Defined tissue compartments of the hair follicle, namely the anagen hair bulb and the stem cell-harboring bulge zone, enjoy a relative state of immune privilege. The protection and restoration of hair follicle immune privilege remains the most fundamental prophylactic and therapeutic challenge in alopecia areata management.

1. Revisiting the role of
immune privilege (IP)
in alopecia areata (AA) pathobiology
Ralf Paus
University of Manchester, Manchester, UK
No relevant conflicts of interest

2. Alopecia areata (AA)
= Hair follicle (HF) cycling disorder
Perifollicular inflammatory infiltrates
1. Attack only anagen hair bulb
(anagen III-VI)
2. Catapult anagen HFs
prematurely into catagen
3. Central role of MHC class I-based
HF immune privilege (IP)
collapse, CD8+ T cells & IFNg
„No IP collapse, no AA“
 HF damage (dystrophy),
hair shaft breakage & shedding:
Hair loss
V Price

3. WCHR, Miami, Nov 2015: AA pathobiology
• Several different pathobiology pathways lead to AA,
which all coalesce in a stereotypic HF damage response,
the AA phenotype
• This AA response pattern also occurs in healthy HFs
in individuals not genetically predisposed to AA
Most essential prerequisite:
• Immune privilege (IP) collapse of anagen HFs
• AA = „territorial disease“  pathobiology can only be deciphered
by intracutaneous analyses focused on IP collapse & restoration

4. Physiological status:
Immune privilege (IP) of anagen HF
No MHC I
No ß2mg / TAP:
• No autoantigen
presentation
TGFß1/2, aMSH
cortisol, MIF:
• „IP guardians“
create immuno-
inhibitory milieu
Main IP pillars:
Ito T et al. AJP 2004
Billingham & Silvers JID 1971
Westgate et al. JID 1991
Paus et al. Yale J Biol Med 1993

5. Physiological status:
Immune privilege (IP) of anagen HF
No MHC I
No ß2mg / TAP:
• No autoantigen
presentation
TGFß1/2, aMSH
cortisol, MIF:
• „IP guardians“
create immuno-
inhibitory milieu
No MHC class II+ LCs:
antigen-presentation
to CD4+ TCs impaired
Christoph et al. BJD 2000
Low level of NKG2D ligands
(MICA, ULBP3):
Insufficient stimulation of
NKG2D+ cells
(NKs, NKTs, CD8)
Ito T et al. JID 2008
Petukhova et al. Nature 2010
Main IP pillars:
Ito T et al. AJP 2004
Billingham & Silvers JID 1971
Westgate et al. JID 1991
Paus et al. Yale J Biol Med 1993

6. Frontiers in HF immune privilege
No MHC I
No ß2mg / TAP:
• No autoantigen
presentation
TGFß1/2, aMSH, IGF-1,
SST, cortisol, MIF,
CGRP, PD-L1, VIP :
„IP guardians“
create immuno-
inhibitory milieu
No MHC class II+ LCs:
antigen-presentation
to CD4+ TCs impaired
Low level of NKG2D ligands
(MICA, ULBP3):
Insufficient stimulation of
NKG2D+ cells
(NKs, NKTs, CD8)
Immunoinhibitory
function of peri-
follicular mast cells
in HF physiology
Bertolini et al. PLoS ONE 2014
Kinori etal. EXD 2012, Breitkopf et al. JID 2013, Wang et al. JID 2014, Bertolini et al. BJD 2016

7. Role of mast cells (MCs) in AA ?

Lesional AA skin:
• much more &
hyperproliferating MCs
• more direct interatcions
between MCs & CD8+ TCs
Compared to non-lesional AA skin &
to healthy human scalp skin
Pink cells= MCs, Brown cells= CD8+ T-cells
Bertolini et al. PLoS ONE 2014
Collab.: K McElwee, A Gilhar et al.

