Exosomes in uence the engraftment of tumor cell lines

in athymic mice BALB/c nude

Roman A. Kovalev

, Vladimir S. Burdakov

, Elena Yu. Varfolomeeva

, Elena V. Semenova

and Michael V. Filatov

1,2

Division of Molecular and Radiation Biophysics, National Research Center “Kurchatov Institute” B.P.

Konstantinov Petersburg Nuclear Physics Institute; Gatchina, Russia;

Saint-Petersburg State Research Institute of Phthisiopulmonology of the Ministry of Healthcare of the

Russian Federation; St. Petersburg, Russia

ABSTRACT

The strain of mice with immunode cit provides a popular experimental model in oncological, immunological, and

transplantation studies. Athymic nude mice (Nudes) are usually used to explore the engraftment, growth, invasive in l-

tration, and potency to develop metastases for different types of cancer, as well as to test in preclinical assessment new

anti-tumor medical products. Nude-mutation in Foxn1 gene results in the congenital thymus dysgenesis, and, as a con-

sequence, in the lack of T-lymphocytes. In its turn, it leads to the suppression of the immune functions which involve

T-cells. It is generally accepted that Nudes’ immune de ciency provides an opportunity to engraft in them immunologi-

cally incompatible human tumors that are xenogeneic for mice.In our study we have shown that there are some types

of tumors that cannot be engrafted in BALB/c Nude mice. In particular, we have discovered that tumor cells de cient in

p53 gene are not able to grow in Nude mice.Furthermore, we have used in vitro model system to demonstrate that the

exosomes released by the  broblasts of BALB/c Nude mice suppress the growth of the very same tumor cell lines that

cannot be engrafted in these mice, but the exosomes do not affect the growth of the tumor cells that can be engrafted

successfully. One of the possible mechanisms of tumor growth suppression in BALB/c Nude mice is the transfer of the

exosomes containing wild type p53 protein from the surrounding tissues to tumor cells with gene p53 damage.

KEY WORDS: BALB/C NUDE MICE, CELL LINES, EXOSOMES, PROTEIN P53

535

ARTICLE INFORMATION:

Corresponding Authors:  l_53@mail.ru

Received 10

Oct, 2018

Accepted after revision 11

Dec, 2018

BBRC Print ISSN: 0974-6455

Online ISSN: 2321-4007 CODEN: USA BBRCBA

Thomson Reuters ISI ESC / Clarivate Analytics USA

Mono of Clarivate Analytics and Crossref Indexed

Journal Mono of CR

NAAS Journal Score 2018: 4.31 SJIF 2017: 4.196

Online Contents Available at: http//www.bbrc.in/

DOI: 10.21786/bbrc/11.4/1

Medical

Communication

Biosci. Biotech. Res. Comm. 11(4): 535-540 (2018)

Roman A. Kovalev et al.

536 EXOSOMES INFLUENCE THE ENGRAFTMENT OF TUMOR CELL LINES IN ATHYMIC MICE BALB/C NUDE BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS

INTRODUCTION

Experimental models of human diseases play an impor-

tant role both in understanding what biological and

genetic factors in uence phenotypic characteristics

of the disease and in developing the strategy for its

treatment.When using in vitro cell models for human

tumor research one should take into consideration that,

though these studies are necessary, they are reduction-

ist, as often the parameters of the cell culture cannot be

equated with the tumor growth inside the host’s body

where the tumor has speci c micro-environment. Mod-

eling human cancer with the help of genetically  awed

mice is an alternative way to study carcinogenesis in

vivo. It provides an opportunity to detect cell and molec-

ular changes taking place at the initiation of cancer and

during its further progression (Stakleff, 2003, Shultz,

2014; Boone, 2015).

One of the mechanisms preventing formation and

growth of the malignant tumors in human body is

immune system, and in particular, cell immunity based

on the activity of T-lymphocytes attacking tumor cells.

To research anti-tumor activity of the immune system

the mice with de cient T-cell immunity are widely used.

