific tests indicated in the corresponding
figure legends. In all cases where a test for normally distributed data was used, normal
distribution was confirmed with the Kolmogorov–Smirnov test (α = 0.05). No methods were
used to predetermine sample size. All experiments presented in the manuscript were
repeated at least in 3 independent experiments/biological replicates. All P and n values as
well as statistical test used is reported in the figure legends. Mice were assigned to groups
randomly (drug treatments) or according to genotype. Analyses were not blinded, and no
datapoints were removed from the analyses.
References
Albert, V., and Hall, M.N. (2015). mTOR signaling in cellular and organismal energetics. Curr Opin Cell Biol 33, 55-66.
Bedogni, B., Welford, S.M., Cassarino, D.S., Nickoloff, B.J., Giaccia, A.J., and Powell, M.B. (2005). The hypoxic microenvironment of the skin contributes to Akt-mediated melanocyte transformation. Cancer Cell 8, 443-454.
Bentzinger, C.F., Romanino, K., Cloetta, D., Lin, S., Mascarenhas, J.B., Oliveri, F., Xia, J., Casanova, E., Costa, C.F., Brink, M., et al. (2008). Skeletal muscle-specific ablation of raptor, but not of rictor, causes metabolic changes and results in muscle dystrophy. Cell Metab 8, 411-424.
Binek, A., Rojo, D., Godzien, J., Ruperez, F.J., Nunez, V., Jorge, I., Ricote, M., Vazquez, J., and Barbas, C. (2019). Flow Cytometry Has a Significant Impact on the Cellular Metabolome. J Proteome Res 18, 169-181.
Blanpain, C., and Fuchs, E. (2014). Stem cell plasticity. Plasticity of epithelial stem cells in tissue regeneration. Science 344, 1242281.
Boveris, A., and Chance, B. (1973). The mitochondrial generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen. Biochem J 134, 707-716.
Castilho, R.M., Squarize, C.H., Chodosh, L.A., Williams, B.O., and Gutkind, J.S. (2009). mTOR mediates Wnt-induced epidermal stem cell exhaustion and aging. Cell Stem Cell 5, 279-289.
35
Chacón-Martínez, C.A., Klose, M., Niemann, C., Glauche, I., and Wickström, S.A. (2016). Hair follicle stem cell cultures reveal self-organizing plasticity of stem cells and their progeny. EMBO J 36, 151-164.
Chacón-Martínez, C.A., Koester, J., and Wickström, S.A. (2018). Signaling in the stem cell niche: regulating cell fate, function and plasticity. Development 145.
DeBerardinis, R.J., and Cheng, T. (2010). Q's next: the diverse functions of glutamine in metabolism, cell biology and cancer. Oncogene 29, 313-324.
Ding, X., Bloch, W., Iden, S., Ruegg, M.A., Hall, M.N., Leptin, M., Partridge, L., and Eming, S.A. (2016). mTORC1 and mTORC2 regulate skin morphogenesis and epidermal barrier formation. Nat Commun 7, 13226.
Ding, X., Willenborg, S., Bloch, W., Wickström, S.A., Wagle, P., Brodesser, S., Roers, A., Jais, A., Brüning, J.C., Hall, M.N., Rüegg, M.A., Eming, S.A., (2020). Epidermal mTORC2 controls lipid synthesis and filaggrin processing in epidermal barrier formation. J Allergy Clin Immunol 145, 283-300.e8.
Eltschinger, S., and Loewith, R. (2016). TOR Complexes and the Maintenance of Cellular Homeostasis. Trends Cell Biol 26, 148-159.
Evans, N.T., and Naylor, P.F. (1967). The systemic oxygen supply to the surface of human skin. Respir Physiol 3, 21-37.
Flores, A., Schell, J., Krall, A.S., Jelinek, D., Miranda, M., Grigorian, M., Braas, D., White, A.C., Zhou, J.L., Graham, N.A., et al. (2017). Lactate dehydrogenase activity drives hair follicle stem cell activation. Nat Cell Biol 19, 1017-1026.
Georgakoudi, I., and Quinn, K.P. (2012). Optical imaging using endogenous contrast to assess metabolic state. Annu Rev Biomed Eng 14, 351-367.
Gonzales, K.A.U., and Fuchs, E. (2017). Skin and Its Regenerative Powers: An Alliance between Stem Cells and Their Niche. Dev Cell 43, 387-401.
Gross, M.I., Demo, S.D., Dennison, J.B., Chen, L., Chernov-Rogan, T., Goyal, B., Janes, J.R., Laidig, G.J., Lewis, E.R., Li, J., et al. (2014). Antitumor activity of the glutaminase inhibitor CB-839 in triple-negative breast cancer. Mol Cancer Ther 13, 890-901.
Hafner, M., Wenk, J., Nenci, A., Pasparakis, M., Scharffetter-Kochanek, K., Smyth, N., Peters, T., Kess, D., Holtkotter, O., Shephard, P., et al. (2004). Keratin 14 Cre transgenic mice authenticate keratin 14 as an oocyte-expressed protein. Genesis 38, 176-181.
Higgins, C.A., Westgate, G.E., and Jahoda, C.A. (2009). From telogen to exogen: mechanisms underlying formation and subsequent loss of the hair club fiber. J Invest Dermatol 129, 2100-2108.
Hildyard, J.C., Ammala, C., Dukes, I.D., Thomson, S.A., and Halestrap, A.P. (2005). Identification and characterisation of a new class of highly specific and potent inhibitors of the mitochondrial pyruvate carrier. Biochim Biophys Acta 1707, 221-230.
