13 Pereira U, Garcia-Le Gal C, Le Gal G et al. ExpDermatol 2010: 19: 796–799.

14 Pereira U, Boulais N, Lebonvallet N et al. BrJ Dermatol 2010: 163: 70–77.

15 Pereira U, Boulais N, Lebonvallet N et al. ExpDermatol 2010: 19: 931–935.

16 Lebonvallet N, Jeanmaire C, Danoux L et al. EurJ Dermatol 2010: 20: 671–684.

17 Gotz C, Pfeiffer R, Tigges J et al. Exp Dermatol2012: 21: 358–363.

18 Zimmermann K, Hein A, Hager U et al. Nat Pro-toc 2009: 4: 174–196.

19 Wellnitz S A, Lesniak D R, Gerling G J et al. JNeurophysiol 2010: 103: 3378–3388.

20 Lebonvallet N, Boulais N, Le Gall C et al. ExpDermatol 2012: 21: 156–158.

Supporting InformationAdditional Supporting Information may be found inthe online version of this article:

Figure S1. Graphic of neuronal activity as measuredby the macropatch technique on the nerve fibres of theskin explant following 10 days of culture. The graphrepresents the continuous trace of record before, duringand after addition of capsaicin. Three phases wasshowed: a phase with a basal activity before addition ofcapsaicin, a phase with an instable signal just after cap-saicin addition and a third phase with activity charac-terised by the apparition of numerous additional spikescompared with the basal activity.

DOI: 10.1111/exd.12301

www.wileyonlinelibrary.com/journal/EXDLetter to the Editor

Origin of Langerhans cells in normal skin and chronic GVHD afterhematopoietic stem-cell transplantation

Rafiq Andani1, Ivan Robertson2, Kelli P. A. MacDonald3, Simon Durrant4, Geoffrey R. Hill3,4 andKiarash Khosrotehrani1

1Experimental Dermatology Laboratory, UQ Centre for Clinical Research, The University of Queensland, Brisbane, QLD, Australia; 2Department of

Dermatology, The Royal Brisbane and Women’s Hospital, Brisbane, QLD, Australia; 3Department of Immunology, QIMR-Berghofer Medical

Research Institute, Brisbane, QLD, Australia; 4Bone Marrow Transplantation Unit, The Royal Brisbane and Women’s Hospital, Brisbane, QLD,

Australia

Correspondence: Kiarash Khosrotehrani, MD, PhD, FACD, The University of Queensland Centre for Clinical Research, Royal Brisbane Hospital

Building 71/918, Herston, QLD 4029, Australia, Tel: +61733466077, Fax: +61733465598, e-mail: k.khosrotehrani@uq.edu.au

Abstract: Chronic graft-versus-host disease (cGVHD) is a

common complication following allogeneic stem-cell

transplantation (SCT). Past studies have implicated the persistence

of host antigen-presenting cells (APCs) in GVHD. Our objective

was to determine the frequency of host Langerhans cells (LCs) in

normal skin post-SCT and ask if their persistence could predict

cGVHD. Biopsies of normal skin from 124 sex-mismatched

T-cell-replete allogenic SCT recipients were taken 100 days post-

transplant. Patients with acute GVHD and those with <9 months

of follow-up were excluded and prospective follow-up information

was collected from remaining 22 patients. CD1a staining and X

and Y chromosome in-situ hybridization were performed to label

LCs and to identify their host or donor origin. At 3 months,

59 � 5% of LCs were host derived. The density of LCs and the

proportion of host-derived LCs were similar between patients that

did or did not develop cGVHD. Most LCs in the skin remained

of host origin 3 months after SCT regardless of cGVHD status.

This finding is in line with the redundant role of LCs in acute

GVHD initiation uncovered in recent experimental models.

Key words: GVHD – hematopoietic stem cell transplantation –

Langerhans cells – antigen presenting cells – chimerism

Accepted for publication 3 December 2013

BackgroundChronic graft-versus-host disease (cGVHD) is a common compli-

cation following allogeneic stem-cell transplantation (SCT) (1). Its

immunological triggers are still under investigation (2,3). Previous

studies have suggested a major role for host-derived antigen-

presenting cells (APCs) in the development of GVHD (4,5).