Athymic mice line BALB/c Nude is generally accepted as

a model for this kind of experiments. The main advan-

tage of this experimental model is natural immunode-

pression resulting from the functional deletion in Foxn1

gene. As the result of the deletion, thymus is either dete-

riorated or absent, and, as a consequence, the number

of T-cells is negligible, and, therefore, cell immunity is

inhibited (Pignata,1996; Frank, 1999; Zuklys, 2016). In

cancer research a large number of various tumor cell

lines are used for inoculation hence, it is comparatively

easy to engraft a tumor being studied in the mice just

inoculating the host-animal with the tumor cells. It is

believed that BALB/c Nude mice can host even immu-

nologically incompatible, xenogneic for mice human

tumors, (Liebman, 2007; Zeineldin, 2014; Conrad, 2015).

However, in our study we discovered that not all tumor

grafts could grow in BALB/c Nude immune de cient mice.

We were not able to reach a positive result of inoculation

with primary glioma lines developed in our laboratory.

The fact led us to the idea of an alternative mechanism

of tumor growth inhibition – the mechanism which did

not involve the immune system. If the mechanism exists,

then even though Nude mice lack T-lymphocyte mediated

cell immunity, the mechanism will account for inhibition

of the growth of inoculated human tumor cells.Taking

into consideration the growing amount of the data that

point to participation of exosomes in formation, develop-

ment and inhibition of cancer (Webber, 2015; Yu, 2015),

we suggested that the discovered phenomenon of tumor

growth inhibition in BALB/c Nude mice was possibly a

result of the in uence of the exosomes formed in the tis-

sues surrounding the tumor.

Exosomes are extracellular membrane nano vesicles

that are released from multivesicle bodies into extracel-

lular space through exocytosis (Huang, 2013). The abil-

ity of the exosomes to be ef cient transporters of exog-

enous proteins and RNA into cells-recipients has been

reported (Shtam, 2013; Haney, 2015; Srivastava, 2016).

Earlier in our experiments in vitro we demonstrated that

the exosomes released from the cells, which contained

wild type of the p53 protein, were able to inhibit the

growth of the cancer cells lacking the same protein. Fur-

thermore, the exosomes both released by the cell lines

and isolated from human blood plasma carried one of

the main tumor suppressors - protein p53 (Burdakov,

2017; Jorgensen, 2015, Burdakov, 2017).

In this study we now present some arguments in

favor of our assumption that one of the factors inhibit-

ing inoculated tumor growth in BALB/c Nude athymic

mice may be negative effect of exosomes on tumor cells.

The effect may be based on the reaction of the inocu-

lated tumor cells de cient in p53 gene to the wild type

p53 protein, which exists in exosomes released by the

surrounding tissues.

MATERIALS AND METHODS

CELL LINES AND CULTIVATION CONDITIONS

In our study we used both transplantable and primary

human cell lines: HT-1080 ( brosarcoma), Gl-V, Gl-R,

Gl-Sh (primary cultures of glioma cells, developed in

our laboratory), transplantable culture of rat cells C6 (rat

brain glioma), and  broblasts of BALB/c Nude mice. Cells

were cultivated in DMEM/F12 medium (Biolot, Russia)

supplemented with 10% fetal bovine serum (Biolot, Rus-

sia), without antibiotics, under 5% CO

at 37°C. When

tumor cells of each of the lines were cultivated with iso-

lated exosomes, the exosomes had  nal concentration in

the culture of 10

13

–10

exosomes/ml.

Isolation of exosomes in the systems in vitro

Conditional medium was collected from growing cell cul-

ture. Then, it was consequently centrifuged at 2,000 g and

at 20,000 g to eliminate dead cells and their fragments.

The procedure was repeated until the amount of the puri-

 ed conditional medium reached 500 ml. This amount

of collected puri ed conditional medium was used to

isolate exosomes by ultracentrifuging. Beckman coulter

ultracentrifuge was used (45Ti rotor) at 100,000 g for 2

hours. After that, the residue was suspended in 100 ml of

PBS and underwent the second ultracentrifugation at the

same conditions. The exosomes obtained as the result of

the second sedimentation were used in the experiments.

Roman A. Kovalev et al.

BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS EXOSOMES INFLUENCE THE ENGRAFTMENT OF TUMOR CELL LINES IN ATHYMIC MICE BALB/C NUDE 537

Evaluation of the cell survival rate

The cells were suspended in Versene/Trypsin solution

(Biolot, Russia), and the survival rate was evaluated

through direct calculation of the cells with the help of

vision-based automated cell counter Scepter (Millipore,

USA). To visualize cell survival, the cells of Gl-V, Gl-R,

Gl-Sh and C6 lines were seeded into 12-well plates in the

concentration equal for all the cell lines (10,000 cells per a

well). After that we added equal amounts of the exosomes

harvested from the conditional medium of the  broblasts

of BALB/c Nude mice to all the cell lines. As the control

cells formed a monolayer, all the cells in the well plates

were stained with Crystal Violet (Fluka AG, Germany). The

data were analyzed by Student’s t test. Data are shown as

the mean ± SEM (standard error of the mean).

Animal care and engraftment rate evaluation

Adult BALB/c Nude mice, 7-8 weeks of age and 18-20

grams of bodyweight, were purchased from the laboratory

animal breeding center “Pushchino” (Pushchino, Russia).

All procedures for mouse care and use were conducted

in accordance with the National Standard of the Rus-

sian Federation GOST R 53434-2009: Principles of good

laboratory practice (introduced 01.03.2010 by Federal

Agency for Technical Regulation and Metrology, pub-

lished by Standardinform, Moscow, 2010). The protocol

was approved by scienti c committee of the Division of

Molecular and Radiation Biophysics of National Research

Center “Kurchatov Institute” B.P.Konstantinov St Peters-

burg Nuclear Physics Institute (Gatchina, Russia).

Animals were housed in groups of three per cage

(males and females were housed separately) with ad libi-

tum access to food and water in standard polycarbonate

cages (width: 300mm, depth: 400mm, height: 200mm)

and maintained under standard conditions of tempera-

ture (22 ± 2°C), humidity (55 ± 10%), arti cial 12-hour

light-dark cycle (lights on at 8:00 A.M., lights off at 8:00

P.M.). The mice were fed on the nutrient extruded and

granulated food designed for experimental rodents (LLC

«Laboratorkorm», Russia).

Tumor inoculations were performed through intra-

muscularly injections of 150,000 to 1,000,000 tumor

cells. To estimate the engraftment rate for each tumor

cell line, not less than 20 BALB/c Nude immune de -

cient athymic mice were used.

RESULTS AND DISCUSSION

Evaluation of engraftment in BALB/c Nude immune

de cient athymic mice for tumor cell lines

To evaluate the engraftment rate we took 5 different

tumor cell lines, which were cultivated in our laboratory.

Two of them were standard transplantable cell lines:

HT-1080, C6. The rest three lines (GL-V, GL-R, GL-Sh)

were primary lines of human gliomas, developed in our

laboratory. In the experiments with each of the cell lines

not less than 20 BALB/c Nude immune de cient athymic

mice were used. In the  rst experiment engraftment

was performed through intramuscularly injections of

150,000 tumor cells per animal. As the result there were

two polar outcomes: transplantable tumor lines HT-1080

and C6 engrafted in 100% of the animals; tumor lines

GL-V, GL-R did not engraft at all ( g. 1). Engraftment

rate of the latter lines (GL-V, GL-R) did not change with

the increase in the amount of inoculated cells up to

1,000,000 cells per animal.

Analysis of survival rate of the tumor cell lines after

treatment with the exosomes from the  broblasts of

BALB/c Nude mice

We assumed that negative results of engraftment of

some of the tumors in BALB/c Nude athymic mice

stemmed from the ability of exosomes to inhibit tumor

cell growth. To con rm the assumption, we decided to

compare the engraftment rate obtained in the experi-

ments with the survival rate of the same tumor cell lines

after they had been treated with the exosomes from the

 broblasts of BALB/c Nude mice. To carry out the com-

parison, we used standard ultracentrifugation methods

described previously, (Lässer et al., 2012 and Shtam et al

2013) to isolate exosomes from the condensed cultural

medium after growing  broblasts of BALB/c Nude mice

in it. Cells from the analyzed lines were seeded onto

24-well plates in the equal concentration. Then, equal

amounts of the isolated exosomes were added to all of

the tested tumor cell lines. In 10 days after the exosomes

had been added, the number of living cells was calcu-

lated for each of the studied tumor cell lines.

The results of the calculation are presented in  g. 2.