Hsu, Y.C., Li, L., and Fuchs, E. (2014). Emerging interactions between skin stem cells and their niches. Nat Med 20, 847-856.
Hsu, Y.C., Pasolli, H.A., and Fuchs, E. (2011). Dynamics between stem cells, niche, and progeny in the hair follicle. Cell 144, 92-105.
Ito, K., and Ito, K. (2016). Metabolism and the Control of Cell Fate Decisions and Stem Cell Renewal. Annu Rev Cell Dev Biol 32, 399-409.
36
Jin, L., Li, D., Alesi, G.N., Fan, J., Kang, H.B., Lu, Z., Boggon, T.J., Jin, P., Yi, H., Wright, E.R., et al. (2015). Glutamate dehydrogenase 1 signals through antioxidant glutathione peroxidase 1 to regulate redox homeostasis and tumor growth. Cancer Cell 27, 257-270.
Jo, H., Mondal, S., Tan, D., Nagata, E., Takizawa, S., Sharma, A.K., Hou, Q., Shanmugasundaram, K., Prasad, A., Tung, J.K., et al. (2012). Small molecule-induced cytosolic activation of protein kinase Akt rescues ischemia-elicited neuronal death. Proc Natl Acad Sci U S A 109, 10581-10586.
Jones, J.D., Ramser, H.E., Woessner, A.E., and Quinn, K.P. (2018). In vivo multiphoton microscopy detects longitudinal metabolic changes associated with delayed skin wound healing. Commun Biol 1, 198.
Joost, S., Annusver, K., Jacob, T., Sun, X., Dalessandri, T., Sivan, U., Sequeira, I., Sandberg, R., and Kasper, M. (2020). The Molecular Anatomy of Mouse Skin during Hair Growth and Rest. Cell Stem Cell 26, 441-457 e447.
Joost, S., Jacob, T., Sun, X., Annusver, K., La Manno, G., Sur, I., and Kasper, M. (2018). Single-Cell Transcriptomics of Traced Epidermal and Hair Follicle Stem Cells Reveals Rapid Adaptations during Wound Healing. Cell Rep 25, 585-597 e587.
Kobayashi, C.I., and Suda, T. (2012). Regulation of reactive oxygen species in stem cells and cancer stem cells. J Cell Physiol 227, 421-430.
Kolenc, O.I., and Quinn, K.P. (2019). Evaluating Cell Metabolism Through Autofluorescence Imaging of NAD(P)H and FAD. Antioxid Redox Signal 30, 875-889.
Laplante, M., and Sabatini, D.M. (2012). mTOR signaling in growth control and disease. Cell 149, 274-293.
Lay, K., Kume, T., and Fuchs, E. (2016). FOXC1 maintains the hair follicle stem cell niche and governs stem cell quiescence to preserve long-term tissue-regenerating potential. Proc Natl Acad Sci U S A 113, E1506-1515.
Li, J.L., Goh, C.C., Keeble, J.L., Qin, J.S., Roediger, B., Jain, R., Wang, Y., Chew, W.K., Weninger, W., and Ng, L.G. (2012). Intravital multiphoton imaging of immune responses in the mouse ear skin. Nat Protoc 7, 221-234.
Llufrio, E.M., Wang, L., Naser, F.J., and Patti, G.J. (2018). Sorting cells alters their redox state and cellular metabolome. Redox Biol 16, 381-387.
Loewith, R., and Hall, M.N. (2011). Target of rapamycin (TOR) in nutrient signaling and growth control. Genetics 189, 1177-1201.
Majmundar, A.J., Wong, W.J., and Simon, M.C. (2010). Hypoxia-inducible factors and the response to hypoxic stress. Mol Cell 40, 294-309.
Means, A.L., Xu, Y., Zhao, A., Ray, K.C., and Gu, G. (2008). A CK19(CreERT) knockin mouse line allows for conditional DNA recombination in epithelial cells in multiple endodermal organs. Genesis 46, 318-323.
Miranda, M., Christofk, H., Jones, L.D., and Lowry, W.E., (2018). Topical inhibition of the electron transport chain can stimulate the hair cycle. J Invest Dermatol 138, 968-972.
Miyamoto, K., Araki, K.Y., Naka, K., Arai, F., Takubo, K., Yamazaki, S., Matsuoka, S., Miyamoto, T., Ito, K., Ohmura, M., et al. (2007). Foxo3a is essential for maintenance of the hematopoietic stem cell pool. Cell Stem Cell 1, 101-112.
37
Muller-Rover, S., Handjiski, B., van der Veen, C., Eichmuller, S., Foitzik, K., McKay, I.A., Stenn, K.S., and Paus, R. (2001). A comprehensive guide for the accurate classification of murine hair follicles in distinct hair cycle stages. J Invest Dermatol 117, 3-15.
Nowak, J.A., Polak, L., Pasolli, H.A., and Fuchs, E. (2008). Hair follicle stem cells are specified and function in early skin morphogenesis. Cell Stem Cell 3, 33-43.
Oburoglu, L., Tardito, S., Fritz, V., de Barros, S.C., Merida, P., Craveiro, M., Mamede, J., Cretenet, G., Mongellaz, C., An, X., et al. (2014). Glucose and glutamine metabolism regulate human hematopoietic stem cell lineage specification. Cell Stem Cell 15, 169-184.