Although experimental models mostly addressed acute GVHD, a

similar role for cGVHD has also been suspected. Langerhans cells

(LC) from the epidermis have been shown to persist frequently

post-SCT in mice and their depletion has been reported to protect

from GVHD although these findings were contradicted in more

accurate models (5–8).Clinical studies, however, are sparse. Unlike murine studies,

they have reported that the majority of LCs in the skin after SCT

were donor derived (9). Thus, the turnover of LCs after SCT and

their role in the development of cGVHD in humans, as opposed

to aGVHD in mice, is less clear (8).

Questions addressedWhat is the frequency of host-derived LCs in normal skin after

SCT. Can LC persistence be predictive of cGVHD?

Experimental designAll recipients of SCT followed from 2005 to 2009 at the Bone

Marrow Transplantation Unit of the Royal Brisbane’s Hospital, a

tertiary teaching hospital in Australia, were considered. Patients

had thorough and multidisciplinary clinical follow-up with routine

skin biopsy examination of normal and abnormal skin. During

this period, 124 sex-mismatched allogeneic SCT recipients were

approached as approved by the Metro North District and the Uni-

ª 2013 John Wiley & Sons A/S. Published by John Wiley & Sons LtdExperimental Dermatology, 2014, 23, 58–77 75

Letter to the Editor

versity of Queensland Human Research Ethics Committees.

Demographic information, initial diagnosis and conditioning

regimen were recorded. Prospective follow-up information during

the first 2 years post-transplantation was collected regarding the

occurrence of acute or cGVHD. Patients lost from follow-up or

deceased before the end of first 9 months were excluded. Finally,

given the effect of acute GVHD on the frequency of host LCs (9),

patients diagnosed with acute GVHD in any organ before the

100-day skin biopsy were not considered.

From the 124 sex-mismatched allogeneic SCT recipients, 81

consented to the study or were deceased while fulfilling inclusion

criteria. Among these, 42 were excluded because they developed

acute GVHD before the skin biopsy. Seventeen were excluded due

to short follow-up. The final cohort comprised 22 recipients of

allogeneic sex-mismatched SCT with no GVHD in any organ until

a first biopsy on normal skin (Table 1). During follow-up, 6

patients developed cutaneous cGVHD. These were younger and

less likely to have acute myelogenic leukaemia as initial diagnosis

(both non-significant). There were no differences in conditioning,

donor type or gender between those that did or did not develop

cGVHD.

We next obtained archival paraffin-embedded 3–5-mm punch

biopsies from normal skin and stained sections as previously

described with anti-CD1a (clone O10, Dako) and X and Y chro-

mosome FISH (combined Vysis CEP Y (DYZ1) Spectrum Green

and a CEPX (DXZ1) Spectrum Orange, Abbott) sequentially

(10–12). Cells were then retrieved and evaluated for their origin

on the same microscope. A section was analysed if at least 70% of

its nucleated cells had a clear FISH signal. Each CD1a+ cell was

counted and then assessed for its origin as compared to keratino-

cytes (Fig. 1).

ResultsCD1a staining, as opposed to isotype control, identified a density of

8.3 � 1.1 LC/mm of epidermal section. In 74% of the CD1a+ cells,

Table 1. Patient characteristics

Patient number Diagnosis Conditioning Donor type Transplant type Age at transplant Patient gender Graft gender

1 AML Flu/Mel Matched sibling PBSC 58 M F2 AML CY/TBI Mismatched UD PBSC 23 F M3 MDS Flu/Mel/MTX Matched sibling PBSC 58 F M4 AML Flu/Mel Matched UD PBSC 52 F M5 AML Flu/Mel Matched UD PBSC 63 F M6 MDS CY/TBI Matched sibling PBSC 50 F M7 AML Flu/Mel Matched sibling PBSC 61 M F8 Myelofibrosis Flu/Mel Matched sibling PBSC 60 M F9 NHL Flu/Cy Matched UD PBSC 58 M F