Addition of the exosomes isolated from the  broblasts

of BALB/c Nude mice did not affect the survivability

of the tumor cells HT-1080 and GL-Sh. However, exo-

some addition to the cells of the tumor lines GL-R и

GL-V resulted in signi cant decrease in their survivabil-

ity ( g. 2). Survivability of the glial cells of GL-R cell

line decreased by almost three times, and survivability

of the cells of the cell line GL-V decreased by more than

10 times. The obtained results point to the ability of the

BALB/c Nude cell exosomes to inhibit the growth of the

target cells GL-R and GL-V in in vitro experiments. It

correlates with the negative results of in vivo inocula-

tion of the cells of these tumor lines to BALB/c Nude

immune de cient mice.

Exosomes, participating in inter-cellular communica-

tion by transporting RNA and speci c proteins between

cells, are able to cause both functional and epigenetic

Roman A. Kovalev et al.

538 EXOSOMES INFLUENCE THE ENGRAFTMENT OF TUMOR CELL LINES IN ATHYMIC MICE BALB/C NUDE BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS

FIGURE 1. Engraftment rates of the tumor cell lines in BALB/c Nude immune

de cient athymic mice

(a) Table of engraftment rates of the tumor cell lines C6, HT-1080, GL-Sh, GL-R

и GL-V.

(b) Examples of the engrafted tumors HT-1080 and GL-Sh.

FIGURE 2. Survival rates of the tumor cell lines C6, HT-1080, GL-Sh, GL-R and GL-V after treatment with the

exosomes from the  broblasts of BALB/c Nude mice

(a) Histograms of survival rates of the tumor cell lines after treatment with the exosomes from the  broblasts of

BALB/c Nude mice.

(b) Survival of the cell lines C6, HT-1080, GL-Sh, GL-R and GL-V after coculturing with the exosomes from the

 broblasts of BALB/c Nude mice demonstrated by means of Crystal Violet staining.

changes in the recipient cells (Grange, 2011; Kobayashi,

2014). Both discovery of p53 oncosupressor in the

exosomes of the human blood plasma and their ability

to affect proliferation of p53-negative cells con rm their

important role in control over oncogenesis (Burdakov,

2017; Jorgensen, 2015). Protein p53 is the main par-

ticipant of many signaling pathways through which cell

reactions to stress are regulated. The protein is able to stop

cell-division cycle and/or to cause apoptosis, thus pre-

venting uncontrollable cell division – reproduction of the

cells with damaged genome. In human malignant tumors

p53 gene is the one that most often is found mutated

BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS EXOSOMES INFLUENCE THE ENGRAFTMENT OF TUMOR CELL LINES IN ATHYMIC MICE BALB/C NUDE 539

Roman A. Kovalev et al.

(Levine, 2009; Vogelstein, 2013; Kandoth, 2013). There

are a number of mouse model studies demonstrating that

restoration of the function of wild type p53 may result in

tumor regression (Ventura, 2007; Xue, 2007).

We assume that negative effect of the exosomes con-

taining wild type p53 protein on inoculated tumor cells

may be one of the factors which prevent the engraftment

of the tumor in BALB/c Nude athymic mice.

It is a common knowledge that any experimental

model based on the usage of the athymus Nude mice

has its restrictions: immunode ciency is severe, but not

absolute (there still exists humoral adaptive immune

system and undamaged congenital immunity). Hence,

despite almost complete lack of functioning T-lympho-

cytes, both congenital immune response and high activ-

ity of NK-cells are able to restrict the speed of engraft-

ment and to decrease metastatic potency of majority of

tumors (Shultz, 2005; Shultz, 2014; Szadvari, 2016).

However, the data on the engraftment rate of differ-

ent tumor cell lines we present in this study point to a

clear division of the tumors into two groups: the tumors

engrafting in BALB/c Nude mice and the tumors which

cannot be engrafted in BALB/c Nude mice. Basing on

the vast statistics we discovered no variations in the

engraftment rate for each of the tumors used in the study

( g.1a). These facts cannot be explained only by exist-

ence of the rudimentary immunity in immune de cient

mice. We assume that one of the possible mechanisms of

the engraftment inhibition involves systemic in uence

of mouse’s own exosomes on the inoculated tumor cells.