Peyssonnaux, C., Boutin, A.T., Zinkernagel, A.S., Datta, V., Nizet, V., and Johnson, R.S. (2008). Critical role of HIF-1alpha in keratinocyte defense against bacterial infection. J Invest Dermatol 128, 1964-1968.
Quinn, K.P., Sridharan, G.V., Hayden, R.S., Kaplan, D.L., Lee, K., and Georgakoudi, I. (2013). Quantitative metabolic imaging using endogenous fluorescence to detect stem cell differentiation. Sci Rep 3, 3432.
Rezvani, H.R., Ali, N., Nissen, L.J., Harfouche, G., de Verneuil, H., Taieb, A., and Mazurier, F. (2011). HIF-1alpha in epidermis: oxygen sensing, cutaneous angiogenesis, cancer, and non-cancer disorders. J Invest Dermatol 131, 1793-1805.
Riedl, J., Flynn, K.C., Raducanu, A., Gartner, F., Beck, G., Bosl, M., Bradke, F., Massberg, S., Aszodi, A., Sixt, M., et al. (2010). Lifeact mice for studying F-actin dynamics. Nat Methods 7, 168-169.
Robinson, M.M., McBryant, S.J., Tsukamoto, T., Rojas, C., Ferraris, D.V., Hamilton, S.K., Hansen, J.C., and Curthoys, N.P. (2007). Novel mechanism of inhibition of rat kidney-type glutaminase by bis-2-(5-phenylacetamido-1,2,4-thiadiazol-2-yl)ethyl sulfide (BPTES). Biochem J 406, 407-414.
Sarbassov, D.D., Guertin, D.A., Ali, S.M., and Sabatini, D.M. (2005). Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 307, 1098-1101.
Scarpulla, R.C., Vega, R.B., and Kelly, D.P. (2012). Transcriptional integration of mitochondrial biogenesis. Trends Endocrinol Metab 23, 459-466.
Schwaiger, M., Rampler, E., Hermann, G., Miklos, W., Berger, W., and Koellensperger, G. (2017). Anion-Exchange Chromatography Coupled to High-Resolution Mass Spectrometry: A Powerful Tool for Merging Targeted and Non-targeted Metabolomics. Anal Chem 89, 7667-7674.
Shyh-Chang, N., Daley, G.Q., and Cantley, L.C. (2013). Stem cell metabolism in tissue development and aging. Development 140, 2535-2547.
Stacpoole, P.W. (1989). The pharmacology of dichloroacetate. Metabolism 38, 1124-1144.
Stegen, S., Laperre, K., Eelen, G., Rinaldi, G., Fraisl, P., Torrekens, S., Van Looveren, R., Loopmans, S., Bultynck, G., Vinckier, S., et al. (2019). HIF-1alpha metabolically controls collagen synthesis and modification in chondrocytes. Nature 565, 511-515.
Sterky, F.H., Lee, S., Wibom, R., Olson, L., and Larsson, N.G. (2011). Impaired mitochondrial transport and Parkin-independent degeneration of respiratory chain-deficient dopamine neurons in vivo. Proc Natl Acad Sci U S A 108, 12937-12942.
Sun, R.C., and Denko, N.C. (2014). Hypoxic regulation of glutamine metabolism through HIF1 and SIAH2 supports lipid synthesis that is necessary for tumor growth. Cell Metab 19, 285-292.
38
Tassone, B., Saoncella, S., Neri, F., Ala, U., Brusa, D., Magnuson, M.A., Provero, P., Oliviero, S., Riganti, C., and Calautti, E. (2017). Rictor/mTORC2 deficiency enhances keratinocyte stress tolerance via mitohormesis. Cell Death Differ 24, 731-746.
Vagnozzi, A.N., Reiter, J.F., and Wong, S.Y. (2015). Hair follicle and interfollicular epidermal stem cells make varying contributions to wound regeneration. Cell Cycle 14, 3408-3417.
Varone, A., Xylas, J., Quinn, K.P., Pouli, D., Sridharan, G., McLaughlin-Drubin, M.E., Alonzo, C., Lee, K., Munger, K., and Georgakoudi, I. (2014). Endogenous two-photon fluorescence imaging elucidates metabolic changes related to enhanced glycolysis and glutamine consumption in precancerous epithelial tissues. Cancer Res 74, 3067-3075.
Wick, A.N., Drury, D.R., Nakada, H.I., and Wolfe, J.B. (1957). Localization of the primary metabolic block produced by 2-deoxyglucose. J Biol Chem 224, 963-969.
Wong, J.M., Malec, P.A., Mabrouk, O.S., Ro, J., Dus, M., and Kennedy, R.T. (2016). Benzoyl chloride derivatization with liquid chromatography-mass spectrometry for targeted metabolomics of neurochemicals in biological samples. J Chromatogr A 1446, 78-90.
Xiang, Y., Stine, Z.E., Xia, J., Lu, Y., O'Connor, R.S., Altman, B.J., Hsieh, A.L., Gouw, A.M., Thomas, A.G., Gao, P., et al. (2015). Targeted inhibition of tumor-specific glutaminase diminishes cell-autonomous tumorigenesis. J Clin Invest 125, 2293-2306.