10 AML Flu/Mel Matched UD PBSC 56 M F11 AML CY/TBI Matched UD PBSC 32 F M12 AML Flu/Mel Matched sibling PBSC 66 F M13 AML CY/TBI Mismatched UD PBSC 33 F M14 MDS Flu/Mel Matched UD PBSC 54 F M15 AML CY/TBI Matched sibling PBSC 49 F M16 AML CY/TBI Matched UD PBSC 49 F M17 AA CY/ATG/TBI Matched UD BM 18 M F18 MDS Flu/Mel Matched sibling PBSC 58 M F19 AML CY/TBI Matched sibling PBSC 25 F M20 AML Flu/Mel Matched sibling PBSC 30 F M21 NHL Flu/TBI/Zevalin Matched UD PBSC 62 F M22 NHL Flu/CY Matched UD PBSC 50 F M

Sex-mismatch stem-cell transplant recipients fulfilling inclusion and exclusion criteria are detailed here.AML, Acute myeloid leukaemia; MDS, Myelodysplastic syndrome; NHL, Non-Hodgkin lymphoma; Flu, Fludarabine; Mel, Melphalan; Cy, Cyclophosphamide, MTX, Metho-trexate, TBI, Total body irradiation; UD, Volunteer unrelated donor; PBSC, Peripheral blood stem cell; BM, Bone marrow. M, male, F, female.

Figure 1. The host or donor origin of CD1a+ epidermal Langerhans cells (LCs).Paraffin-embedded slides were stained for CD1a (red, red arrow) to identifyposition of LCs in the epidermis (first and second columns). Same slides were thensubjected to X (red) and Y (green) chromosome FISH (third column). Positions ofLCs were retrieved and their origin was determined as compared to host-derivedkeratinocytes. Examples of host- or donor-derived LCs are given in male- orfemale-HSCT recipients. The X and Y chromosome of the identified Langerhans cellare indicated by a red and a green arrow, respectively. All sections werecounterstained with DAPI. First column images at 1009 magnification. Second andthird column at 4009 magnification.

76ª 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Experimental Dermatology, 2014, 23, 58–77

Letter to the Editor

X and/or Y chromosomes could be clearly visualised to determine

their origin (Table S1). Among these, 59 � 5% were host and

41 � 5% were donor-derived at 3 months. The density of LCs was

not different between patients that did or did not develop cGVHD

during follow-up (P = 0.7 Mann–Whitney test). The proportion of

host-derived LCs was also similar between patients that did or did

not develop cGVHD (62 � 10% vs 58 � 7%, P = 1). Finally, none

of the other clinical variables such as conditioning regimen, initial

diagnosis or donor type influenced the origin of LC.

ConclusionHost-derived APCs have been suspected to initiate acute GVHD

but their frequency after allogeneic SCT has been controversial. In

our study, most LCs were of host origin in accordance with previ-

ous studies that evaluated the persistence of host APCs after SCT

in both mice and humans (6,13,14). In mouse models, LCs

remained of host origin for 12 months after SCT (6). In patients,

70–75% of dermal dendritic cells remained of host origin 30 days

post-SCT (13). In hand allograft recipients, the long-term persis-

tence and self-renewal of LCs have been suggested (14). Therefore,

in the absence of precipitating events such as acute GVHD as per

our study, host-derived LCs seem able to persist and self-renew in

the first 100 days.

However, contradicting studies evaluated LCs in SCT recipients

following their migration out of skin explants. Migrating cells,

representing only 20% of total LCs, were mostly of donor origin

(9). This was also true for dermal dendritic cell and macrophage

populations (15). Because our study directly assessed the origin of

CD1a+ cells in the epidermis without in vitro manipulation, it

more directly reflects the proportions of donor- and recipient-

derived LCs. Also, our study excluded patients with any form of

acute GVHD possibly resulting in higher levels of host-derived

LCs. Finally, the effect of conditioning regimen on host-LC levels

was not observed here (9,15). However, reduced conditioning in

these studies included anti-CD52 (Campath�) that might more

directly affect skin leucocyte populations even if it does not

directly target LCs (16).

The importance of persisting host-derived APCs in GVHD was

suggested by studies where removal of LCs could prevent acute

GVHD (4). In the present study, we could not observe any associ-

ation between the frequency of host-derived LCs and cGVHD.

Although our sample size might not be sufficient to definitively

exclude this association, our results indicate that the origin of LCs

is unlikely to be clinically significant as a predictor. In accordance

with our findings, host LCs are dispensable for the occurrence of

acute GVHD (8), as are other professional host APC (5,17).