It is necessary to remark that the cells with the wild type

of p53 protein (HT-1080) had 100% engraftment rate,

while GL-V cells lacking endogenous p53 protein (Kova-

lev, 2015) could not be engrafted at all.

Then we examined the sensitivity of all the tumor cell

lines used for inoculation to the in uence of the exosomes

isolated from the  broblasts of BALB/c Nude mice. The

results we received in the experiments in vitro showed

that cells of the GL-R cell line and especially p53-nega-

tive GL-V line were sensitive to such in uence: the cells

from abovementioned tumor cell lines died as the result of

addition of the exosomes from the  broblasts of BALB/c

Nude mice to the tumor cell culture.The results obtained

support the hypothesis that exosomes circulating in the

body are able to perform a defense function controlling

oncogenesis through onco suppressor p53.

Con ict of interest disclosures: We have no con ict of

interest in any part of this article.

REFERENCES

Boone JD, Dobbin ZC, Straughn JM jr., Buchsbaum DJ: (2015)

Ovarian and cervical cancer patient derived xenografts: The

past, present, and future. Gynecol Oncol 138: 486-491.

Burdakov VS, Kovalev RA, Pantina RA, Varfolomeeva EYu,

Makarov EM., Filatov MV: (2017) Exosomes transfer p53

between cells and can suppress growth and proliferation of

p53-dR. Tsitologiia 59 (9): 588-594.

Conrad SJ, El-Aswad M, Kurban E, Jeng D, Tripp BC, Nutting

C, Eversole R, Mackenzie C, Essani K: (2015) Oncolytic tana-

poxvirus expressing FliC causes regression of human colorec-

tal cancer xenografts in nude mice. J Exp Clin Cancer Res 34

(1): 19.

Frank J, Pignata C, Panteleyev AA, Prowse DM, Baden H,

Weiner L, Gaetaniello L, Ahmad W, Pozzi N, Cserhalmi-Fried-

man PB, Aita VM, Uyttendaele H, Gordon D, Ott J, Brissette JL,

Christiano AM

: (1999) Exposing the human nude phenotype.

Nature 398 (6727): 473-474.

Grange C, Tapparo M, Collino F, Vitillo L, Damasco C, Deregi-

bus MC, Tetta C, Bussolati B, Camussi G: (2011) Microvesicles

released from human renal cancer stem cells stimulate angio-

genesis and formation of lung premetastatic niche. Cancer Res

71: 5346-5356.

Haney M J, Klyachko NL, Zhao Y, Gupta R, Plotnikova EG,

He Z, PatelT, Piroyan A, Sokolsky M, Kabanov AV, Batrakova

EV: (2015) Exosomes as drug delivery vehicles for Parkinson’s

disease therapy. J Control Release 207: 18-30.

Huang SH, Li Y, Zhang J, Rong J, Ye S: (2013) Epidermal

growth factor receptor-containing exosomes induce tumor-

speci c regulatory T cells. Cancer Invest 31: 330-335.

Jørgensen MM, Bæk R, Varming K: (2015) Potentials and capa-

bilities of the extracellular vesicle (EV) array. J Extracell Vesi-

cles 4: 26048.

Kandoth C, McLellan MD, Vandin F, Ye K, Niu B, Lu C, Xie M,

Zhang Q, McMichael JF, Wyczalkowski MA, Leiserson MDM,

Miller CA, Welch JS, Walter MJ, Wendl MC, Ley TJ, Wilson

RK, Raphael BJ, Ding L: (2013) Mutational landscape and

signi cance across 12 major cancer types. Nature 502: 333–

339.

Kobayashi M, Salomon C, Tapia J, Illanes SE, Mitchell MD,

Rice GE: (2014) Ovarian cancer cell invasiveness is associated

with discordant exosomal sequestration of Let-7 miRNA and

miR-200. J Transl Med 12: 4.

Kovalev RA, Shtam TA, Karelov DV, Burdakov VS, Volnits-

kiy AV, Makarov EM, Filatov MV: (2015) Histone deacetylase

inhibitors cause the TP53-dependent induction of p21/Waf1

in tumor cells carrying mutations in TP53. Tsitologiia 57 (3):

204-211.