Figure 1 Click here to access/download;Figure;FIgure 1-FINAL.tif
Figure 2 Click here to access/download;Figure;Figure 2 -FINAL-01.tif
Figure 3 Click here to access/download;Figure;Figure 3 -FINAL-01.tif
Figure 4 Click here to access/download;Figure;Figure 4 -FINAL-01.tif
Figure 5 Click here to access/download;Figure;Figure 5-FINAL.tif
Figure 6 Click here to access/download;Figure;Figure 6 -FINAL.tif
Figure 7 Click here to access/download;Figure;Figure 7-FINAL-01.tif
KEY RESOURCES TABLE Kim et al.
REAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
Monoclonal rat anti-CD34-eFluor660
Thermo Fisher Scientific Cat.#50-0341-82
Monoclonal rat anti-CD49f-PE/Cyanine7
Biolegend Cat.#313622
7AAD Thermo Fisher Scientific Cat.#A1310
Fixable Viability Dye eFluor506 Thermo Fisher Scientific Cat.#65-0866-14
Annexin V-PE Biolegend Cat.#640908
Monoclonal rabbit anti-Rictor Cell Signaling Cat.#2114
Polyclonal rabbit anti-AKT Cell Signaling Cat.#9272
Polyclonal rabbit anti-AKT-pS473
Cell Signaling Cat.#9271
Monoclonal rabbit anti-S6-pS240/244
Cell Signaling Cat.#5364
Polyclonal rabbit anti-beta-actin Sigma-Aldrich Cat.#MAB1501 and A2066
Polyclonal rabbit anti-CD34 Manuel Koch, University of Cologne
N/A
Monoclonal rat Anti-CD34 Invitrogen Cat.#14-0341-82
Polyclonal rabbit anti-K16 Abcam Cat.#ab76414
Polyclonal rabbit anti-K6 Biolegend Cat.#905701
Monoclonal mouse anti-Sox9 Santa Cruz Cat.#sc-166505
Monoclonal mouse anti-Periostin Santa Cruz Cat.#sc-398631
Monoclonal mouse anti-BrdU BD Biosciences Cat.#555627
Polyclonal rabbit anti-Hif1 Novus Biologicals Cat #NB100-479
Chemicals, Peptides, and Recombinant Proteins
Insulin Sigma-Aldrich Cat.#11882
EGF Sigma-Aldrich Cat.#E9644
Transferrin Sigma-Aldrich Cat.#T5158
Phosphoethanolamine Sigma-Aldrich Cat.#P0503
Ethanolamine Sigma-Aldrich Cat.#E0135
Hydrocortisone Calbiochem Cat.#386698
L-Glutamine Thermo Fisher Scientific Cat.#A2916801
D-Glucose Sigma Cat.#80-99-7
Y27632 Miltenyi Biotec Cat.#130-104-169
Mouse recombinant VEGF Miltenyi Biotec Cat.#130-094-086
Human recombinant FGF-2 Miltenyi Biotec Cat.#130-093-838
AKT inhibitor Merck Cat.#124005
CB-839 Cayman Chemical Cat.#22038
Sodium dichloroacetate Sigma-Aldrich Cat.#347795
UK5099 Cayman Chemical Cat.#16980
Antimycin A Sigma-Aldrich Cat.#A8674
Rotenone Sigma-Aldrich Cat.#R8875
Alpha-ketoglutaric acid Sigma-Aldrich Cat.#K1128
U-13C6-D-Glucose Cambridge Isotope Laboratories Cat.#CLM-1822-H-0.1 13C5-L-Glutamine Cambridge Isotope Laboratories Cat.#CLM-1396-1
BrdU BD Pharmingen Cat.#550891
BPTES Cayman Chemical Cat.#19284
Glow-in-the-dark hair dye Manic Panic, Hot Pink N/A
H2O CHROMASOLV LC-MS Ultra
Honeywell Riedel-de Haen Cat.#14263
Acetonitrile OPTIMA LC/MS Fisher Scientific Cat.#A955
Methanol OPTIMA LC/MS Fisher Scientific Cat.#A456
Key Resource Table
Isopropanol OPTIMA LC/MS Fisher Scientific Cat.#A461
Formic acid Sigma-Aldrich Cat.#5438040100
Ammonium carbonate VWR Cat.#21217.260
Sodium carbonate BioXtra Sigma-Aldrich Cat.#S7795
Benzoyl chloride ACS 99% Sigma-Aldrich Cat.#259950
MSTFA Silylation Reagent Macherey-Nagel Cat.#701270.201
Methoxyamine hydrochloride Sigma-Aldrich Cat.#226904
U-13C15N Amino acid Mix Std Cambridge Isotope Laboratories Cat.#MSK-A2-1.2
ATP-13C10 sodium salt solution Sigma-Aldrich Cat.#710695
Citric acid-2,2,4,4-d4 Sigma-Aldrich Cat.#485438
Critical Commercial Assays
Click-iT Plus EdU Alexa Fluor488 Flow Cytometry Assay Kit
Thermo Fisher Scientific Cat.#C10633
Agilent Seahorse XF Cell Mito Stress Test Kit
Agilent Technologies Cat.#103015-100
Agilent Seahorse XF Glycolysis Stress Test Kit
Agilent Technologies Cat.#103020-100
Qubit dsDNA HS Assay Kit Thermo Fisher Scientific Cat.#Q32851
RNeasy Plus Mini Kit Qiagen Cat.#74104
RNeasy Plus Micro Kit Qiagen Cat.#74004
SuperScript IV VILO Master Mix Thermo Fisher Scientific Cat.#11755-250
DyNAmo Color Flash SYBR Green Mix
Thermo Fisher Scientific Cat.#F416
Pierce BCA Protein Assay Thermo Fisher Scientific Cat.#23227
HypoxyprobeKit Hypoxyprobe Inc Cat.# HP1-100Kit