Indeed, non-hematopoietic host-derived APCs play an essential

role in the occurrence of GVHD, likely making the LCs in isola-

tion a redundant cell population for the initiation of acute GVHD

(5,18). More specifically, chronic GVHD may be initiated by both

host- and donor-derived APCs (19). It has also been proposed

that LCs may not initiate GVHD but could help maintain the

immune response upon strong toll-like receptor activation (20).

Although this possibility cannot be formally ruled out, it is con-

tradicted by recent findings highlighting the tolerogenic functions

of LCs (21).

In conclusion, we here determined that, contrary to previous

reports, most LCs in the epidermis 3 months after SCT are of host

origin. Despite limitations, our results also suggest that the fre-

quency of host-derived LCs is not associated with the occurrence

of cGVHD, in accordance with recent experimental data.

Author contributionsAll authors had full access to all of the data in the study and take

responsibility for the integrity of the data and the accuracy of the

data analysis. Study concept and design: KK and GH. Acquisition

of data: RA, SD and IR. Analysis and interpretation of data: RA,

KK and KM. Drafting of the manuscript: RA, KK, GH, KM and

IR. Critical revision of the manuscript: all authors. Statistical

analysis: RA and KK. Obtained funding: KK and IR. Administra-

tive, technical or material support: KK, GH, IR and SD and Study

supervision: KK.

AcknowledgementThis study was funded by The Australasian College of Dermatology Science

fund. KK was supported by the NHMRC fellowship 1023371.

Conflict of interestFinancial disclosure: None to report.

References1 Higman M A, Vogelsang G B. Br J Haematol

May 2004: 125: 435–454.2 Blazar B R, Murphy W J, Abedi M. Nat Rev

Immunol Jun 2012: 12: 443–458.3 Lee S J. Blood 2005: 105: 4200–4206.4 Shlomchik W D, Couzens M S, Tang C B et al.

Science 1999: 285: 412–415.5 Koyama M, Kuns R D, Olver S D et al. Nat Med

Jan 2012: 18: 135–142.6 Merad M, Hoffmann P, Ranheim E et al. Nat

Med May 2004: 10: 510–517.7 Durakovic N, Bezak K B, Skarica M et al.

J Immunol 2006: 177: 4414–4425.8 Li H, Kaplan D H, Matte-Martone C et al. Blood

2011: 117: 697–707.9 Collin M P, Hart D N, Jackson G H et al. J Exp

Med 2006: 203: 27–33.

10 Parant O, Dubernard G, Challier J C et al. LabInvest Aug 2009: 89: 915–923.

11 Khosrotehrani K, Stroh H, Bianchi D W et al.Biotechniques Feb 2003: 34: 242–244.

12 Johnson K L, Zhen D K, Bianchi D W. Biotech-niques Dec 2000: 29: 1220–1224.

13 Bogunovic M, Ginhoux F, Wagers A et al. J ExpMed 2006: 203: 2627–2638.

14 Kanitakis J, Morelon E, Petruzzo P et al. ExpDermatol Feb 2011: 20: 145–146.

15 Haniffa M, Ginhoux F, Wang X N et al. J ExpMed 2009: 206: 371–385.

16 Jordan M B, McClain K L, Yan X et al. PediatrBlood Cancer Mar 2005: 44: 251–254.

17 Li H, Demetris A J, McNiff J et al. J Immunol2012: 188: 3804–3811.

18 Toubai T, Tawara I, Sun Y et al. Blood 2012:119: 3844–3853.

19 Anderson B E, McNiff J M, Jain D et al. Blood2005: 105: 2227–2234.

20 Bennett C L, Fallah-Arani F, Conlan T et al.Blood 2011: 117: 7063–7069.

21 Shklovskaya E, O’Sullivan B J, Ng L G et al.Proc Natl Acad Sci USA 2011: 108: 18049–18054.

Supporting InformationAdditional Supporting Information may be found inthe online version of this article:Table S1. Patient outcome and frequency of host

derived Langerhans cells.

ª 2013 John Wiley & Sons A/S. Published by John Wiley & Sons LtdExperimental Dermatology, 2014, 23, 58–77 77

Letter to the Editor