Lässer C, Eldh M, Lötvall J: (2012) Isolation and characterization

of RNA-containing exosomes. J Vis Exp 59: e3037.

Levine AJ, Oren M: (2009) The  rst 30 years of p53: growing

ever more complex. Nat Rev Cancer 9: 749–758.

Liebman MA, Roche MI, Williams BR, Kim J, Pageau SC, Sha-

ron J: (2007) Antibody treatment of human tumor xenografts

elicits active anti-tumor immunity in nude mice. Immunol Lett

114 (1): 16–22.

Pignata C, Fiore M, Guzzetta V, Castaldo A, Sebastio G, Porta

Guarino A: (1996) Congenital Alopecia and nail dystrophy

540 EXOSOMES INFLUENCE THE ENGRAFTMENT OF TUMOR CELL LINES IN ATHYMIC MICE BALB/C NUDE BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS

Roman A. Kovalev et al.

associated with severe functional T-cell immunode ciency in

two sibs. Am J Med Genet 65 (2): 167–170.

Shtam TA, Kovalev RA, Varfolomeeva EYU, Makarov EM,

Kil YUV, Filatov MV: (2013) Exosomes are natural carriers of

exogenous siRNA to human cells in vitro. Cell Commun Signal

11: 88.

Shultz LD, Goodwin N, Ishikawa F, Hosur V, Lyons BL, Greiner

DL: (2014) Human cancer growth and therapy in NOD/

SCID/IL2Rnull (NSG) mice. Cold Spring Harb Protoc 694-

708.

Shultz LD, Lyons BL, Burzenski LM, Gott B, Chen X, Chal-

eff S, Kotb M, Gillies SD, King M,

Mangada J, Greiner DL,

Handgretinger R: (2005) Human lymphoid and myeloid cell

devel opment in NOD/LtSz-scid IL2Rnull mice engrafted with

mobilized human hemopoietic stem cells. J Immunol 174:

6477-6489.

Srivastava A, Babu A, Filant J, Moxley KM, Ruskin R, Dha-

nasekaran D, Sood AK, McMeekin S, Ramesh R: (2016) Exploi-

tation of exosomes as nanocarriers for gene-, chemo-, and

immune-therapy of cancer. J Biomed Nanotechnol 12: 1159-

1173.

Stakleff KDS, von Gruenigen VE: (2003) Rodent models for

ovarian cancer research. Int J Gynecol Cancer 13: 405-412.

Szadvari I, Krizanova O, Babula P: (2016) Athymic nude mice

as an experimental model for cancer treatment. Physiol Res 65

(Supplementum 4): 441-453.

Ventura A, Kirsch DG, McLaughlin ME, Tuveson DA, Grimm J,

Lintault L, Newman J, Reczek EE, Weissleder R, Jacks T: (2007)

Restoration of p53 function leads to tumour regression in vivo.

Nature 445: 661–665.

Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz

LA jr, Kinzler KW: (2013) Cancer genome landscapes. Science

339: 1546–1558.

Webber J, Yeung V, Clayton A: (2015) Extracellular vesicles as

modulators of the cancer microenvironment. Semin Cell Dev

Biol 40: 27-34.

Xue W, Zender L, Miething C, Dickins RA, Hernando E, Krizh-

anovsky V, Cordon-Cardo C, Lowe SW: (2007) Senescence and

tumour clearance is triggered by p53 restoration in murine

liver carcinomas. Nature 445: 656–660.

Yu S, Cao H, Shen B, Feng J: (2015) Tumor-derived exosomes

in cancer progression and treatment failure. Oncotarget 6:

37151-37168.

Zeineldin M, Jensen D, Paranjape SR, Parelkar NK, Jokar I,

Vielhauer GA, Neufeld KL: (2014) Human cancer xenografts in

outbred Nude mice can be confounded by polymorphisms in a

modi er of tumorigenesis. Genetics 197 (4): 1365–1376.

Zuklys S, Handel A, Zhanybekova S, Govani F, Keller M, Maio

S, Mayer CE, Teh HY, Hafen K, Gallone G, Barthlott T, Pont-

ing CP, Holländer GA: (2016)

Foxn1 regulates key target genes

essential for T cell development in postnatal thymic epithelial

cells. Nat Immunol 17: 1206-1215.