ERCC Spike-in Mix Invitrogen Cat.# 4456740
Experimental Models: Organisms/Strains
Mouse: C57BL/6 J Jackson Laboratories JAX: 000664
Mouse: B6.RictorTm1loxP -Tg(Krt14-Cre)1Cgn/J
Bentzinger et al., 2008 Hafner et al., 2004
N/A
Mouse: B6.RictorTm1loxP -Krt19tm1(cre/ERT)Ggu/J
Bentzinger et al., 2008 Means et al., 2008
N/A
Mouse: B6-Tg(CAG-EGFP)1Osb/J
Riedl et al., 2010 N/A
Oligonucleotides
Primers for qPCR, see Table S1 This paper N/A
Software
Fiji Schindelin et al., 2012 http://imagej.net/Fiji
Graphpad Prism 8 GraphPad N/A
TraceFinder v.4.1 Thermo Fisher Scientific N/A
FreeStyle 1.6 Thermo Fisher Scientific N/A
Mass Lynx v4.1 SCN 950 Waters N/A
TargetLynx XS v 4.1 SCN 959 Waters N/A
Ingenuity Pathway Analysis Qiagen N/A
Panther Mi et al, 2019 Pantherdb.org
FlowJo Becton, Dickinson and Company Flowjo.com
Leica LAS X Leica N/A
BD FACSDivaTM Software Becton, Dickinson and Company N/A
Supplementary Figure Legends
Supplementary Fig. S1 (Related to Fig. 1) Characterization of in vivo and 3C organoid HFSCs
(A) Representative immunofluorescence images of 3C cell cultures grown for 12 days
and stained with antibodies against Periostin and Keratin 6 (top row), Periostin and
Keratin 16 (2nd row), CD34 and Sox9 (3rd row) or Periostin and CD34 (bottom row).
Scale bars 50 µm.
(B) Freshly isolated FACS-purified epidermal cells: CD34+/α6+ (HFSCs), CD34-
/α6+/Sca1- (HF progenitors), CD34-/α6+/Sca1+ (IFE progenitors) and 3C cultured,
FACS purified 3C-CD34+/α6+ and 3C-CD34-/α6+ cells were subjected for RT-qPCR
analysis of ORS (Sox9, Barx2) and IFE (Il1rs2, Ly6d) identity genes (mean ±SEM; n=4
mice/condition; *P< 0.05, ANOVA/Fisher’s).
(C) FACS purified 3C-CD34+/α6+ and 3C-CD34-/α6+ cells were subjected for RT-
qPCR analysis of a panel of wound-induced genes (mean ±SEM; n=4 mice/condition;
ns=not significant, Mann-Whitney).
(D) Representative images and quantification of colonies formed by FACS purified 3C-
CD34+/α6+ and 3C-CD34-/α6+ cells plated on feeder cells (n=6 cultures from 2 mice;
**P< 0.0025, Student’s t-test).
(E) Freshly isolated epidermal cells and 3C cultures grown for 14 days were FACS
sorted into pure populations of CD34+/α6+ and CD34-/α6+ for qPCR analysis of genes
involved in mitochondrial biogenesis (mean ±SD; n=4 mice/condition; *P< 0.05, Mann
Whitney).
(F) Relative amounts of key glycolysis metabolites quantified from CD34+/α6+ and
CD34-/α6+ cells by LC-MS, normalized to CD34- cells (mean ±SD; n=4 mice/condition).
Supplementary Fig. S2 (Related to Fig. 2) Redox metabolite ratios from HFSCs and progenitors
(A-C) LC-MS quantification of key redox metabolite ratios from FACS-purified CD34+/-
cells (mean ±SD; n=4 mice/condition; **P< 0.05, * P< 0.05, Mann Whitney).
Supplementary Fig. S3 (Related to Fig. 3) Effects of low pO2 on HFSCs
Supplemental Text and Figures
(A) qRT-PCR analysis of key HFSC signature genes from FACS purified CD34+/- cells
from 3C cultures in 20% or 2% O2 (mean ±SD, n=3 mice/condition; *p < 0.05, Mann-
Whitney, all statistical comparisons are to 20% O2 CD34-).
(B) Quantification of EdU incorporation and AnnexinV staining by flow cytometry. 3C
cultures in 20% or 2% O2 were incubated with 10µM EdU for 24h. (n=3-4 mice).
(C) Work flow and representative immunofluorescence images of 3C cultures stained
with antibodies against CD34 and visualized for GFP expression. GFP+/CD34+/α6+
and GFP-/CD34-/α6+ cells from LifeAct-GFP and wild type mice were FACS purified,
mixed in 1:1 ratio and cultured for 12 days under 20% or 2% O2 before fixation and
immunofluorescence analyses (n=3 mice/condition).
Supplementary Fig. S4 (Related to Fig. 4) Low pO2 and metabolism control HFSC state
(A) Representative western blot analysis of HIF1α from 3C organoids subjected to 12h
of 20% and 2% O2.
(B) RT-qPCR analysis of key HIF1α target genes from 3D-3C cultured cells subjected
to 12h of 20% and 2% O2 followed by FACS-purification of CD34+/- cells (mean ±SD,
n=4 mice/condition; *P< 0.05 **P<0.0097, ANOVA/Fisher’s).