8. MCs in healthy perifollicular
human scalp skin
MCs in lesional AA skin
Perifollicular MCs switch from an immuno-inhibitory
to a pro-inflammatory phenotype in lesional AA skin
MHCI
Heparin-
Histamine
c-Kit
PD-L1
OX40L
TGFβ-1
Tryptase
IL-10
CD30L
Mast Cell
4-1BBL
ICAM-1
c-Kit
MHCI
Heparin-
Histamine
4-1BBL
OX40L
PD-L1
IL-10
TGFβ-1
Tryptase
CD30L
Mast Cell
Ag?ICAM-1
Bertolini et al.
PLoS ONE 2014
 Does this MCs switch promote & maintain IP collapse ?
 Do these MCs regulate (auto-)antigen specific CD8+ T cells in AA ?
Mice: possible in principle  Stelekati/Paus/Bulfone-Paus et al. Immunity 2009
IP guardians up
IP guardians down
Profinflammatory signals up

9. Frontiers in HF immune privilege
No MHC I
No ß2mg / TAP:
• No autoantigen
presentation
TGFß1/2, aMSH, IGF-1,
SST, cortisol, MIF,
CGRP, PD-L1, VIP :
• „IP guardians“
create immuno-
inhibitory milieu
No MHC class II+ LCs:
antigen-presentation
to CD4+ TCs impaired
Low level of NKG2D ligands
(MICA, ULBP3):
Insufficient stimulation of
NKG2D+ cells
(NKs, NKTs, CD8, gdTCs)
 Y Uchida / M Bertolini
immunoinhibitory function
perifollicular mast cellsPeripheral tolerance
to anagen HF-associated
(auto-)antigens
Role of Tregs ?
 M Rosenblum
Oelert et al Exp Dermatol 2016
AIRE gene ?
Kumar et al. AJP 2011:
HF co-local isation with K17

10.  Multiple factors promote IP collapse in AA
MHC I, ß2mg, TAP :
(Auto-)antigen
presentation
TGFß1/2, aMSH, IGF-1
Cortisol, VIP-R, MIF :
Autoimmmunity-
promoting milieu
Mast cells
switch to
Proinflammatory
phenotype
Promotion of
autoantigen-specific
CD8+ T cells responses ?
Key role of excessive
IFNg signaling via JAK
IFNg-inducible chemokines
e.g. CXCL10, CXCR3
Ito T et al. JDS 2013; Dai et al JI 2016
Xing et al. Nat Med 2014

11.  Multiple factors promote IP collapse in AA
MHC I, ß2mg, TAP :
(Auto-)antigen
presentation
TGFß1/2, aMSH, IGF-1
Cortisol, VIP-R, MIF :
Autoimmmunity-
promoting milieu
Mast cells switch to
Proinflammatory phenotype
Promotion of autoantigen-specific
CD8+ T cells responses ?
Key role of excessive
IFNg signaling via JAK
Loss of peripheral tolerance
to HF (auto-)antigens
IFNg-inducible chemokines
e.g. CXCL10, CXCR3
Ito T et al. JDS 2013; Dai et al JI 2016
Xing et al. Nat Med 2014
Perceived stress &
neurogenic inflammation:
SP, mast cells
Peters et al. AJP 2007 (human)
Siebenhaar et al. JID 2007 (mice)
MHC II :
Secondary autoimmune responses
epitope spreading ?

12.  Multiple factors promote IP collapse in AA
MHC I, ß2mg, TAP :
(Auto-)antigen
presentation
TGFß1/2, aMSH, IGF-1
Cortisol, VIP-R, MIF :
Autoimmmunity-
promoting milieu
MHC II :
Secondary autoimmune
responses epitope spreading ?
„stressed“ HFs:
NKG2D ligands
MICA, (ULBP3): Li/Sinclair et al. JID 2016
Activation of
NKG2D+ cells
NKs, CD8+ TCs, gdTCs
Mast cells switch to
proinflammatory phenotype
Key role of excessive
IFNg signaling via JAK
Loss of peripheral tolerance
to HF (auto-)antigens
AA
IFNg-inducible chemokines
e.g. CXCL10, CXCR3
AIRE SNPs ?
Wengraf/Tazi-Ahnini 2008
Treg defect ?
Perceived stress &
neurogenic inflammation:
SP, mast cells

13. gd T-cells and HF biology
• gd T-cell number fluctuates significantly during murine hair cycle
Paus et al. BJD 1994 + JID 1998
• Only extremely few gd T-cells in/around human HFs
Christoph et al. BJD 2000
• gd TC KO mice show catagen retardation, followed by acceleration
of HF cycling Klöpper et al. JID 2013
• FGF9-secreting gd T-cells: key role in wounding-induced HF
neogenesisis Gay/Cotsarelis et al. 2013
• gd T-cells are involved in stress surveilance in murine skin
Hayday, 2009
HYPOTHESIS:
Can excessive stress immunosurveillance activities of
gd T-cells induce HF IP collapse and
thus contribute to AA pathogenesis ?