(C) FACS analyses of HFSC content from 10-day 3C cultured cells treated with 100
µM CoCl2 for 3 days (n= 4 mice/condition).
(D) Absolute numbers of viable cells from 3C cultures treated with ETC inhibitors
Antimycin A (AA) or Rotenone (Rot) (mean±SD; n=3 mice/condition).
(E) FACS analyses of HFSC content from 10-day 3C cultured cells treated with 350
µM UK5099 for 2 days (n= 4 mice/condition).
(F) FACS analyses of HFSC content from 10-day 3C cultured cells treated with 100
µM 2-DG for 2 days (n= 3 mice/condition).
(G) Quantification of HFSC content from 3C cultures where the culture medium was
supplemented with either 15 mM glucose or 15 mM galactose for 12 days (n=3 mice).
(H) Quantification of EdU incorporation from 10-day 3C cultured cells treated with FX11
for 2 days. EdU was added 12 h prior to sample collection (n = 4 mice/condition).
(I) LC/MS quantification of glutamate and glutamine concentrations from FACS-purified
CD34+/- cells (mean±SD; n = 3 mice/condition; *P< 0.05; Mann Whitney).
Supplementary Fig. S5 (Related to Fig. 5) TORC1 and TORC2 in HFSC state regulation
(A) RT-qPCR analyses of key HFSC signature genes from CTR and RictorEKO cells
cultured for 14 days in either 20% or 2% O2 (mean±SD; n=3 mice condition; *p < 0.02,
**p < 0.005, ns=not significant; ANOVA/Holm-Sidak).
(B) Representative immunofluorescence images of cells isolated from control (CTR)
and RictorEKO mice, 3C cultured for 12 days, and stained with antibodies against
Periostin and Keratin 6, Periostin and Keratin 16 or Periostin and Sox9. Scale bars 50
µm.
(C) Representative immunofluorescence images of FACS-purified CD34+/α6+ and
CD34-/α6- cells isolated from CTR and RictorEKO mice, 3C cultured for 7 days and
stained with antibodies against Periostin and Keratin 6. Scale bars 50 µm.
(D) FACS analyses of HFSC content from 10-day 3C cultured cells treated with
indicated concentrations of rapamycin for 2 days (n=3 mice/condition).
(E) FACS analyses of HFSC content from 10-day 3C cultured cells treated with SC79
for 3 days to activate Akt (n=4 mice/condition; P=0.052, Mann-Whitney).
Supplementary Fig. S6 (Related to Fig. 6) Glutamine metabolism controls HFSC state
(A) Quantification of EdU incorporation from 10-day 3C cultured control (CTR) and
RictorEKO cells treated with 20 nM CB839 glutaminase inhibitor for 2 days. EdU was
added 24 h prior to sample collection and analyzed by flow cytometry (n = 4 (CTR)/3
(RictorEKO) mice; *P<0.015, ANOVA/Holm-Sidak).
(B) Flow cytometry analyses of Annexin-positive apoptotic cells from 10-day 3C
cultured control (CTR) and RictorEKO cells treated with 20 nM CB839 for 2 days (n = 4
(CTR)/3 (RictorEKO) mice).
(C) FACS-purified CD34+ cells from RicEKO and CTR mice cultured in 3C for 3 days
with or without CB839 after which CD34+ cells (%) were analyzed via FACS (n=4
mice/condition). ns=not significant
Supplementary Fig. S7 (Related to Fig. 7) Rictor regulates HFSC and hair follicle cycling
(A) Representative images of H/E stained control (CTR) and RictorEKO skin tissue at
P69. Scale bars 200 µm.
(B) Representative images and quantification of back skin whole mounts stained for
CD34 (magenta) and DAPI (blue) to identify bulges. Note two bulges in CTR but not in
and RictorEKO hair follicles (white arrows; n=4 mice/genotype; *p=0.0286, Mann-
Whitney, scale bars 100 µm).
(C) Representative UV and brightfield images of back skin of CTR and RictorEKO mice
dyed with UV-light reactive hair dye during their first telogen hair cycle (P 21) and
imaged at indicated time points (n=3 mice/genotype).
(D) Quantification of BrdU-positive cells in ORS control (CTR) and RictorEKO mice (n=4
mice/genotype).
(E) Representative images of anagen hair follicles stained for Barx2 and CD34 to mark
the hair follicle bulge and ORS. Scale bars 100 µm.
(F, G) Representative images (F) and quantification (G) of Tunel staining to identify
apoptotic cells (white arrows) in catagen hair follicles shows no difference between
CTR RicEKO (n=3 mice/genotype; scale bars 100 µm).
(H) RT-qPCR analyses of Rictor gene expression in FACS-purified CD34+/a6+ and
CD34-/a6+ cells from CTR and Rictor-K19-CreERT mice (mean±SD; n = 4
mice/genotype; *P< 0.05; Ratio paired t-test).
(I, J) Representative images (I) and quantification (J) of hair cycle stage in back skin
of mice entering second postnatal telogen (n = 5 mice/genotype).
(K, L) Representative images of H/E staining (K) and quantification (L) of CTR and
Rictor-K19-CreERT hair follicles with 2 bulges (n = 4 mice; *P< 0.05; Mann Whitney;
scale bars 100 µm).
(M) Representative images of Hypoxyprobe DAB staining in telogen and anagen hair
follicles shows intense signal in bulge and upper ORS of anagen hair follicles (n=2
mice/stage; scale bars 100 µm).
(N) Quantification of hair follicle (HF) density of CTR and RicEKO mice at indicated time
points (n= 3-4 mice/time point).
(O) Quantification of CD34+ HFSCs in CTR and RicEKO mice at P52 (n= 7 mice (CTR)/6
mice (RicEKO)).
0
1
2
3
4
3PG Pyruvate Lactate Phosphoenol Pyruvate
A
TFAM PCG1a0
0.2
0.4
0.6
0.8
1.0
Fol
d C
hang
e
*
*
TFAM PGC1a
* *
Periostin K6 DAPI Merge
Periostin K16 DAPI Merge
CD34 Sox9 DAPI Merge
Periostin CD34 DAPI Merge
In vivo 3C cultureE
0
0.2
0.4
0.6
0.8
1.0
Fol
d C
hang
e
CD34- α6+
CD34+ α6+
FCD34- α6+
CD34+ α6+
Acetyl CoA
SUPPLEMENTARY FIGURE S1
Sox9 Barx2 Il1r2 Ly6d0
2
4
6
10
15
20
25
Fol
d ch
ange
CD34+/α6+ (HFSC)
CD34-/α6+/Sca1- (HF)
CD34-/α6+/Sca1+ (IFE)
3C-CD34+/α6+
3C-CD34-/α6+
ORS IFE
Tgln2 Anxa2 Thbs Tnc0
0.5
1.0
1.5
2.0
Fol
d ch
ange
3C-CD34+/α6+3C-CD34-/α6+
BC
*
*
D
3C-C
D34
+/α
6+3C
-CD
34-/α
6+
0
0.5
1.0
1.5
Colony size3C-CD34+/α6+3C-CD34-/α6+
0
5
10
15
20
25
Colony number
Col
ony
num
ber/
wel
l
Sur
face
are
a (m
m2 )
****
**
**
*
*
ns
ns
ns
ns
In vivo
Fol
d C
hang
e
CD34- CD34+0
0.02
0.04
0.06
0.08
FA
D/F
AD
+N
AD
H
**
A
CD34- CD34+
B C
CD34- α6+CD34+ α6+
AMP/ATP ATP/ADP0
0.5
1.0
1.5
Rat
io
SA
M/S
AH
Rat
io
0
20
40
60
*
*
*
SUPPLEMENTARY FIGURE S2
In vivo
0
20
40
60
80
100
EdU
+ c
ells
(%
)
0
2
4
6
8
10
Ann
exin
+ c
ells
(%
)
A B
20% O2 2% O2 20% O2 2% O2 20% O2 2% O220% O2 2% O2
0
20
40
60
80
100
EdU
+ C
D34
+ c
ells
(%
)
0
20
40
60
80
100
EdU
+ C
D34
- ce
lls (
%)
Dkk3 Id2 Nfatc1 Tcf30
2
4
6
8
10
Fol
d C
hang
e
C
20%
O2
2% O
2
FACS purification
GFP-CD34− α6+
Mix 1:1
3C culture 14 days
ImmunofluorescenceGFP / CD34
GFP+CD34+ α6+
GFP CD34 DAPI Merge
20% O2 / CD34- α6+
20% O2 / CD34+ α6+
2% O2 / CD34- α6+2% O2 / CD34+ α6+
*
*
*
**
*
**
SUPPLEMENTARY FIGURE S3
BA
Glutamate Glutamine0
1
2
3
4
5
Fol
d ch
ange
**
E
3C FX110
20
40
60
80
100
Ed
U+
cel
ls (
%)
3C-CD34+ α6+
3C-CD34- α6+
CD34+ α6+
CD34- α6+
3C+UK5099
3C-Ctrl0
20
40
60
80
100
CD
34+ɑ
6+ c
ells
(%
)
*
Hif1α
Actin
pO2:
C
*
D
Fol
d c
hang
e
G
3C AA Rot0
50000
100000
150000
200000
Live
cel
lsF
3C-Glucose
3C-Galactose
0
20
40
60
80
100
CD
34+ɑ
6+ c
ells
(%
)
3C+ 2-DG3C-Ctrl0
20
40
60
80
100
CD
34+ɑ
6+ c
ells
(%
)
H I
SUPPLEMENTARY FIGURE S4
VEGFPGF
GLUT1
PAI1
MM
P2
PDK2GLS
0
0.5
1.0
1.5
2.0
10
20
30
4020% O2 3C-CD34+20% O2 3C-CD34-
2% O2 3C-CD34+2% O2 3C-CD34-
20% 2%
3C-Crtl
3C-CoCl2
CD
34+ɑ
6+ c
ells
(%
)**
****
**
40
60
80
100
**
**
**
**
*
A 3
2
1
0
Fol
d ch
ange
CTR 20% O2
CTR 2% O2
RicEKO 20% O2
RicEKO 2% O2
Sox9 Nfac1 Dkk3
Periostin Keratin 6 DAPI Merge
Periostin Keratin 16 DAPI Merge
Periostin Sox9 DAPI Merge
CT
RR
icE
KO
CT
RR
icE
KO
CT
RR
icE
KO
B
Periostin Keratin 6 DAPI Merge
CT
RR
icE
KO
CT
RR
icE
KO
Pur
ified
CD
34- α
6+ c
ells
Pur
ified
CD
34+
α6+
cel
ls
C
D
CD
34+
α6+
cel
ls (
%)
0
20
40
60
80
3C 1 nM 10 nM
SUPPLEMENTARY FIGURE S5
3C+Rapamycin
40
60
80
100
CD
34+