14. N=15-21 HFs/group from 6-7 AA patients or healthy donors.
Vd1+T-cells infiltrate in /around AA hair bulbs
A
HS = healthy skin
NL = non-lesional AA skin
AA = lesional AA skin
… also seen in
experimentally induced
AA in human skin
xenotransplants in vivo
(Gilhar‘s humanized
AA mouse model)
Youhei Uchida et al
unpublished

15. Vd1+T-cells show increased NKG2D
& IFNg expression within AA lesions
E
Do Vd1+T cells operate as
stress sentinels around human HFs ?
NKG2D IFNg
HS = healthy skin, NL = non-lesional AA skin
AA = lesional AA skin

16. Ki67/TUNEL
staining
• Reduced hair matrix
& outer root sheath
KC proliferation
• Massive outer root
sheath KC apoptosis
• Signs of direct HF
cytotoxicity
incl.
enhanced LDH release
loss of intercellular adhesion
pigmentary abnormalities
(e.g. melanin clumping)
Co-culture of “stressed” scalp HFs with autologous
dermal Vd1+T-cells induces HF dystrophy ex vivo

17. Autologous dermal Vd1+T-cells induce HF IP collapse
and premature catagen in “stressed” HFs
• ↑ MHC class I; ↑ MICA
• ↓ IP guardians: αMSH and TGFß2 = AA phenocopy !
• premature catagen, HF dystrophy
MHC class I αMSH TGFβ2
Hair cycle

18. Autologous dermal Vd1+T-cells become activated and
up-regulate IFNg and NKG2D when co-cultured with
“stressed” human scalp HFs
Isotype control; non-co-cultured Vd1 T cells
Vd1 T cells co-cultured with “stressed”, healthy (non-AA) scalp HFs
n= 3 experiments from 3 donors
None of the above phenomena seen with
„non-stressed“ human scalp HFs
Abrogated by anti- IFNg, CD1d or MICA antibodies

19. Conclusions 1
• New physiological role of human gd T cells (Vd1)
in human skin biology:
 scouting for
distress signals
from "stressed" HFs

20. Conclusions 2
• These physiological stress sentinel functions of dermal Vd1 T cells
can become pathological by inducing IP collapse & promoting AA
AA
response
pattern

21. Youhei Uchida
Marta Bertolini
Jennifer Gherardini
Jérémy Chéret
et al.
Special thanks to:
Amos Gilhar, Haifa
Alfredo Rossi, Rome
Acknowledgements
gd TC study
Kagoshima Univ
Japan
Jeremy Jen

22. • Therapeutically, protecting & restoring HF IP is crucial
& has the widest success chance among AA therapies
 works in all AA pathobiology scenarios, but is not curative
• Antagonizing IFNg/JAK signalling makes perfect sense, but is
not the only sensible strategy & is also not curative
• Identifying the elusive CD8+/MHC class I-presented (auto-)
antigens in AA and restoring peripheral tolerance
remains critical for curative AA therapy
• Yet, this strategy will counteract only one of several IP collapse
pathways leading to the AA response pattern
 understand roles of Tregs, NKTregs, chemokines, AIRE in human AA
target also mast cells, NK cells, excessive gdTCs activities
restore IP guardians (aMSH, VIP, CGRP; antagonize SP)
Multimodal AA therapy needed !

24. T cell "induced-self" MHC class I/peptide complexes may enable "de
novo" tolerance induction to neo-antigens outside of the thymus
Oelert T, Gilhar A, Paus R. Exp Dermatol. 2016 Oct [Epub ahead of print]
HF epithelium can present self-antigens to cognate CD8+ T cells,
incl. hair cycle-dependent immunogenic proteins including HF-specific neo-antigens.
HYPOTHESIS:
IFNg derived from antigen-specific T cells spotting self-peptides may induce
& alter self-antigen presentation ("induced-self").
Initially, this silences autoreactive T cells, including neo-epitope-specific T cells.
Since the thymus cannot recapitulate neo-epitopes evolving in the periphery,
HF-specific induction of MHC molecules that present these neo-epitopes to
self-peptide-reactive CD8+ T cells is a key element of self-tolerance.
Subsequently, however, the local perpetuation and modification of this crosstalk
induces HF IP collapse and AA
 complements thymus-based regulation models of self/non-self-discrimination:
 "induced-self" in HF maintains peripheral self-tolerance in the case of "danger"
 Can backfire, leading to AA