α6+
cel
ls (
%)
3C 3C+SC79
p=0.0571E
CTRRicEKO
0
20
40
60
80
100
+ C
B83
9
EdU
+ c
ells
(%
) ns
*A B
CD34+ α6+ CD34- α6+ 0
5
10
15CTRRicEKO
Ann
exin
+ c
ells
(%
)
+ C
B83
9
CD34+ α6+ CD34- α6+
+ C
B83
9
+ C
B83
9
nsns
SUPPLEMENTARY FIGURE S6
C
20
40
60
80
100
CTR
RicEKO
CTR + CB839
RicEKO + CB839
ns
Day0 Day3
CD
34+
α6+
cel
ls (
%)
A
C D
CTR Telogen (P69) RICEKO Telogen (P69)
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Fol
d ch
ange
CTR
Ric-K19-CreERT
P26
P39
P48
P53
P65
UV Brightfield Merge UV Brightfield Merge
CTR RicEKO
p21 p32 p520
5
10
15
HF
/mm
CTR
M CD34- α6+CD34+ α6+
I
0
50
100
150catagen/anagen
telogen
CTR
Ric-K19-CreERT
Bac
k sk
in a
rea
(%)
CTR
Ric-K19
-Cre
CTR (P50) Ric-K19-CreERT (P50)
CTR Ric-K19-CreER0
20
40
60
80
100
*
((( )))K
CTR
0
10
20
CD
34+
cel
ls (
%)
O
RicEKO
RicEKO
*
0
20
40
60
80
100
HF
with
2 b
ulge
s (%
)
DA
PI C
D34
CTR Telogen (P62) RicEKO Telogen (P62)B
*
CTR RicEKO
0
5
10
15
20
Brd
U+
cells
in O
RS
CTR RicEKO
E
HF
with
2 b
ulge
s (%
)
CTR Catagen (P40) RicEKO Catagen (P52)
DA
PI T
UN
EL
0
5
10
15
TU
NE
L +
cel
ls /
low
er H
F
CTR RicEKO
P46
P46
F
CT
R A
nage
n (P
32)
Ric
EK
O A
nage
n (P
40)
BARX2 DAPISOX9
Bulge
ORS
ORSBulge
*
*
*
*
G
J L
H
SUPPLEMENTARY FIGURE S7
N
Neg
ativ
e C
TR
Hyp
oxyp
robe
(te
loge
n)
Hyp
oxyp
robe
(an
agen
)
ORS
Bulge
Supplementary Table S1 Primers for RT-qPCR (Related to STAR Methods) Gene Forward Sequence Reverse Sequence
S18 GATCCCAGACTGGTTCCTGA GTCTAGACCGTTGGCCAGAA
Actb TCAAGATCATTGCTCCTCCTG TACTTCTGCTTGCTGATCCAC
Gls CACACCCCGTCCCGGACTTTTTC TTAAGAGCAGCTCCCGTAGCATC
Sox9 AGGAAGCTGGCAGACCAGTA TCCACGAAGGGTCTCTTCTC
Dkk3 ATGCTATGCACCCGAGACAG GAACAGCAGGCCTCTTTGGA
Tcf3 CTCAGCAGCAAATCCAAGAGGCAGAG TGGGAAGACGCAGGGCTATCACAAG
Nfatc1 GGTGCTGTCTGGCCATAACT CCAGGGAATTTGGCTTGCAC
Id2 ATCCCCCAGAACAAGAAGGT TGTCCAGGTCTCTGGTGATG
TFAM AAGGATGATTCGGCTCAGG GGCTTTGAGACCTAACTGG
PCG1a AGCCGTGACCACTGACAACGAG GCTGCATGGTTCTGAGTGCTAAG
Rictor GAGAAAGCTGGGCCATCTGA AACCCGGCTGCTCTTACTTC
Vegfa ACGTCAGAGAGCAACATCAC CTGTCTTTCTTTGGTCTGCATTC
Glut1 CTTCCAGTATGTGGAGCAACT CTCGGGTGTCTTGTCACTTT
Pdk2 TACTTCCTGGACCGCTTCTA CAAAGATGAGGGTGTGTTGATTG
Pai1 CTCCACAGCCTTTGTCATCT ATTGTCTCTGTCGGGTTGTG
Mmp2 CAGGGAATGAGTACTGGGTCTATT ACTCCAGTTAAAGGCAGCATCTAC
Pgf TGCTGGTCATGAAGCTGTTC GGACACAGGACGGACTGAAT
Barx2 ACTCAGCTGCAAGTGAAGAC CTGTCCACCTTTAAGGACCATC
Ly6d ACCGTCACCTCAGTGGA CTCTCGTTGCATAGGTCAGTC
Il1r2 CTCCAGCAGTTCCCATAGTT GTGGGATGCGTTTCTGAATG
Anxa2 CAAGACCAAAGGAGTGGATGAG CTGGGCAGGTGTCTTCAATAG
Tagln2 AGAGGTGCAGCAGAAGATTG AAGCTTGCACAGAACCGT
Tnc TGGTCTTCTGATAGCCAACATC GTGTAACTTCTGGCACTCTCTC
Thbs1 CCTGGACTTGCTGTAGGTTATG GTAGGACTGGGTGACTTGTTTC
Ercc1 AGTAGACGGCCATAGCAGGA TTCGGGTCAACATCGAGCAA
Ercc2 AGCTGGCTCTTTAATGGGAGG CCTGCAATGGGTCTCCAACA